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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1ba7309d-8287-489b-95e6-eb46c0b2030e | bottom-up-approaches-for-multi-person-pose | 2112.11834 | null | https://arxiv.org/abs/2112.11834v1 | https://arxiv.org/pdf/2112.11834v1.pdf | Bottom-up approaches for multi-person pose estimation and it's applications: A brief review | Human Pose Estimation (HPE) is one of the fundamental problems in computer vision. It has applications ranging from virtual reality, human behavior analysis, video surveillance, anomaly detection, self-driving to medical assistance. The main objective of HPE is to obtain the person's posture from the given input. Among different paradigms for HPE, one paradigm is called bottom-up multi-person pose estimation. In the bottom-up approach, initially, all the key points of the targets are detected, and later in the optimization stage, the detected key points are associated with the corresponding targets. This review paper discussed the recent advancements in bottom-up approaches for the HPE and listed the possible high-quality datasets used to train the models. Additionally, a discussion of the prominent bottom-up approaches and their quantitative results on the standard performance matrices are given. Finally, the limitations of the existing methods are highlighted, and guidelines of the future research directions are given. | ['Thong Duy Nguyen', 'Milan Kresović'] | 2021-12-22 | null | null | null | null | ['multi-person-pose-estimation'] | ['computer-vision'] | [ 2.28117057e-03 -2.09082127e-01 1.00633673e-01 -2.30220795e-01
-3.20811361e-01 -1.03212364e-01 2.66177177e-01 1.53111964e-01
-5.47964394e-01 5.27869403e-01 9.39520597e-02 3.88999671e-01
-2.14134574e-01 -3.74102384e-01 -3.53912592e-01 -6.02344751e-01
-1.87005818e-01 6.76261723e-01 2.02513203e-01 -4.61935610e-01
2.27667272e-01 6.04654431e-01 -1.72756445e+00 -6.24173321e-02
4.36528862e-01 9.62468445e-01 -2.43312761e-01 8.49135935e-01
4.59414721e-01 2.18197070e-02 -6.58674479e-01 -6.64456546e-01
2.77736485e-01 -5.49547374e-02 -2.10966885e-01 3.46624821e-01
7.11416245e-01 -2.01484755e-01 -2.63016433e-01 9.86878991e-01
8.33148956e-01 4.02841955e-01 6.28915071e-01 -1.53098893e+00
-2.94172037e-02 -2.56431520e-01 -6.86412156e-01 3.70924532e-01
9.65531230e-01 -1.16133668e-01 6.88475728e-01 -1.31217122e+00
6.52410150e-01 1.43738508e+00 9.28461015e-01 6.87354505e-01
-8.24831009e-01 -2.95773625e-01 3.30055773e-01 4.84680951e-01
-1.47945023e+00 -2.03689560e-01 7.50135601e-01 -6.24470294e-01
5.40387273e-01 5.38199902e-01 9.61602211e-01 1.02104557e+00
2.14897260e-01 8.99782360e-01 1.01883078e+00 -4.26244020e-01
-5.38397282e-02 2.94266462e-01 4.84587610e-01 7.97315538e-01
5.21707833e-01 1.48937836e-01 -7.65550435e-01 -3.74600679e-01
6.13074362e-01 7.21842572e-02 -3.26707773e-02 -7.55128980e-01
-1.10550833e+00 6.40909374e-01 2.88445532e-01 -4.99840677e-02
-4.40626651e-01 -3.54893208e-01 4.22590137e-01 -2.14086443e-01
1.16310924e-01 2.51185685e-01 -2.21519679e-01 -3.79267856e-02
-8.30146134e-01 7.75234282e-01 5.68035781e-01 9.45432544e-01
2.80051768e-01 -2.77277440e-01 -4.20186013e-01 7.01264501e-01
4.84753013e-01 5.14852226e-01 7.69308582e-03 -6.38070464e-01
6.40558898e-01 6.05367959e-01 4.22445685e-01 -1.42690587e+00
-8.42196107e-01 -3.26654404e-01 -7.02094972e-01 9.25621018e-02
5.63384771e-01 -3.10568988e-01 -6.15462720e-01 1.30321968e+00
6.71679795e-01 -3.15934457e-02 -3.26304495e-01 1.09969127e+00
9.12003875e-01 3.56035531e-01 -1.00282945e-01 -3.03406745e-01
1.72446895e+00 -8.73014927e-01 -7.51233518e-01 -6.31626368e-01
3.44976559e-02 -5.21553576e-01 5.46329975e-01 4.75754887e-01
-8.76227140e-01 -7.83316731e-01 -7.92314887e-01 2.68208325e-01
-2.39166051e-01 5.45532048e-01 3.99432302e-01 7.49269307e-01
-5.62644422e-01 3.75046164e-01 -7.35398114e-01 -6.22689188e-01
1.23118825e-01 5.02030849e-01 -4.48699564e-01 1.36508450e-01
-1.10624063e+00 1.07064104e+00 1.43795639e-01 5.99712729e-01
-4.56858486e-01 -2.11064011e-01 -8.74640584e-01 -3.79217595e-01
6.17169321e-01 -7.48084128e-01 9.59569514e-01 -2.99213946e-01
-1.17430794e+00 9.15510654e-01 -3.38393152e-01 -3.18856180e-01
9.12853599e-01 -6.57188594e-01 -3.93965274e-01 3.65751386e-02
1.55216023e-01 5.51889002e-01 8.06677580e-01 -1.04683518e+00
-7.71266937e-01 -8.17555964e-01 -1.87006906e-01 5.86897254e-01
1.42365415e-03 1.73682302e-01 -1.03647304e+00 -4.05275822e-01
1.72215432e-01 -1.24735332e+00 -4.76970732e-01 -6.63498044e-02
-7.73587048e-01 -2.20477819e-01 6.72967613e-01 -7.71751225e-01
1.34899271e+00 -2.01647997e+00 5.33992052e-01 5.12485445e-01
2.09342569e-01 1.54328033e-01 3.35390806e-01 4.55702245e-01
1.43363938e-01 -5.89179575e-01 2.40902007e-01 -4.99596506e-01
-1.01916261e-01 -1.48489892e-01 1.36547759e-01 6.76581740e-01
-3.21926475e-01 7.91073382e-01 -5.74879944e-01 -5.78583777e-01
6.45715356e-01 3.92787755e-01 -2.91487157e-01 5.36415987e-02
3.97120535e-01 5.84978580e-01 -3.00498188e-01 6.33661926e-01
5.28488994e-01 -5.96261621e-02 4.86370437e-02 -4.03768688e-01
-3.85799538e-03 -2.63002962e-01 -1.72536790e+00 1.11336231e+00
2.76491344e-01 5.23829281e-01 3.28309797e-02 -6.26965165e-01
7.65684783e-01 2.07493722e-01 6.88087344e-01 -3.39204729e-01
1.69174671e-01 8.60117227e-02 -7.01223239e-02 -4.73683119e-01
7.18403578e-01 3.03801626e-01 -2.32368603e-01 -1.14547104e-01
-2.96361357e-01 3.99441540e-01 2.94324100e-01 -7.15029016e-02
3.80838871e-01 8.87077749e-02 7.91295290e-01 2.06611589e-01
8.63090932e-01 -2.11584866e-02 5.43766320e-01 5.56638300e-01
-6.89939678e-01 6.25496089e-01 1.02045186e-01 -6.73445344e-01
-8.44749391e-01 -9.11703885e-01 -6.27356470e-02 8.70626032e-01
2.94607729e-01 -5.49097776e-01 -9.69924688e-01 -4.78019297e-01
2.19740078e-01 1.74822956e-01 -7.79337883e-01 -1.61672290e-02
-6.68159842e-01 -8.46075356e-01 2.75977045e-01 6.84946060e-01
5.71368158e-01 -9.53496873e-01 -9.44899023e-01 4.41175625e-02
-4.39251870e-01 -1.31600261e+00 -4.42143768e-01 -5.01043677e-01
-8.31336021e-01 -1.20125234e+00 -9.63132262e-01 -5.69575608e-01
8.86441410e-01 3.62057000e-01 6.95432067e-01 -1.08909138e-01
-4.59207088e-01 6.80090785e-01 -5.69611005e-02 -5.71610212e-01
1.59282863e-01 -3.19740891e-01 7.16472030e-01 2.06806988e-01
6.29800856e-01 1.93348736e-01 -6.68075085e-01 5.97742796e-01
6.37926087e-02 -1.14274263e-01 4.49771672e-01 5.89990497e-01
6.74576819e-01 -7.78598785e-02 1.05899066e-01 -5.98378062e-01
5.43055594e-01 1.94309741e-01 -4.68059123e-01 1.60655662e-01
-4.33263600e-01 -4.96654034e-01 2.01265290e-01 -4.84416634e-01
-6.02889597e-01 3.69640678e-01 -1.77690282e-01 -3.81814808e-01
-2.25962803e-01 1.08764909e-01 -2.33373001e-01 -2.44454414e-01
7.97254682e-01 2.09204569e-01 5.64002581e-02 -4.66914654e-01
4.07030210e-02 3.97876531e-01 5.91458023e-01 -1.59271657e-01
7.87799478e-01 4.39888209e-01 2.27068976e-01 -1.25411403e+00
-8.90001059e-01 -8.73319030e-01 -9.15651679e-01 -7.39973187e-01
8.67280066e-01 -5.86515188e-01 -9.40598845e-01 6.58059120e-01
-1.02512133e+00 4.54618633e-01 -2.78422870e-02 4.74666029e-01
-5.59051454e-01 5.31125903e-01 -3.91505480e-01 -1.15699375e+00
-3.31021279e-01 -1.14269257e+00 1.21975219e+00 5.28070509e-01
-4.91901487e-01 -7.86090076e-01 4.46396358e-02 7.37941146e-01
-1.77502483e-01 3.92599106e-01 2.32170656e-01 -6.19238853e-01
-1.72951311e-01 -8.34561229e-01 2.27572009e-01 -1.36974677e-02
-4.83977012e-02 -1.23092242e-01 -6.54493570e-01 -6.67291284e-01
-1.67794392e-01 2.07063407e-01 2.37178743e-01 8.17744613e-01
7.72618830e-01 -1.52152821e-01 -8.18032742e-01 4.62134272e-01
1.03665328e+00 2.05683619e-01 3.97739202e-01 6.29573941e-01
8.60468686e-01 9.47043180e-01 1.03048348e+00 4.22252685e-01
3.96771550e-01 1.22491872e+00 1.80767357e-01 -1.21475291e-02
1.98217779e-01 -1.41254291e-01 2.95247257e-01 3.05468827e-01
-6.61416948e-01 1.05448700e-01 -6.55640483e-01 7.24223778e-02
-2.00928760e+00 -1.11617649e+00 -3.99810880e-01 2.40053225e+00
-1.40270859e-01 1.62722185e-01 7.45354831e-01 3.82708728e-01
1.03050649e+00 3.56132118e-03 -5.64469874e-01 3.71241048e-02
2.20525756e-01 -4.63086516e-01 3.49234939e-01 3.03595752e-01
-1.55597937e+00 6.32570446e-01 7.11068583e+00 2.34863475e-01
-6.70037866e-01 -1.17548622e-01 1.63630918e-01 -6.60615116e-02
8.42600048e-01 -5.25602877e-01 -1.45733142e+00 3.52687150e-01
4.44364429e-01 -4.07513194e-02 1.23000294e-01 1.04151058e+00
3.17737460e-01 -3.56491327e-01 -1.12927687e+00 1.51470351e+00
4.19599712e-01 -7.72031486e-01 -1.65179998e-01 2.31568903e-01
3.06070685e-01 -5.24497509e-01 4.94706668e-02 1.70833021e-01
-6.37627721e-01 -7.08910346e-01 6.16089344e-01 5.69212914e-01
4.20908481e-01 -9.24417198e-01 7.70560682e-01 5.29453397e-01
-1.53665352e+00 -2.32901394e-01 -2.25394145e-01 -1.65678076e-02
3.83406103e-01 2.18226701e-01 -6.05421305e-01 6.13104582e-01
8.22779119e-01 6.35053813e-01 -7.78800547e-01 1.31865299e+00
-1.77771002e-01 -3.98120843e-02 -2.22632244e-01 -1.49840161e-01
-2.71791548e-01 -3.25858027e-01 1.01797283e+00 1.21840668e+00
6.43006936e-02 -1.04034813e-02 5.61850786e-01 2.86741346e-01
5.71043670e-01 1.37351513e-01 -4.39438075e-01 5.22888124e-01
1.87531620e-01 1.34088385e+00 -6.19913816e-01 -2.64239818e-01
-4.15024281e-01 1.13653302e+00 6.02734685e-02 1.81669027e-01
-9.19273496e-01 -3.37670743e-01 6.94714963e-01 3.18579167e-01
-3.03322077e-02 -2.55551845e-01 -3.23430412e-02 -1.05205667e+00
2.49394074e-01 -1.04471576e+00 5.78744054e-01 -6.17227614e-01
-1.05335557e+00 3.53532255e-01 3.67917091e-01 -1.59428465e+00
-4.24257964e-01 -9.86803710e-01 -4.01798934e-01 6.49195731e-01
-6.33085251e-01 -9.86372888e-01 -5.42548478e-01 6.33371651e-01
6.91850185e-01 -2.85926968e-01 7.32521117e-01 4.32002872e-01
-8.95768881e-01 6.60090148e-01 -2.43948892e-01 4.01210755e-01
4.89329726e-01 -1.05516720e+00 4.26751822e-01 1.15411210e+00
1.50695071e-01 8.51188302e-01 1.06512105e+00 -8.14712942e-01
-1.34346783e+00 -7.34203935e-01 8.10120702e-01 -5.95242500e-01
1.06756017e-01 -3.87178302e-01 -5.85787952e-01 6.78771675e-01
-3.68474960e-01 -2.32458748e-02 5.22714138e-01 2.45964423e-01
1.66006699e-01 -7.70470425e-02 -1.10185683e+00 8.09153736e-01
8.57612491e-01 -4.92622629e-02 -6.20955884e-01 3.91004741e-01
-2.27115989e-01 -8.61922383e-01 -5.04432976e-01 2.06248865e-01
9.45664763e-01 -1.10194290e+00 1.42661512e+00 -5.70078552e-01
-1.97066218e-01 -3.40207756e-01 3.52932550e-02 -1.21885455e+00
-6.02273345e-01 -4.18471545e-01 -5.44203162e-01 7.36602783e-01
1.25425056e-01 -4.12020653e-01 1.13520420e+00 7.11057365e-01
1.79419547e-01 -8.19347858e-01 -8.97331774e-01 -6.37351096e-01
-6.44181669e-01 -2.71902591e-01 1.15764618e-01 2.05069229e-01
-4.59604561e-02 4.86266226e-01 -8.26405048e-01 4.28572118e-01
1.01970828e+00 -1.69475853e-01 1.13748896e+00 -1.57430470e+00
-2.74469048e-01 -1.78824246e-01 -8.46362472e-01 -1.15729880e+00
-4.14223194e-01 -2.89539099e-01 -5.80707472e-03 -1.76953769e+00
3.44548672e-01 1.34445176e-01 4.21174616e-02 -1.43765779e-02
-4.81567353e-01 2.19001025e-01 3.26364636e-01 1.61154866e-01
-5.85632086e-01 1.29368141e-01 1.13208330e+00 8.91603716e-03
-3.71141374e-01 7.97014534e-01 -5.55420697e-01 1.00539744e+00
5.53811789e-01 -2.11953580e-01 -1.79100916e-01 2.18461409e-01
4.27974053e-02 -2.95577198e-02 5.93324065e-01 -1.31501257e+00
4.22863305e-01 -7.42611438e-02 8.01663458e-01 -1.03028178e+00
9.45399880e-01 -9.15208220e-01 1.56506762e-01 7.94577181e-01
1.18457772e-01 4.78934020e-01 2.04814468e-02 6.64444029e-01
-4.24816720e-02 -5.89538291e-02 8.67086649e-01 -1.38258025e-01
-8.95258427e-01 4.09941912e-01 -3.11344206e-01 -4.79375795e-02
1.46904480e+00 -6.81299388e-01 9.68756229e-02 -3.59668285e-01
-1.03539753e+00 2.90461034e-01 4.80346178e-04 5.90110958e-01
8.19883525e-01 -1.34617829e+00 -4.91010785e-01 1.58490613e-01
2.55730033e-01 -2.74215043e-01 4.22464997e-01 1.16127706e+00
-2.99783885e-01 6.63654268e-01 -3.77156466e-01 -8.24020863e-01
-2.06439734e+00 6.14794970e-01 5.47411501e-01 -1.76127374e-01
-6.49850845e-01 7.02350140e-01 1.89313311e-02 -3.29106182e-01
4.37689155e-01 1.70793459e-01 -8.30998898e-01 1.55476332e-01
7.48023570e-01 9.40357387e-01 3.59997340e-02 -1.25968850e+00
-9.23774302e-01 9.85485613e-01 9.64981504e-03 -2.57170111e-01
9.84156966e-01 -2.27265343e-01 1.70911744e-01 4.53209668e-01
6.42436326e-01 8.37540329e-02 -1.05253017e+00 -4.89792302e-02
9.42312647e-03 -5.20156920e-01 -4.15396750e-01 -5.15844584e-01
-7.39573002e-01 7.87834585e-01 8.65218878e-01 -6.61088228e-02
9.06885207e-01 -2.67439812e-01 3.30257088e-01 4.17371929e-01
7.56628513e-01 -1.40258944e+00 1.27594769e-01 3.63350034e-01
8.91494513e-01 -1.22553527e+00 3.02866787e-01 -7.35360384e-01
-8.89485061e-01 9.49379802e-01 9.61165607e-01 -8.45056325e-02
5.69373071e-01 -3.12665924e-02 1.33730397e-02 -3.15201849e-01
-8.83951038e-02 -1.78357437e-01 8.72691214e-01 7.32069135e-01
2.08469152e-01 5.59682101e-02 -2.80440331e-01 4.26854074e-01
-2.92782158e-01 -2.62054145e-01 1.69118747e-01 8.61944139e-01
-3.80892396e-01 -8.46029401e-01 -1.09952247e+00 4.68170494e-01
-3.56680930e-01 4.98724848e-01 -5.87071955e-01 8.50171387e-01
1.20365046e-01 1.02251279e+00 -1.78758204e-01 -5.81851840e-01
1.10630786e+00 6.08996898e-02 6.55563951e-01 -5.80501318e-01
-4.89148885e-01 8.36091042e-02 -1.85538772e-02 -5.83624125e-01
-2.74762928e-01 -1.16141093e+00 -9.75677073e-01 -2.31317967e-01
-1.37818843e-01 6.90581426e-02 5.18625319e-01 8.86985600e-01
1.94044337e-01 4.35330659e-01 2.98403174e-01 -1.31617999e+00
-3.48646671e-01 -8.24743330e-01 -4.19254571e-01 4.64586079e-01
3.01058799e-01 -1.12764800e+00 -2.16524482e-01 -3.02831829e-01] | [7.058037281036377, -0.9709742665290833] |
0be430a5-4feb-422d-a509-196e7d3d3964 | a-two-stage-method-for-non-extreme-value-salt | 2206.05520 | null | https://arxiv.org/abs/2206.05520v2 | https://arxiv.org/pdf/2206.05520v2.pdf | A Two-stage Method for Non-extreme Value Salt-and-Pepper Noise Removal | There are several previous methods based on neural network can have great performance in denoising salt and pepper noise. However, those methods are based on a hypothesis that the value of salt and pepper noise is exactly 0 and 255. It is not true in the real world. The result of those methods deviate sharply when the value is different from 0 and 255. To overcome this weakness, our method aims at designing a convolutional neural network to detect the noise pixels in a wider range of value and then a filter is used to modify pixel value to 0, which is beneficial for further filtering. Additionally, another convolutional neural network is used to conduct the denoising and restoration work. | ['Bing Zeng', 'Yike Liu', 'Renwei Yang'] | 2022-06-11 | null | null | null | null | ['salt-and-pepper-noise-removal'] | ['computer-vision'] | [ 2.17246667e-01 -5.81482708e-01 1.75754428e-01 -2.86224514e-01
3.59966755e-02 -2.45931879e-01 2.43294481e-02 -8.87708738e-02
-6.54963791e-01 5.92678249e-01 2.81707317e-01 -3.45670521e-01
2.85395026e-01 -1.31953859e+00 -2.38809466e-01 -1.01038444e+00
2.10176557e-01 -5.98022461e-01 5.71347117e-01 -5.64770162e-01
3.99796844e-01 4.25011158e-01 -1.27653480e+00 2.39423215e-01
5.81073642e-01 8.47951770e-01 3.36833477e-01 6.35254920e-01
-4.91256475e-01 7.48877466e-01 -9.85122442e-01 -6.10549003e-02
6.54757023e-01 -6.07321441e-01 -2.34958559e-01 -2.19027922e-02
-6.18863665e-02 -6.45456433e-01 -3.61207068e-01 1.81242096e+00
7.12827682e-01 4.04120348e-02 2.82405287e-01 -8.51847887e-01
-8.28613043e-01 6.39327109e-01 -7.01463699e-01 2.50509053e-01
-1.07475780e-01 2.01447695e-01 3.35107625e-01 -5.11609972e-01
2.58015305e-01 1.29703903e+00 1.03812671e+00 2.94540495e-01
-6.45193398e-01 -8.50681365e-01 -9.74404141e-02 1.69614092e-01
-1.27406430e+00 -1.08441636e-01 9.07826543e-01 6.40835613e-02
3.45422179e-01 2.16888145e-01 6.13628924e-01 5.28455436e-01
4.15531009e-01 3.66631508e-01 1.25613403e+00 -4.95507866e-01
2.01464683e-01 -2.25979358e-01 -2.08595335e-01 9.39651392e-03
2.94231862e-01 -1.81479186e-01 6.57038912e-02 2.82850862e-01
8.73718023e-01 4.13150072e-01 -5.14524579e-01 3.84464264e-01
-9.88909602e-01 7.65916526e-01 5.88903844e-01 6.18589163e-01
-5.91633976e-01 -3.97630110e-02 5.23169160e-01 5.30271947e-01
-5.52638136e-02 2.33326003e-01 -4.33208406e-01 -3.70433033e-02
-1.02973783e+00 6.98771998e-02 8.32087040e-01 3.43746126e-01
8.09720814e-01 2.51938105e-01 -5.53120561e-02 7.51941979e-01
5.44647276e-01 5.02239048e-01 5.08140802e-01 -8.61875892e-01
3.00809205e-01 5.07157385e-01 1.77790657e-01 -1.14691973e+00
-1.73688039e-01 -2.45907798e-01 -1.02405751e+00 7.51049757e-01
4.32343930e-01 -3.14936966e-01 -1.45738971e+00 1.01012909e+00
1.22788727e-01 2.90210664e-01 3.49654704e-01 1.09360635e+00
1.09418523e+00 9.51899827e-01 8.63969848e-02 -3.04581910e-01
1.27054727e+00 -4.96853828e-01 -1.39663577e+00 -1.42043412e-01
2.75134116e-01 -1.18299544e+00 9.09729540e-01 6.18612468e-01
-6.82951808e-01 -6.85312867e-01 -1.40327322e+00 1.23659700e-01
-5.30056417e-01 -5.27371354e-02 2.81048626e-01 9.84986246e-01
-6.78148150e-01 6.75470471e-01 -4.77697730e-01 -1.99581027e-01
3.48257661e-01 1.16621144e-01 -8.55769143e-02 -4.86526377e-02
-1.56533587e+00 9.82488036e-01 5.65278113e-01 9.45159674e-01
-3.39301169e-01 -1.21228561e-01 -5.53446472e-01 1.99860677e-01
1.55675799e-01 1.37754098e-01 1.10621488e+00 -1.16774166e+00
-1.38691676e+00 3.19460869e-01 1.03560053e-01 -3.73811722e-01
4.98673737e-01 1.12967957e-02 -7.15911865e-01 1.36934416e-02
-2.41889387e-01 1.01621665e-01 6.43150747e-01 -1.27354252e+00
-5.52042842e-01 -1.82637393e-01 -1.17425524e-01 1.12735555e-02
-2.24238738e-01 8.21770206e-02 -4.62741643e-01 -6.91767991e-01
6.49014950e-01 -4.89784122e-01 -4.65150684e-01 1.73268870e-01
-3.15843046e-01 1.35977700e-01 1.14873385e+00 -9.85625744e-01
1.25672436e+00 -2.25535393e+00 -8.01815391e-01 5.40003538e-01
-1.08212680e-01 1.00208271e+00 8.00101906e-02 4.23339903e-01
5.34354942e-03 4.13650811e-01 -1.31564111e-01 4.43261892e-01
-3.47471058e-01 2.47326091e-01 4.22450639e-02 3.61890793e-01
1.55484639e-02 2.22806364e-01 -7.11536646e-01 -4.85152334e-01
3.85583192e-01 6.43662810e-01 -7.26654828e-02 -2.59195939e-02
6.74707964e-02 3.58763449e-02 -3.64484429e-01 6.13627017e-01
1.55347085e+00 3.94349664e-01 1.90582693e-01 -6.11836314e-01
-4.71261859e-01 -9.09842327e-02 -1.68029678e+00 9.20876443e-01
-2.15339623e-02 5.63814163e-01 3.37088138e-01 -9.39473033e-01
1.41357946e+00 3.77926499e-01 2.92350709e-01 -6.99786603e-01
5.80620766e-01 7.63180628e-02 3.20902199e-01 -8.43529284e-01
3.37737590e-01 -2.54835069e-01 5.13910949e-01 -4.31005191e-03
-6.90215170e-01 -2.20181018e-01 2.19323918e-01 -2.81423748e-01
8.96866322e-01 -3.27528678e-02 2.13806376e-01 -5.67259900e-02
6.35491669e-01 -1.35849282e-01 1.08759642e+00 7.56727099e-01
-6.01388276e-01 6.42669141e-01 5.03087759e-01 -5.14998376e-01
-1.23878396e+00 -8.10641587e-01 -2.61923194e-01 5.17723441e-01
3.82188201e-01 -9.60417539e-02 -7.01912284e-01 -2.41194800e-01
-1.04486994e-01 3.32554489e-01 -5.73854566e-01 -9.33832005e-02
-5.29234588e-01 -8.42164218e-01 4.39782888e-01 5.32505333e-01
1.33254707e+00 -1.40194070e+00 -4.92814064e-01 4.12139833e-01
5.86198643e-02 -6.58632755e-01 -1.67972237e-01 2.69374341e-01
-7.91849613e-01 -9.16723549e-01 -5.94949305e-01 -9.55121338e-01
6.27711833e-01 3.64762485e-01 5.33382833e-01 5.92740119e-01
-3.06030642e-02 -5.35854459e-01 -7.36791730e-01 -7.97839999e-01
-5.66904783e-01 -2.67357200e-01 -5.37723005e-01 -1.70063242e-01
6.92999423e-01 -3.93361568e-01 -8.45813096e-01 1.46575123e-01
-1.27211988e+00 -3.30699056e-01 7.25726008e-01 5.88064253e-01
2.86655664e-01 9.06596899e-01 5.39016902e-01 -6.90195382e-01
7.87933648e-01 -2.77413905e-01 -4.97227341e-01 1.35913119e-01
-4.79764163e-01 -8.76529068e-02 8.20983887e-01 -4.75224972e-01
-1.11208451e+00 8.66466239e-02 -6.95695698e-01 3.22380960e-02
-1.95527256e-01 5.72092772e-01 -2.47482106e-01 2.94302534e-02
4.80909556e-01 6.22083470e-02 2.93309182e-01 -6.23652041e-01
-6.20690808e-02 1.12076461e+00 4.70491379e-01 2.75378972e-01
8.80131960e-01 3.72943878e-01 -1.05450869e-01 -7.10674345e-01
-3.79061908e-01 -2.07218409e-01 -3.37099046e-01 -2.34825045e-01
1.02775204e+00 -6.62786365e-01 -7.62889922e-01 8.62175465e-01
-9.91009712e-01 -1.72470219e-03 1.65143579e-01 6.63902044e-01
1.98483482e-01 6.08390689e-01 -8.38687718e-01 -9.86767769e-01
-4.91244882e-01 -1.05188811e+00 1.10273764e-01 9.68862414e-01
2.90637016e-01 -7.44128883e-01 -2.06661567e-01 -2.30576042e-02
8.76572430e-01 1.67200360e-02 4.88156676e-01 -4.12275136e-01
-1.99088931e-01 -3.39998752e-01 -3.80356729e-01 7.85652399e-01
4.28209782e-01 4.32848006e-01 -6.39848948e-01 -3.86413932e-02
2.92048663e-01 1.83166474e-01 9.20916080e-01 4.28644359e-01
1.05229545e+00 -1.62618801e-01 4.43177298e-02 4.91477281e-01
1.82855284e+00 8.19942236e-01 1.55470967e+00 8.07277560e-01
3.01470488e-01 -2.91069355e-02 5.34240246e-01 3.76480550e-01
1.57730468e-02 1.38019398e-01 7.10939765e-01 -5.75605452e-01
-9.41067282e-03 -8.10643137e-02 5.06478213e-02 4.29917961e-01
-2.05742028e-02 -2.21467033e-01 -5.49976945e-01 4.98546094e-01
-1.46066189e+00 -1.06030464e+00 -4.21790153e-01 1.98775017e+00
1.13709319e+00 3.54700983e-01 -4.04283404e-01 6.33424997e-01
1.02647543e+00 1.83026269e-01 -3.85940105e-01 -5.88008881e-01
-2.49442711e-01 3.03494453e-01 8.43163133e-01 2.32402161e-01
-1.17938256e+00 6.45400882e-01 7.03895712e+00 3.64188015e-01
-1.58876610e+00 -3.90062660e-01 3.85249704e-01 5.03734291e-01
-4.38597165e-02 1.47821140e-02 -5.46019435e-01 6.51274979e-01
3.93565118e-01 3.64455938e-01 3.42046201e-01 6.36003494e-01
6.63472056e-01 -3.32753807e-01 1.37499735e-01 7.79135108e-01
-2.40444750e-01 -9.04238641e-01 -1.70996860e-01 -2.82589644e-01
6.46954656e-01 -3.50570232e-01 -2.64810592e-01 1.19246326e-01
3.50286424e-01 -9.63856697e-01 2.59393156e-01 7.74556458e-01
-3.58008444e-02 -6.91172898e-01 1.56902862e+00 1.88408494e-01
-9.72434878e-01 -4.08243425e-02 -7.70332336e-01 -3.16790521e-01
2.77137943e-02 8.80058765e-01 -6.19041622e-01 1.47115916e-01
1.09861696e+00 3.39435130e-01 -3.69866282e-01 1.46650314e+00
-3.73746604e-01 7.55284846e-01 -3.21820855e-01 -8.98428187e-02
2.60011882e-01 -3.17714989e-01 8.74352530e-02 1.33904040e+00
4.53981131e-01 2.68941745e-02 3.95850651e-02 5.36363602e-01
6.97990507e-03 3.33194226e-01 -2.29099795e-01 7.51056075e-02
6.49660707e-01 1.10116172e+00 -8.91838849e-01 -4.80546474e-01
-5.28746545e-01 8.09693277e-01 -4.78931755e-01 5.82399428e-01
-7.32810974e-01 -1.04382610e+00 3.94485712e-01 2.15761755e-02
6.98771715e-01 -1.26676202e-01 -6.20361686e-01 -6.96087539e-01
-1.14541709e-01 -9.17313397e-01 9.38548818e-02 -9.16208625e-01
-1.16983676e+00 2.75338709e-01 -5.27747333e-01 -1.33038008e+00
4.49539542e-01 -6.30762100e-01 -1.07601333e+00 1.22365439e+00
-1.48819304e+00 -6.95861340e-01 -5.45328140e-01 4.72473294e-01
1.97720319e-01 2.04599813e-01 4.66968209e-01 4.58908707e-01
-5.60766459e-01 3.01349968e-01 2.60653496e-01 5.61730564e-01
7.93361485e-01 -9.27276134e-01 1.28671795e-01 1.14466107e+00
-5.84018469e-01 7.11994112e-01 1.02223861e+00 -8.18397403e-01
-1.13910091e+00 -6.91921592e-01 7.31822371e-01 5.49129009e-01
2.18009442e-01 2.20219001e-01 -1.34677529e+00 4.10864383e-01
4.67223167e-01 6.33561984e-02 1.45247161e-01 -4.20601130e-01
-1.09834289e-02 -3.74492824e-01 -1.66459084e+00 6.92416847e-01
2.69272804e-01 -2.16779746e-02 -6.13700092e-01 -6.28866628e-02
3.51179272e-01 -4.12853688e-01 -8.74474049e-01 5.06962478e-01
5.61980724e-01 -1.00211668e+00 7.99717247e-01 -1.34233907e-01
3.42788488e-01 -1.05031776e+00 -7.55704269e-02 -1.13382542e+00
-2.01137736e-01 -2.91066349e-01 6.16396904e-01 1.29220855e+00
2.26992726e-01 -7.19851017e-01 7.11914837e-01 2.28968501e-01
7.95454159e-02 -4.56517935e-01 -5.47549248e-01 -3.64982307e-01
-8.39658827e-02 -1.66453034e-01 8.21261346e-01 8.97460580e-01
-4.11263496e-01 -1.20885208e-01 -4.28521782e-01 3.01704079e-01
2.83860475e-01 -4.11953777e-01 5.71268320e-01 -1.04485810e+00
2.76881039e-01 -3.51671785e-01 -4.48561162e-01 -8.50430369e-01
-5.22174120e-01 -1.65704802e-01 3.32288384e-01 -2.02252054e+00
-2.07790539e-01 -3.04205686e-01 -5.41929424e-01 5.95379531e-01
-3.48011017e-01 5.96421480e-01 2.26943687e-01 1.54843464e-01
2.12113168e-02 2.79944003e-01 1.57843065e+00 -1.14075623e-01
-3.14384699e-01 -2.94119418e-02 -7.61955261e-01 8.46516728e-01
1.11574328e+00 -3.26268017e-01 6.61849184e-03 -2.17436433e-01
1.99699074e-01 -2.71858960e-01 1.30631611e-01 -1.08315277e+00
1.27507940e-01 -3.59712899e-01 8.11134696e-01 -8.80387843e-01
5.68270404e-03 -1.20719492e+00 1.31317571e-01 7.02828288e-01
7.40413088e-03 -3.51220220e-02 -5.57969660e-02 1.52723655e-01
-3.26605201e-01 -8.12597036e-01 9.55278516e-01 -4.34798539e-01
-9.84153390e-01 -7.02412501e-02 -5.68873703e-01 -5.16653061e-01
8.46236348e-01 -7.26840913e-01 -3.73831570e-01 -5.77384710e-01
-5.75468659e-01 -1.65513030e-03 3.63629133e-01 1.33235022e-01
5.71762204e-01 -1.32257462e+00 -6.53098106e-01 4.10611331e-01
-4.27804053e-01 -3.07497233e-01 3.13571304e-01 5.49730778e-01
-1.26187038e+00 -4.66656744e-01 -4.44315493e-01 -2.04909563e-01
-1.30519593e+00 3.48257601e-01 6.37138367e-01 9.65471119e-02
-6.75273955e-01 4.45623726e-01 -6.19401634e-01 -2.34841496e-01
2.24033579e-01 -3.76854300e-01 -6.30231082e-01 -1.01082288e-01
8.69412184e-01 3.37166160e-01 -8.00177455e-02 -4.06212032e-01
-4.44733351e-02 5.95793903e-01 8.40891600e-02 9.79649276e-02
1.21699524e+00 -1.91580746e-02 -4.32836980e-01 9.43520442e-02
1.07611835e+00 2.68531322e-01 -1.27654958e+00 -2.05442265e-01
-3.32285523e-01 -6.84265077e-01 2.38843516e-01 -8.09889972e-01
-1.38946652e+00 5.83086967e-01 1.06959260e+00 5.75837791e-01
1.55630207e+00 -7.00999379e-01 9.70155060e-01 3.17362875e-01
-3.63733083e-01 -1.41592979e+00 -3.28440368e-01 5.45811296e-01
6.35516644e-01 -1.19317579e+00 1.82437778e-01 -3.36996585e-01
-4.82490867e-01 1.60562313e+00 5.90753555e-01 -3.23946893e-01
9.39244509e-01 7.20036089e-01 8.31555605e-01 1.05373465e-01
-7.67008960e-02 -2.88354397e-01 -2.86987543e-01 7.12196052e-01
5.65577149e-01 -2.95157313e-01 -9.12872076e-01 3.23573887e-01
-1.76675707e-01 3.07775617e-01 8.08961630e-01 1.28459370e+00
-1.14163089e+00 -9.68770385e-01 -9.38097358e-01 4.72613305e-01
-8.15174162e-01 -1.88906565e-01 1.13348946e-01 6.37197733e-01
5.11429608e-01 1.28803837e+00 -2.43286435e-02 -3.52951050e-01
4.06194299e-01 -1.75626367e-01 -9.22657102e-02 -1.04501650e-01
-7.60565817e-01 1.88418120e-01 -2.54512638e-01 -1.13597617e-01
-4.18856412e-01 -2.52314448e-01 -1.66192925e+00 -5.23177505e-01
-3.64707232e-01 -7.20411539e-03 7.98443794e-01 9.15165663e-01
-3.68637174e-01 5.53851128e-01 6.91250503e-01 -4.06418800e-01
-4.51022178e-01 -1.16791475e+00 -7.29833066e-01 6.28215134e-01
3.75080615e-01 -1.53573722e-01 -6.08047366e-01 7.15909079e-02] | [11.288561820983887, -2.4656546115875244] |
d095a7c0-b433-4a10-9a48-36f4e2de82de | sibblings-similarity-driven-building-block | 2306.04817 | null | https://arxiv.org/abs/2306.04817v1 | https://arxiv.org/pdf/2306.04817v1.pdf | SiBBlInGS: Similarity-driven Building-Block Inference using Graphs across States | Interpretable methods for extracting meaningful building blocks (BBs) underlying multi-dimensional time series are vital for discovering valuable insights in complex systems. Existing techniques, however, encounter limitations that restrict their applicability to real-world systems, like reliance on orthogonality assumptions, inadequate incorporation of inter- and intra-state variability, and incapability to handle sessions of varying duration. Here, we present a framework for Similarity-driven Building Block Inference using Graphs across States (SiBBlInGS). SiBBlInGS employs a graph-based dictionary learning approach for BB discovery, simultaneously considers both inter- and intra-state relationships in the data, can extract non-orthogonal components, and allows for variations in session counts and duration across states. Additionally, SiBBlInGS allows for cross-state variations in BB structure and per-trial temporal variability, can identify state-specific vs state-invariant BBs, and offers both supervised and data-driven approaches for controlling the level of BB similarity between states. We demonstrate SiBBlInGS on synthetic and real-world data to highlight its ability to provide insights into the underlying mechanisms of complex phenomena and its applicability to data in various fields. | ['Adam S. Charles', 'Gal Mishne', 'Noga Mudrik'] | 2023-06-07 | null | null | null | null | ['dictionary-learning'] | ['methodology'] | [ 1.19133763e-01 -6.91430271e-01 -6.33617878e-01 -1.79073662e-01
-3.71504098e-01 -8.35708499e-01 5.53626955e-01 5.78377843e-01
2.73646504e-01 5.37164748e-01 1.58915684e-01 -5.93927860e-01
-8.05686414e-01 -3.42150509e-01 -4.08295095e-01 -6.57880366e-01
-1.03031814e+00 7.61540756e-02 1.40330583e-01 -2.95876712e-01
1.08436078e-01 7.92081892e-01 -1.30251741e+00 -9.24953073e-02
4.35944051e-01 7.63576388e-01 -1.32128581e-01 4.50159788e-01
1.68460265e-01 4.60788369e-01 -6.58125281e-01 2.79962569e-01
3.45773369e-01 -5.44045925e-01 -1.54168084e-01 1.67992711e-01
7.57992174e-03 1.79540947e-01 -5.86144626e-01 6.12052381e-01
2.09658071e-01 -7.01906234e-02 6.13325417e-01 -1.66515779e+00
-1.42230108e-01 2.58454204e-01 -5.48196375e-01 7.40560055e-01
2.03942582e-01 3.77203852e-01 1.00416481e+00 -1.30513489e-01
3.83866787e-01 1.07796335e+00 7.06575036e-01 2.16371417e-01
-2.09088159e+00 -4.36008453e-01 3.65915835e-01 3.17863747e-02
-1.42138886e+00 -5.86659372e-01 9.80985463e-01 -6.29361212e-01
8.40936661e-01 6.12119496e-01 9.63348627e-01 1.02376533e+00
6.79591179e-01 5.07995546e-01 1.30929732e+00 -6.27045929e-02
6.02027655e-01 -3.83546174e-01 3.98856789e-01 4.07450199e-01
4.36020166e-01 4.99107897e-01 -7.83134937e-01 -6.04766369e-01
7.78092563e-01 7.20192194e-02 -2.26360843e-01 -8.61612320e-01
-1.35979724e+00 5.35480797e-01 -8.02620649e-02 1.16875470e-01
-3.11989486e-01 -3.22783552e-02 5.21697700e-01 4.59771127e-01
1.08360827e-01 6.14028215e-01 -6.33443654e-01 -2.99335837e-01
-8.96402061e-01 1.65422201e-01 9.09317434e-01 8.82821620e-01
5.94894528e-01 4.36900795e-01 -2.80158430e-01 3.56485844e-01
1.59995407e-01 6.56832397e-01 2.22482726e-01 -7.62745798e-01
2.72844940e-01 6.36007011e-01 -4.25371528e-02 -9.86235440e-01
-7.15253055e-01 -4.66638625e-01 -9.71985519e-01 -2.99942076e-01
3.01585495e-01 -7.32741654e-02 -8.86874616e-01 2.13975000e+00
3.71797800e-01 2.65783340e-01 -2.29841098e-01 5.94972730e-01
2.29952976e-01 4.71613050e-01 -5.29838353e-02 -8.38178515e-01
1.23450899e+00 -8.12844262e-02 -8.40675652e-01 -3.17630649e-01
1.64265037e-01 -3.60025376e-01 1.04114282e+00 1.40608534e-01
-7.42838800e-01 -2.10289627e-01 -9.61920619e-01 7.38644421e-01
-1.98135212e-01 -2.91354090e-01 7.95888007e-01 5.06809175e-01
-7.12272525e-01 5.52785218e-01 -1.27500665e+00 -4.34958041e-01
1.44547030e-01 3.83853167e-01 3.85467485e-02 2.57259637e-01
-1.10120976e+00 7.12683022e-01 1.81712985e-01 -1.12870978e-02
-1.08715379e+00 -8.62518191e-01 -8.20333242e-01 1.12988621e-01
4.30745363e-01 -4.44401890e-01 6.64374709e-01 -2.22241804e-01
-1.20506799e+00 6.10832311e-02 -2.53961861e-01 -1.78127065e-01
-1.84682131e-01 4.08670276e-01 -1.15152562e+00 -1.50801897e-01
6.89346939e-02 -2.73491472e-01 9.64381337e-01 -9.94568944e-01
-1.29136834e-02 -4.81737852e-01 -1.66133374e-01 -9.14708674e-02
-2.66238838e-01 -3.55182618e-01 -2.71277755e-01 -6.95287228e-01
4.78155226e-01 -1.06087244e+00 -2.94826537e-01 -2.75484592e-01
-5.81318855e-01 2.72381663e-01 8.20931971e-01 -3.86071593e-01
1.80487359e+00 -2.13424850e+00 4.13425595e-01 5.42822003e-01
2.11653575e-01 3.43953669e-02 -1.87153742e-01 1.25229597e+00
-9.67835356e-03 -9.41728726e-02 -1.98825896e-01 9.73816067e-02
-6.76462278e-02 5.45736849e-01 -3.77128720e-01 5.66621065e-01
2.11118430e-01 5.47937095e-01 -9.22562242e-01 -1.29415795e-01
4.08282429e-01 4.36010100e-02 -1.79871708e-01 1.72964111e-01
-1.17241908e-02 6.02137387e-01 -3.89167517e-01 6.72041893e-01
3.44960332e-01 -4.41108972e-01 6.55247092e-01 -4.51983213e-01
-2.00305521e-01 3.15145999e-01 -1.40171754e+00 1.31452787e+00
-3.23930413e-01 5.22221267e-01 2.40253955e-02 -1.22346961e+00
7.30226696e-01 8.69170874e-02 9.30387199e-01 -8.12700868e-01
-1.72950141e-02 -1.58321932e-01 1.99220791e-01 -4.14670229e-01
6.75995797e-02 -3.63505073e-02 -3.52523804e-01 6.28297150e-01
7.32931942e-02 -1.46937311e-01 4.68759269e-01 3.07821989e-01
1.52607429e+00 -3.40473056e-01 4.98017669e-01 -6.68211997e-01
1.29635453e-01 -1.72378644e-01 7.99747825e-01 4.78892088e-01
-4.21430409e-01 3.35087329e-02 6.94285870e-01 -4.65507179e-01
-9.51519787e-01 -1.45798206e+00 -4.45285052e-01 6.34290099e-01
5.24158418e-01 -7.56172001e-01 2.97870450e-02 -1.22687362e-01
4.51423258e-01 5.66557646e-01 -7.29892671e-01 -6.51641071e-01
-9.33055580e-02 -1.16139317e+00 4.12464172e-01 4.13518012e-01
-1.01118445e-01 -5.39746583e-01 -5.60342491e-01 4.41692740e-01
-8.97990614e-02 -1.25108397e+00 -4.05483484e-01 6.78371191e-01
-1.08321261e+00 -1.14767957e+00 9.76020917e-02 -5.08150645e-02
5.41728735e-01 2.59099722e-01 1.20436764e+00 -4.24731135e-01
-7.21352577e-01 5.84590673e-01 1.18307754e-01 -1.10691570e-01
-3.33147615e-01 -2.62636006e-01 4.56813902e-01 8.03206675e-03
2.35714510e-01 -7.91960239e-01 -5.51066160e-01 5.56774795e-01
-8.38483572e-01 -3.09233606e-01 4.65834171e-01 7.48256326e-01
5.15790522e-01 1.68892980e-01 6.42555773e-01 -4.68515068e-01
9.49323535e-01 -5.99988043e-01 -6.30501032e-01 2.66718388e-01
-1.11165261e+00 2.90762752e-01 3.74445856e-01 -6.54060662e-01
-3.53500634e-01 -1.86314836e-01 5.90945780e-01 -4.23949659e-01
2.02986598e-02 7.04216599e-01 1.54105514e-01 1.98169038e-01
6.59680009e-01 5.90049148e-01 7.37371072e-02 -1.17150933e-01
2.08362088e-01 3.04785192e-01 1.90648541e-01 -8.72966468e-01
7.09748149e-01 4.77666020e-01 2.60047615e-01 -9.78864372e-01
-4.91297364e-01 -6.58317864e-01 -6.64325416e-01 -2.23088533e-01
4.03354019e-01 -9.90469396e-01 -8.89873803e-01 2.49429733e-01
-4.60056931e-01 -3.69440258e-01 -3.89996707e-01 3.06760609e-01
-5.27228534e-01 2.91865140e-01 -5.97557962e-01 -9.37899888e-01
-1.23159871e-01 -7.63403535e-01 9.64667380e-01 -8.42695013e-02
-6.19708538e-01 -1.16389894e+00 4.76217538e-01 -4.34960201e-02
3.77147794e-01 5.30800939e-01 1.12595809e+00 -3.89480978e-01
-3.26951832e-01 -2.54153907e-01 1.51951000e-01 -7.70160332e-02
6.88193142e-01 1.38295814e-01 -4.95381355e-01 -8.61363232e-01
-3.11351061e-01 -1.93699282e-02 1.50966287e-01 4.15970057e-01
8.61040711e-01 -3.21100980e-01 -3.54468673e-01 3.54741544e-01
1.29964459e+00 2.99347132e-01 2.05192268e-01 1.48957416e-01
5.68330050e-01 3.93712044e-01 3.78829628e-01 8.68833601e-01
4.49574649e-01 8.29263389e-01 3.37138832e-01 -7.56649002e-02
2.29396194e-01 -1.21188909e-01 4.69589949e-01 1.07976735e+00
2.64341950e-01 -5.62397055e-02 -8.35041106e-01 8.32110941e-01
-1.74838519e+00 -1.03609395e+00 -2.23608509e-01 2.52896762e+00
7.59594142e-01 3.14123094e-01 4.61172074e-01 2.30640993e-01
5.53032160e-01 3.58271360e-01 -1.09323168e+00 -3.85954410e-01
7.95117617e-02 -5.76347820e-02 6.59113944e-01 8.20744038e-02
-7.60533810e-01 3.99945557e-01 7.42693090e+00 4.21185791e-01
-1.26629221e+00 -2.52431780e-01 1.20308869e-01 -1.43233791e-01
-2.18456402e-01 1.87721074e-01 -5.25010109e-01 5.28260052e-01
1.26696503e+00 -4.67546970e-01 7.31433690e-01 1.40490100e-01
6.45336330e-01 7.70700723e-03 -1.25331080e+00 8.49645436e-01
-2.27895960e-01 -1.02927923e+00 -1.80873513e-01 1.44128650e-01
6.74890637e-01 1.11712739e-01 -1.21854782e-01 1.16354607e-01
5.40163279e-01 -6.10979676e-01 4.36189502e-01 2.32147872e-01
5.61704874e-01 -4.19109404e-01 1.87801063e-01 3.26557726e-01
-1.51182330e+00 -2.54109383e-01 6.47377595e-02 -1.44034714e-01
1.71563029e-01 9.33508873e-01 -5.49811006e-01 5.62143505e-01
5.76181293e-01 7.83713698e-01 -3.91533226e-01 7.07084179e-01
1.97654754e-01 8.60313296e-01 -5.23210287e-01 1.21846259e-01
-6.35395199e-02 -2.96471208e-01 6.96767926e-01 8.60884547e-01
5.14005274e-02 -1.05622932e-01 4.70005304e-01 7.30506837e-01
7.59632349e-01 -4.20020729e-01 -6.40461445e-01 -5.36680520e-01
9.16410685e-01 9.51996028e-01 -8.09249938e-01 -2.20250979e-01
-2.99192637e-01 5.19379497e-01 -3.40776406e-02 5.75822234e-01
-7.41379976e-01 -1.37443781e-01 1.15996146e+00 3.36835235e-01
2.25532174e-01 -8.15706611e-01 -3.02884251e-01 -1.45342314e+00
-1.62920013e-01 -1.22966921e+00 5.94720602e-01 -3.51242155e-01
-1.30501890e+00 3.24000537e-01 3.00546885e-01 -1.60200822e+00
-2.28792310e-01 -1.49010897e-01 -5.54113686e-01 6.71424329e-01
-1.07057440e+00 -8.83332968e-01 -5.77421635e-02 9.36615527e-01
4.59771693e-01 -7.23436102e-03 9.09582019e-01 1.94726363e-01
-1.05203617e+00 1.73423171e-01 4.00702477e-01 -2.17558026e-01
3.91962528e-01 -1.12986279e+00 4.66890872e-01 8.41043413e-01
2.75539786e-01 9.95006323e-01 1.05697441e+00 -6.45205319e-01
-1.95785296e+00 -8.82896602e-01 2.96088338e-01 -5.27502775e-01
1.21514738e+00 -5.68871260e-01 -8.07074904e-01 6.47847414e-01
-2.58005112e-01 -2.69569829e-02 1.12301946e+00 6.28203094e-01
-3.96011800e-01 -3.78660232e-01 -7.79985726e-01 8.24636817e-01
9.54572499e-01 -7.46244907e-01 -2.77018875e-01 1.79680392e-01
9.51990783e-02 -2.44630620e-01 -1.24263442e+00 4.25835192e-01
5.55613935e-01 -6.07928038e-01 9.80620742e-01 -8.99583876e-01
-8.18957761e-02 -5.40850520e-01 -1.86261073e-01 -1.46008193e+00
-5.37051201e-01 -7.12628067e-01 -5.65787792e-01 1.10577524e+00
2.78743058e-01 -5.45320988e-01 4.02986407e-01 7.70222187e-01
1.41368866e-01 -6.07846498e-01 -8.90859485e-01 -1.25292671e+00
-3.77210230e-01 -3.84301752e-01 7.05590904e-01 1.19842899e+00
4.04840767e-01 4.30199772e-01 -5.16965806e-01 5.48645437e-01
8.61317277e-01 5.25776148e-01 8.55411112e-01 -1.24653351e+00
-4.47337389e-01 -2.37033278e-01 -5.95056534e-01 -7.28793621e-01
-2.59300768e-01 -4.87663388e-01 -6.30718097e-02 -1.09348047e+00
9.61632431e-02 -6.17063463e-01 -5.79373896e-01 4.80758041e-01
3.74379312e-03 -7.91576654e-02 4.80586551e-02 4.31880385e-01
-4.49852258e-01 6.81344867e-01 1.00481665e+00 -2.15205267e-01
-5.14394164e-01 9.20187905e-02 -5.95982730e-01 1.95564926e-01
6.59167826e-01 -5.20187736e-01 -8.01026165e-01 6.71001524e-02
-1.09069701e-02 5.34811854e-01 3.45784515e-01 -8.92317414e-01
-5.80445044e-02 -7.09605277e-01 8.48642811e-02 -3.55881989e-01
2.86334425e-01 -9.86733198e-01 6.04955018e-01 6.07094944e-01
-2.73483485e-01 2.72473127e-01 6.33118331e-01 1.12505019e+00
-1.01249143e-01 4.65570033e-01 3.53257596e-01 2.42763907e-01
-6.25046611e-01 3.94208699e-01 -5.33543229e-01 1.31964982e-01
1.14280593e+00 -2.14558169e-01 -2.51241028e-01 -4.27817881e-01
-7.71374047e-01 4.00217384e-01 2.99883872e-01 6.58491850e-01
1.33155584e-01 -1.39430356e+00 -2.01452360e-01 4.39885646e-01
4.61794049e-01 -5.88183820e-01 3.35430682e-01 9.50478554e-01
-1.14913681e-04 3.30196679e-01 -2.86827922e-01 -1.05011511e+00
-1.24020338e+00 6.98710740e-01 1.48097530e-01 -3.44004899e-01
-3.54658961e-01 1.05859660e-01 -2.59374492e-02 -3.30818146e-01
-1.62377626e-01 -3.60366493e-01 1.79020286e-01 1.58453584e-01
1.07929260e-01 3.04681182e-01 1.37896866e-01 -4.85098302e-01
-5.63652873e-01 4.90239382e-01 1.73478961e-01 1.25329703e-01
1.27203047e+00 -3.80719572e-01 2.74912640e-02 1.10513973e+00
9.11541164e-01 -2.21077070e-01 -1.38774240e+00 -1.61147878e-01
2.81578097e-02 -3.88986737e-01 2.06562385e-01 -7.20542192e-01
-9.56984580e-01 5.82611501e-01 6.32080495e-01 7.55355179e-01
1.17421567e+00 -1.95966572e-01 4.27461207e-01 2.29593158e-01
5.87645113e-01 -9.05487299e-01 2.93310476e-03 1.28340766e-01
5.92194259e-01 -9.78758931e-01 3.95862222e-01 -3.32615882e-01
-5.86908698e-01 9.38037336e-01 3.66146415e-01 2.19468653e-01
9.63847280e-01 3.99422348e-01 -2.18387544e-02 -5.78127563e-01
-1.09131849e+00 -1.21670082e-01 2.00308815e-01 4.64792192e-01
1.15574591e-01 2.67904669e-01 -2.37937361e-01 -1.22479843e-02
2.11001694e-01 -2.59503990e-01 4.99954641e-01 1.24801373e+00
2.38944627e-02 -1.10855913e+00 -2.82694429e-01 7.30484307e-01
-1.28662109e-01 3.31316963e-02 -1.18718959e-01 8.18355858e-01
-4.04142618e-01 1.01379704e+00 1.21215276e-01 -5.39603114e-01
6.26345158e-01 -7.20814914e-02 3.54677767e-01 -5.46470225e-01
-3.25135440e-01 2.60847479e-01 5.65306172e-02 -7.72249222e-01
-2.88508326e-01 -9.82363343e-01 -8.65809321e-01 -5.99985719e-01
-4.40175354e-01 1.52338549e-01 6.34417295e-01 8.75275075e-01
9.81706142e-01 8.65319788e-01 9.41153526e-01 -5.37254393e-01
-5.09758949e-01 -6.83045208e-01 -9.65839088e-01 6.01151943e-01
5.47890186e-01 -9.24057662e-01 -5.99340200e-01 1.41862646e-01] | [6.9464111328125, 3.3612661361694336] |
703e9687-a3f2-4000-93ae-beb87bc14984 | ovarnet-towards-open-vocabulary-object | 2301.09506 | null | https://arxiv.org/abs/2301.09506v1 | https://arxiv.org/pdf/2301.09506v1.pdf | OvarNet: Towards Open-vocabulary Object Attribute Recognition | In this paper, we consider the problem of simultaneously detecting objects and inferring their visual attributes in an image, even for those with no manual annotations provided at the training stage, resembling an open-vocabulary scenario. To achieve this goal, we make the following contributions: (i) we start with a naive two-stage approach for open-vocabulary object detection and attribute classification, termed CLIP-Attr. The candidate objects are first proposed with an offline RPN and later classified for semantic category and attributes; (ii) we combine all available datasets and train with a federated strategy to finetune the CLIP model, aligning the visual representation with attributes, additionally, we investigate the efficacy of leveraging freely available online image-caption pairs under weakly supervised learning; (iii) in pursuit of efficiency, we train a Faster-RCNN type model end-to-end with knowledge distillation, that performs class-agnostic object proposals and classification on semantic categories and attributes with classifiers generated from a text encoder; Finally, (iv) we conduct extensive experiments on VAW, MS-COCO, LSA, and OVAD datasets, and show that recognition of semantic category and attributes is complementary for visual scene understanding, i.e., jointly training object detection and attributes prediction largely outperform existing approaches that treat the two tasks independently, demonstrating strong generalization ability to novel attributes and categories. | ['Weidi Xie', 'Jianqi Chen', 'Yan Gao', 'Xu Tang', 'Yao Hu', 'XiaoLong Jiang', 'Keyan Chen'] | 2023-01-23 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chen_OvarNet_Towards_Open-Vocabulary_Object_Attribute_Recognition_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_OvarNet_Towards_Open-Vocabulary_Object_Attribute_Recognition_CVPR_2023_paper.pdf | cvpr-2023-1 | ['open-vocabulary-object-detection', 'open-vocabulary-attribute-detection'] | ['computer-vision', 'computer-vision'] | [ 4.21472728e-01 3.37329805e-01 -3.60325783e-01 -5.79867303e-01
-8.01180363e-01 -8.49169254e-01 7.45367289e-01 2.11230755e-01
-4.79273617e-01 3.52248818e-01 1.70260921e-01 -1.78448468e-01
-1.14939723e-03 -6.23435676e-01 -1.07734036e+00 -4.92786855e-01
-3.74755710e-02 8.15007091e-01 -1.03414478e-02 2.14671925e-01
1.77009869e-02 3.58696848e-01 -2.01058006e+00 5.89419365e-01
5.17364085e-01 1.68916488e+00 2.34257862e-01 5.90941668e-01
-8.60905647e-02 9.73168969e-01 2.75597945e-02 -7.09438741e-01
4.56189305e-01 9.92888212e-02 -1.08891678e+00 4.10521716e-01
8.60895991e-01 -2.31749669e-01 -2.45571032e-01 8.18301857e-01
3.97286713e-01 6.64256886e-03 9.73169684e-01 -1.34827769e+00
-1.06072378e+00 4.53933984e-01 -2.81298488e-01 -1.29385889e-01
1.10536069e-01 4.12666321e-01 1.37213457e+00 -1.22266829e+00
6.71891689e-01 1.32064474e+00 4.28397149e-01 6.40110254e-01
-1.51844704e+00 -5.88566422e-01 6.84193313e-01 3.65979373e-01
-1.60514390e+00 -5.21844149e-01 6.33692861e-01 -5.30374587e-01
8.87926936e-01 8.87126252e-02 5.44120133e-01 1.43812072e+00
-6.81484222e-01 1.15242743e+00 9.47198391e-01 -3.28402370e-01
2.48088047e-01 3.57338876e-01 1.40397727e-01 5.93827784e-01
2.69492827e-02 -4.05883864e-02 -3.03434521e-01 -9.29010566e-03
1.70599878e-01 3.55775803e-02 1.04490751e-02 -9.14539874e-01
-1.45658433e+00 7.67583787e-01 6.63300872e-01 -2.22851887e-01
-4.69628572e-01 5.05540520e-02 5.03483713e-01 2.70956844e-01
6.06328368e-01 3.74391466e-01 -6.41607940e-01 2.80411154e-01
-6.13307536e-01 2.36780763e-01 6.45763814e-01 1.41144776e+00
8.66189301e-01 -3.40664089e-01 -4.26231831e-01 1.02061510e+00
1.68382078e-01 7.39650905e-01 1.35511488e-01 -7.39834368e-01
4.92957592e-01 6.20249391e-01 -9.26901922e-02 -5.53730130e-01
-2.81768918e-01 -4.93450016e-01 -6.85218275e-01 -3.11086457e-02
2.29547113e-01 1.72015950e-01 -1.14220953e+00 1.77919245e+00
4.40963298e-01 1.87101126e-01 2.68834472e-01 9.37384665e-01
9.82766926e-01 4.75167572e-01 6.18117630e-01 1.55436978e-01
1.67005289e+00 -1.06404126e+00 -3.13184738e-01 -5.86008489e-01
3.78257036e-01 -4.41838771e-01 1.11583555e+00 1.39514789e-01
-8.23057115e-01 -7.19643235e-01 -9.08227086e-01 -3.34856838e-01
-6.78573847e-01 4.06989187e-01 6.77387834e-01 1.54277906e-01
-9.02914226e-01 1.60373136e-01 -4.14171278e-01 -5.11823058e-01
1.16808593e+00 2.65849262e-01 -4.81682688e-01 -2.58466005e-01
-7.19735622e-01 7.29985058e-01 6.68542683e-01 -2.70149615e-02
-1.38872337e+00 -7.72782564e-01 -1.03634262e+00 1.07918471e-01
6.80183530e-01 -8.66033196e-01 1.12517881e+00 -1.30025268e+00
-1.05293584e+00 1.30446553e+00 -7.41293579e-02 -5.11260986e-01
4.91894424e-01 -1.68941334e-01 -1.26020864e-01 1.49920285e-01
2.49485940e-01 1.35796130e+00 1.00339699e+00 -1.36999524e+00
-8.16946328e-01 -5.04027724e-01 2.64164358e-01 2.86526978e-01
-4.55699205e-01 -1.89138249e-01 -6.34451926e-01 -5.61232328e-01
1.57534909e-02 -9.78150606e-01 -9.56580415e-02 4.34154570e-01
-6.02606118e-01 -3.93303394e-01 6.36151373e-01 -3.67758542e-01
4.99657333e-01 -2.42824364e+00 2.16379836e-01 9.14267302e-02
2.82489002e-01 3.62964183e-01 -5.34253597e-01 1.78762954e-02
-8.95489529e-02 -7.98521414e-02 -2.30232835e-01 -5.92414081e-01
1.61291525e-01 3.13787788e-01 -7.31453240e-01 3.47316712e-01
6.76318884e-01 1.32264400e+00 -8.36620867e-01 -5.96690834e-01
1.05535991e-01 2.91398197e-01 -5.83193421e-01 4.12383348e-01
-6.42210543e-01 3.15811694e-01 -5.65113425e-01 1.06950176e+00
5.80573797e-01 -5.10831892e-01 4.80009131e-02 -4.46493715e-01
2.48119488e-01 8.06965008e-02 -1.20726466e+00 1.56875658e+00
-5.19987047e-01 5.38951576e-01 -7.12146461e-02 -1.18966854e+00
7.53434420e-01 1.19058691e-01 2.47579336e-01 -6.87862515e-01
8.32823887e-02 4.64280732e-02 -4.21983242e-01 -6.20095849e-01
1.86611891e-01 1.35154262e-01 -6.19719177e-02 2.18377784e-01
6.10747218e-01 4.48629521e-02 1.16390444e-01 2.89062351e-01
7.40652263e-01 3.48930210e-01 2.89302051e-01 1.95789337e-02
5.53782880e-01 7.54463300e-02 2.84636855e-01 9.65062797e-01
-2.38447130e-01 6.37630582e-01 4.93575007e-01 -5.94186485e-01
-1.28912306e+00 -1.04820001e+00 -4.09813404e-01 1.61723650e+00
1.73846260e-01 -1.92670092e-01 -4.95035768e-01 -1.06823015e+00
3.78735036e-01 5.99053323e-01 -9.05947685e-01 2.11029202e-02
-1.23139381e-01 -4.53028023e-01 4.29273814e-01 8.38378608e-01
3.19700688e-01 -1.12882936e+00 -3.23887408e-01 -7.92978927e-02
-2.81308413e-01 -1.51914060e+00 -2.17297256e-01 3.75250071e-01
-4.14935231e-01 -1.22146368e+00 -4.94193375e-01 -9.05374825e-01
7.48690367e-01 2.11573109e-01 1.23873711e+00 -9.61750522e-02
-4.05985415e-01 7.97401071e-01 -4.22818750e-01 -6.27112389e-01
-2.25442365e-01 5.37782609e-02 -9.12019610e-02 4.71054226e-01
4.84969616e-01 -3.32261652e-01 -6.68902040e-01 2.45013863e-01
-6.52240872e-01 1.69632450e-01 9.95089054e-01 7.30041802e-01
1.00880337e+00 -7.14059651e-01 5.42807043e-01 -8.71726990e-01
-1.32278297e-02 -6.56929970e-01 -5.62110364e-01 4.37783390e-01
-5.35776854e-01 5.13099581e-02 5.51302493e-01 -6.33283973e-01
-9.47580278e-01 5.45497954e-01 -1.47752492e-02 -7.25771010e-01
-5.71629167e-01 2.87123583e-02 -3.44625682e-01 8.94225482e-03
6.00119710e-01 3.59241098e-01 7.97817782e-02 -4.14945662e-01
9.42856491e-01 8.17612112e-01 7.92305231e-01 -4.96786177e-01
8.28671038e-01 8.13484132e-01 -1.55775145e-01 -5.15024483e-01
-1.48160076e+00 -6.95123017e-01 -8.61777604e-01 -1.95836332e-02
1.05265677e+00 -1.42927551e+00 -8.53721440e-01 2.16742471e-01
-1.04393709e+00 -2.94118434e-01 -4.54150915e-01 3.02691430e-01
-8.49367797e-01 1.28957136e-02 -2.93685406e-01 -5.65324426e-01
-3.48085403e-01 -9.17022705e-01 1.42495894e+00 -1.21772140e-02
9.03823003e-02 -6.76109552e-01 -3.68876815e-01 6.14892125e-01
1.67690039e-01 -2.08393317e-02 8.83870482e-01 -1.16142476e+00
-8.19178879e-01 -1.65975496e-01 -7.21405566e-01 4.54560757e-01
-2.80907393e-01 -3.82040977e-01 -1.53858471e+00 -2.58552700e-01
-5.96478462e-01 -1.08421648e+00 1.09582078e+00 8.94486308e-02
1.67192578e+00 -4.42306995e-01 -5.02805412e-01 9.69122112e-01
1.34529424e+00 -4.36316997e-01 2.63529152e-01 2.42367476e-01
7.77204275e-01 7.62895167e-01 7.01367140e-01 3.33718151e-01
6.33204877e-01 5.69370687e-01 7.76284218e-01 -1.51494801e-01
-5.00237048e-01 -4.23854232e-01 5.30299619e-02 1.14373997e-01
1.97306991e-01 -2.12511212e-01 -6.44989014e-01 8.23857963e-01
-1.81800437e+00 -7.48467445e-01 2.25910708e-01 1.95980310e+00
7.83351004e-01 -5.33341803e-02 9.70635563e-02 -4.30288821e-01
5.87368548e-01 8.76636878e-02 -8.40969563e-01 1.20184459e-01
-3.60993057e-01 -2.86588538e-02 6.05754972e-01 3.24131967e-03
-1.56380296e+00 1.02782774e+00 5.62440872e+00 8.29867840e-01
-8.59371126e-01 1.48004606e-01 8.76740813e-01 -2.23044693e-01
-9.48774591e-02 3.65582220e-02 -8.88770938e-01 2.70305425e-01
5.38880467e-01 2.91329801e-01 3.62973690e-01 1.16728294e+00
-5.00379443e-01 2.86541313e-01 -1.55185509e+00 9.22386169e-01
2.84193099e-01 -1.31716728e+00 4.19492573e-01 -1.30292028e-01
5.31392097e-01 2.72952110e-01 1.29729167e-01 6.33274019e-01
3.74853015e-01 -9.23223197e-01 9.66836870e-01 3.76268119e-01
1.00677276e+00 -4.14771438e-01 5.53406000e-01 2.26825505e-01
-1.17372024e+00 -5.77553332e-01 -5.70508420e-01 2.12183222e-01
-1.63718000e-01 1.50128588e-01 -8.65757465e-01 4.15894926e-01
8.07143807e-01 8.87811720e-01 -8.96137476e-01 9.31818962e-01
-2.74991870e-01 6.31647050e-01 -1.09518602e-01 1.79883704e-01
2.98777103e-01 9.62897763e-02 3.71928900e-01 1.20010781e+00
-1.46649182e-01 1.20767564e-01 3.82067651e-01 1.07831919e+00
-2.50092268e-01 1.78979654e-02 -4.69494790e-01 1.06550395e-01
6.01497829e-01 1.40400159e+00 -5.91816127e-01 -4.69509125e-01
-6.74908280e-01 8.49319160e-01 6.84233487e-01 4.02526855e-01
-7.37917781e-01 -1.49487898e-01 7.42833078e-01 -1.30711555e-01
7.85272539e-01 2.44798169e-01 -1.19622909e-01 -1.15209782e+00
9.93652269e-02 -7.34789670e-01 6.76834762e-01 -9.41898048e-01
-1.63423920e+00 4.52506870e-01 -6.01284578e-02 -1.17585576e+00
-8.65595713e-02 -9.32110071e-01 -1.47862449e-01 6.23903215e-01
-1.90863609e+00 -1.81162918e+00 -4.72474337e-01 7.43204474e-01
7.34286547e-01 -1.74004346e-01 7.05295026e-01 3.06070566e-01
-3.51879239e-01 6.26396596e-01 -5.62666021e-02 3.56970161e-01
8.56521308e-01 -1.23423398e+00 3.68314534e-01 5.27070642e-01
4.31531698e-01 1.58898503e-01 4.08168644e-01 -4.55893815e-01
-1.44940186e+00 -1.66615593e+00 5.67052007e-01 -7.98652589e-01
6.82700813e-01 -9.14465547e-01 -9.38514948e-01 8.22041392e-01
-6.93242475e-02 7.40011513e-01 5.05019486e-01 2.75423825e-01
-7.39269257e-01 -1.44663960e-01 -8.92602623e-01 3.13393205e-01
1.40340316e+00 -6.77657127e-01 -5.00458658e-01 6.58176720e-01
1.00527871e+00 -2.11920977e-01 -6.51666760e-01 4.39774364e-01
5.01031399e-01 -3.20001245e-01 1.34419882e+00 -1.03035247e+00
3.32736045e-01 -1.93888023e-01 -4.99657601e-01 -1.00830030e+00
-4.19512331e-01 -7.27102682e-02 -1.13037013e-01 1.40354371e+00
4.66463745e-01 -3.49945068e-01 4.96397078e-01 3.63918722e-01
-5.84150106e-02 -7.71248400e-01 -7.77094901e-01 -7.07795978e-01
-1.65094271e-01 -6.15787566e-01 5.35887778e-01 7.91490853e-01
-5.49153924e-01 5.24422228e-01 -2.95512259e-01 3.48941863e-01
7.02055156e-01 4.38946068e-01 8.09910893e-01 -1.25120604e+00
-2.57992595e-01 -3.16275954e-01 -4.78867114e-01 -1.09827733e+00
5.51272094e-01 -1.17192161e+00 1.34073123e-01 -1.29097044e+00
6.69496655e-01 -7.24246264e-01 -4.02095526e-01 8.68326366e-01
-2.10909128e-01 5.20652235e-01 3.05647820e-01 5.21156073e-01
-1.22046340e+00 7.50046134e-01 9.62255836e-01 -4.98638868e-01
2.88905144e-01 -9.64646600e-03 -8.20523262e-01 6.50774837e-01
1.83030471e-01 -3.79863769e-01 -4.01420861e-01 -4.11157250e-01
1.77373856e-01 -2.32209563e-01 9.69464123e-01 -7.58495986e-01
8.29074457e-02 -2.64713597e-02 5.05376756e-01 -5.04407346e-01
2.63227582e-01 -9.19041634e-01 -4.88028556e-01 -4.93747108e-02
-7.52782345e-01 -4.93637413e-01 8.51716772e-02 1.05971038e+00
-8.76290426e-02 2.54174918e-02 7.05084085e-01 1.06380820e-01
-1.23335683e+00 6.35116577e-01 6.45363703e-02 2.54055351e-01
1.26866913e+00 -2.21426278e-01 -4.39933717e-01 -1.61625654e-01
-8.44691694e-01 6.20006144e-01 3.78963232e-01 7.09231377e-01
4.32368129e-01 -1.22184265e+00 -7.67569184e-01 2.47355625e-01
8.94413292e-01 1.25798613e-01 2.42186144e-01 6.40228391e-01
-1.89116746e-02 4.12051857e-01 -4.88230996e-02 -9.23019171e-01
-1.02342415e+00 1.30078387e+00 1.57058209e-01 2.33063012e-01
-5.54578781e-01 1.00206053e+00 6.65485144e-01 -7.33724654e-01
5.71661294e-01 7.67009482e-02 -3.43514472e-01 3.24790418e-01
4.26070660e-01 -1.33150175e-01 1.26995340e-01 -6.91098630e-01
-4.50689077e-01 5.17437398e-01 -8.75009820e-02 3.61989647e-01
1.36188447e+00 -3.59597474e-01 7.52292052e-02 1.96515426e-01
1.20341969e+00 -3.58141661e-01 -1.62186182e+00 -8.25906396e-01
-9.96744633e-02 -3.23759377e-01 -1.68973953e-01 -1.00575387e+00
-9.45678294e-01 9.03045833e-01 8.06359172e-01 -9.49895009e-02
1.10041428e+00 6.77338958e-01 3.97526145e-01 7.01289177e-01
1.06331043e-01 -1.03630817e+00 1.78477004e-01 3.16844344e-01
7.25995600e-01 -1.59259212e+00 -1.11352868e-01 -3.99559438e-01
-1.03024840e+00 8.88980091e-01 7.88500249e-01 1.17203236e-01
4.80386943e-01 -1.37448773e-01 -5.02493456e-02 -1.75390333e-01
-1.05840862e+00 -6.92773998e-01 5.82784474e-01 6.96537495e-01
-4.80961129e-02 7.77603611e-02 3.73693466e-01 5.94880879e-01
1.71420485e-01 -2.01949224e-01 1.24212921e-01 5.58433831e-01
-3.75457168e-01 -6.84573174e-01 -8.56261551e-02 4.79748875e-01
-1.19516835e-01 -2.68907070e-01 -4.03519809e-01 6.25395536e-01
3.98365498e-01 7.52773166e-01 4.67852682e-01 -7.10030422e-02
2.89565891e-01 2.31643319e-01 1.93509787e-01 -6.53410792e-01
-1.75699830e-01 -3.30788940e-01 1.34985507e-01 -7.47167885e-01
-2.99401373e-01 -6.60820723e-01 -7.33563900e-01 3.81758392e-01
-2.57107973e-01 -1.93313181e-01 6.29049659e-01 9.62174773e-01
7.06467867e-01 3.63795757e-01 5.25931597e-01 -7.39331603e-01
-8.44266951e-01 -8.55472326e-01 -2.50681907e-01 8.98412585e-01
5.30974925e-01 -8.28501046e-01 -2.68285036e-01 2.69106388e-01] | [9.999810218811035, 1.7878491878509521] |
50dd84b0-1ce3-445c-8ba8-b118437cac03 | learning-hierarchical-dynamics-with-spatial | null | null | https://dl.acm.org/doi/abs/10.1145/3503161.3548322 | https://dl.acm.org/doi/abs/10.1145/3503161.3548322 | Learning Hierarchical Dynamics with Spatial Adjacency for Image Enhancement | In various real-world image enhancement applications, the degradations are always non-uniform or non-homogeneous and diverse, which challenges most deep networks with fixed parameters during the inference phase. Inspired by the dynamic deep networks that adapt the model structures or parameters conditioned on the inputs, we propose a DCP-guided hierarchical dynamic mechanism for image enhancement to adapt the model parameters and features from local to global as well as to keep spatial adjacency within the region. Specifically, channel-spatial-level, structure-level, and region-level dynamic components are sequentially applied. Channel-spatial-level dynamics obtain channel- and spatial-wise representation variations, and structure-level dynamics enable modeling geometric transformations and augment sampling locations for the varying local features to better describe the structures. In addition, a novel region-level dynamic is proposed to generate spatially continuous masks for dynamic features which capitalizes on the Dark Channel Priors (DCP). The proposed region-level dynamics benefit from exploiting the statistical differences between distorted and undistorted images. Moreover, the DCP-guided region generations are inherently spatial coherent which facilitates capturing local coherence of the images. The proposed method achieves state-of-the-art performance and generates visually pleasing images for multiple enhancement tasks,i.e. , image dehazing, image deraining and low-light image enhancement. The codes are available at https://github.com/DongLiangSXU/HDM. | ['WangMeng Zuo', 'Wenjian Wang', 'Jiaying Liu', 'Wenqi Ren', 'Bin Wang', 'Yudong Liang'] | 2022-08-10 | null | null | null | acmmm-2022-8 | ['image-dehazing', 'rain-removal', 'image-enhancement', 'low-light-image-enhancement'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 2.65327632e-01 -5.25557995e-01 3.51883750e-03 -2.59460509e-01
-3.47196907e-01 -4.21379149e-01 4.26996887e-01 -1.69932976e-01
-2.27416843e-01 6.86316431e-01 3.61421913e-01 2.87686684e-03
-2.28750139e-01 -8.91183615e-01 -6.40050054e-01 -1.31529987e+00
-4.27225754e-02 -7.00723410e-01 3.48355353e-01 -3.27370048e-01
2.68518865e-01 6.50605619e-01 -1.63362408e+00 3.57342362e-01
1.10871649e+00 1.08145499e+00 4.60029572e-01 7.91801333e-01
1.96443945e-01 4.66353655e-01 -3.80186141e-01 -4.70896512e-02
3.19886655e-01 -4.14154530e-01 -1.58892542e-01 3.73900443e-01
2.65100032e-01 -5.25122821e-01 -7.46899724e-01 1.29406595e+00
6.06139183e-01 3.31751615e-01 4.59453374e-01 -1.00186932e+00
-1.10069668e+00 1.18226908e-01 -9.09721434e-01 3.71328533e-01
-1.85313836e-01 6.81467712e-01 4.34438765e-01 -7.92518616e-01
4.78230566e-01 1.20653784e+00 3.86440754e-01 3.02080005e-01
-1.14965391e+00 -7.56802440e-01 3.81331235e-01 6.62989855e-01
-1.49107373e+00 -5.77023983e-01 1.16062629e+00 -1.12597026e-01
4.19855803e-01 2.17304289e-01 6.04938567e-01 8.89657140e-01
4.82993126e-01 5.84578574e-01 1.24371243e+00 -3.84621412e-01
-6.66157529e-03 -7.97678828e-02 -2.24197179e-01 5.51200092e-01
5.74235357e-02 5.19911766e-01 -4.91893291e-01 2.87215084e-01
1.13952959e+00 2.66558435e-02 -7.32055843e-01 -2.06050649e-01
-1.01965809e+00 3.76192629e-01 6.51646316e-01 3.09529811e-01
-5.31212807e-01 1.32877186e-01 1.24215364e-01 -2.79892646e-02
2.98577011e-01 1.58946127e-01 -3.30318272e-01 1.09543115e-01
-8.52688432e-01 8.52041990e-02 -2.39601694e-02 8.93423438e-01
9.66717839e-01 2.94452041e-01 -6.82631791e-01 9.09652531e-01
1.88208058e-01 5.11938632e-01 3.43729049e-01 -1.07174027e+00
2.93677479e-01 2.79773921e-01 2.86523759e-01 -1.26559877e+00
-2.01401189e-01 -8.20299566e-01 -1.52221632e+00 2.80919552e-01
8.77689868e-02 -5.76124080e-02 -1.07058001e+00 1.80102539e+00
3.57349932e-01 2.87430257e-01 -1.96213454e-01 9.18895781e-01
5.26983082e-01 8.56982887e-01 7.86637887e-02 -4.39627320e-01
1.32094991e+00 -8.51634622e-01 -1.04273283e+00 -1.30835464e-02
-8.67100507e-02 -7.80882776e-01 1.12815249e+00 3.91554862e-01
-1.36779583e+00 -9.53676403e-01 -1.09292364e+00 -7.97339976e-02
-1.67977899e-01 2.65449703e-01 2.39371508e-01 4.20574248e-01
-1.24946642e+00 3.82213891e-01 -6.77386582e-01 2.68011928e-01
4.95995224e-01 7.09581003e-02 -1.91434901e-02 -4.40568507e-01
-1.32082975e+00 4.82195109e-01 3.97115409e-01 5.34294546e-01
-9.85777378e-01 -6.85832620e-01 -7.95570076e-01 2.27521300e-01
2.29655743e-01 -6.29374266e-01 6.97773635e-01 -1.06590807e+00
-1.59215415e+00 5.02107501e-01 -3.94765347e-01 -9.28442851e-02
4.82751399e-01 2.46499121e-01 -7.22068250e-01 3.04487586e-01
-2.67908841e-01 5.51326871e-01 1.25564170e+00 -1.55673611e+00
-5.51522851e-01 -1.68147281e-01 -2.98520755e-02 3.76557678e-01
-5.24306655e-01 -9.78898164e-03 -7.14404404e-01 -9.92838323e-01
9.22821015e-02 -4.91225630e-01 -1.27626672e-01 3.88891727e-01
-3.78670871e-01 5.13564348e-01 1.08858788e+00 -1.02675140e+00
1.57857728e+00 -2.37689734e+00 7.78514892e-02 6.72380775e-02
2.72582501e-01 3.07547510e-01 -4.21820760e-01 2.53120773e-02
-6.55116960e-02 2.89375503e-02 -5.43033183e-01 -1.38394788e-01
-1.94152609e-01 5.16827032e-02 -3.73754092e-02 4.22974199e-01
2.18801111e-01 8.47270191e-01 -7.76625991e-01 -4.27554101e-01
5.46730876e-01 9.68482256e-01 -4.19078499e-01 1.54313117e-01
-1.36968400e-02 6.98921263e-01 -2.19894901e-01 5.84178925e-01
1.33231843e+00 -9.50248167e-02 -1.21101484e-01 -6.08060896e-01
-2.91138887e-01 -3.59923124e-01 -1.20110798e+00 1.24113500e+00
-6.64915562e-01 5.94746709e-01 4.21557158e-01 -6.44563437e-01
8.73844862e-01 8.28634650e-02 2.31940776e-01 -9.50474024e-01
1.03482053e-01 4.46431106e-03 1.65964465e-03 -4.21205521e-01
5.37865102e-01 2.67660003e-02 3.01733106e-01 6.83966279e-02
-2.03684136e-01 6.51832968e-02 -3.34362797e-02 -1.00719251e-01
6.19282424e-01 -8.78506824e-02 1.79586217e-01 -1.46933734e-01
8.50265563e-01 -6.62564218e-01 7.94303894e-01 6.26287401e-01
-2.68904150e-01 8.80701125e-01 1.70285553e-01 -2.55037069e-01
-1.07418048e+00 -1.10536909e+00 -2.65495390e-01 7.82659352e-01
7.48499691e-01 1.20831802e-01 -8.25306773e-01 -1.60155937e-01
-4.44867969e-01 7.23697782e-01 -5.80366492e-01 -3.97131324e-01
-5.32901347e-01 -8.18605840e-01 2.06318140e-01 2.35845223e-01
1.21948481e+00 -9.47297573e-01 -2.61206508e-01 3.77685547e-01
-2.89772540e-01 -1.00134480e+00 -8.00046027e-01 -1.22304186e-01
-6.00368202e-01 -7.84988225e-01 -1.05410182e+00 -7.24064410e-01
7.69604981e-01 3.84141982e-01 8.41142058e-01 1.23530865e-01
-3.61614466e-01 -7.04423636e-02 -2.83553362e-01 -2.84648053e-02
-2.49009982e-01 -2.69770056e-01 -3.34510654e-01 5.77589452e-01
-2.32466459e-01 -9.15531576e-01 -1.22465730e+00 4.97657388e-01
-1.36613226e+00 2.55862892e-01 7.21410215e-01 1.16568220e+00
7.21054077e-01 7.26284325e-01 2.64690816e-01 -2.64348358e-01
5.93023300e-01 -2.14299276e-01 -5.40105164e-01 3.20303559e-01
-3.90278399e-01 -1.31162480e-01 6.68051600e-01 -7.42677569e-01
-1.59461117e+00 -2.68948764e-01 -2.66210791e-02 -5.69901884e-01
-2.52838939e-01 1.56735390e-01 -6.76496506e-01 -2.46993244e-01
5.76183021e-01 6.33574963e-01 -2.42102489e-01 -2.57242382e-01
3.61121505e-01 4.92374182e-01 6.83275640e-01 -4.55244362e-01
1.00062943e+00 7.01789975e-01 -9.15588215e-02 -8.21390450e-01
-5.36842883e-01 -2.17784550e-02 -4.84331340e-01 -3.93568486e-01
8.84646237e-01 -8.94568503e-01 -4.09680486e-01 1.24722099e+00
-1.08017850e+00 -6.77241981e-01 -3.20935160e-01 1.20662369e-01
-1.98090792e-01 4.56707358e-01 -7.87737310e-01 -6.22887433e-01
-1.70108944e-01 -1.25762689e+00 9.19413924e-01 7.22223639e-01
6.13184988e-01 -9.26367402e-01 -3.86432558e-01 -1.87467754e-01
8.90169621e-01 1.56054035e-01 9.93073225e-01 4.80638146e-01
-1.06802750e+00 1.42021909e-01 -6.18417382e-01 5.92598081e-01
3.87215346e-01 8.53436068e-02 -8.94450068e-01 -3.79555315e-01
-1.58868842e-02 2.75855124e-01 8.08511317e-01 9.20145869e-01
1.67941070e+00 -4.86744404e-01 -3.09078544e-02 1.10740054e+00
1.38462293e+00 3.48595053e-01 1.12252057e+00 3.60735685e-01
6.40872419e-01 4.46622014e-01 4.99991089e-01 5.91146171e-01
1.38857797e-01 8.28347087e-01 4.09345418e-01 -6.48164630e-01
-5.30607700e-01 -2.47692972e-01 1.57977104e-01 4.85083073e-01
6.11277744e-02 -5.10259807e-01 -4.42306668e-01 5.69653630e-01
-1.35313034e+00 -9.37527716e-01 1.26111418e-01 2.07210231e+00
9.48760688e-01 -1.72104180e-01 -4.17230934e-01 -1.08244076e-01
1.05694270e+00 5.44616580e-01 -8.62380207e-01 2.45454125e-02
-7.28097796e-01 8.64092186e-02 6.46014512e-01 6.12459362e-01
-9.42339778e-01 7.86943018e-01 4.91498232e+00 1.32990599e+00
-1.33671343e+00 2.14996576e-01 1.10590029e+00 -3.90986986e-02
-5.29969215e-01 -1.00158215e-01 -4.82889622e-01 7.81574190e-01
3.01097691e-01 -1.68359950e-01 6.14876330e-01 2.24506915e-01
6.60330415e-01 -1.26133502e-01 -2.64652371e-01 1.10454261e+00
-1.54666170e-01 -1.36634159e+00 7.67141208e-02 1.87921282e-02
1.06638944e+00 -4.27253038e-01 6.08873665e-01 -1.13392733e-01
3.52468975e-02 -8.56872022e-01 7.16146886e-01 8.39061081e-01
1.02678418e+00 -9.00571048e-01 5.12153268e-01 9.14586335e-02
-1.28863895e+00 -4.06618118e-01 -2.89004624e-01 4.58372980e-01
4.83045995e-01 7.25239098e-01 5.37603013e-02 6.04207158e-01
9.34401274e-01 8.29411685e-01 -5.76503813e-01 1.00851929e+00
-4.69064444e-01 4.17142451e-01 1.13828033e-01 5.14739811e-01
5.08586764e-02 -3.52950871e-01 6.08374357e-01 1.19139218e+00
4.94952857e-01 2.65255541e-01 -1.80229947e-01 9.69624758e-01
4.57962677e-02 -3.29576641e-01 -6.68462068e-02 4.00027454e-01
5.74715853e-01 1.18921411e+00 -5.04851341e-01 -2.37593830e-01
-8.04676861e-02 1.15798390e+00 -3.05359989e-01 8.83023560e-01
-9.00549710e-01 -5.68612874e-01 7.48973191e-01 2.88251936e-01
4.36526805e-01 -2.15279430e-01 -2.02177018e-01 -1.04278314e+00
2.54757106e-01 -1.00869238e+00 9.30794701e-02 -1.05951464e+00
-1.19847286e+00 7.10591435e-01 -1.39488187e-02 -1.33160734e+00
3.35157603e-01 -3.56569827e-01 -7.77545989e-01 1.13409710e+00
-1.91209102e+00 -1.14033723e+00 -7.81125247e-01 8.60561669e-01
3.91616762e-01 6.09880239e-02 1.34651765e-01 4.59000826e-01
-6.62799180e-01 6.92087591e-01 3.19345802e-01 -3.91972438e-02
8.74429166e-01 -7.22306848e-01 9.59331766e-02 1.31213963e+00
-4.76910084e-01 5.19242704e-01 5.68215370e-01 -4.91294324e-01
-9.46731627e-01 -1.24765229e+00 1.94007620e-01 2.43236601e-01
3.43684047e-01 -1.85412183e-01 -1.21388888e+00 8.26634392e-02
2.36790106e-01 2.46427655e-01 -1.03865772e-01 -5.20980835e-01
-1.97551385e-01 -4.42422837e-01 -1.29145420e+00 7.64393687e-01
1.02122331e+00 -5.99095762e-01 2.95740366e-01 1.31196335e-01
8.10466468e-01 -5.58905840e-01 -6.46537364e-01 4.38565165e-01
4.43504572e-01 -1.19768846e+00 1.08832526e+00 4.08785269e-02
4.78061110e-01 -7.72925615e-01 -8.45123678e-02 -1.23580635e+00
-5.59413791e-01 -8.51867437e-01 -1.99991703e-01 1.38896263e+00
8.34949464e-02 -7.56896257e-01 5.24645865e-01 4.05722499e-01
-3.33960861e-01 -7.66358376e-01 -7.46142566e-01 -5.51080406e-01
-1.56053528e-01 -1.28120124e-01 7.45075822e-01 8.64999712e-01
-9.32466984e-01 -3.04526120e-01 -6.25401437e-01 6.34582520e-01
7.16541409e-01 -7.34743252e-02 3.68209571e-01 -4.97425288e-01
-4.86857474e-01 -5.43935955e-01 -3.19485635e-01 -1.26339853e+00
-5.79914153e-02 -3.31028283e-01 5.73525392e-02 -1.25145507e+00
1.64888620e-01 -4.70842272e-01 -4.71993148e-01 1.99037790e-01
-4.29030180e-01 3.18408579e-01 7.93549195e-02 1.22906983e-01
-1.98555231e-01 8.94832373e-01 1.74838495e+00 -2.48835415e-01
-4.75434691e-01 -9.25943777e-02 -7.30495751e-01 4.12011176e-01
8.04406226e-01 -8.69160891e-02 -4.41223443e-01 -6.22872889e-01
-1.66945174e-01 -3.92768160e-02 6.62954569e-01 -1.03155839e+00
2.03645200e-01 -3.69886041e-01 7.24577188e-01 -4.67160255e-01
4.19263512e-01 -6.75694346e-01 2.69882560e-01 2.75453150e-01
-3.23749900e-01 -9.61218923e-02 2.93583304e-01 6.36947811e-01
-5.43715656e-01 1.12577237e-01 1.37820864e+00 -2.94453111e-02
-7.52681971e-01 6.93821371e-01 -2.18897939e-01 -8.06542858e-02
8.35816324e-01 -4.69550431e-01 -4.73849535e-01 -6.49561048e-01
-4.69017118e-01 4.45437431e-02 6.53455853e-01 2.95589894e-01
7.99078703e-01 -1.25126529e+00 -7.86719561e-01 4.27254766e-01
-6.46202639e-02 4.51391600e-02 1.37055111e+00 8.48823071e-01
-4.79951441e-01 -1.12580404e-01 -4.31452006e-01 -5.85411131e-01
-1.18025589e+00 5.56732297e-01 6.76139474e-01 -1.39234513e-01
-5.13912439e-01 9.33213234e-01 7.00641692e-01 5.44856302e-02
4.98267300e-02 -1.62399277e-01 -3.56799811e-02 -3.41666728e-01
7.64707208e-01 2.51161873e-01 -1.58014223e-01 -5.95300376e-01
1.42907470e-01 7.05999732e-01 -4.86563779e-02 -5.78205809e-02
1.21214485e+00 -6.79180622e-01 -1.48287967e-01 2.72988174e-02
1.10119271e+00 6.33333102e-02 -1.92070663e+00 -4.80040789e-01
-6.53642654e-01 -9.03398335e-01 4.58847135e-01 -8.68105412e-01
-1.43428338e+00 1.05844390e+00 1.07556975e+00 -1.77887306e-01
2.00628328e+00 -3.35743487e-01 8.01786065e-01 -3.03353518e-01
5.75633310e-02 -1.06504643e+00 2.73965597e-01 3.34958076e-01
1.03917873e+00 -8.18031907e-01 2.59341039e-02 -4.07185495e-01
-5.54240704e-01 8.87567222e-01 6.16589963e-01 1.61013559e-01
5.67137420e-01 2.18226850e-01 2.04414027e-04 1.18906803e-01
-5.25494695e-01 2.00121850e-03 3.70982379e-01 8.90234470e-01
4.00301963e-02 -1.10572450e-01 6.83966577e-02 3.02240878e-01
2.73442566e-01 -2.62334019e-01 4.31714088e-01 4.35997605e-01
-2.97672540e-01 -9.13531005e-01 -4.91361856e-01 1.08201943e-01
-1.53224185e-01 -3.99904400e-01 1.92627266e-01 7.24453986e-01
4.68118042e-01 9.07814980e-01 -4.54359651e-02 -2.54872382e-01
2.80404925e-01 -4.91344690e-01 3.53796273e-01 -7.08079934e-02
-2.84358650e-01 2.76319295e-01 -3.50255728e-01 -4.78849649e-01
-2.60347903e-01 -5.04202724e-01 -8.21529865e-01 -2.57685870e-01
-1.84255466e-01 -2.76222140e-01 4.61079270e-01 4.16307300e-01
6.35029435e-01 7.97265172e-01 9.97502089e-01 -1.15894485e+00
-8.73719901e-02 -8.26455355e-01 -7.55864859e-01 3.33926529e-01
6.33862913e-01 -6.39085174e-01 -5.72895825e-01 1.88547641e-01] | [11.003433227539062, -2.2082693576812744] |
d0022905-961c-4a43-af19-7ab2187d2223 | discover-and-cure-concept-aware-mitigation-of | 2305.00650 | null | https://arxiv.org/abs/2305.00650v2 | https://arxiv.org/pdf/2305.00650v2.pdf | Discover and Cure: Concept-aware Mitigation of Spurious Correlation | Deep neural networks often rely on spurious correlations to make predictions, which hinders generalization beyond training environments. For instance, models that associate cats with bed backgrounds can fail to predict the existence of cats in other environments without beds. Mitigating spurious correlations is crucial in building trustworthy models. However, the existing works lack transparency to offer insights into the mitigation process. In this work, we propose an interpretable framework, Discover and Cure (DISC), to tackle the issue. With human-interpretable concepts, DISC iteratively 1) discovers unstable concepts across different environments as spurious attributes, then 2) intervenes on the training data using the discovered concepts to reduce spurious correlation. Across systematic experiments, DISC provides superior generalization ability and interpretability than the existing approaches. Specifically, it outperforms the state-of-the-art methods on an object recognition task and a skin-lesion classification task by 7.5% and 9.6%, respectively. Additionally, we offer theoretical analysis and guarantees to understand the benefits of models trained by DISC. Code and data are available at https://github.com/Wuyxin/DISC. | ['James Zou', 'Linjun Zhang', 'Mert Yuksekgonul', 'Shirley Wu'] | 2023-05-01 | null | null | null | null | ['object-recognition', 'skin-lesion-classification'] | ['computer-vision', 'medical'] | [ 1.39091656e-01 2.32011974e-01 -2.80967325e-01 -6.28123343e-01
7.01270590e-04 -3.55605066e-01 2.28831172e-01 1.31241187e-01
-6.65452033e-02 8.12477052e-01 -1.59310639e-01 -4.43224460e-01
-5.03179848e-01 -6.36349261e-01 -8.41592312e-01 -7.14639425e-01
-1.85302541e-01 2.82601029e-01 -7.83264041e-02 1.09801121e-01
4.84963544e-02 1.19437084e-01 -1.64789164e+00 4.74844187e-01
1.16137683e+00 9.85979736e-01 -7.90170729e-02 8.77717435e-02
3.70116867e-02 1.03255475e+00 -5.24583399e-01 -5.56833386e-01
2.09954068e-01 -3.60384315e-01 -5.88461280e-01 -1.03096098e-01
3.31777871e-01 -6.34916782e-01 -1.13869295e-01 1.05202854e+00
3.03313024e-02 -4.03076977e-01 6.33855760e-01 -1.78815353e+00
-1.25528169e+00 9.69274342e-01 -5.35285294e-01 1.06949113e-01
-2.55145758e-01 2.53296643e-01 9.59216475e-01 -6.65549994e-01
5.84876165e-02 1.10821283e+00 8.61709416e-01 9.27214921e-01
-1.18039382e+00 -8.92171979e-01 3.82548720e-01 4.65346813e-01
-1.32235360e+00 -2.05529526e-01 4.58437711e-01 -4.46904331e-01
4.95671511e-01 6.08910620e-01 5.95981658e-01 1.50064349e+00
-4.74179238e-02 7.70992160e-01 1.07926810e+00 -1.12228915e-01
3.31247002e-01 3.31485599e-01 3.04370224e-01 5.39582670e-01
7.64692307e-01 2.36096531e-01 -5.65120041e-01 1.08368546e-01
5.05835056e-01 2.67986536e-01 -1.25921875e-01 -4.85646129e-02
-9.52442884e-01 6.87312365e-01 6.58847690e-01 -3.47177647e-02
-2.80379862e-01 1.52696744e-01 1.68919519e-01 4.28173095e-02
6.02946997e-01 6.34101152e-01 -4.25758570e-01 1.49479762e-01
-4.58010137e-01 1.16833627e-01 6.20238483e-01 1.03084540e+00
3.93971294e-01 -5.32025546e-02 5.64304693e-03 7.74391592e-01
1.85894325e-01 1.93030834e-01 3.20568711e-01 -7.42177010e-01
1.85582370e-01 8.77688527e-01 -8.50568637e-02 -9.35742378e-01
-3.15152973e-01 -7.35779762e-01 -9.86459315e-01 1.35268793e-01
2.55086839e-01 -3.43889277e-03 -9.92187262e-01 1.60729647e+00
2.50982583e-01 4.88992184e-01 -1.68501306e-02 1.01933300e+00
9.47352231e-01 2.62471028e-02 3.38587224e-01 3.18839073e-01
1.27057540e+00 -1.02199435e+00 -6.72374249e-01 -4.64984387e-01
7.10469007e-01 -2.72492230e-01 1.05059159e+00 4.84664708e-01
-7.79004037e-01 -3.85546356e-01 -1.27317989e+00 1.15543805e-01
-4.35040832e-01 2.83112556e-01 1.03218782e+00 6.62421942e-01
-9.60671186e-01 6.93871856e-01 -1.04319251e+00 -3.34187120e-01
9.05787587e-01 3.82994801e-01 -2.38624468e-01 -1.15645453e-01
-1.04611039e+00 8.30966771e-01 3.97779584e-01 3.62164289e-01
-8.11734140e-01 -9.96469617e-01 -4.90106821e-01 7.15940520e-02
3.63420367e-01 -6.14440918e-01 1.08411062e+00 -1.11266947e+00
-8.43947887e-01 8.26360464e-01 -1.27452582e-01 -7.53507793e-01
7.87486970e-01 -5.63149750e-01 -3.06379706e-01 -6.71539530e-02
4.15958092e-02 6.84954762e-01 4.59454834e-01 -1.39277947e+00
-6.42639637e-01 -3.48708063e-01 5.06897457e-02 -8.15998316e-02
-5.17109573e-01 -1.74054042e-01 -9.22442079e-02 -4.68866706e-01
2.88689017e-01 -8.39946330e-01 -2.00967312e-01 4.92032468e-01
-6.52736306e-01 -1.93495691e-01 8.85383725e-01 -5.11426866e-01
1.07018161e+00 -2.12289262e+00 -3.25729162e-01 3.29689542e-03
8.96171033e-01 3.04400563e-01 -1.05372846e-01 5.09769656e-02
-2.48806089e-01 5.16795218e-01 -8.57048035e-02 -4.30095285e-01
-1.72043499e-02 5.23552656e-01 -3.93319070e-01 4.57642645e-01
7.30012357e-01 6.53784394e-01 -8.39806736e-01 -3.25966895e-01
2.45465517e-01 4.08650160e-01 -6.12098217e-01 1.35110859e-02
-1.14708275e-01 3.93092006e-01 -5.45094848e-01 8.32145631e-01
8.01205873e-01 -5.05569637e-01 2.74837375e-01 4.17743437e-02
1.82705224e-01 2.84362823e-01 -8.85247946e-01 9.17718887e-01
-2.09396571e-01 7.70991921e-01 -3.82222235e-01 -1.20399249e+00
9.97881353e-01 6.96536303e-02 3.32951605e-01 -5.42623103e-01
3.41623276e-02 2.62679309e-01 2.37672910e-01 -6.58475399e-01
2.27086306e-01 1.33428484e-01 3.09670359e-01 1.52737334e-01
-2.07918808e-01 5.23890316e-01 -1.24477141e-01 5.37455268e-02
1.21704662e+00 4.23692912e-03 2.57475555e-01 -3.17195356e-01
8.62524360e-02 -1.68841600e-01 7.38344371e-01 9.35232282e-01
-4.03449208e-01 4.73690242e-01 6.20781481e-01 -7.68509746e-01
-8.68648231e-01 -1.27995002e+00 -4.92424637e-01 8.39075506e-01
4.36165124e-01 -3.21595520e-01 -7.46079087e-01 -6.72886014e-01
2.92485088e-01 9.61767554e-01 -1.15065467e+00 -5.52573025e-01
-5.30522503e-02 -9.08441126e-01 5.95059574e-01 1.00173211e+00
5.30768394e-01 -7.92402446e-01 -5.46277344e-01 -5.43930335e-03
-3.83610040e-01 -9.31494951e-01 2.80259520e-01 2.88285255e-01
-8.17498565e-01 -1.04927075e+00 4.32520621e-02 -4.32536602e-01
9.25202191e-01 2.21854031e-01 1.07303250e+00 8.77253473e-01
-3.37436110e-01 -2.18461901e-01 -3.94750804e-01 -7.39989042e-01
-3.97328973e-01 -6.05019368e-02 2.21016124e-01 -7.26885274e-02
8.83818507e-01 -6.82459474e-01 -7.44176269e-01 6.82115018e-01
-8.63670170e-01 4.19631749e-01 7.48404622e-01 9.39014196e-01
4.15943086e-01 4.36468869e-02 8.32991540e-01 -8.68718445e-01
2.99178243e-01 -8.99117947e-01 -4.06311095e-01 2.54690289e-01
-8.19898009e-01 7.72234946e-02 6.79011881e-01 -6.75824702e-01
-7.02283859e-01 -5.25486805e-02 1.29314929e-01 -4.94212717e-01
-4.33997214e-01 3.64691764e-01 3.54035795e-02 2.07485110e-01
9.81641650e-01 8.08046535e-02 -2.50824034e-01 -5.65963984e-01
-1.51245622e-02 7.26585329e-01 4.21606719e-01 -5.71040630e-01
6.80032372e-01 5.12693405e-01 -1.65836230e-01 -3.37915927e-01
-1.18478692e+00 -2.11687595e-01 -6.37514114e-01 -2.85053939e-01
6.22981191e-01 -9.91271734e-01 -7.47702479e-01 2.26086274e-01
-1.01090002e+00 -3.31953913e-01 5.36095910e-02 4.50812876e-01
-3.15516412e-01 -5.08032665e-02 -4.75972027e-01 -1.02337003e+00
-1.32813066e-01 -8.59365463e-01 8.50508809e-01 3.17447215e-01
-5.41935325e-01 -9.14187789e-01 -5.89489937e-01 7.01122344e-01
2.47594431e-01 4.44921076e-01 7.84379005e-01 -8.62951279e-01
-7.41881013e-01 -1.58003047e-01 -3.66879970e-01 3.75400633e-01
2.62468636e-01 1.96434781e-01 -1.25050044e+00 -2.59468309e-03
-6.58694729e-02 -3.62344235e-01 9.22488511e-01 3.26004684e-01
1.83812106e+00 -6.22088671e-01 -5.74112952e-01 6.76909387e-01
1.12571967e+00 1.52725250e-01 7.45708466e-01 3.89061391e-01
5.75037181e-01 7.65669107e-01 5.10451078e-01 5.02066255e-01
3.71497840e-01 2.60715574e-01 7.68909693e-01 -1.73091128e-01
7.26175755e-02 -3.65274072e-01 3.33649572e-03 2.79808074e-01
-3.12857300e-01 -2.63732404e-01 -1.12528408e+00 7.06411123e-01
-2.03770351e+00 -9.38667655e-01 -4.04227644e-01 1.91934967e+00
7.15217650e-01 1.82716101e-01 -1.46400094e-01 1.81816518e-01
6.05461597e-01 -3.14267814e-01 -8.89280498e-01 -8.19336846e-02
-1.12400189e-01 -4.51264903e-02 4.55392152e-01 1.42212091e-02
-1.09138465e+00 6.70399427e-01 5.75554180e+00 6.03656828e-01
-1.02927148e+00 -2.11294182e-03 1.09214878e+00 -2.08805725e-01
-2.11559176e-01 -1.55896336e-01 -6.27412081e-01 5.79594851e-01
7.46523380e-01 -9.47531015e-02 3.49287629e-01 1.27697098e+00
1.63385063e-01 -3.31816971e-02 -1.47107518e+00 5.36172926e-01
4.20193095e-03 -1.17101073e+00 -9.03351679e-02 1.99300461e-02
6.70534074e-01 -2.71059591e-02 4.31174666e-01 1.50324106e-01
4.84520048e-01 -1.46812999e+00 8.20235074e-01 4.26513076e-01
4.86368090e-01 -5.75685918e-01 8.55956495e-01 3.53437245e-01
-7.55060256e-01 -2.43404642e-01 -4.41878885e-01 -2.63405204e-01
-5.51732838e-01 6.52758777e-01 -1.14782667e+00 4.56274509e-01
1.14226580e+00 8.21586311e-01 -8.61809552e-01 1.08731198e+00
-4.18494672e-01 1.03479826e+00 -3.43848914e-01 -2.26374269e-01
1.06525257e-01 4.24764790e-02 2.96991855e-01 8.05111289e-01
2.40438968e-01 2.08104905e-02 -1.70596823e-01 1.52087855e+00
4.47919592e-02 -3.31063211e-01 -4.67014462e-01 -8.57122168e-02
5.21953940e-01 1.11185527e+00 -5.67333937e-01 -2.06714883e-01
-1.75582841e-01 7.47545958e-01 4.70115483e-01 3.36484373e-01
-1.20546496e+00 -4.89825103e-03 9.12828267e-01 2.67301738e-01
2.13766359e-02 1.42643154e-01 -8.07968199e-01 -9.40175176e-01
3.35344136e-01 -7.22905338e-01 3.77321452e-01 -8.39741707e-01
-1.63580525e+00 5.84517837e-01 -9.36098248e-02 -1.38662970e+00
1.41859993e-01 -7.99871027e-01 -6.28811240e-01 7.41294324e-01
-1.39519477e+00 -1.09118271e+00 -6.37301803e-01 2.34191954e-01
3.14273328e-01 6.40691966e-02 7.52892435e-01 2.07893834e-01
-7.35953212e-01 8.79899204e-01 1.78870350e-01 2.21599147e-01
5.98494172e-01 -1.23494601e+00 3.15036386e-01 7.57665575e-01
-8.00819099e-02 8.04205716e-01 7.76880980e-01 -5.95217466e-01
-8.55406106e-01 -1.38151252e+00 6.18985474e-01 -7.84995496e-01
7.35973001e-01 -4.56941307e-01 -1.25246227e+00 8.04646790e-01
-1.45099387e-01 6.51863292e-02 9.80989754e-01 4.70523655e-01
-6.13290608e-01 -1.98296860e-01 -1.15451300e+00 8.64587963e-01
1.31926095e+00 -6.28434867e-02 -3.63917351e-01 5.45198619e-01
7.54269183e-01 -3.52377564e-01 -6.75017536e-01 4.73084301e-01
6.24091566e-01 -1.06844795e+00 8.45711112e-01 -1.02685654e+00
1.03680336e+00 -2.70462930e-01 -5.84039949e-02 -1.21412385e+00
-3.60739470e-01 -1.17640093e-01 -4.97853234e-02 1.02105212e+00
5.98543525e-01 -9.19757247e-01 8.81801963e-01 8.78119111e-01
-1.20401137e-01 -1.07553005e+00 -7.36358225e-01 -8.92357767e-01
1.07713692e-01 -4.70801622e-01 8.38644028e-01 1.14749372e+00
-1.04193814e-01 2.36793365e-02 -4.90665227e-01 5.57181537e-01
8.18613887e-01 4.77851089e-03 4.71491694e-01 -1.27618194e+00
-2.38416195e-01 -3.31965029e-01 -5.96406102e-01 -5.50739527e-01
3.16102840e-02 -6.60834968e-01 -4.76785675e-02 -1.27811944e+00
3.38120699e-01 -6.52501047e-01 -3.95794302e-01 9.93575454e-01
-5.31514406e-01 8.77429098e-02 5.79982065e-02 2.25237370e-01
-5.62469602e-01 4.63841498e-01 1.01237810e+00 -1.10028990e-01
1.53205171e-01 1.47179008e-01 -1.21545637e+00 7.53407359e-01
1.33862221e+00 -7.30458021e-01 -4.97571081e-01 -5.61835885e-01
6.10082224e-02 -5.01467884e-01 1.05082774e+00 -1.11039829e+00
-3.83396680e-03 -4.38592017e-01 7.41145968e-01 -4.41812038e-01
2.97837853e-01 -7.70160735e-01 2.41443500e-01 5.98472357e-01
-5.16261399e-01 -2.23194897e-01 3.70721847e-01 6.01700842e-01
7.69902915e-02 -1.17244706e-01 7.09491014e-01 9.81052220e-02
-4.75323439e-01 2.54426599e-01 -9.58178043e-02 -7.88918287e-02
1.09388196e+00 -2.18779728e-01 -7.58419096e-01 -1.84888154e-01
-5.89444399e-01 4.67704415e-01 2.20553190e-01 6.48791969e-01
5.89413881e-01 -1.24109471e+00 -7.45699525e-01 2.11296365e-01
3.74994636e-01 1.20168298e-01 4.57929134e-01 8.40990961e-01
-5.33680797e-01 3.44766110e-01 -2.35293835e-01 -8.19216967e-01
-1.28470302e+00 4.48222578e-01 5.53265512e-01 9.72495079e-02
-4.69804645e-01 8.81526113e-01 4.81266886e-01 -3.85121047e-01
3.25019538e-01 -5.38167655e-01 3.33827026e-02 -3.93145651e-01
6.60681307e-01 7.41965026e-02 -3.66400578e-03 -8.95729363e-02
-3.98455471e-01 -9.26995724e-02 -3.23572934e-01 4.92309630e-01
1.49002111e+00 1.35905057e-01 1.33534637e-03 3.62127721e-01
7.93195426e-01 -5.13503969e-01 -1.40256059e+00 -1.01599455e-01
2.14126520e-02 -5.18695116e-01 -1.92696139e-01 -1.23028648e+00
-1.06570029e+00 9.25039828e-01 6.19298458e-01 2.85181999e-01
1.23060012e+00 6.50196522e-02 2.82106578e-01 4.69624877e-01
7.27751106e-02 -8.85461330e-01 3.50652821e-02 -2.31904052e-02
1.01310480e+00 -1.38072431e+00 -1.71563160e-02 -6.68600976e-01
-7.65716136e-01 8.92043352e-01 1.15830112e+00 8.88942704e-02
4.84195292e-01 2.79343784e-01 4.03425246e-02 -1.28890932e-01
-1.08282018e+00 6.06033392e-02 2.72301078e-01 7.15042233e-01
2.03572735e-01 3.80784005e-01 2.08963919e-02 1.04737556e+00
-3.52704525e-01 9.99662355e-02 4.55167413e-01 5.52109003e-01
-1.45872638e-01 -6.51366532e-01 -3.19437534e-01 7.72100270e-01
-2.52460361e-01 -2.69311041e-01 -7.07055271e-01 8.42516005e-01
4.18262154e-01 9.86589730e-01 1.99578062e-01 -6.84722662e-01
1.56769127e-01 -1.60089850e-01 7.71806017e-02 -4.56521958e-01
-3.23407561e-01 -5.01120269e-01 9.11378339e-02 -4.96976078e-01
-4.07511353e-01 -5.34792542e-01 -1.06063414e+00 -6.16439462e-01
-3.34793836e-01 -1.83343470e-01 3.30969542e-01 8.21289420e-01
4.83663440e-01 6.80978477e-01 4.70292956e-01 -1.05796665e-01
-6.76975250e-01 -8.61472964e-01 -4.24013764e-01 5.76769948e-01
3.36248368e-01 -9.95336235e-01 -6.13638580e-01 1.48297936e-01] | [8.9321928024292, 5.453731060028076] |
e6739af3-61af-48d7-bb7b-fde6a0150ea2 | de-pacrr-exploring-layers-inside-the-pacrr | 1706.08746 | null | http://arxiv.org/abs/1706.08746v2 | http://arxiv.org/pdf/1706.08746v2.pdf | DE-PACRR: Exploring Layers Inside the PACRR Model | Recent neural IR models have demonstrated deep learning's utility in ad-hoc
information retrieval. However, deep models have a reputation for being black
boxes, and the roles of a neural IR model's components may not be obvious at
first glance. In this work, we attempt to shed light on the inner workings of a
recently proposed neural IR model, namely the PACRR model, by visualizing the
output of intermediate layers and by investigating the relationship between
intermediate weights and the ultimate relevance score produced. We highlight
several insights, hoping that such insights will be generally applicable. | ['Kai Hui', 'Andrew Yates'] | 2017-06-27 | null | null | null | null | ['ad-hoc-information-retrieval'] | ['natural-language-processing'] | [ 5.08248173e-02 3.02309453e-01 -3.66739810e-01 -3.49682480e-01
-6.33766472e-01 -5.02600849e-01 7.26314604e-01 1.88035518e-01
-3.45767617e-01 2.17716962e-01 5.06495893e-01 -5.63554168e-01
-5.07773161e-01 -5.08082569e-01 -4.11498070e-01 -5.31794846e-01
-2.91359484e-01 2.83534646e-01 -1.90569550e-01 -6.93833709e-01
3.69943380e-01 4.16871935e-01 -1.46915638e+00 5.42327344e-01
3.78072888e-01 1.17571080e+00 -2.07812056e-01 4.05030280e-01
6.79746121e-02 1.13123202e+00 -7.86849737e-01 -5.17053068e-01
1.24172606e-01 -1.56328037e-01 -1.06862307e+00 -7.80966818e-01
3.17675859e-01 -3.70070964e-01 -3.59855920e-01 8.05004597e-01
4.06341106e-01 2.14135364e-01 7.07134306e-01 -8.27033162e-01
-1.07591915e+00 6.82744265e-01 -2.97841728e-01 4.01798487e-01
1.33114398e-01 -1.19360434e-02 1.81030273e+00 -7.32768238e-01
6.43389702e-01 1.33903384e+00 5.28133214e-01 5.70103705e-01
-9.47145224e-01 -3.81491244e-01 3.35855871e-01 1.05002865e-01
-1.17134655e+00 -4.16633338e-01 7.09382772e-01 -2.83360273e-01
8.69183064e-01 2.22355112e-01 2.71003693e-01 9.65175152e-01
3.11576515e-01 1.09107339e+00 7.06716835e-01 -5.14373183e-01
2.70050298e-02 4.66833375e-02 5.86693585e-01 5.36072612e-01
2.69858867e-01 4.57282424e-01 -4.80271161e-01 -2.12688968e-01
5.23292959e-01 -1.10018179e-01 -1.11551419e-01 -1.15768254e-01
-9.18696284e-01 9.81145680e-01 1.05288613e+00 6.44153535e-01
-3.16691935e-01 3.48593265e-01 3.03096771e-01 6.02995634e-01
4.85981613e-01 8.91217232e-01 -2.92467564e-01 2.82981485e-01
-6.05817199e-01 2.35943109e-01 3.62077117e-01 4.16986525e-01
4.03116405e-01 -1.80173025e-01 -3.61771405e-01 8.82244289e-01
5.33515275e-01 -1.86390787e-01 4.02736098e-01 -7.80937076e-01
1.46925732e-01 5.61785400e-01 1.78723827e-01 -1.00110793e+00
-6.37999296e-01 -5.18244922e-01 -6.02270424e-01 2.41252080e-01
2.20571801e-01 -2.22201154e-01 -6.63218200e-01 1.64923334e+00
-3.03327471e-01 -4.30148721e-01 2.53747404e-01 1.14421117e+00
9.47473407e-01 6.20433986e-01 4.86041695e-01 1.71537548e-01
1.19800138e+00 -8.87950659e-01 -5.08173168e-01 -4.95628685e-01
6.63888812e-01 -5.68895638e-01 9.78529930e-01 2.49135032e-01
-1.24265194e+00 -5.15389085e-01 -1.12238443e+00 -4.23823178e-01
-3.84402782e-01 1.63365945e-01 9.41190243e-01 2.78497547e-01
-1.26804256e+00 6.54995739e-01 -4.73029971e-01 -2.25591779e-01
2.99234509e-01 5.82692683e-01 1.12864308e-01 8.90791789e-02
-1.58250046e+00 1.22810948e+00 2.52625287e-01 5.21948993e-01
-6.98727846e-01 -3.10477614e-01 -4.62087184e-01 5.02191722e-01
1.25959694e-01 -7.19916523e-01 1.62444711e+00 -8.14843178e-01
-1.25122309e+00 5.78725576e-01 4.81835715e-02 -3.48597795e-01
1.53222278e-01 -2.87409723e-01 -3.65462542e-01 -2.86483355e-02
-4.41282809e-01 8.28731179e-01 4.10420626e-01 -1.49802685e+00
-4.71349090e-01 -5.51507652e-01 3.62252742e-01 3.65428627e-01
-4.40829784e-01 3.48878354e-01 -2.27156222e-01 -6.17327988e-01
-9.09829587e-02 -8.18137288e-01 -3.08166593e-01 -1.69989854e-01
-5.52578643e-03 -7.88663626e-01 2.77077049e-01 -3.83896858e-01
1.30668104e+00 -2.24138665e+00 -1.56550512e-01 2.88904101e-01
2.62310058e-01 3.80615234e-01 -4.48698640e-01 5.24165928e-01
-1.25720605e-01 4.21248853e-01 2.57519037e-01 -2.48818882e-02
9.68056098e-02 -2.69822329e-01 -4.46915001e-01 8.64219144e-02
3.98946464e-01 1.36661410e+00 -8.60600710e-01 -1.37658939e-01
-1.82576314e-01 7.68850327e-01 -2.18808860e-01 5.75650064e-03
-4.47603524e-01 -7.99368173e-02 -6.27165377e-01 4.32387352e-01
1.22862577e-01 -5.32083571e-01 4.06866521e-01 -8.45513642e-02
-6.68039992e-02 7.60954142e-01 -5.83743632e-01 1.48892629e+00
-1.43047601e-01 9.88094151e-01 3.37094665e-02 -8.87340426e-01
8.77367973e-01 4.53760356e-01 2.88307458e-01 -1.02428269e+00
3.52726966e-01 -8.14805478e-02 3.07115883e-01 -8.01350623e-02
9.23725009e-01 -2.04279318e-01 5.95912822e-02 8.96500170e-01
-1.56365722e-01 3.79174769e-01 -1.62069104e-03 1.89988583e-01
8.26562285e-01 -1.26334786e-01 -4.65432853e-02 -2.22208589e-01
1.09369814e-01 1.04971722e-01 2.02388838e-02 9.09482002e-01
-1.85348839e-02 6.36431754e-01 6.19428992e-01 -7.89808571e-01
-8.16926241e-01 -7.95394719e-01 -2.17046455e-01 1.51308322e+00
-6.67046681e-02 -3.79167527e-01 -3.20000410e-01 -5.92788160e-01
-1.76099017e-01 5.29133379e-01 -8.26454163e-01 -5.01401246e-01
-3.79387587e-01 -9.70695972e-01 4.00636047e-01 4.43358392e-01
2.10789945e-02 -1.35624719e+00 -5.84377050e-01 4.57942067e-03
1.00588938e-02 -6.93880498e-01 -4.28485125e-02 3.18590492e-01
-9.92747962e-01 -8.36782753e-01 -6.81620598e-01 -8.81019890e-01
5.36359429e-01 5.39102376e-01 1.34353209e+00 5.25489211e-01
-2.14686189e-02 2.82382667e-01 -2.43224576e-01 -4.85577583e-01
-3.33169192e-01 3.22305769e-01 -1.25469089e-01 -3.26363862e-01
5.57627797e-01 -1.43805251e-01 -8.54808271e-01 2.08509520e-01
-1.05905664e+00 -2.46186346e-01 8.13185573e-01 6.95957959e-01
-8.93737376e-02 -1.86146855e-01 7.12257922e-01 -8.11183333e-01
1.40389955e+00 -5.94622910e-01 -3.56535167e-01 5.04380524e-01
-1.07534564e+00 2.01851368e-01 2.68076330e-01 -1.09624766e-01
-1.11752331e+00 -5.60445309e-01 -6.16797507e-02 5.20799719e-02
2.00881749e-01 8.70165825e-01 4.48269784e-01 1.15941368e-01
9.66817915e-01 -2.97119349e-01 -1.94529086e-01 -6.32144988e-01
3.57474744e-01 6.68704331e-01 2.14171037e-01 -6.02011919e-01
4.83013779e-01 2.00603440e-01 -3.69242162e-01 -4.17000920e-01
-1.03350163e+00 -3.52669924e-01 -3.28746170e-01 4.46998468e-03
6.26066804e-01 -8.89640510e-01 -8.78068209e-01 -1.49488281e-02
-1.21907496e+00 -3.23918194e-01 -1.90437868e-01 2.55442977e-01
-1.66171923e-01 1.68549106e-01 -8.93342555e-01 -6.52034402e-01
-5.88357270e-01 -8.86900663e-01 7.28273332e-01 2.43945286e-01
-4.18208838e-01 -1.03846228e+00 2.01129511e-01 2.93433845e-01
6.52142346e-01 -3.75990033e-01 1.12877524e+00 -7.58937478e-01
-4.19705808e-01 -3.82638335e-01 -5.01353145e-01 2.05519721e-01
-2.99598962e-01 -1.51437335e-02 -1.23981178e+00 -1.49208590e-01
-3.08497488e-01 -4.15904194e-01 1.07291281e+00 4.28303063e-01
1.01836574e+00 -2.68510967e-01 -2.52016723e-01 1.53853714e-01
1.40993512e+00 2.00502798e-01 7.38766253e-01 6.02585077e-01
3.48089963e-01 7.37513483e-01 2.91474491e-01 1.29864201e-01
3.65861803e-01 6.60976887e-01 3.48601580e-01 -2.97517836e-01
3.82711105e-02 -1.48444816e-01 2.81531513e-01 4.40239042e-01
-2.42260009e-01 -3.49510163e-01 -9.54315126e-01 3.28442395e-01
-1.83187032e+00 -8.59494567e-01 2.29064941e-01 2.08191371e+00
6.72792614e-01 2.03012139e-01 -5.12491278e-02 -1.94057107e-01
4.31098163e-01 3.58687580e-01 -4.44289535e-01 -6.61627352e-01
-1.54737920e-01 1.53540581e-01 5.80803351e-03 3.40719372e-01
-8.38942587e-01 8.19805324e-01 8.25640965e+00 1.97072938e-01
-1.05628657e+00 -1.94495142e-01 8.26786637e-01 -1.30914867e-01
-6.18316412e-01 -5.38766123e-02 -6.58338070e-01 7.49668712e-03
1.04733396e+00 3.70974354e-02 4.00009274e-01 6.29479885e-01
-7.08656162e-02 1.92607656e-01 -1.25176561e+00 4.51874197e-01
-7.13621974e-02 -1.39286447e+00 3.29484999e-01 2.66718771e-03
4.92655605e-01 4.40441281e-01 3.39406937e-01 5.13354182e-01
7.82231390e-01 -1.18324184e+00 4.64484960e-01 2.78055996e-01
3.44685733e-01 -5.23846030e-01 7.75964856e-01 1.80590078e-02
-4.86271530e-01 -4.32302326e-01 -5.30274868e-01 -3.29473883e-01
-2.19057024e-01 1.76596969e-01 -7.61626542e-01 2.01267079e-01
6.06043875e-01 4.99953657e-01 -7.09381998e-01 8.21939051e-01
-2.30410621e-01 2.61235327e-01 2.34342627e-02 -1.61227331e-01
5.72978079e-01 2.26647351e-02 5.99490268e-05 1.12813759e+00
1.20592400e-01 1.08670853e-01 -2.58070409e-01 8.64766896e-01
-3.20279866e-01 5.30194268e-02 -6.39833212e-01 -2.99960464e-01
1.42184108e-01 1.34918714e+00 -4.55251783e-01 -3.50658409e-02
-4.26132858e-01 5.32960773e-01 5.06452680e-01 6.68968379e-01
-4.01327074e-01 -2.19164714e-01 6.70032918e-01 -1.06903613e-01
-4.23402525e-02 -6.30281270e-02 -3.74452412e-01 -9.33241367e-01
-8.84702131e-02 -7.15401292e-01 6.49928093e-01 -1.13690984e+00
-1.17132485e+00 5.59338033e-01 -2.45126262e-01 -8.38788211e-01
-2.64124572e-01 -7.18035698e-01 -6.04495227e-01 7.39734888e-01
-1.75339019e+00 -9.26977277e-01 -4.41059507e-02 2.23057240e-01
1.40014648e-01 -1.43973857e-01 9.62394297e-01 6.04895800e-02
-5.83687425e-01 7.23273575e-01 2.97316223e-01 1.80344626e-01
6.36355519e-01 -9.03260708e-01 5.16743660e-01 2.76813477e-01
5.33901989e-01 1.03193712e+00 5.46420097e-01 -1.91645380e-02
-1.39449811e+00 -4.93053645e-01 8.84172380e-01 -6.12444937e-01
5.63183188e-01 4.64660302e-02 -8.13587785e-01 7.61367142e-01
4.29170072e-01 -2.48962790e-01 6.49774611e-01 8.28325868e-01
-6.01522803e-01 -1.30189076e-01 -7.39077866e-01 7.16961205e-01
6.89076364e-01 -5.11060297e-01 -7.22898304e-01 1.96318761e-01
6.25155866e-01 6.26901016e-02 -8.39894891e-01 5.13601899e-01
9.19948459e-01 -8.85203242e-01 1.25487423e+00 -9.02385533e-01
7.40087807e-01 7.16163218e-02 2.31036067e-01 -1.36625707e+00
-7.12394774e-01 -3.29195797e-01 -4.37694266e-02 7.40877450e-01
7.57667601e-01 -3.51253122e-01 6.33979321e-01 8.67436707e-01
-2.65459679e-02 -8.03788006e-01 -6.02504611e-01 -3.25080872e-01
4.20853615e-01 -4.35197532e-01 4.75667417e-01 9.88321006e-01
5.50556540e-01 8.61862123e-01 -1.85381234e-01 -6.89334124e-02
1.43497080e-01 -7.76741579e-02 2.30341598e-01 -1.59359360e+00
5.67683484e-03 -9.65578914e-01 2.40682773e-02 -8.69529128e-01
5.96176349e-02 -8.81567955e-01 -1.28704965e-01 -1.70563757e+00
3.12182009e-01 -6.14219844e-01 -1.17439723e+00 6.36824191e-01
-1.57054871e-01 3.95073056e-01 2.39478692e-01 4.50138897e-01
-8.15006614e-01 2.17666686e-01 1.14953351e+00 -3.79058242e-01
-2.61450201e-01 -1.15769312e-01 -1.43331957e+00 3.35184127e-01
1.07277775e+00 -2.86376983e-01 -3.88507098e-01 -9.89924908e-01
9.16929722e-01 -1.36362627e-01 3.08837473e-01 -6.30228043e-01
3.25378567e-01 -9.39983409e-03 4.87105072e-01 -2.35494301e-01
2.20912606e-01 -6.10837579e-01 -2.69682854e-01 1.86546221e-01
-1.13829148e+00 1.17756359e-01 3.03409666e-01 4.11963820e-01
-3.69449943e-01 -3.31596315e-01 4.31887299e-01 -3.22883487e-01
-5.91756701e-01 1.45694539e-01 -2.94645876e-01 -2.31515154e-01
3.96953464e-01 6.32530302e-02 -4.89254117e-01 -5.31980395e-01
-4.49063778e-01 3.14521402e-01 2.48845026e-01 8.65728796e-01
5.40329039e-01 -1.13922358e+00 -5.76984942e-01 -1.62980277e-02
2.84287989e-01 -2.97236443e-01 5.22198416e-02 4.60820317e-01
-5.28125912e-02 9.16359067e-01 -6.63720742e-02 -1.17502667e-01
-9.20204461e-01 6.24902606e-01 3.21721464e-01 -3.80302399e-01
-5.35942018e-01 7.49691308e-01 3.67035031e-01 -1.83041811e-01
5.44217706e-01 -7.25276172e-02 -7.08955824e-01 1.29992366e-01
7.95501113e-01 6.28408641e-02 1.38547584e-01 -1.92596927e-01
-1.32468402e-01 3.87240738e-01 -5.70780754e-01 -1.86936021e-01
1.53741503e+00 -5.87570341e-03 -1.26324549e-01 3.05841416e-01
1.08419323e+00 -3.79763603e-01 -8.90047014e-01 -4.24392045e-01
3.81137431e-01 7.18047619e-02 4.95830387e-01 -1.21831059e+00
-1.03475440e+00 1.12791157e+00 6.16219103e-01 2.94951141e-01
9.60346401e-01 1.07748799e-01 5.93719482e-01 7.22375989e-01
-2.10806802e-02 -1.12124789e+00 -2.69040540e-02 5.96960664e-01
7.79592276e-01 -1.29653907e+00 1.47411808e-01 3.51371646e-01
-5.62951028e-01 1.11627257e+00 4.02802557e-01 -5.02344333e-02
5.38216293e-01 -6.03313372e-02 2.99468905e-01 -6.65296614e-01
-1.24427021e+00 -2.12058634e-01 5.61736405e-01 2.38585949e-01
1.08821881e+00 -1.84278533e-01 -3.55950654e-01 4.12598819e-01
4.62775305e-02 3.42718735e-02 1.05213128e-01 6.76276922e-01
-4.63890433e-01 -1.06795251e+00 -1.89467862e-01 2.70872235e-01
-7.38652885e-01 -3.65651160e-01 -6.90617204e-01 5.49443185e-01
-4.02497768e-01 8.58165979e-01 2.27842741e-02 -4.50617343e-01
1.43680677e-01 3.14638168e-01 2.57870287e-01 -3.95234555e-01
-8.56163859e-01 -2.55487412e-01 6.68780208e-02 -3.03790748e-01
-1.61184967e-01 -2.02429026e-01 -8.76110256e-01 -1.52929470e-01
-1.41517669e-01 3.61426920e-01 7.84561336e-01 8.02607238e-01
5.65251172e-01 4.15363014e-01 5.56525290e-01 -5.11940956e-01
-7.69258916e-01 -1.08622360e+00 -2.57500142e-01 2.07649857e-01
4.10931408e-01 -3.06447804e-01 -2.78871685e-01 -4.78886336e-01] | [11.455577850341797, 7.5906548500061035] |
d98d7ed0-59e4-495d-980b-5cb1ab8586a0 | real-time-video-quality-representation | 1602.00489 | null | http://arxiv.org/abs/1602.00489v2 | http://arxiv.org/pdf/1602.00489v2.pdf | Real Time Video Quality Representation Classification of Encrypted HTTP Adaptive Video Streaming - the Case of Safari | The increasing popularity of HTTP adaptive video streaming services has
dramatically increased bandwidth requirements on operator networks, which
attempt to shape their traffic through Deep Packet Inspection (DPI). However,
Google and certain content providers have started to encrypt their video
services. As a result, operators often encounter difficulties in shaping their
encrypted video traffic via DPI. This highlights the need for new traffic
classification methods for encrypted HTTP adaptive video streaming to enable
smart traffic shaping. These new methods will have to effectively estimate the
quality representation layer and playout buffer. We present a new method and
show for the first time that video quality representation classification for
(YouTube) encrypted HTTP adaptive streaming is possible. We analyze the
performance of this classification method with Safari over HTTPS. Based on a
large number of offline and online traffic classification experiments, we
demonstrate that it can independently classify, in real time, every video
segment into one of the quality representation layers with 97.18% average
accuracy. | ['Ofir Pele', 'Ofer Hadar', 'Ofir Trabelsi', 'Ran Dubin', 'Itay Richman', 'Amit Dvir'] | 2016-02-01 | null | null | null | null | ['traffic-classification'] | ['miscellaneous'] | [ 2.80614763e-01 -2.71677434e-01 -3.76586229e-01 -3.66063803e-01
-7.73043692e-01 -1.03512418e+00 7.89046213e-02 -1.03614427e-01
-1.18638210e-01 3.89607549e-01 -1.69818848e-02 -9.09554005e-01
-7.11577982e-02 -7.87573099e-01 -3.16640198e-01 -4.12352860e-01
-3.50005388e-01 3.45579058e-01 8.00774992e-01 8.32573324e-02
2.13856146e-01 6.85745955e-01 -1.70763838e+00 1.00313604e+00
4.00216192e-01 1.71725965e+00 -2.63715446e-01 1.21577775e+00
-4.00358737e-01 1.07065952e+00 -7.08395243e-01 -7.43914247e-01
6.02142036e-01 1.92708552e-01 -8.69756281e-01 1.58360004e-01
5.87358594e-01 -1.17539799e+00 -7.30747402e-01 7.42313206e-01
3.04287851e-01 -3.04438740e-01 1.89487934e-01 -1.65700614e+00
1.52600124e-01 5.65615833e-01 -2.03320757e-01 6.74645364e-01
4.46863353e-01 3.57199460e-01 1.10692441e+00 -1.06578171e-01
4.07003462e-01 8.29674006e-01 4.72617388e-01 3.90936524e-01
-1.07118440e+00 -7.88539827e-01 -2.28356689e-01 6.38484657e-01
-1.03479064e+00 -6.91421330e-01 3.81103545e-01 -2.23567158e-01
8.71392906e-01 6.68319941e-01 6.65818036e-01 8.83019924e-01
-2.38899752e-01 8.23701084e-01 5.31949639e-01 1.10858858e-01
2.88447410e-01 2.18591928e-01 -2.39215031e-01 3.92077118e-01
-1.47402316e-01 -1.97523698e-01 -2.12013438e-01 -5.21064222e-01
5.19771159e-01 -2.20978647e-01 -4.03995782e-01 -2.06817001e-01
-6.82527721e-01 4.51988786e-01 -2.59716004e-01 -1.56189010e-01
-3.86902004e-01 2.66244411e-01 1.01466703e+00 5.81244588e-01
3.28436524e-01 8.14605281e-02 -4.45387989e-01 -1.05351865e+00
-1.11119914e+00 2.25701313e-02 1.13716435e+00 1.24452150e+00
4.75555301e-01 1.23207323e-01 -2.19778884e-02 7.82299697e-01
5.51517587e-03 5.84870040e-01 -8.26328918e-02 -1.83372390e+00
6.79686725e-01 1.61226183e-01 1.00420430e-01 -8.50241303e-01
2.61043280e-01 2.59024173e-01 -2.92935878e-01 4.04514551e-01
4.55866039e-01 -2.36866456e-02 -2.57054955e-01 9.76841033e-01
9.20853689e-02 2.96795577e-01 -3.30586046e-01 6.19180858e-01
4.16332530e-03 7.89814651e-01 1.16615608e-01 -2.15567261e-01
1.22809923e+00 -6.53288901e-01 -4.38128740e-01 5.62444568e-01
4.98966426e-01 -7.34482467e-01 9.08586502e-01 9.87951636e-01
-1.47348249e+00 -2.62242794e-01 -8.16380739e-01 3.63072455e-01
2.51731444e-02 -4.24794883e-01 4.55341220e-01 1.25472569e+00
-9.24215555e-01 6.01359844e-01 -6.38898730e-01 -7.43805915e-02
6.85718238e-01 5.60352981e-01 -2.04315916e-01 -5.98200038e-02
-9.14662838e-01 -1.35174766e-01 -5.68942912e-03 -6.22481465e-01
-9.00319397e-01 -9.63885367e-01 -2.60728121e-01 5.33926487e-01
3.13215584e-01 -2.32499376e-01 1.51007783e+00 -1.23344791e+00
-1.57107508e+00 5.51316500e-01 4.52256761e-02 -6.27640128e-01
4.69974726e-01 2.16895156e-02 -7.27617145e-01 1.03194761e+00
-2.86920458e-01 3.47459525e-01 1.10932362e+00 -9.11634564e-01
-1.29825532e+00 1.31521141e-02 4.72302556e-01 -4.81276155e-01
-9.10763681e-01 3.97631705e-01 -3.23592931e-01 -3.27617675e-01
-4.23562467e-01 -4.73392010e-01 4.18024689e-01 4.32071656e-01
6.96817487e-02 7.78533053e-03 1.22097361e+00 -5.42226672e-01
1.62858665e+00 -2.34406543e+00 -6.23161376e-01 5.29447794e-01
6.41347528e-01 8.29248667e-01 -1.05649747e-01 3.21864307e-01
3.26693170e-02 6.70233369e-01 2.29248673e-01 2.67165620e-02
1.43236652e-01 2.26428941e-01 -6.36765659e-01 2.18151078e-01
-1.29486188e-01 2.30779469e-01 -8.53408158e-01 -4.69863117e-01
2.00709894e-01 3.82859558e-01 -9.99595463e-01 5.00333607e-01
-1.03410318e-01 -8.61757398e-02 -3.12310606e-01 7.07191765e-01
9.86856699e-01 -1.08307742e-01 3.25654507e-01 -3.92622203e-01
-1.50974661e-01 4.98815358e-01 -8.85875165e-01 8.83568347e-01
-5.30405402e-01 9.27654564e-01 5.38178682e-01 -8.86535645e-01
3.76322508e-01 8.10745239e-01 1.13997900e+00 -6.54215693e-01
7.17259943e-02 3.78485233e-01 -4.12261724e-01 -8.54183793e-01
5.44822693e-01 2.16441467e-01 3.90446872e-01 4.61139798e-01
-2.23930478e-01 2.03535080e-01 1.77732646e-01 1.94510728e-01
1.70026290e+00 -4.32277292e-01 -3.03899139e-01 2.42959931e-01
6.18443191e-01 -2.64017195e-01 4.17570144e-01 5.41377544e-01
-7.10982144e-01 5.41795909e-01 8.45328391e-01 -4.64012980e-01
-1.34775388e+00 -1.43392658e+00 6.51391968e-02 1.12883031e+00
-2.84630477e-01 -7.09136248e-01 -8.85750473e-01 -7.07737088e-01
-7.41580427e-02 6.32656932e-01 1.06641866e-01 4.39927541e-02
-7.39651024e-01 1.75168857e-01 9.71413851e-01 3.63457203e-01
5.91656148e-01 -7.26378858e-01 -2.92340577e-01 3.74927521e-01
-2.06436217e-01 -1.86410582e+00 -5.89689136e-01 -5.26295006e-01
-7.58863211e-01 -1.28314114e+00 -2.86367267e-01 -4.07748729e-01
9.24323350e-02 5.41968405e-01 1.01370609e+00 3.66403311e-01
-1.81373596e-01 1.03249252e+00 -5.80056310e-01 2.43651301e-01
-8.19131017e-01 1.98956639e-01 -1.50543228e-01 3.10766131e-01
4.15984154e-01 -9.07270789e-01 -7.19444215e-01 6.83176458e-01
-1.07867217e+00 -4.23790306e-01 -6.54883757e-02 1.15042359e-01
2.10038915e-01 2.93327659e-01 3.38929832e-01 -7.44434714e-01
5.66403925e-01 -7.85741568e-01 -6.34518206e-01 9.34522301e-02
-4.80994523e-01 -4.22691345e-01 8.48579347e-01 -3.90643299e-01
-7.24245489e-01 -5.47912121e-01 -5.41055024e-01 -6.87668502e-01
-2.29654655e-01 -2.03270644e-01 -1.11040458e-01 -3.25608939e-01
4.26085591e-01 -4.83226664e-02 2.46490121e-01 -2.72026867e-01
1.33502958e-02 1.33073878e+00 2.48820007e-01 -4.15329069e-01
6.01890564e-01 7.57573366e-01 -1.89164087e-01 -1.03539455e+00
-1.81006938e-01 -5.21644175e-01 -1.71791062e-01 -4.48863894e-01
6.21349633e-01 -4.76403564e-01 -1.35508943e+00 2.74293065e-01
-1.04947138e+00 -3.56047988e-01 -4.75241542e-01 2.16258883e-01
-7.98507512e-01 9.25439715e-01 -8.49099398e-01 -7.16246784e-01
-2.48338535e-01 -1.25992155e+00 8.73120546e-01 -2.79356986e-01
-1.36755988e-01 -7.52704084e-01 -4.54910398e-01 7.31988072e-01
8.57261419e-01 -3.87301505e-01 7.58384168e-01 -4.57114488e-01
-9.15432155e-01 -2.67152458e-01 -6.15089417e-01 8.45409453e-01
1.97418943e-01 5.99912047e-01 -1.01737547e+00 -2.47900456e-01
1.85097735e-02 -1.14698574e-01 4.16695595e-01 6.38186187e-02
2.00334525e+00 -6.03533447e-01 1.39370665e-01 9.51907516e-01
1.09356070e+00 1.58707231e-01 1.09419751e+00 2.94113487e-01
3.46670121e-01 2.58655101e-01 4.99923229e-01 8.67352307e-01
1.37314066e-01 5.33828437e-01 6.61017239e-01 4.94996279e-01
-4.64798771e-02 7.86554143e-02 6.72092080e-01 6.44011378e-01
-1.37343407e-01 -7.67358243e-01 -5.04738569e-01 4.17513549e-02
-1.20266414e+00 -1.52604890e+00 -3.08649912e-02 2.37213016e+00
4.89027470e-01 4.37719077e-01 4.32752669e-01 6.70172572e-01
5.69063187e-01 1.20608360e-01 -2.35313684e-01 -8.02577674e-01
1.74701735e-01 4.17087823e-01 9.33985114e-01 1.52170762e-01
-8.25169265e-01 5.28363109e-01 6.81783438e+00 1.15422380e+00
-1.15191829e+00 -1.23992421e-01 4.86082524e-01 -2.39673946e-02
-4.55396503e-01 -1.50982097e-01 -5.22811174e-01 8.03003073e-01
1.47931707e+00 -1.94119260e-01 8.44751477e-01 7.35202134e-01
5.65678179e-01 1.40035123e-01 -7.56346226e-01 1.13294399e+00
-3.87280047e-01 -1.46613991e+00 3.72061759e-01 2.94915646e-01
9.52818841e-02 -2.49827113e-02 8.39992985e-02 1.45858258e-01
-2.70293146e-01 -4.40488636e-01 4.63780612e-01 2.15069816e-01
1.17489862e+00 -4.39514816e-01 5.12159050e-01 -1.96519960e-02
-1.26601958e+00 -6.98594928e-01 -1.76555306e-01 2.50691861e-01
6.20905817e-01 3.61763895e-01 -7.46680439e-01 -1.59400016e-01
8.13764572e-01 4.24614429e-01 -2.78486103e-01 1.34075928e+00
3.60295862e-01 1.20281017e+00 -3.72840315e-01 3.39283019e-01
3.42167169e-03 -2.49046572e-02 5.78682482e-01 1.08026481e+00
4.56328392e-01 1.40357286e-01 -1.12968877e-01 3.28074902e-01
-3.32690269e-01 6.70387894e-02 -4.16580290e-01 -3.37087840e-01
5.19319415e-01 1.12934971e+00 -2.97529548e-01 -6.27353132e-01
-6.27691865e-01 8.85078967e-01 -2.75486916e-01 3.21475089e-01
-1.21882248e+00 -3.76997143e-01 1.35541606e+00 9.26585972e-01
6.01970017e-01 -6.07980371e-01 2.53527522e-01 -9.12025511e-01
1.05167128e-01 -1.07285976e+00 2.45779142e-01 -6.17601395e-01
-1.13393044e+00 4.25713837e-01 -4.34207208e-02 -1.40260768e+00
-6.97820708e-02 -7.86657214e-01 -3.73199373e-01 1.30694225e-01
-1.57454860e+00 -4.52501148e-01 -4.59222287e-01 7.46004045e-01
5.90284348e-01 -3.26509655e-01 5.44779301e-01 1.19784796e+00
-1.38913199e-01 8.57331634e-01 1.03941098e-01 2.87118405e-02
5.82913995e-01 -8.20926845e-01 3.98939103e-01 2.18187466e-01
-1.91981912e-01 6.77141873e-03 4.51520264e-01 -2.83791840e-01
-1.59199512e+00 -7.03573763e-01 3.20174485e-01 -5.98304905e-02
8.74137700e-01 -2.78000265e-01 -9.18087184e-01 3.99673373e-01
1.95633397e-01 1.36525482e-01 9.40635681e-01 -6.84916914e-01
-5.30667245e-01 -6.41829371e-01 -1.44919884e+00 2.79820025e-01
1.02095020e+00 -7.92107701e-01 4.17965263e-01 3.51228565e-01
7.01990724e-01 6.27757609e-02 -1.19875717e+00 1.87219635e-01
7.75343478e-01 -1.48948765e+00 8.40983927e-01 -7.82721579e-01
-5.93027323e-02 -7.17070245e-04 -4.35832441e-01 -4.60239530e-01
-1.43420115e-01 -1.10669827e+00 -3.57902169e-01 1.20707572e+00
1.36953548e-01 -7.59719551e-01 1.14747024e+00 8.20970714e-01
1.10479392e-01 -4.50274885e-01 -9.46324468e-01 -9.04841840e-01
-5.20819962e-01 -9.15575922e-01 8.29135597e-01 3.36069047e-01
-1.08280882e-01 -2.82493144e-01 -6.67783469e-02 1.48751244e-01
6.18570447e-01 -4.51331705e-01 7.81742752e-01 -8.65918994e-01
-3.08868766e-01 -4.85883623e-01 -9.44160998e-01 -1.43916881e+00
5.26930355e-02 -4.74776179e-01 -5.11169851e-01 -7.53508866e-01
-2.83390760e-01 -6.50205910e-01 -9.99702588e-02 -3.76526788e-02
5.55967450e-01 1.42454088e-01 7.73863494e-02 1.29643008e-01
-7.23559737e-01 1.27979875e-01 7.65030801e-01 -3.69350426e-02
2.68927604e-01 4.55125034e-01 -1.67330638e-01 3.86246920e-01
1.28149784e+00 -2.59575993e-01 -5.81099808e-01 -4.86325771e-01
4.57635105e-01 1.62927851e-01 3.16650927e-01 -1.25114310e+00
6.27947822e-02 -1.62261903e-01 -3.40859383e-01 -1.45703554e-01
4.08249587e-01 -1.27037716e+00 -1.77072603e-02 7.98395947e-02
-1.53293073e-01 1.47454768e-01 3.76138538e-02 4.77939457e-01
-1.41545981e-01 -1.83513835e-01 6.49998128e-01 1.91294372e-01
-6.59168899e-01 7.63696134e-01 -1.05088389e+00 3.49646568e-01
9.88517225e-01 -6.00531161e-01 -3.34447831e-01 -7.68528998e-01
-6.18502259e-01 5.23417108e-02 7.00000405e-01 1.16339542e-01
7.25781560e-01 -1.00758517e+00 -3.81286949e-01 4.59277332e-01
-1.85774669e-01 -7.95208395e-01 2.44871765e-01 7.75033891e-01
-1.20502710e+00 1.41020522e-01 -1.50319487e-01 -5.10320187e-01
-1.55807912e+00 7.76401699e-01 2.62504667e-01 3.63704562e-02
-4.40213174e-01 4.09042150e-01 -3.68108958e-01 3.73722881e-01
3.67014825e-01 -7.74140656e-02 1.55617639e-01 -1.60760894e-01
8.29391658e-01 1.04858649e+00 1.30804121e-01 -4.32039708e-01
4.01751027e-02 1.60756424e-01 1.51388034e-01 -5.58471307e-03
1.04940581e+00 -4.31018442e-01 -1.00097716e-01 -1.04738198e-01
1.69876528e+00 2.63854116e-01 -1.12410533e+00 2.49016851e-01
-2.57737726e-01 -1.12724853e+00 -1.38972392e-02 -6.55878857e-02
-1.63846600e+00 1.07897830e+00 5.83860874e-01 1.02110589e+00
1.32598543e+00 -4.47251141e-01 1.70556748e+00 7.48280212e-02
5.88777184e-01 -1.05912197e+00 -1.17761068e-01 1.45636380e-01
1.63944319e-01 -7.64994264e-01 -2.60680139e-01 -8.66025925e-01
-3.16630423e-01 1.59140170e+00 2.23026231e-01 8.96230713e-02
9.28584933e-01 5.20729482e-01 -1.31223619e-01 -8.80130678e-02
-1.03014851e+00 2.37570167e-01 -5.14195144e-01 6.51723206e-01
1.73916280e-01 9.55337211e-02 8.16438645e-02 -6.21123463e-02
7.75343105e-02 2.53036559e-01 8.30253363e-01 7.42723644e-01
-6.75637066e-01 -1.48175383e+00 -1.26281083e-01 8.33476722e-01
-8.63496065e-01 -6.77568791e-03 2.55620897e-01 2.37448901e-01
-2.22368017e-01 1.09774268e+00 3.61797333e-01 -6.89904809e-01
2.33509511e-01 -8.40477571e-02 1.88561544e-01 -1.06515385e-01
-6.98822796e-01 -4.22358215e-01 4.99895275e-01 -1.02255607e+00
-1.48113938e-02 -5.98099530e-01 -1.13576293e+00 -1.14816189e+00
2.71682232e-03 2.72335202e-01 7.06001520e-01 3.53431612e-01
5.87150455e-01 1.07760519e-01 1.09858990e+00 -4.79663312e-01
-6.35311007e-01 -1.43460259e-01 -6.80504739e-01 4.56736386e-01
3.67343456e-01 -2.49684557e-01 -7.91792214e-01 1.24180503e-01] | [5.072526931762695, 7.24387788772583] |
a4fdd04e-f9e3-4fe3-95ba-46f16ff855e2 | multi-grained-chinese-word-segmentation-with | null | null | https://aclanthology.org/2020.coling-main.183 | https://aclanthology.org/2020.coling-main.183.pdf | Multi-grained Chinese Word Segmentation with Weakly Labeled Data | In contrast with the traditional single-grained word segmentation (SWS), where a sentence corresponds to a single word sequence, multi-grained Chinese word segmentation (MWS) aims to segment a sentence into multiple word sequences to preserve all words of different granularities. Due to the lack of manually annotated MWS data, previous work train and tune MWS models only on automatically generated pseudo MWS data. In this work, we further take advantage of the rich word boundary information in existing SWS data and naturally annotated data from dictionary example (DictEx) sentences, to advance the state-of-the-art MWS model based on the idea of weak supervision. Particularly, we propose to accommodate two types of weakly labeled data for MWS, i.e., SWS data and DictEx data by employing a simple yet competitive graph-based parser with local loss. Besides, we manually annotate a high-quality MWS dataset according to our newly compiled annotation guideline, consisting of over 9,000 sentences from two types of texts, i.e., canonical newswire (NEWS) and non-canonical web (BAIKE) data for better evaluation. Detailed evaluation shows that our proposed model with weakly labeled data significantly outperforms the state-of-the-art MWS model by 1.12 and 5.97 on NEWS and BAIKE data in F1. | ['Min Zhang', 'Bowei Zou', 'Zhenghua Li', 'Chen Gong'] | 2020-12-01 | null | null | null | coling-2020-8 | ['chinese-word-segmentation'] | ['natural-language-processing'] | [ 3.22818190e-01 2.02576861e-01 -3.02006572e-01 -5.32068193e-01
-1.02631390e+00 -5.82794726e-01 1.82204545e-01 2.61448205e-01
-8.80737543e-01 7.63536930e-01 2.68725723e-01 -5.47589183e-01
4.29664463e-01 -6.92780316e-01 -7.01500893e-01 -4.54360515e-01
4.41434562e-01 5.21562040e-01 7.05009997e-01 -3.74066114e-01
1.36675954e-01 -8.51826072e-02 -7.61314988e-01 3.74845892e-01
1.32805240e+00 6.55393362e-01 6.37755036e-01 3.99169654e-01
-5.57086587e-01 1.45231932e-01 -6.28593147e-01 -7.57888496e-01
1.55906156e-01 -7.11464763e-01 -1.11445963e+00 1.60811156e-01
2.06306696e-01 1.33173451e-01 1.12822138e-01 1.17026424e+00
3.64258140e-01 1.49151266e-01 2.56391168e-01 -5.07297754e-01
-7.90178537e-01 1.27163684e+00 -7.74049282e-01 2.91963130e-01
1.05033301e-01 8.64651650e-02 1.48410308e+00 -8.00117254e-01
8.74505579e-01 1.03784788e+00 5.56529105e-01 7.08381116e-01
-1.13806260e+00 -4.52349246e-01 6.67460382e-01 -8.50079581e-02
-1.19731259e+00 -1.90054029e-01 7.11006939e-01 -1.12077616e-01
9.44828868e-01 2.31800720e-01 5.46486855e-01 1.24624395e+00
-1.29579246e-01 1.29004943e+00 1.07536149e+00 -8.17527592e-01
2.03518927e-01 -4.95178364e-02 8.74301851e-01 5.58144450e-01
2.93883383e-01 -5.34908295e-01 -5.01355350e-01 7.91980326e-02
1.61520511e-01 -4.85834301e-01 -4.12367851e-01 1.45221397e-01
-1.15973854e+00 9.10292089e-01 -4.36705127e-02 6.28196537e-01
-1.75504044e-01 -4.26251680e-01 4.80574578e-01 1.92556635e-01
7.39848852e-01 4.08691257e-01 -9.40864086e-01 -1.07642323e-01
-9.93791163e-01 3.01265847e-02 7.86186814e-01 1.17853844e+00
7.48194754e-01 1.49030849e-01 -3.00613850e-01 9.09235597e-01
2.55669415e-01 2.44006664e-01 1.00228238e+00 -1.85859472e-01
1.01750422e+00 6.18235409e-01 -1.17428109e-01 -5.54212868e-01
-3.13613981e-01 -5.82351267e-01 -6.83098435e-01 -5.17462730e-01
4.86752599e-01 -3.67338359e-01 -1.36029303e+00 1.79535544e+00
2.89391190e-01 1.11017562e-01 2.18074948e-01 8.71032655e-01
1.02148008e+00 7.54273355e-01 2.06333891e-01 -1.76012367e-01
1.41662204e+00 -1.30201030e+00 -7.36244500e-01 -5.56079447e-01
7.77895689e-01 -6.15354061e-01 1.76372457e+00 2.85554200e-01
-8.11958194e-01 -5.47262967e-01 -9.96896744e-01 2.28439700e-02
-4.57755089e-01 1.47277653e-01 3.92422229e-01 7.23286688e-01
-6.41166329e-01 3.66205156e-01 -1.00740600e+00 -2.68959135e-01
2.36952379e-01 1.37861744e-01 -1.97685093e-01 -1.30326420e-01
-1.57279146e+00 7.66463935e-01 8.80792797e-01 1.23611234e-01
-6.32676661e-01 -5.80517828e-01 -1.03674626e+00 -8.48662779e-02
7.05321372e-01 -3.88550490e-01 1.18300629e+00 -9.38969612e-01
-1.34954298e+00 9.70574975e-01 -9.85080153e-02 -6.12344444e-01
2.76922196e-01 -3.99116963e-01 -4.79365885e-01 -4.79655638e-02
2.76860684e-01 4.27980393e-01 5.35690963e-01 -1.19582379e+00
-5.43015301e-01 -2.40123868e-01 -6.81977421e-02 4.32371274e-02
-4.76960957e-01 1.89497977e-01 -7.64094353e-01 -9.55141544e-01
-1.03630565e-01 -6.97347522e-01 -4.30377394e-01 -9.03853238e-01
-9.78043854e-01 -3.26552182e-01 4.34605986e-01 -8.25670600e-01
1.50179756e+00 -1.98835897e+00 1.86055407e-01 4.29326817e-02
-9.00551155e-02 5.30213654e-01 -5.17404199e-01 4.04557943e-01
9.97368172e-02 3.99082601e-01 -6.10442042e-01 -6.72595561e-01
-2.98655946e-02 6.25210464e-01 -4.12292294e-02 1.17618576e-01
4.14596587e-01 9.73506272e-01 -1.04035795e+00 -6.63675129e-01
-2.62977749e-01 -2.58997768e-01 -2.84009010e-01 2.59733140e-01
-4.01893467e-01 4.52785492e-01 -5.61881721e-01 6.16262078e-01
5.26816130e-01 -1.53345047e-02 3.06044728e-01 1.94060095e-02
-6.20973110e-02 5.35627782e-01 -1.07529736e+00 1.98688507e+00
-3.27946514e-01 -3.71534191e-02 2.13472228e-02 -1.07236350e+00
8.45617771e-01 2.27764934e-01 1.02882206e-01 -6.13833368e-01
9.74048972e-02 3.77481848e-01 1.22800589e-01 -3.52442056e-01
7.41562247e-01 -2.02503681e-01 -5.27272999e-01 3.03507030e-01
5.17525017e-01 -2.31776126e-02 6.35626554e-01 1.85193673e-01
9.89730895e-01 1.19839057e-01 1.73317745e-01 -4.74418879e-01
4.61784780e-01 2.83679366e-01 1.03877246e+00 6.25020444e-01
-4.37454022e-02 8.79256666e-01 3.99673194e-01 -1.00551024e-01
-8.09341550e-01 -8.55114698e-01 -5.83827794e-02 1.22103000e+00
2.77057409e-01 -6.61440909e-01 -1.13184774e+00 -1.19741547e+00
-4.65083450e-01 9.55615044e-01 -4.26907957e-01 1.06340185e-01
-8.71043384e-01 -1.25725126e+00 7.33638585e-01 4.40418035e-01
5.00693202e-01 -1.18672955e+00 -7.35626072e-02 3.01975608e-01
-4.11963195e-01 -1.25290632e+00 -9.31276917e-01 3.75888973e-01
-4.61681426e-01 -9.79452670e-01 -6.57984316e-01 -1.10016239e+00
6.42108619e-01 2.36738613e-03 1.26022112e+00 1.26582868e-02
9.65952203e-02 -2.15038791e-01 -1.05321109e+00 -1.48667365e-01
-4.90164608e-01 7.97793031e-01 -2.08580539e-01 9.73950997e-02
4.49384481e-01 -5.36290370e-02 -3.45109224e-01 1.64762810e-01
-1.08986723e+00 3.04610729e-01 7.32642651e-01 1.08888316e+00
8.72168720e-01 -2.90017971e-03 7.58255243e-01 -1.56137466e+00
6.68521464e-01 -4.64745313e-01 -4.31845933e-01 3.65500122e-01
-6.14526570e-01 1.56054959e-01 7.68146336e-01 -5.63016713e-01
-1.23568308e+00 3.51940282e-02 -6.34344578e-01 3.41999352e-01
-2.67595232e-01 8.61334443e-01 -4.02005732e-01 3.47841740e-01
4.72890735e-01 3.24268490e-01 -5.71546257e-01 -9.20978725e-01
4.39676434e-01 9.43357229e-01 6.53666139e-01 -7.70256281e-01
6.49985433e-01 -1.71162799e-01 -7.31380999e-01 -7.14926600e-01
-1.32224309e+00 -6.23987079e-01 -8.14697146e-01 3.35878789e-01
1.02168202e+00 -7.27651238e-01 1.99852884e-01 5.94946921e-01
-1.24385941e+00 -4.66143638e-01 -3.79827082e-01 3.12829345e-01
-7.65872595e-04 6.71840906e-01 -1.00586224e+00 -5.34598231e-01
-5.60601413e-01 -9.33224618e-01 1.04806840e+00 1.19344555e-01
-2.54671752e-01 -8.36836874e-01 8.29642192e-02 5.78338921e-01
1.56614427e-02 3.02808285e-01 8.21640193e-01 -1.16207767e+00
-4.73444574e-02 -2.70277560e-01 2.36272737e-02 6.25177324e-01
1.42953128e-01 -2.37210914e-01 -5.23132205e-01 -2.66113579e-01
-1.58376619e-02 -4.82017100e-01 1.05826032e+00 5.01519218e-02
8.74045372e-01 -2.94898361e-01 -9.67116561e-03 2.78470904e-01
1.40188777e+00 1.47279829e-01 3.37627113e-01 4.60553437e-01
8.87948155e-01 5.49569547e-01 8.35145831e-01 1.04815550e-01
3.69390458e-01 2.79368252e-01 1.49891645e-01 -1.80526510e-01
-1.12024270e-01 -4.40088838e-01 3.53273541e-01 1.51993906e+00
2.04986572e-01 -5.61865687e-01 -9.94934082e-01 8.98693144e-01
-1.84491467e+00 -2.48752251e-01 -4.43196476e-01 1.84244823e+00
1.43482280e+00 6.65127575e-01 -1.19033530e-02 -1.21191582e-02
8.34489226e-01 4.21880275e-01 -3.41292858e-01 -3.55713338e-01
-3.65870565e-01 3.70540559e-01 6.06414020e-01 5.18714726e-01
-1.11363947e+00 1.46452701e+00 5.05623817e+00 1.34540558e+00
-7.79092133e-01 2.67017424e-01 8.08542192e-01 3.78502131e-01
-5.72505713e-01 -3.59615982e-02 -1.21889555e+00 8.36784601e-01
9.58502829e-01 1.37017295e-01 1.97228920e-02 7.18959272e-01
1.21873580e-01 -4.36993083e-03 -7.02592611e-01 3.63310963e-01
2.03031879e-02 -1.19421387e+00 2.73896027e-02 -3.60334903e-01
8.95639122e-01 3.33560437e-01 -4.34253246e-01 3.96707892e-01
5.35423815e-01 -7.33424842e-01 8.45923185e-01 -4.03028317e-02
7.69855380e-01 -5.44412017e-01 1.21877933e+00 6.94303274e-01
-1.13447261e+00 5.21049082e-01 -2.98663855e-01 1.24471389e-01
3.57874632e-01 6.12174690e-01 -4.33365196e-01 8.79225314e-01
4.69018906e-01 4.90369201e-01 -6.98046565e-01 7.01015592e-01
-6.17742240e-01 1.36571622e+00 -2.74542093e-01 -1.56081930e-01
7.36187100e-01 -2.86854059e-01 2.43026033e-01 1.67945898e+00
-7.34552294e-02 1.64234042e-01 4.42203850e-01 6.22879326e-01
-1.78624630e-01 3.80337894e-01 5.18707857e-02 -1.45526409e-01
3.71281236e-01 1.34381843e+00 -1.04450142e+00 -4.02438790e-01
-7.39207447e-01 1.02455592e+00 4.61133689e-01 2.96592563e-01
-5.02404749e-01 -6.02234483e-01 2.30942532e-01 -5.86697869e-02
4.88088489e-01 -3.32673401e-01 -3.81171942e-01 -1.42698467e+00
1.13814445e-02 -8.65176260e-01 5.51319122e-01 -2.41025925e-01
-1.48034334e+00 1.12993765e+00 -1.08443603e-01 -9.46548522e-01
-1.05172448e-01 -5.23722887e-01 -7.26752639e-01 9.15794492e-01
-1.84902775e+00 -1.28672838e+00 -1.99393439e-03 3.47480059e-01
9.56786692e-01 -8.45163912e-02 7.11686313e-01 1.93188712e-01
-9.84674990e-01 6.90202177e-01 5.28953336e-02 6.01746380e-01
5.72234452e-01 -1.53382146e+00 1.06739080e+00 1.32859421e+00
4.70635235e-01 5.43596685e-01 5.95140696e-01 -8.08003843e-01
-9.88523960e-01 -1.21817052e+00 1.05586803e+00 -2.79295668e-02
8.29752207e-01 -6.51609123e-01 -1.16811538e+00 6.21450663e-01
3.22580367e-01 1.34536158e-02 7.22749412e-01 1.01668254e-01
-7.67765492e-02 1.75757974e-01 -8.80587220e-01 5.35596550e-01
1.04789245e+00 -1.96923450e-01 -1.06009829e+00 3.07590663e-01
1.23736250e+00 -5.94850004e-01 -6.25214458e-01 2.83157259e-01
-5.15719242e-02 -5.29126883e-01 4.11812842e-01 -8.02777946e-01
3.02331477e-01 -2.11886525e-01 -9.61137041e-02 -1.56909084e+00
-1.25286311e-01 -6.83875322e-01 2.34349906e-01 1.61114192e+00
8.52466762e-01 -4.57338065e-01 9.41822231e-01 3.44539225e-01
-6.89419866e-01 -8.43892276e-01 -9.14079547e-01 -8.81860137e-01
1.49760112e-01 -6.03799045e-01 5.84341884e-01 9.41174507e-01
-7.34968297e-03 7.14629650e-01 -3.20645750e-01 4.90572490e-02
2.56095141e-01 3.53721470e-01 2.80182689e-01 -8.70671391e-01
-4.64617431e-01 -2.27748618e-01 3.69292730e-03 -1.32515121e+00
3.93252015e-01 -9.17090476e-01 4.84915614e-01 -1.63449359e+00
9.14916992e-02 -4.85332996e-01 -3.75461340e-01 7.12655962e-01
-6.75768316e-01 3.82775187e-01 5.01051731e-02 -8.46002921e-02
-8.28019977e-01 6.56918645e-01 1.30260253e+00 -2.03234199e-02
-1.87353328e-01 2.10723542e-02 -7.90664375e-01 7.61676192e-01
8.62997770e-01 -5.19091487e-01 -1.60245463e-01 -7.00495303e-01
-5.30421697e-02 -2.13740915e-01 -3.11163187e-01 -3.78482699e-01
7.83944204e-02 -3.94043356e-01 -5.73213026e-02 -5.27669370e-01
-1.61351159e-01 -4.32573885e-01 -4.58678931e-01 3.64560664e-01
-3.14571053e-01 -1.56917185e-01 2.72410493e-02 5.52347720e-01
-4.73394632e-01 -7.47412980e-01 6.19116306e-01 -3.14927816e-01
-8.91475022e-01 2.79793471e-01 -2.33455449e-01 7.55977869e-01
7.95475781e-01 -1.22351378e-01 -2.78412908e-01 1.42508000e-01
-5.46956241e-01 4.75836933e-01 1.18988134e-01 4.14930463e-01
3.85621011e-01 -1.04043150e+00 -9.08055186e-01 1.94458202e-01
1.92578182e-01 5.97267926e-01 6.52785003e-02 4.79079008e-01
-4.00086820e-01 1.67568386e-01 1.51695579e-01 -2.00308174e-01
-1.08503389e+00 3.09680074e-01 -1.57558948e-01 -5.69388151e-01
-6.22145891e-01 9.98281240e-01 -1.65989790e-02 -5.62605023e-01
7.19834715e-02 -6.79734111e-01 -1.72257304e-01 1.11550437e-02
2.43714795e-01 -3.17887589e-02 1.92910641e-01 -6.40486956e-01
-3.51580292e-01 4.21254933e-01 -2.65733331e-01 6.30512685e-02
1.37249839e+00 -1.77847058e-01 -7.60694891e-02 3.83762598e-01
9.18237627e-01 3.51518512e-01 -1.06125677e+00 -4.49096918e-01
7.30804026e-01 -1.55609205e-01 -1.61893427e-01 -7.88891554e-01
-1.03723359e+00 5.90391099e-01 -1.67042390e-01 1.19706124e-01
9.66407835e-01 1.69697687e-01 1.48892748e+00 9.67397839e-02
2.43786409e-01 -1.41963577e+00 -7.22347572e-02 6.70483291e-01
4.43666190e-01 -1.36642396e+00 -3.61667961e-01 -6.71871305e-01
-9.92591858e-01 8.82871568e-01 5.63949108e-01 -1.32352158e-01
4.31596369e-01 3.37720722e-01 1.17499545e-01 7.07926601e-02
-3.50107372e-01 -4.30951476e-01 4.30483043e-01 4.66120154e-01
3.72845501e-01 4.98430096e-02 -8.41185570e-01 1.40616071e+00
-2.81309664e-01 -4.41814214e-01 4.27454621e-01 7.81563222e-01
-5.71833193e-01 -1.46793282e+00 1.38825998e-01 4.65369731e-01
-7.88223207e-01 -4.89740759e-01 -2.91340649e-01 7.12750554e-01
2.39935800e-01 1.00643110e+00 -2.87490636e-01 -3.12952667e-01
4.72852707e-01 1.29788414e-01 -1.97636504e-02 -1.21745050e+00
-7.29604721e-01 1.85197204e-01 3.86852831e-01 -1.37071386e-01
-3.26982826e-01 -6.06500685e-01 -1.45931995e+00 2.03856975e-01
-6.38476789e-01 5.91592669e-01 4.31519806e-01 1.25915253e+00
-7.90264457e-02 5.23178697e-01 4.20361191e-01 -4.10091221e-01
-5.49610317e-01 -1.20167303e+00 -7.54840791e-01 5.52280962e-01
8.03647041e-02 -4.52671908e-02 -1.87496439e-01 7.33517036e-02] | [9.997364044189453, 10.10886287689209] |
bf85c76a-a4a7-4c59-842c-e5c5086b7c37 | emotion-recognition-in-audio-and-video-using | 2006.08129 | null | https://arxiv.org/abs/2006.08129v1 | https://arxiv.org/pdf/2006.08129v1.pdf | Emotion Recognition in Audio and Video Using Deep Neural Networks | Humans are able to comprehend information from multiple domains for e.g. speech, text and visual. With advancement of deep learning technology there has been significant improvement of speech recognition. Recognizing emotion from speech is important aspect and with deep learning technology emotion recognition has improved in accuracy and latency. There are still many challenges to improve accuracy. In this work, we attempt to explore different neural networks to improve accuracy of emotion recognition. With different architectures explored, we find (CNN+RNN) + 3DCNN multi-model architecture which processes audio spectrograms and corresponding video frames giving emotion prediction accuracy of 54.0% among 4 emotions and 71.75% among 3 emotions using IEMOCAP[2] dataset. | ['Mandeep Singh', 'Yuan Fang'] | 2020-06-15 | null | null | null | null | ['video-emotion-recognition', 'multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'computer-vision', 'speech'] | [-3.32921416e-01 -3.38009655e-01 1.79933965e-01 -5.20787835e-01
-3.19259316e-01 -4.29102272e-01 2.71119118e-01 -3.32905203e-02
-4.55981880e-01 6.12116337e-01 4.38239396e-01 -9.12578255e-02
2.30151609e-01 -3.63757551e-01 -2.24207386e-01 -2.74333745e-01
4.78496477e-02 3.33161056e-02 -2.73981512e-01 -3.25661421e-01
1.84463143e-01 6.31187558e-01 -1.67448103e+00 7.61008441e-01
2.52284557e-01 1.48684013e+00 -1.67990446e-01 1.46580970e+00
-4.32479382e-01 1.27181649e+00 -9.99197364e-01 -2.71049976e-01
-1.24035120e-01 -1.33347094e-01 -1.14576817e+00 -1.37863178e-02
-2.31501564e-01 -1.39482692e-01 -2.93668747e-01 6.70884788e-01
7.58879542e-01 5.61284900e-01 7.12236702e-01 -1.52043569e+00
-7.12455750e-01 1.51131973e-01 -3.07984799e-01 4.89220619e-01
4.09263700e-01 -3.91179770e-01 4.33784276e-01 -7.94846892e-01
2.02434286e-01 1.25880539e+00 4.56210971e-01 7.77630508e-01
-3.72967035e-01 -6.44260824e-01 -2.39053711e-01 7.10050941e-01
-1.14062011e+00 -5.39618671e-01 6.59155250e-01 -1.51718736e-01
1.85512531e+00 3.46541822e-01 6.67758346e-01 1.23949826e+00
3.43741477e-01 7.61285186e-01 1.09214723e+00 -3.76821935e-01
3.97919193e-02 3.33283633e-01 2.29031339e-01 3.14804763e-01
-6.76055610e-01 5.39100580e-02 -7.36247420e-01 1.59759074e-01
3.18420410e-01 3.73689272e-02 7.43333250e-02 7.21900046e-01
-8.65795732e-01 6.44216001e-01 6.95249215e-02 4.63500142e-01
-7.56868303e-01 1.85884759e-01 1.07589138e+00 7.37946928e-01
4.92272198e-01 3.12710218e-02 -1.01433325e+00 -9.48875189e-01
-5.46387255e-01 -3.34825695e-01 8.21976006e-01 5.66227973e-01
3.16880912e-01 8.01324904e-01 4.69224691e-01 1.25250947e+00
3.14906269e-01 3.53188783e-01 1.00993967e+00 -8.50155711e-01
6.10035881e-02 4.57639664e-01 -3.45978081e-01 -1.13123417e+00
-4.96092796e-01 -1.63482934e-01 -1.05244136e+00 2.42653206e-01
-1.08043432e-01 -5.32276154e-01 -1.03092229e+00 1.15257990e+00
1.27355829e-01 2.82799721e-01 7.18982041e-01 8.19210827e-01
1.13836360e+00 1.23797357e+00 3.30370486e-01 -9.88317579e-02
1.50029302e+00 -9.74260211e-01 -9.25175130e-01 -1.01071158e-02
5.39206803e-01 -9.18246448e-01 6.08173192e-01 8.75311375e-01
-8.63431752e-01 -6.18915081e-01 -8.87007713e-01 -3.29743288e-02
-8.94399941e-01 2.06511036e-01 6.43720329e-01 7.52583265e-01
-1.30539656e+00 8.03680196e-02 -5.67623854e-01 -6.09650373e-01
2.62475222e-01 6.56280100e-01 -6.76645398e-01 5.06569326e-01
-1.23900306e+00 9.99163270e-01 4.31287080e-01 -2.75089666e-02
-7.35975504e-01 -8.07323754e-02 -4.54390258e-01 1.48549363e-01
-9.27089378e-02 -1.77022487e-01 1.35773671e+00 -1.67081249e+00
-1.92035580e+00 5.39416373e-01 -1.90141618e-01 -5.76447427e-01
-3.02283704e-01 -1.29006684e-01 -1.14496863e+00 3.83380592e-01
-7.43181765e-01 9.46734905e-01 8.14724088e-01 -6.41878545e-01
-8.03904355e-01 -3.77167970e-01 -2.76046842e-01 2.02866763e-01
-5.61968386e-01 6.18634880e-01 1.05120137e-01 -4.39272076e-01
-8.33497569e-02 -7.88040876e-01 1.59626871e-01 -4.32890385e-01
5.03820181e-02 -4.35146153e-01 1.34994972e+00 -9.61894274e-01
1.01560736e+00 -2.04645777e+00 -2.10577413e-01 -1.36821702e-01
1.97878908e-02 5.39175391e-01 -8.58430043e-02 4.68327373e-01
-3.90172541e-01 2.53733039e-01 4.52584833e-01 -1.92833766e-01
-6.72442093e-02 4.13934618e-01 -8.49950761e-02 1.93487685e-02
3.71166348e-01 6.35417581e-01 -2.63459861e-01 -2.66921580e-01
4.36240107e-01 1.18218005e+00 -2.89336383e-01 2.49384761e-01
1.73368424e-01 2.79388636e-01 -2.99178064e-01 8.21375012e-01
3.49277884e-01 8.23577344e-02 -3.75825465e-01 -1.35806561e-01
-9.18333326e-03 -8.61877576e-02 -9.99739945e-01 1.43021870e+00
-7.44842768e-01 1.25757873e+00 9.05450508e-02 -1.45692670e+00
1.16021240e+00 1.19188333e+00 3.90288323e-01 -6.37987256e-01
5.95894039e-01 2.39425562e-02 2.95146685e-02 -1.02662897e+00
5.73997557e-01 -2.58375615e-01 -1.46006122e-02 2.90868282e-01
5.26003838e-01 3.14739347e-01 -3.91802937e-01 -1.91981927e-01
1.00501835e+00 -5.45037746e-01 3.78987491e-01 2.60862052e-01
6.96288824e-01 -9.32430774e-02 2.55760074e-01 2.72978783e-01
-7.72839844e-01 3.48496437e-01 1.93383873e-01 -8.26270759e-01
-8.97451580e-01 -5.41178942e-01 3.22457552e-01 1.41079402e+00
-5.85833609e-01 -1.27230719e-01 -5.03214121e-01 -3.47231537e-01
-6.08912289e-01 3.64554733e-01 -3.95343810e-01 -1.72944367e-01
-1.15152270e-01 -4.62697268e-01 8.56745839e-01 6.72087967e-01
6.88814938e-01 -1.45798385e+00 -7.26561189e-01 3.25719148e-01
-9.14789960e-02 -1.24982786e+00 1.01730093e-01 3.66388857e-01
-7.79363632e-01 -5.69506109e-01 -4.53566760e-01 -9.19921339e-01
4.60622460e-02 -6.80093914e-02 9.80008125e-01 6.70817941e-02
-4.07900572e-01 6.43865049e-01 -8.36604655e-01 -8.51187348e-01
-3.69705856e-01 -1.76914290e-01 8.36188495e-02 1.55116376e-02
6.22515976e-01 -5.78761995e-01 -3.35208625e-01 -2.34969914e-01
-8.13590050e-01 -1.85843781e-01 4.48732466e-01 6.65409923e-01
1.44332826e-01 2.51109719e-01 9.08185959e-01 -4.78558056e-02
1.06301498e+00 -7.81975985e-01 1.23580277e-01 1.11803450e-01
-2.38378957e-01 -2.86434472e-01 5.11762083e-01 -6.09264374e-01
-1.05274689e+00 1.14197105e-01 -7.29355931e-01 -5.02785325e-01
-9.13142562e-01 5.61685920e-01 2.62102276e-01 8.11131597e-02
3.89873773e-01 1.78686231e-01 1.30873220e-02 -1.93221360e-01
2.37593390e-02 1.45566463e+00 3.27594310e-01 -1.16161950e-01
-3.65951926e-01 1.86806247e-01 -4.22855258e-01 -1.29544234e+00
-3.95334691e-01 -4.37763244e-01 -1.81465209e-01 -6.74715579e-01
1.26372933e+00 -1.01793933e+00 -1.11368573e+00 5.82844198e-01
-1.38432407e+00 4.05566953e-02 3.21438700e-01 8.67263794e-01
-3.33794653e-01 -3.87974158e-02 -9.21584964e-01 -1.31371939e+00
-7.81777859e-01 -1.00907969e+00 8.25239539e-01 5.09275138e-01
-5.11883080e-01 -1.03737414e+00 -1.29901603e-01 3.11365157e-01
8.13338518e-01 1.21191949e-01 4.09897238e-01 -1.15350115e+00
1.53222203e-01 -2.86513388e-01 -9.61185172e-02 6.72803640e-01
-1.31834233e-02 3.03868949e-01 -1.25863206e+00 3.07102233e-01
1.87847823e-01 -6.66755617e-01 5.45775235e-01 3.78519893e-01
1.27615941e+00 -2.84736305e-01 1.51304200e-01 2.22916409e-01
1.20895660e+00 9.74227190e-01 7.26935327e-01 1.98364705e-01
2.45321706e-01 6.00305021e-01 2.16740042e-01 5.14518559e-01
4.39525366e-01 2.24032283e-01 3.27606946e-01 1.46391869e-01
1.26203090e-01 4.95223463e-01 5.06448507e-01 1.33672166e+00
2.78656725e-02 -5.31293333e-01 -9.90892828e-01 4.36333090e-01
-1.57736254e+00 -1.19083011e+00 1.13284132e-02 1.28923213e+00
3.50167423e-01 -2.14240998e-01 1.50119558e-01 5.43001473e-01
6.08952522e-01 -5.84614575e-02 -3.79765153e-01 -1.48016477e+00
-2.02325135e-02 4.47436094e-01 -2.08939582e-01 4.97535437e-01
-1.06958032e+00 9.70747948e-01 6.40603161e+00 5.98736346e-01
-1.74263251e+00 2.12713078e-01 8.99236977e-01 -3.43333006e-01
3.70586693e-01 -6.51644170e-01 -4.03240144e-01 3.46164197e-01
1.74911904e+00 -5.00338115e-02 6.03175700e-01 9.89798307e-01
2.76956141e-01 -3.72726358e-02 -4.18909967e-01 1.62642848e+00
2.38840386e-01 -1.08520210e+00 -1.43518940e-01 -2.79424876e-01
3.62113297e-01 2.71990508e-01 -8.13700445e-03 7.49105752e-01
-1.18165784e-01 -1.31625581e+00 2.40408063e-01 5.38882494e-01
3.88983905e-01 -1.28308165e+00 9.55798030e-01 2.14756712e-01
-1.16210675e+00 -2.09572390e-01 -2.56636471e-01 -4.19458747e-01
1.06996089e-01 1.71938971e-01 -8.81286860e-01 2.37988278e-01
1.16078496e+00 5.51765680e-01 -4.14229371e-02 5.48022330e-01
3.30626637e-01 7.28088379e-01 -3.41713220e-01 -4.72626984e-01
3.52766722e-01 4.34433520e-02 1.47446513e-01 1.64958417e+00
6.08260393e-01 5.03209293e-01 -2.17441067e-01 -4.08509895e-02
-1.00335099e-01 2.93364435e-01 -6.18376434e-01 -4.11724091e-01
2.97655255e-01 1.46323812e+00 -6.93851531e-01 -5.84810555e-01
-4.34980512e-01 1.14581764e+00 2.11902931e-01 3.32463771e-01
-7.84053981e-01 -5.27298570e-01 1.01699853e+00 -8.20910335e-01
4.25785817e-02 -2.67320186e-01 -8.28558430e-02 -7.66989827e-01
-2.95723975e-01 -9.79257405e-01 3.12450588e-01 -1.24731970e+00
-9.70601499e-01 1.20130134e+00 -4.35142279e-01 -8.24179351e-01
-3.78857046e-01 -9.68636751e-01 -6.29552007e-01 5.68248034e-01
-1.01239705e+00 -8.68620992e-01 -2.89381206e-01 8.38811338e-01
1.05970275e+00 -5.63512444e-01 1.16753268e+00 4.90497082e-01
-5.01245558e-01 2.00836241e-01 -1.01239793e-01 2.43987277e-01
5.97771347e-01 -9.23903883e-01 -1.65356681e-01 4.37010825e-01
2.75917649e-01 1.59242973e-01 6.04658246e-01 -1.25920296e-01
-1.38861489e+00 -7.75571704e-01 8.66573572e-01 -1.43380404e-01
4.45996821e-01 2.61992551e-02 -7.48345554e-01 4.39893961e-01
9.89496589e-01 -1.73199847e-01 1.14942884e+00 -7.27202222e-02
-1.47267491e-01 -1.33078471e-01 -1.23617411e+00 4.21350479e-01
2.52799422e-01 -8.02957654e-01 -3.63363862e-01 -4.34623174e-02
6.78829670e-01 -2.28851318e-01 -1.04844046e+00 1.65807575e-01
6.39530182e-01 -1.06135714e+00 7.63261735e-01 -8.24207067e-01
3.38286042e-01 1.60036907e-01 -5.81133604e-01 -1.44841933e+00
2.25057930e-01 -4.07324076e-01 -3.98564160e-01 1.13843429e+00
3.66806686e-01 -5.58343291e-01 7.00589597e-01 5.95574558e-01
-2.69703776e-01 -8.80691230e-01 -9.49273050e-01 -3.34727138e-01
-2.68416941e-01 -9.94480312e-01 4.68293399e-01 9.83544052e-01
2.08295882e-01 6.11511230e-01 -6.09197140e-01 1.51521685e-02
-2.86202848e-01 -5.09666681e-01 3.89938712e-01 -1.22462428e+00
1.67062551e-01 -3.94342512e-01 -7.68882096e-01 -6.48915350e-01
4.23742563e-01 -3.82613212e-01 -3.86382967e-01 -1.52343082e+00
-1.57296628e-01 3.03376615e-01 -4.09995347e-01 6.14128292e-01
3.38180184e-01 1.74236923e-01 7.84380641e-03 -4.56111938e-01
-6.83837652e-01 5.40670872e-01 8.37060690e-01 6.96932673e-02
-8.45306367e-02 -2.48385414e-01 -4.63260531e-01 7.12744117e-01
1.44329548e+00 -3.00330490e-01 -3.89840215e-01 -4.12014514e-01
2.65691161e-01 4.26273882e-01 1.72155291e-01 -1.16875422e+00
3.76194566e-01 1.84391439e-01 8.71448040e-01 -7.01969743e-01
8.43603194e-01 -1.09823465e+00 1.14388272e-01 1.86442062e-01
-2.94553161e-01 5.16996920e-01 5.48030734e-01 3.73581409e-01
-7.73295462e-01 -1.63406953e-01 5.50266981e-01 -1.02235377e-01
-1.17738688e+00 1.93548769e-01 -1.07567656e+00 -4.31120306e-01
9.36749995e-01 -2.66554683e-01 -2.15420604e-01 -9.14279699e-01
-1.18771207e+00 -1.39234588e-01 -2.64058024e-01 8.60229731e-01
9.22252178e-01 -1.30575299e+00 -4.03254628e-01 1.05970465e-01
-3.13159138e-01 -6.34550869e-01 4.03469265e-01 6.54278457e-01
-4.98930305e-01 6.09848619e-01 -5.72121441e-01 -3.16604495e-01
-1.76246989e+00 2.27268249e-01 5.10451198e-01 6.70167208e-02
-4.33191620e-02 9.75714624e-01 -5.14809430e-01 -4.50236410e-01
4.38229322e-01 -2.24433690e-01 -6.67190492e-01 7.06008822e-02
7.45554149e-01 5.68731487e-01 1.51158452e-01 -7.76341319e-01
-4.25755620e-01 3.16475242e-01 3.22122835e-02 -3.40261281e-01
1.41636300e+00 -1.73109591e-01 -1.77007623e-03 6.20342374e-01
1.64001238e+00 -5.12412369e-01 -6.22311175e-01 4.36357796e-01
5.86997792e-02 1.11760922e-01 2.93621927e-01 -1.08867764e+00
-1.22327268e+00 1.30011117e+00 1.10128009e+00 4.99666929e-01
1.60341144e+00 -2.87958771e-01 9.01913345e-01 6.01106286e-01
2.56988779e-02 -1.55751646e+00 4.55544412e-01 9.97166276e-01
7.86207616e-01 -1.56678927e+00 -6.66964471e-01 2.73587763e-01
-1.13328683e+00 1.42030919e+00 6.43899143e-01 -6.05716705e-02
9.42490280e-01 4.39821929e-01 4.94482636e-01 -3.19670051e-01
-1.29578972e+00 -7.36328885e-02 1.17934443e-01 6.68027401e-01
8.55202496e-01 7.37287030e-02 3.07900626e-02 7.72983551e-01
-2.32855648e-01 6.58485154e-03 3.97386104e-01 8.72767031e-01
-5.21995783e-01 -9.29833353e-01 -5.12941658e-01 5.31406105e-01
-1.01432407e+00 3.52992080e-02 -5.57034910e-01 4.47956413e-01
-7.37721473e-02 1.52096128e+00 2.35837996e-01 -7.75091350e-01
2.22073421e-01 5.87076008e-01 1.00364134e-01 -1.45398140e-01
-9.91339207e-01 -1.52846090e-02 3.63872707e-01 -4.06175852e-01
-6.97823763e-01 -3.10778558e-01 -1.49059618e+00 -3.82247359e-01
-7.63260946e-02 2.20917642e-01 1.29662192e+00 9.77367938e-01
6.94155395e-01 7.34605014e-01 4.03315872e-01 -6.17423117e-01
1.45004645e-01 -1.23930073e+00 -4.36033726e-01 9.48890075e-02
3.48022103e-01 -2.47070149e-01 -3.52793843e-01 3.31387073e-01] | [13.484542846679688, 5.448517322540283] |
07ba10f3-6569-4f71-95b8-aac16cbb8c76 | edge-augmentation-for-large-scale-sketch | 2202.13164 | null | https://arxiv.org/abs/2202.13164v2 | https://arxiv.org/pdf/2202.13164v2.pdf | Edge Augmentation for Large-Scale Sketch Recognition without Sketches | This work addresses scaling up the sketch classification task into a large number of categories. Collecting sketches for training is a slow and tedious process that has so far precluded any attempts to large-scale sketch recognition. We overcome the lack of training sketch data by exploiting labeled collections of natural images that are easier to obtain. To bridge the domain gap we present a novel augmentation technique that is tailored to the task of learning sketch recognition from a training set of natural images. Randomization is introduced in the parameters of edge detection and edge selection. Natural images are translated to a pseudo-novel domain called "randomized Binary Thin Edges" (rBTE), which is used as a training domain instead of natural images. The ability to scale up is demonstrated by training CNN-based sketch recognition of more than 2.5 times larger number of categories than used previously. For this purpose, a dataset of natural images from 874 categories is constructed by combining a number of popular computer vision datasets. The categories are selected to be suitable for sketch recognition. To estimate the performance, a subset of 393 categories with sketches is also collected. | ['Ondrej Chum', 'Giorgos Tolias', 'Nikos Efthymiadis'] | 2022-02-26 | null | null | null | null | ['sketch-recognition'] | ['computer-vision'] | [ 2.87373155e-01 -2.89744467e-01 -1.76094770e-01 -4.72012609e-01
-3.68297756e-01 -8.52000296e-01 8.53810608e-01 -2.11870268e-01
-4.89546895e-01 5.53805292e-01 4.78570955e-03 -1.58011615e-01
1.45017803e-01 -1.00153410e+00 -6.19582236e-01 -3.54305953e-01
3.56655642e-02 4.38456684e-01 1.63880274e-01 -7.89978802e-02
7.03304350e-01 9.36205506e-01 -1.63318837e+00 4.95895386e-01
4.99520719e-01 1.03438234e+00 3.21624219e-01 5.44710279e-01
-5.31292677e-01 1.45960823e-01 -5.35401821e-01 -3.10647249e-01
5.02682447e-01 -3.86581749e-01 -4.40834999e-01 2.63965964e-01
1.00221431e+00 -4.58863884e-01 -2.47664317e-01 7.92638659e-01
3.04374218e-01 1.42233849e-01 8.64901006e-01 -1.45820856e+00
-9.05746281e-01 4.38826621e-01 -2.38181427e-01 -2.04784542e-01
1.99009597e-01 -1.00195989e-01 8.60060871e-01 -1.36703455e+00
1.03949690e+00 1.09865630e+00 7.06000388e-01 8.57074082e-01
-1.28347313e+00 -1.03298056e+00 -7.21879378e-02 3.73810381e-02
-1.55387831e+00 -3.54219645e-01 1.43993950e+00 -2.86399275e-01
6.29493117e-01 1.05189227e-01 6.83062077e-01 1.35831463e+00
-4.65246469e-01 8.48652899e-01 1.12406862e+00 -6.89726412e-01
3.99763852e-01 3.73197556e-01 -3.48594151e-02 7.39322305e-01
2.47338399e-01 -2.10526888e-03 -4.04258400e-01 -1.55448705e-01
1.22263563e+00 1.96101427e-01 -1.72020756e-02 -8.44476104e-01
-1.31191921e+00 7.74362504e-01 5.49822152e-01 5.31463861e-01
-1.98606849e-01 7.69980019e-03 5.11501491e-01 5.76165915e-01
1.33172184e-01 5.72708488e-01 -2.89111137e-01 1.69281960e-01
-1.07790911e+00 1.80514157e-01 7.92206287e-01 1.08289206e+00
8.67938757e-01 1.76468983e-01 3.51968817e-02 1.14264834e+00
-2.99721479e-01 4.20528978e-01 4.46858734e-01 -6.21669650e-01
4.23203796e-01 8.73106599e-01 9.94568691e-02 -9.58252847e-01
-3.27946842e-02 1.97619968e-03 -1.10115027e+00 5.23458958e-01
7.67430305e-01 2.76843876e-01 -1.16949224e+00 1.58359218e+00
-2.04053372e-02 -4.67796959e-02 -8.29605088e-02 8.34724128e-01
7.51772523e-01 3.50857079e-01 5.67697063e-02 3.52898479e-01
1.19690263e+00 -9.86500382e-01 -3.02094400e-01 6.74882531e-02
5.45341410e-02 -6.96089387e-01 1.47474229e+00 5.64780533e-01
-7.24280894e-01 -9.66631949e-01 -1.13653433e+00 -6.18841313e-02
-9.38892901e-01 6.60740137e-01 9.16857243e-01 7.13447630e-01
-9.81485307e-01 7.84223139e-01 -9.69577357e-02 -5.52057266e-01
7.43203461e-01 2.77049422e-01 -8.06149006e-01 -2.19471171e-01
-8.37077796e-01 5.89357495e-01 2.35308424e-01 -1.92241341e-01
-7.22789824e-01 -5.29453397e-01 -7.70881057e-01 7.47247040e-02
1.75630808e-01 -4.22508895e-01 6.17155075e-01 -1.23184538e+00
-1.40890193e+00 1.02760351e+00 1.79847315e-01 -7.88485855e-02
7.13719606e-01 1.92967460e-01 -2.33311057e-01 2.72903740e-01
-2.67984737e-02 1.21973383e+00 1.44542611e+00 -1.39362586e+00
-4.05757129e-01 -1.81651458e-01 -1.52666390e-01 -1.75389603e-01
-6.35450423e-01 -1.91554368e-01 -5.04203916e-01 -1.18521047e+00
-4.69047539e-02 -9.27820921e-01 -3.35718915e-02 4.95107770e-01
3.33374180e-03 -2.92049319e-01 9.38077867e-01 -3.19273889e-01
7.87867308e-01 -1.96285379e+00 6.54933276e-03 4.37146306e-01
2.07088351e-01 2.75261223e-01 -5.78747451e-01 4.27790254e-01
-1.40944928e-01 2.11702064e-01 -4.28868860e-01 -4.61968422e-01
3.62969302e-02 2.52196342e-01 -4.69910353e-01 3.37340161e-02
5.02264857e-01 9.15432811e-01 -7.59561062e-01 -4.58600014e-01
2.95754731e-01 2.01964706e-01 -4.98926401e-01 3.83648604e-01
-1.74307212e-01 3.63848746e-01 -3.27752054e-01 9.66269374e-01
8.09993327e-01 6.70071691e-02 -5.54603562e-02 -1.12697482e-01
8.10328349e-02 -5.84152229e-02 -1.34363258e+00 1.94263983e+00
-4.99758571e-01 6.96347177e-01 -4.97659333e-02 -1.00646448e+00
1.45962346e+00 1.80718794e-01 1.76778153e-01 -3.94206583e-01
-4.82264385e-02 2.06547305e-01 -2.39263251e-01 -2.53187299e-01
5.28594553e-01 -2.45691061e-01 -1.52182519e-01 6.34123445e-01
3.72276545e-01 -4.76583987e-01 3.76379430e-01 -1.00011164e-02
9.04296577e-01 2.19431475e-01 9.33238417e-02 -3.53638887e-01
6.15369081e-01 1.24371216e-01 9.24412906e-02 8.83352578e-01
-1.36218175e-01 9.21612978e-01 1.62919372e-01 -9.27960932e-01
-1.63175046e+00 -1.13420880e+00 -2.05640867e-01 1.15112281e+00
-1.77398890e-01 -2.20774278e-01 -4.58507001e-01 -9.04441059e-01
4.11476165e-01 2.97016978e-01 -9.00912523e-01 -8.66061077e-03
-5.47499657e-01 -1.63335383e-01 7.47604787e-01 6.52142882e-01
6.22483730e-01 -1.53041625e+00 -3.04133922e-01 -1.11912847e-01
2.64588714e-01 -9.91639078e-01 -5.75385809e-01 -3.79240327e-02
-9.16512847e-01 -1.00366485e+00 -1.14194441e+00 -1.09972823e+00
1.13785768e+00 3.51913065e-01 1.07659793e+00 2.63905913e-01
-6.34473622e-01 5.37012100e-01 -3.57542485e-01 -2.79484272e-01
-4.42198992e-01 8.55108276e-02 3.91472131e-02 1.25914156e-01
3.24000537e-01 -6.79375529e-01 -4.08973515e-01 2.91797072e-01
-1.00394702e+00 1.79294571e-01 9.67407703e-01 1.23725545e+00
1.45223469e-01 -2.88925350e-01 7.73004651e-01 -7.91551709e-01
6.93439484e-01 -1.00740634e-01 -4.41532850e-01 3.52009684e-01
-4.43266809e-01 2.89881974e-01 1.07700598e+00 -1.04460847e+00
-9.18720543e-01 3.84961098e-01 5.33580892e-02 -6.42117620e-01
-6.46676302e-01 3.26010138e-02 -1.76197495e-02 -4.57328349e-01
7.04552531e-01 3.22500974e-01 4.09119241e-02 -6.17454708e-01
5.96791446e-01 7.12545633e-01 5.85404575e-01 -7.87316740e-01
9.40774322e-01 3.23020369e-01 -5.49510643e-02 -1.16702580e+00
-1.54263884e-01 -4.08986807e-01 -9.85463381e-01 -1.85115516e-01
2.72751242e-01 -3.83209080e-01 -2.79727668e-01 4.70535189e-01
-9.51312602e-01 -3.77790928e-01 -3.91295135e-01 2.03454867e-01
-4.12536830e-01 4.74377304e-01 -4.35449988e-01 -7.03008890e-01
-4.97560501e-01 -7.01516628e-01 1.15316975e+00 1.42695144e-01
-8.64521265e-02 -7.66438127e-01 -1.09911859e-02 -2.68105045e-02
6.20859623e-01 1.15766771e-01 9.49063241e-01 -5.37787139e-01
-5.43579161e-01 -5.56932032e-01 -5.43436766e-01 4.49183375e-01
1.71522230e-01 7.65492320e-02 -1.02607107e+00 -3.13449442e-01
-2.43016228e-01 -6.41389072e-01 9.65462863e-01 -1.80627644e-01
1.30678010e+00 5.76510914e-02 -2.85653621e-01 4.37181354e-01
1.42401040e+00 1.20414361e-01 6.56058669e-01 1.78459749e-01
6.07300997e-01 4.38270420e-01 2.14790374e-01 4.25668746e-01
-4.20951098e-02 5.02967715e-01 -8.67192820e-02 -1.70791522e-01
-3.68729025e-01 -5.97458780e-01 -7.67013803e-02 4.63327199e-01
-1.12086825e-01 7.96988010e-02 -7.76037455e-01 6.74637616e-01
-1.25019372e+00 -8.23235154e-01 5.19920528e-01 2.29886556e+00
7.40354300e-01 1.11330397e-01 2.32652560e-01 3.92316073e-01
7.88490832e-01 -1.15179038e-02 -3.42236817e-01 -2.39245355e-01
1.93473231e-02 5.18437207e-01 1.63848579e-01 1.96686491e-01
-1.15484786e+00 1.12010419e+00 6.16712141e+00 8.27795804e-01
-1.38713408e+00 -4.57459867e-01 4.10840511e-01 5.58676660e-01
-1.40986249e-01 -7.29525881e-03 -5.14959455e-01 3.48420173e-01
3.83702695e-01 2.28876829e-01 6.50957286e-01 1.20008588e+00
-4.58834708e-01 -6.25444129e-02 -1.26337790e+00 1.28271759e+00
2.50167161e-01 -1.47524393e+00 5.35358608e-01 -2.37790421e-01
6.90526962e-01 -4.33589548e-01 -5.81830218e-02 4.48531777e-01
7.47808293e-02 -9.95485663e-01 5.22591591e-01 5.77280164e-01
1.25556314e+00 -5.56713104e-01 3.61696571e-01 2.83012450e-01
-1.37066245e+00 -5.11324070e-02 -6.66945517e-01 -7.37044513e-02
-3.54159355e-01 6.67360574e-02 -9.51526821e-01 2.65135020e-01
5.76626182e-01 6.57520711e-01 -9.33310330e-01 9.83860493e-01
-1.11233129e-03 2.08601892e-01 -4.50811356e-01 -1.85795486e-01
1.36006653e-01 -4.16216284e-01 1.64518505e-01 1.41300297e+00
2.88558632e-01 -1.23590373e-01 -2.16422658e-02 1.11438513e+00
-4.83736098e-01 8.38275999e-02 -1.14848900e+00 -3.12207162e-01
5.13816535e-01 1.49796188e+00 -1.05297041e+00 -6.33554101e-01
-3.27564716e-01 1.27117598e+00 5.02043307e-01 1.31567448e-01
-4.40152258e-01 -7.14435220e-01 1.86582744e-01 -1.67045876e-01
3.56023520e-01 -4.06502813e-01 -4.05068606e-01 -1.30556548e+00
8.64237547e-02 -6.84966385e-01 2.14304134e-01 -7.15542734e-01
-1.62210441e+00 7.19878495e-01 -8.34858567e-02 -1.30127633e+00
-1.01130866e-01 -9.28646624e-01 -8.73882055e-01 7.46830404e-01
-1.32952642e+00 -1.46750212e+00 -7.52356768e-01 5.72707772e-01
7.45457888e-01 -4.58818644e-01 1.15666842e+00 4.98641193e-01
-1.93274543e-01 8.22407365e-01 -2.11850796e-02 5.40756047e-01
9.00910497e-01 -1.30723310e+00 7.55748332e-01 5.34509361e-01
4.60840672e-01 6.23895109e-01 1.87061116e-01 -5.30379713e-01
-1.42947793e+00 -8.85813653e-01 7.82572985e-01 -4.29744750e-01
4.66040850e-01 -8.45488667e-01 -9.42401528e-01 2.83464909e-01
-1.33998290e-01 3.62624645e-01 3.81907791e-01 -3.03851396e-01
-7.41350055e-01 -1.56723797e-01 -1.36883426e+00 6.71624362e-01
1.19150841e+00 -8.30393493e-01 -6.84446275e-01 -6.57839850e-02
-4.52966727e-02 5.82385808e-02 -7.72841573e-01 2.69398093e-01
9.90884423e-01 -6.28508568e-01 1.20052159e+00 -6.89703941e-01
4.05037314e-01 -7.27993436e-03 -2.05501858e-02 -1.10010791e+00
-2.04428673e-01 -3.49469006e-01 2.84603745e-01 1.35587037e+00
1.18374139e-01 -3.56717259e-01 1.25766599e+00 4.57355559e-01
2.27059469e-01 -3.71176988e-01 -6.17682040e-01 -8.55491400e-01
2.69518942e-01 -2.39376947e-01 4.12486583e-01 1.16583157e+00
-5.12994118e-02 3.61288518e-01 -3.01377982e-01 -4.62591708e-01
6.26234829e-01 5.45444310e-01 1.08553481e+00 -1.48072827e+00
2.53673196e-01 -7.31885493e-01 -4.27634418e-01 -9.83966053e-01
1.95326149e-01 -9.51729596e-01 -1.53776482e-01 -1.17956448e+00
1.20352022e-02 -8.97793889e-01 -1.27541348e-01 4.99861449e-01
9.31245387e-02 6.82423532e-01 2.93782741e-01 2.38200679e-01
-3.10716420e-01 5.10709286e-01 1.34680736e+00 -3.11899096e-01
-9.15178284e-02 -7.90714175e-02 -2.63825864e-01 4.50554729e-01
7.44838059e-01 -2.36740872e-01 -2.71325737e-01 -4.04503979e-02
-4.04034972e-01 5.87440655e-02 3.02971661e-01 -1.22147787e+00
2.09761560e-01 -8.76617953e-02 7.29410291e-01 -6.01378441e-01
3.83497030e-01 -1.06524205e+00 -1.35692693e-02 2.76672184e-01
-5.55696428e-01 3.16653779e-04 2.47744977e-01 4.60663736e-01
-1.49956942e-01 -3.74604225e-01 8.17237139e-01 -3.43581557e-01
-9.97496605e-01 2.47431204e-01 -7.86438808e-02 -2.75688559e-01
8.47459078e-01 -4.24291313e-01 -3.67635302e-02 -2.47550607e-01
-6.14595950e-01 -3.08811992e-01 7.46939838e-01 6.78796411e-01
9.99485910e-01 -1.68272388e+00 -5.53615093e-01 7.26511478e-01
4.22518134e-01 -3.34033400e-01 -2.19477993e-02 1.11749686e-01
-5.36398172e-01 3.52682769e-01 -9.93453383e-01 -3.42568010e-01
-1.13103116e+00 1.00424945e+00 -5.03034554e-02 9.70191136e-02
-8.59781206e-01 6.17824316e-01 -3.19043025e-02 -6.29205227e-01
4.40103352e-01 -4.44061309e-01 -1.16606183e-01 -4.04905193e-02
4.87741351e-01 1.99277580e-01 -1.37824729e-01 -3.40260684e-01
-5.16018793e-02 6.34771705e-01 1.05362236e-01 -3.79087515e-02
1.39070892e+00 2.80097932e-01 -1.19501181e-01 3.03050160e-01
1.23944998e+00 1.20689394e-02 -1.40091741e+00 -2.31471106e-01
2.35242322e-01 -7.25216329e-01 -3.68207872e-01 -7.69936442e-01
-9.96939480e-01 9.87076402e-01 6.42894983e-01 4.78548184e-02
9.20622349e-01 -2.16967404e-01 6.17369235e-01 7.30327308e-01
6.09677732e-01 -1.00681221e+00 3.37749302e-01 3.31252307e-01
1.28499937e+00 -1.50752389e+00 -2.55028218e-01 -1.23552375e-01
-5.67217290e-01 1.60575521e+00 4.85510856e-01 -6.88376486e-01
5.27780890e-01 1.92435682e-01 -1.39386475e-01 -2.63652793e-04
-3.39882612e-01 -1.65803567e-01 3.61862034e-01 7.11059153e-01
9.52472016e-02 -5.10076471e-02 -1.67592868e-01 5.19768715e-01
1.51152119e-01 2.17704862e-01 3.62646788e-01 8.60888720e-01
-1.95380330e-01 -1.45445120e+00 -5.29555738e-01 5.12458563e-01
1.03477724e-01 2.48874202e-02 -6.90782070e-01 9.40950036e-01
1.26843214e-01 3.57337654e-01 9.65900496e-02 -3.95618141e-01
2.59807020e-01 4.05310899e-01 5.91971278e-01 -3.88405263e-01
-4.13260967e-01 -4.55688804e-01 -1.25363141e-01 -1.56980231e-02
-2.00382143e-01 -2.22032532e-01 -7.60881841e-01 -1.62306681e-01
-1.92487910e-01 1.03565998e-01 8.11437368e-01 5.66594481e-01
1.32326454e-01 4.02720571e-02 5.81194937e-01 -1.52008259e+00
-4.68865454e-01 -9.75939155e-01 -5.48612952e-01 9.63709354e-01
1.24277785e-01 -8.77969563e-01 -2.84511298e-01 1.83747292e-01] | [11.768340110778809, 0.44448143243789673] |
7a7df4ba-1abe-41c2-8923-326ecd2bda4b | a-simple-approach-to-jointly-rank-passages | 2109.10497 | null | https://arxiv.org/abs/2109.10497v2 | https://arxiv.org/pdf/2109.10497v2.pdf | A Simple Approach to Jointly Rank Passages and Select Relevant Sentences in the OBQA Context | In the open book question answering (OBQA) task, selecting the relevant passages and sentences from distracting information is crucial to reason the answer to a question. HotpotQA dataset is designed to teach and evaluate systems to do both passage ranking and sentence selection. Many existing frameworks use separate models to select relevant passages and sentences respectively. Such systems not only have high complexity in terms of the parameters of models but also fail to take the advantage of training these two tasks together since one task can be beneficial for the other one. In this work, we present a simple yet effective framework to address these limitations by jointly ranking passages and selecting sentences. Furthermore, we propose consistency and similarity constraints to promote the correlation and interaction between passage ranking and sentence selection.The experiments demonstrate that our framework can achieve competitive results with previous systems and outperform the baseline by 28\% in terms of exact matching of relevant sentences on the HotpotQA dataset. | ['Chitta Baral', 'Shuguang Chen', 'Man Luo'] | 2021-09-22 | null | https://aclanthology.org/2022.naacl-srw.23 | https://aclanthology.org/2022.naacl-srw.23.pdf | naacl-acl-2022-7 | ['passage-ranking'] | ['natural-language-processing'] | [-1.70229405e-01 -1.11576907e-01 2.96810180e-01 -5.03109753e-01
-1.43314946e+00 -7.23941088e-01 5.67730725e-01 2.90703237e-01
-5.80046117e-01 7.73349285e-01 3.89335394e-01 -1.20118544e-01
-2.51135170e-01 -6.18872523e-01 -4.96779770e-01 -3.65382701e-01
2.30723143e-01 5.98195374e-01 9.23998356e-01 -7.51493454e-01
5.22709429e-01 5.07021248e-02 -1.68880010e+00 6.35160685e-01
1.26127374e+00 6.70958817e-01 4.80871886e-01 8.76351297e-01
-3.20411056e-01 1.09790170e+00 -8.99562538e-01 -4.53277707e-01
4.08421233e-02 -7.06286013e-01 -1.31794310e+00 -3.23364824e-01
6.84318066e-01 -3.68839771e-01 -2.35156566e-01 7.50046909e-01
9.34228480e-01 5.32931745e-01 4.40207928e-01 -9.16875958e-01
-6.13799453e-01 4.97722119e-01 -1.22682333e-01 7.08324313e-01
6.87842131e-01 -1.08750202e-01 1.38037062e+00 -8.76141131e-01
3.63643885e-01 1.19195294e+00 1.02440462e-01 6.93684161e-01
-6.94415212e-01 -1.99359521e-01 1.95724875e-01 5.40574551e-01
-1.00721991e+00 -5.61029315e-01 7.14842081e-01 -8.47984850e-02
9.88340139e-01 8.10734093e-01 1.29100457e-01 8.67027700e-01
-8.17642286e-02 9.91961598e-01 7.98420906e-01 -4.63006377e-01
4.30622920e-02 2.29809225e-01 5.87608933e-01 4.29731876e-01
-3.32504421e-01 -3.81501406e-01 -6.87355638e-01 -1.54363438e-01
1.17323063e-02 -3.32206368e-01 -3.48488420e-01 -2.66421363e-02
-9.47212458e-01 8.15146565e-01 3.31675112e-01 3.15924376e-01
-9.31773111e-02 -3.00221413e-01 4.63123292e-01 6.91332936e-01
2.47362405e-01 1.11440766e+00 -5.51652491e-01 -2.18917996e-01
-7.39686549e-01 6.26122355e-01 9.44538593e-01 8.12260151e-01
4.92079616e-01 -5.79149246e-01 -1.07109666e+00 1.24339199e+00
1.60314366e-01 4.99555677e-01 4.33170348e-01 -1.02568781e+00
7.70618737e-01 6.90019369e-01 3.40699732e-01 -8.44148040e-01
-2.78828681e-01 -2.60475129e-01 -1.87220976e-01 -4.77606088e-01
3.77706558e-01 -1.24304472e-02 -6.27387524e-01 1.51164198e+00
3.37479144e-01 -3.91044408e-01 3.50710824e-02 9.13367987e-01
1.43360257e+00 8.45428109e-01 9.91567001e-02 -2.43372470e-01
1.43779910e+00 -1.43087280e+00 -8.21682334e-01 -2.66172171e-01
7.28459835e-01 -1.15175760e+00 1.51788676e+00 6.80963695e-02
-1.34789753e+00 -4.79239076e-01 -9.00162876e-01 -5.83649039e-01
-1.63544402e-01 9.02301967e-02 7.98176005e-02 1.75429881e-01
-9.93872106e-01 3.56027037e-01 -2.04858288e-01 -2.62734026e-01
-1.31977601e-02 1.33535266e-01 1.42777145e-01 -1.25288934e-01
-1.55948937e+00 1.22124374e+00 2.43461244e-02 -1.12219922e-01
-7.76518404e-01 -6.23991609e-01 -5.49769163e-01 3.28914016e-01
3.80231291e-01 -7.11714566e-01 1.53084230e+00 -4.14287239e-01
-1.50828242e+00 6.89785719e-01 -3.33145469e-01 -4.28247362e-01
3.71129304e-01 -6.96408153e-01 -2.16036811e-01 6.03992105e-01
3.19525301e-01 6.21326566e-01 4.60721403e-01 -7.85738885e-01
-6.66827559e-01 -2.23844051e-01 5.19886434e-01 6.69275522e-01
-3.78116965e-01 3.84340316e-01 -6.00390911e-01 -1.95051357e-01
4.95153293e-03 -5.35305619e-01 -1.76004902e-01 -4.86510783e-01
-3.71825486e-01 -7.77153075e-01 7.74265409e-01 -6.98168933e-01
1.33524430e+00 -1.91973042e+00 2.35324576e-02 -2.96007633e-01
1.56242549e-02 4.85288799e-01 -3.11778516e-01 7.03052342e-01
4.75628018e-01 7.12284744e-02 1.54862314e-01 -2.00537667e-01
9.11023542e-02 7.72237703e-02 -5.65759242e-01 -3.31146382e-02
3.25666845e-01 8.36988688e-01 -1.10637629e+00 -7.80972540e-01
2.46168673e-02 1.92913711e-01 -5.06125271e-01 4.92468476e-01
-4.16435421e-01 3.50384325e-01 -7.54173517e-01 1.84721380e-01
3.48552823e-01 -1.17008314e-01 -3.38936359e-01 3.50200653e-01
1.56167477e-01 1.01909792e+00 -7.99289346e-01 1.49265611e+00
-4.47492033e-01 4.32198137e-01 -2.16484189e-01 -5.77549279e-01
9.77323413e-01 3.93164188e-01 1.94850750e-02 -1.11833489e+00
3.61225903e-02 2.71954328e-01 -3.01662590e-02 -1.00607121e+00
7.56174088e-01 1.66861545e-02 -1.00675628e-01 5.18011034e-01
1.81078345e-01 -3.61801237e-01 7.30262041e-01 5.56516826e-01
1.24794233e+00 -1.33130252e-01 -6.48213476e-02 -2.85455436e-01
9.95626688e-01 -2.07465347e-02 2.89986551e-01 1.06973088e+00
-4.45939004e-01 7.17729867e-01 5.05678594e-01 -2.41933823e-01
-7.72471428e-01 -9.87282276e-01 6.62134662e-02 1.40362513e+00
2.39476353e-01 -3.62706453e-01 -6.55252695e-01 -1.09390163e+00
-4.39003199e-01 9.21507001e-01 -3.71965915e-01 -1.82478219e-01
-7.03613520e-01 -3.75334173e-01 3.47600430e-01 2.30206236e-01
4.44692880e-01 -1.27435374e+00 -5.08857012e-01 8.20664316e-02
-7.76008844e-01 -1.09232044e+00 -5.48478305e-01 4.97454312e-03
-6.68509722e-01 -1.10692453e+00 -5.52372575e-01 -9.06848729e-01
4.71944392e-01 5.40223479e-01 1.53032708e+00 4.85162109e-01
1.73564300e-01 3.25117856e-01 -7.43996322e-01 -4.10685331e-01
-3.78161430e-01 4.46529478e-01 -3.83568138e-01 -3.91190648e-01
6.17800474e-01 -2.92518944e-01 -6.81000113e-01 4.21851426e-01
-8.83263469e-01 -1.05350226e-01 3.55483770e-01 8.63312244e-01
2.31771812e-01 -2.80178934e-01 1.01655638e+00 -7.69648731e-01
1.07698810e+00 -3.14606071e-01 -2.34026074e-01 6.78860247e-01
-2.56748885e-01 1.66289788e-02 7.24462926e-01 -2.72322372e-02
-1.03683293e+00 -3.93524796e-01 -5.48405707e-01 6.45458177e-02
-9.75731239e-02 3.34640801e-01 -6.03064783e-02 1.39951885e-01
7.17953265e-01 2.17760697e-01 -2.64845371e-01 -5.22186697e-01
2.49282822e-01 6.40228570e-01 1.42217442e-01 -4.76566762e-01
6.05433524e-01 9.93318558e-02 -4.44558740e-01 -6.42957151e-01
-1.40647614e+00 -8.80543768e-01 -5.12245536e-01 -1.23955399e-01
6.18595839e-01 -6.82723880e-01 -4.84573185e-01 -9.75168049e-02
-1.17923117e+00 1.57146618e-01 -3.03595930e-01 1.72654718e-01
-3.21250319e-01 4.65379685e-01 -7.12444603e-01 -8.43001008e-01
-5.39821029e-01 -1.09735990e+00 9.54035759e-01 3.90915900e-01
-2.06133068e-01 -6.65174246e-01 1.86205715e-01 9.14625406e-01
4.04078603e-01 -3.53134602e-01 8.76281261e-01 -1.04486835e+00
-5.37055254e-01 -3.36797893e-01 2.81975861e-03 4.32499617e-01
-2.54720181e-01 -1.18209966e-01 -1.00649965e+00 -1.23476252e-01
2.07893908e-01 -7.40478277e-01 9.86659050e-01 3.34140994e-02
8.74195695e-01 -2.18514293e-01 1.66928902e-01 -1.00861378e-01
1.13782358e+00 1.76851004e-01 6.69028223e-01 4.59358901e-01
3.49347800e-01 9.48708951e-01 8.47915173e-01 9.42030028e-02
4.88944590e-01 6.40354991e-01 2.58800209e-01 2.65475780e-01
-1.62035033e-01 -3.05644095e-01 3.29851180e-01 1.14039874e+00
5.99511862e-01 -2.79631972e-01 -6.94673955e-01 8.23506653e-01
-1.74432731e+00 -1.11145222e+00 -3.23722094e-01 2.15125990e+00
1.01344168e+00 8.41894075e-02 8.69812742e-02 3.69821638e-02
4.81667429e-01 1.99420944e-01 -1.57889336e-01 -5.50114274e-01
-1.64985716e-01 7.93034136e-02 -3.74790698e-01 8.07991505e-01
-1.09439743e+00 8.95632923e-01 6.57302904e+00 8.64971697e-01
-4.97168630e-01 1.20220475e-01 6.21065378e-01 -2.88645327e-01
-4.20755893e-01 -1.45784989e-01 -9.00917411e-01 3.52936029e-01
1.02506006e+00 -1.01623505e-01 1.06945664e-01 5.68311810e-01
1.67023972e-01 -2.94630319e-01 -1.03928256e+00 3.76307279e-01
3.16822261e-01 -9.71433878e-01 1.84674054e-01 -6.77020013e-01
6.30441487e-01 -2.84743570e-02 -9.18645784e-02 7.37753153e-01
2.28499293e-01 -9.36768591e-01 4.63782758e-01 2.64357328e-01
-2.11930588e-01 -7.59095311e-01 9.67835367e-01 6.47814929e-01
-6.33939326e-01 -7.86777288e-02 -5.29811382e-01 -3.93280149e-01
8.53615254e-02 3.69262010e-01 -8.52611899e-01 5.33777773e-01
9.39813256e-01 2.07487315e-01 -9.49824691e-01 1.27664793e+00
-6.15679741e-01 7.05621719e-01 -1.14921167e-01 -6.31988466e-01
3.47432345e-01 1.81695074e-02 5.91183841e-01 8.60861659e-01
4.69692275e-02 1.25356913e-01 1.28578052e-01 6.50898218e-01
-2.64841430e-02 5.22671402e-01 -4.23706055e-01 2.13496163e-01
5.03650725e-01 1.07872939e+00 -3.99414390e-01 -3.28383058e-01
-4.01939154e-01 7.92890429e-01 6.59890890e-01 3.37642759e-01
-6.22713327e-01 -6.58440769e-01 2.41817057e-01 -2.29895771e-01
3.39167029e-01 9.61986259e-02 -2.20951617e-01 -1.13663507e+00
5.26841044e-01 -9.79607046e-01 6.02699757e-01 -7.50069737e-01
-1.40775657e+00 7.90310442e-01 -2.70348340e-01 -1.12909949e+00
-1.02016389e-01 -1.85200721e-01 -5.68213463e-01 8.53338599e-01
-1.91704333e+00 -7.54848540e-01 -3.20235384e-03 5.23352981e-01
8.06581497e-01 2.64387190e-01 8.35090995e-01 3.98820728e-01
-5.51962316e-01 4.96620089e-01 1.15657210e-01 9.67321470e-02
9.92400646e-01 -1.41115034e+00 6.70647696e-02 9.68310118e-01
3.07312101e-01 6.40581667e-01 9.77459788e-01 -2.54686058e-01
-1.04916286e+00 -6.41475737e-01 1.73660970e+00 -8.72588277e-01
3.45324397e-01 -2.06858754e-01 -9.80536103e-01 1.31897911e-01
5.70112824e-01 -3.29755068e-01 8.34987938e-01 4.00289059e-01
-3.31320554e-01 -2.36427099e-01 -9.02153850e-01 7.75540769e-01
6.57737017e-01 -5.23410618e-01 -1.08314037e+00 6.36904120e-01
1.01144528e+00 -5.14062464e-01 -4.95108902e-01 3.68690401e-01
4.17382345e-02 -1.13676310e+00 1.04906642e+00 -7.69358993e-01
8.47369015e-01 -2.74898171e-01 -2.64573768e-02 -1.13622475e+00
-1.54149696e-01 -4.63209897e-01 -1.18112274e-01 1.38264251e+00
6.60348892e-01 -1.03950992e-01 4.48238730e-01 6.09088719e-01
-3.17745566e-01 -1.05644846e+00 -1.01730549e+00 -5.61053514e-01
2.85117477e-01 7.33089494e-03 5.43991804e-01 5.12321830e-01
1.67411655e-01 9.56386089e-01 -2.76003957e-01 1.38744831e-01
2.20645070e-02 3.21910977e-01 6.17037714e-01 -1.00141990e+00
-1.52509078e-01 -4.19599354e-01 1.79521546e-01 -1.33328128e+00
1.15840808e-02 -6.58247590e-01 3.55423838e-01 -1.80073595e+00
4.34592247e-01 -1.24906376e-01 -4.08673763e-01 7.19964653e-02
-8.82266819e-01 -2.96256883e-04 3.41746598e-01 1.88563719e-01
-1.41848707e+00 7.72267818e-01 1.49076068e+00 -2.85490043e-02
-1.46352515e-01 1.35521619e-02 -8.89850736e-01 4.35047030e-01
7.42454052e-01 -2.67247260e-01 -7.14084268e-01 -8.67105603e-01
2.57937133e-01 2.12774128e-01 1.77186221e-01 -8.54089737e-01
3.76955181e-01 1.00701816e-01 7.31109381e-02 -7.89695799e-01
3.01144987e-01 -4.92447555e-01 -8.70844424e-01 1.30978584e-01
-8.65576208e-01 1.86477989e-01 9.01768506e-02 4.44182992e-01
-4.94591504e-01 -5.67017436e-01 5.20510137e-01 -2.59926021e-01
-5.83145142e-01 3.81972082e-02 -2.88007110e-01 6.03828907e-01
8.48180711e-01 2.34559104e-01 -7.28981256e-01 -6.39337003e-01
-3.90529275e-01 9.34064329e-01 -2.14198208e-03 7.37223625e-01
6.91221476e-01 -1.09398293e+00 -9.35307384e-01 -2.68341362e-01
2.53383964e-01 -5.41722849e-02 5.08943498e-01 7.88716614e-01
-4.82224673e-01 7.51809716e-01 1.24887116e-01 -4.07704473e-01
-1.44412506e+00 3.88693988e-01 3.74336213e-01 -7.57446885e-01
-3.19851995e-01 1.21789122e+00 -5.71525283e-02 -6.76106870e-01
4.26294506e-01 5.59695214e-02 -9.01281297e-01 8.58321972e-03
9.29736555e-01 2.18114510e-01 3.68914217e-01 -4.34721470e-01
-2.18323499e-01 1.57896325e-01 -4.62611705e-01 -1.23510538e-02
1.04622459e+00 -4.99144942e-01 -2.89435744e-01 3.77107024e-01
1.08779836e+00 -6.77597523e-02 -7.17420459e-01 -3.99488866e-01
3.51779431e-01 -5.34024477e-01 -1.26070946e-01 -1.15067780e+00
-4.97786641e-01 1.01393664e+00 1.39933005e-01 5.34873128e-01
1.08662963e+00 5.73311709e-02 1.04333234e+00 7.08062112e-01
1.66719288e-01 -1.23590863e+00 3.40454280e-01 8.75525713e-01
1.08016574e+00 -1.30640638e+00 -1.47976458e-01 -4.93905127e-01
-7.19116867e-01 9.03733313e-01 1.07937956e+00 -6.81569055e-02
2.48865888e-01 -4.04371232e-01 5.02273202e-01 -3.49653512e-01
-1.23656094e+00 -3.06228340e-01 5.01834691e-01 1.13190494e-01
7.92985559e-01 -2.99252868e-01 -8.01155746e-01 5.18673897e-01
-3.45778853e-01 -4.51119125e-01 2.81185716e-01 8.78945053e-01
-8.33764195e-01 -1.13287520e+00 -1.95289612e-01 4.67456698e-01
-6.51574016e-01 -2.17036292e-01 -6.74037576e-01 4.45884913e-01
-2.96182483e-01 1.47365773e+00 -2.13070467e-01 -1.77271500e-01
6.00290060e-01 3.13584805e-01 2.40238234e-01 -8.19088519e-01
-1.08134902e+00 -1.99953884e-01 4.50713694e-01 -3.81463826e-01
-3.13178480e-01 -4.40967858e-01 -8.61609757e-01 -7.60235339e-02
-6.03079855e-01 7.93941855e-01 9.28169191e-02 1.18679166e+00
5.00101388e-01 6.30473912e-01 7.51246929e-01 -9.39759687e-02
-1.09661114e+00 -1.12535524e+00 -3.54896545e-01 5.75418055e-01
3.38226229e-01 -3.61192852e-01 -4.06371236e-01 -4.90787804e-01] | [11.370942115783691, 8.027669906616211] |
fa48258a-0cd7-497a-b2fc-eba75a6c5817 | wac-a-corpus-of-wikipedia-conversations-for | 2003.06190 | null | https://arxiv.org/abs/2003.06190v1 | https://arxiv.org/pdf/2003.06190v1.pdf | WAC: A Corpus of Wikipedia Conversations for Online Abuse Detection | With the spread of online social networks, it is more and more difficult to monitor all the user-generated content. Automating the moderation process of the inappropriate exchange content on Internet has thus become a priority task. Methods have been proposed for this purpose, but it can be challenging to find a suitable dataset to train and develop them. This issue is especially true for approaches based on information derived from the structure and the dynamic of the conversation. In this work, we propose an original framework, based on the Wikipedia Comment corpus, with comment-level abuse annotations of different types. The major contribution concerns the reconstruction of conversations, by comparison to existing corpora, which focus only on isolated messages (i.e. taken out of their conversational context). This large corpus of more than 380k annotated messages opens perspectives for online abuse detection and especially for context-based approaches. We also propose, in addition to this corpus, a complete benchmarking platform to stimulate and fairly compare scientific works around the problem of content abuse detection, trying to avoid the recurring problem of result replication. Finally, we apply two classification methods to our dataset to demonstrate its potential. | ['Georges Linares', 'Richard Dufour', 'Vincent Labatut', 'Noé Cecillon'] | 2020-03-13 | wac-a-corpus-of-wikipedia-conversations-for-1 | https://aclanthology.org/2020.lrec-1.173 | https://aclanthology.org/2020.lrec-1.173.pdf | lrec-2020-5 | ['abuse-detection'] | ['natural-language-processing'] | [-4.70495634e-02 2.24764515e-02 -1.16531372e-01 -1.76782325e-01
-5.49131393e-01 -7.91746914e-01 8.49341869e-01 6.47139549e-01
-4.86508399e-01 8.19798827e-01 3.85750622e-01 -1.86147094e-01
-4.19942215e-02 -4.95550632e-01 -2.36119911e-01 -5.05806088e-01
-1.28390685e-01 4.88186598e-01 4.06662196e-01 -4.71946001e-01
6.28499210e-01 3.31242114e-01 -1.74825943e+00 4.85428065e-01
9.65051472e-01 6.02193415e-01 2.07793698e-01 5.02281070e-01
-3.76221299e-01 8.13358605e-01 -9.48938787e-01 -7.65576124e-01
-2.33858183e-01 -5.97117662e-01 -1.15308881e+00 6.25099018e-02
2.83419579e-01 -7.39213526e-02 2.72439271e-02 9.76864278e-01
4.04882640e-01 1.03532663e-02 3.20475340e-01 -1.31953168e+00
3.76725309e-02 8.85215223e-01 -2.05141380e-01 4.32299405e-01
5.66087723e-01 -7.62516037e-02 9.84029651e-01 -4.69473511e-01
1.18579626e+00 7.95381367e-01 5.62351525e-01 3.98997128e-01
-9.69284654e-01 -3.30313593e-01 -1.40694097e-01 5.69038808e-01
-9.48348284e-01 -3.62938643e-01 8.75834584e-01 -7.04105496e-01
8.24587882e-01 5.56252778e-01 6.66736007e-01 1.86600745e+00
-2.41495103e-01 6.91830516e-01 1.12868524e+00 -5.40342808e-01
1.12512052e-01 4.62769091e-01 3.46958905e-01 2.75977284e-01
1.93289623e-01 -7.31826663e-01 -4.66520250e-01 -6.36653244e-01
-8.13038573e-02 -3.14374983e-01 -4.64071900e-01 -1.57965362e-01
-9.16037798e-01 8.62858474e-01 -5.04759438e-02 1.10046053e+00
-1.48164049e-01 -3.49147439e-01 9.49163020e-01 3.97700787e-01
6.09717488e-01 6.62115037e-01 -2.85420418e-01 -8.66966605e-01
-1.04348576e+00 5.35276830e-01 1.46271336e+00 5.57640791e-01
6.59895718e-01 -5.88994980e-01 1.29377827e-01 1.06577182e+00
1.19656540e-01 -5.54817133e-02 4.65716749e-01 -7.00612366e-01
5.69538176e-01 8.06688666e-01 6.67434111e-02 -1.34197176e+00
-6.13998771e-01 -3.34883392e-01 -6.83018625e-01 -1.11234583e-01
8.05275857e-01 -1.74611211e-01 4.41497415e-02 1.49117529e+00
4.83227670e-01 -1.04592144e-01 -3.53384584e-01 6.50510728e-01
5.89504898e-01 3.89443785e-01 -2.45679617e-01 -5.17559946e-01
1.36537385e+00 -8.09600592e-01 -9.31628287e-01 1.45115703e-01
9.21693206e-01 -9.41287339e-01 8.53581369e-01 7.13253796e-01
-8.20430696e-01 1.78527068e-02 -9.07678246e-01 4.14605625e-02
-6.32841527e-01 -4.06601280e-01 2.96141446e-01 6.54019058e-01
-8.01675260e-01 7.61624813e-01 -4.39713329e-01 -6.14153206e-01
1.09910078e-01 -4.81372885e-02 -5.03568888e-01 3.03156793e-01
-1.29747844e+00 1.02845120e+00 3.23875010e-01 9.77199823e-02
-3.01222622e-01 -4.66357976e-01 -3.80180329e-01 -1.68369964e-01
5.27216315e-01 -7.89692402e-02 1.29131734e+00 -1.21459746e+00
-1.40724325e+00 1.05014372e+00 1.40833214e-01 -4.16661590e-01
8.05012167e-01 -1.39690518e-01 -3.75353426e-01 6.80259839e-02
-3.73902842e-02 -1.79397464e-01 5.43711066e-01 -1.10288680e+00
-4.20647085e-01 -2.65709758e-01 1.26723677e-01 -3.72968376e-01
-6.71999872e-01 5.78116834e-01 -3.52410883e-01 -5.63071907e-01
-4.62893337e-01 -9.41767335e-01 -8.03384930e-04 -3.12578648e-01
-5.42755246e-01 -4.33287829e-01 8.86346579e-01 -1.01821029e+00
1.73100066e+00 -1.87957919e+00 5.52346230e-01 2.00845480e-01
3.39111149e-01 4.81481969e-01 2.31603354e-01 1.16517246e+00
6.80025890e-02 3.88492703e-01 -4.71835077e-01 -7.95559585e-01
-4.68283705e-02 2.50885695e-01 -1.86811537e-01 5.04161716e-01
-2.19872415e-01 4.03204113e-01 -9.58902359e-01 -5.54917753e-01
-1.20744608e-01 2.81991750e-01 -5.70534170e-01 4.63630289e-01
-3.30781341e-01 6.38930678e-01 -3.09168309e-01 2.35560521e-01
5.97136796e-01 -1.29742086e-01 4.49976146e-01 2.90588975e-01
-4.64231789e-01 5.14967382e-01 -9.34150755e-01 1.77475441e+00
-4.10956472e-01 9.70291138e-01 4.76416409e-01 -9.98425782e-01
7.46688008e-01 3.71343374e-01 5.02856016e-01 -4.40230638e-01
3.15845847e-01 4.21206981e-01 1.63056165e-01 -9.86991167e-01
5.96751451e-01 2.31767207e-01 -2.12840945e-01 8.14844847e-01
1.62401870e-01 2.63603002e-01 5.79337239e-01 4.19850856e-01
1.29221964e+00 -1.79837406e-01 3.90981644e-01 -1.83728233e-01
7.87557781e-01 -1.77887917e-01 2.40988716e-01 5.77211678e-01
-2.41787121e-01 5.07211208e-01 1.17567956e+00 -4.29328412e-01
-1.08898914e+00 -2.95508116e-01 -2.87901968e-01 9.76752818e-01
-2.71430880e-01 -9.23404574e-01 -1.08775854e+00 -9.23423886e-01
-1.41947523e-01 5.37680507e-01 -6.21028066e-01 3.42136413e-01
-7.52672493e-01 -7.18693733e-01 6.84337258e-01 -2.85117179e-01
2.90235013e-01 -1.16736734e+00 -4.29908633e-01 4.03125435e-01
-8.16125989e-01 -1.18358636e+00 1.16616338e-01 -2.09053010e-01
-4.17920917e-01 -1.35665941e+00 -4.13612932e-01 -1.22338317e-01
8.11790228e-02 -1.60050709e-02 1.14283967e+00 6.69135571e-01
-1.77438512e-01 2.13820159e-01 -9.28763270e-01 -2.72481889e-01
-9.90535855e-01 5.57145417e-01 -1.73767179e-01 1.87514246e-01
1.79692701e-01 -8.80653858e-01 -3.53627503e-01 4.82166111e-01
-1.09490943e+00 -2.29708850e-01 6.24360405e-02 6.25693738e-01
-3.17340881e-01 -5.28902829e-01 5.91617465e-01 -1.27987552e+00
8.01916242e-01 -1.06638479e+00 -2.72239059e-01 -3.55375074e-02
-4.71539557e-01 -3.23780090e-01 5.91735721e-01 -2.70969510e-01
-8.86427701e-01 -5.72343946e-01 -6.80077732e-01 7.09491819e-02
-2.00424030e-01 4.92576897e-01 -4.33626398e-02 1.97694495e-01
8.50647032e-01 -5.05768694e-02 1.37951627e-01 -9.38577473e-01
5.02860323e-02 1.21382356e+00 2.25327183e-02 -3.66728157e-01
5.44909060e-01 3.34212869e-01 -4.16252017e-01 -1.28504872e+00
-7.05199122e-01 -8.25833917e-01 -7.24987924e-01 -3.29169989e-01
6.57617092e-01 -3.12525898e-01 -6.32886589e-01 7.02119231e-01
-1.51823390e+00 -2.79722363e-01 1.25044227e-01 7.78735504e-02
-3.68300676e-01 1.00471056e+00 -7.55705297e-01 -9.12169278e-01
-2.60773271e-01 -9.78831112e-01 6.41554117e-01 -2.53890455e-01
-6.51466131e-01 -1.03019392e+00 6.20083451e-01 6.05831683e-01
6.32898152e-01 4.52453643e-01 6.46414280e-01 -1.14937103e+00
-2.83299685e-01 -1.07948966e-01 1.25805646e-01 4.69062448e-01
-1.01351254e-01 3.28966796e-01 -1.03955543e+00 -5.48972227e-02
1.71592221e-01 -3.37107986e-01 6.35936022e-01 -4.86559391e-01
1.12796342e+00 -5.66113651e-01 -2.83482224e-01 -1.01349931e-02
1.01595843e+00 -3.36940557e-01 8.61235082e-01 6.42932951e-01
4.29253370e-01 1.03944421e+00 3.71884078e-01 6.89540446e-01
2.06751600e-01 1.09548295e+00 6.09919667e-01 4.34460402e-01
1.30534530e-01 4.88610901e-02 2.66224861e-01 1.09305024e+00
-2.28545845e-01 -5.57485521e-01 -1.02278233e+00 3.46452534e-01
-1.96115732e+00 -1.14085042e+00 -7.34427869e-01 2.03629279e+00
8.60105097e-01 1.45921931e-01 4.97947335e-01 4.00904596e-01
5.83883464e-01 3.30977410e-01 3.11000407e-01 -6.34091675e-01
-9.77060385e-03 -7.87652880e-02 -1.16213910e-01 4.61093426e-01
-8.86870205e-01 3.49069953e-01 5.57728052e+00 7.73572564e-01
-9.51696754e-01 4.79440212e-01 2.49812409e-01 2.79981531e-02
-2.50980526e-01 1.13461660e-02 -7.06154287e-01 1.00769281e+00
1.22397065e+00 2.35231906e-01 3.54608774e-01 8.00944269e-01
2.93898374e-01 -2.30136007e-01 -8.59560728e-01 8.80116761e-01
4.37708080e-01 -1.24481249e+00 -4.29997832e-01 1.80488065e-01
3.78770053e-01 1.81343049e-01 -5.64935386e-01 1.71943799e-01
-3.25831503e-01 -6.16305768e-01 5.72954297e-01 4.89209473e-01
1.50263816e-01 -4.78233159e-01 1.00327933e+00 7.85468042e-01
-4.45524186e-01 -6.11842833e-02 -4.06531151e-03 -2.48856619e-01
3.62097114e-01 8.46499503e-01 -8.25404763e-01 5.40482700e-01
6.94092155e-01 7.99353242e-01 -6.94613397e-01 1.23300934e+00
-1.98585317e-01 8.48875582e-01 -3.50974679e-01 -3.17720026e-01
1.00914665e-01 -2.59354621e-01 9.90405619e-01 1.68422580e+00
1.11772977e-01 -3.86367530e-01 -1.32371679e-01 5.46100855e-01
-1.52582809e-01 4.96737182e-01 -6.13942742e-01 -5.89778647e-02
2.86854595e-01 1.60046780e+00 -5.53416908e-01 -1.29410625e-01
-3.93499702e-01 9.14258182e-01 7.71421611e-01 -1.64812788e-01
-6.86226189e-01 -1.70038193e-01 6.76436126e-01 3.84157658e-01
7.10234568e-02 -2.72112340e-01 2.81599760e-01 -1.30411816e+00
2.64703780e-01 -1.12231469e+00 4.66724813e-01 -2.57323682e-01
-1.39938116e+00 6.74657941e-01 9.48297381e-02 -9.84299719e-01
-4.84409451e-01 -3.98427069e-01 -6.34901702e-01 5.21392882e-01
-1.33117330e+00 -8.47516418e-01 -3.48303646e-01 2.71398872e-01
4.23440576e-01 2.54560769e-01 7.75862396e-01 6.89842463e-01
-6.79466605e-01 3.69147092e-01 9.97334942e-02 1.24565490e-01
8.55995536e-01 -1.20108676e+00 4.65197638e-02 4.87213492e-01
9.22098011e-02 4.44696873e-01 1.01866162e+00 -4.68073606e-01
-1.18663597e+00 -4.93730813e-01 1.29430807e+00 -9.29908812e-01
1.18158352e+00 -8.57610106e-01 -1.39682770e+00 4.46490169e-01
5.26505470e-01 -4.48730648e-01 8.99995327e-01 4.33973372e-01
-4.21874911e-01 2.18567029e-01 -1.04261863e+00 3.20939958e-01
1.14447606e+00 -4.43615466e-01 -6.85896575e-01 6.11493468e-01
2.33793080e-01 -3.46409798e-01 -7.53646255e-01 -6.66075647e-02
4.35512304e-01 -1.48772192e+00 3.52131605e-01 -4.76571530e-01
5.96850097e-01 1.44966155e-01 3.56769085e-01 -1.24220181e+00
1.36045784e-01 -9.94641602e-01 -1.31617248e-01 1.69741440e+00
3.61682057e-01 -6.81196809e-01 4.61212486e-01 3.37638825e-01
-8.41121450e-02 -5.43014109e-01 -1.25437474e+00 -3.66008371e-01
1.03420699e-02 -4.89842862e-01 1.51934370e-01 1.18038368e+00
5.59891760e-01 3.62507015e-01 -5.47815144e-01 -4.16854322e-01
1.50580898e-01 -9.03468430e-02 1.10272551e+00 -1.55336821e+00
-4.08390075e-01 -6.12654448e-01 -4.72541183e-01 -5.45462728e-01
2.70837307e-01 -8.79725099e-01 -2.80680865e-01 -1.20696163e+00
2.94162363e-01 -3.97436708e-01 2.45696664e-01 7.40719438e-02
9.91332829e-02 1.75511867e-01 1.64267138e-01 4.12051499e-01
-7.74879694e-01 2.25094363e-01 9.26164448e-01 2.63447225e-01
3.60571630e-02 9.98495240e-03 -3.74680549e-01 7.57964730e-01
8.68079424e-01 -5.46904266e-01 7.14282542e-02 5.62619865e-02
7.96598494e-01 -2.74921179e-01 1.71639144e-01 -8.22746396e-01
1.10239625e-01 1.58660889e-01 -4.35538411e-01 -3.53791744e-01
2.75879115e-01 -7.24927485e-01 8.57820734e-02 2.98663855e-01
-2.72529215e-01 6.23371592e-03 -1.27960175e-01 5.57308435e-01
-2.07510918e-01 -8.66459191e-01 5.62750816e-01 -2.28127688e-01
-1.73383176e-01 -1.63467899e-01 -7.06966281e-01 2.57329673e-01
9.95290279e-01 1.11242250e-01 -4.12749320e-01 -4.29556578e-01
-7.23769546e-01 3.54206972e-02 5.66424191e-01 6.29157126e-01
3.11556682e-02 -7.44093359e-01 -7.89603531e-01 -1.98501214e-01
2.88014263e-01 -4.36667383e-01 3.64888251e-01 1.13421786e+00
-5.24491429e-01 1.39589041e-01 -1.67422011e-01 -4.26249593e-01
-1.52006483e+00 6.37859285e-01 2.05285147e-01 -6.21126711e-01
-4.45290178e-01 3.56623203e-01 -5.93193173e-01 -4.14382875e-01
1.58835843e-01 -2.98312632e-03 -6.16641223e-01 7.28661001e-01
9.19735312e-01 6.29573345e-01 4.21847761e-01 -6.74560368e-01
-1.77371532e-01 -2.16121040e-02 -5.98158613e-02 4.62635122e-02
1.54736495e+00 -1.98813662e-01 -7.20889390e-01 7.48950243e-01
1.35154164e+00 5.83647907e-01 -5.06762862e-01 -3.89288776e-02
4.06158060e-01 -5.73004246e-01 -4.09684360e-01 -6.02353156e-01
-6.33546233e-01 7.29497671e-01 3.31403501e-02 1.13649678e+00
5.91549814e-01 1.05763197e-01 6.85023308e-01 1.54765606e-01
3.81194532e-01 -1.10300815e+00 1.48639172e-01 5.82202852e-01
1.10422885e+00 -1.21053255e+00 -1.04664870e-01 -5.78193843e-01
-3.48146319e-01 1.31480873e+00 3.23668808e-01 3.68110053e-02
6.37876570e-01 1.10375002e-01 -7.31327534e-02 -2.93850273e-01
-6.09904110e-01 -5.71701974e-02 -7.30572939e-02 2.39785329e-01
4.79022592e-01 -3.03064853e-01 -1.03539836e+00 4.47734922e-01
-1.19224668e-01 -7.35579431e-02 8.94987702e-01 8.16829383e-01
-3.01739454e-01 -1.68754971e+00 -2.29959011e-01 3.54685873e-01
-8.04814517e-01 2.08135590e-01 -8.19226444e-01 9.08387959e-01
7.25836307e-02 1.00076997e+00 -1.21097542e-01 -1.75093278e-01
2.32276157e-01 3.58891994e-01 1.82187334e-01 -5.76581478e-01
-1.18107295e+00 -2.95639098e-01 7.68779397e-01 -5.28261423e-01
-7.38858998e-01 -9.09908414e-01 -6.00285292e-01 -5.65012932e-01
-4.04058307e-01 4.91539508e-01 8.59574616e-01 9.55843866e-01
4.12057847e-01 -1.51112117e-02 6.25940979e-01 -8.20702255e-01
-4.92452741e-01 -1.25394666e+00 -3.82003993e-01 8.59229922e-01
2.16750368e-01 -5.81010818e-01 -7.44749904e-01 -2.92706698e-01] | [8.601151466369629, 10.374651908874512] |
a4f73cdb-1719-44c2-808c-2e9b8f336097 | the-geometry-of-multilingual-language-model | 2205.10964 | null | https://arxiv.org/abs/2205.10964v2 | https://arxiv.org/pdf/2205.10964v2.pdf | The Geometry of Multilingual Language Model Representations | We assess how multilingual language models maintain a shared multilingual representation space while still encoding language-sensitive information in each language. Using XLM-R as a case study, we show that languages occupy similar linear subspaces after mean-centering, evaluated based on causal effects on language modeling performance and direct comparisons between subspaces for 88 languages. The subspace means differ along language-sensitive axes that are relatively stable throughout middle layers, and these axes encode information such as token vocabularies. Shifting representations by language means is sufficient to induce token predictions in different languages. However, we also identify stable language-neutral axes that encode information such as token positions and part-of-speech. We visualize representations projected onto language-sensitive and language-neutral axes, identifying language family and part-of-speech clusters, along with spirals, toruses, and curves representing token position information. These results demonstrate that multilingual language models encode information along orthogonal language-sensitive and language-neutral axes, allowing the models to extract a variety of features for downstream tasks and cross-lingual transfer learning. | ['Benjamin K. Bergen', 'Zhuowen Tu', 'Tyler A. Chang'] | 2022-05-22 | null | null | null | null | ['xlm-r'] | ['natural-language-processing'] | [-5.51729441e-01 -1.05519213e-01 -6.59558058e-01 -3.93927336e-01
-7.83586085e-01 -1.18682837e+00 8.84561658e-01 2.60223359e-01
-6.25517845e-01 3.13931525e-01 1.05379760e+00 -5.01592934e-01
-2.82022525e-02 -3.80583078e-01 -5.54218709e-01 -4.18472499e-01
-2.71352857e-01 4.04492915e-01 -3.78734231e-01 -3.65227282e-01
2.84668915e-02 3.81280929e-01 -1.08218491e+00 5.86036444e-01
6.70400798e-01 -7.00310767e-02 1.55836195e-01 7.48549625e-02
-5.01251459e-01 3.82408768e-01 -3.69469255e-01 -1.89992845e-01
2.81161904e-01 -2.28229329e-01 -4.41384822e-01 -3.95996839e-01
6.16954207e-01 2.30580688e-01 -3.54715824e-01 9.58908498e-01
1.49629295e-01 -2.55443960e-01 1.03146982e+00 -9.10843968e-01
-1.13551593e+00 1.31404006e+00 -5.13524711e-01 1.41968623e-01
3.43729734e-01 1.48958951e-01 1.48970938e+00 -1.19392800e+00
1.11842990e+00 1.82240748e+00 7.45402694e-01 3.38090390e-01
-1.98176253e+00 -6.56296253e-01 5.45997858e-01 -2.38952875e-01
-1.49300528e+00 -5.51802099e-01 7.27942288e-01 -1.11767209e+00
1.17304134e+00 3.16563845e-01 7.59029567e-01 1.08995235e+00
5.16129196e-01 6.93403959e-01 1.18958247e+00 -3.80128175e-01
-1.35187179e-01 4.46255386e-01 3.85524929e-01 4.04475629e-01
2.38290921e-01 1.66646764e-01 -7.59857118e-01 -2.20666155e-01
4.72614437e-01 -1.12661548e-01 -8.30590576e-02 -7.15868473e-01
-1.86571848e+00 8.33406627e-01 4.10751492e-01 5.83647013e-01
-1.14282861e-01 -4.80309688e-03 5.95262647e-01 3.62443686e-01
6.04235709e-01 6.33054674e-01 -7.23303616e-01 2.68225163e-01
-1.02337742e+00 1.42484397e-01 3.45932424e-01 1.12208343e+00
9.64134693e-01 3.73910755e-01 -2.52544641e-01 1.10587883e+00
5.83166301e-01 5.49389422e-01 6.91180944e-01 -6.09957099e-01
5.70144117e-01 5.63194454e-01 -3.68401915e-01 -4.75093722e-01
-7.05006778e-01 -5.57727635e-01 -6.05788469e-01 -1.67248443e-01
4.78745699e-01 -1.04569323e-01 -5.83027124e-01 2.36893225e+00
-2.00277671e-01 -5.92980742e-01 3.14684331e-01 6.49424255e-01
4.88439649e-01 6.51405573e-01 5.84305465e-01 -7.83077478e-02
1.62798703e+00 -5.26286125e-01 -4.54249680e-01 -2.14467987e-01
1.25427270e+00 -8.66438687e-01 1.33384657e+00 -1.82727680e-01
-1.04879904e+00 -5.42403817e-01 -6.91682756e-01 -5.42213738e-01
-5.97350359e-01 1.81230009e-01 5.25432467e-01 4.87873822e-01
-1.22283590e+00 1.83186278e-01 -7.32265174e-01 -5.60686946e-01
-8.98385420e-02 8.74395203e-03 -6.93186402e-01 2.03223765e-01
-1.39919770e+00 8.97917211e-01 3.88584405e-01 -3.27657670e-01
-7.77264357e-01 -1.17040062e+00 -1.23853993e+00 1.46235108e-01
-4.23816293e-01 -4.46818888e-01 8.83276284e-01 -6.13242626e-01
-9.56556737e-01 1.06388748e+00 -1.85293734e-01 -8.41534883e-02
4.05554533e-01 1.01883024e-01 -5.52003682e-01 -3.71344835e-01
4.37399864e-01 9.90889251e-01 2.97689587e-01 -1.19487512e+00
-3.33435267e-01 -5.18783629e-01 -3.81298006e-01 4.56056029e-01
-4.83834505e-01 5.50502121e-01 1.47246872e-03 -8.96508157e-01
5.70237875e-01 -1.04368818e+00 -8.19945186e-02 -1.53220668e-01
-3.95663530e-01 -2.93081939e-01 1.99223757e-01 -8.49694192e-01
1.20949054e+00 -2.12828302e+00 1.95199355e-01 2.76618004e-01
2.44088098e-01 -5.22605896e-01 -3.08561355e-01 6.57342911e-01
-5.15285790e-01 4.77937430e-01 -1.64171346e-02 -2.98019379e-01
2.76865959e-01 4.51404266e-02 -4.82323468e-01 7.35940576e-01
2.91427910e-01 9.18281555e-01 -9.17751372e-01 -3.72572213e-01
1.04650326e-01 6.33383512e-01 -8.25865090e-01 -4.93161112e-01
2.05771569e-02 3.90851498e-01 7.02665076e-02 4.15998161e-01
4.78968978e-01 3.73380363e-01 4.64214712e-01 1.67485494e-02
-5.87378263e-01 7.31153429e-01 -8.17412674e-01 1.66102707e+00
-4.45088208e-01 8.01424146e-01 1.43790007e-01 -5.19885302e-01
1.02552617e+00 2.49224111e-01 1.72002792e-01 -3.94176215e-01
-3.87049466e-01 3.48429382e-01 5.95056176e-01 1.89282559e-02
7.14984298e-01 -3.10141534e-01 -4.73612010e-01 5.56402564e-01
3.73838335e-01 -7.71465898e-02 1.03844754e-01 3.05934310e-01
4.98963177e-01 1.79180354e-01 2.56617486e-01 -1.08014810e+00
2.54211158e-01 -6.58231452e-02 6.25261128e-01 3.38161290e-01
-1.43020809e-01 2.30170041e-01 7.80242264e-01 -3.57745945e-01
-1.24559426e+00 -1.49408031e+00 -7.17241228e-01 1.64715254e+00
-4.93654698e-01 -8.38127553e-01 -3.62742335e-01 -2.85418242e-01
2.93850690e-01 1.18291986e+00 -6.99487507e-01 -2.86970407e-01
-5.95052183e-01 -5.46962500e-01 6.55673921e-01 4.72239166e-01
-5.10770261e-01 -9.42004681e-01 -3.31005864e-02 -3.81801277e-02
-5.01631722e-02 -6.58031702e-01 -5.91804266e-01 6.15817964e-01
-8.55498135e-01 -3.99424314e-01 -9.42152560e-01 -8.52341831e-01
6.60982609e-01 1.52350888e-01 1.36077368e+00 -5.07176101e-01
1.74368955e-02 2.75835246e-01 1.93186387e-01 -2.84943283e-01
-6.51483119e-01 2.68174887e-01 5.04002988e-01 -3.75136077e-01
5.27873755e-01 -2.51975596e-01 -1.20359562e-01 1.39792934e-01
-4.67866808e-01 -1.41725376e-01 2.89693683e-01 6.41850889e-01
5.15719056e-01 -6.48601413e-01 2.45990679e-01 -6.64697349e-01
7.66482413e-01 -6.91578209e-01 -4.08766508e-01 2.91786730e-01
-3.14766496e-01 3.99610639e-01 5.01653910e-01 -4.09556359e-01
-7.85702348e-01 -8.30583498e-02 3.66056353e-01 -2.73568958e-01
-6.58627227e-02 5.16404986e-01 -5.68446144e-02 6.16106093e-01
8.10442090e-01 3.27436209e-01 -1.40056387e-01 -6.42094910e-01
9.09435570e-01 5.16972363e-01 2.07857594e-01 -1.03757203e+00
6.79732442e-01 2.26842508e-01 -5.15480101e-01 -1.08174992e+00
-2.26827815e-01 -4.47711200e-01 -1.21024251e+00 1.82021052e-01
9.10083473e-01 -1.39562392e+00 -2.48979583e-01 3.84532474e-02
-1.06448853e+00 -9.50318947e-02 -4.46504623e-01 8.21169734e-01
-4.82401580e-01 6.43353723e-03 -7.51009762e-01 -5.46319127e-01
-1.97799616e-02 -1.35127687e+00 1.12688231e+00 -4.63128537e-01
-7.76124001e-01 -1.24198699e+00 4.00951654e-01 -1.83967710e-01
1.33651823e-01 -2.90281624e-01 1.50312734e+00 -8.43169451e-01
-3.37866902e-01 2.52075762e-01 -7.09530199e-03 -6.88018501e-02
1.21942185e-01 6.17215671e-02 -8.52225542e-01 -5.67404568e-01
-4.23296869e-01 -1.50583282e-01 1.10156786e+00 4.93673235e-01
5.41733980e-01 -2.58396685e-01 -4.08339769e-01 8.45055997e-01
9.56581652e-01 -9.48926434e-02 5.46959080e-02 2.30489790e-01
7.27936745e-01 1.07899964e+00 -1.74769029e-01 -1.45905271e-01
5.40044725e-01 5.57242453e-01 -4.82449353e-01 -1.72112420e-01
-4.82826754e-02 -6.04224741e-01 8.52520883e-01 1.23631275e+00
4.23668683e-01 3.67957443e-01 -1.39235258e+00 1.00251257e+00
-1.44719720e+00 -8.54106963e-01 -3.77626717e-02 2.16843605e+00
1.07868612e+00 1.49092181e-02 1.33388728e-01 -4.33972269e-01
5.42514563e-01 2.51165271e-01 -4.27147746e-01 -5.88501692e-01
-5.25243998e-01 -2.69949615e-01 3.82286400e-01 8.28322828e-01
-7.11956739e-01 1.48886251e+00 7.44845295e+00 3.80851746e-01
-1.19225657e+00 3.45768481e-02 4.43030149e-01 -2.49291137e-01
-9.81538475e-01 1.05191231e-01 -7.72038937e-01 1.75120786e-01
9.34423625e-01 -5.39571404e-01 3.39925587e-01 8.30174327e-01
2.22895950e-01 4.20501143e-01 -1.52140749e+00 7.05616236e-01
-7.00251460e-02 -1.07746673e+00 4.14684683e-01 2.13520572e-01
7.23194063e-01 4.53115553e-01 3.47300947e-01 5.80056310e-01
7.18687832e-01 -1.07195568e+00 1.22250450e+00 4.21643317e-01
1.18972445e+00 -5.78898549e-01 4.64817360e-02 1.20029084e-01
-1.37679219e+00 1.01612844e-01 -6.32901549e-01 -8.19728225e-02
-7.94096515e-02 1.04840532e-01 -9.80956078e-01 1.23480313e-01
5.69076896e-01 8.87001932e-01 -7.08040059e-01 3.90388429e-01
-1.44199103e-01 5.86377144e-01 -8.86380149e-04 3.31506431e-01
2.37431139e-01 -3.10791790e-01 5.74264169e-01 1.76539600e+00
3.19869369e-01 -5.22649527e-01 2.63264835e-01 1.15553093e+00
2.12660059e-02 6.11945987e-01 -1.02520275e+00 -1.98046535e-01
5.27087629e-01 1.03623211e+00 -6.61355376e-01 -2.98736304e-01
-5.61030388e-01 4.45785046e-01 5.32488763e-01 6.55949473e-01
-3.64814073e-01 -8.59669819e-02 1.34998047e+00 1.68173447e-01
-3.15811098e-01 -6.32039785e-01 -5.04115522e-01 -1.42087996e+00
-2.79171437e-01 -8.14129829e-01 2.99062401e-01 -5.48797131e-01
-1.26161814e+00 5.08707821e-01 9.42442939e-02 -1.06775165e+00
-5.01371741e-01 -8.45943332e-01 -1.18224174e-01 1.40691388e+00
-1.02328348e+00 -1.29321063e+00 6.31109834e-01 6.82343483e-01
3.92417848e-01 -7.23936796e-01 1.01523006e+00 -7.78204203e-02
-5.17650247e-01 7.54024148e-01 5.41446149e-01 5.33781290e-01
9.46446061e-01 -1.38297975e+00 6.69644237e-01 5.44314384e-01
7.86414802e-01 1.50742435e+00 4.24863905e-01 -7.48691797e-01
-1.08486378e+00 -9.68320370e-01 1.18635213e+00 -7.91583300e-01
1.03518462e+00 -9.93739545e-01 -9.36786532e-01 1.29549825e+00
3.54617506e-01 -2.30544031e-01 1.12631238e+00 9.42120731e-01
-1.07306576e+00 1.16683789e-01 -5.77896595e-01 1.02139103e+00
9.93342876e-01 -1.18813276e+00 -7.68246830e-01 4.11165982e-01
8.76183510e-01 1.14096522e-01 -8.42633963e-01 7.39210227e-04
5.66397607e-01 -6.52969301e-01 8.78159285e-01 -1.10932517e+00
2.72082150e-01 -2.21310899e-01 -4.78203893e-01 -1.84386408e+00
-7.83408344e-01 -3.18703562e-01 6.55937731e-01 1.17883050e+00
8.04568648e-01 -6.34169400e-01 1.71294004e-01 3.25567484e-01
-1.92116991e-01 -3.21188390e-01 -9.95392799e-01 -7.45790958e-01
1.01162708e+00 -5.41914344e-01 5.27837813e-01 1.49847472e+00
4.94479775e-01 6.10382855e-01 1.80986032e-01 -2.55718499e-01
4.79817301e-01 -8.39285553e-02 2.75111437e-01 -1.17392111e+00
1.23759776e-01 -9.28551137e-01 -4.07346904e-01 -6.35437727e-01
6.03484392e-01 -1.95358086e+00 -4.01765376e-01 -1.30965924e+00
4.83570784e-01 -4.71012503e-01 -4.98834580e-01 6.79294825e-01
1.02365509e-01 -1.65320158e-01 5.57107985e-01 5.48124075e-01
-8.27810317e-02 5.20701349e-01 8.59925628e-01 -1.36367753e-01
-3.33706141e-01 -6.78021491e-01 -8.94954324e-01 9.79676962e-01
6.12696588e-01 -3.13811302e-01 -3.81719619e-01 -7.53256440e-01
2.17224643e-01 -2.33545586e-01 -3.20766494e-02 -5.56676805e-01
-2.70941913e-01 -1.73631236e-01 7.83348322e-01 -4.59262609e-01
1.34151086e-01 -5.45086205e-01 -2.68985201e-02 6.39144957e-01
-7.54702389e-01 4.97795880e-01 4.15530622e-01 1.00180253e-01
-1.61238194e-01 3.37501734e-01 6.46991551e-01 -2.15510845e-01
-4.44650114e-01 -1.42105073e-01 -5.72455764e-01 2.71999806e-01
8.55791271e-01 -9.70383734e-03 -3.68023664e-01 -1.61818326e-01
-8.68042409e-01 1.78280339e-01 6.20349765e-01 9.17381763e-01
1.46389097e-01 -1.64486754e+00 -1.18141961e+00 6.42198324e-01
5.94729662e-01 -7.41935968e-01 -8.33016634e-02 4.39266056e-01
-2.31512412e-01 7.86835551e-01 -2.93393254e-01 -9.59172845e-01
-9.75094438e-01 6.31341100e-01 3.54064673e-01 -1.18315227e-01
-3.22436064e-01 8.16365659e-01 9.14961934e-01 -1.27906013e+00
-9.57138315e-02 -7.68465519e-01 -1.57801583e-01 5.27438521e-01
1.64493874e-01 2.48424355e-02 -2.29399532e-01 -1.17233360e+00
-5.91704965e-01 6.34598315e-01 -2.55954593e-01 -4.57887411e-01
1.03560531e+00 -2.29230225e-02 -3.53285283e-01 1.14373446e+00
1.05791926e+00 5.69317818e-01 -8.98799360e-01 -3.39996576e-01
4.02361810e-01 8.89086723e-03 -2.64925420e-01 -7.04047501e-01
-5.74581146e-01 1.12935567e+00 3.75408024e-01 -1.68792129e-01
3.24082047e-01 2.50990063e-01 -1.81450769e-02 1.45518035e-01
3.10874879e-01 -1.00584579e+00 -4.47461158e-01 8.79382730e-01
1.22505832e+00 -8.93381655e-01 -2.44643077e-01 -7.56234974e-02
-6.80306971e-01 8.64821672e-01 5.69381237e-01 1.97880134e-01
7.67313182e-01 2.86046296e-01 4.45354998e-01 6.40887097e-02
-8.64239752e-01 6.57441542e-02 5.11018753e-01 3.90727073e-01
1.33446801e+00 7.69841194e-01 -4.05239463e-01 4.67302471e-01
-9.64403093e-01 -9.67844009e-01 1.40673742e-01 2.49065906e-01
-1.16511315e-01 -1.13436818e+00 -7.30590284e-01 2.98848122e-01
-2.15263873e-01 -6.73930287e-01 -4.62101221e-01 1.00394952e+00
1.71394736e-01 5.18371224e-01 6.79072976e-01 -1.72293276e-01
1.19084559e-01 7.77668953e-01 2.34496847e-01 -9.12840843e-01
-5.64801335e-01 4.27032024e-01 -7.89683610e-02 -2.31366381e-01
2.36425236e-01 -1.14906967e+00 -1.26147878e+00 -7.40697831e-02
1.92247078e-01 4.93343398e-02 5.68476140e-01 4.66557622e-01
3.50421339e-01 4.86792654e-01 1.56920910e-01 -7.25053847e-01
-5.34098744e-01 -1.28908968e+00 -7.04477727e-01 5.12972951e-01
3.92327905e-01 -3.83157641e-01 -2.17428207e-01 -4.37337458e-02] | [10.916180610656738, 9.875699996948242] |
b00f755c-8bde-4c4b-b3d6-b41692c2d3ef | review-on-dna-strand-algebra-and-its | 1903.04260 | null | https://arxiv.org/abs/1903.04260v2 | https://arxiv.org/pdf/1903.04260v2.pdf | Review on DNA Strand Algebra and its Application | Several technological limitations of traditional silicon based computing are leading towards the paradigm shift, from silicon to carbon, in computational world. Among the unconventional modes of computing evolved in past several decades, DNA computing has been considered to be quite promising in solving computational and reasoning problems by using DNA strands. Along with the sequential operations, the huge parallelism of DNA computing methodologies engaging numerous numbers of DNA strands induce the consideration of concurrent high-level formalisms. In this paper we have reviewed the algebraic explanation of concurrent DNA processes using DNA strand algebra, process calculus and DNA strand graph. We have demonstrated the application of syntax and semantics of the illustrated methodologies in the domains of reasoning and theorem proving with resolution refutation. Finally, we have presented DNA cryptography as one of the prominent areas for the future scope of research work where DNA strand algebra can be used as formal modelling tool to authenticate the security, logic and reasoning of the existing protocols. | ['Kumar S. Ray', 'Mandrita Mondal'] | 2019-03-04 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [ 5.64621806e-01 4.93099719e-01 2.58150309e-01 4.60621640e-02
2.95413464e-01 -9.67017174e-01 1.33959150e+00 4.60153192e-01
-3.94245565e-01 6.23729706e-01 9.89657938e-02 -1.03793228e+00
-2.50725657e-01 -1.03520060e+00 -4.55442637e-01 -8.25774312e-01
-2.02136561e-01 4.73121226e-01 4.92145807e-01 -2.82155901e-01
7.77643025e-01 5.30165493e-01 -1.63853157e+00 3.30634028e-01
1.75305367e-01 4.92727101e-01 -3.91119331e-01 1.09177077e+00
-4.95063215e-01 8.84262502e-01 -2.43149310e-01 -6.61573708e-01
1.16566680e-01 -4.80075210e-01 -7.74795830e-01 -4.26243752e-01
-4.66098398e-01 2.08096266e-01 -2.74851412e-01 9.97089624e-01
1.46319121e-01 -5.47113895e-01 4.90323305e-01 -1.67189193e+00
-2.22754732e-01 9.32797849e-01 -2.95560867e-01 -1.41168050e-02
6.76257789e-01 3.50708932e-01 4.14874226e-01 -5.16179621e-01
9.91089821e-01 1.18907678e+00 5.56256294e-01 5.77526152e-01
-9.33714986e-01 -4.55875307e-01 -5.12635231e-01 1.58190608e-01
-1.25312638e+00 -1.67292245e-02 3.55426639e-01 -3.89578432e-01
1.18468654e+00 4.98962402e-01 8.94714832e-01 8.22467506e-01
9.81864095e-01 2.11163551e-01 1.39208663e+00 -6.42290413e-01
4.17586058e-01 1.32474333e-01 6.55822754e-01 6.39186740e-01
1.12255752e+00 3.57086927e-01 -4.70208317e-01 -5.87436795e-01
3.34326357e-01 -1.18796244e-01 2.11925492e-01 -3.62980574e-01
-1.49314094e+00 4.46488440e-01 -4.59604919e-01 5.93872905e-01
-1.04809076e-01 7.75238633e-01 8.62832308e-01 7.72649497e-02
-4.63267267e-01 -9.19784382e-02 -8.68144333e-02 -2.19906196e-01
-2.91025609e-01 6.02806330e-01 1.24959350e+00 8.86399865e-01
4.58328605e-01 -1.57574303e-02 2.89168119e-01 -8.74588072e-01
9.54182327e-01 4.38733459e-01 -8.86414647e-02 -8.76448035e-01
-1.62872285e-01 8.40526402e-01 -2.93576494e-02 -7.41520524e-01
-2.91320115e-01 -1.07370488e-01 -7.12445676e-01 4.96818572e-01
5.95210314e-01 5.14638498e-02 -7.26035476e-01 9.72522259e-01
4.24418658e-01 1.35321423e-01 8.87920439e-01 4.60939080e-01
3.11089098e-01 9.33722198e-01 2.79333115e-01 -1.80284351e-01
1.68558836e+00 -1.47484332e-01 -7.64625728e-01 5.72854698e-01
5.88187099e-01 -6.51447177e-01 5.40819354e-02 7.66227245e-01
-1.18160403e+00 1.56840146e-01 -1.24593461e+00 2.14301571e-01
-5.69753528e-01 -8.99988532e-01 1.21571517e+00 1.40327168e+00
-1.19483006e+00 8.63867551e-02 -7.54654408e-01 -4.31203187e-01
9.97336358e-02 5.74294448e-01 -3.73326428e-02 -7.72163644e-02
-1.15051234e+00 7.12417364e-01 7.35939205e-01 2.56783664e-01
-8.57639074e-01 -4.29652095e-01 -2.63248205e-01 1.30159959e-01
4.35549676e-01 -6.96882784e-01 8.62076879e-01 -3.46813172e-01
-1.36902487e+00 9.54801857e-01 3.01867247e-01 -7.17558324e-01
4.72869992e-01 6.47977948e-01 -4.67087865e-01 -5.91692589e-02
-4.65188980e-01 1.93057299e-01 1.82549860e-02 -1.18293154e+00
-4.34411198e-01 -4.08956796e-01 9.30154603e-03 -4.29750979e-01
4.34786350e-01 2.22098872e-01 3.34046483e-01 1.92848146e-01
2.41456464e-01 -9.98856485e-01 -2.97588199e-01 -2.09387064e-01
-7.33918771e-02 -3.68395716e-01 6.07413828e-01 1.56606019e-01
6.68842018e-01 -1.83179545e+00 1.08665548e-01 8.15794051e-01
2.21401036e-01 2.72719622e-01 4.82390791e-01 1.30440950e+00
2.06421211e-01 4.92426246e-01 -4.21827473e-02 9.27321255e-01
4.38744456e-01 1.62460670e-01 -2.96321601e-01 6.02494895e-01
-1.08836308e-01 8.30642104e-01 -8.02615106e-01 -3.58707786e-01
1.33011304e-02 4.98204589e-01 -2.82977462e-01 -3.25212419e-01
-6.82333708e-01 2.27663130e-01 -3.78387749e-01 5.69573581e-01
8.90820682e-01 6.73582628e-02 1.00118232e+00 8.54895860e-02
-6.68586910e-01 -2.35926434e-01 -1.47947907e+00 1.67712724e+00
2.79027998e-01 3.57616186e-01 -1.76916987e-01 -2.99316317e-01
8.81778836e-01 7.89115965e-01 2.06308708e-01 -3.68240148e-01
1.74996525e-01 4.70528573e-01 4.66028929e-01 -4.49134380e-01
3.33038688e-01 -3.30810755e-01 -1.66316867e-01 8.45086873e-01
-4.25760597e-01 -2.61324376e-01 3.11810791e-01 3.96677524e-01
1.13317418e+00 5.53018093e-01 5.63731849e-01 -6.18131042e-01
1.05401158e+00 2.56135851e-01 4.08434093e-01 5.67245066e-01
-2.38393813e-01 -4.23124373e-01 8.48908544e-01 -8.44302118e-01
-1.44326472e+00 -1.16578913e+00 2.74620354e-01 2.69752949e-01
2.74952292e-01 -7.15907395e-01 -8.12224448e-01 -5.51646017e-03
-1.51557967e-01 3.83608311e-01 -1.38335237e-02 1.29506543e-01
-6.46607876e-01 -7.81305671e-01 1.39064717e+00 1.89202338e-01
3.74037355e-01 -8.01413238e-01 -1.00501025e+00 2.54240364e-01
3.68707746e-01 -8.88343334e-01 6.80286407e-01 7.33043700e-02
-6.93560779e-01 -1.13265538e+00 3.14924978e-02 -3.96242976e-01
4.14686888e-01 -2.83217449e-02 6.63708389e-01 8.28226268e-01
-4.82879162e-01 3.74633789e-01 -1.87236443e-01 -8.56278360e-01
-9.80977893e-01 -3.56406957e-01 -1.12079673e-01 -4.24687743e-01
4.60769743e-01 -5.50996125e-01 -5.32631874e-01 -3.02816123e-01
-1.27702177e+00 1.86725393e-01 4.09090906e-01 1.86478809e-01
-2.66629428e-01 2.56618559e-01 1.67421743e-01 -1.03919125e+00
4.29863334e-01 -5.20293951e-01 -8.38898301e-01 2.86011964e-01
-7.24909663e-01 3.39572310e-01 3.62266541e-01 1.58294961e-01
-1.15379679e+00 -1.50380209e-01 1.87749684e-01 8.70946050e-01
-2.01551855e-01 5.88843465e-01 -1.55877411e-01 -1.78862989e-01
5.06402016e-01 3.10474575e-01 2.26517349e-01 4.83503491e-01
2.80227989e-01 3.64704430e-01 1.46334127e-01 -9.16738570e-01
7.35530436e-01 8.83892119e-01 1.05383492e+00 -5.17592669e-01
5.13561249e-01 -4.89669070e-02 -2.93272167e-01 -5.02863050e-01
7.80591846e-01 -3.55642140e-01 -1.75412130e+00 6.78472161e-01
-1.57744575e+00 1.59694850e-01 1.65777624e-01 1.07296258e-01
-4.10239309e-01 6.53992832e-01 -6.81136906e-01 -1.62109768e+00
-1.86840817e-01 -1.02141595e+00 6.46324933e-01 2.45442107e-01
-4.65709507e-01 -8.40187252e-01 4.34357166e-01 2.05413163e-01
1.91435874e-01 3.52240056e-01 1.31126571e+00 -4.50265408e-01
-1.41638970e+00 -1.34145647e-01 -1.16507992e-01 -6.10780537e-01
-4.37438101e-01 4.06275660e-01 -8.95324290e-01 -8.01635757e-02
-2.01085925e-01 2.00431719e-01 -5.98057453e-03 -5.77030063e-01
3.08372796e-01 6.79020733e-02 -3.74073446e-01 -9.89164263e-02
2.00035930e+00 7.62641788e-01 1.34967303e+00 4.95087475e-01
2.59890463e-02 6.72017694e-01 2.97190756e-01 4.17506576e-01
2.33861804e-01 1.76507056e-01 2.07128271e-01 6.45209551e-01
1.04140500e-02 -1.37327224e-01 2.94628382e-01 7.36124158e-01
-6.28339529e-01 -2.48557672e-01 -1.75061071e+00 -4.40445822e-03
-1.77285576e+00 -1.22828043e+00 -8.58292103e-01 1.67195559e+00
5.19443095e-01 1.05725557e-01 2.16805115e-02 7.19441116e-01
6.65471375e-01 -4.29409087e-01 1.52247906e-01 -7.92581975e-01
-8.61392356e-03 4.02256936e-01 6.28783226e-01 4.13676709e-01
-2.39871129e-01 5.61066806e-01 6.24894905e+00 3.16574991e-01
-7.57832468e-01 -2.19280319e-03 -7.63715729e-02 7.25598037e-01
-8.82504463e-01 8.39771926e-01 -5.87788165e-01 4.15301800e-01
1.53128326e+00 -4.26849931e-01 4.54773039e-01 2.74467375e-02
3.64240967e-02 -7.23078251e-01 -1.00746620e+00 4.34322000e-01
-2.15570793e-01 -1.66293478e+00 1.69164777e-01 1.54198229e-01
4.35814053e-01 -5.61340809e-01 -3.47477347e-01 -2.37120643e-01
3.65053385e-01 -1.02579939e+00 8.92962039e-01 7.08538711e-01
1.42798189e-03 -8.36061835e-01 8.39887023e-01 3.74096513e-01
-9.19941843e-01 -1.24848880e-01 -8.93793479e-02 -6.34532392e-01
2.84983329e-02 2.92505503e-01 -8.54633212e-01 9.19851780e-01
1.73953190e-01 -3.68999302e-01 1.05416097e-01 7.72513866e-01
1.13630176e-01 2.84588903e-01 -1.62012190e-01 -7.68175602e-01
2.47706473e-01 -6.62676990e-01 3.65655094e-01 1.33406651e+00
3.51819992e-01 4.04682845e-01 -3.03487599e-01 6.90481603e-01
5.86887479e-01 -2.21281797e-01 -9.84825730e-01 -4.64000851e-01
3.42064321e-01 1.01455903e+00 -1.39315701e+00 -2.74042845e-01
-4.61422205e-01 3.05094212e-01 -4.75243300e-01 4.25180942e-02
-5.90031207e-01 -1.98674619e-01 3.25910956e-01 2.51363873e-01
-2.69841313e-01 -9.34591532e-01 -7.00493455e-01 -6.22812390e-01
-4.71174151e-01 -9.30316389e-01 6.71797693e-02 -6.28295302e-01
-5.35471201e-01 6.49597570e-02 8.42587203e-02 -4.49636340e-01
-3.77113894e-02 -6.13079071e-01 -3.90336514e-01 6.50127470e-01
-1.08364856e+00 -1.26029849e+00 1.26337856e-01 2.74131559e-02
-2.70982742e-01 -1.76704854e-01 9.94715095e-01 3.17109004e-02
-5.49187232e-03 -3.33611041e-01 -3.72342467e-01 -2.11119130e-01
-8.01250041e-02 -8.50867450e-01 6.40496016e-01 9.85840976e-01
-1.90288872e-01 1.05073249e+00 1.17971504e+00 -1.05010331e+00
-2.46499562e+00 -3.43417257e-01 1.08534551e+00 -3.82963061e-01
9.08863604e-01 -7.38607943e-01 -5.16499639e-01 6.21422470e-01
5.99931955e-01 -4.87187028e-01 5.80470562e-01 -5.78329742e-01
-3.91926825e-01 5.08973897e-02 -1.23688531e+00 9.32398319e-01
9.25363541e-01 -6.04486585e-01 -3.40488344e-01 3.92913818e-03
5.53184032e-01 -1.70785338e-01 -8.08573246e-01 6.19985424e-02
7.89228618e-01 -1.17302394e+00 6.89943373e-01 -4.64976102e-01
-8.04651063e-03 -8.38083208e-01 -1.47494584e-01 -2.13908300e-01
8.59905556e-02 -1.15410626e+00 1.26865849e-01 1.27583122e+00
2.47262761e-01 -8.34353387e-01 1.16241777e+00 8.51985812e-01
2.38458976e-01 -2.26276070e-01 -8.10060859e-01 -5.48927009e-01
3.53303887e-02 -3.75313342e-01 7.18414903e-01 6.23183906e-01
7.77828217e-01 -5.71318753e-02 6.49290830e-02 1.31060928e-01
9.10886884e-01 -1.98617131e-01 6.40896142e-01 -1.28798807e+00
-1.60561472e-01 -1.44089296e-01 -9.90471005e-01 -3.80981411e-03
-8.39173719e-02 -1.07209218e+00 -1.60655543e-01 -1.33960783e+00
4.78637278e-01 -2.86173671e-01 3.90652716e-01 -1.42669052e-01
5.59652209e-01 4.09636684e-02 3.91209155e-01 -8.78080055e-02
-1.91783801e-01 -3.44561845e-01 9.37882602e-01 -5.45060672e-02
5.44657350e-01 -2.06587926e-01 -3.72820020e-01 3.47062081e-01
8.80560100e-01 -7.24656224e-01 -5.52779973e-01 1.65465057e-01
1.46438253e+00 4.19299692e-01 4.46316957e-01 -9.86538172e-01
9.36427176e-01 -5.64529717e-01 -1.77806422e-01 -4.91643816e-01
4.23689298e-02 -1.19355595e+00 1.08965099e+00 1.25030124e+00
-1.45081878e-01 4.40595746e-01 -1.14419274e-01 6.67783141e-01
2.98763543e-01 -5.50770223e-01 3.35489780e-01 -4.45658952e-01
-6.64803267e-01 -1.44118607e-01 -1.36768532e+00 -4.62620378e-01
1.63951302e+00 -6.42819464e-01 -6.22384965e-01 2.82002389e-01
-3.64742190e-01 -1.47110149e-01 6.80027544e-01 -1.49945661e-01
5.21314085e-01 -6.96850061e-01 -3.92315775e-01 9.14194658e-02
-3.02967489e-01 -4.44393069e-01 9.58591923e-02 6.53512776e-01
-1.35235727e+00 8.88714612e-01 -6.20786309e-01 -4.78161335e-01
-1.48023689e+00 8.68185878e-01 2.97908280e-02 7.93808792e-03
-3.41012120e-01 1.32537549e-02 -4.02325153e-01 -7.67550319e-02
5.11435643e-02 -2.83954263e-01 2.64423937e-01 -4.93768483e-01
5.22244155e-01 8.04766238e-01 -8.74126423e-03 -5.87198399e-02
-6.38293326e-01 3.72075766e-01 3.03936064e-01 -6.78779662e-01
1.01235032e+00 -7.89575353e-02 -1.11355364e+00 5.57045996e-01
2.94675440e-01 -1.50714293e-02 -4.41167384e-01 4.39699948e-01
8.14915717e-01 3.56973261e-02 -5.76648831e-01 -8.27335954e-01
-3.29316296e-02 7.28777468e-01 4.02116746e-01 2.92539179e-01
5.01878262e-01 -3.79002005e-01 3.75804394e-01 6.26267731e-01
1.01745307e+00 -9.05593455e-01 -3.10727000e-01 4.37086582e-01
1.65162891e-01 -3.33484888e-01 1.78325474e-01 -9.63370085e-01
-1.28734306e-01 1.62269926e+00 6.44653589e-02 -8.38467926e-02
4.45517570e-01 1.33423090e+00 2.33684834e-02 -3.84688020e-01
-9.36918020e-01 1.14883840e-01 -7.75207102e-01 8.90353739e-01
6.29575908e-01 2.42812466e-02 -1.03676033e+00 2.58160353e-01
1.35474399e-01 6.26383007e-01 1.38732874e+00 1.59972072e+00
-5.12489617e-01 -1.91562963e+00 -7.91202486e-01 -4.08339649e-01
-6.39432669e-01 5.85930869e-02 -5.41464746e-01 1.02959549e+00
2.74843097e-01 7.81854153e-01 -1.64745986e-01 -1.38897032e-01
-1.92607418e-01 3.36073905e-01 9.59570765e-01 -2.74652183e-01
-1.03311408e+00 -5.88176727e-01 3.91129762e-01 -1.37148798e-01
-6.18774414e-01 -8.56674612e-01 -2.08847046e+00 -9.81093109e-01
8.64687562e-02 5.40044308e-01 1.24107683e+00 9.67536807e-01
3.41063663e-02 3.20138425e-01 -2.55066138e-02 -2.27610007e-01
-2.76945561e-01 -8.87733921e-02 -4.75144535e-01 -2.15049669e-01
-3.63446802e-01 -1.10793561e-01 1.80182472e-01 9.08202752e-02] | [5.7572407722473145, 4.640433311462402] |
ef5b92a2-79e7-4190-83a3-7031a1e843bf | standoff-tracking-using-dnn-based-mpc-with | 2212.10945 | null | https://arxiv.org/abs/2212.10945v1 | https://arxiv.org/pdf/2212.10945v1.pdf | Standoff Tracking Using DNN-Based MPC with Implementation on FPGA | This work studies the standoff tracking problem to drive an unmanned aerial vehicle (UAV) to slide on a desired circle over a moving target at a constant height. We propose a novel Lyapunov guidance vector (LGV) field with tunable convergence rates for the UAV's trajectory planning and a deep neural network (DNN)-based model predictive control (MPC) scheme to track the reference trajectory. Then, we show how to collect samples for training the DNN offline and design an integral module (IM) to refine the tracking performance of our DNN-based MPC. Moreover, the hardware-in-the-loop (HIL) simulation with an FPGA@200MHz demonstrates that our method is a valid alternative to embedded implementations of MPC for addressing complex systems and applications which is impossible for directly solving the MPC optimization problems. | ['Shiji Song', 'Keyou You', 'Xingchen Li', 'Fei Dong'] | 2022-12-21 | null | null | null | null | ['trajectory-planning'] | ['robots'] | [ 2.73440629e-02 1.67831883e-01 -3.47591281e-01 1.99379697e-01
2.14371040e-01 -8.88006330e-01 3.39828342e-01 -2.41354987e-01
-2.12682888e-01 8.44898582e-01 -9.90053475e-01 -9.97881770e-01
-3.43675375e-01 -6.94858193e-01 -8.60772014e-01 -7.55003572e-01
-5.35932660e-01 3.05253658e-02 1.35168180e-01 -4.64860290e-01
-9.27819833e-02 9.77564394e-01 -1.16826808e+00 -7.35989511e-01
7.06013083e-01 1.13922286e+00 2.03039855e-01 1.05720031e+00
1.19199848e+00 5.57874739e-01 -2.79867083e-01 2.98243195e-01
9.79431450e-01 6.92547187e-02 -3.89584959e-01 -1.74839795e-02
4.39299256e-01 -8.44862998e-01 -4.96946871e-01 1.08453286e+00
3.94956857e-01 3.88610572e-01 3.47149044e-01 -1.29999030e+00
-7.09527805e-02 -7.22790696e-03 -6.13534115e-02 -8.81876275e-02
-3.34455818e-01 6.66177392e-01 3.47138345e-01 -1.53527305e-01
4.31470603e-01 8.36020172e-01 8.88222814e-01 5.07996500e-01
-9.53809083e-01 -3.07707548e-01 1.44073278e-01 -2.56497920e-01
-1.42117631e+00 3.23561057e-02 2.43568644e-01 -5.03681779e-01
7.71360099e-01 2.56618559e-01 1.11553049e+00 4.57374513e-01
8.58893096e-01 3.26399744e-01 3.20156693e-01 -8.62988085e-02
3.54360759e-01 -2.19633415e-01 -4.03975427e-01 8.10799122e-01
6.48988903e-01 8.97786260e-01 6.36830449e-01 -6.24651797e-02
1.35738301e+00 2.21425556e-02 -6.40713036e-01 -3.28914523e-01
-1.18094897e+00 8.86962056e-01 6.82855129e-01 -2.61571646e-01
-7.10966349e-01 6.48346186e-01 5.08315004e-02 4.97866988e-01
-3.05636257e-01 8.76945972e-01 -8.64299417e-01 3.68764736e-02
-9.37107801e-01 8.46429646e-01 1.01726782e+00 1.44378066e+00
4.06507730e-01 9.18291748e-01 9.16080270e-03 -2.57199824e-01
3.33084822e-01 8.01492393e-01 1.33902863e-01 -1.21204507e+00
-6.05979860e-02 5.99553168e-01 6.72392905e-01 -1.06654727e+00
-6.60761356e-01 -5.90828478e-01 -9.40379381e-01 7.30207920e-01
1.59768611e-01 -1.16675711e+00 -5.16447246e-01 1.39244211e+00
6.04029059e-01 -1.04214847e-01 1.43669248e-01 1.26665306e+00
-2.17570037e-01 9.88641620e-01 -4.53905463e-01 -2.95656532e-01
7.09325135e-01 -9.18504000e-01 -4.03426021e-01 -1.11714192e-01
8.39483917e-01 -2.94828832e-01 4.49885964e-01 4.76791233e-01
-8.88016284e-01 -6.23940766e-01 -1.43896949e+00 3.52633804e-01
-2.02798888e-01 9.07654524e-01 2.33668968e-01 1.65299162e-01
-1.15056241e+00 1.11561990e+00 -1.04330230e+00 -1.46312490e-01
-1.48938015e-01 8.90944183e-01 1.43455282e-01 6.37211442e-01
-1.15261388e+00 1.05593395e+00 7.06296623e-01 4.24507499e-01
-1.34390008e+00 -8.60498905e-01 -6.43378615e-01 -2.30884522e-01
5.35030246e-01 -7.68120527e-01 1.33950841e+00 -1.02507079e+00
-2.19791961e+00 -4.71537672e-02 4.70686555e-01 -1.05277085e+00
2.91409254e-01 -3.24821144e-01 2.08961759e-02 1.16568074e-01
-5.91343403e-01 7.83149719e-01 1.16841900e+00 -6.30706787e-01
-9.57949162e-01 7.06906170e-02 4.18630600e-01 9.90430266e-02
3.28272767e-02 -6.24150157e-01 2.13888422e-01 -4.28224653e-01
-3.98766369e-01 -1.46015584e+00 -6.94499850e-01 2.01273322e-01
-2.88911849e-01 1.91315606e-01 1.16958153e+00 -3.85436118e-01
1.33424830e+00 -1.58668995e+00 3.90705168e-01 5.76326512e-02
-6.51621819e-02 7.87325799e-01 3.95536982e-03 3.50048661e-01
2.56364644e-01 -4.33898836e-01 4.98525891e-03 4.84497845e-01
-1.89041525e-01 -4.12258580e-02 -6.68778121e-01 7.52327383e-01
2.99104273e-01 7.18396366e-01 -8.09948862e-01 1.12368278e-01
1.56780049e-01 3.46685797e-01 -6.21220648e-01 1.26224726e-01
-6.36799157e-01 3.42284739e-01 -6.16154969e-01 4.81762558e-01
7.19045758e-01 2.11790740e-01 2.03784764e-01 -1.22041360e-01
-9.41773176e-01 -1.67191207e-01 -9.67631578e-01 9.67917383e-01
-2.98932076e-01 8.81486416e-01 8.75981271e-01 -5.47768295e-01
1.09779382e+00 9.65400189e-02 2.64388263e-01 1.52480990e-01
9.23662663e-01 1.37663946e-01 2.01817304e-02 1.12959944e-01
7.68901467e-01 4.93245125e-02 2.51494765e-01 -2.01001987e-01
-9.37882885e-02 -4.84571904e-01 -1.86796024e-01 -3.36821496e-01
8.78031611e-01 1.71419069e-01 3.56068254e-01 -7.07444072e-01
8.34807098e-01 7.86156774e-01 5.96347153e-01 1.86097890e-01
-2.70372063e-01 -3.00433040e-01 2.48818174e-01 -7.32713401e-01
-1.07950842e+00 -4.66481984e-01 1.57286569e-01 4.48556989e-01
2.02451140e-01 -3.27435732e-02 -7.28073061e-01 -7.11840987e-02
7.09844530e-02 7.89428890e-01 -1.14906684e-01 -2.94760495e-01
-5.63444138e-01 -1.69099212e-01 4.34464514e-01 4.66043025e-01
4.49008316e-01 -2.29434699e-01 -1.53007793e+00 4.34176296e-01
6.48860633e-01 -9.12314892e-01 -4.66107696e-01 2.38338038e-01
-1.02145112e+00 -1.18637633e+00 -3.69201601e-01 -9.11899090e-01
6.54719949e-01 1.82192773e-01 1.92354605e-01 -1.20436065e-02
-6.15740158e-02 5.61924458e-01 -7.06813531e-03 -6.22955143e-01
-2.50430673e-01 -5.88730955e-03 6.86000109e-01 -3.97361696e-01
-2.70966887e-01 -1.42325044e-01 -4.46279168e-01 4.11387205e-01
-5.72550058e-01 1.33756682e-01 3.57406199e-01 7.87280142e-01
8.31207812e-01 3.53363723e-01 1.54217333e-01 1.90849736e-01
6.14607215e-01 -2.58648675e-02 -2.10103893e+00 3.95452380e-02
-6.26222908e-01 -2.68295079e-01 1.35096705e+00 -8.14715087e-01
-1.59205258e-01 7.65247226e-01 4.20222372e-01 -1.16273808e+00
2.74538994e-01 1.57394052e-01 1.78958729e-01 -9.40090895e-01
3.31621885e-01 1.83274224e-01 5.72793245e-01 2.81796366e-01
2.64503330e-01 4.10781592e-01 5.84988117e-01 -1.24803454e-01
1.08539093e+00 1.52550980e-01 8.74697924e-01 -9.98759329e-01
-3.57070208e-01 1.25382587e-01 -1.00452185e+00 -4.37221974e-01
7.48529792e-01 -1.06646025e+00 -1.49313176e+00 1.20383821e-01
-1.31427109e+00 -8.00314248e-01 -3.02111536e-01 6.43235505e-01
-8.95333588e-01 -1.15458786e-01 -4.23699886e-01 -8.73962581e-01
-6.91349328e-01 -1.20959115e+00 7.26354301e-01 5.68848133e-01
1.65901691e-01 -9.89277601e-01 1.48482233e-01 -8.01583946e-01
5.95468700e-01 8.12779546e-01 3.69006932e-01 4.69629560e-03
-9.41900790e-01 -4.93293315e-01 3.03300142e-01 3.77796620e-01
-3.49292129e-01 4.33039308e-01 -3.26488942e-01 -1.02705836e+00
1.90477684e-01 -5.39363958e-02 5.95280491e-02 7.54165053e-01
5.78928888e-01 -9.04060543e-01 -5.37664115e-01 1.01834965e+00
1.81935263e+00 5.97594500e-01 7.06353039e-02 4.17531520e-01
5.83326697e-01 1.44697353e-01 1.12090325e+00 4.13952023e-01
1.87102199e-01 5.71342945e-01 9.59538221e-01 2.71432728e-01
6.96245790e-01 -1.62475944e-01 6.89437568e-01 3.53855908e-01
-1.43428534e-01 -1.24106258e-01 -5.97000241e-01 3.64059687e-01
-1.66878760e+00 -5.49275994e-01 -2.89148301e-01 2.32622528e+00
3.84751856e-01 -3.55163723e-01 -6.36748821e-02 -1.01604804e-01
7.02276766e-01 -2.16899604e-01 -6.90676689e-01 -6.16633356e-01
4.27282155e-01 -2.69241869e-01 1.46386909e+00 5.76604187e-01
-1.51146531e+00 9.40330148e-01 5.97202063e+00 6.13080680e-01
-1.74175334e+00 -5.83501101e-01 -3.09597421e-02 -6.51395693e-02
5.43918490e-01 3.90327051e-02 -1.27806830e+00 1.92854881e-01
1.45532751e+00 -5.69719017e-01 8.76827776e-01 1.47463787e+00
5.38062632e-01 2.00209886e-01 -8.33860695e-01 6.23941720e-01
-5.20781279e-01 -1.42531669e+00 -3.06090005e-02 2.65193403e-01
6.93567097e-01 -1.60902049e-02 6.58615232e-02 3.18262428e-01
2.74161190e-01 -5.96616268e-01 7.77159810e-01 4.75268096e-01
6.94303334e-01 -1.03767049e+00 6.11791492e-01 7.84760952e-01
-1.21635973e+00 -4.48377877e-01 -9.10353184e-01 -3.98875087e-01
2.68141210e-01 -8.24113190e-02 -1.30044687e+00 5.92046857e-01
3.03097162e-02 6.14813685e-01 -9.28223282e-02 1.14816105e+00
-1.70660555e-01 3.61057878e-01 -4.21325505e-01 -7.36157835e-01
1.01662970e+00 -3.90740842e-01 1.18115437e+00 6.11163080e-01
5.78580022e-01 2.95015574e-01 3.55697751e-01 9.72332716e-01
5.07253528e-01 -3.93436253e-01 -8.22935641e-01 -2.94601887e-01
1.91315264e-01 1.36213970e+00 -2.17368528e-01 -1.44106131e-02
2.35468045e-01 4.56962764e-01 -2.77716786e-01 1.31275104e-02
-1.05715132e+00 -8.77961099e-01 8.57178807e-01 5.19892052e-02
4.66382474e-01 -8.86838734e-01 2.36544207e-01 -6.32781625e-01
-4.12780941e-01 -6.26810610e-01 -3.69073927e-01 -5.97862422e-01
-5.01989543e-01 6.26280904e-01 -1.70472339e-01 -2.10696816e+00
-6.00547850e-01 -9.95952725e-01 -6.07722461e-01 8.21296871e-01
-1.40373206e+00 -8.23618829e-01 -2.05102697e-01 4.26066309e-01
2.71729171e-01 -3.27280939e-01 6.89060569e-01 -2.12845922e-01
-4.89909947e-01 1.12103067e-01 2.21919015e-01 -1.94351286e-01
2.17356235e-01 -9.02993143e-01 2.44096443e-01 1.14580989e+00
-8.51252615e-01 5.27383804e-01 8.29805851e-01 -7.53827035e-01
-2.39900613e+00 -1.93649614e+00 7.18995705e-02 -1.08749911e-01
9.51625824e-01 2.92847067e-01 -4.18184191e-01 1.00790250e+00
2.03701407e-01 -1.04980335e-01 -2.70853966e-01 -1.12749445e+00
8.12669218e-01 -1.78906798e-01 -1.09786785e+00 7.28798389e-01
3.36209536e-01 3.28273296e-01 -9.45134535e-02 2.57925868e-01
1.34041810e+00 -9.54512298e-01 -9.23615038e-01 6.02916121e-01
5.72852075e-01 -1.95487335e-01 6.64431751e-01 -5.85346818e-01
8.10016226e-03 -8.53592277e-01 -1.45635054e-01 -1.59868371e+00
-4.55144435e-01 -1.32387340e+00 -3.38738978e-01 3.60796601e-01
4.73930985e-02 -6.47648156e-01 5.67218661e-01 2.21609890e-01
-4.57373887e-01 -9.77819443e-01 -7.36820042e-01 -1.13171220e+00
1.28324389e-01 -6.26917183e-03 4.28145558e-01 3.89931083e-01
-2.15996191e-01 4.46135700e-02 -3.91119808e-01 1.24431741e+00
5.18661380e-01 2.87013818e-02 9.35681999e-01 -1.10090280e+00
-1.19725745e-02 -1.30938560e-01 -4.30127412e-01 -1.25630856e+00
8.56942963e-03 -5.46406269e-01 2.35723808e-01 -1.19306815e+00
-1.09712565e+00 -1.34842649e-01 4.77120340e-01 2.09876940e-01
4.90421027e-01 -7.58868098e-01 2.61192471e-01 -4.61034365e-02
-1.63514838e-01 5.47216356e-01 1.50056112e+00 -2.12380633e-01
-5.62487721e-01 5.40458798e-01 -7.49547854e-02 7.61852384e-01
8.17931771e-01 -1.25257373e-02 -6.25868678e-01 -4.63974148e-01
-2.35105544e-01 4.81222779e-01 5.67648232e-01 -1.65772116e+00
6.21223867e-01 -3.85936797e-01 2.17956409e-01 -1.11959898e+00
1.59837529e-01 -1.27406383e+00 4.64908890e-02 1.16923881e+00
3.23813632e-02 3.24759036e-01 6.70109451e-01 5.17644465e-01
-6.60739988e-02 -6.96493313e-02 1.10801589e+00 4.61241126e-01
-5.64634204e-01 7.61701465e-01 -6.14314020e-01 -4.64603335e-01
1.52790284e+00 1.22601569e-01 -3.05460304e-01 -2.17283592e-01
-3.34691048e-01 7.25912452e-01 6.30309045e-01 1.48276305e-02
5.85426629e-01 -1.16170454e+00 -2.44339649e-02 2.84605086e-01
-6.65895462e-01 1.22161299e-01 6.64279312e-02 8.81492257e-01
-1.10033727e+00 1.25529158e+00 -4.77554947e-01 -5.91381907e-01
-9.00748849e-01 7.35466480e-01 8.44510376e-01 2.60975957e-02
-4.05825019e-01 4.27010953e-01 -3.03139746e-01 -6.29385531e-01
1.34890407e-01 -8.66162062e-01 1.04393892e-01 -4.32168424e-01
5.42592287e-01 6.47641599e-01 -2.40551725e-01 -1.55009151e-01
-7.48016015e-02 5.77225029e-01 6.22493744e-01 1.58394426e-01
1.21919656e+00 1.54589385e-01 -3.60365845e-02 -7.89734498e-02
8.25279057e-01 -6.36834681e-01 -1.79791546e+00 6.38112187e-01
-4.37669128e-01 -2.79072858e-02 4.56926793e-01 -3.85760576e-01
-9.05903518e-01 6.24893069e-01 5.15326500e-01 1.74572825e-01
1.03366148e+00 -1.15866971e+00 6.96127594e-01 9.58262205e-01
8.10731411e-01 -9.80869174e-01 -6.58515275e-01 1.08631968e+00
9.19066548e-01 -5.90048313e-01 1.09908536e-01 -1.23870462e-01
-5.97391009e-01 1.80302250e+00 7.07837045e-01 -8.61115873e-01
7.47462094e-01 4.32008058e-01 -3.74303371e-01 2.57215947e-01
-1.12521040e+00 1.05979428e-01 8.44361782e-02 5.28908491e-01
-3.17609102e-01 2.29986444e-01 -1.89529583e-01 4.12448078e-01
-1.60664380e-01 3.87605041e-01 9.89145160e-01 1.03053534e+00
-7.65403390e-01 -4.57732707e-01 -4.20589298e-01 3.92295308e-02
3.77167277e-02 2.44308561e-01 -1.52328331e-02 1.30976129e+00
2.08083242e-02 5.46124935e-01 1.45511419e-01 -7.59146214e-01
4.63981479e-01 -4.89556849e-01 4.44072276e-01 -2.45778531e-01
-5.48858047e-01 -9.93042663e-02 -2.11738989e-01 -7.96058297e-01
7.53017515e-02 -2.55055487e-01 -1.26167500e+00 -3.33373457e-01
-4.30247992e-01 -6.49217516e-02 7.08858788e-01 5.02973855e-01
5.53170919e-01 4.23898339e-01 7.95791447e-01 -1.21286643e+00
-1.02356589e+00 -5.92498720e-01 -5.71935534e-01 -1.06909716e+00
7.83870697e-01 -7.38655746e-01 -4.29564804e-01 -2.07545459e-02] | [4.933393478393555, 2.063222885131836] |
5fdd845c-0e11-4ac7-a2ef-4f268f227674 | metal-artifact-reduction-in-2d-ct-images-with | 2109.13483 | null | https://arxiv.org/abs/2109.13483v1 | https://arxiv.org/pdf/2109.13483v1.pdf | Metal Artifact Reduction in 2D CT Images with Self-supervised Cross-domain Learning | The presence of metallic implants often introduces severe metal artifacts in the X-ray CT images, which could adversely influence clinical diagnosis or dose calculation in radiation therapy. In this work, we present a novel deep-learning-based approach for metal artifact reduction (MAR). In order to alleviate the need for anatomically identical CT image pairs (i.e., metal artifact-corrupted CT image and metal artifact-free CT image) for network learning, we propose a self-supervised cross-domain learning framework. Specifically, we train a neural network to restore the metal trace region values in the given metal-free sinogram, where the metal trace is identified by the forward projection of metal masks. We then design a novel FBP reconstruction loss to encourage the network to generate more perfect completion results and a residual-learning-based image refinement module to reduce the secondary artifacts in the reconstructed CT images. To preserve the fine structure details and fidelity of the final MAR image, instead of directly adopting CNN-refined images as output, we incorporate the metal trace replacement into our framework and replace the metal-affected projections of the original sinogram with the prior sinogram generated by the forward projection of the CNN output. We then use the filtered backward projection (FBP) algorithms for final MAR image reconstruction. We conduct an extensive evaluation on simulated and real artifact data to show the effectiveness of our design. Our method produces superior MAR results and outperforms other compelling methods. We also demonstrate the potential of our framework for other organ sites. | ['Lei Xing', 'Wei Zhao', 'Hongyi Ren', 'Xiaomeng Li', 'Zhicheng Zhang', 'Lequan Yu'] | 2021-09-28 | null | null | null | null | ['metal-artifact-reduction'] | ['medical'] | [ 4.96600896e-01 1.30228832e-01 2.72299677e-01 -1.84283987e-01
-9.46491539e-01 -2.81195324e-02 1.64216429e-01 -3.20874929e-01
-2.24051625e-01 6.94541335e-01 3.33972871e-01 -1.95244417e-01
-2.41381273e-01 -8.01824868e-01 -8.69220912e-01 -8.92390907e-01
4.19576287e-01 2.15039879e-01 3.64122212e-01 2.79548168e-02
-6.48749396e-02 6.10229790e-01 -5.82057774e-01 5.65397859e-01
9.99143898e-01 9.78688717e-01 6.02848172e-01 1.62184685e-01
2.34855294e-01 1.04327893e+00 -3.52405757e-01 -1.16932370e-01
5.13931096e-01 -6.91503346e-01 -7.37901092e-01 1.34155482e-01
2.62001976e-02 -3.88455570e-01 -7.03484356e-01 8.86090517e-01
7.16811359e-01 -5.87626137e-02 5.78570604e-01 -5.27772903e-01
-5.99489868e-01 6.11858428e-01 -8.43049705e-01 1.12519801e-01
-1.48289233e-01 2.19981238e-01 2.16032621e-02 -1.02062893e+00
5.67660928e-01 5.80236316e-01 1.07458591e+00 7.40267992e-01
-1.11829901e+00 -8.01369488e-01 -3.10833603e-01 -9.83013026e-03
-1.18163717e+00 -2.33460758e-02 1.06429517e+00 -3.75524044e-01
4.38705355e-01 4.52131659e-01 6.23956501e-01 7.76980460e-01
7.42253900e-01 4.75301504e-01 1.07005739e+00 -4.24142689e-01
5.12592234e-02 -1.47081017e-01 -2.09906429e-01 8.33787441e-01
2.36834720e-01 2.12393641e-01 -8.66649300e-02 -2.89501756e-01
1.25775230e+00 5.13253748e-01 -8.90188634e-01 -3.75820965e-01
-1.10825527e+00 4.87859935e-01 9.26486671e-01 4.91351336e-01
-7.40796924e-01 3.34657431e-01 1.23922132e-01 -1.75873712e-01
3.95828813e-01 4.51537281e-01 -1.46096274e-01 5.94984531e-01
-9.00900304e-01 8.69668797e-02 2.46419787e-01 6.73077941e-01
3.40663463e-01 1.95399165e-01 -4.01826948e-01 9.76373553e-01
4.06005085e-01 3.20075274e-01 8.39508176e-01 -7.12524414e-01
2.68870831e-01 5.10081291e-01 -2.44427510e-02 -6.20552957e-01
-4.75929856e-01 -6.52519047e-01 -1.10678625e+00 4.11858320e-01
3.44936848e-01 -1.80277854e-01 -1.46368349e+00 1.31292081e+00
3.91791493e-01 4.35529113e-01 -1.13549612e-01 1.15920985e+00
7.27931023e-01 4.69128668e-01 8.06482211e-02 -4.02066886e-01
1.18167353e+00 -9.39842343e-01 -8.38560104e-01 -1.96870998e-01
5.77247679e-01 -8.29973042e-01 9.33010817e-01 2.24352136e-01
-1.46975410e+00 -2.51030892e-01 -1.27962172e+00 1.51029110e-01
4.23254728e-01 4.73467559e-02 3.51337820e-01 3.31907511e-01
-6.66541338e-01 1.04473901e+00 -1.21173835e+00 3.55277956e-01
7.60806561e-01 4.29063708e-01 -6.36466369e-02 -2.46205524e-01
-1.00030053e+00 9.81049597e-01 -2.58810427e-02 1.94518462e-01
-9.94833469e-01 -1.33844161e+00 -7.07017362e-01 -2.91572392e-01
2.00155064e-01 -8.55142176e-01 1.35198164e+00 -1.20399022e+00
-1.41901731e+00 5.57277977e-01 1.44259214e-01 -2.82220066e-01
7.04931676e-01 -6.55594319e-02 -3.51138204e-01 1.16422556e-01
9.77806672e-02 1.65733740e-01 7.00566292e-01 -1.65665936e+00
-1.94200262e-01 -3.49865526e-01 -6.13607883e-01 2.63001949e-01
-4.56031188e-02 -5.18165678e-02 -4.84463483e-01 -1.10796237e+00
6.42435014e-01 -9.48419452e-01 -7.42666066e-01 4.46046323e-01
-3.14202219e-01 5.98834991e-01 4.69648153e-01 -1.00620210e+00
1.07477927e+00 -2.12528110e+00 -2.23096743e-01 4.66260105e-01
3.89421254e-01 -2.48652864e-02 6.58130273e-02 -2.55897969e-01
-5.36653101e-01 -1.65208057e-01 -9.49212015e-01 -2.01020837e-01
-7.81438053e-01 1.19049147e-01 9.11957994e-02 7.14963615e-01
-3.08842603e-02 9.32518423e-01 -7.99984634e-01 -4.75090593e-01
2.45558038e-01 6.96040928e-01 -4.47310388e-01 2.21267879e-01
1.31448910e-01 9.25473869e-01 -5.30082524e-01 3.07160288e-01
1.06787312e+00 -4.48238879e-01 1.59772649e-01 -5.86505294e-01
-7.18044639e-02 1.27426341e-01 -8.52699459e-01 2.03075695e+00
-7.22092271e-01 1.18687138e-01 7.10573420e-02 -4.66798693e-01
5.43045342e-01 5.27225614e-01 1.06920266e+00 -1.04144382e+00
2.47780025e-01 4.36309516e-01 1.64398998e-01 -4.21620518e-01
7.33674970e-03 -7.54471421e-01 4.74978298e-01 4.79074836e-01
-3.14184427e-01 -2.89511263e-01 -6.50101244e-01 -1.22421779e-01
1.08629274e+00 -3.22225615e-02 -2.02053070e-01 -3.51927221e-01
3.94017786e-01 7.39005506e-02 5.76252162e-01 4.55885738e-01
-1.88385677e-02 1.34640551e+00 -6.14148192e-02 -5.17236114e-01
-1.23271024e+00 -1.33154702e+00 -4.84034598e-01 2.15999722e-01
3.08212548e-01 1.27410203e-01 -8.62922192e-01 -7.85883725e-01
-4.28037345e-01 5.24938166e-01 -6.37882352e-01 -3.98430407e-01
-1.26072013e+00 -1.07508814e+00 2.98188269e-01 7.64969885e-01
4.70763922e-01 -1.12024534e+00 -4.42370176e-01 3.98970693e-01
-3.72267365e-01 -7.23140657e-01 -6.78404629e-01 3.81377220e-01
-1.15983605e+00 -9.31570172e-01 -1.14084065e+00 -1.00108743e+00
9.86761570e-01 1.16037622e-01 9.85366344e-01 4.28515643e-01
-4.10147130e-01 -1.68889239e-01 -2.17939746e-02 -1.53198719e-01
-6.05540752e-01 -6.10337734e-01 -3.10422808e-01 -9.33208466e-02
-5.35761178e-01 -5.05610943e-01 -1.20326328e+00 4.50481176e-01
-1.13954759e+00 4.27201778e-01 5.77429891e-01 1.12058794e+00
9.24425244e-01 1.27860263e-01 3.14033270e-01 -9.41714644e-01
4.37999129e-01 -2.98198611e-01 -3.73826504e-01 1.61510617e-01
-4.69537824e-01 1.02593243e-01 5.98647356e-01 -5.11184931e-01
-1.36659396e+00 3.91440451e-01 -4.21979517e-01 -5.11066556e-01
3.24434906e-01 2.74080247e-01 1.49363345e-02 -4.63924348e-01
1.00999367e+00 3.48099262e-01 3.47321033e-02 -2.39123896e-01
-1.00774817e-01 3.29149872e-01 5.69903433e-01 -2.98857152e-01
6.85139298e-01 7.82025993e-01 2.83519854e-03 -3.16888869e-01
-7.68966913e-01 -1.27102658e-01 -3.71644497e-01 -1.81186184e-01
9.24264431e-01 -7.93665826e-01 -3.44163805e-01 4.32588577e-01
-1.21101987e+00 -4.77520287e-01 -4.77381855e-01 6.83209598e-01
-5.71594894e-01 5.85792422e-01 -9.27463353e-01 -2.51373917e-01
-7.35628903e-01 -1.59000409e+00 7.65567899e-01 1.48217484e-01
-1.31935105e-02 -7.23982513e-01 1.39070287e-01 1.76143482e-01
3.95190805e-01 3.11472654e-01 1.10088694e+00 -1.52498111e-01
-5.78868091e-01 5.89024536e-02 -2.73676068e-01 3.86191130e-01
3.50588292e-01 -7.42740810e-01 -8.79558384e-01 -1.15829796e-01
6.32809818e-01 5.98237775e-02 6.49222910e-01 8.23865712e-01
1.48351669e+00 2.42601428e-02 -4.27401692e-01 9.55816567e-01
1.63429642e+00 5.26607275e-01 9.54587936e-01 2.89350122e-01
8.09674859e-01 1.70765683e-01 2.54863054e-01 4.15756375e-01
-1.63401768e-01 5.62047541e-01 3.40182990e-01 -5.32584667e-01
-8.37596476e-01 -2.03181639e-01 -2.45639607e-01 5.49144149e-01
-1.26279444e-01 1.66108515e-02 -9.41493213e-01 5.00387013e-01
-1.65944040e+00 -4.96391207e-01 -4.49460477e-01 2.18051767e+00
9.88459110e-01 -1.24810070e-01 -4.69207823e-01 -4.28612716e-02
6.90440834e-01 -3.04685920e-01 -4.82165903e-01 2.53712952e-01
1.51916251e-01 7.47314274e-01 6.52980030e-01 5.29902875e-01
-7.88404703e-01 4.52638537e-01 6.29344034e+00 7.78034031e-01
-1.24972045e+00 5.57975948e-01 6.57309890e-01 -5.22078313e-02
-5.28346121e-01 -2.31964767e-01 -6.23903610e-03 4.60708022e-01
4.92874861e-01 2.07342193e-01 2.39721745e-01 5.12926280e-01
4.31636304e-01 -1.18613094e-01 -8.72212350e-01 9.85473633e-01
-1.32444918e-01 -1.61965477e+00 -1.37881353e-01 -1.58368185e-01
9.30711448e-01 -3.93198840e-02 -7.20862746e-02 -1.93416476e-01
2.44611293e-01 -9.98912394e-01 6.97358906e-01 4.55540419e-01
8.73808384e-01 -6.93350315e-01 5.96964717e-01 5.57192191e-02
-8.14859986e-01 4.59571704e-02 -3.09705675e-01 5.82895577e-01
3.34196597e-01 7.71536708e-01 -9.24458861e-01 8.28437328e-01
6.90926373e-01 3.64310861e-01 -2.78678447e-01 1.26070964e+00
-5.22130169e-02 4.05402899e-01 -1.78187534e-01 4.67060804e-01
-3.52191925e-02 -7.71730915e-02 2.22269133e-01 9.36561286e-01
3.21936965e-01 4.09889430e-01 -1.31026924e-01 1.10876501e+00
-1.37293428e-01 -1.43639557e-02 -2.81860828e-01 8.56815219e-01
8.20968747e-02 1.04393005e+00 -8.10880005e-01 -3.10130209e-01
-4.02191550e-01 1.21604300e+00 -1.03144851e-02 3.53837520e-01
-1.03276920e+00 7.67205879e-02 1.72776654e-01 6.13411844e-01
-4.37187925e-02 1.35539368e-01 -1.02127039e+00 -8.06555986e-01
1.31431133e-01 -6.05054975e-01 3.10423106e-01 -9.03892338e-01
-1.25174844e+00 9.06677067e-01 -2.92098910e-01 -1.44679999e+00
1.36911333e-01 -2.68926173e-01 -8.85797799e-01 1.07717144e+00
-1.46194196e+00 -9.82022762e-01 -5.63101113e-01 7.56118715e-01
4.51898187e-01 2.69605070e-01 4.26104397e-01 7.39992559e-01
-1.93175659e-01 4.99715000e-01 1.19513288e-01 1.40236378e-01
6.24587715e-01 -9.01176035e-01 4.11638431e-02 7.54009664e-01
-3.78200501e-01 2.90498674e-01 4.18431878e-01 -1.14761376e+00
-1.16831958e+00 -1.33292449e+00 1.15916580e-01 -2.67359823e-01
4.26332474e-01 -1.44029101e-02 -9.60150063e-01 6.35814846e-01
1.39052302e-01 4.95913565e-01 3.76332641e-01 -7.70219326e-01
1.79596767e-01 -2.69647054e-02 -1.43166244e+00 6.80345416e-01
8.21318150e-01 -1.21756777e-01 -4.36716288e-01 4.49597597e-01
7.34596968e-01 -8.39258969e-01 -9.73619938e-01 7.98525989e-01
4.59998220e-01 -6.78650856e-01 1.14622641e+00 -2.44840816e-01
7.47004032e-01 -4.01878834e-01 2.26873696e-01 -1.42609715e+00
-6.32581830e-01 -2.29985714e-01 3.40353251e-01 6.38302982e-01
5.36994457e-01 -3.26987982e-01 1.09587729e+00 7.55830109e-01
-8.66929114e-01 -8.94825399e-01 -9.65783775e-01 -4.21957374e-01
4.41137910e-01 -4.27721918e-01 4.21824217e-01 9.69047725e-01
-2.76852280e-01 -3.02430004e-01 -2.37830892e-01 4.02324468e-01
6.70919240e-01 2.07736008e-02 9.63363796e-02 -6.55680418e-01
-4.45859104e-01 -1.50432065e-01 1.29754826e-01 -7.82405615e-01
-3.30994219e-01 -1.02334213e+00 4.40735817e-01 -1.59227765e+00
4.75997835e-01 -7.09215939e-01 -4.59828317e-01 4.40569371e-01
-1.91454485e-01 6.03767097e-01 1.95601340e-02 3.16108793e-01
1.00252166e-01 5.44479311e-01 2.19234395e+00 -1.32157117e-01
-3.13099504e-01 8.86160657e-02 -4.79211450e-01 8.79508793e-01
5.46289623e-01 -7.83824384e-01 -1.61823586e-01 -5.56752563e-01
-1.36427447e-01 3.92061174e-01 5.25915444e-01 -1.14697742e+00
1.77560553e-01 2.86153890e-03 9.22789931e-01 -5.29896379e-01
1.37804374e-01 -1.24137759e+00 4.80020821e-01 9.26292837e-01
-1.64666697e-01 -2.35565290e-01 2.62303919e-01 1.98617339e-01
6.17607832e-02 -3.80630642e-01 1.31831646e+00 -3.67956102e-01
4.86952327e-02 4.75793183e-01 -2.74619073e-01 -3.30813348e-01
8.85090649e-01 -2.25148350e-01 6.18106201e-02 1.77703835e-02
-8.32828045e-01 -1.54036745e-01 3.47663373e-01 7.35633299e-02
1.07289791e+00 -1.53005624e+00 -6.40642703e-01 2.73899019e-01
-1.00817211e-01 3.41312319e-01 7.21631646e-01 1.11495769e+00
-9.87035871e-01 -7.42245689e-02 -1.87651485e-01 -5.19439876e-01
-8.79106045e-01 5.59027612e-01 9.18962598e-01 -6.21925473e-01
-1.18667030e+00 6.27872050e-01 7.99519300e-01 -3.20428520e-01
-5.66199310e-02 -3.88599128e-01 2.28085369e-01 -8.83361340e-01
5.21520436e-01 1.05671883e-01 5.75490296e-01 -2.71388143e-01
-2.10176110e-01 5.72038531e-01 -2.18781114e-01 -2.39569303e-02
1.46348226e+00 6.67941719e-02 9.70861092e-02 -1.95089638e-01
1.03588462e+00 -3.60205919e-02 -1.50404131e+00 -2.09868431e-01
-3.16559941e-01 -4.56568360e-01 3.58329654e-01 -1.12239659e+00
-1.62749040e+00 4.92913902e-01 1.16527665e+00 -5.62274158e-01
1.28621089e+00 8.34686384e-02 1.06010902e+00 -3.27368677e-01
1.53428957e-01 -7.27022409e-01 3.69256973e-01 1.61968283e-02
1.09695899e+00 -8.93606901e-01 2.83643812e-01 -7.75267780e-01
-6.38911188e-01 9.32686865e-01 7.46472001e-01 -2.99815685e-01
6.73941314e-01 6.56713367e-01 1.85914472e-01 -4.64612305e-01
-9.88224223e-02 2.78730839e-01 2.94109732e-01 6.30743504e-01
5.23216724e-01 -1.15212440e-01 -4.44076031e-01 8.47944975e-01
8.86502787e-02 1.37494311e-01 5.95887423e-01 1.16456580e+00
-1.92487195e-01 -9.20886397e-01 -6.53455257e-01 3.98017794e-01
-5.81323266e-01 -3.06303769e-01 2.18460914e-02 5.27778149e-01
9.13646221e-02 5.07228732e-01 -2.36865565e-01 -1.19139142e-01
5.61411798e-01 -2.73211718e-01 7.66080856e-01 -6.21041477e-01
-9.83486354e-01 4.68464553e-01 -3.58527869e-01 -5.30746579e-01
7.39288470e-03 -2.81171083e-01 -1.65259266e+00 1.02350771e-01
-4.88971114e-01 -2.01349676e-01 6.27189934e-01 7.91648388e-01
8.10770839e-02 1.09606040e+00 7.93870449e-01 -7.91140854e-01
-3.45899969e-01 -7.77583897e-01 -5.79386890e-01 5.83841801e-01
2.15177491e-01 -4.82470870e-01 -3.21816765e-02 2.01277863e-02] | [13.508736610412598, -2.5503976345062256] |
40147488-89f3-4697-8bf6-7707387c95fe | causal-based-supervision-of-attention-in | 2305.13115 | null | https://arxiv.org/abs/2305.13115v1 | https://arxiv.org/pdf/2305.13115v1.pdf | Causal-Based Supervision of Attention in Graph Neural Network: A Better and Simpler Choice towards Powerful Attention | In recent years, attention mechanisms have demonstrated significant potential in the field of graph representation learning. However, while variants of attention-based GNNs are setting new benchmarks for numerous real-world datasets, recent works have pointed out that their induced attentions are less robust and generalizable against noisy graphs due to the lack of direct supervision. In this paper, we present a new framework that utilizes the tool of causality to provide a powerful supervision signal for the learning process of attention functions. Specifically, we estimate the direct causal effect of attention on the final prediction and then maximize such effect to guide attention to attend to more meaningful neighbors. Our method can serve as a plug-and-play module for any canonical attention-based GNNs in an end-to-end fashion. Extensive experiments on a wide range of benchmark datasets illustrated that, by directly supervising attention with our method, the model is able to converge faster with a clearer decision boundary, and thus yields better performances. | ['Xuan Song', 'Shi Han', 'Qiang Fu', 'Lun Du', 'Jiyuan Chen', 'Hongjun Wang'] | 2023-05-22 | null | null | null | null | ['graph-representation-learning'] | ['methodology'] | [ 1.25785619e-01 3.88347685e-01 -3.68997753e-01 -2.50123858e-01
-3.08817416e-01 -1.48946181e-01 6.09981358e-01 1.31336942e-01
-1.39545038e-01 6.46780968e-01 3.47874165e-01 -3.78961295e-01
-4.01252598e-01 -9.57907677e-01 -7.75164306e-01 -5.13601124e-01
-1.99629560e-01 4.29058135e-01 1.72325253e-01 -3.70260745e-01
1.20805614e-01 3.18272978e-01 -1.09036255e+00 -1.88993320e-01
8.47883224e-01 6.95310414e-01 2.32912675e-01 3.36735189e-01
1.17948152e-01 1.09004879e+00 -3.75958592e-01 -4.93889451e-01
6.75708801e-02 -4.05479550e-01 -9.52629983e-01 -1.20673530e-01
2.91131794e-01 4.16463055e-02 -6.81765497e-01 1.06932783e+00
4.74511802e-01 3.29713017e-01 4.23914284e-01 -1.14433372e+00
-1.18138349e+00 9.87961054e-01 -7.31635451e-01 3.82847786e-01
2.91913748e-01 2.44521171e-01 1.65460896e+00 -6.74707830e-01
5.00631928e-01 1.58381164e+00 5.81765115e-01 4.95995075e-01
-1.53629708e+00 -5.07650971e-01 6.45823836e-01 3.78214896e-01
-9.94806230e-01 -1.06160641e-01 9.24848020e-01 -2.75844429e-03
8.90454888e-01 1.42433494e-01 4.56812143e-01 1.20821559e+00
-7.60189677e-03 9.83988583e-01 7.30318666e-01 -3.37024868e-01
-6.33994564e-02 -2.41497606e-01 1.67134255e-01 7.34703481e-01
2.75873721e-01 -1.41846220e-04 -3.00531685e-01 -1.28133699e-01
8.82730782e-01 1.97123021e-01 -5.53435624e-01 -3.32102835e-01
-1.06173015e+00 9.93981242e-01 1.29554439e+00 3.28486472e-01
-4.35606778e-01 5.91981292e-01 2.31576055e-01 1.45490006e-01
6.20306134e-01 6.14509284e-01 -3.43785405e-01 1.49205074e-01
-3.78945619e-01 9.69753936e-02 6.71394706e-01 5.46830118e-01
6.91344738e-01 -1.13945425e-01 -5.17285168e-01 8.40907097e-01
3.31126511e-01 4.98382486e-02 2.54881948e-01 -5.72889566e-01
4.60740864e-01 7.46190608e-01 -2.96281427e-01 -1.22403586e+00
-4.41087842e-01 -7.42182255e-01 -9.28767383e-01 -3.82529907e-02
2.90842026e-01 -1.60271116e-02 -1.14427018e+00 2.01892185e+00
2.70514935e-01 5.29335856e-01 -3.78433853e-01 1.04433215e+00
5.76793373e-01 5.20621181e-01 2.37045452e-01 2.25966930e-01
1.11483026e+00 -1.15680814e+00 -6.47130191e-01 -4.27678883e-01
4.87061530e-01 -2.37529680e-01 1.26255906e+00 2.49711215e-01
-8.26673448e-01 -4.90513623e-01 -8.20997119e-01 -1.86632067e-01
-1.46905690e-01 -2.96923757e-01 9.55889285e-01 2.82394171e-01
-1.14119041e+00 8.83431375e-01 -8.93281460e-01 -5.63484311e-01
5.62371194e-01 4.50120181e-01 -2.20729709e-01 -8.39725584e-02
-1.29983723e+00 8.48660529e-01 2.77897596e-01 2.68389583e-01
-8.85616839e-01 -6.10636532e-01 -8.64640772e-01 4.31431711e-01
6.47386968e-01 -9.67778742e-01 9.55304921e-01 -7.91979373e-01
-1.37452388e+00 3.53460699e-01 2.91127078e-02 -6.63200676e-01
3.35275084e-01 -2.61755675e-01 -1.15785941e-01 -9.96422619e-02
1.13258392e-01 4.83080089e-01 7.30800271e-01 -1.04812694e+00
-3.78187209e-01 -3.14176828e-01 3.70896488e-01 -8.08414966e-02
-4.77444798e-01 -1.27691031e-01 -7.40885019e-01 -7.51377225e-01
-8.78105983e-02 -8.44773412e-01 -6.92902386e-01 -1.56851932e-01
-5.25877237e-01 -4.60280418e-01 7.57814825e-01 -2.89121389e-01
1.33786345e+00 -1.99955964e+00 3.92603606e-01 2.89987355e-01
6.04923368e-01 3.83902609e-01 -2.98218042e-01 4.96521235e-01
-2.14708805e-01 4.05995607e-01 -1.49424359e-01 -3.17217439e-01
-1.24276735e-01 2.85870284e-01 -2.37669468e-01 3.19483757e-01
5.05367517e-01 1.18438208e+00 -1.26090896e+00 -1.00006022e-01
1.17848158e-01 5.86989105e-01 -6.48607075e-01 2.34555155e-01
-3.84964854e-01 4.43510085e-01 -6.59462512e-01 3.36563587e-01
1.68644145e-01 -6.86795592e-01 4.28642809e-01 1.46211714e-01
4.47113752e-01 4.52302426e-01 -8.63792539e-01 1.52725065e+00
-5.23943186e-01 6.36868894e-01 -2.66185421e-02 -1.11016309e+00
9.43292856e-01 9.36450530e-03 1.35600775e-01 -5.48315644e-01
2.19252214e-01 -2.10816577e-01 2.64496744e-01 -3.00627649e-01
3.45031977e-01 1.84999481e-01 1.08894400e-01 3.81537408e-01
1.15919590e-01 3.38059247e-01 3.03419679e-01 3.96434844e-01
1.32520723e+00 -3.75621207e-02 2.35046312e-01 -1.37504399e-01
5.21238029e-01 -4.02163357e-01 4.95269507e-01 8.61788690e-01
-1.66650489e-03 5.50125599e-01 8.35962534e-01 -4.39107329e-01
-7.43951559e-01 -6.17667377e-01 2.53116190e-01 1.31751239e+00
1.57217067e-02 -5.34142911e-01 -4.52713490e-01 -9.57403421e-01
7.12783933e-02 6.01712763e-01 -9.09794867e-01 -4.89791870e-01
-5.15882075e-01 -7.66843975e-01 3.32920879e-01 8.93465936e-01
1.88514918e-01 -1.22341597e+00 -1.91842169e-02 3.23394239e-01
-3.60385552e-02 -9.05878603e-01 -4.71881956e-01 2.35384837e-01
-7.93930054e-01 -1.11562502e+00 -6.65475905e-01 -6.78189635e-01
6.43495560e-01 3.63916010e-01 1.18935323e+00 4.53306228e-01
-7.30719939e-02 1.43709481e-01 -3.34288388e-01 -1.92344233e-01
-1.84239119e-01 3.61056060e-01 -2.68395513e-01 2.37193584e-01
1.19047545e-01 -8.44987690e-01 -5.59862077e-01 2.90256649e-01
-8.30889583e-01 -2.53606379e-01 6.06846213e-01 1.16851747e+00
2.77714700e-01 -3.21223170e-01 7.64800370e-01 -1.17565441e+00
9.08076346e-01 -7.37950265e-01 -6.02942228e-01 1.19062886e-01
-8.40961576e-01 4.55837518e-01 9.01416183e-01 -1.45271659e-01
-7.41406739e-01 -2.42751524e-01 -7.54777044e-02 -6.34988666e-01
9.31190476e-02 8.83401930e-01 -2.06486419e-01 -3.60368565e-02
5.85784853e-01 -1.32533655e-01 -7.97129124e-02 -5.81425607e-01
4.98436093e-01 1.60781249e-01 4.22180504e-01 -5.17382026e-01
7.98381805e-01 3.52403462e-01 8.20041746e-02 -4.21872169e-01
-8.63569379e-01 -4.43237782e-01 -3.59625250e-01 -1.50819437e-03
6.12630188e-01 -7.02836871e-01 -7.39351571e-01 6.65962920e-02
-9.91088390e-01 -3.55281800e-01 -1.08415186e-01 2.17857599e-01
-2.01510370e-01 3.94968361e-01 -5.95112264e-01 -5.94469488e-01
-2.80477464e-01 -1.04724741e+00 8.72427583e-01 1.66903302e-01
-1.10959984e-01 -1.29908359e+00 4.33473736e-02 1.73887551e-01
3.63830447e-01 1.75872371e-01 1.02172363e+00 -6.78303182e-01
-7.46187091e-01 -1.98017180e-01 -4.06198561e-01 1.92534104e-01
9.06967819e-02 6.13255613e-02 -8.76520216e-01 -3.06601763e-01
-4.78082865e-01 -1.07118919e-01 1.38156021e+00 3.12596232e-01
1.33598709e+00 -2.83184081e-01 -4.63684112e-01 6.78515136e-01
1.39173222e+00 -2.41961733e-01 7.00118363e-01 1.74089983e-01
9.76259768e-01 2.69782543e-01 3.55221093e-01 1.51674941e-01
4.90953892e-01 6.55934274e-01 9.11844552e-01 -4.06009465e-01
-2.33561143e-01 -4.71958965e-01 2.61545807e-01 5.07594705e-01
-2.90735692e-01 -6.17865682e-01 -8.61617982e-01 6.67587280e-01
-2.23267794e+00 -8.06051850e-01 -2.48101890e-01 2.07209206e+00
5.44953465e-01 2.49148071e-01 1.41676694e-01 3.17655019e-02
8.44621420e-01 4.69683558e-01 -4.78053629e-01 -1.70210600e-01
1.99139088e-01 1.99856684e-01 4.04450595e-01 3.77746314e-01
-9.33639586e-01 1.03558993e+00 6.48560762e+00 6.88526392e-01
-1.15293455e+00 -1.05415188e-01 6.02838755e-01 1.37843594e-01
-4.70868081e-01 1.46093130e-01 -4.87255067e-01 3.34365517e-01
7.47962534e-01 -1.55010551e-01 5.29392779e-01 8.05186093e-01
2.26972565e-01 3.03826094e-01 -1.22741687e+00 5.98222136e-01
-1.95815176e-01 -1.34954751e+00 1.47655711e-01 1.77245155e-01
7.93870509e-01 2.56465018e-01 1.08238198e-02 3.69324952e-01
7.57586181e-01 -1.07572925e+00 4.69915062e-01 4.14538562e-01
4.45544720e-01 -6.91205800e-01 7.52129316e-01 2.23699048e-01
-1.09550464e+00 -2.11298153e-01 -5.22676885e-01 -4.06132400e-01
-7.64987692e-02 5.99080741e-01 -1.19016695e+00 7.63940096e-01
4.31818485e-01 9.83734787e-01 -6.95846498e-01 1.14155722e+00
-7.98679292e-01 8.29490960e-01 -4.74456139e-02 -3.15218896e-01
4.05223191e-01 -1.33862346e-01 6.00927949e-01 1.03900480e+00
-9.41610560e-02 -1.43890545e-01 2.62990654e-01 1.04226923e+00
-4.99218762e-01 -6.77565811e-03 -5.50381541e-01 -2.33488664e-01
2.13570878e-01 1.35659921e+00 -8.64915609e-01 -1.37020081e-01
-4.40463871e-01 6.30286515e-01 9.19134259e-01 5.18995762e-01
-1.05654633e+00 -3.46465409e-01 5.83007991e-01 1.81832969e-01
6.58874035e-01 -9.56727043e-02 -2.32270300e-01 -1.06906593e+00
-1.54929951e-01 -6.61850274e-01 5.10176539e-01 -6.73023582e-01
-1.50168598e+00 5.75479567e-01 -3.49421263e-01 -6.93425894e-01
-4.01856489e-02 -5.63425779e-01 -9.69996512e-01 8.78848791e-01
-1.64026713e+00 -1.08839381e+00 -3.17542106e-01 5.13305426e-01
3.35540295e-01 2.52974063e-01 6.15338147e-01 2.80982256e-01
-8.44309628e-01 6.13781869e-01 5.14674373e-02 3.28544408e-01
5.92321634e-01 -1.43783236e+00 5.97500443e-01 9.98662889e-01
4.80716676e-01 6.94708347e-01 6.61438763e-01 -5.11605263e-01
-1.32641912e+00 -1.27895617e+00 6.28132105e-01 -3.57315451e-01
9.31269825e-01 -4.43642884e-01 -9.85338390e-01 9.34216619e-01
1.91954821e-01 2.01818049e-01 2.75856286e-01 7.59284854e-01
-4.02058452e-01 -2.23399729e-01 -6.61382616e-01 5.12383819e-01
1.24549758e+00 -3.60291004e-01 -3.20539325e-01 2.14622065e-01
8.61994565e-01 -2.71585226e-01 -6.16573691e-01 2.84259886e-01
1.58459246e-01 -8.26831818e-01 8.47744644e-01 -8.01124752e-01
6.21486425e-01 -2.35578433e-01 3.68872315e-01 -1.54261637e+00
-9.17147458e-01 -7.47852504e-01 -2.29661778e-01 1.26890337e+00
5.80059111e-01 -7.67840326e-01 5.84737122e-01 3.26980948e-01
-1.93345502e-01 -1.01495945e+00 -5.76380312e-01 -5.91573119e-01
-2.51996517e-01 -3.92194659e-01 5.74702442e-01 8.77014637e-01
-2.45221723e-02 6.25757754e-01 -5.09458959e-01 2.37818688e-01
3.76142144e-01 1.34765640e-01 7.64424980e-01 -1.38066721e+00
-4.42097664e-01 -5.58721602e-01 -5.67338586e-01 -1.19794118e+00
3.01955074e-01 -1.19787002e+00 2.21614242e-02 -1.76345718e+00
3.11884552e-01 -4.24426675e-01 -7.64216006e-01 7.48352468e-01
-8.04109693e-01 1.21989645e-01 3.31322998e-01 -3.69980708e-02
-7.03119040e-01 6.22695625e-01 1.28114891e+00 -2.10430935e-01
-1.03256828e-03 2.95155067e-02 -1.19481981e+00 5.68608582e-01
5.53132355e-01 -5.65487027e-01 -4.91808057e-01 -6.01665258e-01
2.77400076e-01 -2.75238454e-01 5.30943751e-01 -6.82018340e-01
2.52666056e-01 -6.47463202e-02 1.88561127e-01 -6.44513965e-02
2.02560890e-02 -6.89520478e-01 -1.66988745e-01 2.47688800e-01
-4.32074219e-01 -1.55044720e-01 -3.50586958e-02 1.01587772e+00
-1.41328663e-01 9.66251548e-03 5.52948594e-01 -6.12787604e-02
-6.76054180e-01 5.87338746e-01 5.37053235e-02 2.37116218e-01
7.65548468e-01 9.31759775e-02 -2.21739709e-01 -5.98012626e-01
-6.46774650e-01 5.09257138e-01 2.72779137e-01 5.40237188e-01
3.94617081e-01 -1.30233943e+00 -7.32477725e-01 9.19127613e-02
-7.71764219e-02 7.34343678e-02 -1.12493388e-01 8.17440689e-01
-4.26598452e-03 4.19148564e-01 2.45554280e-02 -6.01369381e-01
-9.22388256e-01 7.79539466e-01 2.32136577e-01 -7.05228031e-01
-6.05039954e-01 1.07482219e+00 4.36688989e-01 -2.12148711e-01
3.65720540e-01 -5.76747537e-01 -1.94270581e-01 -1.53243348e-01
4.14684087e-01 2.29791343e-01 7.30119720e-02 -2.20900178e-01
-3.64638507e-01 2.07393467e-01 -9.77850780e-02 3.82159472e-01
1.61342537e+00 2.61547361e-02 -4.39794771e-02 1.53918013e-01
9.17627871e-01 -8.21503252e-02 -1.39675617e+00 -3.51716548e-01
2.18611494e-01 -4.91360039e-01 1.40267253e-01 -7.39955664e-01
-1.46208251e+00 7.77739823e-01 -3.31578515e-02 6.72642946e-01
1.14300394e+00 1.93181545e-01 6.68884933e-01 1.26599818e-01
1.43958241e-01 -5.58082163e-01 1.22683659e-01 5.76928258e-01
8.37431908e-01 -1.20515275e+00 -1.19023800e-01 -4.68934774e-01
-4.07330781e-01 1.02187943e+00 6.01827502e-01 -3.15081954e-01
4.71177638e-01 -1.54679250e-02 -3.00637066e-01 -3.75068784e-01
-1.07590020e+00 -4.11104143e-01 3.55015755e-01 5.79259634e-01
5.75785995e-01 -9.02315155e-02 -1.73837245e-01 5.39254427e-01
1.63375899e-01 -9.04045999e-02 4.98273402e-01 4.62332547e-01
-2.01928794e-01 -1.12873399e+00 -1.19952269e-01 3.08090478e-01
-4.69415545e-01 -3.16099823e-01 -4.41051215e-01 8.37145686e-01
-2.03696072e-01 8.32511723e-01 -1.28844246e-01 -3.07508975e-01
4.61194247e-01 -1.67972356e-01 3.42380941e-01 -7.37888277e-01
-6.32666409e-01 -7.91463107e-02 1.52617991e-01 -7.13226497e-01
-2.48548061e-01 -3.06335539e-01 -1.13490868e+00 -2.60261774e-01
-6.31412148e-01 1.77772000e-01 1.77287892e-01 8.93526614e-01
6.58308923e-01 9.93184388e-01 6.76914632e-01 -8.98597300e-01
-5.75951695e-01 -1.01406479e+00 -3.44525069e-01 4.90649134e-01
3.25550377e-01 -7.16037691e-01 -2.73082256e-01 -3.17479491e-01] | [7.198451519012451, 6.1893815994262695] |
58d11aa5-4918-430e-b4d3-df2cd5e05156 | active-learning-for-multilingual-semantic | 2301.12920 | null | https://arxiv.org/abs/2301.12920v3 | https://arxiv.org/pdf/2301.12920v3.pdf | Active Learning for Multilingual Semantic Parser | Current multilingual semantic parsing (MSP) datasets are almost all collected by translating the utterances in the existing datasets from the resource-rich language to the target language. However, manual translation is costly. To reduce the translation effort, this paper proposes the first active learning procedure for MSP (AL-MSP). AL-MSP selects only a subset from the existing datasets to be translated. We also propose a novel selection method that prioritizes the examples diversifying the logical form structures with more lexical choices, and a novel hyperparameter tuning method that needs no extra annotation cost. Our experiments show that AL-MSP significantly reduces translation costs with ideal selection methods. Our selection method with proper hyperparameters yields better parsing performance than the other baselines on two multilingual datasets. | ['Gholamreza Haffari', 'Zhuang Li'] | 2023-01-30 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 1.34337798e-01 3.77005547e-01 -7.31624901e-01 -7.78499722e-01
-1.68138087e+00 -7.77602255e-01 1.20197572e-01 -1.21160284e-01
-5.86702585e-01 1.18788600e+00 3.41929704e-01 -5.24012625e-01
2.38356948e-01 -7.94138014e-01 -8.50034416e-01 -4.31416124e-01
4.91508037e-01 8.51530015e-01 3.85413289e-01 -2.59744644e-01
-9.93767008e-02 2.35672487e-04 -8.80853176e-01 5.73711455e-01
1.56836522e+00 3.55063081e-01 6.68338954e-01 1.00411437e-01
-7.90841520e-01 4.09426451e-01 -5.87710500e-01 -8.12391222e-01
4.83963072e-01 -5.22226751e-01 -1.11578774e+00 -2.00349763e-01
2.70266026e-01 -1.46659076e-01 4.07210201e-01 1.01770413e+00
5.64161897e-01 4.81018461e-02 7.20974058e-02 -1.06057274e+00
-4.48436260e-01 1.62274790e+00 -3.37818086e-01 9.30500776e-03
4.06697750e-01 -6.03308789e-02 1.27462447e+00 -1.01991081e+00
9.40295637e-01 1.54100788e+00 5.17163217e-01 7.33164370e-01
-1.20098639e+00 -8.98829222e-01 2.89758831e-01 3.23412180e-01
-1.05524206e+00 -7.19802737e-01 8.04298222e-01 1.09035738e-01
1.23606586e+00 1.91062167e-01 2.47154117e-01 1.16302657e+00
-2.90321201e-01 1.00069106e+00 1.16971123e+00 -9.84471262e-01
2.76466787e-01 2.39232436e-01 2.34020963e-01 6.93540335e-01
3.24316561e-01 -3.03346366e-01 -7.37322569e-01 -3.52040023e-01
3.32896024e-01 -7.56573915e-01 3.62505764e-02 -1.78629652e-01
-1.21047199e+00 8.95036101e-01 -2.13706776e-01 8.20234492e-02
-1.03168547e-01 -2.59604305e-01 3.22150558e-01 4.57500368e-01
6.21121645e-01 6.19166136e-01 -1.09835291e+00 -1.31329849e-01
-4.69758183e-01 5.71637005e-02 8.42706203e-01 1.43379164e+00
1.19743836e+00 -2.24919707e-01 2.30107576e-01 1.27554548e+00
3.04021060e-01 5.96152425e-01 3.70371312e-01 -9.88028526e-01
1.20963776e+00 8.67136717e-01 -1.70385074e-02 -3.26665103e-01
-2.44984567e-01 -2.87889242e-02 1.76093102e-01 -4.02604938e-01
1.98775828e-01 -5.80001116e-01 -3.44969749e-01 1.90094471e+00
4.38941300e-01 -3.36939245e-01 3.16593856e-01 5.65580726e-01
9.68525410e-01 7.44709194e-01 3.58429044e-01 -5.32068312e-01
1.08393991e+00 -1.13404894e+00 -9.83682096e-01 -4.71191794e-01
1.01031113e+00 -9.81381893e-01 1.65281749e+00 9.40727964e-02
-9.91590738e-01 -2.31580451e-01 -1.04334652e+00 -1.27673924e-01
-1.64790034e-01 3.13191235e-01 1.05539274e+00 6.50868833e-01
-8.75104547e-01 1.54811963e-01 -9.22951400e-01 -4.10851508e-01
1.97168469e-01 3.65746319e-01 -3.78687292e-01 1.53534815e-01
-1.40326369e+00 8.68653655e-01 6.55586720e-01 -2.79662311e-01
-3.32971424e-01 -5.67437768e-01 -1.06455398e+00 -1.99455634e-01
5.86037219e-01 -4.92422163e-01 1.47888517e+00 -1.31394029e+00
-1.85357451e+00 7.70757079e-01 -4.85412836e-01 -1.68338582e-01
4.40686405e-01 -5.07330298e-01 -2.42615864e-01 -1.15810940e-02
4.26421523e-01 8.94205451e-01 3.20782453e-01 -1.26801753e+00
-6.70695305e-01 -1.68183818e-01 1.86389193e-01 6.07830048e-01
-3.40578914e-01 3.82552624e-01 -7.09316969e-01 -5.26865363e-01
3.47542256e-01 -1.08532727e+00 -2.40120739e-01 -5.29267132e-01
-3.31743479e-01 -4.10213858e-01 4.92971867e-01 -6.65071070e-01
1.28929770e+00 -1.88598156e+00 1.72113657e-01 -2.32741490e-01
-4.86681581e-01 7.71879554e-02 -3.15138459e-01 4.65662450e-01
2.44946644e-01 1.26891360e-01 -4.94336903e-01 -4.60849047e-01
-5.76980077e-02 5.48030138e-01 -2.92183667e-01 -1.42267764e-01
1.55726284e-01 8.54182303e-01 -9.95789766e-01 -1.04397452e+00
-1.72540292e-01 -1.30085304e-01 -6.57839179e-01 2.59717166e-01
-5.07111907e-01 5.00273883e-01 -5.26436925e-01 8.18247318e-01
5.94538748e-01 -1.71758607e-02 7.53155649e-01 -1.04013689e-01
-8.01062137e-02 8.85480523e-01 -9.74874079e-01 2.04894543e+00
-7.88869858e-01 1.98885918e-01 -1.88915774e-01 -7.40081966e-01
1.05834317e+00 3.34340096e-01 2.29859591e-01 -7.89908886e-01
-1.88100696e-01 7.38229990e-01 8.97388607e-02 -5.96125960e-01
5.07102966e-01 1.04852699e-01 -4.68399942e-01 5.14271677e-01
2.12821156e-01 1.46257922e-01 4.45484281e-01 1.19037226e-01
7.62694120e-01 6.80727065e-01 4.45092589e-01 -3.93520623e-01
5.22655606e-01 4.29821700e-01 1.39285755e+00 5.44858456e-01
-2.66352624e-01 1.06834859e-01 3.99810791e-01 -2.12018833e-01
-8.60141456e-01 -9.03351724e-01 -5.85320666e-02 1.54870486e+00
1.90867037e-01 -4.26031768e-01 -9.83238697e-01 -1.32600212e+00
-4.61701334e-01 8.44299138e-01 9.25929770e-02 2.71289140e-01
-1.39356804e+00 -1.07280016e+00 6.95604205e-01 4.37510073e-01
6.66304529e-01 -1.11821532e+00 -2.08263516e-01 2.87495494e-01
-7.61703074e-01 -1.29959500e+00 -2.76328951e-01 2.18705490e-01
-8.37929308e-01 -9.12062109e-01 -5.11925966e-02 -1.31028521e+00
7.80903518e-01 -1.17990628e-01 1.21916366e+00 -2.19439581e-01
3.61839384e-01 -2.53895938e-01 -7.43804812e-01 -2.80070573e-01
-6.75118566e-01 5.22292376e-01 -2.18663827e-01 -3.36524069e-01
5.97824931e-01 -1.33244827e-01 -9.05727372e-02 1.73543692e-01
-4.07396466e-01 3.22331995e-01 5.97706199e-01 7.45114982e-01
7.63577998e-01 -3.01976562e-01 8.02740276e-01 -1.54037273e+00
5.84124327e-01 -2.90074915e-01 -5.32340229e-01 6.77390456e-01
-5.23227394e-01 2.08267838e-01 8.00484776e-01 -3.56801838e-01
-1.62578046e+00 5.32581389e-01 -1.70024708e-01 4.37761694e-01
1.83552355e-02 3.25562656e-01 -9.25282419e-01 2.62645960e-01
5.65320611e-01 -6.30725771e-02 -5.17582178e-01 -7.36292899e-01
4.08798397e-01 7.73779273e-01 1.75188005e-01 -1.09024465e+00
4.56284165e-01 -9.67395231e-02 -4.60866839e-01 -4.55974489e-01
-7.35189199e-01 -1.37692750e-01 -8.93652439e-01 1.57854781e-01
5.46079278e-01 -8.90536427e-01 1.46513358e-01 2.73205712e-02
-1.29305303e+00 -4.79659528e-01 3.06952540e-02 6.56051159e-01
-5.19443095e-01 4.50957417e-01 -8.21636617e-01 -4.46917951e-01
-5.80243051e-01 -1.58726263e+00 9.81835485e-01 -7.57757947e-02
-6.17174327e-01 -9.60292459e-01 2.11925536e-01 7.49946117e-01
1.81564599e-01 -1.83334842e-01 1.35971844e+00 -9.42826211e-01
-6.22293830e-01 4.87185597e-01 9.21208560e-02 -1.65990286e-03
3.38453472e-01 -6.57008514e-02 -5.87358236e-01 -1.13586150e-01
-2.73826063e-01 -3.66551429e-01 5.08923471e-01 1.47135347e-01
7.16088295e-01 -5.91110826e-01 -2.62822270e-01 6.01739824e-01
1.48675358e+00 6.85880303e-01 2.57755458e-01 7.23681867e-01
5.03029406e-01 6.82569563e-01 9.47737455e-01 -1.86409846e-01
6.01531982e-01 7.59648621e-01 -2.64508799e-02 -1.06187709e-01
-4.76882160e-02 -3.58148068e-01 7.26258099e-01 1.47102964e+00
4.27991778e-01 -1.54288650e-01 -1.12983608e+00 4.52942431e-01
-1.93533587e+00 -4.39571202e-01 1.09840579e-01 2.07473826e+00
1.60561597e+00 1.92609906e-01 -1.18503734e-01 -2.68045425e-01
8.34936082e-01 -3.77263874e-02 -5.06178558e-01 -5.19850969e-01
-3.83612186e-01 3.71046931e-01 4.78529334e-01 9.96641695e-01
-1.22740877e+00 1.62183857e+00 6.48835564e+00 6.73482835e-01
-9.53342319e-01 4.74254429e-01 4.13758576e-01 7.74031729e-02
-6.46878600e-01 5.26135325e-01 -1.21759927e+00 5.14239371e-01
7.62342572e-01 -1.13249533e-01 1.35036051e-01 9.90446806e-01
2.27451891e-01 2.00087488e-01 -1.09079683e+00 6.93697155e-01
-8.01585615e-02 -1.19072485e+00 3.85503471e-01 -4.28155035e-01
5.99455893e-01 -5.17178923e-02 -5.12639284e-01 4.64189321e-01
6.21477187e-01 -4.34030086e-01 7.20711589e-01 -1.25189021e-01
7.00188458e-01 -7.31650770e-01 7.16644228e-01 3.74803901e-01
-1.03172827e+00 4.48338129e-02 -4.56865191e-01 2.72292197e-01
4.67314988e-01 3.03681403e-01 -1.00973988e+00 5.43026030e-01
3.88490707e-01 4.70691592e-01 -6.08807504e-01 5.52395105e-01
-5.90969741e-01 1.01306331e+00 -1.66952118e-01 -1.25216618e-01
3.60617563e-02 -2.79114902e-01 7.30243206e-01 1.56935644e+00
1.20692879e-01 -1.47864506e-01 5.61882257e-01 6.45672381e-01
-1.51247233e-01 7.19095290e-01 -3.92099768e-01 4.89283390e-02
1.19034851e+00 1.07659912e+00 -7.16661990e-01 -4.84541535e-01
-5.12669802e-01 8.02061856e-01 6.02494061e-01 2.23188415e-01
-6.64760232e-01 -4.33988929e-01 2.69547909e-01 -1.55899972e-01
-1.22098997e-01 -1.89949453e-01 -3.70654374e-01 -1.36340296e+00
2.68790662e-01 -1.21027803e+00 5.16227663e-01 -3.80640417e-01
-1.13529003e+00 8.08596671e-01 2.45470032e-01 -1.14530432e+00
-3.21868300e-01 -3.82803738e-01 -2.65492171e-01 7.08240390e-01
-1.43835294e+00 -1.15388024e+00 2.54321188e-01 3.18144798e-01
1.11407316e+00 -3.68417442e-01 9.79556203e-01 4.91854399e-01
-8.92126560e-01 9.61095989e-01 -3.19730788e-01 3.78181577e-01
8.47298741e-01 -1.28144491e+00 5.94109058e-01 1.09248006e+00
4.21583876e-02 9.40216243e-01 4.97974575e-01 -7.54244745e-01
-1.33882236e+00 -9.28587556e-01 1.50461841e+00 -1.52990997e-01
5.33519745e-01 -4.87067729e-01 -8.80330384e-01 1.00951564e+00
4.11538571e-01 -5.53895235e-01 8.99436176e-01 1.78440243e-01
-3.55608761e-01 -2.50761598e-01 -1.11778748e+00 6.85713649e-01
1.15643108e+00 -3.73543054e-01 -9.84346867e-01 4.11872447e-01
1.14254236e+00 -3.62356961e-01 -6.29730284e-01 6.10374749e-01
2.45976374e-01 -3.91777843e-01 5.99255443e-01 -5.95822275e-01
3.10363859e-01 -1.10144235e-01 -4.18227613e-01 -1.36070788e+00
-1.80534229e-01 -8.19823802e-01 2.03955829e-01 1.67298508e+00
1.22916222e+00 -7.68396497e-01 6.98055744e-01 5.01331806e-01
-3.56749654e-01 -6.60247922e-01 -8.40235472e-01 -7.17756510e-01
1.94602847e-01 -1.73990831e-01 8.94070685e-01 1.27676356e+00
1.80298805e-01 6.90895259e-01 -4.74152900e-02 2.58189104e-02
7.73283422e-01 2.25188866e-01 6.74586773e-01 -8.38621438e-01
-5.22378385e-01 -7.63992891e-02 2.27735087e-01 -9.95877028e-01
5.37016153e-01 -1.10219848e+00 1.54908001e-01 -1.54181397e+00
2.60491222e-01 -9.25263941e-01 -5.79136685e-02 8.76488507e-01
-5.30357182e-01 -3.12016845e-01 4.20903489e-02 3.26105148e-01
-6.14729464e-01 2.89570838e-01 9.45856214e-01 -4.50028479e-02
-5.31456470e-01 -3.29764485e-01 -6.71077311e-01 7.41405964e-01
1.06588566e+00 -6.48593307e-01 -6.49425030e-01 -1.03111446e+00
4.98558104e-01 1.65721431e-01 -5.76782763e-01 -3.99719000e-01
-7.82719627e-02 -5.01913488e-01 -4.11992967e-02 -6.57864809e-01
-9.00235549e-02 -5.44473827e-01 1.00102983e-01 1.58617079e-01
-7.16076076e-01 3.30118477e-01 9.40527208e-03 8.63143522e-03
-1.11015446e-01 -4.52348888e-01 7.38039732e-01 -4.82731283e-01
-1.04433262e+00 2.08088700e-02 -1.63253739e-01 2.41902202e-01
6.04976654e-01 6.71916306e-02 -4.18774188e-01 1.67988390e-01
-5.62809050e-01 2.07511634e-01 4.25421715e-01 6.60727441e-01
2.01402634e-01 -1.24743330e+00 -7.75390208e-01 1.78916946e-01
1.79159820e-01 1.12575904e-01 -3.52558672e-01 4.83010679e-01
-8.39311779e-01 4.24548984e-01 -8.74873474e-02 -5.01749039e-01
-1.38377273e+00 4.94012311e-02 1.41901150e-01 -3.05580527e-01
-6.04512334e-01 8.40323269e-01 -4.02406231e-02 -1.15941000e+00
1.15978554e-01 2.78420700e-03 -3.27373922e-01 1.60562750e-02
1.33452624e-01 3.35712880e-01 2.45094493e-01 -6.37311041e-01
-3.99281859e-01 3.50515127e-01 -2.37542197e-01 -3.42614323e-01
1.25637698e+00 -3.38732958e-01 -3.99439037e-01 4.51485515e-01
1.09974885e+00 4.53307152e-01 -9.51667309e-01 -5.99617302e-01
6.51676774e-01 -3.22595000e-01 -3.07941467e-01 -9.77738142e-01
-8.76413405e-01 3.36059213e-01 2.34017238e-01 -3.99911016e-01
9.92107987e-01 4.76650335e-02 1.01149738e+00 6.87889636e-01
9.60116923e-01 -1.45522153e+00 -2.34065861e-01 6.25821590e-01
6.23300672e-01 -1.24896300e+00 -7.49063119e-02 -9.11277592e-01
-7.54981935e-01 9.09778655e-01 9.25349832e-01 2.82266945e-01
2.46688500e-01 4.76682723e-01 6.55803084e-01 2.66764432e-01
-1.00306797e+00 -5.85646881e-03 -2.69019514e-01 3.17131877e-01
6.91877067e-01 3.00697386e-01 -1.08587468e+00 7.68870115e-01
-5.39209306e-01 -3.72388065e-01 2.73084730e-01 1.11843419e+00
-4.07343686e-01 -1.85798514e+00 -7.66702518e-02 -5.74885346e-02
-5.00606716e-01 -2.13427201e-01 -7.88871288e-01 7.40405619e-01
3.27472568e-01 9.97159898e-01 -2.93284431e-02 -5.60987741e-02
3.00064236e-01 5.42399287e-01 5.46421707e-01 -9.34334338e-01
-6.47915959e-01 2.16762513e-01 7.21943200e-01 -4.63063449e-01
-5.90108752e-01 -6.86941624e-01 -1.65909266e+00 1.54765949e-01
-4.44663048e-01 3.44845861e-01 4.96475726e-01 9.15385604e-01
3.48696589e-01 -4.25962210e-02 6.66750252e-01 -5.41604310e-02
-4.52975541e-01 -9.28542316e-01 4.51894626e-02 3.88452232e-01
-2.58832961e-01 -5.19970179e-01 -1.97517931e-01 1.53085113e-01] | [11.256261825561523, 10.061027526855469] |
06e3c1c7-d219-4af0-bde5-9d3c2d77e900 | zero-shot-causal-learning | 2301.12292 | null | https://arxiv.org/abs/2301.12292v2 | https://arxiv.org/pdf/2301.12292v2.pdf | Zero-shot causal learning | Predicting how different interventions will causally affect a specific individual is important in a variety of domains such as personalized medicine, public policy, and online marketing. There are a large number of methods to predict the effect of an existing intervention based on historical data from individuals who received it. However, in many settings it is important to predict the effects of novel interventions (\emph{e.g.}, a newly invented drug), which these methods do not address. Here, we consider zero-shot causal learning: predicting the personalized effects of a novel intervention. We propose CaML, a causal meta-learning framework which formulates the personalized prediction of each intervention's effect as a task. CaML trains a single meta-model across thousands of tasks, each constructed by sampling an intervention, along with its recipients and nonrecipients. By leveraging both intervention information (\emph{e.g.}, a drug's attributes) and individual features~(\emph{e.g.}, a patient's history), CaML is able to predict the personalized effects of novel interventions that do not exist at the time of training. Experimental results on real world datasets in large-scale medical claims and cell-line perturbations demonstrate the effectiveness of our approach. Most strikingly, CaML's zero-shot predictions outperform even strong baselines trained directly on data from the test interventions. | ['Jure Leskovec', 'Sara Oblak', 'Michihiro Yasunaga', 'Anja Šurina', 'Yining Chen', 'Yusuf Roohani', 'Michael Moor', 'Hamed Nilforoshan'] | 2023-01-28 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [ 6.62474573e-01 1.50578737e-01 -9.52166021e-01 -1.89363226e-01
-7.70305634e-01 -2.74705917e-01 7.43704498e-01 6.60139561e-01
-2.21166208e-01 1.10682917e+00 7.47914910e-01 -4.84161168e-01
-3.50185513e-01 -9.06455159e-01 -1.17243242e+00 -5.78257859e-01
-1.62804946e-01 5.77473223e-01 -1.51653007e-01 5.57596125e-02
2.33541146e-01 -1.61625575e-02 -1.22055423e+00 5.81334710e-01
9.07254100e-01 4.39067364e-01 5.89030869e-02 5.50238252e-01
1.85040697e-01 7.18470693e-01 -3.47232074e-01 -3.56376112e-01
-1.97637275e-01 -4.65858221e-01 -6.20098829e-01 -3.03258479e-01
3.54251526e-02 -4.47314948e-01 -4.52708572e-01 5.72401643e-01
7.88036764e-01 -6.77545667e-02 8.97934854e-01 -1.03287148e+00
-7.78146923e-01 9.77971256e-01 -5.36230206e-01 2.39050880e-01
5.09567440e-01 4.64695245e-01 1.34367204e+00 -3.91481102e-01
6.73098803e-01 1.21712971e+00 7.34007835e-01 6.69984043e-01
-1.62811041e+00 -8.76099288e-01 3.30835193e-01 1.12204358e-01
-5.55013597e-01 -4.45858598e-01 5.99750936e-01 -7.58167922e-01
6.12330317e-01 1.02660693e-01 2.59952426e-01 1.70319784e+00
6.84511125e-01 7.09864557e-01 9.79693711e-01 -3.61070782e-02
4.31154877e-01 -2.66411453e-01 2.12076157e-01 3.93931001e-01
1.77480467e-03 3.31873029e-01 -5.10157526e-01 -9.02338624e-01
2.96132892e-01 5.10764122e-01 -2.51951516e-01 3.52414623e-02
-1.19296336e+00 9.02175426e-01 1.72690675e-01 1.11809783e-01
-7.54308462e-01 3.43665510e-01 4.23937798e-01 5.91695420e-02
7.17550635e-01 4.89475727e-01 -9.29535866e-01 6.41809329e-02
-3.90526861e-01 4.06386584e-01 5.44421196e-01 7.44483054e-01
5.11317551e-01 -5.61279655e-01 -6.00863338e-01 6.09260976e-01
-1.62784293e-01 2.63032198e-01 4.97086644e-01 -4.60072786e-01
2.47329071e-01 5.09684563e-01 3.59417856e-01 -6.80174410e-01
-6.12272024e-01 -1.68199345e-01 -8.61416042e-01 -4.61481363e-01
2.64088452e-01 -5.16936123e-01 -9.01822507e-01 2.03622723e+00
4.10192817e-01 9.93521929e-01 -2.49965817e-01 3.42187345e-01
9.03012753e-01 5.08883834e-01 6.62625790e-01 -6.07601106e-01
1.09846258e+00 -3.07383597e-01 -6.09744430e-01 -1.34555578e-01
7.42147326e-01 -3.79844785e-01 8.41231942e-01 7.67652020e-02
-7.23117828e-01 -1.47836551e-01 -5.46688735e-01 4.57972258e-01
-3.88985604e-01 -2.25457355e-01 9.25767660e-01 3.51962388e-01
-4.90736365e-01 1.04851615e+00 -3.69025528e-01 -3.00015897e-01
8.39621246e-01 5.37087560e-01 -2.14739874e-01 -1.63301721e-01
-1.54146791e+00 4.19158787e-01 1.00141726e-01 -5.72168708e-01
-1.20023024e+00 -1.64808130e+00 -4.10059959e-01 2.06595361e-01
5.71018934e-01 -1.21366179e+00 1.29174423e+00 -6.85391068e-01
-1.22920990e+00 4.60383803e-01 -1.90003961e-01 -5.40456891e-01
2.41301879e-01 3.80952545e-02 -4.56314474e-01 -2.82475263e-01
3.83861244e-01 1.43579662e-01 7.44070828e-01 -7.55785763e-01
-9.07206833e-01 -5.98092258e-01 3.24247405e-02 -1.77612886e-01
-2.64895856e-01 -7.08720013e-02 -1.07523529e-02 -5.76116145e-01
-7.54723489e-01 -9.25878465e-01 -5.25913656e-01 -5.15700400e-01
-7.13717997e-01 -4.91121322e-01 3.37498605e-01 -4.77725476e-01
1.44931042e+00 -1.79826379e+00 1.83348000e-01 -1.77840069e-01
5.43957651e-01 1.87214255e-01 -1.88902795e-01 7.70861149e-01
-3.92253011e-01 4.55247521e-01 -7.99422711e-02 1.03851059e-03
-1.46651715e-01 -4.48624231e-03 -4.93383408e-01 4.78600740e-01
3.57655473e-02 1.06713784e+00 -1.13895738e+00 -2.39631817e-01
1.08622769e-02 2.33404458e-01 -8.45199764e-01 1.20863929e-01
-7.90036678e-01 6.23038173e-01 -9.36714888e-01 3.90396416e-01
3.51229422e-02 -5.69842100e-01 3.36304784e-01 2.54458804e-02
2.40318224e-01 2.38625243e-01 -6.03015244e-01 1.38739848e+00
-2.16939047e-01 -1.17577925e-01 -7.21366167e-01 -1.04932868e+00
2.15310097e-01 6.84562325e-01 1.11695063e+00 -5.25729835e-01
8.42581317e-02 -2.49846540e-02 9.74788740e-02 -5.39600849e-01
-2.31004730e-01 -5.26079059e-01 -3.49881053e-01 6.59222543e-01
-4.05081846e-02 7.92531550e-01 -1.68227062e-01 2.31818274e-01
1.81013095e+00 -3.20380718e-01 6.90403104e-01 -3.02734543e-02
1.44243568e-01 4.68905866e-02 8.84756207e-01 1.12961185e+00
-1.31928533e-01 1.19677410e-01 6.33729935e-01 -4.75404710e-01
-9.37777460e-01 -8.59738648e-01 -2.26431221e-01 1.34967506e+00
-2.75084287e-01 -3.39731902e-01 -3.91439617e-01 -7.76606917e-01
5.89311242e-01 7.11085737e-01 -1.13142073e+00 -4.52583283e-01
-2.98020452e-01 -1.19053316e+00 2.06771150e-01 2.01715037e-01
-2.89172754e-02 -9.51427341e-01 -7.98304901e-02 6.11248910e-01
-8.17619860e-02 -6.02214813e-01 -6.97630644e-01 -1.10385165e-01
-7.33000338e-01 -1.35063052e+00 -4.90233153e-01 -3.35070252e-01
3.47500384e-01 -6.33896291e-02 1.12644696e+00 -9.13532451e-02
-3.21689159e-01 4.55274075e-01 -2.08827510e-01 -7.73010373e-01
-5.43395340e-01 -2.44349968e-02 2.80532241e-01 1.59252524e-01
3.54537606e-01 -7.08021224e-01 -8.92800212e-01 3.17664705e-02
-5.36574543e-01 -1.26851983e-02 5.98745108e-01 1.02693474e+00
5.89921772e-01 1.65017113e-01 1.35835540e+00 -1.86257303e+00
7.68131852e-01 -1.05184305e+00 -1.80172935e-01 2.84536034e-01
-8.86134982e-01 -4.57883589e-02 8.17027926e-01 -9.24017251e-01
-1.05015194e+00 -8.02273899e-02 7.02384710e-02 -1.81065053e-01
-2.28204906e-01 6.96594894e-01 -1.39497399e-01 6.20564342e-01
7.44223595e-01 -5.16669415e-02 -1.70278445e-01 -5.71912110e-01
5.18401384e-01 5.75072348e-01 2.16609240e-01 -5.16622782e-01
3.68278682e-01 2.04032168e-01 7.66088739e-02 -2.56289363e-01
-9.96186078e-01 -4.34734434e-01 -3.28820199e-02 7.64002139e-03
8.33896756e-01 -8.63895118e-01 -1.32141590e+00 2.55977392e-01
-1.00563359e+00 -5.02203286e-01 -1.13575824e-01 5.43224096e-01
-4.75650191e-01 -2.41694987e-01 -6.47930205e-01 -5.06756485e-01
-3.50516111e-01 -9.61932719e-01 8.97348225e-01 5.66359796e-02
-3.92252266e-01 -1.26713657e+00 3.82443547e-01 2.04652399e-01
-2.41245572e-02 4.66131479e-01 1.57570827e+00 -8.69284332e-01
-4.41793501e-01 -3.88270438e-01 1.74445868e-01 -3.08105767e-01
5.00299573e-01 -2.02283591e-01 -7.78254509e-01 -1.26466617e-01
-3.26360106e-01 2.73380782e-02 8.72486532e-01 9.59997892e-01
1.25440156e+00 -7.30306268e-01 -9.90415215e-01 1.92579508e-01
1.09383881e+00 5.36573708e-01 3.47440153e-01 -3.27633947e-01
5.20698845e-01 3.38954061e-01 4.81276989e-01 8.28407466e-01
2.76494563e-01 6.41419768e-01 2.94324726e-01 -8.03353563e-02
2.02399746e-01 -6.22848988e-01 2.19582498e-01 -9.21677202e-02
4.67434675e-02 -4.17911261e-01 -6.51628971e-01 4.03765976e-01
-2.00889683e+00 -1.23683012e+00 -1.14960119e-01 2.45345330e+00
1.35023832e+00 4.32147235e-02 3.58259767e-01 -5.58913171e-01
7.84016848e-01 -8.92577991e-02 -1.30328667e+00 -2.42797583e-01
1.25213014e-02 4.31658387e-01 6.01366282e-01 3.04350227e-01
-1.18149126e+00 7.38683999e-01 6.22634220e+00 8.19934964e-01
-8.54748130e-01 3.05880100e-01 1.08158946e+00 -6.68688491e-02
-5.30833781e-01 5.46440333e-02 -8.17798257e-01 7.83613265e-01
1.09746480e+00 -8.07492435e-01 3.77408206e-01 4.15700316e-01
5.59280694e-01 1.04399577e-01 -1.56989050e+00 4.21430796e-01
-2.32043028e-01 -1.85482836e+00 5.51793240e-02 2.90591121e-01
1.09459341e+00 -8.01854953e-02 2.13826939e-01 4.87901449e-01
9.07279909e-01 -1.06682324e+00 -4.89374287e-02 8.27854455e-01
8.43107283e-01 -6.13830328e-01 3.21016401e-01 6.00700378e-01
-6.35443389e-01 -4.84936297e-01 -1.29095480e-01 -1.91767782e-01
2.93809958e-02 9.25549746e-01 -1.22899652e+00 3.88242126e-01
2.51897395e-01 1.09492505e+00 -1.75640255e-01 8.17936659e-01
-3.43659669e-02 9.46216762e-01 2.35981405e-01 -1.20263353e-01
-2.08184227e-01 1.61920175e-01 3.22620839e-01 8.51023912e-01
1.75240770e-01 5.72747290e-01 1.88640833e-01 6.93161547e-01
-5.47218263e-01 3.00488591e-01 -8.55200350e-01 -4.19376373e-01
3.80968541e-01 8.04630160e-01 2.14665104e-02 -4.75865632e-01
-4.47275698e-01 6.78666711e-01 1.32308334e-01 4.49801356e-01
-9.20888484e-01 2.66929328e-01 8.68536174e-01 3.41298699e-01
-1.26372322e-01 6.89581454e-01 -2.95731664e-01 -7.92376697e-01
-6.79914653e-01 -8.99484515e-01 7.47672856e-01 -5.68800926e-01
-1.75398314e+00 -1.06096581e-01 -1.24050528e-01 -1.02720237e+00
-2.92383462e-01 -1.97441399e-01 -7.65791416e-01 1.00332952e+00
-8.97286057e-01 -1.03976381e+00 4.10221130e-01 4.33804184e-01
4.66581225e-01 -1.49706751e-01 9.26002145e-01 2.97432929e-01
-8.46081734e-01 5.18433452e-01 9.75351110e-02 -1.34235933e-01
1.01205468e+00 -1.06171811e+00 1.98199034e-01 3.10715795e-01
-3.75972182e-01 6.64997518e-01 7.70228207e-01 -1.21830344e+00
-1.44488311e+00 -1.36428034e+00 9.62131858e-01 -6.93769813e-01
8.45807612e-01 -3.93111594e-02 -8.15998614e-01 9.43746269e-01
-7.87732229e-02 -3.45998108e-01 1.10802150e+00 6.42999828e-01
-3.04424673e-01 -1.05542727e-01 -1.04939878e+00 8.43277395e-01
1.14383447e+00 -3.58964324e-01 -3.35272342e-01 9.77588534e-01
1.16707790e+00 -6.79431707e-02 -1.12229109e+00 5.11651874e-01
5.68883717e-01 -5.61265707e-01 1.16520655e+00 -1.62093508e+00
9.92105603e-01 2.60798395e-01 2.63136566e-01 -1.52012110e+00
-6.50767982e-01 -6.30901575e-01 -1.45191640e-01 9.40926135e-01
6.65979147e-01 -8.17530870e-01 4.90084141e-01 1.03782034e+00
7.33215362e-02 -9.30506527e-01 -8.15830231e-01 -2.30768979e-01
1.48997813e-01 -2.21431136e-01 8.27007532e-01 1.21520793e+00
2.90596128e-01 7.60543227e-01 -8.34377170e-01 3.26274991e-01
6.38060629e-01 2.94222653e-01 7.96979487e-01 -1.45479929e+00
-7.32770979e-01 -4.25311446e-01 -4.50187139e-02 -5.66251934e-01
3.89656752e-01 -8.76585245e-01 -2.55155057e-01 -1.48004532e+00
8.83405864e-01 -5.17350793e-01 -5.03506064e-01 6.75077736e-01
-8.36472988e-01 -3.97633761e-01 -3.39362293e-01 5.44895530e-02
-2.39869416e-01 3.77470851e-01 1.19970918e+00 -3.31665426e-01
-4.26929742e-01 4.47900236e-01 -1.02140224e+00 5.62394977e-01
6.74817622e-01 -6.87346816e-01 -4.79998678e-01 1.56561375e-01
1.98206410e-01 6.71100378e-01 3.46805662e-01 -2.88105041e-01
2.63781041e-01 -7.19389498e-01 2.36514449e-01 -1.61167055e-01
-5.30347452e-02 -3.50093156e-01 5.78014076e-01 7.33789146e-01
-7.57892609e-01 -2.39936858e-01 -1.92723408e-01 1.23684657e+00
4.51178402e-01 1.95939496e-01 4.91844893e-01 -1.18294150e-01
-3.51092398e-01 8.82046461e-01 -1.62645504e-01 3.49369757e-02
9.55637157e-01 4.86249417e-01 -5.91896355e-01 -3.18052232e-01
-8.73042405e-01 2.80441791e-01 6.29294142e-02 4.12628412e-01
4.30823237e-01 -1.10417306e+00 -8.16614330e-01 -1.86404079e-01
2.43777305e-01 -3.78963798e-01 6.16795599e-01 9.30605292e-01
5.32704651e-01 3.93983006e-01 2.81976938e-01 -1.69907928e-01
-1.05281150e+00 1.12700009e+00 1.54082403e-01 -4.94255096e-01
-5.58025181e-01 5.96768320e-01 7.49027431e-01 -1.14428334e-01
4.84557375e-02 -4.74763997e-02 -2.07456976e-01 -1.42732151e-02
6.02710366e-01 1.66277885e-01 -3.53446066e-01 -1.03367507e-01
-1.32895216e-01 9.50996764e-03 -3.50267798e-01 2.13594168e-01
1.55863810e+00 2.60917544e-01 -6.19015694e-02 7.72642672e-01
9.92624044e-01 -1.34725943e-01 -1.05929911e+00 -5.67940295e-01
-1.09689564e-01 -3.28668386e-01 -8.75331834e-02 -1.17588067e+00
-6.85819149e-01 5.14139056e-01 4.50033724e-01 -2.09255159e-01
1.06051266e+00 2.30904192e-01 6.98605657e-01 -3.10845152e-02
4.21788275e-01 -9.01010633e-01 5.00625931e-02 -1.62212953e-01
6.30567193e-01 -1.16413915e+00 1.02859095e-01 -2.06312478e-01
-4.26143289e-01 4.15095806e-01 1.65673286e-01 4.24569100e-02
8.38416994e-01 -1.35692626e-01 -6.75059021e-01 -9.26046148e-02
-1.36192667e+00 -1.28399536e-01 2.54142642e-01 4.04397696e-01
4.12526369e-01 7.00618625e-01 -5.52742004e-01 1.13294554e+00
4.25706893e-01 4.35361624e-01 5.05329192e-01 5.15098691e-01
-2.77202845e-01 -1.28671110e+00 -9.04313177e-02 1.30976260e+00
-4.80824828e-01 -4.19322193e-01 -3.45372826e-01 3.94873202e-01
3.37389976e-01 9.27857101e-01 -3.11709762e-01 -4.97262478e-01
4.72202122e-01 1.66746527e-01 1.57934621e-01 -8.29005897e-01
-6.37461841e-01 3.97940576e-02 4.95155901e-02 -5.13430834e-01
-2.65689671e-01 -9.13978994e-01 -1.09053874e+00 -6.41714752e-01
1.88745677e-01 -1.76596969e-01 8.83960128e-02 1.07473242e+00
6.75263703e-01 7.93876231e-01 1.02368724e+00 -1.75132081e-01
-6.66610241e-01 -7.88696051e-01 -6.01193070e-01 7.19070673e-01
2.76661307e-01 -7.11244702e-01 -2.17233002e-01 2.40479857e-01] | [8.0846586227417, 5.495830535888672] |
ed3c58f6-a40e-4480-9571-7c35b4081ef6 | music-generation-using-an-lstm | 2203.12105 | null | https://arxiv.org/abs/2203.12105v1 | https://arxiv.org/pdf/2203.12105v1.pdf | Music Generation Using an LSTM | Over the past several years, deep learning for sequence modeling has grown in popularity. To achieve this goal, LSTM network structures have proven to be very useful for making predictions for the next output in a series. For instance, a smartphone predicting the next word of a text message could use an LSTM. We sought to demonstrate an approach of music generation using Recurrent Neural Networks (RNN). More specifically, a Long Short-Term Memory (LSTM) neural network. Generating music is a notoriously complicated task, whether handmade or generated, as there are a myriad of components involved. Taking this into account, we provide a brief synopsis of the intuition, theory, and application of LSTMs in music generation, develop and present the network we found to best achieve this goal, identify and address issues and challenges faced, and include potential future improvements for our network. | ['Alexander Neuwirth', 'Reagan Strelow', 'David Hunger', 'Kevin Adams', 'Lucas Gral', 'Michael Conner'] | 2022-03-23 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 4.83314842e-01 2.31111851e-02 4.36621755e-02 -6.64828625e-03
-7.68233716e-01 -4.64509159e-01 4.89230216e-01 -3.43528211e-01
-5.83964884e-02 7.34651268e-01 7.09758103e-01 -3.53494465e-01
1.17426679e-01 -7.31640279e-01 -7.22608805e-01 -3.54478836e-01
-2.18256205e-01 7.67846033e-02 -4.88899648e-01 -3.42521191e-01
5.28959572e-01 3.45215052e-01 -1.49541342e+00 7.55554557e-01
3.77736568e-01 1.03353477e+00 4.07421380e-01 8.48525465e-01
-3.22511315e-01 1.46877646e+00 -1.01908290e+00 -3.56231660e-01
-9.53366086e-02 -7.67353475e-01 -9.71962810e-01 -2.69165218e-01
-2.16264445e-02 -2.86607832e-01 -2.74776161e-01 6.38092875e-01
6.87812924e-01 2.43870884e-01 4.17301297e-01 -1.05168033e+00
-5.90564430e-01 1.34747994e+00 -1.61506273e-02 6.08549118e-02
5.42854548e-01 1.03983134e-01 1.04911518e+00 -8.71355891e-01
4.68212247e-01 9.63381886e-01 9.00256336e-01 7.62356400e-01
-9.95355666e-01 -9.42273259e-01 -8.56838524e-02 8.37639049e-02
-1.05558002e+00 -5.86718261e-01 7.62131572e-01 -2.94828922e-01
1.34093237e+00 1.56702578e-01 8.50925446e-01 1.64222431e+00
4.60830063e-01 1.01620722e+00 7.70739257e-01 -5.43539286e-01
-3.24699245e-02 -2.38047808e-01 -3.80304217e-01 2.75480449e-01
-3.40564042e-01 1.99374706e-02 -8.41204643e-01 -2.09285572e-01
9.49588358e-01 -1.98566288e-01 9.97141302e-02 5.08950531e-01
-1.32940221e+00 7.07301795e-01 3.23517025e-01 6.01609647e-01
-5.74759841e-01 6.28753901e-01 4.98744249e-01 3.28493536e-01
3.73258561e-01 7.95991004e-01 -2.49887481e-01 -7.27436006e-01
-1.28458118e+00 4.02443618e-01 8.38295519e-01 6.66364014e-01
1.21059798e-01 8.07912529e-01 -2.76432753e-01 9.23365712e-01
-4.55175759e-03 9.62628797e-02 9.35986519e-01 -8.89631152e-01
4.14491981e-01 1.31308809e-02 -2.02974193e-02 -9.53677356e-01
-2.19872624e-01 -6.31423175e-01 -9.37649012e-01 4.25591953e-02
1.15261473e-01 -5.16257703e-01 -6.95649862e-01 1.68119454e+00
-4.76265460e-01 4.98633325e-01 1.38399312e-02 5.24465084e-01
6.37011707e-01 8.99317920e-01 9.55488384e-02 -1.64799079e-01
9.01426852e-01 -9.23868656e-01 -7.35445678e-01 -2.20192820e-01
4.89091486e-01 -1.09718955e+00 8.70490372e-01 4.63993669e-01
-1.32520020e+00 -6.98036253e-01 -1.00808060e+00 2.25829184e-02
-2.37170517e-01 3.23017240e-02 6.69543207e-01 4.88416135e-01
-1.05502558e+00 1.13829696e+00 -6.78828537e-01 -2.41000578e-01
2.04985607e-02 4.01156396e-01 3.39110643e-01 6.17743850e-01
-1.36286592e+00 7.58934438e-01 4.31753576e-01 2.71175086e-01
-6.62680209e-01 -4.32424068e-01 -2.85839349e-01 1.85084656e-01
-7.40442947e-02 -8.50959957e-01 1.79000103e+00 -1.30241477e+00
-1.82126606e+00 4.08935279e-01 -3.03259522e-01 -1.00793350e+00
1.81243762e-01 -2.68026680e-01 -6.21680677e-01 -3.18514198e-01
-2.20130384e-01 6.97163463e-01 7.49195695e-01 -7.38498628e-01
-4.64443535e-01 9.20690745e-02 -2.16997892e-01 1.95092365e-01
-3.41075391e-01 2.78140694e-01 2.17347696e-01 -1.27570057e+00
-2.30518505e-01 -9.95638132e-01 -2.92466283e-01 -6.59085095e-01
-5.33812106e-01 -2.66080856e-01 5.34600437e-01 -7.30625927e-01
1.59983432e+00 -1.74075079e+00 -1.70258701e-01 5.08071706e-02
-1.29208729e-01 2.45592833e-01 -1.96159273e-01 8.60053003e-01
-2.28853270e-01 3.71178776e-01 1.32652238e-01 -2.74054468e-01
-4.68492284e-02 -1.33869261e-01 -1.10795569e+00 -5.23494303e-01
3.87781560e-02 1.19871151e+00 -7.76165545e-01 -3.10220160e-02
1.94250681e-02 5.98714232e-01 -1.96723804e-01 1.67385757e-01
-6.68676853e-01 3.98574352e-01 -2.50499189e-01 2.92816073e-01
-2.24979863e-01 -3.60492617e-01 8.70470330e-02 1.32691175e-01
-2.47370422e-01 8.98233712e-01 -8.88471365e-01 1.73943579e+00
-6.43716633e-01 1.00406623e+00 -5.55928171e-01 -6.09770358e-01
1.13754761e+00 7.61167467e-01 5.17290354e-01 -5.90855658e-01
5.25902063e-02 4.12446201e-01 7.21222311e-02 -4.12194729e-01
9.09084678e-01 -4.65166897e-01 1.66794024e-02 9.20042992e-01
-1.53563514e-01 -8.17717835e-02 -8.93896967e-02 -4.38409392e-03
1.01033199e+00 2.61395067e-01 2.56086171e-01 4.15948987e-01
5.87717481e-02 -6.81352690e-02 2.42598176e-01 8.81538630e-01
2.29174852e-01 7.20194578e-01 1.51181057e-01 -5.10481000e-01
-1.13551581e+00 -6.23840690e-01 5.30982614e-01 1.09086466e+00
-7.17815995e-01 -5.03990829e-01 -7.42617011e-01 6.50041364e-03
-4.00931686e-01 1.16580415e+00 -3.24487239e-01 -9.87998322e-02
-6.73078775e-01 -4.16931659e-01 1.13191116e+00 5.75703859e-01
1.44300625e-01 -1.93379521e+00 -5.52082479e-01 7.71344006e-01
-5.98466218e-01 -7.81993091e-01 -4.78313178e-01 2.25504655e-02
-1.13413525e+00 -3.82819116e-01 -8.85759115e-01 -8.52416813e-01
-1.06323220e-01 7.13424981e-02 1.25174356e+00 -1.66022062e-01
-1.21909045e-02 2.16436192e-01 -3.50546628e-01 -8.35321963e-01
-7.22724736e-01 3.16784471e-01 -1.23095363e-02 -1.75239131e-01
2.99878180e-01 -8.91622424e-01 -5.44153571e-01 -2.09066555e-01
-8.91592562e-01 3.75655711e-01 5.95964551e-01 6.20583057e-01
2.06867576e-01 -1.43633336e-01 9.83330429e-01 -6.12257063e-01
1.31138182e+00 -4.27886963e-01 4.51829918e-02 1.12955436e-01
-4.88507420e-01 -1.44151570e-02 7.24622071e-01 -4.81646478e-01
-8.31657410e-01 6.03846982e-02 -5.20347953e-01 -3.18385184e-01
1.11582398e-01 8.73123705e-01 6.30844116e-01 2.83106357e-01
7.37387061e-01 5.01213133e-01 -1.22075602e-01 -4.16267484e-01
2.64299154e-01 8.37059557e-01 3.31879556e-01 -3.86258006e-01
3.71534884e-01 6.23886473e-02 -2.27892011e-01 -8.39141309e-01
-8.44286919e-01 1.16290145e-01 -3.96239042e-01 -4.64367092e-01
5.03990233e-01 -7.98632503e-01 -1.04037154e+00 4.21148092e-01
-1.38687968e+00 -6.13219023e-01 -5.36813796e-01 3.05519283e-01
-9.06087339e-01 -5.42900190e-02 -9.87877905e-01 -1.07800126e+00
-8.87541175e-01 -7.73962259e-01 7.16922939e-01 7.66133741e-02
-1.03587210e+00 -1.10794699e+00 1.28306001e-01 1.56612709e-01
6.99291825e-01 3.15506220e-01 9.85406160e-01 -4.52701569e-01
-5.04210174e-01 -9.17176530e-02 1.33471996e-01 2.47271731e-01
5.08075804e-02 -1.10308329e-05 -1.13888502e+00 4.48317118e-02
1.02577388e-01 -4.35707003e-01 7.37825334e-01 5.38521469e-01
1.37912428e+00 -5.31398535e-01 2.63948608e-02 2.39471525e-01
9.97048378e-01 4.63391781e-01 8.26476038e-01 2.83573866e-01
5.95824003e-01 4.15299296e-01 2.86425620e-01 6.37307465e-01
3.23831201e-01 5.89379191e-01 1.08298473e-01 3.92919444e-02
-2.63133556e-01 -8.17737758e-01 7.72681832e-01 1.39714563e+00
-2.46086672e-01 -3.31983536e-01 -7.17114449e-01 3.60210210e-01
-1.71487105e+00 -1.44197214e+00 -3.91678475e-02 2.08909035e+00
8.17443669e-01 2.38691330e-01 3.62450272e-01 3.54997069e-01
4.95437890e-01 3.09959620e-01 -5.17489135e-01 -7.05981135e-01
-5.69027513e-02 6.23835981e-01 4.28621173e-02 1.68550521e-01
-5.77748835e-01 1.12721753e+00 7.78504896e+00 8.85618567e-01
-1.54041982e+00 -2.15947554e-01 6.38660192e-01 -1.83491021e-01
-4.68442827e-01 -1.43545419e-01 -5.36028028e-01 4.88529384e-01
1.53220582e+00 -3.57914239e-01 6.95301414e-01 5.42326748e-01
7.42311239e-01 3.20988655e-01 -9.56283212e-01 1.02684319e+00
8.00631344e-02 -1.75727570e+00 4.66732204e-01 -4.64617424e-02
7.44738579e-01 8.37275609e-02 3.63402486e-01 3.36326092e-01
3.10443610e-01 -1.41003788e+00 8.57706964e-01 7.32746720e-01
7.46391952e-01 -8.50136578e-01 2.72565931e-01 2.85537511e-01
-1.15435791e+00 -2.23728985e-01 -1.71428099e-01 -7.37250447e-01
3.99316579e-01 6.19417369e-01 -1.08688164e+00 1.44567832e-01
2.80901402e-01 7.86331236e-01 -4.46396284e-02 9.62965906e-01
-7.58664086e-02 8.70826185e-01 1.00236647e-01 -3.38789493e-01
2.65935510e-01 -7.47009367e-02 4.29569244e-01 1.22069097e+00
9.02140021e-01 3.62180248e-02 -1.07836440e-01 1.02379620e+00
-2.13182509e-01 5.92040680e-02 -7.71906555e-01 -7.19448507e-01
3.89368594e-01 9.81250346e-01 -5.83314359e-01 -3.40916634e-01
4.98641208e-02 1.12255132e+00 7.94787779e-02 4.88539457e-01
-7.11505115e-01 -4.46025670e-01 5.57938516e-01 3.21706422e-02
4.05937843e-02 -3.53383929e-01 -6.13704622e-01 -9.97647524e-01
-2.23170295e-01 -1.12832749e+00 1.35381550e-01 -1.29110444e+00
-1.01463354e+00 7.51593530e-01 -6.51529253e-01 -1.42679310e+00
-1.01800609e+00 -2.45147973e-01 -8.57115269e-01 9.15128946e-01
-1.00347817e+00 -1.09448349e+00 6.79924265e-02 2.83146113e-01
9.17127132e-01 -2.28335142e-01 1.26536345e+00 1.66533872e-01
-2.80513585e-01 4.13008541e-01 5.39152361e-02 9.39678326e-02
4.69362348e-01 -8.09337795e-01 1.01004088e+00 4.50873047e-01
6.63032591e-01 7.21715629e-01 8.00752282e-01 -7.13106394e-01
-1.24977589e+00 -1.05707753e+00 1.29650760e+00 -1.84263319e-01
7.63862431e-01 -2.77403831e-01 -4.78230894e-01 1.03714001e+00
3.57962191e-01 -1.02226853e+00 8.40961039e-01 1.15869530e-01
-8.15472081e-02 8.57894123e-02 -5.40676713e-01 9.90393221e-01
9.99375939e-01 -7.57931113e-01 -4.07699853e-01 2.36987054e-01
7.34369755e-01 -4.86237168e-01 -7.45249093e-01 2.06175610e-01
1.00982857e+00 -1.06668937e+00 8.24877441e-01 -4.38924760e-01
8.39613914e-01 9.07239094e-02 1.44589394e-01 -1.42325711e+00
-2.01492712e-01 -1.18593574e+00 -2.68794894e-01 1.15493464e+00
5.91929376e-01 -2.99934685e-01 9.44139779e-01 4.69067842e-01
-2.27478713e-01 -7.48549163e-01 -4.46037084e-01 -5.83446801e-01
-1.54613107e-02 -9.66829598e-01 5.16832530e-01 7.16915309e-01
3.20712745e-01 5.19340754e-01 -9.98967171e-01 -5.17232835e-01
1.31035194e-01 1.68044791e-01 6.71673298e-01 -1.03535986e+00
-4.07360375e-01 -6.71675444e-01 -3.96363959e-02 -1.30151260e+00
1.43438593e-01 -9.58520532e-01 -9.83645208e-03 -1.55267441e+00
-4.30573616e-03 -2.61632085e-01 -4.24523711e-01 3.63519460e-01
2.22774372e-01 4.26959723e-01 5.28502762e-01 2.53726244e-01
-1.57639667e-01 3.47075522e-01 1.07891476e+00 -8.07134211e-02
-4.75072324e-01 4.39648867e-01 -8.51875544e-01 6.25231743e-01
9.85304713e-01 -3.77729714e-01 -3.08261067e-01 -3.74853522e-01
6.76465094e-01 4.99323100e-01 1.08686298e-01 -1.20284510e+00
3.68643284e-01 8.35858844e-03 5.87526083e-01 -6.77950203e-01
7.56557703e-01 -3.38493705e-01 4.91451949e-01 5.33998132e-01
-9.07322943e-01 1.60898417e-01 2.55103856e-01 2.69173533e-01
-3.42119128e-01 -2.58191645e-01 2.55279958e-01 -2.68519938e-01
-5.24955809e-01 5.51088015e-03 -8.82924497e-01 -2.63843358e-01
5.13917565e-01 -3.44183177e-01 2.17718825e-01 -1.16093147e+00
-8.07761610e-01 -2.04618722e-01 -2.30443683e-02 7.11293876e-01
8.39206696e-01 -1.44198215e+00 -4.60395902e-01 1.44182071e-02
-2.87213057e-01 -5.38855672e-01 1.90748081e-01 4.26441103e-01
-2.69562215e-01 5.69250762e-01 -2.22641259e-01 -2.33681258e-02
-1.18897271e+00 2.34234825e-01 2.53087997e-01 -3.63639921e-01
-5.24304509e-01 7.68743396e-01 -4.65643913e-01 -1.42893061e-01
3.81630361e-01 -3.97665054e-01 -3.49641591e-01 4.46253344e-02
7.08103776e-01 4.98174369e-01 -1.06759496e-01 -3.48939747e-01
2.99064457e-01 3.36488903e-01 8.38255882e-02 -4.70251977e-01
1.34455943e+00 -2.39447430e-02 -1.74172595e-01 1.13222837e+00
6.77405655e-01 -1.81310266e-01 -7.82598734e-01 -5.53951412e-02
2.64756739e-01 1.00614302e-01 -2.73594499e-01 -8.87397230e-01
-7.55036592e-01 1.03299582e+00 2.61378706e-01 5.19088626e-01
9.70838845e-01 -5.41093588e-01 1.56350517e+00 3.77846837e-01
3.35023254e-01 -1.11387515e+00 3.23521376e-01 9.14628267e-01
9.57415998e-01 -5.97845733e-01 -3.69014114e-01 2.83661783e-01
-6.60808206e-01 1.60812700e+00 1.30356789e-01 -2.05641761e-01
3.73010159e-01 1.91512525e-01 1.19468883e-01 8.63369927e-02
-1.02627170e+00 1.24580741e-01 1.85665339e-01 2.63656497e-01
1.10660541e+00 1.29353717e-01 3.24865058e-02 3.74383122e-01
-8.58849168e-01 5.35117269e-01 6.52162492e-01 4.76147383e-01
-3.92676204e-01 -1.23094451e+00 -3.84891927e-01 6.82140946e-01
-7.15290010e-01 -3.60708773e-01 -5.69480479e-01 1.15388922e-01
-1.23073407e-01 1.16683221e+00 3.38346139e-02 -9.06250358e-01
4.98347059e-02 3.82981449e-01 2.56752938e-01 -6.21050954e-01
-1.01190412e+00 1.04171030e-01 2.04745680e-01 -2.20060512e-01
-2.71387190e-01 -3.22621375e-01 -1.23835742e+00 -5.20703614e-01
4.11162451e-02 3.37135382e-02 7.65789628e-01 9.51902270e-01
5.64456582e-01 7.92789996e-01 5.99032700e-01 -1.17040861e+00
-4.33492362e-01 -1.16298592e+00 -5.26724935e-01 9.26032364e-02
1.32399872e-01 2.14168966e-01 3.30792405e-02 2.05378339e-01] | [15.926700592041016, 5.510984420776367] |
09f6c39e-2912-4793-991e-13e52af2cf00 | u-dudonet-unpaired-dual-domain-network-for-ct | 2103.04552 | null | https://arxiv.org/abs/2103.04552v1 | https://arxiv.org/pdf/2103.04552v1.pdf | U-DuDoNet: Unpaired dual-domain network for CT metal artifact reduction | Recently, both supervised and unsupervised deep learning methods have been widely applied on the CT metal artifact reduction (MAR) task. Supervised methods such as Dual Domain Network (Du-DoNet) work well on simulation data; however, their performance on clinical data is limited due to domain gap. Unsupervised methods are more generalized, but do not eliminate artifacts completely through the sole processing on the image domain. To combine the advantages of both MAR methods, we propose an unpaired dual-domain network (U-DuDoNet) trained using unpaired data. Unlike the artifact disentanglement network (ADN) that utilizes multiple encoders and decoders for disentangling content from artifact, our U-DuDoNet directly models the artifact generation process through additions in both sinogram and image domains, which is theoretically justified by an additive property associated with metal artifact. Our design includes a self-learned sinogram prior net, which provides guidance for restoring the information in the sinogram domain, and cyclic constraints for artifact reduction and addition on unpaired data. Extensive experiments on simulation data and clinical images demonstrate that our novel framework outperforms the state-of-the-art unpaired approaches. | ['S. Kevin Zhou', 'Cheng Peng', 'Jiajun Fu', 'Yuanyuan Lyu'] | 2021-03-08 | null | null | null | null | ['metal-artifact-reduction'] | ['medical'] | [ 2.95916677e-01 -1.41058583e-04 3.93756963e-02 -3.17055553e-01
-9.18988287e-01 -8.69459137e-02 1.68491676e-01 -1.62738204e-01
-1.86395556e-01 8.65790069e-01 6.04982138e-01 -8.32642689e-02
-3.13445091e-01 -5.33706844e-01 -8.16614151e-01 -7.67914712e-01
3.28484103e-02 1.23682640e-01 7.22556934e-02 1.18421465e-01
-8.46647024e-02 -8.36761203e-03 -3.66063505e-01 2.25803062e-01
1.31062973e+00 7.84791946e-01 2.11129636e-01 1.73778951e-01
3.23959082e-01 1.23863840e+00 -3.63028914e-01 8.84967223e-02
3.60506117e-01 -9.20427322e-01 -5.29637694e-01 2.36227152e-05
7.08825290e-02 -8.30634713e-01 -1.03861988e+00 1.07120943e+00
1.02189696e+00 -1.68751910e-01 7.04626858e-01 -9.22946453e-01
-8.04884315e-01 8.16118181e-01 -6.33337796e-01 2.43221849e-01
6.91937059e-02 3.84076297e-01 2.46432364e-01 -9.17667270e-01
4.91004527e-01 5.11971831e-01 9.97197390e-01 5.19116879e-01
-1.15093029e+00 -7.51329660e-01 -2.18977362e-01 8.34475830e-02
-1.25462520e+00 -1.07788108e-01 9.64899778e-01 -5.16986668e-01
6.99521303e-01 -7.02672303e-02 5.80934465e-01 1.44283521e+00
4.92680013e-01 1.05365217e+00 1.11339760e+00 -1.84971064e-01
3.05035830e-01 -2.36584902e-01 -1.27066791e-01 5.01036823e-01
3.18261325e-01 4.17632282e-01 -6.48344100e-01 -1.19002096e-01
1.26757097e+00 1.96615919e-01 -6.97377086e-01 -4.80461091e-01
-1.24738157e+00 5.09354055e-01 4.62274879e-01 1.80125594e-01
-4.29374337e-01 1.92772299e-01 5.93566775e-01 9.51007158e-02
3.98846149e-01 3.97245377e-01 -2.28245735e-01 2.82851774e-02
-1.11787069e+00 1.65429816e-01 3.15615714e-01 1.25378501e+00
1.86860070e-01 4.46538597e-01 -5.81184447e-01 8.31396937e-01
2.58605152e-01 4.75436956e-01 8.62069011e-01 -5.99966466e-01
4.59915787e-01 2.46208057e-01 7.87730142e-03 -5.75029850e-01
-5.39991021e-01 -7.78958023e-01 -1.23579931e+00 1.63073882e-01
2.81424552e-01 -2.90842712e-01 -1.31500864e+00 1.77586329e+00
-1.95690364e-01 5.85441411e-01 -3.84384133e-02 1.29885876e+00
8.44736755e-01 1.72573879e-01 -4.17148769e-02 -2.74024010e-01
1.08524311e+00 -1.01321995e+00 -1.18549287e+00 -2.70578623e-01
5.78433394e-01 -6.38574481e-01 7.51052082e-01 5.68644285e-01
-1.31511927e+00 -3.88832569e-01 -1.35693300e+00 -1.98133171e-01
1.06290996e-01 8.74525383e-02 4.45433617e-01 4.62936521e-01
-6.75659060e-01 7.82885075e-01 -1.25372434e+00 2.43887529e-01
7.56625056e-01 3.18239719e-01 -5.27570024e-02 -2.39508048e-01
-1.43550098e+00 1.11425936e+00 -7.10376799e-02 2.97245175e-01
-1.39453495e+00 -1.08330309e+00 -1.08475482e+00 -9.46560130e-02
3.56798351e-01 -9.25378740e-01 1.33816111e+00 -7.16058731e-01
-1.50807297e+00 3.04629445e-01 9.33256671e-02 -6.41560972e-01
9.29905772e-01 -3.33247811e-01 -4.83433247e-01 2.58341342e-01
3.05532396e-01 1.84288830e-01 8.38648081e-01 -1.14811575e+00
1.46794707e-01 -2.00723991e-01 -5.07171452e-01 2.17542604e-01
-1.83886573e-01 -1.91217825e-01 -3.31426144e-01 -1.13874459e+00
3.60053658e-01 -6.25495315e-01 -5.01246035e-01 3.38540733e-01
-4.74652857e-01 4.78752226e-01 6.75685406e-01 -1.04875290e+00
1.23060393e+00 -2.21592164e+00 -7.66314939e-02 2.64002025e-01
6.71626806e-01 1.68880448e-01 3.70020457e-02 1.60237953e-01
-6.03815317e-01 -1.61742255e-01 -9.42109168e-01 -3.45718056e-01
-2.59259403e-01 1.03708975e-01 1.87079813e-02 7.12169230e-01
1.68618992e-01 8.86480391e-01 -1.22422910e+00 -5.21103144e-01
3.07576865e-01 3.81868601e-01 -5.91920376e-01 2.81733483e-01
1.85993671e-01 9.36881065e-01 -4.86442268e-01 2.67892271e-01
1.04733765e+00 -4.60653722e-01 2.63164073e-01 -2.71001756e-01
1.96132623e-02 4.63153154e-01 -8.92310023e-01 2.33792901e+00
-5.16768277e-01 4.69850481e-01 -1.80077016e-01 -7.91420281e-01
5.35892010e-01 7.32011139e-01 1.01998258e+00 -1.02049077e+00
1.87952518e-01 4.39315796e-01 2.20631629e-01 -7.62848794e-01
-2.08251178e-02 -4.96713638e-01 1.64030179e-01 4.45281327e-01
2.52149045e-01 -6.94969073e-02 -3.14384937e-01 3.68062973e-01
1.39295316e+00 2.61830509e-01 1.71621665e-01 -3.44335407e-01
1.58135951e-01 -1.48646191e-01 9.35868800e-01 9.09150600e-01
-4.85401571e-01 1.24935615e+00 3.25617373e-01 -2.41897687e-01
-1.08464277e+00 -1.46509111e+00 -4.28513020e-01 -1.17390506e-01
4.58024502e-01 -2.86410213e-01 -6.86724544e-01 -8.09035122e-01
-2.50767440e-01 5.58130860e-01 -4.88103181e-01 -4.33932990e-01
-7.15846956e-01 -9.77915525e-01 6.10552013e-01 1.07485318e+00
7.47081220e-01 -7.18979657e-01 -4.18187410e-01 4.72557694e-01
-4.68142092e-01 -1.15114665e+00 -7.90412307e-01 4.29550678e-01
-9.85772848e-01 -1.01992178e+00 -9.45876896e-01 -6.70693755e-01
8.24621975e-01 7.15296715e-03 8.35870087e-01 -1.32524341e-01
-5.29296160e-01 4.18392159e-02 -3.14163119e-01 -4.02388781e-01
-2.45508596e-01 -5.24885714e-01 2.65335944e-02 -8.76723155e-02
3.87889333e-02 -7.84685373e-01 -1.10315955e+00 2.35667527e-01
-1.07512498e+00 2.74311930e-01 6.71133816e-01 1.16814148e+00
4.96863246e-01 -1.16167791e-01 7.36311734e-01 -9.92806971e-01
4.55094904e-01 -5.56486011e-01 -1.92218795e-01 -8.23030174e-02
-6.19685769e-01 1.91045955e-01 6.93219781e-01 -2.78392464e-01
-1.36937249e+00 1.67183071e-01 -1.49821579e-01 -7.02831626e-01
2.03521177e-01 4.35636848e-01 -4.31434158e-03 -2.53822040e-02
8.35140884e-01 4.67780381e-01 1.38910457e-01 -4.35044706e-01
2.37450097e-02 4.32471097e-01 5.92779040e-01 -3.76825333e-01
6.77631855e-01 5.87910831e-01 -8.91055018e-02 -3.85160536e-01
-7.35808730e-01 -2.51892567e-01 -5.11239350e-01 -5.77916317e-02
1.00515807e+00 -1.01747870e+00 -1.72796190e-01 7.12623119e-01
-1.19933105e+00 -3.21351141e-01 -4.32806551e-01 8.75789881e-01
-6.52078569e-01 8.53686929e-01 -8.18695188e-01 -5.00799119e-01
-3.31338704e-01 -1.38418889e+00 7.53334105e-01 -4.35873047e-02
-1.07478216e-01 -9.12711143e-01 -4.61905152e-02 3.83991674e-02
5.57117522e-01 3.83164018e-01 9.19821084e-01 -4.44346011e-01
-5.14882684e-01 8.91664028e-02 -3.68814468e-01 6.48170054e-01
3.75740796e-01 -8.45209599e-01 -9.31091487e-01 -2.04112604e-01
7.24187195e-01 -2.09784403e-01 8.46200109e-01 8.93010497e-01
1.63469040e+00 -8.10396485e-03 -2.27702409e-01 8.99598181e-01
1.47619081e+00 2.97680020e-01 1.07539690e+00 1.21165030e-01
7.21877277e-01 -1.78538844e-01 2.17941061e-01 4.71167177e-01
1.72527567e-01 3.66100848e-01 4.16186512e-01 -5.32504201e-01
-6.31914914e-01 -2.75766164e-01 3.37143876e-02 8.17570865e-01
-6.49907440e-02 -9.62462351e-02 -8.35847497e-01 5.74437439e-01
-1.94612753e+00 -5.85772514e-01 -4.28007543e-01 2.15088367e+00
1.02055573e+00 7.22764358e-02 -4.32558745e-01 3.97020057e-02
5.93399644e-01 1.10982388e-01 -7.00017929e-01 1.79079100e-01
1.48580521e-02 4.33655590e-01 6.64419174e-01 3.41983974e-01
-9.33472335e-01 2.94422686e-01 6.67991686e+00 7.30177343e-01
-8.95165145e-01 4.14843231e-01 4.04816270e-01 -1.07527494e-01
-2.40554258e-01 -1.39246225e-01 2.54174937e-02 7.65745580e-01
5.70166290e-01 3.18803750e-02 1.75137281e-01 5.81019104e-01
5.61082423e-01 -8.64974931e-02 -1.36533344e+00 1.07338977e+00
1.27034083e-01 -1.30067790e+00 -3.08986574e-01 -2.24485233e-01
8.78306329e-01 4.73626703e-03 -3.32927406e-02 1.18315510e-01
1.11516312e-01 -8.53758872e-01 5.38893819e-01 6.06296778e-01
1.02575326e+00 -2.38792211e-01 8.82371306e-01 1.41980797e-01
-7.75190711e-01 1.87377855e-01 -2.12798528e-02 1.60590410e-01
3.69798839e-01 9.31878686e-01 -7.45350003e-01 1.02916551e+00
4.61788207e-01 1.00745225e+00 -2.45679736e-01 1.28927076e+00
-5.06271660e-01 6.38397872e-01 -2.06719577e-01 6.32817686e-01
2.37560589e-02 -1.03281982e-01 6.48104250e-01 1.04155838e+00
2.05407530e-01 2.13856444e-01 -3.64371762e-02 1.20155966e+00
4.92274165e-02 -5.34243464e-01 -4.88062561e-01 4.21820015e-01
2.48296291e-01 7.56837845e-01 -3.86962503e-01 -4.27748203e-01
-6.02380872e-01 1.21235824e+00 -2.65583396e-01 4.83281314e-01
-1.17759526e+00 -3.48202586e-01 2.42721304e-01 2.34060481e-01
-1.18651323e-01 -2.18969852e-01 -9.90951955e-01 -1.28494656e+00
1.91891685e-01 -7.41250277e-01 2.83054501e-01 -8.88441443e-01
-1.52711713e+00 5.27383864e-01 -1.26112968e-01 -2.03584337e+00
4.88792621e-02 -5.39162695e-01 -6.55353606e-01 1.00397646e+00
-1.73986816e+00 -9.49426413e-01 -4.05995697e-01 7.88426638e-01
5.00947416e-01 1.31631836e-01 5.56141973e-01 7.17175543e-01
-4.41438615e-01 6.16156399e-01 2.16069609e-01 4.82118011e-01
1.07994092e+00 -1.29915202e+00 7.38311931e-02 1.12386012e+00
-4.56830472e-01 5.87204158e-01 4.72112715e-01 -9.72884953e-01
-1.27975941e+00 -1.16990113e+00 3.18013072e-01 -3.70140433e-01
5.06058455e-01 -2.02958494e-01 -8.61513257e-01 7.53908694e-01
4.47764426e-01 4.79330003e-01 6.52998745e-01 -3.87661040e-01
-1.57971516e-01 1.03039429e-01 -1.13428903e+00 4.26964492e-01
1.11621857e+00 -4.17309463e-01 -6.63720548e-01 4.48131025e-01
5.89995682e-01 -9.34854090e-01 -8.14817965e-01 6.30130649e-01
2.30631113e-01 -8.61821711e-01 8.57261837e-01 -3.13217551e-01
1.10025513e+00 -3.16808105e-01 2.80235440e-01 -1.53122556e+00
-2.99594998e-01 -5.94655335e-01 -9.50564444e-02 6.31302595e-01
3.36862624e-01 -5.73019266e-01 6.65766180e-01 5.31213939e-01
-9.47038174e-01 -6.63606584e-01 -9.36533391e-01 -8.91904473e-01
1.76773563e-01 -5.83339691e-01 3.67030680e-01 1.13732207e+00
1.51975065e-01 -2.86763050e-02 -6.44761503e-01 5.10075986e-01
8.85203838e-01 -1.78283945e-01 8.81416947e-02 -4.88675475e-01
-5.03561854e-01 -1.06763847e-01 -1.53678820e-01 -1.14880264e+00
-4.81965274e-01 -9.66674030e-01 4.11353618e-01 -1.58427346e+00
4.09733713e-01 -3.62363547e-01 -4.92915213e-01 5.86223602e-01
-1.80744782e-01 1.50977761e-01 -1.09666020e-01 1.50821343e-01
-2.17851028e-01 8.18563998e-01 1.78294182e+00 -1.74088180e-01
-1.32208303e-01 -3.90050113e-01 -5.96963823e-01 8.48642826e-01
4.36404347e-01 -7.16311336e-01 -4.23374057e-01 -8.65024149e-01
-9.43709165e-02 5.24646521e-01 5.18664122e-01 -1.11413121e+00
4.35762107e-01 1.95751667e-01 7.35459864e-01 -5.45006216e-01
1.48485348e-01 -9.65022564e-01 -1.53543964e-01 5.47917008e-01
-2.16194317e-01 -3.62383127e-01 2.22142741e-01 6.07381165e-01
-3.66061151e-01 -9.00434926e-02 9.36548233e-01 -1.67549446e-01
-2.29065090e-01 5.60891747e-01 -4.10005093e-01 6.14560358e-02
6.70688212e-01 -2.80119270e-01 -2.54291981e-01 -3.42770934e-01
-1.04757035e+00 3.83390218e-01 -1.43261552e-01 1.61439523e-01
9.43769634e-01 -1.44278586e+00 -7.72127569e-01 4.49162960e-01
-1.31923715e-02 4.19824749e-01 8.05119574e-01 1.33841240e+00
-4.95334059e-01 4.48579788e-02 -1.78057492e-01 -6.20988309e-01
-4.63465452e-01 4.54737097e-01 6.09999120e-01 -5.58309495e-01
-1.10113466e+00 6.15822315e-01 5.52414536e-01 -1.21205896e-01
1.04223877e-01 -4.11521941e-01 3.72267574e-01 -5.66342711e-01
4.03872550e-01 3.01762819e-01 3.33662868e-01 -2.93995813e-02
-3.73618513e-01 2.49243423e-01 -2.70797342e-01 -6.13444075e-02
1.43808615e+00 -3.76596786e-02 8.70111361e-02 -1.50076523e-01
1.13850665e+00 -2.81960845e-01 -1.43911886e+00 -2.48011827e-01
-3.84665102e-01 -4.86248851e-01 2.87037820e-01 -1.28540921e+00
-1.26588058e+00 8.92180204e-01 7.76789606e-01 -3.06172997e-01
1.48095417e+00 -4.50739831e-01 9.68938828e-01 -2.88811684e-01
3.27254325e-01 -1.07865810e+00 4.02583241e-01 2.16310725e-01
9.91639733e-01 -1.23938072e+00 1.52819470e-01 -6.74284875e-01
-7.35997796e-01 1.01800263e+00 7.40786910e-01 -4.18063045e-01
7.14791596e-01 7.29088306e-01 1.33043109e-02 -2.83663124e-01
-1.15103245e-01 2.85131961e-01 6.87423050e-02 6.07000887e-01
3.80695581e-01 -1.05774090e-01 -6.97451770e-01 1.13269246e+00
2.26704001e-01 4.76250023e-01 8.01348627e-01 1.19470584e+00
1.72692910e-01 -1.04091966e+00 -3.64525676e-01 5.51930666e-01
-3.92150432e-01 -4.02890980e-01 1.15001433e-01 7.21629858e-01
1.94048911e-01 5.97352087e-01 -3.08703184e-01 -2.41006583e-01
3.42871308e-01 -2.28782222e-01 6.24293149e-01 -7.20984042e-01
-4.88189489e-01 4.33548957e-01 -1.04801066e-01 -6.28568709e-01
-1.96787253e-01 -5.19988298e-01 -1.48735535e+00 2.85399288e-01
-3.65062445e-01 -1.26333669e-01 3.44703704e-01 1.03108847e+00
2.47380540e-01 1.19805801e+00 6.60489619e-01 -6.29158556e-01
-6.90200925e-01 -1.02902567e+00 -8.45572889e-01 4.22953308e-01
5.60489237e-01 -6.20281816e-01 -2.96118230e-01 1.44868940e-01] | [13.526665687561035, -2.531423568725586] |
a0135a54-6968-49ff-abf9-0da5a0d7507c | adversarial-attacks-on-audio-source | 2010.03164 | null | https://arxiv.org/abs/2010.03164v3 | https://arxiv.org/pdf/2010.03164v3.pdf | Adversarial attacks on audio source separation | Despite the excellent performance of neural-network-based audio source separation methods and their wide range of applications, their robustness against intentional attacks has been largely neglected. In this work, we reformulate various adversarial attack methods for the audio source separation problem and intensively investigate them under different attack conditions and target models. We further propose a simple yet effective regularization method to obtain imperceptible adversarial noise while maximizing the impact on separation quality with low computational complexity. Experimental results show that it is possible to largely degrade the separation quality by adding imperceptibly small noise when the noise is crafted for the target model. We also show the robustness of source separation models against a black-box attack. This study provides potentially useful insights for developing content protection methods against the abuse of separated signals and improving the separation performance and robustness. | ['Yuki Mitsufuji', 'Shota Inoue', 'Naoya Takahashi'] | 2020-10-07 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [ 6.16658449e-01 -2.93130040e-01 2.43731381e-04 2.60552853e-01
-1.09145713e+00 -9.88866389e-01 2.39095598e-01 -1.12473689e-01
-9.95303094e-02 4.75427896e-01 1.84374571e-01 -3.85075867e-01
-3.31439227e-01 -6.17713809e-01 -5.27070045e-01 -1.07597148e+00
-4.36684132e-01 -5.03273487e-01 1.39610514e-01 -7.06804842e-02
-1.94370389e-01 7.12698936e-01 -1.20702720e+00 3.48042622e-02
9.18475151e-01 8.08824599e-01 -3.58985692e-01 9.62939739e-01
5.62759280e-01 6.70570731e-01 -1.27376592e+00 -5.32462537e-01
4.43062097e-01 -7.67372251e-01 -4.63205159e-01 -9.79208797e-02
2.05905378e-01 -2.22389773e-01 -6.18907809e-01 1.55973029e+00
1.01544571e+00 2.00551841e-02 5.20866454e-01 -1.40114534e+00
-5.38267314e-01 8.72681320e-01 -4.01412785e-01 3.37767929e-01
3.20367724e-01 -1.56381242e-02 7.63880134e-01 -3.06462675e-01
-3.79532166e-02 1.07547319e+00 6.07788026e-01 7.18263924e-01
-1.21294248e+00 -1.12095618e+00 -1.43694714e-01 2.32070372e-01
-1.44848180e+00 -8.23721170e-01 1.37291825e+00 6.79857098e-03
1.76984757e-01 7.95676053e-01 2.36019999e-01 1.30175161e+00
-1.77171037e-01 7.56838262e-01 9.63388324e-01 -5.37249982e-01
2.52103835e-01 8.92414525e-03 -1.89631894e-01 3.91654789e-01
2.81456620e-01 2.75853544e-01 -5.66530764e-01 -6.57767594e-01
5.24515450e-01 -3.67861480e-01 -8.54570329e-01 -2.04002019e-02
-8.19995224e-01 6.08295262e-01 3.51403207e-01 4.26859736e-01
-3.82725638e-03 5.92522398e-02 4.04219270e-01 4.17499125e-01
3.57179374e-01 5.19607246e-01 -2.11410284e-01 1.39035553e-01
-9.24556196e-01 1.12474486e-01 8.19944620e-01 6.19460106e-01
-6.30442426e-02 1.05133390e+00 1.08153172e-01 9.79843616e-01
8.25074017e-02 8.60168755e-01 3.66879910e-01 -8.28350902e-01
3.26648057e-01 -1.58246055e-01 -4.69400436e-02 -1.48595798e+00
-1.02148764e-01 -7.66249955e-01 -8.88624072e-01 2.88258880e-01
6.00075781e-01 -5.70139825e-01 -5.14625728e-01 1.80857790e+00
1.49893418e-01 6.25563562e-01 4.57027435e-01 8.58299434e-01
6.37099862e-01 7.18665540e-01 -2.13163555e-01 -3.59671652e-01
1.05781507e+00 -7.54184008e-01 -9.04281795e-01 -1.99051186e-01
1.78204507e-01 -9.15929258e-01 6.74520195e-01 7.04983711e-01
-1.05975401e+00 -3.27733934e-01 -1.45155728e+00 6.21722639e-01
-1.49515579e-02 -6.04335256e-02 2.72811145e-01 1.47641623e+00
-7.37216532e-01 4.04370695e-01 -5.90147018e-01 3.20029408e-01
3.57243180e-01 5.29869795e-01 -2.16525987e-01 2.34091580e-01
-1.45090258e+00 3.72093141e-01 -8.59753937e-02 2.62319148e-01
-1.19496596e+00 -4.84447986e-01 -7.10136712e-01 1.34709388e-01
3.68820816e-01 -2.78614461e-01 1.01812530e+00 -1.16833699e+00
-1.68570149e+00 3.19405913e-01 2.48267382e-01 -8.22081208e-01
4.82085437e-01 -3.66746098e-01 -1.07424462e+00 4.41983849e-01
-3.69994968e-01 -5.68291843e-02 1.65638208e+00 -1.56166506e+00
-2.18016222e-01 -1.50438339e-01 -3.75213549e-02 -1.35635495e-01
-9.16245699e-01 4.34460998e-01 -6.25950992e-02 -1.43104136e+00
-1.17591359e-01 -8.95913780e-01 -6.39442503e-02 -2.26758465e-01
-7.10952044e-01 6.17641509e-01 9.51215208e-01 -7.88748205e-01
1.26300406e+00 -2.42312813e+00 1.89891949e-01 3.76773924e-01
-1.96647886e-02 8.39923322e-01 -3.22254777e-01 3.01221550e-01
-2.24279910e-01 2.08173499e-01 -4.67818141e-01 -1.22458734e-01
-1.06975310e-01 -1.57056391e-01 -9.91928101e-01 7.42689252e-01
-8.17087889e-02 4.49096262e-01 -6.27281606e-01 -3.07151861e-02
-2.46555023e-02 9.03318644e-01 -4.29439604e-01 4.08045411e-01
2.84909606e-01 4.25915271e-01 -3.95849168e-01 5.65692782e-01
8.30325544e-01 4.70486104e-01 9.46407765e-02 -1.87970370e-01
4.85558957e-01 1.21875800e-01 -1.31070900e+00 9.84600902e-01
-2.96877176e-01 8.10580552e-01 7.71544039e-01 -9.79983568e-01
8.65933299e-01 7.24471688e-01 2.39306003e-01 -1.37370273e-01
4.81182605e-01 1.96984962e-01 2.52462715e-01 -3.49725962e-01
1.74486652e-01 -1.60704151e-01 -1.82328954e-01 2.98250288e-01
-1.10688858e-01 -4.28819135e-02 -4.77325201e-01 2.40431819e-02
9.17987704e-01 -4.29416478e-01 -2.01704912e-02 -9.15546864e-02
6.78291440e-01 -6.48850322e-01 6.17994010e-01 7.42845654e-01
-3.48828197e-01 5.16105473e-01 2.56174386e-01 2.18580425e-01
-4.74159330e-01 -1.08454990e+00 -3.49626243e-02 9.25133944e-01
1.62441537e-01 -3.49947453e-01 -1.11581028e+00 -5.85717201e-01
-1.20290108e-01 4.87798214e-01 -2.46576056e-01 -6.75593615e-01
-6.72217786e-01 -8.13807607e-01 1.57081306e+00 3.91297787e-01
5.37978590e-01 -6.59080327e-01 -6.20271415e-02 5.00202999e-02
-1.99963361e-01 -1.17236054e+00 -4.79711950e-01 1.56464502e-01
-6.59049809e-01 -9.03565943e-01 -8.13509047e-01 -7.55575299e-01
4.63415891e-01 4.79677796e-01 3.68731618e-01 1.00966103e-01
2.70602815e-02 4.25771445e-01 -3.41329575e-01 -6.23262823e-01
-8.81189942e-01 -3.48367929e-01 4.64905620e-01 4.48282152e-01
-1.27219379e-01 -8.38344097e-01 -3.36159468e-01 4.42419857e-01
-1.13732171e+00 -6.70836091e-01 4.90148067e-02 5.67553997e-01
2.93869879e-02 7.97466636e-01 8.59233677e-01 -3.53512466e-01
9.62028801e-01 -3.10230762e-01 -4.84792650e-01 4.36767302e-02
-2.00288445e-01 -2.14007005e-01 1.10524786e+00 -8.93916786e-01
-8.57283592e-01 -2.53411472e-01 -3.59487593e-01 -4.55717027e-01
-1.96349308e-01 1.85331509e-01 -6.36759877e-01 -6.07977629e-01
6.79249167e-01 3.82794172e-01 -1.27232715e-01 -5.84884822e-01
3.64470482e-01 7.67184377e-01 7.02401578e-01 -6.14805698e-01
1.31102073e+00 4.24894005e-01 -1.38367951e-01 -1.26923573e+00
-4.69640166e-01 -4.55912538e-02 -1.26711875e-01 -2.18560472e-01
5.27285635e-01 -8.49560797e-01 -6.58138275e-01 9.97684181e-01
-1.05596304e+00 -1.14525072e-01 2.16336340e-01 3.99890989e-01
-2.77827531e-01 7.48068631e-01 -6.77352548e-01 -1.07754886e+00
-3.89396220e-01 -1.10111928e+00 4.69629347e-01 5.43431602e-02
-2.59729102e-02 -9.83655632e-01 -1.39560178e-02 3.35568607e-01
4.47983533e-01 3.07672620e-01 7.48812914e-01 -9.15096164e-01
-3.57964456e-01 -3.30383658e-01 2.69148439e-01 7.70180643e-01
3.37654471e-01 -1.58935413e-02 -1.29484618e+00 -5.17139494e-01
5.60946941e-01 -5.99119551e-02 6.61059320e-01 2.37672150e-01
1.01648355e+00 -7.21265376e-01 -8.05112347e-02 8.29823494e-01
1.06681478e+00 5.20952761e-01 6.33725822e-01 2.44365975e-01
6.07938230e-01 5.97317994e-01 1.08681535e-02 2.19474331e-01
-6.12476707e-01 6.78424180e-01 3.71027410e-01 -1.68143436e-01
-8.46641734e-02 -8.34227875e-02 7.23935664e-01 9.53250587e-01
1.72593053e-02 -6.18088245e-01 -4.98936206e-01 4.78982329e-01
-1.34078026e+00 -1.07067835e+00 8.61886889e-02 2.24034238e+00
9.64734137e-01 3.04144740e-01 2.87614465e-01 8.61319542e-01
8.79681110e-01 5.30834198e-01 -1.87237650e-01 -2.97717750e-01
-4.07691598e-01 2.49419749e-01 4.41605359e-01 5.85274577e-01
-1.29031336e+00 6.94790661e-01 7.03371906e+00 1.23181176e+00
-1.12742293e+00 3.25405933e-02 2.28703544e-01 -2.61072427e-01
-2.14048512e-02 -3.62804741e-01 -4.38417107e-01 5.08593798e-01
9.25728858e-01 -2.52245277e-01 5.15955865e-01 5.70275724e-01
3.57521288e-02 4.46085483e-01 -6.17660165e-01 7.90387392e-01
2.62876242e-01 -9.60266411e-01 2.05864999e-02 -2.02061683e-01
3.64116251e-01 -6.36893094e-01 2.69787550e-01 -2.38081068e-01
1.22495420e-01 -8.64455640e-01 7.73700595e-01 -1.28474887e-02
5.49759507e-01 -1.19504488e+00 3.28652591e-01 1.66595280e-01
-1.06974661e+00 -2.70691723e-01 -2.39153683e-01 2.34551653e-02
1.51393399e-01 5.25496960e-01 -3.83139223e-01 5.71705341e-01
4.85083431e-01 2.81696200e-01 -3.48331720e-01 1.04789901e+00
-5.62609911e-01 1.25963569e+00 -2.02969849e-01 1.88851163e-01
-6.88616559e-02 6.16967957e-03 1.28300214e+00 1.29243100e+00
1.51305228e-01 -1.26404524e-01 -1.22677922e-01 5.89686573e-01
-1.30066425e-01 -8.21621865e-02 -9.05632496e-01 4.01142500e-02
6.92557991e-01 9.40276563e-01 -7.28626966e-01 1.57894894e-01
7.85738230e-04 1.01902568e+00 -1.92571089e-01 6.64916456e-01
-9.41077828e-01 -9.44023609e-01 8.99646461e-01 -2.22403347e-01
3.01680803e-01 -2.17802793e-01 -3.19586277e-01 -1.07917738e+00
-7.13554174e-02 -1.57950461e+00 3.29839081e-01 -4.37625140e-01
-1.20546961e+00 9.17143822e-01 -2.45748818e-01 -1.42261863e+00
4.73304763e-02 -2.94321716e-01 -7.21910894e-01 4.80233461e-01
-1.14877117e+00 -8.10377896e-01 3.56941074e-01 8.85398567e-01
1.78192213e-01 -5.47010958e-01 8.92697096e-01 2.84925699e-01
-7.65999794e-01 1.12393069e+00 3.15269232e-01 5.40358126e-01
6.58699989e-01 -8.01771760e-01 3.06463510e-01 1.40678203e+00
4.58544672e-01 7.37527609e-01 1.09174585e+00 -4.67144102e-01
-1.25827217e+00 -9.63131666e-01 2.08128676e-01 -1.14667397e-02
8.24894667e-01 -6.73719704e-01 -1.13879204e+00 2.45152310e-01
5.11831701e-01 -3.03078055e-01 9.08588707e-01 -4.07228798e-01
-6.99430406e-01 -2.48205110e-01 -1.11786950e+00 8.13547909e-01
6.61713481e-01 -9.04392898e-01 -6.73490107e-01 5.76507337e-02
9.74132836e-01 -1.06130689e-01 -5.68725705e-01 2.75485724e-01
4.26657140e-01 -8.09394836e-01 1.52706110e+00 -6.13804102e-01
2.13576853e-02 -3.52620959e-01 -2.31066629e-01 -1.29257691e+00
-9.27032903e-02 -1.32366252e+00 -2.91075855e-01 1.62214851e+00
2.69754261e-01 -6.33924365e-01 4.28664148e-01 3.52006227e-01
2.34934300e-01 -1.99255973e-01 -1.10116243e+00 -1.20367026e+00
1.21698819e-01 -6.82805061e-01 5.91293871e-01 1.03731763e+00
6.70840144e-02 3.27234454e-02 -9.57466960e-01 7.82575309e-01
8.25390339e-01 -1.99600667e-01 6.82967007e-01 -7.04523206e-01
-5.25291026e-01 -3.84998560e-01 -5.56982279e-01 -7.57096529e-01
4.15244222e-01 -7.49039710e-01 -6.50719479e-02 -6.57743335e-01
-4.85859990e-01 -1.75695598e-01 -4.74406213e-01 2.16008365e-01
-2.99540818e-01 4.88068223e-01 3.00323546e-01 4.94149234e-03
-8.11029673e-02 4.81685579e-01 7.24921823e-01 -3.14201862e-01
-1.94961727e-01 3.91427070e-01 -7.36996293e-01 9.32821870e-01
8.45468104e-01 -9.83786523e-01 -5.83568871e-01 -2.15003431e-01
-9.68355387e-02 1.04518138e-01 4.82419819e-01 -1.24858224e+00
5.96518852e-02 4.96923132e-03 -4.02426161e-02 5.76758161e-02
3.56237054e-01 -1.21765959e+00 4.29252759e-02 4.55610543e-01
-4.98073220e-01 -5.21969855e-01 5.53003550e-01 8.81605387e-01
-2.59176075e-01 -3.74001622e-01 9.25452828e-01 4.89728779e-01
4.93053012e-02 -6.73760548e-02 -5.46470404e-01 3.95167582e-02
8.98215294e-01 -1.74290501e-02 -2.75288522e-01 -7.14893639e-01
-7.41635323e-01 -3.36650163e-01 1.96422905e-01 3.75030398e-01
6.88814044e-01 -1.31027162e+00 -6.98114514e-01 5.09205580e-01
-4.59265023e-01 -6.69780552e-01 2.51542032e-01 5.63691854e-01
-1.43191412e-01 -1.26981884e-02 -1.27437219e-01 -1.77872732e-01
-1.73554969e+00 8.94079149e-01 5.21370232e-01 1.77261904e-01
-4.34643447e-01 9.37006712e-01 1.02528222e-01 9.64625254e-02
6.61203086e-01 1.17985971e-01 -1.09240785e-01 -1.74333811e-01
8.59426916e-01 7.01131761e-01 -6.82315528e-02 -7.07143009e-01
-4.26877379e-01 5.23603797e-01 3.46469045e-01 -1.99913308e-01
1.16053033e+00 -6.82520196e-02 -7.39990100e-02 2.44120453e-02
1.32823789e+00 9.02394533e-01 -1.01696503e+00 -2.07638562e-01
-2.35857666e-01 -7.19070554e-01 1.12820372e-01 -5.13864160e-01
-1.37109768e+00 9.50690866e-01 5.35913825e-01 8.55733156e-01
1.51800466e+00 -4.22008157e-01 9.50849593e-01 1.94770947e-01
2.00455695e-01 -6.46612227e-01 1.13102332e-01 5.33895828e-02
8.69911730e-01 -5.62285423e-01 -9.18294564e-02 -5.35009325e-01
-5.07435203e-01 9.83325958e-01 5.70447966e-02 -2.00941965e-01
7.13918328e-01 7.98344672e-01 3.26253891e-01 2.47476950e-01
-1.65458336e-01 2.29769289e-01 5.45073271e-01 1.05646431e+00
2.81315893e-02 -8.03263579e-03 6.56559095e-02 1.04223621e+00
-2.58117616e-01 -4.70968306e-01 5.69109023e-01 7.53619671e-01
-4.35263991e-01 -1.21335566e+00 -1.06938279e+00 -1.58146366e-01
-1.15139508e+00 -1.28241047e-01 -7.12728083e-01 2.96475947e-01
-2.37263277e-01 1.52523005e+00 -5.98339200e-01 -6.28543258e-01
2.24459440e-01 6.24046139e-02 1.71544284e-01 7.06964592e-03
-7.47968435e-01 4.49913114e-01 2.05841854e-01 -7.95537308e-02
-5.83630919e-01 -3.99794281e-01 -9.30025816e-01 -3.43183279e-01
-6.52160764e-01 3.24540108e-01 2.78819680e-01 7.49214709e-01
3.75094473e-01 5.84629178e-01 9.62698042e-01 -6.21282697e-01
-6.75952435e-01 -5.45384407e-01 -7.93304086e-01 1.59205735e-01
8.39724123e-01 -4.55984384e-01 -8.79833877e-01 1.69914782e-01] | [14.016773223876953, 5.831924915313721] |
b2c4adce-d052-432e-b536-164198048e0f | alto-a-large-scale-dataset-for-uav-visual | 2207.12317 | null | https://arxiv.org/abs/2207.12317v1 | https://arxiv.org/pdf/2207.12317v1.pdf | ALTO: A Large-Scale Dataset for UAV Visual Place Recognition and Localization | We present the ALTO dataset, a vision-focused dataset for the development and benchmarking of Visual Place Recognition and Localization methods for Unmanned Aerial Vehicles. The dataset is composed of two long (approximately 150km and 260km) trajectories flown by a helicopter over Ohio and Pennsylvania, and it includes high precision GPS-INS ground truth location data, high precision accelerometer readings, laser altimeter readings, and RGB downward facing camera imagery. In addition, we provide reference imagery over the flight paths, which makes this dataset suitable for VPR benchmarking and other tasks common in Localization, such as image registration and visual odometry. To the author's knowledge, this is the largest real-world aerial-vehicle dataset of this kind. Our dataset is available at https://github.com/MetaSLAM/ALTO. | ['Sebastian Scherer', 'Howie Choset', 'Ji Zhang', 'Peng Yin', 'Ivan Cisneros'] | 2022-07-19 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [-3.49991083e-01 -6.29285932e-01 -3.93896520e-01 -1.82600096e-01
-1.21288322e-01 -1.11634040e+00 7.90399790e-01 3.86263393e-02
-4.67097253e-01 5.69328249e-01 2.99380068e-02 -4.35122848e-01
-2.01812778e-02 -7.79020667e-01 -4.23565894e-01 -3.59243095e-01
-3.40774298e-01 3.49077106e-01 3.03206563e-01 -4.36316550e-01
1.28545001e-01 9.02729034e-01 -1.47317672e+00 -5.31590462e-01
3.90655011e-01 9.54563677e-01 1.03830606e-01 8.18760812e-01
7.21398711e-01 5.37724495e-01 -1.41190574e-01 3.76502648e-02
7.75906801e-01 -7.04631731e-02 -6.81358635e-01 1.79316640e-01
9.48583663e-01 -5.46804309e-01 -6.06599510e-01 9.23486650e-01
2.54766077e-01 3.99207562e-01 3.04379255e-01 -1.51219404e+00
-5.60899675e-01 -9.81554240e-02 -4.02672231e-01 3.16930145e-01
6.66153431e-01 5.38693428e-01 8.67790103e-01 -6.20190382e-01
8.48584712e-01 5.27150154e-01 9.18914258e-01 -6.06789961e-02
-8.50809515e-01 -2.07663834e-01 -1.15126051e-01 -1.26568168e-01
-1.85948586e+00 -4.53161210e-01 2.73178518e-01 -8.66104960e-01
9.44483995e-01 1.00797638e-01 1.00799859e+00 8.70374680e-01
9.46515352e-02 1.14406452e-01 7.82741010e-01 -1.23874426e-01
3.05740952e-01 -2.24995330e-01 -2.70885319e-01 1.08390999e+00
5.60231388e-01 4.73919928e-01 -3.14177811e-01 -1.93263933e-01
9.21597660e-01 3.02163005e-01 -5.95744312e-01 -8.13227713e-01
-1.64861488e+00 5.30832946e-01 9.10450935e-01 -1.17900390e-02
-4.66117322e-01 9.03395563e-02 1.40459370e-02 2.28181913e-01
8.18617046e-02 3.73782694e-01 -4.10728693e-01 -2.71977395e-01
-1.04586482e+00 2.44542569e-01 5.68114340e-01 1.44508731e+00
1.03536808e+00 1.43597662e-01 6.06370986e-01 1.43197238e-01
5.10723948e-01 9.86659050e-01 4.51034814e-01 -1.08805645e+00
5.19403040e-01 7.24384785e-01 6.07727587e-01 -1.40150690e+00
-4.99165863e-01 1.36267960e-01 -4.96626645e-01 2.28405118e-01
2.29483783e-01 -4.30839390e-01 -7.43432224e-01 1.07493234e+00
3.28186631e-01 4.68658537e-01 1.10867992e-01 1.40953493e+00
8.77644181e-01 4.88445759e-01 -3.40694457e-01 3.90240818e-01
1.15919125e+00 -7.71902502e-01 -2.30250314e-01 -5.32660127e-01
9.28459466e-01 -6.04868889e-01 5.73800266e-01 -1.93989322e-01
-3.33730668e-01 -4.84473944e-01 -1.22218001e+00 5.05834669e-02
-7.24179983e-01 6.65798426e-01 5.94023526e-01 2.53955036e-01
-1.38325298e+00 4.04511988e-01 -9.50532436e-01 -9.78011608e-01
2.36512005e-01 2.02262744e-01 -8.91288340e-01 -1.16863446e-02
-9.60463107e-01 8.27322125e-01 3.07605892e-01 2.13363677e-01
-9.55731928e-01 -3.72534364e-01 -1.29544187e+00 -6.41866624e-01
-5.34937605e-02 -5.74149668e-01 1.03530931e+00 -5.71991980e-01
-1.20170224e+00 1.02462101e+00 2.28301175e-02 -7.72800088e-01
2.62201458e-01 -8.09459016e-02 -5.06712496e-01 -1.40478853e-02
3.40555370e-01 7.53456593e-01 5.12875497e-01 -8.43695402e-01
-7.77753055e-01 -5.61355889e-01 1.82166889e-01 3.81311804e-01
3.11442524e-01 -2.41244614e-01 -1.90883294e-01 -3.55784118e-01
2.68805802e-01 -1.31204987e+00 -2.73330688e-01 -1.36729218e-02
-2.33256012e-01 5.98940969e-01 7.84310043e-01 -5.34620762e-01
8.37010324e-01 -2.31470013e+00 -1.07596040e-01 4.02400717e-02
-2.52547413e-02 2.02014580e-01 -5.05267940e-02 6.26424372e-01
7.40668401e-02 -8.31696242e-02 -1.46210477e-01 -1.85700789e-01
-1.25403017e-01 2.24506408e-01 -6.45046234e-01 1.05895674e+00
-2.23755047e-01 8.89105976e-01 -1.10462892e+00 -6.71832636e-02
6.99114799e-01 1.58262983e-01 -1.04317002e-01 -5.87931871e-02
2.61123553e-02 5.18139601e-01 -2.35622853e-01 1.08516073e+00
7.11290240e-01 1.12759836e-01 -2.55011152e-02 -1.14671066e-02
-7.62040734e-01 1.36737555e-01 -1.11128926e+00 1.76458454e+00
-2.28895247e-02 1.12074375e+00 -9.57288221e-02 -2.59148180e-01
1.03252625e+00 -4.13803663e-03 3.66542161e-01 -2.53551364e-01
3.91642541e-01 5.19173779e-03 -5.88361740e-01 -6.07268773e-02
1.10242653e+00 3.31439286e-01 -9.87795740e-02 -3.15502018e-01
1.18105769e-01 -4.84991163e-01 2.20222831e-01 -1.33135423e-01
9.94451046e-01 1.48398414e-01 7.49384224e-01 -1.92440748e-01
2.70371914e-01 9.98342633e-01 5.02024531e-01 3.76257509e-01
-5.37041485e-01 7.45871961e-01 -1.13322377e-01 -8.39316308e-01
-8.65623891e-01 -1.05969226e+00 -1.26393244e-01 3.62577081e-01
6.32589042e-01 -7.70195425e-01 -3.75670373e-01 -2.20649049e-01
2.20720127e-01 1.26187548e-01 -3.78181636e-01 2.93224484e-01
-4.28923778e-02 -4.52403128e-01 7.45291173e-01 5.52675664e-01
8.28610003e-01 -4.89958078e-01 -7.89011598e-01 -1.32042840e-01
-1.95112199e-01 -1.36141992e+00 -2.06955865e-01 -9.69811380e-02
-9.41557586e-01 -1.37003636e+00 -3.82167339e-01 -5.42525232e-01
6.77976668e-01 7.97834873e-01 8.25833857e-01 -9.09513831e-02
-3.29668760e-01 7.07581222e-01 -2.36748934e-01 -1.83052182e-01
4.71169114e-01 -1.18134264e-02 5.70747256e-01 -2.24184200e-01
4.65406150e-01 -2.64924914e-01 -4.57280725e-01 6.29836679e-01
-3.18627328e-01 -4.28810984e-01 5.56076281e-02 4.78831917e-01
8.41121316e-01 3.75025272e-02 -7.69788563e-01 1.54087301e-02
6.81381002e-02 -5.80582321e-01 -1.51480532e+00 -2.56741911e-01
-2.26504534e-01 -7.05843985e-01 4.49281693e-01 1.13078326e-01
-2.56260395e-01 4.94567066e-01 3.24976325e-01 -5.22575676e-01
-8.14570427e-01 4.01957363e-01 2.69146144e-01 -6.39879763e-01
8.54392886e-01 1.60391256e-01 -2.37096578e-01 -8.92366245e-02
3.91589999e-01 6.66028202e-01 7.47475624e-01 1.84513032e-01
1.01748753e+00 6.21640086e-01 2.33497173e-01 -1.33797693e+00
-4.19844419e-01 -8.58020484e-01 -1.11399496e+00 -2.16495246e-01
8.81061792e-01 -1.49720407e+00 -4.97966617e-01 4.40874457e-01
-7.85249233e-01 -6.62412703e-01 -1.86365455e-01 7.92931199e-01
-6.04693532e-01 1.69067457e-01 -2.20543906e-01 -4.77477342e-01
1.36378750e-01 -9.47837532e-01 1.08504617e+00 3.67006391e-01
-2.29537874e-01 -1.15144920e+00 5.62395871e-01 1.26082236e-02
9.84790027e-02 6.53689384e-01 -2.83675998e-01 -7.24288896e-02
-8.73307288e-01 -4.95509207e-01 2.93089785e-02 -5.82506657e-02
1.65407553e-01 2.81795293e-01 -6.76439166e-01 -5.84204793e-01
-5.95173836e-01 -1.50993183e-01 7.11587846e-01 3.94955426e-01
1.91154480e-01 -1.87407598e-01 -5.98037601e-01 1.01121962e+00
1.65509498e+00 1.23461597e-01 5.04329085e-01 7.01496065e-01
7.52821863e-01 -3.25758220e-03 9.66263831e-01 5.67734301e-01
8.00018132e-01 8.37461770e-01 8.67601752e-01 -1.52120844e-01
2.10662708e-01 -6.11247718e-01 2.97846675e-01 2.34111384e-01
-4.07056987e-01 -1.84389010e-01 -1.46725261e+00 7.81491160e-01
-1.78157544e+00 -9.29761171e-01 -3.63196075e-01 2.40729594e+00
-8.98197368e-02 -4.91601974e-01 -2.55131349e-02 -3.20056021e-01
4.04154390e-01 6.19435072e-01 -9.12893116e-02 1.27197191e-01
-1.78217106e-02 -4.92005199e-01 1.29808271e+00 6.19753063e-01
-1.78005385e+00 1.28607142e+00 6.25327921e+00 -4.73568812e-02
-1.22030210e+00 -1.13881074e-01 -2.62714386e-01 1.18179277e-01
3.58771563e-01 3.17128658e-01 -1.04063606e+00 3.26610059e-01
1.05094421e+00 3.01631857e-02 6.35807455e-01 1.24446642e+00
1.19407929e-03 -4.47641939e-01 -5.23928404e-01 1.13471484e+00
6.06742762e-02 -1.36280799e+00 -3.10327142e-01 3.30280781e-01
8.16985369e-01 8.97704542e-01 6.14748672e-02 -2.65512377e-01
6.16558254e-01 -6.63516998e-01 6.84030592e-01 3.55259567e-01
8.20855677e-01 -6.24423027e-01 9.38796878e-01 2.71004975e-01
-1.59600258e+00 -3.41284871e-02 -6.73603237e-01 -4.52665299e-01
1.85830817e-01 2.91896224e-01 -9.22882438e-01 6.88036680e-01
9.20048296e-01 1.51963961e+00 -8.81150126e-01 1.28368616e+00
-4.79687572e-01 2.28944361e-01 -3.96164984e-01 3.81388694e-01
5.65469682e-01 -6.40186787e-01 7.63804793e-01 7.57923663e-01
7.08320320e-01 2.09700957e-01 4.92943406e-01 4.01005030e-01
1.94931343e-01 -2.38112211e-01 -1.87327945e+00 -1.30814314e-01
6.84215188e-01 1.27785695e+00 -6.03814602e-01 -5.66134043e-02
-1.65916324e-01 9.98964667e-01 -3.43051963e-02 2.50943780e-01
-8.83290291e-01 -4.67379868e-01 1.36109424e+00 2.21838027e-01
1.28015548e-01 -1.12180758e+00 9.74288657e-02 -1.25024343e+00
-2.28493169e-01 -2.22529918e-01 4.55639437e-02 -1.12861204e+00
-5.95311999e-01 6.88374698e-01 -1.50434971e-01 -1.94494033e+00
-2.93527544e-01 -8.48712206e-01 -4.23286587e-01 7.33004570e-01
-1.47504807e+00 -1.06960154e+00 -9.01074588e-01 6.31451905e-01
9.60603449e-03 -4.41590518e-01 8.95807922e-01 -5.64241186e-02
-5.01392186e-01 -1.83283955e-01 2.16777176e-01 4.70785499e-01
6.33347034e-01 -1.03970718e+00 9.91698384e-01 1.18182647e+00
4.51204032e-01 7.20781863e-01 4.32968289e-01 -7.60438085e-01
-1.66481221e+00 -1.53903866e+00 8.11789811e-01 -1.03298259e+00
8.30190063e-01 -3.05578187e-02 -2.19859257e-01 1.39693773e+00
6.38632327e-02 3.75439733e-01 3.76965553e-01 -3.27848971e-01
2.23535880e-01 -9.15332660e-02 -1.10592461e+00 6.36677504e-01
1.05768800e+00 -5.31287313e-01 -4.43654180e-01 4.01105136e-01
2.63647825e-01 -1.05268097e+00 -8.99893522e-01 2.94626713e-01
4.02717054e-01 -1.01552427e+00 1.13703918e+00 -9.15886611e-02
-1.04452036e-02 -1.06732225e+00 -5.61385214e-01 -1.52313793e+00
-5.48253834e-01 -3.45903695e-01 3.28416228e-01 9.52673018e-01
2.06310526e-01 -9.63705182e-01 4.59384173e-01 2.25282744e-01
-7.59873241e-02 -7.05664232e-02 -7.91145980e-01 -9.86972272e-01
-7.41588593e-01 -2.48460308e-01 7.77574599e-01 9.83958304e-01
-1.20978899e-01 -1.93092123e-01 -3.54938984e-01 9.76114690e-01
4.68986571e-01 2.42110550e-01 1.13775778e+00 -1.32350075e+00
4.97544646e-01 1.30478675e-02 -1.26801455e+00 -1.09485948e+00
1.31193593e-01 -7.88453937e-01 6.24597706e-02 -1.53124869e+00
-8.25487673e-01 -3.04272205e-01 4.25869048e-01 6.76485002e-01
6.43755198e-01 6.68636858e-01 1.67785704e-01 3.36181492e-01
-5.61044097e-01 4.34010804e-01 5.64845145e-01 1.13697164e-02
-1.14918999e-01 -5.89604564e-02 -9.60070491e-02 8.76578152e-01
1.07077336e+00 -3.76767099e-01 -3.26153487e-01 -5.56364417e-01
-1.20933570e-01 4.74319346e-02 8.88871133e-01 -1.53795588e+00
4.23478991e-01 -3.26741993e-01 5.91641128e-01 -7.56970465e-01
5.13387978e-01 -1.08490276e+00 4.10969347e-01 3.70204270e-01
3.85434628e-01 7.21433580e-01 4.42908347e-01 4.72964168e-01
-4.93179232e-01 1.77391753e-01 5.95029652e-01 -2.64253378e-01
-1.50936162e+00 6.22216880e-01 -5.95693886e-01 3.65345478e-02
1.27564085e+00 -3.81547928e-01 -6.08731449e-01 -2.17637241e-01
-1.81524828e-01 3.31737041e-01 1.39698660e+00 4.89883512e-01
6.33126497e-01 -1.51391625e+00 -1.96476728e-01 7.35108316e-01
5.45192659e-01 -1.29552215e-01 -8.08757022e-02 1.02849185e+00
-1.16317368e+00 8.17152202e-01 -5.92983842e-01 -7.39589095e-01
-1.06533504e+00 4.85089779e-01 3.84260356e-01 5.89875638e-01
-6.53710544e-01 4.52649355e-01 -5.17515302e-01 -7.49567211e-01
-2.86815792e-01 -6.29541874e-01 -1.08101018e-01 3.93065345e-03
6.81605697e-01 5.19463539e-01 1.58952754e-02 -1.27613509e+00
-8.60434115e-01 1.00143814e+00 7.63186276e-01 -1.21962048e-01
8.43237877e-01 -4.24310207e-01 -1.34832948e-01 3.83000225e-01
8.54785979e-01 -2.21935846e-02 -1.23117673e+00 1.35267466e-01
-2.35551875e-02 -9.05167639e-01 3.51926148e-01 -2.93414146e-01
-7.94305921e-01 6.91116810e-01 7.09840655e-01 -4.25268821e-02
8.23899329e-01 -3.62777621e-01 4.74587798e-01 6.83566511e-01
1.14881015e+00 -5.97387314e-01 -7.14990675e-01 1.00524032e+00
6.73013151e-01 -1.39356828e+00 9.20886844e-02 -3.94191928e-02
-8.42481971e-01 9.59282160e-01 5.56278110e-01 -3.30273360e-01
4.55069691e-01 1.41849905e-01 3.60581130e-01 -8.56535733e-02
-2.38824129e-01 -6.67864799e-01 1.90277353e-01 1.03134513e+00
3.23364958e-02 2.88878024e-01 4.09397274e-01 -1.01274863e-01
-4.91113871e-01 1.75003871e-01 5.84681094e-01 1.09915698e+00
-4.89141256e-01 -5.58112621e-01 -4.30628896e-01 -8.42760354e-02
3.51005375e-01 -8.61922503e-02 -5.98059058e-01 9.17474866e-01
6.77775666e-02 8.46025825e-01 1.18841104e-01 -6.12684786e-01
5.29757261e-01 -2.49013230e-01 1.79959059e-01 -3.66429538e-01
-1.83555946e-01 -6.92842543e-01 1.42138973e-01 -9.15257215e-01
-2.91895419e-01 -6.53639138e-01 -1.04006839e+00 -7.30247021e-01
6.91286996e-02 -6.97820634e-02 6.81682706e-01 4.81167257e-01
6.56491876e-01 -2.29438022e-02 5.31289756e-01 -1.39180720e+00
-1.04231887e-01 -5.99088073e-01 -9.51151907e-01 -2.21719727e-01
7.78467059e-01 -8.37122262e-01 -2.47825980e-01 1.71814442e-01] | [7.3542962074279785, -1.977980613708496] |
23e42ff8-00ae-4667-91b9-313c6be60aea | line-graph-enhanced-amr-to-text-generation | null | null | https://aclanthology.org/2020.acl-main.67 | https://aclanthology.org/2020.acl-main.67.pdf | Line Graph Enhanced AMR-to-Text Generation with Mix-Order Graph Attention Networks | Efficient structure encoding for graphs with labeled edges is an important yet challenging point in many graph-based models. This work focuses on AMR-to-text generation {--} A graph-to-sequence task aiming to recover natural language from Abstract Meaning Representations (AMR). Existing graph-to-sequence approaches generally utilize graph neural networks as their encoders, which have two limitations: 1) The message propagation process in AMR graphs is only guided by the first-order adjacency information. 2) The relationships between labeled edges are not fully considered. In this work, we propose a novel graph encoding framework which can effectively explore the edge relations. We also adopt graph attention networks with higher-order neighborhood information to encode the rich structure in AMR graphs. Experiment results show that our approach obtains new state-of-the-art performance on English AMR benchmark datasets. The ablation analyses also demonstrate that both edge relations and higher-order information are beneficial to graph-to-sequence modeling. | ['Kai Yu', 'Ruisheng Cao', 'Zhi Chen', 'Yanbin Zhao', 'Su Zhu', 'Lu Chen'] | 2020-07-01 | null | null | null | acl-2020-6 | ['graph-to-sequence'] | ['natural-language-processing'] | [ 6.64266467e-01 6.66740239e-01 -3.65990072e-01 -2.47297123e-01
-3.47886711e-01 -3.97545457e-01 6.63307130e-01 3.43611658e-01
1.16850346e-01 5.35646856e-01 6.57136977e-01 -7.20786154e-01
2.45763421e-01 -1.31842828e+00 -8.66021931e-01 -1.22934945e-01
-1.17367186e-01 3.64952087e-01 -2.10294113e-01 -6.10708356e-01
1.75917327e-01 2.66292781e-01 -1.09090686e+00 4.54018533e-01
8.89005303e-01 5.06368756e-01 3.08460236e-01 8.33254457e-01
-8.11935961e-01 1.32411480e+00 -5.02533019e-01 -7.07927763e-01
-1.26394629e-01 -6.22383475e-01 -1.02889383e+00 2.17023417e-01
1.85173318e-01 -1.35703713e-01 -9.63521421e-01 1.24073160e+00
2.68090010e-01 2.22088471e-01 7.32875347e-01 -1.11241698e+00
-1.43382037e+00 1.31769121e+00 -6.14817202e-01 1.64549515e-01
7.32216537e-01 2.68762052e-01 1.50870955e+00 -5.74541986e-01
6.67231739e-01 1.59284163e+00 2.96223015e-01 6.29665971e-01
-9.46799994e-01 -3.85372818e-01 7.20625341e-01 2.98727214e-01
-1.32804751e+00 -8.69207829e-02 1.03046167e+00 -1.93431973e-01
1.43104661e+00 3.10518712e-01 8.23276758e-01 1.24145472e+00
6.24806769e-02 8.19455981e-01 4.23334926e-01 -4.60003436e-01
-6.41361624e-02 -5.67770422e-01 3.92343283e-01 1.10691607e+00
5.02973557e-01 -3.10945034e-01 -2.83589929e-01 6.23567961e-02
8.94985855e-01 -1.37921080e-01 -3.65514308e-01 1.28934637e-01
-1.05000043e+00 1.09775984e+00 7.82220662e-01 3.72938097e-01
-4.26890522e-01 5.72698772e-01 4.22621131e-01 2.55461276e-01
5.95149100e-01 5.30325234e-01 1.07877264e-02 2.82838941e-02
-2.87846357e-01 1.51696587e-02 6.94542766e-01 1.44621372e+00
7.86695302e-01 6.61256015e-01 -4.97002512e-01 7.01451838e-01
3.46098959e-01 1.99213609e-01 4.89784390e-01 -3.00898254e-01
8.55092406e-01 9.94843006e-01 -3.63810480e-01 -1.31487131e+00
-2.00561017e-01 -4.86997843e-01 -1.25370967e+00 -6.10860705e-01
-1.09607667e-01 2.45610205e-03 -1.16398239e+00 1.66489995e+00
-1.22573756e-01 4.22856420e-01 1.97358713e-01 6.84286118e-01
1.35087144e+00 8.52014363e-01 8.93986747e-02 3.34978588e-02
1.35353947e+00 -1.14071381e+00 -9.47544277e-01 -6.90701485e-01
9.61080253e-01 -2.10566714e-01 1.24258590e+00 -1.96490660e-01
-1.05218983e+00 -5.68323255e-01 -1.02288413e+00 -2.91396379e-01
-3.85921657e-01 -1.21358454e-01 9.08145785e-01 3.00058842e-01
-1.23426473e+00 4.80262995e-01 -4.05843973e-01 -2.25359678e-01
3.34740996e-01 6.74586296e-02 -3.09761941e-01 -3.58642608e-01
-1.59125447e+00 5.90503573e-01 7.94402003e-01 2.62578428e-01
-6.57900035e-01 -3.76516879e-01 -1.59954774e+00 3.24131429e-01
6.03435755e-01 -1.01322055e+00 9.94978726e-01 -8.37397456e-01
-1.30030382e+00 7.52550244e-01 -3.51744652e-01 -6.93701744e-01
5.61621748e-02 -1.22938305e-02 -4.83217210e-01 2.15395868e-01
-7.91598484e-02 5.07778525e-01 5.67588627e-01 -1.06239653e+00
1.15218898e-02 -2.05580413e-01 3.19690734e-01 3.86967748e-01
-2.92022020e-01 -2.20968381e-01 -4.89262342e-01 -1.08962607e+00
1.53643247e-02 -7.07351744e-01 -4.55214292e-01 -6.48650348e-01
-8.25831175e-01 -5.26086390e-01 3.70101482e-01 -7.69856393e-01
1.70280576e+00 -1.60009694e+00 3.87612909e-01 1.32968992e-01
5.76589763e-01 1.86073795e-01 -5.98230064e-01 8.57343495e-01
-2.92501748e-01 6.68419242e-01 -3.40081632e-01 -5.45999967e-02
6.12472594e-02 2.78319448e-01 -4.94606435e-01 -3.57243828e-02
5.24235368e-01 1.69674289e+00 -1.15289164e+00 -3.64359647e-01
1.04590483e-01 3.86125028e-01 -4.15471286e-01 4.47374642e-01
-5.62314093e-01 2.87382036e-01 -6.25024259e-01 2.50223517e-01
3.78878146e-01 -7.24143744e-01 3.98397386e-01 -2.19512179e-01
6.17641807e-01 5.01214862e-01 -6.92925870e-01 1.97402787e+00
-5.38862050e-01 4.16732639e-01 -5.22249222e-01 -1.07420647e+00
1.14595962e+00 2.11371690e-01 -1.82424448e-02 -5.78463018e-01
-1.36089742e-01 -8.70591924e-02 2.75647014e-01 -2.98941135e-01
8.00400674e-01 5.60401119e-02 -7.63070062e-02 4.17952359e-01
-7.96396565e-03 -5.38830198e-02 3.93584311e-01 8.18443358e-01
1.20187795e+00 7.41953626e-02 5.47963977e-01 -6.40252233e-02
6.37133300e-01 -3.15996438e-01 2.39798144e-01 7.95204103e-01
4.26454067e-01 5.01117766e-01 7.74376452e-01 -4.43745703e-01
-9.48367894e-01 -6.56095326e-01 6.61411881e-01 7.20346868e-01
4.63675410e-02 -9.28894103e-01 -6.65060341e-01 -8.09841752e-01
-2.95419216e-01 1.25862610e+00 -6.39966309e-01 -6.12417817e-01
-7.69043744e-01 -5.89860201e-01 5.81827521e-01 7.12078512e-01
1.47374898e-01 -1.27203822e+00 1.39221787e-01 3.90666664e-01
-4.86108154e-01 -1.44342434e+00 -7.88344145e-01 -3.70020807e-01
-7.98919201e-01 -7.69328833e-01 -6.48089290e-01 -8.00056458e-01
9.44439113e-01 2.57530063e-01 1.59602737e+00 5.59933424e-01
-1.22631662e-01 4.21726644e-01 -7.55932450e-01 -1.83424354e-01
-7.51730561e-01 2.86712915e-01 -5.71735442e-01 1.28184944e-01
3.17878187e-01 -5.92184901e-01 -2.86854059e-01 -1.99252844e-01
-9.45196569e-01 4.55127597e-01 6.09084606e-01 6.97949350e-01
5.80644786e-01 -6.37761876e-02 7.81801283e-01 -1.28627336e+00
9.81374621e-01 -4.88759398e-01 -2.68909752e-01 3.23697984e-01
-5.03664315e-01 4.39235300e-01 9.31901097e-01 -2.09203780e-01
-9.53071237e-01 -2.60790348e-01 -2.64042199e-01 -3.28280002e-01
-4.83221598e-02 8.63043129e-01 -3.57877761e-01 3.39330196e-01
3.81485730e-01 5.83934724e-01 -2.53850281e-01 -1.52320921e-01
9.70642269e-01 4.23887283e-01 2.95686603e-01 -7.49595284e-01
6.46304607e-01 1.29537150e-01 2.50410557e-01 -9.40912962e-01
-8.97607625e-01 -2.64183432e-01 -4.23671395e-01 8.54348689e-02
9.76958573e-01 -9.74001050e-01 -4.66822863e-01 1.63870573e-01
-1.39869761e+00 -3.55842292e-01 -2.06341431e-01 1.33717075e-01
-4.92512405e-01 8.27079237e-01 -9.65175033e-01 -8.56545925e-01
-8.02076757e-01 -1.03436828e+00 1.18461323e+00 -7.17044845e-02
-1.49773091e-01 -1.20662451e+00 -3.18446994e-01 2.12025836e-01
1.50528491e-01 3.17622155e-01 1.29333854e+00 -7.00411558e-01
-7.69426405e-01 -4.26666662e-02 -4.00232136e-01 -2.20169015e-02
2.40224063e-01 -3.31398487e-01 -5.53453326e-01 -2.68821150e-01
-5.35091221e-01 -2.09840357e-01 1.02942109e+00 1.72719538e-01
1.54320717e+00 -7.02251911e-01 -3.42469364e-01 4.57950443e-01
1.16138017e+00 -3.21271345e-02 9.11764026e-01 -2.73370266e-01
1.49711680e+00 4.84185308e-01 2.04153866e-01 1.09332614e-01
6.08640492e-01 4.63570088e-01 5.95344782e-01 -1.51379138e-01
-4.25402433e-01 -9.17990685e-01 3.80828559e-01 1.37335539e+00
-6.98003173e-02 -8.53051722e-01 -8.16310465e-01 5.05197167e-01
-1.86975288e+00 -7.17537701e-01 -4.77444261e-01 1.88457274e+00
5.76998651e-01 9.09352079e-02 -1.74045831e-01 -1.52529240e-01
9.10913169e-01 6.25350714e-01 -2.68662512e-01 -5.36617935e-01
-7.26670921e-02 2.21979216e-01 2.88318634e-01 7.09931374e-01
-6.59431815e-01 1.31315553e+00 5.25860167e+00 7.44254529e-01
-5.51200509e-01 -2.91037083e-01 5.82155347e-01 5.62353373e-01
-9.58366156e-01 4.16786037e-02 -5.39766848e-01 2.40573689e-01
9.52607334e-01 -4.35351342e-01 6.28002226e-01 6.43782079e-01
-9.70335826e-02 7.22424328e-01 -1.13336408e+00 9.92689252e-01
2.06908971e-01 -1.47384465e+00 1.03100967e+00 5.80707304e-02
4.78297174e-01 -2.30165347e-01 -3.93821239e-01 4.44172025e-01
7.04373360e-01 -1.44834149e+00 4.43812907e-01 4.84185010e-01
9.66270983e-01 -9.06902671e-01 6.09648705e-01 1.79031432e-01
-1.86682165e+00 1.78453729e-01 -6.01573169e-01 -1.81832865e-01
4.11564469e-01 6.05416298e-01 -9.20483887e-01 1.32453454e+00
-5.27151413e-02 1.14610124e+00 -6.48025274e-01 3.73463154e-01
-8.19547117e-01 6.73651695e-01 2.84895003e-01 -3.71990860e-01
4.11087036e-01 -4.79951978e-01 6.59827948e-01 1.39662349e+00
3.08181554e-01 1.75635949e-01 3.53219748e-01 1.23307240e+00
-4.78125244e-01 1.74916595e-01 -1.18939495e+00 -7.40473986e-01
2.84339249e-01 1.11934447e+00 -7.90535271e-01 -5.22759140e-01
-4.77201581e-01 1.16006804e+00 7.45893598e-01 4.94230896e-01
-5.64878583e-01 -5.28741896e-01 3.99338573e-01 1.15452215e-01
1.54318467e-01 -3.87459517e-01 5.50010130e-02 -1.39784014e+00
-8.89343992e-02 -9.81156766e-01 6.21415377e-01 -9.93178368e-01
-1.43979728e+00 7.89105654e-01 -4.75451676e-03 -7.52768993e-01
-4.60910767e-01 -5.27212918e-01 -6.83526874e-01 1.01713634e+00
-1.40773058e+00 -1.47481155e+00 -1.58688009e-01 2.85518408e-01
7.24708915e-01 -1.81949005e-01 9.28305626e-01 1.91680749e-03
-3.07998478e-01 5.41551828e-01 -7.24856555e-01 5.78074515e-01
-2.09710181e-01 -1.43570900e+00 1.39754462e+00 1.10117912e+00
6.89100266e-01 7.86586165e-01 4.00644630e-01 -9.67665732e-01
-1.70605886e+00 -1.37343180e+00 7.89041519e-01 -2.69224882e-01
7.33624578e-01 -6.47630453e-01 -1.16004932e+00 1.17739701e+00
3.73434603e-01 -4.58452217e-02 5.29428363e-01 1.42212614e-01
-5.17563343e-01 3.71048570e-01 -4.50850993e-01 1.08344746e+00
1.75405276e+00 -7.01070666e-01 -5.57026684e-01 4.00870889e-01
1.50851834e+00 -4.13836449e-01 -7.20691085e-01 3.49371493e-01
-4.54112813e-02 -3.40859771e-01 9.01972353e-01 -1.14968348e+00
6.77889824e-01 -2.57242948e-01 -6.25889599e-02 -1.61199152e+00
-4.56474692e-01 -9.96311307e-01 -4.77355242e-01 1.23511934e+00
4.75930214e-01 -6.68603361e-01 5.90343714e-01 7.25932270e-02
-3.51185441e-01 -7.06547201e-01 -3.95058364e-01 -6.18273199e-01
-8.60165805e-03 -4.21445459e-01 8.06512833e-01 1.03544259e+00
-1.25119546e-02 1.19929385e+00 -4.84571248e-01 7.74814561e-02
5.81675231e-01 2.21341953e-01 6.73795640e-01 -1.07768321e+00
-4.65968430e-01 -3.61484647e-01 -4.00651902e-01 -1.33826256e+00
7.50855029e-01 -1.59102428e+00 -1.83415383e-01 -2.39960480e+00
1.78791881e-01 9.41295624e-02 -8.89911354e-02 4.38687027e-01
-6.72299981e-01 -3.87545079e-01 1.45178333e-01 -3.22413981e-01
-4.80170518e-01 9.17378366e-01 1.55748701e+00 -4.46988136e-01
1.17274083e-01 -3.49864155e-01 -9.51005697e-01 4.14801091e-01
7.12716103e-01 -2.95252174e-01 -9.93851840e-01 -6.75316930e-01
5.44399977e-01 2.48631805e-01 2.14005455e-01 -4.75889444e-01
-1.10005051e-01 -2.46859297e-01 -9.95266959e-02 -4.16453451e-01
1.29285306e-01 -4.34003472e-01 9.19080600e-02 3.52161527e-01
-6.26104593e-01 4.14649993e-01 2.22037658e-02 9.34078574e-01
-2.23002091e-01 -1.67464629e-01 3.09635192e-01 -5.76323867e-01
-8.04861069e-01 5.72770000e-01 -1.84603289e-01 4.44505215e-01
5.11070192e-01 7.32036531e-02 -4.21545208e-01 -1.06536579e+00
-5.54838777e-01 2.36260399e-01 1.28584772e-01 7.55434036e-01
8.70053589e-01 -1.45974898e+00 -1.00297606e+00 -2.73575261e-02
4.21323478e-01 2.19087094e-01 1.07542671e-01 2.03362212e-01
-3.28173339e-01 3.28824371e-01 1.99251965e-01 -2.77771890e-01
-1.04191780e+00 8.96579087e-01 1.42169610e-01 -6.75898850e-01
-9.13538992e-01 7.36756742e-01 5.71860790e-01 -5.18650651e-01
-1.00522831e-01 -4.32841510e-01 -3.95286858e-01 -4.37060654e-01
5.35787046e-01 8.62521008e-02 -1.98867604e-01 -8.33551705e-01
-1.21892206e-02 3.46944511e-01 -2.52696216e-01 2.35775754e-01
1.02983081e+00 -5.62336035e-02 -3.79465550e-01 3.04527313e-01
1.14129877e+00 -1.70198113e-01 -6.31520271e-01 -2.85429031e-01
2.01331094e-01 -3.00190657e-01 -1.80132464e-01 -3.95985872e-01
-9.51951683e-01 1.19540870e+00 -2.67299235e-01 2.92333007e-01
8.44238460e-01 4.71261367e-02 9.78385866e-01 4.75762784e-01
1.80416510e-01 -5.62905312e-01 4.10753071e-01 7.33643770e-01
1.24597013e+00 -7.91886270e-01 -1.42881140e-01 -9.28587973e-01
-7.67914295e-01 1.07973683e+00 7.54638970e-01 2.59435587e-02
2.07708940e-01 -4.78888266e-02 -5.05608022e-01 -2.98377305e-01
-8.09478581e-01 -5.42076886e-01 6.02485657e-01 6.64271414e-01
5.95146537e-01 1.57831863e-01 -3.18943143e-01 6.48562491e-01
-1.94346592e-01 -1.95287481e-01 7.08847523e-01 5.93617082e-01
-2.84164310e-01 -1.09265709e+00 3.16200733e-01 5.87597489e-01
-4.34600592e-01 -6.29922867e-01 -7.51201391e-01 5.13180375e-01
-6.52502537e-01 9.69099879e-01 -7.91778672e-04 -3.92948955e-01
2.33771563e-01 3.91675048e-02 4.74673092e-01 -1.06642497e+00
-2.52313435e-01 -2.14789554e-01 6.49526596e-01 -3.60263795e-01
-2.64420230e-02 9.07889456e-02 -1.66645133e+00 -1.95133403e-01
-8.76409709e-02 1.55571625e-01 2.30334044e-01 6.52992845e-01
7.28368342e-01 1.08016205e+00 2.99879879e-01 -4.71973032e-01
-3.55001092e-01 -1.12567210e+00 -5.73725522e-01 7.70704389e-01
5.67573905e-02 -1.97156012e-01 -7.55354166e-02 -6.13955893e-02] | [10.299758911132812, 8.3108549118042] |
ded1fb3c-10c0-4471-acd2-34f820a7647b | cross-modal-interaction-networks-for-query | 1906.02497 | null | https://arxiv.org/abs/1906.02497v2 | https://arxiv.org/pdf/1906.02497v2.pdf | Cross-Modal Interaction Networks for Query-Based Moment Retrieval in Videos | Query-based moment retrieval aims to localize the most relevant moment in an untrimmed video according to the given natural language query. Existing works often only focus on one aspect of this emerging task, such as the query representation learning, video context modeling or multi-modal fusion, thus fail to develop a comprehensive system for further performance improvement. In this paper, we introduce a novel Cross-Modal Interaction Network (CMIN) to consider multiple crucial factors for this challenging task, including (1) the syntactic structure of natural language queries; (2) long-range semantic dependencies in video context and (3) the sufficient cross-modal interaction. Specifically, we devise a syntactic GCN to leverage the syntactic structure of queries for fine-grained representation learning, propose a multi-head self-attention to capture long-range semantic dependencies from video context, and next employ a multi-stage cross-modal interaction to explore the potential relations of video and query contents. The extensive experiments demonstrate the effectiveness of our proposed method. | ['Zhenxin Xiao', 'Zhou Zhao', 'Zhijie Lin', 'Zhu Zhang'] | 2019-06-06 | null | null | null | null | ['moment-retrieval'] | ['computer-vision'] | [ 5.00031784e-02 -6.10152781e-01 -4.90588665e-01 -3.65569025e-01
-1.02780378e+00 -4.32291120e-01 6.98255062e-01 2.64401808e-02
-4.04863626e-01 1.44681916e-01 8.29273939e-01 7.44419098e-02
-3.09337109e-01 -3.41407180e-01 -6.33038402e-01 -5.30694485e-01
-3.36600579e-02 1.33821201e-02 5.33636689e-01 -2.33549803e-01
3.18922371e-01 1.13821626e-01 -1.62196612e+00 4.61451679e-01
5.95992506e-01 1.12526453e+00 3.49358976e-01 4.76409018e-01
-2.44661227e-01 1.04905331e+00 -4.05066729e-01 -7.53757060e-02
-2.46116281e-01 -5.70837498e-01 -7.72751451e-01 7.53971608e-03
2.11884618e-01 -3.48251045e-01 -5.13899803e-01 9.64746714e-01
5.66770256e-01 4.20315057e-01 4.95394766e-01 -1.20948195e+00
-5.11950970e-01 4.85452294e-01 -5.27474999e-01 5.77983677e-01
7.33192205e-01 6.89913854e-02 1.14230418e+00 -7.60468483e-01
6.80898011e-01 1.50666463e+00 -2.81684659e-02 2.18846664e-01
-6.13693237e-01 -4.88559932e-01 5.75740099e-01 5.81423879e-01
-1.60817420e+00 -3.58845174e-01 1.21568668e+00 -3.11634868e-01
7.32868612e-01 2.57257074e-01 4.62225020e-01 1.07180035e+00
1.18337758e-01 1.11094165e+00 6.51803613e-01 -2.81213373e-01
-1.22569740e-01 -1.65574893e-01 2.20694527e-01 6.29948139e-01
-4.96558845e-01 -2.23000124e-01 -6.63344145e-01 -7.01264068e-02
7.26748705e-01 2.17606187e-01 -2.27936611e-01 -1.48244426e-01
-1.45235252e+00 7.76342213e-01 3.89115274e-01 7.13083684e-01
-3.49491984e-01 2.12503970e-01 7.57636189e-01 2.25467622e-01
3.23590398e-01 1.30123153e-01 -4.04258788e-01 -9.83524695e-02
-7.25074708e-01 1.99811399e-01 5.42475402e-01 1.13153207e+00
8.67603958e-01 -3.66465509e-01 -4.79034603e-01 7.89265335e-01
5.12458086e-01 3.60939384e-01 6.55880392e-01 -7.13255644e-01
5.83704233e-01 7.05672204e-01 -2.34955728e-01 -1.39676023e+00
-2.56873339e-01 -7.81254377e-03 -4.34838057e-01 -7.10819185e-01
-2.46874452e-01 -2.42326744e-02 -6.32681847e-01 1.97379315e+00
3.59562844e-01 4.57734793e-01 9.62355360e-02 1.08428824e+00
1.15933549e+00 7.46357441e-01 4.64959860e-01 -3.69744837e-01
1.59062684e+00 -1.11373889e+00 -7.52036452e-01 -1.97715908e-01
6.40133142e-01 -7.99379349e-01 1.02426577e+00 -2.20982641e-01
-9.65348601e-01 -7.24317551e-01 -7.84755647e-01 -4.67893273e-01
-5.00315189e-01 -1.51586551e-02 5.64886510e-01 -1.30978480e-01
-7.35306680e-01 -1.31362751e-01 -5.42469859e-01 -5.09272814e-01
-9.81914218e-06 3.00688505e-01 -3.86688709e-01 -4.24272746e-01
-1.52583420e+00 3.75046134e-01 6.42420352e-01 1.62659615e-01
-8.05969834e-01 -3.00874531e-01 -9.61728215e-01 -3.09785455e-02
9.92153943e-01 -8.06752741e-01 1.00236142e+00 -8.85143757e-01
-1.04458022e+00 7.02690780e-01 -5.64565003e-01 -6.18830882e-02
1.04455493e-01 -4.47028756e-01 -4.48272526e-01 5.37888527e-01
2.23845735e-01 7.33739614e-01 6.60401821e-01 -1.04291582e+00
-8.40236306e-01 -5.76352954e-01 5.04472792e-01 6.79915845e-01
-2.07080618e-01 3.65300804e-01 -1.35050118e+00 -8.18226099e-01
5.61094433e-02 -7.24581659e-01 -3.59265693e-02 -4.68506902e-01
-3.03116173e-01 -6.23095036e-01 9.83958542e-01 -6.03721797e-01
1.68975914e+00 -2.26286364e+00 4.02538776e-01 -7.75610134e-02
-2.73615252e-02 -4.27106060e-02 -2.70168453e-01 5.37194192e-01
1.26836300e-01 2.63477955e-02 2.27026537e-01 -2.78626740e-01
-4.14472595e-02 1.89524636e-01 -3.42085242e-01 1.31072059e-01
9.90677997e-02 1.21421564e+00 -9.23771262e-01 -9.85342622e-01
2.35080689e-01 5.94908357e-01 -3.94343704e-01 4.87728298e-01
-4.57251161e-01 4.42245930e-01 -9.49770510e-01 8.26074541e-01
1.31907403e-01 -3.45150471e-01 -1.52520537e-01 -8.49539280e-01
9.55847874e-02 -3.97370346e-02 -8.77113223e-01 2.24007344e+00
-4.34805661e-01 2.66088277e-01 -8.24142769e-02 -9.12515461e-01
5.97844481e-01 3.08306098e-01 7.63244927e-01 -9.17803884e-01
1.25662491e-01 -2.16753948e-02 -2.99862683e-01 -1.16166580e+00
4.03838515e-01 3.27651724e-02 -3.77313107e-01 3.16640794e-01
1.89673945e-01 4.30458516e-01 2.34305114e-01 4.42572057e-01
8.36889863e-01 2.58681059e-01 2.26631999e-01 2.92064548e-02
8.37874174e-01 -3.74785900e-01 4.87310052e-01 4.78967100e-01
-3.12767655e-01 4.51724231e-01 4.60344255e-01 -1.05579801e-01
-3.75516623e-01 -4.40573364e-01 3.84875327e-01 1.75355017e+00
7.05371737e-01 -5.63014269e-01 -5.06079614e-01 -7.33661592e-01
-2.75441647e-01 3.25663775e-01 -5.34054577e-01 -4.02845860e-01
-5.09459615e-01 -4.55492079e-01 3.43681812e-01 5.63552797e-01
5.99860787e-01 -1.21450269e+00 -3.97328496e-01 4.65781279e-02
-6.90617859e-01 -1.44000542e+00 -1.04409075e+00 -3.43052059e-01
-5.39539039e-01 -1.16649735e+00 -5.64120471e-01 -9.59759295e-01
2.04043791e-01 5.34910858e-01 1.04026246e+00 2.15520263e-01
-6.37795702e-02 8.68069768e-01 -6.37701094e-01 8.84450451e-02
2.56658047e-01 1.34436399e-01 -2.07153112e-01 2.86530435e-01
7.15788484e-01 -4.74266410e-01 -7.58744299e-01 3.90061378e-01
-1.20389557e+00 -5.97709902e-02 7.43712544e-01 6.02336407e-01
7.48501003e-01 3.22890542e-02 5.75266719e-01 -6.47311449e-01
6.09807611e-01 -7.66674280e-01 -5.64766750e-02 6.25345767e-01
-9.47000608e-02 -1.92809757e-02 1.81295991e-01 -5.41104257e-01
-1.09014416e+00 -2.68691909e-02 -7.41427913e-02 -7.79169202e-01
-2.73885757e-01 9.06543314e-01 -6.56488836e-01 1.47622824e-01
5.40291145e-02 5.29809415e-01 -4.59244579e-01 -3.70933831e-01
4.98367667e-01 4.59144861e-01 6.74836695e-01 -7.33795047e-01
4.66311932e-01 4.60107088e-01 -8.14492851e-02 -6.90949321e-01
-9.06820178e-01 -9.44871783e-01 -6.32895172e-01 -2.72747546e-01
1.31895375e+00 -1.16777420e+00 -8.17348003e-01 1.12378970e-01
-1.10293090e+00 1.35927677e-01 2.76841909e-01 6.26915693e-01
-6.58568144e-01 6.86290503e-01 -5.14395654e-01 -6.80701554e-01
-3.10961902e-01 -1.25371718e+00 1.53942931e+00 2.91855603e-01
1.73496351e-01 -1.05513430e+00 -2.45169532e-02 5.83472669e-01
1.24969892e-01 1.44961014e-01 9.18169916e-01 -7.72147834e-01
-7.20805883e-01 -1.37934402e-01 -4.95604217e-01 -4.77059484e-02
9.26866159e-02 -1.48597866e-01 -6.79113388e-01 -9.57980454e-02
1.27378091e-01 -4.48838085e-01 8.69977415e-01 2.84314126e-01
9.93074715e-01 -2.35296890e-01 -4.99437749e-01 6.10405326e-01
1.23610306e+00 1.96087822e-01 4.00985211e-01 8.68033618e-02
9.66269612e-01 7.30337918e-01 9.55934525e-01 4.11578208e-01
7.36499965e-01 6.56159103e-01 3.66603702e-01 1.71619862e-01
1.60548866e-01 -4.47439998e-01 3.11374336e-01 1.03281021e+00
1.89343672e-02 -2.38090441e-01 -6.52872384e-01 4.27438945e-01
-2.15244484e+00 -1.09877932e+00 4.39033151e-01 1.80604005e+00
4.42625254e-01 2.00374760e-02 4.22936045e-02 -4.22719330e-01
7.75796711e-01 5.32851040e-01 -6.52262390e-01 3.56889546e-01
-1.03512563e-01 -6.17651522e-01 -4.31956500e-02 3.11371028e-01
-1.28758645e+00 1.06970847e+00 5.31090736e+00 1.04232860e+00
-1.06015587e+00 1.11097068e-01 3.57234806e-01 -3.40722241e-02
-4.64833885e-01 9.29552168e-02 -8.02583098e-01 5.83971262e-01
6.44547522e-01 -1.58647537e-01 3.15346718e-01 6.19017005e-01
2.57988811e-01 -6.08717604e-03 -1.13450217e+00 1.26096094e+00
6.88417137e-01 -1.00679123e+00 3.52591872e-01 -1.27559960e-01
3.90673071e-01 -7.89258629e-03 -1.80197597e-01 6.11475408e-01
-3.68273824e-01 -5.71212053e-01 5.49299955e-01 8.88242722e-01
5.90602040e-01 -7.54596770e-01 6.13193452e-01 4.47937608e-01
-1.69342911e+00 -1.23273246e-01 -1.34204132e-02 4.28849787e-01
3.20997953e-01 -1.71070024e-02 -7.39692003e-02 8.41110885e-01
5.81568301e-01 1.00107193e+00 -7.64036179e-01 8.26023161e-01
1.08479232e-01 2.70917773e-01 -3.11564237e-01 1.17546260e-01
4.49315041e-01 3.06055658e-02 4.49412018e-01 1.33165443e+00
1.14464067e-01 3.43677044e-01 5.90676963e-01 4.22377199e-01
-3.72526422e-02 3.77447248e-01 -4.63567197e-01 -2.77836502e-01
4.27642226e-01 1.04937148e+00 -6.76375329e-01 -3.74091148e-01
-8.43602479e-01 9.06854153e-01 1.45812556e-01 7.08287060e-01
-9.92092848e-01 -3.05686682e-01 4.03018385e-01 -2.02356413e-01
3.05565387e-01 -1.78168267e-01 4.69705462e-01 -1.48271608e+00
2.23491564e-01 -7.82142878e-01 7.88586795e-01 -9.01121438e-01
-1.26470947e+00 5.68688750e-01 2.69523829e-01 -1.23310232e+00
-4.44134086e-01 -1.42665550e-01 -5.77342689e-01 6.32103801e-01
-1.66530883e+00 -1.62182617e+00 -3.94655615e-01 1.09757984e+00
8.16737950e-01 9.49234609e-03 5.35641968e-01 6.16828501e-01
-5.54105759e-01 4.42669183e-01 -2.93256044e-01 1.47095397e-01
7.80321658e-01 -6.80159867e-01 -2.43261918e-01 7.20793962e-01
1.94068238e-01 8.16297948e-01 3.47324044e-01 -6.14261687e-01
-1.78286600e+00 -1.01793969e+00 7.29306698e-01 -4.03375417e-01
6.29622281e-01 -1.83327496e-01 -8.99755538e-01 8.22125793e-01
1.14997417e-01 1.00437552e-01 7.82385468e-01 1.04371414e-01
-5.30869782e-01 -1.56305000e-01 -4.98945922e-01 4.87366915e-01
1.04964912e+00 -1.09496939e+00 -7.19308436e-01 2.87044406e-01
1.33545363e+00 -2.41440147e-01 -7.94208586e-01 5.87503195e-01
6.09818459e-01 -8.50049496e-01 1.03387332e+00 -7.25484908e-01
4.22902316e-01 -3.79943699e-01 -4.73974466e-01 -5.38243711e-01
-1.88954085e-01 -4.33095962e-01 -3.17612201e-01 1.54995787e+00
-7.51688406e-02 3.94530520e-02 4.74130690e-01 6.19574845e-01
-9.01665390e-02 -8.97038281e-01 -7.57034302e-01 -3.73500824e-01
-3.50117475e-01 -6.63885593e-01 5.57458162e-01 8.65513921e-01
5.27936555e-02 8.15268219e-01 -4.61031049e-01 2.13538423e-01
2.16900796e-01 4.25567180e-01 6.96769953e-01 -8.40125203e-01
-1.89735442e-01 -3.88467550e-01 -5.44462502e-01 -1.53389263e+00
3.95608008e-01 -5.13336420e-01 4.14496958e-02 -1.42109561e+00
5.90726674e-01 -2.83438489e-02 -6.19623005e-01 9.41352844e-02
-4.35759097e-01 -6.84744492e-02 2.54305899e-01 4.79249567e-01
-1.56694722e+00 6.58872128e-01 1.09773242e+00 -6.40749037e-02
-1.35448053e-01 -2.93543905e-01 -7.96573818e-01 6.56985223e-01
1.84919968e-01 -1.20143384e-01 -9.16408956e-01 -6.98973179e-01
2.08997652e-01 5.49966931e-01 3.98822606e-01 -8.78979445e-01
5.12181938e-01 -3.56817722e-01 1.39347285e-01 -8.01916301e-01
4.28074956e-01 -9.69066739e-01 -1.64614931e-01 -1.22632317e-01
-5.13504088e-01 1.61803499e-01 -3.79693769e-02 9.98013794e-01
-7.72773147e-01 1.04787210e-02 1.67411491e-01 -1.98126376e-01
-1.26749492e+00 6.07399881e-01 -1.11205718e-02 2.15086892e-01
1.03934062e+00 8.29478726e-02 -2.08980888e-01 -5.18033564e-01
-7.99201131e-01 6.03510618e-01 3.21142703e-01 8.86793017e-01
5.00441730e-01 -1.46452558e+00 -3.09831113e-01 -3.03697307e-02
5.32661200e-01 -1.31265789e-01 7.40094721e-01 7.29832768e-01
6.54063001e-02 6.34945095e-01 1.83837593e-01 -6.64530873e-01
-1.25876868e+00 9.57127273e-01 8.93598795e-02 -3.17336380e-01
-3.25801343e-01 8.41209769e-01 6.51806593e-01 7.89444223e-02
3.73063087e-01 -2.79227436e-01 -6.03272021e-01 2.34336480e-01
5.84931076e-01 -2.94688549e-02 -4.02509004e-01 -1.27666605e+00
-4.09431785e-01 9.13768888e-01 -1.11821294e-01 -2.34571714e-02
8.68277550e-01 -6.98038757e-01 -2.42185667e-01 6.41241074e-01
1.67854357e+00 -3.42636764e-01 -1.03107131e+00 -5.71641743e-01
-5.91129810e-02 -4.55508411e-01 -2.95762476e-02 -4.01289940e-01
-1.03613281e+00 8.62667918e-01 3.69259655e-01 4.40430306e-02
1.49072587e+00 4.34601784e-01 9.39800322e-01 5.06778955e-01
2.32215777e-01 -8.38696063e-01 3.34864229e-01 4.52598721e-01
8.54334652e-01 -1.40103328e+00 -5.68727925e-02 -3.02778453e-01
-6.43061340e-01 8.92567754e-01 8.04107249e-01 1.22269861e-01
8.08687985e-01 -2.62850791e-01 1.32615035e-02 -4.21449274e-01
-8.79734159e-01 -5.19998193e-01 6.01588905e-01 3.00124168e-01
4.31739807e-01 -3.98315579e-01 -2.07550868e-01 7.57094979e-01
3.49715382e-01 1.07362695e-01 -2.28075817e-01 1.04239142e+00
-3.57758135e-01 -7.62969911e-01 -1.08860612e-01 2.21202061e-01
-5.00010610e-01 2.74339109e-03 -2.82580525e-01 5.42808771e-01
1.87412173e-01 9.03372765e-01 2.54711825e-02 -5.04121959e-01
3.76204044e-01 1.29097581e-01 2.01843739e-01 -4.65345860e-01
-3.96483928e-01 6.34776711e-01 -3.33290696e-01 -8.46840560e-01
-9.63317335e-01 -4.93271470e-01 -1.20158923e+00 1.56376928e-01
-2.67357647e-01 2.06017375e-01 3.33962053e-01 1.27039289e+00
6.48149371e-01 4.62007105e-01 5.04645288e-01 -7.22003579e-01
-2.79913135e-02 -9.30360734e-01 -3.99791926e-01 7.41581023e-01
2.80280203e-01 -6.86983526e-01 -2.04448938e-01 2.07272768e-01] | [10.308823585510254, 0.9614923000335693] |
a82464fe-e311-4fcb-90fc-341edeff8c97 | epution-at-semeval-2018-task-2-emoji | null | null | https://aclanthology.org/S18-1071 | https://aclanthology.org/S18-1071.pdf | EPUTION at SemEval-2018 Task 2: Emoji Prediction with User Adaption | This paper describes our approach, called EPUTION, for the open trial of the SemEval- 2018 Task 2, Multilingual Emoji Prediction. The task relates to using social media {---} more precisely, Twitter {---} with its aim to predict the most likely associated emoji of a tweet. Our solution for this text classification problem explores the idea of transfer learning for adapting the classifier based on users{'} tweeting history. Our experiments show that our user-adaption method improves classification results by more than 6 per cent on the macro-averaged F1. Thus, our paper provides evidence for the rationality of enriching the original corpus longitudinally with user behaviors and transferring the lessons learned from corresponding users to specific instances. | ['Liyuan Zhou', 'Tom Gedeon', 'Hanna Suominen', 'Qiongkai Xu'] | 2018-06-01 | null | null | null | semeval-2018-6 | ['twitter-sentiment-analysis'] | ['natural-language-processing'] | [-2.57338136e-01 1.89731762e-01 -1.35864317e-01 -5.62143266e-01
-3.52863520e-01 -1.30148515e-01 5.56616485e-01 1.41140342e-01
-6.62562490e-01 8.03801477e-01 3.28241318e-01 -3.01835895e-01
2.54865587e-01 -5.76742530e-01 -6.25237823e-01 -1.09204493e-01
-1.06764838e-01 4.50128078e-01 4.31372561e-02 -3.33851218e-01
-7.16455579e-02 -7.51875639e-02 -1.29315186e+00 4.16728050e-01
9.34241593e-01 8.89271677e-01 -7.17063919e-02 8.46481323e-01
-2.96334270e-03 9.57662284e-01 -3.41594964e-01 -1.12531221e+00
-2.11284533e-01 -4.49733913e-01 -1.00761342e+00 -2.36277714e-01
2.56715477e-01 -1.24733008e-01 -3.27571094e-01 8.05049479e-01
3.29234600e-01 3.41079980e-01 1.00466549e+00 -9.44484472e-01
-9.19881523e-01 8.76738489e-01 -3.75387579e-01 5.04146099e-01
2.82458544e-01 -2.20700368e-01 9.24957097e-01 -7.74870813e-01
6.42001748e-01 9.18563128e-01 1.08106303e+00 7.80751407e-01
-1.18834913e+00 -5.46313405e-01 -1.07709378e-01 4.65439230e-01
-1.04948449e+00 -4.81730849e-01 8.00823271e-01 -7.37251997e-01
9.19456542e-01 5.21669626e-01 2.58825988e-01 1.60686994e+00
-9.73046422e-02 1.05845499e+00 1.19036102e+00 -5.31600893e-01
-8.20453614e-02 7.30728447e-01 7.39639401e-01 7.58766055e-01
-1.70532525e-01 -2.70034730e-01 -4.46702391e-01 -4.05122936e-01
-2.58535057e-01 -2.56782323e-01 3.48382324e-01 5.38464546e-01
-4.27446723e-01 9.64786530e-01 9.51715559e-02 6.79042518e-01
-2.84690112e-01 -2.03222170e-01 7.67668664e-01 4.83621448e-01
1.33531952e+00 6.47965848e-01 -1.16355443e+00 -5.14826953e-01
-5.18246531e-01 -1.18854672e-01 1.21642530e+00 3.96251172e-01
8.47678423e-01 -1.09776713e-01 -3.85711044e-01 1.24163687e+00
-6.43260330e-02 3.22485924e-01 6.64426148e-01 -5.88757396e-01
2.35835239e-01 3.18722844e-01 1.04247041e-01 -8.57168257e-01
-5.81384540e-01 -1.63593471e-01 -2.83373177e-01 -5.64950883e-01
4.07103688e-01 -1.09717238e+00 -2.62418836e-01 1.71253741e+00
2.79619217e-01 4.82771635e-01 -4.02191460e-01 7.59540126e-02
6.83472514e-01 4.05078828e-01 7.51355946e-01 -3.10581475e-02
1.27146053e+00 -1.04725814e+00 -5.91986001e-01 -7.18264282e-02
1.19351113e+00 -6.03480458e-01 1.34657061e+00 -1.89459622e-02
-8.37370276e-01 -3.83706272e-01 -4.95752752e-01 4.96733524e-02
-8.39945316e-01 2.69340605e-01 8.20965350e-01 1.01573408e+00
-8.41009915e-01 7.94412553e-01 -4.80159819e-01 -6.89414561e-01
3.07825953e-01 3.88858855e-01 -1.91028148e-01 3.69001389e-01
-1.44905984e+00 1.23506904e+00 -2.43696962e-02 -3.74450266e-01
1.56637549e-01 -9.37954247e-01 -6.13966465e-01 1.76054060e-01
-4.72475179e-02 -4.69696730e-01 1.35267889e+00 -1.36599195e+00
-2.10234022e+00 1.12759173e+00 -2.43640676e-01 -4.82654214e-01
4.66679573e-01 -2.93792427e-01 -8.23111773e-01 -3.23199600e-01
1.04725808e-01 -9.99855921e-02 6.79836750e-01 -9.66323733e-01
-7.19001889e-01 -2.17845991e-01 -2.83078432e-01 -3.27690423e-01
-1.16756475e+00 3.74461949e-01 -2.42644057e-01 -4.35041279e-01
-1.16930258e+00 -1.03374052e+00 8.56917724e-02 -1.00573766e+00
-3.00263673e-01 -7.63159394e-01 5.87483525e-01 -1.23676014e+00
1.67639017e+00 -2.05900049e+00 1.49074242e-01 2.32627824e-01
1.11580692e-01 3.78611743e-01 -4.57139090e-02 2.97431350e-01
8.43235105e-02 3.14540803e-01 1.01350635e-01 -8.56802881e-01
-4.69887033e-02 -1.82916746e-01 -1.55458316e-01 1.83705449e-01
-1.86025754e-01 1.00690579e+00 -1.03109682e+00 -2.74335146e-01
9.01449397e-02 2.29360417e-01 -6.91291273e-01 4.08353418e-01
6.92040697e-02 5.74905396e-01 -3.73993993e-01 -1.19538888e-01
1.80681065e-01 -1.63134634e-01 2.44521841e-01 -2.94026472e-02
1.78357530e-02 2.28458628e-01 -1.86433583e-01 9.68821764e-01
-9.71838951e-01 8.76811981e-01 -9.35382769e-02 -8.00114691e-01
6.55837178e-01 2.22352207e-01 6.46891713e-01 -7.21781552e-01
4.89284992e-01 9.85830128e-02 1.18720820e-02 -8.82629395e-01
1.98226929e-01 8.66786167e-02 -3.09754550e-01 3.84462088e-01
2.92810321e-01 5.47434926e-01 1.41615883e-01 7.32663646e-02
1.15226543e+00 -4.41714227e-02 1.64725423e-01 -3.64569902e-01
5.99518955e-01 -4.08574909e-01 1.12171151e-01 1.00112498e+00
-1.37958676e-01 -4.45757471e-02 3.72877777e-01 -2.49513000e-01
-9.12560165e-01 -6.11602485e-01 -1.98404342e-01 1.86479497e+00
-6.26103461e-01 -4.44834054e-01 -1.04120409e+00 -1.18586135e+00
8.08456391e-02 1.15347743e+00 -1.26827097e+00 -4.36789960e-01
-3.50380301e-01 -8.48108113e-01 5.39449632e-01 2.74137825e-01
3.06116313e-01 -1.01874077e+00 3.32754329e-02 1.57863662e-01
-2.30234891e-01 -1.05830336e+00 -4.97557223e-01 -1.31301731e-01
-2.98742384e-01 -8.19746733e-01 -4.86974210e-01 -6.34031832e-01
4.08437103e-01 -2.85951346e-01 1.24188054e+00 4.11594212e-01
-5.15833050e-02 8.01112592e-01 -5.83324790e-01 -4.51554567e-01
-4.26021248e-01 7.02192426e-01 1.59803778e-01 6.59567237e-01
9.02839720e-01 -7.88260400e-01 -3.00147086e-01 5.29352017e-02
-2.55611449e-01 -1.66945353e-01 1.56168371e-01 7.56770074e-01
-4.43535894e-01 -5.86278021e-01 8.79049778e-01 -1.66844738e+00
7.34054983e-01 -1.08197618e+00 1.13637224e-01 3.61867845e-01
-6.28048778e-01 -3.16549838e-01 8.00438762e-01 -6.49730384e-01
-1.42940485e+00 -1.20125346e-01 -4.59004462e-01 1.21032126e-01
-3.76795858e-01 4.91218030e-01 3.12497854e-01 7.72418976e-02
7.31358826e-01 -9.64448527e-02 -3.67647558e-01 -7.25727975e-01
3.73686820e-01 1.04504418e+00 8.01847354e-02 -6.46765709e-01
3.37771654e-01 3.30843404e-02 -4.87555712e-01 -1.01855862e+00
-1.32441115e+00 -4.88126338e-01 -7.11207032e-01 -3.30270499e-01
1.09087646e+00 -6.40159845e-01 -1.20162320e+00 4.33295667e-01
-9.95653927e-01 -9.64871705e-01 1.49782240e-01 2.48813704e-01
-4.04659122e-01 1.76758125e-01 -1.19760525e+00 -9.41151261e-01
-4.36435193e-01 -5.87578833e-01 6.17391348e-01 1.79348022e-01
-7.39795446e-01 -1.57634938e+00 4.19031978e-01 4.84076083e-01
7.37796783e-01 -1.36563778e-01 1.01383567e+00 -1.39514124e+00
6.64163053e-01 -5.19640207e-01 4.21402539e-04 3.94456923e-01
7.98196122e-02 5.30313291e-02 -1.13192058e+00 1.70756623e-01
-2.62345225e-01 -4.21878994e-01 6.50753319e-01 -4.37134579e-02
1.62188220e+00 -4.54121351e-01 -3.20895284e-01 4.26326059e-02
7.93287635e-01 -1.11646712e-01 3.21841747e-01 1.30629778e-01
7.49215722e-01 6.89409912e-01 8.83845165e-02 5.83160639e-01
8.58690441e-01 7.80827641e-01 -2.78396219e-01 1.06470831e-01
-8.95507410e-02 -3.24173689e-01 3.52695584e-01 1.03628314e+00
-1.48961231e-01 -1.11277744e-01 -7.41294146e-01 3.72031778e-01
-1.89275491e+00 -9.43422794e-01 -3.03926885e-01 1.94494033e+00
1.09615040e+00 -1.89588711e-01 5.87533176e-01 -4.51846451e-01
7.08245456e-01 -2.06073791e-01 -2.10274756e-01 -5.97620606e-01
3.16349298e-01 4.28594142e-01 4.11584258e-01 9.47897613e-01
-9.28629994e-01 1.10883665e+00 6.13682604e+00 7.60710835e-01
-1.06173730e+00 6.86976969e-01 1.01738405e+00 2.39149407e-02
-7.85958543e-02 -3.37574333e-01 -8.65068614e-01 9.91173327e-01
1.69076788e+00 -1.05706714e-01 7.40039289e-01 6.58113658e-01
-9.84668508e-02 -8.68040249e-02 -1.12915337e+00 5.90017855e-01
3.00896764e-01 -1.05702865e+00 -2.77640283e-01 -1.11903235e-01
8.49868774e-01 2.21963942e-01 2.80382067e-01 1.21023703e+00
5.23631871e-01 -5.87796926e-01 3.14255476e-01 9.07050073e-01
4.68588948e-01 -5.16279280e-01 5.97849131e-01 4.51439112e-01
-4.54698384e-01 -1.35491014e-01 7.20527081e-04 -1.81030974e-01
-6.49498850e-02 4.67051357e-01 -9.78540361e-01 6.06703684e-02
6.69060886e-01 8.60086858e-01 -1.01334679e+00 6.71681762e-01
4.83481251e-02 1.41016948e+00 -1.06063224e-01 -4.07340974e-01
-1.16366260e-01 -2.00617686e-01 1.99544519e-01 1.70394874e+00
1.54584721e-01 -1.61835514e-02 -1.20141573e-01 2.36167654e-01
-5.10171950e-01 5.30620039e-01 -6.56806886e-01 -1.36059925e-01
2.17952311e-01 1.40790772e+00 -2.86362052e-01 -4.64881063e-01
-6.37267292e-01 1.28655195e+00 1.14579117e+00 3.95044059e-01
-7.37068117e-01 -5.03282130e-01 3.58431488e-01 -5.90425543e-02
2.79361248e-01 6.68058246e-02 -7.24831522e-01 -1.30011916e+00
-5.91076732e-01 -4.79120314e-01 4.79006708e-01 -5.97675264e-01
-1.97416794e+00 4.97185975e-01 -3.83820176e-01 -4.36636537e-01
-3.03817660e-01 -6.36663735e-01 -9.06514049e-01 7.82479048e-01
-1.13455927e+00 -1.08167589e+00 1.85258389e-01 4.98255074e-01
4.28109258e-01 -3.00180346e-01 1.07655215e+00 6.82114065e-01
-8.57379138e-01 1.03117406e+00 3.13632548e-01 3.43604267e-01
8.71915162e-01 -1.38792336e+00 2.39651963e-01 2.16424912e-01
1.03122080e-02 4.23554361e-01 9.73995686e-01 -6.60952091e-01
-9.98324215e-01 -9.34679985e-01 1.53085411e+00 -1.24359059e+00
1.34420693e+00 -5.28631628e-01 -7.15105295e-01 1.02512252e+00
4.44810539e-01 -2.33535960e-01 1.31482065e+00 9.12031233e-01
-3.45013976e-01 5.68457274e-03 -1.17691910e+00 8.28627050e-01
8.91759872e-01 -9.05656993e-01 -3.44767332e-01 6.04902685e-01
5.39607167e-01 2.78150551e-02 -1.21426666e+00 9.67907533e-02
4.05593067e-01 -8.06012094e-01 5.83162665e-01 -1.33755004e+00
6.97030425e-01 7.49690294e-01 6.72294348e-02 -1.66666377e+00
-4.39364821e-01 -5.99580526e-01 -3.87556255e-01 1.56517804e+00
7.38080204e-01 -6.07632041e-01 6.47367239e-01 1.04012609e+00
7.95734376e-02 -5.33135116e-01 -8.16686809e-01 -2.08489403e-01
2.77836800e-01 -4.53066766e-01 1.78609833e-01 1.43876517e+00
3.65079969e-01 8.49775732e-01 -8.49321723e-01 -1.77926779e-01
2.23115340e-01 -4.59778219e-01 8.04460406e-01 -1.35756266e+00
-5.63603878e-01 -5.88340640e-01 1.05148315e-01 -5.51968575e-01
9.36789632e-01 -1.02285719e+00 -2.72417814e-01 -9.04744327e-01
3.49299878e-01 -2.92634428e-01 -2.29574710e-01 3.61129910e-01
-6.00628316e-01 3.55486095e-01 1.72354251e-01 -3.68855834e-01
-1.01100338e+00 4.69375998e-01 7.41154432e-01 -1.71642881e-02
-2.75758713e-01 3.88827175e-01 -5.83505571e-01 7.05796719e-01
7.41555274e-01 -4.08585668e-01 1.26955137e-01 -2.33115405e-01
5.36906183e-01 -9.65972021e-02 -3.70179676e-02 -5.43658912e-01
-6.08769804e-02 1.37655720e-01 1.76031128e-01 1.79610118e-01
3.04679126e-01 -8.89630497e-01 -2.28510499e-01 1.93595856e-01
-7.29677260e-01 -2.53449023e-01 2.48512045e-01 5.36033273e-01
3.59482020e-01 -5.26067495e-01 4.57526177e-01 4.54148985e-02
-3.95947576e-01 4.26012874e-01 -6.79131866e-01 1.86912075e-01
7.54973888e-01 4.79339987e-01 -2.53669888e-01 -5.43994725e-01
-1.31471002e+00 7.11853579e-02 -1.03798158e-01 6.14717066e-01
-3.48697096e-01 -1.07650721e+00 -5.95094919e-01 -1.47420377e-01
1.77048102e-01 -1.30889201e+00 3.60691249e-01 1.00585055e+00
2.94295788e-01 2.80811009e-03 4.83655073e-02 5.37113985e-03
-1.16228759e+00 2.57031828e-01 4.38586920e-01 -3.87567043e-01
-1.29976749e-01 8.79134417e-01 -2.81029046e-01 -7.73155391e-01
-8.07659999e-02 4.68128681e-01 -4.93487239e-01 2.24267900e-01
5.32983840e-01 7.71388471e-01 1.01271696e-01 -6.96207166e-01
6.18320219e-02 -1.23296157e-01 -3.35945368e-01 -5.19032218e-02
1.45453560e+00 -3.02434802e-01 -3.18079501e-01 9.63407636e-01
1.58255064e+00 3.87343287e-01 -8.42304647e-01 -4.22451586e-01
4.75799948e-01 -1.64151296e-01 -2.06327394e-01 -1.10646832e+00
-7.60216415e-01 6.29278541e-01 4.81923312e-01 5.94377339e-01
3.13885778e-01 1.10325240e-01 1.00432825e+00 5.83766818e-01
1.26044840e-01 -1.22909653e+00 9.61958617e-03 1.04794216e+00
4.39914435e-01 -1.39812434e+00 -5.50280511e-01 -3.46957403e-03
-8.86864066e-01 9.70664322e-01 9.05915082e-01 -1.60079598e-01
1.27859497e+00 -2.23066613e-01 -1.41131029e-01 -3.31486650e-02
-5.73930800e-01 -1.45707414e-01 5.74214399e-01 4.48962212e-01
7.57548451e-01 4.21755284e-01 -6.52723551e-01 1.16246378e+00
-1.73749521e-01 -2.21063226e-01 3.41182947e-01 3.67916465e-01
-5.09160571e-02 -1.08180130e+00 2.27728397e-01 1.09430861e+00
-7.61105835e-01 -1.07798345e-01 -5.46370804e-01 3.74775141e-01
2.37415526e-02 1.14975095e+00 9.63590220e-02 -8.94677520e-01
3.48036736e-01 7.79773772e-01 2.23766014e-01 -8.36994410e-01
-1.12941229e+00 -6.22850299e-01 6.68120742e-01 -1.78621471e-01
-3.45234752e-01 -8.99692476e-01 -7.70833611e-01 -4.27384198e-01
5.01141623e-02 3.16402376e-01 5.16509354e-01 1.51407111e+00
5.83505094e-01 3.07815760e-01 9.66299236e-01 -8.54521692e-01
-4.05035913e-01 -1.32702816e+00 -3.59673202e-01 8.30204129e-01
1.28089339e-01 -4.41487074e-01 -3.87166351e-01 -5.79836629e-02] | [9.10595989227295, 10.335111618041992] |
a77ecb32-9867-4750-9e6b-e3e3e93a8666 | 3d-fully-convolutional-neural-networks-with | 2102.07280 | null | https://arxiv.org/abs/2102.07280v2 | https://arxiv.org/pdf/2102.07280v2.pdf | 3D Fully Convolutional Neural Networks with Intersection Over Union Loss for Crop Mapping from Multi-Temporal Satellite Images | Information on cultivated crops is relevant for a large number of food security studies. Different scientific efforts are dedicated to generating this information from remote sensing images by means of machine learning methods. Unfortunately, these methods do not take account of the spatial-temporal relationships inherent in remote sensing images. In our paper, we explore the capability of a 3D Fully Convolutional Neural Network (FCN) to map crop types from multi-temporal images. In addition, we propose the Intersection Over Union (IOU) loss function for increasing the overlap between the predicted classes and ground reference data. The proposed method was applied to identify soybean and corn from a study area situated in the US corn belt using multi-temporal Landsat images. The study shows that our method outperforms related methods, obtaining a Kappa coefficient of 91.8%. We conclude that using the IOU loss function provides a superior choice to learn individual crop types. | ['Alfred Stein', 'Mariana Belgiu', 'Sina Mohammadi'] | 2021-02-15 | null | null | null | null | ['security-studies'] | ['miscellaneous'] | [ 3.56054455e-01 -1.27873391e-01 -3.09014887e-01 -4.42669958e-01
-3.08556437e-01 -6.28871858e-01 4.85061616e-01 3.20193350e-01
-3.83346885e-01 7.83019364e-01 -2.77589053e-01 -7.93416977e-01
-3.50537509e-01 -1.39581621e+00 -8.09329867e-01 -7.47650623e-01
-2.76881218e-01 -1.20689757e-01 -1.56916007e-01 -3.33638251e-01
-1.39261544e-01 7.75186419e-01 -1.61061621e+00 1.58522844e-01
9.46631610e-01 1.03149986e+00 7.17660427e-01 3.83723795e-01
-2.72995774e-02 1.64890409e-01 -8.39795098e-02 9.59434435e-02
5.57479262e-01 -1.92282513e-01 -5.49104571e-01 5.30567653e-02
3.69818389e-01 -5.44405520e-01 3.28330159e-01 1.26909244e+00
2.37172768e-01 -5.43863103e-02 8.03384364e-01 -9.89150524e-01
-5.87668598e-01 8.01450133e-01 -7.17733264e-01 3.05756042e-03
-1.94976196e-01 -1.01320855e-01 8.55037987e-01 -6.50823414e-01
2.90433019e-01 1.09386837e+00 9.02145147e-01 -1.62063122e-01
-1.19201839e+00 -7.70240605e-01 2.45690882e-01 -6.87125996e-02
-1.51350152e+00 6.79755723e-03 5.40508032e-01 -6.61730528e-01
7.47758806e-01 1.62853345e-01 6.01585329e-01 3.31940740e-01
2.41329864e-01 5.96930683e-01 1.26353860e+00 -5.88898480e-01
-1.20348828e-02 -1.68155767e-02 1.22613296e-01 4.16171908e-01
4.80025232e-01 5.51536620e-01 1.41742125e-01 -4.66388790e-03
9.01920676e-01 1.34199619e-01 4.43326421e-02 -3.82158250e-01
-8.44406784e-01 1.02467430e+00 1.08728981e+00 6.17615342e-01
-8.76373887e-01 -1.27333611e-01 2.72255018e-02 1.79810360e-01
9.01784301e-01 4.75384712e-01 -6.45057023e-01 7.54088461e-01
-1.09719801e+00 3.07680905e-01 3.01048189e-01 6.67404771e-01
1.04385591e+00 -1.37257390e-02 2.02250659e-01 7.26483881e-01
4.11181986e-01 8.03288400e-01 -3.84367369e-02 -6.67729974e-01
1.55251056e-01 6.71152830e-01 2.10723996e-01 -1.25344443e+00
-4.29669559e-01 -4.46585923e-01 -9.66916680e-01 4.94669706e-01
2.63024539e-01 -3.13076645e-01 -9.43395257e-01 1.39247286e+00
8.40760022e-02 -2.51476914e-01 1.02717370e-01 8.61466348e-01
6.10279500e-01 7.41425633e-01 5.27894437e-01 -1.83749273e-02
1.09134579e+00 -2.98943192e-01 -5.76461017e-01 -5.52617274e-02
4.84600365e-01 -5.23007810e-01 6.99488044e-01 6.18626401e-02
-5.47383070e-01 -6.94011629e-01 -1.06598270e+00 4.96037066e-01
-9.69159305e-01 8.08011651e-01 9.34526861e-01 4.74905789e-01
-9.38894093e-01 6.58790946e-01 -1.00511587e+00 -6.29573047e-01
4.28780019e-01 6.04638942e-02 -3.49080920e-01 8.06313008e-02
-1.14055049e+00 9.28842247e-01 1.01689351e+00 6.64457381e-01
-7.10876822e-01 -5.74824333e-01 -9.85725343e-01 1.28159046e-01
1.68506533e-01 2.95955893e-02 7.09988773e-01 -1.25597930e+00
-1.02300251e+00 8.72346997e-01 2.15256661e-01 -5.33854902e-01
2.82614619e-01 -3.69283736e-01 -2.86380112e-01 -1.56227779e-02
1.97359174e-01 7.32415497e-01 2.69819230e-01 -1.29046082e+00
-8.14062119e-01 -5.47766268e-01 1.54131129e-01 6.04402199e-02
-1.62683174e-01 -8.12858120e-02 4.55582887e-01 -7.28385031e-01
5.29952884e-01 -8.94880533e-01 -3.46801579e-01 1.54933408e-01
-1.05250917e-01 1.00866109e-01 8.24884534e-01 -8.28273475e-01
9.76560652e-01 -2.07288361e+00 -3.76294732e-01 2.79697120e-01
-2.38360882e-01 5.96798301e-01 -1.86527863e-01 2.65682220e-01
-3.62815782e-02 2.91233808e-01 -6.62697077e-01 5.58666766e-01
-1.16470776e-01 1.78077117e-01 -2.11696506e-01 3.34081918e-01
6.00370526e-01 7.65144765e-01 -8.61189365e-01 -1.70787051e-01
5.24174035e-01 5.01681566e-01 -1.29803494e-01 3.64062451e-02
-3.06335002e-01 9.23973992e-02 -6.82049453e-01 7.49688447e-01
1.23781395e+00 -7.87435547e-02 3.20914626e-01 -2.21670493e-01
-6.22506976e-01 -2.75335819e-01 -9.34765637e-01 1.26398194e+00
-2.98564881e-01 4.88244206e-01 -1.73376184e-02 -1.13557339e+00
1.23354101e+00 2.67508119e-01 3.77011359e-01 -5.05445778e-01
-8.11262056e-02 3.53793770e-01 -1.02541789e-01 -3.63874763e-01
5.60615897e-01 2.47559920e-02 2.45374173e-01 -1.69104580e-02
-9.16142017e-02 -7.10800709e-03 1.56419203e-01 -5.25831044e-01
3.52309257e-01 6.53046846e-01 4.97230411e-01 -6.52653694e-01
3.78927618e-01 4.43106771e-01 4.87931132e-01 6.69017076e-01
-1.82606474e-01 3.60706449e-01 2.24672511e-01 -7.68565774e-01
-1.04320991e+00 -9.41225350e-01 -4.78152275e-01 1.00214398e+00
-1.54966056e-01 4.66431677e-01 -4.80751216e-01 -5.28944671e-01
2.56399542e-01 7.11360812e-01 -6.30636871e-01 2.10796803e-01
-3.36652517e-01 -1.30611503e+00 4.49343473e-01 5.56570590e-01
1.05773222e+00 -1.13819802e+00 -9.48594987e-01 3.17271113e-01
-2.20675632e-01 -8.63123298e-01 2.43874177e-01 2.69126415e-01
-9.85295475e-01 -1.05176091e+00 -9.38802302e-01 -6.30321443e-01
5.71721315e-01 4.97738272e-01 9.11556542e-01 -2.95454502e-01
-1.80843964e-01 -2.10844189e-01 -6.08894765e-01 -5.99317133e-01
-4.58768487e-01 4.75589216e-01 -4.56705779e-01 -1.51961416e-01
8.12117517e-01 -3.25132340e-01 -4.00883675e-01 1.91696718e-01
-9.70539153e-01 -9.78308031e-04 7.98905671e-01 6.67035818e-01
5.33428133e-01 3.53921443e-01 6.44831240e-01 -5.70813715e-01
9.68498066e-02 -7.36782134e-01 -1.05165958e+00 4.61449057e-01
-6.74052536e-01 -9.94271487e-02 3.44624296e-02 -1.24660745e-01
-1.11488545e+00 5.43770134e-01 4.58036102e-02 7.70063698e-02
-6.39268816e-01 1.06710851e+00 -8.20667595e-02 4.79700752e-02
6.16042197e-01 2.40860246e-02 -9.73882079e-02 -4.76618886e-01
6.73126653e-02 6.75881505e-01 3.67408067e-01 -1.04424678e-01
5.26320040e-01 2.59865642e-01 2.76113525e-02 -1.08823538e+00
-7.64927864e-01 -4.36857074e-01 -9.52319145e-01 -2.36259654e-01
9.81260359e-01 -1.03299165e+00 -2.89790481e-01 6.81973279e-01
-9.80897009e-01 -3.89923602e-01 1.11460820e-01 8.90656173e-01
-1.99569240e-01 -1.16004862e-01 -1.48702130e-01 -9.25373793e-01
-3.21559012e-01 -8.74526143e-01 9.51451540e-01 2.55770296e-01
3.06663752e-01 -8.69128048e-01 -9.15682241e-02 -4.26722467e-02
6.65658653e-01 7.29726076e-01 9.45039690e-01 -2.54159719e-01
-3.60060126e-01 -7.21323639e-02 -7.82467008e-01 4.80770141e-01
6.33272946e-01 1.97611585e-01 -1.11163616e+00 -1.92704543e-01
-4.84680563e-01 -1.55043945e-01 1.17633641e+00 1.00626695e+00
7.66784847e-01 -2.24601384e-02 -2.74400681e-01 6.07127428e-01
1.90363491e+00 5.24235845e-01 5.14518559e-01 4.44111556e-01
5.35266399e-01 9.39954340e-01 8.81447911e-01 3.60123426e-01
1.43921092e-01 2.31945246e-01 7.23738790e-01 -4.57904845e-01
2.86509842e-01 -1.39769256e-01 5.14677279e-02 -1.31794572e-01
-2.85603285e-01 -2.11649150e-01 -1.19920766e+00 7.39274919e-01
-1.86102760e+00 -1.06711471e+00 -1.51816592e-01 2.11797333e+00
4.53618824e-01 -1.61817178e-01 -7.54339844e-02 1.34271383e-01
8.21921110e-01 2.65673876e-01 -4.86676306e-01 -9.03331861e-02
-5.32592773e-01 1.02993242e-01 1.15496898e+00 3.24625194e-01
-1.79428732e+00 1.13133049e+00 6.34061384e+00 4.16210681e-01
-1.36694181e+00 -3.19357604e-01 7.01301932e-01 6.02570415e-01
3.07330620e-02 -8.52703005e-02 -6.28052771e-01 -6.77175447e-02
6.43200636e-01 2.20437706e-01 8.12602267e-02 6.39114141e-01
3.89050752e-01 -3.84741157e-01 -4.80469793e-01 1.82878852e-01
-2.99654871e-01 -1.03107500e+00 2.60329582e-02 7.52131566e-02
8.37918818e-01 1.08740442e-01 -1.16049089e-02 -2.11897190e-03
6.20324492e-01 -1.05857682e+00 3.95747125e-01 4.93789494e-01
6.84374988e-01 -7.10342884e-01 1.01091552e+00 3.23714256e-01
-1.40339339e+00 -1.17613740e-01 -4.77120787e-01 -2.03364566e-01
-3.32073420e-01 6.27273560e-01 -8.34292710e-01 7.57624447e-01
8.56495619e-01 7.51616955e-01 -4.92674768e-01 8.98419499e-01
7.95101188e-03 6.33793414e-01 -4.94125456e-01 1.15601510e-01
6.56629682e-01 -4.26511824e-01 1.11473024e-01 1.04038644e+00
7.06288576e-01 1.82693675e-01 2.40923584e-01 1.11021125e+00
3.87316614e-01 1.40043691e-01 -9.96610343e-01 -2.41361097e-01
2.97345459e-01 1.04480672e+00 -8.46451223e-01 9.71567780e-02
-1.57339782e-01 5.03563702e-01 -2.04642102e-01 3.10423762e-01
-4.00600940e-01 -4.72726077e-01 3.89988840e-01 -7.73627013e-02
4.17771131e-01 -3.31492543e-01 -1.16093889e-01 -7.24100828e-01
-4.49420400e-02 -4.79294240e-01 1.39184684e-01 -8.62920046e-01
-1.07143056e+00 5.35183251e-01 4.44533199e-01 -1.10947406e+00
-1.77593812e-01 -8.60737205e-01 -2.92717159e-01 1.23466027e+00
-2.07060719e+00 -1.61372757e+00 -5.60849249e-01 3.70932743e-02
2.11217195e-01 -1.57336906e-01 1.05311179e+00 8.12324435e-02
-3.35853577e-01 -8.46158713e-03 4.05348182e-01 1.72383368e-01
2.68681318e-01 -1.13176417e+00 2.18526319e-01 8.91196847e-01
-3.85504514e-01 1.42172471e-01 5.59596539e-01 -6.50983095e-01
-7.72349775e-01 -1.62382925e+00 1.08552289e+00 3.69045645e-01
3.69569957e-01 2.20620707e-01 -8.74424040e-01 6.19110048e-01
3.15252356e-02 4.17574169e-03 5.25565743e-01 -6.46759942e-02
-1.53609648e-01 -3.86465877e-01 -1.41327143e+00 4.34253961e-02
4.81179029e-01 -2.88514405e-01 -1.95895180e-01 1.70668319e-01
4.34073687e-01 1.73390999e-01 -8.60701680e-01 8.29277813e-01
8.43940973e-01 -7.78901994e-01 8.83578598e-01 -2.79730111e-01
7.15826094e-01 -2.34449372e-01 -6.26835763e-01 -1.47422910e+00
-5.42519987e-01 2.28735864e-01 7.66600728e-01 9.91498768e-01
4.67833638e-01 -4.58217233e-01 5.88136196e-01 6.37835860e-02
7.79008269e-02 -1.92276046e-01 -2.94559091e-01 -7.09875762e-01
4.87667382e-01 -9.87010598e-02 9.33055222e-01 9.95746255e-01
-6.17064416e-01 -2.93792903e-01 1.98283065e-02 7.50297725e-01
5.75867057e-01 2.18149230e-01 3.30253839e-01 -1.52608585e+00
2.48380512e-01 -4.70358640e-01 -1.93451896e-01 -3.00555289e-01
2.39251435e-01 -6.52726114e-01 1.18394539e-01 -1.43913186e+00
9.75919068e-02 -8.31306815e-01 -4.75762218e-01 8.70232046e-01
-2.29604498e-01 1.87052399e-01 1.06733188e-01 -1.05684632e-02
4.01774347e-01 3.17798555e-01 9.51548338e-01 -2.95549393e-01
-3.19304615e-01 2.10454956e-01 -3.62200052e-01 6.85225964e-01
1.22358012e+00 -2.81081647e-01 -2.20605478e-01 -6.52034700e-01
1.56626865e-01 2.72206943e-02 6.99626029e-01 -9.70729589e-01
-3.93962473e-01 -5.73434412e-01 4.18819040e-01 -1.04393220e+00
-1.15240894e-01 -1.03702557e+00 3.85316789e-01 5.56372344e-01
-4.21810120e-01 -1.34078473e-01 5.01706302e-01 1.67418987e-01
-2.31269777e-01 -1.92437604e-01 6.89568579e-01 -3.23772252e-01
-9.98945773e-01 2.63297200e-01 -3.14228058e-01 -7.68312037e-01
9.61511135e-01 -8.53346810e-02 -1.77146494e-01 9.16455835e-02
-5.33112943e-01 2.30985954e-01 1.91425741e-01 4.38711047e-01
3.28528702e-01 -1.34406197e+00 -1.05344748e+00 5.17442048e-01
3.51591289e-01 -1.88950256e-01 4.29274552e-02 3.05816323e-01
-1.04043651e+00 7.33285487e-01 -7.62704432e-01 -6.07978284e-01
-1.09803820e+00 2.55319208e-01 4.95745897e-01 -3.03091735e-01
-2.34348297e-01 5.54549754e-01 4.90252897e-02 -8.05621743e-01
-9.79350135e-02 -5.56256652e-01 -6.63288713e-01 2.97015637e-01
2.40165919e-01 1.84389934e-01 4.93695699e-02 -7.87630081e-01
-2.13336632e-01 3.59441638e-01 3.23703229e-01 4.46894616e-02
1.60980558e+00 8.02304316e-03 -7.44799972e-02 5.32572210e-01
9.63860452e-01 -6.63604021e-01 -1.31486666e+00 -2.75658518e-01
3.43911767e-01 -4.33680564e-01 5.05030513e-01 -9.76977348e-01
-1.21231067e+00 8.29849720e-01 1.16499507e+00 4.07818019e-01
1.36549222e+00 -4.41838443e-01 1.97640181e-01 5.22525251e-01
3.50163311e-01 -8.20811033e-01 -7.95078635e-01 4.49655354e-01
7.83385396e-01 -1.79228330e+00 7.91754797e-02 -2.97094077e-01
-2.52787501e-01 1.16035569e+00 4.49886560e-01 -1.60177648e-01
9.43485558e-01 1.29119053e-01 3.00615430e-01 -2.34050676e-02
-1.64008543e-01 -6.68893754e-01 9.73754749e-02 7.42288291e-01
7.57237792e-01 6.94814920e-01 -3.80975842e-01 7.58113414e-02
1.08996749e-01 3.14013720e-01 1.75049260e-01 1.03336108e+00
-6.12068772e-01 -9.61337388e-01 -5.01807868e-01 5.06072819e-01
-3.39127928e-01 -1.99283808e-01 -1.67967871e-01 8.90077055e-01
3.43693137e-01 9.51750696e-01 1.52014509e-01 -1.48509338e-01
1.03964679e-01 -1.22011438e-01 2.53860503e-01 -4.17840153e-01
-4.95093167e-01 2.06209764e-01 -2.33936980e-01 -1.36131540e-01
-1.02293777e+00 -6.99994683e-01 -7.35738456e-01 -2.22578660e-01
-5.68975508e-01 -3.79332937e-02 9.83733118e-01 7.90930092e-01
1.00757390e-01 3.62071097e-01 8.26102018e-01 -1.03117192e+00
-5.52870631e-01 -1.13272810e+00 -9.23324764e-01 -6.80212528e-02
5.54679036e-01 -6.03033721e-01 -9.84396636e-02 2.65991110e-02] | [9.448062896728516, -1.579661250114441] |
6ce09e4f-8f7d-4613-bb07-c778119023b9 | texture-classification-using-block-intensity | 2002.01154 | null | https://arxiv.org/abs/2002.01154v1 | https://arxiv.org/pdf/2002.01154v1.pdf | Texture Classification using Block Intensity and Gradient Difference (BIGD) Descriptor | In this paper, we present an efficient and distinctive local descriptor, namely block intensity and gradient difference (BIGD). In an image patch, we randomly sample multi-scale block pairs and utilize the intensity and gradient differences of pairwise blocks to construct the local BIGD descriptor. The random sampling strategy and the multi-scale framework help BIGD descriptors capture the distinctive patterns of patches at different orientations and spatial granularity levels. We use vectors of locally aggregated descriptors (VLAD) or improved Fisher vector (IFV) to encode local BIGD descriptors into a full image descriptor, which is then fed into a linear support vector machine (SVM) classifier for texture classification. We compare the proposed descriptor with typical and state-of-the-art ones by evaluating their classification performance on five public texture data sets including Brodatz, CUReT, KTH-TIPS, and KTH-TIPS-2a and -2b. Experimental results show that the proposed BIGD descriptor with stronger discriminative power yields 0.12% ~ 6.43% higher classification accuracy than the state-of-the-art texture descriptor, dense microblock difference (DMD). | ['Yuting Hu', 'Ghassan AlRegib', 'Zhen Wang'] | 2020-02-04 | null | null | null | null | ['texture-classification'] | ['computer-vision'] | [-1.02091972e-02 -9.65123534e-01 -3.70549589e-01 -2.91673005e-01
-1.03326392e+00 -1.39984444e-01 7.29660511e-01 1.66919142e-01
-1.75592259e-01 4.33934093e-01 2.62936354e-02 1.58878788e-01
-3.13732415e-01 -7.42150366e-01 -1.77839234e-01 -1.29949498e+00
-2.91917861e-01 2.76339576e-02 5.49867034e-01 -1.21626586e-01
6.75014734e-01 8.04352582e-01 -1.67923486e+00 6.27997279e-01
3.54373962e-01 1.85256660e+00 1.40528128e-01 2.94256091e-01
8.58278573e-02 6.35429084e-01 -3.74001712e-01 1.36236683e-01
2.04963252e-01 -8.15385655e-02 -5.92069030e-01 -4.02198732e-02
8.12782824e-01 -3.77439022e-01 -3.01773399e-01 8.33332181e-01
6.43937111e-01 2.75092442e-02 1.08152664e+00 -8.99757564e-01
-7.58481383e-01 -2.38872319e-01 -1.08720529e+00 5.85155547e-01
3.71885061e-01 -3.00085485e-01 9.02558148e-01 -1.13065672e+00
8.61600757e-01 1.28632832e+00 7.77101040e-01 -1.42176554e-01
-1.21182704e+00 -5.08807957e-01 -2.83781916e-01 5.06642640e-01
-1.72089183e+00 -2.22466335e-01 8.44052911e-01 -3.81515205e-01
9.68659878e-01 3.87817770e-01 5.59721231e-01 9.08352673e-01
9.38983619e-01 7.88644791e-01 1.65102470e+00 -4.72645879e-01
1.06801167e-01 -4.12218839e-01 4.07838702e-01 9.31554198e-01
6.99039772e-02 2.65720692e-02 -6.95079386e-01 -4.56134677e-01
1.11410034e+00 3.53206486e-01 3.71892899e-02 -4.45667118e-01
-1.20962882e+00 8.16047847e-01 4.00367737e-01 4.84778196e-01
-5.79161406e-01 -1.62436053e-01 5.15942812e-01 6.67783678e-01
8.82624388e-01 -3.21016908e-01 -1.48879692e-01 -1.65025130e-01
-7.12388575e-01 2.30036467e-01 5.22958279e-01 4.14716750e-01
1.04870820e+00 -3.04896414e-01 -2.82245696e-01 1.38005805e+00
1.42159536e-01 8.87915909e-01 6.68743074e-01 -4.63802069e-01
2.39762470e-01 3.36420298e-01 -1.87339753e-01 -1.89667118e+00
-1.43702030e-01 1.01065964e-01 -1.16329169e+00 2.24170029e-01
3.06858569e-01 8.01855624e-01 -7.19893456e-01 8.89844120e-01
4.96172279e-01 8.82338509e-02 -2.84211308e-01 9.86451626e-01
9.12491322e-01 6.53201401e-01 -1.94569260e-01 1.87796131e-01
1.38209248e+00 -9.53760982e-01 -3.82437557e-01 1.86070532e-01
5.35281181e-01 -1.11240339e+00 7.67466247e-01 2.67586082e-01
-5.47239423e-01 -9.35191810e-01 -9.94708061e-01 1.48198307e-01
-3.19791734e-01 2.43033901e-01 5.35393059e-01 2.69462347e-01
-7.80891955e-01 5.99298120e-01 -8.76357317e-01 -5.41009724e-01
1.89835936e-01 9.50652957e-02 -8.95649552e-01 -4.49930251e-01
-8.45030248e-01 5.08631647e-01 -1.98955730e-01 -1.26840919e-01
-5.06622791e-01 -4.51283604e-01 -9.93866503e-01 -2.38426059e-01
-2.47225285e-01 2.16536410e-02 4.77394074e-01 -5.88675141e-01
-1.35803378e+00 1.13631976e+00 -3.21794838e-01 8.35686922e-02
1.63224917e-02 1.52223423e-01 -4.32796359e-01 5.25888443e-01
2.98158169e-01 4.06409711e-01 1.12219429e+00 -9.17104244e-01
-6.19960487e-01 -4.14784938e-01 -4.45064753e-01 1.42109066e-01
-1.40653640e-01 1.99237969e-02 -4.47500110e-01 -8.83096814e-01
4.90603566e-01 -7.83146203e-01 2.59422392e-01 1.69946961e-02
-1.77173167e-01 -4.18828547e-01 1.09707415e+00 -6.57613695e-01
1.26908231e+00 -2.48175883e+00 2.76282560e-02 5.43975353e-01
1.11838810e-01 2.58126818e-02 -5.50840139e-01 5.46135306e-01
2.32158452e-02 -3.14971715e-01 4.60311830e-01 5.49786054e-02
-1.36456177e-01 1.04032196e-01 -2.46446282e-02 7.59200215e-01
1.30678251e-01 6.83856070e-01 -5.51455140e-01 -5.93905032e-01
6.10165656e-01 4.90545064e-01 -2.05876738e-01 1.35239121e-03
6.33253336e-01 3.13931592e-02 -8.71424317e-01 1.11560667e+00
1.16419649e+00 -1.77879468e-01 -3.39686796e-02 -6.71359658e-01
-2.29262896e-02 -1.71169072e-01 -1.18252766e+00 1.31783712e+00
-3.14474672e-01 5.77265263e-01 -2.57609934e-01 -1.21546149e+00
1.41327882e+00 -1.79051980e-01 5.09764969e-01 -1.20191061e+00
-1.23006530e-01 5.31161964e-01 -5.19745350e-01 -4.78342831e-01
3.54008794e-01 1.68719023e-01 -1.88860074e-01 3.37469369e-01
1.03971116e-01 2.30259851e-01 -8.39214548e-02 -3.20609808e-01
1.21773994e+00 -1.83994666e-01 4.85243618e-01 -7.59802461e-01
8.21740329e-01 -2.17902943e-01 2.17958793e-01 8.28359127e-01
-2.48605967e-01 6.65224195e-01 5.42390227e-01 -8.48790646e-01
-9.33559000e-01 -9.41609383e-01 -6.18860900e-01 1.04783452e+00
4.92318690e-01 -2.67079085e-01 -3.63987774e-01 -5.13856947e-01
5.06276190e-01 -4.36945379e-01 -9.86821294e-01 7.10656047e-02
-5.42030036e-01 -5.57252944e-01 2.05267742e-01 3.56147349e-01
8.39759707e-01 -7.84097135e-01 -2.32921317e-01 4.53334488e-02
9.97935385e-02 -7.22272813e-01 -6.70566857e-01 1.28545672e-01
-6.80464268e-01 -8.65085185e-01 -9.92281854e-01 -1.15301466e+00
5.49468935e-01 7.79506028e-01 7.97773480e-01 -1.42387256e-01
-5.62880337e-01 2.42443934e-01 -6.37993515e-01 5.11202872e-01
5.96294738e-02 2.57507693e-02 -1.21569082e-01 2.98775285e-01
5.64306438e-01 -2.82666445e-01 -8.75512779e-01 8.89087737e-01
-7.41939306e-01 -1.65723398e-01 7.98278928e-01 1.45792985e+00
1.15481973e+00 -1.38694406e-01 1.38751239e-01 -2.86980510e-01
6.11200631e-01 -4.36656207e-01 -3.72829944e-01 2.43292078e-01
-3.55513722e-01 6.96360171e-02 3.74920398e-01 -5.00940084e-01
-8.02625597e-01 -4.93612409e-01 6.27040714e-02 -3.52766395e-01
-2.22736016e-01 6.15877330e-01 3.02377880e-01 -6.13053501e-01
4.07240957e-01 5.19914031e-01 1.19753405e-01 -7.63208807e-01
-9.49834958e-02 9.71032321e-01 1.70543790e-01 -7.30564952e-01
4.13657844e-01 6.23160481e-01 1.41469225e-01 -1.06672585e+00
-5.04594684e-01 -8.24738324e-01 -7.08942652e-01 2.47485247e-02
6.33489549e-01 -8.97104681e-01 -4.50143129e-01 1.12598503e+00
-6.06533587e-01 -2.98935294e-01 -3.04835588e-02 3.84588033e-01
-3.29405427e-01 4.25027251e-01 -8.49535048e-01 -1.99840188e-01
-2.96612561e-01 -1.17717075e+00 1.76804972e+00 1.27638012e-01
8.77294317e-02 -1.16634786e+00 1.92561790e-01 1.10925823e-01
6.24863267e-01 4.91822749e-01 8.77005696e-01 -2.32588544e-01
-3.49240661e-01 -6.48426533e-01 -6.65317893e-01 4.45258558e-01
4.21640247e-01 -8.59789774e-02 -4.86402631e-01 -5.00486553e-01
-2.65599549e-01 -4.31073219e-01 9.25429523e-01 5.64146757e-01
1.16466868e+00 -2.53167562e-02 -5.13921082e-01 7.18504429e-01
1.60431480e+00 1.74003795e-01 6.28467560e-01 5.89076281e-01
5.42420387e-01 2.04720512e-01 9.23874676e-01 6.08212471e-01
2.13557377e-01 1.11459708e+00 -2.39474580e-01 -1.90817237e-01
-2.82447219e-01 3.96259874e-03 2.64072925e-01 1.05727947e+00
-2.75301516e-01 2.18376130e-01 -6.72156930e-01 3.08180571e-01
-1.69938207e+00 -1.17671525e+00 6.06995933e-02 2.04167914e+00
6.46824360e-01 -2.14310631e-01 -1.32791504e-01 8.62652138e-02
6.70900047e-01 6.54096842e-01 -1.20611466e-01 -5.17573178e-01
-3.93308431e-01 6.69062734e-01 4.58747029e-01 2.78995544e-01
-1.43819821e+00 7.62518704e-01 5.98404980e+00 1.85326779e+00
-1.56168997e+00 4.04090062e-02 8.48447740e-01 4.38061297e-01
1.53001666e-01 -4.07237828e-01 -7.69259810e-01 4.04685438e-01
5.68546355e-01 2.74869978e-01 1.49311483e-01 7.73652196e-01
-8.87584761e-02 -4.48063910e-01 -5.98878026e-01 1.34674335e+00
2.19564050e-01 -1.19359934e+00 6.90745786e-02 3.11800152e-01
8.07751656e-01 2.90180463e-02 3.56503874e-01 7.29721785e-02
-2.83113182e-01 -8.79422009e-01 4.79118884e-01 6.14418209e-01
1.02060461e+00 -5.55361390e-01 9.70303595e-01 -3.87413234e-01
-1.66983974e+00 1.95042685e-01 -8.21141720e-01 5.68269119e-02
-5.07174551e-01 6.89420223e-01 -7.03327209e-02 5.68381011e-01
9.50076163e-01 1.21033096e+00 -7.03623295e-01 8.23848605e-01
2.58826733e-01 4.00885969e-01 -3.25706303e-01 -1.75658390e-01
5.68281889e-01 -2.59959519e-01 2.23041818e-01 1.15334690e+00
4.43533659e-01 -9.77686271e-02 1.88992158e-01 1.69157863e-01
3.82950634e-01 3.78888488e-01 -5.77708125e-01 2.95651197e-01
5.03024518e-01 1.45947385e+00 -6.75446868e-01 -3.32525045e-01
-5.88998139e-01 1.13539696e+00 1.20185860e-01 1.54867679e-01
-4.74751145e-01 -7.12846100e-01 8.00880194e-01 -1.36178927e-02
7.83104300e-01 -2.20029756e-01 8.32635164e-02 -1.13099957e+00
8.66587386e-02 -9.11166370e-01 3.72150153e-01 -6.08959258e-01
-1.78403032e+00 4.61450875e-01 -2.69199722e-02 -1.56019914e+00
-4.00318531e-03 -8.96694124e-01 -3.43113601e-01 9.70154047e-01
-1.31633794e+00 -1.38500714e+00 -6.37909055e-01 7.97838092e-01
2.63443261e-01 -4.17401075e-01 7.82771587e-01 3.07620674e-01
-3.35454971e-01 7.46769071e-01 8.94216120e-01 1.40429541e-01
9.15553749e-01 -8.37958217e-01 1.50051787e-01 9.06721950e-02
-2.20035195e-01 2.92322010e-01 6.06997050e-02 -4.84353155e-01
-1.58966696e+00 -7.33934522e-01 6.98571265e-01 -5.08028269e-02
7.92824090e-01 -1.87967896e-01 -7.61661649e-01 2.36184634e-02
-1.24408543e-01 5.98251402e-01 6.72431409e-01 -1.25827596e-01
-5.94383419e-01 -6.69286668e-01 -1.29708469e+00 1.06278121e-01
7.48948157e-01 -8.30426872e-01 -2.59797454e-01 1.00794546e-01
-1.11999907e-01 -2.99073309e-01 -1.36983955e+00 3.42632860e-01
1.15787721e+00 -1.17268062e+00 1.05232537e+00 -1.30453855e-01
4.51852202e-01 -1.02235205e-01 -7.97728419e-01 -9.74534392e-01
-7.52377570e-01 -4.31781076e-02 3.40368956e-01 9.76430058e-01
-2.36695126e-01 -8.11510742e-01 4.80742931e-01 -2.55399287e-01
1.95432886e-01 -1.14788246e+00 -1.33174801e+00 -9.23806489e-01
1.58437788e-02 -6.49838755e-03 4.97179925e-01 9.13058460e-01
7.76066482e-02 -2.42601931e-01 -3.45578283e-01 -1.54863462e-01
8.20184767e-01 4.86920089e-01 6.70205534e-01 -9.89005387e-01
-2.47686841e-02 -5.28056920e-01 -1.38295865e+00 -1.21604884e+00
-3.32346372e-02 -7.07951546e-01 -3.37535977e-01 -9.88772452e-01
5.63410878e-01 -7.27674067e-01 -7.14782715e-01 4.99236852e-01
1.06716261e-03 8.60048950e-01 -2.32232094e-01 6.12621605e-01
-5.39058149e-01 4.53515828e-01 1.37997282e+00 -4.47398901e-01
6.60713241e-02 -5.19852281e-01 -2.70331144e-01 2.37735763e-01
3.62091660e-01 -2.78602272e-01 1.33224472e-01 -1.83417037e-01
-4.54610854e-01 -2.83469707e-02 3.78184080e-01 -1.07931721e+00
-1.08530246e-01 -2.36621052e-01 6.62508965e-01 -5.27318835e-01
2.60519952e-01 -4.11346346e-01 4.03531231e-02 4.35688078e-01
-3.60344700e-03 -2.80334931e-02 7.45824724e-02 4.07395214e-01
-9.04189229e-01 6.80451468e-02 9.08740819e-01 3.35118949e-01
-1.03819525e+00 3.90282303e-01 -3.10233116e-01 -4.57841575e-01
1.03858280e+00 -5.09069145e-01 -5.49585700e-01 5.68865724e-02
-3.38496685e-01 -4.64068800e-01 5.52375197e-01 5.33094883e-01
8.83787632e-01 -2.02124524e+00 -6.86276495e-01 8.53733897e-01
4.40263450e-01 -6.89530551e-01 6.02643311e-01 1.25191855e+00
-9.26685452e-01 4.64138150e-01 -7.10607111e-01 -1.11415124e+00
-1.51671338e+00 1.41606510e-01 1.67629749e-01 -4.28453326e-01
-6.23597205e-01 7.80440032e-01 3.79286528e-01 -8.14408958e-02
-2.09149614e-01 -3.34155053e-01 3.82448360e-02 2.27041826e-01
6.62341058e-01 4.08649564e-01 2.51524597e-01 -9.73090947e-01
-6.04048193e-01 1.40573049e+00 -2.87685454e-01 2.85094321e-01
1.37072039e+00 -9.43444669e-02 -3.73392731e-01 2.00023964e-01
1.88794053e+00 -6.92880377e-02 -9.21113908e-01 -4.34596837e-01
-3.14781874e-01 -9.53104913e-01 2.70593166e-01 -2.77610421e-01
-1.15429568e+00 6.74910784e-01 1.20049000e+00 2.16443211e-01
1.08351171e+00 1.54014692e-01 8.26676488e-01 3.14402908e-01
7.16023326e-01 -9.25165236e-01 3.10091883e-01 5.30988812e-01
9.35827911e-01 -1.22975612e+00 6.93181306e-02 -1.67554870e-01
-5.15086710e-01 1.34613752e+00 1.91565946e-01 -5.88440359e-01
1.06329811e+00 7.74889216e-02 1.95614517e-01 -2.37491474e-01
-4.74232674e-01 9.88245681e-02 5.82399964e-01 5.90840638e-01
3.04219246e-01 1.64577156e-01 -3.67172748e-01 -3.97317819e-02
4.28838909e-01 -1.39637336e-01 -2.65095681e-01 1.10610127e+00
-6.01663888e-01 -1.15645552e+00 -5.78490496e-01 7.33796835e-01
-3.13095152e-01 6.25772923e-02 7.43708834e-02 7.27270246e-01
-1.68929659e-02 5.72837293e-01 3.23789686e-01 -7.09180415e-01
4.78288792e-02 -1.99731484e-01 7.05804884e-01 6.94490131e-03
-6.59168735e-02 3.44678402e-01 -1.21199846e-01 -8.57439697e-01
-5.70037723e-01 -8.08730960e-01 -5.89947581e-01 -6.13147974e-01
-2.45579079e-01 4.78542671e-02 3.96004856e-01 6.14655316e-01
4.78851199e-01 -5.53131811e-02 1.01934981e+00 -1.34965777e+00
-5.13267457e-01 -1.07144988e+00 -1.33975840e+00 7.79897451e-01
2.21790716e-01 -1.05505753e+00 -3.46326292e-01 -7.58102983e-02] | [10.432111740112305, -0.33212170004844666] |
e854b298-c8c8-4905-858b-a427cbd54eb8 | bert-based-classification-system-for | 2109.02975 | null | https://arxiv.org/abs/2109.02975v1 | https://arxiv.org/pdf/2109.02975v1.pdf | BERT based classification system for detecting rumours on Twitter | The role of social media in opinion formation has far-reaching implications in all spheres of society. Though social media provide platforms for expressing news and views, it is hard to control the quality of posts due to the sheer volumes of posts on platforms like Twitter and Facebook. Misinformation and rumours have lasting effects on society, as they tend to influence people's opinions and also may motivate people to act irrationally. It is therefore very important to detect and remove rumours from these platforms. The only way to prevent the spread of rumours is through automatic detection and classification of social media posts. Our focus in this paper is the Twitter social medium, as it is relatively easy to collect data from Twitter. The majority of previous studies used supervised learning approaches to classify rumours on Twitter. These approaches rely on feature extraction to obtain both content and context features from the text of tweets to distinguish rumours and non-rumours. Manually extracting features however is time-consuming considering the volume of tweets. We propose a novel approach to deal with this problem by utilising sentence embedding using BERT to identify rumours on Twitter, rather than the usual feature extraction techniques. We use sentence embedding using BERT to represent each tweet's sentences into a vector according to the contextual meaning of the tweet. We classify those vectors into rumours or non-rumours by using various supervised learning techniques. Our BERT based models improved the accuracy by approximately 10% as compared to previous methods. | ['Amitava Datta', 'Ghulam Mubashar Hassan', 'Rini Anggrainingsih'] | 2021-09-07 | null | null | null | null | ['rumour-detection'] | ['natural-language-processing'] | [-1.72761485e-01 1.48744807e-01 -1.33087590e-01 -1.79212019e-01
-5.78432381e-02 -3.86605978e-01 9.95077729e-01 9.38505948e-01
-4.70100313e-01 8.18651557e-01 6.10615849e-01 -3.80204529e-01
3.48429471e-01 -1.13143861e+00 -9.18365270e-02 -3.84387851e-01
-4.51851673e-02 1.04566924e-01 2.73259044e-01 -9.03659582e-01
7.17895567e-01 3.85799050e-01 -1.61061192e+00 4.00561571e-01
5.84870577e-01 7.47553349e-01 -1.25920564e-01 5.96059263e-01
-6.33861601e-01 1.27743459e+00 -9.71481085e-01 -3.77261758e-01
-2.27713332e-01 -4.81327444e-01 -6.82242811e-01 2.41939411e-01
-1.88823998e-01 3.77560742e-02 8.13713744e-02 9.97891724e-01
3.23793083e-01 -6.25622943e-02 6.06420875e-01 -9.79094446e-01
-4.81759578e-01 6.29997432e-01 -6.61227107e-01 5.03298163e-01
7.65899003e-01 -4.26181614e-01 7.59971082e-01 -7.48861909e-01
4.86253589e-01 1.14054585e+00 7.12051630e-01 -1.39970956e-02
-1.09880137e+00 -6.05086088e-01 -1.24009848e-01 2.03977793e-01
-1.09644830e+00 -3.69158387e-01 9.28792715e-01 -7.75327325e-01
6.02506876e-01 6.65853918e-01 9.02006805e-01 9.19741869e-01
4.96277511e-01 6.00833237e-01 1.44334233e+00 -4.48978782e-01
1.50049046e-01 9.01437283e-01 2.96831608e-01 3.73159558e-01
8.23986009e-02 -5.16571045e-01 -5.34356296e-01 -5.35051644e-01
-2.34280508e-02 1.83674872e-01 -1.70638993e-01 3.47879678e-01
-9.83407676e-01 1.35213900e+00 4.27457958e-01 6.52589798e-01
-7.95130551e-01 -5.33493221e-01 5.89234352e-01 8.13350201e-01
1.24101853e+00 6.37830675e-01 -2.13949010e-01 3.05161048e-02
-9.73000467e-01 2.89126039e-01 1.08421016e+00 1.45351157e-01
8.10939372e-01 -2.43431702e-01 2.57293969e-01 1.00440681e+00
2.27933094e-01 3.25036764e-01 7.56232262e-01 3.84640694e-02
1.69621736e-01 8.93222153e-01 3.38663220e-01 -2.07591200e+00
-5.44798076e-01 -1.81034431e-01 -8.07576299e-01 -9.11153154e-04
1.11662842e-01 -4.11589205e-01 -4.70745862e-01 9.97343540e-01
5.05877078e-01 -3.23943883e-01 -1.15594625e-01 6.37355864e-01
7.51448333e-01 8.98481071e-01 -2.42369369e-01 -5.41665614e-01
1.32382190e+00 -3.68061870e-01 -1.11364293e+00 -1.13120414e-01
5.62079430e-01 -1.13377607e+00 6.25688136e-01 3.40515167e-01
-6.82203948e-01 -3.90864238e-02 -1.08963418e+00 2.60288388e-01
-9.61595535e-01 -5.60266972e-01 3.63499850e-01 7.25388765e-01
-7.12132514e-01 7.00143039e-01 -2.95531929e-01 -3.66866171e-01
2.80820221e-01 2.50944883e-01 -2.48140633e-01 2.89982527e-01
-1.57967043e+00 1.03073514e+00 1.46932527e-01 -3.45461518e-02
1.31124467e-01 -4.47474681e-02 -9.39271688e-01 -3.78056943e-01
2.10903376e-01 -1.65193185e-01 9.82239366e-01 -1.28394699e+00
-1.40828681e+00 7.20881760e-01 -2.71238357e-01 -4.89400238e-01
5.07102907e-01 3.62439565e-02 -6.02622926e-01 -3.09288735e-03
1.94724366e-01 -1.32384509e-01 1.20189333e+00 -8.91977251e-01
-5.66310048e-01 -4.00431365e-01 -2.49433648e-02 -1.20487332e-01
-6.05195224e-01 5.02026379e-01 3.76349211e-01 -4.84697700e-01
1.85696036e-01 -8.46676528e-01 -1.68079495e-01 -6.72535121e-01
-5.64513147e-01 -3.32872123e-01 1.13757455e+00 -6.13414645e-01
1.31074154e+00 -1.73143673e+00 -3.03240836e-01 7.96214268e-02
5.21452606e-01 4.01242226e-01 5.29325783e-01 9.46996212e-01
1.47433534e-01 4.47241157e-01 1.23536110e-01 -2.20939815e-01
-2.84842998e-01 1.65895253e-01 -4.50054348e-01 8.00357997e-01
2.29920849e-01 4.65264201e-01 -1.06744039e+00 -3.53065014e-01
1.73206627e-01 5.41003525e-01 -2.62446523e-01 9.55878198e-02
4.84063625e-02 2.84348071e-01 -5.10420144e-01 2.45673344e-01
5.62870860e-01 -8.64268690e-02 4.77849618e-02 2.01288685e-01
-5.46833575e-01 7.85845339e-01 -8.24042022e-01 2.85592884e-01
-5.85868895e-01 1.01240218e+00 3.39259990e-02 -8.71309519e-01
1.10696018e+00 5.03315210e-01 2.68696904e-01 -2.82084495e-01
5.24678469e-01 1.27236336e-01 -2.72492617e-01 -6.54806912e-01
9.51093793e-01 -6.25776887e-01 -3.21282357e-01 6.62530243e-01
-5.46528339e-01 -1.09575868e-01 1.52585626e-01 2.31054768e-01
8.71002793e-01 -8.33133638e-01 8.00781190e-01 -2.15303317e-01
6.85159564e-01 -4.43330035e-02 1.94360852e-01 4.42373961e-01
-1.05307788e-01 3.35068762e-01 8.27380478e-01 -5.61720610e-01
-7.31590986e-01 -2.99636006e-01 -1.88518599e-01 1.12924147e+00
-2.35520542e-01 -4.79231149e-01 -4.07077104e-01 -6.41149580e-01
1.44623309e-01 7.22463906e-01 -8.94835830e-01 1.61241725e-01
-2.55098283e-01 -1.20465648e+00 6.97993711e-02 -2.06301019e-01
2.66393214e-01 -1.14326501e+00 -2.05375820e-01 4.49266434e-01
-1.78801700e-01 -7.84876347e-01 -1.17828332e-01 -9.80864391e-02
-5.02603114e-01 -9.63836849e-01 -4.64418083e-01 -3.85286450e-01
7.51169205e-01 4.14755374e-01 7.39982247e-01 1.63771734e-01
1.20087683e-01 -2.18124241e-01 -7.80285060e-01 -8.41481864e-01
-7.90197194e-01 5.67685217e-02 1.26928821e-01 4.67898220e-01
4.41795945e-01 -5.01282871e-01 -2.58035392e-01 2.40553036e-01
-1.13559115e+00 -1.87832579e-01 1.58578128e-01 6.12259686e-01
-2.09022135e-01 1.11899994e-01 9.45543468e-01 -1.35650122e+00
1.33785486e+00 -1.10793912e+00 -1.25635341e-01 -6.48187280e-01
-5.30382216e-01 -3.88071984e-01 6.21721506e-01 -3.91578704e-01
-5.15245855e-01 -4.29784387e-01 -1.67420611e-01 4.09763992e-01
-7.30898157e-02 9.89840329e-01 6.08129323e-01 1.80629343e-02
7.21486270e-01 2.04553142e-01 3.13744158e-01 -3.01270038e-01
2.88502057e-03 1.49453700e+00 -3.98520321e-01 4.57928687e-01
6.27073884e-01 6.59684598e-01 -3.79525155e-01 -1.65543532e+00
-1.12191737e+00 -9.33056951e-01 -4.46656764e-01 -3.12400281e-01
5.93681157e-01 -4.64206696e-01 -6.28932774e-01 3.34441036e-01
-1.17973888e+00 5.11319757e-01 1.94051892e-01 2.88796842e-01
5.93482628e-02 3.80892932e-01 -7.22144723e-01 -1.16242695e+00
-3.56371552e-01 -6.16613328e-01 2.66734511e-01 -1.48539260e-01
-7.76292682e-01 -1.18478119e+00 3.11632931e-01 5.21489143e-01
8.11092794e-01 4.68198359e-01 4.47902024e-01 -9.63716567e-01
1.12616517e-01 -7.35739946e-01 -2.10798919e-01 3.84611011e-01
4.91724610e-01 -1.75983235e-01 -9.36777830e-01 -1.02947377e-01
5.09785533e-01 -2.66345203e-01 7.93114781e-01 -1.01897024e-01
4.05013621e-01 -9.04058516e-01 -4.68807295e-02 -3.01661342e-01
1.06790411e+00 -2.31109202e-01 5.36203325e-01 6.81175232e-01
4.53966349e-01 9.99255121e-01 4.01233107e-01 6.31646574e-01
3.67793679e-01 3.49223107e-01 5.22814691e-01 1.93835735e-01
5.53515375e-01 -2.67312583e-02 7.25874662e-01 1.26015532e+00
1.15043126e-01 -1.75562203e-01 -7.55218804e-01 7.40500212e-01
-1.67659521e+00 -1.19037199e+00 -6.22065783e-01 2.09872580e+00
8.35067630e-01 3.35521579e-01 2.91756421e-01 5.81876040e-01
6.69018984e-01 6.40722334e-01 2.68351585e-01 -9.64202344e-01
-1.28768915e-02 -2.73312420e-01 4.24890161e-01 8.05296898e-01
-1.02769601e+00 4.86215055e-01 5.56029940e+00 2.01838925e-01
-1.60419393e+00 2.84933895e-01 3.69193196e-01 -1.87515598e-02
-1.15908533e-01 -2.78909326e-01 -6.73207521e-01 6.01162672e-01
1.14362872e+00 -2.25653425e-01 -6.04952984e-02 5.19625962e-01
7.41407812e-01 -2.78863966e-01 -4.33094263e-01 6.38043046e-01
4.89276558e-01 -1.22724092e+00 -1.91523150e-01 1.06123425e-01
7.14438796e-01 2.22384423e-01 -1.45469993e-01 2.20050439e-01
-2.20798224e-01 -8.34783852e-01 4.34642136e-01 2.60304242e-01
-6.99188411e-02 -8.11462939e-01 1.22151184e+00 5.96724689e-01
-5.44220388e-01 6.47331029e-02 -3.56083483e-01 -9.24256384e-01
3.77101958e-01 1.15219867e+00 -1.46946967e+00 8.55053589e-02
2.92243987e-01 7.48961687e-01 -4.11666363e-01 7.30685234e-01
-2.17589587e-01 7.39238560e-01 -1.28460914e-01 -6.82844698e-01
4.16334361e-01 -7.04577714e-02 8.30289721e-01 1.46845162e+00
5.90591645e-03 -3.01732600e-01 1.25337362e-01 3.64478379e-01
-8.01177993e-02 5.85001469e-01 -7.71117210e-01 -3.71476203e-01
2.80771200e-02 1.18011820e+00 -7.97306716e-01 -3.69717538e-01
-3.42836827e-01 9.15416360e-01 2.03880340e-01 -1.39744535e-01
-4.14670140e-01 -4.95215267e-01 4.07283694e-01 8.84706855e-01
6.99348971e-02 -1.19808733e-01 -2.83856802e-02 -1.06167042e+00
-2.82165520e-02 -8.11978042e-01 5.48955500e-02 -1.40838549e-01
-1.39502478e+00 7.44020522e-01 -2.63065755e-01 -8.28531444e-01
-3.13946247e-01 -2.39162356e-01 -7.35078156e-01 6.72633529e-01
-1.64562488e+00 -7.12880492e-01 3.27294655e-02 2.02211946e-01
5.24129629e-01 1.79927275e-01 7.99001455e-01 2.25167632e-01
-3.72031569e-01 -2.40141843e-02 9.15240794e-02 1.59978211e-01
6.42721236e-01 -1.27580488e+00 1.57579780e-01 2.46384412e-01
-1.62853926e-01 6.62604749e-01 1.29155135e+00 -5.70745111e-01
-1.03591788e+00 -9.36436594e-01 1.78210306e+00 -4.22458261e-01
1.26780558e+00 -3.23299974e-01 -1.03611827e+00 4.58925307e-01
4.53627557e-01 -3.58138472e-01 1.00702381e+00 4.67420906e-01
-1.31266862e-01 2.08867744e-01 -1.17812335e+00 5.81977487e-01
6.71522245e-02 -4.84684169e-01 -9.92121100e-01 6.35309577e-01
3.02154332e-01 2.18403801e-01 -5.65027118e-01 -1.88589618e-01
3.44565153e-01 -1.17871106e+00 4.10166144e-01 -4.68546152e-01
5.47827959e-01 -1.17196828e-01 1.60737753e-01 -1.58700609e+00
-3.96370627e-02 -7.13382602e-01 2.51428962e-01 1.14116979e+00
5.35230935e-01 -1.13523543e+00 6.51221395e-01 9.57334712e-02
4.19554800e-01 -7.33768284e-01 -5.36390483e-01 -2.69426405e-01
-1.73092350e-01 -4.05225396e-01 2.33577222e-01 1.45123434e+00
5.56188226e-01 7.55311966e-01 -6.34621263e-01 -3.16270329e-02
3.77711326e-01 3.62485461e-02 8.08138907e-01 -1.43514800e+00
1.63938895e-01 -4.66366291e-01 -7.17499912e-01 -5.12597084e-01
-9.06882659e-02 -7.00128734e-01 -1.66816071e-01 -1.38718212e+00
1.59405097e-02 -2.54603773e-01 -3.45768295e-02 6.05439283e-02
7.91116804e-02 5.59433520e-01 -3.08907088e-02 3.16296667e-01
-3.95199567e-01 4.09250081e-01 9.82595205e-01 -1.20317630e-01
-4.56100821e-01 2.35825822e-01 -8.03554416e-01 1.04431963e+00
1.04976189e+00 -7.28505015e-01 -6.82253689e-02 2.90265858e-01
1.03289735e+00 -1.98550195e-01 1.08686507e-01 -4.89017576e-01
7.16466457e-02 -7.58226067e-02 1.61606167e-02 -5.64419329e-01
3.53883237e-01 -4.53568906e-01 -3.27964932e-01 3.74393612e-01
-1.98035300e-01 8.78679603e-02 3.21017727e-02 5.50493360e-01
-5.19777656e-01 -3.16310018e-01 6.52285159e-01 -1.15358748e-01
-3.46402340e-02 -3.23583007e-01 -1.38178825e+00 -2.74427831e-01
9.09069657e-01 2.95185409e-02 -1.88279256e-01 -9.61487651e-01
-7.96073914e-01 1.47544220e-03 3.67158353e-01 4.99091655e-01
6.67342544e-01 -8.49404871e-01 -9.24297512e-01 1.33080870e-01
-3.43152210e-02 -4.74147409e-01 -1.11467302e-01 1.28276360e+00
-4.36015904e-01 2.79704392e-01 4.15638648e-03 -1.17836848e-01
-1.54073739e+00 4.54428017e-01 -9.29619297e-02 -1.52475581e-01
-3.56377631e-01 5.20500004e-01 -3.92330647e-01 -4.31396991e-01
-2.70134181e-01 4.70094662e-03 -9.41259921e-01 9.95984316e-01
1.09001100e+00 4.80295092e-01 1.40754119e-01 -1.09021246e+00
-9.53012705e-02 -4.59585451e-02 -5.35693645e-01 1.66900957e-03
1.38002658e+00 -2.84896582e-01 -7.73840368e-01 1.00128460e+00
1.49784851e+00 6.54232323e-01 9.66783613e-03 -6.65404182e-03
2.06191882e-01 -5.19945979e-01 4.55837071e-01 -2.04759210e-01
-4.87343639e-01 5.22440195e-01 -6.87415898e-02 1.35377419e+00
5.19756377e-01 -1.60452463e-02 9.43966031e-01 1.96320578e-01
4.85075153e-02 -9.64431942e-01 -7.44946375e-02 6.93070710e-01
1.03824556e+00 -1.54575086e+00 2.56828398e-01 -4.95289594e-01
-5.84361494e-01 1.01938927e+00 -2.96292067e-01 -2.15715304e-01
1.03609943e+00 -9.95303988e-02 4.08463866e-01 -4.30305958e-01
-6.10507309e-01 -1.06445178e-01 8.06513056e-02 3.70969027e-02
8.30679357e-01 1.44941822e-01 -1.00087261e+00 2.54162759e-01
-4.65316296e-01 -1.65533751e-01 1.13574564e+00 1.15791321e+00
-9.21833098e-01 -1.05867851e+00 -6.49975419e-01 7.38707006e-01
-1.18125963e+00 -1.52334264e-02 -8.46226692e-01 5.24450123e-01
-5.02140224e-02 1.50025606e+00 -1.60028249e-01 -6.11906528e-01
-1.84551664e-02 9.52463970e-02 -2.19495803e-01 -8.41051579e-01
-7.81714857e-01 -2.81560302e-01 5.85126400e-01 -1.01114996e-01
-5.37629306e-01 -7.86993086e-01 -8.71538997e-01 -7.89526165e-01
-6.29102111e-01 4.95010525e-01 9.52001274e-01 1.01260448e+00
1.86978370e-01 2.03082532e-01 1.02436602e+00 -9.09127295e-01
-5.82721293e-01 -1.26770747e+00 -5.43767214e-01 5.61329901e-01
7.42806554e-01 -5.32967091e-01 -7.66187906e-01 -3.36930096e-01] | [8.270001411437988, 10.097012519836426] |
13c3e948-d889-4510-a962-df561d2a7c34 | end-to-end-trainable-attentive-decoder-for | null | null | https://aclanthology.org/E17-2119 | https://aclanthology.org/E17-2119.pdf | End-to-End Trainable Attentive Decoder for Hierarchical Entity Classification | We address fine-grained entity classification and propose a novel attention-based recurrent neural network (RNN) encoder-decoder that generates paths in the type hierarchy and can be trained end-to-end. We show that our model performs better on fine-grained entity classification than prior work that relies on flat or local classifiers that do not directly model hierarchical structure. | ['Hinrich Sch{\\"u}tze', 'Ulli Waltinger', 'Sanjeev Karn'] | 2017-04-01 | null | null | null | eacl-2017-4 | ['morphological-tagging'] | ['natural-language-processing'] | [-3.27135444e-01 6.26319289e-01 -3.82280141e-01 -5.53569615e-01
-8.06257427e-01 -4.14415449e-01 4.80708867e-01 3.23488384e-01
-4.64703411e-01 9.78054643e-01 9.14389849e-01 -5.87616682e-01
1.81016803e-01 -1.14488041e+00 -9.22053277e-01 1.91125691e-01
-1.34344876e-01 6.36948824e-01 1.43668398e-01 -2.53578573e-01
1.86142623e-02 6.76893368e-02 -1.38342428e+00 7.92044997e-01
7.64117599e-01 9.85261798e-01 -1.01816475e-01 1.07875597e+00
-1.00334264e-01 1.52183974e+00 -8.18854213e-01 -5.90947211e-01
-3.27268481e-01 -1.11237802e-01 -1.35155201e+00 -5.39619863e-01
4.80827034e-01 -4.34691787e-01 -6.37803555e-01 4.99069422e-01
2.99811274e-01 5.82168587e-02 9.63722408e-01 -7.48895228e-01
-1.30097294e+00 1.05410361e+00 1.38173744e-01 4.63167399e-01
1.64845511e-01 -6.02373242e-01 1.58487928e+00 -8.65229130e-01
8.67490053e-01 1.10898888e+00 1.07893491e+00 6.01000726e-01
-1.19375718e+00 -3.76530170e-01 4.30733293e-01 2.06836134e-01
-1.40737796e+00 -3.00438792e-01 1.98948056e-01 -3.50451201e-01
1.85225761e+00 -2.04873011e-01 1.71243146e-01 1.30496430e+00
3.89823824e-01 7.49438465e-01 4.95553583e-01 -3.11902761e-01
-9.82887205e-03 -3.89208645e-01 5.17962933e-01 9.38637793e-01
3.79259825e-01 4.21318151e-02 -1.33593203e-02 -3.30352746e-02
9.39265072e-01 -2.92453974e-01 1.73855349e-01 6.71608970e-02
-1.05171847e+00 8.70092988e-01 1.04792225e+00 3.89534384e-01
-5.33991337e-01 2.80679315e-01 5.62639058e-01 3.58823329e-01
5.16224563e-01 8.95100057e-01 -9.26618457e-01 -3.89138162e-02
-9.72983420e-01 3.53420079e-02 1.00768721e+00 1.16032004e+00
4.90402073e-01 1.77964065e-02 -4.53467071e-01 7.98892379e-01
2.56846279e-01 -1.04411215e-01 7.15875804e-01 -7.20168650e-01
4.62695986e-01 7.89557874e-01 7.83131719e-02 -4.55120325e-01
-6.81167722e-01 -7.18529522e-01 -8.48037302e-01 -3.23560774e-01
1.54511491e-02 -3.57958704e-01 -9.60677624e-01 1.55739152e+00
-1.36870578e-01 2.84179300e-01 2.41568491e-01 3.03681076e-01
1.20593774e+00 7.28795290e-01 5.51968634e-01 3.44465584e-01
1.30668950e+00 -1.44873726e+00 -4.32721496e-01 -2.79558059e-02
8.90586674e-01 -8.35908949e-02 6.22647405e-01 -2.82891661e-01
-1.22959077e+00 -6.05028927e-01 -1.06156015e+00 -7.95820534e-01
-9.96246338e-01 1.88173130e-01 5.92269778e-01 4.08718705e-01
-1.23656392e+00 6.00128770e-01 -9.14141417e-01 -1.62768170e-01
2.52518296e-01 4.27643329e-01 -4.17236418e-01 5.29778421e-01
-1.51068008e+00 1.09725499e+00 6.73218787e-01 1.64374381e-01
-6.78537667e-01 -7.55599618e-01 -1.24181950e+00 4.19365883e-01
-1.05448425e-01 -1.44767225e+00 1.64724195e+00 -4.18890297e-01
-1.43134069e+00 6.45891964e-01 -4.49007094e-01 -9.11864460e-01
-8.27006698e-02 -2.21326128e-01 -3.23605269e-01 -2.81028032e-01
3.78013462e-01 7.38135993e-01 2.69169807e-01 -8.16363394e-01
-1.02639616e+00 -3.94166000e-02 1.50274977e-01 1.72359973e-01
-1.50467768e-01 1.67357773e-01 -1.47846416e-01 -9.26892638e-01
-5.00687242e-01 -6.06076896e-01 -4.48772341e-01 -8.28573406e-01
-7.39651263e-01 -6.68771267e-01 3.62745792e-01 -6.38070166e-01
1.66508257e+00 -1.56226313e+00 3.29265028e-01 -1.91172242e-01
2.35041469e-01 9.26333293e-02 -1.24604374e-01 3.16953361e-01
-8.83449763e-02 6.18083537e-01 8.52827132e-02 -4.00598079e-01
3.34233522e-01 -8.30179779e-04 -3.72741044e-01 -3.77751291e-02
3.11088949e-01 1.48779917e+00 -7.26033032e-01 -3.66107851e-01
-1.29012272e-01 3.46227556e-01 -8.53834629e-01 5.39766192e-01
-2.92552143e-01 -1.09777175e-01 -2.97166735e-01 6.10743344e-01
1.43511266e-01 -6.63376153e-01 1.57004863e-01 -2.33507052e-01
-2.00417951e-01 1.07244527e+00 -5.65162539e-01 1.35682499e+00
-9.42421973e-01 6.01224184e-01 -2.06816152e-01 -7.01871753e-01
4.93394852e-01 5.23202002e-01 -2.66961187e-01 -5.41803539e-01
-1.36067998e-03 2.78621733e-01 -2.14385793e-01 -3.53515297e-02
1.07358623e+00 5.79562820e-02 -9.89048421e-01 2.16671661e-01
4.33099747e-01 6.08183920e-01 5.49488626e-02 7.25072697e-02
1.32315242e+00 1.02353334e-01 5.52246451e-01 -8.75070840e-02
9.91858244e-02 -8.03025365e-02 4.86539930e-01 1.03270400e+00
1.49746820e-01 3.61007839e-01 2.08781689e-01 -4.95654553e-01
-1.42764246e+00 -9.20426130e-01 -3.81255746e-01 1.95957851e+00
-5.67528643e-02 -6.73265934e-01 -6.72402740e-01 -9.56060708e-01
3.17835063e-02 6.10134721e-01 -1.00334287e+00 5.01406565e-02
-7.27473915e-01 -3.36943120e-01 1.14029217e+00 1.27862954e+00
5.47929585e-01 -1.29903257e+00 -2.23218769e-01 6.19721889e-01
-2.68244356e-01 -1.12741458e+00 -3.57642770e-01 5.83362520e-01
-7.12270379e-01 -6.27938807e-01 -7.18581200e-01 -1.22485638e+00
5.55481732e-01 -4.27171707e-01 1.72971106e+00 -6.00181185e-02
6.99747875e-02 -2.02874213e-01 -3.61602008e-01 2.65013985e-02
-3.80866915e-01 1.30960131e+00 -3.08037788e-01 -5.45576870e-01
3.92257929e-01 -3.54141533e-01 -3.34257126e-01 2.81932592e-01
-2.95797855e-01 9.70023051e-02 6.88167274e-01 1.13852596e+00
2.52494693e-01 -1.45077221e-02 9.24248934e-01 -1.45368993e+00
7.33129323e-01 -8.28859210e-01 -4.03716505e-01 5.14088750e-01
-4.74659353e-01 5.67648530e-01 8.80373478e-01 1.35893300e-01
-1.28679860e+00 -3.81005436e-01 -7.71291435e-01 -1.64846659e-01
-6.08677149e-01 7.41001725e-01 1.09373219e-01 3.38989973e-01
5.61790764e-01 -1.46724463e-01 -1.15164971e+00 -6.19319916e-01
6.03408694e-01 9.20689762e-01 7.68363357e-01 -5.33540606e-01
5.05002379e-01 -2.02278897e-01 -4.37409163e-01 -4.91020799e-01
-1.13804150e+00 -3.14504236e-01 -9.89906311e-01 6.30005360e-01
1.18581462e+00 -1.19724345e+00 -7.19477594e-01 1.34767383e-01
-1.35587537e+00 -6.54670119e-01 -5.34562692e-02 7.82613978e-02
-4.09659505e-01 -2.88009763e-01 -1.54221725e+00 -4.27668422e-01
-6.33672297e-01 -5.54304719e-01 1.47937143e+00 1.09858334e-01
-5.34255564e-01 -1.29570973e+00 -4.22869623e-02 -5.88926636e-02
5.72446287e-01 -3.66478227e-02 1.13035870e+00 -7.43997276e-01
-6.18890584e-01 -1.19457409e-01 -4.23523098e-01 -3.15186262e-01
-8.95226449e-02 -1.08004518e-01 -5.72798252e-01 3.81191522e-02
-9.20332432e-01 -3.92268658e-01 1.24310291e+00 3.41521263e-01
1.08358681e+00 -8.63154352e-01 -6.38048947e-01 8.36699486e-01
1.22503459e+00 -5.99545725e-02 6.45484507e-01 5.40574491e-01
9.76070583e-01 3.93575609e-01 2.46277992e-02 7.49045312e-02
1.20814896e+00 6.86737359e-01 1.63791422e-02 -2.30496511e-01
-4.52262647e-02 -8.40138137e-01 2.03467011e-01 7.29949176e-01
-1.56680956e-01 -3.68677527e-01 -7.29320586e-01 9.13147748e-01
-1.94028854e+00 -1.16623855e+00 5.53537309e-02 1.48326051e+00
9.21365321e-01 2.75316745e-01 2.72672206e-01 -3.51928055e-01
6.95961416e-01 8.77705682e-03 -3.66724402e-01 -8.67519379e-01
2.31346022e-02 5.17974257e-01 7.85543382e-01 6.40632451e-01
-1.44885230e+00 1.59347820e+00 7.87178946e+00 4.63034987e-01
-7.70173788e-01 7.49320583e-03 4.57516432e-01 3.31041574e-01
-3.36447716e-01 -4.19730157e-01 -1.29604983e+00 3.18311304e-01
1.53357458e+00 9.52903833e-03 4.97131050e-02 1.05228281e+00
-4.08002704e-01 7.39186943e-01 -1.29532993e+00 3.27006787e-01
-3.57939988e-01 -1.75049436e+00 1.70665935e-01 1.80550128e-01
7.26406872e-01 3.72648865e-01 -1.91314131e-01 9.22636926e-01
1.32945693e+00 -1.43822455e+00 7.10754395e-01 4.57844496e-01
8.74346197e-01 -6.75703585e-01 8.79198849e-01 2.27885440e-01
-1.64210474e+00 -2.89731473e-01 -4.15376574e-01 4.68110899e-03
9.87656415e-02 4.03904654e-02 -1.07768631e+00 5.44545829e-01
5.08952379e-01 8.07697475e-01 -7.26085246e-01 8.42100620e-01
-4.54179674e-01 6.77976310e-01 -1.12156663e-02 -1.77846119e-01
3.29257220e-01 7.14382589e-01 1.33112624e-01 1.90720987e+00
8.64429623e-02 1.13641277e-01 1.29484639e-01 6.93759024e-01
-8.64846885e-01 -1.60483375e-01 -5.76388538e-01 4.86710593e-02
7.93086886e-01 1.01831198e+00 -4.67310965e-01 -5.59601963e-01
-5.38104951e-01 1.16104269e+00 1.07481241e+00 2.44588226e-01
-6.11319005e-01 -7.54581869e-01 4.12403554e-01 -1.21086851e-01
1.02819026e+00 -1.89890414e-01 -3.03811461e-01 -1.62665963e+00
-4.33879763e-01 -4.80562776e-01 9.39856589e-01 -6.90727949e-01
-1.57608056e+00 1.18419039e+00 -3.59328836e-01 -6.18777335e-01
-8.95762205e-01 -6.98815882e-01 -5.15099883e-01 5.71914971e-01
-1.51984763e+00 -1.50908661e+00 9.17243958e-02 4.82251108e-01
5.63359320e-01 4.81248945e-02 1.31573510e+00 5.13580501e-01
-8.00690711e-01 1.30436480e+00 -1.24534417e-03 8.37234616e-01
3.31565380e-01 -1.89958346e+00 1.08854091e+00 4.47509468e-01
6.78441301e-02 9.30783331e-01 2.99334079e-01 -6.90561652e-01
-6.52572811e-01 -1.67196429e+00 1.55716670e+00 -7.51700163e-01
5.45015037e-01 -6.07420862e-01 -8.82447481e-01 1.50740576e+00
4.61916357e-01 8.15682188e-02 7.91668832e-01 9.63976324e-01
-8.17675412e-01 4.47927862e-01 -8.29876423e-01 4.24515814e-01
1.19397712e+00 -9.31289852e-01 -1.12761533e+00 -2.65339762e-01
1.04221404e+00 -5.07185280e-01 -1.44803429e+00 5.68280697e-01
6.48916602e-01 -3.98854136e-01 8.98478746e-01 -1.17305839e+00
7.01619208e-01 1.42998304e-02 -1.65050834e-01 -1.62237215e+00
-9.06282008e-01 -2.49739021e-01 -5.58305740e-01 1.08040822e+00
1.18945229e+00 -4.54462051e-01 8.15441906e-01 4.39935327e-01
-3.74242991e-01 -7.81166196e-01 -5.00004530e-01 -4.50423539e-01
5.43362260e-01 5.03179319e-02 8.20202529e-01 8.95170808e-01
3.01635802e-01 9.86544967e-01 -3.69941950e-01 3.09691817e-01
2.12590069e-01 2.72018582e-01 2.83308893e-01 -1.24559975e+00
-2.88548827e-01 -7.24020183e-01 -4.69202846e-01 -1.34126234e+00
6.30892098e-01 -9.03765321e-01 4.70126495e-02 -1.94505072e+00
7.88100362e-02 -3.49284410e-01 -5.33831358e-01 6.52356386e-01
-2.92903602e-01 4.04613078e-01 -1.63196608e-01 2.08067462e-01
-1.25845182e+00 4.75390524e-01 5.07612228e-01 -9.84481499e-02
2.04842929e-02 -2.28960887e-01 -1.03746915e+00 5.62396646e-01
3.42228234e-01 -3.80918443e-01 2.81447649e-01 -5.84671199e-01
4.92568672e-01 1.98439851e-01 -2.99191661e-02 -6.76851273e-01
4.79848653e-01 4.77184057e-01 7.39338398e-01 -5.75115979e-01
2.05157809e-02 -2.66899198e-01 -3.24452490e-01 4.71849293e-02
-1.03538799e+00 2.45940760e-01 3.29576023e-02 4.68091160e-01
-4.45752531e-01 -1.40670180e-01 1.65688232e-01 -3.15999597e-01
-7.80677199e-01 4.20779377e-01 -7.73519337e-01 2.68181413e-01
5.13207495e-01 1.34224042e-01 -5.56077480e-01 -3.68358999e-01
-9.70180988e-01 3.75827014e-01 4.32586282e-01 4.10760701e-01
1.57966912e-01 -1.53259754e+00 -8.45615983e-01 -1.37532756e-01
1.50742054e-01 -3.21361260e-03 -1.06835417e-01 1.13483317e-01
-3.90210867e-01 1.04840088e+00 -1.28256351e-01 -3.83750573e-02
-8.99152577e-01 5.98099530e-01 7.22749472e-01 -9.02883708e-01
-4.07189161e-01 1.00132215e+00 2.99024075e-01 -7.48772085e-01
2.81458706e-01 -5.47819316e-01 -4.49284762e-01 4.41817157e-02
3.20855200e-01 2.07804441e-01 1.43358380e-01 -7.23053455e-01
-3.10792804e-01 3.25986594e-01 -5.12493432e-01 1.02294900e-01
1.36088085e+00 -1.34724841e-01 -2.83662807e-02 4.70470071e-01
1.35680187e+00 -3.20088789e-02 -8.90870571e-01 -2.93829322e-01
4.58657205e-01 3.13004255e-01 -6.52847067e-02 -9.30990696e-01
-3.71603459e-01 8.30419898e-01 -1.13997839e-01 5.85363388e-01
7.23327816e-01 2.20027328e-01 1.10484815e+00 7.42386043e-01
4.14457351e-01 -8.58295262e-01 -3.34852010e-01 1.25714397e+00
5.05379677e-01 -8.68851483e-01 -4.35640723e-01 -2.53635645e-01
-4.49901134e-01 1.02392232e+00 7.30530977e-01 -3.89225423e-01
6.97306275e-01 5.74724913e-01 -2.55603999e-01 -8.00141916e-02
-1.45248771e+00 -4.63736475e-01 5.12844384e-01 3.95258963e-01
1.01337397e+00 2.35004619e-01 2.77859598e-01 1.09357929e+00
-7.05873489e-01 1.34827033e-01 3.50018531e-01 6.43343151e-01
-3.64666224e-01 -1.02961826e+00 2.43068442e-01 6.05384648e-01
-9.53787148e-01 -5.21914244e-01 -4.51451659e-01 5.02134621e-01
7.93435946e-02 7.41914630e-01 5.41845143e-01 -4.87621665e-01
4.03406709e-01 2.16035485e-01 3.66434485e-01 -9.97992039e-01
-9.91705239e-01 -3.44408453e-01 8.53011906e-01 -4.30953175e-01
-3.91773097e-02 -2.34449029e-01 -1.27643538e+00 -3.49208236e-01
-3.55507493e-01 4.93516415e-01 2.76760817e-01 9.75421727e-01
9.33852673e-01 9.70965862e-01 5.49496114e-01 -8.02590191e-01
-3.19288857e-02 -1.14019549e+00 -2.07315639e-01 1.94318190e-01
6.34489954e-01 -4.76004034e-01 1.36695579e-01 -1.15233712e-01] | [9.725520133972168, 9.42722225189209] |
798c221b-77bb-46f0-9de0-b99373751158 | chairs-towards-full-body-articulated-human | 2212.10621 | null | https://arxiv.org/abs/2212.10621v2 | https://arxiv.org/pdf/2212.10621v2.pdf | Full-Body Articulated Human-Object Interaction | Fine-grained capturing of 3D HOI boosts human activity understanding and facilitates downstream visual tasks, including action recognition, holistic scene reconstruction, and human motion synthesis. Despite its significance, existing works mostly assume that humans interact with rigid objects using only a few body parts, limiting their scope. In this paper, we address the challenging problem of f-AHOI, wherein the whole human bodies interact with articulated objects, whose parts are connected by movable joints. We present CHAIRS, a large-scale motion-captured f-AHOI dataset, consisting of 16.2 hours of versatile interactions between 46 participants and 81 articulated and rigid sittable objects. CHAIRS provides 3D meshes of both humans and articulated objects during the entire interactive process, as well as realistic and physically plausible full-body interactions. We show the value of CHAIRS with object pose estimation. By learning the geometrical relationships in HOI, we devise the very first model that leverage human pose estimation to tackle the estimation of articulated object poses and shapes during whole-body interactions. Given an image and an estimated human pose, our model first reconstructs the pose and shape of the object, then optimizes the reconstruction according to a learned interaction prior. Under both evaluation settings (e.g., with or without the knowledge of objects' geometries/structures), our model significantly outperforms baselines. We hope CHAIRS will promote the community towards finer-grained interaction understanding. We will make the data/code publicly available. | ['Yixin Chen', 'Zhiyuan Zhang', 'Siyuan Huang', 'Yixin Zhu', 'He Wang', 'Jieming Cui', 'Zhexuan Cao', 'Tengyu Liu', 'Nan Jiang'] | 2022-12-20 | null | null | null | null | ['human-object-interaction-detection'] | ['computer-vision'] | [ 3.16280164e-02 1.64344132e-01 -6.20234273e-02 4.35804687e-02
-2.81608105e-01 -5.95297456e-01 5.25227666e-01 -6.14666224e-01
-3.63096595e-02 3.33808482e-01 6.86279237e-01 2.27238595e-01
1.42763972e-01 -3.36226016e-01 -9.14838135e-01 -2.83347338e-01
-5.82082085e-02 1.14572716e+00 1.74951151e-01 -1.81029812e-01
-3.38287689e-02 5.06155610e-01 -1.64474905e+00 3.49595159e-01
6.02204084e-01 6.93686485e-01 1.27941906e-01 9.64394152e-01
5.93380272e-01 5.61565816e-01 -4.72020328e-01 -4.11894441e-01
3.87184113e-01 -2.45360047e-01 -9.06941354e-01 4.96707380e-01
5.75111508e-01 -5.98686516e-01 -5.51673889e-01 2.63810426e-01
6.00910187e-01 5.00496805e-01 6.41869903e-01 -1.24240541e+00
-1.29014701e-01 2.98374176e-01 -5.50789714e-01 -3.22125882e-01
9.21825051e-01 6.88745260e-01 8.99192095e-01 -1.07114959e+00
1.06347358e+00 1.62960041e+00 5.68124175e-01 7.62860000e-01
-1.20137215e+00 -2.99207330e-01 2.71489769e-01 1.16274364e-01
-1.38388157e+00 -5.14011800e-01 5.93827903e-01 -6.54855609e-01
9.84047055e-01 4.43107367e-01 1.35763776e+00 1.56531644e+00
1.51418462e-01 1.22343659e+00 5.59259713e-01 -3.06957424e-01
-1.71552420e-01 -4.98150527e-01 -6.05197996e-02 8.08814049e-01
1.70387179e-01 -1.15301125e-01 -8.82542253e-01 -6.02724291e-02
1.21832156e+00 -8.18996727e-02 -3.54252189e-01 -8.95062447e-01
-1.68501866e+00 3.19692582e-01 1.09444290e-01 -1.04455508e-01
-3.85851711e-01 4.91856754e-01 1.48845509e-01 -4.07431662e-01
1.00867294e-01 1.63871318e-01 -3.63701135e-01 -3.23015332e-01
-4.31408584e-01 9.00954843e-01 8.99711967e-01 1.16744065e+00
7.94957578e-02 -1.47010773e-01 -2.71122545e-01 4.98093516e-01
3.47378880e-01 5.71559310e-01 -1.54472366e-01 -1.42005861e+00
6.33315384e-01 4.52125221e-01 5.20751178e-01 -1.00618327e+00
-5.50565541e-01 4.66634333e-03 -4.78573650e-01 1.47028282e-01
7.27593899e-01 1.29304861e-03 -7.59043276e-01 1.61058652e+00
8.94603670e-01 -2.35910062e-02 -4.15296346e-01 1.31397045e+00
7.74966121e-01 2.96474993e-01 2.37032920e-02 3.07191163e-01
1.55229461e+00 -1.25273776e+00 -6.25863075e-01 -3.13255608e-01
2.90157616e-01 -5.63081920e-01 1.21410716e+00 5.01425803e-01
-1.66629505e+00 -6.75513446e-01 -6.29271626e-01 -5.13572156e-01
1.83793411e-01 1.34541944e-01 7.87350833e-01 2.68954068e-01
-5.16356289e-01 5.31749904e-01 -1.07613838e+00 -4.39060420e-01
3.00231516e-01 5.57966053e-01 -5.88896275e-01 1.15920417e-01
-7.47892499e-01 8.22217047e-01 -1.66333392e-02 3.99522632e-01
-1.17329907e+00 -6.67413712e-01 -9.98423338e-01 -2.72922069e-01
7.19455540e-01 -1.52238095e+00 1.36318862e+00 -5.52573085e-01
-1.55306470e+00 9.67666686e-01 -7.20507652e-02 -1.79235205e-01
1.12027180e+00 -8.82946849e-01 1.48687124e-01 2.47072473e-01
-1.74594745e-01 8.34116399e-01 8.01204085e-01 -1.30643022e+00
-1.50133118e-01 -5.19535601e-01 2.19447523e-01 6.97632194e-01
2.01831296e-01 -2.71628112e-01 -9.68333006e-01 -7.31650710e-01
2.50289422e-02 -1.37375176e+00 -1.20450579e-01 5.52573860e-01
-7.33366489e-01 -3.43632996e-02 8.25425267e-01 -8.15220654e-01
8.64577472e-01 -1.83781493e+00 1.00192881e+00 1.09250389e-01
2.95558065e-01 3.43321003e-02 7.90360346e-02 3.38562220e-01
1.60159618e-01 -1.21755190e-01 -6.70816153e-02 -6.89378679e-01
2.52051473e-01 2.50156760e-01 -2.01929286e-02 5.39702892e-01
-2.40031779e-01 1.42511272e+00 -6.87709987e-01 -5.54844737e-01
5.33198416e-01 7.15337217e-01 -7.34349847e-01 5.22422373e-01
-3.45107615e-01 1.12723982e+00 -5.42795539e-01 7.91312873e-01
2.07241878e-01 -3.01438600e-01 1.98413715e-01 -5.72905660e-01
2.06783399e-01 -9.54662263e-02 -1.43514609e+00 2.07648253e+00
-3.14685374e-01 2.25687265e-01 2.36376315e-01 -4.73077804e-01
3.14807445e-01 3.79588187e-01 7.50834227e-01 -1.44305468e-01
1.75532028e-01 -1.79177746e-01 -6.17990270e-02 -6.82413161e-01
1.21092997e-01 4.88977171e-02 -1.32093042e-01 2.35245481e-01
-3.14929307e-01 -4.15315181e-01 -2.27176428e-01 5.06506413e-02
8.87331128e-01 8.81911695e-01 3.56026441e-01 5.39280660e-02
1.85779691e-01 -6.04225397e-02 3.29575129e-02 6.15310788e-01
-6.99947104e-02 8.64507616e-01 8.42810422e-03 -5.60176015e-01
-1.07751381e+00 -1.34025705e+00 2.48632595e-01 1.08160496e+00
3.06343555e-01 -3.83012056e-01 -9.06860769e-01 -3.36822748e-01
2.19802305e-01 3.58005673e-01 -6.96118951e-01 3.09589189e-02
-9.85460639e-01 -1.33254200e-01 3.03864896e-01 8.15229177e-01
4.18355644e-01 -1.28918839e+00 -1.08135462e+00 5.39128333e-02
-6.93357289e-01 -1.09824717e+00 -8.54355633e-01 -5.45543015e-01
-7.51787484e-01 -1.15233123e+00 -8.22038710e-01 -3.52141082e-01
5.14769077e-01 2.03372970e-01 1.20700443e+00 2.01866671e-01
-8.12862694e-01 8.04851174e-01 -2.09777340e-01 -1.79244466e-02
-9.86965094e-03 -2.30441958e-01 1.74551696e-01 -3.88148576e-02
-5.00735044e-01 -6.97777033e-01 -9.77007866e-01 6.96728110e-01
-3.99238348e-01 6.72997832e-01 2.84827054e-01 4.05838400e-01
6.30427718e-01 -4.43376392e-01 -3.56350422e-01 -5.23384750e-01
-7.36279860e-02 -1.07539691e-01 -1.46635575e-02 1.62950143e-01
4.06656504e-01 -2.15324700e-01 1.03221927e-02 -7.44132638e-01
-1.26330602e+00 4.02704209e-01 1.00691222e-01 -5.46463192e-01
-2.69264400e-01 -1.42109729e-02 -6.09891236e-01 4.71786335e-02
5.20365655e-01 -2.49517009e-01 -1.49344236e-01 -5.29773653e-01
5.75923026e-01 1.86185371e-02 8.77146125e-01 -9.82042253e-01
8.28153193e-01 7.96707094e-01 8.00313652e-02 -8.14343810e-01
-7.34019339e-01 -3.42222154e-01 -1.12397909e+00 -5.28831244e-01
1.16873753e+00 -8.12119663e-01 -1.39841461e+00 6.66116714e-01
-1.18698478e+00 -8.76234889e-01 -2.33387962e-01 5.18458903e-01
-1.08730686e+00 3.81960124e-01 -6.64215088e-01 -8.91509295e-01
-1.85508981e-01 -1.10300338e+00 1.74167347e+00 -3.03200006e-01
-8.42891455e-01 -7.59721100e-01 -1.26620784e-01 1.08410358e+00
-2.33222350e-01 8.61560583e-01 5.21436334e-01 -9.13879368e-03
-7.31253266e-01 -5.55666946e-02 4.20162678e-01 -3.10347855e-01
-8.85805786e-02 5.91887161e-02 -5.93034446e-01 -3.29852641e-01
-4.51088786e-01 -6.06433809e-01 3.49526048e-01 5.59872448e-01
1.30792940e+00 -3.75363946e-01 -3.46196413e-01 6.52607739e-01
5.98544776e-01 -2.25615039e-01 6.76763237e-01 3.44099961e-02
1.31789172e+00 1.05481625e+00 8.47780943e-01 7.68703461e-01
5.43194473e-01 1.28988349e+00 5.28044283e-01 7.10958466e-02
-3.77854854e-01 -4.53281254e-01 3.77037644e-01 4.72281277e-01
-8.46573114e-01 -3.24101031e-01 -9.91801322e-01 3.19142818e-01
-1.90201461e+00 -9.07572031e-01 -2.92967647e-01 2.14214778e+00
6.00121617e-01 8.71223658e-02 5.44969440e-01 -2.20419273e-01
3.94393355e-01 -1.91790443e-02 -8.35971117e-01 2.72430032e-01
2.17730358e-01 8.28275755e-02 -1.60606112e-02 6.46918952e-01
-8.79559517e-01 1.07544565e+00 5.55572557e+00 3.41517895e-01
-6.07640624e-01 -1.06847793e-01 5.25051653e-02 -5.10769844e-01
-8.93670321e-02 -8.36743489e-02 -6.81268692e-01 4.21619490e-02
4.78315949e-01 2.43915275e-01 4.07284319e-01 5.65402269e-01
3.61705989e-01 -7.53423497e-02 -1.45430422e+00 1.04238117e+00
-1.76028553e-02 -1.03157306e+00 1.38380289e-01 2.47721210e-01
6.33031130e-01 -4.55649108e-01 1.32585187e-02 6.38075843e-02
2.71099329e-01 -1.13289809e+00 1.28562558e+00 9.53879356e-01
6.88747227e-01 -3.99997503e-01 1.61978945e-01 5.71895182e-01
-1.55087233e+00 6.60058558e-02 3.72366905e-01 -2.73849159e-01
5.99818170e-01 -1.59739759e-02 -4.01167989e-01 5.30349076e-01
7.51233041e-01 6.85905755e-01 -3.16736311e-01 6.00542903e-01
-1.50538787e-01 3.17496687e-01 -3.78435731e-01 4.54998672e-01
-2.92693436e-01 -1.88512310e-01 7.63027072e-01 9.76046205e-01
-1.47654125e-02 4.87206280e-01 5.58528960e-01 6.86377525e-01
1.90269664e-01 -6.02096841e-02 -5.03432989e-01 8.96303281e-02
1.87104404e-01 1.03639817e+00 -6.69553399e-01 -3.96535844e-01
-8.53790902e-03 1.32343292e+00 2.34746218e-01 4.05451983e-01
-1.06277883e+00 3.37970018e-01 7.54433990e-01 6.05948508e-01
2.02601165e-01 -6.44281864e-01 -2.06191108e-01 -1.35128284e+00
2.65100300e-01 -1.14954937e+00 2.94409722e-01 -1.04991341e+00
-8.19373429e-01 4.53978479e-02 5.76393127e-01 -9.96481538e-01
-3.41166079e-01 -5.22473931e-01 -3.22666675e-01 5.27885914e-01
-2.91575462e-01 -1.61000371e+00 -7.79286325e-01 7.68611789e-01
9.31117892e-01 5.76722205e-01 5.13567328e-01 -1.31697431e-01
-4.00514245e-01 2.68839777e-01 -7.74660826e-01 -5.03114425e-02
5.61189890e-01 -1.04889798e+00 6.94968343e-01 2.74759889e-01
2.74303675e-01 7.98799217e-01 8.82622540e-01 -8.83905292e-01
-1.85936153e+00 -8.08704495e-01 4.26685840e-01 -1.21249032e+00
2.88251430e-01 -7.79856741e-01 -6.14128709e-01 1.06029367e+00
-3.11674625e-01 2.76241452e-01 3.27078700e-01 -8.89457762e-02
-1.41821533e-01 4.31611329e-01 -8.80526602e-01 1.08151352e+00
2.01399994e+00 -2.73163289e-01 -6.03614450e-01 5.49452841e-01
5.20913422e-01 -1.02606988e+00 -9.91115034e-01 5.03605723e-01
1.28368485e+00 -7.09208071e-01 1.50019765e+00 -8.37058067e-01
3.45052481e-01 -3.60915720e-01 -1.36916146e-01 -7.20409453e-01
-1.67985514e-01 -7.48627305e-01 -7.26621091e-01 5.18169343e-01
-2.24919841e-01 6.66405112e-02 9.42110360e-01 9.29506600e-01
-5.09827062e-02 -8.48784387e-01 -6.84201002e-01 -5.65024376e-01
-3.55914772e-01 -5.72089851e-01 2.09840447e-01 4.50468570e-01
-2.00570464e-01 9.17188153e-02 -8.63764524e-01 2.00415686e-01
7.27311194e-01 1.48018733e-01 1.63981175e+00 -1.20584965e+00
-7.74495959e-01 -1.81183442e-01 -2.79167265e-01 -1.39997280e+00
3.01775366e-01 -3.59155089e-01 7.86825642e-02 -1.44367206e+00
3.94961715e-01 -3.52057293e-02 6.85184360e-01 5.50979972e-01
-2.03893170e-01 4.20424253e-01 4.35187072e-01 4.18980688e-01
-6.52335584e-01 6.40696168e-01 1.81929827e+00 1.17092825e-01
-2.52295166e-01 1.67269185e-01 -9.88826603e-02 1.14870298e+00
3.14489067e-01 -2.65860669e-02 -4.39598560e-01 -3.24732721e-01
-1.09380536e-01 3.31212163e-01 1.13263655e+00 -1.03488147e+00
1.59481186e-02 -4.01204139e-01 6.41955376e-01 -5.61573803e-01
8.69699895e-01 -7.55806684e-01 1.08430529e+00 6.64784849e-01
-3.23764145e-01 -9.67419222e-02 7.89834559e-02 6.40543997e-01
4.56557781e-01 3.54280621e-01 3.86069924e-01 -4.13521081e-01
-5.08907497e-01 5.31374395e-01 -6.68810979e-02 1.22019939e-01
1.03194439e+00 -5.14872611e-01 2.00117230e-01 -5.87351561e-01
-1.36124110e+00 1.75798073e-01 6.59087658e-01 5.97083390e-01
6.52770877e-01 -1.18599832e+00 -6.42030299e-01 -1.29100401e-02
6.39756070e-03 3.58461827e-01 4.16564375e-01 9.47616935e-01
-4.40166652e-01 1.93120822e-01 -2.12265268e-01 -7.74647951e-01
-1.66277611e+00 3.17186207e-01 3.04013342e-01 -1.05834737e-01
-1.11109054e+00 7.28586495e-01 7.37953126e-01 -4.96224225e-01
3.34495932e-01 -3.17907453e-01 1.11301601e-01 -3.88177693e-01
3.80805492e-01 7.72136867e-01 -4.08031285e-01 -1.02530849e+00
-3.11388731e-01 9.44936395e-01 2.37367794e-01 -3.60600054e-01
1.07079720e+00 -8.65479633e-02 3.17748249e-01 5.17501295e-01
8.16599429e-01 1.20850317e-01 -1.78716683e+00 3.97970565e-02
-5.09399354e-01 -6.98576689e-01 -7.08023012e-01 -7.22988307e-01
-7.95730054e-01 8.84986520e-01 1.25040382e-01 -6.48916900e-01
6.49755180e-01 3.98916334e-01 8.45266819e-01 4.30791825e-01
7.92196929e-01 -8.35618854e-01 7.58647561e-01 2.94506818e-01
1.47493243e+00 -8.71875942e-01 1.38800681e-01 -8.01523685e-01
-8.20489049e-01 7.87927389e-01 9.21660900e-01 -1.20267840e-02
2.10260779e-01 1.38414711e-01 -3.68794292e-01 -4.14811194e-01
-5.59879959e-01 4.04185802e-03 8.72008681e-01 6.65696979e-01
2.75880277e-01 1.68321669e-01 5.58283851e-02 4.58104402e-01
-3.86080295e-01 -6.81099072e-02 -2.59003993e-02 9.75581110e-01
-2.02610001e-01 -9.14390862e-01 -5.76985359e-01 7.64341950e-02
-6.90571591e-02 4.22360659e-01 -7.06388652e-01 1.06433511e+00
2.60996878e-01 6.56156361e-01 -2.24933788e-01 -1.66940823e-01
8.61783087e-01 -8.00911784e-02 1.01612401e+00 -6.65988505e-01
-5.18247008e-01 6.24702647e-02 2.08051026e-01 -1.22977591e+00
-3.84189129e-01 -9.92247105e-01 -1.50386441e+00 -3.98666829e-01
-2.87712142e-02 -4.13080186e-01 3.20646882e-01 8.77687275e-01
2.61479139e-01 3.27000856e-01 -1.02301836e-01 -1.80591428e+00
-4.73755389e-01 -7.65124619e-01 -3.48077655e-01 9.76864636e-01
2.03858674e-01 -1.19477987e+00 -7.95389339e-02 3.86671871e-01] | [6.9067864418029785, -0.8814417123794556] |
9e08f62b-003c-47c0-b94b-ef5abb642ce8 | ax-mabsa-a-framework-for-extremely-weakly | 2211.03837 | null | https://arxiv.org/abs/2211.03837v1 | https://arxiv.org/pdf/2211.03837v1.pdf | AX-MABSA: A Framework for Extremely Weakly Supervised Multi-label Aspect Based Sentiment Analysis | Aspect Based Sentiment Analysis is a dominant research area with potential applications in social media analytics, business, finance, and health. Prior works in this area are primarily based on supervised methods, with a few techniques using weak supervision limited to predicting a single aspect category per review sentence. In this paper, we present an extremely weakly supervised multi-label Aspect Category Sentiment Analysis framework which does not use any labelled data. We only rely on a single word per class as an initial indicative information. We further propose an automatic word selection technique to choose these seed categories and sentiment words. We explore unsupervised language model post-training to improve the overall performance, and propose a multi-label generator model to generate multiple aspect category-sentiment pairs per review sentence. Experiments conducted on four benchmark datasets showcase our method to outperform other weakly supervised baselines by a significant margin. | ['Mingxue Wang', 'Sourav Dutta', 'Walid Magdy', 'Sabyasachi Kamila'] | 2022-11-07 | null | null | null | null | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 6.38631284e-01 2.81668931e-01 -6.15682542e-01 -8.07526529e-01
-9.09049511e-01 -5.91091394e-01 8.29291284e-01 4.55625564e-01
-5.41946113e-01 6.48694754e-01 3.26299906e-01 -4.01277244e-01
4.92518067e-01 -6.71422601e-01 -2.55095989e-01 -4.02511150e-01
5.40597618e-01 3.26077729e-01 6.20082617e-02 -5.28110385e-01
4.42975312e-01 -4.20920253e-01 -1.31466365e+00 6.28571033e-01
4.98544455e-01 6.51228487e-01 -2.36891583e-01 4.97896582e-01
-3.68121147e-01 8.23886454e-01 -6.66152775e-01 -8.45881939e-01
3.86412106e-02 -5.55931449e-01 -6.70634389e-01 4.00913566e-01
-1.51059683e-02 1.88728735e-01 7.59797215e-01 9.33670223e-01
4.81300443e-01 -9.01949257e-02 9.40698445e-01 -1.08861732e+00
-3.56604248e-01 8.56329620e-01 -9.10614014e-01 -7.40953758e-02
3.78828168e-01 -2.64851987e-01 1.61010706e+00 -1.16438448e+00
4.75174010e-01 9.97366846e-01 7.80505717e-01 3.97744000e-01
-8.27435911e-01 -5.91423333e-01 4.70882595e-01 -1.16222046e-01
-9.69566643e-01 -1.00889556e-01 1.01022017e+00 -2.78757870e-01
1.03130400e+00 6.52931184e-02 5.47679365e-01 8.39411020e-01
3.10837835e-01 7.98049927e-01 1.47289157e+00 -8.31724167e-01
3.42799693e-01 6.40802681e-01 5.28652787e-01 4.40486997e-01
2.78153181e-01 -5.62805831e-01 -6.60412669e-01 -2.58891642e-01
-9.22685936e-02 -1.05579034e-03 3.25042367e-01 -1.17032342e-01
-1.08269107e+00 1.29716289e+00 -9.62954611e-02 1.62117243e-01
-2.41041631e-01 -1.49107531e-01 6.61340415e-01 4.76388186e-01
1.17702496e+00 4.86515820e-01 -9.32164967e-01 -3.07489000e-02
-7.78029740e-01 1.26578823e-01 1.01127017e+00 9.20216441e-01
1.03381777e+00 -2.62902915e-01 -1.94119424e-01 1.01774728e+00
5.01436949e-01 6.72396243e-01 6.69121802e-01 -3.92516226e-01
4.15881664e-01 9.07159448e-01 -1.64023742e-01 -9.10788298e-01
-4.34667438e-01 -4.15203601e-01 -5.14743745e-01 -1.08152218e-01
-2.74046153e-01 -4.42707509e-01 -9.24886167e-01 1.20317709e+00
3.50250214e-01 -2.20982939e-01 8.23220760e-02 5.13389587e-01
9.07512426e-01 2.91127205e-01 2.56645679e-01 -1.95019156e-01
1.63701534e+00 -1.34736300e+00 -7.23014712e-01 -5.55687308e-01
8.82049203e-01 -1.04333198e+00 1.37519860e+00 4.86289889e-01
-6.07790768e-01 -1.58050373e-01 -1.00697362e+00 2.68634826e-01
-5.65542400e-01 1.95042506e-01 9.65450168e-01 9.91702437e-01
-8.26164126e-01 8.93069282e-02 -5.52437663e-01 -3.45098734e-01
3.91728133e-01 2.51962990e-01 -1.97849855e-01 1.89088434e-01
-1.06602263e+00 6.83575034e-01 -1.16651244e-02 -4.14541364e-01
-2.97482520e-01 -3.38954240e-01 -1.15791965e+00 -2.95525253e-01
3.00658226e-01 -5.87136745e-01 1.43533540e+00 -1.37379277e+00
-1.48332572e+00 1.03941417e+00 -4.74770218e-01 -3.04719985e-01
-7.36572742e-02 -2.03905806e-01 -3.51218641e-01 9.35080275e-02
5.98187804e-01 5.83563566e-01 8.65863025e-01 -1.27668667e+00
-7.54016459e-01 -5.77862300e-02 1.05388999e-01 4.93400753e-01
-8.38598788e-01 4.75229502e-01 -2.44488791e-01 -8.43430638e-01
-2.01902509e-01 -1.08202231e+00 -7.80007303e-01 -5.96712530e-01
-5.03415823e-01 -3.46483290e-01 4.17898953e-01 -4.69763428e-02
1.04666317e+00 -1.73064792e+00 -4.60125864e-01 2.51163542e-01
-3.58318798e-02 -3.47813740e-02 -1.50959402e-01 3.55068564e-01
-3.22073251e-02 3.04159731e-01 -4.53911185e-01 -8.66124034e-01
-6.78669587e-02 -2.45252028e-02 -5.07600665e-01 1.37671798e-01
4.44098383e-01 9.20133710e-01 -1.06280470e+00 -5.20782709e-01
-1.82679459e-01 3.19666952e-01 -6.33529961e-01 8.41152295e-02
-2.26490065e-01 4.31259908e-02 -4.44226652e-01 8.02531719e-01
3.30083817e-01 -3.61264050e-01 1.29303396e-01 3.88025539e-04
2.69576579e-01 6.16424024e-01 -8.42258930e-01 1.31778705e+00
-7.87654936e-01 3.11701030e-01 -3.83352607e-01 -9.52877522e-01
1.01205873e+00 2.52645552e-01 5.04892528e-01 -2.04059288e-01
2.97776431e-01 1.64022654e-01 -1.19195245e-01 -1.88867867e-01
7.59558260e-01 -7.72438705e-01 -5.34719646e-01 1.03371990e+00
9.99076962e-02 -5.13643324e-01 5.34837902e-01 2.92939276e-01
1.00189233e+00 6.27449155e-02 5.93609035e-01 -1.99025795e-01
6.09968424e-01 2.57050097e-01 3.68188024e-01 6.53468847e-01
-1.47158086e-01 1.11538374e+00 6.06052220e-01 -2.06128001e-01
-7.89368391e-01 -3.61470252e-01 -6.12344407e-02 1.41062713e+00
-1.09212124e-03 -8.85967314e-01 -6.26397431e-01 -1.25113773e+00
-3.69418204e-01 6.75208986e-01 -6.58344388e-01 4.46015224e-03
-1.86435223e-01 -1.37940550e+00 1.72139555e-01 3.40388834e-01
-2.73833284e-03 -1.42002428e+00 -1.69379085e-01 2.00972572e-01
-7.38943294e-02 -1.28131628e+00 -4.89365250e-01 5.19078434e-01
-6.93741262e-01 -9.25530136e-01 -5.33828735e-01 -9.22293603e-01
9.26244497e-01 3.32821637e-01 1.47193635e+00 -4.53112498e-02
2.41781890e-01 2.84980237e-01 -8.87792170e-01 -9.51098800e-01
-4.19738263e-01 6.41498268e-01 2.55000032e-02 1.85939312e-01
1.02843344e+00 -2.44055629e-01 -5.22383571e-01 1.42794535e-01
-9.55399394e-01 1.10402079e-02 6.77616775e-01 6.85852766e-01
6.15152717e-01 -2.49005109e-01 9.64691460e-01 -1.85214484e+00
9.05810118e-01 -6.93738580e-01 1.91170931e-01 -3.15681174e-02
-1.08192825e+00 -2.29600340e-01 7.31771469e-01 -3.64506096e-01
-8.87987018e-01 9.27506685e-02 -2.75076866e-01 3.54903400e-01
-1.84135661e-01 8.77120256e-01 2.10803896e-02 3.08000088e-01
5.45614004e-01 -4.64188866e-02 -2.11913344e-02 -1.31603166e-01
4.61226970e-01 9.53516722e-01 -2.37749398e-01 -2.10344538e-01
7.52335072e-01 5.39262533e-01 -3.63588184e-01 -5.29953897e-01
-1.64626682e+00 -1.01837671e+00 -5.41273415e-01 -1.01793669e-01
5.92606544e-01 -1.26881659e+00 -3.34804244e-02 2.57449061e-01
-8.79235566e-01 -1.33474827e-01 -3.37768018e-01 3.04575831e-01
-2.15197057e-01 2.08457544e-01 -5.56174576e-01 -6.04451180e-01
-7.00728893e-01 -9.78913367e-01 1.54314125e+00 2.68537067e-02
-6.55420184e-01 -1.26037812e+00 4.26259935e-01 6.43455088e-01
3.48874956e-01 -8.99035409e-02 5.07844210e-01 -1.09063864e+00
1.31135702e-01 -5.91449976e-01 7.05623254e-02 4.85562265e-01
4.60278779e-01 -2.61372209e-01 -9.49540615e-01 -4.36607860e-02
9.28944424e-02 -5.59061944e-01 1.02723086e+00 1.29661411e-01
7.48842776e-01 -2.34540418e-01 -1.66038170e-01 1.80902973e-01
1.28364110e+00 -2.54575610e-01 2.19938532e-01 5.49443662e-01
8.73721302e-01 7.72811353e-01 9.78011191e-01 4.35589790e-01
8.11078131e-01 2.68528163e-01 -1.00328987e-02 -3.31543386e-01
9.16610360e-02 2.45089624e-02 7.25104213e-01 1.40182996e+00
3.30060452e-01 -2.72400767e-01 -6.37769520e-01 9.37185466e-01
-1.71650493e+00 -6.58033729e-01 -1.92734823e-01 1.65490687e+00
1.28145289e+00 4.65360492e-01 1.42764002e-01 2.42970720e-01
5.56067407e-01 3.56882155e-01 -1.77616864e-01 -7.62814701e-01
-3.02295744e-01 5.88843048e-01 3.99855584e-01 4.39844459e-01
-1.38350368e+00 1.19393671e+00 6.06203604e+00 7.15675652e-01
-8.80176306e-01 2.65771121e-01 9.43494380e-01 -2.39576418e-02
-7.63197422e-01 2.35076502e-01 -1.09384990e+00 2.54797995e-01
8.54725897e-01 -7.25126117e-02 -2.99158424e-01 1.04659474e+00
2.57611275e-01 -1.37949750e-01 -5.76244473e-01 5.16223252e-01
6.13079667e-01 -8.77221704e-01 3.02484930e-01 -5.40280975e-02
1.20306206e+00 8.49776343e-02 -1.30583486e-02 4.66506958e-01
3.83674681e-01 -8.57728660e-01 4.32617992e-01 5.85606061e-02
6.93134964e-01 -8.33083212e-01 1.16643441e+00 8.20709020e-02
-1.04901183e+00 3.20153862e-01 -2.41853505e-01 -3.42173040e-01
3.12744558e-01 1.05186951e+00 -9.16262925e-01 2.34897777e-01
4.95081872e-01 1.09769928e+00 -8.53974462e-01 4.14877236e-01
-6.26021326e-01 1.12129819e+00 -1.31976515e-01 -4.34394836e-01
2.14823917e-01 -1.48098290e-01 2.10822523e-01 1.54795563e+00
-1.77699812e-02 -2.46226251e-01 3.73200238e-01 8.26519057e-02
-2.17955172e-01 8.09357226e-01 -5.78308463e-01 1.64300516e-01
-7.42350444e-02 1.78808236e+00 -1.28930783e+00 -5.12216568e-01
-8.39461744e-01 8.04406106e-01 8.41854438e-02 5.46167605e-02
-3.61604840e-01 -5.00860929e-01 3.40832621e-01 -5.98415583e-02
4.82185453e-01 1.71510652e-01 -8.01525533e-01 -1.49540806e+00
8.20553973e-02 -1.08659863e+00 3.24899018e-01 -3.94210875e-01
-1.52663076e+00 8.16949010e-01 -2.40502104e-01 -1.44934356e+00
-4.40143734e-01 -6.39412403e-01 -7.36559212e-01 5.47418118e-01
-1.96621346e+00 -1.35863578e+00 4.14277427e-02 1.74470201e-01
7.16212153e-01 -5.01177013e-01 9.57614481e-01 1.07062936e-01
-3.38579774e-01 5.54716945e-01 -2.55088955e-01 3.93442139e-02
1.12773228e+00 -1.56620765e+00 5.31387508e-01 7.74858892e-01
2.85381913e-01 7.20538974e-01 7.15578020e-01 -7.36263096e-01
-8.42828751e-01 -1.27948213e+00 1.41959906e+00 -8.28369498e-01
7.96875119e-01 -4.62271750e-01 -5.20337820e-01 4.04571265e-01
5.98475575e-01 -1.98773861e-01 1.50824797e+00 4.58767653e-01
-2.97697306e-01 3.11510880e-02 -8.53919864e-01 5.75510740e-01
4.98353213e-01 -4.97876614e-01 -4.48330969e-01 5.20170212e-01
8.00428391e-01 1.91794224e-02 -4.92885828e-01 5.93852699e-01
3.25046539e-01 -6.45375609e-01 5.59192598e-01 -4.84978795e-01
9.53579366e-01 -4.56504464e-01 5.23575172e-02 -1.30957007e+00
1.15351744e-01 -4.46656138e-01 4.87188660e-02 1.47411728e+00
1.00367749e+00 -4.66907561e-01 9.68094945e-01 4.45733637e-01
3.51603068e-02 -1.02008414e+00 -2.80027777e-01 -2.64813244e-01
4.05595580e-04 -7.86450088e-01 3.81752998e-01 8.18423092e-01
3.32734644e-01 1.16012597e+00 -3.20745260e-01 -2.61677831e-01
3.43870580e-01 4.01941627e-01 8.09880853e-01 -8.57505322e-01
-2.87438750e-01 -3.75380814e-01 -1.25865400e-01 -7.41268933e-01
3.86905670e-01 -1.00448811e+00 3.40380698e-01 -1.56308234e+00
5.97199202e-01 -3.34145993e-01 -4.99539822e-01 6.60291791e-01
-8.96608233e-01 1.02532387e+00 -1.37372717e-01 -2.99133407e-03
-1.10420239e+00 5.46187222e-01 8.13949943e-01 -1.88275725e-01
-1.85815379e-01 3.66894096e-01 -1.44524372e+00 9.30179358e-01
9.89323318e-01 -7.88268328e-01 -5.25542200e-01 -4.14591916e-02
8.94988894e-01 -7.84504235e-01 -3.07379067e-01 -4.38903749e-01
-8.94597694e-02 -5.50181009e-02 -1.15507111e-01 -3.77268225e-01
-2.58019343e-02 -5.80479860e-01 -6.81894302e-01 3.19074951e-02
-5.97275078e-01 1.03127949e-01 -1.80024385e-01 5.09102166e-01
-4.60065812e-01 -5.59845030e-01 4.20322865e-01 -1.75933450e-01
-4.46703017e-01 2.18334466e-01 -5.88556349e-01 1.97123721e-01
8.13781559e-01 1.36912510e-01 7.98900500e-02 -5.68121135e-01
-3.35575998e-01 1.36471376e-01 4.85875756e-01 6.22554719e-01
4.62450057e-01 -1.11809015e+00 -8.09351146e-01 4.61065583e-02
5.61680377e-01 -8.42870027e-02 -4.49802309e-01 4.66798544e-01
-2.06428051e-01 3.34333390e-01 5.13798535e-01 -3.30731302e-01
-1.26746428e+00 2.27484778e-01 -2.09721401e-01 -6.82149589e-01
-1.82593286e-01 6.95297897e-01 4.99849841e-02 -8.76149714e-01
-1.85534000e-01 -4.93871011e-02 -8.97752821e-01 4.10055727e-01
6.03938341e-01 -2.76042551e-01 3.16396058e-01 -8.64631355e-01
-3.84681433e-01 7.14815021e-01 -2.65335649e-01 -4.31332499e-01
1.44618952e+00 -2.50998497e-01 -2.77884275e-01 8.08570385e-01
1.23209047e+00 4.06115532e-01 -6.03356421e-01 -2.39143729e-01
3.16996932e-01 -3.80815454e-02 -1.49399906e-01 -6.77558780e-01
-9.43504035e-01 5.72072446e-01 6.96107792e-03 4.15192246e-01
1.00400245e+00 1.25594899e-01 8.10796678e-01 4.54351932e-01
1.22867338e-01 -1.19786131e+00 2.03743204e-01 7.54092336e-01
4.46082264e-01 -1.81270874e+00 3.61174971e-01 -5.65376163e-01
-1.23653495e+00 8.30243289e-01 6.17276311e-01 -2.26325855e-01
1.02088106e+00 3.55020195e-01 7.34335065e-01 -2.77906597e-01
-9.81458068e-01 -5.32445490e-01 4.09304231e-01 3.52292717e-01
1.06152630e+00 -8.12196657e-02 -7.17810273e-01 8.62889707e-01
-4.50814277e-01 -2.91157901e-01 7.39943385e-01 1.07167375e+00
-3.54502380e-01 -1.68478966e+00 1.03008393e-02 9.82676506e-01
-1.17962635e+00 -5.96742451e-01 -5.84435225e-01 1.92225069e-01
-7.94838592e-02 1.32455575e+00 -1.56242326e-01 -3.26606989e-01
1.83393955e-01 1.57680791e-02 -1.42790332e-01 -1.43555045e+00
-1.02492619e+00 1.93530276e-01 3.73772144e-01 -4.47264966e-03
-9.95520830e-01 -8.27013075e-01 -1.08143997e+00 3.18592459e-01
-5.93149722e-01 3.13891530e-01 7.93075383e-01 1.29202008e+00
2.88251251e-01 3.33776861e-01 1.04333520e+00 -4.83987123e-01
-1.94012806e-01 -1.14446795e+00 -4.08258647e-01 5.41565239e-01
2.14075625e-01 -2.73749620e-01 -5.65736294e-01 4.12335753e-01] | [11.332261085510254, 6.7481842041015625] |
0b5f839b-dddb-422e-897e-87eb71f6f778 | mixed-type-wafer-classification-for-low | 2303.13974 | null | https://arxiv.org/abs/2303.13974v1 | https://arxiv.org/pdf/2303.13974v1.pdf | Mixed-Type Wafer Classification For Low Memory Devices Using Knowledge Distillation | Manufacturing wafers is an intricate task involving thousands of steps. Defect Pattern Recognition (DPR) of wafer maps is crucial for determining the root cause of production defects, which may further provide insight for yield improvement in wafer foundry. During manufacturing, various defects may appear standalone in the wafer or may appear as different combinations. Identifying multiple defects in a wafer is generally harder compared to identifying a single defect. Recently, deep learning methods have gained significant traction in mixed-type DPR. However, the complexity of defects requires complex and large models making them very difficult to operate on low-memory embedded devices typically used in fabrication labs. Another common issue is the unavailability of labeled data to train complex networks. In this work, we propose an unsupervised training routine to distill the knowledge of complex pre-trained models to lightweight deployment-ready models. We empirically show that this type of training compresses the model without sacrificing accuracy despite being up to 10 times smaller than the teacher model. The compressed model also manages to outperform contemporary state-of-the-art models. | ['Srivatsan K', 'Anurima Dey', 'Nitish Shukla'] | 2023-03-24 | null | null | null | null | ['type'] | ['speech'] | [ 3.02213252e-01 4.38325182e-02 -2.15988055e-01 -4.70352054e-01
-7.02884734e-01 -5.10776877e-01 7.46409819e-02 5.03746271e-01
2.79247910e-01 4.09986377e-01 -5.25413215e-01 -6.19545579e-01
-5.07513024e-02 -7.71622777e-01 -7.84029186e-01 -4.47026402e-01
2.75056422e-01 8.23113918e-01 2.67106500e-02 7.63463303e-02
4.28959996e-01 9.48207796e-01 -1.43390274e+00 5.98360121e-01
5.23597836e-01 1.15341234e+00 4.70322281e-01 6.66074753e-01
-1.54972285e-01 5.13694286e-01 -8.58959138e-01 -1.61038026e-01
2.32182890e-01 -6.39344603e-02 -2.85706758e-01 2.69448429e-01
4.78213549e-01 -3.05599183e-01 -1.57896489e-01 8.36820841e-01
1.54437125e-01 -3.84783179e-01 5.76187432e-01 -7.61437178e-01
-1.26585126e-01 3.72749358e-01 -4.45254892e-01 -5.47303585e-03
-3.00635789e-02 3.59062403e-02 6.93310320e-01 -7.64244854e-01
4.18732047e-01 1.03895938e+00 5.51737070e-01 3.61296982e-01
-1.04473472e+00 -5.92060983e-01 -1.32850245e-01 -2.23047927e-01
-1.06849802e+00 -5.99465132e-01 6.83622301e-01 -3.07255775e-01
1.20054770e+00 2.93145329e-02 3.02887112e-01 9.15288091e-01
5.50018072e-01 5.58231175e-01 6.78543448e-01 -4.68853384e-01
3.95101964e-01 -5.66361248e-02 2.22015589e-01 1.02221894e+00
7.02497900e-01 -1.65347144e-01 -4.99113321e-01 -8.72034058e-02
9.65792000e-01 6.25397146e-01 3.10873002e-01 -2.98787653e-01
-8.02644610e-01 6.41769171e-01 2.84054399e-01 4.30663258e-01
-1.67375281e-01 2.45327070e-01 2.85185456e-01 4.42694157e-01
2.41002947e-01 8.82128477e-01 -8.15764487e-01 -3.15124273e-01
-9.82453644e-01 -6.08608909e-02 9.47462678e-01 1.00293136e+00
9.23974395e-01 7.19334781e-02 2.79192090e-01 7.45553493e-01
4.77985069e-02 2.86367685e-01 1.88462093e-01 -7.38208652e-01
4.48333949e-01 8.83778214e-01 1.31185770e-01 -9.41144943e-01
-2.92754441e-01 -5.45790792e-01 -9.70689833e-01 1.62024826e-01
2.42337674e-01 -8.69225487e-02 -1.26158190e+00 8.57880056e-01
2.52361652e-02 -9.04283077e-02 -2.40125403e-01 2.17652813e-01
3.06744605e-01 6.82761550e-01 -4.31620479e-01 6.67207092e-02
9.50224280e-01 -8.41077983e-01 -5.49309015e-01 -5.77001929e-01
7.45008886e-01 -7.72217751e-01 6.39414608e-01 8.64819884e-01
-1.10828686e+00 -6.35264099e-01 -1.54225552e+00 -2.24718871e-03
-5.05534470e-01 5.69726229e-01 1.03951430e+00 7.74819553e-01
-8.87078106e-01 1.22185075e+00 -1.18786550e+00 -1.97503433e-01
6.76459730e-01 8.44956636e-01 -2.83522040e-01 -8.42557609e-01
-1.64164186e-01 6.73697889e-01 7.91942179e-02 2.34548628e-01
-9.88744617e-01 -9.72176075e-01 -8.23083043e-01 2.25009605e-01
4.01279271e-01 -2.85516053e-01 1.15581262e+00 -4.73305762e-01
-1.39669645e+00 5.06940603e-01 -4.55933154e-01 -2.56039530e-01
1.37093097e-01 -9.93449390e-02 -4.84711170e-01 -9.14315432e-02
-2.08662271e-01 4.96188521e-01 1.14568341e+00 -1.20087147e+00
-5.21353900e-01 -3.38928163e-01 4.72040214e-02 -4.31680411e-01
-2.59539366e-01 -1.58407852e-01 -5.12718558e-01 -1.45634696e-01
5.82040071e-01 -6.83790803e-01 -3.15020591e-01 -7.35252798e-02
-5.43194473e-01 -4.68133949e-02 1.03685272e+00 -6.15991592e-01
9.20491338e-01 -2.22881103e+00 -3.31384659e-01 2.95095205e-01
3.40582997e-01 2.17049763e-01 -1.35313030e-02 4.61023301e-01
-6.96431175e-02 1.33184269e-01 3.75751853e-02 -4.16453123e-01
-4.23445888e-02 2.29474053e-01 7.60143739e-04 2.71422803e-01
6.41085744e-01 7.64491141e-01 -7.05538869e-01 -2.49690786e-02
2.10183933e-01 8.65324493e-03 -5.15504897e-01 2.57354863e-02
-3.54521960e-01 2.15863690e-01 -2.93476343e-01 1.35504568e+00
6.63417339e-01 -5.52757561e-01 4.23457742e-01 -3.02840292e-01
1.97658360e-01 3.79456460e-01 -8.82410526e-01 1.67041373e+00
-6.05412543e-01 5.78799546e-01 1.44954458e-01 -1.13764262e+00
1.18278480e+00 2.22445074e-02 2.01922104e-01 -4.79186773e-01
2.13515520e-01 5.15712559e-01 2.49722049e-01 -2.73405761e-01
5.58062434e-01 -1.30289435e-01 -1.46625146e-01 2.91157097e-01
-5.78320865e-03 -4.07997519e-01 2.70296663e-01 -2.58875191e-01
1.83923483e+00 -3.77210230e-01 -2.25045711e-01 -7.92547539e-02
-1.46932483e-01 -1.98061690e-01 6.82983816e-01 6.37747586e-01
2.54106522e-01 4.87346828e-01 6.73227251e-01 -6.01233304e-01
-1.05395103e+00 -1.06405699e+00 5.29193059e-02 2.56221712e-01
-1.01927713e-01 -3.30402672e-01 -4.17515248e-01 -7.44835913e-01
2.16169268e-01 5.15859485e-01 -2.89289445e-01 -1.75533205e-01
-5.05140603e-01 -7.24641442e-01 7.76650086e-02 8.50776017e-01
1.23572625e-01 -7.10924327e-01 -3.65943521e-01 2.14375138e-01
5.18442631e-01 -1.10109341e+00 -3.67724188e-02 6.75690413e-01
-1.36645484e+00 -9.68197346e-01 -1.25992462e-01 -8.99224639e-01
1.36820364e+00 2.54220277e-01 1.22970080e+00 2.58862764e-01
-8.16189826e-01 -6.36725426e-02 -1.64629966e-01 -5.80687225e-01
-6.46482408e-01 1.72746360e-01 2.11957637e-02 -3.55414063e-01
4.81682420e-01 -5.31657040e-01 -3.18101496e-01 3.09571713e-01
-8.26571822e-01 -1.19801857e-01 1.04601610e+00 8.48296583e-01
6.55540407e-01 8.28554869e-01 5.43275237e-01 -9.90201116e-01
2.81005144e-01 -2.55023807e-01 -7.29296148e-01 1.81069672e-01
-3.37008089e-01 -4.74277623e-02 5.38355052e-01 -4.12676543e-01
-6.00603759e-01 2.80847400e-01 -9.65090119e-04 -5.98040283e-01
-1.80567011e-01 5.48579335e-01 -2.66722977e-01 -2.22203001e-01
1.70854017e-01 -3.22327971e-01 -5.09009175e-02 -5.22146285e-01
-3.28501463e-01 5.35966277e-01 8.23182091e-02 -5.35884500e-01
7.97113955e-01 3.36383104e-01 2.08542258e-01 -9.02713835e-01
-7.34338880e-01 -3.50424647e-01 -6.91004515e-01 4.66571711e-02
4.15282935e-01 -6.86801732e-01 -5.73729098e-01 3.05732727e-01
-1.13046420e+00 -4.59840506e-01 -1.99049756e-01 1.69898067e-02
-9.63855386e-02 1.06897764e-01 -8.32647264e-01 -5.45727015e-01
3.65214646e-02 -1.22145426e+00 1.20895290e+00 1.76623136e-01
-3.43162000e-01 -8.43497694e-01 -6.24798477e-01 4.17660356e-01
4.67218071e-01 1.14782229e-01 1.45354319e+00 -5.27554512e-01
-9.76229608e-01 -8.89014244e-01 -1.91578597e-01 4.64444935e-01
6.60395920e-01 5.57523891e-02 -8.73630762e-01 -4.50020581e-01
1.70685291e-01 -2.80630201e-01 7.53790617e-01 4.10591274e-01
1.59226644e+00 3.41201842e-01 -6.48349166e-01 2.14223400e-01
1.45072353e+00 4.13823992e-01 5.62078238e-01 -2.40453467e-01
6.83207631e-01 4.83024269e-01 6.02659464e-01 2.37404838e-01
-2.28931412e-01 2.12897420e-01 4.68322754e-01 -1.32102340e-01
1.24101289e-01 -1.12045221e-01 2.64732301e-01 1.11841536e+00
3.52135986e-01 -3.07696164e-01 -1.03046572e+00 5.47295809e-01
-1.33820057e+00 -3.21156561e-01 2.27360398e-01 2.13440824e+00
4.35727090e-01 6.28785849e-01 -6.13775253e-01 2.78116405e-01
6.12241149e-01 -1.15330681e-01 -7.65612245e-01 -4.78331715e-01
2.96507180e-01 5.11633873e-01 5.71460485e-01 2.09833067e-02
-9.26783860e-01 7.04732120e-01 6.34408998e+00 8.27740550e-01
-1.22079682e+00 -1.20230680e-02 9.77187455e-01 -2.62273490e-01
-1.51190301e-02 -9.41435620e-02 -1.17001092e+00 2.58336633e-01
1.19941974e+00 4.66446072e-01 2.34613389e-01 9.58385289e-01
-1.78412527e-01 -3.08223367e-01 -1.83694124e+00 1.07834673e+00
1.13628320e-01 -1.66688466e+00 -2.23342210e-01 3.30055803e-01
9.36667860e-01 -1.87648252e-01 1.97355766e-02 2.85857767e-01
2.78447002e-01 -1.26640177e+00 2.45848149e-01 3.02496970e-01
6.81112766e-01 -8.88963580e-01 9.42494154e-01 1.55212745e-01
-8.61422658e-01 -4.67513055e-01 -8.40352714e-01 -2.91105080e-02
-6.56404868e-02 1.22580028e+00 -1.12191308e+00 5.50044239e-01
4.42959428e-01 6.95881248e-01 -3.91996264e-01 8.04277122e-01
2.14923918e-01 5.04482269e-01 -4.21608746e-01 1.78392269e-02
2.59750728e-02 8.00578743e-02 -5.23365997e-02 7.41555631e-01
6.90707684e-01 -4.05861646e-01 2.21669257e-01 9.53565121e-01
-4.35158551e-01 -5.51092923e-01 -6.48518622e-01 -7.46886134e-01
4.49565858e-01 1.22424507e+00 -1.12674856e+00 -8.13781023e-02
-2.52717912e-01 8.69406223e-01 1.77593559e-01 -1.00952476e-01
-4.45952088e-01 -2.05314428e-01 6.14446223e-01 3.95080328e-01
7.53721356e-01 -5.04904807e-01 -4.71604705e-01 -6.47682488e-01
2.52384543e-01 -8.93407702e-01 -3.42625350e-01 -4.39321220e-01
-1.34676158e+00 3.24774534e-01 -4.88574445e-01 -1.12990403e+00
-4.65703122e-02 -1.05931294e+00 -6.32711887e-01 7.40864813e-01
-1.46097374e+00 -8.28014135e-01 -2.71932006e-01 -2.93730915e-01
7.61387646e-01 -2.67034769e-01 9.29572821e-01 4.27924126e-01
-8.58709335e-01 5.35143793e-01 2.33358398e-01 -1.27668828e-01
6.07686639e-01 -9.83214021e-01 5.62747538e-01 6.60327256e-01
-1.31595656e-01 6.14152491e-01 4.40147787e-01 -9.67681050e-01
-2.07427692e+00 -1.23644006e+00 7.62200296e-01 -4.51542199e-01
5.68458736e-01 -5.69430590e-01 -6.63892090e-01 5.95172703e-01
-2.40037307e-01 1.30891040e-01 5.51128864e-01 3.23039204e-01
-1.34713249e-02 -4.63904589e-01 -1.22196448e+00 3.41946095e-01
7.31639624e-01 -5.20444572e-01 1.71145767e-01 5.81792712e-01
6.78554595e-01 -4.98901516e-01 -9.05763090e-01 5.34557819e-01
2.54454911e-01 -7.66219676e-01 7.13382125e-01 -1.84715480e-01
8.61056864e-01 -4.89291847e-02 -2.38623410e-01 -1.10225308e+00
-1.00359820e-01 -4.31945026e-01 -3.79815936e-01 1.12794518e+00
6.40145361e-01 -4.66182470e-01 1.41821098e+00 6.42530441e-01
-4.86347258e-01 -1.02982950e+00 -6.64140821e-01 -9.55755591e-01
-2.31033862e-01 -3.58087748e-01 6.17954671e-01 8.28110218e-01
-3.70667994e-01 2.22044289e-02 8.69238004e-02 5.19815922e-01
3.32306534e-01 4.99747843e-02 5.81841290e-01 -1.61713338e+00
-5.69350421e-01 -1.39456853e-01 -5.05841792e-01 -9.63654816e-01
1.05829872e-02 -5.95870912e-01 2.40156353e-01 -1.46135509e+00
-1.42810792e-01 -1.01734090e+00 -3.40811610e-01 5.32421350e-01
2.43039683e-01 -2.20963079e-02 -1.79669648e-01 -1.07324034e-01
-3.51709247e-01 1.15678109e-01 1.05237889e+00 -5.64841986e-01
-1.06732301e-01 2.35603135e-02 -5.98035514e-01 6.95013881e-01
9.35335815e-01 -6.18120313e-01 -4.30060476e-01 -6.80944920e-01
2.51831919e-01 1.95950136e-01 1.59256551e-02 -1.39458823e+00
3.18085760e-01 1.96149588e-01 6.56605244e-01 -9.34646368e-01
5.99019945e-01 -8.07181597e-01 1.03845380e-01 5.20273626e-01
2.17588663e-01 1.24226399e-02 5.32877207e-01 5.99757373e-01
-1.61216021e-01 -7.20373094e-01 5.30971229e-01 -2.90465713e-01
-2.70743370e-01 2.30742231e-01 -1.93288311e-01 -7.47158527e-01
8.00451756e-01 -2.85023987e-01 -1.56113476e-01 1.38103619e-01
-4.65435952e-01 -7.36117363e-02 6.04629457e-01 3.15929681e-01
6.86972260e-01 -8.26412141e-01 -2.11026505e-01 5.49072504e-01
-1.19740054e-01 4.07614440e-01 4.78717417e-01 4.34459537e-01
-7.91008234e-01 5.43071508e-01 -5.32915294e-02 -5.75881898e-01
-1.11757815e+00 3.53140503e-01 7.41912946e-02 -7.73982346e-01
-4.98176306e-01 1.02902746e+00 -7.29622273e-03 -3.52392793e-01
1.59772947e-01 -7.53788769e-01 4.53748673e-01 -2.34369308e-01
5.30260444e-01 1.94946393e-01 7.18902290e-01 4.11229879e-01
-5.99614419e-02 2.66275108e-01 -7.38790274e-01 5.46522379e-01
1.58705950e+00 4.60463613e-01 -2.77743399e-01 5.64147711e-01
1.04942966e+00 -1.89243332e-01 -1.21932769e+00 7.81935155e-02
1.11100130e-01 -2.99209177e-01 2.16102391e-01 -7.91940212e-01
-1.19036543e+00 9.78212118e-01 5.05223691e-01 1.31812319e-01
9.32899356e-01 -1.41486362e-01 9.47621465e-01 1.03790915e+00
7.22469330e-01 -1.11194611e+00 3.32871735e-01 3.64899337e-01
5.96232772e-01 -1.32388222e+00 1.33499756e-01 -8.38889182e-01
-1.01537099e-02 1.34705174e+00 5.31886816e-01 -1.44842759e-01
6.90640986e-01 8.29383254e-01 -3.32159519e-01 -4.44945931e-01
-7.33968079e-01 5.38330138e-01 -2.09685132e-01 5.53484976e-01
2.46635169e-01 -5.43318205e-02 5.30406296e-01 4.00594592e-01
1.82643440e-02 1.26734106e-02 4.71027762e-01 1.58047628e+00
-4.17184949e-01 -1.54115283e+00 -1.95835382e-01 1.15596652e+00
-2.81152457e-01 -4.49012630e-02 -9.18645561e-02 5.57842493e-01
2.24584907e-01 9.75423038e-01 4.01735634e-01 -3.50554943e-01
7.93098211e-02 4.49725725e-02 8.71512055e-01 -1.04922903e+00
-4.09851819e-01 -1.35985807e-01 1.68462053e-01 -5.03022730e-01
1.36996955e-01 -6.26225591e-01 -9.74122226e-01 -2.58210152e-01
-3.94377470e-01 -3.45489502e-01 1.13983405e+00 1.08207965e+00
5.52539587e-01 1.15248549e+00 6.08148634e-01 -1.06486392e+00
-4.82435137e-01 -8.51644993e-01 -7.64210999e-01 -3.42882574e-01
1.71998233e-01 -6.70992196e-01 -1.50807172e-01 5.30258752e-02] | [7.3041768074035645, 2.0023796558380127] |
484e6e4b-c66b-4a0e-a397-ffba5a87afbd | unsupervised-feature-learning-for | 1904.04221 | null | https://arxiv.org/abs/1904.04221v2 | https://arxiv.org/pdf/1904.04221v2.pdf | Unsupervised Feature Learning for Environmental Sound Classification Using Weighted Cycle-Consistent Generative Adversarial Network | In this paper we propose a novel environmental sound classification approach incorporating unsupervised feature learning from codebook via spherical $K$-Means++ algorithm and a new architecture for high-level data augmentation. The audio signal is transformed into a 2D representation using a discrete wavelet transform (DWT). The DWT spectrograms are then augmented by a novel architecture for cycle-consistent generative adversarial network. This high-level augmentation bootstraps generated spectrograms in both intra and inter class manners by translating structural features from sample to sample. A codebook is built by coding the DWT spectrograms with the speeded-up robust feature detector (SURF) and the K-Means++ algorithm. The Random Forest is our final learning algorithm which learns the environmental sound classification task from the clustered codewords in the codebook. Experimental results in four benchmarking environmental sound datasets (ESC-10, ESC-50, UrbanSound8k, and DCASE-2017) have shown that the proposed classification approach outperforms the state-of-the-art classifiers in the scope, including advanced and dense convolutional neural networks such as AlexNet and GoogLeNet, improving the classification rate between 3.51% and 14.34%, depending on the dataset. | ['Alessandro Lameiras Koerich', 'Patrick Cardinal', 'Mohammad Esmaeilpour'] | 2019-04-08 | null | null | null | null | ['environmental-sound-classification', 'sound-classification'] | ['audio', 'audio'] | [ 3.38757008e-01 -1.31473094e-01 5.37851989e-01 -2.35773996e-01
-8.21422279e-01 -4.23032820e-01 5.11699498e-01 -8.00998658e-02
-4.58236575e-01 3.08131278e-01 1.72531664e-01 -8.96220505e-02
9.77678448e-02 -1.00152898e+00 -8.16117764e-01 -9.47759032e-01
-4.55653310e-01 6.35800585e-02 1.32931665e-01 -2.48674732e-02
1.03669159e-01 4.53822017e-01 -2.11992717e+00 4.21487182e-01
6.24974310e-01 1.28022003e+00 8.67994204e-02 1.11216569e+00
-1.09791785e-01 6.11071169e-01 -6.84520066e-01 -1.61646590e-01
4.64485198e-01 -2.96224028e-01 -4.49846268e-01 -3.48601133e-01
5.17079353e-01 2.95384713e-02 -3.06674153e-01 9.22220588e-01
7.62419045e-01 2.93911904e-01 6.22558892e-01 -1.36702204e+00
-5.66144049e-01 7.15504289e-01 -1.52792662e-01 6.30251169e-02
1.35590583e-01 1.91558719e-01 9.97602582e-01 -1.10956478e+00
3.62490743e-01 1.15666378e+00 1.09668756e+00 4.25828576e-01
-1.04124045e+00 -1.21093357e+00 -1.97575942e-01 3.90976518e-01
-1.52560258e+00 -4.70789850e-01 1.06682777e+00 -4.73996013e-01
1.24670434e+00 3.61338258e-01 9.78171825e-01 8.92292261e-01
-2.90950388e-02 4.93428439e-01 1.08789980e+00 -7.12257564e-01
6.31064892e-01 -1.51015773e-01 -1.31255180e-01 7.44008601e-01
-1.84544832e-01 4.85983133e-01 -6.86648548e-01 -3.27673584e-01
3.50241154e-01 -2.25203246e-01 -1.10249192e-01 4.90016602e-02
-7.20656335e-01 1.01281309e+00 6.74054503e-01 1.81884542e-01
-4.24304277e-01 4.65619653e-01 6.26715660e-01 2.54601210e-01
6.87086821e-01 3.25221181e-01 -4.15901512e-01 -1.21038489e-01
-1.24506581e+00 3.15281689e-01 6.90392792e-01 7.75327623e-01
7.13036239e-01 8.10304821e-01 1.54483661e-01 8.15596104e-01
5.25152981e-01 6.73430264e-01 8.44094336e-01 -5.95136166e-01
3.14921618e-01 2.64232218e-01 -2.68853605e-01 -1.09129965e+00
-3.43313605e-01 -6.81855500e-01 -1.04990196e+00 4.09260750e-01
-9.02994275e-02 -6.80955052e-02 -1.20957649e+00 1.46709132e+00
3.13306034e-01 9.33622301e-01 1.55834481e-01 4.16442931e-01
1.21807504e+00 9.43363786e-01 -5.74659482e-02 7.66206980e-02
9.83761787e-01 -8.95138204e-01 -6.28649592e-01 -2.46403813e-02
2.92519778e-01 -7.66942620e-01 9.80553508e-01 5.01473546e-01
-6.62386715e-01 -1.04323637e+00 -1.24375188e+00 3.42430264e-01
-6.49105012e-01 3.36728424e-01 3.24964970e-01 9.27561045e-01
-1.12670457e+00 4.18902308e-01 -8.27370405e-01 9.61002260e-02
5.57826042e-01 3.83396596e-01 -1.98585376e-01 2.21973047e-01
-1.37699628e+00 3.49795282e-01 3.80250245e-01 -7.52702579e-02
-1.26164877e+00 -9.56923783e-01 -8.97862315e-01 -3.57289277e-02
-3.77840072e-01 -2.75391400e-01 1.04193807e+00 -7.72527218e-01
-1.74239135e+00 5.65427244e-01 1.86773911e-01 -8.70273173e-01
2.48624653e-01 -1.16086543e-01 -6.30868077e-01 1.71990365e-01
6.35902435e-02 7.23627687e-01 1.43890488e+00 -1.13340139e+00
-6.42798901e-01 -1.08367458e-01 -4.78111178e-01 -1.34882271e-01
-4.46419269e-01 -9.27094072e-02 1.56692982e-01 -1.10064459e+00
1.66017100e-01 -9.71804023e-01 -2.02200457e-01 -1.59030557e-01
-3.79272550e-01 -1.39375553e-01 1.19791496e+00 -6.49640977e-01
1.18786645e+00 -2.46536541e+00 -1.04901910e-01 3.75484496e-01
-7.96969756e-02 4.60750043e-01 -1.85285166e-01 3.22342485e-01
-4.46272671e-01 4.02037911e-02 -4.77664977e-01 -6.58020556e-01
6.01219758e-03 7.16789439e-02 -6.29269421e-01 4.99103099e-01
2.91727632e-01 4.97638196e-01 -8.50713193e-01 -7.72416890e-02
2.66492486e-01 7.07960606e-01 -8.76697719e-01 3.76954466e-01
-1.30073383e-01 2.49682844e-01 5.89240454e-02 5.13754725e-01
7.24359870e-01 4.79828060e-01 -4.46215779e-01 -3.83953787e-02
-1.82962134e-01 2.88187176e-01 -1.55937350e+00 1.75633836e+00
-7.66928136e-01 6.82574928e-01 -7.72545487e-02 -1.16472805e+00
1.33332860e+00 3.99489224e-01 3.33416492e-01 -3.17563117e-01
-5.79871759e-02 3.49284112e-01 -1.21479720e-01 -4.63856459e-01
5.08135796e-01 -5.72599247e-02 -1.53619513e-01 3.22642624e-01
5.06519437e-01 -6.83083713e-01 -2.64558464e-01 -1.37037739e-01
8.76530886e-01 2.21490003e-02 -8.89484808e-02 -1.74337059e-01
4.77669209e-01 -1.32827729e-01 2.48638913e-01 6.09497368e-01
1.26565192e-02 6.38833344e-01 -2.61111706e-01 -4.79556024e-01
-1.09117794e+00 -7.79940307e-01 -4.24461961e-01 1.16938066e+00
-5.53706288e-01 -3.26464415e-01 -1.02732468e+00 -4.12271887e-01
-4.04533744e-02 7.95574665e-01 -9.05396640e-01 -4.59934413e-01
-4.54411328e-01 -6.55758798e-01 1.15398431e+00 5.90459108e-01
7.90345490e-01 -1.36551785e+00 -5.52435875e-01 5.59877157e-01
4.06046510e-02 -1.01837671e+00 -1.71644583e-01 7.20590830e-01
-6.38324320e-01 -7.85274565e-01 -3.74928176e-01 -8.93674552e-01
2.34074503e-01 -1.70632154e-01 8.75777185e-01 -9.32707116e-02
-5.72141469e-01 2.94176579e-01 -7.71789789e-01 -8.45913529e-01
-7.96908677e-01 -4.97490615e-02 3.05951357e-01 3.14825803e-01
1.84820294e-01 -8.92009556e-01 -4.18104738e-01 -2.08793700e-01
-8.65050673e-01 -2.13380039e-01 2.96531975e-01 9.63460386e-01
7.18059659e-01 2.18251467e-01 6.77987576e-01 -3.06981772e-01
3.72112960e-01 -3.79172295e-01 -4.20313954e-01 -3.92640740e-01
-3.66511643e-01 -2.81801634e-02 9.91044641e-01 -4.37999249e-01
-6.45742476e-01 4.53843951e-01 -6.80613339e-01 -6.29447818e-01
-4.26495194e-01 1.51291072e-01 -2.83779353e-02 9.64649580e-03
9.58348870e-01 7.79407322e-01 -2.29766041e-01 -5.21396875e-01
6.24373376e-01 1.12006557e+00 8.76368761e-01 -2.47662127e-01
1.18298483e+00 5.05435646e-01 -2.56974399e-01 -1.14498222e+00
-6.12908661e-01 -3.48648697e-01 -7.65568376e-01 -2.25915030e-01
1.00735748e+00 -1.16179168e+00 -7.81634226e-02 8.02828074e-01
-1.03952432e+00 -2.10839316e-01 -8.37490141e-01 4.38144803e-01
-5.91310799e-01 2.21561491e-02 -2.44557887e-01 -8.53072464e-01
-6.68414474e-01 -1.19171023e+00 1.06278241e+00 9.51405987e-03
-6.41680285e-02 -6.50323808e-01 2.93549865e-01 -9.93485376e-02
5.03686011e-01 4.32488918e-01 7.62201369e-01 -8.16892743e-01
-1.28675237e-01 -2.25192398e-01 3.96474272e-01 8.98211837e-01
-6.67599402e-03 2.38315225e-01 -1.54562747e+00 -2.23838225e-01
-1.64187640e-01 -3.06533962e-01 1.08816445e+00 1.91740140e-01
1.38256252e+00 -3.80556941e-01 3.32633313e-03 1.00237393e+00
1.36373329e+00 3.30431521e-01 6.00051939e-01 3.92684013e-01
4.56247538e-01 -1.55072296e-02 3.71873438e-01 4.93237853e-01
3.35347056e-02 5.43444514e-01 8.35578024e-01 1.11544982e-01
-5.75748265e-01 -3.34387273e-01 4.63036835e-01 1.13456464e+00
-7.89672285e-02 9.10348520e-02 -8.88207555e-01 6.82134271e-01
-1.20737314e+00 -1.03272212e+00 -1.67963356e-01 1.87858331e+00
8.77259910e-01 -8.41061771e-03 -7.20012784e-02 1.17716217e+00
5.98993242e-01 1.29964471e-01 -2.44042829e-01 -4.64811265e-01
1.69512421e-01 1.10242069e+00 7.81260207e-02 2.73807347e-01
-1.41250908e+00 9.68665957e-01 5.67575788e+00 9.78687286e-01
-1.31755102e+00 2.48545557e-01 2.26096004e-01 1.33385822e-01
1.21247746e-01 -5.63312471e-01 -7.92382538e-01 5.25052965e-01
1.33292162e+00 2.30788812e-01 4.17702317e-01 1.06644142e+00
-2.22623590e-02 4.00034964e-01 -7.15336561e-01 9.90135431e-01
1.76545843e-01 -1.40301538e+00 -2.51448769e-02 -3.15896541e-01
8.42951179e-01 2.59795964e-01 4.11185503e-01 4.49420065e-01
3.66993129e-01 -9.53883946e-01 1.09681499e+00 4.08941060e-01
1.17187107e+00 -1.05746222e+00 5.79613090e-01 1.62982374e-01
-1.60257208e+00 -3.95864785e-01 -4.83936906e-01 1.02484003e-01
-4.10725564e-01 4.93351340e-01 -9.02077317e-01 4.48687166e-01
1.18960226e+00 5.16573489e-01 -4.61644560e-01 9.54966724e-01
-1.38025627e-01 1.22225261e+00 -4.48141873e-01 -5.94728813e-02
4.46980417e-01 1.56781644e-01 5.97325683e-01 1.51021409e+00
5.24197638e-01 -6.48074271e-03 -2.30333865e-01 4.76798117e-01
-1.65227160e-01 1.80054039e-01 -5.63912988e-01 1.22098170e-01
7.30883300e-01 1.09753895e+00 -5.23679435e-01 -3.98722410e-01
-9.40121263e-02 6.66780829e-01 -1.23177245e-02 -5.66989034e-02
-8.48957777e-01 -6.35624409e-01 8.03319216e-01 -1.41893048e-02
7.94299245e-01 -6.45885915e-02 -1.74081922e-01 -6.80081308e-01
-2.19140664e-01 -8.43506634e-01 3.77845243e-02 -6.42696440e-01
-9.71525252e-01 9.05251503e-01 -4.43160653e-01 -1.68465757e+00
-3.18729073e-01 -3.71810466e-01 -6.70806587e-01 6.30558908e-01
-1.61361241e+00 -1.13295984e+00 -5.09620368e-01 7.50695050e-01
6.65747821e-01 -8.04139018e-01 1.38317263e+00 1.77859232e-01
-1.75414607e-01 8.31246614e-01 5.83552599e-01 3.42720807e-01
3.37659895e-01 -1.27418840e+00 6.37431264e-01 6.70617402e-01
6.11191273e-01 5.18164970e-02 5.45346916e-01 -1.95673019e-01
-1.18811846e+00 -1.81804395e+00 5.45529068e-01 -1.06910780e-01
5.49033284e-01 -5.65503240e-01 -7.77456641e-01 2.37815171e-01
8.62733051e-02 5.25938690e-01 9.37754095e-01 -5.67220211e-01
-6.97513938e-01 -3.64659876e-01 -1.35174620e+00 -4.62126136e-02
8.04449141e-01 -7.39313304e-01 -6.52697206e-01 3.03187847e-01
9.43744898e-01 -4.27935362e-01 -8.69184136e-01 1.77505776e-01
5.11061251e-01 -8.84321749e-01 1.13741505e+00 -5.65181792e-01
5.80106080e-01 -3.33125830e-01 -5.88164449e-01 -1.67162895e+00
-2.54575878e-01 -5.25100946e-01 -2.29108751e-01 1.14078903e+00
2.61725783e-01 -3.63727480e-01 4.84035164e-01 -3.41463029e-01
-6.70093060e-01 -6.03487909e-01 -1.45715439e+00 -7.41832674e-01
1.32531837e-01 -1.22568715e+00 9.70423460e-01 8.93006980e-01
-4.94840622e-01 -4.46700342e-02 -2.15773493e-01 5.69260418e-01
5.64670384e-01 -9.64276418e-02 7.83737183e-01 -1.26883459e+00
-2.36631706e-01 -1.87726319e-01 -7.98935652e-01 -5.50764799e-01
2.27241278e-01 -1.16421866e+00 2.10869446e-01 -1.14218783e+00
-4.14526463e-01 -4.85696822e-01 -3.16884905e-01 6.82540357e-01
1.10273473e-01 7.50681937e-01 2.33332798e-01 -4.33647372e-02
-3.41858901e-02 7.19403327e-01 9.03032541e-01 -2.76877880e-01
-2.87630737e-01 6.89985156e-02 -1.99795187e-01 8.04601729e-01
8.62423420e-01 -8.13327551e-01 -1.20995812e-01 -1.52206361e-01
-1.78656086e-01 -3.21447104e-01 6.06872380e-01 -1.64270341e+00
2.19382226e-01 4.55783695e-01 3.54220688e-01 -5.92738390e-01
3.75197321e-01 -9.65872884e-01 1.03835657e-01 6.96704030e-01
-3.74252647e-01 -2.09361598e-01 4.01795954e-01 7.53946066e-01
-3.86871278e-01 -2.27629945e-01 1.19094169e+00 5.96466176e-02
-5.50890028e-01 2.87420213e-01 -4.91097808e-01 -5.47703281e-02
8.60646427e-01 -2.36516178e-01 1.96909681e-01 -2.64759243e-01
-1.02957070e+00 -5.13707578e-01 -2.71375179e-01 4.58687007e-01
6.92306399e-01 -1.56881011e+00 -9.10263121e-01 7.82611191e-01
-6.15558885e-02 1.67117864e-01 1.94269091e-01 2.78088033e-01
-4.87687916e-01 1.65718570e-01 -1.15081623e-01 -7.16224253e-01
-1.19065022e+00 9.15768892e-02 5.06534636e-01 -3.43095027e-02
-6.26179934e-01 1.24879766e+00 -3.38953108e-01 -5.56814253e-01
4.72964674e-01 -6.86451316e-01 -2.76326001e-01 2.23038003e-01
5.71713388e-01 3.66439521e-01 4.94522005e-01 -8.41256976e-01
-3.08153182e-01 5.87713182e-01 3.96800548e-01 -6.40395284e-02
1.91580153e+00 4.45555538e-01 7.12091103e-02 2.45784804e-01
1.55237556e+00 -6.09030835e-02 -1.10807037e+00 -2.02158421e-01
-5.19578040e-01 -2.56951183e-01 2.66294688e-01 -6.01424217e-01
-1.41624033e+00 1.15172946e+00 1.12220263e+00 4.73450631e-01
1.31507528e+00 -3.27169985e-01 7.16427982e-01 4.16744828e-01
2.40750074e-01 -9.97602463e-01 2.43382737e-01 6.58158660e-01
1.19346333e+00 -6.80975199e-01 -3.07184130e-01 7.18161315e-02
-2.69154191e-01 1.21483016e+00 2.57420987e-01 -6.77435100e-01
1.20323658e+00 7.39714146e-01 1.22704059e-01 2.93081123e-02
-5.37553012e-01 -1.68749675e-01 2.37021983e-01 9.25976753e-01
1.08373433e-01 3.30654196e-02 3.74608755e-01 9.05613363e-01
-1.02801216e+00 -3.05186331e-01 3.77730578e-01 6.71502292e-01
-5.16173184e-01 -7.57363439e-01 -5.96410871e-01 4.25710320e-01
-4.17884916e-01 -3.17510724e-01 -9.76897255e-02 6.21658385e-01
6.14754021e-01 9.01923776e-01 3.14842194e-01 -8.79660606e-01
3.54460150e-01 4.62758631e-01 -1.90258641e-02 -6.15713239e-01
-9.23112512e-01 4.36979868e-02 -2.33993620e-01 -2.90011048e-01
-4.22669262e-01 -6.87259078e-01 -1.32744515e+00 2.61523455e-01
-5.72206676e-01 3.24169993e-01 9.21759069e-01 6.24698758e-01
2.27993324e-01 8.58706295e-01 1.14021814e+00 -1.12358809e+00
-6.53986990e-01 -1.26265180e+00 -5.31479657e-01 8.79076347e-02
4.29867178e-01 -5.25781274e-01 -6.66119576e-01 4.80896443e-01] | [15.191203117370605, 5.246728897094727] |
facb32f4-91a3-46af-81cd-360eabd23f87 | dshgt-dual-supervisors-heterogeneous-graph | 2306.01376 | null | https://arxiv.org/abs/2306.01376v1 | https://arxiv.org/pdf/2306.01376v1.pdf | DSHGT: Dual-Supervisors Heterogeneous Graph Transformer -- A pioneer study of using heterogeneous graph learning for detecting software vulnerabilities | Vulnerability detection is a critical problem in software security and attracts growing attention both from academia and industry. Traditionally, software security is safeguarded by designated rule-based detectors that heavily rely on empirical expertise, requiring tremendous effort from software experts to generate rule repositories for large code corpus. Recent advances in deep learning, especially Graph Neural Networks (GNN), have uncovered the feasibility of automatic detection of a wide range of software vulnerabilities. However, prior learning-based works only break programs down into a sequence of word tokens for extracting contextual features of codes, or apply GNN largely on homogeneous graph representation (e.g., AST) without discerning complex types of underlying program entities (e.g., methods, variables). In this work, we are one of the first to explore heterogeneous graph representation in the form of Code Property Graph and adapt a well-known heterogeneous graph network with a dual-supervisor structure for the corresponding graph learning task. Using the prototype built, we have conducted extensive experiments on both synthetic datasets and real-world projects. Compared with the state-of-the-art baselines, the results demonstrate promising effectiveness in this research direction in terms of vulnerability detection performance (average F1 improvements over 10\% in real-world projects) and transferability from C/C++ to other programming languages (average F1 improvements over 11%). | ['Xi Zheng', 'Jun Yin', 'Xiaowei Huang', 'Xin Chen', 'Yuzhe Tian', 'Jianping Zhang', 'Rui Xu', 'Tiehua Zhang'] | 2023-06-02 | null | null | null | null | ['vulnerability-detection'] | ['miscellaneous'] | [ 9.44399610e-02 1.21116325e-01 -5.06677449e-01 4.29846300e-03
-4.87071484e-01 -7.89330721e-01 3.42808664e-01 6.06048167e-01
7.01625198e-02 7.38036186e-02 -1.25940964e-01 -1.16062939e+00
5.26640303e-02 -1.02909279e+00 -7.08652377e-01 -8.32355469e-02
-4.87164140e-01 -2.03497469e-01 5.62400699e-01 -3.54846776e-01
5.01306653e-01 1.08210340e-01 -1.21998405e+00 3.23374748e-01
8.62992942e-01 6.65267050e-01 -3.60954195e-01 5.23498774e-01
-3.62912804e-01 1.09582365e+00 -7.03850865e-01 -8.32987607e-01
2.39278764e-01 4.31954116e-02 -7.62994885e-01 -3.19372475e-01
4.88762528e-01 9.44611803e-03 -3.38842183e-01 1.69722056e+00
1.06163517e-01 -3.75319988e-01 2.69740671e-01 -1.42922246e+00
-9.08520341e-01 1.23017430e+00 -9.74380851e-01 1.05508432e-01
3.46870750e-01 2.66418666e-01 1.29382610e+00 -5.33527851e-01
5.27995229e-01 1.08159137e+00 9.47874248e-01 5.17312229e-01
-1.12482023e+00 -6.63759291e-01 2.15094507e-01 2.57744819e-01
-1.19957471e+00 2.34207120e-02 9.78511691e-01 -8.40680659e-01
1.47952497e+00 7.37323388e-02 2.17480451e-01 1.31002748e+00
4.03625011e-01 4.81858283e-01 7.01493979e-01 -5.40630400e-01
1.97158426e-01 1.82186812e-01 7.33204305e-01 1.09956837e+00
6.73445165e-01 7.32974634e-02 3.21518630e-02 -5.35238564e-01
2.37766713e-01 3.82265300e-02 -3.13576490e-01 -5.18591523e-01
-8.40292454e-01 1.00026906e+00 3.70487958e-01 3.92758071e-01
1.85670659e-01 2.37063825e-01 9.93288100e-01 5.06162703e-01
2.92340845e-01 6.43087149e-01 -4.93482083e-01 -8.47208872e-02
-7.20907629e-01 1.21346392e-01 9.62959290e-01 1.01737881e+00
7.42667079e-01 3.58954549e-01 4.97506522e-02 4.95463610e-01
3.99769515e-01 1.48219496e-01 3.58139724e-01 -2.46853784e-01
9.30178821e-01 1.39949679e+00 -5.09655893e-01 -1.44034767e+00
-3.08689296e-01 -4.54328090e-01 -7.03366339e-01 3.29526395e-01
1.91296369e-01 -6.78394362e-02 -7.24934042e-01 1.56931448e+00
4.83779684e-02 2.25961491e-01 -6.60928190e-02 3.17326576e-01
7.89508641e-01 2.27668405e-01 3.92969558e-03 2.21459508e-01
1.33267212e+00 -7.61165321e-01 -1.85246408e-01 -3.91772449e-01
9.63392675e-01 -2.34697461e-01 1.11652291e+00 4.67101723e-01
-3.62089932e-01 -4.07919526e-01 -1.26851952e+00 3.82495880e-01
-6.49098694e-01 -1.81124195e-01 8.29510570e-01 1.18545485e+00
-1.01069105e+00 4.76028174e-01 -8.00261557e-01 -1.67457715e-01
5.86368084e-01 1.91143736e-01 -4.76350904e-01 -1.56954993e-02
-1.26408792e+00 4.80695546e-01 6.30967677e-01 -1.64638177e-01
-1.25057685e+00 -7.32523680e-01 -1.17073357e+00 2.35477731e-01
8.93012285e-01 -2.65002400e-01 9.20231640e-01 -7.28267193e-01
-8.97772431e-01 7.62838185e-01 5.25400639e-01 -3.55532795e-01
3.70138474e-02 -4.63086106e-02 -7.11310565e-01 -2.58272827e-01
-1.51488066e-01 -2.80504465e-01 7.45053709e-01 -1.13107371e+00
-3.74600232e-01 -2.46167853e-01 6.54838204e-01 -7.88194418e-01
-9.08269823e-01 5.62502027e-01 -7.18216449e-02 -5.08636713e-01
-5.02671659e-01 -9.14611459e-01 -1.52849048e-01 -4.87625957e-01
-8.20726156e-01 -4.35145438e-01 7.36346126e-01 -7.06766903e-01
1.78357840e+00 -2.13849354e+00 6.38619736e-02 2.60484517e-01
6.98983192e-01 6.31941557e-01 -3.10910851e-01 4.61259514e-01
-4.79227304e-01 6.91286147e-01 -5.07961392e-01 2.33136062e-02
2.27326035e-01 -2.43005842e-01 -4.58428770e-01 5.38467884e-01
2.20091820e-01 9.10630643e-01 -9.96407032e-01 -4.12003487e-01
-2.09250584e-01 1.71646118e-01 -7.53646553e-01 1.46512210e-01
-3.44279468e-01 -3.12284887e-01 -5.33852279e-01 9.00809467e-01
4.31300461e-01 -3.67894083e-01 4.15769517e-01 2.93013901e-01
9.43548828e-02 3.11488003e-01 -1.02700078e+00 1.51784849e+00
-6.77565336e-01 5.81530690e-01 -1.73352696e-02 -1.08852160e+00
1.04931378e+00 3.32731336e-01 3.60123697e-03 -1.92480817e-01
5.33831865e-02 8.34882855e-02 3.71895015e-01 -6.68346047e-01
3.30749661e-01 3.56851697e-01 -4.26809013e-01 5.93527019e-01
2.23363042e-02 8.01424384e-02 3.13581795e-01 4.00699466e-01
1.76377881e+00 -4.11106832e-02 3.26532990e-01 -3.45176905e-01
6.73938751e-01 -1.38047468e-02 6.40937090e-01 7.07565904e-01
-1.03797205e-01 1.62044510e-01 1.27380109e+00 -6.45043910e-01
-7.59845674e-01 -6.15040541e-01 2.57704049e-01 1.02016366e+00
-2.24511266e-01 -9.68198955e-01 -8.07204604e-01 -1.52144122e+00
1.81328207e-02 5.68380177e-01 -6.85844958e-01 -5.95082283e-01
-8.07333171e-01 -7.13504374e-01 9.54979599e-01 6.91282511e-01
1.89122230e-01 -1.23324549e+00 -3.77164871e-01 1.26760527e-01
2.22080141e-01 -1.20292461e+00 -2.88422853e-01 7.81055689e-02
-4.73296583e-01 -1.55387735e+00 9.07490924e-02 -7.58077383e-01
5.46720445e-01 8.16151351e-02 1.51855540e+00 6.88277483e-01
-4.38767195e-01 2.94036325e-02 -5.45287728e-01 -3.49685162e-01
-7.79973626e-01 3.75392199e-01 -2.76176006e-01 -1.96452990e-01
5.04215062e-01 -5.17699420e-01 -1.07803151e-01 1.08973779e-01
-1.00726771e+00 -6.18115187e-01 4.56042439e-01 5.25801480e-01
-7.23678544e-02 4.47545052e-01 3.63297582e-01 -1.33071542e+00
8.28216016e-01 -7.24951327e-01 -1.07676995e+00 4.22774523e-01
-7.10119784e-01 8.44013542e-02 9.73550797e-01 -3.99497151e-01
-6.77117467e-01 -3.54240835e-01 6.50344463e-03 -3.43472540e-01
-1.09924987e-01 9.46087778e-01 -3.76881093e-01 -3.16742390e-01
1.07310784e+00 1.94705930e-02 -3.56132418e-01 -1.40706986e-01
2.82425195e-01 4.59271818e-01 3.05593580e-01 -9.85178292e-01
1.11265695e+00 -1.16107844e-01 -7.89904967e-02 -4.23561305e-01
-4.09474671e-01 -1.61064744e-01 -2.48153254e-01 1.78466097e-01
6.24245942e-01 -5.95996857e-01 -6.21880293e-01 4.99896407e-01
-1.12362421e+00 -3.18430871e-01 2.08331689e-01 -4.99487780e-02
4.26062718e-02 9.10947680e-01 -6.15191162e-01 -5.85599065e-01
-4.41879779e-01 -1.49987543e+00 8.29592645e-01 -6.18588440e-02
-2.00584516e-01 -1.10848141e+00 3.23307574e-01 1.18787728e-01
3.94031048e-01 5.04471183e-01 1.59600389e+00 -8.24427903e-01
-6.07591808e-01 -4.79744703e-01 -2.33932883e-01 4.88019049e-01
1.18184268e-01 4.16336060e-01 -6.98031068e-01 -4.78416145e-01
-2.29904398e-01 -3.15967679e-01 8.46921742e-01 -2.47630507e-01
1.26448309e+00 -3.34514350e-01 -4.78393108e-01 5.77623308e-01
1.63698697e+00 9.57472175e-02 5.31615853e-01 5.40610492e-01
1.08311391e+00 5.53575695e-01 -8.93649552e-03 1.70509875e-01
2.74893403e-01 5.59962869e-01 9.43586588e-01 2.89447099e-01
1.10662140e-01 -2.09146127e-01 6.09928608e-01 5.58744133e-01
4.59496155e-02 -2.28657052e-01 -1.60971212e+00 7.35825300e-01
-1.72885883e+00 -7.72826433e-01 -2.01325193e-01 2.04747438e+00
7.82770395e-01 4.95759815e-01 1.04481928e-01 2.67541766e-01
8.15290868e-01 3.26464146e-01 -2.35533357e-01 -5.26100636e-01
2.66873062e-01 2.13667557e-01 2.76793838e-01 3.02596062e-01
-1.18325579e+00 9.06178415e-01 5.23059940e+00 7.24973083e-01
-1.11179340e+00 -7.20695555e-02 2.15208322e-01 4.14434999e-01
-5.05665779e-01 3.55717212e-01 -7.51558065e-01 3.02221298e-01
1.05890572e+00 -2.53835648e-01 4.14189219e-01 1.20372272e+00
-4.84630466e-01 5.56314230e-01 -1.15505600e+00 5.23876250e-01
9.12469998e-02 -1.18575275e+00 -1.84434101e-01 1.11398883e-01
6.50595188e-01 2.40297094e-02 -4.79943678e-03 8.50976884e-01
8.16439450e-01 -1.07911730e+00 7.02655375e-01 -2.49362867e-02
6.90116763e-01 -6.28847480e-01 8.29545796e-01 2.27638960e-01
-1.37308538e+00 -4.47990656e-01 -3.34902406e-01 4.34709378e-02
-3.68091583e-01 7.30303764e-01 -7.95785248e-01 7.61598051e-01
7.14485168e-01 8.58210385e-01 -1.00939727e+00 7.41335034e-01
-6.70991123e-01 8.24134529e-01 2.55554378e-01 -2.42254362e-02
1.56211674e-01 1.95693582e-01 3.18015486e-01 1.41483080e+00
1.58560395e-01 -4.20961440e-01 3.45358074e-01 1.09254384e+00
-3.00296128e-01 -4.66647968e-02 -8.68897676e-01 -4.46133345e-01
2.84813672e-01 1.57850814e+00 -6.82958484e-01 -4.90468591e-02
-9.10298586e-01 -2.34501399e-02 5.97064674e-01 2.36995623e-01
-9.24232185e-01 -8.10456753e-01 6.56500697e-01 1.46271929e-01
3.98495913e-01 -2.29924247e-01 -1.01848356e-01 -1.32257819e+00
2.81230271e-01 -1.44481766e+00 4.73079383e-01 -1.42219603e-01
-1.22586024e+00 1.02539527e+00 -4.71482165e-02 -1.02258122e+00
-1.67443305e-01 -8.39301825e-01 -9.89435971e-01 7.77829528e-01
-1.44890618e+00 -1.21558142e+00 -2.05582231e-01 4.68544811e-01
1.68855518e-01 -6.67521477e-01 6.88307881e-01 3.21716577e-01
-8.73959243e-01 1.00407326e+00 -4.07606870e-01 7.40175426e-01
4.01958615e-01 -1.43574774e+00 1.13133109e+00 1.32840252e+00
2.85685450e-01 1.24581814e+00 4.91251618e-01 -8.11736763e-01
-1.58660865e+00 -1.22851336e+00 4.98661369e-01 -7.36154735e-01
1.31417131e+00 -6.97507441e-01 -1.24628556e+00 7.51787424e-01
1.95017502e-01 4.33157384e-01 5.49598992e-01 4.06738460e-01
-1.24133289e+00 7.48329684e-02 -9.36230540e-01 4.97190952e-01
1.12434769e+00 -8.57607782e-01 -3.61488581e-01 1.94279298e-01
8.56491446e-01 -1.79707021e-01 -8.62409353e-01 5.11207879e-01
2.53148913e-01 -1.07954621e+00 8.31905484e-01 -1.02006137e+00
6.93065166e-01 -3.12654197e-01 -6.68228567e-02 -1.21856511e+00
-3.32644880e-01 -6.21458352e-01 -3.76035929e-01 1.35458851e+00
6.48901343e-01 -6.43412173e-01 8.06746066e-01 2.85192907e-01
-3.60126436e-01 -7.25630581e-01 -4.89495933e-01 -8.18493545e-01
1.97956592e-01 -6.30984724e-01 8.19299042e-01 1.30129266e+00
3.77896965e-01 2.02460989e-01 -9.24669132e-02 3.70447546e-01
6.26828551e-01 3.26551914e-01 7.34492242e-01 -1.33064544e+00
-7.50631809e-01 -7.63194144e-01 -6.98450804e-01 -2.27564052e-01
8.15114975e-01 -1.17870688e+00 -2.44054645e-01 -1.25898993e+00
2.52310485e-01 -2.41913140e-01 -4.98518407e-01 9.48004603e-01
-4.23924059e-01 -2.59619594e-01 -3.93293099e-03 -2.40405455e-01
-5.40980279e-01 9.86443833e-02 5.70309103e-01 -7.58738101e-01
1.09893486e-01 -4.00855243e-02 -1.01697290e+00 7.00205147e-01
6.13087833e-01 -7.05320418e-01 -5.42728603e-01 -4.90961224e-01
8.50928903e-01 2.48576589e-02 2.22101450e-01 -8.62658560e-01
5.13055595e-03 -2.00534999e-01 -5.37747204e-01 2.27143586e-01
-6.26425028e-01 -6.24870062e-01 -6.29069135e-02 5.18486917e-01
-3.10477316e-01 3.04653406e-01 3.50724190e-01 5.73713064e-01
-3.06258231e-01 -5.69532812e-01 4.93036479e-01 -2.52810508e-01
-9.45677638e-01 4.76807952e-01 -7.16250837e-02 2.68438488e-01
9.74849343e-01 4.94743660e-02 -9.82007980e-01 1.15513392e-01
-1.84388265e-01 -3.98641936e-02 6.45786285e-01 8.87125671e-01
5.21288991e-01 -9.06825066e-01 -6.87386870e-01 2.97512233e-01
5.67068458e-01 -3.71778905e-01 5.18097468e-02 4.72082227e-01
-4.82954741e-01 2.26624325e-01 -1.38411537e-01 -3.33563924e-01
-1.26568842e+00 1.11428213e+00 2.50625759e-01 -6.77728176e-01
-6.05653346e-01 7.26427495e-01 2.28625342e-01 -5.90989530e-01
1.60561681e-01 -4.54275548e-01 -3.35396707e-01 -2.79486150e-01
4.34063315e-01 1.54755712e-01 3.45459610e-01 -2.25160629e-01
-6.67008579e-01 5.10322928e-01 -3.13234299e-01 6.24303699e-01
1.37795174e+00 8.30838680e-01 -5.20207584e-01 -1.38678759e-01
1.16982710e+00 2.54914522e-01 -6.79151177e-01 -1.14304513e-01
4.77628857e-01 -4.28388655e-01 -1.02450863e-01 -7.09083915e-01
-1.41701889e+00 1.01879859e+00 1.88596666e-01 7.71299958e-01
8.86438847e-01 -6.26286343e-02 4.59228069e-01 4.32741731e-01
7.33844221e-01 -4.17438507e-01 2.26801470e-01 5.75415432e-01
5.74488163e-01 -1.23658884e+00 -1.12656560e-02 -4.10946876e-01
-2.56064355e-01 1.10459781e+00 9.92249727e-01 -1.49788976e-01
6.36778712e-01 5.75158238e-01 -8.48589242e-02 -5.01297057e-01
-7.11065888e-01 2.06588984e-01 2.11313799e-01 6.40505731e-01
6.46111548e-01 3.41463797e-02 5.36569115e-03 6.96159542e-01
1.00260362e-01 -4.68938679e-01 8.66007328e-01 1.04901147e+00
-2.02946767e-01 -1.39710462e+00 -1.22603849e-01 4.19389993e-01
-7.25065947e-01 -3.87484878e-01 -4.71728355e-01 9.17025745e-01
1.97532997e-02 9.49676871e-01 -6.59320593e-01 -7.17779219e-01
5.06671369e-01 -9.53753218e-02 2.84364939e-01 -1.09272230e+00
-1.12591219e+00 -5.64956486e-01 1.07402705e-01 -6.13891184e-01
5.66993579e-02 -3.58793110e-01 -8.63649011e-01 -1.51874095e-01
-3.43811035e-01 2.20474169e-01 4.27349567e-01 5.41927755e-01
3.55666101e-01 8.30871224e-01 4.90863532e-01 -1.82262838e-01
-7.46951759e-01 -7.88940907e-01 -4.62743521e-01 4.72381979e-01
2.39644676e-01 -6.40214145e-01 -5.03819346e-01 -1.39910057e-01] | [7.158489227294922, 7.770480155944824] |
d99076b6-fd59-461c-9206-cb93a366939b | a-modified-pinn-approach-for-identifiable | 2208.01169 | null | https://arxiv.org/abs/2208.01169v2 | https://arxiv.org/pdf/2208.01169v2.pdf | A Modified PINN Approach for Identifiable Compartmental Models in Epidemiology with Applications to COVID-19 | A variety of approaches using compartmental models have been used to study the COVID-19 pandemic and the usage of machine learning methods with these models has had particularly notable success. We present here an approach toward analyzing accessible data on Covid-19's U.S. development using a variation of the "Physics Informed Neural Networks" (PINN) which is capable of using the knowledge of the model to aid learning. We illustrate the challenges of using the standard PINN approach, then how with appropriate and novel modifications to the loss function the network can perform well even in our case of incomplete information. Aspects of identifiability of the model parameters are also assessed, as well as methods of denoising available data using a wavelet transform. Finally, we discuss the capability of the neural network methodology to work with models of varying parameter values, as well as a concrete application in estimating how effectively cases are being tested for in a population, providing a ranking of U.S. states by means of their respective testing. | ['HongKun Zhang', 'Panayotis G. Kevrekidis', 'Connor M Kennedy', 'Haoran Hu'] | 2022-08-01 | null | null | null | null | ['epidemiology'] | ['medical'] | [ 2.55424738e-01 -2.11902246e-01 -2.40222305e-01 -1.95548072e-01
-7.71164238e-01 -3.64962399e-01 6.66864753e-01 2.16725603e-01
-7.10213482e-01 1.13875580e+00 2.38991454e-01 -8.30815971e-01
-7.51870990e-01 -7.81174600e-01 -5.15260637e-01 -8.88768792e-01
-2.13881105e-01 9.58948076e-01 -1.24503843e-01 -3.82631123e-01
2.31281504e-01 6.45007670e-01 -1.22201180e+00 -5.08391932e-02
6.96280062e-01 6.76500201e-01 -1.82568252e-01 8.14029515e-01
5.75801404e-03 6.92507267e-01 -6.99091434e-01 -1.70036584e-01
6.32405803e-02 2.06479002e-02 -5.61060488e-01 -5.94311953e-01
-2.35488221e-01 -5.75581908e-01 -1.62176818e-01 5.50563991e-01
7.73469687e-01 1.20416068e-01 1.05113411e+00 -1.06610155e+00
-2.51960665e-01 2.32434958e-01 -4.71225902e-02 6.46592438e-01
2.34795555e-01 1.51116654e-01 3.92306954e-01 -3.82480681e-01
5.34111261e-01 1.24288702e+00 1.27225363e+00 3.04505259e-01
-1.45273328e+00 -6.47958994e-01 -8.14504772e-02 6.61150068e-02
-1.19139874e+00 -1.90424278e-01 4.82720375e-01 -8.32347751e-01
1.38407230e+00 6.53496608e-02 6.11141324e-01 1.06687689e+00
3.80749792e-01 2.21815214e-01 1.20427465e+00 -4.15629357e-01
3.12582016e-01 8.28902870e-02 1.97895110e-01 4.89247710e-01
4.86332715e-01 4.65148538e-01 -1.17418319e-01 -7.84067869e-01
6.86915636e-01 7.17691630e-02 1.39019951e-01 -2.63707954e-02
-5.68443239e-01 1.26555717e+00 8.03699493e-02 3.39718223e-01
-4.71811801e-01 1.47329986e-01 5.09192467e-01 -5.42305820e-02
9.28416729e-01 3.08516443e-01 -8.77668381e-01 1.23156562e-01
-1.05546832e+00 5.21621525e-01 8.22742343e-01 2.64246255e-01
3.99238259e-01 3.37387532e-01 2.57141702e-02 6.27671838e-01
3.98556709e-01 6.33267283e-01 -6.57485873e-02 -1.08027148e+00
3.31789434e-01 1.47294730e-01 2.21986994e-01 -7.47323751e-01
-7.45404184e-01 -2.59795934e-01 -7.61947453e-01 5.14169782e-02
5.87496102e-01 -8.09003830e-01 -9.76091206e-01 1.90705276e+00
2.15535745e-01 3.10608953e-01 6.05766252e-02 4.56958450e-02
3.46366137e-01 8.01576257e-01 3.57559055e-01 -3.93988818e-01
1.16709769e+00 -8.79601911e-02 -7.52675653e-01 1.12400860e-01
5.66803992e-01 -3.17072690e-01 2.98037797e-01 5.55288434e-01
-1.08727551e+00 -2.59330869e-01 -6.32823765e-01 4.26927418e-01
-9.06506181e-01 -6.12651825e-01 6.40922964e-01 7.93866634e-01
-1.10041869e+00 6.99403405e-01 -1.06522572e+00 -4.28941011e-01
2.88423270e-01 6.34580553e-01 8.62952918e-02 1.13221079e-01
-1.55377793e+00 1.40409064e+00 5.19564986e-01 9.87177044e-02
-7.20139742e-01 -9.63982284e-01 -7.70189285e-01 2.92056194e-03
1.48895159e-01 -6.93631113e-01 1.04709911e+00 -5.38668096e-01
-8.36586118e-01 3.85511130e-01 -2.50714142e-02 -6.85607433e-01
3.56586069e-01 2.18185157e-01 -5.93221188e-01 1.64757878e-01
-1.55518994e-01 4.53753948e-01 3.20229292e-01 -1.08268261e+00
-6.41112208e-01 -3.76905531e-01 -7.27044269e-02 -2.10813507e-01
-5.64137213e-02 5.33296824e-01 1.62043720e-01 -6.04772866e-01
-4.27900642e-01 -6.24794126e-01 -5.08673012e-01 -3.33746463e-01
-1.71280891e-01 -3.86516839e-01 5.48814833e-01 -1.21375453e+00
1.20597720e+00 -1.50344896e+00 -2.78119236e-01 6.64773643e-01
1.30687952e-01 3.31923813e-01 -8.31927545e-03 9.25824821e-01
-6.21502399e-02 2.43506283e-01 -5.46339989e-01 -4.62631918e-02
2.32111830e-02 5.85449755e-01 -1.91342423e-03 3.70706558e-01
4.71129328e-01 4.84636813e-01 -6.44692481e-01 -3.58087629e-01
1.41124234e-01 1.02136183e+00 -5.93289316e-01 5.33839650e-02
2.66509922e-03 5.20939410e-01 -6.30107164e-01 4.89490300e-01
6.44297481e-01 -5.10361195e-02 2.60748804e-01 3.46203655e-01
-1.88395828e-01 1.70945942e-01 -8.04895580e-01 8.01376998e-01
-3.75356019e-01 3.81266862e-01 1.16635650e-01 -1.36526835e+00
6.40205204e-01 5.60431182e-01 7.29965746e-01 -4.82623249e-01
6.35094270e-02 2.77214795e-01 2.51949400e-01 -5.98209918e-01
1.99887045e-02 -4.99465048e-01 2.45967507e-01 6.41997635e-01
-4.08397168e-02 1.60149261e-01 1.92627698e-01 -9.61519256e-02
1.03952467e+00 3.12590525e-02 4.33940262e-01 -4.79810894e-01
4.11386490e-01 2.47059450e-01 5.26107609e-01 9.15422261e-01
-3.98776203e-01 2.68826559e-02 6.37614131e-01 -6.42120957e-01
-1.09984756e+00 -8.25478017e-01 -6.09257579e-01 7.66234875e-01
-8.09962094e-01 2.53497481e-01 -9.62057531e-01 -2.69752175e-01
6.21120185e-02 9.52320874e-01 -1.02806199e+00 -6.27534240e-02
-7.14811444e-01 -1.46336234e+00 7.41134584e-01 5.31851172e-01
-8.48679710e-03 -1.22070789e+00 -5.54996729e-01 4.64556843e-01
1.67640671e-01 -5.00400782e-01 1.65744618e-01 6.01376534e-01
-8.83511841e-01 -1.13725722e+00 -7.69112945e-01 -4.50280547e-01
5.68228841e-01 -3.72781008e-01 1.05911279e+00 1.48495987e-01
-2.72238255e-01 5.20972967e-01 5.08254133e-02 -8.20154369e-01
-8.16910625e-01 8.79713334e-03 2.97453076e-01 -6.05280817e-01
7.12327957e-01 -3.00021768e-01 -6.04070485e-01 -8.86519179e-02
-1.07060134e+00 -5.17185986e-01 1.46816820e-01 9.02630746e-01
3.18716496e-01 4.10966694e-01 9.39324319e-01 -9.75350380e-01
7.99641669e-01 -9.60735619e-01 -8.46497834e-01 1.56987667e-01
-7.59001315e-01 -1.91867962e-01 4.39725935e-01 -3.01965356e-01
-1.12148404e+00 -2.76950359e-01 -4.28834200e-01 5.28800376e-02
-2.48489425e-01 8.44640553e-01 3.14425111e-01 5.65979779e-02
5.32939017e-01 -5.66621050e-02 3.16247433e-01 -7.02610254e-01
-2.12339655e-01 6.23639405e-01 2.52629548e-01 -6.06882036e-01
6.74686611e-01 3.18886817e-01 1.83574587e-01 -6.75884843e-01
-4.49772060e-01 -3.78101617e-01 -5.38194120e-01 -8.84652361e-02
9.46507871e-01 -7.26259530e-01 -5.82219601e-01 5.17726481e-01
-1.10283160e+00 -3.23124915e-01 -6.47558570e-02 6.38573229e-01
-6.94790423e-01 -3.62913720e-02 -6.70439184e-01 -1.28376722e+00
-1.25237435e-01 -1.22898328e+00 5.97376287e-01 5.38591370e-02
-1.06439747e-01 -1.81303084e+00 6.91543043e-01 3.32269967e-01
6.72047794e-01 3.92981380e-01 1.40476799e+00 -1.30568981e+00
-4.08433527e-02 -3.19807529e-01 -3.25565115e-02 3.22180301e-01
3.42674032e-02 3.40319276e-02 -1.00494552e+00 -5.84512591e-01
2.38630742e-01 -9.65988338e-02 5.78421354e-01 1.00464964e+00
8.96713197e-01 -5.58096051e-01 -2.93184280e-01 3.94561589e-01
1.71136594e+00 7.41785526e-01 3.60942662e-01 3.89576703e-01
1.17527433e-01 9.99684930e-01 1.45598114e-01 3.81819725e-01
1.33578867e-01 4.22769845e-01 2.30309665e-01 -2.78843194e-01
4.95318264e-01 -1.30544215e-01 9.97467637e-02 5.88035107e-01
-2.86777586e-01 -3.49737048e-01 -1.32121956e+00 3.80710483e-01
-1.43343949e+00 -1.22069931e+00 8.44660476e-02 2.13006210e+00
7.14592874e-01 2.21280083e-01 3.91160399e-01 7.82144070e-02
6.62425518e-01 -1.17253348e-01 -5.27612150e-01 -7.90801525e-01
7.16258213e-02 4.44171697e-01 6.63215756e-01 5.87658644e-01
-9.39332128e-01 2.13188663e-01 8.45795631e+00 8.33546102e-01
-6.66253746e-01 1.71035767e-01 1.00602829e+00 2.98947603e-01
-2.48472705e-01 -1.19265832e-01 -6.18496537e-01 4.86386746e-01
1.53346348e+00 2.11632147e-01 4.71108496e-01 1.32845461e-01
7.99266696e-01 -9.28504243e-02 -6.12793028e-01 2.22189471e-01
-1.25547171e-01 -1.39624178e+00 -8.81609693e-03 1.30330816e-01
6.91296458e-01 2.04630002e-01 5.42116202e-02 2.09485665e-01
3.62261534e-01 -1.19874465e+00 1.38670921e-01 8.39771569e-01
6.98640287e-01 -1.15727854e+00 1.08325934e+00 4.21408504e-01
-6.56779766e-01 -1.91111058e-01 -2.38233268e-01 -1.61590889e-01
1.21508032e-01 2.38040417e-01 -8.56167197e-01 3.82494181e-01
4.46412474e-01 1.48602128e-01 -1.69395298e-01 1.20250487e+00
3.85891318e-01 8.94697964e-01 -5.72651923e-01 9.64369718e-03
4.00199920e-01 -6.22886084e-02 3.76368701e-01 1.34636605e+00
4.08824593e-01 1.26962394e-01 -1.43328393e-02 6.85795248e-01
5.25570214e-01 -3.54221985e-02 -8.15045118e-01 -2.47208942e-02
4.59127069e-01 6.10082328e-01 -5.81722856e-01 -2.79212981e-01
-3.33215356e-01 8.60412121e-02 -5.38453162e-02 6.06072366e-01
-5.20497203e-01 -3.34597170e-01 5.34780860e-01 1.74670622e-01
1.74628124e-01 3.40035409e-02 9.58637372e-02 -5.72226286e-01
-4.72759217e-01 -8.75647485e-01 6.30235493e-01 -6.42323136e-01
-1.42577434e+00 3.63951802e-01 7.40689218e-01 -6.02773786e-01
-5.49500525e-01 -1.05507481e+00 -7.74718702e-01 1.09097791e+00
-1.40057385e+00 -8.52751017e-01 6.50442779e-01 2.84691751e-01
2.21186101e-01 -4.01702911e-01 8.97359073e-01 2.20980316e-01
-6.55612230e-01 3.23562711e-01 7.11127043e-01 -1.59381241e-01
1.44241661e-01 -1.16551018e+00 1.68659002e-01 4.78129596e-01
-5.51432490e-01 7.26494491e-01 9.68607187e-01 -8.85750890e-01
-9.26426768e-01 -8.11345220e-01 1.01228988e+00 -4.10268217e-01
7.46515572e-01 -1.96438029e-01 -7.58913755e-01 6.43014848e-01
2.82081395e-01 -5.41471422e-01 8.46380532e-01 1.57042798e-02
-1.91075094e-02 1.25469074e-01 -1.72230732e+00 3.19325477e-01
5.11766493e-01 -4.15444911e-01 -6.66999578e-01 7.44141281e-01
4.87701416e-01 2.08941758e-01 -8.26449752e-01 3.36604834e-01
4.51322645e-01 -9.17932212e-01 1.20890141e+00 -1.26709199e+00
-7.17277527e-02 1.71211630e-01 -1.53604448e-01 -1.15091014e+00
-2.84291267e-01 -7.04277754e-01 5.12190759e-02 9.90279913e-01
4.23713446e-01 -1.07190073e+00 6.54895484e-01 5.80457687e-01
2.86512405e-01 -9.83735323e-01 -1.35769546e+00 -7.03928053e-01
6.00452304e-01 -5.30750394e-01 6.36726439e-01 7.90165484e-01
-3.51939976e-01 -2.76226103e-01 -3.39991152e-01 1.10946797e-01
6.40196025e-01 -7.40738153e-01 1.66862532e-01 -1.43027198e+00
-2.92687505e-01 -4.18673068e-01 -2.98560113e-01 -5.05089723e-02
-9.54430667e-04 -4.59165603e-01 -2.30217203e-01 -1.40216565e+00
4.34352934e-01 -5.34740210e-01 -6.40583277e-01 2.63895601e-01
6.94587156e-02 -6.33395556e-03 1.57028660e-02 2.82839742e-02
2.60337740e-01 -1.70238510e-01 7.18122840e-01 -5.02802432e-02
-8.56701732e-02 3.01842421e-01 -6.03450894e-01 8.94826055e-01
1.11049139e+00 -6.61014378e-01 -5.61766028e-01 -1.50091067e-01
3.01488221e-01 2.30440259e-01 5.33766925e-01 -7.54449248e-01
-7.26540573e-03 -3.40701431e-01 6.90934777e-01 -8.04640591e-01
3.37859005e-01 -9.38346803e-01 5.46753287e-01 9.42043662e-01
-2.98603028e-01 5.86657643e-01 3.86656851e-01 5.14032125e-01
2.73294777e-01 -4.89376277e-01 7.83149600e-01 -1.57415330e-01
3.29009593e-02 1.85160890e-01 -7.74951279e-01 1.32581875e-01
7.67009854e-01 -1.44902229e-01 -2.29106396e-01 -4.45899248e-01
-8.70502174e-01 2.32452542e-01 1.22295573e-01 -2.40610853e-01
3.60460252e-01 -8.46624494e-01 -8.00983131e-01 1.77706495e-01
-4.09276307e-01 -4.97868329e-01 3.53670329e-01 9.13717508e-01
-7.58193433e-01 8.66984129e-01 -1.68238774e-01 -2.83114076e-01
-7.20264614e-01 7.59796917e-01 7.55466759e-01 -6.59850895e-01
-1.61356464e-01 4.26352501e-01 1.47128791e-01 -6.15832150e-01
1.91291139e-01 -8.17935914e-02 -4.56515014e-01 2.79798388e-01
4.36765462e-01 6.13560617e-01 -1.03961214e-01 -6.87164962e-01
-4.23121482e-01 3.88737172e-01 -9.18751881e-02 -1.17854089e-01
1.64128470e+00 5.99810220e-02 -1.19505711e-01 2.89748460e-01
1.12568367e+00 -2.43995890e-01 -1.34032750e+00 1.28922388e-02
9.16042700e-02 2.83643156e-01 8.55282396e-02 -1.22789657e+00
-8.55063677e-01 9.35798168e-01 8.15670371e-01 3.75908613e-01
1.10582232e+00 -3.92788053e-01 4.71934438e-01 4.10995215e-01
6.02440387e-02 -1.09124935e+00 -5.93836904e-01 3.58984560e-01
4.30800498e-01 -1.00422144e+00 1.45706609e-01 1.80153951e-01
-1.78929180e-01 1.06725800e+00 -1.22970827e-01 -1.20714158e-01
1.07345200e+00 2.75558740e-01 9.01224017e-02 -3.45669329e-01
-6.73786640e-01 6.61025718e-02 1.20256647e-01 1.13838875e+00
1.16704211e-01 -3.69814150e-02 -3.86834353e-01 3.78047854e-01
2.04274133e-01 5.12139834e-02 4.36184525e-01 9.19150829e-01
-4.24326390e-01 -1.03157520e+00 -6.39884174e-01 7.96805680e-01
-8.34325850e-01 -1.89955577e-01 -1.75310865e-01 1.23391223e+00
2.94017017e-01 8.06396782e-01 7.98730701e-02 2.84602977e-02
1.12231456e-01 3.63128662e-01 9.03817266e-02 -2.76516736e-01
-5.98909140e-01 2.09176168e-01 2.51542538e-01 -1.41062111e-01
-7.19275892e-01 -9.39685106e-01 -9.42065477e-01 -5.12214184e-01
-1.05851553e-01 2.93634623e-01 7.28795171e-01 1.14234948e+00
-1.24046892e-01 2.28648663e-01 5.38070440e-01 -9.12854493e-01
-8.69770586e-01 -6.82359159e-01 -6.89105272e-01 -1.91186652e-01
6.21627510e-01 -8.05485249e-01 -5.22188723e-01 -3.88044626e-01] | [6.037579536437988, 4.327662467956543] |
fa48525f-2807-43ad-975b-e24617c48577 | face-reconstruction-with-variational | 2112.02139 | null | https://arxiv.org/abs/2112.02139v1 | https://arxiv.org/pdf/2112.02139v1.pdf | Face Reconstruction with Variational Autoencoder and Face Masks | Variational AutoEncoders (VAE) employ deep learning models to learn a continuous latent z-space that is subjacent to a high-dimensional observed dataset. With that, many tasks are made possible, including face reconstruction and face synthesis. In this work, we investigated how face masks can help the training of VAEs for face reconstruction, by restricting the learning to the pixels selected by the face mask. An evaluation of the proposal using the celebA dataset shows that the reconstructed images are enhanced with the face masks, especially when SSIM loss is used either with l1 or l2 loss functions. We noticed that the inclusion of a decoder for face mask prediction in the architecture affected the performance for l1 or l2 loss functions, while this was not the case for the SSIM loss. Besides, SSIM perceptual loss yielded the crispest samples between all hypotheses tested, although it shifts the original color of the image, making the usage of the l1 or l2 losses together with SSIM helpful to solve this issue. | ['Eric A. Antonelo', 'Rafael S. Toledo'] | 2021-12-03 | null | null | null | null | ['face-reconstruction'] | ['computer-vision'] | [ 8.62175301e-02 5.55925131e-01 1.55408204e-01 -3.38663846e-01
-6.76700324e-02 -2.39002839e-01 7.90725410e-01 -3.76693875e-01
-2.51114070e-01 7.35767663e-01 2.34706495e-02 1.78218931e-01
-2.34924644e-01 -8.25725973e-01 -1.01781547e+00 -1.09674144e+00
2.18190834e-01 2.95247376e-01 -5.27908355e-02 1.60375778e-02
-1.54543575e-02 8.18213403e-01 -2.23440576e+00 4.25274789e-01
4.34673131e-01 1.08636141e+00 3.50001603e-01 2.89839476e-01
-1.29422054e-01 6.20521545e-01 -5.97925305e-01 -6.99474931e-01
3.90577048e-01 -5.37909567e-01 -2.09631741e-01 1.32913172e-01
5.67365289e-01 -3.30939442e-01 -3.76200043e-02 9.52281117e-01
3.71690452e-01 -6.26322301e-03 9.77218747e-01 -1.16018879e+00
-3.53881389e-01 3.06353033e-01 -3.07723075e-01 -3.38849068e-01
1.51645958e-01 1.53164968e-01 8.81146967e-01 -1.04875958e+00
7.96494067e-01 1.51994705e+00 4.61479992e-01 6.72282934e-01
-1.40254271e+00 -6.62945032e-01 -1.68187007e-01 1.76093370e-01
-1.46454954e+00 -6.57762647e-01 8.37186158e-01 -4.50329930e-01
7.82947242e-01 2.26220161e-01 3.90679538e-01 1.06201732e+00
-1.53420056e-04 5.26268899e-01 1.07182527e+00 -6.06491685e-01
4.20043990e-02 6.48755491e-01 -5.16516149e-01 6.26757979e-01
9.18390900e-02 2.34154552e-01 -2.13791534e-01 1.24665149e-01
8.53931010e-01 -4.63380873e-01 -2.19358236e-01 -4.34961766e-01
-4.82272506e-01 9.72890735e-01 3.26599896e-01 5.07420003e-01
-4.41700399e-01 9.26378667e-02 7.72025660e-02 3.10520321e-01
6.29235387e-01 4.12565887e-01 -2.96221018e-01 3.22766364e-01
-1.07190084e+00 5.38846059e-03 4.04448241e-01 4.58483040e-01
8.48972917e-01 3.66447896e-01 -3.38548005e-01 9.88409936e-01
3.08595002e-01 2.15781659e-01 1.45097539e-01 -1.04889619e+00
1.70576155e-01 4.30510044e-01 -3.96854542e-02 -9.43222463e-01
-2.99332857e-01 -4.28748071e-01 -6.52976871e-01 6.93065047e-01
6.63551450e-01 8.20079073e-03 -8.13467443e-01 1.89909661e+00
2.82595813e-01 2.02754036e-01 4.25338522e-02 9.44589019e-01
5.60876191e-01 7.02290893e-01 -1.11351080e-01 -3.48010004e-01
1.07512665e+00 -4.43678409e-01 -1.03984439e+00 -1.02591321e-01
1.90323502e-01 -8.47593367e-01 1.10771167e+00 5.73395312e-01
-1.25702858e+00 -8.24801087e-01 -1.05323780e+00 6.05275854e-02
-1.10004775e-01 5.69545507e-01 3.15766215e-01 1.00662243e+00
-1.25704694e+00 7.82703400e-01 -6.14993453e-01 -9.09950808e-02
5.71417093e-01 3.86768311e-01 -3.11646044e-01 5.55356815e-02
-8.72313738e-01 1.04938626e+00 1.19159579e-01 1.46873653e-01
-9.14175332e-01 -6.20208383e-01 -6.70883298e-01 4.11616385e-01
2.59605706e-01 -2.29230776e-01 5.87327421e-01 -1.36175621e+00
-1.65770674e+00 1.00037456e+00 -1.50861591e-01 -5.04454255e-01
9.16528165e-01 -1.44054499e-02 -2.70152211e-01 8.35417360e-02
-3.74858707e-01 1.07584953e+00 1.40523458e+00 -1.46613741e+00
-5.53273745e-02 -3.91032070e-01 -1.69566318e-01 -1.70524523e-01
-2.15529621e-01 -2.27338850e-01 -1.98656544e-01 -6.42826676e-01
-1.35116309e-01 -7.86890149e-01 2.65009552e-01 2.69230157e-01
-2.75453448e-01 -9.62653905e-02 8.44512522e-01 -8.94468009e-01
6.88714921e-01 -2.49905562e+00 2.74759650e-01 3.84342641e-01
-8.83847177e-02 1.12714730e-01 -9.15596634e-02 1.77821130e-01
-3.96456867e-01 -3.01346686e-02 -2.10249454e-01 -6.55535460e-01
9.34963375e-02 2.69424111e-01 -2.86657006e-01 6.40242159e-01
5.18304825e-01 4.72153246e-01 -1.16776392e-01 -3.58946204e-01
3.67227703e-01 9.48798597e-01 -9.04568791e-01 1.70628086e-01
-4.83442634e-01 5.04487038e-01 1.38467059e-01 1.02911659e-01
9.34251189e-01 1.66454345e-01 7.84354955e-02 -1.61613837e-01
-7.71419555e-02 -2.33525522e-02 -1.11136174e+00 1.29975712e+00
-4.88702774e-01 8.58143806e-01 4.08591419e-01 -8.27879310e-01
1.15854716e+00 4.96499717e-01 4.16922539e-01 -8.75328362e-01
1.72580943e-01 1.94845662e-01 8.70418027e-02 -3.18465531e-01
1.96142778e-01 -3.80865246e-01 7.28058457e-01 1.45381376e-01
3.09736371e-01 -1.98453411e-01 1.00357696e-01 -3.20459366e-01
3.44327033e-01 2.59881347e-01 -2.69338936e-01 -3.67006838e-01
7.71716237e-01 -6.29905760e-01 3.43111783e-01 2.32789412e-01
2.39047185e-01 8.53091776e-01 8.54518592e-01 -2.00717613e-01
-1.17085350e+00 -9.85271573e-01 -4.52705652e-01 7.27977812e-01
-2.82673091e-01 3.73179391e-02 -1.08736062e+00 -4.10306424e-01
-2.33009249e-01 1.13348567e+00 -5.78982294e-01 -2.96674490e-01
-3.94691527e-01 -6.27547622e-01 5.12488127e-01 -2.09620502e-02
3.00488740e-01 -1.37878799e+00 -5.64885080e-01 -1.51055068e-01
5.65151609e-02 -9.29264128e-01 1.75404586e-02 3.75869423e-02
-7.55976796e-01 -9.77858782e-01 -9.15335119e-01 -5.02681673e-01
6.87271476e-01 -4.36597675e-01 8.86336267e-01 1.94684654e-01
-3.34908426e-01 1.99907482e-01 -2.65483260e-01 -3.17408770e-01
-5.52839637e-01 -2.44445398e-01 -4.38350439e-03 3.63106281e-01
1.39619550e-02 -4.86210227e-01 -4.13513392e-01 1.28369480e-01
-9.63664412e-01 -1.16333596e-01 2.57691264e-01 7.77387738e-01
4.13757980e-01 7.26727769e-02 4.37318295e-01 -6.62725866e-01
3.80755901e-01 -1.52371705e-01 -9.07305956e-01 8.70210603e-02
-6.10315204e-01 3.58089268e-01 4.80616122e-01 -3.90944898e-01
-1.24277508e+00 -9.34726074e-02 -6.09021544e-01 -7.16509163e-01
-1.06417380e-01 -2.08677091e-02 -4.53464359e-01 3.98950912e-02
5.74105978e-01 7.93469846e-02 3.41934502e-01 -5.43410420e-01
1.65797561e-01 2.43026733e-01 7.62515813e-02 -2.04939499e-01
5.11333048e-01 5.33690393e-01 1.36270583e-01 -1.13623214e+00
-1.93203554e-01 3.33905518e-01 -3.54338050e-01 -3.11025947e-01
1.04284120e+00 -5.97213149e-01 -6.66132212e-01 2.94892341e-01
-1.10593295e+00 -2.67052203e-01 -6.70292497e-01 5.72464705e-01
-6.68133795e-01 1.19022958e-01 -3.00104797e-01 -9.89467025e-01
1.04176983e-01 -1.46633887e+00 7.73867249e-01 -4.29831930e-02
-4.46606129e-02 -7.44231761e-01 -2.30481878e-01 2.49864176e-01
3.65200073e-01 -6.97482079e-02 1.20680857e+00 -2.36590862e-01
-4.51938927e-01 1.47949532e-01 -6.24675490e-02 8.59712481e-01
-6.13080114e-02 2.38629594e-01 -1.60021102e+00 -2.84199059e-01
2.90892839e-01 -2.27846444e-01 9.77964818e-01 6.18733168e-01
1.33507180e+00 -3.85582656e-01 1.74263552e-01 7.68300116e-01
1.29125214e+00 1.42197624e-01 1.14223742e+00 -4.53744195e-02
3.80955517e-01 1.15476298e+00 1.80044651e-01 4.06503558e-01
-3.02918851e-01 8.91830027e-01 7.27373779e-01 -2.44203314e-01
-4.00109082e-01 -1.52051061e-01 5.64062297e-01 5.55085540e-02
-3.50517705e-02 -1.48798674e-01 -4.28854614e-01 3.20625603e-01
-1.27877533e+00 -8.43196571e-01 -7.42901936e-02 2.34751201e+00
4.87401217e-01 1.37248546e-01 -5.16499802e-02 3.32253247e-01
6.29559577e-01 1.43909290e-01 -3.80663902e-01 -5.77368498e-01
-4.16263431e-01 3.68072778e-01 2.56462753e-01 7.11257815e-01
-7.41684198e-01 8.29599500e-01 5.72525072e+00 9.72085953e-01
-1.48655617e+00 9.60220024e-02 6.83092356e-01 -2.53838986e-01
-4.64850128e-01 -3.31312060e-01 -5.47125459e-01 5.41588247e-01
8.32073331e-01 5.18432677e-01 4.22204465e-01 6.09669745e-01
1.23779811e-01 -1.19981200e-01 -9.65741515e-01 9.36001241e-01
-2.72368006e-02 -1.24398232e+00 1.08208582e-01 1.93024397e-01
6.64103210e-01 -3.71719807e-01 4.89640236e-01 1.52295921e-02
-3.65155220e-01 -1.23764038e+00 9.24200058e-01 5.17358899e-01
9.26114500e-01 -9.73132074e-01 5.68827331e-01 2.68810868e-01
-5.34284353e-01 -5.66861555e-02 -4.62126434e-01 6.56500161e-02
-3.47988978e-02 6.23879194e-01 -9.33226228e-01 1.73338816e-01
5.34506679e-01 3.00044328e-01 -5.30663371e-01 6.79530382e-01
-2.94684917e-01 5.08781672e-01 -2.64999062e-01 3.86125505e-01
-4.61786613e-02 -4.84390408e-01 7.03510702e-01 9.01487648e-01
5.85224807e-01 -2.73040324e-01 -3.76177639e-01 1.30445850e+00
-8.25532302e-02 3.65659781e-02 -5.87298155e-01 -4.77788821e-02
3.04573663e-02 9.31251526e-01 -5.37647009e-01 3.78206372e-02
-2.12222487e-01 9.13297355e-01 2.09494215e-03 4.09837782e-01
-7.24878669e-01 3.29366587e-02 6.53285980e-01 6.61233306e-01
5.78696847e-01 1.33762434e-01 -2.59092867e-01 -8.21842074e-01
1.98478415e-03 -7.40930021e-01 2.07176767e-02 -8.68458390e-01
-1.03764486e+00 6.85423970e-01 1.11131981e-01 -9.91289556e-01
-4.46929574e-01 -8.07854891e-01 -3.69757771e-01 9.94066238e-01
-1.44283020e+00 -9.32438195e-01 1.74631417e-01 5.40991724e-01
3.33330333e-01 -3.98299217e-01 6.85393631e-01 4.56373006e-01
-4.15873230e-01 7.06740260e-01 -2.95891659e-03 -2.14259118e-01
5.01264870e-01 -9.23628747e-01 -1.27184838e-01 6.82758689e-01
3.17081749e-01 3.67520630e-01 1.09563351e+00 -3.57226074e-01
-8.40280890e-01 -6.42700791e-01 7.55408823e-01 -9.40679610e-02
-8.00763369e-02 -4.79389787e-01 -1.00828278e+00 5.04389524e-01
1.78240046e-01 -1.98499367e-01 2.90794820e-01 -1.27120987e-01
-2.44821876e-01 -1.31567702e-01 -1.59650886e+00 4.73603189e-01
7.41327882e-01 -5.29148817e-01 -1.38846368e-01 4.80122417e-02
3.90140384e-01 -7.45498091e-02 -6.66367769e-01 4.14124995e-01
5.42094588e-01 -1.32645416e+00 1.00821292e+00 -1.91438302e-01
6.19450927e-01 -1.22545972e-01 -1.54568538e-01 -1.21207213e+00
1.49550578e-02 1.45759359e-01 4.10310887e-02 1.29917157e+00
6.03567362e-01 -5.51743507e-01 8.97904038e-01 3.76300216e-01
6.39403760e-02 -4.84257549e-01 -1.23289168e+00 -5.04440725e-01
9.07190219e-02 -3.55021626e-01 4.23242986e-01 6.98779404e-01
-7.33659148e-01 -9.23232362e-03 -5.07050812e-01 1.89100616e-02
4.63000178e-01 -2.71842331e-01 5.49592197e-01 -1.28181446e+00
-3.96212548e-01 -5.09889066e-01 -3.09225738e-01 -6.04005933e-01
4.27649468e-01 -7.55644262e-01 -5.09153195e-02 -1.07577956e+00
-2.10724503e-01 -4.14050043e-01 2.22635344e-02 2.97378331e-01
1.30657613e-01 2.82350749e-01 4.05708969e-01 -9.46548432e-02
2.55754620e-01 8.60268295e-01 1.19800854e+00 -1.39477104e-01
-1.22348718e-01 8.55498165e-02 -2.54751146e-01 6.69267654e-01
5.53206146e-01 -4.82714951e-01 -5.41454732e-01 -4.72370088e-01
2.63454407e-01 3.36199999e-03 4.47375745e-01 -9.97265816e-01
-7.69288018e-02 1.32817164e-01 7.76695788e-01 -3.86871040e-01
1.00146735e+00 -1.09875739e+00 4.51636761e-01 5.27826548e-01
-2.26199254e-01 -3.61051202e-01 3.92821819e-01 7.21162856e-02
-2.09715694e-01 -4.39554870e-01 1.14848149e+00 8.09525177e-02
-3.96540493e-01 -1.26502201e-01 -3.09750557e-01 -4.90977615e-01
9.15466309e-01 -3.83026123e-01 2.36322910e-01 -4.49398220e-01
-7.30939984e-01 -2.62176991e-01 5.57106555e-01 3.71680856e-01
6.20338321e-01 -9.92088914e-01 -8.54712129e-01 8.07424545e-01
-1.81578785e-01 -2.08043575e-01 3.64964128e-01 6.95811927e-01
-5.18963635e-01 2.95199454e-01 -5.83723843e-01 -5.25961816e-01
-1.30585051e+00 6.37698710e-01 5.15254974e-01 -7.74562806e-02
-4.94575083e-01 1.04164159e+00 4.54086363e-01 -3.68788660e-01
4.12235618e-01 -1.92765459e-01 -3.98754328e-01 3.35884571e-01
4.03835803e-01 3.49956572e-01 2.10769355e-01 -8.19328308e-01
-1.88574806e-01 4.90168720e-01 2.67091930e-01 -2.85927653e-01
1.45170951e+00 6.64086118e-02 -2.20680743e-01 3.82532179e-01
1.16645885e+00 1.75463855e-01 -1.57814002e+00 2.27225617e-01
-1.65193066e-01 -5.09373128e-01 2.16017872e-01 -6.76391482e-01
-1.43390024e+00 1.13110471e+00 9.02804673e-01 3.89538123e-03
1.21858811e+00 -1.58004344e-01 2.30406344e-01 -2.27692917e-01
2.68940181e-01 -9.42712784e-01 2.76608206e-02 2.96134114e-01
1.11693585e+00 -1.00773537e+00 -6.71173260e-02 -4.37903374e-01
-6.93300962e-01 9.43142295e-01 4.20527875e-01 -1.61090702e-01
6.07253075e-01 6.25695288e-02 -1.29444823e-01 -2.45322078e-01
-5.87741911e-01 -6.31572679e-02 4.15824592e-01 3.33742440e-01
3.90235573e-01 -9.38100275e-03 -4.13394123e-01 3.25023979e-01
-2.65388668e-01 -1.29523814e-01 3.42849284e-01 2.71973878e-01
-1.59716785e-01 -1.18703759e+00 -5.67328453e-01 4.46803898e-01
-3.93933505e-01 2.13735834e-01 -3.80855769e-01 7.62745619e-01
5.71517408e-01 6.97785378e-01 3.61119479e-01 -6.37447760e-02
2.17034906e-01 3.48222196e-01 7.93579519e-01 -4.30501014e-01
-5.55765688e-01 2.36419216e-01 -1.45727962e-01 -3.02503169e-01
-2.58032829e-01 -6.09281421e-01 -1.03742301e+00 -2.37223521e-01
-2.77585208e-01 -1.47926295e-02 9.37090278e-01 8.20554614e-01
3.49642374e-02 5.23766100e-01 5.20688295e-01 -8.84290814e-01
-4.58847165e-01 -9.67309535e-01 -8.03499937e-01 4.17312771e-01
3.84184986e-01 -9.07602727e-01 -5.00791788e-01 7.08995610e-02] | [12.496417999267578, -0.1027742400765419] |
3d0b7474-acaa-42f0-a820-e4a7565c064c | comparative-study-of-pre-trained-bert-models | 2305.15722 | null | https://arxiv.org/abs/2305.15722v2 | https://arxiv.org/pdf/2305.15722v2.pdf | Comparative Study of Pre-Trained BERT Models for Code-Mixed Hindi-English Data | The term "Code Mixed" refers to the use of more than one language in the same text. This phenomenon is predominantly observed on social media platforms, with an increasing amount of adaptation as time goes on. It is critical to detect foreign elements in a language and process them correctly, as a considerable number of individuals are using code-mixed languages that could not be comprehended by understanding one of those languages. In this work, we focus on low-resource Hindi-English code-mixed language and enhancing the performance of different code-mixed natural language processing tasks such as sentiment analysis, emotion recognition, and hate speech identification. We perform a comparative analysis of different Transformer-based language Models pre-trained using unsupervised approaches. We have included the code-mixed models like HingBERT, HingRoBERTa, HingRoBERTa-Mixed, mBERT, and non-code-mixed models like AlBERT, BERT, and RoBERTa for comparative analysis of code-mixed Hindi-English downstream tasks. We report state-of-the-art results on respective datasets using HingBERT-based models which are specifically pre-trained on real code-mixed text. Our HingBERT-based models provide significant improvements thus highlighting the poor performance of vanilla BERT models on code-mixed text. | ['Raviraj Joshi', 'Gauri Takawane', 'Abhishek Phaltankar', 'Varad Patwardhan', 'Aryan Patil'] | 2023-05-25 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-4.45464164e-01 -1.43790856e-01 2.97527850e-01 -3.52166593e-01
-8.29437852e-01 -6.89380944e-01 5.94488144e-01 5.10692179e-01
-4.45905477e-01 3.63200188e-01 5.92638738e-02 -2.82874644e-01
4.52639401e-01 -3.70315611e-01 -4.02008176e-01 -3.05703938e-01
3.59735191e-02 2.87169158e-01 -9.15020406e-02 -7.68393099e-01
2.51603246e-01 -2.18938012e-02 -1.13987648e+00 8.42518270e-01
9.85636175e-01 3.52525890e-01 3.68478507e-01 9.06128824e-01
-3.68020207e-01 1.54928207e+00 -3.64000231e-01 -1.05014741e+00
-7.21580982e-02 -3.31719846e-01 -7.71170735e-01 -1.17936693e-01
-5.49985170e-02 5.67077063e-02 1.71256519e-03 1.23813951e+00
1.82565019e-01 -2.55945474e-01 8.06517601e-01 -1.25181043e+00
-7.83924520e-01 8.42734218e-01 -4.73326266e-01 1.40981853e-01
6.01878226e-01 -1.21957898e-01 9.32207108e-01 -1.14458585e+00
8.72921526e-01 1.05089438e+00 8.62069428e-01 5.26020825e-01
-8.90198052e-01 -6.69567525e-01 5.82039170e-02 8.16323087e-02
-1.55606198e+00 -2.19615906e-01 8.35200012e-01 -1.11598933e+00
1.44871843e+00 -9.74404588e-02 3.25921744e-01 1.24461067e+00
4.59935606e-01 9.80179131e-01 1.20819020e+00 -6.58642232e-01
-1.44892186e-01 7.29139984e-01 1.92946300e-01 7.72941768e-01
-3.86234045e-01 -5.08764029e-01 -5.87778389e-01 -3.78993116e-02
-4.32518721e-01 7.85222128e-02 5.87516204e-02 1.98985443e-01
-9.30489540e-01 1.20490217e+00 1.77858904e-01 6.59785807e-01
-1.87551305e-01 -2.04124793e-01 8.55198026e-01 5.71321011e-01
8.44518304e-01 2.34401330e-01 -4.61915463e-01 -7.32195675e-01
-1.26958096e+00 7.89768249e-02 8.84346962e-01 1.34668636e+00
7.47138083e-01 -7.88116530e-02 2.61317581e-01 1.22242844e+00
3.63739312e-01 5.18309057e-01 7.61880577e-01 -3.48944962e-01
6.43151522e-01 8.60561430e-01 -3.11646819e-01 -1.01267874e+00
-4.68253702e-01 -3.70643526e-01 -6.99366868e-01 -2.13283934e-02
1.95852548e-01 -2.50717938e-01 -5.75944543e-01 1.36085427e+00
-2.13961586e-01 -5.23333490e-01 2.75732726e-01 3.01677614e-01
8.35598111e-01 8.42739403e-01 6.15856163e-02 1.48287490e-01
1.23668849e+00 -1.35345519e+00 -5.68731308e-01 -5.57834744e-01
1.02234173e+00 -1.21505249e+00 1.10296869e+00 6.06420517e-01
-6.88174248e-01 -4.46335852e-01 -8.22626472e-01 -2.13816583e-01
-9.11240399e-01 4.11155403e-01 1.37412727e-01 9.47176516e-01
-1.02844894e+00 7.79655203e-02 -4.68356282e-01 -6.77839994e-01
1.18485764e-01 9.82305855e-02 -5.67419767e-01 -1.57512754e-01
-1.11639714e+00 8.60857189e-01 1.18735075e-01 -1.21137828e-01
-8.31642270e-01 -2.07506359e-01 -1.10697711e+00 -2.55988359e-01
-6.32834807e-02 3.26793790e-01 1.20204210e+00 -1.23898697e+00
-1.14887333e+00 1.50503397e+00 -2.95606732e-01 -4.95335728e-01
4.74174112e-01 -3.60852063e-01 -8.08075666e-01 -3.55094038e-02
2.79633284e-01 3.79287034e-01 8.88742864e-01 -1.05097008e+00
-3.89987737e-01 -5.03280684e-02 9.62212756e-02 -1.60198241e-01
-5.93293726e-01 7.03185558e-01 6.47625886e-03 -6.76355779e-01
-4.49579686e-01 -1.18868148e+00 2.14399025e-01 -3.50009501e-01
-4.82996792e-01 -7.65878856e-02 8.12918186e-01 -1.00704670e+00
1.59569228e+00 -2.65462089e+00 6.14265539e-02 -1.17139267e-02
2.21661791e-01 3.33680421e-01 -1.34758413e-01 9.42442536e-01
-1.23940773e-01 3.24285895e-01 -3.86753142e-01 -6.49301827e-01
-1.97123718e-02 1.26642317e-01 -3.66588444e-01 3.90937716e-01
3.74652803e-01 6.26113117e-01 -9.60363328e-01 -4.14550364e-01
-3.57559919e-02 3.88072580e-01 -8.83370817e-01 4.55563873e-01
6.89221397e-02 1.26529813e-01 -6.67331368e-02 6.78282201e-01
5.48623323e-01 3.46176215e-02 -4.24685329e-02 4.35771912e-01
-5.50565660e-01 2.38472655e-01 -5.51675200e-01 1.50434101e+00
-1.01838684e+00 1.02806127e+00 2.86320269e-01 -6.84243679e-01
1.02624750e+00 2.61184752e-01 -1.70580223e-02 -3.60253036e-01
2.97883004e-01 4.77054477e-01 7.58532360e-02 -8.22977722e-01
6.71602249e-01 -1.93469793e-01 -4.85859990e-01 2.82692850e-01
1.86538368e-01 -3.56003314e-01 4.83655483e-01 5.44522703e-01
1.08754599e+00 -1.44399345e-01 5.35572112e-01 -5.33440948e-01
9.74685431e-01 1.62310585e-01 6.90814853e-02 5.02344191e-01
-3.71900469e-01 7.49924302e-01 6.94309413e-01 -1.19705714e-01
-8.91358852e-01 -7.37823367e-01 -4.34002355e-02 1.56622350e+00
-3.67397666e-01 -9.49032843e-01 -8.31576049e-01 -6.81958914e-01
-3.44974518e-01 8.14606488e-01 -6.65584147e-01 -8.82429257e-02
-1.73886135e-01 -6.27584398e-01 1.01227367e+00 8.61955211e-02
5.59355736e-01 -8.10613751e-01 -4.70397353e-01 2.92805225e-01
-3.28681141e-01 -1.18283856e+00 -4.11279827e-01 4.73410964e-01
-1.18117295e-02 -9.11382437e-01 -5.87574720e-01 -9.70797956e-01
4.78839517e-01 -9.19115469e-02 1.20105886e+00 4.06930715e-01
-1.65591285e-01 3.99376452e-01 -9.31030929e-01 -5.87601185e-01
-1.04382551e+00 4.10250157e-01 -2.82364756e-01 5.77023439e-02
7.11234987e-01 -7.47214407e-02 1.27072558e-01 -1.42009854e-01
-8.64008129e-01 5.58598973e-02 3.63943249e-01 8.34893703e-01
-4.54561502e-01 -1.98808968e-01 1.85425013e-01 -1.22657692e+00
6.62228346e-01 -9.02784586e-01 -2.97942180e-02 1.22625872e-01
-2.49294803e-01 -7.56979138e-02 1.02917314e+00 -4.96923476e-01
-9.69037533e-01 1.05352238e-01 -6.29520535e-01 4.08220279e-04
-1.45366877e-01 8.05415988e-01 2.52834737e-01 2.36649588e-02
7.62292922e-01 2.65624762e-01 -1.57824248e-01 -3.17554533e-01
1.67578325e-01 1.45097935e+00 7.92036951e-02 -4.74659562e-01
7.62414455e-01 3.11180055e-01 -7.02186108e-01 -1.22450984e+00
-7.16193795e-01 -8.80143464e-01 -9.76135314e-01 -3.40905339e-01
1.22067070e+00 -1.23838174e+00 -1.44093767e-01 1.06407487e+00
-1.40423536e+00 -4.35453296e-01 4.16181803e-01 1.46943435e-01
-4.56427783e-02 3.67827207e-01 -1.13068330e+00 -9.33869600e-01
-6.57599419e-02 -1.29278731e+00 1.07203746e+00 -3.67506623e-01
-5.19695580e-01 -1.26854527e+00 3.08863312e-01 3.26910228e-01
4.93186712e-01 -4.93205898e-02 1.15405929e+00 -1.08293498e+00
2.10145473e-01 -2.01303348e-01 -1.09612025e-01 5.73373973e-01
-1.41889557e-01 3.83870155e-01 -1.11611474e+00 -3.03347081e-01
6.68296637e-03 -6.12554491e-01 5.22942662e-01 -3.78647804e-01
4.54999954e-01 -3.01174432e-01 1.17362574e-01 2.58418530e-01
1.32324958e+00 1.54177412e-01 3.34300518e-01 4.36861783e-01
8.00151885e-01 6.45736337e-01 3.50256681e-01 6.18970215e-01
6.87836289e-01 3.77238393e-01 2.13644207e-01 1.22076944e-01
2.06898704e-01 -1.67114899e-01 8.76738846e-01 1.67854798e+00
3.35208803e-01 -1.86062336e-01 -1.45588303e+00 7.23419309e-01
-1.62727320e+00 -7.91930556e-01 -5.88901281e-01 1.54713118e+00
9.57873523e-01 2.03030393e-01 7.58669227e-02 -4.14171889e-02
7.58218527e-01 1.49823159e-01 4.42794636e-02 -9.89449620e-01
-2.46547461e-01 -6.92212582e-02 1.05622143e-01 5.94956040e-01
-1.21970868e+00 1.01013708e+00 5.63934803e+00 8.58096719e-01
-1.30101275e+00 5.60338438e-01 5.27683198e-01 1.40080944e-01
-1.05082326e-01 1.84249151e-02 -6.08352721e-01 5.40120661e-01
1.24373281e+00 -2.24752545e-01 6.74454451e-01 1.23456633e+00
-1.61701441e-01 -4.47395630e-02 -9.92456734e-01 1.29628575e+00
5.92045009e-01 -7.50976145e-01 -3.13127667e-01 -2.19257429e-01
7.39736795e-01 4.75356132e-01 -6.21803440e-02 8.93651664e-01
2.81671494e-01 -7.54720092e-01 1.32034457e+00 -9.18473229e-02
5.30730844e-01 -6.38832629e-01 9.34478223e-01 5.07477164e-01
-1.17639875e+00 -2.46811494e-01 -1.98443234e-01 -3.85420233e-01
-1.84473857e-01 6.89958036e-01 -4.39477026e-01 3.01317364e-01
6.93415701e-01 1.11161864e+00 -1.15503240e+00 4.12468374e-01
-3.59786451e-01 7.47012913e-01 -3.67928781e-02 -2.98083901e-01
4.99161541e-01 -8.04372281e-02 3.96155715e-01 2.01819396e+00
3.61916244e-01 -2.79202670e-01 1.87258258e-01 7.38990188e-01
-6.14176691e-02 5.43056965e-01 -7.48709142e-01 -4.62542146e-01
2.96912808e-02 1.48804986e+00 -7.90705383e-01 -4.04724598e-01
-1.03376722e+00 1.19259012e+00 5.47552526e-01 1.07112840e-01
-8.51820946e-01 -6.30770743e-01 3.82753938e-01 1.38807416e-01
-2.08893027e-02 -3.72908086e-01 1.04037680e-01 -1.43442619e+00
-1.90677717e-01 -1.14146781e+00 3.93007547e-01 -1.01152909e+00
-1.48084760e+00 1.15795743e+00 -1.89316720e-01 -1.18481278e+00
-4.34381396e-01 -1.03236389e+00 -4.85443264e-01 5.88404238e-01
-1.31200361e+00 -1.24978447e+00 -9.33374166e-02 6.34197652e-01
8.48630846e-01 -5.51303446e-01 7.18849659e-01 6.37781024e-01
-5.04749835e-01 4.57382411e-01 2.92431533e-01 7.38633037e-01
8.90481174e-01 -1.18815267e+00 2.07973868e-01 1.14767456e+00
3.44413221e-02 8.41059744e-01 6.74514532e-01 -5.65801263e-01
-1.17764616e+00 -9.88563895e-01 1.32264721e+00 -6.05413139e-01
1.11325622e+00 -1.10047472e+00 -8.45212996e-01 6.34200335e-01
6.25308216e-01 -3.61290723e-01 8.45972419e-01 5.77256717e-02
-7.12284803e-01 3.09875727e-01 -1.10868013e+00 4.63644594e-01
6.39534116e-01 -1.28405619e+00 -6.24952912e-01 3.51600736e-01
3.89309108e-01 -1.73767939e-01 -5.29637277e-01 -2.32093021e-01
1.76495209e-01 -1.28509450e+00 2.42840752e-01 -3.04296583e-01
9.76974845e-01 -2.82810070e-02 -1.61338806e-01 -1.37918139e+00
-1.03517853e-01 -6.16774976e-01 5.19368887e-01 1.56580055e+00
5.46970487e-01 -4.18716013e-01 7.13511556e-02 5.33034325e-01
-1.30962253e-01 -3.32501918e-01 -7.30856895e-01 -6.10607445e-01
5.64018428e-01 -5.97476244e-01 1.34462699e-01 1.31598604e+00
5.44297278e-01 2.83499151e-01 -2.86395788e-01 -2.61592865e-01
7.83503056e-02 -9.78365839e-02 6.32085264e-01 -9.63865101e-01
-1.38734207e-01 -3.99315029e-01 -5.23173273e-01 -6.46269262e-01
6.85207665e-01 -1.34377134e+00 2.32880846e-01 -1.14636517e+00
3.92491847e-01 3.35199125e-02 1.89263552e-01 6.02109373e-01
9.59898755e-02 4.69276130e-01 3.74205858e-01 2.19346941e-01
-5.64334989e-01 3.23661417e-01 3.10663551e-01 -3.55192512e-01
3.87710370e-02 -3.17537099e-01 -4.87139881e-01 9.47894037e-01
5.96591234e-01 -7.78060079e-01 -1.96177661e-01 -3.59446824e-01
9.02136087e-01 -5.01890421e-01 1.04855917e-01 -1.04785442e+00
1.07273050e-01 2.52215922e-01 -5.21239452e-02 -1.38272047e-01
-9.38402861e-03 -8.01228821e-01 -2.39565164e-01 5.90497315e-01
-3.61675411e-01 4.36306834e-01 2.56241590e-01 -6.35345057e-02
-5.66114008e-01 -7.03678787e-01 9.00107741e-01 -3.75605017e-01
-8.26969683e-01 -2.81510770e-01 -1.28260267e+00 4.43100035e-01
9.25623178e-01 1.43184423e-01 -3.03791434e-01 -5.73697031e-01
-5.11105955e-01 -8.67170617e-02 5.19620717e-01 7.88397431e-01
4.46572393e-01 -9.18000937e-01 -7.97069490e-01 2.81689793e-01
5.70550323e-01 -6.98087394e-01 4.31126356e-02 8.54866624e-01
-9.07909691e-01 2.48137400e-01 -1.95632622e-01 -3.49295676e-01
-1.30706048e+00 6.75178587e-01 1.50928706e-01 -2.72040486e-01
-1.08004339e-01 7.97319591e-01 -3.56919132e-02 -9.85468328e-01
-2.48074815e-01 -1.82862356e-01 -2.17078313e-01 3.35245699e-01
4.38199848e-01 2.10694537e-01 1.25934303e-01 -1.28526139e+00
-6.92734122e-01 5.48868954e-01 -1.81509390e-01 -1.07524402e-01
1.26593375e+00 -2.76911855e-01 -7.58384645e-01 8.91648769e-01
1.56001699e+00 5.97431064e-01 -4.66675758e-01 2.17001468e-01
2.27491781e-01 -3.29434097e-01 -2.09681034e-01 -6.54460132e-01
-7.92903423e-01 1.20905507e+00 3.43785226e-01 3.09298068e-01
9.09155607e-01 7.25675225e-02 5.23053825e-01 4.37502056e-01
5.21439195e-01 -1.30035996e+00 6.25338731e-03 1.14621162e+00
7.88822174e-01 -1.22574043e+00 -4.02837485e-01 -3.89464408e-01
-1.01108003e+00 1.23866820e+00 5.30627608e-01 -3.04832477e-02
1.06047142e+00 5.01775503e-01 4.96900290e-01 -1.11362383e-01
-7.16640174e-01 -2.49991283e-01 8.08419958e-02 3.12455863e-01
1.09052658e+00 -9.84110460e-02 -2.00826064e-01 7.68303752e-01
-4.84420061e-01 -3.25729609e-01 9.44653094e-01 1.14627922e+00
-1.09985299e-01 -1.14174616e+00 -2.66258895e-01 3.10961753e-01
-8.26645434e-01 -6.11851275e-01 -6.62760556e-01 7.62366712e-01
2.37737224e-01 1.38092351e+00 -1.81663930e-01 -6.01022601e-01
4.61613052e-02 4.69531417e-01 -7.33736008e-02 -1.01496673e+00
-1.23894942e+00 -8.29473361e-02 5.87182492e-02 -1.40719503e-01
-5.48043251e-01 -6.50530100e-01 -1.22982311e+00 -3.42435300e-01
-5.05913831e-02 3.33367944e-01 7.10589588e-01 1.02061868e+00
1.50563076e-01 2.17646375e-01 3.18314672e-01 -5.66450953e-01
3.30445431e-02 -1.02313924e+00 -6.17663026e-01 8.21946502e-01
4.17738110e-01 -2.50803769e-01 -6.05156362e-01 3.17682177e-01] | [9.247820854187012, 10.455854415893555] |
0c0753a3-3736-483e-9e0f-52849c4d33ea | mer-2023-multi-label-learning-modality | 2304.08981 | null | https://arxiv.org/abs/2304.08981v1 | https://arxiv.org/pdf/2304.08981v1.pdf | MER 2023: Multi-label Learning, Modality Robustness, and Semi-Supervised Learning | Over the past few decades, multimodal emotion recognition has made remarkable progress with the development of deep learning. However, existing technologies are difficult to meet the demand for practical applications. To improve the robustness, we launch a Multimodal Emotion Recognition Challenge (MER 2023) to motivate global researchers to build innovative technologies that can further accelerate and foster research. For this year's challenge, we present three distinct sub-challenges: (1) MER-MULTI, in which participants recognize both discrete and dimensional emotions; (2) MER-NOISE, in which noise is added to test videos for modality robustness evaluation; (3) MER-SEMI, which provides large amounts of unlabeled samples for semi-supervised learning. In this paper, we test a variety of multimodal features and provide a competitive baseline for each sub-challenge. Our system achieves 77.57% on the F1 score and 0.82 on the mean squared error (MSE) for MER-MULTI, 69.82% on the F1 score and 1.12 on MSE for MER-NOISE, and 86.75% on the F1 score for MER-SEMI, respectively. Baseline code is available at https://github.com/zeroQiaoba/MER2023-Baseline. | ['JianHua Tao', 'Björn W. Schuller', 'Guoying Zhao', 'Erik Cambria', 'Meng Wang', 'Jiangyan Yi', 'Bin Liu', 'Ye Liu', 'Jinming Zhao', 'Licai Sun', 'Haiyang Sun', 'Zheng Lian'] | 2023-04-18 | null | null | null | null | ['multimodal-emotion-recognition', 'multi-label-learning', 'multimodal-emotion-recognition'] | ['computer-vision', 'methodology', 'speech'] | [-1.37446463e-01 -4.48665082e-01 1.37609690e-01 -5.87320983e-01
-1.17565775e+00 -4.63960975e-01 2.57097542e-01 -1.57384202e-01
-7.03739405e-01 4.00289446e-01 9.41962972e-02 2.61591405e-01
5.78831434e-01 -1.96514383e-01 -5.74484944e-01 -5.39214373e-01
-1.10633232e-01 -2.28754848e-01 -4.02816206e-01 -2.50552773e-01
-1.87007353e-01 7.09875152e-02 -1.43857920e+00 6.11168861e-01
9.19598043e-01 1.51063037e+00 -3.17297310e-01 7.94564426e-01
3.89695227e-01 4.93288815e-01 -4.97682422e-01 -7.05413699e-01
-7.08280355e-02 -5.07471621e-01 -5.11370480e-01 -9.77721345e-03
2.30452329e-01 -4.54202741e-01 -4.47349459e-01 1.21398687e+00
9.55434263e-01 3.13090473e-01 5.59303701e-01 -1.74801981e+00
-4.83500093e-01 1.27877861e-01 -6.54147625e-01 -2.25715801e-01
6.08149648e-01 2.53377259e-01 6.37303889e-01 -1.23912334e+00
5.58948338e-01 1.21501076e+00 7.00874209e-01 8.18870604e-01
-7.91335344e-01 -9.09499943e-01 -2.07548798e-03 3.37397367e-01
-1.39720273e+00 -7.12475181e-01 5.98886847e-01 -1.33241832e-01
6.51563525e-01 3.15029919e-01 1.75082937e-01 1.77261901e+00
-2.23388113e-02 1.12879014e+00 1.18959296e+00 -2.21139461e-01
1.57790661e-01 1.83323935e-01 3.54320183e-02 6.63284063e-01
-3.13508630e-01 -4.43488397e-02 -7.25776553e-01 9.77753252e-02
4.32684451e-01 -2.18250290e-01 -2.40816355e-01 1.05215870e-01
-1.03723729e+00 7.35868812e-01 1.60539001e-01 1.03104234e-01
-2.04807803e-01 -4.90951687e-02 9.07462537e-01 5.50511539e-01
1.61330849e-01 5.82659319e-02 -2.76097447e-01 -7.27493346e-01
-5.05840600e-01 -4.13368903e-02 7.53283978e-01 8.08261514e-01
2.67131299e-01 1.60082355e-01 6.93658963e-02 1.26766992e+00
2.57307172e-01 7.29660928e-01 5.88864386e-01 -1.05722082e+00
5.20015717e-01 4.30616587e-01 1.00547761e-01 -1.07975399e+00
-5.28033674e-01 -1.62324160e-01 -1.28039348e+00 -1.69942424e-01
2.10614115e-01 -5.80444753e-01 -5.81493199e-01 1.86994815e+00
1.63361326e-01 -1.64228547e-02 2.67293900e-01 1.05282962e+00
1.28303540e+00 8.12377274e-01 7.34705850e-02 -2.05185071e-01
1.23010802e+00 -1.21590114e+00 -8.70510459e-01 -3.08736432e-02
5.65607667e-01 -8.51920545e-01 1.23735714e+00 7.13235676e-01
-1.00770307e+00 -5.94629407e-01 -9.21349049e-01 1.36697218e-01
-3.62911999e-01 5.65111578e-01 4.05866206e-01 8.29405665e-01
-8.73552442e-01 1.86988622e-01 -7.10892975e-01 -3.02664489e-01
2.28450611e-01 3.11306238e-01 -6.90320432e-01 -2.39924714e-01
-1.42288220e+00 6.06840014e-01 -4.51013520e-02 3.86756361e-01
-9.45847929e-01 -2.17088476e-01 -9.57037389e-01 -6.86370209e-02
2.43316680e-01 -4.03226465e-02 1.20733654e+00 -1.18498540e+00
-1.58995295e+00 8.61340463e-01 -1.30005538e-01 8.79613459e-02
5.77941656e-01 -5.69611132e-01 -1.05800760e+00 2.73452818e-01
-1.38600111e-01 8.26436758e-01 7.56179631e-01 -1.13947701e+00
-4.42366898e-01 -3.08358997e-01 -2.99502671e-01 3.17629457e-01
-6.54867887e-01 2.92247385e-01 -8.14608812e-01 -5.59727907e-01
-2.75296748e-01 -1.11251450e+00 3.07663530e-02 -2.02426657e-01
-4.16514128e-01 -2.14471251e-01 6.45527661e-01 -8.17623854e-01
1.17048597e+00 -2.60149693e+00 -2.61404756e-02 6.80225864e-02
-9.11628008e-02 2.81331420e-01 -6.73277318e-01 4.18343663e-01
-2.18703344e-01 1.83992654e-01 9.75209177e-02 -5.99082768e-01
4.30704743e-01 -2.74102896e-01 -6.83938488e-02 3.24456245e-01
2.32834548e-01 9.98498738e-01 -5.78068972e-01 -1.68333143e-01
-1.29639246e-02 6.51680112e-01 -2.39482507e-01 4.47587609e-01
3.26846331e-01 2.22301260e-01 -1.02942355e-01 1.02657998e+00
7.86018848e-01 -9.74450931e-02 7.94963464e-02 -4.40713793e-01
1.44079357e-01 -3.96555573e-01 -1.49003899e+00 1.53942227e+00
-2.07515478e-01 7.60989785e-01 3.66821587e-01 -8.85121524e-01
8.63070130e-01 5.15485466e-01 4.15153027e-01 -9.09439862e-01
3.78252685e-01 1.64059833e-01 -2.95408130e-01 -8.57097447e-01
5.75551867e-01 6.16241172e-02 -5.36898077e-01 2.52030760e-01
2.24795654e-01 3.68942320e-01 1.33913562e-01 3.41966331e-01
8.56552303e-01 -1.40642926e-01 -5.74877784e-02 3.20145100e-01
3.55907410e-01 -4.40729767e-01 6.53034985e-01 5.44298887e-01
-7.45726049e-01 6.26587868e-01 5.18552184e-01 -1.82414427e-01
-6.26546800e-01 -9.28915203e-01 1.15344644e-01 1.20944226e+00
1.68430552e-01 -3.85874718e-01 -8.22757304e-01 -6.58642709e-01
-2.64501542e-01 3.24671537e-01 -5.98597765e-01 -3.65498632e-01
1.21355522e-02 -7.77388453e-01 9.88821805e-01 5.43989003e-01
8.66293192e-01 -1.00874567e+00 -4.06347275e-01 -2.56942660e-01
-7.51469016e-01 -1.21575654e+00 -6.67768061e-01 -2.22053200e-01
-3.47789824e-01 -9.01224911e-01 -8.80657017e-01 -6.84031546e-01
4.36829627e-01 6.30223453e-02 9.05872941e-01 -1.18213393e-01
-2.05750257e-01 6.94728911e-01 -6.20810688e-01 -3.00168395e-01
7.67691359e-02 -1.89088151e-01 3.36621225e-01 3.69406044e-01
4.45341378e-01 -1.01564117e-01 -4.41418052e-01 5.60250580e-01
-9.29478765e-01 -2.12242857e-01 5.15562057e-01 9.92579401e-01
3.66392225e-01 -1.88651513e-02 8.13010395e-01 -2.81805605e-01
7.95248270e-01 -5.25220275e-01 -1.33543491e-01 3.59309047e-01
-1.99142337e-01 -5.75597763e-01 3.81045043e-01 -8.74169171e-01
-8.98141503e-01 -6.65353239e-02 -3.27051610e-01 -4.00540888e-01
-3.90227646e-01 6.33160770e-01 -4.60157096e-01 5.68198822e-02
4.07294810e-01 2.65322208e-01 -1.48680240e-01 -1.71934366e-01
1.86065942e-01 1.07101738e+00 5.70493460e-01 -4.90992546e-01
2.91002542e-01 1.36722729e-01 -3.92422497e-01 -9.04599547e-01
-6.25334501e-01 -3.45146388e-01 -7.85948485e-02 -6.14648223e-01
8.32911313e-01 -1.28760362e+00 -1.03309500e+00 8.75193417e-01
-8.76820385e-01 -4.04693544e-01 3.79403383e-01 5.81663311e-01
-2.70877063e-01 4.78146732e-01 -8.64253342e-01 -9.51909244e-01
-4.62185740e-01 -1.27887249e+00 9.71640289e-01 5.98354578e-01
-2.36421466e-01 -7.66644239e-01 -8.23257864e-02 6.09929621e-01
4.58483696e-01 3.70072305e-01 2.26294309e-01 -5.70239484e-01
4.94490378e-02 -4.01891142e-01 -3.58385623e-01 6.99560523e-01
-2.54708856e-01 8.65989551e-02 -1.09582210e+00 -4.40624952e-01
-2.17850834e-01 -1.22516263e+00 6.20742261e-01 7.36816674e-02
1.10390151e+00 1.17695458e-01 8.44934508e-02 5.11401772e-01
9.47620451e-01 3.47116262e-01 8.24590325e-01 1.93277985e-01
4.65776205e-01 4.61492777e-01 7.54497051e-01 5.73179841e-01
5.55292189e-01 5.90098798e-01 2.87935108e-01 -1.29684299e-01
3.46206248e-01 -6.05480187e-02 7.57140398e-01 1.02322805e+00
1.00998625e-01 -2.62168825e-01 -8.15165222e-01 4.23032165e-01
-1.84246016e+00 -8.41569602e-01 2.91945115e-02 1.98169208e+00
8.04499865e-01 -2.36436039e-01 3.16661775e-01 -2.20185257e-02
6.44520879e-01 5.74898757e-02 -5.53013325e-01 -4.36752439e-01
-3.78260165e-01 -1.87221393e-01 -2.83854187e-01 2.54915774e-01
-1.41108322e+00 7.13918269e-01 5.47002125e+00 8.34794343e-01
-1.25834048e+00 2.93238852e-02 9.10865545e-01 -1.84700668e-01
1.18778259e-01 -6.45873547e-01 -6.26361072e-01 7.38680243e-01
1.06720972e+00 1.53627917e-01 4.36962813e-01 8.44353318e-01
2.82086730e-02 -9.40871909e-02 -8.07972014e-01 1.64276087e+00
4.55891758e-01 -6.66355312e-01 -2.12292686e-01 -1.88929871e-01
8.17968369e-01 -8.15922841e-02 3.71983796e-01 7.99399495e-01
-1.96904793e-01 -1.28622174e+00 3.48798037e-01 6.31662965e-01
1.03358293e+00 -1.18754888e+00 1.11008930e+00 1.85537234e-01
-1.15652657e+00 8.51923153e-02 -2.54952341e-01 2.13808000e-01
5.83280474e-02 4.31000769e-01 2.88114753e-02 5.38799465e-01
9.54516828e-01 4.82025921e-01 -4.40370500e-01 8.79564464e-01
-1.01759039e-01 6.32630408e-01 -2.24635825e-01 -1.63828433e-01
1.53249905e-01 -1.33712038e-01 2.94653207e-01 1.57304764e+00
2.74537027e-01 1.25974834e-01 9.34609100e-02 2.54287481e-01
-6.24396443e-01 1.49921492e-01 -2.57010847e-01 -3.70452106e-01
4.59118545e-01 1.58428049e+00 -2.59028316e-01 -2.51972467e-01
-3.66989672e-01 1.33620322e+00 2.24181309e-01 5.64072311e-01
-1.10237837e+00 -8.70674193e-01 5.95765591e-01 -8.10985982e-01
7.36305118e-02 -7.17095882e-02 1.57823153e-02 -1.54470587e+00
1.01419255e-01 -1.56947935e+00 4.61732656e-01 -9.56377685e-01
-1.43786597e+00 8.19068193e-01 -4.63434517e-01 -1.17443657e+00
-9.89141092e-02 -7.16830850e-01 -4.82041895e-01 6.54766202e-01
-1.03980207e+00 -9.40021217e-01 -6.17453456e-01 7.20925987e-01
3.68735582e-01 -3.70755166e-01 1.00518322e+00 6.02985442e-01
-1.08325768e+00 1.35205698e+00 2.59570003e-01 3.94783914e-01
1.26430237e+00 -7.94105768e-01 -8.37348104e-02 7.30872393e-01
-4.97868471e-02 4.27686661e-01 3.54894131e-01 -3.42742890e-01
-1.68686199e+00 -9.22981143e-01 7.47917712e-01 -2.78787792e-01
5.21471083e-01 -4.72956628e-01 -7.51044214e-01 3.02143186e-01
2.84940243e-01 -3.06982901e-02 9.92998421e-01 3.01250547e-01
-5.97937226e-01 -4.90464754e-02 -1.12302768e+00 7.39623070e-01
5.22000134e-01 -6.95063293e-01 -5.41470610e-02 2.84249000e-02
4.44899678e-01 -4.31844085e-01 -1.13050747e+00 5.98281443e-01
8.81948292e-01 -8.56028259e-01 5.43932855e-01 -6.78768635e-01
6.03515327e-01 5.25594465e-02 -5.76820970e-01 -1.27023494e+00
6.84595183e-02 -6.54837430e-01 -1.72713742e-01 1.35437167e+00
3.51377219e-01 -3.12956810e-01 6.46270871e-01 8.98906291e-01
5.72508872e-02 -9.85025704e-01 -8.42151701e-01 -6.68056130e-01
-1.95558652e-01 -6.19286656e-01 7.59317949e-02 1.20183837e+00
1.95122585e-01 2.74986058e-01 -7.58747995e-01 3.74134146e-02
4.83505756e-01 -1.69363499e-01 9.63452041e-01 -6.42542541e-01
-9.70276594e-02 -4.61362600e-01 -1.69723600e-01 -1.05231404e+00
5.85852154e-02 -3.27392817e-01 -2.51247548e-03 -1.10510230e+00
4.63278204e-01 1.79927245e-01 -4.70636755e-01 5.73620379e-01
-2.88433313e-01 5.46295524e-01 2.95149088e-01 -1.81153387e-01
-1.29471195e+00 1.08585477e+00 8.52407157e-01 -3.04344352e-02
-1.16851754e-01 -2.76052147e-01 -7.37237930e-01 6.50259554e-01
9.55177784e-01 -1.96555734e-01 -8.14000219e-02 -3.03492993e-01
2.20203683e-01 -9.97322705e-03 2.99384087e-01 -9.05402362e-01
1.60231069e-01 1.00206956e-03 6.36012614e-01 -4.58314091e-01
7.08595932e-01 -6.36079669e-01 -1.74220994e-01 1.54490799e-01
-4.04008687e-01 2.88671792e-01 5.30309379e-01 3.02894354e-01
-4.81845319e-01 3.07454960e-03 6.59205556e-01 3.74999225e-01
-8.73418093e-01 1.43564537e-01 -4.03932959e-01 1.88204110e-01
8.74331713e-01 9.54239294e-02 -5.02468944e-01 -9.17541146e-01
-7.44840622e-01 5.63914716e-01 2.71349788e-01 6.27225161e-01
9.52666819e-01 -1.65897775e+00 -6.93688750e-01 5.85491396e-02
2.40316540e-01 -6.04017019e-01 9.00465310e-01 1.06587970e+00
-1.53957471e-01 7.27227926e-02 -3.62710536e-01 -3.38471293e-01
-1.65254629e+00 1.56425878e-01 3.84807616e-01 -8.18395093e-02
2.11870726e-02 9.74254966e-01 -7.16142803e-02 -6.74718380e-01
6.77766323e-01 3.08059216e-01 6.38397783e-02 9.42342207e-02
9.26385641e-01 5.48502862e-01 -1.74085766e-01 -8.49858105e-01
-4.75136012e-01 2.52437472e-01 -2.98347883e-02 -3.11792582e-01
1.18661106e+00 -2.38572389e-01 2.18852144e-02 5.41854799e-01
1.49233949e+00 -1.16367534e-01 -1.20379353e+00 -6.69501424e-02
-4.21550989e-01 -1.80848703e-01 -6.64458349e-02 -1.34934616e+00
-1.01766860e+00 1.10121012e+00 9.72587407e-01 -6.26814887e-02
1.46835268e+00 -3.38659972e-01 8.27733278e-01 5.10453641e-01
-2.91860523e-03 -1.29006422e+00 4.74917978e-01 7.47810423e-01
1.06116176e+00 -1.59876382e+00 -4.21185851e-01 -9.66463909e-02
-1.18924642e+00 1.01260579e+00 7.21062005e-01 2.75110126e-01
3.69125724e-01 1.62917584e-01 5.09574592e-01 5.06137833e-02
-7.92134702e-01 3.86633095e-03 5.18993795e-01 4.24915761e-01
4.95815158e-01 8.80443305e-02 2.48757787e-02 1.24185503e+00
1.43026888e-01 -9.71678738e-03 2.09520265e-01 6.62221313e-01
-1.24849685e-01 -8.67176056e-01 -3.81915271e-01 2.43512511e-01
-4.11174595e-01 5.91362640e-02 -5.63522756e-01 7.77046621e-01
4.63242903e-02 1.31520498e+00 -3.13868940e-01 -9.96060848e-01
4.26477402e-01 1.60406217e-01 2.12864459e-01 -1.63188241e-02
-3.66115689e-01 2.74964362e-01 3.25705230e-01 -7.83366501e-01
-2.06021711e-01 -6.77322090e-01 -1.00878596e+00 -3.83664280e-01
-2.66377069e-02 2.89932370e-01 7.77508557e-01 5.90756953e-01
5.57090282e-01 2.62650490e-01 7.07415998e-01 -8.26972604e-01
-5.82644105e-01 -1.11909270e+00 -4.38057721e-01 5.83102345e-01
1.06586674e-02 -4.76303548e-01 -2.06086248e-01 -3.01789679e-03] | [13.255877494812012, 5.129270076751709] |
01c0eaa7-737c-4b16-bc5e-3ce4a8746b22 | metamorphosis-task-oriented-privacy-cognizant | 2305.07815 | null | https://arxiv.org/abs/2305.07815v1 | https://arxiv.org/pdf/2305.07815v1.pdf | MetaMorphosis: Task-oriented Privacy Cognizant Feature Generation for Multi-task Learning | With the growth of computer vision applications, deep learning, and edge computing contribute to ensuring practical collaborative intelligence (CI) by distributing the workload among edge devices and the cloud. However, running separate single-task models on edge devices is inefficient regarding the required computational resource and time. In this context, multi-task learning allows leveraging a single deep learning model for performing multiple tasks, such as semantic segmentation and depth estimation on incoming video frames. This single processing pipeline generates common deep features that are shared among multi-task modules. However, in a collaborative intelligence scenario, generating common deep features has two major issues. First, the deep features may inadvertently contain input information exposed to the downstream modules (violating input privacy). Second, the generated universal features expose a piece of collective information than what is intended for a certain task, in which features for one task can be utilized to perform another task (violating task privacy). This paper proposes a novel deep learning-based privacy-cognizant feature generation process called MetaMorphosis that limits inference capability to specific tasks at hand. To achieve this, we propose a channel squeeze-excitation based feature metamorphosis module, Cross-SEC, to achieve distinct attention of all tasks and a de-correlation loss function with differential-privacy to train a deep learning model that produces distinct privacy-aware features as an output for the respective tasks. With extensive experimentation on four datasets consisting of diverse images related to scene understanding and facial attributes, we show that MetaMorphosis outperforms recent adversarial learning and universal feature generation methods by guaranteeing privacy requirements in an efficient way for image and video analytics. | ['Anupam Das', 'Md Yusuf Sarwar Uddin', 'Zhouyu Li', 'Md Adnan Arefeen'] | 2023-05-13 | null | null | null | null | ['scene-understanding', 'edge-computing'] | ['computer-vision', 'time-series'] | [ 2.65613317e-01 4.01288383e-02 8.64467695e-02 -5.45998216e-01
-7.20408142e-01 -7.51046777e-01 4.49566603e-01 -2.28890374e-01
-4.22568470e-01 5.29182613e-01 -2.65321761e-01 -2.57496238e-01
7.58536085e-02 -8.13756108e-01 -8.76161993e-01 -8.95118296e-01
1.87415481e-01 9.67256445e-03 -1.62216112e-01 3.08238775e-01
-2.14404583e-01 4.46428537e-01 -1.46341753e+00 5.88595867e-01
6.54029548e-01 1.46128535e+00 -1.10889815e-01 5.76981783e-01
1.84486270e-01 6.04419768e-01 -6.75399244e-01 -9.41939175e-01
9.27051961e-01 -1.40605643e-01 -4.59727198e-01 -9.08079818e-02
4.58840966e-01 -6.00457489e-01 -3.50317448e-01 1.31222892e+00
5.00540614e-01 -9.36226547e-02 3.79997194e-01 -1.94325626e+00
-7.11715937e-01 2.66370028e-01 -4.88796622e-01 -1.99028298e-01
-1.66263446e-01 4.00806636e-01 7.51532674e-01 -5.36962032e-01
3.92491668e-01 6.86226606e-01 6.05077565e-01 9.36861992e-01
-9.49278116e-01 -1.16295111e+00 1.85072925e-02 -5.44222631e-02
-1.38311672e+00 -4.86617535e-01 6.83876574e-01 -2.88880646e-01
5.98413050e-01 5.71235836e-01 4.70945865e-01 1.31519914e+00
5.43775260e-01 8.49956572e-01 1.09800470e+00 3.22859555e-01
4.48013157e-01 3.95738631e-01 -2.41054729e-01 6.60122097e-01
3.05270046e-01 1.62306592e-01 -7.53566444e-01 -4.23186421e-01
5.42504668e-01 4.30548728e-01 -1.33080274e-01 -2.62717247e-01
-6.67000055e-01 6.95835114e-01 2.59466559e-01 -1.91927895e-01
-3.51771057e-01 2.21034601e-01 7.79405475e-01 6.45401299e-01
5.50480723e-01 1.82881132e-01 -8.29733729e-01 1.43218741e-01
-8.43744278e-01 2.26952881e-01 7.75190890e-01 1.08710110e+00
1.02825880e+00 -1.77341521e-01 -4.91098970e-01 -1.21877175e-02
5.37019409e-02 3.79610181e-01 3.03201914e-01 -8.58198106e-01
3.56700033e-01 5.19573569e-01 -8.43390897e-02 -1.10232186e+00
-1.22816227e-01 -3.30597818e-01 -1.07656300e+00 5.18913507e-01
3.65497857e-01 -6.60176218e-01 -7.43495882e-01 1.91278982e+00
4.59790736e-01 3.41612041e-01 2.41253659e-01 1.06476557e+00
6.44207478e-01 3.53192657e-01 1.04726829e-01 5.71828112e-02
1.48983657e+00 -8.35696340e-01 -5.56615591e-01 -6.41769916e-02
5.62607944e-01 -5.06636024e-01 7.98690736e-01 3.28473270e-01
-6.98770821e-01 -3.41327608e-01 -8.83830070e-01 -2.62214959e-01
-4.12037671e-01 1.16660029e-01 9.69495416e-01 9.90451992e-01
-1.03146088e+00 3.13560635e-01 -6.80588186e-01 1.25420958e-01
1.12543917e+00 7.04481483e-01 -6.26242876e-01 -7.93719515e-02
-9.38207090e-01 5.35055138e-02 1.96426976e-02 -1.17095962e-01
-1.17571104e+00 -1.02246869e+00 -7.65320778e-01 2.89446145e-01
4.01019573e-01 -1.15788674e+00 7.84398317e-01 -1.25734055e+00
-1.51699269e+00 9.20439959e-01 1.15939617e-01 -5.83752871e-01
1.05040562e+00 -2.76282638e-01 -2.54820585e-01 -1.33544635e-02
3.28570902e-02 5.15523374e-01 1.45163953e+00 -1.00762773e+00
-7.80276299e-01 -8.21897209e-01 9.00480598e-02 8.83241668e-02
-5.47302246e-01 1.07747326e-02 -3.89621526e-01 -6.11083925e-01
-4.98173356e-01 -1.05163109e+00 -2.31789589e-01 4.71539617e-01
-5.90657353e-01 4.08770852e-02 1.43784213e+00 -4.61380839e-01
7.32968867e-01 -2.49914241e+00 -1.80887327e-01 2.27033645e-01
6.87681973e-01 3.69282246e-01 -3.47927213e-02 -2.07661614e-01
3.29475135e-01 2.53350884e-01 -8.65647793e-02 -8.94589603e-01
-1.18457377e-01 1.03141040e-01 -2.17513904e-01 4.43818331e-01
1.72082379e-01 1.09036827e+00 -7.36269474e-01 -1.74305409e-01
-3.99696715e-02 4.28678185e-01 -7.29376733e-01 3.01542222e-01
-7.08958730e-02 6.19890809e-01 -7.21284509e-01 6.07600987e-01
1.05543029e+00 -5.33600897e-02 -1.21179901e-01 -1.55773163e-01
3.50852966e-01 -2.71484762e-01 -8.96677494e-01 1.63497043e+00
-6.46430492e-01 5.03754377e-01 4.26768303e-01 -7.94200301e-01
8.00439596e-01 3.83387357e-01 5.34987688e-01 -3.08573693e-01
3.35282385e-01 8.62443522e-02 -3.05522919e-01 -3.78231078e-01
2.72169828e-01 4.45958227e-02 -3.50433767e-01 5.67155182e-01
4.80985604e-02 1.62563592e-01 -7.12604284e-01 9.07756463e-02
1.38450110e+00 -2.14830860e-01 -2.02244252e-01 -8.23502168e-02
2.52659172e-01 -2.53818721e-01 9.27197278e-01 6.64502203e-01
-5.32047868e-01 5.37292659e-01 4.57609266e-01 -6.42840147e-01
-9.09714162e-01 -9.56979573e-01 2.38554001e-01 9.99518871e-01
8.29943866e-02 -3.49387676e-01 -7.87390590e-01 -1.23030603e+00
2.96582222e-01 3.63521218e-01 -7.20538378e-01 -4.65469867e-01
8.08598101e-02 -5.44650018e-01 8.36696446e-01 2.89592534e-01
7.46676803e-01 -7.86168873e-01 -9.70608234e-01 -2.56043464e-01
2.47553349e-01 -1.26914811e+00 -7.41615653e-01 2.23612413e-01
-3.69695067e-01 -1.09637392e+00 -4.60992038e-01 -4.07998860e-01
8.01937044e-01 2.47797698e-01 6.05771244e-01 -2.61543334e-01
-4.52234983e-01 2.99090952e-01 -1.79557368e-01 -5.98415017e-01
-1.22267261e-01 -2.79867323e-03 8.82985219e-02 6.97587788e-01
5.01904249e-01 -6.16677165e-01 -6.83866918e-01 9.95956957e-02
-1.07515454e+00 7.29250163e-03 5.30040681e-01 9.02882040e-01
5.32233059e-01 9.19996575e-02 4.98296291e-01 -1.11784208e+00
4.83475149e-01 -6.91108227e-01 -6.28660321e-01 1.84769064e-01
-5.28357148e-01 -1.01460189e-01 1.11874473e+00 -3.94090742e-01
-1.17197347e+00 3.89645159e-01 2.04077348e-01 -1.06441629e+00
-9.19677913e-02 -1.10254757e-01 -7.37811387e-01 -3.21386606e-01
4.70726043e-01 2.94100970e-01 2.18625292e-01 -7.26697892e-02
3.59289229e-01 7.50365376e-01 4.03273880e-01 -3.70347142e-01
8.09914529e-01 7.80628681e-01 2.13249549e-01 -2.53459901e-01
-6.06617749e-01 -2.05653131e-01 -6.31803647e-02 -6.07007481e-02
1.00123501e+00 -1.18887889e+00 -1.05573881e+00 7.71421611e-01
-1.09432137e+00 -4.04537795e-03 -4.11429554e-01 2.63698369e-01
-5.19063294e-01 2.10518688e-01 -3.28129172e-01 -6.66529059e-01
-7.79434144e-01 -1.35299194e+00 1.21033072e+00 4.30313170e-01
1.27112374e-01 -6.95363879e-01 -4.45767432e-01 5.72696269e-01
4.35824364e-01 3.11426580e-01 4.22679782e-01 -8.81056607e-01
-8.25202823e-01 -3.07073146e-01 -2.18458697e-01 5.65324664e-01
1.84645891e-01 -3.34597707e-01 -1.36133456e+00 -4.51468885e-01
3.42948735e-01 -4.13109034e-01 7.02949226e-01 2.03297853e-01
1.65488863e+00 -6.06087387e-01 -2.27869749e-01 1.39402986e+00
1.32700443e+00 4.59249765e-02 4.76234078e-01 5.39260693e-02
9.43173289e-01 4.83246088e-01 3.53257209e-01 7.63549984e-01
2.75434673e-01 3.29230398e-01 6.35907114e-01 -1.67604029e-01
2.80464321e-01 -2.72783607e-01 3.95566821e-01 -5.33149876e-02
2.94168293e-01 -1.70046464e-01 -4.17809844e-01 4.52893019e-01
-1.98763049e+00 -8.19248021e-01 6.40662089e-02 2.14445519e+00
6.23459160e-01 -2.71990687e-01 -2.04768270e-01 -3.69586468e-01
5.86805344e-01 9.60735306e-02 -1.08945477e+00 -6.39367402e-01
-1.99716594e-02 1.66786194e-01 8.10642958e-01 -1.88962728e-01
-1.20386553e+00 9.13476169e-01 4.81618738e+00 8.26753020e-01
-1.29174924e+00 4.53368962e-01 1.15084171e+00 -3.36152613e-01
-4.98532146e-01 -1.58227503e-01 -5.90196311e-01 5.97834051e-01
6.58034205e-01 -4.47685629e-01 4.32888925e-01 1.27510345e+00
-7.13093653e-02 2.69539773e-01 -1.13371146e+00 1.28809655e+00
-1.83615848e-01 -1.26025844e+00 5.34546748e-02 1.76887944e-01
8.87665629e-01 4.91426177e-02 5.21128714e-01 7.78850392e-02
4.26896632e-01 -9.81654763e-01 5.22941947e-01 2.80118912e-01
1.03462362e+00 -9.94515181e-01 6.74044073e-01 3.49624574e-01
-8.75176609e-01 -1.63578585e-01 -3.73687178e-01 2.99941003e-02
-1.20197192e-01 8.82171929e-01 -4.78084624e-01 7.61267543e-01
9.53039110e-01 2.92933792e-01 -1.90930814e-01 5.51384211e-01
-9.28462446e-02 1.95173785e-01 -3.90869081e-01 2.62498558e-01
9.23685282e-02 -1.67160690e-01 6.45256400e-01 8.05864871e-01
3.96556914e-01 -1.52790233e-01 1.49313629e-01 1.08133948e+00
-5.13775468e-01 3.82143520e-02 -1.01861930e+00 5.81858046e-02
4.55067724e-01 1.68106771e+00 -2.59648621e-01 -2.89009921e-02
-5.45569658e-01 1.55260837e+00 3.03780317e-01 2.82265514e-01
-8.61917078e-01 -2.53148079e-01 1.47351432e+00 -2.58306473e-01
2.07917646e-01 -5.29581308e-03 -6.25747621e-01 -1.17750967e+00
2.63994664e-01 -8.06961477e-01 3.52630019e-01 -2.26975814e-01
-1.42233241e+00 5.29194176e-01 -5.08707464e-01 -1.07967198e+00
-1.38052285e-01 -3.47018927e-01 -8.19148958e-01 1.01117885e+00
-1.38754249e+00 -1.56114888e+00 -3.17811340e-01 1.13859451e+00
1.26670286e-01 -5.66681445e-01 7.66887844e-01 3.66018921e-01
-6.57269835e-01 1.35590434e+00 5.55418879e-02 2.80644387e-01
7.57310510e-01 -8.37135851e-01 4.05453205e-01 1.00737751e+00
1.43947572e-01 3.20105284e-01 1.69745967e-01 -5.78560293e-01
-1.74683082e+00 -1.65462303e+00 4.49538857e-01 -3.06159198e-01
3.79756004e-01 -6.73203468e-01 -5.57500303e-01 7.34930098e-01
-2.36921921e-01 7.28559971e-01 1.04320979e+00 -2.74217188e-01
-5.45304894e-01 -2.43349925e-01 -1.68522835e+00 5.01403570e-01
1.16762435e+00 -8.52742851e-01 3.07943553e-01 8.27215388e-02
8.66898954e-01 -2.61447579e-01 -6.99444592e-01 1.79023191e-01
5.22454143e-01 -1.08418405e+00 5.32003224e-01 -8.25568736e-01
3.81885022e-01 -2.72817582e-01 -1.15561023e-01 -1.11550295e+00
5.52307069e-02 -1.23170459e+00 -2.15314314e-01 1.40799034e+00
2.11122885e-01 -8.35924208e-01 1.21107602e+00 1.14054406e+00
-4.30362001e-02 -6.87051952e-01 -1.21155632e+00 -7.38505304e-01
-1.29238263e-01 -5.51378071e-01 9.72142756e-01 1.07531595e+00
-4.73488808e-01 -6.13410212e-02 -5.46618879e-01 4.17916536e-01
7.35128939e-01 6.28155693e-02 1.06756687e+00 -8.79122853e-01
-3.49291623e-01 -3.55113111e-02 -6.64135396e-01 -6.20013356e-01
4.25145656e-01 -1.06286716e+00 -2.96084404e-01 -7.66349375e-01
1.72937438e-01 -6.34267747e-01 -2.99612969e-01 7.43075192e-01
-1.86601356e-01 1.97211921e-01 4.90716338e-01 1.43016785e-01
-4.16452467e-01 5.18276691e-01 1.16059041e+00 -1.72434062e-01
-5.43328822e-02 3.53895664e-01 -1.04015255e+00 5.07392526e-01
7.91268468e-01 -6.31880939e-01 -6.57170892e-01 -6.33625805e-01
1.51570365e-01 -2.03317702e-01 7.07649767e-01 -1.00543058e+00
5.45067370e-01 2.05461849e-02 4.19547141e-01 8.73180013e-03
1.72459677e-01 -1.29163802e+00 3.60249788e-01 4.02630061e-01
-1.07113287e-01 -2.21833974e-01 1.50772318e-01 7.30815947e-01
-5.48660718e-02 2.01930866e-01 4.86282110e-01 3.85039151e-02
-6.56130552e-01 1.02027607e+00 -6.21163100e-02 -3.40192430e-02
1.63753426e+00 -2.27458358e-01 -2.13365972e-01 -3.24389696e-01
-5.70611179e-01 3.43231440e-01 6.23189151e-01 4.26873863e-01
6.00935519e-01 -1.14499664e+00 -6.91533506e-01 6.93874121e-01
1.36823893e-01 3.21266949e-01 4.30909127e-01 4.87481147e-01
-8.14397112e-02 -9.15486291e-02 -3.64707232e-01 -3.18505347e-01
-1.28958285e+00 6.46017313e-01 2.61425763e-01 -3.08713138e-01
-4.74037796e-01 1.21808183e+00 6.98301256e-01 -2.78972208e-01
9.21742842e-02 2.11492889e-02 4.45721835e-01 -1.45325989e-01
6.00612462e-01 1.87834129e-01 1.00088462e-01 -2.92684585e-01
-2.81703949e-01 6.66296333e-02 -2.75243908e-01 2.33597547e-01
1.09465981e+00 -1.60322562e-01 2.57587293e-03 -2.02468216e-01
1.56440413e+00 6.02084734e-02 -1.78806329e+00 -1.21389531e-01
-5.62463522e-01 -6.72570765e-01 2.50785530e-01 -6.12668216e-01
-1.75179613e+00 6.80332839e-01 7.56872296e-01 -5.43088838e-02
1.54638243e+00 -1.74720705e-01 1.10046613e+00 9.22699645e-02
6.67378128e-01 -9.64064121e-01 -1.76568091e-01 5.47469221e-02
3.56115490e-01 -1.38927364e+00 -3.56262773e-01 -5.83445787e-01
-8.68149042e-01 8.18279803e-01 7.93280780e-01 7.55361542e-02
9.06563699e-01 7.02663720e-01 1.13720581e-01 -1.11972302e-01
-6.35299742e-01 2.15938091e-01 5.89740239e-02 8.78959954e-01
-3.43725383e-01 2.25399956e-01 -8.68675299e-03 1.15329885e+00
-1.42435487e-02 1.51959226e-01 1.63512856e-01 7.25531280e-01
3.30072433e-01 -1.06030214e+00 -7.66448602e-02 6.41262591e-01
-8.83670866e-01 -4.35040332e-03 -3.97735119e-01 2.69018561e-01
5.05597770e-01 8.14513803e-01 2.60065407e-01 -7.86126435e-01
-1.23642512e-01 -7.84037262e-02 -7.68144103e-03 -4.68002498e-01
-1.01821256e+00 -5.69192350e-01 -1.89922646e-01 -8.92591476e-01
2.57614776e-02 -5.93513846e-01 -1.08055842e+00 -5.09564161e-01
-1.52868763e-01 -2.27934837e-01 7.85695374e-01 8.01981807e-01
1.10402811e+00 4.06510562e-01 1.04557061e+00 -5.48240006e-01
-6.44211888e-01 -2.44010627e-01 -7.25395858e-01 6.02360785e-01
2.24316746e-01 -2.41819665e-01 -3.87616843e-01 -4.89379279e-02] | [5.868804931640625, 6.754838943481445] |
2d540731-1484-463f-93ed-05ba7c7f5042 | a-lightweight-nms-free-framework-for-real | 2205.12458 | null | https://arxiv.org/abs/2205.12458v1 | https://arxiv.org/pdf/2205.12458v1.pdf | A Lightweight NMS-free Framework for Real-time Visual Fault Detection System of Freight Trains | Real-time vision-based system of fault detection (RVBS-FD) for freight trains is an essential part of ensuring railway transportation safety. Most existing vision-based methods still have high computational costs based on convolutional neural networks. The computational cost is mainly reflected in the backbone, neck, and post-processing, i.e., non-maximum suppression (NMS). In this paper, we propose a lightweight NMS-free framework to achieve real-time detection and high accuracy simultaneously. First, we use a lightweight backbone for feature extraction and design a fault detection pyramid to process features. This fault detection pyramid includes three novel individual modules using attention mechanism, bottleneck, and dilated convolution for feature enhancement and computation reduction. Instead of using NMS, we calculate different loss functions, including classification and location costs in the detection head, to further reduce computation. Experimental results show that our framework achieves over 83 frames per second speed with a smaller model size and higher accuracy than the state-of-the-art detectors. Meanwhile, the hardware resource requirements of our method are low during the training and testing process. | ['Yang Zhang', 'Ye Hu', 'Bo Wu', 'Huilin Pan', 'Yang Zhou', 'Guodong Sun'] | 2022-05-25 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [-9.83583629e-02 -5.48348844e-01 3.34225416e-01 -1.45303771e-01
-5.80836892e-01 1.14273980e-01 1.22202158e-01 -3.11335754e-02
-6.44982100e-01 1.54955238e-01 -3.82608086e-01 -2.81896532e-01
1.73296556e-01 -7.73394048e-01 -4.52586800e-01 -6.50094092e-01
3.10939103e-01 -7.31651112e-03 1.16891444e+00 -1.48126587e-01
5.17320096e-01 6.64690673e-01 -1.58515799e+00 2.30318815e-01
8.11247408e-01 1.39279807e+00 6.52720213e-01 6.01800442e-01
7.15203062e-02 7.99443126e-01 -7.63105571e-01 -1.39257386e-01
9.02690738e-02 4.00952473e-02 -4.45796102e-01 3.74462664e-01
2.45997116e-01 -6.77175045e-01 -1.01555490e+00 9.69176054e-01
8.12806785e-01 -2.48425245e-01 4.09766108e-01 -1.44014144e+00
-4.14581239e-01 8.31214339e-02 -8.73299599e-01 5.26251078e-01
-1.08860031e-01 3.03369910e-01 3.84965152e-01 -1.15770030e+00
-2.02687737e-02 9.90066886e-01 9.50689018e-01 3.38594049e-01
-5.15841186e-01 -6.75946176e-01 8.33504274e-02 7.70153642e-01
-1.50245893e+00 -4.42705870e-01 5.98312795e-01 -1.57015324e-01
9.53981400e-01 2.27089092e-01 5.46458542e-01 3.80899370e-01
4.06320870e-01 9.00373638e-01 5.05569816e-01 -5.23708880e-01
1.02481671e-01 -2.81531334e-01 2.47049510e-01 1.20212340e+00
4.48623806e-01 -2.53105253e-01 -3.94252747e-01 2.81185567e-01
1.18583179e+00 5.01141429e-01 -4.18492168e-01 2.11735368e-01
-1.02368963e+00 5.90828359e-01 4.00735110e-01 7.46624172e-02
-3.77475977e-01 3.15163881e-01 6.44714534e-01 1.46382660e-01
2.30730101e-01 -3.38017702e-01 -4.35792953e-01 -2.03987006e-02
-8.24109077e-01 3.94508280e-02 3.13346863e-01 1.02052152e+00
6.21540368e-01 2.54410237e-01 -4.53275293e-01 7.97784209e-01
2.54485816e-01 4.77686971e-01 5.11183739e-01 -6.17326736e-01
4.95256543e-01 7.32888639e-01 -2.39148680e-02 -7.89010108e-01
-7.12040603e-01 -5.03307223e-01 -8.71674061e-01 3.18389624e-01
4.60966397e-03 -1.36664942e-01 -1.20433939e+00 8.53326023e-01
1.73222497e-01 3.39682728e-01 -3.32561165e-01 1.00633609e+00
7.46402800e-01 3.98882121e-01 -2.51040220e-01 -8.81166533e-02
1.64823365e+00 -1.32251763e+00 -7.41741776e-01 -1.70399338e-01
7.32567430e-01 -9.62834835e-01 8.10854197e-01 4.21261668e-01
-1.10427368e+00 -7.70573139e-01 -1.22391999e+00 -1.28264472e-01
2.62335241e-02 6.34650826e-01 5.89667916e-01 5.66046357e-01
-8.51709843e-01 4.92137671e-01 -9.71439540e-01 -1.13730326e-01
5.23200154e-01 3.67895693e-01 -1.25138998e-01 -2.15787932e-01
-9.65207875e-01 9.19407189e-01 8.06872696e-02 5.02595186e-01
-8.83382499e-01 -4.23563033e-01 -7.40369022e-01 2.43876502e-01
3.34997594e-01 -4.51231211e-01 1.36518466e+00 -4.82737660e-01
-1.17441070e+00 3.39385390e-01 -5.09966873e-02 -3.63177568e-01
4.93386149e-01 -3.21263671e-01 -4.17435110e-01 3.12041163e-01
8.60651135e-02 3.08448762e-01 8.68362010e-01 -8.46628010e-01
-1.29843330e+00 -7.10419491e-02 -1.76336765e-02 1.39593899e-01
-5.38514376e-01 2.31206685e-01 -9.40241694e-01 -5.06889999e-01
5.17056227e-01 -5.93414009e-01 -2.34983414e-01 2.88221896e-01
-2.40578339e-01 -3.97777587e-01 1.28257656e+00 -6.74757302e-01
1.26769090e+00 -2.18886781e+00 -7.43173659e-01 -1.07527018e-01
1.87077999e-01 5.82423687e-01 1.65993214e-01 1.48629561e-01
1.16794296e-01 -5.06898582e-01 1.59720302e-01 -3.08417529e-01
-2.94563711e-01 9.44267437e-02 3.26232128e-02 7.14697778e-01
4.83729631e-01 6.05780125e-01 -5.39049804e-01 -8.00954342e-01
6.14245236e-01 3.50002050e-01 -5.51479340e-01 -2.27717664e-02
4.69106048e-01 -1.79942455e-02 -5.30290604e-01 8.74414742e-01
1.05223560e+00 -3.68852280e-02 -3.60046685e-01 -6.69658720e-01
-4.54027295e-01 -2.42540110e-02 -1.30974746e+00 1.65389848e+00
-3.35742176e-01 5.33643186e-01 1.64118201e-01 -1.11505151e+00
8.82814825e-01 1.70961127e-01 3.40924144e-01 -7.80135036e-01
3.54033947e-01 1.15855545e-01 -1.65553644e-01 -9.17495787e-01
5.06723583e-01 1.19362630e-01 1.11317188e-01 5.16570620e-02
4.41887565e-02 3.16210181e-01 2.59266466e-01 -1.06940247e-01
1.49470210e+00 -2.09977310e-02 -2.85099953e-01 -5.01609466e-04
5.83879590e-01 -4.76455763e-02 8.94496977e-01 6.41859770e-01
-4.45405781e-01 5.59079230e-01 1.58679605e-01 -5.67814946e-01
-7.94132113e-01 -8.44529688e-01 -8.02930966e-02 7.13350415e-01
4.75607008e-01 -7.62586817e-02 -7.75911987e-01 -6.08373582e-01
4.93889935e-02 1.28354475e-01 -1.89467445e-01 -3.29553396e-01
-8.38451147e-01 -7.46697009e-01 7.00943470e-01 9.35192287e-01
9.42868948e-01 -8.82125139e-01 -8.29721212e-01 3.32534879e-01
-1.55729115e-01 -1.19898522e+00 -5.73213518e-01 6.63060546e-02
-8.67780387e-01 -1.03748322e+00 -7.28545785e-01 -1.34583735e+00
8.95196438e-01 8.39128971e-01 6.59436405e-01 6.12380505e-01
-8.66742671e-01 -1.28440008e-01 -3.55121851e-01 -4.69504774e-01
2.92436361e-01 -3.44172388e-01 -9.62873548e-02 -1.26281470e-01
2.33427331e-01 -2.04273358e-01 -1.00556171e+00 3.49107265e-01
-6.70805156e-01 7.96066895e-02 1.08013451e+00 9.45321560e-01
2.71402657e-01 3.99596006e-01 3.99110734e-01 -3.86373490e-01
3.56859684e-01 -4.57961336e-02 -6.09409392e-01 1.22217052e-01
-5.18828690e-01 -3.60789329e-01 4.72053468e-01 -1.48171857e-01
-9.91247177e-01 6.75979406e-02 -3.11218351e-01 -5.90284228e-01
-7.17319101e-02 1.60126641e-01 1.86369270e-01 -4.23031181e-01
2.53577203e-01 4.42259163e-01 -7.23977163e-02 -4.64382082e-01
-1.31691813e-01 1.08029878e+00 5.98469138e-01 -6.12650514e-02
6.58738852e-01 5.82857668e-01 1.25775501e-01 -8.73530567e-01
-5.17583430e-01 -5.79198539e-01 -4.66144025e-01 -4.03612584e-01
6.11217678e-01 -1.21758008e+00 -1.13609958e+00 8.74283671e-01
-1.33011723e+00 2.57025957e-01 8.15125331e-02 6.55432343e-01
-2.67488182e-01 6.28516436e-01 -1.06426466e+00 -8.99387300e-01
-7.63104200e-01 -1.15892255e+00 1.17497754e+00 4.48438197e-01
6.50294662e-01 -2.76173592e-01 -5.92717826e-01 3.54352981e-01
4.25560206e-01 -9.18002576e-02 5.33892393e-01 -1.31598964e-01
-6.74810767e-01 -4.86553222e-01 -7.21967518e-01 4.78106230e-01
8.57128203e-02 2.66968049e-02 -8.29600453e-01 -3.72234255e-01
3.12797166e-02 6.01445064e-02 1.20451522e+00 4.59347546e-01
1.33584464e+00 1.72527805e-01 -5.05335033e-01 4.19807851e-01
1.54397845e+00 3.56591970e-01 8.07617307e-01 3.81444752e-01
7.43060112e-01 2.40858316e-01 9.25581396e-01 6.70282125e-01
4.64684397e-01 4.58971739e-01 5.35633206e-01 -4.85539854e-01
-1.89666748e-01 2.71985918e-01 3.22520375e-01 1.03704286e+00
-7.77834505e-02 -8.72650668e-02 -8.61069441e-01 5.54402053e-01
-2.02514195e+00 -8.22358847e-01 -5.10567427e-01 1.97970796e+00
3.99375439e-01 6.30749166e-01 -8.84737372e-02 6.27123177e-01
1.11340439e+00 -4.13110286e-01 -3.56794298e-01 -1.30889155e-02
2.62140363e-01 1.95738360e-01 7.49715507e-01 4.23181355e-02
-1.30391419e+00 6.65629566e-01 5.66656065e+00 1.04213977e+00
-1.04355395e+00 2.26755857e-01 2.93090612e-01 5.55418618e-02
5.06717205e-01 -1.79567114e-01 -8.14809084e-01 4.78705168e-01
6.10356271e-01 3.83982152e-01 -9.32338461e-02 8.89822602e-01
3.99945080e-01 -1.89786196e-01 -5.58144510e-01 1.06099951e+00
1.30851135e-01 -1.42610776e+00 -2.23529175e-01 -3.70785683e-01
1.72766596e-01 -5.56507371e-02 -3.46530348e-01 2.68640697e-01
-3.15261692e-01 -3.70812207e-01 9.54636157e-01 4.79224294e-01
6.61580265e-01 -1.15421832e+00 1.00334787e+00 1.96442664e-01
-1.42945635e+00 -3.83295715e-01 -6.25292063e-01 -1.51650086e-01
2.25209460e-01 8.08739185e-01 -4.62618798e-01 7.01496482e-01
8.70522618e-01 3.68414640e-01 -4.10110414e-01 1.32178676e+00
-2.81354133e-02 2.88510054e-01 -8.14175084e-02 -4.99794027e-03
1.89128965e-01 2.88388133e-01 5.87506518e-02 1.30459762e+00
4.11569566e-01 -1.68189004e-01 5.65906584e-01 2.97714442e-01
2.32992664e-01 -4.03504223e-01 -1.93558913e-02 5.79368174e-01
5.81286132e-01 1.53931236e+00 -8.17512512e-01 -2.11298168e-01
-9.46602345e-01 1.16733372e+00 1.22720107e-01 -5.40307239e-02
-1.35699141e+00 -1.13082039e+00 5.20503402e-01 1.62034869e-01
6.05423510e-01 -2.98127919e-01 -1.57073095e-01 -8.34540427e-01
1.76709846e-01 -3.57773095e-01 9.82702672e-02 -6.70043528e-01
-9.76102054e-01 4.58799183e-01 -5.98731816e-01 -1.50600684e+00
6.60088301e-01 -6.98771000e-01 -8.77298295e-01 7.85035551e-01
-1.82943046e+00 -1.35223663e+00 -7.03928709e-01 7.36706614e-01
8.74082386e-01 5.85459396e-02 3.24878216e-01 8.08304131e-01
-1.13860035e+00 7.62690425e-01 -1.27821073e-01 3.83031160e-01
6.65878654e-01 -8.26052487e-01 5.91635525e-01 1.11565220e+00
-4.94112432e-01 2.93169677e-01 3.49287689e-01 -5.76812744e-01
-1.59400046e+00 -1.30633140e+00 5.86031437e-01 3.19037199e-01
4.08973277e-01 -6.96019232e-02 -6.73391283e-01 3.45912755e-01
-9.61919129e-02 3.76605064e-01 3.17389555e-02 -3.84127617e-01
1.53358057e-01 -3.14835668e-01 -1.20059133e+00 4.50438350e-01
9.14667666e-01 -1.79391503e-01 -3.28669459e-01 4.25672889e-01
7.18844950e-01 -5.00520289e-01 -5.62011659e-01 4.83846128e-01
4.25121903e-01 -8.74564648e-01 8.70667458e-01 -2.48607993e-02
4.86497059e-02 -8.37717474e-01 2.04642102e-01 -7.12805212e-01
-8.43773425e-01 -3.65844399e-01 -1.11321539e-01 1.21999443e+00
1.14942320e-01 -5.93154073e-01 8.27436268e-01 2.23785326e-01
-9.03865993e-01 -9.52650785e-01 -1.02073312e+00 -6.88330591e-01
-6.95840597e-01 -3.60210449e-01 2.79705942e-01 4.97958869e-01
-1.87719494e-01 2.31033042e-01 -2.40379184e-01 6.29942298e-01
5.98494768e-01 -7.11566135e-02 4.69246477e-01 -1.16585124e+00
4.72043231e-02 -1.13911644e-01 -9.03011978e-01 -1.03425205e+00
-5.48029184e-01 -3.22468370e-01 4.27317023e-01 -1.67661464e+00
1.99625805e-01 -3.53779525e-01 -4.96017456e-01 6.34015620e-01
-2.18436792e-01 3.18099499e-01 -7.39483982e-02 1.86201379e-01
-7.54661918e-01 4.46840912e-01 1.28014302e+00 1.11167453e-01
2.79966086e-01 -2.10917115e-01 -3.46402466e-01 9.63507533e-01
6.99308753e-01 -4.29705650e-01 -4.78870682e-02 -6.93272710e-01
-4.75716323e-01 -8.33710358e-02 4.37032104e-01 -1.54122388e+00
7.32089043e-01 8.83383006e-02 7.08378315e-01 -9.39053237e-01
2.73948878e-01 -7.70812154e-01 -5.15878916e-01 1.14768946e+00
4.37012941e-01 4.17682141e-01 1.61867246e-01 5.76594174e-01
-2.40606651e-01 -2.78687835e-01 8.40608120e-01 -5.23267612e-02
-1.09607136e+00 2.30568945e-01 -5.99379122e-01 -3.79740149e-01
1.25915337e+00 -3.61970872e-01 -6.90798581e-01 1.99471250e-01
-2.50371546e-01 4.16709960e-01 9.43865404e-02 3.39325279e-01
1.02034068e+00 -1.26680684e+00 -6.38573050e-01 3.96513283e-01
4.79374677e-02 9.72461700e-02 7.73809910e-01 1.06724107e+00
-1.08795917e+00 4.54555750e-01 -3.45795870e-01 -5.67328095e-01
-1.29570317e+00 3.96170050e-01 8.83097798e-02 -1.07313052e-01
-8.93123507e-01 1.09042192e+00 1.04092300e-01 1.60498902e-01
3.21152955e-01 -3.72662097e-01 -9.52257887e-02 -2.71012455e-01
6.93163574e-01 7.61021316e-01 4.64224219e-01 -3.31438601e-01
-6.00764334e-01 5.85721374e-01 -3.66456926e-01 3.08654219e-01
1.11924875e+00 -1.51753590e-01 1.25218973e-01 -2.11856678e-01
7.25202322e-01 -5.18741667e-01 -1.41799307e+00 -8.78342986e-02
-1.18406884e-01 -5.13826311e-01 6.33141875e-01 -3.68908584e-01
-1.43451250e+00 7.43129611e-01 9.39649463e-01 1.11660339e-01
1.46080387e+00 -3.81636798e-01 1.22630656e+00 1.95749074e-01
3.29856336e-01 -1.37192392e+00 1.62630960e-01 4.97713447e-01
5.88720381e-01 -1.00695145e+00 9.62874666e-03 -6.97897732e-01
-4.38034266e-01 1.26868725e+00 1.01744187e+00 -4.12182778e-01
5.86205661e-01 5.90202689e-01 -8.63253996e-02 -1.62568301e-01
-7.14131713e-01 -4.13803309e-01 -8.26974511e-02 4.08911377e-01
2.27759197e-01 -2.44911909e-01 -4.06642824e-01 6.73100591e-01
5.65910161e-01 1.12860180e-01 5.31585991e-01 1.56795800e+00
-1.08666587e+00 -6.86056852e-01 -3.95923316e-01 6.29938543e-01
-5.16703129e-01 -1.04214542e-01 4.41178590e-01 5.69002807e-01
4.84818667e-01 1.24209857e+00 2.72977889e-01 -8.57600808e-01
7.32774377e-01 -3.74275953e-01 5.39742112e-01 -4.19984639e-01
-5.05269587e-01 1.39055122e-02 -1.54661182e-02 -5.82117140e-01
-8.59945044e-02 -3.76001596e-01 -1.53709400e+00 -3.87664855e-01
-1.04597282e+00 -1.43989399e-01 7.17472196e-01 9.60115433e-01
4.05167103e-01 1.05136049e+00 8.68794322e-01 -8.96861732e-01
-4.22281861e-01 -1.01395774e+00 -6.86795771e-01 -6.41516596e-02
3.44539672e-01 -8.14370155e-01 -8.88254866e-02 -7.00657964e-02] | [8.730669975280762, -0.4907466471195221] |
316fae90-d33c-4388-8cae-30b3f996ec5e | guided-mcmc-for-sparse-bayesian-models-to | null | null | https://openreview.net/forum?id=Yc64t25hseP | https://openreview.net/pdf?id=Yc64t25hseP | GUIDED MCMC FOR SPARSE BAYESIAN MODELS TO DETECT RARE EVENTS IN IMAGES SANS LABELED DATA | Detection of rare events in images is a challenging task because of two main problems, the first problem is the lack of labeled data for rare category class and the second problem is a highly imbalanced data problem. Training models in this scenario becomes hard. Unsupervised methods do not apply as we need to detect rare events automatically. Rule-based methods seem to be the only viable solution, but it is tedious to come up with a set of rules covering all corner cases. Even the recently popular zero-shot learning techniques required to be pre-trained on auxiliary datasets. In the given scenario, we propose an approach to provide little guidance from experts as an input into a hierarchical Bayesian model. The guidance influences the Markov chain Monte Carlo (MCMC) based inference technique of the model. After the steady-state is obtained for the underlying Markov chain, it is possible to compute the posterior probability of the presence of the rare event in a given image. The proposed method neither needs any labeled data nor required pre-training, unlike zero-shot learning. The proposed technique has been observed to outperform the state-of-the-art unsupervised image classification techniques. | ['Mrinal Das', 'Gaurav Jain'] | 2021-09-29 | null | null | null | null | ['unsupervised-image-classification'] | ['computer-vision'] | [ 5.01059890e-01 8.04529041e-02 -2.45169476e-02 -3.78558487e-01
-6.69906557e-01 -2.27753185e-02 6.15696013e-01 6.43440068e-01
-6.61195874e-01 8.47118318e-01 -2.96303719e-01 -2.24861473e-01
-4.38076854e-01 -8.57145369e-01 -4.17109102e-01 -9.95326579e-01
1.31866589e-01 9.32953835e-01 5.74272275e-01 2.27872923e-01
2.69388556e-01 3.23954046e-01 -2.12024570e+00 2.48771399e-01
9.19018745e-01 8.71033847e-01 4.05433089e-01 6.68269455e-01
-3.52365047e-01 9.31202412e-01 -4.29896176e-01 -1.42590508e-01
1.18241675e-01 -6.35520637e-01 -5.10801375e-01 4.53205138e-01
-4.62078080e-02 -1.62352487e-01 3.29557300e-01 1.07171845e+00
5.39757669e-01 3.09130967e-01 1.12680328e+00 -1.08045709e+00
4.03795719e-01 1.38493970e-01 -5.53959370e-01 3.04435045e-01
1.78748176e-01 1.76359802e-01 5.84201455e-01 -7.55457938e-01
7.31334209e-01 8.85724545e-01 3.13876122e-01 2.40898535e-01
-1.07402778e+00 -3.27052504e-01 -9.00066718e-02 5.66717565e-01
-1.33474982e+00 -3.68371755e-01 4.50494707e-01 -7.35355139e-01
7.31834114e-01 4.31507565e-02 5.43548822e-01 9.36244309e-01
3.25180255e-02 3.56435001e-01 1.34361136e+00 -8.16233993e-01
8.56873631e-01 3.30372870e-01 1.61855385e-01 5.17566979e-01
4.99162287e-01 5.90055024e-05 -3.79892051e-01 -2.77197629e-01
2.94933110e-01 2.31719509e-01 1.43204689e-01 -4.97699767e-01
-6.50943398e-01 8.33298266e-01 -1.79693297e-01 3.28803897e-01
-7.58452713e-01 -1.99336722e-01 4.83450085e-01 7.00735226e-02
4.52962428e-01 -4.23300732e-03 -1.51921079e-01 -1.57660708e-01
-1.37648892e+00 1.79085508e-01 8.83192480e-01 3.73884469e-01
7.57578015e-01 -1.21740997e-01 -2.83600371e-02 6.34354472e-01
2.45500535e-01 1.51132986e-01 3.30728561e-01 -4.95455593e-01
5.88929020e-02 5.54030597e-01 2.68495739e-01 -7.72404134e-01
-1.07807390e-01 -4.24064279e-01 -7.76505351e-01 5.67787826e-01
8.60804200e-01 -1.40650392e-01 -1.18469834e+00 1.23401654e+00
5.81685662e-01 1.47374600e-01 -1.63778067e-01 5.27959883e-01
3.65943640e-01 7.74747550e-01 2.38613784e-01 -8.81125271e-01
1.21950233e+00 -3.97774696e-01 -8.59524190e-01 -2.32838634e-02
3.36929262e-01 -8.65525782e-01 5.60275197e-01 5.92991173e-01
-7.27582455e-01 -3.42811286e-01 -8.72303247e-01 6.56596661e-01
-4.47690606e-01 2.89971735e-02 3.75451267e-01 7.13797033e-01
-5.14199376e-01 6.04001582e-01 -8.42859924e-01 -7.69399881e-01
3.72977078e-01 2.38750100e-01 -1.54474229e-01 -2.92648435e-01
-8.58209372e-01 1.10177481e+00 7.52845109e-01 4.55792323e-02
-1.09537911e+00 -1.53692275e-01 -6.21668816e-01 4.81065139e-02
7.28251994e-01 -4.42747384e-01 1.05615449e+00 -9.33154464e-01
-1.38458955e+00 8.45593214e-01 -2.40694270e-01 -4.50865656e-01
8.69731426e-01 9.91530940e-02 -5.84108979e-02 3.36347759e-01
-7.27707669e-02 2.68109113e-01 1.11221564e+00 -1.26921403e+00
-6.18032098e-01 -1.18672773e-01 -3.06308031e-01 5.70433401e-02
-9.75957664e-04 1.14438690e-01 -1.23436362e-01 -3.54986072e-01
-1.27926124e-02 -6.21413350e-01 -4.01030570e-01 -1.71918482e-01
-2.07652643e-01 -2.70592839e-01 5.52173734e-01 -3.48651171e-01
9.90135968e-01 -1.93761241e+00 -2.95993179e-01 2.94504583e-01
-1.59684241e-01 3.19951117e-01 5.21676481e-01 4.73319083e-01
-2.09261626e-01 -2.90447742e-01 -3.42561096e-01 -1.15035631e-01
-1.80748641e-01 3.37577909e-01 -1.39478996e-01 5.86552858e-01
2.50634432e-01 2.13389192e-02 -8.94612491e-01 -9.64264572e-01
5.95243871e-01 2.99821973e-01 -4.12493348e-01 3.39337736e-01
-4.02709424e-01 4.27225411e-01 -2.72258878e-01 2.17691958e-01
6.75455987e-01 -3.61769021e-01 3.48167181e-01 8.88863355e-02
-1.55220926e-01 6.85778484e-02 -1.77971876e+00 1.16811061e+00
-1.65564671e-01 9.52892751e-02 -4.40466076e-01 -1.42257094e+00
8.21892440e-01 5.95781744e-01 4.36630249e-01 -3.85481745e-01
4.16695863e-01 3.73736769e-01 1.52471796e-01 -8.07419181e-01
4.07945849e-02 -6.36022091e-01 1.45286858e-01 5.11343956e-01
3.86981875e-01 2.71752402e-02 5.58365822e-01 1.37301117e-01
1.09882534e+00 2.77108848e-01 8.07814717e-01 -1.70140162e-01
4.67872083e-01 8.05358142e-02 7.20644653e-01 9.52448249e-01
-3.28895636e-02 5.67788959e-01 6.50884032e-01 -4.61738229e-01
-1.11264396e+00 -8.37918222e-01 -2.06178680e-01 6.69879317e-01
-1.63083643e-01 -2.10842222e-01 -7.69217908e-01 -5.55391908e-01
-4.79731888e-01 7.38459885e-01 -5.36144197e-01 1.06025092e-01
-7.91299343e-02 -1.16169286e+00 -2.61229873e-01 -1.57452703e-01
3.08905572e-01 -1.09517789e+00 -1.13580346e+00 3.99941146e-01
-1.12217709e-01 -8.22538018e-01 3.74744892e-01 6.15716040e-01
-9.94745374e-01 -1.07372355e+00 -6.77196801e-01 -3.25052291e-01
9.37761784e-01 -1.00321092e-01 9.25934374e-01 2.40818821e-02
-6.18419051e-01 5.73958978e-02 -4.14233059e-01 -5.40503323e-01
-5.38690090e-01 -1.73317850e-01 5.23911789e-02 2.61792541e-01
4.75267947e-01 -4.84470725e-01 -5.14337301e-01 3.61411810e-01
-8.95293951e-01 -1.85512856e-01 6.73674345e-01 6.97176695e-01
7.18533278e-01 4.92961854e-01 6.43138647e-01 -1.06927490e+00
1.16255932e-01 -4.08614725e-01 -6.57747865e-01 3.46404403e-01
-4.36678529e-01 1.00429982e-01 5.46702147e-01 -4.59074795e-01
-1.25126958e+00 4.84382391e-01 -1.56241670e-01 -1.90300286e-01
-7.67706513e-01 3.87765735e-01 -1.46057099e-01 2.53652960e-01
6.97305918e-01 9.70528200e-02 -2.48769652e-02 -4.87579226e-01
-2.79899444e-02 6.42888069e-01 3.53803098e-01 -3.20750117e-01
4.63272423e-01 6.10285103e-01 1.24201134e-01 -1.04104507e+00
-9.71233964e-01 -8.06147873e-01 -7.19206333e-01 -5.82856953e-01
8.89357150e-01 -6.36467755e-01 -4.91037488e-01 3.79327148e-01
-9.87403333e-01 -3.26603241e-02 -4.26020563e-01 5.48028111e-01
-7.21948683e-01 4.05533224e-01 -6.53173774e-02 -1.37353516e+00
5.94072305e-02 -9.68109131e-01 4.97357488e-01 2.69089460e-01
-3.01951051e-01 -6.20697796e-01 1.87175110e-01 2.63372302e-01
1.28679261e-01 3.35456133e-01 8.56863022e-01 -8.20322692e-01
-4.66845363e-01 -6.91509128e-01 8.11982229e-02 2.29046196e-01
1.67741925e-02 1.54118985e-01 -1.11017597e+00 5.26489653e-02
2.99130917e-01 -3.72777134e-01 8.06084812e-01 5.40390730e-01
9.91642058e-01 -3.63774225e-02 -2.17814744e-01 -1.01904981e-01
1.61009669e+00 2.23005101e-01 8.24629962e-01 1.38254687e-01
1.26649186e-01 8.98593485e-01 8.91578197e-01 7.21672952e-01
-5.41915447e-02 5.15007377e-01 3.35846275e-01 1.56957224e-01
1.25132844e-01 -6.33017123e-02 5.08310869e-02 5.77186465e-01
-1.66567385e-01 -3.39399308e-01 -1.05496716e+00 6.08708620e-01
-2.00379229e+00 -1.36125839e+00 -2.74315596e-01 2.62345052e+00
7.46442854e-01 6.41578376e-01 2.21899211e-01 6.61370993e-01
8.04995060e-01 -1.80399254e-01 -2.30780855e-01 -1.82837248e-01
3.85853320e-01 8.54047984e-02 3.42883646e-01 4.64539617e-01
-1.06466746e+00 4.27239269e-01 5.65844631e+00 1.09253812e+00
-7.52943099e-01 2.04439595e-01 6.37372792e-01 -8.82121734e-03
1.06447205e-01 3.37641060e-01 -8.02480936e-01 6.67594850e-01
9.53625441e-01 1.32979646e-01 -1.33395568e-01 6.67555153e-01
4.38726455e-01 -1.11722577e+00 -8.00755501e-01 1.07780480e+00
1.68436468e-01 -8.48362803e-01 -4.76179905e-02 1.08506948e-01
6.52796626e-01 -3.32462817e-01 -4.31643277e-01 9.70364362e-02
1.96773842e-01 -7.05184221e-01 4.33375448e-01 8.23992312e-01
5.00191271e-01 -9.20255721e-01 9.59436893e-01 7.73475826e-01
-6.73160553e-01 -4.24196869e-02 -4.84262317e-01 -2.64802366e-01
2.78336197e-01 1.18404531e+00 -8.84689271e-01 3.76390755e-01
4.91891176e-01 1.72261164e-01 -4.61033344e-01 1.52777970e+00
-1.68206185e-01 8.01741064e-01 -4.86072361e-01 -5.59421517e-02
1.22340083e-01 -3.05769499e-02 4.22766536e-01 9.73898768e-01
3.05779517e-01 3.60742547e-02 1.82656199e-01 3.56767267e-01
4.92472053e-01 1.81338951e-01 -6.68758631e-01 1.15245037e-01
-4.06457558e-02 1.33708048e+00 -1.50632179e+00 -6.30357146e-01
-2.32735842e-01 7.59641945e-01 1.44142240e-01 6.44503683e-02
-5.12161553e-01 -2.52792537e-01 -1.28883749e-01 2.97778577e-01
5.24003088e-01 2.64134612e-02 -4.81182411e-02 -1.00605178e+00
-2.33100757e-01 -8.74222517e-01 6.91460311e-01 -5.55242956e-01
-1.17553902e+00 3.14584434e-01 2.52244085e-01 -1.26190472e+00
-7.22079515e-01 -4.34855610e-01 -7.27984428e-01 7.53599644e-01
-1.27118397e+00 -4.77265418e-01 -2.58338183e-01 5.05901635e-01
6.68617666e-01 1.92044284e-02 8.06611180e-01 4.16134596e-01
-3.67956460e-01 -1.65308952e-01 7.98508301e-02 -3.33966911e-01
6.97278500e-01 -1.18845141e+00 -4.76495445e-01 1.01160550e+00
-9.29713622e-02 2.13088334e-01 1.33644748e+00 -9.22593117e-01
-7.68802226e-01 -6.51603639e-01 1.12368524e+00 -1.96178123e-01
4.21826899e-01 -2.71501541e-01 -1.03058791e+00 9.01964456e-02
5.98989725e-02 1.60954088e-01 7.70986259e-01 -6.98576421e-02
2.06046164e-01 -3.93110327e-02 -1.10343111e+00 1.42471597e-01
4.54927593e-01 -1.91781029e-01 -6.80177212e-01 3.72782677e-01
-1.03863902e-01 4.65141423e-02 -3.80841285e-01 1.93631023e-01
3.75683606e-01 -1.02862799e+00 5.90714216e-01 -4.39505368e-01
2.96343476e-01 -5.87117314e-01 7.55828097e-02 -9.48039174e-01
7.94531778e-02 -3.95754069e-01 1.88131183e-02 1.21812379e+00
3.35267901e-01 -2.88323998e-01 7.39854217e-01 3.91237170e-01
4.55109239e-01 -4.74460512e-01 -1.02961564e+00 -7.31939435e-01
-5.79062343e-01 -4.82350767e-01 -7.99985453e-02 7.44636059e-01
-1.79396376e-01 2.44224653e-01 -4.11391824e-01 1.03501573e-01
1.11330295e+00 7.78098106e-02 7.25800514e-01 -1.53441751e+00
-4.56707537e-01 7.71766901e-02 -5.48877060e-01 -2.71175534e-01
-2.44304866e-01 -4.77662712e-01 3.13700676e-01 -1.73387170e+00
5.28592646e-01 -2.93497950e-01 -1.81685835e-01 2.12313622e-01
-2.34615505e-01 1.69242844e-02 -4.66759577e-02 1.83287263e-01
-7.49609292e-01 3.01091552e-01 6.25852883e-01 1.35485858e-01
9.40806326e-03 2.83973515e-01 -1.34642825e-01 1.00977325e+00
6.82889044e-01 -9.52097178e-01 -5.38137138e-01 3.80441040e-01
5.48965633e-01 1.83725759e-01 4.19594824e-01 -1.05637443e+00
4.37828928e-01 -1.37390062e-01 3.40819597e-01 -8.66581023e-01
9.60149243e-02 -9.90985811e-01 3.60329419e-01 6.25154316e-01
-5.23399701e-03 -4.28496182e-01 -1.44656122e-01 8.04895401e-01
-1.87137425e-01 -9.28118169e-01 1.14577043e+00 -4.46906149e-01
-6.62741125e-01 1.78843457e-03 -7.93654144e-01 -1.66275963e-01
1.27306807e+00 -2.96620399e-01 1.27317369e-01 -3.97531867e-01
-1.06686032e+00 -1.51219303e-02 2.39080653e-01 -2.41444156e-01
3.53787869e-01 -8.24941635e-01 -4.60248470e-01 -1.01156332e-01
1.49479255e-01 -1.20250899e-02 5.38973629e-01 9.99926746e-01
-5.01232564e-01 7.19079822e-02 -2.64128983e-01 -7.42707789e-01
-1.14866662e+00 7.27911174e-01 1.04763694e-01 -4.83492911e-01
-5.57180703e-01 3.69284689e-01 -1.46266952e-01 -6.62132725e-02
2.86627561e-01 6.33384213e-02 -4.40740228e-01 6.02161348e-01
6.37927711e-01 4.61770236e-01 2.63882011e-01 -3.54376763e-01
-3.32634211e-01 3.70079130e-01 -4.86399308e-02 -2.87177742e-01
1.52063560e+00 -8.72171223e-02 4.37309891e-02 7.71837533e-01
6.50544643e-01 -3.87728840e-01 -1.07243371e+00 -4.57728468e-02
1.42740041e-01 -5.34853995e-01 9.64361727e-02 -6.87648237e-01
-4.36846882e-01 1.00912559e+00 7.95226932e-01 3.17381889e-01
1.10263216e+00 -1.16282903e-01 8.27239305e-02 2.97654897e-01
3.88341278e-01 -1.37317085e+00 -5.10242991e-02 1.40266433e-01
2.86007106e-01 -1.48809934e+00 3.02109897e-01 -4.05366182e-01
-4.64495569e-01 1.01679099e+00 2.45856658e-01 -7.59048685e-02
7.52544999e-01 2.11657628e-01 -1.77523285e-01 -2.71421105e-01
-7.96190262e-01 -5.25253177e-01 5.47013879e-02 4.40145314e-01
2.55905032e-01 -1.96918309e-01 -6.69692934e-01 8.04395750e-02
2.82022148e-01 4.74467844e-01 6.03455126e-01 1.13891208e+00
-7.03268349e-01 -1.07089949e+00 -5.55170834e-01 7.49527514e-01
-7.60566056e-01 1.93637893e-01 1.07762799e-01 5.76403618e-01
4.89713252e-01 9.15057898e-01 1.25517935e-01 2.35070080e-01
6.75182343e-02 4.52990443e-01 5.11670411e-01 -7.22536027e-01
-1.54287323e-01 2.90437341e-01 7.35175833e-02 -2.50461102e-01
-7.29228139e-01 -9.38498795e-01 -7.98525572e-01 6.78872466e-02
-6.31166816e-01 2.25131512e-01 7.43057072e-01 1.09338415e+00
4.49565146e-03 3.94164771e-01 4.41129386e-01 -8.52931440e-01
-4.57038939e-01 -7.91554213e-01 -7.14228034e-01 4.59206343e-01
4.04937081e-02 -9.10633564e-01 -5.25490940e-01 3.39756489e-01] | [8.453564643859863, 2.2599825859069824] |
b1d9d19f-a681-42f8-9ae1-2abf4e5aca80 | visible-thermal-uav-tracking-a-large-scale | 2204.04120 | null | https://arxiv.org/abs/2204.04120v1 | https://arxiv.org/pdf/2204.04120v1.pdf | Visible-Thermal UAV Tracking: A Large-Scale Benchmark and New Baseline | With the popularity of multi-modal sensors, visible-thermal (RGB-T) object tracking is to achieve robust performance and wider application scenarios with the guidance of objects' temperature information. However, the lack of paired training samples is the main bottleneck for unlocking the power of RGB-T tracking. Since it is laborious to collect high-quality RGB-T sequences, recent benchmarks only provide test sequences. In this paper, we construct a large-scale benchmark with high diversity for visible-thermal UAV tracking (VTUAV), including 500 sequences with 1.7 million high-resolution (1920 $\times$ 1080 pixels) frame pairs. In addition, comprehensive applications (short-term tracking, long-term tracking and segmentation mask prediction) with diverse categories and scenes are considered for exhaustive evaluation. Moreover, we provide a coarse-to-fine attribute annotation, where frame-level attributes are provided to exploit the potential of challenge-specific trackers. In addition, we design a new RGB-T baseline, named Hierarchical Multi-modal Fusion Tracker (HMFT), which fuses RGB-T data in various levels. Numerous experiments on several datasets are conducted to reveal the effectiveness of HMFT and the complement of different fusion types. The project is available at here. | ['Xiang Ruan', 'Huchuan Lu', 'Dong Wang', 'Jie Zhao', 'Pengyu Zhang'] | 2022-04-08 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zhang_Visible-Thermal_UAV_Tracking_A_Large-Scale_Benchmark_and_New_Baseline_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zhang_Visible-Thermal_UAV_Tracking_A_Large-Scale_Benchmark_and_New_Baseline_CVPR_2022_paper.pdf | cvpr-2022-1 | ['rgb-t-tracking'] | ['computer-vision'] | [-3.63935456e-02 -8.10184956e-01 -3.11096609e-02 -2.39442542e-01
-8.25013638e-01 -8.84185553e-01 3.92662078e-01 -4.91750866e-01
-3.76132339e-01 5.27522326e-01 -3.60616833e-01 -3.76057066e-02
6.06718846e-02 -4.82297719e-01 -7.00173318e-01 -1.08767748e+00
6.56049028e-02 1.47011176e-01 6.05145633e-01 -1.93611294e-01
-3.90108109e-01 3.33239019e-01 -1.80247700e+00 1.26846850e-01
6.62383795e-01 1.62407613e+00 3.68628114e-01 7.17305899e-01
1.83271855e-01 3.13398093e-01 -5.89925945e-01 -4.94920343e-01
7.85733163e-01 -6.73479261e-03 -1.24579273e-01 3.28651518e-02
7.44104922e-01 -3.84609222e-01 -1.68153852e-01 1.01491547e+00
5.47972441e-01 1.05390899e-01 1.58104137e-01 -1.53737819e+00
-2.79660016e-01 5.76598756e-02 -7.53493547e-01 2.97854662e-01
3.05935770e-01 7.68273950e-01 5.67952275e-01 -5.56135833e-01
3.42668355e-01 1.12031817e+00 9.77569580e-01 5.35674393e-01
-8.55968654e-01 -9.58123565e-01 2.46437922e-01 3.63717191e-02
-1.26635551e+00 -3.16734821e-01 3.85186791e-01 -5.86681068e-01
6.15347803e-01 4.51569796e-01 1.04396069e+00 1.15654302e+00
3.10682029e-01 6.57743871e-01 1.29367626e+00 1.23286165e-01
-1.45120144e-01 -3.24118763e-01 -1.89697623e-01 6.66149676e-01
3.54148805e-01 6.85249865e-01 -6.52030587e-01 -1.71827480e-01
5.28498530e-01 1.27903908e-01 -3.79234225e-01 -4.03547674e-01
-1.52517426e+00 2.87469119e-01 5.92445076e-01 -4.29410562e-02
-1.65680349e-01 5.20734072e-01 3.77875566e-01 8.37302655e-02
3.17949504e-01 3.96339782e-02 -5.31158566e-01 -1.79214939e-01
-8.03675830e-01 1.71147138e-02 1.24695696e-01 1.35370123e+00
7.05391467e-01 4.13687378e-01 -3.73331070e-01 2.11810023e-01
6.81761563e-01 1.19472098e+00 2.40768716e-01 -1.03015304e+00
3.79074842e-01 4.78814423e-01 2.60695130e-01 -6.90597296e-01
-3.50278407e-01 -4.92560863e-01 -8.87674928e-01 2.88349241e-01
3.43726367e-01 -2.71584779e-01 -9.73321855e-01 1.59499776e+00
7.22437620e-01 4.04122561e-01 -2.17451435e-02 1.29456127e+00
1.00705957e+00 5.25577366e-01 1.57785624e-01 -2.54628181e-01
1.40930390e+00 -8.84574294e-01 -7.01422393e-01 -2.68341273e-01
3.98601621e-01 -7.60399878e-01 6.09450221e-01 1.62626430e-01
-6.00222647e-01 -8.75186861e-01 -8.20881546e-01 3.18116546e-01
-3.50785464e-01 2.16846526e-01 7.28650510e-01 9.41036403e-01
-8.11808109e-01 2.11466998e-01 -1.01430714e+00 -3.71660113e-01
3.89457524e-01 3.15931886e-01 -2.24054560e-01 -7.48108178e-02
-1.06675744e+00 6.31379783e-01 4.38968033e-01 4.01030451e-01
-1.19366181e+00 -8.10269058e-01 -6.36919796e-01 -6.31242335e-01
6.33892298e-01 -8.33304822e-01 9.33770239e-01 -6.26105845e-01
-1.27001143e+00 5.71835518e-01 1.02966987e-02 -3.73154730e-01
6.06375456e-01 -4.16690081e-01 -3.96106094e-01 -8.38660076e-03
6.93520531e-02 8.00048351e-01 1.01703620e+00 -1.29051244e+00
-1.02882493e+00 -5.28155744e-01 -3.41924243e-02 1.29538924e-01
-1.75113425e-01 3.13293263e-02 -7.54721463e-01 -6.51869476e-01
1.26116024e-02 -1.19961429e+00 -1.65610716e-01 1.36628643e-01
-1.92404270e-01 2.83190552e-02 1.23278081e+00 -4.63614941e-01
8.66636217e-01 -2.12705874e+00 7.50999153e-02 -3.29008728e-01
2.33859986e-01 2.67851681e-01 8.06054845e-02 -1.38452545e-01
3.42334777e-01 -2.18219921e-01 -5.74665517e-02 -4.01431739e-01
5.04752900e-03 2.29568973e-01 -1.52480006e-01 6.66767001e-01
-2.46049255e-01 1.17480862e+00 -7.61539817e-01 -6.65566683e-01
5.80088496e-01 3.09460253e-01 -3.49764042e-02 1.35178313e-01
-4.04596835e-01 8.96352410e-01 -7.08711147e-01 1.31991041e+00
6.96057320e-01 -2.11518377e-01 -3.70554954e-01 -7.45130956e-01
-2.91729361e-01 -4.90176767e-01 -1.00559390e+00 1.84359789e+00
6.83864281e-02 6.22457981e-01 2.27708593e-01 -2.78210312e-01
6.97771132e-01 2.25184217e-01 9.30214882e-01 -6.08186007e-01
4.24711317e-01 -4.18726318e-02 -3.23430300e-01 -2.31250972e-01
7.82396197e-01 4.97195646e-02 -2.41164774e-01 -6.68838546e-02
-1.23724155e-01 5.07673286e-02 1.80564329e-01 -3.13397050e-02
9.60561752e-01 7.12757111e-01 -1.64956734e-01 1.39561314e-02
2.41044372e-01 3.62339288e-01 1.09479129e+00 6.29883587e-01
-6.89021587e-01 5.51961541e-01 -3.09074938e-01 -4.39613402e-01
-8.37134898e-01 -9.04163420e-01 -1.24019431e-03 1.13641655e+00
6.95924282e-01 -3.31605017e-01 -3.73761684e-01 -4.42519903e-01
1.28614172e-01 1.44644827e-01 -5.42062104e-01 1.57981906e-02
-4.31721330e-01 -1.02388310e+00 6.68346524e-01 5.94649971e-01
8.76631737e-01 -5.42410672e-01 -1.25579751e+00 -6.53207824e-02
-4.88686144e-01 -1.55940819e+00 -3.50426406e-01 1.05700269e-01
-8.09296489e-01 -1.08397830e+00 -5.01980245e-01 -3.71116549e-02
4.46064435e-02 6.47993624e-01 8.95336151e-01 -2.09322169e-01
-2.52404422e-01 7.18391895e-01 -5.89297056e-01 -3.17614943e-01
1.55603085e-02 -2.62247205e-01 3.24299425e-01 -2.90013719e-02
7.53712356e-02 1.10099822e-01 -6.09069109e-01 6.93602979e-01
-6.65537059e-01 4.57160920e-02 5.86114764e-01 4.99301136e-01
8.72850001e-01 -5.08738905e-02 -1.35332584e-01 -6.95729405e-02
-2.87045777e-01 -6.73635527e-02 -1.12768793e+00 4.47943121e-01
-2.29090422e-01 -3.84198785e-01 3.53296757e-01 -5.04504621e-01
-9.99298692e-01 4.16313469e-01 2.64491081e-01 -1.20930958e+00
-1.50231570e-01 4.51099910e-02 -2.86632717e-01 -5.28155804e-01
3.88652712e-01 2.72304177e-01 -1.65787652e-01 -3.49562019e-01
3.63546371e-01 2.15369001e-01 9.27923024e-01 -6.15943253e-01
1.17263186e+00 6.54883921e-01 4.00871336e-02 -5.14102399e-01
-8.63638461e-01 -5.42368293e-01 -5.33706188e-01 -8.28722417e-01
1.10499346e+00 -1.33210897e+00 -9.45619702e-01 5.28135717e-01
-7.99961567e-01 -3.62106204e-01 -2.25949317e-01 5.80596089e-01
-4.15139645e-01 2.85035759e-01 -3.42412651e-01 -8.42848897e-01
-3.22133809e-01 -1.38844216e+00 1.47452784e+00 3.71950567e-01
6.25969529e-01 -4.53339249e-01 -3.87926288e-02 6.78414404e-01
3.44951779e-01 5.98423660e-01 -1.89518243e-01 2.44745016e-02
-1.10141158e+00 -4.14877273e-02 -1.58968657e-01 5.85136786e-02
1.26181524e-02 1.59911796e-01 -1.14263570e+00 -6.64431632e-01
-1.61063269e-01 -2.90934235e-01 9.65937734e-01 6.53848648e-01
9.74095523e-01 3.05147648e-01 -5.95840454e-01 1.07072878e+00
1.39668107e+00 1.59102738e-01 3.31279427e-01 4.78809029e-01
1.20118463e+00 1.51747137e-01 1.47092438e+00 4.85186189e-01
4.31320101e-01 9.63373244e-01 8.38726699e-01 2.40659714e-03
-1.72787666e-01 2.48170897e-01 6.20637953e-01 5.61551392e-01
-4.71709490e-01 -3.11470330e-01 -8.49057913e-01 2.34744430e-01
-1.82085693e+00 -9.35186267e-01 -4.21993583e-01 2.21781445e+00
4.36098039e-01 -4.58164001e-03 2.42243379e-01 -2.16380969e-01
7.42334127e-01 2.51984447e-01 -6.72828913e-01 7.07243919e-01
-3.59168738e-01 -4.72823501e-01 1.06622148e+00 -8.23486969e-02
-1.46549606e+00 9.83440578e-01 5.62223434e+00 1.08425736e+00
-1.01124024e+00 2.92454660e-01 2.80201346e-01 -2.26344228e-01
2.68152565e-01 -3.72291654e-02 -1.16492510e+00 6.76408172e-01
8.45228851e-01 2.33496860e-01 1.24786593e-01 7.26341367e-01
-1.18722031e-02 -2.47884706e-01 -6.84583008e-01 1.24275053e+00
-2.17811003e-01 -9.93340671e-01 -3.61267090e-01 1.03520922e-01
6.43351555e-01 4.47824121e-01 1.86241463e-01 1.75770909e-01
4.22783405e-01 -5.21345198e-01 1.07236505e+00 4.85659719e-01
9.34222460e-01 -4.70319778e-01 7.31095612e-01 6.24843463e-02
-2.16084099e+00 -1.35494545e-01 -4.65129584e-01 4.31626737e-01
2.56795973e-01 3.16114366e-01 -2.58104056e-01 1.29263425e+00
1.34771240e+00 1.02910614e+00 -8.95632684e-01 1.19205499e+00
1.72938824e-01 2.85115838e-01 -6.38161480e-01 2.86610633e-01
1.16628990e-01 -1.32535398e-01 5.51496446e-01 8.96432579e-01
6.04290187e-01 2.90896744e-01 6.79563522e-01 4.20156211e-01
2.93909907e-01 -5.47702491e-01 -4.24821645e-01 1.52689099e-01
5.34694731e-01 1.61221170e+00 -7.47645497e-01 -2.30438754e-01
-3.96137148e-01 6.10424995e-01 -2.73940653e-01 1.48247957e-01
-1.38796341e+00 4.35607165e-01 9.29081798e-01 -2.57360905e-01
5.31658709e-01 -4.60798323e-01 7.09557720e-03 -1.09101272e+00
-1.17207728e-01 -7.42362142e-01 3.14585090e-01 -9.73423660e-01
-1.15162218e+00 6.01094782e-01 3.59322615e-02 -1.92138469e+00
9.36001614e-02 -5.24172664e-01 6.90122396e-02 5.76469600e-01
-1.47338283e+00 -1.65600073e+00 -9.51264143e-01 6.95088148e-01
3.85865957e-01 -1.40958592e-01 3.59984368e-01 5.50336480e-01
-7.05448270e-01 4.07652438e-01 1.09757669e-01 4.80540060e-02
8.36699665e-01 -9.50127959e-01 1.00603752e-01 1.06340408e+00
-6.35319203e-02 1.59835055e-01 7.22958088e-01 -8.99831355e-01
-2.03306341e+00 -1.48736823e+00 -2.60573953e-01 -8.12395573e-01
5.89016974e-01 -1.55016512e-01 -5.02759039e-01 6.77714348e-01
1.34083956e-01 4.89630073e-01 4.31807846e-01 -4.31479394e-01
-3.13124806e-02 -3.96102309e-01 -8.81519318e-01 4.98972125e-02
9.89090681e-01 -1.15651950e-01 -3.97606120e-02 2.01751709e-01
9.88240719e-01 -8.66501570e-01 -1.33305073e+00 8.72740090e-01
6.63448751e-01 -1.04563081e+00 1.23056757e+00 1.43175960e-01
-2.98061460e-01 -9.91876483e-01 -4.42490369e-01 -9.95129287e-01
-2.28874847e-01 -4.57059324e-01 -2.11990908e-01 1.32862270e+00
-1.71399176e-01 -4.20858085e-01 7.61286616e-01 4.75842535e-01
-4.58402991e-01 -3.94192100e-01 -1.10010111e+00 -1.09124506e+00
-5.82412064e-01 -5.54562747e-01 7.39232898e-01 7.69255400e-01
-9.52949941e-01 -1.40242249e-01 -6.84656143e-01 5.15092313e-01
1.13848305e+00 5.07346869e-01 1.05574465e+00 -1.16914392e+00
-1.31112799e-01 -1.53123438e-01 -4.73328441e-01 -1.10998261e+00
-2.04564661e-01 -3.72372448e-01 2.65702158e-01 -1.08325994e+00
1.73454925e-01 -6.32659912e-01 -2.32381687e-01 4.12124872e-01
-2.37749606e-01 4.97738779e-01 4.74560499e-01 4.43317324e-01
-1.19107640e+00 9.23931360e-01 1.40583444e+00 -1.80222049e-01
1.50159284e-01 -5.72437327e-03 -2.23443750e-02 5.05253315e-01
5.58122039e-01 -3.93796593e-01 -8.67772773e-02 -6.15733027e-01
-7.05303326e-02 2.92394489e-01 6.94755733e-01 -1.44566917e+00
5.40240049e-01 -3.09318900e-01 6.12090826e-01 -1.23357451e+00
7.74998069e-01 -1.19581807e+00 7.72897422e-01 4.59657848e-01
2.95121014e-01 3.75004828e-01 4.09161270e-01 8.36286604e-01
-6.96802065e-02 4.02579695e-01 6.56509936e-01 3.00301556e-02
-1.24714458e+00 7.66869187e-01 9.12390277e-02 -5.82614094e-02
1.31916809e+00 -5.57894945e-01 -4.97548163e-01 1.30602971e-01
-3.88711482e-01 5.40894508e-01 8.24009418e-01 5.30489326e-01
4.78859007e-01 -1.46091807e+00 -4.84132022e-01 -6.37535527e-02
3.29490721e-01 1.64792955e-01 4.23679501e-01 1.17957616e+00
-1.94134220e-01 3.72320384e-01 -4.04751629e-01 -1.22273612e+00
-1.44968212e+00 5.94718933e-01 3.90293032e-01 7.67621100e-02
-6.03994608e-01 7.79391825e-01 2.16219544e-01 -2.49391109e-01
1.96766436e-01 -4.52956349e-01 1.50008112e-01 2.00934000e-02
3.92479956e-01 3.52693111e-01 -6.19119480e-02 -1.00905049e+00
-6.52311981e-01 8.43650877e-01 4.24251944e-01 1.22582808e-01
9.47910309e-01 -4.43673193e-01 1.77523866e-01 3.62973243e-01
6.34031713e-01 -4.43206668e-01 -1.87164116e+00 -1.96861908e-01
-2.90513515e-01 -8.05105805e-01 1.04478449e-01 -5.05052209e-01
-1.55678332e+00 5.55527747e-01 1.17305422e+00 4.94398959e-02
1.24796081e+00 -8.24531391e-02 8.58353376e-01 2.03214735e-01
8.25984955e-01 -8.32441628e-01 4.28440189e-03 3.55629563e-01
2.47604668e-01 -1.57492459e+00 2.18813047e-01 -3.85416716e-01
-5.78894079e-01 7.67661333e-01 9.66620266e-01 4.39435929e-01
8.55814517e-02 4.54493761e-01 1.55906379e-01 -1.81516543e-01
-7.35409558e-01 -6.66899800e-01 2.79850781e-01 7.36096561e-01
8.48493800e-02 1.05962686e-01 3.92501086e-01 1.42791823e-01
-1.51805729e-02 -1.62637740e-01 1.16108000e-01 1.05396807e+00
-4.76036578e-01 -7.41699934e-01 -9.18895364e-01 3.56176257e-01
-2.25977629e-01 2.48420835e-01 -1.23322815e-01 7.51085877e-01
3.87542188e-01 1.05648720e+00 -3.02543998e-01 -7.83748567e-01
1.53635770e-01 -3.36906403e-01 5.86491466e-01 -9.90013629e-02
-8.22582304e-01 4.41273093e-01 -1.43194064e-01 -1.02248585e+00
-1.05199194e+00 -8.06444764e-01 -9.41395760e-01 -5.89107513e-01
-6.07959092e-01 -1.67113140e-01 6.76925361e-01 8.81855071e-01
2.20123351e-01 8.14137220e-01 3.10554922e-01 -1.11380100e+00
-9.81897563e-02 -7.59112239e-01 -3.56559873e-01 1.90231264e-01
4.21089292e-01 -1.12410283e+00 -1.02086775e-01 1.94243431e-01] | [6.448387622833252, -2.1819286346435547] |
96316df3-74c6-4b58-82b5-b6d8dbad32a6 | random-walk-on-multiple-networks | 2307.01637 | null | https://arxiv.org/abs/2307.01637v1 | https://arxiv.org/pdf/2307.01637v1.pdf | Random Walk on Multiple Networks | Random Walk is a basic algorithm to explore the structure of networks, which can be used in many tasks, such as local community detection and network embedding. Existing random walk methods are based on single networks that contain limited information. In contrast, real data often contain entities with different types or/and from different sources, which are comprehensive and can be better modeled by multiple networks. To take advantage of rich information in multiple networks and make better inferences on entities, in this study, we propose random walk on multiple networks, RWM. RWM is flexible and supports both multiplex networks and general multiple networks, which may form many-to-many node mappings between networks. RWM sends a random walker on each network to obtain the local proximity (i.e., node visiting probabilities) w.r.t. the starting nodes. Walkers with similar visiting probabilities reinforce each other. We theoretically analyze the convergence properties of RWM. Two approximation methods with theoretical performance guarantees are proposed for efficient computation. We apply RWM in link prediction, network embedding, and local community detection. Comprehensive experiments conducted on both synthetic and real-world datasets demonstrate the effectiveness and efficiency of RWM. | ['Xiang Zhang', 'Xiao Liu', 'Jun Huan', 'Xiong Yu', 'Yaowei Yan', 'Yuchen Bian', 'Dongsheng Luo'] | 2023-07-04 | null | null | null | null | ['link-prediction', 'local-community-detection', 'community-detection', 'network-embedding'] | ['graphs', 'graphs', 'graphs', 'methodology'] | [-1.50375634e-01 6.70470819e-02 -5.94204843e-01 -1.30681455e-01
2.44931281e-01 -5.70319712e-01 4.59227294e-01 5.02416730e-01
-2.04945013e-01 7.51524329e-01 1.77112117e-01 -2.83758640e-01
-5.75605333e-01 -1.52312696e+00 -2.19264314e-01 -4.75962251e-01
-7.43600607e-01 7.44380414e-01 6.93270445e-01 -5.56844249e-02
4.08522412e-02 3.34287256e-01 -6.59237623e-01 -2.58855850e-01
6.50605083e-01 2.45968401e-01 3.03650588e-01 4.26916122e-01
-4.69119877e-01 6.14087224e-01 -4.92879182e-01 -4.03876722e-01
1.01434007e-01 -1.36435896e-01 -7.51425803e-01 -2.49564469e-01
-5.30678928e-01 -1.69651404e-01 -7.53408194e-01 1.01138902e+00
4.31411028e-01 -8.75807106e-02 4.63692516e-01 -1.79122674e+00
-7.65718222e-01 1.27239871e+00 -9.23130989e-01 2.39137918e-01
5.53241253e-01 -3.43942314e-01 1.27858090e+00 -8.08897495e-01
8.27426076e-01 1.56210613e+00 7.55649745e-01 1.78419694e-01
-1.26739883e+00 -8.55738461e-01 3.51059884e-01 9.75152329e-02
-1.77821636e+00 -1.21493284e-02 6.98858857e-01 -2.61660248e-01
4.14269239e-01 3.09854299e-01 7.25996196e-01 5.52374125e-01
-1.90398172e-01 4.88955975e-01 6.15954280e-01 -1.54973060e-01
-2.33272508e-01 3.22857231e-01 1.26743004e-01 7.06944048e-01
8.10659945e-01 -3.89306724e-01 -4.06942248e-01 -4.94383186e-01
8.84461820e-01 4.77257580e-01 -4.34149325e-01 -4.59576905e-01
-1.53705537e+00 1.06252873e+00 9.91716385e-01 2.04074740e-01
-1.91030651e-01 1.76077098e-01 2.95528293e-01 4.80936021e-01
1.16403908e-01 -2.53324979e-03 -1.01428263e-01 3.00241083e-01
-5.44276893e-01 -2.61060297e-01 9.38280463e-01 1.17857957e+00
9.80033755e-01 -4.28513110e-01 2.15583518e-01 7.30268419e-01
7.78214097e-01 5.68825960e-01 1.53073907e-01 -4.72402245e-01
7.30914176e-01 9.16653454e-01 -2.50710011e-01 -1.91895437e+00
-3.98274809e-01 -2.75858879e-01 -1.48614168e+00 -5.27166188e-01
-1.70332581e-01 -1.56741709e-01 -2.65876055e-01 1.66937232e+00
5.72149456e-01 5.24788022e-01 5.37283868e-02 5.00104308e-01
9.10688162e-01 8.61466169e-01 -2.21063301e-01 -2.41617352e-01
9.75989223e-01 -9.92885172e-01 -5.28574526e-01 -4.41597868e-03
6.24691129e-01 -5.12635469e-01 4.21089530e-01 -4.59641576e-01
-8.31053555e-01 -1.09774508e-01 -7.74450958e-01 4.63187546e-01
-3.81548405e-01 -3.80808026e-01 6.57229125e-01 3.28960180e-01
-1.20077693e+00 5.77725947e-01 -5.61160684e-01 -5.55082142e-01
3.07118684e-01 4.76560444e-01 -4.32588160e-01 -2.14074969e-01
-1.61180496e+00 4.19585109e-01 5.97230613e-01 2.54409254e-01
-2.76982009e-01 -3.18823606e-01 -7.34974682e-01 2.33032092e-01
5.39102912e-01 -6.70869648e-01 3.47989112e-01 -3.41387212e-01
-6.55080795e-01 3.20850343e-01 -2.69963026e-01 -4.54738826e-01
2.61352122e-01 4.01952744e-01 -8.92485440e-01 3.03402156e-01
3.69640619e-01 3.62732828e-01 1.96011826e-01 -1.02098525e+00
-5.48094571e-01 1.78431302e-01 1.89947218e-01 -4.97536324e-02
-7.79384375e-01 8.88157561e-02 -5.24919748e-01 -4.26005870e-01
3.68641973e-01 -8.49127412e-01 -3.47188562e-01 2.20588416e-01
-8.22284818e-01 -2.61940658e-01 8.21509600e-01 -6.12312816e-02
1.79064167e+00 -1.81438529e+00 2.21457243e-01 1.11147058e+00
9.37506080e-01 1.98141143e-01 -2.65763313e-01 9.30920243e-01
2.99115852e-02 6.17472827e-01 -4.88307141e-02 3.30908634e-02
-1.79949597e-01 3.07750285e-01 3.20126191e-02 3.93077224e-01
-9.70994160e-02 8.65196109e-01 -1.14461577e+00 -8.84949744e-01
-4.92403954e-02 2.82047778e-01 -3.44393343e-01 -1.06893718e-01
4.62460160e-01 -2.06296016e-02 -7.02381909e-01 4.85141337e-01
6.99970484e-01 -9.69362974e-01 7.19651878e-01 -1.67248130e-01
1.97234243e-01 4.19540629e-02 -1.62884331e+00 9.47818279e-01
-1.53047547e-01 6.63049638e-01 5.87883890e-02 -9.36718643e-01
1.05539584e+00 2.37354323e-01 5.24405837e-01 -4.17793393e-02
-2.61550188e-01 3.00710350e-01 1.70079023e-01 -2.09080756e-01
2.36401007e-01 3.57001781e-01 3.18066686e-01 7.40240812e-01
-3.24944854e-01 7.49937832e-01 4.65989649e-01 8.26634765e-01
1.20582414e+00 -8.53412211e-01 5.21020472e-01 -2.65516609e-01
4.89320129e-01 -3.55204314e-01 6.89338207e-01 6.69089913e-01
-1.03662580e-01 9.43114683e-02 6.59312367e-01 -2.10007101e-01
-7.26045132e-01 -1.07133031e+00 1.28223777e-01 5.81089318e-01
6.96990073e-01 -7.16019690e-01 -8.04212466e-02 -5.23284376e-01
3.17571700e-01 -9.27984789e-02 -4.73178715e-01 -1.35939941e-01
-4.75611627e-01 -8.01024795e-01 3.74914020e-01 2.40726590e-01
4.19652402e-01 -9.28791106e-01 3.42801303e-01 4.90599155e-01
-4.49542016e-01 -9.57034111e-01 -6.41010582e-01 -5.10518849e-01
-7.44852304e-01 -1.40265965e+00 -4.65637088e-01 -1.06796861e+00
9.63457584e-01 8.27288628e-01 1.09436083e+00 6.91565931e-01
-6.55002221e-02 2.03581154e-01 -3.57830256e-01 1.64944008e-01
-2.30058283e-01 4.24293369e-01 3.43360215e-01 1.72940999e-01
2.57831365e-01 -9.24445510e-01 -7.36284971e-01 6.70868039e-01
-7.82498360e-01 4.64154445e-02 7.12627411e-01 6.64315403e-01
2.77105331e-01 4.38118160e-01 6.67056382e-01 -1.06204569e+00
9.26989198e-01 -1.09861577e+00 -3.56205732e-01 5.14637887e-01
-8.74674857e-01 -3.51262279e-02 3.07184339e-01 -5.97882211e-01
-3.31009477e-01 -6.10032618e-01 2.43104964e-01 -6.38273284e-02
5.41133404e-01 9.29300606e-01 -1.32793114e-01 -2.26340771e-01
1.69880599e-01 1.22336358e-01 2.20976211e-02 -2.40729347e-01
2.36051574e-01 6.16797328e-01 -1.60000980e-01 -3.31130922e-01
1.28758836e+00 3.65893364e-01 3.46168995e-01 -6.86882854e-01
-1.46771997e-01 -6.73779964e-01 -4.82209206e-01 -2.27811664e-01
2.24865943e-01 -8.88934791e-01 -8.81875277e-01 3.28286171e-01
-1.06874979e+00 8.83069336e-02 2.52706796e-01 4.92414385e-01
4.08472449e-01 5.50948262e-01 -8.51656139e-01 -6.21306837e-01
-3.21602166e-01 -6.91201389e-01 1.24517784e-01 2.38707662e-01
-2.49518260e-01 -1.65740299e+00 6.30615503e-02 -3.45458061e-01
4.46743459e-01 2.09473118e-01 9.09302652e-01 -6.75352097e-01
-8.30281198e-01 -3.08068812e-01 -5.99114597e-01 -3.69803756e-01
5.48798144e-01 3.00441444e-01 -6.92090094e-02 -5.13534725e-01
-9.43407476e-01 2.51494020e-01 6.79536164e-01 -3.53928320e-02
8.32091570e-01 -6.93588257e-01 -1.15698814e+00 3.63538593e-01
1.50242543e+00 -2.89693445e-01 5.31515360e-01 3.40202838e-01
9.14445341e-01 7.46523499e-01 4.54716742e-01 5.02616644e-01
6.58244193e-01 3.66539150e-01 6.06335104e-01 -8.54629725e-02
1.71302512e-01 -7.37181664e-01 1.18819460e-01 1.40407550e+00
4.47016656e-02 -6.04696035e-01 -9.09673512e-01 6.60906315e-01
-1.99822938e+00 -1.13172221e+00 -5.08586049e-01 1.99460280e+00
7.03057170e-01 3.26195538e-01 3.02621901e-01 1.66411698e-02
1.54004335e+00 2.52150357e-01 -5.13093233e-01 2.84919828e-01
-2.78183311e-01 -4.69046891e-01 5.98270178e-01 6.28579617e-01
-6.00552797e-01 6.08893931e-01 6.09501934e+00 8.62673104e-01
-6.01192713e-01 -1.19238190e-01 2.22011328e-01 1.21093594e-01
-8.05708349e-01 3.80936801e-01 -7.37432539e-01 7.43565619e-01
6.37753129e-01 -5.89490294e-01 4.58361894e-01 6.96687102e-01
-1.40469987e-03 3.31676304e-01 -6.74481273e-01 8.18385243e-01
-4.84362632e-01 -1.63288379e+00 3.24935168e-01 3.80181104e-01
6.55583143e-01 1.34127647e-01 -3.41600567e-01 -1.02941580e-01
7.96538830e-01 -8.03221107e-01 1.29584176e-02 4.40020204e-01
7.08969414e-01 -7.84757733e-01 6.65801883e-01 3.65381032e-01
-2.11241531e+00 -5.88054694e-02 -4.99196202e-01 1.37522668e-01
3.84681195e-01 9.15863335e-01 -7.72737801e-01 8.30973923e-01
5.09653449e-01 1.23089278e+00 -4.59859490e-01 1.15553212e+00
-1.80162549e-01 4.01267081e-01 -6.70975566e-01 -5.36556363e-01
1.37768567e-01 -4.95245427e-01 6.70705378e-01 1.01228142e+00
3.08433354e-01 -8.55633020e-02 4.31101531e-01 6.10918880e-01
-4.73516047e-01 2.54443407e-01 -6.91928506e-01 -1.04968734e-01
1.46704471e+00 1.44138706e+00 -9.40593481e-01 -2.49324530e-01
-5.20817041e-01 5.50468087e-01 5.80004394e-01 4.86546487e-01
-6.47291005e-01 -9.32608664e-01 6.46023750e-01 5.05021930e-01
1.70752838e-01 -2.71107167e-01 4.25105572e-01 -1.08406699e+00
-3.13716121e-02 -3.68088603e-01 5.16289413e-01 -5.77670753e-01
-1.52056956e+00 6.38235748e-01 -1.62944958e-01 -1.31500506e+00
2.76064962e-01 -1.31310433e-01 -1.03907275e+00 8.20114434e-01
-1.47554040e+00 -9.15631711e-01 -2.80193120e-01 7.27963865e-01
-3.19954306e-01 -1.71313748e-01 6.63061857e-01 4.32344913e-01
-7.09227264e-01 6.32235706e-01 4.47232008e-01 6.27085924e-01
5.44585705e-01 -1.03242040e+00 4.61427659e-01 8.20103645e-01
8.41181278e-02 1.12559080e+00 3.40106070e-01 -8.58802080e-01
-1.11273634e+00 -1.31117034e+00 1.01436162e+00 2.14989990e-01
1.14167762e+00 -2.18484804e-01 -9.40112412e-01 7.95014739e-01
3.20757180e-03 1.58023044e-01 8.46694112e-01 3.07111323e-01
-2.47854725e-01 -2.36034855e-01 -1.23261893e+00 8.87820125e-01
1.35090172e+00 -4.74806249e-01 -1.27015868e-02 4.00747210e-01
9.66717303e-01 2.20075399e-01 -1.09603357e+00 2.07304880e-01
4.80409831e-01 -6.06247663e-01 1.24180484e+00 -4.82224733e-01
9.88066569e-02 -5.19405484e-01 1.03921257e-01 -1.46193540e+00
-6.26196861e-01 -7.67694116e-01 -4.31524456e-01 1.58299935e+00
4.83876407e-01 -1.28802001e+00 6.97622716e-01 1.95597872e-01
7.43005097e-01 -8.13567758e-01 -7.42065370e-01 -7.20313847e-01
-3.53216052e-01 1.70654967e-01 1.13499618e+00 1.20382500e+00
1.55280098e-01 5.37244081e-01 -4.30382699e-01 3.63959134e-01
8.30693603e-01 2.72518635e-01 7.99488246e-01 -1.79673052e+00
-6.66185617e-02 -4.61907268e-01 -5.97568154e-01 -1.27032316e+00
3.63345072e-02 -1.04749346e+00 -5.59834421e-01 -1.96040332e+00
5.36170542e-01 -1.00352275e+00 -3.51655543e-01 2.40817294e-01
-1.78462580e-01 -9.35359150e-02 -2.28934027e-02 7.76463270e-01
-7.55267560e-01 3.63964319e-01 1.15048611e+00 -1.60422251e-01
-2.00969949e-01 2.10098043e-01 -7.05323279e-01 6.73871636e-01
9.38112140e-01 -6.78597629e-01 -6.19696140e-01 -2.95798182e-01
6.69503391e-01 1.69194698e-01 2.61596054e-01 -6.05443120e-01
5.57691216e-01 -1.44148886e-01 -2.73360610e-02 -6.89673543e-01
1.19984366e-01 -9.44801450e-01 4.71249104e-01 7.11216927e-01
-3.46588969e-01 3.48912686e-01 -4.72981840e-01 1.25361645e+00
-1.21407814e-01 -1.44025519e-01 3.64224017e-01 -6.06314987e-02
-5.32360196e-01 9.17796373e-01 -7.67340139e-02 7.65489861e-02
1.09896302e+00 -2.15610415e-01 -4.88343835e-01 -6.47736609e-01
-7.36942530e-01 9.42281961e-01 3.77071321e-01 3.56091678e-01
8.45037699e-01 -1.80722570e+00 -7.70380795e-01 -1.81819037e-01
1.84367180e-01 -4.42159846e-02 1.99860036e-02 8.57185721e-01
-5.14045894e-01 1.04439244e-01 5.42392023e-02 -4.70221341e-01
-1.63704646e+00 6.23228014e-01 1.55887827e-01 -4.75656956e-01
-4.23954576e-01 5.94221294e-01 -1.95774600e-01 -7.25590289e-01
-5.25307469e-03 2.43280873e-01 -4.24904913e-01 8.13951418e-02
4.58124191e-01 6.25782430e-01 -7.41162777e-01 -5.49381733e-01
-6.15598917e-01 6.23162389e-01 -1.66603953e-01 4.60383892e-02
1.19958651e+00 -5.83575964e-01 -8.43128562e-01 4.77358222e-01
1.38346255e+00 2.50705451e-01 -3.74470234e-01 -8.94359350e-01
7.19248503e-02 -7.14571774e-01 -4.14643228e-01 -2.39331815e-02
-1.31636810e+00 6.28284514e-01 -1.99816702e-03 7.55473912e-01
7.28325009e-01 9.96351801e-03 7.96367288e-01 6.41787648e-01
6.06653035e-01 -7.37169862e-01 -1.93492603e-02 2.24975988e-01
2.99311012e-01 -1.07612741e+00 2.06872016e-01 -9.45324600e-01
-3.14677566e-01 1.12156761e+00 6.27360165e-01 1.45904481e-01
1.20322204e+00 -1.77325875e-01 -4.48522747e-01 -2.05231652e-01
-6.58799469e-01 -1.09957956e-01 -2.64464885e-01 6.00973368e-01
3.57896909e-02 2.02819705e-01 -1.60960749e-01 1.47776693e-01
2.12721154e-01 -4.57294822e-01 6.76000834e-01 7.54899979e-01
-6.74431980e-01 -1.40834713e+00 -1.74627993e-02 6.59746885e-01
-2.93716677e-02 -2.76632071e-01 -2.99144566e-01 7.75357485e-01
-2.32955873e-01 1.02863550e+00 -1.30158085e-02 -6.44843221e-01
2.06043366e-02 -5.24372041e-01 -1.55133560e-01 -5.97661853e-01
-3.03990990e-02 -4.04428154e-01 7.58204237e-02 -5.19250073e-02
-5.65520883e-01 -4.11866039e-01 -1.29857934e+00 -1.19315851e+00
-7.56936789e-01 5.99676490e-01 1.17490880e-01 5.00533998e-01
5.08789420e-01 4.23003703e-01 1.10646236e+00 -8.41291621e-02
-1.13515481e-01 -8.05571079e-01 -9.03376818e-01 1.78768292e-01
2.50550985e-01 -6.72626197e-01 -4.55899209e-01 -5.52145183e-01] | [7.226092338562012, 5.949730396270752] |
15d5e3bc-5437-41b4-9edd-e462b8b9430e | joint-discriminative-and-generative-learning | 1904.07223 | null | https://arxiv.org/abs/1904.07223v3 | https://arxiv.org/pdf/1904.07223v3.pdf | Joint Discriminative and Generative Learning for Person Re-identification | Person re-identification (re-id) remains challenging due to significant intra-class variations across different cameras. Recently, there has been a growing interest in using generative models to augment training data and enhance the invariance to input changes. The generative pipelines in existing methods, however, stay relatively separate from the discriminative re-id learning stages. Accordingly, re-id models are often trained in a straightforward manner on the generated data. In this paper, we seek to improve learned re-id embeddings by better leveraging the generated data. To this end, we propose a joint learning framework that couples re-id learning and data generation end-to-end. Our model involves a generative module that separately encodes each person into an appearance code and a structure code, and a discriminative module that shares the appearance encoder with the generative module. By switching the appearance or structure codes, the generative module is able to generate high-quality cross-id composed images, which are online fed back to the appearance encoder and used to improve the discriminative module. The proposed joint learning framework renders significant improvement over the baseline without using generated data, leading to the state-of-the-art performance on several benchmark datasets. | ['Xiaodong Yang', 'Zhiding Yu', 'Jan Kautz', 'Yi Yang', 'Zhedong Zheng', 'Liang Zheng'] | 2019-04-15 | joint-discriminative-and-generative-learning-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Zheng_Joint_Discriminative_and_Generative_Learning_for_Person_Re-Identification_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Zheng_Joint_Discriminative_and_Generative_Learning_for_Person_Re-Identification_CVPR_2019_paper.pdf | cvpr-2019-6 | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 2.28796706e-01 -1.13520086e-01 2.02907532e-01 -5.88368654e-01
-6.05844140e-01 -5.35470009e-01 8.19958806e-01 -2.19282299e-01
-3.94137889e-01 4.59080428e-01 3.39368880e-01 3.34883749e-01
3.67138952e-01 -6.61177099e-01 -7.19084620e-01 -6.99599743e-01
5.22686839e-01 4.66978639e-01 -1.64207906e-01 7.51087293e-02
-1.10239081e-01 2.30134487e-01 -1.59154236e+00 3.56524438e-02
8.78851771e-01 7.11908996e-01 6.68584257e-02 4.76320446e-01
7.28392377e-02 4.22428489e-01 -4.37235415e-01 -8.55470777e-01
2.53899068e-01 -6.34647965e-01 -3.35122973e-01 3.70149970e-01
6.75190687e-01 -5.26364744e-01 -6.64820492e-01 1.13645744e+00
7.19084918e-01 2.21306860e-01 6.16797745e-01 -1.23680222e+00
-1.17782617e+00 2.51873493e-01 -6.64144576e-01 -5.27558150e-03
2.76568413e-01 2.07881391e-01 8.79871428e-01 -1.03278172e+00
4.23848003e-01 1.36327624e+00 4.95958835e-01 9.46870387e-01
-1.42893231e+00 -8.25096786e-01 3.61332566e-01 2.97183514e-01
-1.40799880e+00 -5.83368778e-01 8.79653990e-01 -5.08732021e-01
6.17807031e-01 -9.07628238e-02 5.93995392e-01 1.38603139e+00
-2.47591749e-01 7.99682856e-01 8.92276943e-01 -3.44571769e-01
3.93772162e-02 2.96915650e-01 -6.38113394e-02 4.16563869e-01
2.89106876e-01 1.32368088e-01 -3.19882840e-01 2.04731315e-01
7.64965892e-01 4.79762554e-01 -2.16621339e-01 -3.93053859e-01
-9.26640928e-01 6.23119295e-01 5.31711400e-01 6.04124367e-02
-2.93870986e-01 -1.48901008e-02 2.99765587e-01 -6.28675520e-02
4.38468069e-01 4.01949603e-03 -3.57721522e-02 5.58963008e-02
-8.86123240e-01 3.24567854e-01 3.07083368e-01 8.10322940e-01
7.51153827e-01 -9.37021822e-02 -4.88492399e-01 1.25081909e+00
4.72492486e-01 6.13040507e-01 5.04455268e-01 -3.69989574e-01
5.66838861e-01 6.83751225e-01 8.89308155e-02 -9.69269276e-01
-1.46068027e-03 -6.24960303e-01 -1.01846385e+00 7.78460279e-02
1.68220028e-01 -4.03416343e-02 -9.24005985e-01 2.02666116e+00
3.45230997e-01 3.58047426e-01 4.26198170e-02 9.16037917e-01
6.37610734e-01 5.22114277e-01 2.00237244e-01 1.93527341e-01
1.37969089e+00 -1.36237168e+00 -6.39735699e-01 -3.68338376e-01
2.80350715e-01 -7.72472441e-01 7.25408196e-01 5.04238233e-02
-9.35785770e-01 -1.16605186e+00 -1.09009862e+00 -1.03732601e-01
-2.09414884e-01 5.27009428e-01 -5.07786451e-03 6.16357327e-01
-9.73775446e-01 2.98128039e-01 -6.50586247e-01 -3.92823279e-01
4.48098481e-01 3.09208721e-01 -6.33051455e-01 -4.10660028e-01
-9.36939538e-01 6.52753830e-01 3.29289794e-01 1.63290128e-01
-9.15944457e-01 -4.95451689e-01 -9.65891957e-01 -4.62608971e-02
-8.46621208e-03 -9.59922433e-01 9.63336527e-01 -1.18467152e+00
-1.51523244e+00 8.93362999e-01 -2.63591200e-01 -1.94771022e-01
5.97612917e-01 -2.49188155e-01 -5.39933741e-01 -1.26444399e-01
1.78043887e-01 6.78207278e-01 1.14948952e+00 -1.45420098e+00
-6.83465362e-01 -5.46590626e-01 -2.23574843e-02 2.39153013e-01
-5.56808770e-01 -8.01645890e-02 -9.13490534e-01 -9.92349803e-01
-3.46059889e-01 -1.27567732e+00 5.76057248e-02 -5.88073097e-02
-4.13156718e-01 -2.51353472e-01 7.70140767e-01 -9.06087041e-01
1.06032050e+00 -2.29834437e+00 4.48263168e-01 -5.75911105e-02
2.13873148e-01 5.03312647e-01 -3.18978339e-01 2.39868581e-01
-4.31627959e-01 -2.22466841e-01 -9.73032564e-02 -1.02177083e+00
1.24502651e-01 -2.83582453e-02 -1.52211398e-01 4.57042336e-01
2.31098741e-01 9.72072721e-01 -8.87913346e-01 -2.72092819e-01
3.81122351e-01 7.24300086e-01 -6.69122756e-01 6.88471377e-01
2.62013912e-01 6.19973958e-01 -1.53252125e-01 5.25284290e-01
9.80301499e-01 -1.91636875e-01 8.73985514e-02 -5.14363229e-01
1.87144428e-02 -1.23446450e-01 -1.16583681e+00 1.70437098e+00
-4.88636345e-01 3.61882687e-01 -1.61539391e-01 -1.00443101e+00
9.97031569e-01 1.58425540e-01 2.84266174e-01 -7.68961132e-01
-1.01793565e-01 7.49737918e-02 -1.24523193e-01 -2.52156347e-01
3.68495017e-01 -1.77970622e-02 -8.82809460e-02 5.59795737e-01
2.77223676e-01 6.10201240e-01 1.36700004e-01 1.29006982e-01
6.77967966e-01 3.30534309e-01 3.10514728e-03 1.87453747e-01
8.23750377e-01 -4.92818177e-01 5.26515365e-01 3.75188053e-01
-1.24756396e-01 8.41763377e-01 -1.56807452e-01 -3.97702545e-01
-1.31651437e+00 -1.22301006e+00 7.69334063e-02 1.00303674e+00
3.80128384e-01 -2.92826861e-01 -8.14471066e-01 -9.16311443e-01
2.32879743e-02 4.09629971e-01 -7.58126020e-01 -4.46243942e-01
-5.71156383e-01 -6.75814450e-01 4.08531576e-01 7.79553592e-01
8.82782638e-01 -7.63388455e-01 4.10904475e-02 2.42951527e-01
-2.95549840e-01 -1.16701341e+00 -9.82370853e-01 -2.91341245e-01
-5.57303429e-01 -8.80725384e-01 -1.17761636e+00 -8.21988761e-01
1.04737806e+00 5.35096169e-01 8.16210687e-01 -8.33397955e-02
-3.72201741e-01 4.38973844e-01 -4.15736228e-01 -3.72276120e-02
-2.85903603e-01 8.98839999e-03 2.12625906e-01 6.91580951e-01
4.57312226e-01 -4.63081509e-01 -8.97021770e-01 3.12590569e-01
-8.48262906e-01 1.50473848e-01 6.65121377e-01 1.01070380e+00
6.66593254e-01 -1.63118958e-01 4.42593962e-01 -6.74089253e-01
2.52753794e-01 -3.07401031e-01 -3.86323065e-01 3.54970664e-01
-5.83843172e-01 2.12119967e-01 7.11044133e-01 -4.53993022e-01
-1.33182716e+00 1.88197657e-01 -2.36390591e-01 -4.92281824e-01
-2.35390440e-01 6.33512018e-03 -6.50743783e-01 1.04716010e-01
2.32302710e-01 4.03071970e-01 -1.53313473e-01 -8.05518448e-01
4.84259516e-01 8.96531165e-01 7.95055747e-01 -4.86063987e-01
1.16034901e+00 4.87849027e-01 -3.67468804e-01 -4.93889242e-01
-7.40524948e-01 -4.21399117e-01 -8.30022156e-01 -2.14230537e-01
1.02383375e+00 -1.31601799e+00 -1.20971173e-01 8.74381721e-01
-1.08781266e+00 -8.43211263e-02 -2.49884069e-01 3.19846869e-01
-2.03390107e-01 4.85026777e-01 -3.71546894e-01 -5.80142558e-01
-4.22622085e-01 -1.22066236e+00 1.28199077e+00 6.60721183e-01
8.92372727e-02 -8.97800088e-01 3.70596170e-01 4.86321926e-01
2.76532650e-01 5.30803613e-02 5.48553228e-01 -5.64695239e-01
-4.54045922e-01 -4.08704937e-01 -5.28399765e-01 5.00944853e-01
4.88441497e-01 -2.31070206e-01 -1.20386708e+00 -6.87111497e-01
-4.79939491e-01 -1.49281383e-01 8.56298685e-01 -8.56385902e-02
1.13700724e+00 -2.48378009e-01 -4.58568811e-01 9.16628242e-01
1.42066181e+00 4.96302880e-02 7.56521165e-01 3.46982777e-01
1.10084653e+00 5.34640670e-01 2.94704229e-01 4.63543087e-01
6.97656631e-01 1.15945423e+00 1.41335577e-01 -3.28864366e-01
-5.42011201e-01 -6.42965376e-01 4.96879160e-01 6.18486822e-01
-1.25759691e-01 -2.29269743e-01 -5.00791550e-01 6.03565812e-01
-1.86385775e+00 -1.10559869e+00 1.30872309e-01 2.24319911e+00
7.01735198e-01 -2.57338256e-01 4.25230831e-01 -8.72987211e-02
1.09061873e+00 7.94707984e-02 -7.41438329e-01 5.40967323e-02
2.07860135e-02 2.84333900e-02 2.64102906e-01 3.64259005e-01
-1.15998459e+00 7.48394370e-01 5.29699945e+00 6.55814588e-01
-9.24219549e-01 1.72420546e-01 5.12706161e-01 4.72107064e-03
-1.45917729e-01 -1.77080095e-01 -1.11577153e+00 8.43120635e-01
6.86227560e-01 -2.79539712e-02 4.08403754e-01 8.62216949e-01
3.44401598e-02 3.07949781e-01 -1.23634958e+00 1.46662259e+00
4.74269897e-01 -9.39642727e-01 2.43711293e-01 2.77548641e-01
8.82598221e-01 -4.38671559e-01 2.34859869e-01 3.19719315e-01
3.27796698e-01 -8.97661448e-01 7.36576140e-01 6.59314752e-01
1.01084852e+00 -8.75159860e-01 8.08081150e-01 -3.26543418e-03
-1.48638356e+00 -5.77453785e-02 -3.30463678e-01 3.58736008e-01
2.16658592e-01 4.69406009e-01 -4.94158417e-01 7.24464893e-01
7.36951113e-01 9.83639121e-01 -9.04826820e-01 9.84647274e-01
-2.62740701e-01 1.89276293e-01 1.14168338e-01 5.11566937e-01
-2.31108576e-01 -2.69279122e-01 3.22467685e-01 1.34366798e+00
3.66374046e-01 -3.23717147e-01 2.03102961e-01 8.06737959e-01
-2.25857243e-01 -1.17290668e-01 -3.29446882e-01 6.87402636e-02
3.45736355e-01 1.32314456e+00 -3.25808883e-01 -4.74748045e-01
-6.47856951e-01 1.75462472e+00 4.98543888e-01 4.72794712e-01
-1.01499808e+00 -2.90120542e-01 8.73210371e-01 -1.62853077e-02
4.50761944e-01 -8.02445188e-02 1.38249502e-01 -1.47220492e+00
1.74349651e-01 -9.56615031e-01 3.88678074e-01 -4.82024640e-01
-1.71362889e+00 6.76098883e-01 -4.25050817e-02 -1.28830218e+00
-3.86042625e-01 -5.18771529e-01 -5.01066923e-01 1.15870857e+00
-1.57407093e+00 -1.67448509e+00 -7.91239679e-01 7.08749354e-01
5.93981326e-01 -3.48199934e-01 6.66327000e-01 6.15527213e-01
-1.01016247e+00 1.23684108e+00 2.41362005e-01 4.66754198e-01
1.14689875e+00 -1.28842080e+00 5.58065593e-01 1.16031277e+00
1.31443709e-01 6.76534593e-01 3.47783893e-01 -5.77767491e-01
-1.08719552e+00 -1.45109570e+00 7.22188771e-01 -3.91093075e-01
1.76628783e-01 -5.12774467e-01 -8.29165101e-01 6.53003633e-01
7.85004571e-02 2.60833085e-01 8.95189583e-01 -1.25680581e-01
-7.27773368e-01 -3.98140490e-01 -1.04048228e+00 4.94240344e-01
1.19368708e+00 -7.24584639e-01 -4.96886075e-01 -8.26537162e-02
2.57108539e-01 -8.87760967e-02 -6.61222160e-01 1.27401710e-01
7.39562690e-01 -7.96738625e-01 1.19234967e+00 -3.51351470e-01
3.09415638e-01 -5.96614659e-01 -5.74340820e-02 -1.41611707e+00
-6.76910281e-01 -2.26268455e-01 -1.51852727e-01 1.74078619e+00
1.74895755e-03 -5.27731955e-01 6.22498572e-01 7.17578590e-01
5.60511872e-02 -2.90541977e-01 -6.76349580e-01 -7.85449922e-01
-1.70712322e-01 5.64318448e-02 7.03216195e-01 6.88497186e-01
-4.01330084e-01 4.27689075e-01 -7.33217061e-01 1.67625427e-01
9.50143874e-01 1.82957262e-01 1.12791800e+00 -1.06723404e+00
-6.19057178e-01 -1.85637757e-01 -4.22584236e-01 -1.42313290e+00
1.46101668e-01 -9.68796492e-01 7.67440349e-02 -1.50243771e+00
7.25498796e-01 -3.96299511e-01 -2.49720767e-01 4.66262311e-01
-6.71890199e-01 5.47960699e-01 3.49879146e-01 3.70823443e-01
-5.54341912e-01 8.84604752e-01 1.06843877e+00 -3.91866267e-01
-6.53549284e-02 -2.07848042e-01 -8.47115695e-01 4.76939976e-01
5.84909499e-01 -2.83768088e-01 -4.27099705e-01 -6.11714125e-01
-2.30563402e-01 -5.44844627e-01 6.59658074e-01 -1.13955402e+00
1.97027862e-01 2.21825629e-01 8.30018044e-01 -3.45050484e-01
3.19127619e-01 -7.31412768e-01 2.71312475e-01 2.07940713e-01
-2.86600679e-01 1.03694484e-01 2.21506990e-02 7.47829735e-01
-2.57095844e-01 -2.30472609e-01 1.00484705e+00 6.31083325e-02
-6.04445994e-01 6.39390051e-01 1.53173774e-01 -5.54365441e-02
9.88761425e-01 -2.50161022e-01 -1.02331102e-01 -2.44769678e-01
-5.87194026e-01 -6.46325434e-03 7.58696556e-01 7.68162906e-01
6.12656236e-01 -1.74030757e+00 -7.83124983e-01 4.14360791e-01
3.78369361e-01 -1.77588060e-01 6.50545120e-01 4.50229049e-01
3.14289108e-02 1.78344816e-01 -3.12920213e-01 -4.81095374e-01
-1.26752245e+00 7.32040346e-01 3.80431235e-01 -3.55786681e-01
-5.64135671e-01 8.66485119e-01 5.71723998e-01 -2.96653539e-01
1.63865164e-01 2.64261693e-01 -2.09080681e-01 9.63420123e-02
8.68391514e-01 3.11045617e-01 -1.49680644e-01 -1.09905410e+00
-3.49450022e-01 7.58935511e-01 -3.24473768e-01 -1.27002433e-01
1.21774065e+00 -3.17365170e-01 2.18829677e-01 1.38220236e-01
1.38126230e+00 -6.68463036e-02 -1.58914006e+00 -4.07076180e-01
-4.43981439e-01 -7.30126321e-01 -2.09248200e-01 -6.47524357e-01
-1.17960835e+00 8.90809417e-01 1.01895726e+00 -1.96132839e-01
1.09386611e+00 -8.94968864e-03 9.91055489e-01 -1.14432961e-01
2.68728942e-01 -1.10744572e+00 4.53410745e-01 1.22946158e-01
7.90986240e-01 -1.32209444e+00 -1.27990082e-01 -1.54508635e-01
-6.84983015e-01 9.82651114e-01 8.03865731e-01 -1.70639038e-01
4.01039809e-01 2.76153255e-03 -1.30720824e-01 2.77076542e-01
-2.86939770e-01 -3.26584607e-01 4.53356028e-01 9.49327171e-01
4.04656142e-01 1.26840370e-02 1.68783799e-01 7.87588179e-01
4.40037921e-02 -8.30313005e-03 -5.15207648e-02 4.69544202e-01
-1.74237844e-02 -1.59526229e+00 -4.55870748e-01 1.31960034e-01
-1.75729021e-01 1.21126138e-01 -3.86866570e-01 4.29786980e-01
4.76635277e-01 8.35823476e-01 2.18465015e-01 -6.48002267e-01
4.71687794e-01 -7.88654089e-02 5.41446269e-01 -6.18396759e-01
-2.66261280e-01 -1.14705777e-02 -1.75312623e-01 -3.63631248e-01
-4.68900055e-01 -8.72883976e-01 -7.49280870e-01 -1.29583716e-01
-1.52503043e-01 -2.27331501e-02 3.72416317e-01 8.47941041e-01
5.07205248e-01 5.05909383e-01 8.04624796e-01 -1.00515091e+00
-4.24023479e-01 -9.83137071e-01 -3.67798090e-01 9.04900908e-01
3.58633578e-01 -7.53622591e-01 -2.08671659e-01 4.31243747e-01] | [14.706496238708496, 1.004082441329956] |
fd59246d-ae84-42a4-b8f1-0c85215d8171 | image-coding-via-perceptually-inspired-graph | 2303.01674 | null | https://arxiv.org/abs/2303.01674v1 | https://arxiv.org/pdf/2303.01674v1.pdf | Image Coding via Perceptually Inspired Graph Learning | Most codec designs rely on the mean squared error (MSE) as a fidelity metric in rate-distortion optimization, which allows to choose the optimal parameters in the transform domain but may fail to reflect perceptual quality. Alternative distortion metrics, such as the structural similarity index (SSIM), can be computed only pixel-wise, so they cannot be used directly for transform-domain bit allocation. Recently, the irregularity-aware graph Fourier transform (IAGFT) emerged as a means to include pixel-wise perceptual information in the transform design. This paper extends this idea by also learning a graph (and corresponding transform) for sets of blocks that share similar perceptual characteristics and are observed to differ statistically, leading to different learned graphs. We demonstrate the effectiveness of our method with both SSIM- and saliency-based criteria. We also propose a framework to derive separable transforms, including separable IAGFTs. An empirical evaluation based on the 5th CLIC dataset shows that our approach achieves improvements in terms of MS-SSIM with respect to existing methods. | ['Antonio Ortega', 'Eduardo Pavez', 'Samuel Fernández-Menduiña'] | 2023-03-03 | null | null | null | null | ['ms-ssim'] | ['computer-vision'] | [ 5.88927746e-01 -2.94046968e-01 -2.73769259e-01 -4.02888447e-01
-7.27741122e-01 -4.92645115e-01 3.71986449e-01 3.18879455e-01
-2.30954230e-01 5.03502011e-01 2.30934322e-01 -3.14560801e-01
-3.09885353e-01 -6.97451770e-01 -6.38615310e-01 -5.54764211e-01
-1.51403844e-01 -2.34326854e-01 5.40157437e-01 -1.42829478e-01
4.83361989e-01 4.46558416e-01 -1.44114625e+00 2.02763930e-01
1.24968207e+00 1.21497273e+00 5.41267991e-01 5.68826079e-01
1.78919196e-01 4.46364343e-01 -5.23332298e-01 -4.36904609e-01
4.60563123e-01 -7.25859284e-01 -4.19202179e-01 -8.91065411e-03
4.48995531e-01 -2.01758027e-01 -3.83910835e-01 1.33292270e+00
6.06446147e-01 -4.96247597e-02 7.25054741e-01 -1.01960588e+00
-5.73750675e-01 4.02998894e-01 -4.31870013e-01 3.25289190e-01
2.28992358e-01 -2.01215521e-02 1.19878864e+00 -9.12903488e-01
5.20898700e-01 1.10647202e+00 4.98521775e-01 8.87176692e-02
-1.39085996e+00 -4.70114738e-01 -1.98983997e-01 7.49288023e-01
-1.61667585e+00 -5.36424518e-01 9.44511533e-01 -1.91038340e-01
7.70145535e-01 4.06880140e-01 6.28219247e-01 5.20026088e-01
2.89690554e-01 6.07704639e-01 1.05208242e+00 -6.24700427e-01
4.94109899e-01 -7.93067291e-02 -6.03298068e-01 6.28539801e-01
3.10523123e-01 1.48931861e-01 -6.50468349e-01 1.57430008e-01
8.70910525e-01 -4.28333282e-01 -8.39095473e-01 -7.45426953e-01
-1.21997392e+00 5.95413685e-01 7.03064978e-01 3.94917727e-01
-2.41191328e-01 2.24786341e-01 1.90856069e-01 4.09093022e-01
5.27733684e-01 5.13234317e-01 -3.05313617e-02 -1.90949962e-01
-1.35318553e+00 -5.47758043e-02 5.33561110e-01 8.59736264e-01
8.05632532e-01 5.07152490e-02 -4.77916211e-01 8.53937805e-01
2.81135172e-01 5.09396434e-01 3.10891122e-01 -8.70827794e-01
5.28920889e-01 3.99222642e-01 -9.98974219e-02 -1.31784260e+00
1.39003620e-02 -6.03002846e-01 -7.75668323e-01 3.26707661e-01
2.04612300e-01 2.85042107e-01 -7.15696812e-01 1.66468143e+00
6.44943416e-02 8.42972621e-02 -1.57105803e-01 1.19296038e+00
3.52612764e-01 5.12257993e-01 -3.82137656e-01 -3.20174605e-01
1.05005765e+00 -7.36802757e-01 -6.72446609e-01 -8.16362575e-02
4.80843693e-01 -8.95460427e-01 1.06338155e+00 5.45348883e-01
-1.33005178e+00 -5.33351421e-01 -1.57066560e+00 -2.02280674e-02
-1.12896666e-01 1.74716115e-01 1.94721222e-01 9.63454783e-01
-1.39133406e+00 7.76137710e-01 -6.06123090e-01 -1.83094919e-01
4.92296338e-01 1.63713336e-01 9.35559794e-02 -8.91316459e-02
-8.91209245e-01 7.79112458e-01 3.40304673e-01 -2.18997523e-01
-8.30048621e-01 -7.29875684e-01 -6.74852073e-01 3.31400663e-01
2.98196077e-01 -4.25527424e-01 6.85493231e-01 -9.88780499e-01
-1.46493447e+00 5.54746032e-01 -1.95132747e-01 -5.83596647e-01
5.35066247e-01 1.90172553e-01 -4.25913841e-01 5.55305302e-01
-2.26709798e-01 6.34920061e-01 1.17572212e+00 -1.18132389e+00
-6.02577865e-01 -7.94308782e-02 2.21683964e-01 2.12730870e-01
-5.75323641e-01 -6.28668740e-02 -4.78368074e-01 -1.11022782e+00
4.39105421e-01 -6.47246301e-01 5.13369180e-02 4.78147864e-01
-3.19256634e-01 3.41805130e-01 7.78039157e-01 -7.78491616e-01
1.65525007e+00 -2.13575649e+00 3.45581263e-01 3.05526197e-01
3.29362124e-01 4.21099752e-01 -1.63303301e-01 1.34655252e-01
-2.40384918e-02 2.04890803e-01 -4.79207605e-01 -2.03650966e-01
-7.04651000e-03 -2.20214278e-02 -5.14450893e-02 5.05099118e-01
2.30484620e-01 8.61909688e-01 -8.38294983e-01 -5.11423230e-01
2.04648003e-01 4.32665974e-01 -7.11041927e-01 -4.42358144e-02
4.24097441e-02 2.19745025e-01 -2.48411894e-01 4.30304468e-01
8.72493684e-01 -1.66978493e-01 1.28874555e-02 -5.45324385e-01
-1.05877183e-01 3.41169000e-01 -9.41281617e-01 1.74288678e+00
-5.18001437e-01 9.86092925e-01 -1.40204310e-01 -1.06191909e+00
1.12092292e+00 1.77896142e-01 3.57072681e-01 -9.55272079e-01
-3.31616439e-02 5.24357915e-01 9.79170203e-02 -9.26569179e-02
5.40148258e-01 1.30348444e-01 2.20412523e-01 3.34384352e-01
-2.79398356e-03 -2.05604270e-01 2.14187726e-01 7.72893131e-02
1.09028780e+00 -1.02918029e-01 2.93052405e-01 -5.25446296e-01
7.25172818e-01 -5.16046584e-01 3.83458585e-01 4.37782049e-01
-1.13623478e-01 9.73436415e-01 4.00934100e-01 -7.77219832e-02
-1.35443676e+00 -1.28322041e+00 -1.42142102e-01 5.68603098e-01
5.92818737e-01 -4.72763538e-01 -8.24541926e-01 -6.84785187e-01
-3.33038956e-01 6.44313037e-01 -1.07156612e-01 -4.02471513e-01
-5.21040916e-01 -3.51413012e-01 4.20747578e-01 1.52167782e-01
6.35833025e-01 -3.79641294e-01 -9.59541023e-01 2.20683873e-01
-3.32647890e-01 -1.17584062e+00 -8.40833068e-01 1.03156164e-01
-9.69642997e-01 -8.09984803e-01 -1.02901626e+00 -5.29292166e-01
6.96752071e-01 4.34555560e-01 9.94645834e-01 9.66862440e-02
-2.12627009e-01 2.59963691e-01 -6.86183453e-01 1.02298046e-02
-4.44729865e-01 -1.17489628e-01 -3.69954050e-01 2.76613712e-01
-2.61977941e-01 -9.09416437e-01 -9.65392947e-01 5.49661458e-01
-1.01298058e+00 2.01239392e-01 6.07804775e-01 7.04861045e-01
7.21144557e-01 1.55048773e-01 3.69458795e-01 -3.30135912e-01
6.11189485e-01 -6.72929129e-03 -5.32397747e-01 5.16278744e-01
-8.58183742e-01 4.63149041e-01 7.31081665e-01 -3.30917418e-01
-7.48746693e-01 -2.61600971e-01 8.28746408e-02 -5.53720534e-01
3.21493030e-01 3.26793522e-01 -2.65244573e-01 -5.56961596e-01
6.17732942e-01 4.69924033e-01 -2.63347208e-01 -3.18015665e-01
8.17736685e-02 3.69551122e-01 4.17158604e-01 -3.22906852e-01
9.14985061e-01 4.85955358e-01 2.34660149e-01 -6.79836988e-01
-1.97823167e-01 -2.98332512e-01 -4.07787830e-01 -3.37626338e-01
6.06664956e-01 -7.69275784e-01 -2.44760305e-01 1.98164195e-01
-1.14141285e+00 -1.02710970e-01 -3.66281897e-01 6.51548266e-01
-7.38756955e-01 6.65884197e-01 -1.98384076e-01 -5.85011840e-01
-1.19910151e-01 -1.35921049e+00 9.80180860e-01 1.06069185e-02
1.00147165e-01 -9.41938877e-01 -3.00168872e-01 -7.78389350e-02
6.54115200e-01 1.68854818e-01 1.06815255e+00 -1.29558891e-01
-7.68359065e-01 2.37708256e-01 -4.30474818e-01 5.12259066e-01
2.13254690e-01 -2.25576282e-01 -8.31985056e-01 -5.17585993e-01
2.20693592e-02 1.16191678e-01 8.78973782e-01 5.11242628e-01
1.23229635e+00 -2.33582571e-01 -7.53490850e-02 9.80332255e-01
1.67063022e+00 3.29909921e-01 1.03054917e+00 7.45702535e-02
6.02292538e-01 2.05086723e-01 6.85257673e-01 6.18006587e-01
1.80964798e-01 1.19483793e+00 5.32677054e-01 -1.30636096e-01
-6.11858726e-01 -3.50725412e-01 4.52268362e-01 6.95832312e-01
-1.24113701e-01 -4.55905706e-01 -6.35860562e-01 4.34751272e-01
-1.59597099e+00 -8.44133794e-01 -6.95125083e-04 2.44614291e+00
7.01923966e-01 2.04279631e-01 -5.05856015e-02 6.00391388e-01
8.19228768e-01 2.58079201e-01 -3.75010282e-01 -4.46202755e-01
-3.22397172e-01 4.99710232e-01 8.54913294e-01 4.56125319e-01
-8.17621469e-01 5.54457963e-01 6.06115437e+00 1.36387706e+00
-1.32074678e+00 1.13762029e-01 7.03450084e-01 1.22940846e-01
-6.68835998e-01 2.06491709e-01 -4.07458276e-01 5.32288730e-01
7.49814808e-01 -5.21037996e-01 6.13944352e-01 4.34604943e-01
5.14384151e-01 -1.59391642e-01 -1.09203804e+00 1.21660173e+00
1.79880038e-01 -1.22099936e+00 1.41000703e-01 1.86549813e-01
7.84805298e-01 -5.22513568e-01 4.33790445e-01 -2.68078953e-01
-2.87596464e-01 -8.62885296e-01 1.09068072e+00 4.08651114e-01
1.07240355e+00 -7.60623455e-01 3.07796478e-01 -5.76739796e-02
-1.52165651e+00 8.13016966e-02 -3.27197164e-01 3.70732158e-01
1.27065986e-01 7.81580389e-01 -8.37729275e-01 7.04289734e-01
5.94567060e-01 8.30754817e-01 -7.00063884e-01 1.43548787e+00
-2.05236748e-01 5.10822594e-01 -2.10823640e-01 3.91995255e-03
5.81098236e-02 -1.24470316e-01 8.09090257e-01 1.04684293e+00
9.83369648e-01 -2.15366006e-01 -2.30160967e-01 1.14521515e+00
1.23860322e-01 1.92763641e-01 -2.10497782e-01 1.70775533e-01
5.46356499e-01 7.87591875e-01 -8.68664265e-01 -7.53903687e-02
-2.46470734e-01 1.34191227e+00 -2.27536112e-01 3.57933372e-01
-1.01260984e+00 -4.87876028e-01 6.17543101e-01 3.24419409e-01
7.34993994e-01 -2.84994841e-01 -3.02678734e-01 -9.44369853e-01
3.60237092e-01 -8.60956132e-01 1.10151153e-02 -6.93407297e-01
-7.53635764e-01 5.40381372e-01 7.68570751e-02 -1.80653632e+00
-1.16492026e-01 -3.77930582e-01 -3.34704548e-01 7.77616680e-01
-1.80179369e+00 -7.71001577e-01 -4.17194664e-01 7.03305960e-01
5.62528431e-01 -1.46773420e-02 3.34045768e-01 5.62271595e-01
-7.25198537e-02 9.78165627e-01 1.93631977e-01 -2.94898242e-01
6.72835648e-01 -1.05364871e+00 2.77444899e-01 1.11056602e+00
2.88742602e-01 1.35235459e-01 8.28316033e-01 -3.78659576e-01
-1.34896111e+00 -1.07442832e+00 6.39315605e-01 2.72314399e-01
2.92800665e-01 -2.46043369e-01 -8.62128735e-01 2.03344468e-02
-1.69394594e-02 1.34483770e-01 3.67085010e-01 -5.55047095e-01
-4.52780783e-01 -3.02206159e-01 -1.27580559e+00 6.87341392e-01
1.33177602e+00 -5.07073760e-01 -2.27168612e-02 -1.33139580e-01
7.47813523e-01 -1.92181274e-01 -8.55354428e-01 3.96477461e-01
3.19725722e-01 -1.46645951e+00 1.10656607e+00 2.67320782e-01
4.56550360e-01 -6.25939846e-01 -4.43700969e-01 -1.43945324e+00
-4.20797378e-01 -6.27814233e-01 -8.58065411e-02 1.02576053e+00
4.19328332e-01 -4.54531252e-01 5.98184824e-01 -5.83508834e-02
-1.88487113e-01 -6.34654641e-01 -1.19800818e+00 -1.13441694e+00
-3.62116486e-01 -5.08649290e-01 6.77874744e-01 5.44213414e-01
-7.52162412e-02 -1.22697145e-01 -2.88109899e-01 1.61793381e-01
6.32062256e-01 -1.30551890e-01 3.28396767e-01 -9.44317579e-01
-5.44996738e-01 -7.02753723e-01 -1.01010776e+00 -1.42915404e+00
-2.63989031e-01 -9.35971081e-01 7.67415166e-02 -1.26539838e+00
-6.98349848e-02 -5.53407371e-01 -3.55812281e-01 6.05840096e-03
7.09763393e-02 4.12358612e-01 3.38742405e-01 1.83554247e-01
-4.47038770e-01 7.24482417e-01 1.33539033e+00 -2.80643046e-01
-9.22684371e-02 -2.76484042e-01 -4.48596776e-01 3.69182378e-01
6.31888509e-01 -3.92335355e-01 -6.98786259e-01 -4.18563128e-01
2.49440536e-01 -5.05666211e-02 3.90414119e-01 -1.54181659e+00
2.38910526e-01 -4.31201700e-03 2.40862399e-01 -3.44784945e-01
3.35483491e-01 -9.29211080e-01 1.60788000e-01 5.52023709e-01
-4.41292733e-01 -4.77433167e-02 3.69868204e-02 5.95440924e-01
-4.26026136e-01 -3.46768826e-01 8.81313801e-01 3.68556559e-01
-7.61065066e-01 1.73106819e-01 1.63614396e-02 -7.80052692e-02
1.02021027e+00 -6.41185045e-01 -9.92886946e-02 -6.56275928e-01
-2.11380899e-01 -3.36323351e-01 6.60973191e-01 2.49347895e-01
1.07186687e+00 -1.53352070e+00 -8.20687771e-01 2.08939925e-01
1.90358609e-01 -4.79780495e-01 5.58852144e-02 8.91697884e-01
-6.62265122e-01 3.32677096e-01 -3.06479484e-01 -7.83171475e-01
-1.39738071e+00 4.85009909e-01 2.29333684e-01 -3.26702058e-01
-4.18941617e-01 7.39810288e-01 2.60077924e-01 4.89342809e-01
4.63160127e-02 -5.24211466e-01 6.19248822e-02 -1.93685502e-01
4.54066843e-01 3.21881890e-01 2.89504975e-01 -6.71557546e-01
-2.74986982e-01 7.93828428e-01 2.68258393e-01 -3.47325027e-01
1.14787495e+00 -3.95120203e-01 1.96896613e-01 1.45176038e-01
1.49215627e+00 7.77285732e-03 -1.26908898e+00 -2.76677996e-01
-3.27764452e-02 -1.09581327e+00 3.73274326e-01 -5.70838511e-01
-1.32466722e+00 8.90453398e-01 9.04650390e-01 3.08159620e-01
1.80270731e+00 -2.37561747e-01 6.77829087e-01 -1.63424805e-01
6.35485172e-01 -1.04510820e+00 3.99777263e-01 1.86474875e-01
1.03523684e+00 -8.96629333e-01 7.48948380e-02 -6.44146264e-01
-3.07188034e-01 1.23973632e+00 -1.61201973e-02 2.52931528e-02
5.18560827e-01 2.93450534e-01 -4.74726588e-01 2.50065535e-01
-5.37238300e-01 -2.95076251e-01 6.61486745e-01 7.12404847e-01
4.03484464e-01 -1.16085745e-02 -7.11590052e-01 1.82595253e-02
-2.34778672e-01 -2.34447062e-01 4.01425093e-01 7.38682568e-01
-5.45822263e-01 -1.21946228e+00 -3.91986430e-01 2.69568324e-01
-2.65448987e-01 -2.88657814e-01 -1.08478814e-01 3.54766488e-01
2.15468392e-01 1.06633139e+00 2.28221100e-02 -6.91389859e-01
1.96284845e-01 -4.80175734e-01 7.37270117e-01 -2.31988266e-01
-2.10754663e-01 1.48773612e-02 -9.84775871e-02 -6.90591276e-01
-4.65896964e-01 -4.18630749e-01 -9.92961407e-01 -3.33800733e-01
-3.71374100e-01 -1.15687534e-01 6.85212970e-01 5.39486468e-01
4.03182983e-01 5.90000093e-01 9.99845266e-01 -7.85334289e-01
-3.50686550e-01 -3.97409916e-01 -5.06022930e-01 3.46361995e-01
3.65868777e-01 -5.52727878e-01 -4.63939518e-01 2.41624594e-01] | [11.44328498840332, -1.8280086517333984] |
c9300839-ff15-4005-99e8-c9f74be46f51 | complex-query-answering-with-neural-link-1 | 2011.03459 | null | https://arxiv.org/abs/2011.03459v4 | https://arxiv.org/pdf/2011.03459v4.pdf | Complex Query Answering with Neural Link Predictors | Neural link predictors are immensely useful for identifying missing edges in large scale Knowledge Graphs. However, it is still not clear how to use these models for answering more complex queries that arise in a number of domains, such as queries using logical conjunctions ($\land$), disjunctions ($\lor$) and existential quantifiers ($\exists$), while accounting for missing edges. In this work, we propose a framework for efficiently answering complex queries on incomplete Knowledge Graphs. We translate each query into an end-to-end differentiable objective, where the truth value of each atom is computed by a pre-trained neural link predictor. We then analyse two solutions to the optimisation problem, including gradient-based and combinatorial search. In our experiments, the proposed approach produces more accurate results than state-of-the-art methods -- black-box neural models trained on millions of generated queries -- without the need of training on a large and diverse set of complex queries. Using orders of magnitude less training data, we obtain relative improvements ranging from 8% up to 40% in Hits@3 across different knowledge graphs containing factual information. Finally, we demonstrate that it is possible to explain the outcome of our model in terms of the intermediate solutions identified for each of the complex query atoms. All our source code and datasets are available online, at https://github.com/uclnlp/cqd. | ['Michael Cochez', 'Pasquale Minervini', 'Daniel Daza', 'Erik Arakelyan'] | 2020-11-06 | complex-query-answering-with-neural-link | https://openreview.net/forum?id=Mos9F9kDwkz | https://openreview.net/pdf?id=Mos9F9kDwkz | iclr-2021-1 | ['complex-query-answering'] | ['knowledge-base'] | [ 1.63819984e-01 4.28004324e-01 -3.77695084e-01 -2.61808902e-01
-1.07548845e+00 -6.54727757e-01 2.24761397e-01 5.27230978e-01
-4.21948373e-01 1.09529650e+00 -1.59901217e-01 -6.09606445e-01
-5.15059531e-01 -1.20001781e+00 -1.31290829e+00 -2.82217935e-02
-2.22045407e-01 1.02909625e+00 3.18728864e-01 -3.90262872e-01
5.16059324e-02 2.98748016e-01 -1.51366484e+00 2.11194709e-01
9.33483362e-01 1.19269717e+00 -7.62792751e-02 4.94380176e-01
-2.35156327e-01 7.87693381e-01 -3.78730386e-01 -9.74132538e-01
1.51788503e-01 3.73413297e-03 -1.18821561e+00 -6.17490053e-01
5.75264513e-01 -1.67364236e-02 -3.03236097e-01 1.11470711e+00
3.76071244e-01 3.95892747e-02 3.69256854e-01 -1.07210720e+00
-5.09570241e-01 7.27518678e-01 -2.19882607e-01 2.39586115e-01
6.58726037e-01 -1.05942348e-02 1.82217360e+00 -8.11355710e-01
7.24599183e-01 1.11352086e+00 3.93438011e-01 1.25990018e-01
-1.23732293e+00 -4.86810207e-01 1.72516443e-02 6.40213430e-01
-1.58082867e+00 -3.73959392e-01 6.24441504e-01 -8.47891197e-02
1.44239140e+00 4.12424356e-01 3.67181867e-01 7.40356326e-01
-8.41447040e-02 7.65217662e-01 5.21358311e-01 -3.57800364e-01
2.34242976e-02 1.24196922e-02 2.52942353e-01 1.20335770e+00
5.08171916e-01 2.50956658e-02 -6.10889852e-01 -4.78903651e-01
3.24609071e-01 -3.24638963e-01 -3.34682047e-01 -3.23976845e-01
-1.05199492e+00 8.86006355e-01 6.86750293e-01 -5.28788427e-03
-4.44016486e-01 5.14621854e-01 2.49005675e-01 3.73383135e-01
2.36950397e-01 7.59129524e-01 -7.54758596e-01 -9.99892056e-02
-6.37949526e-01 3.47031832e-01 1.33607984e+00 1.04879570e+00
1.10616004e+00 -3.35656345e-01 -1.34091243e-01 6.64375365e-01
4.70364913e-02 2.53595412e-01 -1.13506056e-01 -8.16526949e-01
6.58247232e-01 8.18292677e-01 2.58740634e-01 -1.06047130e+00
-3.98223728e-01 -6.97295189e-01 -5.27954340e-01 -3.25344145e-01
5.16654670e-01 -1.37469605e-01 -7.32457519e-01 1.73232627e+00
3.90886575e-01 1.23710074e-01 7.63576431e-03 7.48770773e-01
8.20347726e-01 5.36876500e-01 -1.71549991e-01 -1.26176417e-01
1.43040156e+00 -9.22365546e-01 -4.84340012e-01 -2.96018362e-01
8.23667288e-01 -4.81921941e-01 1.16081810e+00 4.27945495e-01
-1.32153416e+00 -9.69241038e-02 -8.55364203e-01 -2.35689729e-01
-6.75031304e-01 -2.96236239e-02 8.17451954e-01 2.99305528e-01
-1.10451698e+00 4.44033325e-01 -4.69323695e-01 -1.99042767e-01
3.27552050e-01 7.64713228e-01 -1.58465028e-01 -3.86666179e-01
-1.59128714e+00 8.31474125e-01 6.96019530e-01 -2.38908431e-03
-7.35183120e-01 -6.40284002e-01 -7.96341121e-01 3.82896960e-01
1.21646833e+00 -9.45290923e-01 1.16442049e+00 -3.99153948e-01
-1.05570138e+00 6.79028690e-01 -4.95307624e-01 -4.97617751e-01
2.16464639e-01 -2.67202109e-01 -2.43649170e-01 -3.54916928e-03
1.18259706e-01 4.43919301e-01 3.57431263e-01 -1.11984766e+00
-6.34706974e-01 -3.40169400e-01 4.77074951e-01 9.10615996e-02
-7.94746950e-02 -1.65518001e-01 -1.02051342e+00 -2.32315823e-01
-8.71681869e-02 -7.89341748e-01 -1.38309494e-01 -2.19609961e-01
-8.40951145e-01 -5.16327739e-01 3.10998976e-01 -6.13202751e-01
1.45065153e+00 -1.67279673e+00 3.18175286e-01 6.01265192e-01
3.22207481e-01 8.88456628e-02 -7.18790442e-02 5.89097798e-01
1.73894897e-01 3.26006800e-01 -3.75621438e-01 4.51925269e-04
2.64518380e-01 3.97877306e-01 -1.97566018e-01 -1.05633341e-01
3.01682055e-01 1.34501970e+00 -9.14749920e-01 -4.45822388e-01
-3.45554709e-01 1.75718263e-01 -6.64752305e-01 -7.14048073e-02
-9.83386457e-01 -1.25387251e-01 -4.91709828e-01 9.39752042e-01
2.74523765e-01 -7.56235838e-01 3.72347146e-01 -2.54586071e-01
4.72130775e-01 3.90744090e-01 -1.19786036e+00 1.73211586e+00
-4.29267168e-01 3.20112735e-01 1.44449407e-02 -1.06463480e+00
6.60658538e-01 -8.09429884e-02 1.89944714e-01 -9.25118327e-01
4.94811870e-02 4.45316464e-01 -1.34629056e-01 -4.75966215e-01
5.31699181e-01 1.66851431e-01 -1.93981647e-01 2.47291327e-01
4.84798998e-02 -5.30002154e-02 6.89752221e-01 3.71564806e-01
1.46521890e+00 -2.10064858e-01 8.46919194e-02 1.62729934e-01
5.42380869e-01 3.16491872e-01 3.86884511e-01 9.52555895e-01
4.15549070e-01 1.36797369e-01 8.14205527e-01 -4.67837334e-01
-6.58698082e-01 -8.77291501e-01 2.14845598e-01 1.13789511e+00
1.77086011e-01 -5.82710683e-01 -3.96236837e-01 -6.38487935e-01
3.11100096e-01 8.53705645e-01 -2.68636137e-01 -7.71458074e-02
-7.15294540e-01 -6.75505579e-01 6.18073583e-01 3.10712337e-01
3.16907018e-01 -1.09583306e+00 -2.91544735e-01 2.89506465e-01
-3.80907446e-01 -1.20891500e+00 -1.85031325e-01 2.94201434e-01
-7.39716470e-01 -1.40534997e+00 -1.24996774e-01 -7.18743205e-01
6.33568645e-01 -2.13309988e-01 1.60402536e+00 4.25189972e-01
-5.01247644e-01 3.71171981e-01 -1.47960618e-01 -3.18125159e-01
-7.73970485e-02 3.83551210e-01 -3.74689370e-01 -2.22412765e-01
3.83809537e-01 -4.49814945e-01 -4.82188672e-01 1.91916853e-01
-9.08065856e-01 -3.33185464e-01 7.57296443e-01 8.92683923e-01
8.92920792e-01 2.74356902e-01 4.43420708e-01 -1.20501566e+00
9.97756481e-01 -5.85680783e-01 -9.47914720e-01 5.94312012e-01
-8.26716661e-01 5.19435346e-01 6.54541790e-01 -2.19946280e-02
-6.61787331e-01 7.73056224e-02 -5.23915626e-02 -2.82419175e-01
1.35136113e-01 1.10898423e+00 7.04236748e-03 -4.40072306e-02
8.36994171e-01 1.92435741e-01 -3.77216309e-01 -2.87890643e-01
6.21562421e-01 1.08102873e-01 6.60053074e-01 -1.07177532e+00
7.54998326e-01 2.48571068e-01 2.15821803e-01 -3.45415324e-01
-9.31750000e-01 -5.13517499e-01 -1.78947344e-01 1.91617742e-01
4.51757848e-01 -7.10499942e-01 -1.01892710e+00 -4.20288779e-02
-1.20941734e+00 -2.85522193e-01 -2.37330258e-01 1.68735132e-01
-3.26535076e-01 3.02714050e-01 -5.37126780e-01 -5.97098470e-01
-5.76652825e-01 -1.06088042e+00 9.25170362e-01 1.04761615e-01
-1.11256361e-01 -8.92580807e-01 -5.81352040e-02 6.03356004e-01
3.84412706e-01 2.21703544e-01 1.47487104e+00 -8.00968289e-01
-1.14266706e+00 -4.41835046e-01 -4.02250737e-01 4.47734073e-03
-2.33520478e-01 -2.12178037e-01 -5.62230110e-01 -1.04958065e-01
-5.45961320e-01 -6.67530060e-01 8.65528941e-01 8.86401236e-02
1.21573865e+00 -6.31137550e-01 -4.25053030e-01 4.86806929e-01
1.70437574e+00 -4.01830859e-02 4.96345401e-01 2.00062335e-01
5.50193191e-01 4.54598963e-01 4.80795056e-01 3.33292842e-01
6.25470340e-01 6.06589496e-01 9.02976513e-01 2.50912935e-01
1.82973146e-01 -3.59910876e-01 -1.11470103e-01 2.31084779e-01
-1.22319357e-02 -4.80267465e-01 -1.15035856e+00 8.08941841e-01
-1.98031747e+00 -7.78003991e-01 -1.48740813e-01 2.27468252e+00
1.11485219e+00 3.99347216e-01 -1.27176456e-02 -8.09161440e-02
4.49577928e-01 -8.18941742e-03 -7.90313900e-01 -2.30869696e-01
-3.37872244e-02 5.84370553e-01 5.95388353e-01 7.98177421e-01
-7.73404479e-01 1.12691891e+00 5.88519621e+00 8.88589680e-01
-8.24610233e-01 -9.70412344e-02 3.95022810e-01 -2.80201167e-01
-6.24717236e-01 1.92534730e-01 -7.53350139e-01 9.91863757e-02
8.42219114e-01 -2.20942453e-01 9.15012419e-01 6.74560845e-01
-2.83796459e-01 -2.44942889e-01 -1.24267411e+00 8.32708478e-01
-8.12704191e-02 -1.57615602e+00 8.46041143e-02 -3.78683433e-02
5.97685754e-01 1.79546326e-01 -1.02775522e-01 5.02589941e-01
6.70008183e-01 -1.13634884e+00 3.61520648e-01 5.07368147e-01
6.41318679e-01 -6.53376758e-01 6.68011487e-01 4.12615895e-01
-1.08873212e+00 -1.56925887e-01 -2.81022787e-01 8.97074267e-02
3.88898305e-03 7.20782697e-01 -1.13848066e+00 9.05143440e-01
6.48779571e-01 2.29102224e-01 -5.89309454e-01 1.01646829e+00
-4.05295163e-01 4.88195062e-01 -6.65646613e-01 -3.64228308e-01
1.34359315e-01 3.99452560e-02 3.87535006e-01 9.83640730e-01
1.89048797e-01 9.13756043e-02 -3.75380144e-02 1.07915711e+00
-6.17641747e-01 1.14654005e-01 -5.38150251e-01 -2.21204400e-01
5.46781898e-01 1.08112431e+00 -2.63935119e-01 -3.42792332e-01
-3.40384632e-01 7.38970160e-01 8.73842001e-01 6.68027997e-01
-8.28059077e-01 -5.54049432e-01 1.97560310e-01 1.08397603e-01
3.50296587e-01 -4.49471734e-02 -8.41842294e-02 -1.00348580e+00
5.26371479e-01 -8.77707958e-01 9.75061297e-01 -7.42546797e-01
-1.23044479e+00 4.40065563e-01 9.95327979e-02 -2.93178588e-01
-4.52549636e-01 -6.43628895e-01 -5.02903402e-01 8.62345278e-01
-1.80068755e+00 -8.27226937e-01 -2.85104990e-01 5.89623332e-01
8.85066763e-02 6.04123771e-02 7.21876621e-01 3.89375925e-01
-5.19851089e-01 6.59978449e-01 -1.93427876e-02 4.58794646e-02
3.03474396e-01 -1.32360029e+00 2.82725036e-01 5.23942947e-01
4.09056932e-01 5.56092680e-01 5.51297843e-01 -6.32279515e-01
-1.83685327e+00 -9.04949009e-01 1.06607902e+00 -3.53193134e-01
6.94418967e-01 -8.61844495e-02 -9.65106130e-01 7.37919807e-01
-5.98040223e-03 7.37375543e-02 4.76764053e-01 5.84982932e-01
-4.55544949e-01 -1.68581963e-01 -1.04753435e+00 4.22625393e-01
1.29306710e+00 -3.76032293e-01 -3.72678488e-01 6.63291335e-01
8.63259614e-01 -5.14539480e-01 -9.08136964e-01 5.07999063e-01
3.02423120e-01 -7.63943195e-01 1.10511243e+00 -9.14508283e-01
3.75016212e-01 -1.30802691e-01 -2.23259568e-01 -9.73959863e-01
-2.07686082e-01 -6.34423316e-01 -5.43306768e-01 5.77320039e-01
1.07826316e+00 -8.13719928e-01 1.05365419e+00 6.53315961e-01
3.00778132e-02 -1.32025480e+00 -1.04781079e+00 -5.67158937e-01
-2.64565498e-01 -5.00891268e-01 6.23117268e-01 8.23885381e-01
-1.39608039e-02 5.32259703e-01 -4.53069769e-02 4.01247978e-01
5.76252937e-01 3.32186013e-01 7.62758136e-01 -1.21280444e+00
-6.13046288e-01 -5.39713025e-01 -2.79531907e-02 -1.05556881e+00
2.21665606e-01 -1.13166320e+00 -2.95416772e-01 -2.01068068e+00
-5.53484187e-02 -4.90370512e-01 -2.74663597e-01 8.91510725e-01
-2.17441961e-01 -1.47552535e-01 -8.71431530e-02 2.38585919e-02
-8.96185875e-01 4.07047868e-01 1.08622432e+00 -4.03998256e-01
-5.25258929e-02 -1.09733753e-01 -8.26382160e-01 3.97196740e-01
6.79511607e-01 -5.24689138e-01 -4.46659863e-01 -5.84238410e-01
1.02737188e+00 2.68364012e-01 5.48430026e-01 -6.82325721e-01
6.63436651e-01 -1.17134690e-01 -1.25137642e-01 -5.66531479e-01
4.97770548e-01 -7.33446240e-01 1.62141368e-01 4.71661627e-01
-2.44843259e-01 1.21621035e-01 2.97176480e-01 7.26761222e-01
-3.29858482e-01 -2.12627873e-01 1.46876797e-01 -3.26254427e-01
-7.40589321e-01 2.10744396e-01 4.32672769e-01 4.25713897e-01
8.15710187e-01 1.63635805e-01 -6.26784265e-01 -5.45499682e-01
-6.65358245e-01 7.62757599e-01 6.43875599e-02 1.61612377e-01
7.00810492e-01 -1.00562811e+00 -6.54567599e-01 -9.84178632e-02
1.55178934e-01 4.39804971e-01 -4.35361266e-02 7.73318470e-01
-6.17500007e-01 7.43419945e-01 2.60428160e-01 -2.94670165e-01
-8.65709841e-01 5.64343452e-01 4.14953619e-01 -7.16001332e-01
-1.52444318e-02 1.08764350e+00 -2.98510611e-01 -5.28808296e-01
4.79259700e-01 -3.22511464e-01 2.43868932e-01 -7.74291903e-02
-2.68304087e-02 3.45778704e-01 1.44905418e-01 -1.53073192e-01
-4.04945672e-01 4.14051801e-01 -1.44613087e-01 1.33126944e-01
1.22823167e+00 3.92880827e-01 -2.91900337e-01 -1.00453727e-01
1.11522412e+00 5.27338125e-03 -5.63423216e-01 -5.69785297e-01
9.34661254e-02 -2.57873446e-01 -1.62192032e-01 -1.15044498e+00
-1.06091666e+00 5.04029870e-01 1.46446638e-02 3.65061522e-01
1.08909214e+00 3.40539515e-01 8.82873118e-01 1.16423857e+00
4.81063336e-01 -9.54257607e-01 -7.86639526e-02 5.01415253e-01
8.82362962e-01 -1.28965199e+00 1.07966028e-01 -6.14951193e-01
-2.76261926e-01 8.05787265e-01 5.94099164e-01 1.32760510e-01
3.84932458e-01 -1.68551788e-01 -4.14048344e-01 -7.85352528e-01
-1.03523731e+00 -3.89274180e-01 3.70206177e-01 1.75947040e-01
1.01354539e-01 -6.89357370e-02 -2.41140738e-01 4.99127895e-01
-3.29318672e-01 -3.59000303e-02 1.37872532e-01 8.87611389e-01
-3.23732018e-01 -1.22260141e+00 -1.64610863e-01 8.24504018e-01
-3.94274622e-01 -3.22247177e-01 -6.78013086e-01 8.33090484e-01
-8.33932012e-02 9.44354594e-01 -3.85379046e-01 -3.65694851e-01
5.97404718e-01 2.89604515e-01 3.64132553e-01 -5.81654727e-01
-4.58890945e-01 -3.31370652e-01 6.60111487e-01 -6.06462836e-01
-4.36185719e-03 -2.49905944e-01 -1.43911791e+00 -3.00074875e-01
-4.43569243e-01 2.70053327e-01 4.03872013e-01 6.92420959e-01
6.71900272e-01 4.17461395e-01 4.70285751e-02 -1.28065184e-01
-7.43656397e-01 -5.77308655e-01 -2.45862424e-01 3.16581517e-01
2.87107397e-02 -5.22664845e-01 -2.23817796e-01 -2.73270071e-01] | [9.413775444030762, 7.690106391906738] |
67b51468-7a1e-4b68-a1b6-ee5c92d1edaa | a-convolutional-neural-network-of-low | 2301.09861 | null | https://arxiv.org/abs/2301.09861v4 | https://arxiv.org/pdf/2301.09861v4.pdf | A convolutional neural network of low complexity for tumor anomaly detection | The automated detection of cancerous tumors has attracted interest mainly during the last decade, due to the necessity of early and efficient diagnosis that will lead to the most effective possible treatment of the impending risk. Several machine learning and artificial intelligence methodologies has been employed aiming to provide trustworthy helping tools that will contribute efficiently to this attempt. In this article, we present a low-complexity convolutional neural network architecture for tumor classification enhanced by a robust image augmentation methodology. The effectiveness of the presented deep learning model has been investigated based on 3 datasets containing brain, kidney and lung images, showing remarkable diagnostic efficiency with classification accuracies of 99.33%, 100% and 99.7% for the 3 datasets respectively. The impact of the augmentation preprocessing step has also been extensively examined using 4 evaluation measures. The proposed low-complexity scheme, in contrast to other models in the literature, renders our model quite robust to cases of overfitting that typically accompany small datasets frequently encountered in medical classification challenges. Finally, the model can be easily re-trained in case additional volume images are included, as its simplistic architecture does not impose a significant computational burden. | ['Dimitrios-Panagiotis Papageorgiou', 'Pantelis Dogoulis', 'Vasileios E. Papageorgiou'] | 2023-01-24 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 3.58911425e-01 5.57633162e-01 1.13348506e-01 -2.79485315e-01
-4.56143171e-01 -1.54743334e-02 8.36274683e-01 5.81136048e-01
-9.17601407e-01 9.06766295e-01 -2.75517642e-01 -3.87548476e-01
-3.62049758e-01 -6.06072426e-01 -3.44033182e-01 -8.53145599e-01
-2.64310271e-01 5.10890186e-01 3.48373540e-02 1.06596850e-01
2.59839714e-01 9.86623406e-01 -1.57249212e+00 -1.02422722e-01
1.04958701e+00 1.10745132e+00 3.58189076e-01 5.83919406e-01
-1.63913131e-01 8.76274824e-01 -3.26462239e-01 -3.12179238e-01
1.30947560e-01 -3.73112261e-02 -7.37697780e-01 1.92659840e-01
1.12061091e-01 -3.05985153e-01 3.91847156e-02 6.71878636e-01
3.44741434e-01 -2.22025037e-01 9.92866039e-01 -8.48682523e-01
-1.35212881e-03 3.84291559e-01 -3.66690725e-01 2.79775709e-01
-1.28468305e-01 1.00832187e-01 4.00738895e-01 -6.76030457e-01
5.24869025e-01 4.32674289e-01 6.58232033e-01 3.76757860e-01
-1.01752293e+00 -4.32593286e-01 -3.80308062e-01 1.84538051e-01
-1.23662424e+00 -2.04862088e-01 5.89024782e-01 -6.25898957e-01
7.23764479e-01 2.35338897e-01 6.56674683e-01 9.45206106e-01
3.91291708e-01 3.52405876e-01 1.23219943e+00 -5.02849758e-01
1.61078587e-01 6.91681504e-01 9.22313333e-02 7.99827337e-01
6.35119915e-01 -2.54713139e-03 2.08457544e-01 5.54529168e-02
4.80879605e-01 9.42090154e-02 -1.21685572e-01 -3.20818514e-01
-8.51497054e-01 7.79927552e-01 5.37175536e-01 9.94994521e-01
-6.08082950e-01 -5.00187613e-02 6.67951286e-01 -1.50020853e-01
3.66528213e-01 2.81015992e-01 -1.34974837e-01 1.75390109e-01
-9.61067617e-01 3.69526371e-02 5.33160090e-01 3.75044435e-01
1.03309572e-01 1.78330585e-01 6.95906132e-02 6.30127132e-01
2.21392617e-01 8.69959220e-02 7.37100124e-01 -1.14635728e-01
1.13268718e-01 8.53511214e-01 -3.63226496e-02 -8.69303167e-01
-9.13680136e-01 -9.93135810e-01 -1.23233438e+00 3.46539557e-01
4.04442966e-01 1.37839615e-01 -1.02727473e+00 1.34700549e+00
3.66117954e-01 2.14419439e-02 2.15568304e-01 4.62926507e-01
8.71936560e-01 1.69823065e-01 3.86519730e-01 -4.46502045e-02
1.49595296e+00 -4.37235653e-01 -6.71316564e-01 2.04884008e-01
7.40448236e-01 -6.84545279e-01 6.50403082e-01 3.91258568e-01
-9.31078136e-01 -2.52917469e-01 -1.18073022e+00 2.30908841e-01
-5.89782476e-01 4.77154404e-01 7.30349600e-01 8.98912609e-01
-9.41387892e-01 6.25579774e-01 -7.72771657e-01 -5.78539610e-01
7.93711603e-01 6.48859501e-01 -6.04419351e-01 9.28896889e-02
-8.58899951e-01 1.27665555e+00 6.77361667e-01 2.66002417e-01
-7.80383408e-01 -7.06909716e-01 -4.21781629e-01 9.63883400e-02
1.59126341e-01 -5.98743081e-01 9.11648870e-01 -8.25786471e-01
-1.15823662e+00 1.07937694e+00 1.59438446e-01 -8.78024042e-01
8.75522971e-01 1.87875986e-01 -2.16101184e-01 2.46159643e-01
-3.21965128e-01 4.33993429e-01 5.32448530e-01 -9.98823881e-01
-5.40549576e-01 -5.40514171e-01 -3.50881547e-01 -1.60271034e-03
-5.61581373e-01 -1.32939205e-01 -3.36728618e-02 -6.27340436e-01
-2.57742126e-02 -8.21376801e-01 -3.60818416e-01 -6.11055195e-02
-2.99897105e-01 4.83517498e-02 6.21561706e-01 -6.76131964e-01
7.59757578e-01 -1.82364082e+00 -1.13999359e-01 3.72249871e-01
2.11255088e-01 4.77049053e-01 3.21408987e-01 2.37449542e-01
-1.57882184e-01 4.40391600e-02 -4.99954164e-01 -2.78715104e-01
-3.68875057e-01 1.84895635e-01 3.41652811e-01 6.77027881e-01
3.89987141e-01 6.01072907e-01 -3.06848884e-01 -6.40403569e-01
5.41879535e-01 7.88019121e-01 -5.50690293e-02 2.11444050e-01
2.08538607e-01 4.32005495e-01 -3.02809000e-01 5.45993149e-01
5.17723083e-01 -1.36800617e-01 -1.71990037e-01 -1.15669876e-01
-1.37276083e-01 -2.38037392e-01 -8.04600477e-01 1.19476402e+00
-4.98486429e-01 5.08606136e-01 2.11142236e-03 -1.42481625e+00
1.02247441e+00 6.50882363e-01 6.89298272e-01 -7.76432455e-01
6.25044048e-01 3.43082458e-01 3.02704424e-01 -6.33817911e-01
1.57435924e-01 -3.15810621e-01 3.52959126e-01 3.43206227e-02
-3.86608616e-02 1.31923705e-01 1.27692774e-01 -7.41885230e-02
9.23647702e-01 -2.36446813e-01 6.00960732e-01 -3.47398281e-01
1.11394966e+00 3.15322578e-02 1.06884286e-01 4.64739829e-01
-4.02341068e-01 2.77432591e-01 3.08408380e-01 -5.63353956e-01
-1.01440632e+00 -5.74696541e-01 -5.74246585e-01 1.78114250e-01
-4.41933751e-01 3.46104324e-01 -6.92976296e-01 -4.67712730e-01
-2.19193682e-01 4.80970263e-01 -9.00166035e-01 -7.53196776e-02
-4.19997692e-01 -1.16000509e+00 5.26248038e-01 3.29172313e-01
6.07702672e-01 -1.10058773e+00 -9.93496358e-01 3.13488811e-01
1.69568211e-01 -1.24201012e+00 6.13103092e-01 3.13461125e-01
-1.36410999e+00 -1.29235840e+00 -9.88175035e-01 -6.69151545e-01
8.77447724e-01 -2.06813753e-01 8.71805072e-01 4.50206697e-01
-9.75757539e-01 2.21083760e-01 -3.40154827e-01 -7.16206133e-01
-6.84991658e-01 1.80167377e-01 -3.11172873e-01 5.97158596e-02
3.28133881e-01 -4.89228487e-01 -5.45055389e-01 -1.55328974e-01
-1.00326180e+00 6.27203807e-02 1.01977217e+00 9.31877553e-01
3.12634856e-01 -3.25824767e-02 8.26128900e-01 -1.19128168e+00
4.88905460e-01 -5.29931843e-01 -6.38104737e-01 -6.42273799e-02
-6.75938845e-01 -3.85128707e-02 7.63147712e-01 -3.76433283e-02
-9.67283309e-01 2.79077351e-01 -4.01304156e-01 -6.55704513e-02
-6.10562265e-01 4.86053467e-01 3.78271699e-01 -4.31502372e-01
5.81732094e-01 1.93855971e-01 1.53759569e-01 -2.56427199e-01
-1.87309831e-01 4.30323094e-01 4.41004664e-01 2.25622095e-02
8.16654384e-01 5.10513246e-01 6.55785561e-01 -9.35035050e-01
-3.53068829e-01 -4.17942315e-01 -8.44517469e-01 -3.10588449e-01
8.51862192e-01 -5.67250490e-01 -6.52257562e-01 4.51050043e-01
-8.49874139e-01 2.55500048e-01 -3.84413116e-02 6.12608373e-01
-3.22780252e-01 3.55568379e-01 -3.68755668e-01 -1.13415551e+00
-8.35916340e-01 -1.30331814e+00 4.36217070e-01 3.65628988e-01
-1.45596728e-01 -1.10807252e+00 -3.24085563e-01 3.05143267e-01
8.23468506e-01 8.24302673e-01 1.08305645e+00 -1.05923545e+00
-4.41128105e-01 -6.95615649e-01 -2.87961483e-01 2.10884154e-01
6.70001060e-02 1.10563882e-01 -1.22992289e+00 -1.54904246e-01
-3.01020462e-02 -3.02159101e-01 8.29205573e-01 4.31595981e-01
1.12335873e+00 5.02543077e-02 -3.93142909e-01 3.45785856e-01
1.61529207e+00 4.81134146e-01 5.34271002e-01 5.65000951e-01
3.14074874e-01 5.76031864e-01 3.64265978e-01 4.21453267e-01
-2.83888690e-02 3.06166083e-01 9.52846587e-01 -5.21490812e-01
5.37165217e-02 5.38352847e-01 -3.43812913e-01 4.92869288e-01
-3.64367157e-01 -4.12843470e-03 -1.00680566e+00 5.57107449e-01
-1.18596578e+00 -7.89228618e-01 -2.81352192e-01 2.26412749e+00
2.99160004e-01 3.54778886e-01 -1.69112496e-02 6.08150959e-01
3.90536875e-01 -1.70000210e-01 -1.56521708e-01 -5.56852698e-01
5.31048737e-02 4.55780387e-01 5.20543277e-01 1.62304774e-01
-1.27714968e+00 2.69585550e-01 5.54931021e+00 5.64039648e-01
-1.28235340e+00 5.18373447e-03 8.68188024e-01 1.17930382e-01
1.45370916e-01 -4.49583083e-01 -4.52715009e-01 3.52922469e-01
1.03754425e+00 7.50311464e-02 -1.61148623e-01 7.72378743e-01
3.99305940e-01 -5.45922995e-01 -7.70371795e-01 7.52045572e-01
8.58302861e-02 -1.26680970e+00 -1.69434190e-01 1.50389522e-01
2.62250960e-01 -8.36567655e-02 2.33896613e-01 1.07235543e-01
-3.92115891e-01 -1.23588312e+00 8.39401558e-02 5.51179051e-01
4.71033990e-01 -1.06664705e+00 1.29736865e+00 4.88675445e-01
-8.72860610e-01 -2.46156409e-01 -8.61930996e-02 1.45423070e-01
-8.02255794e-02 5.82450092e-01 -1.59613037e+00 7.74377227e-01
5.35426676e-01 2.47192606e-01 -7.53771305e-01 1.27368510e+00
2.79206455e-01 5.06134987e-01 -3.25306445e-01 -2.51149684e-01
4.56362873e-01 -1.31403297e-01 3.17649901e-01 1.18141162e+00
3.90422374e-01 -1.45005882e-01 -2.49884874e-01 7.17042983e-01
1.64010331e-01 6.77206814e-01 -6.93817914e-01 -6.39502928e-02
7.17316493e-02 1.62185538e+00 -1.13456249e+00 -3.22423369e-01
-3.38884622e-01 6.66731238e-01 2.85749882e-01 -1.87367842e-01
-7.80923367e-01 -2.22720176e-01 2.00099335e-03 3.55812818e-01
8.30233693e-02 2.14045241e-01 -3.47880810e-01 -6.83841646e-01
-2.24320993e-01 -4.94089156e-01 3.66903543e-01 -3.16060036e-01
-7.73102045e-01 1.04122543e+00 -1.18254207e-01 -1.06301665e+00
-3.08910102e-01 -6.87134504e-01 -6.87181056e-01 8.30307782e-01
-1.66824257e+00 -1.10733235e+00 -4.89554822e-01 3.12077165e-01
3.97095442e-01 -2.90866554e-01 1.11167431e+00 4.80977684e-01
-6.99946880e-01 6.17798030e-01 -1.59569327e-02 -1.67701449e-02
3.72327000e-01 -1.08820152e+00 -2.96343982e-01 7.72618055e-01
-3.10717344e-01 2.61998296e-01 6.65377855e-01 -3.39542985e-01
-9.99043405e-01 -9.59171653e-01 8.74931216e-01 1.50753325e-03
4.18152243e-01 -9.91349667e-02 -9.92764056e-01 1.98617756e-01
2.52649605e-01 1.26870319e-01 7.73676932e-01 -2.58963078e-01
1.59969985e-01 -3.74023244e-02 -1.48182046e+00 3.94567728e-01
2.19371602e-01 1.66121144e-02 -4.02728707e-01 3.23907495e-01
-6.67190354e-04 -1.37186423e-01 -9.80446637e-01 7.73261130e-01
4.39064413e-01 -1.20822752e+00 8.43199611e-01 -4.22475576e-01
3.56828600e-01 6.12948239e-02 3.66192967e-01 -8.68334711e-01
-7.67931119e-02 7.29039982e-02 1.36171192e-01 9.99825120e-01
6.02019310e-01 -6.33372486e-01 1.06078029e+00 6.49625003e-01
4.80806455e-02 -1.29702246e+00 -1.05892098e+00 -3.98767143e-01
5.21971937e-03 -2.59801775e-01 1.70864925e-01 8.39500010e-01
-3.97976398e-01 -1.38636038e-01 -2.33612075e-01 1.44449500e-02
6.08935237e-01 -2.90047973e-01 5.39464295e-01 -1.59927797e+00
-2.03469433e-02 -6.27588689e-01 -7.71843672e-01 2.30043054e-01
-1.99700788e-01 -8.64373565e-01 -5.03810942e-01 -1.40299451e+00
3.60400677e-01 -5.26706874e-01 -4.24466878e-01 3.50746721e-01
-4.59477445e-03 3.82474095e-01 5.14744548e-03 3.31178084e-02
2.53656089e-01 3.25307280e-01 1.04559326e+00 -1.44180295e-03
3.73265855e-02 9.50805172e-02 -3.93800527e-01 7.91146278e-01
1.03855491e+00 -4.08457637e-01 -3.40617150e-01 1.96839377e-01
-2.50845164e-01 -1.81043223e-02 3.67872149e-01 -1.33365273e+00
1.17950320e-01 2.31261417e-01 7.72795141e-01 -6.02062166e-01
3.42572063e-01 -1.34626746e+00 3.04886341e-01 1.00934315e+00
-2.13013679e-01 -1.69339832e-02 3.14551562e-01 3.47276390e-01
-2.42590547e-01 -7.04060733e-01 1.15798306e+00 -2.80455202e-01
-4.56391543e-01 2.22340122e-01 -5.68916976e-01 -5.94320476e-01
1.63008809e+00 -5.04736662e-01 1.56126350e-01 -8.51096362e-02
-8.05321515e-01 -1.31625712e-01 8.51020813e-02 6.37563542e-02
4.29388344e-01 -8.82853329e-01 -6.53133988e-01 2.75501069e-02
1.18367396e-01 -9.31959823e-02 4.87427354e-01 1.30780339e+00
-9.36970651e-01 5.94916642e-01 -4.77312177e-01 -5.66875279e-01
-1.54436803e+00 6.08662426e-01 4.82222885e-01 -6.63482547e-01
-7.67656565e-01 6.13599539e-01 -1.27968013e-01 -1.38901640e-02
4.69005048e-01 -3.78558248e-01 -7.89283216e-01 6.75592571e-02
3.99410695e-01 4.00160044e-01 4.43346381e-01 -5.56784034e-01
-2.41073877e-01 2.75714308e-01 -4.75467354e-01 2.51578987e-01
1.33685207e+00 2.61413276e-01 8.50093290e-02 1.45798191e-01
1.05054820e+00 -3.18876565e-01 -5.93489051e-01 9.65132490e-02
3.07456613e-01 -1.50326476e-01 2.84732491e-01 -9.74808395e-01
-1.04435349e+00 9.44731176e-01 9.57896173e-01 1.40771568e-01
1.36225271e+00 -3.77462655e-01 4.18527603e-01 2.38569096e-01
1.95789970e-02 -8.88921380e-01 -3.00905615e-01 1.14398018e-01
7.60271311e-01 -1.47929084e+00 1.93985954e-01 -4.25745517e-01
-3.98741841e-01 1.33980119e+00 4.76224363e-01 -2.10188285e-01
4.85949606e-01 4.00055647e-01 5.00358790e-02 -2.24072069e-01
-5.50651848e-01 -2.06719950e-01 1.73666552e-01 4.85029131e-01
5.93299925e-01 -5.71421906e-02 -6.71861231e-01 4.36779231e-01
1.64650768e-01 2.01291442e-01 5.95166802e-01 9.54286397e-01
-4.94909674e-01 -8.77777278e-01 -4.80100244e-01 6.79322779e-01
-9.22340393e-01 1.66278571e-01 -2.86715597e-01 1.44972527e+00
3.46170515e-02 4.56922770e-01 -4.11251225e-02 1.78212989e-02
1.77893326e-01 1.28473058e-01 3.08340847e-01 -1.16788939e-01
-8.53129685e-01 4.51864786e-02 -7.87536800e-02 -1.89303175e-01
-6.51961744e-01 -3.87320608e-01 -1.28520358e+00 6.28124624e-02
-3.71477753e-01 2.40504220e-02 1.05029428e+00 9.92054462e-01
4.83993217e-02 7.84122288e-01 3.98171216e-01 -6.57019675e-01
-4.61266965e-01 -1.16706848e+00 -4.22368914e-01 4.05648410e-01
1.55416518e-01 -8.38563025e-01 -3.00354004e-01 -1.25120014e-01] | [15.18183422088623, -2.688488721847534] |
bb4de3ca-01d3-42e3-95ac-dabfe1db554d | evaluating-inter-bilingual-semantic-parsing | 2304.13005 | null | https://arxiv.org/abs/2304.13005v2 | https://arxiv.org/pdf/2304.13005v2.pdf | Evaluating Inter-Bilingual Semantic Parsing for Indian Languages | Despite significant progress in Natural Language Generation for Indian languages (IndicNLP), there is a lack of datasets around complex structured tasks such as semantic parsing. One reason for this imminent gap is the complexity of the logical form, which makes English to multilingual translation difficult. The process involves alignment of logical forms, intents and slots with translated unstructured utterance. To address this, we propose an Inter-bilingual Seq2seq Semantic parsing dataset IE-SEMPARSE for 11 distinct Indian languages. We highlight the proposed task's practicality, and evaluate existing multilingual seq2seq models across several train-test strategies. Our experiment reveals a high correlation across performance of original multilingual semantic parsing datasets (such as mTOP, multilingual TOP and multiATIS++) and our proposed IE-SEMPARSE suite. | ['Anoop Kunchukuttan', 'Vivek Gupta', 'Divyanshu Aggarwal'] | 2023-04-25 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 2.42320850e-01 3.34943026e-01 -1.00013480e-01 -5.97509444e-01
-1.37849915e+00 -1.07060635e+00 3.13854754e-01 -7.69732371e-02
-3.73638213e-01 1.34511983e+00 4.12906796e-01 -6.59700572e-01
2.15988889e-01 -8.16420853e-01 -9.31367457e-01 4.99970913e-02
4.21186358e-01 9.48229969e-01 1.02363609e-01 -6.48425639e-01
1.00333847e-01 -2.11919680e-01 -8.71785581e-01 6.67881548e-01
1.34244657e+00 5.75887263e-01 5.81529438e-01 4.84555960e-01
-8.16845715e-01 6.97194695e-01 -6.01366520e-01 -7.97105968e-01
2.86686271e-01 -7.96999097e-01 -1.29355371e+00 -6.00752532e-01
4.21357393e-01 1.47798806e-01 3.39381516e-01 1.27711463e+00
4.69842881e-01 -2.38519549e-01 1.06969044e-01 -1.06308007e+00
-7.28884757e-01 1.20385981e+00 -2.04826772e-01 8.08553770e-02
8.05331767e-01 -1.42229542e-01 1.25977564e+00 -6.66566670e-01
1.32958210e+00 1.55433846e+00 8.21273506e-01 7.82950222e-01
-8.87733519e-01 -7.01271355e-01 -4.32203487e-02 1.99181251e-02
-9.93805408e-01 -3.90159518e-01 3.00251544e-01 -1.12595633e-01
1.40147889e+00 -1.21017642e-01 1.04253553e-01 1.26882076e+00
2.77917653e-01 7.89988697e-01 1.58620024e+00 -5.16053319e-01
1.77077249e-01 7.01133460e-02 6.41930774e-02 6.74232960e-01
-4.83276695e-02 -7.96385109e-02 -6.46568179e-01 2.18488425e-01
4.38421965e-01 -7.82152832e-01 3.71366531e-01 2.81617671e-01
-1.21766210e+00 7.59949267e-01 -1.49377331e-01 3.47000867e-01
-8.00703689e-02 -9.66820717e-02 6.49678349e-01 6.13693357e-01
4.21715885e-01 4.71034884e-01 -1.06418490e+00 -5.37064016e-01
-4.55237299e-01 3.37436527e-01 9.77508664e-01 1.35815465e+00
7.33083844e-01 3.35873328e-02 1.52967453e-01 1.06891048e+00
1.12954177e-01 7.49195576e-01 4.86870259e-01 -8.98726583e-01
1.24224722e+00 3.44733417e-01 -2.80660927e-01 -5.42784214e-01
-2.96482027e-01 1.37294540e-02 -2.55752027e-01 -3.58749896e-01
5.53010643e-01 -4.64003205e-01 -5.33776224e-01 2.02576280e+00
2.32470304e-01 -2.85398602e-01 7.62617648e-01 5.66757083e-01
7.98866510e-01 8.30122411e-01 6.96192324e-01 -2.02704910e-02
1.63073480e+00 -9.90087748e-01 -6.40296221e-01 -6.48767710e-01
6.85801864e-01 -9.54249561e-01 1.08955181e+00 1.55403912e-01
-1.31795192e+00 -4.79126602e-01 -7.73899257e-01 -3.21726829e-01
-5.95567822e-01 -1.12357706e-01 7.75635123e-01 5.44221461e-01
-1.13991463e+00 3.27496886e-01 -5.36444247e-01 -7.36914873e-01
5.27169593e-02 1.20530821e-01 -6.21056676e-01 -1.73742965e-01
-1.52331567e+00 9.94513035e-01 8.82795751e-01 -5.51642440e-02
-5.73780239e-01 -6.16069078e-01 -9.79052722e-01 -4.03612018e-01
1.93235621e-01 -5.94115555e-01 1.42357945e+00 -1.17722130e+00
-1.37458992e+00 1.10554385e+00 -2.39947706e-01 -4.40047592e-01
4.45919961e-01 -4.96069252e-01 -5.76661825e-01 -3.43781486e-02
7.67399132e-01 1.14781249e+00 1.63236946e-01 -8.98709297e-01
-8.50565314e-01 -5.44251502e-01 -1.13680102e-01 4.23363149e-01
3.91731590e-01 6.78686082e-01 -2.07550555e-01 -6.48015738e-01
1.94274515e-01 -8.89739275e-01 -2.99300134e-01 -8.71175528e-01
-4.02645916e-01 -3.35748553e-01 3.35418880e-01 -1.21496201e+00
8.48701417e-01 -1.69590497e+00 -3.76979471e-03 -3.65104407e-01
-6.09703958e-01 -8.32929555e-03 -3.98270011e-01 5.78410387e-01
-1.30575791e-01 3.40349972e-01 -4.27610695e-01 -1.08659036e-01
3.13115604e-02 6.52461469e-01 -3.93994391e-01 -2.83824086e-01
5.77470124e-01 1.19751215e+00 -1.08183098e+00 -7.44033396e-01
-1.04771987e-01 5.33571914e-02 -6.63281262e-01 2.09372312e-01
-6.64436460e-01 6.49744570e-01 -7.15559006e-01 8.56366932e-01
6.08931363e-01 3.37659270e-01 6.52243495e-01 2.09917262e-01
-2.76293814e-01 1.01531374e+00 -8.04141104e-01 2.30179572e+00
-6.08859837e-01 2.02819347e-01 5.26867015e-03 -1.02182257e+00
8.05878580e-01 4.30124849e-01 8.93511474e-02 -1.31311059e+00
-2.05447346e-01 7.14355946e-01 -2.16663823e-01 -4.51651394e-01
7.26195157e-01 -4.71514940e-01 -7.38423586e-01 1.54188097e-01
4.70580429e-01 -1.56988963e-01 2.93816507e-01 3.74643244e-02
9.27951038e-01 9.22213137e-01 3.34289104e-01 -4.93833452e-01
3.89303416e-01 5.49710393e-01 8.55301917e-01 4.79784310e-01
-1.25989839e-01 3.88580322e-01 5.48192441e-01 -4.88912612e-01
-9.76948559e-01 -1.21484828e+00 -4.79210839e-02 1.22899401e+00
-1.70206547e-01 -2.33027861e-01 -1.10831749e+00 -1.00253940e+00
-4.76062804e-01 8.65430117e-01 -1.21983066e-01 6.61590099e-01
-1.23215425e+00 -8.25976908e-01 1.06578243e+00 4.10124362e-01
6.62372947e-01 -1.54303765e+00 -1.37734115e-01 7.41333783e-01
-6.74696922e-01 -1.85360503e+00 -4.90394328e-03 2.56241947e-01
-9.66022313e-01 -1.06143141e+00 -5.14337905e-02 -1.20559013e+00
3.32134277e-01 -3.89229596e-01 1.58296072e+00 -5.68515599e-01
6.20469972e-02 -4.40925397e-02 -5.02711475e-01 -4.20063555e-01
-8.68305504e-01 3.61830890e-01 -2.66258627e-01 -7.21337497e-01
5.10704637e-01 -4.07588661e-01 -8.96636173e-02 -3.21830921e-02
-6.65499389e-01 4.39591587e-01 4.96675283e-01 6.40575588e-01
6.38779819e-01 -4.04535979e-01 9.16881979e-01 -1.21574795e+00
4.67796177e-01 -5.55140078e-01 -6.78268731e-01 5.23428380e-01
-2.48056248e-01 1.99097186e-01 8.89568210e-01 3.36518168e-01
-1.54443324e+00 -2.89905481e-02 -6.40009940e-01 6.52842700e-01
-4.29633558e-01 4.88651305e-01 -4.57962543e-01 5.63034177e-01
3.17061692e-01 2.29280710e-01 -4.12281960e-01 -4.86111492e-01
5.12280285e-01 4.95988041e-01 8.24079514e-01 -1.22739172e+00
5.43308675e-01 -1.21336907e-01 -2.73190081e-01 -6.75336778e-01
-9.69392538e-01 -6.80599585e-02 -6.29381120e-01 3.19671184e-01
1.34482825e+00 -1.19933391e+00 -2.65308529e-01 2.94103891e-01
-1.62943542e+00 -4.72128868e-01 -4.07738611e-02 1.93952411e-01
-6.26365662e-01 1.60529971e-01 -8.85889292e-01 -4.07769114e-01
-5.52512467e-01 -1.06260967e+00 1.26351249e+00 1.29602537e-01
-4.33691233e-01 -1.29793978e+00 2.92507023e-01 7.93166041e-01
3.29755276e-01 2.26180032e-01 1.28430688e+00 -8.47758770e-01
-5.48322320e-01 3.97419512e-01 -3.35755974e-01 2.29173392e-01
-1.06846817e-01 -4.35932904e-01 -5.94980776e-01 1.68688044e-01
-3.92853647e-01 -6.81433439e-01 2.92061210e-01 4.07818221e-02
4.39042598e-01 -2.81101018e-01 5.16754203e-02 4.71126080e-01
1.70909166e+00 4.37008977e-01 6.58814549e-01 5.54579496e-01
5.36744416e-01 8.43528986e-01 8.05659831e-01 -1.37092605e-01
1.01816869e+00 4.82647091e-01 -3.63496281e-02 1.37042289e-03
-1.71513334e-01 -7.67956495e-01 8.14126134e-01 1.44945061e+00
3.62427145e-01 -2.03641146e-01 -1.09639895e+00 5.99211156e-01
-1.61698377e+00 -5.50962150e-01 -2.55433053e-01 1.71498752e+00
1.20674586e+00 -7.33811557e-02 -3.23461741e-01 -5.76066494e-01
4.34641868e-01 3.24226357e-02 -8.03044289e-02 -9.46532130e-01
-4.80899930e-01 8.35824192e-01 4.88329828e-01 7.61360049e-01
-8.47297907e-01 1.87145758e+00 6.44444561e+00 8.62178028e-01
-8.12063456e-01 3.82584453e-01 7.23168254e-01 6.29468977e-01
-4.80489820e-01 3.52632076e-01 -1.19361472e+00 4.16573197e-01
1.18669426e+00 4.29395698e-02 5.13886988e-01 6.12249434e-01
1.80314362e-01 -9.31023732e-02 -1.10364950e+00 5.23970366e-01
-8.11537504e-02 -1.14431739e+00 2.15777367e-01 -4.85336155e-01
8.39335918e-01 4.49721426e-01 -4.27403390e-01 7.12285459e-01
9.25836146e-01 -1.11805785e+00 6.74797356e-01 -5.43356389e-02
8.97927403e-01 -5.56545675e-01 7.08138347e-01 2.26049662e-01
-1.14102423e+00 3.44104767e-01 -3.55225891e-01 6.66208491e-02
4.73893940e-01 2.22600713e-01 -7.90365815e-01 8.90544593e-01
5.43708801e-01 5.66893041e-01 -4.67055529e-01 5.97501919e-02
-5.98321259e-01 6.45833611e-01 -3.70337442e-02 1.41422689e-01
4.75673020e-01 -5.41732490e-01 3.23874027e-01 1.48724163e+00
5.01420140e-01 5.99240698e-02 4.83542651e-01 7.60639191e-01
-7.19826147e-02 5.78939259e-01 -6.78954899e-01 -2.10421354e-01
4.81854558e-01 1.04647303e+00 -6.74786985e-01 -4.62201059e-01
-6.17852330e-01 1.17555368e+00 5.08736849e-01 1.47829235e-01
-7.31418729e-01 -9.93831679e-02 5.63929856e-01 -3.62885654e-01
-4.92341518e-02 -2.98893034e-01 -4.24612343e-01 -1.11078990e+00
1.10468037e-01 -1.38593209e+00 5.13761342e-01 -8.59952927e-01
-1.14890301e+00 7.53168344e-01 -1.27632558e-01 -6.29978240e-01
-4.33298677e-01 -7.01050818e-01 -3.60307544e-01 9.22926307e-01
-1.67609012e+00 -1.58540845e+00 1.88804835e-01 4.95044380e-01
1.07911384e+00 -2.90204585e-01 1.12393498e+00 3.74415636e-01
-3.99665982e-01 4.58947510e-01 -3.85679841e-01 3.30488950e-01
7.65072763e-01 -1.45003319e+00 9.79311824e-01 8.21952224e-01
9.77113657e-03 4.77319598e-01 6.49083734e-01 -8.47586811e-01
-1.46314812e+00 -1.20982444e+00 1.59995258e+00 -6.38130605e-01
8.75199735e-01 -4.17808622e-01 -5.95981598e-01 9.18869913e-01
6.00686014e-01 -4.82600689e-01 5.98577321e-01 9.09100547e-02
-4.09994155e-01 1.12581700e-01 -1.00320160e+00 6.02745771e-01
1.34548104e+00 -6.20056868e-01 -9.40186083e-01 3.61361623e-01
1.00008917e+00 -5.28264642e-01 -8.76309335e-01 3.74862045e-01
3.72715980e-01 -8.35953414e-01 7.28012741e-01 -8.25074732e-01
8.71017754e-01 -8.00131708e-02 -4.92938071e-01 -1.45537448e+00
3.24622214e-01 -6.82356000e-01 8.02098989e-01 1.56017768e+00
9.74472702e-01 -6.70496166e-01 6.63326979e-01 1.80690467e-01
-4.39241499e-01 -1.13601513e-01 -1.03686655e+00 -7.59140253e-01
5.91101706e-01 -6.99826360e-01 4.37593102e-01 1.28890073e+00
1.08158752e-01 6.87903941e-01 -9.38490927e-02 4.39425148e-02
4.76544768e-01 5.03604226e-02 6.32741213e-01 -8.84872913e-01
-4.65760469e-01 -8.35878775e-02 7.04019377e-03 -8.02727520e-01
5.91663659e-01 -1.27686751e+00 1.53308660e-01 -1.56279671e+00
2.18015045e-01 -6.95180655e-01 -7.37998113e-02 7.30864227e-01
-1.81616530e-01 9.71776247e-03 4.83216457e-02 -2.06543237e-01
-7.17428446e-01 2.93325543e-01 1.00764394e+00 2.48182654e-01
-8.77983868e-02 -7.78959453e-01 -7.14253783e-01 6.02466226e-01
8.13025534e-01 -6.55107617e-01 -1.68669701e-01 -1.00015914e+00
6.72743618e-01 5.85049152e-01 -4.19554301e-02 -8.32655013e-01
-7.61385411e-02 -2.76147366e-01 -9.01382975e-03 -4.97052044e-01
-3.13588679e-01 -4.18263704e-01 1.40931889e-01 2.81678379e-01
-2.70583749e-01 7.12770462e-01 4.00646776e-01 -1.11252084e-01
-5.75964272e-01 -1.00953221e-01 4.97990489e-01 -5.80916882e-01
-1.13796282e+00 -1.60708223e-02 -2.34977514e-01 8.37447822e-01
6.76085591e-01 -1.57725081e-01 -6.23082995e-01 8.33327472e-02
-5.56479454e-01 3.13771397e-01 2.82054067e-01 8.57969940e-01
1.37083143e-01 -1.10438824e+00 -8.50725472e-01 1.05010390e-01
1.05256878e-01 -6.44937973e-04 7.26724416e-02 6.74962401e-01
-9.21194553e-01 6.69146538e-01 -5.22319674e-01 -3.38249952e-01
-9.55887318e-01 1.05228275e-01 1.63086057e-01 -6.41597688e-01
-2.84051627e-01 6.54293656e-01 2.36619323e-01 -1.27274561e+00
-3.79924327e-01 -2.82830030e-01 5.32520041e-02 -2.43019968e-01
1.76174596e-01 1.39857963e-01 4.86891270e-02 -7.06165671e-01
-3.75104010e-01 4.96747494e-01 5.56880943e-02 -3.14837873e-01
1.31784403e+00 -1.06159076e-01 -5.19340217e-01 1.61314115e-01
1.05609131e+00 2.76355535e-01 -7.97668815e-01 -5.67583218e-02
7.21682131e-01 1.63271651e-01 -7.20268369e-01 -1.18389165e+00
-5.76864958e-01 6.70945227e-01 2.58067757e-01 -3.39477867e-01
8.41592371e-01 6.00481108e-02 1.25586033e+00 3.71023327e-01
8.45701098e-01 -1.59129393e+00 -2.99172014e-01 1.29793632e+00
5.08884013e-01 -1.38362348e+00 -7.11329043e-01 -5.88187933e-01
-7.21278250e-01 7.79222846e-01 7.15064704e-01 1.28946185e-01
8.93766433e-02 4.52771038e-01 5.32345772e-01 -1.12694658e-01
-7.72287786e-01 8.23151693e-03 -2.88478255e-01 6.71104550e-01
9.20977831e-01 3.31596851e-01 -7.12631762e-01 7.88394213e-01
-7.96126664e-01 -1.41569883e-01 1.84126303e-01 9.16159809e-01
-2.56054044e-01 -1.71815038e+00 -1.17560789e-01 -2.37069547e-01
-1.17408419e+00 -6.10104799e-01 -4.42957401e-01 9.74683940e-01
2.55617291e-01 9.10371125e-01 -1.30960988e-02 9.89625454e-02
3.01319689e-01 5.09344339e-01 4.54985827e-01 -7.64831126e-01
-7.27230728e-01 -9.06165913e-02 7.52638102e-01 -7.76522338e-01
-3.36671829e-01 -7.55428612e-01 -1.59297228e+00 -1.52964517e-02
3.24059993e-01 3.26197326e-01 9.24466074e-01 1.16096318e+00
2.39291027e-01 3.25089306e-01 2.20008120e-01 -9.26604122e-02
-3.22487503e-01 -8.34096193e-01 -1.08076610e-01 4.94786203e-01
-4.58686084e-01 5.72113842e-02 1.90233231e-01 1.91025630e-01] | [10.769173622131348, 9.661900520324707] |
a984734e-91a3-40c8-9a34-3f022a136491 | deep-learning-based-vulnerability-detection-1 | 2212.01254 | null | https://arxiv.org/abs/2212.01254v1 | https://arxiv.org/pdf/2212.01254v1.pdf | Deep-Learning-based Vulnerability Detection in Binary Executables | The identification of vulnerabilities is an important element in the software development life cycle to ensure the security of software. While vulnerability identification based on the source code is a well studied field, the identification of vulnerabilities on basis of a binary executable without the corresponding source code is more challenging. Recent research [1] has shown, how such detection can be achieved by deep learning methods. However, that particular approach is limited to the identification of only 4 types of vulnerabilities. Subsequently, we analyze to what extent we could cover the identification of a larger variety of vulnerabilities. Therefore, a supervised deep learning approach using recurrent neural networks for the application of vulnerability detection based on binary executables is used. The underlying basis is a dataset with 50,651 samples of vulnerable code in the form of a standardized LLVM Intermediate Representation. The vectorised features of a Word2Vec model are used to train different variations of three basic architectures of recurrent neural networks (GRU, LSTM, SRNN). A binary classification was established for detecting the presence of an arbitrary vulnerability, and a multi-class model was trained for the identification of the exact vulnerability, which achieved an out-of-sample accuracy of 88% and 77%, respectively. Differences in the detection of different vulnerabilities were also observed, with non-vulnerable samples being detected with a particularly high precision of over 98%. Thus, the methodology presented allows an accurate detection of 23 (compared to 4 [1]) vulnerabilities. | ['Dominik Binder', 'Andreas Schaad'] | 2022-11-25 | null | null | null | null | ['vulnerability-detection'] | ['miscellaneous'] | [ 1.31311059e-01 -4.58768010e-03 1.60759836e-01 -9.02138203e-02
-5.76881886e-01 -7.23166823e-01 4.18444484e-01 6.57329559e-01
-2.62322307e-01 2.16053560e-01 -2.37540573e-01 -9.29193199e-01
-1.65977761e-01 -1.06329060e+00 -3.11659753e-01 -5.94081819e-01
-2.10833147e-01 7.70295709e-02 1.21901274e-01 -3.84653121e-01
4.27285045e-01 5.71554601e-01 -1.61668277e+00 2.84810126e-01
3.44029933e-01 9.21685696e-01 -4.86706495e-02 8.52789283e-01
-5.05996644e-01 5.78419387e-01 -9.92847681e-01 -3.61483306e-01
1.88852668e-01 -9.20726731e-02 -7.88134992e-01 -5.24494588e-01
1.86803967e-01 -1.00319840e-01 -1.71529669e-02 1.34462214e+00
1.57456324e-01 -1.86027423e-01 5.80114424e-01 -8.70732605e-01
-4.62050706e-01 8.70738566e-01 -1.91831037e-01 3.24528456e-01
5.24031043e-01 -1.76413119e-01 9.87017393e-01 -5.96050024e-01
3.28097790e-01 8.75444770e-01 7.16529250e-01 4.66909707e-01
-1.07129133e+00 -5.04653871e-01 -3.24670136e-01 1.47521898e-01
-1.37504423e+00 1.70073640e-02 6.65011168e-01 -8.38118672e-01
1.29676819e+00 2.41533682e-01 3.12205970e-01 1.23686910e+00
4.00999904e-01 -9.01257247e-02 9.00812685e-01 -6.00303710e-01
2.09418267e-01 5.17483890e-01 6.72683299e-01 4.42136288e-01
4.81333315e-01 2.00131536e-01 3.44539762e-01 -4.27964926e-01
2.21989036e-01 8.37229639e-02 -1.91297568e-02 6.76523596e-02
-6.70725167e-01 1.14689231e+00 2.55768567e-01 1.16727161e+00
-2.52659172e-01 -2.27924168e-01 1.00234282e+00 6.71875715e-01
3.53135079e-01 5.92179894e-01 -3.87677401e-01 -2.74817586e-01
-9.45332408e-01 -4.59416732e-02 8.74529898e-01 3.42683494e-01
6.95484161e-01 4.89724904e-01 4.76041473e-02 4.90170032e-01
1.68457866e-01 3.14286053e-01 7.82656193e-01 -8.52438882e-02
4.04371053e-01 1.03683782e+00 -2.77317494e-01 -1.38756800e+00
-4.66115475e-01 -2.79812366e-01 -6.82230949e-01 7.05107689e-01
2.52368599e-01 -2.40077168e-01 -6.84756041e-01 1.44412482e+00
2.51093004e-02 -1.52326643e-01 3.39023560e-01 2.15738013e-01
6.76790237e-01 6.85240269e-01 2.65017077e-02 8.19954202e-02
1.12973142e+00 -2.64708757e-01 -5.01005650e-01 1.37780607e-01
7.17930734e-01 -6.25877321e-01 6.08363450e-01 3.55317295e-01
-4.67296839e-01 -6.18498385e-01 -1.18172109e+00 4.96097386e-01
-1.11737251e+00 8.97762477e-02 2.23883778e-01 1.38750672e+00
-1.29011011e+00 5.90947807e-01 -3.87651205e-01 -3.96486968e-01
-1.90830693e-01 4.06071603e-01 -5.17109692e-01 4.68208253e-01
-1.32912993e+00 1.01719487e+00 7.36710846e-01 5.53933159e-02
-8.36875856e-01 -4.17377651e-01 -9.25370395e-01 2.58568466e-01
4.52971496e-02 3.33398432e-01 6.74128711e-01 -1.04114246e+00
-1.14324784e+00 8.03882778e-01 2.39709482e-01 -5.87361157e-01
1.30237132e-01 1.12355262e-01 -7.90325522e-01 -1.23373065e-02
-6.61815107e-02 -2.04804227e-01 8.98093998e-01 -6.82970226e-01
-4.69751686e-01 -1.20682076e-01 2.72002459e-01 -7.13376939e-01
-9.77948427e-01 4.14402366e-01 3.19202960e-01 -4.74194795e-01
-3.21543545e-01 -7.00805128e-01 -6.50520027e-02 -8.62313986e-01
-5.09366095e-01 -1.14996731e-01 8.58337522e-01 -1.07660854e+00
1.62127388e+00 -2.26464128e+00 8.39611515e-02 5.57806373e-01
2.98499018e-01 7.19059527e-01 -1.82889953e-01 6.11930907e-01
-6.83516622e-01 3.80310357e-01 -4.76592302e-01 -1.04651162e-02
-2.27305163e-02 -2.76210964e-01 -6.99352980e-01 3.55705410e-01
1.64004371e-01 5.42583525e-01 -5.19510150e-01 6.44467920e-02
2.69513041e-01 7.03412831e-01 -5.21527119e-02 3.60378563e-01
1.22436099e-01 -1.62218511e-01 -3.92267436e-01 7.17387021e-01
3.74644816e-01 1.93249449e-01 1.49464458e-01 2.47296140e-01
-4.16052639e-01 -2.03994825e-03 -9.68974948e-01 9.08115506e-01
-6.92232549e-01 8.09993386e-01 -3.21471870e-01 -9.50197816e-01
1.49215937e+00 7.16111183e-01 1.95899710e-01 -4.29462105e-01
3.41635138e-01 5.17632961e-01 2.41620764e-01 -4.50202376e-01
3.45263720e-01 3.71802360e-01 -3.02351296e-01 5.70775688e-01
1.28583927e-02 2.57012069e-01 2.32666999e-01 -1.36330843e-01
1.34946990e+00 -2.57878691e-01 3.77760828e-01 -2.78020054e-01
9.74005878e-01 -2.26796418e-01 -5.19188121e-02 6.64631665e-01
-4.17486057e-02 2.99015701e-01 8.22942853e-01 -6.93004847e-01
-1.04047823e+00 -5.54639697e-01 -1.90323859e-01 6.43447697e-01
-5.21767139e-01 -3.44171494e-01 -1.04720461e+00 -8.44686210e-01
-1.69034258e-01 6.18265867e-01 -9.81481314e-01 -3.07702184e-01
-5.04272044e-01 -8.38706374e-01 8.75351071e-01 2.01187536e-01
1.07574537e-01 -1.40539098e+00 -1.07666934e+00 2.35369682e-01
3.81314814e-01 -8.52465034e-01 1.88707441e-01 3.54761958e-01
-5.60333431e-01 -1.29865396e+00 -5.07507443e-01 -3.13011885e-01
4.75144893e-01 -2.51131535e-01 8.71444166e-01 3.72039646e-01
-6.34121180e-01 2.65678674e-01 -7.09028721e-01 -1.47993803e-01
-1.06359339e+00 4.07745577e-02 -2.33106875e-05 4.27075662e-02
5.34413159e-01 -4.09030348e-01 5.18763736e-02 -1.78286076e-01
-1.03698206e+00 -8.91826689e-01 6.35138035e-01 5.99679649e-01
1.77267920e-02 2.55842417e-01 3.51677835e-01 -7.27718771e-01
8.19423556e-01 -8.03935230e-01 -7.48796940e-01 2.59178668e-01
-5.98415196e-01 2.26762444e-02 8.44867706e-01 -3.77444774e-01
-7.28994489e-01 -7.19076544e-02 -6.60566688e-01 -3.79649669e-01
-4.05990541e-01 6.96813345e-01 7.32756555e-02 -6.08046293e-01
7.07112014e-01 2.77832896e-01 7.94954505e-03 -4.30672914e-01
-2.13974267e-02 6.42767310e-01 -1.29268097e-03 -2.42518231e-01
8.90414476e-01 -1.71729133e-01 4.17182185e-02 -9.18321908e-01
-1.42016470e-01 -3.59624863e-01 -5.75534105e-01 -2.07021415e-01
8.82411599e-01 -2.57172078e-01 -4.35417444e-01 5.22427917e-01
-1.24545479e+00 -9.56720300e-03 -5.23119681e-02 1.09468162e-01
7.73247182e-02 6.27536297e-01 -5.08385062e-01 -1.01290464e+00
-5.93775988e-01 -1.51055264e+00 5.79255581e-01 -4.72065397e-02
-4.94947314e-01 -1.16206622e+00 3.59186292e-01 -3.01123530e-01
6.17787778e-01 5.79940736e-01 1.29076862e+00 -1.21734750e+00
1.59011465e-02 -8.45192432e-01 -2.45128293e-02 5.83633184e-01
2.82690048e-01 4.66502279e-01 -1.08404350e+00 -3.56428266e-01
1.97950825e-01 -5.19323759e-02 6.14216685e-01 -2.79189982e-02
8.02706361e-01 -2.85718162e-02 -1.51230782e-01 3.42690974e-01
1.74957728e+00 6.52347863e-01 6.43362522e-01 7.00399101e-01
6.27854347e-01 7.78047085e-01 3.00864428e-01 2.83387423e-01
-2.23716810e-01 7.16182828e-01 9.24663126e-01 1.68927014e-01
2.60826111e-01 3.09601665e-01 6.98391497e-01 5.20749450e-01
-2.90179309e-02 2.15928294e-02 -1.36875856e+00 4.65391666e-01
-1.21283102e+00 -9.11245048e-01 -3.93978417e-01 2.49514771e+00
3.75616282e-01 3.26719999e-01 2.11267427e-01 7.78784990e-01
8.78211617e-01 2.04618454e-01 -7.26826023e-03 -1.08995342e+00
3.55993062e-02 3.48492950e-01 4.10296381e-01 4.11396176e-01
-1.27890491e+00 5.74184000e-01 6.33089209e+00 7.07867086e-01
-1.49629617e+00 -1.53051298e-02 4.62711662e-01 3.31878155e-01
-1.41515538e-01 -2.11580023e-01 -9.53746557e-01 5.89552462e-01
1.55711162e+00 3.40948137e-03 5.11675887e-02 1.03348219e+00
-1.73538178e-01 1.65286824e-01 -7.30536342e-01 6.08382881e-01
2.52738744e-01 -1.01721668e+00 2.85650492e-02 1.00785851e-01
2.89183259e-01 -3.25711876e-01 3.31098974e-01 4.16041106e-01
-1.18907437e-01 -1.06905031e+00 6.59002304e-01 4.16473597e-01
7.27565825e-01 -1.05807209e+00 1.08201838e+00 1.11501120e-01
-1.28769135e+00 -6.75199866e-01 -3.53245884e-01 -3.82650197e-02
-2.41782621e-01 5.92025280e-01 -8.16327214e-01 6.97011411e-01
8.80468965e-01 4.76584285e-01 -9.00611281e-01 6.60970271e-01
-9.81999785e-02 5.07926047e-01 8.46577436e-02 -2.87429512e-01
4.82445806e-01 -6.00522496e-02 6.44680917e-01 1.48492849e+00
5.16741097e-01 -5.88799715e-01 -3.07459772e-01 1.01899016e+00
5.51212072e-01 2.41158098e-01 -8.41626763e-01 -3.40380281e-01
1.94296941e-01 1.40789473e+00 -8.42221022e-01 -1.59319013e-01
-4.35985863e-01 4.69056785e-01 2.91914344e-01 1.27978399e-01
-6.33560896e-01 -9.65328515e-01 7.13465393e-01 -2.45370328e-01
4.03075308e-01 7.14663640e-02 -2.08013177e-01 -7.92422056e-01
4.38979976e-02 -9.99921203e-01 3.08315516e-01 -6.55851960e-02
-9.16784048e-01 1.33292544e+00 -7.20711499e-02 -1.10912764e+00
-6.43052578e-01 -9.65694904e-01 -1.00251973e+00 1.29280221e+00
-1.07813847e+00 -9.02461350e-01 -2.95071173e-02 1.25828132e-01
2.30216518e-01 -7.60720372e-01 1.25352681e+00 2.34575123e-01
-8.90019834e-01 6.71857953e-01 1.52227700e-01 3.43921214e-01
2.38737822e-01 -1.23618722e+00 7.53559232e-01 1.21504259e+00
2.70410068e-02 7.68904209e-01 6.16597176e-01 -7.46126950e-01
-9.99874592e-01 -1.01206219e+00 1.01854193e+00 -5.07765293e-01
9.91308630e-01 -2.95538157e-01 -1.20859170e+00 3.89249474e-01
1.26110911e-01 -3.11639786e-01 6.67808294e-01 -1.37157410e-01
-6.58583581e-01 2.83165395e-01 -1.16871297e+00 3.23322386e-01
1.22418992e-01 -8.57328713e-01 -5.36632180e-01 -1.08143128e-01
6.26534283e-01 2.52480735e-03 -9.52824175e-01 3.64603221e-01
3.78540516e-01 -1.34932315e+00 1.11335158e+00 -3.61569315e-01
4.78276789e-01 -2.25905292e-02 -1.96863443e-01 -9.14404392e-01
-1.64170593e-01 -1.58528630e-02 -1.52889475e-01 1.19250572e+00
4.64550346e-01 -8.32977116e-01 4.82369661e-01 3.80686671e-01
1.69535756e-01 -7.02449381e-01 -9.28945005e-01 -6.39864445e-01
1.43380210e-01 -4.88125235e-01 6.92474902e-01 9.70705509e-01
4.54379525e-03 -3.30103487e-01 -1.06307410e-01 2.69792229e-01
4.21549737e-01 -8.71503130e-02 2.10822254e-01 -1.32394326e+00
-3.26384604e-01 -8.58566999e-01 -9.97120321e-01 2.39560053e-01
7.46974587e-01 -7.04756260e-01 -3.09924543e-01 -8.19434106e-01
-1.13581493e-01 -2.14021191e-01 -6.10897124e-01 4.54281449e-01
-5.94834834e-02 3.18614602e-01 -6.95990864e-04 -6.84931502e-02
2.12146640e-01 -2.52387017e-01 3.80116478e-02 -3.48397136e-01
1.12509830e-02 3.11033338e-01 -3.49968910e-01 5.34932435e-01
8.15812826e-01 -4.45674568e-01 -1.97124854e-01 -7.29719102e-02
4.84181523e-01 -2.77733374e-02 3.69809508e-01 -1.14039063e+00
-2.77119949e-02 4.21377629e-01 1.25558242e-01 -2.06074655e-01
6.06196411e-02 -7.91981101e-01 2.66713083e-01 8.58064175e-01
-2.81447500e-01 3.03036183e-01 5.78868032e-01 2.82482535e-01
-3.49165231e-01 -1.14672148e+00 5.33571124e-01 -1.46566153e-01
-9.52237725e-01 1.76812951e-02 -6.48714542e-01 -5.19526780e-01
1.13846552e+00 -3.20503205e-01 -6.20041676e-02 5.52759618e-02
-5.92879295e-01 -6.48996949e-01 3.70393008e-01 4.86098468e-01
7.77307093e-01 -9.66114283e-01 -5.31448185e-01 3.05871665e-01
1.52925193e-01 -7.65610456e-01 3.27815294e-01 3.48840833e-01
-5.17900765e-01 5.98097205e-01 -5.39889693e-01 -1.10533319e-01
-1.60851181e+00 8.44923794e-01 4.53942865e-01 -2.93518394e-01
-4.88820821e-01 7.46693015e-01 -2.93031484e-01 -3.81418824e-01
1.49162367e-01 -5.32222772e-03 -1.14601099e+00 2.81430274e-01
8.34204614e-01 4.06268537e-01 4.07284141e-01 -9.53936398e-01
-4.04921561e-01 7.73426235e-01 -2.24003289e-02 1.25803709e-01
1.14273787e+00 5.25039017e-01 -6.21519983e-01 5.75480342e-01
1.43469810e+00 -5.30385645e-03 -4.55092490e-01 2.44115219e-01
4.06131655e-01 -2.16552764e-01 -6.26259372e-02 -6.64655209e-01
-9.68013823e-01 1.12356448e+00 8.55520844e-01 9.12891507e-01
1.02370298e+00 -4.91887659e-01 3.62707913e-01 2.92881399e-01
4.18160945e-01 -4.12546456e-01 -2.83252120e-01 6.33272529e-01
6.03149116e-01 -1.17045391e+00 -4.20736104e-01 -4.07469235e-02
-8.46775547e-02 1.75243855e+00 2.75915116e-01 -1.06484145e-01
5.93657255e-01 4.52758700e-01 6.27084896e-02 -1.58146054e-01
-3.11467975e-01 2.60987841e-02 1.95708558e-01 5.44129610e-01
6.31757617e-01 1.22965544e-01 -2.36164451e-01 3.76152903e-01
-1.90490317e-02 -6.25455379e-01 8.12039137e-01 8.28638792e-01
-3.15006852e-01 -1.33749270e+00 -6.56887829e-01 4.15181637e-01
-8.19859743e-01 -2.37400264e-01 -4.57822204e-01 7.43069887e-01
-6.65478641e-03 9.24215019e-01 -1.14143088e-01 -8.47882330e-01
2.48027459e-01 1.82462111e-01 -1.07771076e-01 -7.75455236e-01
-1.27245951e+00 -6.14879012e-01 -2.05819353e-01 -4.60596770e-01
1.72212813e-02 -6.05789959e-01 -7.37005115e-01 -1.09099925e-01
-1.77475195e-02 4.17186558e-01 9.43531811e-01 9.09906209e-01
4.43251021e-02 8.58972430e-01 7.04848349e-01 -7.52038121e-01
-6.82821572e-01 -8.56549501e-01 -5.31735659e-01 2.20723420e-01
4.85843420e-01 -5.60265958e-01 -7.95554280e-01 -2.31472999e-01] | [7.062564849853516, 7.774383068084717] |
833446a8-cfed-4b9a-a979-b7d1181d29fb | galaxy-image-deconvolution-for-weak | 2211.01567 | null | https://arxiv.org/abs/2211.01567v3 | https://arxiv.org/pdf/2211.01567v3.pdf | Galaxy Image Deconvolution for Weak Gravitational Lensing with Unrolled Plug-and-Play ADMM | Removing optical and atmospheric blur from galaxy images significantly improves galaxy shape measurements for weak gravitational lensing and galaxy evolution studies. This ill-posed linear inverse problem is usually solved with deconvolution algorithms enhanced by regularisation priors or deep learning. We introduce a so-called "physics-informed deep learning" approach to the Point Spread Function (PSF) deconvolution problem in galaxy surveys. We apply algorithm unrolling and the Plug-and-Play technique to the Alternating Direction Method of Multipliers (ADMM), in which a neural network learns appropriate hyperparameters and denoising priors from simulated galaxy images. We characterise the time-performance trade-off of several methods for galaxies of differing brightness levels as well as our method's robustness to systematic PSF errors and network ablations. We show an improvement in reduced shear ellipticity error of 38.6% (SNR=20)/45.0% (SNR=200) compared to classic methods and 7.4% (SNR=20)/33.2% (SNR=200) compared to modern methods. | ['Emma Alexander', 'Tianao Li'] | 2022-11-03 | null | null | null | null | ['image-deconvolution'] | ['computer-vision'] | [ 5.93769713e-04 -1.48383398e-02 8.31527948e-01 -1.69938520e-01
-5.23481190e-01 -8.28270555e-01 9.81849074e-01 -7.33138978e-01
-5.53418458e-01 7.42141962e-01 4.40656900e-01 -4.07862127e-01
-4.00059432e-01 -4.08522695e-01 -6.78155661e-01 -1.16450846e+00
-1.11577265e-01 4.99815047e-01 2.36652732e-01 1.05305091e-01
5.54017305e-01 4.24087495e-01 -1.06399286e+00 -1.83934227e-01
5.95740497e-01 7.03076661e-01 4.63349819e-01 1.10680842e+00
1.79559380e-01 5.26353002e-01 -2.53184497e-01 -3.26992333e-01
4.74618405e-01 -3.44999671e-01 -7.05916226e-01 3.95064175e-01
4.68892872e-01 -5.49804926e-01 -6.71120644e-01 1.24148428e+00
7.01300919e-01 2.72321463e-01 6.94795072e-01 -4.43385780e-01
-5.79332590e-01 6.47704676e-02 -8.58643413e-01 4.63817924e-01
-4.40418482e-01 7.31811941e-01 3.90564322e-01 -1.06492090e+00
3.03542405e-01 1.23863244e+00 9.42883730e-01 2.05020174e-01
-1.49685585e+00 -7.40833879e-02 -6.88054919e-01 -1.15195744e-01
-1.02198958e+00 -6.42606139e-01 2.85900563e-01 -8.96815896e-01
6.47312701e-01 1.86896980e-01 1.73099607e-01 6.36071205e-01
9.30247903e-02 -1.27736479e-02 1.36541808e+00 -4.44890767e-01
8.20473880e-02 1.02024794e-01 -1.04781874e-01 4.19502020e-01
6.71045721e-01 5.08865118e-01 -2.33724833e-01 -2.13371798e-01
1.43179560e+00 -6.01497948e-01 -5.36369979e-01 1.73605293e-01
-1.48179173e+00 7.93987632e-01 3.22205871e-01 -2.31239468e-01
-2.53512889e-01 4.68535841e-01 2.24347226e-02 -2.66803987e-02
6.93800211e-01 9.82120693e-01 -4.65624720e-01 -9.99826565e-03
-9.59239185e-01 5.38080096e-01 7.89100707e-01 4.91317600e-01
7.88178861e-01 2.97998875e-01 2.25738380e-02 9.84903276e-01
5.79823375e-01 7.50852764e-01 3.78779978e-01 -1.79393613e+00
-3.36101055e-02 -1.74440265e-01 4.98186707e-01 -6.60534441e-01
-2.43977770e-01 -9.41017270e-01 -8.24660599e-01 6.58747554e-01
9.96205926e-01 -3.67109418e-01 -6.74778759e-01 1.29784608e+00
1.50583118e-01 4.60260898e-01 4.29802239e-02 1.20736098e+00
7.51358449e-01 4.75288540e-01 -6.16429746e-01 -1.94969609e-01
1.43953049e+00 -8.57983649e-01 -2.44974896e-01 -7.49710441e-01
-1.00294307e-01 -1.23132181e+00 7.26606011e-01 4.47136670e-01
-1.33759463e+00 -3.78383875e-01 -9.44669068e-01 1.36491079e-02
3.68397176e-01 7.10600615e-02 3.38770062e-01 6.12087965e-01
-1.20729685e+00 1.09564126e+00 -6.09147966e-01 4.33824956e-02
1.29182220e-01 3.68320584e-01 -9.11762267e-02 3.23059797e-01
-4.81790602e-01 6.56269789e-01 1.62597403e-01 -6.42744005e-02
-1.01649773e+00 -1.03788650e+00 -3.74560177e-01 -2.84914523e-02
2.03989238e-01 -9.16999042e-01 1.66319728e+00 -7.49276161e-01
-1.70450842e+00 9.75228012e-01 -1.60713848e-02 -6.39966846e-01
5.40379107e-01 -3.95447642e-01 1.69775449e-02 1.88406691e-01
-1.02379292e-01 2.06549704e-01 1.07213426e+00 -1.33443391e+00
-4.01044250e-01 -1.16196506e-01 -1.30661756e-01 5.21187298e-02
3.39522392e-01 2.42449373e-01 -3.32160264e-01 -6.56895936e-01
3.26113075e-01 -1.03023458e+00 -2.56347954e-01 -3.24592218e-02
-5.55024333e-02 4.77047205e-01 5.63578129e-01 -1.14080584e+00
3.96663815e-01 -1.94033289e+00 6.20187484e-02 -2.27799103e-01
5.11214435e-01 3.50483149e-01 7.90479183e-02 7.51311257e-02
-1.77351505e-01 -3.34481537e-01 -4.37428981e-01 -2.97124237e-01
-1.98429480e-01 5.08418567e-02 -1.91260353e-02 8.67725551e-01
-1.30624529e-02 5.17055571e-01 -8.65381539e-01 3.06394368e-01
3.53292406e-01 5.95579505e-01 -6.57700241e-01 5.61733663e-01
2.52791252e-02 8.76009107e-01 3.04201692e-01 2.04286203e-01
1.34340036e+00 -2.27787644e-01 -9.51509401e-02 -3.98197770e-01
-5.93768239e-01 1.59979448e-01 -1.20818353e+00 1.25699842e+00
-3.52103263e-01 1.11374116e+00 7.57299364e-01 -6.89493299e-01
7.57932365e-01 1.83577672e-01 -1.53902829e-01 -3.17190617e-01
8.62163827e-02 4.21191305e-01 2.40203053e-01 -7.12831855e-01
3.83097917e-01 -5.89507103e-01 7.01505661e-01 2.85943121e-01
3.31476897e-01 -4.41861212e-01 -2.05961689e-01 1.00573108e-01
9.85712230e-01 9.73588452e-02 -1.24986000e-01 -8.80515516e-01
5.39157748e-01 -3.41601014e-01 2.43849441e-01 1.02886832e+00
-2.27648560e-02 1.08579361e+00 6.57547951e-01 -3.51224810e-01
-1.59814596e+00 -8.27289402e-01 -2.52756894e-01 7.71736801e-01
-2.80288905e-01 1.33240134e-01 -7.91564226e-01 4.16870043e-02
1.10984817e-01 6.85051262e-01 -3.37297827e-01 -4.83824266e-03
-6.35310113e-01 -1.38659394e+00 4.41137969e-01 1.38359159e-01
7.73660719e-01 -5.20739198e-01 -2.26083279e-01 9.06100962e-03
1.00673690e-01 -1.03505063e+00 -2.20488861e-01 -1.19417489e-01
-7.56406248e-01 -9.61939275e-01 -9.61950302e-01 -2.66015768e-01
5.99677920e-01 5.03498197e-01 1.22916210e+00 -1.91393763e-01
9.28177685e-02 3.15845698e-01 1.34310529e-01 -8.13959762e-02
-6.71453416e-01 -5.21084845e-01 2.70034045e-01 1.31393015e-01
-2.72323638e-01 -6.30242944e-01 -1.02819574e+00 5.90021968e-01
-6.40811324e-01 -9.71126109e-02 6.38160765e-01 9.32290375e-01
-1.28381684e-01 2.61111017e-02 1.71563342e-01 -5.01924694e-01
3.60418677e-01 -3.41212183e-01 -1.37263680e+00 -3.97709221e-01
-7.33483672e-01 3.17556471e-01 2.24448919e-01 -3.92846107e-01
-1.71917450e+00 -2.62086123e-01 -6.94327205e-02 -3.39595765e-01
-4.28209454e-02 1.83927938e-01 1.71875447e-01 -7.92786777e-01
1.06334722e+00 -1.15725910e-03 4.23608691e-01 -8.29973221e-01
3.26413542e-01 5.58409214e-01 1.36765122e+00 -4.70616430e-01
1.13323593e+00 8.14687371e-01 2.27523297e-01 -9.88415062e-01
-6.99621081e-01 -6.38846219e-01 -4.57732499e-01 -2.84935385e-01
6.88997030e-01 -9.43524301e-01 -8.01993072e-01 7.94697344e-01
-1.10318720e+00 -3.19347888e-01 -6.17900006e-02 9.32438612e-01
-4.04671818e-01 7.57852316e-01 -6.88489616e-01 -7.69699097e-01
-1.22755542e-01 -1.22124505e+00 7.64864624e-01 5.54504335e-01
3.16974342e-01 -1.30665708e+00 -6.02808408e-02 4.51290578e-01
9.29284632e-01 1.21355958e-01 3.37396771e-01 4.46478324e-03
-5.90231240e-01 1.62036002e-01 -8.62141907e-01 7.32179224e-01
-1.64203882e-01 -2.33634129e-01 -1.34054279e+00 -5.82768917e-01
8.89310658e-01 1.86822385e-01 1.17896545e+00 1.00966358e+00
8.53028715e-01 -5.02151191e-01 6.54312819e-02 1.12531257e+00
1.52787673e+00 -1.67837545e-01 8.44202340e-01 5.48060179e-01
6.10960066e-01 6.96229458e-01 7.91076943e-03 4.12019670e-01
-3.17112654e-01 4.89702761e-01 5.51535547e-01 -9.82833933e-03
-5.94501853e-01 5.66052198e-01 1.11635573e-01 4.17506188e-01
-4.04806823e-01 4.19178978e-02 -8.17431688e-01 5.29103577e-01
-1.43738627e+00 -7.49266267e-01 -9.33924496e-01 2.54759026e+00
7.13404894e-01 1.30472958e-01 -1.04499593e-01 -2.91994929e-01
7.96324849e-01 -8.15189853e-02 -2.55319208e-01 -1.88274030e-02
-1.65901721e-01 2.38245595e-02 1.10189438e+00 9.26725447e-01
-9.70621765e-01 3.60248417e-01 6.19418478e+00 6.22983992e-01
-9.43847537e-01 2.86640733e-01 6.53806567e-01 -5.24873212e-02
-7.20023364e-02 2.77749479e-01 -6.09460771e-01 6.17824912e-01
9.99193788e-01 1.40037909e-01 8.48198771e-01 2.88529396e-01
5.88243306e-01 -5.02629519e-01 -5.95156074e-01 1.13482964e+00
-9.23559368e-02 -1.56668568e+00 -7.86998451e-01 3.78494561e-01
8.84173036e-01 4.46478188e-01 5.62990606e-02 -2.25944877e-01
1.69775397e-01 -8.75423133e-01 5.02178550e-01 7.63416827e-01
8.99025798e-01 -5.45075774e-01 8.65302145e-01 1.38511866e-01
-3.78798127e-01 1.08023718e-01 -6.76323235e-01 -2.96202689e-01
2.38314137e-01 1.06681025e+00 -1.12327147e+00 3.72449964e-01
7.34090388e-01 2.76562929e-01 -5.13267100e-01 1.48690319e+00
-7.50600696e-02 1.06894219e+00 -3.29164714e-01 4.67073113e-01
4.43747163e-01 -5.82025528e-01 1.10716629e+00 1.45828915e+00
4.68299508e-01 6.01858497e-02 -7.46180415e-01 9.97204363e-01
2.21989341e-02 -7.24158764e-01 -1.04527362e-01 3.17737848e-01
7.40468204e-02 1.54118681e+00 -7.11949468e-01 -3.04155380e-01
-4.89676595e-01 6.84084058e-01 -3.12642545e-01 6.05206430e-01
-5.39806485e-01 -3.45731378e-01 7.01533198e-01 5.40472329e-01
5.22625864e-01 -3.97787273e-01 -2.64717400e-01 -9.89242494e-01
-4.93348449e-01 -5.98907351e-01 -9.33151245e-02 -1.18865371e+00
-1.33601785e+00 5.67019843e-02 7.40989223e-02 -8.88564587e-01
-3.55745142e-04 -1.13562953e+00 -8.92488241e-01 1.17796230e+00
-1.13987470e+00 -7.75152504e-01 -5.48343182e-01 -1.08463787e-01
4.23151523e-01 -4.31285024e-01 1.03823222e-01 2.02549860e-01
-2.17124864e-01 -1.39110446e-01 8.65523338e-01 -1.29481018e-01
7.65364647e-01 -1.69087589e+00 7.18156815e-01 1.09575093e+00
-5.48339248e-01 5.17210066e-01 1.32490361e+00 -4.55109864e-01
-1.42944276e+00 -9.86729145e-01 4.85583454e-01 -3.47086668e-01
9.75210965e-01 -9.97684672e-02 -1.22190332e+00 5.83554924e-01
6.82663441e-01 2.53427893e-01 1.15371786e-01 -3.26924860e-01
-4.60819826e-02 4.01748680e-02 -1.14594483e+00 4.12355334e-01
6.17618203e-01 -2.59085596e-01 -6.81423068e-01 4.65345025e-01
4.57302332e-01 -4.41781551e-01 -5.05529463e-01 3.10405254e-01
4.68333781e-01 -1.20083833e+00 1.26004577e+00 -2.20323309e-01
2.81899869e-01 -5.82959771e-01 -8.46956521e-02 -1.47698236e+00
-8.65340114e-01 -1.21465290e+00 2.25730632e-02 1.06995046e+00
1.42972702e-02 -8.75680983e-01 6.06998324e-01 1.08322583e-01
-3.99293005e-01 9.44015086e-02 -8.67501140e-01 -6.77767038e-01
6.14302903e-02 -1.01451173e-01 4.40410860e-02 7.72657037e-01
-8.22619915e-01 3.38988334e-01 -4.21927720e-01 5.45387864e-01
1.21123672e+00 -2.77201116e-01 6.58199966e-01 -1.08676386e+00
-8.82521212e-01 -4.30151433e-01 -8.36994722e-02 -1.12554276e+00
-3.73747647e-01 -5.50914943e-01 2.16799885e-01 -1.14039063e+00
2.45449901e-01 5.23682265e-03 2.51042098e-01 -3.31041425e-01
-9.25042629e-02 4.01565760e-01 -2.53639638e-01 2.68027425e-01
3.11152898e-02 8.52872501e-04 1.05327737e+00 1.10819891e-01
1.42527327e-01 1.94664299e-01 -5.80990434e-01 9.92250323e-01
7.47156680e-01 -3.02699536e-01 -1.47013083e-01 -6.94623947e-01
3.18563819e-01 3.80091853e-02 9.74000692e-01 -1.00057995e+00
1.40814468e-01 -4.78135496e-02 5.16592801e-01 -2.00265020e-01
2.78239518e-01 -3.49640474e-02 3.15482974e-01 4.38747287e-01
1.37043223e-01 -4.30117697e-01 3.19365680e-01 4.36623871e-01
2.24074364e-01 -7.69525945e-01 1.27050865e+00 -6.52063012e-01
-3.52347583e-01 -2.82838166e-01 -5.29092968e-01 1.53140143e-01
2.48254880e-01 2.25426495e-01 -5.96340477e-01 -5.02724707e-01
-5.54134727e-01 -3.26234192e-01 4.33907956e-01 -2.62267709e-01
3.48018706e-01 -9.47361410e-01 -1.14988422e+00 1.59177363e-01
-3.99998844e-01 4.35272567e-02 8.39335322e-02 1.13313031e+00
-1.21787286e+00 1.48514867e-01 6.57717511e-02 -6.41610801e-01
-9.61865544e-01 1.89209670e-01 9.38074231e-01 2.97878683e-01
-3.71680856e-01 1.55328739e+00 5.24674296e-01 -4.11030293e-01
-2.40853146e-01 6.98643476e-02 1.61298722e-01 -3.80141705e-01
6.91968322e-01 1.03029490e+00 2.36067951e-01 -4.73576993e-01
-3.66699062e-02 3.66473645e-01 1.60202384e-01 -3.47722292e-01
1.31139922e+00 -3.81037533e-01 -6.01984143e-01 -3.07165273e-02
9.88629580e-01 1.98302090e-01 -1.93282020e+00 -2.54306883e-01
-5.73233739e-02 -9.07383025e-01 7.17840493e-01 -8.32883716e-01
-9.75728929e-01 6.84035063e-01 6.54802084e-01 3.69685203e-01
7.13458896e-01 1.91360280e-01 4.94527251e-01 1.43169433e-01
-4.19328570e-01 -7.69195199e-01 6.98314980e-02 6.51122630e-01
1.10053933e+00 -1.17099082e+00 2.81934649e-01 -1.27082750e-01
-2.59562314e-01 1.19734406e+00 2.88142115e-01 -4.66614068e-01
5.01722395e-01 3.80214721e-01 -4.00689654e-02 -3.43417972e-01
-3.80462080e-01 2.68400218e-02 4.05349940e-01 5.27066231e-01
2.97251374e-01 -1.80655122e-01 -1.99373707e-01 1.82933614e-01
-1.56736597e-01 -4.22179878e-01 8.99327815e-01 3.98291856e-01
-8.34536910e-01 -7.38348663e-01 -1.16351581e+00 4.75850195e-01
-4.43531007e-01 -1.86964184e-01 4.67454530e-02 3.22093725e-01
1.61876976e-01 8.31102550e-01 1.47835821e-01 1.19650170e-01
-2.89713386e-02 -2.17720807e-01 6.04789793e-01 -2.85915047e-01
-3.10928285e-01 7.57193863e-01 2.78193086e-01 6.50272965e-02
-3.16203445e-01 -8.91088903e-01 -8.20679843e-01 -5.63913882e-01
-3.61228108e-01 9.76731405e-02 7.29765713e-01 8.88164222e-01
1.25206813e-01 4.68381077e-01 4.81878042e-01 -1.59163487e+00
-5.58534920e-01 -1.08935833e+00 -6.47551775e-01 -7.82770142e-02
6.88722253e-01 -7.01106369e-01 -1.12360883e+00 2.21626207e-01] | [11.638808250427246, -2.724001884460449] |
0a4d890b-1eec-40b3-9212-ffe455ba801c | ego-vehicle-speed-estimation-using-3d | 2212.05432 | null | https://arxiv.org/abs/2212.05432v1 | https://arxiv.org/pdf/2212.05432v1.pdf | Ego Vehicle Speed Estimation using 3D Convolution with Masked Attention | Speed estimation of an ego vehicle is crucial to enable autonomous driving and advanced driver assistance technologies. Due to functional and legacy issues, conventional methods depend on in-car sensors to extract vehicle speed through the Controller Area Network bus. However, it is desirable to have modular systems that are not susceptible to external sensors to execute perception tasks. In this paper, we propose a novel 3D-CNN with masked-attention architecture to estimate ego vehicle speed using a single front-facing monocular camera. To demonstrate the effectiveness of our method, we conduct experiments on two publicly available datasets, nuImages and KITTI. We also demonstrate the efficacy of masked-attention by comparing our method with a traditional 3D-CNN. | ['Thariq Khalid', 'Athul M. Mathew'] | 2022-12-11 | null | null | null | null | ['vehicle-speed-estimation'] | ['computer-vision'] | [-3.41410637e-01 -2.31095552e-02 -2.26517305e-01 -4.86501426e-01
-1.26324192e-01 -4.05845016e-01 4.77016807e-01 -6.53080106e-01
-5.23811281e-01 2.86237568e-01 -2.48104230e-01 -7.30215669e-01
3.55110586e-01 -5.92996299e-01 -8.06344867e-01 -3.06959242e-01
3.27813685e-01 -2.00307220e-01 3.51750612e-01 -3.61720622e-01
4.08765450e-02 8.36032867e-01 -1.91425025e+00 -2.50416160e-01
5.62450469e-01 1.28648758e+00 2.84455359e-01 8.08623970e-01
3.51646572e-01 6.80489302e-01 -4.74202931e-01 -1.66504830e-02
4.27530736e-01 2.45114356e-01 -1.10454634e-01 -8.69588107e-02
4.53181177e-01 -8.75763774e-01 -7.35608697e-01 9.72523212e-01
2.80461669e-01 -2.46778205e-01 1.68652534e-01 -1.70493460e+00
-2.21033037e-01 4.47985530e-03 -3.79616350e-01 3.43411416e-01
-5.49349003e-02 5.91348290e-01 3.18874806e-01 -6.04527354e-01
2.57188231e-01 1.15262353e+00 4.25529689e-01 5.98947287e-01
-7.87703216e-01 -9.57388282e-01 2.72381812e-01 5.63806832e-01
-1.40195584e+00 -9.15164649e-01 9.31710184e-01 -4.10941124e-01
1.15685403e+00 -2.04674751e-01 3.53973001e-01 1.09681523e+00
3.86734307e-01 5.94469309e-01 8.75613153e-01 2.45801106e-01
2.47707441e-01 2.49230951e-01 1.31027132e-01 5.33312500e-01
3.87788266e-01 5.34009457e-01 -3.60996872e-01 5.75908840e-01
5.61620355e-01 5.87822264e-03 2.69545075e-02 -4.01987076e-01
-1.15216720e+00 5.94942212e-01 5.40954292e-01 -2.65399754e-01
-3.72189254e-01 5.58484495e-01 2.32249901e-01 3.17748219e-01
3.41176718e-01 5.24447113e-02 -3.87325525e-01 -3.31189215e-01
-4.70847696e-01 1.13613605e-01 3.14780116e-01 1.49298739e+00
8.81849408e-01 4.26369339e-01 2.15081215e-01 8.51556882e-02
3.13900083e-01 8.98508012e-01 3.33604485e-01 -1.16422474e+00
3.45565617e-01 5.01951337e-01 1.73325315e-01 -8.04272115e-01
-5.67946613e-01 -3.30910802e-01 -4.84706640e-01 6.83375776e-01
4.74838912e-03 -2.33196184e-01 -7.96486080e-01 1.41595757e+00
2.06903264e-01 2.22877845e-01 2.54740059e-01 1.28148341e+00
7.70601153e-01 2.25136817e-01 -3.39307547e-01 2.88398892e-01
1.10169578e+00 -9.22817409e-01 -7.53488898e-01 -5.63287735e-01
4.76579666e-01 -2.63303906e-01 8.40970635e-01 1.82149947e-01
-7.48109341e-01 -9.77502465e-01 -1.52411413e+00 -2.54052162e-01
-5.33811152e-01 1.51682526e-01 5.29284179e-01 7.44659305e-01
-1.04643500e+00 1.28105611e-01 -9.57728446e-01 -1.17912143e-01
2.72700161e-01 5.07795513e-01 -3.70383143e-01 2.52346545e-01
-1.06985390e+00 1.02610898e+00 1.22980796e-01 2.32629687e-01
-1.10782623e+00 -5.07108033e-01 -9.82951999e-01 -6.48093373e-02
1.72233418e-01 -6.06026590e-01 1.51647139e+00 -5.37445188e-01
-1.82991552e+00 5.50215006e-01 -3.63608152e-01 -7.38529265e-01
4.89148229e-01 -3.82340103e-01 -6.13358319e-01 3.01020201e-02
1.03302803e-02 1.00500011e+00 9.50979590e-01 -8.74442220e-01
-8.49924564e-01 -4.61149424e-01 2.93534070e-01 -1.19124763e-01
-1.15142420e-01 -3.08218777e-01 -4.77813393e-01 1.78358570e-01
-9.48004574e-02 -9.15741086e-01 -6.89670220e-02 2.47558188e-02
-1.96118206e-01 6.82505174e-03 1.55568027e+00 -2.48029560e-01
7.01813936e-01 -2.33715320e+00 -5.65865576e-01 -1.45635143e-01
5.34789801e-01 4.17110324e-01 -2.49202605e-02 -2.52064854e-01
1.25024810e-01 -2.50883639e-01 4.53084558e-01 -3.07346135e-01
2.09218115e-02 2.75244806e-02 -3.41419578e-01 7.23187029e-01
5.28796613e-01 1.06414306e+00 -6.93197608e-01 -2.29691699e-01
6.98276818e-01 5.77514708e-01 -4.89546120e-01 1.15688324e-01
2.70210486e-03 4.23579007e-01 -4.49476749e-01 7.54530907e-01
7.66929328e-01 1.64341450e-01 -3.36127430e-01 -2.64081419e-01
-5.88318110e-01 4.08521712e-01 -7.77529359e-01 1.27299726e+00
-5.30200839e-01 1.34243715e+00 2.64381945e-01 -7.60546207e-01
9.16370332e-01 1.01039730e-01 6.64145276e-02 -1.19137239e+00
7.42882967e-01 6.41879346e-03 -4.49343175e-02 -5.42870164e-01
7.08742380e-01 3.30290884e-01 -1.40692145e-01 -7.46920779e-02
-3.44625324e-01 -7.61881471e-02 -1.20094776e-01 -7.14614764e-02
1.02232337e+00 1.09966442e-01 -7.37069920e-02 -7.76053667e-02
4.45176482e-01 1.71958685e-01 5.82275510e-01 4.59375352e-01
-8.01848471e-01 1.56876013e-01 3.68962973e-01 -5.02287984e-01
-1.12824225e+00 -1.02060747e+00 -1.36357755e-03 5.07054567e-01
5.12638032e-01 -2.20236275e-02 -6.71182930e-01 -5.28317213e-01
2.64439672e-01 7.04320967e-01 -3.56812656e-01 -2.97203064e-01
-4.46628749e-01 1.40067443e-01 5.92283785e-01 9.69390333e-01
8.14177275e-01 -2.49066547e-01 -1.18148148e+00 6.98407665e-02
1.04629546e-02 -1.78251290e+00 -4.35691476e-01 -1.17675625e-01
-5.25731206e-01 -8.66964281e-01 -5.30880541e-02 -5.09107172e-01
3.99028152e-01 1.11407685e+00 6.58214688e-01 -2.23839507e-01
7.67540857e-02 1.57077894e-01 7.66552100e-03 -9.58417535e-01
-2.07195133e-01 1.51975721e-01 4.80579197e-01 9.73652378e-02
8.74544144e-01 -3.13096553e-01 -8.20846796e-01 5.63777387e-01
-2.00279146e-01 2.84203202e-01 6.33704960e-01 2.26811171e-01
3.64120901e-01 -5.48190549e-02 6.43875480e-01 -8.74194801e-02
2.61236936e-01 -3.79532367e-01 -1.40469491e+00 -5.74331105e-01
-7.16218174e-01 -1.30213276e-01 5.83103895e-01 -2.10434914e-01
-7.46817291e-01 3.09002966e-01 8.26293230e-02 -8.97021413e-01
-3.51834744e-01 -1.34151466e-02 -3.86305869e-01 -1.02611132e-01
3.06319982e-01 -1.28492281e-01 4.12908286e-01 -5.63805597e-03
3.92010301e-01 1.27156651e+00 8.53583395e-01 2.80597180e-01
8.24101627e-01 9.04222906e-01 5.48363738e-02 -1.10960364e+00
-4.30154055e-01 -5.25968850e-01 -4.65268821e-01 -5.46851158e-01
1.01534057e+00 -1.58206522e+00 -1.36392128e+00 4.26882088e-01
-1.15254986e+00 -3.19246799e-01 7.25829154e-02 8.04537177e-01
-6.40470207e-01 2.51045562e-02 -2.84086883e-01 -7.36934960e-01
-1.63646072e-01 -1.34851694e+00 1.14195764e+00 3.50012720e-01
-8.00288003e-03 -5.90565264e-01 -2.40012273e-01 4.95450586e-01
7.68306613e-01 4.05522399e-02 1.15869880e-01 -3.08384988e-02
-9.73079741e-01 -4.91003603e-01 -5.39634883e-01 3.23143810e-01
2.10578814e-02 -6.83531165e-02 -1.43759775e+00 -2.00795427e-01
-4.78317402e-02 -7.83883706e-02 8.18134248e-01 3.90251905e-01
1.14987648e+00 1.50734529e-01 -5.22060513e-01 8.54624689e-01
1.11603081e+00 2.28435278e-01 6.65154219e-01 3.05365622e-01
7.39542365e-01 4.05417264e-01 6.77219927e-01 2.54567936e-02
7.43840992e-01 6.43428266e-01 7.55925655e-01 -1.05128214e-01
-1.39804721e-01 -3.04365456e-01 5.48515618e-01 5.55623651e-01
1.87041819e-01 -1.16272263e-01 -6.56101763e-01 7.48298585e-01
-1.63646197e+00 -9.03096020e-01 -2.42010579e-01 1.87753916e+00
9.90414545e-02 5.20301044e-01 -3.01772263e-02 4.46000062e-02
5.03304243e-01 -1.02314994e-01 -9.47180986e-01 -5.41213274e-01
1.29131794e-01 -2.70442456e-01 1.21143496e+00 4.62460816e-01
-1.12292254e+00 1.04015815e+00 6.51815510e+00 1.86681181e-01
-1.53406692e+00 1.55759647e-01 2.42765620e-01 -2.94741303e-01
-1.31004192e-02 -1.85800359e-01 -1.26069522e+00 4.20865476e-01
1.41531086e+00 -1.70594186e-01 3.11772019e-01 1.14945507e+00
6.26364410e-01 -1.11334272e-01 -1.12495744e+00 1.02304459e+00
-4.29828167e-02 -1.08687186e+00 -5.30186594e-01 2.87933677e-01
4.19743866e-01 5.49532592e-01 9.22236890e-02 4.20790195e-01
1.29894689e-01 -8.63444388e-01 8.59629631e-01 2.30877370e-01
1.04743862e+00 -7.55943120e-01 7.84112990e-01 4.97283548e-01
-1.29369926e+00 -1.54492721e-01 -3.47697914e-01 -5.70156097e-01
2.68956900e-01 3.01981091e-01 -1.08887649e+00 7.86832497e-02
7.34390676e-01 6.47479594e-01 -3.56115758e-01 7.26689279e-01
-1.85436100e-01 3.44125450e-01 -2.92820096e-01 -2.00281292e-01
2.96839327e-01 1.20374054e-01 4.98983860e-01 8.35048616e-01
2.16806591e-01 -1.65561616e-01 -2.85913974e-01 1.08508432e+00
-2.23340858e-02 -7.07173467e-01 -1.18678904e+00 1.14906952e-01
3.84699106e-01 1.29437816e+00 -2.05721512e-01 -1.80508882e-01
-8.23290169e-01 6.18004024e-01 1.12309061e-01 3.87645185e-01
-1.24855566e+00 -6.26534820e-01 1.51581967e+00 3.08271408e-01
4.81106311e-01 -8.36652517e-01 -2.81100273e-01 -8.42342198e-01
2.46337242e-02 -3.11609745e-01 -4.71623868e-01 -9.52500582e-01
-5.57260394e-01 5.83128572e-01 -9.37789604e-02 -1.48864555e+00
-1.13089435e-01 -8.34889948e-01 -4.01888430e-01 7.44569480e-01
-2.09954381e+00 -9.46233153e-01 -8.60644341e-01 6.14827096e-01
4.80581790e-01 -2.58481026e-01 2.23409802e-01 4.13500607e-01
-7.03213751e-01 6.03421509e-01 -1.84564382e-01 5.67084141e-02
6.00578547e-01 -8.37480664e-01 1.08329487e+00 9.28634763e-01
-6.62231028e-01 2.63079077e-01 6.79765463e-01 -3.91536415e-01
-1.97239840e+00 -1.46904731e+00 6.24129832e-01 -7.18838930e-01
5.16697168e-01 -4.78318155e-01 -4.08326954e-01 8.93368185e-01
3.41147631e-01 1.21480465e-01 1.36368707e-01 -4.78379726e-01
-2.19135940e-01 -6.04199409e-01 -9.21341717e-01 6.46633565e-01
9.92627442e-01 -5.76102376e-01 -3.10745358e-01 -1.90592170e-01
1.06701243e+00 -5.12277126e-01 -3.56422216e-01 3.78554612e-01
7.70219266e-01 -8.20933878e-01 6.91856802e-01 -3.82256895e-01
-3.40113677e-02 -7.62315214e-01 -2.27284923e-01 -1.22296953e+00
-2.15576142e-01 -5.83667517e-01 -2.47831017e-01 6.81756914e-01
2.15694800e-01 -1.08050621e+00 9.97692823e-01 4.74566162e-01
-5.41765511e-01 -2.52151936e-01 -9.83675003e-01 -1.01709712e+00
-2.77521402e-01 -9.72043991e-01 8.77093613e-01 3.56206834e-01
-8.79696012e-02 4.83943492e-01 -2.27996245e-01 7.23771334e-01
6.01716697e-01 -1.92708313e-01 1.29941487e+00 -1.03644562e+00
2.38711327e-01 -4.48583692e-01 -9.37849164e-01 -1.49580336e+00
2.76250899e-01 -3.31620276e-01 2.98746198e-01 -1.15443575e+00
-2.88623303e-01 -5.23911119e-02 -1.23072669e-01 4.25517261e-02
2.77202904e-01 2.23395988e-01 -1.53566465e-01 -2.06597045e-01
-6.14749014e-01 6.34410799e-01 1.03639805e+00 -3.15038741e-01
-1.20650150e-01 -6.96763694e-02 -6.11828387e-01 6.01540625e-01
1.01817954e+00 -6.77614436e-02 -5.14559329e-01 -8.26148391e-01
2.50170175e-02 -8.21096227e-02 7.39122748e-01 -1.32805526e+00
6.32360756e-01 -3.94551605e-02 2.44805291e-01 -1.06763577e+00
5.46813667e-01 -9.25734401e-01 -1.67370543e-01 4.51843262e-01
2.72223085e-01 3.25330466e-01 5.04450858e-01 4.74354148e-01
-1.20335296e-01 5.35164595e-01 6.14163935e-01 3.08744162e-01
-1.19442821e+00 3.49799573e-01 -7.44657874e-01 -3.50293636e-01
1.33191943e+00 -4.95349497e-01 -7.28960931e-01 -4.60602969e-01
-6.15692977e-03 5.71390212e-01 6.36682689e-01 9.13078785e-01
8.78495753e-01 -1.30518985e+00 -3.29416215e-01 8.13563287e-01
4.30103689e-01 -3.91131677e-02 2.43629023e-01 8.25032353e-01
-5.91708422e-01 1.20200992e+00 -3.23046178e-01 -9.22381401e-01
-9.56631541e-01 8.79569411e-01 5.21190703e-01 7.82313287e-01
-5.56571364e-01 3.25482935e-01 8.58363435e-02 -2.95576423e-01
2.14611024e-01 -6.53046429e-01 -7.15998188e-02 -5.45001566e-01
6.61744773e-01 4.72927272e-01 2.91456133e-01 -7.46748626e-01
-4.77030814e-01 5.24709761e-01 9.31344926e-02 3.63065936e-02
6.37559772e-01 -5.57789326e-01 6.47253215e-01 2.83738345e-01
1.25300503e+00 -1.76898092e-01 -1.75292134e+00 3.25054228e-02
-4.92023557e-01 -3.90245616e-01 7.76799858e-01 -9.40371156e-02
-1.15952098e+00 9.79746461e-01 8.76636803e-01 2.65656784e-02
9.84272540e-01 -8.11749101e-02 9.48785543e-01 4.35787559e-01
4.62279320e-01 -1.09840977e+00 -3.66156429e-01 5.11515558e-01
4.26574379e-01 -1.55344069e+00 -4.96753514e-01 -2.66434461e-01
-4.26235139e-01 8.47002923e-01 1.03231430e+00 1.45927250e-01
7.88790524e-01 4.54584569e-01 4.41078395e-01 -2.28938714e-01
-1.02002752e+00 -5.97733438e-01 3.99993733e-02 7.15336084e-01
-5.84798567e-02 1.27287924e-01 3.79117548e-01 3.14358801e-01
-3.02666306e-01 1.57205418e-01 6.76708460e-01 8.81399274e-01
-3.95661592e-01 -4.10094321e-01 -1.68255672e-01 9.79666561e-02
-1.18830688e-01 4.71342385e-01 -1.69865429e-01 8.79074931e-01
1.19583786e-01 1.40956366e+00 5.51753819e-01 -8.57186675e-01
6.86416507e-01 -2.80637980e-01 7.76279420e-02 -1.61473945e-01
7.65688345e-02 -3.72716188e-01 9.56752449e-02 -9.54509616e-01
-2.00053886e-01 -4.67640281e-01 -1.08922887e+00 -7.34266281e-01
-2.64009893e-01 -4.69180137e-01 1.21657789e+00 8.20139825e-01
8.79835129e-01 7.36043990e-01 8.38351369e-01 -1.00594091e+00
-4.45186615e-01 -7.66387522e-01 -3.12820762e-01 -6.10397570e-02
8.48411143e-01 -8.10658514e-01 -2.80483902e-01 -1.29126728e-01] | [7.935519695281982, -1.2891311645507812] |
c3de6181-72a2-40c3-b858-7e280bfa0985 | imface-a-nonlinear-3d-morphable-face-model | 2203.14510 | null | https://arxiv.org/abs/2203.14510v2 | https://arxiv.org/pdf/2203.14510v2.pdf | ImFace: A Nonlinear 3D Morphable Face Model with Implicit Neural Representations | Precise representations of 3D faces are beneficial to various computer vision and graphics applications. Due to the data discretization and model linearity, however, it remains challenging to capture accurate identity and expression clues in current studies. This paper presents a novel 3D morphable face model, namely ImFace, to learn a nonlinear and continuous space with implicit neural representations. It builds two explicitly disentangled deformation fields to model complex shapes associated with identities and expressions, respectively, and designs an improved learning strategy to extend embeddings of expressions to allow more diverse changes. We further introduce a Neural Blend-Field to learn sophisticated details by adaptively blending a series of local fields. In addition to ImFace, an effective preprocessing pipeline is proposed to address the issue of watertight input requirement in implicit representations, enabling them to work with common facial surfaces for the first time. Extensive experiments are performed to demonstrate the superiority of ImFace. | ['Liming Chen', 'Di Huang', 'Hongyu Yang', 'Mingwu Zheng'] | 2022-03-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zheng_ImFace_A_Nonlinear_3D_Morphable_Face_Model_With_Implicit_Neural_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zheng_ImFace_A_Nonlinear_3D_Morphable_Face_Model_With_Implicit_Neural_CVPR_2022_paper.pdf | cvpr-2022-1 | ['face-model'] | ['computer-vision'] | [ 8.80654305e-02 1.29870951e-01 -1.60431728e-01 -6.44396842e-01
-2.31498063e-01 -3.99664670e-01 4.86232460e-01 -5.87703228e-01
2.19956025e-01 4.24812287e-01 9.37360823e-02 1.71367958e-01
-4.66959132e-03 -7.48391986e-01 -4.76102978e-01 -5.85003495e-01
-6.63109943e-02 1.47714645e-01 -4.67963070e-01 -2.74070770e-01
-2.74049444e-03 9.72596407e-01 -1.66224289e+00 1.85166016e-01
7.17357755e-01 1.06618488e+00 -3.30402762e-01 2.60373861e-01
-3.25006366e-01 4.40569520e-01 -1.43194005e-01 -5.68571985e-01
5.14734864e-01 -2.16492057e-01 -3.79701465e-01 2.77343214e-01
9.87489283e-01 -6.15434647e-01 -3.39607596e-01 9.52155709e-01
3.81864786e-01 5.90457171e-02 7.66531289e-01 -1.30074120e+00
-1.27937782e+00 -6.85221255e-02 -5.48621058e-01 -3.74355942e-01
2.05558896e-01 1.43579304e-01 8.38659108e-01 -1.30976963e+00
5.70897937e-01 1.66068411e+00 8.75527859e-01 9.51570809e-01
-1.38025057e+00 -9.68922019e-01 3.13537478e-01 -1.83411464e-01
-1.59048641e+00 -6.49052739e-01 1.19847357e+00 -4.14621025e-01
5.74490368e-01 2.66044527e-01 7.18390822e-01 1.09920394e+00
-1.57951131e-01 6.16278946e-01 1.21620762e+00 -1.60106614e-01
1.11913809e-03 1.04421929e-01 -1.11443125e-01 1.02224946e+00
1.38792157e-01 5.05772121e-02 -4.95200604e-01 -2.40583614e-01
1.41546595e+00 1.83769226e-01 -2.75462508e-01 -5.29842257e-01
-5.94673395e-01 7.60227740e-01 5.23384213e-01 4.82415222e-02
-2.70020694e-01 2.19559133e-01 -2.01840065e-02 1.27220407e-01
7.63175070e-01 3.41243923e-01 -4.54575688e-01 1.13488525e-01
-7.50253022e-01 4.83360976e-01 5.74131131e-01 9.85468447e-01
1.23015416e+00 2.96539277e-01 8.29723403e-02 9.46157932e-01
4.52608347e-01 4.32107151e-01 2.34023258e-02 -1.23409021e+00
9.97687504e-02 8.45147669e-01 -1.39026061e-01 -1.32252717e+00
-4.93787304e-02 -3.10671240e-01 -1.01265740e+00 6.18441820e-01
1.73170984e-01 2.29218870e-01 -1.07867897e+00 1.92907727e+00
4.65343267e-01 6.29743516e-01 -2.50101119e-01 8.79881203e-01
8.86851370e-01 5.15393317e-01 -8.55717361e-02 8.01570565e-02
1.16621327e+00 -5.91845572e-01 -7.84149587e-01 3.35322209e-02
3.09494197e-01 -4.48720127e-01 1.06976402e+00 1.12111501e-01
-1.16410112e+00 -6.15710258e-01 -9.17463541e-01 -5.57750106e-01
-2.76312262e-01 5.24184294e-02 8.12045395e-01 4.04566258e-01
-1.24227941e+00 6.21815860e-01 -9.33174729e-01 1.01248091e-02
8.32840860e-01 6.64272964e-01 -6.99464977e-01 -1.65381119e-01
-1.02263772e+00 6.68612957e-01 -2.70027697e-01 4.25775588e-01
-3.76426280e-01 -9.87310231e-01 -1.18552709e+00 -9.63371620e-02
-9.88055021e-02 -7.43014514e-01 7.54201710e-01 -1.07726216e+00
-1.79801953e+00 1.14514065e+00 -2.57454544e-01 2.91135013e-01
4.58383828e-01 -2.14746535e-01 -1.23872250e-01 -8.75168294e-02
-2.92681903e-01 7.86547244e-01 1.11016178e+00 -1.35766602e+00
1.15340956e-01 -7.17297137e-01 1.11184746e-01 1.76076442e-01
-5.07360518e-01 -1.37184247e-01 -5.54966509e-01 -8.67308676e-01
7.24092498e-02 -7.70678163e-01 -9.08201039e-02 9.21612561e-01
-3.74704562e-02 -1.49496272e-01 1.07186401e+00 -6.80624545e-01
8.86758685e-01 -2.25435805e+00 5.38968861e-01 2.22272158e-01
5.17914176e-01 2.81963229e-01 -4.15539742e-01 3.87649536e-02
2.74501108e-02 3.15371573e-01 -3.82947236e-01 -7.54940569e-01
1.08757228e-01 4.80750084e-01 -2.47732252e-01 4.57794815e-01
8.24456334e-01 1.11360347e+00 -5.42968988e-01 -3.09666514e-01
-6.08673841e-02 1.09514499e+00 -8.40393543e-01 3.37539405e-01
-1.64035872e-01 5.70748568e-01 -5.31424582e-01 8.36818337e-01
1.22095418e+00 9.53128934e-03 -8.38811547e-02 -4.70754355e-01
4.73973304e-02 -7.29217455e-02 -1.06712103e+00 1.77054107e+00
-4.32071269e-01 4.23449934e-01 6.47334874e-01 -7.94937551e-01
1.22674191e+00 1.46481454e-01 3.59203786e-01 -5.16886353e-01
2.66826868e-01 1.76195905e-01 -5.51891088e-01 -3.74479264e-01
3.06003153e-01 -3.89622867e-01 1.77169800e-01 1.13520421e-01
7.57969320e-02 -3.45975935e-01 -6.05738819e-01 -2.02548146e-01
6.45999968e-01 4.84075665e-01 -1.09876171e-02 -2.89893746e-01
5.68629444e-01 -5.93523860e-01 8.98097694e-01 -1.80400804e-01
-4.34479080e-02 6.14193439e-01 1.93790421e-01 -7.05921054e-01
-6.70678616e-01 -1.08684635e+00 -2.68817604e-01 7.65785158e-01
1.78519741e-01 -3.03168327e-01 -6.53773189e-01 -5.72108507e-01
3.69033217e-01 3.00360054e-01 -9.23250973e-01 -1.79104775e-01
-7.25310266e-01 -3.17892134e-01 5.55781126e-01 5.75681448e-01
4.59857643e-01 -6.48424566e-01 -9.04925019e-02 -1.94130525e-01
3.08256716e-01 -9.66601253e-01 -6.32086337e-01 -3.93627465e-01
-9.36314702e-01 -9.39403832e-01 -5.36626220e-01 -1.05207336e+00
1.00717080e+00 4.07151245e-02 7.30134964e-01 3.84802043e-01
-3.50827068e-01 3.56845051e-01 2.25687423e-03 -1.85599059e-01
-2.35650703e-01 -2.95135647e-01 1.55481115e-01 3.79325867e-01
2.78534144e-01 -9.09936190e-01 -5.29452801e-01 1.97660461e-01
-9.74901497e-01 1.21120177e-01 4.05694485e-01 8.85549426e-01
6.31637812e-01 -3.04652125e-01 4.11399335e-01 -8.81586611e-01
5.49212992e-01 -3.35839748e-01 -4.63373303e-01 1.43224880e-01
-3.83748204e-01 1.82888493e-01 3.93262833e-01 -6.97241306e-01
-1.13634872e+00 1.50476238e-02 -2.26002291e-01 -9.72228348e-01
-8.07346478e-02 2.75062889e-01 -5.17704785e-01 -5.44267654e-01
4.08354044e-01 -4.43860814e-02 4.84509110e-01 -7.51427948e-01
6.16868615e-01 4.78279650e-01 4.94199216e-01 -8.21550131e-01
1.06801724e+00 4.84486461e-01 6.82635903e-02 -9.00979936e-01
-5.15282989e-01 9.99825224e-02 -8.25490952e-01 -5.42427273e-03
5.96800208e-01 -9.74016368e-01 -5.07566154e-01 7.28663325e-01
-1.20539045e+00 -2.63800681e-01 -2.69119531e-01 1.68718442e-01
-4.77909625e-01 2.92999953e-01 -7.67312169e-01 -6.16616786e-01
-2.63077974e-01 -9.50330436e-01 1.19328630e+00 2.76181698e-01
-1.55192301e-01 -1.05781531e+00 5.69638833e-02 1.39653265e-01
4.27169889e-01 6.69833004e-01 9.62714493e-01 4.32195654e-03
-6.71961129e-01 -1.45766707e-02 -2.75294334e-01 4.88622099e-01
7.17516243e-01 3.10792446e-01 -1.18200445e+00 -3.93218756e-01
2.82027032e-02 -3.35972071e-01 6.45013571e-01 -6.41029775e-02
1.43733680e+00 -5.91542959e-01 -1.75236374e-01 1.32717347e+00
1.17784715e+00 -1.30438581e-01 6.01026177e-01 -2.19472125e-01
1.01121247e+00 6.38939559e-01 1.91551760e-01 2.89816856e-01
5.64050257e-01 6.55564964e-01 4.58913237e-01 -3.13187718e-01
-3.46574426e-01 -4.71265614e-01 3.54265451e-01 1.03642964e+00
-3.11369985e-01 3.36264372e-01 -4.56029713e-01 2.86811084e-01
-1.42878842e+00 -7.32343912e-01 3.65958005e-01 1.80416095e+00
9.90638673e-01 -3.21754962e-01 -3.12130511e-01 -1.37455285e-01
4.98574257e-01 1.68353274e-01 -7.62643576e-01 -4.49345529e-01
-1.08205847e-01 7.64051974e-01 -1.00470133e-01 5.72629571e-01
-9.17658150e-01 1.02953899e+00 6.17689419e+00 5.86779058e-01
-1.50500929e+00 -4.62536253e-02 5.74844778e-01 5.94312139e-02
-7.76452720e-01 -1.24174349e-01 -6.01796746e-01 1.57407105e-01
2.78342396e-01 4.60007749e-02 5.72287440e-01 7.07854390e-01
-9.77477431e-02 7.41210878e-01 -1.18751800e+00 1.24672258e+00
1.78361595e-01 -1.41954279e+00 5.23172557e-01 1.86695039e-01
7.07651436e-01 -5.80308259e-01 3.46282423e-01 2.29744434e-01
1.24395341e-02 -1.39667666e+00 5.60495913e-01 7.33756781e-01
1.35658276e+00 -5.91754973e-01 2.02507600e-01 1.15926638e-01
-1.25841022e+00 1.14223942e-01 -2.40433857e-01 -2.34726995e-01
-5.22327945e-02 1.00130618e-01 -5.30182183e-01 4.02931541e-01
5.20223141e-01 1.04276168e+00 -3.56015056e-01 4.39304233e-01
-1.87665880e-01 2.16114387e-01 -2.35569552e-01 2.81945109e-01
-2.10761473e-01 -5.19202352e-01 3.06485265e-01 8.31565619e-01
4.66828555e-01 3.83167058e-01 -4.62622084e-02 1.30736887e+00
-4.32940423e-01 1.60970527e-03 -7.18096912e-01 -7.75253922e-02
4.37787980e-01 1.37657094e+00 -6.76824898e-02 1.04653589e-01
-4.65977937e-01 1.04412413e+00 7.10189581e-01 5.46811044e-01
-6.12770617e-01 -1.18722431e-01 1.39230049e+00 3.96676123e-01
1.83017999e-01 -6.17636383e-01 -2.18853623e-01 -1.37069702e+00
2.58164480e-02 -8.04904878e-01 -1.37320399e-01 -5.47939062e-01
-1.61416531e+00 5.95601678e-01 -8.07557404e-02 -9.29188848e-01
6.08352497e-02 -8.52497756e-01 -5.69008648e-01 1.02271676e+00
-1.72112906e+00 -1.74774909e+00 -5.02902865e-01 6.93899393e-01
1.56274453e-01 3.90619561e-02 1.07346988e+00 3.33835363e-01
-5.45412064e-01 1.00444186e+00 -3.13568443e-01 2.00540558e-01
5.82999229e-01 -7.92913079e-01 4.87083852e-01 4.02104974e-01
9.04765502e-02 9.00617182e-01 1.78402767e-01 -4.76689816e-01
-1.80967343e+00 -1.27352154e+00 4.46431875e-01 -4.56164986e-01
2.32734591e-01 -6.66535676e-01 -1.34353006e+00 8.16537201e-01
-5.18464819e-02 4.85679418e-01 8.30512941e-01 1.07650447e-03
-7.38538325e-01 -2.95355022e-01 -1.29277563e+00 6.43663704e-01
1.49147022e+00 -9.18440402e-01 -4.74875271e-01 -2.77469039e-01
6.94323540e-01 -5.39380312e-01 -1.08277428e+00 7.20736563e-01
7.33649015e-01 -7.47339547e-01 1.04248035e+00 -7.27885544e-01
2.22510412e-01 -2.58507729e-01 -1.28489330e-01 -1.23129654e+00
-5.86886823e-01 -7.48223007e-01 -5.29292226e-01 1.46496415e+00
-1.04310147e-01 -7.88602293e-01 9.02477384e-01 8.76085162e-01
-2.24172100e-02 -1.12351787e+00 -8.84936333e-01 -4.33493525e-01
3.89085442e-01 -6.46462888e-02 1.09620094e+00 1.32307720e+00
-4.81957614e-01 1.55833550e-02 -3.55579138e-01 2.07118541e-01
5.45294046e-01 2.14990094e-01 8.90833437e-01 -1.46878135e+00
-4.00219969e-02 -4.98792291e-01 -6.80264413e-01 -1.23136258e+00
5.93959570e-01 -1.13337100e+00 -2.37079278e-01 -1.09873331e+00
-9.91952568e-02 -7.32184470e-01 -2.56611377e-01 6.21252358e-01
-1.57249644e-01 4.06719744e-01 7.66982958e-02 4.55064774e-02
1.27257228e-01 1.16995931e+00 1.65945232e+00 -2.58996636e-01
-1.02389753e-01 -2.43307367e-01 -9.03121114e-01 7.53734231e-01
6.03420496e-01 2.94308290e-02 -5.73895812e-01 -7.36176550e-01
-3.89375724e-02 -3.34012836e-01 5.34029305e-01 -4.60326463e-01
-9.53304172e-02 -4.13838148e-01 6.21523440e-01 -8.39501545e-02
7.10151076e-01 -7.30368674e-01 3.36047709e-01 -5.37663512e-02
-1.76692590e-01 1.96642075e-02 4.22206759e-01 3.82064134e-01
-2.25402892e-01 2.99418449e-01 7.39443958e-01 1.56975668e-02
-6.01335406e-01 9.68993664e-01 2.53229111e-01 -4.89003882e-02
9.65187073e-01 -3.87183636e-01 -1.40699735e-02 -2.51600325e-01
-5.80283105e-01 4.63701971e-02 9.82506216e-01 5.78569770e-01
9.49849129e-01 -1.86642253e+00 -6.98601961e-01 9.14982378e-01
1.04569718e-01 4.04181331e-01 1.91545039e-01 3.79237503e-01
-5.96568167e-01 -3.12302530e-01 -4.50705111e-01 -3.92781943e-01
-1.22148895e+00 2.76357114e-01 4.45132047e-01 2.05761522e-01
-5.07114947e-01 1.02177072e+00 4.43489403e-01 -7.40964890e-01
5.00922985e-02 -3.92378002e-01 -6.32824982e-03 -1.47615954e-01
4.52880651e-01 1.55188292e-01 -1.57308623e-01 -9.66493785e-01
-2.66548485e-01 1.16584599e+00 5.15033454e-02 2.14903250e-01
1.38137412e+00 5.56952730e-02 -3.79058540e-01 1.83036715e-01
1.49080813e+00 1.64419696e-01 -1.68789577e+00 -3.21216077e-01
-3.09717208e-01 -8.01975131e-01 4.71741706e-02 -4.25361127e-01
-1.31974971e+00 1.18451631e+00 4.18601453e-01 -2.87360013e-01
1.20360398e+00 -1.87690124e-01 8.01232159e-01 3.13102193e-02
2.80601203e-01 -6.36497080e-01 1.20473251e-01 4.01563793e-01
1.24888980e+00 -9.14570212e-01 -1.26850568e-02 -5.75526297e-01
-2.61449546e-01 1.23196197e+00 8.22236896e-01 -2.16463655e-01
9.26502883e-01 3.32592905e-01 1.68122560e-01 -1.61542505e-01
-5.20274639e-01 2.69741565e-01 5.74702799e-01 7.49555469e-01
3.33274454e-01 -1.48902880e-02 1.35272786e-01 5.93777955e-01
-7.71868080e-02 8.98856074e-02 -6.14629909e-02 7.82740474e-01
1.70738306e-02 -1.40250945e+00 -1.12045899e-01 1.85902461e-01
-1.22559182e-01 2.24480137e-01 -4.96906042e-01 7.67031074e-01
4.27150428e-01 3.26635212e-01 2.06723481e-01 -4.66952145e-01
4.09056783e-01 6.79433271e-02 8.10118258e-01 -6.25365257e-01
-2.03664124e-01 -1.63268343e-01 -2.80138344e-01 -6.35083497e-01
-4.66614366e-01 -5.02601564e-01 -1.31679094e+00 -3.85993421e-01
-1.26377478e-01 -2.55128473e-01 4.96598005e-01 5.92441618e-01
7.82956719e-01 1.27290711e-01 7.81843185e-01 -1.32759225e+00
-6.37887478e-01 -5.98798573e-01 -6.33246601e-01 7.95655310e-01
5.49051166e-01 -1.10867178e+00 -3.23300928e-01 5.90021051e-02] | [12.952497482299805, -0.07988473027944565] |
7ba5dbaa-9d6a-4e03-b6c5-602a89617060 | a-novel-ecg-denoising-scheme-using-the | 2207.11819 | null | https://arxiv.org/abs/2207.11819v1 | https://arxiv.org/pdf/2207.11819v1.pdf | A Novel ECG Denoising Scheme Using the Ensemble Kalman Filter | Monitoring of electrocardiogram (ECG) provides vital information as well as any cardiovascular anomalies. Recent advances in the technology of wearable electronics have enabled compact devices to acquire personal physiological signals in the home setting; however, signals are usually contaminated with high level noise. Thus, an efficient ECG filtering scheme is a dire need. In this paper, a novel method using Ensemble Kalman Filter (EnKF) is developed for denoising ECG signals. We also intensively explore various filtering algorithms, including Savitzky-Golay (SG) filter, Ensemble Empirical mode decomposition (EEMD), Normalized Least-Mean-Square (NLMS), Recursive least squares (RLS) filter, Total variation denoising (TVD), Wavelet and extended Kalman filter (EKF) for comparison. Data from the MIT-BIH Noise Stress Test database were used. The proposed methodology shows the average signal to noise ratio (SNR) of 10.96, the Percentage Root Difference of 150.45, and the correlation coefficient of 0.959 from the modified MIT-BIH database with added motion artifacts. | ['Hung Cao', 'Tadesse Ghirmai', 'Manoj Vishwanath', 'Michael P. H. Lau', 'Samir Malhotra', 'Daniel Jilani', 'Hoang Vuong', 'Sadaf Sarafan'] | 2022-07-24 | null | null | null | null | ['ecg-denoising'] | ['medical'] | [ 3.57713193e-01 -4.31237668e-01 4.93851304e-01 -2.14069158e-01
-5.10204434e-01 -2.75833100e-01 -1.18268162e-01 1.83741361e-01
-5.26296377e-01 1.16717255e+00 1.14424646e-01 -1.78402156e-01
-4.71893966e-01 -2.31397852e-01 -3.36342342e-02 -9.36580300e-01
-3.72501761e-01 -6.19176209e-01 -1.98648334e-01 5.67489602e-02
2.18642615e-02 3.28696430e-01 -1.12617958e+00 -8.65965411e-02
1.14384282e+00 1.16148686e+00 2.38395315e-02 8.61106515e-01
6.88918531e-01 2.07065299e-01 -7.32673526e-01 -3.09260916e-02
1.45565256e-01 -9.08832252e-01 -2.56487697e-01 -2.67000854e-01
-3.21980000e-01 -4.31679547e-01 -6.32110834e-02 8.18162143e-01
1.12706053e+00 2.25098386e-01 5.89644551e-01 -6.79811239e-01
1.53161362e-01 9.38383043e-02 -5.30724525e-01 5.94062805e-01
3.43633890e-01 1.26499953e-02 -3.65509279e-02 -7.56013215e-01
4.05150920e-01 5.88949144e-01 1.26104236e+00 1.45911753e-01
-1.35728526e+00 -5.76272786e-01 -6.06517971e-01 2.94987589e-01
-1.60638487e+00 -3.82319331e-01 7.50810385e-01 -2.39541367e-01
6.85977638e-01 6.60861909e-01 8.78486574e-01 8.25767577e-01
9.49709117e-01 -7.13394135e-02 1.62780023e+00 -3.87743980e-01
9.75824594e-02 4.08308059e-02 3.51075828e-01 2.46540457e-01
6.69189692e-01 1.80692971e-01 -4.12149757e-01 -4.58057433e-01
8.31107855e-01 6.25920147e-02 -5.16505301e-01 2.30670020e-01
-1.20847046e+00 2.30272681e-01 -2.86858141e-01 5.99244714e-01
-9.10036266e-01 2.42115445e-02 6.07135296e-01 6.22102916e-01
2.98937678e-01 8.51810724e-02 -5.65850198e-01 -3.60713303e-01
-8.95516217e-01 1.40626296e-01 8.59593093e-01 4.69746232e-01
2.48767063e-01 3.56029898e-01 -2.43306994e-01 5.49850523e-01
4.95420158e-01 7.85537601e-01 6.30028903e-01 -9.50224280e-01
1.52333891e-02 3.98300774e-02 3.50487828e-01 -1.14496744e+00
-6.57623410e-01 -8.64883721e-01 -1.34707022e+00 -1.17815264e-01
2.97961622e-01 -6.36420012e-01 -2.50675380e-01 1.34112406e+00
3.31620514e-01 5.75027347e-01 9.99911800e-02 7.62045741e-01
7.02221811e-01 5.43472767e-01 1.32243410e-01 -1.05673194e+00
1.27929938e+00 3.67212631e-02 -1.42224193e+00 8.32196623e-02
3.08218360e-01 -8.07348967e-01 3.95577610e-01 1.03246427e+00
-1.00438559e+00 -8.48556995e-01 -1.02207553e+00 5.65445185e-01
2.23705679e-01 2.86450416e-01 1.28864497e-01 9.99080122e-01
-7.02285111e-01 1.13498962e+00 -8.71634543e-01 -3.37510824e-01
1.78295728e-02 2.10970104e-01 -2.75586605e-01 3.88999760e-01
-1.21978343e+00 8.05505395e-01 1.19446695e-01 4.41040397e-01
-3.49316180e-01 -4.65734631e-01 -4.02785420e-01 -1.67348400e-01
-1.62486255e-01 -6.04944885e-01 7.29507744e-01 -7.84904003e-01
-1.53705180e+00 4.23127055e-01 -2.41166025e-01 -3.96744549e-01
3.13770622e-01 -4.03198868e-01 -8.99627030e-01 1.14149466e-01
-3.87032807e-01 -6.19980991e-01 9.51497078e-01 -6.56165421e-01
-6.51441514e-02 -7.01862395e-01 -8.67610574e-01 8.77123699e-02
5.78102618e-02 -4.59329132e-03 4.49325144e-01 -7.82534480e-01
7.09690750e-01 -6.23893499e-01 -2.34984055e-01 -5.61908782e-01
-3.07720341e-02 3.59305561e-01 3.53534460e-01 -1.34368086e+00
1.85216391e+00 -2.47095346e+00 -1.29628241e-01 6.66554511e-01
-1.67806849e-01 3.20570976e-01 5.13844311e-01 5.08871436e-01
-1.10414073e-01 -2.24937230e-01 -2.65455157e-01 2.04023927e-01
-4.81299728e-01 1.78778306e-01 1.35458052e-01 8.87278497e-01
-4.52981025e-01 4.60423529e-01 -6.12441063e-01 -2.59991676e-01
3.17465156e-01 5.44717014e-01 -4.29871045e-02 2.00575903e-01
8.98941875e-01 1.06405246e+00 -2.45989352e-01 4.69064683e-01
7.10333467e-01 3.20173174e-01 2.05585763e-01 -7.46514678e-01
-2.09642321e-01 -1.81177422e-01 -1.91847324e+00 1.62634575e+00
-3.53451110e-02 3.43387157e-01 3.22265029e-01 -9.15974259e-01
1.34963369e+00 8.83501232e-01 5.39721787e-01 -3.69664192e-01
3.32489640e-01 2.83795804e-01 1.65072620e-01 -1.09718919e+00
-6.49741888e-02 -3.32500070e-01 4.59661365e-01 2.42821842e-01
-9.80896652e-02 2.99729049e-01 -1.32672921e-01 -2.39249200e-01
1.02072215e+00 1.96008459e-01 8.37981522e-01 -6.96635962e-01
7.14496136e-01 -5.51648557e-01 7.37582982e-01 6.82814598e-01
-4.57132310e-01 4.50984567e-01 3.08266990e-02 -3.13035250e-01
-6.42193019e-01 -8.24174821e-01 -4.12122637e-01 3.55245396e-02
-9.33684781e-02 -3.46382380e-01 -6.13155842e-01 2.15994194e-01
-1.63982049e-01 4.00643378e-01 -4.44580391e-02 -3.01538050e-01
-5.16166329e-01 -9.28434193e-01 6.85035646e-01 2.80769795e-01
5.85422754e-01 -7.85659671e-01 -1.02582133e+00 7.49340653e-01
-1.70032427e-01 -6.15399718e-01 -4.00427952e-02 2.34514251e-02
-1.59323537e+00 -8.44537973e-01 -6.48410618e-01 -1.74271509e-01
1.87592655e-01 -2.04745978e-01 8.34356725e-01 -5.41104451e-02
-5.80950856e-01 3.86536181e-01 -3.01297724e-01 -4.86006498e-01
-1.62608758e-01 -5.81977785e-01 4.89053428e-01 2.91525275e-01
3.80543798e-01 -9.58031058e-01 -9.78535235e-01 2.26001143e-01
-3.93481016e-01 -4.74480122e-01 5.43803275e-01 8.47829700e-01
5.35843194e-01 1.51487634e-01 1.02615690e+00 -5.72431922e-01
1.16051209e+00 -2.59329021e-01 -3.77277404e-01 -1.54341072e-01
-9.54700708e-01 -6.06190979e-01 4.33192194e-01 -2.97233224e-01
-9.49996471e-01 -1.45988027e-02 -3.55731845e-01 6.55699819e-02
-2.63643503e-01 4.35930103e-01 -2.02393249e-01 7.48319030e-02
8.92058790e-01 4.37475145e-01 2.17209354e-01 -5.91090858e-01
-2.77103096e-01 8.04839253e-01 7.71687508e-01 -3.33495617e-01
3.33772540e-01 1.11786678e-01 2.06903890e-01 -1.27143955e+00
-1.85573339e-01 -6.04555070e-01 -5.52033603e-01 -3.63456368e-01
9.38350499e-01 -9.68144953e-01 -8.37740004e-01 7.07476139e-01
-9.32670474e-01 1.32193089e-01 -9.97345969e-02 1.20785272e+00
-3.41767192e-01 6.59063935e-01 -6.18024349e-01 -1.42754674e+00
-9.49210227e-01 -6.24172688e-01 1.24261491e-01 1.92529902e-01
-6.02208495e-01 -7.01993346e-01 -1.47762736e-02 -5.42880408e-02
4.22591478e-01 1.07649863e+00 2.35941902e-01 -2.17201561e-01
1.32325515e-01 -4.60151106e-01 3.26638401e-01 8.19512606e-01
2.63171375e-01 -2.97236383e-01 -8.77376437e-01 -3.51078421e-01
7.76701927e-01 4.24794227e-01 2.05800220e-01 8.26671064e-01
7.78170824e-01 -1.06217332e-01 -1.19576655e-01 5.09767294e-01
1.51646793e+00 6.66679442e-01 1.10340548e+00 -4.96945996e-03
2.34866887e-01 2.35457756e-02 4.83533710e-01 8.46399128e-01
-2.25626737e-01 1.61207482e-01 -1.52057081e-01 -8.12004358e-02
3.88109177e-01 2.61589795e-01 2.23157197e-01 1.01794124e+00
-5.70295453e-01 9.95814279e-02 -4.90938723e-01 1.19502060e-01
-1.64692283e+00 -9.30850387e-01 -1.02647817e+00 2.59326959e+00
5.79253972e-01 2.64398661e-02 1.24334946e-01 9.37170744e-01
5.60819209e-01 -3.43295425e-01 -4.34406340e-01 -4.63140190e-01
-1.63642988e-01 4.85178053e-01 6.35769069e-01 2.88986832e-01
-9.00114119e-01 -2.31801391e-01 5.59033823e+00 3.71906132e-01
-7.99737692e-01 3.49366874e-01 2.84033656e-01 4.50154513e-01
3.76621664e-01 -5.43427877e-02 -3.12170953e-01 7.32087314e-01
1.35756314e+00 -2.03775123e-01 3.39616120e-01 4.45989192e-01
7.39990652e-01 -4.51852977e-01 -4.37207490e-01 1.53117514e+00
-1.13454655e-01 -5.06236374e-01 -8.53626549e-01 -3.65641452e-02
4.68430042e-01 -6.11536205e-01 -3.95198643e-01 -1.05968907e-01
-1.03285396e+00 -6.25988722e-01 1.39968559e-01 1.33070421e+00
7.19203889e-01 -8.79237294e-01 1.04761791e+00 4.79655415e-01
-1.07129860e+00 -6.70980290e-02 -3.39750320e-01 -4.04418588e-01
2.53026843e-01 1.32037377e+00 -2.09351748e-01 7.98501670e-01
6.14128172e-01 7.07969964e-01 -1.66606441e-01 1.26765633e+00
3.60356048e-02 9.55443501e-01 -5.09089351e-01 2.02662230e-01
-4.68571454e-01 -4.80113298e-01 8.14707339e-01 8.91111732e-01
6.18072569e-01 9.51968431e-01 -2.04541594e-01 3.82664591e-01
5.68557680e-01 3.77902955e-01 -2.67422050e-01 3.41848910e-01
3.63415599e-01 1.04979384e+00 -5.90004027e-01 -4.54744965e-01
-2.86319762e-01 7.07634330e-01 -8.25088024e-01 2.19196782e-01
-6.03940487e-01 -7.69917071e-01 3.10058862e-01 1.44386500e-01
-3.51438493e-01 -2.81740338e-01 -5.63150227e-01 -9.04857397e-01
-5.22568077e-02 -1.08398378e+00 3.08700234e-01 -5.65389097e-01
-1.06657720e+00 4.57281291e-01 -5.59906624e-02 -1.38335311e+00
-3.29030305e-02 -1.67517602e-01 -5.66854239e-01 1.50258243e+00
-7.28995442e-01 -7.06441849e-02 -4.28169042e-01 7.16744959e-01
1.56312123e-01 7.67642558e-02 1.03142011e+00 7.02337503e-01
-4.13812488e-01 3.51289630e-01 2.10569814e-01 -3.45371634e-01
6.28046572e-01 -1.03475523e+00 -2.54590660e-01 8.44348252e-01
-5.69363952e-01 9.07673240e-01 1.07263517e+00 -7.60489941e-01
-1.40668714e+00 -5.87683797e-01 9.39195216e-01 -3.25690722e-03
7.56650642e-02 2.04318479e-01 -9.81052876e-01 1.12012714e-01
1.70143306e-01 -1.12659261e-01 7.59495437e-01 -2.80254513e-01
6.33714616e-01 -6.46536112e-01 -1.45449996e+00 1.04598910e-01
4.62329924e-01 -3.02544713e-01 -6.34628534e-01 -7.52943978e-02
-1.16817459e-01 -3.56276482e-01 -1.61156654e+00 6.42432392e-01
1.09874260e+00 -1.12567925e+00 7.73501515e-01 -1.86791211e-01
-3.54779065e-01 -4.90815967e-01 7.30404407e-02 -1.08708668e+00
-3.82504433e-01 -1.11288273e+00 -1.61808506e-01 1.27683485e+00
4.91143502e-02 -9.55541492e-01 2.84296870e-01 5.67358792e-01
-1.13134496e-02 -6.90033674e-01 -8.12036693e-01 -6.89270258e-01
-7.83131003e-01 -3.21355343e-01 8.10824558e-02 6.13470912e-01
1.61571205e-01 3.32673669e-01 -8.43099177e-01 4.45134118e-02
8.67731035e-01 -4.93912578e-01 4.41074699e-01 -1.45663726e+00
-4.32432026e-01 4.65075791e-01 -5.05852580e-01 -5.87313890e-01
-6.13326490e-01 -2.43000969e-01 -3.74895453e-01 -1.50301337e+00
-2.29974836e-01 3.45882550e-02 -7.71087646e-01 -2.21654221e-01
-3.08165759e-01 2.02131405e-01 -1.26236781e-01 1.56574920e-02
9.80772823e-02 -1.66992692e-03 9.25662398e-01 4.42631036e-01
-6.59362495e-01 5.90324640e-01 -3.44629973e-01 6.77763343e-01
8.27483296e-01 -4.94182825e-01 -4.77201521e-01 3.20752025e-01
1.80868655e-01 6.34814203e-01 1.06269486e-01 -1.35184872e+00
-7.48391403e-03 4.49416846e-01 9.02144074e-01 -8.35774541e-01
2.97368675e-01 -9.41588640e-01 1.08817708e+00 9.95639443e-01
8.09227675e-02 3.54987860e-01 1.05264291e-01 5.28796196e-01
-8.82405415e-02 -2.19905257e-01 7.03829765e-01 -2.57228166e-01
-1.34502918e-01 -1.18561037e-01 -7.74726033e-01 -2.84143567e-01
8.07954907e-01 -6.78768873e-01 1.63020283e-01 -2.52959281e-01
-1.20330834e+00 -1.63750306e-01 -2.16841564e-01 -3.38145047e-01
7.50874221e-01 -1.17935371e+00 -7.69397616e-01 4.72985595e-01
-3.22342902e-01 -4.86463785e-01 8.41296852e-01 1.64889050e+00
-5.50347745e-01 -8.39150045e-03 -3.26155692e-01 -4.85130996e-01
-1.47429907e+00 8.71121809e-02 3.01722139e-01 3.26700024e-02
-6.72073245e-01 5.87397158e-01 -6.24568284e-01 5.20963490e-01
7.76301026e-02 -1.98965788e-01 -4.54477996e-01 1.79600641e-01
7.28037536e-01 1.07917356e+00 2.56350309e-01 -3.88684750e-01
-4.43392098e-01 7.23118424e-01 6.31155968e-01 -2.71652471e-02
9.99693573e-01 -5.60497761e-01 -3.28014642e-01 6.61845207e-01
9.41190898e-01 -3.96624655e-02 -6.62297547e-01 1.33674875e-01
-6.58440683e-03 -3.25344563e-01 1.27819136e-01 -8.45851541e-01
-6.08920157e-01 7.27257967e-01 1.44981921e+00 2.60346085e-01
1.67020154e+00 -9.73800957e-01 8.90462518e-01 2.95079529e-01
3.86585772e-01 -1.10184729e+00 -4.68654096e-01 -8.08804333e-02
5.67671835e-01 -4.80177999e-01 3.45716536e-01 -1.34479374e-01
-4.19923633e-01 1.20560706e+00 -4.22944017e-02 -3.98003459e-01
1.14222956e+00 2.35284433e-01 1.55014604e-01 1.55952245e-01
-3.74809235e-01 6.54825047e-02 -4.81150579e-04 6.37845755e-01
7.68966138e-01 1.94324926e-02 -1.36939657e+00 9.33300376e-01
1.40518710e-01 6.24528825e-01 4.52834100e-01 1.03044271e+00
-5.24159729e-01 -7.22034812e-01 -7.70983994e-01 7.89510310e-01
-1.10516310e+00 1.39378875e-01 3.10105503e-01 3.13212126e-01
2.31223673e-01 1.26388371e+00 -4.13095117e-01 -2.19100788e-01
6.05928779e-01 2.26338968e-01 6.38613105e-01 -1.24824412e-01
-8.07663918e-01 6.53781295e-01 2.26415172e-02 -5.34748733e-01
-5.14875591e-01 -7.70982563e-01 -9.71296132e-01 1.34111628e-01
-4.28075016e-01 3.83470297e-01 8.81124794e-01 8.67226005e-01
2.03144774e-01 7.60723293e-01 4.60133463e-01 -4.31469738e-01
-5.76379597e-01 -1.31427157e+00 -1.24380076e+00 2.38426387e-01
3.85943502e-01 -3.41566741e-01 -4.10760731e-01 3.43635708e-01] | [14.044398307800293, 3.0761325359344482] |
d2272059-ddc6-4c10-9fa3-52b7a70702d9 | bidirectional-attention-flow-for-machine | 1611.01603 | null | http://arxiv.org/abs/1611.01603v6 | http://arxiv.org/pdf/1611.01603v6.pdf | Bidirectional Attention Flow for Machine Comprehension | Machine comprehension (MC), answering a query about a given context
paragraph, requires modeling complex interactions between the context and the
query. Recently, attention mechanisms have been successfully extended to MC.
Typically these methods use attention to focus on a small portion of the
context and summarize it with a fixed-size vector, couple attentions
temporally, and/or often form a uni-directional attention. In this paper we
introduce the Bi-Directional Attention Flow (BIDAF) network, a multi-stage
hierarchical process that represents the context at different levels of
granularity and uses bi-directional attention flow mechanism to obtain a
query-aware context representation without early summarization. Our
experimental evaluations show that our model achieves the state-of-the-art
results in Stanford Question Answering Dataset (SQuAD) and CNN/DailyMail cloze
test. | ['Ali Farhadi', 'Minjoon Seo', 'Hannaneh Hajishirzi', 'Aniruddha Kembhavi'] | 2016-11-05 | null | null | null | null | ['cloze-test'] | ['natural-language-processing'] | [ 7.06652030e-02 -1.34085402e-01 -8.73952061e-02 -6.00775063e-01
-1.09355378e+00 -5.32499969e-01 5.18904150e-01 5.73382556e-01
-4.49967265e-01 5.99814892e-01 8.10487390e-01 -4.58438843e-01
4.66894284e-02 -6.26774371e-01 -7.13256776e-01 -1.96110979e-01
9.40909088e-02 2.91736990e-01 5.37389219e-01 -2.14560047e-01
4.60098267e-01 1.31470971e-02 -1.12768793e+00 8.24144900e-01
9.84365702e-01 6.81186080e-01 5.55650830e-01 1.22517347e+00
-6.73905730e-01 1.24664855e+00 -8.72506320e-01 -3.91057372e-01
-5.53965569e-01 -5.89548707e-01 -1.51182461e+00 -3.35193932e-01
8.29838812e-01 -4.07515615e-01 -3.39415580e-01 8.03421855e-01
4.69771147e-01 3.79989445e-01 2.10300580e-01 -6.80615723e-01
-1.09100473e+00 8.07284355e-01 -5.35097122e-01 9.85048831e-01
7.77014256e-01 1.20764293e-01 1.36247408e+00 -8.98867130e-01
5.04179001e-01 1.44278133e+00 3.15497756e-01 4.05661792e-01
-1.06054676e+00 -3.65412921e-01 8.38813305e-01 7.13464797e-01
-7.76575625e-01 4.93323617e-02 8.42093766e-01 -2.23924056e-01
1.39440000e+00 4.49839801e-01 4.09057766e-01 1.18188882e+00
3.15924406e-01 1.10107291e+00 6.39049351e-01 -3.19356620e-01
1.50625512e-01 -4.50356394e-01 1.11402905e+00 4.08217251e-01
-2.25020945e-01 -5.35733044e-01 -4.97854233e-01 -2.58289009e-01
2.11587578e-01 -9.62425843e-02 -5.05716205e-01 4.59871799e-01
-1.21763790e+00 7.71988213e-01 8.22258651e-01 4.26285148e-01
-6.24214292e-01 4.01170522e-01 4.26071256e-01 3.42216730e-01
4.61773068e-01 4.05575335e-01 -4.27912921e-01 -1.00711353e-01
-6.58606827e-01 3.62955779e-01 8.39260459e-01 1.10752928e+00
6.66970611e-01 -3.86713177e-01 -9.62303400e-01 7.57335901e-01
1.47713155e-01 8.28161910e-02 5.12148440e-01 -7.79398501e-01
9.19078052e-01 6.44558191e-01 6.33921027e-02 -9.78624165e-01
-2.60909319e-01 -3.58372301e-01 -7.61394918e-01 -5.63553572e-01
-9.32132900e-02 -1.70171320e-01 -9.77782667e-01 1.85917401e+00
2.72694439e-01 3.42774540e-01 2.14264113e-02 9.70628381e-01
1.29491770e+00 1.14163375e+00 6.18524492e-01 -2.41610721e-01
1.64930677e+00 -1.49624765e+00 -1.05001831e+00 -4.44875032e-01
1.64243355e-01 -5.32005370e-01 1.44819677e+00 1.40865251e-01
-1.25990939e+00 -8.17967772e-01 -9.13632989e-01 -9.98463333e-01
-4.17538226e-01 -3.25726777e-01 3.66183668e-01 -9.98933986e-03
-9.73095655e-01 2.99152225e-01 -5.13179183e-01 -4.07094896e-01
3.19824547e-01 1.66373104e-01 -1.52813205e-02 -3.32303017e-01
-1.48035800e+00 6.71351910e-01 1.03308283e-01 2.33634725e-01
-9.43818390e-01 -6.76777899e-01 -7.14302778e-01 5.10320067e-01
4.81678993e-01 -1.09854150e+00 1.61768019e+00 -7.86310613e-01
-1.13922918e+00 5.58473527e-01 -7.19106734e-01 -3.08687001e-01
3.38645689e-02 -7.75285363e-01 -2.79268354e-01 3.40227425e-01
2.81481475e-01 6.86455369e-01 5.50175250e-01 -1.04298270e+00
-6.47426605e-01 -3.44423681e-01 3.97459209e-01 3.38530719e-01
1.72222257e-01 2.28372544e-01 -8.06946874e-01 -5.96294641e-01
-1.47003904e-01 -3.49621147e-01 -1.51132002e-01 -6.29884541e-01
-3.59490365e-01 -8.06029558e-01 1.04446495e+00 -9.80474234e-01
1.90430236e+00 -1.93703985e+00 4.67755884e-01 -5.70917904e-01
1.15828916e-01 2.15955615e-01 -2.85695553e-01 7.80849695e-01
-6.38210252e-02 3.52592915e-01 -4.64159220e-01 -5.50096512e-01
-2.39704385e-01 2.48333201e-01 -6.44475162e-01 -1.93559110e-01
4.19999808e-01 1.26375663e+00 -1.27897930e+00 -4.81102228e-01
-1.83341190e-01 2.33255073e-01 -6.63119495e-01 5.63023388e-01
-7.96269417e-01 2.06080705e-01 -6.06510937e-01 4.77398127e-01
3.47184420e-01 -4.39087927e-01 -1.37289613e-01 -7.43267387e-02
6.39323369e-02 5.03631294e-01 -5.66285253e-01 1.99495745e+00
-4.86497790e-01 7.74671018e-01 1.63994923e-01 -7.12101042e-01
5.34015238e-01 2.69064903e-01 -2.58088529e-01 -5.68608701e-01
6.73077330e-02 -3.32538933e-01 -1.65602058e-01 -1.11237502e+00
7.69016683e-01 1.98582038e-01 -2.03498155e-01 3.53682905e-01
4.98294383e-02 1.46227151e-01 2.19844982e-01 4.42840934e-01
1.05851960e+00 1.53241744e-02 2.92174935e-01 -2.43236467e-01
8.69069934e-01 4.85438332e-02 3.57510805e-01 1.01837349e+00
-3.02213788e-01 7.41874635e-01 8.77374351e-01 -3.10813397e-01
-6.58691764e-01 -7.20944941e-01 4.17016864e-01 1.72852993e+00
1.85016826e-01 -4.23572659e-01 -7.50021398e-01 -8.72406304e-01
-6.39630258e-02 1.04424727e+00 -9.36281204e-01 -8.74656439e-02
-9.43130851e-01 -3.47957224e-01 2.55104899e-01 7.36681223e-01
8.16091597e-01 -1.48770487e+00 -5.75403392e-01 4.96759027e-01
-4.62478310e-01 -8.49035919e-01 -8.75435293e-01 2.20855102e-02
-8.94984424e-01 -8.62385392e-01 -6.64467812e-01 -1.16391993e+00
2.82089919e-01 3.50615144e-01 1.51403439e+00 3.20831060e-01
-6.78889640e-03 3.39562058e-01 -3.99017006e-01 -2.54758298e-01
8.61836821e-02 3.24244380e-01 -7.85817981e-01 5.03525995e-02
4.79435861e-01 -5.42334080e-01 -8.16533804e-01 1.23323351e-02
-9.76262927e-01 -6.95051206e-03 3.77518028e-01 9.38357711e-01
7.24154472e-01 -6.06942534e-01 1.09161520e+00 -1.15230370e+00
1.18308651e+00 -8.81017089e-01 -1.60754025e-01 6.22622073e-01
-4.74005938e-02 3.27091827e-03 6.54366434e-01 -4.16343600e-01
-1.24562562e+00 -4.28286165e-01 -1.94953203e-01 -3.87076110e-01
-2.05995202e-01 8.67910683e-01 -1.63112298e-01 7.48222053e-01
5.18015444e-01 2.02498794e-01 -6.57014191e-01 -5.24000287e-01
7.58416355e-01 6.97192073e-01 7.88348794e-01 -5.81542790e-01
-1.42384777e-02 9.17713568e-02 -5.59229434e-01 -6.33651078e-01
-1.24320281e+00 -6.89920008e-01 -3.62819135e-01 -1.07406527e-01
1.40804374e+00 -6.14038408e-01 -7.41408408e-01 1.46855384e-01
-1.56073916e+00 -3.13960463e-01 4.69623366e-03 3.60755511e-02
-2.04376683e-01 4.70116168e-01 -7.77285457e-01 -7.24799812e-01
-8.19708705e-01 -1.03119206e+00 1.22461760e+00 5.71113288e-01
-3.17520291e-01 -8.58842850e-01 1.25550047e-01 5.19287825e-01
5.01139045e-01 7.73113593e-02 1.16350973e+00 -6.98872924e-01
-7.81356931e-01 -1.51241198e-02 -3.52112055e-01 4.64317203e-03
-4.96203154e-02 -1.18407898e-01 -1.02007759e+00 -1.75957575e-01
1.23473518e-01 -3.29059005e-01 1.24091101e+00 2.91466594e-01
1.63719940e+00 -3.57258737e-01 -3.05611461e-01 2.37788543e-01
1.25509894e+00 2.44176850e-01 4.83042657e-01 -8.99507776e-02
5.89600623e-01 5.01117170e-01 4.54756737e-01 -2.08292156e-02
6.46259904e-01 1.56055495e-01 4.70511168e-01 3.57930601e-01
-1.44090429e-01 -3.43120337e-01 -8.61248299e-02 9.20644820e-01
3.61690551e-01 -6.56350732e-01 -7.99882472e-01 8.87943089e-01
-1.98555803e+00 -1.11807704e+00 -2.80232310e-01 1.75200176e+00
8.65273595e-01 1.53196827e-01 -1.29891977e-01 -4.31064397e-01
8.52379441e-01 4.91433322e-01 -7.35669553e-01 -6.59500062e-01
5.46614826e-02 -4.66164574e-02 -2.49978215e-01 1.05456376e+00
-9.57487822e-01 1.05948436e+00 6.68950081e+00 5.11914194e-01
-9.51424241e-01 2.14703456e-01 8.07774007e-01 -9.80622321e-02
-5.55582345e-01 -4.58172895e-02 -5.77246964e-01 4.46147799e-01
8.74560356e-01 -2.32063085e-01 2.11974144e-01 7.40880430e-01
5.15823392e-03 -3.56690198e-01 -1.19942391e+00 6.84266627e-01
2.04289928e-01 -1.26247597e+00 3.28015387e-01 -4.47315156e-01
6.21787906e-01 -1.27382306e-02 -1.84021726e-01 8.64876390e-01
-5.57708405e-02 -1.01280093e+00 2.62614131e-01 6.72862828e-01
5.41346371e-01 -6.91493213e-01 7.98834622e-01 5.45083165e-01
-1.04665077e+00 -2.49509275e-01 -3.09462488e-01 -1.74404591e-01
4.71884131e-01 1.51643485e-01 -4.74272400e-01 6.89028263e-01
8.20466161e-01 4.47480977e-01 -7.29137838e-01 8.77814651e-01
-3.05420011e-01 9.72968757e-01 -1.01031423e-01 -5.74948668e-01
5.68464637e-01 2.46671345e-02 6.01242423e-01 1.59286153e+00
-2.23370716e-01 7.29595602e-01 -3.03221699e-02 9.00217772e-01
-3.67114723e-01 1.51186407e-01 -1.26428187e-01 8.96177143e-02
3.93648714e-01 9.99863327e-01 -2.52408177e-01 -8.19724798e-01
-4.42289203e-01 1.00956881e+00 7.67260313e-01 6.91892087e-01
-8.40008378e-01 -6.98099732e-01 4.37429726e-01 -3.37670028e-01
4.09105092e-01 -4.72556539e-02 -5.01512825e-01 -1.09656191e+00
1.78957373e-01 -6.13546431e-01 8.34623396e-01 -1.29579186e+00
-1.27006316e+00 7.94655621e-01 8.96043926e-02 -5.61944366e-01
5.85317947e-02 -4.41730879e-02 -1.26406968e+00 1.31552017e+00
-1.68556368e+00 -9.16397572e-01 -6.45817757e-01 4.56192791e-01
1.24678469e+00 3.60165060e-01 5.80985248e-01 9.29581001e-02
-7.35270500e-01 3.98538917e-01 -3.80655885e-01 2.17242949e-02
5.27087212e-01 -1.57343805e+00 5.47663748e-01 8.07896495e-01
-1.01758912e-01 9.63002563e-01 6.63167715e-01 -8.12963605e-01
-1.15077019e+00 -9.62427199e-01 1.44703460e+00 -6.69470608e-01
4.85531002e-01 -2.51550764e-01 -1.47431183e+00 8.54189754e-01
1.03885257e+00 -3.47316116e-01 5.78424573e-01 2.12902829e-01
-2.15986714e-01 -1.00796431e-01 -7.42706776e-01 6.65197790e-01
9.22717810e-01 -6.45813048e-01 -1.51962221e+00 2.47118950e-01
1.53096068e+00 -4.21217948e-01 -5.08979917e-01 3.29203725e-01
7.69694373e-02 -8.64331901e-01 7.68798470e-01 -1.14172745e+00
7.17855573e-01 -1.16072297e-01 -7.37745464e-02 -1.10017526e+00
-6.66237652e-01 -6.48570418e-01 -4.36184496e-01 1.25232494e+00
3.83097231e-01 -1.89898059e-01 5.02397478e-01 5.61543405e-01
-4.13702458e-01 -1.04868126e+00 -8.43432784e-01 -4.90843840e-02
1.26799598e-01 -3.64536822e-01 6.45987868e-01 8.63346219e-01
1.59729868e-01 1.10921919e+00 -9.22446698e-02 2.64115661e-01
1.26725182e-01 3.57609540e-01 3.76098782e-01 -7.31997073e-01
-1.70671210e-01 -6.73623025e-01 2.27292776e-01 -1.72008908e+00
1.20303676e-01 -6.40369892e-01 6.57706782e-02 -1.99828529e+00
2.43823975e-01 2.20533639e-01 -3.17297876e-01 1.43580392e-01
-1.04154122e+00 -5.16536236e-01 -1.24749411e-02 2.43470091e-02
-1.01276600e+00 5.67765176e-01 1.30952942e+00 -2.00332865e-01
-2.00894043e-01 -2.65782148e-01 -7.85231173e-01 2.61276573e-01
6.10619068e-01 -1.80763751e-01 -5.54858088e-01 -1.02997553e+00
2.12049797e-01 4.38145250e-01 1.99774280e-01 -5.68014979e-01
5.93086481e-01 -3.63399722e-02 3.40693295e-01 -1.12241793e+00
1.60344347e-01 -5.86424768e-01 -3.74447793e-01 2.04532593e-01
-1.03658712e+00 4.27943438e-01 2.60649234e-01 9.82234001e-01
-5.26448429e-01 -2.05905527e-01 2.82932222e-01 -1.89888686e-01
-6.78086698e-01 3.00399542e-01 -2.81712383e-01 4.28699374e-01
6.86803222e-01 3.39916170e-01 -6.19756997e-01 -5.08549035e-01
-9.69829917e-01 9.99379992e-01 -2.59309679e-01 7.14697957e-01
5.14454305e-01 -1.26674652e+00 -7.92816520e-01 -2.12985188e-01
1.06736407e-01 4.43036646e-01 6.81212306e-01 3.66853207e-01
-5.40809572e-01 6.22789681e-01 2.89104521e-01 -6.52478218e-01
-1.14259791e+00 9.74889159e-01 3.06567252e-01 -3.95548642e-01
-4.95869696e-01 1.21661317e+00 4.63843346e-01 -3.84499803e-02
4.48733628e-01 -8.59188080e-01 -7.26121843e-01 1.69498041e-01
9.63699222e-01 1.69676319e-01 -1.30999371e-01 -2.58101732e-01
-2.26746336e-01 5.87424159e-01 -2.79822737e-01 -3.62877883e-02
1.00027287e+00 -3.46181363e-01 -3.03352684e-01 5.71137547e-01
1.26705766e+00 -2.62949884e-01 -1.10948277e+00 -3.11154872e-01
8.40147138e-02 -2.25868493e-01 -6.96666762e-02 -1.07067358e+00
-6.46869242e-01 1.06623971e+00 -5.71988970e-02 6.15098715e-01
1.21184540e+00 2.81263441e-01 9.41563606e-01 4.28898156e-01
-3.02062422e-01 -8.01083326e-01 3.47888738e-01 7.04267740e-01
1.53144646e+00 -1.11602533e+00 -3.14727336e-01 -2.93757349e-01
-6.62719309e-01 9.86410916e-01 1.08604610e+00 -2.88998574e-01
5.68298519e-01 -3.05768400e-01 -2.28818431e-02 -3.21352690e-01
-1.38299119e+00 -1.29287422e-01 2.83808053e-01 1.99813381e-01
5.88336885e-01 -2.06014752e-01 -3.58804494e-01 1.03619015e+00
1.82182137e-02 -1.04307905e-01 3.59581530e-01 9.05505896e-01
-7.76353478e-01 -6.22833252e-01 -2.10745320e-01 2.84253776e-01
-5.59372067e-01 -3.10228646e-01 -6.12828791e-01 5.55794656e-01
-1.92127183e-01 1.19119406e+00 1.42758653e-01 -9.66742709e-02
7.32765615e-01 5.22939563e-01 3.98657173e-02 -7.78869450e-01
-1.15296173e+00 2.18728371e-02 1.19233854e-01 -6.22810483e-01
-1.80413961e-01 -4.55338717e-01 -1.16030157e+00 8.69649649e-02
-9.02410895e-02 3.10390234e-01 2.20923036e-01 8.46545458e-01
7.09348142e-01 8.96359563e-01 3.37179512e-01 -3.93308431e-01
-3.79781336e-01 -1.42654300e+00 2.46028602e-01 5.17870486e-01
7.46663690e-01 -8.43724608e-02 -1.97617471e-01 -3.65316197e-02] | [11.239821434020996, 8.176207542419434] |
15e9d9c2-1b3f-4f22-8662-6091be2310a1 | unsupervised-extractive-summarization-with | 2211.04698 | null | https://arxiv.org/abs/2211.04698v1 | https://arxiv.org/pdf/2211.04698v1.pdf | Unsupervised Extractive Summarization with Heterogeneous Graph Embeddings for Chinese Document | In the scenario of unsupervised extractive summarization, learning high-quality sentence representations is essential to select salient sentences from the input document. Previous studies focus more on employing statistical approaches or pre-trained language models (PLMs) to extract sentence embeddings, while ignoring the rich information inherent in the heterogeneous types of interaction between words and sentences. In this paper, we are the first to propose an unsupervised extractive summarizaiton method with heterogeneous graph embeddings (HGEs) for Chinese document. A heterogeneous text graph is constructed to capture different granularities of interactions by incorporating graph structural information. Moreover, our proposed graph is general and flexible where additional nodes such as keywords can be easily integrated. Experimental results demonstrate that our method consistently outperforms the strong baseline in three summarization datasets. | ['Di Yin', 'Siyu An', 'Ye Liu', 'Chen Lin'] | 2022-11-09 | null | null | null | null | ['sentence-embeddings', 'sentence-embeddings', 'unsupervised-extractive-summarization', 'extractive-summarization'] | ['methodology', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 3.95982057e-01 5.60813725e-01 -4.11016196e-01 -1.58868313e-01
-7.21257448e-01 -5.38333058e-01 8.62482369e-01 8.06648791e-01
-4.10695195e-01 5.91953516e-01 1.24755394e+00 -1.14101633e-01
-7.32165650e-02 -7.56630301e-01 -3.96417677e-01 -4.21832174e-01
-1.28505602e-01 1.85730204e-01 3.85702327e-02 -4.45898205e-01
6.16993666e-01 2.06159860e-01 -1.02778399e+00 1.17530480e-01
1.39341652e+00 2.43802741e-01 3.27174127e-01 9.35817182e-01
-5.64711511e-01 6.85251772e-01 -8.78689110e-01 -4.12822932e-01
-1.97349772e-01 -6.17563009e-01 -7.32035756e-01 2.89734393e-01
5.36372066e-01 -1.65279463e-01 -5.97141802e-01 1.00049198e+00
6.90382421e-01 3.23023111e-01 9.43741739e-01 -6.34846389e-01
-9.94679034e-01 1.23550940e+00 -6.29671931e-01 3.41233522e-01
4.73917753e-01 -1.75223708e-01 1.64990556e+00 -9.12006915e-01
7.52579927e-01 1.23778689e+00 3.02148402e-01 4.08356160e-01
-1.10005057e+00 -5.16049229e-02 5.28592229e-01 1.39492899e-01
-1.02601159e+00 -4.47648168e-01 1.27210438e+00 -1.00610308e-01
1.25853252e+00 4.86240327e-01 5.46272993e-01 1.14059865e+00
3.86577368e-01 1.12394357e+00 3.01742703e-01 -5.68418860e-01
1.77870601e-01 -1.02140270e-01 6.65790975e-01 6.29552364e-01
7.33818293e-01 -7.91010380e-01 -5.88682532e-01 -2.03173354e-01
3.29135299e-01 1.06757529e-01 -4.10179973e-01 -2.45673120e-01
-1.14598584e+00 1.02139807e+00 3.01244438e-01 5.29734731e-01
-6.39620006e-01 1.36925774e-02 7.72683918e-01 1.58787161e-01
7.81311750e-01 8.68091583e-01 -3.64560820e-02 4.63141426e-02
-1.11669362e+00 9.63120684e-02 9.35231090e-01 1.19159806e+00
6.37954473e-01 2.85250992e-01 -6.89123333e-01 7.96902895e-01
2.77140029e-02 1.27811655e-01 6.48892105e-01 -4.91780639e-01
9.57992077e-01 9.02820587e-01 -2.55091131e-01 -1.15887499e+00
-2.60871381e-01 -4.59551007e-01 -1.05422950e+00 -7.27975488e-01
-4.48928624e-01 -3.19851249e-01 -7.07270145e-01 1.13198543e+00
-1.39949381e-01 -9.04254615e-02 3.16768914e-01 4.67042059e-01
1.37463295e+00 8.29189479e-01 -1.35469332e-01 -3.41092080e-01
1.15061116e+00 -1.27078688e+00 -1.00079799e+00 -3.19906652e-01
7.18336165e-01 -5.17192602e-01 1.00288010e+00 -9.27129835e-02
-1.15888023e+00 -2.99801886e-01 -1.23591554e+00 -3.41483176e-01
-3.00057679e-01 6.90912381e-02 5.52338243e-01 2.43043512e-01
-1.15210652e+00 5.01940608e-01 -6.49780273e-01 -6.06601715e-01
3.40628743e-01 2.21153751e-01 -2.55722195e-01 -7.39309341e-02
-1.17420089e+00 7.35963285e-01 7.64054298e-01 -8.24299753e-02
-3.24549288e-01 -4.64602411e-01 -1.28851330e+00 4.85706061e-01
6.33530796e-01 -8.63356829e-01 8.52478743e-01 -7.89499938e-01
-1.44942188e+00 2.85610199e-01 -2.95303792e-01 -5.38748443e-01
-5.24503971e-03 -3.51029426e-01 -2.07704291e-01 6.23204589e-01
2.94028595e-02 2.98831969e-01 6.38695776e-01 -1.35978270e+00
-1.30205512e-01 -2.06292018e-01 1.80169478e-01 5.29119968e-01
-8.81207049e-01 6.84331805e-02 -4.54888076e-01 -9.92895722e-01
-2.01002210e-01 -5.21295965e-01 -4.97809440e-01 -8.03620815e-01
-8.92726302e-01 -5.50930619e-01 8.22370291e-01 -8.25234115e-01
1.80368495e+00 -1.89835405e+00 6.06126666e-01 -1.92610189e-01
5.21456659e-01 4.03613776e-01 -4.62324709e-01 1.30564618e+00
3.33018243e-01 4.50640678e-01 -3.65253866e-01 -5.30146599e-01
1.83610886e-01 1.63918406e-01 -2.91628361e-01 9.29757059e-02
4.28948224e-01 1.34935617e+00 -1.18484068e+00 -8.70902181e-01
-6.41684905e-02 3.42450798e-01 -5.54030001e-01 1.90994665e-01
-1.60678968e-01 1.09510086e-02 -7.83321917e-01 3.61027002e-01
4.62465405e-01 -1.64850876e-01 4.05215532e-01 -1.25189364e-01
1.92095935e-02 4.98646796e-01 -7.61508107e-01 1.67639709e+00
-4.11208957e-01 5.95764160e-01 -1.45486683e-01 -1.17702365e+00
8.35563421e-01 2.17482835e-01 3.20560426e-01 -1.54191807e-01
9.82408002e-02 -3.32996100e-01 -1.01544894e-01 -5.30256033e-01
1.25990033e+00 2.39765480e-01 -4.66421664e-01 6.16095662e-01
3.50589067e-01 -3.55466932e-01 5.26942551e-01 1.00039637e+00
1.24203575e+00 -3.57179105e-01 7.29920328e-01 -3.36225450e-01
3.47848892e-01 -1.67004585e-01 3.66612107e-01 8.62516761e-01
7.44952559e-02 7.29549587e-01 7.45585620e-01 2.05655947e-01
-8.18904340e-01 -1.00176275e+00 4.10302281e-01 1.04110169e+00
6.60316497e-02 -1.17029500e+00 -5.96033514e-01 -9.21665132e-01
-1.61801487e-01 1.00659823e+00 -4.67966050e-01 -2.72278249e-01
-8.42530668e-01 -4.57344443e-01 3.05091590e-01 5.42502701e-01
2.34508201e-01 -1.09241605e+00 -2.17839107e-01 3.01487386e-01
-1.88059106e-01 -1.09948385e+00 -9.71175432e-01 -8.92544445e-03
-9.26498294e-01 -7.56236434e-01 -5.72633147e-01 -9.50964153e-01
8.15153122e-01 5.34427106e-01 1.04941273e+00 2.78555062e-02
-6.31645843e-02 6.48214638e-01 -8.25275242e-01 -3.99169683e-01
-3.19233596e-01 5.27941525e-01 -1.59561634e-01 -7.03006461e-02
1.89246446e-01 -5.57208836e-01 -5.06353438e-01 -7.02816188e-01
-1.16024506e+00 5.66969030e-02 8.09616089e-01 8.91669095e-01
1.87740847e-01 -2.63208579e-02 8.76163304e-01 -1.18666232e+00
1.36955690e+00 -4.82460469e-01 2.62907445e-01 4.93977219e-01
-4.34614897e-01 2.70057887e-01 8.14321816e-01 -3.83194178e-01
-1.16119814e+00 -5.31770825e-01 5.13173640e-02 1.38833061e-01
7.46519938e-02 1.03850853e+00 -1.60550892e-01 4.63567436e-01
4.22612250e-01 4.73883837e-01 -2.97602773e-01 -2.65744984e-01
7.34077811e-01 7.64518678e-01 1.92476213e-01 -3.32444996e-01
7.18921840e-01 1.62683085e-01 -2.07488567e-01 -1.50355065e+00
-7.33979702e-01 -6.88392699e-01 -8.45788777e-01 6.69723959e-04
7.84164250e-01 -7.46316016e-01 1.33232519e-01 -1.38123542e-01
-1.32883060e+00 4.15585525e-02 -4.02288795e-01 4.27549899e-01
-1.78562224e-01 1.05088246e+00 -6.84856415e-01 -6.33908927e-01
-8.31948459e-01 -6.82834268e-01 1.22109342e+00 2.41280615e-01
-4.12302881e-01 -1.32749677e+00 2.71724984e-02 1.59983501e-01
3.77015114e-01 3.19931477e-01 9.83965874e-01 -1.03315878e+00
-4.08966899e-01 -4.77151483e-01 -1.17758080e-01 2.66938537e-01
5.31269312e-01 1.16695218e-01 -4.98912990e-01 -3.10466707e-01
-2.31482476e-01 -1.74401641e-01 1.35696983e+00 4.44946051e-01
8.61298800e-01 -8.10282946e-01 -3.42142403e-01 9.36479568e-02
1.16791081e+00 -3.61546278e-01 4.54573512e-01 7.13103786e-02
1.09104586e+00 6.41413033e-01 2.32035831e-01 5.27042687e-01
6.14286244e-01 1.87426075e-01 7.24490285e-02 -1.61047827e-03
-1.78531677e-01 -4.63059068e-01 6.10745728e-01 1.66546154e+00
-5.07244281e-03 -7.24514008e-01 -5.25458634e-01 6.66123807e-01
-1.95278323e+00 -1.14092112e+00 -9.95801389e-02 1.66334760e+00
8.85024726e-01 1.07771695e-01 -1.07054628e-01 -8.94013792e-02
7.65192330e-01 8.62099946e-01 -2.24553853e-01 -7.57190883e-01
-2.20383719e-01 2.19537407e-01 3.48756731e-01 5.34404755e-01
-9.30409670e-01 1.14884806e+00 5.81600142e+00 7.34192193e-01
-6.69170260e-01 -2.02833563e-01 4.08994630e-02 -1.05947532e-01
-8.51042509e-01 2.85641458e-02 -6.64714277e-01 1.09717190e-01
8.66959691e-01 -6.88138306e-01 -8.39986652e-02 4.67130214e-01
3.78652424e-01 9.10292193e-02 -8.29145849e-01 5.12912512e-01
6.49296045e-01 -1.45618570e+00 6.27929509e-01 -9.80281308e-02
8.29305172e-01 -1.76358104e-01 -3.08379352e-01 5.13893008e-01
4.41577494e-01 -7.07016170e-01 1.80922151e-01 4.62706536e-01
3.27899277e-01 -7.72259593e-01 7.11556375e-01 3.70832831e-01
-1.15460503e+00 1.33535624e-01 -4.63537395e-01 -1.24762140e-04
2.88035721e-01 5.53714752e-01 -8.78066719e-01 1.17651641e+00
1.07951209e-01 1.11081886e+00 -8.52855682e-01 7.15205550e-01
-5.35196006e-01 8.20635915e-01 2.56022185e-01 -5.87800980e-01
4.08082932e-01 -3.27521503e-01 9.03694391e-01 1.71396935e+00
1.53518945e-01 2.63451301e-02 2.87719816e-01 4.10653502e-01
-4.29267645e-01 5.69598675e-01 -9.78347361e-01 -5.95638454e-01
2.95323431e-01 1.40333009e+00 -7.36555278e-01 -5.07925034e-01
-4.57337111e-01 1.14591849e+00 5.57411432e-01 5.28441310e-01
-3.26375604e-01 -7.83683121e-01 1.75376266e-01 -1.81501076e-01
5.31642556e-01 -6.45226598e-01 -1.53126642e-01 -1.58061838e+00
5.65097742e-02 -6.09763086e-01 4.08143133e-01 -3.36241156e-01
-1.32915866e+00 5.97114623e-01 9.32672173e-02 -9.16475236e-01
-2.25377113e-01 -1.68424416e-02 -1.15484977e+00 4.73124832e-01
-1.47948384e+00 -1.27921784e+00 2.48465702e-01 1.24581501e-01
1.08158839e+00 -4.77442816e-02 6.56360090e-01 -2.92344272e-01
-7.38359034e-01 5.53302169e-01 1.61615029e-01 1.85062155e-01
5.24887681e-01 -1.47205687e+00 4.26961601e-01 1.12925482e+00
3.68056357e-01 9.64722693e-01 8.32879424e-01 -7.66501665e-01
-1.65876508e+00 -1.26020610e+00 1.33232462e+00 -2.59944528e-01
7.93720782e-01 -5.74409664e-01 -9.62758720e-01 6.73979163e-01
1.08426821e+00 -5.82874835e-01 1.04695630e+00 2.67852962e-01
-1.59339607e-01 2.58307159e-01 -6.54083788e-01 1.05540454e+00
9.78754759e-01 -4.15044785e-01 -1.22309399e+00 3.13958496e-01
1.21686840e+00 5.06819896e-02 -8.15965116e-01 3.16475838e-01
4.27406356e-02 -4.52026963e-01 7.25689113e-01 -7.35450089e-01
6.69783413e-01 7.86084458e-02 -7.64006656e-03 -1.67962408e+00
-3.17917764e-01 -6.75097466e-01 -5.32581329e-01 1.63231897e+00
4.19278502e-01 -5.22230864e-01 4.89705622e-01 2.26392880e-01
-5.48480630e-01 -6.73631132e-01 -4.80582893e-01 -5.94899893e-01
2.58674286e-02 1.09558083e-01 3.57752413e-01 9.14246380e-01
5.48551023e-01 8.88537884e-01 -2.90409714e-01 9.41969678e-02
4.47167337e-01 3.16733092e-01 7.87017763e-01 -1.15103316e+00
-1.81487665e-01 -7.24115491e-01 -2.83502400e-01 -1.09696043e+00
6.62376285e-01 -1.19561374e+00 -2.48370972e-02 -2.40226960e+00
4.95407164e-01 4.93208855e-01 -2.96652108e-01 6.02371581e-02
-6.22458816e-01 -4.37194139e-01 1.42084718e-01 -8.16993341e-02
-9.81269479e-01 1.18390858e+00 1.30452013e+00 -5.00557005e-01
-3.99124414e-01 -2.23549187e-01 -1.27720475e+00 4.98804688e-01
8.49824548e-01 -3.33690822e-01 -6.12310290e-01 -4.40407395e-01
-3.67727019e-02 -1.15528367e-02 -1.78605303e-01 -4.68716443e-01
4.28210437e-01 -1.70764461e-01 -1.40006840e-01 -6.69496298e-01
-1.59108564e-02 -2.29065061e-01 -6.99911058e-01 1.45327255e-01
-6.71925902e-01 8.54989663e-02 -1.88459735e-03 8.51392269e-01
-4.10212696e-01 -4.16300565e-01 1.36329800e-01 -1.07416809e-01
-4.63620991e-01 2.22373426e-01 -4.40350831e-01 2.95737326e-01
6.66513503e-01 -5.58317788e-02 -3.13330919e-01 -6.11238241e-01
-3.53674173e-01 4.06922519e-01 3.42493325e-01 5.81434309e-01
9.66007352e-01 -1.15866399e+00 -9.25570965e-01 -2.43067250e-01
2.26679057e-01 4.85484190e-02 1.77737087e-01 5.78171194e-01
-3.02147418e-01 4.14569587e-01 2.08998546e-01 -8.29922110e-02
-1.23887026e+00 5.51343143e-01 -4.28189427e-01 -5.21105468e-01
-9.78985488e-01 5.17539561e-01 2.42094379e-02 -2.40584001e-01
5.28375804e-02 -4.02473539e-01 -6.25744343e-01 4.36330080e-01
4.42244887e-01 3.19289476e-01 -1.72706068e-01 -6.37963295e-01
-2.72480994e-02 2.71073222e-01 -4.27573711e-01 1.18794352e-01
1.66239941e+00 -2.61553377e-01 -3.97319108e-01 4.56832826e-01
1.19490099e+00 3.92208636e-01 -6.74836516e-01 -3.09361279e-01
4.58360344e-01 -6.14390709e-02 3.91504355e-02 -9.17635113e-02
-6.92988753e-01 7.41993487e-01 -5.08190513e-01 5.21774709e-01
9.30202842e-01 1.83507472e-01 9.19539154e-01 6.47603869e-01
-8.58236253e-02 -1.12770844e+00 3.49390924e-01 5.38269520e-01
1.16555130e+00 -9.39262331e-01 3.65402311e-01 -3.30464482e-01
-9.63129640e-01 1.21271813e+00 4.87927556e-01 -4.00918722e-01
3.59622777e-01 -4.58345041e-02 -3.10699850e-01 -3.65624398e-01
-8.08022559e-01 -3.66232634e-01 5.40175736e-01 4.78805333e-01
5.83979130e-01 -9.84560512e-03 -8.50711286e-01 6.92742169e-01
-9.51699391e-02 -7.10395873e-01 1.06674707e+00 1.11319542e+00
-6.42479539e-01 -1.18229663e+00 1.53108209e-01 8.12217295e-01
-4.45036620e-01 -3.90360415e-01 -9.20103610e-01 6.33657992e-01
-6.65351987e-01 1.17330813e+00 -8.82473588e-02 -1.59632698e-01
3.50329787e-01 3.87546048e-02 3.51700425e-01 -1.24938238e+00
-5.60177147e-01 1.16054095e-01 3.44681859e-01 -3.30799036e-02
-4.35156554e-01 -6.49386227e-01 -1.25476992e+00 9.47251469e-02
-4.17721301e-01 4.21920180e-01 3.13251019e-01 7.77896821e-01
6.58209026e-01 8.43462467e-01 6.66970313e-01 -1.03491080e+00
-4.66346502e-01 -1.27604830e+00 -6.77348554e-01 3.03705126e-01
3.95288140e-01 -1.90124512e-01 -4.46161389e-01 -1.49210185e-01] | [12.5450439453125, 9.582670211791992] |
ad48b55c-0749-4816-b972-90ae784ec41b | neural-extractive-summarization-with | null | null | https://aclanthology.org/2020.emnlp-main.295 | https://aclanthology.org/2020.emnlp-main.295.pdf | Neural Extractive Summarization with Hierarchical Attentive Heterogeneous Graph Network | Sentence-level extractive text summarization is substantially a node classification task of network mining, adhering to the informative components and concise representations. There are lots of redundant phrases between extracted sentences, but it is difficult to model them exactly by the general supervised methods. Previous sentence encoders, especially BERT, specialize in modeling the relationship between source sentences. While, they have no ability to consider the overlaps of the target selected summary, and there are inherent dependencies among target labels of sentences. In this paper, we propose HAHSum (as shorthand for Hierarchical Attentive Heterogeneous Graph for Text Summarization), which well models different levels of information, including words and sentences, and spotlights redundancy dependencies between sentences. Our approach iteratively refines the sentence representations with redundancy-aware graph and delivers the label dependencies by message passing. Experiments on large scale benchmark corpus (CNN/DM, NYT, and NEWSROOM) demonstrate that HAHSum yields ground-breaking performance and outperforms previous extractive summarizers. | ['Shi Wang', 'Cong Cao', 'Fang Fang', 'Hengzhu Tang', 'Yanan Cao', 'Ruipeng Jia'] | null | null | null | null | emnlp-2020-11 | ['extractive-document-summarization'] | ['natural-language-processing'] | [ 4.42260146e-01 5.56871533e-01 -5.19726276e-01 -4.59921151e-01
-5.17126799e-01 -1.81001291e-01 4.18761522e-01 7.66519308e-01
-7.33961677e-03 9.25564587e-01 1.15325177e+00 1.05324797e-02
-1.56299517e-01 -8.52504194e-01 -4.59440976e-01 -3.48436594e-01
-2.04270348e-01 2.17947647e-01 2.95589268e-01 -5.17588735e-01
4.14913148e-01 5.82242198e-02 -9.95017409e-01 5.51048219e-01
9.90610659e-01 5.25243282e-01 3.18653196e-01 8.59847546e-01
-7.51156688e-01 1.42242837e+00 -9.82818305e-01 -5.75298250e-01
-3.16061348e-01 -6.97443664e-01 -1.16516948e+00 2.94350207e-01
6.06790185e-01 -1.67697087e-01 -8.19901884e-01 1.18878484e+00
4.70416129e-01 9.18138698e-02 7.38982797e-01 -1.02837706e+00
-6.82774425e-01 1.64435542e+00 -7.66521394e-01 5.61838210e-01
3.88477445e-01 -4.36379939e-01 1.64620388e+00 -4.36133862e-01
6.98837221e-01 1.33852017e+00 4.85482991e-01 4.90982592e-01
-9.07595456e-01 -3.28121424e-01 5.63724875e-01 -8.17930419e-03
-9.64449883e-01 -5.15316784e-01 9.58075881e-01 -2.36387830e-02
1.27638674e+00 6.37319744e-01 5.13385177e-01 1.16557598e+00
5.29692054e-01 1.11635494e+00 3.37758482e-01 -2.35795557e-01
-3.38110398e-03 -1.47469878e-01 9.42985773e-01 7.65911400e-01
6.16276443e-01 -8.04130793e-01 -9.45110500e-01 -4.24907319e-02
2.13342905e-01 4.35602851e-02 -4.43539679e-01 1.99670985e-01
-1.03574944e+00 7.60639489e-01 5.15015423e-01 4.20989841e-01
-4.66589302e-01 2.03081727e-01 8.33916366e-01 3.71141255e-01
8.84429932e-01 5.02562642e-01 -3.10066313e-01 2.63587326e-01
-1.10792828e+00 -1.23188190e-01 9.80211318e-01 1.53291988e+00
7.39520967e-01 3.16390097e-01 -8.13790500e-01 7.52808511e-01
4.29056473e-02 8.53635073e-02 6.63037479e-01 -4.99425113e-01
9.41296458e-01 9.17612731e-01 -6.03978574e-01 -1.38582885e+00
-6.26740932e-01 -9.97404039e-01 -1.52511632e+00 -6.92172945e-01
-6.54879570e-01 -3.22842956e-01 -7.22492278e-01 1.43725526e+00
-1.07656620e-01 3.81786488e-02 3.47459972e-01 4.26839411e-01
1.67887342e+00 8.93072903e-01 -2.40007788e-02 -7.70455897e-01
1.19728327e+00 -1.42404938e+00 -1.05379820e+00 -4.99588311e-01
6.56838238e-01 -4.10957605e-01 5.64323366e-01 7.59084299e-02
-1.22341442e+00 -3.21688265e-01 -1.20025349e+00 -3.23970765e-01
-7.53477728e-03 8.81484002e-02 8.26926708e-01 1.43078730e-01
-1.31203580e+00 8.47694337e-01 -5.59230268e-01 -4.44522023e-01
5.55685282e-01 2.81861871e-01 -2.82689720e-01 2.08294570e-01
-1.16226506e+00 6.56409919e-01 7.74046302e-01 2.52138823e-02
-5.19320965e-01 -3.10249150e-01 -1.15445054e+00 5.61861873e-01
4.59841698e-01 -1.06023622e+00 1.11713266e+00 -8.23385298e-01
-1.24829960e+00 4.49911565e-01 -3.48973453e-01 -7.69650578e-01
-5.88625371e-02 1.47802085e-02 -2.91568696e-01 3.71769100e-01
2.60075033e-01 5.43603837e-01 7.47202933e-01 -1.00317657e+00
-6.59187376e-01 -2.98453093e-01 7.86612108e-02 5.47394753e-01
-7.06805825e-01 -8.06491897e-02 -2.99327821e-01 -6.50024056e-01
1.75782666e-01 -2.45136991e-01 -4.52677846e-01 -8.77376020e-01
-1.29152632e+00 -5.82342207e-01 4.92672712e-01 -7.37865388e-01
1.68446934e+00 -1.84406412e+00 4.39137816e-01 -1.66532561e-01
7.67243028e-01 -3.70922200e-02 -3.02588582e-01 9.09841955e-01
1.46626309e-01 3.77801627e-01 -3.71432513e-01 -6.01598084e-01
-7.16880187e-02 6.77243471e-02 -4.39197391e-01 -1.34942576e-03
2.23391369e-01 9.94167030e-01 -1.16519833e+00 -1.03321671e+00
-3.05855721e-01 -6.99588209e-02 -2.09207654e-01 3.75055112e-02
-2.20601588e-01 -1.32419318e-01 -7.51816392e-01 3.87181967e-01
3.86810601e-01 -3.46643627e-01 2.45545924e-01 -3.03695142e-01
1.67870462e-01 7.36903906e-01 -6.95607126e-01 1.86182094e+00
-1.88158929e-01 7.15217292e-01 -8.58802162e-03 -1.14635885e+00
9.61478949e-01 1.51534095e-01 3.20969611e-01 -2.80178934e-01
1.38983846e-01 -1.38682783e-01 -1.61600456e-01 -5.61622381e-01
1.02082193e+00 1.94635838e-01 -3.03615540e-01 3.03718656e-01
5.49737215e-01 -2.36949459e-01 7.84783483e-01 1.15058494e+00
1.57560039e+00 -3.31589311e-01 5.90639174e-01 -3.75541568e-01
4.62607056e-01 6.29003942e-02 5.13945401e-01 9.10745561e-01
3.72126788e-01 6.73649728e-01 1.05098903e+00 8.03805143e-02
-6.35247827e-01 -7.02012479e-01 3.55349272e-01 1.00726688e+00
1.80157736e-01 -1.18043768e+00 -5.88351369e-01 -9.44979370e-01
-3.48181486e-01 8.49866807e-01 -4.89021391e-01 -3.94919813e-01
-5.74876606e-01 -7.32837439e-01 3.77157182e-01 4.13018346e-01
4.43810433e-01 -1.15659785e+00 -4.76046540e-02 3.22246253e-01
-2.72501141e-01 -9.65165079e-01 -7.75472760e-01 2.39408478e-01
-1.03632128e+00 -9.55724239e-01 -2.60254323e-01 -1.07922852e+00
6.58925951e-01 5.58932602e-01 1.34610224e+00 3.31205040e-01
-3.23792920e-02 2.15487346e-01 -6.24352217e-01 -3.27978849e-01
-5.58677733e-01 7.28233635e-01 -2.17806622e-01 -1.68282092e-01
-5.52500002e-02 -7.86424220e-01 -3.60653490e-01 -5.48394322e-01
-8.64086509e-01 3.81806105e-01 9.85242665e-01 7.41486669e-01
2.10072055e-01 1.66316420e-01 1.05742538e+00 -1.31846976e+00
1.19787467e+00 -4.99005407e-01 1.81079194e-01 4.45982844e-01
-4.79738653e-01 2.82611161e-01 8.34327340e-01 4.92159761e-02
-9.84986901e-01 -3.07796597e-01 -1.11710705e-01 1.67896658e-01
2.34723777e-01 9.35100496e-01 -2.60550552e-03 4.48299229e-01
5.37772477e-01 5.29224157e-01 -2.84394026e-01 -2.89769322e-01
3.41958553e-01 7.95557976e-01 3.87522429e-01 -1.84515938e-01
5.71601510e-01 1.03493042e-01 7.80448392e-02 -1.19277847e+00
-1.26252294e+00 -5.82443535e-01 -5.01561880e-01 -1.75707787e-02
6.61607206e-01 -7.95489073e-01 -3.81194174e-01 4.71477630e-03
-1.56888771e+00 3.49885762e-01 -4.71303850e-01 1.71612754e-01
-8.37544650e-02 8.36647868e-01 -8.68308067e-01 -5.66012800e-01
-1.14076936e+00 -7.24850357e-01 1.07190001e+00 2.81267107e-01
-3.02173376e-01 -9.78034079e-01 -6.42095655e-02 1.09024584e-01
2.83424854e-01 1.43168643e-01 1.09785712e+00 -1.03386080e+00
-4.01401997e-01 -2.24299446e-01 -1.60420507e-01 4.06294286e-01
3.25861156e-01 -1.86427720e-02 -5.52287936e-01 -2.46512800e-01
-7.92245045e-02 -2.64348745e-01 1.61365259e+00 4.52803344e-01
1.02102494e+00 -8.84247303e-01 -3.45386356e-01 9.07313600e-02
1.09155214e+00 -3.51966769e-01 4.33722913e-01 -2.36641571e-01
8.72670889e-01 7.35611796e-01 1.65549055e-01 2.82713860e-01
4.76969093e-01 2.30109971e-02 4.73967761e-01 6.98623061e-02
-2.98446804e-01 -4.75698620e-01 4.37660962e-01 1.80969715e+00
2.09930867e-01 -7.55062163e-01 -2.53773153e-01 3.69518459e-01
-2.03819489e+00 -1.04904985e+00 -5.40213168e-01 1.59646332e+00
1.03123438e+00 4.41792756e-01 -2.43550241e-01 -3.37171666e-02
8.88573289e-01 8.35271358e-01 -3.36257935e-01 -3.77718747e-01
-3.26415867e-01 1.73707753e-02 3.63319367e-01 5.07661879e-01
-9.68088627e-01 1.01866460e+00 5.83876181e+00 9.22980189e-01
-5.56622446e-01 -1.53811455e-01 5.93327045e-01 -7.28829727e-02
-6.05805695e-01 1.11741498e-01 -8.21281731e-01 2.55700886e-01
7.83909798e-01 -6.45581126e-01 -1.60434395e-02 6.13538504e-01
6.78885132e-02 1.09720416e-01 -1.05680859e+00 5.54799795e-01
6.45659745e-01 -1.69637263e+00 6.03861988e-01 -5.14419317e-01
7.95695305e-01 -1.00783654e-01 -5.21138966e-01 6.24821782e-01
3.96352649e-01 -8.19326639e-01 3.80950630e-01 4.91246998e-01
2.04012111e-01 -7.27553427e-01 9.82970536e-01 6.55213296e-01
-1.15743721e+00 1.69661731e-01 -8.31806421e-01 -1.19592644e-01
6.46449775e-02 6.94249034e-01 -8.32328618e-01 1.15567577e+00
2.85588145e-01 1.39326668e+00 -8.05161893e-01 7.58927047e-01
-4.64749604e-01 7.11618781e-01 1.78277075e-01 -6.02063775e-01
2.26856098e-01 -1.99506119e-01 8.50400269e-01 1.51644230e+00
1.35293528e-01 4.90071662e-02 3.24143320e-01 4.72886771e-01
-6.70693934e-01 3.86193067e-01 -5.98172009e-01 -1.29676297e-01
4.34497565e-01 1.54280901e+00 -7.20577896e-01 -5.93261957e-01
-2.08862573e-01 9.38810468e-01 5.27809083e-01 3.18240225e-01
-3.35473239e-01 -6.74611032e-01 -4.06864956e-02 -1.70555905e-01
1.21771684e-02 1.58176757e-02 -1.40605792e-01 -1.34472561e+00
1.14907093e-01 -6.40933931e-01 4.66083348e-01 -6.45790577e-01
-1.23431206e+00 8.28302443e-01 -4.69329581e-02 -9.76312578e-01
-9.10433531e-02 -2.00272375e-03 -1.04480648e+00 3.75719100e-01
-1.44256532e+00 -1.13365376e+00 -1.90373659e-01 1.48944020e-01
1.24445641e+00 -2.42110685e-01 6.95673764e-01 -1.61746681e-01
-9.52553988e-01 2.84666896e-01 -1.18083656e-01 1.71021521e-01
4.22107816e-01 -1.40679967e+00 4.04825777e-01 1.00963306e+00
2.28638574e-01 5.62758684e-01 9.83288407e-01 -8.81799281e-01
-1.17410862e+00 -1.34412563e+00 1.18062997e+00 5.27363494e-02
7.26490974e-01 -2.93220282e-01 -7.76440859e-01 9.07927394e-01
1.05518126e+00 -5.71354687e-01 5.17926931e-01 1.99884906e-01
-4.47387397e-02 -1.75664455e-01 -5.87010682e-01 6.35425270e-01
1.35721362e+00 -1.05098277e-01 -9.13025141e-01 7.75707901e-01
1.45388305e+00 -1.71107754e-01 -4.97178257e-01 3.26765537e-01
-2.42091417e-01 -7.92861462e-01 5.87288678e-01 -9.87782001e-01
9.27981079e-01 1.13829508e-01 2.00821340e-01 -1.62437129e+00
-5.31497478e-01 -7.31051207e-01 -4.10015672e-01 1.70047092e+00
5.61149240e-01 -2.75640011e-01 7.54716516e-01 -1.26967177e-01
-8.44377756e-01 -7.44251013e-01 -5.94867706e-01 -4.33232397e-01
-3.62505078e-01 8.62298012e-02 4.45982754e-01 8.12948644e-01
5.10012209e-01 1.43497956e+00 -4.74210411e-01 1.91733874e-02
6.09375000e-01 2.24316359e-01 4.81495947e-01 -1.33041978e+00
-1.37361228e-01 -6.58984840e-01 -3.73004943e-01 -1.23706663e+00
6.04978502e-01 -1.31866002e+00 2.06560548e-02 -2.44635296e+00
6.41230583e-01 2.96651870e-01 -8.38756263e-02 3.40244144e-01
-3.89880896e-01 -4.07391936e-01 -1.80683956e-01 1.45327643e-01
-1.29316199e+00 9.32706416e-01 1.27418518e+00 -7.03287184e-01
-1.01394102e-01 1.52684554e-01 -1.15578771e+00 7.07733512e-01
7.88578391e-01 -5.81133068e-01 -6.63154960e-01 -5.13969362e-01
4.46022451e-01 3.05480540e-01 -1.19408354e-01 -7.22751081e-01
8.11716735e-01 -3.53691285e-03 -5.36872959e-03 -1.00741792e+00
7.66196251e-02 -3.50107342e-01 -4.12040532e-01 3.76413673e-01
-9.32018340e-01 5.35636628e-03 -2.64045805e-01 8.71784270e-01
-3.80023420e-01 -6.76045716e-01 2.02128500e-01 -2.77192026e-01
-4.52676296e-01 3.65114629e-01 -3.87414783e-01 3.46151531e-01
5.04495144e-01 2.07627073e-01 -6.82388306e-01 -5.46688557e-01
-4.20886129e-01 7.02639461e-01 -1.85788736e-01 3.10855746e-01
8.65945876e-01 -1.09492624e+00 -1.18087673e+00 -2.34622881e-01
-6.78191483e-02 3.96800727e-01 3.87980521e-01 6.47038341e-01
-3.27750862e-01 3.32951486e-01 1.67670488e-01 -2.27059230e-01
-1.32071590e+00 4.26794171e-01 -2.56174635e-02 -6.13834143e-01
-8.65483403e-01 8.13551664e-01 3.34358066e-02 -7.22824708e-02
2.83991575e-01 -4.63656187e-01 -8.01197827e-01 2.55463481e-01
5.37256360e-01 4.24637288e-01 6.17131256e-02 -3.86357844e-01
-1.54285850e-02 3.68426368e-02 -5.32130361e-01 4.95605618e-01
1.44638813e+00 -2.69501001e-01 -7.93727219e-01 4.57321405e-01
1.13255501e+00 1.15353949e-01 -5.12660742e-01 -4.44242597e-01
3.04137170e-01 2.07798272e-01 4.86252457e-03 -1.07111275e-01
-9.95193660e-01 6.34420872e-01 -5.92157543e-01 7.46296108e-01
9.77582276e-01 1.67753205e-01 8.80189836e-01 7.40334928e-01
4.21762355e-02 -9.98076439e-01 2.94736862e-01 6.15781426e-01
1.02357745e+00 -9.51192439e-01 4.69841391e-01 -8.06318223e-01
-6.32678151e-01 1.21603847e+00 6.14564538e-01 -2.40234882e-01
2.51355082e-01 5.02625480e-02 -6.35684073e-01 -5.34303725e-01
-1.12061119e+00 -1.34185031e-01 2.60061115e-01 2.48908848e-01
4.63039160e-01 -7.47254938e-02 -6.28491104e-01 9.16662753e-01
-4.01058942e-01 -6.50098443e-01 1.00934792e+00 7.28148818e-01
-9.44854140e-01 -7.01882780e-01 2.36884385e-01 1.08921504e+00
-4.90078598e-01 -4.55411792e-01 -7.87157714e-01 5.65763593e-01
-4.86969441e-01 1.22205234e+00 -2.40429062e-02 -4.86030430e-01
3.42100829e-01 -1.53397813e-01 9.08423439e-02 -1.23021662e+00
-7.20156789e-01 -6.65903538e-02 5.80202222e-01 -1.72514804e-02
-4.76833463e-01 -4.01971549e-01 -1.35397243e+00 -2.49044716e-01
-4.82535541e-01 5.44445932e-01 3.06393713e-01 9.06379044e-01
4.33442324e-01 1.21705675e+00 7.95106351e-01 -5.85294962e-01
-8.42072427e-01 -1.44924891e+00 -7.58395076e-01 2.60068566e-01
4.00812477e-01 9.85640883e-02 -4.82540578e-01 -3.21509928e-01] | [12.563172340393066, 9.534866333007812] |
fd7ae812-47a5-46f8-9248-ce03bb5ba80f | a-hybrid-persian-sentiment-analysis-framework | 1909.13568 | null | https://arxiv.org/abs/1909.13568v1 | https://arxiv.org/pdf/1909.13568v1.pdf | A Hybrid Persian Sentiment Analysis Framework: Integrating Dependency Grammar Based Rules and Deep Neural Networks | Social media hold valuable, vast and unstructured information on public opinion that can be utilized to improve products and services. The automatic analysis of such data, however, requires a deep understanding of natural language. Current sentiment analysis approaches are mainly based on word co-occurrence frequencies, which are inadequate in most practical cases. In this work, we propose a novel hybrid framework for concept-level sentiment analysis in Persian language, that integrates linguistic rules and deep learning to optimize polarity detection. When a pattern is triggered, the framework allows sentiments to flow from words to concepts based on symbolic dependency relations. When no pattern is triggered, the framework switches to its subsymbolic counterpart and leverages deep neural networks (DNN) to perform the classification. The proposed framework outperforms state-of-the-art approaches (including support vector machine, and logistic regression) and DNN classifiers (long short-term memory, and Convolutional Neural Networks) with a margin of 10-15% and 3-4% respectively, using benchmark Persian product and hotel reviews corpora. | ['Kia Dashtipour', 'Bin Kong', 'Amir Hussain', 'Mandar Gogate', 'Jingpeng Li', 'Fengling Jiang'] | 2019-09-30 | null | null | null | null | ['persian-sentiment-anlysis'] | ['natural-language-processing'] | [ 1.32809639e-01 -9.14536566e-02 -5.60286343e-01 -5.83750367e-01
-7.26653710e-02 -8.11138511e-01 5.71515739e-01 8.27656746e-01
-2.99102992e-01 7.46236861e-01 1.41875207e-01 -6.50021255e-01
5.57040647e-02 -1.10896814e+00 -3.14823776e-01 -5.90750873e-01
-4.50914577e-02 2.33173579e-01 -1.69311315e-01 -8.41427863e-01
6.58606172e-01 2.61559188e-01 -1.34376013e+00 6.30310893e-01
8.10164809e-01 1.48245788e+00 -2.67744124e-01 4.44874644e-01
-7.52898932e-01 1.30508351e+00 -3.57575476e-01 -7.46495366e-01
-4.22537811e-02 2.95700487e-02 -7.25240886e-01 -4.28644657e-01
-1.18359387e-01 1.92422077e-01 3.86514634e-01 1.18132734e+00
1.29866689e-01 -7.87554160e-02 7.23718226e-01 -9.76775646e-01
-8.76434624e-01 7.56338894e-01 -5.23544967e-01 2.07421213e-01
3.40967149e-01 -4.05557007e-01 1.55363953e+00 -7.66039550e-01
4.06900793e-01 1.09447289e+00 6.76965952e-01 9.38139856e-03
-7.08524585e-01 -7.40049422e-01 5.95475554e-01 1.57984570e-01
-8.64945173e-01 -7.38654956e-02 9.52811599e-01 -4.87529993e-01
1.36695838e+00 -2.36187920e-01 6.71252131e-01 8.39336038e-01
5.83034337e-01 5.66696763e-01 1.07653034e+00 -4.91230518e-01
2.51632452e-01 4.55494583e-01 5.00722826e-01 4.47434753e-01
3.19867283e-01 -3.73768270e-01 -6.62378430e-01 -3.85503694e-02
-2.03473959e-03 2.36468781e-02 4.94967252e-01 -1.69093255e-02
-8.08109045e-01 1.19850981e+00 3.96668911e-01 4.00016069e-01
-5.91850698e-01 -3.38758230e-01 7.33559608e-01 3.64918679e-01
6.68144107e-01 7.20654011e-01 -1.04420233e+00 -9.34607349e-03
-6.99547648e-01 1.39540032e-01 1.00153697e+00 5.77661753e-01
6.36179090e-01 -2.23819278e-02 1.55639380e-01 8.49855125e-01
2.47771859e-01 4.54869390e-01 9.52742457e-01 -1.36436433e-01
4.39862460e-01 1.14343143e+00 2.24524885e-02 -1.61159992e+00
-7.52090871e-01 -4.89133149e-01 -7.16064036e-01 -1.14980869e-01
-1.16087804e-02 -3.70930880e-01 -6.94742620e-01 1.29638183e+00
1.96373194e-01 -1.82977274e-01 6.31398082e-01 5.91311157e-01
7.68292129e-01 7.09620297e-01 3.59904878e-02 -1.87568560e-01
1.49425781e+00 -8.67849827e-01 -6.57200098e-01 -4.75594491e-01
9.50094700e-01 -8.53537440e-01 9.11259949e-01 7.23931491e-01
-6.79128945e-01 -4.87080336e-01 -1.32804990e+00 -5.73500767e-02
-1.10743070e+00 3.62664610e-02 8.28732967e-01 4.80469614e-01
-7.49923825e-01 6.30941212e-01 -5.86790144e-01 -2.80195951e-01
3.85631770e-01 7.42688715e-01 -2.83194512e-01 3.67264986e-01
-1.45634782e+00 9.71390009e-01 3.71230006e-01 3.49124074e-01
-1.55732766e-01 -3.10702115e-01 -1.07583499e+00 -2.68386342e-02
2.25920662e-01 -7.56475553e-02 1.26566780e+00 -1.36618328e+00
-1.70057404e+00 6.70262158e-01 -1.83446601e-01 -8.11054289e-01
-1.31329089e-01 -4.44510341e-01 -7.72648096e-01 -9.07407422e-03
1.33070454e-01 4.48084235e-01 6.05382979e-01 -7.26476848e-01
-9.13963437e-01 -3.82522255e-01 3.18606824e-01 1.18127123e-01
-7.69226909e-01 1.56730905e-01 8.54689255e-02 -3.87811512e-01
-5.59989735e-02 -1.05434132e+00 -2.60717452e-01 -7.29356706e-01
-5.59084713e-01 -4.31431741e-01 7.26715863e-01 -3.57549697e-01
1.28343976e+00 -1.87754703e+00 -1.40876874e-01 3.91422212e-01
-6.71820566e-02 3.18940222e-01 3.57576311e-02 2.53014654e-01
-2.57507265e-01 1.64005727e-01 -1.20764576e-01 -1.11729532e-01
-4.68859673e-02 7.86241144e-02 -5.27252793e-01 2.55762845e-01
5.57288885e-01 6.79979801e-01 -7.87855625e-01 -3.03729504e-01
2.66476154e-01 2.59777576e-01 -5.39818048e-01 -1.39699906e-01
-3.44879001e-01 2.74091978e-02 -3.95559698e-01 8.57776046e-01
4.91308540e-01 -3.49591523e-01 5.79573035e-01 -1.91793516e-01
-3.19714874e-01 7.84959316e-01 -8.34606826e-01 1.19161344e+00
-8.88052285e-01 6.02458537e-01 -2.58100480e-01 -1.36414015e+00
1.31148040e+00 9.86093655e-02 3.05849463e-01 -8.74575198e-01
4.46037084e-01 3.33656311e-01 1.45127639e-01 -3.53205919e-01
7.03567863e-01 -2.77865797e-01 -5.36222339e-01 2.15220347e-01
-1.58133428e-03 1.86770782e-01 4.89462942e-01 -8.41802955e-02
6.68041885e-01 -1.32762985e-02 5.97027659e-01 -3.80329788e-01
1.02185559e+00 2.06073552e-01 4.27365839e-01 2.33039528e-01
1.20934010e-01 9.76761356e-02 1.12769556e+00 -7.30144024e-01
-7.30381966e-01 -4.48583096e-01 -8.69102329e-02 1.37001407e+00
3.59493904e-02 -4.52959180e-01 -4.17245924e-01 -8.54658663e-01
-2.24749818e-02 7.89370418e-01 -7.32913196e-01 -2.92474497e-02
-5.08218467e-01 -9.76608157e-01 1.72513276e-01 6.48928583e-01
3.77445400e-01 -1.37482643e+00 -3.92544895e-01 3.30073744e-01
1.24471068e-01 -1.24434412e+00 1.89972863e-01 6.54150307e-01
-7.33409524e-01 -9.42308724e-01 -1.68762982e-01 -9.08376992e-01
5.15218914e-01 -1.55649036e-01 1.12365949e+00 -2.51099616e-01
1.26908824e-01 -3.21249247e-01 -6.21883869e-01 -5.91257513e-01
-2.43156970e-01 5.70911705e-01 1.22566134e-01 1.16707876e-01
9.76723313e-01 -5.63829124e-01 -5.22561193e-01 5.85429966e-02
-6.66915298e-01 -3.28912318e-01 7.09883630e-01 8.57110381e-01
3.73237669e-01 2.03931481e-01 9.44927871e-01 -1.31383014e+00
9.04917240e-01 -7.02668309e-01 -7.14940488e-01 -3.98225076e-02
-9.09634590e-01 1.34698033e-01 1.10731173e+00 -1.77022472e-01
-9.13306594e-01 -4.14413065e-02 -1.60790727e-01 1.94966227e-01
-3.20754126e-02 1.16289675e+00 1.36048079e-01 2.27132857e-01
5.64242005e-01 -8.56291950e-02 -2.83934563e-01 -1.79166362e-01
4.52346444e-01 8.98599029e-01 1.68534115e-01 -2.52896100e-01
2.10280851e-01 4.47981626e-01 -1.06026180e-01 -5.52891314e-01
-1.21957600e+00 -5.21501482e-01 -8.03980529e-01 2.03530118e-01
9.14488971e-01 -8.77433538e-01 -6.56437039e-01 3.85565311e-01
-1.12378049e+00 1.34972125e-01 1.55796811e-01 3.05120349e-01
-1.47586882e-01 -1.45249413e-02 -5.54999053e-01 -7.70250022e-01
-5.58813393e-01 -9.10856128e-01 8.53727400e-01 2.70996988e-01
-5.87024212e-01 -1.20182431e+00 -4.26957831e-02 1.82936296e-01
1.59935623e-01 1.43141106e-01 1.21082425e+00 -1.14314938e+00
2.55468488e-01 -4.72823381e-01 -1.77164704e-01 7.88286448e-01
2.19767943e-01 -5.07863984e-03 -8.31711233e-01 2.49541894e-01
-7.94908330e-02 -4.31482196e-01 7.30612755e-01 4.10785466e-01
8.98902774e-01 -4.74286228e-01 -1.65926948e-01 1.16755717e-01
1.33483148e+00 5.72641790e-01 4.70426023e-01 7.02019095e-01
6.67127848e-01 9.79436219e-01 7.04127908e-01 3.64945084e-01
6.38763309e-01 1.76490411e-01 4.67764378e-01 -6.16954602e-02
6.56184435e-01 -1.95544176e-02 5.01556456e-01 1.03447676e+00
1.89897999e-01 -2.25269929e-01 -9.39388216e-01 6.51853383e-01
-1.82644284e+00 -6.71819150e-01 -2.72969306e-02 1.78908288e+00
7.99505830e-01 8.04590881e-01 -4.84904312e-02 4.78931427e-01
5.89780748e-01 2.40717307e-01 -5.39667845e-01 -1.12656879e+00
-1.70505702e-01 3.80506128e-01 5.34758925e-01 3.67090344e-01
-1.33657575e+00 1.24285376e+00 5.26392746e+00 3.80371720e-01
-1.58152711e+00 -4.46165614e-02 7.15530634e-01 1.26154482e-01
-7.52084181e-02 -2.55551606e-01 -1.03604889e+00 1.79230243e-01
1.13233554e+00 9.06084701e-02 8.15622434e-02 1.11360431e+00
1.80243161e-02 3.26734781e-02 -9.33267772e-01 7.41618156e-01
2.33561710e-01 -1.30736423e+00 -4.41857837e-02 -2.31796771e-01
7.86090732e-01 6.14313707e-02 2.65409827e-01 5.34220695e-01
3.39917958e-01 -1.13736713e+00 5.16442657e-01 1.07765988e-01
2.05167770e-01 -1.09694850e+00 1.24330652e+00 4.89193462e-02
-1.15949547e+00 -4.37770039e-01 -2.49116793e-01 -5.30547857e-01
4.82589798e-03 6.98571861e-01 -8.79369497e-01 4.53711897e-01
7.30313599e-01 1.20734131e+00 -4.52592820e-01 7.47480765e-02
-4.20339882e-01 5.26212871e-01 -1.07263945e-01 -6.29439056e-01
6.00186646e-01 -2.35177040e-01 4.24093269e-02 1.34689474e+00
1.07789740e-01 -1.81938425e-01 1.26257483e-02 3.23607415e-01
-1.16759509e-01 6.24868870e-01 -6.85054302e-01 -3.68486017e-01
7.34567642e-03 1.55232430e+00 -1.06663918e+00 -4.38731998e-01
-6.99499726e-01 6.33489847e-01 2.39858255e-01 1.03234023e-01
-5.89394689e-01 -6.18956387e-01 5.92532277e-01 -1.14618279e-02
5.46371222e-01 -1.09998889e-01 -7.03665078e-01 -1.04927492e+00
-1.29489107e-02 -9.57513213e-01 3.64989847e-01 -5.46984017e-01
-1.40165114e+00 9.53608871e-01 -3.63778025e-01 -1.18706691e+00
-4.67250973e-01 -1.22683108e+00 -4.58640873e-01 6.51072443e-01
-1.88861310e+00 -1.05930424e+00 2.05622330e-01 4.41416413e-01
4.29458320e-01 -4.17899936e-01 9.10104573e-01 2.54719049e-01
-5.01393795e-01 3.53738815e-01 -8.44597816e-05 2.18888700e-01
5.67659974e-01 -1.20425701e+00 2.29233831e-01 5.14290690e-01
-4.01573814e-02 7.33247995e-01 6.41329348e-01 -5.22118092e-01
-1.28147030e+00 -1.06728959e+00 1.33701861e+00 -1.24323264e-01
1.09912741e+00 -4.57629234e-01 -4.51743364e-01 4.26011801e-01
3.54769647e-01 -1.70203187e-02 1.00717211e+00 5.98956883e-01
-4.20783639e-01 -3.76209974e-01 -9.35625732e-01 6.07852459e-01
3.30045134e-01 -4.14135903e-01 -6.65485620e-01 3.14278871e-01
7.35715449e-01 -6.00817800e-02 -6.73754811e-01 4.88987356e-01
7.91492879e-01 -9.13446367e-01 5.71515381e-01 -7.99372852e-01
1.01236284e+00 -2.80960977e-01 -3.45498919e-01 -1.24662292e+00
4.55879280e-03 -2.97791541e-01 4.07499820e-02 9.49195743e-01
1.02318478e+00 -6.75174236e-01 6.17995203e-01 1.73150703e-01
-8.68907794e-02 -9.72006798e-01 -4.67439234e-01 -2.67023385e-01
1.10953316e-01 -7.16898143e-01 5.22748828e-01 9.84910429e-01
3.55386555e-01 8.78488362e-01 -1.20654568e-01 5.37847243e-02
-1.06687039e-01 4.25000042e-01 2.90649563e-01 -1.39440143e+00
1.20050184e-01 -6.53329253e-01 -5.57308257e-01 -7.32928634e-01
5.72149932e-01 -8.00945520e-01 -1.59285024e-01 -1.52597737e+00
-4.09584969e-01 -2.32875288e-01 -7.00603843e-01 5.72510719e-01
2.37457722e-01 2.73174673e-01 -8.29841718e-02 -1.19119272e-01
-5.93853414e-01 2.39186049e-01 7.51874149e-01 -2.27355123e-01
-4.49775696e-01 -3.64203714e-02 -1.27483010e+00 1.20457125e+00
1.05108047e+00 -4.73206043e-01 -3.31613153e-01 -7.96933472e-02
1.19471717e+00 -3.43674421e-01 -1.25450253e-01 -6.15091562e-01
2.24242881e-01 -1.17658757e-01 3.19161534e-01 -7.90300667e-01
9.18722972e-02 -6.56800628e-01 -7.03462541e-01 3.63587469e-01
-5.22943497e-01 3.92027974e-01 3.16604555e-01 2.54160136e-01
-6.45905495e-01 -1.87623829e-01 4.77742761e-01 -1.34007528e-01
-6.62381589e-01 -9.43547115e-02 -5.28499246e-01 -2.91317493e-01
8.76761615e-01 4.83633950e-02 -2.62272149e-01 -1.61938459e-01
-5.06014168e-01 2.26441920e-01 -8.77316669e-02 6.82756245e-01
3.74082714e-01 -1.07311535e+00 -3.22390437e-01 1.94537967e-01
1.01585850e-01 -6.42580912e-02 -1.79196969e-02 7.80493081e-01
-5.26069283e-01 8.08567166e-01 -6.16613962e-02 -3.30362946e-01
-1.04366398e+00 5.65358877e-01 1.46970317e-01 -5.20905077e-01
-8.52517486e-02 7.67118037e-01 -1.22547373e-02 -5.81673443e-01
-8.43118206e-02 -7.61164844e-01 -1.09955132e+00 7.84563959e-01
3.67726892e-01 4.27736975e-02 4.67327178e-01 -7.65386641e-01
-6.37961984e-01 6.57105505e-01 -3.85770559e-01 -3.77982594e-02
1.47608411e+00 1.01824619e-01 -4.62476104e-01 6.74231887e-01
1.19500852e+00 1.75115585e-01 -4.50234622e-01 7.01830983e-02
3.13093662e-01 2.69010216e-01 1.81773305e-01 -9.08167541e-01
-9.92662132e-01 9.66163993e-01 2.53507763e-01 6.37458861e-01
1.15108836e+00 -2.76198417e-01 8.04391682e-01 8.94632578e-01
-5.73497862e-02 -1.30077577e+00 -2.99751144e-02 9.79782522e-01
5.37932575e-01 -1.34492326e+00 -2.13515526e-03 -1.30901977e-01
-8.56933475e-01 1.41726553e+00 6.36849523e-01 -4.52574283e-01
1.00749111e+00 3.48450392e-01 2.99238831e-01 -2.62674451e-01
-8.49577427e-01 -3.28022130e-02 1.56307191e-01 2.07059860e-01
7.89968252e-01 4.88014221e-02 -4.37646389e-01 1.06952059e+00
-6.43278241e-01 1.84834120e-03 3.93716067e-01 9.23978388e-01
-3.07729095e-01 -1.13923144e+00 -9.25482810e-02 5.80907702e-01
-9.45218623e-01 -6.22119129e-01 -3.44585389e-01 5.38264990e-01
3.32385808e-01 1.23122036e+00 2.14096606e-01 -4.42860872e-01
3.15836281e-01 4.81649339e-02 -5.95733598e-02 -7.68632472e-01
-8.35908353e-01 3.26267108e-02 3.97201896e-01 -2.76751429e-01
-8.34789753e-01 -6.43610656e-01 -1.36455262e+00 -1.59646481e-01
-3.47663462e-01 9.66285616e-02 9.34937298e-01 1.19818032e+00
3.17142308e-01 5.47759414e-01 7.39558399e-01 -3.89124095e-01
-1.98673040e-01 -9.83833909e-01 -3.98318827e-01 4.10829596e-02
3.05230707e-01 -6.24738455e-01 -3.04697156e-01 8.54827687e-02] | [11.165939331054688, 7.006159782409668] |
59d020a5-51f9-4b3b-b9d6-bc5cc4ce5388 | smoothnets-optimizing-cnn-architecture-design | 2205.04095 | null | https://arxiv.org/abs/2205.04095v1 | https://arxiv.org/pdf/2205.04095v1.pdf | SmoothNets: Optimizing CNN architecture design for differentially private deep learning | The arguably most widely employed algorithm to train deep neural networks with Differential Privacy is DPSGD, which requires clipping and noising of per-sample gradients. This introduces a reduction in model utility compared to non-private training. Empirically, it can be observed that this accuracy degradation is strongly dependent on the model architecture. We investigated this phenomenon and, by combining components which exhibit good individual performance, distilled a new model architecture termed SmoothNet, which is characterised by increased robustness to the challenges of DP-SGD training. Experimentally, we benchmark SmoothNet against standard architectures on two benchmark datasets and observe that our architecture outperforms others, reaching an accuracy of 73.5\% on CIFAR-10 at $\varepsilon=7.0$ and 69.2\% at $\varepsilon=7.0$ on ImageNette, a state-of-the-art result compared to prior architectural modifications for DP. | ['Georgios Kaissis', 'Daniel Rueckert', 'Alexander Ziller', 'Nicolas W. Remerscheid'] | 2022-05-09 | null | null | null | null | ['image-classification-with-dp'] | ['computer-vision'] | [ 1.41719386e-01 1.73699677e-01 1.09049208e-01 -4.69012946e-01
-7.36325681e-01 -5.05756497e-01 6.48739517e-01 -1.16037086e-01
-9.82479095e-01 8.68802249e-01 -3.88856053e-01 -3.79810244e-01
-6.05905382e-03 -5.41791797e-01 -9.34928715e-01 -8.68744671e-01
-2.82298028e-01 -4.70211841e-02 1.18819639e-01 6.35763183e-02
-1.16457865e-01 6.10631049e-01 -1.20660830e+00 2.16373384e-01
4.58265752e-01 1.45100927e+00 -3.66574377e-01 4.68868554e-01
4.29108620e-01 6.86006784e-01 -6.71803594e-01 -1.05307508e+00
7.01915383e-01 -8.96805674e-02 -4.44018900e-01 -5.27589381e-01
8.21429968e-01 -4.50046867e-01 -3.52147996e-01 1.29029381e+00
7.16136217e-01 -8.48798528e-02 4.74560678e-01 -1.04326630e+00
-5.46102583e-01 7.17049897e-01 -3.08823705e-01 2.57041901e-01
-2.78241366e-01 3.48763347e-01 1.15569901e+00 -6.31978869e-01
4.44985390e-01 1.02345049e+00 8.73550653e-01 7.53729165e-01
-1.61270642e+00 -8.27872396e-01 1.89356226e-02 -1.77596621e-02
-1.42832482e+00 -5.81943154e-01 4.06276852e-01 -1.41585201e-01
1.02396226e+00 1.98182464e-01 3.82346183e-01 1.52460694e+00
1.88671798e-01 9.14767385e-01 1.24639082e+00 -1.27496868e-01
5.06835759e-01 3.08548898e-01 1.78461328e-01 4.53638434e-01
2.62114465e-01 3.46050054e-01 -4.18797165e-01 -1.68095708e-01
3.92537206e-01 -1.97050408e-01 -3.68861616e-01 -3.36135715e-01
-6.36058688e-01 7.91147351e-01 5.87447047e-01 1.37612164e-01
-3.36411119e-01 5.24950027e-01 5.92277467e-01 6.13479614e-01
5.30569375e-01 3.95471156e-01 -6.92847550e-01 -1.29177809e-01
-1.05959344e+00 4.39521521e-01 8.30450475e-01 7.45675027e-01
6.48969829e-01 2.51749694e-01 -1.44723356e-01 6.09914660e-01
2.13552281e-01 3.05052787e-01 2.89072335e-01 -1.08308375e+00
3.50587487e-01 4.30581868e-01 -2.71905005e-01 -7.61128008e-01
-1.80292517e-01 -1.11170805e+00 -1.15107405e+00 5.13049006e-01
6.04874551e-01 -5.28219759e-01 -6.87616944e-01 1.90546191e+00
1.21630379e-03 3.10041636e-01 3.37126672e-01 6.08734429e-01
6.23288453e-01 2.77252793e-01 6.28760755e-02 2.10784480e-01
1.01704872e+00 -7.83968210e-01 -2.07904071e-01 -4.04774874e-01
7.02220678e-01 -3.17952573e-01 8.96252096e-01 5.01460195e-01
-1.19632292e+00 -3.22684973e-01 -1.27430224e+00 2.75742002e-02
-3.23882848e-01 1.00656368e-01 3.89788300e-01 1.12285233e+00
-1.50225008e+00 1.11794305e+00 -8.90304983e-01 1.89439565e-01
1.10540497e+00 5.65750539e-01 -4.42586362e-01 -1.79226603e-02
-1.21141303e+00 5.94096184e-01 3.88024062e-01 3.92420404e-02
-9.22673821e-01 -1.25739932e+00 -5.56805551e-01 3.24230909e-01
9.51750726e-02 -5.51278293e-01 1.20198262e+00 -9.73690987e-01
-1.53011990e+00 1.18683469e+00 3.59990031e-01 -1.57361817e+00
1.10545623e+00 -2.53632903e-01 -2.78180063e-01 -6.07444271e-02
-4.79835153e-01 7.70665526e-01 9.04967964e-01 -9.65718567e-01
-6.30716145e-01 -4.84668911e-01 7.98746422e-02 -2.30343178e-01
-4.67971653e-01 -2.07423553e-01 -2.05679223e-01 -4.78519112e-01
-3.54296237e-01 -9.81361449e-01 -1.86800748e-01 1.09228097e-01
-4.02636230e-01 6.83583245e-02 8.44886720e-01 -4.60476011e-01
9.66235876e-01 -2.35483313e+00 -1.89330265e-01 3.32038373e-01
4.29178178e-01 7.40876019e-01 -9.33474898e-02 5.28145768e-02
-5.48707992e-02 2.32143506e-01 -4.65222776e-01 -8.39046180e-01
4.31201369e-01 1.18046016e-01 -1.48686543e-01 6.17970109e-01
1.99354738e-01 8.63482475e-01 -5.47769010e-01 7.81160146e-02
-5.34313312e-03 8.14884901e-01 -7.99875677e-01 -1.78770825e-01
-8.88083428e-02 -1.48870191e-02 -1.31034866e-01 2.40117207e-01
1.06753838e+00 -1.09062411e-01 -1.57143045e-02 1.01993896e-01
8.62476826e-02 1.55931056e-01 -1.02407110e+00 1.58182764e+00
-1.88089997e-01 8.28313887e-01 2.38795191e-01 -1.03246963e+00
9.52137351e-01 2.01237738e-01 2.20911458e-01 -6.87638760e-01
4.08352196e-01 3.64070028e-01 8.22687745e-02 1.31292820e-01
1.92182466e-01 7.68922791e-02 2.20747203e-01 2.26758420e-01
1.97888613e-01 2.75831193e-01 -1.26235887e-01 -4.44409102e-02
1.34000647e+00 -3.51391107e-01 -1.53702602e-01 -6.08427286e-01
4.68322307e-01 -7.45492637e-01 3.01790774e-01 9.82673585e-01
-5.07985234e-01 6.05645001e-01 9.77159858e-01 -4.30972725e-01
-7.58259892e-01 -9.16300237e-01 -3.62146795e-01 9.49377656e-01
-5.17516553e-01 -1.92903191e-01 -1.20804203e+00 -9.44911897e-01
-1.03979409e-02 8.34453523e-01 -8.35222960e-01 -3.71667385e-01
-2.88174957e-01 -9.82156336e-01 9.60748553e-01 4.00518268e-01
9.75582719e-01 -8.55686784e-01 -7.81230092e-01 -4.00219187e-02
6.40420556e-01 -1.13996232e+00 -1.39434904e-01 4.49069083e-01
-6.95652306e-01 -6.33856118e-01 -8.70857835e-01 -4.24065083e-01
4.16325629e-01 -2.35175416e-01 1.27363765e+00 8.15977901e-02
6.35037571e-02 1.37755334e-01 -8.29098746e-02 -5.14652669e-01
-3.66839409e-01 6.34628296e-01 -2.64136672e-01 2.17715055e-01
3.55973482e-01 -8.50293815e-01 -8.96554887e-01 1.52405918e-01
-1.05282927e+00 -3.25699002e-01 5.74455559e-01 8.99392247e-01
5.09818196e-01 -2.29387939e-01 3.91042173e-01 -1.12074351e+00
6.00510001e-01 -3.50420058e-01 -7.62186468e-01 -1.55860677e-01
-8.26909363e-01 1.65166229e-01 7.62505412e-01 -1.59036964e-01
-8.15066457e-01 1.52784791e-02 -5.44853628e-01 -5.96231759e-01
-2.05179349e-01 1.87089160e-01 -1.49045557e-01 -2.68722743e-01
9.51445103e-01 1.21717356e-01 6.64108917e-02 -5.60296953e-01
2.08747730e-01 3.38839620e-01 4.25910830e-01 -2.95192391e-01
4.92211372e-01 5.16249359e-01 1.93090662e-01 -7.15513527e-01
-6.58824682e-01 1.24081902e-01 -2.97003150e-01 2.77614921e-01
6.23526275e-01 -9.93847609e-01 -7.88335621e-01 8.21408927e-01
-7.92158425e-01 -6.65479124e-01 -5.59220850e-01 3.09391111e-01
-2.98679501e-01 6.67031296e-03 -5.03039658e-01 -5.57347655e-01
-6.13015950e-01 -1.10782874e+00 6.89856410e-01 -2.43255701e-02
3.74404192e-02 -1.22452927e+00 -2.51286298e-01 4.75130565e-02
7.81991124e-01 4.80622053e-01 6.35506690e-01 -1.04289031e+00
-4.66321558e-01 -3.93356413e-01 -3.71510476e-01 1.13513589e+00
-4.84839946e-01 -1.55001134e-01 -1.58283699e+00 -5.58746397e-01
2.38974378e-01 -2.96643823e-01 1.03742504e+00 4.33013767e-01
1.22091079e+00 -5.22436202e-01 7.99732432e-02 1.01610231e+00
1.54199994e+00 1.23572478e-03 9.16308522e-01 4.10904408e-01
3.19863230e-01 1.77117020e-01 -1.63216367e-01 4.53523785e-01
-2.09024232e-02 4.67800796e-01 8.05652499e-01 -1.43403411e-01
3.09338067e-02 -4.77090254e-02 3.17944765e-01 1.29513547e-01
2.16154903e-01 -1.31000146e-01 -8.49946499e-01 4.02832299e-01
-1.42388320e+00 -7.55807519e-01 -8.84630084e-02 2.08804917e+00
7.14769363e-01 6.40503466e-01 1.47768082e-02 7.84079581e-02
2.31069252e-01 4.56011683e-01 -6.80802643e-01 -5.43210268e-01
-3.59975308e-01 5.54717779e-01 9.98304307e-01 1.56562120e-01
-1.08084488e+00 8.69471550e-01 5.82543421e+00 9.46360350e-01
-1.32165134e+00 1.75704524e-01 1.22836971e+00 -3.52080196e-01
-8.20432156e-02 -4.28672642e-01 -8.46569777e-01 5.97656071e-01
1.39826107e+00 -8.94168019e-03 2.75058001e-01 1.12735558e+00
-3.23191166e-01 2.67311901e-01 -1.30134714e+00 9.85419452e-01
-1.83434904e-01 -1.33726120e+00 -3.26003522e-01 2.30868444e-01
9.14990604e-01 5.85203052e-01 6.57614350e-01 3.56953442e-01
4.27044332e-01 -1.11906862e+00 6.80854261e-01 1.87039748e-01
6.80153787e-01 -9.67327356e-01 9.02478337e-01 2.34062389e-01
-5.50420761e-01 7.51302615e-02 -3.74146998e-01 -3.93154938e-03
-2.84065366e-01 6.71779156e-01 -6.47571564e-01 2.88994044e-01
9.68719482e-01 6.92379177e-01 -5.95963955e-01 9.56491351e-01
-2.19323650e-01 7.85648227e-01 -4.82357085e-01 6.48848414e-02
5.48593819e-01 1.21364435e-02 4.00148690e-01 1.23300743e+00
3.20138365e-01 -3.48633140e-01 -6.39583349e-01 8.67064655e-01
-6.63903177e-01 -1.22091718e-01 -4.14189577e-01 3.27648818e-01
2.75173724e-01 9.55251873e-01 -2.30653167e-01 -2.23082870e-01
-2.56286263e-01 1.04851186e+00 3.36484849e-01 3.03831011e-01
-9.00346577e-01 -2.54296601e-01 9.00947988e-01 1.05708223e-02
6.65095747e-01 7.38338828e-02 -2.44689777e-01 -9.07638252e-01
2.04688281e-01 -9.86638904e-01 3.72803837e-01 -1.76384106e-01
-1.18532968e+00 7.64098644e-01 -1.95717767e-01 -8.13265920e-01
-1.18997663e-01 -7.16897070e-01 -3.86904776e-01 8.50284874e-01
-1.58468390e+00 -7.73768127e-01 8.94578695e-02 5.68191469e-01
-9.36611965e-02 -3.19908977e-01 8.65007460e-01 4.74723548e-01
-6.08344376e-01 1.37815106e+00 6.09940350e-01 1.46891803e-01
2.99736112e-01 -1.20838797e+00 8.97031307e-01 8.13745856e-01
1.56199515e-01 3.61864954e-01 5.81394315e-01 1.42976046e-01
-1.10289919e+00 -1.28007853e+00 6.94850206e-01 -3.45747411e-01
5.02057254e-01 -4.86810088e-01 -1.02141631e+00 8.17392588e-01
2.73920655e-01 4.32607412e-01 5.88878036e-01 -2.11524770e-01
-5.61568201e-01 -4.72351879e-01 -1.54007649e+00 6.32200956e-01
1.09233260e+00 -4.95715320e-01 -1.25058353e-01 1.62826896e-01
7.06693888e-01 -5.38810670e-01 -8.93003643e-01 1.75335705e-01
5.19383311e-01 -1.53641975e+00 9.21986222e-01 -4.09854144e-01
2.93066084e-01 3.67046058e-01 -2.75358677e-01 -1.28766370e+00
-1.18364673e-02 -9.68810558e-01 -1.07696593e-01 1.08283401e+00
6.72791481e-01 -1.20296669e+00 1.28624070e+00 9.55302477e-01
-8.62098113e-03 -9.18460608e-01 -1.32153416e+00 -9.04658973e-01
4.54603851e-01 -4.70547855e-01 6.31904304e-01 7.99131572e-01
-7.58744657e-01 -5.65129817e-02 -3.27725232e-01 1.24261044e-02
7.72668183e-01 -6.08145535e-01 6.84113503e-01 -1.10907924e+00
-3.29405963e-01 -6.38518333e-01 -5.72941005e-01 -9.41892505e-01
3.70076776e-01 -9.13812876e-01 -4.70029294e-01 -7.74195313e-01
-3.29851985e-01 -3.96323234e-01 -7.35718548e-01 4.18840408e-01
2.67139584e-01 4.16812152e-01 3.30348283e-01 -1.19059570e-01
-3.58461887e-01 4.88583654e-01 7.89163828e-01 -3.80277820e-02
-8.59789774e-02 1.59398437e-01 -7.51960456e-01 5.88750124e-01
1.01457930e+00 -6.13182902e-01 -3.70597631e-01 -7.23658919e-01
1.88100308e-01 -4.75303382e-01 5.89803934e-01 -1.43165112e+00
-6.65235445e-02 4.59778786e-01 1.74819350e-01 -2.01415554e-01
4.28336352e-01 -9.70458508e-01 2.15170592e-01 6.55185580e-01
-5.42912960e-01 -6.43708333e-02 4.66459990e-01 4.23452735e-01
-1.65434275e-02 -1.17023528e-01 1.07087386e+00 4.99278195e-02
-4.10329223e-01 4.71577138e-01 3.99891194e-03 2.64613599e-01
9.78188694e-01 -6.92413226e-02 -3.10829192e-01 -1.63200438e-01
-4.66542453e-01 -2.63662878e-02 3.76609057e-01 5.71373627e-02
3.96153122e-01 -1.09685731e+00 -7.80441999e-01 5.24811685e-01
-3.00483227e-01 6.27622753e-02 4.30693507e-01 7.39347339e-01
-6.37029231e-01 3.83360952e-01 -1.26371115e-01 -5.78166008e-01
-1.14079249e+00 5.84763624e-02 7.81012654e-01 -4.23383057e-01
-5.13829768e-01 1.44057512e+00 2.09170766e-02 -2.74394035e-01
5.07211804e-01 -5.10842741e-01 2.22386852e-01 -1.50859311e-01
4.07965153e-01 3.59973639e-01 5.19315183e-01 -4.19274867e-01
-3.68382871e-01 7.66188977e-03 -3.84886205e-01 -1.57081801e-02
1.50058603e+00 3.16921055e-01 1.22744821e-01 9.56444442e-02
1.80713236e+00 -4.41274613e-01 -1.90364134e+00 -3.48561823e-01
-1.98687464e-01 -3.81187469e-01 2.38017082e-01 -7.90027142e-01
-1.55371785e+00 1.11566484e+00 1.04119468e+00 3.30770314e-01
1.06567538e+00 -4.60925162e-01 6.45833731e-01 4.74924475e-01
3.26014668e-01 -8.77894759e-01 -2.87422448e-01 5.64247906e-01
7.43376791e-01 -1.19843280e+00 -5.25759868e-02 2.15346932e-01
-4.35053855e-01 6.01534545e-01 3.19831103e-01 -4.81318057e-01
9.52546895e-01 2.50114948e-01 -6.41414374e-02 -4.98718806e-02
-8.26848745e-01 4.23512191e-01 1.85323328e-01 4.40642357e-01
1.63906179e-02 -7.90442377e-02 -1.18825771e-02 6.29776537e-01
-3.99154097e-01 1.36418976e-02 7.41012469e-02 8.93949986e-01
8.73143002e-02 -9.49054599e-01 1.08903438e-01 4.02900308e-01
-1.11990559e+00 -2.79131234e-01 -3.80659103e-01 9.24916327e-01
2.17781588e-02 4.72183079e-01 1.07280694e-01 -4.05317307e-01
3.79091084e-01 1.03026316e-01 1.07270837e-01 -9.03678089e-02
-1.07416773e+00 -4.94292200e-01 3.76004577e-02 -6.49318457e-01
-1.41594827e-01 -7.42987096e-01 -9.51366186e-01 -5.65534294e-01
-5.03181741e-02 -1.06228784e-01 7.67438829e-01 5.99943221e-01
7.17070580e-01 3.64840806e-01 5.12517214e-01 -6.10730708e-01
-9.45202053e-01 -6.15971327e-01 -6.87592924e-01 4.42154884e-01
5.30126691e-01 -1.89305753e-01 -7.36320198e-01 -2.94692069e-01] | [5.890960216522217, 6.95286750793457] |
2eeeab83-18e8-4d0b-b367-245382c08b27 | label-embedding-for-image-classification | 1503.08677 | null | http://arxiv.org/abs/1503.08677v2 | http://arxiv.org/pdf/1503.08677v2.pdf | Label-Embedding for Image Classification | Attributes act as intermediate representations that enable parameter sharing
between classes, a must when training data is scarce. We propose to view
attribute-based image classification as a label-embedding problem: each class
is embedded in the space of attribute vectors. We introduce a function that
measures the compatibility between an image and a label embedding. The
parameters of this function are learned on a training set of labeled samples to
ensure that, given an image, the correct classes rank higher than the incorrect
ones. Results on the Animals With Attributes and Caltech-UCSD-Birds datasets
show that the proposed framework outperforms the standard Direct Attribute
Prediction baseline in a zero-shot learning scenario. Label embedding enjoys a
built-in ability to leverage alternative sources of information instead of or
in addition to attributes, such as e.g. class hierarchies or textual
descriptions. Moreover, label embedding encompasses the whole range of learning
settings from zero-shot learning to regular learning with a large number of
labeled examples. | ['Cordelia Schmid', 'Zeynep Akata', 'Florent Perronnin', 'Zaid Harchaoui'] | 2015-03-30 | null | null | null | null | ['zero-shot-action-recognition', 'multi-label-zero-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 4.62361991e-01 2.76620686e-01 -7.56567717e-01 -9.33792591e-01
-6.46103144e-01 -4.30444926e-01 1.02386904e+00 4.23235536e-01
-5.90345144e-01 6.95110440e-01 2.36861035e-01 1.42969757e-01
-1.37676597e-01 -8.43020737e-01 -6.62083983e-01 -8.39637518e-01
-4.49142791e-02 7.30072320e-01 2.90068518e-03 -3.19341682e-02
6.99273450e-03 1.69311687e-01 -1.96677637e+00 3.13097149e-01
3.06182653e-01 1.37119401e+00 -4.63868268e-02 3.14852893e-01
-1.76782623e-01 5.51469624e-01 -1.14046462e-01 -5.25577188e-01
3.90307277e-01 -1.12643637e-01 -7.27577984e-01 2.89168715e-01
6.55564308e-01 -2.79430598e-01 -2.18633160e-01 9.28918958e-01
1.87155187e-01 2.02135265e-01 1.14650297e+00 -1.91797650e+00
-8.51979375e-01 4.08793330e-01 -1.00207426e-01 -1.26980141e-01
1.50362387e-01 7.72013068e-02 1.67449474e+00 -1.02671409e+00
7.37096548e-01 1.18018758e+00 4.55114752e-01 6.14484966e-01
-1.72341871e+00 -5.84536791e-01 1.70584098e-01 3.92055035e-01
-1.06220007e+00 -3.70515019e-01 5.45224011e-01 -9.13285375e-01
6.03417635e-01 -5.83028309e-02 5.28165579e-01 1.37652647e+00
-5.87943010e-02 5.70040703e-01 1.06336689e+00 -3.74207258e-01
5.75246930e-01 6.23233020e-01 3.40701044e-01 4.94201213e-01
1.62424162e-01 1.97453097e-01 -7.40747452e-01 -3.61405998e-01
1.93865359e-01 3.08153540e-01 -9.96322110e-02 -1.18029571e+00
-1.28749025e+00 1.20088530e+00 5.35632133e-01 -1.77091792e-01
-3.04723084e-01 1.82502702e-01 4.44404066e-01 4.88884360e-01
4.87502962e-01 6.24579608e-01 -4.86842901e-01 4.20830309e-01
-4.82217222e-01 6.75677881e-02 5.96604049e-01 1.29402888e+00
1.27890801e+00 -1.76174164e-01 -3.16442221e-01 8.85901093e-01
3.04984212e-01 3.23252887e-01 4.91371870e-01 -1.06587625e+00
1.70295715e-01 5.75619817e-01 4.11388986e-02 -6.39169455e-01
-4.62955236e-02 -2.68438429e-01 -6.36954963e-01 4.82683867e-01
2.03816205e-01 3.81420791e-01 -9.22992647e-01 2.01319647e+00
4.43878025e-01 3.15160543e-01 1.55331016e-01 6.41543031e-01
5.67425370e-01 5.06329119e-01 5.00210881e-01 -1.46839889e-02
1.38348353e+00 -8.29258621e-01 -5.91352403e-01 -3.89976710e-01
6.90151393e-01 -3.27218354e-01 1.21600008e+00 7.12408945e-02
-6.64357781e-01 -5.36786258e-01 -1.36676860e+00 -9.84461680e-02
-7.89416134e-01 -9.73038301e-02 6.47001326e-01 4.93297309e-01
-8.34568977e-01 7.71871507e-01 -3.65494370e-01 -5.03943861e-01
6.43456697e-01 3.23900402e-01 -7.54379570e-01 -1.29177704e-01
-1.15692067e+00 7.74250805e-01 5.27147174e-01 -6.21868074e-01
-1.07886565e+00 -8.86989176e-01 -1.08918452e+00 1.40261039e-01
4.29721683e-01 -6.93684280e-01 1.02004516e+00 -1.05122983e+00
-9.83478010e-01 1.36891258e+00 2.89487332e-01 -5.44570565e-01
3.63475770e-01 -7.72721991e-02 -1.63591474e-01 1.07518531e-01
2.92422563e-01 1.06205690e+00 1.15500355e+00 -1.27637339e+00
-8.47392142e-01 -4.49409395e-01 2.51874357e-01 1.85948834e-01
-6.72570288e-01 -4.71459746e-01 1.28909618e-01 -5.46439052e-01
-4.58784215e-02 -1.16972828e+00 -5.42048030e-02 6.12423837e-01
-4.01750535e-01 -2.69821376e-01 6.89469397e-01 2.66531110e-02
7.11642802e-01 -2.32110715e+00 2.41569877e-01 1.69712812e-01
3.31801593e-01 -1.15825936e-01 -2.10395098e-01 3.80278438e-01
-2.82155335e-01 1.23677582e-01 -2.71483570e-01 -3.01205933e-01
3.30299228e-01 4.83998179e-01 -3.50084841e-01 6.08350992e-01
2.66243696e-01 5.68992496e-01 -9.15158272e-01 -5.51156759e-01
1.95397794e-01 4.09040838e-01 -4.79942292e-01 3.74687552e-01
-2.47668058e-01 2.06316322e-01 -3.54597747e-01 5.55451155e-01
1.41515747e-01 -4.69220102e-01 4.57751416e-02 -1.92083105e-01
3.97995710e-01 3.04194330e-03 -1.19257939e+00 1.56486619e+00
-4.16363418e-01 3.64489913e-01 -3.56893092e-01 -9.12433565e-01
8.16200972e-01 3.63131642e-01 5.05827546e-01 -2.97802746e-01
7.16243498e-03 7.75048658e-02 -1.80790544e-01 -2.35751346e-01
2.50903249e-01 -1.99035689e-01 -2.57262349e-01 6.14303112e-01
7.14116216e-01 1.61276028e-01 1.57004684e-01 3.53854269e-01
7.79740810e-01 1.33853212e-01 7.62630522e-01 -3.22680801e-01
3.30096215e-01 -1.94046780e-01 6.05998874e-01 7.43735850e-01
-2.03271762e-01 3.22963506e-01 4.99645263e-01 -6.16449535e-01
-1.52132142e+00 -1.09182525e+00 -5.32234550e-01 1.72431684e+00
8.19111094e-02 -4.07684237e-01 -3.97231847e-01 -8.06727529e-01
2.99498320e-01 7.18952775e-01 -1.10255885e+00 -5.43147266e-01
3.16524692e-02 -4.11341161e-01 5.29064015e-02 5.91218770e-01
1.01187833e-01 -9.95870352e-01 -7.67765462e-01 5.97099513e-02
4.98689935e-02 -1.06395149e+00 -3.36019039e-01 6.83645844e-01
-6.33427680e-01 -1.13410902e+00 -4.10528988e-01 -7.61453032e-01
5.54395556e-01 3.58838551e-02 1.26938498e+00 1.19617628e-02
-3.91356647e-01 5.88652670e-01 -4.72523481e-01 -2.53548533e-01
-3.80820125e-01 -6.62588840e-03 1.79365486e-01 3.38034987e-01
4.48845357e-01 -5.90039968e-01 -6.46319032e-01 3.28528851e-01
-7.24077463e-01 -3.85913812e-02 5.49477875e-01 1.26433039e+00
7.00366199e-01 -6.18710220e-01 6.61462545e-01 -1.21606791e+00
1.00778461e-01 -6.37306631e-01 -4.09539014e-01 4.20400947e-01
-8.96417379e-01 3.37486506e-01 3.08545291e-01 -6.84398353e-01
-6.91415668e-01 4.02139187e-01 3.31317544e-01 -3.94866556e-01
-3.37016463e-01 1.61157131e-01 -3.57574582e-01 1.43465921e-01
6.76014125e-01 -4.96661551e-02 1.25035748e-01 -3.11779469e-01
8.31922650e-01 5.08601189e-01 4.90131408e-01 -6.57312512e-01
6.29191399e-01 4.57666814e-01 1.86219558e-01 -5.60495555e-01
-1.09608245e+00 -5.90891838e-01 -9.64406729e-01 -6.76355734e-02
9.79435086e-01 -8.38407695e-01 -4.31750476e-01 -8.01320076e-02
-6.55260026e-01 -1.29376680e-01 -7.28619576e-01 4.99388188e-01
-1.00331354e+00 -1.25585511e-01 -4.74554002e-01 -4.34734702e-01
2.24836338e-02 -1.13310492e+00 9.80206192e-01 4.29498740e-02
-2.92879581e-01 -9.61124241e-01 -7.32824858e-03 5.94465015e-03
1.67031750e-01 2.32519150e-01 1.54056370e+00 -1.31374788e+00
-4.99949396e-01 -2.73226351e-01 -8.10769498e-02 2.44127691e-01
1.71465799e-01 -2.54368395e-01 -1.37330103e+00 -4.86669868e-01
-2.98378438e-01 -9.34699476e-01 9.90257978e-01 1.31549224e-01
1.10248506e+00 -3.38601291e-01 -4.32395697e-01 6.62582457e-01
1.43879199e+00 -3.24357785e-02 2.58493513e-01 5.23650050e-01
5.07100224e-01 7.71668136e-01 7.38149762e-01 6.25483513e-01
1.71715409e-01 7.84380019e-01 6.76685274e-01 2.18532942e-02
-4.26251218e-02 -3.07918072e-01 -2.09896997e-01 2.03291163e-01
2.97516763e-01 -6.11466095e-02 -8.00484121e-01 3.89340699e-01
-1.73004794e+00 -8.60229969e-01 3.23760241e-01 2.67010570e+00
9.06325877e-01 1.01409599e-01 1.20321669e-01 1.17559731e-01
7.76771069e-01 9.44257155e-02 -7.98541725e-01 -7.83311725e-02
-6.93860427e-02 -1.79332420e-01 3.46433222e-01 3.18895936e-01
-1.47354162e+00 7.53748894e-01 5.95839500e+00 4.04224426e-01
-7.30699658e-01 1.46942064e-01 6.66566968e-01 3.36289369e-02
-3.01049143e-01 2.81648878e-02 -8.50049376e-01 3.37569535e-01
8.61757398e-01 -3.55485827e-01 2.67737240e-01 1.16116726e+00
-6.93911135e-01 2.23242804e-01 -1.91179979e+00 8.03834260e-01
1.37550279e-01 -1.18793011e+00 2.26309121e-01 2.47931257e-01
5.41755259e-01 -1.57122999e-01 3.62019449e-01 4.73415643e-01
5.61583459e-01 -1.03998387e+00 6.50791466e-01 3.32208365e-01
1.08987808e+00 -6.06364071e-01 6.31161451e-01 1.86067268e-01
-9.77214456e-01 -3.96161199e-01 -6.99698746e-01 3.09788167e-01
-9.36795250e-02 1.44324586e-01 -9.38963413e-01 9.46968049e-02
6.50897384e-01 9.78577614e-01 -7.93551087e-01 9.01029050e-01
-1.41298920e-01 3.72930199e-01 1.14053078e-01 3.57076794e-01
3.01944464e-01 -1.23906992e-01 1.88014120e-01 9.13612306e-01
1.48832306e-01 9.24486369e-02 5.20657539e-01 5.66704512e-01
-3.16269845e-01 7.48918280e-02 -9.20329690e-01 3.63140665e-02
7.02804625e-01 1.32895815e+00 -4.72198963e-01 -5.67959905e-01
-5.72459638e-01 8.11736584e-01 4.90958542e-01 3.79138350e-01
-5.38190186e-01 -1.28688008e-01 8.62131476e-01 -2.01584101e-02
2.92738199e-01 2.56801546e-01 -5.91081157e-02 -8.39471161e-01
-4.09664035e-01 -6.22956276e-01 7.12332785e-01 -7.29028821e-01
-1.67371154e+00 5.36170304e-01 1.02894165e-01 -1.54898643e+00
-3.35425556e-01 -7.26835608e-01 -2.46175006e-01 5.83656847e-01
-1.52419662e+00 -1.23605573e+00 -3.37965280e-01 5.83992422e-01
5.22276938e-01 -5.42284966e-01 1.19946325e+00 1.39136270e-01
-9.07622650e-02 6.94936037e-01 1.25557408e-01 4.17148955e-02
9.77662563e-01 -1.45865846e+00 -1.57607161e-02 1.06436566e-01
4.32059377e-01 3.85599047e-01 8.49060476e-01 -3.45477849e-01
-9.23205495e-01 -1.19485557e+00 7.48820066e-01 -4.62552190e-01
6.44785523e-01 -5.20956159e-01 -1.06998873e+00 9.17914450e-01
1.45591319e-01 5.26429296e-01 1.25645387e+00 3.13394994e-01
-9.91190434e-01 -1.95037305e-01 -1.14383769e+00 2.34854162e-01
9.36513841e-01 -7.90906131e-01 -5.50416946e-01 4.09844309e-01
7.88934231e-01 2.51940548e-01 -1.00231969e+00 3.03075194e-01
6.29045606e-01 -5.85691810e-01 1.25435770e+00 -1.16430402e+00
6.25278950e-01 -5.80138899e-02 -7.44261146e-01 -1.60534418e+00
-5.82676291e-01 2.25829147e-02 4.65267850e-03 1.25479305e+00
4.63995636e-01 -5.24690628e-01 6.49562061e-01 6.02051318e-01
5.53808622e-02 -7.37346530e-01 -9.03312027e-01 -9.08888519e-01
7.12046120e-03 9.28426534e-03 6.50247097e-01 1.18936861e+00
-2.22308487e-01 7.48341262e-01 -5.29766679e-01 -5.10466434e-02
1.02929831e+00 1.13002285e-01 5.24241686e-01 -1.84697247e+00
-1.41082525e-01 -1.51177451e-01 -8.93931687e-01 -3.28455627e-01
5.52985311e-01 -1.30421400e+00 -7.26355612e-02 -1.12155843e+00
4.58795607e-01 -6.90355361e-01 -5.81069291e-01 7.56945729e-01
-6.79801963e-03 3.25644016e-01 1.43672973e-01 4.19060618e-01
-6.23496056e-01 7.56169438e-01 7.99098432e-01 -3.85346174e-01
3.36903840e-01 -8.74494538e-02 -4.59389359e-01 7.22716033e-01
6.50188386e-01 -7.65150428e-01 -6.87272310e-01 -5.61045408e-02
2.38328576e-01 2.99426215e-03 4.78005707e-01 -7.71797121e-01
9.27023515e-02 -2.36065626e-01 5.60997248e-01 -1.87295169e-01
6.86087310e-01 -1.15918326e+00 -8.61966014e-02 3.10666859e-01
-1.07364786e+00 -1.37273133e-01 -4.28870380e-01 1.12959087e+00
-1.26610830e-01 -3.25680017e-01 1.05211627e+00 -7.58340284e-02
-9.97112870e-01 6.58694863e-01 4.47758511e-02 2.35606790e-01
1.31170499e+00 -2.72761345e-01 -3.74692619e-01 -2.82864600e-01
-1.03280962e+00 2.44392276e-01 6.90789580e-01 4.42528635e-01
4.83118981e-01 -1.77411675e+00 -5.56595504e-01 3.26925486e-01
9.63189185e-01 -4.85101759e-01 -1.06803760e-01 3.85615706e-01
2.05273852e-01 1.04970992e-01 -6.44112051e-01 -7.82194316e-01
-1.16332531e+00 9.92171943e-01 1.05192803e-01 1.26521826e-01
-6.96373522e-01 7.85464644e-01 5.43423772e-01 -5.82942605e-01
4.64820534e-01 3.06029558e-01 -4.15320367e-01 5.05072236e-01
4.86827821e-01 1.32637531e-01 -9.19823200e-02 -7.43983269e-01
-1.84428245e-01 3.44265491e-01 -1.90044135e-01 -5.63477390e-02
1.32899976e+00 -8.16768631e-02 1.93084896e-01 9.97615993e-01
1.40843630e+00 -7.21342206e-01 -1.44867897e+00 -7.01109946e-01
3.26796770e-01 -7.22565711e-01 -1.22494355e-01 -6.96169198e-01
-7.59782553e-01 1.01649439e+00 9.12059844e-01 1.54522182e-02
7.58709073e-01 3.46512794e-01 1.42115816e-01 6.79587424e-01
3.72845143e-01 -1.21816647e+00 4.95148093e-01 1.06174313e-01
7.07461655e-01 -1.59991264e+00 1.28462538e-02 -2.82669395e-01
-9.02909875e-01 9.87602472e-01 6.64303422e-01 -5.14636189e-02
6.87781751e-01 -1.16004616e-01 -5.25700003e-02 -6.18494600e-02
-1.16851664e+00 -2.50489891e-01 3.64977568e-01 7.78965771e-01
3.02876711e-01 1.76264971e-01 1.36131048e-01 3.66886497e-01
-1.00447778e-02 -3.96250069e-01 4.26113844e-01 8.53540659e-01
-6.75343633e-01 -1.08398807e+00 4.14967276e-02 7.56266892e-01
-1.30767882e-01 4.92281727e-02 -1.78568482e-01 7.10710406e-01
1.23413578e-01 6.68511629e-01 3.06276411e-01 -2.24676803e-01
9.37322229e-02 5.63001633e-01 2.29961589e-01 -9.73284602e-01
-2.50928968e-01 -2.45150611e-01 1.45591469e-03 -5.48138738e-01
-3.64160925e-01 -5.94647169e-01 -7.47970104e-01 6.20638169e-02
-1.77210167e-01 2.99109872e-02 5.23079634e-01 7.61252820e-01
1.29996628e-01 2.57051438e-01 7.05071449e-01 -6.82872832e-01
-9.35576200e-01 -7.32555747e-01 -8.25386107e-01 8.82013381e-01
4.72217917e-01 -1.05569029e+00 -5.06499410e-01 2.36464635e-01] | [9.973225593566895, 2.3623909950256348] |
d1408da0-8cf4-4999-9ff8-0bc55256639b | two-stage-autoencoder-neural-network-for-3d | 2306.04987 | null | https://arxiv.org/abs/2306.04987v2 | https://arxiv.org/pdf/2306.04987v2.pdf | Convolutional Recurrent Neural Network with Attention for 3D Speech Enhancement | 3D speech enhancement can effectively improve the auditory experience and plays a crucial role in augmented reality technology. However, traditional convolutional-based speech enhancement methods have limitations in extracting dynamic voice information. In this paper, we incorporate a dual-path recurrent neural network block into the U-Net to iteratively extract dynamic audio information in both the time and frequency domains. And an attention mechanism is proposed to fuse the original signal, reference signal, and generated masks. Moreover, we introduce a loss function to simultaneously optimize the network in the time-frequency and time domains. Experimental results show that our system outperforms the state-of-the-art systems on the dataset of ICASSP L3DAS23 challenge. | ['Mou Wang', 'Jianfeng Chen', 'Yafei Jia', 'Siwei Huang', 'Jisheng Bai', 'Han Yin'] | 2023-06-08 | null | null | null | null | ['speech-enhancement'] | ['speech'] | [-5.61522394e-02 -1.27121434e-01 1.78415582e-01 -1.08376145e-01
-1.07505214e+00 5.61100505e-02 1.21432729e-01 -3.77683222e-01
-4.85935032e-01 3.44769120e-01 7.98590600e-01 -2.81302512e-01
5.09728119e-02 -4.09826547e-01 -5.01039445e-01 -3.19946736e-01
-5.59840947e-02 -7.08392859e-01 8.22022334e-02 -4.12260205e-01
-2.94059604e-01 2.91892499e-01 -1.51862454e+00 2.33340994e-01
5.72384596e-01 1.08165383e+00 3.05055231e-01 9.20391083e-01
6.09596185e-02 7.68110573e-01 -6.83836281e-01 -1.65561095e-01
1.06281333e-01 -3.79770577e-01 -1.52650073e-01 -2.80210376e-02
1.44847438e-01 -4.89758700e-01 -1.11880922e+00 1.00590718e+00
1.38115013e+00 4.81747895e-01 7.77033046e-02 -8.23749363e-01
-8.17594588e-01 7.01630712e-01 -4.45287585e-01 5.35364389e-01
4.38514352e-01 2.97773397e-03 7.11908519e-01 -1.11718726e+00
2.46084988e-01 1.14702296e+00 7.97298133e-01 4.70831901e-01
-6.61890805e-01 -7.10594654e-01 3.81412923e-01 3.29577476e-01
-1.04563832e+00 -7.59304047e-01 1.07424033e+00 2.14302670e-02
1.23378479e+00 1.46532238e-01 6.67575717e-01 1.19208586e+00
-6.82028681e-02 1.16777134e+00 7.13554680e-01 -2.52634555e-01
-6.90349117e-02 -2.01037273e-01 -1.64151210e-02 3.88988137e-01
-6.26405776e-01 6.75852478e-01 -8.88691127e-01 2.37046987e-01
7.75602221e-01 -2.27383986e-01 -6.03000641e-01 3.40855211e-01
-9.51945066e-01 4.13729697e-01 4.05436099e-01 3.94781888e-01
-4.97022480e-01 2.89163291e-01 5.20791650e-01 3.07145894e-01
7.39406824e-01 2.75666952e-01 -6.92655146e-01 -5.03041387e-01
-7.13492930e-01 -4.26490307e-02 2.18560398e-01 5.57123721e-01
-1.41921178e-01 6.85620844e-01 -2.42554083e-01 1.28402150e+00
4.51089859e-01 3.29349965e-01 7.14525580e-01 -8.05106878e-01
2.59248823e-01 -1.42303243e-01 -7.50421435e-02 -6.79080725e-01
-4.58707929e-01 -9.34878707e-01 -6.99819446e-01 7.00875446e-02
-2.14188874e-01 -4.05023128e-01 -1.09142983e+00 1.69686973e+00
4.27800924e-01 7.31711507e-01 5.22878096e-02 1.13869786e+00
1.18656909e+00 6.45126700e-01 -2.56393194e-01 -2.63245791e-01
1.26638234e+00 -1.15288699e+00 -1.46222794e+00 -1.59418970e-01
-6.82948753e-02 -8.35613847e-01 1.00506651e+00 3.52143079e-01
-1.45050657e+00 -9.16107476e-01 -1.26449645e+00 -1.88951761e-01
-8.67571160e-02 3.36143255e-01 5.10739982e-01 7.87522018e-01
-1.18061852e+00 3.96538556e-01 -8.75092864e-01 2.93201655e-01
8.82654935e-02 2.39154845e-01 -2.61699438e-01 4.06842649e-01
-1.53941143e+00 6.51521981e-01 -7.42560700e-02 3.62121791e-01
-6.91520631e-01 -8.34771693e-01 -1.22908485e+00 2.16287553e-01
2.11205348e-01 -5.68806767e-01 1.64051378e+00 -2.22924262e-01
-2.04841423e+00 2.80771613e-01 -1.85387105e-01 -6.88726664e-01
3.50793034e-01 -5.67738116e-01 -9.15842891e-01 1.32533625e-01
-2.58389413e-01 5.62994897e-01 8.68419588e-01 -8.88336599e-01
-6.55191720e-01 -2.13707849e-01 -4.04667221e-02 3.32035035e-01
-4.10291672e-01 5.53396605e-02 -7.10374534e-01 -1.16477323e+00
4.67645258e-01 -5.36752582e-01 -2.88405001e-01 -7.71080181e-02
-9.46789160e-02 2.03903019e-01 8.35182786e-01 -1.20218396e+00
1.33141005e+00 -2.70319796e+00 -2.84195356e-02 -1.68415099e-01
2.47810576e-02 4.12018389e-01 -3.19699645e-01 -1.95763797e-01
-2.68008679e-01 -1.72754973e-01 -7.05050379e-02 -7.55160689e-01
1.05595432e-01 -1.49931982e-01 -4.35140312e-01 3.00473154e-01
2.22700417e-01 6.24374926e-01 -8.01485062e-01 -1.28338948e-01
4.70336527e-01 1.27685142e+00 -7.05013514e-01 5.87842427e-02
2.75880754e-01 4.25969660e-01 6.09120168e-02 4.16414142e-01
6.28267527e-01 3.64038527e-01 -3.14197600e-01 -3.28818083e-01
-1.44970506e-01 8.39343011e-01 -1.19914472e+00 1.97441649e+00
-9.13098514e-01 8.35093915e-01 2.98939496e-01 -4.47874457e-01
8.13711286e-01 8.30117106e-01 5.75248837e-01 -1.06656277e+00
3.00760061e-01 2.52189398e-01 4.91389818e-03 -4.28105175e-01
7.15642333e-01 -1.84760883e-01 2.10907206e-01 3.04912150e-01
3.87159526e-01 -1.07531391e-01 -2.88684309e-01 -2.06258163e-01
9.42166805e-01 -1.78460274e-02 -7.32913464e-02 6.68612778e-01
3.50952327e-01 -9.91707444e-01 5.12104034e-01 4.95747030e-01
-4.94421333e-01 8.52637410e-01 7.40923788e-05 1.05788954e-01
-6.44950926e-01 -1.08182120e+00 -9.25409943e-02 1.09392297e+00
8.99529830e-03 -2.76721537e-01 -4.31006819e-01 -4.71719146e-01
-1.93974793e-01 5.95427394e-01 -4.44802463e-01 -5.39927244e-01
-4.79825288e-01 -3.11651200e-01 8.23004782e-01 7.90056586e-01
6.98738456e-01 -9.98994231e-01 -3.10899317e-01 5.38473666e-01
-4.36958730e-01 -1.27175939e+00 -9.58483994e-01 2.57544279e-01
-5.94365537e-01 -3.95191044e-01 -8.88320088e-01 -6.59760118e-01
4.82008159e-02 4.82651949e-01 5.47166049e-01 -2.24332720e-01
-1.37413498e-02 4.41956311e-01 -4.24959689e-01 -5.51069856e-01
4.30381186e-02 -7.10077658e-02 3.20573561e-02 -9.07222405e-02
-6.60399860e-03 -8.34040165e-01 -5.36408365e-01 1.79574952e-01
-6.30771935e-01 -1.80696577e-01 4.62979138e-01 9.12812233e-01
5.05198359e-01 2.22389162e-01 1.03822267e+00 1.06509747e-02
8.85468543e-01 -1.02574758e-01 -2.03590125e-01 -2.38090888e-01
-1.93618551e-01 -1.80336103e-01 -8.30471423e-03 -6.97019577e-01
-1.35746515e+00 1.17531242e-02 -8.83510709e-01 -5.93866110e-01
1.60476670e-01 7.11147904e-01 -4.30269003e-01 9.58900079e-02
4.20634270e-01 2.51768947e-01 -1.29949346e-01 -6.79540098e-01
3.36871415e-01 9.88705575e-01 9.03528929e-01 -1.11625912e-02
3.70418698e-01 3.74783188e-01 -4.63224411e-01 -9.59933758e-01
-6.75766051e-01 -6.70766115e-01 1.98630635e-02 -3.48820180e-01
5.93164027e-01 -1.15958548e+00 -5.92304826e-01 5.61670303e-01
-1.23087609e+00 -2.30167583e-01 -4.77302164e-01 1.05890048e+00
-4.44228739e-01 2.03000709e-01 -8.25596690e-01 -1.09247732e+00
-4.95937496e-01 -1.15581954e+00 1.04531515e+00 3.84460926e-01
2.60548331e-02 -4.60219264e-01 -7.20953345e-02 2.19579756e-01
6.08570278e-01 -3.50730360e-01 2.48797700e-01 -2.54674166e-01
-2.22620800e-01 -2.80633360e-01 1.34878010e-02 5.55208623e-01
2.09324971e-01 -2.95181215e-01 -1.42582750e+00 -7.44791748e-03
3.15336436e-01 1.57401532e-01 1.06403589e+00 7.68974662e-01
1.13501859e+00 7.52216717e-03 1.03478730e-01 7.78443336e-01
5.48649371e-01 5.50365269e-01 9.77851570e-01 2.67676953e-02
3.03059906e-01 2.24993557e-01 5.12048304e-01 4.18657362e-01
2.76636064e-01 7.12209702e-01 2.06165627e-01 -3.75864625e-01
-7.08359778e-01 -3.52888942e-01 5.19375324e-01 1.09166598e+00
2.44127318e-01 -5.95280379e-02 -4.27208304e-01 7.21561849e-01
-1.57109988e+00 -9.07754481e-01 2.96719551e-01 1.91768217e+00
9.79670882e-01 2.70410329e-01 -6.98158816e-02 5.75469851e-01
6.87700212e-01 2.34371677e-01 -5.23823619e-01 -9.72844362e-02
-2.33300179e-01 5.44511318e-01 7.88153857e-02 6.77520096e-01
-1.19674993e+00 6.41097367e-01 6.47679090e+00 7.68428981e-01
-1.51351559e+00 2.94302762e-01 2.20939428e-01 -4.75752264e-01
-1.51469797e-01 -7.50390589e-01 -3.66127402e-01 2.27107376e-01
9.59523857e-01 -6.01693094e-02 4.19682115e-01 5.59184730e-01
3.89253885e-01 3.06050360e-01 -6.40893877e-01 1.11361158e+00
5.02917580e-02 -9.01711643e-01 -5.01630843e-01 -5.40098958e-02
4.63544577e-01 9.05951560e-02 5.26990592e-01 6.09598279e-01
-1.23435352e-02 -8.24638844e-01 9.00906861e-01 4.40220714e-01
9.21718955e-01 -9.07547116e-01 5.12033463e-01 -5.96988164e-02
-1.47699535e+00 -1.71840623e-01 9.33273062e-02 3.04253288e-02
5.57901144e-01 5.96774399e-01 -6.88038290e-01 3.99275541e-01
8.26106071e-01 6.01001740e-01 5.13783433e-02 1.32985175e+00
-4.53891098e-01 4.93676424e-01 -2.56824732e-01 3.27768713e-01
-9.92198009e-03 3.30848038e-01 9.47229505e-01 1.07782960e+00
4.69652116e-01 3.40081871e-01 -3.10946137e-01 4.03231263e-01
-2.00734735e-01 -2.84779012e-01 -3.54233414e-01 -1.26750499e-01
3.62640887e-01 9.05258179e-01 -2.76750714e-01 8.45603719e-02
-3.76394719e-01 7.74014831e-01 -2.50205547e-01 4.76811290e-01
-9.37831819e-01 -7.73061275e-01 9.71716702e-01 -2.76122063e-01
5.02900898e-01 -4.63402539e-01 -2.56856233e-01 -1.00048387e+00
7.03906640e-02 -9.02779222e-01 1.43423826e-01 -1.07484770e+00
-8.92614126e-01 8.35705698e-01 -6.82151496e-01 -1.31040156e+00
-1.45651847e-01 -5.22460639e-01 -3.66483212e-01 9.15537596e-01
-1.64168859e+00 -9.70325470e-01 -9.85428765e-02 5.91433167e-01
7.02447236e-01 -7.12265000e-02 6.34263575e-01 1.02015924e+00
-3.66864294e-01 8.89251411e-01 -3.57383221e-01 1.73356459e-01
5.83547413e-01 -1.03435802e+00 7.70636678e-01 9.93375182e-01
2.89520025e-01 4.88109291e-01 6.78273380e-01 -4.82099503e-01
-1.11087430e+00 -9.94019866e-01 6.10571921e-01 1.46005407e-01
4.71313953e-01 -2.01219112e-01 -9.59036231e-01 3.83276314e-01
1.58230752e-01 1.73410013e-01 7.42324412e-01 1.12908989e-01
-2.15967715e-01 3.44770551e-02 -8.78913999e-01 6.80900335e-01
1.09029996e+00 -1.17137611e+00 -4.82369930e-01 -1.95298031e-01
1.45862472e+00 -9.63450015e-01 -7.19337344e-01 6.00593626e-01
7.44782209e-01 -6.42943501e-01 1.21815813e+00 -4.82183456e-01
2.41348371e-01 -3.85019302e-01 -3.48739892e-01 -1.60760665e+00
5.96647598e-02 -9.33145702e-01 -6.10755563e-01 1.17872095e+00
4.45425570e-01 -4.69369441e-01 4.23046023e-01 -5.01308008e-04
-7.10057139e-01 -7.16986716e-01 -1.08995366e+00 -5.83383858e-01
-5.51755786e-01 -9.88916397e-01 6.22204185e-01 6.25645757e-01
5.53736128e-02 2.38756552e-01 -5.57941079e-01 4.42965508e-01
1.73838645e-01 -4.67018992e-01 3.95610213e-01 -7.36745834e-01
-4.45318669e-01 -4.37716395e-01 -3.14160138e-01 -1.31037629e+00
7.00302720e-02 -3.72390062e-01 2.93264389e-01 -1.51704919e+00
-5.38292408e-01 -8.67856108e-03 -7.12032855e-01 2.56523281e-01
-3.32565755e-01 1.59012869e-01 1.86768085e-01 -5.26953936e-01
-2.35469460e-01 1.15204692e+00 1.34889162e+00 -1.99331939e-01
-6.80472434e-01 2.43649989e-01 -6.11937284e-01 5.83622277e-01
5.92700064e-01 -2.00242043e-01 -4.37077105e-01 -5.40183544e-01
9.08222422e-02 4.89572734e-01 3.28071743e-01 -1.04284143e+00
3.47876310e-01 4.11335617e-01 1.88803107e-01 -9.70097959e-01
1.00345862e+00 -7.98427582e-01 -9.31019336e-02 1.81900591e-01
-4.08697963e-01 -2.21324220e-01 6.77228153e-01 5.44135153e-01
-5.72187960e-01 2.17801496e-01 6.48039043e-01 2.60824472e-01
-4.25134361e-01 3.37658316e-01 -2.61124045e-01 -3.30596328e-01
3.65347952e-01 3.43904570e-02 -2.35067278e-01 -7.53105462e-01
-1.02717423e+00 8.13686848e-02 -3.83653045e-01 6.20056272e-01
9.79683518e-01 -1.54483032e+00 -6.24974608e-01 2.90747136e-01
-3.11429441e-01 -2.76736796e-01 7.71821260e-01 9.71736193e-01
9.57534090e-02 1.41661540e-01 -5.82796112e-02 -3.00060421e-01
-1.05754566e+00 2.26369917e-01 8.89226198e-01 -3.79327498e-02
-7.81630278e-01 1.14946032e+00 -8.80550370e-02 -4.44702238e-01
8.79457593e-01 -4.07251894e-01 -4.22334701e-01 1.06913246e-01
8.53201210e-01 3.68380010e-01 3.04910958e-01 -5.15254438e-01
-1.14776194e-01 1.08944014e-01 4.18892279e-02 -7.89960802e-01
1.42865944e+00 -3.75095278e-01 4.16128546e-01 3.48026991e-01
1.27282774e+00 2.01502070e-01 -1.26742530e+00 -4.42604065e-01
-4.35621351e-01 -3.83963346e-01 7.89883912e-01 -9.26137269e-01
-1.44607043e+00 1.08648121e+00 9.18680191e-01 4.10821661e-02
1.52467728e+00 -2.65998304e-01 1.20674312e+00 5.76313101e-02
-1.89149715e-02 -1.08497751e+00 4.61035669e-01 7.91520596e-01
1.22197556e+00 -9.49818075e-01 -3.90693754e-01 -3.46870512e-01
-4.86821711e-01 8.09053421e-01 5.43768346e-01 2.67180741e-01
8.82891297e-01 6.23743415e-01 3.47487688e-01 4.08864953e-02
-6.26272678e-01 -5.40274918e-01 4.21674043e-01 7.36019611e-01
5.36813319e-01 -2.41529509e-01 8.15049857e-02 1.16194439e+00
-5.08433819e-01 -9.14329514e-02 3.48330855e-01 7.70148277e-01
-2.53350616e-01 -7.43554473e-01 -4.60707873e-01 3.30961458e-02
-6.58340216e-01 -2.87754476e-01 -9.79914516e-02 4.63414878e-01
-1.12496912e-01 1.18731749e+00 -1.98573079e-02 -6.99534178e-01
7.78051615e-01 8.08854848e-02 3.34878802e-01 -1.66499913e-01
-7.48253822e-01 7.31907845e-01 1.37591854e-01 -4.22833622e-01
-1.95655823e-01 -5.41766882e-01 -1.36368620e+00 1.91689134e-01
-6.44413292e-01 -6.16224520e-02 1.00820518e+00 7.10241675e-01
5.01871467e-01 1.47252309e+00 7.21265674e-01 -9.39492524e-01
-3.72127980e-01 -1.17757809e+00 -5.58193624e-01 -1.71539813e-01
8.86595964e-01 -6.39825404e-01 -3.17493111e-01 -7.62003586e-02] | [14.941435813903809, 5.935564041137695] |
e3924672-5318-417c-ab83-474d3b9b16de | conan-a-complementary-neighboring-based | null | null | https://aclanthology.org/2020.coling-main.177 | https://aclanthology.org/2020.coling-main.177.pdf | CoNAN: A Complementary Neighboring-based Attention Network for Referring Expression Generation | Daily scenes are complex in the real world due to occlusion, undesired lighting conditions, etc. Although humans handle those complicated environments well, they evoke challenges for machine learning systems to identify and describe the target without ambiguity. Most previous research focuses on mining discriminating features within the same category for the target object. One the other hand, as the scene becomes more complicated, human frequently uses the neighbor objects as complementary information to describe the target one. Motivated by that, we propose a novel Complementary Neighboring-based Attention Network (CoNAN) that explicitly utilizes the visual differences between the target object and its highly-related neighbors. These highly-related neighbors are determined by an attentional ranking module, as complementary features, highlighting the discriminating aspects for the target object. The speaker module then takes the visual difference features as an additional input to generate the expression. Our qualitative and quantitative results on the dataset RefCOCO, RefCOCO+, and RefCOCOg demonstrate that our generated expressions outperform other state-of-the-art models by a clear margin. | ['Jialin Wu', 'Hanbin Ko', 'Jungjun Kim'] | 2020-12-01 | null | null | null | coling-2020-8 | ['referring-expression-generation'] | ['computer-vision'] | [-4.84971553e-02 -2.07162142e-01 -2.66239375e-01 -7.71382272e-01
-3.50366950e-01 -5.27691066e-01 6.16350889e-01 2.23308980e-01
-3.27581257e-01 4.89299476e-01 3.76040816e-01 4.46834624e-01
-8.40325356e-02 -5.06402910e-01 -3.57468516e-01 -7.70659924e-01
1.95693791e-01 2.44050696e-01 -3.58561128e-02 -3.87604833e-01
3.63358676e-01 5.48940659e-01 -2.01212883e+00 5.33372879e-01
7.25122988e-01 1.24534452e+00 1.79953933e-01 3.35366800e-02
-2.88222522e-01 6.27087533e-01 -6.28409028e-01 -1.87254623e-01
1.85894072e-01 -3.78506750e-01 -4.97126460e-01 2.21784666e-01
5.79780996e-01 1.09924898e-01 3.81755196e-02 1.14734483e+00
3.79479975e-01 3.51959020e-01 8.71222734e-01 -1.34691834e+00
-8.71141970e-01 2.81153172e-01 -6.71546876e-01 1.12635478e-01
4.22401071e-01 1.07792422e-01 1.10756028e+00 -1.02300560e+00
6.92267418e-01 1.25973320e+00 -6.13306975e-03 5.30912697e-01
-1.10971034e+00 -7.87969172e-01 6.66849732e-01 5.57877541e-01
-1.57985067e+00 -2.97462165e-01 1.02963436e+00 -4.68147427e-01
5.90232849e-01 3.92519474e-01 7.20643878e-01 1.16044331e+00
-8.77453685e-02 6.82614744e-01 1.19837332e+00 -2.23224655e-01
2.25803256e-01 5.31850934e-01 2.70271629e-01 4.66066033e-01
-1.78661793e-02 8.29404965e-02 -5.98164618e-01 1.24905430e-01
1.82723880e-01 3.03108186e-01 -4.61884946e-01 -4.29079831e-01
-1.12712991e+00 6.83373451e-01 9.06645358e-01 6.52581751e-01
-5.43031216e-01 -8.26270208e-02 1.70825124e-01 -3.48353237e-02
4.36845750e-01 5.97133636e-01 -2.06196457e-01 6.04649633e-02
-6.50798798e-01 5.56437820e-02 4.80652660e-01 1.04328394e+00
1.05769193e+00 -3.62539351e-01 -5.45041382e-01 8.25392962e-01
2.08359599e-01 2.83340096e-01 3.76551986e-01 -8.46224606e-01
1.77042902e-01 8.22050512e-01 2.02448711e-01 -1.36289012e+00
-2.15197951e-01 -6.48571849e-01 -8.24002504e-01 2.86588103e-01
3.07955663e-03 2.37437770e-01 -7.56101608e-01 1.78716290e+00
4.19092536e-01 -3.52578551e-01 -9.80215222e-02 1.42817640e+00
1.00536084e+00 6.51120603e-01 3.49180281e-01 -1.15512140e-01
1.35867524e+00 -9.98416126e-01 -6.78825259e-01 -3.47334057e-01
3.75543326e-01 -6.63317323e-01 1.17474174e+00 2.42307588e-01
-4.39820021e-01 -7.00641096e-01 -9.57654595e-01 -4.43832986e-02
-5.09643614e-01 3.66432548e-01 7.69661367e-01 1.03441902e-01
-8.33527863e-01 2.18557626e-01 -1.12904213e-01 -4.91642416e-01
4.42128211e-01 9.59029570e-02 -5.34849644e-01 -1.31523833e-01
-1.09616506e+00 8.90306711e-01 2.63877749e-01 2.69661099e-01
-6.45983219e-01 -4.37810153e-01 -9.25693810e-01 1.38539806e-01
2.94970930e-01 -3.97726387e-01 8.40736270e-01 -1.60212016e+00
-1.30224264e+00 1.14708543e+00 -2.91462898e-01 5.80129921e-02
2.57972777e-01 -2.74386913e-01 -5.28478563e-01 -9.20325052e-03
3.71398747e-01 9.82935727e-01 7.82974601e-01 -1.55131853e+00
-6.24784946e-01 -4.13528442e-01 2.41075635e-01 4.13822323e-01
-2.47848794e-01 1.24039582e-03 -4.81827497e-01 -5.22292495e-01
2.09162999e-02 -7.45691538e-01 -1.14209995e-01 2.51065761e-01
-5.37608087e-01 -5.95186770e-01 9.08809364e-01 -4.04537693e-02
1.08253837e+00 -2.63341498e+00 1.09824829e-01 2.06013009e-01
3.32582831e-01 9.90136862e-02 -3.44761550e-01 2.42084950e-01
-2.06274390e-01 7.75993317e-02 7.35106617e-02 -8.55478644e-02
-6.17225170e-02 -8.46248027e-03 -1.39322340e-01 2.97112972e-01
3.86225611e-01 6.07142448e-01 -1.04017580e+00 -5.61149299e-01
1.80343553e-01 4.60565656e-01 -2.78517336e-01 3.80861193e-01
-8.79802257e-02 5.00811875e-01 -5.22665977e-01 6.69567227e-01
5.95120311e-01 -2.25250974e-01 1.56787783e-02 -5.91520309e-01
-1.95576966e-01 -2.15586320e-01 -8.72437418e-01 1.32231498e+00
-6.17836416e-01 7.36501873e-01 7.08528683e-02 -9.83172536e-01
1.25573874e+00 -1.75535694e-01 1.76706642e-01 -9.02791560e-01
3.10875118e-01 2.32391864e-01 1.11646004e-01 -7.87865937e-01
3.27814192e-01 -1.12668656e-01 -2.22791046e-01 9.78904404e-03
-2.04754576e-01 1.51707545e-01 3.00359905e-01 1.57595545e-01
6.50003612e-01 4.70237695e-02 5.71719646e-01 -3.20912510e-01
6.37820542e-01 -8.10531899e-02 7.37850666e-01 4.68003213e-01
-4.43552494e-01 6.43941224e-01 2.18384266e-01 -6.29408836e-01
-5.81150711e-01 -1.04600990e+00 4.86779325e-02 1.30538547e+00
6.27897859e-01 -2.56825835e-01 -3.86033118e-01 -8.00064147e-01
-3.59239392e-02 7.43735135e-01 -9.33709502e-01 -1.19681522e-01
-2.11730510e-01 -1.76519334e-01 -6.36449605e-02 5.55256426e-01
7.04934776e-01 -1.27956700e+00 -8.11337411e-01 -1.75875623e-03
-3.05895120e-01 -9.98966396e-01 -4.80416477e-01 1.59461394e-01
-1.59064725e-01 -1.02265501e+00 -7.18734264e-01 -9.05821025e-01
8.76583457e-01 3.80669773e-01 1.15653884e+00 1.44764051e-01
-4.08939302e-01 2.16089442e-01 -4.12003249e-01 -5.13830066e-01
2.00993687e-01 -3.46019626e-01 -6.64694160e-02 4.77065235e-01
7.60246217e-01 -3.26293796e-01 -5.87149262e-01 3.30615550e-01
-5.81419587e-01 2.84860522e-01 6.75692499e-01 7.31192648e-01
7.09634423e-01 -4.33417596e-02 2.97024250e-01 -6.55721009e-01
3.97543639e-01 -5.26006699e-01 -3.70746344e-01 2.70637482e-01
-1.43161774e-01 2.95027960e-02 7.34499097e-01 -5.93683064e-01
-1.03893542e+00 2.33751774e-01 4.81876731e-01 -5.24710774e-01
-4.12813604e-01 3.86376113e-01 -4.50628936e-01 2.97732383e-01
5.38126767e-01 7.11159334e-02 -2.78019845e-01 -1.59419075e-01
2.74189830e-01 6.84769571e-01 3.25416565e-01 -2.94368982e-01
7.55928278e-01 5.26255429e-01 -3.74199003e-02 -8.83628249e-01
-1.35134435e+00 -6.09612226e-01 -6.71234012e-01 -4.66562033e-01
9.74983096e-01 -7.88259506e-01 -7.23961592e-01 2.05938727e-01
-1.41942298e+00 8.00610706e-03 -5.60839325e-02 3.73018503e-01
-1.14681721e-01 1.99451037e-02 3.15709524e-02 -8.88772547e-01
-7.63873681e-02 -9.42599416e-01 1.20514369e+00 5.26521385e-01
-4.94370997e-01 -4.51990634e-01 -2.65308261e-01 3.34929764e-01
4.26610619e-01 2.91559339e-01 1.05890584e+00 -6.01420999e-01
-5.11202157e-01 -3.19328420e-02 -5.76159358e-01 2.02501088e-01
3.61501396e-01 6.16771095e-02 -9.43573594e-01 7.44622722e-02
-1.97926566e-01 -2.77414978e-01 7.62953758e-01 8.02948847e-02
1.40425539e+00 -2.68547177e-01 -4.31991041e-01 4.37590092e-01
1.17863083e+00 3.32658887e-01 3.58130157e-01 1.99575484e-01
6.19834006e-01 1.12164068e+00 1.03701365e+00 3.27734649e-01
2.25104213e-01 7.37570465e-01 5.14809787e-01 -2.97837377e-01
9.57112089e-02 8.23120698e-02 2.33121663e-01 4.54389840e-01
-1.39488941e-02 -2.98200190e-01 -6.76258087e-01 6.58227503e-01
-1.74289823e+00 -9.71158206e-01 6.46187738e-02 1.96658576e+00
6.63868487e-01 -1.22959822e-01 -1.83608726e-01 -1.29151478e-01
1.08216333e+00 2.17278227e-01 -4.58179832e-01 -2.74676591e-01
-3.06482822e-01 -1.68277338e-01 -6.67233840e-02 4.21244502e-02
-1.22692263e+00 8.31908882e-01 5.27378225e+00 7.37848341e-01
-1.40626550e+00 -2.20834881e-01 9.94588196e-01 -2.36006275e-01
-2.27668405e-01 -1.39095485e-01 -6.85309649e-01 3.77408624e-01
2.61509299e-01 -1.44355536e-01 2.81124592e-01 1.06851912e+00
2.47581810e-01 -2.06351370e-01 -1.33988082e+00 1.31633019e+00
3.92460674e-01 -9.26612437e-01 2.17073068e-01 -1.00844562e-01
5.98926306e-01 -4.29553211e-01 2.09169254e-01 3.27112764e-01
-3.45080681e-02 -1.15139425e+00 7.73304284e-01 7.50371695e-01
6.85272634e-01 -7.24063218e-01 9.19929564e-01 1.30016327e-01
-1.17680335e+00 -1.01378344e-01 -2.94887960e-01 -1.38898432e-01
-1.92096338e-01 6.09911203e-01 -5.72936177e-01 3.28791410e-01
1.03936219e+00 7.75055289e-01 -7.41585016e-01 8.82910430e-01
-4.51223254e-01 2.02593505e-01 1.60127301e-02 -4.86432463e-01
3.61961812e-01 -1.76742971e-01 4.81631815e-01 1.09783351e+00
3.89037251e-01 2.50915408e-01 3.09245080e-01 1.07315195e+00
-3.77630182e-02 3.06964755e-01 -6.29959047e-01 -3.83129492e-02
3.79220545e-01 1.61026633e+00 -5.69223821e-01 -3.43592048e-01
-2.81925261e-01 1.11186326e+00 5.02714932e-01 4.46791351e-01
-7.73473084e-01 -6.01684511e-01 6.73926473e-01 9.80667025e-02
2.58375138e-01 1.83528215e-01 9.56157688e-03 -9.50950563e-01
2.03085050e-01 -7.85900772e-01 2.89283842e-01 -1.00911605e+00
-1.66727960e+00 8.79348993e-01 -9.76536125e-02 -1.49012244e+00
4.52628285e-02 -5.96790791e-01 -6.86830044e-01 8.57769728e-01
-1.45256960e+00 -1.21695542e+00 -8.91187072e-01 4.25682217e-01
5.22389412e-01 -2.88852379e-02 7.93624520e-01 1.45280704e-01
-3.94199103e-01 4.00242746e-01 -2.72446126e-01 2.13315696e-01
8.07437122e-01 -1.02057636e+00 -3.50255966e-01 5.33548415e-01
1.95494130e-01 6.86747015e-01 6.58723354e-01 -3.67546201e-01
-9.03158903e-01 -1.07075799e+00 1.09626079e+00 -1.05408423e-01
4.03928846e-01 -5.12660623e-01 -7.86128879e-01 2.34840140e-01
2.17112869e-01 4.01916653e-01 7.78545201e-01 1.43837214e-01
-7.09161103e-01 -2.92749941e-01 -9.66581941e-01 7.15734303e-01
1.14839292e+00 -4.11519915e-01 -4.71972942e-01 -1.07353581e-02
5.27436793e-01 1.20208478e-02 -3.50105464e-01 3.81237984e-01
4.95201141e-01 -1.17535770e+00 5.61054051e-01 -4.72315282e-01
5.58176458e-01 -6.23522580e-01 -3.15893084e-01 -1.43018162e+00
-5.74089706e-01 -2.08053991e-01 4.37315345e-01 1.35507476e+00
3.46771598e-01 -3.56126696e-01 5.47131240e-01 3.99342477e-01
-5.52316718e-02 -7.33421922e-01 -9.14043546e-01 -6.35824025e-01
-4.49144483e-01 -1.91761866e-01 6.56187236e-01 1.02164805e+00
-1.13916129e-01 5.15737474e-01 -2.51318038e-01 -1.18191857e-02
3.39308053e-01 7.38506675e-01 6.30079091e-01 -1.39432788e+00
-2.76534297e-02 -5.55508018e-01 -6.79782033e-01 -9.56669807e-01
2.87332207e-01 -8.01937759e-01 2.61167228e-01 -1.34566128e+00
4.15206522e-01 -3.97456169e-01 -5.43955088e-01 4.49712276e-01
-1.35843322e-01 3.68935883e-01 3.43058437e-01 1.72169637e-02
-9.71372902e-01 8.25532377e-01 1.25150037e+00 -5.89293957e-01
-1.29454434e-01 -2.18368798e-01 -9.76955831e-01 8.83192897e-01
5.86232126e-01 -2.17575565e-01 -1.99936852e-01 -3.60125363e-01
1.64165441e-02 -4.28613245e-01 4.43874776e-01 -1.11603618e+00
6.61339685e-02 -4.89585042e-01 7.07405150e-01 -5.81140935e-01
4.46551204e-01 -1.05622113e+00 -1.40923038e-01 2.27918759e-01
-5.81294656e-01 -5.53375892e-02 -3.53315193e-03 3.80488306e-01
-5.11175454e-01 -2.97571681e-02 7.83783972e-01 -1.35052353e-01
-1.15100276e+00 2.59384602e-01 -3.53181809e-01 -7.32352957e-04
1.15048838e+00 -2.25545049e-01 -3.59905124e-01 -6.98184788e-01
-7.08138108e-01 2.25939333e-01 3.05848837e-01 7.61897564e-01
7.27995098e-01 -1.57405984e+00 -6.33008480e-01 1.33086860e-01
7.43404031e-01 -2.01529607e-01 3.71228337e-01 6.71951771e-01
5.10712042e-02 2.51943529e-01 -3.62512887e-01 -6.15542829e-01
-1.55726135e+00 7.30749488e-01 2.95130551e-01 7.12498501e-02
-1.57470942e-01 9.09529924e-01 8.47040832e-01 -1.22680545e-01
2.14429274e-01 -5.52792698e-02 -5.76366544e-01 2.95035213e-01
4.55179125e-01 1.34385884e-01 -3.53637874e-01 -1.05127418e+00
-6.11533284e-01 7.10962653e-01 -8.92711356e-02 1.63014278e-01
1.20698619e+00 -1.24253713e-01 -2.26166934e-01 5.58314860e-01
1.31366253e+00 -3.25792246e-02 -9.42389190e-01 -3.33635747e-01
-3.31880972e-02 -6.15110993e-01 -1.80844814e-01 -9.30587828e-01
-1.15594161e+00 9.86058235e-01 6.96199834e-01 -5.23001961e-02
1.33295631e+00 2.20837653e-01 1.40832588e-01 4.69653845e-01
3.78648013e-01 -1.14219761e+00 4.20173109e-01 4.79679912e-01
1.28922784e+00 -1.35670364e+00 -1.90134063e-01 -6.58966660e-01
-8.49784076e-01 9.82018411e-01 1.14328909e+00 5.80694564e-02
5.23607850e-01 -1.93428725e-01 4.79053140e-01 -2.92229086e-01
-6.81071877e-01 -3.66491258e-01 5.16690493e-01 7.30346978e-01
6.31171644e-01 5.19534275e-02 -1.59746706e-01 6.18931472e-01
-1.78795025e-01 -4.66038227e-01 1.19356573e-01 7.30686545e-01
-4.29341108e-01 -7.98542321e-01 -3.87305409e-01 3.44449162e-01
-2.33394101e-01 -4.52088639e-02 -7.33017325e-01 7.16083467e-01
4.64623034e-01 1.05118620e+00 4.00398940e-01 -4.24449891e-01
3.20886403e-01 -3.24784040e-01 1.06375724e-01 -6.96654975e-01
-6.36253774e-01 -5.09641729e-02 6.55758157e-02 -7.00730443e-01
-6.68120742e-01 -4.23095226e-01 -1.30423450e+00 2.01102018e-01
-1.53754309e-01 2.47745201e-01 4.16812032e-01 7.85561919e-01
5.70529163e-01 4.44157541e-01 7.88127244e-01 -9.48101878e-01
-1.34568512e-01 -8.79847229e-01 -5.10527968e-01 9.77740943e-01
2.19525337e-01 -8.91118228e-01 -4.03052092e-01 -1.68828532e-01] | [10.039467811584473, 1.8852379322052002] |
6497fd96-27e4-4d58-9ae8-9d2e31673df9 | semi-supervised-learning-from-street-view | 2307.02574 | null | https://arxiv.org/abs/2307.02574v1 | https://arxiv.org/pdf/2307.02574v1.pdf | Semi-supervised Learning from Street-View Images and OpenStreetMap for Automatic Building Height Estimation | Accurate building height estimation is key to the automatic derivation of 3D city models from emerging big geospatial data, including Volunteered Geographical Information (VGI). However, an automatic solution for large-scale building height estimation based on low-cost VGI data is currently missing. The fast development of VGI data platforms, especially OpenStreetMap (OSM) and crowdsourced street-view images (SVI), offers a stimulating opportunity to fill this research gap. In this work, we propose a semi-supervised learning (SSL) method of automatically estimating building height from Mapillary SVI and OSM data to generate low-cost and open-source 3D city modeling in LoD1. The proposed method consists of three parts: first, we propose an SSL schema with the option of setting a different ratio of "pseudo label" during the supervised regression; second, we extract multi-level morphometric features from OSM data (i.e., buildings and streets) for the purposed of inferring building height; last, we design a building floor estimation workflow with a pre-trained facade object detection network to generate "pseudo label" from SVI and assign it to the corresponding OSM building footprint. In a case study, we validate the proposed SSL method in the city of Heidelberg, Germany and evaluate the model performance against the reference data of building heights. Based on three different regression models, namely Random Forest (RF), Support Vector Machine (SVM), and Convolutional Neural Network (CNN), the SSL method leads to a clear performance boosting in estimating building heights with a Mean Absolute Error (MAE) around 2.1 meters, which is competitive to state-of-the-art approaches. The preliminary result is promising and motivates our future work in scaling up the proposed method based on low-cost VGI data, with possibilities in even regions and areas with diverse data quality and availability. | ['Martin Werner', 'Alexander Zipf', 'Hongchao Fan', 'Gefei Kong', 'Gabriel Dax', 'Zhendong Yuan', 'Hao Li'] | 2023-07-05 | null | null | null | null | ['object-detection', 'pseudo-label'] | ['computer-vision', 'miscellaneous'] | [ 1.54088512e-01 2.67595649e-01 3.92786235e-01 -4.45290476e-01
-8.54769349e-01 -1.10743947e-01 7.06026971e-01 6.04057372e-01
-5.24418831e-01 8.72519255e-01 -2.47559603e-02 -4.07968134e-01
-2.06452891e-01 -1.68664515e+00 -7.37817168e-01 -5.99389553e-01
-1.72609910e-01 9.45555449e-01 3.56474280e-01 -2.52364308e-01
5.59935383e-02 8.08075011e-01 -2.01131177e+00 -1.25790477e-01
1.16086757e+00 8.52125764e-01 5.44112861e-01 5.38327575e-01
-3.58344108e-01 3.49015772e-01 -2.46514872e-01 1.42452680e-02
2.96832889e-01 1.23930201e-01 -7.64384925e-01 3.25211696e-02
5.29631972e-01 -3.78765643e-01 2.63624489e-01 4.83168751e-01
6.01188481e-01 -2.39930481e-01 7.05470264e-01 -1.07542777e+00
-1.14047274e-01 5.19421458e-01 -2.82148719e-01 -1.64441764e-01
6.56718686e-02 2.35237643e-01 7.86990106e-01 -7.63586760e-01
4.58721936e-01 7.93370008e-01 9.17293012e-01 -6.50757998e-02
-1.29847074e+00 -3.93068045e-01 3.40533108e-02 -1.09334871e-01
-1.95336092e+00 -1.14277370e-01 4.84083384e-01 -8.28824878e-01
7.72892475e-01 4.53323781e-01 1.00681078e+00 4.72571462e-01
-2.95772016e-01 4.87737149e-01 1.37310338e+00 -4.59424108e-01
4.38571513e-01 1.27663165e-01 -3.09211284e-01 6.08686686e-01
5.58185339e-01 3.57772633e-02 -7.26833642e-02 4.55615968e-02
5.62057674e-01 -1.17816068e-01 -3.17633376e-02 -3.22906524e-01
-9.23281908e-01 9.25763130e-01 8.01271856e-01 4.54146951e-01
-4.73643124e-01 -1.73097588e-02 -4.59501296e-02 -3.18938613e-01
8.23286712e-01 -3.08937877e-02 -4.24538404e-01 3.27993512e-01
-1.49089444e+00 4.34194982e-01 4.25506443e-01 5.17806828e-01
1.12240899e+00 9.41904187e-02 1.73642084e-01 7.69040465e-01
6.69741094e-01 9.77209032e-01 8.85310397e-02 -4.12641019e-01
7.61712849e-01 9.97315884e-01 5.05316593e-02 -9.06835437e-01
-9.20267701e-01 -3.32684577e-01 -7.36833870e-01 3.24866891e-01
5.44656813e-01 1.78336322e-01 -8.38391542e-01 1.31981957e+00
6.02334142e-01 -7.02619851e-02 -1.25095099e-01 6.89755201e-01
1.13152909e+00 4.12608624e-01 1.90468237e-01 3.19151968e-01
1.11017776e+00 -3.26438278e-01 5.37268594e-02 -2.68746406e-01
8.57070148e-01 -2.44226709e-01 1.02590144e+00 -6.04130793e-03
-6.62544727e-01 -5.39947212e-01 -1.06064951e+00 6.27789274e-02
-9.20166612e-01 4.81099188e-01 5.25245965e-01 8.52942348e-01
-1.01777625e+00 5.67492485e-01 -7.38528669e-01 -6.65228009e-01
4.60689962e-01 2.73424119e-01 -2.99811929e-01 2.06723735e-01
-1.09667301e+00 8.10904622e-01 4.51079607e-01 3.63036484e-01
-5.81564426e-01 -7.87362576e-01 -9.36492741e-01 -7.49963820e-02
-9.79160368e-02 -6.11991167e-01 4.09384280e-01 -2.76867777e-01
-9.87420380e-01 1.34450483e+00 2.38365471e-01 -6.10127926e-01
8.47582281e-01 1.06618159e-01 -3.41394454e-01 -2.07220897e-01
6.31400883e-01 8.52121651e-01 3.63272905e-01 -1.50428116e+00
-7.80969262e-01 -7.33239949e-01 -2.78794885e-01 -2.30416939e-01
-1.30456127e-02 -4.05059516e-01 4.27455157e-02 -1.04685724e-01
4.60632175e-01 -8.94549012e-01 -2.71594346e-01 -2.46015489e-01
-3.45717669e-01 1.03507839e-01 3.77784312e-01 -9.56540346e-01
1.21446574e+00 -1.87018335e+00 -3.89514059e-01 5.50308764e-01
1.29620239e-01 3.51070255e-01 3.23131174e-01 4.34780270e-01
1.10411681e-01 1.87787309e-01 -7.38089263e-01 -1.92599848e-01
-1.82557881e-01 9.55173150e-02 3.86841632e-02 3.38679552e-01
8.94333124e-02 8.37732017e-01 -7.98976660e-01 -6.61806703e-01
9.76196766e-01 6.48372173e-01 -2.17341408e-01 -6.34688139e-02
-9.27302092e-02 6.68330431e-01 -2.51792192e-01 6.75429761e-01
1.14889205e+00 -1.52385114e-02 7.18818307e-02 5.14894687e-02
-7.36707807e-01 5.38549758e-03 -1.46041906e+00 1.36758327e+00
-8.26223969e-01 4.35479850e-01 -1.60712361e-01 -6.66362762e-01
1.46356797e+00 3.45222577e-02 5.30786216e-01 -6.04334712e-01
-7.65925944e-02 5.97539902e-01 -4.60566103e-01 -2.74438053e-01
5.25975108e-01 -1.54978521e-02 2.44954824e-02 -5.92616424e-02
-1.24805257e-01 -4.94515628e-01 8.50838050e-02 -1.60690129e-01
6.70230806e-01 3.72035772e-01 2.49501750e-01 -4.32778835e-01
9.53111291e-01 1.85448498e-01 3.93731631e-02 4.69631344e-01
8.19598883e-03 7.61582732e-01 4.09251563e-02 -8.35185766e-01
-1.19573104e+00 -1.04720736e+00 -5.53193629e-01 5.44870079e-01
-2.96073377e-01 -2.12077022e-01 -9.05825555e-01 -2.07002878e-01
1.88997656e-01 7.37226665e-01 -4.81787235e-01 5.13996422e-01
-4.09345984e-01 -1.10607696e+00 6.13386571e-01 3.45438302e-01
8.64585578e-01 -8.04648578e-01 -8.46256971e-01 2.56852180e-01
-2.43294984e-01 -1.11651242e+00 5.71683228e-01 1.04850002e-01
-9.01254237e-01 -9.46394145e-01 -5.32468498e-01 -2.53589183e-01
4.39616233e-01 5.48693351e-02 1.11547339e+00 1.95914209e-01
-2.72623539e-01 -2.33678818e-02 -2.03442261e-01 -4.55535144e-01
-2.71823108e-01 5.93702734e-01 -3.68206918e-01 7.11599663e-02
2.51997203e-01 -8.52772534e-01 -6.96838558e-01 3.31907749e-01
-6.47681177e-01 4.21598315e-01 3.90837133e-01 8.65853205e-02
7.13816643e-01 -1.92963947e-02 4.14829552e-01 -7.75813818e-01
-6.87979385e-02 -4.83327210e-01 -9.70361590e-01 -5.96679151e-02
-8.19657385e-01 3.20854634e-02 4.16471273e-01 3.53605002e-01
-9.23624158e-01 3.63908470e-01 -7.68682361e-01 3.38566333e-01
-6.92858040e-01 4.29299653e-01 -3.77953589e-01 -8.38909745e-02
8.21762562e-01 5.24044558e-02 -1.94283023e-01 -5.91190040e-01
3.27266335e-01 9.13170099e-01 3.45692426e-01 -5.25465965e-01
1.02547860e+00 8.56054187e-01 3.27331543e-01 -1.13255298e+00
-6.69649184e-01 -4.94545758e-01 -1.01763546e+00 -4.41807508e-01
9.19581532e-01 -1.06925917e+00 -4.38874692e-01 5.10931909e-01
-7.45653987e-01 -4.92702186e-01 -3.46260995e-01 3.02827090e-01
-4.77755636e-01 1.54190391e-01 -9.56384093e-02 -1.10441828e+00
-3.76335382e-01 -9.78232324e-01 1.40799797e+00 1.80032924e-01
1.36860847e-01 -7.24397004e-01 3.88911009e-01 6.46010220e-01
6.00081086e-01 8.55064869e-01 5.81215322e-01 -2.88254917e-01
-8.45848680e-01 -1.84559494e-01 -4.24553931e-01 1.67290986e-01
-5.69058396e-02 2.37088129e-01 -1.25542784e+00 2.43694067e-01
-6.51255190e-01 1.07614145e-01 9.00221825e-01 4.94633228e-01
6.92565441e-01 1.17931567e-01 -2.42993906e-01 5.91297686e-01
1.90005577e+00 -3.00946176e-01 8.79314423e-01 9.49039698e-01
7.50403404e-01 7.52419233e-01 5.19575834e-01 6.50352955e-01
8.39618742e-01 7.26860523e-01 8.28389049e-01 -2.98580527e-01
-2.27126077e-01 -5.37515044e-01 -1.98900267e-01 2.61191905e-01
-6.59957707e-01 2.47742444e-01 -1.41227722e+00 6.91351831e-01
-1.69700074e+00 -8.39604020e-01 -9.54842627e-01 2.69166780e+00
1.84400916e-01 1.60734519e-01 2.11597711e-01 4.60763961e-01
5.30099809e-01 2.37235248e-01 -2.14244928e-02 -9.75106284e-02
-3.13160717e-01 3.39278936e-01 9.46554065e-01 6.44971430e-01
-1.29324698e+00 9.19418275e-01 4.84837103e+00 6.19013846e-01
-1.18151700e+00 2.64104635e-01 4.33198422e-01 1.02497108e-01
-1.86382696e-01 -2.51362706e-03 -9.80473638e-01 4.16097790e-01
1.31793606e+00 3.92382592e-01 1.98265180e-01 1.08179665e+00
5.67693293e-01 -3.42449307e-01 -2.74768233e-01 6.51356459e-01
-3.24875891e-01 -1.45293510e+00 -3.03096294e-01 5.04838645e-01
6.01135433e-01 4.93460208e-01 -6.06333852e-01 1.14860810e-01
1.99058041e-01 -7.47201324e-01 8.46179545e-01 5.60953677e-01
9.06657994e-01 -6.89505100e-01 7.92261541e-01 4.82724637e-01
-1.74316907e+00 5.84137207e-03 -2.25245878e-01 -9.37962905e-02
3.49322349e-01 8.88860285e-01 -9.57889557e-01 8.04825902e-01
7.58791864e-01 2.66281992e-01 -1.01951206e+00 1.09994233e+00
-4.28236634e-01 5.50750375e-01 -7.65006721e-01 1.70206308e-01
3.18782926e-01 -4.35815305e-01 5.80881089e-02 1.14203525e+00
4.69975322e-01 -3.01823288e-01 1.96266919e-02 9.66568708e-01
3.43106061e-01 5.43753564e-01 -7.70825446e-01 5.17877638e-01
3.11243415e-01 1.36796880e+00 -1.12041628e+00 -1.50748566e-01
-2.03155443e-01 4.29050565e-01 3.30299139e-01 -2.42371157e-01
-7.41680920e-01 -1.18059441e-01 2.38911927e-01 9.31192517e-01
2.04051167e-01 -2.69354343e-01 -4.86809075e-01 -8.36952388e-01
-4.47463393e-02 -7.66771063e-02 1.86038196e-01 -9.21612263e-01
-6.52244866e-01 4.47961718e-01 1.90671340e-01 -1.17927659e+00
-1.59263059e-01 -5.09665728e-01 -4.26858395e-01 8.30310106e-01
-1.68045866e+00 -1.77073884e+00 -8.84722114e-01 2.17684254e-01
2.59267390e-01 2.37117216e-01 8.60311091e-01 5.31972408e-01
-3.64760369e-01 -1.04301743e-01 2.11011961e-01 9.24189687e-02
2.63119377e-02 -1.26156461e+00 5.68472981e-01 5.57969451e-01
1.04917757e-01 4.38826531e-02 4.81062591e-01 -7.21252680e-01
-5.69801331e-01 -1.40893030e+00 1.22361636e+00 -4.76127237e-01
4.07359421e-01 -3.05787325e-01 -7.13044226e-01 5.28690755e-01
-6.08809710e-01 -5.13367504e-02 4.77801174e-01 7.74852708e-02
-2.62299217e-02 -3.59194219e-01 -1.52340305e+00 1.45904213e-01
1.01088941e+00 -2.35798240e-01 -1.89830542e-01 4.36822236e-01
3.38541150e-01 -1.46003172e-01 -1.11202860e+00 5.58692396e-01
5.49706757e-01 -1.44403112e+00 1.20427954e+00 1.68143526e-01
4.64498490e-01 -5.86262524e-01 -4.30702865e-01 -9.80101764e-01
-1.88735202e-02 2.95813650e-01 3.47356588e-01 1.40440631e+00
5.16960442e-01 -6.25133514e-01 8.89938056e-01 6.26027763e-01
-1.94882646e-01 -5.28284669e-01 -1.14004028e+00 -6.60788476e-01
1.34806767e-01 -7.39952266e-01 1.08708191e+00 4.91001934e-01
-6.81152940e-01 -4.07508165e-02 -4.69917357e-02 5.53547919e-01
6.21360242e-01 1.16483308e-01 1.11435378e+00 -1.90355623e+00
2.92853326e-01 -2.28306964e-01 -7.27594674e-01 -2.38345295e-01
-4.13504653e-02 -8.06098163e-01 -1.59347162e-01 -1.95925689e+00
-1.88553825e-01 -9.09175158e-01 2.08427384e-01 2.80889750e-01
1.82179883e-01 3.44850510e-01 4.47169170e-02 1.84961274e-01
6.63538873e-02 5.15957057e-01 5.24737477e-01 -1.16018012e-01
-3.05977672e-01 1.88011110e-01 -1.58028826e-01 7.84324110e-01
9.16983128e-01 -4.90300387e-01 -5.85975125e-02 -1.43303201e-01
3.00822288e-01 -2.72040665e-01 6.58731401e-01 -1.55315304e+00
-2.15298399e-01 6.01588115e-02 3.50485206e-01 -1.22427356e+00
2.17807949e-01 -8.82738709e-01 5.90592146e-01 5.74980736e-01
3.42798471e-01 -4.42462057e-01 8.84525627e-02 1.72413498e-01
8.48850533e-02 -3.66464078e-01 8.33879650e-01 -3.96850169e-01
-6.12877667e-01 3.19831312e-01 -1.77157342e-01 -2.15331882e-01
8.38809729e-01 -5.36918223e-01 -8.29146802e-02 1.98999457e-02
-7.13143706e-01 -5.04350848e-02 6.79234385e-01 5.93397673e-03
3.20338547e-01 -1.12460220e+00 -9.60740387e-01 3.57994944e-01
4.51147199e-01 3.23477745e-01 4.61575210e-01 6.47440851e-01
-9.74767506e-01 6.05243742e-01 -1.50276184e-01 -8.09595406e-01
-1.05872691e+00 3.22510511e-01 3.46717834e-01 -4.62380946e-01
-7.20450163e-01 3.76214772e-01 -3.27215850e-01 -1.13125038e+00
-4.44643438e-01 -4.86789227e-01 -3.80513400e-01 2.71139532e-01
2.66248941e-01 5.15237868e-01 4.93592709e-01 -1.16329694e+00
-4.51824278e-01 7.91227698e-01 8.21805656e-01 -1.54799640e-01
1.61120951e+00 -2.53455788e-01 3.19461524e-02 5.23075044e-01
7.32745945e-01 -2.95849815e-02 -9.12590325e-01 1.01444945e-01
1.75961360e-01 -6.16822839e-01 2.41029814e-01 -5.39067268e-01
-9.52295184e-01 7.88067698e-01 1.03761244e+00 1.72186330e-01
7.10385084e-01 5.95431216e-02 4.29914474e-01 5.33395149e-02
8.67322743e-01 -1.17304373e+00 -7.34014630e-01 -1.90624278e-02
7.61939228e-01 -1.43859422e+00 1.66157022e-01 -4.82714206e-01
-2.36082539e-01 9.52311039e-01 3.20247054e-01 7.73347244e-02
7.71449089e-01 4.93300706e-02 -1.54984772e-01 -4.32450503e-01
1.07904665e-01 -1.09996033e+00 -1.25641120e-03 8.93322706e-01
3.42144161e-01 5.34130335e-01 -3.84911984e-01 2.29398727e-01
-5.29402137e-01 2.32204318e-01 3.31334710e-01 6.99073493e-01
-7.65213490e-01 -8.80773902e-01 -7.91396141e-01 4.58606482e-01
2.71711200e-01 -1.55582041e-01 2.49178428e-03 1.05582130e+00
5.92489183e-01 9.35323894e-01 2.13591680e-02 -2.48763070e-01
4.12314981e-01 7.41281435e-02 2.21649766e-01 -3.85263175e-01
-4.53257322e-01 -2.76166230e-01 2.01264188e-01 -4.12852854e-01
-5.26547968e-01 -8.32899988e-01 -1.06205344e+00 -4.39278364e-01
-2.22568005e-01 -1.75803259e-01 1.32844830e+00 9.52910781e-01
1.24932446e-01 2.17887893e-01 5.37373543e-01 -1.17533815e+00
2.01809779e-01 -9.97013152e-01 -7.16865003e-01 2.87075430e-01
-1.66919380e-01 -8.41052294e-01 -4.05601501e-01 -2.50881732e-01] | [8.956365585327148, -1.9258956909179688] |
952a043c-d0c1-4a9b-b218-281d18549f0a | a-robust-pedestrian-detection-approach-for | 2210.10489 | null | https://arxiv.org/abs/2210.10489v1 | https://arxiv.org/pdf/2210.10489v1.pdf | A Robust Pedestrian Detection Approach for Autonomous Vehicles | Nowadays, utilizing Advanced Driver-Assistance Systems (ADAS) has absorbed a huge interest as a potential solution for reducing road traffic issues. Despite recent technological advances in such systems, there are still many inquiries that need to be overcome. For instance, ADAS requires accurate and real-time detection of pedestrians in various driving scenarios. To solve the mentioned problem, this paper aims to fine-tune the YOLOv5s framework for handling pedestrian detection challenges on the real-world instances of Caltech pedestrian dataset. We also introduce a developed toolbox for preparing training and test data and annotations of Caltech pedestrian dataset into the format recognizable by YOLOv5. Experimental results of utilizing our approach show that the mean Average Precision (mAP) of our fine-tuned model for pedestrian detection task is more than 91 percent when performing at the highest rate of 70 FPS. Moreover, the experiments on the Caltech pedestrian dataset samples have verified that our proposed approach is an effective and accurate method for pedestrian detection and can outperform other existing methodologies. | ['Asadollah Shahbahrami', 'Ali Tourani', 'Bahareh Ghari'] | 2022-10-19 | null | null | null | null | ['pedestrian-detection'] | ['computer-vision'] | [-2.35853344e-01 -2.39066422e-01 2.06816196e-01 -4.16190922e-01
-6.19109690e-01 -2.12476462e-01 4.62540478e-01 -1.99323017e-02
-8.06141496e-01 6.52688861e-01 -4.51848239e-01 -4.40417141e-01
5.07297635e-01 -7.26949692e-01 -5.19148290e-01 -6.82890892e-01
3.78379971e-01 1.30520603e-02 9.96641278e-01 -4.90839511e-01
1.36016294e-01 6.29661620e-01 -1.98473716e+00 1.48323923e-01
8.84447038e-01 8.16879630e-01 3.63037586e-01 9.05885816e-01
2.09633574e-01 3.49172086e-01 -8.54995131e-01 -7.21326411e-01
5.27835965e-01 1.61608949e-01 -2.89845675e-01 5.91520071e-02
8.17946196e-01 -4.84945089e-01 -4.42141205e-01 9.61977065e-01
7.81584203e-01 9.39478576e-02 3.63358200e-01 -1.53419733e+00
-3.21462840e-01 -1.72216386e-01 -6.99348450e-01 5.81339777e-01
2.38189012e-01 8.01694036e-01 3.63762587e-01 -8.16090524e-01
1.46218613e-01 1.32672775e+00 6.51934683e-01 4.59159046e-01
-7.93380737e-01 -7.52776206e-01 9.67669412e-02 6.69650495e-01
-1.43348134e+00 -4.52055484e-01 2.73795813e-01 -5.36436677e-01
7.74025381e-01 2.07647249e-01 6.40635550e-01 9.19167638e-01
2.35780880e-01 9.12868083e-01 1.08474517e+00 -4.07093942e-01
7.35048130e-02 6.53300583e-01 5.57375133e-01 6.68120027e-01
6.78104997e-01 3.27520847e-01 -2.31103212e-01 3.03645521e-01
3.72211277e-01 -3.47814560e-01 1.82305902e-01 -2.31717199e-01
-9.14570212e-01 6.90896869e-01 3.62258226e-01 -6.66341186e-02
-1.74612120e-01 -1.66978136e-01 6.41411722e-01 -1.32811507e-02
1.20092317e-01 -1.79771677e-01 -3.32977295e-01 -9.05504674e-02
-5.28409541e-01 4.57333922e-01 2.81929135e-01 1.31435692e+00
5.60984731e-01 1.60088554e-01 -2.71678627e-01 6.31679118e-01
2.01311767e-01 8.84705305e-01 1.34355769e-01 -6.02862597e-01
5.80850840e-01 6.53421700e-01 4.77513582e-01 -9.56571043e-01
-3.20718706e-01 -3.42787236e-01 -5.79089224e-01 7.11004615e-01
7.97876894e-01 -2.10552871e-01 -7.57825017e-01 1.10231161e+00
5.79641461e-01 4.38323170e-02 3.98745313e-02 1.14933920e+00
1.09903789e+00 6.99159861e-01 6.37784541e-01 2.20305651e-01
1.75830197e+00 -1.03746712e+00 -4.34588790e-01 -3.70764434e-01
5.12706757e-01 -1.03827536e+00 9.65884268e-01 2.74656922e-01
-6.46449208e-01 -1.24743891e+00 -1.11390126e+00 -1.32434428e-01
-5.73832214e-01 7.95530200e-01 4.25371766e-01 1.36125302e+00
-9.06510770e-01 -1.94311172e-01 -5.24634123e-01 -8.14987481e-01
2.82039702e-01 2.36938804e-01 -4.15113926e-01 -1.59892607e-02
-9.54541445e-01 1.16575503e+00 4.49849725e-01 3.67857158e-01
-8.16531539e-01 -4.39119041e-01 -7.08409309e-01 -1.30147070e-01
4.21915680e-01 -5.48465669e-01 1.04334152e+00 -4.27852571e-01
-1.32421923e+00 9.91017342e-01 -1.47841558e-01 -9.02419031e-01
8.13107491e-01 -3.47171128e-01 -6.27870679e-01 -7.03928620e-02
2.44273037e-01 8.73861015e-01 6.84799194e-01 -9.80464637e-01
-1.45204496e+00 -1.22748725e-01 2.73438752e-01 -7.46724531e-02
-1.35625586e-01 1.99208289e-01 -4.87857521e-01 -3.67440656e-02
-4.97263551e-01 -9.80475664e-01 -2.79409230e-01 1.67411089e-01
-2.86494702e-01 -4.37246382e-01 1.07524574e+00 -6.24797821e-01
7.72397697e-01 -2.06583238e+00 -5.64683199e-01 -1.76088184e-01
5.59030212e-02 1.13260508e+00 -1.03373952e-01 3.66586521e-02
2.45152920e-01 -5.24002373e-01 1.62055895e-01 -2.15273008e-01
-8.43160450e-02 -6.83635473e-03 -4.45200168e-02 3.64288479e-01
2.26778537e-01 6.33519888e-01 -8.05279732e-01 -6.88404918e-01
9.82138991e-01 6.12153649e-01 -2.46608108e-01 3.44817847e-01
3.23786050e-01 2.99750298e-01 -3.69845241e-01 6.98600769e-01
9.68051076e-01 5.09783089e-01 -4.79439318e-01 -2.81017125e-01
-5.82025766e-01 -3.52037013e-01 -1.32946026e+00 8.11086237e-01
-1.84246749e-01 9.86841321e-01 -1.89607129e-01 -9.01037216e-01
1.01626182e+00 -6.26276881e-02 1.95722580e-02 -8.25083733e-01
2.51777649e-01 -2.58093840e-03 3.93524356e-02 -6.97616696e-01
8.74988139e-01 4.18775439e-01 2.06156354e-03 -3.22778702e-01
-3.68341841e-02 2.01189294e-01 5.89727342e-01 1.10958725e-01
7.56757081e-01 -2.81957984e-02 4.14048672e-01 -2.70487726e-01
1.25767398e+00 3.21963519e-01 4.09013063e-01 9.35582757e-01
-1.00635505e+00 2.64813423e-01 8.48397687e-02 -7.73134530e-01
-1.13930655e+00 -1.10017598e+00 -9.05203819e-02 9.36051071e-01
2.31192887e-01 -1.89281866e-01 -1.08167768e+00 -5.68738997e-01
-6.47660568e-02 6.81834459e-01 -3.10214579e-01 -7.55384658e-03
-5.74223101e-01 -8.50090563e-01 5.75674832e-01 6.22200727e-01
1.06758189e+00 -8.07277381e-01 -9.88132775e-01 1.32109925e-01
-1.35654360e-01 -1.70076180e+00 -1.56097963e-01 -3.92015100e-01
-2.59297788e-01 -1.42167139e+00 -8.60001564e-01 -8.60897064e-01
4.83174205e-01 1.01417840e+00 6.83463454e-01 -4.44685780e-02
-8.43592763e-01 3.17351043e-01 -2.75982201e-01 -6.11206353e-01
-3.18213403e-01 3.32147144e-02 -8.67941529e-02 8.50828141e-02
8.18816781e-01 3.26436400e-01 -8.51521730e-01 7.17224002e-01
-1.92634284e-01 -3.10899634e-02 5.85842848e-01 5.04848480e-01
3.07500333e-01 1.02675200e-01 6.08042121e-01 -4.86895055e-01
3.17895949e-01 -8.31449553e-02 -1.20092380e+00 8.32677931e-02
-3.81257862e-01 -3.65109026e-01 4.61958200e-01 -6.51748925e-02
-1.36713231e+00 4.97304231e-01 -4.37325269e-01 2.22134441e-01
-7.37851977e-01 -4.24545884e-01 -2.67589897e-01 -3.96630824e-01
7.79660881e-01 1.25500411e-01 -2.05849022e-01 -2.23441258e-01
4.85747337e-01 8.58764529e-01 8.49763572e-01 -3.99671406e-01
7.82904029e-01 4.51595366e-01 -3.94130824e-03 -1.27566826e+00
-6.67095065e-01 -7.61702418e-01 -8.45705092e-01 -5.91169238e-01
1.07175279e+00 -1.10009742e+00 -1.12034369e+00 7.81120896e-01
-1.10679030e+00 7.15919659e-02 1.44572919e-02 4.05569762e-01
-1.63002968e-01 6.03281915e-01 -2.84196883e-01 -1.15995908e+00
-3.17768246e-01 -1.36085641e+00 1.11638498e+00 5.31385660e-01
2.40250364e-01 -5.58098912e-01 -5.04340410e-01 7.29736209e-01
4.53120559e-01 2.83214986e-01 2.87099570e-01 -8.65838379e-02
-7.81253576e-01 -5.98550439e-01 -7.81696022e-01 4.57210183e-01
-1.92175016e-01 2.00873464e-01 -1.16016340e+00 -1.19585559e-01
-4.40736413e-01 1.33206472e-01 8.32072020e-01 3.65542263e-01
7.88736224e-01 3.25426251e-01 -3.85236204e-01 1.45280138e-01
1.16262805e+00 3.35367024e-01 6.80072665e-01 7.12735534e-01
5.25894046e-01 5.64023495e-01 1.11828792e+00 3.09751868e-01
6.20924532e-01 9.27977502e-01 3.87347162e-01 -1.08150229e-01
-4.20251310e-01 -6.42764121e-02 3.59301209e-01 6.42571449e-02
-2.06259221e-01 -1.69333257e-02 -8.49726617e-01 4.81408000e-01
-1.86821115e+00 -1.10773396e+00 -8.39326143e-01 2.19474912e+00
2.43111625e-01 4.38364029e-01 6.46327674e-01 3.62431228e-01
9.14792180e-01 -1.98405713e-01 -3.01855713e-01 -3.67491841e-01
-1.15352102e-01 -1.71620369e-01 7.66183794e-01 4.19451922e-01
-1.49378228e+00 1.23646140e+00 6.25780058e+00 7.28839934e-01
-9.06374872e-01 7.96005651e-02 5.14142513e-01 3.39333177e-01
8.81908417e-01 -3.93705875e-01 -1.53785002e+00 5.69475710e-01
1.03769898e+00 -1.49779573e-01 -1.76678002e-01 1.29820526e+00
5.67413390e-01 -4.22615349e-01 -6.30096018e-01 1.08382010e+00
-4.79530171e-02 -9.32040274e-01 -1.30799949e-01 -1.86499298e-01
4.67106491e-01 -3.10055405e-01 1.17473178e-01 5.17697275e-01
3.66649508e-01 -6.81332529e-01 6.94636643e-01 2.57526219e-01
5.15722811e-01 -8.55048418e-01 1.23600590e+00 4.18480098e-01
-1.49851716e+00 -1.01147499e-02 -9.15351272e-01 -6.71574548e-02
2.65498489e-01 3.39582950e-01 -1.14543056e+00 3.41818124e-01
7.22025514e-01 4.49842304e-01 -1.10386670e+00 1.51722288e+00
-2.81795323e-01 3.88822883e-01 -1.86804503e-01 -1.79372638e-01
2.69561946e-01 9.73272778e-04 4.64832962e-01 1.58897543e+00
1.88753262e-01 2.18734760e-02 5.49989007e-02 3.77684742e-01
6.32352769e-01 6.01047911e-02 -5.27084470e-01 5.42770922e-01
3.05037588e-01 1.54289186e+00 -6.40724182e-01 -5.18584728e-01
-5.73647857e-01 8.09861720e-01 -1.61216147e-02 2.46500522e-01
-1.17945266e+00 -2.65005529e-01 7.89842546e-01 2.00923994e-01
3.33872348e-01 -4.44405496e-01 -3.43160152e-01 -7.74787903e-01
-1.77964363e-02 -5.72132647e-01 2.88353384e-01 -8.31629217e-01
-8.00800025e-01 7.25049675e-01 6.10899925e-02 -1.33016956e+00
1.08835131e-01 -1.12978530e+00 -4.80478704e-01 8.10115576e-01
-1.62649858e+00 -1.33659244e+00 -8.91027510e-01 6.29030943e-01
8.74774575e-01 -4.87567276e-01 4.29155558e-01 6.68291330e-01
-8.14326048e-01 8.14102888e-01 -3.30183238e-01 1.15662351e-01
6.30913496e-01 -1.00257349e+00 7.36937404e-01 1.30165684e+00
-5.19098878e-01 2.67517686e-01 1.01916933e+00 -2.76640952e-01
-1.24340820e+00 -1.36328936e+00 7.35210001e-01 -5.35866082e-01
3.60559255e-01 -2.06177309e-01 -5.16515732e-01 4.98819798e-01
1.52943075e-01 5.72031811e-02 2.63020486e-01 -4.04389650e-01
-1.50541872e-01 -4.72753584e-01 -1.35278976e+00 5.25112510e-01
9.18336928e-01 -1.10429049e-01 -5.00935018e-01 2.78680384e-01
2.84603179e-01 -5.01654983e-01 -3.70153487e-01 1.00389533e-01
5.45926392e-01 -1.20854294e+00 1.28284371e+00 -3.77132058e-01
-4.29369397e-02 -8.26912522e-01 -1.65290043e-01 -1.04330027e+00
-1.76822022e-01 -1.87586784e-01 1.67096168e-01 1.22263360e+00
1.97940785e-02 -7.47752428e-01 7.98368216e-01 6.64125323e-01
-2.38114938e-01 -2.39638537e-01 -9.34847772e-01 -8.92386794e-01
-3.09635371e-01 -4.26116049e-01 3.12606782e-01 1.69062793e-01
-5.00372827e-01 2.13454649e-01 -5.55625141e-01 5.13382792e-01
9.99318004e-01 -4.71173942e-01 1.59944940e+00 -1.05838132e+00
1.40439048e-01 -1.80338755e-01 -1.20202279e+00 -1.15585887e+00
-1.78414315e-01 -2.41225168e-01 1.18042415e-04 -1.52126169e+00
9.70778316e-02 -3.05844754e-01 -3.56432609e-02 1.21274233e-01
-4.25889790e-01 3.93729210e-01 2.65878737e-01 -1.64377108e-01
-6.16058648e-01 2.77647287e-01 9.43617463e-01 -1.32400155e-01
1.98173895e-01 3.02648902e-01 -5.20030260e-01 7.03786135e-01
7.12845922e-01 -1.65598601e-01 -1.15587592e-01 -1.45998701e-01
-4.82160062e-01 -2.45788157e-01 7.91358411e-01 -1.53227019e+00
3.59248608e-01 -4.87087965e-02 4.70373362e-01 -1.05767715e+00
4.75199014e-01 -7.96772361e-01 -2.22621202e-01 7.56690562e-01
2.01754063e-01 4.16267253e-02 4.87349719e-01 4.71813351e-01
-1.57541689e-02 9.35593545e-02 1.27647805e+00 2.19110250e-01
-1.52035379e+00 1.29703447e-01 -6.22065485e-01 -2.30760649e-01
1.58013451e+00 -3.95806134e-01 -6.60747528e-01 -5.49907424e-02
-3.85458589e-01 3.29020143e-01 1.19079076e-01 7.68799365e-01
6.63671374e-01 -1.09486890e+00 -7.76619852e-01 3.79964471e-01
3.89281124e-01 -3.18132043e-01 3.66665572e-01 6.05313361e-01
-6.75237954e-01 9.19277191e-01 -6.63433969e-01 -8.86017263e-01
-1.75474501e+00 6.43875062e-01 3.49233449e-01 3.04934442e-01
-7.21313238e-01 6.36136711e-01 3.14503223e-01 -9.32396203e-02
1.69899568e-01 -2.73281902e-01 -4.07785922e-01 -2.49507189e-01
8.61425459e-01 8.64714265e-01 5.50238155e-02 -1.13894868e+00
-4.75638330e-01 5.32729149e-01 -9.82799008e-02 2.43336201e-01
8.26029837e-01 -3.57102543e-01 5.69458365e-01 2.67523006e-02
7.20253050e-01 -2.22137898e-01 -1.31802225e+00 3.56831700e-02
-1.65460795e-01 -7.44301677e-01 -1.63262069e-01 -6.55787885e-01
-8.82731736e-01 9.79817450e-01 1.20700347e+00 7.56115243e-02
7.08117604e-01 -2.82565653e-01 9.48343694e-01 5.00156701e-01
5.59068978e-01 -1.11217833e+00 -2.17165172e-01 3.61456424e-01
5.17752290e-01 -1.70556200e+00 -1.84433401e-01 -7.06746459e-01
-7.15318441e-01 1.14550221e+00 1.04993951e+00 -1.66352734e-01
3.14011872e-01 2.72297338e-02 3.07603538e-01 1.91522300e-01
-2.26223379e-01 -6.24128699e-01 2.60094464e-01 1.13073814e+00
2.33091459e-01 2.95463920e-01 -4.24236238e-01 2.13035136e-01
-2.53202856e-01 -1.56203896e-01 4.53344941e-01 5.65844178e-01
-9.57978308e-01 -8.64145875e-01 -6.97650254e-01 9.65999663e-02
-8.85080770e-02 2.94047922e-01 -4.94246110e-02 1.03459358e+00
4.13732111e-01 1.31265450e+00 -3.15194488e-01 -2.76561528e-01
7.82282472e-01 -2.16331005e-01 2.38464937e-01 -2.60973543e-01
-1.90304354e-01 -4.26053643e-01 1.38272822e-01 -4.57939595e-01
-3.20814341e-01 -6.25944495e-01 -8.86143446e-01 -5.57143033e-01
-2.14860275e-01 -4.41745110e-02 7.92511284e-01 9.86841619e-01
2.74481744e-01 5.18063247e-01 3.49102765e-01 -8.88868332e-01
-3.29383641e-01 -6.82254434e-01 -1.95315763e-01 2.63556510e-01
1.98871195e-01 -9.94007528e-01 -2.07873769e-02 1.29927531e-01] | [7.9616618156433105, -0.8374708890914917] |
9bd8fcb2-90a1-4a0e-b095-b6f20dcf23c1 | optimizing-neural-networks-through-activation | 2304.03374 | null | https://arxiv.org/abs/2304.03374v1 | https://arxiv.org/pdf/2304.03374v1.pdf | Optimizing Neural Networks through Activation Function Discovery and Automatic Weight Initialization | Automated machine learning (AutoML) methods improve upon existing models by optimizing various aspects of their design. While present methods focus on hyperparameters and neural network topologies, other aspects of neural network design can be optimized as well. To further the state of the art in AutoML, this dissertation introduces techniques for discovering more powerful activation functions and establishing more robust weight initialization for neural networks. These contributions improve performance, but also provide new perspectives on neural network optimization. First, the dissertation demonstrates that discovering solutions specialized to specific architectures and tasks gives better performance than reusing general approaches. Second, it shows that jointly optimizing different components of neural networks is synergistic, and results in better performance than optimizing individual components alone. Third, it demonstrates that learned representations are easier to optimize than hard-coded ones, creating further opportunities for AutoML. The dissertation thus makes concrete progress towards fully automatic machine learning in the future. | ['Garrett Bingham'] | 2023-04-06 | null | null | null | null | ['automl'] | ['methodology'] | [ 1.54315010e-01 2.44666502e-01 -5.40562630e-01 -4.56170410e-01
-4.48636591e-01 -5.48210204e-01 1.66676387e-01 -9.04136449e-02
-3.60694975e-01 7.11178184e-01 -3.16146873e-02 -2.65949816e-01
-4.40635949e-01 -4.06005234e-01 -5.13882995e-01 -7.25699544e-01
-1.73601165e-01 4.78756964e-01 -2.82586277e-01 -1.42795041e-01
3.79187554e-01 9.35484409e-01 -1.52936268e+00 3.24952543e-01
3.62843990e-01 6.96243048e-01 5.10597005e-02 7.08471417e-01
-1.91577956e-01 6.51685357e-01 -7.16121733e-01 -9.75828171e-02
1.46971896e-01 -6.79458529e-02 -7.85086572e-01 -5.23592941e-02
5.55979311e-01 2.43389115e-01 -5.17454557e-02 7.73390889e-01
4.43408757e-01 3.76891404e-01 6.41857564e-01 -1.37704980e+00
-5.32229364e-01 9.88435984e-01 -1.75849870e-01 1.87577307e-01
-2.21409053e-01 7.58808702e-02 1.08737516e+00 -7.50839055e-01
4.21282388e-02 1.17415988e+00 9.39884126e-01 7.44075835e-01
-1.43580270e+00 -8.02483618e-01 4.80275989e-01 2.01608822e-01
-1.29146266e+00 -6.97174430e-01 9.16621387e-01 -3.90323550e-01
1.50241697e+00 2.09466934e-01 7.38230407e-01 1.03626215e+00
-1.59998208e-01 1.15875697e+00 6.21359468e-01 -5.86650610e-01
1.59611315e-01 3.83381277e-01 5.01888096e-01 7.52867818e-01
5.36775053e-01 1.11003578e-01 -3.07817578e-01 6.67987689e-02
7.34058201e-01 -3.81095767e-01 -3.83871756e-02 -2.71938056e-01
-7.61084855e-01 8.73678267e-01 1.29786074e-01 6.89496398e-01
-2.94058859e-01 3.73092175e-01 1.90689027e-01 2.74216086e-01
3.87839740e-03 1.41412866e+00 -7.97747016e-01 -1.48836330e-01
-1.18014920e+00 1.78098291e-01 8.68994832e-01 7.16138899e-01
1.03279614e+00 7.39638627e-01 -3.52238975e-02 1.20508456e+00
3.43415529e-01 1.77906957e-02 5.25670230e-01 -1.43404758e+00
3.24368626e-01 7.78760970e-01 -3.66112530e-01 -9.88616407e-01
-8.03292096e-01 -8.22940469e-01 -6.65279210e-01 2.40707129e-01
3.65504891e-01 -4.32087481e-01 -7.58833051e-01 1.77552521e+00
-1.65150583e-01 -1.19735338e-01 2.39897221e-02 3.36411446e-01
7.48707235e-01 7.21217573e-01 9.94615853e-02 -1.52363852e-01
9.61449325e-01 -1.17501414e+00 -6.01721048e-01 -6.44689381e-01
7.15932727e-01 -4.19597089e-01 8.22048604e-01 6.38039172e-01
-1.33899796e+00 -4.16019410e-01 -1.20886409e+00 1.56362876e-01
-5.35362899e-01 4.44022864e-01 1.19541693e+00 9.79516149e-01
-1.08210886e+00 7.17953503e-01 -8.76201570e-01 -1.02110118e-01
4.63793904e-01 8.48689377e-01 -7.44554028e-02 3.56503159e-01
-1.01600623e+00 1.14299405e+00 7.55339205e-01 1.89125821e-01
-5.72360158e-01 -7.90732384e-01 -9.28494513e-01 2.50949919e-01
2.32830763e-01 -4.64751273e-01 1.31023455e+00 -1.21260726e+00
-1.63464463e+00 5.42392373e-01 -5.50448857e-02 -4.59741116e-01
-2.34322011e-01 3.13657662e-03 -2.81212360e-01 -3.98603678e-02
-4.91749942e-01 1.00835490e+00 8.21573794e-01 -1.21661389e+00
-5.03105938e-01 -1.49218753e-01 -2.40101125e-02 1.60003647e-01
-9.59749758e-01 -1.34783864e-01 -7.10461736e-01 -7.05015182e-01
-1.46372750e-01 -6.72318101e-01 -4.61380005e-01 -3.27425241e-01
-3.39104682e-01 -4.13073570e-01 4.73649502e-01 -2.21660569e-01
1.59771490e+00 -1.72468197e+00 3.17661673e-01 5.22038937e-01
1.95204034e-01 3.93263191e-01 -4.90573287e-01 5.40936105e-02
-4.65002686e-01 2.19781026e-01 -3.41888145e-02 -2.26511136e-01
9.94002447e-03 3.88028800e-01 1.66869923e-01 9.56784114e-02
2.78832793e-01 9.86425519e-01 -4.97665495e-01 -4.48194206e-01
1.75124913e-01 4.43137944e-01 -5.80859482e-01 -2.33024105e-01
-7.86670521e-02 -1.59736767e-01 -3.64412010e-01 9.23994303e-01
1.18941307e-01 -3.29620659e-01 2.06621557e-01 -3.76291662e-01
-9.04828403e-03 1.22846656e-01 -1.28932464e+00 1.30972397e+00
-4.91307497e-01 1.03942823e+00 2.93267608e-01 -1.51930749e+00
1.04933691e+00 2.66086251e-01 6.89132273e-01 -3.14376444e-01
3.14196169e-01 -1.19677521e-01 9.57034435e-03 -6.08883977e-01
3.92523915e-01 1.70097485e-01 2.12596878e-01 5.96351266e-01
3.99914593e-01 -1.31312758e-01 5.54911494e-01 -2.98114985e-01
1.02664030e+00 3.70540023e-02 1.81925833e-01 -2.41637930e-01
4.21932638e-01 4.94563095e-02 5.15077353e-01 9.22841549e-01
8.67182165e-02 1.56301767e-01 3.61553997e-01 -4.72140580e-01
-9.43522811e-01 -6.53061509e-01 -1.56368464e-01 1.49713850e+00
-4.75695878e-01 -2.58562863e-01 -1.00867474e+00 -3.78255069e-01
1.68645799e-01 8.17582190e-01 -6.98369920e-01 -1.05322860e-01
-7.35500991e-01 -1.18089104e+00 8.94273400e-01 7.73410380e-01
2.29105130e-01 -1.04226875e+00 -3.81204098e-01 2.59212404e-01
1.67670861e-01 -9.23099279e-01 1.75780077e-02 6.15309119e-01
-1.55719602e+00 -9.13566768e-01 -4.67040628e-01 -9.26383793e-01
7.91026890e-01 3.41160968e-03 1.19241786e+00 2.02137202e-01
-4.02273208e-01 3.66222203e-01 1.35620028e-01 -4.11396950e-01
-2.28566974e-01 7.53996551e-01 -1.72346942e-02 -4.11415398e-01
4.11302596e-01 -6.67227507e-01 1.84324495e-02 6.68128133e-02
-6.50190890e-01 -5.53934427e-04 8.66157174e-01 8.20451438e-01
3.47909898e-01 2.42446646e-01 6.54497802e-01 -9.32292342e-01
1.06917644e+00 -1.58470079e-01 -5.57244301e-01 4.53805953e-01
-1.27297843e+00 5.45234084e-01 4.23793972e-01 -6.79674387e-01
-9.37994897e-01 3.09627473e-01 -8.30363482e-02 -3.00134122e-01
-3.04633409e-01 3.87323499e-01 7.63848126e-02 -2.45205894e-01
1.00223279e+00 -1.55498058e-01 1.38574809e-01 -5.56940615e-01
1.82015270e-01 2.88483381e-01 2.92474508e-01 -9.80577230e-01
6.03875875e-01 -1.53469875e-01 -7.95206800e-02 -8.90343726e-01
-6.64516687e-01 -2.04693958e-01 -4.89913106e-01 -1.95112318e-01
4.16978836e-01 -6.04461968e-01 -8.95919144e-01 1.56792581e-01
-1.01664770e+00 -5.76830983e-01 -3.66489679e-01 5.02119303e-01
-4.54800248e-01 -1.54796094e-01 -6.52916193e-01 -8.54440153e-01
-2.34582096e-01 -1.46242285e+00 5.80676317e-01 4.69198257e-01
-8.29299271e-01 -1.44406784e+00 6.01287521e-02 3.17280293e-01
4.38127398e-01 -2.38021091e-01 1.00952923e+00 -8.31084013e-01
-2.72118717e-01 -3.74575377e-01 5.08716367e-02 4.74645734e-01
-6.64170533e-02 2.41071060e-01 -1.10845137e+00 -2.07921088e-01
-3.47675681e-01 -3.72605443e-01 7.99449861e-01 6.21672094e-01
1.44255030e+00 -5.12029409e-01 -4.03008789e-01 8.46232712e-01
1.23652923e+00 4.37126338e-01 5.37442029e-01 7.61883736e-01
5.61633408e-01 8.17287028e-01 3.39354612e-02 1.47954896e-01
1.24931164e-01 5.14807284e-01 2.03705072e-01 -2.89389431e-01
-7.48353079e-02 2.13798314e-01 3.70783746e-01 6.20847642e-01
-3.88019383e-01 9.03683975e-02 -9.25851583e-01 4.26122040e-01
-1.68037701e+00 -1.02487779e+00 1.30949840e-01 1.75869679e+00
1.00147450e+00 1.06033601e-01 1.41467806e-02 4.32710737e-01
5.63716054e-01 1.72260888e-02 -6.99811578e-01 -7.20212281e-01
-2.47120321e-01 6.35102689e-01 5.93274117e-01 8.02760661e-01
-1.04386568e+00 1.03043234e+00 8.73237801e+00 8.56767297e-01
-9.32295680e-01 -2.11504951e-01 5.94543517e-01 -5.39914012e-01
-2.60041118e-01 -2.69389987e-01 -1.24969721e+00 -3.58104371e-02
9.06240344e-01 4.79039401e-02 8.06347728e-01 1.19863653e+00
2.36244664e-01 1.28450677e-01 -1.24483526e+00 9.08597052e-01
1.34673655e-01 -1.74830627e+00 -6.25025034e-02 -8.60849395e-02
8.75427425e-01 -2.12008581e-01 6.25297725e-02 4.97576594e-01
4.81228173e-01 -1.27022994e+00 3.00419211e-01 5.01275539e-01
2.24908769e-01 -8.47954988e-01 5.71883917e-01 9.09799337e-02
-9.32197928e-01 -5.06673872e-01 -3.14130694e-01 -1.16710894e-01
-3.32487524e-01 4.75786746e-01 -7.43493795e-01 -2.53009982e-02
6.19817853e-01 7.57082462e-01 -8.29868257e-01 1.06473577e+00
-4.66812402e-01 7.40778089e-01 -4.00052100e-01 -5.10996222e-01
2.15017170e-01 -9.87746194e-03 3.58311594e-01 1.65849936e+00
2.10836306e-02 -2.67995119e-01 -1.85970008e-01 9.22853172e-01
-3.82053107e-02 -6.03907704e-02 -6.56968534e-01 -2.80815095e-01
5.79921722e-01 1.36442578e+00 -5.25069296e-01 -1.19183592e-01
1.37852142e-02 3.53697449e-01 4.36363041e-01 7.85531282e-01
-4.93280709e-01 -5.77416956e-01 8.11045647e-01 -2.48717621e-01
-3.21581624e-02 -1.99856654e-01 -8.76172543e-01 -9.09617305e-01
-3.48263651e-01 -1.08373511e+00 3.58971566e-01 -7.14030445e-01
-9.60863829e-01 3.57458442e-01 2.67724365e-01 -6.23614013e-01
-4.34631348e-01 -1.04336476e+00 -6.50137842e-01 5.42103231e-01
-1.11186707e+00 -7.08431363e-01 5.45353023e-03 3.90597880e-01
6.68763459e-01 -5.50107419e-01 1.00606370e+00 2.47973233e-01
-1.22880924e+00 1.02000475e+00 1.35718826e-02 8.15712214e-02
3.91470045e-01 -9.53857183e-01 -2.04323772e-02 6.29386365e-01
2.42313400e-01 8.80257189e-01 5.15649021e-01 -1.76811934e-01
-1.43249905e+00 -7.16719747e-01 7.09705591e-01 -2.57937193e-01
6.02615178e-01 1.23160258e-02 -8.30016017e-01 5.74859262e-01
2.05700606e-01 -7.72086918e-01 8.58416319e-01 6.96612239e-01
-1.76797196e-01 -3.25431764e-01 -8.11577499e-01 7.50086904e-01
6.31973326e-01 -1.79102287e-01 -4.35565442e-01 2.19288006e-01
2.08160818e-01 -2.19125748e-01 -9.23855364e-01 4.95952398e-01
7.55843759e-01 -6.21841252e-01 1.24432385e+00 -8.85372519e-01
1.98456079e-01 -5.62579148e-02 6.83391988e-02 -1.22979414e+00
-5.56740046e-01 -6.41436398e-01 -4.59159940e-01 1.09640169e+00
1.01074862e+00 -5.34157097e-01 1.34042811e+00 9.70476866e-01
-3.23709637e-01 -1.03530049e+00 -5.58537900e-01 -7.11104810e-01
1.61751583e-01 -8.94292474e-01 5.14354289e-01 1.09750319e+00
1.07799657e-01 4.75558937e-01 -5.71148768e-02 -8.71135592e-02
5.57353914e-01 -2.34734803e-01 4.61636603e-01 -1.21492636e+00
-4.43442971e-01 -1.35546219e+00 -2.26583570e-01 -1.02508295e+00
3.73873889e-01 -9.09988761e-01 -9.63666514e-02 -1.40110648e+00
-6.14800043e-02 -5.71264982e-01 -4.18875962e-01 1.14700651e+00
3.11848428e-02 2.62202352e-01 -1.86617319e-02 -1.42608434e-02
-2.20559463e-01 9.92771760e-02 1.02176583e+00 -3.70043427e-01
-4.66706246e-01 9.79792252e-02 -9.76238191e-01 8.75735164e-01
1.27495527e+00 -3.95754933e-01 -5.44558406e-01 -6.70044780e-01
3.68659347e-01 -2.73156345e-01 -7.09555596e-02 -1.06932950e+00
4.34616476e-01 -4.21475977e-01 7.22914100e-01 -3.29673797e-01
5.08836627e-01 -6.73148751e-01 -5.10611050e-02 2.50306100e-01
-7.47079134e-01 8.22231397e-02 6.38144076e-01 -1.41896512e-02
-7.78164417e-02 -8.98933470e-01 8.42234492e-01 -1.58727035e-01
-5.51194966e-01 2.14678906e-02 -8.46983314e-01 -1.72333181e-01
8.12348247e-01 -7.15837002e-01 -1.17251463e-01 -2.18658656e-01
-9.46335614e-01 3.85002911e-01 5.31221405e-02 3.64674449e-01
4.73394930e-01 -1.17245400e+00 -3.01175207e-01 2.84656495e-01
-2.72565365e-01 -2.37705842e-01 -2.18472794e-01 6.62552893e-01
-3.10231388e-01 5.65965116e-01 -2.29301497e-01 -2.31520340e-01
-1.29658377e+00 1.66236535e-01 5.86950421e-01 -3.47931415e-01
-6.87095672e-02 1.12932944e+00 -3.38121623e-01 -4.95283604e-01
9.71949100e-01 -1.06790371e-01 -5.24310768e-01 2.01212868e-01
3.88658822e-01 5.50121665e-01 7.05083162e-02 -1.14877276e-01
-2.37871557e-01 7.10579634e-01 -1.63471863e-01 -2.35939547e-02
1.57391286e+00 4.33658481e-01 -1.53441485e-02 3.75248462e-01
1.05132556e+00 -3.95988882e-01 -1.09114790e+00 1.13143660e-02
1.80866331e-01 1.54595643e-01 2.65939027e-01 -7.24069893e-01
-1.43414986e+00 9.81473267e-01 4.26667035e-01 8.71165246e-02
1.33241832e+00 -2.50523806e-01 1.18052535e-01 1.28706789e+00
-8.63792598e-02 -1.57192338e+00 2.17761308e-01 7.49991357e-01
7.13130355e-01 -9.88280237e-01 3.99106771e-01 -2.60563027e-02
-4.95692998e-01 1.65456343e+00 7.63026536e-01 -6.87636528e-03
5.94873965e-01 7.10506558e-01 -1.02209195e-01 -1.49643496e-01
-8.76902580e-01 4.77322042e-02 3.88378739e-01 7.11203277e-01
6.99244201e-01 -2.41826549e-01 3.76384631e-02 6.55864775e-01
-4.36889857e-01 -2.04563484e-01 1.26793116e-01 7.57250786e-01
-6.20813131e-01 -1.42170596e+00 -5.47036827e-01 5.40948212e-01
-4.55987304e-01 -1.32839635e-01 -4.28888589e-01 9.39645886e-01
3.65410978e-03 8.82973552e-01 -1.04002193e-01 -6.00648165e-01
2.39027336e-01 2.87388831e-01 7.05255032e-01 -6.24361098e-01
-1.01213598e+00 1.25021022e-02 3.17213684e-01 -4.23185319e-01
-2.57588804e-01 -4.96287823e-01 -1.17155969e+00 -1.41561151e-01
-4.38062638e-01 2.76593179e-01 9.94401693e-01 9.69886720e-01
2.31064588e-01 7.56868601e-01 2.95990080e-01 -1.10975361e+00
-4.07542974e-01 -8.28000069e-01 -2.27180392e-01 -2.92142242e-01
2.45091710e-02 -6.59110785e-01 -3.23350042e-01 2.05926239e-01] | [8.548623085021973, 3.3467648029327393] |
2c8f45c5-25d3-49a5-9697-818e15a92d90 | generative-adversarial-networks-for-image | 2204.04707 | null | https://arxiv.org/abs/2204.04707v2 | https://arxiv.org/pdf/2204.04707v2.pdf | Generative Adversarial Networks for Image Augmentation in Agriculture: A Systematic Review | In agricultural image analysis, optimal model performance is keenly pursued for better fulfilling visual recognition tasks (e.g., image classification, segmentation, object detection and localization), in the presence of challenges with biological variability and unstructured environments. Large-scale, balanced and ground-truthed image datasets, however, are often difficult to obtain to fuel the development of advanced, high-performance models. As artificial intelligence through deep learning is impacting analysis and modeling of agricultural images, data augmentation plays a crucial role in boosting model performance while reducing manual efforts for data preparation, by algorithmically expanding training datasets. Beyond traditional data augmentation techniques, generative adversarial network (GAN) invented in 2014 in the computer vision community, provides a suite of novel approaches that can learn good data representations and generate highly realistic samples. Since 2017, there has been a growth of research into GANs for image augmentation or synthesis in agriculture for improved model performance. This paper presents an overview of the evolution of GAN architectures followed by a systematic review of their application to agriculture (https://github.com/Derekabc/GANs-Agriculture), involving various vision tasks for plant health, weeds, fruits, aquaculture, animal farming, plant phenotyping as well as postharvest detection of fruit defects. Challenges and opportunities of GANs are discussed for future research. | ['Yanbo Huang', 'Yuzhen Lu', 'Dong Chen', 'Ebenezer Olaniyi'] | 2022-04-10 | null | null | null | null | ['plant-phenotyping', 'image-augmentation'] | ['computer-vision', 'computer-vision'] | [ 7.87279129e-01 2.89067198e-02 -1.65769443e-01 -7.33432025e-02
-2.23537743e-01 -9.37954426e-01 3.24061096e-01 4.13415611e-01
7.04339966e-02 5.13655543e-01 -4.33574289e-01 -5.99504352e-01
2.34447956e-01 -9.96302724e-01 -7.88736224e-01 -1.02345908e+00
2.12747380e-01 1.64696127e-01 -3.41640025e-01 -2.57211000e-01
-9.02861729e-02 7.92951584e-01 -1.61780584e+00 1.56754419e-01
9.46035981e-01 1.05881166e+00 4.92912799e-01 9.69916940e-01
-2.54843235e-01 3.63835871e-01 -7.28049099e-01 -2.56201535e-01
1.79180309e-01 -7.42748857e-01 -3.90887499e-01 3.47439378e-01
2.06231996e-02 -2.07738742e-01 1.80006772e-01 9.83921647e-01
4.99804407e-01 -5.04492581e-01 5.87172329e-01 -1.31618881e+00
-9.77829218e-01 3.38947564e-01 -8.37415099e-01 -2.58437932e-01
-3.51602584e-01 4.12380219e-01 3.42621356e-01 -5.57747424e-01
3.73643458e-01 8.56752694e-01 6.04583502e-01 5.37348926e-01
-1.42819524e+00 -6.06579661e-01 -1.48512110e-01 -1.08981907e-01
-8.63278389e-01 -1.89174041e-01 5.60320854e-01 -3.10793012e-01
3.53852868e-01 4.09588516e-01 9.58424687e-01 1.01816404e+00
3.54518294e-01 9.13653195e-01 1.10868227e+00 -5.06983578e-01
2.02316537e-01 -1.68691918e-01 -3.41618359e-01 3.56188238e-01
4.39573586e-01 2.12316260e-01 -2.64931828e-01 1.24555327e-01
9.26510870e-01 5.81253730e-02 -2.60960460e-01 -3.43752772e-01
-1.01243079e+00 8.94025743e-01 6.22727215e-01 1.40628174e-01
-8.67686212e-01 8.38877708e-02 3.09437424e-01 1.40749335e-01
3.86743695e-01 6.71970725e-01 -6.88692868e-01 2.34138519e-01
-8.99434626e-01 1.79287285e-01 5.08508027e-01 1.01796341e+00
5.96350968e-01 6.54646993e-01 -2.05215756e-02 7.37034440e-01
1.65373281e-01 9.26630855e-01 2.55509198e-01 -9.24752295e-01
-2.99662441e-01 6.64898694e-01 1.40363807e-02 -1.02331650e+00
-1.65343136e-01 -4.34561700e-01 -1.31297374e+00 4.38628465e-01
3.40264618e-01 -1.07135683e-01 -1.43796635e+00 1.60880196e+00
4.93023753e-01 -1.91694312e-02 2.52354443e-01 6.02077961e-01
8.02352846e-01 8.65434408e-01 3.00320178e-01 -1.27939716e-01
1.39716303e+00 -5.33545017e-01 -7.68160701e-01 -4.41785634e-01
4.97476339e-01 -8.99786890e-01 7.65106201e-01 5.26449859e-01
-8.50143671e-01 -5.39912820e-01 -1.00396502e+00 2.51326025e-01
-7.57952273e-01 4.20115471e-01 9.32207406e-01 7.85226762e-01
-7.28515983e-01 3.11584979e-01 -9.85411048e-01 -6.09753728e-01
1.06322706e+00 3.42550814e-01 -3.62258911e-01 -3.26482803e-01
-5.92196286e-01 7.82309771e-01 3.54438394e-01 6.48348927e-01
-1.25911200e+00 -8.10597301e-01 -8.85440111e-01 -1.17944025e-01
1.35590568e-01 -5.55627704e-01 8.82397950e-01 -1.25885212e+00
-1.57500660e+00 1.09591460e+00 9.09169540e-02 -7.00934350e-01
1.36852965e-01 -2.47450009e-01 -7.71783739e-02 -1.60372451e-01
-6.53629750e-02 1.14374399e+00 8.73785257e-01 -1.53961492e+00
-5.31525671e-01 -5.93390882e-01 -1.90344229e-01 -1.64596364e-01
-1.51351750e-01 -1.19444638e-01 1.19053826e-01 -8.50604892e-01
5.59900515e-02 -1.02441096e+00 -4.80655611e-01 1.80546805e-01
-3.05110544e-01 4.99952495e-01 1.22537029e+00 -9.72114027e-01
3.95568669e-01 -1.94253898e+00 -4.52497303e-02 -1.51255518e-01
1.90572497e-02 9.89840209e-01 -4.38150167e-01 3.28898132e-01
-6.39180988e-02 2.66890883e-01 -5.95923662e-01 3.20772767e-01
-5.66832781e-01 2.69629776e-01 -2.35848606e-01 3.81779909e-01
8.94345582e-01 1.34458303e+00 -7.94929624e-01 -1.01405777e-01
4.90705311e-01 6.47269547e-01 -1.29203603e-01 4.47455406e-01
-4.23969567e-01 8.67807448e-01 -3.37733060e-01 1.26074564e+00
1.00922227e+00 9.82918544e-04 5.01400381e-02 -2.25662008e-01
2.24029049e-02 -6.11312270e-01 -6.58524036e-01 1.41128087e+00
-3.87572020e-01 6.19557381e-01 4.94114697e-01 -1.22863865e+00
1.18338752e+00 1.21215962e-01 4.48120773e-01 -5.63820899e-01
2.58659065e-01 1.01069614e-01 4.89556566e-02 -2.74132878e-01
1.24484494e-01 1.84102625e-01 2.42314130e-01 -6.51566908e-02
7.02101886e-02 -4.83558655e-01 1.15994163e-01 -2.78742999e-01
7.39417970e-01 4.79667157e-01 2.01647311e-01 -1.61350161e-01
2.58644521e-01 6.04885042e-01 5.13729930e-01 2.60138094e-01
-2.39163935e-01 6.89978421e-01 4.17791784e-01 -3.82193863e-01
-1.18166208e+00 -6.39947176e-01 1.24696214e-02 8.39976549e-01
-9.94121656e-02 1.55498788e-01 -1.03894734e+00 -4.27185595e-01
-4.94669303e-02 5.35876036e-01 -6.74754262e-01 -4.17053044e-01
-4.02661085e-01 -1.32580769e+00 7.75302052e-01 5.53342700e-01
7.76187181e-01 -1.39996219e+00 -9.65691149e-01 2.79058844e-01
3.24114636e-02 -1.17280817e+00 3.19174111e-01 7.06566751e-01
-9.21413660e-01 -1.04305494e+00 -1.09207237e+00 -8.54801357e-01
7.11630821e-01 2.42740929e-01 1.15465915e+00 3.55834253e-02
-8.00926864e-01 5.86069487e-02 -6.19154453e-01 -1.29461122e+00
-8.14836323e-01 8.32877383e-02 -4.83663142e-01 -3.72909874e-01
3.20724756e-01 -2.63892233e-01 -6.67009354e-01 1.95704758e-01
-1.33705223e+00 7.83611760e-02 9.36487198e-01 1.34545147e+00
8.81806850e-01 -4.09956686e-02 8.06965411e-01 -9.62936938e-01
2.38237292e-01 -3.71039480e-01 -1.01276326e+00 4.08469111e-01
-5.12493074e-01 -3.71190935e-01 4.88388330e-01 -4.46203977e-01
-8.96473289e-01 2.48358265e-01 -1.23914011e-01 -1.28838331e-01
-6.40331268e-01 6.99033976e-01 -4.24198329e-01 -2.33130693e-01
7.49601245e-01 2.33907789e-01 5.00919163e-01 -8.30095783e-02
3.58372897e-01 4.70001817e-01 5.30911446e-01 -1.51913300e-01
7.13677287e-01 2.76169449e-01 2.58758008e-01 -1.27694070e+00
-6.08527064e-01 -3.68645787e-02 -5.84642649e-01 -8.29937458e-02
8.64724338e-01 -8.34791481e-01 -3.56599987e-01 1.13187218e+00
-9.48740840e-01 -7.63313890e-01 -4.86735195e-01 2.14525387e-01
-2.81688541e-01 -2.83464864e-02 -2.84424216e-01 -7.33545482e-01
-6.68147206e-01 -1.13725305e+00 1.20405638e+00 4.55635160e-01
7.71865845e-02 -7.11935699e-01 -3.06859016e-01 4.02532339e-01
6.32710397e-01 9.64142382e-01 9.68173504e-01 -1.14139467e-01
-4.06353176e-01 -5.38553894e-01 -1.93238333e-01 6.99160576e-01
4.17430520e-01 4.39791799e-01 -1.16677058e+00 -1.39279008e-01
-1.70617491e-01 -5.73153853e-01 6.92763269e-01 8.85162652e-01
1.23440552e+00 9.45890043e-03 -2.36352175e-01 8.17145407e-01
1.31971467e+00 4.61889327e-01 8.32472622e-01 6.67300299e-02
7.20082581e-01 7.38027513e-01 1.00330341e+00 2.19791323e-01
-1.46540135e-01 1.06895030e-01 1.12727976e+00 -7.98385739e-01
-1.52276799e-01 2.12508254e-02 -7.49593193e-04 1.78557292e-01
3.91004753e-04 -6.60419762e-01 -8.11803758e-01 6.49395049e-01
-1.40075243e+00 -5.73879361e-01 -2.37738892e-01 2.01852632e+00
5.40651381e-01 -3.85444641e-01 -1.04596794e-01 3.10840100e-01
7.10921884e-01 -2.37583488e-01 -1.05590796e+00 -4.59769815e-01
-5.75408399e-01 4.52361465e-01 9.45815384e-01 -1.04847476e-01
-1.16631663e+00 1.38909197e+00 6.07908010e+00 5.96325099e-01
-1.45009136e+00 -3.44313383e-01 1.19529879e+00 5.59935093e-01
2.85803396e-02 -1.36863455e-01 -4.88743156e-01 1.08666658e-01
7.74753749e-01 4.59809512e-01 3.59187543e-01 6.86396599e-01
1.55486867e-01 -2.95289844e-01 -6.18364155e-01 6.00669622e-01
-1.90829948e-01 -1.32824171e+00 6.39774948e-02 1.23398066e-01
8.03663373e-01 -1.81768402e-01 1.88021794e-01 -2.67208591e-02
4.43964541e-01 -1.33668280e+00 2.91386396e-01 1.60759985e-02
8.25631499e-01 -8.01290452e-01 9.82886076e-01 2.30516225e-01
-9.02823210e-01 8.10334235e-02 -4.36775148e-01 3.80051881e-01
-1.81035116e-01 7.03604341e-01 -7.66943991e-01 5.80934525e-01
8.20335567e-01 5.84992945e-01 -6.83418155e-01 7.44482160e-01
-2.30770394e-01 9.75238740e-01 -2.50826120e-01 2.72918642e-01
2.01124609e-01 -4.30355012e-01 1.39542252e-01 8.62481475e-01
3.91488552e-01 4.28236872e-02 -1.95431709e-02 9.19888377e-01
1.51338093e-02 -1.09117396e-01 -7.56289959e-01 -7.19601631e-01
2.93111652e-01 1.49063110e+00 -1.21372604e+00 1.38644263e-01
-3.94385159e-02 1.03986096e+00 -2.40833282e-01 3.66274416e-01
-7.46683419e-01 -2.98445642e-01 6.46909475e-01 -9.19948518e-02
2.41876796e-01 -3.01786691e-01 -3.70068550e-01 -5.14197290e-01
-2.97797978e-01 -1.33843446e+00 4.22121510e-02 -6.56282604e-01
-9.10983562e-01 3.65669459e-01 -2.35155210e-01 -8.07471156e-01
-5.20066768e-02 -8.99708927e-01 -4.49211985e-01 9.12367582e-01
-1.56312335e+00 -1.80057216e+00 -8.33111942e-01 -3.46770487e-03
5.96057653e-01 -2.05907956e-01 1.27520978e+00 -1.94921009e-02
-5.53336084e-01 2.57034540e-01 2.32906625e-01 6.60617277e-03
4.41619158e-01 -9.90719974e-01 5.00210941e-01 8.66852164e-01
1.19625535e-02 5.66165224e-02 7.49390781e-01 -5.90456784e-01
-1.66449416e+00 -1.41994715e+00 1.53729677e-01 -1.26445787e-02
1.92542851e-01 -2.35479489e-01 -8.86706233e-01 5.12405396e-01
2.70922393e-01 1.32913291e-01 7.40728736e-01 -5.56740403e-01
2.65987903e-01 -1.79114223e-01 -1.45263171e+00 3.61988664e-01
4.88935620e-01 7.09389225e-02 4.27331358e-01 3.19411427e-01
5.06419003e-01 -4.41628873e-01 -8.14152479e-01 6.67080104e-01
5.28719246e-01 -5.77943802e-01 9.71608698e-01 -5.97709000e-01
5.88340402e-01 -3.14807266e-01 -5.41589744e-02 -1.47416091e+00
-1.35622308e-01 -4.20192480e-01 -6.71193236e-04 1.42343974e+00
1.96086302e-01 -5.17946661e-01 1.01701427e+00 5.74526563e-02
-5.52227721e-02 -7.91456699e-01 -1.93520010e-01 -5.92109442e-01
2.60808736e-01 2.48174481e-02 8.07437599e-01 8.08054149e-01
-8.87451053e-01 4.02951725e-02 -2.63536632e-01 3.16524714e-01
4.64262038e-01 2.95714028e-02 9.54064608e-01 -1.14899588e+00
1.06742054e-01 -3.35041434e-01 -6.18149638e-01 -4.60684240e-01
-4.29069484e-03 -5.42036653e-01 5.04808649e-02 -1.69972777e+00
-1.71281606e-01 -4.53274637e-01 2.07644805e-01 6.05111122e-01
-2.85938382e-01 5.43441892e-01 1.79121201e-03 -1.86836660e-01
4.63665903e-01 3.53859335e-01 1.46326673e+00 -4.27538484e-01
-2.50430852e-01 8.94787014e-02 -8.66075814e-01 5.16509593e-01
1.32747185e+00 -2.97857136e-01 -5.03961623e-01 -6.38734281e-01
3.36537696e-02 -2.33907178e-01 5.24152279e-01 -9.32732522e-01
-4.07583058e-01 -3.71561706e-01 8.24068069e-01 -5.46128213e-01
1.31935209e-01 -9.84624803e-01 3.94962668e-01 8.57147217e-01
-1.88784063e-01 5.49279079e-02 5.81722021e-01 3.94027770e-01
-3.35032016e-01 -2.04731971e-01 1.03507149e+00 -3.07671577e-01
-6.65723920e-01 3.73220146e-01 -4.98574406e-01 -3.20003599e-01
1.40819776e+00 -3.67760330e-01 -1.49213880e-01 -1.61450624e-01
-4.97326612e-01 2.85663545e-01 4.52619731e-01 3.36313426e-01
5.23479879e-01 -9.34299171e-01 -1.02621531e+00 3.64557236e-01
1.30028009e-01 6.02363110e-01 2.34978661e-01 5.30943692e-01
-1.07823098e+00 1.33675069e-01 -6.40070140e-01 -8.88567924e-01
-1.30863225e+00 4.93403941e-01 1.06694773e-01 -2.27571130e-01
-6.92673847e-02 7.75439501e-01 2.82855004e-01 -3.79234433e-01
-8.74167085e-02 -4.56721097e-01 -3.28555673e-01 4.87844013e-02
2.78933436e-01 2.59587556e-01 3.21994573e-01 -3.81759197e-01
2.40771417e-02 3.37724626e-01 1.91623881e-01 4.79901075e-01
1.53750396e+00 2.09659413e-01 -6.16269931e-02 1.63384646e-01
6.42388284e-01 -5.78046739e-01 -1.22179031e+00 3.55625063e-01
-1.71306133e-01 -3.51636708e-01 1.97721973e-01 -1.03780735e+00
-1.44187379e+00 1.09037423e+00 1.08276296e+00 3.21862698e-01
1.51522541e+00 -3.91357362e-01 6.24331474e-01 1.87340260e-01
2.19084114e-01 -6.67505145e-01 -1.82476342e-02 2.18997344e-01
9.89063203e-01 -1.51159441e+00 -1.98050648e-01 -5.16186118e-01
-6.79641962e-01 9.70437765e-01 6.12979591e-01 4.28798981e-02
4.75675493e-01 5.93992233e-01 4.76936281e-01 -8.34403262e-02
-2.02502415e-01 -2.49170586e-01 -1.88449502e-01 1.34389210e+00
6.80681288e-01 1.65893957e-01 9.56826657e-02 1.90506890e-01
7.86131546e-02 8.20455775e-02 3.27243477e-01 1.32949555e+00
-2.41249949e-01 -1.21660960e+00 -5.20715535e-01 6.01573110e-01
-5.77193856e-01 1.24294115e-02 -5.92202604e-01 6.89139605e-01
2.49633431e-01 9.46213186e-01 -1.31544501e-01 1.02342248e-01
1.08854316e-01 -1.39154181e-01 5.18186212e-01 -5.28013170e-01
-6.10288978e-01 7.09957778e-02 -3.20980996e-01 -2.23029271e-01
-5.69892347e-01 -5.72115123e-01 -7.94252992e-01 -1.64020494e-01
-6.38648331e-01 -1.85842067e-01 1.24105322e+00 5.33266604e-01
4.77675498e-01 8.31212997e-01 5.85671365e-01 -8.44383776e-01
-2.18387693e-01 -1.00260234e+00 -5.03187180e-01 -6.72797784e-02
1.58248678e-01 -3.37389350e-01 6.92708641e-02 5.81485629e-01] | [9.158320426940918, -1.5244324207305908] |
4a00d991-63da-4bcc-ba06-3e1e5441556d | deep-reinforcement-learning-for-unsupervised | 1801.00054 | null | http://arxiv.org/abs/1801.00054v3 | http://arxiv.org/pdf/1801.00054v3.pdf | Deep Reinforcement Learning for Unsupervised Video Summarization with Diversity-Representativeness Reward | Video summarization aims to facilitate large-scale video browsing by
producing short, concise summaries that are diverse and representative of
original videos. In this paper, we formulate video summarization as a
sequential decision-making process and develop a deep summarization network
(DSN) to summarize videos. DSN predicts for each video frame a probability,
which indicates how likely a frame is selected, and then takes actions based on
the probability distributions to select frames, forming video summaries. To
train our DSN, we propose an end-to-end, reinforcement learning-based
framework, where we design a novel reward function that jointly accounts for
diversity and representativeness of generated summaries and does not rely on
labels or user interactions at all. During training, the reward function judges
how diverse and representative the generated summaries are, while DSN strives
for earning higher rewards by learning to produce more diverse and more
representative summaries. Since labels are not required, our method can be
fully unsupervised. Extensive experiments on two benchmark datasets show that
our unsupervised method not only outperforms other state-of-the-art
unsupervised methods, but also is comparable to or even superior than most of
published supervised approaches. | ['Tao Xiang', 'Yu Qiao', 'Kaiyang Zhou'] | 2017-12-29 | null | null | null | null | ['unsupervised-video-summarization', 'supervised-video-summarization'] | ['computer-vision', 'computer-vision'] | [ 2.64177948e-01 -2.60503348e-02 -4.81180340e-01 -5.05741179e-01
-8.32130671e-01 -3.29101294e-01 5.47733426e-01 2.12447599e-01
-3.25403899e-01 9.30876970e-01 7.90246844e-01 1.57631665e-01
2.10007787e-01 -4.88884062e-01 -9.07322645e-01 -5.87997556e-01
-1.13771804e-01 2.41524801e-01 2.87133783e-01 3.51773709e-01
3.31438750e-01 1.70235693e-01 -1.58507562e+00 3.32040876e-01
1.03777230e+00 9.26092625e-01 4.92504239e-01 9.23906505e-01
2.88212061e-01 1.46472466e+00 -8.21884811e-01 -3.99044529e-02
-2.06803292e-01 -8.72079134e-01 -6.79412961e-01 5.23010910e-01
4.74350810e-01 -1.00819921e+00 -6.74996912e-01 9.25367951e-01
5.61657667e-01 5.95716596e-01 7.37953067e-01 -1.09038329e+00
-6.84104919e-01 8.80329311e-01 -4.44495052e-01 2.44369909e-01
6.02419257e-01 2.34391704e-01 1.25671721e+00 -5.21818757e-01
7.77179122e-01 1.09241581e+00 2.18121290e-01 6.27561927e-01
-1.08880758e+00 -2.78087229e-01 4.26694006e-01 1.56652838e-01
-7.87632883e-01 -5.84645152e-01 6.64042413e-01 -2.17767417e-01
8.23465765e-01 2.19545886e-01 7.33602047e-01 1.15993714e+00
2.98801512e-02 1.47882211e+00 2.86129296e-01 -2.13979296e-02
5.90700686e-01 -2.84982055e-01 -5.82435727e-02 6.58603609e-01
1.66326225e-01 -4.70686108e-01 -7.15138078e-01 -1.27424002e-01
7.14518011e-01 2.75553793e-01 -3.76796007e-01 -3.71197283e-01
-1.45371807e+00 6.96596742e-01 2.94109762e-01 -9.71020013e-02
-8.85313690e-01 4.32992518e-01 6.91874981e-01 -8.45824252e-04
3.18185359e-01 3.48280966e-01 1.02795668e-01 -4.00354147e-01
-1.26582623e+00 4.41947073e-01 5.65252423e-01 9.72464800e-01
6.40399992e-01 2.44049430e-01 -8.12881052e-01 8.47482085e-01
-1.36490248e-03 4.04387027e-01 5.85334718e-01 -1.58631849e+00
3.34275007e-01 3.92873019e-01 2.52090544e-01 -8.39652538e-01
-2.36774497e-02 -2.10340142e-01 -8.31151426e-01 -1.49594352e-01
-2.74343133e-01 -3.40982974e-01 -7.96806276e-01 1.82976818e+00
-1.65481210e-01 2.57988036e-01 1.60604060e-01 1.04642320e+00
1.01315570e+00 1.14440835e+00 8.84231403e-02 -5.96622884e-01
8.31644893e-01 -1.33195305e+00 -6.74644172e-01 -3.30122048e-03
2.65725017e-01 -2.67547339e-01 9.88830149e-01 3.51200223e-01
-1.41402376e+00 -5.66633344e-01 -9.62337792e-01 7.01621175e-02
5.36172748e-01 3.83232117e-01 2.52773583e-01 -5.89651018e-02
-1.15785062e+00 8.94023120e-01 -9.68485832e-01 -4.80311513e-01
6.62611246e-01 1.14029139e-01 -7.66932145e-02 8.81103724e-02
-8.13438952e-01 3.08230519e-01 5.26824713e-01 -3.64055276e-01
-1.30695713e+00 -2.51830906e-01 -9.37732220e-01 4.62087780e-01
6.16227686e-01 -9.89885092e-01 1.57965839e+00 -1.00098741e+00
-1.72790694e+00 2.94212371e-01 -3.69863480e-01 -7.26974249e-01
4.80124593e-01 -3.46693575e-01 -8.81279111e-02 7.61826396e-01
4.08796966e-01 1.09639239e+00 8.52874219e-01 -1.08843994e+00
-9.08030629e-01 1.92233250e-01 1.13934316e-01 5.14799774e-01
-4.81171817e-01 -1.25030607e-01 -6.53264463e-01 -7.30299115e-01
-4.84901160e-01 -6.73826694e-01 -3.32620919e-01 -3.36850822e-01
-7.18014836e-01 -4.01215613e-01 5.39618075e-01 -7.03153431e-01
1.51186919e+00 -2.10960960e+00 3.96199703e-01 -2.10415483e-01
4.07153606e-01 2.50685155e-01 -1.74416244e-01 5.15889645e-01
4.02873755e-01 1.31179830e-02 -1.98782131e-01 -3.79523396e-01
1.15710355e-01 -1.61550213e-02 -3.00904423e-01 5.87621704e-02
2.46709853e-01 6.79643273e-01 -1.22362709e+00 -5.60724437e-01
1.27420843e-01 9.38796028e-02 -7.88066447e-01 5.77603817e-01
-5.83550930e-01 4.39086974e-01 -5.76780498e-01 4.34610486e-01
2.85157382e-01 -5.35959363e-01 1.73700556e-01 -1.54879391e-01
-5.63537516e-02 3.28739971e-01 -8.80006611e-01 1.73177218e+00
-1.80997908e-01 7.28894532e-01 -4.27191645e-01 -1.12705290e+00
7.40046620e-01 2.53743023e-01 4.38098013e-01 -5.58951318e-01
5.76989278e-02 5.07865697e-02 -5.02127647e-01 -7.13587821e-01
8.82390380e-01 3.73749286e-01 -1.67068258e-01 7.14649916e-01
1.65555343e-01 1.22208141e-01 8.20285141e-01 7.92276621e-01
1.27250671e+00 2.93023705e-01 3.07606071e-01 9.48272794e-02
3.40451688e-01 -1.85284078e-01 8.68417323e-01 9.95541692e-01
-9.55748409e-02 6.79957867e-01 1.06358254e+00 -1.23623222e-01
-1.09497404e+00 -1.07978475e+00 6.07139111e-01 1.23360944e+00
2.88606554e-01 -6.74747705e-01 -8.93128216e-01 -8.15099299e-01
-2.01514244e-01 8.52642238e-01 -2.85774827e-01 -3.44623059e-01
-4.47115839e-01 -2.42314767e-02 9.09601822e-02 6.09244525e-01
5.53132534e-01 -1.49755287e+00 -8.71810436e-01 2.71769345e-01
-3.85983825e-01 -1.08423865e+00 -9.22644019e-01 -2.11447135e-01
-8.66941392e-01 -7.30843008e-01 -1.04828537e+00 -8.40493679e-01
6.74522519e-01 4.65550095e-01 1.12007034e+00 -1.78298041e-01
2.01068684e-01 4.11598951e-01 -5.06745875e-01 -5.22910133e-02
-4.38343555e-01 1.88797385e-01 6.12845309e-02 3.34841497e-02
-1.36870528e-02 -5.32836020e-01 -8.03473651e-01 -5.81856258e-02
-9.99458015e-01 2.16383338e-01 5.93044579e-01 6.70516074e-01
7.62547135e-01 -8.03423859e-03 9.61884975e-01 -9.83467698e-01
8.10743093e-01 -4.67862070e-01 -1.31043687e-01 2.63788015e-01
-5.69718555e-02 6.39183596e-02 1.06421602e+00 -3.47769499e-01
-1.00112331e+00 -2.28799786e-02 1.22741312e-01 -6.51813507e-01
-1.08115233e-01 5.75723290e-01 -7.28183091e-02 7.73425102e-01
3.23173136e-01 5.01699209e-01 -5.13218716e-02 -1.47382081e-01
3.18287939e-01 6.14992678e-01 7.85091102e-01 -2.92684317e-01
2.76370287e-01 2.99303561e-01 -4.96839672e-01 -7.51714885e-01
-8.07668746e-01 -3.17525893e-01 -1.91000506e-01 -4.37860906e-01
8.29707921e-01 -1.12443519e+00 -7.03949392e-01 3.44197512e-01
-9.42009687e-01 -4.79268998e-01 -4.10059512e-01 5.54628670e-01
-9.94665623e-01 6.13664508e-01 -6.13482177e-01 -6.81841016e-01
-4.31478262e-01 -1.16327786e+00 9.55767214e-01 7.18620420e-01
-4.09741044e-01 -6.86200976e-01 -1.05083592e-01 1.02353372e-01
2.99817532e-01 2.87914485e-01 6.08385205e-01 -6.21787727e-01
-6.73939705e-01 -3.05017736e-02 -5.59861250e-02 5.77081382e-01
2.23809063e-01 4.32135642e-01 -4.08225626e-01 -2.80643016e-01
-4.56626028e-01 -6.13130867e-01 1.26805663e+00 7.73909986e-01
1.47616315e+00 -7.95598567e-01 -5.93234338e-02 3.72871399e-01
1.10360324e+00 1.16310611e-01 5.79563320e-01 2.20030025e-01
6.61117792e-01 3.53334516e-01 8.34460974e-01 8.80249083e-01
5.99504173e-01 2.73632795e-01 4.54902798e-01 1.37480274e-01
-2.84257699e-02 -5.26167989e-01 7.00659454e-01 7.05030739e-01
4.84086946e-02 -6.24119222e-01 -2.50459462e-01 5.80762148e-01
-2.32910562e+00 -1.40042150e+00 3.84354681e-01 2.13750720e+00
8.82539809e-01 3.20424169e-01 7.01697886e-01 -2.90871501e-01
9.03613210e-01 5.88155329e-01 -9.61962819e-01 -3.09220523e-01
-3.73299047e-02 -2.32480481e-01 9.25146416e-02 2.46007338e-01
-1.13426602e+00 1.05590165e+00 5.93024206e+00 7.15300739e-01
-1.02095842e+00 -4.57653105e-01 8.63481939e-01 -5.30794084e-01
-4.74592865e-01 -1.84162304e-01 -3.39626580e-01 7.26055384e-01
7.67204404e-01 -4.67737108e-01 4.53326464e-01 8.78793359e-01
5.83921671e-01 -3.08003902e-01 -1.21247613e+00 9.08197880e-01
2.51756191e-01 -1.64913690e+00 5.40498972e-01 -3.36108446e-01
1.10107338e+00 -6.96831495e-02 -5.78479022e-02 3.02418083e-01
5.17699838e-01 -5.89524984e-01 8.61041546e-01 7.64468074e-01
5.45056939e-01 -9.63941157e-01 4.93818671e-01 4.23966676e-01
-9.16828394e-01 -1.63917437e-01 -5.14135480e-01 6.56455979e-02
4.03028101e-01 5.31239629e-01 -5.48369706e-01 3.88639539e-01
5.19599080e-01 1.27057052e+00 -3.97451639e-01 1.17817998e+00
-4.76198316e-01 7.53911138e-01 1.67228326e-01 -3.36975604e-01
3.58235419e-01 -2.03828126e-01 6.75659835e-01 1.39526570e+00
3.84066373e-01 1.35916114e-01 5.79674423e-01 6.10736907e-01
-5.41031063e-01 2.97725368e-02 -3.56041998e-01 -4.07180160e-01
6.61341250e-01 1.19982755e+00 -6.93697512e-01 -9.70355511e-01
-1.41970515e-01 1.15748918e+00 3.01543176e-01 5.10315180e-01
-9.38239455e-01 -5.70522904e-01 2.79253036e-01 -1.23322792e-01
6.44085884e-01 -1.71791296e-03 7.00643212e-02 -1.33135521e+00
9.39496756e-02 -8.21683347e-01 3.46823275e-01 -9.31376159e-01
-1.01528502e+00 6.48203194e-01 -8.62926170e-02 -1.41303265e+00
-5.49352109e-01 1.57591477e-01 -1.05692267e+00 2.53603965e-01
-1.41431177e+00 -6.50791168e-01 -3.98496836e-01 2.21179903e-01
9.63241816e-01 -3.72551501e-01 3.53549272e-01 1.18556246e-02
-9.38992918e-01 3.68023694e-01 2.52870262e-01 -5.32035679e-02
8.92250836e-01 -1.31316674e+00 2.16649905e-01 9.26854730e-01
-1.66461796e-01 3.40019464e-01 8.44412863e-01 -6.42786741e-01
-1.08314455e+00 -1.19233084e+00 6.33992195e-01 2.14041010e-01
3.86093944e-01 8.80536065e-02 -7.46162117e-01 5.22038102e-01
4.09161389e-01 -2.68040419e-01 6.53741241e-01 -3.66722137e-01
-4.41751033e-02 -1.46282837e-01 -7.72509336e-01 7.15832591e-01
1.03949392e+00 -1.70702115e-01 -5.38858950e-01 3.23345244e-01
8.54787350e-01 -1.85303763e-01 -5.14143229e-01 1.83773249e-01
4.32773799e-01 -1.25122094e+00 6.57073736e-01 -5.66531897e-01
1.07999969e+00 -1.71762407e-01 3.62252370e-02 -1.54191351e+00
-4.77285266e-01 -8.23851526e-01 -4.86343980e-01 1.40285754e+00
1.78955495e-01 -8.70389268e-02 7.38102138e-01 3.61807168e-01
-4.03408021e-01 -7.96103001e-01 -4.39809918e-01 -5.75772703e-01
-6.32523894e-01 2.11977974e-01 4.72717285e-01 4.93397981e-01
8.13951120e-02 4.71473187e-01 -6.34026885e-01 -1.03819817e-01
6.16692364e-01 3.34575057e-01 8.42462838e-01 -9.15066779e-01
-2.32555896e-01 -6.45587862e-01 -1.56400278e-01 -1.52780259e+00
2.98036665e-01 -5.90268672e-01 2.26185024e-01 -2.10122895e+00
7.15598404e-01 1.58162177e-01 -2.51059264e-01 2.53127575e-01
-4.60174978e-01 6.69718534e-02 2.26591900e-01 3.57725054e-01
-1.68418550e+00 9.24778461e-01 1.19585681e+00 -1.48759872e-01
-3.88318986e-01 1.02683567e-01 -9.70618904e-01 7.67482102e-01
6.66326642e-01 -1.71673283e-01 -5.32983601e-01 -4.61281687e-01
-4.65903943e-03 5.11393785e-01 4.63855453e-02 -1.09178567e+00
2.89416015e-01 -1.84250638e-01 3.76846075e-01 -9.14341450e-01
-2.64956560e-02 -1.69342399e-01 -2.28464961e-01 3.14817965e-01
-8.58038425e-01 -1.77728847e-01 -3.18693101e-01 7.35745430e-01
-2.94418871e-01 -2.38985077e-01 6.40781820e-01 -1.34649694e-01
-7.18338430e-01 5.24401784e-01 -6.72456205e-01 8.19601119e-02
9.38818336e-01 -2.00863764e-01 -6.06164522e-02 -9.48448420e-01
-6.51001215e-01 7.34276056e-01 6.25521004e-01 2.35562757e-01
6.76954985e-01 -1.51007450e+00 -9.41126406e-01 -2.43212849e-01
1.53923631e-02 2.03246012e-01 4.99635190e-01 3.74163598e-01
-6.09481514e-01 2.02338457e-01 -2.47201383e-01 -5.99612713e-01
-1.09055424e+00 4.71501440e-01 -2.01949015e-01 -2.21164316e-01
-6.58940315e-01 6.30533278e-01 2.19999447e-01 1.64867148e-01
5.95537066e-01 -2.89010674e-01 -5.00661433e-01 1.55994385e-01
6.86862826e-01 6.48598611e-01 -4.98604238e-01 -4.38433349e-01
-4.43870351e-02 7.80155733e-02 -4.10701156e-01 9.28445756e-02
1.56260777e+00 -2.88316041e-01 1.23064540e-01 4.29688483e-01
1.11452305e+00 -1.69825047e-01 -1.89280331e+00 -2.93130785e-01
-3.55854034e-01 -3.69024634e-01 -2.21106946e-01 -5.45046210e-01
-9.89640296e-01 5.80868781e-01 -1.53783172e-01 3.68728369e-01
1.34602559e+00 -1.58598181e-02 1.08563066e+00 4.75911736e-01
-2.09948868e-02 -1.25123262e+00 6.76767886e-01 3.75315249e-01
7.65528738e-01 -1.11914313e+00 1.82948932e-01 3.04637533e-02
-1.36794281e+00 1.07877600e+00 6.49127662e-01 -4.71774817e-01
-8.36818814e-02 -1.98276162e-01 -4.47766334e-01 8.60499665e-02
-1.00258577e+00 6.78165331e-02 8.46369117e-02 4.13437426e-01
3.07445258e-01 5.87356277e-04 -2.47838810e-01 5.63242376e-01
-3.50402817e-02 2.04402432e-01 8.50956917e-01 8.06172490e-01
-8.63658130e-01 -7.27547884e-01 3.07397805e-02 7.65097797e-01
-3.12895238e-01 2.27726728e-01 -2.56411970e-01 1.16819382e-01
-4.25863326e-01 8.10360610e-01 2.20163271e-01 -2.64476895e-01
2.22315803e-01 -4.05932277e-01 2.39382505e-01 -6.93432927e-01
-1.78241566e-01 3.38384122e-01 8.92478153e-02 -5.63791454e-01
-5.43226421e-01 -7.73782313e-01 -1.41488886e+00 -3.92957270e-01
-4.18068282e-02 3.55747014e-01 1.67122260e-01 6.94223166e-01
5.54487526e-01 8.92438412e-01 1.11336601e+00 -1.28547370e+00
-6.78696871e-01 -8.39173675e-01 -5.47085524e-01 4.60508943e-01
4.67812479e-01 -2.82418609e-01 -1.50242463e-01 2.47715443e-01] | [10.477422714233398, 0.4199371337890625] |
2fd9da36-03fd-45ee-bc9e-657a1ba1232b | fair-lending-needs-explainable-models-for | 1809.04684 | null | http://arxiv.org/abs/1809.04684v1 | http://arxiv.org/pdf/1809.04684v1.pdf | Fair lending needs explainable models for responsible recommendation | The financial services industry has unique explainability and fairness
challenges arising from compliance and ethical considerations in credit
decisioning. These challenges complicate the use of model machine learning and
artificial intelligence methods in business decision processes. | ['Jiahao Chen'] | 2018-09-12 | null | null | null | null | ['explainable-models'] | ['computer-vision'] | [-1.34220526e-01 5.27702868e-01 -7.34305441e-01 -8.81774187e-01
3.37942958e-01 -2.18870223e-01 5.50934255e-01 -4.88938950e-02
-3.45929444e-01 6.71558321e-01 1.92249656e-01 -1.10442781e+00
-3.15113217e-01 -5.85229099e-01 1.10955015e-01 -1.53551757e-01
8.59487057e-02 7.50025570e-01 -8.01902533e-01 1.06612593e-01
9.01539564e-01 3.84205401e-01 -8.43939066e-01 5.41281700e-01
6.29157782e-01 8.59486699e-01 -8.34643185e-01 3.42055202e-01
-2.28755906e-01 1.83612347e+00 -3.18542302e-01 -1.17677414e+00
6.13907099e-01 -1.24537304e-01 -1.06235290e+00 9.64182243e-02
-4.08227295e-01 -7.76023507e-01 -1.86900571e-02 7.60142028e-01
-5.46335243e-03 1.52116969e-01 1.17138684e+00 -1.87607181e+00
-9.85666156e-01 8.72575462e-01 -4.50811774e-01 3.39443952e-01
2.70264924e-01 2.87152082e-01 1.40234470e+00 -6.67166710e-01
3.05618942e-01 1.18713915e+00 5.08271754e-01 9.20243204e-01
-1.22267377e+00 -7.99430311e-01 1.35619223e-01 5.32964110e-01
-6.68847382e-01 -5.98504782e-01 5.51666617e-01 -9.39550877e-01
1.33780968e+00 2.02980354e-01 6.30360246e-01 8.68101180e-01
4.00411636e-01 4.60006922e-01 1.28131211e+00 -5.62021494e-01
5.39129496e-01 2.90636569e-01 2.50771701e-01 -7.68577009e-02
9.17966366e-01 8.23374927e-01 -3.71817023e-01 -4.91922498e-01
6.80605710e-01 2.14303389e-01 3.75087082e-01 -1.93546027e-01
-6.70750916e-01 1.46888721e+00 2.72837937e-01 -6.28962740e-02
-4.01560843e-01 1.71101764e-01 2.70604372e-01 7.40433156e-01
-2.33144481e-02 8.95658791e-01 -5.22151172e-01 -2.71431446e-01
-6.60051167e-01 6.54766977e-01 1.40608823e+00 7.26494193e-01
2.97474742e-01 2.05579489e-01 3.53438675e-01 2.06494376e-01
9.84746158e-01 -9.27723199e-02 1.13942362e-02 -1.55249977e+00
5.85511863e-01 3.82168531e-01 4.97959018e-01 -1.09990275e+00
-1.19347140e-01 1.46821782e-01 -5.23965597e-01 9.05211151e-01
5.00526845e-01 -1.73762649e-01 -3.21961731e-01 6.33072853e-01
-1.84552103e-01 -6.20047033e-01 2.61463314e-01 7.98219800e-01
2.24679172e-01 1.28436148e-01 6.71681583e-01 -4.77142543e-01
8.50851655e-01 -8.13910007e-01 -9.94261622e-01 -5.74334204e-01
5.64596951e-01 -3.35564107e-01 4.38211471e-01 5.82372725e-01
-1.09362042e+00 -3.94724868e-02 -8.59876573e-01 7.35120475e-02
1.83314458e-01 -8.05612803e-01 1.11843264e+00 9.00161028e-01
-2.29197055e-01 9.45955336e-01 -5.69858730e-01 1.78382397e-01
1.03763342e+00 4.83220607e-01 -2.07533076e-01 1.72960997e-01
-9.81284380e-01 1.49231172e+00 2.61125416e-01 1.47322804e-01
-1.77887991e-01 -3.40884149e-01 -4.93672431e-01 3.59582931e-01
3.12561691e-01 -6.86101794e-01 1.37260902e+00 -1.33547866e+00
-1.17951238e+00 7.73583949e-01 3.82547110e-01 -8.31729949e-01
9.40895617e-01 -5.00671156e-02 -3.05935472e-01 -3.11965853e-01
-7.59330615e-02 3.80417369e-02 5.15849471e-01 -9.62157309e-01
-8.84344041e-01 -3.52203548e-01 -2.56805450e-01 1.12342082e-01
2.19654709e-01 5.96483946e-01 9.30081308e-01 -4.78964597e-01
2.62373328e-01 -4.21400785e-01 -6.50129974e-01 -3.19092751e-01
1.55561864e-01 -8.27468932e-02 3.97660643e-01 -8.07735860e-01
1.04314196e+00 -1.68700886e+00 -5.07247031e-01 4.89371270e-01
3.55744362e-01 -1.64538234e-01 5.80794036e-01 2.78981179e-01
-1.48781717e-01 1.13997519e+00 3.56478393e-01 1.81677401e-01
5.78721046e-01 4.16178495e-01 -1.51990533e-01 2.72160083e-01
4.29349631e-01 9.31800306e-01 -4.20597881e-01 -5.28700948e-01
3.54984075e-01 -3.16464454e-01 -5.85330367e-01 4.32703823e-01
2.05827698e-01 3.48268896e-01 -7.63393402e-01 9.35300350e-01
5.51963866e-01 8.08418840e-02 5.00939667e-01 4.86577719e-01
-4.46893796e-02 5.64841449e-01 -9.43894684e-01 5.76036572e-01
3.64728458e-02 3.52154702e-01 1.72838703e-01 -1.13671660e+00
8.34417820e-01 4.67500120e-01 2.86236376e-01 -7.35942066e-01
3.16555858e-01 2.10459813e-01 6.28995776e-01 -6.62020326e-01
4.39907879e-01 -1.22358060e+00 2.09644228e-01 1.04583287e+00
-5.09218097e-01 -5.39938569e-01 -2.91703880e-01 -3.74914467e-01
5.22010148e-01 1.24497458e-01 8.79911542e-01 -4.63674247e-01
-4.38941196e-02 6.60485476e-02 1.19412363e+00 4.75389361e-01
-8.30889881e-01 3.77050400e-01 8.55541646e-01 -1.24830449e+00
-8.93154740e-01 -3.73541862e-01 -2.44728044e-01 8.32351327e-01
-3.31149966e-01 2.48955831e-01 -3.94873649e-01 -9.66811776e-01
8.05737495e-01 1.38756919e+00 -5.81646323e-01 -6.95571676e-02
-3.24434668e-01 -5.22854805e-01 1.10451311e-01 7.17894435e-01
-1.25363350e-01 -8.80435586e-01 -9.18494284e-01 5.95611155e-01
3.05896103e-01 -9.07507598e-01 2.09786266e-01 5.79657257e-01
-9.84540164e-01 -1.25532377e+00 2.11682782e-01 3.25548928e-03
2.68641859e-01 -2.80858576e-01 1.26779377e+00 1.12933688e-01
-1.08666562e-01 -9.97064263e-02 -2.28658497e-01 -1.04131460e+00
-1.03530550e+00 -3.37279052e-01 -1.61747202e-01 -3.79596621e-01
1.18423796e+00 -2.24764064e-01 -1.38423860e-01 5.48587739e-01
-3.54273140e-01 -2.10003346e-01 7.68462867e-02 9.34955180e-01
7.84499687e-04 2.70806968e-01 9.20633018e-01 -1.08985746e+00
1.06724918e+00 -7.72028387e-01 -4.78476852e-01 2.75171846e-01
-1.80801070e+00 -2.59693027e-01 7.11813644e-02 8.73394869e-03
-1.21817267e+00 -4.83370125e-01 5.76415956e-01 1.55352131e-01
-2.60115981e-01 5.93337595e-01 -2.22029105e-01 -1.57073513e-01
6.14120245e-01 -9.96138096e-01 3.09442192e-01 -1.36892334e-01
-4.30747010e-02 9.17764008e-01 -1.92258164e-01 -4.72852588e-01
4.04543102e-01 1.92782711e-02 -7.14860633e-02 6.77273124e-02
-4.45139766e-01 1.26259118e-01 -5.69761336e-01 -1.73825949e-01
7.08588064e-01 -5.05459547e-01 -9.19515908e-01 1.48796394e-01
-8.00646305e-01 7.74858966e-02 -5.45410395e-01 6.84138775e-01
-6.90157652e-01 1.54260144e-01 -6.99785411e-01 -1.33767450e+00
-1.75700173e-01 -8.75918090e-01 -3.80494446e-01 3.28545541e-01
-1.14881384e+00 -1.07250309e+00 -4.00219321e-01 1.26150429e+00
1.85802877e-01 5.97972274e-01 1.20774043e+00 -1.22800994e+00
-3.48640591e-01 -6.67764723e-01 -6.38838783e-02 5.31207502e-01
1.69706881e-01 2.78322339e-01 -7.69902885e-01 8.05956349e-02
5.46477199e-01 -4.20169622e-01 2.62104988e-01 4.48774517e-01
6.07917726e-01 -7.88630128e-01 2.07513720e-01 2.36064106e-01
1.18213439e+00 7.52745152e-01 3.84495676e-01 9.05373693e-01
4.00299877e-01 1.57130480e+00 8.74902427e-01 9.31979358e-01
5.36258101e-01 1.66272968e-01 2.13504389e-01 2.79439718e-01
9.00654197e-01 -2.18070727e-02 -4.31386800e-03 6.50534406e-02
-6.75906718e-01 1.33534759e-01 -1.49473131e+00 1.21042870e-01
-2.31586623e+00 -1.34159744e+00 -2.70892531e-01 1.79890728e+00
4.84863669e-01 3.93182963e-01 1.29119875e-02 4.28521931e-01
4.63235587e-01 -1.84963658e-01 -5.55744052e-01 -1.61404455e+00
-8.00682325e-03 -1.47384375e-01 5.36441803e-01 4.55289871e-01
-8.66088510e-01 5.71918845e-01 8.04514790e+00 -5.68184368e-02
-1.20207787e-01 -1.64706126e-01 1.49166405e+00 -3.60209137e-01
-7.49660850e-01 2.62607664e-01 -1.99304983e-01 2.35464782e-01
1.04877675e+00 -6.63266718e-01 4.72860307e-01 1.06134713e+00
6.17015481e-01 -1.25406496e-02 -1.61676288e+00 4.93958622e-01
-4.27671731e-01 -1.20336831e+00 -1.34150967e-01 4.54604089e-01
8.76426876e-01 -5.98423123e-01 -2.07557087e-03 9.54866037e-03
9.32481110e-01 -1.98773968e+00 9.73535240e-01 4.23616111e-01
-1.89238228e-02 -9.43168998e-01 1.20206261e+00 5.58146276e-02
-2.51976252e-01 -7.65142500e-01 -4.33250040e-01 -1.08202696e+00
1.58435449e-01 1.13442197e-01 -7.89948940e-01 2.13497162e-01
6.93418980e-01 6.56771302e-01 1.99353412e-01 3.02880257e-01
-2.36951202e-01 5.90229809e-01 4.64475363e-01 7.17650354e-02
1.79808289e-01 -7.48576760e-01 -1.52279705e-01 1.01047230e+00
-1.76420480e-01 3.13452840e-01 -3.81002992e-01 1.27569258e+00
2.40723416e-01 5.93256950e-03 -5.78753114e-01 -3.14586282e-01
2.17224270e-01 7.34643817e-01 -4.91854697e-01 1.56525791e-01
-7.42812335e-01 1.70034453e-01 -9.81618613e-02 2.97804266e-01
-2.89686173e-01 2.07218900e-01 1.02084947e+00 2.96626717e-01
-1.64577812e-02 -4.66298684e-02 -1.46019554e+00 -1.03885460e+00
-4.41724882e-02 -1.39120889e+00 6.21869206e-01 -3.00014734e-01
-1.57322264e+00 -5.91175212e-03 -2.75419354e-01 -8.78682613e-01
-8.18658233e-01 -7.48944938e-01 -8.57533813e-01 1.14003181e+00
-1.53251898e+00 -1.07893753e+00 3.65185976e-01 9.61699989e-03
-8.46479684e-02 -7.51792729e-01 8.76859546e-01 -4.96228099e-01
-4.37805742e-01 1.83764949e-01 9.67591852e-02 -1.31431371e-01
4.11504537e-01 -1.36724532e+00 7.86532998e-01 4.82463211e-01
-3.01287532e-01 6.46873653e-01 5.87858617e-01 -4.52882677e-01
-9.16322291e-01 -4.84986275e-01 1.55073476e+00 -8.17268729e-01
7.76002109e-01 1.95903644e-01 -8.75951707e-01 8.41687918e-01
9.43606421e-02 -3.15746218e-01 1.29854774e+00 2.07296923e-01
-3.28210741e-01 1.00815799e-02 -1.51282072e+00 1.52986243e-01
5.18700063e-01 -6.35819256e-01 -1.24350238e+00 -2.60899872e-01
2.53076971e-01 3.13839167e-01 -1.02580118e+00 -3.37705463e-02
8.16635847e-01 -9.98428941e-01 7.71101296e-01 -1.53319108e+00
6.80150688e-01 4.17401612e-01 -8.39797631e-02 -4.91231382e-01
-7.88383842e-01 -9.16623235e-01 1.43472608e-02 1.24023616e+00
7.30028033e-01 -7.36524224e-01 1.15541112e+00 2.37920904e+00
5.80196798e-01 -4.95628566e-01 -1.15587604e+00 -4.01920974e-01
8.61964941e-01 -4.80667114e-01 1.07472622e+00 1.56109774e+00
5.88759601e-01 -2.29747325e-01 -3.97710085e-01 -3.95667821e-01
9.96996760e-01 1.98592082e-01 3.94825011e-01 -1.68066514e+00
-5.70069961e-02 -7.47485220e-01 -3.35243493e-01 3.17654051e-02
2.87073404e-01 -4.47459817e-01 -6.00435853e-01 -1.31130719e+00
9.48877335e-02 -4.42052633e-01 -5.32717526e-01 4.22229975e-01
-1.03180319e-01 -6.90060139e-01 7.04808235e-01 4.82887268e-01
1.03305154e-01 2.87386507e-01 9.04136717e-01 3.86357680e-02
-1.79739416e-01 1.99014664e-01 -1.41207957e+00 9.59752738e-01
1.09560370e+00 -7.34916747e-01 -2.05960289e-01 -2.94915676e-01
5.60334027e-01 2.87170619e-01 3.24098021e-01 -1.39462054e-01
-7.77963921e-02 -1.35081851e+00 2.69949406e-01 2.36597151e-01
-3.89860541e-01 -1.17349899e+00 6.00420415e-01 9.69035864e-01
-7.11930215e-01 8.29882175e-02 -3.56255233e-01 3.55797023e-01
-4.00998682e-01 -7.08122253e-01 6.63957536e-01 -4.21813965e-01
-4.15246516e-01 -3.68985645e-02 -8.38199437e-01 -2.35239081e-02
1.18814492e+00 -7.49723136e-01 -2.20885813e-01 -4.25452441e-01
-8.77246320e-01 4.00536776e-01 5.67519546e-01 3.68473411e-01
4.43911791e-01 -1.20264804e+00 -8.74617279e-01 1.82774410e-01
-2.05941200e-02 -1.25547498e-01 -3.32629532e-01 2.84822673e-01
-8.09334517e-01 3.89684498e-01 -9.09273803e-01 5.24012566e-01
-8.23010266e-01 4.07593966e-01 6.73268855e-01 -1.26032472e-01
-6.40226826e-02 5.62875032e-01 -1.00416290e-02 -3.49221602e-02
-1.37296282e-02 -2.05773979e-01 -5.33886194e-01 1.92829996e-01
3.02619368e-01 9.51175094e-01 -2.54841954e-01 -3.04848731e-01
-4.29947615e-01 -1.49022669e-01 -2.78506488e-01 -4.30635542e-01
1.42864943e+00 -5.80916740e-02 -1.15171902e-01 2.27133945e-01
3.19799036e-01 -6.96707070e-01 -1.16948521e+00 1.50197074e-01
1.02588332e+00 -1.07528079e+00 2.34352738e-01 -9.24568713e-01
-8.13120544e-01 7.99417973e-01 -1.85792729e-01 5.92676759e-01
5.30018449e-01 -4.92873877e-01 2.50127241e-02 4.64815855e-01
2.13165179e-01 -1.79908669e+00 -5.50877750e-01 -6.59275576e-02
9.17061269e-01 -1.55747914e+00 3.51731241e-01 -1.46393120e-01
-1.35597730e+00 1.30522466e+00 5.35826623e-01 1.52631044e-01
9.40538108e-01 2.52908677e-01 6.72448158e-01 8.06884840e-03
-9.31728780e-01 6.69491589e-01 6.37765750e-02 9.35468376e-01
6.91969991e-01 6.39449120e-01 -6.78479373e-01 1.59752107e+00
-3.68610680e-01 3.44497412e-01 8.84730160e-01 1.12434649e+00
-1.69985905e-01 -1.41377234e+00 -3.10895354e-01 7.12972581e-01
-1.03548694e+00 -7.09108934e-02 -7.84680128e-01 8.22959542e-01
6.24870695e-03 1.46190727e+00 1.93683431e-01 -4.51452345e-01
1.94088385e-01 1.43749684e-01 -2.53759772e-01 -4.24923390e-01
-1.30883312e+00 8.71804729e-03 6.04712486e-01 -5.74098825e-01
-5.30633867e-01 -1.30146980e+00 -1.12596238e+00 -8.96088362e-01
-4.31137711e-01 2.61029899e-01 5.10751843e-01 9.47214901e-01
1.09494120e-01 3.44642662e-02 8.16333532e-01 -2.81597786e-02
-1.33193886e+00 -6.40812278e-01 -1.12218618e+00 5.65338373e-01
-2.06534620e-02 -4.02420044e-01 -6.61532938e-01 2.88825184e-01] | [8.78154182434082, 5.664770603179932] |
42fabae5-31f9-4b59-b80e-2f0399ccda7c | scriptworld-text-based-environment-for | 2307.03906 | null | https://arxiv.org/abs/2307.03906v1 | https://arxiv.org/pdf/2307.03906v1.pdf | ScriptWorld: Text Based Environment For Learning Procedural Knowledge | Text-based games provide a framework for developing natural language understanding and commonsense knowledge about the world in reinforcement learning based agents. Existing text-based environments often rely on fictional situations and characters to create a gaming framework and are far from real-world scenarios. In this paper, we introduce ScriptWorld: a text-based environment for teaching agents about real-world daily chores and hence imparting commonsense knowledge. To the best of our knowledge, it is the first interactive text-based gaming framework that consists of daily real-world human activities designed using scripts dataset. We provide gaming environments for 10 daily activities and perform a detailed analysis of the proposed environment. We develop RL-based baseline models/agents to play the games in Scriptworld. To understand the role of language models in such environments, we leverage features obtained from pre-trained language models in the RL agents. Our experiments show that prior knowledge obtained from a pre-trained language model helps to solve real-world text-based gaming environments. We release the environment via Github: https://github.com/Exploration-Lab/ScriptWorld | ['Ashutosh Modi', 'Umang Pandey', 'Areeb Ahmad', 'Abhinav Joshi'] | 2023-07-08 | null | null | null | null | ['natural-language-understanding', 'text-based-games'] | ['natural-language-processing', 'playing-games'] | [-2.14713752e-01 6.06955774e-03 2.55174756e-01 -1.64087236e-01
-2.11075857e-01 -6.26829207e-01 9.90183949e-01 -3.87667656e-01
-4.08611804e-01 8.09712589e-01 3.10052663e-01 -4.36080605e-01
2.93776989e-01 -1.11508632e+00 -6.72504544e-01 -3.48273814e-02
-7.41969943e-02 7.45017886e-01 2.79202402e-01 -1.08585238e+00
2.94953793e-01 1.03464395e-01 -1.59217083e+00 6.00412667e-01
6.94869339e-01 2.12749138e-01 5.54244578e-01 1.06590164e+00
3.47943511e-03 2.07986283e+00 -7.74902999e-01 -4.19566333e-01
3.57063085e-01 -8.53051841e-01 -1.11341715e+00 -1.02625191e-01
-2.23557487e-01 -8.79876316e-01 -8.33724439e-01 7.95717597e-01
3.68771106e-01 6.30621195e-01 3.91460806e-01 -1.43263447e+00
-6.22958362e-01 9.63087082e-01 8.63400325e-02 6.82599917e-02
8.63709629e-01 9.67303991e-01 7.81129897e-01 -2.83964723e-01
7.84789324e-01 1.35724247e+00 2.16323853e-01 9.86492455e-01
-7.00739086e-01 -8.08791876e-01 7.77534842e-02 3.82451653e-01
-8.58546495e-01 -3.20713550e-01 8.40681553e-01 -4.30116892e-01
1.30688524e+00 6.45160675e-02 1.08877969e+00 1.82115924e+00
2.20450297e-01 1.07896805e+00 1.50460660e+00 -3.09924901e-01
2.93527573e-01 -1.88949838e-04 -2.38619328e-01 1.06246412e+00
-3.93804520e-01 5.86888015e-01 -8.93538356e-01 1.65238291e-01
1.29426777e+00 -1.00544266e-01 1.85345754e-01 -2.29818195e-01
-1.31860113e+00 1.02779043e+00 2.73287028e-01 2.61188716e-01
-2.63778478e-01 5.74041069e-01 6.49793923e-01 3.42505276e-01
1.75133392e-01 9.23557341e-01 -7.07721114e-02 -9.45189476e-01
-2.60457516e-01 6.42064750e-01 1.01080918e+00 9.72759068e-01
4.89330292e-01 2.60473698e-01 -1.25920415e-01 6.73289955e-01
3.05568397e-01 3.84930193e-01 6.44677997e-01 -1.20520401e+00
3.94580483e-01 5.85783482e-01 2.47111514e-01 -6.95864618e-01
-4.84045833e-01 -1.27372310e-01 -7.31708705e-02 3.54276299e-01
5.13100088e-01 -4.44558412e-01 -5.77842891e-01 1.83868515e+00
1.47172343e-02 6.71938181e-01 2.40512371e-01 8.15191805e-01
1.15813732e+00 7.04627752e-01 3.21657002e-01 2.92872488e-01
1.22242725e+00 -1.14289820e+00 -4.11751360e-01 -5.21155357e-01
8.47514629e-01 -2.63777286e-01 1.95341134e+00 4.12312061e-01
-9.33546126e-01 -3.73937368e-01 -8.66373241e-01 1.62091069e-02
-5.19946992e-01 -1.76486000e-01 1.06315136e+00 5.88177383e-01
-8.78912330e-01 3.78659695e-01 -1.01422310e+00 -7.24534869e-01
2.67212033e-01 -3.33204046e-02 3.77133191e-02 6.17163852e-02
-1.46989894e+00 1.22156549e+00 6.03266060e-01 -3.44235629e-01
-1.72019601e+00 -3.32695991e-01 -1.10088968e+00 -4.31016892e-01
8.33284616e-01 -6.31547987e-01 1.68569386e+00 -7.94602752e-01
-2.24816799e+00 1.05509055e+00 3.16596776e-01 -4.25214350e-01
7.34585464e-01 -3.08844715e-01 -1.46420849e-02 -1.13026045e-01
1.45248368e-01 5.08783162e-01 2.11817130e-01 -1.20833027e+00
-3.49820167e-01 2.55771301e-05 1.09032702e+00 5.43985128e-01
1.56384602e-01 3.65186073e-02 -2.18277648e-01 -5.21309614e-01
-6.42073214e-01 -1.04387116e+00 -3.45831960e-01 -3.40395659e-01
-2.23478273e-01 -1.59429550e-01 2.67320842e-01 -3.12983841e-01
7.73267269e-01 -1.76350665e+00 -2.16233298e-01 -2.18133405e-01
3.55031610e-01 1.47221446e-01 -2.87950009e-01 9.05989945e-01
2.95363754e-01 -9.65041965e-02 3.71833920e-01 9.33252182e-03
4.22244519e-01 4.35626209e-01 -2.29424596e-01 -1.57672450e-01
-6.48079515e-02 1.24481273e+00 -1.46047032e+00 -3.41774911e-01
9.31905389e-01 1.20446824e-01 -7.87913203e-01 5.38811147e-01
-7.31724501e-01 6.38394356e-01 -6.48934007e-01 1.28531933e-01
5.31012826e-02 -8.39058831e-02 2.28673086e-01 5.22607744e-01
1.81908473e-01 7.05130756e-01 -9.34283435e-01 2.07277703e+00
-8.73312533e-01 6.51844740e-01 -5.54701388e-01 -5.34799993e-01
6.01321280e-01 1.12167992e-01 8.13751742e-02 -9.42447186e-01
4.38592136e-01 -2.34710664e-01 1.17795698e-01 -7.40966618e-01
6.38601840e-01 -5.67374706e-01 -3.69216055e-01 7.89141476e-01
1.95554018e-01 -7.58677661e-01 4.44217294e-01 4.63747054e-01
1.30536604e+00 6.92434967e-01 4.44444090e-01 1.11550679e-02
1.08776368e-01 3.66641939e-01 6.49144650e-02 1.20169663e+00
-3.39430958e-01 3.24552618e-02 4.08906281e-01 -6.50600135e-01
-7.35823512e-01 -1.03087246e+00 6.56758130e-01 1.88789570e+00
1.76427722e-01 -6.26623511e-01 -7.38669991e-01 -3.04140121e-01
-6.03708923e-01 1.38444245e+00 -7.17827082e-01 -2.54160255e-01
-4.55959231e-01 -3.67711097e-01 9.75567520e-01 4.46276993e-01
8.60527515e-01 -1.86667848e+00 -1.17030585e+00 2.62595922e-01
-4.12766099e-01 -1.19817924e+00 -1.20406263e-02 1.37739852e-01
-5.13761520e-01 -1.12052429e+00 -1.97816808e-02 -4.91553873e-01
4.43235971e-02 -2.42301486e-02 1.65158987e+00 1.67296842e-01
-1.86720505e-01 5.19216239e-01 -6.08786404e-01 -6.77743793e-01
-7.50275671e-01 -2.29955629e-01 -9.30345356e-02 -7.73201406e-01
4.26846385e-01 -7.17834234e-01 -1.97705045e-01 2.35322386e-01
-5.85081279e-01 7.86051631e-01 2.15004370e-01 5.93914270e-01
-2.07714945e-01 -6.61616996e-02 9.47547853e-02 -8.97687137e-01
1.03355396e+00 -5.04768550e-01 -4.65002298e-01 -9.24314260e-02
1.15024716e-01 -1.68971077e-03 6.84327900e-01 -5.65150797e-01
-1.11565661e+00 -3.65416765e-01 -1.61275968e-01 1.14495410e-02
-5.73790491e-01 5.36037445e-01 -5.83124487e-03 2.49075264e-01
1.05933070e+00 4.87228990e-01 -3.94572973e-01 1.50044844e-01
6.40082955e-01 3.81312340e-01 4.45183009e-01 -1.28578866e+00
6.64948046e-01 1.54814482e-01 -4.62831259e-01 -5.90506792e-01
-4.97934878e-01 -1.17258027e-01 -2.81442851e-01 -6.61151409e-01
7.17907190e-01 -9.07507777e-01 -1.24250221e+00 6.08439326e-01
-8.48232746e-01 -1.43914211e+00 -3.95739168e-01 4.87442046e-01
-1.18855703e+00 -1.87324166e-01 -8.77242804e-01 -9.33755517e-01
1.78805783e-01 -1.13223708e+00 7.23251462e-01 2.53463417e-01
-4.79329377e-01 -9.72054660e-01 4.81309921e-01 6.59978986e-01
2.49317706e-01 3.27691972e-01 6.53794289e-01 -6.96715534e-01
-3.36666971e-01 2.90411919e-01 1.27337277e-01 -2.31513396e-01
8.59503299e-02 -2.88754672e-01 -1.04389143e+00 2.41310507e-01
2.93086842e-02 -1.34549057e+00 1.58530653e-01 1.88260656e-02
8.43864083e-01 -1.76003724e-01 1.42672166e-01 2.72413731e-01
1.13279378e+00 5.70900440e-01 9.06829476e-01 7.07449198e-01
4.87092048e-01 3.21228147e-01 6.62673175e-01 8.29075515e-01
8.43186677e-01 4.37647730e-01 4.15040791e-01 1.14219666e-01
1.85315400e-01 -8.32065463e-01 7.52268374e-01 4.23204452e-01
-3.30484658e-01 -3.36005121e-01 -1.18189728e+00 5.29450357e-01
-1.79857123e+00 -1.31039298e+00 3.42967153e-01 1.53659403e+00
1.17002797e+00 3.62252653e-01 3.16535234e-01 -4.25496876e-01
2.06397712e-01 4.32965279e-01 -7.17855513e-01 -5.37138045e-01
2.20816329e-01 3.24133039e-01 -2.28873178e-01 7.88850427e-01
-9.28937972e-01 1.72174335e+00 5.76483965e+00 8.48759174e-01
-9.71906960e-01 1.08862467e-01 2.72266358e-01 -3.43008548e-01
7.50944614e-02 -2.58065462e-01 -9.46098492e-02 -4.55697142e-02
8.30617547e-01 -3.77410561e-01 8.99917603e-01 9.53927875e-01
6.30499780e-01 -2.22271129e-01 -1.26552880e+00 9.24614906e-01
4.13004644e-02 -1.20959079e+00 2.32417762e-01 -7.23362109e-03
5.23028731e-01 3.83661211e-01 3.94722968e-02 1.06773388e+00
1.54593599e+00 -1.33539641e+00 8.60898018e-01 2.44672388e-01
4.41781074e-01 -7.47673512e-01 4.93344784e-01 7.22613931e-01
-7.79389679e-01 1.41914845e-01 -1.36302412e-01 -9.04776394e-01
-1.76914752e-01 -4.14966196e-01 -9.76968646e-01 1.49576351e-01
4.19736236e-01 7.69461513e-01 -4.71903652e-01 3.70697320e-01
-5.79342186e-01 7.61093080e-01 -1.87303647e-01 -5.99504530e-01
5.53273499e-01 -1.81534603e-01 4.04129058e-01 1.08916259e+00
-1.65610284e-01 7.60100543e-01 4.81995881e-01 1.27632225e+00
-9.46188867e-02 1.81833860e-02 -1.04232872e+00 -4.14285451e-01
1.23558857e-01 1.03838837e+00 -7.33368993e-01 -3.90105337e-01
-3.86543512e-01 9.79761660e-01 3.76069367e-01 2.38610134e-01
-1.09939361e+00 -1.77631691e-01 7.41876245e-01 3.85879278e-02
-3.57525110e-01 -3.89574856e-01 -1.26645923e-01 -1.47468555e+00
-4.36343580e-01 -1.53140938e+00 -1.51683008e-02 -1.29810965e+00
-1.09281850e+00 5.94217777e-01 1.85189262e-01 -1.01090777e+00
-7.27591991e-01 -7.03327477e-01 -8.67878914e-01 4.33835953e-01
-9.18634355e-01 -1.29911840e+00 -4.86135662e-01 9.49640036e-01
9.54019308e-01 -4.10145700e-01 1.02591765e+00 -4.65600103e-01
-1.81841448e-01 6.10057153e-02 -2.00370982e-01 5.65535963e-01
4.72573966e-01 -1.29731786e+00 7.34626174e-01 5.33606470e-01
3.45013082e-01 7.77396560e-01 1.04880834e+00 -7.03005314e-01
-1.29968607e+00 -7.84630179e-01 1.17301181e-01 -8.53407621e-01
1.00361490e+00 -7.26803720e-01 -4.13242579e-01 1.08808446e+00
5.37575185e-01 -5.33452928e-01 8.46607089e-01 2.24515438e-01
-2.80097574e-01 6.22449815e-01 -9.70315158e-01 1.29894543e+00
1.47042155e+00 -8.39591444e-01 -9.87476468e-01 4.51201469e-01
3.79064232e-01 -6.72387540e-01 -4.11757171e-01 -2.07414597e-01
5.29588580e-01 -1.12658215e+00 7.79870331e-01 -9.53172982e-01
1.12644124e+00 -3.11017573e-01 -1.22520365e-01 -1.59300041e+00
-4.38402109e-02 -7.79754221e-01 3.79171431e-01 6.81860745e-01
2.00122893e-01 -4.67778474e-01 9.16551292e-01 6.82052732e-01
1.37312084e-01 -3.49645942e-01 -4.01209027e-01 -6.39799833e-01
3.22938949e-01 -1.16935933e+00 5.05661130e-01 1.10316646e+00
6.12419665e-01 3.96730542e-01 -3.57039541e-01 -1.72617659e-01
3.36242348e-01 -1.64088771e-01 1.32242906e+00 -6.87803328e-01
-7.60216713e-01 -4.22648698e-01 -2.66438603e-01 -1.04634750e+00
4.30715024e-01 -7.32922375e-01 1.71468481e-01 -1.58112788e+00
4.99124140e-01 -1.39597073e-01 1.14956662e-01 5.51700354e-01
3.67321111e-02 2.09874555e-01 5.28298616e-01 -8.70834664e-02
-1.02966714e+00 7.03070760e-01 1.42608488e+00 -2.34107021e-02
-3.37875277e-01 -4.30831045e-01 -6.29446924e-01 1.06671154e+00
1.09663069e+00 -2.20984936e-01 -9.27021682e-01 -4.73123521e-01
5.04686713e-01 7.96048567e-02 5.92918992e-01 -1.17197251e+00
1.12959586e-01 -8.11327040e-01 -1.59068883e-03 -9.18006450e-02
4.72443968e-01 -3.70737225e-01 -8.08136910e-02 4.58509922e-01
-6.18428051e-01 -2.12522462e-01 5.63508689e-01 1.30147785e-01
6.56007379e-02 -9.09010023e-02 4.88484800e-01 -8.03488016e-01
-1.07880926e+00 -1.73883244e-01 -1.09458494e+00 5.65261304e-01
1.08840501e+00 -2.38620088e-01 -3.59824151e-01 -1.09735060e+00
-8.45754087e-01 2.88746059e-01 5.41316926e-01 4.88805085e-01
6.09769166e-01 -1.12667370e+00 -8.05574179e-01 -1.78649038e-01
3.04682106e-01 -3.11844796e-01 1.19847566e-01 -3.52834933e-03
-1.02342653e+00 1.79013893e-01 -8.21258187e-01 -7.73163065e-02
-1.03652561e+00 3.74439418e-01 7.23479152e-01 -5.53350329e-01
-5.73738813e-01 8.85415316e-01 6.31308556e-01 -9.54293668e-01
-1.56931445e-01 -5.03563881e-01 -3.00425649e-01 -5.35352409e-01
6.73091412e-01 1.37783498e-01 -4.09598887e-01 -4.57703501e-01
-1.59176558e-01 6.49115397e-03 2.85447501e-02 -7.62000799e-01
1.42861652e+00 1.13131680e-01 2.50835240e-01 8.48103821e-01
3.95792037e-01 -3.12606871e-01 -1.23372507e+00 -1.51567593e-01
-3.45251054e-01 -3.75523537e-01 -2.35938787e-01 -1.19052577e+00
-6.75479114e-01 9.12378728e-01 2.49450058e-01 -1.67879194e-01
7.87190855e-01 -3.55332671e-03 5.10243416e-01 9.13550079e-01
1.12669170e+00 -1.14614046e+00 6.25467956e-01 1.05291879e+00
1.06816423e+00 -1.17880380e+00 -3.56438816e-01 -5.58765456e-02
-1.16204298e+00 7.66945362e-01 1.10598886e+00 -4.59741503e-01
2.55213171e-01 3.91571522e-01 3.21958989e-01 -4.00260150e-01
-1.10331106e+00 -4.45216298e-01 -3.31119627e-01 1.04611182e+00
5.28144002e-01 3.50899071e-01 1.83264971e-01 7.95114815e-01
-9.49328005e-01 3.87806565e-01 8.70126069e-01 1.19596016e+00
-2.29006380e-01 -8.28722537e-01 -1.98159963e-01 1.60351500e-01
-2.55150706e-01 -2.35270917e-01 -8.91175926e-01 9.17583823e-01
1.79023370e-02 1.24271226e+00 -3.74505758e-01 -6.32330835e-01
3.94817054e-01 4.57860120e-02 8.66241693e-01 -8.79942477e-01
-9.29450929e-01 -4.10639763e-01 4.03195113e-01 -9.83208716e-01
-2.57442385e-01 -3.64801913e-01 -1.45237267e+00 -7.76881993e-01
1.86868221e-01 -1.06410883e-01 1.18019901e-01 1.15382266e+00
-1.82190537e-01 5.59105158e-01 3.93677428e-02 -1.05970550e+00
-1.33524343e-01 -1.04530501e+00 -4.66832578e-01 5.01107037e-01
-2.36717254e-01 -7.92837679e-01 3.88782732e-02 4.55061384e-02] | [3.8920204639434814, 1.2813148498535156] |
2bbce5e4-5f03-4d46-a7df-c2251acdea19 | x-trans2cap-cross-modal-knowledge-transfer | 2203.00843 | null | https://arxiv.org/abs/2203.00843v3 | https://arxiv.org/pdf/2203.00843v3.pdf | X-Trans2Cap: Cross-Modal Knowledge Transfer using Transformer for 3D Dense Captioning | 3D dense captioning aims to describe individual objects by natural language in 3D scenes, where 3D scenes are usually represented as RGB-D scans or point clouds. However, only exploiting single modal information, e.g., point cloud, previous approaches fail to produce faithful descriptions. Though aggregating 2D features into point clouds may be beneficial, it introduces an extra computational burden, especially in inference phases. In this study, we investigate a cross-modal knowledge transfer using Transformer for 3D dense captioning, X-Trans2Cap, to effectively boost the performance of single-modal 3D caption through knowledge distillation using a teacher-student framework. In practice, during the training phase, the teacher network exploits auxiliary 2D modality and guides the student network that only takes point clouds as input through the feature consistency constraints. Owing to the well-designed cross-modal feature fusion module and the feature alignment in the training phase, X-Trans2Cap acquires rich appearance information embedded in 2D images with ease. Thus, a more faithful caption can be generated only using point clouds during the inference. Qualitative and quantitative results confirm that X-Trans2Cap outperforms previous state-of-the-art by a large margin, i.e., about +21 and about +16 absolute CIDEr score on ScanRefer and Nr3D datasets, respectively. | ['Shuguang Cui', 'Zhen Li', 'Guanbin Li', 'Yao Guo', 'Yinghong Liao', 'Xu Yan', 'Zhihao Yuan'] | 2022-03-02 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Yuan_X-Trans2Cap_Cross-Modal_Knowledge_Transfer_Using_Transformer_for_3D_Dense_Captioning_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Yuan_X-Trans2Cap_Cross-Modal_Knowledge_Transfer_Using_Transformer_for_3D_Dense_Captioning_CVPR_2022_paper.pdf | cvpr-2022-1 | ['dense-captioning', '3d-dense-captioning'] | ['computer-vision', 'computer-vision'] | [-1.27466366e-01 4.04114604e-01 -3.09159577e-01 -4.91473675e-01
-1.07658863e+00 -6.40323162e-01 7.43943632e-01 5.99717684e-02
-8.21912810e-02 4.93818194e-01 -5.53892180e-02 -2.79075652e-02
2.64848135e-02 -8.41537058e-01 -1.33053350e+00 -6.85630620e-01
4.31018710e-01 9.28522050e-01 -2.04324480e-02 -1.42173305e-01
-1.99772343e-01 6.55300200e-01 -1.60629082e+00 1.50977103e-02
7.93989003e-01 1.27160859e+00 2.44839042e-01 8.20079967e-02
-6.53467417e-01 3.84916693e-01 -2.76825219e-01 -6.51320755e-01
2.47023433e-01 1.60495248e-02 -6.28447413e-01 2.32331455e-01
6.96343422e-01 -5.44528663e-01 -4.73256826e-01 1.05308104e+00
3.80506814e-01 -1.19519103e-02 5.93075514e-01 -1.39905667e+00
-8.62942278e-01 2.62736738e-01 -5.64887583e-01 -4.21248108e-01
5.21455586e-01 3.10061336e-01 9.24375951e-01 -1.02612305e+00
4.98910815e-01 1.28735161e+00 5.92914164e-01 6.48330510e-01
-1.40876424e+00 -7.62271166e-01 2.81908095e-01 4.10656258e-02
-1.59206057e+00 -3.99477109e-02 1.14322603e+00 -3.66401553e-01
8.28260303e-01 1.99238256e-01 7.54266679e-01 1.14273238e+00
-3.75007063e-01 1.01532280e+00 9.47501838e-01 6.94043934e-02
8.10825601e-02 2.88844526e-01 -2.31016427e-01 6.54046953e-01
1.27876848e-02 1.22237578e-01 -5.75676918e-01 1.17140114e-01
1.01664114e+00 1.15561202e-01 -1.25725955e-01 -8.14872265e-01
-1.48959959e+00 8.14703763e-01 1.19350052e+00 -5.74362800e-02
-6.48948371e-01 2.22489536e-01 1.71622366e-01 -5.96112683e-02
5.41749597e-01 3.99910122e-01 -3.88808787e-01 -5.53822704e-02
-7.86937058e-01 2.31057346e-01 2.35366017e-01 1.48262775e+00
1.14756334e+00 -2.16882393e-01 3.59113030e-02 5.32131135e-01
5.00228643e-01 1.16927230e+00 -8.28242004e-02 -8.24416697e-01
7.08280861e-01 8.33211899e-01 4.17188667e-02 -8.64218712e-01
-1.58864036e-01 -2.92286992e-01 -1.02918124e+00 -4.92492169e-02
3.12832370e-02 2.67025918e-01 -1.20295346e+00 1.78142369e+00
6.22611165e-01 2.98259914e-01 1.78693905e-01 1.31822789e+00
1.16851532e+00 7.33982563e-01 1.33081555e-01 2.29935497e-01
1.38774681e+00 -8.09215188e-01 -4.22773004e-01 -1.73453212e-01
3.81976932e-01 -5.39196789e-01 1.00129056e+00 -1.19816039e-04
-1.03468406e+00 -6.91194534e-01 -8.28161657e-01 -3.18194419e-01
-4.22321171e-01 -1.52845979e-01 7.92526126e-01 1.73185691e-01
-7.26693034e-01 1.63552627e-01 -8.81239653e-01 -2.73564667e-01
6.22507036e-01 3.50013882e-01 -8.29588711e-01 -4.35181260e-01
-1.06417763e+00 8.67118895e-01 4.16629285e-01 2.16623656e-02
-9.44640458e-01 -1.13030529e+00 -1.16961694e+00 -6.85528815e-02
1.67192802e-01 -1.08821476e+00 1.25960279e+00 -3.47368091e-01
-1.16974139e+00 1.13678324e+00 -1.70140304e-02 -1.66386783e-01
4.75871354e-01 -2.12649629e-01 -4.89563821e-03 3.92978400e-01
1.36174679e-01 1.41619229e+00 5.09080172e-01 -1.72873294e+00
-3.11356753e-01 -3.64939839e-01 4.91842628e-01 4.09371525e-01
7.00344816e-02 -6.01036370e-01 -9.11691070e-01 -3.46225619e-01
4.51108098e-01 -1.01749408e+00 2.53211521e-02 4.91123736e-01
-4.92628604e-01 -2.75627524e-01 8.18621278e-01 -3.81848097e-01
2.11418077e-01 -2.20453501e+00 2.14926228e-01 6.71101436e-02
1.95089757e-01 7.27279037e-02 -1.05632827e-01 3.89604978e-02
-1.92069747e-02 5.83538262e-04 -3.24357182e-01 -6.57123804e-01
4.11617160e-01 5.94095588e-01 -4.47923511e-01 4.17549103e-01
7.70988464e-01 1.23862529e+00 -1.10830200e+00 -7.35011458e-01
7.43688583e-01 9.72532809e-01 -5.44906735e-01 4.48741287e-01
-5.99483490e-01 6.03349149e-01 -6.97151303e-01 7.54584134e-01
9.95791793e-01 -5.99411249e-01 -5.36047041e-01 -5.62440455e-01
7.18344515e-03 2.46967316e-01 -7.55930603e-01 2.45829248e+00
-7.83460379e-01 3.53842139e-01 -1.28817707e-01 -9.27862704e-01
1.01318800e+00 4.37887758e-01 5.32380342e-01 -8.69530499e-01
3.64256389e-02 1.46394938e-01 -6.63824499e-01 -3.84405553e-01
2.75287628e-01 -3.67043495e-01 -3.18823844e-01 2.29913652e-01
2.37038523e-01 -8.85343254e-01 -2.28023633e-01 2.47923568e-01
6.92189515e-01 4.70099211e-01 -2.11559370e-01 3.54571283e-01
3.37799489e-01 3.65583569e-01 2.19152480e-01 4.73923683e-01
1.51958138e-01 1.04382122e+00 2.26826742e-01 -2.65569866e-01
-1.18821454e+00 -1.40791893e+00 -1.77908167e-01 3.79908442e-01
6.08484089e-01 -6.00650832e-02 -3.89573753e-01 -7.32612371e-01
3.31156671e-01 7.61094153e-01 -6.33815408e-01 -3.26929778e-01
-3.27238709e-01 -2.86067635e-01 3.95366251e-01 6.12361193e-01
7.46348441e-01 -7.66071141e-01 -2.82973826e-01 -5.07539101e-02
-3.20906401e-01 -1.48521471e+00 -2.05552489e-01 1.20318778e-01
-9.72064376e-01 -7.97322571e-01 -8.36599708e-01 -6.03444278e-01
9.90350544e-01 4.22933638e-01 1.26307654e+00 -1.47548497e-01
1.34212658e-01 4.09222662e-01 -4.35191900e-01 -3.35302263e-01
-1.97782874e-01 -1.55073954e-02 -4.84076068e-02 -2.99577475e-01
5.44041038e-01 -7.45129108e-01 -4.92782801e-01 2.15943933e-01
-8.29383612e-01 6.27159119e-01 9.96288121e-01 6.17758214e-01
1.00333977e+00 -3.83070886e-01 1.29152730e-01 -4.94228095e-01
-2.52789617e-01 -5.02220988e-01 -5.49929678e-01 8.20703432e-02
-2.49733314e-01 1.54318765e-01 5.01922965e-01 -3.86283815e-01
-9.47788715e-01 4.01149422e-01 -2.70292819e-01 -1.18719184e+00
-3.57104748e-01 4.10214782e-01 -4.57952917e-01 -1.47149161e-01
3.88576627e-01 4.94995445e-01 5.56428954e-02 -6.29415393e-01
5.81230998e-01 5.24579763e-01 7.66636729e-01 -6.34824038e-01
1.41735923e+00 6.35124326e-01 5.61885945e-02 -4.89636928e-01
-1.15470755e+00 -4.04583842e-01 -6.80422962e-01 -1.28270611e-01
1.03657651e+00 -1.46457934e+00 -8.80247355e-01 2.16241658e-01
-1.34191918e+00 8.07790235e-02 -2.72222877e-01 5.94318449e-01
-7.34251797e-01 -6.27079904e-02 -2.42875978e-01 -4.48121727e-01
-3.08183819e-01 -1.05800903e+00 1.76971483e+00 2.09672928e-01
1.15347177e-01 -7.04080105e-01 -1.35039493e-01 6.39129221e-01
1.22030906e-01 5.34881473e-01 8.93905342e-01 -3.62717211e-01
-9.97678518e-01 -2.37873182e-01 -7.23038554e-01 2.59901941e-01
1.52094096e-01 -3.92631382e-01 -1.11494911e+00 6.08882383e-02
-4.20819908e-01 -5.29199302e-01 4.13991749e-01 -3.51550467e-02
1.16230905e+00 -2.58823913e-02 -2.53726155e-01 8.29877198e-01
1.38747549e+00 -2.14998350e-01 3.86053056e-01 1.26518399e-01
1.13347530e+00 4.28782314e-01 7.99248159e-01 2.30312273e-01
7.30254054e-01 7.08476424e-01 9.45915759e-01 -2.94706792e-01
-2.75335938e-01 -7.30331302e-01 -1.08974189e-01 7.50312150e-01
1.84046566e-01 -2.36238563e-03 -9.48358357e-01 6.36004508e-01
-1.66902459e+00 -5.61336815e-01 1.20071195e-01 1.82613122e+00
9.74172890e-01 1.56169221e-01 -2.35680163e-01 -4.01120931e-02
5.88822842e-01 1.08375095e-01 -7.55989254e-01 2.41055712e-01
-1.20396547e-01 -1.61581300e-02 3.93237531e-01 2.74683923e-01
-9.32272255e-01 8.62548828e-01 4.56724548e+00 6.39911652e-01
-1.22913420e+00 -2.52806284e-02 3.10500592e-01 -2.15424955e-01
-6.30658746e-01 -1.81880280e-01 -6.72238827e-01 4.41612363e-01
5.73392093e-01 4.15721685e-02 7.66356140e-02 9.35099423e-01
-2.72316858e-02 2.19821185e-01 -1.42326939e+00 1.41655850e+00
-3.69301327e-02 -1.47307014e+00 4.44453657e-01 -6.51391447e-02
6.23292327e-01 2.35178694e-01 5.24767004e-02 4.48157191e-01
1.14415251e-01 -8.86205971e-01 8.36766005e-01 6.18272305e-01
9.55281675e-01 -7.50714898e-01 7.30360091e-01 3.70637298e-01
-1.00738895e+00 4.63340759e-01 -2.70615906e-01 3.08863968e-01
4.44107085e-01 7.26809025e-01 -8.09625328e-01 9.14911628e-01
7.85672188e-01 7.97720730e-01 -3.29996109e-01 1.00056624e+00
-2.87031561e-01 2.74053901e-01 -7.16159463e-01 1.16499081e-01
5.58344007e-01 -1.07367197e-02 5.09195089e-01 8.06253970e-01
4.02090460e-01 3.10915887e-01 2.95341890e-02 1.20106614e+00
-1.49106458e-01 -3.90410036e-01 -5.70498347e-01 2.86503416e-02
4.73973989e-01 1.08930576e+00 -1.99979678e-01 -2.74250895e-01
-4.47278619e-01 8.42615306e-01 2.27847949e-01 1.87377155e-01
-1.00686693e+00 -2.00893536e-01 8.30937803e-01 1.03913635e-01
4.51625347e-01 -2.83318222e-01 -3.27586442e-01 -1.11306345e+00
2.60798812e-01 -4.34938312e-01 -1.01909801e-01 -1.40590680e+00
-1.41396511e+00 6.19374037e-01 2.66506314e-01 -1.50545299e+00
-2.26343498e-01 -4.43909377e-01 -1.84565827e-01 9.53146040e-01
-1.88394916e+00 -1.68356884e+00 -7.81102777e-01 6.71843767e-01
2.43312195e-01 3.12748432e-01 7.81852961e-01 4.80276287e-01
-2.45379433e-01 4.53947783e-01 -2.86311597e-01 1.11441210e-01
5.11465013e-01 -1.35013592e+00 4.73763704e-01 2.56883204e-01
2.25492775e-01 4.17267442e-01 7.28442013e-01 -4.12593693e-01
-1.72305965e+00 -1.28809488e+00 7.12186992e-01 -4.94954646e-01
3.69145840e-01 -5.34600675e-01 -1.06650877e+00 4.52088147e-01
-1.90125540e-01 3.94554138e-01 4.27879989e-01 -9.07600597e-02
-5.91875196e-01 -2.09950030e-01 -1.25628829e+00 3.62708747e-01
1.13495815e+00 -7.63392687e-01 -6.21229351e-01 4.72023875e-01
1.24426532e+00 -8.89816403e-01 -1.18241620e+00 5.23219228e-01
1.63716093e-01 -5.15357971e-01 1.25241303e+00 -2.97234535e-01
6.19203329e-01 -5.65812528e-01 -3.47250938e-01 -1.23665655e+00
-3.65012214e-02 -2.21159682e-01 -2.94747829e-01 1.25235283e+00
2.84759372e-01 -2.29997262e-01 9.72633898e-01 6.11462831e-01
-4.26025242e-01 -6.78756237e-01 -1.07464731e+00 -6.50185823e-01
7.12862331e-03 -5.88670015e-01 1.01455641e+00 9.32880700e-01
-5.90158463e-01 2.90290296e-01 -8.82241875e-02 6.35190427e-01
8.32473040e-01 4.56487417e-01 9.09369111e-01 -1.26017118e+00
5.66093549e-02 -1.18109904e-01 -6.48510277e-01 -1.52324045e+00
3.57526004e-01 -9.21429932e-01 1.69323325e-01 -1.70140803e+00
-1.08736577e-02 -8.13512802e-01 -6.79139569e-02 7.39485085e-01
7.58076534e-02 5.48406601e-01 3.15171689e-01 1.38003230e-01
-6.33078456e-01 1.05877399e+00 1.62674928e+00 -4.46660131e-01
7.28617236e-02 -1.88835442e-01 -6.53163671e-01 4.62149918e-01
4.51541632e-01 -4.47777182e-01 -3.91719311e-01 -9.07593071e-01
2.07882255e-01 1.80360019e-01 9.01185274e-01 -9.12884235e-01
1.93694755e-01 -8.55531394e-02 4.97908473e-01 -1.13646865e+00
8.54279339e-01 -1.21431077e+00 2.81860769e-01 1.97339058e-03
1.82329805e-03 -1.64312229e-01 3.82795036e-01 4.77789015e-01
-4.48463947e-01 2.25160360e-01 5.97941518e-01 -9.91825834e-02
-6.80192769e-01 7.16190040e-01 5.04418194e-01 -2.17181563e-01
1.00682437e+00 -2.57551163e-01 -2.40243748e-01 -2.31853679e-01
-5.51011384e-01 5.65188646e-01 6.57051206e-01 6.48313403e-01
7.00170517e-01 -1.75538898e+00 -6.00350440e-01 2.21433342e-01
6.11782372e-01 1.08460987e+00 5.55944502e-01 7.46868551e-01
-4.57776606e-01 5.79371274e-01 -2.45269969e-01 -1.19243872e+00
-8.17210019e-01 5.93317032e-01 2.17611641e-01 1.35713056e-01
-7.31623054e-01 8.86603832e-01 2.92243302e-01 -7.99367547e-01
2.67878115e-01 -2.75075763e-01 2.37387449e-01 -1.32657632e-01
1.85416609e-01 -3.47203821e-01 4.93824296e-02 -8.85238409e-01
-5.50021410e-01 9.02674258e-01 -4.12434340e-02 1.10750869e-01
1.47176671e+00 -1.33021221e-01 9.03885812e-02 3.33666831e-01
1.50583577e+00 -5.37328243e-01 -1.51506174e+00 -4.95644480e-01
-4.88709778e-01 -5.78061402e-01 2.20421880e-01 -6.50543451e-01
-1.21382833e+00 9.97791111e-01 3.96277964e-01 -6.82087168e-02
1.04675531e+00 5.58375537e-01 9.02242959e-01 4.37875777e-01
4.69325811e-01 -3.49794030e-01 2.67026853e-02 2.59263009e-01
9.43312764e-01 -1.53473735e+00 -4.08160798e-02 -5.64128995e-01
-6.08910441e-01 6.27660513e-01 9.08731401e-01 -1.70819908e-02
4.64762181e-01 -1.90573618e-01 4.55218703e-02 -3.07824135e-01
-6.07037961e-01 -2.20156655e-01 4.19546694e-01 7.62244463e-01
-6.53045550e-02 -3.07286493e-02 5.73133767e-01 3.26683193e-01
-4.21858072e-01 -1.98871881e-01 -2.47515365e-02 6.52055562e-01
-1.22647233e-01 -7.76073873e-01 -4.19474959e-01 1.31720230e-01
1.76491380e-01 4.84262928e-02 -2.79162407e-01 1.00307369e+00
2.58524984e-01 5.28136015e-01 4.58991051e-01 -4.20132875e-01
6.32555008e-01 -2.12163091e-01 5.66109478e-01 -5.50002992e-01
-2.29789391e-01 -3.10997069e-01 -2.87091941e-01 -6.73729181e-01
-6.66622281e-01 -4.95789230e-01 -1.49290764e+00 -4.07539517e-01
-2.29195550e-01 2.42830947e-01 9.23873246e-01 9.15670395e-01
4.51270640e-01 4.01829958e-01 5.83236217e-01 -1.19247377e+00
-3.46198171e-01 -7.49398470e-01 -2.93423414e-01 5.66469908e-01
5.74464202e-01 -9.44597483e-01 -3.00297529e-01 5.59075968e-03] | [8.229923248291016, -3.2383687496185303] |
7a54a15d-ad16-4bd5-8295-4751c7f473c9 | learning-analytical-posterior-probability-for | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Fang_Learning_Analytical_Posterior_Probability_for_Human_Mesh_Recovery_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Fang_Learning_Analytical_Posterior_Probability_for_Human_Mesh_Recovery_CVPR_2023_paper.pdf | Learning Analytical Posterior Probability for Human Mesh Recovery | Despite various probabilistic methods for modeling the uncertainty and ambiguity in human mesh recovery, their overall precision is limited because existing formulations for joint rotations are either not constrained to SO(3) or difficult to learn for neural networks. To address such an issue, we derive a novel analytical formulation for learning posterior probability distributions of human joint rotations conditioned on bone directions in a Bayesian manner, and based on this, we propose a new posterior-guided framework for human mesh recovery. We demonstrate that our framework is not only superior to existing SOTA baselines on multiple benchmarks but also flexible enough to seamlessly incorporate with additional sensors due to its Bayesian nature. The code is available at https://github.com/NetEase-GameAI/ProPose. | ['Weidong Zhang', 'Jiefeng Li', 'Qing Shuai', 'Yinghui Fan', 'Kang Chen', 'Qi Fang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['human-mesh-recovery'] | ['computer-vision'] | [ 1.13753617e-01 2.26136804e-01 -4.39358771e-01 -1.68178007e-01
-1.12378502e+00 -1.98053703e-01 1.94229826e-01 -3.98615748e-01
1.68966223e-02 6.27355635e-01 6.15350425e-01 4.34370972e-02
-3.06672156e-01 -6.53978646e-01 -8.52408350e-01 -4.69483614e-01
5.69519885e-02 8.03688049e-01 1.79080844e-01 5.83637680e-04
-1.95961595e-01 4.82336938e-01 -1.20961535e+00 -1.64010331e-01
4.65779096e-01 6.35071754e-01 -1.06971040e-01 5.02212226e-01
6.21639729e-01 3.79913181e-01 -1.24824665e-01 -1.82310462e-01
1.95998669e-01 -7.76630193e-02 -6.49469435e-01 -1.49284214e-01
4.46697533e-01 -6.60433054e-01 -7.20537305e-01 8.33487332e-01
7.00484216e-01 1.51665539e-01 8.68669808e-01 -9.96312976e-01
-3.97456735e-01 6.14644706e-01 -7.35497296e-01 -1.23093940e-01
5.31745374e-01 -1.30584881e-01 1.00003576e+00 -9.66896832e-01
6.78284883e-01 1.27937961e+00 8.46065342e-01 5.96378326e-01
-1.19432712e+00 -6.92245007e-01 9.99565609e-03 -1.78077340e-01
-1.63152349e+00 -7.00589001e-01 8.33343029e-01 -4.26028967e-01
4.45190519e-01 2.13062793e-01 7.93267429e-01 1.45080554e+00
3.80592912e-01 1.00737250e+00 8.26334536e-01 -6.08564802e-02
1.52103126e-01 -5.92958212e-01 3.17930393e-02 8.02317142e-01
3.06195289e-01 2.38647759e-01 -8.32869887e-01 -3.96483272e-01
1.20118237e+00 -4.72466499e-02 -1.67026430e-01 -5.83671749e-01
-1.27962494e+00 6.23543441e-01 3.61208022e-01 -2.94750273e-01
-4.34332907e-01 7.40511715e-01 1.41824707e-01 -4.56293613e-01
3.93941224e-01 -2.23160721e-02 -2.86113650e-01 -2.01090723e-01
-1.13809454e+00 6.98818088e-01 7.79812634e-01 9.64167774e-01
4.19333965e-01 -6.06234185e-02 -8.99228454e-02 8.76031280e-01
7.79682934e-01 9.19209242e-01 -1.74629733e-01 -1.42808175e+00
1.22525819e-01 -9.88032222e-02 3.88295241e-02 -1.24188364e+00
-6.05654418e-01 -5.07160068e-01 -9.20854867e-01 1.01666272e-01
6.26009703e-01 -2.06331581e-01 -1.08132219e+00 1.88274360e+00
4.97826725e-01 3.26344699e-01 -5.22455692e-01 1.06273091e+00
8.61443758e-01 2.96404362e-01 -1.81651965e-01 1.93559542e-01
1.41636240e+00 -7.44278431e-01 -6.74264491e-01 -4.79565471e-01
-1.65213540e-01 -7.42802083e-01 7.49959469e-01 5.54192185e-01
-1.19752026e+00 -1.82152390e-01 -9.42017913e-01 -1.53973356e-01
2.94166386e-01 2.25243028e-02 7.44521499e-01 5.67041278e-01
-9.29222524e-01 6.02991283e-01 -1.46344125e+00 -2.88179487e-01
3.83023351e-01 3.48875761e-01 -1.76474944e-01 -3.61425802e-02
-1.04178882e+00 6.84085310e-01 2.21322589e-02 3.76251459e-01
-8.64986598e-01 -4.97634888e-01 -9.64294434e-01 -2.50595093e-01
4.71716970e-01 -1.14548397e+00 1.31748617e+00 -6.06396906e-02
-1.77307212e+00 4.79445159e-01 -2.71077722e-01 -2.38930449e-01
6.43551052e-01 -5.39331615e-01 1.87535971e-01 3.21728945e-01
1.02693304e-01 8.51902425e-01 7.80489087e-01 -1.27895534e+00
-1.20733477e-01 -3.22987527e-01 -4.54720370e-02 1.07197031e-01
6.37978539e-02 -1.76361859e-01 -9.60340977e-01 -1.12499058e+00
5.40920377e-01 -1.44505250e+00 -4.73868519e-01 3.33393157e-01
-7.61646330e-01 8.99905935e-02 2.57229775e-01 -9.45419729e-01
1.00800788e+00 -1.66195798e+00 4.03915048e-01 5.76107204e-01
2.47650579e-01 -2.72417545e-01 1.26021937e-01 1.87934443e-01
2.76428580e-01 -1.20621935e-01 -3.72455597e-01 -6.26125574e-01
1.19305059e-01 3.83835912e-01 -1.51369825e-01 6.96884632e-01
-1.51986822e-01 6.72426939e-01 -8.65068555e-01 -5.99858880e-01
4.12236452e-01 9.76241469e-01 -7.75862336e-01 -1.03763141e-01
-7.05778822e-02 6.36853099e-01 -6.47731125e-01 1.07895017e+00
6.40068293e-01 -3.97904485e-01 3.39533418e-01 -5.18914342e-01
4.93438333e-01 7.79464245e-02 -1.57970393e+00 2.02533007e+00
-5.65206446e-02 4.04064730e-02 2.42408380e-01 -5.73214054e-01
5.03511488e-01 4.63645339e-01 9.90064204e-01 -1.99292287e-01
2.47794449e-01 1.97147205e-01 -2.98544019e-01 -1.31284818e-01
3.98053557e-01 -1.36942834e-01 -1.21767901e-01 4.42883134e-01
-5.13648055e-03 -2.07097784e-01 -1.41252786e-01 1.58535689e-01
1.14021146e+00 6.99965239e-01 2.06457183e-01 -2.31393740e-01
-1.29823655e-01 -3.15035254e-01 9.16454017e-01 6.61385715e-01
-1.95023954e-01 1.22211623e+00 1.37553334e-01 -8.58608051e-04
-8.03145885e-01 -1.66184592e+00 -4.46340054e-01 6.69171333e-01
2.04497471e-01 -6.00648701e-01 -6.55652285e-01 -3.50941092e-01
2.19719931e-01 2.48825178e-01 -5.75561643e-01 7.84069076e-02
-6.11160457e-01 -7.01909900e-01 5.75290859e-01 8.52475405e-01
1.30255390e-02 -5.99882543e-01 -4.10155207e-01 2.44182885e-01
-3.86345714e-01 -1.16808021e+00 -4.79042858e-01 -4.29201312e-02
-9.81596470e-01 -1.10760212e+00 -9.49652910e-01 -1.61927506e-01
4.92008448e-01 1.38912676e-02 1.06605375e+00 -2.96677416e-03
-2.14997694e-01 8.44629288e-01 -8.82218182e-02 -3.16957176e-01
-1.54185861e-01 1.56252980e-01 2.33325049e-01 -4.72730130e-01
-1.44700378e-01 -9.49753523e-01 -1.00213420e+00 3.73447031e-01
-6.81807339e-01 1.22864649e-01 5.31313419e-01 6.59229100e-01
9.20056403e-01 -2.68186241e-01 4.14069235e-01 -7.31779754e-01
3.18939060e-01 -6.63333118e-01 -2.61884481e-01 -8.04508626e-02
-5.01138389e-01 9.85184312e-02 -1.65749118e-01 -3.70043904e-01
-1.00652158e+00 3.50935042e-01 -4.37458754e-01 -4.20735657e-01
-1.32250056e-01 6.35504663e-01 2.99696326e-02 -5.77765377e-03
5.33592045e-01 -2.50427276e-01 1.80248380e-01 -5.96968949e-01
3.79817009e-01 3.08152139e-01 7.94652522e-01 -1.34289110e+00
8.38883281e-01 8.37771893e-01 1.03849046e-01 -6.66168392e-01
-7.95073271e-01 -3.08082700e-01 -5.28724134e-01 -3.11360151e-01
7.92823792e-01 -1.02959776e+00 -5.75425923e-01 3.83068383e-01
-9.96034324e-01 -3.18097174e-01 -1.61123909e-02 6.92737997e-01
-8.58029008e-01 6.79504812e-01 -1.05406499e+00 -8.04954529e-01
-4.62787330e-01 -1.31108975e+00 1.31191373e+00 1.49404302e-01
-5.53312480e-01 -8.28730106e-01 2.13387191e-01 6.60222232e-01
2.86308289e-01 5.49261868e-01 4.51931715e-01 2.87988540e-02
-4.97841984e-01 -2.26423234e-01 -1.13581441e-01 1.61353990e-01
3.69885229e-02 1.74104258e-01 -7.35146046e-01 -2.60147274e-01
-2.40477115e-01 -2.42102370e-01 6.02189600e-01 8.51929486e-01
1.01595426e+00 1.44476909e-02 -5.49484134e-01 6.36866093e-01
1.01684952e+00 -4.80625659e-01 6.18289828e-01 2.22004235e-01
8.18707049e-01 3.29739064e-01 5.24473727e-01 6.27650380e-01
8.01700115e-01 9.53157425e-01 4.64887828e-01 3.10962349e-02
-3.93515736e-01 -2.14497343e-01 2.24715427e-01 1.00813568e+00
-5.21752536e-01 -8.69534239e-02 -1.03892303e+00 6.60184860e-01
-1.98822665e+00 -6.00026965e-01 -7.29806954e-03 2.02666593e+00
1.05373096e+00 2.10498229e-01 1.59421608e-01 -2.23356798e-01
4.52372700e-01 1.27830520e-01 -5.23205996e-01 2.99374402e-01
1.36679277e-01 4.77693647e-01 6.43909633e-01 4.87687171e-01
-1.02220953e+00 6.94394588e-01 7.29760408e+00 7.86625922e-01
-6.63586080e-01 2.18449786e-01 1.86757639e-01 -2.27872953e-01
-3.31491232e-01 1.37996450e-02 -7.69235432e-01 2.06569478e-01
5.98637760e-01 2.39153668e-01 4.16637719e-01 5.94098330e-01
1.63492933e-01 -1.96046636e-01 -1.01139486e+00 9.04936552e-01
-4.25712503e-02 -1.04212606e+00 -2.12764144e-01 2.45431855e-01
5.89176893e-01 3.54713559e-01 -2.12549567e-02 -8.98023173e-02
6.16012573e-01 -9.53468025e-01 9.94610131e-01 8.43578219e-01
4.98342425e-01 -5.86903930e-01 3.89371812e-01 2.14334011e-01
-1.25099587e+00 4.35881764e-01 -1.78254753e-01 -3.42030898e-02
5.94895184e-01 8.40319812e-01 -6.06977940e-01 6.84676468e-01
7.76037037e-01 7.08281994e-01 -2.77169764e-01 1.08313727e+00
-5.31671941e-01 7.56042004e-01 -9.34887707e-01 4.91055340e-01
-3.82418752e-01 1.70582682e-02 7.36823916e-01 9.13293958e-01
5.01016140e-01 -7.48625919e-02 3.74448776e-01 7.76001751e-01
3.72095555e-02 -8.94004256e-02 -3.90516639e-01 3.15374166e-01
7.38681972e-01 1.05764997e+00 -9.51917470e-01 6.11017644e-02
-2.67934799e-01 7.33424008e-01 2.32393712e-01 4.95095909e-01
-1.01293433e+00 1.79050177e-01 5.89419186e-01 4.05635446e-01
9.99456570e-02 -6.16075337e-01 -4.34085488e-01 -1.20365703e+00
1.48112327e-01 -8.52966130e-01 4.80225265e-01 -7.33603477e-01
-1.43930387e+00 1.99762136e-02 5.14197469e-01 -1.06116390e+00
-3.62197936e-01 -7.17428386e-01 -2.17940003e-01 4.37883139e-01
-1.07248199e+00 -1.55316830e+00 -1.72394246e-01 5.38348198e-01
1.63096920e-01 3.36391747e-01 8.22049975e-01 5.04585862e-01
-4.92291600e-01 5.86935759e-01 -1.20878130e-01 1.46785691e-01
8.65081370e-01 -1.10254645e+00 3.36306125e-01 7.85195529e-01
1.20401599e-01 8.50078464e-01 9.94357049e-01 -1.00364149e+00
-1.67137289e+00 -5.30929744e-01 6.06699996e-02 -6.39314234e-01
5.79066336e-01 -2.29257226e-01 -6.09396398e-01 9.93695915e-01
-1.29773289e-01 8.93262774e-02 6.53207779e-01 5.62497318e-01
-3.01894754e-01 1.55518889e-01 -9.36412990e-01 6.86231613e-01
1.21784055e+00 -3.75943482e-01 -4.69229549e-01 3.32891107e-01
3.89252901e-01 -1.04357910e+00 -1.39629519e+00 9.09823596e-01
9.43212688e-01 -6.79373384e-01 1.42475545e+00 1.49523839e-02
3.81366104e-01 -4.72329378e-01 -4.52504516e-01 -8.16444099e-01
-2.41057396e-01 -5.24532020e-01 -6.14036083e-01 9.36882436e-01
1.49651080e-01 -4.96479094e-01 1.12509286e+00 6.53754890e-01
-9.78161767e-02 -8.70464027e-01 -1.19561446e+00 -5.74365199e-01
2.26782218e-01 -8.12199533e-01 4.20069337e-01 7.46438205e-01
-2.23085910e-01 1.90963596e-02 -8.88488352e-01 5.26027858e-01
1.05460727e+00 -1.55203775e-01 8.73062015e-01 -1.12001014e+00
-7.42629588e-01 -2.75218517e-01 -1.96964458e-01 -1.21893847e+00
4.26365659e-02 -7.22217560e-01 3.67283434e-01 -1.73725891e+00
3.67803186e-01 -5.31075120e-01 -2.08377570e-01 5.81264079e-01
-2.51593534e-02 4.97459263e-01 -1.28685772e-01 3.26009512e-01
-3.59952390e-01 5.74921131e-01 1.08425319e+00 2.05869116e-02
2.59649813e-01 7.81699941e-02 -5.71254194e-01 1.01484644e+00
5.51852763e-01 -5.41792989e-01 -3.36562365e-01 -5.33805728e-01
4.75834608e-01 2.65694320e-01 6.80654705e-01 -1.25363445e+00
2.70853639e-01 -1.96794599e-01 4.07789975e-01 -8.16663146e-01
5.85000038e-01 -6.27438068e-01 5.80863655e-01 2.63839006e-01
9.59909558e-02 -3.05479646e-01 5.19913435e-02 7.31477201e-01
1.47961020e-01 -1.10306241e-01 6.29644394e-01 3.02673019e-02
-2.50129730e-01 5.77073276e-01 -4.25313801e-01 4.39224802e-02
4.86628175e-01 -5.03088124e-02 -4.65139598e-02 -5.12010038e-01
-1.00937307e+00 1.44948080e-01 6.33047581e-01 3.03841084e-01
5.89381218e-01 -1.44686830e+00 -6.73221767e-01 -2.69840479e-01
-8.11096206e-02 5.02640665e-01 4.90438670e-01 9.86427605e-01
-6.33596718e-01 -2.53682714e-02 -6.76634163e-02 -7.38915741e-01
-9.56331074e-01 1.72966719e-01 3.87310684e-01 -2.73078769e-01
-8.83987844e-01 7.42799580e-01 -2.89793070e-02 -7.31570303e-01
3.60978454e-01 -3.14027816e-01 3.46767157e-01 -4.62283194e-01
1.82125047e-01 4.51245308e-01 -1.55196473e-01 -7.37017989e-01
-4.68495220e-01 7.46450305e-01 2.40466222e-01 -3.61449510e-01
1.26611114e+00 -1.45235837e-01 3.87974046e-02 2.33486533e-01
7.02973306e-01 4.58681017e-01 -1.35358143e+00 -4.01872396e-01
-1.50055215e-01 -2.90665418e-01 3.40874121e-02 -5.43397129e-01
-1.31425428e+00 4.59558696e-01 4.70297158e-01 -5.00814915e-01
8.12901676e-01 2.26711318e-01 9.89045620e-01 3.09741080e-01
6.50774479e-01 -1.00478542e+00 -4.94646318e-02 3.79869819e-01
9.54931974e-01 -1.02614832e+00 6.24799311e-01 -8.13449383e-01
-2.25743547e-01 1.00276482e+00 3.51569593e-01 -3.97682190e-01
1.20761609e+00 7.01437473e-01 7.81580806e-02 -4.50634897e-01
-2.91544348e-01 7.09539801e-02 5.78651011e-01 6.86779916e-01
5.98968327e-01 1.87383115e-01 -1.55211464e-01 6.85091853e-01
-2.97970414e-01 2.16629699e-01 2.45588273e-01 1.15361691e+00
-1.18995287e-01 -1.17424154e+00 -5.17251194e-01 3.42688411e-01
-6.24458492e-01 1.52279973e-01 1.34905532e-01 6.08051181e-01
-1.69243962e-01 6.50777996e-01 -3.22390527e-01 -3.76996905e-01
2.23234907e-01 -3.77943754e-01 7.55957842e-01 -6.10831797e-01
1.97217643e-01 4.49984998e-01 2.40962386e-01 -7.91220427e-01
-5.48259139e-01 -9.12320495e-01 -1.38848782e+00 -2.08592951e-01
-2.19108075e-01 -3.21278334e-01 4.57481086e-01 9.00905788e-01
2.92349249e-01 4.87884223e-01 3.33746187e-02 -1.40581751e+00
-4.41388935e-01 -8.41824293e-01 -5.63542187e-01 1.94420516e-01
-8.56278166e-02 -1.25706100e+00 -9.25845206e-02 -1.07247993e-01] | [7.086912631988525, -1.0712344646453857] |
e8af4cb2-5f1c-40f8-9769-aa6e99966992 | a-novel-heap-based-pilot-assignment-for-full | 2007.04787 | null | https://arxiv.org/abs/2007.04787v1 | https://arxiv.org/pdf/2007.04787v1.pdf | A Novel Heap-based Pilot Assignment for Full Duplex Cell-Free Massive MIMO with Zero-Forcing | This paper investigates the combined benefits of full-duplex (FD) and cell-free massive multiple-input multipleoutput (CF-mMIMO), where a large number of distributed access points (APs) having FD capability simultaneously serve numerous uplink and downlink user equipments (UEs) on the same time-frequency resources. To enable the incorporation of FD technology in CF-mMIMO systems, we propose a novel heapbased pilot assignment algorithm, which not only can mitigate the effects of pilot contamination but also reduce the involved computational complexity. Then, we formulate a robust design problem for spectral efficiency (SE) maximization in which the power control and AP-UE association are jointly optimized, resulting in a difficult mixed-integer nonconvex programming. To solve this problem, we derive a more tractable problem before developing a very simple iterative algorithm based on inner approximation method with polynomial computational complexity. Numerical results show that our proposed methods with realistic parameters significantly outperform the existing approaches in terms of the quality of channel estimate and SE. | ['Oh-Soon Shin', 'Björn Ottersten', 'Symeon Chatzinotas', 'Shree Krishna Sharma', 'Octavia A. Dobre', 'Van-Dinh Nguyen', 'Hieu V. Nguyen'] | 2020-07-08 | null | null | null | null | ['robust-design'] | ['miscellaneous'] | [-5.78833483e-02 2.31355369e-01 -3.33288983e-02 4.65696752e-01
-7.19490111e-01 -5.25217831e-01 -2.10200876e-01 -4.26387399e-01
-4.46513742e-02 1.57944202e+00 -2.56229222e-01 -8.39436948e-01
-3.93557042e-01 -6.61971688e-01 -3.49535257e-01 -1.14107680e+00
-1.40249699e-01 -1.08510643e-01 -3.07383150e-01 2.57371664e-02
-1.08764179e-01 3.00964445e-01 -6.75193369e-01 -7.11195946e-01
1.27732110e+00 1.31974900e+00 3.60634416e-01 5.32358229e-01
2.62793511e-01 4.16815937e-01 -6.52479351e-01 -1.69426560e-01
4.06279504e-01 -5.07916093e-01 -1.08542211e-01 4.55850154e-01
-3.81902456e-01 -2.58250684e-01 -5.53986549e-01 5.78491747e-01
1.12961555e+00 1.55736178e-01 5.60519218e-01 -1.33120430e+00
-3.65211219e-01 7.35910609e-02 -9.75206435e-01 2.55840272e-01
1.09842516e-01 -1.87627912e-01 7.21365988e-01 -1.08768320e+00
1.31689206e-01 9.98197794e-01 5.91001153e-01 -1.03874639e-01
-6.76237941e-01 -6.44430757e-01 -3.13018769e-01 -4.01607379e-02
-1.92700553e+00 -6.44079506e-01 1.52445942e-01 -7.34027997e-02
4.33590889e-01 4.82264340e-01 5.70578039e-01 -1.44920498e-01
-1.14170261e-01 5.20999908e-01 6.17662787e-01 -8.41032147e-01
2.36501127e-01 1.07863538e-01 -1.76039711e-01 9.22785342e-01
1.00256550e+00 -4.26182389e-01 -2.96046883e-02 -4.83645439e-01
1.17564642e+00 -3.15936655e-01 -7.14761078e-01 -1.76269948e-01
-1.17372441e+00 4.72999007e-01 1.01308748e-01 3.40934038e-01
-6.77637994e-01 6.83759153e-02 -4.98627335e-01 3.12232345e-01
2.40559369e-01 4.03689221e-02 -6.76595420e-02 3.20917398e-01
-9.11112309e-01 8.06052312e-02 8.02232742e-01 1.40323615e+00
3.85793984e-01 2.84555286e-01 -6.71534359e-01 8.13178360e-01
6.45609438e-01 1.09528852e+00 -5.74127436e-01 -1.01711500e+00
8.96284580e-01 -9.32251960e-02 7.91822970e-01 -9.23856854e-01
-3.70619535e-01 -1.59554768e+00 -1.08519471e+00 -6.38271213e-01
3.70799661e-01 -1.19086480e+00 -2.61322647e-01 1.51439202e+00
2.83246905e-01 4.38317090e-01 8.48650038e-02 8.03776562e-01
1.02413416e-01 8.46729994e-01 -3.68951440e-01 -1.08035660e+00
1.07799971e+00 -8.34181249e-01 -1.19103849e+00 -8.97244290e-02
7.51315951e-01 -5.95008135e-01 2.02551946e-01 2.59166211e-01
-1.46341503e+00 -1.65974289e-01 -1.42129958e+00 5.44473827e-01
3.59176040e-01 7.22040236e-01 3.54036003e-01 1.34921741e+00
-7.68620074e-01 -2.48056799e-01 -9.86967757e-02 -1.40473574e-01
5.50546288e-01 4.87782896e-01 1.75073326e-01 -2.62883365e-01
-1.01079583e+00 4.78236794e-01 3.27288993e-02 2.52826363e-01
3.38703208e-02 -7.26566911e-01 -4.94092524e-01 4.25127894e-01
5.21015286e-01 -1.24079347e+00 9.51866567e-01 -1.18186735e-01
-1.34653986e+00 6.44725412e-02 -3.94719154e-01 -1.73732385e-01
1.77672967e-01 4.55850586e-02 -4.96062100e-01 1.12431392e-01
1.12566918e-01 -3.15998614e-01 3.88786942e-01 -1.31594741e+00
-1.04602015e+00 -3.46938908e-01 1.54987872e-01 3.60025346e-01
-3.91029119e-01 -2.38410711e-01 -4.69738513e-01 -7.77448833e-01
1.08148679e-01 -8.58353496e-01 -5.28854311e-01 -2.16569871e-01
-5.39537191e-01 3.56838405e-01 4.10568416e-01 -6.28621161e-01
1.96960521e+00 -2.21262980e+00 1.77186266e-01 5.66301107e-01
3.30329426e-02 3.31659317e-01 3.69139880e-01 5.46940684e-01
4.92863774e-01 1.79893449e-02 -3.26065496e-02 -2.80148655e-01
2.76221093e-02 5.01684472e-02 4.92896102e-02 5.96137226e-01
-3.12214911e-01 5.85791290e-01 -5.88876486e-01 -1.88639820e-01
9.60001200e-02 -8.10720995e-02 -7.04202592e-01 1.46894857e-01
3.42508405e-01 5.95888972e-01 -8.26556444e-01 7.52099991e-01
1.26053190e+00 -4.95745629e-01 2.30894566e-01 -1.92522496e-01
-4.19715494e-01 -4.82164055e-01 -1.67890155e+00 1.27023065e+00
-7.24464178e-01 8.31491053e-02 6.71222150e-01 -1.15524364e+00
3.32858920e-01 6.39885306e-01 4.55168098e-01 -7.34696090e-01
3.57504755e-01 6.14568651e-01 -3.13847885e-03 -3.30850691e-01
2.46819183e-01 -1.78043216e-01 -1.16478369e-01 4.12797034e-01
-4.09581736e-02 5.84707975e-01 2.12246150e-01 1.70440912e-01
9.57980931e-01 -5.41662633e-01 8.92792523e-01 -4.96925205e-01
9.43417251e-01 -5.68697095e-01 7.52426386e-01 8.15519273e-01
-5.94482981e-02 8.36997200e-03 1.28149435e-01 3.83281201e-01
-7.88962841e-01 -9.82235610e-01 -2.47903049e-01 4.09311026e-01
5.93190432e-01 7.02766255e-02 -3.63286704e-01 1.35888129e-01
3.00633103e-01 7.29713619e-01 1.44434255e-02 2.73137271e-01
-1.11742564e-01 -1.09912527e+00 2.95660049e-01 -8.55925959e-03
7.76419938e-01 3.38705719e-01 4.76430319e-02 5.31261206e-01
-3.14684480e-01 -1.23864007e+00 -4.95331198e-01 -1.05959445e-01
-2.99268126e-01 -6.97808206e-01 -1.42163491e+00 -6.29734278e-01
7.90534258e-01 1.01826799e+00 4.70383644e-01 -5.31920120e-02
-6.87494641e-03 5.32725990e-01 -1.15922287e-01 -3.10910881e-01
3.54570895e-01 7.19945654e-02 2.79739797e-01 4.17088330e-01
-3.49484146e-01 -5.90851009e-01 -7.52196670e-01 6.42298102e-01
-1.58052385e-01 -1.68345451e-01 7.10365772e-01 7.06088305e-01
3.83695632e-01 3.11935455e-01 1.34165287e+00 -3.14260781e-01
8.09344947e-01 -9.63177443e-01 -3.80742222e-01 5.49650848e-01
-4.13329117e-02 -1.75486609e-01 5.21018565e-01 3.33452880e-01
-1.19768548e+00 -1.87596515e-01 -4.28176709e-02 -7.78589323e-02
5.96122563e-01 5.60776055e-01 -4.71643388e-01 -7.85943091e-01
3.44199657e-01 2.46496007e-01 -3.90043020e-01 -5.56109324e-02
3.57946962e-01 1.20392239e+00 2.51329809e-01 -6.44217551e-01
1.14078081e+00 3.21838617e-01 5.12082815e-01 -1.20200753e+00
-8.61349761e-01 -6.86246514e-01 -2.03762054e-01 -2.45918080e-01
1.72090128e-01 -1.52101588e+00 -7.49258101e-01 4.48637418e-02
-9.99064624e-01 2.23189130e-01 2.88498729e-01 7.34528542e-01
-3.72623503e-01 7.27311015e-01 -1.82077974e-01 -1.47704554e+00
-2.53832698e-01 -4.69871104e-01 6.69979155e-01 4.75105613e-01
3.81245911e-01 -7.53028929e-01 -3.29770386e-01 3.04216683e-01
6.18596435e-01 3.13751459e-01 5.34542143e-01 3.52939069e-01
-8.16415548e-01 -2.14537960e-02 -4.36587930e-01 -1.19485758e-01
3.21938813e-01 -6.60191655e-01 -5.67476511e-01 -6.28855646e-01
-1.67552173e-01 4.40716684e-01 5.73060922e-02 7.81878293e-01
7.88330734e-01 -2.97797531e-01 -8.61679614e-01 5.63130021e-01
1.57957411e+00 3.20130408e-01 7.01715231e-01 2.52044620e-03
2.34707683e-01 -3.84660438e-02 1.09785688e+00 1.18401074e+00
4.31475043e-01 7.77498126e-01 3.46463658e-02 -2.28687048e-01
3.81015599e-01 3.96503627e-01 -2.40527794e-01 7.89425492e-01
-2.88686961e-01 -9.61735845e-01 -1.64941117e-01 4.41332370e-01
-2.02486324e+00 -7.64595211e-01 -4.85752076e-01 2.42449713e+00
2.12026432e-01 -2.67891973e-01 1.86829552e-01 3.78814936e-01
9.00109053e-01 -4.28109348e-01 -3.57048362e-01 1.35797977e-01
-4.08876836e-01 9.25624743e-03 1.09920704e+00 3.03386211e-01
-8.87532651e-01 1.10681638e-01 5.30630875e+00 1.12417293e+00
-5.21583200e-01 2.39498362e-01 6.06200278e-01 2.01572683e-02
-3.18387479e-01 -3.27328622e-01 -5.58240891e-01 6.02727830e-01
7.47949004e-01 -5.80052018e-01 3.45676214e-01 2.64890671e-01
8.59727383e-01 -3.91851157e-01 -2.88607538e-01 1.40041351e+00
-3.24637592e-01 -1.40290928e+00 -3.57195050e-01 5.00318706e-01
8.56313467e-01 -7.04181015e-01 -1.88623592e-01 1.41093254e-01
-3.54059577e-01 -4.71289009e-01 3.81773025e-01 7.31711388e-01
7.94854820e-01 -9.40364420e-01 6.68963373e-01 3.43401402e-01
-1.50484097e+00 -7.07770288e-01 -2.23155171e-01 -2.77592480e-01
8.66196036e-01 1.17854309e+00 -2.13050246e-01 1.32332301e+00
-5.61373308e-02 8.92183930e-02 3.47867757e-01 1.78440261e+00
1.02519795e-01 4.47210222e-01 -5.79539418e-01 6.98494837e-02
1.17902763e-01 -4.16830212e-01 6.50384307e-01 8.71395171e-01
1.32338834e+00 6.36932313e-01 1.09708026e-01 5.41886628e-01
-3.84544209e-02 4.63066280e-01 -2.80612409e-01 2.10888639e-01
1.15864122e+00 1.25806594e+00 -2.61163324e-01 -7.07256198e-02
-8.52628231e-01 7.88154781e-01 -2.41708815e-01 7.18209267e-01
-9.40713227e-01 -6.95707738e-01 7.35597074e-01 1.06472805e-01
4.47114527e-01 -8.51124048e-01 -7.31781900e-01 -1.03356338e+00
2.13597089e-01 -2.16497287e-01 1.29603118e-01 -4.19953823e-01
-8.25503349e-01 -1.36437416e-01 -6.28851056e-01 -1.55018508e+00
8.12160149e-02 -2.11764812e-01 -5.13993502e-01 1.20310712e+00
-1.77462876e+00 -8.32189858e-01 -1.72604308e-01 4.79263008e-01
-1.90029949e-01 -2.72817194e-01 5.90381503e-01 1.16497099e+00
-9.66673315e-01 8.93978477e-01 7.90490389e-01 -4.98545140e-01
2.84986496e-01 -7.68658280e-01 -3.53162199e-01 1.00285137e+00
-6.96466923e-01 5.87027729e-01 6.04561031e-01 -2.63464481e-01
-1.59175122e+00 -9.46129918e-01 7.04416275e-01 6.13649487e-01
3.10794652e-01 -1.63119391e-01 -7.15049207e-02 1.34237319e-01
9.84115377e-02 5.68042733e-02 1.06138134e+00 -3.30107063e-01
7.52400994e-01 1.23736769e-01 -1.27970505e+00 7.73905158e-01
1.20951438e+00 3.31687704e-02 3.27320755e-01 3.62188637e-01
3.84735703e-01 -4.22626287e-01 -1.06014431e+00 2.63720959e-01
4.35710728e-01 -4.11028773e-01 1.09564447e+00 3.04432437e-02
-5.07668555e-01 -7.78573155e-01 -5.31859577e-01 -1.21546316e+00
-7.58932531e-01 -1.10085523e+00 -7.07762480e-01 1.12647355e+00
3.68296921e-01 -7.35886395e-01 4.66470838e-01 2.86631465e-01
-1.74372673e-01 -8.17746818e-01 -9.94576871e-01 -9.16418195e-01
-3.64740878e-01 6.22244552e-02 2.32905164e-01 6.29658282e-01
2.42638126e-01 6.33434713e-01 -7.78233647e-01 9.07997489e-01
8.36360335e-01 -2.81350285e-01 8.14985633e-01 -1.11922312e+00
-5.74115038e-01 9.37076434e-02 1.00971563e-02 -1.68586254e+00
-3.14648062e-01 -4.91558433e-01 -3.73246342e-01 -1.79070139e+00
-2.43097857e-01 -7.90026963e-01 -9.84257087e-02 -1.86790209e-02
-3.66415769e-01 9.49017256e-02 8.88963416e-02 4.19193096e-02
-5.73113322e-01 7.28178740e-01 1.51104438e+00 2.52696455e-01
-1.38564706e-01 6.33354425e-01 -1.03119481e+00 1.62511151e-02
5.40444791e-01 6.58663958e-02 -5.74753582e-01 -1.42898932e-01
6.70863688e-03 7.69263983e-01 -5.44692688e-02 -1.45870328e+00
1.77589074e-01 -7.30356649e-02 1.50744796e-01 -4.71013010e-01
4.34656829e-01 -1.04599786e+00 1.76115021e-01 4.87406015e-01
6.22718453e-01 -7.70649135e-01 1.53497741e-01 8.89309466e-01
1.57501221e-01 -7.25595653e-02 7.25271285e-01 3.80375952e-01
-2.50726551e-01 4.82696444e-01 -7.98086703e-01 -2.77175367e-01
1.48638070e+00 -3.97027612e-01 -4.01244938e-01 -8.39160800e-01
-7.60884225e-01 7.37078130e-01 -3.40329409e-01 -2.19478175e-01
-1.79142803e-02 -1.41873586e+00 -6.05753958e-01 -2.18596175e-01
-1.81302264e-01 -5.29717982e-01 7.55958915e-01 1.25301385e+00
-2.11519226e-01 1.37858367e+00 3.89926247e-02 -3.37016732e-01
-7.30180264e-01 2.45131515e-02 3.90923560e-01 -9.77382958e-02
7.96596631e-02 6.80057704e-01 1.92615436e-03 7.41382390e-02
-1.57598890e-02 1.84369281e-01 2.27105785e-02 -1.94245249e-01
2.98511535e-01 7.27146327e-01 2.51380000e-02 -4.79328096e-01
-4.11775224e-02 5.55584073e-01 4.18403894e-01 8.21516961e-02
9.13096189e-01 -1.08618402e+00 1.13172069e-01 -3.69735956e-01
8.08812022e-01 5.32718062e-01 -8.79574358e-01 -4.32890117e-01
-5.01612782e-01 -8.18589151e-01 2.34652698e-01 -4.56746757e-01
-8.93251598e-01 2.61778444e-01 4.87632036e-01 3.54679644e-01
1.11958539e+00 -4.93056685e-01 9.93373454e-01 4.83201712e-01
1.07581031e+00 -1.07366729e+00 -3.90641719e-01 3.01388532e-01
4.78807569e-01 -5.56736469e-01 5.22516668e-02 -1.14614880e+00
4.09543738e-02 9.89629030e-01 3.10756415e-01 3.11215401e-01
7.89933145e-01 1.44014597e-01 -5.38577735e-01 1.99134961e-01
-2.43668243e-01 -7.17136502e-01 -6.07802160e-02 4.66349959e-01
4.22375113e-01 1.81558713e-01 -1.12454689e+00 1.06384909e+00
1.82084382e-01 1.91615924e-01 5.92728436e-01 9.12815511e-01
-8.80371630e-01 -1.13391924e+00 -5.92975557e-01 7.23357081e-01
-4.22471136e-01 -3.38096684e-03 4.99090344e-01 4.12758261e-01
2.36138448e-01 1.31161690e+00 -2.25047484e-01 -1.16309272e-02
4.06202883e-01 -5.29566228e-01 6.27589107e-01 -3.45735997e-01
5.27599216e-01 1.90569922e-01 6.44343197e-01 1.63504571e-01
-2.67575920e-01 -4.58477974e-01 -1.21937251e+00 -2.77251601e-01
-7.44336963e-01 4.37240869e-01 3.26041847e-01 1.40036476e+00
6.01497948e-01 8.29128325e-01 1.29334879e+00 -6.01453304e-01
-2.25024834e-01 -5.82590282e-01 -1.16749656e+00 -7.08103001e-01
9.96948704e-02 -8.81348014e-01 -2.99533814e-01 -5.10268986e-01] | [6.181392669677734, 1.4286829233169556] |
9e31b00a-19cf-4a41-8cb8-b8c4b1c114ce | tdnn-a-two-stage-deep-neural-network-for | null | null | https://aclanthology.org/P18-1100 | https://aclanthology.org/P18-1100.pdf | TDNN: A Two-stage Deep Neural Network for Prompt-independent Automated Essay Scoring | Existing automated essay scoring (AES) models rely on rated essays for the target prompt as training data. Despite their successes in prompt-dependent AES, how to effectively predict essay ratings under a prompt-independent setting remains a challenge, where the rated essays for the target prompt are not available. To close this gap, a two-stage deep neural network (TDNN) is proposed. In particular, in the first stage, using the rated essays for non-target prompts as the training data, a shallow model is learned to select essays with an extreme quality for the target prompt, serving as pseudo training data; in the second stage, an end-to-end hybrid deep model is proposed to learn a prompt-dependent rating model consuming the pseudo training data from the first step. Evaluation of the proposed TDNN on the standard ASAP dataset demonstrates a promising improvement for the prompt-independent AES task. | ['Le Sun', 'Ben He', 'Kai Hui', 'Cancan Jin'] | 2018-07-01 | null | null | null | acl-2018-7 | ['automated-essay-scoring'] | ['natural-language-processing'] | [ 7.17706755e-02 -1.42793894e-01 -1.22211568e-01 -7.83119977e-01
-1.11802864e+00 -5.38649678e-01 3.16344917e-01 4.54587638e-01
-6.26478314e-01 5.31255484e-01 3.02743524e-01 -1.32691786e-01
6.62260130e-02 -5.52718461e-01 -1.15510948e-01 -4.28792655e-01
6.30508900e-01 5.68937421e-01 -6.43560067e-02 -4.85935718e-01
6.80446923e-01 -4.98448014e-02 -1.16328442e+00 4.93145138e-01
1.08320999e+00 1.20593226e+00 9.04596001e-02 9.51219201e-01
-1.01621933e-01 8.71746838e-01 -9.37448502e-01 -5.12762070e-01
4.35045660e-01 -5.75798333e-01 -6.71210051e-01 -3.26495737e-01
6.18202925e-01 -5.87140322e-01 -4.45799798e-01 6.51654899e-01
6.04396939e-01 4.35720652e-01 6.99716032e-01 -9.21116710e-01
-9.50558126e-01 6.00258887e-01 -2.96009570e-01 1.08908176e-01
4.73302752e-01 1.53089985e-01 1.57030725e+00 -1.13350010e+00
5.98701276e-02 4.31805849e-01 3.92063320e-01 6.50318146e-01
-7.69094050e-01 -6.87698841e-01 6.49017096e-02 5.16526043e-01
-8.70392919e-01 -2.87211001e-01 1.02988183e+00 -1.37318969e-01
5.55491745e-01 9.81072187e-02 5.41739643e-01 1.25919235e+00
1.38998553e-01 1.16909730e+00 1.26041269e+00 -4.09091562e-01
4.18818414e-01 1.85065389e-01 8.04373205e-01 3.39247257e-01
-3.14606845e-01 2.31005296e-01 -9.38463330e-01 -1.77127942e-01
1.63791731e-01 6.43754005e-03 -2.67293990e-01 1.36977687e-01
-7.54672945e-01 1.01857555e+00 4.78770494e-01 3.00653223e-02
-7.07043827e-01 -3.69032621e-01 3.43944371e-01 7.73240983e-01
5.95714808e-01 9.41000879e-01 -6.19466126e-01 -3.81339163e-01
-1.41946530e+00 5.78481436e-01 1.10524726e+00 2.97142446e-01
4.46808875e-01 -4.82513718e-02 -9.39727485e-01 1.03661633e+00
5.12303896e-02 9.76065397e-02 7.86633670e-01 -4.85863507e-01
7.54629850e-01 8.49338472e-01 4.32203412e-01 -5.44669688e-01
-5.26198864e-01 -6.32697105e-01 -4.90846962e-01 2.08056524e-01
4.78575349e-01 -2.77602673e-01 -6.73611701e-01 1.67039740e+00
2.03855321e-01 -5.33822551e-02 1.20606951e-01 1.23657012e+00
1.10534263e+00 7.32330322e-01 6.80375025e-02 -1.12703539e-01
1.28002560e+00 -1.42488921e+00 -4.97707665e-01 -2.51119047e-01
5.35007954e-01 -6.43149734e-01 1.55198336e+00 7.35258877e-01
-1.13674963e+00 -5.92541397e-01 -1.28547883e+00 -2.96056390e-01
-1.12213060e-01 7.50213623e-01 3.67937535e-01 4.46714282e-01
-1.01656389e+00 3.60409617e-01 -2.73471326e-01 1.25534520e-01
2.48025104e-01 2.52000749e-01 2.41427831e-02 -1.60980914e-02
-1.51149142e+00 8.20537865e-01 -5.57197630e-02 8.01800638e-02
-9.03043032e-01 -1.00046933e+00 -5.93782246e-01 6.18318617e-01
9.97035727e-02 -3.92414123e-01 1.74227011e+00 -1.10028124e+00
-2.04017901e+00 6.34747326e-01 1.21385075e-01 -3.50118130e-01
4.64477628e-01 -4.24395889e-01 -1.21203884e-01 7.94831514e-02
-1.17554426e-01 2.51877636e-01 6.99496686e-01 -8.06001961e-01
-6.69774234e-01 -2.71401495e-01 4.43766922e-01 6.80098653e-01
-8.79028618e-01 -8.26071948e-02 4.25978713e-02 -5.21772027e-01
-3.10928196e-01 -8.05818439e-01 -1.31417453e-01 -3.08267921e-01
-2.52792954e-01 -7.41803169e-01 3.59717578e-01 -7.35787630e-01
1.57312632e+00 -1.78863752e+00 1.35040551e-01 7.77114332e-02
3.10703099e-01 4.27502722e-01 -5.85194290e-01 2.52716988e-01
-1.13903537e-01 -4.39139873e-01 -3.17702405e-02 -5.51458180e-01
5.64620383e-02 -5.98173320e-01 -3.49841565e-01 3.91551526e-03
3.10353160e-01 7.77817726e-01 -9.55762744e-01 3.24952044e-02
-1.91760346e-01 -8.75932947e-02 -4.63221580e-01 9.09605622e-01
-1.54690549e-01 2.64371365e-01 -3.58146131e-01 3.63638073e-01
4.41048443e-01 -2.66433001e-01 -2.90756017e-01 7.57984221e-02
-2.67386977e-02 8.04125130e-01 -7.87721217e-01 1.72529888e+00
-6.60990417e-01 4.03405219e-01 -1.18909948e-01 -7.15019822e-01
1.33661699e+00 4.72984999e-01 4.46930051e-01 -9.72666442e-01
1.55336618e-01 2.32760176e-01 -1.46649880e-02 -3.56976688e-01
9.65052009e-01 -1.05717339e-01 -3.32677573e-01 1.16776669e+00
2.95931660e-02 -1.66650563e-01 2.16030747e-01 4.26924855e-01
1.49467993e+00 -1.06375672e-01 2.53305078e-01 1.04065269e-01
7.19945431e-01 6.18014112e-02 2.82414973e-01 6.28993094e-01
-3.66299689e-01 9.36823070e-01 4.72370982e-01 -6.67467773e-01
-7.66514897e-01 -8.16794634e-01 1.93172455e-01 1.48564601e+00
-1.24665998e-01 -3.38371366e-01 -7.34031796e-01 -1.09456742e+00
-3.22715819e-01 8.44583273e-01 -7.18781173e-01 -4.86006051e-01
-3.61957252e-01 -6.69704318e-01 3.40575784e-01 4.65200424e-01
3.96059483e-01 -1.26964462e+00 -8.69586289e-01 4.02599961e-01
-3.60242128e-01 -6.70079708e-01 -8.81578267e-01 6.24996066e-01
-5.54550111e-01 -1.11127186e+00 -8.92838418e-01 -4.64152545e-01
5.54742157e-01 2.86357731e-01 1.27510238e+00 2.36467779e-01
3.30093235e-01 2.01532245e-01 -5.52133381e-01 -4.03115630e-01
-2.54932255e-01 5.21705091e-01 -2.12451201e-02 8.58537182e-02
7.53456414e-01 -2.92929828e-01 -8.22910845e-01 6.14218749e-02
-6.81065381e-01 2.70852167e-03 6.29022956e-01 1.23397863e+00
1.52724117e-01 -6.04543507e-01 9.94738400e-01 -8.35866332e-01
1.45826626e+00 -5.69599688e-01 -3.24155003e-01 1.49360403e-01
-9.32400048e-01 -3.89299206e-02 1.30755210e+00 -6.19527102e-01
-1.12777889e+00 -7.12194294e-02 -2.61075556e-01 -1.16590962e-01
8.34953785e-02 9.07847762e-01 2.88348973e-01 2.56207585e-01
9.52699244e-01 2.04537168e-01 -1.39437467e-01 -3.41248244e-01
-1.44939706e-01 8.27864945e-01 4.84479457e-01 -6.22319043e-01
6.11033738e-01 -3.33131343e-01 -3.28311712e-01 -1.95877589e-02
-1.67789364e+00 -5.86015046e-01 -5.63101888e-01 -2.65550494e-01
4.07950908e-01 -9.54598486e-01 -5.51868558e-01 5.38249314e-01
-1.15783739e+00 -4.34955359e-01 -3.32460433e-01 3.94748151e-01
-3.16416681e-01 6.34179860e-02 -6.30799294e-01 -7.04379499e-01
-1.04845059e+00 -1.01805425e+00 8.04396212e-01 6.72063887e-01
-6.10756099e-01 -8.63889575e-01 6.17310166e-01 6.41441166e-01
4.36088413e-01 -3.74399811e-01 8.62803519e-01 -1.42825925e+00
-9.19842124e-02 -6.26135409e-01 -1.48746505e-01 6.23384178e-01
1.39115648e-02 -2.97413558e-01 -1.03485298e+00 -1.90386936e-01
2.73186415e-01 -1.09223127e+00 6.52902305e-01 1.85203567e-01
1.18205822e+00 -2.53461838e-01 5.45540035e-01 5.21167517e-01
9.55383480e-01 -1.59764186e-01 2.02319562e-01 3.97021383e-01
4.06592995e-01 3.79772007e-01 9.19151902e-01 5.87521970e-01
6.42868638e-01 6.10304892e-01 1.89619333e-01 1.45436704e-01
6.32453188e-02 -1.25162005e-01 7.83030152e-01 1.11526012e+00
2.94220626e-01 -1.41552702e-01 -6.65241599e-01 4.23285782e-01
-1.65137684e+00 -5.95982194e-01 -2.28555113e-01 2.37543178e+00
1.14060557e+00 5.13914376e-02 3.97236012e-02 2.32180372e-01
2.79545099e-01 3.34576339e-01 -8.06636393e-01 -6.58289969e-01
1.43899783e-01 6.07080936e-01 -2.67223984e-01 2.96273917e-01
-9.24953878e-01 6.12577856e-01 5.65915632e+00 5.05742848e-01
-1.27412522e+00 1.95918441e-01 5.65184712e-01 -4.18513238e-01
-2.04681188e-01 -2.29959249e-01 -6.69520736e-01 5.33851564e-01
1.33929396e+00 -3.06667686e-01 3.88670236e-01 8.49330127e-01
4.02396291e-01 -1.93101820e-02 -1.18079817e+00 7.40899146e-01
2.12018833e-01 -8.43844056e-01 -1.45688966e-01 -3.85852754e-01
6.79557502e-01 -2.13255823e-01 3.00096333e-01 9.04053450e-01
3.80122721e-01 -8.59650612e-01 6.87448442e-01 2.90046453e-01
8.35551679e-01 -8.09987903e-01 8.80463123e-01 5.24320662e-01
-7.70651937e-01 -3.53335649e-01 -5.30127525e-01 -4.27289665e-01
-1.94624797e-01 5.96031904e-01 -1.00384033e+00 4.43276197e-01
2.63064265e-01 5.77095151e-01 -8.12768459e-01 1.06438470e+00
-5.77014804e-01 1.02190578e+00 5.51779531e-02 -3.03568631e-01
4.17585403e-01 -1.85565934e-01 2.33334541e-01 9.73364174e-01
3.39657336e-01 3.52572680e-01 1.55303469e-02 8.03613782e-01
-5.75714469e-01 3.24558645e-01 1.27481101e-02 1.28898859e-01
4.78924721e-01 1.76322329e+00 -1.73596516e-01 -1.42326087e-01
-3.33605647e-01 9.95834291e-01 8.19518209e-01 2.35250339e-01
-7.49412477e-01 -5.13205647e-01 1.04785636e-01 -3.17954928e-01
-3.22357953e-01 1.94518343e-01 -8.28752160e-01 -1.16730893e+00
-4.06226609e-03 -1.22023892e+00 5.72021961e-01 -7.06912875e-01
-1.43381989e+00 6.16724133e-01 -6.52604043e-01 -1.36547220e+00
-3.14540476e-01 -5.19135118e-01 -1.31007349e+00 1.28514647e+00
-1.62703502e+00 -8.33746791e-01 -7.09418774e-01 4.99769479e-01
9.17180896e-01 -4.40512419e-01 9.15953100e-01 6.86504021e-02
-4.30226326e-01 1.14035165e+00 6.43093735e-02 1.43799126e-01
1.24116850e+00 -1.71839142e+00 2.09371164e-01 6.96413577e-01
1.88666284e-01 4.04462755e-01 6.54132426e-01 -2.58624166e-01
-1.12086558e+00 -8.91835928e-01 1.02988350e+00 -5.74381948e-01
4.58623916e-01 -1.44109786e-01 -1.13119447e+00 1.35181278e-01
2.17913702e-01 -3.54703784e-01 8.71001482e-01 4.50547814e-01
-4.29610223e-01 -1.84742853e-01 -9.74771440e-01 5.62368572e-01
2.62026042e-01 -5.25281668e-01 -6.71983182e-01 2.66356796e-01
3.70181739e-01 -6.14620924e-01 -7.16239572e-01 -4.15333509e-02
7.00934887e-01 -7.25077569e-01 6.03911459e-01 -5.42030513e-01
1.13240874e+00 1.57985061e-01 3.55152071e-01 -1.96656942e+00
-3.06258321e-01 -4.37514365e-01 -3.70774209e-01 9.53324437e-01
4.02708381e-01 -5.22196144e-02 1.03408718e+00 8.28147411e-01
-4.28153515e-01 -1.12759626e+00 -8.18292201e-01 -1.99813113e-01
1.93079337e-01 -1.16445705e-01 6.02473199e-01 7.84105241e-01
2.73828179e-01 8.01110089e-01 -4.23201054e-01 -1.96225762e-01
3.72842640e-01 1.59045756e-01 6.80143654e-01 -1.48753822e+00
-2.15839267e-01 -5.15983403e-01 4.31180328e-01 -1.25263155e+00
2.27398708e-01 -8.47722888e-01 3.90519589e-01 -1.42101693e+00
3.84893984e-01 -4.75202024e-01 -7.60348380e-01 4.90353316e-01
-8.89765799e-01 -5.45929447e-02 1.87573850e-01 9.82362404e-02
-6.77552164e-01 8.64376187e-01 1.23530686e+00 -2.97373146e-01
-4.42623943e-01 4.12129939e-01 -1.05219638e+00 2.34533489e-01
7.02159345e-01 -5.96466780e-01 -5.83562493e-01 -2.56236285e-01
3.30932945e-01 4.49490100e-01 -6.51479438e-02 -8.97520483e-01
5.66138268e-01 -2.45281890e-01 3.23183239e-01 -6.27888083e-01
3.43753994e-01 -4.98115242e-01 -7.33890653e-01 1.06921509e-01
-1.13903272e+00 1.95763737e-01 -2.21521348e-01 3.57267380e-01
-2.64370382e-01 -7.00500190e-01 7.52382636e-01 1.64660618e-01
-1.25150591e-01 4.41865414e-01 -1.29128456e-01 6.12309948e-02
5.45964003e-01 -5.62891103e-02 -3.66980880e-01 -6.03023529e-01
-3.13343436e-01 4.45260495e-01 1.56282932e-01 6.76987827e-01
7.88165033e-01 -1.23385906e+00 -1.16740608e+00 3.85621823e-02
2.82447904e-01 4.13333811e-02 4.65961963e-01 7.21257508e-01
-6.39706617e-03 3.28504592e-01 -2.28707016e-01 -9.63053182e-02
-1.10789728e+00 1.08464979e-01 1.99547470e-01 -1.05035639e+00
-1.59577116e-01 9.34184670e-01 1.18762217e-01 -6.43092215e-01
2.05113724e-01 6.48882193e-03 -6.14622891e-01 9.27805752e-02
8.36997449e-01 1.63806498e-01 6.08121395e-01 -6.22948445e-02
2.54155993e-01 -1.78570122e-01 -6.76090240e-01 -1.87399507e-01
1.46526706e+00 5.69516160e-02 2.11354420e-01 5.56593180e-01
8.47953618e-01 5.58514008e-03 -1.46549749e+00 -3.41275990e-01
1.35525063e-01 -2.97829717e-01 2.62952924e-01 -1.46981359e+00
-1.06897688e+00 9.91995096e-01 2.89778978e-01 -8.31307992e-02
1.14442265e+00 -7.70317674e-01 1.01571381e+00 5.07636011e-01
8.28029290e-02 -1.28285491e+00 6.66539371e-01 9.65496838e-01
9.50986087e-01 -1.31461406e+00 -2.03784019e-01 3.91420007e-01
-9.23040867e-01 1.46649134e+00 1.12294793e+00 -2.86038727e-01
1.71260402e-01 -3.69243659e-02 3.64845365e-01 -2.12890521e-01
-1.19349682e+00 4.89324033e-01 5.12510598e-01 1.68371856e-01
7.15453804e-01 5.48044406e-02 -5.34957707e-01 1.58828163e+00
-3.32906127e-01 1.14215881e-01 9.54586327e-01 6.51970804e-01
-2.66423672e-01 -1.05989099e+00 -2.70210952e-01 9.69042420e-01
-4.40959305e-01 -2.63281763e-01 -5.72561085e-01 -1.61951650e-02
-4.80859756e-01 1.03241074e+00 -2.91685462e-01 -6.59378409e-01
4.39863235e-01 4.09998268e-01 1.30130202e-01 -8.24173391e-01
-1.46159518e+00 -7.10509777e-01 -9.54075754e-02 -2.84631252e-01
1.70791060e-01 -5.36606312e-01 -9.58182931e-01 -2.23916575e-01
-3.02748621e-01 5.49661458e-01 6.83708668e-01 1.01181662e+00
1.55273810e-01 6.50359392e-01 1.16431510e+00 -6.52697861e-01
-1.45030308e+00 -1.24506557e+00 -5.51936209e-01 4.71441269e-01
3.45727921e-01 -4.02424812e-01 -5.03060341e-01 -3.12428623e-01] | [11.294940948486328, 9.319923400878906] |
b7d6396f-db4e-4fe7-b331-bcb874e9a4f4 | collaborative-receptive-field-learning | 1402.0170 | null | http://arxiv.org/abs/1402.0170v1 | http://arxiv.org/pdf/1402.0170v1.pdf | Collaborative Receptive Field Learning | The challenge of object categorization in images is largely due to arbitrary
translations and scales of the foreground objects. To attack this difficulty,
we propose a new approach called collaborative receptive field learning to
extract specific receptive fields (RF's) or regions from multiple images, and
the selected RF's are supposed to focus on the foreground objects of a common
category. To this end, we solve the problem by maximizing a submodular function
over a similarity graph constructed by a pool of RF candidates. However,
measuring pairwise distance of RF's for building the similarity graph is a
nontrivial problem. Hence, we introduce a similarity metric called
pyramid-error distance (PED) to measure their pairwise distances through
summing up pyramid-like matching errors over a set of low-level features.
Besides, in consistent with the proposed PED, we construct a simple
nonparametric classifier for classification. Experimental results show that our
method effectively discovers the foreground objects in images, and improves
classification performance. | ['Shu Kong', 'Qiang Yang', 'Zhuolin Jiang'] | 2014-02-02 | null | null | null | null | ['object-categorization'] | ['computer-vision'] | [ 3.25316519e-01 -3.18304598e-01 -7.36753121e-02 -6.97562575e-01
-7.30173290e-01 -4.85718340e-01 3.17459613e-01 -8.22648183e-02
-2.45313376e-01 3.82342875e-01 1.80468373e-02 2.52663583e-01
-6.59585774e-01 -7.81127870e-01 -6.76366806e-01 -9.76736248e-01
-1.93154812e-01 1.23239132e-02 6.95548952e-01 1.55642509e-01
5.27961314e-01 5.08333445e-01 -1.47730601e+00 4.20424163e-01
9.08145428e-01 1.26798522e+00 4.96541113e-01 2.72017181e-01
-1.43593177e-01 8.76698792e-01 -5.58445275e-01 -1.33974895e-01
3.59137297e-01 -3.66200745e-01 -7.34182000e-01 4.03688192e-01
4.28942025e-01 -9.57644582e-02 -2.21052036e-01 1.32821929e+00
2.28278920e-01 3.40061456e-01 8.18863690e-01 -1.33018851e+00
-9.13526058e-01 4.34104234e-01 -7.38313675e-01 4.67861354e-01
8.30364525e-02 -2.36395016e-01 1.07269657e+00 -1.11386752e+00
4.29697871e-01 1.43869412e+00 4.52546507e-01 2.38541454e-01
-1.35258996e+00 -7.29589880e-01 5.29655933e-01 3.32656413e-01
-1.91730058e+00 -2.57625520e-01 9.36201513e-01 -4.02958959e-01
1.00522630e-01 4.10293102e-01 2.54007250e-01 6.94453418e-01
1.99006930e-01 7.39214957e-01 1.26516426e+00 -6.86276779e-02
3.35727125e-01 1.97561786e-01 7.43039101e-02 6.36954904e-01
2.58470088e-01 -5.30970693e-01 -3.52614671e-01 -2.65456676e-01
8.47873747e-01 3.65281343e-01 -4.33782429e-01 -7.49227285e-01
-1.30377352e+00 9.01471555e-01 7.55477667e-01 3.78865540e-01
-1.35396481e-01 -1.85169280e-01 -6.70496598e-02 4.51862216e-02
2.81873792e-01 2.18842048e-02 -3.61568838e-01 7.16798484e-01
-7.38085628e-01 6.59554452e-02 4.56023067e-01 9.46259975e-01
1.10825717e+00 -5.19958794e-01 -3.68412077e-01 9.21035528e-01
5.09747267e-01 3.75565827e-01 2.36767411e-01 -8.81404579e-01
4.15063024e-01 8.10628951e-01 8.46738275e-03 -1.56923556e+00
-3.96892056e-02 -4.01892930e-01 -9.43703473e-01 1.25669166e-01
3.37829709e-01 1.27844527e-01 -6.79736674e-01 1.55379033e+00
5.92082262e-01 3.62720191e-01 -6.41088262e-02 1.03337848e+00
7.19059110e-01 6.92963064e-01 -9.19679552e-02 -3.52520049e-01
1.13060236e+00 -7.96425462e-01 -3.52551460e-01 -8.26554745e-02
2.24464431e-01 -6.11540616e-01 6.72173917e-01 1.70122221e-01
-7.13858187e-01 -8.48793566e-01 -8.59765828e-01 2.33839750e-01
-1.71757117e-01 1.66538239e-01 6.15381837e-01 2.92447746e-01
-8.59347403e-01 6.16085708e-01 -3.63544524e-01 -9.58500989e-03
6.67880356e-01 5.23248255e-01 -1.96887299e-01 -2.39768147e-01
-7.26060808e-01 3.36704284e-01 3.49822313e-01 2.19512925e-01
-8.75000000e-01 -4.67407793e-01 -5.79024911e-01 -1.56477183e-01
4.64961469e-01 -3.10512573e-01 5.97080290e-01 -1.15986538e+00
-1.09155905e+00 7.27466583e-01 -1.39730319e-01 1.71801988e-02
3.24671507e-01 1.80184320e-01 -4.84330326e-01 1.27617493e-01
3.10680479e-01 5.73336482e-01 1.05806172e+00 -1.54294837e+00
-8.67252231e-01 -4.53111172e-01 2.01165825e-01 1.44831955e-01
-3.41832519e-01 2.62647480e-01 -3.75948191e-01 -5.83265305e-01
5.80912769e-01 -5.26531458e-01 -3.00657809e-01 5.63860834e-02
-1.66444644e-01 -6.04172170e-01 8.63727450e-01 -5.55529714e-01
9.93442059e-01 -2.25242186e+00 3.02443087e-01 4.98838842e-01
4.49733764e-01 -1.73340470e-01 -1.41910449e-01 -1.43956497e-01
2.91740805e-01 -9.54444930e-02 -2.98116803e-01 8.70548636e-02
-2.21309900e-01 1.94712058e-01 -3.16673785e-01 6.48625314e-01
3.73109400e-01 7.09351540e-01 -9.30448055e-01 -9.48926210e-01
-4.34738630e-03 3.33437145e-01 -3.88940215e-01 1.76261604e-01
1.77660212e-02 5.81397712e-01 -9.86335576e-01 7.28359699e-01
1.16237199e+00 -4.46382642e-01 7.40365684e-02 -4.12102252e-01
2.51558255e-02 -1.68102726e-01 -1.54718494e+00 1.58553731e+00
-7.18918815e-02 1.98084131e-01 1.62732527e-01 -1.32930708e+00
1.28667486e+00 -3.42682362e-01 4.60765451e-01 -2.54598856e-01
1.28705591e-01 2.64042169e-01 4.68517765e-02 -2.59342939e-01
-2.17106324e-02 2.26991847e-01 6.08886741e-02 1.33489013e-01
1.61090568e-01 3.29186201e-01 2.27742698e-02 -2.46903654e-02
1.03785002e+00 5.79866990e-02 1.12550914e-01 -5.91890216e-01
1.05050933e+00 -3.60945851e-01 9.39261317e-01 8.16542387e-01
-2.34450698e-01 6.00301802e-01 1.41761154e-01 -3.45284253e-01
-6.18543267e-01 -1.23398304e+00 -4.15694952e-01 1.18675661e+00
8.04587901e-01 -9.69287902e-02 -6.42100573e-01 -1.00259113e+00
1.08947180e-01 -4.36125435e-02 -6.20441139e-01 -8.27976465e-02
-6.79683864e-01 -6.63989305e-01 1.35894820e-01 5.67548394e-01
8.11752021e-01 -8.61569285e-01 -1.68133914e-01 1.62637588e-02
-2.73900390e-01 -9.19111490e-01 -6.96083009e-01 5.69454059e-02
-7.80225277e-01 -1.01194882e+00 -9.33173001e-01 -1.16839564e+00
9.54649925e-01 8.02764237e-01 9.22825634e-01 -6.37217164e-02
-1.99177831e-01 2.13506043e-01 -3.88323724e-01 -1.30512074e-01
7.29945600e-02 -2.52323717e-01 6.35682866e-02 7.12065995e-01
4.55540776e-01 -5.32180429e-01 -7.93571353e-01 9.39310491e-01
-7.18969941e-01 -2.71608710e-01 9.31589365e-01 5.83851099e-01
9.97297466e-01 2.67575383e-01 4.05081868e-01 -3.00604582e-01
2.61483103e-01 -5.20164609e-01 -8.10492158e-01 5.81118584e-01
-6.64021671e-02 -5.80490492e-02 5.42023838e-01 -8.36815655e-01
-9.38647330e-01 5.47785126e-02 5.37361503e-01 -5.88802576e-01
-5.46029359e-02 1.05091758e-01 -4.67798948e-01 -4.81794059e-01
3.92325401e-01 3.96271944e-01 -2.72130787e-01 -4.25621659e-01
1.90000743e-01 7.72416353e-01 4.01932985e-01 -4.77790087e-01
1.05349457e+00 6.51768923e-01 1.05147831e-01 -6.09874308e-01
-1.07013464e+00 -5.97346127e-01 -6.75907612e-01 -2.58532912e-01
8.95337343e-01 -8.72992337e-01 -6.78814650e-01 1.96398318e-01
-1.05515873e+00 1.88678518e-01 -3.17371525e-02 7.17816591e-01
-3.48051548e-01 4.31691736e-01 -1.95185959e-01 -8.23877156e-01
-2.30595972e-02 -9.81266081e-01 1.15957749e+00 4.87813205e-01
2.68905967e-01 -6.99025035e-01 -1.89063892e-01 3.06237370e-01
7.23097101e-02 1.30230322e-01 6.66630626e-01 -5.20697355e-01
-1.09488845e+00 1.03052482e-01 -6.33978546e-01 5.60511351e-01
3.72463018e-01 -1.71331540e-01 -9.18928564e-01 -3.29460710e-01
2.06223875e-01 -5.17070815e-02 9.45795417e-01 5.05387187e-01
1.64204907e+00 -1.14565104e-01 -5.61440349e-01 8.33184421e-01
1.30129600e+00 3.56400400e-01 3.61997306e-01 8.20804089e-02
7.13842392e-01 6.39728665e-01 8.68316352e-01 2.58809060e-01
8.96905214e-02 5.00782847e-01 2.17903569e-01 5.92895746e-02
6.90455884e-02 -1.12591691e-01 3.07444334e-01 8.68648469e-01
-1.99754879e-01 -9.47463233e-03 -5.38346469e-01 2.01035976e-01
-1.88496721e+00 -8.38709772e-01 8.57256874e-02 2.25110912e+00
4.76952314e-01 5.48039861e-02 2.36532446e-02 -8.11269209e-02
1.32410038e+00 2.45887768e-02 -5.59463263e-01 2.72351861e-01
-2.68235356e-01 -7.46855289e-02 5.08023977e-01 1.62589088e-01
-1.31520283e+00 5.69222093e-01 5.80482674e+00 1.23964012e+00
-8.58875990e-01 1.07474597e-02 6.36324286e-01 3.22276920e-01
-6.38338551e-02 1.27742112e-01 -1.11238265e+00 6.36838555e-01
1.30410060e-01 -1.09826684e-01 4.29146618e-01 8.79366875e-01
-4.33197469e-02 6.85142949e-02 -9.48117137e-01 1.05442858e+00
2.41178915e-01 -1.08521760e+00 9.94176939e-02 8.96480903e-02
8.12791109e-01 -1.83039099e-01 1.93123072e-02 1.81784555e-01
1.15197793e-01 -8.58120799e-01 5.99995434e-01 6.48834407e-01
4.80595767e-01 -7.14060187e-01 5.84516704e-01 3.45886558e-01
-1.61424577e+00 -3.07902992e-01 -1.03828931e+00 2.50409603e-01
-3.69232625e-01 5.82474887e-01 -4.54332173e-01 5.30392408e-01
9.62007642e-01 7.42806733e-01 -7.41980731e-01 1.10739577e+00
1.42045766e-01 2.61194944e-01 -1.85394794e-01 -1.62417144e-01
2.48476285e-02 -5.18245637e-01 3.16642404e-01 1.00072110e+00
2.97840476e-01 1.12179287e-01 4.60594058e-01 1.05119240e+00
-1.36058554e-01 3.88053268e-01 -5.08974969e-01 3.04607749e-01
5.74120820e-01 1.58268023e+00 -1.13680232e+00 -1.44335747e-01
-4.90876257e-01 1.03113866e+00 3.42542112e-01 2.54218370e-01
-1.08459091e+00 -5.65219939e-01 6.85001373e-01 -1.60387367e-01
5.72669744e-01 -1.04521833e-01 8.21720883e-02 -1.14072359e+00
4.21357393e-01 -7.04813004e-01 3.60034645e-01 -6.17811143e-01
-1.57519984e+00 3.23372841e-01 3.94884124e-02 -1.49613976e+00
5.84491551e-01 -6.14991009e-01 -6.28213286e-01 7.27166295e-01
-1.39772117e+00 -1.13251710e+00 -6.22365832e-01 1.00402915e+00
3.56854022e-01 -1.13322876e-01 3.61211956e-01 2.09136531e-01
-2.62300134e-01 3.71407688e-01 2.64746189e-01 3.27776343e-01
5.55508196e-01 -1.09873343e+00 -1.51896290e-02 7.03434289e-01
2.48037070e-01 5.91936827e-01 1.64736047e-01 -5.31931818e-01
-1.32486773e+00 -1.29681838e+00 6.02487743e-01 -3.87169868e-01
5.73306322e-01 -5.07056415e-01 -1.05111444e+00 4.31836665e-01
-3.40026975e-01 4.54086185e-01 5.86683035e-01 -8.39307904e-02
-4.12662119e-01 -4.73318398e-01 -1.23077297e+00 3.26415151e-01
1.50348222e+00 -3.17257673e-01 -6.78477705e-01 5.15175879e-01
6.60500228e-01 7.36174285e-02 -9.22967315e-01 6.12021148e-01
2.91017562e-01 -8.69707227e-01 1.18995023e+00 -3.72199863e-01
1.72469378e-01 -7.83785343e-01 -5.24810493e-01 -1.00295651e+00
-8.18962157e-01 -2.88619906e-01 1.95229441e-01 1.54482079e+00
-7.00025260e-02 -7.13165820e-01 6.92112923e-01 -5.94852194e-02
7.25332648e-02 -6.09813690e-01 -8.92449796e-01 -9.54759061e-01
-1.26088470e-01 -3.71776102e-03 5.93666911e-01 8.14727783e-01
-5.17064869e-01 2.25575224e-01 -9.57029685e-02 6.46524072e-01
1.04449618e+00 6.14534378e-01 6.02093875e-01 -1.55761099e+00
-3.73901099e-01 -3.24019432e-01 -7.41792679e-01 -1.22035885e+00
1.33192256e-01 -1.03747964e+00 1.58065483e-01 -1.27486265e+00
5.69626451e-01 -6.75729215e-01 -5.89905679e-01 1.19083367e-01
-2.24858493e-01 2.40990713e-01 2.12180782e-02 5.75813115e-01
-9.80400264e-01 5.65722704e-01 1.43214726e+00 -3.23293686e-01
-1.39583886e-01 9.15335938e-02 -7.55274296e-01 8.86918008e-01
5.95821738e-01 -4.63451236e-01 -3.10264170e-01 -2.57851630e-01
-2.52275854e-01 -2.23262653e-01 4.15325493e-01 -1.07952511e+00
3.46503377e-01 -4.76432800e-01 7.59298682e-01 -6.64816976e-01
1.53260797e-01 -7.46929049e-01 -5.62207699e-02 1.80085972e-01
-2.61314243e-01 -2.90407300e-01 -3.80415440e-01 7.99899101e-01
-2.12440923e-01 -2.33993530e-02 9.74332809e-01 2.33014021e-02
-9.00417924e-01 5.52079618e-01 -2.72261221e-02 4.72763926e-02
1.24738240e+00 -3.47247303e-01 -1.51970014e-01 7.15124458e-02
-5.43779314e-01 2.80432731e-01 2.38903821e-01 4.60760176e-01
7.11531222e-01 -1.48674774e+00 -6.41050875e-01 3.33130509e-01
2.92540014e-01 1.40114039e-01 1.39194608e-01 7.79069841e-01
-1.25831723e-01 1.58844620e-01 -1.86767623e-01 -1.00214076e+00
-1.33908319e+00 6.45764172e-01 3.07344705e-01 1.28801793e-01
-4.08446014e-01 1.08658826e+00 7.84488916e-01 -3.21738720e-01
1.79187953e-01 -3.24866056e-01 -3.75348181e-01 1.11264000e-02
5.90700388e-01 3.76837939e-01 -1.49820521e-01 -7.45900989e-01
-6.63798511e-01 9.09666717e-01 -5.94282858e-02 3.31425756e-01
1.17670345e+00 -3.09360445e-01 -3.04130256e-01 3.13990146e-01
1.38411295e+00 8.83881375e-02 -1.41209888e+00 -5.28247178e-01
6.31789565e-02 -7.80984282e-01 -3.85264643e-02 -2.10100248e-01
-1.06131804e+00 3.94853503e-01 7.50290096e-01 9.43263248e-02
1.25171149e+00 3.38845879e-01 3.89933676e-01 5.57692766e-01
5.37094116e-01 -1.04512477e+00 2.23456249e-01 2.07046390e-01
7.80200303e-01 -1.28742826e+00 -1.14328869e-01 -7.22127736e-01
-3.99215013e-01 1.08391011e+00 6.06952429e-01 -3.95850509e-01
7.47661352e-01 -3.98656428e-02 -1.27759248e-01 1.26667051e-02
-3.18395138e-01 -2.52282023e-01 4.87517118e-01 7.05991745e-01
1.52266547e-01 -2.30839495e-02 -2.80824453e-01 5.68103671e-01
1.96412116e-01 -1.53008699e-01 -4.74785715e-02 7.53862143e-01
-8.19323123e-01 -8.40837777e-01 -5.59527755e-01 5.74299395e-01
-3.06546450e-01 2.32223913e-01 -2.57558465e-01 3.56814772e-01
5.95104694e-01 1.05277991e+00 1.90089568e-01 -5.79753757e-01
1.17344260e-01 -3.76114696e-01 5.23254991e-01 -5.65476120e-01
-5.80248944e-02 8.39366838e-02 -5.99651039e-01 -4.94677305e-01
-6.12796605e-01 -6.36106968e-01 -1.00978923e+00 3.39718461e-02
-5.94326675e-01 2.04897746e-01 3.87930840e-01 8.31711352e-01
8.02474841e-02 2.81519383e-01 1.22034907e+00 -7.71219254e-01
-7.59882092e-01 -6.63652480e-01 -1.05344200e+00 4.20062065e-01
7.54464120e-02 -9.64204729e-01 -6.08513415e-01 -4.11146805e-02] | [9.77925968170166, 1.9640717506408691] |
1e017724-cc9a-4fc3-80c1-b7a8c94901a2 | um-checker-a-hybrid-system-for-english | null | null | https://aclanthology.org/W13-3605 | https://aclanthology.org/W13-3605.pdf | UM-Checker: A Hybrid System for English Grammatical Error Correction | null | ['Long-Yue Wang', 'Xiaodong Zeng', 'Lidia S. Chao', 'Junwen Xing', 'Derek F. Wong'] | 2013-08-01 | null | null | null | ws-2013-8 | ['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] | [-7.285709857940674, 3.7072830200195312] |
b440cc54-513d-4ff3-b55b-59d6ba33eb64 | medlocker-a-transferable-adversarial | 2303.09858 | null | https://arxiv.org/abs/2303.09858v2 | https://arxiv.org/pdf/2303.09858v2.pdf | MedLocker: A Transferable Adversarial Watermarking for Preventing Unauthorized Analysis of Medical Image Dataset | The collection of medical image datasets is a demanding and laborious process that requires significant resources. Furthermore, these medical datasets may contain personally identifiable information, necessitating measures to ensure that unauthorized access is prevented. Failure to do so could violate the intellectual property rights of the dataset owner and potentially compromise the privacy of patients. As a result, safeguarding medical datasets and preventing unauthorized usage by AI diagnostic models is a pressing challenge. To address this challenge, we propose a novel visible adversarial watermarking method for medical image copyright protection, called MedLocker. Our approach involves continuously optimizing the position and transparency of a watermark logo, which reduces the performance of the target model, leading to incorrect predictions. Importantly, we ensure that our method minimizes the impact on clinical visualization by constraining watermark positions using semantical masks (WSM), which are bounding boxes of lesion regions based on semantic segmentation. To ensure the transferability of the watermark across different models, we verify the cross-model transferability of the watermark generated on a single model. Additionally, we generate a unique watermark parameter list each time, which can be used as a certification to verify the authorization. We evaluate the performance of MedLocker on various mainstream backbones and validate the feasibility of adversarial watermarking for copyright protection on two widely-used diabetic retinopathy detection datasets. Our results demonstrate that MedLocker can effectively protect the copyright of medical datasets and prevent unauthorized users from analyzing medical images with AI diagnostic models. | ['Shiji Zhao', 'Huazhu Fu', 'Xingxing Wei', 'Bangzheng Pu'] | 2023-03-17 | null | null | null | null | ['diabetic-retinopathy-detection'] | ['medical'] | [ 8.66702914e-01 4.04053450e-01 -1.52669698e-01 3.99691388e-02
-4.36755419e-01 -9.97762620e-01 7.06321076e-02 4.04406905e-01
-3.00200731e-01 3.54031444e-01 -1.27935603e-01 -7.10432529e-01
8.17825124e-02 -8.10373247e-01 -7.19891489e-01 -5.51024973e-01
6.36328384e-02 -2.18124956e-01 2.42126569e-01 3.44183594e-01
1.86366007e-01 6.42294645e-01 -9.01378334e-01 3.13133419e-01
9.72286940e-01 1.12672961e+00 -2.78460085e-01 4.46345657e-01
2.70259380e-01 2.99211353e-01 -1.10184073e+00 -6.06258631e-01
8.02420318e-01 -2.46864036e-01 -4.97584134e-01 1.19256176e-01
4.83417898e-01 -7.04561472e-01 -3.28112513e-01 1.31398010e+00
5.28921008e-01 -6.94443285e-01 4.05688941e-01 -1.44142520e+00
-7.02717721e-01 2.23283470e-01 -7.42055535e-01 -2.63397872e-01
3.82954739e-02 4.15960103e-01 5.78507781e-01 -3.35863195e-02
8.22829843e-01 5.82041621e-01 4.93135214e-01 5.46332121e-01
-1.27069879e+00 -1.09029496e+00 -1.53084144e-01 -7.62321502e-02
-1.42008054e+00 -3.53121936e-01 7.41482794e-01 -6.72023952e-01
-3.47580737e-03 7.01840341e-01 6.38266444e-01 9.04284477e-01
5.34153521e-01 3.05612862e-01 1.08950067e+00 -2.97593027e-01
1.55585229e-01 2.06386805e-01 -7.15819150e-02 9.48314369e-01
9.27503288e-01 2.01725662e-01 -3.88873100e-01 -6.82017863e-01
7.26575971e-01 -5.78081869e-02 -7.03278959e-01 -4.57538009e-01
-1.18536711e+00 5.03590703e-01 1.31242439e-01 -2.71399580e-02
5.75627722e-02 2.82236785e-02 2.34711781e-01 1.81336850e-02
-6.33312166e-02 3.16487759e-01 -2.05584824e-01 5.51385581e-01
-7.52362072e-01 -4.23469730e-02 3.83971810e-01 8.14868569e-01
2.02533484e-01 -4.15722698e-01 -1.96624935e-01 9.05024409e-02
5.57271361e-01 5.75264812e-01 3.47183973e-01 -7.19660699e-01
3.26592088e-01 7.87521243e-01 8.60187933e-02 -1.34229469e+00
-2.72252760e-03 -1.62428692e-01 -1.02894819e+00 6.57602310e-01
4.33499753e-01 -3.79139371e-02 -1.02745068e+00 1.49306190e+00
5.33200979e-01 -1.63174495e-02 -5.12143224e-03 5.82541704e-01
4.94975269e-01 2.64230907e-01 3.59483719e-01 -6.96720108e-02
1.66623843e+00 -3.45217079e-01 -8.79190743e-01 4.97059636e-02
5.17105758e-01 -8.92112195e-01 7.62885630e-01 3.20481628e-01
-8.37333143e-01 -2.18226030e-01 -1.42950022e+00 -1.44617306e-02
-2.79317141e-01 3.30161080e-02 1.65152073e-01 1.13219512e+00
-6.09663010e-01 3.09825003e-01 -8.25937986e-01 3.46592277e-01
1.01905215e+00 4.43920285e-01 -4.39679354e-01 1.96350981e-02
-1.19029415e+00 4.78889495e-01 2.68830746e-01 -8.73979852e-02
-4.84631568e-01 -1.02280247e+00 -8.94568264e-01 -6.00722767e-02
1.14772290e-01 -6.37270391e-01 6.57646418e-01 -8.57919991e-01
-8.75630438e-01 1.16464269e+00 3.14996988e-01 -5.24862289e-01
1.07372308e+00 9.04378444e-02 -5.60129344e-01 4.14208680e-01
-9.47241560e-02 6.06432140e-01 1.30736101e+00 -1.42517138e+00
-4.47193623e-01 -3.57840687e-01 -1.49526536e-01 -5.31088531e-01
-6.11956656e-01 -4.72430028e-02 -2.62105167e-01 -1.21320224e+00
-8.79444927e-02 -1.00520515e+00 -1.59449205e-01 9.43153322e-01
-6.55040860e-01 7.02197909e-01 1.04849577e+00 -1.09387052e+00
1.43689942e+00 -2.70700836e+00 -3.60029429e-01 6.76716149e-01
4.94346589e-01 4.45875585e-01 7.63300136e-02 -2.00968623e-01
1.07138522e-01 6.30648732e-01 -7.04855859e-01 2.43640572e-01
-1.38223782e-01 2.91190203e-02 -5.34451902e-01 7.70549834e-01
8.29096064e-02 9.40881431e-01 -3.86514843e-01 -7.43551970e-01
1.78864300e-02 5.34446418e-01 -5.17307103e-01 5.45948744e-02
-2.39555210e-01 4.30970341e-01 -5.91703296e-01 7.40068734e-01
9.95123625e-01 -4.18342054e-01 3.98555189e-01 -3.06234390e-01
3.30769420e-01 -3.75413060e-01 -1.18185973e+00 1.20197928e+00
1.47375211e-01 4.63076413e-01 4.83325273e-02 -3.35516423e-01
6.51089609e-01 5.29153526e-01 4.91743237e-01 -4.49449867e-01
1.18937008e-01 1.91549584e-01 -3.45606208e-02 -4.84760344e-01
2.10590735e-01 5.42482808e-02 -4.41106968e-02 5.08470714e-01
-7.83165276e-01 1.57853797e-01 -3.62677217e-01 8.44182149e-02
9.26433802e-01 -9.70638841e-02 2.82307267e-01 -1.33575723e-02
3.74788105e-01 6.57769665e-02 6.33185565e-01 4.44798976e-01
-5.19075453e-01 5.79475224e-01 5.90530753e-01 -1.77514106e-01
-8.76918674e-01 -1.00630772e+00 -3.22197676e-01 1.02538690e-01
3.03706825e-01 -4.13535871e-02 -8.82470071e-01 -1.06421030e+00
3.43800068e-01 4.62693304e-01 -5.90342581e-01 -3.11311543e-01
-4.56328273e-01 -4.57131892e-01 1.01441169e+00 2.13227540e-01
5.24789155e-01 -7.78300881e-01 -1.19531846e+00 -5.15047833e-02
-1.31341755e-01 -1.05711770e+00 -9.94267881e-01 -4.69730765e-01
-6.75829470e-01 -1.55984342e+00 -7.54583359e-01 -7.54919946e-01
1.27108705e+00 1.21902972e-02 3.68632317e-01 4.62076128e-01
-7.70498753e-01 3.09925318e-01 -9.66644883e-02 -6.76492989e-01
-8.89415860e-01 -2.86662877e-01 -3.95597816e-01 2.06671923e-01
-1.70524552e-01 -1.59787580e-01 -8.61385584e-01 3.95922065e-01
-1.49147427e+00 2.57822853e-02 3.00059885e-01 4.48849499e-01
6.84576631e-01 1.21975385e-01 2.44401470e-01 -9.89634991e-01
4.36539769e-01 -1.09159149e-01 -1.06363833e+00 5.47077477e-01
-7.25922883e-01 -1.93301842e-01 3.79217833e-01 -5.47862351e-01
-5.83462894e-01 1.36936486e-01 3.98527354e-01 -3.87151778e-01
1.94539689e-02 2.87282407e-01 -4.25696790e-01 -4.80889022e-01
4.37643945e-01 -9.43587124e-02 2.82391280e-01 -2.30718270e-01
2.25142196e-01 7.90297687e-01 5.33008993e-01 -1.67132497e-01
1.11433029e+00 7.58278728e-01 1.30463094e-01 -4.75177735e-01
-2.86264420e-01 4.02729623e-02 -2.21096441e-01 -2.36038700e-01
8.59607577e-01 -5.88683426e-01 -7.55730867e-01 7.20713556e-01
-1.26244760e+00 1.07000165e-01 -2.07754925e-01 4.12336364e-02
-3.37873958e-02 7.81507909e-01 -4.42751199e-01 -3.83862764e-01
-4.51720417e-01 -1.24233711e+00 6.70997858e-01 -5.06283157e-03
-3.90641153e-01 -6.86600089e-01 -3.23367774e-01 8.04337919e-01
2.10227415e-01 1.00288641e+00 1.31638789e+00 -3.87147605e-01
-9.89795148e-01 -5.01638710e-01 -1.79741815e-01 3.33022416e-01
3.60474765e-01 2.05923557e-01 -8.38540673e-01 -5.42313516e-01
6.57000691e-02 1.13449484e-01 4.60388184e-01 2.84470797e-01
1.29458916e+00 -6.57160640e-01 -4.10285264e-01 7.16031969e-01
1.17523050e+00 3.66258949e-01 8.24244678e-01 4.19331431e-01
6.04710639e-01 7.37576187e-01 4.66403157e-01 4.46300745e-01
-5.87930121e-02 3.37157786e-01 4.43211198e-01 -3.62816453e-01
3.23946364e-02 -1.77416861e-01 1.54753968e-01 3.59226644e-01
3.65657687e-01 -2.59496748e-01 -8.57068181e-01 3.81680131e-01
-1.42334759e+00 -4.56905037e-01 -3.23192291e-02 2.32601166e+00
1.06808972e+00 7.09701255e-02 -3.25091153e-01 1.63991213e-01
8.20155442e-01 2.58177426e-02 -5.84486365e-01 -1.32977730e-02
2.93681584e-02 1.39581889e-01 9.54186499e-01 2.10385829e-01
-1.09478533e+00 5.25311768e-01 6.11443806e+00 6.30600572e-01
-1.24546456e+00 1.28577515e-01 6.90692425e-01 7.03564361e-02
-4.53151315e-01 -1.20361872e-01 -2.80596316e-01 6.82711542e-01
4.85125214e-01 -3.17785025e-01 3.13094780e-02 4.85115856e-01
2.87663788e-01 -4.41424660e-02 -8.07738304e-01 5.76963305e-01
-1.12571329e-01 -1.53903103e+00 2.60332108e-01 5.17802596e-01
5.51867366e-01 -7.22814500e-01 4.76884753e-01 -7.27956653e-01
5.21384366e-02 -9.09462810e-01 7.90296316e-01 4.79022950e-01
1.24554956e+00 -6.72842860e-01 5.10183334e-01 -5.72366603e-02
-7.83921480e-01 2.66594857e-01 1.53277442e-02 8.03519726e-01
1.91035882e-01 4.56858188e-01 -7.24930584e-01 4.71645355e-01
3.49436551e-01 2.07732663e-01 -6.26493216e-01 9.77533460e-01
-3.63846242e-01 4.34934139e-01 -1.56623170e-01 6.33936405e-01
-3.21042091e-01 4.82438877e-02 6.49457097e-01 7.47780800e-01
2.84821332e-01 4.51040678e-02 1.01035625e-01 8.65741193e-01
-3.81059200e-01 1.16605081e-01 -5.99859059e-01 -3.84454101e-01
3.84439290e-01 7.34921753e-01 -6.51058793e-01 -6.77483231e-02
-2.24374861e-01 9.52665567e-01 -5.18647909e-01 2.26145864e-01
-1.00501692e+00 -3.76436979e-01 6.21658862e-01 3.84765506e-01
1.65329412e-01 -1.46578133e-01 -6.65034354e-01 -8.56765807e-01
3.64056319e-01 -1.15614176e+00 5.95292330e-01 -5.83574355e-01
-9.14464712e-01 3.11924070e-01 -3.29146743e-01 -1.62543130e+00
5.02909422e-01 -4.32145447e-01 -8.31291750e-02 8.16607237e-01
-1.73351848e+00 -1.31184721e+00 -2.11247981e-01 6.97575986e-01
-3.04251939e-01 -1.04113899e-01 9.03571904e-01 3.47926617e-01
-4.14842993e-01 1.02258825e+00 -1.48095250e-01 4.54447180e-01
8.06620061e-01 -5.86370647e-01 3.45741153e-01 9.47617173e-01
-1.37915656e-01 7.75888443e-01 3.84881556e-01 -1.05570900e+00
-1.25669646e+00 -1.19636869e+00 5.06853342e-01 -2.65049607e-01
5.85278034e-01 -6.10282384e-02 -1.15155447e+00 6.37296736e-01
-5.89153282e-02 1.85173482e-01 1.19355226e+00 -9.25406158e-01
-4.48418349e-01 8.28384161e-02 -1.62353230e+00 6.65789485e-01
4.70634431e-01 -5.18872917e-01 -2.90691435e-01 9.84388441e-02
9.00135815e-01 -4.14559066e-01 -1.05046451e+00 4.32617188e-01
8.09004247e-01 -5.28431416e-01 9.10290003e-01 -5.87700069e-01
3.77064377e-01 -6.78901672e-01 9.84620228e-02 -5.34606874e-01
2.13165611e-01 -7.58401752e-01 -1.05033107e-01 1.12763512e+00
3.42900693e-01 -9.49748635e-01 6.93846762e-01 9.86424863e-01
5.10968328e-01 -2.72923619e-01 -1.14691007e+00 -5.88151574e-01
-1.44600272e-01 -1.81811988e-01 7.50685275e-01 1.36071229e+00
-1.58532202e-01 -6.79294586e-01 -3.12629968e-01 8.33481610e-01
8.44220400e-01 -1.35613725e-01 6.11215770e-01 -1.03269851e+00
-8.62243921e-02 -2.64086336e-01 -6.91514075e-01 -3.51225734e-01
-1.47635996e-01 -8.38520527e-01 -3.95268083e-01 -1.11052561e+00
-2.86933184e-02 -5.51420689e-01 -3.52436900e-01 7.09585428e-01
-1.69213548e-01 4.39753294e-01 3.90435725e-01 5.34723580e-01
2.26759359e-01 -1.36338072e-02 1.38687253e+00 -5.13791382e-01
-2.87391078e-02 -1.85862333e-01 -6.26324594e-01 6.93188071e-01
7.38289893e-01 -8.92964184e-01 -3.24109405e-01 -2.12608829e-01
1.21277168e-01 -2.16837510e-01 7.57331312e-01 -7.41181672e-01
1.92079470e-01 -4.02859300e-01 2.54965067e-01 -2.57649362e-01
-2.60669559e-01 -1.39933622e+00 5.51309168e-01 9.32228029e-01
-2.63880193e-01 -3.68529081e-01 3.56631845e-01 6.56430960e-01
2.81617958e-02 -1.99076280e-01 8.59938860e-01 1.87556684e-01
-2.24596858e-01 3.00855786e-01 -1.94713846e-01 -1.25494629e-01
1.55052102e+00 -3.82227004e-01 -5.40772438e-01 2.29281873e-01
-5.78345478e-01 2.67019868e-01 9.38505709e-01 1.92386612e-01
7.31307924e-01 -9.69066143e-01 -5.18067479e-01 5.83071113e-01
2.85510093e-01 -5.13208993e-02 2.20006883e-01 4.52198207e-01
-8.38215768e-01 -2.21090436e-01 -1.69686288e-01 -3.17751348e-01
-1.84760725e+00 8.34422469e-01 3.16879213e-01 -1.44212231e-01
-8.98945212e-01 1.52004704e-01 5.88893294e-02 1.25816256e-01
2.73213536e-01 -2.61061639e-01 9.02396366e-02 4.42008525e-02
6.67693496e-01 2.11219653e-01 -1.15085535e-01 -3.88646364e-01
-2.77495384e-01 5.26869237e-01 -1.41601592e-01 -9.34808403e-02
8.92003655e-01 -1.62403435e-02 -1.28224835e-01 -3.37197661e-01
9.92373407e-01 5.14974952e-01 -1.26498973e+00 7.37451017e-02
-1.82798043e-01 -7.19930232e-01 -4.45936136e-02 -9.77709115e-01
-1.35551977e+00 6.04568720e-01 8.53987277e-01 2.59134501e-01
1.17028356e+00 -4.19459641e-01 1.09533262e+00 -2.45367438e-01
3.11080247e-01 -7.04281747e-01 -3.50873977e-01 -5.80691397e-01
7.89447308e-01 -8.95225942e-01 3.29414159e-01 -7.51494944e-01
-5.89589596e-01 8.92246127e-01 1.32136017e-01 4.58850861e-01
7.02082753e-01 3.73257190e-01 4.77659076e-01 -6.94080815e-02
1.06325723e-01 5.19831896e-01 4.63283837e-01 6.52599275e-01
-1.31208226e-01 2.13587120e-01 -3.42300445e-01 6.02587283e-01
-3.18058953e-02 3.06031317e-01 6.58523858e-01 1.50078249e+00
-5.58721051e-02 -1.23324645e+00 -7.97351599e-01 2.81373352e-01
-6.80136681e-01 3.34572457e-02 -4.05957639e-01 5.58011770e-01
4.22310650e-01 6.96313322e-01 -3.33766967e-01 -5.13912067e-02
3.68086874e-01 -3.86090875e-02 4.56269085e-02 -2.48542249e-01
-6.54567599e-01 1.16233369e-02 -1.60241961e-01 -2.85318613e-01
-2.71835834e-01 -3.98837447e-01 -1.27313960e+00 -3.00302833e-01
-2.66165018e-01 -3.68139952e-01 5.63223541e-01 5.73966324e-01
6.91984653e-01 6.25642657e-01 5.52751780e-01 1.01250149e-01
-5.66071987e-01 -3.01435627e-02 -4.58489776e-01 4.61632788e-01
6.15516901e-01 -2.54725873e-01 -1.99353904e-01 7.92629719e-01] | [6.014008045196533, 7.067345142364502] |
b7df76b1-d461-4330-a0c6-14ec4c3c1bea | chinese-word-segmentation-with-heterogeneous | 2201.08975 | null | https://arxiv.org/abs/2201.08975v1 | https://arxiv.org/pdf/2201.08975v1.pdf | Chinese Word Segmentation with Heterogeneous Graph Neural Network | In recent years, deep learning has achieved significant success in the Chinese word segmentation (CWS) task. Most of these methods improve the performance of CWS by leveraging external information, e.g., words, sub-words, syntax. However, existing approaches fail to effectively integrate the multi-level linguistic information and also ignore the structural feature of the external information. Therefore, in this paper, we proposed a framework to improve CWS, named HGNSeg. It exploits multi-level external information sufficiently with the pre-trained language model and heterogeneous graph neural network. The experimental results on six benchmark datasets (e.g., Bakeoff 2005, Bakeoff 2008) validate that our approach can effectively improve the performance of Chinese word segmentation. Importantly, in cross-domain scenarios, our method also shows a strong ability to alleviate the out-of-vocabulary (OOV) problem. | ['Qi Su', 'Jun Wang', 'Xuemei Tang'] | 2022-01-22 | null | null | null | null | ['chinese-word-segmentation'] | ['natural-language-processing'] | [-9.04890224e-02 5.12115024e-02 -4.37390417e-01 -3.50409806e-01
-6.50184453e-01 -6.17883801e-01 1.51543334e-01 8.49558040e-02
-6.09549344e-01 4.70129669e-01 1.94421276e-01 -6.70356572e-01
3.75533968e-01 -9.51156139e-01 -4.68488783e-01 -2.57911801e-01
4.63602483e-01 1.95611700e-01 5.87458849e-01 -3.21579427e-01
2.09599599e-01 3.01750842e-02 -7.68909872e-01 1.20123792e-02
1.52122962e+00 6.91006958e-01 5.07528424e-01 1.44043759e-01
-7.84152508e-01 5.50453961e-01 -5.81657231e-01 -3.47061485e-01
9.53513980e-02 -3.49786311e-01 -8.98281574e-01 -1.69945091e-01
9.69049111e-02 -2.26985440e-01 -1.04749143e-01 1.07524192e+00
4.59123284e-01 -6.57400638e-02 1.60080656e-01 -7.87751138e-01
-8.64016473e-01 1.21699512e+00 -6.13443375e-01 -4.30635177e-02
-4.57470715e-02 6.59821108e-02 1.36211419e+00 -9.17954385e-01
7.52096355e-01 1.08547497e+00 7.67384589e-01 3.45043957e-01
-7.29519367e-01 -7.98430383e-01 7.73699880e-01 1.36539564e-01
-1.21869159e+00 1.53848708e-01 8.23674321e-01 -2.31844366e-01
9.73213434e-01 6.76399618e-02 6.26727998e-01 8.88719499e-01
-2.07129434e-01 1.34482145e+00 1.01631331e+00 -5.96803248e-01
-8.41267630e-02 -2.42463395e-01 7.83696234e-01 5.84749043e-01
4.96729821e-01 -5.41148901e-01 -9.20145214e-02 2.26317286e-01
4.20822471e-01 -2.28777885e-01 -1.88822374e-01 -3.89615782e-02
-7.38833725e-01 1.03298473e+00 3.10609788e-01 7.20342577e-01
-7.47373840e-03 -6.67281970e-02 3.66817594e-01 -1.81793123e-02
7.34971762e-01 5.47919154e-01 -8.80863369e-01 2.24207230e-02
-8.09958100e-01 -1.08984992e-01 7.93989241e-01 1.32197297e+00
7.77515888e-01 2.56283194e-01 -1.60723597e-01 8.02510858e-01
4.25597906e-01 2.70201236e-01 7.20192611e-01 -5.13414979e-01
8.75999212e-01 8.27908874e-01 -3.96863371e-01 -9.24580932e-01
-4.71178889e-01 -6.24650121e-01 -6.78054452e-01 -6.56083047e-01
3.23586166e-01 -5.76630235e-01 -1.26039338e+00 1.86182475e+00
2.53359348e-01 1.43417507e-01 -4.69856039e-02 6.39859796e-01
1.16343820e+00 6.42140329e-01 3.84302914e-01 8.32176954e-02
1.08951986e+00 -1.33779407e+00 -9.88463700e-01 -6.16747081e-01
9.43720937e-01 -8.54609549e-01 1.27727950e+00 3.43829319e-02
-8.75519276e-01 -5.81589520e-01 -9.32570934e-01 -7.93369338e-02
-7.00530648e-01 2.61415690e-01 8.12495530e-01 7.63835728e-01
-9.70580041e-01 4.13145483e-01 -7.71373391e-01 -1.53707966e-01
3.63084555e-01 3.75268072e-01 6.91152120e-04 -2.12791026e-01
-1.71182299e+00 5.70495188e-01 7.94839799e-01 1.27340123e-01
-4.58074331e-01 -7.40453124e-01 -1.07060075e+00 7.27843866e-02
9.28306639e-01 -3.87263507e-01 1.13427746e+00 -1.03478670e+00
-1.48727107e+00 6.38315558e-01 -9.45514254e-03 -2.22519293e-01
2.55640030e-01 -5.12161672e-01 -4.57074493e-01 -5.27102835e-02
1.00773625e-01 4.88977998e-01 4.44976151e-01 -1.18642318e+00
-4.30471182e-01 -4.21096802e-01 2.47329339e-01 1.25761151e-01
-5.94722688e-01 8.30723792e-02 -1.03284991e+00 -8.09696674e-01
6.37488812e-02 -7.26250827e-01 -3.77505213e-01 -7.22201824e-01
-7.64302731e-01 -5.58695257e-01 8.59720170e-01 -9.54364955e-01
1.67863500e+00 -2.01241422e+00 1.18323080e-01 1.49197459e-01
2.28310958e-01 9.35096264e-01 -5.23548484e-01 5.72738945e-01
2.49653980e-01 7.17915833e-01 -3.59508693e-01 -3.92117053e-01
8.41721818e-02 3.33777517e-01 5.22591434e-02 1.61004975e-01
3.70587677e-01 1.37797391e+00 -7.80378938e-01 -7.38903224e-01
-9.89846438e-02 2.85983294e-01 -4.14992779e-01 2.02520356e-01
-3.79500002e-01 1.62170634e-01 -7.63760746e-01 5.67323387e-01
7.96914637e-01 -3.96562487e-01 4.52138335e-01 1.00380942e-01
-1.50202796e-01 4.76123065e-01 -1.06459832e+00 1.74669385e+00
-4.20319766e-01 3.48804563e-01 1.80514380e-01 -9.59046900e-01
7.33442545e-01 4.32930738e-02 2.03990683e-01 -7.99582303e-01
1.39363617e-01 2.33603641e-01 1.72172561e-01 -3.29861343e-01
6.14117086e-01 2.43485168e-01 -1.39721841e-01 2.12934390e-01
1.19061731e-01 1.48276910e-01 3.51164192e-01 5.51223576e-01
7.69453883e-01 3.01756084e-01 3.46606761e-01 -2.94239342e-01
6.23538971e-01 1.21790521e-01 7.68579900e-01 6.95165157e-01
-2.57060349e-01 5.70912302e-01 5.38732529e-01 1.10388093e-01
-6.70315444e-01 -6.23534441e-01 7.43608177e-02 1.17773783e+00
3.75550479e-01 -6.93502486e-01 -1.21549606e+00 -9.84637082e-01
-1.83590516e-01 5.46355963e-01 -3.93259019e-01 -2.16327146e-01
-9.21364248e-01 -6.99159682e-01 7.89287746e-01 9.30140793e-01
5.80422878e-01 -1.12147558e+00 -5.78711256e-02 3.85995686e-01
-3.02961290e-01 -1.55384159e+00 -6.14067852e-01 -2.42912918e-02
-7.92359054e-01 -8.00915778e-01 -6.51585281e-01 -1.12047505e+00
4.77110744e-01 4.64842677e-01 1.15893376e+00 5.04409969e-01
-1.19495623e-01 1.12054765e-01 -8.47779632e-01 -3.20280820e-01
-1.78459376e-01 7.50436366e-01 -6.34110391e-01 -2.63284445e-01
5.90839565e-01 6.44486696e-02 -4.36861962e-01 1.15585484e-01
-9.53603745e-01 1.28468528e-01 5.87138712e-01 7.75035024e-01
4.71012354e-01 -4.98394668e-02 7.58792818e-01 -1.32476771e+00
7.50891387e-01 -3.73379618e-01 -5.95244765e-01 3.57696086e-01
-6.44391835e-01 -2.01692089e-01 7.41021872e-01 -4.28163618e-01
-1.19274974e+00 -2.87932038e-01 -4.30012316e-01 2.39680082e-01
-1.87998816e-01 1.13925922e+00 -6.23574853e-01 6.12583309e-02
-1.11296095e-01 1.63047817e-02 -5.03050029e-01 -7.39851892e-01
4.64619011e-01 7.12684274e-01 5.52334264e-02 -5.11895299e-01
6.84864044e-01 1.24198124e-01 -5.35910010e-01 -7.37330914e-01
-1.07969379e+00 -7.20511794e-01 -8.42667758e-01 1.68928981e-01
1.18870759e+00 -1.05703425e+00 -7.55324215e-02 9.08551574e-01
-1.20307493e+00 -5.80900788e-01 1.50790319e-01 3.65958631e-01
2.54572779e-01 8.17518234e-01 -9.49829459e-01 -6.09788001e-01
-4.45351034e-01 -1.07046974e+00 8.99313331e-01 4.13009763e-01
4.87817749e-02 -1.20735204e+00 -2.72432715e-01 4.39094633e-01
5.54653347e-01 1.74567282e-01 1.08883631e+00 -9.17713940e-01
-5.69935620e-01 -1.32444009e-01 -4.42334175e-01 3.70491266e-01
1.39855132e-01 -6.27627373e-02 -6.80435121e-01 -1.98926494e-01
-3.25786233e-01 -1.83934793e-01 1.14103520e+00 3.21774542e-01
1.12360203e+00 5.93513250e-02 -3.17191750e-01 5.43705046e-01
1.55867517e+00 2.54724205e-01 4.39519793e-01 3.75253379e-01
1.22504163e+00 5.63148558e-01 7.02605367e-01 1.30020633e-01
6.97839916e-01 3.53015333e-01 4.00392294e-01 -5.38645804e-01
-3.07627261e-01 -2.79160500e-01 2.97237813e-01 1.38032520e+00
6.84713349e-02 -5.62364399e-01 -1.13601124e+00 6.82070911e-01
-1.78446245e+00 -2.27591947e-01 -5.27782202e-01 1.58815587e+00
9.41357791e-01 3.31367165e-01 -2.00686648e-01 -2.71507025e-01
8.43452096e-01 3.96290034e-01 -5.59222221e-01 -2.83967406e-01
-3.53443652e-01 4.17455107e-01 5.30124605e-01 5.20770550e-01
-1.20819271e+00 1.91498995e+00 5.67482328e+00 1.17001331e+00
-8.37095916e-01 1.68658942e-01 4.94369328e-01 6.66713297e-01
-6.54692829e-01 -1.26425340e-03 -1.16883099e+00 4.01599109e-01
5.57241082e-01 1.38821602e-01 1.72404438e-01 6.62975192e-01
-1.32438019e-01 1.30774125e-01 -3.81659895e-01 2.82154322e-01
5.16976081e-02 -1.07076108e+00 -5.87673746e-02 4.37809713e-02
8.87780249e-01 3.18661988e-01 -8.91860500e-02 5.89526772e-01
6.40070736e-01 -8.43280971e-01 4.81696844e-01 -1.33439898e-01
6.33420348e-01 -7.92236805e-01 9.76853251e-01 4.54588681e-01
-1.38258314e+00 2.77251035e-01 -2.85133451e-01 -9.22891591e-03
3.59397143e-01 5.11112809e-01 -4.59378451e-01 8.73730838e-01
3.86698008e-01 7.38246918e-01 -8.31196249e-01 6.68019593e-01
-6.96938872e-01 1.14849317e+00 -1.37670979e-01 -1.56128362e-01
8.80088627e-01 -5.15615940e-01 1.23165287e-01 1.57791567e+00
6.81236060e-03 7.89536685e-02 4.05207813e-01 7.61278987e-01
-2.49403700e-01 5.96192300e-01 -3.33809555e-01 -4.39760178e-01
2.88455188e-01 1.25097334e+00 -9.35407162e-01 -3.95791680e-01
-8.22377920e-01 6.01521552e-01 5.57669520e-01 5.71462452e-01
-8.35205436e-01 -6.16117060e-01 4.64182824e-01 -3.34202588e-01
6.33718014e-01 -5.47176003e-01 -4.08959419e-01 -1.24668682e+00
3.05587966e-02 -9.42280471e-01 3.72655332e-01 -2.59756893e-01
-1.17430186e+00 5.84405839e-01 -2.39708647e-01 -7.02692270e-01
1.38331935e-01 -6.56741381e-01 -6.61782205e-01 8.91662538e-01
-1.95336628e+00 -1.49205387e+00 -7.72572532e-02 5.19455314e-01
6.76154912e-01 -2.31526103e-02 4.77833003e-01 3.40027124e-01
-9.59999740e-01 7.27459967e-01 2.67154500e-02 7.57431269e-01
5.83640873e-01 -1.41971326e+00 9.33576524e-01 1.16513872e+00
2.26476759e-01 8.14769745e-01 1.68919772e-01 -1.04334354e+00
-1.31112182e+00 -1.10186815e+00 1.02883947e+00 -2.59268992e-02
8.27371538e-01 -5.55471897e-01 -1.08573830e+00 7.05585063e-01
5.31151712e-01 -2.24188894e-01 8.69356394e-01 2.67983973e-01
-3.64324600e-01 2.22115144e-01 -6.21099114e-01 6.75590277e-01
1.16147792e+00 -3.72790337e-01 -6.67786241e-01 2.50054479e-01
1.44193006e+00 -5.30263722e-01 -6.85275018e-01 3.82993042e-01
8.48015100e-02 -5.69844365e-01 8.60117197e-01 -1.01676166e+00
4.54667866e-01 8.25488567e-03 2.35870089e-02 -1.32083261e+00
-2.12417394e-01 -3.88639450e-01 -3.62752154e-02 1.54093802e+00
6.84903920e-01 -6.54646099e-01 5.96911967e-01 5.06684542e-01
-5.92759252e-01 -7.36949444e-01 -5.13479471e-01 -7.87289143e-01
5.00177324e-01 -5.27792752e-01 8.03273022e-01 1.19205260e+00
-1.55050755e-01 4.61625189e-01 -3.40100735e-01 7.34680817e-02
2.26921782e-01 4.18732017e-01 5.64830840e-01 -1.15011311e+00
-2.24040866e-01 -4.67898816e-01 9.66785029e-02 -1.30945742e+00
5.82262635e-01 -9.68829334e-01 -1.74587835e-02 -1.91980588e+00
1.07129887e-01 -4.84704345e-01 -4.31669563e-01 4.06280965e-01
-7.95495749e-01 6.55679265e-03 4.30794597e-01 -2.18810916e-01
-8.44587088e-01 3.22709262e-01 1.64958608e+00 -7.20413029e-02
-1.51692390e-01 -3.05067271e-01 -9.23707902e-01 9.09394801e-01
1.15904677e+00 -3.60102028e-01 -1.42648697e-01 -9.64292586e-01
1.20170355e-01 -3.71507883e-01 -2.63462245e-01 -4.71504033e-01
2.81604201e-01 -2.99847305e-01 2.52648946e-02 -7.95269430e-01
-6.54321164e-02 -5.74626148e-01 -5.77406704e-01 4.43329781e-01
-1.71317860e-01 -6.50073513e-02 2.25128874e-01 6.72788978e-01
-4.20105934e-01 -4.83823270e-01 4.70353425e-01 -3.94835055e-01
-1.07866752e+00 3.98852527e-01 -3.76117289e-01 6.64864838e-01
8.24912846e-01 2.48628147e-02 -4.28258806e-01 -1.46876708e-01
-4.14655477e-01 5.21305084e-01 1.39484927e-01 5.54899812e-01
3.61786306e-01 -1.13475370e+00 -5.18847167e-01 1.71134114e-01
9.29409172e-03 2.65169472e-01 1.58697814e-01 6.53867662e-01
-5.02516389e-01 4.80238885e-01 2.12136224e-01 -3.56290847e-01
-1.10776675e+00 4.74824488e-01 -4.06932905e-02 -7.80871212e-01
-5.17846286e-01 8.91529739e-01 1.76304594e-01 -5.16192377e-01
2.03043044e-01 -4.98353451e-01 -4.57377434e-01 2.34133765e-01
1.74455971e-01 1.10103101e-01 -4.38373983e-02 -6.55701518e-01
-2.95077711e-01 7.74732351e-01 -3.03479224e-01 1.31388739e-01
1.31623113e+00 -2.59356946e-01 -1.82840824e-01 1.60542950e-01
1.21305239e+00 2.19577044e-01 -9.28984046e-01 -5.92693508e-01
5.35669684e-01 -1.84200928e-01 1.20321780e-01 -7.20921636e-01
-1.46679688e+00 1.12090766e+00 -6.70545995e-02 3.02693337e-01
1.06120348e+00 -1.03027299e-01 1.45985115e+00 2.32294917e-01
2.21819207e-01 -1.60931182e+00 3.27700675e-02 8.97854507e-01
1.58510208e-01 -1.42314708e+00 -1.92315221e-01 -8.28257859e-01
-8.02157044e-01 9.97672200e-01 9.77291167e-01 -8.72313008e-02
6.49980724e-01 2.38943473e-01 2.53681213e-01 -9.56668407e-02
-2.56006688e-01 -7.11841226e-01 3.68505239e-01 4.64192480e-01
5.46091855e-01 2.83551306e-01 -7.54494071e-01 9.34854090e-01
1.83170900e-01 -3.42418104e-01 3.69064242e-01 8.32400918e-01
-5.13567328e-01 -1.41047168e+00 1.64963469e-01 2.63195157e-01
-6.36518538e-01 -5.75042963e-01 -7.68138528e-01 1.19652176e+00
1.77454561e-01 1.01899183e+00 -2.30859920e-01 -2.01990768e-01
2.64652580e-01 -1.02704819e-02 -1.64185148e-02 -7.74047494e-01
-8.77703309e-01 4.01711345e-01 2.01930732e-01 -4.86867696e-01
-4.29940194e-01 -4.07829285e-01 -1.51538622e+00 3.11105959e-02
-7.15877116e-01 9.54602212e-02 5.76815724e-01 1.18261361e+00
2.38532081e-01 6.35194600e-01 2.91787386e-01 -7.55133927e-02
-4.02552605e-01 -9.20548558e-01 -5.16385615e-01 4.56979424e-01
-1.49018606e-02 -2.45724052e-01 -1.93521455e-01 -2.81164944e-01] | [10.021080017089844, 10.125632286071777] |
3460079e-6a69-4f60-9248-ad423fecc8a1 | relational-representation-learning-in | 2205.02411 | null | https://arxiv.org/abs/2205.02411v1 | https://arxiv.org/pdf/2205.02411v1.pdf | Relational Representation Learning in Visually-Rich Documents | Relational understanding is critical for a number of visually-rich documents (VRDs) understanding tasks. Through multi-modal pre-training, recent studies provide comprehensive contextual representations and exploit them as prior knowledge for downstream tasks. In spite of their impressive results, we observe that the widespread relational hints (e.g., relation of key/value fields on receipts) built upon contextual knowledge are not excavated yet. To mitigate this gap, we propose DocReL, a Document Relational Representation Learning framework. The major challenge of DocReL roots in the variety of relations. From the simplest pairwise relation to the complex global structure, it is infeasible to conduct supervised training due to the definition of relation varies and even conflicts in different tasks. To deal with the unpredictable definition of relations, we propose a novel contrastive learning task named Relational Consistency Modeling (RCM), which harnesses the fact that existing relations should be consistent in differently augmented positive views. RCM provides relational representations which are more compatible to the urgent need of downstream tasks, even without any knowledge about the exact definition of relation. DocReL achieves better performance on a wide variety of VRD relational understanding tasks, including table structure recognition, key information extraction and reading order detection. | ['Bo Ren', 'Yinsong Liu', 'Deqiang Jiang', 'Yunfei Wu', 'Haoyu Cao', 'Yiqing Hu', 'Yan Zheng', 'Xin Li'] | 2022-05-05 | null | null | null | null | ['key-information-extraction'] | ['natural-language-processing'] | [ 1.57241791e-01 2.06358075e-01 -5.65427423e-01 -2.85668224e-01
-3.01661700e-01 -6.80223405e-01 9.11054075e-01 6.79170132e-01
2.83317864e-01 3.69569004e-01 5.84790349e-01 -7.29506791e-01
-6.75353825e-01 -9.44189787e-01 -6.29723310e-01 -1.56252041e-01
1.58518199e-02 4.50182080e-01 2.10581109e-01 -6.33358240e-01
3.64074469e-01 7.11388409e-01 -1.38188338e+00 9.19967830e-01
6.92484915e-01 1.07099390e+00 2.08198681e-01 3.03870827e-01
-5.69430530e-01 1.39932287e+00 -6.73035562e-01 -5.25762856e-01
-1.06909037e-01 -5.04869558e-02 -1.20322013e+00 -1.12575606e-01
2.56574452e-01 -2.44804814e-01 -5.53491473e-01 7.10901260e-01
5.92273101e-02 1.13375828e-01 7.84741998e-01 -1.21216905e+00
-1.15372133e+00 8.65485132e-01 -4.94207650e-01 2.54380703e-01
7.20917881e-01 -4.30342585e-01 1.54719329e+00 -9.84735668e-01
9.55016077e-01 1.45726812e+00 5.14552355e-01 3.16659473e-02
-1.09373534e+00 -2.71551639e-01 6.56826258e-01 5.15443802e-01
-1.42683661e+00 -1.68573394e-01 7.49899030e-01 -3.58756393e-01
1.14569438e+00 5.61344206e-01 2.83014804e-01 1.10704470e+00
-5.66075323e-03 7.51866639e-01 1.03932822e+00 -4.46035773e-01
-8.98728222e-02 1.77469656e-01 3.85961622e-01 4.38830733e-01
4.27074999e-01 -3.67350250e-01 -8.56304288e-01 1.62851244e-01
7.17707276e-01 1.18209533e-01 -2.50080884e-01 -5.62924385e-01
-1.36011505e+00 4.00591731e-01 6.35852039e-01 3.32597077e-01
-4.63152267e-02 -3.16662699e-01 4.18259561e-01 3.27150047e-01
4.09369543e-02 5.03281236e-01 -5.34955680e-01 1.42534539e-01
-2.05571786e-01 9.37054083e-02 6.93817735e-01 1.25248528e+00
7.30775356e-01 -4.78742242e-01 -2.10011438e-01 7.79052138e-01
4.61845309e-01 6.25176579e-02 3.55593637e-02 -4.36901271e-01
1.09236193e+00 1.35202622e+00 -2.24386245e-01 -1.57898962e+00
-6.00865960e-01 -3.92613351e-01 -1.06916296e+00 -1.58207178e-01
2.39422187e-01 5.75276971e-01 -4.63883549e-01 1.27191484e+00
3.93671691e-01 -2.63773113e-01 3.65336329e-01 6.35274470e-01
1.36615670e+00 4.23001319e-01 -9.18449312e-02 -1.00883715e-01
1.44818735e+00 -5.66539288e-01 -9.80047941e-01 -2.73711264e-01
8.85924280e-01 -7.12180316e-01 1.00049734e+00 3.50223571e-01
-5.23856223e-01 -6.08628213e-01 -1.22314179e+00 -5.68886638e-01
-9.09249008e-01 -2.01552249e-02 1.04664171e+00 2.99571902e-01
-7.89938927e-01 2.78649271e-01 -4.26303893e-01 -3.42836291e-01
4.80580002e-01 1.28799468e-01 -6.98891461e-01 -3.27415109e-01
-1.19824767e+00 1.07791436e+00 4.05654728e-01 5.00932157e-01
-2.27819309e-01 -5.60150504e-01 -8.52574646e-01 4.22226638e-03
9.77842748e-01 -4.55304742e-01 8.57536256e-01 -2.06471514e-02
-8.26619148e-01 6.87392414e-01 -6.55415729e-02 -1.36354849e-01
3.11196417e-01 -4.98464793e-01 -5.29717565e-01 -9.46915299e-02
-1.98581889e-02 1.65919063e-03 3.90717655e-01 -1.79771209e+00
-3.14846784e-01 -5.54982901e-01 4.17678714e-01 4.94332761e-01
-1.13016114e-01 -1.94241658e-01 -7.62726843e-01 -6.12117767e-01
6.58622384e-01 -4.81572628e-01 7.73185790e-02 -2.15710849e-01
-9.64271486e-01 -4.97295260e-01 1.02050424e+00 -5.55545092e-01
1.45249546e+00 -1.97181141e+00 2.15623513e-01 4.10155743e-01
5.39276004e-01 7.87996799e-02 4.82243896e-02 8.38787973e-01
-1.77269116e-01 2.12881014e-01 2.44503275e-01 7.44069042e-03
2.12252233e-02 3.51181954e-01 -8.12891245e-01 -4.28778119e-02
2.61631429e-01 1.07593501e+00 -7.98152268e-01 -5.59847355e-01
3.69708426e-02 4.24799144e-01 -1.18961275e-01 2.17143178e-01
-3.18744093e-01 2.25017682e-01 -4.78571802e-01 8.63242626e-01
7.24662602e-01 -6.10055029e-01 7.72840202e-01 -6.54366553e-01
1.96727633e-01 6.21238649e-01 -1.65742230e+00 1.46812046e+00
-2.91719586e-01 5.95421493e-01 -4.55353856e-01 -9.46325779e-01
1.19832432e+00 -4.90770079e-02 2.82360762e-01 -9.60367203e-01
-2.13476986e-01 -2.50442773e-01 -1.85006797e-01 -6.34634495e-01
9.71662581e-01 4.22528088e-01 5.08551747e-02 5.77970862e-01
-1.57380089e-01 -5.57829626e-02 9.20068622e-02 8.44952524e-01
1.08200002e+00 3.27220231e-01 6.49184525e-01 1.53802782e-01
4.90244031e-01 -1.28915504e-01 3.22012275e-01 6.59029007e-01
3.58150601e-01 4.84944314e-01 1.08508122e+00 -4.49299872e-01
-4.28806782e-01 -1.05192208e+00 -1.51073307e-01 1.08459377e+00
6.04345500e-01 -1.10695016e+00 1.37744576e-01 -8.77446771e-01
1.77429214e-01 3.52840275e-01 -6.22803330e-01 -9.57130268e-02
-6.20140910e-01 -4.83188778e-01 2.49037802e-01 8.63068759e-01
1.69690400e-01 -9.36565340e-01 -2.10915118e-01 -7.55085498e-02
-1.59037009e-01 -1.51265430e+00 2.19446406e-01 4.10177439e-01
-7.51883686e-01 -1.44071484e+00 1.62423596e-01 -6.42679870e-01
6.84298754e-01 5.78174531e-01 1.51067448e+00 3.81821156e-01
-8.22669566e-02 4.79983807e-01 -6.13336921e-01 -1.44841924e-01
-2.82986462e-01 2.76222341e-02 -2.32559606e-01 -2.74256319e-01
3.33276033e-01 -3.29464436e-01 -2.82783985e-01 3.96439582e-01
-8.66440237e-01 2.31957257e-01 5.84715188e-01 6.62072301e-01
6.97402835e-01 2.97646612e-01 4.21359479e-01 -1.20230579e+00
8.37876737e-01 -4.48506415e-01 -3.87357771e-01 9.82655048e-01
-6.75322115e-01 2.54608393e-01 3.89249921e-01 -2.21593350e-01
-1.20040452e+00 -4.49186563e-01 3.54204148e-01 -2.48128437e-02
5.87749295e-02 7.68338621e-01 -5.02453446e-01 4.68145013e-01
6.41778886e-01 -4.88313697e-02 -3.82429808e-01 -4.97289002e-01
7.27987587e-01 3.78917813e-01 5.40927172e-01 -7.93297470e-01
8.71406794e-01 3.52198929e-01 3.33565593e-01 -4.59940583e-01
-9.44558144e-01 -3.95524979e-01 -9.48650599e-01 5.98058887e-02
6.22372329e-01 -9.21598613e-01 -9.47185695e-01 7.85027456e-04
-1.25206828e+00 -1.02628931e-01 -1.23763263e-01 1.95945919e-01
-1.53497502e-01 4.71579999e-01 -4.52988416e-01 -7.83554494e-01
2.95250062e-02 -9.55836058e-01 9.06952322e-01 2.82533616e-02
-3.68875772e-01 -9.10687268e-01 -1.68016374e-01 4.22424018e-01
5.40434793e-02 5.02398461e-02 1.61014366e+00 -6.59522712e-01
-1.17715001e+00 1.07094064e-01 -7.49444127e-01 -1.26636103e-01
4.38335925e-01 2.78609246e-02 -7.45106041e-01 1.81409359e-01
-4.40133452e-01 -5.54544866e-01 7.84764588e-01 -1.89260229e-01
1.16810489e+00 -1.93512663e-01 -6.26222074e-01 4.03394639e-01
1.15371931e+00 3.56041104e-01 6.58781707e-01 3.81392837e-01
1.26975048e+00 1.02455068e+00 8.57861280e-01 3.83442640e-01
9.99446094e-01 8.38304520e-01 5.50688565e-01 5.19674830e-02
-1.75378934e-01 -5.95620811e-01 -1.38111740e-01 5.71631432e-01
-2.14530200e-01 -3.85292768e-01 -1.07748640e+00 1.30727872e-01
-2.01890755e+00 -9.18563843e-01 -3.25093925e-01 1.84705281e+00
8.26423407e-01 1.53736025e-01 -3.24383318e-01 4.43010867e-01
4.56999362e-01 3.40753853e-01 -3.70678812e-01 -1.76458001e-01
-3.15958560e-01 -1.88066855e-01 -7.61009380e-02 2.55742788e-01
-8.84166002e-01 1.01074028e+00 5.63614655e+00 5.38007975e-01
-6.49705768e-01 -4.76116717e-01 5.13505280e-01 5.19951224e-01
-6.83251560e-01 1.13338262e-01 -9.59690571e-01 -2.01825276e-01
1.10901825e-01 2.69669682e-01 3.34179491e-01 6.43105745e-01
-3.25799555e-01 -2.64626533e-01 -1.50830448e+00 1.26120734e+00
1.20433353e-01 -1.58748078e+00 4.10330892e-01 -2.76686940e-02
3.46512973e-01 -5.49364328e-01 1.31635740e-01 3.65847945e-01
3.57901573e-01 -1.34077382e+00 5.15659332e-01 7.17828035e-01
6.81798816e-01 -7.06546664e-01 5.50159752e-01 1.25409588e-01
-1.51657057e+00 1.00796655e-01 -3.39568526e-01 3.31120230e-02
-1.24036953e-01 4.94458884e-01 -1.02291512e+00 1.05826306e+00
8.19886923e-01 9.24618065e-01 -9.68659163e-01 4.38682050e-01
-4.60590005e-01 -1.33988678e-01 6.94622025e-02 1.76925406e-01
-2.84870833e-01 -1.76266402e-01 1.53871685e-01 1.00537360e+00
9.57872123e-02 1.29045323e-01 -1.18278027e-01 6.49061620e-01
-2.91107446e-01 1.54815063e-01 -9.11930740e-01 -1.32594243e-01
6.24706566e-01 1.21005166e+00 -8.00762951e-01 -1.37151748e-01
-5.65572917e-01 4.99224693e-01 7.48789966e-01 4.32655811e-01
-4.22028393e-01 -1.87345166e-02 5.23033082e-01 -6.19873367e-02
3.66979480e-01 -3.94394487e-01 -6.34030581e-01 -1.20102453e+00
4.07497376e-01 -1.02874601e+00 6.83458209e-01 -9.13133204e-01
-1.31285441e+00 6.70515299e-01 1.91395476e-01 -1.35127664e+00
-2.17842430e-01 -6.74319148e-01 -2.51235574e-01 5.42355776e-01
-1.59180391e+00 -1.37340295e+00 -4.58946675e-01 6.31787896e-01
2.58261651e-01 -2.38489285e-01 7.54822493e-01 1.55034363e-02
-5.58873177e-01 5.30215919e-01 -3.37524384e-01 1.72823623e-01
7.64601588e-01 -1.32167089e+00 2.64639854e-01 6.69137836e-01
6.01043046e-01 1.09764314e+00 3.66911292e-01 -7.32560515e-01
-1.76443827e+00 -7.36093521e-01 1.07034242e+00 -8.84450674e-01
7.24042773e-01 -4.93049562e-01 -1.26958239e+00 9.35078204e-01
2.28561722e-02 1.95091486e-01 7.24456787e-01 9.06867623e-01
-1.10395253e+00 -2.46887982e-01 -6.26282275e-01 7.96371698e-01
1.52356899e+00 -8.94594669e-01 -8.09434175e-01 2.74745435e-01
8.25575829e-01 -6.50347888e-01 -1.01998079e+00 6.02287233e-01
4.40405101e-01 -9.51286912e-01 1.25863218e+00 -7.48421133e-01
6.35864139e-01 -5.88782430e-01 -4.10102010e-01 -9.21764910e-01
-1.06162392e-01 -3.21659178e-01 -7.13968396e-01 1.66484094e+00
4.87636179e-01 -2.36928046e-01 5.79481721e-01 5.71405292e-01
8.11726004e-02 -9.54938352e-01 -6.44606650e-01 -6.31439507e-01
-3.98746729e-01 -6.48477852e-01 8.69090319e-01 1.31822634e+00
3.81085783e-01 7.48437047e-01 -2.80454546e-01 4.24161285e-01
9.78665203e-02 5.04517496e-01 9.82636213e-01 -1.32225227e+00
-2.81920522e-01 -2.88755178e-01 -3.49058241e-01 -1.27691329e+00
-1.66297369e-02 -1.02798319e+00 -4.41860378e-01 -2.19942117e+00
1.78049207e-01 -8.71872485e-01 -3.30152690e-01 5.54400086e-01
-2.50370085e-01 -2.48297930e-01 2.56379366e-01 3.50144267e-01
-8.24695349e-01 4.11357939e-01 1.37945998e+00 -4.70152378e-01
-3.76183957e-01 -8.03059563e-02 -1.13593078e+00 4.87103999e-01
3.60704690e-01 -1.93076119e-01 -9.06176925e-01 -4.24860299e-01
1.07821739e+00 1.94777548e-01 2.94757426e-01 -5.48274755e-01
3.43536913e-01 -1.99674085e-01 5.72186530e-01 -1.16445518e+00
2.92048395e-01 -9.39396441e-01 7.41329370e-03 -1.53839886e-01
-4.53951299e-01 5.06954432e-01 5.90407178e-02 7.61859119e-01
-3.92140061e-01 1.15253404e-01 -5.73747978e-02 1.14609726e-01
-9.79259789e-01 1.16776014e-02 8.29209164e-02 2.23582983e-02
7.82527924e-01 -2.39468604e-01 -1.04335141e+00 -2.28838801e-01
-5.15809894e-01 1.77573040e-01 6.81658164e-02 7.86670983e-01
8.52586627e-01 -1.26505446e+00 -2.18270689e-01 1.16271928e-01
5.98092377e-01 5.08055270e-01 2.69679993e-01 5.38720131e-01
-2.12049350e-01 4.97422487e-01 -2.59322003e-02 -5.94561756e-01
-1.24434614e+00 8.74434114e-01 -1.08132817e-01 -7.45453775e-01
-7.94817686e-01 6.78778827e-01 2.20417410e-01 -3.97458255e-01
4.61766690e-01 -6.57402456e-01 -8.21359754e-01 4.68105406e-01
5.35848141e-01 1.85733169e-01 4.49351579e-01 -4.80270505e-01
-5.28636575e-01 4.83737886e-01 -4.60180104e-01 3.88693750e-01
1.28877151e+00 -1.87156379e-01 -2.46397108e-01 4.59322602e-01
8.10435236e-01 2.24963874e-01 -1.05230570e+00 -5.63896120e-01
4.97005701e-01 -5.13997495e-01 -3.30764621e-01 -9.51521873e-01
-7.81350493e-01 6.17572248e-01 3.22851725e-03 4.26108748e-01
8.64251733e-01 3.46932083e-01 2.88667649e-01 8.68173659e-01
2.43969291e-01 -8.46387088e-01 3.98669422e-01 7.19917774e-01
1.14527643e+00 -1.41990888e+00 6.34471059e-01 -8.83036733e-01
-9.31703508e-01 1.13271260e+00 8.03499222e-01 5.79815328e-01
5.51557183e-01 3.89989585e-01 1.31906644e-01 -5.17872274e-01
-8.98040533e-01 -3.09288800e-01 6.09282494e-01 9.79527414e-01
5.12697577e-01 -1.66107088e-01 9.80751365e-02 4.72697884e-01
-1.77215606e-01 -4.08520371e-01 2.57459402e-01 9.82136369e-01
8.10656101e-02 -1.22342980e+00 -2.99968332e-01 3.86391342e-01
5.33659607e-02 -1.94309592e-01 -6.24065459e-01 1.02039826e+00
-4.45327461e-02 9.75304067e-01 3.18558961e-02 -4.41130549e-01
5.74419260e-01 -2.19506428e-01 4.57442552e-01 -6.17017269e-01
-3.44028056e-01 -3.09420913e-01 3.35142493e-01 -4.91269439e-01
-4.83457863e-01 -4.55776572e-01 -1.37086976e+00 -1.70526445e-01
-1.54079720e-01 -2.13577718e-01 3.03979039e-01 1.21459591e+00
3.07302237e-01 7.82934785e-01 3.94320011e-01 -1.59471497e-01
-2.85815179e-01 -6.47893548e-01 -4.31131363e-01 5.24684250e-01
3.08673382e-01 -8.43797088e-01 1.31020740e-01 -7.98611492e-02] | [9.175243377685547, 8.216641426086426] |
a64c8ceb-32cd-4872-bb60-b26e71c3f725 | promptclass-weakly-supervised-text | 2305.13723 | null | https://arxiv.org/abs/2305.13723v1 | https://arxiv.org/pdf/2305.13723v1.pdf | PromptClass: Weakly-Supervised Text Classification with Prompting Enhanced Noise-Robust Self-Training | Recently proposed weakly-supervised text classification settings train a classifier using the label name of each target class as the only supervision. Such weakly-supervised settings have been gaining increasing attention since they can largely reduce human annotation efforts compared to fully-supervised and semi-supervised settings. Most existing methods follow the strategy that first uses the label names as static features to generate pseudo labels, which are then used for classifier training. While reasonable, such a commonly adopted framework suffers from two limitations: (1) words can have different meanings in different contexts, so using label names for context-free matching can induce very noisy pseudo labels; and (2) the errors made in the pseudo label generation stage will directly propagate to the classifier training stage without a chance of being corrected. In this paper, we propose a new method, PromptClass, consisting of two modules: (1) a pseudo label acquisition module that uses zero-shot prompting of pre-trained language models (PLM) to get pseudo labels based on contextualized text understanding, and (2) a noise-robust self-training module that iteratively trains the classifier and updates pseudo labels by utilizing two PLM fine-tuning strategies that regularize each other. Extensive experiments show that PromptClass achieves overall better performance than existing strong baselines on four benchmark datasets and even achieves similar performance to fully-supervised classifiers on sentiment classification tasks. | ['Jiawei Han', 'Yu Zhang', 'Yu Meng', 'Minhao Jiang', 'Yunyi Zhang'] | 2023-05-23 | null | null | null | null | ['pseudo-label', 'sentiment-analysis'] | ['miscellaneous', 'natural-language-processing'] | [ 7.69970655e-01 7.42372870e-02 -3.96746218e-01 -8.72219324e-01
-8.96528602e-01 -6.32955432e-01 7.53424823e-01 4.29532051e-01
-6.77844942e-01 6.73248827e-01 1.02625497e-01 -2.59482831e-01
4.44747269e-01 -6.44892871e-01 -5.35545111e-01 -5.45991957e-01
6.50584042e-01 6.16570473e-01 2.31954157e-01 -2.34482586e-01
2.55541086e-01 -1.96446061e-01 -1.76059294e+00 6.49731159e-01
8.61982167e-01 1.00288463e+00 5.98999374e-02 3.20322871e-01
-5.90880215e-01 9.95098948e-01 -5.66408455e-01 -5.24441600e-01
2.18921602e-02 -4.99275863e-01 -1.13605118e+00 1.46119550e-01
1.56895235e-01 -7.68754482e-02 3.65535259e-01 1.09648848e+00
2.28955835e-01 2.44320825e-01 6.91714168e-01 -1.02156663e+00
-5.23769557e-01 9.62382555e-01 -4.49032187e-01 -2.52835095e-01
3.40141743e-01 -9.49940234e-02 1.22262645e+00 -1.08985913e+00
3.53089362e-01 1.32236671e+00 6.85981214e-01 8.51233900e-01
-1.23786724e+00 -9.21336889e-01 4.14070696e-01 1.54942432e-02
-1.21470237e+00 -2.46532381e-01 7.34906912e-01 -2.56550372e-01
8.09950650e-01 6.43067211e-02 1.19572207e-01 1.34412062e+00
-2.60389149e-01 7.94561207e-01 1.42043138e+00 -9.75818992e-01
3.91369879e-01 6.83009744e-01 6.99581087e-01 5.17107189e-01
-8.71141721e-03 -6.64256439e-02 -5.20581126e-01 -3.22874904e-01
-9.01108906e-02 5.76738603e-02 -1.44962609e-01 3.24098277e-03
-1.09124374e+00 9.44630265e-01 2.02994525e-01 4.21332538e-01
-5.49833737e-02 -1.14733592e-01 5.32175124e-01 3.81727785e-01
8.15154672e-01 4.15947288e-01 -8.51393402e-01 -1.59108602e-02
-7.47947454e-01 -1.93059947e-02 8.43636692e-01 9.66081500e-01
1.07312846e+00 -5.21299064e-01 -2.66007036e-01 1.16640210e+00
4.15157706e-01 3.57920110e-01 8.67291033e-01 -3.90572041e-01
4.62165385e-01 1.02810907e+00 1.59831956e-01 -6.85994148e-01
-4.89800602e-01 -3.22949260e-01 -5.91882765e-01 -1.92061663e-01
3.95784646e-01 -1.56108752e-01 -9.29968357e-01 1.80324173e+00
3.65948766e-01 1.73433185e-01 1.18976817e-01 5.60497224e-01
9.04103398e-01 5.95605493e-01 4.68895465e-01 -2.22700730e-01
1.30922961e+00 -1.39093041e+00 -8.46057415e-01 -6.72758698e-01
1.20350111e+00 -8.27751517e-01 1.49348795e+00 1.82252929e-01
-4.81719971e-01 -5.29191196e-01 -9.80802238e-01 3.67560871e-02
-6.67268753e-01 1.98921427e-01 3.90836895e-01 6.69300437e-01
-7.45140612e-01 3.61284494e-01 -5.18937588e-01 -2.81105608e-01
3.49185914e-01 1.49028078e-01 -1.16260670e-01 -1.38985336e-01
-1.31482399e+00 7.63585389e-01 5.58175981e-01 -1.06889307e-01
-6.80445015e-01 -5.62921524e-01 -9.72283185e-01 -9.73558500e-02
6.14270210e-01 -3.36002350e-01 1.62448585e+00 -1.48475564e+00
-1.50572884e+00 1.29019320e+00 -4.70299333e-01 -5.33219241e-02
3.24150503e-01 -2.49931529e-01 -4.35847551e-01 -2.01536715e-01
3.80315959e-01 6.31320059e-01 9.82138157e-01 -1.43049181e+00
-9.64192927e-01 -2.26996109e-01 -8.01249072e-02 3.31784129e-01
-5.81051528e-01 3.56592312e-02 -3.85622114e-01 -7.72781610e-01
1.30119398e-01 -1.07025170e+00 -2.23300830e-01 -3.40791702e-01
-4.97533977e-01 -6.07303262e-01 8.77868891e-01 -5.53403385e-02
1.19528782e+00 -1.90149415e+00 -3.43313903e-01 7.82946125e-02
-5.61570600e-02 5.04864812e-01 -3.25893879e-01 1.72393411e-01
-2.75306970e-01 1.17228478e-01 -3.90755922e-01 -7.08059430e-01
-1.30442247e-01 2.49135971e-01 -5.69113016e-01 -2.38740519e-02
2.54187435e-01 8.93173873e-01 -1.25097287e+00 -3.23582679e-01
1.46731287e-01 1.45337760e-01 -2.06927925e-01 3.42830539e-01
-3.85789573e-01 4.16643620e-01 -3.40071052e-01 4.23077434e-01
5.87617517e-01 -5.08645594e-01 1.97879344e-01 1.62318684e-02
3.23893785e-01 4.87761855e-01 -1.26792037e+00 1.44950485e+00
-7.35425174e-01 1.37625352e-01 -1.51178807e-01 -1.22165716e+00
9.99866784e-01 2.66399652e-01 3.78631540e-02 -3.86925191e-01
3.26364815e-01 1.97974324e-01 -6.01524353e-01 -3.88909340e-01
3.28867137e-01 -3.05113375e-01 -1.87753767e-01 9.50495958e-01
1.16406426e-01 -2.22204491e-01 2.01774269e-01 3.00740272e-01
9.51000810e-01 5.39391749e-02 2.42405266e-01 -1.34570524e-01
8.42054963e-01 1.20330378e-01 6.59116149e-01 8.89788091e-01
-1.68333918e-01 4.44328964e-01 2.13550240e-01 -4.03308868e-01
-5.34139872e-01 -5.45736551e-01 -1.90194532e-01 1.63887763e+00
2.19218373e-01 -5.30036032e-01 -7.15728104e-01 -1.35075450e+00
-1.23469174e-01 9.33084369e-01 -7.46399939e-01 -3.95245820e-01
-2.60343641e-01 -7.89342105e-01 3.42366606e-01 5.67357242e-01
4.37276751e-01 -1.17688715e+00 -2.19219536e-01 2.16023356e-01
-3.32570285e-01 -1.18345296e+00 -5.47020674e-01 6.78006709e-01
-7.35218406e-01 -1.06828606e+00 -3.10857654e-01 -1.06594431e+00
1.09020770e+00 3.99683416e-01 1.19010043e+00 2.71866351e-01
1.61245719e-01 2.70586610e-01 -7.64909387e-01 -5.18573046e-01
-8.18290532e-01 7.46008158e-02 9.70531106e-02 2.81518012e-01
9.48119223e-01 -1.16061375e-01 -2.46390253e-01 5.48099697e-01
-9.79983032e-01 1.68714657e-01 4.38838243e-01 1.17468560e+00
5.86351037e-01 3.72664221e-02 8.45053494e-01 -1.70836234e+00
5.44655323e-01 -4.55074519e-01 -3.89105409e-01 3.73325735e-01
-1.09361124e+00 1.19520985e-02 9.42146659e-01 -5.64011037e-01
-1.33790362e+00 2.59664387e-01 -1.15692817e-01 1.16247840e-01
-4.28496808e-01 4.57143039e-01 -1.19929299e-01 2.52446830e-01
8.47134709e-01 8.01533312e-02 -2.21652791e-01 -4.35542047e-01
5.49887717e-01 1.17333734e+00 2.85364598e-01 -5.73517203e-01
8.57409060e-01 4.56761777e-01 -5.27283907e-01 -4.01865959e-01
-1.90814745e+00 -7.28417695e-01 -8.48692298e-01 -5.65093867e-02
5.76488793e-01 -1.01422060e+00 -9.51413587e-02 6.44897819e-01
-9.83905613e-01 -4.74262118e-01 -2.94563740e-01 1.98497608e-01
-9.62957814e-02 2.57620305e-01 -6.46699905e-01 -6.28030777e-01
-3.28440040e-01 -1.03011441e+00 1.12998080e+00 4.19476092e-01
-3.93823445e-01 -1.28137445e+00 -9.06017423e-02 6.81491315e-01
3.94577920e-01 -1.97738051e-01 8.72391462e-01 -1.17062652e+00
7.91441202e-02 -4.16149646e-01 -2.16771111e-01 7.48791993e-01
3.62983644e-01 -4.02586550e-01 -1.41232896e+00 -2.53730267e-01
-9.84828472e-02 -9.79205370e-01 8.57154608e-01 -1.01995870e-01
1.07561493e+00 -1.92353010e-01 -5.18622994e-01 1.97754204e-01
1.07309222e+00 -1.88135989e-02 -1.44133950e-02 3.16780299e-01
7.01727152e-01 7.64567494e-01 8.26894760e-01 2.89291590e-01
4.90226030e-01 5.01839638e-01 4.10762988e-02 -1.84026852e-01
-7.14238733e-02 -3.79494518e-01 4.03767645e-01 9.21149313e-01
6.32695794e-01 -5.58014326e-02 -8.40666771e-01 4.70295221e-01
-2.03517914e+00 -4.84996229e-01 -1.00706019e-01 2.13030076e+00
1.46848333e+00 3.43401432e-01 -2.55696684e-01 2.40650818e-01
8.87456238e-01 -2.11697184e-02 -5.46608150e-01 -2.50602156e-01
-2.33834669e-01 3.78863066e-01 1.77266195e-01 5.21572888e-01
-1.29192424e+00 1.18192387e+00 5.71638393e+00 8.13280404e-01
-1.02148116e+00 3.83974940e-01 8.42853606e-01 1.34833083e-01
-2.08987802e-01 2.42259651e-01 -1.00949025e+00 3.39583248e-01
8.02298069e-01 5.31144068e-02 9.45545658e-02 9.82111931e-01
-2.09134370e-02 -1.93238914e-01 -1.26956928e+00 8.61007452e-01
2.79702574e-01 -8.54971707e-01 1.19588628e-01 -4.78020489e-01
9.68503773e-01 -7.86641985e-02 -1.88747048e-01 6.54916823e-01
5.91275930e-01 -8.28578949e-01 7.39714146e-01 1.09494001e-01
7.72725880e-01 -5.82325399e-01 9.34971690e-01 5.92279196e-01
-9.84071910e-01 -2.18965083e-01 -3.10937047e-01 -2.08676800e-01
-4.31080442e-03 8.91137958e-01 -7.49724805e-01 2.25116283e-01
6.11187696e-01 7.87127972e-01 -7.42526054e-01 4.38932508e-01
-8.58200967e-01 1.12179351e+00 -2.21775830e-01 -2.33694628e-01
2.99813151e-01 1.08225890e-01 1.55296102e-01 1.43870747e+00
-3.06008786e-01 1.08287390e-03 4.94732946e-01 6.58149838e-01
-3.22443634e-01 4.11897838e-01 -2.34067053e-01 1.69725701e-01
5.97966969e-01 1.50979877e+00 -9.28603709e-01 -6.86159432e-01
-6.88894272e-01 9.54639614e-01 3.99192274e-01 3.07355016e-01
-4.89384085e-01 -4.42999989e-01 2.47101218e-01 -2.48748392e-01
6.53451607e-02 3.61227959e-01 -3.98117244e-01 -1.29816496e+00
1.13088302e-02 -8.25960100e-01 6.40718877e-01 -5.89118421e-01
-1.53027070e+00 5.73439896e-01 -2.51522511e-01 -1.01185536e+00
-2.01311007e-01 -5.77434897e-01 -5.37530780e-01 6.70084000e-01
-1.76847541e+00 -1.15860641e+00 -4.40470695e-01 4.01762813e-01
7.89063752e-01 -1.75370369e-02 1.23589551e+00 1.81558028e-01
-7.15987802e-01 8.51659656e-01 2.54583415e-02 2.23501295e-01
1.19864464e+00 -1.39018500e+00 1.87109724e-01 6.59363210e-01
2.30567932e-01 4.75268811e-01 4.86409396e-01 -6.39469326e-01
-8.07465136e-01 -1.37561595e+00 1.21889973e+00 -7.27110922e-01
5.64993620e-01 -5.97180784e-01 -1.04498005e+00 6.68959975e-01
2.25415044e-02 1.51946023e-01 1.03103745e+00 2.23700508e-01
-6.85775995e-01 -7.16016293e-02 -9.72932875e-01 3.99079204e-01
7.98963904e-01 -5.82318425e-01 -7.61180639e-01 6.89862430e-01
7.56528437e-01 -1.76902056e-01 -2.95258284e-01 4.17951971e-01
1.86959475e-01 -5.93956769e-01 3.90397161e-01 -5.57183981e-01
2.71759778e-01 -3.23531806e-01 5.49235828e-02 -1.37544537e+00
8.84710532e-03 -5.56096375e-01 1.85662612e-01 1.63795590e+00
7.91394055e-01 -6.27812684e-01 5.96176505e-01 7.30437219e-01
7.95257092e-02 -6.32583261e-01 -6.11708701e-01 -6.22725010e-01
1.23703182e-02 -6.79877400e-01 4.69007730e-01 1.38025248e+00
1.82929248e-01 9.07698452e-01 -7.84870535e-02 -1.19896352e-01
4.24215853e-01 1.56454414e-01 6.12898886e-01 -1.24067557e+00
-9.22505930e-03 -2.51370102e-01 1.29996762e-01 -1.21935081e+00
6.25465512e-01 -1.13727534e+00 4.79865193e-01 -1.16297257e+00
4.23440248e-01 -8.76860380e-01 -4.56060886e-01 1.00104451e+00
-6.64151490e-01 3.94870490e-01 -3.49542052e-01 2.84751713e-01
-9.09839153e-01 4.64924246e-01 8.24034333e-01 -1.53395504e-01
-2.63957977e-01 6.92662522e-02 -9.68528628e-01 9.68044460e-01
5.41672349e-01 -8.50232422e-01 -4.11558568e-01 -3.58188003e-01
3.07604015e-01 -5.49454689e-01 -1.24059290e-01 -5.90075731e-01
1.88812748e-01 -5.81631884e-02 7.69573301e-02 -2.72475868e-01
-1.01624057e-01 -7.88273573e-01 -5.90904713e-01 2.19254568e-01
-8.44761491e-01 -3.23728383e-01 1.85842123e-02 6.76345229e-01
-2.47532725e-01 -6.79576099e-01 9.16752398e-01 -1.09413698e-01
-6.87484741e-01 -3.37112322e-02 -4.83610958e-01 2.85442919e-01
9.87483799e-01 8.94249603e-02 -2.34286934e-01 -2.83488780e-01
-5.50680816e-01 4.37937737e-01 3.93367678e-01 7.67197371e-01
1.76923081e-01 -1.15150654e+00 -6.38726413e-01 1.44398600e-01
6.91127837e-01 2.24210843e-01 -1.93404824e-01 3.75373632e-01
8.98984447e-02 2.65848398e-01 5.89438796e-01 -6.79878950e-01
-1.09610021e+00 5.24761617e-01 1.92552149e-01 -4.67076719e-01
-4.36725676e-01 1.09891522e+00 2.17724338e-01 -9.94337082e-01
4.40090835e-01 -1.48381352e-01 -3.58712435e-01 1.81462511e-01
7.94413865e-01 -1.74166769e-01 2.93793112e-01 -7.37808287e-01
-3.44073683e-01 4.99978870e-01 -4.47456777e-01 3.23527753e-02
1.07778752e+00 -3.05483550e-01 -1.62593797e-01 7.27920473e-01
1.06761563e+00 -1.82653829e-01 -1.00126815e+00 -9.36175048e-01
5.06483316e-01 -1.13171078e-01 1.43432349e-01 -1.14083350e+00
-7.97352195e-01 6.94924116e-01 4.81011778e-01 4.93139997e-02
1.12486458e+00 1.30392894e-01 8.88366044e-01 5.38269401e-01
2.22734362e-01 -1.44821262e+00 2.77073741e-01 7.92879045e-01
3.28507602e-01 -1.56429327e+00 -3.47775728e-01 -5.55923760e-01
-7.12525189e-01 1.06190598e+00 8.08391929e-01 2.89345860e-01
7.89482415e-01 2.99273431e-01 7.42714286e-01 2.31322944e-02
-9.39366877e-01 -1.92139581e-01 2.47004688e-01 3.46265227e-01
5.98816335e-01 -9.44203362e-02 -3.67718905e-01 9.99910355e-01
1.03812274e-02 -5.78531511e-02 2.96160311e-01 1.18780482e+00
-5.07427931e-01 -1.46198654e+00 -2.89658815e-01 5.83992362e-01
-3.80879521e-01 -1.77936554e-01 -5.38035095e-01 2.04253852e-01
2.03235835e-01 1.42366290e+00 -1.94345057e-01 -3.57049853e-01
3.46491992e-01 5.65230072e-01 -3.41731794e-02 -1.24779344e+00
-7.48319745e-01 -2.15227932e-01 5.55885434e-02 -4.18095589e-01
-6.63563490e-01 -5.52164733e-01 -1.33241844e+00 3.30598354e-01
-9.03180003e-01 4.09015208e-01 6.57835782e-01 1.46274841e+00
2.17498571e-01 2.84545094e-01 9.16916668e-01 -4.62125987e-01
-8.53822947e-01 -1.24981976e+00 -3.36676478e-01 9.76487875e-01
9.83851478e-02 -6.61978543e-01 -6.05401039e-01 3.31749856e-01] | [10.666293144226074, 7.3118367195129395] |
885fc4d8-9124-48e9-a457-13f610ee839c | dpw-sdnet-dual-pixel-wavelet-domain-deep-cnns | 1805.10558 | null | http://arxiv.org/abs/1805.10558v1 | http://arxiv.org/pdf/1805.10558v1.pdf | DPW-SDNet: Dual Pixel-Wavelet Domain Deep CNNs for Soft Decoding of JPEG-Compressed Images | JPEG is one of the widely used lossy compression methods. JPEG-compressed
images usually suffer from compression artifacts including blocking and
blurring, especially at low bit-rates. Soft decoding is an effective solution
to improve the quality of compressed images without changing codec or
introducing extra coding bits. Inspired by the excellent performance of the
deep convolutional neural networks (CNNs) on both low-level and high-level
computer vision problems, we develop a dual pixel-wavelet domain deep
CNNs-based soft decoding network for JPEG-compressed images, namely DPW-SDNet.
The pixel domain deep network takes the four downsampled versions of the
compressed image to form a 4-channel input and outputs a pixel domain
prediction, while the wavelet domain deep network uses the 1-level discrete
wavelet transformation (DWT) coefficients to form a 4-channel input to produce
a DWT domain prediction. The pixel domain and wavelet domain estimates are
combined to generate the final soft decoded result. Experimental results
demonstrate the superiority of the proposed DPW-SDNet over several
state-of-the-art compression artifacts reduction algorithms. | ['Shuhua Xiong', 'Xiaohai He', 'Truong Q. Nguyen', 'Linbo Qing', 'Honggang Chen'] | 2018-05-27 | null | null | null | null | ['jpeg-artifact-correction'] | ['computer-vision'] | [ 7.80459702e-01 -3.47367138e-01 -1.43135667e-01 -3.62523884e-01
-5.38253307e-01 3.18099469e-01 1.93078592e-01 -1.40387475e-01
-3.89437199e-01 4.74668413e-01 4.03725863e-01 -7.65907317e-02
1.92749292e-01 -9.69199717e-01 -8.71857643e-01 -6.51791573e-01
-4.87553217e-02 -3.78779799e-01 4.14818913e-01 -1.57464236e-01
2.76395738e-01 1.15065657e-01 -1.30828345e+00 8.04277897e-01
8.33737731e-01 1.62342167e+00 6.01490974e-01 8.22280347e-01
8.80668231e-04 1.08982635e+00 -3.38520825e-01 -2.95447022e-01
4.90878552e-01 -5.39725363e-01 -3.09597820e-01 1.26366958e-01
4.64199066e-01 -1.11341631e+00 -1.15046799e+00 1.68958783e+00
4.35984462e-01 -1.59221441e-01 8.50631818e-02 -6.97147787e-01
-1.01141369e+00 4.72623855e-01 -4.72983420e-01 2.59393126e-01
6.09789565e-02 4.10351068e-01 5.42190373e-01 -8.14733505e-01
5.17963231e-01 1.40989578e+00 7.55911350e-01 2.61650622e-01
-1.02589452e+00 -7.92172909e-01 -3.83846015e-01 6.93336427e-01
-1.15520370e+00 -5.06875336e-01 9.00729597e-01 1.09682620e-01
1.08933997e+00 -1.54577300e-01 7.39838421e-01 8.04745555e-01
5.87326109e-01 5.54218829e-01 7.23281682e-01 -2.23829672e-01
2.27976084e-01 -5.97337544e-01 -2.99547464e-01 5.27237296e-01
3.42040926e-01 1.72047645e-01 -5.11368692e-01 3.18607271e-01
8.70150626e-01 4.42202121e-01 -7.24528670e-01 4.20940295e-02
-1.08736086e+00 6.28623426e-01 7.40412712e-01 2.36884311e-01
-5.19615233e-01 5.98488629e-01 5.68775117e-01 6.10290408e-01
3.51237565e-01 -3.84377152e-01 -1.15894452e-01 -1.17680870e-01
-1.46717680e+00 1.28458172e-01 4.27228898e-01 8.49418104e-01
4.95275825e-01 4.38533008e-01 -7.77415410e-02 1.08718312e+00
3.85141164e-01 4.85133111e-01 8.11037362e-01 -1.24863839e+00
7.96689570e-01 1.97098061e-01 -1.68343395e-01 -1.29901528e+00
1.61900729e-01 -4.28065211e-01 -1.61329961e+00 3.73148024e-01
-2.62210041e-01 4.29739624e-01 -9.18192208e-01 1.25805318e+00
-4.95503008e-01 2.60792345e-01 1.47492096e-01 1.08936191e+00
6.18139803e-01 1.24421966e+00 -1.47477493e-01 -3.18121225e-01
1.23708189e+00 -1.03205585e+00 -9.47700620e-01 -4.03046995e-01
-1.44045977e-02 -8.45583320e-01 5.35909772e-01 5.84468544e-01
-1.64001775e+00 -1.26347744e+00 -1.68280900e+00 -6.06038392e-01
6.61397800e-02 -1.41724627e-02 -1.28178045e-01 6.51414618e-02
-1.00901818e+00 1.03027213e+00 -9.05757070e-01 1.97367698e-01
7.82167912e-01 1.46824390e-01 -1.75206512e-01 -6.27878606e-01
-1.42914259e+00 7.42660940e-01 5.70107043e-01 7.70549104e-02
-1.11587203e+00 -4.54644144e-01 -9.02053297e-01 5.79014480e-01
-1.16741888e-01 -3.97448689e-01 1.18802619e+00 -9.90365565e-01
-1.12882924e+00 6.02442443e-01 6.05947971e-02 -9.39711273e-01
4.46549207e-01 -4.18163165e-02 -7.11237788e-01 6.93780065e-01
-6.51822090e-02 6.75107956e-01 1.42576313e+00 -8.20201695e-01
-7.14951515e-01 -8.53993446e-02 -5.14803231e-01 -6.52641058e-02
-4.09156382e-01 -3.85804653e-01 -5.02792299e-01 -9.99319911e-01
4.00789410e-01 -1.27406731e-01 1.42550856e-01 6.53163493e-01
-1.32450104e-01 2.63284236e-01 1.37498593e+00 -1.20366776e+00
1.33243918e+00 -2.55637956e+00 8.40952247e-02 -3.46230835e-01
3.87716293e-01 6.74906313e-01 -4.26750988e-01 2.10249871e-01
-3.24771494e-01 -7.68514499e-02 -6.68722749e-01 -2.72367388e-01
-2.85524189e-01 1.63786203e-01 -3.72933596e-01 4.54401523e-01
2.74053186e-01 7.09200084e-01 -6.96817458e-01 -3.88170779e-01
5.48505962e-01 7.62363553e-01 -5.26246190e-01 2.86908776e-01
-2.12178275e-01 -1.27013922e-01 -1.14044592e-01 3.68883193e-01
1.36255431e+00 -2.24649325e-01 4.19074530e-03 -6.17809415e-01
-9.61353257e-02 4.34796363e-01 -8.35330546e-01 1.86364877e+00
-4.32151169e-01 1.00301635e+00 3.08558047e-01 -1.20213675e+00
9.49030459e-01 1.69443116e-01 2.29871318e-01 -1.17892349e+00
2.23206431e-01 4.76093560e-01 -2.91893277e-02 -6.28456533e-01
4.58468288e-01 -3.33811045e-01 2.57662922e-01 -2.77768169e-02
1.83626696e-01 -3.20432156e-01 -4.68749404e-02 4.93355729e-02
1.18742919e+00 -7.59605691e-02 2.16789350e-01 7.28271157e-02
6.07730508e-01 -4.40283626e-01 5.93470097e-01 4.29659605e-01
-3.38649333e-01 1.01402867e+00 3.16981375e-01 -6.75056040e-01
-1.59621799e+00 -1.08422089e+00 -2.03245983e-01 4.87958401e-01
4.30117935e-01 -2.06039056e-01 -5.98717153e-01 2.44440008e-02
-1.93777397e-01 6.55208290e-01 -7.04179853e-02 -6.42660141e-01
-7.51922250e-01 -2.02594429e-01 4.09207642e-01 2.53797442e-01
1.70612419e+00 -1.05413210e+00 -6.99573636e-01 6.24952614e-01
-5.27029276e-01 -1.24124515e+00 -5.99610209e-01 2.63962328e-01
-1.01678073e+00 -8.29932868e-01 -9.84726489e-01 -1.20370567e+00
2.49994546e-01 3.65922272e-01 7.06109941e-01 3.37615430e-01
-1.85323745e-01 -3.69272411e-01 -4.52849567e-01 1.04991861e-01
-7.04234242e-01 -4.64265615e-01 -3.91406059e-01 4.59222682e-02
3.78056020e-01 -8.33771408e-01 -1.04978430e+00 -2.46890172e-01
-1.41286862e+00 2.80745268e-01 7.22537339e-01 9.78054881e-01
5.35320818e-01 6.68614089e-01 6.86036050e-02 -4.35582578e-01
6.97073281e-01 -9.89091769e-02 -5.58108151e-01 -5.51967137e-02
-5.04161775e-01 2.81340238e-02 1.12543774e+00 -1.27722785e-01
-8.31304967e-01 -2.13180035e-01 -4.38232362e-01 -6.83057904e-01
1.97454825e-01 3.94137830e-01 -5.99356659e-04 -6.47512600e-02
3.68029445e-01 9.32172120e-01 2.00949803e-01 -6.18341982e-01
-9.01439339e-02 9.46485043e-01 1.14417386e+00 1.20773710e-01
8.04554403e-01 4.27702963e-01 -1.66328743e-01 -5.61978459e-01
-3.85516912e-01 5.71651198e-02 -2.11616591e-01 -2.78852787e-02
1.04081190e+00 -1.24639297e+00 -3.64463955e-01 9.21957850e-01
-1.47532022e+00 -9.23795626e-02 -2.79123694e-01 4.21543181e-01
-5.65086842e-01 8.35830510e-01 -1.23547423e+00 -1.51107639e-01
-5.27720749e-01 -1.43318820e+00 9.13639069e-01 7.00267106e-02
3.19583088e-01 -5.71621776e-01 -4.11052614e-01 -1.34423077e-02
8.27768683e-01 8.52413848e-02 1.13470352e+00 1.27677262e-01
-6.94466054e-01 -2.32811347e-01 -5.45188725e-01 1.12661457e+00
3.46631967e-02 -3.77850860e-01 -6.91708505e-01 -5.19548893e-01
6.40752792e-01 -3.03516895e-01 1.40630376e+00 5.65296710e-01
1.78921235e+00 -4.39474255e-01 1.50419161e-01 1.21947837e+00
1.84612834e+00 2.45331287e-01 1.24206245e+00 2.93024957e-01
3.00989479e-01 -2.58228838e-01 2.04780072e-01 6.62787735e-01
6.26397282e-02 3.14646304e-01 7.69603133e-01 -1.54476851e-01
-7.64085889e-01 -4.39579874e-01 3.86058420e-01 1.07523894e+00
1.73028633e-01 -3.35527003e-01 -4.48292494e-01 2.23850593e-01
-1.47317290e+00 -1.12789500e+00 -2.47730631e-02 2.07049751e+00
9.71921265e-01 4.34873611e-01 -6.82158470e-01 4.21410888e-01
8.36830676e-01 7.36158013e-01 -6.77055776e-01 -3.77069265e-01
-1.60846233e-01 5.35283625e-01 7.01837063e-01 3.85861099e-01
-8.96356761e-01 4.09287989e-01 5.22186613e+00 1.02673924e+00
-1.18565655e+00 7.63723329e-02 7.71629035e-01 -6.71006516e-02
3.49848270e-02 -1.83453768e-01 -2.35215500e-01 1.00522316e+00
8.82070005e-01 1.39765099e-01 7.52238393e-01 6.92382812e-01
4.46553856e-01 -1.45727828e-01 -9.88828361e-01 1.37795007e+00
6.02756217e-02 -1.62659967e+00 3.06218863e-01 -3.69095281e-02
6.89926326e-01 1.01342220e-02 2.35963091e-01 4.52704504e-02
-4.27710824e-02 -9.23944890e-01 9.94940400e-01 2.72434205e-01
1.43419015e+00 -6.53989732e-01 8.75169814e-01 2.82837778e-01
-1.21642244e+00 -2.86475927e-01 -9.18437004e-01 -4.16958258e-02
2.13923424e-01 7.78781712e-01 2.64181923e-02 8.73391256e-02
1.03918529e+00 1.32482219e+00 -1.20807670e-01 8.87269139e-01
-1.50855416e-02 5.09581745e-01 -1.13876471e-02 6.28440857e-01
3.67017984e-01 -4.75278422e-02 3.12711656e-01 1.18059731e+00
6.94377244e-01 2.19557866e-01 -2.60621965e-01 8.58521879e-01
-6.92345500e-01 -7.13831782e-01 -2.85406768e-01 7.44618997e-02
3.89365405e-01 7.92901993e-01 -1.74693748e-01 -7.74027705e-01
-5.58113337e-01 1.40240383e+00 -1.96106538e-01 3.10656816e-01
-6.60134673e-01 -7.54010856e-01 6.59174562e-01 8.09904858e-02
8.18107426e-01 -1.52766123e-01 -2.10198060e-01 -1.16629088e+00
1.48820058e-01 -1.11901736e+00 3.86902341e-03 -1.10980499e+00
-1.01052034e+00 2.55401999e-01 -4.16542858e-01 -1.79069030e+00
2.52689391e-01 -5.04012942e-01 -5.58668554e-01 1.06659579e+00
-2.17157078e+00 -6.00094378e-01 -6.40366733e-01 5.59435725e-01
1.04784417e+00 -1.94869921e-01 4.38899308e-01 4.93874043e-01
-1.74290255e-01 2.20844939e-01 4.64502394e-01 2.69776732e-01
5.58762491e-01 -5.41012406e-01 5.29066265e-01 1.16946924e+00
-6.65734410e-01 1.90171644e-01 6.25175893e-01 -6.79279327e-01
-1.46068192e+00 -1.33856857e+00 6.98011041e-01 9.30725396e-01
2.22470537e-01 2.36289762e-02 -1.11862123e+00 3.66756350e-01
5.29303133e-01 3.80920500e-01 2.35350598e-02 -1.31096423e+00
-4.00546014e-01 -4.45639551e-01 -1.43263030e+00 1.87913090e-01
7.94283450e-01 -6.40089691e-01 -5.58835924e-01 1.68529630e-01
1.00567269e+00 -4.00998443e-01 -6.86393797e-01 2.48994231e-01
3.90499890e-01 -1.37618852e+00 1.11705434e+00 1.66612491e-01
1.52071273e+00 -2.71238714e-01 -5.05554795e-01 -1.29050803e+00
-7.06751943e-01 -2.68000036e-01 -1.12385154e-01 5.90152681e-01
-2.23456040e-01 -3.73323023e-01 4.63833034e-01 6.74603507e-02
-3.74565810e-01 -5.09469151e-01 -1.05865800e+00 -6.15085423e-01
-1.29967600e-01 -4.44627076e-01 5.82878709e-01 5.61735213e-01
-2.22208545e-01 -2.23953009e-01 -7.17246115e-01 2.11540818e-01
9.00079608e-01 -7.61123821e-02 4.15800773e-02 -6.76372707e-01
-4.61200029e-01 -4.33049202e-01 -5.87758124e-01 -1.62072241e+00
-3.03978920e-01 -8.53037953e-01 2.03071505e-01 -1.58381784e+00
1.18039556e-01 2.28122815e-01 -3.76236707e-01 1.41204819e-01
2.78394759e-01 5.57574034e-01 3.98409158e-01 4.49627459e-01
-3.64051968e-01 7.54499555e-01 1.55438316e+00 -6.74166739e-01
3.08864534e-01 -4.99309063e-01 -3.79825801e-01 6.24772072e-01
5.72796285e-01 -5.12468219e-01 -1.71679020e-01 -9.57695186e-01
-6.15514293e-02 4.74131912e-01 5.22795737e-01 -1.52020049e+00
3.22439849e-01 1.57058865e-01 9.23081577e-01 -6.68876350e-01
3.16830397e-01 -1.12176788e+00 5.81440590e-02 1.15502417e+00
-6.02598846e-01 -2.54403532e-01 7.58145824e-02 6.74271703e-01
-7.83504486e-01 -3.06674480e-01 1.36896884e+00 -4.08076614e-01
-8.53048503e-01 3.66982430e-01 -3.39718074e-01 -2.22961694e-01
7.09288299e-01 -5.26994824e-01 -4.25359786e-01 -5.73321879e-01
-3.20975572e-01 -1.28347084e-01 4.64651644e-01 2.38592252e-01
1.42330647e+00 -1.36764205e+00 -8.49503100e-01 7.51406252e-01
-3.31238538e-01 6.35583848e-02 4.74476010e-01 3.00404131e-01
-1.32022536e+00 2.22342014e-01 -7.15095460e-01 -4.40621376e-01
-9.94649410e-01 4.17237014e-01 4.13884193e-01 -1.13069706e-01
-9.18796837e-01 7.74927557e-01 -1.31469503e-01 5.79739988e-01
3.55921119e-01 -6.29422188e-01 2.28749782e-01 -6.28966987e-01
8.97780478e-01 3.68897825e-01 2.11305288e-03 -3.85251373e-01
-7.87394643e-02 6.38458550e-01 -1.93632796e-01 1.42038167e-01
1.66261196e+00 -1.87814981e-01 -3.45649928e-01 -1.86452076e-01
1.82829499e+00 -6.71691895e-01 -1.57816505e+00 -3.51795495e-01
-6.03321910e-01 -7.28530765e-01 5.73745906e-01 -5.58408439e-01
-1.56146932e+00 1.33148301e+00 9.50165212e-01 1.79237098e-01
1.74668884e+00 -5.32007098e-01 1.66473687e+00 3.95078734e-02
9.35858339e-02 -1.17239463e+00 1.95709914e-01 2.09613830e-01
9.96923566e-01 -1.11733747e+00 2.50452787e-01 -1.10423818e-01
-9.32119936e-02 1.67079175e+00 3.53671402e-01 -5.52840889e-01
5.05182087e-01 3.22741687e-01 -3.42292637e-01 1.75309390e-01
-7.10793316e-01 4.63780969e-01 -2.12860033e-01 6.05507195e-01
1.58412769e-01 -1.79776162e-01 -2.76485533e-01 1.93374857e-01
-2.98226643e-02 6.30902231e-01 5.78702509e-01 7.00112939e-01
-8.11357439e-01 -8.73956382e-01 -4.11073059e-01 7.18474090e-01
-4.26590592e-01 -3.96370411e-01 3.51939261e-01 1.03363380e-01
4.53532904e-01 8.36209476e-01 2.77627438e-01 -6.24306262e-01
5.70026375e-02 -2.19745845e-01 1.79633781e-01 -1.38558283e-01
-3.20961952e-01 4.90470715e-02 -5.41465461e-01 -6.98735237e-01
-2.16369659e-01 -9.59429219e-02 -1.20518339e+00 -6.04840577e-01
1.82475045e-01 -4.98671919e-01 5.74485898e-01 5.79726636e-01
2.60131150e-01 7.86056638e-01 7.48081148e-01 -9.74802792e-01
-7.52929807e-01 -9.23612237e-01 -6.14924252e-01 7.11508751e-01
9.45769131e-01 8.51426274e-02 -5.12192905e-01 5.91064990e-01] | [11.366656303405762, -1.644039511680603] |
6599816d-bfac-4653-993b-b2561a8220fa | vinafood21-a-novel-dataset-for-evaluating | 2108.02929 | null | https://arxiv.org/abs/2108.02929v1 | https://arxiv.org/pdf/2108.02929v1.pdf | VinaFood21: A Novel Dataset for Evaluating Vietnamese Food Recognition | Vietnam is such an attractive tourist destination with its stunning and pristine landscapes and its top-rated unique food and drink. Among thousands of Vietnamese dishes, foreigners and native people are interested in easy-to-eat tastes and easy-to-do recipes, along with reasonable prices, mouthwatering flavors, and popularity. Due to the diversity and almost all the dishes have significant similarities and the lack of quality Vietnamese food datasets, it is hard to implement an auto system to classify Vietnamese food, therefore, make people easier to discover Vietnamese food. This paper introduces a new Vietnamese food dataset named VinaFood21, which consists of 13,950 images corresponding to 21 dishes. We use 10,044 images for model training and 6,682 test images to classify each food in the VinaFood21 dataset and achieved an average accuracy of 74.81% when fine-tuning CNN EfficientNet-B0. (https://github.com/nguyenvd-uit/uit-together-dataset) | ['Khang Nguyen', 'Kiet Van Nguyen', 'Nguyen D. Vo', 'Ngoc Ho', 'Vi Nguyen', 'Dung Vo', 'Thuan Q. Nguyen', 'Thuan Trong Nguyen'] | 2021-08-06 | null | null | null | null | ['food-recognition'] | ['computer-vision'] | [-7.09554195e-01 -4.51872230e-01 -3.63328964e-01 -4.57119912e-01
-3.98249090e-01 -6.17646694e-01 1.73506558e-01 4.39741105e-01
-4.54401195e-01 6.21860683e-01 3.03841770e-01 1.43464925e-02
3.42205882e-01 -1.15403092e+00 -6.74042165e-01 -7.56057918e-01
-3.38525802e-01 9.01423171e-02 -4.15384203e-01 -7.33407915e-01
-2.87320703e-01 4.47435211e-03 -1.28538144e+00 1.67903900e-01
1.03033924e+00 8.31613123e-01 4.21875954e-01 4.67915535e-01
2.39929765e-01 5.87370574e-01 -3.09524119e-01 -7.52010524e-01
5.11087596e-01 -5.43889821e-01 -1.68025002e-01 -2.53326595e-02
4.06424791e-01 -5.34874320e-01 -3.08130056e-01 1.28288424e+00
3.47751558e-01 1.39211833e-01 5.72877228e-01 -1.13497591e+00
-1.68877888e+00 9.28197801e-01 -2.31309220e-01 -1.29832834e-01
1.24605231e-01 3.54598790e-01 9.80433166e-01 -8.72359037e-01
5.76811910e-01 9.69128966e-01 5.96757472e-01 3.02212626e-01
-8.45194221e-01 -8.78945589e-01 1.39099270e-01 2.82658309e-01
-1.52886212e+00 -2.50399351e-01 4.94675398e-01 -1.68416008e-01
8.18359077e-01 8.19221511e-02 1.35764182e+00 1.08091187e+00
3.32299173e-01 9.06256735e-01 8.65510464e-01 1.37763232e-01
-1.83968935e-02 2.24058852e-01 -1.98319986e-01 8.57594371e-01
5.26882410e-01 1.73476726e-01 -3.76669355e-02 5.93927860e-01
5.29432952e-01 2.43781805e-01 -9.17574912e-02 6.57993332e-02
-1.42850029e+00 1.21863401e+00 1.42188203e+00 2.54742682e-01
-5.99157512e-01 -2.55581975e-01 5.01272380e-01 3.20477873e-01
3.87003154e-01 4.05284077e-01 -5.77293277e-01 2.69042701e-01
-2.90981948e-01 4.56184208e-01 8.72439146e-01 1.08274925e+00
5.87608337e-01 3.32214415e-01 4.65807468e-01 9.60999489e-01
3.26100796e-01 1.11219072e+00 2.99486935e-01 -6.64507866e-01
2.30977759e-01 2.80165553e-01 1.70868337e-01 -1.45157027e+00
-5.92637599e-01 -3.72547388e-01 -1.19206798e+00 -2.01453492e-01
4.88687813e-01 -2.10539088e-01 -1.01377356e+00 1.57096469e+00
2.86439747e-01 -6.94634855e-01 9.75331217e-02 1.21346855e+00
1.37039077e+00 1.08040071e+00 4.59471822e-01 2.80569136e-01
1.34848952e+00 -1.43002784e+00 -6.12894416e-01 -1.04322590e-01
4.58294988e-01 -7.51064420e-01 1.31325138e+00 6.35865629e-01
-8.42596292e-01 -6.25108540e-01 -1.19770896e+00 -3.50587577e-01
-8.75574827e-01 2.12572739e-01 8.52259457e-01 5.46348631e-01
-7.54691780e-01 5.27103841e-01 -5.62414169e-01 -7.58639574e-01
3.52056861e-01 8.57140273e-02 -3.73424292e-01 -5.88256955e-01
-1.18805265e+00 7.21127570e-01 6.92359626e-01 2.99846351e-01
-1.05759096e+00 -5.41652381e-01 -1.19395149e+00 4.78504077e-02
1.22786745e-01 -2.78222829e-01 1.07290494e+00 -1.07822216e+00
-1.37742460e+00 6.53649330e-01 4.47447181e-01 -1.20513536e-01
1.67997032e-01 -4.29943912e-02 -1.15477538e+00 1.10657752e-01
5.01620889e-01 1.28164530e+00 2.04920739e-01 -1.04528284e+00
-7.73737609e-01 -1.43080354e-01 8.01582113e-02 1.33079693e-01
-2.67043561e-01 -1.37566939e-01 -9.76786390e-02 -5.85475981e-01
-1.72768071e-01 -8.77058566e-01 -2.13346183e-01 5.82070462e-02
-4.12194699e-01 -1.55570894e-01 2.14890629e-01 -9.59390044e-01
8.00634980e-01 -2.45875621e+00 -3.75343531e-01 -4.98894453e-02
4.43725921e-02 4.93294485e-02 -6.44033253e-01 3.66811723e-01
2.51925111e-01 1.38640925e-01 1.52228177e-01 1.55397445e-01
3.00576121e-01 3.14983130e-01 3.82595450e-01 5.96672893e-01
1.57594442e-01 9.29703116e-01 -1.31730223e+00 -3.87977064e-01
4.65537339e-01 7.66466856e-01 -2.99010873e-01 -5.40633276e-02
-2.80305803e-01 2.64226943e-01 -3.83317202e-01 1.26096129e+00
8.57074797e-01 -1.79478630e-01 2.83007413e-01 -4.32576269e-01
-3.66251349e-01 1.07190080e-01 -7.65174389e-01 1.69917893e+00
-2.00956807e-01 2.93666840e-01 1.97375596e-01 -4.66564268e-01
9.77679193e-01 -1.90050304e-02 3.54512125e-01 -1.24994969e+00
6.20530605e-01 5.16322017e-01 1.62314549e-01 -4.89028454e-01
7.22097218e-01 9.00748372e-03 -2.34238252e-01 -1.04108103e-01
-3.02824583e-02 2.28388712e-01 8.30000639e-01 -1.77633435e-01
2.68559277e-01 1.41563654e-01 5.05216360e-01 -8.40526700e-01
2.90818606e-02 5.36708295e-01 7.26868749e-01 3.71209136e-03
-5.32623172e-01 2.78357863e-01 -2.76279598e-01 -9.50270772e-01
-1.20312703e+00 -1.17954564e+00 -1.76041842e-01 1.22005987e+00
2.91895419e-01 -2.80570596e-01 -3.77447754e-01 -3.38652670e-01
-3.97742465e-02 6.77262962e-01 -8.34480345e-01 -5.62368184e-02
-5.69273651e-01 -7.05317795e-01 1.15308233e-01 5.23220420e-01
1.23562586e+00 -1.20603216e+00 -3.35495353e-01 4.45750058e-01
-4.87803280e-01 -4.73801285e-01 -9.67409551e-01 4.08032596e-01
-4.26379204e-01 -9.82770741e-01 -9.13645267e-01 -1.19388616e+00
3.30882162e-01 7.65893221e-01 1.42379057e+00 -4.04915437e-02
-1.48195833e-01 -4.43911463e-01 -7.23261356e-01 -5.82980156e-01
-3.31571639e-01 1.10800818e-01 8.87927637e-02 -2.96535522e-01
8.85695696e-01 -7.03823417e-02 -9.35616732e-01 2.75670916e-01
-7.58968592e-01 -8.20534006e-02 3.33610356e-01 6.98866904e-01
7.29440749e-01 2.91951269e-01 5.53262949e-01 -4.23654169e-01
1.02894448e-01 -9.32896912e-01 -4.08662319e-01 1.58545390e-01
-4.95656490e-01 -5.68025827e-01 9.99556661e-01 -6.95875168e-01
-7.08434224e-01 1.39131919e-01 -5.97658157e-01 8.66541173e-03
-1.19538479e-01 6.07070208e-01 -2.48846412e-01 1.49824157e-01
7.05238819e-01 4.94825877e-02 -2.24730894e-01 -3.77038658e-01
6.37592614e-01 4.65882242e-01 4.84197259e-01 -2.17701733e-01
4.75894481e-01 1.94111839e-01 -5.07101059e-01 -1.05386710e+00
-8.46647024e-01 -2.31308326e-01 -4.32066917e-01 -1.75458610e-01
1.41710711e+00 -1.22347641e+00 -9.49426830e-01 4.31617051e-01
-5.96266866e-01 -6.68649375e-01 -7.59633034e-02 5.93980312e-01
-1.73487852e-03 -1.35725319e-01 -8.16957891e-01 -3.61377686e-01
-3.98109078e-01 -1.09250319e+00 5.71184218e-01 5.29013276e-01
2.54487339e-03 -9.10356343e-01 -1.84374675e-01 1.48347080e-01
4.22502428e-01 5.23451686e-01 8.14096510e-01 -2.04006225e-01
-3.66868228e-01 9.66820195e-02 -4.05176312e-01 3.37010503e-01
4.00087535e-01 5.64904138e-03 -6.33099794e-01 -4.58180875e-01
-4.02694613e-01 -5.69927812e-01 9.04847562e-01 6.29167438e-01
5.69230735e-01 -4.44911242e-01 5.37339263e-02 7.25908339e-01
1.76859212e+00 7.98644602e-01 4.46959168e-01 5.61959028e-01
6.94595218e-01 3.30138296e-01 6.99465930e-01 2.74733692e-01
9.63376999e-01 1.39635220e-01 7.06386507e-01 -5.45271814e-01
-1.20705903e-01 -5.06642222e-01 4.34566766e-01 1.10172451e+00
7.07701407e-03 -5.19988954e-01 -3.09829354e-01 8.01037490e-01
-1.28654420e+00 -7.60263383e-01 -2.80657738e-01 1.89903736e+00
5.38146079e-01 -3.70022744e-01 6.18016899e-01 -2.77064741e-01
3.81752431e-01 -6.37267157e-02 -7.22802818e-01 -2.85003304e-01
-3.78019512e-01 -1.36924490e-01 8.12355101e-01 1.27591401e-01
-1.46474314e+00 7.94666171e-01 6.14702320e+00 4.87612128e-01
-1.14445508e+00 1.38494164e-01 7.92860746e-01 -1.11458113e-03
-9.89529043e-02 -8.08831275e-01 -7.14742422e-01 3.77022952e-01
8.74872684e-01 1.31926864e-01 7.23233342e-01 9.47956681e-01
1.15224145e-01 2.07236275e-01 -7.81207681e-01 9.11514759e-01
1.77634448e-01 -1.19069338e+00 -2.66898453e-01 -8.10665078e-04
8.48740935e-01 3.45162809e-01 2.73474634e-01 5.84154844e-01
8.01213980e-01 -1.10304916e+00 1.22570336e+00 -1.16660111e-01
6.64849401e-01 -9.58423436e-01 5.21057248e-01 -1.97207909e-02
-1.62375700e+00 -3.69859785e-01 -9.08737004e-01 5.84409721e-02
1.40184229e-02 4.61890489e-01 -3.05088401e-01 4.91476923e-01
1.24157631e+00 1.19971526e+00 -4.42670733e-01 9.06881094e-01
-2.03910127e-01 3.50337297e-01 -4.76576120e-01 -5.43365121e-01
6.88973308e-01 -4.93288249e-01 -1.06783174e-01 1.16043556e+00
6.29598916e-01 1.39810860e-01 6.79568648e-01 8.66204500e-01
-1.81621343e-01 5.41286290e-01 -4.85910982e-01 -4.56414700e-01
9.69304815e-02 1.42775118e+00 -1.01534688e+00 -2.29983181e-01
-6.14860892e-01 8.50590110e-01 2.52807979e-02 2.20243201e-01
-9.36542809e-01 -2.09202260e-01 8.21383834e-01 -3.01940918e-01
3.55018705e-01 -3.14395875e-01 2.37607777e-01 -1.32314348e+00
-4.46331054e-01 -1.20265269e+00 4.27381724e-01 -5.16008794e-01
-1.70546210e+00 7.61428773e-01 -3.67171735e-01 -1.05249369e+00
3.20894450e-01 -7.66653657e-01 -1.90172046e-01 5.91461301e-01
-1.55729032e+00 -1.53867984e+00 -5.23799837e-01 6.20117307e-01
9.00648952e-01 6.69371784e-02 9.36123550e-01 6.73908949e-01
-6.49187267e-01 5.46166182e-01 4.40699935e-01 3.34677011e-01
5.70453048e-01 -1.09522498e+00 3.97244990e-01 4.16808873e-01
6.42404556e-02 3.82060200e-01 5.64983010e-01 -7.30198324e-01
-1.69380164e+00 -1.47191727e+00 1.01184762e+00 -7.33202398e-02
5.72882533e-01 -4.96903777e-01 -5.64828634e-01 5.80230296e-01
5.46455204e-01 -5.24147034e-01 7.58697927e-01 5.82743362e-02
-4.74965006e-01 -2.83000052e-01 -1.34079635e+00 6.56685650e-01
8.62656713e-01 8.39920416e-02 1.65436184e-03 6.99673235e-01
8.10570657e-01 -1.37640134e-01 -1.15680909e+00 -9.66511965e-02
8.46815646e-01 -7.53760815e-01 9.36768413e-01 -1.27847627e-01
5.12921035e-01 -1.59119889e-01 -5.98176479e-01 -1.61180353e+00
-9.31205153e-01 9.06230360e-02 6.07353687e-01 1.08384156e+00
6.61347508e-01 -5.90490341e-01 5.19060552e-01 2.41774619e-01
-3.89898509e-01 -3.91130507e-01 -2.15233624e-01 -8.00242543e-01
4.15257752e-01 2.13522449e-01 1.12995565e+00 1.26052201e+00
-3.15187335e-01 2.22856537e-01 -5.87826848e-01 -3.48171988e-03
5.85849881e-01 2.17062905e-01 4.68647093e-01 -9.17102516e-01
-7.85169452e-02 -3.67365867e-01 2.04478219e-01 -9.03326154e-01
-3.20330620e-01 -1.15607464e+00 3.24891716e-01 -1.69378722e+00
1.43161863e-01 -5.24107635e-01 -3.29680294e-01 7.85690486e-01
1.78509817e-01 6.84538364e-01 3.97268206e-01 -5.06359302e-02
-3.27573150e-01 4.26232547e-01 1.72687793e+00 -6.79562688e-01
-2.56655067e-01 -6.08314693e-01 -9.54427361e-01 2.55628914e-01
1.10088420e+00 -2.74059445e-01 -4.35241312e-01 -6.69580698e-01
3.12685966e-01 -2.63935000e-01 1.23378657e-01 -8.29163551e-01
-3.23030442e-01 -3.90640289e-01 7.94435799e-01 -6.21761560e-01
2.25084528e-01 -7.91995764e-01 6.15199804e-01 9.39821422e-01
-1.29758939e-01 2.50521213e-01 1.22050725e-01 3.41172107e-02
-1.07912309e-01 8.12265351e-02 8.13346386e-01 -5.20342588e-01
-1.03905094e+00 5.20826519e-01 -3.72863829e-01 -3.35052133e-01
1.00286603e+00 -9.01478231e-02 -5.31445563e-01 -2.95451254e-01
-5.37664354e-01 4.35322762e-01 5.31764269e-01 5.78328788e-01
5.03322244e-01 -1.63520336e+00 -1.05746067e+00 4.23194945e-01
2.96318322e-01 -2.46385664e-01 5.84748924e-01 4.94345129e-01
-1.06421697e+00 2.85233140e-01 -7.58619964e-01 -2.35484079e-01
-7.29872882e-01 1.16606379e+00 1.02300286e-01 1.31148189e-01
-9.21700537e-01 7.77120352e-01 3.59130323e-01 -4.84656632e-01
-4.16815057e-02 -6.39779389e-01 -1.59689575e-01 -2.81013027e-02
6.97655618e-01 3.12056959e-01 -1.70433745e-01 -7.93287814e-01
-1.46327749e-01 2.77201027e-01 3.52660343e-02 7.20225036e-01
1.38675559e+00 -1.30265594e-01 6.61762357e-02 2.70405293e-01
1.29145730e+00 -4.06005353e-01 -1.14196527e+00 2.65083939e-01
-6.14983082e-01 -6.04220569e-01 -9.56339985e-02 -1.17586124e+00
-1.63237298e+00 6.31071687e-01 7.65169859e-01 3.25051606e-01
1.20815325e+00 -1.67878807e-01 1.30427575e+00 4.49377596e-01
5.01739025e-01 -8.36210251e-01 -1.06736630e-01 2.90863007e-01
7.84136534e-01 -1.53694856e+00 -3.02006304e-01 -4.41122144e-01
-6.44171715e-01 8.17881525e-01 4.68086034e-01 -3.67181987e-01
7.13303149e-01 7.79170319e-02 6.11532271e-01 -9.77480263e-02
-3.45057786e-01 -2.00427204e-01 -7.07925996e-03 9.69166517e-01
4.69380945e-01 7.05583990e-01 -3.27919394e-01 5.09111822e-01
-6.07323647e-01 9.77121592e-02 2.49072447e-01 5.57797611e-01
-5.20648599e-01 -9.31174397e-01 -1.24233007e-01 3.83873463e-01
-1.71478197e-01 -2.93030620e-01 9.42430347e-02 1.26132309e+00
4.99415755e-01 1.40429223e+00 -2.77119782e-02 -4.95303392e-01
3.12479168e-01 -4.55996186e-01 4.09171790e-01 -8.85821134e-02
-9.17109728e-01 3.68594736e-01 3.39118153e-01 -4.30377662e-01
-4.71510738e-01 -6.57821357e-01 -1.09387243e+00 -1.08727419e+00
5.59197403e-02 5.32898977e-02 8.57474089e-01 3.78048927e-01
-6.70692697e-02 1.31834924e-01 5.48561692e-01 -8.17892253e-01
-3.23466748e-01 -9.53552663e-01 -1.23646903e+00 5.12270689e-01
5.70414253e-02 -6.20647371e-01 -2.26924326e-02 7.72360042e-02] | [11.553814888000488, 4.384588241577148] |
13e181cf-4c27-4804-81ea-67a463827505 | memory-efficient-fine-tuning-of-compressed | 2305.14152 | null | https://arxiv.org/abs/2305.14152v1 | https://arxiv.org/pdf/2305.14152v1.pdf | Memory-Efficient Fine-Tuning of Compressed Large Language Models via sub-4-bit Integer Quantization | Parameter-efficient fine-tuning (PEFT) methods have emerged to mitigate the prohibitive cost of full fine-tuning large language models (LLMs). Nonetheless, the enormous size of LLMs impedes routine deployment. To address the issue, we present Parameter-Efficient and Quantization-aware Adaptation (PEQA), a novel quantization-aware PEFT technique that facilitates model compression and accelerates inference. PEQA operates through a dual-stage process: initially, the parameter matrix of each fully-connected layer undergoes quantization into a matrix of low-bit integers and a scalar vector; subsequently, fine-tuning occurs on the scalar vector for each downstream task. Such a strategy compresses the size of the model considerably, leading to a lower inference latency upon deployment and a reduction in the overall memory required. At the same time, fast fine-tuning and efficient task switching becomes possible. In this way, PEQA offers the benefits of quantization, while inheriting the advantages of PEFT. We compare PEQA with competitive baselines in comprehensive experiments ranging from natural language understanding to generation benchmarks. This is done using large language models of up to $65$ billion parameters, demonstrating PEQA's scalability, task-specific adaptation performance, and ability to follow instructions, even in extremely low-bit settings. | ['Dongsoo Lee', 'Se Jung Kwon', 'Kang Min Yoo', 'Joonsuk Park', 'Sungdong Kim', 'Jung Hyun Lee', 'Jeonghoon Kim'] | 2023-05-23 | null | null | null | null | ['model-compression'] | ['methodology'] | [ 1.70483232e-01 -1.68089673e-01 -2.34051466e-01 -4.18363333e-01
-1.09150398e+00 -6.13585174e-01 5.28062880e-01 1.77329883e-01
-7.93404579e-01 7.16802299e-01 2.02417478e-01 -5.10922134e-01
9.25068334e-02 -6.29670858e-01 -8.60592246e-01 -5.12849867e-01
1.62408486e-01 5.57139516e-01 1.84021890e-01 -2.16940030e-01
1.90632313e-01 3.90020043e-01 -1.35526097e+00 2.11168647e-01
8.77891898e-01 9.72318828e-01 3.77701759e-01 7.46060610e-01
-2.33719379e-01 3.78587663e-01 -6.42509460e-01 -5.99283993e-01
2.56152302e-01 1.84736714e-01 -8.10355842e-01 -2.38773838e-01
5.94337523e-01 -4.98535514e-01 -2.69137979e-01 8.52288544e-01
5.97163558e-01 1.45364016e-01 4.33176845e-01 -1.01410198e+00
-5.56228876e-01 8.55467021e-01 -3.64674121e-01 2.79346943e-01
-2.68255502e-01 3.46079707e-01 1.11655700e+00 -7.43825495e-01
2.36751214e-01 1.47496986e+00 5.86671650e-01 5.21532595e-01
-1.47095788e+00 -7.72627950e-01 2.98290402e-01 -1.31504238e-01
-1.52483785e+00 -6.77051187e-01 1.68362215e-01 -2.62693882e-01
1.68953145e+00 6.55213296e-02 4.38188821e-01 8.37078214e-01
2.96505958e-01 6.68006539e-01 5.67995131e-01 -2.80252010e-01
3.58417034e-01 -4.83916178e-02 -3.30976360e-02 8.74863148e-01
2.21021995e-01 -6.63286150e-02 -6.51367545e-01 -2.46128663e-01
7.80584872e-01 -3.09253275e-01 -2.94255167e-02 -1.63686857e-01
-1.27556193e+00 7.59340823e-01 3.94507021e-01 3.84958163e-02
-4.11082476e-01 6.21536970e-01 7.26113856e-01 2.22965658e-01
2.25753546e-01 5.67949355e-01 -5.65765500e-01 -4.08483922e-01
-1.08923459e+00 3.47252131e-01 7.41784930e-01 9.72765863e-01
7.97795832e-01 2.67054021e-01 -4.42006141e-01 8.75414193e-01
1.76152676e-01 5.10174572e-01 6.18790090e-01 -1.23074174e+00
8.43008518e-01 4.63299632e-01 -8.22284892e-02 -6.15881324e-01
-2.91102886e-01 -6.34274364e-01 -1.00702000e+00 -6.18044697e-02
2.00650766e-01 -1.28631771e-01 -8.83124709e-01 2.10982728e+00
1.70748681e-01 -4.05002274e-02 -7.72845522e-02 5.36738396e-01
9.80706215e-02 7.97779441e-01 4.67570633e-01 -3.13978605e-02
1.58108556e+00 -1.10562503e+00 -5.76249003e-01 -6.01139963e-01
8.04687977e-01 -6.47504866e-01 1.58224046e+00 3.61462265e-01
-1.13310027e+00 -6.48961604e-01 -1.11506891e+00 -4.47210670e-01
-2.82575577e-01 2.01393247e-01 8.15114796e-01 7.29455173e-01
-1.19058239e+00 3.69317174e-01 -1.04027045e+00 1.33632854e-01
4.10184681e-01 6.91528499e-01 -2.28471443e-01 -6.83770403e-02
-1.24052238e+00 6.90100014e-01 6.59060538e-01 -1.12193145e-01
-7.51684844e-01 -1.05900085e+00 -7.67062008e-01 5.13575792e-01
2.78002709e-01 -1.03888190e+00 1.58827233e+00 -4.28529978e-01
-1.48800468e+00 3.64573568e-01 -4.20498371e-01 -8.31325889e-01
3.29930753e-01 -4.12985474e-01 -1.72111332e-01 -1.66724727e-01
-6.69950172e-02 1.04455721e+00 9.76309359e-01 -6.94028139e-01
-7.09145904e-01 -1.88069358e-01 1.11648440e-01 2.82846659e-01
-7.34660268e-01 -3.58742267e-01 -8.62638056e-01 -6.62567973e-01
-4.02924657e-01 -9.35840011e-01 -2.94254988e-01 -2.63613731e-01
-2.69011676e-01 -3.32221061e-01 3.54030907e-01 -4.72102970e-01
1.80842447e+00 -2.32676244e+00 9.60335881e-02 7.67455399e-02
3.26349497e-01 5.49513638e-01 -3.90813261e-01 2.46269107e-01
4.17913437e-01 2.04877451e-01 -3.07030290e-01 -7.23582983e-01
3.09881389e-01 3.94097805e-01 -3.35292399e-01 -7.22289681e-02
2.29824036e-01 1.19672120e+00 -7.41686940e-01 -5.41773736e-01
-4.24497575e-02 4.91276473e-01 -9.99685705e-01 -1.58747658e-02
-4.76434231e-01 -1.27762616e-01 -3.51361871e-01 3.58782351e-01
3.63847435e-01 -4.45096552e-01 1.31027728e-01 -4.60138410e-01
2.51772292e-02 5.36453784e-01 -1.13710392e+00 1.83172297e+00
-8.74179602e-01 5.35688221e-01 -5.92651172e-03 -6.49068773e-01
6.38327658e-01 2.01213807e-01 -4.57837284e-02 -8.55786741e-01
-1.31722137e-01 2.91237682e-01 -1.22382581e-01 -9.67627242e-02
8.00122678e-01 1.00087322e-01 -2.39475653e-01 6.54202044e-01
1.72131974e-02 -2.41893053e-01 4.97688055e-01 3.24398458e-01
9.38848674e-01 -8.07591379e-02 2.76819587e-01 -1.75563112e-01
3.58887762e-01 -7.22961724e-02 3.70705515e-01 8.50889504e-01
-6.29304573e-02 -4.97901365e-02 1.98810786e-01 -3.08636039e-01
-1.22727656e+00 -1.04123664e+00 6.09642155e-02 1.56321788e+00
-4.73872215e-01 -8.02246451e-01 -1.06636095e+00 -3.39422792e-01
1.33080080e-01 7.86279500e-01 -4.24046427e-01 -3.72913480e-01
-8.77564251e-01 -6.05878890e-01 7.25473583e-01 6.82114005e-01
6.80812120e-01 -8.17321062e-01 -5.70369482e-01 4.62030262e-01
-2.93187350e-01 -1.11152780e+00 -8.86982977e-01 2.72955447e-01
-1.11249328e+00 -3.53457719e-01 -4.53392029e-01 -6.30935907e-01
5.09023845e-01 1.36192665e-01 1.39546371e+00 -1.03992693e-01
-7.45761320e-02 -1.02225840e-01 2.36824930e-01 -1.78451955e-01
-4.63328063e-01 7.02184856e-01 1.11434758e-01 -3.65550548e-01
3.67971212e-01 -5.08907318e-01 -6.30408227e-01 4.12960239e-02
-9.84376252e-01 1.39883906e-01 8.83711338e-01 8.72119904e-01
8.50758553e-01 6.81081191e-02 5.96890926e-01 -9.17067885e-01
9.03581500e-01 -2.53886759e-01 -9.00179327e-01 4.26620394e-01
-8.42275739e-01 5.97572744e-01 8.72448564e-01 -4.89797711e-01
-1.05347323e+00 7.93298110e-02 -2.63922065e-01 -1.82824627e-01
1.47149682e-01 3.94876331e-01 1.73635373e-03 7.33901039e-02
7.51113176e-01 3.35317045e-01 -1.19160838e-01 -5.33516169e-01
6.53624535e-01 7.44348705e-01 7.90425360e-01 -7.11051404e-01
6.79725409e-01 1.51105106e-01 -2.63069779e-01 -5.24201274e-01
-8.27787042e-01 -2.47776613e-01 -4.80118275e-01 4.74211961e-01
6.79927707e-01 -1.21873450e+00 -7.09073842e-01 2.69669801e-01
-9.28784668e-01 -8.43942702e-01 -2.51945853e-01 3.67716968e-01
-5.04152596e-01 2.68838316e-01 -8.03903043e-01 -4.33456600e-01
-8.33933949e-01 -1.38485217e+00 1.14020586e+00 8.44146963e-03
-3.62918228e-01 -1.00476480e+00 -1.73382640e-01 4.36128199e-01
7.54707098e-01 -5.12632668e-01 1.35962439e+00 -2.47251570e-01
-6.98312581e-01 4.23220433e-02 -3.53738844e-01 3.60908985e-01
-1.22701630e-01 -2.65695304e-01 -8.11658978e-01 -5.07184684e-01
-3.46983492e-01 -3.76656353e-01 7.86552429e-01 2.71916270e-01
1.33730590e+00 -3.35576504e-01 -2.34565496e-01 9.44140434e-01
1.30105662e+00 -8.24567229e-02 2.14492321e-01 3.59260529e-01
6.37001336e-01 -5.93445636e-02 5.13878047e-01 4.11937058e-01
6.33086383e-01 7.21439004e-01 1.69719771e-01 8.76066927e-03
-2.09182531e-01 -4.47430134e-01 4.51921701e-01 1.02467644e+00
2.60579586e-01 -2.98238397e-01 -8.86544347e-01 5.03207922e-01
-1.55090559e+00 -3.53039086e-01 4.53899801e-01 2.23522520e+00
1.30277669e+00 3.93236220e-01 -6.74642175e-02 5.55550680e-02
2.21539482e-01 -1.54475337e-02 -8.93500984e-01 -8.66412640e-01
3.34915705e-02 2.54741818e-01 5.40665269e-01 5.70883036e-01
-9.24335420e-01 1.26158738e+00 6.46752071e+00 1.21459150e+00
-1.11527300e+00 1.64265200e-01 6.30872369e-01 -3.41247261e-01
-3.72789741e-01 -3.79962534e-01 -1.27824008e+00 2.67208457e-01
1.39134157e+00 -1.80617571e-01 9.12351787e-01 7.93388844e-01
-1.00108259e-03 1.41578391e-01 -1.12060678e+00 1.06219637e+00
-2.90634155e-01 -1.50646603e+00 4.58353728e-01 -8.67174845e-03
7.12204516e-01 4.04978216e-01 1.37159199e-01 6.19862318e-01
3.32753211e-01 -9.07278001e-01 7.33535051e-01 6.90184087e-02
1.19565845e+00 -9.18879449e-01 4.90239680e-01 4.23759699e-01
-1.22802532e+00 -2.99819767e-01 -5.61961532e-01 9.76350084e-02
2.91476130e-01 5.06582856e-01 -9.07310605e-01 7.38759637e-02
6.19614840e-01 1.88709125e-01 -6.34898663e-01 6.26174212e-01
-1.07256874e-01 7.08625197e-01 -5.11440754e-01 1.28565252e-01
4.65168297e-01 6.69378340e-02 1.16360463e-01 1.62877381e+00
1.93333328e-01 -1.51110385e-02 -2.44396552e-02 7.48781681e-01
-3.73244733e-01 -3.97218503e-02 -1.72571003e-01 -3.12880993e-01
9.69831288e-01 1.12857497e+00 -3.85767549e-01 -6.62312806e-01
-3.19805712e-01 1.01150179e+00 7.97765315e-01 3.22897196e-01
-7.87520885e-01 -3.74712974e-01 8.82198691e-01 -9.12828892e-02
3.09738606e-01 -5.19766331e-01 -4.01817799e-01 -9.30138111e-01
-5.79889417e-02 -1.03319323e+00 3.20533395e-01 -4.16419983e-01
-8.33282948e-01 6.88526154e-01 -2.33182088e-02 -5.66223204e-01
-4.91318256e-01 -3.00108939e-01 1.36136860e-02 9.59620595e-01
-1.63611948e+00 -9.33706224e-01 -1.38589978e-01 5.42164326e-01
7.34111369e-01 -8.10026377e-02 9.82883334e-01 5.39022744e-01
-7.02543378e-01 1.21893632e+00 2.17229575e-01 -1.76866576e-01
6.92883849e-01 -1.29843581e+00 1.13882995e+00 6.91916108e-01
4.90590930e-02 1.07136118e+00 2.53015757e-01 -4.28723603e-01
-1.47933984e+00 -1.27287626e+00 1.12033916e+00 -2.37390593e-01
5.38095415e-01 -6.47885978e-01 -9.67591286e-01 6.35871768e-01
-7.97715113e-02 -1.66812837e-01 5.90700030e-01 1.50370389e-01
-6.29923820e-01 -2.87267685e-01 -7.68335998e-01 7.51529515e-01
9.47632968e-01 -7.29881406e-01 -3.67412478e-01 3.26724827e-01
1.06116319e+00 -4.48007405e-01 -1.16397560e+00 1.70286726e-02
6.37629211e-01 -6.31842077e-01 1.26458430e+00 -5.17389596e-01
2.53121823e-01 -1.36971354e-01 -2.28484184e-01 -1.29872727e+00
-5.89653254e-01 -7.23424137e-01 -4.23597276e-01 1.21124136e+00
5.86612463e-01 -6.54419422e-01 5.40983975e-01 6.57836318e-01
-1.18855029e-01 -8.64623129e-01 -7.78765023e-01 -7.17914939e-01
1.98227182e-01 -4.78619665e-01 9.66660500e-01 5.22236824e-01
-3.89592260e-01 5.52006662e-01 -1.97101772e-01 9.03263874e-03
5.81167638e-01 -1.04710460e-01 7.31904328e-01 -7.25890219e-01
-6.41975760e-01 -5.71949780e-01 2.07791537e-01 -1.41268814e+00
9.33330655e-02 -7.98977971e-01 4.28172573e-03 -1.27222693e+00
1.77684397e-01 -7.63665736e-01 -3.35772038e-01 6.54806197e-01
-4.93971735e-01 3.61560076e-01 3.34523350e-01 1.99769229e-01
-6.62687838e-01 3.91586781e-01 1.07218468e+00 -9.58222523e-02
-2.50487626e-01 -2.21699744e-01 -6.25854909e-01 5.38313746e-01
9.30360854e-01 -3.65047395e-01 -6.38444424e-01 -1.12039244e+00
3.99587452e-01 -1.65181965e-01 -4.24180180e-03 -9.51015234e-01
3.00068706e-01 1.06305696e-01 1.77526638e-01 -3.79420996e-01
4.25173610e-01 -4.07556027e-01 3.07647679e-02 6.05469584e-01
-4.44125086e-01 3.53428870e-01 5.67271590e-01 4.42116082e-01
-1.82357326e-01 -1.75406747e-02 8.13762546e-01 6.10316545e-03
-7.72413552e-01 3.82561743e-01 -2.16471374e-01 1.85313761e-01
4.79333252e-01 -3.95925641e-02 -2.40842417e-01 -3.02610658e-02
-3.53243440e-01 2.47057885e-01 2.73199320e-01 3.81243527e-01
2.70745665e-01 -1.19676590e+00 -6.68386817e-01 4.56801742e-01
2.09582578e-02 2.35133111e-01 3.13600481e-01 6.94829941e-01
-4.20279592e-01 8.66272330e-01 -4.91046607e-02 -4.95228559e-01
-1.05744267e+00 4.51912731e-01 1.44906253e-01 -5.98476350e-01
-4.67300683e-01 1.17524636e+00 3.95175934e-01 -4.06591088e-01
3.70316535e-01 -5.63606381e-01 1.61420912e-01 -1.46317303e-01
6.52179003e-01 3.34205985e-01 2.43695319e-01 -1.88511074e-01
-9.86836702e-02 3.13960433e-01 -3.81250530e-01 -1.03147306e-01
1.03621578e+00 -2.59078294e-01 1.24409469e-02 1.37956873e-01
1.20633912e+00 -5.91662899e-02 -1.56256545e+00 -6.50060952e-01
-4.84684482e-02 -1.54278442e-01 3.53850335e-01 -9.18269634e-01
-9.59628522e-01 1.06514716e+00 4.53588903e-01 -1.92305624e-01
1.21614909e+00 -2.99421787e-01 1.27949154e+00 6.96138918e-01
3.06086928e-01 -1.09389520e+00 -1.00564681e-01 6.66789293e-01
5.97654998e-01 -1.01293492e+00 -1.82129160e-01 -1.13440387e-01
-5.44240475e-01 8.28965962e-01 4.50910568e-01 3.12677443e-01
2.39212781e-01 5.29756010e-01 -6.84598684e-02 1.92280278e-01
-1.19275856e+00 1.70957088e-01 2.23017126e-01 3.58973593e-01
3.59960735e-01 2.49933064e-01 7.15672299e-02 5.24297297e-01
-5.37565827e-01 1.36946216e-01 -1.97430030e-02 6.21678770e-01
-5.26865304e-01 -1.14988363e+00 -9.60663706e-02 5.37207544e-01
-4.21192884e-01 -5.11697531e-01 1.10977992e-01 6.07805669e-01
-1.38546592e-02 7.11625934e-01 2.03389540e-01 -1.58785582e-01
2.78333992e-01 1.81144461e-01 3.80859584e-01 -4.81519192e-01
-8.56205463e-01 4.06526402e-02 4.73149754e-02 -8.13169062e-01
3.77365261e-01 -2.47863680e-01 -1.27832139e+00 -5.02839267e-01
-1.57989711e-01 1.85627922e-01 8.25830162e-01 7.36177325e-01
8.28075528e-01 5.94147742e-01 1.55758962e-01 -7.49277949e-01
-1.18205643e+00 -8.69385958e-01 -2.03774884e-01 2.00510263e-01
3.57246429e-01 -3.92420232e-01 -1.98618931e-04 5.24496511e-02] | [8.74825382232666, 3.5931732654571533] |
6555ce4e-5b27-4a4b-9a67-ab78554bb3fd | grounded-situation-recognition | 2003.12058 | null | https://arxiv.org/abs/2003.12058v1 | https://arxiv.org/pdf/2003.12058v1.pdf | Grounded Situation Recognition | We introduce Grounded Situation Recognition (GSR), a task that requires producing structured semantic summaries of images describing: the primary activity, entities engaged in the activity with their roles (e.g. agent, tool), and bounding-box groundings of entities. GSR presents important technical challenges: identifying semantic saliency, categorizing and localizing a large and diverse set of entities, overcoming semantic sparsity, and disambiguating roles. Moreover, unlike in captioning, GSR is straightforward to evaluate. To study this new task we create the Situations With Groundings (SWiG) dataset which adds 278,336 bounding-box groundings to the 11,538 entity classes in the imsitu dataset. We propose a Joint Situation Localizer and find that jointly predicting situations and groundings with end-to-end training handily outperforms independent training on the entire grounding metric suite with relative gains between 8% and 32%. Finally, we show initial findings on three exciting future directions enabled by our models: conditional querying, visual chaining, and grounded semantic aware image retrieval. Code and data available at https://prior.allenai.org/projects/gsr. | ['Ali Farhadi', 'Aniruddha Kembhavi', 'Sarah Pratt', 'Mark Yatskar', 'Luca Weihs'] | 2020-03-26 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1987_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123490307.pdf | eccv-2020-8 | ['grounded-situation-recognition', 'situation-recognition'] | ['computer-vision', 'computer-vision'] | [ 5.20053148e-01 3.98708403e-01 -2.95744419e-01 -4.33845997e-01
-1.23224866e+00 -7.70868063e-01 7.31833041e-01 4.14504439e-01
-2.16175184e-01 5.67195952e-01 9.03778553e-01 1.67683568e-02
-1.68756574e-01 -4.04610544e-01 -8.89539480e-01 -2.09281474e-01
-1.78991988e-01 4.38594639e-01 2.35015586e-01 -1.17829688e-01
4.15767819e-01 4.51134257e-02 -1.57266283e+00 7.53736317e-01
8.23383987e-01 1.03562915e+00 4.66513664e-01 4.03436124e-01
-4.72688396e-03 1.47628438e+00 -7.78759599e-01 -3.46460432e-01
-4.57510464e-02 -2.37780288e-01 -1.01229846e+00 2.20822826e-01
1.02041984e+00 -3.72637540e-01 -5.69605291e-01 9.24908161e-01
4.34793204e-01 4.63307500e-01 5.47626197e-01 -1.69028258e+00
-1.13599908e+00 6.80581748e-01 -5.04118741e-01 3.97480816e-01
8.21536839e-01 -5.97471111e-02 1.39779902e+00 -1.14286840e+00
1.23136199e+00 1.32658672e+00 3.72102529e-01 4.88661468e-01
-8.46312344e-01 -5.50148427e-01 4.16285843e-01 3.61619681e-01
-1.48492610e+00 -5.79783320e-01 5.51506460e-01 -3.92371833e-01
1.10973835e+00 1.99385017e-01 3.34694505e-01 1.26213861e+00
-6.59821630e-01 1.10018528e+00 7.00379312e-01 -2.24298835e-01
2.85974771e-01 -7.78728053e-02 2.63219625e-01 8.42409074e-01
4.08548385e-01 -5.45502007e-01 -1.03967977e+00 -1.80608407e-01
6.06565356e-01 8.64687487e-02 -2.36374810e-01 -5.85608542e-01
-1.75465727e+00 4.69742775e-01 7.38258064e-01 2.19212040e-01
-5.57425976e-01 5.45574009e-01 1.87252700e-01 -2.35076204e-01
3.50493550e-01 8.00676048e-01 -2.81264186e-02 1.56185674e-02
-1.00148177e+00 5.35918355e-01 4.51635301e-01 1.49689984e+00
8.56776059e-01 -2.29179576e-01 -5.46585262e-01 6.30173147e-01
1.73156485e-01 6.58361375e-01 1.27478763e-01 -1.27242696e+00
9.95712996e-01 8.63947690e-01 4.40648198e-01 -1.21746016e+00
-3.25398117e-01 -2.65252143e-01 -2.85087168e-01 -5.20471752e-01
2.58006841e-01 -2.21171081e-02 -1.10333586e+00 1.75285089e+00
2.12171271e-01 3.48640621e-01 1.48085669e-01 1.04217505e+00
1.33158660e+00 7.76185691e-01 5.41339099e-01 3.75941932e-01
1.64062595e+00 -1.08370817e+00 -7.47274876e-01 -6.33852243e-01
6.92530751e-01 -3.22585225e-01 1.16626632e+00 -2.57975589e-02
-1.18085337e+00 -1.93882540e-01 -9.91580427e-01 -5.27076364e-01
-6.78362191e-01 2.25505933e-01 7.52523720e-01 -1.48010537e-01
-1.32010722e+00 1.28138557e-01 -4.47598636e-01 -6.08499467e-01
6.92998886e-01 -1.61593840e-01 -5.67914963e-01 -1.87020451e-01
-1.17666996e+00 7.92046010e-01 3.40401828e-01 -2.21574262e-01
-1.20138121e+00 -8.18845749e-01 -1.14664590e+00 1.52119964e-01
7.93197572e-01 -5.24735391e-01 1.26923251e+00 -6.33771181e-01
-2.44638503e-01 1.33026934e+00 -3.12296689e-01 -7.38230228e-01
6.24545105e-03 -3.93536389e-01 -6.03038132e-01 6.71460152e-01
8.67812157e-01 1.08205938e+00 5.24609208e-01 -1.20206463e+00
-7.85589337e-01 -3.68792892e-01 4.07561809e-01 6.12448514e-01
-2.09347859e-01 1.20734945e-01 -5.74738681e-01 -6.02307677e-01
2.02851102e-01 -7.56517947e-01 2.89345514e-02 -1.04168832e-01
-6.23008430e-01 -3.03307146e-01 7.29911625e-01 -8.78112853e-01
1.21053982e+00 -2.34002256e+00 1.22411057e-01 -7.09373038e-03
6.55873179e-01 -2.10701823e-01 -1.97244197e-01 5.77496946e-01
-1.80238545e-01 2.13838264e-01 -6.74973875e-02 -3.51655900e-01
1.95174202e-01 -1.53390467e-01 -5.10451078e-01 2.86324590e-01
1.66969404e-01 1.21318066e+00 -1.25293064e+00 -7.25583971e-01
3.81743945e-02 2.68829942e-01 -3.85038555e-01 8.71285144e-03
-1.74967453e-01 5.05026691e-02 -5.87685764e-01 9.82167184e-01
1.04037479e-01 -6.88672721e-01 1.32527933e-01 -4.22315687e-01
1.50008798e-01 4.87486601e-01 -1.13366556e+00 2.04924130e+00
-2.08982617e-01 8.62391174e-01 -1.22663252e-01 -7.45114505e-01
7.57937908e-01 2.30863586e-01 3.78011018e-01 -9.58550811e-01
-4.03235495e-01 4.45391312e-02 -6.26551330e-01 -4.03944433e-01
1.02450693e+00 2.40986988e-01 -6.50643647e-01 1.96078703e-01
4.33591902e-01 1.64207086e-01 2.61812717e-01 9.66768861e-01
1.43871772e+00 9.25394893e-02 4.45915937e-01 -8.92528892e-02
-1.33799136e-01 4.73563969e-01 3.13587099e-01 8.13879251e-01
-4.26670939e-01 8.23952317e-01 4.96262103e-01 -2.52060235e-01
-7.70576179e-01 -1.13129556e+00 3.28340322e-01 1.32318568e+00
7.18953729e-01 -6.05471253e-01 -6.05928421e-01 -7.96011388e-01
1.30121693e-01 9.77362394e-01 -7.89717674e-01 -1.69900462e-01
-5.60612321e-01 1.04165671e-03 5.71973622e-01 7.99992085e-01
6.62764430e-01 -1.21361351e+00 -6.15855575e-01 -1.22962765e-01
-7.29942501e-01 -1.41253507e+00 -5.68149447e-01 -6.84179217e-02
-3.85448664e-01 -1.23026860e+00 -7.73516417e-01 -8.99740160e-01
6.02548540e-01 7.39110529e-01 1.71652305e+00 -1.02346286e-01
-2.89151013e-01 8.31200778e-01 -4.86652970e-01 -3.80083948e-01
2.24974334e-01 -7.52142891e-02 -3.70823801e-01 -1.89889207e-01
3.63224387e-01 -1.87412351e-01 -8.11552227e-01 3.12900603e-01
-5.93681037e-01 2.98524916e-01 4.06039417e-01 2.72454739e-01
7.45481730e-01 -3.01385671e-01 5.98707378e-01 -7.97812223e-01
4.42141145e-01 -8.71443391e-01 -2.05053315e-01 5.71357489e-01
-3.31274182e-01 -1.55772299e-01 -1.90718591e-01 -1.72194257e-01
-1.00549579e+00 -2.04322696e-01 4.75567430e-01 -6.12781107e-01
-3.23556632e-01 3.08479160e-01 -1.90983251e-01 5.06510556e-01
8.43053818e-01 8.22501257e-02 -4.97148871e-01 -1.81904554e-01
8.53819132e-01 4.86915827e-01 9.93569672e-01 -6.71502709e-01
6.18524849e-01 6.24408245e-01 -2.87465215e-01 -5.45972884e-01
-1.16463375e+00 -1.09078741e+00 2.56073643e-02 -3.45502824e-01
1.17943597e+00 -1.62017727e+00 -2.78821975e-01 1.80660393e-02
-1.09334862e+00 -3.98178279e-01 -5.40585995e-01 5.20524383e-02
-6.48179889e-01 5.39905019e-02 -2.08002836e-01 -5.69144487e-01
-2.76941389e-01 -6.79921031e-01 1.79833615e+00 4.03389893e-02
-4.82081652e-01 -6.62150621e-01 -2.88857490e-01 7.79413760e-01
1.32259101e-01 5.60054183e-01 4.89806056e-01 -7.54101157e-01
-8.04809928e-01 -1.10366851e-01 -5.55530190e-01 -1.76749960e-01
-3.04435007e-02 -4.85594571e-01 -5.69692969e-01 -9.92481411e-02
-5.95547795e-01 -5.31728864e-01 8.22336257e-01 2.49824747e-01
1.09726071e+00 -3.59659165e-01 -6.81971908e-01 2.87503958e-01
1.28795445e+00 1.63415164e-01 7.13765860e-01 4.05746460e-01
9.13215578e-01 6.61858797e-01 1.07600439e+00 3.55341375e-01
7.38689423e-01 6.11519814e-01 6.45513475e-01 -1.95625275e-01
-4.89235520e-01 -6.32834435e-01 1.87987283e-01 -2.52742231e-01
1.77439172e-02 -5.49352169e-01 -1.15452731e+00 1.08207619e+00
-2.13009310e+00 -1.38815558e+00 9.08225551e-02 1.76261854e+00
5.61128139e-01 -2.85978466e-01 1.73504651e-01 -3.34781051e-01
1.08847058e+00 5.48962355e-01 -5.16880751e-01 2.70167500e-01
-2.79810190e-01 -1.68678358e-01 4.54194009e-01 1.61162332e-01
-1.18213832e+00 1.20175827e+00 5.69681358e+00 6.44147992e-01
-5.62489688e-01 3.67578030e-01 5.18517911e-01 -3.92787069e-01
-4.07880217e-01 -2.56437287e-02 -8.43404114e-01 2.82942563e-01
6.13849580e-01 -1.34927198e-01 5.03821611e-01 1.08913791e+00
5.40624596e-02 -2.26419773e-02 -1.12637413e+00 1.34134686e+00
4.81330961e-01 -1.60669935e+00 2.22882226e-01 -1.87549293e-01
9.13790345e-01 6.20029904e-02 4.39897515e-02 3.35262209e-01
3.47230762e-01 -9.67310846e-01 1.33722556e+00 3.28699380e-01
9.02805209e-01 -3.58323723e-01 3.01318526e-01 -1.32050171e-01
-1.26484644e+00 -1.80513248e-01 -4.53494452e-02 2.02919364e-01
2.02852935e-01 2.34757766e-01 -9.31964159e-01 3.35069686e-01
8.46625984e-01 1.00003326e+00 -7.51961291e-01 7.70970047e-01
-3.09640139e-01 4.99188960e-01 -6.85318187e-02 -3.69737633e-02
4.53949004e-01 2.48637587e-01 6.71566904e-01 1.17874181e+00
3.05824637e-01 1.27224758e-01 -1.93993933e-02 9.64398026e-01
-4.35639620e-01 -4.32039090e-02 -8.95567894e-01 -3.31889242e-01
8.34217131e-01 1.21007061e+00 -7.99480915e-01 -6.18362904e-01
-1.93110302e-01 1.04383099e+00 4.29876179e-01 6.71146750e-01
-1.14306998e+00 -4.47012037e-01 5.99317193e-01 4.99086738e-01
2.86894768e-01 -1.06403768e-01 -1.15803443e-02 -1.28047574e+00
1.37796089e-01 -7.88850665e-01 6.10870004e-01 -1.57452595e+00
-9.05640781e-01 2.73420900e-01 3.80405635e-01 -1.04068506e+00
-2.17380568e-01 -3.90363187e-01 -3.13540190e-01 4.69424665e-01
-1.27619946e+00 -1.33668435e+00 -7.76072860e-01 7.50424087e-01
6.22178614e-01 -2.97277980e-02 4.88452554e-01 4.13060933e-01
-2.75494874e-01 2.70354986e-01 -5.47214627e-01 2.69146293e-01
7.63682246e-01 -1.47735548e+00 5.17697871e-01 8.65990043e-01
5.19378483e-01 7.07251966e-01 6.17359459e-01 -7.98385203e-01
-1.18482327e+00 -1.44983649e+00 1.02702808e+00 -8.67386937e-01
6.40057623e-01 -5.95990837e-01 -5.71024001e-01 1.13139415e+00
2.71566480e-01 5.36995977e-02 5.23799539e-01 3.21425676e-01
-7.25743055e-01 2.42332980e-01 -1.19332635e+00 8.03649008e-01
1.89675784e+00 -8.41710389e-01 -8.59589934e-01 6.24713838e-01
9.79883969e-01 -5.43708146e-01 -4.77164119e-01 1.00917704e-01
1.28404006e-01 -6.15404069e-01 1.30398190e+00 -6.57622814e-01
5.60144722e-01 -3.98612350e-01 -6.42115831e-01 -9.95359540e-01
-4.10730392e-01 -5.24401486e-01 -4.06095535e-01 1.32158136e+00
5.15570343e-01 -3.09208602e-01 7.82000959e-01 6.65435135e-01
-4.27893102e-01 -4.80263799e-01 -5.78151703e-01 -6.38083816e-01
-5.79941213e-01 -4.06344652e-01 7.13307440e-01 1.01811647e+00
-1.98428519e-02 4.06698793e-01 -1.15816474e-01 4.37449843e-01
6.40281975e-01 2.57122755e-01 4.87676263e-01 -8.28504324e-01
-6.45880029e-02 -7.26779774e-02 -5.68028748e-01 -9.30680394e-01
2.92675853e-01 -1.01817942e+00 6.62005469e-02 -2.25763226e+00
4.25988734e-01 -2.91499317e-01 -3.97125423e-01 9.11519051e-01
-1.57111019e-01 2.62102634e-01 2.62346655e-01 2.17048988e-01
-1.78810191e+00 1.97093338e-01 9.63666916e-01 -3.76899868e-01
1.95469595e-02 -7.02925801e-01 -1.18024015e+00 5.38073003e-01
6.64221406e-01 -1.54173791e-01 -6.98533714e-01 -6.61188900e-01
3.22185606e-01 -1.67585239e-02 9.24076796e-01 -1.11727417e+00
1.56339914e-01 -4.75559682e-02 1.85296059e-01 -6.76343501e-01
5.03279150e-01 -4.96301770e-01 2.66837366e-02 -2.51698136e-01
-7.30927765e-01 9.74957421e-02 5.13787568e-02 7.33858407e-01
-3.70373636e-01 -2.17764415e-02 9.35702845e-02 -1.96389198e-01
-1.57758045e+00 8.07278231e-02 -9.16817486e-02 6.36511028e-01
1.09488332e+00 -1.85744897e-01 -7.30096698e-01 -5.80195189e-01
-7.99460173e-01 3.69945049e-01 4.21193123e-01 5.86979628e-01
6.90397263e-01 -1.31338286e+00 -4.97627169e-01 -5.05908966e-01
7.55468786e-01 1.54096365e-01 4.51087862e-01 4.02196288e-01
-2.86651522e-01 4.80137318e-01 1.34184465e-01 -4.34006423e-01
-1.20567787e+00 5.71861088e-01 7.74647892e-02 -5.67311496e-02
-4.24011230e-01 9.65783060e-01 5.53719699e-01 -5.83965108e-02
4.11571383e-01 -4.18818176e-01 -8.77779871e-02 2.65433162e-01
5.70605159e-01 4.57391232e-01 -1.17071792e-02 -7.98920929e-01
-6.65862739e-01 3.01901847e-01 6.06858656e-02 -1.04787499e-01
1.15178657e+00 -1.48316383e-01 8.24674517e-02 1.77416131e-01
1.19346321e+00 -3.01332653e-01 -1.10112441e+00 -2.05206588e-01
1.17421128e-01 -2.02013537e-01 3.07083409e-02 -1.05254614e+00
-7.10094392e-01 6.12819731e-01 2.69695550e-01 1.20816901e-01
1.00859189e+00 7.67696500e-01 5.54815054e-01 5.27757764e-01
7.08327651e-01 -9.60142016e-01 4.58831102e-01 9.89683345e-02
1.41534638e+00 -1.23222363e+00 -1.61869884e-01 -5.93961418e-01
-1.03198218e+00 2.79143721e-01 6.98398471e-01 -1.55725423e-02
1.41857699e-01 -1.25339702e-01 -2.83387840e-01 -7.79600263e-01
-7.16017008e-01 -6.48895860e-01 3.50519836e-01 7.12269902e-01
1.73689872e-01 1.51776746e-01 2.20568240e-01 5.60434282e-01
6.75431453e-04 -2.41164669e-01 1.93536922e-01 9.80870366e-01
-3.92653465e-01 -2.32824564e-01 -1.32090315e-01 3.82462561e-01
-3.98746043e-01 -3.38990390e-01 -4.99179631e-01 7.98652112e-01
1.33791000e-01 1.11374629e+00 2.57183999e-01 -6.67379722e-02
5.98995566e-01 -3.56788351e-03 2.73304224e-01 -9.38354731e-01
-2.72697359e-01 -4.53550994e-01 6.09319150e-01 -1.03116632e+00
-5.44260740e-01 -8.05836141e-01 -1.62667477e+00 2.21959442e-01
-2.77684499e-02 9.41081271e-02 5.11945784e-01 6.29389524e-01
8.83808315e-01 4.94075537e-01 8.82967338e-02 -7.32561290e-01
-1.37691170e-01 -7.77113914e-01 -4.67603087e-01 8.91057432e-01
2.19745845e-01 -8.33350778e-01 -3.17943603e-01 2.75978565e-01] | [10.415332794189453, 1.2995322942733765] |
7fe2aad9-65d1-4eaf-8d10-e6a3de92d19e | black-lscdiscovery-shared-task-ualberta-at | null | null | https://aclanthology.org/2022.lchange-1.19 | https://aclanthology.org/2022.lchange-1.19.pdf | black[LSCDiscovery shared task] UAlberta at LSCDiscovery: Lexical Semantic Change Detection via Word Sense Disambiguation | We describe our two systems for the shared task on Lexical Semantic Change Discovery in Spanish. For binary change detection, we frame the task as a word sense disambiguation (WSD) problem. We derive sense frequency distributions for target words in both old and modern corpora. We assume that the word semantics have changed if a sense is observed in only one of the two corpora, or the relative change for any sense exceeds a tuned threshold. For graded change discovery, we follow the design of CIRCE (Pömsl and Lyapin, 2020) by combining both static and contextual embeddings. For contextual embeddings, we use XLM-RoBERTa instead of BERT, and train the model to predict a masked token instead of the time period. Our language-independent methods achieve results that are close to the best-performing systems in the shared task. | ['Grzegorz Kondrak', 'Spencer von der Ohe', 'Daniela Teodorescu'] | null | null | null | null | lchange-acl-2022-5 | ['word-sense-disambiguation'] | ['natural-language-processing'] | [ 1.50423735e-01 -2.03571290e-01 -4.31396216e-01 -2.47771412e-01
-7.23035991e-01 -8.65602732e-01 9.99095857e-01 7.44218469e-01
-1.12887073e+00 8.80445242e-01 5.05495310e-01 -3.45190644e-01
3.10220450e-01 -7.66352296e-01 -3.18007618e-01 -5.42206585e-01
8.23414326e-02 3.41636568e-01 4.61614549e-01 -5.49341977e-01
2.37134680e-01 4.69502211e-02 -1.45945132e+00 2.40810573e-01
5.28682232e-01 4.94185895e-01 3.89712423e-01 5.31736195e-01
-3.85154575e-01 3.34658355e-01 -5.77797592e-01 -2.50214428e-01
9.17092664e-04 -2.88812608e-01 -1.08091033e+00 -5.38430095e-01
3.53279889e-01 4.44170058e-01 -2.33661428e-01 1.19199598e+00
7.01591551e-01 2.62409747e-01 6.44425690e-01 -9.42571998e-01
-9.92549241e-01 9.48888600e-01 -4.82399166e-01 9.74694908e-01
5.66319823e-01 -1.37826115e-01 1.44953060e+00 -1.01302934e+00
1.14596057e+00 1.34059548e+00 7.36117482e-01 4.45349395e-01
-1.37456834e+00 -4.65218931e-01 5.54565609e-01 7.62378097e-01
-1.17711473e+00 -5.54165602e-01 5.45409501e-01 -4.18411344e-01
1.43290281e+00 1.31649733e-01 7.34775662e-01 1.45033264e+00
3.67213070e-01 5.56613445e-01 1.04307115e+00 -7.70867407e-01
4.78382915e-01 -2.32450143e-01 3.62170398e-01 2.38759294e-01
4.47523415e-01 -1.97353706e-01 -7.02346504e-01 -3.66820067e-01
-7.86512867e-02 -2.00167015e-01 -1.56893268e-01 8.53637457e-02
-1.42828882e+00 7.44907260e-01 3.65410894e-02 9.63733375e-01
-4.46841866e-01 1.95890501e-01 6.23224556e-01 6.74655378e-01
8.30445170e-01 7.88861096e-01 -1.26591527e+00 -4.01787341e-01
-8.91581118e-01 4.76834863e-01 5.32287061e-01 6.07757092e-01
7.43668497e-01 -1.99633405e-01 -1.97748616e-01 9.18300569e-01
9.74583477e-02 4.01024491e-01 1.28384185e+00 -5.35755277e-01
-4.71834559e-03 4.32031840e-01 2.97152340e-01 -7.25499392e-01
-4.94492382e-01 -3.55341872e-05 -2.00433433e-01 -1.36606589e-01
2.19465703e-01 4.50471155e-02 -1.08181381e+00 2.00765085e+00
2.66305745e-01 2.22739771e-01 4.55665588e-02 3.16568017e-01
5.38332880e-01 5.30324340e-01 4.41896409e-01 -4.63820249e-01
1.56460667e+00 -3.83691728e-01 -9.27380562e-01 -5.01487315e-01
7.10020065e-01 -7.13477612e-01 1.39630461e+00 4.90224063e-02
-7.72277117e-01 -3.06677550e-01 -1.17215419e+00 -2.22289637e-01
-9.18338180e-01 -3.40987116e-01 3.41387957e-01 8.87382030e-02
-1.10377848e+00 5.43230236e-01 -8.08089912e-01 -9.60245311e-01
-2.74679922e-02 -2.49936253e-01 -1.77733213e-01 1.53263286e-01
-1.94261229e+00 1.11742425e+00 6.07093751e-01 -6.24638021e-01
-4.85156506e-01 -9.00467813e-01 -8.57614338e-01 -2.49742717e-01
3.13146472e-01 -6.20869219e-01 1.33572185e+00 -9.54988360e-01
-1.08821154e+00 1.55990386e+00 -4.17115688e-01 -5.19995749e-01
3.02964374e-02 -1.42797947e-01 -9.62482691e-01 -3.44436496e-01
6.85626268e-01 4.28718805e-01 7.06799448e-01 -8.31640601e-01
-9.02418077e-01 -3.34733188e-01 -1.23337299e-01 2.04708368e-01
-2.00543389e-01 2.76548266e-01 -1.21861078e-01 -8.86373818e-01
-6.10731505e-02 -7.23445952e-01 -2.64426470e-01 -4.49732602e-01
-1.36057898e-01 -6.77681386e-01 4.84378576e-01 -6.92639947e-01
1.68072295e+00 -2.33498454e+00 -1.28657399e-02 -2.93994099e-01
6.36454448e-02 1.14231803e-01 -3.26898068e-01 3.57370734e-01
-4.94924396e-01 4.10226345e-01 -4.14835751e-01 -2.02174827e-01
9.83629674e-02 3.89425844e-01 -5.43101609e-01 3.46028686e-01
7.07435384e-02 8.16460729e-01 -1.43107259e+00 -1.92372710e-01
-2.17965603e-01 -2.31489405e-01 -4.77346450e-01 -1.89634666e-01
-3.65724087e-01 -1.77240193e-01 -2.27932602e-01 4.71490830e-01
2.62593031e-01 1.24274850e-01 5.77188969e-01 -2.00812854e-02
-4.23790038e-01 9.73173141e-01 -1.12762177e+00 1.89019215e+00
-5.62520623e-01 6.61482215e-01 -4.81405467e-01 -6.88000083e-01
6.33536279e-01 3.28302741e-01 4.73986357e-01 -7.30446160e-01
-3.48091155e-01 2.35057294e-01 -3.08803674e-02 -4.53297079e-01
8.17830324e-01 -3.26955229e-01 -6.03268921e-01 3.73587966e-01
1.18682437e-01 -2.35193431e-01 2.87691146e-01 -2.69631390e-02
1.55702698e+00 -5.59684541e-03 1.07592535e+00 -6.11855209e-01
1.76685259e-01 2.03408167e-01 1.00980198e+00 5.86043298e-01
-4.21718329e-01 1.79771051e-01 5.15094578e-01 -3.21503103e-01
-9.31888163e-01 -1.45941043e+00 -2.75687009e-01 1.59543800e+00
1.49104536e-01 -6.86880767e-01 -1.19619831e-01 -7.21569538e-01
3.24687272e-01 1.21492136e+00 -8.25533509e-01 -1.56669840e-01
-6.14649117e-01 -7.96121597e-01 5.78694522e-01 3.37273896e-01
4.19269390e-02 -1.23901570e+00 -4.00054127e-01 6.10504448e-01
-1.74109086e-01 -7.26464033e-01 -5.56768060e-01 4.53055620e-01
-4.31094885e-01 -1.08596313e+00 -2.66223941e-02 -9.97926772e-01
-1.24431811e-01 -7.19882324e-02 1.35263443e+00 -3.55253369e-01
-4.43004936e-01 2.73224145e-01 -5.42467356e-01 -4.20054793e-01
-3.55013281e-01 2.38667965e-01 5.16699791e-01 -5.17695904e-01
8.92268479e-01 -4.91164923e-01 -1.43974006e-01 -3.57053727e-01
-9.29092884e-01 -6.95740223e-01 -6.80378601e-02 8.39795649e-01
6.75748050e-01 -7.85401762e-02 8.73291731e-01 -9.11259830e-01
8.32435310e-01 -8.50740850e-01 -2.13292867e-01 2.13975370e-01
-1.14885616e+00 1.34974971e-01 1.57938495e-01 -6.06000125e-01
-9.88036752e-01 -2.72389084e-01 -1.69500753e-01 9.03751254e-02
-1.92795739e-01 6.53190494e-01 -1.57573164e-01 9.05468762e-01
8.02309036e-01 -9.79315341e-02 -6.82017088e-01 -6.63838387e-01
9.54545259e-01 6.56182587e-01 6.36126339e-01 -5.20075738e-01
4.64402407e-01 3.33139509e-01 -8.64003778e-01 -8.09238315e-01
-8.89763534e-01 -6.74108267e-01 -8.13737929e-01 2.98873276e-01
1.12815225e+00 -9.83892620e-01 3.80791388e-02 4.41857994e-01
-1.34087789e+00 -3.14502865e-01 -6.88662827e-01 2.87091196e-01
-3.97530913e-01 1.35405660e-01 -3.44657212e-01 -2.74462223e-01
-1.17885962e-01 -6.44612610e-01 9.34172153e-01 -2.38196310e-02
-9.19316113e-01 -1.41474771e+00 7.88666844e-01 -5.03796279e-01
3.10405135e-01 3.17062229e-01 1.31161666e+00 -1.10103941e+00
2.06437364e-01 2.36029372e-01 2.95293421e-01 -1.32864555e-02
7.77368903e-01 -1.64594308e-01 -8.83736551e-01 -2.88132697e-01
-6.98931515e-03 -1.14835635e-01 1.21262550e+00 2.67563432e-01
7.98711598e-01 -2.54195809e-01 -5.86095810e-01 2.96801984e-01
1.36631095e+00 3.30110073e-01 2.08872765e-01 7.88883626e-01
3.63078445e-01 1.01550169e-01 4.38940614e-01 4.82109219e-01
5.86909592e-01 5.16645491e-01 -1.70495361e-01 2.60233790e-01
-2.92685539e-01 -3.13776523e-01 7.45976150e-01 1.03029609e+00
5.55366695e-01 -3.11306685e-01 -9.78456020e-01 1.40044808e+00
-1.63282669e+00 -1.15195155e+00 -1.02694638e-01 1.94787776e+00
1.61889994e+00 1.85638398e-01 -1.51133955e-01 -3.18452902e-02
7.48156548e-01 6.90823495e-01 -6.54016316e-01 -6.62176311e-01
-5.24052024e-01 5.93258679e-01 4.71745133e-01 7.61033654e-01
-1.11531889e+00 1.49877608e+00 7.04898834e+00 6.64974868e-01
-8.39277923e-01 6.60262585e-01 2.13446319e-02 -2.32778758e-01
-7.42124677e-01 1.06592737e-01 -6.74085617e-01 4.91713434e-01
9.76830900e-01 -6.36886954e-01 3.31427217e-01 5.39824247e-01
1.98649988e-02 -1.90919578e-01 -1.22923839e+00 8.84029806e-01
2.27704704e-01 -1.00604749e+00 -1.48852736e-01 -3.68355513e-01
8.82028043e-01 3.82639050e-01 -1.73863038e-01 4.67841238e-01
7.63546944e-01 -5.79313636e-01 8.53288531e-01 6.10882103e-01
8.85498762e-01 -3.02538425e-01 5.46054423e-01 -1.98146015e-01
-1.14470541e+00 2.26328328e-01 -2.37319902e-01 -2.35727146e-01
1.49766400e-01 1.05547440e+00 -7.94382632e-01 3.42158377e-01
8.38735759e-01 1.18920720e+00 -7.08757579e-01 5.69970310e-01
-6.16602302e-01 9.60838139e-01 -2.68762738e-01 -5.08465581e-02
-1.62029415e-01 2.13112533e-01 9.88808870e-01 1.42475033e+00
1.78780079e-01 -1.65848240e-01 2.84503609e-01 6.76896513e-01
-2.37060294e-01 1.94202811e-01 -5.29156208e-01 -1.69871628e-01
1.00298870e+00 8.23764384e-01 -4.08573478e-01 -5.67190707e-01
-3.48396152e-01 1.28665543e+00 3.11589897e-01 3.41941416e-01
-4.07208443e-01 -5.49250603e-01 1.52303827e+00 -1.20463856e-01
4.54706430e-01 -1.04607917e-01 -4.07730371e-01 -1.28788412e+00
-8.40748101e-02 -5.07676780e-01 8.24297845e-01 -5.54299831e-01
-1.73752010e+00 3.31502676e-01 -6.42123446e-03 -7.04164386e-01
-3.89191568e-01 -5.91775060e-01 -5.96671522e-01 8.03475618e-01
-1.58137417e+00 -6.16083562e-01 4.05506819e-01 3.34461421e-01
6.83470666e-01 -9.73630697e-02 9.40245569e-01 2.11290985e-01
-3.07143122e-01 4.44902122e-01 5.41515015e-02 -1.22118041e-01
1.33673978e+00 -1.68542695e+00 8.26548934e-01 8.94955158e-01
2.49582499e-01 5.71071744e-01 9.11708355e-01 -8.79782975e-01
-7.21382678e-01 -1.17881310e+00 1.66999149e+00 -7.34311700e-01
1.32320881e+00 -2.33031392e-01 -1.02208281e+00 9.19811845e-01
4.08524424e-01 -2.93503940e-01 6.91659391e-01 5.66623151e-01
-6.10689640e-01 7.66960010e-02 -1.00970459e+00 6.94807768e-01
1.38648820e+00 -8.27066958e-01 -1.64089859e+00 2.49542058e-01
1.29695356e+00 -5.65765724e-02 -7.62600362e-01 2.75594369e-02
1.18748553e-01 -2.65034139e-01 6.33756816e-01 -1.09810722e+00
4.86798864e-03 -4.78436828e-01 -6.14096224e-01 -1.98412192e+00
-5.93474567e-01 -3.54372710e-01 2.02054843e-01 1.39396310e+00
6.26469433e-01 -7.79485881e-01 -1.22757763e-01 -7.82222822e-02
-1.08215399e-01 -1.71660304e-01 -1.26959503e+00 -7.48420894e-01
4.39767718e-01 -6.84554577e-01 8.56928766e-01 1.60060370e+00
2.12243870e-01 5.21127343e-01 3.34060848e-01 1.93070918e-01
-3.15244906e-02 8.36866498e-02 -2.99629942e-02 -1.27531075e+00
1.58867929e-02 -6.35883570e-01 -4.82538342e-01 -5.39517641e-01
6.37509942e-01 -1.25897920e+00 2.08812714e-01 -1.54030526e+00
1.36719018e-01 -1.12934574e-01 -8.57889235e-01 9.44252491e-01
-7.17943430e-01 -7.81001449e-02 -1.54880270e-01 -1.49742293e-04
-4.24885750e-01 3.75647813e-01 5.23631632e-01 -3.51863563e-01
-4.36178088e-01 -6.23205185e-01 -7.48008311e-01 7.32503533e-01
7.09866047e-01 -7.09693253e-01 -8.56711194e-02 -3.43427926e-01
6.87623441e-01 -6.20541096e-01 1.18276887e-01 -5.07270634e-01
-6.41130283e-02 -4.78477120e-01 4.24144557e-03 -2.98689038e-01
-1.96235836e-01 -3.44513357e-01 -1.90488681e-01 4.49173242e-01
-5.98529458e-01 6.14219010e-01 2.62097865e-01 5.28097570e-01
-5.16407527e-02 -1.77021325e-01 6.86738849e-01 8.61908495e-03
-1.30623436e+00 -1.74956378e-02 -6.90456331e-01 5.64927518e-01
8.50104630e-01 2.99243331e-01 -3.18203062e-01 1.53885141e-01
-1.03175628e+00 1.08015373e-01 7.14281857e-01 1.00133860e+00
2.26576373e-01 -1.58353162e+00 -7.56973445e-01 -3.57771106e-02
6.18736565e-01 -5.84075630e-01 -1.66930914e-01 3.29558194e-01
1.14162892e-01 3.12718935e-03 1.21890306e-01 -5.93648031e-02
-8.98554385e-01 7.48754323e-01 2.92949826e-01 -1.84894383e-01
-3.72642398e-01 6.98858619e-01 -1.73286438e-01 -7.47460425e-01
-3.79670531e-01 -7.38444805e-01 -1.60306051e-01 7.93168545e-01
5.02875090e-01 3.98425341e-01 2.92366087e-01 -3.47471535e-01
-8.85242879e-01 2.04156041e-01 -2.83662379e-02 -3.57667595e-01
1.32627797e+00 -3.26871186e-01 -3.97911251e-01 1.29577184e+00
1.13964558e+00 3.26247066e-01 -4.93571341e-01 -5.90661585e-01
5.98164916e-01 -8.68815258e-02 8.79107192e-02 -1.06078064e+00
-3.03498238e-01 1.74864739e-01 7.45356739e-01 2.60191739e-01
8.44274879e-01 4.04928684e-01 8.32244873e-01 2.61907041e-01
2.33508706e-01 -1.42644584e+00 -2.82265842e-01 9.32162285e-01
7.93096542e-01 -1.01955295e+00 -1.83909267e-01 2.56106168e-01
-4.98498857e-01 7.65791595e-01 3.70021731e-01 -9.59568471e-02
8.02586973e-01 2.11541116e-01 2.67822683e-01 -1.92341655e-01
-1.18014312e+00 -5.36829948e-01 8.82608443e-02 5.41826427e-01
4.80451375e-01 5.12572527e-01 -9.31698143e-01 6.91552103e-01
-4.51419204e-01 -2.75282115e-01 5.65241456e-01 8.61925840e-01
-6.72179818e-01 -1.15659106e+00 -1.52701572e-01 5.40561557e-01
-4.07777041e-01 -5.55257440e-01 -4.92333263e-01 7.16850817e-01
4.95564997e-01 9.19587910e-01 4.11224127e-01 -2.93626636e-01
6.12050653e-01 7.84679472e-01 3.10941219e-01 -1.01320195e+00
-3.35132688e-01 -2.39498809e-01 2.35457703e-01 -6.27523959e-01
-3.56776774e-01 -1.25864971e+00 -1.42888057e+00 1.26097381e-01
2.24440679e-01 -5.88980829e-03 4.16095972e-01 1.13498974e+00
3.31960350e-01 5.57707191e-01 4.65429515e-01 -1.45916879e-01
-4.76297945e-01 -1.30301809e+00 -7.12486088e-01 6.86084867e-01
4.41652089e-01 -7.40023971e-01 -3.75373900e-01 3.90241563e-01] | [10.28369140625, 8.998906135559082] |
317bb9ee-a844-43f0-9113-3e4ff84d35bc | speech-synthesis-with-mixed-emotions | 2208.05890 | null | https://arxiv.org/abs/2208.05890v3 | https://arxiv.org/pdf/2208.05890v3.pdf | Speech Synthesis with Mixed Emotions | Emotional speech synthesis aims to synthesize human voices with various emotional effects. The current studies are mostly focused on imitating an averaged style belonging to a specific emotion type. In this paper, we seek to generate speech with a mixture of emotions at run-time. We propose a novel formulation that measures the relative difference between the speech samples of different emotions. We then incorporate our formulation into a sequence-to-sequence emotional text-to-speech framework. During the training, the framework does not only explicitly characterize emotion styles, but also explores the ordinal nature of emotions by quantifying the differences with other emotions. At run-time, we control the model to produce the desired emotion mixture by manually defining an emotion attribute vector. The objective and subjective evaluations have validated the effectiveness of the proposed framework. To our best knowledge, this research is the first study on modelling, synthesizing, and evaluating mixed emotions in speech. | ['Haizhou Li', 'B. W. Schuller', 'Rajib Rana', 'Berrak Sisman', 'Kun Zhou'] | 2022-08-11 | null | null | null | null | ['emotional-speech-synthesis'] | ['speech'] | [ 2.51296580e-01 -2.22945940e-02 7.03091770e-02 -4.07986224e-01
-5.44501066e-01 -5.52176654e-01 6.70723021e-01 -3.44018608e-01
-3.39475200e-02 6.10512495e-01 4.29766864e-01 6.30297810e-02
1.96283340e-01 -4.33297455e-01 -2.61253744e-01 -6.16131723e-01
3.33820313e-01 2.99855918e-01 -4.42137122e-01 -2.56961912e-01
-8.26788098e-02 5.03516734e-01 -1.77052891e+00 4.10211295e-01
9.80522096e-01 1.00078487e+00 1.16753943e-01 1.10773122e+00
-3.71373445e-01 9.74038422e-01 -1.13035882e+00 -4.39630657e-01
-1.03621893e-01 -9.82010901e-01 -4.70860839e-01 4.56837147e-01
-9.59858000e-02 7.15497956e-02 3.19246173e-01 1.13228369e+00
7.72361696e-01 2.97616422e-01 8.44829917e-01 -1.40422261e+00
-4.41567391e-01 7.12194443e-01 3.02115325e-02 -2.71390796e-01
7.01188862e-01 9.32852998e-02 8.26034546e-01 -7.35044956e-01
5.15320599e-01 1.23269689e+00 3.41618806e-01 8.01734447e-01
-1.07290232e+00 -6.68849051e-01 1.31433383e-01 -1.08958088e-01
-1.20764208e+00 -8.33382130e-01 1.26502979e+00 -4.81574476e-01
7.40674555e-01 6.95812643e-01 6.10274613e-01 1.43514562e+00
-3.44050974e-01 8.05392742e-01 1.28779423e+00 -6.18252575e-01
2.75176048e-01 5.03282845e-01 -8.96316245e-02 3.26945841e-01
-6.51689708e-01 1.36349976e-01 -3.44562709e-01 -6.48361742e-02
2.76711911e-01 -5.34008086e-01 -3.24847579e-01 2.96241999e-01
-1.00663733e+00 6.65163636e-01 -4.07529950e-01 5.70722818e-01
-7.41644800e-01 -1.74293637e-01 5.03356993e-01 3.25780839e-01
5.80363154e-01 4.84609753e-01 -1.98939264e-01 -5.84780157e-01
-1.09496832e+00 5.65990573e-03 9.98018384e-01 9.50993061e-01
4.04234827e-01 5.90100527e-01 -2.84407377e-01 1.23710513e+00
1.62260562e-01 4.25962418e-01 6.76006079e-01 -1.01260066e+00
-1.17035918e-02 2.88385421e-01 2.04592481e-01 -9.45780993e-01
-1.87855348e-01 -3.38571966e-01 -7.35816061e-01 1.24541856e-01
-1.30484283e-01 -6.16473377e-01 -6.61973953e-01 1.93707001e+00
2.84248024e-01 1.40643552e-01 5.92763722e-01 8.56642663e-01
9.46072400e-01 1.13635969e+00 1.39695466e-01 -7.92712033e-01
1.18775237e+00 -9.90115821e-01 -1.51983356e+00 -5.01903184e-02
2.85663575e-01 -1.11620271e+00 1.35464382e+00 5.11778116e-01
-1.14478910e+00 -7.40786135e-01 -8.89422417e-01 3.93211335e-01
-1.93937555e-01 5.71619689e-01 2.35309869e-01 1.07956994e+00
-1.00150478e+00 2.12999284e-01 -2.83524662e-01 -4.65686172e-02
-3.39289844e-01 -3.78601067e-02 2.14386778e-03 6.57702744e-01
-1.49127829e+00 8.35912764e-01 1.00415595e-01 -9.19538960e-02
-5.78888178e-01 -3.84796888e-01 -7.54226744e-01 -1.79597765e-01
2.92904973e-01 -4.31104064e-01 1.45458448e+00 -1.82794929e+00
-2.29922223e+00 4.69254702e-01 -4.13556457e-01 -3.53820957e-02
4.18904305e-01 4.26007025e-02 -9.45236027e-01 6.53387234e-02
-2.29883134e-01 6.66514993e-01 9.12393093e-01 -1.72754204e+00
-4.14458841e-01 1.55626595e-01 -1.62336811e-01 3.41945440e-01
-3.97411376e-01 5.89109123e-01 -3.33888799e-01 -9.71720576e-01
-3.84548992e-01 -9.15888965e-01 1.30225765e-02 -5.28133690e-01
-4.28486615e-01 -1.50144994e-01 6.40546679e-01 -5.30777216e-01
1.44378912e+00 -2.23566723e+00 2.98346162e-01 8.66063498e-03
-1.62277251e-01 1.72787696e-01 -1.34538993e-01 3.90515834e-01
-1.86387330e-01 3.18439901e-01 -1.61107466e-01 -6.64561510e-01
2.60408729e-01 2.22765625e-01 -3.73042822e-01 -1.04178451e-02
3.29752862e-01 6.17123842e-01 -7.76241779e-01 -7.68359661e-01
2.04187289e-01 7.12189317e-01 -4.23795164e-01 6.71141505e-01
-3.46157968e-01 4.34954792e-01 -3.46561432e-01 4.19379294e-01
3.83214921e-01 3.73367548e-01 6.97054714e-02 -3.52148801e-01
-2.44583800e-01 1.49088353e-01 -1.19276190e+00 1.13224089e+00
-8.50552678e-01 4.38027769e-01 4.19264697e-02 -6.55543447e-01
1.06696606e+00 8.36846590e-01 3.82860124e-01 -2.07284138e-01
5.82215965e-01 1.37717858e-01 9.91233811e-02 -7.30958223e-01
8.37117553e-01 -5.89982688e-01 -2.62549341e-01 2.74652869e-01
7.60380644e-03 -5.55165768e-01 2.08496004e-01 -3.00571352e-01
4.63327676e-01 -6.81593046e-02 3.75282377e-01 1.45986855e-01
6.35759294e-01 -3.22067022e-01 3.06923598e-01 4.28737223e-01
-1.97130889e-01 5.64613879e-01 5.75958431e-01 2.92646378e-01
-8.84546697e-01 -6.99042380e-01 1.86168820e-01 1.05872619e+00
-1.61176443e-01 -2.57506639e-01 -1.03892469e+00 -3.28724802e-01
-5.95565915e-01 1.20341432e+00 -4.02249694e-01 -7.83418640e-02
-1.05939634e-01 -3.92586499e-01 1.03828764e+00 8.47351253e-02
1.70661837e-01 -1.44757330e+00 -4.92011338e-01 2.70538002e-01
-5.49421728e-01 -1.19883132e+00 -7.13908911e-01 1.06583290e-01
-3.53547364e-01 -4.05338734e-01 -7.58083642e-01 -7.00867593e-01
3.50294918e-01 -3.21600735e-01 9.05499697e-01 -3.54840994e-01
1.94835290e-01 3.15403312e-01 -7.77577341e-01 -4.40937430e-01
-1.17477763e+00 -1.19345352e-01 7.73017183e-02 5.35241544e-01
-6.47508502e-02 -4.73645270e-01 8.92054662e-02 2.37114623e-01
-1.07770336e+00 1.82735592e-01 4.35175210e-01 6.08727276e-01
3.66455585e-01 3.13732654e-01 7.69459307e-01 -5.95260262e-01
1.38527071e+00 -3.40091825e-01 1.00688659e-01 2.65890926e-01
-2.30905488e-01 4.79961932e-03 1.10571897e+00 -1.03776884e+00
-1.35967839e+00 -1.56586617e-02 -5.14703214e-01 -6.16601706e-01
-5.86217880e-01 5.34156203e-01 -6.02819264e-01 3.52064371e-01
2.74799466e-01 2.99013942e-01 -4.65004658e-03 -4.83598299e-02
6.64329529e-01 1.21918190e+00 3.35988253e-01 -7.99327791e-01
3.76163930e-01 -9.54818949e-02 -4.86285448e-01 -1.23211908e+00
-4.70984787e-01 -5.72993346e-02 -1.43226236e-01 -7.99396753e-01
8.77669215e-01 -6.16321981e-01 -6.02167726e-01 5.14510989e-01
-1.31302011e+00 -3.52916420e-01 -2.87145108e-01 4.49025095e-01
-7.54977524e-01 3.63567173e-01 -6.11008048e-01 -1.37275374e+00
-2.23456785e-01 -1.31596303e+00 1.06685627e+00 8.25988874e-02
-7.38665879e-01 -7.43299127e-01 3.53522189e-02 1.06319681e-01
3.74337673e-01 1.78444281e-01 6.38522625e-01 -6.25483990e-01
2.88085043e-01 -4.76655029e-02 2.84859747e-01 7.09801614e-01
4.37465370e-01 5.63754559e-01 -9.46466804e-01 3.42357755e-01
3.04499477e-01 -3.95555168e-01 1.79586008e-01 2.14316234e-01
8.00222695e-01 -6.59539104e-01 3.63584667e-01 3.42282146e-01
8.17038059e-01 7.74507046e-01 5.61385572e-01 -2.49231815e-01
4.65096563e-01 9.29901481e-01 7.13470638e-01 5.54776847e-01
6.98034912e-02 6.31061614e-01 1.32375918e-02 -1.04511783e-01
-7.78468400e-02 -5.28948233e-02 6.40168488e-01 1.28119707e+00
1.09087743e-01 -7.35852480e-01 -5.19281089e-01 3.57519537e-01
-1.45178139e+00 -1.20889199e+00 1.00390457e-01 1.70936966e+00
1.12707329e+00 -4.62552207e-03 1.45120278e-01 3.42884541e-01
8.34058344e-01 4.15004611e-01 -1.82310551e-01 -9.29289162e-01
-1.58498302e-01 1.67885840e-01 -8.67964998e-02 7.79731095e-01
-9.58806396e-01 1.00613046e+00 6.10546160e+00 9.45908189e-01
-1.43081379e+00 -9.62388366e-02 5.81936002e-01 -1.03111096e-01
-5.47593653e-01 -3.09108287e-01 -3.84084642e-01 4.58109230e-01
1.14899659e+00 -2.96252817e-01 6.95196867e-01 6.25954568e-01
6.28342927e-01 4.05756801e-01 -8.89219582e-01 9.16638970e-01
3.04971874e-01 -5.95937073e-01 1.93075165e-01 -2.89132833e-01
5.68901062e-01 -7.22750962e-01 1.28272370e-01 3.65345895e-01
3.70184556e-02 -1.06836224e+00 1.18395519e+00 7.88831949e-01
6.90824747e-01 -1.03887665e+00 4.64782834e-01 4.09913540e-01
-9.88652885e-01 1.88744366e-01 1.46210849e-01 -6.69825971e-02
4.46419865e-01 4.44684297e-01 -8.44750226e-01 4.36104834e-01
1.98065266e-02 2.68400967e-01 -8.60809386e-02 3.33427727e-01
-3.11814576e-01 8.09336483e-01 -4.09267777e-05 -5.59453547e-01
1.35454834e-01 -3.33291650e-01 8.19706321e-01 1.59718597e+00
5.83125591e-01 2.27746606e-01 2.05305293e-01 1.01482904e+00
4.61351052e-02 6.27322018e-01 -4.84271884e-01 -6.50929749e-01
5.15955389e-01 1.21508217e+00 -4.83554006e-01 -5.60979307e-01
-9.38275307e-02 1.20810378e+00 -1.61668018e-01 4.02728587e-01
-1.05775869e+00 -3.82311046e-01 6.98018491e-01 -3.60193521e-01
8.41195229e-03 -5.21181971e-02 -2.94250160e-01 -9.14174020e-01
-1.12393335e-01 -1.45094180e+00 -2.84809858e-01 -1.12900841e+00
-1.19194102e+00 1.18955576e+00 -9.63292420e-02 -1.16815817e+00
-5.38931668e-01 -2.08712608e-01 -5.59799194e-01 1.02329576e+00
-1.01189852e+00 -8.62179756e-01 -9.25270990e-02 3.20248246e-01
7.39182770e-01 -1.44085348e-01 9.26075876e-01 3.33497107e-01
-6.40631974e-01 5.28142154e-01 -2.95853436e-01 -7.83148557e-02
7.34797060e-01 -1.04297912e+00 1.25099700e-02 7.14674056e-01
-5.01842387e-02 4.69007224e-01 1.25944507e+00 -5.75802028e-01
-1.03180087e+00 -8.78553391e-01 9.44652855e-01 8.47232342e-02
5.39539456e-01 -2.47828409e-01 -8.16505373e-01 1.48085222e-01
5.83980441e-01 -5.33380508e-01 8.43807697e-01 -1.45344123e-01
-7.41765648e-02 1.68576360e-01 -1.10240459e+00 8.57334197e-01
7.01062739e-01 -6.69360757e-01 -5.23242056e-01 -1.70847163e-01
1.00743949e+00 -2.60525286e-01 -9.07372952e-01 4.16819066e-01
4.62540835e-01 -9.50738609e-01 6.27850354e-01 -5.03025293e-01
6.59200788e-01 -3.30751508e-01 -3.58815819e-01 -1.73519623e+00
3.96357141e-02 -7.70747900e-01 1.33555189e-01 1.69545901e+00
5.34426868e-01 -3.70384395e-01 3.47862840e-01 3.66710871e-01
-1.24845691e-01 -5.49023628e-01 -5.17005086e-01 -6.46422982e-01
-1.81544811e-01 -7.97033608e-01 7.52469242e-01 1.03480148e+00
7.85165206e-02 4.04103458e-01 -6.50738418e-01 3.68571728e-01
1.60698190e-01 1.14295222e-02 6.54967666e-01 -7.11123705e-01
-3.89443606e-01 -7.80846655e-01 -1.03107445e-01 -7.41537571e-01
6.22106254e-01 -6.10585868e-01 3.76432747e-01 -1.15119493e+00
-2.29544088e-01 -1.56720534e-01 5.86274341e-02 1.78736180e-01
-3.55038643e-01 -1.79193407e-01 3.44197273e-01 -2.61545300e-01
-1.45055041e-01 9.25756872e-01 1.23563731e+00 1.00473175e-02
-3.95795971e-01 4.64078709e-02 -7.27036178e-01 6.48779809e-01
1.12115884e+00 -3.83386642e-01 -6.46223903e-01 1.76963612e-01
6.83364226e-03 5.11556506e-01 5.92970587e-02 -9.61243689e-01
-6.33940250e-02 -4.83269542e-01 -1.28921360e-01 -3.74759287e-01
6.71823084e-01 -7.79074848e-01 6.24045432e-01 4.11167666e-02
-7.05681741e-01 -6.50092289e-02 1.38986945e-01 1.49013614e-02
-5.20043910e-01 -4.63972151e-01 8.99645388e-01 1.19381808e-01
-2.74797291e-01 -1.17020279e-01 -7.47079849e-01 1.18389986e-01
9.73557532e-01 -1.50541022e-01 1.42584950e-01 -9.82482910e-01
-7.53736734e-01 -1.29339829e-01 3.80761504e-01 5.20792961e-01
5.80332518e-01 -1.44256747e+00 -7.28442192e-01 8.03085193e-02
1.28106391e-02 -6.55799091e-01 2.53247470e-01 4.35606033e-01
1.69055313e-02 -9.47782956e-03 5.89981303e-02 -3.05517584e-01
-1.64622986e+00 6.24798656e-01 5.34337640e-01 -7.77531490e-02
1.10009424e-01 6.30161524e-01 -1.01236314e-01 -5.82059145e-01
2.75238156e-01 -1.66845798e-01 -5.53410113e-01 4.30781990e-01
2.37440467e-01 3.23842943e-01 -2.49299288e-01 -1.08526719e+00
-1.16780393e-01 3.67025554e-01 5.15413642e-01 -7.88478792e-01
8.88559878e-01 -1.41225636e-01 -1.17345683e-01 8.97514701e-01
1.28922737e+00 4.51291442e-01 -9.07613397e-01 1.99094862e-01
-1.50754958e-01 -1.10299669e-01 3.27094197e-02 -8.98531735e-01
-8.38403404e-01 7.15266526e-01 3.02354068e-01 4.06868577e-01
1.35839689e+00 -2.82028526e-01 6.14724398e-01 -2.75155026e-02
-3.43346186e-02 -1.18577647e+00 1.71212390e-01 6.64478779e-01
1.15837944e+00 -6.88291013e-01 -5.89353263e-01 -3.89474690e-01
-1.31200469e+00 9.51171100e-01 5.21365821e-01 2.81413347e-01
4.12264049e-01 6.94592416e-01 5.70612550e-01 2.34504864e-01
-9.50708926e-01 -2.66461641e-01 3.53454918e-01 4.64682639e-01
7.71825671e-01 4.57828909e-01 -5.19337237e-01 8.84013236e-01
-6.79542422e-01 -6.82735443e-02 5.92630923e-01 5.01300395e-01
-4.04330820e-01 -1.33798707e+00 -5.84862709e-01 8.04199725e-02
-5.31056821e-01 -1.46512702e-01 -7.89606810e-01 3.04252595e-01
4.03780267e-02 1.55917859e+00 -1.57993883e-01 -6.79244101e-01
6.09023571e-01 5.87976336e-01 2.22773880e-01 -3.04906100e-01
-6.66780710e-01 5.06445825e-01 5.83117485e-01 -8.10646936e-02
-6.46049023e-01 -5.68773568e-01 -1.20263326e+00 -4.80874553e-02
-2.60538518e-01 4.14842427e-01 8.04291189e-01 7.56972790e-01
1.80390865e-01 1.07167757e+00 1.12789607e+00 -8.65764678e-01
-4.56118196e-01 -1.18547070e+00 -5.20870686e-01 4.33637142e-01
2.64700770e-01 -3.22370589e-01 -6.14022493e-01 5.28499596e-02] | [14.613619804382324, 6.45572566986084] |
3d79021f-5fac-402f-81db-71d4e8b29421 | qa-driven-zero-shot-slot-filling-with-weak | null | null | https://aclanthology.org/2021.acl-short.83 | https://aclanthology.org/2021.acl-short.83.pdf | QA-Driven Zero-shot Slot Filling with Weak Supervision Pretraining | Slot-filling is an essential component for building task-oriented dialog systems. In this work, we focus on the zero-shot slot-filling problem, where the model needs to predict slots and their values, given utterances from new domains without training on the target domain. Prior methods directly encode slot descriptions to generalize to unseen slot types. However, raw slot descriptions are often ambiguous and do not encode enough semantic information, limiting the models{'} zero-shot capability. To address this problem, we introduce QA-driven slot filling (QASF), which extracts slot-filler spans from utterances with a span-based QA model. We use a linguistically motivated questioning strategy to turn descriptions into questions, allowing the model to generalize to unseen slot types. Moreover, our QASF model can benefit from weak supervision signals from QA pairs synthetically generated from unlabeled conversations. Our full system substantially outperforms baselines by over 5{\%} on the SNIPS benchmark. | ['Yuan Zhang', 'Panupong Pasupat', 'Dian Yu', 'Qi Li', 'Luheng He', 'Xinya Du'] | 2021-08-01 | null | null | null | acl-2021-5 | ['zero-shot-slot-filling'] | ['natural-language-processing'] | [ 3.42235655e-01 8.83960903e-01 -1.50389820e-01 -7.85018146e-01
-1.17591143e+00 -5.09717524e-01 5.06831169e-01 -1.24415793e-01
-2.67963231e-01 1.22090709e+00 3.50534469e-01 -4.45845455e-01
5.49321055e-01 -7.71396458e-01 -5.19766688e-01 -8.19019377e-02
4.26818818e-01 1.23827767e+00 4.55458790e-01 -7.77502060e-01
-2.43379533e-01 -3.37709129e-01 -1.24803436e+00 5.96741796e-01
1.17184770e+00 6.92030191e-01 5.05407274e-01 4.96340126e-01
-1.08511209e+00 5.47049403e-01 -7.59432316e-01 -4.93019640e-01
-4.08618413e-02 -6.33358359e-01 -1.41573572e+00 4.43638265e-01
-8.47576708e-02 -3.87433589e-01 -2.71140635e-01 6.79037213e-01
1.68748915e-01 4.14560407e-01 3.10852736e-01 -1.23208404e+00
-4.71613497e-01 8.17569792e-01 2.00061247e-01 -1.70026764e-01
6.12093091e-01 5.45335934e-02 1.35781002e+00 -9.37066734e-01
8.23994875e-01 1.58940804e+00 3.62051100e-01 1.22026658e+00
-1.37572837e+00 -2.91185051e-01 2.17793867e-01 -3.77311744e-02
-8.59531105e-01 -7.39273548e-01 6.78987145e-01 -9.39855501e-02
1.19666195e+00 2.41124585e-01 1.06899418e-01 1.16054308e+00
-7.06225336e-01 1.08705652e+00 8.52524281e-01 -5.89332819e-01
4.90612328e-01 3.84879768e-01 3.63080621e-01 3.32935244e-01
-3.64618778e-01 -3.03117186e-01 -4.52032804e-01 -2.56108046e-01
3.65760237e-01 -2.03366384e-01 -3.76150236e-02 -2.91754574e-01
-1.01308405e+00 1.02776825e+00 -2.59580910e-02 2.35251710e-02
-1.82548985e-01 -3.63602459e-01 5.46775520e-01 4.16863024e-01
4.52951133e-01 7.97976553e-01 -7.83424020e-01 -3.53123993e-01
-6.04780555e-01 6.29884660e-01 1.09096754e+00 1.35868573e+00
1.04941463e+00 -1.67362794e-01 -4.82514441e-01 1.36700320e+00
2.11166456e-01 4.67583746e-01 5.20757437e-01 -1.23911011e+00
6.78641558e-01 5.66176414e-01 5.42204797e-01 6.00024313e-02
-3.85732561e-01 1.54246703e-01 -1.15276888e-01 -6.10033870e-01
4.06393558e-01 -3.55315536e-01 -1.15606630e+00 1.95391071e+00
4.97886032e-01 -1.29153594e-01 6.07628942e-01 7.28749752e-01
1.13740349e+00 7.84918129e-01 4.49955940e-01 -1.63200527e-01
1.64905643e+00 -1.27219450e+00 -9.88131583e-01 -7.91155875e-01
8.03835809e-01 -4.57944065e-01 1.66762686e+00 -2.54885226e-01
-1.03517401e+00 -3.65785033e-01 -6.74730241e-01 -2.85833240e-01
-2.32146755e-01 -1.95834920e-01 5.60953796e-01 5.98103940e-01
-9.14506078e-01 1.65660813e-01 -6.67360604e-01 -4.02656078e-01
1.16952740e-01 4.81260195e-02 -1.99571475e-01 -3.35831732e-01
-1.73147523e+00 6.61197543e-01 3.99576902e-01 -3.78531784e-01
-8.16789567e-01 -4.38040853e-01 -1.49065208e+00 1.48602501e-01
7.28601635e-01 -5.24764717e-01 2.00296140e+00 -4.27055180e-01
-1.56228149e+00 8.22195172e-01 -7.30014980e-01 -8.11330020e-01
-2.14432701e-02 3.23889889e-02 -3.66045475e-01 2.28387956e-02
5.34018099e-01 1.12685108e+00 3.93384963e-01 -1.22061110e+00
-4.63479251e-01 -1.10997856e-01 4.58831251e-01 4.14472789e-01
8.16702470e-02 -3.63988578e-01 -3.12131792e-01 -1.87947318e-01
2.95720935e-01 -7.80792952e-01 -3.50589335e-01 -3.76227856e-01
-5.39123535e-01 -5.05798817e-01 7.02077687e-01 -6.87583983e-01
9.26411390e-01 -2.12899780e+00 -1.80875987e-01 -4.58149046e-01
-1.10131390e-01 2.80659556e-01 -3.98957998e-01 7.05381930e-01
4.61037517e-01 -1.42823055e-01 -5.38554192e-01 -7.41155446e-01
4.60513502e-01 8.70010376e-01 -7.72420466e-01 -5.22721291e-01
5.52934885e-01 1.07517195e+00 -9.08021688e-01 -5.05971789e-01
9.36124846e-02 -1.32087827e-01 -8.37060034e-01 6.48109734e-01
-1.14180779e+00 4.93944913e-01 -3.75756234e-01 6.22848392e-01
5.91218710e-01 -2.88166851e-01 5.02743781e-01 2.61459589e-01
4.47878838e-01 1.07577145e+00 -7.22045600e-01 2.12901235e+00
-7.35970080e-01 3.49040955e-01 6.30037785e-02 -1.08737063e+00
1.13396788e+00 5.62681317e-01 6.64568394e-02 -8.09768677e-01
-1.50211394e-01 2.64882863e-01 -2.62887359e-01 -3.46886516e-01
9.15635526e-01 -5.57960153e-01 -6.66768849e-01 4.50948626e-01
3.52257192e-01 -4.77788121e-01 2.46088013e-01 3.63112777e-01
1.04796696e+00 -4.14741151e-02 3.59160565e-02 1.49357811e-01
3.23075622e-01 4.78939861e-01 8.24221134e-01 7.74916589e-01
-3.49420220e-01 4.70505387e-01 5.88307619e-01 -1.52550146e-01
-9.87462044e-01 -1.31903863e+00 -1.55711591e-01 1.45470357e+00
3.59562874e-01 -2.85408288e-01 -8.84076416e-01 -8.93693030e-01
-1.56164378e-01 1.14942551e+00 -2.21942052e-01 5.43155782e-02
-4.62371290e-01 -1.46734670e-01 4.56348687e-01 3.14846277e-01
4.55683082e-01 -1.34350944e+00 -3.54369998e-01 6.71514332e-01
-8.20364118e-01 -1.49596488e+00 -2.59897470e-01 2.55645335e-01
-7.03544915e-01 -7.48815894e-01 -7.28707850e-01 -8.95263970e-01
3.61335248e-01 2.92858809e-01 1.58772910e+00 -8.31928998e-02
3.96009581e-03 2.00490430e-01 -6.53181434e-01 -2.46462077e-01
-7.02035904e-01 2.44716644e-01 -2.04375640e-01 -5.61925590e-01
9.28594589e-01 -3.44815642e-01 -3.19637895e-01 6.35423541e-01
-7.54511297e-01 1.46754429e-01 2.01743737e-01 1.35391700e+00
3.10021371e-01 -4.59055185e-01 1.09959924e+00 -1.26659453e+00
8.02705526e-01 -5.58589041e-01 -2.79349566e-01 3.40543926e-01
-2.40162745e-01 3.72066349e-01 4.69077826e-01 -2.22065553e-01
-1.51625216e+00 -4.39933725e-02 -4.54802543e-01 7.19698817e-02
-3.51869941e-01 2.08889782e-01 -5.18246174e-01 6.44564271e-01
8.42935443e-01 2.80740261e-01 1.04828458e-02 -6.63850427e-01
7.05116034e-01 1.01985550e+00 6.11868024e-01 -9.24755573e-01
2.64338791e-01 1.96983606e-01 -8.13526750e-01 -7.32634723e-01
-1.19437039e+00 -7.15783954e-01 -3.87572944e-01 3.85418206e-01
7.09149539e-01 -9.95434403e-01 -2.04937518e-01 2.50759851e-02
-1.37605989e+00 -6.23330832e-01 -5.93990147e-01 8.36676545e-03
-7.28940725e-01 3.14499825e-01 -8.24920118e-01 -1.00215256e+00
-2.73359358e-01 -1.06587040e+00 1.28476644e+00 3.65433216e-01
-3.84844393e-01 -9.32168126e-01 2.05290750e-01 8.58956456e-01
4.05956089e-01 -4.55963701e-01 8.97916019e-01 -1.20364749e+00
-3.16761971e-01 1.16805561e-01 -1.76098049e-01 1.01099551e-01
1.24987349e-01 -8.86882365e-01 -1.15042055e+00 9.08791646e-02
-1.62449911e-01 -1.04882276e+00 6.97468638e-01 1.32280821e-02
7.80882120e-01 -3.34311545e-01 -2.63650477e-01 -6.92612603e-02
7.47851431e-01 2.98325866e-01 5.07957280e-01 1.69383630e-01
-2.15301607e-02 8.96036029e-01 1.16479909e+00 4.86443102e-01
8.52814555e-01 7.05037057e-01 1.14286363e-01 1.42845772e-02
4.16031182e-02 -7.06741631e-01 3.21240574e-02 4.13570493e-01
9.12202477e-01 -2.99614638e-01 -9.15121198e-01 8.70161593e-01
-1.73038757e+00 -5.44138908e-01 3.18099737e-01 1.72903442e+00
1.28942478e+00 2.08989352e-01 6.92854822e-02 -2.29559824e-01
7.59067237e-01 2.85091102e-01 -7.80959487e-01 -3.66994560e-01
8.74076262e-02 3.18249285e-01 -8.01744387e-02 8.09231520e-01
-9.37853098e-01 1.70066774e+00 6.05670404e+00 5.66421568e-01
-5.73502898e-01 2.66320944e-01 4.47127640e-01 3.25222313e-01
-7.23874807e-01 3.52459759e-01 -9.21548426e-01 2.66501069e-01
1.08237028e+00 -1.91074833e-01 3.42369467e-01 1.06742430e+00
-1.54488221e-01 -1.23187020e-01 -9.04661179e-01 5.96430600e-01
-1.56780958e-01 -1.25283754e+00 -3.84044164e-04 -2.39375785e-01
5.14936686e-01 -1.57754689e-01 -2.14045241e-01 1.12697470e+00
8.02403748e-01 -9.83318925e-01 2.14942321e-01 -8.46396461e-02
9.19464409e-01 -1.93651378e-01 6.75287485e-01 5.31829596e-01
-8.67862880e-01 8.78140628e-02 -8.22532296e-01 -2.27650866e-01
8.02599072e-01 3.10743570e-01 -1.60907710e+00 1.81340992e-01
9.77766216e-02 1.56492814e-01 8.03758204e-02 5.04860878e-01
-3.58155221e-01 6.78751290e-01 -2.13719025e-01 -6.15592375e-02
3.18683803e-01 5.32865375e-02 3.40471685e-01 1.00979638e+00
7.93178752e-02 4.04414564e-01 3.76109153e-01 1.03737521e+00
-1.09055594e-01 -8.64038095e-02 -3.50319415e-01 -3.43234181e-01
1.04785931e+00 1.05812776e+00 -3.94718140e-01 -5.79344988e-01
-4.88538027e-01 1.15943348e+00 2.95395166e-01 3.17748278e-01
-3.98936570e-01 -4.24501121e-01 9.86276805e-01 -1.44194946e-01
1.84397951e-01 -1.06622115e-01 -2.49336749e-01 -1.30359316e+00
1.17004709e-02 -8.47905040e-01 4.17985409e-01 -7.30927587e-01
-1.23861182e+00 8.66640210e-01 -5.36927581e-02 -9.31352258e-01
-1.00632274e+00 -4.02454883e-01 -5.26726782e-01 9.66261446e-01
-1.60451055e+00 -9.35727835e-01 -1.05120882e-01 3.74815881e-01
1.21388912e+00 -1.64344251e-01 1.15525317e+00 -7.15678651e-03
-1.75675243e-01 5.44361949e-01 -3.04061621e-01 2.76586831e-01
7.11150348e-01 -1.26353931e+00 1.20012164e+00 3.88680905e-01
2.03968287e-01 5.07568300e-01 1.02210617e+00 -5.98934889e-01
-7.45793104e-01 -8.92846227e-01 1.24798417e+00 -5.36140144e-01
5.12857318e-01 -7.33451903e-01 -1.30792201e+00 6.76855683e-01
1.74720306e-02 -1.19848520e-01 9.12572563e-01 5.03064811e-01
-4.72468317e-01 4.87249732e-01 -1.28736019e+00 4.13870841e-01
1.16399908e+00 -8.29738796e-01 -1.36903059e+00 4.06425327e-01
1.47637904e+00 -6.35100067e-01 -4.36656535e-01 1.85804740e-01
2.03590635e-02 -7.46009886e-01 8.71199071e-01 -8.90869915e-01
1.94121793e-01 4.20624912e-02 -3.53987962e-01 -1.38635898e+00
2.94845223e-01 -4.79041427e-01 5.97246131e-03 1.24591684e+00
7.90437937e-01 -4.03850973e-01 1.06397235e+00 1.18533647e+00
-3.85236561e-01 -3.47981572e-01 -1.14600527e+00 -8.19560826e-01
1.75231859e-01 -4.22941297e-01 1.00120378e+00 8.57448697e-01
6.65095985e-01 7.74396420e-01 -3.62030119e-01 -1.45136267e-01
2.46707529e-01 1.77639857e-01 9.15838003e-01 -1.09598041e+00
-4.52928334e-01 2.77999490e-01 1.11916304e-01 -1.69527864e+00
6.27528846e-01 -5.30636609e-01 5.92068970e-01 -1.65510762e+00
-1.23248011e-01 -8.68886948e-01 1.75217673e-01 6.58571362e-01
-3.54114026e-01 -9.28614661e-02 1.51717901e-01 1.35704637e-01
-9.11084592e-01 9.05907214e-01 1.05700636e+00 -1.42172113e-01
-5.26767015e-01 6.05050139e-02 -7.11825132e-01 3.11439276e-01
7.16665506e-01 -1.98135197e-01 -8.33885372e-01 -4.30066109e-01
-2.40414262e-01 8.04710150e-01 -5.80034824e-03 -5.95744789e-01
-5.40698171e-02 -3.05535614e-01 -4.53773499e-01 -3.78077686e-01
9.45036590e-01 -3.64145845e-01 -3.86671096e-01 7.00662658e-02
-6.96718395e-01 -4.32035029e-01 2.72953391e-01 4.39910412e-01
-4.36322540e-01 -4.63442564e-01 4.52366620e-01 -4.57145840e-01
-9.66111302e-01 7.24054649e-02 -3.23337048e-01 6.76373541e-01
6.10243261e-01 3.24021764e-02 -4.91009295e-01 -6.40665472e-01
-1.04351068e+00 6.22581780e-01 4.75473732e-01 5.15466928e-01
3.63631636e-01 -1.21906805e+00 -3.83715689e-01 2.74104714e-01
5.19400537e-01 4.38901961e-01 5.46403408e-01 1.26793340e-01
2.13112980e-02 7.66791344e-01 -5.16581871e-02 -5.16891003e-01
-7.62847662e-01 3.88106823e-01 2.48377249e-01 -3.04310679e-01
-4.30194080e-01 9.54935730e-01 3.83732259e-01 -1.18682337e+00
1.84831381e-01 -1.37156025e-01 -1.20585836e-01 -2.09086597e-01
6.41772449e-01 -2.37997964e-01 -1.08724266e-01 -4.25917238e-01
-1.29030153e-01 -3.37215841e-01 -2.58618832e-01 -7.16323376e-01
9.25573111e-01 -5.71189344e-01 2.47052476e-01 5.15056252e-01
8.25108230e-01 -3.41485918e-01 -1.38128388e+00 -8.83754075e-01
3.90687793e-01 -4.45403218e-01 -5.64764738e-01 -9.29684818e-01
-8.98101777e-02 1.03885770e+00 4.79367562e-02 3.31635326e-01
5.72976768e-01 4.41818714e-01 1.41606987e+00 6.99182391e-01
5.96510768e-01 -1.28051078e+00 2.15150312e-01 9.69815195e-01
5.51874042e-01 -1.45862925e+00 -9.24178123e-01 -6.17147982e-01
-1.15567005e+00 7.30213046e-01 8.81547570e-01 4.49943930e-01
1.58973828e-01 -7.95823149e-03 4.24751103e-01 -1.43203400e-02
-1.06014073e+00 -5.25520384e-01 -1.10323153e-01 9.62219834e-01
4.27981555e-01 2.23461613e-01 -1.89425170e-01 9.57048535e-01
-4.52775747e-01 -9.86589342e-02 4.73848671e-01 9.69674468e-01
-1.00620592e+00 -1.59950614e+00 -1.37689084e-01 3.49771827e-01
-2.12191530e-02 -3.52067977e-01 -4.22676086e-01 3.97831470e-01
-3.19191873e-01 1.25130129e+00 2.71955460e-01 -1.80785552e-01
3.15435112e-01 7.62695611e-01 -7.05392733e-02 -1.27720547e+00
4.49554846e-02 -1.82022646e-01 7.16103673e-01 -3.66795868e-01
-5.35404235e-02 -5.86626589e-01 -1.54491520e+00 1.31458178e-01
-2.11958349e-01 9.12617445e-01 5.17682850e-01 1.24110663e+00
4.71972585e-01 2.40718573e-01 4.34884697e-01 -4.06542242e-01
-7.90281892e-01 -1.15413678e+00 -4.89879608e-01 4.94733632e-01
2.15951130e-01 -8.40754986e-01 -1.81968838e-01 -1.48174047e-01] | [12.608088493347168, 7.522870063781738] |
45a861da-18f1-4928-a7bd-ebbc0b57072d | translating-robot-skills-learning | null | null | https://openreview.net/forum?id=NPJ5zWk_IQj | https://openreview.net/pdf?id=NPJ5zWk_IQj | Translating Robot Skills: Learning Unsupervised Skill Correspondences Across Robots | In this paper, we explore how we can endow robots with the ability to learn correspondences between their own skills, and those of morphologically different robots in different domains, in an entirely unsupervised manner. We make the insight that different morphological robots use similar task strategies to solve similar tasks. Based on this insight, we frame learning skill correspondences as a problem of matching distributions of sequences of skills across robots. We then present an unsupervised objective that encourages a learnt skill translation model to match these distributions across domains, inspired by recent advances in unsupervised machine translation. Our approach is able to learn semantically meaningful correspondences between skills across 3 robot domain pairs despite being completely unsupervised. Further, the learnt correspondences enable the transfer of task strategies across robots and domains.
We present dynamic visualizations of our results at https://sites.google.com/view/translatingrobotskills/home. | ['Jean Oh', 'Stuart Anderson', 'Vikash Kumar', 'Aravind Rajeswaran', 'Yixin Lin', 'Tanmay Shankar'] | 2021-09-29 | null | null | null | null | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 3.61398280e-01 3.81776899e-01 1.91645309e-01 -5.13775587e-01
-4.14030612e-01 -1.13993239e+00 6.73384786e-01 8.36071372e-02
-3.68522078e-01 6.55644774e-01 2.00140804e-01 3.88866439e-02
-3.13012004e-01 -5.48654914e-01 -9.58378613e-01 -3.80128622e-01
2.91329384e-01 9.41769719e-01 1.70745440e-02 -4.01768386e-01
3.43845397e-01 2.87401468e-01 -1.43378055e+00 8.88735652e-02
1.09693992e+00 2.19912723e-01 9.13566709e-01 5.11360347e-01
-1.12909950e-01 8.04339767e-01 -1.30423501e-01 -2.49866739e-01
6.15515947e-01 -6.16694808e-01 -1.27392781e+00 8.95951688e-02
4.94347990e-01 -2.14448702e-02 -3.74236703e-01 1.27600050e+00
-9.43503156e-02 2.52001494e-01 1.09454966e+00 -1.44377851e+00
-1.16486859e+00 7.54078507e-01 6.23188615e-02 -2.61630654e-01
5.85926712e-01 1.91275015e-01 8.67702484e-01 -4.79020059e-01
6.90141797e-01 1.28085899e+00 4.92135733e-01 8.45597923e-01
-1.45000517e+00 -4.58653212e-01 -1.54381692e-01 6.39853179e-02
-1.02298081e+00 -3.07136416e-01 6.21490479e-01 -7.50174165e-01
7.57056296e-01 -3.40541333e-01 6.07809961e-01 1.02337337e+00
1.20169200e-01 5.43556213e-01 1.38841832e+00 -4.59026575e-01
-1.73879473e-03 5.26027307e-02 -2.65142530e-01 8.35786998e-01
2.79478729e-01 2.98406750e-01 -7.40921259e-01 2.79467791e-01
1.19240570e+00 9.21551585e-02 6.24762103e-02 -8.63151610e-01
-1.55191505e+00 5.38372397e-01 4.38990653e-01 4.94749427e-01
-1.47114083e-01 3.67067933e-01 1.28526896e-01 9.46447313e-01
-1.04629606e-01 1.27803326e+00 -7.31513500e-01 -1.50299132e-01
-5.66177480e-02 2.20033079e-01 7.81246126e-01 1.61194932e+00
1.27055836e+00 -1.81306630e-01 3.63035470e-01 7.96824157e-01
1.46084920e-01 6.36347651e-01 7.13937819e-01 -1.51579916e+00
1.11178041e-01 6.22752964e-01 2.13400409e-01 -8.07755351e-01
-2.39975497e-01 1.10483266e-01 -2.64779091e-01 2.36778781e-01
7.30938911e-01 2.64124689e-03 -6.66296482e-01 2.18041992e+00
-2.96519976e-03 -1.83900476e-01 4.00095135e-01 6.50714993e-01
2.95847267e-01 2.70829320e-01 3.68094146e-02 3.23428839e-01
1.08776999e+00 -1.02586627e+00 -2.65316755e-01 -5.52910030e-01
8.34424019e-01 -6.57945693e-01 1.57943809e+00 2.18925308e-02
-1.10817730e+00 -7.19882309e-01 -7.85410762e-01 -2.66741961e-01
-4.69438016e-01 -3.92253585e-02 6.91567659e-01 -1.59829799e-02
-1.28724313e+00 1.00050867e+00 -7.60233819e-01 -1.24693203e+00
3.79351050e-01 3.27719688e-01 -6.54144704e-01 1.43400922e-01
-6.24799669e-01 1.34663367e+00 5.34853220e-01 -6.19032621e-01
-1.31008792e+00 -4.13244903e-01 -1.06018949e+00 -1.67086080e-01
2.61537045e-01 -8.84801507e-01 1.83395004e+00 -1.30516291e+00
-1.68241966e+00 1.33000576e+00 1.38763249e-01 -2.23217547e-01
1.86218470e-01 -3.04187089e-01 2.76283741e-01 5.19334599e-02
4.62537169e-01 1.02509534e+00 6.19166732e-01 -1.36377883e+00
-7.76567698e-01 -3.79191160e-01 2.66035944e-01 4.65141594e-01
-2.93371111e-01 -1.58780590e-02 8.43446627e-02 -6.39388561e-01
2.23280177e-01 -1.15223086e+00 -6.56100884e-02 4.78443578e-02
-2.43901722e-02 -2.41631851e-01 1.67923957e-01 -4.37659025e-01
2.09777337e-02 -2.12708378e+00 1.00038898e+00 -1.19073600e-01
6.76508322e-02 -4.91374463e-01 -4.35239583e-01 4.95944738e-01
9.37575623e-02 -1.09262370e-01 -5.72979391e-01 -2.08557472e-01
2.73162812e-01 6.65286720e-01 -2.80215681e-01 3.36421043e-01
2.95181841e-01 9.77843881e-01 -1.48190010e+00 -2.28754789e-01
-1.06625237e-01 -2.40865961e-01 -4.80461478e-01 6.19253755e-01
-3.16878557e-01 1.21359265e+00 -4.89851862e-01 3.92198473e-01
1.47888854e-01 -7.80390427e-02 5.77201188e-01 4.17262703e-01
5.43416031e-02 5.10425448e-01 -6.26714051e-01 2.39557099e+00
-5.38441956e-01 5.78798175e-01 5.03224581e-02 -1.23757041e+00
1.08499980e+00 1.52149796e-01 1.39322788e-01 -4.98997331e-01
2.44683146e-01 5.98305404e-01 1.43807322e-01 -5.11283159e-01
3.95356357e-01 -3.73383522e-01 -4.31519598e-01 7.42431104e-01
5.95840275e-01 -1.10481966e+00 1.09419249e-01 -3.45002487e-02
1.03743470e+00 6.08900428e-01 2.65295774e-01 -6.28512442e-01
9.12360251e-02 4.61914331e-01 4.01238233e-01 7.06146300e-01
-7.29299113e-02 2.66050696e-01 1.74752280e-01 -2.77575642e-01
-1.43345344e+00 -1.48826778e+00 3.10838580e-01 1.38967204e+00
3.31805319e-01 3.88381332e-02 -6.78108633e-01 -5.27037680e-01
5.10169156e-02 5.08594930e-01 -7.43607759e-01 -1.80344552e-01
-6.76303387e-01 4.48247701e-01 6.62712932e-01 5.76794922e-01
3.07186306e-01 -1.50218797e+00 -9.24846232e-01 -2.52527267e-01
-2.78191064e-02 -9.76368010e-01 -2.29688570e-01 5.33874333e-01
-1.12996113e+00 -1.03042305e+00 -7.30065405e-01 -1.62857604e+00
1.00946558e+00 3.52106065e-01 1.24775338e+00 4.33098152e-02
5.25860041e-02 9.25553858e-01 -6.34502649e-01 -4.82279032e-01
-9.57162976e-01 1.79827258e-01 5.92805624e-01 -7.15083838e-01
2.31036469e-01 -1.15227914e+00 -8.35148692e-02 4.09372538e-01
-9.31014061e-01 3.14785033e-01 5.57715833e-01 7.25115240e-01
3.76018554e-01 -2.62399048e-01 2.79721290e-01 -6.45777822e-01
7.65270829e-01 -5.43435633e-01 -5.65217674e-01 3.09473544e-01
-3.29121143e-01 5.87986112e-01 7.11011350e-01 -7.18910694e-01
-8.88355494e-01 3.10113251e-01 5.14441550e-01 -2.46759832e-01
-6.91518903e-01 1.23998977e-01 -1.14553228e-01 8.52827355e-02
7.93155849e-01 4.40265924e-01 3.04199338e-01 -6.55301809e-01
7.90437520e-01 5.34091353e-01 7.25972235e-01 -1.31504679e+00
1.09655786e+00 3.98634225e-01 -3.01163852e-01 -4.45927322e-01
-5.22006273e-01 -2.10070565e-01 -1.16324019e+00 9.81645063e-02
9.36935365e-01 -7.27077305e-01 -3.36998791e-01 6.13379717e-01
-1.07544553e+00 -1.19335330e+00 -7.00991452e-01 4.41761762e-01
-1.45910370e+00 9.81643125e-02 -4.82831150e-01 -1.21597312e-01
2.59434074e-01 -1.08800304e+00 8.29435349e-01 4.80024517e-01
-6.55741870e-01 -1.07051480e+00 3.56287718e-01 -1.00041576e-01
2.36498147e-01 -1.60194397e-01 1.07407188e+00 -8.49740267e-01
-2.94055402e-01 5.94056010e-01 -6.21205382e-02 2.83187181e-01
7.21106946e-01 -4.86859083e-01 -4.15506572e-01 -2.68935114e-01
1.05695248e-01 -7.33764231e-01 2.87942648e-01 -3.96428891e-02
7.01959550e-01 -1.62023202e-01 -2.30513275e-01 5.60037553e-01
1.15460587e+00 1.22824736e-01 3.14373165e-01 5.61891973e-01
6.67992175e-01 1.26475251e+00 7.82517254e-01 -1.41579688e-01
4.64469343e-01 5.50780833e-01 1.50871962e-01 2.55165517e-01
-8.21785778e-02 -6.38090312e-01 4.63518173e-01 9.35284376e-01
-6.14145398e-02 3.69472474e-01 -1.11340094e+00 9.63036597e-01
-1.97010243e+00 -7.12949753e-01 2.90817678e-01 1.77639496e+00
1.18954003e+00 -2.82738537e-01 2.43657812e-01 -4.78761315e-01
7.00544178e-01 -4.53095168e-01 -6.66569591e-01 -7.55255759e-01
7.01398104e-02 3.14979821e-01 3.52421224e-01 4.15461779e-01
-8.80200028e-01 1.37219429e+00 6.34790850e+00 2.81917244e-01
-6.03822887e-01 1.11769930e-01 -1.78438071e-02 3.30029130e-01
-3.23691517e-01 3.38298291e-01 -1.77703679e-01 2.75837332e-01
5.61627150e-01 -3.72761279e-01 9.70201373e-01 7.40413368e-01
-1.42599523e-01 7.76054338e-02 -1.69746494e+00 7.57087409e-01
-1.43656343e-01 -9.15174425e-01 2.78089512e-02 -7.49553442e-02
7.98896372e-01 2.14311704e-02 2.38758907e-01 2.53347069e-01
1.25168395e+00 -1.13519478e+00 9.99308944e-01 4.97953266e-01
6.64209545e-01 -2.84129679e-01 1.85153201e-01 5.08460760e-01
-8.47919524e-01 -2.19704404e-01 -6.43961608e-01 -6.03140593e-01
-1.17123984e-01 -1.84440807e-01 -1.02385688e+00 3.21294546e-01
7.88165748e-01 9.46814775e-01 -3.35212529e-01 5.57687104e-01
-7.72458553e-01 -1.11959845e-01 -7.16225058e-02 -9.88460183e-02
-3.22915427e-02 -4.43878442e-01 4.93368000e-01 9.48273242e-01
5.16501427e-01 5.57638425e-03 1.98058397e-01 1.12625563e+00
-8.38152319e-02 -8.82085636e-02 -1.16316426e+00 -1.99856490e-01
6.97245479e-01 8.79827976e-01 -6.97134614e-01 -2.05243796e-01
-2.61879891e-01 1.28713298e+00 7.71796167e-01 1.28482491e-01
-5.54751992e-01 -1.71825543e-01 1.04472697e+00 -1.94165722e-01
5.78716993e-02 -6.84108913e-01 -3.86527419e-01 -9.54166830e-01
-2.54627973e-01 -1.05652583e+00 9.35926277e-04 -1.12601030e+00
-1.47532141e+00 1.43929169e-01 1.72320604e-01 -1.36054575e+00
-3.46595198e-01 -8.32967222e-01 -3.23158294e-01 6.23672545e-01
-1.12056637e+00 -1.12588978e+00 -1.67474493e-01 6.79293871e-01
7.51629472e-01 -4.51035440e-01 1.00883806e+00 -3.19641769e-01
8.00816491e-02 3.12700570e-01 1.67020380e-01 2.33638272e-01
8.34131300e-01 -1.52137709e+00 5.98770320e-01 5.03675461e-01
4.23434019e-01 8.72997403e-01 8.34819376e-01 -6.10104144e-01
-1.36774731e+00 -8.38627338e-01 6.19979620e-01 -8.55124712e-01
9.45308626e-01 -3.12642813e-01 -7.26051807e-01 1.08610904e+00
4.94304925e-01 -6.25999033e-01 5.73324859e-01 4.97751944e-02
-5.33260584e-01 3.98355186e-01 -1.12929237e+00 8.67301702e-01
1.62632275e+00 -6.94229841e-01 -1.40340364e+00 4.30091828e-01
9.31717873e-01 -3.58377218e-01 -8.89719129e-01 7.59714320e-02
5.43293238e-01 -5.15669584e-01 7.41330624e-01 -8.89346004e-01
8.04555595e-01 -2.96237200e-01 -3.00967157e-01 -1.85691738e+00
-4.19654757e-01 -5.60954154e-01 4.50900763e-01 1.16387928e+00
4.04259115e-01 -8.68281841e-01 4.12650675e-01 2.66992927e-01
-2.79112786e-01 -1.51466280e-01 -4.94299173e-01 -1.15408278e+00
6.59327328e-01 2.31791325e-02 5.73932230e-01 1.28706479e+00
4.43323761e-01 3.30389857e-01 5.03366441e-02 1.56829670e-01
4.11605418e-01 3.22124928e-01 9.39958274e-01 -1.33481157e+00
-5.27340233e-01 -4.75385845e-01 -3.48780125e-01 -1.18560803e+00
6.45607054e-01 -1.24668324e+00 5.15245795e-01 -1.41270351e+00
3.54962379e-01 -6.27475023e-01 3.17368507e-02 7.54948318e-01
2.24963039e-01 1.41262142e-02 4.05802161e-01 6.16286814e-01
-5.04949212e-01 4.21662122e-01 1.34029996e+00 -2.94968975e-03
-1.84598595e-01 -5.12394428e-01 -9.41773474e-01 1.07937706e+00
1.37740421e+00 -8.48874629e-01 -5.00265718e-01 -1.15972257e+00
3.30020189e-01 -4.90670800e-01 2.74023801e-01 -1.03246534e+00
1.91934958e-01 -5.54244936e-01 6.12581335e-02 3.58511299e-01
1.92868263e-02 -8.90184760e-01 -7.80011117e-02 5.58189511e-01
-4.73474443e-01 2.83107966e-01 2.42100954e-01 3.56611401e-01
6.91179261e-02 -3.53314042e-01 7.00598836e-01 -6.21628761e-01
-1.03108418e+00 -1.98663965e-01 -6.01217926e-01 2.85305709e-01
1.06686962e+00 -2.98029959e-01 -2.64062494e-01 -2.91008472e-01
-8.61707628e-01 3.36316943e-01 1.34297013e+00 4.58326697e-01
4.44555134e-01 -1.19835651e+00 -5.90998113e-01 4.87810783e-02
4.25009608e-01 7.15126768e-02 -3.03241491e-01 4.92763102e-01
-6.33391857e-01 1.90876096e-01 -9.59156990e-01 -5.01380384e-01
-9.56580698e-01 4.25624788e-01 3.09274882e-01 1.67076319e-01
-3.75672847e-01 1.00604260e+00 4.70403045e-01 -1.30277359e+00
-1.22804955e-01 -3.91948432e-01 -8.11503530e-02 -4.51874644e-01
-5.83484285e-02 1.82138667e-01 -5.43865383e-01 -5.78618824e-01
-1.81936532e-01 9.59741950e-01 1.77348211e-01 -3.86809021e-01
1.32467568e+00 -2.85620451e-01 -3.95145684e-01 5.25116980e-01
8.22476327e-01 -1.30459353e-01 -1.48392928e+00 -3.96307260e-01
2.10129350e-01 -4.16161537e-01 -9.70408440e-01 -6.03939354e-01
-3.87191862e-01 7.48096526e-01 3.01317453e-01 4.36726883e-02
1.01738119e+00 7.58668363e-01 4.33393329e-01 7.66414404e-01
8.97045135e-01 -1.17062187e+00 5.26477635e-01 1.09029913e+00
8.24740291e-01 -1.11299086e+00 -4.25904781e-01 -3.17450285e-01
-7.02771544e-01 9.32594955e-01 6.22075200e-01 -4.82601285e-01
2.12139040e-02 1.97082996e-01 6.06227964e-02 -1.48620784e-01
-4.37769711e-01 -5.54277122e-01 -1.68876633e-01 1.34683371e+00
2.96114057e-01 1.96592212e-01 -7.08062798e-02 2.95561999e-01
-7.01209188e-01 -1.64939269e-01 5.61821043e-01 1.23148286e+00
-6.67266428e-01 -1.47505629e+00 -1.79575086e-01 1.05353752e-02
1.10047907e-01 1.01301484e-01 -8.91715169e-01 6.84296131e-01
1.21903002e-01 8.07589233e-01 2.43868187e-01 -5.04649878e-01
4.34864700e-01 2.32256263e-01 1.16317177e+00 -1.31458628e+00
-2.57303715e-01 -5.74310303e-01 -1.56050578e-01 -4.45694357e-01
-5.61969936e-01 -7.60246217e-01 -1.48491263e+00 -1.67479023e-01
4.56581235e-01 9.66122523e-02 3.64967376e-01 9.71805751e-01
2.99162686e-01 3.09523880e-01 2.94384718e-01 -9.25047100e-01
-7.09375143e-01 -9.98438418e-01 -6.59403026e-01 7.98313320e-01
1.72228962e-02 -8.54675472e-01 -2.29954824e-01 5.67195714e-01] | [4.387297630310059, 0.9698885083198547] |
2cdfaea4-89f3-48db-a542-41e4810108b5 | swiden-convolutional-neural-networks-for | 1607.08764 | null | http://arxiv.org/abs/1607.08764v1 | http://arxiv.org/pdf/1607.08764v1.pdf | SwiDeN : Convolutional Neural Networks For Depiction Invariant Object Recognition | Current state of the art object recognition architectures achieve impressive
performance but are typically specialized for a single depictive style (e.g.
photos only, sketches only). In this paper, we present SwiDeN : our
Convolutional Neural Network (CNN) architecture which recognizes objects
regardless of how they are visually depicted (line drawing, realistic shaded
drawing, photograph etc.). In SwiDeN, we utilize a novel `deep' depictive
style-based switching mechanism which appropriately addresses the
depiction-specific and depiction-invariant aspects of the problem. We compare
SwiDeN with alternative architectures and prior work on a 50-category Photo-Art
dataset containing objects depicted in multiple styles. Experimental results
show that SwiDeN outperforms other approaches for the depiction-invariant
object recognition problem. | ['Srinivas S. S. Kruthiventi', 'Venkatesh Babu R', 'Ravi Kiran Sarvadevabhatla', 'Shiv Surya'] | 2016-07-29 | null | null | null | null | ['depiction-invariant-object-recognition'] | ['computer-vision'] | [ 3.66649181e-01 -4.39669698e-01 -1.14704400e-01 -5.62309325e-01
1.28950030e-01 -9.93732154e-01 1.03505635e+00 -5.22841632e-01
-1.23577058e-01 3.30777258e-01 -3.95602770e-02 -3.21046382e-01
5.10557555e-02 -8.52040946e-01 -8.51489782e-01 -3.93233865e-01
2.22826183e-01 4.83814895e-01 1.82936236e-01 -2.53274798e-01
4.63027090e-01 1.44084978e+00 -1.61632860e+00 8.70214403e-01
5.33128195e-02 1.06954062e+00 -2.77569026e-01 9.62213993e-01
-6.50586128e-01 6.69484973e-01 -1.09245074e+00 -5.50603211e-01
3.63630414e-01 -1.71918929e-01 -5.49933851e-01 4.49179769e-01
1.67523813e+00 -5.60537100e-01 -4.11964357e-01 7.33226418e-01
3.37385863e-01 -1.29621074e-01 9.82045770e-01 -1.65329695e+00
-1.36756885e+00 3.41615856e-01 -5.29455125e-01 2.38171861e-01
4.24584709e-02 3.21386784e-01 6.71188056e-01 -1.06916583e+00
7.28913069e-01 1.58965039e+00 7.33977497e-01 6.42714858e-01
-1.46245313e+00 -7.17811167e-01 3.93838316e-01 -1.97201651e-02
-1.31676650e+00 -3.25897902e-01 1.16861773e+00 -3.87318254e-01
9.76839185e-01 4.35096413e-01 8.91886115e-01 1.39057791e+00
-2.64100488e-02 1.10713005e+00 1.24135804e+00 -3.54293108e-01
3.07748914e-01 -1.11365303e-01 5.99053912e-02 7.35725164e-01
3.58490705e-01 1.00506609e-02 -3.53307307e-01 2.94386983e-01
1.43112707e+00 6.69057891e-02 6.81240559e-02 -6.52458072e-01
-1.16187787e+00 3.21659893e-01 6.69116259e-01 3.45638454e-01
-4.32296582e-02 5.46105146e-01 5.23350358e-01 5.09158432e-01
1.03815518e-01 5.01501381e-01 -1.52848929e-01 6.58656983e-03
-9.73545551e-01 1.64831713e-01 6.71248436e-01 1.23954821e+00
5.15025616e-01 8.12456965e-01 -3.21253836e-01 9.17314649e-01
-2.10270777e-01 5.52679241e-01 5.24072396e-03 -7.75894582e-01
4.39108282e-01 9.20358062e-01 1.07905902e-01 -1.12535012e+00
-3.49645048e-01 -3.56260985e-01 -1.06986201e+00 7.01229334e-01
2.52256840e-01 1.64461300e-01 -1.42732286e+00 1.27765214e+00
-1.73638999e-01 -2.90568080e-02 -7.16173872e-02 9.40003395e-01
1.41244078e+00 4.77780879e-01 4.78323430e-01 6.65405452e-01
1.40390682e+00 -1.08263147e+00 -5.02536893e-01 -9.43474248e-02
1.39940847e-02 -7.84426212e-01 1.31588197e+00 4.17358488e-01
-1.09832048e+00 -8.80549610e-01 -1.32273257e+00 -3.64115566e-01
-9.66780722e-01 8.43642771e-01 7.00353622e-01 8.77125323e-01
-1.20580840e+00 5.70183694e-01 -1.43318191e-01 -3.60254705e-01
8.59564722e-01 1.73040107e-01 -5.56459069e-01 -2.56347042e-02
-5.54343998e-01 8.59975994e-01 4.72035646e-01 1.25092834e-01
-9.14436877e-01 -7.34218419e-01 -6.90872610e-01 2.80445457e-01
-3.46007310e-02 -7.20790863e-01 1.20184577e+00 -1.74503636e+00
-1.66487360e+00 1.23257387e+00 2.45525628e-01 -1.63753361e-01
8.25724602e-01 -8.97397846e-02 -6.72362030e-01 1.86162964e-01
-3.63849461e-01 1.00065958e+00 1.09957969e+00 -1.82366633e+00
-2.90180415e-01 3.68702635e-02 4.23159450e-01 2.71017384e-02
-2.05196500e-01 1.24059692e-01 -3.07179481e-01 -9.43962157e-01
-2.33188018e-01 -8.66181850e-01 2.60180503e-01 8.69887412e-01
-7.12811708e-01 -3.30406517e-01 1.39129019e+00 -5.29014394e-02
8.60698164e-01 -2.17358875e+00 1.08213477e-01 -2.25785524e-02
6.23236001e-02 7.53562987e-01 -5.45378864e-01 7.46543527e-01
-3.27454925e-01 2.30911151e-01 -2.09557742e-01 -5.25834262e-01
3.34057659e-01 4.79548156e-01 -3.61821651e-01 3.06484938e-01
5.27605414e-01 1.20628071e+00 -4.09576625e-01 -4.20336157e-01
6.08064473e-01 5.37514627e-01 9.76278782e-02 1.28061891e-01
-1.64296091e-01 -4.37074751e-02 -3.93802710e-02 9.63165104e-01
1.04113197e+00 -1.49770439e-01 1.75964668e-01 -6.45763636e-01
-2.10424319e-01 -4.04512882e-01 -1.18817687e+00 1.61793363e+00
-5.40677547e-01 1.12177050e+00 -1.93412021e-01 -7.09365666e-01
1.15283203e+00 -1.45303130e-01 -2.38047704e-01 -5.74230015e-01
1.51698172e-01 -1.26393214e-01 -1.67364419e-01 -2.04561323e-01
4.49560553e-01 -1.03831384e-02 -1.58431791e-02 3.96265149e-01
-6.57487884e-02 -3.10089707e-01 3.98888849e-02 4.39257063e-02
7.55531073e-01 2.89947033e-01 4.33914393e-01 -4.29145426e-01
2.82328188e-01 -4.78214659e-02 -5.71353771e-02 1.08461976e+00
-1.76958546e-01 9.49489236e-01 5.03318727e-01 -1.16666508e+00
-1.40322959e+00 -1.35997617e+00 -3.37079242e-02 1.18924606e+00
1.06973708e-01 4.67446223e-02 -4.00535524e-01 -7.49737918e-01
3.29368263e-01 8.32713902e-01 -1.09420192e+00 1.02209799e-01
-9.82915819e-01 7.59098306e-02 7.87787974e-01 9.96881306e-01
8.78566086e-01 -1.37545598e+00 -9.50633049e-01 -1.14183135e-01
7.57196248e-01 -9.84883666e-01 -2.97727466e-01 -1.84120938e-01
-5.66321671e-01 -9.28655624e-01 -9.40061629e-01 -8.59587550e-01
7.91721106e-01 4.70482022e-01 1.53407550e+00 2.37971634e-01
-3.10230464e-01 6.45946026e-01 -2.03775410e-02 -5.98189175e-01
-1.37091160e-01 -2.00096533e-01 -3.60584557e-01 3.20265144e-01
8.99542216e-03 -5.87887287e-01 -8.00265491e-01 2.05463290e-01
-1.18046558e+00 3.51750314e-01 8.13849986e-01 7.43417621e-01
1.43015623e-01 -4.31552470e-01 2.67728657e-01 -8.76928270e-01
5.33941269e-01 3.02870497e-02 -2.31862620e-01 6.17857695e-01
4.88279983e-02 -2.82830358e-01 7.65730023e-01 -8.91092300e-01
-1.04484701e+00 1.26171067e-01 1.45810440e-01 -8.16150904e-01
-7.51954496e-01 -4.54393849e-02 -1.95601076e-01 -3.50554794e-01
4.14726406e-01 2.97304451e-01 -3.01625013e-01 -4.68999326e-01
6.67543590e-01 3.56224746e-01 5.99729180e-01 -6.04895353e-01
8.67039382e-01 7.78743088e-01 2.19050556e-01 -8.31192791e-01
-5.66839576e-01 2.87369154e-02 -1.00635505e+00 -5.56427002e-01
5.92827857e-01 -5.90449154e-01 -6.79232538e-01 5.51885307e-01
-1.31006968e+00 -4.77116883e-01 -3.82626057e-01 -3.41073871e-01
-5.42131245e-01 2.44137615e-01 -3.37120920e-01 -9.05044973e-01
-2.89755404e-01 -8.54118764e-01 1.53033412e+00 2.61194795e-01
-1.87465698e-01 -8.12729061e-01 -2.57558435e-01 -3.31176519e-01
6.39194489e-01 7.11197197e-01 1.04863346e+00 -3.44404787e-01
-6.75949395e-01 -2.43997946e-01 -1.01679039e+00 3.02261531e-01
8.86981338e-02 5.49214840e-01 -9.01610911e-01 -1.51830107e-01
-8.11047614e-01 -3.46160948e-01 8.82089913e-01 -6.14890642e-02
1.37537444e+00 -4.82550591e-01 -2.05932349e-01 7.56335139e-01
1.72599137e+00 5.41296780e-01 1.00188398e+00 3.72650713e-01
9.98861432e-01 3.83707076e-01 -9.61541310e-02 1.92919850e-01
3.29972394e-02 7.18248188e-01 3.29428792e-01 -5.66191733e-01
-5.88844538e-01 -2.06458271e-01 -1.74554568e-02 2.20000789e-01
-2.02728838e-01 -5.70584893e-01 -8.54938865e-01 4.68803853e-01
-1.48939741e+00 -1.06174576e+00 -8.38638376e-03 1.42039990e+00
2.95579255e-01 -1.81181598e-02 8.01929459e-02 4.18149419e-02
6.33904397e-01 2.81828910e-01 -6.18779898e-01 -1.00130486e+00
-6.02544427e-01 3.89080882e-01 1.41852871e-01 -2.31836766e-01
-1.38440120e+00 9.75327969e-01 6.90217352e+00 8.44244897e-01
-1.39853644e+00 -2.99144179e-01 5.89503407e-01 -6.62820265e-02
-1.19340666e-01 -4.39184099e-01 -4.23825294e-01 3.14344287e-01
2.89781928e-01 3.15385103e-01 2.44038716e-01 1.09125352e+00
-3.69850367e-01 1.03445090e-01 -1.33705652e+00 1.30999172e+00
4.75782365e-01 -1.77167201e+00 8.23408961e-01 -4.36837077e-01
7.88016915e-01 -5.28681576e-01 5.70325255e-01 2.39607990e-01
3.36957991e-01 -1.20341492e+00 1.06069481e+00 7.13193536e-01
1.33382535e+00 -6.91900074e-01 3.08597744e-01 -3.45449388e-01
-1.26204050e+00 -1.11250579e-01 -2.58266568e-01 -2.09917203e-02
-1.39735956e-02 -1.16822787e-01 -5.06698847e-01 4.32822883e-01
7.65138090e-01 7.71791339e-01 -9.77521956e-01 8.29581976e-01
-1.00751929e-02 3.12289715e-01 -9.86285135e-02 -3.19549948e-01
7.40060627e-01 -1.76132806e-02 3.00820798e-01 1.80473590e+00
1.32784754e-01 1.68465748e-01 -2.26534940e-02 1.10662961e+00
-1.79639369e-01 -2.06082731e-01 -9.35358405e-01 -5.68439588e-02
3.45616043e-01 1.13559937e+00 -1.04292238e+00 -6.51318491e-01
-3.13872159e-01 1.32190156e+00 4.46479023e-01 5.67026973e-01
-5.94067931e-01 -3.63267124e-01 6.00095809e-01 -1.26434341e-01
8.06369662e-01 -5.07182539e-01 -5.17844141e-01 -9.49879169e-01
-1.91040203e-01 -8.34516406e-01 1.81882545e-01 -1.31979358e+00
-1.56886685e+00 7.77251065e-01 2.55409312e-02 -1.33499014e+00
2.88889825e-01 -1.21750069e+00 -1.03053463e+00 5.41885853e-01
-1.33712387e+00 -1.94617677e+00 -5.16906321e-01 4.06358659e-01
7.85686791e-01 -3.16370577e-01 7.58220196e-01 7.94606507e-02
-5.77806294e-01 7.27959275e-01 7.67069012e-02 5.36852181e-01
4.17286336e-01 -1.33207786e+00 8.43267083e-01 4.03049737e-01
3.23065013e-01 6.87961578e-01 3.77997130e-01 -4.32834893e-01
-1.55343568e+00 -1.12469828e+00 3.87851924e-01 -4.74112332e-01
4.73137617e-01 -5.26690900e-01 -7.27858603e-01 6.96329892e-01
5.70030510e-01 8.87326524e-02 4.51919764e-01 -1.24027506e-01
-1.05708313e+00 -8.67746100e-02 -1.13256025e+00 1.00117195e+00
1.45600879e+00 -5.22493124e-01 -5.13592064e-01 3.23649347e-02
1.27743036e-01 -4.18575495e-01 -6.86686456e-01 2.56176263e-01
1.13038027e+00 -1.06309485e+00 1.36343062e+00 -1.13266921e+00
5.36644399e-01 -3.07428986e-01 -1.47609666e-01 -1.08490515e+00
-4.17971700e-01 -2.81472653e-01 8.97993683e-04 1.14496613e+00
-3.96604910e-02 -4.07264501e-01 7.34873116e-01 4.05357808e-01
-2.13790670e-01 -6.03529155e-01 -9.55252767e-01 -9.66236174e-01
5.18207662e-02 4.02703099e-02 8.94857824e-01 9.96915758e-01
-8.56873214e-01 1.40384376e-01 -5.05453229e-01 -1.57746896e-01
3.84922445e-01 5.26181161e-01 1.04147756e+00 -1.15150654e+00
1.66646317e-01 -1.00987411e+00 -7.95914650e-01 -7.88108468e-01
9.52704847e-02 -5.14868319e-01 -3.58501881e-01 -1.66653574e+00
7.83981085e-02 -3.39050978e-01 -2.11796194e-01 5.38870275e-01
3.09850156e-01 6.42921805e-01 7.45226324e-01 1.04940474e-01
-6.46924376e-01 2.95668006e-01 1.34261191e+00 -6.80784881e-01
2.68160999e-01 -4.74170655e-01 -3.33735287e-01 6.17612600e-01
8.66074264e-01 9.10359621e-02 -1.89454734e-01 -7.37145662e-01
-1.07887998e-01 -4.56435174e-01 8.75278354e-01 -1.08244801e+00
-3.81451123e-03 -1.81221589e-01 1.08142984e+00 -9.31918800e-01
6.33088052e-01 -9.53220129e-01 4.58344489e-01 3.42350155e-01
-4.93201196e-01 3.61836135e-01 6.24446869e-01 4.25883740e-01
4.15044948e-02 -4.57656384e-02 9.35847759e-01 -3.45614016e-01
-1.03458321e+00 1.86900869e-01 -3.80280942e-01 -3.77392411e-01
8.23765874e-01 -6.17862165e-01 -7.32811987e-01 -2.29611009e-01
-5.96437514e-01 -5.17791033e-01 5.57754517e-01 6.71644747e-01
9.11792040e-01 -1.75032020e+00 -6.52765334e-01 1.14915073e-01
3.17853391e-01 -4.83430415e-01 5.26307225e-01 4.59127650e-02
-9.35024381e-01 4.49962646e-01 -9.09381568e-01 -5.44673324e-01
-1.30454171e+00 7.36899316e-01 3.48276138e-01 3.41869533e-01
-8.89053643e-01 6.81150377e-01 5.37765443e-01 -2.36809015e-01
1.67225972e-01 -4.42849666e-01 -9.94014144e-02 -6.76834658e-02
5.36706924e-01 2.65016139e-01 -1.23535693e-01 -7.57715940e-01
-2.00408205e-01 8.04771245e-01 -4.38040774e-03 2.60910392e-01
1.17774415e+00 3.39920104e-01 5.13997599e-02 5.75961471e-01
1.24388957e+00 -5.23663104e-01 -1.40804768e+00 -7.58611262e-02
-1.22721240e-01 -5.00973821e-01 -2.66011775e-01 -1.22386456e+00
-1.14290881e+00 9.84041452e-01 6.47410393e-01 2.47019142e-01
1.08515167e+00 -3.20498705e-01 2.39012823e-01 6.37754381e-01
1.86455473e-01 -7.46488750e-01 3.98324132e-01 3.23731482e-01
1.82156312e+00 -1.00301838e+00 2.41651163e-01 -2.31791958e-01
-7.09400237e-01 1.65994716e+00 9.25442636e-01 -6.72673166e-01
4.45283741e-01 1.81542993e-01 1.99948758e-01 -4.69741493e-01
-4.14063573e-01 -5.10196872e-02 8.26470077e-01 7.04423964e-01
1.28571317e-01 2.93080658e-01 2.94290274e-01 1.14818767e-01
-1.46335378e-01 -3.99281830e-01 4.58328635e-01 1.05089915e+00
-1.06891342e-01 -6.85832322e-01 -2.65691042e-01 3.66198391e-01
2.08951384e-01 -1.48366196e-02 -1.13490164e+00 1.37351668e+00
1.76707149e-01 2.94241756e-01 5.41823506e-01 -3.76601182e-02
6.71361566e-01 -7.50709651e-03 7.56209552e-01 -3.05805922e-01
-8.18118811e-01 -3.40322465e-01 3.90387356e-01 -6.02877259e-01
-5.76855481e-01 -4.44906354e-01 -4.32031721e-01 -4.59188670e-01
2.65340030e-01 -6.71068549e-01 7.74720967e-01 4.10573334e-01
3.60602349e-01 6.03856206e-01 3.27051520e-01 -1.39381540e+00
-4.52331193e-02 -7.21970499e-01 -3.43437761e-01 7.39670992e-01
3.91059875e-01 -8.98294806e-01 -8.88314843e-03 2.09673792e-02] | [11.653627395629883, 0.40152353048324585] |
d25f2828-2300-4255-98e5-a30e1ab5394e | follow-the-timeline-generating-abstractive | 2301.00867 | null | https://arxiv.org/abs/2301.00867v1 | https://arxiv.org/pdf/2301.00867v1.pdf | Follow the Timeline! Generating Abstractive and Extractive Timeline Summary in Chronological Order | Nowadays, time-stamped web documents related to a general news query floods spread throughout the Internet, and timeline summarization targets concisely summarizing the evolution trajectory of events along the timeline. Unlike traditional document summarization, timeline summarization needs to model the time series information of the input events and summarize important events in chronological order. To tackle this challenge, in this paper, we propose a Unified Timeline Summarizer (UTS) that can generate abstractive and extractive timeline summaries in time order. Concretely, in the encoder part, we propose a graph-based event encoder that relates multiple events according to their content dependency and learns a global representation of each event. In the decoder part, to ensure the chronological order of the abstractive summary, we propose to extract the feature of event-level attention in its generation process with sequential information remained and use it to simulate the evolutionary attention of the ground truth summary. The event-level attention can also be used to assist in extracting summary, where the extracted summary also comes in time sequence. We augment the previous Chinese large-scale timeline summarization dataset and collect a new English timeline dataset. Extensive experiments conducted on these datasets and on the out-of-domain Timeline 17 dataset show that UTS achieves state-of-the-art performance in terms of both automatic and human evaluations. | ['Rui Yan', 'Xiangliang Zhang', 'Xin Gao', 'Dongyan Zhao', 'Zhangming Chan', 'Shen Gao', 'Mingzhe Li', 'Xiuying Chen'] | 2023-01-02 | null | null | null | null | ['timeline-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.34641963e-01 1.82805851e-01 -2.42302567e-01 -2.55679846e-01
-9.33542311e-01 -4.74358410e-01 8.26903522e-01 7.89796710e-01
-3.18488270e-01 8.99496436e-01 1.34199440e+00 -7.16691241e-02
-1.45952538e-01 -7.67948925e-01 -7.86095262e-01 -1.40604302e-01
-1.99392229e-01 2.99786597e-01 2.90005296e-01 -2.55633324e-01
4.61339533e-01 1.31906018e-01 -1.11020231e+00 4.28861678e-01
1.33412039e+00 5.71842849e-01 3.74307454e-01 9.57789063e-01
-3.72359067e-01 1.03517139e+00 -1.26781595e+00 -4.09900457e-01
-3.52264732e-01 -1.00211048e+00 -7.25899756e-01 1.17820479e-01
2.66415983e-01 -4.34534132e-01 -9.91267681e-01 7.81098485e-01
6.02623701e-01 3.72911423e-01 6.06390834e-01 -1.04371893e+00
-4.84176099e-01 1.23231637e+00 -7.15802252e-01 7.30507910e-01
5.21996498e-01 -7.96866715e-02 1.24185693e+00 -2.11565539e-01
9.24419224e-01 1.07177663e+00 2.83778578e-01 2.28353679e-01
-5.17864406e-01 -3.04897666e-01 5.01668155e-01 3.34243119e-01
-7.20858872e-01 -2.18027905e-01 1.03281522e+00 1.16373017e-01
1.33158779e+00 5.06663859e-01 9.20781374e-01 1.04298556e+00
5.89144349e-01 1.32115436e+00 -4.34116609e-02 -5.31197479e-03
1.27027750e-01 -6.08664095e-01 4.46874052e-01 3.38436604e-01
2.13391811e-01 -4.84901905e-01 -7.87020922e-01 -1.62055776e-01
2.25911558e-01 8.69021844e-03 -2.94829130e-01 5.99817216e-01
-1.53399289e+00 6.66585803e-01 4.83098954e-01 4.54440117e-01
-9.52297747e-01 2.00454846e-01 9.38099325e-01 2.80485213e-01
9.16689873e-01 5.11743367e-01 -2.84128070e-01 -1.70201018e-01
-1.36211884e+00 4.26197439e-01 9.35727119e-01 1.09805822e+00
1.58765450e-01 1.10450514e-01 -8.50614846e-01 4.38998401e-01
-3.04331958e-01 3.59983176e-01 7.51641810e-01 -4.32020754e-01
9.84306633e-01 7.29446471e-01 5.86365312e-02 -1.19602656e+00
-4.64787424e-01 -8.22241187e-01 -1.04857504e+00 -9.45384800e-01
-2.53128260e-01 -3.82509142e-01 -8.94809842e-01 1.63763928e+00
9.30351317e-02 5.93920529e-01 1.59412831e-01 5.73571146e-01
1.13662076e+00 1.56128097e+00 -4.47844639e-02 -9.09347236e-01
1.39022100e+00 -1.12354434e+00 -1.18856668e+00 -3.80246222e-01
4.20797974e-01 -3.89078051e-01 7.09108770e-01 -4.73873615e-02
-1.26621556e+00 -4.74209368e-01 -1.26410508e+00 -3.02330345e-01
-1.43368259e-01 5.83877228e-02 3.47570986e-01 -2.57338494e-01
-6.17973983e-01 6.72326088e-01 -9.31636274e-01 -5.96074760e-01
1.92049965e-01 -2.50292361e-01 -5.63384295e-02 2.75655419e-01
-1.40933251e+00 6.71931982e-01 1.05464113e+00 -1.06186777e-01
-5.26919127e-01 -9.02365446e-01 -8.63288462e-01 4.65973020e-01
4.88304108e-01 -1.02516305e+00 1.37616181e+00 -3.49724829e-01
-1.10084951e+00 2.14155808e-01 -4.35486048e-01 -9.28882003e-01
4.58745033e-01 -2.74615318e-01 -6.09551668e-01 4.87545997e-01
3.51637155e-01 2.88954645e-01 5.39105117e-01 -6.10184669e-01
-9.19579208e-01 -1.11400977e-01 -2.11927712e-01 3.72088820e-01
-3.64626527e-01 5.57703972e-02 -7.03908086e-01 -1.04212964e+00
-1.59869045e-01 -2.75952458e-01 -2.85888284e-01 -9.18710113e-01
-9.39018428e-01 -4.36414331e-01 7.38920212e-01 -1.22071052e+00
2.08711505e+00 -1.88666451e+00 4.46955085e-01 -3.72336090e-01
1.95950806e-01 2.32887547e-02 -1.43695876e-01 1.13085985e+00
1.33440301e-01 5.21850884e-02 -5.41582882e-01 -3.53899688e-01
4.88673858e-02 -1.35867134e-01 -8.56898248e-01 8.05299431e-02
2.60441035e-01 1.23615539e+00 -1.36179245e+00 -6.81914032e-01
-1.69345900e-01 8.90163258e-02 -3.34250361e-01 3.93029809e-01
-6.38252497e-01 3.73248398e-01 -7.54741549e-01 1.36875764e-01
1.86449721e-01 -3.24972242e-01 -1.53100371e-01 -4.10565704e-01
-1.72666386e-01 7.07512200e-01 -6.41296506e-01 1.95632470e+00
-2.62584269e-01 7.49071360e-01 -8.15254629e-01 -7.08486617e-01
7.62197971e-01 3.77823621e-01 5.52142560e-01 -7.84780562e-01
-9.94011760e-03 1.13614924e-01 -3.55472475e-01 -4.74206418e-01
1.20955288e+00 3.91085476e-01 -6.10714495e-01 8.15576494e-01
2.11753160e-01 -1.81319281e-01 7.36784637e-01 7.75245309e-01
1.20414138e+00 -1.83817074e-01 7.09205747e-01 2.17942566e-01
3.38631749e-01 2.06121698e-01 2.69105405e-01 7.34795272e-01
4.10803407e-01 7.11302161e-01 8.67036879e-01 -3.03293854e-01
-1.21413457e+00 -6.93791986e-01 5.44039845e-01 8.14279258e-01
1.01213008e-01 -9.64416385e-01 -5.54821670e-01 -8.43494415e-01
-2.90868610e-01 1.43649411e+00 -5.00451088e-01 -4.42133904e-01
-8.74343753e-01 -7.09948003e-01 4.70831215e-01 5.17795801e-01
6.29724383e-01 -1.22873354e+00 -7.99198627e-01 8.28056872e-01
-7.47996032e-01 -9.82688785e-01 -1.01133800e+00 -1.57613292e-01
-7.17215478e-01 -7.34522641e-01 -8.78696024e-01 -6.11649632e-01
3.59006047e-01 1.74112990e-01 1.04245055e+00 -1.15864255e-01
1.54677317e-01 1.81472987e-01 -5.08200884e-01 -7.02407360e-01
-3.66808981e-01 6.49279892e-01 -4.76140290e-01 7.47446641e-02
-1.81567669e-02 -6.53359056e-01 -5.48224986e-01 -4.20699865e-01
-1.19383442e+00 4.61271554e-01 3.56160879e-01 6.52673066e-01
4.04044747e-01 2.65175670e-01 1.21912575e+00 -8.74419510e-01
1.14418626e+00 -9.00945365e-01 -2.26990074e-01 4.82472122e-01
-9.09980312e-02 1.06086992e-01 8.77178431e-01 -3.40896308e-01
-1.17037952e+00 -5.38239300e-01 -4.39324193e-02 -2.46406365e-02
4.33537722e-01 1.23874032e+00 1.62464023e-01 1.21849120e+00
4.21301484e-01 8.08525562e-01 -6.39677763e-01 -3.47996742e-01
4.92090046e-01 6.70185149e-01 8.61717343e-01 -2.15820804e-01
5.14727414e-01 3.12287301e-01 -3.33939433e-01 -8.24503005e-01
-1.10795224e+00 -4.45854664e-01 -2.38173351e-01 -3.56870860e-01
6.85492635e-01 -8.14835906e-01 -1.48770183e-01 3.17595512e-01
-1.60705459e+00 5.09707741e-02 -5.04884720e-01 4.22642916e-01
-5.37805736e-01 5.72530687e-01 -5.62472105e-01 -5.18414140e-01
-9.42833602e-01 -4.14041609e-01 1.27042699e+00 5.36973953e-01
-4.57684726e-01 -9.26023424e-01 1.49215356e-01 -3.86313677e-01
2.08035186e-01 6.41500592e-01 9.32909369e-01 -1.18109500e+00
-4.33738291e-01 -4.77411419e-01 -7.93095753e-02 -2.70114988e-01
3.31045091e-01 1.65214837e-02 -3.49585652e-01 -1.37164220e-01
1.14317194e-01 1.15786538e-01 1.29000640e+00 5.40062845e-01
9.64217544e-01 -8.76532435e-01 -3.70502472e-01 2.82043427e-01
1.09046280e+00 2.81078786e-01 5.29392302e-01 1.17686011e-01
7.37797022e-01 6.28099144e-01 6.80627763e-01 6.75921559e-01
6.22434318e-01 2.29267687e-01 7.15838885e-03 2.50584871e-01
-2.64001667e-01 -7.58853376e-01 3.64882886e-01 1.42907524e+00
2.27418616e-01 -9.80630279e-01 -5.28268754e-01 7.91833699e-01
-2.25592184e+00 -1.42869997e+00 -8.02821442e-02 1.87127757e+00
8.68370175e-01 2.49500751e-01 1.53957173e-01 -2.65003741e-02
8.49686027e-01 8.75110805e-01 -7.35027134e-01 -1.90347791e-01
-2.21102551e-01 -2.88733423e-01 2.35089317e-01 4.01140600e-01
-8.14795077e-01 8.84345829e-01 5.28357029e+00 9.25790429e-01
-1.02473652e+00 -1.30007967e-01 6.27622783e-01 -1.82325795e-01
-7.25766063e-01 -2.22230241e-01 -5.55990458e-01 6.55365109e-01
1.27532995e+00 -1.32180071e+00 1.04995184e-01 3.67919981e-01
4.16537881e-01 6.85069561e-02 -1.05743432e+00 6.31022155e-01
2.76760548e-01 -1.68146539e+00 5.04472613e-01 -4.88143414e-01
8.16539347e-01 -8.37873816e-02 -5.20509362e-01 4.51272607e-01
-5.75137045e-03 -4.64420289e-01 7.75641024e-01 7.14848995e-01
5.89814842e-01 -9.21366274e-01 6.45858765e-01 5.64089119e-01
-1.42346060e+00 1.85364559e-01 -1.84206277e-01 2.16088772e-01
8.19135070e-01 7.05064476e-01 -9.15885270e-01 1.35079861e+00
1.34440243e-01 1.40070236e+00 -5.20066798e-01 1.26371765e+00
-4.11221951e-01 8.77866387e-01 -2.93785155e-01 -3.46871793e-01
4.35990036e-01 -7.62648135e-02 1.13555360e+00 1.48875701e+00
5.69304645e-01 2.60940105e-01 -3.42207998e-02 4.61946875e-01
-2.47118443e-01 6.75594211e-02 -4.64638591e-01 -5.07422686e-01
5.36279619e-01 9.10890460e-01 -9.13755178e-01 -8.34650636e-01
-1.06003270e-01 1.17495656e+00 1.70530468e-01 4.17514473e-01
-1.06122029e+00 -8.76686692e-01 -1.40556395e-01 -2.63765275e-01
3.25168878e-01 -1.42631307e-01 -1.00998301e-03 -1.42794955e+00
1.92738801e-01 -5.83788812e-01 8.15766454e-01 -1.02994180e+00
-1.02861345e+00 8.65954995e-01 2.27634147e-01 -1.22313726e+00
-5.32907367e-01 5.14136076e-01 -1.15973473e+00 6.34041488e-01
-1.43565106e+00 -9.98551786e-01 -1.75429285e-01 1.79608166e-01
1.22059464e+00 4.81286505e-03 1.81296661e-01 7.73781911e-02
-5.81004143e-01 1.11148879e-01 -1.07428126e-01 -1.14876941e-01
5.28263152e-01 -1.22110474e+00 1.16235960e+00 1.26580143e+00
6.79948181e-02 4.77799714e-01 1.08581078e+00 -1.24836361e+00
-1.40186048e+00 -1.48892319e+00 1.51618183e+00 1.88262034e-02
7.92674601e-01 -4.87154961e-04 -9.93145585e-01 7.95508921e-01
7.68660247e-01 -8.77354145e-01 1.63395956e-01 -3.85620624e-01
1.07346093e-02 -5.78225665e-02 -6.40406132e-01 9.32579756e-01
1.18346596e+00 -3.83508652e-01 -1.31559420e+00 6.06994510e-01
1.53394020e+00 -5.74879169e-01 -4.80787605e-01 1.84679687e-01
1.84992820e-01 -4.01334554e-01 5.86438119e-01 -7.56445348e-01
8.93572867e-01 -2.73722142e-01 2.67982304e-01 -1.76709247e+00
-3.69277686e-01 -1.09854841e+00 -6.22694314e-01 1.55213392e+00
2.71712810e-01 -5.07290423e-01 3.30797672e-01 -1.75028369e-01
-7.05776215e-01 -4.62339669e-01 -6.64202213e-01 -5.06308854e-01
-5.05909562e-01 -1.61796376e-01 8.68282378e-01 5.99878788e-01
2.26000994e-01 8.48120749e-01 -5.73925078e-01 7.58634135e-02
3.75260174e-01 4.44466919e-01 5.74728787e-01 -9.20196891e-01
6.49491102e-02 -4.66305315e-01 -3.01183406e-02 -1.15098131e+00
1.55653208e-01 -8.91024351e-01 -7.84665942e-02 -2.26242614e+00
2.72446573e-01 5.62370658e-01 -1.39424145e-01 3.54888180e-04
-6.95595980e-01 -5.95499992e-01 1.05182938e-01 1.37243783e-02
-9.67900097e-01 9.72930014e-01 1.26828420e+00 -4.41558570e-01
-3.46632421e-01 -2.85256631e-03 -8.43204558e-01 2.66699255e-01
6.74366951e-01 -6.18454218e-01 -7.23568380e-01 -4.11167651e-01
3.76442969e-01 7.66047060e-01 -4.09846790e-02 -9.23315585e-01
8.02735507e-01 -1.92491859e-01 -3.01969529e-04 -1.37857246e+00
1.88218653e-02 -4.19911921e-01 2.35760868e-01 5.69666624e-01
-7.92302370e-01 2.99159944e-01 1.53820112e-01 8.69194627e-01
-5.13011634e-01 -8.30037743e-02 5.67178205e-02 -1.39087066e-01
-5.67974985e-01 4.84719932e-01 -1.86012954e-01 3.03180724e-01
7.23774374e-01 1.20556004e-01 -6.89217448e-01 -7.56695271e-01
-1.45679444e-01 6.62395000e-01 -6.33527860e-02 4.80100781e-01
5.81065834e-01 -1.29320538e+00 -1.38943672e+00 -2.94942826e-01
1.28842041e-01 5.90378121e-02 5.30626714e-01 6.15352869e-01
-4.87798721e-01 5.62155545e-01 -9.22807232e-02 -2.38573462e-01
-7.58494079e-01 6.85928583e-01 -1.51859462e-01 -5.98484576e-01
-1.00919223e+00 4.49993938e-01 5.06787822e-02 2.93616503e-01
1.38620988e-01 -6.64671421e-01 -5.32659411e-01 6.41593933e-01
1.06471789e+00 3.85690838e-01 9.15791169e-02 -2.02740610e-01
6.91339001e-02 1.85004801e-01 -2.80772984e-01 -1.53134286e-01
1.62099659e+00 -2.41450533e-01 -2.22536713e-01 6.65277600e-01
1.17373753e+00 1.66788459e-01 -1.03663766e+00 -1.92530170e-01
3.19715500e-01 6.47390541e-03 -2.04052716e-01 -5.43328822e-01
-7.77653754e-01 3.75193506e-01 -6.34901583e-01 6.45038366e-01
1.29787040e+00 9.10379961e-02 1.38690293e+00 3.31989378e-01
-1.05691038e-01 -7.91209102e-01 5.90033643e-02 7.39130497e-01
1.29159713e+00 -6.45938456e-01 2.55604267e-01 -2.11011305e-01
-8.57897222e-01 1.18264496e+00 2.44506776e-01 -1.68815941e-01
1.30147204e-01 -1.49221659e-01 -6.81974411e-01 -2.92773783e-01
-1.14185393e+00 -1.01834543e-01 5.17151237e-01 5.02853729e-02
2.06151471e-01 -1.91111743e-01 -7.67761827e-01 8.75244796e-01
-4.89088595e-01 -1.15395963e-01 7.37986147e-01 8.19565415e-01
-5.77453375e-01 -3.89450878e-01 7.67024085e-02 7.22834527e-01
-4.02187258e-01 -6.52138665e-02 -3.47943693e-01 3.76272559e-01
-7.00261295e-01 9.60039675e-01 1.71594650e-01 -2.20379531e-01
6.06706381e-01 -8.62106401e-03 3.11525445e-02 -5.35622299e-01
-8.92506778e-01 1.71954408e-01 5.24982870e-01 -2.70589381e-01
-9.53902453e-02 -6.55659556e-01 -1.40593171e+00 -2.06099913e-01
1.68498012e-03 5.20388246e-01 5.07614315e-01 7.87644923e-01
5.82389772e-01 1.26181448e+00 7.77579784e-01 -7.38153875e-01
-2.97529966e-01 -1.19873178e+00 -2.79092222e-01 3.15775841e-01
6.07077956e-01 1.21786356e-01 -2.24722534e-01 1.20499067e-01] | [12.558239936828613, 9.504176139831543] |
dad60fb9-97a8-4fb0-a669-f3250f0d947f | predictive-compliance-monitoring-in-process | 2205.05446 | null | https://arxiv.org/abs/2205.05446v3 | https://arxiv.org/pdf/2205.05446v3.pdf | Predictive Compliance Monitoring in Process-Aware Information Systems: State of the Art, Functionalities, Research Directions | Business process compliance is a key area of business process management and aims at ensuring that processes obey to compliance constraints such as regulatory constraints or business rules imposed on them. Process compliance can be checked during process design time based on verification of process models and at runtime based on monitoring the compliance states of running process instances. For existing compliance monitoring approaches it remains unclear whether and how compliance violations can be predicted, although predictions are crucial in order to prepare and take countermeasures in time. This work, hence, analyzes existing literature from compliance monitoring as well as predictive process monitoring and provides an updated framework of compliance monitoring functionalities. Moreover, it raises the vision of a comprehensive predictive compliance monitoring system that integrates existing predicate prediction approaches with the idea of employing PPM with different prediction goals such as next activity or remaining time for prediction and subsequent mapping of the prediction results onto the given set of compliance constraints (PCM). For each compliance monitoring functionality we elicit PCM system requirements and assess their coverage by existing approaches. Based on the assessment, open challenges and research directions realizing a comprehensive PCM system are elaborated. | ['Janik-Vasily Benzin', 'Karolin Winter', 'Stefanie Rinderle-Ma'] | 2022-05-10 | null | null | null | null | ['predictive-process-monitoring'] | ['time-series'] | [ 1.00016177e+00 5.58646798e-01 -3.83611888e-01 -5.03200650e-01
-1.07440785e-01 -5.20365834e-01 9.72283483e-01 7.36160278e-01
1.39131129e-01 2.09378108e-01 3.45724165e-01 -4.32104051e-01
-8.86812925e-01 -1.13742304e+00 1.99038088e-02 1.65785551e-01
2.77884662e-01 8.96971583e-01 4.71860558e-01 1.74781516e-01
5.84883094e-01 5.94606042e-01 -1.53232872e+00 5.85344553e-01
4.95158821e-01 8.42239141e-01 -1.98961809e-01 4.51622576e-01
-4.03555244e-01 1.07044232e+00 -2.31921822e-01 -1.81637883e-01
1.27151594e-01 -2.99467355e-01 -1.05637217e+00 2.36770093e-01
-1.67503208e-01 -4.35670950e-02 2.62502939e-01 8.51270854e-01
-5.14478445e-01 -1.09107107e-01 3.36694807e-01 -1.55823398e+00
-3.04052204e-01 8.46163452e-01 -8.51860940e-02 -1.79471239e-01
8.04074705e-01 5.05665958e-01 1.07855308e+00 -1.92327082e-01
6.68883920e-01 9.06469166e-01 5.47771931e-01 7.63200164e-01
-1.37598050e+00 -3.07895631e-01 3.79911840e-01 2.86422968e-01
-8.22465420e-01 -3.26421946e-01 3.47606808e-01 -8.54704976e-01
1.25096428e+00 7.27077603e-01 6.59121275e-01 7.29256630e-01
3.20516646e-01 2.68249899e-01 1.10524356e+00 -6.42764986e-01
5.11037767e-01 1.64902225e-01 5.54320872e-01 9.96345431e-02
8.27626765e-01 -8.30849074e-03 -4.58764613e-01 -3.97266895e-01
4.86742407e-01 1.15648128e-01 -9.70712304e-02 -3.14269602e-01
-8.24397266e-01 2.54344106e-01 -7.22573996e-01 8.74769270e-01
-5.51266313e-01 1.36897294e-02 3.67381662e-01 2.62253135e-01
-2.80660242e-01 4.85059857e-01 -7.78460801e-01 -4.27522898e-01
-8.37209404e-01 3.80585790e-01 1.52391100e+00 1.21714973e+00
5.62622309e-01 -3.15477669e-01 -7.36344695e-01 -2.13164404e-01
8.06842566e-01 2.24738829e-02 -1.24281282e-02 -1.06194079e+00
2.78060228e-01 1.31431973e+00 5.86939096e-01 -6.95728958e-01
-1.95518821e-01 2.50669032e-01 -2.81946748e-01 -5.21716885e-02
3.41751873e-01 3.83977413e-01 -2.66252488e-01 1.20976162e+00
4.31887358e-02 2.58030184e-02 3.64490822e-02 1.92399770e-01
3.79335918e-02 3.38108033e-01 6.61857665e-01 -9.26728845e-01
1.47447193e+00 -5.18594921e-01 -1.08158600e+00 -1.99147701e-01
4.64211106e-01 -4.05617714e-01 6.58470750e-01 4.92567271e-01
-1.08008254e+00 -4.64358240e-01 -6.93437994e-01 6.63689017e-01
-6.86403289e-02 -4.14705306e-01 5.92011273e-01 1.04380786e+00
-8.03334653e-01 9.69434798e-01 -1.01811874e+00 -7.02969134e-01
1.68681860e-01 5.60810208e-01 -1.48189142e-01 5.66116348e-02
-8.46484244e-01 1.05355263e+00 5.10130584e-01 -6.38998076e-02
-4.12017673e-01 -8.48053694e-01 -6.96568727e-01 4.32178736e-01
5.35391986e-01 -6.01309776e-01 1.57145643e+00 -5.98964632e-01
-1.49182928e+00 6.38419628e-01 -2.29971305e-01 -7.17779160e-01
1.12185466e+00 -2.63618052e-01 -9.05488491e-01 -2.91386873e-01
-1.74704760e-01 -2.62316704e-01 2.55763620e-01 -1.02707624e+00
-1.05950952e+00 -4.49462026e-01 -1.98063016e-01 -3.79456550e-01
-3.14193405e-02 5.68471372e-01 -3.04795176e-01 5.80333769e-01
2.74014652e-01 -5.31065404e-01 -5.56630194e-01 -4.32206333e-01
-4.29728985e-01 -8.25028494e-02 7.25031793e-01 -5.56480289e-01
1.68596017e+00 -1.44706428e+00 -4.23756570e-01 8.97364557e-01
-1.25946060e-01 2.17883363e-01 2.41727099e-01 1.02718508e+00
-1.47855449e-02 5.46456337e-01 -8.99299756e-02 1.38968170e-01
6.47946894e-01 2.11549520e-01 -3.53209376e-01 1.54069081e-01
2.74093598e-01 6.76255405e-01 -5.43083191e-01 -3.42622548e-01
8.45505893e-01 -5.87890223e-02 -2.34541327e-01 3.84817183e-01
-7.71397650e-01 9.02183115e-01 -5.25265038e-01 7.48565674e-01
5.31360984e-01 -1.12621866e-01 1.12067664e+00 2.31659681e-01
-3.92690718e-01 2.88048208e-01 -1.47278571e+00 7.71227539e-01
-1.76083133e-01 -2.61539519e-01 -1.42053068e-01 -6.72151387e-01
1.32900500e+00 8.41012955e-01 1.07567704e+00 -6.10710740e-01
-2.19929770e-01 8.93047377e-02 -4.51436676e-02 -6.55111432e-01
2.01240003e-01 1.59808025e-01 -3.03390939e-02 6.16447508e-01
-5.65312624e-01 7.02487677e-02 6.02797866e-01 -6.39948070e-01
1.60196352e+00 6.51646435e-01 1.02592051e+00 -1.68483362e-01
1.30081916e+00 1.39139205e-01 8.86288643e-01 6.17224932e-01
-4.01830733e-01 -3.62465113e-01 8.74331594e-01 -6.63994908e-01
-9.44520533e-01 -5.14805377e-01 8.62312019e-02 5.70880830e-01
-1.59124210e-01 -8.87659013e-01 -5.11466742e-01 -6.66875482e-01
4.00423910e-03 9.74742413e-01 -3.01732898e-01 1.30445147e-02
-7.16552377e-01 -3.12793404e-01 1.21625990e-01 5.20674825e-01
2.11614281e-01 -1.20698357e+00 -9.12402451e-01 6.51457548e-01
2.61826336e-01 -1.27486408e+00 3.11176091e-01 1.58974889e-03
-9.74385798e-01 -1.72444141e+00 9.10107613e-01 -2.17590675e-01
4.25841361e-01 -4.99788493e-01 1.12098300e+00 1.08961046e-01
2.23117977e-01 4.71706539e-01 -9.06154066e-02 -4.85136122e-01
-9.76268232e-01 -3.31666261e-01 -1.85687050e-01 -1.57620326e-01
1.09405875e+00 -5.17430484e-01 -1.79738760e-01 5.00582755e-01
-6.38261318e-01 -4.94424328e-02 3.27717215e-01 3.81564200e-02
7.58996010e-01 3.35948914e-01 3.90838057e-01 -1.27445841e+00
8.65993321e-01 -4.00513262e-02 -1.01503181e+00 7.63429642e-01
-1.38785231e+00 -8.91729891e-02 3.21619481e-01 4.87385057e-02
-1.19685411e+00 1.98872581e-01 7.77388215e-02 1.20067522e-01
-7.07418025e-01 3.32902044e-01 -6.74853086e-01 5.20217419e-01
1.69499248e-01 1.05148099e-01 -3.95983130e-01 -2.62295842e-01
-4.28003408e-02 2.34641135e-01 6.76407665e-02 -8.94631207e-01
8.55031669e-01 4.71252233e-01 3.34208876e-01 -1.07489683e-01
-4.69353616e-01 -6.82504416e-01 -8.83268714e-01 -4.76880699e-01
8.65850449e-01 -7.30327517e-02 -1.41081965e+00 -2.99776673e-01
-1.17603099e+00 -1.45542637e-01 -6.91641331e-01 -7.73617160e-03
-9.17536914e-01 3.68106604e-01 -6.30361259e-01 -1.46083343e+00
-5.49983025e-01 -8.66819143e-01 6.49700820e-01 -5.22076264e-02
-1.29166389e+00 -8.76381993e-01 2.87250310e-01 5.31185031e-01
4.69841689e-01 2.81004161e-01 1.04377949e+00 -1.21388364e+00
-8.91138315e-01 -4.68859375e-01 1.64910540e-01 2.01235935e-01
6.42259657e-01 4.04610664e-01 -6.69836938e-01 1.64374560e-01
1.77206919e-01 7.09192038e-01 -2.26075321e-01 1.54466733e-01
1.10235858e+00 -5.34581482e-01 -4.36193734e-01 -1.82170510e-01
1.68244040e+00 6.22295320e-01 9.18561578e-01 6.63345397e-01
1.20610692e-01 1.13089406e+00 9.60683167e-01 6.14789069e-01
-7.79596120e-02 5.99037409e-01 2.61167109e-01 7.28171110e-01
5.90670168e-01 -2.39445835e-01 2.02383012e-01 1.85562953e-01
-5.86272001e-01 2.32781127e-01 -1.31028879e+00 7.73281902e-02
-2.29222989e+00 -1.57851374e+00 -9.09684896e-01 2.28446293e+00
5.97224593e-01 3.66105199e-01 -1.52549446e-01 5.86089313e-01
6.94232285e-01 -2.35015243e-01 -9.31744836e-03 -7.18219578e-01
5.04206717e-01 2.58419424e-01 3.98721009e-01 5.73345482e-01
-9.99074638e-01 6.11033261e-01 5.90827322e+00 -1.13855354e-01
-3.53299946e-01 1.06573924e-01 5.65149263e-02 4.06219423e-01
-2.72007644e-01 5.72961271e-01 -1.05026019e+00 1.85799852e-01
1.45539403e+00 -3.97641331e-01 1.17675230e-01 9.04884696e-01
1.14444995e+00 1.19155459e-01 -1.81124341e+00 3.49766463e-01
-5.62620759e-01 -1.47474551e+00 -1.30324602e-01 3.19936961e-01
7.70956099e-01 -6.30173743e-01 -7.27516055e-01 1.20501287e-01
5.69406509e-01 -9.29113925e-01 6.91788137e-01 1.07167721e+00
5.93213588e-02 -4.28727627e-01 7.96140790e-01 2.22186550e-01
-1.26133168e+00 -4.79230553e-01 5.94155937e-02 -1.28258362e-01
3.26049387e-01 4.41411197e-01 -9.80876982e-01 9.19814229e-01
3.75059605e-01 5.47276556e-01 -1.06546946e-01 8.92801166e-01
-6.06555283e-01 5.13330877e-01 1.82509348e-01 2.31457368e-01
-3.29272807e-01 -6.07017398e-01 2.50206023e-01 1.26409411e+00
3.93778414e-01 -4.16930079e-01 3.79823834e-01 1.19907367e+00
7.60221303e-01 7.02643022e-02 -3.75455379e-01 -1.08105540e-01
5.15118480e-01 1.03630531e+00 -3.82521302e-01 -2.66238779e-01
-5.98379374e-01 3.33574086e-01 -3.63109112e-01 1.12868749e-01
-6.38392091e-01 6.63067698e-01 7.61950195e-01 7.24762380e-01
7.85623193e-02 1.01449648e-02 -7.74562538e-01 -5.78209758e-01
2.12421671e-01 -8.01757395e-01 5.55740297e-01 -3.15363586e-01
-1.13846350e+00 8.06139708e-02 -1.44672707e-01 -1.23363972e+00
-2.26475462e-01 -4.67198998e-01 -7.36270785e-01 9.27878737e-01
-1.53641975e+00 -1.40974152e+00 -4.14607525e-01 4.27120805e-01
5.65993674e-02 2.48958245e-01 1.19028938e+00 -9.24857333e-03
-4.24479157e-01 -5.92109740e-01 -6.74475670e-01 -3.76679122e-01
4.88695174e-01 -1.22714686e+00 -6.49564639e-02 1.00304544e+00
-4.37525719e-01 8.85156870e-01 8.13185632e-01 -1.23865700e+00
-1.40002704e+00 -1.45393157e+00 1.48245597e+00 -6.74330354e-01
9.42442656e-01 1.92459822e-01 -1.05237722e+00 1.23542082e+00
7.56094381e-02 -4.20650810e-01 1.02001059e+00 4.35108170e-02
-1.22215763e-01 -3.71496469e-01 -1.29606676e+00 4.65398401e-01
1.05222082e+00 -2.31358379e-01 -7.14905560e-01 2.47214854e-01
3.69924515e-01 1.25494167e-01 -1.50076210e+00 5.80791414e-01
2.77893841e-01 -1.11721122e+00 4.35550719e-01 -1.05693007e+00
2.85628706e-01 -5.73440373e-01 -2.75940329e-01 -1.93598554e-01
-3.59295487e-01 -8.30911934e-01 -6.99070454e-01 1.52409172e+00
2.06237078e-01 -4.77699429e-01 1.07001591e+00 1.60780752e+00
-4.54166904e-03 -4.77063537e-01 -7.31486380e-01 -6.76011980e-01
-5.81817389e-01 -8.67607236e-01 1.14744401e+00 9.95194912e-01
4.95938092e-01 -1.37635469e-01 -1.15038045e-01 4.10715848e-01
5.13975203e-01 1.82387665e-01 7.84617007e-01 -1.83210313e+00
-3.05779457e-01 -5.24077415e-01 -4.12998438e-01 3.13879885e-02
-1.65143088e-01 -4.40048993e-01 -3.17811102e-01 -2.24671578e+00
6.60259649e-02 -9.13849920e-02 -2.63823897e-01 7.08508909e-01
4.24091876e-01 -6.68472171e-01 1.13212459e-01 3.67262393e-01
-5.71199536e-01 -1.41335562e-01 1.02991986e+00 -5.29105291e-02
-5.35927236e-01 5.58402896e-01 -7.12776601e-01 7.14565694e-01
8.36889982e-01 -2.08277553e-01 -3.64185840e-01 4.73803967e-01
5.19231379e-01 3.26892883e-01 8.59159380e-02 -1.18036199e+00
4.98690099e-01 -9.91257071e-01 -6.94818646e-02 -3.54765773e-01
-1.13775559e-01 -1.62247896e+00 1.04158330e+00 1.05684245e+00
-4.36202586e-01 -1.54858917e-01 -2.68080443e-01 6.88434482e-01
-2.19213679e-01 -4.51313078e-01 5.61956167e-01 -1.21609867e-01
-8.95125985e-01 1.97614580e-01 -6.43539131e-01 -7.05628157e-01
1.74725175e+00 -7.12013006e-01 -4.60424542e-01 2.24862937e-02
-1.12060606e+00 2.20597312e-01 3.38789612e-01 2.38011926e-01
2.96144366e-01 -7.84347117e-01 -4.42931384e-01 1.57829896e-01
3.96815091e-01 -2.75166035e-01 5.97340353e-02 1.01577282e+00
-2.68618047e-01 7.21937656e-01 -1.65300459e-01 -4.80948359e-01
-1.48160982e+00 6.69003785e-01 5.42265296e-01 -1.05463326e+00
-5.05231798e-01 -3.80512595e-01 -5.36912739e-01 -3.85523468e-01
6.18525185e-02 -5.99671364e-01 -7.32010901e-01 -3.84127110e-01
7.16994941e-01 2.81178445e-01 2.45413005e-01 -5.09160869e-02
-5.15275836e-01 4.19933200e-01 2.95741171e-01 -2.05152743e-02
1.41813433e+00 -2.51822881e-02 -5.24325907e-01 2.06482962e-01
-1.91539690e-01 2.05579311e-01 -1.04605567e+00 -2.63658911e-01
1.33132839e+00 -5.83994806e-01 -5.64355254e-01 -1.00052774e+00
-4.41223085e-01 2.87372440e-01 2.76683658e-01 4.62822318e-01
8.35759401e-01 -7.39611536e-02 -4.27337876e-03 2.45450884e-01
4.77392524e-01 -1.38553202e+00 -2.27385506e-01 4.25594032e-01
7.09673584e-01 -5.82519650e-01 7.00496882e-02 -1.13385344e+00
-5.77660739e-01 1.15800500e+00 7.93950140e-01 4.28242207e-01
5.11516571e-01 7.07806051e-01 -4.63935167e-01 -2.90971249e-01
-9.83759642e-01 9.66605544e-02 -1.31386250e-01 9.95160341e-01
4.19163495e-01 4.06698197e-01 -5.70291936e-01 1.00764966e+00
1.63755968e-01 4.60684955e-01 5.52748978e-01 1.30065322e+00
-5.44891834e-01 -1.97536337e+00 -4.50966001e-01 5.54914951e-01
-4.12239492e-01 3.62800896e-01 -4.96711373e-01 7.80896902e-01
2.59761125e-01 1.31913853e+00 -2.55693272e-02 -2.15641648e-01
1.02873266e+00 4.90281135e-01 3.32274318e-01 -9.36156988e-01
-1.05393708e+00 -8.99190828e-02 8.29576910e-01 -7.27894008e-01
-5.83937645e-01 -1.19662392e+00 -1.23433673e+00 -3.73085797e-01
1.41444996e-01 5.70021905e-02 4.38610226e-01 1.24907625e+00
-2.32589617e-02 8.34305227e-01 1.12269320e-01 2.60025442e-01
-7.70720124e-01 -7.69598842e-01 -4.38780665e-01 6.15172565e-01
-6.26096308e-01 7.98047557e-02 2.73243994e-01 4.66873646e-01] | [8.599291801452637, 6.074399948120117] |
eba872eb-d2c6-4895-8340-a0afa1b2f3f1 | hypertree-proof-search-for-neural-theorem | 2205.11491 | null | https://arxiv.org/abs/2205.11491v1 | https://arxiv.org/pdf/2205.11491v1.pdf | HyperTree Proof Search for Neural Theorem Proving | We propose an online training procedure for a transformer-based automated theorem prover. Our approach leverages a new search algorithm, HyperTree Proof Search (HTPS), inspired by the recent success of AlphaZero. Our model learns from previous proof searches through online training, allowing it to generalize to domains far from the training distribution. We report detailed ablations of our pipeline's main components by studying performance on three environments of increasing complexity. In particular, we show that with HTPS alone, a model trained on annotated proofs manages to prove 65.4% of a held-out set of Metamath theorems, significantly outperforming the previous state of the art of 56.5% by GPT-f. Online training on these unproved theorems increases accuracy to 82.6%. With a similar computational budget, we improve the state of the art on the Lean-based miniF2F-curriculum dataset from 31% to 42% proving accuracy. | ['Timothée Lacroix', 'Aurélien Rodriguez', 'Gabriel Ebner', 'Amaury Hayat', 'Xavier Martinet', 'Thibaut Lavril', 'Marie-Anne Lachaux', 'Guillaume Lample'] | 2022-05-23 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [ 1.41254038e-01 5.73288262e-01 -6.15367532e-01 -9.26030800e-03
-1.18206489e+00 -1.07761323e+00 5.08253992e-01 1.79061815e-01
2.59769522e-02 1.18609679e+00 -3.82868528e-01 -1.34655821e+00
-3.45306933e-01 -9.11822855e-01 -1.43055928e+00 1.34844020e-01
-4.30620968e-01 7.55557060e-01 5.69990158e-01 -1.68239146e-01
2.73815244e-01 -1.71276703e-01 -1.35284388e+00 7.08969057e-01
1.22944295e+00 5.67209482e-01 -1.31473422e-01 9.88323748e-01
-1.07949905e-01 1.32141364e+00 -5.58243871e-01 -7.92244911e-01
4.74023484e-02 -1.38759300e-01 -1.43758166e+00 -9.26300585e-01
9.47181761e-01 -7.90665686e-01 -4.00838941e-01 1.07042217e+00
-7.14273378e-02 -4.40239102e-01 3.68768901e-01 -1.76311707e+00
-4.03377920e-01 1.56256700e+00 -1.38449565e-01 1.04026049e-01
6.33002520e-01 -1.27827302e-01 1.43744898e+00 -2.70061284e-01
1.05700028e+00 1.04988492e+00 8.30274463e-01 5.84667802e-01
-1.42471850e+00 -1.00316632e+00 -1.02668725e-01 7.43818760e-01
-1.09353888e+00 -3.13809365e-01 4.02728587e-01 -3.20406854e-01
1.42767966e+00 2.29105875e-01 4.67116117e-01 1.06539369e+00
2.44201109e-01 1.07046318e+00 8.91191125e-01 -5.94518781e-01
4.47055921e-02 7.65018389e-02 1.07591495e-01 1.21813262e+00
5.25095880e-01 -3.51995532e-03 -6.05547905e-01 -3.86117220e-01
2.29465291e-01 -7.24982500e-01 -9.93564427e-02 -7.10690975e-01
-1.28579926e+00 7.97945857e-01 4.37878668e-02 1.90593794e-01
4.19720680e-01 4.93862361e-01 7.43818641e-01 7.13061333e-01
3.31925750e-02 8.23902249e-01 -1.11277723e+00 -4.88834709e-01
-1.09633660e+00 8.15813541e-01 1.34957731e+00 1.20260346e+00
4.49075043e-01 -3.66153479e-01 1.59509052e-02 1.87539726e-01
-2.15849921e-01 7.13452458e-01 -1.52696744e-01 -1.38026083e+00
7.29375541e-01 5.63196361e-01 1.06788658e-01 -5.55097282e-01
-4.10304666e-02 -5.70036411e-01 -5.45952395e-02 -1.32535130e-01
8.70627165e-01 -1.02096230e-01 -5.05187213e-01 1.83106387e+00
2.22613201e-01 9.86407548e-02 1.43366545e-01 2.82716483e-01
5.73981822e-01 5.55555642e-01 -3.31939369e-01 -1.39465397e-02
1.02085721e+00 -9.82778013e-01 -2.69897312e-01 1.07542068e-01
1.18774343e+00 -4.36803281e-01 7.28728771e-01 1.06867290e+00
-1.31970644e+00 1.72669277e-01 -1.16640663e+00 -1.17780536e-01
-3.52699935e-01 -2.63935566e-01 1.24110556e+00 4.24118578e-01
-1.21511424e+00 7.00891316e-01 -6.48517072e-01 1.31056588e-02
6.09991431e-01 1.75563619e-01 -1.98600382e-01 -3.63143057e-01
-1.34517956e+00 8.29429090e-01 5.69641590e-01 -6.58312201e-01
-1.33042717e+00 -1.59113026e+00 -9.15643930e-01 2.33612910e-01
7.48866737e-01 -7.20288157e-01 1.97913730e+00 -1.81455314e-01
-1.18428147e+00 6.66177630e-01 -1.39224812e-01 -1.12658739e+00
8.05386543e-01 -3.00503373e-01 -6.40432537e-02 3.02843630e-01
4.37224805e-01 5.30342817e-01 2.59406090e-01 -8.46563280e-01
-7.76723862e-01 2.90158957e-01 7.49866903e-01 -5.24386466e-01
1.58132426e-02 -4.74434830e-02 1.64277092e-01 8.30052327e-03
-1.81531489e-01 -5.65858305e-01 4.21646714e-01 -1.40730008e-01
-2.67729223e-01 -6.34196281e-01 7.31922686e-01 -4.55859780e-01
1.14504325e+00 -1.61697435e+00 4.36570123e-02 2.20249042e-01
4.16563213e-01 -2.69129407e-02 9.60088819e-02 4.97654796e-01
-1.59944762e-02 1.97133049e-01 1.49377612e-02 8.97316486e-02
6.80465996e-01 1.90838706e-02 -7.83675075e-01 1.11100771e-01
1.31084561e-01 1.06493986e+00 -1.51736736e+00 -9.41274047e-01
4.01239432e-02 -3.13258916e-01 -1.11403155e+00 -2.40802273e-01
-1.15135348e+00 -2.22091481e-01 -3.37249428e-01 7.86764860e-01
5.05669832e-01 -5.30592620e-01 4.87416148e-01 3.34900439e-01
2.73583401e-02 8.75685334e-01 -6.89294398e-01 2.32377911e+00
-3.85730416e-01 7.46010840e-01 1.03202082e-01 -9.70397949e-01
3.25283319e-01 1.77483425e-01 2.26726845e-01 -7.49890983e-01
-9.55646038e-02 5.49308479e-01 2.62966573e-01 -2.96812624e-01
2.93193430e-01 2.01455802e-01 -4.42669354e-02 5.70481420e-01
4.73817796e-01 -2.31632113e-01 7.21386611e-01 8.07979465e-01
1.87685215e+00 5.72804868e-01 -1.54978022e-01 -4.58972305e-01
5.15846908e-01 5.58662117e-01 4.27871823e-01 1.12242365e+00
5.72894476e-02 -3.35096776e-01 1.03786290e+00 -5.04633486e-01
-1.08982420e+00 -1.02538121e+00 -3.13588291e-01 1.02432120e+00
-2.47733206e-01 -1.31836545e+00 -7.29447365e-01 -1.22560811e+00
4.52809364e-01 1.00792241e+00 -3.26134652e-01 -6.84095696e-02
-4.82401043e-01 2.14473993e-01 9.88908768e-01 3.18742156e-01
6.22035444e-01 -6.71387136e-01 -6.28758252e-01 8.65384117e-02
-4.45351303e-01 -1.20270324e+00 2.34336466e-01 1.89654857e-01
-6.53350592e-01 -1.65359938e+00 -1.08691759e-01 -5.96350670e-01
2.98576117e-01 -1.77737206e-01 1.49170101e+00 4.06645149e-01
-1.36623129e-01 2.40529567e-01 -3.00550550e-01 -4.09053832e-01
-8.79066169e-01 4.07568693e-01 -1.44506276e-01 -1.14585662e+00
1.49698392e-01 -6.95631921e-01 -1.17221914e-01 -1.10926777e-01
-2.20719144e-01 2.62815863e-01 4.11413282e-01 9.43644881e-01
-9.75814834e-02 2.62473702e-01 2.33814836e-01 -1.16897249e+00
2.91348130e-01 -3.39456886e-01 -1.13263929e+00 5.24239659e-01
-8.49510133e-01 3.96230161e-01 1.05707228e+00 -2.42756322e-01
-5.49213529e-01 -7.00118840e-01 5.13304234e-01 -5.37699282e-01
2.23058760e-01 3.47249687e-01 3.55083019e-01 -5.59575595e-02
8.75683010e-01 1.75694004e-01 -1.42986104e-01 7.92616308e-02
3.23306262e-01 8.35058764e-02 7.92939186e-01 -1.56947625e+00
1.20780647e+00 -1.58216655e-01 3.85528237e-01 1.93911761e-01
-1.07233644e+00 2.98187044e-02 -8.22041780e-02 1.21722877e-01
9.45143253e-02 -7.35845029e-01 -1.36860132e+00 -1.04636410e-02
-1.23510075e+00 -1.09936762e+00 -1.67022366e-02 2.58123308e-01
-6.81234360e-01 3.53834301e-01 -7.09181249e-01 -8.36655676e-01
-4.87982482e-01 -9.45102990e-01 1.03531730e+00 -2.84829319e-01
-4.55356926e-01 -9.34059680e-01 1.34949669e-01 4.34029251e-01
2.18434319e-01 1.90250725e-01 1.38058221e+00 -8.53339612e-01
-8.27493548e-01 1.13056086e-01 -3.37047219e-01 1.62651807e-01
-3.63602310e-01 1.29811689e-02 -9.28227842e-01 -1.64825544e-01
-5.58729291e-01 -8.59589398e-01 6.73901439e-01 -4.61068973e-02
1.10394812e+00 -5.94254434e-01 -5.35530865e-01 4.54865038e-01
1.15821147e+00 -2.72701412e-01 4.03553665e-01 9.42044258e-01
-2.85339262e-02 -1.88182756e-01 7.74184406e-01 6.10357076e-02
5.14996529e-01 1.08502321e-01 4.98789668e-01 5.65342247e-01
1.22924820e-02 -8.52406323e-01 3.24152052e-01 -4.41986360e-02
1.28828049e-01 2.45928258e-01 -1.29847169e+00 6.60869181e-01
-1.84267020e+00 -1.18016362e+00 1.46877617e-01 1.85589004e+00
1.47106743e+00 7.25224376e-01 -1.08335549e-02 3.59439969e-01
1.52647719e-01 -3.44806880e-01 -4.08819377e-01 -6.39444411e-01
2.30477795e-01 8.45582962e-01 4.68657672e-01 7.66649842e-01
-9.30889487e-01 1.04528284e+00 6.86993790e+00 9.17461097e-01
-6.21297538e-01 -1.01151749e-01 -4.40659523e-02 -8.40568841e-02
-6.25391126e-01 5.70454597e-01 -5.68449020e-01 1.89305723e-01
1.11428773e+00 -5.50355077e-01 8.91503572e-01 1.19582951e+00
-5.02541840e-01 -2.51161903e-01 -1.68824184e+00 5.02561450e-01
-1.28326148e-01 -1.75823128e+00 -1.32415906e-01 -1.26195699e-01
7.83972502e-01 -7.15560392e-02 -5.66725992e-02 1.11902940e+00
7.84424126e-01 -1.03063786e+00 8.87139261e-01 6.37307093e-02
9.19846237e-01 -7.84953117e-01 5.75573266e-01 6.55357301e-01
-7.33326793e-01 -2.36146599e-01 6.94302171e-02 -4.10536289e-01
-5.55448949e-01 2.68390924e-01 -1.51605260e+00 8.44362378e-01
5.61134756e-01 7.46713042e-01 -6.12358272e-01 7.45174766e-01
-7.34028935e-01 8.10039163e-01 -5.44854522e-01 -3.30294192e-01
1.88611165e-01 3.52507353e-01 4.46860045e-01 1.16068804e+00
-1.20252669e-01 -7.31178373e-02 6.21679425e-02 1.36975002e+00
-2.92492688e-01 -5.47177792e-01 -5.87148666e-01 -5.39251901e-02
6.83253348e-01 1.21448898e+00 -8.88622701e-02 -6.57444835e-01
-3.58537473e-02 4.57379907e-01 5.96855700e-01 1.06491327e-01
-1.35935616e+00 -5.23179054e-01 1.89075395e-01 9.28141996e-02
3.91111314e-01 -1.55973330e-01 -1.09260425e-01 -1.17092836e+00
3.34816307e-01 -1.23380756e+00 4.25207675e-01 -9.22445536e-01
-6.91850185e-01 -7.29639977e-02 5.09262979e-01 -6.10386550e-01
-5.42467237e-01 -7.22521961e-01 -5.58546185e-01 6.49370432e-01
-1.83916056e+00 -1.12789750e+00 1.69823006e-01 3.14797819e-01
1.77348271e-01 -1.26201078e-01 1.11965537e+00 1.13595881e-01
-6.63135275e-02 1.12247682e+00 -2.30927914e-01 2.47189239e-01
6.52488053e-01 -1.68863261e+00 3.62845033e-01 8.91778529e-01
3.23095508e-02 1.04945064e+00 9.23886001e-01 -4.81829345e-01
-2.06848836e+00 -7.13203847e-01 1.14606750e+00 -6.69377804e-01
1.49116182e+00 -4.69211817e-01 -4.05044168e-01 1.08998346e+00
1.45175859e-01 -2.16740400e-01 6.83296770e-02 8.04036558e-01
-1.08707559e+00 -1.31027624e-01 -1.13430715e+00 6.12376392e-01
1.32845545e+00 -7.22895026e-01 -9.98081744e-01 6.63487673e-01
7.18702018e-01 -9.93595898e-01 -9.82917726e-01 4.15525645e-01
7.57447839e-01 -7.95206845e-01 7.18719482e-01 -9.09605265e-01
1.04400778e+00 -1.41729519e-01 -8.94061252e-02 -8.68320823e-01
-3.16253491e-02 -1.07609367e+00 -7.77148426e-01 9.70802844e-01
5.91549575e-01 -6.68582380e-01 7.35557854e-01 4.71118420e-01
-1.52776614e-01 -9.17442560e-01 -9.43958402e-01 -8.74191523e-01
5.71957707e-01 -6.68778718e-01 4.41648215e-01 9.38756704e-01
9.58058476e-01 3.62463087e-01 4.54395652e-01 1.32957041e-01
8.18223476e-01 5.36707878e-01 1.19213653e+00 -9.90628123e-01
-4.83271122e-01 -5.31570911e-01 -1.89970672e-01 -9.74625707e-01
8.28509748e-01 -1.49988139e+00 -2.53485292e-01 -1.50598526e+00
6.04038775e-01 -4.40030128e-01 -1.05160169e-01 1.00914061e+00
2.60472834e-01 -3.01052183e-01 -6.46536425e-02 -3.57550144e-01
-1.08012879e+00 7.96747208e-02 1.08944106e+00 -4.72405016e-01
4.03487563e-01 -4.15079534e-01 -5.53798258e-01 5.74031770e-01
5.70604444e-01 -2.30538353e-01 -4.49572533e-01 -3.81094366e-01
8.04034233e-01 1.55423269e-01 7.07121789e-01 -1.04821718e+00
4.89960492e-01 -5.95794655e-02 -2.04485562e-02 -6.19060338e-01
-2.72446275e-01 -5.54327250e-01 -1.92132458e-01 6.40466869e-01
-4.22806561e-01 6.87253997e-02 7.04866529e-01 1.07379399e-01
1.85355976e-01 -1.08225815e-01 1.11495204e-01 -2.30473932e-02
-4.80461627e-01 -1.92177638e-01 1.90984309e-02 5.15210450e-01
7.80704916e-01 1.66873142e-01 -9.27584708e-01 -1.05325103e-01
-1.41028881e-01 7.76688397e-01 4.54427987e-01 3.49040926e-02
3.04628938e-01 -8.52424622e-01 -6.25066400e-01 -1.29727870e-01
-6.26090029e-03 2.46948630e-01 -2.01499626e-01 9.89693403e-01
-6.49614811e-01 9.03472424e-01 1.65773720e-01 -6.36280954e-01
-1.22740066e+00 6.39992535e-01 3.81759763e-01 -7.08670080e-01
-7.43229330e-01 7.58152425e-01 -3.57176304e-01 -4.84605312e-01
2.43168592e-01 -9.11040306e-01 6.42726600e-01 -8.24635863e-01
4.39455956e-01 2.47468501e-01 1.43733591e-01 7.11513698e-01
-5.65306902e-01 1.16022639e-01 -3.47707689e-01 -2.56955214e-02
1.29035962e+00 8.40625346e-01 -2.61650592e-01 1.29650906e-01
7.90568531e-01 2.37945512e-01 -6.80783987e-01 -7.09170997e-02
7.95588866e-02 -1.73569754e-01 -1.20444626e-01 -1.33224344e+00
-2.90308565e-01 3.83597106e-01 -4.51438755e-01 2.18946189e-01
6.82633400e-01 8.84417221e-02 7.36844242e-01 1.02314508e+00
1.04511690e+00 -8.29280913e-01 -3.28050226e-01 8.34066808e-01
5.18394053e-01 -9.09569323e-01 2.87742704e-01 -2.82474816e-01
1.04066789e-01 1.30999601e+00 4.28956807e-01 -1.68684512e-01
-7.22925439e-02 6.75711572e-01 -6.59137368e-01 -1.17201358e-02
-1.44688082e+00 4.36726511e-01 -1.40686527e-01 3.04726660e-01
1.90584987e-01 -1.05992973e-01 5.41897789e-02 6.12408340e-01
-8.30125868e-01 4.78560656e-01 3.33400279e-01 1.09810472e+00
-2.81916976e-01 -1.18354225e+00 -2.16130525e-01 1.48815021e-01
-4.72575784e-01 -4.54476982e-01 -3.79475117e-01 1.37555325e+00
-2.85459645e-02 8.13928843e-01 -2.95714319e-01 -2.86547363e-01
-1.83165376e-03 4.09756154e-01 1.44388688e+00 -4.16340232e-01
-2.77482450e-01 -6.60419047e-01 7.48560011e-01 -6.79101527e-01
-7.83319697e-02 -6.91054821e-01 -1.32006574e+00 -1.03132391e+00
-3.46712500e-01 6.46490216e-01 3.05095941e-01 9.49615479e-01
5.89982569e-02 5.80712020e-01 2.75642723e-01 -4.17914242e-02
-1.09195328e+00 -8.14118505e-01 -1.18269004e-01 -4.94805500e-02
4.09373581e-01 -7.12376833e-01 -3.39194924e-01 -1.22346021e-01] | [8.94991683959961, 7.0665364265441895] |
3563b6eb-fa5e-4ec3-8f48-0cee476063d6 | ssncse-nlp-lt-edi-acl2022-hope-speech | null | null | https://aclanthology.org/2022.ltedi-1.30 | https://aclanthology.org/2022.ltedi-1.30.pdf | SSNCSE_NLP@LT-EDI-ACL2022:Hope Speech Detection for Equality, Diversity and Inclusion using sentence transformers | In recent times, applications have been developed to regulate and control the spread of negativity and toxicity on online platforms. The world is filled with serious problems like political & religious conflicts, wars, pandemics, and offensive hate speech is the last thing we desire. Our task was to classify a text into ‘Hope Speech’ and ‘Non-Hope Speech’. We searched for datasets acquired from YouTube comments that offer support, reassurance, inspiration, and insight, and the ones that don’t. The datasets were provided to us by the LTEDI organizers in English, Tamil, Spanish, Kannada, and Malayalam. To successfully identify and classify them, we employed several machine learning transformer models such as m-BERT, MLNet, BERT, XLMRoberta, and XLM_MLM. The observed results indicate that the BERT and m-BERT have obtained the best results among all the other techniques, gaining a weighted F1- score of 0.92, 0.71, 0.76, 0.87, and 0.83 for English, Tamil, Spanish, Kannada, and Malayalam respectively. This paper depicts our work for the Shared Task on Hope Speech Detection for Equality, Diversity, and Inclusion at LTEDI 2021. | ['Senthil Kumar B', 'Thenmozhi Durairaj', 'Josephine Varsha', 'Dhanya Srinivasan', 'Bharathi B'] | null | null | null | null | ltedi-acl-2022-5 | ['hope-speech-detection'] | ['natural-language-processing'] | [-6.11584425e-01 2.56450567e-02 -3.81145477e-01 1.87121212e-01
-5.24771512e-01 -6.90945923e-01 9.97357666e-01 2.28833169e-01
-3.52793306e-01 1.07743943e+00 7.91405559e-01 -4.98199344e-01
-1.61444530e-01 -4.59788233e-01 1.29618160e-02 -4.26440746e-01
5.99364098e-03 2.78277189e-01 -1.38286073e-02 -7.88456917e-01
4.41791058e-01 2.01762736e-01 -1.05226851e+00 3.34344178e-01
1.00651741e+00 7.75246501e-01 -3.67755175e-01 7.20935345e-01
6.97650537e-02 1.44185221e+00 -6.97943151e-01 -9.72760975e-01
-2.17796415e-01 -2.87438482e-01 -8.74234319e-01 -5.28048515e-01
-6.52423277e-02 -8.85043591e-02 -4.46358740e-01 1.14977539e+00
7.61569798e-01 5.97419515e-02 6.46494687e-01 -1.48199391e+00
-7.07978249e-01 8.86853695e-01 -6.59190297e-01 4.53169346e-01
5.38225353e-01 -2.27002859e-01 8.11222315e-01 -6.95892513e-01
9.37496424e-01 1.40683568e+00 5.46079457e-01 5.14642775e-01
-5.35042822e-01 -9.24849749e-01 -6.20541573e-01 2.95152545e-01
-1.34702933e+00 -6.43105507e-01 6.28869355e-01 -4.97376591e-01
1.08012116e+00 6.87191963e-01 5.82953811e-01 1.62024117e+00
1.88436568e-01 7.79755354e-01 1.05020094e+00 -2.89199352e-01
-3.16666923e-02 6.50113285e-01 -1.51447356e-01 5.99207103e-01
-3.56977195e-01 -9.21175703e-02 -8.44302416e-01 -4.39001262e-01
-1.14002600e-01 -5.08503675e-01 -3.31672668e-01 4.78783816e-01
-8.13705862e-01 1.19136906e+00 3.79270501e-02 5.38671970e-01
-1.82273969e-01 -6.07894540e-01 8.20834041e-01 5.26593268e-01
8.50766778e-01 2.36742258e-01 -2.01441765e-01 -8.64242196e-01
-9.41053689e-01 4.50887717e-02 1.07820892e+00 6.05412006e-01
-4.31562178e-02 4.45365310e-02 3.64742279e-02 1.13508546e+00
3.25270355e-01 5.29955685e-01 7.33252525e-01 -5.93158543e-01
5.95297396e-01 3.24926198e-01 3.10001057e-02 -1.43229890e+00
-4.62543130e-01 -4.15305614e-01 -1.09189057e+00 -2.89681882e-01
1.46777160e-03 -4.25674647e-01 -5.41007519e-01 1.49670911e+00
1.17418379e-01 -2.13259518e-01 2.72473365e-01 5.83736539e-01
1.09744978e+00 9.59061265e-01 7.13666007e-02 -5.27326822e-01
9.58893538e-01 -7.89906502e-01 -1.00549090e+00 3.28925662e-02
7.68304706e-01 -1.18166018e+00 9.26341414e-01 7.14715123e-01
-8.81808281e-01 6.46546260e-02 -7.98534036e-01 5.06842360e-02
-4.87754881e-01 -6.73305094e-02 3.22501332e-01 9.85148132e-01
-9.22975719e-01 4.47085679e-01 -1.65413246e-01 -6.38119042e-01
4.40522850e-01 -6.14960790e-02 -4.44627255e-01 4.05121565e-01
-1.49056959e+00 9.99457955e-01 1.69973627e-01 -1.38125017e-01
-6.12197399e-01 -1.67104110e-01 -4.54059482e-01 -1.59419298e-01
1.93663442e-03 1.71412557e-01 7.07526147e-01 -8.30624402e-01
-1.21180451e+00 9.52967763e-01 2.26358503e-01 -5.91308892e-01
4.28864807e-01 -3.07214051e-01 -8.49128187e-01 -1.63320720e-01
-3.47948074e-02 4.01656866e-01 4.72487479e-01 -7.57998466e-01
-5.80524445e-01 -3.53039324e-01 -4.46024388e-02 8.82838368e-02
-7.77710676e-01 7.20325649e-01 3.09424877e-01 -4.96397495e-01
-4.25540179e-01 -6.79894328e-01 4.77574319e-01 -6.79716766e-01
-8.55056763e-01 -3.08225691e-01 1.11943495e+00 -1.25305665e+00
1.81082964e+00 -2.21636963e+00 1.15526892e-01 1.84802458e-01
2.27467477e-01 4.91347909e-01 2.69624412e-01 6.98320508e-01
1.75875366e-01 6.39772475e-01 2.83879012e-01 6.50096759e-02
6.47348985e-02 1.92762852e-01 -1.95184186e-01 5.15657127e-01
-2.72579700e-01 4.72525895e-01 -8.17575753e-01 -4.90078479e-01
-1.27973989e-01 4.65172023e-01 -4.83830273e-01 1.48926184e-01
1.62996501e-01 4.02298905e-02 -2.26865336e-01 7.33105183e-01
2.68076867e-01 -1.93599686e-02 2.53509879e-01 7.41137862e-02
-5.54934978e-01 4.04947102e-01 -6.15054846e-01 8.21966171e-01
-2.48491168e-01 9.97790039e-01 1.52598768e-01 -7.35332429e-01
1.14036286e+00 5.79936743e-01 3.48267436e-01 -8.75311553e-01
4.70504463e-01 4.53680098e-01 2.88522661e-01 -7.57714152e-01
7.16799915e-01 -1.39742419e-01 -2.25545377e-01 2.18970791e-01
-1.86178043e-01 8.37756023e-02 7.76154995e-02 5.35250843e-01
9.20463860e-01 -4.66015100e-01 4.15118217e-01 -2.33018860e-01
6.52661562e-01 -3.41974378e-01 3.57203692e-01 3.40766340e-01
-6.74874306e-01 3.08915079e-01 8.37445796e-01 -1.19405732e-01
-1.13033950e+00 -7.66256630e-01 3.65729779e-02 1.21547127e+00
-2.87528783e-01 -6.29903972e-01 -6.17218018e-01 -5.13526022e-01
-4.41423774e-01 1.08708072e+00 -3.24414730e-01 -4.06967878e-01
-2.76283383e-01 -5.84834099e-01 9.78515863e-01 -1.96110293e-01
9.07800853e-01 -1.00180709e+00 -3.76754433e-01 4.33938066e-03
-7.26035476e-01 -9.50366735e-01 -3.21350813e-01 -5.65639809e-02
-3.03785950e-01 -7.76812017e-01 -6.55786335e-01 -5.43392897e-01
-1.56288430e-01 -1.02642342e-01 8.40914249e-01 -5.22413738e-02
-1.39536839e-02 5.65414429e-02 -5.89514852e-01 -4.65698034e-01
-6.16919696e-01 1.60486430e-01 7.86792859e-02 -1.91104218e-01
2.65228093e-01 -4.25170422e-01 -1.74866304e-01 1.64766133e-01
-6.13514781e-01 -1.73568159e-01 -3.08789518e-02 8.53640020e-01
-3.49801034e-01 1.11766957e-01 7.70470202e-01 -8.46729994e-01
1.05351818e+00 -1.10752308e+00 5.98030649e-02 3.40317823e-02
-3.69094968e-01 -5.04107237e-01 5.87744296e-01 -2.71558702e-01
-1.02031767e+00 -4.64416504e-01 -5.09033918e-01 -3.94405797e-02
1.61289662e-01 7.17836797e-01 -2.71162875e-02 2.89748639e-01
8.95127296e-01 1.22574836e-01 -1.87660724e-01 -4.40225124e-01
2.46233359e-01 1.55591393e+00 2.55710036e-01 -9.67817158e-02
5.08777976e-01 1.00361541e-01 -5.76398253e-01 -1.36510885e+00
-7.21236229e-01 -5.31372845e-01 -2.91528165e-01 -4.82355446e-01
8.75329316e-01 -7.39751339e-01 -5.30489862e-01 6.63787067e-01
-1.21406817e+00 9.28632990e-02 1.44954219e-01 2.48841077e-01
-7.51940012e-02 5.13616562e-01 -7.33407676e-01 -1.28061330e+00
-8.51763368e-01 -5.40692329e-01 3.35640669e-01 2.81292528e-01
-6.61479414e-01 -1.00311947e+00 6.28219247e-02 5.81863999e-01
6.41455948e-01 3.20140749e-01 1.16533983e+00 -1.16013634e+00
4.37266022e-01 -5.09867296e-02 -1.64905876e-01 4.55094010e-01
1.52513534e-01 3.29775155e-01 -7.12717235e-01 -1.33009806e-01
6.31664023e-02 -7.44449973e-01 3.32837075e-01 1.12262657e-02
7.05109477e-01 -1.04992747e+00 1.90930814e-02 8.36654827e-02
8.67557824e-01 4.79014367e-01 9.07466888e-01 5.17367721e-01
2.58101195e-01 5.20704329e-01 5.28721690e-01 8.86770308e-01
3.60306770e-01 4.13633078e-01 4.72323895e-01 1.58633724e-01
2.13240728e-01 -3.64160240e-01 6.33395016e-01 1.20103896e+00
1.55542223e-02 -7.16006398e-01 -9.66041088e-01 5.75959504e-01
-1.58674645e+00 -1.26703691e+00 -2.96241075e-01 2.02596378e+00
7.92978466e-01 2.98020542e-01 5.95676899e-01 5.37672520e-01
5.10379016e-01 3.90830070e-01 1.56606302e-01 -1.10751879e+00
-3.64862204e-01 2.61232890e-02 1.77775905e-01 5.58556557e-01
-1.18588328e+00 9.27581370e-01 5.21509123e+00 1.09330475e+00
-1.25819671e+00 2.35939711e-01 1.03471565e+00 -5.15230410e-02
-3.16208303e-01 -3.58423442e-01 -4.75708514e-01 6.51276648e-01
1.42803514e+00 -4.31441545e-01 4.54917729e-01 6.78717554e-01
5.70278287e-01 -7.83338677e-04 -2.31965676e-01 9.32452142e-01
2.78910697e-01 -1.07255650e+00 -2.77058423e-01 -6.12827204e-02
5.72364688e-01 3.41680288e-01 2.99181610e-01 3.09868455e-01
2.52248764e-01 -1.16066301e+00 6.09000206e-01 2.20515504e-01
5.32970428e-01 -9.85951960e-01 9.00543153e-01 7.42532551e-01
-4.35522377e-01 -2.31858075e-01 -2.53902972e-01 -1.97656937e-02
-6.80590123e-02 5.93518853e-01 -6.56462193e-01 2.87418306e-01
6.85544789e-01 6.86757743e-01 -2.81405389e-01 8.06461692e-01
-2.22161725e-01 1.03281236e+00 -3.68155330e-01 -6.31234348e-01
4.85328197e-01 6.01236634e-02 6.86667144e-01 1.36229897e+00
5.13116956e-01 -4.62990254e-02 -2.39829898e-01 2.99385339e-01
-1.06878698e-01 5.89948475e-01 -7.53396690e-01 -4.42757934e-01
5.00090837e-01 1.15392768e+00 -3.68835956e-01 -1.57087207e-01
-1.80185512e-01 8.06664944e-01 4.96741571e-02 -1.57026369e-02
-1.07688415e+00 -4.86983448e-01 2.49355495e-01 2.10152745e-01
-3.17069262e-01 -1.24277242e-01 -1.11130394e-01 -9.37354326e-01
-4.80310768e-01 -1.31071377e+00 6.15708351e-01 -7.49884307e-01
-1.05727363e+00 7.41083205e-01 -2.30272532e-01 -7.98216522e-01
-3.63073409e-01 -4.81195301e-01 -5.15102386e-01 7.00084567e-01
-9.55013156e-01 -7.06853211e-01 1.98904216e-01 5.42357564e-01
4.92225081e-01 -5.99185884e-01 7.02371299e-01 6.41342282e-01
-5.90574563e-01 5.38337708e-01 2.45853081e-01 2.10020587e-01
4.29370642e-01 -8.63258302e-01 -5.50807826e-02 4.77808595e-01
1.06733225e-01 2.91744918e-01 8.88379216e-01 -5.84214330e-01
-9.84501004e-01 -6.32181883e-01 1.70280206e+00 -3.91093828e-02
9.70253468e-01 -3.46830189e-01 -6.23403490e-01 2.39163339e-01
5.48646271e-01 -7.07963169e-01 8.61088932e-01 1.58522427e-01
-3.55754346e-01 2.89291978e-01 -1.36685383e+00 6.12733185e-01
7.36465752e-01 -5.45651615e-01 -3.64779443e-01 8.06615770e-01
2.75568515e-01 -2.15568483e-01 -7.43542731e-01 -2.48332113e-01
5.51935494e-01 -1.19270909e+00 6.60649300e-01 -9.01530504e-01
8.48255157e-01 4.79624718e-01 -3.11096579e-01 -1.27454138e+00
-1.73578292e-01 -8.94017220e-01 -8.25676024e-02 1.54237306e+00
7.33003438e-01 -5.23758054e-01 6.90126598e-01 3.04425269e-01
5.14291935e-02 -6.35040402e-01 -1.31172156e+00 -5.45634270e-01
3.32174391e-01 -2.52250016e-01 1.16647810e-01 1.28136027e+00
4.79427159e-01 7.29080260e-01 -1.05235517e+00 -4.05367315e-01
1.58001974e-01 -4.98252362e-01 5.06672382e-01 -1.04056609e+00
7.67373517e-02 -3.37599903e-01 -3.57162982e-01 -5.42505622e-01
1.93677074e-03 -9.31607246e-01 -5.66847086e-01 -1.27437127e+00
2.41556600e-01 -1.41625091e-01 1.77815184e-02 5.30100942e-01
2.95725733e-01 2.28873208e-01 2.72774428e-01 2.76355267e-01
-3.88196468e-01 4.74476695e-01 8.59205246e-01 -8.95806700e-02
-1.33440226e-01 2.59997156e-02 -8.52775633e-01 7.69889116e-01
1.26638818e+00 -5.12365758e-01 -3.29961151e-01 7.67053217e-02
5.47739983e-01 3.56391743e-02 9.75941718e-02 -7.32116342e-01
-1.65152568e-02 -1.61404029e-01 6.13693185e-02 -5.35403132e-01
4.37383175e-01 -3.47330958e-01 -9.56663024e-03 7.05843925e-01
-3.20060581e-01 1.83949277e-01 -8.23860914e-02 -3.31895985e-02
-3.09086621e-01 -2.35588461e-01 8.36507499e-01 -9.12235826e-02
-2.57246614e-01 -4.17111665e-02 -7.28113949e-01 4.74849641e-01
6.97809756e-01 2.81925909e-02 -7.49955833e-01 -1.08553910e+00
-6.24793768e-01 6.21537790e-02 -2.47099042e-01 6.57513797e-01
6.94547653e-01 -1.23200357e+00 -9.69098508e-01 -1.17358722e-01
-2.21414313e-01 -1.01463330e+00 2.21609417e-02 9.86420393e-01
-4.72204417e-01 4.28111553e-01 -2.41591945e-01 -9.15295165e-03
-1.61007118e+00 1.67384788e-01 1.81708202e-01 -3.10816675e-01
-1.33521497e-01 8.62318754e-01 -4.63577628e-01 -6.31416559e-01
1.79554045e-01 2.68612444e-01 -5.93617380e-01 3.92232150e-01
2.74890214e-01 8.10440063e-01 -1.73277259e-01 -1.10996366e+00
-4.01477575e-01 -6.31798506e-02 -1.45426737e-02 -8.79582539e-02
1.27203691e+00 -1.85425952e-01 -3.34089965e-01 4.02896225e-01
1.35312164e+00 4.63609576e-01 -1.48258824e-02 2.68918067e-01
1.46345779e-01 -4.91678655e-01 5.69418855e-02 -1.16716957e+00
-9.15436447e-01 8.10626030e-01 4.26180750e-01 8.21018100e-01
9.51355577e-01 -1.96265370e-01 7.94221282e-01 5.28549999e-02
2.74343751e-02 -1.31129646e+00 6.95013031e-02 8.47891390e-01
1.06047130e+00 -9.63188112e-01 -1.64603487e-01 -2.72417367e-01
-8.12455356e-01 1.04946578e+00 2.55846024e-01 4.23699081e-01
7.89754629e-01 9.24923345e-02 1.62835270e-01 -2.87305657e-02
-8.54890287e-01 3.06384385e-01 6.90543875e-02 6.74037933e-01
8.41996372e-01 1.53368711e-01 -8.51055920e-01 6.26326263e-01
-6.48364305e-01 -1.13097690e-01 7.27732003e-01 5.81801534e-01
-7.18511462e-01 -7.43739247e-01 -2.30906531e-01 5.67991257e-01
-9.69625473e-01 -2.62450546e-01 -9.81116295e-01 9.20984030e-01
1.67157687e-02 1.29644966e+00 -1.39804959e-01 -9.25852120e-01
1.27180383e-01 5.34013696e-02 -2.71845132e-01 -9.85371843e-02
-8.72499049e-01 1.54265225e-01 9.12285566e-01 3.08809970e-02
-2.08585501e-01 -4.24489677e-01 -1.03406692e+00 -9.71370578e-01
-3.03532213e-01 5.46694219e-01 6.29652202e-01 7.30511427e-01
1.38457045e-01 2.13495996e-02 8.45197797e-01 -1.24480892e-02
-5.30735970e-01 -1.04509139e+00 -5.30981004e-01 2.37757072e-01
-7.93159846e-03 -2.82268822e-01 -6.14060402e-01 -5.03495038e-01] | [8.926660537719727, 10.646751403808594] |
f146067d-8bec-47a6-b06a-fde0a93355b6 | autoencoders-for-real-time-suep-detection | 2306.13595 | null | https://arxiv.org/abs/2306.13595v2 | https://arxiv.org/pdf/2306.13595v2.pdf | Autoencoders for Real-Time SUEP Detection | Confining dark sectors with pseudo-conformal dynamics can produce Soft Unclustered Energy Patterns, or SUEPs, at the Large Hadron Collider: the production of dark quarks in proton-proton collisions leading to a dark shower and the high-multiplicity production of dark hadrons. The final experimental signature is spherically-symmetric energy deposits by an anomalously large number of soft Standard Model particles with a transverse energy of a few hundred MeV. The dominant background for the SUEP search, if it gets produced via gluon-gluon fusion, is multi-jet QCD events. We have developed a deep learning-based Anomaly Detection technique to reject QCD jets and identify any anomalous signature, including SUEP, in real-time in the High-Level Trigger system of the Compact Muon Solenoid experiment at the Large Hadron Collider. A deep convolutional neural autoencoder network has been trained using QCD events by taking transverse energy deposits in the inner tracker, electromagnetic calorimeter, and hadron calorimeter sub-detectors as 3-channel image data. To tackle the biggest challenge of the task, due to the sparse nature of the data: only ~0.5% of the total ~300 k image pixels have non-zero values, a non-standard loss function, the inverse of the so-called Dice Loss, has been exploited. The trained autoencoder with learned spatial features of QCD jets can detect 40% of the SUEP events, with a QCD event mistagging rate as low as 2%. The model inference time has been measured using the Intel CoreTM i5-9600KF processor and found to be ~20 ms, which perfectly satisfies the High-Level Trigger system's latency of O(100) ms. Given the virtue of the unsupervised learning of the autoencoders, the trained model can be applied to any new physics model that predicts an experimental signature anomalous to QCD jets. | ['Syed Hasan', 'Maurzio Pierini', 'Benedikt Maier', 'Nadezda Chernyavskaya', 'Simranjit Singh Chhibra'] | 2023-06-23 | null | null | null | null | ['anomaly-detection'] | ['methodology'] | [-3.60339403e-01 4.56469879e-02 2.55935937e-01 -3.37370694e-01
-3.69205385e-01 -1.77242085e-01 8.40543389e-01 4.96870205e-02
-7.97620595e-01 5.43982506e-01 -1.11616842e-01 -4.66729224e-01
2.87292004e-01 -7.65311182e-01 -8.78463626e-01 -1.10136700e+00
-1.38086258e-02 1.27057183e+00 4.72921461e-01 4.75095436e-02
-8.47208649e-02 7.38042116e-01 -1.22951818e+00 6.52211130e-01
1.16657794e-01 1.44547677e+00 -2.32910067e-02 7.54278183e-01
1.76221132e-02 6.73183262e-01 -3.13620627e-01 -5.28427482e-01
6.49119854e-01 -5.53899825e-01 -1.85857818e-01 -7.91579843e-01
9.64895356e-03 -2.03855306e-01 -7.53877044e-01 1.49444842e+00
6.19988501e-01 1.40784651e-01 6.53396964e-01 -1.00243855e+00
-2.91015953e-01 5.04696667e-01 -5.98659515e-02 6.54402733e-01
-4.47232366e-01 2.73972720e-01 5.95877290e-01 -1.20326686e+00
5.38965344e-01 1.13947237e+00 4.72301573e-01 4.92213845e-01
-1.13923156e+00 -9.11833048e-01 -1.00473094e+00 1.40478700e-01
-7.93591022e-01 -1.21631436e-01 4.13172930e-01 -5.57717144e-01
1.53243065e+00 1.87367305e-01 7.48191357e-01 1.38396025e+00
1.16949654e+00 2.32576996e-01 1.22987437e+00 -6.84725642e-01
6.77768588e-01 2.03130350e-01 -2.03806207e-01 3.94418150e-01
2.73008227e-01 1.09338927e+00 -3.21965188e-01 -2.88367301e-01
8.42295825e-01 -2.13189460e-02 3.48155349e-01 -2.25245073e-01
-1.11929154e+00 1.16555142e+00 3.78585666e-01 4.61061805e-01
-3.71162109e-02 8.74583870e-02 8.90113354e-01 3.76227289e-01
3.61161292e-01 8.35363865e-01 -6.20437860e-01 5.88719249e-02
-7.39783227e-01 5.44710577e-01 7.36370981e-01 5.33949792e-01
2.13791147e-01 4.69729066e-01 3.39568742e-02 1.64351016e-01
1.41214311e-01 8.75701189e-01 6.03675008e-01 -2.61396199e-01
3.75414878e-04 2.71875620e-01 1.01786986e-01 -4.57804531e-01
-6.39555573e-01 -2.61128753e-01 -9.89930451e-01 7.85056472e-01
4.75072891e-01 -5.33314526e-01 -1.09927845e+00 1.15767527e+00
4.38082516e-01 -2.82290548e-01 3.25278044e-02 1.25859594e+00
4.83484119e-01 7.13104486e-01 7.64370412e-02 -2.08698213e-02
1.28504908e+00 -5.56222558e-01 -3.65189493e-01 -1.53709143e-01
2.89009511e-01 -9.42954242e-01 -8.11261311e-03 3.64784062e-01
-9.09462273e-01 -5.36482155e-01 -1.00536954e+00 8.11534449e-02
-3.10009062e-01 -2.57581055e-01 1.09527636e+00 2.12934747e-01
-4.02697951e-01 9.35879588e-01 -1.14792478e+00 -9.50047150e-02
-1.54416978e-01 3.41595083e-01 -1.17304556e-01 6.14261985e-01
-1.13027239e+00 8.76943111e-01 8.73270929e-01 -2.07853764e-02
-7.43369222e-01 -7.26327300e-01 -3.31537664e-01 1.59076333e-01
-3.84026095e-02 -3.15039933e-01 1.36791337e+00 -1.21463132e+00
-1.20925498e+00 7.60549903e-01 4.87713158e-01 -8.51839423e-01
2.03729063e-01 -1.03161789e-01 -1.26846898e+00 6.26335815e-02
2.68639270e-02 7.88042620e-02 6.50309563e-01 -6.00476861e-01
-6.79820955e-01 -4.21768695e-01 -8.02741408e-01 -3.39937687e-01
4.86948550e-01 8.33725154e-01 -5.40088452e-02 -6.56995416e-01
2.40310088e-01 -1.07280731e+00 -4.53486234e-01 -8.43950033e-01
-3.09401453e-01 4.71987575e-02 5.04828811e-01 -3.31305444e-01
3.19391161e-01 -2.16868210e+00 2.98911929e-02 3.52395296e-01
1.68873623e-01 1.82953864e-01 4.39260751e-01 1.86786503e-01
-4.86348718e-01 -5.72688043e-01 2.25180045e-01 2.88718671e-01
1.83343485e-01 -4.28093411e-02 -4.17338789e-01 4.39478904e-01
-7.26012439e-02 8.73491526e-01 -6.76806748e-01 2.49587655e-01
4.49076116e-01 2.09305048e-01 -3.83360565e-01 2.84882098e-01
-4.00979877e-01 6.57133460e-01 -4.38588858e-01 3.65355492e-01
1.04537332e+00 1.27842531e-01 -1.44266799e-01 -2.27102131e-01
-4.73349363e-01 9.73427594e-02 -9.32815909e-01 1.13352036e+00
8.82700160e-02 4.70143527e-01 4.37356830e-01 -8.73662591e-01
8.83850813e-01 2.19745249e-01 6.35671437e-01 -1.61665857e+00
3.14186156e-01 5.73173404e-01 3.11823577e-01 -1.62636042e-01
5.96684873e-01 -7.55841970e-01 -3.23027223e-01 2.23986387e-01
3.47103417e-01 3.15078467e-01 -1.24786809e-01 1.03842564e-01
1.40900803e+00 -3.01094890e-01 -1.48919001e-01 -4.48245555e-01
-9.97872427e-02 3.36721629e-01 8.75894368e-01 9.25267696e-01
1.02589009e-02 3.39753300e-01 5.26268721e-01 -1.16816533e+00
-1.95089912e+00 -1.24797845e+00 -7.92789459e-01 8.52620661e-01
-8.83771181e-02 -3.36672932e-01 -3.84863794e-01 -5.39218366e-01
1.21023886e-01 9.31281567e-01 -4.76425253e-02 -4.22141880e-01
-6.26768827e-01 -1.37535524e+00 4.22803164e-01 7.59992182e-01
8.43243226e-02 -1.53420186e+00 -4.87750530e-01 5.42197287e-01
7.73793995e-01 -1.17673266e+00 3.69948328e-01 1.14436245e+00
-5.42036176e-01 -9.21525598e-01 1.46198988e-01 -3.08809817e-01
5.30159354e-01 -7.43716836e-01 1.23090541e+00 -4.76371467e-01
-7.63966084e-01 -2.60360688e-01 -3.66045237e-01 -4.78895724e-01
-9.03857529e-01 -9.96054471e-01 8.21626902e-01 -2.68824756e-01
1.15160465e+00 -3.18056762e-01 -6.14963710e-01 -1.52227208e-02
-6.03420734e-01 -8.76219273e-02 9.64620471e-01 1.11658669e+00
6.43573523e-01 2.98749685e-01 1.73177019e-01 -4.49584216e-01
-2.81682819e-01 -4.03039455e-01 -1.24726367e+00 -5.98871827e-01
-3.50809723e-01 4.19046164e-01 1.01063001e+00 -2.53415495e-01
-1.07662237e+00 1.46676540e-01 -4.37466264e-01 -6.06535077e-01
-3.65786761e-01 -3.82118791e-01 -1.23154342e-01 -2.54809588e-01
5.97468674e-01 -8.56563374e-02 -4.87818390e-01 -7.51690805e-01
7.55069777e-02 3.43300343e-01 9.53594863e-01 -4.43816662e-01
6.92269146e-01 2.14465261e-01 3.17182183e-01 -4.62822437e-01
-6.36350334e-01 -4.24220622e-01 -6.24983788e-01 1.61128536e-01
1.36058724e+00 -9.73606050e-01 -7.40256786e-01 2.85140604e-01
-1.06162369e+00 8.88623446e-02 -5.54918349e-01 1.17604578e+00
-3.38701099e-01 1.12638168e-01 -8.39467525e-01 -5.84113717e-01
-4.58827853e-01 -1.07839954e+00 8.25344682e-01 1.39263764e-01
4.04348493e-01 -4.94269878e-01 3.83516043e-01 -1.47119209e-01
3.51974368e-01 4.08373833e-01 1.22592270e+00 -1.44798267e+00
-5.79705298e-01 -3.93059343e-01 -4.23379779e-01 5.17927170e-01
-7.66513467e-01 -5.39754748e-01 -6.59461200e-01 -5.10950863e-01
6.98399425e-01 -2.69200832e-01 1.17407537e+00 4.47048903e-01
1.13196063e+00 8.85265842e-02 -4.00449067e-01 6.62794650e-01
1.61625803e+00 3.74522150e-01 1.56073242e-01 2.08287671e-01
5.69652319e-01 -2.32995376e-01 4.43586797e-01 5.03791034e-01
-5.56947470e-01 9.02530074e-01 5.58481157e-01 -3.09103057e-02
1.82826087e-01 3.72699648e-01 5.57058573e-01 9.13305759e-01
-7.46106207e-02 1.54232979e-01 -7.04618216e-01 1.50778621e-01
-1.67516685e+00 -1.20867550e+00 -5.69949687e-01 2.40577102e+00
5.26189864e-01 5.79041719e-01 -4.02653664e-01 -4.12810117e-01
5.55818915e-01 -3.22078675e-01 -3.34638506e-01 -1.27450740e+00
2.14873821e-01 9.51434433e-01 9.88221288e-01 1.64637268e-01
-1.20236850e+00 5.64047635e-01 5.07081938e+00 7.41290092e-01
-1.12237060e+00 3.57892513e-01 8.55306312e-02 -5.75747490e-01
4.20735329e-02 -2.08028927e-02 -9.98448968e-01 1.11034322e+00
1.44340062e+00 2.29387909e-01 4.81449008e-01 1.03317094e+00
-2.36816913e-01 -1.75062716e-01 -1.17518425e+00 9.66405571e-01
-2.38075793e-01 -1.35581315e+00 -5.08834958e-01 6.22148477e-02
6.34186447e-01 7.52529442e-01 -3.70896757e-01 9.26571786e-01
5.48444808e-01 -9.22250152e-01 8.41440380e-01 4.66579407e-01
6.96529925e-01 -1.18257248e+00 9.95023847e-01 3.00733715e-01
-6.01692259e-01 -1.26549393e-01 -9.71793413e-01 -9.45163444e-02
2.23047450e-01 1.28422070e+00 -8.94563973e-01 2.83633202e-01
9.56560075e-01 -3.28188241e-01 -2.74100125e-01 9.38474000e-01
2.84493387e-01 7.57887661e-01 -6.87672257e-01 -3.01109497e-02
4.80388343e-01 -3.19655210e-01 8.56919169e-01 1.40962565e+00
6.38768911e-01 -1.01052932e-02 3.40624422e-01 7.85569072e-01
-8.78509060e-02 -2.88872927e-01 -7.99535394e-01 -7.61263520e-02
-5.21537550e-02 1.22423565e+00 -7.66284227e-01 -3.03620636e-01
-6.09467387e-01 8.60348523e-01 -6.74334168e-02 -4.23820019e-02
-6.59996808e-01 -2.76774496e-01 4.12669510e-01 3.22399251e-02
5.68140566e-01 -1.14637822e-01 -7.05838352e-02 -1.14887118e+00
6.00381047e-02 -5.69758177e-01 1.39145270e-01 -3.86790693e-01
-1.29654443e+00 6.39864266e-01 -2.78161526e-01 -7.33381510e-01
-5.31567752e-01 -1.28500724e+00 -8.74255180e-01 9.64266241e-01
-8.01160932e-01 -7.77853131e-01 2.86327302e-01 4.14378971e-01
2.86330014e-01 -9.88195240e-01 9.87412035e-01 3.45641762e-01
-3.44552875e-01 1.04613099e-02 1.03645539e+00 2.66127765e-01
6.92524910e-01 -1.79543352e+00 7.50851810e-01 7.17741430e-01
1.03601970e-01 -1.56696245e-01 1.17411041e+00 -9.52300191e-01
-1.75595951e+00 -1.28781867e+00 6.93430066e-01 -4.47171181e-01
8.96645367e-01 -1.04398787e+00 -8.22842181e-01 7.40726054e-01
2.19085082e-01 1.03983271e+00 2.36202285e-01 -1.46662727e-01
9.34775770e-02 1.69547364e-01 -1.06562901e+00 -5.51349372e-02
4.63371068e-01 -3.59290034e-01 -7.45688915e-01 7.88137496e-01
3.80258173e-01 -5.23256421e-01 -6.53595984e-01 7.89320827e-01
3.66148464e-02 -1.21294272e+00 8.20192993e-01 -1.00521719e+00
1.10537328e-01 9.75234713e-03 -4.17615958e-02 -1.29577863e+00
-6.12589478e-01 -2.50761330e-01 -1.98607400e-01 6.27789021e-01
3.44052762e-02 -2.19891429e-01 7.59210885e-01 2.57370740e-01
-2.87677079e-01 -4.36961979e-01 -1.42511284e+00 -7.80001998e-01
4.49283004e-01 -4.83249426e-01 1.31494701e-01 6.42530441e-01
-2.94337839e-01 1.22684039e-01 -3.30660552e-01 4.08571362e-01
6.89740002e-01 1.39557868e-01 1.58834577e-01 -1.39607787e+00
-7.41478562e-01 -1.46652773e-01 -7.35177815e-01 -2.88277477e-01
-9.47086439e-02 -1.33521438e+00 -2.55660135e-02 -4.94588614e-01
1.91280290e-01 -2.56027997e-01 -6.76769793e-01 6.10607788e-02
5.59678674e-01 1.82204694e-01 -1.99096546e-01 -1.38712853e-01
-7.44082093e-01 3.93398225e-01 6.30767643e-01 6.21232428e-02
2.95260280e-01 -3.65587771e-02 3.01379055e-01 1.17957330e+00
4.45321202e-01 -6.89505041e-01 6.76799297e-01 -1.17448628e-01
7.45691895e-01 3.17143351e-01 7.60107458e-01 -1.47849381e+00
2.47203052e-01 1.47223279e-01 1.40386176e+00 -6.52580380e-01
1.78125292e-01 -6.37718439e-01 1.22551464e-01 1.49602756e-01
3.71331610e-02 8.04197788e-02 3.90270829e-01 3.45156252e-01
-2.54224539e-01 -6.54115319e-01 1.05934787e+00 -4.81872022e-01
-7.00406730e-01 9.03383493e-02 -3.88513952e-01 -5.50891794e-02
1.18336189e+00 6.41228974e-01 5.64821325e-02 5.84235191e-01
-8.99241865e-01 -9.96794924e-02 3.58625472e-01 4.31112558e-01
2.21104205e-01 -1.77970231e+00 -5.03149748e-01 4.78523940e-01
-1.06285505e-01 -2.94797540e-01 1.54314667e-01 5.55832863e-01
-5.81139445e-01 6.16587043e-01 -5.91747522e-01 -4.48037505e-01
-7.03786492e-01 9.16677177e-01 3.48869652e-01 -2.67767400e-01
-1.53804362e+00 7.00133204e-01 1.79488465e-01 -3.86306882e-01
-1.63077891e-01 -8.46536234e-02 4.04333711e-01 -4.24256712e-01
7.02162862e-01 6.96953312e-02 6.96970820e-01 -4.28016335e-01
-1.56251103e-01 -2.09758103e-01 -3.41963619e-01 2.31856823e-01
1.26669204e+00 7.66756892e-01 -1.98158517e-01 2.19173342e-01
1.02687252e+00 1.86740175e-01 -1.08781111e+00 -9.80849862e-02
1.18394613e-01 -1.67323127e-01 6.21091843e-01 -1.12840474e+00
-8.99784565e-01 8.42938364e-01 1.05536473e+00 4.15610969e-01
5.04806161e-01 4.62189198e-01 1.02668047e+00 6.13553762e-01
2.29302645e-01 -1.29641116e+00 -2.67549396e-01 7.67840803e-01
7.46940255e-01 -1.37044513e+00 -3.41743439e-01 6.77504718e-01
-5.99999905e-01 1.38630152e+00 6.40862942e-01 -3.34223419e-01
3.62888724e-01 9.02735472e-01 -2.57863849e-01 -7.04916358e-01
-9.43517685e-01 3.85159045e-01 1.54649138e-01 1.72958318e-02
6.90046549e-02 4.82956409e-01 -1.50217518e-01 8.01842988e-01
-2.40864530e-01 -4.32156444e-01 3.51770461e-01 5.36655724e-01
-6.67325318e-01 -9.98830378e-01 -6.27852559e-01 7.68552244e-01
-7.56548464e-01 -1.43672392e-01 1.64190724e-01 5.09984672e-01
4.32843626e-01 3.27793539e-01 6.87353909e-01 3.58495675e-02
1.34852484e-01 4.45598155e-01 4.45261896e-01 -4.40822363e-01
-6.90111816e-01 1.40581533e-01 -2.00559407e-01 -8.78468156e-01
2.83647031e-01 -4.77846444e-01 -1.48269188e+00 -2.18748271e-01
-2.27072746e-01 4.65799510e-01 1.05138278e+00 1.08283544e+00
3.90395597e-02 5.32774687e-01 6.12560034e-01 -1.13636982e+00
-8.91277492e-01 -7.68245757e-01 -8.42988551e-01 4.39345062e-01
1.17847808e-01 -6.28909111e-01 -5.55537939e-01 -7.45790064e-01] | [15.689948081970215, 2.92404842376709] |
4398dc2d-c245-4c3f-8494-f2414f687248 | uncovering-convolutional-neural-network | 1904.08771 | null | http://arxiv.org/abs/1904.08771v1 | http://arxiv.org/pdf/1904.08771v1.pdf | Uncovering convolutional neural network decisions for diagnosing multiple sclerosis on conventional MRI using layer-wise relevance propagation | Machine learning-based imaging diagnostics has recently reached or even
superseded the level of clinical experts in several clinical domains. However,
classification decisions of a trained machine learning system are typically
non-transparent, a major hindrance for clinical integration, error tracking or
knowledge discovery. In this study, we present a transparent deep learning
framework relying on convolutional neural networks (CNNs) and layer-wise
relevance propagation (LRP) for diagnosing multiple sclerosis (MS). MS is
commonly diagnosed utilizing a combination of clinical presentation and
conventional magnetic resonance imaging (MRI), specifically the occurrence and
presentation of white matter lesions in T2-weighted images. We hypothesized
that using LRP in a naive predictive model would enable us to uncover relevant
image features that a trained CNN uses for decision-making. Since imaging
markers in MS are well-established this would enable us to validate the
respective CNN model. First, we pre-trained a CNN on MRI data from the
Alzheimer's Disease Neuroimaging Initiative (n = 921), afterwards specializing
the CNN to discriminate between MS patients and healthy controls (n = 147).
Using LRP, we then produced a heatmap for each subject in the holdout set
depicting the voxel-wise relevance for a particular classification decision.
The resulting CNN model resulted in a balanced accuracy of 87.04% and an area
under the curve of 96.08% in a receiver operating characteristic curve. The
subsequent LRP visualization revealed that the CNN model focuses indeed on
individual lesions, but also incorporates additional information such as lesion
location, non-lesional white matter or gray matter areas such as the thalamus,
which are established conventional and advanced MRI markers in MS. We conclude
that LRP and the proposed framework have the capability to make diagnostic
decisions of... | ['John-Dylan Haynes', 'René M. Giess', 'Klemens Ruprecht', 'Judith Bellmann-Strobl', 'Martin Weygandt', 'Susanna Asseyer', 'Kerstin Ritter', 'Joseph Kuchling', 'Friedemann Paul', 'Fabian Eitel', 'Emily Soehler', 'Alexander U. Brandt', 'Michael Scheel'] | 2019-04-18 | null | null | null | null | ['holdout-set'] | ['computer-vision'] | [ 3.75621855e-01 2.61812061e-01 -1.41138524e-01 -5.47960877e-01
-7.02013671e-01 -2.52647847e-01 5.13426185e-01 5.14082611e-01
-6.82039499e-01 7.19170451e-01 1.39450207e-01 -5.38210988e-01
-5.32616377e-01 -5.67764819e-01 -4.22931701e-01 -7.14524269e-01
-7.70652235e-01 8.12735021e-01 1.39821693e-01 2.61800081e-01
1.07462719e-01 7.39903629e-01 -1.15506089e+00 5.00056028e-01
7.63769031e-01 1.27903020e+00 3.81455153e-01 4.29248989e-01
2.21092954e-01 5.86587131e-01 -3.51651192e-01 -3.59769538e-02
-1.72744896e-02 -9.00500547e-03 -7.84613729e-01 -1.59611955e-01
3.37956935e-01 -2.89619625e-01 -1.58224300e-01 9.38161016e-01
6.92211390e-01 -2.89372653e-01 8.33204806e-01 -7.93773472e-01
-4.91120905e-01 6.70322001e-01 -4.94447768e-01 7.56825566e-01
-3.13179314e-01 4.56517488e-01 6.19211018e-01 -5.42498946e-01
8.54785502e-01 7.72186935e-01 8.32393169e-01 3.26744556e-01
-1.17985797e+00 -3.92570078e-01 -1.03797734e-01 5.57014167e-01
-1.07794309e+00 -1.69567719e-01 1.68661833e-01 -8.88309419e-01
9.52262640e-01 3.23284626e-01 9.63667393e-01 1.04047275e+00
6.69551373e-01 3.05678159e-01 1.56209576e+00 -2.29594082e-01
2.63860136e-01 -8.42261165e-02 2.79139996e-01 5.71615219e-01
8.84854570e-02 8.66777897e-02 6.72246814e-02 -1.69120833e-01
9.23815250e-01 -6.11533262e-02 -1.77981973e-01 -1.10257633e-01
-1.51464295e+00 7.80445457e-01 9.62765336e-01 5.54810047e-01
-6.70416653e-01 -7.15767890e-02 5.01254678e-01 4.71344450e-03
5.03341496e-01 3.09589893e-01 -3.93828750e-01 4.07165021e-01
-1.19520473e+00 1.50020286e-01 2.11622119e-01 -1.37079760e-01
1.65457249e-01 -1.35683522e-01 -4.05311108e-01 8.28959703e-01
1.42908186e-01 3.87097418e-01 7.09020317e-01 -7.40897298e-01
2.59588271e-01 7.41761088e-01 -2.67527699e-01 -8.47414434e-01
-1.11712646e+00 -1.01072598e+00 -1.17829728e+00 4.56363201e-01
4.44710463e-01 -2.61993378e-01 -1.09271669e+00 1.49284935e+00
1.30956724e-01 -3.89610112e-01 -2.60773003e-01 1.26201415e+00
4.41763848e-01 9.85663733e-04 3.31012666e-01 -2.88167130e-02
1.69343781e+00 -4.91343677e-01 -3.27744156e-01 -2.60341704e-01
7.22093999e-01 -1.40519410e-01 7.99746275e-01 3.67517918e-01
-8.66635382e-01 -2.76810765e-01 -1.01553011e+00 1.70179009e-01
-1.92960650e-01 3.61666083e-01 7.19587266e-01 2.73383677e-01
-1.33884823e+00 7.65121698e-01 -9.23003137e-01 -4.97115582e-01
9.19793725e-01 5.86684585e-01 -5.95625937e-01 -1.50762666e-02
-1.19798684e+00 1.56049883e+00 4.29114968e-01 3.35780650e-01
-9.17417407e-01 -8.90890300e-01 -3.22348326e-01 -5.92955574e-02
1.32505773e-02 -9.58255351e-01 7.91509211e-01 -1.00314450e+00
-8.51700962e-01 1.07937479e+00 1.35312900e-01 -6.86352551e-01
7.27545738e-01 -4.19004485e-02 -5.91072381e-01 5.18191814e-01
2.08341330e-01 7.38464892e-01 5.97288132e-01 -8.98144245e-01
-5.05531788e-01 -9.23738301e-01 -1.59475282e-01 -1.09197609e-01
9.66392830e-03 2.93737262e-01 2.86182880e-01 -5.57128131e-01
3.44106019e-01 -7.36625969e-01 -6.31190121e-01 1.35537103e-01
-6.24418437e-01 2.11973879e-02 5.13505876e-01 -1.10644972e+00
8.29715312e-01 -1.69614506e+00 7.11892173e-02 3.98713768e-01
8.44266415e-01 1.67052343e-01 1.25993460e-01 -1.59656122e-01
-5.58399379e-01 2.10102394e-01 -3.00769210e-01 1.24480627e-01
-2.76570946e-01 -3.16631287e-01 3.29613566e-01 6.73277438e-01
4.46712464e-01 1.09231746e+00 -5.97116888e-01 -1.97248533e-01
2.15618238e-01 7.56830394e-01 -2.19491228e-01 -2.15806678e-01
-5.47050163e-02 8.39045763e-01 -1.47137970e-01 5.47482312e-01
4.05984879e-01 -4.71694738e-01 3.87725651e-01 -4.43437845e-01
-6.70766979e-02 1.46619052e-01 -4.97119576e-01 1.32491839e+00
-2.98333704e-01 6.81916595e-01 -8.29454064e-02 -1.19190633e+00
7.48569250e-01 3.00493240e-01 5.51109791e-01 -9.43525612e-01
2.37946928e-01 3.28576595e-01 6.45592034e-01 -5.66721976e-01
-3.80859494e-01 -2.57877231e-01 4.80375051e-01 5.35181344e-01
-2.41639819e-02 6.86658204e-01 -1.46672338e-01 -1.39663681e-01
1.37300801e+00 -1.72041103e-01 2.68935442e-01 -3.95387113e-01
3.78844202e-01 1.54141530e-01 2.76172876e-01 6.72146142e-01
-3.99763644e-01 7.31431484e-01 6.35393381e-01 -6.28815472e-01
-1.18794024e+00 -1.12703061e+00 -5.83834827e-01 6.10868812e-01
-6.81781888e-01 1.67422011e-01 -6.38952494e-01 -6.77376330e-01
-1.74596608e-01 4.97193545e-01 -9.75221753e-01 -1.72457367e-01
-5.52296817e-01 -1.19067419e+00 3.93045276e-01 5.62948883e-01
5.17153621e-01 -1.02968645e+00 -9.32784796e-01 2.32521653e-01
5.51976934e-02 -8.23240221e-01 2.17409700e-01 4.37932968e-01
-1.26182723e+00 -1.25231242e+00 -1.19848645e+00 -6.46239877e-01
6.62400544e-01 -2.91348368e-01 8.56647849e-01 -1.19628929e-01
-6.15010798e-01 -7.88037479e-03 -7.90759921e-02 -1.12696730e-01
-2.22950175e-01 1.15967989e-01 -5.10733575e-03 -2.19935942e-02
2.32109904e-01 -7.11092353e-01 -1.20909131e+00 -2.93563381e-02
-6.93847716e-01 9.89361778e-02 1.02390993e+00 6.21938527e-01
6.04353130e-01 -2.41560742e-01 6.50454700e-01 -6.55685186e-01
4.78817850e-01 -6.68963492e-01 -9.57713351e-02 3.19759220e-01
-6.86374187e-01 -1.88818142e-01 1.38689622e-01 -2.64514446e-01
-6.75941467e-01 -9.05900896e-02 -7.41083175e-02 -9.81010795e-02
-4.62509751e-01 7.94343054e-01 1.88818812e-01 5.05135357e-02
7.40841091e-01 9.90294367e-02 1.88785657e-01 -4.71460015e-01
2.44155321e-02 5.52559137e-01 4.90527272e-01 -9.56539288e-02
9.24196690e-02 4.00398135e-01 1.73951283e-01 -4.70449448e-01
-6.05611086e-01 -2.58244514e-01 -9.13094819e-01 -3.40452224e-01
1.07538486e+00 -5.76772928e-01 -5.81557095e-01 2.47325316e-01
-1.02123570e+00 -3.71042013e-01 -4.94574606e-02 7.83372581e-01
-3.16827685e-01 2.87553221e-02 -4.82868612e-01 -3.81106973e-01
-7.51748621e-01 -1.35061073e+00 6.88536346e-01 -6.98174760e-02
-4.84263957e-01 -1.15629232e+00 4.13944758e-02 4.21496034e-01
7.61995971e-01 4.08220202e-01 1.62292135e+00 -1.03855407e+00
-1.12752885e-01 -3.97737890e-01 -5.96341431e-01 1.38575926e-01
-4.41305935e-02 -2.73589969e-01 -9.60483074e-01 -1.29727259e-01
-3.82855497e-02 -2.88908146e-02 8.99460971e-01 8.23705256e-01
1.05651748e+00 -1.22077547e-01 -3.20702165e-01 3.40053529e-01
1.39684701e+00 2.03518793e-01 6.69960558e-01 7.85654902e-01
4.16468918e-01 8.13568115e-01 -8.10283795e-02 1.91766277e-01
3.07960540e-01 7.26691961e-01 5.61534882e-01 -2.37935707e-01
-1.90948531e-01 5.17503381e-01 9.81581677e-03 3.41915697e-01
-4.62947696e-01 3.85777891e-01 -1.36677325e+00 5.28982341e-01
-1.59142280e+00 -6.22932494e-01 -3.80555600e-01 2.19840813e+00
6.41868770e-01 2.19711393e-01 5.15646040e-02 7.19909184e-03
8.40112448e-01 -2.33148500e-01 -7.39300609e-01 -2.51536906e-01
-2.80088961e-01 4.16382730e-01 4.86170560e-01 7.63370395e-02
-1.13880920e+00 3.81776303e-01 6.21706200e+00 4.06602353e-01
-1.52307451e+00 5.55326521e-01 8.86043012e-01 -2.04037413e-01
4.42051217e-02 -3.17833960e-01 -2.17963189e-01 2.78073668e-01
1.06234646e+00 2.02707902e-01 1.34347230e-01 5.23045540e-01
5.03953159e-01 -3.16584408e-01 -8.96284163e-01 5.37643492e-01
-2.82248527e-01 -1.38010466e+00 3.01834643e-02 7.25768358e-02
3.60606313e-01 4.00175601e-01 1.93673223e-01 2.22856365e-03
-6.97254762e-02 -1.38702965e+00 4.71505672e-01 8.37653279e-01
9.73684311e-01 -4.47864711e-01 1.02260983e+00 1.57940388e-02
-4.77040976e-01 -6.89878613e-02 -6.43994659e-02 1.74648836e-01
-7.10605085e-02 9.59470272e-01 -1.19844782e+00 3.39348257e-01
8.66171837e-01 4.43540573e-01 -7.84539580e-01 1.26916754e+00
-9.41994935e-02 6.49280190e-01 -3.54082435e-02 4.40517783e-01
3.87265861e-01 1.47503972e-01 4.52282697e-01 1.14036250e+00
3.53206038e-01 -1.21281385e-01 -1.82999849e-01 1.08509028e+00
3.71313244e-01 2.10196450e-01 -2.19743565e-01 -2.88329199e-02
-1.21175550e-01 1.40517366e+00 -1.15788269e+00 -2.96425968e-01
-1.14125982e-01 5.67704380e-01 3.53517801e-01 2.42889062e-01
-4.40249532e-01 -1.37654632e-01 1.49987742e-01 5.41006446e-01
6.04012124e-02 -5.34370504e-02 -7.13338315e-01 -8.28745127e-01
1.71461944e-02 -7.09980190e-01 3.44078332e-01 -8.51606548e-01
-1.18387544e+00 8.35794270e-01 -2.68883020e-01 -7.68620729e-01
-1.58072919e-01 -6.94139004e-01 -6.29687309e-01 1.27493489e+00
-1.31652367e+00 -1.16795397e+00 -1.81069493e-01 3.32758933e-01
-1.95095446e-02 -2.74388313e-01 8.90429974e-01 1.99426427e-01
-6.39406681e-01 1.99754983e-01 8.13327208e-02 1.86804384e-01
5.02220511e-01 -1.20214999e+00 -3.13372328e-03 4.11587030e-01
-4.79906678e-01 8.29141140e-01 4.17455196e-01 -8.42651725e-01
-6.05023921e-01 -1.10314715e+00 8.49978209e-01 -5.87771684e-02
8.52377415e-01 2.26041570e-01 -9.74298000e-01 5.94520450e-01
-1.27030566e-01 2.95307487e-02 7.57727087e-01 2.95996424e-02
-3.17721069e-01 2.80059520e-02 -1.32087660e+00 4.65967059e-01
6.08925521e-01 -4.48873758e-01 -5.58225811e-01 4.77344096e-01
2.73593783e-01 2.73047555e-02 -1.15767419e+00 7.14301348e-01
7.04081297e-01 -1.00639212e+00 1.02786076e+00 -9.56891775e-01
6.94091260e-01 1.06109917e-01 -2.02722270e-02 -1.25829720e+00
-4.58379835e-01 2.54895985e-01 2.89094122e-03 6.21407628e-01
5.58006227e-01 -6.81471050e-01 5.81386745e-01 5.51643014e-01
-2.59863704e-01 -1.12764323e+00 -1.07567835e+00 -4.55437690e-01
3.13569486e-01 -5.39869964e-01 1.69911608e-01 8.67042422e-01
-1.40340298e-01 -3.46593969e-02 2.01502532e-01 2.53653020e-01
6.06538117e-01 -1.77581787e-01 -2.26310775e-01 -1.34036219e+00
-1.10024586e-01 -7.29809999e-01 -6.50087893e-01 1.42729506e-01
-1.78614452e-01 -1.38791621e+00 -3.94486159e-01 -1.90904093e+00
5.46492994e-01 -6.77594006e-01 -7.73835480e-01 6.20119631e-01
-1.86789948e-02 5.03560185e-01 5.94334938e-02 5.05471349e-01
-2.90576607e-01 -7.08762258e-02 1.31508482e+00 -2.68359303e-01
-7.97164291e-02 5.96044250e-02 -5.22111416e-01 7.36824155e-01
1.06673789e+00 -4.52173620e-01 -8.21226463e-02 -3.48868608e-01
1.16034925e-01 1.99786853e-02 1.07272696e+00 -1.21870339e+00
-2.40432024e-02 2.39913061e-01 9.96171772e-01 -3.29170495e-01
1.27477080e-01 -5.69961190e-01 1.75609499e-01 9.38362002e-01
-4.60386306e-01 -1.64441288e-01 8.23583379e-02 1.27780363e-01
1.03440419e-01 -1.78779438e-01 7.29874611e-01 -3.97405714e-01
-5.54928601e-01 3.26589674e-01 -5.78816414e-01 -2.43313178e-01
8.84323180e-01 -2.20395833e-01 -4.17805493e-01 -1.41722662e-02
-1.44429958e+00 -1.31581903e-01 -8.57738331e-02 2.00354204e-01
5.86837649e-01 -1.09128094e+00 -7.89734066e-01 -1.07703015e-01
-6.19945265e-02 -4.81067002e-01 6.49152875e-01 1.63296282e+00
-6.67980134e-01 7.59900153e-01 -7.67344236e-01 -8.87581408e-01
-1.06192493e+00 1.94823563e-01 6.27109230e-01 -4.00075138e-01
-8.07867467e-01 5.31588435e-01 2.03684475e-02 -3.82693201e-01
2.26836756e-01 -3.82663697e-01 -6.10388577e-01 2.99420565e-01
7.72170782e-01 3.01214665e-01 4.75615233e-01 -5.61200440e-01
-4.49353099e-01 3.30103278e-01 -3.96346629e-01 -1.08020507e-01
1.60619354e+00 2.15942506e-02 -3.62653762e-01 3.83809000e-01
1.03867853e+00 -6.53378010e-01 -8.78786385e-01 -2.01906964e-01
1.89472735e-01 1.85547382e-01 4.66296136e-01 -1.47818780e+00
-1.31007612e+00 1.15347600e+00 1.38258648e+00 -5.11626191e-02
8.56229484e-01 8.80630016e-02 4.69598144e-01 7.19738603e-02
3.70962828e-01 -6.71893895e-01 -3.07266206e-01 2.17707410e-01
1.08379531e+00 -1.24717700e+00 -6.82829618e-02 -2.79270951e-02
-5.99909842e-01 1.44947922e+00 3.48865449e-01 4.77005318e-02
6.85008466e-01 -3.91690545e-02 2.26375520e-01 -6.57553494e-01
-4.77823347e-01 -1.69342503e-01 5.12437642e-01 1.00742507e+00
4.21582013e-01 3.68080914e-01 -3.69388163e-01 9.48470294e-01
-3.36492993e-02 3.16343755e-01 1.26279294e-01 7.24426806e-01
-5.81018448e-01 -6.58361197e-01 -3.21714401e-01 1.02736425e+00
-5.91924906e-01 -2.29783505e-01 -2.59890676e-01 6.65654957e-01
4.54685479e-01 4.77354079e-01 2.72983879e-01 -3.69668275e-01
1.06025249e-01 2.10994318e-01 4.43142503e-01 -5.44895172e-01
-7.65912116e-01 -1.41981825e-01 4.76009287e-02 -3.90670061e-01
-2.85611510e-01 -7.04571426e-01 -1.30746388e+00 -7.17583895e-02
1.11056745e-01 -2.42407411e-01 8.16906452e-01 1.11477745e+00
3.83524686e-01 9.17822361e-01 8.30048695e-02 -8.64754975e-01
-2.73584843e-01 -1.16650867e+00 -5.88494182e-01 -5.23649110e-03
1.50858924e-01 -7.52949655e-01 -5.80960400e-02 -1.97382882e-01] | [14.210542678833008, -1.8864550590515137] |
6facb969-7da9-4900-a68e-cd389b45ed7a | directionally-constrained-fully-convolutional | 1908.06673 | null | https://arxiv.org/abs/1908.06673v1 | https://arxiv.org/pdf/1908.06673v1.pdf | Directionally Constrained Fully Convolutional Neural Network For Airborne Lidar Point Cloud Classification | Point cloud classification plays an important role in a wide range of airborne light detection and ranging (LiDAR) applications, such as topographic mapping, forest monitoring, power line detection, and road detection. However, due to the sensor noise, high redundancy, incompleteness, and complexity of airborne LiDAR systems, point cloud classification is challenging. In this paper, we proposed a directionally constrained fully convolutional neural network (D-FCN) that can take the original 3D coordinates and LiDAR intensity as input; thus, it can directly apply to unstructured 3D point clouds for semantic labeling. Specifically, we first introduce a novel directionally constrained point convolution (D-Conv) module to extract locally representative features of 3D point sets from the projected 2D receptive fields. To make full use of the orientation information of neighborhood points, the proposed D-Conv module performs convolution in an orientation-aware manner by using a directionally constrained nearest neighborhood search. Then, we designed a multiscale fully convolutional neural network with downsampling and upsampling blocks to enable multiscale point feature learning. The proposed D-FCN model can therefore process input point cloud with arbitrary sizes and directly predict the semantic labels for all the input points in an end-to-end manner. Without involving additional geometry features as input, the proposed method has demonstrated superior performance on the International Society for Photogrammetry and Remote Sensing (ISPRS) 3D labeling benchmark dataset. The results show that our model has achieved a new state-of-the-art level of performance with an average F1 score of 70.7%, and it has improved the performance by a large margin on categories with a small number of points (such as powerline, car, and facade). | ['Tianhe Chi', 'Lina Yang', 'Congcong Wen', 'Xiang Li', 'Ling Peng'] | 2019-08-19 | null | null | null | null | ['line-detection'] | ['computer-vision'] | [ 2.19647348e-01 -4.13005859e-01 -2.86215581e-02 -6.46916032e-01
-5.14878809e-01 -4.28266138e-01 5.21890521e-01 8.95739347e-02
-4.54230309e-01 2.50353038e-01 -3.41102004e-01 -5.10968089e-01
-2.44533777e-01 -1.27973545e+00 -8.20050061e-01 -3.34757745e-01
-6.86371103e-02 3.75106812e-01 2.78187126e-01 -1.06708907e-01
2.25672230e-01 1.45965862e+00 -1.90574431e+00 3.42460051e-02
9.27132249e-01 1.34220874e+00 4.16620702e-01 1.91451654e-01
-4.12109107e-01 -2.22061817e-02 -1.25687018e-01 9.07816961e-02
4.33691919e-01 5.20384967e-01 -4.42529470e-01 1.39404550e-01
7.96968043e-01 -4.84022349e-01 -1.25806019e-01 9.35857296e-01
4.27865028e-01 2.12582529e-01 6.53708816e-01 -1.05483258e+00
-6.85646951e-01 1.20659508e-02 -8.22591603e-01 1.78746588e-03
-2.65643984e-01 5.36111519e-02 8.62050116e-01 -1.43310738e+00
1.98955223e-01 1.37225699e+00 7.42015541e-01 1.50293127e-01
-1.06225133e+00 -1.00804460e+00 2.27348417e-01 9.06402990e-02
-1.84875405e+00 2.81149186e-02 9.31392074e-01 -3.62167716e-01
9.54976737e-01 1.76972002e-01 6.37859941e-01 2.45918453e-01
-2.46933207e-01 4.33702916e-01 1.04964137e+00 -1.81963712e-01
2.85890311e-01 -1.87421381e-01 2.10048556e-02 7.02495575e-01
1.30492672e-01 1.78158268e-01 -1.43105969e-01 -4.40372527e-03
8.86639893e-01 4.66179848e-01 -3.66636254e-02 -3.32475275e-01
-1.01496053e+00 8.21838021e-01 1.22709537e+00 1.05247431e-01
-4.35754359e-01 1.98341176e-01 2.42215078e-02 -9.89826918e-02
5.55179000e-01 -1.23076670e-01 -4.16679651e-01 4.55725163e-01
-8.65453720e-01 1.09780736e-01 2.08017722e-01 1.02469909e+00
1.25150847e+00 -1.16478989e-03 9.28404331e-02 9.22280848e-01
2.89073259e-01 9.92064893e-01 -9.10856724e-02 -9.21404362e-01
5.52972794e-01 9.37618494e-01 -3.00314017e-02 -1.05332029e+00
-5.23550510e-01 -4.35237616e-01 -9.96846139e-01 5.12641072e-01
-1.47962958e-01 2.88772225e-01 -1.19002056e+00 1.38318396e+00
4.93518353e-01 2.57933617e-01 -1.60686806e-01 9.64002073e-01
8.81118953e-01 7.13115275e-01 1.12318441e-01 2.06137121e-01
1.24018407e+00 -3.87334108e-01 -6.85830861e-02 -3.42115104e-01
3.59141916e-01 -5.49523175e-01 1.18380380e+00 -1.88001856e-01
-6.36899114e-01 -7.79738843e-01 -1.11187780e+00 -2.97726154e-01
-5.87037385e-01 3.94354731e-01 6.40171885e-01 3.72514397e-01
-8.32517803e-01 5.75426400e-01 -7.69452214e-01 -4.11680609e-01
9.52691436e-01 4.46652025e-01 -2.52275437e-01 -3.53001505e-01
-7.38023520e-01 6.20028138e-01 1.28118366e-01 3.67341727e-01
-4.38875526e-01 -8.01594377e-01 -7.78131783e-01 3.10101509e-01
1.21157341e-01 -5.70150733e-01 8.44590664e-01 -1.91544324e-01
-1.07355022e+00 7.92959630e-01 -2.67610699e-01 -1.77984849e-01
2.73249447e-01 -1.69451460e-01 -3.36267978e-01 6.32452369e-02
3.80785108e-01 1.10034299e+00 6.46615326e-01 -1.17092288e+00
-9.89113271e-01 -8.22641373e-01 -5.84373251e-02 3.06163043e-01
-2.23200321e-01 -3.59907866e-01 -2.21138269e-01 -3.73858929e-01
6.33270681e-01 -1.01986837e+00 -1.63799718e-01 4.76801544e-01
-1.23955615e-01 -4.08290207e-01 1.15342557e+00 -4.66500409e-02
6.90406740e-01 -2.39109945e+00 -4.44626391e-01 2.75291651e-01
-1.19674066e-02 2.42734104e-01 -5.51063456e-02 6.13223836e-02
-2.09908877e-02 2.82438278e-01 -5.61629891e-01 -1.38678700e-01
-1.11387409e-01 2.46062919e-01 -2.95940101e-01 4.68489468e-01
4.93975401e-01 7.68758118e-01 -6.59048855e-01 -3.24879736e-01
7.64560997e-01 6.71199560e-01 -4.10662204e-01 -1.46160349e-01
-1.48385495e-01 1.18748918e-01 -5.06361127e-01 1.02598846e+00
1.24540496e+00 -1.28824338e-01 -5.15545785e-01 -4.14869100e-01
-5.13863981e-01 -1.91335808e-02 -1.17880177e+00 1.67765474e+00
-8.18426728e-01 6.19630218e-01 -9.76557005e-03 -6.32468283e-01
1.34275973e+00 -1.73182100e-01 5.13957202e-01 -4.66800153e-01
-1.42675534e-01 4.00348186e-01 -2.43577406e-01 -2.53672022e-02
3.22048396e-01 -1.28075957e-01 7.33932629e-02 5.83712049e-02
-2.85071224e-01 -4.20992404e-01 -3.70563298e-01 -2.92685419e-01
7.94951141e-01 -1.47948042e-01 5.77440113e-02 -1.68865100e-01
6.49339199e-01 3.19335818e-01 3.32727909e-01 4.89905864e-01
1.06635161e-01 6.21187687e-01 -2.87286729e-01 -7.53524065e-01
-9.58105505e-01 -1.14513004e+00 -5.59030831e-01 7.04099119e-01
2.38087535e-01 1.78749472e-01 -2.36365482e-01 -3.42579901e-01
4.81629461e-01 5.95816791e-01 -2.39055216e-01 4.68788743e-02
-6.49805248e-01 -4.58862126e-01 4.07290399e-01 7.65809059e-01
1.04786241e+00 -8.35239291e-01 -5.82964301e-01 4.58197035e-02
2.41588920e-01 -1.35401094e+00 -1.24981664e-01 2.23508224e-01
-1.17247593e+00 -8.50731671e-01 -3.84921908e-01 -9.07646418e-01
6.81800306e-01 7.64927685e-01 6.83338761e-01 -9.20740217e-02
-1.66760266e-01 -1.33235365e-01 -1.99183244e-02 -5.12855947e-01
3.06997895e-01 1.67016387e-01 -8.95660371e-02 -3.91454138e-02
8.30485642e-01 -7.66964257e-01 -6.94503367e-01 4.02533174e-01
-7.66280234e-01 -1.05905160e-01 8.10484707e-01 6.41563773e-01
8.88583660e-01 1.75713286e-01 3.98741812e-01 -6.00055933e-01
2.61302859e-01 -1.64190888e-01 -8.83019626e-01 -1.68850571e-01
-5.20580947e-01 -3.39688271e-01 5.73524833e-01 2.29595718e-03
-7.89258242e-01 4.86620754e-01 -2.83062696e-01 -7.48674512e-01
-6.28312886e-01 4.12492424e-01 -2.52229065e-01 -3.79563332e-01
5.92020154e-01 2.45514572e-01 -1.74466446e-01 -5.77929258e-01
4.52451408e-01 1.02745831e+00 5.90928376e-01 -3.74020636e-01
1.14236796e+00 8.92556489e-01 5.42484343e-01 -1.09962738e+00
-6.42354131e-01 -5.33317924e-01 -9.03510690e-01 -6.81255162e-02
7.18703151e-01 -1.19054604e+00 -7.97597229e-01 3.29160392e-01
-1.18468690e+00 1.35995761e-01 -3.31579596e-01 5.40153921e-01
-2.08543405e-01 1.67975843e-01 -1.51071638e-01 -6.82175279e-01
-4.50228989e-01 -1.02114975e+00 1.36812747e+00 2.08280370e-01
3.21439356e-01 -4.68759924e-01 -4.63622600e-01 2.16400102e-01
3.61994654e-01 2.81308681e-01 1.29680657e+00 -3.29575151e-01
-9.59713638e-01 -3.50023210e-01 -7.41398931e-01 4.85844195e-01
2.80019820e-01 -1.50907189e-01 -1.07057667e+00 -6.59263954e-02
-1.97735310e-01 7.15590268e-02 8.44683766e-01 3.74579430e-01
1.44722271e+00 1.42218947e-01 -3.84871721e-01 8.93441200e-01
1.67028463e+00 -1.36891017e-02 2.51171380e-01 2.23819405e-01
8.58057857e-01 4.25995737e-01 6.67979538e-01 2.45223850e-01
4.26302463e-01 4.18762743e-01 9.06570613e-01 -1.83542520e-01
-5.06417938e-02 -3.60690236e-01 -2.27164909e-01 2.85034180e-01
-1.21111728e-01 6.74898848e-02 -9.78623688e-01 6.43204093e-01
-1.57899702e+00 -7.07837880e-01 -2.56477296e-01 2.17893195e+00
1.99366972e-01 5.59717268e-02 -2.59467542e-01 2.06934437e-01
9.39379394e-01 1.28406122e-01 -7.61503220e-01 -2.71320462e-01
2.06268076e-02 4.82442051e-01 8.73663127e-01 2.91597277e-01
-1.16732240e+00 1.04064572e+00 5.08562613e+00 8.84349406e-01
-1.34314311e+00 -1.01657219e-01 2.41321430e-01 6.97890073e-02
-1.95531204e-01 -9.72751081e-02 -9.78365302e-01 3.22190315e-01
4.75343734e-01 2.02455416e-01 2.77077854e-01 1.04413354e+00
3.54099751e-01 1.87758937e-01 -8.41216445e-01 1.10177565e+00
-3.76642078e-01 -1.49641347e+00 3.12380791e-01 3.02640498e-01
6.03921056e-01 5.11782289e-01 4.43544053e-02 2.12652430e-01
-5.92447631e-02 -1.01096213e+00 5.34265459e-01 1.46966800e-01
1.11971354e+00 -9.07656968e-01 5.67797363e-01 5.77359200e-01
-1.49255085e+00 -3.36153239e-01 -8.24309587e-01 -1.24983557e-01
9.71929803e-02 8.82560313e-01 -8.49327862e-01 4.68244165e-01
8.55678260e-01 7.19665885e-01 -4.06728238e-01 9.79774058e-01
-1.65972367e-01 3.61176580e-01 -8.38700652e-01 2.64665224e-02
4.44146484e-01 -3.30153853e-01 3.56288910e-01 9.57042933e-01
6.00857973e-01 3.65312785e-01 2.72506207e-01 9.87439573e-01
-2.65560389e-01 1.81001499e-02 -8.13064814e-01 4.17307973e-01
9.48471963e-01 1.33969820e+00 -7.20175564e-01 -1.92638449e-02
-4.68363494e-01 4.15571094e-01 2.72814244e-01 1.46406353e-01
-5.87146699e-01 -5.62744081e-01 7.21717238e-01 2.90614545e-01
5.34932613e-01 -5.14556170e-01 -6.65717959e-01 -6.78538799e-01
1.23457558e-01 -1.34282604e-01 5.28317392e-02 -8.19118798e-01
-1.18953228e+00 5.12685955e-01 1.67244803e-02 -1.44957602e+00
1.63913876e-01 -7.78785884e-01 -6.01724625e-01 1.14413440e+00
-2.06787038e+00 -1.43353689e+00 -8.82968605e-01 5.78451633e-01
5.52993715e-01 -3.59188281e-02 5.94138324e-01 4.06590998e-01
-1.26329944e-01 2.64663640e-02 -6.21387772e-02 1.75419211e-01
2.33028188e-01 -1.04716980e+00 6.02424562e-01 6.08782172e-01
2.24573333e-02 3.81886631e-01 -8.28918964e-02 -5.14019251e-01
-1.28021228e+00 -1.72469211e+00 6.78236246e-01 3.30922678e-02
3.10113192e-01 -2.71651030e-01 -8.46990407e-01 3.69347066e-01
-5.42360663e-01 4.80753839e-01 4.23559606e-01 -1.70776561e-01
-1.17383450e-01 -5.40260732e-01 -1.37735653e+00 3.72492135e-01
1.37875187e+00 -5.40962338e-01 -3.71805876e-01 4.59804058e-01
7.47163773e-01 -3.56995106e-01 -8.76207888e-01 9.69430864e-01
3.69619161e-01 -7.26481378e-01 1.23303711e+00 -1.22627065e-01
4.10441220e-01 -6.74710214e-01 -6.04515314e-01 -1.02723157e+00
-6.16012394e-01 2.27094293e-01 2.57394463e-01 1.08340371e+00
1.89703032e-01 -7.00794637e-01 8.91807973e-01 2.25938946e-01
-4.94277596e-01 -8.26653302e-01 -1.17012239e+00 -7.60991037e-01
-3.26750167e-02 -6.73639238e-01 1.04760551e+00 6.89392447e-01
-8.56866956e-01 3.67445767e-01 2.84914732e-01 7.72249699e-01
7.29934454e-01 6.86921000e-01 6.89397216e-01 -1.68314826e+00
5.13010859e-01 -4.49351907e-01 -6.61364257e-01 -1.36062634e+00
2.67121106e-01 -1.16337276e+00 -1.38158709e-01 -1.64487994e+00
-3.49969178e-01 -9.30515528e-01 -5.43771349e-02 4.74043101e-01
1.14634790e-01 4.66393054e-01 1.26172453e-01 4.51181173e-01
1.77974120e-01 6.98797047e-01 1.07510555e+00 -3.86468828e-01
-1.64874166e-01 2.15078115e-01 -3.92894715e-01 7.51038313e-01
7.21328259e-01 -3.69284302e-01 -2.60821640e-01 -8.47439885e-01
-4.42621596e-02 -3.22288632e-01 7.23414719e-01 -1.28086305e+00
2.76929200e-01 -1.40522063e-01 6.75695240e-01 -1.37570524e+00
4.69007701e-01 -1.31044090e+00 -4.10220958e-02 3.24308991e-01
6.05573915e-02 -4.69819419e-02 2.50115454e-01 5.79438210e-01
-1.75853044e-01 -6.92386106e-02 8.50611031e-01 -1.75641045e-01
-9.24944997e-01 6.47648156e-01 6.32922873e-02 -4.70139503e-01
9.46352661e-01 -5.26884377e-01 -2.56805867e-01 7.53079355e-02
-4.09454077e-01 3.66247565e-01 5.62568665e-01 3.54022503e-01
9.70449924e-01 -1.44183373e+00 -6.53793573e-01 5.48664749e-01
2.24118635e-01 8.45111430e-01 1.58316061e-01 2.97965556e-01
-7.47150838e-01 4.66939032e-01 -1.50085434e-01 -1.24514318e+00
-9.47874427e-01 3.35277587e-01 1.96851879e-01 4.52762634e-01
-7.83189595e-01 5.80036342e-01 -1.67535450e-02 -7.77366042e-01
-4.38435972e-02 -6.65682673e-01 -1.28039747e-01 -1.77167222e-01
1.60869598e-01 4.04494971e-01 4.07590926e-01 -8.14681113e-01
-5.61465442e-01 1.15261555e+00 2.79581726e-01 3.34966242e-01
1.47379327e+00 -1.90907507e-04 1.06707467e-02 1.70164526e-01
1.22038221e+00 -2.42147341e-01 -1.11764324e+00 -3.71836215e-01
-2.25516349e-01 -7.83581257e-01 4.96762156e-01 -4.62987363e-01
-1.14344406e+00 1.19862711e+00 9.90315735e-01 -3.02262474e-02
9.87847328e-01 -1.00006594e-03 7.31220722e-01 6.06878638e-01
4.52491969e-01 -7.08679259e-01 -5.50186932e-01 5.60239136e-01
6.50503993e-01 -1.30198753e+00 1.45281747e-01 -6.84571981e-01
-1.39017761e-01 1.16809225e+00 5.94881773e-01 -2.29253590e-01
9.06181753e-01 5.31638786e-02 -5.50162233e-02 -3.88752282e-01
-2.63819516e-01 -4.62289184e-01 2.45412648e-01 6.49147928e-01
4.57170121e-02 2.32615694e-01 1.29772052e-01 9.52303857e-02
-2.06630692e-01 1.11854270e-01 4.78321761e-02 8.81925046e-01
-8.13574672e-01 -5.49510717e-01 -4.26684111e-01 6.01360619e-01
2.00315624e-01 -6.46650791e-02 -5.14024422e-02 8.28229904e-01
3.66946369e-01 7.62090683e-01 4.99145597e-01 -4.23740953e-01
6.75425828e-01 -1.35737106e-01 1.43357694e-01 -6.96638763e-01
-1.73552275e-01 -1.30753905e-01 -3.72472793e-01 -3.05825055e-01
-3.83670837e-01 -4.22337234e-01 -1.46653461e+00 -2.99782753e-01
-3.37031692e-01 -1.15456700e-01 1.08990276e+00 8.01010191e-01
6.11803532e-01 2.17882574e-01 9.44854558e-01 -1.19551134e+00
-5.33749580e-01 -9.14254725e-01 -5.81930578e-01 1.48193896e-01
2.65427977e-01 -7.36883223e-01 -2.14200541e-01 -3.86082977e-01] | [8.02676010131836, -3.026393413543701] |
0f457811-bf99-4f48-b2d1-68af4d845322 | mpc-bert-a-pre-trained-language-model-for | 2106.01541 | null | https://arxiv.org/abs/2106.01541v1 | https://arxiv.org/pdf/2106.01541v1.pdf | MPC-BERT: A Pre-Trained Language Model for Multi-Party Conversation Understanding | Recently, various neural models for multi-party conversation (MPC) have achieved impressive improvements on a variety of tasks such as addressee recognition, speaker identification and response prediction. However, these existing methods on MPC usually represent interlocutors and utterances individually and ignore the inherent complicated structure in MPC which may provide crucial interlocutor and utterance semantics and would enhance the conversation understanding process. To this end, we present MPC-BERT, a pre-trained model for MPC understanding that considers learning who says what to whom in a unified model with several elaborated self-supervised tasks. Particularly, these tasks can be generally categorized into (1) interlocutor structure modeling including reply-to utterance recognition, identical speaker searching and pointer consistency distinction, and (2) utterance semantics modeling including masked shared utterance restoration and shared node detection. We evaluate MPC-BERT on three downstream tasks including addressee recognition, speaker identification and response selection. Experimental results show that MPC-BERT outperforms previous methods by large margins and achieves new state-of-the-art performance on all three downstream tasks at two benchmarks. | ['Daxin Jiang', 'Xiubo Geng', 'Can Xu', 'Zhen-Hua Ling', 'Chongyang Tao', 'Jia-Chen Gu'] | 2021-06-03 | null | https://aclanthology.org/2021.acl-long.285 | https://aclanthology.org/2021.acl-long.285.pdf | acl-2021-5 | ['speaker-identification'] | ['speech'] | [ 2.88837522e-01 3.51239532e-01 -3.15098703e-01 -9.67053533e-01
-1.38665235e+00 -4.88109231e-01 4.89014238e-01 -1.18953504e-01
2.54683018e-01 3.44224989e-01 8.94203067e-01 -3.79162133e-01
2.54143566e-01 4.01512720e-02 -2.13037997e-01 -7.55895078e-01
-6.34990409e-02 6.87277913e-01 -4.22705570e-03 -2.84502625e-01
1.70586690e-01 -6.86800405e-02 -1.35127366e+00 9.00756180e-01
6.32148445e-01 1.07406437e+00 1.56661309e-02 8.74765396e-01
-2.44629592e-01 1.24301457e+00 -5.51369190e-01 -5.06479859e-01
-5.08194506e-01 -4.73133057e-01 -1.43280089e+00 1.24199316e-01
9.03274938e-02 -3.92132461e-01 -2.08364353e-01 6.08246148e-01
8.15423727e-01 2.65860289e-01 5.77156126e-01 -1.43555474e+00
-5.91604114e-01 1.41056383e+00 -3.45050126e-01 2.04239964e-01
6.29223824e-01 -3.81301623e-03 1.23822153e+00 -7.61487424e-01
3.58115397e-02 1.99219084e+00 6.98757350e-01 7.65980780e-01
-1.08956254e+00 -6.27286792e-01 4.33140099e-01 5.63141584e-01
-1.09800696e+00 -1.22300148e+00 7.90778220e-01 -7.80973062e-02
1.08248329e+00 6.72265947e-01 -3.65008056e-01 1.23738945e+00
-3.70774090e-01 1.43298960e+00 8.34937036e-01 -2.85600990e-01
-7.43104592e-02 3.70738566e-01 6.85809255e-01 2.75087655e-01
-1.21130872e+00 -2.59511173e-01 -9.47961569e-01 -6.03866339e-01
8.58120096e-04 -6.23286963e-01 -4.58150744e-01 2.81952053e-01
-1.07156742e+00 1.07955253e+00 1.64527684e-01 -2.48920284e-02
-3.89459357e-02 -1.62787423e-01 7.90058196e-01 6.24623358e-01
5.94174027e-01 -1.18231319e-01 -5.77027678e-01 -3.17792684e-01
-3.36560249e-01 2.43876856e-02 1.41235125e+00 9.80848610e-01
5.02737582e-01 -1.02348626e-01 -3.09838146e-01 1.44587207e+00
5.16178906e-01 1.58656850e-01 6.93805039e-01 -9.24084067e-01
6.21743977e-01 4.13194925e-01 -2.53572643e-01 -7.24829018e-01
-6.13994777e-01 -1.60855114e-01 -7.95041978e-01 -7.34735310e-01
2.51288831e-01 -2.54870355e-01 -2.21610457e-01 1.99054122e+00
5.38431108e-01 3.56874436e-01 4.44953322e-01 7.53639460e-01
1.69841039e+00 7.47696757e-01 -1.00403517e-01 -3.84682566e-01
1.39343369e+00 -1.67256749e+00 -8.46138179e-01 -2.39092499e-01
9.23061728e-01 -7.91783214e-01 8.15557122e-01 -2.02233531e-02
-9.78283584e-01 -3.99077237e-01 -4.39135671e-01 -1.02592520e-01
8.72659981e-02 -7.98534676e-02 6.00839496e-01 9.34755147e-01
-1.16568375e+00 -2.20349859e-02 -3.37515622e-01 -2.80489177e-01
-8.85673333e-03 5.62941790e-01 -2.21486643e-01 1.32948771e-01
-1.41144240e+00 7.87249982e-01 -1.87595636e-01 1.90239117e-01
-8.64819527e-01 -5.66171169e-01 -9.52459633e-01 3.36535543e-01
2.04423651e-01 -3.95684063e-01 1.81255198e+00 -9.79631126e-01
-2.08745670e+00 1.06579220e+00 -9.30510223e-01 -4.55433786e-01
1.23466797e-01 3.99853066e-02 -6.55569971e-01 -4.95464653e-02
-1.69373348e-01 6.27328277e-01 9.57744658e-01 -1.20318604e+00
-6.79421008e-01 -4.32146907e-01 -2.01018751e-02 4.21428978e-01
-9.92804989e-02 6.10921204e-01 -3.22776139e-01 -1.74710080e-01
3.50978941e-01 -8.16461146e-01 3.39368939e-01 -8.56625915e-01
-7.89806664e-01 -9.02773321e-01 8.86468112e-01 -7.42632747e-01
1.09048736e+00 -2.11479425e+00 1.74845189e-01 -1.11811034e-01
2.63755828e-01 -1.65173829e-01 -4.87764806e-01 6.52109027e-01
-9.10804123e-02 -1.18985541e-01 5.75561523e-02 -1.07553971e+00
4.25586551e-01 -1.71648428e-01 -5.81521690e-01 5.07631958e-01
-1.27262235e-01 8.31447840e-01 -4.64330941e-01 -3.02727759e-01
-6.18848167e-02 2.94328570e-01 -4.96575743e-01 7.39618838e-01
-1.19327791e-01 7.46769488e-01 -2.32453406e-01 7.08623350e-01
6.34118557e-01 -3.75467271e-01 4.10688788e-01 1.26357660e-01
3.19923788e-01 9.92871821e-01 -8.99067461e-01 1.33777297e+00
-5.57830870e-01 6.90159261e-01 8.18140030e-01 -9.52192247e-01
1.01026952e+00 7.24749446e-01 1.50621682e-01 -2.34552428e-01
1.84526816e-01 4.04227562e-02 1.40345424e-01 -5.24691999e-01
6.03937089e-01 -3.48631591e-02 -3.07487816e-01 1.08085191e+00
-3.31518948e-02 2.82783151e-01 -6.94951117e-01 3.12819600e-01
8.28417778e-01 -5.87158144e-01 3.06009889e-01 -1.71762317e-01
1.13196468e+00 -5.97677052e-01 5.32817900e-01 8.16088259e-01
-7.12171257e-01 4.40373868e-01 6.25474632e-01 -8.81977230e-02
-1.42009750e-01 -7.47671664e-01 -4.25674953e-02 2.13015699e+00
1.70938969e-01 -8.62848163e-02 -7.23865330e-01 -6.94146633e-01
-3.22665945e-02 7.99544156e-01 -3.41038883e-01 -1.33317202e-01
-7.42097616e-01 -6.16782069e-01 8.41379642e-01 4.23367292e-01
3.43650013e-01 -1.16372705e+00 3.56695503e-01 2.52594352e-01
-8.10306787e-01 -1.22895598e+00 -9.34053481e-01 8.36435631e-02
-4.42137480e-01 -7.91018844e-01 -2.88491905e-01 -1.24653971e+00
2.18657702e-01 5.52727520e-01 1.12076294e+00 1.60333216e-01
3.13950568e-01 4.97408181e-01 -4.67310160e-01 -9.25351307e-02
-8.12320948e-01 3.12214941e-01 7.28478581e-02 5.13299048e-01
6.21373773e-01 -5.35919726e-01 -2.78015703e-01 8.96126330e-01
-2.42968827e-01 -1.51765183e-01 1.06442466e-01 1.00537729e+00
-2.05010906e-01 -5.27211130e-01 1.44393063e+00 -9.28179920e-01
8.74136508e-01 -6.68800890e-01 1.75963446e-01 5.21584272e-01
-4.95888405e-02 -2.62232631e-01 3.64183635e-01 -4.42280769e-01
-1.40060866e+00 -1.98131055e-01 -5.70659220e-01 1.68122888e-01
-4.77646440e-01 2.34982580e-01 -5.35188854e-01 3.44489352e-03
1.94827974e-01 5.95856965e-01 1.63202524e-01 -5.89070082e-01
4.76254642e-01 1.48455310e+00 5.36261857e-01 -5.32954037e-01
7.18703046e-02 2.34802842e-01 -1.02042961e+00 -8.32482457e-01
-7.73360550e-01 -9.37529802e-01 -4.86284465e-01 -1.22125596e-01
6.46136045e-01 -9.54430342e-01 -1.42251205e+00 6.99454725e-01
-1.42026174e+00 -3.73779923e-01 5.53280890e-01 2.73160666e-01
-6.76660061e-01 5.35718620e-01 -1.05088413e+00 -1.26015747e+00
-5.78025639e-01 -1.40868187e+00 1.09183145e+00 -2.09409781e-02
-6.36798203e-01 -1.13888597e+00 -2.06834331e-01 1.05983865e+00
6.44941807e-01 -6.90616250e-01 1.16744459e+00 -1.32394695e+00
-3.58651578e-01 1.33712158e-01 -2.58918941e-01 1.67704150e-01
-2.73042899e-02 -5.34877717e-01 -1.60352349e+00 -2.42189080e-01
4.53881174e-01 -8.19720387e-01 8.00690234e-01 4.19863135e-01
1.12793207e+00 -5.43000996e-01 -4.38963711e-01 4.70642090e-01
4.32849228e-01 1.03259444e-01 3.02844852e-01 -1.15271837e-01
5.85059285e-01 1.37758374e+00 1.34487078e-01 2.91417986e-01
1.13856339e+00 8.00737381e-01 4.53920543e-01 3.12791497e-01
1.00867912e-01 -1.38344765e-01 6.97695673e-01 1.21570468e+00
7.06454813e-01 -5.59262037e-01 -6.28490627e-01 5.70488334e-01
-2.23988748e+00 -8.40315282e-01 -1.46046653e-01 1.67941988e+00
8.07655036e-01 -3.93913090e-01 3.58892620e-01 -1.91422448e-01
1.05153024e+00 6.01194620e-01 -6.57805741e-01 -5.99083841e-01
-1.60516366e-01 -6.52350903e-01 -1.94057032e-01 8.80223274e-01
-1.02061641e+00 9.89114642e-01 6.34954691e+00 7.99622893e-01
-9.03303206e-01 4.76454586e-01 9.80067253e-01 7.55413398e-02
-2.00990915e-01 -2.12112948e-01 -1.32214308e+00 3.74401271e-01
1.11152482e+00 4.16189097e-02 4.33744371e-01 8.29558611e-01
1.12457313e-01 1.81723759e-01 -1.65077043e+00 1.17279696e+00
4.76491421e-01 -1.22426713e+00 -3.10080945e-01 -2.86994517e-01
4.22643363e-01 -2.01697014e-02 5.39381504e-02 7.84413636e-01
4.27400500e-01 -1.15732253e+00 4.33822870e-01 1.05028413e-02
1.28797203e-01 -6.40249372e-01 8.12743843e-01 6.27408803e-01
-1.08486950e+00 -2.28590727e-01 -1.63645387e-01 1.59961712e-02
5.51454306e-01 -3.94520648e-02 -1.03274906e+00 2.44463667e-01
4.11942452e-01 4.97311682e-01 1.23846820e-02 5.10642469e-01
-3.29915732e-02 8.66436243e-01 -9.87429097e-02 -1.71174601e-01
2.74799615e-01 1.82969451e-01 6.62953019e-01 1.35116518e+00
-2.75174916e-01 1.25865236e-01 2.04453215e-01 6.82634652e-01
-3.48373830e-01 2.12618902e-01 -2.93615907e-02 3.73844326e-01
8.97687018e-01 1.03153026e+00 -9.24848914e-02 -3.49595726e-01
-4.25244987e-01 8.78150761e-01 5.16100705e-01 3.78828794e-01
-6.51441038e-01 1.11018255e-01 9.60505426e-01 -4.93354142e-01
1.47488296e-01 2.95223534e-01 -2.15882584e-01 -9.51980233e-01
-1.21926069e-01 -1.18534803e+00 7.21169829e-01 -2.25752518e-01
-1.55288911e+00 6.91170037e-01 -3.21219176e-01 -7.47374475e-01
-6.56715333e-01 -2.01866508e-01 -1.06204009e+00 9.43005681e-01
-1.47850490e+00 -1.31296337e+00 1.20972469e-01 6.85190201e-01
1.11095083e+00 -5.01096010e-01 1.12891543e+00 1.49436817e-01
-8.69196236e-01 1.19362748e+00 9.10740718e-02 1.45833150e-01
9.60752666e-01 -9.99801755e-01 3.33980143e-01 2.87344068e-01
-2.53909826e-01 7.04667807e-01 4.25648570e-01 -8.27074796e-02
-1.34100044e+00 -8.53068531e-01 1.26190019e+00 -5.42627096e-01
6.41574323e-01 -5.99069774e-01 -1.06550765e+00 9.57812548e-01
4.65744853e-01 -6.06468797e-01 1.20552647e+00 7.85474598e-01
-3.85888845e-01 3.94870751e-02 -1.00563228e+00 3.51045012e-01
7.81710207e-01 -9.55462754e-01 -5.38598418e-01 7.71215320e-01
1.06888270e+00 -4.98876184e-01 -8.05367470e-01 7.66248405e-02
4.99471664e-01 -1.08869755e+00 9.11960363e-01 -7.59693980e-01
1.88480541e-01 3.38825196e-01 -3.41721773e-01 -1.18197334e+00
-8.36239159e-02 -1.07489073e+00 -2.41170943e-01 1.67379320e+00
5.05913794e-01 -8.84675622e-01 6.82716012e-01 8.83179724e-01
-4.69655603e-01 -6.60591781e-01 -1.26433766e+00 -3.23051900e-01
-6.04233108e-02 -5.09511769e-01 1.00653207e+00 1.14849997e+00
6.75520241e-01 9.13919449e-01 -7.87933707e-01 2.55958378e-01
5.11045754e-01 4.40282375e-01 8.78320813e-01 -7.55146503e-01
-5.07721186e-01 -5.38673103e-01 9.72829908e-02 -2.04477906e+00
9.69790697e-01 -9.61218417e-01 2.97362953e-01 -1.03404856e+00
2.47718900e-01 -4.29795623e-01 1.97648376e-01 2.40029231e-01
-3.47305268e-01 -3.46016586e-01 -1.03768468e-01 4.36173260e-01
-9.20764208e-01 7.65771389e-01 7.06005633e-01 -1.95988029e-01
-3.79689932e-01 6.34941161e-01 -1.03366852e+00 6.50610864e-01
6.85659111e-01 -2.45567039e-01 -2.49970078e-01 -5.15427649e-01
-4.98356551e-01 6.39556050e-01 1.11827709e-01 -6.54992983e-02
4.40768838e-01 2.31587246e-01 -4.30570483e-01 -7.53735185e-01
6.65402830e-01 -2.33863562e-01 -4.93805647e-01 1.05570674e-01
-9.87689734e-01 -2.34428197e-01 -7.60545814e-03 6.75193191e-01
-4.51286584e-01 -2.11757228e-01 5.57040513e-01 7.37367719e-02
-8.75516713e-01 -2.94073299e-02 -5.93837559e-01 -7.15122819e-02
6.05209291e-01 8.45882595e-02 -7.33258665e-01 -1.22685540e+00
-9.25741434e-01 7.14585662e-01 -2.41568506e-01 8.12528670e-01
5.39202571e-01 -1.13534606e+00 -9.77042615e-01 7.68809021e-02
2.08189934e-01 -2.94488907e-01 8.07522655e-01 1.12119174e+00
2.72641867e-01 8.28265727e-01 4.30406213e-01 -9.42275524e-01
-1.82825625e+00 1.25766546e-01 3.57479960e-01 -3.14023346e-01
-3.17535549e-01 1.44173288e+00 6.92565918e-01 -1.30630541e+00
7.97702610e-01 -4.64445502e-02 -4.54484165e-01 2.55947486e-02
9.11751926e-01 2.91427463e-01 -4.06352654e-02 -1.19502234e+00
-5.19191921e-01 -1.88834872e-02 -3.74311656e-01 1.16261609e-01
9.93748844e-01 -7.83039987e-01 -3.52561146e-01 5.08262515e-01
1.56274438e+00 -2.46727511e-01 -8.65738690e-01 -8.37267101e-01
3.28085236e-02 -2.09279969e-01 -1.38415501e-01 -5.98760068e-01
-6.99787498e-01 9.66635287e-01 2.21463919e-01 4.31279480e-01
6.64302111e-01 4.91937041e-01 1.05501378e+00 4.59780097e-01
1.66131943e-01 -8.57858002e-01 8.22120085e-02 7.73493171e-01
1.09192646e+00 -1.72828400e+00 -4.80252147e-01 -6.82286084e-01
-9.99641240e-01 8.51360261e-01 7.46099889e-01 6.41984463e-01
6.08770967e-01 -6.57728836e-02 3.95923167e-01 -5.18640503e-03
-1.25207603e+00 9.78592634e-02 2.60841161e-01 5.92281342e-01
7.54024506e-01 2.96157420e-01 2.52373010e-01 8.91495585e-01
-3.91145498e-01 -9.25465047e-01 2.04249814e-01 4.58275169e-01
-5.63459575e-01 -1.05141246e+00 -4.58475143e-01 3.60603869e-01
-3.64552796e-01 -2.74768651e-01 -7.17494547e-01 2.16790259e-01
-6.73788071e-01 1.72635281e+00 -5.25825769e-02 -5.58250785e-01
-1.17977589e-01 3.70732009e-01 -1.60715014e-01 -6.86342418e-01
-1.08927321e+00 1.73390478e-01 7.80998349e-01 -4.15572137e-01
-4.78597999e-01 -6.88639879e-01 -1.06658149e+00 -5.57107925e-01
-6.51958108e-01 3.37042838e-01 4.53788728e-01 1.21347415e+00
4.41599965e-01 3.81608814e-01 9.96266544e-01 -7.72343457e-01
-1.08663261e+00 -1.47946632e+00 -4.34901237e-01 2.85228968e-01
5.26345730e-01 -2.44297087e-01 -5.87531090e-01 -4.96619008e-02] | [12.58296012878418, 7.720943450927734] |
6ad18787-1b4f-43f9-98e3-cc4771798e60 | seal-semantic-frame-execution-and | 2303.14067 | null | https://arxiv.org/abs/2303.14067v1 | https://arxiv.org/pdf/2303.14067v1.pdf | SEAL: Semantic Frame Execution And Localization for Perceiving Afforded Robot Actions | Recent advances in robotic mobile manipulation have spurred the expansion of the operating environment for robots from constrained workspaces to large-scale, human environments. In order to effectively complete tasks in these spaces, robots must be able to perceive, reason, and execute over a diversity of affordances, well beyond simple pick-and-place. We posit the notion of semantic frames provides a compelling representation for robot actions that is amenable to action-focused perception, task-level reasoning, action-level execution, and integration with language. Semantic frames, a product of the linguistics community, define the necessary elements, pre- and post- conditions, and a set of sequential robot actions necessary to successfully execute an action evoked by a verb phrase. In this work, we extend the semantic frame representation for robot manipulation actions and introduce the problem of Semantic Frame Execution And Localization for Perceiving Afforded Robot Actions (SEAL) as a graphical model. For the SEAL problem, we describe our nonparametric Semantic Frame Mapping (SeFM) algorithm for maintaining belief over a finite set of semantic frames as the locations of actions afforded to the robot. We show that language models such as GPT-3 are insufficient to address generalized task execution covered by the SEAL formulation and SeFM provides robots with efficient search strategies and long term memory needed when operating in building-scale environments. | ['Odest Chadwicke Jenkins', 'Matthew Shannon', 'Daksh Narang', 'Cameron Kisailus'] | 2023-03-24 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 3.99893582e-01 3.15209478e-01 -1.36674598e-01 -3.61371905e-01
-3.35717291e-01 -6.22128427e-01 8.72355402e-01 8.81926715e-02
-3.20354939e-01 5.69993615e-01 2.66831160e-01 -3.41619730e-01
-6.26764536e-01 -6.61273420e-01 -7.08725274e-01 -1.90598071e-01
-3.31304133e-01 4.43305761e-01 4.19227064e-01 -5.72251558e-01
4.22076166e-01 6.46909118e-01 -1.74369407e+00 9.06329527e-02
2.72616327e-01 7.02730060e-01 1.26663208e+00 5.70351720e-01
-2.63464749e-02 9.62981701e-01 -3.59593987e-01 2.82970518e-01
1.78708017e-01 -1.94381937e-01 -1.23544765e+00 3.73809576e-01
-4.11744595e-01 -1.94543883e-01 -3.98226500e-01 1.00119257e+00
-2.57783309e-02 6.21187925e-01 6.08429551e-01 -1.65840983e+00
-7.13616192e-01 7.23338604e-01 3.47269356e-01 -2.41739303e-01
9.04136837e-01 1.00650027e-01 8.10302138e-01 -8.54464889e-01
9.00990129e-01 1.84965122e+00 3.34308743e-01 4.63282347e-01
-8.89008462e-01 1.84639156e-01 1.62124574e-01 1.75319329e-01
-1.21055472e+00 -5.13789713e-01 3.25189441e-01 -4.75738406e-01
1.49228585e+00 6.65569007e-02 5.25444567e-01 7.47741222e-01
5.92335701e-01 5.34800887e-01 7.06586897e-01 -6.81918919e-01
3.33336920e-01 -3.44742507e-01 -5.77679053e-02 6.59534216e-01
1.30116165e-01 -4.32741120e-02 -6.51024342e-01 -4.16168682e-02
1.19770622e+00 7.59436190e-02 4.68559042e-02 -7.44942904e-01
-1.71962178e+00 5.61620593e-01 3.92222017e-01 4.91143167e-01
-6.18102193e-01 6.66324615e-01 3.37798387e-01 2.53013521e-01
-7.54886717e-02 7.65944064e-01 -4.19485658e-01 -3.31726253e-01
1.31082609e-02 6.05450213e-01 9.40666616e-01 1.50190461e+00
6.87556028e-01 -3.51217598e-01 1.44713759e-01 6.55341923e-01
4.18350160e-01 4.93033707e-01 4.14566875e-01 -1.74520063e+00
4.35505629e-01 5.32460392e-01 7.45773613e-01 -1.08074582e+00
-7.20966935e-01 3.74005079e-01 4.74567786e-02 1.70253485e-01
2.44218037e-01 3.63192737e-01 -8.13704491e-01 1.90170276e+00
4.41802591e-01 -4.28112447e-01 5.18153310e-01 7.88039088e-01
4.87760752e-02 4.84629184e-01 2.60842592e-01 -5.00742719e-02
1.57683897e+00 -7.28632212e-01 -8.96734774e-01 -6.95110083e-01
1.00337517e+00 -4.55620557e-01 1.06909823e+00 5.70661910e-02
-1.15823221e+00 -4.04696912e-01 -1.15158546e+00 -5.52078784e-01
-4.51745033e-01 -1.55009031e-01 9.36446428e-01 -1.05643667e-01
-1.25837672e+00 3.77243400e-01 -1.20893109e+00 -8.53240252e-01
2.97428109e-02 3.07003707e-01 -7.43072212e-01 -1.12378813e-01
-1.00735223e+00 1.67539036e+00 7.28222191e-01 2.91073859e-01
-1.21638560e+00 2.76692867e-01 -1.40453231e+00 -1.61926597e-01
6.89862370e-01 -5.95368147e-01 1.62790608e+00 -5.03009021e-01
-1.27294958e+00 7.28516102e-01 -2.36252174e-01 -5.39640486e-01
2.36124825e-02 -3.27186495e-01 4.82870378e-02 2.11320862e-01
6.52266920e-01 9.14844751e-01 6.20151043e-01 -1.35977316e+00
-8.09050560e-01 -4.89429474e-01 7.23597944e-01 7.22081244e-01
3.02753985e-01 1.32512584e-01 -1.43561319e-01 -1.45286113e-01
8.04451823e-01 -1.21937144e+00 -2.47997046e-01 -4.44393903e-02
-4.58151773e-02 -5.22821367e-01 6.17681921e-01 -3.95298094e-01
6.53395891e-01 -2.22146416e+00 6.91271067e-01 -3.98080535e-02
-6.74042404e-02 -5.72159231e-01 -4.79604602e-02 6.71066344e-01
1.55787334e-01 -4.91937622e-02 -4.47839536e-02 -1.90454856e-01
5.25134504e-01 8.12268317e-01 -4.17198479e-01 5.24542093e-01
-2.80127972e-02 7.89382219e-01 -1.27851486e+00 -3.93514186e-01
4.10671502e-01 3.23247425e-02 -3.78037363e-01 1.45417854e-01
-4.36898738e-01 1.66016221e-01 -7.56543934e-01 5.30404747e-01
-1.45200178e-01 2.83564012e-02 5.29527068e-01 2.71765441e-01
-2.69073188e-01 4.58860129e-01 -1.17834151e+00 2.27304077e+00
-5.49055696e-01 3.38205010e-01 4.07550484e-01 -6.64402544e-01
7.63652503e-01 3.73506933e-01 3.44944477e-01 -1.49882495e-01
6.07727189e-03 4.25623626e-01 -5.54912239e-02 -7.46012449e-01
9.40907478e-01 -1.86956376e-01 -6.48885906e-01 3.15203339e-01
3.07285357e-02 -8.19993854e-01 1.77380249e-01 2.02025652e-01
1.20560944e+00 8.60862255e-01 5.17262995e-01 -3.81984651e-01
2.69096017e-01 4.53095019e-01 1.99521452e-01 7.18272209e-01
-2.40717739e-01 -7.16643333e-02 2.63842374e-01 -2.48201296e-01
-9.83545184e-01 -9.72981870e-01 1.35624692e-01 1.45177758e+00
7.90445447e-01 -2.35083938e-01 -7.83055663e-01 1.05949324e-02
-1.45714596e-01 9.80223894e-01 -3.05615783e-01 -9.68526751e-02
-6.75536215e-01 -3.02362721e-03 3.04652393e-01 5.98075867e-01
3.13803256e-01 -1.44072628e+00 -1.71010017e+00 3.22322965e-01
-3.79653633e-01 -1.27343047e+00 -1.99731649e-03 5.04550099e-01
-6.29906356e-01 -1.03406155e+00 -7.36006796e-02 -1.05447495e+00
7.46480227e-01 4.56024230e-01 6.73189998e-01 -1.38400942e-01
-4.07125428e-02 9.74170685e-01 -5.75131178e-01 -4.47120011e-01
-5.03745437e-01 -4.07511771e-01 4.19104457e-01 -6.90661132e-01
5.33092879e-02 -3.92929584e-01 -1.49370894e-01 3.50683481e-01
-7.71977484e-01 3.05470198e-01 3.49917889e-01 7.06388533e-01
4.57014471e-01 2.75501490e-01 2.89758831e-01 3.63925397e-02
7.49947429e-01 -4.43349540e-01 -2.24980503e-01 3.39173973e-01
-3.45564522e-02 8.01589638e-02 -1.40440539e-01 -2.65412033e-01
-1.25229383e+00 8.55460688e-02 4.55397308e-01 -8.00769329e-02
-2.69084573e-01 6.62761509e-01 -2.76972294e-01 2.69463599e-01
7.17931926e-01 1.51531309e-01 1.91772997e-01 -5.50272353e-02
7.48095155e-01 6.06731832e-01 7.40697265e-01 -1.15670347e+00
3.17430526e-01 5.02146661e-01 2.47588024e-01 -6.90440953e-01
-3.92535776e-01 -3.96999180e-01 -8.62029314e-01 -2.16782331e-01
1.08601665e+00 -9.00302827e-01 -7.47720659e-01 1.50503635e-01
-1.38482201e+00 -6.82618439e-01 -3.47195715e-01 4.89478469e-01
-1.55808258e+00 1.14760056e-01 -3.76811445e-01 -1.14208853e+00
5.03467917e-01 -1.43344152e+00 1.41811168e+00 -2.78754979e-01
-7.03385115e-01 -6.70596719e-01 -6.44613504e-01 -6.63739517e-02
1.29884034e-01 2.83052951e-01 9.23327148e-01 -5.18773317e-01
-5.42677641e-01 -8.13238614e-04 7.65738934e-02 -1.76210135e-01
4.30835158e-01 -6.44133747e-01 -2.70268142e-01 -5.54336980e-02
4.49613392e-01 -4.84288871e-01 4.96008508e-02 3.45018446e-01
6.85365736e-01 -1.41760767e-01 -7.11262226e-01 -5.55195427e-03
1.12763667e+00 6.94451630e-01 5.85795522e-01 6.32463872e-01
3.25174987e-01 1.03546703e+00 1.16770625e+00 1.36376277e-01
7.72720337e-01 6.30198181e-01 6.03771091e-01 6.82775378e-01
1.52366519e-01 -2.78038770e-01 4.17581737e-01 2.15737015e-01
-1.11364245e-01 -1.77130848e-01 -1.08464730e+00 6.93639934e-01
-2.30520153e+00 -9.05002534e-01 1.37483597e-01 1.78877127e+00
6.17828012e-01 1.78494498e-01 -3.09787720e-01 5.92817813e-02
6.97116911e-01 -6.44717664e-02 -4.49720562e-01 -2.81419247e-01
3.90868217e-01 -2.79737830e-01 4.38320756e-01 7.48878837e-01
-9.23606396e-01 1.31347239e+00 6.57925797e+00 4.27746207e-01
-4.45191890e-01 2.80157208e-01 -1.52679563e-01 3.81154388e-01
-1.19262666e-01 1.79317400e-01 -5.50529659e-01 -1.33801550e-01
7.94026971e-01 -4.48560789e-02 7.65152335e-01 1.10796320e+00
2.85867751e-01 -5.96105874e-01 -1.57605779e+00 8.87346923e-01
-4.82184067e-02 -9.82943058e-01 -2.04544291e-01 -1.78518310e-01
1.63029909e-01 -1.79037839e-01 -2.55285949e-01 2.03722045e-01
6.30520701e-01 -1.00166810e+00 1.53478646e+00 4.30736035e-01
6.16058826e-01 -2.65927136e-01 2.76578784e-01 6.06327355e-01
-1.22245514e+00 -5.66150904e-01 -3.55081171e-01 -5.20993352e-01
5.93012631e-01 -2.70069301e-01 -9.23745275e-01 5.19921005e-01
4.63841528e-01 4.44186270e-01 2.83562601e-01 2.23997325e-01
-3.62681478e-01 -5.21904409e-01 -3.57313752e-01 -2.63418555e-01
2.23216653e-01 4.28064354e-02 6.98402524e-01 5.97415447e-01
5.26421964e-01 3.69465292e-01 5.83397806e-01 7.57730305e-01
5.55584550e-01 -4.95409101e-01 -9.18555915e-01 -1.51177600e-01
8.91876817e-01 5.93307734e-01 -1.08689559e+00 -3.03740174e-01
-1.05131939e-01 8.97045016e-01 1.72390193e-01 3.63822728e-01
-6.19380176e-01 -2.54961044e-01 7.68204451e-01 -1.39661521e-01
-1.21386722e-01 -1.10602129e+00 -2.07449138e-01 -6.07650995e-01
9.13578123e-02 -7.39086568e-01 -1.73563033e-01 -1.32384622e+00
-6.32053375e-01 2.79272348e-01 8.21189940e-01 -9.44049835e-01
-6.05369329e-01 -9.25941586e-01 1.80283561e-01 7.59356201e-01
-8.69831920e-01 -1.35225165e+00 -1.21691771e-01 5.09891808e-01
1.05097234e+00 1.50186628e-01 1.07186019e+00 -5.68254471e-01
3.87945861e-01 -6.16553366e-01 -6.34455681e-01 -3.67948949e-01
1.16290778e-01 -1.03158450e+00 3.03030819e-01 6.79454744e-01
-2.69512177e-01 1.14127231e+00 1.00255048e+00 -9.67949927e-01
-1.84495151e+00 -5.70377231e-01 8.04054260e-01 -9.26319480e-01
7.07189441e-01 -3.65440726e-01 -3.98458421e-01 1.28483641e+00
-2.44579047e-01 -3.31238091e-01 4.91368324e-02 -6.06406070e-02
-2.88359709e-02 4.87130791e-01 -1.18988967e+00 8.59663725e-01
1.75199819e+00 -7.01285005e-01 -1.52702868e+00 4.74946141e-01
1.10508871e+00 -8.16274703e-01 -7.23174989e-01 3.38564754e-01
5.80659568e-01 -5.20945013e-01 9.76175487e-01 -3.13108742e-01
2.05655068e-01 -6.53096735e-01 -7.96512544e-01 -1.04318500e+00
-3.77104402e-01 -5.33831060e-01 1.74431533e-01 5.43914795e-01
1.47066295e-01 -6.43075943e-01 9.51460283e-03 1.01652920e+00
-6.88735306e-01 -2.16511726e-01 -1.06911373e+00 -7.94683456e-01
-4.08337831e-01 -8.88146877e-01 6.97540641e-01 5.81051469e-01
6.82145834e-01 3.48044299e-02 1.74073443e-01 3.32406163e-01
2.16408119e-01 -1.44978121e-01 5.83884716e-01 -9.72913980e-01
-4.20228727e-02 -2.53524840e-01 -5.89056551e-01 -1.15684414e+00
5.26871681e-01 -7.72425711e-01 9.14219916e-01 -2.06718683e+00
-2.02823624e-01 -7.19183862e-01 1.99974269e-01 8.14696550e-01
4.55616981e-01 -5.22704482e-01 4.92675543e-01 4.62070704e-01
-5.93964875e-01 3.82450521e-01 1.20092690e+00 1.18692219e-01
-2.05672249e-01 -5.02302527e-01 -4.56714243e-01 9.59148526e-01
5.47786176e-01 -1.94558278e-01 -7.89256811e-01 -6.48826659e-01
5.28093934e-01 3.52100283e-01 4.83911067e-01 -8.50719869e-01
3.93240213e-01 -5.82483113e-01 5.09389155e-02 -1.96284547e-01
7.21646428e-01 -1.07063091e+00 3.02088112e-01 5.80985606e-01
-5.48066318e-01 3.19726110e-01 5.45923114e-02 7.06979692e-01
3.88841219e-02 -3.33596438e-01 3.08847893e-02 -5.85454941e-01
-1.42802978e+00 -3.92289877e-01 -7.08127558e-01 -4.75620866e-01
1.30814672e+00 -4.08405513e-01 -2.45446160e-01 -1.70195058e-01
-1.12040496e+00 3.48931432e-01 6.50007367e-01 7.29966164e-01
7.12747633e-01 -1.22920120e+00 7.49642029e-04 -5.77788502e-02
1.54233068e-01 3.06967705e-01 -1.31429151e-01 4.75399733e-01
-6.26616657e-01 5.60989141e-01 -4.18404788e-01 -5.09251654e-01
-7.70916700e-01 5.55224538e-01 1.09765053e-01 3.53725523e-01
-6.90796494e-01 8.26565146e-01 2.47411087e-01 -5.98942578e-01
4.46436480e-02 -8.06762695e-01 -7.74834976e-02 -3.75868767e-01
2.77782142e-01 3.20314258e-01 -4.31219071e-01 -9.67195988e-01
-4.62924480e-01 2.17790470e-01 7.27965951e-01 -5.54900646e-01
9.73530650e-01 -5.90981603e-01 -6.31503403e-01 8.14011514e-01
4.79755998e-01 -5.80312133e-01 -1.29315090e+00 1.48397014e-01
3.12460274e-01 -3.60384703e-01 -3.87374640e-01 -4.63840872e-01
2.57984430e-01 3.26553702e-01 2.15550020e-01 3.55685741e-01
6.06583059e-01 4.15846676e-01 2.84463465e-01 8.90615165e-01
1.39301407e+00 -1.23249412e+00 1.08209901e-01 8.61868739e-01
1.24854934e+00 -7.52228379e-01 -1.76698461e-01 -5.92137396e-01
-4.87668484e-01 1.02862406e+00 5.14246881e-01 1.29251614e-01
3.87033492e-01 2.31892705e-01 -2.98661232e-01 -3.81757200e-01
-6.84050262e-01 -3.09681177e-01 -2.27567673e-01 8.51612389e-01
-2.27222834e-02 3.78213376e-01 -2.19770268e-01 2.50844270e-01
-3.60796005e-01 -4.68481109e-02 4.71552700e-01 1.76818180e+00
-1.02356064e+00 -7.87941158e-01 -5.83054781e-01 -1.81882121e-02
-1.17720105e-01 3.59639883e-01 -1.41557470e-01 9.52925444e-01
2.10873529e-01 1.31522667e+00 1.93060145e-01 -1.63445637e-01
3.98801714e-01 2.79602200e-01 8.78051579e-01 -1.06883442e+00
4.38249819e-02 -1.21603489e-01 3.72981369e-01 -8.87698710e-01
-8.19580853e-01 -8.61380458e-01 -2.09270620e+00 2.78988600e-01
-2.20420927e-01 -3.89446728e-02 1.01380169e+00 1.38413465e+00
1.96406588e-01 5.69480300e-01 -6.84033632e-02 -1.46264231e+00
-6.95775986e-01 -1.01586723e+00 -4.48509574e-01 2.76770115e-01
2.86786705e-01 -1.13948607e+00 -2.23091900e-01 3.16841662e-01] | [4.541940689086914, 0.8160803914070129] |
6f258263-847f-4128-bfa9-8c496fa46a41 | sparse-fuse-dense-towards-high-quality-3d-1 | 2203.09780 | null | https://arxiv.org/abs/2203.09780v2 | https://arxiv.org/pdf/2203.09780v2.pdf | Sparse Fuse Dense: Towards High Quality 3D Detection with Depth Completion | Current LiDAR-only 3D detection methods inevitably suffer from the sparsity of point clouds. Many multi-modal methods are proposed to alleviate this issue, while different representations of images and point clouds make it difficult to fuse them, resulting in suboptimal performance. In this paper, we present a novel multi-modal framework SFD (Sparse Fuse Dense), which utilizes pseudo point clouds generated from depth completion to tackle the issues mentioned above. Different from prior works, we propose a new RoI fusion strategy 3D-GAF (3D Grid-wise Attentive Fusion) to make fuller use of information from different types of point clouds. Specifically, 3D-GAF fuses 3D RoI features from the couple of point clouds in a grid-wise attentive way, which is more fine-grained and more precise. In addition, we propose a SynAugment (Synchronized Augmentation) to enable our multi-modal framework to utilize all data augmentation approaches tailored to LiDAR-only methods. Lastly, we customize an effective and efficient feature extractor CPConv (Color Point Convolution) for pseudo point clouds. It can explore 2D image features and 3D geometric features of pseudo point clouds simultaneously. Our method holds the highest entry on the KITTI car 3D object detection leaderboard, demonstrating the effectiveness of our SFD. Codes are available at https://github.com/LittlePey/SFD. | ['Deng Cai', 'Haifeng Liu', 'Chengqi Deng', 'Chenxi Huang', 'Liang Xie', 'Honghui Yang', 'Liang Peng', 'Xiaopei Wu'] | 2022-03-18 | sparse-fuse-dense-towards-high-quality-3d | http://openaccess.thecvf.com//content/CVPR2022/html/Wu_Sparse_Fuse_Dense_Towards_High_Quality_3D_Detection_With_Depth_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wu_Sparse_Fuse_Dense_Towards_High_Quality_3D_Detection_With_Depth_CVPR_2022_paper.pdf | cvpr-2022-1 | ['depth-completion'] | ['computer-vision'] | [-1.76981091e-01 -4.89408821e-01 5.19414805e-02 -6.18292205e-02
-1.02428436e+00 -6.27412021e-01 6.50985241e-01 -8.44512880e-02
-1.49473533e-01 2.45784938e-01 -9.17555988e-02 -2.99416631e-01
8.80384073e-02 -1.05751133e+00 -7.41446018e-01 -5.74890018e-01
2.63132334e-01 4.60714787e-01 4.59380150e-01 -3.91104698e-01
2.27046773e-01 1.07476139e+00 -2.01324391e+00 9.39314961e-02
9.45167065e-01 9.38864529e-01 5.35923839e-01 3.30689609e-01
-4.27497506e-01 -8.10302272e-02 -1.28978580e-01 -6.32049516e-02
7.41631091e-01 4.17340636e-01 -1.31736383e-01 1.59887642e-01
7.36952245e-01 -5.42783737e-01 -2.71689653e-01 1.00066864e+00
3.95864367e-01 6.55136630e-02 5.89115441e-01 -1.53117907e+00
-4.06097472e-01 -1.57672793e-01 -9.98113632e-01 -3.40915322e-02
2.18285441e-01 4.07141834e-01 7.09663451e-01 -1.61411965e+00
3.36613178e-01 1.46359730e+00 7.53548801e-01 2.29903534e-01
-9.56187606e-01 -9.91326451e-01 2.49599054e-01 -1.26940668e-01
-1.69871449e+00 -2.39093632e-01 9.51460898e-01 -4.37650323e-01
7.78894663e-01 2.76511967e-01 7.34101593e-01 8.23211133e-01
-1.52149722e-01 6.30712152e-01 1.04234600e+00 -1.28574252e-01
-8.11548438e-03 -5.90689108e-02 6.46919012e-03 6.49317920e-01
4.21794683e-01 5.42161882e-01 -5.37093222e-01 -3.55836004e-01
9.59701478e-01 6.83148265e-01 -1.56378329e-01 -6.42575383e-01
-1.27233672e+00 8.42926264e-01 6.63459539e-01 2.14056089e-03
-4.67377663e-01 6.03962019e-02 4.06580456e-02 -3.51050459e-02
5.54790854e-01 -5.85702397e-02 -3.24768335e-01 1.45888537e-01
-7.98152149e-01 3.67442906e-01 2.16647357e-01 1.33836555e+00
1.42021799e+00 -1.13034574e-02 1.35589764e-01 6.21379673e-01
5.73451519e-01 9.80532408e-01 -3.96244824e-02 -1.09946835e+00
4.54791337e-01 8.99255455e-01 1.81632396e-02 -9.86226857e-01
-2.87689418e-01 -3.82285267e-01 -7.52707541e-01 7.05754280e-01
-1.10726006e-01 1.25096038e-01 -1.03561246e+00 1.15903711e+00
8.51497769e-01 4.57597047e-01 2.63246279e-02 1.01926446e+00
1.04618406e+00 5.53527117e-01 -2.08825484e-01 2.80425578e-01
1.31900704e+00 -6.18131399e-01 -1.68955386e-01 -2.79806972e-01
6.09681129e-01 -8.19725633e-01 9.48033094e-01 1.95107639e-01
-7.08442688e-01 -6.07254803e-01 -9.96448874e-01 -2.94101447e-01
-4.72565114e-01 2.02384777e-02 6.45390749e-01 3.39494437e-01
-8.57814372e-01 7.60708451e-02 -7.07268596e-01 -2.85672337e-01
5.72084188e-01 2.39361405e-01 -5.51605344e-01 -4.89796102e-01
-6.74870253e-01 7.57644475e-01 3.10443252e-01 1.16901070e-01
-7.13454247e-01 -1.04065514e+00 -8.97188306e-01 -2.27062285e-01
5.74896395e-01 -9.49963629e-01 9.65067744e-01 -1.39043987e-01
-9.24832165e-01 8.32984269e-01 -3.24145317e-01 -1.54427767e-01
2.67598510e-01 -4.02867496e-01 -4.06073220e-02 8.77356976e-02
3.92636061e-01 8.82252395e-01 7.96744049e-01 -1.66557884e+00
-8.09392095e-01 -6.32064462e-01 4.25229482e-02 3.70074660e-01
-6.01228438e-02 -4.08874273e-01 -6.32349432e-01 -3.71510118e-01
4.81262624e-01 -7.69392490e-01 -3.59543592e-01 2.86799997e-01
-3.53207707e-01 -2.13022515e-01 1.29931021e+00 -1.04971109e-02
7.74427712e-01 -2.48549867e+00 -2.55648553e-01 7.13563934e-02
4.82747287e-01 2.03658193e-01 -3.15129101e-01 3.81652206e-01
8.59730616e-02 5.42935319e-02 -4.15754229e-01 -7.40144193e-01
-1.08489268e-01 4.49596524e-01 -4.04039204e-01 4.54452842e-01
4.31790411e-01 7.63262033e-01 -7.94930100e-01 -5.31336546e-01
7.83355296e-01 8.13426912e-01 -5.45762837e-01 -7.49166310e-02
-6.48793951e-02 3.66964489e-01 -8.51202726e-01 1.22963941e+00
1.45009422e+00 -2.19701920e-02 -6.68998778e-01 -5.09255469e-01
-6.75063133e-01 -1.37208611e-01 -1.19661224e+00 1.92845118e+00
-4.18497920e-01 8.04120004e-02 1.73122585e-01 -4.16039020e-01
1.04412210e+00 -5.17960526e-02 7.11148560e-01 -3.48494023e-01
7.30974153e-02 3.70042324e-01 -5.61480939e-01 -1.38367051e-02
7.26898789e-01 -4.75490801e-02 -1.45501317e-02 7.15878606e-02
-1.24263197e-01 -6.93238318e-01 -2.82219738e-01 3.60442400e-01
1.02755439e+00 2.10990429e-01 1.89991444e-01 6.18953109e-02
4.05224621e-01 3.85950625e-01 4.98262614e-01 7.14667797e-01
-8.80643353e-02 1.03636837e+00 -1.98928565e-01 -2.19267353e-01
-8.46067250e-01 -1.23455846e+00 -2.20542610e-01 4.13598567e-01
4.22403842e-01 -4.33818400e-01 -1.67864248e-01 -6.67672098e-01
4.90576237e-01 5.04026949e-01 -3.88856709e-01 3.46262380e-02
-4.65464622e-01 -3.33658516e-01 2.21084908e-01 4.06743705e-01
5.13430953e-01 -4.44675684e-01 -4.91055399e-01 4.48840186e-02
2.97472738e-02 -1.26631451e+00 -3.03317100e-01 1.10805050e-01
-1.10354197e+00 -1.10778475e+00 -4.12909418e-01 -3.28556329e-01
5.32859206e-01 1.24755728e+00 8.69360268e-01 1.63363993e-01
-1.80251211e-01 6.06880367e-01 -5.44372320e-01 -5.20635128e-01
9.61038768e-02 -1.24446951e-01 1.05873823e-01 -1.17223956e-01
5.73081970e-01 -7.30762362e-01 -5.50152242e-01 3.12963307e-01
-8.66594076e-01 1.60373569e-01 6.87363625e-01 5.46771228e-01
1.03954339e+00 -2.84934908e-01 5.08190617e-02 -4.51537758e-01
3.85385342e-02 -7.19831765e-01 -7.83889413e-01 -2.84614682e-01
-2.43522644e-01 -3.04507345e-01 4.30888459e-02 -1.69094265e-01
-7.52202332e-01 4.71635193e-01 -2.24445015e-01 -1.35046530e+00
-3.57648343e-01 2.10532770e-01 -3.21935683e-01 -3.92799884e-01
5.29168129e-01 2.07615152e-01 1.15426853e-01 -7.61058152e-01
6.27512991e-01 5.74016213e-01 6.13505661e-01 -5.22648335e-01
1.39597225e+00 9.56868827e-01 1.27591148e-01 -8.56762290e-01
-6.92875624e-01 -8.71314049e-01 -7.19335735e-01 -2.05004245e-01
6.92433059e-01 -1.26545322e+00 -5.34879506e-01 3.98728907e-01
-1.27702665e+00 1.14954427e-01 -3.84013981e-01 4.66963679e-01
-3.82475913e-01 4.46910143e-01 -4.92435321e-02 -8.27872753e-01
-2.08569318e-01 -1.10108244e+00 1.69011796e+00 4.11943160e-02
3.91031504e-01 -4.71145064e-01 6.22197725e-02 2.20304817e-01
1.34120539e-01 4.52419519e-01 2.87196279e-01 -2.57139057e-01
-1.11344171e+00 -2.70881891e-01 -5.73658407e-01 2.76294023e-01
2.03996360e-01 1.48868961e-02 -1.01022911e+00 -1.51849777e-01
5.79262301e-02 6.64855540e-02 9.28105414e-01 1.79544806e-01
1.01554871e+00 1.92864493e-01 -6.09853268e-01 8.82704556e-01
1.56567824e+00 -1.66920677e-01 4.49553072e-01 2.55845815e-01
1.10373700e+00 2.63119668e-01 1.05874181e+00 5.13402104e-01
6.76202893e-01 7.51618981e-01 1.11592841e+00 -2.99399465e-01
-2.10838422e-01 -2.42585659e-01 2.38287583e-01 5.50862849e-01
-2.05396205e-01 4.32855450e-02 -9.95205283e-01 5.38237393e-01
-1.75489140e+00 -7.60607779e-01 -5.46046972e-01 2.08518720e+00
1.96942717e-01 -4.91702296e-02 3.49229276e-02 4.05514538e-02
6.04668438e-01 1.18007317e-01 -5.13130426e-01 2.57512987e-01
-2.06352457e-01 2.39658162e-01 6.64895058e-01 4.23159420e-01
-9.93320584e-01 8.85336936e-01 4.95063877e+00 1.03186333e+00
-8.78857732e-01 3.82059187e-01 -2.70866174e-02 -1.91039458e-01
-6.12159669e-01 2.07340866e-01 -1.13837588e+00 3.65374714e-01
3.77992272e-01 -5.42316958e-03 -3.57129313e-02 9.61401701e-01
1.66270480e-01 -1.38904512e-01 -7.56257892e-01 1.24742055e+00
-8.20936710e-02 -1.45314062e+00 9.53366235e-02 4.52500284e-01
5.46383262e-01 6.39474213e-01 2.93493774e-02 2.76999146e-01
1.99684247e-01 -5.57983220e-01 8.05525243e-01 4.99986947e-01
7.94260502e-01 -6.98270500e-01 4.11405057e-01 4.86136526e-01
-1.63588536e+00 -1.69161186e-02 -3.85220289e-01 1.86529029e-02
3.91793966e-01 1.03001869e+00 -7.40447760e-01 1.01476181e+00
8.63464296e-01 9.46838498e-01 -5.29075205e-01 1.21488452e+00
9.25777387e-03 7.90677816e-02 -8.69482279e-01 4.03130949e-01
3.14372391e-01 -1.81721061e-01 8.63514602e-01 9.01303172e-01
8.44971061e-01 3.23935181e-01 4.41420168e-01 1.00646782e+00
2.03184411e-01 -2.14686155e-01 -1.07993436e+00 3.92228395e-01
8.17225814e-01 1.43185937e+00 -4.03491586e-01 -2.70962685e-01
-7.14583218e-01 6.14936590e-01 8.57295990e-02 1.94856554e-01
-6.45783722e-01 -2.12695673e-02 1.07594740e+00 3.25347632e-01
4.41123605e-01 -7.11514235e-01 -3.83435309e-01 -1.14782584e+00
-8.37536380e-02 -4.10248041e-01 1.29304454e-01 -9.83342886e-01
-1.48655117e+00 4.09781367e-01 2.49473602e-01 -1.86213613e+00
1.45181105e-01 -5.68282723e-01 -4.55949932e-01 1.00983357e+00
-1.87430239e+00 -1.47261715e+00 -6.59035623e-01 8.92259002e-01
5.15991211e-01 1.56343773e-01 4.96871889e-01 4.18685049e-01
-2.04156622e-01 1.24499167e-03 -2.29343906e-01 -2.79293299e-01
4.95199561e-01 -8.58672917e-01 4.78467792e-01 8.48148227e-01
1.15221635e-01 5.55836320e-01 2.40287170e-01 -7.69563258e-01
-1.68743026e+00 -1.30248940e+00 3.45253408e-01 -6.70506120e-01
3.62625182e-01 -3.33306581e-01 -9.25037980e-01 4.54050779e-01
-3.48358244e-01 3.60495180e-01 2.39958182e-01 -1.50201291e-01
-3.30162972e-01 -2.19489157e-01 -1.17533505e+00 3.76693606e-01
1.33032286e+00 -5.26393294e-01 -4.85824227e-01 3.12468141e-01
9.60841417e-01 -6.66033566e-01 -6.44586504e-01 8.16844702e-01
1.41259924e-01 -1.13310480e+00 1.37548959e+00 2.07445383e-01
5.50763682e-02 -1.01043618e+00 -5.39972007e-01 -9.64397728e-01
-1.82211161e-01 -1.90744817e-01 -2.92388439e-01 1.14664578e+00
-4.93598282e-02 -8.49380076e-01 7.38940716e-01 5.90289943e-02
-8.72138500e-01 -5.87779880e-01 -1.23914623e+00 -7.90413618e-01
-1.47087947e-01 -8.74651968e-01 1.05281770e+00 8.14254999e-01
-6.63745165e-01 9.04126540e-02 -7.31118098e-02 6.63465321e-01
7.52512157e-01 3.80008936e-01 1.23117030e+00 -1.40716517e+00
1.33181080e-01 -3.77649844e-01 -5.64411759e-01 -1.18620026e+00
-1.43591732e-01 -8.52170050e-01 -1.03682034e-01 -1.44631076e+00
-3.05385161e-02 -8.38744462e-01 -3.95196564e-02 5.96809983e-01
-6.47106618e-02 5.72154462e-01 4.58030313e-01 4.15015191e-01
-1.55556068e-01 7.29077578e-01 1.22713232e+00 -3.97607870e-02
-8.28152969e-02 -3.20185982e-02 -6.54830873e-01 7.22450197e-01
6.22754872e-01 -4.06208754e-01 -1.74000189e-01 -6.65180743e-01
-3.34131829e-02 -7.15170279e-02 9.49453175e-01 -1.25853419e+00
2.54511803e-01 -1.87291890e-01 2.60772824e-01 -1.54199827e+00
9.03225601e-01 -1.13474000e+00 3.14541250e-01 1.23291254e-01
7.28928387e-01 8.90871212e-02 4.28683072e-01 5.90341449e-01
-2.24853784e-01 2.65765153e-02 5.40998936e-01 -2.80414432e-01
-8.70642543e-01 8.78133476e-01 6.66157454e-02 -5.30933261e-01
1.11833763e+00 -5.93108892e-01 -3.32451582e-01 4.55936901e-02
-4.23218340e-01 3.89793664e-01 7.97926903e-01 4.67832625e-01
1.00138795e+00 -1.52275193e+00 -6.40709817e-01 4.59368855e-01
4.41024661e-01 6.82472408e-01 5.71994662e-01 9.15830851e-01
-2.84388334e-01 1.29592702e-01 4.35758345e-02 -1.10789716e+00
-1.06233644e+00 6.40708745e-01 9.23891887e-02 1.48115426e-01
-8.84692252e-01 7.22163200e-01 5.13893843e-01 -6.31419301e-01
-2.30459884e-01 -4.40409899e-01 6.31806105e-02 5.89572676e-02
3.53311151e-01 2.37311602e-01 2.22472876e-01 -7.71864831e-01
-4.98818845e-01 1.04473925e+00 1.42607659e-01 8.75873864e-02
1.26460159e+00 -1.76535577e-01 1.27854824e-01 3.43766809e-01
1.06334019e+00 2.54630327e-01 -1.40760612e+00 -3.06412518e-01
-4.85021532e-01 -8.70656371e-01 2.81551361e-01 -4.18808401e-01
-1.07012486e+00 9.96591449e-01 6.37367725e-01 -1.30059393e-02
1.11779833e+00 1.13712341e-01 7.72393405e-01 1.90284494e-02
7.51158595e-01 -4.18401659e-01 -1.80584207e-01 5.80413163e-01
9.24247146e-01 -1.38187671e+00 1.83175698e-01 -7.67024696e-01
-3.08996171e-01 1.01331747e+00 6.88979447e-01 -2.53717721e-01
6.71611130e-01 2.12612107e-01 -7.21409395e-02 -5.28714597e-01
-3.84880513e-01 -6.87970638e-01 2.07539707e-01 7.97169209e-01
-2.94562310e-01 -6.91249892e-02 1.31865114e-01 2.94429779e-01
1.92137819e-03 -5.99024370e-02 1.87270597e-01 1.03220582e+00
-6.02991939e-01 -1.05146754e+00 -8.53719413e-01 3.69016975e-01
3.89385581e-01 -1.30409062e-01 1.49443997e-02 8.99313748e-01
4.00848180e-01 7.25848019e-01 2.14078322e-01 -6.16562605e-01
5.35582304e-01 -2.98622102e-01 2.82280117e-01 -8.72531533e-01
-2.28291631e-01 1.47141367e-01 -2.68095195e-01 -7.93342292e-01
-4.67631102e-01 -7.75846243e-01 -1.17361569e+00 -3.81335646e-01
-4.84602839e-01 -1.66443869e-01 8.43092799e-01 6.45336628e-01
7.06998646e-01 2.68592268e-01 7.45496690e-01 -1.53428435e+00
-1.74816996e-01 -6.73533976e-01 -4.39361662e-01 -7.10681975e-02
4.77870375e-01 -1.22181845e+00 -5.27615368e-01 -4.70935136e-01] | [7.82658576965332, -2.7329671382904053] |
18ef2e93-bb13-437e-8109-0cbfd46c16e3 | meta-federated-reinforcement-learning-for | 2307.02900 | null | https://arxiv.org/abs/2307.02900v2 | https://arxiv.org/pdf/2307.02900v2.pdf | Meta Federated Reinforcement Learning for Distributed Resource Allocation | In cellular networks, resource allocation is usually performed in a centralized way, which brings huge computation complexity to the base station (BS) and high transmission overhead. This paper explores a distributed resource allocation method that aims to maximize energy efficiency (EE) while ensuring the quality of service (QoS) for users. Specifically, in order to address wireless channel conditions, we propose a robust meta federated reinforcement learning (\textit{MFRL}) framework that allows local users to optimize transmit power and assign channels using locally trained neural network models, so as to offload computational burden from the cloud server to the local users, reducing transmission overhead associated with local channel state information. The BS performs the meta learning procedure to initialize a general global model, enabling rapid adaptation to different environments with improved EE performance. The federated learning technique, based on decentralized reinforcement learning, promotes collaboration and mutual benefits among users. Analysis and numerical results demonstrate that the proposed \textit{MFRL} framework accelerates the reinforcement learning process, decreases transmission overhead, and offloads computation, while outperforming the conventional decentralized reinforcement learning algorithm in terms of convergence speed and EE performance across various scenarios. | ['Xiaoming Tao', 'Zhijin Qin', 'Zelin Ji'] | 2023-07-06 | null | null | null | null | ['meta-learning', 'federated-learning'] | ['methodology', 'methodology'] | [-3.38588029e-01 -5.50620183e-02 -3.94006789e-01 7.50658587e-02
-4.72465575e-01 -3.58721882e-01 -1.39967099e-01 -7.06722820e-03
-3.31808269e-01 1.22995281e+00 -2.28108361e-01 -3.98943007e-01
-4.18696910e-01 -1.15961325e+00 -1.38972938e-01 -1.36232138e+00
-3.76953423e-01 5.40194064e-02 -5.08224726e-01 3.38890613e-03
-9.50452387e-02 3.66222739e-01 -7.50769317e-01 -7.40607083e-01
1.00286067e+00 1.38346922e+00 6.19870901e-01 4.47393686e-01
2.96416879e-02 7.40329623e-01 -6.93900645e-01 -1.49735034e-01
1.25693515e-01 -4.80854332e-01 -4.24272805e-01 1.43560752e-01
-7.62334466e-01 -4.54503000e-01 -5.46901107e-01 1.01165187e+00
9.76697922e-01 1.44217074e-01 1.53679743e-01 -1.38546622e+00
-2.60240167e-01 5.00066400e-01 -3.93808126e-01 9.21439007e-02
-2.46353596e-01 6.70235679e-02 8.24915588e-01 -3.83738458e-01
1.65329337e-01 6.71887755e-01 4.60844934e-01 4.93809879e-01
-8.78097773e-01 -9.03446317e-01 1.77167088e-01 5.54954231e-01
-1.49517143e+00 -4.58404869e-01 8.10823143e-01 2.33587429e-01
5.35765111e-01 3.07282299e-01 9.42577899e-01 3.25421333e-01
3.94401886e-02 5.19499660e-01 7.50200212e-01 -7.10922837e-01
9.21623111e-01 -6.37454493e-03 -6.63937807e-01 7.70995855e-01
7.94105679e-02 2.01918744e-03 -1.00855507e-01 -3.55865747e-01
1.03670144e+00 -5.76656386e-02 -4.24246311e-01 -4.28477764e-01
-9.51312780e-01 5.45598865e-01 4.85035956e-01 2.56253242e-01
-7.93315530e-01 4.42875862e-01 2.45255589e-01 4.48474854e-01
2.62156397e-01 5.49761159e-03 -4.94495541e-01 -1.66616097e-01
-5.44657052e-01 -3.07854265e-01 7.64637589e-01 9.66019094e-01
1.08233142e+00 3.61990601e-01 -3.07474315e-01 9.48719621e-01
3.89853269e-01 8.06888163e-01 3.22175831e-01 -1.44422853e+00
2.96515703e-01 3.27514857e-01 2.06298307e-01 -8.42401087e-01
-5.19177437e-01 -1.14503479e+00 -1.38371873e+00 -7.41833448e-02
-1.82747409e-01 -1.07034767e+00 4.03835066e-02 1.88197911e+00
4.26848412e-01 2.66315132e-01 3.73504311e-01 8.98589969e-01
1.97813153e-01 9.50338244e-01 1.15136892e-01 -1.02916229e+00
8.53754401e-01 -9.60036695e-01 -9.31127369e-01 1.62414998e-01
7.45258689e-01 -3.51322234e-01 3.88724208e-01 1.64909422e-01
-1.11910701e+00 -9.98337194e-02 -1.04221690e+00 8.81431401e-01
8.24381337e-02 4.92188275e-01 6.74401283e-01 9.85422730e-01
-1.13100028e+00 3.91948462e-01 -4.98349637e-01 -2.45587781e-01
5.24155319e-01 6.38142526e-01 4.15105641e-01 1.06770948e-01
-1.19648468e+00 2.87260801e-01 4.71856415e-01 8.59302804e-02
-8.43351424e-01 -4.94265705e-01 -4.67368096e-01 4.97976422e-01
5.81984401e-01 -1.01503205e+00 1.38902926e+00 -9.02691960e-01
-1.99189258e+00 -3.48465294e-01 1.54578820e-01 -3.17084551e-01
3.16403627e-01 4.91179645e-01 -6.49930239e-01 3.29685420e-01
-2.35367611e-01 1.04407869e-01 6.03607833e-01 -1.06112611e+00
-9.07239914e-01 -1.44418478e-01 -7.55492151e-02 6.04211092e-01
-9.64000881e-01 -3.58157754e-01 -5.16266942e-01 -6.87652349e-01
-2.86381185e-01 -8.11619043e-01 -5.62293053e-01 -2.39932224e-01
2.48604473e-02 -1.80975690e-01 1.05051732e+00 -3.66620839e-01
1.45621848e+00 -1.96155500e+00 1.52674653e-02 7.01489687e-01
2.29201883e-01 3.87624428e-02 4.52889968e-03 4.52535391e-01
5.48220754e-01 -2.91699003e-02 2.16119230e-01 1.21114664e-01
-1.71442181e-01 3.94647628e-01 3.46619815e-01 2.79503316e-01
-7.28853583e-01 7.02421010e-01 -9.21544671e-01 -5.59730232e-01
1.41959473e-01 1.06610268e-01 -6.29553795e-01 3.18400681e-01
9.65044089e-03 3.63933682e-01 -9.44080651e-01 5.46231151e-01
5.01478195e-01 -5.95348001e-01 5.97212911e-01 -1.54485583e-01
4.10292298e-03 -3.78608108e-01 -1.21994865e+00 1.48022044e+00
-1.06257057e+00 3.17789584e-01 7.44326711e-01 -1.32410216e+00
8.24072003e-01 7.57626116e-01 1.17552841e+00 -1.00933671e+00
3.46690923e-01 1.97052270e-01 -2.77467281e-01 -3.41021359e-01
-1.69865891e-01 1.04915760e-01 4.63540033e-02 6.07556164e-01
1.27201214e-01 3.96237493e-01 -2.00300738e-02 2.07042634e-01
1.07626319e+00 -2.55731702e-01 3.64686221e-01 -1.18070669e-01
7.28155553e-01 -5.35880983e-01 8.60998213e-01 5.78624427e-01
-4.19809997e-01 -6.11094415e-01 -5.32431863e-02 -2.20873937e-01
-9.40309286e-01 -6.86330497e-01 4.71438468e-01 1.14892697e+00
5.55186212e-01 -1.90598056e-01 -9.22522128e-01 -3.48906338e-01
-3.25878829e-01 4.28792775e-01 -9.63612348e-02 -3.54600519e-01
-3.35105769e-02 -9.35870469e-01 2.80758530e-01 1.13002673e-01
1.06560743e+00 -7.59122431e-01 -5.90595782e-01 6.75894260e-01
-3.69384050e-01 -1.00605750e+00 -4.82323736e-01 2.36117572e-01
-6.27851844e-01 -5.67720056e-01 -7.70323455e-01 -6.98225856e-01
6.25024199e-01 3.34263921e-01 5.85957825e-01 7.77420327e-02
-8.16732496e-02 6.82613909e-01 -2.85334527e-01 -2.19521478e-01
-1.51027322e-01 1.34349629e-01 1.01205513e-01 3.07784975e-01
-3.68045896e-01 -7.71051705e-01 -9.66361701e-01 5.29900312e-01
-5.24971306e-01 -9.78441350e-03 7.65324593e-01 7.52894223e-01
4.96742040e-01 6.44530654e-01 1.27079749e+00 -2.95402527e-01
7.36070156e-01 -5.80681622e-01 -6.03613198e-01 4.07157689e-01
-8.75359356e-01 -1.40926287e-01 1.09984219e+00 -9.49192420e-03
-1.27514553e+00 -1.03050888e-01 1.22420684e-01 -1.13162503e-01
4.11833465e-01 3.05909365e-01 -3.76473069e-01 -5.26335418e-01
3.62778693e-01 5.75134397e-01 1.15381680e-01 -5.90947457e-02
9.50186402e-02 9.76835608e-01 1.31706536e-01 -5.01043737e-01
7.24792421e-01 1.34136066e-01 2.03767359e-01 -6.60899937e-01
-5.95436811e-01 -2.12458059e-01 1.05231747e-01 -6.86316371e-01
1.55956000e-01 -1.20945215e+00 -1.19037795e+00 1.69321463e-01
-6.89076602e-01 -5.59415519e-01 -1.92158390e-02 5.49150765e-01
-8.27068865e-01 2.55180299e-01 -2.48496562e-01 -9.29637671e-01
-7.67116308e-01 -5.52226841e-01 2.10614130e-01 6.12321794e-01
5.02900004e-01 -1.09550571e+00 -2.02218965e-01 1.53146073e-01
8.55247319e-01 1.76856504e-03 8.74327421e-01 1.39416784e-01
-4.80342209e-01 1.02441862e-01 -4.63343374e-02 1.92259952e-01
3.85098755e-01 -3.95737678e-01 -5.20057023e-01 -7.52307475e-01
-4.62978423e-01 -2.56148726e-01 1.04624659e-01 5.20360470e-01
1.62371945e+00 -8.27200532e-01 -3.85429442e-01 5.87445617e-01
1.71647608e+00 4.42216367e-01 3.20621192e-01 2.71643400e-01
1.14586771e-01 -1.04036247e-02 5.32563329e-01 1.26198936e+00
6.04822755e-01 4.51330811e-01 8.61421347e-01 -2.48066485e-01
2.48271018e-01 9.36840400e-02 2.01690540e-01 7.79783607e-01
-2.22800314e-01 -3.77434343e-01 -2.68327057e-01 1.47926882e-01
-2.20385671e+00 -1.03934014e+00 7.40159392e-01 2.07348442e+00
5.39352059e-01 -2.77954102e-01 2.50557780e-01 2.43184716e-01
8.10261190e-01 -2.69236743e-01 -8.46415579e-01 -2.67751753e-01
1.58476382e-01 -2.09707603e-01 6.57616377e-01 1.37999028e-01
-6.56952202e-01 4.19336766e-01 5.18916559e+00 1.08009326e+00
-1.18363452e+00 2.37523094e-01 7.77583301e-01 -1.22949973e-01
-1.36424437e-01 -1.78011060e-01 -8.96714032e-02 6.37735426e-01
1.01377618e+00 -5.12858033e-01 1.15490174e+00 8.87528360e-01
9.21072304e-01 1.17218256e-01 -3.60350400e-01 1.26938426e+00
-6.49536431e-01 -1.54770184e+00 -2.72045702e-01 3.21842611e-01
8.43713641e-01 -5.10101169e-02 -4.38131422e-01 4.49898928e-01
2.38171473e-01 -4.96446192e-01 3.10235173e-01 6.44296765e-01
6.09925628e-01 -1.48034120e+00 5.27701199e-01 6.96721137e-01
-1.32469058e+00 -8.61986458e-01 -3.09980989e-01 4.80602756e-02
1.97156202e-02 7.71127045e-01 -6.91770375e-01 7.59797692e-01
6.55283391e-01 3.30328286e-01 3.71759059e-03 1.27087700e+00
2.06764221e-01 6.42562032e-01 -1.49495721e-01 -4.69263703e-01
2.04727754e-01 -3.74656081e-01 3.81209731e-01 8.86789083e-01
6.79653525e-01 3.90645415e-01 5.59668183e-01 3.44294548e-01
-3.98094475e-01 4.85852450e-01 4.45532054e-02 2.01728776e-01
1.08823025e+00 1.87242234e+00 -4.51073349e-01 -2.59534597e-01
-2.80585408e-01 8.00571620e-01 2.23249927e-01 5.59118271e-01
-7.30476558e-01 -8.53838801e-01 5.37905216e-01 -3.50425154e-01
2.88768619e-01 -2.46510729e-01 -7.98518397e-03 -6.20782316e-01
-1.91935956e-01 -3.71031225e-01 2.72986531e-01 -5.33889413e-01
-7.35684216e-01 1.25353247e-01 -7.57598579e-01 -1.32055843e+00
-2.56593913e-01 2.21986361e-02 -7.06816852e-01 4.98730958e-01
-1.77577162e+00 -7.66956270e-01 -5.73234379e-01 9.98985529e-01
1.99904814e-01 -4.26406384e-01 9.43380237e-01 5.88868618e-01
-9.64117348e-01 7.74288654e-01 7.26175308e-01 -3.42088714e-02
1.87618136e-01 -1.04122281e+00 -8.96601200e-01 3.98115486e-01
-4.90999281e-01 5.03995344e-02 3.53710473e-01 4.70313281e-02
-1.69194329e+00 -1.38341606e+00 1.92125559e-01 1.10514402e+00
3.70619416e-01 1.15375012e-01 -8.83634388e-02 -5.91709577e-02
3.90058547e-01 3.90929818e-01 9.48463619e-01 -2.42986694e-01
5.02506673e-01 -7.54560173e-01 -1.33573329e+00 8.35576355e-01
6.83126390e-01 -7.49730691e-02 8.12380075e-01 3.74675989e-01
4.50414151e-01 8.24074522e-02 -1.23301005e+00 1.28300756e-01
4.25964087e-01 -4.25530910e-01 5.76120496e-01 -8.29650536e-02
-3.96351486e-01 -3.75011921e-01 -2.39027753e-01 -1.71936870e+00
-6.95166111e-01 -1.24122584e+00 -6.97596371e-01 1.14685762e+00
2.81508446e-01 -6.39081657e-01 8.78621876e-01 2.25341141e-01
1.10609189e-01 -9.35289681e-01 -1.21435416e+00 -7.44472563e-01
-4.42131966e-01 -1.81586117e-01 7.25902259e-01 7.18172312e-01
3.53916287e-01 3.25365990e-01 -3.81903261e-01 4.84416813e-01
7.41564691e-01 7.53762275e-02 6.47796392e-01 -8.70050132e-01
-5.02417564e-01 -3.73667687e-01 -1.66764483e-01 -9.43886518e-01
-2.58777216e-02 -7.73953021e-01 -2.45131761e-01 -1.62493145e+00
3.62898409e-02 -9.50797200e-01 -7.20834792e-01 6.57487035e-01
2.27762893e-01 1.60087600e-01 1.34631231e-01 2.34816760e-01
-1.04703116e+00 9.16057348e-01 1.33621919e+00 -8.99029598e-02
-3.32220882e-01 4.14065987e-01 -8.88123691e-01 3.33837420e-01
1.17144454e+00 -1.05222106e-01 -6.32493496e-01 -2.34253421e-01
9.64507237e-02 8.34980249e-01 1.51770607e-01 -1.29502630e+00
4.26831812e-01 -4.55939412e-01 4.62374955e-01 -5.01584671e-02
6.98000640e-02 -1.32773221e+00 -1.10263497e-01 7.25214362e-01
-1.49100302e-02 -4.24348861e-01 -1.86526373e-01 9.01641428e-01
1.50007144e-01 1.16668260e-02 7.51225531e-01 1.91993788e-01
-5.38251281e-01 5.95917404e-01 -7.31749177e-01 -4.30641949e-01
1.46545017e+00 2.52177776e-03 7.55304694e-02 -9.39196646e-01
-8.13113093e-01 9.19527233e-01 -9.85576957e-02 -9.37866122e-02
3.39953899e-01 -1.54667687e+00 -3.35592896e-01 -1.86394483e-01
-2.86277652e-01 -4.23313349e-01 4.48101133e-01 8.32175612e-01
-1.51792020e-01 4.45510179e-01 -4.26687188e-02 -4.63412642e-01
-9.52740252e-01 1.97826102e-01 7.21894205e-01 -1.99714243e-01
-1.12185277e-01 2.55088598e-01 -4.95244235e-01 -2.26851106e-01
3.26364934e-01 3.25685650e-01 -1.05213344e-01 -2.03168333e-01
1.27539799e-01 9.33624148e-01 1.23108774e-01 3.49710509e-02
-8.49795714e-02 2.30743870e-01 2.67042130e-01 3.92350219e-02
1.25679934e+00 -8.73388708e-01 4.88520265e-02 -4.46927458e-01
1.11691308e+00 -7.56029934e-02 -1.33748579e+00 -5.34482718e-01
-4.80507642e-01 -2.02075765e-01 7.86329627e-01 -9.45800900e-01
-1.53932381e+00 1.09148607e-01 9.96853828e-01 3.40485662e-01
1.73288620e+00 -3.90225381e-01 9.24125612e-01 7.51879811e-01
8.16693485e-01 -1.74348748e+00 1.07236244e-01 1.47328630e-01
2.58162916e-01 -8.19967806e-01 -1.22352079e-01 -1.66768894e-01
-4.23496842e-01 1.23337889e+00 6.34611845e-01 3.11089367e-01
7.77334273e-01 1.27723679e-01 -2.55289972e-01 3.51432920e-01
-7.29781687e-01 -5.02356812e-02 -5.73041737e-01 4.94030952e-01
1.33569259e-02 2.86168069e-01 -4.15737987e-01 6.26462758e-01
1.35099530e-01 9.34266746e-02 4.42816049e-01 9.57259595e-01
-9.27882791e-01 -1.17555952e+00 -1.92413330e-01 4.81299728e-01
-2.39898384e-01 2.89369911e-01 3.23960841e-01 2.25827023e-01
1.80722445e-01 1.20304561e+00 -1.01383127e-01 -7.99837336e-02
-2.97570005e-02 -5.10104775e-01 3.54885757e-01 -5.88729158e-02
-4.31917608e-02 4.74255025e-01 -3.44189852e-02 -3.43892634e-01
-3.75907540e-01 -2.03091562e-01 -1.62386644e+00 -2.21290037e-01
-4.67306733e-01 6.42702520e-01 7.92944193e-01 1.06442857e+00
7.29675055e-01 8.50139022e-01 1.81361222e+00 -6.66440070e-01
-6.09759867e-01 -5.35654783e-01 -7.75014222e-01 -4.89145547e-01
4.08338368e-01 -2.89200604e-01 1.48830980e-01 -3.07930589e-01] | [5.94134521484375, 1.6367100477218628] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.