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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
65f7da03-e1de-447e-aed5-e950397cb90a | dionysus-recovering-scene-structures-by | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_Dionysus_Recovering_Scene_Structures_by_Dividing_Into_Semantic_Pieces_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_Dionysus_Recovering_Scene_Structures_by_Dividing_Into_Semantic_Pieces_CVPR_2023_paper.pdf | Dionysus: Recovering Scene Structures by Dividing Into Semantic Pieces | Most existing 3D reconstruction methods result in either detail loss or unsatisfying efficiency. However, effectiveness and efficiency are equally crucial in real-world applications, e.g., autonomous driving and augmented reality. We argue that this dilemma comes from wasted resources on valueless depth samples. This paper tackles the problem by proposing a novel learning-based 3D reconstruction framework named Dionysus. Our main contribution is to find out the most promising depth candidates from estimated semantic maps. This strategy simultaneously enables high effectiveness and efficiency by attending to the most reliable nominators. Specifically, we distinguish unreliable depth candidates by checking the cross-view semantic consistency and allow adaptive sampling by redistributing depth nominators among pixels. Experiments on the most popular datasets confirm our proposed framework's effectiveness. | ['Lei Chen', 'Likang Wang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['3d-reconstruction'] | ['computer-vision'] | [ 2.14905683e-02 2.98691511e-01 -2.20057145e-01 -2.64238387e-01
-9.37856138e-01 -2.65486330e-01 3.46499652e-01 1.29660107e-02
-3.61788213e-01 7.98411548e-01 -5.91791905e-02 -3.01322509e-02
-2.99096823e-01 -8.53539348e-01 -5.80930948e-01 -6.24242902e-01
3.43657620e-02 4.76363569e-01 5.61183155e-01 2.73623206e-02
7.56434202e-01 6.30942106e-01 -1.89052784e+00 1.54645247e-02
1.05281246e+00 1.18356764e+00 5.32327831e-01 1.29884496e-01
-1.65793970e-01 5.21628618e-01 -2.63390422e-01 -2.53255427e-01
4.79781568e-01 -1.05562314e-01 -4.37546641e-01 1.86386779e-01
3.81937146e-01 -6.41228080e-01 1.26615033e-01 1.03729618e+00
6.64866328e-01 -5.04061244e-02 4.34016287e-01 -1.28266919e+00
3.02357711e-02 9.91034806e-02 -7.46356428e-01 9.98693854e-02
3.30733865e-01 -1.24070831e-01 8.34549904e-01 -1.03259850e+00
4.89243597e-01 1.02627063e+00 5.91518879e-01 4.65388656e-01
-9.16622877e-01 -5.18739641e-01 1.12006113e-01 2.69938350e-01
-1.42721105e+00 -7.27118611e-01 1.13901854e+00 -2.70789176e-01
6.33054435e-01 2.35425174e-01 5.59162080e-01 7.32021570e-01
1.44101053e-01 9.25873399e-01 1.16571522e+00 -3.98849666e-01
6.44091725e-01 3.66047591e-01 -1.96102574e-01 7.18402207e-01
4.09135163e-01 2.26672702e-02 -7.98900783e-01 4.17678207e-02
9.76972103e-01 -2.26418793e-01 -1.75208926e-01 -7.53489077e-01
-1.10606897e+00 5.87190390e-01 2.79002100e-01 -4.79146652e-03
-4.20602858e-01 -3.35307978e-02 1.14015713e-01 2.41990704e-02
5.93231618e-01 1.71017081e-01 -4.92905289e-01 -1.60936147e-01
-8.80362928e-01 1.69307753e-01 1.93335012e-01 1.02004290e+00
9.58798409e-01 -1.37642652e-01 4.41770881e-01 7.75032461e-01
2.45626435e-01 2.86803216e-01 1.66963011e-01 -1.12649167e+00
5.10795474e-01 5.83051383e-01 2.78427452e-01 -1.05089891e+00
-3.48729402e-01 -4.38624650e-01 -6.44994795e-01 5.12568057e-01
1.95427567e-01 2.02222675e-01 -6.23429418e-01 1.39198065e+00
6.54288828e-01 6.57787919e-02 -8.77427906e-02 1.05270755e+00
6.83851004e-01 4.09183711e-01 -2.00486213e-01 -7.17864037e-02
1.09729993e+00 -7.65062034e-01 -5.24085641e-01 -3.55282664e-01
3.74969035e-01 -5.43874323e-01 1.07527208e+00 6.29158854e-01
-1.20964718e+00 -4.53958869e-01 -1.04913282e+00 -8.52436200e-02
-3.46540697e-02 1.21033773e-01 7.91918039e-01 6.77601039e-01
-9.12273645e-01 3.21290910e-01 -7.74138510e-01 -8.06365386e-02
4.44163769e-01 2.74958491e-01 -5.68824470e-01 -2.05583926e-02
-7.23845363e-01 7.09412634e-01 3.60874087e-01 1.43403128e-01
-6.41985357e-01 -5.37187219e-01 -6.63359165e-01 -2.81780899e-01
5.72416663e-01 -7.63817906e-01 1.22979069e+00 -7.68760204e-01
-1.32807302e+00 9.06220138e-01 -1.92461222e-01 -2.70019710e-01
8.69524956e-01 -2.18597695e-01 -4.42818440e-02 2.43570819e-01
3.69863719e-01 5.82721949e-01 6.31848156e-01 -1.58106279e+00
-1.07428002e+00 -8.75452757e-01 8.53219181e-02 4.77515697e-01
-2.49185711e-01 -6.67776346e-01 -5.35584927e-01 -1.39497235e-01
1.00076199e+00 -5.84281564e-01 -2.56649047e-01 3.37074876e-01
-2.92093515e-01 -1.10272355e-01 5.27823985e-01 -5.27429581e-01
8.81833732e-01 -1.96323228e+00 -1.10670872e-01 2.40597039e-01
2.97359049e-01 -2.77936310e-01 2.05908194e-01 1.26306742e-01
4.65500444e-01 -2.61607468e-01 -6.70463964e-02 -4.83398050e-01
-1.70239940e-01 1.58878461e-01 1.50196748e-02 6.07799530e-01
-1.10861607e-01 5.54888964e-01 -8.91190350e-01 -8.57483268e-01
6.32358253e-01 4.28613722e-01 -5.19794703e-01 -3.56490649e-02
-6.13220036e-02 4.39194143e-01 -8.21758270e-01 9.65983748e-01
9.52960908e-01 -2.90496387e-02 7.30101839e-02 -2.74121642e-01
-2.03643993e-01 2.58770645e-01 -1.38147008e+00 1.85210586e+00
-6.28834248e-01 2.69480616e-01 3.41555402e-02 -9.00190413e-01
1.03222883e+00 -1.13732599e-01 6.03886127e-01 -1.22417402e+00
1.61571465e-02 4.74207699e-01 -5.20223200e-01 -4.20805216e-01
9.02827024e-01 -1.88338667e-01 1.09083168e-02 6.74173161e-02
-4.14380342e-01 -2.19342336e-01 -5.04040122e-01 9.69035700e-02
6.73191249e-01 4.11988169e-01 4.72512871e-01 -1.85940892e-01
4.98623639e-01 1.51110599e-02 6.73548281e-01 4.31403190e-01
-4.85256761e-01 6.91241503e-01 3.52224052e-01 -4.40567166e-01
-1.05229986e+00 -1.19741940e+00 -1.51522696e-01 3.92955750e-01
7.15352535e-01 -2.12259859e-01 -5.53854644e-01 -7.40020335e-01
-1.03903264e-01 6.19218051e-01 -4.56336349e-01 4.10836637e-02
-2.94832140e-01 -3.50662887e-01 4.74335719e-03 4.50223148e-01
6.47936583e-01 -7.73298979e-01 -1.21136105e+00 7.91378692e-02
-3.43408972e-01 -1.07815087e+00 1.23125881e-01 7.01223388e-02
-1.27652085e+00 -8.87978017e-01 -5.78071952e-01 -4.98228014e-01
8.69066358e-01 8.26037347e-01 1.14656866e+00 -5.05732298e-02
-8.27258173e-03 1.77838951e-01 -3.34053218e-01 -1.31895691e-01
4.90194820e-02 5.96032515e-02 7.36938184e-03 -2.51819104e-01
1.84064373e-01 -6.68510675e-01 -8.85316849e-01 4.15784448e-01
-6.41903460e-01 2.84758955e-01 7.11815357e-01 6.01823032e-01
1.05011427e+00 3.39629233e-01 6.47602022e-01 -8.09082210e-01
2.34669089e-01 -3.81225348e-01 -8.46063972e-01 -3.35869715e-02
-6.83364511e-01 8.36777408e-03 4.18884099e-01 3.27936523e-02
-1.16004765e+00 2.25604936e-01 -2.94533968e-01 -2.42640659e-01
-1.25923365e-01 1.03374302e-01 -5.21699548e-01 5.25479652e-02
5.27677894e-01 2.47809261e-01 -1.52275234e-01 -3.17227811e-01
1.34697080e-01 6.15212739e-01 3.96893829e-01 -4.96920884e-01
5.80711007e-01 8.80309284e-01 1.05050370e-01 -9.13888395e-01
-6.35499239e-01 -5.08835495e-01 -4.59539741e-01 -4.91304934e-01
4.41672176e-01 -1.04168582e+00 -7.64585197e-01 2.23788708e-01
-1.00335789e+00 1.11578539e-01 -2.09384009e-01 4.93227780e-01
-7.57537663e-01 5.14720678e-01 -2.03880653e-01 -1.22234106e+00
-1.82951912e-01 -1.07757545e+00 1.46336567e+00 1.61499605e-01
-9.43316333e-03 -6.16949916e-01 -6.75083995e-02 5.17202556e-01
2.24189777e-02 2.18277812e-01 6.52744830e-01 5.41442223e-02
-1.02269745e+00 -1.74564868e-01 -3.34162027e-01 3.87469195e-02
1.12488948e-01 -4.00847763e-01 -1.24625075e+00 1.09612113e-02
-2.33910661e-02 -1.50091082e-01 6.68857098e-01 5.11363387e-01
1.14128184e+00 1.26363263e-01 -3.20699751e-01 5.67130148e-01
1.64099050e+00 1.80926055e-01 7.58796751e-01 5.78662395e-01
4.97005403e-01 9.35672998e-01 1.00625753e+00 7.14869738e-01
5.25897324e-01 7.28619576e-01 8.36507261e-01 8.70608911e-02
9.08209826e-04 -6.08354092e-01 -2.12817383e-03 6.35238051e-01
-2.46118661e-02 -1.33131281e-01 -6.36025846e-01 6.50872648e-01
-1.76224577e+00 -7.29742229e-01 -9.93604586e-02 2.39914083e+00
4.52161670e-01 4.78767931e-01 1.15683321e-02 4.52728093e-01
3.55671287e-01 4.56829928e-02 -6.02040112e-01 5.51240984e-03
-1.25823423e-01 -1.95631739e-02 6.44230783e-01 5.60116053e-01
-8.03069234e-01 8.03154051e-01 5.35281992e+00 9.89701748e-01
-7.33446896e-01 1.31517306e-01 6.02671444e-01 -7.43677244e-02
-6.58717930e-01 4.14559990e-02 -8.75769854e-01 4.57937747e-01
3.26559961e-01 1.63572773e-01 3.01438551e-02 1.06680393e+00
3.66487741e-01 -7.02845514e-01 -7.77001560e-01 1.27814007e+00
5.95518984e-02 -1.21891046e+00 -1.58539772e-01 1.05581798e-01
6.21736169e-01 -2.76911646e-01 -3.76344100e-02 -1.72628060e-01
-2.15892002e-01 -6.41096175e-01 1.05325127e+00 3.74812335e-01
7.89168537e-01 -9.36134756e-01 7.05208719e-01 6.86376929e-01
-1.04107356e+00 2.01352611e-02 -4.97416317e-01 -3.79257873e-02
2.66740322e-01 1.06352997e+00 -8.41689765e-01 8.05216074e-01
6.67624772e-01 4.58182275e-01 -3.44496489e-01 1.08940136e+00
-2.39430308e-01 -7.73245618e-02 -3.03586215e-01 7.90431201e-02
7.24534169e-02 -2.57782787e-01 3.89695585e-01 6.09389484e-01
5.23748696e-01 5.99645451e-02 -1.19553104e-01 7.27759480e-01
3.01790684e-01 4.80267741e-02 -7.15960503e-01 5.66471517e-01
7.27308095e-01 1.14491320e+00 -1.21783412e+00 -1.63219407e-01
-3.47331941e-01 1.06720400e+00 2.36240655e-01 -1.67509243e-02
-6.93227172e-01 -4.80105095e-02 4.34722602e-01 4.87196714e-01
7.37892687e-02 -3.99082899e-01 -7.49436975e-01 -1.05475914e+00
3.95960540e-01 -4.70818192e-01 2.21417412e-01 -6.97834849e-01
-1.03775561e+00 5.23440003e-01 -9.92698744e-02 -1.62830353e+00
-1.24404870e-03 -3.26947331e-01 -3.10657974e-02 6.71555877e-01
-1.92530596e+00 -1.01056886e+00 -6.16912007e-01 4.69304889e-01
8.64345729e-01 2.55112857e-01 3.51421237e-01 4.90160763e-01
-1.46548092e-01 4.20085609e-01 -2.03420743e-01 -5.73016644e-01
4.07509267e-01 -1.03994000e+00 3.51750225e-01 6.85703933e-01
-3.96120325e-02 1.93450347e-01 6.16403878e-01 -6.21486366e-01
-1.37625313e+00 -6.35353506e-01 9.04247761e-01 -4.50878799e-01
1.47368275e-02 -3.61510426e-01 -6.56869948e-01 2.21205413e-01
-4.19448227e-01 -9.33772549e-02 5.13737202e-02 -1.01803459e-01
1.22979679e-03 -2.07580045e-01 -1.37802923e+00 5.93438983e-01
1.31781948e+00 -4.42484587e-01 -3.40681314e-01 6.80568367e-02
3.73195440e-01 -4.93634939e-01 -3.13133061e-01 5.43013275e-01
8.26192200e-01 -1.64586174e+00 1.10753894e+00 1.39432341e-01
4.58063722e-01 -4.59281594e-01 -4.40248698e-01 -8.31703067e-01
1.20862231e-01 -2.01870680e-01 -8.81681889e-02 9.17052448e-01
1.10979043e-01 -4.76720661e-01 1.35036695e+00 4.38000947e-01
-2.45666698e-01 -7.46284842e-01 -1.13529170e+00 -8.04608643e-01
-2.84703046e-01 -6.39075100e-01 4.43131924e-01 8.33180189e-01
-3.75955492e-01 5.15668131e-02 -4.13401157e-01 4.38898444e-01
9.79274154e-01 1.26122162e-01 1.00704741e+00 -1.36450422e+00
-1.52723134e-01 -2.34885544e-01 -5.06607890e-01 -1.42494035e+00
-7.88798928e-02 -3.24436426e-01 1.95371345e-01 -1.69650960e+00
6.26957491e-02 -1.01544034e+00 -7.73751661e-02 1.11133829e-01
-3.54890898e-02 2.86358029e-01 -1.98249295e-01 2.93871552e-01
-6.37210131e-01 5.80768287e-01 1.22769201e+00 2.72192627e-01
-2.95008808e-01 1.91990480e-01 -5.78328371e-01 8.18586051e-01
8.58623385e-01 -2.36333519e-01 -7.95031905e-01 -5.19885421e-01
4.69492793e-01 2.97706068e-01 4.91676748e-01 -1.08142257e+00
1.77684322e-01 -2.37909466e-01 3.64792287e-01 -1.09550202e+00
6.33418262e-01 -9.74472344e-01 1.74428627e-01 2.72638083e-01
1.05756193e-01 -2.90556680e-02 -4.14270833e-02 5.97727954e-01
-1.02413438e-01 -1.83849752e-01 7.15937376e-01 -2.52487272e-01
-1.16047132e+00 1.44562230e-01 1.03854723e-01 -2.83367425e-01
1.14004898e+00 -9.23330247e-01 -1.26574729e-02 -3.36097538e-01
-4.23705429e-01 5.21646179e-02 7.16861248e-01 3.09040964e-01
9.87720013e-01 -1.11772859e+00 -3.52470696e-01 4.17670399e-01
3.08187336e-01 2.44724542e-01 5.23096025e-01 6.99377775e-01
-4.40562993e-01 2.48550296e-01 -1.89623848e-01 -7.50656366e-01
-1.07718301e+00 2.73621291e-01 1.77947879e-01 1.43353269e-01
-8.60431790e-01 8.05468261e-01 1.66188568e-01 -2.58094221e-01
3.27856302e-01 -1.69176504e-01 -7.08773453e-03 -2.74208467e-02
3.18591386e-01 5.63743532e-01 2.69088924e-01 -3.46141249e-01
-4.65473861e-01 8.01447570e-01 1.43170238e-01 -2.19328299e-01
1.13660312e+00 -6.51906729e-01 2.39431992e-01 3.23088080e-01
9.06984806e-01 1.88716173e-01 -1.48712814e+00 -1.14002645e-01
2.11792719e-02 -9.70224917e-01 3.61327887e-01 -5.19072175e-01
-1.00805879e+00 7.06390560e-01 7.34660983e-01 7.02091306e-02
1.29334152e+00 7.78322667e-02 8.49580288e-01 1.29162490e-01
1.02817965e+00 -1.12991977e+00 -4.01046798e-02 -1.13276929e-01
4.16194737e-01 -1.49412858e+00 2.33969197e-01 -5.28419137e-01
-3.99553657e-01 8.15161526e-01 7.60086775e-01 6.12981431e-02
4.31235462e-01 2.21200377e-01 -1.10111646e-01 -3.17660242e-01
-5.24100482e-01 -8.30235854e-02 -1.55675739e-01 6.95407271e-01
1.88534304e-01 -1.26022696e-02 -5.37942946e-01 4.14633214e-01
-1.22612864e-01 -8.30117017e-02 4.47091728e-01 1.02033770e+00
-7.39737272e-01 -8.74135554e-01 -5.24832845e-01 2.59543449e-01
-8.24668780e-02 1.69835016e-01 -3.64501998e-02 7.72256672e-01
3.01292479e-01 8.61732900e-01 1.54917449e-01 -3.12105328e-01
3.95907551e-01 -3.73353690e-01 6.14862025e-01 -2.73546129e-01
1.41365469e-01 1.50501326e-01 -1.22319078e-02 -7.34885812e-01
-5.76195955e-01 -5.84782362e-01 -1.17695260e+00 -2.81561524e-01
-4.32570934e-01 2.54351422e-02 8.75709593e-01 6.35892451e-01
3.18160594e-01 3.00601900e-01 9.24019873e-01 -6.67354286e-01
-5.07546306e-01 -4.57240760e-01 -6.74468875e-01 1.56793538e-02
3.58371735e-01 -9.60646510e-01 -4.00114417e-01 -3.04268122e-01] | [8.760688781738281, -2.1941099166870117] |
c68b3ae3-9619-42f5-b3c9-c1cc4f9165b6 | 3d-bounding-box-estimation-using-deep | 1612.00496 | null | http://arxiv.org/abs/1612.00496v2 | http://arxiv.org/pdf/1612.00496v2.pdf | 3D Bounding Box Estimation Using Deep Learning and Geometry | We present a method for 3D object detection and pose estimation from a single
image. In contrast to current techniques that only regress the 3D orientation
of an object, our method first regresses relatively stable 3D object properties
using a deep convolutional neural network and then combines these estimates
with geometric constraints provided by a 2D object bounding box to produce a
complete 3D bounding box. The first network output estimates the 3D object
orientation using a novel hybrid discrete-continuous loss, which significantly
outperforms the L2 loss. The second output regresses the 3D object dimensions,
which have relatively little variance compared to alternatives and can often be
predicted for many object types. These estimates, combined with the geometric
constraints on translation imposed by the 2D bounding box, enable us to recover
a stable and accurate 3D object pose. We evaluate our method on the challenging
KITTI object detection benchmark both on the official metric of 3D orientation
estimation and also on the accuracy of the obtained 3D bounding boxes. Although
conceptually simple, our method outperforms more complex and computationally
expensive approaches that leverage semantic segmentation, instance level
segmentation and flat ground priors and sub-category detection. Our
discrete-continuous loss also produces state of the art results for 3D
viewpoint estimation on the Pascal 3D+ dataset. | ['John Flynn', 'Jana Kosecka', 'Dragomir Anguelov', 'Arsalan Mousavian'] | 2016-12-01 | 3d-bounding-box-estimation-using-deep-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Mousavian_3D_Bounding_Box_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Mousavian_3D_Bounding_Box_CVPR_2017_paper.pdf | cvpr-2017-7 | ['vehicle-pose-estimation', 'viewpoint-estimation'] | ['computer-vision', 'computer-vision'] | [-1.45129450e-02 1.70341820e-01 -6.76354915e-02 -5.02734303e-01
-8.11887026e-01 -8.16188276e-01 6.29493237e-01 3.22628021e-01
-7.49965429e-01 3.74085754e-02 -3.56603593e-01 -3.73617522e-02
2.84546584e-01 -4.73886430e-01 -1.03303194e+00 -4.12491500e-01
-8.16062838e-02 9.14793491e-01 8.19892645e-01 2.88242877e-01
3.96891832e-01 1.02627563e+00 -1.51181972e+00 -1.54757619e-01
4.50714946e-01 1.55529606e+00 5.51554039e-02 6.79006636e-01
1.79696381e-02 1.68695822e-01 -6.01843536e-01 -1.99341550e-01
7.56832600e-01 2.20060796e-01 -4.56870139e-01 3.33401322e-01
1.19462228e+00 -6.89495921e-01 2.72492796e-01 8.71368945e-01
3.54428291e-01 -5.82940504e-02 9.76595759e-01 -1.07686031e+00
-6.28113672e-02 -1.62652414e-02 -7.93843865e-01 -1.34181827e-01
2.53730923e-01 2.08653778e-01 8.70441377e-01 -8.70301247e-01
7.88838148e-01 1.46559715e+00 9.36983347e-01 1.42598003e-01
-1.44678736e+00 -4.37277675e-01 4.91453588e-01 -2.52561361e-01
-1.41570258e+00 9.00938548e-03 5.97576380e-01 -7.24184275e-01
1.20174742e+00 -3.53461243e-02 6.94417834e-01 6.77095413e-01
1.42943814e-01 8.92943501e-01 9.82675552e-01 -3.49761218e-01
7.41792768e-02 1.72576353e-01 1.65726081e-01 7.82622576e-01
3.49292696e-01 7.33002871e-02 -5.08738123e-02 1.12269767e-01
6.82982445e-01 -9.61309597e-02 2.09340185e-01 -1.05887485e+00
-1.02466488e+00 6.84535682e-01 6.73470974e-01 -3.48967969e-01
-2.67832220e-01 2.42505401e-01 3.45022976e-01 -2.48996139e-01
6.99072480e-01 2.54399210e-01 -5.94313979e-01 3.90234999e-02
-9.01882052e-01 6.46596432e-01 7.47204781e-01 1.08733082e+00
8.67833257e-01 -2.17585430e-01 -1.51818573e-01 6.34757340e-01
5.06833434e-01 7.52909660e-01 -2.62860507e-01 -1.14697313e+00
4.02357996e-01 7.35256612e-01 3.51854891e-01 -8.64417136e-01
-5.99777341e-01 -5.63701630e-01 -5.60862459e-02 7.40678310e-01
7.75524378e-01 2.21850842e-01 -1.31149673e+00 1.51482749e+00
8.63407671e-01 -3.78473848e-01 -4.95128423e-01 1.17754149e+00
6.47031009e-01 2.68679142e-01 -1.91572487e-01 4.24960703e-01
1.34785080e+00 -8.13513279e-01 5.95342778e-02 -3.96204710e-01
5.14003396e-01 -7.53055692e-01 7.92183161e-01 4.05186236e-01
-1.11325037e+00 -5.53183079e-01 -1.08652258e+00 -3.42879474e-01
-4.31507856e-01 3.27005059e-01 4.42072392e-01 7.27841496e-01
-7.69909382e-01 5.04312754e-01 -1.04637861e+00 -1.93504840e-01
7.65522718e-01 4.86538291e-01 -1.69500053e-01 8.15109685e-02
-4.27138984e-01 1.29840112e+00 5.41863978e-01 -3.81924175e-02
-8.05991352e-01 -8.71406078e-01 -9.97455657e-01 -2.16804221e-01
6.36673391e-01 -5.59564471e-01 1.29621816e+00 -5.14121234e-01
-1.34439921e+00 1.27974224e+00 2.47939959e-01 -6.78229690e-01
1.04966259e+00 -5.14515758e-01 5.30372024e-01 7.29114860e-02
1.24694929e-01 1.15033293e+00 1.03244722e+00 -1.16362154e+00
-6.70652926e-01 -6.99329019e-01 4.71330322e-02 3.11338931e-01
2.04697222e-01 -2.13671342e-01 -5.77403784e-01 -2.77097106e-01
4.48334068e-01 -9.43924010e-01 -1.80640623e-01 6.09744906e-01
-5.06402254e-01 -3.43810797e-01 9.83235002e-01 -6.19968653e-01
4.67815757e-01 -1.99090493e+00 5.19612432e-02 1.03320211e-01
2.10990697e-01 8.26261938e-02 6.67128041e-02 -2.83179671e-01
2.48697564e-01 -1.12750214e-02 -3.21092963e-01 -5.89106083e-01
3.01664591e-01 -2.01229349e-01 -7.89886564e-02 8.13368917e-01
5.49315572e-01 7.60743737e-01 -6.93331122e-01 -4.09896672e-01
5.57346463e-01 7.16592908e-01 -7.61307955e-01 1.74634010e-01
-5.30134678e-01 4.31599654e-02 -2.96464801e-01 7.64714658e-01
9.14958537e-01 -7.28038028e-02 -3.87042880e-01 -3.50431085e-01
-1.38772175e-01 3.58569801e-01 -1.23696923e+00 1.56200600e+00
-3.60514015e-01 5.15471637e-01 3.05549186e-02 -7.80794322e-01
1.12115479e+00 -2.92241752e-01 4.85634297e-01 -2.64265239e-01
3.96702498e-01 3.68252218e-01 -2.23502234e-01 3.43621485e-02
4.00024205e-01 2.54618004e-03 -6.79162368e-02 3.16391170e-01
1.38347387e-01 -9.88365948e-01 -1.38914194e-02 -4.18462455e-02
7.40569711e-01 8.07494402e-01 3.03378642e-01 -1.89328581e-01
1.93976566e-01 1.53123334e-01 3.11995596e-01 6.95364416e-01
-3.11365843e-01 8.26020122e-01 6.62169337e-01 -5.41587651e-01
-1.35421741e+00 -1.19489217e+00 -4.00162369e-01 5.92175484e-01
4.21633005e-01 -9.65459943e-02 -8.06173682e-01 -1.14964843e+00
5.14832258e-01 5.80493152e-01 -5.56989312e-01 1.22677989e-01
-6.12141013e-01 -5.66961408e-01 2.23755062e-01 6.94913805e-01
4.06253815e-01 -4.61475521e-01 -9.37242568e-01 9.08681229e-02
2.25309193e-01 -1.49391627e+00 -4.72028077e-01 4.88940060e-01
-8.65921021e-01 -1.10095060e+00 -5.02779782e-01 -4.79813099e-01
6.61669195e-01 5.53917140e-03 1.13557100e+00 -3.08776408e-01
-6.53188944e-01 2.24069580e-01 -1.72018871e-01 -7.93365359e-01
-2.06553474e-01 3.57588530e-02 3.29306424e-02 -2.46579945e-01
5.44998527e-01 -1.37926847e-01 -5.72286904e-01 3.54889572e-01
-4.22539383e-01 -2.48036329e-02 5.71497619e-01 5.16166210e-01
9.43035603e-01 -1.83301270e-01 -1.45295318e-02 -6.71775937e-01
-2.29005411e-01 -4.86621074e-03 -1.21950662e+00 -1.77620143e-01
-4.03621733e-01 6.39317632e-02 1.87184557e-01 -6.24279976e-01
-7.43863285e-01 7.50807881e-01 -8.65624100e-02 -6.23210371e-01
-4.24346209e-01 -3.47241044e-01 -3.95529792e-02 -1.26522794e-01
7.17957616e-01 -2.60250568e-01 1.23745464e-01 -6.14857197e-01
3.67920816e-01 4.61193621e-01 4.86133248e-01 -3.69641721e-01
9.62093890e-01 7.08012700e-01 3.56344134e-01 -7.07285225e-01
-1.21138501e+00 -4.89679337e-01 -1.09308374e+00 -2.25847676e-01
1.17602801e+00 -1.02315593e+00 -7.65080810e-01 4.80957091e-01
-1.31181073e+00 -3.30272496e-01 -3.61542046e-01 5.38151264e-01
-7.30476737e-01 1.16152696e-01 -3.41553211e-01 -9.63649452e-01
-1.97162956e-01 -1.15274608e+00 1.78159046e+00 -2.54540920e-01
-3.45926046e-01 -5.65468669e-01 -3.90647352e-01 4.33463961e-01
3.51260826e-02 5.91047049e-01 7.25836575e-01 -5.18135369e-01
-9.06220794e-01 -5.73760688e-01 -5.34747422e-01 5.33421278e-01
-1.98622674e-01 1.02715623e-02 -9.56274271e-01 -1.34991378e-01
-2.35366505e-02 -3.72919589e-01 9.91102219e-01 6.41739249e-01
9.50999260e-01 1.87545732e-01 -3.22695255e-01 7.04516828e-01
1.34184670e+00 -2.91169226e-01 -4.04784493e-02 2.86011547e-01
8.18376780e-01 6.67348444e-01 9.04671967e-01 3.31239194e-01
3.04236472e-01 7.86329746e-01 8.20955276e-01 -6.16113804e-02
-2.58037269e-01 -1.12517983e-01 1.43397227e-01 5.72529584e-02
1.50289740e-02 2.78245937e-02 -8.65421355e-01 2.78726578e-01
-1.46296334e+00 -3.41003776e-01 -2.02747405e-01 2.26715779e+00
4.65382695e-01 7.06154346e-01 5.21984398e-01 2.05809064e-02
5.79857767e-01 5.03481179e-03 -9.57474828e-01 -2.55597800e-01
7.02435672e-02 2.46385373e-02 9.79202449e-01 5.51505446e-01
-1.38803744e+00 9.86577094e-01 6.14936113e+00 7.05495834e-01
-1.00276232e+00 -8.11110064e-02 4.83403891e-01 -3.43766898e-01
1.94118321e-01 -9.34992880e-02 -1.46043789e+00 1.68287069e-01
3.12685102e-01 4.54447091e-01 -2.27446202e-02 1.19297874e+00
6.11076392e-02 -2.93922782e-01 -1.41002214e+00 8.31064999e-01
8.09126422e-02 -9.25136089e-01 -6.37101680e-02 1.66652501e-01
5.42906523e-01 2.81604826e-01 7.45336711e-03 1.06140062e-01
1.36225000e-01 -7.36302555e-01 1.41178989e+00 1.69836730e-01
6.93309665e-01 -7.15126395e-01 5.94926834e-01 5.67887306e-01
-9.59983170e-01 -5.07337600e-02 -3.10668349e-01 5.43440990e-02
-1.89251192e-02 5.58812976e-01 -1.34944785e+00 6.74263537e-02
8.64957213e-01 5.84740758e-01 -6.74213350e-01 1.17059505e+00
-2.33293489e-01 2.00705007e-01 -8.87237549e-01 -2.53386229e-01
2.66996980e-01 -7.69069791e-02 9.36631799e-01 1.27104378e+00
-2.75913579e-03 -2.61195660e-01 3.05892855e-01 1.27931261e+00
-1.88110061e-02 -6.76798224e-02 -4.51063335e-01 2.61475980e-01
3.68780673e-01 1.15598619e+00 -1.06703627e+00 -2.15511829e-01
-1.64976835e-01 6.10687733e-01 3.71101499e-01 -1.09959040e-02
-7.90544629e-01 -3.13914806e-01 6.54473305e-01 4.23879415e-01
8.07256699e-01 -5.17547488e-01 -5.30277729e-01 -9.30909753e-01
3.83901417e-01 -4.15367067e-01 1.00852683e-01 -6.52372420e-01
-1.10822403e+00 4.38812822e-01 4.02135253e-01 -1.11442125e+00
-3.34946334e-01 -1.10401416e+00 -1.86738241e-02 7.88370788e-01
-1.39301407e+00 -1.18067384e+00 -2.82167077e-01 -9.15215909e-02
6.13707125e-01 2.43577629e-01 3.91709089e-01 -3.17341648e-03
-1.80548877e-01 3.12646627e-01 -3.59811813e-01 5.61471134e-02
4.64893699e-01 -1.52447104e+00 6.57506168e-01 6.00247741e-01
7.97892809e-02 1.67940944e-01 6.64644837e-01 -5.87856293e-01
-1.26146722e+00 -1.03880811e+00 4.83059168e-01 -1.02084851e+00
2.59123862e-01 -9.97835398e-01 -7.73570478e-01 5.38393736e-01
-5.80557227e-01 2.62052506e-01 -5.84499761e-02 -5.03693260e-02
-6.03734374e-01 -1.17474332e-01 -1.32108772e+00 2.93909073e-01
1.03441548e+00 -2.14258566e-01 -5.65907180e-01 4.12514746e-01
7.96283305e-01 -7.54685521e-01 -6.82112992e-01 7.08235562e-01
7.55835414e-01 -9.61037815e-01 1.28435373e+00 -3.64447743e-01
-1.14007071e-02 -3.83843988e-01 -1.81191117e-01 -8.26694012e-01
3.10206376e-02 -2.33662218e-01 -1.16811022e-01 6.62737012e-01
3.41224015e-01 -3.18566024e-01 1.11592197e+00 5.67679048e-01
-8.29613507e-02 -8.22043598e-01 -9.34734046e-01 -1.01811719e+00
7.76406229e-02 -7.74075150e-01 3.32892954e-01 3.18251550e-01
-9.41896081e-01 1.49242565e-01 1.56130061e-01 2.56513357e-01
1.09497643e+00 3.48526359e-01 9.68736053e-01 -1.39082193e+00
-1.17315739e-01 -6.74137771e-01 -8.46387744e-01 -1.45290577e+00
1.13335356e-01 -7.87562728e-01 3.11227858e-01 -1.08768225e+00
6.11159280e-02 -3.78905743e-01 1.21327981e-01 3.50918472e-01
7.23416060e-02 5.77514827e-01 1.52285904e-01 -7.38324374e-02
-5.62234163e-01 3.33821774e-01 1.11833966e+00 -1.07672855e-01
-3.70120466e-01 4.49664965e-02 -1.39314651e-01 1.11185873e+00
4.08959568e-01 -4.77013111e-01 9.65411216e-02 -4.62493420e-01
-6.91940486e-02 -4.06292409e-01 6.90149069e-01 -8.64340067e-01
-1.21961348e-01 8.88866782e-02 7.84800887e-01 -1.24727118e+00
6.89492345e-01 -8.63842607e-01 -4.42420483e-01 4.18044031e-01
-1.43819734e-01 -5.65951943e-01 2.73628324e-01 4.08079773e-01
2.79491037e-01 -1.19419411e-01 1.13823295e+00 -7.58996159e-02
-5.53145468e-01 4.04601067e-01 4.54653576e-02 7.90028423e-02
1.05449891e+00 -7.25451887e-01 2.01908886e-01 1.99321546e-02
-7.21636057e-01 1.67915896e-01 6.01108193e-01 5.16527295e-01
5.39234042e-01 -1.05752361e+00 -6.95019603e-01 3.40757430e-01
-7.49999797e-03 6.19795203e-01 -2.69725680e-01 7.31287241e-01
-7.45550692e-01 3.26373190e-01 -1.26382597e-02 -1.35207903e+00
-1.34413290e+00 4.36404616e-01 5.78498483e-01 1.35075897e-01
-6.63338721e-01 9.23389494e-01 2.83918053e-01 -5.97688794e-01
5.35687089e-01 -7.22812116e-01 2.49080524e-01 8.94476697e-02
1.91302598e-01 2.42527217e-01 1.66788161e-01 -6.08705819e-01
-5.40605605e-01 9.47136760e-01 -5.55218682e-02 2.84495428e-02
1.42638659e+00 -1.23860829e-01 2.50452280e-01 4.30300921e-01
1.41812468e+00 -3.13670844e-01 -1.94026780e+00 -2.05707431e-01
3.20442319e-02 -6.94174886e-01 1.82283312e-01 -7.33526051e-01
-7.99440324e-01 9.85246956e-01 7.63329685e-01 5.93400002e-02
6.18644476e-01 2.60544956e-01 5.52810550e-01 4.77989852e-01
2.35618114e-01 -1.11716127e+00 1.45876929e-01 5.96523643e-01
7.58848786e-01 -1.46345830e+00 3.21367085e-01 -6.82831407e-01
-1.88897729e-01 1.08330894e+00 6.82526708e-01 -3.99318665e-01
5.17708063e-01 3.93105745e-01 2.82361098e-02 -6.98313862e-02
-2.87131786e-01 -3.40901375e-01 7.23172843e-01 5.31582713e-01
1.27503455e-01 -9.34655368e-02 9.44676399e-02 2.41932943e-01
-3.22987199e-01 -4.89939123e-01 9.84812453e-02 8.10360551e-01
-5.81318855e-01 -7.16417074e-01 -5.01362681e-01 3.26758355e-01
-4.18708712e-01 2.76883543e-01 -5.81929147e-01 1.11532211e+00
3.32701176e-01 3.39506567e-01 3.65191191e-01 -8.77633616e-02
5.57154059e-01 -4.19496931e-02 8.62445712e-01 -7.93811619e-01
-2.18574524e-01 2.65882254e-01 -6.11683615e-02 -6.96210742e-01
-2.40981489e-01 -7.45314896e-01 -1.15937245e+00 9.87083688e-02
-6.39195502e-01 -4.07304168e-01 1.22335815e+00 7.91876853e-01
5.66185005e-02 1.20818868e-01 3.56984377e-01 -1.50720549e+00
-8.29662204e-01 -7.95155048e-01 -3.75377476e-01 3.93131226e-01
4.30094957e-01 -9.72745359e-01 -4.41420794e-01 -1.19295254e-01] | [7.622430324554443, -2.734161376953125] |
2a140041-1ab5-479d-9ee7-6331beaed57c | do-multilingual-language-models-capture | 2203.09904 | null | https://arxiv.org/abs/2203.09904v1 | https://arxiv.org/pdf/2203.09904v1.pdf | Do Multilingual Language Models Capture Differing Moral Norms? | Massively multilingual sentence representations are trained on large corpora of uncurated data, with a very imbalanced proportion of languages included in the training. This may cause the models to grasp cultural values including moral judgments from the high-resource languages and impose them on the low-resource languages. The lack of data in certain languages can also lead to developing random and thus potentially harmful beliefs. Both these issues can negatively influence zero-shot cross-lingual model transfer and potentially lead to harmful outcomes. Therefore, we aim to (1) detect and quantify these issues by comparing different models in different languages, (2) develop methods for improving undesirable properties of the models. Our initial experiments using the multilingual model XLM-R show that indeed multilingual LMs capture moral norms, even with potentially higher human-agreement than monolingual ones. However, it is not yet clear to what extent these moral norms differ between languages. | ['Kristian Kersting', 'Alexander Fraser', 'Jindřich Libovický', 'Patrick Schramowski', 'Björn Deiseroth', 'Katharina Hämmerl'] | 2022-03-18 | null | null | null | null | ['xlm-r'] | ['natural-language-processing'] | [-2.56621540e-01 2.30303809e-01 -1.05229467e-01 -3.84204984e-01
-8.54870856e-01 -7.54487097e-01 7.08244383e-01 5.66934571e-02
-9.04146731e-01 1.14884996e+00 5.81634998e-01 -3.35375428e-01
2.61332095e-01 -5.82048953e-01 -7.14914620e-01 -5.29750466e-01
1.93723828e-01 7.89554417e-01 -4.24363405e-01 -3.70285243e-01
8.31760988e-02 -8.30103457e-02 -1.14228439e+00 3.05725873e-01
1.26871228e+00 -2.94471104e-02 2.66292226e-02 2.60137796e-01
1.31310254e-01 1.01118875e+00 -6.64529204e-01 -9.00000930e-01
2.21090168e-01 -4.64823037e-01 -6.26759410e-01 -3.11191559e-01
4.70373333e-01 -1.92194372e-01 2.86995769e-01 1.31058145e+00
4.74752843e-01 -1.59866974e-01 1.00348175e+00 -8.32826018e-01
-1.04346919e+00 1.29804349e+00 -5.09788454e-01 -8.25124010e-02
1.90455943e-01 1.72058553e-01 1.02956939e+00 -6.74793005e-01
9.08681750e-01 1.78126967e+00 6.46339834e-01 7.84176350e-01
-1.48800778e+00 -7.90934563e-01 -1.31465290e-02 -2.70368103e-02
-1.30308831e+00 -6.24573767e-01 8.17530811e-01 -6.15468740e-01
7.77745128e-01 3.66644524e-02 5.00040412e-01 1.66817844e+00
3.38297904e-01 3.52090329e-01 1.81711090e+00 -3.77277553e-01
1.32101417e-01 6.83553457e-01 -5.75787723e-02 2.70405680e-01
5.28439105e-01 1.71461523e-01 -4.47249055e-01 -2.52249658e-01
4.22839671e-01 -7.59562314e-01 -2.10661471e-01 -2.15986684e-01
-1.10454774e+00 1.07105923e+00 3.20989877e-01 6.34018481e-01
-3.17124367e-01 5.77528104e-02 6.07645094e-01 7.24105477e-01
5.87574840e-01 8.29584897e-01 -6.33204043e-01 -1.96762495e-02
-6.92751408e-01 1.93971813e-01 6.80037975e-01 5.10097861e-01
7.57153273e-01 3.92017365e-02 2.18353018e-01 1.13703120e+00
1.60728022e-01 5.61971009e-01 4.81029958e-01 -9.52950776e-01
4.52158213e-01 2.37263218e-01 1.08284183e-01 -1.08451641e+00
-2.84223080e-01 -3.88900548e-01 -6.57177925e-01 2.32298240e-01
6.05308354e-01 -3.11285466e-01 -3.72617692e-01 2.34001470e+00
-5.51512018e-02 -5.94538808e-01 4.51244801e-01 9.03798938e-01
2.44305789e-01 6.04495585e-01 6.33831084e-01 -1.65917456e-01
1.16700435e+00 -3.55105847e-01 -4.51412320e-01 -6.06241226e-01
1.12632740e+00 -9.02324855e-01 1.35454524e+00 1.78570375e-01
-1.16955340e+00 -2.60134757e-01 -9.75160837e-01 -1.49925083e-01
-4.59796488e-01 -3.81605290e-02 6.66593790e-01 7.97074378e-01
-1.13894188e+00 5.23591816e-01 -4.92869765e-01 -3.62498224e-01
2.87023604e-01 1.37657389e-01 -5.10751843e-01 -2.29756638e-01
-1.72041893e+00 1.63628030e+00 2.64400482e-01 2.96887103e-02
-6.66230917e-01 -5.49237549e-01 -8.93822134e-01 -2.54705340e-01
4.05286215e-02 -5.87594450e-01 7.92554140e-01 -1.69894111e+00
-9.63308215e-01 1.37779331e+00 2.06890777e-01 -1.48757502e-01
7.74449825e-01 -2.14941457e-01 -1.75273359e-01 -4.71687138e-01
4.96634215e-01 8.79611790e-01 3.34785908e-01 -1.44005609e+00
-2.97603101e-01 -3.45030248e-01 5.81699684e-02 6.71530545e-01
-4.22314107e-01 3.98730725e-01 2.42267624e-01 -5.76989114e-01
-4.94352281e-01 -8.78554821e-01 -9.22335684e-02 -3.11311036e-01
-1.51738092e-01 -2.03581944e-01 -1.44656986e-01 -9.46587026e-01
6.14496708e-01 -1.86404932e+00 3.17423552e-01 -1.51707739e-01
-1.82176918e-01 -6.90874606e-02 -2.38256842e-01 2.35746980e-01
3.99031490e-02 5.43081880e-01 -5.25640510e-02 -1.85740769e-01
-6.60810843e-02 3.32871377e-01 -8.92096311e-02 5.56963801e-01
3.04424971e-01 6.06792927e-01 -9.07819927e-01 -7.33716726e-01
-1.07943341e-01 5.61680317e-01 -9.67119098e-01 -1.60004243e-01
-6.77596964e-03 5.74638188e-01 -1.60920650e-01 1.55228466e-01
5.49890280e-01 3.70926768e-01 4.93299752e-01 6.67318404e-02
-8.08963701e-02 7.25009859e-01 -5.93430817e-01 1.43697786e+00
-6.64185822e-01 5.97007692e-01 1.13189489e-01 -7.29837239e-01
7.45261073e-01 2.86044180e-01 2.22460255e-02 -8.10331762e-01
2.76119173e-01 6.45692647e-01 4.86280352e-01 -3.66092771e-01
6.10631287e-01 -9.45603549e-01 -4.77934182e-01 5.39217353e-01
-4.47507761e-02 -1.80978358e-01 5.43610193e-02 1.53380334e-01
4.91267145e-01 1.18723005e-01 8.43083560e-02 -1.00800908e+00
4.64119613e-01 5.29468507e-02 1.00954998e+00 4.81767267e-01
-4.05405104e-01 1.88045442e-01 8.12262416e-01 -2.59029299e-01
-1.27206635e+00 -1.02690101e+00 -4.27779734e-01 1.14474249e+00
-2.60244280e-01 2.03395542e-02 -7.17939138e-01 -6.38653934e-01
-8.67642611e-02 1.42117536e+00 -7.86096036e-01 -4.23564881e-01
-4.36757475e-01 -9.92636085e-01 8.22540760e-01 2.49361143e-01
-3.82825807e-02 -1.03169000e+00 -3.61713171e-01 9.61890444e-02
-3.56414288e-01 -8.38340998e-01 -1.11275375e-01 1.76522329e-01
-6.30574584e-01 -7.20680237e-01 -5.89310229e-01 -8.21558952e-01
7.92754412e-01 -3.26173872e-01 1.55020654e+00 -6.44667298e-02
2.55865276e-01 -2.70024035e-02 -1.94681019e-01 -5.35994887e-01
-1.22586238e+00 2.18147293e-01 3.91100764e-01 -4.22732085e-01
7.50322998e-01 -3.03638130e-01 -1.89561229e-02 1.22900769e-01
-6.46208227e-01 1.43118843e-01 3.10619295e-01 8.11805487e-01
-6.97126314e-02 -3.66391331e-01 9.59667265e-01 -1.20710397e+00
7.98289299e-01 -5.70566952e-01 -4.17150587e-01 3.21337700e-01
-4.85144615e-01 9.61756557e-02 7.84337103e-01 -3.79075110e-01
-1.09204495e+00 -6.01490319e-01 9.89328790e-03 1.19188517e-01
-7.08509460e-02 5.98380446e-01 -1.10199384e-01 5.28954864e-01
9.18072879e-01 -2.20739424e-01 -4.90263365e-02 -3.26114684e-01
3.22015435e-01 8.15219820e-01 9.74144712e-02 -1.16222799e+00
6.66718543e-01 1.95453465e-01 -6.58298314e-01 -8.00211251e-01
-7.81190515e-01 3.11952412e-01 -7.24726379e-01 -1.40592426e-01
9.15687382e-01 -1.29663420e+00 -1.26624987e-01 2.74479777e-01
-1.09854305e+00 -5.94447970e-01 -1.12286903e-01 6.99527919e-01
-3.71606141e-01 7.15064853e-02 -8.47730875e-01 -8.22903395e-01
-2.69933343e-01 -1.17265892e+00 5.31821549e-01 -3.20116282e-02
-8.02235842e-01 -1.35945630e+00 2.82903850e-01 5.43920040e-01
4.04369920e-01 -1.75007693e-02 1.45743024e+00 -6.75688684e-01
8.57848767e-03 5.35065532e-02 8.12641084e-02 6.57171488e-01
3.01179010e-02 1.82040840e-01 -1.02234030e+00 -4.79358375e-01
2.01319173e-01 -1.09738243e+00 6.21567011e-01 1.28809199e-01
1.90698072e-01 -3.22427571e-01 9.03659835e-02 3.20875049e-01
1.34947896e+00 -1.52923033e-01 5.35642624e-01 3.05198997e-01
6.20138109e-01 1.18765473e+00 4.19502288e-01 -1.74502999e-01
6.72134638e-01 3.27974260e-01 -1.58067659e-01 1.35160005e-02
2.32493103e-01 -3.07927668e-01 1.00347066e+00 1.51333213e+00
2.85387654e-02 2.09269896e-01 -1.08961952e+00 7.88514674e-01
-1.55329144e+00 -9.16725218e-01 -3.98415588e-02 2.28898025e+00
1.41446924e+00 1.11583255e-01 -2.16997266e-01 -3.36001247e-01
7.01292813e-01 1.86436176e-01 -3.15614432e-01 -9.73353744e-01
-6.15864277e-01 -2.47468963e-01 3.49707782e-01 7.91578174e-01
-7.52585769e-01 1.32144880e+00 6.43291759e+00 5.32796264e-01
-1.09412360e+00 1.78838134e-01 9.23676193e-01 -2.78293312e-01
-7.92333841e-01 1.17413953e-01 -4.64350551e-01 2.79867440e-01
1.04177737e+00 -4.73867685e-01 4.82296169e-01 7.38137364e-01
1.51394904e-01 -4.52676043e-03 -1.29506123e+00 4.43240672e-01
2.13539660e-01 -6.36254430e-01 -7.55218938e-02 6.69812635e-02
9.06326115e-01 3.02209884e-01 -2.25646440e-02 7.34286249e-01
8.68870258e-01 -1.13318980e+00 1.01090276e+00 1.43252984e-01
7.83527553e-01 -9.49598253e-01 8.18790138e-01 5.61291873e-01
-2.44664639e-01 2.20330596e-01 -9.12946463e-01 -3.90430212e-01
1.25617394e-02 3.91933471e-01 -4.81819242e-01 1.43484071e-01
4.81417745e-01 4.54096138e-01 -7.03442812e-01 1.62042663e-01
-4.51811969e-01 6.05603933e-01 -1.95545256e-01 -1.15067132e-01
1.95587113e-01 -3.12552392e-01 3.36831778e-01 9.72589612e-01
4.23821881e-02 -4.41518366e-01 -5.36264926e-02 1.12343764e+00
-2.54805237e-01 5.17295122e-01 -1.00921893e+00 -9.12679657e-02
2.79828548e-01 1.07301354e+00 -2.90390640e-01 -2.62838095e-01
-6.32910311e-01 8.08764637e-01 8.56884480e-01 2.27322981e-01
-6.81678414e-01 1.58338964e-01 7.52390802e-01 -9.38043520e-02
-5.18851101e-01 -1.66824713e-01 -3.93279523e-01 -1.52482879e+00
-2.89853573e-01 -1.35095143e+00 2.51117617e-01 -6.88583136e-01
-1.81650770e+00 3.62075180e-01 -1.17931306e-01 -6.14165366e-01
-3.89813989e-01 -6.08839273e-01 -2.52540261e-01 1.16243708e+00
-1.33147657e+00 -1.25064814e+00 6.18187010e-01 4.29732949e-01
2.16086194e-01 -1.52112916e-01 1.01542497e+00 4.36837643e-01
-4.97916490e-01 7.41448998e-01 1.19746968e-01 3.92944306e-01
1.22149956e+00 -1.06813192e+00 1.45643786e-01 5.70640922e-01
-1.28639769e-02 8.33395004e-01 9.06634808e-01 -7.42338598e-01
-7.53588855e-01 -7.81180441e-01 1.36965728e+00 -8.86454105e-01
8.45027328e-01 -4.93949354e-01 -1.07566237e+00 9.97212648e-01
5.39539933e-01 -8.10152829e-01 1.00624287e+00 6.30044580e-01
-7.23538399e-01 1.21543311e-01 -1.06763160e+00 1.20923162e+00
7.61001408e-01 -7.07137346e-01 -9.26618278e-01 2.63000697e-01
5.66359878e-01 2.81322271e-01 -8.70291293e-01 2.09636301e-01
3.81080449e-01 -1.02254283e+00 5.75389445e-01 -9.14156497e-01
8.56952369e-01 2.15910226e-01 -2.65206486e-01 -1.78225017e+00
-3.80817622e-01 -1.75213307e-01 9.59489226e-01 1.63112974e+00
7.87246048e-01 -9.44023073e-01 2.48752177e-01 8.04871440e-01
1.82885483e-01 -2.97263592e-01 -9.86797929e-01 -6.27641678e-01
1.38437188e+00 -4.11549717e-01 2.42490828e-01 1.62121665e+00
1.29129112e-01 6.16314113e-01 -3.90611172e-01 -6.54783472e-02
8.79939020e-01 -1.16596527e-01 5.07446229e-01 -1.18402660e+00
-9.74636301e-02 -4.35359508e-01 -1.04044430e-01 -5.37182810e-03
6.95832789e-01 -1.24428678e+00 7.69307986e-02 -1.13341820e+00
5.49613535e-01 -7.73736119e-01 -2.07744807e-01 3.45554054e-01
-3.59543592e-01 3.62555459e-02 2.89419740e-01 2.42494404e-01
-1.65913448e-01 5.06683171e-01 9.95824099e-01 1.18770137e-01
1.77864343e-01 -6.91367865e-01 -1.15639150e+00 1.12577128e+00
9.31488752e-01 -8.32345843e-01 -2.82873124e-01 -9.04332638e-01
6.05641842e-01 -4.30254936e-01 2.11103544e-01 -7.02022731e-01
-3.00665289e-01 -4.14454132e-01 3.39021593e-01 1.99668348e-01
1.15869142e-01 -5.53075016e-01 -4.92546745e-02 5.74504972e-01
-7.12060094e-01 8.38378817e-02 8.82498324e-02 -3.35470587e-02
-5.57531901e-02 -2.53282905e-01 9.78627682e-01 -4.45100754e-01
-9.44083109e-02 -3.82858157e-01 -4.92987365e-01 5.10781586e-01
8.73043895e-01 3.17952633e-01 -5.69423735e-01 -3.55174005e-01
-2.58194715e-01 2.24704638e-01 1.28863537e+00 6.99258268e-01
-9.38752666e-02 -1.37535560e+00 -1.31840432e+00 3.61008868e-02
1.82555541e-01 -5.04808486e-01 2.48038962e-01 6.29304051e-01
-5.40776074e-01 2.73097575e-01 -5.22722960e-01 -2.57871896e-01
-8.80017698e-01 5.67180932e-01 3.91102493e-01 -2.53956854e-01
-9.46908072e-02 8.82266223e-01 4.34209019e-01 -9.87051129e-01
-3.21813136e-01 2.94182748e-01 -1.94161668e-01 4.28248525e-01
-5.40065393e-03 -9.03870631e-03 -4.17117059e-01 -1.12474585e+00
-5.43168366e-01 5.17799318e-01 -2.88694888e-01 -5.64855218e-01
1.26459837e+00 3.73262651e-02 -4.29309219e-01 8.40498984e-01
1.16590834e+00 4.14296925e-01 -9.13025737e-01 9.04558152e-02
-1.15250172e-02 -3.11556250e-01 -2.02926919e-01 -8.21231663e-01
-6.99041069e-01 1.05179441e+00 1.21851727e-01 -1.54573560e-01
6.08050823e-01 -4.14323509e-02 3.41834068e-01 1.39519826e-01
7.06552148e-01 -1.59901965e+00 -3.21233630e-01 5.69158494e-01
1.00857365e+00 -1.18023336e+00 -9.76422504e-02 9.82067138e-02
-1.11743987e+00 5.24276018e-01 7.66292930e-01 -1.46176055e-01
2.01646805e-01 2.69963056e-01 7.83259988e-01 7.81483874e-02
-1.16690493e+00 1.18642025e-01 -1.85446441e-01 5.94857275e-01
1.18174708e+00 3.46624017e-01 -7.41993964e-01 4.50601518e-01
-5.95471144e-01 -3.40912312e-01 8.50285172e-01 5.58727145e-01
-6.13033473e-02 -1.27934432e+00 -5.71053922e-01 1.97917730e-01
-6.89552069e-01 -4.53628212e-01 -7.79170215e-01 8.75392318e-01
2.38371804e-01 7.84281194e-01 9.29012746e-02 -1.89470276e-01
1.09098598e-01 5.17411768e-01 3.32083076e-01 -5.81127644e-01
-8.89688492e-01 2.13249773e-02 5.04060984e-01 -1.38554081e-01
-2.27367461e-01 -9.13888812e-01 -8.67267668e-01 -5.56552052e-01
-1.54155508e-01 8.73995721e-02 4.85069036e-01 6.47023499e-01
1.84102863e-01 1.81928575e-01 3.64295840e-01 -6.91183746e-01
-9.29490924e-01 -1.14291346e+00 -5.73733985e-01 8.97767901e-01
-1.24665923e-01 -3.11113775e-01 -7.31225848e-01 -1.43483669e-01] | [9.986949920654297, 10.21242618560791] |
06ccd4f0-2f1a-4890-8796-460578b1dd05 | boosting-few-shot-text-classification-via | 2303.16764 | null | https://arxiv.org/abs/2303.16764v1 | https://arxiv.org/pdf/2303.16764v1.pdf | Boosting Few-Shot Text Classification via Distribution Estimation | Distribution estimation has been demonstrated as one of the most effective approaches in dealing with few-shot image classification, as the low-level patterns and underlying representations can be easily transferred across different tasks in computer vision domain. However, directly applying this approach to few-shot text classification is challenging, since leveraging the statistics of known classes with sufficient samples to calibrate the distributions of novel classes may cause negative effects due to serious category difference in text domain. To alleviate this issue, we propose two simple yet effective strategies to estimate the distributions of the novel classes by utilizing unlabeled query samples, thus avoiding the potential negative transfer issue. Specifically, we first assume a class or sample follows the Gaussian distribution, and use the original support set and the nearest few query samples to estimate the corresponding mean and covariance. Then, we augment the labeled samples by sampling from the estimated distribution, which can provide sufficient supervision for training the classification model. Extensive experiments on eight few-shot text classification datasets show that the proposed method outperforms state-of-the-art baselines significantly. | ['Xianchao Zhang', 'Hong Yu', 'Hongyang Chen', 'Xiao-Ming Wu', 'Fenglong Ma', 'Siyang Zhao', 'Xiaotong Zhang', 'Feng Zhang', 'Han Liu'] | 2023-03-26 | null | null | null | null | ['few-shot-image-classification', 'few-shot-text-classification'] | ['computer-vision', 'natural-language-processing'] | [ 3.23903680e-01 -2.12938249e-01 -4.31534350e-01 -6.02069557e-01
-8.06295455e-01 -2.22658291e-01 7.12614357e-01 1.52512029e-01
-4.67730343e-01 8.53587389e-01 7.68015087e-02 2.54001975e-01
5.68643175e-02 -7.66009867e-01 -4.85685170e-01 -9.06973481e-01
5.95548451e-01 4.84293431e-01 4.91445541e-01 3.35770249e-02
4.58691508e-01 -1.23894177e-02 -1.91905665e+00 1.86775878e-01
8.80292296e-01 1.12808406e+00 3.85624677e-01 2.73836493e-01
-5.70663452e-01 6.05569124e-01 -7.31181085e-01 -2.67966390e-01
5.03673702e-02 -6.86371744e-01 -4.51425642e-01 3.52611393e-01
2.92044312e-01 -5.83512664e-01 -2.10427687e-01 1.37639368e+00
4.04945374e-01 7.00461268e-01 1.23487902e+00 -1.29442954e+00
-7.94303596e-01 3.60709459e-01 -8.82570684e-01 1.82102010e-01
1.41348913e-01 -7.91427866e-02 6.78920865e-01 -1.05081797e+00
6.51674628e-01 1.23568809e+00 4.27229255e-01 6.16533160e-01
-9.65384960e-01 -8.47820044e-01 1.98452085e-01 3.43843967e-01
-1.46668518e+00 -5.64068198e-01 7.09792852e-01 -3.89202923e-01
4.20091957e-01 -1.67137936e-01 2.97047913e-01 1.11526716e+00
-1.86020002e-01 9.52933490e-01 8.22946787e-01 -5.92073143e-01
5.63045025e-01 5.82220733e-01 4.77252841e-01 4.13965076e-01
1.96239382e-01 -4.63870704e-01 -2.69697517e-01 -2.27517769e-01
1.86429694e-01 4.80169028e-01 -3.45965117e-01 -6.17389560e-01
-1.00130486e+00 9.86837506e-01 3.31960857e-01 2.46068403e-01
-1.79260761e-01 -1.33466855e-01 6.68025792e-01 5.90726845e-02
6.18772388e-01 -4.55822684e-02 -2.67843693e-01 4.63849911e-03
-8.22670519e-01 8.27854406e-03 4.73634362e-01 1.19340229e+00
1.04242766e+00 -2.20039785e-01 -5.32616377e-01 1.26995623e+00
1.33943781e-01 4.47450966e-01 8.40531409e-01 -5.45450985e-01
5.29732645e-01 3.46824467e-01 1.15366623e-01 -8.86460662e-01
3.07129592e-01 5.23448572e-04 -7.98275650e-01 -2.07147062e-01
3.44448119e-01 -8.01092908e-02 -1.07416034e+00 1.45876133e+00
4.82143462e-01 2.46984750e-01 8.67546797e-02 7.00695872e-01
6.10232830e-01 8.00749004e-01 7.50096962e-02 -4.38096136e-01
1.22406685e+00 -9.67283428e-01 -8.24288189e-01 -3.67043167e-01
6.00655556e-01 -5.56548476e-01 1.31209934e+00 4.60216813e-02
-4.13389951e-01 -5.38476527e-01 -1.13639712e+00 2.04051480e-01
-4.67894346e-01 2.54681800e-02 4.11927342e-01 5.32671809e-01
-3.34785432e-01 5.70547938e-01 -4.94944096e-01 -5.93204200e-01
7.72231758e-01 -2.30544686e-01 -1.61886662e-01 -5.41298509e-01
-1.27492380e+00 5.65392613e-01 7.17634201e-01 -3.18588108e-01
-7.40170896e-01 -4.94521856e-01 -8.79626334e-01 2.53692985e-01
5.63445687e-01 -3.68681043e-01 1.21599257e+00 -8.93697560e-01
-1.14636052e+00 7.03297555e-01 -2.59450912e-01 -2.36523047e-01
4.65274900e-01 6.76830485e-02 -1.56330958e-01 1.98129162e-01
4.07072306e-01 5.90348542e-01 1.29423273e+00 -1.06957006e+00
-7.17895389e-01 -4.56186265e-01 -3.59925807e-01 2.43611038e-01
-8.73867035e-01 -2.13286906e-01 -4.74067390e-01 -4.68476295e-01
5.91343194e-02 -5.60786486e-01 1.55092552e-02 3.16564977e-01
-1.66689411e-01 -3.46953124e-01 9.39243436e-01 -2.48634994e-01
1.02744603e+00 -2.42527699e+00 -3.00472885e-01 -1.30408168e-01
-7.40715787e-02 1.89658254e-01 6.46151742e-03 1.52192533e-01
2.09998935e-01 -3.02910209e-01 -1.82221219e-01 -1.61486372e-01
-6.12838082e-02 1.41310483e-01 -4.73630041e-01 2.88686067e-01
2.78264564e-02 5.55221498e-01 -1.06730628e+00 -6.52064264e-01
3.21732014e-01 1.58739358e-01 -1.80847526e-01 2.87453890e-01
-2.37214044e-01 6.15086406e-02 -5.50617039e-01 3.49610686e-01
7.29399502e-01 -2.83563286e-01 1.38824936e-02 -2.04286590e-01
4.78888899e-01 -1.17103688e-01 -1.29174542e+00 1.44711256e+00
-2.55801648e-01 5.26590407e-01 -3.91338319e-01 -1.40688944e+00
1.15223682e+00 6.39546812e-02 1.73821613e-01 -3.40720922e-01
3.58254105e-01 -9.06842500e-02 -2.76130468e-01 -6.31390393e-01
3.09214205e-01 -4.54660773e-01 -6.13014773e-02 6.04638755e-01
3.12553048e-01 -2.45707080e-01 2.35197634e-01 2.18478963e-01
6.50211453e-01 -1.16998494e-01 6.03889942e-01 -4.61454093e-02
3.64837497e-01 -1.38367578e-01 5.18340826e-01 8.96043301e-01
-5.36077142e-01 5.99968374e-01 3.36206794e-01 -1.34366080e-01
-9.14178789e-01 -1.08734477e+00 -3.11454952e-01 1.27669835e+00
2.27774024e-01 -7.05480203e-02 -6.94199860e-01 -9.08189535e-01
-1.45115450e-01 9.62739408e-01 -8.35813820e-01 -4.82813358e-01
1.42702714e-01 -5.79365313e-01 1.37267277e-01 5.44389606e-01
6.30846620e-01 -8.77269149e-01 -3.94932508e-01 1.59208149e-01
-2.76060820e-01 -9.29672360e-01 -6.10279679e-01 1.14982843e-01
-7.37374961e-01 -1.10029376e+00 -1.16440046e+00 -8.03471267e-01
8.50242853e-01 8.64771962e-01 5.87199330e-01 -1.15188122e-01
-3.84816498e-01 2.12527469e-01 -4.67270434e-01 -4.56299812e-01
-2.98148721e-01 -1.96707904e-01 8.15419033e-02 1.74030811e-01
9.03862536e-01 -2.08328262e-01 -4.62394476e-01 2.52104878e-01
-9.19225752e-01 -2.16526732e-01 2.45226622e-01 1.18696678e+00
4.10316586e-01 1.26938865e-01 8.70932639e-01 -1.10931075e+00
6.48821771e-01 -6.04549766e-01 -2.48492405e-01 4.05602127e-01
-5.26693463e-01 -1.35944739e-01 8.34833682e-01 -8.29043865e-01
-1.30597055e+00 -3.23334187e-02 3.76695454e-01 -7.11046398e-01
-2.69352317e-01 3.50966066e-01 -1.86163515e-01 3.55008215e-01
8.07025373e-01 5.05388141e-01 1.24830276e-01 -1.66896999e-01
5.15472114e-01 1.20580566e+00 2.91787177e-01 -4.81989086e-01
5.49223363e-01 6.32549822e-01 -3.92634958e-01 -1.06318569e+00
-1.34639645e+00 -9.08733070e-01 -8.08114648e-01 -5.47540523e-02
6.73338473e-01 -7.97813833e-01 5.78873120e-02 5.70704699e-01
-9.65386987e-01 1.61311835e-01 -3.39425802e-01 6.25751138e-01
-5.11954188e-01 5.68334997e-01 -3.75778198e-01 -8.73896718e-01
-3.06875765e-01 -9.36720848e-01 9.67878461e-01 4.40462232e-01
-7.65575469e-02 -9.48823333e-01 -3.01627554e-02 1.95073962e-01
1.51534036e-01 -2.88909405e-01 1.07019472e+00 -9.03149366e-01
-6.82734922e-02 -5.51464975e-01 -4.45915729e-01 4.56372589e-01
5.03850937e-01 -2.45941013e-01 -1.18112540e+00 -3.79582167e-01
2.52197117e-01 -7.08066881e-01 9.23726082e-01 4.04614031e-01
1.25618899e+00 8.22998434e-02 -5.10303020e-01 2.36211568e-01
1.17639625e+00 1.31396964e-01 5.41720867e-01 5.47939241e-02
4.85091656e-01 5.93506634e-01 1.16858840e+00 6.28851116e-01
9.41548273e-02 3.30649316e-01 -1.78195938e-01 4.25801575e-01
1.69562064e-02 -3.96761268e-01 2.30252687e-02 6.81699514e-01
3.87369901e-01 -7.17368573e-02 -5.38227439e-01 4.64732379e-01
-1.89264667e+00 -1.08369458e+00 4.04695809e-01 2.33787489e+00
9.39785540e-01 1.77385390e-01 -1.54665619e-01 1.54871792e-02
1.25207984e+00 1.14632919e-01 -8.38979423e-01 1.65764302e-01
3.24580908e-01 -8.46069083e-02 6.74193129e-02 6.06932328e-04
-1.13848889e+00 9.50592339e-01 6.04139996e+00 1.17335665e+00
-8.64737213e-01 2.55091526e-02 5.48928022e-01 1.87268928e-02
6.52685091e-02 -6.34661838e-02 -1.07513785e+00 5.70828378e-01
5.09826124e-01 -6.70686603e-01 1.77828491e-01 1.14617980e+00
-2.65684351e-02 -2.31418803e-01 -1.15277481e+00 1.16003764e+00
5.45639873e-01 -1.11079025e+00 3.85166526e-01 -2.30373144e-01
8.14286649e-01 -3.10966790e-01 -9.42886844e-02 7.75934875e-01
4.49036881e-02 -5.37419498e-01 3.63564283e-01 4.89045709e-01
8.61241996e-01 -6.42294407e-01 8.00964475e-01 6.99945092e-01
-9.87021506e-01 -1.05189510e-01 -1.09645772e+00 -2.81360615e-02
-1.85302123e-01 5.06485343e-01 -9.34070885e-01 1.13569878e-01
6.65337563e-01 9.75219071e-01 -5.78631341e-01 9.82697964e-01
-4.71117720e-02 4.74965602e-01 -8.82004946e-02 -3.96528840e-01
6.29174039e-02 -2.86163837e-01 1.74288675e-01 8.06230009e-01
3.90496790e-01 5.00978529e-02 2.13137373e-01 6.69868767e-01
-2.61068016e-01 2.61157602e-01 -7.45244324e-01 -9.33890492e-02
7.05615699e-01 1.15331352e+00 -9.63635325e-01 -9.27737653e-01
-6.01136088e-01 1.02051616e+00 5.47765374e-01 4.05962616e-01
-7.03033864e-01 -8.01966250e-01 2.71894068e-01 -1.95128188e-01
4.44955111e-01 3.33627552e-01 -2.90974826e-02 -1.41451418e+00
-2.50510536e-02 -6.11603200e-01 6.75601900e-01 -8.60431850e-01
-1.75451481e+00 1.54374376e-01 1.46213099e-01 -1.51054919e+00
-2.92103320e-01 -4.24978524e-01 -6.88807547e-01 6.68882847e-01
-1.30359316e+00 -8.13474774e-01 -4.98512805e-01 5.36352575e-01
1.03592479e+00 -2.84502417e-01 8.11039507e-01 1.39590994e-01
-3.84919077e-01 6.56807005e-01 5.55178404e-01 1.62563846e-01
1.11865985e+00 -8.51608932e-01 2.51470469e-02 5.65618157e-01
-2.05855444e-02 5.41684389e-01 6.41338170e-01 -5.76318443e-01
-9.21753943e-01 -1.12935412e+00 3.11765760e-01 -2.30803281e-01
7.76630580e-01 -3.19940567e-01 -1.39277220e+00 5.72541237e-01
-8.92570466e-02 2.06601515e-01 7.83224225e-01 -7.42335320e-02
-3.43343824e-01 -8.88662636e-02 -1.18779588e+00 6.19822860e-01
7.54224360e-01 -5.07465005e-01 -9.35336173e-01 4.89219546e-01
5.88895082e-01 9.80082676e-02 -3.76975328e-01 -1.77782495e-02
4.03367549e-01 -7.79409230e-01 7.57171571e-01 -6.76465869e-01
4.08383399e-01 -1.07787468e-01 -4.00323510e-01 -1.53149259e+00
-1.54836461e-01 6.58245683e-02 -4.79824468e-02 1.51542044e+00
1.29997373e-01 -6.50214016e-01 7.28815079e-01 6.53919935e-01
3.00253749e-01 -4.22204822e-01 -7.53874838e-01 -7.61154950e-01
1.01425789e-01 -2.48267025e-01 3.95596623e-01 1.12436342e+00
1.52394116e-01 6.28861725e-01 -4.32320267e-01 -3.18590134e-01
8.64299774e-01 1.99682087e-01 9.71561313e-01 -1.25300384e+00
-2.04706505e-01 -6.51387125e-02 -5.30075788e-01 -1.13321006e+00
4.28685993e-01 -7.84763515e-01 5.35506427e-01 -1.35577011e+00
8.52323174e-01 -2.67394245e-01 -3.25078994e-01 2.42537260e-01
-5.65218031e-01 1.24439076e-01 7.23452941e-02 3.49044770e-01
-6.96065605e-01 9.71069455e-01 1.27102923e+00 -4.33192044e-01
1.69705246e-02 1.57454908e-01 -5.94042718e-01 8.57464194e-01
5.40052831e-01 -5.65526366e-01 -7.62500763e-01 5.73773962e-03
-3.65155160e-01 -1.25744477e-01 2.06816122e-02 -8.13522041e-01
3.40290695e-01 -3.99762481e-01 6.44439936e-01 -6.96232796e-01
3.62925023e-01 -7.17389584e-01 -5.40350914e-01 2.95705557e-01
-5.02641380e-01 -8.34879160e-01 -1.12785555e-01 1.10137105e+00
-1.68566525e-01 -8.57296288e-01 1.08202362e+00 -1.51842266e-01
-8.46894264e-01 3.94188166e-01 -2.79494852e-01 3.24059665e-01
1.29976475e+00 -3.56506914e-01 -3.91602516e-01 -4.37656641e-01
-4.39257115e-01 2.88206309e-01 4.66162890e-01 6.11550331e-01
6.94115639e-01 -1.34185708e+00 -3.84507447e-01 1.99348852e-01
7.75591373e-01 -1.56476021e-01 3.34082574e-01 4.61529911e-01
2.84694918e-02 8.72735307e-02 -1.23174578e-01 -7.43458092e-01
-1.07902277e+00 1.02803934e+00 1.00523680e-01 2.17616156e-01
-6.89086318e-01 6.04392529e-01 4.39657062e-01 -4.21481192e-01
3.04712892e-01 9.34630409e-02 -3.84629637e-01 3.76547515e-01
9.52378392e-01 2.96094745e-01 -2.38622785e-01 -3.25030506e-01
-1.94041044e-01 4.79593962e-01 -5.26095867e-01 1.22959077e-01
1.11269653e+00 -4.33138877e-01 2.50140131e-01 1.02930236e+00
1.32052076e+00 -5.97289860e-01 -1.31691909e+00 -6.10043108e-01
-1.46093771e-01 -8.36741090e-01 -2.28910195e-03 -2.87873626e-01
-6.97053611e-01 1.22682226e+00 6.02681279e-01 7.51815811e-02
7.19086707e-01 -1.27861267e-02 7.55026221e-01 6.69030309e-01
4.33094949e-01 -1.41058707e+00 2.58734643e-01 3.76334399e-01
3.48587841e-01 -1.59819186e+00 6.85724616e-03 -4.23228472e-01
-7.93776691e-01 1.14357197e+00 7.83711374e-01 2.94009168e-02
8.22490096e-01 -2.50622422e-01 6.42586946e-02 5.82180209e-02
-6.48674607e-01 -1.21424921e-01 9.77245718e-02 7.82067180e-01
2.39570886e-01 -1.31950974e-01 -1.83210298e-01 5.70390046e-01
2.82641411e-01 1.93468645e-01 5.72336435e-01 1.02495837e+00
-8.01265538e-01 -5.86006820e-01 -2.95972258e-01 1.00229168e+00
-1.49969131e-01 -4.70788144e-02 -3.39367613e-02 4.07938689e-01
-3.68415296e-01 8.00942421e-01 1.59284398e-01 -3.40183312e-03
1.52505755e-01 2.94272512e-01 3.91845942e-01 -1.04523456e+00
3.96922916e-01 9.85468924e-03 -2.96260387e-01 7.59162307e-02
-3.09416503e-01 -5.91248810e-01 -1.18298972e+00 7.52993450e-02
-8.15726995e-01 2.75034577e-01 3.97702575e-01 1.19905519e+00
1.04279786e-01 3.45706165e-01 8.53342056e-01 -8.43680859e-01
-1.15849602e+00 -1.30631196e+00 -8.62967134e-01 7.70620525e-01
5.26352637e-02 -9.90972102e-01 -7.44751334e-01 2.21823022e-01] | [10.192517280578613, 3.372305154800415] |
f2c4264d-da52-4aaa-8b92-2c7249a7da64 | a-generative-approach-to-zero-shot-and-few | 1801.09086 | null | http://arxiv.org/abs/1801.09086v1 | http://arxiv.org/pdf/1801.09086v1.pdf | A Generative Approach to Zero-Shot and Few-Shot Action Recognition | We present a generative framework for zero-shot action recognition where some
of the possible action classes do not occur in the training data. Our approach
is based on modeling each action class using a probability distribution whose
parameters are functions of the attribute vector representing that action
class. In particular, we assume that the distribution parameters for any action
class in the visual space can be expressed as a linear combination of a set of
basis vectors where the combination weights are given by the attributes of the
action class. These basis vectors can be learned solely using labeled data from
the known (i.e., previously seen) action classes, and can then be used to
predict the parameters of the probability distributions of unseen action
classes. We consider two settings: (1) Inductive setting, where we use only the
labeled examples of the seen action classes to predict the unseen action class
parameters; and (2) Transductive setting which further leverages unlabeled data
from the unseen action classes. Our framework also naturally extends to
few-shot action recognition where a few labeled examples from unseen classes
are available. Our experiments on benchmark datasets (UCF101, HMDB51 and
Olympic) show significant performance improvements as compared to various
baselines, in both standard zero-shot (disjoint seen and unseen classes) and
generalized zero-shot learning settings. | ['Anurag Mittal', 'Piyush Rai', 'Vinay Kumar Verma', 'M Shiva Krishna Reddy', 'Ashish Mishra', 'Arulkumar S'] | 2018-01-27 | null | null | null | null | ['zero-shot-action-recognition', 'few-shot-action-recognition'] | ['computer-vision', 'computer-vision'] | [ 4.35808271e-01 1.00152954e-01 -6.64106607e-01 -3.75270247e-01
-5.39891005e-01 -4.18816149e-01 8.15403402e-01 -3.62177074e-01
-2.33277231e-01 7.16659367e-01 4.86616611e-01 2.46663064e-01
2.48541180e-02 -8.25948656e-01 -7.08074093e-01 -1.08159006e+00
2.10056141e-01 7.42134988e-01 4.05640811e-01 -1.16532400e-01
-7.46736750e-02 -2.16728915e-03 -1.86828232e+00 4.24181253e-01
3.57672125e-01 9.62760329e-01 -1.22059710e-01 7.52517700e-01
-1.02282278e-01 1.21505535e+00 -6.00638986e-01 -5.24091199e-02
3.53951544e-01 -6.50674641e-01 -6.10835850e-01 6.83629811e-01
1.69201225e-01 -5.63804865e-01 -6.18288100e-01 8.55707347e-01
1.57202631e-01 6.13593996e-01 9.78043556e-01 -1.68455160e+00
-7.83758581e-01 1.57509491e-01 -4.58242744e-01 1.23498224e-01
3.16927105e-01 3.31887811e-01 1.01947057e+00 -7.35743701e-01
7.12869942e-01 1.23123372e+00 3.69039141e-02 7.98594117e-01
-1.18978202e+00 -4.69949603e-01 3.50493759e-01 6.62886143e-01
-1.17177653e+00 -4.48328137e-01 5.56399107e-01 -6.19425476e-01
9.22384441e-01 -1.52086928e-01 6.83838785e-01 1.48454082e+00
2.24929061e-02 1.03247368e+00 6.22309625e-01 -5.16202092e-01
8.72841060e-01 -7.03429431e-02 2.63489544e-01 4.03567255e-01
-6.51625469e-02 1.65031537e-01 -5.04217625e-01 -3.44999343e-01
5.99558592e-01 5.81574678e-01 -1.20216742e-01 -9.10932779e-01
-1.07097137e+00 1.01316059e+00 1.84086323e-01 -2.36505121e-02
-3.82037252e-01 1.24367476e-01 3.02133262e-01 1.49595886e-01
2.41884038e-01 -7.16645494e-02 -3.74336421e-01 -1.74012542e-01
-2.21351177e-01 3.11297588e-02 5.81654549e-01 1.05267143e+00
1.06818032e+00 2.05190629e-02 -5.31163156e-01 9.38840628e-01
1.76457778e-01 5.55384815e-01 8.67999136e-01 -9.72682297e-01
3.37925225e-01 6.78416431e-01 3.81554008e-01 -3.19966346e-01
1.83210835e-01 2.94933975e-01 -3.63939315e-01 1.73320502e-01
4.33327436e-01 -8.54954347e-02 -1.60116434e+00 1.77979839e+00
4.14296180e-01 6.12076163e-01 3.43652338e-01 6.45640194e-01
7.20558763e-01 6.72020495e-01 2.46649176e-01 -1.73436403e-01
1.04584587e+00 -8.70319605e-01 -7.00998783e-01 -3.72659087e-01
7.01254904e-01 -1.65777415e-01 9.91152287e-01 1.89062089e-01
-5.13163030e-01 -4.64730769e-01 -9.86352026e-01 1.65998757e-01
-4.81627733e-01 2.16805562e-02 5.65078914e-01 3.40129197e-01
-5.10538399e-01 4.76696700e-01 -1.03243113e+00 -4.48719889e-01
6.45805359e-01 1.97591066e-01 -3.78625810e-01 -4.28417414e-01
-1.11219728e+00 4.71619636e-01 6.32992446e-01 -3.50953013e-01
-1.54905200e+00 -3.48374605e-01 -1.03453553e+00 3.92268524e-02
9.27975655e-01 -3.57847691e-01 1.28597879e+00 -1.15504169e+00
-1.15633321e+00 6.49451971e-01 6.49560019e-02 -2.97497243e-01
7.81923011e-02 -7.43677840e-02 -4.40857917e-01 2.52915561e-01
2.20379829e-01 4.98875469e-01 9.21963215e-01 -1.04933524e+00
-8.29984605e-01 -3.72366756e-01 2.55846739e-01 3.19108605e-01
-3.38848442e-01 -2.77486980e-01 -2.99040705e-01 -2.75310963e-01
-1.10421211e-01 -1.06057549e+00 -8.03164467e-02 1.93195075e-01
-4.68085885e-01 -2.65655160e-01 8.96195412e-01 -5.04039861e-02
9.02192652e-01 -2.33359098e+00 2.79822424e-02 -1.47234946e-01
-1.26565456e-01 2.47117996e-01 -3.00628304e-01 3.97489071e-01
-2.73231357e-01 -3.28248024e-01 -9.77782980e-02 4.70786877e-02
-9.25027300e-03 8.28207850e-01 -4.94326591e-01 3.53744566e-01
6.88892603e-02 8.49005997e-01 -1.08825040e+00 -3.41192245e-01
4.98589516e-01 3.54120612e-01 -2.94636250e-01 4.08460796e-01
-3.09386224e-01 2.14215726e-01 -6.23793840e-01 6.62744284e-01
6.85149804e-02 -3.23757946e-01 2.62813091e-01 6.21781498e-02
5.86858273e-01 -9.33323279e-02 -1.39885092e+00 1.38490748e+00
-1.03237577e-01 1.17971480e-01 -6.83676958e-01 -1.02366912e+00
6.42419338e-01 4.43079531e-01 5.51591337e-01 -1.47924528e-01
9.70588252e-02 -3.79813254e-01 1.09036595e-01 -6.37855232e-01
-9.15923491e-02 -4.50330198e-01 -2.14875638e-01 6.17747188e-01
6.69062018e-01 3.90105337e-01 3.58603030e-01 3.26853782e-01
1.20008433e+00 3.33592057e-01 6.95605576e-01 3.51698309e-01
1.42358303e-01 -9.53170955e-02 8.98584485e-01 8.23905528e-01
-4.10628408e-01 6.08043075e-01 5.44572115e-01 -5.66467881e-01
-8.59534204e-01 -1.27842712e+00 -4.81957663e-03 1.38306105e+00
1.40920147e-01 -4.77493167e-01 -4.68105078e-01 -8.89005423e-01
-1.48044184e-01 9.20410275e-01 -1.10032165e+00 -5.67195415e-01
-1.77904088e-02 -6.59861028e-01 -4.71789315e-02 1.01896334e+00
3.26202631e-01 -1.22949326e+00 -6.11903965e-01 7.06619844e-02
3.24500096e-03 -1.11418641e+00 -3.36636186e-01 3.29110861e-01
-6.12608492e-01 -1.51208472e+00 -4.30076838e-01 -4.13380355e-01
6.83814466e-01 3.12419057e-01 7.80399978e-01 -2.18975946e-01
-5.62139571e-01 9.74298477e-01 -5.33804655e-01 -4.27631468e-01
-2.18405530e-01 -6.07302666e-01 1.78180426e-01 6.43710852e-01
8.14024746e-01 -4.46760267e-01 -4.79274333e-01 3.34594220e-01
-9.36472893e-01 -2.76754081e-01 4.27605331e-01 8.99244130e-01
8.12664151e-01 2.18177112e-04 3.81940514e-01 -1.01881182e+00
-1.67045202e-02 -7.97698736e-01 -2.14221682e-02 6.44260645e-01
-2.40345761e-01 -3.13819833e-02 5.67579031e-01 -9.38829541e-01
-1.12585449e+00 4.25522894e-01 5.18171370e-01 -9.47362065e-01
-5.80698192e-01 4.52117547e-02 -5.98886788e-01 4.66852605e-01
7.15426445e-01 3.41462761e-01 -6.79969043e-02 -3.90789360e-01
5.14482796e-01 7.15328693e-01 4.69675094e-01 -4.95220959e-01
5.93303263e-01 6.98624969e-01 -1.28601596e-01 -8.21881711e-01
-1.05618393e+00 -6.66035712e-01 -9.06783223e-01 -2.36985773e-01
1.11722875e+00 -9.16277051e-01 -1.97770029e-01 5.27657032e-01
-5.27309120e-01 -3.70858699e-01 -8.66607428e-01 5.57184160e-01
-8.85824025e-01 2.64654726e-01 -3.21480274e-01 -9.35083032e-01
2.29625717e-01 -1.08511496e+00 1.20935202e+00 2.57238388e-01
-9.10657495e-02 -9.92155969e-01 2.04986677e-01 2.57924736e-01
-1.82754174e-01 2.45472282e-01 1.01978457e+00 -1.07884336e+00
-4.00046915e-01 -5.91701746e-01 4.27369982e-01 5.32347739e-01
4.88438636e-01 9.51453578e-03 -9.93541956e-01 -1.88534796e-01
-1.05218120e-01 -9.12508488e-01 8.62678528e-01 3.78007114e-01
1.02613509e+00 -2.10029542e-01 -4.77500290e-01 1.52701333e-01
1.20047963e+00 4.73376483e-01 7.41329372e-01 -1.29516453e-01
6.87593043e-01 2.75704056e-01 7.39185691e-01 7.43948579e-01
1.00576766e-01 6.24004483e-01 3.60965937e-01 2.36045435e-01
1.12905726e-01 -5.88045180e-01 3.84057432e-01 -1.26906568e-02
-1.30746573e-01 -3.94026667e-01 -6.04052961e-01 5.56163311e-01
-2.12477827e+00 -1.29214358e+00 3.74076247e-01 2.55532622e+00
6.20310605e-01 -5.59573120e-04 4.02738661e-01 7.60068595e-02
9.04394150e-01 2.73613334e-01 -1.06206584e+00 2.71351775e-03
1.13056026e-01 2.02938750e-01 1.45022362e-01 2.14111909e-01
-1.46433902e+00 8.00930440e-01 6.26763916e+00 5.71937501e-01
-5.76941192e-01 5.68766780e-02 4.32812512e-01 -4.22124833e-01
1.39875948e-01 1.56200871e-01 -9.02335286e-01 4.67813134e-01
8.40985835e-01 -3.48750591e-01 2.89231420e-01 1.11920524e+00
-2.67490745e-01 -1.53558388e-01 -1.41988647e+00 8.64502251e-01
3.23321134e-01 -8.87091279e-01 2.02773735e-01 2.43124545e-01
6.99547708e-01 -1.06927939e-01 -1.02514341e-01 8.79173636e-01
8.01848233e-01 -7.42002010e-01 2.15493768e-01 5.40746093e-01
6.86526299e-01 -5.85355878e-01 3.07152331e-01 6.22953117e-01
-8.67366254e-01 -4.70453054e-01 -6.89068913e-01 -8.85570943e-02
-4.90223728e-02 -1.56368334e-02 -6.78871393e-01 8.16332921e-02
5.37461042e-01 1.08056641e+00 -3.57081950e-01 8.90932679e-01
-5.08754432e-01 6.61304951e-01 -1.55955041e-02 1.85396746e-01
2.08087772e-01 -1.02630198e-01 2.96675831e-01 4.57618624e-01
9.47398841e-02 5.55834234e-01 5.89369416e-01 4.17552322e-01
6.45721629e-02 6.91317429e-04 -7.03964293e-01 -1.76486433e-01
2.97067940e-01 1.12937152e+00 -6.02715611e-01 -7.00937212e-01
-8.04823160e-01 8.97604406e-01 2.77548313e-01 6.27327502e-01
-7.23169327e-01 -7.23745227e-02 8.88977051e-01 4.90281219e-03
6.33986056e-01 3.63114595e-01 5.24552345e-01 -1.38153768e+00
-3.03786993e-01 -5.35903156e-01 9.20730829e-01 -8.00014198e-01
-1.53105664e+00 2.04367936e-01 3.53114128e-01 -1.68591487e+00
-5.65607965e-01 -7.00400174e-01 -7.64669955e-01 5.73083520e-01
-9.91889894e-01 -1.01671112e+00 -1.02923587e-01 8.15620244e-01
6.50378227e-01 -3.81769538e-01 1.11793947e+00 -2.42360428e-01
-5.04234731e-01 2.60499239e-01 2.88799495e-01 3.25584143e-01
5.37582576e-01 -1.18678045e+00 2.73609105e-02 5.29831648e-01
4.25456047e-01 3.39469492e-01 4.08058375e-01 -6.97737336e-01
-1.17222381e+00 -1.25341964e+00 4.82209176e-01 -5.70863664e-01
6.75956011e-01 -4.62872237e-01 -9.48877692e-01 1.20484030e+00
-1.75744236e-01 5.90826571e-01 1.22974181e+00 1.19207397e-01
-4.01840478e-01 5.81017211e-02 -9.76504207e-01 4.42519128e-01
9.65475798e-01 -3.90038073e-01 -8.91503036e-01 6.69405699e-01
3.25952858e-01 -8.83769840e-02 -5.07424235e-01 1.28277838e-01
4.51412201e-01 -7.28298545e-01 1.03109312e+00 -1.55777168e+00
4.09343719e-01 -2.28011161e-01 -5.01847804e-01 -1.38941586e+00
-4.15156305e-01 -5.74142672e-03 -4.37129706e-01 9.17123616e-01
1.34004593e-01 -3.45597565e-01 7.51626372e-01 1.07835197e+00
6.87164813e-02 -7.29569316e-01 -9.97567058e-01 -1.02875948e+00
-1.89769030e-01 -2.74890184e-01 5.29337883e-01 7.55429506e-01
3.01668439e-02 5.49227297e-01 -8.14525306e-01 -1.79362267e-01
5.85114002e-01 1.95339426e-01 7.99420774e-01 -1.13656306e+00
-7.38709629e-01 3.44222903e-01 -9.73927021e-01 -8.40856552e-01
3.29903364e-01 -6.87761962e-01 2.24692836e-01 -1.42619419e+00
7.31480122e-01 1.13026150e-01 -5.45581281e-01 1.05525661e+00
-2.51837730e-01 1.14976704e-01 1.01436943e-01 3.10194254e-01
-8.72315466e-01 8.13615799e-01 1.16217780e+00 -1.47533923e-01
-1.19446583e-01 1.34740099e-01 -5.37077904e-01 8.90068769e-01
4.37962145e-01 -4.66736436e-01 -1.01016259e+00 -8.52441192e-02
-4.67381507e-01 1.40693516e-01 4.17131245e-01 -9.29569781e-01
-3.09939700e-04 -7.54378319e-01 5.38293898e-01 -3.06467295e-01
7.27095425e-01 -7.45798290e-01 -4.75175679e-02 7.84311667e-02
-6.50451303e-01 -7.77151883e-01 -4.15453106e-01 1.24375713e+00
6.44824728e-02 -2.98955977e-01 8.21044028e-01 -2.22074956e-01
-1.12897635e+00 5.71964622e-01 -3.16142440e-01 3.46850485e-01
1.54101765e+00 -4.85707313e-01 -3.99404645e-01 -5.49279153e-01
-1.22929692e+00 1.53124973e-01 4.67069894e-01 6.44226789e-01
6.71643496e-01 -1.63308036e+00 -3.01884443e-01 2.32658461e-01
6.33799076e-01 -2.95557201e-01 3.38996112e-01 4.37525243e-01
4.20195788e-01 1.17285766e-01 -3.71218681e-01 -5.21812856e-01
-1.06646335e+00 1.06913614e+00 2.99266964e-01 9.60059166e-02
-7.49386430e-01 6.27611220e-01 7.40362704e-01 -1.27974257e-01
3.23058307e-01 1.23832747e-01 -2.67051190e-01 -3.24373692e-02
8.41366947e-01 3.75467211e-01 -3.06609362e-01 -9.87090468e-01
-3.78230572e-01 1.23931229e-01 -2.17976138e-01 1.07159493e-02
1.41431069e+00 2.05253601e-01 4.87397373e-01 1.02903116e+00
1.22380197e+00 -7.59218037e-01 -1.55422950e+00 -5.31660557e-01
-2.86855191e-01 -7.50845790e-01 -1.50286540e-01 -8.47724855e-01
-1.01709378e+00 8.74090970e-01 6.64202809e-01 -2.25008890e-01
9.47856903e-01 3.60769331e-01 4.57850069e-01 3.53206486e-01
4.65802759e-01 -1.33462489e+00 4.35555100e-01 1.65857866e-01
5.24535775e-01 -1.32574642e+00 -2.08331421e-01 -2.83890396e-01
-1.17951894e+00 1.01305997e+00 7.59442985e-01 -2.87116796e-01
6.91897333e-01 6.07259460e-02 -1.02877682e-02 -3.03848624e-01
-1.07914376e+00 -4.92543668e-01 2.95847982e-01 7.94625700e-01
9.97164175e-02 1.84109762e-01 1.83008075e-01 7.49351263e-01
4.38551128e-01 1.64610609e-01 6.01935685e-01 1.23001266e+00
-7.08841503e-01 -1.00696325e+00 -1.47645384e-01 7.32067645e-01
-2.50688667e-04 1.95363507e-01 -5.08525848e-01 6.65754318e-01
1.92879006e-01 9.36478913e-01 1.88368365e-01 -3.71729076e-01
3.34499359e-01 5.14346719e-01 5.25969982e-01 -1.25219274e+00
2.67145038e-01 4.83561829e-02 -1.36056125e-01 -6.81906402e-01
-3.75324935e-01 -7.56821811e-01 -1.27112579e+00 4.29214239e-01
-1.84784368e-01 -3.78533080e-02 4.25753780e-02 1.21035981e+00
1.29692122e-01 1.56484723e-01 7.44904459e-01 -6.46713138e-01
-8.97845685e-01 -9.94793475e-01 -1.02799499e+00 8.00620377e-01
1.03468955e-01 -1.25710523e+00 -6.39512181e-01 4.32289958e-01] | [8.59445571899414, 0.992862343788147] |
3973abee-d499-4267-8a49-b0d0cefc4523 | y-2-net-fcrn-for-acoustic-echo-and-noise | 2103.17189 | null | https://arxiv.org/abs/2103.17189v2 | https://arxiv.org/pdf/2103.17189v2.pdf | Y$^2$-Net FCRN for Acoustic Echo and Noise Suppression | In recent years, deep neural networks (DNNs) were studied as an alternative to traditional acoustic echo cancellation (AEC) algorithms. The proposed models achieved remarkable performance for the separate tasks of AEC and residual echo suppression (RES). A promising network topology is a fully convolutional recurrent network (FCRN) structure, which has already proven its performance on both noise suppression and AEC tasks, individually. However, the combination of AEC, postfiltering, and noise suppression to a single network typically leads to a noticeable decline in the quality of the near-end speech component due to the lack of a separate loss for echo estimation. In this paper, we propose a two-stage model (Y$^2$-Net) which consists of two FCRNs, each with two inputs and one output (Y-Net). The first stage (AEC) yields an echo estimate, which - as a novelty for a DNN AEC model - is further used by the second stage to perform RES and noise suppression. While the subjective listening test of the Interspeech 2021 AEC Challenge mostly yielded results close to the baseline, the proposed method scored an average improvement of 0.46 points over the baseline on the blind testset in double-talk on the instrumental metric DECMOS, provided by the challenge organizers. | ['Tim Fingscheidt', 'Maximilian Strake', 'Jan Franzen', 'Ernst Seidel'] | 2021-03-31 | null | null | null | null | ['acoustic-echo-cancellation', 'acoustic-echo-cancellation'] | ['medical', 'speech'] | [ 1.82878584e-01 -1.09557472e-01 7.81643033e-01 -1.17046669e-01
-8.66399348e-01 -1.97493941e-01 5.50244749e-01 -3.35355252e-02
-7.06763327e-01 3.69854122e-01 5.41069567e-01 -3.02305400e-01
-3.11155561e-02 -2.47198179e-01 -5.19765258e-01 -9.15359616e-01
-5.14759608e-02 -3.20584744e-01 2.15341315e-01 -3.78953815e-01
-1.24473631e-01 3.31059456e-01 -1.65528524e+00 2.31277823e-01
7.90181577e-01 1.14674890e+00 2.18399197e-01 8.33958387e-01
4.89862310e-03 5.69382250e-01 -7.89242804e-01 -3.51769775e-01
1.70370638e-01 -5.48037052e-01 -1.28269941e-01 -5.71985662e-01
3.37382138e-01 -1.07044555e-01 -4.76183504e-01 1.15335155e+00
1.16738188e+00 3.23423773e-01 2.77186543e-01 -6.88486934e-01
-5.08624874e-02 8.29987228e-01 -1.70746550e-01 4.76064682e-02
7.60173425e-02 7.15582222e-02 7.66879320e-01 -1.18035686e+00
7.52252787e-02 9.68552351e-01 9.43877041e-01 6.42890871e-01
-1.02809691e+00 -9.05112386e-01 -9.55322534e-02 1.48893327e-01
-1.20524454e+00 -9.40336049e-01 8.22679043e-01 -3.25099342e-02
1.06379771e+00 3.48069370e-01 5.52244306e-01 1.16180897e+00
-1.01309359e-01 5.98943591e-01 1.06084716e+00 -5.39398074e-01
4.19069916e-01 7.95161426e-02 1.33628249e-01 1.16874069e-01
-1.58347964e-01 6.32265747e-01 -6.73183680e-01 6.70588315e-02
3.45402002e-01 -3.15144986e-01 -7.07068026e-01 1.10916436e-01
-7.57537484e-01 3.85419369e-01 5.41322351e-01 5.99836707e-01
-5.87534606e-01 2.25684986e-01 3.59474629e-01 6.00978971e-01
5.17142534e-01 2.66401976e-01 -1.41126737e-01 -1.83144823e-01
-1.21408808e+00 2.63614953e-01 9.27640676e-01 4.28392529e-01
1.66840836e-01 7.33938873e-01 -1.95467144e-01 1.18510807e+00
2.70687193e-01 6.80125535e-01 6.48920953e-01 -6.37388587e-01
4.05665934e-01 -1.50213167e-01 1.43814702e-02 -9.85060334e-01
-5.72837472e-01 -1.24137580e+00 -1.17013788e+00 4.86993194e-01
9.34247375e-02 -4.01362628e-01 -1.00348723e+00 1.93151069e+00
3.21877841e-03 1.66655943e-01 1.77728653e-01 1.13153100e+00
8.30239952e-01 7.15972781e-01 6.62989616e-02 -1.47420540e-01
9.08349633e-01 -1.03870440e+00 -1.00127089e+00 -2.78022468e-01
3.13504934e-01 -8.64634156e-01 5.89789271e-01 5.40985525e-01
-1.33491099e+00 -7.62597144e-01 -1.21741581e+00 2.94158369e-01
-3.18584144e-01 6.38498738e-02 -3.77568603e-02 9.09508586e-01
-1.57454407e+00 5.63944340e-01 -5.35214722e-01 4.49783430e-02
-2.36167818e-01 3.45329762e-01 -2.51565844e-01 1.30095661e-01
-1.37284827e+00 7.25349247e-01 -1.41423777e-01 6.71691656e-01
-1.01416373e+00 -6.85344934e-01 -6.55369818e-01 4.28317070e-01
2.36924767e-01 -3.39912742e-01 1.18236160e+00 -8.53653491e-01
-1.78939867e+00 3.84795934e-01 -6.44648522e-02 -9.08176124e-01
5.61742902e-01 -6.85052872e-01 -1.05468369e+00 -9.25584659e-02
-3.73641640e-01 4.19734210e-01 1.15276039e+00 -1.14396417e+00
-5.86165369e-01 -1.06859125e-01 -3.80364954e-01 2.08267346e-01
-4.05609161e-02 1.30767405e-01 -4.09326315e-01 -9.58154202e-01
3.45455319e-01 -6.44643724e-01 -1.75257385e-01 -5.29014766e-01
-3.94196808e-01 1.12181365e-01 6.01954639e-01 -1.06816733e+00
1.36613464e+00 -2.56915617e+00 1.44800499e-01 2.80049682e-01
1.35320798e-01 8.79249811e-01 -4.07091379e-01 3.10406506e-01
-6.17399573e-01 -7.53054023e-02 -2.35316321e-01 -6.57665491e-01
-5.48762120e-02 -4.99589771e-01 -2.39282563e-01 3.63832504e-01
8.89947563e-02 4.75544721e-01 -4.91567343e-01 2.77650595e-01
2.69323736e-01 8.00208867e-01 -4.19378549e-01 2.75801867e-01
1.99420720e-01 5.07012382e-02 3.06703776e-01 2.58533210e-01
9.70354617e-01 5.42156637e-01 -4.89348955e-02 -1.05974965e-01
-4.42043960e-01 3.59151125e-01 -1.41373193e+00 1.38051653e+00
-5.95816553e-01 7.77941287e-01 7.96682239e-01 -7.54040122e-01
1.03847039e+00 8.50210190e-01 9.82622281e-02 -1.07023752e+00
6.11933060e-02 7.41119921e-01 3.49907249e-01 -8.65088627e-02
3.32217515e-01 -2.21183240e-01 2.95667768e-01 2.75114179e-01
2.91859061e-01 3.43798823e-03 -1.60891548e-01 2.45278999e-02
1.20480311e+00 -2.82789618e-01 -1.52085051e-01 -2.54730731e-01
8.63968253e-01 -7.37300396e-01 5.67119896e-01 8.94781590e-01
-2.56984562e-01 8.34031463e-01 1.40844062e-01 -1.97115280e-02
-7.29838073e-01 -9.58267510e-01 2.19344676e-01 7.90742517e-01
-2.15986133e-01 -3.59751493e-01 -8.86415899e-01 -2.00363100e-01
-3.35905880e-01 8.70553553e-01 -1.95677608e-01 -2.91643411e-01
-7.89853454e-01 -4.60837156e-01 8.05038929e-01 3.49890649e-01
5.88304818e-01 -1.23313475e+00 -3.81026506e-01 5.00435233e-01
-2.00558215e-01 -9.85538960e-01 -5.32488823e-01 6.35836303e-01
-5.20813644e-01 -4.70084548e-01 -1.18520248e+00 -7.47947574e-01
2.17457876e-01 1.87685654e-01 7.26710618e-01 -7.05386177e-02
1.03450015e-01 1.58764273e-01 -3.52514744e-01 -5.35535455e-01
-4.59156007e-01 -7.72909671e-02 3.63100678e-01 2.13835463e-01
2.53509302e-02 -8.05943429e-01 -7.61070371e-01 2.84443796e-01
-8.38767707e-01 -2.77950138e-01 7.52331972e-01 7.75450349e-01
1.92792296e-01 3.61866765e-02 7.15756774e-01 -4.01231706e-01
8.88812065e-01 -3.24271023e-02 -6.35626316e-01 -8.65679830e-02
-5.29471338e-01 -9.22759250e-02 6.76305115e-01 -3.30559999e-01
-1.20571768e+00 -1.51673153e-01 -8.41939449e-01 -4.25113857e-01
-5.33081070e-02 4.04822081e-01 -1.34901792e-01 -1.73162557e-02
4.31859165e-01 3.72077912e-01 -4.15700898e-02 -8.77855539e-01
-3.25543806e-02 7.21911669e-01 7.96242893e-01 1.04480274e-01
9.35746789e-01 -1.96969762e-01 -2.87264496e-01 -9.26930070e-01
-4.37099546e-01 -4.94331270e-01 -9.20396894e-02 -2.05221757e-01
6.32711291e-01 -9.87348914e-01 -6.42770290e-01 9.53710794e-01
-1.11827898e+00 -2.78220117e-01 -2.84956634e-01 8.29440832e-01
-1.08279316e-02 1.03123605e-01 -4.77839500e-01 -1.20744181e+00
-6.94223285e-01 -1.07250595e+00 4.96560752e-01 3.15163016e-01
1.58059765e-02 -5.76748967e-01 1.40764862e-01 -1.14158474e-01
1.18438625e+00 -3.33185285e-01 6.54164433e-01 -7.19783723e-01
-1.36746705e-01 -3.27764779e-01 -1.51107516e-02 9.43905771e-01
-2.00001970e-01 -5.01861095e-01 -1.46653950e+00 -3.36975545e-01
5.38336277e-01 -2.60736197e-02 1.14905441e+00 5.47104716e-01
8.51610839e-01 5.90426736e-02 1.63303703e-01 6.59855187e-01
1.26112187e+00 4.81553704e-01 8.79161298e-01 8.08597449e-03
1.11197695e-01 4.70754474e-01 -5.37742041e-02 -2.14153994e-03
-2.87392795e-01 6.82986736e-01 3.30286682e-01 -4.33584303e-01
-7.11512983e-01 -8.96844715e-02 6.70698524e-01 1.30858994e+00
-3.59899066e-02 -4.24726367e-01 -6.08670294e-01 4.80309397e-01
-1.31649184e+00 -8.07594419e-01 -9.47583914e-02 2.32568312e+00
5.25044084e-01 3.08164924e-01 -1.71664536e-01 5.21701276e-01
6.76809907e-01 3.17363113e-01 -3.35533530e-01 -6.23625934e-01
-3.87015522e-01 6.25615835e-01 3.76969278e-01 6.61918283e-01
-8.49644184e-01 5.82304537e-01 6.21263409e+00 8.76351237e-01
-1.26793432e+00 8.96985307e-02 3.82895619e-01 -1.46419093e-01
-1.14522971e-01 -3.35105985e-01 -5.94799221e-01 2.53671914e-01
1.34315860e+00 3.43733996e-01 5.31648755e-01 3.56028497e-01
2.72316813e-01 -4.89091128e-02 -6.35762334e-01 1.00009799e+00
3.25274318e-02 -8.67788553e-01 -1.72572106e-01 -1.30636722e-01
2.76458234e-01 2.38694534e-01 2.40853846e-01 6.08686328e-01
-1.22540690e-01 -9.88790393e-01 8.27339411e-01 5.89058936e-01
8.14253330e-01 -7.87986636e-01 1.00499940e+00 2.07470685e-01
-1.09605992e+00 -2.58948803e-01 -1.59648091e-01 1.89855486e-01
2.15598777e-01 7.47570038e-01 -4.02946025e-01 6.09102845e-01
8.01148176e-01 1.13775805e-02 -2.42776066e-01 1.21567893e+00
-2.26316974e-01 9.77365613e-01 -3.31853092e-01 7.28056878e-02
2.46820852e-01 3.52814533e-02 1.03629506e+00 1.44436479e+00
3.26920182e-01 7.43658468e-02 -5.77676654e-01 6.77644074e-01
-2.76716173e-01 -6.03687316e-02 -1.48126245e-01 5.26202582e-02
2.89353818e-01 1.15294290e+00 -2.10139781e-01 5.49938194e-02
-3.03986907e-01 9.28733826e-01 -4.05673645e-02 6.89660490e-01
-5.46339214e-01 -1.00588346e+00 4.78005230e-01 -1.13323502e-01
3.72995406e-01 -7.08179697e-02 -5.42234741e-02 -6.74498796e-01
1.70840859e-01 -1.09188581e+00 -2.16062918e-01 -6.56332970e-01
-8.63279045e-01 1.05565333e+00 -5.19690394e-01 -1.09857440e+00
-3.54679972e-01 -4.99787718e-01 -7.80764759e-01 1.44745266e+00
-1.53351974e+00 -3.92956883e-01 -7.41884410e-02 5.17906964e-01
3.93498600e-01 -4.05843407e-01 9.26842034e-01 7.30027914e-01
-5.23803711e-01 8.77056956e-01 1.71543375e-01 8.91359672e-02
6.15586519e-01 -1.15793335e+00 4.84550685e-01 1.25027025e+00
5.92629984e-02 4.68218833e-01 9.36407089e-01 -4.97960627e-01
-9.64962244e-01 -8.33848119e-01 1.07880580e+00 3.92024994e-01
2.98529327e-01 -5.46434402e-01 -8.85064960e-01 1.81134837e-03
3.24722350e-01 -7.38881454e-02 2.84700632e-01 9.63621959e-03
-2.28326231e-01 -4.39006686e-01 -8.38925302e-01 5.33715189e-01
8.14316154e-01 -6.85445726e-01 -3.47599000e-01 -4.43996668e-01
9.15347934e-01 -3.14342231e-01 -2.89702445e-01 4.50879753e-01
6.16749585e-01 -1.20908034e+00 8.82644355e-01 -9.82075781e-02
9.44838002e-02 -2.15967804e-01 -3.48523140e-01 -1.59209704e+00
-1.21575728e-01 -9.93436337e-01 -1.80477202e-02 1.27640283e+00
5.99441171e-01 -5.14015019e-01 3.97719145e-01 2.23399341e-01
-5.03422379e-01 -3.18608284e-01 -8.94875169e-01 -7.65673697e-01
-2.35846043e-01 -7.81373203e-01 2.57619202e-01 1.78041577e-01
-4.07172561e-01 2.77128100e-01 -6.11380517e-01 2.05217481e-01
5.28849363e-01 -5.20273924e-01 3.04320723e-01 -1.01073313e+00
-4.38017458e-01 -6.31238282e-01 -1.43750340e-01 -1.00168765e+00
-2.29502589e-01 -6.14693880e-01 4.47902113e-01 -1.23663390e+00
-4.57326412e-01 -6.50376752e-02 -6.76236868e-01 -3.32294665e-02
-4.51855250e-02 8.53191540e-02 5.02883971e-01 -2.03228086e-01
-1.16153948e-01 6.94487631e-01 7.23312557e-01 -9.59990695e-02
-4.64461505e-01 4.97579247e-01 -5.47929525e-01 5.75480878e-01
5.87644339e-01 -5.05971551e-01 -1.40917778e-01 -5.24654090e-01
3.55399922e-02 4.27904904e-01 3.19325954e-01 -1.52417564e+00
6.34687126e-01 8.63839149e-01 2.46949375e-01 -7.40015924e-01
6.09005570e-01 -8.30969274e-01 1.94573209e-01 5.84544897e-01
-5.27148843e-01 -1.59204215e-01 3.79838049e-01 5.29204071e-01
-7.01888263e-01 -1.92142084e-01 8.01867545e-01 1.60350248e-01
-2.73389369e-01 -1.40825287e-01 -5.97421885e-01 -2.04564676e-01
1.11230887e-01 1.37043938e-01 6.51357323e-02 -7.47059703e-01
-9.27379310e-01 -7.37444162e-02 -4.21222538e-01 3.30201477e-01
7.10914135e-01 -1.05076730e+00 -1.07956219e+00 4.28145409e-01
-4.37128603e-01 -3.71126682e-01 5.79607546e-01 8.90420437e-01
-9.63981003e-02 4.54460651e-01 1.49375061e-02 -3.84719461e-01
-1.09619892e+00 9.32026803e-02 7.75590241e-01 -2.95756876e-01
-7.66168773e-01 1.02519834e+00 1.61706582e-01 -3.58298928e-01
9.48490858e-01 -1.48066446e-01 -2.82582968e-01 -5.43248802e-02
7.60729492e-01 4.93978351e-01 6.29002333e-01 -4.26756740e-01
-2.78112769e-01 3.75831693e-01 -7.06844265e-03 -6.72153294e-01
1.42826140e+00 -2.28710666e-01 5.18462919e-02 3.47388744e-01
1.23859763e+00 3.71173561e-01 -7.97748685e-01 -3.59050393e-01
1.70651972e-02 7.42611215e-02 5.44567168e-01 -1.27601337e+00
-1.32948458e+00 1.09639704e+00 1.18707347e+00 1.27345189e-01
1.58898163e+00 -5.63805938e-01 8.21617424e-01 2.19235882e-01
-1.95888534e-01 -1.01107097e+00 -1.39123842e-01 7.16205239e-01
1.14349163e+00 -8.06425273e-01 -5.60306787e-01 1.04812220e-01
-3.49665731e-01 1.10338938e+00 2.94829369e-01 -1.11101225e-01
9.72416818e-01 6.27567947e-01 2.92461723e-01 8.35172739e-03
-6.40685856e-01 -1.76881120e-01 3.62458974e-01 4.25354034e-01
5.10410190e-01 -8.61298133e-07 -2.82906681e-01 8.36313307e-01
-2.90896595e-01 -2.11824730e-01 1.17058784e-01 4.30039585e-01
-4.21462834e-01 -8.66786480e-01 -3.47316563e-01 1.52728856e-01
-5.94185233e-01 -3.79019916e-01 -3.23082238e-01 6.28428400e-01
-6.98652342e-02 1.28804040e+00 -9.03174467e-03 -6.72343254e-01
7.38556564e-01 2.20224604e-01 -2.00885475e-01 -1.03120856e-01
-1.18348563e+00 5.96716762e-01 1.02522902e-01 -6.11180067e-01
-1.91949293e-01 -4.86292869e-01 -9.08916831e-01 -3.75214443e-02
-3.81972700e-01 3.53145599e-01 9.53208327e-01 7.02562451e-01
1.49443761e-01 9.97031808e-01 7.11102784e-01 -7.14699805e-01
-5.79579949e-01 -1.34174085e+00 -8.16562712e-01 9.36035514e-02
8.40395093e-01 -1.43115103e-01 -7.35455155e-01 -4.24171597e-01] | [15.036421775817871, 5.93364143371582] |
84287a5a-12c1-4d9e-ae6a-086b3536d27f | extractive-topical-summarization-with-aspects | null | null | https://openreview.net/forum?id=dTwZya8uNnS | https://openreview.net/pdf?id=dTwZya8uNnS | Extractive Topical Summarization With Aspects | Extractive summarization is a task of highlighting the most important parts of the text. We introduce a new approach to extractive summarization task using hidden topical structure and information about aspects of the text. Experimental results on CNN/DailyMail demonstrate that our approach generates more accurate summarizations than baseline methods, achieving state-of-the-art results in terms of ROUGE metric. Additionally, we show that aspect information is extremely important in extractive summarization scenario. | ['Anonymous'] | 2021-09-17 | null | null | null | acl-arr-september-2021-9 | ['extractive-summarization'] | ['natural-language-processing'] | [ 2.53800094e-01 5.81745088e-01 -2.90366411e-01 -8.84426683e-02
-1.17508006e+00 -5.71794927e-01 7.62031138e-01 6.54488802e-01
-4.44551080e-01 1.08722496e+00 1.50110793e+00 1.12612180e-01
1.43262178e-01 -5.23920059e-01 -5.62169671e-01 -2.26515040e-01
-7.17808958e-03 2.24151745e-01 8.79960041e-03 -5.43203175e-01
9.69705522e-01 1.40685961e-01 -1.01012206e+00 6.90812469e-01
1.15594149e+00 2.93956876e-01 -2.42954455e-02 1.14409280e+00
-4.12369221e-01 6.79508388e-01 -1.50374126e+00 -4.08608943e-01
-4.93584536e-02 -6.25297010e-01 -1.09044516e+00 1.29552573e-01
1.08499372e+00 -5.14216781e-01 -3.51553679e-01 1.00270784e+00
7.51160562e-01 2.97172785e-01 7.61096656e-01 -5.83551347e-01
-6.24391019e-01 1.18604374e+00 -8.48943591e-01 6.70277059e-01
4.84491289e-01 -1.36478931e-01 1.49999869e+00 -7.02594697e-01
8.40565622e-01 1.20474827e+00 2.45059520e-01 5.21198094e-01
-6.77564144e-01 -2.13008001e-01 3.34319174e-01 -4.37548384e-02
-6.51592076e-01 -5.35794854e-01 8.22145164e-01 4.17042300e-02
1.42728972e+00 7.29767382e-01 6.24186635e-01 8.27012002e-01
6.98036611e-01 1.36710107e+00 3.18697184e-01 -1.08675532e-01
-2.26884224e-02 -2.68934488e-01 7.14171827e-01 6.89712763e-01
6.78113699e-01 -9.17462289e-01 -9.74945247e-01 -2.09957197e-01
3.53569575e-02 -1.73863277e-01 -3.64587069e-01 4.53523368e-01
-1.43567455e+00 6.84131324e-01 2.22400993e-01 2.94402421e-01
-8.50944221e-01 2.10909918e-01 9.67405140e-01 9.26774815e-02
1.02459729e+00 9.45726275e-01 -1.35236263e-01 -2.92231053e-01
-1.47404695e+00 3.90713573e-01 1.01887202e+00 1.24255037e+00
4.68016535e-01 3.49347234e-01 -9.95661378e-01 7.65334904e-01
-3.45505446e-01 4.24460441e-01 4.50777888e-01 -9.15490806e-01
1.02059686e+00 7.43056953e-01 -5.19375736e-03 -8.64944935e-01
-6.12117983e-02 -7.08173633e-01 -1.16810322e+00 -5.73464096e-01
-7.02955246e-01 -5.12724280e-01 -9.52865124e-01 9.73707497e-01
-2.27895096e-01 -2.87816852e-01 3.23917300e-01 3.66833657e-01
1.62499917e+00 1.07588375e+00 -1.35208473e-01 -6.02299511e-01
1.37961125e+00 -1.38292480e+00 -1.15809810e+00 -2.39856675e-01
2.18236804e-01 -8.28494906e-01 7.02521086e-01 2.51668811e-01
-1.50247347e+00 -3.32561880e-01 -1.24605274e+00 -3.35277945e-01
-1.51598305e-01 5.15360475e-01 4.49926198e-01 4.92299944e-02
-1.20769405e+00 7.07794011e-01 -4.73273903e-01 -4.97087985e-01
6.49532020e-01 1.68604314e-01 -3.78320634e-01 4.59215492e-01
-7.96604931e-01 5.96516311e-01 8.29801679e-01 -1.82510957e-01
-6.78725123e-01 -6.99886441e-01 -9.14810359e-01 5.13525128e-01
3.45335156e-01 -1.22197068e+00 1.54806983e+00 -4.59285140e-01
-1.28131855e+00 5.61274230e-01 -5.90482831e-01 -9.17553246e-01
2.83288091e-01 -9.42518711e-01 -1.67257115e-02 8.59344006e-01
2.83725709e-01 6.69023037e-01 8.64165545e-01 -1.16083550e+00
-8.71974647e-01 -1.23759814e-01 4.08369750e-02 4.76163447e-01
-4.64080215e-01 1.06576644e-02 -3.02941859e-01 -9.07060087e-01
-2.84414738e-01 -3.60999107e-01 -2.90219277e-01 -9.36033845e-01
-1.22240531e+00 -4.45042670e-01 9.82420564e-01 -1.21826947e+00
1.68180704e+00 -1.71865392e+00 2.55233794e-01 -3.58581662e-01
7.77590692e-01 4.33821797e-01 -5.50585501e-02 1.03825331e+00
2.07521975e-01 5.87626755e-01 -2.95546293e-01 -4.92201388e-01
2.27273237e-02 -2.91932195e-01 -8.11730087e-01 -6.54217154e-02
3.00500840e-01 1.28107154e+00 -9.20039952e-01 -8.48961592e-01
-1.42658621e-01 1.51999801e-01 -3.13579530e-01 1.10787690e-01
-2.65972167e-01 1.30040586e-01 -7.39336848e-01 3.76869202e-01
3.93349826e-01 1.00814238e-01 -3.50230485e-01 -1.27253324e-01
-1.81274250e-01 5.86847544e-01 -3.21277529e-01 1.65384042e+00
-2.41757095e-01 1.12652552e+00 -3.22344929e-01 -7.14826405e-01
7.11338580e-01 3.71840507e-01 2.13529587e-01 -2.40887806e-01
1.49521410e-01 -1.09477915e-01 -3.87129277e-01 -3.69142950e-01
1.67661214e+00 2.88011014e-01 -2.89496094e-01 7.21356213e-01
2.33332828e-01 -3.12185466e-01 7.99778163e-01 1.04992414e+00
1.03410065e+00 -3.24931920e-01 8.88276219e-01 -4.11529213e-01
3.41027021e-01 1.27387345e-01 2.84377962e-01 1.02683640e+00
1.23063497e-01 7.39422023e-01 1.00070083e+00 -1.28068015e-01
-1.07314634e+00 -7.25706160e-01 5.65319359e-01 9.71901357e-01
-1.84624195e-01 -1.03958189e+00 -1.03564346e+00 -1.07903421e+00
-2.98650652e-01 1.23263931e+00 -7.34726250e-01 -2.13666260e-02
-1.08314896e+00 -4.47263390e-01 5.01864612e-01 5.66379011e-01
7.73369372e-01 -1.26952803e+00 -6.57140136e-01 7.07613528e-02
-4.67167825e-01 -1.02507842e+00 -9.62423146e-01 -3.34726602e-01
-1.35325015e+00 -8.15517545e-01 -9.40336704e-01 -7.25329995e-01
8.10278475e-01 6.14737570e-01 1.24443662e+00 9.97007638e-02
3.97699811e-02 4.26801056e-01 -4.82099384e-01 -8.18219721e-01
-4.90528375e-01 7.18443930e-01 -5.58859706e-01 -3.42273325e-01
2.34426092e-02 -4.73611921e-01 -8.18567038e-01 -7.02544689e-01
-9.86395240e-01 1.42563179e-01 8.93796325e-01 4.24347997e-01
2.92091906e-01 -2.97564924e-01 8.43467236e-01 -1.20920575e+00
1.59470570e+00 -2.25698799e-01 3.35124642e-01 2.77127326e-01
-3.30875844e-01 2.94429690e-01 5.61612070e-01 -3.84467132e-02
-1.18385530e+00 -5.24333179e-01 1.41284671e-02 3.45085531e-01
5.00364155e-02 5.45353889e-01 2.57050037e-01 6.76385581e-01
4.67059016e-01 6.89430475e-01 -3.13798070e-01 -4.48551774e-01
4.24721897e-01 6.87989891e-01 6.34313941e-01 -2.38576874e-01
5.86005867e-01 6.28626525e-01 -2.07347080e-01 -1.40962684e+00
-1.09377861e+00 -8.55176449e-01 -7.15981603e-01 -2.63876289e-01
6.20611846e-01 -7.10349143e-01 -3.38174611e-01 1.43966973e-01
-1.65418732e+00 3.08145583e-01 -6.80645347e-01 4.67729867e-02
-4.13620472e-01 9.11777198e-01 -5.32841265e-01 -4.35806304e-01
-1.71296966e+00 -6.56434536e-01 1.57336652e+00 5.35761714e-01
-6.58827126e-01 -8.99014413e-01 1.25819981e-01 3.01880598e-01
3.23718429e-01 4.84823763e-01 8.97134542e-01 -1.14349091e+00
-3.97123545e-01 -4.21386331e-01 -1.95029989e-01 3.81657481e-01
1.19056903e-01 1.49389848e-01 -5.53349078e-01 -2.03159928e-01
-1.38945937e-01 6.89439774e-02 1.69422388e+00 5.88754177e-01
8.32159162e-01 -1.02010310e+00 -2.27832511e-01 4.38727029e-02
1.03930521e+00 -2.87296236e-01 7.12248385e-01 6.41660616e-02
6.89308107e-01 5.02276778e-01 6.22966886e-01 4.79633301e-01
2.75499433e-01 1.88971274e-02 1.79130450e-01 5.96567094e-02
-4.85604912e-01 -3.31999242e-01 4.33972567e-01 1.50152397e+00
-2.35812385e-02 -5.49091160e-01 -4.48467046e-01 9.17974532e-01
-2.07641077e+00 -1.25159371e+00 -3.37270141e-01 1.67207193e+00
8.56518805e-01 1.79674134e-01 1.03514805e-01 -2.18706369e-01
7.96500683e-01 9.47510660e-01 -2.76515365e-01 -8.97682607e-01
-1.91940904e-01 1.38749316e-01 2.94435054e-01 5.48717201e-01
-1.01346648e+00 1.28514671e+00 7.09690475e+00 7.38193870e-01
-8.32182646e-01 -7.94601440e-02 4.03455675e-01 -2.26151034e-01
-3.56579185e-01 -1.78554207e-01 -9.03246999e-01 1.15012020e-01
6.04981542e-01 -8.21853936e-01 -3.66527826e-01 5.94404221e-01
4.28384721e-01 -3.80186141e-01 -9.29923117e-01 5.83560050e-01
8.40343595e-01 -1.74111974e+00 8.84021342e-01 -3.91109437e-02
1.14035881e+00 -1.47860572e-01 -1.87872350e-01 2.76967585e-01
1.59506708e-01 -7.29330420e-01 5.20363927e-01 5.94977856e-01
5.27911425e-01 -9.83332753e-01 9.38671172e-01 4.20319766e-01
-8.41859043e-01 4.00575727e-01 -4.96087879e-01 1.54004142e-01
3.98723036e-01 6.44259989e-01 -1.10869515e+00 8.50125253e-01
2.67712563e-01 1.06928170e+00 -8.54784906e-01 1.01987028e+00
-5.97299218e-01 7.51779675e-01 1.69745803e-01 -4.69649345e-01
4.93402183e-01 1.42992184e-01 1.28003931e+00 1.88308811e+00
1.43965662e-01 1.25075921e-01 -2.46429041e-01 4.62584406e-01
-6.06597364e-01 3.75354439e-01 -7.22759545e-01 -2.46749192e-01
5.06327115e-02 1.31255448e+00 -7.45170891e-01 -8.36716354e-01
2.35322371e-01 1.09507811e+00 1.36272699e-01 4.71296817e-01
-3.39832187e-01 -9.14026856e-01 -3.17198709e-02 -2.53178775e-01
5.85491180e-01 -3.08156371e-01 -3.69168460e-01 -1.36498129e+00
1.93613708e-01 -8.49643052e-01 3.60965937e-01 -6.05135620e-01
-7.39492178e-01 7.09435284e-01 2.29978621e-01 -1.00502026e+00
-3.05790365e-01 1.31298915e-01 -1.29487920e+00 3.63853693e-01
-1.42651320e+00 -1.12906468e+00 -2.58828998e-01 -7.77389780e-02
1.62762368e+00 -3.30058038e-01 5.47218502e-01 -4.85095620e-01
-4.38922882e-01 2.73209274e-01 5.64375408e-02 4.32561114e-02
7.60751009e-01 -1.60789633e+00 9.80482876e-01 1.01921856e+00
1.29606515e-01 8.00170600e-01 1.09781134e+00 -9.58561003e-01
-1.17335653e+00 -1.15510094e+00 1.23319638e+00 -4.26084429e-01
2.83500522e-01 -6.42732903e-03 -5.64690530e-01 7.35433459e-01
1.42561996e+00 -1.26095307e+00 8.27378452e-01 8.03557858e-02
-1.42112067e-02 1.10288687e-01 -6.79677367e-01 9.21641052e-01
8.27996254e-01 -4.90850993e-02 -1.48477173e+00 4.89130199e-01
1.18530405e+00 -3.67170125e-01 -4.14619982e-01 3.21443051e-01
3.87757480e-01 -7.68372357e-01 6.42227232e-01 -8.98702502e-01
1.07174051e+00 1.55208081e-01 3.55433166e-01 -1.69351709e+00
-1.04546987e-01 -1.17000806e+00 -6.59042656e-01 1.39924753e+00
4.47712481e-01 -3.97231400e-01 6.24546826e-01 -1.41886428e-01
-5.54943740e-01 -7.33160079e-01 -5.46938062e-01 -2.59885490e-01
-9.64126363e-02 2.94131696e-01 3.68398219e-01 5.04210949e-01
2.24079370e-01 1.02441716e+00 -3.55134994e-01 -3.63042206e-01
5.08991659e-01 2.06715018e-01 9.90429461e-01 -8.57846558e-01
3.21057320e-01 -7.58507252e-01 -1.00722410e-01 -1.24228525e+00
4.44869906e-01 -7.12765038e-01 3.06775235e-02 -2.55083179e+00
8.11837137e-01 1.04597485e+00 2.76057702e-02 1.19529024e-01
-5.40187180e-01 -2.99045239e-02 4.24786389e-01 1.76526278e-01
-1.32449031e+00 8.33461821e-01 1.41918111e+00 -6.13280475e-01
-1.65632591e-01 3.54952849e-02 -1.27118933e+00 6.47318959e-01
9.38334882e-01 -4.27900851e-01 -2.82796800e-01 -5.23039997e-01
2.43124992e-01 -3.48023102e-02 -1.29637212e-01 -7.55131781e-01
4.58569586e-01 1.42514333e-01 8.94169323e-03 -1.50512540e+00
7.61884302e-02 7.03286827e-02 -8.02035093e-01 3.69658858e-01
-8.50399613e-01 2.53009677e-01 3.36931169e-01 6.25268877e-01
-5.02811551e-01 -5.14613509e-01 1.94707543e-01 -2.48654261e-01
-4.05283064e-01 1.03782956e-02 -5.35755515e-01 4.95769471e-01
5.42045832e-01 -3.45193520e-02 -7.64015496e-01 -1.00111485e+00
-2.43811443e-01 4.11190808e-01 2.37247441e-03 4.28740799e-01
1.02577734e+00 -8.55810225e-01 -1.43607688e+00 -4.39470381e-01
-1.06271833e-01 3.30469497e-02 2.09355667e-01 6.41457200e-01
-8.47853243e-01 8.72135460e-01 -1.04887553e-01 -2.22073928e-01
-1.67862964e+00 1.35344476e-01 -3.45803589e-01 -6.64230824e-01
-1.02623558e+00 5.29096961e-01 1.91887662e-01 1.39179096e-01
1.89236283e-01 -4.08229560e-01 -7.48407483e-01 5.35330713e-01
8.46981883e-01 7.49044418e-01 -1.48739843e-02 -5.42347729e-01
1.37558222e-01 3.54423732e-01 -7.40042746e-01 -1.02451466e-01
1.46831012e+00 -1.66416332e-01 -5.55211008e-01 2.44777873e-01
1.07969332e+00 5.13385713e-01 -7.14680076e-01 -1.10695526e-01
2.99551971e-02 1.11038787e-02 -7.56908581e-02 -6.79268539e-01
-6.01775169e-01 9.00662541e-01 -5.12869000e-01 4.30241466e-01
1.03839493e+00 -1.61985494e-02 1.35341311e+00 9.63776529e-01
-3.61977428e-01 -1.34845531e+00 3.12427342e-01 8.16914916e-01
1.53153205e+00 -9.33063090e-01 8.35176826e-01 -3.52719665e-01
-1.03416872e+00 1.23119557e+00 3.41166556e-01 -4.58769143e-01
9.73897353e-02 -1.53363347e-01 -2.30456963e-01 -5.40338635e-01
-1.02708399e+00 -1.37490369e-02 6.65699661e-01 2.67364711e-01
5.21082103e-01 -1.38484254e-01 -8.61646950e-01 4.72988486e-01
-6.79465830e-01 -5.60381353e-01 9.63391364e-01 8.97209585e-01
-9.57367122e-01 -4.88127708e-01 -4.17490937e-02 7.64437735e-01
-1.01431835e+00 -3.62326771e-01 -1.20315373e+00 7.36246705e-01
-9.09063697e-01 9.02179599e-01 -1.15013696e-01 6.22591488e-02
4.29360449e-01 4.78664339e-02 1.90024734e-01 -1.13623023e+00
-1.04665804e+00 2.27700040e-01 6.18022799e-01 -2.87482023e-01
-2.77229697e-01 -4.35161948e-01 -1.18534291e+00 -3.12107146e-01
-1.60656378e-01 5.66500604e-01 6.47976637e-01 8.34873736e-01
7.18966842e-01 1.03733218e+00 4.43481416e-01 -9.25305486e-01
-6.54401362e-01 -1.48817396e+00 -3.03337306e-01 1.90917328e-01
6.38625681e-01 4.08718079e-01 -1.39343679e-01 2.08182365e-01] | [12.544774055480957, 9.50643539428711] |
2cdabe5a-607c-4f0c-9605-24864fa531f7 | amigo-sparse-multi-modal-graph-transformer | 2303.00865 | null | https://arxiv.org/abs/2303.00865v2 | https://arxiv.org/pdf/2303.00865v2.pdf | AMIGO: Sparse Multi-Modal Graph Transformer with Shared-Context Processing for Representation Learning of Giga-pixel Images | Processing giga-pixel whole slide histopathology images (WSI) is a computationally expensive task. Multiple instance learning (MIL) has become the conventional approach to process WSIs, in which these images are split into smaller patches for further processing. However, MIL-based techniques ignore explicit information about the individual cells within a patch. In this paper, by defining the novel concept of shared-context processing, we designed a multi-modal Graph Transformer (AMIGO) that uses the celluar graph within the tissue to provide a single representation for a patient while taking advantage of the hierarchical structure of the tissue, enabling a dynamic focus between cell-level and tissue-level information. We benchmarked the performance of our model against multiple state-of-the-art methods in survival prediction and showed that ours can significantly outperform all of them including hierarchical Vision Transformer (ViT). More importantly, we show that our model is strongly robust to missing information to an extent that it can achieve the same performance with as low as 20% of the data. Finally, in two different cancer datasets, we demonstrated that our model was able to stratify the patients into low-risk and high-risk groups while other state-of-the-art methods failed to achieve this goal. We also publish a large dataset of immunohistochemistry images (InUIT) containing 1,600 tissue microarray (TMA) cores from 188 patients along with their survival information, making it one of the largest publicly available datasets in this context. | ['Ali Bashashati', 'Blake Gilks', 'Alexander Baras', 'Hossein Farahani', 'Haoyang Mi', 'Puria Azadi Moghadam', 'Ramin Nakhli'] | 2023-03-01 | null | null | null | null | ['multiple-instance-learning'] | ['methodology'] | [ 4.11927551e-01 1.63839430e-01 -3.52273062e-02 -1.00412793e-01
-1.36834192e+00 -4.23125178e-01 4.32668179e-01 6.39685571e-01
-3.52403075e-01 5.64562559e-01 2.24545226e-01 -3.09429795e-01
-2.23853156e-01 -8.14792871e-01 -4.92507011e-01 -1.32056129e+00
-1.56380668e-01 5.97896039e-01 3.47870976e-01 -1.53189316e-01
2.55566230e-03 6.42144620e-01 -1.23809707e+00 8.68980825e-01
4.92260844e-01 9.53662872e-01 1.76781103e-01 7.85334289e-01
-1.85913324e-01 6.56212330e-01 -2.30293527e-01 -1.04903221e-01
-1.08803101e-01 -8.26474130e-02 -8.62565041e-01 4.58151735e-02
5.61324894e-01 2.86921382e-01 -2.38978386e-01 6.89099371e-01
5.03788054e-01 -5.76836109e-01 7.04112947e-01 -9.63917255e-01
1.12047851e-01 5.07214844e-01 -6.96454108e-01 7.69358650e-02
-9.63602066e-02 1.77625880e-01 1.11085522e+00 -7.12359965e-01
9.89979982e-01 6.90144658e-01 8.89522433e-01 3.86520863e-01
-1.61050105e+00 -4.82803613e-01 -1.31956367e-02 1.26740158e-01
-1.51083064e+00 -2.30316550e-01 5.15230596e-01 -4.76938754e-01
9.79320228e-01 5.94208777e-01 8.15044641e-01 7.57543027e-01
7.74912179e-01 6.13250494e-01 1.52255666e+00 -4.33747143e-01
1.28118932e-01 -1.72138140e-01 4.30169761e-01 9.05061126e-01
4.50754352e-02 -2.44039387e-01 -4.20748979e-01 -3.42815757e-01
4.25430566e-01 3.34005326e-01 -1.95263073e-01 -2.23218873e-01
-1.47720098e+00 6.31196380e-01 6.39358580e-01 6.10049069e-01
-1.65363401e-01 4.45538759e-02 3.60516459e-01 1.68170884e-01
5.64555824e-01 1.55011579e-01 -4.30026442e-01 4.52492952e-01
-9.90425467e-01 -1.63339615e-01 6.89529002e-01 2.48851970e-01
7.64238656e-01 -6.90896988e-01 -4.28892523e-01 5.07482409e-01
1.15521833e-01 -8.51089135e-02 5.83080113e-01 -4.96860296e-01
-7.19638020e-02 9.99772191e-01 -4.86942798e-01 -8.18039715e-01
-8.94705713e-01 -6.85308695e-01 -1.21536613e+00 1.58119842e-01
6.19670093e-01 2.27384344e-01 -1.13498318e+00 1.52495885e+00
3.80679965e-01 4.71002012e-01 -1.07766591e-01 5.46646297e-01
9.41402197e-01 3.25797528e-01 2.83428371e-01 -3.40219975e-01
1.62762153e+00 -7.14881361e-01 -3.77261132e-01 -8.09624884e-03
1.10678267e+00 -5.33144653e-01 7.10740924e-01 3.61286491e-01
-7.15669215e-01 -8.09295550e-02 -8.10295463e-01 6.89830557e-02
-4.78378952e-01 -2.51369197e-02 9.35562909e-01 3.83741170e-01
-1.39067769e+00 5.79084516e-01 -9.16879237e-01 -5.87923765e-01
6.41809821e-01 6.13703907e-01 -8.58418941e-01 -1.70452923e-01
-6.29363894e-01 6.35836601e-01 -4.94269803e-02 -1.65365383e-01
-9.93446350e-01 -1.33552706e+00 -5.64312220e-01 3.84317711e-02
2.68712401e-01 -9.73023772e-01 6.84351325e-01 -5.62902510e-01
-1.10452855e+00 1.33917999e+00 -3.78630280e-01 -3.24688166e-01
3.82987440e-01 4.87283409e-01 -1.39001191e-01 3.48530948e-01
-1.23313420e-01 5.59612155e-01 4.18298542e-01 -1.14337420e+00
-7.08116651e-01 -7.52302885e-01 -2.77745813e-01 -1.01414785e-01
-4.01535034e-01 -3.12859178e-01 -5.03111064e-01 -4.80541795e-01
2.14178652e-01 -1.00114632e+00 -4.99058068e-01 6.07984737e-02
-6.24526501e-01 -9.19720605e-02 6.10885680e-01 -5.41965783e-01
1.06214678e+00 -2.12710047e+00 3.81952316e-01 3.79053980e-01
5.78548014e-01 -1.69176295e-01 -2.03822091e-01 4.56571102e-01
-1.43880576e-01 2.53489792e-01 -1.78344131e-01 -6.23894274e-01
-3.20805550e-01 1.65604174e-01 1.88519880e-01 7.19432831e-01
-5.99735044e-02 9.78942156e-01 -8.04123104e-01 -8.88112009e-01
6.78527281e-02 5.04754663e-01 -4.90547776e-01 1.09502204e-01
4.64578792e-02 4.61336672e-01 -1.53172135e-01 7.52958298e-01
4.18637246e-01 -6.49730682e-01 5.26412725e-01 -4.91043657e-01
1.28970429e-01 -1.73989147e-01 -8.00898373e-01 1.54663885e+00
-3.12726945e-01 3.37892354e-01 3.29044849e-01 -9.14317906e-01
3.63165051e-01 2.52694279e-01 1.06185460e+00 -3.74581009e-01
6.97455704e-02 1.67432889e-01 3.03207021e-02 -2.13877574e-01
-1.66733980e-01 -3.61046910e-01 -6.64451122e-02 2.90828258e-01
4.07623835e-02 -1.36986226e-02 1.60781875e-01 3.00742179e-01
1.67764115e+00 -4.66174453e-01 3.96263897e-01 -4.96361911e-01
5.94770789e-01 1.80652410e-01 6.09074712e-01 6.21280730e-01
-1.14614785e-01 7.49902844e-01 7.37859905e-01 -4.14808452e-01
-6.86592758e-01 -9.47753310e-01 -2.69984365e-01 8.94208729e-01
-1.88676432e-01 -5.09010792e-01 -5.51303506e-01 -8.33448052e-01
1.37681201e-01 3.96586433e-02 -1.12863421e+00 1.24822952e-01
-3.94263417e-01 -1.48912644e+00 3.67943764e-01 2.23026186e-01
2.07788683e-02 -5.03181875e-01 -1.80236340e-01 2.31302276e-01
-1.55120239e-01 -1.00031388e+00 -3.13550830e-01 3.23612213e-01
-9.48782802e-01 -1.37882185e+00 -5.52140176e-01 -9.26982105e-01
1.06690180e+00 2.91063458e-01 1.18691337e+00 4.71787959e-01
-1.01908398e+00 1.32753357e-01 -1.29932791e-01 -2.73497254e-01
-5.02812505e-01 1.79072365e-01 -5.08240879e-01 2.62616724e-01
1.00855812e-01 -5.23358047e-01 -7.75584459e-01 1.85364157e-01
-9.10285234e-01 3.93171310e-01 9.63978291e-01 1.21674025e+00
1.24245811e+00 1.15473114e-01 3.45641375e-01 -1.33043516e+00
2.53982134e-02 -4.66566622e-01 -2.71077484e-01 4.37164456e-01
-4.72219378e-01 -1.67847365e-01 5.03645658e-01 -8.82130638e-02
-7.33739913e-01 3.30822170e-01 -1.79847002e-01 4.95764911e-02
-4.76504341e-02 7.27567017e-01 6.47805408e-02 -4.62269336e-01
3.77664566e-01 1.27045006e-01 3.52615267e-01 -2.99143344e-01
-7.98010603e-02 2.80034810e-01 2.38044024e-01 -8.99274945e-02
4.63224083e-01 8.79671633e-01 7.91831970e-01 -6.79432154e-01
-9.38756168e-01 -7.83120453e-01 -7.12021828e-01 -9.00415182e-02
7.65726566e-01 -7.18882799e-01 -7.94932544e-01 7.08760679e-01
-4.73878980e-01 -5.59960365e-01 -7.96620846e-02 5.10562770e-02
-3.88566464e-01 3.65926147e-01 -1.18690050e+00 -2.66382575e-01
-4.53759074e-01 -1.17695689e+00 1.46145284e+00 -1.39083311e-01
4.38772105e-02 -1.23528957e+00 2.42424294e-01 4.49264169e-01
4.48200017e-01 7.39834547e-01 1.31685281e+00 -5.49471617e-01
-5.07052660e-01 -3.72988433e-01 -1.57543749e-01 -2.48947874e-01
2.97453433e-01 1.80971190e-01 -1.02501106e+00 -5.63278258e-01
-3.98154378e-01 -8.63826498e-02 1.31136107e+00 4.68241245e-01
1.40185714e+00 -4.19062227e-02 -1.07162917e+00 7.60842264e-01
1.85382080e+00 -4.84724879e-01 7.02369034e-01 1.97433516e-01
6.73306942e-01 6.61550522e-01 3.70430022e-01 1.30439043e-01
4.63105410e-01 5.89621484e-01 5.36485434e-01 -6.12256348e-01
-3.17830831e-01 1.07985221e-01 -1.27337920e-02 5.26247978e-01
1.97697841e-02 -7.27209225e-02 -1.10125482e+00 6.20746672e-01
-1.75354898e+00 -6.76773787e-01 -2.60590971e-01 2.02959323e+00
8.58075738e-01 -8.65318626e-03 -5.01174740e-02 1.78752095e-01
2.88374305e-01 -5.35695516e-02 -2.90456444e-01 2.21437678e-01
-2.12175637e-01 2.65511215e-01 5.32302022e-01 4.28627133e-01
-1.13349187e+00 6.26233876e-01 6.78151035e+00 9.96873438e-01
-1.19423246e+00 7.41251409e-02 1.25230837e+00 -3.53749134e-02
-2.66415179e-01 -1.68437317e-01 -7.74005592e-01 1.59985557e-01
9.01786327e-01 -5.02051339e-02 -5.13803139e-02 4.29587334e-01
5.05292267e-02 -1.98326364e-01 -1.13002336e+00 7.04625547e-01
-5.70178181e-02 -1.76635826e+00 2.45815497e-02 5.22990167e-01
6.30074561e-01 2.02026546e-01 -1.03299953e-01 8.80374387e-02
3.05616677e-01 -1.21524298e+00 2.41987727e-04 6.92762375e-01
8.53758752e-01 -4.52356189e-01 1.16630089e+00 2.82452703e-01
-1.06873095e+00 -3.33173536e-02 -1.75866991e-01 2.77842969e-01
-1.98296309e-01 9.74078894e-01 -1.15660119e+00 7.87416875e-01
5.45175552e-01 7.93483853e-01 -1.05568731e+00 9.70381141e-01
4.20728058e-01 6.06881380e-01 -3.52674276e-01 2.92264581e-01
5.96261397e-02 1.92810267e-01 1.78168342e-01 1.40168917e+00
4.67194110e-01 1.14945576e-01 2.40498468e-01 2.09920198e-01
6.15728945e-02 1.96215361e-01 -2.02267081e-01 2.32819006e-01
2.84794811e-02 1.90108347e+00 -9.46063876e-01 -2.08853483e-01
-3.76846045e-01 5.98421395e-01 5.79274833e-01 6.01977967e-02
-4.54206496e-01 1.56296253e-01 5.08848786e-01 4.20615822e-01
1.24672003e-01 1.37856394e-01 -3.34353268e-01 -9.45336938e-01
-3.89504761e-01 -5.49756646e-01 9.51168656e-01 -2.71407723e-01
-1.65270078e+00 4.77627516e-01 -3.81114423e-01 -1.18037951e+00
2.09412113e-01 -8.44951212e-01 -5.90762794e-01 6.97742820e-01
-1.93139243e+00 -1.68695712e+00 -5.66724837e-01 5.97126842e-01
3.40277031e-02 5.70452921e-02 1.12668061e+00 1.02170959e-01
-6.01422191e-01 6.01464272e-01 1.12854078e-01 5.04764430e-02
8.36124420e-01 -1.35350037e+00 -3.14882040e-01 3.27454537e-01
-5.92408367e-02 4.20380026e-01 5.00866354e-01 -4.61223304e-01
-1.72392964e+00 -1.19869351e+00 7.98765182e-01 -4.58127350e-01
8.98572743e-01 -3.52897614e-01 -8.83028448e-01 6.89049304e-01
3.64922807e-02 4.41595852e-01 1.39221358e+00 1.17548399e-01
-8.85765851e-02 -3.96455675e-01 -1.23324132e+00 5.60266614e-01
8.78735423e-01 -2.35519558e-01 -2.32588686e-02 5.59384167e-01
4.09601808e-01 -4.51042920e-01 -1.29504132e+00 5.34760296e-01
3.72480929e-01 -1.01495743e+00 9.19016898e-01 -3.62632871e-01
2.58958846e-01 -3.31575572e-01 -5.83748929e-02 -1.24378192e+00
-6.50005043e-01 -1.38065398e-01 8.55806395e-02 1.08654821e+00
5.60937643e-01 -8.10654938e-01 9.80992913e-01 2.52459258e-01
-3.73074859e-01 -1.23009491e+00 -1.15151262e+00 -3.03713709e-01
2.22085699e-01 -1.37116149e-01 4.13164943e-01 6.79219902e-01
3.08805734e-01 2.92795636e-02 1.07300080e-01 1.70923829e-01
8.15241516e-01 3.16868842e-01 5.59296310e-01 -1.35123134e+00
-3.92596871e-01 -5.40933490e-01 -6.28838599e-01 -1.70903444e-01
1.68224156e-01 -1.31151462e+00 -3.09537619e-01 -1.64985478e+00
7.99070060e-01 -5.74224353e-01 -8.07818532e-01 7.17735112e-01
-3.76869023e-01 7.60721922e-01 -3.47867012e-02 3.95215064e-01
-6.07149005e-01 -9.05878693e-02 1.32374561e+00 -4.75321412e-01
2.12310821e-01 -2.92486042e-01 -8.59919667e-01 4.68875051e-01
5.46245337e-01 -2.78289706e-01 3.32209244e-02 3.26259732e-02
3.39459069e-02 1.98612079e-01 4.68255848e-01 -1.01143992e+00
5.09599268e-01 -1.52076289e-01 5.90757668e-01 -5.64836800e-01
2.72317976e-01 -7.63017893e-01 5.42582452e-01 6.70936227e-01
-2.19799489e-01 -4.43322480e-01 2.01873988e-01 5.30587018e-01
-2.79537320e-01 3.49404991e-01 7.52733648e-01 -2.06489325e-01
-5.04881203e-01 6.29726350e-01 -3.20657253e-01 -5.03301442e-01
1.34688413e+00 -1.06770560e-01 -6.51789486e-01 1.42287016e-01
-1.01772439e+00 3.65814835e-01 6.15755081e-01 -1.89966038e-01
4.02022511e-01 -1.01678932e+00 -1.01695418e+00 1.29284650e-01
4.25256371e-01 4.59148921e-02 7.67785311e-01 1.32991791e+00
-5.08244932e-01 4.75914598e-01 -4.51000109e-02 -1.00546730e+00
-1.60586989e+00 5.32366216e-01 5.41687906e-01 -1.16991353e+00
-6.23704553e-01 7.75370836e-01 6.23429000e-01 -3.46227497e-01
-1.34226397e-01 -2.53717661e-01 -3.26239318e-01 1.18317463e-01
5.28426707e-01 -6.39072210e-02 4.28526312e-01 -4.95355308e-01
-4.81845647e-01 7.29467332e-01 -3.83966655e-01 3.41070980e-01
1.44550097e+00 8.22527185e-02 -5.83756566e-01 4.11320627e-01
1.23363972e+00 -5.02326861e-02 -1.04920959e+00 -1.51120171e-01
-8.22817907e-03 -1.99054241e-01 2.64095634e-01 -8.58961046e-01
-1.19660032e+00 5.97246647e-01 4.71080482e-01 1.47234574e-01
1.32743990e+00 2.07128584e-01 5.70169628e-01 -8.74764994e-02
5.81400394e-01 -7.58343220e-01 1.12657305e-02 1.40573129e-01
5.76546550e-01 -1.20438993e+00 8.29200149e-02 -8.96375716e-01
-2.97461927e-01 1.07692337e+00 3.01531553e-01 1.01704849e-03
8.13221633e-01 5.58689654e-01 2.00936005e-01 -4.78632003e-01
-1.15110862e+00 -8.91505331e-02 1.81307450e-01 3.89811069e-01
5.23596287e-01 3.04993898e-01 -1.13497391e-01 6.07533455e-01
-7.62924105e-02 9.51609239e-02 4.11085963e-01 8.15497994e-01
-2.16588825e-01 -1.08374023e+00 -2.48481825e-01 7.67481685e-01
-7.03317761e-01 -8.68465528e-02 -3.55299741e-01 8.54772627e-01
-5.44856451e-02 6.06367707e-01 -7.18118772e-02 -2.58812189e-01
1.71836361e-01 -5.18405661e-02 4.31618005e-01 -6.46528780e-01
-7.65185118e-01 1.93914503e-01 -9.04794261e-02 -6.12182915e-01
-3.69108438e-01 -7.10654855e-01 -1.28361964e+00 -1.81621686e-01
-4.51864786e-02 -8.71208161e-02 4.11386371e-01 8.31746221e-01
4.09229368e-01 7.83460319e-01 6.33126199e-01 -7.17032194e-01
-9.88605544e-02 -7.19821513e-01 -8.06758761e-01 4.26876277e-01
4.75604892e-01 -2.97966391e-01 -6.06412053e-01 6.17588796e-02] | [15.116127967834473, -2.934734344482422] |
c6585cfe-94e1-4e1f-829e-aee4ad6cff9b | tsgp-two-stage-generative-prompting-for | 2211.13515 | null | https://arxiv.org/abs/2211.13515v1 | https://arxiv.org/pdf/2211.13515v1.pdf | TSGP: Two-Stage Generative Prompting for Unsupervised Commonsense Question Answering | Unsupervised commonsense question answering requires mining effective commonsense knowledge without the rely on the labeled task data. Previous methods typically retrieved from traditional knowledge bases or used pre-trained language models (PrLMs) to generate fixed types of knowledge, which have poor generalization ability. In this paper, we aim to address the above limitation by leveraging the implicit knowledge stored in PrLMs and propose a two-stage prompt-based unsupervised commonsense question answering framework (TSGP). Specifically, we first use knowledge generation prompts to generate the knowledge required for questions with unlimited types and possible candidate answers independent of specified choices. Then, we further utilize answer generation prompts to generate possible candidate answers independent of specified choices. Experimental results and analysis on three different commonsense reasoning tasks, CommonsenseQA, OpenBookQA, and SocialIQA, demonstrate that TSGP significantly improves the reasoning ability of language models in unsupervised settings. Our code is available at: https://github.com/Yueqing-Sun/TSGP. | ['Qi Shi', 'Le Qi', 'Yu Zhang', 'Yueqing Sun'] | 2022-11-24 | null | null | null | null | ['answer-generation'] | ['natural-language-processing'] | [ 2.97085226e-01 4.39662576e-01 -9.25846249e-02 -4.04764444e-01
-5.83256483e-01 -6.68661714e-01 4.97612894e-01 1.77646488e-01
-3.84780943e-01 8.89159560e-01 2.44157538e-01 -6.37181520e-01
-1.67433962e-01 -1.22538257e+00 -4.31025147e-01 -8.50974843e-02
6.58923984e-01 4.86760408e-01 3.20620179e-01 -6.68070257e-01
3.97093415e-01 -2.08636194e-01 -1.34604943e+00 3.73054862e-01
1.68550408e+00 6.64117575e-01 2.46667117e-01 4.55986768e-01
-7.55386233e-01 1.73098302e+00 -4.79433089e-01 -7.26170421e-01
-1.57964125e-01 -6.94351256e-01 -1.34079444e+00 -4.49044168e-01
-2.99066678e-02 -4.07558888e-01 -2.76028663e-01 1.19208717e+00
2.41771042e-01 4.71995056e-01 5.68383813e-01 -1.22541165e+00
-1.43892848e+00 1.28586495e+00 2.56055057e-01 3.47232193e-01
6.71167791e-01 2.59637773e-01 1.28499389e+00 -7.13468611e-01
5.70223749e-01 1.29071617e+00 2.92175770e-01 1.01529956e+00
-9.49661851e-01 -7.65427530e-01 -1.57866683e-02 7.85641372e-01
-1.10095012e+00 -4.71217960e-01 1.19259787e+00 -2.78508484e-01
1.00467372e+00 2.53650725e-01 4.49281693e-01 1.13803375e+00
-2.89635479e-01 1.07761192e+00 1.44507742e+00 -7.74597585e-01
4.04348582e-01 1.54229313e-01 6.93852186e-01 7.18877792e-01
1.45934626e-01 -1.36228785e-01 -5.83523273e-01 -2.93520153e-01
5.10910988e-01 -1.31096348e-01 -2.65004158e-01 6.88093379e-02
-1.17151213e+00 1.11245394e+00 2.85658777e-01 4.09474880e-01
-3.19110215e-01 4.28956300e-02 2.91467249e-01 4.66598362e-01
2.08665580e-01 7.88199306e-01 -4.28711534e-01 -2.05366865e-01
-6.93466306e-01 2.49597624e-01 9.61469531e-01 1.04081404e+00
8.66104722e-01 -7.36356676e-02 -3.38906825e-01 8.75596941e-01
2.51708865e-01 6.75515294e-01 8.30346048e-01 -1.10713005e+00
5.55835128e-01 1.00989640e+00 2.19239127e-02 -1.03875268e+00
-1.86982378e-01 -1.64443385e-02 -3.78145844e-01 -6.07136250e-01
3.23458791e-01 -2.15075091e-01 -6.44406974e-01 1.89047492e+00
3.92810762e-01 5.67082837e-02 4.40182328e-01 7.94054866e-01
1.22860408e+00 5.46543241e-01 2.83242345e-01 -6.17365763e-02
1.29283988e+00 -9.04075325e-01 -9.04084206e-01 -3.59124482e-01
8.76835585e-01 -3.21229577e-01 1.77162790e+00 1.46879882e-01
-7.51440763e-01 -2.27632046e-01 -7.93011308e-01 -4.94022220e-01
-5.46527326e-01 -3.59498933e-02 8.27561438e-01 4.28370535e-01
-7.80928969e-01 1.58900663e-01 -4.12545502e-01 -1.58849776e-01
4.77890670e-01 -7.35046417e-02 2.14370564e-01 -3.64298522e-01
-1.76102328e+00 8.34240675e-01 8.08076978e-01 -8.70662034e-02
-7.00397491e-01 -5.81869006e-01 -9.81544256e-01 1.77464988e-02
9.51285720e-01 -8.01349044e-01 1.35697556e+00 -9.25361037e-01
-1.69691384e+00 8.78744543e-01 -2.89728135e-01 -4.70385909e-01
2.73797512e-01 -3.50452840e-01 -3.91585350e-01 2.62425542e-01
3.32872897e-01 5.04302859e-01 4.41266358e-01 -1.06479192e+00
-1.75233021e-01 -1.43253714e-01 6.38604999e-01 3.55579630e-02
-3.56293857e-01 -1.94930062e-02 1.57350942e-01 -6.26364470e-01
-5.73436124e-03 -7.66579568e-01 5.50610758e-02 -4.97014791e-01
-6.61777794e-01 -6.76639020e-01 2.70901799e-01 -6.11983955e-01
1.27225232e+00 -1.78211129e+00 -1.09428450e-01 7.70577565e-02
2.82600909e-01 1.73812971e-01 -1.77493423e-01 3.62564713e-01
1.15734339e-01 2.03243151e-01 -3.61214906e-01 2.42369816e-01
3.15555513e-01 3.91807407e-01 -8.31772149e-01 -5.28366029e-01
1.74256891e-01 1.32666147e+00 -1.46416593e+00 -7.89010346e-01
7.66451210e-02 -2.43092358e-01 -6.32865906e-01 3.10551047e-01
-7.93152452e-01 2.42211688e-02 -8.07965279e-01 7.02968538e-01
3.94184649e-01 -6.00719690e-01 1.56004518e-01 8.24257657e-02
3.82678926e-01 7.39641428e-01 -7.90077567e-01 1.55719399e+00
-4.68526781e-01 3.30933779e-01 -6.54971659e-01 -8.25459123e-01
9.08586621e-01 1.47795856e-01 -1.50445640e-01 -6.65808141e-01
1.48057476e-01 2.73363560e-01 1.22001581e-01 -8.09362829e-01
4.18594152e-01 -6.60149813e-01 -1.08039491e-01 6.91395342e-01
1.20262727e-01 -3.99821401e-01 3.39148223e-01 6.21061623e-01
1.28328729e+00 3.58772799e-02 4.65782315e-01 -1.46943524e-01
6.11322641e-01 3.93116325e-01 4.11908329e-01 8.86906683e-01
-3.27366620e-01 -5.66539802e-02 5.00456214e-01 -7.89535046e-02
-4.60365891e-01 -1.15023685e+00 3.09914052e-01 1.30619490e+00
2.84623086e-01 -4.77504879e-01 -5.22021830e-01 -7.96142995e-01
-2.58378357e-01 1.70591676e+00 -4.86190110e-01 -4.05776858e-01
-4.42511350e-01 -3.68087083e-01 7.82765031e-01 3.94871891e-01
7.93682694e-01 -1.68612778e+00 -6.63608670e-01 2.52341688e-01
-7.52401114e-01 -1.11913943e+00 -1.06099524e-01 1.22409500e-01
-7.70634711e-01 -1.24859822e+00 -1.06678724e-01 -5.66267371e-01
4.77944821e-01 7.23908022e-02 1.31231189e+00 2.45045990e-01
1.09754235e-01 5.73921442e-01 -7.61174142e-01 -4.27111924e-01
-4.86531854e-01 -9.82244387e-02 -8.91457349e-02 -3.89503032e-01
1.11683786e+00 -8.23779464e-01 -2.58682072e-01 -1.26680642e-01
-1.01412570e+00 1.48466349e-01 3.01668912e-01 7.26709247e-01
3.71614307e-01 -8.46954659e-02 1.01804948e+00 -9.96450841e-01
1.19776726e+00 -8.75875890e-01 -1.80018753e-01 5.76621115e-01
-4.55574542e-01 2.70569772e-01 8.15141082e-01 -4.45664406e-01
-1.27838659e+00 -4.86293793e-01 6.14893176e-02 -2.78158456e-01
-2.11540118e-01 7.08788037e-01 -2.03776434e-01 4.07223761e-01
9.02371526e-01 6.03359461e-01 -3.18466634e-01 -2.42656410e-01
6.55832291e-01 5.90325296e-01 4.52378243e-01 -1.22570503e+00
8.78069758e-01 2.58046329e-01 -5.51086664e-01 -4.71881062e-01
-1.42912972e+00 -2.80892402e-01 -2.33534262e-01 8.30233935e-03
8.23498666e-01 -5.87896943e-01 -4.74367410e-01 1.60776913e-01
-1.20531452e+00 -5.63103914e-01 -4.04394656e-01 2.42961228e-01
-6.36060834e-01 4.36942279e-01 -5.98087668e-01 -8.11358988e-01
-5.54665804e-01 -5.71309984e-01 4.65811223e-01 3.55750889e-01
-4.22132343e-01 -1.06382418e+00 4.11733240e-02 8.35184216e-01
3.58699739e-01 -2.06541456e-02 1.39707184e+00 -1.08555317e+00
-4.85383004e-01 1.75204366e-01 -1.87456131e-01 3.79042298e-01
2.91997612e-01 -3.08038235e-01 -8.37362289e-01 5.50108969e-01
1.01753637e-01 -1.06256187e+00 8.40106428e-01 -1.40515015e-01
1.35988069e+00 -5.08974016e-01 1.56388938e-01 1.22560754e-01
1.06214154e+00 -3.73354331e-02 6.28729761e-01 4.74450260e-01
4.86811996e-01 5.36035240e-01 5.88722289e-01 4.37147975e-01
9.47243631e-01 4.15441431e-02 4.93758731e-03 6.24421656e-01
3.81308212e-03 -6.48553252e-01 4.11703050e-01 9.22622263e-01
-4.58373949e-02 1.37567380e-02 -1.15974891e+00 8.64654839e-01
-1.76796138e+00 -1.11533237e+00 7.98054636e-02 1.52311122e+00
1.62404621e+00 -7.53644034e-02 -2.42462128e-01 6.43285960e-02
6.06671870e-01 -2.64186040e-02 -8.29545736e-01 -2.79915869e-01
-1.34062141e-01 5.01687050e-01 -2.45005444e-01 6.34544075e-01
-6.44233584e-01 1.38946390e+00 5.18343258e+00 9.96017039e-01
-6.08161330e-01 1.37579918e-01 1.21479258e-01 -7.66704977e-02
-9.75998878e-01 2.64970720e-01 -4.00129616e-01 3.77204835e-01
8.09637010e-01 -4.95932728e-01 5.85337102e-01 8.72676730e-01
-7.68092722e-02 -7.49963596e-02 -9.88904059e-01 9.40043092e-01
5.68889901e-02 -1.21499193e+00 5.30500948e-01 -4.61001694e-01
5.57662904e-01 -2.10767210e-01 -3.33635032e-01 9.75171745e-01
7.14226484e-01 -8.62198114e-01 5.19133985e-01 6.73442721e-01
3.21787983e-01 -5.59030414e-01 5.30122340e-01 7.19643235e-01
-5.90143442e-01 -6.38319924e-02 -4.03003752e-01 -3.52838844e-01
1.97514072e-02 6.63171887e-01 -8.01567495e-01 4.12127942e-01
3.77930999e-01 3.71173084e-01 -7.11990058e-01 3.78091156e-01
-9.68019187e-01 1.07734692e+00 -2.96411902e-01 -5.50617278e-01
1.44736722e-01 3.75780948e-02 4.13055271e-01 8.93896520e-01
8.72609988e-02 8.42598617e-01 1.32452667e-01 1.50592911e+00
-2.29681045e-01 1.79941028e-01 -3.12868625e-01 -5.26355684e-01
9.22999322e-01 9.88431096e-01 -4.42230403e-01 -6.81063712e-01
-9.30489749e-02 8.54446411e-01 6.48110509e-01 4.22742784e-01
-7.94082761e-01 -6.66665137e-01 1.96349248e-01 -7.62834474e-02
-2.57974584e-02 -5.69558106e-02 -2.76369184e-01 -1.54216659e+00
1.42843813e-01 -9.71913278e-01 5.91222644e-01 -1.07091784e+00
-1.66365075e+00 2.67562300e-01 2.18259469e-01 -5.84895551e-01
-3.66331905e-01 -5.63438833e-01 -7.75773108e-01 6.73416853e-01
-1.74185419e+00 -1.05678689e+00 -3.91567290e-01 8.67813706e-01
5.18836617e-01 -5.25560230e-02 8.98959279e-01 -3.07278514e-01
-3.25715244e-01 5.24146497e-01 -3.58369231e-01 3.76996547e-01
5.13532221e-01 -1.27381408e+00 -1.41732236e-02 6.08170688e-01
1.70529500e-01 1.15718293e+00 6.43097520e-01 -7.69971251e-01
-1.20885360e+00 -8.64539802e-01 1.11946821e+00 -7.34449983e-01
9.92218614e-01 -1.76934637e-02 -1.12391031e+00 8.22935700e-01
2.59746104e-01 -2.73242503e-01 1.29524946e+00 2.13886872e-01
-7.20569611e-01 3.39218616e-01 -1.11149609e+00 7.99504220e-01
1.10595632e+00 -8.34747672e-01 -1.63451445e+00 4.23950255e-01
1.07631016e+00 -7.66152516e-02 -4.65026557e-01 1.37696877e-01
1.19354678e-02 -7.45180130e-01 5.48348188e-01 -8.67879331e-01
9.13506925e-01 -3.76901180e-01 -3.95279288e-01 -1.17426300e+00
-9.75571722e-02 -4.99105662e-01 -5.21666288e-01 1.22838163e+00
4.28918898e-01 -7.67148077e-01 3.95461112e-01 9.45817232e-01
-3.65198553e-02 -6.03576958e-01 -7.97498643e-01 -7.16217995e-01
2.41088763e-01 -6.87309325e-01 6.90038562e-01 1.34554172e+00
4.61104184e-01 7.37527549e-01 1.30424008e-01 -3.46786119e-02
5.49159884e-01 4.41636294e-01 5.52168369e-01 -9.85062242e-01
-3.80331814e-01 -1.57737836e-01 1.11889876e-01 -1.08779156e+00
6.21485174e-01 -1.21006954e+00 -1.86292194e-02 -1.65026653e+00
4.27283078e-01 -4.90708709e-01 -3.54558498e-01 9.34359014e-01
-6.83512688e-01 -1.76295742e-01 1.77182972e-01 4.18747403e-02
-8.75765085e-01 8.29048455e-01 1.27971911e+00 -1.05516508e-01
-1.90015867e-01 -5.35557568e-01 -1.13402724e+00 8.52865219e-01
1.15648353e+00 -3.95973235e-01 -8.80853891e-01 -4.56101209e-01
6.29081309e-01 -3.01013678e-01 7.17658401e-01 -7.04811335e-01
2.03784063e-01 -6.90864265e-01 -4.18730378e-02 -3.55257034e-01
1.18181542e-01 -3.78658235e-01 -5.71761549e-01 1.35697693e-01
-7.07758069e-01 -2.42232099e-01 1.29385844e-01 3.72653157e-01
-1.17357217e-01 -5.33252716e-01 2.94241935e-01 -3.79242122e-01
-9.64839518e-01 -2.38738865e-01 -1.80171639e-01 7.43396342e-01
7.77046919e-01 -6.39864504e-02 -6.09449089e-01 -4.03285503e-01
-6.17048979e-01 4.16559190e-01 1.53607577e-01 4.55835044e-01
8.93888175e-01 -1.37982702e+00 -6.48186266e-01 -3.52193803e-01
4.93753701e-01 3.31917666e-02 2.74538696e-01 5.39222181e-01
-2.79860437e-01 4.99432832e-01 7.59015754e-02 -2.50544716e-02
-7.98984110e-01 6.29200518e-01 2.36309454e-01 -2.21665725e-01
-3.27274263e-01 9.80461061e-01 -9.35686678e-02 -9.08581853e-01
-3.12981814e-01 -5.03641367e-01 -3.05154949e-01 -2.35285312e-01
5.84883928e-01 2.07351208e-01 -4.32532758e-01 -1.49810731e-01
-3.20302427e-01 9.21171978e-02 -1.22739278e-01 -1.75439920e-02
1.08075166e+00 -8.63673072e-03 -3.68083954e-01 6.32401884e-01
5.65883279e-01 7.85065442e-02 -4.36699748e-01 -6.36202276e-01
3.78548324e-01 -2.73835450e-01 -1.15733825e-01 -1.27335715e+00
-4.23598439e-01 7.69475639e-01 -1.46217048e-01 1.60185337e-01
9.70568001e-01 3.17309022e-01 1.00487435e+00 1.06904459e+00
4.94953841e-01 -1.12267482e+00 3.61659527e-01 9.37821329e-01
8.36941957e-01 -1.19274223e+00 -4.69535977e-01 -3.88162404e-01
-6.86839461e-01 1.00518942e+00 9.69555199e-01 -6.04143925e-03
3.28570217e-01 -1.74992412e-01 1.76550835e-01 -3.81844193e-01
-8.83599699e-01 -3.77094746e-01 8.14157799e-02 5.14738977e-01
5.13755262e-01 1.92516789e-01 -3.39811146e-01 1.20016968e+00
-6.71628058e-01 2.20001325e-01 4.49408025e-01 1.07868171e+00
-7.59443879e-01 -8.40675414e-01 -2.58474052e-01 5.28524339e-01
-3.76822352e-02 -7.05518603e-01 -9.49141502e-01 3.92416209e-01
9.95727330e-02 1.27564359e+00 -5.20883501e-01 -2.82030880e-01
1.50864571e-01 6.17775023e-01 6.22954428e-01 -8.82064760e-01
-2.78202623e-01 -8.39709818e-01 2.21955121e-01 -4.34073448e-01
-4.88411307e-01 -1.98491186e-01 -1.72971618e+00 -2.71809459e-01
-4.52015370e-01 3.84481311e-01 -4.52924818e-02 1.40456152e+00
2.74630964e-01 3.70745689e-01 8.84575024e-02 1.85825571e-01
-9.74955618e-01 -1.15697610e+00 -2.50930965e-01 4.01966035e-01
1.35779843e-01 -5.89553595e-01 -4.74967003e-01 1.69430584e-01] | [10.105185508728027, 8.010886192321777] |
20956aa8-4983-47d3-8a83-4638a50f8524 | hst-mrf-heterogeneous-swin-transformer-with | 2304.04614 | null | https://arxiv.org/abs/2304.04614v1 | https://arxiv.org/pdf/2304.04614v1.pdf | HST-MRF: Heterogeneous Swin Transformer with Multi-Receptive Field for Medical Image Segmentation | The Transformer has been successfully used in medical image segmentation due to its excellent long-range modeling capabilities. However, patch segmentation is necessary when building a Transformer class model. This process may disrupt the tissue structure in medical images, resulting in the loss of relevant information. In this study, we proposed a Heterogeneous Swin Transformer with Multi-Receptive Field (HST-MRF) model based on U-shaped networks for medical image segmentation. The main purpose is to solve the problem of loss of structural information caused by patch segmentation using transformer by fusing patch information under different receptive fields. The heterogeneous Swin Transformer (HST) is the core module, which achieves the interaction of multi-receptive field patch information through heterogeneous attention and passes it to the next stage for progressive learning. We also designed a two-stage fusion module, multimodal bilinear pooling (MBP), to assist HST in further fusing multi-receptive field information and combining low-level and high-level semantic information for accurate localization of lesion regions. In addition, we developed adaptive patch embedding (APE) and soft channel attention (SCA) modules to retain more valuable information when acquiring patch embedding and filtering channel features, respectively, thereby improving model segmentation quality. We evaluated HST-MRF on multiple datasets for polyp and skin lesion segmentation tasks. Experimental results show that our proposed method outperforms state-of-the-art models and can achieve superior performance. Furthermore, we verified the effectiveness of each module and the benefits of multi-receptive field segmentation in reducing the loss of structural information through ablation experiments. | ['Jin Zhang', 'Hongfang Gong', 'Xiaofei Huang'] | 2023-04-10 | null | null | null | null | ['skin-lesion-segmentation', 'lesion-segmentation', 'long-range-modeling'] | ['medical', 'medical', 'natural-language-processing'] | [ 4.62262779e-01 2.06911445e-01 -2.06169412e-01 -1.02351226e-01
-1.02027977e+00 -2.84101039e-01 7.93580040e-02 1.25511080e-01
-3.07425499e-01 3.32719356e-01 2.89149851e-01 -4.61885370e-02
-1.94608957e-01 -9.31853652e-01 -6.78994596e-01 -1.04452133e+00
-1.84021033e-02 -1.15102999e-01 7.06644416e-01 -1.67461276e-01
1.29799426e-01 3.25418591e-01 -1.09314954e+00 9.09585893e-01
1.23266852e+00 1.17896545e+00 8.55921865e-01 2.21491754e-01
-2.52797395e-01 6.56409681e-01 -3.28417629e-01 -1.78294778e-01
5.86211644e-02 -1.81845695e-01 -9.63336051e-01 -5.18226065e-02
1.22660600e-01 -1.21088423e-01 -1.96696818e-01 1.16516531e+00
7.48529017e-01 -1.05294056e-01 4.99897838e-01 -6.92806661e-01
-6.30243003e-01 7.04499245e-01 -6.72709525e-01 3.05266738e-01
-6.57996684e-02 -7.69661646e-03 7.03218699e-01 -5.51729560e-01
3.20370018e-01 1.17193353e+00 8.13885093e-01 4.33676511e-01
-8.73807967e-01 -5.81854582e-01 2.63371825e-01 2.18747005e-01
-1.21894240e+00 9.30365771e-02 6.79919481e-01 -2.94094414e-01
1.00013363e+00 5.09650707e-01 7.63337433e-01 7.30345786e-01
6.51216209e-01 1.14045453e+00 1.02638173e+00 -1.53635368e-01
-2.78945774e-01 4.67293933e-02 2.98498213e-01 8.00324857e-01
-1.39872417e-01 -1.27649799e-01 -4.25978340e-02 3.98918875e-02
1.05177033e+00 2.91283429e-01 -4.33955222e-01 1.60908714e-01
-1.28683496e+00 5.95184505e-01 1.32350993e+00 8.36284399e-01
-4.87964779e-01 -5.58254384e-02 2.82419652e-01 -1.39370233e-01
4.68902022e-01 2.75864422e-01 -2.48787507e-01 6.12052560e-01
-7.86157608e-01 -2.01889694e-01 8.81512687e-02 5.45216322e-01
4.88382936e-01 -3.94011497e-01 -8.35925817e-01 1.08832729e+00
2.27726460e-01 1.60270423e-01 1.01309109e+00 -5.90836704e-01
3.18364203e-01 9.56997037e-01 -4.28428948e-01 -8.46282959e-01
-4.27652299e-01 -8.35044384e-01 -1.17800486e+00 -1.84780672e-01
1.33103400e-01 2.40829945e-01 -1.41207945e+00 1.54622066e+00
1.57615840e-01 4.16940004e-01 1.17581589e-02 9.60835338e-01
1.26197934e+00 5.86850345e-01 2.80489236e-01 -1.49441911e-02
1.49597371e+00 -1.22641325e+00 -6.17845535e-01 -2.54119754e-01
5.56243062e-01 -8.07368398e-01 1.02970707e+00 -9.45755187e-03
-1.08100104e+00 -6.91290081e-01 -9.35039282e-01 -1.38614044e-01
-4.11232322e-01 1.26373231e-01 6.43274963e-01 3.04371715e-01
-1.10123956e+00 5.04535079e-01 -8.87809396e-01 -1.40457585e-01
9.42319274e-01 6.60548806e-01 -3.05707574e-01 -1.60369590e-01
-1.37041175e+00 7.30969250e-01 2.34303489e-01 3.96366954e-01
-7.62539148e-01 -9.99810517e-01 -8.35299730e-01 9.68997106e-02
1.29035950e-01 -7.90765822e-01 9.87504423e-01 -8.78067970e-01
-1.20380032e+00 7.12860644e-01 -5.28002866e-02 -2.65945524e-01
1.90900832e-01 2.67770737e-01 -1.65514693e-01 3.82345259e-01
2.85496712e-01 1.07246399e+00 7.44248807e-01 -9.68730569e-01
-6.16279960e-01 -5.49977005e-01 -8.86602513e-03 3.89879555e-01
-3.69576812e-01 -3.01465482e-01 -4.71121788e-01 -8.63624752e-01
3.82228523e-01 -5.66237152e-01 -3.82762998e-01 -1.83157220e-01
-4.49117303e-01 -2.86675125e-01 8.58018696e-01 -1.04646611e+00
1.31739271e+00 -2.22110891e+00 1.98606640e-01 2.58054733e-01
3.23120564e-01 3.06446612e-01 -2.04229325e-01 -2.85598695e-01
-1.35177344e-01 3.88228893e-01 -4.74067062e-01 -5.22358492e-02
-3.53890926e-01 4.48794737e-02 1.56717017e-01 2.85309523e-01
1.85643613e-01 1.25636840e+00 -6.21644199e-01 -8.22356582e-01
4.47453767e-01 6.81404471e-01 -5.83533764e-01 2.81287264e-02
1.52722523e-01 6.12362683e-01 -7.18278229e-01 9.89056349e-01
7.68161833e-01 -3.27709228e-01 -3.14642161e-01 -7.71147251e-01
4.05662321e-02 -8.30116123e-02 -7.27010131e-01 1.86815155e+00
-6.03984594e-01 6.41026720e-02 3.13969970e-01 -1.13243294e+00
4.86953259e-01 3.16141993e-01 7.88211107e-01 -9.31793034e-01
3.93137693e-01 2.05686986e-01 2.25032896e-01 -7.56794870e-01
-3.14639546e-02 -2.45521396e-01 -8.75412747e-02 -1.29542887e-01
7.78327510e-02 1.26302466e-01 -1.12365507e-01 -8.49775225e-02
9.77903128e-01 -1.44633129e-01 -6.21627793e-02 -2.53253937e-01
7.24515676e-01 -1.76247433e-01 6.49008274e-01 5.28714061e-01
-2.75818378e-01 7.57005930e-01 2.62477428e-01 -1.43321082e-01
-4.65272069e-01 -9.97558773e-01 -3.98585737e-01 6.98097825e-01
5.00684202e-01 -8.17629993e-02 -8.98516297e-01 -8.10860157e-01
-1.39747664e-01 2.10163340e-01 -9.23085749e-01 -4.08595204e-01
-5.52614927e-01 -1.14478779e+00 4.46993530e-01 7.01283216e-01
9.75209415e-01 -1.24620676e+00 -5.52989542e-01 3.84057254e-01
-6.36767507e-01 -7.73301125e-01 -7.17614114e-01 1.27844647e-01
-8.44695270e-01 -9.19049680e-01 -1.17239320e+00 -1.16479874e+00
8.85925293e-01 2.41766289e-01 5.91348946e-01 1.69334128e-01
-6.25227749e-01 -3.06514129e-02 -2.85586417e-01 -1.87596664e-01
-1.33478805e-01 1.93908289e-01 -7.76080072e-01 1.19940139e-01
-7.07492232e-02 -3.28010380e-01 -8.95741522e-01 4.72138077e-01
-1.16112006e+00 2.03943163e-01 1.15788364e+00 1.16279495e+00
9.01069760e-01 3.39883626e-01 4.28417653e-01 -7.40132809e-01
4.39887524e-01 -2.49962509e-01 -2.11224154e-01 4.37030256e-01
-5.03220186e-02 -2.31689796e-01 4.14900422e-01 -3.65511775e-01
-1.20040143e+00 -1.82122558e-01 -4.44412798e-01 -2.72528201e-01
1.36378676e-01 4.78549123e-01 -2.65038311e-01 -4.45602894e-01
2.62021214e-01 3.19616973e-01 1.89665243e-01 -3.46917659e-01
-9.22571942e-02 7.61704385e-01 2.97248691e-01 -1.77959248e-01
2.35305786e-01 4.29713875e-01 -2.71554768e-01 -6.47727609e-01
-7.71794021e-01 -5.69860697e-01 -4.21426028e-01 -2.21383218e-02
1.26283944e+00 -8.35902333e-01 -5.66961348e-01 8.16234350e-01
-8.86321664e-01 -2.02558428e-01 -2.14651883e-01 3.97759974e-01
-3.16347957e-01 3.34624738e-01 -9.68536377e-01 -1.85326129e-01
-6.72591984e-01 -1.64946365e+00 1.20196903e+00 5.41674376e-01
2.81168997e-01 -9.32463527e-01 -3.80158067e-01 6.25118673e-01
4.90520418e-01 1.64405748e-01 1.07159495e+00 -2.17785850e-01
-6.64263248e-01 -1.39189273e-01 -4.55318213e-01 5.66457212e-01
1.73096552e-01 -5.65644443e-01 -9.05152261e-01 -4.00422513e-01
-6.46581575e-02 -1.04728518e-02 1.38151217e+00 1.10732043e+00
1.58589232e+00 1.91292372e-02 -7.95066714e-01 9.41466630e-01
1.46191227e+00 2.73234874e-01 9.14232850e-01 2.04015762e-01
9.09906149e-01 4.73930746e-01 5.43481767e-01 -1.63686723e-01
2.33385369e-01 3.99313033e-01 3.98039043e-01 -8.57058644e-01
-5.48519254e-01 -1.22659646e-01 6.44576252e-02 6.87744796e-01
1.81683190e-02 -5.37694618e-02 -6.69332445e-01 7.46587813e-01
-1.75764656e+00 -7.28160799e-01 1.92071944e-01 1.87275040e+00
7.54138887e-01 1.10713936e-01 -2.06558809e-01 6.21184111e-02
7.66248703e-01 -1.10532098e-01 -5.60331047e-01 -8.98583606e-02
-1.83393836e-01 3.09242040e-01 6.91496313e-01 3.83282185e-01
-1.33550334e+00 8.52808237e-01 5.11696005e+00 1.40290868e+00
-1.29701722e+00 4.26663667e-01 7.70726144e-01 3.46314788e-01
-2.80270576e-01 -3.96437287e-01 -6.96265161e-01 5.59136331e-01
1.56749532e-01 3.65789741e-01 1.52467340e-01 3.48348469e-01
-9.40188244e-02 -1.28559038e-01 -3.93825442e-01 7.40903914e-01
-2.75592599e-02 -1.33590662e+00 4.49538194e-02 -2.52166740e-03
5.48995256e-01 6.84613809e-02 2.51983821e-01 2.43836433e-01
-2.21011683e-01 -1.09543538e+00 3.15778404e-01 4.58748877e-01
8.18837345e-01 -5.82645416e-01 1.02864707e+00 1.67337224e-01
-1.28875816e+00 -3.88573050e-01 -4.39522564e-01 5.42989790e-01
-1.56167313e-01 5.51592469e-01 -7.70599246e-01 7.77045727e-01
9.19015527e-01 1.06066060e+00 -7.27834165e-01 1.40025151e+00
2.53311228e-02 4.22815263e-01 -4.34872895e-01 2.08291113e-01
4.76443529e-01 7.63745755e-02 4.53036726e-01 1.11589122e+00
2.64686137e-01 7.64787048e-02 3.09485167e-01 8.61032426e-01
1.38159007e-01 2.25683630e-01 -1.73856586e-01 3.04768860e-01
5.34017943e-02 1.50963950e+00 -1.03839052e+00 -3.02733690e-01
-2.94813991e-01 7.92105675e-01 -1.82990301e-02 3.69847596e-01
-8.53949606e-01 -4.54339564e-01 1.93495423e-01 2.49683231e-01
3.79799843e-01 5.41548967e-01 -3.50352228e-01 -1.03634298e+00
-8.90050491e-04 -5.65454483e-01 5.36508143e-01 -5.12713730e-01
-1.21127379e+00 7.85671711e-01 -1.60271466e-01 -1.18935800e+00
5.98379791e-01 -3.41957390e-01 -6.36527717e-01 1.00301969e+00
-1.90504587e+00 -1.72871625e+00 -3.70829850e-01 7.51794159e-01
4.73962337e-01 1.25082016e-01 6.61560655e-01 4.34094697e-01
-6.62736535e-01 7.59304404e-01 -2.29725614e-01 1.66495949e-01
5.59720039e-01 -1.16299772e+00 -2.50741750e-01 6.74290359e-01
-5.02998114e-01 4.12985533e-01 -1.10982873e-01 -7.71830320e-01
-8.17072153e-01 -1.39681983e+00 3.31206411e-01 -2.72156764e-02
1.75730228e-01 -1.02680884e-01 -9.79001105e-01 3.86804134e-01
8.77942964e-02 1.80224642e-01 6.74273551e-01 -3.32643062e-01
1.44836202e-01 -3.23745042e-01 -1.48278117e+00 2.99598932e-01
6.45260632e-01 -2.55263060e-01 -5.83254457e-01 1.70041770e-01
8.81246626e-01 -4.97402489e-01 -1.25158751e+00 9.25561428e-01
4.55613673e-01 -6.62370622e-01 1.21396887e+00 1.64473593e-01
2.67517716e-01 -4.18953896e-01 -1.71504356e-02 -1.27271807e+00
-7.22289681e-01 -1.07220612e-01 5.88482797e-01 1.01360846e+00
3.81512970e-01 -6.14274025e-01 6.27866328e-01 9.93391573e-02
-7.56325066e-01 -1.10390925e+00 -9.85107124e-01 -1.63124710e-01
2.74306297e-01 -9.59159061e-02 5.61891377e-01 6.38954163e-01
-1.03902735e-01 1.27666205e-01 6.12001531e-02 2.69040793e-01
5.81192970e-01 1.08679950e-01 -2.68793046e-01 -7.89068758e-01
-1.40299961e-01 -7.05869436e-01 -3.49838495e-01 -7.96947479e-01
4.96990979e-03 -1.40874088e+00 -3.89758497e-02 -1.91736186e+00
3.73626620e-01 -5.50383747e-01 -8.04351270e-01 7.43246853e-01
-3.70045513e-01 5.93029618e-01 -1.60129573e-02 1.15894876e-01
-4.65696990e-01 3.96651536e-01 2.15881705e+00 -5.37485719e-01
-2.21440569e-01 7.47745484e-02 -8.62097800e-01 5.87811828e-01
6.17852211e-01 -1.34108409e-01 -3.77830446e-01 -4.46625829e-01
-4.25999522e-01 4.82191332e-02 6.23047233e-01 -1.09143555e+00
2.42090851e-01 3.47222149e-01 8.87484372e-01 -6.96624160e-01
1.26205206e-01 -8.79203796e-01 -1.87794909e-01 6.58983827e-01
-2.31382608e-01 -5.84212184e-01 4.78902221e-01 2.87681520e-01
-4.86031681e-01 8.88614506e-02 1.04883969e+00 -4.37428087e-01
-6.93620801e-01 5.43069839e-01 -2.82478273e-01 -2.76605934e-01
1.18363309e+00 -3.29792500e-01 -1.79812029e-01 3.46936852e-01
-1.00662577e+00 4.84163940e-01 -1.32228285e-02 4.22814876e-01
7.87615359e-01 -1.11894000e+00 -6.13249600e-01 6.30235672e-01
1.04830986e-04 2.69414008e-01 9.62547779e-01 1.38881707e+00
-4.63729858e-01 3.76830220e-01 -3.19920152e-01 -9.25274432e-01
-1.18284667e+00 3.90507758e-01 8.42389524e-01 -6.43039644e-01
-6.79151118e-01 1.07052004e+00 7.60965168e-01 -4.15400684e-01
2.35676780e-01 -7.94905603e-01 -5.64606369e-01 -3.73456925e-02
4.67235029e-01 -9.63383075e-03 3.34292799e-01 -5.99699259e-01
-4.27139372e-01 1.04701173e+00 -4.12529320e-01 4.79356319e-01
1.04918015e+00 -6.88235089e-02 -4.49237108e-01 -2.44841054e-01
1.28161371e+00 -5.09240091e-01 -9.56680119e-01 -2.09488377e-01
-6.08599484e-01 -1.37534499e-01 6.35334849e-01 -1.26345038e+00
-1.38072586e+00 1.06309175e+00 1.02669358e+00 -4.97130826e-02
1.57704318e+00 1.40566751e-01 1.39271903e+00 -3.24444145e-01
1.76773891e-01 -7.21103430e-01 -4.32354845e-02 2.02668849e-02
7.48125315e-01 -1.05555713e+00 -4.13402855e-01 -7.92839706e-01
-4.63465542e-01 8.33781004e-01 7.72749901e-01 -9.38797444e-02
8.81830513e-01 4.04208839e-01 4.87088449e-02 -1.67648792e-01
-2.51030654e-01 -2.09007144e-01 5.56015968e-01 5.70734560e-01
2.73408532e-01 1.85087010e-01 -2.30966493e-01 1.02340090e+00
2.21058428e-01 9.94741917e-03 -3.24451327e-02 7.55316615e-01
-4.26057309e-01 -1.07641912e+00 -2.92696506e-01 7.67253339e-01
-7.10094213e-01 -2.26780713e-01 2.17136648e-02 4.84517723e-01
6.62582695e-01 7.27539539e-01 -4.21009474e-02 -4.41341937e-01
3.22673768e-01 -4.87979740e-01 6.26257420e-01 -5.39600730e-01
-1.06706774e+00 7.72341192e-01 -4.51320589e-01 -5.42993844e-01
-3.57502311e-01 -4.06580806e-01 -1.37495506e+00 3.18493605e-01
-3.71261179e-01 5.94322607e-02 3.16490531e-01 8.79719615e-01
1.82768911e-01 1.22005916e+00 5.31454265e-01 -6.06508970e-01
-1.77777752e-01 -1.00881493e+00 -5.50547898e-01 2.99805999e-01
3.81511241e-01 -6.44111454e-01 -1.14259988e-01 -1.95455715e-01] | [14.68004035949707, -2.685971260070801] |
e5a74832-ba67-4250-9a44-504b8c1b59b0 | identification-of-stormwater-control | 2305.18630 | null | https://arxiv.org/abs/2305.18630v1 | https://arxiv.org/pdf/2305.18630v1.pdf | Identification of stormwater control strategies and their associated uncertainties using Bayesian Optimization | Dynamic control is emerging as an effective methodology for operating stormwater systems under stress from rapidly evolving weather patterns. Informed by rainfall predictions and real-time sensor measurements, control assets in the stormwater network can be dynamically configured to tune the behavior of the stormwater network to reduce the risk of urban flooding, equalize flows to the water reclamation facilities, and protect the receiving water bodies. However, developing such control strategies requires significant human and computational resources, and a methodology does not yet exist for quantifying the risks associated with implementing these control strategies. To address these challenges, in this paper, we introduce a Bayesian Optimization-based approach for identifying stormwater control strategies and estimating the associated uncertainties. We evaluate the efficacy of this approach in identifying viable control strategies in a simulated environment on real-world inspired combined and separated stormwater networks. We demonstrate the computational efficiency of the proposed approach by comparing it against a Genetic algorithm. Furthermore, we extend the Bayesian Optimization-based approach to quantify the uncertainty associated with the identified control strategies and evaluate it on a synthetic stormwater network. To our knowledge, this is the first-ever stormwater control methodology that quantifies uncertainty associated with the identified control actions. This Bayesian optimization-based stormwater control methodology is an off-the-shelf control approach that can be applied to control any stormwater network as long we have access to the rainfall predictions, and there exists a model for simulating the behavior of the stormwater network. | ['Branko Kerkez', 'Abhiram Mullapudi'] | 2023-05-29 | null | null | null | null | ['bayesian-optimization'] | ['methodology'] | [ 1.64821267e-01 9.97756049e-03 1.38587326e-01 -3.96279842e-02
-3.10314790e-04 -6.91353559e-01 3.35552007e-01 6.42395318e-01
-1.16352044e-01 9.71538424e-01 -8.52186084e-02 -5.47700226e-01
-8.75076056e-01 -1.41835058e+00 -3.65757495e-01 -9.34906006e-01
-3.39874744e-01 4.36153889e-01 4.53861594e-01 -5.12523770e-01
3.88829321e-01 9.01717126e-01 -1.86014986e+00 -5.58357835e-01
9.55682456e-01 6.60945296e-01 2.23751396e-01 6.62277877e-01
-1.82751082e-02 2.10960791e-01 -5.23872197e-01 5.04723370e-01
3.43407124e-01 -2.56366521e-01 -1.86809316e-01 -2.47628585e-01
-1.57307416e-01 -3.36362422e-01 9.18100998e-02 6.84303701e-01
6.78937078e-01 4.06621575e-01 7.13042736e-01 -1.15493762e+00
4.60958272e-01 8.03897917e-01 -2.90583551e-01 3.03275973e-01
1.19923055e-01 3.21264297e-01 4.76745516e-01 -2.24103332e-01
1.63402915e-01 9.97968733e-01 2.92865098e-01 -8.99936855e-02
-1.14268935e+00 -9.85965669e-01 2.99597800e-01 -1.78272620e-01
-1.27153933e+00 -2.66002446e-01 1.98723674e-01 -5.43822408e-01
9.82428670e-01 2.18850568e-01 1.07981944e+00 2.71253943e-01
5.88298202e-01 5.53646199e-02 1.00670862e+00 -5.61568439e-01
8.10223341e-01 -3.49512517e-01 7.55495057e-02 1.20345980e-01
7.06498384e-01 7.39158869e-01 -4.10602808e-01 -2.88510799e-01
8.89764577e-02 -1.59015000e-01 -3.70972514e-01 -7.71635696e-02
-4.04545248e-01 7.55361795e-01 3.00121069e-01 2.30602428e-01
-5.86888850e-01 3.14435452e-01 -7.54401535e-02 5.24407551e-02
4.67180282e-01 5.97821712e-01 -2.12299019e-01 2.68832803e-01
-1.39904559e+00 5.26306272e-01 1.14652646e+00 4.77902353e-01
6.63483679e-01 4.66030389e-01 -1.84549719e-01 1.00893028e-01
6.54260278e-01 1.23363972e+00 -1.14729315e-01 -7.06089258e-01
2.00906143e-01 3.93595934e-01 5.97579420e-01 -1.00541890e+00
-6.13551557e-01 -2.21220747e-01 -5.26868463e-01 5.01917720e-01
2.60576904e-01 -7.95869052e-01 -9.67463493e-01 1.41687477e+00
4.62166369e-01 5.73461950e-01 1.26734093e-01 4.98243421e-01
5.63471019e-02 1.09864020e+00 1.87840879e-01 -5.65719366e-01
7.73029447e-01 1.66402519e-01 -8.41956258e-01 4.83912341e-02
1.30641147e-01 -6.07954204e-01 2.75300592e-01 1.38272569e-01
-5.90002298e-01 2.19635367e-01 -1.23666310e+00 1.37369561e+00
-6.32928193e-01 -4.30346936e-01 2.43302062e-01 9.45783675e-01
-7.62512028e-01 4.76844072e-01 -1.16686320e+00 -5.69532871e-01
-1.87186450e-01 2.94461817e-01 3.74129951e-01 2.04816923e-01
-1.56103265e+00 1.33464706e+00 3.74578625e-01 6.16080821e-01
-1.19500363e+00 -1.06794739e+00 -7.27529585e-01 4.24791962e-01
3.23863119e-01 -2.68237740e-01 8.09952497e-01 -2.50239819e-01
-1.77114320e+00 -2.79179186e-01 2.85517070e-02 -5.16929448e-01
4.25179899e-01 -5.76883182e-02 -2.29842886e-01 -9.42392722e-02
-6.82176948e-02 4.48776856e-02 6.33288622e-01 -1.33932436e+00
-8.52299690e-01 -2.49580100e-01 -5.21806106e-02 9.74497199e-02
-2.10066184e-01 -6.13634028e-02 6.34874940e-01 -5.17278492e-01
-3.91055159e-02 -1.02240443e+00 -6.76685870e-01 -1.99784189e-01
-1.18423738e-01 2.94629604e-01 7.17775822e-01 -1.84740767e-01
1.46432376e+00 -1.52425110e+00 9.63668991e-03 1.02672565e+00
-4.01279241e-01 3.48012060e-01 1.09785788e-01 9.53710973e-01
3.55247825e-01 1.49161831e-01 -7.30975866e-01 3.82728070e-01
5.68348765e-02 4.24267948e-01 -6.21332288e-01 4.63916779e-01
1.13929175e-01 -1.24441776e-02 -7.00812757e-01 -9.39380303e-02
5.91341615e-01 4.09477800e-01 -2.24721447e-01 5.70553601e-01
-4.68175262e-01 2.34120131e-01 -8.33417296e-01 4.88759637e-01
8.94978702e-01 6.15174234e-01 4.91763592e-01 2.64388114e-01
-6.86528921e-01 -6.15001917e-01 -1.76668549e+00 4.88975763e-01
-7.64761865e-01 2.84779340e-01 3.41358930e-01 -9.52538610e-01
1.27216411e+00 3.63618374e-01 5.53837895e-01 -1.04234047e-01
9.94983837e-02 3.61549646e-01 -3.95298079e-02 -7.39374518e-01
3.90894055e-01 -2.25934803e-01 -9.13520455e-02 6.27975404e-01
-4.06805009e-01 -7.83338726e-01 4.06244129e-01 -1.89871520e-01
9.62010086e-01 -2.43747756e-02 2.45981216e-01 -7.73333132e-01
4.13453579e-01 -1.22308932e-01 5.00743866e-01 7.94067979e-01
-1.15201948e-02 -5.21909446e-02 2.49578685e-01 -2.85842240e-01
-5.42897284e-01 -9.54310179e-01 -5.15342414e-01 8.10910881e-01
3.14676017e-01 2.07356498e-01 -4.13468987e-01 1.99203357e-01
4.53637660e-01 1.16250896e+00 -4.96176332e-01 5.21456264e-03
-4.73573238e-01 -1.39966452e+00 6.69797599e-01 8.74997005e-02
2.93682724e-01 -7.11949885e-01 -1.05913961e+00 8.76082599e-01
3.53925020e-01 -5.07425845e-01 2.37664983e-01 6.15972400e-01
-9.12221849e-01 -1.26599801e+00 -2.33249709e-01 -4.53359261e-02
4.14467722e-01 -4.24707942e-02 8.56867611e-01 1.08549021e-01
-2.72839129e-01 2.52108306e-01 -3.46406430e-01 -6.38716877e-01
-5.00067830e-01 1.61638543e-01 -1.21876143e-01 -1.43506289e-01
-1.89139470e-02 -6.82769358e-01 -5.65257192e-01 4.13678855e-01
-1.19934118e+00 -6.18844926e-01 9.89720374e-02 2.35770106e-01
3.72321278e-01 4.10494655e-01 9.60183620e-01 -5.46271503e-01
8.21166635e-01 -9.35225666e-01 -1.38137627e+00 6.83709025e-01
-7.92781174e-01 2.72395790e-01 3.40710372e-01 -4.66493070e-02
-1.12812066e+00 6.42944634e-01 2.11625621e-01 2.64815152e-01
-1.70863360e-01 7.77239680e-01 -1.22783542e-01 -1.79953322e-01
4.14053589e-01 -8.20458755e-02 -2.70134479e-01 -1.50545910e-01
2.24606872e-01 5.69446504e-01 1.53282583e-01 -7.41785645e-01
1.01751316e+00 4.23045516e-01 4.63363171e-01 -9.48022842e-01
-5.66086769e-01 -3.11838239e-01 -4.70028192e-01 -6.70119405e-01
6.54353738e-01 -6.48937285e-01 -7.19969690e-01 5.35642326e-01
-8.44392836e-01 -2.93057382e-01 4.10010032e-02 4.59744394e-01
-2.63065666e-01 -2.04550296e-01 2.64390469e-01 -1.37800348e+00
-4.21054304e-01 -1.06356716e+00 4.14443493e-01 4.15856600e-01
-1.34354271e-02 -9.97094452e-01 5.77388525e-01 -3.71171981e-01
9.21920836e-01 8.29915285e-01 6.33143067e-01 -4.30392623e-02
-7.51989365e-01 -1.00835867e-01 3.29843581e-01 -1.87185824e-01
1.87658936e-01 6.97219193e-01 -5.21304488e-01 -3.89785737e-01
-1.79632008e-01 1.78355053e-01 7.56978035e-01 6.09459579e-01
5.03231347e-01 -1.77380294e-01 -3.67351502e-01 4.30154055e-01
1.84189951e+00 4.07520443e-01 3.41890097e-01 5.72209060e-01
-5.43191470e-02 8.55540335e-01 7.54185617e-01 1.07163310e+00
1.28790855e-01 4.50594217e-01 9.20038640e-01 3.47963631e-01
4.09024447e-01 4.89520878e-02 3.70442897e-01 5.97152784e-02
1.93552211e-01 -7.87971258e-01 -1.29842830e+00 6.29332185e-01
-1.82555699e+00 -9.70366716e-01 -1.43250972e-01 2.19222164e+00
5.49620688e-01 -7.37265721e-02 -4.52164203e-01 1.45412624e-01
1.00254130e+00 1.89386696e-01 -4.60278153e-01 -5.08015811e-01
1.16747834e-01 5.55300117e-01 1.03248346e+00 8.33192468e-01
-8.61804843e-01 6.81785524e-01 6.25253057e+00 4.57859665e-01
-1.25503993e+00 -4.02449310e-01 3.84562351e-02 -1.30661532e-01
-4.60231841e-01 3.25663269e-01 -1.06196070e+00 3.44958454e-01
1.24619806e+00 -5.44081569e-01 3.18176180e-01 2.72118777e-01
1.11338687e+00 -7.12353826e-01 -4.91474390e-01 -1.73163638e-01
-4.00649607e-01 -1.30845022e+00 -9.25951153e-02 -1.72133163e-01
7.83193171e-01 -4.07747664e-02 -3.96099836e-01 -1.93137303e-01
9.51442897e-01 -8.88018906e-01 5.58368623e-01 7.78601229e-01
-9.63159502e-02 -9.94987845e-01 8.43185067e-01 2.72540510e-01
-1.34850848e+00 -3.26736569e-01 -8.71434137e-02 -1.62380695e-01
4.56727833e-01 8.76717865e-01 -6.32042527e-01 3.43063653e-01
6.70957625e-01 2.94697851e-01 -9.35123563e-02 1.43149507e+00
-5.50865173e-01 8.83632839e-01 -1.01917374e+00 -3.65167618e-01
3.48507106e-01 -1.67165399e-01 7.20785499e-01 8.62527847e-01
6.81893647e-01 3.57739568e-01 3.37982595e-01 6.01350427e-01
5.94351649e-01 1.56834692e-01 -6.05304420e-01 3.70557755e-01
1.01624072e+00 8.07392955e-01 -6.50764346e-01 3.38393934e-02
2.22273380e-01 -2.34821826e-01 -4.16539967e-01 4.22790706e-01
-7.24759459e-01 -7.01839387e-01 7.21371293e-01 2.25248545e-01
1.16596237e-01 -1.01631552e-01 -2.86257654e-01 -5.17302275e-01
-3.38524967e-01 -3.73106033e-01 3.28254133e-01 -5.22744238e-01
-9.76168811e-01 3.18748295e-01 6.70954585e-01 -9.86033797e-01
-3.66534472e-01 -3.27850804e-02 -9.33821857e-01 9.00844216e-01
-1.88961804e+00 -5.80686629e-01 -4.39875692e-01 2.17776582e-01
-1.59704372e-01 -1.81437377e-02 7.38168359e-01 -1.11330105e-02
-7.37358093e-01 -8.69475976e-02 6.99264705e-01 -5.39541125e-01
3.83872926e-01 -8.79582942e-01 -2.55965203e-01 1.14116633e+00
-7.50956416e-01 1.54927015e-01 1.17009485e+00 -8.48025680e-01
-1.23772216e+00 -1.28793132e+00 5.17538846e-01 3.23809654e-01
9.76088405e-01 -5.38742496e-03 -7.96493769e-01 1.41148701e-01
1.31255686e-01 -3.08091074e-01 4.52073246e-01 -3.18559706e-01
2.64139801e-01 -2.88740873e-01 -1.40322900e+00 3.58424515e-01
2.33953983e-01 1.17360577e-01 -5.01093507e-01 2.06507996e-01
4.49269742e-01 -1.42618820e-01 -9.99320090e-01 7.27886617e-01
5.28900325e-01 -5.64341247e-01 4.97044176e-01 -2.47072026e-01
-2.62396559e-02 -5.55858195e-01 -2.79938161e-01 -1.59026146e+00
2.05689177e-01 -6.21145129e-01 2.54857719e-01 1.31428158e+00
4.54665035e-01 -8.21058095e-01 6.73473775e-01 6.83494270e-01
1.14535742e-01 -4.71510649e-01 -1.21954060e+00 -6.06200695e-01
3.04188013e-01 -2.44801208e-01 9.06608701e-01 4.87459570e-01
-1.97190389e-01 -5.56778967e-01 -9.24912319e-02 1.08057809e+00
8.71060967e-01 3.86073321e-01 4.15425122e-01 -1.16447258e+00
8.40159506e-02 -1.81660712e-01 -4.31993194e-02 5.93271889e-02
3.07797253e-01 -3.43206465e-01 4.64908183e-01 -1.43114841e+00
-4.17216480e-01 -6.96653247e-01 -1.13899805e-01 5.84632754e-01
-4.49345112e-02 -4.89825785e-01 1.29391178e-01 -7.94717297e-02
3.36408854e-01 8.24046195e-01 6.03977561e-01 -6.04420416e-02
-6.68458581e-01 2.86553770e-01 -7.00639933e-02 5.45727074e-01
9.86319363e-01 -1.02726817e+00 -4.98439133e-01 -1.57153010e-01
4.02977705e-01 4.33890581e-01 2.01134197e-02 -1.13507414e+00
3.64237100e-01 -8.66949081e-01 -2.64289975e-01 -6.47006691e-01
-2.86391646e-01 -1.21153271e+00 6.49408817e-01 1.08294952e+00
-1.77972212e-01 -1.82549089e-01 4.34881955e-01 8.18900228e-01
-2.40423679e-01 -3.91419351e-01 1.02325833e+00 3.43546234e-02
-5.63073099e-01 1.53185919e-01 -1.17466283e+00 -4.02282104e-02
1.43887949e+00 5.34699634e-02 -4.35045302e-01 -2.83836693e-01
-6.16229117e-01 9.81373131e-01 2.94600148e-02 1.86159149e-01
4.38531339e-01 -4.21087682e-01 -8.58698905e-01 9.44033265e-02
-1.93728134e-01 -3.14117908e-01 3.89723331e-02 5.86278923e-03
-9.68224406e-01 1.66876823e-01 -3.82841021e-01 -3.86102408e-01
-6.53601766e-01 8.45741108e-02 9.08674002e-01 -2.60573208e-01
-1.11302733e-01 2.46237278e-01 -4.66379225e-01 -3.72087985e-01
5.49266264e-02 -3.76616657e-01 -3.31232160e-01 6.27261817e-01
3.76757473e-01 5.49249649e-01 1.99584723e-01 -1.87741742e-01
-4.32743937e-01 7.89986670e-01 7.30607986e-01 -4.05045658e-01
1.69857824e+00 -1.94926232e-01 -2.49664515e-01 2.54251268e-02
2.77831942e-01 -2.27417931e-01 -1.24260223e+00 2.44561598e-01
4.61583920e-02 -4.36591953e-01 6.24313831e-01 -6.45957649e-01
-1.30729139e+00 4.94133025e-01 8.83094132e-01 1.81355208e-01
1.17298377e+00 -8.19918454e-01 3.21292192e-01 7.40275621e-01
2.81546801e-01 -1.09376526e+00 -4.83513415e-01 3.65090579e-01
8.08263183e-01 -7.90780842e-01 2.95717388e-01 -2.73952186e-01
-2.35093813e-02 1.28299105e+00 4.15698051e-01 -3.46801817e-01
1.48517203e+00 8.91961515e-01 4.51257788e-02 -9.14571285e-02
-7.09559500e-01 -3.04757863e-01 -5.79906940e-01 4.48584557e-01
-1.70591295e-01 3.39430183e-01 -3.85178536e-01 6.96563870e-02
9.38178822e-02 6.16612434e-02 9.51325893e-01 1.15804303e+00
-1.03649628e+00 -1.05767715e+00 -8.40990067e-01 4.08200562e-01
-5.35240993e-02 -1.44392535e-01 1.24980718e-01 6.88816369e-01
6.19406588e-02 1.42408085e+00 -1.19045898e-01 1.06192335e-01
5.90183616e-01 -1.49649918e-01 -1.94344327e-01 -4.72793639e-01
-9.09595788e-01 -3.25308174e-01 7.77088664e-03 -2.13955387e-01
-5.22280931e-01 -7.09441960e-01 -1.24764621e+00 -3.45532417e-01
-5.04563391e-01 5.21180987e-01 9.38134074e-01 9.04685140e-01
1.39698341e-01 3.21855485e-01 9.87266362e-01 -9.58938360e-01
-7.44554102e-01 -6.06561422e-01 -8.57766509e-01 -5.10208547e-01
7.92386085e-02 -9.72390532e-01 -8.12610090e-01 -3.97987872e-01] | [5.5793843269348145, 2.630479097366333] |
a4c144ff-74a8-4a3e-a089-9df36df51488 | video-moment-retrieval-with-text-query | 2106.13566 | null | https://arxiv.org/abs/2106.13566v1 | https://arxiv.org/pdf/2106.13566v1.pdf | Video Moment Retrieval with Text Query Considering Many-to-Many Correspondence Using Potentially Relevant Pair | In this paper we undertake the task of text-based video moment retrieval from a corpus of videos. To train the model, text-moment paired datasets were used to learn the correct correspondences. In typical training methods, ground-truth text-moment pairs are used as positive pairs, whereas other pairs are regarded as negative pairs. However, aside from the ground-truth pairs, some text-moment pairs should be regarded as positive. In this case, one text annotation can be positive for many video moments. Conversely, one video moment can be corresponded to many text annotations. Thus, there are many-to-many correspondences between the text annotations and video moments. Based on these correspondences, we can form potentially relevant pairs, which are not given as ground truth yet are not negative; effectively incorporating such relevant pairs into training can improve the retrieval performance. The text query should describe what is happening in a video moment. Hence, different video moments annotated with similar texts, which contain a similar action, are likely to hold the similar action, thus these pairs can be considered as potentially relevant pairs. In this paper, we propose a novel training method that takes advantage of potentially relevant pairs, which are detected based on linguistic analysis about text annotation. Experiments on two benchmark datasets revealed that our method improves the retrieval performance both quantitatively and qualitatively. | ['Tatsuya Harada', 'Yusuke Mukuta', 'Sho Maeoki'] | 2021-06-25 | null | null | null | null | ['moment-retrieval', 'text-annotation'] | ['computer-vision', 'natural-language-processing'] | [ 2.17829108e-01 -3.57737422e-01 -4.55261648e-01 -4.72909957e-01
-1.00530958e+00 -6.23508334e-01 7.56771266e-01 3.58014435e-01
-3.94927979e-01 6.47352338e-01 3.40734184e-01 2.57793069e-01
1.16837295e-02 -6.72520697e-01 -7.01017439e-01 -6.60748541e-01
1.53251141e-01 4.28005636e-01 3.92010301e-01 -1.12594679e-01
3.30119818e-01 -1.43936515e-01 -1.49073684e+00 6.43899918e-01
3.72720182e-01 1.12936676e+00 2.88198143e-01 1.15018405e-01
-2.74679601e-01 1.00694752e+00 -6.85313642e-01 -4.93591011e-01
1.59208179e-01 -4.90865260e-01 -7.19916523e-01 1.11744210e-01
5.11792958e-01 -5.01628458e-01 -2.81422138e-01 1.26540244e+00
2.41125777e-01 3.79863292e-01 6.09453142e-01 -1.40498948e+00
-3.63503993e-01 5.83653212e-01 -6.53376102e-01 1.90271452e-01
7.40327895e-01 -3.79638493e-01 1.37060535e+00 -1.11652064e+00
7.53775299e-01 1.15614378e+00 1.59697905e-01 3.44737887e-01
-4.10638899e-01 -7.96662509e-01 1.84670120e-01 3.10897827e-01
-1.49403012e+00 -3.14867765e-01 1.02801561e+00 -2.53764331e-01
4.31262314e-01 3.83452505e-01 9.87009168e-01 7.74700105e-01
4.00818102e-02 1.26029730e+00 4.98735934e-01 -4.28174078e-01
-1.95106521e-01 7.52100125e-02 -1.69826657e-01 5.34841895e-01
-1.78154245e-01 -3.78854930e-01 -5.15087962e-01 -3.78145720e-03
8.28259110e-01 3.27978134e-01 -4.68821645e-01 -2.58461297e-01
-1.76865077e+00 6.64603531e-01 2.44882628e-01 5.99874377e-01
-2.53452510e-01 -8.62561688e-02 7.80085146e-01 3.65353405e-01
3.66684645e-01 1.66307837e-01 -2.34990850e-01 -5.06488755e-02
-7.37176061e-01 2.37375766e-01 6.15348160e-01 1.01577854e+00
1.12287343e+00 -6.91790640e-01 -4.12470609e-01 7.94087589e-01
3.71219784e-01 5.38092792e-01 6.85258687e-01 -8.09949040e-01
8.59059930e-01 9.09187853e-01 2.66443580e-01 -1.82916415e+00
1.11105010e-01 1.69244319e-01 -7.04854548e-01 -9.63063657e-01
1.34605214e-01 1.65401146e-01 -5.76629758e-01 1.57892251e+00
3.12230617e-01 4.91616398e-01 1.91992685e-01 1.23676860e+00
1.11797798e+00 1.01184475e+00 -5.30375466e-02 -7.27519095e-01
1.22724807e+00 -9.07253027e-01 -9.48681235e-01 -1.34368211e-01
8.15546036e-01 -1.25337780e+00 9.86102581e-01 9.08490345e-02
-8.68082047e-01 -3.97386432e-01 -6.95847034e-01 1.54889524e-01
-2.57434011e-01 3.03861797e-01 1.34497821e-01 -9.99647528e-02
-6.99065387e-01 3.58299643e-01 -3.05453748e-01 -4.79505062e-01
-2.08090991e-01 1.46770999e-01 -3.93475980e-01 -1.22383229e-01
-1.60012913e+00 4.67202723e-01 7.33329475e-01 2.67965849e-02
-5.95178127e-01 8.72919802e-03 -6.91084206e-01 -2.24104255e-01
6.98448062e-01 -2.87783086e-01 1.29738915e+00 -1.17278254e+00
-6.97986603e-01 1.07387698e+00 -4.08117205e-01 -1.25869051e-01
5.19784451e-01 -1.46021515e-01 -7.42030680e-01 6.81757689e-01
4.92714673e-01 8.94454837e-01 8.02639186e-01 -1.16074920e+00
-9.50178802e-01 8.30620974e-02 2.94657141e-01 6.85037613e-01
-7.37044036e-01 2.70806789e-01 -1.08743834e+00 -8.13241243e-01
4.58169043e-01 -7.63983488e-01 2.20846996e-01 1.13252050e-03
-3.53846371e-01 -4.98950362e-01 1.05937159e+00 -3.24673980e-01
1.44015110e+00 -2.15577817e+00 -1.59543663e-01 2.32499033e-01
2.23106537e-02 4.80938889e-02 -1.40563652e-01 5.44494152e-01
8.85122688e-04 1.92052737e-01 1.92923576e-01 -9.97608807e-03
-1.92923814e-01 3.84420007e-01 -4.95900542e-01 1.46980271e-01
1.05146922e-01 7.33596265e-01 -1.39206481e+00 -1.12263167e+00
3.25778872e-01 1.08728454e-01 -8.49997327e-02 3.03775638e-01
-2.88732320e-01 3.84874851e-01 -8.04794788e-01 5.85471094e-01
2.78416812e-01 -1.85006425e-01 -1.28608076e-02 -6.89163804e-01
1.61397412e-01 -7.32854009e-02 -9.88370299e-01 1.32844138e+00
-1.51581421e-01 7.74453521e-01 -5.08502245e-01 -1.27483487e+00
9.01059389e-01 6.47047043e-01 8.36882830e-01 -7.77496696e-01
1.87251896e-01 3.15720111e-01 -4.04062599e-01 -9.35205698e-01
7.60451496e-01 -6.30971268e-02 -4.03958559e-02 4.99274701e-01
-2.31687948e-01 -1.45942971e-01 4.61010814e-01 2.30505183e-01
5.42813957e-01 4.64021973e-02 2.81748772e-01 1.39043704e-01
8.64566565e-01 -4.61399779e-02 8.12638700e-01 4.93267328e-01
-2.91260004e-01 6.08505845e-01 5.83453059e-01 -4.03363585e-01
-9.29353893e-01 -5.71996331e-01 1.51589215e-01 9.17218864e-01
7.18832016e-01 -9.05519307e-01 -4.12455231e-01 -7.92710721e-01
-5.39625466e-01 -7.48881884e-03 -4.94451761e-01 -1.08630151e-01
-5.97400963e-01 -1.51917458e-01 3.68837446e-01 2.45345786e-01
6.18246317e-01 -1.06221545e+00 -2.81334579e-01 6.09408505e-02
-8.73796225e-01 -1.20446885e+00 -8.01366687e-01 -4.39859867e-01
-6.32357717e-01 -1.41702569e+00 -9.89981055e-01 -1.19362462e+00
8.14460456e-01 9.02446151e-01 1.07552230e+00 4.46013063e-01
2.72398561e-01 4.61471111e-01 -8.67466331e-01 -1.73015162e-01
-3.65147531e-01 -4.98695880e-01 6.68254718e-02 2.13238478e-01
7.07740724e-01 3.85128567e-03 -7.36634314e-01 6.51089072e-01
-1.12928855e+00 2.12280318e-01 4.19679672e-01 9.15223420e-01
9.77324426e-01 1.65010139e-01 3.34903002e-01 -4.46404546e-01
4.23201770e-01 -5.70989072e-01 -1.15937993e-01 6.00074112e-01
-4.28612456e-02 -3.49543512e-01 6.69928432e-01 -7.90697694e-01
-5.66048920e-01 4.96585667e-02 2.71833628e-01 -9.58668172e-01
4.29323316e-02 9.66950357e-01 -1.12192355e-01 1.54626772e-01
2.96668500e-01 2.72288084e-01 -4.03169096e-01 4.96554524e-02
-8.97740647e-02 7.95865417e-01 3.72894883e-01 -6.36589229e-01
5.13625741e-01 4.08380032e-01 -1.19776405e-01 -8.54865968e-01
-1.08264291e+00 -1.02151763e+00 -4.52136129e-01 -6.22686744e-01
5.98030925e-01 -9.56037164e-01 -1.43248156e-01 2.82814950e-01
-1.12691593e+00 2.15860605e-01 1.22161150e-01 6.40032649e-01
-4.76747274e-01 8.04088116e-01 -3.90645534e-01 -5.55562019e-01
-2.21317530e-01 -1.13026571e+00 1.39483082e+00 2.87312537e-01
-1.93048060e-01 -1.00010765e+00 -2.07698792e-01 9.18562338e-02
-3.32657605e-01 -1.01713045e-03 7.03092396e-01 -1.01851344e+00
-3.19538325e-01 -5.34409583e-01 -9.48185176e-02 1.61426321e-01
4.10333633e-01 3.68457466e-01 -5.08079410e-01 -2.32213661e-01
3.79575454e-02 -2.81859964e-01 4.60416973e-01 1.61160976e-01
9.28897500e-01 -6.27736807e-01 -5.45507073e-01 -4.68714014e-02
1.02653754e+00 3.72629821e-01 5.38697422e-01 3.05424750e-01
7.08759725e-01 7.08433509e-01 1.54019368e+00 3.29263955e-01
2.01157391e-01 7.43850112e-01 2.57420391e-01 1.08894899e-01
4.15554881e-01 -4.76525843e-01 3.62126559e-01 1.14345467e+00
1.17877483e-01 -5.06201982e-01 -8.97082090e-01 6.96969509e-01
-2.27518606e+00 -1.19535339e+00 -2.61612713e-01 2.27941442e+00
9.34494317e-01 4.29344550e-02 -3.07327285e-02 2.24670321e-01
1.21645427e+00 2.83939570e-01 -3.60959560e-01 4.13252443e-01
-8.11656937e-02 -5.63433051e-01 -2.70450860e-01 -1.16389513e-01
-1.31124818e+00 8.20214450e-01 5.06762362e+00 1.06213224e+00
-1.11647916e+00 -7.48545676e-02 4.91383672e-01 2.27092020e-02
-3.05249482e-01 1.43659666e-01 -6.76799715e-01 6.47023439e-01
3.41527045e-01 -5.04898846e-01 -2.58878097e-02 7.68799543e-01
2.65402555e-01 -1.07565880e-01 -1.19950211e+00 1.31374192e+00
4.22493011e-01 -1.14685524e+00 6.51369274e-01 -4.30350780e-01
9.26239908e-01 -3.24889153e-01 -1.48956895e-01 2.58681148e-01
-3.70249540e-01 -4.06024456e-01 8.07651818e-01 4.35233146e-01
6.62119389e-01 -7.40810454e-01 1.06102479e+00 5.65816998e-01
-1.61976540e+00 2.19085768e-01 -5.48930228e-01 1.90262616e-01
1.40584648e-01 5.56227207e-01 -5.07492840e-01 7.31353641e-01
7.17359781e-01 1.46213317e+00 -3.65344584e-01 1.08896327e+00
-2.55597711e-01 1.42669827e-01 -2.85286754e-01 -1.89173684e-01
2.86396980e-01 -2.85329998e-01 5.15207708e-01 9.46969271e-01
5.20490825e-01 3.71202558e-01 6.02638423e-01 4.20498550e-01
-2.89231539e-01 5.68500400e-01 -7.53364921e-01 -1.91186503e-01
6.89861655e-01 1.22454739e+00 -8.46509635e-01 -7.14526772e-01
-7.22747087e-01 8.26652110e-01 -2.75124819e-03 3.85067999e-01
-8.34903955e-01 -2.89145321e-01 1.57485425e-01 -2.14179441e-01
-1.85849041e-01 3.10815543e-01 4.77963835e-01 -1.51264131e+00
3.69514555e-01 -6.47974133e-01 5.36874950e-01 -1.16746306e+00
-1.49021316e+00 4.07088488e-01 5.46703041e-02 -2.12340856e+00
-2.67763406e-01 -1.31584629e-01 -5.38031340e-01 3.07107627e-01
-1.27844632e+00 -1.01183510e+00 -4.88002181e-01 8.62270057e-01
5.55770516e-01 4.47469577e-02 5.15885532e-01 4.69597965e-01
-2.90464699e-01 3.20386112e-01 -7.00482503e-02 5.30109107e-01
1.13712394e+00 -7.20160604e-01 -2.63806283e-01 8.22811544e-01
3.57642561e-01 7.28079319e-01 4.61661458e-01 -8.24561000e-01
-1.08088028e+00 -1.19673073e+00 9.97489989e-01 -8.11123624e-02
8.95025313e-01 2.32226893e-01 -9.64068055e-01 7.61457980e-01
1.91291776e-02 -1.11175790e-01 7.20943332e-01 -4.17550445e-01
-3.13213170e-02 1.17947534e-02 -7.07278490e-01 7.42059648e-01
7.32832789e-01 -9.71762180e-01 -9.86495376e-01 8.61160338e-01
5.22416651e-01 -5.39287388e-01 -7.11784065e-01 5.09032845e-01
3.77653301e-01 -8.02471817e-01 8.03244174e-01 -2.33646408e-01
7.46867478e-01 -4.81901646e-01 -1.81391180e-01 -1.00952804e+00
3.42905939e-01 -3.47083479e-01 -2.06259295e-01 1.54710054e+00
-1.30544780e-02 -1.79627016e-01 4.20438498e-01 4.48910356e-01
-2.91492380e-02 -5.58270335e-01 -7.79230416e-01 -8.39826405e-01
-2.75535077e-01 -2.35740378e-01 5.65957606e-01 1.40778244e+00
4.10432696e-01 3.25993866e-01 -5.28870404e-01 -1.63692489e-01
7.60423616e-02 3.23689520e-01 6.71870410e-01 -1.15649390e+00
2.22289979e-01 -3.93747211e-01 -5.63140631e-01 -1.53730750e+00
3.49273175e-01 -7.05296516e-01 3.73438209e-01 -1.42170823e+00
6.93623185e-01 -2.76833773e-01 -2.58519232e-01 5.64127743e-01
-3.97629797e-01 5.13706863e-01 8.78136531e-02 8.45712662e-01
-9.36768115e-01 3.62347484e-01 1.41880834e+00 -2.53469884e-01
-4.06849384e-02 -2.03935787e-01 -1.52617574e-01 8.60195994e-01
5.59533060e-01 -3.87162060e-01 -4.53853756e-01 -3.56936991e-01
6.00638211e-01 3.62168282e-01 3.40704083e-01 -8.19863856e-01
3.11812431e-01 -3.93441349e-01 1.53575316e-01 -8.18814337e-01
4.22739476e-01 -1.10452545e+00 4.71596001e-03 2.02555820e-01
-4.57169056e-01 -1.05658516e-01 -2.56963968e-01 6.26267076e-01
-8.55820239e-01 -3.85831833e-01 2.52027988e-01 -1.44214153e-01
-1.09884644e+00 4.46556956e-01 -9.24961790e-02 3.54430676e-02
1.17720330e+00 -4.50310498e-01 -3.03618014e-01 -7.20018446e-01
-5.23266077e-01 4.89486516e-01 6.20390832e-01 5.83877802e-01
7.64793217e-01 -1.71203172e+00 -4.30596322e-01 -1.60178095e-01
5.48196733e-01 4.75651510e-02 5.49203753e-02 8.71800125e-01
-1.92898303e-01 4.44923192e-01 9.13571343e-02 -8.05582941e-01
-1.50831497e+00 6.08907819e-01 1.10576510e-01 -4.68434626e-03
-3.88832241e-01 5.40287435e-01 3.35528731e-01 2.39265338e-02
2.48038501e-01 -2.62208641e-01 -6.83514118e-01 4.73646939e-01
7.04074264e-01 -2.00117916e-01 -2.06760630e-01 -1.22514999e+00
-2.26276949e-01 1.06676924e+00 -6.59275353e-02 1.25725746e-01
8.14632356e-01 -3.36422682e-01 -3.87007415e-01 6.99091852e-01
1.47648287e+00 -5.46993725e-02 -8.29046130e-01 -4.70198482e-01
-5.72526194e-02 -8.14804852e-01 -1.54001132e-01 -2.31871426e-01
-1.27234054e+00 8.83614302e-01 2.01032370e-01 2.53825247e-01
1.27103937e+00 5.59227653e-02 9.99715388e-01 6.49845719e-01
4.88353878e-01 -1.11245847e+00 4.42881465e-01 5.14917135e-01
6.13296866e-01 -1.49352491e+00 -1.07327521e-01 -5.62446296e-01
-6.91420853e-01 1.25023973e+00 7.16869712e-01 1.59774035e-01
2.58360147e-01 -4.37968642e-01 -1.23155874e-03 -2.40632161e-01
-6.02445006e-01 -1.82440639e-01 5.43955922e-01 1.38667643e-01
4.30716783e-01 -1.94115222e-01 -5.48093319e-01 3.88045818e-01
1.72589853e-01 4.74407524e-02 2.95306027e-01 1.03251278e+00
-4.66965795e-01 -9.00664032e-01 -2.72954315e-01 5.57410836e-01
-4.50398713e-01 1.42974649e-02 -4.93913591e-01 6.13187730e-01
-1.16525017e-01 1.10618103e+00 2.69534349e-01 -5.55904508e-01
1.09855622e-01 -1.03402533e-01 1.37561783e-01 -5.42773247e-01
-3.31048489e-01 3.88798773e-01 -1.04332715e-01 -3.77173215e-01
-9.73410368e-01 -4.25389558e-01 -1.59115720e+00 -2.94172466e-01
-6.41185403e-01 5.42656779e-01 -2.38233469e-02 1.27777958e+00
-4.28817887e-03 9.60377604e-02 7.68726885e-01 -6.21252537e-01
-6.17016433e-03 -7.40181983e-01 -3.66061658e-01 8.54315162e-01
1.66700408e-01 -6.67189419e-01 -4.65030253e-01 4.19334143e-01] | [10.129617691040039, 0.7494707703590393] |
b9a20abc-0462-4f0b-87a5-52610738d74b | learning-large-margin-sparse-embeddings-for | 2307.04541 | null | https://arxiv.org/abs/2307.04541v1 | https://arxiv.org/pdf/2307.04541v1.pdf | Learning Large Margin Sparse Embeddings for Open Set Medical Diagnosis | Fueled by deep learning, computer-aided diagnosis achieves huge advances. However, out of controlled lab environments, algorithms could face multiple challenges. Open set recognition (OSR), as an important one, states that categories unseen in training could appear in testing. In medical fields, it could derive from incompletely collected training datasets and the constantly emerging new or rare diseases. OSR requires an algorithm to not only correctly classify known classes, but also recognize unknown classes and forward them to experts for further diagnosis. To tackle OSR, we assume that known classes could densely occupy small parts of the embedding space and the remaining sparse regions could be recognized as unknowns. Following it, we propose Open Margin Cosine Loss (OMCL) unifying two mechanisms. The former, called Margin Loss with Adaptive Scale (MLAS), introduces angular margin for reinforcing intra-class compactness and inter-class separability, together with an adaptive scaling factor to strengthen the generalization capacity. The latter, called Open-Space Suppression (OSS), opens the classifier by recognizing sparse embedding space as unknowns using proposed feature space descriptors. Besides, since medical OSR is still a nascent field, two publicly available benchmark datasets are proposed for comparison. Extensive ablation studies and feature visualization demonstrate the effectiveness of each design. Compared with state-of-the-art methods, MLAS achieves superior performances, measured by ACC, AUROC, and OSCR. | ['Jicong Zhang', 'Lu Xu', 'Mingyuan Liu'] | 2023-07-10 | null | null | null | null | ['medical-diagnosis', 'open-set-learning'] | ['medical', 'miscellaneous'] | [ 3.17594260e-01 3.10258597e-01 -3.55050951e-01 -1.79813817e-01
-5.48039079e-01 -3.25492144e-01 2.45609090e-01 3.15325022e-01
-1.49463147e-01 6.70290232e-01 8.55669379e-02 -9.51554254e-02
-3.66515905e-01 -6.75046086e-01 -3.55174989e-01 -9.41503465e-01
-6.20515551e-03 2.16751292e-01 1.18869700e-01 -7.35447556e-02
-5.02479151e-02 5.43955564e-01 -1.61408043e+00 2.83681333e-01
1.07552660e+00 1.20908058e+00 -7.40692317e-02 -4.78265658e-02
-2.29381755e-01 3.89624596e-01 -5.88495970e-01 -2.46709540e-01
2.18579158e-01 -7.72398189e-02 -4.62083548e-01 3.13916355e-02
1.43277034e-01 -4.26308475e-02 -5.84172845e-01 1.05950952e+00
5.06248355e-01 -1.64749652e-01 7.18482912e-01 -1.22148752e+00
-8.94632220e-01 3.28652740e-01 -4.84663427e-01 5.14421538e-02
2.45528594e-01 8.48976001e-02 8.48297656e-01 -1.16072333e+00
4.76185203e-01 8.92243087e-01 6.61549032e-01 5.22804439e-01
-9.52268243e-01 -7.08613813e-01 1.73370838e-01 1.46985620e-01
-1.67150509e+00 -1.93333805e-01 8.38890254e-01 -2.66020775e-01
4.91111904e-01 6.35834038e-01 5.63515544e-01 1.04072845e+00
2.52170116e-01 8.62828851e-01 9.60643828e-01 -2.73276150e-01
2.50046164e-01 4.49612051e-01 2.84149706e-01 6.44094050e-01
5.23636460e-01 2.02585116e-01 -3.86652231e-01 -2.43415132e-01
5.94070733e-01 5.36486566e-01 -7.21286178e-01 -4.32701737e-01
-1.34567702e+00 6.85312331e-01 7.82149494e-01 2.98043907e-01
-1.19122431e-01 -6.83524609e-01 3.27312946e-01 1.54881239e-01
3.62982571e-01 4.56329674e-01 -3.94133568e-01 2.97613293e-01
-8.00097823e-01 -1.36140585e-01 4.90537316e-01 8.17353964e-01
5.53660870e-01 -2.56005824e-01 -2.78333157e-01 1.06117594e+00
1.90082073e-01 3.34429443e-01 9.94068503e-01 -2.25538671e-01
3.28283250e-01 1.15153408e+00 -2.86714464e-01 -1.07006192e+00
-4.29341316e-01 -9.08565283e-01 -1.32795036e+00 -2.13148803e-01
3.17482710e-01 2.50445545e-01 -8.94764960e-01 1.45398819e+00
5.31115294e-01 4.11232740e-01 1.58215299e-01 8.26023996e-01
1.02393007e+00 2.50639051e-01 -2.37970605e-01 -3.09175611e-01
1.54767203e+00 -7.85112262e-01 -7.91415870e-01 -8.50103199e-02
7.01048195e-01 -4.55084950e-01 9.76383448e-01 4.47988659e-01
-3.67113769e-01 -4.16522324e-01 -1.20126820e+00 6.90856427e-02
-5.69301605e-01 4.24759418e-01 6.74867630e-01 5.05007327e-01
-6.50496304e-01 5.06348610e-01 -7.18081594e-01 -9.16105285e-02
9.29910421e-01 2.91545302e-01 -6.63364470e-01 -4.26753432e-01
-1.14656782e+00 4.96849447e-01 1.97113514e-01 2.03576788e-01
-5.50241232e-01 -8.24080110e-01 -9.73479569e-01 1.53988540e-01
3.02983642e-01 -5.19355416e-01 4.49582040e-01 -6.21877372e-01
-1.06476188e+00 7.59675562e-01 8.39989260e-02 -2.65536606e-01
5.36463678e-01 -1.80657148e-01 -6.05515718e-01 9.33957547e-02
1.11573227e-05 4.20401216e-01 9.31154788e-01 -9.98710811e-01
-3.21376175e-01 -8.01171601e-01 -2.28372023e-01 1.33069679e-01
-9.58353877e-01 -5.43366313e-01 -1.82528436e-01 -9.27718580e-01
7.18270600e-01 -9.33484674e-01 -2.16044739e-01 4.59960312e-01
-6.09478056e-01 -2.65995413e-01 1.02136135e+00 -4.42392051e-01
1.53546929e+00 -2.53484035e+00 7.20923115e-03 3.08851510e-01
5.94053268e-01 3.59815329e-01 -5.39935604e-02 -8.85869637e-02
-3.90674829e-01 -3.64654548e-02 -3.29564571e-01 4.83729504e-02
-2.59011596e-01 1.72013760e-01 -2.74608254e-01 6.13837421e-01
4.53695118e-01 7.37356305e-01 -8.57932270e-01 -4.47610825e-01
2.34154925e-01 5.04349649e-01 -4.15026337e-01 -9.00952145e-04
2.42684111e-01 2.09969133e-01 -4.06814724e-01 1.07660615e+00
7.82143474e-01 -5.35711765e-01 -4.89191562e-02 -3.55654240e-01
2.86945999e-01 -2.23779231e-01 -1.35979176e+00 1.49060309e+00
-2.05326423e-01 3.73147190e-01 -2.39273384e-01 -1.19748008e+00
1.15787137e+00 1.57812402e-01 4.50545102e-01 -3.70791465e-01
1.20332614e-01 4.72750306e-01 -9.46367532e-02 -4.92300808e-01
1.50864884e-01 2.33116485e-02 1.71879739e-01 -1.27552703e-01
2.34440956e-02 3.85535091e-01 -2.60204196e-01 7.89055601e-02
1.26395142e+00 -3.42099339e-01 5.45305312e-01 -2.14239821e-01
7.05433071e-01 -2.70553023e-01 7.47861981e-01 2.74827421e-01
-4.29488719e-01 6.85941935e-01 3.65669161e-01 -5.52554071e-01
-5.00388920e-01 -1.17335904e+00 -7.93916225e-01 4.73202467e-01
5.00895262e-01 -2.26713672e-01 -3.12420636e-01 -1.06723845e+00
2.36513317e-01 3.35417807e-01 -8.68694305e-01 -4.53094691e-01
-2.88128644e-01 -7.93803275e-01 3.54404539e-01 5.18333793e-01
2.15276882e-01 -6.96682930e-01 -3.97552222e-01 -5.29578030e-02
-7.62852058e-02 -9.92352664e-01 -4.78662759e-01 2.47175500e-01
-9.20415044e-01 -1.35356283e+00 -9.95329976e-01 -7.66095757e-01
1.07771242e+00 3.60248029e-01 7.49982059e-01 2.14995906e-01
-7.66683638e-01 -6.75194114e-02 -5.91790497e-01 -4.31473821e-01
-2.14704707e-01 -1.62392482e-01 2.50478148e-01 2.99024373e-01
5.04324436e-01 -3.77641916e-01 -8.21502030e-01 5.25536835e-01
-8.82455170e-01 -1.11223139e-01 8.32833886e-01 1.24438989e+00
7.83652902e-01 4.91772555e-02 6.53963864e-01 -7.77506590e-01
2.56452441e-01 -6.58549368e-01 -1.28884986e-01 3.58407468e-01
-6.36627316e-01 -6.24055648e-03 6.97563529e-01 -6.83340311e-01
-3.69127065e-01 4.87373881e-02 -1.19588925e-02 -6.77794695e-01
-2.27883846e-01 2.76319712e-01 -3.11161846e-01 -3.24373618e-02
7.11190403e-01 4.05513376e-01 2.50112653e-01 -3.44689548e-01
7.93705508e-02 1.01040947e+00 2.64648080e-01 -1.74199194e-01
7.63305902e-01 6.72104120e-01 -1.33770570e-01 -8.91270399e-01
-9.07290280e-01 -6.35399342e-01 -3.90904039e-01 4.92054038e-02
5.37667572e-01 -9.84116256e-01 -2.83663571e-01 2.50366747e-01
-7.02185750e-01 2.99357355e-01 -5.42363822e-01 5.16305268e-01
-3.15660201e-02 3.31849903e-01 -2.00338036e-01 -5.16509295e-01
-3.64669651e-01 -1.23270726e+00 1.26788807e+00 3.69837046e-01
-1.76821113e-01 -8.52439642e-01 -2.00297788e-01 3.85508180e-01
3.17214906e-01 3.28100413e-01 8.59036863e-01 -1.14270699e+00
-1.89829022e-01 -5.88416576e-01 -2.12439492e-01 4.22174722e-01
5.04087329e-01 -2.43987471e-01 -1.01173973e+00 -5.00608027e-01
-2.34967005e-02 -2.96581894e-01 8.25359404e-01 2.32162818e-01
1.58918107e+00 -2.76284367e-01 -7.53667176e-01 7.19576180e-01
1.08861721e+00 9.93348472e-03 6.31424785e-01 2.39816725e-01
5.83164752e-01 4.84600157e-01 6.22024655e-01 5.40521502e-01
1.44260377e-01 4.87668663e-01 3.40413302e-01 -2.10798234e-01
-1.24052100e-01 -1.40643626e-01 2.72751618e-02 7.83188820e-01
3.83979887e-01 8.17401856e-02 -7.44001031e-01 3.75626266e-01
-1.49481249e+00 -6.08209789e-01 6.75717965e-02 2.18684101e+00
7.31631994e-01 2.54117996e-01 -3.20627600e-01 7.29165673e-01
6.07115090e-01 4.10953052e-02 -8.41244876e-01 3.59103054e-01
-3.54453743e-01 7.62628242e-02 2.29191974e-01 -2.77505014e-02
-1.10626376e+00 2.86526024e-01 4.77753448e+00 1.13680053e+00
-1.35658002e+00 2.79332437e-02 8.43661666e-01 -1.42139588e-02
-2.88498402e-01 -2.92859405e-01 -8.58905494e-01 5.59191108e-01
3.72807264e-01 6.54888451e-02 -4.20241943e-03 1.00248504e+00
-1.72201008e-01 1.70691893e-01 -1.31453049e+00 1.25437665e+00
3.22436750e-01 -1.33804703e+00 2.07814276e-01 1.08350463e-01
5.15816212e-01 -2.37018391e-01 1.75893158e-01 4.71718729e-01
-3.79582763e-01 -1.00788295e+00 2.73256272e-01 3.21660817e-01
1.08552766e+00 -4.51555312e-01 9.72953737e-01 3.84187996e-01
-1.18003750e+00 -4.36379761e-01 -3.60551268e-01 3.16106647e-01
-3.43897849e-01 8.81551325e-01 -6.34327412e-01 6.84218109e-01
5.25409222e-01 9.10474658e-01 -7.33384371e-01 1.06911182e+00
4.05942556e-03 4.51261610e-01 -3.15138102e-01 8.87543857e-02
-4.59810980e-02 -6.27315864e-02 5.77179372e-01 7.95048058e-01
3.43495637e-01 2.39278778e-01 1.70663387e-01 7.72523820e-01
-1.53145054e-02 1.58768699e-01 -5.54233015e-01 -5.89585630e-03
6.64530098e-01 1.22567904e+00 -6.64285183e-01 -1.20782852e-01
-2.25791395e-01 9.34097409e-01 1.10272951e-01 1.20018803e-01
-7.04566181e-01 -6.18458152e-01 6.19207740e-01 3.03070366e-01
3.13787945e-02 2.58998275e-01 -3.98295313e-01 -1.40279591e+00
3.38218212e-01 -9.42758441e-01 6.58093274e-01 -2.55559415e-01
-1.37772155e+00 6.64068282e-01 -4.31749195e-01 -1.83035934e+00
5.71048737e-01 -7.30543852e-01 -3.46629709e-01 3.95426810e-01
-1.66210806e+00 -1.00339329e+00 -6.16769850e-01 4.04659361e-01
7.20716119e-01 -4.33050394e-01 8.91358793e-01 5.31356931e-01
-8.36264849e-01 1.10804331e+00 2.83518642e-01 2.27710620e-01
6.48616791e-01 -9.49035645e-01 -2.17224091e-01 4.64323133e-01
2.23864332e-01 6.52104259e-01 3.74188930e-01 -3.61363649e-01
-1.34217131e+00 -1.21427357e+00 5.14237404e-01 -3.53075624e-01
5.34641922e-01 -3.38870794e-01 -1.33072829e+00 2.24091426e-01
-6.76238596e-01 8.13461840e-01 9.58594680e-01 -2.23817751e-02
-5.34931421e-01 -3.03616941e-01 -1.19495451e+00 3.11380446e-01
9.03618634e-01 -3.88567299e-01 -4.59486842e-01 5.71176410e-01
7.84329593e-01 -5.07460356e-01 -1.03179276e+00 7.62478352e-01
5.03936529e-01 -7.35623181e-01 9.88518119e-01 -5.35166442e-01
3.93176943e-01 -3.19441468e-01 -3.30308229e-01 -1.20806050e+00
-2.06331208e-01 -2.45635286e-01 -2.92728454e-01 1.01920235e+00
3.57289523e-01 -8.14122200e-01 8.49295557e-01 2.52096921e-01
-3.13293815e-01 -1.45732415e+00 -1.01987350e+00 -8.28151882e-01
1.37252947e-02 -7.75023252e-02 6.68025613e-01 1.25739205e+00
7.25970343e-02 1.75055154e-02 -4.91139665e-02 3.80179822e-01
4.85265225e-01 1.62952375e-02 3.45506459e-01 -1.40233207e+00
-1.65915444e-01 -3.15588295e-01 -7.34266579e-01 -7.47812867e-01
-7.61765707e-03 -1.03321385e+00 -2.32421771e-01 -1.27677310e+00
1.97423175e-01 -9.25117135e-01 -7.27967143e-01 5.28037429e-01
-3.96967441e-01 1.97169334e-01 -4.13405560e-02 4.33748335e-01
-5.93504429e-01 8.64006817e-01 1.15237832e+00 -3.55959386e-01
-2.40857154e-01 -6.61361031e-04 -8.35833192e-01 7.62198806e-01
5.13447821e-01 -4.01743472e-01 -3.67026538e-01 -1.29389033e-01
-1.86480179e-01 5.47671430e-02 2.96541601e-01 -1.18927729e+00
8.00217986e-02 1.70320589e-02 5.03222823e-01 -4.78077859e-01
2.24999323e-01 -8.96041870e-01 4.65270691e-02 7.28771091e-01
-2.97622561e-01 -4.53403711e-01 1.05381556e-01 8.65946651e-01
-4.21429783e-01 5.12623116e-02 9.00805414e-01 2.28913739e-01
-4.76842701e-01 4.86866236e-01 2.65362203e-01 6.13395460e-02
1.35608017e+00 -5.35912573e-01 -3.36591542e-01 3.05258453e-01
-7.72825658e-01 2.82008290e-01 2.13866487e-01 6.50680661e-01
1.01495445e+00 -1.35046136e+00 -6.38293386e-01 6.94762230e-01
6.18479252e-01 2.73968667e-01 4.87759769e-01 1.00641942e+00
-2.74330765e-01 2.78306454e-01 1.51585877e-01 -8.07461023e-01
-1.28893542e+00 4.97536391e-01 1.69435069e-01 -2.84901917e-01
-7.48756826e-01 1.02596974e+00 3.81543070e-01 -5.07760644e-01
4.22878712e-01 -2.87497014e-01 -2.97540665e-01 7.22907409e-02
8.48114908e-01 1.67600513e-01 3.02892625e-01 -4.21872765e-01
-5.53967595e-01 4.21420127e-01 -2.88844496e-01 8.06278706e-01
1.32128811e+00 1.82473108e-01 1.06773965e-01 2.85027683e-01
1.32390630e+00 -1.64720714e-01 -9.69003618e-01 -4.52870935e-01
-1.93319380e-01 -5.59289813e-01 6.19516447e-02 -7.49292016e-01
-9.76641238e-01 8.08159530e-01 1.05541193e+00 5.54048046e-02
1.17886984e+00 4.93961163e-02 8.07337940e-01 2.77944744e-01
2.57795274e-01 -8.45685959e-01 3.52854908e-01 2.34656066e-01
8.83065104e-01 -1.54571915e+00 6.37446642e-02 -5.89241683e-01
-6.02341652e-01 1.13561261e+00 7.03652859e-01 9.91768390e-02
9.17668700e-01 2.32569605e-01 -7.95851499e-02 -1.20248675e-01
-6.01224542e-01 1.96251735e-01 3.76123130e-01 5.33214033e-01
2.21392229e-01 1.66405022e-01 -1.98943794e-01 1.02497208e+00
-9.55874696e-02 -2.27759108e-02 2.17035070e-01 7.52868295e-01
-4.24314529e-01 -7.05943942e-01 -4.54899967e-01 1.02180910e+00
-2.20976233e-01 7.59878010e-02 -1.03151821e-01 5.99689484e-01
3.85485709e-01 5.86201251e-01 5.73406816e-02 -5.62316477e-01
3.56945276e-01 -1.78410485e-01 4.73136157e-02 -5.73413789e-01
-1.09303094e-01 -2.71769345e-01 -4.39835072e-01 -4.64811414e-01
1.34715721e-01 -4.49983925e-01 -1.16833603e+00 2.72326171e-01
-8.46451461e-01 1.63072765e-01 4.26243007e-01 7.83653498e-01
4.92394120e-01 6.82979167e-01 8.54534149e-01 -3.98633868e-01
-1.04988348e+00 -7.93107688e-01 -6.79452002e-01 4.95955467e-01
5.14394283e-01 -9.51315165e-01 -6.29781663e-01 -2.05129892e-01] | [9.712736129760742, 2.9170374870300293] |
7961c356-911b-469d-a015-0718ae91a77d | cmt-interpretable-model-for-rapid-recognition | 2210.16584 | null | https://arxiv.org/abs/2210.16584v1 | https://arxiv.org/pdf/2210.16584v1.pdf | CMT: Interpretable Model for Rapid Recognition Pneumonia from Chest X-Ray Images by Fusing Low Complexity Multilevel Attention Mechanism | Chest imaging plays an essential role in diagnosing and predicting patients with COVID-19 with evidence of worsening respiratory status. Many deep learning-based diagnostic models for pneumonia have been developed to enable computer-aided diagnosis. However, the long training and inference time make them inflexible. In addition, the lack of interpretability reduces their credibility in clinical medical practice. This paper presents CMT, a model with interpretability and rapid recognition of pneumonia, especially COVID-19 positive. Multiple convolutional layers in CMT are first used to extract features in CXR images, and then Transformer is applied to calculate the possibility of each symptom. To improve the model's generalization performance and to address the problem of sparse medical image data, we propose Feature Fusion Augmentation (FFA), a plug-and-play method for image augmentation. It fuses the features of the two images to varying degrees to produce a new image that does not deviate from the original distribution. Furthermore, to reduce the computational complexity and accelerate the convergence, we propose Multilevel Multi-Head Self-Attention (MMSA), which computes attention on different levels to establish the relationship between global and local features. It significantly improves the model performance while substantially reducing its training and inference time. Experimental results on the largest COVID-19 dataset show the proposed CMT has state-of-the-art performance. The effectiveness of FFA and MMSA is demonstrated in the ablation experiments. In addition, the weights and feature activation maps of the model inference process are visualized to show the CMT's interpretability. | ['Chenyang Xue', 'Mengxing Huang', 'Guanjun Wang', 'Sufen Ren', 'Shengchao Chen'] | 2022-10-29 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 1.62896484e-01 -2.53061652e-01 4.41141659e-03 -3.88310373e-01
-5.70551097e-01 -5.29498272e-02 -6.89777359e-02 -4.03510183e-02
-2.40076199e-01 5.16698241e-01 2.64466196e-01 -4.29840773e-01
-4.01104003e-01 -6.04159355e-01 -4.27332520e-01 -8.93221736e-01
1.64537385e-01 5.95664382e-01 4.62279133e-02 3.36885095e-01
-2.29973063e-01 5.40860057e-01 -1.39173830e+00 7.69815385e-01
1.11420488e+00 1.08847129e+00 6.94858134e-01 7.01093137e-01
-4.36091498e-02 1.17657936e+00 -4.34257776e-01 -5.68542071e-02
-1.80585384e-02 -4.22735304e-01 -5.77882826e-01 1.01459891e-01
1.80075362e-01 -7.68780828e-01 -5.37650526e-01 7.16915548e-01
8.02207410e-01 -2.01457247e-01 8.25568497e-01 -1.12627578e+00
-6.07839465e-01 3.44446659e-01 -6.05401337e-01 6.64396226e-01
-1.36359716e-02 2.64352798e-01 6.12833261e-01 -1.04433811e+00
1.01183258e-01 1.14969444e+00 8.37508738e-01 5.12416661e-01
-5.97635329e-01 -7.47873068e-01 1.89135969e-02 4.73252416e-01
-1.32708490e+00 1.52345106e-01 3.89100909e-01 -4.79958922e-01
9.62092698e-01 4.45977092e-01 7.91291118e-01 9.57934260e-01
1.80271432e-01 7.95711637e-01 8.19260538e-01 -6.78845271e-02
-1.20922342e-01 -8.10178835e-03 6.04671538e-02 9.95588779e-01
3.40568125e-01 -1.31876424e-01 -2.92366356e-01 -2.06587866e-01
9.41012502e-01 6.88854814e-01 -5.80122113e-01 2.61937212e-02
-1.24124742e+00 8.54129195e-01 7.51201689e-01 3.20337385e-01
-6.46673381e-01 -8.96832570e-02 2.53090322e-01 -3.08799803e-01
1.68822944e-01 2.79385775e-01 -4.11319792e-01 2.11730853e-01
-6.94095016e-01 9.38080344e-03 2.76826441e-01 4.92859274e-01
1.81543499e-01 -3.29774767e-02 -5.79522431e-01 9.00388896e-01
2.38601059e-01 7.27132976e-01 8.65135193e-01 -8.26816738e-01
5.26306510e-01 8.34633172e-01 -2.31912136e-01 -8.97156656e-01
-6.26057446e-01 -6.70683324e-01 -1.22426879e+00 -1.06532313e-01
-4.23407517e-02 -1.94849655e-01 -1.08853638e+00 1.53634286e+00
2.47370020e-01 4.44701612e-01 -3.31156589e-02 8.71911228e-01
9.55480695e-01 6.20376945e-01 4.15102951e-02 -1.46580353e-01
1.54377460e+00 -1.00128222e+00 -7.70523429e-01 -2.00174525e-01
5.66964090e-01 -3.30879718e-01 9.16545391e-01 1.45150989e-01
-9.12439167e-01 -5.17201960e-01 -7.66606033e-01 -4.04492132e-02
-1.03013366e-01 2.98654437e-01 4.76885170e-01 2.08678469e-01
-8.11214685e-01 3.25490266e-01 -1.04586399e+00 -8.20769146e-02
8.84222031e-01 4.55645561e-01 -1.62054703e-01 -3.15972656e-01
-1.23283124e+00 9.51121688e-01 3.21513087e-01 7.23319575e-02
-7.04944253e-01 -8.26127529e-01 -7.39382863e-01 4.73623276e-01
8.83322731e-02 -1.31566656e+00 1.15605986e+00 -5.64001858e-01
-8.82956266e-01 5.01207888e-01 -3.82479161e-01 -3.11182082e-01
5.08559585e-01 -1.69032544e-01 -2.16555670e-01 4.67326999e-01
1.18380979e-01 6.28291488e-01 9.31413531e-01 -1.10519207e+00
-8.30910265e-01 -3.95317644e-01 -2.56412715e-01 3.88832062e-01
-4.65541482e-01 -9.02274847e-02 -4.72495824e-01 -6.91254795e-01
1.11836076e-01 -7.25931823e-01 -2.58379847e-01 2.35811602e-02
-2.79404789e-01 -1.77499101e-01 1.03117537e+00 -7.54827619e-01
1.32691383e+00 -2.26386213e+00 -1.52714118e-01 1.43845484e-01
6.44586027e-01 5.10800540e-01 2.22218353e-02 -1.94717064e-01
-1.85213238e-01 1.85197368e-01 -5.21683633e-01 -2.33522102e-01
-4.23560053e-01 4.55138683e-01 8.61290321e-02 1.80719256e-01
5.67475319e-01 1.03892255e+00 -6.84637249e-01 -7.56079912e-01
3.77694905e-01 8.54900956e-01 -8.02587450e-01 6.17816329e-01
1.68327034e-01 7.13566542e-01 -6.46715879e-01 4.56360191e-01
5.50518513e-01 -1.07817495e+00 5.72550707e-02 -2.94359952e-01
1.86893240e-01 8.21507722e-02 -8.28770399e-01 1.13177514e+00
-4.17539865e-01 1.96132779e-01 -1.50239944e-01 -8.80266368e-01
4.09483403e-01 5.61501563e-01 6.34983897e-01 -3.07872117e-01
3.96526009e-01 1.05106659e-01 2.45433524e-01 -1.07990289e+00
-1.32346302e-01 -1.32327944e-01 5.34577608e-01 3.19328249e-01
-3.34328324e-01 1.09614236e-02 -7.83878714e-02 9.21815261e-02
1.03773856e+00 -4.85304475e-01 3.57482553e-01 -3.08581181e-02
6.32791460e-01 -2.91305751e-01 5.46216190e-01 6.89226329e-01
-5.49664546e-04 9.75505531e-01 1.75988480e-01 -2.67196357e-01
-7.71978319e-01 -1.00143778e+00 -3.07348460e-01 5.56053877e-01
-1.09430343e-01 6.38198256e-02 -7.64412642e-01 -7.78347909e-01
6.48506507e-02 4.70122695e-01 -8.25411558e-01 -3.49950075e-01
-6.50645018e-01 -1.01449454e+00 2.72967041e-01 1.04432273e+00
6.85530961e-01 -1.16449177e+00 -8.74244452e-01 5.38633987e-02
-5.47746956e-01 -9.40127730e-01 -5.11302829e-01 1.22152418e-01
-8.99310410e-01 -1.18546140e+00 -8.89396727e-01 -9.32701468e-01
9.66161728e-01 3.12480032e-01 7.34737337e-01 4.31750000e-01
-5.02761006e-01 2.68676043e-01 -1.60904273e-01 -4.64594573e-01
-2.39930674e-01 -1.22721814e-01 -7.81999901e-02 4.81667854e-02
1.92541718e-01 -3.51349950e-01 -8.98620903e-01 1.27238482e-01
-1.06011951e+00 1.65994450e-01 7.98819900e-01 1.13022327e+00
7.31447816e-01 1.95113167e-01 6.84107244e-01 -9.04008150e-01
5.81486166e-01 -7.78283119e-01 -4.51955199e-03 1.22873992e-01
-6.27358496e-01 -8.88416469e-02 6.90375507e-01 -3.54676902e-01
-1.13453579e+00 -8.69265497e-02 -1.86983615e-01 -8.90666068e-01
1.14609012e-02 6.21668279e-01 7.72077367e-02 2.57831693e-01
3.15157384e-01 2.65929937e-01 1.98220059e-01 -4.81927782e-01
-8.19332600e-02 9.14874196e-01 5.97916842e-01 -1.07446685e-01
5.25263429e-01 3.29967111e-01 -1.09562226e-01 -4.25628453e-01
-1.08326006e+00 -5.35740852e-01 -4.29332912e-01 4.13441807e-02
1.15177023e+00 -9.95660126e-01 -6.63394392e-01 6.36872411e-01
-1.09852040e+00 -1.12003721e-02 -2.66458124e-01 6.89471126e-01
-2.79652804e-01 3.05309623e-01 -6.98706269e-01 -5.63320220e-01
-6.44794405e-01 -1.35242951e+00 8.13399434e-01 2.96020716e-01
-1.49936005e-01 -8.59620333e-01 -1.60336122e-01 5.60888827e-01
4.68107581e-01 2.60600019e-02 1.25508475e+00 -7.51541257e-01
-4.83593464e-01 -1.84528366e-01 -6.33449793e-01 5.53850234e-01
4.31402057e-01 -1.57800540e-01 -9.70794380e-01 -3.31675708e-01
3.71305496e-01 -1.52246237e-01 8.55758548e-01 6.67324901e-01
1.78131628e+00 -4.12765741e-01 -4.03605253e-01 7.38034248e-01
1.14592779e+00 5.42205393e-01 4.34908926e-01 -7.86487386e-03
9.67775643e-01 3.17848921e-01 2.58981138e-01 5.48381627e-01
5.10483503e-01 1.72248378e-01 5.45832694e-01 -5.02987981e-01
-3.70301664e-01 -5.72299212e-02 -1.23766512e-01 1.15993857e+00
-1.40346110e-01 -3.75315577e-01 -1.00132501e+00 5.24802148e-01
-1.75432515e+00 -8.63685250e-01 -9.06001478e-02 1.76873064e+00
8.13609183e-01 -1.94895923e-01 -3.01631778e-01 1.40283674e-01
8.42280030e-01 -1.40282556e-01 -7.36129999e-01 -3.00563779e-02
2.78600790e-02 3.76705438e-01 6.31388724e-02 2.23715186e-01
-1.01346266e+00 1.69987947e-01 6.16341734e+00 6.19483948e-01
-1.26835728e+00 2.57929683e-01 8.59465957e-01 -3.72747295e-02
-2.08182395e-01 -7.06740439e-01 -4.68913972e-01 6.63159847e-01
5.77253163e-01 2.23300412e-01 2.60639876e-01 7.69550264e-01
2.36981824e-01 1.23719908e-01 -8.95488560e-01 1.07492769e+00
1.10650100e-01 -1.27180529e+00 2.28665307e-01 -1.97987646e-01
6.41490221e-01 1.86439812e-01 2.58960992e-01 1.18215762e-01
-6.97152764e-02 -1.18719029e+00 3.07681888e-01 4.71878976e-01
1.01377535e+00 -6.43608868e-01 1.12356496e+00 3.08465391e-01
-1.17073512e+00 -3.87282878e-01 -1.91397458e-01 2.89724559e-01
2.80281086e-03 4.52291906e-01 -1.12168884e+00 4.74528879e-01
9.36922610e-01 6.07048810e-01 -5.84680438e-01 9.75888968e-01
-1.84552848e-01 5.94778955e-01 -3.75605941e-01 7.81711638e-02
1.03874058e-01 1.20693773e-01 3.29122394e-01 1.21049941e+00
5.14600635e-01 5.70991099e-01 1.71623617e-01 9.47312117e-01
-4.48601544e-02 -1.81813501e-02 -4.53469664e-01 1.67184547e-01
4.47003067e-01 1.05394912e+00 -3.68776649e-01 -6.99767649e-01
-4.56447601e-01 7.55825639e-01 2.15217441e-01 3.34165931e-01
-1.07448030e+00 -1.94376260e-01 3.31330687e-01 1.97663903e-01
4.33759451e-01 4.79811609e-01 -5.89606881e-01 -1.06156659e+00
-3.50249335e-02 -9.50343549e-01 5.59990764e-01 -8.96279573e-01
-1.18754613e+00 8.17129374e-01 -1.48272872e-01 -1.11658728e+00
-1.74774438e-01 -4.43787605e-01 -6.59249663e-01 8.81961286e-01
-1.57266521e+00 -8.83390903e-01 -7.27283776e-01 7.82627285e-01
5.13554454e-01 -4.46259193e-02 8.00400138e-01 5.22750437e-01
-7.37522840e-01 6.04141295e-01 2.64998386e-03 4.24461924e-02
2.83353686e-01 -9.60697532e-01 4.42912616e-02 6.90051675e-01
-4.64385986e-01 5.60011804e-01 1.55186832e-01 -6.11772418e-01
-8.08340728e-01 -1.50324976e+00 6.70577466e-01 -3.38226348e-01
1.71943963e-01 3.75353724e-01 -1.28915167e+00 4.96251762e-01
-1.58786476e-02 1.19286142e-01 6.55109227e-01 -5.30296564e-01
-2.04052150e-01 6.52621984e-02 -1.29382110e+00 3.61770868e-01
8.81415546e-01 -3.89241904e-01 -7.16487706e-01 3.26196283e-01
7.62308061e-01 -3.19005758e-01 -9.30350542e-01 9.53773320e-01
4.21988994e-01 -8.15134764e-01 1.08011770e+00 -5.70526242e-01
6.26327217e-01 -2.40629435e-01 1.91593921e-04 -1.02511966e+00
-7.47196496e-01 3.58712003e-02 -2.68239051e-01 7.46118486e-01
3.02216977e-01 -8.72742474e-01 6.01597071e-01 4.17098701e-01
-2.25596204e-01 -1.32536006e+00 -7.79785573e-01 -2.94614971e-01
-1.27916947e-01 -1.81443155e-01 6.55173719e-01 1.00723636e+00
-2.14349464e-01 3.37520331e-01 -1.29997566e-01 3.03722084e-01
2.63251364e-01 7.78490975e-02 7.38235638e-02 -1.09911120e+00
-2.82760561e-01 -4.11688805e-01 -2.91955732e-02 -7.84093499e-01
-2.53040284e-01 -1.00503898e+00 -4.74301204e-02 -1.94943035e+00
7.35993028e-01 -4.72286016e-01 -7.11171091e-01 6.70467973e-01
-7.86757529e-01 1.36333108e-01 9.11213160e-02 3.26080114e-01
-2.44446725e-01 3.80845189e-01 1.59848583e+00 -1.83716714e-01
3.40980769e-04 4.55941260e-02 -6.72319472e-01 8.57516170e-01
9.56097007e-01 -5.89684069e-01 -6.06202722e-01 -5.26738465e-01
-2.67609477e-01 1.66243762e-01 3.90279233e-01 -9.83237207e-01
4.20952886e-02 -4.55960147e-02 7.79352248e-01 -8.41987908e-01
3.51127654e-01 -9.06564474e-01 1.19727917e-01 9.12343800e-01
-1.30785897e-01 2.25389749e-01 1.57941446e-01 4.50839579e-01
-2.31022716e-01 -1.74660340e-01 9.83732939e-01 -2.03355819e-01
-1.72630966e-01 6.75206363e-01 -5.12870669e-01 6.05207160e-02
8.40697467e-01 -2.15563662e-02 -2.01299250e-01 -1.96184024e-01
-4.63575244e-01 2.89290220e-01 6.04685694e-02 1.07188582e-01
9.86316085e-01 -1.17713630e+00 -7.97446668e-01 5.06218016e-01
-4.65062037e-02 5.86483121e-01 5.95691204e-01 1.22977483e+00
-7.25524127e-01 4.65992659e-01 2.54461132e-02 -8.69057000e-01
-1.29283893e+00 6.60483837e-01 4.94001687e-01 -5.46392083e-01
-8.09272766e-01 1.03081524e+00 8.75224769e-01 -2.39797071e-01
3.24670702e-01 -6.16094410e-01 -3.83541107e-01 -2.05597341e-01
6.64872885e-01 2.28327408e-01 2.86092125e-02 -2.64276564e-01
-4.52719927e-01 5.19781888e-01 -2.82443523e-01 4.06037599e-01
1.25365841e+00 2.95498259e-02 -1.94587991e-01 5.56508265e-03
1.19995201e+00 -2.85116553e-01 -1.12203002e+00 -1.78051099e-01
-5.22009254e-01 -4.08729345e-01 9.52784047e-02 -9.91428077e-01
-1.44788802e+00 1.09208834e+00 9.28122699e-01 -4.68263403e-02
1.37187362e+00 -6.45720437e-02 1.05350947e+00 2.46725336e-01
-1.71694383e-01 -5.23493767e-01 2.41472572e-01 3.28736186e-01
8.81260395e-01 -1.22824419e+00 -1.34484410e-01 -3.16552937e-01
-8.02724659e-01 9.43012357e-01 6.51045263e-01 2.13726178e-01
6.41886294e-01 3.72185886e-01 1.71501130e-01 -2.72682637e-01
-6.36973739e-01 8.99404287e-02 3.14562529e-01 6.01388752e-01
1.42853007e-01 7.34843090e-02 -1.84302256e-02 9.75980461e-01
1.19472101e-01 8.78704637e-02 1.43020332e-01 7.53454149e-01
-5.27745783e-01 -4.95088547e-01 -4.38737959e-01 1.00885201e+00
-6.88851118e-01 -4.25912708e-01 -9.38090980e-02 5.94660044e-01
3.64315987e-01 7.83485234e-01 7.83964247e-02 -4.32633311e-01
1.88973665e-01 1.06226265e-01 3.56052279e-01 -5.44527113e-01
-6.02598667e-01 1.32055670e-01 -2.00576663e-01 -2.51204997e-01
-2.50420958e-01 -4.90080148e-01 -1.74817860e+00 -1.08184159e-01
-4.27515864e-01 8.78373012e-02 3.02519172e-01 9.18219745e-01
5.62218189e-01 1.17066777e+00 6.90508544e-01 -3.87336135e-01
-5.82139552e-01 -9.87824678e-01 -3.51607502e-01 3.83004844e-01
5.76514065e-01 -6.61047637e-01 -3.61462206e-01 1.50443643e-01] | [15.428973197937012, -1.8695292472839355] |
b85885c3-de17-4bac-994b-c9aeaddc5105 | learnable-graph-matching-a-practical-paradigm | 2303.15414 | null | https://arxiv.org/abs/2303.15414v1 | https://arxiv.org/pdf/2303.15414v1.pdf | Learnable Graph Matching: A Practical Paradigm for Data Association | Data association is at the core of many computer vision tasks, e.g., multiple object tracking, image matching, and point cloud registration. Existing methods usually solve the data association problem by network flow optimization, bipartite matching, or end-to-end learning directly. Despite their popularity, we find some defects of the current solutions: they mostly ignore the intra-view context information; besides, they either train deep association models in an end-to-end way and hardly utilize the advantage of optimization-based assignment methods, or only use an off-the-shelf neural network to extract features. In this paper, we propose a general learnable graph matching method to address these issues. Especially, we model the intra-view relationships as an undirected graph. Then data association turns into a general graph matching problem between graphs. Furthermore, to make optimization end-to-end differentiable, we relax the original graph matching problem into continuous quadratic programming and then incorporate training into a deep graph neural network with KKT conditions and implicit function theorem. In MOT task, our method achieves state-of-the-art performance on several MOT datasets. For image matching, our method outperforms state-of-the-art methods with half training data and iterations on a popular indoor dataset, ScanNet. Code will be available at https://github.com/jiaweihe1996/GMTracker. | ['Zhaoxiang Zhang', 'Naiyan Wang', 'Zehao Huang', 'JiaWei He'] | 2023-03-27 | null | null | null | null | ['point-cloud-registration', 'multiple-object-tracking', 'graph-matching'] | ['computer-vision', 'computer-vision', 'graphs'] | [-1.62032619e-01 -1.16365403e-01 -4.27017897e-01 -4.71137673e-01
-6.40826404e-01 -3.16352695e-01 6.27724379e-02 -1.90270729e-02
-4.09671664e-01 3.79531085e-01 -1.58786029e-01 -2.83894658e-01
-3.84984523e-01 -7.51407266e-01 -1.04286432e+00 -4.84277964e-01
-3.65496166e-02 7.06679523e-01 2.32031241e-01 -1.16872072e-01
-2.63053458e-02 4.41107213e-01 -1.19894707e+00 -2.24832371e-01
8.65149677e-01 1.09133565e+00 2.06955940e-01 4.82944757e-01
-1.64846689e-01 6.85026169e-01 2.10606586e-02 -4.02830124e-01
5.46347260e-01 1.06110133e-01 -6.81310713e-01 -2.52589080e-02
1.00713658e+00 -2.81303197e-01 -1.05369139e+00 1.28529131e+00
6.18394971e-01 1.75840572e-01 1.55357704e-01 -1.74915648e+00
-9.06172752e-01 2.89188296e-01 -8.01263750e-01 -8.31726640e-02
8.95031020e-02 1.03454955e-01 1.01747191e+00 -9.32238698e-01
4.10364568e-01 9.90232110e-01 8.58128607e-01 3.89415830e-01
-1.10652053e+00 -1.05811465e+00 2.78009266e-01 3.91110510e-01
-1.40754020e+00 -2.93439299e-01 9.40992355e-01 -4.63752091e-01
6.70307934e-01 1.30758256e-01 8.00518751e-01 4.11163568e-01
2.83232369e-02 7.67290354e-01 5.55584431e-01 1.01332583e-01
-2.56117642e-01 -4.26126570e-01 1.26070663e-01 1.10673392e+00
2.59961724e-01 1.93833485e-01 -2.25184157e-01 -1.33777142e-01
7.78826654e-01 4.14802134e-01 -2.50324517e-01 -7.78022349e-01
-1.34302402e+00 8.10817719e-01 1.02028859e+00 -3.14268321e-02
-2.28229687e-01 6.41845524e-01 2.60882586e-01 3.03682625e-01
2.99348116e-01 4.26919712e-03 -3.57615829e-01 4.35430884e-01
-6.44248784e-01 3.53745520e-01 6.87588096e-01 1.29316020e+00
1.25625956e+00 -2.10314631e-01 3.16872112e-02 6.80129528e-01
6.44892097e-01 6.50251329e-01 -3.76360351e-03 -9.32072401e-01
7.29718029e-01 6.61144555e-01 -1.76755771e-01 -1.36781919e+00
-4.98720616e-01 -4.05002773e-01 -1.03779876e+00 -3.52070145e-02
3.97589505e-01 -1.12761244e-01 -9.29107010e-01 1.74585485e+00
5.69703519e-01 5.32648683e-01 -2.92797297e-01 1.21895504e+00
1.18174005e+00 5.16509235e-01 -1.63036376e-01 1.62637889e-01
1.21484041e+00 -1.43934846e+00 -5.57274282e-01 -3.68113458e-01
6.30107284e-01 -6.89429879e-01 8.14143658e-01 -5.90961091e-02
-9.27413940e-01 -4.61468935e-01 -8.76901090e-01 -1.97798848e-01
-2.68373042e-01 -5.91681376e-02 8.84422719e-01 9.31452885e-02
-9.38903153e-01 4.50351864e-01 -7.44080424e-01 -2.51239449e-01
5.07270038e-01 8.71228576e-01 -4.74498481e-01 -3.24896932e-01
-9.94790316e-01 4.55241948e-01 2.76428968e-01 3.41551393e-01
-5.81042886e-01 -7.05091894e-01 -1.05063462e+00 -1.40757978e-01
6.28839076e-01 -1.22392547e+00 9.57054257e-01 -7.68083334e-01
-1.20043218e+00 9.13034439e-01 -9.64835286e-02 -1.43836856e-01
4.82108682e-01 -2.67338365e-01 -2.44430006e-01 -2.19818100e-01
2.33057857e-01 6.20050967e-01 4.80095983e-01 -1.04106402e+00
-5.39740801e-01 -3.55865270e-01 2.31461316e-01 7.41705447e-02
-1.29733831e-01 -1.14764281e-01 -1.06762147e+00 -5.09220839e-01
3.09683144e-01 -1.20992529e+00 -4.82213199e-01 4.65277702e-01
-4.26300198e-01 -3.31704766e-01 9.87617195e-01 -5.40385365e-01
1.03423595e+00 -2.05805564e+00 1.53332457e-01 2.68126696e-01
6.35201454e-01 1.16786540e-01 -3.10048640e-01 1.59785286e-01
-1.99467558e-02 -2.06173822e-01 -2.58524746e-01 -4.73541468e-01
2.00873792e-01 2.06402421e-01 1.46554872e-01 9.10732627e-01
-1.13535300e-01 1.20156491e+00 -1.02748799e+00 -6.42934918e-01
3.16467822e-01 4.39421088e-01 -6.36516273e-01 2.50121921e-01
-1.35820657e-01 5.18068671e-01 -4.98213232e-01 6.90415919e-01
8.98489714e-01 -6.39837265e-01 -1.40041068e-01 -5.39993286e-01
3.53613943e-02 3.64360549e-02 -1.44061911e+00 2.22502255e+00
-3.09115976e-01 6.07007980e-01 2.57156193e-01 -1.27047634e+00
7.90295660e-01 -8.34643543e-02 9.37742412e-01 -6.99261785e-01
5.69738485e-02 2.44079322e-01 4.03161906e-02 -4.26030844e-01
3.90733123e-01 3.25350910e-01 2.31511965e-02 8.11825618e-02
-7.52279535e-02 1.88728988e-01 1.11714015e-02 1.61104053e-01
9.77135360e-01 -5.53687988e-03 5.49012758e-02 -1.69974118e-01
4.60690558e-01 7.23207742e-02 8.22718740e-01 6.07097328e-01
-1.47981748e-01 6.58133268e-01 7.52777280e-03 -6.53533459e-01
-8.22208643e-01 -9.76513386e-01 5.68919517e-02 8.63050997e-01
7.95476496e-01 -3.86412382e-01 -3.75279665e-01 -7.05963552e-01
2.18440145e-01 -9.78251994e-02 -4.25928801e-01 -8.61961115e-03
-9.39959586e-01 -5.04768908e-01 3.05447400e-01 4.83967304e-01
6.58869147e-01 -8.60229015e-01 -1.10076405e-01 2.96977729e-01
-2.83271044e-01 -1.28518009e+00 -1.02150702e+00 -2.24931657e-01
-6.81597352e-01 -1.29644644e+00 -5.48911333e-01 -1.01267993e+00
8.53835762e-01 6.10851586e-01 1.21652114e+00 6.04074955e-01
-3.00870538e-01 3.27115864e-01 -1.75713450e-02 -1.98350325e-01
3.05177182e-01 2.79093206e-01 -8.14398974e-02 -1.10906865e-02
2.56105572e-01 -6.80238426e-01 -9.43550885e-01 6.05364501e-01
-6.33784413e-01 7.14421868e-02 6.00168049e-01 8.09867024e-01
8.83373678e-01 -3.61008555e-01 1.44665167e-01 -6.52513266e-01
1.55865207e-01 -4.10827219e-01 -9.53019917e-01 2.14781791e-01
-5.76213241e-01 7.73486942e-02 3.96194309e-01 -4.91973877e-01
-3.08541417e-01 4.38284546e-01 -1.02453999e-01 -9.88271892e-01
8.14193264e-02 6.06126487e-01 -2.65554816e-01 -7.75851011e-01
2.09314823e-01 9.17315707e-02 7.43557066e-02 -2.40805313e-01
3.82784665e-01 2.77100623e-01 6.28845096e-01 -5.30374467e-01
1.13902998e+00 4.51648057e-01 1.98830664e-01 -4.58317131e-01
-8.81932437e-01 -8.75168800e-01 -5.02162635e-01 -2.69188344e-01
8.53081286e-01 -1.03960252e+00 -1.13297856e+00 4.84549522e-01
-1.18941534e+00 -5.28420985e-01 1.31822243e-01 6.03253782e-01
-4.06043142e-01 4.64922935e-01 -4.43076134e-01 -2.83898205e-01
-5.22460103e-01 -1.23341858e+00 1.12909424e+00 4.23838109e-01
3.47655028e-01 -1.02490163e+00 1.10053279e-01 4.27344292e-01
3.94005507e-01 3.35617751e-01 4.58575726e-01 -3.12482804e-01
-1.13542998e+00 -7.82444701e-02 -6.46421134e-01 -1.45763159e-01
5.11850044e-02 -1.59719050e-01 -5.01214266e-01 -5.47846735e-01
-4.84413087e-01 -2.01067448e-01 9.38092709e-01 4.54903305e-01
1.41769755e+00 -3.92310828e-01 -6.19604647e-01 1.41654742e+00
1.60190392e+00 -1.06517173e-01 5.12346923e-01 4.05879647e-01
1.39754689e+00 4.53361183e-01 6.11053169e-01 3.50195020e-02
1.05307686e+00 8.83050084e-01 8.70103419e-01 -3.50765169e-01
-1.24190360e-01 -1.81248635e-01 1.60754975e-02 8.57989788e-01
6.02040142e-02 -1.12134330e-01 -1.01844716e+00 5.02391279e-01
-2.45712066e+00 -7.24391878e-01 -3.72637898e-01 2.16417766e+00
3.09208810e-01 -8.20345283e-02 1.82350695e-01 -6.07563198e-01
9.76674676e-01 1.86030433e-01 -7.42334425e-01 2.95166582e-01
9.61072296e-02 -2.03253269e-01 1.00771058e+00 4.81460899e-01
-1.37394583e+00 8.96885097e-01 4.86104298e+00 6.75231099e-01
-1.22398102e+00 1.22240052e-01 2.58107752e-01 3.07123959e-02
-2.08401725e-01 1.91766202e-01 -7.10840523e-01 4.51601475e-01
1.88283175e-01 3.68543155e-02 7.42034972e-01 7.98881412e-01
-1.91262141e-01 3.76474023e-01 -1.07078075e+00 1.46668959e+00
-4.04763222e-02 -1.45463109e+00 -2.80622631e-01 2.24505574e-01
5.57188928e-01 5.41055202e-01 -1.31556585e-01 2.47324258e-01
3.51146966e-01 -9.12186742e-01 5.07319987e-01 4.11412150e-01
7.02196240e-01 -5.10851920e-01 6.22367322e-01 2.36723557e-01
-1.79137957e+00 1.22252099e-01 -5.46476305e-01 3.06744240e-02
2.61534721e-01 5.52401125e-01 -2.82780766e-01 9.17898238e-01
8.02014649e-01 8.99448752e-01 -4.97512966e-01 1.54400325e+00
-8.03289376e-03 2.08212644e-01 -6.12419903e-01 8.09015185e-02
2.58431345e-01 -4.06860918e-01 6.18977726e-01 1.01748443e+00
2.92305678e-01 -2.78003439e-02 6.24127328e-01 8.59008670e-01
-3.91114235e-01 -7.45311677e-02 -6.89107776e-01 2.23383799e-01
4.72531587e-01 1.43094802e+00 -5.31510115e-01 -9.47827771e-02
-6.36685669e-01 8.04185092e-01 5.57441711e-01 3.38904589e-01
-1.01658869e+00 -5.02796471e-01 8.07175934e-01 -6.45230990e-03
3.31191927e-01 -3.09489101e-01 -9.42436159e-02 -1.28595078e+00
2.60009855e-01 -6.95688963e-01 4.34929490e-01 -5.48346639e-01
-1.50235391e+00 4.21638608e-01 -1.18246511e-01 -1.29741120e+00
5.05478717e-02 -4.68469471e-01 -8.46972585e-01 6.51215017e-01
-1.57382870e+00 -1.37508214e+00 -8.72067928e-01 9.23178017e-01
1.19737305e-01 7.09542073e-03 3.90079379e-01 9.45006669e-01
-5.86646378e-01 7.67006516e-01 -2.48542428e-02 6.71031117e-01
7.47958004e-01 -1.14417100e+00 6.37743235e-01 7.85838723e-01
1.56476364e-01 4.89836901e-01 3.32380325e-01 -5.73114812e-01
-1.73918152e+00 -1.45592570e+00 6.09969556e-01 -2.60559708e-01
7.77645588e-01 -4.67128336e-01 -9.13916171e-01 8.16577911e-01
9.86290872e-02 7.61181891e-01 3.17864239e-01 1.47994250e-01
-3.36718082e-01 -4.21762943e-01 -8.41143131e-01 4.68262792e-01
1.60416532e+00 -4.08053279e-01 -1.15051687e-01 4.98313636e-01
1.00880218e+00 -1.02598131e+00 -8.51626635e-01 6.04775310e-01
5.46087682e-01 -5.42437971e-01 1.20217943e+00 -6.15975618e-01
9.13970396e-02 -7.11015165e-01 -1.58990264e-01 -1.07472348e+00
-5.54849207e-01 -7.06637144e-01 -1.77852526e-01 1.10770988e+00
3.71343642e-01 -8.77128780e-01 1.06531191e+00 5.66214979e-01
-3.69886100e-01 -9.75333869e-01 -8.31816256e-01 -7.04102337e-01
-2.81352729e-01 -2.45242476e-01 7.80340493e-01 1.30693853e+00
-5.31149030e-01 3.97019684e-01 -5.02159834e-01 4.72620904e-01
1.07935202e+00 4.81104046e-01 1.28408003e+00 -1.10242760e+00
-4.00527567e-01 -4.08879012e-01 -6.01099789e-01 -1.35852206e+00
2.72414029e-01 -1.01724732e+00 9.05222967e-02 -1.76917732e+00
1.61726162e-01 -9.34987009e-01 -3.25096071e-01 5.80577135e-01
-2.59109080e-01 1.74265295e-01 4.65382099e-01 3.31481606e-01
-8.99989784e-01 4.88297909e-01 1.33300078e+00 -5.17826974e-01
-1.30059451e-01 5.69285192e-02 -5.76692283e-01 5.11039436e-01
7.86851346e-01 -7.31718719e-01 -2.29869410e-01 -8.51596117e-01
3.00591469e-01 1.62921712e-01 7.17420042e-01 -8.96514535e-01
8.00389647e-01 -3.94828826e-01 4.84605469e-02 -6.78036749e-01
2.05571249e-01 -1.13500488e+00 4.47376728e-01 5.21037519e-01
7.31481761e-02 4.19508755e-01 1.65670201e-01 6.30173028e-01
-2.13791132e-01 1.01922072e-01 5.91487288e-01 5.61036058e-02
-9.25478220e-01 1.29530489e+00 4.39411938e-01 2.14973778e-01
9.76173639e-01 -1.62706912e-01 -5.04092336e-01 -1.97812930e-01
-4.74812746e-01 7.66274214e-01 5.69525242e-01 4.64855194e-01
5.51362991e-01 -1.69025230e+00 -7.41514742e-01 -7.46876076e-02
1.71427950e-01 6.15279257e-01 2.79857785e-01 1.25920081e+00
-5.56623220e-01 9.91094336e-02 -3.11242957e-02 -9.42052066e-01
-1.24236906e+00 5.96459508e-01 5.70702434e-01 -3.02444786e-01
-7.27240980e-01 8.22226942e-01 4.42786098e-01 -8.69749784e-01
4.44400102e-01 -7.04769865e-02 1.60913229e-01 -4.00192946e-01
1.10455096e-01 1.24078915e-01 -9.36180279e-02 -5.85410833e-01
-6.22427702e-01 1.03988385e+00 7.56951198e-02 5.57889938e-01
1.28061318e+00 9.67530534e-03 -3.10665548e-01 -1.03134930e-01
1.48989928e+00 -1.00361995e-01 -1.15074134e+00 -5.05314350e-01
-3.54069293e-01 -5.66502869e-01 1.37672713e-02 -2.30020359e-01
-1.69547212e+00 6.67933166e-01 6.19560003e-01 -4.25434522e-02
1.00617635e+00 1.29373744e-01 1.13407075e+00 6.50342703e-01
3.63775611e-01 -7.51868308e-01 -6.79110130e-03 4.99699324e-01
7.29354799e-01 -1.60927689e+00 1.02186106e-01 -6.04221225e-01
-1.73671022e-01 9.75190818e-01 9.92225528e-01 -4.00294304e-01
9.17706132e-01 1.38248652e-01 9.04858410e-02 -4.43290651e-01
-4.01529044e-01 -3.46533030e-01 5.81456959e-01 5.23272216e-01
1.57525450e-01 3.70127372e-02 -6.04343675e-02 1.76849753e-01
-7.10920766e-02 -1.05514992e-02 -1.60913467e-02 6.57487333e-01
-1.20859198e-01 -9.58335578e-01 -2.00143650e-01 5.15820622e-01
-2.41827354e-01 -1.13615207e-01 -1.20187491e-01 8.97918165e-01
1.30371407e-01 6.91299736e-01 3.71767767e-02 -5.89450002e-01
5.50975442e-01 -6.66685820e-01 4.42677617e-01 -3.44835192e-01
-4.41314846e-01 -1.71693459e-01 -1.18700378e-01 -7.98247337e-01
-6.15455389e-01 -4.57890630e-01 -1.28027773e+00 -6.98077083e-01
-6.15826547e-01 -4.01345780e-04 4.66619462e-01 9.08121347e-01
5.88217735e-01 5.68932950e-01 4.10166025e-01 -9.05139863e-01
-1.72232404e-01 -5.26227713e-01 -1.22778021e-01 4.22036529e-01
5.78420460e-01 -7.15398431e-01 -5.30309938e-02 -3.10453981e-01] | [7.180649757385254, 6.39756441116333] |
187197f7-9889-4503-bd8f-1e0bee01b069 | the-ncte-transcripts-a-dataset-of-elementary | 2211.11772 | null | https://arxiv.org/abs/2211.11772v2 | https://arxiv.org/pdf/2211.11772v2.pdf | The NCTE Transcripts: A Dataset of Elementary Math Classroom Transcripts | Classroom discourse is a core medium of instruction - analyzing it can provide a window into teaching and learning as well as driving the development of new tools for improving instruction. We introduce the largest dataset of mathematics classroom transcripts available to researchers, and demonstrate how this data can help improve instruction. The dataset consists of 1,660 45-60 minute long 4th and 5th grade elementary mathematics observations collected by the National Center for Teacher Effectiveness (NCTE) between 2010-2013. The anonymized transcripts represent data from 317 teachers across 4 school districts that serve largely historically marginalized students. The transcripts come with rich metadata, including turn-level annotations for dialogic discourse moves, classroom observation scores, demographic information, survey responses and student test scores. We demonstrate that our natural language processing model, trained on our turn-level annotations, can learn to identify dialogic discourse moves and these moves are correlated with better classroom observation scores and learning outcomes. This dataset opens up several possibilities for researchers, educators and policymakers to learn about and improve K-12 instruction. The dataset can be found at https://github.com/ddemszky/classroom-transcript-analysis. | ['Heather Hill', 'Dorottya Demszky'] | 2022-11-21 | null | null | null | null | ['elementary-mathematics'] | ['reasoning'] | [-1.80355027e-01 3.18582982e-01 -7.47811317e-01 -6.09209359e-01
-1.20949113e+00 -1.04145753e+00 5.82863390e-01 1.07909262e+00
-2.87329942e-01 3.42797786e-01 1.29552877e+00 -9.15791512e-01
-2.26280272e-01 -7.24158347e-01 -6.38718963e-01 -2.83005267e-01
5.17788172e-01 7.83397928e-02 -9.75373238e-02 -1.52747974e-01
4.95221764e-01 1.75048411e-01 -1.49026930e+00 5.78192294e-01
1.37471688e+00 -1.21236280e-01 -4.64215002e-04 8.35056722e-01
-3.19556415e-01 1.43352127e+00 -6.72641098e-01 -3.30042928e-01
-1.84804574e-01 -4.17548180e-01 -1.22441792e+00 -2.97920376e-01
1.19437981e+00 -6.61931157e-01 -3.94349337e-01 8.92134845e-01
3.19019228e-01 5.88282883e-01 2.45567068e-01 -8.22383463e-01
-9.04435515e-01 1.17762268e+00 -3.91779169e-02 5.51005006e-01
6.15851879e-01 2.62562752e-01 9.80247557e-01 -4.41468298e-01
8.17070961e-01 1.13069248e+00 2.14347482e-01 5.75851798e-01
-1.04891658e+00 -9.62028563e-01 9.72850248e-02 4.28128481e-01
-5.53194165e-01 -6.32155061e-01 3.76280874e-01 -1.09452724e+00
4.88045514e-01 1.50100052e-01 1.08175504e+00 1.18396354e+00
8.19962174e-02 1.00932622e+00 1.57031310e+00 -3.64944547e-01
-1.31678686e-01 5.12907766e-02 9.06679034e-01 7.21914828e-01
-8.39959905e-02 -2.91379333e-01 -9.01272416e-01 8.29147026e-02
1.49616674e-01 2.62744203e-02 -2.88890749e-01 1.57125175e-01
-1.14257967e+00 8.56072903e-01 3.20589468e-02 2.52433032e-01
3.14584434e-01 -1.48063704e-01 3.91520709e-01 6.02070630e-01
5.03112257e-01 6.44359469e-01 -2.43931323e-01 -1.18805039e+00
-5.90359330e-01 4.27253455e-01 7.21450567e-01 7.26003051e-01
3.03798378e-01 -2.27605805e-01 -2.31806487e-01 7.75797307e-01
2.61881739e-01 3.55001897e-01 3.63830686e-01 -1.30642438e+00
7.11267054e-01 9.22124863e-01 -3.36033672e-01 -4.70950812e-01
-2.00767264e-01 1.17103443e-01 7.33441338e-02 -1.16695523e-01
8.68365943e-01 -1.12091064e-01 -5.31131446e-01 1.69956017e+00
7.38315403e-01 1.39388874e-01 -1.69104978e-01 3.75583142e-01
1.65228653e+00 4.49516326e-01 2.36457095e-01 -1.02422453e-01
1.48830068e+00 -7.89418519e-01 -1.11896598e+00 3.84204596e-01
1.23523235e+00 -1.23656762e+00 1.18292689e+00 3.86312544e-01
-1.35075676e+00 -3.34596664e-01 -3.30825239e-01 -7.43059278e-01
-4.66025978e-01 -3.04332316e-01 2.24639446e-01 7.51471281e-01
-9.82868493e-01 2.51001507e-01 -1.06053495e+00 -2.66578794e-01
7.61176467e-01 1.57667562e-01 -3.81022274e-01 -4.57962975e-02
-3.82376462e-01 5.98842680e-01 -3.85923445e-01 -4.25761431e-01
-1.13777375e+00 -1.45778513e+00 -8.65076602e-01 1.55373931e-01
4.36866999e-01 5.41235581e-02 1.87564063e+00 -1.70565560e-01
-1.88682950e+00 9.53335047e-01 -2.34210700e-01 1.27571896e-01
1.76815480e-01 -5.10995626e-01 3.55486661e-01 -5.61459213e-02
2.11770326e-01 4.05110002e-01 -2.35589802e-01 -5.13885677e-01
-6.65924430e-01 -5.18061578e-01 -1.78793892e-02 2.17863873e-01
-5.16104102e-01 3.65788549e-01 5.13682008e-01 -3.21846753e-01
1.34381771e-01 -7.14268744e-01 1.95497811e-01 -5.64946413e-01
-1.46091163e-01 -9.94356096e-01 8.12954426e-01 -8.93336594e-01
1.18619907e+00 -1.96172154e+00 -1.28712982e-01 -1.63013756e-01
6.04828179e-01 2.43767187e-01 1.43057364e-03 6.09212935e-01
-1.07217692e-01 4.29689556e-01 5.68001509e-01 -1.73853651e-01
6.84150383e-02 -1.52741387e-01 -3.82240832e-01 5.59080184e-01
-3.03147346e-01 8.36557150e-01 -1.44601858e+00 -2.28907079e-01
5.85404992e-01 3.43877450e-02 -7.34763980e-01 6.45475030e-01
-7.63955191e-02 8.50467622e-01 -5.42561114e-01 4.26917255e-01
1.24016337e-01 1.35420904e-01 -3.10634095e-02 7.28993118e-01
-8.60862255e-01 1.44601476e+00 -4.98000771e-01 1.53705299e+00
-3.28491867e-01 1.40199459e+00 2.47866496e-01 -7.64295042e-01
7.85591841e-01 4.77123737e-01 3.79807711e-01 -5.45499861e-01
2.06833884e-01 -1.62276298e-01 3.03874999e-01 -1.05090022e+00
5.67844927e-01 4.11166459e-01 -6.34943992e-02 8.95501375e-01
1.46447286e-01 -6.81706548e-01 1.43851876e-01 4.61124390e-01
1.40601325e+00 -2.06976086e-01 5.01959994e-02 -4.97740537e-01
1.31641269e-01 1.88344702e-01 5.28684914e-01 4.18921858e-01
-6.48978949e-01 -1.48637339e-01 5.77238142e-01 -3.37230831e-01
-6.61350429e-01 -1.08726692e+00 -1.57109588e-01 1.70746171e+00
-7.32896686e-01 -7.00649858e-01 -8.84176970e-01 -3.23209554e-01
-1.58399403e-01 8.25480044e-01 -3.23879093e-01 3.58540565e-02
-6.67084932e-01 3.28872651e-01 4.68047917e-01 1.60618946e-01
3.91378909e-01 -7.68649817e-01 -3.29651266e-01 -1.94637273e-02
-5.56330621e-01 -1.06822157e+00 -4.05020326e-01 -1.69569641e-01
-9.04803872e-01 -1.22641551e+00 1.64314032e-01 -9.31829631e-01
4.65979695e-01 4.41193640e-01 1.05614102e+00 3.94845009e-01
-1.04798719e-01 9.02073622e-01 -3.33968401e-01 -7.42504299e-01
-8.23222160e-01 2.80736715e-01 2.01341704e-01 -1.01749384e+00
1.04006386e+00 -6.28678501e-01 -4.02512610e-01 -2.16967270e-01
-4.53469485e-01 2.31470257e-01 -2.29727983e-01 5.34672558e-01
3.58140841e-02 -4.42652673e-01 4.65454221e-01 -9.27829325e-01
7.00078011e-01 -5.00859261e-01 -8.99915934e-01 5.67899905e-02
-4.89721596e-01 -4.59375918e-01 4.44258779e-01 -2.40841210e-01
-1.16917014e+00 -4.31759804e-01 -1.56005055e-01 1.49625698e-02
-7.26307809e-01 3.42229903e-01 2.36256376e-01 7.49028400e-02
3.82474184e-01 -3.91012132e-01 -2.35172734e-01 -6.06777906e-01
3.85624737e-01 9.45541382e-01 6.40600622e-01 -9.92135525e-01
6.02499306e-01 -7.78328162e-03 -3.05874377e-01 -9.87463474e-01
-1.23299325e+00 -6.10456944e-01 -7.95256615e-01 -6.17972791e-01
1.10930061e+00 -1.35074139e+00 -1.61833000e+00 2.13113561e-01
-7.81117260e-01 -1.12447453e+00 -3.46914470e-01 8.61517310e-01
-2.46316880e-01 -2.56810158e-01 -1.09152031e+00 -5.43856680e-01
9.88530368e-02 -1.48702002e+00 5.11021256e-01 7.69348681e-01
-5.60479283e-01 -1.08380532e+00 4.38624203e-01 1.51064062e+00
-1.60637140e-01 9.83071234e-03 1.22885847e+00 -1.01608396e+00
-7.39352524e-01 5.54762900e-01 4.29256290e-01 1.11154005e-01
6.16280809e-02 7.59131312e-01 -9.79341805e-01 -3.59009989e-02
-1.33019939e-01 -8.04713547e-01 3.37529421e-01 3.60284865e-01
1.29188478e+00 -6.55097365e-01 1.92771539e-01 2.47396693e-01
8.49414766e-01 1.17373608e-01 6.44111633e-02 3.80322814e-01
9.51590359e-01 9.21508312e-01 4.83349800e-01 8.16029608e-02
1.03845060e+00 1.88290626e-01 -3.76863033e-02 6.43328488e-01
-2.69270837e-01 -5.42830288e-01 7.83824801e-01 1.89754808e+00
1.42337203e-01 5.19740991e-02 -1.52734780e+00 1.23436546e+00
-1.41168749e+00 -1.19989324e+00 -6.88085973e-01 1.81910610e+00
1.35862052e+00 -3.57434630e-01 1.06938995e-01 -2.57711530e-01
2.35805288e-01 1.12382300e-01 -4.57143784e-02 -1.08864379e+00
6.46924973e-01 6.76606894e-01 1.10438548e-01 9.26812768e-01
-5.71038604e-01 1.04683363e+00 5.96295118e+00 6.43965900e-01
-9.09889758e-01 3.42644244e-01 4.73379850e-01 -5.88104367e-01
-6.49314165e-01 -7.91738033e-02 -9.47244346e-01 2.61055380e-01
1.57605827e+00 -3.84886831e-01 3.99237812e-01 4.16094422e-01
6.75772250e-01 -2.21927077e-01 -1.17510211e+00 2.20505312e-01
-2.77591556e-01 -1.52115905e+00 -5.44474900e-01 2.45956644e-01
1.30167603e+00 1.75727263e-01 3.05174232e-01 4.83539313e-01
1.26327181e+00 -1.16960669e+00 4.39289451e-01 1.61261901e-01
3.23300451e-01 -5.63528359e-01 2.68338948e-01 4.36355084e-01
-8.64323854e-01 -3.21460664e-01 9.50068142e-03 -6.62759781e-01
-6.31165326e-01 2.12776735e-01 -1.41025496e+00 -6.13888092e-02
7.95201421e-01 7.07521617e-01 -4.61261719e-01 6.58848345e-01
-6.78312540e-01 1.65942943e+00 -2.62599774e-02 -1.45314962e-01
3.16713095e-01 -4.74990666e-01 4.25408244e-01 8.56220901e-01
1.31715909e-01 9.38347816e-01 2.07365751e-01 7.24156976e-01
-5.06241500e-01 2.27596208e-01 -7.99081326e-01 -3.81210655e-01
1.22458518e+00 1.25130415e+00 -5.44973791e-01 -1.58680558e-01
-5.83692729e-01 -1.15333013e-01 6.01546407e-01 2.90998787e-01
-2.11776003e-01 1.38266489e-01 1.15116835e+00 1.47082642e-01
-3.49075139e-01 -4.33371991e-01 -5.09041786e-01 -9.03189540e-01
-4.22863543e-01 -1.50323927e+00 2.60956258e-01 -4.39147383e-01
-9.04308736e-01 -6.16290569e-01 5.61909704e-03 -5.94390750e-01
-2.24452317e-01 -4.27050948e-01 -9.76178944e-01 4.32555884e-01
-1.08031678e+00 -6.31581664e-01 -5.19592881e-01 3.92055094e-01
7.78255701e-01 -8.90541896e-02 6.63327634e-01 2.25175038e-01
-9.91104126e-01 6.10025287e-01 3.11409444e-01 5.00221431e-01
1.22320449e+00 -1.44654810e+00 4.44303565e-02 5.02243221e-01
2.57966280e-01 7.34051466e-01 5.75141430e-01 -5.54862320e-01
-1.49564183e+00 -8.82610381e-01 1.07112491e+00 -1.18208933e+00
9.71233249e-01 -4.01230037e-01 -9.98373628e-01 1.29939985e+00
6.96547925e-01 -5.97597897e-01 1.51716781e+00 5.41626215e-01
-1.40841112e-01 2.63631701e-01 -8.52351069e-01 6.48940980e-01
9.23157275e-01 -1.02941108e+00 -1.11966586e+00 6.05535209e-01
1.00713241e+00 -1.02048719e+00 -1.78929555e+00 -8.23038220e-02
3.61243218e-01 -8.72723639e-01 6.13932252e-01 -8.59691858e-01
1.13964641e+00 3.79386395e-01 2.88475275e-01 -1.27491891e+00
7.07202032e-02 -7.26066113e-01 -7.64515717e-03 1.56525707e+00
2.36139745e-02 -3.64966065e-01 7.70956337e-01 9.65558887e-01
-6.18265986e-01 -8.41252029e-01 -8.69858801e-01 -1.43862486e-01
9.44838166e-01 -2.46317878e-01 8.33917022e-01 1.69744432e+00
5.04138470e-01 1.62812680e-01 6.34887457e-01 -1.87828559e-02
2.61692196e-01 1.45364970e-01 1.12200451e+00 -1.16989470e+00
4.28450257e-01 -8.09284270e-01 -1.08163469e-01 -9.10173953e-01
5.95736325e-01 -1.05692625e+00 -1.25575989e-01 -1.37886512e+00
4.56416994e-01 -5.03997982e-01 3.36310685e-01 7.21375942e-01
-3.77761275e-01 -3.73237848e-01 1.40450805e-01 -3.76479506e-01
-5.80017030e-01 4.75471675e-01 1.62172055e+00 1.97890043e-01
-3.50777715e-01 -3.07387173e-01 -6.29570365e-01 8.56508672e-01
7.73149788e-01 -5.31870306e-01 -3.88314694e-01 -6.26025021e-01
-1.06294975e-02 9.83248875e-02 1.37614638e-01 -5.61308801e-01
3.21911007e-01 -8.67184579e-01 1.41014665e-01 -6.25435352e-01
-2.05805838e-01 -3.95781815e-01 -5.42443693e-01 1.81289256e-01
-1.05046356e+00 1.07311845e-01 3.10959846e-01 -1.99414715e-01
-1.45279199e-01 -3.85642946e-01 3.95985186e-01 1.15704887e-01
-1.80244938e-01 -1.83683634e-03 -8.22270751e-01 5.02572834e-01
9.50915575e-01 8.60764906e-02 -1.21299946e+00 -5.63800395e-01
-3.98529291e-01 6.08750522e-01 4.19919074e-01 5.67531466e-01
2.64813364e-01 -1.24411213e+00 -8.48681688e-01 1.81494094e-02
-2.68175811e-01 4.83044595e-01 2.65453517e-01 8.17184806e-01
-3.47421199e-01 7.26953030e-01 -8.04132894e-02 -6.40497029e-01
-1.78463209e+00 -1.12031393e-01 -7.41053224e-02 -1.53687492e-01
-4.68104124e-01 1.20903313e+00 -5.69513291e-02 -9.27513421e-01
2.78264523e-01 -9.10120368e-01 -3.55551213e-01 3.04666460e-01
6.94324732e-01 8.57073545e-01 -1.05444737e-01 -4.03169125e-01
3.68274271e-01 8.53156373e-02 -3.90895568e-02 -7.22817518e-03
1.58490729e+00 -1.89037263e-01 -2.11691067e-01 8.31065178e-01
1.15755916e+00 7.01873183e-01 -6.99190140e-01 -1.58187807e-01
7.00664371e-02 -6.93599820e-01 7.34380037e-02 -5.83368480e-01
-6.72437549e-01 1.17170882e+00 2.87956119e-01 1.65307567e-01
5.80520391e-01 6.76267296e-02 7.68573046e-01 3.26302797e-01
-2.22332507e-01 -1.03197336e+00 3.49618644e-01 8.59235823e-01
3.19160432e-01 -1.46582675e+00 5.73490448e-02 -3.67539465e-01
9.86611024e-02 1.03941107e+00 1.03867865e+00 3.55252475e-01
3.45083207e-01 2.02339798e-01 3.07505310e-01 -4.56783772e-01
-1.20114064e+00 1.99799418e-01 2.65327632e-01 3.21829349e-01
1.28443360e+00 5.22725105e-01 -1.91383556e-01 4.09754574e-01
-1.00968647e+00 -3.35946560e-01 1.15660322e+00 7.93080926e-01
-7.55792260e-01 -1.19844663e+00 -7.39139974e-01 6.80537879e-01
-6.51590824e-01 -4.15890962e-02 -6.31855547e-01 7.33438075e-01
-2.39887256e-02 1.36491811e+00 3.80770177e-01 -1.28436521e-01
1.06727906e-01 5.02740204e-01 5.69642842e-01 -1.08634555e+00
-1.28548396e+00 -5.40774941e-01 2.11945638e-01 -5.10258079e-01
-2.76380777e-01 -1.08216178e+00 -1.27810287e+00 -1.07874024e+00
-2.14232609e-01 3.34817529e-01 4.58287507e-01 9.73275959e-01
9.50776190e-02 7.80681789e-01 4.45726484e-01 -2.25666109e-02
-4.59684253e-01 -8.86022210e-01 1.32074118e-01 1.57917380e-01
6.23604894e-01 -3.67332637e-01 -1.70350328e-01 4.53110375e-02] | [11.384937286376953, 8.251638412475586] |
87a0ff5d-63f8-4187-835b-4a32d3c0f3b3 | an-adaptive-parameter-estimation-for-guided | 1609.01380 | null | http://arxiv.org/abs/1609.01380v1 | http://arxiv.org/pdf/1609.01380v1.pdf | An Adaptive Parameter Estimation for Guided Filter based Image Deconvolution | Image deconvolution is still to be a challenging ill-posed problem for
recovering a clear image from a given blurry image, when the point spread
function is known. Although competitive deconvolution methods are numerically
impressive and approach theoretical limits, they are becoming more complex,
making analysis, and implementation difficult. Furthermore, accurate estimation
of the regularization parameter is not easy for successfully solving image
deconvolution problems. In this paper, we develop an effective approach for
image restoration based on one explicit image filter - guided filter. By
applying the decouple of denoising and deblurring techniques to the
deconvolution model, we reduce the optimization complexity and achieve a simple
but effective algorithm to automatically compute the parameter in each
iteration, which is based on Morozov's discrepancy principle. Experimental
results demonstrate that the proposed algorithm outperforms many
state-of-the-art deconvolution methods in terms of both ISNR and visual
quality. | ['Zhongbo Zhang', 'Hang Yang', 'Yujing Guan'] | 2016-09-06 | null | null | null | null | ['image-deconvolution'] | ['computer-vision'] | [ 1.89563796e-01 -5.76207519e-01 5.82474709e-01 -9.98882353e-02
-5.93742251e-01 -4.58853960e-01 3.33077013e-01 -6.57321453e-01
-3.60547870e-01 8.96566331e-01 1.84082717e-01 -1.58787087e-01
-3.65891457e-01 -1.84804246e-01 -4.45731640e-01 -1.14717841e+00
2.10578859e-01 3.31475511e-02 1.99398901e-02 2.11688615e-02
4.66922790e-01 3.88832211e-01 -1.33905232e+00 -2.68080860e-01
1.39695692e+00 9.31613684e-01 7.13332117e-01 5.69039583e-01
3.56177543e-03 7.47141361e-01 -3.61102104e-01 -1.21468954e-01
2.67098576e-01 -8.70266020e-01 -7.40103900e-01 4.70457882e-01
1.99969038e-01 -7.11263657e-01 -4.54611927e-01 1.55557632e+00
4.69486535e-01 1.97374940e-01 7.14666307e-01 -5.20719886e-01
-8.77518535e-01 2.30851769e-02 -6.95618868e-01 3.64169359e-01
1.06461227e-01 5.70715852e-02 5.03839970e-01 -1.09390831e+00
3.02846819e-01 8.78207147e-01 7.16549098e-01 3.28737110e-01
-1.21260035e+00 -2.93803066e-01 -2.50593990e-01 3.40815216e-01
-1.30620015e+00 -6.33103192e-01 6.75257981e-01 -5.24070382e-01
2.61781394e-01 2.11167023e-01 3.41582984e-01 2.77115345e-01
2.11313307e-01 3.87616843e-01 1.49159706e+00 -5.01325607e-01
1.99629158e-01 -1.00860551e-01 1.98362708e-01 5.02827108e-01
3.53234529e-01 1.10163681e-01 -2.49363393e-01 -1.55587405e-01
1.15536225e+00 1.31211981e-01 -1.04854739e+00 -1.88122228e-01
-9.33197558e-01 6.79766715e-01 3.42521578e-01 3.09745014e-01
-6.66617155e-01 1.34136885e-01 -6.11808449e-02 2.44024336e-01
6.77802980e-01 4.34063375e-01 -4.84180897e-02 2.41180398e-02
-1.14431226e+00 2.30195165e-01 6.72278762e-01 5.38844764e-01
7.54613459e-01 5.88619150e-02 -3.43863219e-02 1.04030108e+00
3.51137012e-01 6.51127756e-01 3.13730657e-01 -1.38019180e+00
-1.00662909e-01 2.52216030e-02 6.86214805e-01 -7.33814716e-01
-1.05105005e-01 -5.69933534e-01 -1.02568400e+00 4.65829790e-01
5.58651865e-01 -1.34684697e-01 -9.10381377e-01 1.23141372e+00
1.34800702e-01 4.31385219e-01 1.31498173e-01 1.48861420e+00
5.68179786e-01 7.36085713e-01 -5.32176554e-01 -5.68292797e-01
1.32900608e+00 -9.62745070e-01 -1.07864976e+00 -2.46147171e-01
-4.91029993e-02 -1.14529681e+00 6.13311589e-01 4.82785463e-01
-1.24799848e+00 -2.94819057e-01 -9.72787082e-01 -9.34186131e-02
3.95953566e-01 2.16695845e-01 4.56320941e-01 5.20191252e-01
-9.47408199e-01 7.73963392e-01 -6.05189621e-01 -1.16971895e-01
3.44053507e-01 4.77190353e-02 -2.56367892e-01 -4.74693805e-01
-7.59894192e-01 1.01915407e+00 4.25393060e-02 4.92017090e-01
-7.91172147e-01 -6.84824049e-01 -5.44578195e-01 1.19263008e-01
9.12632123e-02 -7.55619168e-01 1.19851208e+00 -7.98376977e-01
-1.74393284e+00 6.00293875e-01 -4.36071575e-01 -3.41025412e-01
4.38264489e-01 -4.27966505e-01 -7.42520839e-02 5.77196658e-01
-1.09974615e-01 -1.41615957e-01 1.40343058e+00 -1.60596454e+00
-3.08687836e-01 -2.49248981e-01 -1.54564112e-01 3.17496091e-01
-2.05708086e-03 9.29501131e-02 -3.52705657e-01 -5.62221289e-01
4.74084496e-01 -6.36608005e-01 -3.71635616e-01 2.80962229e-01
-1.91186666e-01 4.96048391e-01 6.56487525e-01 -1.23249185e+00
9.91792858e-01 -2.23943853e+00 1.68133706e-01 -1.40478000e-01
4.41936851e-01 4.33280528e-01 5.82673512e-02 3.40192527e-01
-3.86679694e-02 -3.44713360e-01 -7.53063738e-01 -2.50999093e-01
-4.20003772e-01 9.22750086e-02 -3.23415697e-01 1.01713300e+00
-1.69552192e-01 5.89053571e-01 -7.43268371e-01 -2.07821932e-02
4.18643683e-01 7.22702384e-01 -4.41079557e-01 5.14321804e-01
2.94592321e-01 8.88934851e-01 -2.62067616e-01 2.44466633e-01
1.29793966e+00 -4.34542507e-01 -1.59602519e-02 -3.31389964e-01
-4.41487700e-01 -1.62580356e-01 -1.20928299e+00 1.37787974e+00
-4.57468778e-01 7.44812548e-01 6.18218064e-01 -1.13203716e+00
7.55179167e-01 4.02911454e-01 3.34966481e-01 -3.60844433e-01
1.26604423e-01 4.98897165e-01 -2.13114753e-01 -8.08646917e-01
3.10514718e-01 -6.71380520e-01 7.05328345e-01 3.33817303e-01
-5.71472105e-03 -3.02825719e-01 8.99350345e-02 -1.26799405e-01
8.77657950e-01 -6.39163479e-02 2.16850415e-01 -6.26659572e-01
8.01495671e-01 -1.77331641e-01 3.05420339e-01 7.20709801e-01
-4.06361595e-02 1.15818644e+00 4.65333797e-02 -2.15665147e-01
-1.07843471e+00 -9.07030106e-01 -3.78847063e-01 2.03167439e-01
5.82213044e-01 1.61221340e-01 -1.12520301e+00 3.81630957e-02
-2.88901299e-01 4.52266812e-01 -1.77197903e-01 1.88048378e-01
-5.24462819e-01 -1.07406294e+00 5.92148490e-02 -6.55338541e-02
8.43090773e-01 -4.86984074e-01 -3.88961285e-01 2.62053102e-01
-5.83208382e-01 -1.26088631e+00 -6.13443315e-01 -2.01612025e-01
-1.05913091e+00 -1.30519211e+00 -1.51594734e+00 -7.60827303e-01
1.00939310e+00 9.76785719e-01 7.28393972e-01 2.94996560e-01
-2.05068395e-01 3.02130491e-01 -2.11379319e-01 4.83828783e-02
-3.78646582e-01 -7.00900018e-01 -1.59157351e-01 3.18825424e-01
-1.34223118e-01 -6.40838981e-01 -9.94246602e-01 4.67492431e-01
-9.60813344e-01 5.14774099e-02 5.91434419e-01 1.15151155e+00
3.33040923e-01 4.04029727e-01 3.21321219e-01 -4.57975745e-01
8.23410034e-01 -1.33171901e-01 -7.99180627e-01 5.66617548e-02
-6.62907541e-01 7.66400248e-02 6.13099158e-01 -2.45891377e-01
-1.49909079e+00 -8.04276392e-02 2.84987111e-02 -6.24518514e-01
1.17730118e-01 5.03624916e-01 3.50649238e-01 -5.12302935e-01
6.44158423e-01 6.30586445e-01 4.91429627e-01 -8.64445865e-01
3.12190622e-01 6.59336388e-01 9.47534800e-01 -2.29015648e-01
1.04246163e+00 7.85463750e-01 -1.62468776e-01 -1.21650791e+00
-8.07015896e-01 -7.70797014e-01 -4.33834791e-01 -3.01207006e-01
8.19598615e-01 -1.00728452e+00 -7.70091295e-01 9.14630413e-01
-1.43257415e+00 -2.47466993e-02 5.26643395e-02 7.67975688e-01
-4.99612868e-01 9.47200537e-01 -8.49924326e-01 -8.32198918e-01
-2.63884425e-01 -1.16232407e+00 6.20920777e-01 4.30897862e-01
4.15718168e-01 -1.08734059e+00 -2.13187151e-02 5.14494538e-01
7.56051660e-01 -2.04987407e-01 4.49394643e-01 3.17333430e-01
-8.42764854e-01 -3.22406851e-02 -7.48097301e-01 7.53966689e-01
1.89751893e-01 -4.26315308e-01 -1.03494430e+00 -4.16425765e-01
9.95311975e-01 1.67230710e-01 9.23470676e-01 8.40795517e-01
9.63491380e-01 -3.30402285e-01 -3.68900672e-02 9.42142546e-01
1.72626591e+00 -7.44201764e-02 9.57203925e-01 1.17889740e-01
4.05714482e-01 3.85372639e-01 3.56650829e-01 3.77377063e-01
-3.62810381e-02 4.92109239e-01 4.51023906e-01 -1.96700498e-01
-3.98281485e-01 3.94487888e-01 2.14484766e-01 9.21086013e-01
-5.58245480e-01 -4.23131362e-02 -4.72344130e-01 5.52222610e-01
-1.77522671e+00 -9.32260692e-01 -5.19368649e-01 2.26179862e+00
8.62635434e-01 -4.32652861e-01 -6.11548245e-01 2.12079082e-02
8.80986273e-01 -5.07251322e-02 -3.51389080e-01 2.26973400e-01
-1.97311163e-01 -1.07069258e-02 6.35808647e-01 8.13916326e-01
-9.14243996e-01 5.38302898e-01 6.88810492e+00 6.55548394e-01
-1.05887008e+00 1.99265122e-01 3.28029424e-01 3.67236972e-01
-1.83999881e-01 6.72568455e-02 -3.01475585e-01 5.47584474e-01
5.33800960e-01 -2.44281724e-01 9.44585025e-01 3.54712248e-01
5.68456173e-01 -4.71172601e-01 -6.53475583e-01 1.38150144e+00
6.02533296e-02 -1.31981611e+00 -3.69261026e-01 5.42839505e-02
7.39514828e-01 -8.23559910e-02 -1.52870238e-01 -5.14981389e-01
-7.11835772e-02 -9.61231112e-01 5.08438468e-01 7.95996010e-01
5.44341147e-01 -4.11468148e-01 8.47155392e-01 3.83407682e-01
-6.28958941e-01 4.45850603e-02 -5.24380326e-01 -2.07849562e-01
5.09125531e-01 1.12685597e+00 -1.36216328e-01 6.18939340e-01
7.11504459e-01 6.55856848e-01 -2.54709031e-02 1.65751326e+00
-3.52316409e-01 5.59072793e-01 1.93735342e-02 4.83237624e-01
1.06557086e-02 -9.40489888e-01 7.93675184e-01 8.92315805e-01
7.84537673e-01 6.30778372e-01 -2.97316998e-01 1.08205593e+00
1.03018545e-01 -2.56931335e-01 -2.67862171e-01 1.84052885e-01
1.24161288e-01 1.24478769e+00 -5.50665617e-01 -4.70165834e-02
-4.23596591e-01 1.31106973e+00 -1.53491125e-01 8.27999353e-01
-6.17901862e-01 -2.49787495e-01 7.74611592e-01 -2.06399187e-02
4.89054233e-01 -2.96695679e-01 -2.19389319e-01 -1.39649057e+00
1.28717169e-01 -7.24349201e-01 -1.51038721e-01 -1.07685268e+00
-1.37597382e+00 5.96132934e-01 -3.33642453e-01 -1.43228531e+00
5.54117635e-02 -7.15998352e-01 -6.49838030e-01 1.25965130e+00
-2.07705784e+00 -6.87874258e-01 -4.95974511e-01 5.93992531e-01
4.08593595e-01 1.55255049e-01 5.86803794e-01 3.88617605e-01
-3.61645728e-01 -1.07856929e-01 8.71197283e-01 -1.99872151e-01
6.74892485e-01 -9.21428323e-01 -9.27856937e-02 1.14756143e+00
-4.71519560e-01 6.56826854e-01 1.21696079e+00 -3.32207352e-01
-1.53095412e+00 -4.64807630e-01 4.79214519e-01 1.61263138e-01
6.79234385e-01 5.47316894e-02 -1.24365628e+00 2.40040556e-01
3.82813543e-01 1.57741413e-01 3.01299393e-01 -3.44877750e-01
1.33198127e-02 -2.50193346e-02 -1.16086900e+00 2.95924306e-01
6.56789005e-01 -4.74272043e-01 -6.65372849e-01 3.54225397e-01
3.05727243e-01 -4.30683404e-01 -6.30108774e-01 9.46626440e-02
3.09215128e-01 -1.16026378e+00 1.12657034e+00 9.70441699e-02
5.63263893e-01 -6.07996881e-01 1.66264791e-02 -1.50399578e+00
-4.59884554e-01 -9.74295974e-01 -3.08235008e-02 9.78802621e-01
-2.19548959e-03 -8.77607286e-01 3.86230558e-01 3.33074749e-01
-2.40383625e-01 -3.77305478e-01 -8.34553063e-01 -8.32664728e-01
-2.28716865e-01 2.15168744e-02 3.33177969e-02 7.77522743e-01
-2.14967862e-01 1.03882395e-01 -7.60126412e-01 5.78960121e-01
1.30825675e+00 3.26534092e-01 4.02870268e-01 -1.13672853e+00
-3.55955571e-01 -3.42897952e-01 -1.22332178e-01 -1.57696569e+00
-1.01813413e-01 -5.17069221e-01 4.72803086e-01 -1.73095083e+00
3.54808688e-01 -1.69012845e-01 2.29974780e-02 -1.79955125e-01
-3.24323416e-01 2.89833874e-01 4.91692014e-02 7.06403315e-01
2.15090718e-02 6.41116858e-01 1.67126358e+00 9.91765130e-03
3.77839878e-02 3.08064446e-02 -7.11783051e-01 7.79043734e-01
4.89055425e-01 -3.12374949e-01 -3.83603632e-01 -6.60348475e-01
-2.33752340e-01 3.11479211e-01 5.79368830e-01 -7.80205548e-01
4.50666606e-01 -1.15293741e-01 1.42532483e-01 -2.90050268e-01
5.10139108e-01 -7.62095094e-01 2.30106115e-01 4.31558162e-01
1.39931202e-01 -5.49002171e-01 6.55738171e-03 5.81234813e-01
-4.37789828e-01 -6.67961299e-01 1.20808673e+00 -2.24839181e-01
-5.70007086e-01 1.24205172e-01 -3.78700674e-01 -8.06985721e-02
5.77729106e-01 -1.95193559e-01 -3.90816391e-01 -5.50845265e-01
-5.09504080e-01 -1.26718760e-01 5.29442132e-01 -1.87403128e-01
8.59610975e-01 -9.97085929e-01 -9.72124338e-01 8.46431870e-03
-3.75896156e-01 -1.82762623e-01 5.71693301e-01 1.18053031e+00
-9.45848644e-01 8.24491978e-02 -1.22408554e-01 -5.30485749e-01
-1.08287871e+00 4.03026879e-01 5.95773399e-01 2.54211202e-02
-9.03555632e-01 7.66113222e-01 5.51419139e-01 -4.20174561e-02
-9.64019001e-02 -6.97367778e-03 -3.73378694e-02 -3.26618999e-01
9.81182337e-01 5.73056519e-01 -1.02012485e-01 -6.07341230e-01
-2.44807396e-02 8.35514605e-01 2.02587277e-01 -9.88416672e-02
1.52743411e+00 -6.90631151e-01 -6.43360913e-01 -8.91013965e-02
1.04590106e+00 1.17141223e-02 -1.57752728e+00 -2.64907479e-01
-3.15959245e-01 -8.29068303e-01 5.14003873e-01 -5.61621904e-01
-1.09615576e+00 8.41431856e-01 5.94154596e-01 2.39860639e-01
1.32060170e+00 -2.79287100e-01 8.71673942e-01 1.60138234e-01
1.43308342e-01 -8.05885553e-01 -5.39959893e-02 4.17463750e-01
1.04098737e+00 -1.23183465e+00 3.13618124e-01 -7.99759090e-01
-1.16339006e-01 1.21969485e+00 4.45947330e-03 -2.15213984e-01
7.71095335e-01 2.33009644e-02 2.15882763e-01 -8.67311209e-02
-1.42318711e-01 -7.30607733e-02 3.11306238e-01 4.91703600e-01
3.06071877e-01 -2.56125420e-01 -5.92605412e-01 3.27589095e-01
3.02624375e-01 2.54481614e-01 8.80219221e-01 5.57071149e-01
-8.07039917e-01 -7.81475425e-01 -9.44350481e-01 1.47484720e-01
-6.03740573e-01 -2.53963441e-01 4.26769912e-01 2.74622083e-01
-3.40484440e-01 1.07980609e+00 -3.13167214e-01 2.32308462e-01
3.02499067e-02 -3.70898485e-01 6.39379323e-01 -2.35578597e-01
1.41708180e-02 3.61595869e-01 -1.55123383e-01 -1.94016576e-01
-6.55378044e-01 -5.24070501e-01 -1.17078567e+00 -3.83102924e-01
-6.19065046e-01 2.45568588e-01 7.05380023e-01 1.09256053e+00
2.02324450e-01 2.53946960e-01 8.06330502e-01 -1.17824376e+00
-7.45355308e-01 -8.25054526e-01 -9.78768349e-01 2.59804130e-01
7.49051988e-01 -5.91393352e-01 -7.83165395e-01 3.75287950e-01] | [11.603854179382324, -2.692882776260376] |
0e96c489-0fe1-4fd7-8498-06ff23a09809 | deep-convolutional-sparse-coding-network-for | 2103.05946 | null | https://arxiv.org/abs/2103.05946v1 | https://arxiv.org/pdf/2103.05946v1.pdf | Deep Convolutional Sparse Coding Network for Pansharpening with Guidance of Side Information | Pansharpening is a fundamental issue in remote sensing field. This paper proposes a side information partially guided convolutional sparse coding (SCSC) model for pansharpening. The key idea is to split the low resolution multispectral image into a panchromatic image related feature map and a panchromatic image irrelated feature map, where the former one is regularized by the side information from panchromatic images. With the principle of algorithm unrolling techniques, the proposed model is generalized as a deep neural network, called as SCSC pansharpening neural network (SCSC-PNN). Compared with 13 classic and state-of-the-art methods on three satellites, the numerical experiments show that SCSC-PNN is superior to others. The codes are available at https://github.com/xsxjtu/SCSC-PNN. | ['Chunxia Zhang', 'Junmin Liu', 'Lu Huang', 'Zixiang Zhao', 'Kai Sun', 'Jiangshe Zhang', 'Shuang Xu'] | 2021-03-10 | null | null | null | null | ['pansharpening'] | ['computer-vision'] | [ 6.62233114e-01 -3.98222625e-01 -1.83290705e-01 -1.44918039e-01
-3.29664528e-01 -4.13911492e-01 3.84643108e-01 -5.75637579e-01
-2.21597865e-01 7.42151141e-01 3.72221559e-01 -2.51959443e-01
-5.28531849e-01 -1.15365529e+00 -5.97258449e-01 -1.01295650e+00
1.35373101e-01 -1.88440066e-02 -1.08959973e-01 -6.57042861e-01
9.32483897e-02 6.83380365e-01 -1.35543334e+00 2.15321407e-01
9.99920607e-01 9.72623050e-01 3.95507812e-01 5.77773690e-01
8.79479423e-02 5.66541374e-01 4.77155745e-02 1.64147034e-01
8.31181347e-01 -4.37402248e-01 -6.59887016e-01 2.44906545e-01
2.98160613e-01 -3.57755899e-01 -5.47563851e-01 1.60633028e+00
2.79894680e-01 3.04213256e-01 3.36232543e-01 -9.96019363e-01
-1.07665145e+00 5.73200285e-01 -1.09586823e+00 2.66651422e-01
-2.58684009e-01 -2.71164894e-01 6.50875449e-01 -6.27320409e-01
4.47452158e-01 9.17483151e-01 8.31908286e-01 4.59582603e-04
-7.74324119e-01 -6.46251976e-01 -2.53093541e-01 2.36274526e-01
-1.37983525e+00 8.16500187e-02 8.80551577e-01 -2.79388845e-01
6.77311420e-01 4.66986239e-01 8.35380077e-01 2.87029982e-01
2.32228249e-01 5.73498845e-01 1.20156097e+00 -3.50657314e-01
-8.17407444e-02 -3.14318627e-01 2.64406949e-01 5.26099026e-01
5.68320096e-01 6.59726083e-01 9.79870465e-03 -3.64227071e-02
1.10287941e+00 3.23248923e-01 -7.73317754e-01 -9.75609049e-02
-8.36376965e-01 1.02582359e+00 1.02653062e+00 4.92302179e-01
-6.68627501e-01 -1.31286383e-01 1.09056816e-01 4.30114001e-01
4.80840534e-01 2.65078038e-01 -5.32147467e-01 4.11433160e-01
-1.22404385e+00 1.78200558e-01 3.67511749e-01 6.30390048e-01
1.27294827e+00 4.23286051e-01 3.65707904e-01 9.92539346e-01
2.73034304e-01 8.08056831e-01 5.92219889e-01 -1.07317543e+00
3.07218164e-01 6.04681492e-01 1.45881986e-02 -1.14374733e+00
-4.57751781e-01 -7.21670449e-01 -1.36314964e+00 2.55846322e-01
-4.72794831e-01 -5.43900728e-01 -1.04164886e+00 1.20344138e+00
1.08047277e-01 4.91700396e-02 3.03261757e-01 1.30484796e+00
9.55370545e-01 1.34773123e+00 -4.66472447e-01 -5.82082681e-02
1.13584006e+00 -1.24555922e+00 -7.22064316e-01 -3.08279425e-01
1.78304940e-01 -7.48291016e-01 3.44846845e-01 4.58428770e-01
-4.50936258e-01 -6.92135930e-01 -1.08373559e+00 3.84690315e-02
-5.53366125e-01 3.42644304e-01 9.70308602e-01 4.13196266e-01
-1.09359479e+00 6.34297490e-01 -6.39849961e-01 -3.04519475e-01
3.40746939e-01 -1.21911056e-02 -4.67481852e-01 -1.30129993e-01
-1.28473508e+00 6.97104573e-01 8.93843412e-01 5.58434188e-01
-6.64425373e-01 -4.52575564e-01 -8.45066130e-01 1.06612980e-01
1.09722033e-01 -2.21441492e-01 1.11026585e+00 -1.37740362e+00
-1.50742936e+00 5.98132372e-01 3.21351975e-01 -3.74310076e-01
2.43976843e-02 -3.98836255e-01 -8.41526091e-01 3.64682972e-01
5.41660301e-02 4.70057815e-01 9.47099984e-01 -1.33992803e+00
-5.54495394e-01 -1.90978646e-01 -1.75746873e-01 4.11378741e-01
4.86688241e-02 2.11089384e-02 -3.22151750e-01 -8.76754940e-01
5.09271741e-01 -7.31186271e-01 -4.95706826e-01 -7.24024186e-03
-2.44679540e-01 3.58154118e-01 1.17885578e+00 -1.13608813e+00
7.85157740e-01 -2.16617084e+00 1.66863784e-01 2.04302490e-01
-1.85821265e-01 5.34398735e-01 -4.75820094e-01 5.21919847e-01
-7.30312884e-01 -1.39935315e-01 -8.16911459e-01 3.02487135e-01
-3.83309573e-01 2.51381069e-01 -3.69087130e-01 8.73914301e-01
-1.80483431e-01 8.45677793e-01 -6.33118153e-01 -1.39576882e-01
3.36307913e-01 5.60111284e-01 -1.50904417e-01 -5.27249947e-02
-1.14171773e-01 3.29406798e-01 -3.92740935e-01 7.91138351e-01
1.55141521e+00 -8.74847993e-02 -2.21764371e-01 -2.66526997e-01
-5.65619588e-01 -6.23469472e-01 -1.18004239e+00 1.56813824e+00
-2.29075044e-01 6.18422329e-01 5.44493437e-01 -1.15630078e+00
1.05537260e+00 1.47455305e-01 4.37122136e-01 -8.50571275e-01
2.48507723e-01 2.11079657e-01 -3.35027725e-01 -3.11862648e-01
8.14303577e-01 -3.55467588e-01 5.12602270e-01 3.95642638e-01
-2.24038381e-02 -5.18623829e-01 -3.26203741e-02 3.96043509e-02
1.53114706e-01 1.32335410e-01 4.60950822e-01 -3.95425200e-01
6.53887153e-01 2.49541819e-01 7.15891063e-01 4.86335069e-01
-2.63334997e-02 7.66525924e-01 -9.35760699e-03 -6.56826496e-01
-8.88695240e-01 -5.57834923e-01 -2.07809910e-01 6.48541272e-01
1.59686267e-01 1.81238845e-01 -5.30625582e-01 -1.68005466e-01
-1.72903746e-01 5.56960821e-01 -6.03380740e-01 1.49126686e-02
-3.17605376e-01 -1.04063213e+00 2.75339872e-01 2.52068877e-01
1.33170199e+00 -1.11923397e+00 -5.58813930e-01 7.50279352e-02
-1.86035156e-01 -5.95013380e-01 -2.32844800e-01 -4.28507365e-02
-1.07993281e+00 -1.08472383e+00 -1.02431798e+00 -6.78076863e-01
3.83604020e-01 9.02927458e-01 7.04962432e-01 -1.14537567e-01
-4.55775559e-02 8.46001878e-02 -7.35441625e-01 -1.72430530e-01
-7.43738860e-02 -2.34946445e-01 -3.20514053e-01 3.85639906e-01
2.17372939e-01 -7.72182226e-01 -5.78077853e-01 -1.82762235e-01
-1.34602523e+00 2.79576868e-01 1.00758636e+00 6.50476754e-01
7.93346524e-01 7.43561387e-01 -2.80011129e-02 -7.88308084e-01
2.50596762e-01 -4.81274545e-01 -1.03425229e+00 4.76289466e-02
-5.57939053e-01 -3.38125587e-01 4.78465140e-01 -1.22128865e-02
-1.38380015e+00 1.99328199e-01 -2.70892769e-01 -4.57987696e-01
-2.60803133e-01 1.04627860e+00 6.70152009e-02 -5.32943666e-01
7.69585252e-01 7.82674193e-01 -8.97840783e-02 -8.74920666e-01
4.60643649e-01 8.58240128e-01 8.48451614e-01 -1.38265770e-02
1.01496553e+00 7.50723541e-01 -1.53215349e-01 -1.15044713e+00
-1.07714379e+00 -6.70461893e-01 -5.09683728e-01 -2.50983953e-01
8.86023879e-01 -1.28946030e+00 1.00735888e-01 8.64995539e-01
-8.61659467e-01 -4.67500299e-01 -2.61155963e-01 5.11263311e-01
-3.84392202e-01 6.54621005e-01 -5.87468803e-01 -4.41725940e-01
-6.40248001e-01 -6.67576909e-01 9.59631622e-01 6.84670150e-01
6.44249797e-01 -7.99042165e-01 5.61268330e-01 3.68444204e-01
6.72503352e-01 4.54983652e-01 4.00822818e-01 -9.88070965e-02
-3.22150856e-01 -2.12445736e-01 -6.34214997e-01 7.44860530e-01
2.05368295e-01 -8.09843689e-02 -6.65673375e-01 -4.43691552e-01
4.04367387e-01 -2.41810739e-01 1.29025793e+00 7.21355915e-01
1.27103913e+00 -4.46240276e-01 1.98745728e-03 1.36452711e+00
2.14254594e+00 1.56266049e-01 1.06662536e+00 6.22410893e-01
6.10021591e-01 2.90229693e-02 3.98783833e-01 2.95738220e-01
-4.60693650e-02 1.23290725e-01 7.39073813e-01 -4.29115981e-01
1.11833774e-03 -1.73974261e-01 8.49014670e-02 6.43520236e-01
-5.49864352e-01 -1.94172151e-02 -8.39958608e-01 5.59201777e-01
-1.76022553e+00 -1.15587068e+00 -4.51723665e-01 1.59226382e+00
5.86705267e-01 -4.31967586e-01 -5.82859814e-01 1.56908408e-01
8.86540651e-01 8.75293016e-01 -3.26618075e-01 -3.04754227e-01
-5.55441201e-01 3.67929131e-01 9.98636603e-01 7.66985059e-01
-1.49408531e+00 8.37460220e-01 5.40110540e+00 8.27130020e-01
-1.28274155e+00 2.13280812e-01 2.13703305e-01 5.18789053e-01
-2.19449937e-01 7.29969144e-02 -2.34126106e-01 1.54947549e-01
5.67426682e-01 -1.43269241e-01 6.70412898e-01 9.14757013e-01
1.17810950e-01 -2.29511768e-01 -8.62761140e-02 1.10148478e+00
1.11609682e-01 -1.42969322e+00 2.18347326e-01 -1.30552873e-01
1.23390794e+00 7.04813004e-01 3.90041023e-02 1.62841529e-01
3.36566240e-01 -7.68414319e-01 4.19783294e-01 7.08881140e-01
8.87931168e-01 -6.89727247e-01 9.14198577e-01 2.70663619e-01
-1.28351235e+00 -1.90354481e-01 -6.78377926e-01 -2.09932134e-01
-5.70365451e-02 8.37459862e-01 1.62131637e-01 1.31306040e+00
8.65183711e-01 1.33746946e+00 -3.37877601e-01 1.26095855e+00
-4.73620623e-01 7.43527949e-01 -2.45666847e-01 5.13319850e-01
6.14193141e-01 -8.00229430e-01 5.13766766e-01 1.15058804e+00
4.78871018e-01 6.59101546e-01 3.16074677e-02 7.47965157e-01
1.91235378e-01 -8.73967931e-02 -4.75474805e-01 -7.16127455e-02
1.35689616e-01 1.41892421e+00 -5.95734000e-01 -4.59466606e-01
-4.65065122e-01 1.00533760e+00 -3.09718132e-01 7.23339021e-01
-7.07991362e-01 -5.84840178e-01 2.57890761e-01 -6.06289387e-01
6.34060740e-01 -6.86025172e-02 -1.35786816e-01 -1.40755451e+00
-4.39366758e-01 -1.03115535e+00 3.34411800e-01 -1.20929384e+00
-9.75191236e-01 6.22577727e-01 -7.99740925e-02 -1.54986703e+00
5.09009659e-01 -6.93603396e-01 -7.07993805e-01 1.12302136e+00
-2.11238050e+00 -1.22947180e+00 -7.33693779e-01 8.35461557e-01
3.75161558e-01 -2.28509918e-01 8.07611763e-01 1.51700582e-02
-4.64107066e-01 -3.16819757e-01 7.57551849e-01 1.32639080e-01
5.13803884e-02 -7.02756464e-01 1.19820602e-01 1.35121679e+00
-1.12384081e-01 6.92212880e-02 6.91475809e-01 -6.64451420e-01
-1.11726630e+00 -1.54090178e+00 5.43878317e-01 5.06477058e-01
5.77716053e-01 4.15297210e-01 -8.63198817e-01 8.15868676e-01
5.24053693e-01 5.45315184e-02 4.99549478e-01 -4.42387670e-01
-3.53560090e-01 -1.61334068e-01 -1.07613361e+00 1.20064765e-01
4.86384094e-01 -4.12071139e-01 -8.49586546e-01 4.40472960e-01
6.60650909e-01 -4.04027253e-01 -5.68461657e-01 3.69265497e-01
3.01870406e-01 -9.85266030e-01 8.15442085e-01 -1.58052728e-01
8.46410394e-01 -4.75401998e-01 -5.03875315e-01 -1.38735926e+00
-9.50270712e-01 -2.39118725e-01 2.26888150e-01 4.23647523e-01
1.19713791e-01 -8.07978094e-01 6.37621224e-01 -3.22765976e-01
-2.62138277e-01 -2.93586612e-01 -7.26623714e-01 -7.89638162e-01
2.10640892e-01 -8.53531808e-02 7.49186277e-01 1.36616659e+00
-5.15486002e-01 2.32210811e-02 -5.39353549e-01 7.84816384e-01
6.63574219e-01 8.67135167e-01 1.66884080e-01 -1.18386996e+00
-3.12312543e-01 -4.08840060e-01 -1.54876262e-02 -8.07836592e-01
-9.50749964e-02 -7.98357844e-01 -3.86210755e-02 -1.85540199e+00
2.20549911e-01 -7.70879015e-02 -3.85969728e-01 7.55618751e-01
1.44869119e-01 4.09693003e-01 1.06106535e-01 3.05136800e-01
-2.86838487e-02 8.60648513e-01 1.31491983e+00 -3.72618467e-01
-2.93605894e-01 -5.95481992e-02 -6.61551654e-01 5.93832493e-01
1.27465391e+00 -2.87094563e-01 -1.42648965e-01 -8.06147575e-01
2.32703865e-01 3.50084335e-01 4.94695812e-01 -1.18874383e+00
2.02860802e-01 -3.77790332e-01 1.61220044e-01 -8.91694129e-01
2.20796481e-01 -9.51791644e-01 5.20635545e-01 7.62424529e-01
1.24367721e-01 -3.60135138e-01 3.40469599e-01 3.82884085e-01
-6.42770648e-01 -2.80688703e-01 9.45009172e-01 -4.18227464e-01
-1.18014371e+00 5.14843524e-01 -3.34527373e-01 -5.01114488e-01
6.86055183e-01 -9.23774019e-02 -5.11752605e-01 -3.89754444e-01
-4.81458724e-01 1.21067278e-01 2.15684593e-01 9.61963981e-02
7.05045402e-01 -1.28255737e+00 -8.68786991e-01 4.56256658e-01
-6.55887127e-02 1.40538532e-02 7.93816805e-01 6.07452035e-01
-1.18372214e+00 5.12362659e-01 -6.58262193e-01 -1.18688457e-01
-8.32823634e-01 5.19000232e-01 7.61083066e-01 -5.66390157e-02
-7.92677462e-01 9.68978524e-01 -2.73142960e-02 -8.43646705e-01
-3.13576251e-01 -1.72698960e-01 -4.55516934e-01 -9.10809264e-02
5.73438108e-01 2.97553271e-01 5.94088100e-02 -9.14513648e-01
-1.23300649e-01 7.64779031e-01 2.68176526e-01 -1.34908743e-02
1.90700722e+00 2.28970740e-02 -7.30285704e-01 -2.31712446e-01
1.15675771e+00 -1.74774751e-01 -1.26500857e+00 -4.33032036e-01
-5.01659632e-01 -5.02970695e-01 7.06161857e-01 -8.87301624e-01
-1.70371556e+00 6.94827259e-01 8.25999558e-01 8.19744617e-02
1.64783943e+00 -4.52175826e-01 7.03135252e-01 5.08728087e-01
2.12086499e-01 -1.01691580e+00 -5.69733739e-01 7.06454277e-01
1.33486450e+00 -1.13361824e+00 3.88585180e-01 -2.00749204e-01
-5.69691181e-01 1.23621440e+00 2.61626393e-01 -4.98377502e-01
1.06653106e+00 -6.58764616e-02 1.04092732e-01 -4.71243173e-01
-1.29299462e-01 -1.99007347e-01 -5.08448519e-02 5.25620043e-01
-2.27906592e-02 3.38493228e-01 -3.62862587e-01 3.70163560e-01
-3.79000992e-01 2.35769674e-01 6.95508122e-01 7.75751650e-01
-6.74633026e-01 -6.74201965e-01 -7.18058169e-01 4.75048453e-01
-1.45395845e-01 -2.82638937e-01 -4.99399379e-02 6.71762764e-01
4.12513703e-01 7.16277599e-01 -6.41678721e-02 -3.17542344e-01
1.48211524e-01 -3.36229920e-01 -1.38571426e-01 -3.72787267e-01
-9.00316238e-02 8.61561000e-02 -2.17793122e-01 -5.38985372e-01
-9.89097953e-01 -5.67764819e-01 -8.47629786e-01 -2.02586234e-01
-3.11685950e-01 3.52524400e-01 4.50691551e-01 6.73098624e-01
1.08828112e-01 2.65074760e-01 7.26423502e-01 -1.19143915e+00
-2.31364191e-01 -1.04887509e+00 -1.19847023e+00 -1.33605450e-01
7.76006520e-01 -2.80037493e-01 -3.75699133e-01 2.97169108e-02] | [10.160844802856445, -1.9290993213653564] |
d1a5b951-5eaf-4a99-be93-2b469e823b15 | benchmarking-commonsense-knowledge-base | 2109.07679 | null | https://arxiv.org/abs/2109.07679v1 | https://arxiv.org/pdf/2109.07679v1.pdf | Benchmarking Commonsense Knowledge Base Population with an Effective Evaluation Dataset | Reasoning over commonsense knowledge bases (CSKB) whose elements are in the form of free-text is an important yet hard task in NLP. While CSKB completion only fills the missing links within the domain of the CSKB, CSKB population is alternatively proposed with the goal of reasoning unseen assertions from external resources. In this task, CSKBs are grounded to a large-scale eventuality (activity, state, and event) graph to discriminate whether novel triples from the eventuality graph are plausible or not. However, existing evaluations on the population task are either not accurate (automatic evaluation with randomly sampled negative examples) or of small scale (human annotation). In this paper, we benchmark the CSKB population task with a new large-scale dataset by first aligning four popular CSKBs, and then presenting a high-quality human-annotated evaluation set to probe neural models' commonsense reasoning ability. We also propose a novel inductive commonsense reasoning model that reasons over graphs. Experimental results show that generalizing commonsense reasoning on unseen assertions is inherently a hard task. Models achieving high accuracy during training perform poorly on the evaluation set, with a large gap between human performance. We will make the data publicly available for future contributions. Codes and data are available at https://github.com/HKUST-KnowComp/CSKB-Population. | ['Bin He', 'Yangqiu Song', 'Hongming Zhang', 'Shibo Hao', 'Sehyun Choi', 'Weiqi Wang', 'Tianqing Fang'] | 2021-09-16 | null | https://aclanthology.org/2021.emnlp-main.705 | https://aclanthology.org/2021.emnlp-main.705.pdf | emnlp-2021-11 | ['knowledge-base-population'] | ['natural-language-processing'] | [ 1.41159177e-01 7.15781629e-01 -2.70387709e-01 -4.21669483e-01
-7.03710020e-01 -5.99950731e-01 7.27889657e-01 4.34676707e-01
-3.36767197e-01 1.14666367e+00 4.32535887e-01 -3.80266130e-01
-6.13655746e-02 -1.03519988e+00 -8.93048584e-01 -1.68806836e-01
2.42371038e-01 7.79203475e-01 4.73703623e-01 -5.71484685e-01
1.82927445e-01 3.49594690e-02 -1.10758507e+00 6.15781307e-01
1.05653572e+00 8.52759004e-01 -2.23480448e-01 3.31407666e-01
-6.71180338e-02 1.54977775e+00 -6.41943097e-01 -1.12648702e+00
-2.27508262e-01 -4.44938809e-01 -1.27802742e+00 -6.66399121e-01
5.30018747e-01 -1.03358798e-01 -3.41319114e-01 1.33636653e+00
3.58496606e-01 3.12454700e-01 6.60839140e-01 -1.64389670e+00
-1.33131444e+00 1.25719702e+00 1.76990733e-01 5.86293638e-01
6.28379643e-01 1.37071028e-01 1.52571690e+00 -7.54012227e-01
8.82373691e-01 1.19460475e+00 6.56836331e-01 8.84616077e-01
-1.24930060e+00 -6.80308282e-01 -1.30396098e-01 8.50698531e-01
-1.44906616e+00 -5.41164458e-01 7.92470276e-01 -1.92627162e-01
1.52640796e+00 4.66628104e-01 5.94970942e-01 1.59629989e+00
-1.27417102e-01 7.97544837e-01 1.14202619e+00 -4.36367631e-01
5.39813697e-01 8.57921094e-02 6.21736348e-01 7.75042415e-01
6.64223611e-01 -3.93058926e-01 -6.66155875e-01 -3.89581710e-01
3.43798071e-01 -3.91418755e-01 -5.92851281e-01 -2.08512187e-01
-1.45283937e+00 7.82086253e-01 7.44837761e-01 3.53358150e-01
-4.22255605e-01 3.42206538e-01 4.57736790e-01 2.67368466e-01
2.81672448e-01 6.75828099e-01 -5.11606336e-01 -1.28599957e-01
-7.15270340e-01 5.25942683e-01 1.25208974e+00 1.01983011e+00
4.73859519e-01 -2.36234859e-01 -3.38185877e-01 6.86224997e-01
1.49191007e-01 3.53551745e-01 4.89037335e-01 -8.47112358e-01
7.28244722e-01 7.70976305e-01 8.48443359e-02 -1.24288762e+00
-3.42792511e-01 -1.69882312e-01 -6.80921972e-01 -5.33066094e-01
3.95970941e-01 3.07209305e-02 -5.43813586e-01 1.99640131e+00
3.07092756e-01 2.35573053e-01 4.28938240e-01 7.69588590e-01
1.33467317e+00 4.21161115e-01 2.79576570e-01 -5.67630418e-02
1.43376768e+00 -7.15714872e-01 -7.75521219e-01 -4.13775951e-01
6.42534018e-01 7.32492432e-02 1.33976448e+00 1.71117172e-01
-9.84058738e-01 6.62513748e-02 -8.26023281e-01 -5.77775180e-01
-6.64254367e-01 -1.24264278e-01 6.74901664e-01 2.20376164e-01
-1.04307032e+00 2.19759598e-01 -5.23511589e-01 -3.93003404e-01
5.80179989e-01 -2.61027902e-01 -1.88933268e-01 -2.69213915e-01
-2.09477663e+00 1.39671123e+00 1.13407648e+00 3.60817611e-02
-8.07764351e-01 -6.79029405e-01 -1.16475880e+00 1.41651571e-01
8.48974288e-01 -7.75031328e-01 9.96959746e-01 -5.80647290e-01
-9.03075039e-01 1.19216275e+00 -2.59508584e-02 -7.98688650e-01
5.32124996e-01 1.63809992e-02 -7.52015471e-01 1.10190012e-01
3.83780748e-01 4.11184281e-01 4.69723254e-01 -1.03988552e+00
-1.02030843e-01 -2.41626233e-01 5.07077932e-01 -1.34631217e-01
-1.79207608e-01 -1.65545881e-01 5.23661971e-02 -4.93490905e-01
-5.11179678e-02 -7.24560261e-01 2.46366620e-01 -2.60076761e-01
-8.99216056e-01 -6.25977635e-01 1.43215671e-01 -7.31600165e-01
1.29207253e+00 -1.86169481e+00 9.60059911e-02 5.68628637e-03
2.90720433e-01 -8.02518129e-02 1.60775676e-01 3.59202594e-01
-1.85977936e-01 3.53544235e-01 -3.33788306e-01 1.85988545e-01
4.63685155e-01 3.41642439e-01 -7.41038799e-01 7.88843110e-02
2.89430171e-01 1.24903989e+00 -1.29582214e+00 -7.09836364e-01
-3.19493264e-01 3.68873626e-02 -4.94762272e-01 -2.37677038e-01
-6.07736826e-01 -3.06565255e-01 -2.88125694e-01 7.60400355e-01
2.17146948e-01 -6.78769886e-01 1.36524990e-01 -4.68883693e-01
6.83775663e-01 6.32489681e-01 -8.61168385e-01 1.50384879e+00
-5.09054102e-02 4.83728498e-01 -5.38404286e-01 -9.23136294e-01
6.38342798e-01 1.71486497e-01 -3.83300334e-01 -4.00586963e-01
5.58033399e-02 1.49058461e-01 2.55704224e-01 -5.22772908e-01
6.29648685e-01 -6.22215927e-01 -2.50702679e-01 2.21069902e-01
3.99707228e-01 -4.54459965e-01 5.41133702e-01 7.95118034e-01
1.37423897e+00 -1.76219821e-01 7.68112540e-01 -2.93160677e-01
4.10670012e-01 4.74224478e-01 5.48907697e-01 7.20043659e-01
-3.11869085e-01 1.31839380e-01 9.12853062e-01 -2.55066931e-01
-6.81569695e-01 -1.18147361e+00 -1.36607900e-01 9.10690010e-01
1.42402843e-01 -6.48367882e-01 -3.73974621e-01 -8.98518324e-01
1.61751732e-01 1.77268398e+00 -9.41166461e-01 -2.58000374e-01
-1.42361790e-01 -5.54566741e-01 1.05274701e+00 6.33530259e-01
6.07976437e-01 -1.52070177e+00 -3.57624739e-01 6.97305351e-02
-7.84359455e-01 -1.36254036e+00 9.22696963e-02 1.94556296e-01
-4.00445789e-01 -1.45240390e+00 9.49390978e-02 -3.76370788e-01
5.73259413e-01 -2.89455384e-01 1.52599180e+00 2.40909457e-01
-2.06110537e-01 3.93445253e-01 -4.02807713e-01 -5.27612984e-01
-4.72505063e-01 -2.53736883e-01 -6.55305162e-02 -4.57932442e-01
8.24404895e-01 -4.46088403e-01 -8.52974206e-02 -1.25718012e-01
-9.12937820e-01 2.54805014e-02 -4.89267150e-05 8.13516080e-01
3.83557588e-01 1.97147980e-01 7.73006976e-01 -9.49483514e-01
9.70890582e-01 -9.57525492e-01 -2.13387698e-01 4.31445628e-01
-4.42941248e-01 1.96378417e-02 6.05185270e-01 -2.12979525e-01
-1.03428829e+00 -7.41461933e-01 8.50644633e-02 -2.23420084e-01
2.43502334e-02 8.56778085e-01 -6.14396781e-02 5.77967227e-01
1.20402646e+00 2.90549576e-01 -3.19508880e-01 1.17749304e-01
5.27090073e-01 2.38432288e-01 5.57867467e-01 -1.15054750e+00
6.63431108e-01 2.75594354e-01 -1.89450607e-01 -5.61549723e-01
-1.47797549e+00 -1.71887249e-01 -3.31718862e-01 8.69837850e-02
7.56674945e-01 -9.31290925e-01 -5.93893290e-01 -1.17044307e-01
-1.24732971e+00 -6.24113083e-01 -5.07813215e-01 2.40730852e-01
-5.36536694e-01 2.73388118e-01 -8.42275679e-01 -4.73920017e-01
-4.36480105e-01 -5.73785543e-01 6.09285235e-01 -1.49484158e-01
-7.01453269e-01 -1.28029907e+00 5.24751097e-02 6.75322056e-01
1.86655283e-01 3.89238417e-01 1.23641729e+00 -1.31394315e+00
-2.86788166e-01 -1.20855041e-01 -3.71455342e-01 3.80222589e-01
-1.74114272e-01 -7.31745884e-02 -8.60911965e-01 2.08303452e-01
-2.78170556e-01 -1.01531804e+00 1.04703116e+00 -1.20773725e-01
1.17735279e+00 -5.39853215e-01 -2.47001156e-01 4.03608419e-02
1.38035822e+00 -4.58698153e-01 7.21022666e-01 3.96637410e-01
4.79379773e-01 4.27798450e-01 5.09815753e-01 3.36524338e-01
9.14227784e-01 1.13276049e-01 4.46823567e-01 5.85669994e-01
-5.10319546e-02 -3.78720969e-01 3.32534462e-01 4.98230457e-01
-4.00938869e-01 -4.05515611e-01 -1.38171649e+00 9.07651126e-01
-1.83210397e+00 -1.56130183e+00 -1.27858162e-01 1.59151185e+00
1.50253046e+00 2.68500507e-01 -1.84090912e-01 6.47744387e-02
6.49145663e-01 5.74681163e-02 -4.97003198e-01 -2.46050581e-01
-3.05725366e-01 1.27634823e-01 -8.85301605e-02 6.10790074e-01
-6.42443299e-01 1.16488552e+00 5.12152195e+00 8.93869340e-01
-5.41046143e-01 2.07782298e-01 2.21363917e-01 -6.92126825e-02
-7.24757969e-01 1.46009311e-01 -7.54397631e-01 4.20787513e-01
8.61506343e-01 -3.86054128e-01 5.39964974e-01 8.86492729e-01
-4.84356970e-01 -7.88083300e-02 -1.35297287e+00 8.79903734e-01
4.20349091e-01 -1.48260367e+00 2.94874430e-01 -3.17849845e-01
5.45320988e-01 2.27995276e-01 -4.36294079e-01 9.52781379e-01
7.54153967e-01 -9.74194825e-01 9.33508158e-01 5.50717831e-01
5.64010441e-01 -3.55642855e-01 6.62888169e-01 5.94581246e-01
-7.29870915e-01 7.11085498e-02 -4.39790756e-01 -1.03817299e-01
-4.93374132e-02 8.47415984e-01 -9.90350246e-01 4.34829175e-01
6.65003955e-01 7.45570362e-01 -9.12083983e-01 3.58011097e-01
-9.93788838e-01 7.30537951e-01 -1.59008458e-01 -4.49752808e-01
3.75230536e-02 3.15125614e-01 5.51545084e-01 1.37267566e+00
7.63164880e-03 5.62708974e-01 -2.16636099e-02 1.61368036e+00
-4.80593532e-01 -1.47372395e-01 -5.77317178e-01 -3.44873816e-01
6.35554254e-01 1.11752701e+00 -4.20172691e-01 -8.30567598e-01
-3.21478471e-02 7.91917026e-01 1.01659393e+00 3.58988792e-01
-1.06591797e+00 -2.76371002e-01 2.36815587e-01 -1.13886036e-01
-1.78074557e-02 2.33745426e-01 5.95678501e-02 -1.58252621e+00
1.84762657e-01 -6.37285113e-01 7.82214224e-01 -1.35241926e+00
-1.88431215e+00 4.31142867e-01 3.10488760e-01 -4.60631251e-01
-3.04386884e-01 -7.38212407e-01 -3.68335485e-01 6.93111777e-01
-1.57243097e+00 -1.22413242e+00 -5.81152976e-01 9.28743064e-01
2.09368587e-01 5.12455925e-02 9.07173634e-01 -2.96482623e-01
-3.94320935e-01 3.32860172e-01 -5.26736677e-01 4.75935519e-01
7.82941818e-01 -1.52672255e+00 2.43074253e-01 5.65871894e-01
4.65604961e-02 1.09335768e+00 8.80583644e-01 -1.04510343e+00
-1.15408897e+00 -1.06302679e+00 1.10401976e+00 -1.04490364e+00
1.31386018e+00 -1.90478832e-01 -1.18804407e+00 1.33552980e+00
1.92156896e-01 5.26861250e-01 8.91317606e-01 4.64376658e-01
-8.65403771e-01 3.81439269e-01 -1.32188129e+00 7.74057806e-01
1.32192242e+00 -7.24636734e-01 -1.52190852e+00 4.69140112e-01
6.89890504e-01 -4.90877151e-01 -8.74827147e-01 1.51401207e-01
1.03741489e-01 -6.25192106e-01 7.54209459e-01 -1.15878177e+00
9.66607153e-01 -2.98536628e-01 -4.75576818e-01 -1.52346742e+00
-2.28184521e-01 -1.98742360e-01 -4.53299373e-01 1.16782737e+00
6.97009623e-01 -7.79251456e-01 3.09679121e-01 9.19361472e-01
7.56016225e-02 -5.10290861e-01 -8.67449939e-01 -8.38444233e-01
-4.03160555e-03 -6.45205617e-01 4.78555590e-01 1.59642601e+00
6.93511069e-01 7.11625576e-01 2.75423914e-01 1.34796754e-01
7.47189462e-01 1.34468228e-01 3.74969363e-01 -1.21040273e+00
-1.89633593e-01 -2.42655337e-01 -4.02728647e-01 -2.53359556e-01
7.89722800e-01 -1.51252019e+00 -2.65065767e-02 -1.95275819e+00
6.26561284e-01 -3.42377186e-01 -2.58334458e-01 1.02425969e+00
-3.66325825e-01 5.80980554e-02 -6.24059141e-02 -5.14600202e-02
-8.32707644e-01 4.95477796e-01 1.13305163e+00 -3.26949358e-01
2.20378384e-01 -7.62068570e-01 -8.15656424e-01 8.69551659e-01
9.71092582e-01 -3.03320497e-01 -5.91898084e-01 -1.76441103e-01
9.66980398e-01 -7.86928907e-02 1.16104090e+00 -6.60657465e-01
3.39017332e-01 -3.82076561e-01 1.99585095e-01 -2.70880967e-01
2.88365155e-01 -5.32406926e-01 -8.78824145e-02 1.86811343e-01
-7.24596858e-01 -1.97202742e-01 2.86642164e-01 6.40221953e-01
-1.21449873e-01 -2.91635752e-01 4.01603937e-01 -2.87956595e-01
-9.43848073e-01 -2.06496000e-01 3.82165074e-01 9.95355248e-01
9.00119305e-01 8.73835534e-02 -9.54136550e-01 -2.18129560e-01
-9.65742350e-01 1.51216453e-02 3.30381989e-01 1.69023857e-01
7.79160917e-01 -1.24268937e+00 -9.22895849e-01 -4.87834781e-01
5.61149478e-01 3.38032357e-02 2.59721428e-01 9.61566746e-01
-3.44663888e-01 2.73830384e-01 -1.75831154e-01 -2.33981106e-02
-8.32241952e-01 7.18062341e-01 1.91152424e-01 -1.83842734e-01
-5.38697481e-01 1.01914847e+00 -2.77971625e-01 -5.69187403e-01
-2.05611736e-01 -6.02850854e-01 -6.13315366e-02 6.55019656e-03
5.47890544e-01 2.58423239e-01 1.25988182e-02 -4.19907629e-01
-6.59134328e-01 -2.50962675e-01 -7.79915377e-02 1.16723560e-01
1.28399968e+00 2.74695396e-01 -6.42989635e-01 5.58600426e-01
6.01035297e-01 1.11811295e-01 -3.42138767e-01 -4.04341936e-01
9.12554860e-02 -1.77522406e-01 -1.57712683e-01 -1.25404704e+00
-4.83049482e-01 5.34541249e-01 -5.30508697e-01 3.76869977e-01
5.20347178e-01 4.15002584e-01 4.27967489e-01 8.24540675e-01
4.39910680e-01 -1.04823744e+00 1.07396789e-01 7.94329762e-01
1.43616331e+00 -1.36075342e+00 8.28780755e-02 -4.85291511e-01
-9.38033223e-01 8.73698533e-01 8.16070795e-01 -8.19308916e-04
2.71349549e-01 4.98794802e-02 -5.11432350e-01 -6.89617634e-01
-1.03846276e+00 -5.72354458e-02 2.89875239e-01 4.15983707e-01
4.58005846e-01 2.50976384e-01 -2.16754004e-01 1.06805789e+00
-5.26240528e-01 3.79895307e-02 6.96741104e-01 7.84389019e-01
-3.02490771e-01 -2.40144670e-01 -1.43755719e-01 6.31748974e-01
-1.46261796e-01 -6.84241056e-01 -8.42913687e-01 9.30511773e-01
-6.75620511e-02 8.59323621e-01 -3.98759156e-01 -2.04682257e-02
2.87701219e-01 4.76609319e-01 6.49838686e-01 -7.49795079e-01
-2.73363531e-01 -1.02723336e+00 7.76942372e-01 -5.84357440e-01
-4.18116242e-01 -4.69158828e-01 -1.57573915e+00 -5.29377520e-01
-3.15669000e-01 1.00414231e-01 1.32044673e-01 1.12036920e+00
1.84492737e-01 5.29879153e-01 -4.76808369e-01 -7.87204504e-02
-8.19451451e-01 -9.61894095e-01 -5.09492815e-01 7.28976727e-01
1.18327834e-01 -6.76582992e-01 -5.88502407e-01 2.08856285e-01] | [9.905088424682617, 8.126697540283203] |
35432d63-239e-4d53-aed6-375aaa9c3444 | ceres-pretraining-of-graph-conditioned | null | null | https://openreview.net/forum?id=r_2z0-tufhS | https://openreview.net/pdf?id=r_2z0-tufhS | CERES: Pretraining of Graph-Conditioned Transformer for Semi-Structured Session Data | User sessions empower many search and recommendation tasks on a daily basis. Such session data are semi-structured, which encode heterogeneous relations between queries and products, and each item is described by the unstructured text. Despite recent advances in self-supervised learning for text or graphs, there lack of self-supervised learning models that can effectively capture both intra-item semantics and inter-item interactions for semi-structured sessions. To fill this gap, we propose CERES, a graph-based transformer model for semi-structured session data. CERES learns representations that capture both inter- and intra-item semantics with (1) a graph-conditioned masked language pretraining task that jointly learns from item text and item-item relations; and (2) a graph-conditioned transformer architecture that propagates inter-item contexts to item-level representations. We pretrained CERES using ~468 million Amazon sessions and find that CERES outperforms strong pretraining baselines by up to 9% in three session search and entity linking tasks. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['session-search'] | ['natural-language-processing'] | [ 3.04909348e-01 3.27694625e-01 -7.53578126e-01 -7.10115910e-01
-4.59831536e-01 -8.35481822e-01 7.44864941e-01 5.64811766e-01
-1.09553300e-01 3.10212702e-01 5.94364405e-01 -3.68182153e-01
-2.14347348e-01 -8.27219784e-01 -8.98318052e-01 1.35080501e-01
-3.06832850e-01 9.83140171e-01 1.88862726e-01 -5.30466020e-01
-3.26455444e-01 -1.13128945e-01 -1.11349821e+00 7.91728616e-01
7.43872225e-01 1.00758672e+00 4.56410386e-02 5.92264652e-01
-6.42302275e-01 8.93034399e-01 -1.67527407e-01 -6.97428167e-01
1.01687955e-02 -6.35485411e-01 -9.24722731e-01 2.49643117e-01
6.31822705e-01 2.97414921e-02 -5.41550636e-01 6.30546749e-01
5.49515225e-02 2.90652603e-01 7.14317679e-01 -1.21341646e+00
-1.31085718e+00 1.25536335e+00 -5.03138542e-01 2.13482618e-01
6.03237212e-01 -5.50125614e-02 1.87625074e+00 -6.22270405e-01
8.56887281e-01 1.06253076e+00 5.97072482e-01 3.70905817e-01
-1.60058784e+00 -5.10810196e-01 4.59838092e-01 -1.54689059e-01
-1.09865344e+00 -1.61169663e-01 6.60416484e-01 -2.52687365e-01
1.61898208e+00 8.50742683e-02 7.01995552e-01 1.49634004e+00
8.32222402e-02 8.47842813e-01 6.74822628e-01 -1.84867293e-01
-2.79118001e-01 3.60251784e-01 7.19104230e-01 9.54027891e-01
4.72972244e-01 -7.36889020e-02 -9.95331466e-01 -2.07529187e-01
8.12627435e-01 1.56504720e-01 2.45808035e-01 -6.90633357e-01
-8.36165309e-01 9.22773182e-01 8.07561457e-01 4.27788377e-01
-2.55582541e-01 2.16179103e-01 4.67441112e-01 7.07754672e-01
5.94398022e-01 6.31264627e-01 -6.46016598e-01 3.85607094e-01
-7.61545360e-01 1.08601682e-01 1.22009826e+00 1.49783754e+00
7.88453817e-01 -4.20497730e-02 -1.49650425e-01 8.68772566e-01
6.30862057e-01 4.65170622e-01 3.77562582e-01 -1.98926494e-01
7.11746991e-01 8.90678227e-01 -2.30611756e-01 -9.13592994e-01
-4.58653659e-01 -7.62649775e-01 -6.13067269e-01 -8.85505855e-01
1.52679920e-01 2.16753736e-01 -9.86123204e-01 1.79700601e+00
-2.39351038e-02 9.34526771e-02 -1.91553026e-01 6.11484468e-01
1.52878165e+00 3.31505686e-01 5.93002379e-01 5.02375420e-03
1.45191681e+00 -1.31878185e+00 -8.37878942e-01 -6.16369128e-01
8.64901245e-01 -4.83494848e-01 1.27121031e+00 -2.66633600e-01
-1.14523375e+00 -5.23354113e-01 -8.62379849e-01 -4.41633314e-01
-8.70523572e-01 -1.74740046e-01 1.11483359e+00 1.77813366e-01
-1.09955037e+00 3.30525666e-01 -6.36526644e-01 -7.52603412e-01
3.47912580e-01 5.46344697e-01 -3.61995608e-01 6.85982108e-02
-1.39044201e+00 5.08094847e-01 1.65261388e-01 -2.85347342e-01
-5.35895646e-01 -1.12734127e+00 -1.26470256e+00 3.84738237e-01
4.86188859e-01 -1.07302678e+00 1.19828260e+00 -8.07295799e-01
-1.04498875e+00 1.14629376e+00 -3.16570014e-01 -6.76155865e-01
-2.31780663e-01 -2.04259679e-01 -8.65912020e-01 -2.68738955e-01
1.21034041e-01 2.38968357e-01 6.16917133e-01 -1.10745192e+00
-2.06436992e-01 -5.56299448e-01 1.90928474e-01 5.36779225e-01
-4.41642761e-01 -9.37149003e-02 -9.06736791e-01 -6.34839833e-01
-2.30521727e-02 -9.27724183e-01 -3.51766646e-02 -3.86284560e-01
-6.34355128e-01 -6.31147504e-01 3.82516444e-01 -4.16829288e-01
1.40527391e+00 -1.94047356e+00 2.20243841e-01 3.08181167e-01
4.47211385e-01 -2.84648649e-02 -5.05964220e-01 9.04135466e-01
-7.50468671e-02 3.83294597e-02 3.58698636e-01 -8.35413218e-01
3.03698570e-01 2.50985354e-01 -3.53579283e-01 -5.97759262e-02
-8.78720172e-03 1.63759577e+00 -9.85890806e-01 -3.57166588e-01
-6.72604740e-02 2.90824711e-01 -4.94550973e-01 4.95147169e-01
-5.54720640e-01 2.59504877e-02 -4.81733590e-01 7.33161569e-01
1.88298061e-01 -1.14240110e+00 6.01680994e-01 -4.51614529e-01
6.82519376e-01 9.31881309e-01 -7.80364394e-01 2.24025226e+00
-6.05976701e-01 1.74374893e-01 -3.93345058e-02 -6.84966862e-01
4.71372128e-01 1.22403152e-01 6.10098600e-01 -1.11496842e+00
-2.41233528e-01 -1.48537718e-02 -4.92508680e-01 -3.43751192e-01
8.63023221e-01 2.28100330e-01 -4.90536869e-01 9.57548797e-01
7.65160263e-01 6.61076754e-02 2.36082390e-01 1.16989446e+00
1.21125627e+00 7.06160665e-02 1.10957466e-01 -9.47959721e-02
1.65475890e-01 -1.36647835e-01 -5.78030944e-02 8.92321408e-01
3.89681548e-01 1.81502551e-01 2.95156598e-01 -2.12236539e-01
-6.53977811e-01 -1.25090146e+00 4.53214735e-01 1.92107618e+00
2.87871361e-01 -8.86324644e-01 -5.90508394e-02 -1.12517428e+00
6.09834731e-01 5.93865454e-01 -8.08847547e-01 -3.52568835e-01
-1.91034734e-01 -1.71267956e-01 1.73964590e-01 8.09962094e-01
1.14166029e-01 -1.15490305e+00 5.21908343e-01 2.07239673e-01
-4.28007990e-02 -1.32587385e+00 -1.18750966e+00 3.61585170e-01
-7.51016200e-01 -1.10936022e+00 -1.84643090e-01 -9.89398956e-01
5.92797160e-01 5.34403563e-01 2.06714487e+00 1.68304741e-01
-2.00434417e-01 7.82474041e-01 -2.09404781e-01 -3.05415113e-02
-1.14844620e-01 5.07568121e-01 -1.05141528e-01 -2.31475681e-01
8.48659754e-01 -7.13495970e-01 -7.32145846e-01 3.07960421e-01
-6.44385338e-01 4.50560637e-02 5.20822585e-01 6.05134487e-01
5.02155781e-01 -5.34491539e-01 5.74272335e-01 -1.95621836e+00
8.82019222e-01 -9.88303542e-01 -3.28199476e-01 7.23249257e-01
-1.05794358e+00 1.01139292e-01 2.57274479e-01 -4.02583957e-01
-9.66580093e-01 -7.71916881e-02 2.57208228e-01 -1.43618166e-01
5.54940552e-02 7.97991574e-01 1.51703954e-01 7.61118755e-02
8.39042902e-01 1.62097842e-01 -2.59603947e-01 -5.68129063e-01
9.60219562e-01 4.40738827e-01 2.49221310e-01 -3.94051969e-01
7.84006596e-01 1.45810083e-01 -3.49927276e-01 -3.97957087e-01
-1.49231446e+00 -1.05342007e+00 -4.05509651e-01 2.64497340e-01
7.71939576e-01 -1.26596725e+00 -7.18052983e-01 -8.30783136e-03
-6.38573885e-01 -6.69758379e-01 -6.60117567e-01 1.40373707e-01
-3.20717245e-01 1.88639507e-01 -1.03302109e+00 -3.49699736e-01
-6.19076550e-01 -3.72288436e-01 1.26709509e+00 1.52052298e-01
-3.95084351e-01 -1.61797380e+00 1.47147477e-01 6.71590269e-01
6.19773209e-01 -3.13192546e-01 8.63007069e-01 -1.08547056e+00
-4.38094109e-01 -1.06397152e-01 -4.23193842e-01 -1.05921201e-01
3.76657456e-01 -6.87055171e-01 -5.16412497e-01 -3.94101083e-01
-6.15606368e-01 -8.08204353e-01 9.96136248e-01 -2.89530475e-02
1.01376963e+00 -3.43189389e-01 -6.09912574e-01 5.85559249e-01
1.20236778e+00 -4.21440870e-01 7.16593713e-02 -2.03006476e-01
1.02626753e+00 3.22599143e-01 1.88540906e-01 5.99353351e-02
9.36667621e-01 6.91866696e-01 1.45080000e-01 -1.52362958e-01
-5.18007517e-01 -1.03304172e+00 1.73844635e-01 8.78642499e-01
3.57481629e-01 -5.69448292e-01 -3.25748712e-01 6.18044078e-01
-2.03235483e+00 -9.04361784e-01 -1.50779426e-01 1.91416156e+00
1.09407341e+00 1.00483827e-01 3.58440161e-01 -7.91639805e-01
1.74019098e-01 4.36778128e-01 -9.01309907e-01 -2.37152770e-01
-2.70138770e-01 2.76594281e-01 5.18767238e-01 4.64087963e-01
-1.04490232e+00 1.19986451e+00 6.42852497e+00 4.41066206e-01
-6.11798286e-01 8.65323767e-02 2.44904449e-03 9.08858888e-03
-8.97872746e-01 -1.79451212e-01 -9.22968388e-01 3.23709011e-01
1.01839268e+00 -3.56052727e-01 6.08536005e-01 7.47205079e-01
-3.59963745e-01 4.85756397e-01 -1.47874236e+00 9.25602019e-01
3.67369771e-01 -1.44243693e+00 2.66075820e-01 1.86777607e-01
9.44731414e-01 1.86192289e-01 2.59013146e-01 8.64052057e-01
1.05204439e+00 -1.09086776e+00 1.22653998e-01 3.06135714e-01
7.77466893e-01 -5.31119704e-02 2.96897620e-01 7.16180131e-02
-1.64040124e+00 1.25349119e-01 -5.16047850e-02 4.32047814e-01
1.89603463e-01 2.10710868e-01 -9.35333490e-01 8.12496424e-01
5.09638250e-01 1.34021926e+00 -8.72003973e-01 6.18423522e-01
-3.54286790e-01 7.98525035e-01 -2.95970321e-01 -3.01423464e-02
3.26394349e-01 -2.97564775e-01 1.48171261e-01 1.40554440e+00
-1.12050667e-01 -4.50737821e-03 3.25929880e-01 8.64331663e-01
-6.90397501e-01 2.18084320e-01 -8.76716793e-01 -8.19489300e-01
4.27283704e-01 1.31226993e+00 -4.75422144e-01 -3.72099668e-01
-8.52840066e-01 1.26765323e+00 6.94380879e-01 7.64040828e-01
-5.35567462e-01 -3.21819112e-02 6.38475478e-01 4.88909215e-01
4.32983637e-01 -2.61045784e-01 -1.21733665e-01 -1.48424804e+00
-3.17449898e-01 -7.69459426e-01 1.12047756e+00 -7.30277121e-01
-1.99336994e+00 5.64863384e-01 -2.18431935e-01 -6.37361348e-01
-5.62989652e-01 -3.32951427e-01 -4.10727322e-01 8.27480137e-01
-1.43951929e+00 -1.74798167e+00 -1.46784320e-01 1.05079067e+00
6.67763889e-01 -3.14924717e-01 1.09450376e+00 2.61877477e-01
-1.04797043e-01 9.47441399e-01 -2.29412869e-01 2.27842778e-01
7.65702784e-01 -1.54459417e+00 6.86979234e-01 1.59470633e-01
9.96294796e-01 1.13543510e+00 4.57462162e-01 -8.37273717e-01
-1.68329692e+00 -1.13062179e+00 1.34345067e+00 -8.80060852e-01
1.11249840e+00 -8.71797383e-01 -9.41721141e-01 1.44496560e+00
5.42588949e-01 2.23972738e-01 1.23233163e+00 1.24338579e+00
-9.97893035e-01 1.30564151e-02 -6.43361986e-01 3.67417663e-01
1.74196315e+00 -1.24363112e+00 -7.54270375e-01 9.38350737e-01
1.13041031e+00 -4.39621121e-01 -1.09342611e+00 1.13398343e-01
2.77311772e-01 -4.35606033e-01 1.09133375e+00 -1.35980642e+00
1.41486630e-01 2.99460500e-01 -2.20167101e-03 -1.27361131e+00
-5.57851732e-01 -7.22106278e-01 -8.21210504e-01 1.12255752e+00
9.19234455e-01 -4.11770999e-01 1.03606856e+00 6.14402831e-01
-4.08054665e-02 -5.83770931e-01 -1.25087321e-01 -6.83980823e-01
-3.34039778e-01 -3.00842263e-02 5.31124592e-01 1.09608150e+00
3.07219535e-01 1.21985447e+00 -3.31712157e-01 -4.90656868e-02
7.26861894e-01 6.50129974e-01 6.00894988e-01 -1.28562176e+00
-7.13927448e-01 -1.47970259e-01 -7.35288672e-03 -1.43766320e+00
4.71511692e-01 -1.28217089e+00 -3.95948082e-01 -2.04524565e+00
5.43155909e-01 -4.36240017e-01 -6.24076247e-01 6.15148067e-01
-3.37820537e-02 2.04731271e-01 -2.62000769e-01 1.25408635e-01
-1.48674965e+00 3.96852821e-01 1.02597106e+00 -1.93117589e-01
-2.89688528e-01 1.10423371e-01 -1.03586912e+00 1.48761481e-01
2.82422423e-01 -2.13090077e-01 -1.05900240e+00 -7.96605408e-01
6.98190749e-01 1.63399875e-02 -6.04078174e-03 -1.33294955e-01
6.07928157e-01 1.03694305e-01 1.64702997e-01 -4.51242179e-01
4.37768459e-01 -8.08317721e-01 2.34218761e-02 -1.05124533e-01
-9.95900273e-01 -1.90837517e-01 8.65985826e-02 1.15617418e+00
-1.33825809e-01 4.34609205e-01 -1.07713416e-01 -2.13386953e-01
-7.27128208e-01 7.19857156e-01 2.03278497e-01 3.22480530e-01
4.78398293e-01 7.19287097e-02 -4.58120525e-01 -6.86029196e-01
-1.18855512e+00 5.57110429e-01 6.21189401e-02 9.50567603e-01
2.26464212e-01 -1.39194608e+00 -3.21605295e-01 2.04418838e-01
5.26386321e-01 -3.36115330e-01 2.71724313e-01 2.83697724e-01
2.99584568e-01 7.59845793e-01 2.45279849e-01 -4.34484005e-01
-1.32070279e+00 7.35185504e-01 3.07333916e-02 -8.18533957e-01
-3.79954606e-01 1.29477465e+00 2.25542709e-01 -6.36532187e-01
4.44591522e-01 -9.84948948e-02 -3.35786849e-01 1.01936590e-02
1.49196818e-01 -3.29166561e-01 4.48157564e-02 -4.67786878e-01
-2.65873224e-01 2.67265737e-01 -5.52228510e-01 1.88041344e-01
1.45271599e+00 -3.35645527e-01 -3.75310332e-02 6.52418613e-01
1.38074076e+00 -6.14755675e-02 -8.17911685e-01 -1.07242155e+00
2.70276934e-01 -2.29399011e-01 -8.27950835e-02 -1.04202855e+00
-1.02083969e+00 3.59984964e-01 -1.60893742e-02 6.27750337e-01
7.38645077e-01 6.61630869e-01 9.49532926e-01 5.48171520e-01
4.83454973e-01 -7.91010261e-01 5.16625702e-01 4.87882316e-01
8.62710118e-01 -1.44182050e+00 1.24416091e-01 -7.69180477e-01
-9.26804066e-01 5.17294943e-01 8.03544343e-01 -1.36295289e-01
1.07570577e+00 -7.10782930e-02 -3.15602690e-01 -8.69819164e-01
-1.14119935e+00 -3.87575328e-01 9.50273871e-01 4.04443622e-01
9.32397604e-01 -8.99298023e-03 -3.25058997e-02 8.34261298e-01
-2.32932717e-02 -9.89135206e-02 -3.15753192e-01 6.23338938e-01
1.88968368e-02 -1.35408831e+00 8.26363385e-01 7.44769096e-01
-7.04246238e-02 -4.59615320e-01 -7.19645619e-01 5.11966467e-01
-3.36382985e-01 9.52557504e-01 1.46146566e-01 -4.54275548e-01
5.12558937e-01 2.59912252e-01 4.11435753e-01 -1.31868565e+00
-1.10157299e+00 -1.71246812e-01 4.48818058e-01 -9.84665990e-01
-2.02451929e-01 -4.28919524e-01 -1.07513797e+00 -2.22853377e-01
-3.04119468e-01 4.82244194e-01 5.04184425e-01 7.63771355e-01
8.72607052e-01 6.62052453e-01 4.08061147e-01 -1.87069878e-01
-5.89767218e-01 -9.86817241e-01 -8.97099972e-01 9.30153489e-01
-4.58405241e-02 -3.31178576e-01 -2.18698055e-01 -6.72149435e-02] | [10.953429222106934, 7.308383941650391] |
359f6e88-3772-4bfd-bf6c-ad32c35105eb | discovering-stochastic-partial-differential | 2306.15873 | null | https://arxiv.org/abs/2306.15873v1 | https://arxiv.org/pdf/2306.15873v1.pdf | Discovering stochastic partial differential equations from limited data using variational Bayes inference | We propose a novel framework for discovering Stochastic Partial Differential Equations (SPDEs) from data. The proposed approach combines the concepts of stochastic calculus, variational Bayes theory, and sparse learning. We propose the extended Kramers-Moyal expansion to express the drift and diffusion terms of an SPDE in terms of state responses and use Spike-and-Slab priors with sparse learning techniques to efficiently and accurately discover the underlying SPDEs. The proposed approach has been applied to three canonical SPDEs, (a) stochastic heat equation, (b) stochastic Allen-Cahn equation, and (c) stochastic Nagumo equation. Our results demonstrate that the proposed approach can accurately identify the underlying SPDEs with limited data. This is the first attempt at discovering SPDEs from data, and it has significant implications for various scientific applications, such as climate modeling, financial forecasting, and chemical kinetics. | ['Souvik Chakraborty', 'Rajdip Nayek', 'Tapas Tripura', 'Yogesh Chandrakant Mathpati'] | 2023-06-28 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [ 2.35100567e-01 -5.96472800e-01 2.90080100e-01 1.20441265e-01
-5.77487826e-01 -3.84207904e-01 6.54625475e-01 -1.41034245e-01
-1.99669182e-01 1.11665320e+00 -2.64519334e-01 -1.99313059e-01
-2.83694565e-01 -3.62952977e-01 -4.21384484e-01 -1.36624384e+00
-1.74682543e-01 3.97409469e-01 1.61505744e-01 5.19221947e-02
4.98584419e-01 8.95077825e-01 -1.38624620e+00 -1.13739625e-01
7.69678175e-01 1.26064146e+00 3.11660856e-01 6.21640027e-01
-4.35825408e-01 9.55279529e-01 -2.89672881e-01 2.55754590e-01
-6.78949570e-03 -4.41457838e-01 -2.25774944e-01 -1.83351919e-01
-4.21240509e-01 1.59241736e-01 -2.49368459e-01 9.78404403e-01
3.17833215e-01 3.35539579e-01 1.15396571e+00 -9.13452864e-01
-4.91752237e-01 1.33111045e-01 -6.82631135e-01 6.96817756e-01
5.10661071e-03 2.25867510e-01 6.86429381e-01 -9.96192873e-01
3.91390353e-01 1.05368364e+00 7.48416901e-01 5.38760245e-01
-1.40142822e+00 -6.64988875e-01 1.12764560e-01 -9.46034342e-02
-1.73486876e+00 -3.65765601e-01 9.12528098e-01 -8.66718113e-01
9.11399543e-01 6.61077201e-02 5.01454055e-01 1.02667582e+00
5.62390566e-01 8.48573327e-01 1.46162856e+00 -1.11910634e-01
1.00128186e+00 3.58285569e-02 1.73632115e-01 4.32363808e-01
3.30846816e-01 9.29784402e-02 -4.85891640e-01 -5.65424263e-01
8.51215422e-01 1.58888474e-01 -1.62126243e-01 -3.54615510e-01
-6.80499196e-01 7.71075308e-01 -3.23279142e-01 3.69219065e-01
-9.59652305e-01 1.65786594e-01 -1.45583516e-02 -2.13280305e-01
5.01661897e-01 2.55040199e-01 -4.11271602e-01 -2.07666382e-02
-1.20194864e+00 5.36549747e-01 1.13685191e+00 6.42740369e-01
7.27703691e-01 6.35733902e-01 1.20473251e-01 4.30634528e-01
3.78663391e-01 1.13729239e+00 2.45274082e-01 -1.25182223e+00
-6.24260381e-02 -1.68109804e-01 5.54278970e-01 -6.74696863e-01
-1.88833207e-01 -3.05156857e-01 -1.19445503e+00 -1.19082905e-01
6.42034486e-02 -5.14113367e-01 -6.48914635e-01 1.35486197e+00
2.09176064e-01 9.61293101e-01 3.85307789e-01 7.10426748e-01
4.53408629e-01 1.05993652e+00 -2.27202162e-01 -7.13780224e-01
6.60813153e-01 1.27170961e-02 -1.02582705e+00 1.95313409e-01
1.82656944e-01 -3.70507270e-01 3.11503530e-01 3.25555116e-01
-1.07771671e+00 1.02891894e-02 -6.74016237e-01 3.94171864e-01
-2.92773964e-03 -2.02753365e-01 7.04572856e-01 3.82648349e-01
-7.99961567e-01 6.39903605e-01 -1.30696595e+00 -1.49653703e-01
2.34140426e-01 -1.03831440e-02 3.46558481e-01 2.02568144e-01
-9.18401003e-01 5.45877039e-01 -1.92728341e-01 1.54229149e-01
-1.07404649e+00 -8.53374064e-01 -6.78034306e-01 7.74851814e-02
-2.39317166e-03 -6.47868812e-01 8.82768869e-01 -2.23518893e-01
-1.77113903e+00 4.99506682e-01 -7.98572958e-01 -6.29659355e-01
7.40937516e-02 2.17325717e-01 -3.67125869e-01 5.01463823e-02
-6.86206529e-03 -1.72772959e-01 6.92722559e-01 -1.26705217e+00
-3.27215642e-01 -3.39865029e-01 -5.43545544e-01 -1.49390116e-01
2.72901416e-01 -1.45844579e-01 4.24810857e-01 -5.28866529e-01
5.60683191e-01 -7.59883225e-01 -3.55966032e-01 -1.82160780e-01
-1.57768130e-01 -6.86839670e-02 6.82170510e-01 -4.54499543e-01
1.01221406e+00 -2.06074929e+00 3.66590887e-01 3.41641247e-01
2.76467223e-02 1.72408689e-02 2.87003458e-01 5.79411387e-01
2.15893120e-01 -5.51095828e-02 -8.64043176e-01 -3.85181308e-01
-1.96118236e-01 3.49388480e-01 -6.95527494e-01 6.24731421e-01
2.88877547e-01 5.17883718e-01 -8.01274896e-01 -1.31861836e-01
6.11526370e-02 6.23332083e-01 -1.26430526e-01 -1.64046474e-02
-1.91489637e-01 8.17564547e-01 -7.51209676e-01 4.98197079e-01
8.00599694e-01 -2.67244756e-01 -1.69107750e-01 2.84883112e-01
-5.32257318e-01 -6.16186298e-03 -1.55094504e+00 1.32784808e+00
-3.80730629e-01 5.06350100e-01 3.34906250e-01 -1.36104321e+00
1.20133805e+00 4.20521021e-01 9.58659887e-01 -4.11309093e-01
-6.49943724e-02 3.59254777e-01 -2.90691733e-01 -5.25328398e-01
1.21247642e-01 -7.16269612e-01 3.54372174e-01 6.79278374e-01
1.61560267e-01 -3.63938272e-01 1.80327639e-01 1.35324076e-01
9.13980782e-01 3.30324434e-02 1.22104228e-01 -7.93859780e-01
7.83653915e-01 -3.70649397e-01 8.94190729e-01 6.29274547e-01
-1.71672195e-01 2.52498895e-01 5.44711947e-01 -4.14905697e-01
-8.88833761e-01 -1.08568478e+00 -3.15522850e-01 2.41059110e-01
5.11785001e-02 9.12897587e-02 -5.16898572e-01 3.17627132e-01
2.23137587e-01 1.02207661e+00 -4.59333092e-01 -3.94039713e-02
-2.86120623e-01 -1.17041910e+00 4.81349289e-01 2.92236596e-01
3.06216508e-01 -6.40548170e-01 -4.62982535e-01 3.69557709e-01
5.84799564e-03 -1.09483480e+00 -3.19242701e-02 2.81688899e-01
-1.13466311e+00 -8.34061027e-01 -9.70786631e-01 -3.90611023e-01
2.91391879e-01 -1.01882480e-01 6.66366041e-01 -6.73209071e-01
-3.95111263e-01 6.79122210e-01 2.24626482e-01 -5.48610210e-01
-2.81440616e-01 -4.51458514e-01 4.64061201e-01 4.75850731e-01
4.31897730e-01 -8.75801146e-01 -4.17886496e-01 6.29614294e-02
-7.61496484e-01 -5.57328880e-01 1.61397040e-01 4.91293877e-01
1.05293643e+00 3.29178751e-01 3.62892658e-01 -6.41501963e-01
6.86813772e-01 -7.42708623e-01 -1.04462636e+00 -4.94806953e-02
-5.74113369e-01 4.21709925e-01 2.64621854e-01 -2.94122696e-01
-1.37635684e+00 1.27633095e-01 -1.87844317e-02 -6.48080409e-01
-1.15374647e-01 5.96091509e-01 4.58070308e-01 -8.78148451e-02
4.39491779e-01 8.96466792e-01 -5.83751686e-02 -6.82330132e-01
-1.96580529e-01 4.21011537e-01 3.21421206e-01 -8.38018239e-01
3.83666128e-01 8.90377402e-01 5.52621543e-01 -1.44152784e+00
-3.72587293e-01 -5.53767323e-01 -4.62838024e-01 8.41624737e-02
7.92068839e-01 -8.60619307e-01 -9.09667015e-01 7.98172534e-01
-1.20430195e+00 -2.46714965e-01 -4.47996020e-01 6.45882666e-01
-6.65900946e-01 3.15822601e-01 -8.38791132e-01 -1.75999558e+00
-1.39309108e-01 -7.06431389e-01 1.02540636e+00 1.64483562e-01
5.90805188e-02 -1.31846726e+00 6.22853160e-01 -2.46968955e-01
6.11486256e-01 2.55219102e-01 7.58745074e-01 -3.54157060e-01
-6.23072326e-01 -1.54206514e-01 1.45407826e-01 3.64207715e-01
-3.66752371e-02 2.92257577e-01 -8.97736013e-01 -1.41321868e-01
7.47561276e-01 9.98286903e-02 1.00689912e+00 9.73798335e-01
8.09558570e-01 -4.49955054e-02 -3.63186151e-01 6.10447705e-01
1.50124729e+00 4.54603046e-01 3.32758158e-01 -4.14719671e-01
3.55321825e-01 3.54640663e-01 -1.70340389e-01 9.74873662e-01
2.30752841e-01 5.67456149e-02 -1.00844204e-02 3.97385925e-01
4.97307718e-01 -9.69619863e-03 4.30940628e-01 9.86321211e-01
-1.58271164e-01 -8.87317955e-02 -1.06819916e+00 5.73791921e-01
-1.73123050e+00 -1.22453904e+00 -2.52257019e-01 2.01629043e+00
3.74949425e-01 -2.63493478e-01 -2.04259977e-01 -2.08958481e-02
6.09725356e-01 1.68679669e-01 -8.05014729e-01 -3.28650922e-01
-4.50065404e-01 4.80597168e-01 5.35451591e-01 5.96561074e-01
-6.80348396e-01 5.51070273e-01 7.23110580e+00 3.80207092e-01
-1.27815259e+00 2.25351512e-01 3.08970034e-01 5.54402657e-02
-4.48583990e-01 -6.87005445e-02 -9.30103362e-01 6.14163637e-01
1.02073431e+00 -3.71789575e-01 6.12993419e-01 3.22533250e-01
4.79063660e-01 -3.36153597e-01 -8.29348981e-01 1.18837619e+00
-1.79759830e-01 -1.34044898e+00 -2.73749799e-01 4.28993553e-02
1.08243346e+00 1.67422965e-01 3.01426686e-02 -8.29263777e-02
4.38312680e-01 -7.37248480e-01 2.91413486e-01 1.28280318e+00
3.03513944e-01 -4.52844322e-01 6.49348080e-01 6.36349440e-01
-1.15874016e+00 3.16286832e-02 -4.79029864e-01 -3.70481193e-01
3.61931413e-01 1.09233713e+00 -1.93778500e-01 1.39235750e-01
5.56343138e-01 8.74404132e-01 3.91764879e-01 9.34106231e-01
2.63888929e-02 9.15807545e-01 -6.28843606e-01 -1.81412309e-01
1.20512106e-01 -6.86310947e-01 8.94156992e-01 9.23894525e-01
7.87294626e-01 5.00689387e-01 -2.04235405e-01 1.46170461e+00
4.26911384e-01 -1.59645036e-01 -6.12815142e-01 -4.64024961e-01
4.14710194e-01 6.55482233e-01 -7.22497821e-01 -2.70230353e-01
-3.35136443e-01 7.76864290e-01 -4.16161060e-01 1.09233499e+00
-4.84520942e-01 -1.50698975e-01 9.31311309e-01 4.42607887e-02
7.23645031e-01 -4.85361516e-01 -2.74693340e-01 -1.37861478e+00
-1.91997901e-01 -2.52156168e-01 1.80709928e-01 -6.62629604e-01
-1.40341532e+00 1.34848058e-01 -9.78875011e-02 -6.94164991e-01
-2.39635527e-01 -5.67500770e-01 -7.62748778e-01 1.32424355e+00
-1.52582514e+00 -2.97765195e-01 2.43558556e-01 6.45447433e-01
2.03112200e-01 -3.23706269e-01 8.51248384e-01 2.53500324e-02
-7.48595774e-01 -3.30876648e-01 1.02439427e+00 -3.44644248e-01
-4.36402783e-02 -1.09239018e+00 2.10532188e-01 7.01104701e-01
-9.20184050e-03 5.57017267e-01 1.08337939e+00 -6.73497140e-01
-1.54614806e+00 -7.58350551e-01 6.70007586e-01 -2.16797784e-01
8.94819558e-01 -2.64218301e-01 -1.12959349e+00 3.27904165e-01
-1.45010754e-01 1.74474582e-01 6.19669855e-01 -1.53691500e-01
1.97001740e-01 -2.10833639e-01 -1.13098812e+00 -6.74286187e-02
5.16968131e-01 -8.37113023e-01 -4.52344984e-01 4.47629601e-01
-2.25238018e-02 -1.41125396e-01 -7.56694734e-01 4.82546315e-02
3.40155900e-01 -5.26767910e-01 7.52859354e-01 -5.92166483e-01
1.75865188e-01 -4.03539449e-01 -4.18437302e-01 -1.22957802e+00
-3.40780437e-01 -8.22042167e-01 -5.54421306e-01 9.70218837e-01
1.36742741e-01 -8.05401981e-01 5.55076480e-01 7.29750097e-01
1.62869766e-01 -6.54052198e-01 -1.14799786e+00 -9.17800486e-01
2.73000240e-01 -4.70415980e-01 3.17969710e-01 7.94501781e-01
-2.29706883e-01 1.91674661e-03 -4.13596421e-01 5.64177275e-01
1.04234385e+00 1.01169050e-01 1.83968022e-01 -1.48537469e+00
-4.11319107e-01 -3.81364763e-01 -2.38720387e-01 -9.72088814e-01
3.27073276e-01 -6.28851235e-01 2.22374961e-01 -1.28660691e+00
1.24079697e-01 -3.08562487e-01 -7.23475039e-01 -4.19398665e-01
8.87016505e-02 -3.55814219e-01 -1.61819249e-01 4.43770945e-01
-3.11234593e-01 8.21132660e-01 7.70932376e-01 3.21123190e-02
-1.56330869e-01 4.65064585e-01 -2.77972370e-01 4.32058811e-01
6.47941828e-01 -6.49075449e-01 -3.18468332e-01 -1.49445206e-01
1.50616154e-01 3.18276703e-01 3.82088900e-01 -1.01710486e+00
5.26278019e-01 -3.64500910e-01 4.88548800e-02 -5.61656475e-01
3.48430336e-01 -4.45082664e-01 2.88217872e-01 4.14436042e-01
-1.16327167e-01 -9.99400690e-02 2.85362154e-01 9.49466169e-01
-2.85285145e-01 -1.66111216e-01 7.86151350e-01 -2.59554774e-01
-5.46788096e-01 3.79252017e-01 -9.70867753e-01 1.22153372e-01
9.56618428e-01 1.04508661e-01 2.30404809e-01 -4.98269886e-01
-9.07818377e-01 2.75785357e-01 5.42257614e-02 -3.49081039e-01
6.68498158e-01 -9.13336873e-01 -6.60923958e-01 4.11079019e-01
-4.95127350e-01 -1.01648167e-01 5.21145582e-01 1.13226640e+00
-3.14240634e-01 6.39047682e-01 -2.78292950e-02 -8.57527077e-01
-5.88903844e-01 5.33389390e-01 4.80521291e-01 -6.16269074e-02
-4.35505062e-01 8.09512734e-01 1.91372633e-01 -1.45793647e-01
-6.53360877e-03 -3.57593209e-01 7.83238932e-02 3.22248526e-02
4.93714660e-01 7.37003207e-01 -1.78400606e-01 -5.66544294e-01
-4.60563183e-01 6.31123483e-01 4.11465138e-01 -3.66587996e-01
1.40149701e+00 -2.64327198e-01 -2.19862640e-01 1.12729204e+00
9.07151580e-01 -2.56974310e-01 -1.54850912e+00 -4.85291809e-01
9.91915688e-02 2.41526365e-02 4.03135657e-01 -2.63865411e-01
-9.31035936e-01 1.31634271e+00 6.29399419e-01 2.29529694e-01
7.70966291e-01 -7.06236809e-02 6.94238007e-01 4.77717698e-01
4.33307290e-01 -1.06609499e+00 -2.90973127e-01 6.86931789e-01
5.15553236e-01 -8.24256361e-01 -8.74790475e-02 6.14542328e-03
-5.74273467e-01 9.66495574e-01 -7.80432075e-02 -2.98252046e-01
1.13673663e+00 5.54832757e-01 -2.16753319e-01 -1.61501959e-01
-9.69866931e-01 -1.26206160e-01 -9.60935354e-02 4.24505025e-01
2.44028464e-01 -6.38457239e-02 -1.39830485e-01 6.31165504e-01
5.92571020e-01 2.97860771e-01 5.22242844e-01 9.92234111e-01
-3.83436203e-01 -5.68723321e-01 -2.76248038e-01 4.16460246e-01
-4.09444392e-01 -7.50250593e-02 -7.24533126e-02 2.75683761e-01
-2.99694419e-01 5.42985320e-01 2.83819228e-01 2.33234391e-01
9.89490151e-02 6.08745396e-01 2.85360724e-01 -4.17058736e-01
3.34649324e-01 2.43939713e-01 -5.59454679e-01 -4.01453167e-01
-6.15301132e-01 -1.23774719e+00 -1.26121140e+00 -1.59192309e-01
-6.23358823e-02 3.93768996e-01 7.84101307e-01 1.25708807e+00
4.30354565e-01 3.77549112e-01 6.42503381e-01 -7.76135266e-01
-4.85523671e-01 -6.95232511e-01 -1.23600507e+00 -8.86398777e-02
5.37274480e-01 -8.07475030e-01 -7.30138719e-01 -6.30569756e-02] | [6.647765159606934, 3.5862555503845215] |
1d434869-0bd0-497f-bc3c-1e1e0e0af9bd | a-literature-review-of-3d-face-reconstruction | 2110.09299 | null | https://arxiv.org/abs/2110.09299v2 | https://arxiv.org/pdf/2110.09299v2.pdf | A Review of 3D Face Reconstruction From a Single Image | 3D face reconstruction is a challenging problem but also an important task in the field of computer vision and graphics. Recently, many researchers put attention to the problem and a large number of articles have been published. Single image reconstruction is one of the branches of 3D face reconstruction, which has a lot of applications in our life. This paper is a review of the recent literature on 3D face reconstruction from a single image. | ['Hanxin Wang'] | 2021-10-13 | null | null | null | null | ['3d-face-reconstruction', 'face-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 1.71296492e-01 -2.98573315e-01 -4.13310900e-02 -3.57230872e-01
1.71390802e-01 -2.46411860e-02 3.95523101e-01 -5.94464481e-01
3.22188204e-03 3.03300232e-01 -5.84102459e-02 -2.91391145e-02
2.34502792e-01 -7.27266550e-01 -2.54508674e-01 -5.83710074e-01
2.30624110e-01 2.23350123e-01 2.03797057e-01 -8.02689940e-02
2.98983276e-01 9.17232871e-01 -1.79657519e+00 -1.16230167e-01
2.06302956e-01 9.83697772e-01 1.28106818e-01 1.42483905e-01
-6.23946428e-01 -1.99519806e-02 -3.38005275e-01 -5.10232270e-01
1.99090317e-01 -7.82854199e-01 -3.45202595e-01 6.60743356e-01
1.93191260e-01 -2.18377247e-01 -9.34954882e-02 9.73914146e-01
8.84809971e-01 -2.26300225e-01 5.73439956e-01 -1.13327456e+00
-4.81217325e-01 -2.51095086e-01 -1.14397430e+00 1.60829812e-01
6.70654714e-01 -3.27859432e-01 -1.01038620e-01 -1.12719142e+00
4.52319205e-01 1.67267418e+00 4.95907009e-01 6.60190761e-01
-8.48936915e-01 -8.36430013e-01 -9.19265226e-02 3.23106170e-01
-1.38097703e+00 -4.34807718e-01 9.89295065e-01 -3.68851751e-01
5.94874203e-01 2.02769473e-01 8.10375988e-01 5.60793519e-01
1.99932292e-01 4.78561580e-01 1.39777803e+00 -6.65221334e-01
-1.32080063e-01 2.48096287e-01 -1.06305212e-01 7.26031959e-01
2.17971757e-01 1.13374822e-01 -3.95249754e-01 -9.05991793e-02
1.22800529e+00 1.79768518e-01 -3.31258863e-01 -2.05628991e-01
-5.58974266e-01 6.62165880e-01 -9.61514376e-03 3.20941031e-01
-1.47779435e-01 -1.84590459e-01 1.60469368e-01 2.37230062e-01
6.77548170e-01 -5.48032284e-01 3.85942832e-02 1.17047606e-02
-5.78135312e-01 1.10650174e-01 6.90012753e-01 7.07286775e-01
4.88489538e-01 2.56897688e-01 4.25709963e-01 8.18360090e-01
6.04675472e-01 6.84560537e-01 2.50767410e-01 -6.77029490e-01
4.43037003e-02 6.36699617e-01 -2.65066773e-01 -1.06285822e+00
-1.80691049e-01 -4.39452082e-02 -7.76282370e-01 5.55766642e-01
8.96193832e-02 1.19227402e-01 -5.73491096e-01 9.75913107e-01
8.76795113e-01 2.63810813e-01 -3.23064893e-01 8.71522248e-01
1.43284118e+00 7.01780260e-01 -3.50413144e-01 -7.39772975e-01
1.27005243e+00 -5.58898866e-01 -1.05028820e+00 -9.09941867e-02
-3.58783215e-01 -1.19735384e+00 4.34588671e-01 3.49231809e-01
-1.13608253e+00 -5.46951354e-01 -6.45682335e-01 -1.66127477e-02
9.66872647e-02 -2.70705521e-01 5.00428379e-01 8.81462336e-01
-9.18748021e-01 2.84760356e-01 -3.91473591e-01 -6.37920797e-01
4.07513410e-01 4.39125866e-01 -7.68825769e-01 -2.58509755e-01
-5.34382582e-01 1.06771564e+00 2.87897531e-02 2.03471467e-01
-3.75165105e-01 -6.05054805e-03 -6.62431419e-01 -3.79220754e-01
3.31703305e-01 -5.43021083e-01 9.15271044e-01 -5.18239319e-01
-1.62194419e+00 1.52485037e+00 -6.16825879e-01 3.77158195e-01
3.09400737e-01 1.10591855e-02 -4.14471567e-01 -4.93113771e-02
-2.17904449e-01 -3.67935310e-04 1.17222917e+00 -1.04204762e+00
-3.48410666e-01 -1.05735791e+00 -4.51627284e-01 1.98172674e-01
6.46602288e-02 3.99369001e-01 -6.40672565e-01 -3.54329765e-01
3.60508859e-01 -6.97183132e-01 9.84866098e-02 3.92907530e-01
1.29881844e-01 -4.81832713e-01 1.17477751e+00 -4.27521557e-01
9.87765849e-01 -2.28863430e+00 2.20569298e-01 -3.16558927e-02
-8.50725845e-02 4.66062665e-01 1.23575166e-01 4.22321707e-01
-9.67387035e-02 1.18005775e-01 -2.55259186e-01 -4.61459577e-01
-4.53220844e-01 1.81545570e-01 -1.74448431e-01 8.43853772e-01
-2.28918761e-01 4.63014990e-01 -3.71270925e-01 -6.06109023e-01
5.51069438e-01 7.87763834e-01 -3.07590783e-01 3.54251415e-01
3.10002953e-01 6.93388045e-01 -6.68908656e-01 7.08898723e-01
1.10007441e+00 1.95940554e-01 -1.44448308e-02 -7.21865371e-02
-4.17488277e-01 -4.68748122e-01 -1.28828931e+00 1.30851829e+00
-1.53106734e-01 3.13512176e-01 5.09623230e-01 -1.08584011e+00
1.28563845e+00 5.42741060e-01 5.04499853e-01 -4.32632446e-01
5.02609193e-01 2.53223717e-01 -2.78448910e-01 -8.31519902e-01
-1.07934885e-01 -7.56350875e-01 3.87389332e-01 5.93872666e-01
-1.92138761e-01 -5.75265706e-01 -2.46943578e-01 -3.81031305e-01
4.02280420e-01 1.08421803e-01 7.78883219e-01 3.81660871e-02
8.86041701e-01 -3.27253371e-01 5.54252446e-01 -2.31546640e-01
-2.67139405e-01 8.71402323e-01 1.61521569e-01 -8.00849020e-01
-7.63343990e-01 -7.75528848e-01 -4.51654553e-01 1.81867316e-01
3.87877613e-01 -1.57339618e-01 -5.95799088e-01 -4.55322087e-01
2.26778224e-01 -6.18574247e-02 -6.21094227e-01 2.09860995e-01
-7.32223570e-01 -5.77809453e-01 1.60703789e-02 1.35024577e-01
9.33294535e-01 -1.23648608e+00 -5.06051242e-01 -1.10810868e-01
9.95311663e-02 -1.07868910e+00 -2.96100825e-01 -4.57890004e-01
-1.21404564e+00 -1.15905988e+00 -9.35270011e-01 -1.05130947e+00
1.01624250e+00 8.71000350e-01 8.87130916e-01 5.08019626e-01
-5.27136445e-01 5.80144823e-01 -2.68802434e-01 -5.93858480e-01
-1.77911744e-01 -4.22015399e-01 2.32830673e-01 1.19090721e-01
3.97091985e-01 -5.55270195e-01 -2.82454759e-01 6.26282334e-01
-6.31857991e-01 -1.87240928e-01 3.93128544e-01 5.40274441e-01
9.17377710e-01 1.85644746e-01 4.74820167e-01 -8.49807501e-01
4.94158268e-01 -2.88634539e-01 -8.33797395e-01 2.99655311e-02
-2.29499534e-01 -2.38801017e-01 3.60431999e-01 -1.08204953e-01
-1.20425415e+00 1.67546600e-01 -5.34196019e-01 -4.31881219e-01
-3.35539758e-01 1.15017071e-01 -4.33852345e-01 -5.59243739e-01
2.63530254e-01 4.58479732e-01 7.63163209e-01 -9.85728681e-01
-1.81939170e-01 7.25060403e-01 1.37213856e-01 -2.30651140e-01
7.30777681e-01 7.29977965e-01 4.82522666e-01 -1.27027225e+00
-5.72552025e-01 -3.76200199e-01 -6.45221889e-01 -5.05738795e-01
7.00982273e-01 -4.40401614e-01 -8.72640014e-01 5.30423045e-01
-1.23371029e+00 2.59018689e-01 6.18630573e-02 5.05643070e-01
-3.47208917e-01 4.45438832e-01 -1.78471074e-01 -9.75593746e-01
-4.48581398e-01 -1.28730714e+00 9.22569633e-01 6.58143222e-01
2.37399057e-01 -9.49125051e-01 1.17107771e-01 3.27300191e-01
1.81034505e-01 3.26871336e-01 7.52768517e-01 3.00066710e-01
-4.09381688e-01 -2.84573346e-01 -2.08444104e-01 1.97453097e-01
4.02124792e-01 7.88056478e-02 -1.04165006e+00 -1.39483005e-01
6.72405839e-01 6.50133342e-02 4.96472389e-01 5.53520620e-01
1.09222209e+00 1.50614232e-01 -5.30383289e-01 5.76098025e-01
1.49762535e+00 4.77442056e-01 7.35705912e-01 -1.11289494e-01
3.39495510e-01 7.44382203e-01 5.16918361e-01 4.54637557e-01
6.62143156e-02 6.44848585e-01 5.43395102e-01 8.90862420e-02
-2.11563423e-01 -1.68416992e-01 3.81607339e-02 8.76632869e-01
-5.12917697e-01 2.67361164e-01 -6.39631033e-01 -5.47318794e-02
-1.40934741e+00 -1.09065127e+00 -2.59094894e-01 2.20542669e+00
3.60388875e-01 -2.77972668e-01 1.95516258e-01 5.30268610e-01
1.00042653e+00 4.26938906e-02 -4.17148411e-01 -3.60048831e-01
5.37382066e-02 2.06761688e-01 -1.84393018e-01 4.11157966e-01
-7.69009769e-01 7.50650465e-01 7.71140671e+00 6.29004121e-01
-1.39961135e+00 -5.07588238e-02 3.97647470e-01 3.03214163e-01
-6.67968541e-02 -7.54877999e-02 -1.05766070e+00 4.67904747e-01
1.29908249e-01 -1.98780388e-01 1.98462576e-01 8.07366490e-01
1.47150233e-01 -6.10323012e-01 -6.65585339e-01 1.65155470e+00
6.15359366e-01 -9.97821391e-01 1.99342191e-01 4.46979433e-01
4.80432034e-01 -6.90692544e-01 2.26653084e-01 -1.42731726e-01
-6.00224853e-01 -1.32905948e+00 2.27414355e-01 4.59386349e-01
1.08530712e+00 -8.85059595e-01 6.07117832e-01 7.29427099e-01
-1.27530372e+00 3.05883765e-01 -6.11866415e-01 -3.92211795e-01
4.05448824e-01 7.76378572e-01 -6.01298034e-01 3.40505391e-01
7.82695234e-01 6.94028080e-01 -6.01804033e-02 1.28741086e+00
-1.40436620e-01 1.81986272e-01 -1.95832714e-01 -1.05964109e-01
-3.51644725e-01 -6.18894219e-01 5.29175341e-01 6.27921641e-01
3.85627031e-01 7.35626280e-01 8.19111615e-02 5.03611028e-01
-2.86128521e-02 2.72960186e-01 -1.05675590e+00 1.39968678e-01
1.11446194e-01 1.11275983e+00 -7.88940012e-01 8.78656730e-02
-6.86335802e-01 7.70057321e-01 5.38527817e-02 -2.17944473e-01
-3.90578836e-01 -8.60287845e-02 4.86981541e-01 4.65616614e-01
1.66754108e-02 -3.91585112e-01 -9.88637805e-02 -8.43020499e-01
-2.02130914e-01 -5.18483162e-01 2.87222713e-01 -5.22301733e-01
-9.96475399e-01 4.94010717e-01 1.86794266e-01 -1.25623941e+00
4.38000187e-02 -7.04194665e-01 -7.19583750e-01 9.82685089e-01
-1.62344360e+00 -7.02925503e-01 -4.69745427e-01 6.10675871e-01
7.87607610e-01 -3.41260433e-01 9.78024483e-01 2.14216277e-01
-6.37504220e-01 1.77190289e-01 -8.74559358e-02 -1.78314835e-01
5.10688245e-01 -4.79574651e-01 2.82237262e-01 4.11620289e-01
-1.12232141e-01 5.17885804e-01 4.89477545e-01 -5.76398194e-01
-1.78603661e+00 -4.90006387e-01 7.80397654e-01 -1.56226218e-01
-2.92147428e-01 -2.30955184e-02 -7.62685657e-01 3.14455688e-01
2.12271139e-01 3.72004807e-01 7.09973276e-01 -3.85471314e-01
1.77762970e-01 -2.34753326e-01 -1.68400180e+00 2.73626208e-01
1.10728312e+00 -1.65663317e-01 -4.98039782e-01 -5.63945621e-03
-9.57259238e-02 -5.06186366e-01 -6.05182052e-01 4.59459782e-01
7.53872871e-01 -1.36396635e+00 9.35224593e-01 1.43222794e-01
-1.00381419e-01 -3.62214744e-01 6.60850257e-02 -1.08823240e+00
-2.11172640e-01 -3.18632782e-01 -8.30532983e-02 9.31973994e-01
-5.15720606e-01 -5.77416301e-01 8.59604239e-01 3.31312269e-01
1.50450900e-01 -9.17047381e-01 -1.03954101e+00 -4.58162427e-01
-1.40597403e-01 4.76219505e-02 5.15267730e-01 5.52130878e-01
-2.98183650e-01 4.42740858e-01 -2.31438965e-01 -3.12052786e-01
1.03114200e+00 5.03746271e-01 8.80268991e-01 -1.66066706e+00
3.70569259e-01 -3.91887009e-01 -4.99035418e-01 -9.13571715e-01
2.15116456e-01 -8.86139095e-01 -5.13393998e-01 -1.74938631e+00
1.29639179e-01 -2.89791316e-01 6.20905221e-01 1.05080597e-01
2.18806162e-01 5.65984130e-01 -1.11068403e-02 2.56789476e-01
1.31428093e-01 5.55595577e-01 1.79012215e+00 2.30132744e-01
2.64136661e-02 3.20677012e-01 -5.21169364e-01 8.35392058e-01
6.98123336e-01 -2.72967577e-01 -2.31105179e-01 -4.71364796e-01
-2.05950722e-01 1.60551876e-01 6.24096021e-02 -5.04628539e-01
-3.93819660e-02 -2.17556193e-01 6.33723557e-01 -8.92969429e-01
6.92568779e-01 -1.17838657e+00 4.41421241e-01 4.09268945e-01
6.41845465e-01 1.08431481e-01 1.03066370e-01 2.91632891e-01
-3.03315490e-01 -4.47000802e-01 1.15119100e+00 -6.95604801e-01
-8.73004556e-01 6.26966536e-01 -1.71613291e-01 -3.30634475e-01
1.40529525e+00 -6.75135493e-01 2.31018022e-01 -1.41824871e-01
-5.31571448e-01 -1.76389530e-01 4.59121525e-01 2.08955541e-01
1.20189583e+00 -1.40627575e+00 -7.30957091e-01 7.03303039e-01
-3.81677240e-01 2.54223589e-02 2.75463641e-01 6.05329990e-01
-6.74006462e-01 3.33029687e-01 -4.88880724e-01 -6.95423543e-01
-1.68412924e+00 4.16042984e-01 2.05885783e-01 5.08843124e-01
-8.01341832e-01 6.60396159e-01 1.73921391e-01 -1.84286118e-01
1.98012933e-01 3.99282008e-01 -5.71445048e-01 -6.05791770e-02
8.90715122e-01 5.61000764e-01 -2.44704232e-01 -1.20704639e+00
-4.56427872e-01 1.46856070e+00 1.92193672e-01 1.71246496e-03
1.29805005e+00 -2.04450026e-01 -4.51213658e-01 3.17986846e-01
1.23023760e+00 -2.15275474e-02 -7.78079808e-01 -9.03244317e-02
-5.54176092e-01 -1.10084379e+00 4.07459140e-02 -1.00555465e-01
-1.29153907e+00 1.17840469e+00 5.34116864e-01 1.65194198e-01
1.16423583e+00 1.02576964e-01 5.68336070e-01 -2.11182743e-01
7.98749268e-01 -6.92803919e-01 2.16925204e-01 4.43756998e-01
1.30528629e+00 -1.31189156e+00 3.95067006e-01 -9.08992410e-01
-2.41033062e-01 1.22180641e+00 4.87038374e-01 -2.61878788e-01
1.26458848e+00 4.37892467e-01 1.58464253e-01 -1.66297063e-01
-1.10769318e-02 -1.21168755e-01 5.70969954e-02 7.26968884e-01
6.91572428e-01 -3.33530784e-01 -7.29976594e-01 4.37604636e-01
3.95784006e-02 1.19963415e-01 1.44925803e-01 1.03768444e+00
-5.94344854e-01 -1.42141473e+00 -9.44838047e-01 2.48841330e-01
-4.22423929e-01 5.01472533e-01 -2.37634480e-01 5.55427194e-01
7.89796561e-02 7.43703783e-01 -1.19543917e-01 -1.61547482e-01
4.58664596e-01 -3.06918975e-02 1.13002253e+00 -6.89986169e-01
-1.92069039e-01 1.43326432e-01 -4.73826259e-01 -2.83535719e-01
-6.87711537e-01 -6.73019290e-01 -1.17297697e+00 -5.91986358e-01
-3.91250610e-01 2.24215746e-01 9.51037586e-01 7.26508677e-01
3.34091634e-01 -5.91650121e-02 1.00265801e+00 -1.04591072e+00
-3.87605056e-02 -6.74897373e-01 -1.05815566e+00 1.28654405e-01
1.75929755e-01 -1.07375991e+00 -2.20869020e-01 -4.90940455e-03] | [13.131649017333984, 0.44553282856941223] |
8e060556-02cd-4708-9de0-1abd7f846080 | supervised-dictionary-learning-with-auxiliary | 2206.06774 | null | https://arxiv.org/abs/2206.06774v1 | https://arxiv.org/pdf/2206.06774v1.pdf | Supervised Dictionary Learning with Auxiliary Covariates | Supervised dictionary learning (SDL) is a classical machine learning method that simultaneously seeks feature extraction and classification tasks, which are not necessarily a priori aligned objectives. The goal of SDL is to learn a class-discriminative dictionary, which is a set of latent feature vectors that can well-explain both the features as well as labels of observed data. In this paper, we provide a systematic study of SDL, including the theory, algorithm, and applications of SDL. First, we provide a novel framework that `lifts' SDL as a convex problem in a combined factor space and propose a low-rank projected gradient descent algorithm that converges exponentially to the global minimizer of the objective. We also formulate generative models of SDL and provide global estimation guarantees of the true parameters depending on the hyperparameter regime. Second, viewed as a nonconvex constrained optimization problem, we provided an efficient block coordinate descent algorithm for SDL that is guaranteed to find an $\varepsilon$-stationary point of the objective in $O(\varepsilon^{-1}(\log \varepsilon^{-1})^{2})$ iterations. For the corresponding generative model, we establish a novel non-asymptotic local consistency result for constrained and regularized maximum likelihood estimation problems, which may be of independent interest. Third, we apply SDL for imbalanced document classification by supervised topic modeling and also for pneumonia detection from chest X-ray images. We also provide simulation studies to demonstrate that SDL becomes more effective when there is a discrepancy between the best reconstructive and the best discriminative dictionaries. | ['Weixin Yao', 'Hanbaek Lyu', 'Joowon Lee'] | 2022-06-14 | null | null | null | null | ['pneumonia-detection', 'document-classification'] | ['medical', 'natural-language-processing'] | [ 1.54194251e-01 9.46771912e-03 -3.51315141e-01 -1.85528278e-01
-1.21258318e+00 -4.04853910e-01 6.64461404e-02 9.57842991e-02
-1.58275634e-01 6.52785361e-01 1.51745632e-01 -1.28097028e-01
-6.62991047e-01 -4.83307898e-01 -6.79786682e-01 -1.17990005e+00
-1.40924647e-01 5.52936137e-01 -4.87725139e-01 2.45760810e-02
-1.56853925e-02 3.30581814e-01 -1.11500585e+00 -2.50388741e-01
9.57120597e-01 9.03390706e-01 2.10923940e-01 5.31593382e-01
2.57640332e-01 4.96945471e-01 -2.95190752e-01 -7.23849908e-02
2.81359643e-01 -5.79305887e-01 -5.36521733e-01 6.19489849e-01
3.48705530e-01 1.12894863e-01 -1.03533700e-01 1.12817359e+00
5.10014296e-01 1.53517365e-01 9.66674805e-01 -1.18747735e+00
-4.91465241e-01 1.42019525e-01 -6.81789696e-01 1.97250947e-01
8.02014843e-02 -4.01567012e-01 1.12157381e+00 -1.18693507e+00
4.87611383e-01 8.80085051e-01 5.47222197e-01 2.84359217e-01
-1.29241574e+00 -4.82445210e-01 1.03778504e-01 -1.05276152e-01
-1.53368855e+00 -3.23269457e-01 8.36937487e-01 -5.55447519e-01
5.78400314e-01 4.35038745e-01 4.15977925e-01 8.15773129e-01
1.79597512e-01 8.62419963e-01 9.77636874e-01 -6.71839595e-01
3.63905758e-01 1.64902657e-01 2.65186280e-01 1.03195202e+00
5.65459669e-01 -6.13037013e-02 -7.50284970e-01 -5.01450360e-01
7.22852945e-01 1.07640907e-01 -4.91465837e-01 -6.02356791e-01
-1.11909270e+00 1.25649405e+00 -1.07343420e-01 2.61345655e-01
-4.53281164e-01 6.62214011e-02 -2.15964671e-02 1.17179550e-01
7.02695310e-01 2.44661584e-01 -3.61822575e-01 6.67822361e-02
-9.72488761e-01 6.07180521e-02 8.20562720e-01 7.29674339e-01
7.30024517e-01 2.15170458e-01 2.64425445e-02 9.10368979e-01
5.64284444e-01 9.71631706e-01 3.00418913e-01 -8.25729728e-01
4.43373442e-01 3.03487539e-01 1.73003934e-02 -1.14700115e+00
-2.56994992e-01 -7.99952984e-01 -9.58619297e-01 -2.14322105e-01
2.64679313e-01 -1.29323125e-01 -4.72544253e-01 1.89371443e+00
3.58585805e-01 1.68232441e-01 -1.49024099e-01 8.68144095e-01
2.97982097e-01 6.98786080e-01 -4.07969087e-01 -9.67230082e-01
1.05421925e+00 -6.35199487e-01 -6.70719922e-01 -2.04461068e-01
6.43553197e-01 -7.03595698e-01 9.77432787e-01 4.84263688e-01
-1.02529287e+00 -1.61481172e-01 -8.32185924e-01 2.21616834e-01
2.80593604e-01 3.78251702e-01 5.83316088e-01 6.02714419e-01
-9.14518058e-01 2.46668831e-01 -1.02082646e+00 1.94397178e-02
1.48448274e-01 4.06948328e-01 -3.06839049e-01 -1.69778883e-01
-7.45730639e-01 5.40608943e-01 2.22860649e-02 8.83458778e-02
-8.38720679e-01 -5.97887576e-01 -8.46385896e-01 -1.26681104e-01
4.59612638e-01 -5.95788240e-01 7.67939389e-01 -7.44891822e-01
-1.13841653e+00 8.88815939e-01 -6.30234897e-01 -1.81293860e-01
9.10357162e-02 -1.90547571e-01 -1.53923213e-01 2.03772768e-01
3.87484044e-01 -1.30470991e-01 1.10510993e+00 -1.16832006e+00
-3.68074626e-01 -5.29560030e-01 -2.62724280e-01 3.45476240e-01
-4.63826418e-01 -2.16055542e-01 -4.31057245e-01 -9.44837451e-01
4.61065680e-01 -9.49093580e-01 -3.96175623e-01 -6.34776428e-02
-3.06075811e-01 -1.16857931e-01 4.13251758e-01 -6.91720963e-01
1.41100669e+00 -2.35084295e+00 5.91780543e-01 5.55421114e-01
2.65816927e-01 -9.81487036e-02 1.31638438e-01 3.89933884e-01
-8.33151937e-02 -1.72718555e-01 -5.33286631e-01 -5.68929732e-01
1.07824001e-02 3.94540071e-01 -4.07089800e-01 1.04943871e+00
-1.55859828e-01 4.85930622e-01 -8.38344276e-01 -4.30985332e-01
2.43927892e-02 4.62045968e-01 -6.58048689e-01 6.03523999e-02
-1.75378043e-02 4.53813672e-01 -4.98957992e-01 6.98193610e-01
5.05878568e-01 -5.04735172e-01 1.98791251e-01 -1.37371480e-01
-4.16062027e-03 -1.03964694e-01 -1.56603527e+00 1.62565267e+00
-5.18233895e-01 4.76412475e-01 3.97969246e-01 -1.47248459e+00
8.35198641e-01 2.96119273e-01 9.90828335e-01 -2.22353905e-01
1.01840049e-02 4.14382279e-01 -5.19985318e-01 -4.29423928e-01
-2.87233461e-02 -5.24590015e-01 -1.25130773e-01 5.89364469e-01
2.86510959e-02 8.32267404e-02 6.20948970e-02 2.80974060e-02
9.69737709e-01 -4.13448095e-01 4.83474135e-01 -5.93670428e-01
4.62626725e-01 -2.62459666e-01 6.51020110e-01 8.44042122e-01
2.11483434e-01 6.92873001e-01 4.43964660e-01 -1.30378187e-01
-7.54224718e-01 -1.08706534e+00 -4.87934887e-01 9.55458701e-01
9.46088433e-02 -2.88011104e-01 -4.83576566e-01 -6.37427747e-01
-6.80247620e-02 5.65679252e-01 -4.47082430e-01 -6.99079782e-02
-4.69058037e-01 -1.28930664e+00 9.44535732e-02 2.60979027e-01
1.70883656e-01 -4.68211621e-01 -4.21200305e-01 1.51867077e-01
-1.70204505e-01 -8.22599649e-01 -7.67681897e-01 5.20767152e-01
-8.71685147e-01 -1.14106894e+00 -7.27735221e-01 -9.16531503e-01
1.12121880e+00 3.37529093e-01 9.15121436e-01 -4.88953516e-02
-2.07286924e-01 7.07364440e-01 -2.27580607e-01 -2.11748451e-01
-1.22731611e-01 -1.59943938e-01 4.13908750e-01 3.17658365e-01
1.65854886e-01 -5.35759270e-01 -4.57609326e-01 4.27887797e-01
-8.96123290e-01 -1.94792271e-01 4.33773100e-01 1.10288894e+00
1.12149811e+00 2.08243400e-01 4.76222038e-01 -7.29776323e-01
6.13343418e-01 -5.84954023e-01 -6.31296158e-01 2.97804743e-01
-8.56407881e-01 2.60237426e-01 5.58729947e-01 -2.71084368e-01
-6.31998837e-01 8.28363970e-02 2.34534536e-02 -5.61588049e-01
2.66457260e-01 7.99152970e-01 -1.99369609e-01 7.20306560e-02
7.44698703e-01 5.95324218e-01 -2.50083990e-02 -6.28093660e-01
2.47813538e-01 6.29605353e-01 3.99066329e-01 -7.65492201e-01
7.61534572e-01 7.34726489e-01 1.92790762e-01 -9.85656917e-01
-1.20336413e+00 -7.84283578e-01 -4.93717134e-01 5.03575243e-03
5.75209498e-01 -9.23126578e-01 -5.19199193e-01 -2.82603363e-03
-6.95319235e-01 -1.82881579e-01 -5.64071000e-01 7.75765300e-01
-7.35219121e-01 4.61300582e-01 -2.31584474e-01 -8.73321414e-01
-2.60874748e-01 -9.96224999e-01 1.08566833e+00 -1.81004748e-01
-1.09128952e-01 -1.29667246e+00 3.64559025e-01 3.10429603e-01
-7.58410944e-03 2.35962838e-01 9.45865870e-01 -6.20150685e-01
-3.11410695e-01 -2.65044779e-01 7.25586489e-02 6.19166195e-01
2.48574778e-01 -4.54960644e-01 -4.94339824e-01 -7.18772233e-01
6.58968747e-01 -7.06101432e-02 7.58381903e-01 7.32161522e-01
1.03504539e+00 -6.97092175e-01 -3.90292972e-01 8.56455803e-01
1.52388704e+00 1.29027724e-01 8.51150453e-02 -1.42399609e-01
6.81983650e-01 1.76083639e-01 4.64268386e-01 7.91483521e-01
1.31445974e-01 7.29610384e-01 1.82005152e-01 -1.02619380e-01
2.96369977e-02 -1.91422906e-02 3.92918885e-01 1.18448699e+00
7.41367117e-02 -1.43734872e-01 -7.50127137e-01 7.16610610e-01
-1.83722436e+00 -7.72078753e-01 -1.29169434e-01 2.55169392e+00
8.30318570e-01 -1.77196279e-01 5.19306585e-02 2.98593193e-01
7.05400646e-01 9.22537073e-02 -5.23763478e-01 -6.18081540e-02
-1.52196914e-01 6.19681180e-01 4.35421824e-01 6.89833403e-01
-1.01833367e+00 4.02042747e-01 6.20376110e+00 9.85585332e-01
-1.00752437e+00 3.19902033e-01 3.48369569e-01 -2.02997595e-01
-4.75858927e-01 -2.91542783e-02 -8.22748125e-01 4.89989370e-01
5.43401599e-01 -2.04062954e-01 4.09643978e-01 8.74342978e-01
3.04993868e-01 -1.06921643e-01 -1.01685011e+00 1.20911753e+00
4.13732052e-01 -1.15983725e+00 -2.18711779e-01 4.56013232e-01
9.91702557e-01 -1.80197194e-01 2.89892048e-01 -9.19543207e-02
7.55608603e-02 -9.40682292e-01 4.11722660e-01 3.50384176e-01
8.14485729e-01 -7.89245069e-01 3.97297770e-01 5.61907351e-01
-9.79397118e-01 -1.31741807e-01 -4.04247046e-01 1.08653426e-01
1.48453400e-01 1.11257899e+00 -5.57864726e-01 3.90955448e-01
3.46971214e-01 8.20147395e-01 -5.64588197e-02 8.54014039e-01
-1.99062735e-01 7.01277375e-01 -6.56373203e-01 1.49091497e-01
-9.02506581e-04 -4.94165272e-01 8.44167292e-01 1.13412607e+00
4.62555379e-01 3.19229484e-01 4.23092306e-01 5.62190235e-01
-3.55647951e-02 3.54934663e-01 -4.51780289e-01 1.64169505e-01
3.57865721e-01 9.21943009e-01 -6.89946532e-01 3.95511277e-03
-3.23689818e-01 9.09243286e-01 2.13863924e-01 3.84744287e-01
-7.58689165e-01 3.33105959e-02 5.88926554e-01 1.54657483e-01
4.96498793e-01 -4.21269566e-01 -3.03585202e-01 -1.56019819e+00
2.49370530e-01 -9.75923002e-01 5.26544571e-01 -2.74311304e-01
-1.29070807e+00 3.93589944e-01 9.66931060e-02 -1.25859261e+00
-3.42712075e-01 -5.24207652e-01 -2.90757269e-01 6.66354835e-01
-1.22361088e+00 -8.32574785e-01 1.99270830e-03 9.17164624e-01
4.53809261e-01 -3.50856423e-01 9.30745780e-01 3.68829161e-01
-6.12051427e-01 5.21534920e-01 7.63533533e-01 -8.44815746e-02
4.16596472e-01 -1.20394289e+00 -5.51095545e-01 9.69383717e-01
3.96251082e-01 5.26490450e-01 7.34162688e-01 -5.97563803e-01
-1.39878476e+00 -1.01555717e+00 9.77109134e-01 -2.40377560e-01
5.53465486e-01 -1.79869622e-01 -6.18217647e-01 6.71920240e-01
-2.96077311e-01 1.82398766e-01 1.04204583e+00 6.98519573e-02
3.31325317e-03 -1.01259872e-01 -9.05956566e-01 1.96546629e-01
8.97951245e-01 -5.45012236e-01 -3.08146685e-01 7.62856007e-01
1.88854113e-01 -4.55653965e-01 -8.17690194e-01 2.91987687e-01
3.26753736e-01 -6.63345218e-01 1.01978326e+00 -6.16338313e-01
2.52498955e-01 -1.63132235e-01 -6.22246921e-01 -1.03821421e+00
-2.31202453e-01 -6.94224060e-01 -3.96449268e-01 7.38822579e-01
2.21905470e-01 -4.60348189e-01 8.38618100e-01 2.91190445e-01
-9.32295099e-02 -1.10463023e+00 -1.17409861e+00 -8.28499019e-01
-6.48730397e-02 -5.67862570e-01 -2.33424932e-01 9.30501521e-01
-6.84364289e-02 3.02925736e-01 -6.28165901e-01 3.02075118e-01
9.77214932e-01 3.25013459e-01 3.65223557e-01 -9.95414853e-01
-7.84847319e-01 -7.36206174e-02 -2.29019582e-01 -1.30822098e+00
3.45525831e-01 -1.19811153e+00 -1.60119701e-02 -1.41498148e+00
3.70022416e-01 -6.39423490e-01 -2.43840098e-01 3.20715547e-01
-1.25062719e-01 -1.85748562e-02 1.07292328e-02 4.32308286e-01
-4.19066548e-01 4.70454246e-01 9.91582394e-01 -1.21918231e-01
-3.49881083e-01 4.01045293e-01 -8.58744442e-01 7.27079391e-01
4.13346082e-01 -8.09600949e-01 -5.67596614e-01 -4.24019873e-01
3.67129505e-01 3.04014534e-01 2.39564523e-01 -5.78033924e-01
7.61278346e-02 -2.90271342e-01 6.82948604e-02 -4.60920036e-01
3.15786362e-01 -7.34759808e-01 1.40801743e-01 3.96915793e-01
-3.18072408e-01 -1.51742265e-01 -3.79139006e-01 7.17097342e-01
-1.67120263e-01 -6.14772856e-01 7.58980095e-01 1.23718470e-01
-8.71320069e-02 3.46221060e-01 -3.56973678e-01 1.61607414e-01
8.93919051e-01 4.96019386e-02 2.25417748e-01 -5.84979951e-01
-9.74310815e-01 -3.05842701e-03 1.95684224e-01 -1.23213731e-01
6.58419847e-01 -1.33141422e+00 -8.79641175e-01 3.49399120e-01
-2.31311083e-01 2.64026999e-01 3.83719918e-03 1.00874269e+00
-2.46056020e-01 5.36476433e-01 5.18122852e-01 -8.04676175e-01
-1.03704488e+00 5.09968877e-01 2.57765919e-01 -4.91234004e-01
-4.06212181e-01 1.00126278e+00 2.36681923e-01 -1.17184579e-01
2.29284212e-01 -9.82277617e-02 2.01625496e-01 5.61989695e-02
2.52321482e-01 4.10249680e-01 1.58238381e-01 -7.20102847e-01
-4.62462008e-01 6.28344595e-01 1.89804584e-01 -2.46589243e-01
1.35351205e+00 -2.60207176e-01 -2.18333021e-01 3.55831414e-01
1.63111711e+00 3.11912298e-01 -1.04729474e+00 -4.15822178e-01
-2.26491675e-01 -3.87859583e-01 2.39919618e-01 -3.00970852e-01
-1.15518117e+00 5.71561038e-01 5.04959464e-01 1.68379918e-01
1.19440043e+00 2.61913210e-01 5.26266694e-01 2.68068045e-01
3.02792042e-01 -9.64685261e-01 3.42609644e-01 2.95670152e-01
8.96857321e-01 -1.07657480e+00 2.98235089e-01 -4.56293344e-01
-3.43243748e-01 1.00979435e+00 3.46816741e-02 -2.52736241e-01
1.08040440e+00 4.08047698e-02 -2.53933460e-01 -3.19157183e-01
-5.17429769e-01 -3.99996005e-02 5.00953853e-01 3.79951239e-01
2.90668935e-01 8.66087303e-02 -6.22160077e-01 6.83266580e-01
-1.88930541e-01 -1.94925398e-01 1.93994388e-01 9.19761658e-01
-3.40406060e-01 -1.03691542e+00 -5.54066837e-01 5.65788150e-01
-5.20033240e-01 -1.81993097e-02 1.11566991e-01 5.23497045e-01
1.50581703e-01 8.46461117e-01 -1.05719551e-01 5.59879793e-03
8.00326690e-02 -1.54229533e-02 5.57522118e-01 -7.42512941e-01
1.73990250e-01 5.98857105e-01 -2.58804172e-01 -2.34155789e-01
-4.56884831e-01 -9.27568197e-01 -1.16442215e+00 6.52897730e-02
-6.19618773e-01 3.56403261e-01 5.84095597e-01 1.19868577e+00
8.62096474e-02 1.18531123e-01 1.07783210e+00 -4.12759721e-01
-7.24172175e-01 -5.06965160e-01 -9.22669590e-01 2.32875273e-01
3.88365239e-01 -5.85567057e-01 -6.71905637e-01 2.40804449e-01] | [7.139898300170898, 4.440162181854248] |
9ab73237-0696-4934-896c-02757ac1ff70 | troika-multi-path-cross-modal-traction-for | 2303.15230 | null | https://arxiv.org/abs/2303.15230v1 | https://arxiv.org/pdf/2303.15230v1.pdf | Troika: Multi-Path Cross-Modal Traction for Compositional Zero-Shot Learning | Recent compositional zero-shot learning (CZSL) methods adapt pre-trained vision-language models (VLMs) by constructing trainable prompts only for composed state-object pairs. Relying on learning the joint representation of seen compositions, these methods ignore the explicit modeling of the state and object, thus limiting the exploitation of pre-trained knowledge and generalization to unseen compositions. With a particular focus on the universality of the solution, in this work, we propose a novel paradigm for CZSL models that establishes three identification branches (i.e., Multi-Path) to jointly model the state, object, and composition. The presented Troika is our implementation that aligns the branch-specific prompt representations with decomposed visual features. To calibrate the bias between semantically similar multi-modal representations, we further devise a Cross-Modal Traction module into Troika that shifts the prompt representation towards the current visual content. We conduct extensive experiments on three popular benchmarks, where our method significantly outperforms existing methods in both closed-world and open-world settings. | ['Donglin Wang', 'Yiliang Lv', 'Yutong Feng', 'Biao Gong', 'Siteng Huang'] | 2023-03-27 | null | null | null | null | ['compositional-zero-shot-learning'] | ['computer-vision'] | [ 3.80710453e-01 -2.88953986e-02 -3.66349041e-01 -9.74091738e-02
-6.94128036e-01 -6.26059651e-01 1.20721495e+00 1.94370877e-02
-2.31860518e-01 7.09744543e-02 4.18666631e-01 -9.35827568e-02
2.66532838e-01 -5.28810143e-01 -7.06222653e-01 -6.69772983e-01
5.42150378e-01 4.02271420e-01 6.03057444e-01 -1.25825047e-01
9.52543318e-02 1.14725903e-01 -1.83718717e+00 7.13452697e-01
4.00074691e-01 9.58631039e-01 3.81067187e-01 5.78623652e-01
-5.01148283e-01 1.11921751e+00 -6.18594326e-02 -3.25877726e-01
2.44302377e-01 -4.51200634e-01 -6.37011945e-01 4.57690060e-01
9.51390505e-01 -2.74164528e-01 -3.83267641e-01 9.95082736e-01
2.31298521e-01 1.85477212e-01 8.03191304e-01 -1.39251995e+00
-7.99287200e-01 6.63446426e-01 -4.46625441e-01 -7.10639125e-03
2.30358586e-01 5.87019742e-01 1.42160964e+00 -1.00099635e+00
1.04028988e+00 1.17770898e+00 3.17847997e-01 8.71229291e-01
-1.56354117e+00 -3.55097443e-01 8.11721027e-01 5.80836892e-01
-1.16533542e+00 -7.21995831e-01 9.33196723e-01 -7.07310617e-01
8.68856072e-01 1.70815252e-02 5.53477705e-01 1.58437753e+00
-2.83094734e-01 1.13418531e+00 1.27142036e+00 -5.41475177e-01
2.92462409e-01 3.88795771e-02 3.70983690e-01 8.49907219e-01
-6.19758666e-03 2.11564139e-01 -7.39041746e-01 -1.15228303e-01
6.07990801e-01 5.93885072e-02 -1.81383535e-01 -1.07221258e+00
-1.44447422e+00 7.22352445e-01 2.86094099e-01 2.53068089e-01
-2.77913660e-01 2.62695789e-01 3.96917075e-01 1.07540354e-01
2.00746190e-02 1.25053093e-01 -2.89027870e-01 2.69991964e-01
-7.69130528e-01 2.13565633e-01 6.14879608e-01 1.06683934e+00
1.03416812e+00 1.45546585e-01 -7.67341018e-01 6.85362101e-01
4.52220470e-01 2.41354734e-01 4.89640832e-01 -8.39244008e-01
2.49781966e-01 7.24850714e-01 -9.37384367e-02 -3.83566976e-01
-1.25727445e-01 -2.91124344e-01 -3.90121222e-01 3.78547721e-02
2.82709271e-01 2.19929665e-01 -1.29087555e+00 2.00292921e+00
3.33682805e-01 6.01632535e-01 1.54582337e-01 8.60279918e-01
9.98114765e-01 6.30449533e-01 4.86411929e-01 -6.95833936e-02
1.39289534e+00 -1.42107248e+00 -5.73898077e-01 -3.22760433e-01
2.95623571e-01 -4.88580763e-01 1.34262025e+00 8.75441134e-02
-9.90369678e-01 -5.63089490e-01 -8.86448920e-01 -2.73479879e-01
-4.67604667e-01 6.36654347e-02 5.42299628e-01 1.38034418e-01
-9.81079519e-01 2.68992096e-01 -7.12143600e-01 -4.93547380e-01
2.72944719e-01 -1.46343961e-01 -3.45913559e-01 -1.58127427e-01
-8.35191011e-01 7.91447759e-01 3.86366069e-01 -2.42442936e-01
-1.50259042e+00 -7.83814609e-01 -1.05791616e+00 8.84055346e-02
7.18584776e-01 -9.00656819e-01 1.41295731e+00 -1.15046620e+00
-1.53056777e+00 9.24975157e-01 -3.13310146e-01 -2.71619588e-01
2.24568054e-01 -5.45624783e-03 -1.95718512e-01 2.06764430e-01
6.02912381e-02 8.87401521e-01 1.32690692e+00 -1.75781465e+00
-5.98719001e-01 -2.17786178e-01 3.03841531e-01 1.31532280e-02
-2.75134891e-01 3.29538458e-03 -5.34635544e-01 -5.09624124e-01
6.93737194e-02 -7.40515053e-01 -7.51758665e-02 3.29487115e-01
-3.17107409e-01 -3.31952721e-01 7.13564456e-01 -3.08435291e-01
8.68445694e-01 -2.19786906e+00 4.68369126e-01 -2.27343082e-01
1.36727065e-01 2.45567381e-01 -5.97200215e-01 5.69739461e-01
-3.29362787e-02 -3.43696922e-01 -1.69790208e-01 -7.43618965e-01
2.08384290e-01 4.45526838e-01 -6.28988802e-01 3.93539906e-01
4.12945956e-01 1.00008655e+00 -9.89612281e-01 -6.59090340e-01
2.47683659e-01 2.95012116e-01 -5.92794478e-01 4.50350702e-01
-6.89426422e-01 3.23279411e-01 -7.42217619e-03 6.53698146e-01
2.67128080e-01 -4.25064743e-01 2.79072911e-01 -4.50228125e-01
-1.89001575e-01 2.42964283e-01 -9.86735642e-01 2.01821947e+00
-4.52108294e-01 2.86115795e-01 -1.17561035e-01 -9.49472785e-01
4.90430802e-01 3.59646142e-01 3.41361582e-01 -6.19040430e-01
1.31903529e-01 -1.41003430e-01 -2.95213331e-02 -5.78311563e-01
3.10325205e-01 -2.50067592e-01 1.03777340e-02 5.27756929e-01
7.23475754e-01 1.12333953e-01 1.71732545e-01 4.47912037e-01
8.11789513e-01 5.77992976e-01 4.14133340e-01 -1.83081627e-02
5.17287791e-01 -1.26424149e-01 4.81660783e-01 6.59255922e-01
-3.64394426e-01 6.77353740e-01 3.79622340e-01 -2.76719093e-01
-9.73361015e-01 -1.34887683e+00 3.52915555e-01 1.52757144e+00
2.88393706e-01 -5.73104084e-01 -3.52853119e-01 -7.68206120e-01
-1.08776100e-01 8.65683138e-01 -7.32016861e-01 -2.76929677e-01
-5.35995722e-01 -8.24543759e-02 3.97191793e-01 5.70176899e-01
1.06930954e-03 -1.07926822e+00 -8.35870385e-01 7.61486264e-03
-3.74184549e-01 -1.41404068e+00 -4.87664998e-01 2.48186499e-01
-6.31282032e-01 -1.02644873e+00 -5.37171960e-01 -9.09651160e-01
4.95027095e-01 4.29397523e-01 1.02854884e+00 -1.04925774e-01
-2.03273714e-01 7.99504578e-01 -3.62732828e-01 -1.62664920e-01
-4.20935303e-01 -1.67035922e-01 -7.47215375e-02 5.25504172e-01
1.91341549e-01 -4.43152070e-01 -4.49560851e-01 -1.15035996e-01
-8.15818369e-01 4.63385582e-01 5.91187119e-01 8.05188894e-01
6.05602264e-01 -6.41769707e-01 2.80089557e-01 -6.77654564e-01
1.64398730e-01 -3.58321190e-01 -3.72363269e-01 5.91073036e-01
-4.68133092e-01 4.09919918e-01 4.99355197e-01 -9.37626064e-01
-1.19703722e+00 2.94400781e-01 2.62512296e-01 -1.04817653e+00
-3.70206147e-01 1.68509364e-01 -2.39591181e-01 1.69844747e-01
5.14634192e-01 5.48516452e-01 -1.71338484e-01 -6.50375724e-01
9.69420314e-01 1.30153760e-01 7.59131074e-01 -8.12191308e-01
8.03199947e-01 7.63143241e-01 -1.27165094e-01 -6.65319800e-01
-9.18013096e-01 -7.51572847e-01 -5.22609830e-01 -2.49666065e-01
8.77157271e-01 -1.00609076e+00 -4.69294310e-01 3.09522808e-01
-1.31699169e+00 -4.20873076e-01 -5.30420423e-01 1.33736059e-01
-8.19167018e-01 5.23323298e-01 -5.32522440e-01 -7.02556193e-01
-2.37257153e-01 -1.11234415e+00 1.26852906e+00 -8.86349007e-02
-2.61555374e-01 -6.86556935e-01 2.59465188e-01 1.84845850e-01
1.17928214e-01 -9.83913913e-02 1.23118174e+00 -5.81557512e-01
-7.22803414e-01 2.30039805e-01 -4.16123301e-01 1.43868849e-01
-1.39920592e-01 3.12416237e-02 -1.25377274e+00 -8.99061486e-02
-6.75250292e-02 -5.79291880e-01 1.21447742e+00 -1.98924616e-02
9.73113537e-01 -2.05363154e-01 -2.12700173e-01 4.68535423e-01
1.53805602e+00 -1.47688553e-01 4.42658335e-01 1.86749578e-01
9.00438070e-01 8.27277958e-01 3.36304128e-01 2.66551197e-01
7.15226710e-01 6.93883777e-01 6.03354752e-01 2.86728382e-01
-6.63618565e-01 -6.35793388e-01 6.56720996e-01 7.34370470e-01
-2.13957299e-02 -5.94673231e-02 -6.50239170e-01 8.22401524e-01
-2.02766562e+00 -1.09830964e+00 1.41586021e-01 1.88276088e+00
9.86143589e-01 -1.42214596e-01 7.57986233e-02 -2.27531821e-01
4.91139412e-01 5.70720553e-01 -5.12448907e-01 -6.95373341e-02
-4.79349568e-02 2.39001885e-01 1.52527004e-01 4.03529614e-01
-1.03433180e+00 1.19783771e+00 5.78216457e+00 6.25014603e-01
-1.24666774e+00 4.26950037e-01 -8.35499913e-02 -1.37713820e-01
-5.11169732e-01 3.48774433e-01 -7.65687704e-01 1.53734401e-01
6.92841053e-01 1.69595778e-01 4.79261220e-01 7.76600182e-01
-1.30915895e-01 1.27718747e-01 -1.56486034e+00 1.04333496e+00
3.97993803e-01 -1.43942332e+00 4.37517405e-01 -2.45150298e-01
7.89884329e-01 -1.47394734e-02 1.59141690e-01 5.35799563e-01
3.60370845e-01 -5.22030711e-01 1.32403278e+00 6.85867786e-01
6.74675882e-01 -7.22622778e-03 1.84793007e-02 4.33441579e-01
-1.38304675e+00 -2.47211516e-01 -2.43201852e-02 3.92682068e-02
2.06838548e-01 -1.52036563e-01 -5.37201166e-01 5.64111888e-01
3.76185536e-01 7.19103158e-01 -6.20485961e-01 8.13419998e-01
-3.03542793e-01 3.61566246e-01 9.39874128e-02 2.78934330e-01
2.79099643e-01 5.70696518e-02 5.78096747e-01 1.11619031e+00
-6.43096343e-02 -2.37740621e-01 5.22130847e-01 1.21375108e+00
1.58045009e-01 -7.85241500e-02 -4.04665947e-01 -2.01988950e-01
1.07138127e-01 1.25753641e+00 -5.17957389e-01 -5.88503957e-01
-6.94906473e-01 9.09256756e-01 5.64552784e-01 5.92306852e-01
-7.35658288e-01 3.02488297e-01 6.84372604e-01 4.24704291e-02
6.91907287e-01 -2.68310040e-01 -1.33530095e-01 -1.60435808e+00
-2.34025851e-01 -9.23362076e-01 3.74045253e-01 -6.51112139e-01
-1.40326500e+00 3.99661034e-01 1.80880353e-01 -1.18554688e+00
-2.70188153e-01 -6.14863873e-01 -6.85074925e-01 5.67870259e-01
-1.71882617e+00 -1.85967112e+00 -2.77455568e-01 8.55612040e-01
8.93224001e-01 -2.37401146e-02 7.93387949e-01 7.50817135e-02
-5.35840750e-01 4.49441314e-01 -4.81695890e-01 -8.09449032e-02
8.60979974e-01 -1.07789981e+00 2.65698969e-01 9.57020223e-01
4.79868799e-01 7.46812642e-01 8.16832304e-01 -5.30628979e-01
-1.61100030e+00 -1.12074256e+00 6.73732996e-01 -6.21093690e-01
9.80776906e-01 -6.77994132e-01 -8.41732323e-01 8.15476239e-01
2.21603215e-01 2.76210278e-01 5.95209837e-01 3.85424010e-02
-1.15289748e+00 6.56701997e-02 -6.09829187e-01 8.11065555e-01
1.18831134e+00 -9.88072634e-01 -7.82369912e-01 1.27809122e-02
8.61578763e-01 -9.52604488e-02 -3.40004802e-01 3.21362466e-01
6.57455027e-01 -8.86568189e-01 1.00641811e+00 -9.38013196e-01
4.88675565e-01 -4.87867981e-01 -4.33459044e-01 -1.00632477e+00
-4.26790953e-01 -1.84681356e-01 -7.20536768e-01 1.15850532e+00
1.10733740e-01 -2.26243064e-01 4.85717237e-01 2.80250221e-01
-2.49047011e-01 -6.49461389e-01 -6.91894233e-01 -8.23505402e-01
-1.42494455e-01 -4.64002043e-01 4.03476834e-01 8.29143405e-01
-1.29534468e-01 6.96099341e-01 -3.91524881e-01 1.68422788e-01
6.87135577e-01 5.05173028e-01 7.77806044e-01 -1.13984871e+00
-5.85998178e-01 -5.01424253e-01 -3.23079497e-01 -1.05834043e+00
5.12350440e-01 -9.42225754e-01 2.80089647e-01 -1.52996731e+00
4.95850444e-01 -1.24355048e-01 -4.65958148e-01 7.71785676e-01
-1.26494482e-01 3.97614054e-02 5.30351222e-01 2.81143367e-01
-9.29878414e-01 7.67711878e-01 1.19587553e+00 -5.33020258e-01
1.49435387e-03 -4.14512098e-01 -6.10710680e-01 6.85746610e-01
3.30315948e-01 -8.80389810e-02 -6.08605742e-01 -5.67911446e-01
-2.67070271e-02 -3.62167299e-01 7.76541829e-01 -7.62981594e-01
3.25186491e-01 -3.37453574e-01 -9.26692113e-02 -4.90588188e-01
5.87202728e-01 -9.28164542e-01 2.04130001e-02 3.33259732e-01
-6.66467369e-01 -4.07036871e-01 -1.81502290e-02 9.33013618e-01
3.25039737e-02 -2.31916323e-01 7.84023941e-01 -2.31849566e-01
-1.31758082e+00 4.63761955e-01 -2.29851454e-01 -1.62606448e-01
1.12964439e+00 -1.21343374e-01 -5.62671304e-01 -2.06237108e-01
-8.15555215e-01 1.29452094e-01 6.55750692e-01 6.71407759e-01
6.25157595e-01 -1.28507829e+00 -3.12701672e-01 2.43467942e-01
9.01244044e-01 -1.67257294e-01 4.37013656e-01 8.13244164e-01
1.06060423e-01 2.46702060e-01 -2.48380110e-01 -6.29186332e-01
-1.10500658e+00 1.09845126e+00 2.61853844e-01 -7.14522377e-02
-9.21169817e-01 6.53402269e-01 7.21566200e-01 -3.15961689e-01
3.42413753e-01 -4.32190895e-01 -1.27932712e-01 1.49220079e-01
5.82850695e-01 1.24168739e-01 -2.72742331e-01 -9.87715662e-01
-2.18155697e-01 5.27299941e-01 -2.60467350e-01 -2.98293114e-01
9.79805529e-01 -1.33779094e-01 -7.40331933e-02 9.69995499e-01
9.89002228e-01 -3.17775398e-01 -1.55240655e+00 -7.16574848e-01
8.93159136e-02 -2.40071967e-01 -8.81235451e-02 -5.76298952e-01
-8.26345325e-01 1.05931497e+00 5.41383326e-01 -1.96827352e-01
7.49205589e-01 4.41313475e-01 6.20237231e-01 2.93094188e-01
3.98438543e-01 -8.12451482e-01 4.69519675e-01 5.96969783e-01
7.02293932e-01 -1.18176818e+00 -3.52516383e-01 -1.72850952e-01
-7.79803038e-01 9.43997264e-01 8.31344366e-01 -1.44298330e-01
3.27352673e-01 1.74836013e-02 5.15947044e-02 -2.03956991e-01
-1.13741243e+00 -6.47002399e-01 3.18920225e-01 7.32837558e-01
7.32701793e-02 4.98495400e-02 5.45102777e-03 6.77883983e-01
3.65998060e-01 -1.90514848e-01 3.45276952e-01 9.86917973e-01
-4.81517166e-01 -1.08095181e+00 -1.26991168e-01 7.48568848e-02
1.06852964e-01 -1.53034329e-01 -4.51818466e-01 4.89231229e-01
2.34964326e-01 6.69981778e-01 1.26602739e-01 -2.52967119e-01
1.49957225e-01 4.93864000e-01 7.33556747e-01 -9.77462709e-01
-3.79027396e-01 9.65749174e-02 -9.93422046e-02 -6.80875838e-01
-6.92448735e-01 -5.93878865e-01 -1.14054811e+00 2.39514634e-01
-1.43534571e-01 -4.76021141e-01 4.99379396e-01 1.16208684e+00
2.50180393e-01 6.11153901e-01 1.87399462e-01 -9.30434763e-01
-6.77314997e-01 -7.91745603e-01 -2.42255092e-01 6.80167139e-01
5.22298992e-01 -1.01466608e+00 -1.48971394e-01 2.95090497e-01] | [10.229567527770996, 2.1892056465148926] |
27507987-c14e-496e-bc62-49d04dd23916 | from-semi-supervised-to-omni-supervised-room | 2301.13865 | null | https://arxiv.org/abs/2301.13865v1 | https://arxiv.org/pdf/2301.13865v1.pdf | From Semi-supervised to Omni-supervised Room Layout Estimation Using Point Clouds | Room layout estimation is a long-existing robotic vision task that benefits both environment sensing and motion planning. However, layout estimation using point clouds (PCs) still suffers from data scarcity due to annotation difficulty. As such, we address the semi-supervised setting of this task based upon the idea of model exponential moving averaging. But adapting this scheme to the state-of-the-art (SOTA) solution for PC-based layout estimation is not straightforward. To this end, we define a quad set matching strategy and several consistency losses based upon metrics tailored for layout quads. Besides, we propose a new online pseudo-label harvesting algorithm that decomposes the distribution of a hybrid distance measure between quads and PC into two components. This technique does not need manual threshold selection and intuitively encourages quads to align with reliable layout points. Surprisingly, this framework also works for the fully-supervised setting, achieving a new SOTA on the ScanNet benchmark. Last but not least, we also push the semi-supervised setting to the realistic omni-supervised setting, demonstrating significantly promoted performance on a newly annotated ARKitScenes testing set. Our codes, data and models are released in this repository. | ['Hongbin Zha', 'Yurong Chen', 'Guyue Zhou', 'Hao Zhao', 'Xiaoxue Chen', 'Pengfei Li', 'Beiwen Tian', 'Huan-ang Gao'] | 2023-01-31 | null | null | null | null | ['set-matching', 'motion-planning'] | ['computer-vision', 'robots'] | [ 2.52140373e-01 1.88487805e-02 -1.10258400e-01 -5.20636022e-01
-8.33991528e-01 -7.66448140e-01 4.73179936e-01 2.16232777e-01
-3.70150447e-01 4.59505051e-01 1.80458769e-01 -2.14083314e-01
-3.40142518e-01 -5.54099679e-01 -8.46690655e-01 -7.54105449e-01
3.38056982e-02 6.16972983e-01 2.25837499e-01 9.54431089e-05
3.58424544e-01 5.76227546e-01 -1.44132400e+00 -2.84819484e-01
1.12692845e+00 9.26598012e-01 5.28910518e-01 4.28036809e-01
-3.78616452e-02 2.30441049e-01 -1.52251408e-01 -7.03153312e-02
4.26798314e-01 1.73944175e-01 -6.39589667e-01 2.30240494e-01
1.10675204e+00 1.64286811e-02 -2.75896639e-01 8.46723258e-01
7.26386249e-01 2.00340465e-01 6.79566801e-01 -1.50610363e+00
-3.34542483e-01 3.10584217e-01 -6.99228227e-01 -2.52363801e-01
1.98966995e-01 4.08673845e-02 1.13121700e+00 -8.53312850e-01
5.78341544e-01 9.14846480e-01 8.58124018e-01 2.14090317e-01
-1.32574904e+00 -4.79291886e-01 5.72519183e-01 -2.28058361e-02
-1.46062183e+00 -4.90552157e-01 9.18192744e-01 -4.51158017e-01
5.84962785e-01 1.96797699e-01 4.97888774e-01 1.09778523e+00
-2.87661850e-01 9.38488483e-01 1.15256941e+00 -4.35136944e-01
6.31892562e-01 1.68427750e-02 1.18662909e-01 9.60249603e-01
3.15246105e-01 -2.46304110e-01 -5.65636039e-01 -1.12515064e-02
4.70138788e-01 -1.16921872e-01 -4.58677858e-01 -1.23603916e+00
-1.44774282e+00 5.73479891e-01 5.94379961e-01 1.76103152e-02
-1.19436033e-01 2.75219411e-01 9.00581777e-02 -3.05945486e-01
3.12431365e-01 5.14278412e-01 -5.15641868e-01 -7.36107677e-02
-1.03858805e+00 2.52716482e-01 7.16066837e-01 1.46243000e+00
7.86368906e-01 -4.25279111e-01 -1.89118996e-01 8.12103331e-01
3.86204869e-01 7.24921584e-01 -1.46488994e-01 -1.15740538e+00
7.02126801e-01 4.32673186e-01 2.78248310e-01 -1.13308024e+00
-6.87945604e-01 -6.10819876e-01 -6.64343834e-01 1.65018469e-01
5.95750988e-01 1.13905974e-01 -7.84768820e-01 1.78805780e+00
4.47998494e-01 1.31741673e-01 -2.81763047e-01 8.44586492e-01
5.58506370e-01 1.84499800e-01 -2.24450901e-01 -4.77561951e-02
1.03996444e+00 -1.41019499e+00 -5.02138674e-01 -5.13862073e-01
6.20275199e-01 -4.84273583e-01 1.30595696e+00 4.56322938e-01
-6.66214943e-01 -4.31134939e-01 -1.20175493e+00 -1.37518778e-01
-3.35509747e-01 3.79520535e-01 7.82023966e-01 6.29376829e-01
-1.07344854e+00 4.86185282e-01 -7.49123096e-01 -6.72730803e-01
5.79238415e-01 1.80537313e-01 -2.46252254e-01 -4.88243431e-01
-4.39150035e-01 6.79543793e-01 1.35356337e-01 1.82147264e-01
-4.49535966e-01 -6.75475597e-01 -9.82647419e-01 -2.03403994e-01
6.47216439e-01 -8.11115086e-01 1.02129745e+00 -3.96673948e-01
-1.27429259e+00 6.58121943e-01 -2.60836512e-01 -2.18552366e-01
5.93305051e-01 -2.87649810e-01 7.19386190e-02 -1.25626385e-01
3.24153602e-01 9.02413845e-01 6.55419827e-01 -1.73092520e+00
-5.88213682e-01 -4.63825792e-01 1.10622689e-01 3.39898556e-01
-3.06805372e-01 -7.63591230e-01 -7.97295511e-01 -3.19990307e-01
4.88615602e-01 -1.25409389e+00 -3.62498164e-01 3.94145489e-01
-6.62503779e-01 -1.98184475e-01 6.00086927e-01 -3.81537259e-01
1.06571972e+00 -2.06072831e+00 1.48474514e-01 3.02889287e-01
2.18142524e-01 -2.36406818e-01 -9.36903283e-02 2.86922157e-01
3.22799623e-01 -1.94559842e-01 -6.20328486e-01 -9.62059557e-01
3.54344875e-01 2.48765171e-01 -1.72164887e-01 7.34395921e-01
-9.61482003e-02 8.74544501e-01 -1.13679481e+00 -6.92264020e-01
5.07161021e-01 1.77017272e-01 -6.31537855e-01 -4.17770557e-02
-2.27245092e-01 4.22104836e-01 -3.84384394e-01 8.49087059e-01
9.11720097e-01 -2.97010988e-01 6.55928701e-02 -3.47710133e-01
-3.17435741e-01 -7.49736419e-03 -1.21385717e+00 2.50519705e+00
-5.28001130e-01 6.01288676e-01 7.09718019e-02 -7.12838829e-01
7.28756607e-01 -3.09272885e-01 6.60564601e-01 -5.17456055e-01
-1.41542554e-01 1.20629124e-01 -4.73957986e-01 -4.42694545e-01
6.50774598e-01 3.53186637e-01 -2.30216384e-01 1.05214536e-01
-2.17997923e-01 -3.75702530e-01 -2.31796950e-02 9.28488448e-02
1.24427485e+00 7.48161197e-01 1.95900306e-01 -3.49229604e-01
2.77004004e-01 3.55221748e-01 4.19614285e-01 7.45630562e-01
-4.02009308e-01 9.62216020e-01 3.50207798e-02 -3.78450640e-02
-9.64695573e-01 -1.09214902e+00 -1.90781564e-01 8.53248954e-01
5.01435161e-01 -4.69268113e-01 -7.04579353e-01 -9.99264896e-01
6.60209805e-02 7.17465520e-01 -4.83702928e-01 2.46465981e-01
-4.61097300e-01 -3.94067675e-01 2.74527133e-01 6.36805415e-01
4.36551034e-01 -7.43133187e-01 -7.42906511e-01 -2.96277925e-02
-4.37499374e-01 -1.41108203e+00 -6.45473540e-01 4.09596413e-01
-4.76897746e-01 -1.02554977e+00 -6.12098575e-01 -7.50763834e-01
7.98710704e-01 7.05115974e-01 1.11216688e+00 -1.75473347e-01
-1.13605723e-01 8.75315547e-01 -2.83075184e-01 -4.00725424e-01
2.20396653e-01 4.62494403e-01 7.14010522e-02 -1.89089239e-01
8.43044296e-02 -5.70141017e-01 -7.52677202e-01 2.37124205e-01
-3.34007859e-01 2.91145921e-01 6.92184746e-01 5.12070417e-01
8.52378130e-01 -3.02262902e-02 1.44965306e-01 -5.08797586e-01
8.80951881e-02 -2.41179839e-01 -6.82683110e-01 2.47926012e-01
-8.59326482e-01 2.34392956e-01 4.72305119e-01 -7.91775584e-02
-8.33054483e-01 6.42816424e-01 1.27721667e-01 -5.05545497e-01
-3.43968809e-01 2.04421327e-01 -5.35672605e-01 -1.64931402e-01
4.01362747e-01 4.57359366e-02 -1.88062429e-01 -3.70927930e-01
6.47000372e-01 4.70386058e-01 6.84414804e-01 -5.92547536e-01
1.14100313e+00 7.89920986e-01 2.87281990e-01 -8.45156312e-01
-9.81384635e-01 -9.13078725e-01 -1.05687714e+00 -2.82343715e-01
8.33529294e-01 -7.70997584e-01 -6.66817307e-01 2.32882604e-01
-1.11677265e+00 -5.83631158e-01 -2.16359869e-01 2.44394287e-01
-7.63021648e-01 5.46498954e-01 8.12759995e-02 -9.57774818e-01
-1.28067452e-02 -1.07277369e+00 1.54040658e+00 1.87326074e-02
-3.99327558e-03 -9.92946625e-01 2.32297182e-01 3.80346030e-01
1.25021040e-01 3.88437212e-01 5.70029080e-01 -1.63443178e-01
-8.53664696e-01 4.09990847e-02 -3.71760637e-01 6.03694655e-02
6.58727586e-02 -3.36918443e-01 -1.24573290e+00 -2.22614139e-01
-4.59162682e-01 -8.60227868e-02 9.95393991e-01 4.84924436e-01
1.34221709e+00 2.92928945e-02 -6.11028910e-01 7.74168670e-01
1.55106926e+00 -3.58904064e-01 3.51453781e-01 6.32812262e-01
1.21123433e+00 8.41307759e-01 9.03221011e-01 3.26446474e-01
9.10836399e-01 1.02182257e+00 8.07173550e-01 -8.29228312e-02
-2.44519949e-01 -4.25625443e-01 8.95116031e-02 7.48606205e-01
1.07838169e-01 -2.73374796e-01 -1.07518530e+00 5.28361142e-01
-2.15687752e+00 -6.32295310e-01 -3.82853836e-01 2.34711027e+00
5.49728811e-01 7.89511055e-02 -5.13463467e-03 8.14636648e-02
3.31759572e-01 3.16962034e-01 -5.52158594e-01 2.84691930e-01
-1.03368647e-01 -1.25138029e-01 1.05491138e+00 3.55135769e-01
-1.50930429e+00 9.17773843e-01 5.22489405e+00 7.12609410e-01
-7.92603791e-01 6.18823282e-02 3.39035362e-01 5.02473898e-02
-3.42294008e-01 1.43560134e-02 -8.18476677e-01 2.60656059e-01
4.72322971e-01 4.23560917e-01 6.23026848e-01 1.19730413e+00
2.44028494e-01 -2.44619310e-01 -1.21128118e+00 1.31353855e+00
2.57698476e-01 -1.02502584e+00 -4.75723416e-01 3.51366341e-01
6.42664015e-01 9.98484343e-02 8.19778442e-02 1.82556704e-01
2.14205265e-01 -8.71387184e-01 1.05778003e+00 3.56336951e-01
8.37534547e-01 -5.36594033e-01 3.58453959e-01 3.36432248e-01
-1.45606112e+00 -2.49973223e-01 -3.48473817e-01 2.54493535e-01
1.80298865e-01 6.64021850e-01 -8.24742258e-01 7.95006454e-01
7.67660141e-01 8.38918507e-01 -1.00088155e+00 1.26887655e+00
-1.22244805e-01 2.75808632e-01 -4.52311993e-01 1.30236924e-01
2.14157164e-01 -2.75838256e-01 4.10268575e-01 1.17733240e+00
2.90617973e-01 -4.15693700e-01 4.28824067e-01 7.64193416e-01
9.74113941e-02 4.34729159e-02 -8.26651871e-01 4.37248945e-01
7.31918931e-01 1.64879537e+00 -9.38379586e-01 8.02228786e-03
-1.52634934e-01 9.76179361e-01 5.67080319e-01 3.04699302e-01
-9.16580081e-01 -1.53855175e-01 3.84549022e-01 1.36661783e-01
3.57132643e-01 -6.38588607e-01 -7.58440495e-01 -9.65343595e-01
2.11264268e-01 -4.92083371e-01 -6.47391379e-02 -7.88042009e-01
-1.21189845e+00 3.10895741e-01 1.59565940e-01 -1.53405797e+00
1.51427031e-01 -5.95758796e-01 -2.79992312e-01 3.29771489e-01
-1.83844423e+00 -1.44919240e+00 -1.00019872e+00 2.27504939e-01
7.17893362e-01 3.37975800e-01 6.69044614e-01 3.04143250e-01
-6.48338139e-01 5.15242696e-01 5.19495010e-02 -6.87837675e-02
9.15121615e-01 -1.64296138e+00 3.53581280e-01 9.27129090e-01
2.54007936e-01 4.94348079e-01 7.89384663e-01 -5.37075996e-01
-1.52907300e+00 -1.44657171e+00 6.63622200e-01 -8.50490689e-01
3.57730895e-01 -6.71639740e-01 -5.01846731e-01 5.27622342e-01
-8.10341090e-02 7.22054467e-02 3.24866027e-01 2.24150538e-01
-2.35157281e-01 -1.01921551e-01 -9.65797544e-01 7.62622476e-01
1.82001984e+00 -2.83256859e-01 -1.13014884e-01 4.82344031e-01
7.41903722e-01 -5.43065012e-01 -6.42531395e-01 3.52661997e-01
5.08348882e-01 -8.81295919e-01 1.15650475e+00 2.10677847e-01
2.48877600e-01 -6.04827702e-01 -5.51544428e-01 -1.24347496e+00
-2.91202188e-01 -4.55027699e-01 -1.21288568e-01 1.28397822e+00
3.00184101e-01 -3.96787643e-01 9.00688469e-01 5.79378247e-01
-5.90586782e-01 -6.64615989e-01 -8.68655264e-01 -8.62405300e-01
-3.37146431e-01 -7.27453232e-01 4.46373791e-01 8.80899847e-01
-1.69101536e-01 3.51847619e-01 -2.00597420e-01 4.90759581e-01
1.14302564e+00 1.15463510e-01 1.07726598e+00 -1.43117392e+00
-1.57509699e-01 -4.09615785e-01 -4.07519601e-02 -1.40785563e+00
2.37095669e-01 -9.11498249e-01 5.85180342e-01 -1.95634973e+00
9.42464173e-02 -1.01228857e+00 -4.68545258e-02 3.55346799e-01
-3.48943248e-02 1.08734943e-01 1.38879538e-01 2.93167233e-01
-9.98995900e-01 7.14417398e-01 1.07836688e+00 -3.25149208e-01
-2.58957744e-01 -1.05757385e-01 -4.82406855e-01 6.76233292e-01
7.53321528e-01 -2.48987854e-01 -6.47048891e-01 -3.52567375e-01
2.11030185e-01 -3.68093699e-01 5.03089190e-01 -1.30199897e+00
4.48727250e-01 -1.19114615e-01 1.77249223e-01 -1.04099178e+00
4.44975913e-01 -1.01519310e+00 -1.12153381e-01 9.85179618e-02
-2.29389414e-01 -4.26844880e-02 8.34476575e-02 1.00607216e+00
3.73263389e-01 -9.96343344e-02 5.02475679e-01 4.29183356e-02
-9.01898801e-01 5.06358027e-01 1.77982673e-01 -7.13712946e-02
8.25005352e-01 -4.58117902e-01 -3.18279475e-01 -1.10464074e-01
-2.71106839e-01 4.67724293e-01 9.28356230e-01 4.83755291e-01
6.02824211e-01 -1.31110656e+00 -2.98552275e-01 -2.88733398e-03
5.92059910e-01 6.66482270e-01 1.39613032e-01 1.24004197e+00
-4.14991677e-01 5.24298489e-01 2.14391798e-01 -8.92864168e-01
-1.15003312e+00 6.83131933e-01 6.19951449e-02 -1.75783649e-01
-8.20463061e-01 6.36426508e-01 2.39643127e-01 -7.55955160e-01
5.88306725e-01 -5.68409383e-01 -2.70657748e-01 -6.48963451e-02
4.61892970e-02 4.42391157e-01 8.55418146e-02 -5.88546991e-01
-5.94057560e-01 8.07619989e-01 3.28308672e-01 -1.15224533e-01
1.23889518e+00 -5.57780266e-01 1.30933180e-01 5.66724360e-01
1.09785080e+00 3.64917934e-01 -1.62280047e+00 -1.93297088e-01
2.68115759e-01 -4.27728415e-01 4.17916141e-02 -5.83386600e-01
-7.23327696e-01 6.68575287e-01 7.01577902e-01 -1.19394325e-01
7.95228243e-01 5.32000437e-02 6.27913177e-01 5.60132384e-01
8.13414395e-01 -1.17141485e+00 1.69201076e-01 3.23949188e-01
8.47314715e-01 -1.49793422e+00 1.55510262e-01 -5.58215141e-01
-6.01349771e-01 8.96641791e-01 5.83016276e-01 3.42944451e-02
5.05543590e-01 1.37672991e-01 -1.85917452e-01 -3.18579197e-01
-3.26022714e-01 -3.77789885e-01 2.67085671e-01 9.58863676e-01
6.44724304e-03 1.06296562e-01 1.40656054e-01 1.72165528e-01
-3.26732367e-01 -1.60698190e-01 3.86878431e-01 9.66690242e-01
-3.74129206e-01 -8.21994841e-01 -3.76296610e-01 2.82122612e-01
3.04766685e-01 -7.05445409e-02 -1.52038693e-01 6.96489930e-01
1.19561493e-01 9.87325847e-01 2.70749023e-03 -4.59235549e-01
4.61256862e-01 -1.72628447e-01 5.83985507e-01 -5.11581898e-01
1.42582441e-02 5.48050441e-02 1.32177919e-01 -7.45061994e-01
-5.87480724e-01 -9.56323624e-01 -1.04478180e+00 3.29222828e-02
-3.69418561e-01 -2.27749348e-01 9.90434945e-01 7.56004095e-01
3.78102422e-01 6.42785549e-01 6.13964021e-01 -1.30378819e+00
-4.21601057e-01 -7.73356080e-01 -3.72356623e-01 1.55150548e-01
3.29222679e-01 -9.60446179e-01 -2.47516438e-01 -1.88940018e-01] | [8.174554824829102, -2.8444690704345703] |
4af85b3b-1f83-42a4-b233-3cd464137c3e | fooling-vision-and-language-models-despite | 1709.08693 | null | http://arxiv.org/abs/1709.08693v2 | http://arxiv.org/pdf/1709.08693v2.pdf | Fooling Vision and Language Models Despite Localization and Attention Mechanism | Adversarial attacks are known to succeed on classifiers, but it has been an
open question whether more complex vision systems are vulnerable. In this
paper, we study adversarial examples for vision and language models, which
incorporate natural language understanding and complex structures such as
attention, localization, and modular architectures. In particular, we
investigate attacks on a dense captioning model and on two visual question
answering (VQA) models. Our evaluation shows that we can generate adversarial
examples with a high success rate (i.e., > 90%) for these models. Our work
sheds new light on understanding adversarial attacks on vision systems which
have a language component and shows that attention, bounding box localization,
and compositional internal structures are vulnerable to adversarial attacks.
These observations will inform future work towards building effective defenses. | ['Anna Rohrbach', 'Chang Liu', 'Xinyun Chen', 'Dawn Song', 'Xiaojun Xu', 'Trevor Darrell'] | 2017-09-25 | fooling-vision-and-language-models-despite-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Xu_Fooling_Vision_and_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Xu_Fooling_Vision_and_CVPR_2018_paper.pdf | cvpr-2018-6 | ['dense-captioning'] | ['computer-vision'] | [ 3.37929696e-01 4.93573517e-01 4.12885904e-01 -1.84200421e-01
-9.03387189e-01 -1.28447604e+00 9.64078605e-01 -3.44782501e-01
-1.56427562e-01 3.66985530e-01 2.26993531e-01 -8.75465453e-01
5.13107955e-01 -7.59091973e-01 -9.99623895e-01 -4.77795869e-01
2.89963633e-02 2.27764934e-01 3.76725852e-01 -3.85035604e-01
1.02996737e-01 7.77440667e-01 -8.84876251e-01 3.16013813e-01
4.69829112e-01 6.28158092e-01 -5.32287478e-01 1.33668828e+00
2.26002589e-01 1.34004033e+00 -1.07877433e+00 -1.21976674e+00
4.36648369e-01 -1.75077319e-01 -9.19340312e-01 -2.86720041e-02
1.44067776e+00 -5.37096381e-01 -7.93743134e-01 1.29339373e+00
2.26112366e-01 -3.61537158e-01 5.93235373e-01 -1.51197624e+00
-1.42755735e+00 4.23962504e-01 -3.47363561e-01 3.25190306e-01
1.84290364e-01 1.04257739e+00 9.00437415e-01 -3.60452771e-01
3.36571813e-01 1.66587889e+00 6.43577278e-01 1.15915430e+00
-1.25976849e+00 -6.34782255e-01 3.13959062e-01 2.21224576e-01
-9.50819731e-01 -3.72677565e-01 4.89140272e-01 -4.90069121e-01
9.49901998e-01 4.17750955e-01 9.03449804e-02 1.58065557e+00
5.44147909e-01 5.57719290e-01 1.09968019e+00 -9.83173922e-02
1.10461220e-01 8.03103596e-02 1.41868964e-01 8.41604412e-01
2.28022724e-01 5.67240119e-01 -2.86231935e-02 -3.81535292e-01
5.15422285e-01 -4.64969695e-01 -2.85373211e-01 -2.12327570e-01
-6.30344868e-01 1.21269417e+00 8.75015259e-01 -1.63611054e-01
1.25992605e-02 7.76691437e-01 4.27704632e-01 3.35831255e-01
2.23785285e-02 9.45810378e-01 -2.93478668e-01 3.81680429e-01
-2.29314536e-01 3.24773073e-01 9.05246437e-01 8.55751157e-01
4.36041713e-01 4.89374340e-01 -3.31541568e-01 1.85976416e-01
2.47560397e-01 9.83988285e-01 -8.31086338e-02 -1.12862182e+00
3.16586614e-01 1.62801668e-01 -1.21067546e-01 -1.04327428e+00
-2.42299795e-01 -2.41634279e-01 -6.21854484e-01 8.70211899e-01
5.92314005e-01 -3.38628918e-01 -1.27753472e+00 2.04539657e+00
-1.56608477e-01 1.80795029e-01 4.16034400e-01 8.81938219e-01
7.89959252e-01 4.48349565e-01 4.01486725e-01 3.84816915e-01
1.45225203e+00 -9.07339931e-01 -2.60934442e-01 -7.18979537e-01
2.06679702e-01 -9.06274438e-01 9.16101992e-01 1.44655943e-01
-1.16884887e+00 -5.54892361e-01 -9.13330913e-01 -2.22373709e-01
-3.75435323e-01 -6.24920607e-01 6.09422684e-01 1.11963689e+00
-1.12288284e+00 2.11065814e-01 -8.44090044e-01 -3.24894667e-01
8.22837114e-01 2.48781741e-01 -2.31210932e-01 -1.85790077e-01
-1.19166875e+00 1.29073894e+00 -7.12815896e-02 -2.31974304e-01
-1.68612611e+00 -6.10600650e-01 -9.64217305e-01 -8.22875202e-02
1.87386289e-01 -1.11656213e+00 1.41343427e+00 -1.18481719e+00
-1.02838683e+00 8.90686989e-01 2.45160744e-01 -1.08063745e+00
3.80613774e-01 -6.97542354e-02 -2.88902044e-01 4.16204154e-01
-2.82357037e-01 9.17257607e-01 1.22223711e+00 -1.37531960e+00
-6.11719079e-02 -4.77447242e-01 8.51539671e-01 -9.88987088e-02
-2.89356589e-01 3.36838633e-01 2.59064361e-02 -6.97753787e-01
-5.60340822e-01 -1.24446368e+00 -4.12631631e-01 4.63416189e-01
-5.79549074e-01 2.41212785e-01 8.73541236e-01 -6.79685175e-01
4.15942758e-01 -2.11168647e+00 8.00214559e-02 3.19678672e-02
5.26849508e-01 6.40796363e-01 -4.49049145e-01 1.68404236e-01
-2.26723984e-01 5.80867589e-01 -2.64456213e-01 -1.04891136e-01
1.13008626e-01 3.89871508e-01 -1.08551764e+00 4.34023023e-01
7.14210391e-01 1.33764720e+00 -5.99959493e-01 -1.99141651e-01
1.08598188e-01 4.26563621e-01 -7.52212405e-01 3.75849873e-01
-5.55493534e-01 2.18066320e-01 -4.03573781e-01 7.17548192e-01
5.06799042e-01 -1.63287073e-01 -3.52697492e-01 -7.68277347e-02
4.14615691e-01 -1.62942395e-01 -3.07999313e-01 8.93526852e-01
-1.84153289e-01 1.07831562e+00 3.06176066e-01 -5.47794402e-01
5.18114150e-01 1.60092741e-01 -4.31631327e-01 -3.97624344e-01
2.03049287e-01 -3.24430496e-01 3.66324544e-01 -4.39677119e-01
1.93495035e-01 -1.59464732e-01 -2.07345158e-01 1.84416473e-01
-4.42608222e-02 -5.85884273e-01 -2.66913503e-01 6.75553083e-01
1.47471845e+00 -2.88531631e-01 -7.53867850e-02 6.55502751e-02
5.09042621e-01 2.50956714e-01 -1.02065981e-03 1.21831596e+00
-6.79018974e-01 7.75955677e-01 6.47667468e-01 -5.73420107e-01
-1.37385213e+00 -1.44476795e+00 3.93554904e-02 8.56524408e-01
1.02995196e-02 -3.09735090e-01 -1.05158377e+00 -8.58215630e-01
7.65658244e-02 7.36752331e-01 -8.56114030e-01 -7.65667677e-01
-5.18785954e-01 -4.54170823e-01 1.29544806e+00 7.53734112e-01
4.27033812e-01 -1.12372923e+00 -2.94051468e-01 -2.84068942e-01
2.07992971e-01 -1.37012553e+00 -3.83196086e-01 -6.72340319e-02
-4.41135377e-01 -1.27021062e+00 -5.30313313e-01 -6.67899489e-01
5.54696679e-01 4.20184694e-02 1.37359953e+00 1.96523219e-01
-4.21628743e-01 9.26425099e-01 -1.79521039e-01 -9.05314505e-01
-9.17140126e-01 -1.94695398e-01 -1.44565001e-01 -3.31214130e-01
2.93138117e-01 -3.28383327e-01 -4.32925403e-01 1.33077860e-01
-1.10511112e+00 -4.32950377e-01 5.03524899e-01 6.54721320e-01
6.26384616e-02 -4.74904060e-01 3.44686747e-01 -7.61220038e-01
6.68434858e-01 -3.94652456e-01 -8.19106400e-01 2.78080434e-01
-1.48287769e-02 1.75503809e-02 9.68227863e-01 -5.24176836e-01
-6.12036049e-01 4.24459716e-03 -4.25632954e-01 -7.90352404e-01
-4.45224196e-01 -1.82332560e-01 -2.14112118e-01 -6.83947206e-01
1.33746946e+00 5.68476766e-02 3.50386314e-02 1.41380847e-01
8.54437232e-01 2.97089785e-01 8.55338275e-01 -5.57326972e-01
1.48979378e+00 7.20261037e-01 -3.17207687e-02 -8.58941078e-01
-7.65659511e-01 3.00833285e-01 -6.79055080e-02 1.87127084e-01
1.23522854e+00 -7.40692556e-01 -8.78086686e-01 4.36451852e-01
-1.51209223e+00 -2.37371892e-01 -1.97053343e-01 -1.46628451e-02
-7.32945561e-01 7.01252401e-01 -7.59379745e-01 -7.63479054e-01
-3.31201822e-01 -1.22914362e+00 9.01550472e-01 1.89971492e-01
-1.83812037e-01 -9.21951413e-01 -1.39271215e-01 6.78692520e-01
5.41893303e-01 3.97827297e-01 8.97770822e-01 -8.31290126e-01
-1.16101062e+00 -6.16977587e-02 -3.66686642e-01 5.42697132e-01
-4.30917740e-01 1.24762267e-01 -1.30839348e+00 -4.28594589e-01
4.14225459e-02 -9.28255558e-01 9.72306669e-01 6.59911558e-02
1.26895881e+00 -5.15727103e-01 -3.92089365e-03 7.78277338e-01
1.24199963e+00 1.13888942e-01 9.96335745e-01 9.16288868e-02
8.40287805e-01 5.19438565e-01 -3.73827331e-02 -1.60199553e-01
2.23380961e-02 3.63112837e-01 1.14780343e+00 -1.89360559e-01
-2.91587979e-01 -4.92687523e-02 4.81830209e-01 -8.69662464e-02
3.42597723e-01 -5.00310242e-01 -1.05631340e+00 4.51027691e-01
-1.46617472e+00 -1.14918709e+00 -4.70244233e-03 1.61245358e+00
6.99403822e-01 4.51075137e-01 -8.42499211e-02 -3.27043325e-01
4.99609411e-01 2.85155416e-01 -8.58800530e-01 -8.69293332e-01
-2.30821878e-01 4.66231197e-01 5.73455453e-01 7.62622237e-01
-1.44276845e+00 1.29151034e+00 7.27160549e+00 2.93121666e-01
-8.41640949e-01 -6.20176792e-02 5.88649869e-01 2.86746286e-02
-3.37031335e-01 2.29911394e-02 -5.98766744e-01 1.60543561e-01
1.01256168e+00 -1.65756494e-01 6.09725893e-01 9.01754141e-01
-3.99752349e-01 5.51753044e-01 -1.09378219e+00 5.54543614e-01
4.98650521e-01 -1.28209186e+00 4.67398256e-01 1.56644955e-01
6.34644032e-01 2.22792372e-01 5.44669390e-01 3.81210864e-01
1.06225312e+00 -1.44564605e+00 4.77768451e-01 3.23632360e-01
5.98272324e-01 -6.69803500e-01 6.26372278e-01 1.27772763e-01
-5.97706854e-01 -1.95943892e-01 -4.57138479e-01 7.15930685e-02
1.10656440e-01 -2.62609988e-01 -7.36029446e-01 6.22616522e-02
5.76768100e-01 1.07771397e-01 -9.95119870e-01 6.44040644e-01
-4.33122218e-01 6.62547648e-01 -7.91929513e-02 1.10693939e-01
5.36377072e-01 4.44253594e-01 7.09200799e-01 9.50865388e-01
-3.89288872e-01 1.19064435e-01 2.09312066e-01 1.04065084e+00
-1.40490904e-01 -5.16994715e-01 -1.06424344e+00 -1.82323605e-01
7.20584020e-02 9.25866723e-01 -3.19925219e-01 -1.56191587e-01
-5.62117279e-01 9.49307919e-01 3.56825769e-01 4.55223560e-01
-1.15859509e+00 -2.18667790e-01 1.31836700e+00 -1.44688949e-01
3.41614306e-01 -3.22725952e-01 -3.42826068e-01 -1.27254725e+00
-3.03161949e-01 -1.38248610e+00 2.73906529e-01 -1.02290797e+00
-1.71206331e+00 8.30963850e-01 -3.50066632e-01 -5.29319167e-01
-2.91481078e-01 -1.07740092e+00 -8.00784230e-01 9.18369710e-01
-1.26436269e+00 -1.46771181e+00 -2.25983158e-01 8.72426808e-01
3.99522811e-01 -4.13583368e-01 1.06608891e+00 -1.43879905e-01
-4.37501490e-01 1.00004590e+00 -4.59730655e-01 6.18701279e-01
4.98004943e-01 -1.30230486e+00 1.16095972e+00 1.19774032e+00
4.92590785e-01 6.80347204e-01 8.18408787e-01 -4.25951123e-01
-1.33109462e+00 -1.30299532e+00 1.81083500e-01 -1.35480595e+00
9.22822773e-01 -2.85894543e-01 -9.68989730e-01 9.78583992e-01
6.09155297e-01 4.97343428e-02 6.62682056e-01 -1.79216981e-01
-1.05809402e+00 3.04013699e-01 -1.18659484e+00 1.05574083e+00
9.15000796e-01 -1.05019128e+00 -8.47244978e-01 5.46432972e-01
1.11146069e+00 -3.23946685e-01 -5.50277412e-01 3.20073366e-01
3.19756806e-01 -9.12727952e-01 1.42723274e+00 -1.29275680e+00
6.17729425e-01 -3.20012510e-01 -3.30848157e-01 -1.15627789e+00
-4.08570915e-01 -6.22049391e-01 -2.49411821e-01 9.53561842e-01
3.99394274e-01 -6.78584933e-01 6.22981906e-01 8.66478682e-01
-4.54015546e-02 -4.28208798e-01 -6.29355788e-01 -6.98598325e-01
6.75570130e-01 -4.19817001e-01 2.14424804e-01 7.88386405e-01
-6.57245457e-01 4.24682558e-01 -2.92380542e-01 8.04648280e-01
7.12981641e-01 -1.99021131e-01 9.48460460e-01 -7.07126856e-01
-4.73904967e-01 -3.93280596e-01 -9.35577214e-01 -7.13734388e-01
4.14983809e-01 -6.82914734e-01 -3.10260296e-01 -1.18009567e+00
1.77995693e-02 1.12142332e-01 1.60814561e-02 5.55551410e-01
-2.10755914e-01 7.94987559e-01 5.93636811e-01 -9.45293996e-03
-3.18637997e-01 2.18768790e-01 1.06406271e+00 -6.46934390e-01
6.05025232e-01 -1.95863858e-01 -1.16864038e+00 1.09275305e+00
9.56150055e-01 -3.59116644e-01 -4.46984202e-01 -7.76582658e-01
1.88591760e-02 -4.18303251e-01 9.84504640e-01 -8.94337714e-01
1.30531371e-01 -1.68425307e-01 5.27624190e-01 5.99488132e-02
5.67303479e-01 -7.90733874e-01 -4.99078929e-01 6.20765924e-01
-4.75348860e-01 9.54329595e-02 5.41567028e-01 6.12092912e-01
1.20745979e-01 -1.98287666e-02 1.17737055e+00 -3.96268427e-01
-5.86484969e-01 4.07964438e-01 -3.21550608e-01 4.69502985e-01
1.22261465e+00 1.08873539e-01 -9.91283417e-01 -6.59723401e-01
-9.66110110e-01 3.24558973e-01 6.43881977e-01 5.78492343e-01
6.85431182e-01 -9.93784964e-01 -9.74034429e-01 1.44204766e-01
2.26192221e-01 -3.61649305e-01 3.61074321e-02 -2.58970149e-02
-7.81515419e-01 2.42312849e-01 -2.30543122e-01 -4.52944845e-01
-1.43810606e+00 1.29729819e+00 8.00463676e-01 1.24518842e-01
-2.81036735e-01 1.13258183e+00 7.74420738e-01 -9.78431031e-02
3.84519666e-01 2.73347050e-02 4.39170301e-02 -5.73067844e-01
6.79891527e-01 -7.83175826e-02 -4.76637959e-01 -5.85997045e-01
-4.14150447e-01 5.19105315e-01 -3.62658560e-01 1.46034975e-02
6.60467446e-01 2.82194495e-01 -3.23055452e-03 -3.53946835e-01
1.01205850e+00 6.66792989e-02 -1.22280991e+00 -2.98517174e-03
-3.22597802e-01 -4.40621495e-01 -3.11978579e-01 -9.20957923e-01
-8.89397085e-01 1.20272017e+00 5.57764769e-01 5.83924890e-01
9.82708097e-01 3.66089344e-01 6.77030683e-01 5.66916108e-01
7.36161321e-02 -2.20254123e-01 3.55537981e-01 6.58657849e-01
1.25152576e+00 -1.24169588e+00 -2.55041271e-01 -4.36034977e-01
-6.19462550e-01 7.63151526e-01 1.08371699e+00 -5.51216066e-01
4.11272347e-01 4.36334789e-01 4.36901718e-01 -1.36280552e-01
-8.42920482e-01 -2.70922124e-01 1.52270466e-01 1.09632397e+00
-1.89057924e-02 -1.17779486e-01 3.61152977e-01 4.10519391e-02
-5.19581676e-01 -7.17769861e-01 6.09218180e-01 5.91505408e-01
-2.79165655e-01 -9.96543467e-01 -7.57710755e-01 -8.17205105e-03
-8.45920265e-01 -3.57070237e-01 -1.22213936e+00 6.94739044e-01
2.74071433e-02 1.20156014e+00 -4.80192304e-02 -4.27965015e-01
3.17231447e-01 5.72499484e-02 7.00658143e-01 -7.13679850e-01
-8.61886144e-01 -6.46998942e-01 1.22814015e-01 -6.61024690e-01
1.51584744e-01 -1.87116176e-01 -7.16332436e-01 -4.14653510e-01
-1.09581999e-01 -2.28185549e-01 4.28382933e-01 6.99827254e-01
2.91753769e-01 5.48848629e-01 5.26061475e-01 -4.80656326e-01
-9.74900782e-01 -6.12068474e-01 1.05567910e-01 5.45803726e-01
7.45110631e-01 -4.46023606e-02 -6.13432169e-01 2.30442539e-01] | [5.700892448425293, 7.939897060394287] |
f9b3d17a-0194-415b-a338-4745a1a8b887 | are-we-there-yet-evaluating-state-of-the-art | 2007.07455 | null | https://arxiv.org/abs/2007.07455v1 | https://arxiv.org/pdf/2007.07455v1.pdf | Are We There Yet? Evaluating State-of-the-Art Neural Network based Geoparsers Using EUPEG as a Benchmarking Platform | Geoparsing is an important task in geographic information retrieval. A geoparsing system, known as a geoparser, takes some texts as the input and outputs the recognized place mentions and their location coordinates. In June 2019, a geoparsing competition, Toponym Resolution in Scientific Papers, was held as one of the SemEval 2019 tasks. The winning teams developed neural network based geoparsers that achieved outstanding performances (over 90% precision, recall, and F1 score for toponym recognition). This exciting result brings the question "are we there yet?", namely have we achieved high enough performances to possibly consider the problem of geoparsing as solved? One limitation of this competition is that the developed geoparsers were tested on only one dataset which has 45 research articles collected from the particular domain of Bio-medicine. It is known that the same geoparser can have very different performances on different datasets. Thus, this work performs a systematic evaluation of these state-of-the-art geoparsers using our recently developed benchmarking platform EUPEG that has eight annotated datasets, nine baseline geoparsers, and eight performance metrics. The evaluation result suggests that these new geoparsers indeed improve the performances of geoparsing on multiple datasets although some challenges remain. | ['Yingjie Hu', 'Jimin Wang'] | 2020-07-15 | null | null | null | null | ['toponym-resolution'] | ['natural-language-processing'] | [-3.71812433e-01 8.88582468e-02 -3.31934243e-01 -7.26776719e-02
-9.94036973e-01 -5.55221319e-01 9.83781397e-01 7.08504200e-01
-7.44803488e-01 9.86074626e-01 3.15587074e-01 -4.67840955e-02
-4.98877853e-01 -9.65797186e-01 -8.37951005e-01 -6.39821708e-01
-2.02030286e-01 8.52635086e-01 1.68239400e-01 -2.15520248e-01
6.42697811e-01 4.67628270e-01 -1.49888158e+00 3.30548823e-01
7.12611437e-01 6.55527413e-01 2.19111919e-01 5.93076348e-01
-2.15646043e-01 1.98016152e-01 -9.19705272e-01 -1.62615195e-01
6.53600842e-02 1.99212372e-01 -6.42614245e-01 -1.25195861e+00
6.82661116e-01 5.74166000e-01 -3.64019424e-01 1.09720945e+00
6.21395528e-01 -7.46958181e-02 6.90592647e-01 -1.11713052e+00
-1.05035818e+00 1.23660469e+00 -5.21704257e-01 5.45601308e-01
4.12043691e-01 -4.80938584e-01 1.27338052e+00 -9.10863936e-01
1.05314183e+00 9.16966438e-01 1.01511598e+00 1.46243572e-01
-5.45774102e-01 -4.42598015e-01 -1.34427130e-01 4.79053438e-01
-1.98663354e+00 4.57221232e-02 9.15764868e-02 -4.12928462e-01
1.28456986e+00 6.86633825e-01 2.46034220e-01 9.49884892e-01
3.26323330e-01 4.29361522e-01 8.70217443e-01 -3.83080721e-01
3.11407745e-01 -7.31056556e-02 1.81407824e-01 3.67639005e-01
5.13639450e-01 -1.48647502e-01 -7.24845886e-01 -2.13748276e-01
4.13064420e-01 -2.11509049e-01 -2.92803228e-01 1.87341377e-01
-1.81347632e+00 5.12183428e-01 6.87378943e-01 9.28389668e-01
-5.23272038e-01 1.60901118e-02 2.98266232e-01 -5.31785190e-02
5.38777173e-01 1.22803712e+00 -5.91273069e-01 -2.49492183e-01
-1.08219683e+00 5.49397469e-01 1.00914252e+00 8.87207687e-01
3.12561005e-01 -5.42942226e-01 -2.56921202e-01 5.91105580e-01
9.17296577e-03 4.17097926e-01 7.66646862e-01 -2.58464813e-01
7.38071859e-01 7.06298232e-01 1.30170390e-01 -1.40423942e+00
-9.58135605e-01 -4.84465301e-01 -6.99577689e-01 -4.80190217e-01
6.50033176e-01 -1.49774387e-01 -7.86769569e-01 1.40585232e+00
6.12167120e-02 5.47360122e-01 1.20998500e-02 7.88379967e-01
1.60208106e+00 7.97658682e-01 3.81670147e-01 3.02139848e-01
1.63751757e+00 -6.62143409e-01 -7.39244699e-01 3.74512263e-02
7.87429214e-01 -6.17324173e-01 5.89484513e-01 2.44236454e-01
-6.44965112e-01 -1.82797208e-01 -1.24631786e+00 1.71475276e-01
-1.69391358e+00 2.47138783e-01 7.31691062e-01 5.10147154e-01
-1.15869641e+00 9.36850905e-01 -4.60942090e-01 -7.77753353e-01
2.96230358e-03 2.76062101e-01 -7.09200501e-01 2.84000158e-01
-1.84550416e+00 1.28027022e+00 8.81621718e-01 4.12314236e-02
-9.66845825e-02 -1.04391134e+00 -3.77288222e-01 2.28485286e-01
8.59597176e-02 -5.19940078e-01 5.70140302e-01 1.09021194e-01
-8.17685723e-01 1.13806963e+00 3.51058215e-01 -7.01735079e-01
3.81201267e-01 -2.44453087e-01 -9.37148511e-01 -4.60175276e-02
4.43028986e-01 5.21852314e-01 -2.32537985e-01 -4.62566018e-01
-9.27316904e-01 -3.13330054e-01 -3.48072946e-01 2.39761412e-01
-1.04909003e-01 1.58592910e-01 -6.40221775e-01 -5.15584588e-01
9.03316960e-02 -7.80376077e-01 -4.74306904e-02 -6.13984287e-01
-4.80842263e-01 -6.51496232e-01 2.06044361e-01 -7.85459876e-01
1.37160110e+00 -1.87206709e+00 4.00981568e-02 9.67819020e-02
2.94037104e-01 2.38287419e-01 4.64314856e-02 6.50189281e-01
-3.67724448e-01 3.64538103e-01 2.00564936e-01 3.12997490e-01
-5.21727452e-05 -2.33535975e-01 -2.72012889e-01 5.75079381e-01
2.22965493e-03 1.07840860e+00 -1.09269941e+00 -3.47491264e-01
1.60618320e-01 5.13631463e-01 -5.40277399e-02 -1.70673832e-01
-1.16366893e-01 6.87154382e-02 -5.50976872e-01 7.78715909e-01
6.64569795e-01 -3.25501561e-01 1.43936813e-01 -1.00022718e-01
-4.54900414e-01 2.39155874e-01 -1.12556434e+00 1.69306576e+00
-1.45343393e-01 8.01783979e-01 -7.08593905e-01 -9.92784202e-01
8.62764299e-01 4.44531977e-01 7.95163512e-01 -6.81251466e-01
-4.63440381e-02 5.56324065e-01 -3.49256068e-01 -4.82178122e-01
1.03474796e+00 6.43358827e-01 -5.79955220e-01 -4.39884067e-02
2.11454302e-01 2.24776849e-01 4.72502470e-01 6.22368790e-02
1.21002853e+00 1.90017149e-01 5.95580459e-01 -6.07593238e-01
3.96941841e-01 4.56230313e-01 3.60119760e-01 1.17223573e+00
-1.08373024e-01 1.03394866e+00 4.10690069e-01 -7.45318174e-01
-9.50131416e-01 -6.96351111e-01 -3.44840765e-01 1.03952658e+00
1.30249122e-02 -5.29978752e-01 -7.90931940e-01 -4.79221523e-01
1.15808301e-01 6.75217271e-01 -9.73839402e-01 2.14805454e-01
-6.18005633e-01 -1.36261785e+00 1.13634491e+00 3.43794584e-01
4.34840381e-01 -1.03927362e+00 -4.77878571e-01 2.31660679e-01
-2.17596114e-01 -1.00048256e+00 -9.06088352e-02 2.85763353e-01
-3.10729653e-01 -1.33494115e+00 -1.08145344e+00 -6.16689742e-01
2.44578719e-01 9.20422822e-02 1.33118081e+00 -2.91365772e-01
-2.22838059e-01 -3.00089657e-01 -3.64370704e-01 -7.68407702e-01
1.00985885e-01 8.35881710e-01 2.52210088e-02 -5.09567261e-01
9.30226684e-01 -3.55998129e-01 -5.52902341e-01 1.49853781e-01
-4.82551247e-01 -1.50063336e-01 5.49970508e-01 5.57604909e-01
4.94994313e-01 -1.64951324e-01 8.98432612e-01 -1.07584512e+00
4.36488718e-01 -9.02307212e-01 -8.22161973e-01 5.54703772e-01
-6.30837858e-01 6.89016432e-02 4.32214767e-01 6.80423006e-02
-5.92933297e-01 -1.04406670e-01 -3.49481046e-01 9.03647020e-02
-1.44546136e-01 1.18258691e+00 -6.62553683e-02 7.29624042e-03
1.10879695e+00 5.52244261e-02 -6.95657730e-01 -8.45274091e-01
5.97355545e-01 8.49693298e-01 8.48288000e-01 -4.32027221e-01
4.31009322e-01 4.22022752e-02 4.11822973e-03 -5.91966569e-01
-7.38931060e-01 -9.35167015e-01 -5.32893538e-01 1.56134203e-01
8.78526032e-01 -1.21558487e+00 -8.16869617e-01 2.96805412e-01
-1.20672989e+00 2.58994728e-01 -3.15264463e-02 5.35732448e-01
-6.32345304e-02 -1.41999453e-01 -1.57706842e-01 -5.13742678e-02
-4.18398231e-01 -7.04764068e-01 1.17713380e+00 5.89959681e-01
-2.76992053e-01 -1.05708015e+00 2.23471135e-01 -4.11734171e-02
7.15886474e-01 4.57852155e-01 5.48576832e-01 -1.49146843e+00
-3.53404135e-01 -5.86093962e-01 -3.84084105e-01 -7.80554831e-01
8.42065141e-02 -1.28026372e-02 -9.49975610e-01 2.05447018e-01
-6.58169448e-01 3.22535336e-01 9.43570614e-01 3.28574061e-01
1.21043301e+00 -3.23095351e-01 -8.55410576e-01 6.60239458e-01
1.46546888e+00 1.80285946e-01 5.39915562e-01 1.12679005e+00
8.56963992e-01 4.89462674e-01 6.69197500e-01 4.14140671e-02
3.00852805e-01 8.91870737e-01 2.79844910e-01 1.14186727e-01
7.77849108e-02 -1.57483384e-01 -3.67806882e-01 5.79433680e-01
-1.63788408e-01 -7.14801848e-01 -1.41937959e+00 8.32574546e-01
-2.05529690e+00 -8.53127956e-01 -3.25573832e-01 1.98708701e+00
8.72936487e-01 -2.56905675e-01 -8.34013820e-02 -1.21581137e-01
9.43770468e-01 2.54748344e-01 -1.18216000e-01 -2.03119189e-01
-5.46181738e-01 1.28991365e-01 1.14337778e+00 2.78289486e-02
-1.66165626e+00 1.06441641e+00 6.48186636e+00 9.21699822e-01
-1.37683535e+00 -4.92153317e-02 4.12053168e-01 1.51643246e-01
1.44665197e-01 -2.50505418e-01 -1.14761531e+00 6.75217509e-01
1.42213261e+00 -6.07316554e-01 -2.28948109e-02 9.61234033e-01
-3.06871325e-01 1.10667229e-01 -8.88053656e-01 1.01733172e+00
1.44618481e-01 -2.02973318e+00 -1.20586880e-01 -2.34037906e-01
5.85147560e-01 6.02979541e-01 -1.40726209e-01 2.25894526e-01
4.99170214e-01 -1.35068405e+00 4.96532828e-01 6.07237637e-01
7.44739950e-01 -4.68594730e-01 1.03029442e+00 -7.71844992e-04
-9.78762150e-01 3.53390068e-01 -3.81670713e-01 3.29581201e-01
-8.91386718e-02 6.26159191e-01 -1.07606125e+00 1.15027881e+00
1.02120519e+00 9.28607285e-01 -9.10899997e-01 1.65378189e+00
-2.09801272e-01 1.52711466e-01 -5.54514647e-01 -3.02739382e-01
3.69491249e-01 4.15657341e-01 7.14118063e-01 1.51970637e+00
6.62640452e-01 -2.91175216e-01 -3.45306337e-01 5.02174675e-01
-1.38758793e-01 2.80511647e-01 -6.95175529e-01 -1.39353737e-01
8.67175877e-01 1.40473044e+00 -8.48169148e-01 -2.62832642e-01
-1.40644640e-01 3.42423439e-01 9.53259841e-02 7.72277638e-02
-7.86435127e-01 -8.22483182e-01 3.17331135e-01 1.10758394e-02
-3.28963739e-03 2.34635994e-01 -5.13933957e-01 -9.83565688e-01
-3.78047645e-01 -4.93015468e-01 1.04918265e+00 -9.60850298e-01
-1.13883567e+00 7.38528490e-01 1.70620337e-01 -1.16947687e+00
-2.25340173e-01 -7.97034562e-01 -2.61277616e-01 9.70951498e-01
-1.30269647e+00 -1.19937909e+00 -3.70431513e-01 -5.87857626e-02
1.26844542e-02 -4.67967629e-01 1.04393280e+00 7.47972190e-01
-6.26978457e-01 6.44639969e-01 6.20690286e-01 1.93182394e-01
1.18555200e+00 -1.66075087e+00 8.50316405e-01 6.30513668e-01
4.10257876e-01 8.90162706e-01 9.32510018e-01 -6.22813106e-01
-1.21900213e+00 -1.29928780e+00 1.65153551e+00 -8.51282358e-01
9.83495235e-01 -1.91275120e-01 -8.80044878e-01 6.38059020e-01
1.85246274e-01 -5.40965945e-02 4.25767124e-01 1.76648214e-01
-2.11697564e-01 1.76416069e-01 -9.47745085e-01 4.63061571e-01
8.00346613e-01 -1.80091888e-01 -8.39765787e-01 7.57629693e-01
4.56324697e-01 -9.29182112e-01 -1.08379805e+00 2.56168306e-01
5.42179942e-01 -2.02095523e-01 1.34490871e+00 -6.85165763e-01
5.36740005e-01 -4.81426060e-01 -2.99003236e-02 -1.49783826e+00
-2.24819541e-01 -2.18945920e-01 1.50559425e-01 1.20399725e+00
7.40228772e-01 -7.34339118e-01 7.03846276e-01 2.67481685e-01
-2.62424350e-01 -5.22223175e-01 -1.18663764e+00 -8.09097052e-01
4.28359747e-01 1.59714341e-01 9.53055680e-01 1.42778826e+00
3.65592897e-01 4.95804399e-02 -2.09704384e-01 1.51434377e-01
9.13108140e-02 -3.31340320e-02 3.88371468e-01 -1.61102271e+00
3.54953140e-01 -6.04153693e-01 -5.95809698e-01 -1.74601227e-01
-1.62421949e-02 -1.10064495e+00 2.95791663e-02 -1.72204065e+00
-4.30773050e-02 -5.59279144e-01 -7.28951275e-01 8.42553556e-01
-4.18159187e-01 3.91853988e-01 -2.33231276e-01 4.05384392e-01
-7.65385330e-01 -1.90478921e-01 2.89667308e-01 -1.95371062e-01
-1.29567906e-01 -5.16366422e-01 -9.32953238e-01 5.27292609e-01
6.10997915e-01 -6.33909464e-01 2.89227456e-01 -3.91227126e-01
7.90571094e-01 -2.70457983e-01 2.28704512e-01 -1.10603738e+00
6.43688500e-01 -2.06883907e-01 3.37083399e-01 -8.71325135e-01
-7.91803747e-02 -4.42353487e-01 5.03859460e-01 1.80120468e-01
-3.30944121e-01 4.39291924e-01 4.09289777e-01 3.48137081e-01
-2.97541827e-01 -1.48993237e-02 8.53890404e-02 -2.39702880e-01
-1.25381100e+00 2.09580529e-02 -2.68091559e-01 1.33144245e-01
8.68860364e-01 -3.13664898e-02 -1.05957043e+00 4.04742688e-01
-5.36444128e-01 1.60805121e-01 2.22670078e-01 8.89452994e-01
-4.86194119e-02 -1.26450264e+00 -9.41165268e-01 -4.21369880e-01
5.89801311e-01 -2.74645418e-01 1.76108375e-01 6.98413610e-01
-9.86870587e-01 1.23463345e+00 -4.58906859e-01 -4.98232037e-01
-1.10978997e+00 4.69148368e-01 3.27240616e-01 -6.39094532e-01
-7.67338455e-01 9.46690917e-01 -6.00514188e-02 -5.91280580e-01
1.07852861e-01 5.52833416e-02 -9.89694834e-01 4.46387231e-01
6.88555896e-01 2.03627124e-01 6.85863674e-01 -6.12289906e-01
-8.92131269e-01 4.98555541e-01 -4.99182120e-02 5.22626974e-02
1.68485188e+00 3.20957750e-01 -1.32023916e-01 3.63225579e-01
9.57821608e-01 -1.98807921e-02 -1.75008595e-01 -1.43654600e-01
6.27060175e-01 -1.84582546e-01 -1.41488165e-01 -1.12803650e+00
-9.10959363e-01 4.27754790e-01 5.35927892e-01 2.58559346e-01
6.46085262e-01 -4.67664674e-02 3.42014730e-01 5.69962084e-01
2.93161750e-01 -1.20119691e+00 -6.43202126e-01 7.43877709e-01
8.79407883e-01 -9.83081639e-01 1.72024518e-01 -3.48154664e-01
-2.83207923e-01 9.28990960e-01 2.69269317e-01 -4.48343791e-02
5.71411550e-01 -4.85067293e-02 -6.18889518e-02 -5.96185684e-01
-4.38207537e-01 2.47543734e-02 6.41709507e-01 3.83952230e-01
6.66402936e-01 1.64068699e-01 -8.70052040e-01 7.53946304e-01
-5.47333419e-01 2.35210247e-02 3.63329530e-01 7.23580897e-01
-1.70887873e-01 -9.14922118e-01 -2.97657967e-01 4.62035775e-01
-8.67843986e-01 -4.30191338e-01 -6.24535382e-01 1.04465961e+00
7.83566833e-02 4.54687059e-01 -2.20225826e-02 -2.45851487e-01
3.68886471e-01 8.93425569e-02 -4.90450636e-02 -5.45010567e-01
-1.00691819e+00 -6.42541230e-01 2.26428032e-01 -4.93990302e-01
-2.94828087e-01 -5.45624793e-01 -1.00913453e+00 -3.93000126e-01
-1.83543965e-01 8.07688951e-01 1.01316178e+00 8.59039068e-01
6.26771629e-01 8.20449769e-01 4.06745309e-03 -4.77753371e-01
-8.22722819e-03 -1.14148855e+00 -4.83480781e-01 1.76994473e-01
-1.61869645e-01 -7.15537965e-01 -8.07806253e-02 -2.92162150e-01] | [9.154876708984375, 9.034463882446289] |
6a664e89-d084-4f62-87e0-ce98d38acfbc | humble-teachers-teach-better-students-for | 2106.10456 | null | https://arxiv.org/abs/2106.10456v1 | https://arxiv.org/pdf/2106.10456v1.pdf | Humble Teachers Teach Better Students for Semi-Supervised Object Detection | We propose a semi-supervised approach for contemporary object detectors following the teacher-student dual model framework. Our method is featured with 1) the exponential moving averaging strategy to update the teacher from the student online, 2) using plenty of region proposals and soft pseudo-labels as the student's training targets, and 3) a light-weighted detection-specific data ensemble for the teacher to generate more reliable pseudo-labels. Compared to the recent state-of-the-art -- STAC, which uses hard labels on sparsely selected hard pseudo samples, the teacher in our model exposes richer information to the student with soft-labels on many proposals. Our model achieves COCO-style AP of 53.04% on VOC07 val set, 8.4% better than STAC, when using VOC12 as unlabeled data. On MS-COCO, it outperforms prior work when only a small percentage of data is taken as labeled. It also reaches 53.8% AP on MS-COCO test-dev with 3.1% gain over the fully supervised ResNet-152 Cascaded R-CNN, by tapping into unlabeled data of a similar size to the labeled data. | ['Yuting Zhang', 'Yijun Luo', 'Weifeng Chen', 'Yihe Tang'] | 2021-06-19 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Tang_Humble_Teachers_Teach_Better_Students_for_Semi-Supervised_Object_Detection_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Tang_Humble_Teachers_Teach_Better_Students_for_Semi-Supervised_Object_Detection_CVPR_2021_paper.pdf | cvpr-2021-1 | ['semi-supervised-object-detection'] | ['computer-vision'] | [ 3.38639654e-02 4.48456436e-01 -2.90939033e-01 -3.82560790e-01
-1.37163341e+00 -5.61230242e-01 6.85553670e-01 -2.05212999e-02
-8.31809700e-01 6.52072966e-01 -1.54582009e-01 -1.42388493e-01
4.37554091e-01 -4.25595343e-01 -9.57747281e-01 -7.82193959e-01
2.38703415e-01 8.41814101e-01 9.68287647e-01 4.81416658e-02
-2.51359403e-01 4.45744574e-01 -1.41127276e+00 3.49012107e-01
7.33031511e-01 1.08224487e+00 2.32303903e-01 6.75959408e-01
9.80175585e-02 8.78916264e-01 -5.82772434e-01 -2.88883686e-01
4.08806711e-01 -3.45774025e-01 -5.87056458e-01 1.65197492e-01
1.02791953e+00 -2.25657761e-01 -4.31000948e-01 9.75376904e-01
5.97869873e-01 1.76688522e-01 8.57867360e-01 -6.66929066e-01
-8.16402256e-01 8.13081920e-01 -9.28236663e-01 2.85999566e-01
-3.75425965e-01 3.21731925e-01 9.70148146e-01 -1.61652505e+00
6.14129066e-01 1.13719320e+00 5.16432762e-01 8.22503328e-01
-1.28738308e+00 -9.24573183e-01 4.46527898e-01 2.67977834e-01
-1.53315926e+00 -2.79660612e-01 3.56896311e-01 -2.01451749e-01
7.47512758e-01 -7.84651935e-02 5.39289534e-01 1.26612592e+00
-9.23477039e-02 1.19325423e+00 1.21232533e+00 -4.33960021e-01
3.94240022e-01 6.57648087e-01 4.36538398e-01 6.59638047e-01
1.04952358e-01 2.05895334e-01 -3.12475413e-01 4.56605069e-02
3.61200392e-01 -7.56223351e-02 2.11019098e-04 -2.54494011e-01
-7.78307915e-01 8.22090745e-01 8.60995412e-01 -9.69554931e-02
-2.29264826e-01 1.18855394e-01 1.00238338e-01 -1.05259111e-02
8.58328283e-01 1.22797072e-01 -4.45409149e-01 4.62778747e-01
-1.31736147e+00 8.47190917e-02 5.02246737e-01 1.20942998e+00
5.70745945e-01 4.13439304e-01 -5.03868997e-01 9.26339805e-01
5.02689123e-01 5.16812503e-01 4.56250221e-01 -5.81252038e-01
2.81815767e-01 5.35046816e-01 5.71091026e-02 1.34033948e-01
-1.02586888e-01 -1.33530438e+00 -4.16330129e-01 5.03904045e-01
4.88418072e-01 -1.18485630e-01 -1.80016351e+00 1.58876574e+00
4.90997493e-01 4.53222513e-01 -2.47471184e-02 8.80196810e-01
1.05232799e+00 6.69549286e-01 7.96426982e-02 1.11817181e-01
1.30747700e+00 -1.61596310e+00 -2.55154490e-01 -4.29006696e-01
3.61259997e-01 -6.07042253e-01 6.39302015e-01 5.80140650e-01
-1.11177170e+00 -8.28639925e-01 -1.04489362e+00 -1.17291018e-01
-5.04762053e-01 5.96244514e-01 8.54082853e-02 5.43953598e-01
-1.38700736e+00 2.37818077e-01 -6.57670617e-01 -1.34810328e-01
9.34998155e-01 1.89974353e-01 -6.03357628e-02 -3.93586159e-01
-7.04012215e-01 1.04306448e+00 5.07876873e-01 1.28935073e-02
-1.67488468e+00 -8.95860493e-01 -5.83236039e-01 5.08577228e-02
5.89482367e-01 -6.44041896e-02 1.45077431e+00 -9.02133107e-01
-1.36183727e+00 9.48964775e-01 7.31540769e-02 -7.36343980e-01
8.69964421e-01 -3.14611077e-01 -2.22994909e-01 1.09187886e-01
1.54287845e-01 1.46814978e+00 9.06125605e-01 -1.29721856e+00
-1.06085002e+00 -2.69454151e-01 -3.39852571e-01 2.08168715e-01
-3.78946997e-02 -1.43456757e-01 -9.04218137e-01 -6.08892918e-01
8.47133845e-02 -1.02998531e+00 -5.32744229e-01 8.63274038e-02
-5.06282508e-01 -6.36281908e-01 1.18567550e+00 -2.69278884e-01
6.81904137e-01 -1.97299147e+00 4.13242215e-03 2.43250426e-04
5.19493461e-01 6.50012255e-01 -3.20127785e-01 -2.69658804e-01
-2.28543192e-01 -2.70456344e-01 -1.13194034e-01 -6.34479284e-01
-2.86153287e-01 -1.05403764e-02 -4.73071843e-01 7.41054177e-01
5.60953021e-01 1.00642383e+00 -9.02510941e-01 -5.78128636e-01
1.02635548e-01 3.65767658e-01 -3.47285599e-01 2.46919766e-01
-2.49203324e-01 9.85501930e-02 -3.33697736e-01 8.11681092e-01
7.11617172e-01 -4.11216617e-01 -2.30749145e-01 1.43682241e-01
-4.65649590e-02 7.32283667e-02 -1.06928110e+00 1.38949025e+00
-1.15644328e-01 7.79756486e-01 3.61929387e-01 -8.35646689e-01
9.27401125e-01 1.47741079e-01 5.14671840e-02 -4.45679188e-01
1.09726541e-01 1.49999171e-01 -5.87199777e-02 2.03882754e-01
4.17878628e-01 2.67291188e-01 2.59585708e-01 3.56646925e-01
6.75975680e-01 -1.11066885e-01 1.25877991e-01 5.21484554e-01
1.15637231e+00 3.48810762e-01 -4.63966988e-02 -4.34223145e-01
2.98236519e-01 4.21743691e-02 6.90680027e-01 1.02202177e+00
-3.40766132e-01 7.87554681e-01 1.85126979e-02 -3.84211034e-01
-8.86880398e-01 -1.25023866e+00 -4.06899303e-01 1.40609741e+00
-5.35003766e-02 5.22261970e-02 -6.98903263e-01 -1.18095315e+00
-4.91366498e-02 9.89072740e-01 -7.92700410e-01 -2.20543705e-02
-4.76790279e-01 -7.53882170e-01 4.77367818e-01 7.82338321e-01
2.91172445e-01 -1.03750813e+00 -3.62680674e-01 3.60822737e-01
4.63224113e-01 -1.18602931e+00 -5.98383367e-01 8.00007701e-01
-6.78828299e-01 -8.11975479e-01 -1.03946006e+00 -8.22056115e-01
8.48578334e-01 3.85614544e-01 1.08894897e+00 -3.02947432e-01
-5.32293379e-01 4.02513891e-01 -2.93050438e-01 -6.64164722e-01
-3.55883896e-01 9.18390080e-02 -1.58263724e-02 4.54624370e-02
3.73066694e-01 -1.48270741e-01 -6.21370912e-01 3.87887985e-01
-5.99729002e-01 1.80532545e-01 8.44223022e-01 9.65747356e-01
8.50737989e-01 -5.14383554e-01 7.09029019e-01 -1.10136199e+00
-1.78287685e-01 -5.92518151e-01 -7.98348427e-01 3.20085675e-01
-9.19359028e-01 -6.93144575e-02 6.06729031e-01 -8.72204661e-01
-1.32103717e+00 5.26362002e-01 -3.95919904e-02 -8.66939425e-01
-1.93587184e-01 -2.57794499e-01 3.60185504e-01 -2.01923966e-01
1.14014471e+00 2.74587393e-01 -3.61196756e-01 -4.91469026e-01
6.36175990e-01 5.24382532e-01 7.04387188e-01 -5.03107786e-01
9.86673653e-01 5.39020956e-01 -3.25811476e-01 -5.95767438e-01
-1.34970057e+00 -7.85325587e-01 -6.11454546e-01 -3.58994037e-01
8.67513239e-01 -1.45322311e+00 -1.88041300e-01 3.94935787e-01
-9.15981829e-01 -6.21324718e-01 -6.26547575e-01 4.65341151e-01
-1.27878919e-01 -9.23394933e-02 -7.60459661e-01 -8.72466028e-01
-3.30734193e-01 -1.31460202e+00 1.01192141e+00 5.03813326e-01
2.46927664e-01 -6.01807117e-01 -1.29791811e-01 4.67470258e-01
3.83334368e-01 -1.98300019e-01 1.59528684e-02 -1.09594333e+00
-6.76085353e-01 -4.60759521e-01 -4.99619156e-01 6.40389740e-01
-4.99695390e-01 -2.33256835e-02 -1.51761091e+00 -4.78387117e-01
-3.75433028e-01 -1.06066918e+00 1.46013582e+00 3.56986910e-01
1.25333178e+00 1.05860844e-01 -4.72183168e-01 4.46507990e-01
1.24457324e+00 -1.35254830e-01 2.45254830e-01 -3.33182544e-01
7.70407796e-01 4.95934576e-01 5.64798832e-01 4.75864559e-02
3.22381794e-01 5.87364376e-01 7.10785985e-01 -4.83380646e-01
-5.24406016e-01 -3.07426900e-01 4.73272085e-01 5.48596263e-01
1.08276360e-01 -2.48539999e-01 -7.86024392e-01 6.29978955e-01
-1.71518314e+00 -6.93040788e-01 -1.62597761e-01 1.85569274e+00
8.87960315e-01 4.12575126e-01 2.13637009e-01 -4.18629229e-01
7.27801919e-01 6.59055449e-03 -8.64539444e-01 6.62545860e-02
-7.22944587e-02 5.20228028e-01 9.76377428e-01 3.50250214e-01
-1.32828796e+00 1.15011871e+00 6.16890955e+00 1.30641377e+00
-1.01944268e+00 5.50674081e-01 1.06950295e+00 -4.29840833e-01
7.26138661e-03 -2.54845381e-01 -1.54181421e+00 2.63164520e-01
1.13561237e+00 4.44251150e-01 2.40643814e-04 1.35432649e+00
-3.08493841e-02 -2.81467199e-01 -1.07480276e+00 7.16363668e-01
2.25062817e-01 -1.42117953e+00 -2.81739444e-01 1.35894924e-01
1.21363795e+00 8.47735882e-01 3.16628069e-01 8.37426782e-01
9.50717449e-01 -9.88670766e-01 1.22258592e+00 1.56559274e-01
9.84828174e-01 -4.22593743e-01 5.85185230e-01 6.53151631e-01
-1.16465628e+00 -1.14645824e-01 -4.74495947e-01 4.43873793e-01
-3.32053632e-01 5.54400563e-01 -1.04768205e+00 3.09768379e-01
8.95785332e-01 4.92178112e-01 -1.02465546e+00 1.11140633e+00
-5.08475721e-01 1.36338794e+00 -6.30692065e-01 -1.01537190e-01
6.99653745e-01 -1.44637460e-02 2.76251018e-01 1.58404064e+00
3.34976129e-02 4.23830673e-02 4.85673368e-01 1.15613329e+00
-3.93743247e-01 -1.23776257e-01 -1.76942647e-01 4.81318235e-01
4.93405402e-01 1.66112339e+00 -9.87343669e-01 -7.18023181e-01
-3.69732022e-01 7.63534069e-01 7.23770142e-01 4.87371534e-01
-9.91183043e-01 -6.55845553e-02 -1.90775365e-01 -4.79240157e-02
5.36425948e-01 2.53525048e-01 -5.01350015e-02 -7.67517984e-01
-2.16310591e-01 -5.60160816e-01 4.06930327e-01 -6.15968466e-01
-1.19120538e+00 6.31949902e-01 7.41615938e-03 -1.05948436e+00
-8.78113881e-02 -7.11356461e-01 -9.81090486e-01 9.77827132e-01
-1.47466004e+00 -1.25359917e+00 -1.04582064e-01 1.95768863e-01
9.69982803e-01 -2.62580425e-01 5.41570902e-01 1.38235971e-01
-7.60921717e-01 7.15719223e-01 1.46767527e-01 1.40928626e-01
8.00379276e-01 -1.60461235e+00 4.05454755e-01 9.56792712e-01
5.08395016e-01 6.26399592e-02 5.06067038e-01 -3.76125813e-01
-8.92463505e-01 -1.54204810e+00 3.86540532e-01 -7.11899161e-01
4.30494398e-01 -5.87796688e-01 -1.06080627e+00 6.48861408e-01
3.91893536e-01 8.75279963e-01 2.85660714e-01 -8.80722478e-02
-4.38130438e-01 -1.60806060e-01 -1.21152234e+00 3.43440354e-01
6.57424927e-01 -2.58513212e-01 -6.15552187e-01 5.51214814e-01
8.02548885e-01 -6.27727211e-01 -3.93858671e-01 3.24717015e-01
2.70268738e-01 -6.36833370e-01 8.22616518e-01 -3.38491827e-01
-6.23399904e-03 -4.88408089e-01 1.48112029e-01 -1.26906943e+00
-4.40396070e-01 -2.15660185e-01 -2.42684677e-01 1.04740989e+00
8.17154646e-01 -2.45018914e-01 1.13252866e+00 1.04944356e-01
-4.53259856e-01 -1.19327188e+00 -1.05178797e+00 -6.76718712e-01
9.37636569e-02 -5.35824537e-01 -2.10454762e-01 6.61674380e-01
-7.66475141e-01 5.15498340e-01 -7.66629353e-02 1.07484676e-01
9.66007352e-01 -3.04343462e-01 5.27191520e-01 -1.12143362e+00
-3.37463439e-01 -3.41689199e-01 9.04881582e-02 -1.23245251e+00
4.31279093e-02 -1.00130737e+00 3.87148172e-01 -1.18602502e+00
3.82084161e-01 -5.65754414e-01 -4.85916883e-01 6.92332625e-01
-1.37288347e-01 5.96389055e-01 7.16544837e-02 1.11620948e-02
-1.02339971e+00 6.79724097e-01 1.09607995e+00 -3.48039776e-01
-2.18314156e-01 1.96227595e-01 -4.68072444e-01 7.50099242e-01
4.77638930e-01 -7.83594608e-01 -3.49777222e-01 -3.53120901e-02
-3.59215975e-01 -2.74242580e-01 4.47558641e-01 -1.03835654e+00
4.57673848e-01 5.87556046e-03 7.68289626e-01 -1.20482111e+00
4.92895663e-01 -6.37460053e-01 -4.10844058e-01 5.24498880e-01
-4.77593750e-01 -3.62473428e-01 2.65741169e-01 8.79290819e-01
1.42874986e-01 -3.77441406e-01 1.28622437e+00 -1.08600281e-01
-7.44351566e-01 4.17771310e-01 -3.01974595e-01 2.24651888e-01
1.22252035e+00 -7.68574029e-02 -2.85414547e-01 -7.98819289e-02
-8.02277744e-01 6.70313954e-01 5.48804514e-02 3.66469413e-01
6.86913729e-01 -1.09494603e+00 -9.78700638e-01 8.48411489e-03
1.20751552e-01 5.44542491e-01 1.73006788e-01 5.50767660e-01
-1.88105524e-01 3.17475855e-01 1.89600572e-01 -8.84409189e-01
-1.14597380e+00 3.60670775e-01 2.66320735e-01 -3.14891309e-01
-6.20791435e-01 1.37561846e+00 3.94429594e-01 -4.20050710e-01
6.97422087e-01 -9.45977047e-02 -1.47894308e-01 1.50416583e-01
6.01708770e-01 2.35401556e-01 -4.13164422e-02 -4.79644597e-01
-2.94698328e-01 6.37773350e-02 -6.91251755e-01 -2.45158121e-01
1.34163928e+00 2.74692267e-01 4.18873370e-01 3.32835317e-01
8.82057130e-01 -2.92264502e-02 -1.77801573e+00 -6.86198831e-01
-8.46439600e-02 1.65206447e-01 3.81070524e-01 -9.97423112e-01
-1.08417678e+00 8.47776771e-01 9.39645767e-01 -2.71658301e-01
5.85436344e-01 4.57645178e-01 2.02992573e-01 4.24148798e-01
7.19842464e-02 -1.30770230e+00 3.61514360e-01 4.83969718e-01
7.55861878e-01 -1.45147574e+00 8.10504705e-02 -1.07890248e-01
-6.58513367e-01 8.78232777e-01 1.20646989e+00 -3.30200642e-01
5.88074684e-01 2.55849451e-01 1.02917343e-01 -3.26624475e-02
-1.13529336e+00 -3.10226142e-01 6.71791911e-01 3.18561822e-01
1.47507459e-01 2.50079811e-01 3.18368673e-01 3.66935551e-01
2.83188790e-01 -3.81213993e-01 4.62257236e-01 6.54648304e-01
-8.41368020e-01 -6.87764108e-01 -5.01461506e-01 7.25680351e-01
-4.66224194e-01 -8.30221251e-02 -3.01822841e-01 7.74034798e-01
2.70089447e-01 6.88108861e-01 9.56983492e-02 -1.62583753e-01
5.74993826e-02 1.27175763e-01 2.30618119e-01 -1.17468143e+00
-7.00826108e-01 3.30092102e-01 -2.40648687e-01 -4.27061111e-01
-1.31969109e-01 -5.30298114e-01 -1.19121122e+00 2.66878754e-01
-7.44789422e-01 3.94906998e-02 5.43181419e-01 8.40467989e-01
8.79086256e-02 6.93298757e-01 5.45855224e-01 -1.11880445e+00
-9.37481582e-01 -1.40128577e+00 -3.57295454e-01 -6.25983998e-02
2.69171625e-01 -6.01320207e-01 -3.38205248e-01 -3.09846699e-01] | [9.21873664855957, 1.3033549785614014] |
715bdded-c458-4bd8-a30c-9eee355cb846 | training-differentially-private-graph-neural | 2301.00738 | null | https://arxiv.org/abs/2301.00738v1 | https://arxiv.org/pdf/2301.00738v1.pdf | Training Differentially Private Graph Neural Networks with Random Walk Sampling | Deep learning models are known to put the privacy of their training data at risk, which poses challenges for their safe and ethical release to the public. Differentially private stochastic gradient descent is the de facto standard for training neural networks without leaking sensitive information about the training data. However, applying it to models for graph-structured data poses a novel challenge: unlike with i.i.d. data, sensitive information about a node in a graph cannot only leak through its gradients, but also through the gradients of all nodes within a larger neighborhood. In practice, this limits privacy-preserving deep learning on graphs to very shallow graph neural networks. We propose to solve this issue by training graph neural networks on disjoint subgraphs of a given training graph. We develop three random-walk-based methods for generating such disjoint subgraphs and perform a careful analysis of the data-generating distributions to provide strong privacy guarantees. Through extensive experiments, we show that our method greatly outperforms the state-of-the-art baseline on three large graphs, and matches or outperforms it on four smaller ones. | ['Stephan Günnemann', 'Daniel Zügner', 'Lukas Gosch', 'Jan Schuchardt', 'Morgane Ayle'] | 2023-01-02 | null | null | null | null | ['privacy-preserving-deep-learning', 'privacy-preserving-deep-learning'] | ['methodology', 'natural-language-processing'] | [ 1.17005013e-01 5.53116202e-01 -2.51658559e-01 -5.23987830e-01
-8.48281443e-01 -9.38588440e-01 2.22129256e-01 2.22920686e-01
-4.17283952e-01 8.28572035e-01 1.41826615e-01 -6.29574299e-01
2.02051565e-01 -1.19294930e+00 -1.08806241e+00 -8.54033351e-01
-4.65867758e-01 1.97677672e-01 -1.62771538e-01 1.72504615e-02
-2.36954167e-01 6.11912072e-01 -6.51203692e-01 8.64796638e-02
3.36982727e-01 7.34545708e-01 -5.67746460e-01 4.66009200e-01
2.48417869e-01 7.05731511e-01 -4.71907109e-01 -9.65054810e-01
7.40325511e-01 -2.29957238e-01 -7.95655310e-01 -1.69104964e-01
5.97964048e-01 -6.65675342e-01 -9.38921988e-01 1.35762608e+00
5.09622812e-01 -1.03004344e-01 2.61183143e-01 -1.45481884e+00
-9.91359830e-01 8.74142408e-01 -6.29881382e-01 -1.15155876e-01
3.54327410e-02 1.10236667e-01 1.20877206e+00 -3.61015648e-01
6.99555159e-01 8.91338527e-01 8.15167427e-01 7.51729608e-01
-1.31596804e+00 -8.82181168e-01 3.03276271e-01 -3.47904474e-01
-1.35354829e+00 -3.51704866e-01 8.46467495e-01 -2.68616408e-01
5.77782094e-01 4.24152493e-01 4.67457503e-01 1.31139624e+00
1.15426563e-01 8.42908621e-01 7.02561498e-01 7.21938238e-02
4.29092586e-01 1.68676496e-01 1.68187305e-01 9.21366692e-01
7.20372975e-01 8.80568549e-02 -2.68468380e-01 -7.61357844e-01
4.34188545e-01 1.06527992e-01 -4.96574581e-01 -8.98507476e-01
-6.11389995e-01 1.08299077e+00 6.68681324e-01 1.02425283e-02
3.15215327e-02 5.49707949e-01 5.23243129e-01 4.87883091e-01
7.21928060e-01 1.87028289e-01 -4.20630455e-01 5.75187266e-01
-5.05737782e-01 3.94016564e-01 1.15002167e+00 1.03245413e+00
9.12663758e-01 -1.47763491e-01 -8.98565426e-02 1.97859377e-01
2.13340595e-01 2.34793112e-01 -5.20758610e-03 -4.77814168e-01
8.66424680e-01 4.08807933e-01 -3.07492539e-02 -1.44310117e+00
-1.88983932e-01 -5.76946914e-01 -1.01362944e+00 -5.43417595e-03
6.09500110e-01 -6.95351422e-01 -8.06194842e-01 2.19780755e+00
3.75639021e-01 -2.64893062e-02 8.81248992e-03 7.95366466e-01
5.39458215e-01 5.51434636e-01 -1.24099508e-01 9.66728628e-02
7.76748300e-01 -5.73295414e-01 -3.61671329e-01 -4.40459579e-01
9.97685790e-01 2.19155118e-01 7.96142995e-01 -8.40101950e-03
-8.45144451e-01 2.63330400e-01 -1.03996360e+00 -3.45614165e-01
-4.86806333e-01 -3.53228748e-01 1.02150953e+00 1.12006724e+00
-1.32317936e+00 8.20251584e-01 -8.25693727e-01 -1.65041089e-01
1.07013941e+00 5.10251820e-01 -7.84802020e-01 -1.77746281e-01
-1.39517713e+00 2.26493374e-01 1.67012006e-01 1.21187404e-01
-1.14230490e+00 -7.96316445e-01 -1.21042538e+00 2.77129441e-01
3.82762820e-01 -6.30131185e-01 8.40567350e-01 -6.44146025e-01
-1.08807755e+00 1.24590635e+00 1.32037014e-01 -9.86101329e-01
7.94915080e-01 4.11133170e-02 -1.81120709e-02 -2.06287410e-02
-5.06597832e-02 1.15246244e-01 7.18661547e-01 -1.11303735e+00
-2.96130449e-01 -8.27688515e-01 3.22368354e-01 1.38887428e-02
-6.67375028e-01 -1.89003915e-01 -2.64122754e-01 -4.30586100e-01
-2.21307158e-01 -1.04778063e+00 -6.12616539e-01 4.50112730e-01
-9.12290215e-01 1.55508250e-01 7.35926151e-01 -5.27939498e-01
1.01934910e+00 -2.23175406e+00 -2.38043934e-01 4.86030608e-01
8.03171158e-01 2.64441699e-01 -5.17842583e-02 4.55518007e-01
5.00535890e-02 4.33163941e-01 -4.76502180e-01 -5.89865029e-01
1.29782036e-01 1.29728943e-01 -4.68786955e-01 1.08823073e+00
-2.13795096e-01 9.51899052e-01 -9.50828671e-01 -7.38127828e-02
-2.68753588e-01 3.97150367e-01 -6.92874730e-01 1.53763995e-01
-1.13110386e-01 9.49137509e-02 -6.48721814e-01 1.43231750e-01
1.13366318e+00 -5.10710537e-01 3.90111655e-01 3.71722490e-01
6.52441204e-01 1.88289553e-01 -8.71547282e-01 1.58302259e+00
-1.32659391e-01 4.58750725e-01 4.58339959e-01 -9.47776496e-01
7.37449586e-01 1.32329851e-01 3.10243070e-01 -1.92963660e-01
1.10909753e-01 -1.99640393e-02 -3.25508505e-01 -1.35701761e-01
6.99423254e-02 -1.08578116e-01 -3.13638449e-01 8.33727837e-01
-9.96178985e-02 2.88000107e-01 -3.40955436e-01 5.56819379e-01
1.35020578e+00 -4.34265554e-01 -9.06165540e-02 -3.18503827e-01
2.36890316e-01 -5.48820913e-01 6.18370593e-01 9.76182401e-01
-2.94412792e-01 6.67666674e-01 9.72953975e-01 -6.97822511e-01
-7.40846515e-01 -9.27836120e-01 3.30470830e-01 9.16582704e-01
-9.58917663e-02 -4.13392931e-01 -1.08704269e+00 -1.27507949e+00
3.96976352e-01 4.81662601e-01 -8.50078642e-01 -5.13860703e-01
-2.02469543e-01 -8.89918268e-01 6.86960340e-01 1.35958433e-01
4.91625547e-01 -5.65749109e-01 -7.60348663e-02 -1.23966299e-01
1.61761105e-01 -1.05946529e+00 -8.01395416e-01 1.63818851e-01
-5.70951581e-01 -1.07882333e+00 -5.37870646e-01 -6.42973602e-01
1.00107121e+00 3.54011774e-01 9.96985734e-01 2.15952322e-01
-3.80395614e-02 1.60433054e-01 1.31581500e-01 -4.23421472e-01
-3.74893188e-01 4.36067939e-01 -3.24099720e-01 4.01868641e-01
3.09927285e-01 -8.18716586e-01 -5.84975123e-01 -1.55905306e-01
-9.55293179e-01 -3.16242367e-01 5.32471463e-02 5.57074487e-01
4.72581863e-01 1.36779204e-01 1.54492289e-01 -1.73957467e+00
8.52222323e-01 -7.62701929e-01 -8.98034215e-01 1.84973240e-01
-5.05154252e-01 1.73196912e-01 9.11604285e-01 2.86806859e-02
-5.60307980e-01 2.36858711e-01 -5.08629670e-03 -4.32364672e-01
7.92748183e-02 4.14141566e-01 -4.58377510e-01 -4.63563353e-01
7.54263639e-01 4.20826562e-02 5.29805198e-02 -3.27648371e-01
4.93465394e-01 2.49857828e-01 2.80126154e-01 -3.61049950e-01
1.02179742e+00 7.62137055e-01 3.28622013e-01 -4.97367710e-01
-8.18785787e-01 6.31918665e-03 -2.24507421e-01 3.26322407e-01
6.34658158e-01 -9.30338323e-01 -7.27655768e-01 6.71435714e-01
-1.01488698e+00 -4.29684460e-01 -3.17406327e-01 2.62970142e-02
-2.48321831e-01 5.00334680e-01 -7.53616273e-01 -6.51413441e-01
-4.36569721e-01 -8.11976075e-01 8.27682972e-01 -1.36927292e-01
1.44579053e-01 -1.27303529e+00 9.19090025e-03 8.57349560e-02
2.93753833e-01 7.60240257e-01 8.08847845e-01 -1.06600606e+00
-8.02874625e-01 -7.38966107e-01 -4.29992117e-02 4.39634234e-01
5.42023331e-02 -3.03166956e-01 -1.02552545e+00 -6.77485406e-01
2.04270720e-01 -4.21504945e-01 1.03354895e+00 2.24008173e-01
1.54891646e+00 -9.97504115e-01 -4.40943390e-01 1.17416513e+00
1.46002412e+00 -4.46337312e-01 4.82284725e-01 -1.38703480e-01
1.06541455e+00 6.03060007e-01 -1.20447345e-01 4.15046006e-01
4.58831161e-01 2.10892316e-02 7.92845428e-01 -2.72277355e-01
4.92879868e-01 -8.03539872e-01 1.92634732e-01 6.91285310e-03
4.75333750e-01 -6.08390152e-01 -7.35553145e-01 5.28270841e-01
-1.75223219e+00 -7.05818057e-01 -2.72553181e-03 2.41426110e+00
7.64373481e-01 -7.41777988e-03 1.04549833e-01 -2.54603893e-01
7.59802878e-01 6.73910737e-01 -6.93172157e-01 -3.85707527e-01
-1.70083463e-01 8.82783160e-02 1.18953085e+00 6.95666790e-01
-1.24558222e+00 9.47699666e-01 6.31236696e+00 4.96135831e-01
-9.52001214e-01 8.20273068e-03 1.15943515e+00 -2.99185008e-01
-8.91031265e-01 1.38671726e-01 -6.26452327e-01 3.92560869e-01
9.23539221e-01 -5.86154997e-01 6.92379773e-01 1.12742221e+00
-3.10043335e-01 4.74643320e-01 -1.41561580e+00 8.10000241e-01
-1.86849982e-01 -1.45853555e+00 6.38132840e-02 6.18351281e-01
8.30948353e-01 2.65438795e-01 3.84172767e-01 -2.74941325e-02
8.87442291e-01 -1.14106131e+00 3.02016407e-01 5.54835126e-02
6.83911145e-01 -1.09788322e+00 4.81077433e-01 4.16252106e-01
-5.87001622e-01 5.35865575e-02 -6.58999264e-01 1.06654897e-01
-3.16498935e-01 8.42786193e-01 -7.69458711e-01 4.17675346e-01
4.98481363e-01 6.64696157e-01 -4.33115572e-01 5.18555939e-01
-3.59479457e-01 6.58806384e-01 -4.62153345e-01 -8.76336843e-02
3.66467923e-01 -2.13390648e-01 4.78065342e-01 1.01714504e+00
1.21704526e-01 -2.91545372e-02 -1.06922187e-01 1.14173079e+00
-8.96845222e-01 4.00169566e-02 -1.36826146e+00 -2.80417204e-01
3.88834029e-01 1.08668256e+00 -2.09265813e-01 4.69476916e-02
-3.84975135e-01 1.19741356e+00 7.80401111e-01 5.80912769e-01
-4.73043829e-01 -5.48039436e-01 1.01444590e+00 1.73681617e-01
3.12637389e-01 -1.21574439e-01 -1.29455179e-01 -1.32142472e+00
3.47753137e-01 -7.73067653e-01 8.42567921e-01 -2.50062317e-01
-1.53593719e+00 3.72826129e-01 -4.28302705e-01 -7.46590853e-01
-2.04940706e-01 -5.13825595e-01 -6.28446102e-01 1.02557468e+00
-1.34062028e+00 -1.14788401e+00 8.58910680e-02 8.97235930e-01
-7.01666117e-01 -9.36255455e-02 8.57371867e-01 1.84473954e-02
-4.00649607e-01 1.27145982e+00 2.40301371e-01 7.22426176e-01
2.16136202e-01 -1.21415448e+00 1.12966740e+00 1.16020525e+00
3.79157215e-01 6.91732883e-01 4.72418994e-01 -5.31696081e-01
-1.87463129e+00 -1.32213163e+00 8.57353866e-01 -4.18019593e-01
7.37966418e-01 -1.06427312e+00 -9.99934554e-01 1.30335867e+00
-9.03734490e-02 6.19888961e-01 8.36642444e-01 1.07234970e-01
-6.82867885e-01 -2.28015319e-01 -1.64329565e+00 6.96410000e-01
1.25611663e+00 -8.97104919e-01 2.86236346e-01 7.99468994e-01
1.13290262e+00 -5.32596588e-01 -6.64881647e-01 -8.94386619e-02
1.79379225e-01 -1.11400759e+00 7.46938348e-01 -1.16297221e+00
1.41565993e-01 8.45270082e-02 5.98936481e-03 -1.31432056e+00
-1.26403749e-01 -1.34158301e+00 -1.80109978e-01 1.09164858e+00
7.14645267e-01 -1.04149067e+00 1.49490833e+00 1.23722661e+00
4.63724047e-01 -6.67567313e-01 -9.65801716e-01 -5.52932858e-01
5.11009455e-01 -1.85518384e-01 1.02692008e+00 1.07523835e+00
-8.64132121e-02 6.40590563e-02 -6.34555638e-01 6.19023919e-01
9.64732885e-01 1.05168983e-01 9.92252231e-01 -8.84484172e-01
-2.62779862e-01 -3.18199731e-02 -6.59092128e-01 -7.64679790e-01
5.66286266e-01 -1.34477746e+00 -3.23188186e-01 -1.26221073e+00
1.59022734e-01 -2.75610000e-01 -2.73887753e-01 6.01197898e-01
-1.75193802e-01 -7.60990977e-02 -1.16531312e-01 -2.47477531e-01
-3.65435660e-01 5.03824115e-01 1.04916143e+00 -1.49705902e-01
3.91063616e-02 4.06282037e-01 -1.31204355e+00 3.17202628e-01
9.09862876e-01 -7.72375643e-01 -7.03146458e-01 -5.88704407e-01
4.26567674e-01 -1.16662877e-02 4.94567513e-01 -7.21762419e-01
2.45878339e-01 1.68404594e-01 1.05658926e-01 -1.85818061e-01
-1.01144843e-01 -9.54262078e-01 1.94034845e-01 4.03385580e-01
-6.98681355e-01 -2.39287972e-01 4.42246422e-02 1.00194836e+00
1.37590632e-01 7.27099702e-02 8.14691305e-01 -3.15912515e-01
1.33403065e-02 1.19555140e+00 -5.89852221e-03 5.14834166e-01
9.09966886e-01 1.93684846e-01 -3.74909908e-01 -9.84653771e-01
-5.75224400e-01 4.06258732e-01 7.69908249e-01 7.12268278e-02
5.14914870e-01 -1.25441611e+00 -7.59780109e-01 4.27880287e-01
1.99213639e-01 2.68308431e-01 2.18591556e-01 1.48748711e-01
-3.86075318e-01 1.16167970e-01 1.43630013e-01 -1.25732385e-02
-1.15885842e+00 9.63202536e-01 5.94940066e-01 -3.38041306e-01
-7.11606562e-01 1.27602720e+00 5.10998726e-01 -6.23507619e-01
4.68066692e-01 -2.30763316e-01 4.24954087e-01 -4.63686198e-01
5.11306703e-01 2.35079527e-02 -3.03789135e-02 -3.82067949e-01
-2.55631834e-01 -7.38016814e-02 -4.51799929e-01 4.33265604e-02
1.39523745e+00 7.25606456e-02 -3.00375938e-01 -1.77635588e-02
1.97526741e+00 1.48561642e-01 -1.32787502e+00 -3.84092450e-01
-3.68581384e-01 -6.66829884e-01 -9.54430774e-02 -3.10880154e-01
-1.64053917e+00 9.07903135e-01 5.45612946e-02 4.60546017e-01
8.35058093e-01 8.77281372e-03 9.97146487e-01 5.61999083e-01
5.79140604e-01 -6.94152534e-01 -5.99789381e-01 1.73494607e-01
6.02133095e-01 -1.06358874e+00 7.26728067e-02 -4.40530479e-01
-4.05297309e-01 7.18235195e-01 3.17221493e-01 -3.76399606e-01
9.66619730e-01 2.57044077e-01 -1.65246263e-01 -2.53318846e-01
-5.90448022e-01 4.43153024e-01 -6.92203194e-02 7.92185366e-01
6.08760640e-02 1.47155181e-01 8.69028047e-02 5.92447460e-01
-2.14031577e-01 -2.27031589e-01 6.16785169e-01 8.35282207e-01
6.50467426e-02 -1.08115649e+00 2.63060600e-01 5.04070342e-01
-9.51381266e-01 -1.28136158e-01 -8.81011844e-01 7.23646820e-01
-4.25420552e-01 6.75955355e-01 -2.92726278e-01 -4.84973013e-01
3.81327160e-02 -3.23383838e-01 2.11410061e-01 -6.21246994e-01
-5.57771802e-01 -6.49689555e-01 1.36011317e-01 -8.90718043e-01
2.15263397e-01 -6.90431774e-01 -9.29236948e-01 -8.24730515e-01
-8.45354632e-04 1.84256196e-01 4.03861254e-01 4.07449991e-01
7.18761921e-01 -1.70917958e-01 1.04571986e+00 -3.32416117e-01
-1.05937302e+00 -3.05056810e-01 -9.50609446e-01 6.10092044e-01
7.37524867e-01 1.45210877e-01 -7.43832588e-01 -4.79745060e-01] | [6.005266189575195, 7.031274318695068] |
832aa3f0-417c-4204-82d2-2bc4f902e208 | deep-learning-based-fast-and-accurate-3d-ct | 2304.11135 | null | https://arxiv.org/abs/2304.11135v1 | https://arxiv.org/pdf/2304.11135v1.pdf | Deep-Learning-based Fast and Accurate 3D CT Deformable Image Registration in Lung Cancer | Purpose: In some proton therapy facilities, patient alignment relies on two 2D orthogonal kV images, taken at fixed, oblique angles, as no 3D on-the-bed imaging is available. The visibility of the tumor in kV images is limited since the patient's 3D anatomy is projected onto a 2D plane, especially when the tumor is behind high-density structures such as bones. This can lead to large patient setup errors. A solution is to reconstruct the 3D CT image from the kV images obtained at the treatment isocenter in the treatment position. Methods: An asymmetric autoencoder-like network built with vision-transformer blocks was developed. The data was collected from 1 head and neck patient: 2 orthogonal kV images (1024x1024 voxels), 1 3D CT with padding (512x512x512) acquired from the in-room CT-on-rails before kVs were taken and 2 digitally-reconstructed-radiograph (DRR) images (512x512) based on the CT. We resampled kV images every 8 voxels and DRR and CT every 4 voxels, thus formed a dataset consisting of 262,144 samples, in which the images have a dimension of 128 for each direction. In training, both kV and DRR images were utilized, and the encoder was encouraged to learn the jointed feature map from both kV and DRR images. In testing, only independent kV images were used. The full-size synthetic CT (sCT) was achieved by concatenating the sCTs generated by the model according to their spatial information. The image quality of the synthetic CT (sCT) was evaluated using mean absolute error (MAE) and per-voxel-absolute-CT-number-difference volume histogram (CDVH). Results: The model achieved a speed of 2.1s and a MAE of <40HU. The CDVH showed that <5% of the voxels had a per-voxel-absolute-CT-number-difference larger than 185 HU. Conclusion: A patient-specific vision-transformer-based network was developed and shown to be accurate and efficient to reconstruct 3D CT images from kV images. | ['Wei Liu', 'Baoxin Li', 'Steven E. Schild', 'Terence T. Sio', 'Nathan Y. Yu', 'William W. Wong', 'David Liu', 'Zhengliang Liu', 'Jason Holmes', 'Yunze Yang', 'Hongying Feng', 'Yuzhen Ding'] | 2023-04-21 | null | null | null | null | ['anatomy'] | ['miscellaneous'] | [ 2.64344215e-02 3.29377621e-01 6.55297413e-02 -4.18938935e-01
-1.15868890e+00 -5.32597974e-02 1.61284640e-01 1.91408157e-01
-6.02015555e-01 6.27940714e-01 4.58993763e-01 -3.73149842e-01
-1.86399683e-01 -9.01449740e-01 -6.99910998e-01 -1.11213052e+00
7.33290706e-03 7.67176330e-01 2.59385675e-01 3.22167665e-01
-2.72781074e-01 6.69251561e-01 -9.92086291e-01 5.22198975e-01
3.30149978e-01 1.09045923e+00 8.24343503e-01 9.33809817e-01
2.42827684e-01 9.51012731e-01 -3.81747991e-01 1.36232987e-01
4.12334353e-01 -4.12069649e-01 -6.61683738e-01 -7.84679502e-02
5.49775176e-02 -6.75472677e-01 -8.01541984e-01 7.94004142e-01
8.77555192e-01 -1.47632837e-01 7.45655656e-01 -8.58178735e-01
-3.25235039e-01 5.66091418e-01 -6.02565467e-01 2.33567148e-01
1.64992213e-01 5.01084030e-01 2.29670927e-01 -7.13460624e-01
5.33336163e-01 5.13367832e-01 7.37427890e-01 5.34626722e-01
-1.11379838e+00 -3.61227721e-01 -7.95494914e-01 2.74327248e-01
-1.19871819e+00 2.46064752e-01 3.34513277e-01 -5.91250420e-01
1.15068448e+00 4.93127823e-01 1.25617194e+00 8.24204504e-01
9.52121854e-01 2.82245666e-01 1.34333646e+00 -4.31924343e-01
4.10119861e-01 6.51428476e-02 -1.51171044e-01 5.61672509e-01
7.58281797e-02 3.52954715e-01 1.93931684e-01 -2.46739350e-02
1.22042668e+00 2.21735686e-02 -1.06298506e+00 -1.82456568e-01
-1.13104641e+00 1.02659655e+00 9.14996505e-01 6.93082213e-01
-6.81339800e-01 -1.69748702e-04 4.23015267e-01 -1.33932561e-01
1.56700592e-02 2.01502770e-01 -1.85221538e-01 -2.74679363e-02
-7.65081823e-01 -2.81117111e-01 3.95038277e-01 5.85222185e-01
1.22735575e-01 1.81397706e-01 -2.36304760e-01 6.04182899e-01
2.77747214e-01 6.20059669e-01 1.14418721e+00 -5.84068239e-01
5.96119948e-02 3.66728336e-01 -6.06702924e-01 -3.90438110e-01
-7.30191767e-01 -2.77945161e-01 -1.18697560e+00 5.29575825e-01
2.92880952e-01 1.08455576e-01 -1.44268787e+00 1.24535155e+00
4.60421503e-01 -6.72302395e-02 6.57974137e-03 1.15233397e+00
1.11475754e+00 7.00570285e-01 -1.36760250e-01 -4.52593595e-01
1.29354751e+00 -8.81872058e-01 -7.65171349e-01 -2.06386987e-02
7.62540698e-01 -6.95932865e-01 9.52420652e-01 2.14291736e-01
-1.42694533e+00 -4.65473354e-01 -1.11226594e+00 1.86863929e-01
2.20899031e-01 -1.25648931e-01 2.00431556e-01 5.63164890e-01
-1.11434877e+00 6.99648738e-01 -1.06281924e+00 1.00885376e-01
3.61824721e-01 4.45393354e-01 -6.78565681e-01 -3.35353911e-01
-1.08358443e+00 1.28897989e+00 3.62390608e-01 9.66772344e-03
-1.19296825e+00 -1.07533717e+00 -8.80169749e-01 -6.34884164e-02
2.75665987e-02 -8.97829413e-01 1.41999841e+00 -5.94860911e-01
-1.41831315e+00 6.03240013e-01 3.73491138e-01 -3.72343332e-01
6.60321534e-01 3.80759031e-01 -3.25131804e-01 4.12642330e-01
6.79714903e-02 3.99063885e-01 6.12367213e-01 -1.20236719e+00
-2.29936957e-01 -6.61853135e-01 -5.49239755e-01 5.53746879e-01
2.82896012e-01 -2.17430219e-01 -4.15565163e-01 -2.75983006e-01
4.19189155e-01 -8.10873985e-01 -4.32709426e-01 1.69605762e-01
-5.13746083e-01 5.96637726e-01 6.10791802e-01 -9.57962334e-01
6.76031053e-01 -1.73571694e+00 -8.39085653e-02 2.74812847e-01
4.15169984e-01 -2.25778744e-01 3.91518801e-01 -1.50807410e-01
-6.06770098e-01 -3.56469452e-01 -6.44826233e-01 -5.50821349e-02
-6.89656854e-01 3.21828991e-01 4.33109134e-01 8.13102126e-01
-7.18155801e-01 9.80743408e-01 -7.67156601e-01 -5.63526034e-01
5.81545293e-01 8.11708868e-01 -2.05120847e-01 3.67882937e-01
3.55615139e-01 7.65313268e-01 -2.39616171e-01 3.53127658e-01
1.04532230e+00 -2.98135102e-01 3.58120985e-02 -6.15276992e-01
-9.20283720e-02 -1.25632495e-01 -7.28707492e-01 1.69071412e+00
-6.78512037e-01 5.24550259e-01 1.13141067e-01 -3.79940659e-01
3.12419087e-01 6.19362950e-01 9.53146458e-01 -1.03909481e+00
4.59312886e-01 2.53804028e-01 2.69701570e-01 -5.39338887e-01
-2.81316936e-02 -6.21273339e-01 2.68504232e-01 4.38149154e-01
-1.94732502e-01 -9.65683520e-01 -5.05271614e-01 6.10500574e-05
1.11771083e+00 -6.10670745e-01 2.46799007e-01 1.40127605e-02
1.60809353e-01 1.23640671e-01 1.18370719e-01 5.69729984e-01
-1.67016819e-01 1.08306038e+00 1.81383207e-01 -4.23421800e-01
-1.38472199e+00 -1.30885541e+00 -6.91042066e-01 -2.53497481e-01
-2.65766084e-01 1.22547626e-01 -8.93526077e-01 -6.35822594e-01
-3.81185830e-01 1.08791864e+00 -7.29408264e-01 -1.62910551e-01
-6.81380689e-01 -8.36620271e-01 3.01374704e-01 8.44256699e-01
3.15669835e-01 -1.09871888e+00 -1.02602565e+00 2.34027147e-01
-2.63566703e-01 -1.10316408e+00 -3.98788184e-01 6.10051215e-01
-1.11854815e+00 -1.16006494e+00 -1.03863645e+00 -5.85684717e-01
7.87746012e-01 -4.81847227e-02 1.09488809e+00 -2.26847306e-01
-6.38774931e-01 4.61283475e-01 -1.20644294e-01 -2.03347072e-01
-7.49951661e-01 -7.93102682e-01 2.26673946e-01 -6.59485459e-01
-2.31542498e-01 -5.01928985e-01 -8.96599352e-01 9.31878090e-02
-1.10656905e+00 3.17066669e-01 5.45375466e-01 1.31394207e+00
1.04947877e+00 3.17904919e-01 -2.57270783e-01 -3.85738283e-01
3.23225230e-01 -7.92018548e-02 -4.64077950e-01 1.92517079e-02
-4.49524701e-01 -1.73630267e-01 6.63379788e-01 -1.76945105e-01
-8.68519127e-01 1.40098959e-01 -7.06613362e-01 -7.78400898e-01
-1.37607664e-01 2.59567708e-01 1.55501282e-02 -2.51113683e-01
7.24624217e-01 6.17303491e-01 3.13360184e-01 1.41440123e-01
1.46958694e-01 5.98555148e-01 5.57289720e-01 -1.15185659e-02
5.08501947e-01 5.35001814e-01 -2.00758018e-02 -7.97174752e-01
-2.85721272e-01 -1.62085682e-01 -6.43120348e-01 -4.23305720e-01
1.44283199e+00 -7.23838210e-01 -6.97248340e-01 4.72465366e-01
-8.21435273e-01 -5.17578304e-01 -9.29391086e-01 1.39076221e+00
-6.79387391e-01 4.15937632e-01 -1.05890214e+00 -2.26538435e-01
-7.32474208e-01 -1.89253211e+00 6.47416890e-01 3.03294659e-01
7.51653537e-02 -7.37644315e-01 2.53242791e-01 2.34108359e-01
5.26560605e-01 1.08500414e-01 1.18105865e+00 -3.36396903e-01
-2.30576947e-01 -1.70784846e-01 -6.57581314e-02 4.97725099e-01
1.12413116e-01 -4.54523295e-01 -1.01707137e+00 -3.32051039e-01
8.80978703e-01 -4.62525457e-01 3.16956997e-01 1.33128130e+00
1.46988392e+00 1.34774819e-01 -1.78589016e-01 8.48973393e-01
1.75619411e+00 7.02821553e-01 7.64371753e-01 1.47263288e-01
8.20791066e-01 3.09202578e-02 1.22765534e-01 3.84665281e-01
9.78201404e-02 6.73232734e-01 9.00656164e-01 -4.16953146e-01
-2.34503105e-01 -2.55632419e-02 -1.17873862e-01 8.47184360e-01
-4.29596528e-02 3.35022286e-02 -8.95592153e-01 3.18193436e-01
-8.52276385e-01 -7.14480579e-01 -4.52248156e-01 2.29537249e+00
5.94683707e-01 6.32013530e-02 -4.26606238e-01 3.32100451e-01
4.62998956e-01 -1.59580916e-01 -5.03849983e-01 -3.40652943e-01
8.86898562e-02 5.61164081e-01 8.05005908e-01 7.03589737e-01
-5.91508925e-01 -2.91171111e-02 5.79275846e+00 7.99493909e-01
-1.38401353e+00 4.29818213e-01 8.43211472e-01 -1.79468021e-01
-3.30144465e-01 -3.94411236e-01 -2.70859927e-01 4.85580236e-01
9.33503151e-01 2.07541078e-01 1.08380660e-01 9.78341401e-01
2.71860570e-01 -6.55308902e-01 -1.00322342e+00 1.10860801e+00
-4.69283015e-03 -1.19125772e+00 -3.30736905e-01 1.93617731e-01
5.06614566e-01 3.34205300e-01 9.52895433e-02 2.19872445e-01
1.51716605e-01 -1.35846889e+00 4.70206231e-01 1.42352790e-01
1.31227708e+00 -5.09925723e-01 8.53463590e-01 5.15078843e-01
-8.77205968e-01 1.68950588e-01 -3.57557356e-01 8.11910629e-01
5.47268093e-01 7.39404380e-01 -1.49345267e+00 8.19473386e-01
9.96800840e-01 8.69599171e-03 -1.12584710e-01 9.34868813e-01
-5.76614551e-02 2.71561414e-01 -6.20672345e-01 2.97235012e-01
2.13939190e-01 -1.32531330e-01 3.08464587e-01 9.13398802e-01
6.22231126e-01 6.64403081e-01 -2.90819973e-01 5.10480285e-01
2.10185125e-01 -1.15295187e-01 -5.58968544e-01 6.86147332e-01
3.35673951e-02 1.24122250e+00 -7.45856166e-01 -3.65203798e-01
-3.46053749e-01 9.23725247e-01 -2.36962661e-01 -9.50319916e-02
-1.13684762e+00 2.95182556e-01 -7.72611052e-02 3.65293175e-01
1.08785681e-01 2.46709853e-01 -2.68670112e-01 -7.39800096e-01
-1.44667104e-01 -5.90648115e-01 4.98484612e-01 -1.17954516e+00
-1.17498505e+00 1.26280081e+00 1.33284003e-01 -1.22843051e+00
-2.89918602e-01 -6.26820505e-01 -6.24786079e-01 1.01863420e+00
-1.13887227e+00 -7.47816443e-01 -6.67176843e-01 7.73229241e-01
5.41514635e-01 2.76701033e-01 1.07062411e+00 3.34597260e-01
-8.26895088e-02 3.66833687e-01 3.38535935e-01 6.81856945e-02
3.40329319e-01 -1.38958275e+00 -1.40903905e-01 7.41802454e-02
-2.98801094e-01 -8.29493329e-02 4.69848365e-01 -6.58120751e-01
-1.40436554e+00 -9.15925443e-01 3.13682467e-01 -2.19082326e-01
2.57611334e-01 5.27071171e-02 -8.35186422e-01 7.74804711e-01
2.91856915e-01 5.76303661e-01 6.11345291e-01 -8.46872747e-01
3.71667564e-01 1.88935086e-01 -1.80708385e+00 2.43138835e-01
2.52665490e-01 -3.49573135e-01 -4.02268171e-01 4.72276956e-01
5.36511421e-01 -1.31229210e+00 -1.40750182e+00 5.23766994e-01
3.75842541e-01 -1.15690577e+00 1.09712362e+00 6.01689629e-02
6.19063377e-01 -1.66016415e-01 -2.74921238e-01 -1.69387758e+00
-4.02005404e-01 5.65604031e-01 2.47725561e-01 1.49154067e-02
3.02364290e-01 -5.05407691e-01 9.99689281e-01 5.20397305e-01
-7.21302152e-01 -1.16439879e+00 -1.44759667e+00 -3.86936188e-01
1.97183758e-01 -6.37405694e-01 1.90673620e-01 9.25065398e-01
-2.17380419e-01 -2.13080794e-01 2.40012988e-01 3.17962229e-01
5.81864834e-01 -3.49602759e-01 1.83972381e-02 -4.94608074e-01
-7.69944727e-01 -1.80106267e-01 -3.84882212e-01 -7.39820600e-01
-4.90413934e-01 -1.02849555e+00 -2.13528872e-02 -1.80694830e+00
4.97197568e-01 -3.66815567e-01 -7.33278990e-02 2.56469637e-01
3.85985106e-01 1.16590904e-02 2.13714950e-02 5.05344160e-02
6.28255844e-01 6.24009430e-01 1.99857330e+00 -3.33831996e-01
1.71364695e-02 -3.37442867e-02 5.16784526e-02 7.80451655e-01
4.58832473e-01 -5.59847593e-01 -4.77382839e-01 -3.47782195e-01
-5.44211209e-01 9.67075348e-01 3.16351026e-01 -1.18117082e+00
2.84690470e-01 1.48658037e-01 1.11756957e+00 -1.10204327e+00
7.32612491e-01 -1.27796113e+00 5.89731514e-01 8.55825603e-01
1.27699390e-01 2.51025587e-01 2.36941010e-01 1.34653673e-01
-1.46079674e-01 -3.98534954e-01 1.21815813e+00 -6.22082889e-01
-5.85925095e-02 4.57172900e-01 -3.06304991e-01 -3.26952249e-01
1.40695608e+00 -4.27327842e-01 1.09853961e-01 -3.41805875e-01
-1.00400746e+00 -3.42955321e-01 4.40479934e-01 -3.44087869e-01
1.23958457e+00 -1.39425290e+00 -6.52598739e-01 3.38638335e-01
-2.06519797e-01 6.00107789e-01 1.04560196e+00 1.24162996e+00
-1.05802977e+00 4.08891827e-01 -3.27582061e-01 -1.09038627e+00
-1.17314625e+00 4.83185232e-01 9.47412848e-01 -7.41162598e-01
-1.25627613e+00 1.02894247e+00 4.42654699e-01 -3.70327413e-01
-9.09491926e-02 -5.20603120e-01 -8.54699537e-02 -3.77170175e-01
5.03053665e-01 -1.82329297e-01 6.76205635e-01 -6.81076169e-01
-1.14354737e-01 6.74386382e-01 -1.53364748e-01 -3.49638224e-01
1.38809180e+00 3.04324567e-01 3.08054596e-01 1.24696977e-01
1.40395916e+00 -2.80760825e-01 -1.15320122e+00 1.38577685e-01
-8.64882290e-01 -5.34727514e-01 4.94950026e-01 -8.65725935e-01
-1.40930533e+00 7.98472524e-01 1.22845817e+00 -1.39652058e-01
1.34766471e+00 1.66254176e-03 7.11657643e-01 -4.40806240e-01
3.72466892e-01 -6.72371209e-01 -9.77752358e-03 2.72903860e-01
8.81147087e-01 -1.12332630e+00 1.15372039e-01 -2.89367944e-01
-1.06002140e+00 1.22474933e+00 6.80024028e-01 1.42845377e-01
6.79982722e-01 6.62397385e-01 1.40908822e-01 -3.94619733e-01
-3.56912583e-01 5.35977483e-01 -7.80858919e-02 6.28557742e-01
3.02561611e-01 1.98673368e-01 8.24522749e-02 2.45324999e-01
-3.21286738e-01 -7.18583390e-02 9.08278346e-01 1.10513759e+00
-1.89549312e-01 -5.73678970e-01 -8.17159534e-01 5.26577592e-01
-3.30123842e-01 -1.72085866e-01 5.12264192e-01 1.02066731e+00
-7.42182657e-02 2.62905806e-01 9.46605578e-02 -2.59492546e-01
4.40581948e-01 -3.77940595e-01 9.24787521e-01 -3.54058594e-01
-7.65517116e-01 1.58718467e-01 -3.20842505e-01 -5.47557712e-01
2.62828171e-01 -3.58412743e-01 -1.61573899e+00 -1.56775996e-01
-2.44576409e-01 1.16889320e-01 1.10631812e+00 4.49803293e-01
-5.02966464e-01 9.83056664e-01 8.39726806e-01 -8.41461480e-01
-5.75066745e-01 -1.02181399e+00 -9.04364884e-01 5.71413152e-02
2.46775836e-01 -3.46324444e-01 -5.11525095e-01 -3.14666688e-01] | [13.550982475280762, -2.5820529460906982] |
567e2018-8214-4c15-8576-7f336a7ad899 | social-media-writing-style-fingerprint | 1712.04762 | null | http://arxiv.org/abs/1712.04762v3 | http://arxiv.org/pdf/1712.04762v3.pdf | Social Media Writing Style Fingerprint | We present our approach for computer-aided social media text authorship
attribution based on recent advances in short text authorship verification. We
use various natural language techniques to create word-level and
character-level models that act as hidden layers to simulate a simple neural
network. The choice of word-level and character-level models in each layer was
informed through validation performance. The output layer of our system uses an
unweighted majority vote vector to arrive at a conclusion. We also considered
writing bias in social media posts while collecting our training dataset to
increase system robustness. Our system achieved a precision, recall, and
F-measure of 0.82, 0.926 and 0.869 respectively. | ['Juliang Li', 'Himank Yadav'] | 2017-12-11 | null | null | null | null | ['authorship-verification'] | ['natural-language-processing'] | [ 2.65799791e-01 2.57879257e-01 -4.38999593e-01 -3.88794869e-01
-4.45920043e-02 -3.84711266e-01 8.34024906e-01 5.37515700e-01
-6.96112096e-01 7.76431739e-01 8.10790285e-02 -6.54719889e-01
2.67778814e-01 -7.05768466e-01 -2.52480716e-01 -1.30158141e-01
2.01371312e-01 1.04850702e-01 -1.95935354e-01 1.00915648e-01
9.75555837e-01 5.68633735e-01 -1.14359307e+00 3.86140764e-01
8.79407287e-01 7.04665601e-01 -4.52272624e-01 1.10587060e+00
-1.93955511e-01 1.23480093e+00 -8.82457852e-01 -8.32102537e-01
1.26790956e-01 -3.75177562e-01 -6.33124709e-01 -2.50988752e-01
4.49663967e-01 -9.16675553e-02 -2.50704795e-01 1.03613734e+00
4.74001199e-01 -2.38706425e-01 1.16068792e+00 -1.53898299e+00
-8.85302901e-01 8.02720308e-01 -6.40927672e-01 -6.25335127e-02
3.95396411e-01 1.27254695e-01 8.23828399e-01 -8.24363768e-01
4.64579225e-01 1.14084589e+00 9.63970482e-01 4.40336198e-01
-1.01156366e+00 -1.13019228e+00 -8.97111520e-02 -1.46923944e-01
-1.45948625e+00 -4.15065885e-01 6.54522181e-01 -6.05826139e-01
7.66425133e-01 5.56817725e-02 3.89797032e-01 1.03910983e+00
6.01374030e-01 5.28298616e-01 1.28365970e+00 -8.18512201e-01
-6.86496682e-03 4.59410965e-01 4.49847877e-01 1.12613308e+00
5.51440418e-01 -2.02004477e-01 -6.69833481e-01 -5.89931667e-01
4.86684471e-01 -2.38579512e-01 2.93824345e-01 4.42177415e-01
-1.19902921e+00 8.80640686e-01 2.56658763e-01 3.97644520e-01
-2.79660732e-01 -2.03819964e-02 4.17369246e-01 3.80313456e-01
1.34607911e-01 7.01127708e-01 -1.06853046e-01 4.60200816e-01
-1.21413136e+00 4.53850217e-02 1.06819499e+00 7.71998465e-01
4.77915853e-01 5.46811558e-02 -4.75910723e-01 7.79010236e-01
5.83673239e-01 4.04229194e-01 8.97183716e-01 -8.19893301e-01
2.27885813e-01 7.08519518e-01 6.90662041e-02 -1.36995578e+00
-2.85911351e-01 -5.05195737e-01 -1.08683872e+00 1.77710608e-01
5.06194890e-01 -2.68514425e-01 -9.43663001e-01 1.22798240e+00
-3.51274550e-01 -9.83099490e-02 7.66862631e-02 5.17002523e-01
7.30622947e-01 2.95859486e-01 2.43073389e-01 7.01494217e-02
1.37725806e+00 -9.26051140e-01 -8.80869389e-01 5.87049946e-02
6.34082675e-01 -8.87938797e-01 6.86782181e-01 3.02948743e-01
-8.51797462e-01 -5.53173304e-01 -1.27153039e+00 -6.85813874e-02
-6.16676629e-01 4.89881724e-01 3.36144626e-01 8.15254509e-01
-9.40113366e-01 9.13493514e-01 -1.63697436e-01 -1.40707508e-01
5.85202575e-01 4.93721485e-01 -4.16414827e-01 6.92064345e-01
-1.26691329e+00 9.57241356e-01 1.52056694e-01 -9.37420949e-02
-2.11135328e-01 -1.97193876e-01 -5.81515610e-01 -2.42390111e-01
-3.32385808e-01 -4.81606632e-01 1.22016358e+00 -1.10289752e+00
-1.58493114e+00 1.21479928e+00 -3.85619998e-01 -4.54544216e-01
8.47988427e-01 2.33224481e-01 -5.34373522e-01 -2.33756304e-01
4.40995917e-02 6.06157959e-01 7.27685928e-01 -9.83155847e-01
-4.85653847e-01 -2.50606209e-01 -5.84214985e-01 -2.20697373e-01
-6.39341772e-01 1.31264463e-01 9.49238893e-03 -1.01208937e+00
-3.80487852e-02 -8.95183265e-01 -5.77339111e-03 3.06282520e-01
-7.13737190e-01 -3.81482363e-01 3.57464373e-01 -9.53273892e-01
1.56548762e+00 -1.59146178e+00 -3.18199754e-01 7.04482794e-01
5.97995341e-01 2.62841195e-01 1.14037625e-01 4.83208567e-01
7.41826817e-02 5.29718697e-01 -1.36056006e-01 -6.24147475e-01
-7.10605830e-02 -3.60175759e-01 -1.27477095e-01 5.50622225e-01
4.39479277e-02 8.13654959e-01 -7.79536903e-01 -7.65421033e-01
-1.88340589e-01 2.60240823e-01 -9.67673585e-02 3.91315073e-01
2.33923391e-01 -2.22644821e-01 -1.25274077e-01 6.47202849e-01
3.76929432e-01 -3.39692444e-01 9.20306370e-02 2.00473234e-01
-6.16838373e-02 1.60914525e-01 -9.91591454e-01 1.09084022e+00
-2.58755773e-01 1.12831056e+00 -4.39498454e-01 -2.63623983e-01
1.34239244e+00 2.07769036e-01 -9.67520699e-02 -2.44500190e-01
6.09119177e-01 3.44770938e-01 1.41983002e-01 -4.01761025e-01
8.59257638e-01 -4.28097844e-02 -9.62790754e-03 8.15206468e-01
-1.35783911e-01 3.86240929e-01 -6.31830096e-02 4.20167804e-01
7.27463365e-01 -1.76800668e-01 4.52840000e-01 -3.40687841e-01
6.80239379e-01 -2.53774166e-01 3.76044482e-01 1.22793257e+00
-2.69378811e-01 6.50447547e-01 6.91999793e-01 -3.28527123e-01
-1.18837738e+00 -3.81176353e-01 -1.01095758e-01 8.11026812e-01
-2.25613669e-01 -3.20817292e-01 -8.40180397e-01 -8.73764575e-01
2.84392983e-01 7.22415328e-01 -8.71836424e-01 1.76559519e-02
-3.01361948e-01 -3.72961432e-01 1.26326334e+00 3.92962217e-01
3.31934303e-01 -1.12714732e+00 -3.05183470e-01 -4.43697013e-02
2.99458414e-01 -7.47240365e-01 -5.00713289e-01 -1.63902994e-02
-6.07269585e-01 -1.13300455e+00 -1.01684308e+00 -8.60537708e-01
9.37359154e-01 -3.57256323e-01 6.70784295e-01 5.05601048e-01
5.75131699e-02 -3.58149171e-01 4.40185983e-03 -5.27560413e-01
-8.50272238e-01 4.90901679e-01 2.08073035e-01 1.65151820e-01
4.58048582e-01 -3.56937051e-02 -2.70671725e-01 1.97147384e-01
-6.33020341e-01 7.54705444e-02 4.39956427e-01 8.76509309e-01
-3.03721577e-01 -3.49045515e-01 4.29859668e-01 -1.09470606e+00
1.32506120e+00 -4.21077281e-01 -4.27085876e-01 3.24813962e-01
-1.18151808e+00 1.58317342e-01 7.12264001e-01 -7.08552361e-01
-4.85881358e-01 -1.64293125e-01 3.79108451e-02 -7.32152909e-02
1.00850485e-01 4.53904957e-01 2.60989547e-01 -1.51771054e-01
6.10365272e-01 4.07238677e-03 2.85434514e-01 -1.53953806e-01
-1.60820931e-02 1.52304780e+00 3.82398903e-01 -1.22248761e-01
6.68756008e-01 -1.32959113e-01 -1.54562995e-01 -6.47847593e-01
-3.90019387e-01 -5.59064932e-02 -6.94594800e-01 -2.79096961e-01
4.34930205e-01 -6.25726223e-01 -1.31158066e+00 9.52975035e-01
-1.28333044e+00 -1.15268774e-01 2.94924051e-01 2.11727992e-01
9.96474922e-02 3.46247762e-01 -6.93585277e-01 -1.07232797e+00
-7.27522731e-01 -7.07500458e-01 6.42049313e-01 3.16071332e-01
-1.03977036e+00 -9.97422338e-01 -6.43396676e-02 2.71583676e-01
5.12664616e-01 1.36293486e-01 8.27320397e-01 -9.27771866e-01
3.30339164e-01 -7.07026422e-01 -4.65571821e-01 2.35426217e-01
6.90253675e-02 4.45497602e-01 -1.21173894e+00 -4.31831330e-02
-5.04672110e-01 -1.43742427e-01 6.99334919e-01 -8.42659250e-02
1.35317039e+00 -6.72256470e-01 -3.86814982e-01 3.14301699e-01
1.00133324e+00 -6.11308925e-02 6.15143538e-01 4.04494047e-01
8.30291331e-01 6.74225986e-01 -4.07855511e-02 4.53297764e-01
4.39688265e-01 2.50484169e-01 -1.38988897e-01 -4.51037139e-02
3.63717005e-02 -6.29227638e-01 5.62973209e-02 3.65032583e-01
1.76552221e-01 -7.40945637e-01 -1.34526408e+00 2.41456956e-01
-1.65426743e+00 -8.07979643e-01 -3.90100867e-01 2.03027987e+00
8.95173669e-01 3.42238277e-01 1.41572595e-01 5.55845499e-01
9.31197703e-01 4.11148602e-03 -3.96966398e-01 -7.82590568e-01
-2.76604116e-01 -7.22829252e-02 9.47308660e-01 7.85008550e-01
-8.22341502e-01 9.16588783e-01 6.88383770e+00 3.93718123e-01
-1.24640000e+00 -4.01981264e-01 7.04541802e-01 3.01832736e-01
-2.06135079e-01 -3.14380258e-01 -7.77120411e-01 7.56371737e-01
9.37503040e-01 -1.39287025e-01 1.79517269e-01 3.56167704e-01
-2.35586595e-02 -9.21487436e-02 -9.12867486e-01 9.23566997e-01
3.25314432e-01 -1.30377233e+00 2.55648047e-01 1.14579253e-01
7.31382668e-01 -3.43465477e-01 8.31597596e-02 -7.19809681e-02
4.48472202e-01 -1.44629931e+00 8.11468840e-01 8.28504264e-01
8.67032528e-01 -5.40446281e-01 9.33496296e-01 3.69746894e-01
-5.01689911e-01 -6.98705912e-02 1.69280153e-02 -4.36556965e-01
-2.70245910e-01 4.85439241e-01 -1.10055828e+00 -1.15633421e-01
6.21165484e-02 5.49212694e-01 -6.64061964e-01 6.12859726e-01
-3.43780220e-01 5.52009821e-01 -4.05968130e-02 -6.90486133e-01
-1.40060182e-03 3.93386036e-01 1.17917277e-01 1.46309924e+00
2.70592961e-02 -1.27783090e-01 -2.52367347e-01 6.27969444e-01
-7.27412283e-01 2.08217800e-01 -4.40133721e-01 -2.03757182e-01
6.00540221e-01 9.55044329e-01 -7.39979208e-01 -5.56267977e-01
4.45085913e-02 1.22470236e+00 3.50778669e-01 2.10350722e-01
-5.27479649e-01 -8.06254089e-01 2.37336233e-01 3.32731456e-01
-3.77307564e-01 -1.44680291e-01 -1.09230971e+00 -1.20150864e+00
-1.71321422e-01 -9.64610517e-01 3.66181225e-01 -7.58307755e-01
-1.45385194e+00 4.89552796e-01 -5.22517204e-01 -8.96934927e-01
-4.51528430e-02 -8.24793279e-01 -6.89941466e-01 1.34583127e+00
-1.17970467e+00 -7.98877597e-01 -2.33580917e-01 1.10414870e-01
4.13480364e-02 -8.14879894e-01 1.00329876e+00 -1.03640281e-01
-8.13122213e-01 1.32574117e+00 -1.17993549e-01 7.82119274e-01
9.73689497e-01 -1.08394563e+00 4.39327806e-01 5.21943748e-01
-2.75100291e-01 1.16578865e+00 5.60151935e-01 -1.02701092e+00
-1.10460484e+00 -8.90019476e-01 1.79914820e+00 -4.95725900e-01
7.02798903e-01 -3.15879256e-01 -7.68617272e-01 4.07107085e-01
3.89778763e-01 -2.21901938e-01 1.03549850e+00 1.65254995e-01
-5.37802041e-01 2.67474830e-01 -1.50621974e+00 6.15563989e-01
6.72639251e-01 -7.76192486e-01 -5.00766337e-01 5.94099313e-02
-2.44876333e-02 -6.64339736e-02 -8.21425974e-01 8.98798928e-02
1.33158433e+00 -5.01829088e-01 4.53323841e-01 -7.33044326e-01
8.48777473e-01 -2.30955884e-01 1.79246604e-01 -9.20947731e-01
-4.43566591e-01 -5.07414281e-01 -1.01404563e-01 1.07600093e+00
8.80542576e-01 -7.54883587e-01 8.69219959e-01 7.87350059e-01
6.52369142e-01 -5.89520931e-01 -3.57593209e-01 -3.95705372e-01
2.62426585e-01 -5.62116913e-02 5.05183637e-01 1.28616774e+00
2.32599035e-01 3.92613977e-01 -5.71907461e-01 -8.29235762e-02
7.41600394e-01 -2.24848121e-01 6.75607562e-01 -1.53077614e+00
1.60352767e-01 -7.82798648e-01 -3.89647663e-01 -4.50520009e-01
4.61579680e-01 -9.52175915e-01 -2.83833742e-01 -1.28479600e+00
2.75995404e-01 -4.78112131e-01 -4.74225074e-01 5.54662347e-01
-1.80932105e-01 7.32921958e-01 6.59089461e-02 4.67628360e-01
-5.75141720e-02 4.14238721e-02 7.57140160e-01 -2.62717009e-01
-1.65247068e-01 -1.64314955e-01 -8.59639168e-01 6.65707946e-01
1.23054135e+00 -6.35733008e-01 1.97468296e-01 -2.55258441e-01
2.52456397e-01 -1.84897378e-01 1.52668357e-01 -6.35285318e-01
5.41592717e-01 -6.79920092e-02 9.61034954e-01 -2.31465310e-01
-8.95958915e-02 -4.36079502e-01 -1.03133529e-01 8.69769871e-01
-1.07707691e+00 6.04067519e-02 -9.17759091e-02 7.25207478e-02
1.33344606e-01 -3.66552949e-01 7.52443612e-01 7.38246888e-02
-1.31172389e-01 -6.28514588e-02 -7.41403699e-01 -4.98354226e-01
6.45260096e-01 -6.01520479e-01 -2.68841624e-01 -2.77031988e-01
-3.47466618e-01 2.58638591e-01 7.60631084e-01 7.12828398e-01
4.84511793e-01 -1.10283673e+00 -8.55341971e-01 4.70114410e-01
2.94871509e-01 -1.02370679e+00 -4.89705056e-01 3.34711939e-01
-6.79137409e-01 3.04836094e-01 -3.58176738e-01 -1.41777813e-01
-1.55015278e+00 9.35620535e-03 3.82279754e-01 4.82150577e-02
-6.17286144e-03 7.01419175e-01 -7.88907468e-01 -4.91196781e-01
5.76708734e-01 -5.44969516e-04 -4.52950358e-01 1.82769150e-01
5.42099535e-01 6.92221224e-01 -1.50916830e-01 -8.55178177e-01
-4.57427323e-01 3.34915787e-01 -9.32983980e-02 -3.95885259e-01
7.16866970e-01 2.15113744e-01 -2.20067903e-01 6.84114814e-01
1.24416292e+00 2.40094215e-01 -4.22100902e-01 -2.03667849e-01
2.51765288e-02 -2.70544469e-01 2.18089044e-01 -1.05555415e+00
-4.98075426e-01 7.17415810e-01 4.87340450e-01 1.04264706e-01
3.73194277e-01 -6.63753271e-01 4.39243495e-01 4.91208762e-01
-1.81689262e-02 -1.14877975e+00 -3.80250812e-01 6.09416008e-01
6.46275938e-01 -1.35168076e+00 3.07230413e-01 7.25415945e-02
-4.94254202e-01 1.38539767e+00 4.16169971e-01 -6.47174940e-02
4.77985382e-01 3.97107750e-01 4.12526637e-01 2.64300913e-01
-3.76581132e-01 5.19203424e-01 5.22684574e-01 9.28060785e-02
9.90389585e-01 2.06734478e-01 -6.30573571e-01 4.50259954e-01
-4.93504614e-01 2.26957470e-01 5.96529961e-01 9.19984341e-01
-2.74097770e-01 -1.19952786e+00 -5.02790809e-01 7.89447248e-01
-6.98737800e-01 -3.67795788e-02 -1.00651360e+00 2.11192071e-01
2.31259651e-02 1.01952040e+00 3.64299566e-02 -7.88655221e-01
-3.66603732e-02 2.16029465e-01 1.13882676e-01 -3.31532300e-01
-1.14128053e+00 -7.13413537e-01 1.82584986e-01 6.89073130e-02
-9.52244326e-02 -5.47178090e-01 -1.11044681e+00 -7.80489922e-01
-4.38700855e-01 3.30203533e-01 9.40964162e-01 6.65622175e-01
2.40027189e-01 2.74977416e-01 5.99880099e-01 -3.60709280e-01
-3.68448317e-01 -1.30507922e+00 -1.62053436e-01 1.21292442e-01
5.46171784e-01 -3.22402157e-02 -4.09321845e-01 6.69639111e-02] | [9.625229835510254, 10.565197944641113] |
a55f22ba-dfc7-453f-930e-dbaa6d909882 | patient-independent-interictal-epileptiform | 2304.13965 | null | https://arxiv.org/abs/2304.13965v2 | https://arxiv.org/pdf/2304.13965v2.pdf | Patient Independent Interictal Epileptiform Discharge Detection | Epilepsy is a highly prevalent brain condition with many serious complications arising from it. The majority of patients which present to a clinic and undergo electroencephalogram (EEG) monitoring would be unlikely to experience seizures during the examination period, thus the presence of interictal epileptiform discharges (IEDs) become effective markers for the diagnosis of epilepsy. Furthermore, IED shapes and patterns are highly variable across individuals, yet trained experts are still able to identify them through EEG recordings - meaning that commonalities exist across IEDs that an algorithm can be trained on to detect and generalise to the larger population. This research proposes an IED detection system for the binary classification of epilepsy using scalp EEG recordings. The proposed system features an ensemble based deep learning method to boost the performance of a residual convolutional neural network, and a bidirectional long short-term memory network. This is implemented using raw EEG data, sourced from Temple University Hospital's EEG Epilepsy Corpus, and is found to outperform the current state of the art model for IED detection across the same dataset. The achieved accuracy and Area Under Curve (AUC) of 94.92% and 97.45% demonstrates the effectiveness of an ensemble method, and that IED detection can be achieved with high performance using raw scalp EEG data, thus showing promise for the proposed approach in clinical settings. | ['Amitava Datta', 'Ghulam Mubashar Hassan', 'Hezam Albaqami', 'Matthew McDougall'] | 2023-04-27 | null | null | null | null | ['eeg', 'eeg'] | ['methodology', 'time-series'] | [ 1.25899836e-01 -2.75352955e-01 4.90855098e-01 -2.90369928e-01
-6.72312617e-01 -3.97158891e-01 2.38304704e-01 7.10256174e-02
-3.67030084e-01 8.23578477e-01 2.38553509e-02 -2.76513875e-01
-5.07364333e-01 -4.76456881e-01 -3.97987545e-01 -6.48309112e-01
-7.69571900e-01 2.61088550e-01 -1.79481775e-01 3.58043909e-02
3.75806689e-01 6.05283141e-01 -1.35588729e+00 5.43751419e-01
9.28376615e-01 1.20955920e+00 4.87263471e-01 2.35734001e-01
3.02787483e-01 3.41411591e-01 -9.95079398e-01 2.00168312e-01
4.24922891e-02 -1.12272657e-01 -4.98909444e-01 -1.04795396e-01
-9.33432430e-02 -1.55860648e-01 -1.32597595e-01 7.67457664e-01
9.77879107e-01 -2.90899932e-01 8.67331445e-01 -9.54933524e-01
-2.58955777e-01 1.63480610e-01 -2.45355457e-01 6.83016300e-01
5.53002536e-01 -2.53519807e-02 3.88156831e-01 -9.18927193e-01
1.36744916e-01 2.80377686e-01 6.73552990e-01 3.26793849e-01
-9.75515664e-01 -1.19851363e+00 -2.98196644e-01 2.53129244e-01
-1.67113352e+00 -3.07213783e-01 4.45479393e-01 -7.26861954e-01
1.38948691e+00 -5.44832647e-02 9.59660769e-01 1.24572611e+00
1.00332165e+00 3.64397347e-01 1.49357998e+00 -6.75441623e-02
2.15838850e-01 1.60330012e-01 3.32941636e-02 3.13828439e-02
1.75578415e-01 2.98508584e-01 -6.18956506e-01 -1.61957785e-01
5.65279126e-01 1.36982754e-01 -5.82054138e-01 1.33122265e-01
-1.13606250e+00 4.99332041e-01 3.00920814e-01 7.06436276e-01
-8.96649599e-01 -4.98182893e-01 4.31343079e-01 4.32302088e-01
4.33762223e-01 5.41209817e-01 -4.34991807e-01 -1.37592673e-01
-1.30530536e+00 -4.52977121e-02 5.49664736e-01 4.90124613e-01
1.89435586e-01 -1.74811840e-01 -1.85409009e-01 8.82664680e-01
-4.12396789e-02 3.54522645e-01 9.24226165e-01 8.67469981e-02
2.98298180e-01 8.64570975e-01 -1.52990088e-01 -9.63818669e-01
-8.05929005e-01 -8.24408829e-01 -9.85846221e-01 9.36616585e-02
-2.08246171e-01 -3.96451920e-01 -8.39763224e-01 1.34431326e+00
-4.49072719e-01 4.21464205e-01 4.86326516e-02 5.53225815e-01
7.03073859e-01 3.57830286e-01 -2.30148435e-02 -1.95192710e-01
1.22859037e+00 -1.93028942e-01 -5.86280048e-01 -4.48945135e-01
5.09568751e-01 -3.26344997e-01 3.03431839e-01 9.06828940e-01
-7.80402899e-01 -2.77167320e-01 -1.21839786e+00 7.50134110e-01
-4.93623942e-01 2.66726255e-01 4.84732091e-01 6.19085550e-01
-1.11706746e+00 3.87390018e-01 -8.17180276e-01 -2.76030362e-01
6.21688068e-01 1.03538716e+00 -7.29275048e-01 2.42528290e-01
-1.10987473e+00 1.18826330e+00 5.34666061e-01 2.67401248e-01
-8.43910813e-01 -5.38429022e-01 -3.85464877e-01 -3.90215628e-02
-4.12967980e-01 -3.26651961e-01 5.66433609e-01 -7.63124049e-01
-9.17691708e-01 7.35164881e-01 2.99667139e-02 -4.70429122e-01
2.41999552e-01 7.60211200e-02 -7.49015212e-01 3.60620841e-02
-1.39964689e-02 3.69999230e-01 6.50821209e-01 -6.12554252e-01
-7.46265888e-01 -7.85235763e-01 -4.86478478e-01 -4.30778898e-02
-2.95031488e-01 2.36245543e-01 3.59286934e-01 -5.87363958e-01
6.98482990e-02 -7.75083840e-01 2.49639541e-01 -8.05877268e-01
-2.06624717e-02 -3.64031702e-01 7.30634212e-01 -9.87625599e-01
1.29503024e+00 -2.11192131e+00 -1.06764518e-01 3.62759918e-01
1.76107600e-01 2.92684913e-01 1.33405760e-01 1.50977761e-01
-5.52432775e-01 -2.00456634e-01 -4.22895581e-01 3.52813631e-01
-2.21532419e-01 -3.52513939e-01 -7.42409676e-02 5.33780038e-01
2.24890292e-01 8.12529385e-01 -5.54623485e-01 3.30085069e-01
3.11637610e-01 7.80756712e-01 3.04680262e-02 4.85303968e-01
5.88202417e-01 6.93074107e-01 -1.50013417e-01 5.26954830e-01
6.69484615e-01 -9.86305177e-02 -1.23691954e-01 3.26801501e-02
-9.31553021e-02 3.50649148e-01 -1.04590845e+00 1.49901021e+00
-3.24543178e-01 7.58627832e-01 -2.67739534e-01 -1.35603344e+00
1.14949059e+00 8.72009039e-01 3.44965398e-01 -7.65734136e-01
2.98857689e-01 6.88153505e-01 3.90603989e-01 -6.56442761e-01
-4.95239258e-01 1.54823512e-02 2.39437729e-01 3.57497841e-01
1.43146455e-01 3.58511172e-02 -1.07531056e-01 -2.14357480e-01
1.39794827e+00 -3.09288412e-01 3.74939919e-01 -6.07277930e-01
5.25358081e-01 -4.90364641e-01 3.51167142e-01 5.02785504e-01
6.46191388e-02 3.10500443e-01 9.14594010e-02 -5.85992754e-01
-5.89934528e-01 -6.88034594e-01 -8.68365824e-01 4.48687434e-01
-1.90001592e-01 6.81824312e-02 -8.40057135e-01 -3.20323318e-01
-1.64047748e-01 4.92535144e-01 -6.65308595e-01 -1.86540142e-01
-1.02988236e-01 -1.16684985e+00 4.51858848e-01 5.96221507e-01
6.34673417e-01 -1.36867380e+00 -1.10046434e+00 6.27481461e-01
3.93992215e-02 -8.02179098e-01 7.32574165e-02 7.08443105e-01
-7.26731002e-01 -1.01173496e+00 -9.17998612e-01 -1.01191306e+00
4.80300725e-01 -3.45418990e-01 8.32163513e-01 -1.44570902e-01
-5.31535923e-01 2.21676692e-01 -2.37275004e-01 -7.17770338e-01
3.92064713e-02 6.14044257e-02 3.02861392e-01 8.91507044e-02
1.03407264e+00 -9.84619915e-01 -8.70940030e-01 2.28470251e-01
-6.28699839e-01 -2.02899545e-01 8.99667978e-01 8.41690063e-01
3.64435524e-01 2.06777543e-01 1.16250110e+00 -2.53483117e-01
1.00968719e+00 -7.39397049e-01 -2.80076474e-01 2.16537163e-01
-5.99685431e-01 -3.84753436e-01 3.85644883e-01 -3.64394814e-01
-4.76660430e-01 -2.48850405e-01 -2.10834354e-01 -6.73825592e-02
-7.18505979e-01 4.79535639e-01 1.12168910e-02 -2.97382861e-01
3.62162590e-01 5.96910715e-01 -4.01364446e-01 -1.81188047e-01
-8.46278369e-01 1.32402408e+00 5.25268674e-01 7.49006271e-02
-5.67926280e-02 -6.93613710e-03 -1.67614445e-01 -8.09664845e-01
-2.24879876e-01 -5.12315154e-01 -6.40028894e-01 -6.65438473e-02
8.10086608e-01 -1.09230542e+00 -4.45680588e-01 7.52737105e-01
-1.11758065e+00 -7.07710255e-03 5.74286819e-01 9.11266506e-01
-2.24363789e-01 -1.87043041e-01 -1.35408714e-01 -8.17165613e-01
-8.03644419e-01 -1.37402368e+00 8.40431869e-01 2.28964564e-04
-5.54651260e-01 -8.91385913e-01 3.19341980e-02 -7.65403062e-02
5.46542704e-01 3.50617528e-01 8.73482406e-01 -1.24125004e+00
-9.99832153e-02 -5.98269224e-01 -5.26101477e-02 4.72614288e-01
3.92032355e-01 -5.12728333e-01 -1.12440646e+00 -5.84329188e-01
2.82374203e-01 -8.28918591e-02 5.03272414e-01 6.72960520e-01
1.10652614e+00 1.86767846e-01 -5.83926857e-01 7.21885204e-01
1.28925800e+00 9.18065906e-01 6.26650691e-01 4.66463745e-01
1.01078704e-01 3.05168778e-01 -2.48694628e-01 4.15342927e-01
1.44763365e-01 3.04594576e-01 3.73353571e-01 6.36293516e-02
2.29420602e-01 4.08049196e-01 2.91093349e-01 7.31011331e-01
-1.79833174e-01 -1.04322761e-01 -1.23124027e+00 7.76074290e-01
-1.15070784e+00 -7.68451095e-01 6.80849189e-03 2.25033832e+00
4.61017281e-01 8.75567272e-02 -5.81104308e-02 5.11537373e-01
7.56568611e-01 -5.42864442e-01 -5.24023831e-01 -4.14614707e-01
-2.28086933e-01 7.88852096e-01 1.88644186e-01 -2.47082919e-01
-9.46457028e-01 1.44060045e-01 6.50516176e+00 4.14581239e-01
-1.57157862e+00 7.62143657e-02 5.63920975e-01 -8.29986036e-02
3.21453214e-01 -5.21347284e-01 -5.73052227e-01 8.47316086e-01
1.26675761e+00 -1.40790910e-01 5.31881988e-01 2.98213720e-01
2.15498194e-01 -2.11626455e-01 -1.11097586e+00 1.35739160e+00
2.02626929e-01 -1.03202105e+00 -2.46113598e-01 1.42090335e-01
6.59548759e-01 3.31409156e-01 2.24126503e-02 9.36382785e-02
-5.34890354e-01 -1.57343090e+00 2.32055008e-01 6.23058379e-01
1.07971478e+00 -9.86552358e-01 1.16996586e+00 4.80373770e-01
-9.40776706e-01 -3.95425558e-01 -1.70525342e-01 -1.16124392e-01
-1.98980525e-01 4.48776186e-01 -1.03223705e+00 2.41134703e-01
1.08651233e+00 7.52012432e-01 -5.64193130e-01 1.46184862e+00
-1.81246519e-01 5.00059783e-01 -3.02481085e-01 -1.13594890e-01
1.12990826e-01 -3.87065075e-02 4.58332628e-01 1.25018275e+00
1.00470436e+00 2.02214792e-01 -4.26216602e-01 7.59877443e-01
1.22176439e-01 2.13169947e-01 -7.81423569e-01 9.54006158e-04
2.34985098e-01 1.16079497e+00 -6.70891941e-01 3.84021848e-02
-5.51701605e-01 7.99866378e-01 6.25887737e-02 1.31945372e-01
-4.16450232e-01 -6.96705401e-01 1.08457975e-01 -1.26722762e-02
7.51263797e-02 3.70410115e-01 -4.87530261e-01 -9.47917283e-01
2.55581558e-01 -9.90474999e-01 1.38412952e-01 -8.34100664e-01
-1.28145182e+00 1.22795844e+00 -1.66330934e-01 -1.24996972e+00
-1.50330022e-01 -8.55150342e-01 -9.18292940e-01 1.19080865e+00
-1.34108090e+00 -7.22076714e-01 -3.29542786e-01 8.84392321e-01
3.37827891e-01 -5.13433456e-01 1.23434019e+00 3.22189182e-01
-4.64114577e-01 4.02557343e-01 4.70135733e-02 1.40916929e-01
6.18564785e-01 -1.03946912e+00 -9.61840451e-02 5.92775345e-01
-2.14157954e-01 6.91306889e-01 3.61902773e-01 -5.73622882e-01
-1.00010824e+00 -9.74325061e-01 8.57891738e-01 -2.32500672e-01
4.77337837e-01 -2.39687428e-01 -1.02186608e+00 5.41014552e-01
5.14446914e-01 -2.40724370e-01 9.23130870e-01 -2.99779505e-01
1.99704051e-01 -1.16363041e-01 -1.25568175e+00 -5.73125742e-02
7.37100601e-01 -4.66131657e-01 -1.17688441e+00 2.88483769e-01
-4.54472452e-01 -7.10145161e-02 -1.13757670e+00 7.79430926e-01
7.51616538e-01 -1.07896757e+00 5.10909557e-01 -1.07207417e-01
1.01270445e-01 1.29897848e-01 1.29979372e-01 -1.84874558e+00
-2.80163854e-01 -4.56021637e-01 1.69695407e-01 7.47087061e-01
5.77405691e-01 -1.02703679e+00 3.13461304e-01 6.06135249e-01
-1.31446153e-01 -9.61368620e-01 -9.52664554e-01 -6.70197785e-01
1.60049856e-01 -4.43013936e-01 8.90105128e-01 6.32115662e-01
3.63167763e-01 2.46515304e-01 -6.38674805e-03 2.77268618e-01
4.03165787e-01 -1.63820416e-01 8.21253061e-02 -1.54180670e+00
1.94076732e-01 -3.62893075e-01 -9.12964940e-01 -2.72338063e-01
1.68873489e-01 -1.19104588e+00 -2.31803387e-01 -1.57638812e+00
3.15358460e-01 -2.76911259e-01 -8.28979313e-01 4.66631800e-01
1.08335182e-01 3.33811820e-01 -4.78907675e-01 -7.89452437e-03
2.93518156e-02 7.53754601e-02 7.79708445e-01 -1.06708162e-01
-3.22422177e-01 1.27175882e-01 -4.53096122e-01 6.71951771e-01
9.92920399e-01 -5.34056485e-01 -3.52380782e-01 -3.66021395e-01
6.67899698e-02 1.04560696e-01 3.24263394e-01 -1.42182791e+00
4.45917934e-01 5.35424232e-01 1.01476824e+00 -6.27516866e-01
1.04738027e-01 -8.27309787e-01 4.88709688e-01 4.00104642e-01
1.24593638e-03 3.03172231e-01 5.29329956e-01 1.61528483e-01
-3.86695474e-01 -1.85574796e-02 6.52357042e-01 -5.12160733e-02
-6.14489794e-01 3.93068284e-01 -6.70323193e-01 -2.98118353e-01
1.28482819e+00 -6.20529950e-01 6.27641305e-02 -1.54301422e-02
-8.24983001e-01 -1.00045592e-01 -4.74888533e-02 3.60986233e-01
8.60106170e-01 -1.11632252e+00 -8.86858881e-01 8.62361193e-01
1.72089398e-01 -3.53416115e-01 2.19214901e-01 1.37995851e+00
-2.72828311e-01 6.82261884e-01 -7.03274965e-01 -6.67895496e-01
-9.64027882e-01 1.31882966e-01 5.26183426e-01 8.22240934e-02
-8.53604555e-01 1.04632950e+00 8.92084986e-02 -8.46502930e-02
2.64839232e-01 -1.33007571e-01 -5.80086768e-01 1.84415817e-01
8.90507579e-01 1.62159473e-01 8.53437781e-01 -5.25202513e-01
-6.85363591e-01 4.47896689e-01 1.23747028e-02 2.85954654e-01
1.83429098e+00 2.13842869e-01 -3.72163445e-01 2.46920452e-01
1.12365365e+00 -4.19762969e-01 -7.17665017e-01 3.16197604e-01
6.31416636e-03 -1.15797110e-01 4.48820651e-01 -1.27015412e+00
-1.13614547e+00 1.05513155e+00 1.21870732e+00 2.25940987e-01
1.49806869e+00 -2.22719505e-01 5.27512193e-01 3.09656978e-01
7.46478081e-01 -8.74942005e-01 -5.29231906e-01 2.46308118e-01
9.50805545e-01 -9.39117551e-01 -3.81284922e-01 4.10560995e-01
-3.97207528e-01 1.34185553e+00 1.69636577e-01 -5.23060799e-01
8.56755376e-01 5.71028411e-01 -1.68345779e-01 -3.45867574e-01
-6.91126645e-01 4.20776755e-01 4.55614775e-01 6.92488074e-01
5.78440666e-01 2.80318081e-01 -4.46457714e-01 9.74399388e-01
-2.66461790e-01 2.72057086e-01 2.55562633e-01 6.70680821e-01
-2.75875032e-01 -7.98629642e-01 -2.98052818e-01 1.11934710e+00
-8.48730743e-01 -2.43192837e-01 -2.95302778e-01 5.86344123e-01
5.07376432e-01 1.08272457e+00 1.95563987e-01 -3.53190064e-01
3.26078773e-01 3.93763304e-01 5.55017889e-01 -5.55180013e-01
-8.13466549e-01 2.50986423e-02 -3.16021293e-01 -2.84284800e-01
-3.85575354e-01 -6.19418621e-01 -9.97561693e-01 2.50745445e-01
-5.21608889e-01 1.80268362e-01 7.32048869e-01 1.17655790e+00
6.00299478e-01 7.19852149e-01 6.00350797e-01 -7.01434672e-01
-3.47694866e-02 -1.33290124e+00 -9.97407377e-01 1.14297666e-01
4.65126008e-01 -9.59991336e-01 -6.29062116e-01 -9.85123664e-02] | [13.212512969970703, 3.467357873916626] |
8f73a0cd-e4dc-4c95-948e-95215f504455 | glen-general-purpose-event-detection-for | 2303.09093 | null | https://arxiv.org/abs/2303.09093v2 | https://arxiv.org/pdf/2303.09093v2.pdf | GLEN: General-Purpose Event Detection for Thousands of Types | The development of event extraction systems has been hindered by the absence of wide-coverage, large-scale datasets. To make event extraction systems more accessible, we build a general-purpose event detection dataset GLEN, which covers 3,465 different event types, making it over 20x larger in ontology than any current dataset. GLEN is created by utilizing the DWD Overlay, which provides a mapping between Wikidata Qnodes and PropBank rolesets. This enables us to use the abundant existing annotation for PropBank as distant supervision. In addition, we also propose a new multi-stage event detection model specifically designed to handle the large ontology size and partial labels in GLEN. We show that our model exhibits superior performance (~10% F1 gain) compared to both conventional classification baselines and newer definition-based models. Finally, we perform error analysis and show that label noise is still the largest challenge for improving performance. | ['Jiawei Han', 'Heng Ji', 'Martha Palmer', 'Kathryn Conger', 'Sha Li', 'Qiusi Zhan'] | 2023-03-16 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [-3.77152450e-02 4.57647800e-01 -5.09314597e-01 -3.77768427e-01
-1.06785083e+00 -6.39914930e-01 6.13012671e-01 7.22973049e-01
-7.18049169e-01 9.38588381e-01 7.51184404e-01 -8.35819021e-02
-9.56596136e-02 -7.82055020e-01 -6.39398634e-01 -8.96891654e-02
-1.76782966e-01 5.36173582e-01 6.99214220e-01 -5.56320176e-02
-1.29486129e-01 2.29025885e-01 -1.34965312e+00 3.94253224e-01
4.38797504e-01 8.55346441e-01 -1.97199225e-01 1.89864948e-01
5.88967092e-02 1.09692526e+00 -7.05610514e-01 -5.58974564e-01
3.22868489e-02 -3.84721994e-01 -1.09074605e+00 -5.19421935e-01
1.44719556e-01 -2.02313989e-01 -4.49666709e-01 6.10341012e-01
6.97627902e-01 1.40314251e-01 3.87096345e-01 -1.32325983e+00
-2.07134470e-01 1.23863387e+00 -2.46324033e-01 7.46771872e-01
5.27202129e-01 -3.73185366e-01 1.36143827e+00 -7.16710746e-01
1.20774472e+00 8.47989798e-01 8.91665220e-01 2.60031611e-01
-1.00623810e+00 -8.39548707e-01 1.14966549e-01 2.54752129e-01
-1.40140188e+00 -5.73101223e-01 2.63777107e-01 -2.54187226e-01
1.45996869e+00 2.26580858e-01 4.46602643e-01 1.21560931e+00
3.14894691e-02 6.04470074e-01 5.68245113e-01 -6.07933044e-01
1.20508507e-01 -1.46501929e-01 2.77330548e-01 6.94958985e-01
4.22048777e-01 -1.49450824e-01 -8.34609091e-01 -4.28693444e-01
4.03202295e-01 -1.94004983e-01 -2.08256990e-01 5.54304495e-02
-1.30883741e+00 5.60361266e-01 1.15657598e-01 1.77084967e-01
-2.50983417e-01 1.00641951e-01 6.42812967e-01 9.85143930e-02
4.62985963e-01 6.15139544e-01 -8.69141638e-01 -4.11200643e-01
-6.62487388e-01 5.12800872e-01 1.14507449e+00 1.00490689e+00
3.56217623e-01 -3.11441630e-01 -2.68810373e-02 7.70182371e-01
1.97433263e-01 -4.65679169e-02 2.64104098e-01 -1.08657503e+00
7.26262450e-01 7.32237101e-01 1.02804959e-01 -5.97212315e-01
-9.56944227e-01 -2.81604439e-01 -7.61162192e-02 -3.14853579e-01
6.45826817e-01 -2.26182640e-01 -5.70477545e-01 1.93039513e+00
4.33353037e-01 2.47913092e-01 6.86518326e-02 4.85254318e-01
9.77721810e-01 2.83400863e-01 3.91449302e-01 -1.15600340e-01
1.69090807e+00 -6.86865807e-01 -8.15472424e-01 -1.69708237e-01
9.39633608e-01 -4.27048266e-01 5.63886166e-01 2.15834111e-01
-9.86558497e-01 5.94127327e-02 -1.10612464e+00 -2.88111836e-01
-4.66428071e-01 -2.83741146e-01 9.60549772e-01 3.43493879e-01
-3.72122467e-01 3.22949171e-01 -1.09303463e+00 -7.43264437e-01
4.16773319e-01 -6.69279769e-02 -5.28754711e-01 6.45940900e-02
-1.52313447e+00 1.14404118e+00 9.56720471e-01 -5.57232380e-01
-6.56110168e-01 -1.09242499e+00 -9.94858682e-01 1.89763471e-01
7.77665317e-01 -2.91574806e-01 1.35393262e+00 -1.94990952e-02
-8.08643878e-01 7.33219266e-01 2.23138854e-02 -6.33405626e-01
1.58899352e-01 -1.41520813e-01 -9.77274239e-01 2.64697582e-01
5.03975868e-01 4.87830997e-01 -4.54004481e-02 -6.28906071e-01
-1.12888813e+00 2.05201879e-02 1.46049440e-01 1.33552969e-01
-4.49304521e-01 5.96290469e-01 -5.52678525e-01 -7.91139007e-01
-1.15794994e-01 -5.89838624e-01 -1.48813173e-01 -3.44897598e-01
-3.20244998e-01 -4.08828735e-01 4.27992821e-01 -7.31042624e-01
1.65069652e+00 -2.00164175e+00 -2.89210558e-01 5.00210635e-02
1.70422792e-01 -2.90132463e-01 1.75650850e-01 5.75545430e-01
-2.58407503e-01 9.37817022e-02 5.29043674e-02 -1.70864895e-01
1.80164725e-01 3.57026517e-01 -1.99640572e-01 9.33612958e-02
4.05408621e-01 7.19596386e-01 -1.07404625e+00 -7.61197448e-01
-3.50021034e-01 2.60525830e-02 -7.26250589e-01 -2.47410476e-01
-1.22738630e-01 6.62633777e-03 -3.69249672e-01 8.56547356e-01
7.75947794e-02 -4.28464204e-01 5.89617372e-01 -2.21434250e-01
-1.46297902e-01 1.04125321e+00 -1.29260647e+00 1.89875460e+00
2.89829238e-03 4.13343370e-01 -3.28298360e-01 -9.31523979e-01
5.52132010e-01 6.93948805e-01 8.38721931e-01 -4.52696562e-01
2.70048864e-02 2.56073326e-01 -7.19873384e-02 -2.72685975e-01
5.59694350e-01 -1.10622615e-01 -6.06883287e-01 5.58644474e-01
3.70287746e-01 4.44603086e-01 6.74564958e-01 5.16744375e-01
1.78636611e+00 2.53011677e-02 6.51530981e-01 -1.44139081e-01
-1.02457322e-01 4.32702899e-01 1.37667871e+00 6.94476902e-01
-6.86321631e-02 4.63870436e-01 6.72863543e-01 -3.78474563e-01
-6.97830379e-01 -1.02018619e+00 -6.05561018e-01 1.13768804e+00
-1.93453595e-01 -1.00786245e+00 -4.03720289e-01 -1.01848972e+00
7.87137598e-02 6.91368163e-01 -3.99106175e-01 5.84733635e-02
-7.72737503e-01 -1.31015348e+00 1.33816993e+00 8.04415405e-01
3.66142243e-01 -1.00732100e+00 -6.19858205e-01 7.20405042e-01
-6.85962141e-01 -1.45799625e+00 -2.46363088e-01 5.75959921e-01
-4.03267562e-01 -1.31769657e+00 2.67939586e-02 -6.21676505e-01
1.31146228e-02 -4.70226794e-01 1.57492507e+00 -3.94063205e-01
-3.62444520e-01 -9.26826335e-03 -5.89945674e-01 -7.75156200e-01
-2.61996388e-01 4.30439472e-01 1.79326143e-02 -6.22127116e-01
8.75156224e-01 -4.14583355e-01 -2.19329834e-01 3.18564028e-01
-8.31450939e-01 -2.72086769e-01 2.02021688e-01 6.55970871e-01
4.40336734e-01 1.99942127e-01 1.16386461e+00 -1.04971445e+00
2.79455394e-01 -9.74030256e-01 -4.88502949e-01 1.42948642e-01
-7.64390945e-01 -2.15927541e-01 9.28729922e-02 -2.48638302e-01
-1.28895295e+00 -9.87147167e-02 -1.66517764e-01 3.68840575e-01
-1.05645992e-01 9.39604878e-01 -1.80861458e-01 5.96263647e-01
9.26489353e-01 -4.85099822e-01 -3.38179976e-01 -8.01186204e-01
1.73870832e-01 6.09780133e-01 7.84193695e-01 -6.11198962e-01
3.64414692e-01 4.01350349e-01 -3.52042198e-01 -3.29236865e-01
-1.15626371e+00 -4.23590183e-01 -3.92461419e-01 1.65863156e-01
8.56738925e-01 -1.29134345e+00 -4.86663789e-01 5.93667217e-02
-8.93157303e-01 -3.36145163e-01 -5.44829607e-01 6.47211254e-01
-7.47294575e-02 1.33781090e-01 -1.07094204e+00 -3.63305986e-01
-1.31581858e-01 -6.33294404e-01 7.35936344e-01 1.78709012e-02
-5.16896367e-01 -8.85358632e-01 2.39807621e-01 2.93238342e-01
-7.57201016e-02 5.38989782e-01 8.25808167e-01 -1.18611455e+00
-3.23034018e-01 -1.80751175e-01 -2.12105781e-01 -2.04767942e-01
9.03354120e-03 -2.18889147e-01 -6.36655092e-01 1.58037916e-02
-5.29053867e-01 -3.82536948e-01 8.49673688e-01 1.28271906e-02
7.02573180e-01 1.33772120e-01 -7.48177826e-01 3.64354104e-01
1.14839160e+00 2.26653188e-01 4.80666667e-01 8.32593918e-01
4.84714806e-01 3.89485121e-01 7.16366827e-01 7.17400491e-01
9.19424117e-01 7.39344656e-01 9.05193612e-02 -1.47837563e-04
-1.97054967e-01 -3.46915901e-01 1.78426877e-01 6.64510787e-01
4.72287498e-02 -4.88460183e-01 -1.15764427e+00 8.38407934e-01
-1.94357133e+00 -1.08676255e+00 -2.33373284e-01 1.79677320e+00
1.18551970e+00 4.34145957e-01 1.07326500e-01 1.43799797e-01
4.48191971e-01 3.75880301e-03 -3.94103587e-01 7.88433403e-02
-3.40089828e-01 5.26557028e-01 7.14937925e-01 6.11202493e-02
-1.31005871e+00 8.98068070e-01 7.37043810e+00 5.74312687e-01
-4.08948362e-01 4.42652881e-01 -9.65775829e-03 -4.53822374e-01
-4.38279612e-03 1.48845568e-01 -1.17104399e+00 5.56189358e-01
1.36063910e+00 -1.81744412e-01 -3.22760567e-02 5.44566453e-01
2.88575329e-02 -1.48363918e-01 -1.28596425e+00 6.94248855e-01
-2.26986445e-02 -1.70544302e+00 -3.73929381e-01 1.92191005e-01
5.06159723e-01 4.75669563e-01 -7.55860686e-01 6.00382149e-01
8.38029683e-01 -5.76753497e-01 9.72263873e-01 2.51278102e-01
7.36039281e-01 -6.85810089e-01 7.91481793e-01 2.12078735e-01
-1.39416122e+00 -1.92693204e-01 -7.70093948e-02 -9.75655541e-02
6.57807291e-01 7.97855675e-01 -8.15727115e-01 8.33156943e-01
9.73471165e-01 9.97744203e-01 -4.51491892e-01 1.05178821e+00
-2.11530343e-01 8.49793494e-01 -6.80656135e-01 3.97404939e-01
-1.35097787e-01 4.59264398e-01 4.57773894e-01 1.42052698e+00
1.67787507e-01 3.02754223e-01 4.80926394e-01 5.41395664e-01
-4.99731660e-01 -3.11409328e-02 -4.19288009e-01 -7.96869397e-02
9.27346408e-01 1.11749160e+00 -7.00383604e-01 -5.31578243e-01
-6.58516467e-01 5.03609836e-01 4.85194027e-01 3.36073078e-02
-1.07190025e+00 -6.62821412e-01 7.19227731e-01 -9.42442715e-02
1.81144014e-01 9.97683555e-02 6.89267516e-02 -1.44918537e+00
-4.54724021e-02 -7.86512375e-01 1.27738237e+00 -5.05467653e-01
-1.22710669e+00 3.39606881e-01 3.67354691e-01 -8.38094294e-01
-3.39569479e-01 -3.39089394e-01 -2.14306161e-01 4.98413980e-01
-1.32727385e+00 -9.64650929e-01 -1.44405901e-01 4.24929470e-01
2.66877949e-01 -4.40296084e-02 9.45326447e-01 1.06307042e+00
-8.40573013e-01 6.99809134e-01 -2.97567546e-01 5.37144899e-01
1.14920723e+00 -1.35557115e+00 4.93000507e-01 8.90488029e-01
1.06645688e-01 4.66843694e-01 4.98191267e-01 -8.96944463e-01
-8.45564544e-01 -1.13541341e+00 1.25294077e+00 -9.30778027e-01
9.47661996e-01 -5.07622957e-01 -8.39275241e-01 1.28489423e+00
-1.05384201e-01 -4.09959145e-02 9.61856186e-01 5.86663425e-01
-6.27817512e-01 1.57348573e-01 -1.04104948e+00 2.61195570e-01
1.49783790e+00 -3.65832180e-01 -1.04180121e+00 3.29700947e-01
7.52928615e-01 -5.39900482e-01 -1.48957157e+00 5.21530151e-01
4.82939810e-01 -4.02988166e-01 8.62627327e-01 -8.48134875e-01
1.73993453e-01 -2.84172922e-01 -2.59483904e-01 -1.00885558e+00
-2.41496146e-01 -3.78646076e-01 -3.00959557e-01 1.77348173e+00
8.70394349e-01 -7.44256735e-01 5.87014675e-01 7.00546861e-01
-3.95346731e-01 -1.43009678e-01 -9.39067185e-01 -8.00774813e-01
-1.80772603e-01 -7.48068571e-01 7.62400389e-01 1.44907725e+00
6.83313131e-01 4.57147807e-01 -1.20317146e-01 3.71582925e-01
4.73759323e-01 -8.97541568e-02 3.54586869e-01 -1.49493444e+00
-3.20087820e-01 -4.67092656e-02 -3.30942929e-01 -6.41792238e-01
1.12406254e-01 -1.23203623e+00 -7.39574358e-02 -1.61821520e+00
3.32194537e-01 -8.14010561e-01 -6.28828287e-01 1.14081883e+00
-1.30308419e-01 3.07412177e-01 -1.53618872e-01 1.58422366e-01
-8.29558015e-01 1.59634724e-01 4.26349908e-01 2.52702594e-01
-1.74070656e-01 -6.24578834e-01 -8.49903166e-01 8.78385723e-01
6.18231297e-01 -9.43796873e-01 -1.14096768e-01 -5.48290431e-01
4.01605338e-01 5.20050265e-02 1.01526573e-01 -1.00211167e+00
2.13426128e-01 -1.18201924e-02 2.35131264e-01 -6.32211208e-01
1.01983964e-01 -4.54584688e-01 4.20684248e-01 6.48871362e-02
-4.88239527e-01 1.24849588e-01 1.54686093e-01 6.73274934e-01
-2.33716890e-01 -2.39386871e-01 3.32709283e-01 -3.91227268e-02
-7.57988036e-01 1.87296588e-02 -4.22829539e-01 7.10820794e-01
7.62601197e-01 3.66709679e-01 -6.15233004e-01 9.81825739e-02
-8.08712542e-01 2.86433131e-01 2.56902248e-01 4.40373272e-01
-8.26688856e-02 -1.36620629e+00 -7.14865267e-01 -2.67468929e-01
4.72288549e-01 1.43850995e-02 -2.34833181e-01 8.03773284e-01
-3.28843504e-01 1.33482769e-01 -1.93092264e-02 -2.73226760e-02
-9.80689108e-01 2.64539063e-01 -3.78621928e-02 -6.10470057e-01
-1.02335596e+00 7.56333649e-01 -2.97600210e-01 -4.60890889e-01
3.02456915e-01 -4.61253971e-01 -2.03692704e-01 2.92942077e-01
7.66268432e-01 4.67915148e-01 2.62896895e-01 -2.01480210e-01
-6.11084044e-01 -3.11658084e-01 -8.71760398e-02 -2.47034684e-01
1.57737875e+00 6.34199381e-02 -5.61993867e-02 3.11454594e-01
7.65017450e-01 1.61225498e-01 -9.35448945e-01 -4.07697946e-01
8.19354236e-01 1.41815040e-02 -4.45879921e-02 -9.97396529e-01
-7.00914681e-01 6.61678091e-02 1.06836498e-01 3.27358872e-01
8.20405185e-01 2.84937114e-01 9.83838797e-01 3.55289608e-01
5.29798090e-01 -1.15564466e+00 -3.17569375e-01 6.35530889e-01
3.71245831e-01 -1.06268013e+00 7.90340826e-02 -5.75151622e-01
-4.67778951e-01 6.25077546e-01 4.98654574e-01 2.11663440e-01
7.04174459e-01 6.50949359e-01 -1.07702278e-01 -5.07251501e-01
-1.16009665e+00 -4.34810996e-01 3.49098481e-02 3.07895690e-01
4.48588312e-01 -1.89221412e-01 -4.37090904e-01 1.04035401e+00
-1.78692520e-01 1.85199142e-01 4.42496628e-01 1.22298133e+00
-3.48789454e-01 -1.27963161e+00 -9.26121697e-02 6.57187998e-01
-1.05252397e+00 -2.46278301e-01 5.06963693e-02 1.08945727e+00
3.17280650e-01 9.92636025e-01 1.59569666e-01 -5.42032979e-02
6.46788716e-01 4.96050209e-01 2.26055294e-01 -9.04551923e-01
-6.82456791e-01 -8.94161016e-02 1.01760256e+00 -6.61049604e-01
-4.91011739e-01 -9.76049244e-01 -1.61813962e+00 -2.93725487e-02
-3.39245707e-01 2.72984147e-01 3.60112041e-01 1.04450023e+00
5.30076802e-01 7.40458906e-01 -1.38760269e-01 -3.23903978e-01
-1.83274373e-01 -9.60057199e-01 -5.19055188e-01 4.62309748e-01
-2.83666074e-01 -1.02891731e+00 -2.43988410e-01 2.45933533e-01] | [9.136513710021973, 9.209160804748535] |
43c5f0ab-2606-485e-b7ee-98ae2b814562 | improving-fairness-in-adaptive-social | 2302.09298 | null | https://arxiv.org/abs/2302.09298v2 | https://arxiv.org/pdf/2302.09298v2.pdf | Improving Fairness in Adaptive Social Exergames via Shapley Bandits | Algorithmic fairness is an essential requirement as AI becomes integrated in society. In the case of social applications where AI distributes resources, algorithms often must make decisions that will benefit a subset of users, sometimes repeatedly or exclusively, while attempting to maximize specific outcomes. How should we design such systems to serve users more fairly? This paper explores this question in the case where a group of users works toward a shared goal in a social exergame called Step Heroes. We identify adverse outcomes in traditional multi-armed bandits (MABs) and formalize the Greedy Bandit Problem. We then propose a solution based on a new type of fairness-aware multi-armed bandit, Shapley Bandits. It uses the Shapley Value for increasing overall player participation and intervention adherence rather than the maximization of total group output, which is traditionally achieved by favoring only high-performing participants. We evaluate our approach via a user study (n=46). Our results indicate that our Shapley Bandits effectively mediates the Greedy Bandit Problem and achieves better user retention and motivation across the participants. | ['Jichen Zhu', 'Shahin Jabbari', 'Danielle Arigo', 'Santiago Ontañón', 'Diane H. Dallal', 'Thomas B. Fox', 'Jennifer Villareale', 'Robert C. Gray'] | 2023-02-18 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [ 1.20485704e-02 4.98477787e-01 -8.59704852e-01 -2.97505260e-01
-5.24706721e-01 -3.81685436e-01 -4.82167564e-02 -2.69867219e-02
-8.48313212e-01 1.34380770e+00 2.95305282e-01 -5.62521100e-01
-7.88121939e-01 -6.85697079e-01 -4.48931366e-01 -4.43028688e-01
1.62388012e-01 7.55093873e-01 -2.62545735e-01 -3.04037303e-01
4.25221413e-01 2.34767452e-01 -1.46167994e+00 2.14723140e-01
1.20024335e+00 5.95840096e-01 -8.43539834e-02 7.09554434e-01
3.93095985e-02 9.17732179e-01 -2.99494207e-01 -7.68443644e-01
6.94180012e-01 -8.42192888e-01 -8.83137882e-01 -9.50452834e-02
3.28036994e-01 -7.63542295e-01 6.06856607e-02 7.08773553e-01
4.52377319e-01 4.71181273e-01 2.98475444e-01 -1.67080867e+00
-5.74144602e-01 1.09650564e+00 -9.98731136e-01 3.40364613e-02
2.57792145e-01 1.28569588e-01 1.48850274e+00 3.50527167e-01
3.09399933e-01 1.06982076e+00 5.68785965e-01 7.65214980e-01
-1.70075560e+00 -7.19545007e-01 3.91375981e-02 1.67732835e-01
-7.03331411e-01 -5.07242143e-01 6.62223816e-01 -1.93818793e-01
5.54027081e-01 1.10389006e+00 1.28355861e+00 3.11384797e-01
-4.77032989e-01 1.00439799e+00 9.51500475e-01 -6.02413952e-01
4.08060670e-01 1.53653175e-01 5.10141790e-01 1.21587135e-01
8.69553328e-01 5.63027523e-02 -4.85123396e-01 -5.79319894e-01
5.47566414e-01 4.04082574e-02 4.53814585e-03 -5.98321557e-01
-7.91492999e-01 9.17054653e-01 2.18719661e-01 1.42300189e-01
-9.74430919e-01 6.59554183e-01 9.57727283e-02 3.08220178e-01
4.75302637e-01 7.11099982e-01 -1.27473429e-01 -2.61311591e-01
-1.25317657e+00 7.19528615e-01 9.39612806e-01 6.14689410e-01
7.34597921e-01 -3.28734070e-01 -8.09824705e-01 8.14769447e-01
6.14963733e-02 1.95990205e-01 8.04683045e-02 -1.55606508e+00
3.83371681e-01 5.70357859e-01 8.52092028e-01 -5.40014863e-01
-3.59391242e-01 -3.82071674e-01 -4.06425416e-01 3.10950100e-01
6.97557926e-01 -5.46132028e-01 -3.31140369e-01 1.82907152e+00
2.47014329e-01 -3.09795648e-01 -4.71969545e-01 9.54779446e-01
2.39641462e-02 3.17878723e-01 4.03338701e-01 -7.98295736e-01
9.25730526e-01 -8.71837616e-01 -8.70162427e-01 -2.55252391e-01
6.40035510e-01 -2.36110821e-01 1.03192687e+00 2.70819336e-01
-1.63207197e+00 1.80916503e-01 -4.35175478e-01 3.09976190e-01
2.55494356e-01 -3.62372816e-01 8.24749708e-01 1.66750765e+00
-9.09291983e-01 6.14514232e-01 -2.20070273e-01 -3.70261431e-01
8.93808603e-01 6.67594850e-01 1.67195037e-01 3.04727495e-01
-8.23884070e-01 7.13532031e-01 8.06126744e-03 -2.68668413e-01
-2.87574351e-01 -1.07273722e+00 -9.41140503e-02 6.80231512e-01
7.49065518e-01 -9.88208711e-01 1.50461709e+00 -1.38316596e+00
-1.48558867e+00 9.91204560e-01 2.97625750e-01 -4.55532074e-01
9.23464298e-01 -5.53240329e-02 5.06768882e-01 -3.04048717e-01
1.50188416e-01 3.09153110e-01 1.96040690e-01 -1.14423680e+00
-9.03629720e-01 -5.02336740e-01 5.49619198e-01 5.96378684e-01
-5.58807850e-01 2.95368433e-01 3.38521957e-01 -1.33383632e-01
-4.91595894e-01 -1.06492174e+00 -4.75266695e-01 -3.77457440e-01
-4.23926711e-01 -2.05588847e-01 -2.78746523e-02 -3.82365614e-01
1.52802026e+00 -1.85537803e+00 -9.03157666e-02 5.13606787e-01
3.03742737e-01 -5.05600385e-02 1.34831458e-01 3.79748315e-01
1.47962287e-01 4.93913382e-01 8.30944926e-02 -1.57185853e-01
4.22465593e-01 -1.90137967e-01 8.37915540e-02 6.10982120e-01
-8.89049649e-01 8.40252042e-01 -6.18295908e-01 -3.65757763e-01
-1.77761748e-01 -8.43958139e-01 -1.19395149e+00 1.58049271e-01
-9.19779018e-02 -6.72908127e-02 -5.10805547e-01 2.79225200e-01
5.77724993e-01 -8.41297135e-02 3.41028154e-01 5.48430145e-01
-5.08089125e-01 -1.73305154e-01 -1.09607005e+00 1.09161842e+00
-3.84014547e-02 2.80591756e-01 2.88833052e-01 -1.20968068e+00
2.79625356e-01 1.00685507e-01 8.45915198e-01 -6.35488212e-01
3.25905442e-01 7.62811750e-02 2.58595884e-01 -3.19402546e-01
7.29845524e-01 -4.78994548e-01 -6.30634725e-02 1.03413105e+00
-4.44024980e-01 -1.39259677e-02 1.61742315e-01 3.28427196e-01
8.02653551e-01 -2.48314008e-01 8.49572241e-01 -6.41015232e-01
-3.19787450e-02 4.66716230e-01 4.37669694e-01 1.37322712e+00
-8.02636743e-01 7.88087845e-02 8.46405149e-01 -4.52934504e-01
-1.12664509e+00 -6.94621623e-01 4.00958031e-01 1.94025803e+00
4.25780416e-01 2.21992075e-01 -7.16565430e-01 -3.34115922e-01
6.24453723e-01 1.16139519e+00 -7.77917385e-01 -2.12414742e-01
-1.83213264e-01 -6.94199979e-01 2.43245408e-01 -1.01698734e-01
5.31048596e-01 -9.96999800e-01 -1.05828631e+00 1.79209888e-01
-3.50693971e-01 -3.03820372e-01 -8.78743887e-01 -3.86957020e-01
-6.04989111e-01 -9.32342649e-01 -7.41792679e-01 1.85904186e-02
1.91309795e-01 5.84078252e-01 8.96048784e-01 2.82376558e-01
-2.14631245e-01 5.16155064e-01 -1.01514570e-01 -7.78020203e-01
9.43933576e-02 -1.12686053e-01 9.83697549e-02 1.07063703e-01
4.11956817e-01 -5.03996491e-01 -9.69742835e-01 3.40096861e-01
-5.05371749e-01 4.11088139e-01 1.48515597e-01 6.49374306e-01
-6.46821335e-02 -4.99804378e-01 9.89947140e-01 -1.14535856e+00
1.03620505e+00 -5.36720693e-01 -4.09587651e-01 4.38844115e-01
-5.58099270e-01 -4.38747913e-01 3.03353995e-01 -5.57521224e-01
-1.03046250e+00 -1.42886326e-01 5.56615531e-01 1.66381374e-01
5.40252686e-01 2.93364495e-01 1.48011118e-01 -2.99134087e-02
9.32567060e-01 -2.77498066e-01 2.07035467e-01 4.13105190e-02
3.79941106e-01 7.63854325e-01 -1.76566735e-01 -8.50714624e-01
1.13271326e-01 2.66681254e-01 -2.04056323e-01 -6.93423092e-01
-5.99161088e-01 -4.05041873e-01 3.43432665e-01 -7.91965723e-01
4.07324195e-01 -5.97430408e-01 -1.84402156e+00 4.82067503e-02
-6.84360445e-01 -6.01583838e-01 -7.94164896e-01 4.95269656e-01
-1.24006593e+00 -4.99102622e-02 -8.66244435e-02 -1.59632289e+00
-3.51358742e-01 -4.62645918e-01 3.44461381e-01 5.46766639e-01
-5.32774329e-01 -3.74345720e-01 2.37501711e-01 9.73870516e-01
6.00426018e-01 -7.00251758e-02 5.53458929e-01 -6.87086999e-01
-5.65359950e-01 -1.02254242e-01 -1.33144706e-01 -2.44567871e-01
-6.52718619e-02 -2.11123466e-01 -5.78700542e-01 -3.97843212e-01
-4.02548850e-01 -4.64290798e-01 4.78805900e-01 1.11249304e+00
1.12263930e+00 -5.98138332e-01 -2.26318553e-01 8.32317919e-02
1.13986993e+00 3.03079724e-01 5.69312513e-01 4.88509506e-01
1.55621171e-01 8.71715844e-01 4.88065124e-01 9.51011121e-01
3.79648775e-01 5.85848153e-01 5.33918858e-01 1.91316344e-02
5.18337309e-01 -3.14916670e-02 -1.21250756e-01 -3.56102109e-01
-7.66799808e-01 -1.83610320e-01 -6.65231407e-01 6.86627686e-01
-2.33867812e+00 -1.44976020e+00 9.82888117e-02 2.39105844e+00
8.04156244e-01 -2.12655529e-01 1.16535389e+00 -2.08402053e-02
1.04434228e+00 -2.80550122e-01 -7.33703315e-01 -7.61888504e-01
3.76922399e-01 -1.66674361e-01 1.02562189e+00 4.66716170e-01
-4.64247316e-01 6.66685760e-01 6.84974861e+00 8.45738351e-01
-3.41606408e-01 2.05111876e-01 1.22678697e+00 -1.01938426e+00
-4.16852236e-01 3.68374996e-02 -3.05490732e-01 4.03805405e-01
9.02509749e-01 -1.04800522e+00 8.99063766e-01 7.31162310e-01
7.92261481e-01 -4.44705725e-01 -8.97347629e-01 8.89944375e-01
-5.10086715e-01 -1.43474960e+00 -4.25136447e-01 3.34156722e-01
1.06218135e+00 -6.50648475e-01 -2.27100272e-02 3.48022401e-01
9.87434626e-01 -9.17561948e-01 8.87346983e-01 4.67716783e-01
5.44510722e-01 -9.43278551e-01 1.21156394e-01 7.14767873e-01
-2.83584446e-01 -6.06730938e-01 4.30098698e-02 -7.39285648e-01
1.33110210e-01 2.04264536e-01 -4.20833021e-01 1.44688666e-01
5.47308981e-01 -1.09852105e-01 3.19757700e-01 1.48047781e+00
4.62412298e-01 5.63171804e-01 -3.66342396e-01 -5.10662079e-01
2.33225718e-01 -5.64422131e-01 4.08997476e-01 5.70466816e-01
3.17394972e-01 6.37370348e-01 -5.63582685e-03 8.59232783e-01
-3.57708752e-01 5.90210795e-01 -3.41467321e-01 -3.33471037e-02
4.50487137e-01 1.09937310e+00 -6.42994046e-01 -2.74239123e-01
-2.34214440e-01 4.93061334e-01 3.41919512e-01 3.35567504e-01
-7.52183378e-01 1.20499954e-02 8.95767093e-01 3.82954210e-01
-2.18361571e-01 4.74876851e-01 -8.42357516e-01 -7.31382906e-01
-3.65781426e-01 -6.32124603e-01 5.31703472e-01 -5.94104230e-01
-1.10749304e+00 -3.91782254e-01 1.73470423e-01 -7.31361806e-01
1.69880807e-01 1.84263200e-01 -6.43853009e-01 8.30139518e-01
-1.11035895e+00 -8.86305749e-01 -3.18686105e-02 3.38762462e-01
1.41196743e-01 7.87585452e-02 1.92434877e-01 3.17186713e-01
-5.53315759e-01 5.82054913e-01 2.47762218e-01 -7.19091475e-01
6.46549582e-01 -1.12008500e+00 -4.11285460e-01 4.45184916e-01
-4.91804570e-01 5.80424905e-01 1.04820120e+00 -4.91239399e-01
-1.10399413e+00 -5.66126645e-01 6.11465394e-01 -2.42749434e-02
3.67003173e-01 2.16476601e-02 -6.36729747e-02 7.06411600e-01
2.02600867e-01 -4.61933732e-01 1.12363923e+00 6.03556454e-01
3.66621286e-01 -3.23600829e-01 -1.54656112e+00 1.03045022e+00
1.43054283e+00 -1.11392543e-01 -2.27020904e-01 3.16994488e-01
4.55637187e-01 -9.30129811e-02 -2.02362940e-01 -1.71342865e-01
1.00618839e+00 -1.31857586e+00 7.78850675e-01 -1.21388853e+00
3.43362570e-01 3.31545085e-01 6.77523948e-03 -1.44729567e+00
-6.86813414e-01 -1.11327720e+00 3.14847529e-01 7.94235587e-01
3.33836973e-01 -5.15152812e-01 1.42046809e+00 1.63017356e+00
1.44263238e-01 -2.69197553e-01 -1.00864339e+00 -6.43410265e-01
1.98518649e-01 -3.41139138e-02 7.73346782e-01 1.15068722e+00
6.17239058e-01 -1.69160113e-01 -1.03282416e+00 -5.08781016e-01
8.45243812e-01 2.79072523e-01 9.98939931e-01 -1.19140291e+00
-3.34879339e-01 -8.83576274e-01 1.35807455e-01 -6.54502928e-01
-1.03110991e-01 -5.59317768e-01 -8.14293772e-02 -1.54126251e+00
7.86499083e-01 -5.25901377e-01 -3.69159192e-01 5.76229990e-01
-1.86466530e-01 1.66745469e-01 5.52669704e-01 3.87228318e-02
-9.00386393e-01 2.46515095e-01 1.22896659e+00 -2.70573646e-02
-8.65906835e-01 4.64495420e-01 -1.61989141e+00 1.81156829e-01
8.49067926e-01 -4.60601658e-01 -2.79268980e-01 2.79009104e-01
5.77785134e-01 6.29365504e-01 1.06621973e-01 -4.26178545e-01
4.12039220e-01 -1.15246892e+00 -2.13815778e-01 -8.44819397e-02
3.49040776e-02 -1.03928912e+00 6.91076636e-01 6.42301500e-01
-6.72563791e-01 -5.74992299e-01 -4.45735455e-02 4.63040382e-01
4.93293256e-01 -2.32362807e-01 8.33811164e-01 -1.26307532e-01
6.23472687e-03 1.57423973e-01 -5.37123561e-01 -7.50548691e-02
1.64821458e+00 -5.37759900e-01 -3.81690204e-01 -9.55134213e-01
-5.32920480e-01 7.12117612e-01 3.36338729e-01 -1.53525651e-01
-5.27647957e-02 -1.29197717e+00 -9.21314120e-01 -3.22434783e-01
4.26224321e-02 -8.98991346e-01 6.11061215e-01 8.43020141e-01
-4.91400599e-01 2.36555278e-01 -6.19255602e-01 3.27484459e-02
-1.49129760e+00 2.60509312e-01 4.85810876e-01 -2.32785106e-01
8.78977105e-02 6.06415033e-01 1.52274892e-01 2.48028524e-02
2.54861228e-02 2.65222818e-01 -1.06808297e-01 3.57368201e-01
4.29034024e-01 1.04287624e+00 -5.29407084e-01 -1.82102516e-01
-4.89967018e-02 -1.59509182e-01 4.77392040e-02 -2.99176574e-01
1.43922257e+00 -2.95655608e-01 -3.10548514e-01 7.52802864e-02
4.09147054e-01 1.40651107e-01 -1.06407428e+00 -9.56793036e-03
-1.08983040e-01 -1.05277610e+00 3.96658257e-02 -7.48022318e-01
-7.81927764e-01 -8.20171461e-02 2.02177078e-01 9.84421909e-01
1.04256248e+00 -3.51290524e-01 3.76241267e-01 2.05017164e-01
4.21287805e-01 -1.35165966e+00 -2.72863925e-01 -2.32071623e-01
5.31716049e-01 -1.10718250e+00 1.84731886e-01 -9.51788500e-02
-8.89720500e-01 5.41915894e-01 5.50967813e-01 -2.13993728e-01
3.47614884e-01 -4.35708582e-01 -5.58121324e-01 5.20549603e-02
-6.39994025e-01 -4.31114197e-01 -1.75749362e-01 1.26815483e-01
2.85331935e-01 6.21941745e-01 -1.12089932e+00 9.24941480e-01
-2.77579665e-01 4.20735151e-01 8.54848742e-01 8.20308387e-01
-8.10695648e-01 -1.07265568e+00 -4.49250966e-01 1.22053504e+00
-5.97871721e-01 1.90621763e-01 -4.31024045e-01 6.23227656e-01
-1.26329526e-01 1.13220000e+00 6.50224611e-02 1.16771460e-01
5.24520218e-01 2.74242368e-02 4.59665447e-01 -4.02688771e-01
-9.44758296e-01 -1.40704468e-01 3.55386585e-01 -4.53648716e-01
-7.24483609e-01 -7.57625699e-01 -6.81394100e-01 -1.15659857e+00
-3.87931883e-01 4.33974236e-01 3.60552788e-01 7.03543425e-01
4.22979742e-02 4.40726250e-01 6.93209529e-01 -6.50534153e-01
-6.44525468e-01 -5.98826647e-01 -1.19730508e+00 5.35166919e-01
4.45633858e-01 -4.89480644e-01 -2.77819544e-01 -4.18920338e-01] | [4.31564998626709, 3.119462490081787] |
93c194cb-0b1f-4749-8cae-96c6b30f0320 | undecimated-wavelet-transform-for-word | 2307.03679 | null | https://arxiv.org/abs/2307.03679v1 | https://arxiv.org/pdf/2307.03679v1.pdf | Undecimated Wavelet Transform for Word Embedded Semantic Marginal Autoencoder in Security improvement and Denoising different Languages | By combining the undecimated wavelet transform within a Word Embedded Semantic Marginal Autoencoder (WESMA), this research study provides a novel strategy for improving security measures and denoising multiple languages. The incorporation of these strategies is intended to address the issues of robustness, privacy, and multilingualism in data processing applications. The undecimated wavelet transform is used as a feature extraction tool to identify prominent language patterns and structural qualities in the input data. The proposed system may successfully capture significant information while preserving the temporal and geographical links within the data by employing this transform. This improves security measures by increasing the system's ability to detect abnormalities, discover hidden patterns, and distinguish between legitimate content and dangerous threats. The Word Embedded Semantic Marginal Autoencoder also functions as an intelligent framework for dimensionality and noise reduction. The autoencoder effectively learns the underlying semantics of the data and reduces noise components by exploiting word embeddings and semantic context. As a result, data quality and accuracy are increased in following processing stages. The suggested methodology is tested using a diversified dataset that includes several languages and security scenarios. The experimental results show that the proposed approach is effective in attaining security enhancement and denoising capabilities across multiple languages. The system is strong in dealing with linguistic variances, producing consistent outcomes regardless of the language used. Furthermore, incorporating the undecimated wavelet transform considerably improves the system's ability to efficiently address complex security concerns | ['Shreyanth S'] | 2023-07-06 | null | null | null | null | ['denoising', 'word-embeddings'] | ['computer-vision', 'methodology'] | [-1.46323353e-01 -4.44031149e-01 -3.15274112e-02 3.22914380e-03
-3.41137439e-01 -5.35067379e-01 6.09518051e-01 4.56431180e-01
-7.69888997e-01 2.63300002e-01 4.65784967e-01 -1.75443947e-01
-3.56305927e-01 -1.10310912e+00 -2.13958308e-01 -8.45825315e-01
-1.75477728e-01 -3.03757638e-01 -1.71929449e-01 -3.73154312e-01
3.48017991e-01 6.36943877e-01 -1.57343817e+00 2.23704457e-01
7.32953846e-01 9.63078439e-01 2.24213481e-01 4.20733899e-01
-3.33970934e-01 5.01433134e-01 -6.56566441e-01 -4.31553423e-01
2.66093045e-01 1.26203537e-01 -4.57651556e-01 2.05664486e-02
-1.64977863e-01 -2.89542973e-01 -1.82826713e-01 1.29815376e+00
6.07406616e-01 1.13062404e-01 3.61074120e-01 -7.77248442e-01
-9.81161535e-01 3.51599246e-01 -4.47361529e-01 4.53466475e-01
3.06535035e-01 -2.17464089e-01 8.90001774e-01 -9.89332974e-01
3.19107771e-01 1.31991136e+00 7.95404136e-01 6.32415563e-02
-7.77079880e-01 -6.29544556e-01 -6.76609650e-02 2.92757899e-01
-1.47737348e+00 -1.31585389e-01 1.01898646e+00 -3.43091011e-01
8.38919342e-01 2.04374120e-01 4.41012949e-01 8.78518999e-01
4.68662560e-01 3.38229328e-01 1.02696061e+00 -7.27427185e-01
1.27748564e-01 5.61131299e-01 3.09434474e-01 7.87803769e-01
4.15452033e-01 1.97222680e-01 -5.96872628e-01 -4.05269057e-01
2.79997081e-01 3.13058197e-01 -2.45298907e-01 1.09457754e-01
-8.68778765e-01 1.10962319e+00 1.47963911e-01 9.54412103e-01
-6.67399168e-01 -5.15171707e-01 6.35072112e-01 2.09162802e-01
6.47880793e-01 2.19540298e-01 -3.35073262e-01 6.32476211e-02
-7.15761244e-01 -1.39530241e-01 5.03965795e-01 4.62302923e-01
5.47091484e-01 6.26888156e-01 3.42304230e-01 8.98837566e-01
3.56222868e-01 7.06456065e-01 9.32395339e-01 -5.31185985e-01
2.60023683e-01 7.93601453e-01 -3.63304406e-01 -1.71733391e+00
-2.04735801e-01 -4.91714627e-01 -7.54256308e-01 1.24478474e-01
-1.17668711e-01 -1.54361635e-01 -6.69887960e-01 1.61455405e+00
4.39620465e-01 -1.34575188e-01 4.95225608e-01 4.76020426e-01
6.23046100e-01 9.85602260e-01 3.82450372e-01 7.12038428e-02
1.85943878e+00 -1.27512127e-01 -1.09606349e+00 -2.65180692e-03
3.07615042e-01 -9.15680707e-01 7.72746623e-01 3.30810696e-01
-5.48494816e-01 -7.45024800e-01 -1.09074092e+00 1.55287847e-01
-9.59963739e-01 1.95882663e-01 4.76635754e-01 7.96836853e-01
-7.56100714e-01 2.09788829e-01 -6.55146420e-01 -5.23772895e-01
7.13213608e-02 2.65469790e-01 -6.82256520e-01 -3.98240797e-02
-1.57110596e+00 1.00601768e+00 7.50261784e-01 1.68419674e-01
-2.78125852e-01 -5.48302054e-01 -1.24490237e+00 1.83835328e-02
-5.77164888e-02 -9.98741537e-02 3.85009855e-01 -6.64227724e-01
-9.21213567e-01 4.61634040e-01 -4.47555110e-02 -5.55914760e-01
-1.41980380e-01 -1.84642300e-02 -1.11612511e+00 3.96978289e-01
1.08656429e-01 3.61038633e-02 9.82728004e-01 -9.08432186e-01
-8.29741299e-01 -6.91064119e-01 -3.32314432e-01 1.62151963e-01
-1.14689314e+00 7.90967569e-02 -4.01495025e-02 -1.17207909e+00
-4.79962900e-02 -4.35272604e-01 2.11888194e-01 -1.00970916e-01
2.07563236e-01 3.73361520e-02 1.41180253e+00 -1.39884603e+00
1.46401215e+00 -2.55750227e+00 -9.14207324e-02 5.41176200e-01
1.08821690e-03 4.80738103e-01 3.33606489e-02 5.84607244e-01
2.88155768e-02 2.80243363e-02 -3.66115093e-01 8.09918195e-02
-2.00682327e-01 4.58298862e-01 -1.85938418e-01 5.34644485e-01
2.25358024e-01 3.64533931e-01 -4.45104808e-01 -2.79768556e-01
3.71900797e-01 9.61052299e-01 -4.23513085e-01 4.39938307e-02
5.76647222e-01 -1.29366666e-01 -5.30733407e-01 7.75557220e-01
6.56035960e-01 5.46293020e-01 4.87847812e-02 -4.33070332e-01
-1.97209209e-01 -3.77604783e-01 -1.31342387e+00 1.20740223e+00
-6.34323895e-01 5.51535487e-01 2.84196764e-01 -1.17316246e+00
1.26253700e+00 5.01340330e-01 4.37851489e-01 -8.28751683e-01
2.95617580e-01 1.15193523e-01 -1.36721581e-01 -8.62724006e-01
5.20244896e-01 -2.62688994e-01 -1.31229863e-01 1.70162350e-01
2.12495863e-01 3.73974025e-01 -1.10739082e-01 -7.02363402e-02
7.24329710e-01 -3.80310237e-01 3.78892004e-01 -3.97524208e-01
8.92088115e-01 -1.99792594e-01 7.10026979e-01 1.34172976e-01
-2.44713664e-01 -2.36934975e-01 -5.83407804e-02 -5.07890344e-01
-9.88932908e-01 -1.06850970e+00 -1.58430621e-01 8.34637463e-01
-9.07031670e-02 -8.63097236e-02 -5.90270877e-01 -3.31283897e-01
1.37145728e-01 8.22317362e-01 -5.91819584e-01 -4.77177441e-01
-2.69935310e-01 -8.28685641e-01 7.65101731e-01 4.67359900e-01
6.98719680e-01 -8.96224678e-01 -6.47504210e-01 2.13110968e-01
-1.35022432e-01 -9.55050826e-01 -1.98044226e-01 1.00022554e-01
-6.83650196e-01 -1.03380036e+00 -3.10578525e-01 -8.78302693e-01
6.35674179e-01 3.14239115e-01 4.12842959e-01 1.98748168e-02
-4.73331660e-01 6.00703299e-01 -6.42203689e-01 -5.18518388e-01
-4.46706563e-01 -3.30703855e-01 3.76570255e-01 3.86853606e-01
8.05751264e-01 -3.46252769e-01 -2.75553852e-01 -2.47406334e-01
-1.43449938e+00 -7.97042608e-01 4.25410211e-01 9.86826479e-01
3.23270082e-01 1.10334170e+00 6.00203037e-01 -5.63450992e-01
1.24472213e+00 -6.62480056e-01 -3.87673765e-01 2.17376247e-01
-7.23148465e-01 9.49066579e-02 7.43091285e-01 -2.34705925e-01
-1.33843589e+00 -4.48614329e-01 -2.07529142e-01 -2.17891876e-02
-2.18387410e-01 6.94130123e-01 -1.24311589e-01 -1.26345515e-01
3.17516357e-01 5.30292332e-01 3.86949152e-01 -5.52414000e-01
1.87858835e-01 1.02989352e+00 5.20921767e-01 -4.88787621e-01
8.37367773e-01 6.26127005e-01 -3.31915081e-01 -1.33497977e+00
-2.25023136e-01 -5.48440397e-01 -5.08543074e-01 -3.84042673e-02
1.14196587e+00 -9.50895846e-01 -5.35417855e-01 5.19600034e-01
-9.63077903e-01 8.34654450e-01 -1.51722550e-01 7.11504161e-01
2.10189253e-01 7.38424122e-01 -5.49777091e-01 -1.06071615e+00
-6.48114324e-01 -9.76059854e-01 6.42817438e-01 9.46257338e-02
-1.81307256e-01 -1.37869442e+00 -1.99365526e-01 3.06458712e-01
4.70842153e-01 3.22975039e-01 1.17412436e+00 -9.02274668e-01
2.30111808e-01 -6.83773696e-01 1.47524357e-01 8.29074442e-01
4.41618472e-01 -7.27125853e-02 -8.47335100e-01 -2.69429266e-01
4.88305658e-01 1.91024959e-01 5.52076340e-01 -2.84808241e-02
7.62436390e-01 -4.99450892e-01 2.98009545e-01 3.54379058e-01
1.43010581e+00 6.07631564e-01 3.25037748e-01 5.11928201e-01
3.69581133e-01 9.79691148e-01 3.28704834e-01 4.72076476e-01
2.29464129e-01 4.63082492e-02 2.32462808e-01 -1.27577260e-01
1.34692684e-01 -2.46957123e-01 4.82392639e-01 1.18159735e+00
1.46716624e-01 1.19604602e-01 -7.95931101e-01 7.67656505e-01
-1.25361109e+00 -1.16095912e+00 1.92838401e-01 1.75897801e+00
4.35404897e-01 -5.34706190e-02 -1.79170713e-01 7.62774825e-01
7.79329896e-01 1.84516594e-01 -1.14882939e-01 -8.46762598e-01
-1.85559377e-01 5.00127494e-01 5.52057147e-01 4.59706098e-01
-1.12867630e+00 6.92828238e-01 5.27052164e+00 9.47870851e-01
-1.10010195e+00 4.40344438e-02 1.75672039e-01 4.88212287e-01
-2.64902115e-01 -3.51565301e-01 -4.62428689e-01 3.63071054e-01
8.79009068e-01 -1.32388741e-01 3.73673111e-01 7.20036745e-01
4.05116409e-01 -1.08470581e-02 -2.84243912e-01 7.25523889e-01
2.51839876e-01 -1.05907035e+00 2.49722123e-01 4.46189381e-02
6.80182055e-02 -4.23620433e-01 3.49526554e-01 -8.83744098e-03
-1.25677543e-04 -7.67208040e-01 6.07829869e-01 3.85835439e-01
5.02607465e-01 -1.31628060e+00 1.02650797e+00 2.57489711e-01
-1.32096505e+00 -5.82641780e-01 -2.22712144e-01 -4.65302402e-03
2.67758779e-03 4.09618199e-01 -6.31307960e-01 7.87214398e-01
8.67153525e-01 5.60558856e-01 -4.16789323e-01 4.48664755e-01
1.73120365e-01 5.22543311e-01 -2.00506970e-01 1.00998200e-01
3.92244846e-01 -2.88846821e-01 5.83310306e-01 1.32266438e+00
4.93757695e-01 3.06786783e-02 7.19837025e-02 3.74805331e-01
3.39708120e-01 4.03008819e-01 -8.56557012e-01 -3.94514918e-01
5.84905446e-01 9.65088189e-01 -6.47207916e-01 -4.33497205e-02
-4.88726705e-01 7.23513663e-01 -1.66057393e-01 3.60272855e-01
-4.87162530e-01 -7.67322659e-01 8.12386394e-01 -5.14384136e-02
3.36067319e-01 -3.55859429e-01 -3.27408105e-01 -9.60171640e-01
-4.90304157e-02 -1.03061604e+00 7.69111097e-01 -3.28237444e-01
-1.25942636e+00 9.26545799e-01 -4.71256934e-02 -1.14112818e+00
3.33178341e-02 -8.58147860e-01 -3.91986609e-01 9.15986598e-01
-1.46755397e+00 -1.34634113e+00 2.32772887e-01 9.47959065e-01
5.51729679e-01 -8.52006435e-01 1.14069033e+00 5.03606319e-01
-5.44715226e-01 4.30650294e-01 2.16895103e-01 3.60566795e-01
3.03313106e-01 -9.04272556e-01 -2.77135491e-01 1.38195813e+00
1.99455783e-01 1.04430163e+00 5.45216322e-01 -7.40489125e-01
-1.51006186e+00 -9.44064617e-01 9.01776314e-01 1.01348169e-01
8.61384869e-01 -3.37381631e-01 -8.87834370e-01 2.24905476e-01
3.57486010e-01 -3.23006690e-01 1.10776150e+00 -9.45784822e-02
-5.21845520e-01 -2.93109238e-01 -1.56528711e+00 4.51836735e-01
1.43254638e-01 -9.58944857e-01 -8.89220417e-01 -1.00586317e-01
7.21230090e-01 2.78223127e-01 -1.15438271e+00 1.61418125e-01
4.43353415e-01 -5.57332158e-01 1.05559576e+00 -5.47309101e-01
1.05230846e-01 -2.64964879e-01 -6.04013622e-01 -1.10309970e+00
-3.96066755e-01 -7.50336498e-02 2.48041246e-02 1.51842606e+00
1.66057542e-01 -9.05075848e-01 2.16869161e-01 4.65069890e-01
1.67704672e-01 -4.40738231e-01 -9.95105088e-01 -5.41881025e-01
-1.28901124e-01 -6.75144255e-01 5.16063690e-01 1.12143207e+00
-2.16617391e-01 -2.73945361e-01 -5.29182613e-01 5.82716107e-01
7.48814940e-01 -3.04617137e-01 1.41217664e-01 -1.02310348e+00
-5.27427383e-02 -1.75781474e-01 -6.52553082e-01 -4.29596659e-03
2.94011176e-01 -9.12802398e-01 -5.70555866e-01 -1.03148687e+00
-1.96298316e-01 9.42566320e-02 -4.75116134e-01 3.95951599e-01
4.44500037e-02 1.92809150e-01 1.55142128e-01 -1.16536833e-01
3.28753710e-01 5.29811740e-01 5.55696547e-01 -2.12614760e-01
8.48289654e-02 -1.45329818e-01 -9.23029125e-01 9.35464859e-01
8.16007376e-01 -4.37503606e-01 -5.24344862e-01 -5.55179417e-01
-1.88204363e-01 -2.90544897e-01 2.52248913e-01 -7.69571781e-01
1.76065177e-01 3.63689885e-02 4.57740426e-01 -3.38554859e-01
1.87021613e-01 -1.45108879e+00 1.13308370e-01 7.34308481e-01
-9.49019194e-02 4.87773687e-01 4.88540918e-01 7.70922601e-01
-5.95201731e-01 -1.83768794e-01 6.89936221e-01 -1.10553650e-04
-1.17392969e+00 -8.08741897e-02 -6.74169362e-01 -4.26732600e-01
1.12248039e+00 -4.68824953e-01 8.20421726e-02 -2.19027594e-01
-6.55177057e-01 -1.81118652e-04 2.99020968e-02 6.69182479e-01
9.60283458e-01 -1.34507394e+00 -5.82203448e-01 6.96731806e-01
2.40916740e-02 -7.72650540e-01 4.24052894e-01 3.04035962e-01
-7.26525843e-01 2.71980584e-01 -3.02682161e-01 -6.28056303e-02
-1.41854906e+00 5.30462742e-01 -3.05051096e-02 -1.36580959e-01
-6.13601267e-01 5.74906945e-01 -3.14427555e-01 -4.32124823e-01
2.63506562e-01 7.35177323e-02 -5.82827091e-01 3.65962476e-01
5.99222243e-01 4.32851195e-01 6.18198588e-02 -1.08758044e+00
-5.11946857e-01 5.61560154e-01 1.38699552e-02 -7.38302022e-02
1.58478057e+00 -2.19709858e-01 -3.65958244e-01 2.10628420e-01
1.25525248e+00 3.95857275e-01 -4.53843355e-01 -2.56303966e-01
2.70961553e-01 -4.28216815e-01 4.89037871e-01 -6.81982160e-01
-1.12623274e+00 9.08315778e-01 9.19331968e-01 2.77881593e-01
1.51267588e+00 -6.23538256e-01 9.31256652e-01 1.17372148e-01
2.57432818e-01 -1.33345044e+00 -2.51216739e-01 2.82761872e-01
6.56946063e-01 -9.86680627e-01 -1.29375100e-01 1.00579262e-02
-7.88622856e-01 1.29372847e+00 7.74670616e-02 -2.61293519e-02
1.03930449e+00 4.19918448e-01 1.15935691e-01 -1.92524508e-01
-1.78131908e-01 -1.92549620e-02 3.59878540e-01 7.85161734e-01
2.52988964e-01 9.62065160e-02 -5.57377279e-01 7.63507307e-01
-1.47074908e-01 -5.93696952e-01 2.26742923e-01 1.08388281e+00
-5.78761518e-01 -1.16142416e+00 -8.56366158e-01 1.24783322e-01
-8.26509595e-01 -5.52623309e-02 -1.87262475e-01 8.46224368e-01
5.20247817e-01 1.16506457e+00 -1.32210165e-01 -5.60559511e-01
3.28083783e-01 1.67918906e-01 -2.57083327e-01 -8.10455903e-02
-6.28401160e-01 5.69226369e-02 -1.40429929e-01 -2.33264372e-01
-3.72594863e-01 -5.00810266e-01 -1.07742643e+00 -3.46219808e-01
-1.47588089e-01 4.75897312e-01 9.43381429e-01 8.63849640e-01
3.54444474e-01 5.18358052e-01 6.88325047e-01 -8.49334449e-02
-6.09311998e-01 -6.45284653e-01 -7.18441844e-01 5.21266699e-01
3.11319113e-01 -6.06595874e-01 -2.55867690e-01 4.60783057e-02] | [10.175972938537598, 8.694068908691406] |
93296783-85aa-4967-93e3-52df56dd244e | sketch-less-for-more-on-the-fly-fine-grained | 2002.10310 | null | https://arxiv.org/abs/2002.10310v4 | https://arxiv.org/pdf/2002.10310v4.pdf | Sketch Less for More: On-the-Fly Fine-Grained Sketch Based Image Retrieval | Fine-grained sketch-based image retrieval (FG-SBIR) addresses the problem of retrieving a particular photo instance given a user's query sketch. Its widespread applicability is however hindered by the fact that drawing a sketch takes time, and most people struggle to draw a complete and faithful sketch. In this paper, we reformulate the conventional FG-SBIR framework to tackle these challenges, with the ultimate goal of retrieving the target photo with the least number of strokes possible. We further propose an on-the-fly design that starts retrieving as soon as the user starts drawing. To accomplish this, we devise a reinforcement learning-based cross-modal retrieval framework that directly optimizes rank of the ground-truth photo over a complete sketch drawing episode. Additionally, we introduce a novel reward scheme that circumvents the problems related to irrelevant sketch strokes, and thus provides us with a more consistent rank list during the retrieval. We achieve superior early-retrieval efficiency over state-of-the-art methods and alternative baselines on two publicly available fine-grained sketch retrieval datasets. | ['Yi-Zhe Song', 'Ayan Kumar Bhunia', 'Yongxin Yang', 'Timothy M. Hospedales', 'Tao Xiang'] | 2020-02-24 | null | null | null | null | ['sketch-based-image-retrieval', 'on-the-fly-sketch-based-image-retrieval'] | ['computer-vision', 'computer-vision'] | [ 1.42891094e-01 -6.24493837e-01 -3.55883449e-01 -1.33364499e-01
-1.37236524e+00 -7.86121070e-01 7.94058502e-01 -6.95418566e-02
-3.04752499e-01 4.75950539e-01 2.60768086e-01 2.46900603e-01
-4.72501248e-01 -7.95792162e-01 -5.32536924e-01 -5.26055992e-01
3.97486657e-01 5.80332577e-01 6.96006343e-02 -4.44920287e-02
7.65872180e-01 8.66321087e-01 -1.55619752e+00 3.24702322e-01
4.58733231e-01 9.58770454e-01 3.98658961e-01 6.57485008e-01
-1.30725160e-01 4.53262806e-01 -6.89964235e-01 -5.78256905e-01
3.09435248e-01 -1.60970196e-01 -6.96924329e-01 -9.41430330e-02
1.07156742e+00 -1.02469397e+00 -7.58612812e-01 6.65647209e-01
7.00609386e-01 2.80471444e-01 8.80666614e-01 -1.16681612e+00
-1.00745308e+00 1.23882167e-01 -5.89879215e-01 -1.78262845e-01
5.54541230e-01 -7.37327933e-02 1.33835292e+00 -1.08125234e+00
9.72248554e-01 1.38008451e+00 5.06062582e-02 7.75022268e-01
-1.18921494e+00 -7.31493592e-01 3.92468601e-01 1.17545566e-02
-1.68433559e+00 -4.37121868e-01 9.82298553e-01 -1.69590861e-01
5.07437646e-01 2.80729532e-01 6.24760985e-01 1.03527486e+00
-4.53553289e-01 1.37982535e+00 7.14924574e-01 -3.96553427e-01
2.16794044e-01 -3.58216733e-01 -1.36431515e-01 6.59918427e-01
-4.94945347e-02 1.32111698e-01 -8.50839913e-01 -3.16399813e-01
1.22820044e+00 5.14942229e-01 -3.01009789e-02 -7.99051762e-01
-1.19596076e+00 6.77999854e-01 5.01683414e-01 3.56162548e-01
-4.88034129e-01 7.05627143e-01 2.01398611e-01 2.48564497e-01
1.85467839e-01 4.69350368e-01 1.00899175e-01 -6.92767128e-02
-1.43608594e+00 6.58807814e-01 4.68365341e-01 1.07957053e+00
7.06010759e-01 -4.01524663e-01 -7.28125334e-01 1.13559783e+00
1.49509877e-01 6.76690698e-01 -5.21596661e-03 -1.23280001e+00
2.99901575e-01 3.13448995e-01 4.80591536e-01 -9.58898365e-01
5.72289169e-01 -2.77156532e-02 -6.82889283e-01 1.59589857e-01
2.25129634e-01 5.69563210e-01 -8.07859600e-01 1.62033665e+00
-5.55393063e-02 -3.37960459e-02 -2.65209377e-01 1.16547966e+00
7.11942613e-01 5.97293496e-01 2.13465586e-01 2.63358921e-01
1.14671493e+00 -1.11692083e+00 -4.99309957e-01 -4.89736199e-02
-2.99343318e-01 -8.33674610e-01 1.29383910e+00 4.38394278e-01
-1.23922122e+00 -5.33188581e-01 -1.05262816e+00 -3.85219365e-01
-3.19758922e-01 5.20394564e-01 5.90205669e-01 2.46245191e-01
-1.05009258e+00 8.25102806e-01 -3.20881993e-01 -2.74585992e-01
5.66022515e-01 1.53433949e-01 -3.86301309e-01 -6.11965001e-01
-7.92810738e-01 6.47501349e-01 -1.12296537e-01 5.48811667e-02
-1.02346265e+00 -6.71513081e-01 -3.43714327e-01 2.26274639e-01
4.78371769e-01 -9.17000353e-01 1.27762699e+00 -6.97269797e-01
-1.47022128e+00 8.76028657e-01 -2.29938835e-01 1.63422018e-01
7.72931933e-01 -5.77008545e-01 1.31440386e-01 4.64937031e-01
1.68761611e-01 1.10293305e+00 1.35462809e+00 -1.42058659e+00
-2.72197157e-01 -3.04506063e-01 2.35150546e-01 2.51357138e-01
-2.48376071e-01 -2.30709851e-01 -1.23971355e+00 -8.95282567e-01
-1.58270434e-01 -7.91029334e-01 -4.80166776e-03 6.95814371e-01
-1.33334666e-01 -6.89261973e-01 5.84514141e-01 -3.30263227e-01
1.28162456e+00 -2.20458603e+00 3.40925306e-01 2.57820070e-01
1.22152805e-01 2.47219637e-01 -7.16772854e-01 9.49425101e-01
3.70896459e-01 1.75196733e-02 1.61017746e-01 -4.36772585e-01
2.86195815e-01 1.26836568e-01 -6.91858053e-01 7.82340989e-02
2.63366103e-01 1.25104690e+00 -1.14643407e+00 -6.71720386e-01
2.44582295e-01 4.99389172e-01 -4.07100677e-01 4.72432792e-01
-2.55599886e-01 -2.40016021e-02 -8.28453779e-01 7.78039396e-01
5.64399600e-01 -3.35652679e-01 8.05034041e-02 -3.50708842e-01
1.96507782e-01 -2.72123784e-01 -1.09438312e+00 2.41745782e+00
-4.83706802e-01 4.83004749e-01 -1.23597100e-01 -4.25725996e-01
9.02011812e-01 -3.34284492e-02 2.96223581e-01 -1.05287921e+00
-3.60134393e-01 1.74968436e-01 -8.05136740e-01 -1.56273648e-01
9.01273966e-01 2.45805979e-02 -1.38061866e-01 7.28102803e-01
-1.93467885e-01 -3.86940271e-01 2.04042405e-01 6.25044882e-01
9.58009601e-01 4.79050785e-01 9.67725739e-02 4.48050760e-02
3.13254893e-01 -3.27133566e-01 -1.21383309e-01 1.20289600e+00
6.79233745e-02 8.08842599e-01 1.84804931e-01 -5.12079775e-01
-1.13725543e+00 -1.36577809e+00 3.21025431e-01 1.27316809e+00
4.72786099e-01 -4.05553162e-01 -4.38750625e-01 -4.76049185e-01
3.62031788e-01 2.66392261e-01 -5.80070436e-01 4.33738120e-02
-5.24923027e-01 2.80939490e-01 4.29935098e-01 4.99300092e-01
3.74518216e-01 -1.36994016e+00 -6.39251709e-01 6.34296238e-02
-1.95409909e-01 -6.51429117e-01 -1.01217556e+00 -5.20486236e-01
-7.34913766e-01 -1.05405986e+00 -1.48195469e+00 -7.36079693e-01
8.33023846e-01 7.38060951e-01 1.43708682e+00 5.46767294e-01
-6.66058481e-01 8.14853787e-01 -2.74802774e-01 1.75051659e-01
1.38300300e-01 4.80123162e-02 -4.30278510e-01 -2.22418875e-01
6.21501617e-02 -1.82504267e-01 -1.14141893e+00 2.03720629e-01
-1.09008873e+00 -9.65442508e-02 8.79217684e-01 1.02295744e+00
8.16845298e-01 -3.54535669e-01 7.37379551e-01 -5.73232293e-01
1.00232160e+00 1.00157738e-01 -4.49930221e-01 8.56221855e-01
-5.39715528e-01 3.70994806e-01 2.82329708e-01 -5.03162742e-01
-7.78124332e-01 2.68330537e-02 1.58247814e-01 -9.32449043e-01
1.77547246e-01 4.82270196e-02 1.35877803e-01 -1.97136793e-02
3.32949400e-01 3.49936128e-01 -2.62600034e-01 -5.42097807e-01
8.32915008e-01 5.25164843e-01 5.23841381e-01 -1.04773617e+00
7.83710003e-01 3.80517662e-01 -1.63314432e-01 -6.60923123e-01
-6.74695253e-01 -6.88295245e-01 -5.05134523e-01 -3.27842295e-01
4.33907121e-01 -8.63787293e-01 -1.10012925e+00 2.52382547e-01
-1.18210590e+00 -3.23449284e-01 -2.29439169e-01 -1.75128654e-01
-6.31330192e-01 5.48125386e-01 -4.60482150e-01 -1.03673327e+00
-6.89386725e-01 -9.44384336e-01 1.90638983e+00 7.30849504e-02
-1.53754488e-01 -3.20327967e-01 1.50411308e-01 8.12838227e-02
4.30859596e-01 -1.27712503e-01 9.43326473e-01 9.71801579e-02
-1.29822803e+00 -3.34530145e-01 -7.71145403e-01 -1.01622164e-01
-6.91433474e-02 7.53438100e-02 -8.02319050e-01 -5.29438794e-01
-1.00639856e+00 -6.50868058e-01 1.07087064e+00 1.26005784e-01
1.31717741e+00 -1.65360466e-01 -3.08784634e-01 1.49034351e-01
1.66930974e+00 -5.57587445e-02 6.89420760e-01 2.38100961e-02
3.91119927e-01 3.73998165e-01 9.11308289e-01 4.35052693e-01
1.69382170e-01 8.67777467e-01 1.27776057e-01 2.94438712e-02
-3.28634560e-01 -8.38025510e-01 -2.57480472e-01 2.37956509e-01
3.43155749e-02 -3.07715803e-01 -2.07879603e-01 7.15861082e-01
-1.93297863e+00 -1.10180914e+00 7.77046263e-01 2.36606622e+00
6.96199298e-01 -2.35223979e-01 9.73163098e-02 -9.13184136e-02
5.21949053e-01 4.75677222e-01 -6.79097950e-01 1.46883831e-03
1.41371205e-01 4.08197999e-01 5.02357446e-02 4.41251129e-01
-8.78608704e-01 1.19224262e+00 6.13466406e+00 1.17906010e+00
-8.86091053e-01 -4.22928095e-01 4.02318984e-01 -3.26750398e-01
-4.91163254e-01 -1.70137823e-01 -4.62994874e-01 1.19576789e-01
1.11120701e-01 -5.62784448e-02 9.92200255e-01 7.76456654e-01
-2.41005331e-01 -1.55256405e-01 -1.40668690e+00 1.47712898e+00
2.13379025e-01 -1.40338337e+00 6.11087203e-01 -1.99288353e-01
6.32882714e-01 -3.81112605e-01 1.30872831e-01 2.66427279e-01
1.54919282e-01 -9.39050734e-01 8.29416990e-01 9.24528599e-01
1.36513782e+00 -7.68047690e-01 -3.54921110e-02 5.29143848e-02
-1.22493505e+00 6.22020848e-02 -4.01826262e-01 3.24882299e-01
-1.20686568e-01 1.68479994e-01 -4.75351244e-01 5.44822991e-01
4.36577260e-01 6.13595843e-01 -5.38444817e-01 9.77212667e-01
-1.56415761e-01 -2.45992944e-01 -6.86935484e-02 -3.30640286e-01
2.47965947e-01 -1.43684298e-01 1.73664019e-01 1.06280506e+00
2.54628330e-01 2.23410845e-01 2.33257920e-01 8.57110202e-01
-4.62806433e-01 1.16323091e-01 -6.74407959e-01 -3.27102095e-01
8.17638099e-01 1.24134910e+00 -5.80189764e-01 -4.84556049e-01
-4.51411456e-02 1.51803434e+00 6.50193632e-01 5.34514368e-01
-2.93297112e-01 -5.01531720e-01 4.39413041e-01 -9.22500789e-02
4.32576209e-01 -1.32750958e-01 2.14976400e-01 -1.04917479e+00
1.49595961e-01 -7.77221799e-01 3.27878952e-01 -1.19435513e+00
-1.68585563e+00 4.37949985e-01 -1.89131200e-01 -1.13246536e+00
-3.36446166e-01 -2.76117235e-01 -3.80746186e-01 9.21017170e-01
-1.85119712e+00 -1.40607274e+00 -4.73933905e-01 6.55486226e-01
8.00930858e-01 7.69915059e-02 9.65222776e-01 4.57801163e-01
-6.14137463e-02 6.70894682e-01 2.42352728e-02 1.02152623e-01
1.16008794e+00 -1.15554059e+00 5.30579209e-01 3.28120887e-01
3.86328816e-01 8.79702747e-01 2.24276423e-01 -6.13063872e-01
-1.76731002e+00 -7.05531716e-01 7.52085626e-01 -3.81242812e-01
3.33353817e-01 -2.41686106e-01 -5.66103697e-01 7.30711296e-02
-4.45813499e-03 3.50461379e-02 1.52719781e-01 9.83757526e-02
-6.85548186e-01 -3.06408614e-01 -9.56509829e-01 9.09155071e-01
1.16012084e+00 -1.14201725e+00 -4.42569375e-01 1.63644597e-01
2.82091647e-01 -2.14776292e-01 -6.50023818e-01 -1.11330390e-01
1.34106863e+00 -6.59989238e-01 1.50939405e+00 -5.07746398e-01
5.74372053e-01 -2.19113439e-01 -2.21202403e-01 -9.54857171e-01
-3.62198859e-01 -5.97756267e-01 -7.30465949e-02 9.51171994e-01
-1.74921274e-01 4.33635227e-02 9.47980165e-01 6.38051331e-01
4.96154070e-01 -7.77346134e-01 -5.75805843e-01 -6.65589333e-01
-1.30433679e-01 2.93376017e-03 6.91797912e-01 3.74418437e-01
-2.89362431e-01 2.07996443e-01 -6.13489747e-01 -1.89513639e-01
8.03928971e-01 7.28933275e-01 9.39439058e-01 -1.23644483e+00
-1.56059682e-01 -6.56243980e-01 -5.43045551e-02 -1.52693737e+00
1.52083755e-01 -5.84174514e-01 1.95628047e-01 -1.70651054e+00
5.59259057e-01 -6.18032277e-01 -4.12795424e-01 4.11121875e-01
-3.05673897e-01 5.19528806e-01 6.07101738e-01 5.69102705e-01
-1.07771885e+00 5.48471570e-01 1.47649789e+00 -3.62903535e-01
4.14856263e-02 -2.82781959e-01 -6.25051439e-01 -2.11375300e-02
2.12590024e-01 -2.59783030e-01 -5.47941566e-01 -4.95632380e-01
2.93180227e-01 5.04908741e-01 6.13256156e-01 -4.73429739e-01
4.11119610e-01 -7.59564415e-02 6.44109845e-01 -1.01654434e+00
6.44717216e-01 -7.43275583e-01 8.17361102e-02 1.70919806e-01
-6.78786397e-01 3.21352147e-02 -1.01509422e-01 8.12179148e-01
-1.30602345e-01 -2.68365204e-01 4.40901697e-01 -2.59253889e-01
-6.85881197e-01 5.92992842e-01 6.58115298e-02 -7.95128196e-02
6.18433952e-01 4.35686037e-02 -3.22533667e-01 -5.20972729e-01
-4.02455628e-01 1.28607109e-01 6.88242316e-01 6.20551765e-01
8.89741182e-01 -1.67711914e+00 -5.72897673e-01 4.13954332e-02
4.32914317e-01 -3.68264854e-01 3.79441082e-01 -5.97476922e-02
-2.84093827e-01 6.23520851e-01 -2.64727563e-01 -2.43779719e-01
-1.24390161e+00 8.34551454e-01 -8.73787105e-02 -3.98839653e-01
-6.09516084e-01 6.70542538e-01 2.19571918e-01 -1.25272842e-02
4.94470596e-01 9.64090154e-02 2.14723200e-01 8.60718936e-02
6.22801602e-01 3.94079238e-01 -9.26461816e-02 -8.78971070e-02
-2.16930985e-01 9.03305888e-01 -2.60312051e-01 -3.50057006e-01
1.25023282e+00 -6.18385710e-03 1.34402305e-01 1.52258560e-01
1.17011833e+00 2.57181581e-02 -1.46817851e+00 -3.03901315e-01
-1.59147292e-01 -1.00749815e+00 -6.98199729e-03 -1.15900540e+00
-8.81495833e-01 8.61452579e-01 5.21210074e-01 -1.86941996e-01
9.95635986e-01 -3.83376032e-02 8.41123283e-01 8.54963243e-01
6.02544904e-01 -9.54501688e-01 6.25634074e-01 -7.30702281e-02
1.29283369e+00 -1.11294472e+00 3.19696128e-01 5.25961779e-02
-4.95723307e-01 1.20528138e+00 2.87466645e-01 -5.13616681e-01
2.25255772e-01 -2.93325812e-01 -1.45431548e-01 -3.89750600e-01
-6.86819732e-01 -1.29981488e-01 6.19230628e-01 2.59495497e-01
2.25962937e-01 4.02906649e-02 -1.09799333e-01 9.25047994e-02
4.62201446e-01 3.37425917e-01 -1.29936680e-01 9.67279494e-01
-2.54179090e-01 -1.45063114e+00 -1.28843278e-01 4.34440672e-01
-9.09745842e-02 -1.77402094e-01 -6.97159231e-01 5.65109849e-01
-6.18679523e-01 6.15053773e-01 -6.15542494e-02 1.68881342e-01
4.62660253e-01 4.26587537e-02 1.02175367e+00 -2.90790498e-01
-3.33263338e-01 1.25851393e-01 -2.40804330e-01 -7.61024356e-01
-3.66948843e-01 -3.23611200e-01 -6.59048438e-01 -2.97052205e-01
-2.76714236e-01 5.86925671e-02 6.29929304e-01 5.43875754e-01
5.52300096e-01 9.44194943e-02 6.38311684e-01 -1.33001602e+00
-5.62106729e-01 -6.17165983e-01 -5.36859035e-01 5.85544765e-01
4.46224064e-01 -6.67096496e-01 4.05748710e-02 -1.52248055e-01] | [11.651276588439941, 0.5857167840003967] |
ea7b53bc-79c7-43a6-aede-31692316e13b | structural-bias-for-aspect-sentiment-triplet | 2209.00820 | null | https://arxiv.org/abs/2209.00820v1 | https://arxiv.org/pdf/2209.00820v1.pdf | Structural Bias for Aspect Sentiment Triplet Extraction | Structural bias has recently been exploited for aspect sentiment triplet extraction (ASTE) and led to improved performance. On the other hand, it is recognized that explicitly incorporating structural bias would have a negative impact on efficiency, whereas pretrained language models (PLMs) can already capture implicit structures. Thus, a natural question arises: Is structural bias still a necessity in the context of PLMs? To answer the question, we propose to address the efficiency issues by using an adapter to integrate structural bias in the PLM and using a cheap-to-compute relative position structure in place of the syntactic dependency structure. Benchmarking evaluation is conducted on the SemEval datasets. The results show that our proposed structural adapter is beneficial to PLMs and achieves state-of-the-art performance over a range of strong baselines, yet with a light parameter demand and low latency. Meanwhile, we give rise to the concern that the current evaluation default with data of small scale is under-confident. Consequently, we release a large-scale dataset for ASTE. The results on the new dataset hint that the structural adapter is confidently effective and efficient to a large scale. Overall, we draw the conclusion that structural bias shall still be a necessity even with PLMs. | ['Dawei Song', 'Wei Wu', 'Jingang Wang', 'Fang Ma', 'Lei Ren', 'Chen Zhang'] | 2022-09-02 | null | https://aclanthology.org/2022.coling-1.585 | https://aclanthology.org/2022.coling-1.585.pdf | coling-2022-10 | ['aspect-sentiment-triplet-extraction'] | ['natural-language-processing'] | [ 7.33883157e-02 1.95250586e-01 -3.33701462e-01 -4.39349413e-01
-9.01483119e-01 -5.91722250e-01 6.30951881e-01 2.08050847e-01
-7.54362762e-01 5.21619976e-01 4.88075763e-01 -6.09228730e-01
2.03853101e-01 -5.48669994e-01 -6.56090260e-01 -3.78169656e-01
2.97826082e-01 5.51690876e-01 3.52914274e-01 -4.50094700e-01
3.22215110e-01 3.83522242e-01 -1.28638637e+00 4.87484276e-01
6.05533659e-01 7.94346035e-01 -7.23487362e-02 2.74066448e-01
-4.37140405e-01 6.44260466e-01 -5.73624313e-01 -9.88439143e-01
1.77990586e-01 7.46903345e-02 -9.45970416e-01 -1.39546752e-01
6.29593074e-01 -1.56530857e-01 4.33071434e-01 7.91029036e-01
4.30657685e-01 -3.88294578e-01 4.53885138e-01 -1.01502895e+00
-6.57526925e-02 7.45939970e-01 -6.52254879e-01 1.33389026e-01
4.18757796e-02 9.18853730e-02 1.60346103e+00 -1.07366800e+00
6.60269380e-01 8.77943277e-01 7.13668883e-01 2.96907455e-01
-1.30686593e+00 -4.35289919e-01 6.72638476e-01 -5.63662462e-02
-1.03889060e+00 -6.03297591e-01 5.48046589e-01 -2.27077931e-01
1.47663379e+00 3.80642176e-01 4.25137609e-01 1.18542409e+00
-6.15955740e-02 9.46364462e-01 1.22677791e+00 -5.19345820e-01
-2.35629804e-03 5.89630187e-01 4.34482545e-01 6.83087230e-01
7.68519521e-01 -1.22520842e-01 -6.02448106e-01 -3.43106598e-01
1.34271905e-02 -4.20716286e-01 -6.92724884e-02 -3.48351330e-01
-9.45371807e-01 8.33149433e-01 1.65337712e-01 5.31291723e-01
-2.24442884e-01 1.36640063e-03 6.01208925e-01 2.71908283e-01
5.17894983e-01 6.49549663e-01 -8.65027905e-01 -3.62949461e-01
-1.09859455e+00 2.75495470e-01 1.02174091e+00 7.52297819e-01
6.38974190e-01 -1.25369191e-01 1.29980564e-01 6.12561941e-01
1.48765668e-01 4.65198874e-01 1.64955541e-01 -7.27998257e-01
7.25737572e-01 8.21587741e-01 -6.59333095e-02 -7.73400366e-01
-3.93597543e-01 -5.89751363e-01 -6.06167138e-01 2.14487687e-01
4.96959597e-01 -4.98130471e-02 -5.96646607e-01 1.75816584e+00
2.91537106e-01 -1.37207612e-01 1.13113230e-04 7.61749327e-01
5.36212444e-01 4.56416577e-01 4.31484953e-02 -7.14122429e-02
1.64030230e+00 -8.79499555e-01 -5.94761252e-01 -6.46892428e-01
9.60143149e-01 -9.67740953e-01 1.37022996e+00 4.99560714e-01
-1.01466167e+00 -1.25943407e-01 -1.20178080e+00 -2.90271137e-02
-3.37286592e-01 9.73794237e-02 9.05791163e-01 8.14552188e-01
-9.23665702e-01 4.01496381e-01 -9.20733333e-01 -4.59421456e-01
1.36505023e-01 3.73647809e-01 -3.90661299e-01 3.73442173e-02
-9.10960972e-01 1.07429576e+00 2.55477875e-01 3.74239273e-02
-2.00953841e-01 -6.29575968e-01 -6.07441545e-01 -4.55761626e-02
6.66247845e-01 -7.14250922e-01 1.16224968e+00 -1.05605602e+00
-1.29176354e+00 9.51615512e-01 -4.45093364e-01 -5.58265567e-01
3.78817797e-01 -5.14225304e-01 -2.26872861e-01 -1.23239152e-01
-4.47792280e-03 4.89936292e-01 6.35238588e-01 -1.27347338e+00
-6.60262883e-01 -2.83193171e-01 3.01043242e-01 1.34635955e-01
-4.47908729e-01 3.21023047e-01 -4.04403746e-01 -5.05261540e-01
-9.21369493e-02 -1.11849844e+00 -1.93493813e-01 -4.82096642e-01
-1.59398615e-01 -2.62940973e-01 4.44501728e-01 -6.20832920e-01
1.54278696e+00 -2.06951690e+00 1.37613313e-02 7.69734755e-02
-1.29363433e-01 4.84285086e-01 -1.96260244e-01 3.58984023e-01
-6.26816228e-02 2.89476693e-01 -4.43728447e-01 -5.76439142e-01
-2.23844033e-03 2.94032365e-01 -3.88917893e-01 1.43333763e-01
2.90125072e-01 1.02068889e+00 -5.59630871e-01 -4.15116817e-01
-2.11087510e-01 3.64173681e-01 -1.04480469e+00 1.92119092e-01
-3.84130418e-01 -5.57073392e-02 -2.18263417e-01 5.28169215e-01
5.89043379e-01 -3.60571235e-01 4.52331960e-01 -3.28359574e-01
-1.90585643e-01 9.81160700e-01 -1.19012320e+00 1.58051062e+00
-5.18872857e-01 2.90840030e-01 8.23890641e-02 -6.78039432e-01
6.23860538e-01 2.33556211e-01 1.34822413e-01 -6.85510874e-01
-4.42106910e-02 5.02531767e-01 2.43103489e-01 -1.47957027e-01
8.35548282e-01 -2.37651616e-01 -1.72828332e-01 6.60523653e-01
-1.07654937e-01 -4.53021452e-02 2.10591167e-01 3.18677664e-01
1.07041204e+00 2.89574593e-01 2.98230499e-01 -3.74108613e-01
5.17468095e-01 -1.04989717e-02 4.70870107e-01 3.33761722e-01
9.26769301e-02 5.22323906e-01 6.82033956e-01 -3.87490988e-01
-8.29963446e-01 -7.14148045e-01 1.06342897e-01 1.12050998e+00
-3.42714787e-01 -9.39957500e-01 -6.98276639e-01 -1.16748846e+00
-3.26864809e-01 8.99567962e-01 -4.22083795e-01 2.56239921e-01
-9.49365854e-01 -1.08366418e+00 4.27904874e-01 4.85297650e-01
3.04152966e-01 -9.84762967e-01 -7.03020275e-01 2.50020057e-01
-2.72234708e-01 -1.36321580e+00 -3.94234806e-01 4.01479185e-01
-9.45932329e-01 -9.75255370e-01 -2.39115864e-01 -4.92608845e-01
4.89517510e-01 1.81785122e-01 1.60273397e+00 1.30060151e-01
2.33059511e-01 1.31715760e-01 -3.17074537e-01 -3.77320766e-01
-3.05230081e-01 6.25480890e-01 -4.35164660e-01 -2.03220323e-01
6.10334814e-01 -6.58064663e-01 -6.73472703e-01 2.70589143e-01
-9.18605268e-01 5.39899729e-02 7.77454197e-01 6.61711037e-01
3.50492001e-01 -3.26555163e-01 3.82837027e-01 -1.51864159e+00
3.13789576e-01 -1.33218646e-01 -5.91049373e-01 9.41154510e-02
-1.04941702e+00 3.56695473e-01 4.92330402e-01 -1.35604620e-01
-1.14809334e+00 -2.88170308e-01 -5.41750371e-01 2.26079866e-01
-4.13670465e-02 5.98293483e-01 -4.57938254e-01 4.91392016e-01
4.26357627e-01 -2.51694419e-03 -2.11748585e-01 -6.62469387e-01
3.90344054e-01 5.37620425e-01 2.74095312e-02 -8.25045705e-01
5.58897316e-01 6.65566266e-01 -4.19383459e-02 -7.36898482e-01
-1.08839667e+00 -4.92587149e-01 -4.81709689e-01 5.14154971e-01
5.77480614e-01 -1.09319627e+00 -1.75930306e-01 1.77936301e-01
-1.27761066e+00 -2.53923982e-01 -2.95398235e-01 3.10417879e-02
-1.33485198e-01 5.56224346e-01 -6.77968800e-01 -4.98441100e-01
-6.39505863e-01 -1.25793862e+00 1.09081972e+00 -3.41669202e-01
-6.81506336e-01 -8.51501048e-01 1.40211312e-03 6.22018814e-01
6.57320857e-01 -8.26851949e-02 1.00546837e+00 -8.23479533e-01
-7.95379877e-01 -2.78654471e-02 -2.76968479e-01 4.85649973e-01
-4.06280458e-02 -8.58097821e-02 -1.08643019e+00 -3.38563055e-01
5.92692308e-02 -1.39486045e-01 7.76613235e-01 -7.41114840e-02
6.45856202e-01 -3.87739569e-01 4.93117869e-02 5.00552118e-01
1.50383711e+00 -3.08614790e-01 5.52720964e-01 6.43654406e-01
6.27292216e-01 7.14047074e-01 5.75779617e-01 2.49651119e-01
6.66002154e-01 8.01853597e-01 2.46141091e-01 -7.57125244e-02
-1.64803952e-01 -2.84106791e-01 6.52940452e-01 1.28052688e+00
2.49461122e-02 -1.02508105e-01 -7.27762282e-01 5.66071987e-01
-1.56096339e+00 -5.24220228e-01 -3.15053254e-01 2.22661519e+00
8.35046411e-01 6.68674469e-01 1.38380751e-01 1.49239361e-01
1.12002797e-01 4.87741381e-01 -3.15564759e-02 -7.88551509e-01
-2.61065662e-01 3.14996302e-01 4.89777446e-01 5.25432587e-01
-7.47829974e-01 1.00905740e+00 5.79344273e+00 6.85275733e-01
-1.27848887e+00 2.50135183e-01 5.80877542e-01 -9.95852575e-02
-6.44438326e-01 5.38373351e-01 -1.21443129e+00 3.74791443e-01
1.21619487e+00 1.08440325e-01 3.04987401e-01 6.97838128e-01
-1.48839936e-01 -1.89173102e-01 -1.09366286e+00 6.21934116e-01
1.59359664e-01 -9.83343542e-01 1.12989359e-01 2.30257988e-01
4.33108568e-01 3.54557961e-01 3.94907221e-03 4.10826504e-01
1.33511573e-01 -8.42144907e-01 5.73142052e-01 -2.72261468e-03
6.30391598e-01 -7.25522816e-01 9.54506099e-01 4.49981004e-01
-1.03006434e+00 3.47273797e-01 -3.28926355e-01 -8.36019441e-02
1.94080845e-01 7.30605125e-01 -9.36820090e-01 5.97343266e-01
4.09678429e-01 4.11031157e-01 -8.55474114e-01 4.97886449e-01
-4.54338700e-01 7.64353693e-01 -4.22825903e-01 -2.74713859e-02
4.40000534e-01 -1.67529657e-01 4.62437153e-01 1.51764762e+00
1.79647654e-01 -3.80051672e-01 -2.78992772e-01 4.20802504e-01
-4.86238264e-02 4.80441839e-01 -5.94342649e-01 -5.19806147e-02
6.78477064e-02 1.41902363e+00 -6.97974682e-01 -3.33558649e-01
-9.07044411e-01 7.74923742e-01 5.24231911e-01 7.35145062e-02
-6.57265067e-01 1.73050538e-01 6.11496508e-01 3.57587963e-01
5.36938071e-01 -6.16618581e-02 -5.81807554e-01 -1.36170185e+00
2.73841709e-01 -1.31160474e+00 6.35106862e-01 -4.32650417e-01
-9.57525253e-01 8.38280439e-01 -8.35731700e-02 -9.26410317e-01
-1.48645505e-01 -6.98687792e-01 -2.48936549e-01 6.84605479e-01
-1.92585540e+00 -1.10470986e+00 1.81314692e-01 1.45024419e-01
5.42763829e-01 6.53855726e-02 9.28051114e-01 3.64084244e-01
-4.20083433e-01 7.85115421e-01 -4.53530312e-01 -3.94399390e-02
8.19894016e-01 -1.36172616e+00 5.74972153e-01 9.97550905e-01
5.71597159e-01 9.64994729e-01 7.58406341e-01 -4.20521408e-01
-1.40179598e+00 -8.91354740e-01 1.45013666e+00 -1.09605253e+00
7.55373418e-01 -4.44758475e-01 -1.02031446e+00 7.46563792e-01
5.37558258e-01 -1.74070254e-01 7.55037904e-01 5.65999091e-01
-6.45385623e-01 -1.10927850e-01 -7.53481925e-01 5.98405063e-01
9.45330203e-01 -5.53188980e-01 -9.26165938e-01 2.46792682e-04
7.73025334e-01 -1.51545703e-01 -6.14487827e-01 6.56691611e-01
5.43388188e-01 -1.13185751e+00 7.16848910e-01 -3.88663203e-01
4.59822387e-01 -1.72049984e-01 -3.67074281e-01 -1.11745834e+00
7.58654922e-02 -4.99585450e-01 -5.04779667e-02 1.57946622e+00
9.75921035e-01 -7.98872471e-01 9.79233682e-01 8.20894957e-01
1.34830430e-01 -9.33264375e-01 -9.20406640e-01 -6.81407452e-01
2.58071929e-01 -7.47128487e-01 5.86893737e-01 7.16904342e-01
-6.88042417e-02 9.74490523e-01 -3.79740924e-01 3.53938863e-02
3.21526289e-01 3.79391462e-01 9.93251741e-01 -9.82051373e-01
-5.15628755e-01 -4.08946246e-01 7.80097097e-02 -9.86723661e-01
1.82692811e-01 -9.52775717e-01 -3.34369034e-01 -1.44169509e+00
3.21175218e-01 -5.85769892e-01 -4.53482449e-01 3.73136193e-01
-3.35263371e-01 1.31433025e-01 3.16346377e-01 4.91354354e-02
-6.30204499e-01 3.43864650e-01 7.99442410e-01 2.92584803e-02
9.23166424e-02 5.65650389e-02 -1.13216281e+00 8.57692182e-01
6.42967582e-01 -6.73941433e-01 -1.64851546e-01 -4.54019040e-01
8.29516053e-01 -3.79572719e-01 -6.58567995e-02 -4.86031055e-01
9.24402028e-02 2.00207502e-01 -2.78696388e-01 -5.88957131e-01
2.70993471e-01 -1.02854943e+00 -9.37315896e-02 2.80973345e-01
-5.08106910e-02 3.77465039e-01 3.76192808e-01 1.68873429e-01
-3.48226815e-01 -3.15219730e-01 5.30658782e-01 -1.53402194e-01
-4.32514071e-01 2.20851749e-02 -6.69670254e-02 3.73841256e-01
3.76169682e-01 8.76481757e-02 -2.65239507e-01 -1.45089841e-02
-1.29907548e-01 -1.29071221e-01 6.37883186e-01 3.41518462e-01
2.16347933e-01 -8.72604549e-01 -5.96696615e-01 2.10863724e-01
2.54409939e-01 -1.44013360e-01 -1.41607583e-01 8.93597841e-01
-1.96085632e-01 6.20848775e-01 4.14176911e-01 -4.79521304e-01
-1.47317278e+00 5.25418282e-01 -3.34278308e-02 -8.92882705e-01
-5.52071035e-01 8.21770847e-01 1.93366766e-01 -5.95835924e-01
4.31265645e-02 -6.52172685e-01 -4.70631495e-02 3.49946111e-01
3.07071269e-01 4.64139171e-02 5.85368752e-01 -6.23353362e-01
-4.81663197e-01 4.91498232e-01 -2.02724874e-01 -2.05428615e-01
1.51219606e+00 -1.16622530e-01 -2.81116188e-01 4.41093862e-01
1.12320828e+00 7.26629078e-01 -8.93282235e-01 -3.89648885e-01
6.31834447e-01 -2.88978457e-01 -1.03305392e-01 -1.00091076e+00
-9.76215363e-01 1.02478540e+00 2.57985502e-01 -1.34118617e-01
1.03946900e+00 -9.91973579e-02 9.03052330e-01 4.80496407e-01
5.17258704e-01 -9.53197479e-01 -1.13347530e-01 6.13629937e-01
7.41327047e-01 -1.29476190e+00 3.46727401e-01 -4.78813082e-01
-8.29862893e-01 6.80514634e-01 5.44017494e-01 -3.42781134e-02
5.32690048e-01 6.87060118e-01 1.60358354e-01 -2.17328325e-01
-1.18870926e+00 -1.83163315e-01 2.97299117e-01 1.35110067e-02
8.12733710e-01 2.03268789e-02 -5.16991496e-01 5.57038665e-01
-3.07288706e-01 -2.05676466e-01 4.13846016e-01 9.78432715e-01
-1.68300077e-01 -1.68797612e+00 -1.52817145e-01 3.20441335e-01
-9.80764747e-01 -5.67221880e-01 -3.03765863e-01 1.00132680e+00
-2.27169786e-02 7.62816668e-01 -3.26014042e-01 -3.84952761e-02
5.82018554e-01 2.32045308e-01 4.34270114e-01 -7.68979132e-01
-9.90160167e-01 1.61146134e-01 6.05908155e-01 -6.99456453e-01
-3.89053643e-01 -6.42643392e-01 -7.92228878e-01 -1.05384298e-01
-3.60336423e-01 3.34159464e-01 7.87111044e-01 9.66261625e-01
4.34862047e-01 2.10816711e-01 2.61735588e-01 -2.86312342e-01
-5.06470025e-01 -9.44517970e-01 -1.31205201e-01 3.08532864e-01
1.47790164e-01 -2.19997510e-01 -5.37277102e-01 -3.09640706e-01] | [11.357250213623047, 6.793923377990723] |
5918af26-9418-4816-bbce-990c3bc40e91 | a-general-recipe-for-the-analysis-of | 2303.06058 | null | https://arxiv.org/abs/2303.06058v1 | https://arxiv.org/pdf/2303.06058v1.pdf | A General Recipe for the Analysis of Randomized Multi-Armed Bandit Algorithms | In this paper we propose a general methodology to derive regret bounds for randomized multi-armed bandit algorithms. It consists in checking a set of sufficient conditions on the sampling probability of each arm and on the family of distributions to prove a logarithmic regret. As a direct application we revisit two famous bandit algorithms, Minimum Empirical Divergence (MED) and Thompson Sampling (TS), under various models for the distributions including single parameter exponential families, Gaussian distributions, bounded distributions, or distributions satisfying some conditions on their moments. In particular, we prove that MED is asymptotically optimal for all these models, but also provide a simple regret analysis of some TS algorithms for which the optimality is already known. We then further illustrate the interest of our approach, by analyzing a new Non-Parametric TS algorithm (h-NPTS), adapted to some families of unbounded reward distributions with a bounded h-moment. This model can for instance capture some non-parametric families of distributions whose variance is upper bounded by a known constant. | ['Junya Honda', 'Kazuya Suzuki', 'Dorian Baudry'] | 2023-03-10 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [-3.46100777e-02 1.51281267e-01 -7.21169472e-01 -2.33037889e-01
-9.06093121e-01 -1.02708924e+00 2.36400411e-01 1.95795566e-01
-3.65704536e-01 1.24309790e+00 -5.88928834e-02 -7.62586057e-01
-8.80662978e-01 -7.30941951e-01 -9.12682474e-01 -9.02187169e-01
1.11043835e-02 9.98460531e-01 8.93470719e-02 5.27673662e-02
2.66717523e-01 9.08240378e-01 -1.36000896e+00 -1.04578128e-02
7.39354730e-01 1.23178625e+00 -1.46793410e-01 7.73447335e-01
-2.25256577e-01 5.24628758e-01 -4.16075528e-01 -6.60488904e-01
3.64975214e-01 -6.67806804e-01 -9.69287455e-01 -7.95854777e-02
-2.57961363e-01 -2.89594948e-01 2.35574245e-01 1.01885748e+00
1.38527378e-01 1.61633402e-01 9.68204677e-01 -1.20562804e+00
-2.13108927e-01 1.08764994e+00 -7.23509669e-01 3.39580268e-01
1.66018426e-01 -4.63338792e-01 1.00102246e+00 3.23402852e-01
3.77799392e-01 1.07359838e+00 4.59840298e-01 4.24870014e-01
-1.28983998e+00 -3.96437228e-01 1.38265237e-01 1.30418584e-01
-9.50919271e-01 -1.84928030e-01 5.23340523e-01 -2.42061466e-01
3.15623432e-01 7.18948960e-01 5.60472548e-01 9.02393281e-01
-1.91582561e-01 9.37472403e-01 1.23292100e+00 -7.15657055e-01
6.27515495e-01 2.90181220e-01 4.15201992e-01 1.97710127e-01
8.93725276e-01 1.63315028e-01 3.64835225e-02 -4.86939192e-01
6.79643512e-01 -1.53238982e-01 -2.14736745e-01 -6.47330165e-01
-7.21888006e-01 1.03135288e+00 -2.45568193e-02 5.09883940e-01
-5.40265501e-01 1.70455784e-01 3.03364486e-01 3.17343980e-01
5.68678975e-01 1.70055609e-02 -4.83408183e-01 -1.56547695e-01
-8.98881376e-01 2.83696771e-01 1.20128834e+00 1.01004875e+00
5.67040265e-01 -3.56173337e-01 -5.91482937e-01 4.41214919e-01
-1.20934919e-01 5.92635214e-01 -1.04379822e-02 -5.72006464e-01
6.45870388e-01 -1.78562313e-01 1.10099268e+00 -3.56712788e-01
-4.28553075e-01 -6.95912898e-01 -4.86001343e-01 -1.82779834e-01
1.05305171e+00 -2.34865382e-01 -4.03622329e-01 1.82596898e+00
3.65269512e-01 -1.74055710e-01 -1.55619532e-01 7.60027647e-01
-2.35962749e-01 3.90968919e-01 -2.79890120e-01 -9.11862075e-01
9.03827608e-01 -7.30678976e-01 -6.58604085e-01 1.44025758e-01
6.66565716e-01 -5.31710804e-01 8.49812269e-01 6.01349235e-01
-1.14631438e+00 2.84401238e-01 -6.40118599e-01 4.78389889e-01
6.28219172e-02 1.13770001e-01 7.03395247e-01 1.33290291e+00
-4.91293728e-01 7.36558676e-01 -6.85026050e-01 -2.38303810e-01
3.12794417e-01 1.19034871e-01 7.36960247e-02 2.95648668e-02
-6.10159516e-01 5.04707456e-01 4.16201323e-01 8.17368552e-02
-7.52675056e-01 -3.53074282e-01 -1.38188750e-01 2.98793823e-01
4.05991197e-01 -7.19063580e-01 1.48690689e+00 -1.33821225e+00
-1.64445400e+00 6.59737408e-01 -1.67985976e-01 -7.14176416e-01
1.17289019e+00 -3.15610953e-02 5.70800975e-02 1.89403206e-01
-2.03410804e-01 -5.76032758e-01 8.14672053e-01 -9.08884943e-01
-4.51354593e-01 -5.41615009e-01 4.27878499e-01 -1.99251607e-01
-1.12129532e-01 1.41199395e-01 2.51168281e-01 -3.15040052e-01
-1.82654262e-02 -8.86484444e-01 -2.77221233e-01 -4.05984849e-01
-5.07546246e-01 -1.02500342e-01 -2.52744645e-01 -2.15689287e-01
1.09108627e+00 -1.82408726e+00 1.66605398e-01 5.47225356e-01
-6.65085733e-01 -1.51957512e-01 -1.14537962e-02 5.51953077e-01
-1.34332478e-02 1.49624839e-01 9.56661329e-02 -1.02495551e-01
3.63539040e-01 2.59810150e-01 -4.06017274e-01 8.25981736e-01
-4.09375191e-01 5.34275115e-01 -6.81934178e-01 -1.59525275e-01
-1.74117833e-01 -2.99225956e-01 -6.35558486e-01 8.54994357e-02
-5.75066388e-01 2.96324462e-01 -8.74958694e-01 2.50620216e-01
8.59781563e-01 -2.13010654e-01 2.83966988e-01 2.16268539e-01
-2.21208811e-01 -3.28056216e-02 -1.38671780e+00 1.01296604e+00
-4.22438294e-01 -5.20456955e-02 2.99877971e-01 -1.40479302e+00
5.98372698e-01 1.32022128e-01 4.66607749e-01 -1.91439673e-01
6.71684623e-01 3.95953655e-01 -2.23350674e-01 -3.74563009e-01
1.44165143e-01 -6.61769271e-01 3.58050093e-02 7.31412232e-01
-3.82631898e-01 3.15703332e-01 3.41812015e-01 -2.12196425e-01
9.40662622e-01 -2.04475254e-01 2.81171560e-01 -4.62805659e-01
3.93811435e-01 -2.24490479e-01 2.11002871e-01 1.32576764e+00
5.31249456e-02 3.18497300e-01 9.94991362e-01 -2.36836448e-01
-1.12608624e+00 -8.54067862e-01 -1.81317583e-01 1.07691824e+00
2.05443334e-02 2.69576430e-01 -6.58029675e-01 -5.54843187e-01
2.68632740e-01 1.04347980e+00 -9.81109440e-01 1.07763000e-01
6.56393729e-03 -9.67049479e-01 2.13986501e-01 2.22842529e-01
2.35468760e-01 -5.91094315e-01 -6.55798137e-01 1.08238973e-01
-6.29668757e-02 -8.14496040e-01 -2.23333508e-01 4.50342476e-01
-9.15493965e-01 -1.24826646e+00 -8.80403101e-01 -1.22066885e-01
3.63464952e-01 1.61194146e-01 8.57242525e-01 -4.40714180e-01
2.23474219e-01 5.04518390e-01 -6.05182171e-01 -5.40375710e-01
-2.48547107e-01 2.01656833e-01 -1.73230201e-01 2.51770496e-01
-4.00588103e-02 -4.97007459e-01 -4.27522689e-01 4.11991537e-01
-1.08850920e+00 -2.63659686e-01 5.25551856e-01 6.58902168e-01
6.23938024e-01 -1.31069748e-02 5.77689111e-01 -8.94701004e-01
5.64099848e-01 -6.86704457e-01 -1.01859391e+00 6.16202176e-01
-3.34356785e-01 4.43897247e-01 6.62334681e-01 -5.96159220e-01
-1.00342536e+00 -5.18441871e-02 1.42889798e-01 -2.56309390e-01
1.33302705e-02 5.02593279e-01 -9.28712785e-02 4.99111600e-02
5.41089714e-01 1.00925691e-01 -1.33321509e-01 -6.02993727e-01
3.46118897e-01 6.15290880e-01 1.24069594e-01 -1.14893126e+00
5.10828137e-01 5.66335857e-01 3.75553966e-01 -4.92300272e-01
-1.28990662e+00 -3.47101122e-01 -1.11202180e-01 1.19218707e-01
2.25800917e-01 -1.67870492e-01 -1.13818479e+00 1.23075746e-01
-9.22186971e-01 -3.89153570e-01 -3.20900500e-01 8.13655436e-01
-1.35343766e+00 4.05293912e-01 -4.30260122e-01 -1.66355371e+00
-1.37914449e-01 -6.43626213e-01 5.39587915e-01 8.46612230e-02
3.22517008e-01 -9.07827377e-01 1.58166587e-01 2.36144826e-01
3.71134341e-01 2.13688016e-01 9.34301674e-01 -9.00411308e-01
-3.94126356e-01 -1.40553564e-01 -2.08154023e-01 2.21168146e-01
-9.33183506e-02 -2.14200050e-01 -3.88523519e-01 -3.83970350e-01
-5.36676720e-02 -1.23217009e-01 7.30402946e-01 9.35478687e-01
1.35123432e+00 -6.37602389e-01 -4.11150187e-01 4.95336086e-01
1.47124362e+00 2.33133793e-01 3.66327941e-01 7.28600740e-01
-2.41594657e-01 3.71491432e-01 8.00247014e-01 9.63083923e-01
-1.94944441e-01 5.71623623e-01 6.44368410e-01 4.94467020e-01
8.67777288e-01 4.99857105e-02 3.77874752e-03 -9.66631845e-02
-2.85820305e-01 -4.05717522e-01 -3.35006803e-01 6.57393813e-01
-1.98005629e+00 -1.15046120e+00 -2.16651961e-01 2.81563067e+00
9.02503908e-01 1.46147832e-01 8.93516898e-01 1.80336401e-01
1.01321840e+00 -2.30080754e-01 -6.35459840e-01 -8.38389456e-01
-1.35568097e-01 3.87607008e-01 1.12426174e+00 5.18509150e-01
-8.69250298e-01 3.08838189e-01 6.75231409e+00 1.03373170e+00
-6.73680246e-01 1.66017205e-01 5.28359056e-01 -4.20714080e-01
-4.26025987e-01 6.98192641e-02 -7.36780882e-01 6.49501145e-01
9.77266490e-01 -5.80311477e-01 6.45106196e-01 1.04749727e+00
2.15057179e-01 -2.56161869e-01 -9.81865883e-01 7.46771932e-01
-4.85522479e-01 -1.13156474e+00 -1.91874936e-01 2.16869846e-01
7.80766308e-01 -4.16406006e-01 9.02618244e-02 1.13848858e-02
4.98345584e-01 -7.89522350e-01 7.95036376e-01 5.48745632e-01
5.10324478e-01 -1.14940047e+00 8.29628527e-01 6.90331280e-01
-4.91485864e-01 -3.57862025e-01 -4.55139071e-01 -3.68489623e-02
5.99123873e-02 8.08053017e-01 -2.92422354e-01 9.65427339e-01
3.12010705e-01 -9.54126194e-02 1.70912594e-01 1.64524257e+00
7.66020417e-02 7.30964839e-01 -9.27654922e-01 -5.00467718e-01
4.23514187e-01 -3.70382756e-01 5.36640227e-01 1.00231373e+00
6.31087840e-01 7.58599490e-02 -2.41732568e-01 4.96067703e-01
3.01474929e-01 4.37006742e-01 -1.17671221e-01 -1.71550792e-02
3.60843539e-01 8.48371148e-01 -7.84233630e-01 -1.23526432e-01
-1.82447061e-01 6.77199721e-01 2.75401831e-01 3.08120102e-01
-1.10120523e+00 -3.32576782e-01 4.13737237e-01 1.78447098e-01
8.32008660e-01 8.03552568e-02 -1.56245455e-01 -1.12681651e+00
1.47546560e-01 -4.54235733e-01 7.41182506e-01 -2.97696829e-01
-1.29217231e+00 3.10558170e-01 4.58428413e-01 -9.47011828e-01
-4.60231483e-01 -4.98050392e-01 -2.99392074e-01 7.16926038e-01
-1.52512145e+00 -8.95090222e-01 2.89833754e-01 7.86291718e-01
-1.49809673e-01 8.55562761e-02 5.45069337e-01 5.64697087e-02
-5.72704375e-01 7.00668633e-01 9.48094428e-01 -5.01663506e-01
9.45747495e-02 -1.10694313e+00 -5.38629293e-01 5.26801109e-01
-6.03606254e-02 3.30064833e-01 1.36722720e+00 -2.19754204e-01
-1.36290598e+00 -7.89336860e-01 3.64394993e-01 1.63560793e-01
9.29384828e-01 6.12152889e-02 -3.79642159e-01 9.65214372e-01
-2.39539236e-01 -1.00607477e-01 6.31072223e-01 3.06926697e-01
-2.56445318e-01 -1.89524472e-01 -1.46401918e+00 1.06168687e-01
9.53616142e-01 8.31814930e-02 -3.08545977e-01 6.45435572e-01
1.96903989e-01 -2.80143291e-01 -6.17896438e-01 1.87728882e-01
7.94348359e-01 -1.36548567e+00 5.21930695e-01 -1.08763027e+00
2.20193062e-02 7.82160312e-02 -2.45129421e-01 -1.17591715e+00
-2.62488663e-01 -1.02152967e+00 -6.49411008e-02 1.01748073e+00
2.58736253e-01 -9.07105625e-01 8.68111491e-01 5.21503627e-01
3.94151896e-01 -7.98111498e-01 -1.40042019e+00 -1.32769108e+00
3.65653247e-01 -5.47093332e-01 7.38996208e-01 4.61998135e-01
1.19323954e-01 -1.47730678e-01 -5.26643932e-01 1.13254055e-01
5.95734596e-01 6.73349977e-01 6.57277524e-01 -1.20672214e+00
-8.83816779e-01 -5.21873593e-01 -1.59924269e-01 -1.15725350e+00
2.67696828e-01 -4.52740043e-01 -3.92576396e-01 -1.24890435e+00
3.53679240e-01 -6.22923434e-01 -3.84045660e-01 8.42687190e-02
3.68570447e-01 -2.70490170e-01 -6.45939559e-02 -6.56068791e-03
-6.49080455e-01 3.92533571e-01 1.08852232e+00 1.52087703e-01
-2.03661889e-01 9.76281047e-01 -7.68135548e-01 3.62192512e-01
8.91945839e-01 -4.98522103e-01 -3.95549655e-01 -4.29183096e-02
5.79008281e-01 5.97917736e-01 1.09183639e-01 -5.74378610e-01
-1.34770930e-01 -6.33474946e-01 -9.09630731e-02 -7.39882648e-01
-3.43779288e-03 -8.99532437e-01 4.41652894e-01 4.35260117e-01
-5.33748746e-01 -4.11675930e-01 1.46459892e-01 7.11765528e-01
2.84033328e-01 -8.52508485e-01 8.22155416e-01 5.59110828e-02
3.47773969e-01 1.02822624e-01 -3.19392443e-01 -1.15828253e-01
1.19848561e+00 -9.24825743e-02 -3.21125269e-01 -7.40093648e-01
-1.01327872e+00 3.56083848e-02 1.67042613e-01 -3.23086947e-01
-4.88717854e-02 -1.17723680e+00 -4.87295240e-01 -1.64152592e-01
-1.42635405e-01 -5.36932588e-01 2.69944131e-01 1.24879289e+00
-3.79536778e-01 6.69418693e-01 -7.06587359e-02 -8.82080272e-02
-1.10234249e+00 1.04808617e+00 3.64101797e-01 -4.30392385e-01
-1.95632726e-02 4.65373188e-01 -1.14848129e-01 2.61894286e-01
1.67295888e-01 -5.51605523e-01 1.48144782e-01 -2.65388917e-02
4.93524015e-01 6.15390122e-01 6.02363348e-02 -1.29039764e-01
-1.47488907e-01 3.46042186e-01 -4.10043746e-02 -2.49055699e-01
1.42722619e+00 -1.85959056e-01 -1.36258677e-01 3.88337553e-01
8.43919814e-01 1.64054826e-01 -9.63887751e-01 -2.06107676e-01
1.48741715e-02 -4.18222398e-01 -1.53431877e-01 -7.06710339e-01
-9.28061306e-01 4.32398140e-01 2.41601050e-01 9.73856390e-01
1.20799327e+00 2.42467269e-01 3.12373459e-01 4.46062744e-01
7.93474793e-01 -8.95068169e-01 -6.42045200e-01 2.81919271e-01
8.14744711e-01 -5.62441766e-01 -3.16571966e-02 -2.94928312e-01
-2.50911742e-01 1.38042259e+00 -1.41747788e-01 -1.42639428e-01
5.45425355e-01 1.61722958e-01 -6.38175011e-01 4.99113470e-01
-5.30504942e-01 -6.58958972e-01 -3.11515126e-02 3.70852470e-01
1.09159715e-01 2.60647506e-01 -1.08697343e+00 9.25326288e-01
-2.60112852e-01 1.90580353e-01 5.02676427e-01 6.33463085e-01
-4.98189270e-01 -1.09438372e+00 -5.19933760e-01 5.48216462e-01
-8.74246299e-01 2.15131894e-01 -1.98497474e-01 8.31590235e-01
-5.62662244e-01 1.00276828e+00 -1.14090927e-01 1.84734672e-01
2.04313621e-01 -9.87684876e-02 1.03835142e+00 -1.63854524e-01
-2.20182434e-01 1.40323490e-01 3.21903348e-01 -1.90491244e-01
-4.89112407e-01 -7.95513809e-01 -6.23629272e-01 -6.23513341e-01
-6.52973294e-01 7.88203120e-01 5.88087499e-01 1.10786247e+00
-5.09629399e-02 -4.63357195e-02 7.80396879e-01 -5.74832320e-01
-1.35468209e+00 -9.96434748e-01 -1.19927776e+00 4.96099070e-02
4.03469920e-01 -7.04460204e-01 -6.44079685e-01 -5.98766267e-01] | [4.532642364501953, 3.2965569496154785] |
ff3485cf-eb6b-421a-bc6d-cae084080f26 | bevfusion-multi-task-multi-sensor-fusion-with | 2205.13542 | null | https://arxiv.org/abs/2205.13542v2 | https://arxiv.org/pdf/2205.13542v2.pdf | BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird's-Eye View Representation | Multi-sensor fusion is essential for an accurate and reliable autonomous driving system. Recent approaches are based on point-level fusion: augmenting the LiDAR point cloud with camera features. However, the camera-to-LiDAR projection throws away the semantic density of camera features, hindering the effectiveness of such methods, especially for semantic-oriented tasks (such as 3D scene segmentation). In this paper, we break this deeply-rooted convention with BEVFusion, an efficient and generic multi-task multi-sensor fusion framework. It unifies multi-modal features in the shared bird's-eye view (BEV) representation space, which nicely preserves both geometric and semantic information. To achieve this, we diagnose and lift key efficiency bottlenecks in the view transformation with optimized BEV pooling, reducing latency by more than 40x. BEVFusion is fundamentally task-agnostic and seamlessly supports different 3D perception tasks with almost no architectural changes. It establishes the new state of the art on nuScenes, achieving 1.3% higher mAP and NDS on 3D object detection and 13.6% higher mIoU on BEV map segmentation, with 1.9x lower computation cost. Code to reproduce our results is available at https://github.com/mit-han-lab/bevfusion. | ['Song Han', 'Daniela Rus', 'Huizi Mao', 'Xinyu Yang', 'Alexander Amini', 'Haotian Tang', 'Zhijian Liu'] | 2022-05-26 | null | null | null | null | ['scene-segmentation', '3d-multi-object-tracking'] | ['computer-vision', 'computer-vision'] | [-5.77804670e-02 -4.31333125e-01 -9.56964120e-02 -4.48129684e-01
-7.86128581e-01 -9.28248644e-01 6.20084763e-01 1.39702737e-01
-4.69253421e-01 1.50820902e-02 -1.93611339e-01 -3.30523729e-01
2.60129180e-02 -8.86053443e-01 -7.99251497e-01 -4.25487548e-01
5.66862285e-01 3.70156318e-01 8.22339892e-01 -6.07021034e-01
2.00288236e-01 7.69768953e-01 -2.19645166e+00 1.97116777e-01
7.26085067e-01 1.12923181e+00 5.70566714e-01 7.04799831e-01
-1.29278183e-01 5.06703444e-02 -2.48127744e-01 -3.76096487e-01
5.39970696e-01 3.14376175e-01 -3.44779074e-01 -8.81419480e-02
1.02818644e+00 -3.45164359e-01 -2.41718978e-01 1.09249473e+00
3.93521518e-01 -2.31179427e-02 4.42143381e-01 -1.63799214e+00
-1.58633977e-01 -1.45445675e-01 -8.87549579e-01 2.50426590e-01
2.05692038e-01 1.89249396e-01 9.35480177e-01 -1.01548982e+00
3.79587948e-01 1.33938372e+00 6.00330472e-01 2.39784434e-01
-1.08233154e+00 -7.48221636e-01 1.42899007e-01 1.54736519e-01
-1.55086768e+00 -5.19546032e-01 5.12226164e-01 -4.87577915e-01
1.07938325e+00 4.70856041e-01 5.87175488e-01 6.47228479e-01
3.14041734e-01 6.43096030e-01 1.09481251e+00 1.17757292e-02
-1.12206854e-01 1.05017588e-01 2.96685666e-01 6.70089066e-01
6.51341498e-01 1.77539647e-01 -7.59714246e-01 1.94419995e-01
5.64236403e-01 3.38847011e-01 3.21463235e-02 -6.35917425e-01
-1.32216656e+00 8.09668124e-01 6.17054582e-01 2.86824424e-02
-3.12456898e-02 2.66655982e-01 2.07930997e-01 2.27363631e-01
4.29028183e-01 2.66329013e-02 -3.61230344e-01 7.94443674e-03
-7.42656469e-01 3.86553466e-01 2.27309570e-01 1.23272550e+00
1.20703089e+00 -1.61137983e-01 1.53678417e-01 4.45834130e-01
4.20253575e-01 1.18283069e+00 -7.96682090e-02 -1.25752640e+00
4.37179118e-01 8.07467043e-01 -9.30112973e-02 -8.63379419e-01
-4.99673337e-01 -1.90028086e-01 -6.63914859e-01 6.05620682e-01
-5.45845740e-03 1.86691582e-01 -9.97893631e-01 1.21963501e+00
5.57948232e-01 5.72559386e-02 -4.15709615e-02 9.10958111e-01
1.03299534e+00 4.56328124e-01 -3.36227357e-01 5.02156436e-01
1.66977072e+00 -8.97108436e-01 -3.82903367e-01 -5.75334728e-01
4.37357128e-01 -8.39051485e-01 8.83630276e-01 1.55850872e-01
-7.64115632e-01 -7.16526508e-01 -1.33691716e+00 -5.38837075e-01
-6.73095822e-01 -1.04781747e-01 7.17478514e-01 7.24517167e-01
-1.22745776e+00 6.86465576e-02 -8.19463372e-01 -3.80527705e-01
4.90400046e-01 4.39036518e-01 -5.55262506e-01 -3.06889296e-01
-6.54136777e-01 9.56010759e-01 4.02513355e-01 -2.70169616e-01
-6.32598460e-01 -8.13437819e-01 -1.00433409e+00 -1.57114550e-01
6.43147945e-01 -9.28962231e-01 1.15663528e+00 -3.84979732e-02
-1.13216031e+00 1.03581142e+00 -4.36346024e-01 -4.01626617e-01
2.96785325e-01 -4.87614036e-01 -2.17853650e-01 5.13840280e-02
2.68528372e-01 1.05482256e+00 8.63403916e-01 -1.27188027e+00
-1.12480807e+00 -7.31184065e-01 1.14922211e-01 3.75745267e-01
4.17758264e-02 -1.93481699e-01 -5.69776177e-01 -7.30639771e-02
3.94836336e-01 -1.03583729e+00 -2.46371984e-01 1.45284206e-01
2.39335024e-03 -1.74194813e-01 1.27992153e+00 -1.11513296e-02
5.78701854e-01 -2.43069148e+00 3.28652648e-04 -8.05506781e-02
7.40634203e-01 1.37661576e-01 -3.53833474e-02 2.22283855e-01
4.81499076e-01 -4.44866680e-02 -1.84556022e-01 -6.40259564e-01
5.35898656e-03 3.24156106e-01 -3.35733533e-01 5.83864450e-01
7.42323548e-02 1.11748731e+00 -6.85826957e-01 -3.78921717e-01
8.60359251e-01 4.36938852e-01 -4.68338519e-01 -1.12642340e-01
-2.62607634e-02 2.76484311e-01 -3.62488002e-01 8.47597539e-01
1.15520680e+00 -1.63789671e-02 -4.26149666e-01 -3.68999928e-01
-4.99825448e-01 2.98907813e-02 -1.23662353e+00 2.03772378e+00
-3.94851476e-01 6.35050237e-01 3.70793521e-01 -4.87220675e-01
9.32781875e-01 -2.16247723e-01 5.50077975e-01 -7.10851967e-01
2.49473795e-01 3.87326151e-01 -4.73579317e-01 5.61872981e-02
9.05717731e-01 1.79266378e-01 -3.88858318e-01 -2.17009205e-02
1.57808706e-01 -8.90927374e-01 -3.84365022e-02 1.81863233e-01
8.30437541e-01 7.90422335e-02 2.49672636e-01 -3.28174770e-01
3.64220202e-01 2.36670807e-01 4.03924376e-01 6.24090672e-01
-1.83826849e-01 6.04352713e-01 -1.60252959e-01 -1.75578058e-01
-8.08015645e-01 -1.39644098e+00 -2.49191061e-01 9.63708520e-01
7.26727009e-01 -6.31055653e-01 -3.63553792e-01 -3.97828102e-01
5.04338384e-01 5.47938287e-01 -1.56412050e-01 -3.43031511e-02
-3.03926617e-01 -4.19947267e-01 3.52570027e-01 4.43533570e-01
6.60554051e-01 -2.37503603e-01 -1.12288678e+00 -1.24967210e-01
-6.77243769e-02 -1.44070256e+00 -1.88118234e-01 2.88753837e-01
-7.68115759e-01 -9.42836046e-01 -2.16714054e-01 -3.48711759e-01
2.22977355e-01 1.20263410e+00 1.02617681e+00 -3.14329267e-01
-2.01440960e-01 4.39901710e-01 -2.99218893e-01 -8.00188601e-01
3.13307121e-02 2.64912862e-02 1.21740207e-01 -2.45737687e-01
6.23851836e-01 -5.14291406e-01 -4.79281902e-01 3.59339207e-01
-6.69141829e-01 2.59828269e-01 5.69231808e-01 3.07114750e-01
9.61585343e-01 -2.90466577e-01 -8.28170255e-02 -4.21726674e-01
3.13783363e-02 -2.19572842e-01 -9.65015233e-01 -2.12120757e-01
-5.51811099e-01 -1.23353839e-01 3.42022628e-02 1.44349054e-01
-6.28013849e-01 3.47356737e-01 -1.83609635e-01 -7.74184644e-01
-3.55398446e-01 -1.81033969e-01 -1.98779598e-01 -3.45165908e-01
4.69334036e-01 6.78625256e-02 -2.97542010e-02 -5.16099870e-01
7.65278757e-01 7.75478899e-01 7.73647547e-01 -2.88075149e-01
9.65197742e-01 1.03515673e+00 2.92463273e-01 -8.52318048e-01
-7.38965571e-01 -1.00464511e+00 -9.21366632e-01 -2.44130641e-01
1.03023016e+00 -1.36105001e+00 -7.82891810e-01 4.52439189e-01
-1.13514328e+00 -5.03062978e-02 -3.21446449e-01 4.17804718e-01
-5.51514268e-01 9.41169634e-02 -1.06208704e-01 -5.43931425e-01
-2.29743257e-01 -1.30725169e+00 1.72990572e+00 1.59146801e-01
1.84121698e-01 -3.59332323e-01 -1.32136539e-01 5.22086203e-01
2.50450790e-01 2.98061967e-01 3.35964918e-01 -1.80580035e-01
-9.78483319e-01 -1.14273019e-01 -5.77384114e-01 2.77831554e-01
-1.12507395e-01 -9.73103195e-02 -1.37627041e+00 -2.83882201e-01
-2.36996822e-02 2.40908358e-02 1.11922228e+00 2.74114490e-01
7.78406620e-01 4.91383731e-01 -5.76599836e-01 9.83076632e-01
1.60064960e+00 -1.00789353e-01 2.14405507e-01 1.47358790e-01
1.02602458e+00 5.22372186e-01 8.10770810e-01 1.86839625e-01
9.39285576e-01 8.03367376e-01 8.39278460e-01 -1.41851440e-01
-4.25032169e-01 -1.41521022e-01 3.41405064e-01 6.73537910e-01
8.54931772e-03 -1.94403931e-01 -1.10676515e+00 5.17068565e-01
-1.80785382e+00 -5.55064917e-01 -6.25202417e-01 2.03813314e+00
1.35453165e-01 1.88900232e-01 2.92638261e-02 -9.89571959e-02
5.40592790e-01 4.24896598e-01 -5.32898366e-01 -2.20543206e-01
-2.75804639e-01 1.45977065e-01 1.15033686e+00 5.29301405e-01
-1.21228302e+00 1.12998998e+00 5.40238190e+00 8.02946568e-01
-1.05699241e+00 5.79920709e-01 1.36618474e-02 -3.44226301e-01
-2.76048988e-01 -7.60570318e-02 -1.25221038e+00 2.26027802e-01
8.55260193e-01 -9.36301202e-02 1.09962210e-01 8.36828530e-01
-4.60870825e-02 -3.01608562e-01 -8.94416869e-01 1.49582648e+00
1.36081383e-01 -1.51423883e+00 -1.15786623e-02 4.24882919e-01
5.25364935e-01 9.23676372e-01 6.78016022e-02 -2.57803388e-02
3.11619967e-01 -6.54082716e-01 1.06066704e+00 2.26158589e-01
9.28098738e-01 -7.45465338e-01 6.21350586e-01 5.30990005e-01
-1.65373611e+00 2.04699598e-02 -4.11641598e-01 -1.25783965e-01
2.98731089e-01 6.88775539e-01 -4.53567058e-01 8.16286623e-01
1.05294108e+00 8.57752919e-01 -8.82950604e-01 7.89487958e-01
8.70577842e-02 -1.36211306e-01 -7.52636850e-01 2.24990875e-01
2.82840729e-01 -2.53134698e-01 7.74119854e-01 1.01769114e+00
5.87115943e-01 -7.94524625e-02 3.12177122e-01 5.96800625e-01
7.50616938e-03 -2.93886662e-01 -1.10287082e+00 4.97659028e-01
5.69021761e-01 1.34521890e+00 -7.90756404e-01 -2.63138652e-01
-6.24901116e-01 1.00621176e+00 -7.03708604e-02 -6.45664632e-02
-9.08532739e-01 -3.65169525e-01 1.13488460e+00 1.71236739e-01
5.40256143e-01 -7.03833699e-01 -6.53974295e-01 -1.06067216e+00
1.27945393e-01 -3.06961268e-01 5.87144606e-02 -7.24372268e-01
-8.68755102e-01 4.73727405e-01 1.57999218e-01 -1.29636443e+00
1.09558932e-01 -7.14674830e-01 -1.43614262e-01 7.30640709e-01
-1.84092367e+00 -1.45990276e+00 -7.23168790e-01 7.26071417e-01
5.15153885e-01 1.08465537e-01 5.00643969e-01 3.77782971e-01
-2.03370407e-01 1.12200134e-01 -8.57430696e-02 -4.04276788e-01
6.56098604e-01 -1.17976987e+00 8.33994150e-01 1.06374848e+00
2.52998054e-01 1.56262308e-01 4.50785577e-01 -4.46888089e-01
-1.80144048e+00 -1.20625043e+00 5.82799256e-01 -1.07483888e+00
4.83578444e-01 -7.80576706e-01 -5.56823850e-01 5.95957696e-01
2.95082815e-02 1.18767120e-01 3.92740369e-01 -8.58247280e-02
-4.81367797e-01 -4.51712519e-01 -1.10938966e+00 4.16590750e-01
1.42368317e+00 -5.64758241e-01 -4.80611026e-01 1.87372535e-01
1.06792521e+00 -6.81645572e-01 -7.62664676e-01 7.62411416e-01
4.19326067e-01 -1.15493739e+00 1.32752109e+00 1.87938377e-01
-1.74160853e-01 -7.97894239e-01 -7.59694815e-01 -9.04721081e-01
-3.65181059e-01 -2.98498422e-01 -1.24406904e-01 8.73967826e-01
6.51554391e-02 -8.28068614e-01 6.50525451e-01 -2.51451228e-03
-6.05466545e-01 -3.71754438e-01 -1.19440782e+00 -8.43177199e-01
-2.67202020e-01 -8.56009722e-01 8.89617145e-01 6.00956798e-01
-6.75770283e-01 5.33798099e-01 5.47570176e-02 5.49210727e-01
7.55554199e-01 4.41045463e-01 1.09868741e+00 -1.51829636e+00
2.21081927e-01 -5.71226895e-01 -7.09758162e-01 -1.27456307e+00
-1.80108756e-01 -1.04557443e+00 -5.79716414e-02 -1.55880487e+00
-1.23167835e-01 -3.45081896e-01 -4.66953777e-02 4.18455392e-01
5.91374971e-02 5.86017847e-01 5.66046238e-01 1.76022112e-01
-7.42389381e-01 4.53687638e-01 1.02720392e+00 -6.93603382e-02
-2.32002363e-02 -1.05075017e-01 -6.79557085e-01 7.93830872e-01
9.10607934e-01 -3.81185412e-01 -4.06939149e-01 -7.97048688e-01
1.24842644e-01 -2.91159153e-01 7.01631546e-01 -1.32376838e+00
3.02441567e-01 -9.94450897e-02 1.76145539e-01 -1.18143916e+00
8.98723006e-01 -9.66584444e-01 1.29961625e-01 4.14071530e-01
5.54558694e-01 3.09296221e-01 4.91060227e-01 5.39939702e-01
-2.08035082e-01 1.85362265e-01 7.89079905e-01 -3.55905108e-02
-1.30556250e+00 4.59215671e-01 -6.39673546e-02 -2.64364064e-01
1.30166495e+00 -5.50655067e-01 -5.95700264e-01 1.82352439e-01
-2.05240622e-01 4.31898236e-01 8.81261110e-01 6.67629719e-01
7.68588185e-01 -1.23432672e+00 -7.23494470e-01 5.49072385e-01
5.38930178e-01 5.12342513e-01 4.67389792e-01 8.95399928e-01
-5.47994912e-01 6.79178894e-01 -2.57066220e-01 -1.18916750e+00
-1.40836084e+00 4.97663021e-01 1.11183681e-01 3.06807816e-01
-7.47175395e-01 8.48762751e-01 2.91213334e-01 -5.33891082e-01
-2.69313991e-01 -6.21774971e-01 9.88852084e-02 1.03795953e-01
4.17300075e-01 4.70740467e-01 3.55454087e-01 -9.18717861e-01
-6.40369177e-01 1.01055861e+00 1.53358400e-01 -5.09270048e-03
1.14236331e+00 -4.80452389e-01 -1.78910419e-02 5.77596486e-01
1.08755088e+00 1.90188423e-01 -1.39184618e+00 -1.56354681e-01
-2.39397854e-01 -6.79805338e-01 5.40636957e-01 -3.71616155e-01
-8.57216477e-01 1.01023388e+00 7.85968781e-01 6.35704622e-02
9.28499341e-01 3.65707904e-01 8.92371535e-01 3.23975176e-01
8.99546027e-01 -8.03749204e-01 -2.78925389e-01 6.04589581e-01
6.70917392e-01 -1.54121315e+00 8.37726817e-02 -7.36430168e-01
-5.52218139e-01 6.90269828e-01 5.44832647e-01 -6.50613829e-02
6.53122842e-01 4.68689293e-01 6.98135793e-03 -5.79852223e-01
-4.51386213e-01 -7.09952414e-01 3.16065192e-01 7.71666288e-01
-2.95965403e-01 3.54181200e-01 3.63201171e-01 1.90529481e-01
-3.11131597e-01 -3.41293007e-01 2.22408921e-01 9.36155915e-01
-7.83946514e-01 -9.35260773e-01 -4.32062060e-01 5.09146512e-01
1.98225617e-01 8.03707074e-03 -1.46030948e-01 7.84114301e-01
5.13053954e-01 1.03166270e+00 3.31180573e-01 -7.64973760e-01
6.20090663e-01 -1.91869646e-01 4.95318502e-01 -7.40443349e-01
-1.94622949e-01 2.68597044e-02 -1.33339837e-01 -1.02610362e+00
-4.17342156e-01 -7.80603468e-01 -1.14666343e+00 -5.99708319e-01
-2.61063933e-01 -4.04217541e-01 1.02475917e+00 5.33449113e-01
7.84040391e-01 4.23732191e-01 4.63001549e-01 -1.17777872e+00
-1.50752649e-01 -4.60224211e-01 -4.84423220e-01 1.91716570e-02
5.27815163e-01 -1.01977360e+00 -1.59468904e-01 -2.09235907e-01] | [7.834085941314697, -2.5369772911071777] |
68405ff9-f46a-408f-91d3-50da74b1366d | active-inference-in-hebbian-learning-networks | 2306.05053 | null | https://arxiv.org/abs/2306.05053v2 | https://arxiv.org/pdf/2306.05053v2.pdf | Active Inference in Hebbian Learning Networks | This work studies how brain-inspired neural ensembles equipped with local Hebbian plasticity can perform active inference (AIF) in order to control dynamical agents. A generative model capturing the environment dynamics is learned by a network composed of two distinct Hebbian ensembles: a posterior network, which infers latent states given the observations, and a state transition network, which predicts the next expected latent state given current state-action pairs. Experimental studies are conducted using the Mountain Car environment from the OpenAI gym suite, to study the effect of the various Hebbian network parameters on the task performance. It is shown that the proposed Hebbian AIF approach outperforms the use of Q-learning, while not requiring any replay buffer, as in typical reinforcement learning systems. These results motivate further investigations of Hebbian learning for the design of AIF networks that can learn environment dynamics without the need for revisiting past buffered experiences. | ['Gert Cauwenberghs', 'Georges Gielen', 'Francky Catthoor', 'André Bourdoux', 'Ilja Ocket', 'Lars Keuninckx', 'Tim Verbelen', 'Ali Safa'] | 2023-06-08 | null | null | null | null | ['q-learning', 'openai-gym'] | ['methodology', 'playing-games'] | [ 2.28611007e-01 1.46679834e-01 2.80299067e-01 -1.19393758e-01
3.34432960e-01 -2.85856068e-01 8.12903762e-01 -2.81561673e-01
-6.23749197e-01 1.07490730e+00 2.00885981e-02 -5.10193557e-02
-5.56641877e-01 -8.86255980e-01 -1.05955470e+00 -1.21430099e+00
-1.86181724e-01 6.88496172e-01 3.53963405e-01 -3.77876401e-01
2.87862182e-01 4.77407932e-01 -1.80028224e+00 -4.68106847e-03
8.62949610e-01 5.60803890e-01 7.06170857e-01 9.94475842e-01
3.73034686e-01 1.10983109e+00 -6.71126246e-01 3.58434558e-01
2.66865015e-01 -7.42249489e-01 -3.60751599e-01 -1.11123897e-01
-1.27975568e-01 -2.80614972e-01 -4.20948744e-01 7.14886129e-01
3.86436164e-01 7.15245783e-01 6.15498483e-01 -1.08553243e+00
-3.77461880e-01 7.16342032e-01 3.61754030e-01 4.75856155e-01
-2.59119384e-02 8.05363774e-01 5.86647987e-01 -1.91919953e-01
4.50387239e-01 1.11800587e+00 2.49736860e-01 8.43213737e-01
-1.56166887e+00 -6.35149300e-01 2.04208106e-01 4.24093485e-01
-1.00687969e+00 -3.63888919e-01 5.19403577e-01 -3.73803884e-01
1.22962141e+00 2.74632359e-03 1.42810500e+00 1.43519759e+00
8.84482145e-01 5.52895904e-01 1.59369779e+00 -4.27229375e-01
9.80199099e-01 -8.06982070e-02 -6.54395744e-02 5.75873256e-01
2.37599209e-01 9.49929893e-01 -8.79436076e-01 -8.29586685e-02
9.99399602e-01 -5.51908426e-02 -1.30249009e-01 -6.39114439e-01
-8.74712944e-01 6.54938400e-01 4.38666552e-01 1.98018998e-01
-9.23182368e-01 6.95469499e-01 1.55649073e-02 4.99051929e-01
-2.50732526e-02 6.79941595e-01 -2.65594959e-01 -2.52674043e-01
-6.39123678e-01 5.74139476e-01 8.13887954e-01 3.98327172e-01
7.87330449e-01 5.98803520e-01 -2.58072615e-01 3.87573212e-01
4.68241811e-01 6.12013519e-01 6.30113542e-01 -1.14922845e+00
-3.32568496e-01 3.70885104e-01 2.30310127e-01 -4.62273210e-01
-1.68657392e-01 -3.91945988e-01 -3.83496612e-01 6.56252742e-01
1.52277038e-01 -3.99003714e-01 -1.12703812e+00 1.74172664e+00
1.71828212e-03 5.40263772e-01 6.24648146e-02 6.39806271e-01
7.57961348e-02 7.32903123e-01 4.59602475e-02 -2.63480902e-01
5.78353226e-01 -7.43805587e-01 -8.48506808e-01 -4.12688732e-01
-5.30544017e-03 9.77115110e-02 8.33531320e-01 5.48956156e-01
-1.20316839e+00 -7.00630128e-01 -1.11163747e+00 8.19906294e-01
-3.54974359e-01 -4.14281428e-01 5.99846244e-01 6.49857283e-01
-1.47614312e+00 8.34398389e-01 -1.41280162e+00 -3.95268768e-01
1.37047708e-01 5.70361853e-01 1.74194440e-01 3.39799255e-01
-1.19794858e+00 1.33356893e+00 8.85785758e-01 3.31756085e-01
-1.85395920e+00 -3.17520469e-01 -3.46236050e-01 1.95553839e-01
3.10668707e-01 -9.55768824e-01 1.17660642e+00 -9.72476423e-01
-2.32796025e+00 1.98657915e-01 7.69091845e-02 -9.02723014e-01
2.11058170e-01 -3.13641310e-01 -8.11631326e-03 -1.44553021e-01
-5.55274367e-01 7.86211312e-01 9.26056564e-01 -1.23630130e+00
-2.93920845e-01 -2.73070276e-01 -1.05524719e-01 3.64852130e-01
2.26884931e-02 -7.68780231e-01 2.40723208e-01 -3.10860991e-01
-1.77452579e-01 -1.44527006e+00 -4.89839435e-01 -5.50736785e-01
1.82009831e-01 -3.48250926e-01 4.76774454e-01 -1.91517547e-01
8.62480760e-01 -1.82608902e+00 6.28119349e-01 4.74941343e-01
-1.99268162e-01 3.24936211e-01 -1.77917525e-01 6.40246451e-01
7.12440237e-02 -5.55838287e-01 -1.10084236e-01 1.94221422e-01
-4.23086202e-03 8.61707091e-01 -5.44214070e-01 2.61640362e-03
-1.00830436e-01 1.02424657e+00 -8.29198182e-01 1.55439720e-01
1.31229445e-01 2.50477999e-01 -5.77307045e-01 5.06752312e-01
-6.58303082e-01 7.72956789e-01 -2.27382660e-01 -2.92267591e-01
-6.09914623e-02 6.91007078e-02 4.75247949e-01 4.33924615e-01
-1.69831187e-01 2.31384277e-01 -9.78303611e-01 1.69791186e+00
-3.99504840e-01 4.91699070e-01 -3.65710445e-02 -1.25714588e+00
1.21521318e+00 4.27554280e-01 3.07766706e-01 -7.42944062e-01
3.76079857e-01 9.86061096e-02 7.44903386e-01 -5.03700078e-01
1.52813837e-01 -1.51934683e-01 2.21716121e-01 6.36960387e-01
5.80934167e-01 -4.44655031e-01 3.43214810e-01 -1.36360992e-02
1.19494820e+00 6.25976861e-01 -1.01732470e-01 -4.70575958e-01
3.31850886e-01 -1.58977076e-01 5.61833203e-01 1.18150795e+00
-1.73725337e-01 -2.24552751e-01 3.04809242e-01 -5.46716273e-01
-7.66203523e-01 -1.38624477e+00 3.40747803e-01 1.19323039e+00
1.33272141e-01 -1.23100080e-01 -7.93471456e-01 7.48579949e-02
-2.41142645e-01 1.10270941e+00 -8.61947596e-01 -8.28896761e-01
-6.24676406e-01 -8.12146246e-01 -3.30694485e-03 4.65538889e-01
3.41398358e-01 -1.85027289e+00 -1.45948243e+00 5.83036304e-01
2.23513395e-01 -3.81956756e-01 1.11080773e-01 7.16329336e-01
-1.22471452e+00 -6.47550285e-01 -2.35582203e-01 -4.06226456e-01
4.19900686e-01 -4.31684613e-01 9.49422598e-01 -9.81861129e-02
-1.66337281e-01 6.63463771e-01 -7.87932947e-02 -3.71203333e-01
-6.02878034e-01 3.47421989e-02 3.26362193e-01 -2.81416297e-01
2.28238598e-01 -1.11728823e+00 -6.52778327e-01 3.48655181e-03
-8.26956511e-01 1.91563368e-01 7.89989769e-01 1.11617720e+00
5.39012015e-01 2.67505884e-01 6.24446511e-01 -5.89111924e-01
8.30669224e-01 -2.41930991e-01 -7.19847679e-01 2.75296867e-01
-9.33640420e-01 5.98237038e-01 2.70437211e-01 -6.26016617e-01
-1.43164980e+00 6.81012310e-03 9.17695165e-02 -1.56971246e-01
-3.41222942e-01 3.55246872e-01 2.21955433e-01 1.88376367e-01
7.63375580e-01 8.00707996e-01 1.00918368e-01 -1.20971520e-02
3.63664746e-01 1.47568867e-01 4.64993387e-01 -8.17195773e-01
3.81041676e-01 -1.07473256e-02 -1.24772973e-01 -6.00313962e-01
-4.08014238e-01 1.11075222e-01 -4.73245710e-01 -4.94823277e-01
9.54986393e-01 -5.85052192e-01 -9.79416132e-01 6.14020348e-01
-7.63893306e-01 -1.07867932e+00 -6.73976243e-01 4.94510919e-01
-1.35313940e+00 -5.30503571e-01 -7.06605077e-01 -1.02953076e+00
-2.65025705e-01 -1.04829776e+00 1.26523107e-01 6.97620928e-01
-2.83096820e-01 -8.33784759e-01 7.57598698e-01 1.72045171e-01
6.63700461e-01 -1.49766743e-01 8.63927424e-01 -4.32851285e-01
-8.08167458e-01 2.47049838e-01 7.55102634e-01 1.77921191e-01
-1.97793275e-01 -1.36770397e-01 -8.14700842e-01 -5.26667774e-01
1.81070969e-01 -4.50361967e-01 8.49782228e-01 5.73147058e-01
7.66381562e-01 -3.24298829e-01 -1.19407825e-01 6.69574663e-02
1.18785095e+00 8.11322272e-01 7.76026905e-01 3.41092944e-01
1.41823977e-01 3.47225904e-01 6.43086284e-02 3.44816655e-01
1.54463023e-01 2.84008831e-01 6.18641794e-01 5.73292017e-01
1.35841772e-01 -2.91206360e-01 7.70587802e-01 9.42043960e-01
-5.44454873e-01 -7.81008899e-02 -7.19735920e-01 4.44231540e-01
-2.07056022e+00 -1.25767684e+00 5.29986918e-01 2.07032609e+00
9.73367274e-01 4.09931362e-01 1.11652337e-01 -1.36247337e-01
3.17544341e-01 -1.21701695e-01 -1.00571990e+00 -5.77261925e-01
1.95501238e-01 5.68449557e-01 1.43966913e-01 4.34361666e-01
-5.23344696e-01 8.95490706e-01 6.62583828e+00 4.04364884e-01
-1.18645298e+00 8.64959955e-02 2.01294735e-01 -2.13924855e-01
-1.36884868e-01 5.17461039e-02 -6.41709268e-01 4.79910940e-01
1.57134354e+00 -4.63189669e-02 1.01772881e+00 5.81759155e-01
3.34713787e-01 -5.50177753e-01 -8.89749944e-01 4.59173977e-01
-2.09596127e-01 -1.31103122e+00 -2.50987440e-01 4.33759540e-02
8.91873419e-01 2.33731851e-01 2.26232350e-01 5.34126997e-01
1.10934412e+00 -8.90789509e-01 8.08753908e-01 1.20647931e+00
-1.09132022e-01 -6.39785230e-01 3.74228984e-01 8.67558599e-01
-5.34351289e-01 -6.30356133e-01 -2.18235895e-01 -4.30984735e-01
-6.52924925e-02 -7.90401734e-03 -9.73048568e-01 -6.29428998e-02
6.56360984e-01 3.57231438e-01 -4.15808648e-01 1.03762281e+00
-4.32356209e-01 9.41917539e-01 -2.48217836e-01 -5.71164489e-01
1.90819055e-01 -4.34535623e-01 6.48100495e-01 5.64301252e-01
3.33352476e-01 1.55476958e-01 3.77013236e-02 1.23653591e+00
4.43595797e-01 -5.65222979e-01 -5.48871696e-01 -3.10495566e-03
4.45861787e-01 8.73695970e-01 -8.70099247e-01 -3.77397269e-01
4.35751975e-01 7.75678039e-01 4.02961850e-01 6.40186906e-01
-5.19768953e-01 5.52379675e-02 3.47085565e-01 6.83176294e-02
5.38821518e-01 -4.34827149e-01 1.79981142e-02 -9.98106599e-01
-7.44056642e-01 -8.31259727e-01 -1.20388359e-01 -1.00643134e+00
-8.34085405e-01 6.16545796e-01 3.22902501e-01 -5.34845114e-01
-9.53070521e-01 -3.78666103e-01 -7.44769514e-01 8.10841441e-01
-8.50682855e-01 -4.72292662e-01 -1.14070423e-01 6.34224772e-01
4.41600055e-01 -3.72545540e-01 9.91295457e-01 -5.63169241e-01
-7.66281545e-01 3.32842208e-03 2.37785831e-01 -3.23444068e-01
2.20286041e-01 -1.34298944e+00 9.52673890e-03 7.41777122e-01
3.11076641e-01 7.46038675e-01 1.12430286e+00 -6.17761195e-01
-1.29558957e+00 -6.68137133e-01 1.64806936e-02 -3.84448260e-01
3.38821530e-01 -3.28778654e-01 -9.41628516e-01 6.81985736e-01
7.03283370e-01 -2.09502250e-01 4.72691596e-01 6.59911521e-03
1.78130448e-01 -3.66085976e-01 -7.32437789e-01 7.53454924e-01
8.44464123e-01 -4.76172715e-01 -8.76911521e-01 -1.49994165e-01
2.22720042e-01 -1.88240036e-01 -5.87285638e-01 1.50765732e-01
6.60792887e-01 -9.52524781e-01 7.29559422e-01 -5.44430494e-01
-3.58286612e-02 -2.99974322e-01 2.17891589e-01 -2.02803564e+00
-6.11049116e-01 -5.38248003e-01 -5.44437170e-01 4.74188834e-01
1.87945977e-01 -6.30259931e-01 8.05364251e-01 5.15729487e-01
-1.27635136e-01 -6.03562474e-01 -8.75097513e-01 -7.69806862e-01
4.73512001e-02 9.75414440e-02 3.21205020e-01 2.94072419e-01
-2.35948741e-01 4.55468744e-01 -2.90788621e-01 -1.67735592e-01
4.35578376e-01 -1.09016504e-02 6.47407413e-01 -1.05149984e+00
-6.95186496e-01 -3.77581596e-01 -1.02030799e-01 -6.63420141e-01
2.73396015e-01 -6.18534744e-01 7.21309423e-01 -1.31687593e+00
-4.38652560e-02 -3.38306874e-01 -6.41042829e-01 4.22831029e-01
-4.27175872e-02 -2.15991810e-01 2.39689544e-01 5.56954066e-04
-4.56322730e-01 8.92020881e-01 1.27233016e+00 1.34082079e-01
-4.92806256e-01 6.29962608e-02 -2.04837009e-01 6.89242005e-01
9.28113163e-01 -5.29041171e-01 -1.03102744e+00 -2.40539342e-01
3.38353634e-01 2.62336433e-01 4.18870151e-01 -1.37517297e+00
3.97610575e-01 -4.38885003e-01 4.10648376e-01 -3.68566185e-01
3.98955822e-01 -7.49987245e-01 5.74552536e-01 1.16919339e+00
-6.10933244e-01 2.35402554e-01 1.04688644e-01 1.00932038e+00
1.92917526e-01 -2.19022453e-01 6.51777387e-01 -2.72028029e-01
-7.98202693e-01 5.28517626e-02 -1.23777366e+00 -2.97127515e-01
1.01420391e+00 -3.29768866e-01 -1.34699821e-01 -3.05407852e-01
-1.14956367e+00 3.43371779e-02 6.11970350e-02 2.65963197e-01
6.17698848e-01 -1.06510758e+00 -4.57004368e-01 5.26319742e-01
-5.15477300e-01 -2.10977018e-01 2.15491548e-01 7.23818779e-01
-3.59668642e-01 2.74919957e-01 -9.28342104e-01 -4.70324755e-01
-6.24315381e-01 2.86658376e-01 6.50538027e-01 -4.60966468e-01
-4.14487123e-01 6.89980090e-01 -3.72453555e-02 -2.97457039e-01
6.17177933e-02 -1.83982387e-01 -2.67520934e-01 -2.08180398e-01
2.39135027e-01 3.78362447e-01 -5.03735580e-02 -8.88214409e-02
1.27846271e-01 -2.20218435e-01 -3.06671835e-03 -6.07562363e-01
1.54034781e+00 1.65372323e-02 -1.00418925e-01 9.25372303e-01
1.18035384e-01 -8.90931666e-01 -1.71920478e+00 -4.90027033e-02
-1.04638729e-02 2.37103272e-02 1.93297774e-01 -1.01864588e+00
-8.16810668e-01 9.23192322e-01 9.32874084e-01 1.58888429e-01
1.05771184e+00 -4.09147143e-01 3.69851679e-01 7.81429052e-01
6.56707644e-01 -1.33442831e+00 5.63547492e-01 8.09591532e-01
9.86612439e-01 -6.74361885e-01 -2.58432239e-01 6.39367819e-01
-5.71937859e-01 9.66009974e-01 9.95098114e-01 -8.33835363e-01
7.21405029e-01 1.63857594e-01 -2.15784490e-01 -1.30840614e-01
-1.49360693e+00 -1.11773320e-01 1.15574293e-01 6.69254661e-01
-3.33141349e-02 1.53538659e-01 -5.79747483e-02 1.90348476e-01
-3.41138303e-01 3.60730588e-01 5.34548521e-01 1.01232982e+00
-6.94084287e-01 -9.31792736e-01 -2.52441496e-01 5.02961159e-01
1.42911986e-01 2.96768606e-01 3.56581286e-02 5.39255142e-01
3.35219800e-01 5.67939043e-01 3.63379329e-01 -2.22410679e-01
7.67223462e-02 3.25358212e-01 1.03528535e+00 -6.26753092e-01
-7.26202011e-01 6.73248917e-02 -2.84726888e-01 -5.57409883e-01
-3.72858107e-01 -8.22225153e-01 -1.09742355e+00 -1.03434682e-01
-3.12036842e-01 2.19192013e-01 3.82177621e-01 9.38546896e-01
2.75147825e-01 8.71881664e-01 2.66873926e-01 -9.46945786e-01
-7.09748924e-01 -1.24363530e+00 -5.28745711e-01 1.91631749e-01
9.06101763e-02 -8.81632030e-01 -3.55250426e-02 2.99679041e-01] | [4.255686283111572, 1.5647941827774048] |
d32ac06d-3a95-4242-a0c6-1987da7c443f | knowledge-guided-metric-learning-for-few-shot | 2004.01907 | null | https://arxiv.org/abs/2004.01907v1 | https://arxiv.org/pdf/2004.01907v1.pdf | Knowledge Guided Metric Learning for Few-Shot Text Classification | The training of deep-learning-based text classification models relies heavily on a huge amount of annotation data, which is difficult to obtain. When the labeled data is scarce, models tend to struggle to achieve satisfactory performance. However, human beings can distinguish new categories very efficiently with few examples. This is mainly due to the fact that human beings can leverage knowledge obtained from relevant tasks. Inspired by human intelligence, we propose to introduce external knowledge into few-shot learning to imitate human knowledge. A novel parameter generator network is investigated to this end, which is able to use the external knowledge to generate relation network parameters. Metrics can be transferred among tasks when equipped with these generated parameters, so that similar tasks use similar metrics while different tasks use different metrics. Through experiments, we demonstrate that our method outperforms the state-of-the-art few-shot text classification models. | ['Kang Liu', 'Dianbo Sui', 'Yubo Chen', 'Jun Zhao', 'Delai Qiu', 'Binjie Mao'] | 2020-04-04 | null | https://aclanthology.org/2021.naacl-main.261 | https://aclanthology.org/2021.naacl-main.261.pdf | naacl-2021-4 | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 1.99571878e-01 2.05762461e-01 -1.76928803e-01 -3.56379002e-01
-6.62079602e-02 -8.00435767e-02 8.87198150e-01 3.56134355e-01
-8.22229505e-01 6.29996061e-01 3.34763080e-02 9.89531949e-02
4.65460792e-02 -1.16220403e+00 -1.93450227e-01 -4.36399788e-01
6.18134260e-01 5.93366623e-01 4.47470188e-01 -4.74687427e-01
2.71810085e-01 -1.30176976e-01 -1.71456873e+00 3.65046747e-02
9.93983328e-01 9.13878858e-01 3.38659197e-01 3.18455160e-01
-5.38564384e-01 7.85539389e-01 -7.18188822e-01 -7.23613262e-01
-1.56496912e-01 -6.85082078e-01 -9.49119329e-01 -5.66017255e-02
-2.36175522e-01 -8.40526596e-02 -2.29753196e-01 9.93030250e-01
5.35749137e-01 6.61067784e-01 9.07608807e-01 -1.26350164e+00
-6.93801105e-01 8.53300631e-01 -2.26775911e-02 5.83336428e-02
7.48074651e-02 1.41740561e-01 1.10575747e+00 -8.38283837e-01
7.97657728e-01 9.27158117e-01 5.61636508e-01 8.07723224e-01
-1.10436749e+00 -4.86089051e-01 6.07538084e-03 5.00553966e-01
-1.08812332e+00 -2.80607551e-01 8.44604731e-01 -4.48699951e-01
7.94577837e-01 -2.70164400e-01 6.22746050e-01 1.36811614e+00
-7.27384165e-02 8.43098581e-01 7.03551292e-01 -6.84034407e-01
5.86091757e-01 3.44634712e-01 6.39381528e-01 6.10727429e-01
3.93022895e-01 -2.28728592e-01 -5.69152594e-01 2.03268558e-01
3.06419134e-01 3.62751752e-01 -3.74815136e-01 -4.96300668e-01
-1.10302389e+00 1.01567459e+00 5.40273726e-01 8.32052171e-01
-2.40596220e-01 -1.19968809e-01 7.03251719e-01 3.87994826e-01
6.08959138e-01 9.16893899e-01 -3.14790398e-01 -1.30132228e-01
-6.61164999e-01 -1.79110825e-01 9.08080161e-01 8.47168326e-01
8.54180574e-01 -3.07012908e-02 -5.16929150e-01 1.11954308e+00
-5.45760132e-02 4.90922891e-02 1.08490431e+00 -5.06634593e-01
3.35528463e-01 7.40466237e-01 -1.09178610e-01 -7.79607534e-01
-4.93346214e-01 -4.85895902e-01 -1.01029873e+00 -2.01293960e-01
2.09312394e-01 -1.44991264e-01 -8.38172078e-01 1.57887971e+00
1.50089741e-01 3.01403925e-02 3.17255497e-01 7.09852278e-01
9.34970200e-01 4.49184060e-01 2.04363301e-01 -1.86769634e-01
1.31259012e+00 -1.17430580e+00 -8.63471329e-01 -4.16383237e-01
9.88923848e-01 -1.22126468e-01 1.32037997e+00 4.77968901e-03
-6.85797930e-01 -7.46757448e-01 -1.13168097e+00 8.89336914e-02
-8.24397326e-01 -1.55507818e-01 5.57191610e-01 5.57790637e-01
-4.58391547e-01 8.78951311e-01 -4.96986091e-01 -6.34065092e-01
7.06410527e-01 -6.33729771e-02 -1.53912663e-01 -1.19755432e-01
-1.69217455e+00 1.16047537e+00 8.92916799e-01 -3.79262269e-01
-6.47370815e-01 -6.15139902e-01 -9.43912506e-01 3.32793713e-01
6.05595171e-01 -9.36141610e-01 1.20948887e+00 -8.31248760e-01
-1.55747902e+00 8.44428718e-01 1.80869773e-01 -5.13343215e-01
6.48749411e-01 1.25255517e-03 -1.71412528e-01 1.64917573e-01
9.97036770e-02 5.69420934e-01 8.93352509e-01 -1.04501522e+00
-5.35323858e-01 -2.45389909e-01 3.08671277e-02 5.74451461e-02
-1.04491389e+00 -1.85993299e-01 -2.29215696e-01 -6.02967858e-01
-3.99558187e-01 -6.73250139e-01 -1.79604411e-01 -1.45744175e-01
-3.37075382e-01 -6.61273658e-01 5.19238293e-01 -5.47042899e-02
1.11367524e+00 -1.75638485e+00 1.64019257e-01 -3.67287137e-02
2.95476764e-01 7.22973883e-01 -1.92885607e-01 2.40598500e-01
3.02808404e-01 8.59184340e-02 -1.10260285e-01 -2.32691213e-01
1.78960845e-01 1.95750862e-01 -1.54248610e-01 -8.02267045e-02
8.49912688e-03 1.11209106e+00 -1.22991979e+00 -4.07226562e-01
2.18954906e-01 1.68822259e-01 -2.51535833e-01 3.74167055e-01
-2.84136742e-01 1.02266453e-01 -6.17129385e-01 2.46570826e-01
1.71823844e-01 -5.37286043e-01 1.02172852e-01 3.44346911e-02
3.51839364e-01 6.70154318e-02 -7.02186346e-01 1.72279441e+00
-7.65974760e-01 5.46792686e-01 -5.30770004e-01 -1.54091060e+00
1.15753222e+00 3.44501138e-01 -5.42558283e-02 -5.84896863e-01
8.01783085e-01 6.34349734e-02 3.73190939e-02 -5.57040751e-01
2.20144629e-01 -3.66742373e-01 1.64148025e-02 7.81276166e-01
6.32047772e-01 -3.09215993e-01 5.61243057e-01 2.05385059e-01
1.19423187e+00 -8.95125940e-02 5.91488004e-01 -1.24947552e-03
5.11614203e-01 -1.14313155e-01 3.73390645e-01 8.96059573e-01
-3.32520366e-01 2.63730139e-01 4.23596144e-01 -5.96284091e-01
-9.95679915e-01 -5.15453041e-01 -1.07346565e-01 1.51783729e+00
6.23941869e-02 -4.31939185e-01 -8.26280594e-01 -9.47210014e-01
-2.19510540e-01 9.57448542e-01 -9.77229595e-01 -7.07608342e-01
4.79331203e-02 -6.46853209e-01 2.85416484e-01 5.77891052e-01
6.08036339e-01 -1.32296550e+00 -9.00326252e-01 3.46391439e-01
-1.75805733e-01 -1.21210480e+00 -1.54676616e-01 4.83891875e-01
-7.82611668e-01 -1.02101326e+00 -1.13753474e+00 -8.17867935e-01
5.63483953e-01 4.97862369e-01 1.21773362e+00 2.98222691e-01
-3.58195573e-01 3.40201378e-01 -8.24347079e-01 -5.83436787e-01
-5.04077435e-01 4.83501881e-01 9.21314582e-02 8.97405595e-02
9.17878568e-01 -5.24580657e-01 -2.67330587e-01 4.08441514e-01
-1.02647388e+00 2.55497396e-01 3.60438704e-01 1.28514016e+00
1.36716560e-01 2.24571209e-02 8.17452192e-01 -1.11189246e+00
8.33428860e-01 -4.72934783e-01 -1.05272748e-01 4.29758489e-01
-6.97651744e-01 2.60353267e-01 1.04953849e+00 -7.94455290e-01
-1.17353964e+00 -3.11236054e-01 2.27474406e-01 -4.93384212e-01
-2.10388318e-01 6.33966684e-01 -3.29106972e-02 2.58334190e-01
1.05152941e+00 1.28846392e-01 -9.72636715e-02 -2.96293765e-01
5.98491013e-01 8.32785368e-01 1.10505849e-01 -3.82048070e-01
6.33337200e-01 3.18997115e-01 -2.68159121e-01 -8.45895469e-01
-1.52467561e+00 -5.06669521e-01 -9.09759939e-01 -1.18231453e-01
8.05310488e-01 -4.56084281e-01 -4.01944190e-01 5.41423857e-01
-1.14627612e+00 -5.08386731e-01 -4.73292589e-01 4.47100729e-01
-4.84243602e-01 1.48580983e-01 -5.35498977e-01 -6.36977553e-01
-4.39098716e-01 -7.43189871e-01 6.39744163e-01 3.84924144e-01
-2.78831154e-01 -1.21710813e+00 1.04299426e-01 2.66050190e-01
5.45770466e-01 -2.03118935e-01 9.48223114e-01 -1.31328070e+00
7.03507587e-02 -2.44776443e-01 -1.41454473e-01 3.06258738e-01
1.08072720e-01 -4.98750031e-01 -1.06019437e+00 -4.96433899e-02
1.90999880e-01 -8.32850158e-01 1.19711316e+00 -2.31869724e-02
1.13559461e+00 -2.33472437e-02 -4.07573074e-01 3.19359004e-01
8.85102034e-01 -5.33716902e-02 5.08702159e-01 5.30074298e-01
7.24263489e-01 9.92923737e-01 6.32062137e-01 5.29785752e-01
2.78033137e-01 6.20073855e-01 1.61875952e-02 2.62750626e-01
1.52186712e-03 -1.75467536e-01 -4.52515744e-02 8.61197710e-01
-1.80440456e-01 -2.08493263e-01 -1.06881046e+00 4.06729311e-01
-2.00944638e+00 -9.40908194e-01 2.56948203e-01 1.90157223e+00
1.07021451e+00 2.56030440e-01 -1.83843508e-01 3.21166903e-01
9.22199547e-01 1.49968013e-01 -6.84384704e-01 -1.20493799e-01
1.16894402e-01 3.34243536e-01 -4.51932773e-02 -7.35426992e-02
-1.10056877e+00 1.16250348e+00 5.20063543e+00 1.00117171e+00
-8.43727708e-01 2.88723230e-01 4.48232859e-01 9.40837935e-02
-1.04900219e-01 -1.35718405e-01 -7.89813042e-01 4.25895035e-01
9.92954612e-01 -6.43085659e-01 4.01740149e-02 1.03659165e+00
-3.34448546e-01 2.45478123e-01 -1.15902364e+00 9.15815711e-01
4.95701432e-01 -1.11325741e+00 2.76010275e-01 -2.49911547e-01
7.47653306e-01 -2.76753187e-01 -2.65001625e-01 8.21642339e-01
3.34088385e-01 -7.69495666e-01 2.73763210e-01 5.81352830e-01
4.76649314e-01 -7.42620170e-01 1.02291191e+00 5.91816843e-01
-8.45408320e-01 -1.67447418e-01 -9.20334578e-01 -1.15693577e-01
5.33096865e-02 6.67148292e-01 -8.91234636e-01 3.23971689e-01
3.72457355e-01 8.09249103e-01 -6.50135934e-01 8.72424304e-01
-6.66417301e-01 3.34914327e-01 1.84892759e-01 -5.98815143e-01
1.58110410e-01 -7.49317855e-02 1.37623608e-01 1.00171840e+00
3.20388496e-01 1.51207581e-01 1.60006180e-01 9.85551775e-01
-4.33788598e-01 3.98396790e-01 -8.02421153e-01 -1.18106015e-01
2.67677754e-01 1.44069171e+00 -8.85313451e-01 -7.55305409e-01
-4.55948621e-01 1.05587232e+00 6.34178936e-01 1.11168712e-01
-5.19125521e-01 -8.41834307e-01 1.24695301e-01 -2.41708070e-01
2.45335326e-01 1.38700023e-01 -6.02452420e-02 -1.30461359e+00
-2.75094599e-01 -6.51368856e-01 3.74385953e-01 -6.97339475e-01
-1.72802496e+00 7.43958890e-01 -2.05750629e-01 -1.22686994e+00
-3.91696602e-01 -6.22479796e-01 -7.00977027e-01 4.30450797e-01
-1.37915885e+00 -9.62493539e-01 -6.33463025e-01 5.62096655e-01
6.77705467e-01 -5.83893239e-01 9.71676052e-01 -1.62620574e-01
-5.37363470e-01 7.03398108e-01 -2.83908308e-03 4.17546511e-01
7.90953279e-01 -1.09968615e+00 4.95571166e-01 5.51510096e-01
3.23276073e-01 4.94491309e-01 4.99218702e-01 -4.59787697e-01
-9.70793128e-01 -1.15057182e+00 7.96534777e-01 -4.50564444e-01
8.56148362e-01 -3.06194097e-01 -1.28517723e+00 3.56934756e-01
4.42550704e-02 2.07683280e-01 1.03374088e+00 2.96747953e-01
-5.76986790e-01 2.02665314e-01 -8.26051712e-01 7.00442374e-01
1.18626165e+00 -4.36531901e-01 -1.14606285e+00 4.37179089e-01
7.41905868e-01 1.66275293e-01 -7.50524700e-01 1.09733559e-01
3.60100567e-01 -8.25861573e-01 7.45695472e-01 -9.67510819e-01
5.87476313e-01 2.26253211e-01 1.69031769e-01 -1.81249690e+00
-2.71020025e-01 -2.24202275e-01 -2.93640047e-01 1.25547791e+00
3.93535167e-01 -8.27092648e-01 5.84942222e-01 7.29438543e-01
-2.68537793e-02 -4.13986951e-01 -5.85453689e-01 -1.18493021e+00
1.27598107e-01 -3.26540709e-01 4.47505444e-01 1.40757644e+00
4.98065650e-01 1.03410685e+00 -3.43451232e-01 -6.85815215e-01
5.22045135e-01 1.37959108e-01 8.06644142e-01 -1.98235822e+00
-1.15610614e-01 -5.25705636e-01 -4.38209355e-01 -7.02263892e-01
6.90510929e-01 -1.11471581e+00 1.35535583e-01 -1.48027635e+00
5.64943731e-01 -3.00364852e-01 -4.98686880e-01 6.72542810e-01
-6.12369895e-01 2.31343165e-01 3.81945312e-01 1.73949897e-01
-9.05974627e-01 1.02564049e+00 1.28027856e+00 -2.93568850e-01
-2.72373229e-01 9.89862345e-03 -6.49514437e-01 8.84155333e-01
9.29784298e-01 -4.62261260e-01 -4.68565941e-01 -6.14242777e-02
2.66690642e-01 -1.82226688e-01 3.19761783e-02 -1.16784143e+00
4.05972451e-01 -1.94668602e-02 3.79629523e-01 -6.22027218e-02
1.56720161e-01 -6.23214006e-01 -4.77401197e-01 5.01604140e-01
-6.61413491e-01 -6.73288286e-01 -2.79553056e-01 6.25276864e-01
-2.39140868e-01 -8.89075756e-01 8.69494736e-01 -3.11429799e-01
-7.56175160e-01 3.86792302e-01 -3.15333903e-01 4.13012743e-01
9.15750563e-01 9.97730531e-03 -5.42785048e-01 -3.93714696e-01
-6.54819965e-01 1.91887170e-01 4.30024505e-01 7.65890777e-01
4.92357671e-01 -1.43506444e+00 -4.74685758e-01 -8.23953673e-02
7.06679344e-01 -2.94018596e-01 2.38130927e-01 6.35021269e-01
-8.31312407e-03 3.21562767e-01 -2.70669341e-01 -3.55187297e-01
-9.46794331e-01 9.21180308e-01 1.01682499e-01 -3.85596067e-01
-7.43243575e-01 5.55458963e-01 1.96780786e-02 -4.53225881e-01
3.08148086e-01 -2.53793448e-02 -6.37950242e-01 6.01831734e-01
9.46367025e-01 2.45063782e-01 8.97212625e-02 -2.67259032e-01
1.53830238e-02 3.90168995e-01 -1.99140668e-01 6.90322891e-02
1.32541537e+00 1.08364865e-01 3.09719861e-01 6.32676303e-01
1.15572953e+00 -7.44191885e-01 -9.15464759e-01 -6.75775290e-01
7.21869245e-02 -4.17771429e-01 -2.82055438e-02 -5.07793248e-01
-8.57071698e-01 1.29120016e+00 2.94755101e-01 4.50483471e-01
6.48619056e-01 -7.49724880e-02 8.77501547e-01 9.97675776e-01
5.31561196e-01 -1.51639140e+00 6.45588219e-01 9.17950690e-01
5.01130342e-01 -1.49921823e+00 -2.31663376e-01 -2.59023011e-02
-8.46105099e-01 1.19506776e+00 7.68945336e-01 1.43088102e-01
6.00949347e-01 -2.39129812e-01 -1.28545552e-01 -5.11857159e-02
-8.01858962e-01 -6.55804515e-01 2.02376485e-01 7.11728752e-01
3.55837137e-01 -1.83046609e-01 -4.40796942e-01 9.04845476e-01
-1.24711819e-01 1.78291202e-01 5.31938791e-01 9.56488371e-01
-8.36514890e-01 -1.08491302e+00 7.87640139e-02 7.00795054e-01
-1.99815363e-01 -1.89195737e-01 -2.52913326e-01 4.87307221e-01
-1.70895502e-01 9.44335759e-01 -1.25827463e-02 -3.17214668e-01
1.99289024e-01 5.81117868e-01 2.32261106e-01 -1.02311397e+00
-3.87043446e-01 -6.36869967e-01 1.09537356e-02 -2.89953351e-01
-4.08049047e-01 -2.90032495e-02 -1.17069113e+00 -1.37721580e-02
-4.75644916e-01 3.16833586e-01 4.33380336e-01 1.18554235e+00
1.44576937e-01 7.77412295e-01 5.93627691e-01 -9.42485988e-01
-6.28349364e-01 -1.24271333e+00 -6.05373979e-01 8.67597580e-01
-2.64710426e-01 -1.01143909e+00 -6.56013429e-01 4.14012484e-02] | [10.22469425201416, 3.585304021835327] |
49df928f-a0a0-43e6-9dad-e19fc7c65b00 | gcn-alp-addressing-matching-collisions-in | 2103.10600 | null | https://arxiv.org/abs/2103.10600v1 | https://arxiv.org/pdf/2103.10600v1.pdf | GCN-ALP: Addressing Matching Collisions in Anchor Link Prediction | Nowadays online users prefer to join multiple social media for the purpose of socialized online service. The problem \textit{anchor link prediction} is formalized to link user data with the common ground on user profile, content and network structure across social networks. Most of the traditional works concentrated on learning matching function with explicit or implicit features on observed user data. However, the low quality of observed user data confuses the judgment on anchor links, resulting in the matching collision problem in practice. In this paper, we explore local structure consistency and then construct a matching graph in order to circumvent matching collisions. Furthermore, we propose graph convolution networks with mini-batch strategy, efficiently solving anchor link prediction on matching graph. The experimental results on three real application scenarios show the great potentials of our proposed method in both prediction accuracy and efficiency. In addition, the visualization of learned embeddings provides us a qualitative way to understand the inference of anchor links on the matching graph. | ['Xueqi Cheng', 'HuaWei Shen', 'Shanshan Lyu', 'Yongqing Wang', 'Hao Gao'] | 2021-03-19 | null | null | null | null | ['anchor-link-prediction'] | ['graphs'] | [-2.96770543e-01 3.89414161e-01 -6.52097344e-01 -4.66068178e-01
6.35140240e-02 -3.67715538e-01 2.52166599e-01 6.40223384e-01
8.47882554e-02 3.77944738e-01 1.38754115e-01 -2.75460690e-01
-6.99429452e-01 -1.16259682e+00 -4.87004846e-01 -1.64576277e-01
-5.89257061e-01 3.88959557e-01 1.85800746e-01 -3.01856905e-01
1.18457772e-01 2.18295395e-01 -1.12033308e+00 6.84456453e-02
1.04169619e+00 1.09016645e+00 1.41280517e-01 2.10316226e-01
-6.56894147e-01 5.31865537e-01 1.83915123e-02 -9.46038425e-01
2.04871252e-01 9.43286791e-02 -7.08843529e-01 -1.23514637e-01
3.97194296e-01 -1.72218665e-01 -5.40192962e-01 1.30987167e+00
3.29879194e-01 6.28880337e-02 2.89900333e-01 -1.66295969e+00
-9.57518339e-01 9.29596663e-01 -4.70519900e-01 1.15717247e-01
6.74065530e-01 -5.01327813e-01 1.70362401e+00 -6.85538292e-01
6.60882115e-01 1.02177298e+00 1.00787151e+00 -4.09886725e-02
-1.04076540e+00 -6.83731258e-01 4.09712702e-01 4.54239786e-01
-1.55549288e+00 -4.78312895e-02 9.03067470e-01 -4.96392816e-01
4.57043827e-01 2.65372753e-01 8.20596635e-01 6.12848520e-01
-7.66444877e-02 5.06623864e-01 4.09376591e-01 -9.72170383e-02
-3.17424387e-01 3.04525465e-01 3.61051857e-01 1.07866967e+00
3.53825063e-01 -3.10773879e-01 -5.24405479e-01 -4.15964544e-01
5.87585509e-01 4.82372284e-01 -4.01399285e-01 -5.34916580e-01
-8.78829777e-01 8.26851964e-01 8.30723405e-01 4.12325084e-01
4.89578955e-02 -1.15606681e-01 3.59940171e-01 5.79254568e-01
5.93599856e-01 -4.19620499e-02 -2.81338930e-01 4.24270511e-01
-6.52393520e-01 -1.92590170e-02 1.02953720e+00 1.30129647e+00
9.95738864e-01 -5.09036601e-01 1.12847336e-01 7.76519716e-01
6.88093662e-01 -1.08836154e-02 3.76059592e-01 -6.38410628e-01
6.65292978e-01 1.13131213e+00 -5.17376810e-02 -1.83249962e+00
-5.88991106e-01 -5.51052213e-01 -8.81242394e-01 -2.74452418e-01
3.80538970e-01 2.18382366e-02 -1.08976113e-02 1.59034991e+00
4.10553992e-01 6.16366148e-01 -3.64463091e-01 7.53544033e-01
1.14113331e+00 3.18761736e-01 -6.63413480e-02 -2.35568315e-01
9.68308449e-01 -1.03508818e+00 -8.88345838e-01 2.47990623e-01
1.03644824e+00 -7.68535912e-01 8.82798851e-01 -1.24156930e-01
-8.36477399e-01 -5.87056875e-01 -1.00130391e+00 3.97658758e-02
-5.32415152e-01 -2.07843706e-01 1.11122417e+00 2.64900744e-01
-9.60849822e-01 1.08386183e+00 -3.23143154e-01 -6.15784407e-01
3.19714963e-01 5.71229577e-01 -4.02898699e-01 6.62765950e-02
-1.55840874e+00 4.57736760e-01 4.24616277e-01 1.50040433e-01
1.30451426e-01 -7.36685038e-01 -7.63668954e-01 3.08887541e-01
2.25305796e-01 -5.46522737e-01 7.97936618e-01 -8.83384645e-01
-9.83927906e-01 7.70356357e-01 1.45954834e-02 -2.24404350e-01
6.27354860e-01 2.56296229e-02 -1.05652452e+00 -1.88090518e-01
-3.48194577e-02 1.19700260e-01 3.57778221e-01 -1.05833101e+00
-7.16346681e-01 -1.99139118e-01 2.73224682e-01 1.18974656e-01
-8.95444214e-01 -2.93036520e-01 -6.89426303e-01 -4.13779408e-01
4.25889254e-01 -6.21401966e-01 2.60410402e-02 3.29624265e-01
-3.70101035e-01 -6.23319209e-01 7.25419939e-01 -6.09200358e-01
1.73379934e+00 -1.96578825e+00 1.91072319e-02 7.65556335e-01
6.82895958e-01 -1.12424698e-02 -1.59789458e-01 8.02083135e-01
-2.05816090e-01 2.65516549e-01 3.10803652e-01 -3.16679180e-01
9.64734107e-02 9.17974338e-02 -2.18334883e-01 4.20542032e-01
-4.28426951e-01 9.25814688e-01 -9.04794633e-01 -8.89163911e-01
-1.65993050e-01 2.41794020e-01 -5.12142599e-01 2.99880505e-01
-1.32073229e-03 2.53942519e-01 -6.37098610e-01 5.23010015e-01
7.33996451e-01 -6.19935513e-01 5.10323763e-01 -6.36884034e-01
1.44625500e-01 3.15616317e-02 -1.37360263e+00 1.89436483e+00
-3.36562067e-01 4.60886985e-01 2.12439179e-01 -1.07678783e+00
1.11586452e+00 6.72350749e-02 8.40888321e-01 -4.21031803e-01
1.54640719e-01 1.31449223e-01 -2.57447809e-01 -8.09873700e-01
3.60941023e-01 5.66673696e-01 2.95767158e-01 4.32127297e-01
6.56395257e-02 6.77948475e-01 1.20640844e-01 6.19388521e-01
8.52201819e-01 -7.34285712e-02 5.24656139e-02 -1.96686834e-01
4.91579115e-01 -4.80023116e-01 3.13793331e-01 4.50479120e-01
-3.03440671e-02 1.08582966e-01 4.95531231e-01 -5.50253272e-01
-8.13624322e-01 -9.32918847e-01 -9.32198539e-02 1.00961566e+00
5.84337831e-01 -7.36259699e-01 -4.60573614e-01 -7.91358709e-01
3.01410884e-01 1.60622239e-01 -5.62522709e-01 -1.43754691e-01
-4.27338898e-01 -5.32520890e-01 1.12767547e-01 2.56452054e-01
4.20000821e-01 -6.76534474e-01 6.90462947e-01 1.32816717e-01
-2.53315568e-01 -9.73663330e-01 -6.77797019e-01 -4.52288419e-01
-8.89772177e-01 -1.32681072e+00 -1.15860067e-01 -1.00408757e+00
8.15295875e-01 5.37082255e-01 1.10515857e+00 7.69121349e-01
1.74810756e-02 2.60890126e-01 -2.68032342e-01 2.52239019e-01
9.89634097e-02 3.01969916e-01 1.90872595e-01 4.01143044e-01
5.60736060e-01 -1.01726198e+00 -6.14507914e-01 5.23270369e-01
-6.68266594e-01 -7.22341314e-02 2.61315227e-01 5.86816728e-01
2.04992190e-01 9.25881714e-02 6.79448605e-01 -1.04429209e+00
9.42442060e-01 -1.01044083e+00 -5.53218484e-01 5.14471829e-01
-1.04230011e+00 -2.17734173e-01 3.28783542e-01 -4.09211338e-01
-5.41596234e-01 -2.88316071e-01 -1.65953697e-03 -2.31215701e-01
4.59326416e-01 9.55325007e-01 -2.62308329e-01 -2.99892038e-01
4.04536188e-01 -1.33446410e-01 1.72057301e-01 -6.27636313e-01
4.28722054e-01 6.96861684e-01 -8.89515132e-03 -3.33409816e-01
9.38400567e-01 3.92278492e-01 7.70107657e-02 -2.64696062e-01
-8.14621270e-01 -5.96874356e-01 -7.11861312e-01 -3.95735562e-01
5.06998658e-01 -7.57195234e-01 -1.25583494e+00 -1.79984599e-01
-1.05531359e+00 3.85925584e-02 1.82841212e-01 3.49770427e-01
-1.95030287e-01 7.36969948e-01 -5.70203304e-01 -4.10587758e-01
-2.73209810e-01 -5.68971157e-01 4.18185502e-01 1.52359769e-01
-1.76904216e-01 -1.53793645e+00 3.81745435e-02 2.81851351e-01
3.46750677e-01 -2.39658356e-02 8.21779132e-01 -8.74425948e-01
-8.01397681e-01 -4.94225591e-01 -7.92943120e-01 -1.54151618e-01
3.51719350e-01 7.28487298e-02 -6.11673355e-01 -3.16371948e-01
-6.47031486e-01 1.83569029e-01 4.95254189e-01 2.69448478e-02
1.36075616e+00 -4.32215273e-01 -7.42306113e-01 6.79797709e-01
1.37091577e+00 -3.59098643e-01 2.75489658e-01 1.34601995e-01
8.90027344e-01 7.78743386e-01 4.91246790e-01 4.61850494e-01
5.56544840e-01 6.99750662e-01 5.99674046e-01 -1.21833965e-01
1.45580158e-01 -7.11891174e-01 -7.74326324e-02 1.18053770e+00
-3.38020660e-02 -4.16678578e-01 -5.25425136e-01 2.99748927e-01
-2.38207579e+00 -8.34927917e-01 -5.11197329e-01 2.07747793e+00
4.54681158e-01 1.68181986e-01 1.39351234e-01 -1.19249865e-01
1.19659472e+00 -1.04200523e-02 -2.37101257e-01 6.03681356e-02
1.38931006e-01 -3.88371289e-01 5.15522420e-01 7.72703648e-01
-8.57995033e-01 5.05258083e-01 5.40711355e+00 8.43963325e-01
-7.83904791e-01 2.27225602e-01 1.52721897e-01 2.89480001e-01
-6.68277442e-01 5.94987199e-02 -8.85408521e-01 5.62341452e-01
4.37111437e-01 -2.40118012e-01 5.15574396e-01 7.24866450e-01
1.18057564e-01 5.44624567e-01 -1.06360674e+00 9.49507535e-01
-1.56260431e-01 -1.59244335e+00 -2.72591412e-02 5.01023643e-02
6.03992105e-01 -3.64972875e-02 -8.32281187e-02 1.33477628e-01
3.31233531e-01 -7.20316350e-01 1.33067638e-01 7.93445230e-01
6.45402431e-01 -5.41663229e-01 7.67675400e-01 2.54198521e-01
-1.62134802e+00 -4.02452834e-02 -3.38610649e-01 -1.74138695e-01
2.54303664e-01 6.62744761e-01 -8.57629001e-01 7.76962459e-01
6.64206684e-01 1.13319397e+00 -5.73989570e-01 1.20536077e+00
5.81037030e-02 3.29948604e-01 -3.73823345e-01 -1.72567725e-01
1.14383005e-01 -5.52085340e-01 3.26827914e-01 9.10577893e-01
3.69151592e-01 -1.42821848e-01 4.07235950e-01 7.79600441e-01
-3.97030562e-01 6.28359616e-01 -8.14645052e-01 -5.99445440e-02
6.75174892e-01 1.48577654e+00 -5.52868247e-01 -3.06217298e-02
-8.87561917e-01 7.40106225e-01 8.00247848e-01 4.32917148e-01
-7.89622724e-01 -3.68718594e-01 5.31982362e-01 5.87649167e-01
1.16559519e-02 -3.23049277e-01 6.20347969e-02 -1.19740701e+00
1.18866540e-01 -1.37395471e-01 6.44631207e-01 -5.07951140e-01
-1.71920586e+00 2.89106101e-01 -2.23780274e-01 -1.39279556e+00
2.50113159e-01 -4.76671040e-01 -9.36843157e-01 7.61911988e-01
-1.41667652e+00 -1.30263686e+00 -6.57683253e-01 7.63925672e-01
-1.81337744e-02 -2.95078397e-01 8.21968794e-01 8.93909097e-01
-5.92982948e-01 8.94569933e-01 3.02956343e-01 3.29853058e-01
6.31836355e-01 -9.34181035e-01 3.22476327e-01 1.14238277e-01
2.06457078e-01 8.22295904e-01 3.92582417e-01 -8.97653341e-01
-1.19014215e+00 -1.07439065e+00 1.17857099e+00 -1.39585510e-01
1.29546964e+00 -2.76364893e-01 -1.04934323e+00 8.59507442e-01
2.10881099e-01 3.67756784e-01 1.01796937e+00 6.38724267e-01
-4.13021863e-01 -4.31356341e-01 -1.17738235e+00 6.27251744e-01
1.62510216e+00 -4.96433407e-01 -1.25367850e-01 6.96079850e-01
9.89971101e-01 2.60407645e-02 -1.34666622e+00 3.31280440e-01
5.97654104e-01 -6.91159844e-01 9.20937479e-01 -8.81088257e-01
1.17721535e-01 -2.82415092e-01 -9.64622721e-02 -9.99378264e-01
-5.57431459e-01 -6.82862639e-01 -4.20325994e-01 1.57061815e+00
4.66308981e-01 -7.84563541e-01 9.66969550e-01 5.32170355e-01
4.74605225e-02 -7.71451294e-01 -5.87373078e-01 -3.36642027e-01
-5.16498983e-01 -1.42084315e-01 8.50140631e-01 1.42091727e+00
3.33315849e-01 1.94382310e-01 -2.56250054e-01 3.63403469e-01
6.47929430e-01 2.84199774e-01 6.56106174e-01 -1.74975371e+00
-3.78988236e-01 -4.75751698e-01 -5.04882872e-01 -1.09284925e+00
3.26851904e-01 -1.30580103e+00 -8.32687676e-01 -1.61876059e+00
-4.81718928e-02 -9.45068538e-01 -4.66415167e-01 7.67440796e-02
2.22982526e-01 1.11547537e-01 -1.69294938e-01 4.42199171e-01
-8.18890691e-01 2.68901467e-01 1.23250473e+00 -2.93488145e-01
-1.72624052e-01 1.29394680e-01 -4.96928155e-01 6.87487960e-01
7.30756104e-01 -2.90179133e-01 -6.91479325e-01 -4.66571689e-01
9.04931784e-01 7.99281001e-02 2.22526133e-01 -5.86319983e-01
5.58079064e-01 -2.46525668e-02 3.61996293e-02 -4.17704612e-01
9.18110088e-02 -1.36556065e+00 2.74813622e-01 2.39408940e-01
-3.74794453e-01 7.18102828e-02 -3.60253155e-01 9.48635876e-01
-1.62725642e-01 -3.94346088e-01 2.73867488e-01 1.36247650e-01
-6.61365509e-01 8.10901046e-01 3.65139455e-01 -1.50814205e-01
8.25597882e-01 -1.42587155e-01 -1.29724473e-01 -5.55449665e-01
-1.20171916e+00 5.70718288e-01 2.12347552e-01 5.73966622e-01
4.59965706e-01 -1.77662063e+00 -4.10116374e-01 2.62344807e-01
3.43539059e-01 -2.50198394e-01 1.10867880e-01 8.35151374e-01
-3.15165430e-01 5.25585469e-03 6.83350489e-02 -4.41881478e-01
-1.21930492e+00 6.12708032e-01 2.31504947e-01 -1.27940506e-01
-5.20125151e-01 7.20100164e-01 -4.82971102e-01 -6.05805755e-01
6.04599059e-01 1.23538792e-01 -5.63425958e-01 4.15309519e-01
1.98063537e-01 3.57737780e-01 -9.21059549e-02 -3.81920189e-01
3.07064876e-02 6.58334136e-01 -2.05634192e-01 5.14712095e-01
1.15607512e+00 -5.03510416e-01 -4.40386772e-01 3.26729625e-01
1.58554220e+00 1.55379832e-01 -6.69631362e-01 -6.44914865e-01
1.26506403e-01 -5.98860681e-01 -2.10178137e-01 -1.29601941e-01
-1.24152875e+00 5.15839577e-01 5.50130904e-01 8.87147903e-01
6.06633067e-01 5.35966158e-02 9.26616549e-01 5.83878636e-01
2.95739979e-01 -9.26681340e-01 -8.17093924e-02 4.88180630e-02
6.17673397e-01 -1.43325007e+00 7.24555776e-02 -8.31971824e-01
-1.62161142e-01 1.24320757e+00 7.24004507e-01 -2.25807354e-01
1.34149015e+00 -1.75223202e-01 -3.04547668e-01 -3.73352736e-01
-3.34212005e-01 -2.43818928e-02 4.03380394e-01 4.54268724e-01
5.01405895e-01 -6.30316213e-02 -5.17117202e-01 6.42594934e-01
-3.22320592e-03 -8.47689807e-02 9.66310203e-02 4.02240217e-01
-5.48696280e-01 -1.30183077e+00 1.90636680e-01 6.40790343e-01
-1.61474854e-01 -1.09180510e-01 -2.22936153e-01 6.61886752e-01
2.72176713e-01 7.34250009e-01 1.90077692e-01 -7.94430554e-01
1.57848388e-01 -1.25679180e-01 1.45402551e-01 -3.40429842e-01
-2.85626352e-01 -2.35609993e-01 2.16992423e-01 -5.38911700e-01
-4.17356402e-01 -3.30663979e-01 -1.27334630e+00 -7.60604620e-01
-6.83291316e-01 4.09103632e-01 4.91645962e-01 7.44923711e-01
5.18985569e-01 2.71716923e-01 8.53289247e-01 -3.86792004e-01
-2.45277926e-01 -9.32297051e-01 -6.29996359e-01 8.07104647e-01
5.38267791e-02 -7.26927102e-01 -3.72857332e-01 -2.92171866e-01] | [7.327400207519531, 6.250469207763672] |
b6d05947-8ad1-4065-952a-9fbaaecb0190 | keyphrase-extraction-with-incomplete | null | null | https://aclanthology.org/2021.wnut-1.4 | https://aclanthology.org/2021.wnut-1.4.pdf | Keyphrase Extraction with Incomplete Annotated Training Data | Extracting keyphrases that summarize the main points of a document is a fundamental task in natural language processing. Supervised approaches to keyphrase extraction(KPE) are largely developed based on the assumption that the training data is fully annotated. However, due to the difficulty of keyphrase annotating, KPE models severely suffer from incomplete annotated problem in many scenarios. To this end, we propose a more robust training method that learns to mitigate the misguidance brought by unlabeled keyphrases. We introduce negative sampling to adjust training loss, and conduct experiments under different scenarios. Empirical studies on synthetic datasets and open domain dataset show that our model is robust to incomplete annotated problem and surpasses prior baselines. Extensive experiments on five scientific domain datasets of different scales demonstrate that our model is competitive with the state-of-the-art method. | ['Richong Zhang', 'Guanghui Ma', 'Chunming Hu', 'Yanfei Lei'] | null | null | null | null | wnut-acl-2021-11 | ['keyphrase-extraction'] | ['natural-language-processing'] | [ 1.56524166e-01 6.87055513e-02 -4.91394818e-01 -1.18683353e-02
-1.17579317e+00 -9.90354896e-01 7.13168025e-01 5.77514708e-01
-5.52262962e-01 8.77228916e-01 4.13685352e-01 -2.03290567e-01
-7.57767539e-03 -6.00119054e-01 -8.18430066e-01 -3.55255246e-01
1.49953872e-01 1.06528178e-01 2.73709565e-01 -9.49883536e-02
6.19243979e-01 3.04675251e-01 -8.78576517e-01 3.79377931e-01
8.49682987e-01 7.19111145e-01 2.49199141e-02 5.38008213e-01
-4.74758476e-01 1.09759164e+00 -8.15190196e-01 -4.32567328e-01
3.39722067e-01 -2.70539410e-02 -9.89374518e-01 8.34354088e-02
5.36350191e-01 -2.77617216e-01 -5.71717918e-01 1.25472164e+00
3.37482601e-01 -2.15143919e-01 6.84198499e-01 -1.35078514e+00
-6.53618872e-01 8.42423260e-01 -7.51683116e-01 3.51322144e-01
2.02569202e-01 -4.13461328e-01 1.32933509e+00 -1.09977734e+00
7.18168318e-01 1.02184880e+00 4.78780329e-01 2.50189990e-01
-1.08880818e+00 -7.73424625e-01 2.94365823e-01 2.93956362e-02
-1.43500328e+00 -2.28090316e-01 7.61683404e-01 -1.68135107e-01
5.42718649e-01 1.51121914e-01 2.01555431e-01 8.76523495e-01
1.79429621e-01 1.33120501e+00 1.12178278e+00 -4.31766480e-01
2.29412951e-02 1.34963736e-01 5.09011149e-01 6.90373540e-01
5.36462188e-01 -4.42983240e-01 -7.82601714e-01 -5.03350496e-01
5.01176298e-01 -4.44718264e-02 -4.01943088e-01 -2.52015769e-01
-1.45867193e+00 5.78506231e-01 -7.34791160e-02 3.70914415e-02
-3.46535474e-01 -2.54792392e-01 4.83769894e-01 4.03157651e-01
5.41176140e-01 9.49206948e-01 -8.50027025e-01 -5.15998639e-02
-1.10247970e+00 5.87041974e-01 1.04866123e+00 1.08174479e+00
5.57032347e-01 -4.33389664e-01 -4.54026341e-01 7.39419639e-01
-1.91771775e-01 2.95810491e-01 3.09441775e-01 -7.53416777e-01
5.53499579e-01 8.98502827e-01 2.39859805e-01 -1.10208595e+00
-1.57309532e-01 -3.95122796e-01 -8.70996654e-01 -3.79551053e-01
3.60230684e-01 -1.74927250e-01 -8.99486542e-01 1.35253644e+00
1.47277057e-01 3.16940770e-02 8.27362239e-02 4.44195330e-01
1.09275353e+00 8.12601268e-01 8.44439026e-03 -3.67858499e-01
1.51632333e+00 -9.70851064e-01 -1.12586415e+00 -2.27670461e-01
4.04770076e-01 -9.60474133e-01 1.03334069e+00 7.27014720e-01
-7.45952785e-01 -2.47802094e-01 -1.12917387e+00 -3.58543336e-01
-4.95289773e-01 4.29172575e-01 6.35976732e-01 3.09702694e-01
-5.18068969e-01 4.73249286e-01 -5.82504690e-01 -1.40495494e-01
5.62507093e-01 -3.87213379e-02 -4.69323099e-01 -1.33603603e-01
-1.04317629e+00 5.38591862e-01 7.99884439e-01 -3.32820922e-01
-7.38849044e-01 -1.02387667e+00 -6.01208508e-01 1.13149613e-01
1.11452937e+00 -4.20649052e-01 1.28998554e+00 -2.80982137e-01
-1.02387273e+00 8.32624853e-01 -1.34183178e-02 -4.75389600e-01
3.60524803e-01 -7.15232611e-01 -1.96840018e-01 2.59800673e-01
1.67963639e-01 5.01060665e-01 8.97071004e-01 -1.21077693e+00
-7.39679039e-01 -1.06502935e-01 2.09838495e-01 -1.11979470e-02
-8.59494567e-01 9.45474580e-02 -7.12609828e-01 -1.23648298e+00
-1.56733405e-03 -7.97297776e-01 -2.08303854e-01 -4.41870391e-02
-6.90901458e-01 -4.05155778e-01 9.29551482e-01 -7.05922782e-01
1.63449574e+00 -1.89800632e+00 -9.31450864e-04 -1.03817452e-02
5.46801805e-01 3.24645072e-01 -7.23662376e-02 7.23539114e-01
1.68240875e-01 3.56494576e-01 -3.31301928e-01 -8.88122991e-03
-9.82730687e-02 1.15443513e-01 -1.04590225e+00 1.90394714e-01
1.11626267e-01 8.36821795e-01 -1.03881419e+00 -7.25538373e-01
-3.55794579e-01 -6.47462383e-02 -3.53622317e-01 3.62955123e-01
-4.95649427e-01 1.42750695e-01 -8.31108630e-01 7.55933344e-01
5.98978162e-01 -5.99241853e-01 -2.57944223e-02 -4.42623794e-01
1.24776354e-02 3.68135989e-01 -1.27786863e+00 1.67307055e+00
-4.17665839e-02 4.99335259e-01 -2.21987106e-02 -1.06658423e+00
6.46387756e-01 3.07706267e-01 5.62781572e-01 -2.86936194e-01
-5.93894161e-02 1.28298670e-01 -3.77808690e-01 -3.11240315e-01
9.16464150e-01 4.09450203e-01 -3.45300913e-01 6.98833704e-01
1.67138889e-01 -2.74023116e-01 5.05138993e-01 9.34366703e-01
1.20981157e+00 -6.18521795e-02 6.86149597e-01 -3.64718765e-01
5.57323456e-01 1.97091356e-01 6.76495612e-01 1.03061759e+00
-7.09339529e-02 3.50964516e-01 8.70095730e-01 -5.38991392e-01
-1.05316961e+00 -6.27840281e-01 -3.75267006e-02 1.05277932e+00
7.43518099e-02 -1.11633718e+00 -6.45343304e-01 -1.19078612e+00
-2.31911358e-03 3.45852166e-01 -4.56532180e-01 -2.24480722e-02
-3.74914646e-01 -8.70998919e-01 6.87672079e-01 4.43502814e-01
4.12661701e-01 -7.08107531e-01 -2.89449513e-01 9.65067223e-02
-3.54131073e-01 -1.63300371e+00 -5.70120454e-01 1.45675587e-02
-5.49579740e-01 -1.23690438e+00 -8.02700043e-01 -6.79645181e-01
8.45330596e-01 5.57532251e-01 1.28430223e+00 -7.44225904e-02
-2.10373268e-01 3.65798563e-01 -5.37070394e-01 -7.70569623e-01
-3.40092987e-01 4.09786850e-01 -1.64211709e-02 -3.50630701e-01
5.43247819e-01 -2.24075899e-01 -3.21230263e-01 5.89954527e-03
-1.21143150e+00 1.46753788e-01 7.21607387e-01 8.94972622e-01
6.53789043e-01 4.32467490e-01 5.23960114e-01 -1.27790344e+00
9.80659366e-01 -2.99757570e-01 -7.33157635e-01 5.93216419e-01
-8.24833393e-01 2.56230533e-01 7.08535075e-01 -4.91865873e-01
-8.94316077e-01 -1.70418724e-01 3.57537240e-01 -2.56651759e-01
1.41410574e-01 6.91761017e-01 -1.46493062e-01 1.43890837e-02
5.28062999e-01 3.72264713e-01 -4.97349232e-01 -5.95288277e-01
5.29932678e-01 7.09207356e-01 7.35591412e-01 -8.32368851e-01
1.11771786e+00 5.34413040e-01 2.18240656e-02 -8.03956985e-01
-1.43708682e+00 -6.93324089e-01 -4.79973108e-01 1.56374693e-01
3.18429708e-01 -1.28427017e+00 -4.72578764e-01 3.57459038e-01
-1.17631161e+00 1.76696599e-01 -1.28054321e-01 2.15211749e-01
-1.05787784e-01 6.68239117e-01 -7.37160444e-01 -4.42297727e-01
-7.08242059e-01 -7.59896636e-01 1.12569547e+00 2.87589431e-01
-1.73164755e-01 -5.96534193e-01 3.24007198e-02 3.00959349e-01
1.10076834e-02 1.61249518e-01 9.27006245e-01 -1.00846851e+00
-5.97109914e-01 -4.77180690e-01 -3.86619955e-01 3.76484960e-01
3.60793024e-01 9.60381925e-02 -8.29737008e-01 -2.22898990e-01
-1.66322410e-01 -7.17423916e-01 1.16892362e+00 -2.87050903e-02
1.53676212e+00 -6.74034774e-01 -4.16111410e-01 3.53619903e-01
1.07028115e+00 -2.77314574e-01 3.44133854e-01 4.08378333e-01
7.66987145e-01 5.78379214e-01 9.28231776e-01 6.07521594e-01
4.35750157e-01 1.84102338e-02 3.22068594e-02 -6.62579611e-02
2.69102871e-01 -5.57914555e-01 1.40161484e-01 8.20292652e-01
3.54840934e-01 -4.94239062e-01 -8.80798936e-01 6.45788848e-01
-1.87699211e+00 -7.88169980e-01 8.88176188e-02 1.85084558e+00
1.37709379e+00 4.86325502e-01 2.49154004e-03 3.07071149e-01
3.99322242e-01 3.71308029e-01 -3.04548144e-01 6.75989687e-02
-2.05242530e-01 1.44283041e-01 6.71071649e-01 5.80581501e-02
-1.41416454e+00 1.28066742e+00 6.35960674e+00 1.10322523e+00
-7.79087365e-01 -3.02180320e-01 5.66733479e-01 2.42935047e-01
-1.76968724e-01 6.57591457e-03 -1.17338800e+00 2.42048666e-01
5.55972219e-01 -5.48009813e-01 7.16722384e-02 9.32140827e-01
-1.91717252e-01 -2.28411872e-02 -1.09564602e+00 8.57324958e-01
9.95209664e-02 -1.45934021e+00 2.53154784e-01 -1.20505422e-01
8.89117062e-01 -1.80945292e-01 -6.77146912e-02 3.13756168e-01
6.22918487e-01 -6.83348000e-01 5.21941543e-01 2.94417173e-01
6.57826900e-01 -5.47402561e-01 6.39660776e-01 5.36256492e-01
-1.06506705e+00 1.26332536e-01 -4.31328863e-01 5.94992675e-02
-2.34745607e-01 7.62442827e-01 -9.56159711e-01 4.42290217e-01
5.46949089e-01 8.33278239e-01 -7.16051936e-01 7.61597395e-01
-6.19152248e-01 7.41993964e-01 -2.69258589e-01 -8.89810249e-02
2.27396771e-01 1.58897117e-01 4.58267033e-01 1.42136943e+00
-5.05761616e-02 3.29911351e-01 4.21767056e-01 7.57646799e-01
-7.20843971e-01 2.70752609e-01 -5.56116402e-01 -7.80554831e-01
6.54526770e-01 1.54783058e+00 -8.83625805e-01 -5.93175113e-01
-5.40690422e-01 8.41085732e-01 3.33833635e-01 3.47719878e-01
-4.28506970e-01 -6.27349913e-01 3.28628838e-01 -1.68239385e-01
3.72372776e-01 -3.08953702e-01 -1.83725059e-01 -1.47856224e+00
1.97187871e-01 -1.22935581e+00 6.67965114e-01 -5.14008284e-01
-1.42917180e+00 3.25487137e-01 9.59880948e-02 -1.29610789e+00
6.46952987e-02 -5.71250796e-01 -2.59256840e-01 4.89885420e-01
-1.73610246e+00 -1.09792745e+00 -8.68269205e-02 2.75705814e-01
6.97385609e-01 -1.35297135e-01 7.43860424e-01 3.88256535e-02
-4.64195907e-01 4.84024048e-01 1.04249015e-01 6.15041554e-01
1.11154878e+00 -1.40037835e+00 6.32591844e-01 1.10677946e+00
3.64954412e-01 1.03939128e+00 6.17582858e-01 -8.18883061e-01
-1.48951209e+00 -9.92340505e-01 1.00112295e+00 -5.69680035e-01
8.86130273e-01 -3.75294626e-01 -1.12157166e+00 5.90766370e-01
3.82567167e-01 2.36413121e-01 9.24233317e-01 1.80585366e-02
-6.90768063e-01 4.48770560e-02 -8.02920699e-01 6.32441282e-01
6.93347454e-01 -6.79660797e-01 -9.38260198e-01 4.67244983e-01
8.34783494e-01 -6.13501430e-01 -8.42096686e-01 5.47702491e-01
4.44467396e-01 -1.90478995e-01 1.04383695e+00 -8.63622487e-01
7.03614533e-01 -3.31839204e-01 1.02812931e-01 -1.14980567e+00
-1.93504244e-01 -9.18454468e-01 -6.39762342e-01 1.30368078e+00
2.76168376e-01 -7.35706985e-02 6.08802557e-01 5.47131181e-01
2.81049609e-01 -4.94682729e-01 -4.35373843e-01 -6.77194476e-01
6.51524663e-02 -1.53567642e-01 4.17099953e-01 1.23739767e+00
1.84077635e-01 8.01957965e-01 -5.51582575e-01 2.08651900e-01
7.37403929e-01 2.99990118e-01 9.70542014e-01 -1.30235422e+00
-1.21673825e-03 -2.33431578e-01 9.04443935e-02 -1.15032649e+00
3.42249393e-01 -6.76218867e-01 -1.66151866e-01 -1.53103435e+00
5.71580708e-01 -1.26683608e-01 -4.04122740e-01 6.75745130e-01
-7.80607224e-01 3.98067646e-02 8.64270143e-03 3.74658018e-01
-9.02142346e-01 4.00094330e-01 1.20018518e+00 -3.68856817e-01
-6.52825758e-02 -2.59986401e-01 -1.01057112e+00 8.55224967e-01
5.61823249e-01 -7.26730525e-01 -4.00641233e-01 -1.75343305e-01
6.00474834e-01 -1.84105113e-01 1.08973056e-01 -6.15715742e-01
5.64016461e-01 -3.79733443e-01 1.95576638e-01 -8.68502021e-01
-2.10280031e-01 -6.52179599e-01 -5.98984063e-01 1.30470410e-01
-7.14239717e-01 -1.86465483e-03 3.60798478e-01 7.12975085e-01
-2.42119461e-01 -7.75270984e-02 4.25971955e-01 -3.08523864e-01
-6.55274987e-01 4.69706982e-01 -1.49375200e-01 5.89019239e-01
5.11000812e-01 4.65432197e-01 -5.49324930e-01 -1.79933056e-01
-4.61557470e-02 4.46450233e-01 2.85378098e-01 4.71411765e-01
5.38026631e-01 -1.14291215e+00 -8.38353753e-01 -5.65813556e-02
4.64350551e-01 1.75518915e-01 -2.93546945e-01 3.39295924e-01
-3.64348501e-01 5.29891789e-01 1.44921124e-01 -9.17191505e-02
-1.23525846e+00 7.72196472e-01 -2.76841760e-01 -7.39491940e-01
-5.89481115e-01 7.47357368e-01 1.68752104e-01 -3.95635486e-01
5.14406860e-01 -5.01806140e-01 -3.07296306e-01 9.63220075e-02
9.26718414e-01 2.18320817e-01 1.54643089e-01 -2.32667267e-01
-1.96683094e-01 2.07961127e-01 -8.23103547e-01 9.13772509e-02
1.27329445e+00 6.24633469e-02 -1.79770589e-01 2.61455834e-01
9.71441567e-01 3.62545818e-01 -8.62461925e-01 -9.06376123e-01
3.21949750e-01 -3.68114740e-01 2.32129261e-01 -7.24225938e-01
-6.59219563e-01 5.03130436e-01 -1.17362827e-01 1.24345802e-01
9.92654324e-01 -1.23929836e-01 9.95817900e-01 1.07682383e+00
1.46201119e-01 -1.33794713e+00 2.84220695e-01 5.86024463e-01
6.74956739e-01 -1.45714438e+00 6.42512321e-01 -5.43994486e-01
-4.66335416e-01 1.13937092e+00 5.48109949e-01 1.19487204e-01
6.29395664e-01 5.13503313e-01 9.80914235e-02 -2.49932840e-01
-8.21075499e-01 6.25315309e-02 4.75520581e-01 7.75622651e-02
4.33208495e-01 -2.42828563e-01 -4.81737465e-01 7.67426729e-01
-1.18026324e-01 2.18549401e-01 5.95147789e-01 1.45133221e+00
-3.70384723e-01 -1.04709780e+00 -3.82700503e-01 4.94005740e-01
-1.12073803e+00 -3.34663242e-01 -8.39787364e-01 5.64563155e-01
-3.23286533e-01 7.21047103e-01 -4.18652624e-01 -9.89949256e-02
1.26641929e-01 1.82066225e-02 3.01935196e-01 -6.00669622e-01
-4.77154046e-01 2.51780957e-01 -9.75459665e-02 -2.26176679e-01
-4.73688751e-01 -1.93920076e-01 -1.13698411e+00 -1.05436005e-01
-2.87467867e-01 5.19753695e-01 1.41226664e-01 8.84378552e-01
4.38275367e-01 5.02173305e-01 4.95374769e-01 -1.86676979e-01
-7.75441170e-01 -9.35395002e-01 -5.09809554e-01 4.43288326e-01
4.90891010e-01 -2.68484592e-01 -7.52241015e-02 4.57166642e-01] | [12.292301177978516, 8.88737964630127] |
f2db487a-7457-4ee9-87d3-68d161a3e37d | combining-deep-learning-with-geometric | 2005.05481 | null | https://arxiv.org/abs/2005.05481v2 | https://arxiv.org/pdf/2005.05481v2.pdf | Combining Deep Learning with Geometric Features for Image based Localization in the Gastrointestinal Tract | Tracking monocular colonoscope in the Gastrointestinal tract (GI) is a challenging problem as the images suffer from deformation, blurred textures, significant changes in appearance. They greatly restrict the tracking ability of conventional geometry based methods. Even though Deep Learning (DL) can overcome these issues, limited labeling data is a roadblock to state-of-art DL method. Considering these, we propose a novel approach to combine DL method with traditional feature based approach to achieve better localization with small training data. Our method fully exploits the best of both worlds by introducing a Siamese network structure to perform few-shot classification to the closest zone in the segmented training image set. The classified label is further adopted to initialize the pose of scope. To fully use the training dataset, a pre-generated triangulated map points within the zone in the training set are registered with observation and contribute to estimating the optimal pose of the test image. The proposed hybrid method is extensively tested and compared with existing methods, and the result shows significant improvement over traditional geometric based or DL based localization. The accuracy is improved by 28.94% (Position) and 10.97% (Orientation) with respect to state-of-art method. | ['Jingwei Song', 'Andreas Girgensohn', 'Mitesh Patel', 'Chelhwon Kim'] | 2020-05-11 | null | null | null | null | ['image-based-localization'] | ['computer-vision'] | [-2.41197750e-01 -1.23285428e-01 -2.64460415e-01 8.44750926e-02
-6.36631668e-01 -7.28197813e-01 2.85274923e-01 2.36691743e-01
-5.93879342e-01 5.59819877e-01 -6.27646744e-02 -8.71049687e-02
-6.31386414e-02 -6.36168420e-01 -6.83569789e-01 -9.15388525e-01
-1.91213161e-01 3.39252412e-01 4.34068859e-01 -2.74335649e-02
1.73363805e-01 4.57785785e-01 -1.05640781e+00 -1.16043091e-01
8.77087235e-01 8.07666600e-01 4.51834977e-01 3.59605432e-01
-4.36978079e-02 5.52510202e-01 -3.78514558e-01 3.55433732e-01
3.79475653e-01 -1.99600846e-01 -4.78643030e-01 -8.26807097e-02
3.50056469e-01 -2.17156574e-01 -1.25264227e-01 1.28090334e+00
6.78804100e-01 3.48639399e-01 5.16453505e-01 -4.94632810e-01
-1.85348079e-01 2.86472887e-01 -7.43382215e-01 2.51450062e-01
4.59501743e-01 -1.15566388e-01 1.20980725e-01 -7.04948127e-01
9.99603271e-01 8.21731269e-01 9.96291578e-01 2.87465721e-01
-7.41337597e-01 -4.66271281e-01 1.35950863e-01 -1.85909256e-01
-1.35336816e+00 1.64449573e-01 7.22486317e-01 -3.65263939e-01
3.84067684e-01 4.19714563e-02 8.05263281e-01 5.80267608e-01
2.21611843e-01 5.97445667e-01 1.01935506e+00 -5.98974168e-01
3.40246297e-02 3.74854267e-01 3.15438658e-02 1.08408689e+00
6.32497787e-01 3.13684225e-01 2.34493725e-02 -9.09989979e-03
1.02956188e+00 3.78381610e-01 -4.98389989e-01 -9.90825295e-01
-1.44465065e+00 6.59182429e-01 9.72005069e-01 5.44203579e-01
-3.06164056e-01 -9.03004557e-02 3.76036793e-01 -7.41755888e-02
6.63347840e-02 5.27590036e-01 -3.84824961e-01 1.24146163e-01
-8.07117045e-01 -3.04989927e-02 8.88088167e-01 8.88782501e-01
4.89774853e-01 -2.27593095e-03 8.56734067e-02 4.92418706e-01
4.54708308e-01 3.66054714e-01 9.18640256e-01 -4.84893233e-01
3.03082854e-01 8.26244175e-01 1.29748389e-01 -1.09527361e+00
-8.88362408e-01 -9.71750259e-01 -5.68057716e-01 3.82128358e-01
5.85388482e-01 -3.13946992e-01 -1.02859879e+00 1.30102336e+00
7.78858125e-01 3.31832319e-01 -1.68270003e-02 1.04705322e+00
1.07362926e+00 3.77011091e-01 -1.55654535e-01 -7.75898471e-02
1.31500936e+00 -1.20929527e+00 -4.56547499e-01 1.28862873e-01
1.00603402e+00 -1.04139805e+00 5.70626378e-01 5.17364919e-01
-8.32633317e-01 -6.95691407e-01 -1.33668351e+00 2.51383901e-01
-3.56513262e-01 5.40473104e-01 4.25327510e-01 7.46146083e-01
-8.85112882e-01 5.50251901e-01 -1.18615329e+00 -4.99251693e-01
9.51539204e-02 6.30304635e-01 -4.28491026e-01 -2.26776630e-01
-9.17959690e-01 9.86804247e-01 6.34060979e-01 -6.24225661e-02
-7.21096873e-01 -5.04772544e-01 -1.11430907e+00 -3.97574663e-01
3.92898977e-01 -7.24478722e-01 9.23973322e-01 -5.99096715e-01
-1.51492751e+00 7.17722058e-01 1.32805869e-01 -4.23599154e-01
6.92280591e-01 -2.22085893e-01 -1.08650140e-01 2.04391435e-01
1.27873242e-01 5.12683809e-01 4.90153998e-01 -1.26827633e+00
-7.75635421e-01 -3.65140498e-01 8.28040764e-02 7.30771065e-01
-8.10839310e-02 -4.27024126e-01 -7.43740499e-01 -4.72587228e-01
5.60198009e-01 -1.21856666e+00 -2.85663128e-01 2.13204816e-01
-2.02854648e-01 6.83588758e-02 8.35512519e-01 -6.54767931e-01
1.09926105e+00 -2.08328867e+00 -7.73193464e-02 -2.01335438e-02
9.65112001e-02 5.24148285e-01 4.54091609e-01 1.55457392e-01
4.43410166e-02 -4.81891274e-01 2.29234427e-01 -2.07289219e-01
-4.14990008e-01 -4.47756937e-03 4.46249574e-01 1.15668476e+00
-5.02863407e-01 6.32655084e-01 -1.12997496e+00 -7.90484607e-01
8.37874293e-01 5.42771101e-01 -3.50369364e-01 1.72593016e-02
-5.44357598e-02 9.31941330e-01 -6.12816274e-01 6.67399943e-01
9.20846045e-01 -1.88414931e-01 8.47785771e-02 -4.92253125e-01
-3.47401679e-01 -7.07559437e-02 -1.51298022e+00 2.14581800e+00
-5.64872861e-01 1.18606895e-01 -9.92263705e-02 -9.22645748e-01
8.89146447e-01 3.75788927e-01 6.57931626e-01 -4.91284698e-01
3.12684804e-01 4.43957537e-01 1.25820339e-01 -8.42222393e-01
7.22588897e-02 1.88524798e-01 1.40845776e-01 6.01611622e-02
-8.91345926e-03 1.51147172e-01 7.21086711e-02 -2.69095302e-01
6.28987789e-01 7.31409132e-01 6.76802754e-01 -3.52971435e-01
8.71374190e-01 4.45072949e-01 4.95389462e-01 5.82313478e-01
-3.49071205e-01 6.52654946e-01 -1.56535238e-01 -8.30889881e-01
-8.51539910e-01 -6.74364746e-01 -3.46274763e-01 5.05993664e-01
8.50082695e-01 2.97570497e-01 -6.79814219e-01 -7.95888007e-01
-4.89584431e-02 7.46705458e-02 -6.20087385e-01 1.33403674e-01
-9.24976408e-01 -7.62394667e-01 3.60425353e-01 3.16957831e-01
7.10468233e-01 -6.93813264e-01 -7.46496737e-01 2.06444159e-01
1.05122037e-01 -8.72512162e-01 -4.16358709e-01 1.21856235e-01
-1.02978098e+00 -1.32330263e+00 -1.03879118e+00 -1.41561604e+00
9.59643304e-01 3.79664123e-01 6.45522535e-01 1.06081195e-01
-3.80807996e-01 -8.84354413e-02 -3.35144609e-01 -1.88544691e-01
-2.59651244e-01 2.81255454e-01 -9.90805849e-02 -4.04505849e-01
1.62368372e-01 -3.24339755e-02 -1.14064884e+00 3.42927992e-01
-5.03886223e-01 -3.13233398e-02 6.37487411e-01 9.06600773e-01
6.03278100e-01 -1.00257427e-01 2.51694828e-01 -8.97821486e-01
1.96837157e-01 -5.04791141e-01 -9.94557619e-01 9.15139392e-02
-6.45077288e-01 -6.24024309e-02 7.17685759e-01 -6.10460997e-01
-8.99403632e-01 4.21789676e-01 2.86455266e-02 -3.69633079e-01
-2.51375914e-01 5.03883779e-01 4.60590750e-01 -6.82336986e-01
7.38738656e-01 1.02666333e-01 2.26858407e-01 -3.90963763e-01
9.12658647e-02 4.54987079e-01 4.97629732e-01 -1.06840141e-01
6.13650620e-01 6.44951105e-01 1.31067514e-01 -3.94196868e-01
-6.70242548e-01 -9.48970318e-01 -8.63735735e-01 -7.25055337e-02
7.94182479e-01 -1.11363137e+00 -6.93735182e-01 1.27142102e-01
-8.08892071e-01 9.35391337e-02 6.19316660e-02 1.22808683e+00
-1.91772938e-01 5.85006416e-01 -6.96573615e-01 -6.00305140e-01
-5.40186644e-01 -1.36807442e+00 8.37777317e-01 5.01722753e-01
1.48889899e-01 -1.42140841e+00 2.88954526e-01 4.04574238e-02
4.13794547e-01 7.99122512e-01 3.36875081e-01 -7.60879219e-01
-6.00992084e-01 -6.44407153e-01 -1.52723521e-01 -1.42596528e-01
3.11732233e-01 -4.14696097e-01 -6.22448683e-01 -6.83090985e-01
4.53938246e-01 1.40025109e-01 5.39514124e-01 8.36721957e-01
6.69370651e-01 1.96661055e-01 -9.25095916e-01 9.63373721e-01
1.87633979e+00 6.03856862e-01 1.07695125e-02 7.23722219e-01
7.89481103e-01 7.29771927e-02 1.00104558e+00 2.19645932e-01
2.38711819e-01 8.00751686e-01 4.91670579e-01 -4.83143419e-01
-3.44244421e-01 -9.02681202e-02 1.81753952e-02 8.19436073e-01
8.36463273e-02 1.15619272e-01 -6.52259052e-01 4.78235453e-01
-1.74386859e+00 -5.52469492e-01 -6.79287463e-02 2.43430471e+00
6.16078258e-01 7.09596202e-02 -2.46929331e-03 -9.18766484e-02
7.74910510e-01 -7.88979754e-02 -4.84399587e-01 1.23454630e-01
2.59690881e-01 -1.71587393e-02 9.30386305e-01 5.04436672e-01
-1.58407557e+00 5.83028495e-01 5.38346243e+00 7.34198689e-01
-1.44727981e+00 1.32445663e-01 1.94200382e-01 1.15954779e-01
3.64338398e-01 -2.16906965e-01 -9.49362397e-01 3.81087452e-01
2.89393157e-01 6.70481324e-02 1.37846142e-01 9.49707866e-01
1.93754941e-01 -3.32480222e-01 -8.54263842e-01 9.78206813e-01
4.21708912e-01 -1.15065873e+00 -4.33311373e-01 1.10923462e-02
1.14298296e+00 2.34565258e-01 4.05692048e-02 2.32778817e-01
-2.47126296e-02 -7.74887204e-01 2.53990620e-01 3.34724069e-01
7.88792133e-01 -5.11448085e-01 1.14422178e+00 4.86397177e-01
-1.48393106e+00 1.14750236e-01 -5.10451615e-01 1.80836245e-01
1.46854324e-02 2.45211661e-01 -1.29880965e+00 8.03387165e-01
4.02324885e-01 6.91774368e-01 -3.84257644e-01 1.74443805e+00
-2.01277640e-02 1.53119311e-01 -5.38002074e-01 -1.66566193e-01
5.59873343e-01 -1.43679708e-01 5.73279738e-01 9.33209419e-01
5.03067374e-01 -3.43194008e-01 6.77734017e-01 4.76287782e-01
1.36963874e-01 3.71524125e-01 -7.85243511e-01 6.20480895e-01
3.33362341e-01 1.39591384e+00 -9.54594195e-01 -4.18862224e-01
-4.07190949e-01 6.99835002e-01 -7.54212290e-02 1.02170035e-02
-8.23228300e-01 -3.18114728e-01 -2.71287322e-01 2.72256713e-02
3.19587111e-01 -1.18635230e-01 -7.03339800e-02 -1.10544658e+00
-7.72706121e-02 -5.87315619e-01 4.35622662e-01 -4.39161003e-01
-5.52393436e-01 7.63185859e-01 -1.26931742e-01 -1.95603204e+00
-2.69141287e-01 -6.01887643e-01 -5.63127697e-01 8.50429833e-01
-1.59114122e+00 -1.32150102e+00 -7.37480342e-01 3.91004890e-01
6.20885789e-01 -8.44829082e-02 8.57421577e-01 5.55775583e-01
-2.78376520e-01 5.15668094e-01 5.02454817e-01 1.62074864e-01
7.80545950e-01 -1.32980514e+00 -9.47529227e-02 7.25525260e-01
2.45249625e-02 7.50793576e-01 4.70345378e-01 -8.00753474e-01
-1.37962961e+00 -9.72873807e-01 1.93631202e-01 -3.51030260e-01
2.96136796e-01 1.06132492e-01 -5.97150326e-01 5.80268383e-01
6.81424886e-02 3.16857606e-01 3.48576546e-01 -3.44974667e-01
3.02672595e-01 7.78772589e-03 -1.35249305e+00 4.80967909e-01
5.86862743e-01 6.95170015e-02 -4.80702430e-01 4.75099415e-01
6.90790355e-01 -1.23267019e+00 -9.65248406e-01 3.98541600e-01
5.67445695e-01 -7.83893824e-01 1.13250887e+00 1.72809079e-01
-1.70488864e-01 -7.10688710e-01 4.81702268e-01 -1.19379163e+00
-9.08549726e-02 -4.97118413e-01 2.67648131e-01 4.72979009e-01
1.49993837e-01 -6.40340805e-01 1.14029586e+00 -3.32154594e-02
-3.28712493e-01 -8.34600389e-01 -7.33493268e-01 -6.42513156e-01
-8.41047168e-02 3.02193671e-01 3.40112716e-01 1.12046027e+00
9.18100327e-02 -2.17198402e-01 -1.20660789e-01 4.95576978e-01
7.33332753e-01 2.82411307e-01 6.70290947e-01 -1.05755818e+00
-1.22094378e-01 -5.14935739e-02 -5.56993008e-01 -1.07692611e+00
-5.39215207e-01 -9.36588883e-01 -1.34886950e-01 -1.64604688e+00
6.42279815e-03 -7.51253843e-01 -3.14371645e-01 4.72218767e-02
-1.21941306e-02 3.26040447e-01 -1.06735647e-01 3.65589440e-01
-5.86209536e-01 -7.02980608e-02 1.67396700e+00 1.58174202e-01
-4.11656529e-01 3.45973015e-01 -1.16137363e-01 8.95880878e-01
7.21728683e-01 -4.43557113e-01 -3.86690050e-01 -3.23665291e-01
-1.42328948e-01 3.82428080e-01 7.69924298e-02 -1.39494264e+00
6.49473429e-01 2.88805932e-01 6.49533510e-01 -9.08005595e-01
7.79436529e-02 -1.00617421e+00 2.42359802e-01 1.08138871e+00
-2.39714775e-02 3.11453957e-02 9.82522890e-02 4.46916193e-01
-3.49062622e-01 -6.56013906e-01 9.59550321e-01 -5.48792064e-01
-6.54321849e-01 2.34754369e-01 2.25013047e-01 -1.95921168e-01
1.32718229e+00 -5.65753579e-01 -9.15634409e-02 2.75721997e-02
-9.15177643e-01 1.18719801e-01 6.34095669e-01 7.25346133e-02
3.61533433e-01 -1.12403190e+00 -2.23722532e-01 2.99985647e-01
2.89538186e-02 4.49576169e-01 1.10307291e-01 1.39841998e+00
-1.35038078e+00 6.46628201e-01 -5.51120900e-02 -7.84238815e-01
-1.13477468e+00 7.78374553e-01 6.39526725e-01 -3.80045325e-01
-8.35118890e-01 7.03876853e-01 2.16883600e-01 -4.60878849e-01
3.40086728e-01 -6.50374413e-01 -5.65320611e-01 -2.85380185e-01
1.31262884e-01 3.15426856e-01 8.18054751e-02 -5.34016490e-01
-3.72571170e-01 1.05615807e+00 -1.40040427e-01 2.26499721e-01
1.02403784e+00 -2.55242884e-01 2.10598171e-01 1.26947597e-01
1.30098236e+00 1.33299649e-01 -1.09054053e+00 -3.17289919e-01
-1.99076414e-01 -4.17324990e-01 4.60304022e-02 -7.96612382e-01
-1.05281997e+00 7.39203572e-01 1.17695987e+00 -6.10564575e-02
6.94621503e-01 -3.25643301e-01 8.28477800e-01 1.28581583e-01
5.39335310e-01 -7.86619365e-01 -7.98351765e-02 1.61607862e-01
2.33983442e-01 -1.58345544e+00 2.44486660e-01 -4.26343054e-01
-4.10033762e-01 1.28068590e+00 5.72992444e-01 -6.62025034e-01
6.71133280e-01 2.27970168e-01 3.78439516e-01 -3.10836941e-01
3.86343747e-02 -4.23606932e-02 3.40254307e-01 5.63167572e-01
5.25520504e-01 -1.70221761e-01 -5.74493885e-01 -5.53301238e-02
1.32263899e-01 1.50555119e-01 2.44043544e-01 1.17349434e+00
-5.31656384e-01 -9.29341495e-01 -7.17859566e-01 1.37436688e-01
-7.84959555e-01 1.29701778e-01 6.29555702e-01 1.16788733e+00
4.14469212e-01 4.56970096e-01 -1.95802674e-01 9.74712744e-02
2.12728456e-01 -4.24461901e-01 5.15566826e-01 -6.37522757e-01
-7.36256242e-01 5.54374218e-01 -1.77676409e-01 -2.47903645e-01
-4.58093345e-01 -5.59741259e-01 -1.43819797e+00 3.43932301e-01
-7.40892947e-01 2.00384095e-01 8.60839844e-01 7.50303566e-01
-2.93393917e-02 4.59207416e-01 4.36148316e-01 -9.26612139e-01
-5.52598357e-01 -8.40759933e-01 -3.33828449e-01 4.96270388e-01
4.31751430e-01 -9.18489456e-01 -3.87383431e-01 -4.11138088e-02] | [13.961085319519043, -3.1316280364990234] |
d1cf0250-4112-4de0-978d-4d27e3181b08 | multi-source-survival-domain-adaptation | 2212.00424 | null | https://arxiv.org/abs/2212.00424v2 | https://arxiv.org/pdf/2212.00424v2.pdf | Multi-Source Survival Domain Adaptation | Survival analysis is the branch of statistics that studies the relation between the characteristics of living entities and their respective survival times, taking into account the partial information held by censored cases. A good analysis can, for example, determine whether one medical treatment for a group of patients is better than another. With the rise of machine learning, survival analysis can be modeled as learning a function that maps studied patients to their survival times. To succeed with that, there are three crucial issues to be tackled. First, some patient data is censored: we do not know the true survival times for all patients. Second, data is scarce, which led past research to treat different illness types as domains in a multi-task setup. Third, there is the need for adaptation to new or extremely rare illness types, where little or no labels are available. In contrast to previous multi-task setups, we want to investigate how to efficiently adapt to a new survival target domain from multiple survival source domains. For this, we introduce a new survival metric and the corresponding discrepancy measure between survival distributions. These allow us to define domain adaptation for survival analysis while incorporating censored data, which would otherwise have to be dropped. Our experiments on two cancer data sets reveal a superb performance on target domains, a better treatment recommendation, and a weight matrix with a plausible explanation. | ['Carolin Lawrence', 'Ammar Shaker'] | 2022-12-01 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [ 1.28898963e-01 -7.88463578e-02 -5.28433383e-01 -6.08302295e-01
-7.03996360e-01 -6.33727193e-01 3.88762295e-01 6.06596351e-01
-5.15005171e-01 1.24263704e+00 5.84144413e-01 -1.95480749e-01
-5.14635384e-01 -6.57635331e-01 -3.47295284e-01 -9.62404847e-01
-1.15656219e-01 8.10992718e-01 -9.28904265e-02 -8.80185515e-02
-2.26873517e-01 2.66163856e-01 -1.06414545e+00 2.48828456e-02
8.64195943e-01 7.53792048e-01 -1.29791394e-01 4.21038747e-01
-1.34708639e-02 4.21840608e-01 -5.35038769e-01 -3.08293402e-01
-2.58356899e-01 -3.41060430e-01 -9.99595284e-01 -1.73092633e-01
3.23414356e-02 -1.04844412e-02 6.29781336e-02 6.22237384e-01
3.85149539e-01 9.69068259e-02 1.18615627e+00 -1.58959270e+00
-5.39621472e-01 4.62813437e-01 -2.93346494e-01 4.12412211e-02
2.55160570e-01 -1.89528376e-01 9.58018184e-01 -3.01068693e-01
6.72917187e-01 9.79438365e-01 6.56956494e-01 6.78472579e-01
-1.51881051e+00 -2.14785397e-01 -5.83699606e-02 1.92435339e-01
-1.14303672e+00 -2.72289932e-01 4.50883836e-01 -7.35339046e-01
4.48826164e-01 4.25312996e-01 2.46312335e-01 1.42939961e+00
3.26182783e-01 3.67352694e-01 1.08261335e+00 -2.75960028e-01
4.16424334e-01 3.22728962e-01 3.59341413e-01 2.16241315e-01
3.01085293e-01 2.27636382e-01 -2.86322355e-01 -4.19574797e-01
2.25450277e-01 2.99635321e-01 -4.60452527e-01 -6.17194831e-01
-1.33277202e+00 1.05696893e+00 3.09757739e-01 5.98975122e-01
-3.30208272e-01 -2.38744929e-01 5.06363809e-01 5.43437123e-01
7.03327894e-01 2.60651499e-01 -9.59073544e-01 1.11299366e-01
-9.15693462e-01 1.11248054e-01 1.08409882e+00 5.32496035e-01
4.60190922e-01 -4.68499005e-01 -3.89197260e-01 5.80816269e-01
-3.75199392e-02 2.92825967e-01 4.53981966e-01 -4.06804442e-01
1.89535394e-01 5.57136118e-01 2.75903970e-01 -4.43172812e-01
-9.20801461e-01 -5.34290016e-01 -1.16565049e+00 1.81155324e-01
1.08985090e+00 -1.94304913e-01 -8.00531149e-01 2.21498227e+00
3.69847864e-01 1.37005389e-01 3.60831738e-01 8.24741960e-01
4.19360220e-01 1.95836172e-01 2.55742997e-01 -3.21792334e-01
1.61389184e+00 -4.81851965e-01 -5.87108910e-01 -3.13402772e-01
1.04246712e+00 -3.53993267e-01 8.51179540e-01 2.85266399e-01
-5.35929859e-01 1.29800271e-02 -5.19061089e-01 7.13749081e-02
-6.58060014e-01 7.65971690e-02 8.04880381e-01 6.18139923e-01
-7.99052477e-01 7.71275461e-01 -7.83349633e-01 -8.30352962e-01
3.92423391e-01 2.93649197e-01 -5.76437891e-01 -2.02142000e-01
-1.43715084e+00 1.03462183e+00 5.10772407e-01 -3.69393259e-01
-4.83817399e-01 -8.71982932e-01 -6.62513733e-01 2.43946254e-01
4.53078419e-01 -1.20121896e+00 7.64265418e-01 -1.06405163e+00
-8.94445658e-01 9.71379340e-01 -2.16890201e-02 -4.51830298e-01
4.88032937e-01 1.11764446e-01 -4.65476245e-01 -3.20563257e-01
1.32181928e-01 6.92305863e-02 5.89017332e-01 -8.57649326e-01
-7.10964441e-01 -7.53709793e-01 -8.00881460e-02 6.98310509e-02
-5.41611791e-01 -1.27859358e-02 2.59054708e-03 -6.83798671e-01
-3.40712279e-01 -1.01490986e+00 -1.99824005e-01 -6.76294863e-02
-2.49315664e-01 -2.80553550e-01 2.77479172e-01 -6.76019669e-01
1.13828695e+00 -2.19245577e+00 5.73624074e-01 -5.25319017e-02
3.31751764e-01 -2.55693734e-01 1.15669817e-02 4.48229849e-01
-3.48204285e-01 8.78143758e-02 -3.30484927e-01 -3.84026945e-01
-1.07607752e-01 1.38231814e-01 -5.23424707e-02 5.77939928e-01
4.95813377e-02 6.65997922e-01 -7.36997724e-01 -5.15967250e-01
-2.03391016e-01 4.95971680e-01 -3.70159000e-01 1.51037708e-01
-1.70843303e-02 9.03869152e-01 -5.47252953e-01 7.07315922e-01
3.97283226e-01 -6.09763205e-01 2.74706930e-02 9.62526351e-02
3.48310739e-01 -7.73294345e-02 -9.35542524e-01 1.60923409e+00
-2.87979007e-01 1.67582974e-01 -5.90793006e-02 -1.22829711e+00
6.25473917e-01 4.83725011e-01 7.91479468e-01 -2.67285794e-01
2.26881325e-01 3.07786942e-01 -1.32592037e-01 -3.61794442e-01
-1.96882952e-02 -7.33871698e-01 -3.39387387e-01 1.75572366e-01
5.30536566e-03 3.29636425e-01 8.02192017e-02 2.55830456e-02
1.36403215e+00 -1.69827357e-01 5.72267830e-01 -3.32693219e-01
5.50763071e-01 1.88037485e-01 7.98690140e-01 4.45590138e-01
-1.55226573e-01 5.99433541e-01 7.25684762e-01 -4.85370129e-01
-7.36338139e-01 -9.63469207e-01 -6.12236083e-01 1.10404766e+00
-1.76748127e-01 1.20590784e-01 -3.80536348e-01 -8.70528579e-01
1.22241423e-01 7.24035084e-01 -9.46462929e-01 -3.39014530e-01
-1.43293515e-01 -1.21422708e+00 3.53884369e-01 4.75472778e-01
-2.31524736e-01 -9.11307156e-01 -4.37988549e-01 1.98043376e-01
-3.38454515e-01 -7.50276029e-01 -4.92615759e-01 6.50525033e-01
-9.36791897e-01 -1.13578629e+00 -1.14832890e+00 -6.23750150e-01
4.11900908e-01 -3.37480277e-01 1.42653310e+00 -3.89176530e-05
3.68239060e-02 1.88522354e-01 -4.62799966e-01 -2.45864674e-01
-4.72919196e-01 4.33172852e-01 1.40078738e-01 1.53968960e-01
3.30396444e-01 -6.30292058e-01 -6.24427795e-01 2.78844506e-01
-1.02721584e+00 -2.81827357e-02 5.39828122e-01 1.05729103e+00
3.60266596e-01 1.98097304e-02 9.82432246e-01 -1.22530055e+00
3.11940193e-01 -1.01698756e+00 -1.91736281e-01 4.90809083e-01
-7.84080923e-01 2.34321892e-01 7.67742336e-01 -6.24921381e-01
-7.95540810e-01 1.70181125e-01 1.55753225e-01 -1.11577921e-01
-4.46090013e-01 8.25339139e-01 -3.04375648e-01 4.32689607e-01
6.32962286e-01 -2.88486406e-02 8.16452205e-02 -4.99068469e-01
2.30414361e-01 4.77619916e-01 3.28454584e-01 -3.73489320e-01
4.86633629e-01 5.35983443e-01 2.44587809e-01 -3.97266120e-01
-8.68548155e-01 -5.49370348e-01 -8.38945210e-01 2.59862304e-01
9.58482683e-01 -6.94464087e-01 -6.86184347e-01 3.82350743e-01
-8.16642702e-01 -3.21825683e-01 -3.81381154e-01 5.85821271e-01
-5.37609816e-01 1.21474266e-01 -2.49743894e-01 -5.21642625e-01
-1.68895304e-01 -7.86973178e-01 9.45355833e-01 3.42642888e-02
-2.79955864e-01 -1.64333415e+00 3.13807011e-01 2.80285299e-01
4.24434543e-01 6.09092712e-01 1.27070606e+00 -1.17351198e+00
5.68780527e-02 -3.31166416e-01 -5.33514358e-02 2.75883339e-02
4.23203558e-01 -4.58847612e-01 -6.73094690e-01 -5.96390247e-01
-2.59003360e-02 -1.80241406e-01 1.04456663e+00 6.57944322e-01
9.08375680e-01 -2.51247793e-01 -7.48268247e-01 6.28755867e-01
1.26243722e+00 4.79472503e-02 2.55609483e-01 2.27862552e-01
5.06664753e-01 9.59282100e-01 5.72483659e-01 4.98294175e-01
6.24699652e-01 8.40493381e-01 3.50753903e-01 -1.78134844e-01
1.79826036e-01 1.29619867e-01 5.28899990e-02 2.67706543e-01
1.31667957e-01 -6.31007671e-01 -1.06432831e+00 7.24228680e-01
-1.85700071e+00 -7.17178524e-01 -3.02375019e-01 2.58770084e+00
8.99252713e-01 -2.96862274e-02 3.68079603e-01 1.59684613e-01
4.88846511e-01 -8.19425806e-02 -7.94134319e-01 2.68547554e-02
-1.96905971e-01 -7.51847997e-02 6.66398048e-01 2.95334637e-01
-1.24370515e+00 4.18937534e-01 5.29493713e+00 6.35304570e-01
-1.09331906e+00 1.55562431e-01 7.00768530e-01 -2.72758454e-02
-2.57248610e-01 2.55609393e-01 -5.79285204e-01 5.74321210e-01
1.12559557e+00 -3.80523235e-01 1.50677189e-01 5.45794845e-01
7.97080323e-02 -3.76074109e-03 -1.48125565e+00 8.04889202e-01
-5.06301969e-02 -7.78880775e-01 -3.67610097e-01 2.09979504e-01
3.88716102e-01 -2.27057084e-01 -1.50085324e-02 4.25357223e-01
4.10192221e-01 -1.06352532e+00 3.61702740e-01 8.91541541e-01
9.20837224e-01 -4.90518034e-01 8.75677228e-01 4.91394162e-01
-8.40426743e-01 -2.06044078e-01 -8.88667107e-02 1.45311072e-01
2.95364797e-01 8.17578375e-01 -9.33779240e-01 8.48246872e-01
5.49309313e-01 7.38503695e-01 -6.00023568e-01 1.03481638e+00
2.67345816e-01 5.86127222e-01 -1.39497057e-01 2.26739764e-01
-2.20340565e-01 -8.99552703e-02 2.97330707e-01 7.40652204e-01
5.94634771e-01 1.41029015e-01 1.32542685e-01 6.20218515e-01
7.52683580e-02 1.78991735e-01 -5.26051521e-01 3.25932130e-02
7.63127953e-02 1.27949870e+00 -8.02955866e-01 -1.37886122e-01
-5.51936448e-01 1.14396942e+00 2.67833412e-01 3.33728790e-01
-7.28410780e-01 5.35090826e-02 5.67072034e-01 2.94433743e-01
-1.60286427e-01 6.85659945e-02 -4.29042816e-01 -1.35248137e+00
-3.90761018e-01 -5.62154710e-01 9.98119056e-01 -3.87850970e-01
-1.74889970e+00 5.69640756e-01 -3.78777534e-02 -1.20438755e+00
-2.58185178e-01 -3.25872153e-01 -3.36618423e-01 8.99693787e-01
-1.65581822e+00 -1.14999104e+00 -1.42741159e-01 7.48409927e-01
1.76681295e-01 -5.36714792e-02 1.03462446e+00 3.91173124e-01
-5.76189935e-01 4.24520999e-01 5.18064916e-01 -8.99241194e-02
1.22994268e+00 -1.56763482e+00 -5.36637045e-02 3.17188829e-01
9.79108922e-03 1.35468662e-01 7.95518816e-01 -6.52393401e-01
-1.01420975e+00 -1.09364951e+00 1.23897421e+00 -9.33450937e-01
5.79795659e-01 -2.84874707e-01 -1.32780802e+00 5.52973092e-01
-2.99280912e-01 1.37643859e-01 9.39766765e-01 5.51717937e-01
-1.19687408e-01 -1.23545276e-02 -1.26971865e+00 1.94983363e-01
7.52721488e-01 -2.46451125e-01 -5.06471276e-01 5.01580656e-01
5.10094047e-01 -1.46370888e-01 -1.09231031e+00 4.57797647e-01
4.95569378e-01 -6.57088578e-01 9.74297225e-01 -9.20407593e-01
2.90779501e-01 -5.75360134e-02 1.20088356e-02 -1.61758506e+00
-3.66171986e-01 -8.79241526e-02 7.45479017e-02 1.36085784e+00
4.23157364e-01 -7.48831928e-01 7.91862011e-01 8.92529726e-01
2.44390085e-01 -6.03317082e-01 -1.15797412e+00 -9.05678868e-01
5.55122256e-01 4.90010381e-02 6.87404931e-01 1.21321738e+00
-4.55687866e-02 4.93479431e-01 -4.95469987e-01 2.95386165e-01
5.11872828e-01 2.67180949e-01 4.47909534e-01 -1.86298013e+00
-1.81681424e-01 -1.79697022e-01 -1.79202482e-01 -4.29186761e-01
3.47506523e-01 -1.02151430e+00 -2.59985566e-01 -1.57174814e+00
4.70453531e-01 -6.37651145e-01 -5.94799817e-01 6.98094010e-01
-3.34037125e-01 -2.11699650e-01 -1.81779608e-01 2.40366697e-01
-4.03518349e-01 5.56340277e-01 8.73883545e-01 -9.14623886e-02
-3.08154225e-01 4.56426144e-01 -7.36996531e-01 5.83904326e-01
6.86368585e-01 -9.07410979e-01 -2.92902887e-01 -1.13624642e-02
2.45864496e-01 7.36188054e-01 4.76045668e-01 -7.80085266e-01
2.29161754e-01 -4.38126862e-01 1.57182053e-01 -2.69855678e-01
1.86628759e-01 -1.21175349e+00 4.13336694e-01 4.54791576e-01
-2.52444059e-01 -1.83583334e-01 -1.42564118e-01 7.75727987e-01
-1.24813668e-01 -1.11420721e-01 6.93293214e-01 2.00273499e-01
-4.33869600e-01 3.62219602e-01 -9.47315246e-02 2.49645501e-01
1.13549435e+00 -1.40731320e-01 -1.39008299e-01 -2.02536985e-01
-1.11533511e+00 5.13835907e-01 5.87560534e-01 4.85660225e-01
2.57032782e-01 -1.26500559e+00 -9.37750041e-01 7.52616115e-03
3.46736401e-01 -1.42822385e-01 4.42045838e-01 1.05289626e+00
1.35830283e-01 4.17762339e-01 -1.29961163e-01 -3.89056474e-01
-1.32418132e+00 9.46794450e-01 2.77184248e-01 -5.98922193e-01
-4.00434464e-01 4.72746283e-01 3.71314108e-01 -3.16163212e-01
7.92688578e-02 -3.15409452e-01 -5.84381640e-01 6.27888501e-01
1.58555031e-01 2.39682049e-01 1.22138485e-01 -5.77970207e-01
-5.99951267e-01 3.60830337e-01 -4.62228507e-02 1.35406211e-01
1.51719105e+00 -2.28999004e-01 8.58372897e-02 6.65413857e-01
1.03373945e+00 -1.85702384e-01 -1.00599635e+00 -3.58755559e-01
4.23291415e-01 -2.73892879e-01 -8.47145617e-02 -1.15686035e+00
-1.04467165e+00 6.30562067e-01 7.25252211e-01 3.56510699e-01
1.36841762e+00 2.99544156e-01 5.69897175e-01 -1.95505172e-01
3.25236499e-01 -6.38585627e-01 -2.42465422e-01 3.10212344e-01
7.43316531e-01 -1.57299697e+00 -1.92676440e-01 -1.05756149e-01
-7.83127546e-01 9.32753444e-01 8.49373564e-02 3.34924608e-01
7.66075552e-01 9.88380611e-02 -8.60914737e-02 -2.03469560e-01
-8.00310373e-01 -4.08621341e-01 3.14659715e-01 5.84143400e-01
4.96708721e-01 2.66346067e-01 -5.07592261e-01 1.11356890e+00
-7.13933483e-02 2.59827912e-01 4.46757019e-01 5.31267822e-01
-3.98161337e-02 -1.39029109e+00 -4.55377400e-01 7.46784627e-01
-7.50228584e-01 6.15646550e-03 -1.82739034e-01 6.17817819e-01
5.78197651e-02 8.82161379e-01 -1.94137827e-01 3.07118353e-02
4.58515078e-01 3.29274803e-01 4.70306054e-02 -7.42567658e-01
-3.45342994e-01 -2.79063255e-01 -1.14943348e-01 -1.59665868e-01
-4.51653004e-01 -8.12007666e-01 -1.11563265e+00 -1.56530380e-01
-4.35052902e-01 2.34737933e-01 4.88945425e-01 9.20669615e-01
2.42350593e-01 7.14435339e-01 5.54184914e-01 -3.88900250e-01
-6.23028159e-01 -6.95388198e-01 -9.40659404e-01 5.81626058e-01
6.48393154e-01 -7.47405112e-01 -6.38197601e-01 1.69218212e-01] | [7.7536725997924805, 5.546572208404541] |
70e923ba-d885-4b9e-9efa-2f096d8adcb9 | digit-recognition-in-handwritten-weather | 1304.6933 | null | http://arxiv.org/abs/1304.6933v2 | http://arxiv.org/pdf/1304.6933v2.pdf | Digit Recognition in Handwritten Weather Records | This paper addresses the automatic recognition of handwritten temperature
values in weather records. The localization of table cells is based on line
detection using projection profiles. Further, a stroke-preserving line removal
method which is based on gradient images is proposed. The presented digit
recognition utilizes features which are extracted using a set of filters and a
Support Vector Machine classifier. It was evaluated on the MNIST and the USPS
dataset and our own database with about 17,000 RGB digit images. An accuracy of
99.36% per digit is achieved for the entire system using a set of 84 weather
records. | ['Robert Sablatnig', 'Manuel Keglevic'] | 2013-04-25 | null | null | null | null | ['line-detection'] | ['computer-vision'] | [ 1.15979224e-01 -6.56201363e-01 1.56066582e-01 -5.34924686e-01
-1.80363670e-01 -8.16097319e-01 5.83696425e-01 2.59763360e-01
-6.54334307e-01 6.38938010e-01 -4.35249299e-01 -3.30980897e-01
1.12338133e-01 -9.75032508e-01 -2.66813815e-01 -8.12731922e-01
-7.15672150e-02 -2.66705733e-02 2.29894057e-01 -1.11888774e-01
1.04161704e+00 1.25048530e+00 -1.17639422e+00 3.74166906e-01
4.24343497e-01 1.43324447e+00 -2.88241416e-01 1.00146520e+00
-4.71367091e-02 8.00805449e-01 -9.89367366e-01 -3.88728499e-01
5.67005754e-01 -2.03114778e-01 -1.80716664e-01 1.99970737e-01
6.80434942e-01 -5.19225895e-01 -4.92641896e-01 7.12014258e-01
4.85514998e-01 -7.13017657e-02 1.03204560e+00 -9.42967057e-01
-6.30857408e-01 2.71318972e-01 -6.78330243e-01 2.24612981e-01
1.77452534e-01 -1.53281897e-01 2.55985171e-01 -8.88147116e-01
3.68230283e-01 5.91077149e-01 7.84523666e-01 -8.33914652e-02
-9.81746614e-01 -3.19287181e-01 -6.56801403e-01 3.60595644e-01
-1.64100230e+00 -9.66606811e-02 9.84058678e-01 -4.57260221e-01
1.30647731e+00 3.74848902e-01 5.62871516e-01 3.29559445e-01
5.77418327e-01 2.54882365e-01 1.75374281e+00 -8.67742717e-01
5.31982839e-01 2.89815843e-01 6.03637695e-01 8.57257068e-01
5.15576959e-01 -1.38299450e-01 -6.04441226e-01 -4.32540774e-02
9.16990578e-01 -7.07071424e-02 -1.16532587e-01 -1.27518892e-01
-8.36002231e-01 5.86133122e-01 3.13739777e-01 4.90524888e-01
-4.07377481e-01 -1.02615133e-01 2.77508289e-01 1.58224434e-01
5.32531179e-02 3.74078989e-01 -2.33801410e-01 -7.69900009e-02
-1.17166889e+00 -2.13974133e-01 1.13198233e+00 7.97020257e-01
4.52863723e-01 3.85902792e-01 1.96382076e-01 5.68803847e-01
2.50367075e-01 1.12181950e+00 5.99265337e-01 -2.90474683e-01
4.68635112e-01 9.33427572e-01 1.48841366e-01 -1.19775474e+00
-5.69067240e-01 4.13424931e-02 -9.49809194e-01 8.76771092e-01
6.17108464e-01 -2.13304847e-01 -1.32183218e+00 4.39598143e-01
1.84295494e-02 -3.23946625e-01 3.11263442e-01 7.34214664e-01
5.06757915e-01 7.82189012e-01 -7.01766670e-01 -1.15175976e-03
1.26222897e+00 -4.83195245e-01 -8.06785285e-01 1.49953261e-01
1.12447046e-01 -9.06567216e-01 6.71456337e-01 1.07191098e+00
-4.87217635e-01 -5.01409292e-01 -1.61619067e+00 9.10512134e-02
-1.01870644e+00 1.05131257e+00 4.23612893e-01 1.16456234e+00
-6.46867335e-01 5.25470376e-01 -7.83575594e-01 -3.92219037e-01
2.84555316e-01 3.27006876e-01 -3.52925599e-01 1.91382289e-01
-4.36347991e-01 1.12325895e+00 2.98537761e-01 4.49816823e-01
1.29211590e-01 1.16834737e-01 -6.58690929e-01 -2.74423748e-01
-4.04298633e-01 1.93510458e-01 5.55343211e-01 -6.99948609e-01
-1.88927257e+00 8.17463398e-01 8.96542296e-02 -6.77488744e-01
8.62116039e-01 -3.62510122e-02 -7.60704577e-01 6.39028251e-01
-5.27922451e-01 -6.20554835e-02 7.78567970e-01 -7.83696532e-01
-6.12474263e-01 -8.41627359e-01 -9.28407848e-01 -1.24406025e-01
-5.54804981e-01 -2.60158688e-01 -6.48997575e-02 -7.56058693e-01
6.41443908e-01 -6.42334104e-01 2.30332971e-01 -2.63916180e-02
-6.82038128e-01 3.44542384e-01 1.12297237e+00 -1.18721986e+00
8.92501831e-01 -1.94330859e+00 -5.33626854e-01 9.64843154e-01
-5.13228834e-01 4.33714576e-02 5.03836811e-01 3.43233973e-01
2.83276647e-01 -3.93630981e-01 -4.89063025e-01 1.33407965e-01
-1.49354577e-01 -1.27629638e-01 -5.97149968e-01 9.48539495e-01
-4.79304790e-02 2.72578329e-01 -2.23966558e-02 -3.01279485e-01
7.36654520e-01 5.76221585e-01 3.75564635e-01 -1.51333198e-01
5.40881276e-01 -3.58530074e-01 -1.47321656e-01 8.78043711e-01
1.11215961e+00 3.96927714e-01 4.56676399e-03 -3.69333655e-01
-6.89239979e-01 -3.06066453e-01 -1.58365607e+00 1.26248407e+00
-8.54614899e-02 1.01874352e+00 -3.92169952e-01 -5.88128567e-01
1.90037954e+00 -2.07923371e-02 2.81300485e-01 -8.19198668e-01
1.59499153e-01 5.17749667e-01 -2.84404248e-01 -3.73186588e-01
8.31945002e-01 4.06871259e-01 -6.72186911e-02 4.10239190e-01
-2.07964376e-01 -2.72508860e-01 1.70793012e-01 -1.49112672e-01
8.74357045e-01 1.65708095e-01 2.02917516e-01 -5.25500059e-01
5.00562787e-01 4.65788990e-01 1.54148206e-01 6.94897532e-01
-2.54069179e-01 5.88307142e-01 1.17274068e-01 -6.90913677e-01
-1.01292932e+00 -8.77070665e-01 -5.63234866e-01 2.73423910e-01
2.38509223e-01 1.65287778e-01 -7.44366586e-01 -1.61264598e-01
2.69500583e-01 5.73621690e-01 -7.24879920e-01 1.02311388e-01
-6.23654902e-01 -8.44103396e-01 9.04049397e-01 7.82355845e-01
1.27143061e+00 -8.25246274e-01 -1.38076663e+00 1.83215544e-01
5.49620569e-01 -9.88683283e-01 1.73207551e-01 5.61232209e-01
-1.08081675e+00 -1.11543536e+00 -8.75878632e-01 -7.39186883e-01
8.07508349e-01 -2.65724719e-01 4.68733191e-01 -4.22439575e-01
-8.88182998e-01 4.16252136e-01 -2.08543986e-01 -5.54565489e-01
1.67097643e-01 -3.45842205e-02 -4.32281680e-02 1.37491912e-01
7.33990550e-01 6.41041175e-02 -4.80008066e-01 2.29476467e-01
-3.08287144e-01 -5.65850854e-01 5.68882406e-01 4.93026674e-01
4.27432686e-01 7.26722926e-02 3.10895313e-02 -3.17031801e-01
5.06236076e-01 3.86357665e-01 -1.22317147e+00 5.32987356e-01
-7.43968904e-01 1.06746607e-01 6.84876919e-01 -7.24118203e-02
-1.15830636e+00 5.92970133e-01 3.42975944e-01 2.07527027e-01
-3.04733485e-01 -5.42799979e-02 1.26973838e-01 -6.05679035e-01
9.08857465e-01 6.16220534e-01 -2.31046498e-01 -4.27521348e-01
-8.69442057e-03 1.00338304e+00 1.06056547e+00 -1.18691705e-01
6.53949797e-01 6.87105358e-01 2.36317515e-01 -1.28652382e+00
2.81918734e-01 -3.65586698e-01 -1.14297569e+00 -4.11169648e-01
7.75360584e-01 -3.37312102e-01 -6.28644526e-01 1.37839782e+00
-9.67999101e-01 -1.44637123e-01 -1.32397842e-02 7.36732483e-01
-1.90904081e-01 3.70300531e-01 -7.64402270e-01 -1.19127619e+00
-5.56472003e-01 -7.73535311e-01 4.36101437e-01 5.63327730e-01
1.41114220e-01 -7.23848224e-01 9.24163461e-02 -1.25119656e-01
3.08597237e-01 4.15197343e-01 4.59860593e-01 -4.62759435e-01
-2.79072523e-01 -8.02856326e-01 -3.18031274e-02 3.93546611e-01
2.53531665e-01 7.08171785e-01 -1.06697488e+00 3.38058285e-02
-1.35704458e-01 -2.16929838e-01 1.04074955e+00 3.01076829e-01
7.60201037e-01 9.87097323e-02 -1.51242658e-01 6.01294041e-01
1.86181724e+00 7.50865459e-01 7.88777769e-01 8.93647194e-01
5.36134481e-01 2.48581633e-01 4.28825825e-01 7.53534496e-01
-3.32236812e-02 2.99280643e-01 6.63072318e-02 -7.13179335e-02
2.41584435e-01 3.37190270e-01 6.40113465e-03 3.09897304e-01
-2.99190104e-01 -2.73787361e-02 -1.28399003e+00 1.67430550e-01
-1.26958525e+00 -6.39555097e-01 -5.02194047e-01 2.20966983e+00
4.74575490e-01 2.68176824e-01 -8.63324255e-02 9.73451138e-01
7.87485301e-01 -1.60626739e-01 -5.09747863e-01 -8.66052091e-01
-4.32207584e-01 3.14184427e-01 1.37980521e+00 4.32283342e-01
-1.32636666e+00 6.48899317e-01 6.44420767e+00 3.41754258e-02
-1.69073129e+00 -8.56051624e-01 7.56086946e-01 5.27236104e-01
4.95330751e-01 -6.66028082e-01 -9.09038723e-01 2.69161016e-01
4.89556134e-01 1.83315247e-01 8.62324387e-02 7.83719778e-01
-1.19993150e-01 -6.48444712e-01 -4.91886914e-01 1.15966845e+00
5.24476945e-01 -1.04370141e+00 -2.49637052e-01 -1.13460071e-01
5.14056742e-01 -3.36796165e-01 3.26736957e-01 -3.43709081e-01
-1.22367166e-01 -8.08772743e-01 7.85590172e-01 7.94444740e-01
7.92411923e-01 -1.01262367e+00 7.32910097e-01 -9.79582891e-02
-1.04787111e+00 -1.71659216e-01 -5.07550359e-01 -2.07874134e-01
-4.04541910e-01 5.60069859e-01 -9.15557802e-01 2.63992131e-01
6.81677222e-01 2.94776320e-01 -7.52357483e-01 1.23451984e+00
-2.06082165e-01 5.24955153e-01 -7.92702913e-01 -2.54756540e-01
1.20265931e-01 -3.97571564e-01 -1.42433539e-01 1.70391130e+00
5.47901630e-01 1.80598766e-01 -4.54345256e-01 3.77290815e-01
1.98928088e-01 3.33587378e-01 -5.73185265e-01 2.73475349e-01
5.90893030e-01 1.34373009e+00 -1.43407655e+00 -3.75861347e-01
-1.48152113e-01 1.30274177e+00 -2.87011802e-01 2.67419040e-01
-4.10995692e-01 -1.27558851e+00 -8.68756697e-03 -3.37934524e-01
6.71390831e-01 -5.29368997e-01 -1.18391001e+00 -9.12167966e-01
1.75075784e-01 -3.08506131e-01 2.50011355e-01 -6.32464349e-01
-8.57366204e-01 6.31657422e-01 -3.16081375e-01 -1.15849328e+00
-2.70668387e-01 -1.32747996e+00 -5.62959433e-01 1.03808045e+00
-1.28772795e+00 -7.11522222e-01 -5.81106126e-01 7.96014845e-01
7.33934417e-02 -5.38130522e-01 9.06028092e-01 2.07865257e-02
-6.63233638e-01 6.75329506e-01 7.53185093e-01 6.04079187e-01
6.73216045e-01 -1.44305539e+00 3.96591634e-01 1.14125109e+00
6.01453781e-02 2.19940692e-01 4.99647290e-01 -5.29146314e-01
-1.47828794e+00 -8.48539174e-01 7.85083771e-01 -5.96618392e-02
1.30196899e-01 -3.99729818e-01 -8.75505388e-01 3.98599595e-01
1.91853359e-01 1.61028340e-01 6.05439663e-01 -8.14749002e-01
-4.76752460e-01 -5.32798886e-01 -1.77256644e+00 -7.61227980e-02
6.28862083e-02 -4.10409033e-01 -6.48238361e-01 4.21876051e-02
-6.80339754e-01 -5.42045355e-01 -9.88488615e-01 5.53360656e-02
7.95286655e-01 -9.28493023e-01 7.40867734e-01 1.43323734e-01
4.21602204e-02 -4.58516419e-01 -2.17739761e-01 -1.01599002e+00
-2.08370257e-02 -3.87164615e-02 7.46963844e-02 1.14863479e+00
4.72681969e-01 -7.29187012e-01 9.43239391e-01 7.45462596e-01
2.59342581e-01 -2.42671773e-01 -8.37930202e-01 -7.56909192e-01
-2.92719901e-01 -1.19762354e-01 2.87754118e-01 1.04194021e+00
9.60863940e-03 -1.75151721e-01 -1.44927353e-01 2.28233904e-01
9.53816414e-01 3.96730900e-01 3.73475760e-01 -1.05470085e+00
7.45648220e-02 -3.07618558e-01 -9.31733489e-01 -3.37761909e-01
-3.83237153e-01 -2.60417968e-01 -3.34490389e-02 -1.44881737e+00
-5.59333801e-01 1.81849040e-02 -1.84134990e-01 4.63349164e-01
4.75406229e-01 7.09446013e-01 1.25137761e-01 1.99630633e-01
2.35148057e-01 -3.99097726e-02 5.37778974e-01 -2.18742192e-01
-2.51701593e-01 -1.48788139e-01 2.25012466e-01 6.51894271e-01
1.24676800e+00 -3.13765891e-02 1.04474410e-01 -9.20549780e-02
-2.14563116e-01 -3.47523153e-01 1.08625345e-01 -1.53303397e+00
5.42344809e-01 1.28049478e-01 1.35678434e+00 -1.06731737e+00
2.90857345e-01 -8.46971214e-01 -2.82012641e-01 1.04842544e+00
-9.59745273e-02 3.32747817e-01 3.51374418e-01 1.97623402e-01
-2.13764772e-01 -1.70832291e-01 7.89115012e-01 5.67513779e-02
-1.20738614e+00 -4.78326380e-01 -6.58318162e-01 -7.78551996e-01
1.29252064e+00 -7.25097477e-01 -4.04438376e-01 8.35094228e-02
-3.28232735e-01 -4.23314333e-01 4.32592720e-01 -1.51754871e-01
9.42009509e-01 -1.20542514e+00 -4.74279702e-01 4.75692958e-01
1.32315814e-01 -5.20776927e-01 -1.93112418e-01 2.91352838e-01
-1.41586673e+00 4.42225248e-01 -8.36822629e-01 -4.80643302e-01
-1.51110709e+00 1.17708370e-01 6.40037119e-01 1.96484402e-01
-6.08106554e-01 6.50796413e-01 -1.20582187e+00 -6.67545274e-02
3.84655505e-01 -7.80220091e-01 -3.94756436e-01 1.52635187e-01
6.46730363e-01 8.06455553e-01 4.39536542e-01 -4.70534056e-01
-6.90649390e-01 9.85599577e-01 1.43336475e-01 -4.41052973e-01
1.25974393e+00 4.33367640e-01 1.03109730e-02 5.39362192e-01
9.75259066e-01 -3.16561796e-02 -1.27974510e+00 6.88312352e-02
4.60169353e-02 -5.33554018e-01 5.96981458e-02 -1.18033123e+00
-1.06704736e+00 9.58048522e-01 1.19446373e+00 1.55834436e-01
1.21560657e+00 -1.02209115e+00 2.52628595e-01 1.00296509e+00
2.75882304e-01 -1.68673813e+00 -3.48873675e-01 3.78430724e-01
6.69618428e-01 -9.87942040e-01 2.60141283e-01 -1.04978532e-01
-5.57393968e-01 2.12038660e+00 4.12027448e-01 -4.91254032e-01
4.96019691e-01 8.97814751e-01 4.42691028e-01 5.65067381e-02
9.61705372e-02 1.64653763e-01 -4.57366668e-02 5.14634967e-01
4.36098397e-01 1.51823744e-01 -4.71776038e-01 8.91709328e-02
-3.67058009e-01 1.92425147e-01 7.48623013e-01 1.63698077e+00
-6.53967619e-01 -4.62113857e-01 -1.09920752e+00 4.27191675e-01
-3.20871770e-01 1.04305767e-01 -8.08212459e-01 8.68956625e-01
-2.42325947e-01 7.92368114e-01 4.05500114e-01 -5.66121638e-01
4.76446390e-01 1.60779402e-01 7.61235893e-01 2.87015110e-01
-6.03707135e-01 -4.09193397e-01 -2.77202994e-01 -2.35496685e-01
-1.84147149e-01 -6.37885511e-01 -1.44562626e+00 -2.26098716e-01
3.66075360e-03 -1.61459029e-01 1.47500205e+00 5.34850478e-01
-3.98903750e-02 1.68423504e-01 7.61559129e-01 -4.64822084e-01
-6.02066278e-01 -9.29253042e-01 -1.13219249e+00 -8.30494538e-02
1.66871831e-01 -1.68153226e-01 -3.02755088e-01 3.86628181e-01] | [11.841126441955566, 2.655866861343384] |
8bdcbe38-cef3-4988-977f-819ac702c7f1 | octree-transformer-autoregressive-3d-shape | 2111.12480 | null | https://arxiv.org/abs/2111.12480v1 | https://arxiv.org/pdf/2111.12480v1.pdf | Octree Transformer: Autoregressive 3D Shape Generation on Hierarchically Structured Sequences | Autoregressive models have proven to be very powerful in NLP text generation tasks and lately have gained popularity for image generation as well. However, they have seen limited use for the synthesis of 3D shapes so far. This is mainly due to the lack of a straightforward way to linearize 3D data as well as to scaling problems with the length of the resulting sequences when describing complex shapes. In this work we address both of these problems. We use octrees as a compact hierarchical shape representation that can be sequentialized by traversal ordering. Moreover, we introduce an adaptive compression scheme, that significantly reduces sequence lengths and thus enables their effective generation with a transformer, while still allowing fully autoregressive sampling and parallel training. We demonstrate the performance of our model by comparing against the state-of-the-art in shape generation. | ['Leif Kobbelt', 'Gregor Kobsik', 'Moritz Ibing'] | 2021-11-24 | null | null | null | null | ['3d-shape-generation'] | ['computer-vision'] | [ 3.16653848e-01 1.20644487e-01 1.50445491e-01 1.29860610e-01
-8.13092649e-01 -6.57146811e-01 9.56084907e-01 8.88808072e-02
-1.24828011e-01 7.20655680e-01 3.67576957e-01 -3.85495305e-01
8.45288113e-02 -1.08494246e+00 -5.76120019e-01 -6.03097618e-01
5.39188609e-02 9.09139514e-01 1.58875570e-01 -2.47319981e-01
1.79029092e-01 7.93684661e-01 -1.69319403e+00 1.87925071e-01
8.40651929e-01 7.83224881e-01 3.79271954e-01 7.69803345e-01
-3.34798753e-01 4.05332476e-01 -5.58586895e-01 -3.78736943e-01
1.18995287e-01 -4.10977155e-01 -6.30405009e-01 2.48588875e-01
2.42535114e-01 -3.39054644e-01 -2.09691077e-01 4.76787627e-01
5.80679178e-01 7.95431212e-02 9.91867185e-01 -8.76003146e-01
-3.69850755e-01 2.57041931e-01 -4.00050461e-01 -3.64078909e-01
3.57253700e-01 -1.19660020e-01 9.78297234e-01 -7.63308346e-01
8.34543467e-01 1.28251421e+00 4.92439717e-01 6.71297193e-01
-1.45542145e+00 -2.64186531e-01 -1.08005650e-01 -7.19841570e-02
-1.08665407e+00 -6.11693621e-01 7.64049888e-01 -4.27101672e-01
1.18079436e+00 3.01778883e-01 8.21238458e-01 8.68251979e-01
7.55800679e-02 8.32130015e-01 8.34263146e-01 -6.75438881e-01
2.54316241e-01 -8.69479328e-02 -4.66204613e-01 7.14985192e-01
3.01194913e-03 -3.40405516e-02 -2.49540836e-01 -3.10073584e-01
1.12270546e+00 -3.57575774e-01 -1.59269065e-01 -5.55065989e-01
-1.03402400e+00 9.60375130e-01 2.26193815e-01 4.04024184e-01
-3.96277785e-01 3.95473003e-01 2.69907892e-01 1.49718717e-01
6.45063400e-01 4.11468178e-01 -2.74192363e-01 -2.18457028e-01
-1.06628001e+00 7.03102887e-01 7.82401443e-01 1.05106068e+00
4.33572829e-01 3.93376619e-01 -9.48441178e-02 9.51818049e-01
1.22749500e-01 3.89786303e-01 3.76839191e-01 -8.35806549e-01
3.17990512e-01 3.34396988e-01 6.09072708e-02 -8.97802293e-01
-1.20115951e-01 -2.59560645e-01 -9.15499568e-01 2.77613163e-01
3.64657819e-01 1.16986647e-01 -9.43740487e-01 1.43355834e+00
2.88889527e-01 -1.05909035e-02 -1.19524756e-02 3.95471185e-01
4.59524274e-01 9.19413030e-01 -1.09419428e-01 -2.74396807e-01
1.21863198e+00 -6.81033492e-01 -4.20826703e-01 5.07106893e-02
5.45596719e-01 -9.52690184e-01 9.17242825e-01 4.86208618e-01
-1.52300823e+00 -4.30726647e-01 -7.73585320e-01 -1.37595475e-01
-1.54884398e-01 4.17133997e-04 8.13405037e-01 8.73908699e-01
-1.35828149e+00 6.93565071e-01 -8.97205651e-01 -9.16770399e-02
4.23046678e-01 3.23637724e-01 -2.23700866e-01 -3.01750638e-02
-7.68364012e-01 6.94435239e-01 2.79324234e-01 -3.75512481e-01
-5.07649481e-01 -6.98996603e-01 -9.11725700e-01 1.51656032e-01
1.52160272e-01 -9.29738879e-01 1.31014073e+00 -4.10026282e-01
-1.73619616e+00 6.58976793e-01 -4.25204992e-01 -7.07420647e-01
4.80333537e-01 4.13619690e-02 1.60148248e-01 1.75874174e-01
-1.72054693e-01 9.01456714e-01 9.14004922e-01 -1.21598661e+00
-1.92106515e-01 -2.01853126e-01 -4.57586870e-02 3.44467789e-01
-3.11824679e-01 -2.33009070e-01 -5.20090699e-01 -8.55546057e-01
1.14948682e-01 -1.00097978e+00 -6.80253744e-01 -3.14515121e-02
-1.59461215e-01 -2.60345578e-01 6.71702743e-01 -5.11061847e-01
1.06446171e+00 -1.87999725e+00 2.20975801e-01 2.94391185e-01
7.41945133e-02 2.97472745e-01 -1.79931179e-01 6.75058782e-01
9.77844223e-02 2.73758262e-01 -3.17928553e-01 -5.35287976e-01
-9.37229022e-02 3.79848540e-01 -6.34761631e-01 -1.12093212e-02
2.69416660e-01 1.02917755e+00 -7.52094448e-01 -5.73089898e-01
3.16374302e-01 5.82657337e-01 -7.46399879e-01 6.07725009e-02
-5.53356171e-01 4.18421179e-01 -5.57952166e-01 1.26773641e-01
4.84546423e-01 -1.74492717e-01 1.12595424e-01 3.61722102e-03
-1.08146138e-01 3.78008097e-01 -9.40164983e-01 1.63923037e+00
-8.35898757e-01 3.23463053e-01 -2.08709121e-01 -8.68378043e-01
9.50881660e-01 5.58973849e-01 5.44020593e-01 -3.42469960e-01
-8.96143764e-02 3.04621339e-01 -3.05145085e-01 4.64027151e-02
7.45869815e-01 -3.41057748e-01 9.62186828e-02 5.30632257e-01
-2.35565409e-01 -1.04318953e+00 4.46762532e-01 8.06296319e-02
7.99409449e-01 3.64751726e-01 3.62522840e-01 6.20615892e-02
5.23426533e-01 -2.18115430e-02 3.31241228e-02 5.28765440e-01
6.82656109e-01 8.77709091e-01 5.26259303e-01 -2.25956723e-01
-1.43995047e+00 -1.03578210e+00 9.39023122e-02 6.01987243e-01
-3.50572050e-01 -4.95648354e-01 -7.46939123e-01 -3.64850461e-01
-2.03432918e-01 9.76991415e-01 -3.07039499e-01 3.04595798e-01
-7.43911207e-01 -5.45553267e-01 3.83884370e-01 5.68175495e-01
1.66835323e-01 -1.08722341e+00 -6.42762303e-01 4.22340333e-01
-2.05946318e-03 -1.17895508e+00 -3.93871754e-01 -1.97890460e-01
-1.26482975e+00 -7.13435590e-01 -9.80398953e-01 -6.00517094e-01
8.13397229e-01 3.03229451e-01 1.30908763e+00 7.05038309e-02
-3.58449429e-01 5.54325283e-01 -3.56610239e-01 -4.69586045e-01
-9.31431234e-01 1.72714069e-01 -3.17880929e-01 -3.51320922e-01
-3.20853502e-01 -7.83038676e-01 -4.08056170e-01 7.98554420e-02
-1.23950815e+00 4.20581192e-01 6.07059777e-01 8.20531070e-01
5.90576291e-01 -6.94044754e-02 3.51645321e-01 -8.27200055e-01
7.36539721e-01 2.62951385e-02 -7.80117989e-01 -2.22811121e-02
-4.54404414e-01 3.23011070e-01 7.15187490e-01 -1.23259969e-01
-1.06161773e+00 1.20673552e-01 -5.60853124e-01 -4.27942246e-01
-1.95876852e-01 3.27899516e-01 2.36875266e-01 -6.84225783e-02
4.88172650e-01 4.80087817e-01 1.30771905e-01 -3.98252189e-01
4.85100538e-01 4.40661222e-01 1.97823122e-01 -6.45927608e-01
8.33854973e-01 3.42719078e-01 4.78305101e-01 -1.29752922e+00
-3.55010241e-01 -2.52089500e-01 -4.59431857e-01 1.65630311e-01
6.63955390e-01 -6.23956323e-01 -4.74033386e-01 3.38667452e-01
-1.51510251e+00 -2.78421015e-01 -4.84255850e-01 -1.05486708e-02
-1.06547606e+00 5.48646629e-01 -7.20252693e-01 -1.05188322e+00
-5.51811099e-01 -1.00547194e+00 1.31145513e+00 -2.51264811e-01
-2.66476125e-01 -9.32539999e-01 -4.39854600e-02 1.12580776e-01
4.95171100e-01 3.38289320e-01 1.30049813e+00 -1.02050483e-01
-8.25753152e-01 -2.39351153e-01 -1.53572083e-01 1.29053280e-01
-7.87910223e-02 -2.55147964e-01 -7.03677595e-01 -2.02415049e-01
-2.20224440e-01 -2.35046312e-01 6.64519012e-01 4.53125775e-01
1.14379966e+00 -2.92882442e-01 -3.80189598e-01 2.69448757e-01
1.23487675e+00 6.66348189e-02 8.30505908e-01 -6.91363290e-02
4.83664840e-01 5.72845757e-01 2.69019246e-01 5.71731746e-01
1.74794525e-01 9.78230357e-01 1.30624354e-01 -2.23369375e-02
-2.97066420e-01 -4.69004273e-01 1.56255811e-02 1.03310704e+00
-2.62537032e-01 -4.04935271e-01 -8.69978845e-01 6.09317303e-01
-1.68543553e+00 -8.66647243e-01 -2.16986075e-01 2.36419725e+00
6.81190431e-01 3.88432965e-02 8.23992565e-02 3.00925761e-01
2.87774593e-01 2.00875491e-01 -1.85938850e-02 -4.72060978e-01
2.93918904e-02 7.00671077e-01 3.60008657e-01 5.01096189e-01
-7.98134506e-01 9.91979003e-01 6.91371965e+00 1.13288975e+00
-8.47461104e-01 -1.52450442e-01 7.08466232e-01 1.50984451e-01
-6.26892090e-01 -6.65501356e-02 -6.69403315e-01 1.85345754e-01
5.92094123e-01 -2.58374393e-01 3.63061130e-01 7.45636463e-01
2.43518069e-01 3.09562050e-02 -8.48859847e-01 1.16642201e+00
5.07412888e-02 -1.37671208e+00 4.41894710e-01 2.45099559e-01
9.04786408e-01 -3.11604917e-01 3.10518462e-02 8.25972781e-02
1.51930407e-01 -1.17276943e+00 6.32987201e-01 1.47666708e-01
9.17210937e-01 -7.97072351e-01 2.87398756e-01 5.55646360e-01
-1.28758526e+00 2.20723301e-01 -4.86392736e-01 1.29353330e-02
4.17308778e-01 6.84593856e-01 -1.15246177e+00 6.11214340e-01
6.18689917e-02 3.35360944e-01 -1.72876984e-01 1.06369054e+00
-2.44169310e-01 4.68443394e-01 -5.26440203e-01 -1.89210340e-01
2.50199735e-01 -3.27312797e-01 5.64346671e-01 1.06808722e+00
9.22136128e-01 -3.50250606e-03 1.05825163e-01 7.61236906e-01
3.41197513e-02 2.52661645e-01 -9.58250046e-01 -1.06550641e-01
3.07755619e-01 9.60047185e-01 -7.33201921e-01 -2.89810628e-01
-2.73804754e-01 1.00669789e+00 1.58930570e-01 1.98463619e-01
-4.52353477e-01 -9.95283425e-02 3.41415852e-01 5.10098934e-01
5.46357632e-01 -7.89040744e-01 -2.93166637e-01 -9.36850727e-01
-1.12785466e-01 -7.60278821e-01 4.99516912e-02 -7.20582902e-01
-9.93344069e-01 6.55678272e-01 1.10121742e-01 -1.20876431e+00
-8.93648088e-01 -4.69223857e-01 -3.14590782e-01 8.12777042e-01
-1.47724032e+00 -1.33668745e+00 2.77014244e-02 2.85638422e-01
7.10666418e-01 -8.18737596e-02 1.13883984e+00 3.65282744e-02
2.58815527e-01 3.44064713e-01 -2.46344432e-02 -2.98658282e-01
2.34416619e-01 -1.21451843e+00 1.06066906e+00 5.09197593e-01
3.75525415e-01 5.12086809e-01 7.86146879e-01 -6.21025980e-01
-1.42465782e+00 -7.83217728e-01 9.23417628e-01 -2.61470675e-01
3.90747488e-01 -3.50833088e-01 -7.65316963e-01 3.05545330e-01
2.37246491e-02 -4.95906085e-01 3.74560088e-01 -1.96882691e-02
-2.39061549e-01 2.43965700e-01 -9.20550346e-01 8.41570377e-01
1.06640434e+00 -1.76410064e-01 -2.44251519e-01 3.03891808e-01
5.70732772e-01 -6.36891127e-01 -7.18958259e-01 2.54847676e-01
5.35703123e-01 -1.01098895e+00 1.12121081e+00 -1.31820172e-01
6.59413695e-01 -1.02731466e-01 1.04306571e-01 -1.32971406e+00
-2.67685711e-01 -8.59906435e-01 -1.44697577e-01 1.07089710e+00
3.94795030e-01 -4.79747027e-01 1.23424065e+00 3.60205054e-01
-7.70572498e-02 -9.88969624e-01 -8.45934272e-01 -7.60241508e-01
7.19676092e-02 -5.09405017e-01 6.91336274e-01 4.10419405e-01
-2.97398359e-01 3.95219535e-01 -5.19869328e-01 -4.41627175e-01
5.17813683e-01 3.67328972e-01 9.33126032e-01 -1.22221529e+00
-5.32043397e-01 -7.15260327e-01 -3.37252766e-01 -1.57346165e+00
3.10736988e-02 -7.77352154e-01 8.47567394e-02 -1.67367578e+00
-7.90671781e-02 -6.20670438e-01 4.11669284e-01 2.56803781e-01
3.81586179e-02 2.69831121e-01 3.69922012e-01 7.68009648e-02
3.96101810e-02 6.68935895e-01 1.42780077e+00 1.96380749e-01
-2.83946067e-01 1.37611479e-01 -4.42523926e-01 4.90805060e-01
7.56130517e-01 -1.72866821e-01 -6.46191537e-01 -3.84077281e-01
2.42985204e-01 3.94960403e-01 1.49077028e-01 -9.09878492e-01
-1.17776461e-01 1.14734121e-01 3.72307748e-01 -7.89105833e-01
6.41232431e-01 -6.87764227e-01 3.05488259e-01 3.74918967e-01
-2.18233332e-01 1.67345062e-01 3.14333349e-01 5.42761326e-01
-8.96531269e-02 -5.52710593e-01 6.96429014e-01 -3.79672974e-01
-2.34230042e-01 3.43611568e-01 -3.71605873e-01 -1.85555562e-01
7.93988466e-01 -3.77994716e-01 2.57193834e-01 -7.19967067e-01
-4.28667158e-01 -9.11563337e-02 6.90289915e-01 2.68685013e-01
7.95923352e-01 -1.28652775e+00 -7.98107207e-01 2.71348059e-01
-2.51819819e-01 1.84697524e-01 2.20608562e-01 3.87376666e-01
-6.50819778e-01 6.26953423e-01 3.07310410e-02 -5.18537402e-01
-1.41253138e+00 4.31131065e-01 -1.34404808e-01 -8.07365537e-01
-6.69126689e-01 5.11332691e-01 2.97039270e-01 -1.65876314e-01
-4.55943048e-02 -8.29631314e-02 -1.57475933e-01 -6.43724427e-02
4.05736953e-01 3.08793753e-01 1.70412034e-01 -4.38978583e-01
1.23801611e-01 7.65455902e-01 2.17495002e-02 -4.15271640e-01
1.33712053e+00 1.97810099e-01 -1.38523370e-01 1.46606743e-01
8.99059713e-01 3.75027731e-02 -9.18215454e-01 9.95241255e-02
-2.25851938e-01 -4.60769296e-01 -1.23686112e-01 -2.68928260e-01
-6.48545146e-01 9.26547110e-01 8.65739733e-02 4.90901291e-01
9.67192113e-01 -9.26366821e-02 1.27772820e+00 1.90935194e-01
5.57715416e-01 -6.59513354e-01 3.82376648e-02 6.80298984e-01
9.79537070e-01 -6.05910420e-01 1.59326494e-01 -8.11093271e-01
-5.11214435e-01 1.15401649e+00 -1.04247153e-01 -1.05872743e-01
2.95675784e-01 4.29226637e-01 -2.69899070e-01 1.87806383e-01
-8.58870208e-01 -2.26588830e-01 3.85333627e-01 8.01207304e-01
7.06782818e-01 -4.43498045e-02 -3.52826327e-01 -1.24612384e-01
-4.01614010e-01 4.39339690e-02 4.24073875e-01 7.30397046e-01
-3.81227493e-01 -1.77445078e+00 -3.57068866e-01 5.34603298e-01
-4.40520376e-01 -2.09741056e-01 -1.25438720e-01 5.60256064e-01
-2.90852904e-01 6.99925065e-01 1.56613309e-02 1.68052986e-01
2.53863573e-01 9.80448723e-02 9.39664900e-01 -6.41400576e-01
-4.09480393e-01 3.90271544e-01 1.56632170e-01 -2.06892282e-01
-9.34395567e-02 -7.79541731e-01 -9.44161117e-01 -1.08062357e-01
-1.64980516e-01 2.10115328e-01 7.57191420e-01 6.64081752e-01
4.63918805e-01 3.58766735e-01 5.50857902e-01 -1.24325037e+00
-6.47426963e-01 -6.27756715e-01 -3.22378844e-01 2.52497941e-01
3.71468365e-02 -4.83089209e-01 1.67141989e-01 2.06451297e-01] | [8.9655179977417, -3.614137649536133] |
9703bf69-c0df-43d8-abea-c1bec40d6ba2 | recommendation-systems-in-libraries-an | 2303.11746 | null | https://arxiv.org/abs/2303.11746v1 | https://arxiv.org/pdf/2303.11746v1.pdf | Recommendation Systems in Libraries: an Application with Heterogeneous Data Sources | The Reading&Machine project exploits the support of digitalization to increase the attractiveness of libraries and improve the users' experience. The project implements an application that helps the users in their decision-making process, providing recommendation system (RecSys)-generated lists of books the users might be interested in, and showing them through an interactive Virtual Reality (VR)-based Graphical User Interface (GUI). In this paper, we focus on the design and testing of the recommendation system, employing data about all users' loans over the past 9 years from the network of libraries located in Turin, Italy. In addition, we use data collected by the Anobii online social community of readers, who share their feedback and additional information about books they read. Armed with this heterogeneous data, we build and evaluate Content Based (CB) and Collaborative Filtering (CF) approaches. Our results show that the CF outperforms the CB approach, improving by up to 47\% the relevant recommendations provided to a reader. However, the performance of the CB approach is heavily dependent on the number of books the reader has already read, and it can work even better than CF for users with a large history. Finally, our evaluations highlight that the performances of both approaches are significantly improved if the system integrates and leverages the information from the Anobii dataset, which allows us to include more user readings (for CF) and richer book metadata (for CB). | ['Marco Mellia', 'Luca Vassio', 'Greta Vallero', 'Alessandro Speciale'] | 2023-03-21 | null | null | null | null | ['collaborative-filtering'] | ['miscellaneous'] | [-5.18143415e-01 -3.90575714e-02 -2.76174009e-01 -2.54505455e-01
-7.26889610e-01 -7.39042044e-01 9.62766349e-01 5.85444450e-01
-4.87952262e-01 4.19948757e-01 8.01966548e-01 -4.06663150e-01
-5.05276918e-01 -1.13168430e+00 -4.50542033e-01 -1.90506831e-01
2.11479217e-01 4.94983912e-01 2.80424863e-01 -5.94820678e-01
6.69092894e-01 4.79156494e-01 -1.83738399e+00 2.87294865e-01
1.04982543e+00 6.85699105e-01 8.28748524e-01 4.32211787e-01
-2.63420552e-01 3.19092333e-01 -5.50596893e-01 -6.09890759e-01
1.56924024e-01 1.42781466e-01 -5.94109058e-01 -2.16599479e-01
5.59770763e-01 -4.59997594e-01 -4.49772865e-01 7.61478245e-01
4.63756353e-01 4.59339291e-01 4.09533381e-01 -4.64769512e-01
-9.47461188e-01 9.17868614e-01 2.19167620e-02 2.15616271e-01
8.41983140e-01 -3.93228024e-01 1.16495204e+00 -1.00785279e+00
1.08980060e+00 8.49591136e-01 2.73760587e-01 5.47710881e-02
-7.83186138e-01 -4.03716236e-01 1.87961906e-01 2.96782076e-01
-1.29397786e+00 -4.40907389e-01 2.07170889e-01 -4.94020879e-01
7.61311293e-01 8.17090869e-01 6.96638584e-01 7.34318614e-01
-9.89494994e-02 7.52463281e-01 8.83609593e-01 -6.55659616e-01
3.91975522e-01 6.84329212e-01 5.88134229e-01 4.13887277e-02
1.33254260e-01 -2.59821653e-01 -7.10630178e-01 6.12893328e-03
4.87604469e-01 3.44026387e-01 -6.03241503e-01 -2.85045300e-02
-1.08935678e+00 6.15369856e-01 3.44113320e-01 8.42546523e-01
-4.12460417e-01 -6.94073558e-01 -2.45460514e-02 3.31042409e-01
5.90302229e-01 7.67332196e-01 -3.92021090e-01 -2.52537340e-01
-9.80491221e-01 2.81975955e-01 1.24410999e+00 1.00018251e+00
3.62260967e-01 -8.37660134e-01 -2.04558477e-01 9.07417357e-01
2.74563283e-01 3.52384567e-01 6.22631490e-01 -7.50527501e-01
4.15537715e-01 7.46159196e-01 3.30375552e-01 -8.88598561e-01
-2.14747608e-01 -8.09621215e-01 -3.99547428e-01 3.27080786e-02
6.02651000e-01 1.39900714e-01 -1.94256693e-01 9.69027638e-01
-1.62962019e-01 -5.62740147e-01 -2.28913799e-01 8.55032206e-01
1.06856108e+00 3.31000984e-01 -2.01816335e-01 -2.59685159e-01
1.11893177e+00 -5.76950788e-01 -8.51707816e-01 2.78166682e-01
5.19951105e-01 -9.03128862e-01 1.24323046e+00 7.08813012e-01
-1.16544473e+00 -6.07414901e-01 -9.86949682e-01 9.99256223e-02
-7.62371004e-01 3.52898687e-01 6.21879518e-01 7.79256225e-01
-1.20902979e+00 8.81283164e-01 -3.98913920e-01 -8.37612987e-01
2.55837202e-01 2.32830137e-01 -2.05882028e-01 -2.14565545e-01
-8.45553100e-01 8.75802934e-01 -2.04802945e-01 -4.29172188e-01
-1.81410462e-01 -7.54076600e-01 -2.17464507e-01 4.46520358e-01
3.63492221e-01 -3.04770827e-01 9.42992508e-01 -5.30121684e-01
-1.34558678e+00 4.01068240e-01 -3.35524641e-02 -5.08954786e-02
8.15024376e-01 -6.27097130e-01 -3.49837363e-01 -2.00964212e-01
8.94578472e-02 -1.77205265e-01 -8.03364515e-02 -1.06631327e+00
-8.22199643e-01 -6.32625163e-01 2.36068785e-01 2.23160103e-01
-8.12427580e-01 2.14931622e-01 -1.00490332e+00 -4.86162245e-01
1.23351009e-03 -4.49254602e-01 -1.74173433e-02 -3.58418405e-01
-1.86443344e-01 -2.77688146e-01 3.08185041e-01 -1.11414313e+00
1.59901738e+00 -2.04415488e+00 7.43427053e-02 5.76057613e-01
1.00374103e-01 4.64525027e-03 2.10257679e-01 7.87772119e-01
6.27404511e-01 1.86948776e-01 6.24292195e-01 -3.60198826e-01
-4.75745201e-02 2.68878695e-03 -1.48548707e-01 2.90881991e-01
-8.30231130e-01 4.59729165e-01 -8.08432817e-01 -9.54566300e-02
2.40004972e-01 4.00162131e-01 -4.09712911e-01 1.01216897e-01
-6.43319413e-02 2.73481280e-01 -4.85288411e-01 5.70539355e-01
4.87791181e-01 -2.56756842e-01 3.18608999e-01 1.28169969e-01
-5.72112799e-01 4.32563752e-01 -1.34036624e+00 1.68022943e+00
-5.74068367e-01 6.76166177e-01 -2.05098644e-01 -2.54971504e-01
1.07102036e+00 4.38426398e-02 2.83914059e-01 -1.01668262e+00
-1.38029158e-01 1.90631658e-01 -4.60747004e-01 -4.70254868e-01
9.85299289e-01 7.13099480e-01 4.66041654e-01 7.19091058e-01
-1.92433581e-01 4.24316466e-01 3.61683428e-01 7.59001076e-01
1.26078343e+00 1.40252143e-01 -1.28730446e-01 -4.14947867e-01
3.98596048e-01 -2.36594066e-01 -5.02447188e-02 1.03930783e+00
6.37346208e-01 3.73496771e-01 5.38887642e-02 1.47679165e-01
-9.92571056e-01 -8.22132707e-01 -4.12480295e-01 1.40684223e+00
7.39603043e-02 -7.10499883e-01 -5.61444283e-01 -4.69254643e-01
2.06043035e-01 1.07518637e+00 -3.76874685e-01 3.94400835e-01
-1.92135572e-01 -4.34997678e-01 -4.10595864e-01 3.81679386e-01
4.07492101e-01 -6.81810737e-01 -2.33839840e-01 1.92584932e-01
-1.62280992e-01 -5.83367765e-01 -2.55147845e-01 -1.72314808e-01
-7.51660526e-01 -7.89983928e-01 -8.44441593e-01 -1.47026911e-01
3.48647177e-01 3.46501172e-01 1.22315609e+00 3.64637345e-01
7.45737404e-02 7.20230341e-01 -8.55019033e-01 -3.23829979e-01
-1.72855332e-01 5.68858743e-01 -7.46010914e-02 -2.69089550e-01
1.97426945e-01 -4.52018172e-01 -5.65769196e-01 5.06530762e-01
-6.00700736e-01 -3.50797698e-02 2.11717144e-01 3.49696755e-01
2.14344233e-01 -7.65264481e-02 8.09232056e-01 -1.24982285e+00
8.07555437e-01 -7.88693428e-01 -6.22775912e-01 3.78262490e-01
-1.33606887e+00 -3.16091031e-01 3.45420003e-01 -1.59652218e-01
-1.26761866e+00 -3.10704291e-01 -1.27006248e-01 4.59514380e-01
3.09645701e-02 7.63452113e-01 -3.47991973e-01 2.33656801e-02
6.22916996e-01 -1.75095275e-01 -2.91762620e-01 -1.26795459e+00
5.03379405e-01 1.24462330e+00 6.22927845e-01 -2.51172483e-01
3.50266784e-01 1.53530538e-01 -4.05122280e-01 -5.67040920e-01
-5.47705293e-01 -7.96379328e-01 -7.36268163e-01 -6.45960569e-01
3.88922155e-01 -8.24608803e-01 -7.18899131e-01 9.27227736e-02
-4.68155324e-01 -1.88684464e-01 -1.76634982e-01 5.38681626e-01
-1.46743715e-01 1.91210926e-01 -5.24496794e-01 -8.87350202e-01
-2.94800043e-01 -7.22230494e-01 3.96072447e-01 5.65638781e-01
-2.18559816e-01 -8.24091494e-01 -1.11636102e-01 6.01051331e-01
6.27268493e-01 -1.47219881e-01 7.97771335e-01 -1.16975653e+00
-7.11115122e-01 -5.98795593e-01 -2.38054767e-01 1.28919343e-02
-1.80786654e-01 2.89304126e-02 -8.92664492e-01 -3.80871385e-01
-5.19181013e-01 4.54486519e-01 5.05963266e-01 1.61243156e-01
9.30086851e-01 -7.44725689e-02 -3.89710218e-01 1.51592553e-01
1.58755672e+00 1.21358439e-01 8.07725549e-01 9.62586284e-01
4.36419129e-01 6.07195556e-01 7.53825247e-01 8.27052116e-01
6.31691098e-01 8.38460386e-01 3.30336154e-01 1.60484359e-01
-7.43406788e-02 -1.20607980e-01 2.87485838e-01 9.25496757e-01
-6.03284836e-01 -4.69623178e-01 -7.07041740e-01 3.86308491e-01
-1.92969632e+00 -7.50703692e-01 -4.38105375e-01 2.65523314e+00
5.96327960e-01 6.40767887e-02 2.73133218e-01 1.89412162e-01
3.89109880e-01 -4.90868807e-01 -1.79979607e-01 -9.84759331e-02
-1.26942664e-01 4.03496295e-01 7.20539331e-01 4.26304162e-01
-4.15202260e-01 4.74242151e-01 5.88484907e+00 6.02083087e-01
-6.84273541e-01 -9.36021563e-03 3.96641195e-01 -3.36148590e-01
-6.18764400e-01 -7.37836063e-02 -9.01261389e-01 5.63877821e-01
1.13730741e+00 -4.67689812e-01 6.16547883e-01 7.67310560e-01
3.25308591e-01 -7.26213276e-01 -1.05528522e+00 5.55663824e-01
7.36046210e-02 -1.76422894e+00 -3.53066832e-01 4.19322401e-01
5.44274330e-01 1.53543890e-01 -3.31809111e-02 3.08968395e-01
3.35961759e-01 -7.40393102e-01 7.89193034e-01 1.29681230e+00
4.93509471e-01 -7.66652822e-01 8.69510770e-01 6.22424722e-01
-7.64617085e-01 -1.89238876e-01 -3.26639205e-01 -2.20039099e-01
-1.57548457e-01 5.88773012e-01 -5.83695769e-01 9.20401812e-01
1.03636622e+00 8.21052432e-01 -1.06998372e+00 1.38768280e+00
9.27134454e-02 6.44819319e-01 -1.20055467e-01 -3.00195813e-01
-3.81307483e-01 -5.68588555e-01 2.88315296e-01 1.17516923e+00
6.49303794e-01 2.82766521e-01 -2.23939836e-01 5.80267012e-01
-1.56854242e-01 6.15850508e-01 -2.55243331e-01 2.53680736e-01
7.18784213e-01 1.30112612e+00 -5.43339074e-01 -3.94931942e-01
-6.91012442e-01 6.83136284e-01 2.11935297e-01 1.81408688e-01
-2.00915411e-02 -5.63886702e-01 1.87078968e-01 7.00917363e-01
1.29724890e-01 -5.38438782e-02 -4.62224245e-01 -1.07501626e+00
-9.95403081e-02 -7.48167098e-01 3.21032554e-01 -8.92065048e-01
-1.08548856e+00 3.71994376e-01 -2.50806332e-01 -8.36601615e-01
-1.63017381e-02 -3.29961956e-01 -5.66444159e-01 1.19730771e+00
-9.70569730e-01 -8.25925708e-01 -5.59724927e-01 3.11014175e-01
4.47771907e-01 -3.39453161e-01 6.91343606e-01 6.99173689e-01
-5.34305215e-01 5.22009552e-01 9.04056549e-01 -2.46520177e-01
9.56916153e-01 -1.47317731e+00 8.68869498e-02 5.75070918e-01
1.25872314e-01 8.09017181e-01 6.13188863e-01 -7.19232142e-01
-1.58853388e+00 -5.65278709e-01 1.23806250e+00 -8.39082181e-01
4.76357996e-01 -9.22019258e-02 -9.40210044e-01 5.81397533e-01
2.85817087e-01 -8.03348720e-01 9.05672729e-01 7.14643300e-01
4.75136749e-02 2.63653696e-02 -1.12056303e+00 4.68327552e-01
1.00435960e+00 -3.40932131e-01 -4.14243609e-01 4.70833004e-01
1.89934045e-01 -2.42307901e-01 -1.51134658e+00 -2.93044478e-01
6.73453271e-01 -1.14616966e+00 9.40418303e-01 -9.88354832e-02
4.06219840e-01 1.94602720e-02 -3.65155071e-01 -1.14469898e+00
-4.84428972e-01 -3.40275139e-01 1.19358353e-01 1.74843228e+00
4.99518007e-01 -5.17001152e-01 5.93641460e-01 1.08745682e+00
-1.53412268e-01 -6.05191469e-01 -5.38973689e-01 -5.05140483e-01
1.85442660e-02 -4.58686590e-01 9.71169233e-01 8.38461459e-01
4.69591826e-01 1.04205497e-01 8.33493173e-02 -9.31189880e-02
2.23057419e-01 1.89918160e-01 9.30086672e-01 -1.62143862e+00
-4.25943792e-01 -5.40096998e-01 7.02349022e-02 -9.59354997e-01
-4.10707891e-01 -1.06082201e+00 -5.56054592e-01 -1.88116395e+00
4.87985998e-01 -5.66629469e-01 -1.69072270e-01 1.77285984e-01
-3.54439951e-02 -5.71492203e-02 5.03619134e-01 7.72954047e-01
-7.07790732e-01 1.20975070e-01 9.70466614e-01 2.44752422e-01
-4.05189127e-01 3.56480420e-01 -8.78525019e-01 4.77122337e-01
3.73071969e-01 9.10056084e-02 1.24899074e-01 -2.00405657e-01
6.91646159e-01 1.57968670e-01 -1.84489489e-02 -8.53964329e-01
3.56831521e-01 2.53433019e-01 6.73066139e-01 -8.31167638e-01
-3.15416902e-02 -7.77784646e-01 4.34124231e-01 9.68416333e-02
-4.60051805e-01 1.56557150e-02 -1.15225665e-01 4.91466284e-01
1.31271601e-01 -5.42989790e-01 9.93641615e-02 -3.83118028e-03
-3.52219522e-01 -8.91763121e-02 -6.12044930e-01 -5.91940522e-01
4.79913890e-01 3.73556204e-02 -4.38572824e-01 -7.36486614e-01
-9.66575742e-01 5.32464534e-02 7.08117545e-01 3.84191453e-01
1.77355424e-01 -1.09949887e+00 -5.66705704e-01 6.07657656e-02
2.75002450e-01 -4.80836689e-01 -1.51568323e-01 5.92522919e-01
-2.11461216e-01 2.93948382e-01 -1.52064607e-01 -1.38604239e-01
-1.36902535e+00 2.04149276e-01 -2.26699144e-01 -1.64638802e-01
-6.83609605e-01 6.74863935e-01 -4.30024475e-01 -8.65723938e-02
5.42602420e-01 -3.08497772e-02 -1.00142705e+00 4.68290448e-01
6.57236934e-01 9.34308231e-01 4.23080951e-01 -3.17961842e-01
4.10481170e-02 1.36672258e-01 -3.90537739e-01 -3.78371447e-01
1.69364643e+00 -5.61821461e-01 1.70956720e-02 4.02115524e-01
4.93803352e-01 6.13362432e-01 -7.14397192e-01 -5.23951232e-01
3.08214486e-01 -9.29069102e-01 3.17087442e-01 -1.31906903e+00
-9.05241311e-01 3.38475198e-01 6.32030427e-01 4.26395833e-01
1.01323462e+00 1.05101615e-01 4.26623434e-01 4.15191531e-01
6.12831175e-01 -1.11708295e+00 -3.64799052e-01 2.25951567e-01
8.34830463e-01 -1.05624044e+00 3.46966594e-01 -2.06981510e-01
-5.83436310e-01 1.26180398e+00 1.49960667e-01 2.89944232e-01
6.83841646e-01 -9.30128545e-02 -7.90346265e-02 -1.22769944e-01
-6.07813597e-01 -1.98231235e-01 3.05341274e-01 3.94916385e-01
7.43005335e-01 2.62030661e-01 -7.31518626e-01 1.21791112e+00
-3.63751531e-01 1.10826857e-01 8.14378679e-01 7.24811614e-01
-6.15761220e-01 -1.36020982e+00 -2.99112082e-01 1.00431538e+00
-5.00596821e-01 -7.35550150e-02 -3.71547699e-01 6.38952255e-01
-1.22048058e-01 1.13208294e+00 2.34321803e-02 -3.89793277e-01
5.75510919e-01 -2.16866046e-01 3.12941074e-01 -6.72069609e-01
-1.14228249e+00 7.99097344e-02 2.96269059e-01 -4.10391033e-01
9.72069725e-02 -1.10096645e+00 -9.51151729e-01 -6.85500085e-01
-7.47287691e-01 2.44457260e-01 1.09316754e+00 7.63728678e-01
4.63327438e-01 5.01627803e-01 6.77769721e-01 -7.13168740e-01
-3.50630492e-01 -1.03520274e+00 -7.85940111e-01 3.34724993e-01
-2.76048064e-01 -4.24082518e-01 -1.92095101e-01 -3.92092802e-02] | [10.063895225524902, 5.790229320526123] |
d69dbeb7-d8bb-4e5b-b5f9-8b5dddd4d962 | distributional-reinforcement-learning-for-1 | 1912.08517 | null | https://arxiv.org/abs/1912.08517v1 | https://arxiv.org/pdf/1912.08517v1.pdf | Distributional Reinforcement Learning for Energy-Based Sequential Models | Global Autoregressive Models (GAMs) are a recent proposal [Parshakova et al., CoNLL 2019] for exploiting global properties of sequences for data-efficient learning of seq2seq models. In the first phase of training, an Energy-Based model (EBM) over sequences is derived. This EBM has high representational power, but is unnormalized and cannot be directly exploited for sampling. To address this issue [Parshakova et al., CoNLL 2019] proposes a distillation technique, which can only be applied under limited conditions. By relating this problem to Policy Gradient techniques in RL, but in a \emph{distributional} rather than \emph{optimization} perspective, we propose a general approach applicable to any sequential EBM. Its effectiveness is illustrated on GAM-based experiments. | ['Jean-Marc Andreoli', 'Marc Dymetman', 'Tetiana Parshakova'] | 2019-12-18 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [ 4.26195830e-01 1.87897176e-01 -3.62184078e-01 -1.97671294e-01
-1.00781989e+00 -6.27704620e-01 7.45448709e-01 -3.38000990e-03
-7.64076293e-01 1.01401520e+00 5.46278596e-01 -3.72984469e-01
-3.34679335e-02 -6.68098927e-01 -5.68373561e-01 -7.29793012e-01
9.68525559e-02 3.33619535e-01 7.36424774e-02 -2.81914234e-01
3.15498501e-01 4.83910859e-01 -1.40354931e+00 2.79308975e-01
6.13577247e-01 6.77772105e-01 6.21765852e-01 1.01734805e+00
-3.81636590e-01 1.00386739e+00 -5.44283926e-01 -9.47655365e-02
1.11687556e-01 -9.67112780e-01 -8.88760746e-01 -5.21487057e-01
-6.37478083e-02 -1.55292109e-01 3.59681770e-02 1.10464656e+00
8.37749124e-01 6.58273816e-01 7.78600037e-01 -7.27773309e-01
-2.86602139e-01 8.06713641e-01 -2.30905548e-01 3.21336657e-01
2.15518311e-01 9.67247635e-02 1.14972055e+00 -8.17655623e-01
6.16401911e-01 1.17900562e+00 4.90795970e-01 9.97012854e-01
-1.06918931e+00 -3.95793706e-01 1.47409841e-01 3.75447720e-01
-1.05487835e+00 -4.48708773e-01 6.35104418e-01 -9.58658159e-02
1.25196290e+00 6.23666167e-01 7.36036837e-01 1.28400850e+00
-1.24859728e-01 1.37147868e+00 1.30342591e+00 -7.94981241e-01
6.13799512e-01 -1.05535783e-01 -8.27813521e-03 3.47816676e-01
-3.28044206e-01 1.49321005e-01 -7.51468241e-01 -2.40014166e-01
6.24631166e-01 -5.44816375e-01 -1.09560415e-01 -2.04032794e-01
-8.76806676e-01 1.06548667e+00 4.64769416e-02 5.41695595e-01
-6.07052982e-01 4.19838250e-01 6.45630956e-01 2.11620316e-01
5.45696974e-01 5.70691943e-01 -4.85513210e-01 -1.05176544e+00
-1.14115882e+00 5.05358219e-01 6.73474073e-01 8.61722350e-01
3.75976652e-01 4.32993203e-01 -3.77101958e-01 1.09899688e+00
3.41559261e-01 2.67224282e-01 8.82975340e-01 -1.10875821e+00
1.41303867e-01 -1.19573906e-01 1.37178674e-01 -2.85969555e-01
-3.69013488e-01 -3.33247036e-01 -6.64691687e-01 3.36708091e-02
4.17515457e-01 -2.02600688e-01 -6.06494904e-01 1.94831014e+00
1.41484395e-01 2.13096127e-01 9.00831819e-02 6.01098359e-01
3.60462219e-01 8.56761217e-01 3.79104793e-01 -5.60830116e-01
9.82143760e-01 -9.82348502e-01 -7.21035719e-01 -4.20765840e-02
7.34552741e-01 -5.02634227e-01 1.27210510e+00 5.95365465e-01
-1.27469134e+00 -3.92579049e-01 -7.67083943e-01 1.06520159e-03
-2.68698812e-01 -1.11087419e-01 3.92466307e-01 9.27424848e-01
-1.31499386e+00 7.68965006e-01 -7.40418375e-01 -4.54680413e-01
8.02333876e-02 2.36244872e-01 2.99478620e-01 2.25980565e-01
-1.22364008e+00 1.02206159e+00 6.95065141e-01 -6.90081492e-02
-6.97711170e-01 -4.96465504e-01 -7.05799639e-01 2.74681803e-02
2.14374825e-01 -4.63643849e-01 1.58461416e+00 -1.26023710e+00
-2.31179714e+00 5.83100617e-01 -2.45598972e-01 -7.38206208e-01
5.03626525e-01 -3.48554611e-01 -2.33201280e-01 6.43744022e-02
-5.63578308e-01 5.07824242e-01 7.86604702e-01 -9.79562521e-01
-3.89185309e-01 -9.39138681e-02 -2.15108022e-02 3.09956998e-01
-3.31627309e-01 3.95793974e-01 -2.27790698e-02 -8.75782192e-01
-6.20961130e-01 -7.87997842e-01 -3.81180257e-01 -6.33877754e-01
-1.29464775e-01 -3.50543737e-01 2.54659295e-01 -8.04919660e-01
1.44114137e+00 -1.81806862e+00 2.72471100e-01 2.47424796e-01
-5.43038905e-01 7.47624457e-01 -4.64266241e-01 6.19636834e-01
-2.97541544e-02 2.33373180e-01 -4.12265629e-01 -3.44615966e-01
1.76747993e-01 2.90444970e-01 -3.19357723e-01 7.02878460e-02
1.06601112e-01 1.14095807e+00 -1.18772590e+00 -3.89552236e-01
2.63371527e-01 5.00109732e-01 -6.85834110e-01 2.74119526e-01
-5.07365286e-01 5.76185465e-01 -3.92770410e-01 2.26933867e-01
2.68840492e-01 1.93328798e-01 3.05690378e-01 2.03725711e-01
-2.64682919e-01 3.94941270e-01 -9.22693968e-01 1.77919841e+00
-8.21566224e-01 5.04651070e-01 -1.88718766e-01 -1.17687738e+00
8.46844673e-01 3.20928752e-01 6.33251369e-01 -6.76014662e-01
1.28519893e-01 3.25412184e-01 -7.03684464e-02 -3.50671858e-01
8.42756152e-01 -2.30567515e-01 1.52121847e-02 4.74545658e-01
1.89566523e-01 1.88916326e-02 2.54551083e-01 -1.04500778e-01
1.02613711e+00 8.12461257e-01 7.25445688e-01 -5.10467350e-01
5.06697297e-01 -1.24325715e-01 2.65479118e-01 1.00390434e+00
-2.63581076e-03 6.67225540e-01 1.98255032e-01 -8.73062536e-02
-1.24104190e+00 -7.11783767e-01 1.78845301e-01 1.51466179e+00
-4.54840899e-01 -6.13287807e-01 -9.65993583e-01 -5.64074755e-01
-5.36099732e-01 1.33798552e+00 -4.31568295e-01 -2.39810899e-01
-9.07834411e-01 -9.10891354e-01 6.59316421e-01 4.49562460e-01
2.08097786e-01 -1.47410333e+00 -1.05482554e+00 5.59641719e-01
-1.98484316e-01 -8.39694262e-01 -4.89176512e-01 3.41238678e-01
-1.00147903e+00 -4.56592947e-01 -9.43922162e-01 -2.03784987e-01
6.51806146e-02 -3.88116181e-01 1.20206273e+00 -2.75588304e-01
-7.79996160e-03 4.75957334e-01 -6.05888128e-01 -6.68130815e-01
-7.96622634e-01 1.78650320e-01 -1.79490924e-01 -9.96712148e-02
1.98013484e-01 -6.64910436e-01 -4.96500254e-01 -2.56241739e-01
-8.20489526e-01 -1.33605555e-01 5.90840340e-01 9.34278786e-01
6.37188971e-01 -4.37971145e-01 8.01880062e-01 -9.42386389e-01
8.48115981e-01 -3.50606471e-01 -4.23576742e-01 1.45704344e-01
-6.05198085e-01 2.42967442e-01 7.59537876e-01 -4.39252079e-01
-1.13438678e+00 -5.47572896e-02 -8.24526072e-01 -4.15890783e-01
-2.17479188e-02 4.85263497e-01 -2.02970847e-01 2.44803429e-01
5.40023804e-01 6.75573945e-01 -1.79390937e-01 -6.61773264e-01
5.90188503e-01 5.37899673e-01 2.01003268e-01 -5.99420667e-01
1.72593385e-01 -4.36511375e-02 -2.54892670e-02 -1.00256717e+00
-5.40194392e-01 -4.11628634e-01 -4.63795096e-01 -2.77765870e-01
8.05788815e-01 -4.80184734e-01 -6.39572859e-01 2.03403980e-01
-1.00329161e+00 -8.68265629e-01 -8.83952796e-01 5.54024875e-01
-1.27306724e+00 3.67249310e-01 -4.79827434e-01 -1.30797541e+00
-3.52964222e-01 -6.84717476e-01 9.25656319e-01 2.47681942e-02
-5.06904662e-01 -1.08971620e+00 3.60628307e-01 -6.96838796e-02
5.86503327e-01 -8.89856294e-02 9.24919426e-01 -1.01729131e+00
-1.90798894e-01 1.29221737e-01 2.88851738e-01 6.24291956e-01
-2.48985827e-01 -1.64791599e-01 -1.05374062e+00 -2.45723963e-01
2.25741476e-01 -2.86823928e-01 7.20346332e-01 4.61117178e-01
1.32255888e+00 -5.00017822e-01 1.31048664e-01 3.02440763e-01
1.42120326e+00 3.37443769e-01 7.22276568e-01 2.45407462e-01
4.56394762e-01 3.75530034e-01 6.82923853e-01 4.75647420e-01
8.07418302e-02 8.06088090e-01 1.78869367e-01 4.08236235e-01
-1.88055605e-01 -4.25677925e-01 8.69657040e-01 1.21274602e+00
-3.33977312e-01 -5.91483831e-01 -6.06128037e-01 4.94517595e-01
-1.97421777e+00 -1.11481905e+00 1.07589051e-01 2.24596071e+00
9.45999920e-01 -1.00953132e-01 4.98330772e-01 -8.14418048e-02
5.05148292e-01 4.29507107e-01 -4.27425176e-01 -1.04829001e+00
-1.63725644e-01 5.45775771e-01 4.83340442e-01 5.97310901e-01
-7.52116501e-01 9.08856750e-01 6.81404829e+00 1.42133927e+00
-7.92990267e-01 3.86951119e-01 3.74884307e-01 -2.26564154e-01
-5.57596028e-01 -1.11110285e-01 -7.60230899e-01 6.86918437e-01
1.61209619e+00 -1.61756143e-01 6.78397596e-01 6.83111310e-01
4.17575240e-01 -1.71131492e-01 -8.99064422e-01 9.11480308e-01
-1.59713775e-02 -1.12186313e+00 -1.06992468e-01 4.12863195e-02
6.39312506e-01 1.44171223e-01 3.64286415e-02 6.18673682e-01
6.53817058e-01 -1.14878440e+00 6.82136238e-01 6.33604884e-01
5.38729548e-01 -9.34374809e-01 4.80207264e-01 8.07199836e-01
-9.48982120e-01 -3.09229176e-02 -4.39228982e-01 1.01566203e-01
3.66579592e-01 4.04164135e-01 -8.09406996e-01 7.93675840e-01
2.64472574e-01 4.33059782e-01 -3.13390349e-03 8.12813759e-01
-2.06134245e-01 1.09447122e+00 -3.13322157e-01 -5.00888407e-01
4.84172970e-01 -3.16775352e-01 7.37706304e-01 1.70369720e+00
5.93911171e-01 1.43860923e-02 -1.09591790e-01 7.08970428e-01
5.84339686e-02 5.57874382e-01 -5.03253460e-01 -1.71900168e-01
3.85708332e-01 1.04764271e+00 -5.08562565e-01 -3.37774426e-01
-1.44882649e-01 1.04887867e+00 3.10233235e-01 2.87448257e-01
-6.82480097e-01 -2.15880517e-02 2.48636842e-01 -3.35315973e-01
4.28128242e-01 -3.32347840e-01 4.32353057e-02 -9.88775074e-01
-3.82522762e-01 -1.03939307e+00 4.37936604e-01 -6.02092445e-01
-9.50330973e-01 4.16895628e-01 3.33653182e-01 -9.82718408e-01
-1.01762271e+00 -3.37351710e-01 -3.81839961e-01 1.01836550e+00
-1.52556729e+00 -1.00243485e+00 3.37427646e-01 4.29074168e-01
1.01091838e+00 -1.33164093e-01 1.01942110e+00 5.10137752e-02
-2.14651138e-01 4.57771778e-01 5.66915512e-01 -3.29481870e-01
1.90329686e-01 -1.46816921e+00 4.51973885e-01 6.74247622e-01
3.38921756e-01 3.30132872e-01 8.81296277e-01 -4.02776182e-01
-1.42764783e+00 -1.11571395e+00 9.65499878e-01 -3.16152781e-01
4.81470793e-01 -3.48700643e-01 -8.32328320e-01 5.17723262e-01
5.22507071e-01 -4.76296127e-01 6.78115904e-01 -1.41789556e-01
2.14359835e-02 3.50004345e-01 -9.69862103e-01 6.48019195e-01
1.14876699e+00 -5.12246311e-01 -3.89945418e-01 2.77869552e-01
5.88458300e-01 -3.30310255e-01 -8.81242990e-01 2.69806564e-01
4.04448301e-01 -8.98188710e-01 7.49781549e-01 -8.12801063e-01
2.35476568e-02 1.66358635e-01 -2.92109430e-01 -1.60177624e+00
-1.78628843e-02 -1.12891889e+00 -5.12553215e-01 1.24164462e+00
2.36055657e-01 -4.65938240e-01 6.34602129e-01 1.40616268e-01
-1.60755008e-01 -7.45684683e-01 -1.08165824e+00 -1.21197915e+00
4.95048702e-01 -6.78908348e-01 4.72441971e-01 5.46767414e-01
7.95454159e-02 1.79521903e-01 -6.83828056e-01 -5.37304342e-01
2.58106560e-01 -1.55649960e-01 4.78744894e-01 -7.25164950e-01
-6.39670551e-01 -6.99906468e-01 -2.88330037e-02 -1.23077488e+00
3.65415215e-01 -9.55770373e-01 3.67042243e-01 -1.30300534e+00
-7.29903430e-02 -3.25825483e-01 -6.34112418e-01 1.11458898e-01
4.10200469e-02 -7.89948404e-02 4.74255264e-01 1.92754343e-02
-4.62917745e-01 7.16460824e-01 9.61603165e-01 2.72811145e-01
-1.66746691e-01 1.58299625e-01 -3.30456972e-01 6.28241181e-01
1.02242100e+00 -3.97705346e-01 -5.93718529e-01 -9.77453291e-02
3.14329773e-01 4.46964018e-02 1.05267510e-01 -7.52435327e-01
1.59271315e-01 -2.04978406e-01 -6.58713579e-02 -3.73935342e-01
5.50092936e-01 -4.53107864e-01 1.81160197e-01 5.19597471e-01
-8.22623074e-01 1.83899656e-01 1.99061453e-01 5.94427109e-01
-8.62852931e-02 -8.84730756e-01 8.07526529e-01 -3.99132252e-01
-7.79512286e-01 1.82800386e-02 -6.80256009e-01 2.34462366e-01
7.55622447e-01 -4.13613245e-02 2.26141885e-01 -6.04470253e-01
-6.99252486e-01 -1.18035167e-01 1.70046762e-01 1.92899153e-01
4.58024591e-01 -1.21236742e+00 -7.98744261e-01 -1.09751217e-01
-2.99488902e-01 -2.69363433e-01 2.14512557e-01 9.09462392e-01
-2.13690594e-01 4.88102078e-01 -4.74428199e-02 -2.13849604e-01
-1.10854053e+00 6.30590320e-01 2.39776969e-01 -6.40863001e-01
-4.61983830e-01 7.30315268e-01 -8.06934759e-02 -4.56763476e-01
8.40904415e-02 -1.08702898e-01 -3.58301014e-01 1.34578198e-01
4.73817557e-01 4.90477800e-01 2.43674330e-02 -5.41854858e-01
-1.09296985e-01 3.51341426e-01 1.97432801e-01 -6.24049485e-01
1.40945065e+00 -1.61828041e-01 1.85265169e-01 8.94116700e-01
1.20353103e+00 1.54924672e-02 -1.09206533e+00 -2.13744834e-01
3.18513483e-01 -3.04970622e-01 -1.16678469e-01 -8.67397368e-01
-6.10884130e-01 1.07770133e+00 4.27026212e-01 3.66120756e-01
1.14840412e+00 -3.37415010e-01 7.25563407e-01 2.81755716e-01
3.12258631e-01 -1.37225080e+00 -1.62940919e-01 7.33138442e-01
7.83654928e-01 -7.24288166e-01 -3.26754153e-01 6.85393959e-02
-7.69813120e-01 1.30629063e+00 8.16636980e-02 -1.33956857e-02
4.02419329e-01 2.31228247e-01 -3.64466399e-01 1.51237622e-01
-9.05572176e-01 -3.81442100e-01 7.77671039e-02 7.48203993e-01
4.86191273e-01 8.32510740e-02 -6.81159854e-01 5.03160238e-01
-1.76655889e-01 1.39183670e-01 4.00375992e-01 1.05981910e+00
-5.70105374e-01 -1.36669052e+00 -7.55032748e-02 2.78595805e-01
-5.43468893e-01 -4.08029646e-01 -2.94589639e-01 6.91298187e-01
-2.87052184e-01 7.51210868e-01 -1.27188042e-01 -1.76255450e-01
1.68050125e-01 5.31850219e-01 6.23609126e-01 -4.95856762e-01
-8.75909805e-01 2.67243803e-01 3.97410303e-01 -5.80759168e-01
-7.12571323e-01 -9.25429046e-01 -1.05801380e+00 1.26595078e-02
-9.42745209e-02 3.73444438e-01 8.22098911e-01 9.25567985e-01
3.93394351e-01 5.84533215e-01 5.26927531e-01 -6.60641849e-01
-9.12164688e-01 -1.03671837e+00 -5.45186520e-01 7.93122202e-02
4.74365205e-02 -1.85063228e-01 4.00399901e-02 -8.42476040e-02] | [11.895626068115234, 9.174835205078125] |
579cd518-42c6-4539-b814-e33e20c32413 | a-case-study-for-compliance-as-code-with | 2302.01842 | null | https://arxiv.org/abs/2302.01842v1 | https://arxiv.org/pdf/2302.01842v1.pdf | A Case Study for Compliance as Code with Graphs and Language Models: Public release of the Regulatory Knowledge Graph | The paper presents a study on using language models to automate the construction of executable Knowledge Graph (KG) for compliance. The paper focuses on Abu Dhabi Global Market regulations and taxonomy, involves manual tagging a portion of the regulations, training BERT-based models, which are then applied to the rest of the corpus. Coreference resolution and syntax analysis were used to parse the relationships between the tagged entities and to form KG stored in a Neo4j database. The paper states that the use of machine learning models released by regulators to automate the interpretation of rules is a vital step towards compliance automation, demonstrates the concept querying with Cypher, and states that the produced sub-graphs combined with Graph Neural Networks (GNN) will achieve expandability in judgment automation systems. The graph is open sourced on GitHub to provide structured data for future advancements in the field. | ['Vladimir Ershov'] | 2023-02-03 | null | null | null | null | ['coreference-resolution'] | ['natural-language-processing'] | [ 4.80402410e-02 8.82528484e-01 -2.73220241e-01 -3.93197149e-01
-3.61145049e-01 -6.41318977e-01 3.86758953e-01 4.96744990e-01
3.19170170e-02 5.98223805e-01 6.28963411e-01 -8.57297003e-01
-8.02370369e-01 -8.57011080e-01 -1.12044953e-01 2.20162526e-01
-1.52358040e-01 1.07520282e+00 4.32115681e-02 -5.75391173e-01
1.37581080e-01 3.98750991e-01 -8.11989427e-01 5.54628670e-01
7.97058880e-01 5.33944130e-01 -6.63763359e-02 2.94782013e-01
-4.41795856e-01 1.24653733e+00 -3.62140000e-01 -7.50860810e-01
2.85764635e-01 -1.12362497e-01 -1.30276847e+00 -2.61890680e-01
1.28215283e-01 6.81566596e-02 1.45842835e-01 1.19079745e+00
1.10552490e-01 -2.48829141e-01 4.28061396e-01 -8.86593938e-01
-7.37155557e-01 1.28191662e+00 1.30950481e-01 -1.21454902e-01
6.94627404e-01 -2.33559728e-01 1.16821527e+00 -6.95742592e-02
9.40737665e-01 1.18550253e+00 7.26332247e-01 2.38482162e-01
-7.71647274e-01 -8.21093678e-01 -1.42633438e-01 3.02558005e-01
-1.15454936e+00 -1.79532140e-01 6.16639853e-01 -7.27285862e-01
1.53147066e+00 -6.65187836e-05 7.64463723e-01 4.62529570e-01
8.01184252e-02 -1.46815889e-02 9.20707285e-01 -8.85138929e-01
-5.79784450e-04 2.63424963e-01 5.03942072e-01 8.35892439e-01
7.99816847e-01 1.98183641e-01 -1.52717233e-01 -2.26779103e-01
5.58157861e-01 -7.21885502e-01 1.15474902e-01 -2.97144741e-01
-2.69802392e-01 9.39972103e-01 1.52792424e-01 7.35191703e-01
-4.28822815e-01 -2.36755610e-01 4.35273588e-01 2.88502842e-01
1.70959113e-03 6.75681651e-01 -4.85226363e-01 4.45084907e-02
-6.93797886e-01 3.72199923e-01 1.15767896e+00 1.06261873e+00
5.24603605e-01 -8.06871355e-02 1.61563650e-01 3.95936340e-01
7.30131507e-01 1.43000215e-01 2.15034187e-01 -1.05106950e+00
6.73572838e-01 1.49189472e+00 3.39204371e-01 -1.01477003e+00
-7.73336530e-01 -1.99406788e-01 1.29850119e-01 -8.62396434e-02
4.69796807e-01 -3.51772964e-01 -8.18575203e-01 1.11474371e+00
1.57254875e-01 -3.72709721e-01 3.42672229e-01 2.64380157e-01
9.56869483e-01 3.70655954e-02 7.13790834e-01 -1.55124992e-01
1.44638228e+00 -3.96820247e-01 -1.13249171e+00 1.09703697e-01
1.21715379e+00 -3.60274196e-01 2.47477591e-01 3.88049513e-01
-1.01357639e+00 -4.57955122e-01 -1.04286337e+00 1.02178305e-01
-9.36931133e-01 -1.01270430e-01 1.05324662e+00 8.74121189e-01
-7.99117506e-01 4.68120307e-01 -6.24714077e-01 -7.08003879e-01
3.27337831e-01 8.11823487e-01 -4.30102199e-01 1.48418292e-01
-1.58060408e+00 1.36708939e+00 1.10960543e+00 5.61549142e-02
5.04559930e-03 -3.44653249e-01 -1.17990804e+00 -3.09053343e-02
4.27550405e-01 -3.86612207e-01 1.07166600e+00 -8.37253213e-01
-9.32396173e-01 8.82577896e-01 4.91968900e-01 -7.84393489e-01
8.43813941e-02 -7.12845773e-02 -1.26606190e+00 -6.44724146e-02
2.36493275e-01 3.01704288e-01 -2.19562858e-01 -8.68095815e-01
-9.12911415e-01 -4.58830297e-01 1.41469285e-01 -9.21603665e-02
5.66456653e-02 4.82135206e-01 -1.97033063e-01 1.20634353e-02
-1.21362410e-01 -5.91566980e-01 6.96768926e-04 -1.16269028e+00
-6.91673830e-02 -5.75151742e-01 6.31414235e-01 -1.21005011e+00
1.67996907e+00 -1.59813833e+00 -3.58560920e-01 8.36924791e-01
-1.53670937e-01 6.05304420e-01 7.14276433e-02 7.74033189e-01
-4.38092917e-01 2.14740112e-01 1.41664684e-01 7.46804476e-01
2.01064199e-01 3.84366363e-01 -1.25665575e-01 6.70938715e-02
-1.69121996e-01 9.73299026e-01 -7.13074923e-01 -3.49533588e-01
3.35114986e-01 -1.42017886e-01 -3.09116483e-01 -2.01923966e-01
-4.08464164e-01 1.34447768e-01 -6.80731595e-01 3.80671084e-01
2.70854920e-01 3.86641845e-02 1.20590198e+00 -4.13125247e-01
-6.97142333e-02 5.36741018e-01 -1.22864783e+00 1.42611432e+00
5.56774028e-02 -1.41732350e-01 -1.15941785e-01 -1.08765614e+00
1.19445407e+00 6.93927646e-01 4.22823101e-01 -8.18982601e-01
5.11821985e-01 1.75197691e-01 2.49529287e-01 -9.55752790e-01
3.10231268e-01 5.13734780e-02 -3.29559356e-01 2.73880839e-01
1.08412430e-01 3.24466452e-02 5.43540835e-01 2.23933205e-01
8.01107883e-01 5.65394998e-01 8.23464572e-01 -1.12264439e-01
8.55306327e-01 6.97482526e-01 4.50686246e-01 2.95509070e-01
2.98285880e-03 -5.69740653e-01 3.63621265e-01 -7.06429124e-01
-7.90698290e-01 -3.49960625e-01 4.82337698e-02 5.18175423e-01
-4.68628019e-01 -6.99226379e-01 -7.78713465e-01 -8.22404325e-01
-5.23539111e-02 1.20779967e+00 -2.63306946e-01 1.63880020e-01
-5.21930397e-01 -2.53810912e-01 6.69247866e-01 5.53116441e-01
4.89487648e-01 -1.30407655e+00 -5.47002792e-01 4.76369292e-01
-5.09425700e-02 -1.34110677e+00 3.04142475e-01 2.30358079e-01
-7.57183194e-01 -1.72614658e+00 7.03108370e-01 -8.47152174e-01
4.08253759e-01 -6.27723753e-01 8.02245140e-01 5.88843040e-02
2.73726061e-02 3.36221188e-01 -3.32001776e-01 -7.04466999e-01
-8.90736282e-01 -1.11886092e-01 -3.72768372e-01 -8.24940443e-01
1.33611798e+00 -3.97699594e-01 3.15489441e-01 -9.63479280e-02
-7.69916832e-01 -1.42119184e-01 2.42828563e-01 1.79620609e-01
2.41803035e-01 5.13437033e-01 6.63444877e-01 -1.64528012e+00
8.97199988e-01 -2.56417125e-01 -1.06412637e+00 5.16566396e-01
-1.19314861e+00 1.58912987e-01 3.62581998e-01 4.90002751e-01
-1.24059880e+00 2.99431205e-01 6.15871996e-02 1.71690673e-01
-3.57378691e-01 9.99309242e-01 -2.70478189e-01 8.44972059e-02
6.73713684e-01 -4.11709994e-01 -1.40735731e-01 -4.62910414e-01
6.85986459e-01 9.11717951e-01 7.20596075e-01 -5.90971231e-01
6.99550152e-01 -4.51375507e-02 -5.81346899e-02 -1.67795241e-01
-7.45102167e-01 -5.69722295e-01 -8.70955825e-01 -6.28807247e-02
1.05171466e+00 -9.00655866e-01 -7.88419068e-01 -3.63573909e-01
-7.63992727e-01 -1.72396615e-01 -2.29425713e-01 4.63868290e-01
-4.69091922e-01 3.88203323e-01 -3.67257357e-01 -1.01742005e+00
-5.81619680e-01 -4.68294650e-01 2.30780974e-01 2.57409424e-01
-9.89735186e-01 -1.27407122e+00 2.09703773e-01 1.09961140e+00
-5.76321147e-02 3.07271361e-01 1.51305091e+00 -1.44535446e+00
-2.21284807e-01 -3.07545990e-01 6.31826073e-02 2.62017220e-01
3.07608783e-01 -2.59141415e-01 -3.92746150e-01 1.74298495e-01
-2.69658089e-01 -1.23682193e-01 -1.04858249e-01 2.19230101e-01
4.48206097e-01 -3.56510013e-01 -4.44196850e-01 9.64459106e-02
1.53478110e+00 9.69356060e-01 6.90463901e-01 6.86997294e-01
4.28241849e-01 8.11357498e-01 6.59033120e-01 3.54203098e-02
7.26241350e-01 5.85648954e-01 1.95201281e-02 2.46572345e-01
-4.87773567e-02 -3.14515382e-01 -3.71359736e-02 5.94480634e-01
-4.85114753e-01 8.69575888e-02 -1.18693554e+00 5.37033916e-01
-2.04663229e+00 -1.02052212e+00 -3.39690477e-01 1.74511063e+00
6.15075052e-01 3.05669427e-01 1.61777779e-01 2.74574347e-02
5.50263584e-01 -4.65401262e-01 2.00621352e-01 -8.42344642e-01
3.80096771e-02 6.74347878e-01 7.10614681e-01 7.98724473e-01
-9.54858482e-01 1.40417600e+00 6.86259842e+00 1.89684540e-01
-4.41477746e-01 -2.74527241e-02 -1.70496061e-01 4.71280426e-01
-6.55359626e-02 3.62547547e-01 -8.35733473e-01 -1.04446234e-02
1.51740742e+00 -2.00377434e-01 4.98536915e-01 6.81004107e-01
6.53360665e-01 9.15034190e-02 -7.95671523e-01 4.18337613e-01
-1.89132050e-01 -1.66663706e+00 2.46561229e-01 2.16159239e-01
5.31314194e-01 -1.20899804e-01 -6.93606138e-01 4.04130757e-01
1.20494974e+00 -9.54850376e-01 3.73339206e-01 5.18244624e-01
4.34182942e-01 -8.22434664e-01 1.14062989e+00 9.95372981e-03
-1.08672178e+00 -5.74295104e-01 -2.27052405e-01 -4.64740545e-02
1.56256035e-01 -5.78438714e-02 -1.10615301e+00 1.40085578e+00
2.91970551e-01 5.59873283e-01 -7.24989295e-01 5.79956472e-01
-4.95954931e-01 3.46997023e-01 3.91738862e-02 3.00245166e-01
1.25107422e-01 -4.57111150e-01 1.71849832e-01 1.27977026e+00
1.32895917e-01 3.83758634e-01 3.26000780e-01 7.19823182e-01
2.33919159e-01 3.45826983e-01 -6.19035304e-01 -5.13509095e-01
3.98246378e-01 1.10363495e+00 -3.76390606e-01 -3.84748638e-01
-6.96994424e-01 2.02032879e-01 2.57874042e-01 2.68307626e-01
-5.88270366e-01 -1.96034342e-01 -1.10652283e-01 4.09126997e-01
3.73936266e-01 9.30290297e-02 -1.59530431e-01 -5.59168041e-01
-3.93318594e-01 -1.14241338e+00 1.23520958e+00 -1.01407683e+00
-7.47640729e-01 2.50495404e-01 5.07765114e-01 -7.30075598e-01
-4.51986760e-01 -7.72563457e-01 -3.99949461e-01 1.01269531e+00
-1.00098920e+00 -1.48657882e+00 1.18795611e-01 4.40906912e-01
-9.00322795e-02 -3.85145515e-01 1.22581446e+00 3.07173729e-01
-1.39236286e-01 -8.27192049e-03 -6.40809715e-01 4.91736174e-01
2.27931142e-01 -1.15664566e+00 2.68428773e-03 6.73841178e-01
2.79250622e-01 9.07012939e-01 5.10316253e-01 -1.43991435e+00
-8.10978591e-01 -1.19164753e+00 1.40334177e+00 -6.18212819e-01
9.01936412e-01 1.01497911e-01 -5.43701530e-01 1.24058330e+00
4.94812787e-01 -7.41895080e-01 1.13854659e+00 6.36143759e-02
-1.74817294e-01 -3.83163281e-02 -1.15897787e+00 1.94297239e-01
9.43578660e-01 -5.32944083e-01 -1.36748672e+00 4.18568164e-01
2.46241733e-01 -2.89129436e-01 -1.24336159e+00 2.71425962e-01
4.34598863e-01 -6.06679678e-01 4.46843982e-01 -1.20421517e+00
-1.43130228e-01 -6.38242424e-01 -1.43365026e-01 -7.22486317e-01
-5.13032794e-01 -6.22412026e-01 -4.42443341e-02 1.30428791e+00
8.74439180e-01 -5.52168667e-01 6.51582956e-01 9.01165307e-01
-2.00644821e-01 -1.21845953e-01 -6.60569191e-01 -5.38534701e-01
-2.97937512e-01 -5.12936473e-01 6.36413336e-01 1.17235553e+00
7.16246307e-01 5.44162989e-01 -4.95644584e-02 3.57945949e-01
3.18432063e-01 2.60498337e-02 4.28667903e-01 -1.58710587e+00
-1.20077908e-01 -6.79343268e-02 -5.86710036e-01 -6.56093210e-02
1.80780768e-01 -1.21072888e+00 -6.67632878e-01 -2.12406588e+00
-1.93055958e-01 -1.10281356e-01 -1.73592940e-01 9.62855279e-01
5.63930571e-01 -3.72727126e-01 -3.59170027e-02 4.98000532e-02
-2.86969036e-01 -4.00849074e-01 8.86706412e-01 -1.84114441e-01
-5.07899225e-01 -2.97599465e-01 -1.33078504e+00 6.90170169e-01
7.53771067e-01 -4.33848768e-01 -5.83579242e-01 1.08321212e-01
5.72073162e-01 4.65826504e-02 -9.54938829e-02 -9.46534395e-01
3.55626941e-01 -1.68596864e-01 4.05888170e-01 -5.94958663e-01
-3.89721423e-01 -1.27843821e+00 9.97226655e-01 6.49994850e-01
-4.07420605e-01 -3.58395800e-02 3.54260683e-01 1.89475030e-01
-7.74327219e-02 -5.39313853e-01 2.17058674e-01 -4.26293969e-01
-7.34705150e-01 -2.99870580e-01 -1.87020630e-01 -2.67364085e-01
1.00936687e+00 -5.36850572e-01 -3.67644101e-01 -2.62108535e-01
-1.34966528e+00 3.27062488e-01 4.80878279e-02 3.47967714e-01
-1.01941109e-01 -1.05212426e+00 -2.74821043e-01 4.00762260e-02
2.50139311e-02 -2.94100136e-01 -6.97675869e-02 5.49281776e-01
-6.32310748e-01 1.10236430e+00 -3.31705064e-01 1.18749231e-01
-1.40061533e+00 5.70315242e-01 4.90252465e-01 -7.28745699e-01
-5.45642495e-01 -1.38367601e-02 -5.78081131e-01 -3.45571667e-01
4.45185080e-02 -3.31757635e-01 -1.16449022e+00 5.64534729e-03
2.11075902e-01 1.24680968e-02 1.25630125e-01 -7.33608782e-01
-3.71962428e-01 4.20593530e-01 -7.92974979e-02 8.42718259e-02
1.50463176e+00 2.29701214e-02 -1.98161528e-01 -6.17420040e-02
2.89783299e-01 2.10525334e-01 -2.27557272e-01 -7.76115507e-02
8.32867920e-01 1.26680076e-01 -9.81458351e-02 -1.34598708e+00
-8.08249474e-01 1.10866250e-02 5.69805861e-01 2.80253947e-01
8.95355284e-01 -1.42921703e-02 2.82786459e-01 5.58088005e-01
3.45115930e-01 -1.59707403e+00 -7.65941978e-01 5.07023215e-01
6.30212069e-01 -5.78070164e-01 3.53285789e-01 -7.19502151e-01
-6.98151469e-01 1.34340346e+00 4.17452872e-01 1.08392477e-01
5.37025630e-01 3.86714906e-01 2.30092376e-01 -9.62263286e-01
-3.61477256e-01 -3.48990291e-01 2.13858247e-01 9.24370050e-01
5.86590052e-01 2.10930318e-01 -8.92434180e-01 1.01108932e+00
-3.14957559e-01 5.41142404e-01 4.03945744e-01 7.70303130e-01
-3.55140299e-01 -1.72229016e+00 -1.26329288e-01 4.85193431e-01
-6.76765382e-01 -1.39949366e-01 -8.55912685e-01 1.27500021e+00
6.21364236e-01 1.15042925e+00 -2.45804563e-01 -3.00852478e-01
7.76235342e-01 6.16651893e-01 4.27660525e-01 -8.93192053e-01
-1.11726892e+00 2.37282962e-02 1.19453859e+00 -3.97487342e-01
-8.82396221e-01 -4.23549354e-01 -1.48508596e+00 1.44954875e-01
-2.85469651e-01 6.58765912e-01 4.54835862e-01 1.10140669e+00
2.60952473e-01 5.25362611e-01 -2.78229892e-01 4.99850810e-01
-2.23024905e-01 -9.71444368e-01 -5.34406662e-01 5.55248618e-01
-5.79980850e-01 -5.00614047e-01 1.80321038e-01 3.46220762e-01] | [9.397933006286621, 8.510960578918457] |
9b864e5d-39aa-4d2c-8781-34334812ae73 | woce-a-framework-for-clustering-ensemble-by | 1612.06598 | null | http://arxiv.org/abs/1612.06598v1 | http://arxiv.org/pdf/1612.06598v1.pdf | WoCE: a framework for clustering ensemble by exploiting the wisdom of Crowds theory | The Wisdom of Crowds (WOC), as a theory in the social science, gets a new
paradigm in computer science. The WOC theory explains that the aggregate
decision made by a group is often better than those of its individual members
if specific conditions are satisfied. This paper presents a novel framework for
unsupervised and semi-supervised cluster ensemble by exploiting the WOC theory.
We employ four conditions in the WOC theory, i.e., diversity, independency,
decentralization and aggregation, to guide both the constructing of individual
clustering results and the final combination for clustering ensemble. Firstly,
independency criterion, as a novel mapping system on the raw data set, removes
the correlation between features on our proposed method. Then, decentralization
as a novel mechanism generates high-quality individual clustering results.
Next, uniformity as a new diversity metric evaluates the generated clustering
results. Further, weighted evidence accumulation clustering method is proposed
for the final aggregation without using thresholding procedure. Experimental
study on varied data sets demonstrates that the proposed approach achieves
superior performance to state-of-the-art methods. | ['Sheng-Jun Huang', 'Daoqiang Zhang', 'Muhammad Yousefnezhad'] | 2016-12-20 | null | null | null | null | ['clustering-ensemble'] | ['graphs'] | [-8.06169659e-02 -2.98011154e-01 4.16690886e-01 -3.63941252e-01
-3.36477667e-01 -4.30319637e-01 6.69996440e-01 5.96364439e-01
-5.21745086e-01 7.35385716e-01 3.52263838e-01 3.11112881e-01
-5.09004176e-01 -7.10687041e-01 -2.29835048e-01 -1.26443243e+00
1.23269834e-01 5.49834192e-01 3.71711850e-01 4.85837571e-02
5.63523948e-01 9.03808028e-02 -1.88032079e+00 2.85221070e-01
1.34437072e+00 7.44503260e-01 6.21882379e-02 3.15609485e-01
-1.97100520e-01 4.92058307e-01 -6.48136139e-01 -2.12511912e-01
2.36535072e-01 -6.00795329e-01 -4.43793863e-01 3.04479569e-01
-4.93521690e-01 2.85017788e-01 6.68081403e-01 1.02883422e+00
6.36877596e-01 2.94941276e-01 1.09547687e+00 -1.38487911e+00
-7.39452958e-01 7.05658615e-01 -8.43226433e-01 -1.50611073e-01
2.75706798e-01 4.33207937e-02 7.36960292e-01 -7.66539991e-01
3.46022636e-01 1.22770882e+00 4.08406645e-01 1.90460339e-01
-8.32961798e-01 -4.02574539e-01 4.41791266e-02 2.35919014e-01
-1.63767910e+00 -3.74298096e-01 5.84417164e-01 -6.28459513e-01
4.32092965e-01 1.38032824e-01 4.99950111e-01 1.22000962e-01
1.70080602e-01 4.27235514e-01 1.38569379e+00 -6.78010583e-01
7.46476054e-01 2.99950838e-01 2.52308041e-01 4.52783197e-01
6.21748805e-01 -4.72188443e-01 -3.81137609e-01 -3.04099709e-01
2.27902737e-02 2.18074784e-01 7.21529424e-02 -4.59412709e-02
-1.18071055e+00 5.96935868e-01 2.44044349e-01 6.08650386e-01
-6.99504793e-01 -1.68357641e-01 2.82626301e-01 -1.04915723e-01
3.80636483e-01 1.19296215e-01 8.91643837e-02 2.14371428e-01
-1.01813507e+00 -4.89374250e-03 6.77064538e-01 7.49550402e-01
7.94895351e-01 -2.32740283e-01 -3.12456012e-01 6.69311106e-01
4.78430331e-01 5.84155381e-01 7.56306171e-01 -1.03031552e+00
2.63805121e-01 1.24866736e+00 1.94959015e-01 -1.44538653e+00
-3.50941807e-01 -4.22410488e-01 -1.02441478e+00 1.38230965e-01
6.77143857e-02 -3.74844730e-01 -4.62020665e-01 1.41136765e+00
7.55232334e-01 2.39833638e-01 3.17580521e-01 8.03358555e-01
7.81025887e-01 5.76240480e-01 -7.14215040e-02 -6.04753494e-01
1.20707965e+00 -6.12340152e-01 -9.08805609e-01 4.86851752e-01
3.61507654e-01 -8.14280033e-01 5.05436897e-01 5.62295616e-01
-8.24162543e-01 -5.03949046e-01 -1.04921055e+00 4.73491609e-01
-4.82062072e-01 2.17227429e-01 2.67065048e-01 5.80986857e-01
-1.08043039e+00 3.56368035e-01 -3.24221075e-01 -7.36911356e-01
1.34140983e-01 3.09750885e-01 -1.94805652e-01 1.79108724e-01
-9.05653298e-01 7.49716043e-01 7.28048861e-01 2.04889625e-02
-2.58955508e-01 -5.96044511e-02 -1.76444814e-01 7.36034662e-02
4.83879089e-01 -5.70015728e-01 5.20591676e-01 -8.64373326e-01
-1.16253757e+00 6.67309046e-01 -2.01527894e-01 -3.08335841e-01
4.55592990e-01 2.24299133e-01 -6.53869092e-01 1.89260349e-01
4.17129934e-01 3.32505405e-01 5.07610679e-01 -1.81755733e+00
-1.18393409e+00 -5.10649860e-01 -5.18003643e-01 5.13409257e-01
-2.83552378e-01 1.04882466e-02 -2.62646437e-01 -3.41265410e-01
2.97229677e-01 -7.24600375e-01 -3.04600716e-01 -5.59347451e-01
-2.46816754e-01 -6.33047342e-01 5.95546901e-01 -3.54566216e-01
1.53602898e+00 -2.07755160e+00 1.47947237e-01 8.90409946e-01
4.09704357e-01 1.55742735e-01 2.76637852e-01 5.09023011e-01
3.75666231e-01 3.59849066e-01 -4.88153130e-01 -1.13069840e-01
7.79742151e-02 1.44276187e-01 3.06176752e-01 3.61905158e-01
1.07001606e-02 2.32232973e-01 -1.05807316e+00 -1.07719624e+00
2.58477777e-01 -9.18206002e-04 -3.04152012e-01 -1.08686052e-01
1.58103421e-01 4.83369231e-01 -4.19966161e-01 5.47442138e-01
7.43797064e-01 -2.13950068e-01 3.74123484e-01 7.67650604e-02
-4.06343281e-01 -7.37429380e-01 -1.58460522e+00 1.06672680e+00
-1.28362868e-02 -9.95486006e-02 6.40701875e-02 -1.02660060e+00
1.17735207e+00 4.30343390e-01 6.63585305e-01 -1.78826287e-01
5.08817673e-01 5.13051391e-01 7.60886371e-02 -6.57323718e-01
4.09226090e-01 5.15970141e-02 -6.93433508e-02 7.39765942e-01
-1.33284375e-01 1.75846681e-01 6.33160830e-01 3.15862179e-01
8.42216194e-01 -2.90937811e-01 6.05742931e-01 -5.85767806e-01
9.02854979e-01 5.44476695e-02 8.81818771e-01 6.09277666e-01
-1.54972374e-01 4.27774280e-01 1.39955118e-01 -8.33044201e-02
-7.80680835e-01 -7.37511635e-01 1.57731354e-01 8.85008335e-01
3.97853225e-01 -1.85702905e-01 -9.75956380e-01 -6.00927949e-01
-2.32493924e-03 4.10582155e-01 -6.18888259e-01 1.78263545e-01
-1.58944264e-01 -8.25625241e-01 2.58326650e-01 1.38353854e-01
9.02375102e-01 -1.02909040e+00 -3.17682773e-01 2.93303937e-01
-2.50654995e-01 -7.45786607e-01 -2.18672901e-01 -3.22198600e-01
-6.65748060e-01 -1.10131466e+00 -4.81182456e-01 -7.46931493e-01
8.87094676e-01 6.54333174e-01 7.10900784e-01 8.09740067e-01
2.07848758e-01 3.43869746e-01 -8.91354740e-01 -4.60319340e-01
-2.60235637e-01 -1.19069695e-01 3.67137283e-01 6.28064811e-01
7.24202216e-01 -7.26266265e-01 -7.34537482e-01 3.25357258e-01
-7.85056949e-01 -1.25585049e-01 5.75001180e-01 4.42836702e-01
4.36827391e-01 7.96901107e-01 9.03949678e-01 -6.44375563e-01
9.76743340e-01 -8.02544236e-01 -1.39125198e-01 5.04224479e-01
-8.08912635e-01 -2.67049260e-02 8.07272613e-01 -3.70980948e-02
-1.34004629e+00 3.13871801e-02 5.32184958e-01 1.52848497e-01
-2.94382989e-01 5.55885553e-01 -2.70412385e-01 2.82200545e-01
7.51088560e-01 2.93538779e-01 7.46187270e-02 -3.23578387e-01
3.42745811e-01 1.17867291e+00 4.81037825e-01 -7.82662690e-01
6.84105396e-01 8.21137130e-01 -1.18993498e-01 -6.72313690e-01
-3.19002628e-01 -8.17295551e-01 -9.27791774e-01 -7.09675431e-01
1.09207559e+00 -8.13650429e-01 -9.30710912e-01 3.58308494e-01
-1.07383859e+00 4.01677489e-01 9.02166069e-02 5.25175989e-01
-1.10702433e-01 4.62477028e-01 2.00526863e-01 -1.30863965e+00
-3.70174706e-01 -8.91179502e-01 6.97984397e-01 6.18828297e-01
-1.05982937e-01 -8.82312417e-01 -1.44974470e-01 3.69434834e-01
3.76759544e-02 5.85065782e-01 5.96151590e-01 -9.02296484e-01
-2.39192203e-01 -1.86853200e-01 -7.43706822e-02 4.12501723e-01
3.71423066e-01 4.42168981e-01 -6.81259692e-01 -1.77508537e-02
-1.42612320e-03 3.69100749e-01 7.53408432e-01 1.78120136e-01
7.99362898e-01 -3.70730639e-01 -3.12936187e-01 -3.88121456e-02
1.51371574e+00 5.34972370e-01 3.07964891e-01 2.02100605e-01
4.52757388e-01 8.00761580e-01 7.11836696e-01 8.52769911e-01
5.37631929e-01 1.29045159e-01 6.73683807e-02 1.53200477e-01
1.78878710e-01 1.59226179e-01 1.74295649e-01 1.36927116e+00
-6.85065210e-01 -3.45167637e-01 -1.05914485e+00 5.91569722e-01
-2.36004615e+00 -1.07375550e+00 -4.76404220e-01 2.16485476e+00
6.13640726e-01 -8.50068703e-02 1.45835102e-01 5.31170964e-01
1.38236856e+00 -2.46108830e-01 -2.42046848e-01 -1.92803904e-01
-3.11532170e-01 -1.70301616e-01 3.00600082e-01 4.59879309e-01
-8.25433671e-01 6.78399444e-01 5.12725353e+00 1.03192067e+00
-4.99817461e-01 3.19414496e-01 2.86726147e-01 1.12847589e-01
-2.25727692e-01 7.29086474e-02 -6.09706461e-01 9.46025133e-01
3.24302554e-01 -4.90643680e-01 3.46498758e-01 4.05265629e-01
6.02662325e-01 -6.38674855e-01 -6.07120216e-01 8.17180395e-01
1.48574308e-01 -1.01796997e+00 4.70020100e-02 5.11928424e-02
1.24714756e+00 -5.78287899e-01 -2.69457072e-01 -1.03160650e-01
9.02199090e-01 -6.65386140e-01 5.84002674e-01 1.07290196e+00
1.59858003e-01 -8.82077694e-01 1.19278729e+00 6.32241249e-01
-1.21623909e+00 -2.22489968e-01 -2.71692961e-01 -2.53070980e-01
3.57563756e-02 1.00187695e+00 -5.34061670e-01 9.82731104e-01
6.74667895e-01 6.03729963e-01 -8.69750261e-01 9.87651587e-01
-1.20601892e-01 5.56690454e-01 -4.91816729e-01 -5.12007654e-01
1.92981064e-01 -4.70354110e-01 4.32068229e-01 1.16196299e+00
4.33136970e-01 4.52252746e-01 3.16678733e-01 4.22618121e-01
2.43195266e-01 6.21381760e-01 -3.84415507e-01 3.34862500e-01
1.08910358e+00 1.23748863e+00 -1.22089887e+00 -6.83540225e-01
4.16458510e-02 5.26326835e-01 3.34066227e-02 4.16587025e-01
-5.36551833e-01 -3.35744053e-01 -2.13216126e-01 -1.01924226e-01
4.16743904e-02 -1.27694160e-01 -7.21550465e-01 -8.36655915e-01
1.79323629e-01 -7.00284898e-01 4.46197778e-01 -5.76176286e-01
-1.49574459e+00 3.38582039e-01 -2.17856262e-02 -1.54683447e+00
1.57245755e-01 -8.79599825e-02 -1.01932585e+00 6.52021766e-01
-8.91191185e-01 -9.19321656e-01 -5.82722008e-01 7.17041135e-01
1.58467904e-01 -5.69738686e-01 4.79758114e-01 2.74190269e-02
-5.44509172e-01 1.27105221e-01 6.99817479e-01 -1.40485749e-01
8.15255046e-01 -1.19714975e+00 -5.68297803e-01 1.06981397e+00
-3.53289902e-01 7.75340974e-01 6.09115243e-01 -9.12874579e-01
-6.91675425e-01 -8.49341035e-01 8.50043535e-01 -2.53647596e-01
1.98049203e-01 -1.67296715e-02 -7.10213780e-01 1.84209328e-02
4.21452969e-01 -4.26314801e-01 1.01245177e+00 -4.10525538e-02
9.78362784e-02 -3.87877434e-01 -1.48057079e+00 5.30179203e-01
6.56605542e-01 7.90892020e-02 -6.26516044e-01 1.87589198e-01
3.97987694e-01 3.92915070e-01 -9.81518567e-01 2.13397264e-01
5.62759161e-01 -1.22924185e+00 4.79489207e-01 -8.88904035e-02
2.88161218e-01 -1.07895947e+00 -3.19906414e-01 -1.36705947e+00
-5.37497342e-01 -3.03430021e-01 2.84022987e-01 1.74020338e+00
3.36668998e-01 -7.45249391e-01 2.94420421e-01 4.49834049e-01
1.89128086e-01 -3.34767640e-01 -7.63227880e-01 -7.33511746e-01
-8.26811269e-02 6.18064329e-02 1.04965079e+00 1.07515335e+00
2.04132333e-01 2.07608640e-01 -6.62396848e-02 1.41037270e-01
8.95781398e-01 -1.73993155e-01 8.16637158e-01 -1.72286487e+00
1.29361048e-01 -5.21041572e-01 -3.88930440e-01 -1.47834197e-01
-1.54472604e-01 -8.76109660e-01 1.23508669e-01 -1.91430092e+00
2.20646769e-01 -5.27000487e-01 -4.29188102e-01 2.14925453e-01
-4.22936916e-01 1.22402959e-01 3.94851834e-01 6.72600508e-01
-1.01894426e+00 3.52400392e-01 9.23228264e-01 7.61185065e-02
-4.00165886e-01 -2.43127555e-01 -8.77237022e-01 8.40955615e-01
1.06450260e+00 -4.18195486e-01 -2.83385247e-01 -8.92786384e-02
1.47146851e-01 -3.40207785e-01 3.16197351e-02 -1.28531265e+00
7.06402838e-01 -3.23643476e-01 2.75616139e-01 -5.28777540e-01
-2.59324521e-01 -1.04716980e+00 3.46794993e-01 3.25976878e-01
-1.29711613e-01 -2.18369067e-02 -4.34660554e-01 6.53678060e-01
-1.45468310e-01 -2.97383666e-01 7.28135765e-01 -2.33251154e-01
-5.60814500e-01 -2.16747656e-01 -4.84685928e-01 -8.67801756e-02
1.22801244e+00 -4.71023500e-01 -2.71068245e-01 -2.08261698e-01
-6.88296616e-01 7.57885635e-01 5.99531353e-01 1.41493410e-01
2.61367619e-01 -1.56558156e+00 -1.06261361e+00 -2.51902133e-01
1.53986551e-02 -5.49056288e-03 3.10409576e-01 1.00041175e+00
-4.27117795e-01 -2.13095844e-02 -8.37033764e-02 -7.00457573e-01
-1.42478275e+00 3.91753227e-01 -2.36222088e-01 -1.77050605e-01
-1.50033906e-01 1.97354183e-01 -1.93223417e-01 -2.85656840e-01
-4.14083004e-02 1.11714132e-01 -7.11860359e-01 4.01713014e-01
3.51794869e-01 8.14277112e-01 -4.32777070e-02 -8.91926348e-01
-4.99698818e-01 8.30017388e-01 3.12833011e-01 -4.15771425e-01
1.13629067e+00 -4.20870185e-01 -6.42046273e-01 5.31169474e-01
6.66995168e-01 2.58049890e-02 -6.69917881e-01 -3.12436581e-01
1.28252104e-01 -2.30620682e-01 -2.84524143e-01 -8.71211886e-01
-7.19879985e-01 5.06905317e-01 4.31338817e-01 6.65132582e-01
1.14319885e+00 -2.91977048e-01 2.67693698e-01 3.45993310e-01
3.86334032e-01 -1.72920978e+00 -1.16160065e-01 2.41187528e-01
7.21943378e-01 -1.34015357e+00 1.39006302e-01 -1.73322946e-01
-9.44607437e-01 8.19041431e-01 5.31100094e-01 -2.21251816e-01
9.18934464e-01 -1.13811366e-01 -5.25037833e-02 -2.16515869e-01
-5.91136456e-01 -5.53812563e-01 1.26004651e-01 6.09119952e-01
2.85598606e-01 4.67058241e-01 -1.10631001e+00 8.77211273e-01
-1.58270314e-01 9.52951238e-03 5.15643299e-01 8.71552765e-01
-1.07981408e+00 -6.54636204e-01 -8.90592098e-01 4.96261388e-01
-2.42323145e-01 2.51506448e-01 -6.06351495e-01 5.26567042e-01
9.38682675e-01 1.75305641e+00 -6.90522343e-02 -6.17600203e-01
8.64454731e-02 1.42391413e-01 -2.81255171e-02 -4.76060480e-01
-6.95964456e-01 2.88571473e-02 -1.75575718e-01 -6.91239396e-03
-1.17005801e+00 -6.59595549e-01 -1.60322702e+00 -4.88921911e-01
-5.45541108e-01 8.02390397e-01 6.43451929e-01 1.10503924e+00
5.33203125e-01 3.76390815e-01 9.72961903e-01 -6.45644784e-01
6.77934214e-02 -1.07408381e+00 -7.41513431e-01 5.22885501e-01
-2.30982780e-01 -8.39248121e-01 -6.85587704e-01 3.36086988e-01] | [7.619149208068848, 4.559760570526123] |
280ffcfd-dfc4-4a60-8b75-ac1c596466ae | self-aware-feedback-based-self-learning-in | 2205.00029 | null | https://arxiv.org/abs/2205.00029v1 | https://arxiv.org/pdf/2205.00029v1.pdf | Self-Aware Feedback-Based Self-Learning in Large-Scale Conversational AI | Self-learning paradigms in large-scale conversational AI agents tend to leverage user feedback in bridging between what they say and what they mean. However, such learning, particularly in Markov-based query rewriting systems have far from addressed the impact of these models on future training where successive feedback is inevitably contingent on the rewrite itself, especially in a continually updating environment. In this paper, we explore the consequences of this inherent lack of self-awareness towards impairing the model performance, ultimately resulting in both Type I and II errors over time. To that end, we propose augmenting the Markov Graph construction with a superposition-based adjacency matrix. Here, our method leverages an induced stochasticity to reactively learn a locally-adaptive decision boundary based on the performance of the individual rewrites in a bi-variate beta setting. We also surface a data augmentation strategy that leverages template-based generation in abridging complex conversation hierarchies of dialogs so as to simplify the learning process. All in all, we demonstrate that our self-aware model improves the overall PR-AUC by 27.45%, achieves a relative defect reduction of up to 31.22%, and is able to adapt quicker to changes in global preferences across a large number of customers. | ['Chenlei Guo', 'Chengyuan Ma', 'Gustavo Aguilar', 'Clint Solomon Mathialagan', 'Pragaash Ponnusamy'] | 2022-04-29 | null | https://aclanthology.org/2022.naacl-industry.36 | https://aclanthology.org/2022.naacl-industry.36.pdf | naacl-acl-2022-7 | ['self-learning'] | ['natural-language-processing'] | [ 3.45019698e-01 4.60917920e-01 -3.95280756e-02 -4.84829128e-01
-9.25276875e-01 -7.78092802e-01 5.94151556e-01 2.46325821e-01
-2.91939169e-01 7.96291947e-01 3.66585553e-01 -4.90482897e-01
-1.09361541e-02 -6.91612780e-01 -6.41897380e-01 -4.58865225e-01
-7.12942407e-02 9.78911042e-01 1.56035230e-01 -7.26550341e-01
1.10569514e-01 2.71115471e-02 -1.20241630e+00 2.18161702e-01
1.16348743e+00 6.22885644e-01 2.65521705e-02 8.56370628e-01
-7.57316351e-02 7.99084902e-01 -8.27182174e-01 -7.40931034e-01
2.32772291e-01 -4.97500151e-01 -7.61306047e-01 1.06680408e-01
2.64063537e-01 -2.78974682e-01 -2.58979976e-01 6.49391830e-01
4.75315005e-01 2.23079324e-01 5.01970172e-01 -9.42111552e-01
-3.33472967e-01 1.14490926e+00 -2.47560248e-01 1.55337915e-01
3.85310143e-01 6.15126967e-01 1.33352625e+00 -1.37019053e-01
5.76169968e-01 1.30889571e+00 6.06094241e-01 6.61216438e-01
-1.49903548e+00 -4.99702066e-01 5.76328039e-01 1.03051633e-01
-9.92959082e-01 -6.09252155e-01 7.08460033e-01 -1.50920540e-01
1.11787045e+00 3.57601464e-01 5.97226739e-01 1.17861485e+00
3.09267938e-01 6.41848564e-01 9.74888384e-01 -4.03528720e-01
2.93350518e-01 1.58506736e-01 2.28448641e-02 8.98883283e-01
-8.82148743e-06 -5.21822758e-02 -6.43386364e-01 -4.08056915e-01
2.53777564e-01 -2.06118509e-01 1.70784175e-01 -3.58395129e-01
-6.89294994e-01 9.04207349e-01 1.29479885e-01 1.45065546e-01
-3.99230540e-01 3.93388383e-02 1.39872164e-01 6.21153355e-01
4.31926608e-01 1.05187297e+00 -5.43601811e-01 -4.99228656e-01
-4.52777416e-01 3.12051535e-01 1.32928419e+00 8.49155188e-01
7.50928402e-01 -2.36841794e-02 -3.78505141e-01 8.21095765e-01
9.24923830e-03 4.29823875e-01 3.63729119e-01 -1.06208551e+00
5.19439757e-01 8.95276606e-01 -3.77805606e-02 -7.64219165e-01
-4.78479087e-01 -6.66460752e-01 -5.10594666e-01 -1.50354132e-01
3.24006319e-01 -5.89395523e-01 -8.81715357e-01 2.18705392e+00
4.07593697e-01 -1.18703336e-01 6.99668229e-02 6.19502783e-01
1.39689505e-01 3.45332921e-01 -9.90744904e-02 -3.45820516e-01
1.01052928e+00 -9.55876648e-01 -5.99802077e-01 -4.99910653e-01
8.53028357e-01 -6.02025628e-01 1.27083683e+00 3.85208964e-01
-1.08359885e+00 -5.21485880e-02 -1.09539020e+00 2.07619265e-01
4.47971895e-02 -5.37204325e-01 8.49251151e-01 7.36862183e-01
-1.13679826e+00 5.64806104e-01 -7.90257752e-01 -4.46823716e-01
4.22802716e-02 6.59691632e-01 2.49597326e-01 1.69403777e-01
-1.35018098e+00 9.83701766e-01 -1.28694415e-01 -2.60812342e-01
-5.52910984e-01 -6.49976552e-01 -4.39668894e-01 2.02283487e-01
8.91057432e-01 -1.04401791e+00 1.71564496e+00 -8.19017053e-01
-1.95775819e+00 2.69232690e-01 -1.74980551e-01 -6.73843741e-01
5.89683354e-01 -1.44810090e-03 -2.36742437e-01 -3.66041422e-01
-2.00802997e-01 4.18100387e-01 6.45157695e-01 -1.20334458e+00
-6.23357475e-01 -4.01262611e-01 3.81054550e-01 6.44393861e-01
-4.53265756e-01 -5.71024120e-01 -5.02489984e-01 -4.52774882e-01
-1.55021265e-01 -1.41125369e+00 -4.76464152e-01 -6.86643124e-01
-4.00661081e-01 -3.48791480e-01 2.71571785e-01 -2.89731055e-01
1.50556457e+00 -1.67981708e+00 2.81368166e-01 3.69403899e-01
2.76922792e-01 1.36221901e-01 -2.84406990e-01 7.63071775e-01
5.91495752e-01 7.11108372e-02 -2.50337780e-01 -4.87588406e-01
-5.95682897e-02 3.03594202e-01 -2.03111932e-01 -8.08352977e-02
1.13384023e-01 8.02055240e-01 -8.56008530e-01 -3.45264852e-01
-2.14832857e-01 8.90041292e-02 -1.15241849e+00 3.24242562e-01
-7.06894636e-01 5.28676867e-01 -4.56417650e-01 3.39502543e-01
4.19018418e-02 -3.81199747e-01 7.67805696e-01 1.63216710e-01
2.59614825e-01 6.70095265e-01 -8.11881602e-01 1.62757313e+00
-7.98128963e-01 2.35260278e-01 1.66639447e-01 -7.04004467e-01
8.28648925e-01 -6.12062514e-02 5.75186491e-01 -7.50686884e-01
3.18267643e-02 -1.98525619e-02 4.98741180e-01 -3.31821322e-01
5.80821931e-01 -9.36444327e-02 -1.05819531e-01 8.73542249e-01
-1.46459654e-01 -3.11447680e-01 3.19854319e-01 4.45631146e-01
1.41049528e+00 -3.34025294e-01 6.83667883e-03 -1.28651736e-02
2.36868888e-01 -6.54807314e-02 5.23017645e-01 1.13864875e+00
5.87462075e-02 2.79695779e-01 6.78668678e-01 -2.00386852e-01
-9.32050884e-01 -7.24250138e-01 3.28495562e-01 1.49234176e+00
-3.61106247e-02 -5.70084929e-01 -7.57586122e-01 -6.54117227e-01
2.29556426e-01 1.07552540e+00 -5.07262886e-01 -6.81082010e-01
-6.26972675e-01 -7.30425715e-01 4.74796474e-01 1.47216082e-01
2.28051409e-01 -7.93109655e-01 -3.29655558e-01 3.93759936e-01
-9.83958989e-02 -9.89807963e-01 -8.22824001e-01 1.84547424e-01
-8.27961266e-01 -8.74828100e-01 -1.66771263e-01 -4.22000527e-01
3.59738618e-01 1.24676907e-02 1.17500806e+00 1.11420982e-01
4.46020439e-02 5.31697631e-01 -1.94298938e-01 -3.17844063e-01
-9.75844264e-01 7.23847866e-01 2.40106061e-01 -2.46300921e-02
1.15469038e-01 -8.07551086e-01 -4.78607655e-01 3.27551723e-01
-6.41662657e-01 -4.53480193e-03 7.74904847e-01 9.26534355e-01
-5.39095886e-02 -4.15717661e-01 6.34268165e-01 -1.45884597e+00
1.22553647e+00 -4.97951835e-01 -4.51717585e-01 5.76678753e-01
-1.31337535e+00 5.39879560e-01 6.54455781e-01 -6.14383638e-01
-1.24134171e+00 -1.23729497e-01 2.01271884e-02 -2.26430744e-02
3.38262528e-01 5.51347256e-01 2.49182940e-01 1.03195429e-01
8.64208341e-01 1.45200208e-01 2.98324168e-01 -2.25518376e-01
5.86643100e-01 6.98734701e-01 1.86721310e-01 -7.50705361e-01
7.01012194e-01 -8.28034303e-04 -2.19775304e-01 -5.24029613e-01
-7.96875060e-01 -2.57470489e-01 -1.16219446e-01 -1.59019470e-01
3.21231097e-01 -7.18986452e-01 -1.14463592e+00 3.11746389e-01
-8.51487577e-01 -8.19854379e-01 -2.58017957e-01 9.76133570e-02
-5.37950754e-01 1.64615661e-01 -6.71734214e-01 -9.64888215e-01
-5.09937704e-01 -1.17169118e+00 5.86887658e-01 3.03874850e-01
-4.24825430e-01 -1.00410593e+00 2.67780900e-01 6.97002113e-01
7.58049667e-01 -3.26268464e-01 1.06824625e+00 -1.06651831e+00
-6.02925897e-01 -8.61141011e-02 3.22104931e-01 6.10549524e-02
2.82855392e-01 -1.09771267e-01 -7.18557000e-01 -3.65163207e-01
-2.51862984e-02 -4.08150792e-01 5.48103988e-01 -6.48607090e-02
6.88943088e-01 -6.54355645e-01 -2.37058550e-01 1.74709827e-01
7.12952852e-01 3.18877995e-01 2.23525893e-02 8.03241655e-02
5.46423137e-01 6.12197936e-01 3.50249350e-01 5.96068621e-01
8.03588271e-01 9.09099936e-01 2.01218799e-01 2.92912573e-01
4.69003897e-03 -3.09815675e-01 5.35487711e-01 9.65492487e-01
2.33401865e-01 -4.60704446e-01 -9.53073740e-01 2.73305207e-01
-1.96555531e+00 -6.63759172e-01 3.53266597e-01 2.13749194e+00
1.35329485e+00 4.57688034e-01 2.52896696e-01 -4.56297517e-01
4.37778294e-01 1.72130108e-01 -1.08002353e+00 -5.83706319e-01
8.24549794e-02 3.38236354e-02 5.55926502e-01 8.66298795e-01
-5.53597152e-01 1.20688796e+00 6.35670614e+00 4.62684095e-01
-1.10270965e+00 -4.70877290e-02 7.26568878e-01 -3.10687035e-01
-5.51847577e-01 1.10407524e-01 -6.74069762e-01 3.42501193e-01
1.20030248e+00 -3.51804972e-01 1.05098474e+00 5.54357827e-01
5.08079007e-02 -5.22862971e-02 -1.18666852e+00 4.18510795e-01
7.23811910e-02 -1.28208625e+00 1.06264934e-01 8.57921541e-02
8.03620338e-01 4.84038852e-02 1.83139428e-01 7.05967546e-01
1.01758730e+00 -7.35960543e-01 2.92355984e-01 5.01656353e-01
3.49517226e-01 -6.79734886e-01 4.85623747e-01 5.77561259e-01
-5.29268265e-01 -3.45324963e-01 8.89700353e-02 -1.16526112e-01
6.15725107e-02 3.09157073e-01 -1.57735586e+00 2.53993839e-01
1.57191157e-01 -5.17893285e-02 -4.56432015e-01 5.26601613e-01
1.34781944e-02 8.04501891e-01 -4.82654691e-01 -4.06151712e-01
1.07200637e-01 -2.00067446e-01 6.18840992e-01 8.98359060e-01
-1.52593032e-01 3.01897913e-01 3.12835336e-01 5.08570254e-01
-2.72858262e-01 1.06187582e-01 -4.00861263e-01 -1.53444842e-01
6.73142791e-01 1.03019989e+00 -1.70213401e-01 -2.56987721e-01
-7.73242712e-02 7.13443398e-01 6.15549982e-01 4.02835518e-01
-5.84216356e-01 -6.02895953e-02 7.40157485e-01 3.56018916e-02
2.05043256e-01 -2.02009961e-01 -3.68479401e-01 -1.06629241e+00
-2.20553465e-02 -1.36053324e+00 4.03587639e-01 -1.56251982e-01
-1.24893081e+00 5.65060973e-01 -2.28518128e-01 -5.67611337e-01
-7.90267110e-01 3.09914406e-02 -6.95507407e-01 5.71654260e-01
-1.12939477e+00 -9.05228198e-01 5.31383678e-02 1.95832595e-01
6.62410975e-01 -3.38172078e-01 7.93627620e-01 6.58726320e-02
-6.28728211e-01 1.08182931e+00 7.71612898e-02 -3.00109923e-01
9.61841226e-01 -1.19444478e+00 5.75383067e-01 4.58725035e-01
2.84954309e-01 9.05213475e-01 1.11792445e+00 -6.65735543e-01
-1.68547916e+00 -6.52373314e-01 6.37600899e-01 -6.46264493e-01
7.42577672e-01 -4.65994775e-01 -8.17618191e-01 6.31350875e-01
1.69765294e-01 -6.22031629e-01 4.82737690e-01 7.64867663e-01
-3.40593845e-01 -4.32240456e-01 -9.36541438e-01 9.38286126e-01
1.24141371e+00 -3.60144734e-01 -2.00127244e-01 4.95622665e-01
1.02810502e+00 -5.18008947e-01 -8.64000499e-01 4.11673754e-01
4.23844397e-01 -8.29783022e-01 5.59979081e-01 -9.90248561e-01
9.80292484e-02 2.75929898e-01 6.27996922e-02 -1.46268857e+00
-2.14768425e-01 -1.23727548e+00 -3.10226351e-01 1.13247252e+00
8.76600981e-01 -8.06318283e-01 1.11179113e+00 8.80298913e-01
-1.62319481e-01 -1.03500259e+00 -8.22256505e-01 -4.77373183e-01
1.59330724e-03 -1.14258125e-01 4.97849047e-01 7.47742474e-01
2.12268919e-01 9.10107613e-01 -4.16351706e-01 -8.48349091e-03
3.15008491e-01 8.44361335e-02 1.06625950e+00 -1.06019962e+00
-6.72593236e-01 -5.13079524e-01 2.39294007e-01 -1.25033545e+00
7.52968565e-02 -6.56871319e-01 1.61694542e-01 -1.16092098e+00
9.69349965e-02 -7.24906981e-01 -1.18503660e-01 3.02938789e-01
-3.83702099e-01 -3.94815058e-01 2.62918711e-01 2.84059733e-01
-8.17128479e-01 6.56809747e-01 1.19138920e+00 -1.56042889e-01
-6.92995429e-01 3.91175866e-01 -9.45043683e-01 2.68859744e-01
8.73434126e-01 -4.11224514e-01 -7.50755608e-01 -2.40147993e-01
2.79708028e-01 1.36515304e-01 -2.59920985e-01 -7.23650873e-01
5.94222248e-01 -3.02880436e-01 -3.83871198e-01 -9.68576595e-02
4.81770217e-01 -4.53903556e-01 -4.44153436e-02 4.68592435e-01
-8.75125289e-01 2.60669887e-01 9.24277604e-02 8.86545300e-01
2.80184984e-01 9.73435268e-02 3.63445491e-01 -8.18187073e-02
-1.13735408e-01 8.50098133e-02 -4.04798001e-01 3.43271971e-01
7.78543949e-01 1.78299844e-01 -3.67031634e-01 -9.78939891e-01
-5.91722369e-01 6.49739444e-01 4.13486809e-01 5.20025551e-01
2.92059872e-02 -6.63954735e-01 -5.88910460e-01 7.87285417e-02
-4.89421003e-03 -7.09851310e-02 2.47753739e-01 8.29160333e-01
-3.50390039e-02 4.94613349e-01 1.44326448e-01 -4.63107795e-01
-1.11439490e+00 2.77130097e-01 3.78074527e-01 -7.72751331e-01
-1.95976987e-01 8.58818412e-01 -1.19808435e-01 -5.47828853e-01
5.19173682e-01 -1.97703198e-01 1.00587197e-02 5.71396425e-02
-1.14902213e-01 1.59360781e-01 1.57476917e-01 -6.67621382e-03
-1.96172029e-01 -6.92193061e-02 -6.65658414e-01 -4.67427522e-01
1.20742357e+00 -2.12961838e-01 1.95612460e-01 5.32992363e-01
7.56981254e-01 2.16286287e-01 -1.37330461e+00 -5.10564268e-01
1.24098986e-01 -2.33456790e-02 -2.32028708e-01 -1.16160440e+00
-7.77612209e-01 4.20031846e-01 2.98353642e-01 4.95121032e-01
7.18248129e-01 2.59812246e-03 9.25506353e-01 1.02045298e+00
3.61946583e-01 -1.23615563e+00 3.60532433e-01 6.77689075e-01
6.05928004e-01 -1.23958099e+00 7.58533105e-02 -1.82230234e-01
-1.11461926e+00 8.02578866e-01 7.41332829e-01 2.86433041e-01
2.65239805e-01 1.25067979e-01 2.43847057e-01 -1.11426689e-01
-1.55929959e+00 -1.36594415e-01 2.54581533e-02 2.91187197e-01
2.89280266e-01 3.70376483e-02 -2.15395913e-01 3.35947394e-01
-4.11747485e-01 -4.39132690e-01 5.19341528e-01 8.49816024e-01
-3.88766468e-01 -1.42360628e+00 9.68996510e-02 6.70775056e-01
-1.21366508e-01 -1.42392769e-01 -6.82301879e-01 4.11335617e-01
-5.29230535e-01 1.23544586e+00 -2.86137825e-03 -7.85693049e-01
5.51410913e-01 3.02036315e-01 2.63904184e-01 -7.32967198e-01
-9.13392007e-01 -1.68147743e-01 5.78392327e-01 -4.33436632e-01
7.49609992e-02 -8.99245977e-01 -1.11983788e+00 -4.91448820e-01
-3.45608741e-01 2.17034712e-01 6.17761552e-01 9.44984615e-01
8.43489051e-01 6.00075305e-01 8.90473127e-01 -2.10605308e-01
-1.36908901e+00 -1.01300406e+00 5.05700782e-02 2.88779974e-01
2.23650143e-01 -5.45238614e-01 -2.78243423e-01 -4.34158713e-01] | [12.868898391723633, 7.983265399932861] |
799aa1ec-2323-4c65-a6d9-27e84ca89afe | high-fidelity-direct-contrast-synthesis-from | 2212.10817 | null | https://arxiv.org/abs/2212.10817v1 | https://arxiv.org/pdf/2212.10817v1.pdf | High-fidelity Direct Contrast Synthesis from Magnetic Resonance Fingerprinting | Magnetic Resonance Fingerprinting (MRF) is an efficient quantitative MRI technique that can extract important tissue and system parameters such as T1, T2, B0, and B1 from a single scan. This property also makes it attractive for retrospectively synthesizing contrast-weighted images. In general, contrast-weighted images like T1-weighted, T2-weighted, etc., can be synthesized directly from parameter maps through spin-dynamics simulation (i.e., Bloch or Extended Phase Graph models). However, these approaches often exhibit artifacts due to imperfections in the mapping, the sequence modeling, and the data acquisition. Here we propose a supervised learning-based method that directly synthesizes contrast-weighted images from the MRF data without going through the quantitative mapping and spin-dynamics simulation. To implement our direct contrast synthesis (DCS) method, we deploy a conditional Generative Adversarial Network (GAN) framework and propose a multi-branch U-Net as the generator. The input MRF data are used to directly synthesize T1-weighted, T2-weighted, and fluid-attenuated inversion recovery (FLAIR) images through supervised training on paired MRF and target spin echo-based contrast-weighted scans. In-vivo experiments demonstrate excellent image quality compared to simulation-based contrast synthesis and previous DCS methods, both visually as well as by quantitative metrics. We also demonstrate cases where our trained model is able to mitigate in-flow and spiral off-resonance artifacts that are typically seen in MRF reconstructions and thus more faithfully represent conventional spin echo-based contrast-weighted images. | ['Michael Lustig', 'Stella X. Yu', 'Jonathan I. Tamir', 'Fei Tan', 'Ekin Karasan', 'Thomas Amthor', 'Jakob Meineke', 'Mariya Doneva', 'Ke Wang'] | 2022-12-21 | null | null | null | null | ['magnetic-resonance-fingerprinting'] | ['medical'] | [ 6.38308406e-01 1.03960186e-01 1.10372856e-01 -1.77438408e-01
-7.86054373e-01 -5.59039354e-01 5.75059056e-01 -2.60038465e-01
-3.21494460e-01 8.75960231e-01 2.52039731e-03 -3.24977964e-01
-1.88915730e-01 -6.74728453e-01 -7.86145926e-01 -8.54482710e-01
-4.82528627e-01 6.03421926e-01 4.37542528e-01 -1.26264289e-01
4.12647612e-02 4.69389081e-01 -8.37359667e-01 -1.77374128e-02
1.08941329e+00 6.56120241e-01 3.47734541e-01 7.15811610e-01
2.97951102e-01 1.09023213e+00 -3.10465187e-01 -6.31405413e-02
1.66356415e-01 -8.54606688e-01 -8.09571445e-01 -1.61101803e-01
3.81227583e-01 -5.14406741e-01 -5.79123735e-01 1.16893983e+00
5.96835613e-01 2.19270453e-01 6.35157406e-01 -1.03903127e+00
-3.32561702e-01 7.11254060e-01 -4.78767931e-01 3.85836333e-01
3.59221734e-03 4.94775176e-01 -1.50405541e-02 -6.87720537e-01
9.50077951e-01 5.76731086e-01 6.85063958e-01 5.87492704e-01
-1.48280358e+00 -4.69424963e-01 -4.37898159e-01 2.67815422e-02
-9.84214544e-01 -6.79552555e-02 8.32079530e-01 -6.16546333e-01
3.80192995e-01 2.98008293e-01 7.04451501e-01 8.65825057e-01
7.51629174e-01 2.87556261e-01 1.87149167e+00 -3.58557343e-01
1.44580618e-01 -4.59587842e-01 -1.79309577e-01 9.57776725e-01
6.34776056e-02 4.27399904e-01 -7.15695992e-02 -8.14199522e-02
1.04478014e+00 2.55328286e-02 -8.40479434e-01 -5.99861443e-01
-1.56098676e+00 4.71594751e-01 6.12666488e-01 4.73730266e-01
-6.75233841e-01 2.51294345e-01 1.66308239e-01 1.51258081e-01
7.91378617e-02 5.09310246e-01 3.84823591e-01 1.72830045e-01
-1.20049047e+00 9.99992043e-02 3.21187019e-01 5.00766456e-01
4.37866151e-01 3.39015543e-01 -3.59547704e-01 5.25712788e-01
-6.79479539e-03 9.38566327e-01 7.84975946e-01 -1.01053715e+00
-1.73783630e-01 -1.94685534e-01 2.77653392e-02 -8.21001470e-01
-4.29910094e-01 -2.74798840e-01 -9.35545146e-01 3.02310497e-01
4.02073145e-01 2.38380246e-02 -1.32186902e+00 1.80260301e+00
3.72507811e-01 5.00573874e-01 -2.23856926e-01 1.42343462e+00
6.04194164e-01 3.47293198e-01 5.65251634e-02 -4.46745604e-01
9.86726105e-01 -9.45539653e-01 -8.56368423e-01 4.52136360e-02
2.69474059e-01 -5.11565864e-01 8.10365677e-01 2.91660801e-02
-1.63899767e+00 -2.71287501e-01 -1.28540027e+00 5.14801800e-01
1.59415603e-01 -6.49730504e-01 4.18927938e-01 5.52244604e-01
-1.01480448e+00 9.53678787e-01 -1.31469536e+00 4.07483548e-01
2.13654190e-02 3.60872030e-01 -4.71076339e-01 -3.61133814e-01
-1.46828485e+00 1.10956931e+00 1.14154406e-01 4.34720479e-02
-1.34092498e+00 -1.21272588e+00 -6.59382343e-01 -3.19509268e-01
3.46956283e-01 -6.60515189e-01 9.98550475e-01 -7.65971303e-01
-1.44671834e+00 6.37510717e-01 1.92646563e-01 -4.42778021e-01
6.23175740e-01 5.43305576e-01 -5.67290664e-01 7.91241288e-01
2.73115225e-02 5.86195111e-01 1.02749109e+00 -1.31739926e+00
2.33380243e-01 -2.43149132e-01 -1.16245970e-01 -6.13186136e-02
6.06993675e-01 -8.92360508e-02 1.70383304e-01 -6.36887789e-01
3.19046527e-01 -1.05451488e+00 -3.27101618e-01 -7.32540786e-02
-4.76452529e-01 9.49806690e-01 4.14068967e-01 -1.11638784e+00
7.79986620e-01 -1.70091188e+00 -2.39677099e-03 4.91608322e-01
5.10903537e-01 2.28072464e-01 -3.29271823e-01 3.93009279e-03
-5.56755483e-01 -8.39649513e-02 -7.94000328e-01 4.01017040e-01
-3.09668154e-01 -5.08819968e-02 -3.24840434e-02 6.01217747e-01
-8.29745755e-02 1.24496448e+00 -1.34463072e+00 -4.32240099e-01
2.60915041e-01 7.60389149e-01 -4.30702806e-01 2.43811712e-01
2.09296778e-01 1.23028529e+00 -2.05628380e-01 2.21397758e-01
7.94593275e-01 -8.53670314e-02 7.57129729e-01 -5.26012003e-01
-9.44985524e-02 1.14988513e-01 -6.37546480e-01 1.57651103e+00
-6.89660251e-01 2.13132426e-01 2.71601915e-01 -1.09207082e+00
6.10795856e-01 5.08760989e-01 8.83805096e-01 -1.18019557e+00
5.88866659e-02 5.47726214e-01 5.45959175e-01 -3.37294400e-01
7.68941492e-02 -7.32550442e-01 2.98231453e-01 7.96843469e-01
8.34191144e-02 -6.66424513e-01 1.04827777e-01 1.11794569e-01
1.18727064e+00 -2.64650211e-02 -2.86107600e-01 -6.05833828e-01
6.43642545e-01 -2.44603142e-01 2.55859315e-01 7.90973604e-01
-3.03284347e-01 9.29366291e-01 4.00494277e-01 -3.45055133e-01
-1.43379569e+00 -1.28025568e+00 -3.80071551e-01 4.28666845e-02
1.26375064e-01 8.24522451e-02 -9.30361927e-01 -7.33178973e-01
-3.62499237e-01 5.04329026e-01 -6.13720179e-01 -3.74768227e-01
-9.30855691e-01 -9.17496145e-01 3.72491986e-01 3.20490271e-01
3.40693474e-01 -1.03165233e+00 -5.77914357e-01 5.17744303e-01
-5.14701784e-01 -1.05414474e+00 -6.07066870e-01 1.09134875e-01
-1.00679696e+00 -1.02305329e+00 -1.27362609e+00 -4.23943669e-01
7.91102290e-01 3.14598307e-02 1.19263160e+00 1.49345636e-01
-2.74431825e-01 2.71978736e-01 -6.59269318e-02 1.04282402e-01
-8.42357934e-01 -3.11587244e-01 -9.21910163e-03 6.14696229e-03
-5.93806863e-01 -9.66844141e-01 -1.07142138e+00 3.83052349e-01
-1.26150882e+00 2.47838542e-01 3.67617667e-01 1.19088900e+00
9.78591383e-01 -2.35043451e-01 5.11960149e-01 -8.56879354e-01
5.97560167e-01 -3.28819782e-01 -5.18479645e-01 4.03522134e-01
-7.46397257e-01 2.31237665e-01 6.53252304e-01 -5.37894905e-01
-7.74173439e-01 -3.70233804e-01 -9.85173434e-02 -4.79520053e-01
1.85615003e-01 3.37138146e-01 3.21567923e-01 -5.63795805e-01
6.66948736e-01 7.33126640e-01 4.38611090e-01 1.32444113e-01
2.72771325e-02 1.95550203e-01 8.73381257e-01 -6.81404710e-01
7.03145742e-01 6.09711349e-01 3.67218435e-01 -4.66278732e-01
-4.02590692e-01 2.64674366e-01 -4.45587665e-01 -4.94962990e-01
8.10239077e-01 -3.65542591e-01 -5.61759293e-01 6.66504204e-01
-6.43730938e-01 -7.31373489e-01 -3.93800825e-01 7.31509566e-01
-7.41337240e-01 4.51901048e-01 -8.54676783e-01 -2.17045575e-01
-4.42622989e-01 -1.74858880e+00 5.27677655e-01 1.05365917e-01
9.14041400e-02 -1.14684701e+00 1.18543752e-01 2.37401649e-01
1.08103335e+00 7.75285542e-01 1.13972330e+00 -4.76268008e-02
-5.64221084e-01 1.84616655e-01 7.24055097e-02 3.56582254e-01
2.13467535e-02 -6.35281026e-01 -4.56856281e-01 -3.66218954e-01
2.01208204e-01 -3.19264591e-01 4.60130215e-01 7.98289835e-01
1.03314507e+00 7.72068053e-02 7.91070089e-02 8.11506629e-01
1.34043419e+00 5.01809180e-01 8.37279618e-01 -9.21824872e-02
7.08759844e-01 3.20683092e-01 2.02446029e-01 -1.03490151e-01
9.92079228e-02 7.14611411e-01 1.44158378e-01 -1.88469529e-01
-5.74130118e-01 -4.54062164e-01 1.79794021e-02 9.38986361e-01
-2.86522239e-01 3.05003643e-01 -1.04122853e+00 3.85398060e-01
-1.26074898e+00 -9.88593578e-01 -2.02879965e-01 2.30049610e+00
8.64371598e-01 -1.13787979e-01 -1.43690050e-01 -3.80955450e-02
8.49941969e-01 1.12763636e-01 -7.21837163e-01 6.89623654e-02
1.30264536e-01 7.43052065e-01 6.16424382e-01 8.45838428e-01
-6.91786051e-01 3.44845802e-01 6.31493330e+00 3.46452057e-01
-1.80952048e+00 5.79105854e-01 6.39413178e-01 7.78413042e-02
-7.70486355e-01 -3.38089876e-02 2.57930100e-01 6.70220733e-01
1.06149900e+00 -7.19219446e-02 8.91200185e-01 1.53544292e-01
2.37036258e-01 -2.22841620e-01 -6.41975164e-01 8.20598125e-01
7.05837971e-03 -1.44879997e+00 -2.26969197e-01 -5.50293215e-02
8.07270348e-01 4.98544909e-02 7.36062750e-02 2.65290402e-02
6.14888221e-02 -1.09614062e+00 6.76406145e-01 7.42074728e-01
1.47225380e+00 -6.57711446e-01 4.51700687e-01 1.69150263e-01
-5.74711382e-01 5.35479963e-01 7.44594932e-02 8.07423234e-01
5.26693165e-01 7.96669126e-01 -6.93956435e-01 6.18127942e-01
2.93506980e-01 1.34983003e-01 -2.24229231e-01 8.08090329e-01
-3.02686960e-01 6.20192587e-01 -1.83012024e-01 6.00122392e-01
3.61626118e-01 -4.41990614e-01 5.64584672e-01 6.86308980e-01
2.74913877e-01 2.83857495e-01 2.03921702e-02 1.11415994e+00
9.53074098e-02 -2.78671324e-01 -3.39457929e-01 3.24197039e-02
7.04969838e-02 1.29093850e+00 -1.01411211e+00 -4.21984941e-01
-6.52807280e-02 8.52160394e-01 -1.84745863e-01 4.85384405e-01
-9.22395885e-01 -2.58808941e-01 -2.08596170e-01 6.97638750e-01
1.09854257e-02 -1.82172880e-01 1.74611703e-01 -1.32513332e+00
-7.77008459e-02 -1.00102484e+00 -1.98580295e-01 -9.74855483e-01
-9.40215230e-01 1.01093876e+00 2.18112200e-01 -9.19750571e-01
-4.68321562e-01 -1.24358207e-01 -5.81941485e-01 1.10141909e+00
-1.68900526e+00 -7.73964226e-01 -3.30067575e-01 6.23997629e-01
-3.62254769e-01 1.48293972e-01 4.97122973e-01 4.80951518e-01
7.47866333e-02 3.11894715e-01 6.17639273e-02 9.55953076e-02
6.41556680e-01 -1.28270304e+00 1.13076590e-01 8.68576169e-01
-3.77869606e-01 6.27312183e-01 5.40135741e-01 -7.31246471e-01
-1.23473418e+00 -7.50036478e-01 3.26347053e-01 5.43091595e-02
6.22541308e-01 -1.20791860e-01 -9.84626710e-01 5.02622902e-01
5.64057790e-02 8.11144650e-01 2.90940642e-01 -7.82655060e-01
1.63091794e-01 -5.36511950e-02 -1.70020401e+00 2.97891527e-01
6.78164124e-01 -5.26665270e-01 -2.44477600e-01 2.27637798e-01
4.16115761e-01 -7.75819778e-01 -1.25825012e+00 4.05219167e-01
6.50854349e-01 -1.00330770e+00 9.63821352e-01 -3.85470271e-01
6.54622257e-01 -3.28442872e-01 1.73825309e-01 -1.77380168e+00
-2.44413584e-01 -7.47230411e-01 -2.21843999e-02 5.20322084e-01
8.58279616e-02 -7.68756866e-01 5.64736784e-01 6.07793391e-01
-3.03709507e-01 -4.96521086e-01 -7.59378970e-01 -6.77129328e-01
3.94956261e-01 -1.02394067e-01 6.14305496e-01 1.19695067e+00
-3.21864814e-01 -6.21397421e-02 -3.13798547e-01 -5.64947631e-03
1.10503960e+00 1.83011219e-01 2.14600060e-02 -6.09443605e-01
-3.55526805e-01 -2.02553719e-01 -1.43148735e-01 -6.10930145e-01
1.99648738e-01 -1.30127239e+00 6.08930737e-02 -1.13811553e+00
1.80564553e-01 -8.95188093e-01 -4.56512123e-01 1.77894980e-01
-9.02927965e-02 4.98898357e-01 3.55381012e-01 3.03119540e-01
1.39848024e-01 2.76536465e-01 2.29464674e+00 -2.06598431e-01
2.11473167e-01 -3.82655650e-01 -1.47382841e-01 3.10923040e-01
5.02961338e-01 -5.97656369e-01 -5.46319187e-01 -7.15011656e-02
-1.68365046e-01 1.07492137e+00 6.00797355e-01 -1.08409894e+00
-1.40175121e-02 -6.16811141e-02 3.54079813e-01 -8.54158252e-02
-4.02830392e-02 -7.21657693e-01 5.33442020e-01 8.49970877e-01
-2.57586420e-01 2.84858141e-02 -2.30781555e-01 -5.80804534e-02
-4.13934946e-01 -3.20848644e-01 1.26795352e+00 -4.02774006e-01
-2.14811608e-01 5.77651799e-01 -3.43156755e-01 2.36712098e-01
6.85406923e-01 2.02200025e-01 -1.88308492e-01 -5.42901695e-01
-1.09788251e+00 -2.09363058e-01 3.52104157e-01 1.90747660e-02
7.13382125e-01 -1.32017648e+00 -7.24673629e-01 4.40689415e-01
-3.80925089e-01 -2.93914169e-01 8.31536353e-01 1.58393204e+00
-9.90149438e-01 1.94016680e-01 -6.63632333e-01 -6.28923893e-01
-3.61752361e-01 5.41121185e-01 1.07677603e+00 -7.64319539e-01
-7.89878666e-01 1.95212662e-01 1.11955032e-01 -7.04977393e-01
-4.82746929e-01 -1.68292835e-01 2.86543906e-01 -5.33666253e-01
6.11689270e-01 5.51870838e-02 2.92604476e-01 -5.93369246e-01
-3.34905088e-01 3.89667720e-01 1.90199167e-01 -4.31782246e-01
1.18387437e+00 3.72282863e-02 2.30763964e-02 8.60363618e-02
1.01684630e+00 -1.35768101e-01 -1.43261969e+00 -8.19098875e-02
-4.21768248e-01 -3.41702521e-01 5.52136660e-01 -1.15681171e+00
-1.60265732e+00 1.01040411e+00 8.29012632e-01 1.86313745e-02
9.55459654e-01 -4.59349453e-01 1.14023137e+00 -3.61305356e-01
5.95020354e-01 -4.27817136e-01 -3.72243114e-02 1.44482821e-01
8.32761049e-01 -8.67777646e-01 -2.94888943e-01 -2.77475119e-01
-5.53128958e-01 1.02529752e+00 3.31357926e-01 -2.70994455e-01
5.36092162e-01 7.34322488e-01 1.75915599e-01 -1.58215046e-01
-2.48321339e-01 3.28019083e-01 2.05047399e-01 8.29535067e-01
4.18810964e-01 1.55049235e-01 -3.47505987e-01 7.10581616e-02
1.47394557e-02 2.04154283e-01 7.26032853e-01 8.53768051e-01
1.55248255e-01 -1.15960944e+00 -3.92856061e-01 3.26742321e-01
-5.66139281e-01 -1.97827861e-01 3.62005025e-01 6.47992849e-01
-1.51878029e-01 2.71327138e-01 1.33289099e-01 -2.68138554e-02
1.80036232e-01 4.08720933e-02 1.16925406e+00 -1.44608662e-01
-6.78874493e-01 -4.97591458e-02 -2.72214741e-01 -6.84251428e-01
-5.48532844e-01 -4.51503992e-01 -1.50404632e+00 -3.37939024e-01
1.01132587e-01 2.55408645e-01 8.94657671e-01 7.20076621e-01
3.43674831e-02 7.52289951e-01 8.09234321e-01 -8.96903872e-01
-4.68159288e-01 -8.57664108e-01 -7.95557439e-01 7.19479382e-01
3.89252573e-01 -6.75708115e-01 -1.97655991e-01 -9.40474495e-02] | [13.648326873779297, -2.3429038524627686] |
b17e91ec-93cc-48de-a703-2e70f36a77af | action-graphs-weakly-supervised-action | 2002.01449 | null | https://arxiv.org/abs/2002.01449v1 | https://arxiv.org/pdf/2002.01449v1.pdf | Action Graphs: Weakly-supervised Action Localization with Graph Convolution Networks | We present a method for weakly-supervised action localization based on graph convolutions. In order to find and classify video time segments that correspond to relevant action classes, a system must be able to both identify discriminative time segments in each video, and identify the full extent of each action. Achieving this with weak video level labels requires the system to use similarity and dissimilarity between moments across videos in the training data to understand both how an action appears, as well as the sub-actions that comprise the action's full extent. However, current methods do not make explicit use of similarity between video moments to inform the localization and classification predictions. We present a novel method that uses graph convolutions to explicitly model similarity between video moments. Our method utilizes similarity graphs that encode appearance and motion, and pushes the state of the art on THUMOS '14, ActivityNet 1.2, and Charades for weakly supervised action localization. | ['Hedvig Kjellström', 'Yong Jae Lee', 'Maheen Rashid'] | 2020-02-04 | null | null | null | null | ['weakly-supervised-action-localization'] | ['computer-vision'] | [ 2.23626092e-01 -2.08494022e-01 -6.23486221e-01 -3.85480314e-01
-3.14951539e-01 -6.86782777e-01 5.62216043e-01 3.67483824e-01
-2.18426794e-01 2.55780578e-01 7.03761697e-01 6.29910082e-02
7.36040547e-02 -5.44716954e-01 -5.91949940e-01 -4.39197093e-01
-5.26764512e-01 1.41053334e-01 5.39862990e-01 5.91548942e-02
2.48554721e-01 5.31497777e-01 -1.35046875e+00 8.53436530e-01
4.67762984e-02 1.02168977e+00 1.66436777e-01 9.70312357e-01
-6.42648200e-03 1.52916920e+00 -2.98411787e-01 1.92168251e-01
1.91602502e-02 -7.21723318e-01 -9.77702498e-01 4.04512703e-01
8.84424508e-01 -4.83704835e-01 -8.51834357e-01 8.05478632e-01
-2.37511359e-02 5.13565361e-01 7.05051541e-01 -1.44020510e+00
-5.12048006e-01 4.61306661e-01 -3.32982123e-01 7.53071964e-01
7.12637305e-01 9.20433328e-02 1.04366553e+00 -4.65976089e-01
9.23974752e-01 9.57662761e-01 6.61294520e-01 4.58048582e-01
-9.63384628e-01 -3.79339486e-01 3.11969399e-01 5.91941357e-01
-1.21915948e+00 -4.46285367e-01 7.55491495e-01 -6.42431140e-01
1.40559983e+00 -1.09129421e-01 9.52896476e-01 9.78138924e-01
1.19416259e-01 9.62550342e-01 5.53782821e-01 -2.40294337e-01
2.25232303e-01 -4.95763481e-01 1.08441733e-01 1.05151415e+00
-4.27837014e-01 -3.53875518e-01 -7.51044333e-01 -8.78070667e-03
8.18650305e-01 2.30626449e-01 -2.26661429e-01 -5.15451372e-01
-1.42477345e+00 4.82981563e-01 3.52773726e-01 6.68101907e-01
-3.87651950e-01 7.44262695e-01 5.78244567e-01 1.48820594e-01
5.54185450e-01 2.57883906e-01 -3.52661997e-01 -6.41013920e-01
-1.13915658e+00 6.25620335e-02 8.07172358e-01 8.58375132e-01
8.67293894e-01 -1.82190493e-01 -3.30081254e-01 4.91115004e-01
5.49958237e-02 -2.09138803e-02 4.03991282e-01 -1.52090192e+00
2.43344679e-01 7.83044815e-01 -1.26906946e-01 -1.05202913e+00
-2.97158062e-01 -4.39965352e-03 -1.91452280e-01 4.59549241e-02
5.74412584e-01 2.48595521e-01 -9.85153556e-01 1.77311242e+00
1.70824304e-01 8.03285658e-01 -2.14910597e-01 7.16751814e-01
5.70852935e-01 5.33180296e-01 2.31154695e-01 -2.93063419e-03
9.87584352e-01 -8.55036259e-01 -4.20696139e-01 -1.38561696e-01
1.15687263e+00 -5.16318202e-01 8.83287013e-01 -7.49633834e-02
-9.91289854e-01 -7.57238925e-01 -8.15065801e-01 -8.59902278e-02
-3.87778908e-01 1.32403761e-01 8.66994560e-01 1.48934349e-01
-1.24833858e+00 9.73712802e-01 -1.22333801e+00 -8.13623011e-01
5.00092030e-01 2.25616574e-01 -6.77079737e-01 -1.03655316e-01
-8.87693584e-01 7.18618751e-01 3.49255383e-01 -3.89806807e-01
-1.14766419e+00 -3.76236260e-01 -1.14184737e+00 7.73962587e-02
2.08499819e-01 -2.05411628e-01 1.26533306e+00 -1.47050750e+00
-9.55719233e-01 7.12685049e-01 -2.18919337e-01 -7.71825671e-01
1.32814750e-01 -2.83073410e-02 -2.56794482e-01 6.67277098e-01
9.14855450e-02 8.80172253e-01 6.35170162e-01 -5.13565242e-01
-9.45928693e-01 -2.39646375e-01 6.42621756e-01 3.87786508e-01
-3.65127146e-01 5.80510385e-02 -6.44549787e-01 -4.67102796e-01
7.26700500e-02 -9.24558103e-01 2.17418354e-02 1.66870341e-01
1.11649756e-03 -4.30794507e-01 1.04895627e+00 -7.04694331e-01
1.03902340e+00 -2.38197637e+00 -3.45807858e-02 1.25811368e-01
2.47524589e-01 -9.25549567e-02 -2.97402740e-01 5.41066885e-01
-1.24059901e-01 -1.23529144e-01 2.44403556e-01 -2.04422325e-01
-4.37080152e-02 4.18640643e-01 -3.48610841e-02 8.67671132e-01
-1.95494574e-03 8.52659404e-01 -1.24832451e+00 -7.15640664e-01
3.93362880e-01 4.86137718e-01 -6.02615595e-01 4.50586304e-02
-3.66098046e-01 2.82880276e-01 -4.06424820e-01 7.43003130e-01
-1.81255683e-01 -3.74533743e-01 1.46345526e-01 -4.48518068e-01
-4.31330577e-02 2.88062543e-01 -9.50958431e-01 2.05356288e+00
-1.07573599e-01 1.00851274e+00 -2.61012673e-01 -1.32981884e+00
2.23900124e-01 4.27025616e-01 1.23488736e+00 -5.41507423e-01
-1.17675640e-01 -1.39659062e-01 -5.50344102e-02 -6.07711136e-01
2.99233109e-01 1.99791402e-01 6.78622723e-02 6.45199418e-01
5.08565784e-01 3.16353351e-01 7.37571299e-01 4.58524406e-01
1.74788547e+00 6.60460114e-01 1.83379427e-01 -1.45858321e-02
3.64929855e-01 1.30447611e-01 4.27592278e-01 6.51479185e-01
-4.43516403e-01 4.56923604e-01 5.87197721e-01 -5.56452274e-01
-8.07376325e-01 -9.04272437e-01 5.10505557e-01 1.48763680e+00
1.73925877e-01 -7.61890709e-01 -6.97279751e-01 -1.11672115e+00
-9.44422930e-02 3.04839015e-01 -8.93032253e-01 -2.99007714e-01
-5.04162967e-01 1.31078675e-01 5.67691088e-01 9.42575753e-01
4.97455478e-01 -1.12715244e+00 -5.11765242e-01 1.52556464e-01
-3.06800663e-01 -1.18754959e+00 -9.29005802e-01 2.08833545e-01
-8.64669204e-01 -1.42064762e+00 -4.28650320e-01 -9.16452408e-01
6.94640398e-01 3.79671603e-01 1.11097682e+00 -3.88416871e-02
-3.30464631e-01 1.01870394e+00 -5.55876255e-01 3.63155067e-01
-2.94578731e-01 -3.11764121e-01 1.52450316e-02 9.57626551e-02
6.10888004e-01 -6.96543157e-01 -6.26109481e-01 3.14278930e-01
-6.47751093e-01 -3.28115211e-03 4.07272249e-01 3.85691375e-01
7.95912862e-01 4.55928706e-02 -2.36883387e-01 -4.94473457e-01
1.34430021e-01 -4.19880927e-01 -1.24466233e-01 3.92719716e-01
-1.70614749e-01 6.92685097e-02 4.55072701e-01 -7.21966982e-01
-5.92919230e-01 5.66788971e-01 1.87679738e-01 -7.89199471e-01
-4.19500202e-01 4.21798527e-01 1.09340250e-01 -1.59018993e-01
6.70150101e-01 4.04153109e-01 -2.72827446e-01 -2.93768525e-01
5.38023531e-01 3.71850044e-01 6.26757801e-01 -3.90057623e-01
2.51541793e-01 8.16188097e-01 3.07101794e-02 -8.33227992e-01
-8.24577570e-01 -9.62411523e-01 -1.08503604e+00 -5.46475828e-01
1.24153924e+00 -9.20581102e-01 -6.41665399e-01 1.91348448e-01
-1.00632584e+00 -6.62722588e-01 -4.57580060e-01 7.41612077e-01
-9.34697211e-01 6.32707596e-01 -7.56555378e-01 -3.92809480e-01
1.80711687e-01 -8.17680120e-01 1.06265819e+00 -6.19700700e-02
-5.97570956e-01 -1.09852636e+00 2.09585100e-01 2.03496262e-01
5.13470955e-02 2.68611580e-01 6.03071749e-01 -5.99461138e-01
-4.86927152e-01 -3.40828180e-01 -1.02499574e-01 5.21202922e-01
2.57616222e-01 1.43272787e-01 -5.62344611e-01 -2.30711654e-01
-3.29482973e-01 -4.81988192e-01 8.37111950e-01 6.69745684e-01
1.15876329e+00 -2.30731025e-01 -4.16489482e-01 6.87557161e-01
9.79751766e-01 1.41353875e-01 5.39642513e-01 -1.20566092e-01
9.08908665e-01 4.33408588e-01 5.16728401e-01 3.61425817e-01
2.96156019e-01 6.28258348e-01 3.00283045e-01 -5.95249273e-02
-4.44050699e-01 -4.23169851e-01 7.10032701e-01 4.78098214e-01
-4.21189755e-01 -1.36091605e-01 -6.92974806e-01 5.93263507e-01
-2.05350137e+00 -1.56770802e+00 2.43386198e-02 1.91948378e+00
5.01013637e-01 2.41484433e-01 1.68236941e-01 -2.01510072e-01
6.31217599e-01 5.17498672e-01 -5.34742594e-01 -2.05154747e-01
2.44327664e-01 1.92186534e-01 6.42583430e-01 2.88443983e-01
-1.44843578e+00 1.02101994e+00 6.93913031e+00 6.36290491e-01
-9.55180764e-01 1.07676841e-01 3.67817879e-01 -1.77939788e-01
2.28334650e-01 2.75931358e-01 -3.93226177e-01 2.89410591e-01
9.89153028e-01 5.73938387e-03 5.77011943e-01 1.06285572e+00
3.08868736e-01 -2.27448374e-01 -1.55803454e+00 9.63261247e-01
3.77041727e-01 -1.41924787e+00 -1.73485830e-01 -3.43281366e-02
7.63833582e-01 2.51777470e-01 -3.54828954e-01 2.19776541e-01
2.85236776e-01 -9.09798384e-01 7.21040785e-01 6.45739377e-01
5.49302876e-01 -4.82070625e-01 2.70680845e-01 9.25330669e-02
-1.68594790e+00 7.42884679e-03 -1.62947088e-01 -4.05891180e-01
-2.73445882e-02 1.41379327e-01 -8.39055240e-01 -3.55207585e-02
5.46229899e-01 1.50543547e+00 -6.07849717e-01 8.37985456e-01
-3.19364250e-01 6.75888300e-01 -1.36389151e-01 2.99211163e-02
4.59876567e-01 1.16904005e-01 1.47212386e-01 1.31892538e+00
2.13985220e-01 3.07912799e-03 7.70179927e-01 4.24480110e-01
-6.40040785e-02 -1.52734056e-01 -7.04420269e-01 -5.96600711e-01
1.84587743e-02 1.05592275e+00 -9.65669751e-01 -6.64044023e-01
-7.36141324e-01 1.32205415e+00 2.70573944e-01 4.54100013e-01
-8.93522441e-01 -2.27026522e-01 1.03560328e+00 2.38395110e-01
1.96341887e-01 -6.72607064e-01 3.31837386e-01 -1.23516643e+00
-1.48313940e-01 -6.09391510e-01 6.32445455e-01 -9.25693333e-01
-9.03610289e-01 1.82715595e-01 -5.55353463e-02 -1.44832122e+00
-5.94304979e-01 -5.70640385e-01 -5.79697967e-01 4.38824326e-01
-8.14717472e-01 -1.29481733e+00 -3.32422525e-01 9.63005126e-01
7.45046496e-01 8.70911181e-02 7.12744296e-01 1.35199606e-01
3.81837599e-02 1.17632404e-01 -1.58756226e-01 5.12226164e-01
5.61151206e-01 -1.32177258e+00 3.33626628e-01 8.13483775e-01
9.05174315e-01 4.48841631e-01 5.08020878e-01 -7.50194073e-01
-1.35605502e+00 -9.43635523e-01 8.85188103e-01 -6.87312901e-01
9.00821865e-01 -2.92014748e-01 -6.57720149e-01 9.71646786e-01
-6.13904372e-03 4.28705454e-01 6.93844020e-01 3.00880894e-02
-5.29832721e-01 -7.65523091e-02 -7.24310100e-01 5.15618563e-01
1.52575946e+00 -1.17046344e+00 -5.46581566e-01 6.82096243e-01
3.83457154e-01 -2.15026349e-01 -6.96465850e-01 1.63033098e-01
6.02634966e-01 -8.49029601e-01 9.16540921e-01 -8.38369489e-01
3.92494708e-01 -3.90775830e-01 -1.93735927e-01 -9.46652949e-01
-5.30021846e-01 -3.53265136e-01 -4.45855170e-01 7.97859251e-01
1.29062861e-01 8.45441446e-02 1.08331978e+00 4.53959733e-01
-2.80581027e-01 -4.98672068e-01 -7.84475029e-01 -7.75695145e-01
-5.21973550e-01 -6.97394252e-01 9.27076414e-02 1.07074547e+00
4.29723173e-01 -1.89666846e-03 -2.38571420e-01 1.04220338e-01
3.28922004e-01 2.08214909e-01 6.04345202e-01 -7.47250974e-01
-3.86129647e-01 -4.09079939e-01 -1.12303758e+00 -1.13976789e+00
4.20466870e-01 -9.79373872e-01 2.12331012e-01 -1.79229391e+00
2.62812704e-01 1.85668394e-01 -6.83166683e-01 8.74396384e-01
3.75350863e-01 4.94588792e-01 4.63072360e-02 2.91075170e-01
-1.21412838e+00 4.92904410e-02 8.37056041e-01 -2.32243255e-01
-7.37537816e-02 -1.47065267e-01 -1.48481727e-01 9.08116162e-01
6.13549292e-01 -2.63498724e-01 -6.88278854e-01 -3.72697979e-01
-5.16443923e-02 1.43319443e-01 4.93956774e-01 -1.45781934e+00
4.52142298e-01 -3.37370336e-01 6.16072595e-01 -4.78588134e-01
4.25860614e-01 -7.59972513e-01 3.03700734e-02 4.38428313e-01
-7.85115421e-01 4.98857237e-02 -1.57284945e-01 8.07274282e-01
-3.45513225e-01 1.69588551e-01 6.00787580e-01 -4.32642609e-01
-1.39964461e+00 6.15836740e-01 -5.01413226e-01 -8.63765627e-02
1.09524822e+00 -3.29381347e-01 -3.40655185e-02 -6.10997558e-01
-1.08569551e+00 -1.71945971e-02 7.68626690e-01 4.55677658e-01
6.19156599e-01 -1.42115092e+00 -2.03144163e-01 -7.93251991e-02
3.14694822e-01 -8.19821656e-01 2.37877756e-01 9.06513572e-01
-4.51343566e-01 1.86524749e-01 -3.06347728e-01 -4.66946095e-01
-1.45725107e+00 5.02962768e-01 3.08158785e-01 -1.57622918e-01
-6.50838315e-01 8.58384252e-01 3.02939594e-01 2.93072999e-01
4.54642236e-01 -5.65083981e-01 -2.12807834e-01 -7.19406083e-02
5.77142715e-01 1.79615274e-01 -2.61360377e-01 -9.72690046e-01
-6.00183070e-01 3.35243881e-01 2.63432264e-01 -9.18415934e-02
1.18437290e+00 -1.20916376e-02 -9.67312977e-02 5.28206468e-01
1.48319125e+00 -4.39539850e-01 -1.72599447e+00 -1.45439178e-01
7.99697042e-02 -5.55145204e-01 9.31926146e-02 -5.56874573e-01
-1.11212718e+00 9.39665675e-01 5.76397240e-01 1.65475637e-01
9.96935427e-01 3.31661701e-01 8.15830350e-01 3.82068276e-01
3.25241268e-01 -1.31713164e+00 5.72362483e-01 5.80705285e-01
4.70339715e-01 -8.49757612e-01 -7.59954751e-02 -1.39534511e-02
-5.99842727e-01 1.29423046e+00 5.72396755e-01 -3.43358040e-01
5.97220659e-01 5.05258469e-03 -2.35762730e-01 -5.39128661e-01
-6.43142462e-01 -5.25401890e-01 5.44370651e-01 7.03737795e-01
6.35953605e-01 -6.74685910e-02 -2.51209021e-01 -1.09636791e-01
4.55795199e-01 9.27268192e-02 2.13267371e-01 1.22080159e+00
-5.73848069e-01 -9.10103440e-01 1.87847644e-01 5.76286674e-01
-5.48930049e-01 1.18149992e-03 -6.55574560e-01 6.47216916e-01
3.98603857e-01 7.95079231e-01 3.25770080e-01 -6.86308801e-01
7.26092532e-02 2.17430606e-01 6.67494059e-01 -9.16248918e-01
-2.63897687e-01 9.78884473e-03 1.50941253e-01 -1.28003526e+00
-9.87306058e-01 -7.96418607e-01 -1.50753438e+00 -9.59363356e-02
1.28826678e-01 -6.92300173e-03 5.61617017e-01 1.13464403e+00
3.22058946e-01 3.05093139e-01 3.67720813e-01 -9.32822704e-01
-5.01426533e-02 -8.90822351e-01 -8.30571473e-01 8.59991431e-01
2.00076640e-01 -5.52610278e-01 -2.25430474e-01 6.95005536e-01] | [8.56916618347168, 0.7304630279541016] |
2d05b2e8-a7a5-4f79-8d4d-da35a82490e8 | an-original-framework-for-wheat-head | 2009.11977 | null | https://arxiv.org/abs/2009.11977v1 | https://arxiv.org/pdf/2009.11977v1.pdf | An original framework for Wheat Head Detection using Deep, Semi-supervised and Ensemble Learning within Global Wheat Head Detection (GWHD) Dataset | In this paper, we propose an original object detection methodology applied to Global Wheat Head Detection (GWHD) Dataset. We have been through two major architectures of object detection which are FasterRCNN and EfficientDet, in order to design a novel and robust wheat head detection model. We emphasize on optimizing the performance of our proposed final architectures. Furthermore, we have been through an extensive exploratory data analysis and adapted best data augmentation techniques to our context. We use semi supervised learning to boost previous supervised models of object detection. Moreover, we put much effort on ensemble to achieve higher performance. Finally we use specific post-processing techniques to optimize our wheat head detection results. Our results have been submitted to solve a research challenge launched on the GWHD Dataset which is led by nine research institutes from seven countries. Our proposed method was ranked within the top 6% in the above mentioned challenge. | ['Rabah Attia', 'Wided Souidene', 'Fares Fourati'] | 2020-09-24 | null | null | null | null | ['head-detection'] | ['computer-vision'] | [ 1.97802201e-01 1.78865552e-01 2.63611913e-01 -3.01346898e-01
-5.14270961e-01 -3.55490535e-01 4.32246655e-01 2.86378115e-01
-3.85402948e-01 4.40433145e-01 -6.69980794e-02 -2.05248833e-01
4.55943011e-02 -9.29928720e-01 -7.21441984e-01 -4.49496388e-01
-9.32284519e-02 4.07603353e-01 5.47600806e-01 -4.17043239e-01
2.96308815e-01 5.54382563e-01 -1.71761620e+00 2.44749770e-01
7.83655167e-01 1.00868559e+00 7.30815172e-01 1.00628710e+00
1.40886933e-01 5.32653749e-01 -4.47225302e-01 -2.46542588e-01
2.63454050e-01 -1.35286912e-01 -7.45361865e-01 3.35729599e-01
2.24193797e-01 -1.72742710e-01 3.94304186e-01 9.04514372e-01
7.41461813e-01 -1.33934990e-01 6.94126666e-01 -1.06552196e+00
-5.74809611e-01 7.70695090e-01 -8.63779306e-01 1.42804280e-01
8.90847370e-02 5.31645529e-02 9.73274410e-01 -1.02252197e+00
4.90895778e-01 1.04234338e+00 8.65325391e-01 3.20373178e-02
-1.03019524e+00 -4.81281549e-01 3.96484822e-01 3.68980438e-01
-1.34101629e+00 -4.00079072e-01 5.10146201e-01 -2.25094557e-01
7.91095376e-01 5.58497488e-01 6.65759921e-01 7.23746479e-01
5.52278059e-03 1.19145668e+00 1.10770857e+00 -8.69770288e-01
1.81392521e-01 2.21483767e-01 5.41346192e-01 7.81556666e-01
4.59537774e-01 3.15112434e-02 -3.90743762e-02 5.14990650e-02
4.50261503e-01 -1.46051094e-01 -6.10442609e-02 -3.31355214e-01
-8.24487507e-01 9.16894555e-01 6.72232389e-01 3.65950495e-01
-6.66536272e-01 -3.80301684e-01 7.21614808e-02 -1.01468734e-01
3.92537206e-01 4.44921046e-01 -6.91241205e-01 5.28411627e-01
-8.58213603e-01 5.40849827e-02 9.75284994e-01 8.81971180e-01
4.50032979e-01 -1.71471894e-01 -1.87652066e-01 8.07631910e-01
4.87573445e-01 4.66136664e-01 1.60116896e-01 -5.39637148e-01
4.35593612e-02 8.58863294e-01 2.72021413e-01 -1.00578654e+00
-8.20462167e-01 -6.42848074e-01 -6.25067532e-01 2.60488480e-01
2.62816161e-01 -2.45450109e-01 -1.15072227e+00 1.23159707e+00
2.06144959e-01 3.45955640e-02 5.66436797e-02 8.28049898e-01
9.66301143e-01 6.41672194e-01 2.60380477e-01 -7.87252411e-02
1.80872071e+00 -9.91420388e-01 -6.68831170e-01 -2.02753887e-01
8.86416495e-01 -1.22135532e+00 8.40530336e-01 6.88358307e-01
-8.30953836e-01 -7.10430086e-01 -1.46982574e+00 2.54494607e-01
-7.17751920e-01 8.45117390e-01 5.30420840e-01 6.08976245e-01
-7.82762468e-01 2.17883393e-01 -8.65433753e-01 -1.07148254e+00
2.15593562e-01 4.08637434e-01 -1.52239695e-01 1.22727029e-01
-7.42203534e-01 1.17296219e+00 8.19677770e-01 4.75372106e-01
-8.94931972e-01 -2.76833087e-01 -4.95236397e-01 -4.99884002e-02
6.24465227e-01 -3.62995684e-01 1.08075833e+00 -5.27044058e-01
-1.27980053e+00 1.05824149e+00 1.97573751e-01 -7.72082925e-01
3.80286217e-01 -5.73059797e-01 -4.81070608e-01 -4.09890175e-01
-8.66012052e-02 6.91789746e-01 3.25382590e-01 -1.28130794e+00
-1.13901591e+00 -5.15349329e-01 -2.67482102e-01 -2.63545979e-02
-4.45905894e-01 2.46375605e-01 -4.33246791e-01 -6.75589561e-01
2.92855918e-01 -9.69300508e-01 -2.62056917e-01 -3.40029687e-01
-6.32554114e-01 -9.35983211e-02 1.02443993e+00 -7.81517029e-01
9.80533421e-01 -1.74620855e+00 1.13085747e-01 -8.25403258e-02
-1.22954533e-01 7.26339579e-01 -1.37554944e-01 2.57552892e-01
2.32533626e-02 -2.44118825e-01 -1.91012517e-01 -3.46534252e-01
-8.07818770e-02 1.08955733e-01 -3.27029914e-01 4.44300771e-01
5.54987848e-01 7.73733079e-01 -3.69925052e-01 -4.67890352e-01
4.47976559e-01 3.09412032e-01 -2.23218381e-01 4.60272014e-01
-2.41238013e-01 -1.94608122e-01 -3.86892110e-01 1.03664005e+00
1.22716272e+00 1.02355480e-01 -1.52527392e-02 -5.34173071e-01
-5.62489212e-01 -3.51873606e-01 -1.38157046e+00 1.31843758e+00
-2.00811684e-01 1.91968396e-01 1.70021504e-01 -8.75038505e-01
1.09149802e+00 -2.38787588e-02 2.82466620e-01 -3.96269590e-01
4.81946856e-01 1.49351656e-01 -6.43943250e-02 -3.49342734e-01
5.80997944e-01 6.51154935e-01 1.43898308e-01 7.85819441e-02
1.68399200e-01 1.72253385e-01 4.44929630e-01 6.41582385e-02
8.18283617e-01 4.87292796e-01 4.49397922e-01 -6.02770984e-01
4.70038801e-01 4.71190631e-01 3.51971120e-01 9.11924899e-01
-3.66640687e-01 7.04290390e-01 2.91677266e-01 -5.09748995e-01
-7.05693245e-01 -7.12197542e-01 -1.99120894e-01 1.56029856e+00
-6.83073252e-02 -2.11028486e-01 -7.53477156e-01 -7.07641482e-01
-2.63879925e-01 5.90770960e-01 -7.21580744e-01 1.29288072e-02
-6.18614912e-01 -1.59588826e+00 5.95555723e-01 6.79439604e-01
5.63990176e-01 -1.32584679e+00 -8.31030905e-01 4.74744886e-01
1.73810616e-01 -1.26141942e+00 1.21307634e-01 9.28883374e-01
-5.36760569e-01 -1.12854564e+00 -9.59358215e-01 -1.05852282e+00
3.61787230e-01 3.15678090e-01 1.02317894e+00 -1.77250013e-01
-9.35474932e-01 -1.48940414e-01 -7.87109017e-01 -1.31842887e+00
-2.08377659e-01 5.57980955e-01 -3.06706965e-01 -1.67454153e-01
4.42251384e-01 5.22920825e-02 -4.73489970e-01 4.38740432e-01
-7.43847072e-01 4.98046726e-02 8.73567402e-01 4.40013289e-01
4.05833840e-01 -1.94175482e-01 3.55315000e-01 -8.85403097e-01
2.21457392e-01 -2.64948487e-01 -1.07564723e+00 4.76531029e-01
-4.88600373e-01 -2.50550747e-01 1.43006191e-01 -2.69004852e-01
-1.09672785e+00 5.83359480e-01 -2.93643445e-01 2.25508556e-01
-3.57356668e-01 1.56474903e-01 -3.22498202e-01 -2.05109417e-01
6.99844897e-01 -8.21365416e-02 -2.33662948e-01 -6.27847552e-01
4.66602981e-01 6.21197939e-01 4.56006318e-01 -1.00449353e-01
6.59673631e-01 3.95172954e-01 -1.16675548e-01 -8.09315324e-01
-1.01176023e+00 -6.08112454e-01 -8.31235707e-01 5.92618808e-02
9.86122310e-01 -9.48898852e-01 -5.58973610e-01 9.20628369e-01
-1.02391672e+00 -9.21909288e-02 2.03405485e-01 2.60919392e-01
6.19359873e-03 2.25003287e-01 -4.91287827e-01 -1.09667158e+00
-6.47227705e-01 -7.82416940e-01 1.03319180e+00 4.72608596e-01
1.85916707e-01 -5.84965587e-01 -2.51355171e-02 -3.05107348e-02
4.11454380e-01 2.04996198e-01 2.79446393e-01 -6.83238447e-01
-4.12143499e-01 -4.87193882e-01 -6.19796693e-01 9.03831795e-02
1.19029827e-01 3.46338421e-01 -1.11026978e+00 -3.92219275e-01
-1.56351089e-01 -1.86875433e-01 1.26102829e+00 6.69178307e-01
8.60624909e-01 3.80232781e-01 -6.09746397e-01 5.52467167e-01
1.38385153e+00 1.39242247e-01 3.41634393e-01 6.68188095e-01
5.72208464e-01 7.17960060e-01 1.22495687e+00 5.20412862e-01
4.41370696e-01 8.04967344e-01 7.47334421e-01 -5.98898530e-01
-2.38473326e-01 1.83196962e-01 2.75812358e-01 4.57017809e-01
-2.09933236e-01 -4.76738065e-01 -8.20197999e-01 5.83559215e-01
-2.13447595e+00 -4.21668321e-01 -6.08219922e-01 1.77625883e+00
2.29607806e-01 3.82165343e-01 4.03393567e-01 9.94835347e-02
6.31884158e-01 -1.36637285e-01 -6.90977275e-02 -2.35277787e-01
-3.80866230e-01 1.85291305e-01 7.37591922e-01 1.62832245e-01
-1.65473557e+00 1.26332152e+00 6.41869020e+00 4.39991117e-01
-9.01430845e-01 -4.78134304e-02 1.96759358e-01 2.90169835e-01
6.01518095e-01 -9.47630405e-02 -1.28229213e+00 8.33765790e-02
7.44906008e-01 3.06667984e-01 -1.99992791e-01 7.73185372e-01
-3.06496508e-02 -2.81348258e-01 -5.44704437e-01 2.80432224e-01
8.41565803e-02 -8.77727687e-01 -2.04410091e-01 -1.90288201e-02
3.89309973e-01 8.22430924e-02 -1.63100973e-01 4.98032570e-01
6.55957341e-01 -7.64295757e-01 4.91176158e-01 1.84598342e-01
-1.07237794e-01 -7.21987724e-01 1.07418096e+00 4.52881634e-01
-1.42942142e+00 -3.88587445e-01 -7.55162477e-01 5.17134890e-02
2.85856780e-02 6.43546999e-01 -1.12466991e+00 6.68968022e-01
8.33674669e-01 2.20960185e-01 -1.08804333e+00 1.32699656e+00
-6.68762624e-02 5.25864363e-01 -4.70378131e-01 -1.66117728e-01
2.46546835e-01 1.02136977e-01 3.53756577e-01 1.46748912e+00
2.81939030e-01 -1.92319974e-02 4.77831751e-01 5.69433153e-01
2.46266991e-01 2.70272166e-01 -2.88697511e-01 2.00406745e-01
1.02784984e-01 1.78557169e+00 -1.10929525e+00 -9.59410220e-02
-1.52207479e-01 7.32335687e-01 1.93773046e-01 -1.06633380e-01
-9.43417192e-01 -4.55468088e-01 9.40717906e-02 -1.91925302e-01
5.97046018e-01 -4.11581295e-03 -1.61893860e-01 -8.09600651e-01
-3.04769486e-01 -5.05369365e-01 6.22344077e-01 -8.12989652e-01
-8.82861733e-01 6.88281655e-01 -1.37686491e-01 -7.50555754e-01
1.12393744e-01 -9.69276071e-01 -5.49120843e-01 6.79814160e-01
-1.47200966e+00 -1.77808237e+00 -6.44722342e-01 2.36237615e-01
7.51428783e-01 -1.89750090e-01 1.01311481e+00 4.57438469e-01
-1.03063703e+00 4.30490106e-01 -1.71038404e-01 -1.20657077e-02
6.79569602e-01 -1.26937473e+00 6.32888556e-01 1.23250198e+00
2.28775024e-01 1.40638292e-01 8.78769934e-01 -5.80921292e-01
-1.46382523e+00 -1.19877481e+00 6.95739508e-01 -4.53486145e-01
5.86178422e-01 -3.97945672e-01 -9.43694174e-01 6.93567038e-01
3.09668988e-01 -1.49574935e-01 3.28004748e-01 1.02051377e-01
1.07779980e-01 -4.13011410e-04 -1.01188171e+00 2.15435892e-01
6.63285196e-01 2.38056317e-01 -5.24397910e-01 4.27471608e-01
5.59228659e-01 -1.68201819e-01 -4.25105274e-01 8.14197302e-01
4.17010665e-01 -6.86786056e-01 8.25883865e-01 -5.46211720e-01
2.57973000e-02 -5.16827524e-01 -3.42488617e-01 -1.17044210e+00
-5.54467261e-01 -1.73351541e-01 -1.85488284e-01 1.56166577e+00
2.91134804e-01 -3.01382631e-01 9.27078009e-01 -2.43383154e-01
-1.93105608e-01 -3.42138886e-01 -3.60746607e-02 -4.66921687e-01
-3.38412046e-01 -3.11986268e-01 4.47041154e-01 6.55967474e-01
-6.53806210e-01 2.09846020e-01 -4.42811608e-01 7.07864523e-01
7.91195631e-01 4.20940220e-01 7.69030809e-01 -1.42023885e+00
-7.39741549e-02 -2.98220694e-01 -2.38123834e-01 -5.45112371e-01
-4.16333228e-01 -5.92064917e-01 2.73249745e-01 -1.40602934e+00
4.09516424e-01 -7.99399987e-02 -6.61674321e-01 6.85557783e-01
-3.95691186e-01 6.60733223e-01 3.05046082e-01 -4.99481075e-02
-4.57133561e-01 7.81318452e-03 7.71911442e-01 3.29168513e-02
-2.73938537e-01 2.68479735e-01 -8.26517463e-01 7.75843203e-01
8.62731934e-01 -4.33810264e-01 1.09409027e-01 -4.10290241e-01
-3.52130383e-01 -5.62264264e-01 5.83181560e-01 -1.10203040e+00
1.43226489e-01 1.04273126e-01 7.48021364e-01 -1.05405664e+00
3.30675840e-01 -8.56351912e-01 -1.52510032e-01 7.95792341e-01
4.03013863e-02 -1.07771106e-01 3.41579318e-01 9.90340393e-03
1.02158070e-01 -2.06497386e-01 8.95047128e-01 -7.23777339e-02
-1.18423605e+00 -1.64180949e-01 -2.62948781e-01 -5.51348865e-01
1.23309565e+00 -4.21927013e-02 -2.10281417e-01 2.20493212e-01
-9.84857023e-01 4.22677904e-01 1.30727634e-01 6.69439375e-01
2.56951779e-01 -8.61490786e-01 -9.66829121e-01 1.37328491e-01
2.93516666e-01 -4.89770442e-01 -2.40786329e-01 8.38676512e-01
-6.78463101e-01 4.56550866e-01 -6.35402083e-01 -6.84531808e-01
-1.51916742e+00 4.84413117e-01 -3.47896554e-02 -3.40030491e-01
-4.33168054e-01 8.54814172e-01 -1.47967068e-02 -5.96007109e-01
5.50391793e-01 -2.56178170e-01 -7.99107909e-01 4.40776557e-01
5.07368565e-01 5.02956927e-01 4.29213703e-01 -5.12628019e-01
-5.78616858e-01 5.26382387e-01 -2.97256231e-01 3.65814894e-01
1.42181826e+00 1.24853827e-01 2.38913357e-01 2.93917283e-02
7.55716860e-01 -3.71493042e-01 -8.04781497e-01 1.05002835e-01
4.01666820e-01 -4.02322412e-02 3.28498840e-01 -9.59110796e-01
-8.85853231e-01 4.83428985e-01 1.49494863e+00 5.78728437e-01
1.21496534e+00 -1.36949122e-01 4.18066710e-01 4.87829417e-01
3.41015488e-01 -8.94333661e-01 -2.65204459e-01 5.71269155e-01
6.50319338e-01 -1.58060634e+00 1.22625567e-01 -6.73097551e-01
-4.01813000e-01 9.35970068e-01 8.04110289e-01 -2.80392528e-01
4.69193548e-01 7.30984688e-01 2.55858779e-01 -1.77244008e-01
-5.66054285e-01 -8.67598712e-01 3.45979422e-01 6.51917696e-01
6.01751089e-01 1.25066072e-01 -4.56963360e-01 5.68631530e-01
8.12486932e-02 2.55684078e-01 2.52328128e-01 1.23268127e+00
-9.24281180e-01 -1.03396559e+00 -7.05282211e-01 3.20280433e-01
-3.57887238e-01 1.76509470e-01 -6.56475842e-01 1.06170034e+00
2.70614892e-01 9.70450819e-01 -3.09891731e-01 -2.68652707e-01
7.93503702e-01 -2.11248264e-01 4.40010041e-01 -4.51110929e-01
-7.19644785e-01 1.03267461e-01 5.08777462e-02 -3.01417261e-01
-2.60532618e-01 -4.44111824e-01 -9.59309995e-01 -4.66585793e-02
-8.32210362e-01 -9.80547741e-02 9.01126504e-01 6.78415239e-01
1.72523156e-01 7.71864951e-01 2.93320626e-01 -1.04855990e+00
-4.23175424e-01 -1.25410891e+00 -5.37994385e-01 -9.01743025e-02
-5.76491132e-02 -5.38725972e-01 3.02047562e-02 4.67393287e-02] | [9.018762588500977, -1.1284117698669434] |
27725fdb-35df-45cc-ad92-6c8952500d2f | balanced-adversarial-training-balancing | null | null | https://openreview.net/forum?id=jB4K33NvZsd | https://openreview.net/pdf?id=jB4K33NvZsd | Balanced Adversarial Training: Balancing Tradeoffs Between Oversensitivity and Undersensitivity in NLP Models | Traditional (\emph{oversensitive}) adversarial examples involve finding a small perturbation that does not change an input's true label but confuses the classifier into outputting a different prediction. \emph{Undersensitive} adversarial examples are the opposite---the adversary's goal is to find a small perturbation that changes the true label of an input while preserving the classifier's prediction. Adversarial training and certified robust training have shown some effectiveness in improving the robustness of machine learnt models to oversensitive adversarial examples. However, recent work has shown that using these techniques to improve robustness for image classifiers may make a model more vulnerable to undersensitive adversarial examples. We demonstrate the same phenomenon applies to NLP models, showing that training methods that improve robustness to synonym-based attacks (oversensitive adversarial examples) tend to increase a model's vulnerability to antonym-based attacks (undersensitive adversarial examples) for both natural language inference and paraphrase identification tasks. To counter this phenomenon, we introduce \textit{Balanced Adversarial Training} which incorporates contrastive learning to increase robustness against both over- and undersensitive adversarial examples. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['paraphrase-identification'] | ['natural-language-processing'] | [ 6.11116946e-01 2.95857161e-01 6.75369874e-02 -4.73065525e-01
-1.08056772e+00 -1.44426441e+00 7.17786729e-01 1.54482886e-01
-3.59648943e-01 6.56038582e-01 6.55913055e-02 -5.40463626e-01
1.08073756e-01 -9.20665145e-01 -8.86230767e-01 -6.99462354e-01
2.29570210e-01 3.78437489e-01 1.03686132e-01 -5.18675745e-01
3.66626710e-01 8.58062506e-01 -1.06232429e+00 7.92082429e-01
5.57608128e-01 4.00147706e-01 -5.85895956e-01 7.69323289e-01
1.45703584e-01 1.16331446e+00 -1.25054550e+00 -9.73258674e-01
7.40497112e-01 -3.68904501e-01 -1.06670487e+00 -4.83892739e-01
1.06892014e+00 -1.82520822e-02 -2.92508870e-01 1.55638444e+00
5.89757442e-01 2.51487549e-02 7.26758182e-01 -1.42314029e+00
-1.09198022e+00 8.35720301e-01 -1.91499144e-01 4.91695225e-01
6.02299035e-01 6.25830054e-01 9.40290451e-01 -4.00971800e-01
5.45362830e-01 1.49191952e+00 7.73870766e-01 1.00787127e+00
-1.43253124e+00 -1.00727642e+00 -2.93923519e-03 4.71892692e-02
-1.12481499e+00 -3.47964436e-01 8.00289273e-01 -2.05168292e-01
7.22691894e-01 8.91504884e-01 8.89360681e-02 1.61531878e+00
3.81037951e-01 3.55840832e-01 1.47816062e+00 -3.25066954e-01
9.05924067e-02 5.19812703e-01 -7.18509108e-02 3.11398953e-01
6.55665109e-03 5.93236208e-01 -2.14229338e-02 -5.05334914e-01
3.42759602e-02 -2.41283774e-01 -3.59598398e-01 -1.49707841e-02
-9.65999424e-01 9.55756605e-01 7.37828612e-01 3.45013469e-01
1.23727024e-02 3.99212819e-03 6.81985021e-01 9.24682260e-01
1.86103404e-01 1.67147493e+00 -4.60109800e-01 3.14848840e-01
-5.36500573e-01 5.06080747e-01 7.70746768e-01 7.18888760e-01
4.62746501e-01 2.38262534e-01 -3.54376644e-01 7.17192590e-01
-3.31550151e-01 7.81485736e-01 7.60224044e-01 -8.71075034e-01
6.91031039e-01 3.62583637e-01 -7.63155669e-02 -1.19287825e+00
-2.67720371e-01 -3.37082446e-01 -4.47645843e-01 4.93733823e-01
7.56484389e-01 -9.38073248e-02 -7.82676339e-01 2.21806622e+00
-5.73227182e-02 -1.76902711e-01 4.81248677e-01 6.88113153e-01
6.17721796e-01 6.06282592e-01 4.30896670e-01 3.38914432e-02
9.90322590e-01 -4.69382674e-01 -2.61587292e-01 -5.09258628e-01
8.12128544e-01 -8.31498861e-01 1.63873994e+00 2.08270445e-01
-1.02770507e+00 -4.59764957e-01 -1.26762784e+00 1.13839336e-01
-9.36335981e-01 -8.72379899e-01 2.48340800e-01 1.19145882e+00
-5.22346020e-01 7.08790302e-01 -1.48241416e-01 -1.18739963e-01
4.31117833e-01 5.54580927e-01 -6.25918925e-01 -1.41301394e-01
-1.76954699e+00 1.46443272e+00 5.31303883e-01 -5.19592881e-01
-7.20293224e-01 -1.00490069e+00 -8.04301441e-01 -1.10979915e-01
8.61922354e-02 -4.44747925e-01 1.17749298e+00 -1.71435583e+00
-9.89513218e-01 1.22789514e+00 4.33947980e-01 -6.89600766e-01
7.25125074e-01 5.44931665e-02 -7.71292806e-01 -8.04058388e-02
4.12177946e-03 5.60891747e-01 1.16556096e+00 -1.27389812e+00
6.12138137e-02 -4.31398153e-01 3.89391899e-01 1.82947204e-01
-5.90569437e-01 2.05985501e-01 4.48263019e-01 -1.20056677e+00
-3.44418228e-01 -1.10504532e+00 -3.11565995e-02 -2.55661458e-01
-7.89451659e-01 1.63451567e-01 9.12543476e-01 -4.23692822e-01
1.02791309e+00 -2.05477524e+00 -5.94315045e-02 4.25973952e-01
-7.86209181e-02 7.42240191e-01 -3.45572710e-01 3.31760049e-01
-7.98417926e-01 4.66647655e-01 -4.32477355e-01 2.47770131e-01
-1.61798354e-02 3.79695952e-01 -9.00135815e-01 4.55857486e-01
2.43937269e-01 8.70898485e-01 -9.57882404e-01 -2.55458713e-01
1.49792165e-01 3.99974324e-02 -5.43388307e-01 1.64073363e-01
-1.10130407e-01 1.49008065e-01 -1.38510466e-01 3.99013668e-01
5.80966353e-01 3.43457192e-01 -1.08572289e-01 -3.99213433e-02
6.97383761e-01 -1.86244268e-02 -8.35789144e-01 8.51839662e-01
-5.35662293e-01 6.49225593e-01 -4.24625009e-01 -9.01832342e-01
8.00126553e-01 1.75504863e-01 -1.36546612e-01 -5.53197384e-01
6.02391921e-02 2.71823913e-01 3.69167626e-01 -5.89397073e-01
3.21042299e-01 -7.85813272e-01 -4.10048991e-01 4.72002059e-01
-2.87136823e-01 -6.33936465e-01 -3.37768614e-01 4.95299846e-01
1.19938672e+00 -3.05622518e-01 2.73597449e-01 -1.84810132e-01
8.35517347e-01 2.69344226e-02 5.37167311e-01 1.13380134e+00
-3.69770944e-01 5.37885070e-01 4.62146670e-01 -4.73478049e-01
-1.24282372e+00 -1.49020553e+00 -6.55741096e-02 1.13842487e+00
-9.02750865e-02 -6.26312420e-02 -9.20064211e-01 -1.19981670e+00
4.69533890e-01 1.33332515e+00 -8.59658003e-01 -9.53213453e-01
-5.27426124e-01 -7.73128211e-01 1.39096916e+00 5.51397622e-01
3.60034794e-01 -1.24823701e+00 -1.45253867e-01 -1.77862450e-01
2.05759685e-02 -8.69019568e-01 -5.37505150e-01 2.41573557e-01
-5.14141798e-01 -1.10688770e+00 -2.21776843e-01 -4.90243286e-01
8.77242327e-01 -4.72095788e-01 1.20992374e+00 7.44325221e-02
-1.45218104e-01 1.44972444e-01 -2.33695194e-01 -6.38794661e-01
-1.37721419e+00 -1.61190212e-01 2.84466922e-01 -3.16347331e-01
4.18101072e-01 -4.54671323e-01 1.41935861e-02 4.53003615e-01
-1.34007335e+00 -6.01595461e-01 2.02616677e-01 8.73347998e-01
2.33660460e-01 -4.54963930e-02 8.73758912e-01 -1.57499838e+00
9.52364862e-01 -5.15951991e-01 -1.58690572e-01 3.27740073e-01
-5.78269243e-01 2.07889169e-01 1.58691990e+00 -1.08924699e+00
-7.52300382e-01 -1.18586995e-01 -3.37973803e-01 -5.14216840e-01
-1.97923273e-01 -6.27551973e-02 -2.99590170e-01 -4.60539430e-01
1.55576813e+00 1.16674872e-02 -1.80110484e-01 -6.49892017e-02
3.78039837e-01 4.76251304e-01 8.93747449e-01 -6.43222332e-01
1.44172740e+00 1.65212512e-01 -8.20394084e-02 -4.15756822e-01
-1.04040933e+00 1.18469857e-01 -2.83935040e-01 2.92284396e-02
4.11092639e-01 -5.36669135e-01 -3.59425783e-01 3.94320965e-01
-9.38490808e-01 -1.99648693e-01 -7.17201173e-01 3.76384184e-02
-6.36318147e-01 3.66252542e-01 -3.88698816e-01 -2.53544271e-01
-2.28117168e-01 -1.10157871e+00 4.49993432e-01 -1.33994788e-01
-7.52578855e-01 -1.06027114e+00 -2.76754554e-02 5.29260993e-01
4.37956929e-01 5.69642901e-01 1.27785230e+00 -1.64491308e+00
6.28980324e-02 -6.59617364e-01 1.32481351e-01 8.68177891e-01
1.30144149e-01 -1.02605790e-01 -1.27569377e+00 -3.56320858e-01
2.60308832e-01 -5.92960119e-01 4.79830563e-01 -3.59023392e-01
1.18788385e+00 -9.13199902e-01 -1.10038286e-02 6.08529747e-01
1.27382255e+00 1.12412371e-01 8.01932991e-01 4.61369336e-01
7.75041163e-01 5.74807405e-01 3.73674661e-01 -2.93887407e-01
-4.33836758e-01 6.65633023e-01 5.20661831e-01 -1.55700937e-01
3.28086056e-02 -3.53180915e-01 5.85430861e-01 -7.40811555e-03
5.02796769e-01 -2.52814293e-01 -8.30426991e-01 3.37470621e-01
-1.25534034e+00 -1.48330379e+00 1.36623740e-01 2.25811863e+00
1.33606982e+00 4.72928047e-01 4.02665930e-03 3.76394004e-01
7.64973283e-01 2.57130623e-01 -7.98650920e-01 -1.19702113e+00
-3.67301345e-01 3.65024120e-01 6.05931520e-01 7.02466965e-01
-1.11852026e+00 1.23997462e+00 6.33828259e+00 9.20188963e-01
-9.70036507e-01 3.83613408e-02 6.46782756e-01 -2.41970196e-01
-6.51861489e-01 -1.86224252e-01 -1.56504288e-01 5.46851218e-01
8.71337593e-01 -2.86726147e-01 4.77065623e-01 9.88222539e-01
-2.35274673e-01 4.09678042e-01 -1.44404960e+00 6.34121180e-01
3.18018049e-01 -1.23382974e+00 4.25392658e-01 -1.55532256e-01
8.80516529e-01 -2.21025646e-01 4.56531107e-01 4.81429309e-01
7.41071820e-01 -1.29138350e+00 5.37504971e-01 1.53485954e-01
7.16961145e-01 -1.09682071e+00 7.89250374e-01 3.33325386e-01
-4.03918922e-01 -3.41266334e-01 -3.56200129e-01 -1.07142203e-01
-4.44825381e-01 1.98773101e-01 -1.00347435e+00 -8.45635086e-02
4.00516093e-01 8.88753459e-02 -9.36572134e-01 2.61264861e-01
-5.26666582e-01 4.60940629e-01 -3.28263342e-02 1.02761373e-01
2.82790035e-01 2.80603051e-01 8.76831412e-01 1.23642814e+00
-4.18591708e-01 2.13641208e-02 4.91277725e-02 8.73404503e-01
-3.58103067e-01 1.55708596e-01 -1.33802843e+00 4.47029620e-02
6.64480746e-01 7.33806014e-01 -2.74338096e-01 -4.08131391e-01
6.34181648e-02 1.18811321e+00 3.09075326e-01 3.17825317e-01
-8.23845088e-01 -6.10276937e-01 7.92129159e-01 -9.75023806e-02
-3.32405090e-01 4.25849050e-01 -4.33011353e-01 -8.95685136e-01
-1.07533149e-01 -1.65884984e+00 6.39899135e-01 -8.72889996e-01
-1.61748421e+00 4.88349259e-01 -2.04436898e-01 -1.05630600e+00
-3.10699701e-01 -5.39710045e-01 -7.87341952e-01 1.09849000e+00
-9.02992785e-01 -1.11964023e+00 2.52923220e-01 9.47270036e-01
3.66960555e-01 -4.48596209e-01 1.08412313e+00 -2.62255520e-01
-2.22821593e-01 1.32278085e+00 -7.00068623e-02 5.13158500e-01
1.05220211e+00 -1.46747565e+00 5.62488198e-01 1.09063911e+00
1.80753931e-01 8.72222126e-01 1.01927400e+00 -5.53192198e-01
-9.68264520e-01 -1.17935550e+00 8.86461675e-01 -1.10731268e+00
8.77342463e-01 -1.29478052e-01 -1.07061100e+00 8.94133627e-01
-1.15523890e-01 8.44308292e-04 8.09980869e-01 -1.74434915e-01
-1.23970091e+00 -9.61985737e-02 -1.92945898e+00 9.57627952e-01
7.57364213e-01 -1.12296641e+00 -1.07448530e+00 6.75264776e-01
7.92871952e-01 -3.16657901e-01 -8.50398839e-01 3.51533592e-01
3.52394879e-01 -8.14427614e-01 1.30481505e+00 -1.47828662e+00
4.53552336e-01 -5.30957803e-02 -3.15998673e-01 -1.50475776e+00
-3.11890870e-01 -6.75831020e-01 2.68975973e-01 1.09724665e+00
5.75414538e-01 -8.93618941e-01 4.88397479e-01 7.41461217e-01
1.17303498e-01 -3.56494784e-01 -7.91483641e-01 -1.02228463e+00
7.68713117e-01 -4.48190928e-01 5.68057537e-01 1.50928867e+00
1.98837563e-01 7.26413727e-02 -1.52099103e-01 4.89443570e-01
3.07239145e-01 -2.50763208e-01 6.15038931e-01 -7.06264973e-01
-2.00774863e-01 -3.05577010e-01 -7.15467691e-01 4.50782701e-02
6.42541766e-01 -1.20330131e+00 -8.07636157e-02 -6.14035368e-01
-7.87729397e-02 -4.87492085e-01 -3.45425636e-01 7.21444011e-01
-5.48924625e-01 5.75002193e-01 4.71730322e-01 1.68579996e-01
1.58628840e-02 -2.01571345e-01 9.16931272e-01 -4.16726887e-01
1.91050753e-01 3.09462905e-01 -1.09135425e+00 8.06716144e-01
8.62198234e-01 -8.85385334e-01 -5.60917377e-01 -1.55428275e-01
4.04681802e-01 -4.06108379e-01 4.76943731e-01 -8.68917584e-01
-5.82937486e-02 -1.44988045e-01 5.99602222e-01 3.93799394e-01
1.63236067e-01 -8.98967981e-01 -1.93620939e-02 8.60786796e-01
-1.09166992e+00 4.74419117e-01 5.08256674e-01 4.34289604e-01
1.39167234e-01 -4.53240901e-01 1.28655267e+00 -2.23292693e-01
-4.52323884e-01 4.75737117e-02 -3.31341922e-01 6.19740844e-01
1.08950543e+00 -1.63839221e-01 -7.01027572e-01 -3.56230974e-01
-8.00396323e-01 -1.03698134e-01 7.55051374e-01 4.84316558e-01
4.73702252e-01 -1.17636335e+00 -7.79543698e-01 2.81439632e-01
1.63822800e-01 -6.42486870e-01 -8.14973861e-02 1.51748329e-01
-3.93762022e-01 -4.02842127e-02 -4.21784669e-01 -1.46099359e-01
-1.62834430e+00 1.18308413e+00 7.38043725e-01 -2.07151994e-01
-2.21700817e-01 1.19401026e+00 1.13884300e-01 -8.81190717e-01
1.10821620e-01 7.42939562e-02 1.67184353e-01 -1.46021113e-01
4.43699300e-01 2.12627798e-01 1.05200991e-01 -6.00636184e-01
-4.26496595e-01 3.03865314e-01 -4.94001687e-01 4.54400703e-02
7.28225350e-01 5.23571491e-01 3.56368497e-02 2.36548349e-01
1.48501945e+00 3.76600474e-01 -6.90062582e-01 -1.73659697e-01
-1.56317696e-01 -6.01747453e-01 -4.23144042e-01 -1.29875791e+00
-7.44938672e-01 7.31434286e-01 6.00096107e-01 3.37109149e-01
1.03212857e+00 -3.43031108e-01 4.93742436e-01 6.60123408e-01
3.86695862e-02 -9.33611572e-01 1.69796616e-01 2.55443931e-01
9.50444043e-01 -1.24691999e+00 1.08852185e-01 -3.03297669e-01
-1.00178659e+00 9.50180829e-01 7.95889616e-01 -2.16380134e-01
1.90381825e-01 2.85971940e-01 4.51659054e-01 -4.53196466e-02
-3.15686584e-01 4.67348665e-01 2.41220042e-01 9.06508029e-01
-2.70203035e-02 9.91829857e-02 8.87771398e-02 1.54065657e-02
-6.64492846e-01 -6.65151000e-01 5.54731488e-01 6.91426039e-01
1.74809359e-02 -1.12969661e+00 -7.88253009e-01 4.17334795e-01
-6.82830155e-01 -4.19368476e-01 -1.06081963e+00 9.50857997e-01
1.84201211e-01 8.42604935e-01 -8.87255967e-02 -6.19861662e-01
4.70952719e-01 3.99365604e-01 4.60438371e-01 -8.29670250e-01
-1.41660845e+00 -9.51021850e-01 1.30033150e-01 -4.82512921e-01
-1.33404089e-02 -5.37573576e-01 -1.10524440e+00 -5.30424476e-01
-1.31509691e-01 7.10746571e-02 3.60719383e-01 9.69815373e-01
6.51381686e-02 3.23249876e-01 9.36515987e-01 -2.44355246e-01
-1.32068741e+00 -7.09242463e-01 -1.80091724e-01 1.19488215e+00
3.60123128e-01 -1.66309863e-01 -8.20546806e-01 -5.84606677e-02] | [5.963688850402832, 8.080445289611816] |
b728edcb-0614-48c8-9838-4f6bb998656d | unicorn-a-unified-backdoor-trigger-inversion | 2304.02786 | null | https://arxiv.org/abs/2304.02786v1 | https://arxiv.org/pdf/2304.02786v1.pdf | UNICORN: A Unified Backdoor Trigger Inversion Framework | The backdoor attack, where the adversary uses inputs stamped with triggers (e.g., a patch) to activate pre-planted malicious behaviors, is a severe threat to Deep Neural Network (DNN) models. Trigger inversion is an effective way of identifying backdoor models and understanding embedded adversarial behaviors. A challenge of trigger inversion is that there are many ways of constructing the trigger. Existing methods cannot generalize to various types of triggers by making certain assumptions or attack-specific constraints. The fundamental reason is that existing work does not consider the trigger's design space in their formulation of the inversion problem. This work formally defines and analyzes the triggers injected in different spaces and the inversion problem. Then, it proposes a unified framework to invert backdoor triggers based on the formalization of triggers and the identified inner behaviors of backdoor models from our analysis. Our prototype UNICORN is general and effective in inverting backdoor triggers in DNNs. The code can be found at https://github.com/RU-System-Software-and-Security/UNICORN. | ['Shiqing Ma', 'Juan Zhai', 'Kai Mei', 'Zhenting Wang'] | 2023-04-05 | null | null | null | null | ['backdoor-attack'] | ['adversarial'] | [ 2.16678575e-01 -7.13270307e-02 -3.88334453e-01 -1.39662758e-01
-1.45396307e-01 -1.45846593e+00 6.56921506e-01 -6.20034039e-01
-1.41548142e-01 3.39728862e-01 -1.75536111e-01 -1.04871106e+00
5.12244068e-02 -9.37874496e-01 -1.18267131e+00 -6.48226976e-01
-5.48888631e-02 -1.80812761e-01 5.18701017e-01 -3.09527576e-01
-3.46547365e-02 5.44735253e-01 -1.31762028e+00 2.42855519e-01
2.49227807e-01 6.19804025e-01 -1.38140664e-01 6.20048165e-01
-5.78538887e-02 6.72259450e-01 -9.51615155e-01 -4.61286902e-01
6.50500655e-01 -2.27087274e-01 -6.15945101e-01 -8.53954613e-01
3.86435449e-01 -7.96471715e-01 -7.47991979e-01 1.22145855e+00
2.57465929e-01 -4.70499873e-01 -6.59572557e-02 -1.95623779e+00
-3.75716120e-01 1.38657308e+00 -5.89690134e-02 2.82150686e-01
-6.29216433e-02 2.03675255e-01 7.01274276e-01 -4.46885049e-01
4.89806503e-01 1.10067141e+00 5.25617242e-01 1.09320366e+00
-1.08289552e+00 -1.17224848e+00 3.60163271e-01 -2.85839159e-02
-1.14487863e+00 -3.05755258e-01 7.55540252e-01 -3.42598408e-01
9.18279231e-01 7.67627954e-01 5.45041621e-01 1.94459498e+00
3.77293348e-01 7.65811563e-01 8.75476003e-01 -1.68832600e-01
2.40902781e-01 -4.41810079e-02 5.10644197e-01 4.32418078e-01
5.16167343e-01 8.67899358e-01 -2.11869076e-01 -6.76052153e-01
6.78517699e-01 1.43631622e-01 -6.60763979e-02 1.58248439e-01
-8.29049826e-01 8.96067679e-01 1.50671422e-01 2.61533767e-01
2.03507334e-01 8.11339319e-01 6.28515303e-01 2.65968591e-01
-1.89596415e-01 3.30289871e-01 -4.93301213e-01 -1.00159422e-01
-6.20479941e-01 4.13676500e-01 9.56284463e-01 9.15947795e-01
6.20079577e-01 3.35771233e-01 7.59950429e-02 -7.90010691e-02
2.24533096e-01 5.08306324e-01 1.20128788e-01 -6.57387257e-01
1.67305604e-01 3.41752023e-01 -2.11842120e-01 -8.15896511e-01
-1.32537574e-01 -3.64965558e-01 -3.14249992e-01 8.70894343e-02
4.13888931e-01 -6.46827817e-01 -1.02415597e+00 2.28528070e+00
3.21120143e-01 9.21516299e-01 3.11337244e-02 6.37067854e-01
7.20317483e-01 4.48665977e-01 -2.56341547e-01 2.76956767e-01
1.37294853e+00 -6.01450145e-01 -7.42148221e-01 -2.12563738e-01
5.72127938e-01 -3.80338311e-01 9.90856469e-01 2.95865774e-01
-4.66572165e-01 7.65777612e-03 -1.24522376e+00 2.96191543e-01
-7.36560941e-01 -1.67409614e-01 8.38591278e-01 1.33763659e+00
-1.01825809e+00 1.53582558e-01 -8.31100106e-01 1.35078669e-01
2.57204950e-01 7.55816281e-01 -1.26750633e-01 3.70184779e-01
-1.78525543e+00 4.26832438e-01 5.16222835e-01 3.02760974e-02
-1.80930626e+00 -8.72785151e-01 -6.71346545e-01 -3.73933196e-01
7.47979641e-01 -2.83868879e-01 1.24021542e+00 -8.35526228e-01
-1.24600387e+00 6.03530467e-01 2.09412560e-01 -7.92145193e-01
4.88360822e-01 -3.68207663e-01 -5.23295045e-01 -1.18970141e-01
-1.88735411e-01 3.03111196e-01 9.65426087e-01 -1.23426795e+00
-2.31766179e-01 -1.81375593e-01 7.88842678e-01 -6.04194999e-01
-3.82911325e-01 4.17451411e-01 -3.71101320e-01 -8.14990342e-01
-3.98717254e-01 -1.12860870e+00 -3.18541557e-01 -3.83252442e-01
-1.02601266e+00 3.68495323e-02 1.38937557e+00 -1.59275964e-01
1.34223568e+00 -2.29121828e+00 -2.82501519e-01 5.13261020e-01
5.25880277e-01 4.40273553e-01 -1.40794486e-01 4.79057580e-01
-3.03127795e-01 6.09611154e-01 1.86391920e-02 1.45338830e-02
1.93223357e-01 5.77054381e-01 -1.35541129e+00 3.81522536e-01
-9.70916897e-02 7.95567691e-01 -8.23863506e-01 1.23581350e-01
5.03194379e-03 4.92134541e-01 -6.83186948e-01 1.16123386e-01
-4.25122499e-01 2.54881848e-02 -5.14867306e-01 6.44327819e-01
7.72999704e-01 2.54991561e-01 1.74843594e-01 -2.46653091e-02
-1.61993191e-01 2.63425529e-01 -1.26324868e+00 9.92114723e-01
1.50521528e-02 6.23377919e-01 -7.25677088e-02 -5.31269073e-01
7.37749100e-01 5.12693048e-01 1.79153129e-01 5.70174158e-02
5.12078285e-01 2.05550089e-01 5.14787808e-03 -2.64279395e-01
1.92354783e-01 3.66335839e-01 -3.54326010e-01 5.92207134e-01
-7.30286017e-02 4.99677539e-01 -1.48222014e-01 2.67263860e-01
1.55001247e+00 -3.20940107e-01 -7.18199164e-02 -2.89970160e-01
2.76459247e-01 -2.39134744e-01 9.01583314e-01 1.15294528e+00
-2.95125097e-01 7.31590763e-02 1.01352382e+00 -4.96846467e-01
-5.44790328e-01 -1.38071883e+00 2.71187327e-03 9.18195784e-01
1.44829005e-01 -6.04525030e-01 -1.16066170e+00 -8.20108950e-01
2.49955580e-02 7.75923908e-01 -8.97168517e-01 -5.72519660e-01
-4.17771608e-01 -4.06148225e-01 1.83813584e+00 5.72482288e-01
5.18933475e-01 -8.60305965e-01 -8.76480699e-01 -7.66711161e-02
-2.50767358e-03 -1.20141375e+00 -6.05859876e-01 4.27735597e-01
-6.52911544e-01 -1.16802168e+00 9.73234400e-02 -4.81478721e-01
7.54966140e-01 9.69045162e-02 7.81080544e-01 2.33775318e-01
-3.09663802e-01 2.86685199e-01 -9.47330594e-02 -8.13288629e-01
-4.78173107e-01 -1.03119485e-01 4.61234599e-01 -4.85762767e-02
5.19283056e-01 -5.76010227e-01 -2.67763466e-01 6.34092987e-01
-1.51619196e+00 -4.53449577e-01 1.97298244e-01 5.75185955e-01
2.49395534e-01 1.12570472e-01 3.58375698e-01 -1.02325082e+00
8.14802170e-01 -5.44140697e-01 -1.06733787e+00 2.28510648e-01
-1.90776363e-01 8.91251042e-02 7.00550497e-01 -9.17101800e-01
-3.54797810e-01 8.77717808e-02 -2.71707982e-01 -1.00503039e+00
-2.25168675e-01 1.05203465e-01 -5.51749825e-01 -2.17467681e-01
8.55267286e-01 8.74311253e-02 -1.64578915e-01 -1.04020722e-01
4.22584504e-01 1.55673906e-01 4.46906865e-01 -9.54335332e-01
1.16410363e+00 6.58128202e-01 -2.28279740e-01 -5.15552282e-01
-2.58592039e-01 1.94646776e-01 8.05295184e-02 -3.67325932e-01
6.46077514e-01 -6.21934295e-01 -9.83508646e-01 7.44197369e-01
-1.31106913e+00 -5.82204938e-01 -4.74823490e-02 3.18198726e-02
-7.14532137e-02 1.22535586e-01 -6.72995448e-01 -7.86333859e-01
-2.29826644e-01 -1.64583707e+00 5.76673508e-01 1.81622595e-01
-1.43799499e-01 -9.30723369e-01 -1.42652802e-02 -1.77139536e-01
3.27810496e-01 5.30188918e-01 7.82980800e-01 -9.22007501e-01
-7.31889963e-01 -1.57784119e-01 4.12814140e-01 2.66465902e-01
6.35819435e-02 4.54576790e-01 -1.36975014e+00 -1.73517019e-01
1.91268295e-01 -1.33060040e-02 6.93124473e-01 2.28044629e-01
1.36821103e+00 -7.75361061e-01 -5.71903527e-01 8.45518887e-01
1.35197556e+00 4.85593736e-01 7.18343377e-01 3.94435287e-01
5.98888993e-01 2.96452314e-01 3.40088899e-03 3.38382214e-01
3.68932225e-02 4.22813147e-01 1.08191502e+00 6.23689294e-02
6.67557657e-01 -4.85129356e-01 9.85824168e-01 3.62285562e-02
2.73830771e-01 -1.70250729e-01 -1.03916085e+00 4.22607422e-01
-1.55525482e+00 -1.00806069e+00 -3.41356099e-02 1.99518371e+00
9.18574691e-01 3.58751655e-01 2.41737917e-01 1.46392271e-01
8.14550281e-01 1.59190521e-01 -6.02954388e-01 -6.63253486e-01
-1.06972791e-01 1.29637375e-01 9.08917069e-01 6.83086991e-01
-1.04684520e+00 1.16773236e+00 6.45731497e+00 7.80966520e-01
-1.43615842e+00 1.75358146e-01 3.61866504e-01 -1.87801376e-01
-9.27868307e-01 3.61869454e-01 -1.30430186e+00 5.02781630e-01
9.49986935e-01 -2.11129874e-01 7.72036016e-01 7.59871125e-01
-3.50783989e-02 5.80100238e-01 -1.33699417e+00 6.18706882e-01
-1.59227908e-01 -1.59600055e+00 1.17069311e-01 3.28071654e-01
4.53431398e-01 1.42861813e-01 4.22426939e-01 1.37473911e-01
7.41320908e-01 -9.37754333e-01 1.15280986e+00 9.74988043e-02
7.61424184e-01 -8.29725564e-01 3.10144752e-01 2.57310331e-01
-9.41531837e-01 -4.10675853e-01 -1.49833247e-01 1.49648622e-01
-3.08693081e-01 4.34436381e-01 -8.10935199e-01 1.95967793e-01
7.94517994e-01 2.20064744e-01 -3.72102320e-01 3.90641659e-01
-5.85440636e-01 1.11349809e+00 -7.60507643e-01 -4.10040747e-03
2.90743411e-01 2.26553008e-01 8.93725276e-01 9.90599751e-01
1.11392505e-01 -1.50885195e-01 -9.23118815e-02 1.28815687e+00
-1.18959025e-01 -6.54767871e-01 -1.26201701e+00 -1.34213626e-01
8.99246573e-01 1.02998340e+00 -6.51748776e-01 9.99320671e-02
-1.49605453e-01 3.50963026e-01 -1.90931380e-01 6.17862105e-01
-1.50212765e+00 -3.27013016e-01 1.23451030e+00 1.75650015e-01
-6.29077777e-02 -3.70837897e-01 -3.44489157e-01 -9.10968304e-01
7.12855486e-03 -1.28042662e+00 2.55116045e-01 -2.29732990e-01
-8.96087229e-01 6.46437883e-01 4.22315568e-01 -9.37308311e-01
-1.95105299e-02 -7.44960606e-01 -1.04484653e+00 4.31420982e-01
-8.58344436e-01 -9.60123718e-01 1.25359833e-01 1.12442029e+00
-5.75312451e-02 -2.96710879e-01 8.84113491e-01 3.74122381e-01
-1.01839828e+00 1.28137136e+00 -2.11633757e-01 4.31017935e-01
4.00229394e-01 -9.11234140e-01 7.36906350e-01 1.53446639e+00
6.89882413e-02 1.34019721e+00 9.15440440e-01 -7.24637985e-01
-1.75780618e+00 -1.17712867e+00 1.97427735e-01 -6.19871259e-01
9.14296150e-01 -1.09899902e+00 -5.99854052e-01 1.10427296e+00
1.07098557e-01 -1.12854270e-02 8.95182133e-01 -2.49808371e-01
-8.20501685e-01 -1.70533791e-01 -1.18440223e+00 1.27206182e+00
1.11564994e+00 -7.22310781e-01 -1.79137141e-01 3.24651673e-02
1.22501504e+00 -5.69259048e-01 -3.82392615e-01 2.56265819e-01
5.24009407e-01 -6.79260373e-01 9.80548263e-01 -6.59038007e-01
4.07881327e-02 -5.30977428e-01 -3.34892362e-01 -6.71238124e-01
2.20123768e-01 -1.21308887e+00 -4.50052679e-01 1.23197997e+00
4.42491621e-01 -1.34263563e+00 7.24807680e-01 5.34575462e-01
-3.66269574e-02 -8.01558435e-01 -1.11341667e+00 -1.04189706e+00
1.28369331e-02 -9.97430742e-01 1.14808714e+00 9.93097246e-01
-4.01044399e-01 -2.48791367e-01 -3.11419874e-01 7.58792639e-01
7.76896894e-01 -5.57319999e-01 8.55459571e-01 -8.15574527e-01
-3.64350289e-01 -3.23966265e-01 -2.94389516e-01 -7.04663813e-01
3.62586707e-01 -9.56516445e-01 -2.72268414e-01 -6.96429908e-01
-3.92099172e-01 -4.74495590e-01 -4.65164930e-01 1.02327442e+00
2.75138348e-01 4.83138114e-03 2.95033932e-01 2.58649290e-02
-2.34710276e-02 9.21605304e-02 6.92694783e-01 -1.72497317e-01
-8.18184391e-02 -1.96826421e-02 -9.68038023e-01 7.42133796e-01
1.33728492e+00 -8.44397783e-01 -8.14150572e-01 -5.74259222e-01
5.62250197e-01 -4.03175950e-01 8.33225608e-01 -8.16873014e-01
2.83524811e-01 -2.80656099e-01 -2.38249972e-01 -4.66716170e-01
6.71675354e-02 -9.88492608e-01 4.60933298e-01 7.64892042e-01
-2.66670138e-01 3.18730384e-01 6.74034715e-01 4.21933949e-01
2.30835438e-01 -1.78130209e-01 2.63486415e-01 7.49908015e-02
-7.79932082e-01 3.52159768e-01 -6.62868738e-01 1.31007031e-01
1.32623529e+00 -3.10667902e-01 -7.84383655e-01 -8.83589089e-02
-4.16765630e-01 8.65984187e-02 3.68793160e-01 7.74731398e-01
8.31927061e-01 -1.23189521e+00 -1.96133509e-01 6.29948676e-01
-9.12446901e-02 -2.33146250e-01 -7.34121129e-02 4.55112427e-01
-5.54330170e-01 2.42429435e-01 -1.09048836e-01 -4.45622474e-01
-1.36499929e+00 5.23507178e-01 7.88378060e-01 -1.67070121e-01
-4.00398284e-01 1.04136086e+00 4.06822652e-01 -5.37581265e-01
4.00310189e-01 -5.30975044e-01 2.27950945e-01 -3.71364683e-01
6.23668313e-01 8.53709951e-02 -2.68562973e-01 -1.02268860e-01
-6.48211837e-01 2.41570488e-01 -2.60325909e-01 -2.33904228e-01
8.95003498e-01 6.56597853e-01 -3.38734746e-01 1.17146060e-01
9.69354808e-01 1.04846865e-01 -1.05219162e+00 -3.56430933e-02
-1.87913507e-01 -3.07600379e-01 -2.31904089e-01 -5.19752741e-01
-1.18509781e+00 6.35850489e-01 5.11482954e-01 1.52511194e-01
1.16771770e+00 -7.21164569e-02 8.16534102e-01 6.40744030e-01
6.59140766e-01 -7.01094866e-01 -1.13368109e-01 8.20006609e-01
7.74618924e-01 -2.82510459e-01 -6.70650303e-01 -4.08083230e-01
-4.68654595e-02 9.31516528e-01 8.80617321e-01 -2.95676321e-01
1.02408969e+00 1.16762924e+00 2.42622137e-01 -1.43486425e-01
-8.47442508e-01 3.75361979e-01 -3.86046708e-01 5.57115912e-01
-1.68361917e-01 2.67263323e-01 -5.21421693e-02 1.01791573e+00
-3.75383675e-01 -1.65802628e-01 7.42544889e-01 9.94667590e-01
4.64048684e-02 -1.48305559e+00 -6.55289829e-01 2.22690076e-01
-9.09664512e-01 -2.00934812e-01 -6.59979939e-01 7.23773539e-01
5.06352782e-01 9.98848021e-01 -2.18471274e-01 -1.24473214e+00
1.57990366e-01 -1.51887164e-01 1.51615128e-01 -5.24047971e-01
-9.75095034e-01 -3.00321788e-01 -3.15939449e-02 -7.70820439e-01
2.04730764e-01 -3.73748153e-01 -1.38428617e+00 -6.02989972e-01
-3.15071851e-01 3.52897458e-02 5.10269105e-01 6.34381592e-01
3.23363364e-01 6.05366707e-01 6.53759181e-01 -2.38428697e-01
-6.02261841e-01 -1.04290970e-01 -3.24271649e-01 6.77432399e-05
4.30142790e-01 -7.03735471e-01 -5.32817602e-01 -2.40390480e-01] | [5.676504135131836, 7.664111137390137] |
d094ca14-e329-4951-95cc-6dfa6002a916 | user-controllable-recommendation-against | 2204.13844 | null | https://arxiv.org/abs/2204.13844v1 | https://arxiv.org/pdf/2204.13844v1.pdf | User-controllable Recommendation Against Filter Bubbles | Recommender systems usually face the issue of filter bubbles: overrecommending homogeneous items based on user features and historical interactions. Filter bubbles will grow along the feedback loop and inadvertently narrow user interests. Existing work usually mitigates filter bubbles by incorporating objectives apart from accuracy such as diversity and fairness. However, they typically sacrifice accuracy, hurting model fidelity and user experience. Worse still, users have to passively accept the recommendation strategy and influence the system in an inefficient manner with high latency, e.g., keeping providing feedback (e.g., like and dislike) until the system recognizes the user intention. This work proposes a new recommender prototype called UserControllable Recommender System (UCRS), which enables users to actively control the mitigation of filter bubbles. Functionally, 1) UCRS can alert users if they are deeply stuck in filter bubbles. 2) UCRS supports four kinds of control commands for users to mitigate the bubbles at different granularities. 3) UCRS can respond to the controls and adjust the recommendations on the fly. The key to adjusting lies in blocking the effect of out-of-date user representations on recommendations, which contains historical information inconsistent with the control commands. As such, we develop a causality-enhanced User-Controllable Inference (UCI) framework, which can quickly revise the recommendations based on user controls in the inference stage and utilize counterfactual inference to mitigate the effect of out-of-date user representations. Experiments on three datasets validate that the UCI framework can effectively recommend more desired items based on user controls, showing promising performance w.r.t. both accuracy and diversity. | ['Tat-Seng Chua', 'Liqiang Nie', 'Fuli Feng', 'Wenjie Wang'] | 2022-04-29 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [-1.26502842e-01 -5.21203466e-02 -5.65313756e-01 -4.35959369e-01
-9.36717615e-02 -9.49052095e-01 3.98037106e-01 8.19991529e-02
-3.72232646e-01 4.90172923e-01 4.47728604e-01 -6.65245295e-01
-2.37111762e-01 -1.13622689e+00 -5.48361480e-01 -2.27355540e-01
1.39184684e-01 8.89960229e-02 2.21128181e-01 -3.19040120e-01
5.97195089e-01 1.37986258e-01 -1.81473124e+00 4.36633259e-01
1.25379181e+00 7.68009782e-01 8.25752094e-02 4.10241604e-01
2.04309225e-02 6.39833272e-01 -5.89189112e-01 -9.04520154e-02
2.81299800e-01 -2.43576646e-01 -2.83192426e-01 -3.07268530e-01
8.18803534e-02 -7.51021802e-01 -8.37011188e-02 9.92226124e-01
2.45390266e-01 3.77905279e-01 2.32887685e-01 -1.28833866e+00
-1.19252408e+00 9.85537767e-01 -1.62669316e-01 1.95439979e-01
5.73259890e-01 2.02212602e-01 1.14525616e+00 -7.16967642e-01
2.94230580e-01 1.21594334e+00 3.17734540e-01 5.45246303e-01
-1.27384448e+00 -9.46617007e-01 1.07533181e+00 -1.20943353e-01
-1.27201664e+00 -3.56470495e-01 3.26601565e-01 -5.11665165e-01
5.34083605e-01 1.19529605e+00 5.73617697e-01 1.02777874e+00
1.36532769e-01 6.71304107e-01 7.49173999e-01 -4.59115021e-02
5.54593742e-01 5.02383113e-01 4.19232547e-01 1.45804107e-01
6.02820456e-01 4.36258018e-01 -2.82663256e-01 -4.71863627e-01
7.15512991e-01 4.29655254e-01 -5.69785058e-01 2.19731182e-02
-8.43499184e-01 8.92369092e-01 4.23800677e-01 -1.84281857e-03
-5.62143445e-01 -4.24643874e-01 2.40491703e-02 4.07711238e-01
2.92724997e-01 9.39034283e-01 -7.14041173e-01 -5.72502352e-02
-6.15007699e-01 3.66722345e-01 7.50950575e-01 8.03211153e-01
4.94319350e-01 -1.61642238e-01 -7.58079112e-01 6.00839674e-01
4.77537692e-01 4.69395339e-01 5.79118013e-01 -6.76407516e-01
-1.24659808e-02 7.17112005e-01 6.91219687e-01 -1.09024334e+00
-1.42511025e-01 -5.83810925e-01 -4.28642631e-01 1.40844241e-01
2.58471459e-01 -2.85283983e-01 -6.11495495e-01 1.83640683e+00
3.86921853e-01 9.65345129e-02 -3.25406879e-01 1.18085575e+00
3.70139182e-01 6.03613317e-01 1.31138757e-01 -5.30280828e-01
1.23302305e+00 -4.73565757e-01 -8.08780134e-01 1.95374456e-03
4.78383541e-01 -5.90755522e-01 1.48367667e+00 6.19559586e-01
-7.02452123e-01 -4.57663894e-01 -8.78619194e-01 4.78050292e-01
-8.79597366e-02 -3.48157547e-02 5.41705847e-01 6.73345625e-01
-6.06913388e-01 6.96143150e-01 -6.10453725e-01 -6.93362653e-02
-6.23415411e-02 3.58523011e-01 3.39256346e-01 3.77242088e-01
-1.49715507e+00 4.93092030e-01 -9.66178079e-04 3.35774757e-02
-4.88852501e-01 -1.15559447e+00 -4.34847623e-01 3.79378676e-01
7.49369740e-01 -5.58984518e-01 1.43409777e+00 -9.96945381e-01
-1.57961214e+00 -1.75851032e-01 2.86884069e-01 -2.06455097e-01
7.22496510e-01 -5.03771007e-01 -7.00785935e-01 -7.05190122e-01
-7.37749189e-02 1.30077824e-01 8.14145088e-01 -1.15728855e+00
-9.12717879e-01 -3.31399381e-01 5.00993729e-01 2.56199598e-01
-5.46527445e-01 -2.47086033e-01 -2.59105533e-01 -6.38763607e-01
-2.33249575e-01 -1.03585994e+00 -3.42385888e-01 -3.87170941e-01
-3.42684358e-01 -1.71637386e-01 6.98172450e-01 -1.01822473e-01
2.16043282e+00 -1.93518651e+00 -4.73279536e-01 3.87138218e-01
8.27323124e-02 3.50795239e-01 -4.06233743e-02 2.61844158e-01
1.03146784e-01 5.74854493e-01 5.75014412e-01 2.77471721e-01
1.67889595e-02 1.05640218e-01 -6.18077159e-01 1.93513602e-01
-3.84736210e-01 4.76777911e-01 -8.72369826e-01 2.46885151e-01
2.23908573e-01 8.60811844e-02 -1.21788895e+00 5.00463247e-01
-5.49122334e-01 3.06280404e-01 -7.20088542e-01 2.66703576e-01
6.21397674e-01 -3.30592841e-01 3.18542719e-01 -1.90653160e-01
-7.52040088e-01 6.32051408e-01 -1.43125379e+00 8.41542125e-01
-5.71313977e-01 -1.31510735e-01 4.81009297e-02 -3.66717994e-01
6.76743209e-01 1.31070092e-01 2.12028578e-01 -7.17728674e-01
-3.24263558e-04 -1.36894912e-01 7.95482248e-02 -3.28415751e-01
5.34480572e-01 5.18381238e-01 1.64681375e-02 7.15167820e-01
-6.42528355e-01 7.00960875e-01 -2.63570417e-02 3.80634159e-01
8.20051074e-01 -9.39251576e-03 4.44782883e-01 -4.38087225e-01
4.66028512e-01 -3.51765722e-01 7.71577954e-01 1.22498381e+00
9.20108333e-02 3.31006758e-02 2.74115145e-01 -2.66556919e-01
-3.84465337e-01 -8.16451311e-01 1.07320502e-01 1.75563383e+00
2.85981327e-01 -7.20296443e-01 -3.33921194e-01 -8.25486898e-01
3.60829651e-01 1.36183488e+00 -7.85009146e-01 -4.78219599e-01
-9.42027047e-02 -3.63096267e-01 -2.15547755e-02 3.27115506e-01
1.03254765e-01 -7.71376193e-01 -6.65691197e-01 3.64712030e-01
-8.24617594e-02 -2.82247484e-01 -1.08011699e+00 -3.51153761e-01
-6.21588826e-01 -9.34642315e-01 -6.06944077e-02 2.65909046e-01
7.39485741e-01 7.61020601e-01 8.19948494e-01 4.66599464e-01
3.18562418e-01 2.39440367e-01 -4.48407352e-01 -2.99725920e-01
-2.47262210e-01 -1.81116298e-01 2.95212120e-01 1.53564677e-01
2.24664524e-01 -4.89380062e-01 -1.07169807e+00 8.94792199e-01
-8.23051572e-01 -3.64709795e-02 2.31309146e-01 5.67625463e-01
3.29173625e-01 -4.64206338e-02 8.74766469e-01 -1.33880877e+00
1.17130458e+00 -8.23387444e-01 -8.72949302e-01 2.50272155e-01
-1.54964733e+00 9.16592479e-02 1.01635075e+00 -9.29827571e-01
-1.13561213e+00 -3.79637033e-01 1.48829952e-01 -3.02757263e-01
5.13068326e-02 5.46212971e-01 -2.82753021e-01 4.15581882e-01
8.42828095e-01 -1.80531532e-01 -3.18993390e-01 -4.80886638e-01
6.35118961e-01 9.52286303e-01 1.63839713e-01 -4.35263157e-01
4.55220491e-01 1.79511026e-01 -1.03970957e+00 -2.52764344e-01
-9.99098718e-01 -3.91917944e-01 1.04879573e-01 -4.01733220e-01
2.24841908e-01 -7.29721367e-01 -1.14393640e+00 -1.79030687e-01
-6.08906388e-01 -8.61071199e-02 -3.92373651e-02 4.56763893e-01
7.30499998e-02 -1.60117773e-03 -3.40358585e-01 -1.05632377e+00
-4.45709765e-01 -9.81706619e-01 3.14358503e-01 6.11743331e-01
-6.00903928e-01 -6.06227040e-01 -6.37463406e-02 7.97487572e-02
6.68414831e-01 -2.68699020e-01 6.81330919e-01 -9.89734232e-01
-5.86720467e-01 -2.62378752e-01 8.48938152e-02 -8.04601312e-02
4.06814158e-01 2.81681627e-01 -8.13354313e-01 -4.56977248e-01
-1.94186315e-01 3.94170195e-01 3.67859304e-01 2.93930113e-01
1.24941444e+00 -1.12262988e+00 -3.25734407e-01 1.87253430e-01
1.02764857e+00 5.54579735e-01 4.10624534e-01 2.09610358e-01
3.74438465e-01 3.53715241e-01 9.13322151e-01 9.15773928e-01
2.98366994e-01 6.87640190e-01 4.61473376e-01 1.78960532e-01
4.15090680e-01 -6.74680293e-01 5.47040462e-01 3.56241792e-01
3.06409728e-02 -4.53133941e-01 -1.59234673e-01 4.42816578e-02
-2.15713787e+00 -9.09107327e-01 -9.25795510e-02 2.79708862e+00
7.86866426e-01 4.17466313e-01 2.18879327e-01 -1.60237119e-01
6.77872777e-01 -1.50334567e-01 -9.48782980e-01 -5.72439730e-01
5.82338452e-01 -3.87069017e-01 5.03428102e-01 6.15409613e-01
-6.84278846e-01 6.82801604e-01 5.30701065e+00 2.91406125e-01
-1.22661269e+00 -1.27558649e-01 5.24966657e-01 -3.52947116e-01
-8.55174899e-01 1.02182515e-01 -8.42786252e-01 9.00248647e-01
9.01974499e-01 -4.79833782e-01 6.70072258e-01 9.30949867e-01
8.37083817e-01 -1.76339783e-02 -1.23225868e+00 5.12367129e-01
-3.50602388e-01 -1.37401450e+00 1.74452171e-01 2.20878467e-01
4.88789260e-01 -2.89967746e-01 1.43411487e-01 6.62865579e-01
6.30686522e-01 -5.88468075e-01 8.46372783e-01 6.70499682e-01
5.17624438e-01 -8.06719005e-01 3.37347120e-01 7.26898551e-01
-8.14447105e-01 -3.45680326e-01 -2.43197873e-01 -4.42176551e-01
-8.40541497e-02 4.68069464e-01 -7.46848166e-01 3.06571215e-01
5.09495378e-01 3.02936614e-01 -2.89300472e-01 7.34575927e-01
-3.37226808e-01 7.79905975e-01 -1.79181844e-01 -2.40296870e-01
-1.97936714e-01 -4.37330455e-01 4.39207882e-01 8.62029314e-01
3.80823225e-01 5.08279085e-01 4.53303546e-01 1.02084720e+00
3.12329773e-02 1.49792627e-01 -2.45707318e-01 -1.75285526e-02
1.06978929e+00 1.16028929e+00 -3.24292898e-01 -3.04553002e-01
-2.34449700e-01 5.41359186e-01 4.88964329e-03 4.45421219e-01
-8.78283083e-01 -1.47742284e-02 1.01269257e+00 5.67772806e-01
1.64420381e-01 3.22802186e-01 -3.07464451e-01 -1.18502545e+00
-3.09747428e-01 -9.41188812e-01 5.85293829e-01 -5.05578101e-01
-1.22034836e+00 2.19506860e-01 -2.60594755e-01 -1.08996356e+00
-8.20009336e-02 5.10117635e-02 -6.85176253e-01 7.10087776e-01
-1.05275142e+00 -3.53485346e-01 -6.86345175e-02 5.12472212e-01
4.60924298e-01 4.20234919e-01 8.22761297e-01 2.98676163e-01
-7.60908186e-01 9.83782172e-01 -3.08549069e-02 -4.77300227e-01
9.91130054e-01 -1.01059520e+00 1.12973973e-01 7.33878195e-01
-3.62751484e-02 1.37354982e+00 8.64529610e-01 -8.47959578e-01
-1.48199224e+00 -1.14964139e+00 5.73999345e-01 -4.52955008e-01
5.36322117e-01 -3.26707542e-01 -1.13878441e+00 6.63079202e-01
-2.97577798e-01 -2.56052792e-01 9.81732368e-01 5.98108530e-01
-4.57337797e-01 -1.36571348e-01 -1.07646561e+00 9.83482122e-01
8.85349452e-01 -1.49288431e-01 -4.70391184e-01 8.99944529e-02
8.65587771e-01 -3.59399229e-01 -7.76196480e-01 2.95521766e-01
9.25455451e-01 -9.64983463e-01 7.42647290e-01 -8.89586329e-01
-9.18546021e-02 -4.78765011e-01 1.18550971e-01 -1.38177598e+00
-8.00963461e-01 -9.49899435e-01 -2.52565503e-01 1.34914660e+00
5.80111563e-01 -7.16404498e-01 3.86842191e-01 1.41031897e+00
6.10392094e-02 -5.85357726e-01 -1.63668036e-01 -4.14157063e-01
-4.89848703e-01 -3.86861324e-01 1.06973886e+00 9.73764360e-01
2.05526516e-01 4.36456203e-01 -6.22392118e-01 5.35656154e-01
1.48584992e-01 4.66286808e-01 7.55180657e-01 -1.27812135e+00
-3.02446574e-01 -4.10256267e-01 4.86073077e-01 -1.30273092e+00
-4.11932588e-01 -5.76074243e-01 -1.27814963e-01 -1.12141705e+00
-6.75259456e-02 -6.76167011e-01 -7.70083129e-01 4.64604557e-01
-4.68325734e-01 -4.69067782e-01 2.16366932e-01 4.33508962e-01
-7.10735381e-01 1.83539703e-01 1.17146087e+00 2.75611848e-01
-8.08536947e-01 6.51551485e-01 -1.41474354e+00 6.86374426e-01
8.15796793e-01 -3.47501606e-01 -7.38344848e-01 5.27134351e-03
5.34279525e-01 7.75468647e-02 -7.49730095e-02 -2.85105586e-01
2.92646706e-01 -6.84065402e-01 2.71410197e-01 -4.16618407e-01
-3.39791447e-01 -7.81224787e-01 3.76655579e-01 3.58094901e-01
-7.27969348e-01 5.91575354e-02 -2.87325773e-02 8.03044558e-01
3.86349171e-01 1.45854533e-01 5.22578895e-01 2.28841659e-02
-3.46430838e-01 3.28398913e-01 -6.61286891e-01 -2.00866401e-01
8.43778431e-01 8.42612237e-03 -4.77375597e-01 -5.32339334e-01
-5.75759888e-01 6.66348875e-01 3.48321021e-01 8.93811643e-01
4.28639352e-01 -1.09287739e+00 -3.65899414e-01 5.15337646e-01
2.35523447e-01 -6.16268754e-01 2.57911950e-01 4.99966830e-01
3.07705998e-01 4.57518995e-01 1.61839977e-01 -2.84061581e-01
-1.14539981e+00 7.40333498e-01 2.30214238e-01 -8.66940692e-02
-2.90949762e-01 6.93451941e-01 3.19253951e-01 -3.48905563e-01
4.35156554e-01 -4.94885117e-01 -5.08730948e-01 2.69030273e-01
9.31195557e-01 3.18522781e-01 1.77681353e-02 1.35638922e-01
-2.90625423e-01 -3.62629861e-01 -5.60991049e-01 1.28198206e-01
9.67561305e-01 -4.73966211e-01 1.75264135e-01 2.53022552e-01
4.68650192e-01 5.36094427e-01 -1.29690707e+00 -1.66738361e-01
-4.21208367e-02 -1.08515191e+00 1.71596065e-01 -1.33216763e+00
-8.14629376e-01 7.52046406e-02 4.93433475e-01 6.95587039e-01
1.06442165e+00 -4.73218173e-01 5.52909613e-01 1.12637378e-01
3.13693255e-01 -1.10095048e+00 -2.79581606e-01 2.53720909e-01
7.10724235e-01 -1.01391542e+00 7.97151122e-03 -1.71286836e-01
-7.47908890e-01 7.54126847e-01 9.37451363e-01 1.39954276e-02
6.29352748e-01 1.54176429e-01 -1.17886057e-02 1.09620690e-01
-1.28313637e+00 1.44528031e-01 3.15401137e-01 1.69743627e-01
6.18597269e-01 5.52807033e-01 -6.88064575e-01 1.24864006e+00
-1.24199800e-02 1.05885960e-01 6.39020443e-01 4.78100985e-01
-6.16775632e-01 -1.07468343e+00 -3.92842472e-01 7.23276675e-01
-3.57667327e-01 -5.88260591e-02 -2.87471205e-01 5.27336955e-01
9.60534588e-02 1.36182773e+00 9.82409995e-03 -6.74739182e-01
7.99024761e-01 -3.30909997e-01 -2.19922140e-01 -6.74058914e-01
-9.01495099e-01 2.70047754e-01 1.40289083e-01 -7.71011353e-01
1.62093863e-01 -5.10477126e-01 -1.16893053e+00 -4.46458787e-01
-7.71888137e-01 5.21920204e-01 2.19945148e-01 6.68043196e-01
8.34983528e-01 5.76874673e-01 1.07200027e+00 -2.93445379e-01
-1.10983062e+00 -8.15547824e-01 -3.91203880e-01 4.06683654e-01
3.09502482e-01 -7.69488096e-01 -5.21896660e-01 -3.95402640e-01] | [9.919707298278809, 5.650774002075195] |
3a4dff04-75a2-43d0-8556-a374d5a62a6b | conner-a-cascade-count-and-measurement | null | null | https://aclanthology.org/2021.semeval-1.176 | https://aclanthology.org/2021.semeval-1.176.pdf | CONNER: A Cascade Count and Measurement Extraction Tool for Scientific Discourse | This paper presents our wining contribution to SemEval 2021 Task 8: MeasEval. The purpose of this task is identifying the counts and measurements from clinical scientific discourse, including quantities, entities, properties, qualifiers, units, modifiers, and their mutual relations. This task can be induced to a joint entity and relation extraction problem. Accordingly, we propose CONNER, a cascade count and measurement extraction tool that can identify entities and the corresponding relations in a two-step pipeline model. We provide a detailed description of the proposed model hereinafter. Furthermore, the impact of the essential modules and our in-process technical schemes are also investigated. | ['Yefeng Zheng', 'Xi Chen', 'Zhiyuan Qi', 'Yunyan Zhang', 'Yuejia Xiang', 'Jiarun Cao'] | 2021-08-01 | null | null | null | semeval-2021 | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [ 3.45280886e-01 8.00828338e-01 -3.62516671e-01 -2.44860753e-01
-7.29077697e-01 -6.45753622e-01 7.59642780e-01 1.06260753e+00
-4.52138364e-01 1.13247788e+00 6.29572868e-01 -5.55816531e-01
-4.56025690e-01 -5.35770535e-01 -3.49309325e-01 -1.75764456e-01
-1.00962557e-01 4.57103908e-01 -1.07583925e-01 1.00369819e-01
3.91968578e-01 2.71193624e-01 -9.00320768e-01 5.69329560e-01
8.23287427e-01 6.73746347e-01 -3.13448668e-01 7.47647107e-01
-2.10295051e-01 1.49345493e+00 -7.67300189e-01 -5.56266367e-01
-5.71178079e-01 -6.09357476e-01 -1.16125107e+00 -3.88205111e-01
-8.61858800e-02 3.94627452e-02 9.93154049e-02 1.06269455e+00
4.27345961e-01 -1.44465610e-01 1.02001631e+00 -1.11173320e+00
-2.65628844e-01 1.35332310e+00 -6.40797839e-02 2.97644317e-01
6.27432466e-01 -1.15692846e-01 1.22239816e+00 -8.53585780e-01
1.18864214e+00 1.03199863e+00 6.66317225e-01 4.95225579e-01
-1.01413596e+00 -4.83698934e-01 -2.67276704e-01 1.97864212e-02
-1.42615342e+00 -6.61392331e-01 1.53453007e-01 -5.74847937e-01
1.51372206e+00 6.63604498e-01 5.23558140e-01 8.51740479e-01
5.07173002e-01 5.13160825e-01 9.68441665e-01 -3.29402536e-01
3.84948909e-01 1.65314406e-01 4.47948426e-01 6.33405983e-01
6.30614221e-01 -4.96727914e-01 -6.08502150e-01 -3.89486402e-01
4.66877103e-01 -4.56359744e-01 -1.18420817e-01 3.57015282e-01
-1.33878553e+00 2.01542154e-01 1.72994472e-02 5.24718165e-01
-5.08951545e-01 5.61468117e-03 4.17629570e-01 3.46649587e-02
5.69756925e-01 1.08254814e+00 -8.15826774e-01 -1.68036297e-01
-8.44499648e-01 1.99929476e-01 1.22332835e+00 1.24013460e+00
1.61766052e-01 -9.29912865e-01 -6.19584143e-01 3.51693928e-01
3.73028040e-01 8.77114534e-02 2.89233327e-01 -7.31147528e-01
3.76428694e-01 1.10737514e+00 2.36728370e-01 -8.53545129e-01
-1.01477325e+00 -4.06068236e-01 -6.41905427e-01 -6.80663407e-01
-2.06367150e-02 -4.95106250e-01 -6.98993862e-01 1.34633219e+00
4.88238603e-01 4.41015095e-01 1.45170204e-02 4.19170797e-01
1.85448098e+00 2.56123506e-02 7.69287407e-01 -5.86881399e-01
1.95373058e+00 -5.89853823e-01 -1.58951926e+00 5.97063303e-02
1.05178308e+00 -9.73414063e-01 1.71947211e-01 -4.44573052e-02
-1.17938173e+00 6.17066845e-02 -9.15405154e-01 -4.52106982e-01
-6.57807052e-01 2.09820613e-01 5.74135482e-01 1.98293746e-01
-7.69798279e-01 8.21990490e-01 -8.93219769e-01 -1.83657944e-01
3.65066350e-01 2.53140420e-01 -3.59513938e-01 6.10678136e-01
-1.49413502e+00 1.39308882e+00 5.44256210e-01 1.19185627e-01
-1.16588689e-01 -1.22278321e+00 -1.10534441e+00 2.18767989e-02
3.57208997e-01 -9.80825067e-01 1.20617580e+00 1.62339434e-01
-1.08435488e+00 1.10312700e+00 -2.95262128e-01 -5.79826951e-01
2.78992623e-01 -1.33869544e-01 -6.21092618e-01 3.76720130e-02
2.13904485e-01 9.68028083e-02 -8.14261809e-02 -5.71499050e-01
-7.79748738e-01 -1.28092974e-01 -1.64346725e-01 2.11719945e-01
2.47072112e-02 4.97407377e-01 -5.19306004e-01 -3.58621389e-01
-1.99617408e-02 -4.25677329e-01 -3.86758596e-01 -4.93781239e-01
-1.12923229e+00 -4.52342123e-01 -8.11466724e-02 -7.41235435e-01
2.05836892e+00 -1.80688357e+00 2.01261550e-01 3.24033558e-01
9.96551514e-01 -2.55910933e-01 3.63838077e-01 6.25178874e-01
-3.29566360e-01 4.87919927e-01 -1.38544172e-01 -2.85940647e-01
4.94226217e-02 -1.39450088e-01 1.77852467e-01 4.02777761e-01
5.74349165e-01 1.27591062e+00 -1.18497348e+00 -1.04719377e+00
6.82269558e-02 3.98526251e-01 -9.44999307e-02 1.36775717e-01
-1.49278715e-01 3.40453476e-01 -5.41173220e-01 8.61422062e-01
2.12987527e-01 -5.56848764e-01 5.95865130e-01 -6.44214213e-01
-1.02473453e-01 9.83240604e-01 -1.13780761e+00 1.40810204e+00
-1.96752489e-01 3.41360897e-01 1.59426325e-03 -5.60848534e-01
5.28893292e-01 5.36043465e-01 9.54102516e-01 -2.28484616e-01
4.34396952e-01 3.18015546e-01 6.57158867e-02 -8.05575013e-01
5.61685264e-01 -5.99689558e-02 -2.78810829e-01 1.46309838e-01
1.46600962e-01 -1.44071519e-01 7.38361537e-01 5.27701378e-01
1.55555594e+00 2.35089138e-01 1.33393788e+00 -4.39233243e-01
8.45266521e-01 5.24698198e-02 6.28849447e-01 3.88815254e-01
6.47385120e-02 2.01899350e-01 1.12023878e+00 8.39208886e-02
-7.59256184e-01 -5.75526357e-01 -4.46881652e-01 3.02784026e-01
-3.17400932e-01 -1.21551788e+00 -5.86346626e-01 -7.59383738e-01
1.91525612e-02 8.34173620e-01 -1.14711404e+00 1.66404694e-01
-5.02061427e-01 -1.03152990e+00 7.12904394e-01 3.65376532e-01
-7.82676637e-02 -9.53477085e-01 -6.52615666e-01 3.55449736e-01
-2.61394739e-01 -1.46487558e+00 -1.88917443e-01 8.99165422e-02
-5.93118608e-01 -1.58910084e+00 -1.83470711e-01 -5.03655672e-01
7.27985799e-01 -9.26993906e-01 1.40106452e+00 -5.35717383e-02
-3.36805850e-01 -2.53637582e-02 -1.66342258e-01 -7.45831847e-01
-6.90401554e-01 -3.40294465e-02 -3.92429918e-01 -5.23431540e-01
6.81945145e-01 -5.80415241e-02 -7.66585052e-01 -9.75931212e-02
-6.88396871e-01 3.74013871e-01 6.18365943e-01 5.90730250e-01
7.13323116e-01 -4.31454897e-01 6.16391897e-01 -1.63502324e+00
9.78177905e-01 -6.74609780e-01 -2.34515905e-01 4.84375238e-01
-9.36754465e-01 2.76783220e-02 1.06863178e-01 1.36209494e-02
-7.44631290e-01 1.55596323e-02 -2.56923854e-01 4.77710426e-01
1.16958328e-01 1.04487836e+00 -2.72605896e-01 6.61733747e-01
6.62413418e-01 -2.49947056e-01 -1.28362060e-01 -4.55184758e-01
8.06944907e-01 7.51181006e-01 3.73754889e-01 -3.71346354e-01
2.46442705e-01 1.97695166e-01 3.51210654e-01 -2.35723406e-01
-1.02861929e+00 -8.44787955e-01 -6.55826628e-01 3.90299447e-02
8.84729087e-01 -8.69244277e-01 -9.62132573e-01 7.37574100e-02
-1.37414861e+00 6.37553036e-02 -6.46020830e-01 4.23049212e-01
-9.82845202e-02 1.71428457e-01 -1.02305830e+00 -5.94500661e-01
-8.75624597e-01 -1.00915086e+00 1.00475585e+00 3.05032462e-01
-9.40439224e-01 -1.20605826e+00 1.29059315e-01 1.03436828e-01
2.24321499e-01 5.64807236e-01 8.48970056e-01 -1.19560885e+00
-2.96453922e-03 -2.03065276e-01 -3.41182470e-01 -1.32399917e-01
5.10913134e-01 1.23789400e-01 -4.75403875e-01 3.20819378e-01
-3.13234299e-01 1.89235449e-01 4.58319902e-01 3.42961490e-01
9.41484272e-01 -4.43682075e-01 -8.52224171e-01 2.62266517e-01
9.62202370e-01 -7.96390921e-02 5.16460299e-01 1.36134192e-01
6.48137808e-01 6.86106443e-01 4.94807631e-01 4.70553428e-01
4.58284020e-01 1.99029714e-01 4.45443504e-02 -2.36161828e-01
-2.87430257e-01 -1.12730697e-01 -6.09720983e-02 1.10941553e+00
-1.55184239e-01 -1.49063110e-01 -9.75729406e-01 4.75937754e-01
-1.75469124e+00 -5.93113244e-01 -7.40170836e-01 1.63718951e+00
1.52008712e+00 1.92693964e-01 -3.45324963e-01 -1.54800758e-01
2.88540214e-01 -3.75149876e-01 -2.10921466e-01 -2.34448254e-01
-1.87853426e-01 5.53297460e-01 4.91355330e-01 5.25641918e-01
-1.13894081e+00 7.74980485e-01 6.83211708e+00 7.19431639e-01
-3.99667948e-01 6.72105476e-02 5.89487910e-01 8.44125599e-02
-3.94507945e-01 2.30741072e-02 -9.02954698e-01 2.22290099e-01
1.28062117e+00 -4.78718936e-01 -2.08547682e-01 2.48082832e-01
4.42869842e-01 -1.17717043e-01 -1.41140807e+00 5.98298728e-01
-8.31315890e-02 -1.75540483e+00 -4.57019322e-02 -6.68789297e-02
5.16023278e-01 -2.70834476e-01 -3.73206526e-01 2.77854085e-01
4.04236704e-01 -1.15567267e+00 3.29716921e-01 8.50490391e-01
8.87053311e-01 -1.16335727e-01 1.05169106e+00 -2.30607018e-01
-1.05936015e+00 3.11645687e-01 1.33696914e-01 1.75660804e-01
2.24152833e-01 9.87299800e-01 -1.19289362e+00 1.19724143e+00
2.61582106e-01 1.03891850e+00 -4.78979319e-01 1.07652807e+00
-6.85528696e-01 7.03686059e-01 1.69438254e-02 -2.72975206e-01
-2.68224657e-01 1.06821256e-02 3.90199572e-01 1.81445694e+00
1.48162469e-01 3.57700646e-01 -3.41260880e-01 8.26771915e-01
-5.16321242e-01 6.67277634e-01 -6.24242201e-02 -4.25300777e-01
7.13184655e-01 1.52214348e+00 -8.23435307e-01 -8.86396170e-01
-2.31348440e-01 4.37025666e-01 7.04904944e-02 6.82228282e-02
-7.86782980e-01 -4.67021674e-01 3.46293449e-01 9.23886299e-02
-3.50918680e-01 3.65731984e-01 -6.53454244e-01 -1.08722603e+00
-2.89456993e-01 -9.14225340e-01 8.08089018e-01 -4.96973991e-01
-1.26801920e+00 4.19793367e-01 4.00629751e-02 -1.05489361e+00
-2.89604992e-01 -4.94317800e-01 -1.74972057e-01 9.69891906e-01
-1.10271537e+00 -1.11599815e+00 -7.47886375e-02 4.53837030e-02
1.42517284e-01 2.17742041e-01 1.20141315e+00 4.66667056e-01
-7.10966945e-01 3.21500838e-01 -2.68692464e-01 4.16903019e-01
8.96832228e-01 -1.53948498e+00 4.20459151e-01 5.24580359e-01
-2.72352993e-01 1.12774169e+00 1.03026760e+00 -1.11785054e+00
-1.17484319e+00 -1.07826948e+00 1.72078049e+00 -8.52382779e-01
1.21362603e+00 -2.82271326e-01 -6.07717037e-01 6.85591400e-01
3.03797543e-01 -2.63421118e-01 1.03945196e+00 3.73663813e-01
5.78153282e-02 3.81200373e-01 -9.91423905e-01 4.35370564e-01
8.54318261e-01 -3.11068892e-01 -8.27421427e-01 6.35257721e-01
7.27100670e-01 -9.38181520e-01 -1.57230401e+00 4.90899056e-01
2.47020721e-01 -8.65871608e-02 8.34038377e-01 -9.96383667e-01
7.66692340e-01 -3.47091168e-01 4.30510312e-01 -1.00818074e+00
-4.34285551e-02 -6.84416354e-01 -5.75594008e-01 1.27041340e+00
1.22274184e+00 -4.19044882e-01 2.76161790e-01 8.64028454e-01
-1.25957698e-01 -9.46589112e-01 -8.64188194e-01 -2.23050281e-01
-4.72872481e-02 -3.69503438e-01 6.03192925e-01 1.20629776e+00
8.71285677e-01 8.95641923e-01 -2.54977327e-02 -3.30606513e-02
4.98576373e-01 1.15057430e-03 2.43157089e-01 -1.00762618e+00
-2.91161865e-01 -4.50548977e-01 -1.61521450e-01 -4.05179232e-01
-1.58796191e-01 -1.07501543e+00 -1.25852153e-01 -2.23883724e+00
5.21696270e-01 -1.59851298e-01 -5.60620308e-01 5.24369299e-01
-6.13807321e-01 -1.67979032e-01 -2.30468392e-01 8.09443593e-02
-7.39698291e-01 -1.40794754e-01 1.20017362e+00 -1.78257450e-01
-2.13747218e-01 -1.47197485e-01 -8.98206830e-01 5.14598548e-01
4.94114101e-01 -5.02397418e-01 -9.66914743e-02 8.14899728e-02
7.36530960e-01 -3.09580304e-02 -3.06692272e-02 -2.53693253e-01
4.87050474e-01 -1.40708834e-01 1.28952190e-01 -6.01195812e-01
-4.04672325e-02 -3.73493195e-01 3.82951647e-01 5.39211333e-01
-6.72980011e-01 -5.44059239e-02 2.49276370e-01 4.60785814e-02
-6.68698624e-02 -2.20117375e-01 9.58961397e-02 -2.15442404e-01
-1.48664787e-01 -1.29484937e-01 -3.28615308e-01 4.32222299e-02
7.50051379e-01 5.25188446e-01 -7.20183313e-01 2.72573739e-01
-1.11637497e+00 5.72803199e-01 -2.79613316e-01 1.72606573e-01
4.27753240e-01 -8.69083881e-01 -1.17915213e+00 -4.96073127e-01
6.87555745e-02 4.79047112e-02 1.50696650e-01 1.45384717e+00
-4.14475918e-01 4.93896544e-01 1.42373219e-01 -9.20243859e-02
-1.42368460e+00 4.98805076e-01 4.13483411e-01 -8.72742534e-01
-4.05394733e-01 8.43645811e-01 -1.55237705e-01 -2.61234164e-01
4.11119573e-02 -7.09696651e-01 -8.10792506e-01 3.57278049e-01
6.57091141e-01 3.47611159e-01 2.48092279e-01 -3.76724154e-01
-9.29832518e-01 -1.78028643e-01 -8.43374357e-02 -3.28423013e-03
1.30724943e+00 -8.11866447e-02 -8.83625448e-01 3.99684787e-01
9.59948659e-01 2.11298153e-01 -2.23085016e-01 -6.86248317e-02
6.16242170e-01 5.36810935e-01 -5.31643406e-02 -1.22457576e+00
-5.20258188e-01 3.59028816e-01 -5.29632419e-02 8.20675120e-02
8.30044508e-01 1.99630544e-01 3.63065124e-01 -9.26404521e-02
-2.85048932e-01 -1.06534219e+00 -6.65006816e-01 3.54328632e-01
9.03346598e-01 -8.02447617e-01 6.68687105e-01 -1.05426931e+00
-3.24247777e-01 1.08808744e+00 2.85235047e-01 4.14581925e-01
9.25760865e-01 7.78678119e-01 -1.44907653e-01 -7.58511424e-01
-9.27784085e-01 -1.92324430e-01 7.76541352e-01 1.08572535e-01
1.21305847e+00 2.97083884e-01 -1.22217929e+00 9.63397145e-01
-1.81347832e-01 1.41576186e-01 4.43343997e-01 9.12754774e-01
1.54374018e-01 -1.06378853e+00 -4.03276645e-02 6.99724197e-01
-9.96882856e-01 -5.66613853e-01 -6.50335371e-01 6.42110169e-01
2.73447454e-01 1.00903916e+00 -2.50503421e-01 -3.36675495e-01
7.08960593e-01 -9.26641151e-02 2.61818260e-01 -7.78788984e-01
-1.00323784e+00 3.03815991e-01 8.44792724e-01 -5.62648058e-01
-6.44690394e-01 -7.92371094e-01 -1.77118492e+00 1.34752914e-01
-3.70883912e-01 3.60960335e-01 5.66520214e-01 1.21716785e+00
4.96620864e-01 1.30657744e+00 -1.23067573e-01 2.99306542e-01
-2.05395609e-01 -1.42171907e+00 -2.33193845e-01 4.16059017e-01
-1.28996953e-01 -2.95198053e-01 1.41062560e-02 2.83901602e-01] | [8.495051383972168, 8.717028617858887] |
9fbab5e1-8819-482e-ba1f-296460664f31 | 1000-fps-hdr-video-with-a-spike-rgb-hybrid | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chang_1000_FPS_HDR_Video_With_a_Spike-RGB_Hybrid_Camera_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chang_1000_FPS_HDR_Video_With_a_Spike-RGB_Hybrid_Camera_CVPR_2023_paper.pdf | 1000 FPS HDR Video With a Spike-RGB Hybrid Camera | Capturing high frame rate and high dynamic range (HFR&HDR) color videos in high-speed scenes with conventional frame-based cameras is very challenging. The increasing frame rate is usually guaranteed by using shorter exposure time so that the captured video is severely interfered by noise. Alternating exposures could alleviate the noise issue but sacrifice frame rate due to involving long-exposure frames. The neuromorphic spiking camera records high-speed scenes of high dynamic range without colors using a completely different sensing mechanism and visual representation. We introduce a hybrid camera system composed of a spiking and an alternating-exposure RGB camera to capture HFR&HDR scenes with high fidelity. Our insight is to bring each camera's superiority into full play. The spike frames, with accurate fast motion information encoded, are first reconstructed for motion representation, from which the spike-based optical flows guide the recovery of missing temporal information for middle- and long-exposure RGB images while retaining their reliable color appearances. With the strong temporal constraint estimated from spike trains, both missing and distorted colors cross RGB frames are recovered to generate time-consistent and HFR color frames. We collect a new Spike-RGB dataset that contains 300 sequences of synthetic data and 20 groups of real-world data to demonstrate 1000 FPS HDR videos outperforming HDR video reconstruction methods and commercial high-speed cameras. | ['Boxin Shi', 'Tiejun Huang', 'Chao Xu', 'Liwen Hu', 'Yuchen Hong', 'Chu Zhou', 'Yakun Chang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['video-reconstruction'] | ['computer-vision'] | [ 4.00679231e-01 -6.76598728e-01 3.41353536e-01 4.33096439e-02
-4.01638657e-01 -5.91663837e-01 2.13666603e-01 -8.06076169e-01
-7.04413712e-01 9.12091792e-01 -2.22783566e-01 2.69544631e-01
3.03564310e-01 -5.16213000e-01 -7.91235745e-01 -9.96651351e-01
1.35403365e-01 -3.97359103e-01 4.85393852e-01 -8.02998915e-02
3.12281370e-01 2.03680873e-01 -1.88332236e+00 4.50431257e-01
3.16757232e-01 1.14769566e+00 5.80269992e-01 8.53236794e-01
-9.19432193e-03 9.95824933e-01 -4.27064449e-01 -8.05133283e-02
6.17669582e-01 -6.45020783e-01 -9.05783027e-02 6.49736524e-02
1.90022901e-01 -7.62228251e-01 -1.16965771e+00 9.97547507e-01
2.64820039e-01 -4.60500792e-02 -7.93977007e-02 -1.26751781e+00
-6.51916504e-01 2.23841354e-01 -7.06483066e-01 3.69183660e-01
6.70856357e-01 8.19416106e-01 2.09340513e-01 -6.19450867e-01
9.64217484e-01 9.58686471e-01 3.43005121e-01 9.47337866e-01
-1.45537305e+00 -8.61713290e-01 -1.24972537e-01 4.68050241e-02
-1.28912330e+00 -6.64657235e-01 7.38131166e-01 -2.29855120e-01
8.69670987e-01 6.57419115e-02 1.29557765e+00 1.29581094e+00
4.64352876e-01 1.90014675e-01 1.34230304e+00 3.14786077e-01
2.71383286e-01 -5.57684004e-01 1.04319733e-02 4.34547871e-01
2.30795503e-01 4.27623242e-01 -1.10505033e+00 2.51515716e-01
1.52009761e+00 3.97629827e-01 -9.81284857e-01 2.41470367e-01
-1.77709353e+00 -1.75547823e-02 3.25957686e-01 2.06245378e-01
-2.59228230e-01 5.97955465e-01 -5.66592440e-02 4.73352611e-01
-1.56348601e-01 1.95351318e-01 5.48254512e-03 -5.83689809e-01
-1.00725448e+00 -7.87724629e-02 3.65629643e-01 1.09939861e+00
6.72274649e-01 5.22272766e-01 7.99353644e-02 3.13035905e-01
8.22297335e-02 8.96429062e-01 5.95256925e-01 -1.57940245e+00
1.87441662e-01 1.67876378e-01 2.72976279e-01 -7.58702457e-01
-8.14436376e-02 8.79810974e-02 -1.06179249e+00 4.85320330e-01
6.71679258e-01 1.85823604e-01 -9.51625586e-01 1.80447614e+00
-2.72775501e-01 5.84802151e-01 1.21208012e-01 1.38637435e+00
4.51498747e-01 9.55909073e-01 -3.30381900e-01 -8.08673322e-01
1.12309098e+00 -2.84250528e-01 -1.08632863e+00 -5.01509644e-02
-2.05005720e-01 -4.51799452e-01 1.11547184e+00 6.26241446e-01
-1.42621887e+00 -5.07634223e-01 -1.18898225e+00 -2.54875898e-01
1.17415667e-01 -2.77474731e-01 4.55095291e-01 2.89996803e-01
-1.18047237e+00 6.57961369e-01 -1.01956379e+00 1.69327527e-01
3.42704952e-01 3.10306847e-01 -5.72472274e-01 -4.52825218e-01
-9.56970215e-01 4.09753621e-01 1.04024269e-01 5.95887974e-02
-1.05031812e+00 -8.21314752e-01 -4.46829557e-01 -3.68090183e-01
-2.57833481e-01 -6.61159039e-01 7.83687949e-01 -1.11210740e+00
-1.73065925e+00 8.48089457e-01 -6.74753040e-02 -3.62291068e-01
4.77431118e-01 1.02387875e-01 -3.60185206e-01 7.30791867e-01
-1.66320518e-01 7.07448244e-01 9.44416940e-01 -1.12524474e+00
-2.38662943e-01 -3.10496569e-01 -1.60644621e-01 2.29006764e-02
-2.90739536e-02 -9.76012722e-02 -5.52695096e-01 -5.09611905e-01
2.20187068e-01 -9.33374941e-01 2.01652661e-01 5.44569790e-01
-3.36299241e-02 8.35170507e-01 1.08826828e+00 -4.22862440e-01
9.00327563e-01 -2.37471986e+00 1.95378184e-01 -5.03216386e-01
4.71101761e-01 7.02936351e-02 -1.96522623e-01 1.78185627e-01
5.60324714e-02 -3.79780531e-01 -4.86438364e-01 -7.52554238e-02
-6.92077160e-01 3.07908893e-01 -5.44529080e-01 7.77648687e-01
6.93459064e-02 8.38825226e-01 -1.06728506e+00 -3.20260763e-01
3.58077914e-01 8.71991396e-01 -4.20315892e-01 3.10578734e-01
1.13323443e-01 9.79108572e-01 -1.91736251e-01 7.10926235e-01
8.17179561e-01 -2.54322380e-01 -7.95223117e-02 -4.70503390e-01
-4.05616403e-01 -3.44823956e-01 -9.30770218e-01 2.05988359e+00
-3.49909812e-02 9.42598581e-01 6.57593608e-02 -3.13842326e-01
8.15036356e-01 6.30782768e-02 7.81718194e-01 -1.30558467e+00
1.20766036e-01 3.54714692e-01 -2.15321735e-01 -5.11215270e-01
3.80888104e-01 -2.48447195e-01 1.02921583e-01 2.21295998e-01
-5.32180555e-02 -1.41697377e-01 4.17545587e-02 4.03339326e-01
1.36660612e+00 4.14177835e-01 -2.89886653e-01 -1.20388590e-01
-4.85533550e-02 -3.49050790e-01 7.01577723e-01 4.55725551e-01
-4.44665670e-01 1.07938230e+00 2.13061735e-01 -4.14877623e-01
-1.48870599e+00 -1.33883464e+00 -1.29444823e-01 2.92611063e-01
8.15203309e-01 -2.54340053e-01 -4.03180808e-01 3.41007859e-01
-3.91616255e-01 3.10141563e-01 -4.85950083e-01 -3.06331336e-01
-6.18646383e-01 -7.18764484e-01 5.79887927e-01 2.20122844e-01
7.90244162e-01 -9.33820069e-01 -1.30437994e+00 3.04023057e-01
-4.18531179e-01 -1.36000264e+00 -4.34267908e-01 2.30526328e-01
-7.85502374e-01 -1.02438545e+00 -9.89809275e-01 -4.77985054e-01
5.02828240e-01 8.28964114e-01 8.23367596e-01 -6.63104653e-02
-4.66874570e-01 2.66537160e-01 -3.49778533e-01 3.65408957e-01
4.53390442e-02 -1.02589917e+00 1.20142996e-01 5.19603677e-03
-9.35705379e-02 -8.86146903e-01 -1.02649939e+00 2.93642044e-01
-1.27091110e+00 4.80477959e-01 1.37636736e-01 7.15994895e-01
7.63081074e-01 -3.85602325e-01 1.91564009e-01 -2.32821405e-01
-1.37014300e-01 -1.84628740e-01 -5.59279740e-01 -1.77272826e-01
-2.03130692e-02 -2.25229740e-01 8.89188766e-01 -7.73736715e-01
-8.75068367e-01 1.74942270e-01 2.31250599e-01 -8.58815432e-01
1.83795229e-01 -4.75117624e-01 1.71344191e-01 -2.94293374e-01
3.91871601e-01 7.66593814e-01 1.13773957e-01 1.35271773e-01
1.50956497e-01 3.22199523e-01 1.01764095e+00 -2.17280641e-01
7.56680608e-01 1.11358762e+00 -1.38712868e-01 -7.79432178e-01
-2.89117575e-01 -6.37511387e-02 -1.60263672e-01 -8.24107170e-01
1.03402984e+00 -1.25084424e+00 -1.06942511e+00 1.14979041e+00
-1.02139008e+00 -5.36714435e-01 -3.17425370e-01 6.64497495e-01
-7.30683923e-01 2.72382289e-01 -1.23679447e+00 -7.33151376e-01
8.65703449e-02 -1.25923789e+00 1.10836256e+00 5.39939642e-01
4.00116861e-01 -1.78805768e-01 7.95343705e-03 -8.37064609e-02
5.05454421e-01 3.03806543e-01 3.35261613e-01 6.57937169e-01
-1.07714307e+00 2.34862581e-01 -3.75596762e-01 -9.83613804e-02
3.31125557e-02 2.05380797e-01 -1.16649199e+00 -3.90220016e-01
3.82007807e-01 -3.52215171e-01 9.68884170e-01 5.57758689e-01
1.14083219e+00 1.06251486e-01 1.01814640e-03 1.20460904e+00
1.81225288e+00 5.03127277e-01 1.27530503e+00 1.98282778e-01
7.52289355e-01 2.54259735e-01 2.73248732e-01 4.74522531e-01
1.12287857e-01 5.61209500e-01 4.61626798e-01 3.10906563e-02
-4.03823197e-01 -1.13343209e-01 7.68353999e-01 9.12162602e-01
-3.13630491e-01 -2.46462911e-01 -4.02768314e-01 1.53440863e-01
-1.57955420e+00 -1.22384322e+00 -3.42048734e-01 2.48012495e+00
8.64018798e-01 -4.19895314e-02 1.02421746e-01 2.78973818e-01
8.84862125e-01 -4.20349538e-02 -8.70926261e-01 3.16593319e-01
-9.09337461e-01 -9.37250778e-02 5.46727002e-01 1.43138945e-01
-4.26708341e-01 6.24474585e-01 6.03585148e+00 2.79689252e-01
-1.39819133e+00 -1.00646213e-01 5.21482825e-01 -8.35452557e-01
-5.52051902e-01 9.54097956e-02 -4.27184820e-01 1.04450357e+00
1.04503489e+00 -1.98936108e-02 9.90981400e-01 9.27827805e-02
4.45814252e-01 -4.21677411e-01 -8.25083792e-01 1.83635175e+00
-8.21190551e-02 -1.39427209e+00 -9.64838490e-02 2.05938742e-01
5.82269251e-01 -8.18509609e-02 2.52751529e-01 -3.03542942e-01
-1.52482965e-03 -6.19547606e-01 1.02285421e+00 9.11409140e-01
1.38059998e+00 -4.59147394e-01 1.26011193e-01 -6.80977404e-02
-1.17987061e+00 -3.01519006e-01 -6.35485411e-01 -1.30557343e-01
5.90856433e-01 7.67234325e-01 4.32225883e-01 8.85013565e-02
1.06670487e+00 1.36308670e+00 -3.60009760e-01 8.52746129e-01
3.60338569e-01 1.33366123e-01 -3.03017676e-01 3.64857167e-01
-3.27905491e-02 -4.66749340e-01 5.76661408e-01 7.51101315e-01
6.84721768e-01 7.14079440e-01 -3.35963845e-01 1.18722463e+00
-3.26229073e-03 -8.56265903e-01 -8.74484241e-01 -1.16187163e-01
4.98035371e-01 1.14737868e+00 -9.04767513e-01 -2.13389426e-01
-5.05859911e-01 1.35771787e+00 -8.82416666e-02 5.71797788e-01
-1.14643681e+00 -3.10872138e-01 7.36626148e-01 1.13304503e-01
-4.07095859e-03 -5.43807864e-01 -5.77408373e-02 -1.72857416e+00
9.12829787e-02 -4.81727540e-01 -1.89450100e-01 -1.39634335e+00
-1.02020919e+00 6.60056651e-01 -5.54351032e-01 -1.67465031e+00
3.13210428e-01 -3.88295650e-01 -2.63204575e-01 4.76119876e-01
-1.55965960e+00 -6.39889181e-01 -6.79032266e-01 1.04340279e+00
4.59409267e-01 2.44559005e-01 4.17642295e-01 5.40437579e-01
-6.84167922e-01 6.80953041e-02 2.73397952e-01 -1.20929465e-01
8.92240405e-01 -8.30087662e-01 7.61575922e-02 1.08555913e+00
-2.62988150e-01 5.15076697e-01 5.95986366e-01 -6.16806149e-01
-2.29096794e+00 -9.74597454e-01 -7.54803494e-02 -1.48296908e-01
5.62219501e-01 -3.85208905e-01 -9.45587635e-01 2.93090075e-01
1.31511748e-01 3.42576623e-01 2.61096448e-01 -1.23209858e+00
-4.35469389e-01 -2.87953496e-01 -1.16374815e+00 5.17561078e-01
1.26469219e+00 -7.97700405e-01 -1.38329417e-01 -2.40141645e-01
8.84677887e-01 -1.49326682e-01 -7.16459394e-01 -2.71214768e-02
8.38034213e-01 -1.33908617e+00 8.54867697e-01 2.71608621e-01
6.86733425e-01 -7.75764644e-01 -4.02790397e-01 -8.56380582e-01
-5.69829382e-02 -1.01521325e+00 -4.67607565e-02 1.01988578e+00
-2.60638446e-01 -6.32468224e-01 4.03784871e-01 5.97700059e-01
-1.30279824e-01 -1.50497481e-01 -9.02976990e-01 -7.45854437e-01
-4.66922194e-01 -2.96085179e-01 7.13703632e-02 7.91040182e-01
7.25383461e-02 -1.92532375e-01 -7.22470701e-01 -1.48508430e-01
1.07687509e+00 1.40389174e-01 2.35336781e-01 -6.97114587e-01
-1.81710869e-01 3.77250537e-02 -7.01413572e-01 -1.02957582e+00
-1.83318436e-01 -4.17266399e-01 2.75817811e-01 -8.99374366e-01
5.00059247e-01 2.30939295e-02 -2.96885699e-01 -4.09405009e-04
1.77433744e-01 8.52935970e-01 4.40954208e-01 6.48797631e-01
-6.23283505e-01 5.50226808e-01 1.41799259e+00 1.76793545e-01
-7.30049163e-02 -8.22936356e-01 -1.20629579e-01 4.09315825e-01
2.28369415e-01 -3.34398210e-01 -3.70581061e-01 -7.23515034e-01
3.16604376e-01 7.30125964e-01 7.02145398e-01 -1.43251085e+00
4.37993646e-01 -5.05345426e-02 8.33490908e-01 -1.98911831e-01
5.59567988e-01 -8.15061569e-01 7.75375009e-01 6.87027872e-01
-2.97055364e-01 3.92456234e-01 4.13307501e-03 8.70768487e-01
-1.99586987e-01 4.69021320e-01 1.34057963e+00 -4.17225122e-01
-8.59695852e-01 5.22935867e-01 -5.80319881e-01 9.50133204e-02
1.10913503e+00 -8.02659869e-01 -8.53750408e-01 -3.47284228e-01
-4.42201108e-01 -3.04748595e-01 1.22109127e+00 1.74010366e-01
1.18686759e+00 -1.37638581e+00 -2.82583565e-01 4.45402026e-01
-1.27012981e-02 -3.34584117e-01 7.43318319e-01 8.63126099e-01
-8.24552000e-01 -5.25793172e-02 -9.99412775e-01 -8.70113432e-01
-6.45331800e-01 6.63601518e-01 2.02611044e-01 5.72219908e-01
-1.29685926e+00 7.10049808e-01 2.77274013e-01 6.87363982e-01
-5.64568043e-02 -2.51591325e-01 3.06473494e-01 -2.55281597e-01
9.67676520e-01 3.83728713e-01 -4.41284835e-01 -5.13207436e-01
-3.18820715e-01 1.02436352e+00 4.15507704e-01 -3.69700253e-01
1.38605845e+00 -7.78332412e-01 1.19521983e-01 7.13936210e-01
1.30237603e+00 -3.71886075e-01 -2.08369040e+00 2.16072559e-01
-6.59251630e-01 -9.78648007e-01 1.31091937e-01 -4.47253346e-01
-1.50861895e+00 9.54015195e-01 8.14277172e-01 7.20046367e-03
1.60751927e+00 -2.80776650e-01 1.22941279e+00 6.10174723e-02
8.09538841e-01 -8.01429152e-01 3.74700189e-01 1.17671773e-01
4.48523402e-01 -8.80829632e-01 -4.92307425e-01 -1.05836943e-01
-6.92620575e-01 1.31749737e+00 5.71067274e-01 -4.84568477e-01
9.26572084e-02 8.05516720e-01 -1.02544226e-01 6.19133785e-02
-1.06853008e+00 -2.79281549e-02 -5.30545175e-01 7.33420789e-01
6.83893412e-02 -3.59980583e-01 1.25619069e-01 2.34213710e-01
1.54452711e-01 3.40387255e-01 1.12928283e+00 6.73475742e-01
-3.73344183e-01 -4.96156812e-01 -1.93018183e-01 2.35633507e-01
-3.73456568e-01 1.42645929e-02 -8.71035755e-02 4.82467771e-01
1.90005754e-03 7.39037991e-01 2.14600146e-01 -5.55748522e-01
1.04704380e-01 -2.40272462e-01 9.09462750e-01 5.25332950e-02
-2.72780985e-01 4.17537659e-01 -3.70864540e-01 -1.32732296e+00
-6.33737147e-01 -4.13461626e-01 -1.49793637e+00 -4.73810017e-01
1.74546659e-01 -3.67173880e-01 6.20202541e-01 3.78282547e-01
3.15037429e-01 5.72100818e-01 6.65753305e-01 -1.07393897e+00
2.18546614e-01 -2.11663112e-01 -1.06788504e+00 6.90970600e-01
7.91408062e-01 -4.32880729e-01 -7.31458008e-01 6.08763635e-01] | [8.851215362548828, -1.2586861848831177] |
7e924ef3-ba3e-4d38-8eb4-0b47fdc1c0f6 | assisted-text-annotation-using-active | 2112.11914 | null | https://arxiv.org/abs/2112.11914v1 | https://arxiv.org/pdf/2112.11914v1.pdf | Assisted Text Annotation Using Active Learning to Achieve High Quality with Little Effort | Large amounts of annotated data have become more important than ever, especially since the rise of deep learning techniques. However, manual annotations are costly. We propose a tool that enables researchers to create large, high-quality, annotated datasets with only a few manual annotations, thus strongly reducing annotation cost and effort. For this purpose, we combine an active learning (AL) approach with a pre-trained language model to semi-automatically identify annotation categories in the given text documents. To highlight our research direction's potential, we evaluate the approach on the task of identifying frames in news articles. Our preliminary results show that employing AL strongly reduces the number of annotations for correct classification of even these complex and subtle frames. On the framing dataset, the AL approach needs only 16.3\% of the annotations to reach the same performance as a model trained on the full dataset. | ['Bela Gipp', 'Karsten Donnay', 'Felix Hamborg', 'Franziska Weeber'] | 2021-12-15 | null | null | null | null | ['text-annotation'] | ['natural-language-processing'] | [ 2.56362349e-01 6.62975311e-01 -5.00085592e-01 -5.84536076e-01
-1.32292020e+00 -7.04515815e-01 5.54746270e-01 5.63376248e-01
-8.53327632e-01 7.75347769e-01 3.50696415e-01 -1.61077887e-01
4.63101566e-01 -3.90067101e-01 -5.23231387e-01 -1.75496504e-01
3.34517121e-01 5.11239529e-01 4.50630099e-01 1.33203194e-01
9.83686671e-02 2.06537219e-03 -1.41921473e+00 7.12641597e-01
7.88375318e-01 1.05463886e+00 9.55908298e-02 4.06143814e-01
-3.86999458e-01 1.21862793e+00 -6.56829476e-01 -6.67167127e-01
-2.47523165e-03 -2.27495342e-01 -1.13548398e+00 9.82631817e-02
6.11248612e-01 -1.85681552e-01 1.68896720e-01 7.09626973e-01
4.35833335e-01 8.85369815e-03 2.90681213e-01 -7.25069225e-01
-2.20824316e-01 9.89143312e-01 -4.75286812e-01 3.76773357e-01
3.85782272e-01 -1.55026421e-01 1.30795634e+00 -9.25785482e-01
8.53660643e-01 9.00837421e-01 1.05481398e+00 3.14480364e-01
-1.35850370e+00 -4.16608691e-01 8.98722485e-02 1.01269111e-01
-1.16895270e+00 -8.56478572e-01 6.90415919e-01 -7.20339954e-01
9.23326492e-01 1.18692748e-01 6.01872742e-01 1.23812556e+00
-2.46902287e-01 9.37276721e-01 7.67849028e-01 -9.27547216e-01
1.31739259e-01 1.73354939e-01 5.08670211e-01 7.67993152e-01
3.30900460e-01 -5.74948430e-01 -5.17970860e-01 -3.43827397e-01
2.74812549e-01 -4.63846594e-01 -6.03907891e-02 -2.09793195e-01
-1.26101089e+00 9.17722404e-01 -4.80104387e-02 5.15362680e-01
-1.80215031e-01 7.89142027e-03 6.80230260e-01 1.06844995e-02
1.04525638e+00 8.98517132e-01 -6.84312820e-01 -4.42220479e-01
-1.12315154e+00 4.66866754e-02 9.68798399e-01 7.98340917e-01
5.79269946e-01 -2.64057994e-01 -1.72359824e-01 8.23289871e-01
2.86430210e-01 -1.34841329e-03 1.15405053e-01 -9.51219141e-01
6.21219993e-01 1.03283143e+00 5.50171077e-01 -1.02099264e+00
-4.58101034e-01 -2.55306125e-01 -3.50689471e-01 -2.24826902e-01
8.10445309e-01 -3.98484230e-01 -5.97600818e-01 1.38952649e+00
1.72354653e-01 -2.29815543e-01 -3.42546076e-01 4.37987745e-01
7.84033954e-01 4.86522675e-01 3.57159853e-01 -4.94458944e-01
1.30799878e+00 -9.63324547e-01 -9.82858419e-01 -3.71784419e-01
1.02240050e+00 -9.54045892e-01 1.19535041e+00 2.78047442e-01
-1.05433440e+00 -5.21522939e-01 -6.89248502e-01 -2.29533672e-01
-2.09406048e-01 5.70309281e-01 7.10479438e-01 4.80780691e-01
-7.17017114e-01 1.65522009e-01 -9.26701188e-01 -1.71027780e-01
6.76339626e-01 2.16271624e-01 -2.88788706e-01 1.89587116e-01
-1.15429425e+00 1.02005863e+00 6.77378714e-01 -1.91276044e-01
-6.79037094e-01 -6.28076911e-01 -8.20120633e-01 1.17408246e-01
8.71641338e-01 3.90023477e-02 1.64717889e+00 -1.52756548e+00
-1.17807627e+00 1.00840306e+00 -1.45342767e-01 -5.47207177e-01
5.39178193e-01 -6.80512190e-01 -3.28012764e-01 1.21112376e-01
1.51181012e-01 5.80532312e-01 5.37444711e-01 -1.07805014e+00
-8.18210483e-01 9.15804133e-02 2.91632473e-01 -1.66392520e-01
-6.51075721e-01 4.92807090e-01 -2.57955581e-01 -5.93156815e-01
-3.27624261e-01 -9.20262277e-01 -1.42821670e-01 -6.39614789e-03
-9.50902775e-02 -4.99746829e-01 8.20745885e-01 -8.15408647e-01
1.39122045e+00 -2.27346301e+00 -8.15976486e-02 -1.54859588e-01
4.16476816e-01 3.62598449e-01 1.61869109e-01 1.39139414e-01
1.45949736e-01 2.63320893e-01 -1.42839432e-01 -4.65591043e-01
-1.23523436e-01 2.69591480e-01 -2.66772032e-01 2.55617291e-01
3.11102986e-01 8.39619279e-01 -1.04167128e+00 -7.30171025e-01
-3.43777895e-01 2.68541902e-01 -5.85658789e-01 1.30545869e-01
-5.24643362e-01 5.34003675e-01 -1.82826251e-01 3.31137896e-01
-3.80036980e-02 -4.51929241e-01 5.57546496e-01 -2.16192350e-01
-3.34628433e-01 5.74298739e-01 -9.12241638e-01 1.58230102e+00
-4.46225911e-01 1.16449928e+00 -7.87969008e-02 -1.19848621e+00
9.10423696e-01 7.46253908e-01 5.86836934e-01 -5.24350822e-01
1.68793648e-01 3.27962339e-01 7.03138784e-02 -6.02589130e-01
4.28119093e-01 2.59979188e-01 -3.17648053e-01 6.29287779e-01
3.38829815e-01 4.36756879e-01 5.28371692e-01 1.46502435e-01
1.09803224e+00 7.39861131e-02 4.50793594e-01 -2.91134596e-01
3.14853668e-01 3.54128122e-01 7.53253996e-01 7.17055678e-01
-1.52524680e-01 2.21491337e-01 7.21731186e-01 -8.12772036e-01
-1.36023581e+00 -2.46158078e-01 6.77950382e-02 1.38805485e+00
-3.02704602e-01 -6.43095076e-01 -8.85725975e-01 -1.14638650e+00
-3.75434875e-01 6.87889040e-01 -5.33292294e-01 2.88573414e-01
-8.57376635e-01 -7.20464885e-01 7.43966401e-01 4.60420102e-01
5.28060019e-01 -8.33429992e-01 -8.04750562e-01 2.38834009e-01
-6.13164783e-01 -1.18768525e+00 -7.45399445e-02 1.11616001e-01
-5.12667954e-01 -1.09759152e+00 -2.62760103e-01 -7.50492513e-01
6.48386538e-01 -1.30348012e-01 1.37584078e+00 8.58912095e-02
-1.03177227e-01 2.63612688e-01 -8.09660137e-01 -6.70376658e-01
-4.79871720e-01 3.77808630e-01 -2.63532937e-01 4.30137245e-03
5.85720360e-01 1.15857258e-01 -1.07826635e-01 2.32820317e-01
-6.27168834e-01 1.57412514e-01 2.94187486e-01 6.85984373e-01
3.40213060e-01 -8.47730562e-02 5.80614269e-01 -1.17580283e+00
4.37774271e-01 -3.55849504e-01 -6.29307628e-01 2.11869821e-01
-4.53004032e-01 -1.48265377e-01 3.60566795e-01 -6.02368295e-01
-1.23554146e+00 4.25830960e-01 2.71460153e-02 3.00560474e-01
-9.41070244e-02 6.96564376e-01 2.75865644e-02 1.56571612e-01
9.75545943e-01 -5.04908264e-01 -3.52984607e-01 -7.19553590e-01
1.87432542e-01 7.40653753e-01 2.65343457e-01 -4.07690674e-01
5.10817289e-01 3.02515984e-01 -5.74919581e-01 -5.75556457e-01
-1.87923884e+00 -4.80194837e-01 -1.06085813e+00 -2.74313360e-01
8.59082043e-01 -9.19082999e-01 -1.09496126e-02 7.07951859e-02
-1.35064721e+00 -5.74610949e-01 -4.33542579e-01 5.43630183e-01
-3.14098656e-01 2.86937445e-01 -5.09303749e-01 -7.60957062e-01
-1.76553875e-01 -1.01993287e+00 9.22417581e-01 -7.32358545e-02
-8.02496910e-01 -9.84903991e-01 1.67404488e-01 7.64581263e-01
1.54027209e-01 3.06839377e-01 7.36250401e-01 -1.15675247e+00
-1.92378759e-01 -3.40198934e-01 -9.78003964e-02 2.44814515e-01
-5.18350117e-02 7.79242739e-02 -1.01000082e+00 -6.12599291e-02
-1.60889000e-01 -5.95061123e-01 6.25211000e-01 6.05673678e-02
1.09996402e+00 -5.45659423e-01 -2.46541336e-01 -8.87674242e-02
1.00188315e+00 2.53658682e-01 4.74002302e-01 4.57758158e-01
6.72606289e-01 3.32940549e-01 7.40100920e-01 3.34840566e-01
2.23556206e-01 8.76879990e-01 2.10601799e-02 -2.60422707e-01
-1.10294558e-01 -3.06419842e-02 2.26349667e-01 7.59530783e-01
-3.76143962e-01 -3.36087883e-01 -1.52007914e+00 5.40945470e-01
-1.94321084e+00 -1.13777459e+00 -2.18041345e-01 1.96820021e+00
1.32886553e+00 5.49423993e-01 1.20787174e-01 4.26640302e-01
7.68097639e-01 1.15043353e-02 -6.69731759e-03 -2.22534075e-01
-1.90378400e-03 8.50569159e-02 2.90294170e-01 3.69464129e-01
-1.59251261e+00 7.88014352e-01 6.84748459e+00 7.02396989e-01
-8.52175474e-01 5.63167274e-01 6.97818220e-01 2.44539306e-02
1.25341147e-01 1.07357554e-01 -8.92413318e-01 4.87153500e-01
1.17118025e+00 1.23625830e-01 -1.62954882e-01 1.18889511e+00
2.91246027e-01 -2.27299213e-01 -1.10433245e+00 7.01127172e-01
2.00139672e-01 -1.60385728e+00 -3.02858859e-01 -1.08029982e-02
8.15730155e-01 -6.68844059e-02 -4.59815264e-01 3.49056065e-01
2.92159259e-01 -6.67552054e-01 9.32844818e-01 4.91974264e-01
6.83517516e-01 -6.07577801e-01 1.14552772e+00 4.97062981e-01
-8.42718542e-01 -2.33494043e-01 -1.19950883e-01 -4.70476866e-01
3.55012506e-01 8.70727956e-01 -1.03153253e+00 1.75337464e-01
4.74661291e-01 6.87683165e-01 -7.52191842e-01 1.01568007e+00
-3.30287814e-01 1.11500633e+00 -7.97087699e-02 8.29589814e-02
1.01956137e-01 4.05566990e-01 3.26282859e-01 1.40183759e+00
-7.33453184e-02 1.38990566e-01 6.50298655e-01 4.97868419e-01
-3.23974490e-01 2.06586465e-01 -2.93713748e-01 -3.56479645e-01
4.49146986e-01 1.33330667e+00 -1.09164345e+00 -6.85885012e-01
-5.67069471e-01 4.89195198e-01 5.59426248e-01 -6.37087524e-02
-8.30105722e-01 -2.61266828e-01 -1.62058443e-01 2.68526465e-01
9.68682170e-02 -4.00292814e-01 -3.30183476e-01 -1.18452275e+00
-1.85155347e-01 -8.99373710e-01 3.57556850e-01 -6.27304435e-01
-1.02909338e+00 4.14847285e-01 4.76658419e-02 -9.86214817e-01
-1.06912896e-01 -3.27913702e-01 -2.17215613e-01 2.73108870e-01
-9.69013631e-01 -1.14785719e+00 -3.58079284e-01 -9.26371256e-04
8.27881813e-01 -1.56055182e-01 9.14422750e-01 7.02308238e-01
-6.77901506e-01 4.51944381e-01 -1.42047361e-01 7.40746319e-01
9.18916881e-01 -1.08992350e+00 2.96382725e-01 8.33981633e-01
4.88578260e-01 4.12979782e-01 6.31339848e-01 -6.73818767e-01
-7.97002256e-01 -9.35785770e-01 1.40745950e+00 -8.08133066e-01
8.05821240e-01 -5.23176849e-01 -8.39147985e-01 8.64055037e-01
1.68445408e-01 -5.81060648e-02 1.00492513e+00 3.91801685e-01
-4.24305350e-01 1.56327844e-01 -8.64892185e-01 2.28129551e-01
8.68877172e-01 -5.78444064e-01 -9.67309713e-01 5.34469187e-01
6.69727921e-01 -3.83520901e-01 -9.17145312e-01 4.09653336e-01
4.12552059e-01 -5.60033798e-01 6.10335410e-01 -5.61304569e-01
4.61748660e-01 -7.83363134e-02 4.75953147e-02 -7.98860848e-01
-3.19600075e-01 -4.94156748e-01 -2.56024271e-01 1.45891321e+00
8.84523749e-01 -2.31839478e-01 5.32278836e-01 7.22017467e-01
-2.41864040e-01 -5.53143263e-01 -7.51577735e-01 -4.20998245e-01
-4.49564099e-01 -4.48554128e-01 3.27761145e-03 1.42746043e+00
1.68336809e-01 6.96918249e-01 -3.38175356e-01 -2.46147886e-01
2.21562445e-01 -2.24196792e-01 7.20521569e-01 -1.78953302e+00
-9.70575809e-02 -7.07970411e-02 -2.94735789e-01 -6.28081143e-01
3.69346082e-01 -7.08885610e-01 9.35248360e-02 -1.45701790e+00
3.60944718e-01 -5.66755652e-01 9.97112505e-03 1.00261712e+00
-1.09701224e-01 5.55552185e-01 -4.35950086e-02 4.19740885e-01
-1.10490465e+00 -1.29190311e-02 7.01880872e-01 -9.28320363e-02
-1.06039740e-01 -1.94027379e-01 -5.05839646e-01 1.27559125e+00
8.14475834e-01 -5.67123711e-01 -3.30228180e-01 -5.95748842e-01
6.20937943e-01 -4.00597572e-01 1.68221578e-01 -8.30756187e-01
-5.10220677e-02 -1.61591023e-02 4.77336794e-01 -2.74140239e-01
4.57720198e-02 -7.24456191e-01 2.84400284e-02 2.26181209e-01
-9.08761799e-01 -1.09369136e-01 1.57256231e-01 4.29584771e-01
-2.49082699e-01 -4.59429860e-01 7.42705703e-01 -2.74834424e-01
-7.55624175e-01 -1.69977352e-01 -5.14101505e-01 2.88419455e-01
9.67477322e-01 2.46228222e-02 -4.58526254e-01 -1.46834642e-01
-9.39845443e-01 -2.22017184e-01 3.03556323e-01 3.08431357e-01
-6.31770566e-02 -9.68962908e-01 -4.95891094e-01 -2.09215268e-01
4.17360812e-02 2.52934456e-01 -2.56167918e-01 9.36738491e-01
-5.75381875e-01 4.74106044e-01 7.78183574e-03 -5.01614392e-01
-1.35916018e+00 3.42727005e-01 -5.98503649e-02 -4.58272547e-01
-6.91990137e-01 7.25475013e-01 -3.42964202e-01 -9.13975239e-02
2.74825662e-01 -1.09367877e-01 -6.76005781e-01 7.15910912e-01
5.91811299e-01 4.08528984e-01 7.07888752e-02 -6.82208359e-01
-1.02243803e-01 3.86759490e-02 -1.40936017e-01 -1.49973884e-01
1.46231723e+00 -8.87121186e-02 -3.19076367e-02 6.98456705e-01
7.89734781e-01 3.29145402e-01 -1.14192533e+00 -3.13140392e-01
7.36834407e-01 -3.57156157e-01 2.47586414e-01 -7.96892881e-01
-5.85765004e-01 6.12906158e-01 3.64895493e-01 5.79863131e-01
6.88350022e-01 2.01334476e-01 5.16961694e-01 5.55703521e-01
3.07314545e-01 -1.41275740e+00 2.56369501e-01 5.94295382e-01
5.62959969e-01 -1.30508471e+00 1.87175006e-01 -4.82016772e-01
-6.55550957e-01 1.07556105e+00 5.72927177e-01 1.71181455e-01
2.63378531e-01 2.97746211e-01 6.77657202e-02 -2.73225933e-01
-7.73413658e-01 -7.10396618e-02 3.61395091e-01 2.70403624e-01
9.14768457e-01 -2.80864447e-01 -4.68990624e-01 4.99406040e-01
1.20299704e-01 1.36702284e-01 5.65056264e-01 1.02063549e+00
-7.61581957e-01 -1.06215918e+00 -1.12281635e-01 4.19111490e-01
-8.04047883e-01 -4.38044406e-02 -5.90603054e-01 6.63045228e-01
3.33481491e-01 8.66105676e-01 3.63725424e-01 3.93188559e-02
1.58956915e-01 5.07237673e-01 1.83376804e-01 -1.06775749e+00
-5.05794048e-01 1.81476086e-01 8.17361653e-01 -3.85275692e-01
-1.12544894e+00 -6.28170967e-01 -8.73184443e-01 1.06112286e-01
-5.51257968e-01 3.84308159e-01 3.55575114e-01 1.22462082e+00
3.69594842e-01 2.88968474e-01 2.14321285e-01 -7.91436374e-01
-2.90741682e-01 -1.07885027e+00 1.80069894e-01 3.11144114e-01
-4.14181948e-02 -8.61062229e-01 -2.45404348e-01 6.63934529e-01] | [9.672682762145996, 4.620616912841797] |
4b7d0923-a357-4c27-8654-665baeff40a7 | contrastive-learning-for-self-supervised-pre | 2301.07283 | null | https://arxiv.org/abs/2301.07283v2 | https://arxiv.org/pdf/2301.07283v2.pdf | Contrastive Learning for Self-Supervised Pre-Training of Point Cloud Segmentation Networks With Image Data | Reducing the quantity of annotations required for supervised training is vital when labels are scarce and costly. This reduction is particularly important for semantic segmentation tasks involving 3D datasets, which are often significantly smaller and more challenging to annotate than their image-based counterparts. Self-supervised pre-training on unlabelled data is one way to reduce the amount of manual annotations needed. Previous work has focused on pre-training with point clouds exclusively. While useful, this approach often requires two or more registered views. In the present work, we combine image and point cloud modalities by first learning self-supervised image features and then using these features to train a 3D model. By incorporating image data, which is often included in many 3D datasets, our pre-training method only requires a single scan of a scene and can be applied to cases where localization information is unavailable. We demonstrate that our pre-training approach, despite using single scans, achieves comparable performance to other multi-scan, point cloud-only methods. | ['Jonathan Kelly', 'Edwin G. Ng', 'Brandon Wagstaff', 'Andrej Janda'] | 2023-01-18 | null | null | null | null | ['point-cloud-segmentation'] | ['computer-vision'] | [ 5.11152565e-01 3.05720329e-01 -1.40601639e-02 -6.86505914e-01
-8.85910690e-01 -9.22686815e-01 4.09005463e-01 5.37839532e-01
-6.03385210e-01 4.12294298e-01 -5.57556927e-01 -4.96739864e-01
1.54707313e-01 -7.73628294e-01 -9.31883037e-01 -3.12494338e-01
2.70881563e-01 1.09133613e+00 7.19199061e-01 1.20600194e-01
2.46442124e-01 8.14872146e-01 -1.77614629e+00 -1.79780945e-01
8.27103317e-01 8.62860978e-01 6.09456241e-01 5.12731552e-01
-4.91733879e-01 1.62346795e-01 -2.26492703e-01 -5.47505952e-02
4.72323895e-01 -1.60925910e-01 -9.88229334e-01 7.05152988e-01
5.86456954e-01 -2.76405185e-01 3.50308359e-01 8.04874539e-01
2.36895278e-01 1.57120675e-01 6.57624364e-01 -9.64776874e-01
2.65361547e-01 -8.68931115e-02 -6.82668686e-01 -7.23585784e-02
9.80826765e-02 -6.76700696e-02 7.02061236e-01 -7.36165643e-01
5.20384908e-01 7.47410715e-01 8.40455949e-01 3.04430991e-01
-1.41182065e+00 -3.38412017e-01 -1.28675867e-02 -3.14995050e-01
-1.21967769e+00 -3.72556269e-01 8.76932502e-01 -7.17467368e-01
8.58211815e-01 4.01598103e-02 5.50054550e-01 4.69691217e-01
-5.42511046e-01 4.35132205e-01 1.37312555e+00 -6.80973649e-01
3.33274722e-01 3.93029392e-01 -1.13136232e-01 7.86360383e-01
2.62276322e-01 -1.89336196e-01 -1.80386052e-01 1.90629419e-02
1.06134164e+00 1.38372093e-01 7.16136843e-02 -1.03188419e+00
-1.07237530e+00 7.06560791e-01 4.34452027e-01 2.27075726e-01
-3.15121263e-01 4.36165556e-03 2.82516271e-01 1.39406919e-01
7.53315330e-01 4.55012619e-01 -6.24060690e-01 -1.19223565e-01
-1.33771372e+00 -1.87092856e-01 5.28572559e-01 9.32841122e-01
1.27621615e+00 -5.33395529e-01 8.00072372e-01 9.11957145e-01
1.14484861e-01 5.16889811e-01 4.28339392e-02 -1.07405722e+00
3.30218494e-01 9.08529580e-01 1.97107151e-01 -6.86862290e-01
-4.89334941e-01 -3.04471403e-01 -3.73710960e-01 4.57602590e-01
6.08834088e-01 9.08064321e-02 -1.29496014e+00 1.13491201e+00
6.79272652e-01 -1.98797464e-01 -1.60801187e-01 8.90081763e-01
5.64787090e-01 1.46201685e-01 -1.51426625e-02 -7.04172254e-02
1.05464137e+00 -7.85015225e-01 -6.15077913e-02 -4.59636033e-01
8.15596998e-01 -8.48810077e-01 1.15439618e+00 2.37339690e-01
-8.04774940e-01 -4.07194287e-01 -8.86923432e-01 -1.14701360e-01
-4.39512640e-01 8.12770054e-02 6.16583169e-01 6.54986799e-01
-8.03741455e-01 6.71413422e-01 -1.08918250e+00 -6.01481438e-01
6.87373519e-01 5.34253478e-01 -7.03243196e-01 -2.95692176e-01
-5.32183826e-01 1.07992315e+00 4.19142544e-01 -1.64183661e-01
-5.42537808e-01 -5.05008042e-01 -9.43595946e-01 -2.59479195e-01
6.72923386e-01 -3.01499605e-01 1.35095012e+00 -9.30138290e-01
-1.21940410e+00 1.26855934e+00 -1.02093152e-03 -6.42060628e-03
5.69428086e-01 -1.02276079e-01 2.50318378e-01 4.10354316e-01
3.34150791e-01 8.82491529e-01 8.01918507e-01 -1.69897068e+00
-4.55867320e-01 -6.09112561e-01 2.03878298e-01 4.42803979e-01
-2.66180765e-02 -1.83174625e-01 -7.54968584e-01 -7.40763471e-02
5.22710025e-01 -1.07514894e+00 -5.38032711e-01 9.11847949e-02
-1.62087813e-01 -3.24748047e-02 1.13169348e+00 -3.30070108e-01
2.79278934e-01 -1.98566401e+00 -1.31848857e-01 4.26977962e-01
-5.16905636e-02 1.70730650e-01 1.13499239e-01 3.70234936e-01
4.28792834e-02 2.56773084e-02 -6.55274332e-01 -5.28963983e-01
-3.07715625e-01 6.35756373e-01 1.26567438e-01 5.49340844e-01
3.07242393e-01 5.46001613e-01 -8.69449377e-01 -8.99965823e-01
5.92392683e-01 3.71776402e-01 -4.18713003e-01 1.14623025e-01
-4.30039316e-01 1.01205075e+00 -3.97726148e-01 6.13173127e-01
5.69314659e-01 -4.99181271e-01 7.90158287e-02 -1.04423374e-01
-2.39760503e-01 2.31679246e-01 -1.18297887e+00 2.09586930e+00
-8.81852388e-01 4.01962847e-01 1.53673394e-02 -1.27108252e+00
8.66104186e-01 2.68536389e-01 7.10010827e-01 -3.34658146e-01
2.06938937e-01 4.87240195e-01 -4.12954211e-01 -4.85624552e-01
1.99704722e-01 -3.64157557e-01 1.87194236e-02 7.34729826e-01
1.83920026e-01 -8.61246288e-01 1.77596495e-01 -3.24619859e-02
9.65450943e-01 5.24354219e-01 1.82153016e-01 3.21225561e-02
4.23718810e-01 5.48734307e-01 4.42096621e-01 5.32124221e-01
1.27690271e-01 9.14896190e-01 2.39351079e-01 -3.38879764e-01
-1.35867047e+00 -7.01588213e-01 -3.49425286e-01 6.93007767e-01
4.61121559e-01 -2.41079181e-01 -7.14107811e-01 -1.03152764e+00
-8.24860763e-03 3.41204703e-01 -2.92968124e-01 4.21533853e-01
-3.69897634e-01 -3.13594580e-01 1.46626651e-01 6.18800402e-01
3.41436893e-01 -6.38751686e-01 -9.19310451e-01 9.28719938e-02
2.37024482e-02 -1.32406449e+00 1.16751092e-02 5.60058653e-01
-1.35429978e+00 -1.23643792e+00 -6.42918825e-01 -6.41578019e-01
1.33242691e+00 3.49211514e-01 1.23925662e+00 2.94182956e-01
-1.16553925e-01 7.30998456e-01 -4.86658156e-01 -6.36080861e-01
-8.04143772e-02 3.24209988e-01 -2.14410901e-01 -1.58757016e-01
3.19321096e-01 -7.09427714e-01 -4.36695457e-01 4.14964050e-01
-8.94434392e-01 1.29243881e-01 7.74821341e-01 7.64076352e-01
1.02639091e+00 4.46565077e-02 1.88228026e-01 -1.44206154e+00
-1.86775327e-01 -1.77386895e-01 -8.25283527e-01 -3.19951735e-02
-5.02258539e-01 -6.72785863e-02 2.54686952e-01 -1.89438209e-01
-9.91710126e-01 8.72547388e-01 -1.32067814e-01 -6.32852376e-01
-8.05234671e-01 4.05863702e-01 -6.27801493e-02 -4.17219371e-01
6.30017698e-01 -2.36638799e-01 2.94830352e-01 -7.24484444e-01
2.20372096e-01 7.17666686e-01 4.29278702e-01 -4.21536803e-01
1.01399720e+00 6.73379064e-01 1.31929517e-01 -9.73519146e-01
-1.01148188e+00 -9.76263344e-01 -1.54146469e+00 -2.71839231e-01
8.24516535e-01 -7.21989095e-01 -1.09953508e-01 7.70105794e-02
-1.02729321e+00 -3.72261614e-01 -4.23013717e-01 6.05536222e-01
-6.92933619e-01 4.08716112e-01 -1.01473201e-02 -7.70944297e-01
8.20061639e-02 -1.15954971e+00 1.46781802e+00 3.32122333e-02
-1.11260042e-01 -9.41558540e-01 -3.95692661e-02 7.25572705e-01
1.12180956e-01 3.40229690e-01 5.28011560e-01 -6.60018265e-01
-6.42552972e-01 -5.55419266e-01 -2.58640319e-01 3.59917134e-01
1.57961264e-01 -3.89167100e-01 -1.13360965e+00 -2.17947531e-02
-5.30165806e-02 -5.83587348e-01 4.90770549e-01 1.43120095e-01
1.17852938e+00 3.86445016e-01 -4.32780266e-01 3.11714262e-01
1.49865925e+00 1.26573574e-02 1.74055189e-01 3.17769647e-01
9.18486357e-01 9.11023498e-01 9.49377179e-01 1.74579825e-02
4.47033644e-01 6.50416195e-01 5.71730077e-01 -5.69414258e-01
2.32792944e-01 -1.78977430e-01 -2.81734467e-01 6.01211488e-01
-3.32764655e-01 1.30747557e-01 -1.33551872e+00 6.06464744e-01
-1.72309077e+00 -6.09469533e-01 -2.73936331e-01 2.33408403e+00
6.55529022e-01 1.62017822e-01 1.15703799e-01 3.95743459e-01
5.46152294e-01 -1.66145250e-01 -6.15707397e-01 1.61280204e-02
3.17209870e-01 4.11054254e-01 8.79570067e-01 3.37987185e-01
-1.41727781e+00 9.29796040e-01 6.07896233e+00 3.66330922e-01
-1.12879348e+00 3.41911346e-01 4.54194665e-01 5.42298183e-02
-1.05019055e-01 3.76034439e-01 -3.77971768e-01 2.08880261e-01
7.51243353e-01 5.58906972e-01 6.71677738e-02 1.01481640e+00
1.07124992e-01 -6.46937251e-01 -1.28666556e+00 1.13908672e+00
3.90876830e-02 -1.03435075e+00 -4.94371712e-01 1.94872051e-01
6.34376407e-01 2.75322884e-01 -4.60156262e-01 -8.18456113e-02
1.84516534e-01 -9.16903913e-01 4.90801603e-01 1.85757607e-01
9.09068465e-01 -5.99419713e-01 8.93516421e-01 6.85536265e-01
-9.68009174e-01 2.58905560e-01 -2.56483167e-01 8.17042142e-02
3.95575494e-01 9.13312137e-01 -7.69825220e-01 6.15909457e-01
8.09479535e-01 4.56199616e-01 -6.05909228e-01 1.09604251e+00
-2.10305169e-01 4.64039356e-01 -8.41867805e-01 3.17559898e-01
1.83205709e-01 -2.41146863e-01 3.01664501e-01 8.88529003e-01
2.42189020e-01 1.17960177e-01 5.30124187e-01 6.03821695e-01
8.10759068e-02 4.58021127e-02 -9.62438643e-01 1.37898967e-01
3.10051650e-01 1.39339173e+00 -1.33582318e+00 -2.54254073e-01
-5.02184451e-01 1.01174712e+00 3.12001228e-01 -2.37199273e-02
-4.20313597e-01 -2.66811341e-01 -2.90357843e-02 3.93523872e-01
4.15711403e-01 -6.75882876e-01 -2.84105480e-01 -9.59697664e-01
4.63054813e-02 -4.20810938e-01 7.20053613e-02 -7.04469264e-01
-1.10250080e+00 4.88427758e-01 2.72761226e-01 -1.43454051e+00
-3.41804534e-01 -4.85684663e-01 -3.26840729e-01 7.83566535e-01
-1.63097692e+00 -1.49538291e+00 -5.16269982e-01 3.81829947e-01
3.87040466e-01 2.91513801e-01 7.74357975e-01 4.43659604e-01
-1.18221119e-01 6.21361695e-02 -2.14528099e-01 2.52377447e-02
4.64517027e-01 -1.47142947e+00 9.01513174e-02 5.82486153e-01
3.88939232e-01 2.87485540e-01 4.29937899e-01 -5.90296745e-01
-1.22365713e+00 -9.68143463e-01 5.67885756e-01 -5.98757267e-01
2.30442822e-01 -4.48093057e-01 -9.71121013e-01 4.54300463e-01
-3.12062800e-01 4.22071159e-01 7.26345778e-01 1.98370010e-01
4.04489227e-03 8.64818543e-02 -1.15402591e+00 -4.56850566e-02
1.03005981e+00 -7.03738451e-01 -4.22834188e-01 5.57686448e-01
4.32574630e-01 -7.29578674e-01 -9.57808614e-01 4.74659830e-01
1.08113542e-01 -8.53752494e-01 8.18888545e-01 -6.30007982e-02
1.13641769e-01 -5.70481658e-01 3.59348394e-02 -1.19719088e+00
2.96387225e-01 -5.34618534e-02 4.15850222e-01 1.17149138e+00
5.45927584e-01 -3.87131572e-01 1.20180941e+00 5.13588369e-01
-2.03922644e-01 -4.60087955e-01 -6.94373906e-01 -8.23147535e-01
-1.55006140e-01 -5.65342188e-01 4.52719539e-01 1.02899075e+00
-5.33306241e-01 2.34741271e-01 1.52646184e-01 3.57730210e-01
7.89908648e-01 3.30438137e-01 1.14811122e+00 -1.69505131e+00
-1.24460943e-01 4.76182401e-02 -6.34580970e-01 -8.69066060e-01
6.36971891e-02 -8.83474529e-01 3.65913212e-01 -1.68288457e+00
-1.22657321e-01 -1.10217452e+00 1.90538973e-01 7.03486323e-01
1.51929110e-01 7.21621752e-01 -1.15288094e-01 4.77895319e-01
-4.73764598e-01 1.92506209e-01 1.09056592e+00 3.12981606e-02
-2.91198730e-01 4.20774408e-02 -1.89032674e-01 1.04290974e+00
7.89678872e-01 -6.11010849e-01 -6.49107873e-01 -6.02146745e-01
-7.80699775e-02 -1.44439861e-01 3.90185595e-01 -1.06338465e+00
2.68946499e-01 -9.79008153e-02 3.79201621e-01 -9.23363090e-01
5.25877118e-01 -1.26231229e+00 1.89347357e-01 -2.07940377e-02
1.13791041e-01 -1.80980816e-01 1.47504836e-01 6.10372186e-01
-3.18334311e-01 -7.06479073e-01 7.52549827e-01 -5.56509376e-01
-6.36634886e-01 3.52061749e-01 -5.27562276e-02 -2.54837304e-01
1.23829746e+00 -5.31332135e-01 3.24194312e-01 -1.19763643e-01
-8.72213185e-01 1.55086443e-01 9.87328172e-01 5.84601425e-03
3.94559264e-01 -9.79647219e-01 -1.41337186e-01 2.27687880e-01
2.64496952e-01 1.00165439e+00 6.57712296e-02 7.91644514e-01
-8.67203295e-01 2.55545199e-01 -1.03363916e-01 -1.19201422e+00
-1.39179313e+00 3.20183814e-01 2.47493953e-01 -1.51394457e-01
-7.65486062e-01 5.66462517e-01 -4.50289510e-02 -9.07936037e-01
3.50261889e-02 -9.92100462e-02 -1.85002953e-01 4.08169851e-02
-8.26838464e-02 -4.42651771e-02 5.14260352e-01 -7.38610387e-01
-3.29825193e-01 8.92161071e-01 -5.05760349e-02 -3.39302748e-01
1.60948753e+00 -1.34093598e-01 -4.84687835e-02 6.53793037e-01
1.07334888e+00 -1.88556552e-01 -1.33554041e+00 -2.71893829e-01
2.40805522e-01 -6.73043191e-01 3.90967757e-01 -4.25447375e-01
-9.74125862e-01 1.14613903e+00 4.69636768e-01 1.96237892e-01
1.00628603e+00 3.69467050e-01 6.00921929e-01 3.20905477e-01
7.43701518e-01 -1.10622966e+00 -1.87959131e-02 2.70170480e-01
5.20622432e-01 -1.65708518e+00 2.71758825e-01 -7.87013888e-01
-4.39912707e-01 9.95465040e-01 6.63609922e-01 2.05570236e-02
6.45012677e-01 1.26458272e-01 2.39832327e-01 -4.84963626e-01
-1.91992998e-01 -5.20444036e-01 2.61077285e-01 7.83544898e-01
2.81457633e-01 -2.83841550e-01 -2.13869344e-02 -8.24561492e-02
-8.64133239e-02 -7.05201179e-02 3.03856760e-01 1.53498030e+00
-3.65862936e-01 -1.50342977e+00 -4.32553351e-01 5.73820829e-01
-3.18061441e-01 1.95173666e-01 -4.72983897e-01 8.27061951e-01
2.62752980e-01 7.72635996e-01 5.00278294e-01 -1.66019499e-01
3.84618491e-01 5.72018176e-02 6.80447996e-01 -1.08411384e+00
-2.47968763e-01 2.07810283e-01 3.33918966e-02 -5.00535905e-01
-1.04568374e+00 -7.13833332e-01 -1.24199343e+00 8.64000022e-02
-7.50386119e-01 9.28434581e-02 1.10076785e+00 1.04403698e+00
2.60853440e-01 2.37863079e-01 6.07831001e-01 -1.36357939e+00
-9.60077718e-02 -6.44233346e-01 -3.81299436e-01 4.07744080e-01
1.35712415e-01 -1.10072279e+00 -2.47001380e-01 4.04277444e-01] | [8.110700607299805, -2.827766180038452] |
3ad1bec5-c79e-4bb1-a7c3-44d54ce245dd | a-streamlit-based-artificial-intelligence | 2211.12851 | null | https://arxiv.org/abs/2211.12851v1 | https://arxiv.org/pdf/2211.12851v1.pdf | A Streamlit-based Artificial Intelligence Trust Platform for Next-Generation Wireless Networks | With the rapid development and integration of artificial intelligence (AI) methods in next-generation networks (NextG), AI algorithms have provided significant advantages for NextG in terms of frequency spectrum usage, bandwidth, latency, and security. A key feature of NextG is the integration of AI, i.e., self-learning architecture based on self-supervised algorithms, to improve the performance of the network. A secure AI-powered structure is also expected to protect NextG networks against cyber-attacks. However, AI itself may be attacked, i.e., model poisoning targeted by attackers, and it results in cybersecurity violations. This paper proposes an AI trust platform using Streamlit for NextG networks that allows researchers to evaluate, defend, certify, and verify their AI models and applications against adversarial threats of evasion, poisoning, extraction, and interference. | ['O Gueler', 'U. Cali', 'S. Sarp', 'F. O. Catak', 'M. Kuzlu'] | 2022-10-25 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [-1.19049996e-01 1.80202946e-01 -6.13938808e-01 1.05219208e-01
1.98597848e-01 -7.52406120e-01 3.55437279e-01 4.71624956e-02
2.53474414e-01 7.17613161e-01 -2.58616626e-01 -5.99578798e-01
4.33370434e-02 -1.16967618e+00 -2.41877794e-01 -4.99689639e-01
-5.28226495e-01 9.07070190e-02 1.37447894e-01 -2.14215264e-01
1.22103482e-01 7.01898813e-01 -8.73353660e-01 -7.97004625e-02
7.85103798e-01 1.49288130e+00 -8.71970475e-01 2.18542397e-01
5.04507780e-01 7.24731803e-01 -1.10630369e+00 -6.73686624e-01
5.42105436e-01 -3.39179784e-01 -4.55113590e-01 -3.68423134e-01
-1.61730498e-01 -5.11706889e-01 -6.63537621e-01 1.17714560e+00
3.33393663e-02 -7.62923360e-01 2.09892228e-01 -2.36502695e+00
-3.21260959e-01 1.00479019e+00 -2.96431005e-01 -2.15450987e-01
1.48234352e-01 4.95915055e-01 6.31420970e-01 7.89277479e-02
2.35373691e-01 1.03492904e+00 3.46065968e-01 5.05537093e-01
-8.93550277e-01 -1.88525987e+00 -4.84041758e-02 1.98064998e-01
-1.16472101e+00 -5.04387736e-01 9.18651402e-01 -3.33612859e-02
6.32191122e-01 2.04151839e-01 1.11270881e+00 1.16702318e+00
6.62479043e-01 4.67622280e-01 5.64833164e-01 -3.07196170e-01
7.76866674e-01 3.26729089e-01 -1.40466452e-01 5.82681596e-01
9.41052437e-01 6.77986920e-01 -3.25500786e-01 -6.51826739e-01
3.90499800e-01 -1.36543885e-01 4.82371449e-02 -6.16936199e-02
-9.80269551e-01 6.69191599e-01 5.69356561e-01 2.46680066e-01
-6.91502571e-01 5.49824655e-01 4.87250447e-01 1.89219877e-01
1.75956428e-01 4.86544192e-01 -2.05309317e-01 -1.16933875e-01
-6.36236250e-01 -2.17869863e-01 9.24329102e-01 1.11598039e+00
5.30818701e-01 6.81864738e-01 4.78757560e-01 -2.39446297e-01
5.96392214e-01 8.92428100e-01 2.94247150e-01 -1.08228219e+00
6.23753816e-02 5.62347054e-01 -3.77605736e-01 -1.26808727e+00
-4.62323427e-01 -8.26112807e-01 -1.26248813e+00 2.60684460e-01
-8.31267461e-02 -6.59652352e-01 -2.97319680e-01 1.61948144e+00
8.37297216e-02 8.27354074e-01 4.04627651e-01 2.74132788e-01
2.79785693e-01 4.39916521e-01 8.87757018e-02 -2.77487069e-01
8.12251806e-01 -7.73414612e-01 -8.20925415e-01 -2.13844299e-01
4.20031220e-01 -8.54064748e-02 5.31941429e-02 7.12476254e-01
-8.00435126e-01 -1.71206415e-01 -1.77778316e+00 1.01104569e+00
-3.61938566e-01 -4.41599011e-01 9.75489497e-01 1.51614487e+00
-6.24245703e-01 2.68597931e-01 -3.93085659e-01 -3.31884444e-01
8.03083897e-01 6.17258847e-01 1.79296002e-01 2.98365176e-01
-1.66995990e+00 3.86066109e-01 6.74222827e-01 -4.26323444e-01
-1.01017952e+00 -6.48870349e-01 -4.47393954e-01 2.57429004e-01
2.29543224e-01 -2.74631679e-01 8.31640363e-01 -8.09676290e-01
-1.41431940e+00 4.20186408e-02 8.12083960e-01 -9.68434930e-01
-4.13009822e-02 4.35636714e-02 -1.32441556e+00 5.28715491e-01
-2.14483052e-01 1.36983320e-01 8.31733763e-01 -9.96484697e-01
-6.54293835e-01 -1.86723068e-01 1.14929266e-01 -5.74530482e-01
-8.78894031e-01 4.64369506e-02 4.61549371e-01 -5.72413564e-01
-1.24310426e-01 -1.21746659e+00 -1.88296586e-01 -1.13614075e-01
-5.34023345e-01 2.27649942e-01 1.34186721e+00 -7.81890452e-02
1.38447642e+00 -2.09528303e+00 -7.40061939e-01 9.16043758e-01
4.16659296e-01 7.97004879e-01 1.33730695e-01 6.03558242e-01
5.63845634e-02 7.34968722e-01 5.25264323e-01 7.09250867e-01
-1.94794565e-01 -1.81854233e-01 -4.84529108e-01 3.92354161e-01
-1.44226894e-01 6.82990789e-01 -8.11520696e-01 -1.55551091e-01
1.15276143e-01 -1.07716136e-01 -4.62349772e-01 -4.43728082e-02
-2.08515331e-01 -1.00015819e-01 -7.83495247e-01 8.11975241e-01
5.94886720e-01 -2.66245365e-01 2.70844579e-01 -5.87504148e-01
4.52772938e-02 -2.74973035e-01 -8.80668283e-01 1.04705620e+00
-2.57313281e-01 2.44782805e-01 9.78509262e-02 -6.51114762e-01
1.06522620e+00 7.91732252e-01 9.01138961e-01 -6.07541740e-01
3.93222123e-01 2.81472176e-01 2.08296537e-01 2.56740674e-02
-4.80798393e-01 3.97613615e-01 -3.77311826e-01 1.09388220e+00
-8.06395859e-02 1.16451129e-01 -7.13559762e-02 5.41111231e-01
1.32206166e+00 -6.06222630e-01 3.62457871e-01 -1.03146598e-01
6.94293737e-01 -9.96145383e-02 7.26227820e-01 8.40714157e-01
-4.48453218e-01 -5.71615994e-01 2.74464905e-01 -3.05448264e-01
-6.08774304e-01 -1.00980103e+00 2.65423805e-01 5.77407740e-02
4.51596022e-01 -5.33417940e-01 -7.71439135e-01 -8.90201151e-01
3.37294862e-02 9.08690035e-01 2.08153829e-01 -1.07401490e+00
-1.98113546e-01 -4.31217015e-01 1.34013462e+00 2.44227394e-01
1.24254394e+00 -9.45930660e-01 -4.97152805e-01 5.72021902e-01
2.42637500e-01 -1.44263005e+00 -1.46425813e-01 -1.24034479e-01
-3.48639578e-01 -1.16732442e+00 2.01085538e-01 -1.24491826e-01
1.94962516e-01 2.68200845e-01 5.54875076e-01 3.87887686e-01
-7.28996843e-02 3.07850659e-01 -2.72267759e-01 -8.96613240e-01
-8.65633667e-01 2.55395416e-02 6.88022375e-01 1.57725826e-01
2.97034115e-01 -1.04178488e+00 -4.37380046e-01 4.46366787e-01
-6.94809496e-01 -2.76965469e-01 6.88395381e-01 2.17330635e-01
-3.58938426e-01 8.13935697e-01 1.15452147e+00 -7.06687987e-01
4.83196735e-01 -5.90752065e-01 -8.28584433e-01 5.00367284e-01
-1.17808449e+00 -4.14513737e-01 1.15708876e+00 -4.02321488e-01
-4.85065341e-01 -4.21609461e-01 4.48031664e-01 -1.61290333e-01
3.11919060e-02 3.28360319e-01 -8.19780886e-01 -7.59617686e-01
7.36847460e-01 -7.55028725e-02 2.94303179e-01 3.75501961e-01
1.79303974e-01 9.12384212e-01 2.87946880e-01 -2.12683693e-01
1.76401436e+00 4.99382019e-01 4.16097552e-01 -6.21899664e-01
-6.50309920e-01 4.15707648e-01 1.59570590e-01 -6.34944081e-01
4.46801007e-01 -8.62139583e-01 -1.11825109e+00 8.06949675e-01
-1.02135098e+00 1.71819046e-01 3.07636380e-01 6.04064345e-01
-1.18038476e-01 1.93717822e-01 -6.54034138e-01 -6.35906875e-01
-9.64168787e-01 -6.73616290e-01 -2.15308040e-01 6.84598267e-01
-2.35077798e-01 -6.80228353e-01 -4.90449607e-01 5.76980889e-01
8.54137659e-01 2.90623069e-01 9.96127725e-01 -9.16701674e-01
-7.45157778e-01 -5.41969836e-01 -2.81193405e-01 4.29929256e-01
9.71764550e-02 2.17221200e-01 -6.35208964e-01 -3.46062660e-01
1.02183213e-02 -2.59733230e-01 -2.18683735e-01 -1.33219082e-02
1.02464080e+00 -9.64292884e-01 -4.36811268e-01 7.85908341e-01
1.18407166e+00 9.45061266e-01 7.39199996e-01 1.41081989e-01
3.80143642e-01 2.63131410e-01 7.40487933e-01 5.76470852e-01
-1.61003023e-01 2.15644628e-01 9.78695154e-01 -5.99301755e-02
4.62072670e-01 -2.67115772e-01 3.38663191e-01 1.18446253e-01
3.22326541e-01 -6.44852042e-01 -8.10312927e-01 -3.24548274e-01
-1.63132632e+00 -1.20785642e+00 2.80792713e-01 1.78933167e+00
1.10541098e-01 8.50552201e-01 2.02584282e-01 5.58316946e-01
6.28159404e-01 -9.82282311e-02 -7.73734391e-01 -2.51498669e-01
7.87132159e-02 1.25555634e-01 8.34418476e-01 -5.03248630e-05
-9.17966008e-01 9.99506474e-01 5.85278845e+00 8.04125309e-01
-1.31800485e+00 -1.89684734e-01 5.70130765e-01 1.05631620e-01
-6.53270110e-02 2.84332424e-01 -4.41932708e-01 6.75118983e-01
1.21738338e+00 -8.19509566e-01 4.59985107e-01 8.47026229e-01
3.07649314e-01 4.03108299e-01 -6.93225622e-01 6.64609849e-01
-6.61370829e-02 -1.58156335e+00 3.18528228e-02 3.65045160e-01
4.65298533e-01 -2.15060681e-01 -2.57456988e-01 1.68997616e-01
5.12654662e-01 -7.40705013e-01 4.18576807e-01 1.58685371e-01
5.02955437e-01 -1.52107000e+00 9.19671297e-01 2.11655602e-01
-8.89093101e-01 -5.16094208e-01 1.91122845e-01 2.37049200e-02
1.01557665e-01 6.94572508e-01 -5.31280577e-01 6.70018196e-01
5.47429323e-01 5.65074980e-01 -4.62218046e-01 8.42770398e-01
-3.58922929e-01 9.94144022e-01 -4.84198838e-01 -3.74620140e-01
7.21554533e-02 -2.50454992e-01 5.73360503e-01 4.19356972e-01
2.41802782e-01 1.18401326e-01 5.10737479e-01 8.39085221e-01
-1.46437988e-01 -4.23866242e-01 -5.53806067e-01 -3.86569589e-01
1.29601812e+00 1.34735239e+00 -4.63037223e-01 -1.88575298e-01
-2.50957012e-01 5.32502651e-01 -5.47249079e-01 4.23456162e-01
-9.48692083e-01 -7.83029795e-01 7.65234232e-01 2.05054477e-01
-3.88486773e-01 -1.53928801e-01 -3.55570555e-01 -7.01946497e-01
-2.93001682e-01 -1.19970071e+00 4.73356217e-01 -4.85704005e-01
-1.06469381e+00 5.66895127e-01 -1.66530266e-01 -1.27177465e+00
-1.80471286e-01 5.43718319e-03 -7.62587249e-01 1.26311913e-01
-9.53436434e-01 -1.37032127e+00 -2.68632680e-01 7.69029081e-01
-3.82619739e-01 -1.05898070e+00 1.12346435e+00 2.62197942e-01
-7.66283989e-01 1.00414598e+00 -2.95116603e-01 2.71520108e-01
2.25426838e-01 -4.40581053e-01 2.18544245e-01 1.11238611e+00
-8.36156085e-02 2.70560235e-01 3.58433694e-01 -8.02080452e-01
-1.48459196e+00 -1.00640380e+00 9.08161402e-02 5.18400252e-01
9.92248535e-01 -4.37720656e-01 -1.61240160e-01 5.06805241e-01
1.01362228e-01 1.55059993e-01 9.00020063e-01 -3.03975672e-01
-3.64395738e-01 -8.85607660e-01 -1.59406769e+00 8.88387322e-01
7.40922868e-01 -3.62962991e-01 4.69023168e-01 2.26018563e-01
8.94430995e-01 3.44407260e-01 -7.26876855e-01 3.72486919e-01
5.33031046e-01 -8.39445353e-01 8.64968359e-01 -5.00939250e-01
-2.59679526e-01 -2.52327591e-01 1.46461248e-01 -1.15884912e+00
-3.46901476e-01 -1.25400698e+00 -7.02379584e-01 1.46476948e+00
3.06493729e-01 -1.06562722e+00 1.17057490e+00 4.07321751e-01
3.06523949e-01 -1.57728001e-01 -8.50620091e-01 -1.00064409e+00
-3.24262291e-01 -3.31958145e-01 1.16752136e+00 9.29874003e-01
4.72993970e-01 3.47201288e-01 -5.22002578e-01 4.42728519e-01
1.14175880e+00 -4.01878774e-01 8.65023971e-01 -1.51014161e+00
6.92569911e-02 -4.21166778e-01 -7.79108942e-01 -3.39135319e-01
2.29683608e-01 -6.62314415e-01 -8.19698095e-01 -8.46153021e-01
-4.64035362e-01 -6.32802486e-01 -4.67208534e-01 6.20716929e-01
6.07717216e-01 1.87348261e-01 3.36473167e-01 1.48533955e-01
-4.03988600e-01 3.37543547e-01 6.59819067e-01 -3.81496936e-01
-3.08499094e-02 4.22333091e-01 -8.33050728e-01 7.18730688e-01
1.35192859e+00 -4.91407424e-01 -7.87922323e-01 3.84479702e-01
4.82589193e-03 4.41228747e-01 3.56565386e-01 -1.62762117e+00
6.31891072e-01 -4.02247369e-01 6.66544586e-02 -3.48775446e-01
-7.50167947e-03 -1.59029746e+00 2.16612607e-01 1.05820918e+00
1.60859600e-01 -2.04535976e-01 -2.86292974e-02 3.85884017e-01
1.15157396e-01 6.20868877e-02 6.09428346e-01 3.41285169e-01
-1.55262381e-01 7.26790845e-01 -8.41535926e-01 -2.72802383e-01
1.52487731e+00 -1.16592087e-01 -6.13715589e-01 -7.65840411e-01
-8.81258100e-02 4.38318938e-01 2.46348694e-01 3.59312028e-01
7.26821542e-01 -1.34158647e+00 -6.06635928e-01 4.68466341e-01
1.35115795e-02 -7.25811541e-01 -2.53045559e-01 3.08422804e-01
-4.43802297e-01 2.83268392e-01 -4.56572503e-01 -3.57320271e-02
-1.23732948e+00 6.44125819e-01 2.97932863e-01 -1.65263727e-01
-3.08996916e-01 2.89133489e-02 -5.55100024e-01 1.00638367e-01
5.52960515e-01 5.71003437e-01 7.59167895e-02 -4.04156327e-01
4.29006875e-01 5.11328459e-01 -2.58366048e-01 6.32853433e-02
-6.20800674e-01 2.16250420e-02 -2.66311228e-01 -1.17036216e-01
7.78463840e-01 9.28745195e-02 -9.84079689e-02 -2.11888671e-01
6.14713728e-01 -4.19532321e-02 -3.91768634e-01 -7.17326030e-02
1.03360854e-01 -1.95923656e-01 4.74365503e-01 -1.15149093e+00
-1.49220467e+00 1.73719779e-01 4.70171452e-01 6.22759759e-01
1.24488699e+00 -4.98442590e-01 1.14560068e+00 5.18621504e-01
9.05322611e-01 -8.71052027e-01 4.25214648e-01 1.47763088e-01
1.30489856e-01 -8.41059983e-01 2.70859987e-01 -6.16629899e-01
-1.70067146e-01 1.25841928e+00 1.07670105e+00 -5.69530576e-03
1.36580586e+00 6.09770060e-01 1.52552053e-01 -3.73526096e-01
-6.62421227e-01 6.47362113e-01 -4.03273702e-01 8.79893839e-01
-1.94105387e-01 -5.22014424e-02 3.99029590e-02 6.87005222e-01
-2.53376424e-01 -1.59986049e-01 8.54425550e-01 7.01223135e-01
-3.93135369e-01 -1.24956071e+00 -3.52809548e-01 4.41715509e-01
-3.71626288e-01 1.47316873e-01 -7.01566696e-01 6.58053160e-01
9.98907909e-02 1.43301678e+00 -3.57818276e-01 -1.04253507e+00
-1.81003362e-01 -5.55454493e-01 -1.79533944e-01 -1.68647349e-01
-7.83289433e-01 -1.78676516e-01 2.91492075e-01 -5.95683992e-01
-1.21394835e-01 -2.96876073e-01 -1.34886503e+00 -7.51047075e-01
-5.13440430e-01 5.14070570e-01 6.78192079e-01 7.15686619e-01
3.87371331e-01 3.60849559e-01 1.55678427e+00 1.18860930e-01
-3.65062296e-01 -4.04706240e-01 -7.05227435e-01 -5.50053753e-02
-1.55919284e-01 -2.40181312e-01 -7.21348405e-01 -7.75219560e-01] | [5.464162349700928, 7.273461818695068] |
7d4c36a6-ff8f-48fd-88bc-f7247b285aae | integrating-unsupervised-data-generation-into | 2107.08772 | null | https://arxiv.org/abs/2107.08772v1 | https://arxiv.org/pdf/2107.08772v1.pdf | Integrating Unsupervised Data Generation into Self-Supervised Neural Machine Translation for Low-Resource Languages | For most language combinations, parallel data is either scarce or simply unavailable. To address this, unsupervised machine translation (UMT) exploits large amounts of monolingual data by using synthetic data generation techniques such as back-translation and noising, while self-supervised NMT (SSNMT) identifies parallel sentences in smaller comparable data and trains on them. To date, the inclusion of UMT data generation techniques in SSNMT has not been investigated. We show that including UMT techniques into SSNMT significantly outperforms SSNMT and UMT on all tested language pairs, with improvements of up to +4.3 BLEU, +50.8 BLEU, +51.5 over SSNMT, statistical UMT and hybrid UMT, respectively, on Afrikaans to English. We further show that the combination of multilingual denoising autoencoding, SSNMT with backtranslation and bilingual finetuning enables us to learn machine translation even for distant language pairs for which only small amounts of monolingual data are available, e.g. yielding BLEU scores of 11.6 (English to Swahili). | ['Cristina España-Bonet', 'Josef van Genabith', 'Dietrich Klakow', 'Dana Ruiter'] | 2021-07-19 | null | https://aclanthology.org/2021.mtsummit-research.7 | https://aclanthology.org/2021.mtsummit-research.7.pdf | mtsummit-2021-8 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 3.34164143e-01 1.76368237e-01 -1.37671694e-01 -2.66558290e-01
-1.42514503e+00 -1.11467946e+00 1.17588556e+00 3.04925367e-02
-4.51449335e-01 1.37952340e+00 1.71622396e-01 -9.05166805e-01
2.53267825e-01 -4.13701415e-01 -9.72307265e-01 -3.96103710e-01
2.90663213e-01 1.02170897e+00 -2.90427178e-01 -5.32433927e-01
1.64223816e-02 7.50928819e-02 -8.79302084e-01 1.96227267e-01
1.28133070e+00 9.30120200e-02 3.26581299e-01 8.27093959e-01
-2.88249135e-01 2.08427832e-01 -5.62280834e-01 -6.21913314e-01
5.65725625e-01 -8.68924558e-01 -7.69981444e-01 -1.14168741e-01
5.79938769e-01 -1.05198741e-01 -5.63118830e-02 7.98827171e-01
7.04995871e-01 -3.66133660e-01 5.25133014e-01 -7.03868806e-01
-6.25410855e-01 1.15807521e+00 -4.58460689e-01 1.03751458e-01
5.07401109e-01 7.47788697e-02 8.48438323e-01 -1.12364435e+00
9.42448139e-01 1.26117373e+00 6.09683752e-01 4.16149616e-01
-1.40038490e+00 -4.84309554e-01 -5.17596185e-01 -3.83284718e-01
-1.17143238e+00 -7.05640614e-01 1.94290876e-01 -1.84380949e-01
1.31616223e+00 3.27664584e-01 4.81752723e-01 1.32984912e+00
3.93974006e-01 7.20481813e-01 1.71941543e+00 -1.00803018e+00
-9.04106796e-02 3.07331681e-01 -3.71358901e-01 3.50642473e-01
2.09471285e-01 7.13105202e-02 -5.23066580e-01 -6.05169451e-03
5.71809471e-01 -5.57168126e-01 1.08574465e-01 2.80916989e-01
-2.02334571e+00 6.31866753e-01 -2.12597296e-01 5.14349401e-01
-2.30728835e-01 -2.29379699e-01 5.59095442e-01 1.04245126e+00
7.92569876e-01 6.10682249e-01 -7.85046816e-01 -3.45797449e-01
-1.11697471e+00 6.69139624e-02 8.45035136e-01 1.39381647e+00
8.90084922e-01 1.89602792e-01 -7.30368719e-02 1.00977194e+00
-1.49460420e-01 1.19640541e+00 7.68803060e-01 -6.20745063e-01
9.63812590e-01 2.64414042e-01 -7.23859668e-02 -2.93632209e-01
-2.19977468e-01 -4.08953995e-01 -6.74685657e-01 -4.16847259e-01
4.49070454e-01 -4.83394414e-01 -8.49280894e-01 1.74563611e+00
8.41732845e-02 -3.89340073e-01 7.20018566e-01 6.15891933e-01
4.03464317e-01 8.36375713e-01 -3.04641932e-01 -5.96147895e-01
1.14639664e+00 -9.83160317e-01 -7.01711357e-01 -3.24928254e-01
9.98338282e-01 -1.53060198e+00 1.08193874e+00 1.23637661e-01
-1.22397089e+00 -5.48946679e-01 -8.37471247e-01 2.92112082e-02
-3.17048281e-01 2.32732445e-01 4.45130616e-01 6.19588196e-01
-1.35290682e+00 4.87691849e-01 -8.11245680e-01 -9.00716066e-01
-1.42071530e-01 4.61633354e-01 -6.53030396e-01 -2.73671776e-01
-1.47400260e+00 1.22992158e+00 4.41082031e-01 -1.11075900e-01
-7.22620904e-01 -3.18149686e-01 -8.13855231e-01 -6.76743746e-01
-2.80056335e-02 -6.22396290e-01 1.23949289e+00 -1.14580405e+00
-1.69196737e+00 8.94836962e-01 -1.97987363e-01 -6.85309112e-01
8.90596509e-01 -9.25980806e-02 -4.87698555e-01 -2.00842991e-01
4.51047868e-01 9.25625086e-01 4.67800081e-01 -9.71706808e-01
-4.44407463e-01 -1.89692616e-01 -4.20554966e-01 4.19485420e-01
-2.08883643e-01 3.98895770e-01 -1.69916719e-01 -8.30541968e-01
3.76861840e-02 -1.23435998e+00 -1.89753607e-01 -9.01423931e-01
-5.70812285e-01 9.81716719e-03 5.01221895e-01 -1.10910559e+00
9.27944124e-01 -1.45250738e+00 2.35541582e-01 1.38410190e-02
-3.41055453e-01 3.74436587e-01 -5.13662815e-01 9.55403090e-01
1.34300649e-01 1.26398459e-01 -4.95892763e-01 -4.00874883e-01
-4.73739281e-02 5.82735956e-01 -2.29208507e-02 2.61698246e-01
2.87124395e-01 1.21230793e+00 -9.52754915e-01 -3.91978294e-01
6.33936003e-02 1.92702606e-01 -2.15686530e-01 4.37640771e-02
-1.67679161e-01 7.93745935e-01 -3.29559320e-03 5.63610435e-01
5.38022995e-01 2.68533021e-01 5.99263132e-01 4.08182859e-01
-3.98369014e-01 8.23298752e-01 -6.26051068e-01 1.95528507e+00
-7.90305972e-01 7.60509431e-01 -3.87954563e-01 -5.57004213e-01
1.02477086e+00 5.05895972e-01 1.65905133e-02 -8.74444664e-01
-7.91222975e-02 1.02719200e+00 3.05661887e-01 -2.51467168e-01
5.56958079e-01 -3.18204075e-01 -2.51834601e-01 7.96579301e-01
2.38168746e-01 -2.61781871e-01 5.76645374e-01 1.18706755e-01
1.01681554e+00 2.66086638e-01 3.12182009e-01 -5.84026396e-01
4.06853735e-01 2.59254098e-01 4.35581058e-01 5.53539991e-01
1.56021938e-01 6.28686368e-01 1.37027621e-01 -5.30509651e-02
-1.69085205e+00 -1.00329602e+00 -5.68043441e-02 8.53892028e-01
-5.12647748e-01 -4.32431430e-01 -1.01725709e+00 -6.87132418e-01
-3.05543900e-01 7.56422758e-01 -1.54273406e-01 9.44681093e-02
-1.03233612e+00 -9.79793906e-01 8.39053929e-01 9.22576785e-02
4.31211472e-01 -1.06680131e+00 2.38136023e-01 3.90829355e-01
-5.46794653e-01 -1.20949566e+00 -5.44082701e-01 1.40299618e-01
-1.11818588e+00 -4.14713353e-01 -9.69394922e-01 -8.37158620e-01
5.03814578e-01 6.67158738e-02 1.27310741e+00 -4.75240827e-01
3.33301187e-01 -1.87220454e-01 -3.69351894e-01 -1.30419239e-01
-1.29785430e+00 6.47780061e-01 4.64209795e-01 -2.66797423e-01
3.83220404e-01 -6.27158225e-01 -5.68489805e-02 1.62767127e-01
-6.88133419e-01 3.53650898e-01 1.09452319e+00 9.36647534e-01
4.35832709e-01 -5.42326391e-01 7.51849830e-01 -1.11727870e+00
4.81451273e-01 -4.51252788e-01 -3.32032144e-01 3.94285172e-01
-7.88778841e-01 2.37722605e-01 7.46353507e-01 -5.90463161e-01
-8.55169535e-01 -2.24751353e-01 -1.31595343e-01 -4.72247005e-02
-1.18022479e-01 3.88911754e-01 -3.77341770e-02 1.87183022e-01
7.73132801e-01 5.94094932e-01 -3.86997610e-02 -5.21677792e-01
4.25335199e-01 1.07055604e+00 3.88201892e-01 -5.61659396e-01
8.86541486e-01 -5.02427593e-02 -3.54050010e-01 -6.70924544e-01
-3.64932567e-01 -6.69764727e-02 -8.74751925e-01 1.51044279e-01
6.15983427e-01 -1.04549503e+00 2.30774909e-01 4.07130867e-01
-1.15843952e+00 -5.07501006e-01 -1.97158173e-01 7.51840234e-01
-6.52103126e-01 3.05873811e-01 -8.22010636e-01 -5.18187821e-01
-6.38489842e-01 -1.20281863e+00 1.10551643e+00 -3.01596433e-01
-6.12029493e-01 -1.07918882e+00 3.01911831e-01 4.99820024e-01
4.01114762e-01 8.96222666e-02 9.04498577e-01 -8.67274165e-01
-2.80567765e-01 -8.65567550e-02 3.34562175e-02 4.45783317e-01
4.73460227e-01 -2.48371273e-01 -6.10608995e-01 -6.04620755e-01
-3.34949166e-01 -3.51179659e-01 3.59312385e-01 -3.38483080e-02
-6.97924644e-02 -4.41791952e-01 -2.53620520e-02 2.66795963e-01
1.20145380e+00 1.28987506e-01 3.94899845e-01 4.17226881e-01
7.38988161e-01 7.24446535e-01 4.33181852e-01 5.15340688e-03
6.03703856e-01 6.67339027e-01 -2.93702155e-01 -1.70222402e-01
-3.49913716e-01 -3.68445128e-01 8.22916269e-01 1.72797656e+00
-5.66804362e-03 -1.92300871e-01 -1.17353642e+00 6.79213464e-01
-1.53647435e+00 -5.30248523e-01 -4.31394458e-01 2.30864358e+00
1.30791247e+00 1.56445041e-01 -1.58838034e-02 6.98220134e-02
7.30902851e-01 -2.14136750e-01 -1.49121478e-01 -6.28097057e-01
-5.08920491e-01 3.54854137e-01 8.19761157e-01 7.23688483e-01
-6.75349951e-01 1.39703465e+00 6.14629602e+00 7.94974685e-01
-1.04722214e+00 4.35642153e-01 5.17939150e-01 7.77099356e-02
-5.54088295e-01 3.69685858e-01 -6.19298577e-01 4.23345119e-01
1.54824317e+00 -3.35411280e-02 7.32245028e-01 2.06681371e-01
3.86290014e-01 8.54976401e-02 -1.07485950e+00 7.75256753e-01
-1.19726723e-02 -1.00025105e+00 2.54783481e-01 9.75206196e-02
1.33517480e+00 4.61653024e-01 -1.45328641e-01 3.78091186e-01
4.79886204e-01 -8.14253151e-01 8.05989027e-01 1.83719218e-01
1.08524084e+00 -8.87375236e-01 7.84425437e-01 6.19785130e-01
-7.71178663e-01 4.98222560e-01 -3.44737291e-01 -3.91656533e-02
6.71239421e-02 5.80892682e-01 -1.14789474e+00 9.21008170e-01
1.60551026e-01 6.53995693e-01 -4.73489851e-01 2.36677036e-01
-3.27685058e-01 8.56090784e-01 -4.26243216e-01 1.13958471e-01
4.18002725e-01 -3.65793109e-01 6.57583296e-01 1.42917395e+00
6.92361712e-01 -5.06149650e-01 1.72815457e-01 4.13820714e-01
-1.49582550e-01 5.81964672e-01 -7.56668508e-01 -3.79259497e-01
4.46262211e-01 1.00352836e+00 -5.19492507e-01 -4.97625947e-01
-2.96204507e-01 1.44391537e+00 3.20839643e-01 2.73065776e-01
-2.66290933e-01 -2.64246523e-01 4.50417221e-01 3.31895649e-02
-1.58348367e-01 -4.36709046e-01 -2.60746419e-01 -1.33764160e+00
9.65145528e-02 -1.38793707e+00 -4.14845012e-02 -5.51199913e-01
-1.14168429e+00 9.94431555e-01 -1.96072623e-01 -1.29482412e+00
-8.68610322e-01 -3.92855972e-01 -3.54394466e-01 1.25658560e+00
-1.16902912e+00 -1.36355305e+00 5.05798995e-01 1.62913680e-01
8.32248390e-01 -6.08299553e-01 8.98654878e-01 3.02437067e-01
-4.68467146e-01 7.93784916e-01 5.74666202e-01 3.79805677e-02
1.08276749e+00 -1.35004961e+00 1.17603898e+00 9.42368746e-01
4.28972006e-01 8.96860957e-01 7.00899541e-01 -7.81727493e-01
-1.52790070e+00 -1.18135691e+00 1.92451286e+00 -6.48627102e-01
7.53699124e-01 -7.46085346e-01 -5.91669917e-01 8.19081247e-01
7.15023398e-01 -5.00719190e-01 5.64908624e-01 -8.36532265e-02
-1.96800724e-01 3.54223512e-02 -9.26561892e-01 7.90839553e-01
1.04768145e+00 -6.47321522e-01 -5.24891615e-01 5.99911034e-01
6.98499143e-01 -2.99418390e-01 -1.08485448e+00 3.38132083e-01
4.87362325e-01 -6.86449409e-01 5.94644427e-01 -4.93775427e-01
4.82209682e-01 -2.05509797e-01 -4.24801797e-01 -1.68349993e+00
1.72915757e-01 -1.13348711e+00 3.00424725e-01 1.47022605e+00
1.02417386e+00 -8.92114878e-01 3.11485469e-01 -2.55334586e-01
-2.56299108e-01 -4.26784068e-01 -9.08675969e-01 -1.02878463e+00
4.87863779e-01 -3.00994456e-01 6.08810127e-01 1.19107139e+00
-3.82224191e-03 6.26823127e-01 -6.48963869e-01 -2.22496778e-01
4.67878461e-01 6.25874177e-02 9.34135675e-01 -7.09831655e-01
-3.89404565e-01 -2.07691386e-01 2.79024336e-02 -9.31310952e-01
2.85413474e-01 -1.30863774e+00 -1.07725970e-02 -1.54589975e+00
2.13575169e-01 -4.61922675e-01 1.06755815e-01 3.74209315e-01
-1.10633411e-01 6.55235052e-01 8.07383135e-02 4.19455349e-01
-5.92872649e-02 2.65316933e-01 1.35039198e+00 -6.95091933e-02
-1.65070102e-01 -1.38946399e-01 -4.72671568e-01 3.34381104e-01
1.00912678e+00 -5.76619208e-01 -2.76519328e-01 -7.66741872e-01
7.46361464e-02 2.25696981e-01 -2.73400158e-01 -7.85600841e-01
-2.05651656e-01 -1.89206016e-03 2.79895097e-01 -4.03139770e-01
-9.40501094e-02 -3.44864547e-01 3.26916933e-01 3.98560315e-01
-3.23260456e-01 6.83846593e-01 2.29944155e-01 8.53018165e-02
-2.31392890e-01 2.51200683e-02 4.51233327e-01 -3.33368450e-01
-2.56604671e-01 -1.21939853e-01 -6.13300622e-01 9.29426625e-02
6.25796258e-01 -1.37534365e-01 -4.80847843e-02 -3.32714111e-01
-4.02483046e-01 8.96805748e-02 5.77242970e-01 4.77506250e-01
1.38889223e-01 -1.52991617e+00 -1.36109686e+00 4.40348923e-01
1.28149003e-01 -3.55867505e-01 -2.63390332e-01 1.06477880e+00
-5.76897502e-01 7.53912985e-01 -2.06320286e-01 -5.58227599e-01
-1.06109571e+00 1.39120042e-01 2.42870878e-02 -4.49234158e-01
-2.54001766e-01 4.21245605e-01 -2.95880526e-01 -1.30159926e+00
-4.36902732e-01 -9.36323851e-02 3.50562483e-01 -8.50836635e-02
8.91597345e-02 3.21320176e-01 4.28894430e-01 -9.18794394e-01
-1.45146415e-01 4.00041968e-01 -2.18903840e-01 -6.86046541e-01
9.14256454e-01 -3.86145771e-01 -3.95727187e-01 6.11895978e-01
1.17894697e+00 3.35736871e-01 -6.59763873e-01 -3.36077809e-01
2.78821826e-01 -1.55523270e-01 -4.26793158e-01 -1.01856780e+00
-4.02039230e-01 7.43069530e-01 3.69283915e-01 -1.86707288e-01
9.71816242e-01 -1.55587345e-01 1.14998817e+00 5.46576202e-01
6.09971523e-01 -1.04668021e+00 -4.50360030e-01 8.93289328e-01
6.79770529e-01 -1.34350944e+00 -4.54577774e-01 -7.44456127e-02
-4.76453245e-01 9.70787346e-01 1.92983598e-01 1.71881273e-01
1.68763734e-02 3.49648505e-01 5.08400738e-01 4.61795419e-01
-7.95163751e-01 -1.47939786e-01 1.63189039e-01 4.28132653e-01
8.48301649e-01 2.76778042e-01 -8.56253505e-01 -1.00673303e-01
-7.16626406e-01 -4.10030156e-01 4.17423904e-01 9.15337026e-01
-9.79597792e-02 -1.86225140e+00 -4.80083108e-01 3.62594604e-01
-6.01262033e-01 -6.86555028e-01 -5.61318874e-01 8.80362034e-01
1.21828709e-02 1.01508272e+00 3.21249776e-02 -4.61811274e-01
1.01134457e-01 2.88042068e-01 6.30657673e-01 -6.71653211e-01
-7.87267685e-01 3.66738766e-01 5.26853323e-01 6.42021745e-02
-4.26164001e-01 -8.79013419e-01 -8.73468399e-01 -5.57110369e-01
-2.51797199e-01 3.39759260e-01 7.38499463e-01 1.15913308e+00
1.24843888e-01 1.20508216e-01 6.78523898e-01 -7.39191175e-01
-5.62950492e-01 -1.51687503e+00 -1.25414431e-01 7.41647333e-02
1.25329137e-01 3.01958378e-02 -7.97688290e-02 1.42943233e-01] | [11.581031799316406, 10.369853973388672] |
9812f8e8-9ef5-4a14-bfe2-91c59a789a3c | weavenet-a-differentiable-solver-for-non | null | null | https://openreview.net/forum?id=ktHKpsbsxx | https://openreview.net/pdf?id=ktHKpsbsxx | WeaveNet: A Differentiable Solver for Non-linear Assignment Problems | Assignment, a task to match a limited number of elements, is a fundamental problem in informatics. Traditionally, non-linear assignment is discussed as a combinatorial optimization problem with its calculation complexity. On the other hand, it is often a sub-problem of image processing tasks, such as 3D point cloud matching. This paper proposes WeaveNet, a differentiable solver for diverse non-linear assignment problems. Traditional graph convolutional networks (GCNs) suffer from an over-smoothing problem when characterizing nodes with their relationship. WeaveNet overcomes this problem by forwarding edge-wise features at each layer rather than aggregated node features.
To deal with the exponentially large input space of combinatorial optimization problems, we designed WeaveNet to be highly parameter efficient while characterizing edges through stacked set-encoder with cross-concatenation operations. Experimental results show that WeaveNet approximates two strongly NP-hard variants of stable matching in a comparative performance with the gold standard hand-crafted algorithms under the limited size of problem instances. We have also confirmed that it can boost 3D point cloud matching performance significantly. | ['Yoshitaka Ushiku', 'Naoya Chiba', 'rintaro yanagi', 'Jiaxin Ma', 'Atsushi Hashimoto', 'Shusaku Sone'] | 2021-09-29 | null | null | null | null | ['3d-point-cloud-matching'] | ['computer-vision'] | [ 4.61694002e-01 2.26062179e-01 -1.19081430e-01 -4.19791818e-01
-7.01069474e-01 -5.98685563e-01 4.76362109e-02 5.47484100e-01
-3.63837659e-01 5.79481661e-01 -3.32405716e-01 -3.76610130e-01
-5.50082207e-01 -1.14854562e+00 -1.43547475e+00 -4.30982113e-01
-2.72870094e-01 9.00284231e-01 -1.04796104e-01 -1.00472383e-01
1.65729418e-01 8.43166649e-01 -1.64675915e+00 7.52392262e-02
8.82776499e-01 1.23996341e+00 1.79817021e-01 4.62377429e-01
-6.38029873e-01 1.99383691e-01 -3.11230034e-01 -7.40165055e-01
6.57450974e-01 1.98975638e-01 -1.00348198e+00 9.32218730e-02
1.16927683e+00 5.34465685e-02 -3.40139329e-01 1.16578364e+00
5.22921741e-01 -8.75331983e-02 3.15849364e-01 -1.93890655e+00
-7.97288716e-01 5.70566714e-01 -5.69491804e-01 -7.11568743e-02
2.75022507e-01 -1.55965090e-01 1.26425254e+00 -6.84318781e-01
6.92625880e-01 1.16316724e+00 1.15674555e+00 2.50219584e-01
-1.16079319e+00 -7.02515960e-01 -1.21128313e-01 4.22198713e-01
-1.53147221e+00 1.43294215e-01 6.62567139e-01 -2.87398696e-01
1.17698896e+00 6.82611287e-01 6.18674040e-01 3.59992176e-01
5.45952283e-02 4.27774340e-01 5.99768102e-01 -1.74263090e-01
-5.18754981e-02 -2.53239930e-01 3.04428667e-01 8.22560191e-01
3.74444097e-01 -8.19310099e-02 -3.50023687e-01 -1.98822007e-01
7.60993838e-01 9.57047194e-03 -2.53583223e-01 -6.19494021e-01
-1.39121318e+00 6.30677938e-01 9.02333379e-01 2.21369475e-01
-2.48242244e-01 6.23551548e-01 4.03977245e-01 6.56185746e-01
2.28735760e-01 6.92880869e-01 -6.64559484e-01 2.62838215e-01
-4.96216506e-01 4.70048845e-01 1.00109303e+00 1.61567986e+00
1.15661263e+00 -4.68206912e-01 -3.65652621e-01 4.30591941e-01
-4.70039487e-01 1.97994649e-01 -3.10058743e-01 -7.64148057e-01
7.18294322e-01 1.06810594e+00 -1.22383505e-01 -1.51308024e+00
-6.70936525e-01 -6.16210580e-01 -1.20101094e+00 9.52375233e-02
5.03654361e-01 1.73679069e-01 -7.26911843e-01 1.74349380e+00
5.13253152e-01 3.63575161e-01 -4.34992969e-01 1.04610181e+00
1.02210188e+00 6.09503627e-01 -1.53449729e-01 -1.15581609e-01
1.27160323e+00 -1.04025102e+00 -4.17609751e-01 6.77850693e-02
6.46526515e-01 -7.51399875e-01 9.61106896e-01 2.87254937e-02
-1.30230534e+00 -4.25297022e-01 -8.65886331e-01 -4.79115278e-01
-5.78120887e-01 -3.18510771e-01 1.11984718e+00 4.83328730e-01
-1.15913498e+00 7.75182188e-01 -1.56141549e-01 -1.29892513e-01
6.73248768e-01 9.23457563e-01 -7.86547720e-01 -1.71528131e-01
-9.74536419e-01 9.74286973e-01 2.86664724e-01 2.54881769e-01
1.80874001e-02 -1.21353340e+00 -8.91952038e-01 3.61185521e-01
5.64214289e-01 -9.90287840e-01 8.83914769e-01 -7.19782293e-01
-8.95411015e-01 1.16772890e+00 -1.38677880e-01 -3.44562560e-01
4.80693430e-01 1.17093690e-01 -4.78031002e-02 -1.57669768e-01
-1.59137975e-02 6.91341221e-01 5.03665388e-01 -8.63063455e-01
-4.75057960e-01 -3.99609387e-01 3.25155407e-01 1.80411562e-01
-3.28839034e-01 -1.95791081e-01 -6.69851303e-01 -3.29511106e-01
2.45483413e-01 -9.52890635e-01 -5.02732873e-01 6.78835511e-01
-4.28739429e-01 -5.62897027e-01 5.05403161e-01 -1.84458524e-01
7.46182144e-01 -1.87430477e+00 3.85073364e-01 4.85114008e-01
5.98280191e-01 3.34657216e-03 -3.63695204e-01 2.88125932e-01
-3.05327684e-01 5.59437945e-02 -4.49446797e-01 -3.14037919e-01
4.48439389e-01 3.97604764e-01 3.50019932e-02 6.70133948e-01
5.36260128e-01 1.18955243e+00 -8.72377157e-01 -6.63232923e-01
9.49278548e-02 2.68887341e-01 -6.42257810e-01 1.17895268e-01
-3.68178636e-01 2.19736807e-02 -1.92155018e-01 7.09654450e-01
1.24754655e+00 -6.13168299e-01 -1.63757190e-01 -4.73810464e-01
2.38477718e-02 -7.48730078e-02 -1.39540076e+00 2.19032955e+00
-4.25965786e-01 6.72651529e-01 8.72703716e-02 -1.29794776e+00
1.03785491e+00 -7.98503086e-02 7.23193467e-01 -6.89656019e-01
9.34486538e-02 3.42085451e-01 -1.93312377e-01 -3.88232797e-01
7.57805347e-01 3.14434588e-01 -2.18873501e-01 7.45483115e-02
1.28984144e-02 -2.34050050e-01 1.22166134e-01 -1.67622417e-02
1.26773894e+00 -5.07499389e-02 1.05448678e-01 -3.48288447e-01
4.68324840e-01 3.04074466e-01 3.99604142e-01 7.08910525e-01
2.91365266e-01 6.28304601e-01 6.03797674e-01 -8.12783241e-01
-1.11983836e+00 -9.66675341e-01 -2.75338590e-01 7.71157622e-01
4.92077798e-01 -3.62737983e-01 -4.78537261e-01 -3.43544066e-01
5.35727799e-01 2.08478108e-01 -4.55970496e-01 1.01794511e-01
-9.32034671e-01 -5.93756318e-01 3.62656802e-01 4.85261202e-01
2.21726805e-01 -7.54022002e-01 -3.31884027e-01 3.02397400e-01
1.15699381e-01 -1.26035750e+00 -7.22476304e-01 3.63877296e-01
-5.48764646e-01 -1.20093083e+00 -5.42589247e-01 -1.32034886e+00
9.78823483e-01 3.15223277e-01 1.64881182e+00 4.51397181e-01
-6.75972760e-01 1.15729876e-01 -1.02469191e-01 -3.65472376e-01
-5.98495118e-02 3.30516070e-01 -3.19654912e-01 -2.10612386e-01
5.16333044e-01 -7.77837515e-01 -4.33041692e-01 3.60086381e-01
-7.80713797e-01 8.46971422e-02 5.32961428e-01 6.77052498e-01
1.00908685e+00 -5.00429980e-02 3.41883063e-01 -8.79262507e-01
4.60996896e-01 -4.17394727e-01 -1.13905299e+00 3.84460866e-01
-3.45523864e-01 2.00670332e-01 5.80275238e-01 -4.12388653e-01
-5.53334616e-02 3.78313154e-01 2.37221234e-02 -6.83970869e-01
1.77376181e-01 3.77748132e-01 -2.27065518e-01 -6.64414167e-01
2.94613689e-01 1.31859453e-02 9.94630158e-02 -2.65779138e-01
4.28108901e-01 3.62925887e-01 8.62136722e-01 -7.41710365e-01
1.08795846e+00 2.10542992e-01 8.40219080e-01 -3.85036349e-01
-7.14851439e-01 -6.43072009e-01 -4.97152179e-01 -1.92650184e-01
7.60665596e-01 -5.80521166e-01 -1.41292691e+00 2.79336572e-01
-1.52510333e+00 7.72454217e-02 -2.59014606e-01 1.42671037e-02
-6.22555256e-01 3.33002031e-01 -2.84264863e-01 -3.12066168e-01
-6.42325282e-01 -1.02526510e+00 1.15654838e+00 4.35902700e-02
4.59584258e-02 -7.82588661e-01 -2.27581561e-01 7.10125938e-02
4.12743866e-01 7.98496068e-01 1.14719200e+00 -4.87517357e-01
-1.16473866e+00 -3.02554280e-01 -7.01528192e-01 -2.03111351e-01
-9.95644182e-02 -3.12034637e-01 -6.04085267e-01 -3.03043902e-01
-3.61013979e-01 -9.00691375e-02 5.62114358e-01 1.23631261e-01
1.88165283e+00 -3.42309713e-01 -3.61682296e-01 1.17019165e+00
1.78478420e+00 -5.30531347e-01 5.47469735e-01 2.96889186e-01
8.45862269e-01 5.27669311e-01 3.70387375e-01 2.94282645e-01
4.87807035e-01 7.23271906e-01 1.05904853e+00 -3.46029699e-01
-6.85184821e-02 -1.00647189e-01 -5.05182207e-01 7.23010540e-01
1.38191685e-01 -3.08328956e-01 -7.33166754e-01 4.87953782e-01
-2.00476241e+00 -6.28278375e-01 -6.24408543e-01 2.02538419e+00
6.70578301e-01 -1.23687655e-01 -1.15361869e-01 1.50292918e-01
9.39514577e-01 -1.78114831e-01 -4.56591725e-01 -5.00177383e-01
-3.45460027e-01 5.60553670e-01 1.03017795e+00 2.06663147e-01
-1.10347378e+00 5.36040783e-01 5.54794168e+00 7.23328888e-01
-8.32177877e-01 -6.38266876e-02 1.54370829e-01 5.83472699e-02
-4.73669946e-01 -2.86088288e-01 -6.37652993e-01 4.10482228e-01
2.91367829e-01 -2.70531029e-01 6.44042075e-01 6.49495006e-01
-6.82675838e-01 4.53940719e-01 -1.50720596e+00 1.38790226e+00
-1.43812206e-02 -1.87900984e+00 -1.05370365e-01 4.62762034e-03
6.77930057e-01 -2.98567340e-02 -2.14700308e-02 5.82455285e-02
6.04069680e-02 -1.40632927e+00 5.50822914e-01 4.28995907e-01
9.46806908e-01 -8.41646373e-01 7.60108590e-01 1.12128556e-01
-1.40853691e+00 7.18275160e-02 -7.93913543e-01 4.62120287e-02
1.34929270e-01 6.97523177e-01 -7.23746896e-01 7.89096653e-01
7.16199636e-01 6.14738047e-01 -1.97980553e-01 1.62323022e+00
3.24151427e-01 -1.17052250e-01 -6.03168726e-01 -2.09595218e-01
3.58078152e-01 -3.02025169e-01 6.47308111e-01 1.03452241e+00
5.30298233e-01 6.29128702e-03 6.95342645e-02 1.09587467e+00
-4.46330100e-01 4.95255813e-02 -6.17111504e-01 1.86432570e-01
5.58365166e-01 1.30789971e+00 -4.26233917e-01 5.68674691e-02
-3.29449952e-01 8.47986996e-01 7.54003406e-01 -1.92082390e-01
-1.01542974e+00 -5.75987518e-01 7.34637558e-01 -2.73162080e-03
4.82597888e-01 -2.33655453e-01 -6.13650680e-01 -6.30603909e-01
3.57612044e-01 -4.54480350e-01 4.07398462e-01 -7.82481492e-01
-1.51283991e+00 6.43361568e-01 -2.42243901e-01 -1.38228953e+00
3.78248245e-01 -8.76500010e-01 -3.82175058e-01 9.36333597e-01
-1.75774002e+00 -1.16629720e+00 -7.58654535e-01 7.28925705e-01
4.78785709e-02 8.72578770e-02 9.18487668e-01 8.56092274e-01
-3.73161942e-01 9.00351942e-01 -2.62074679e-01 -1.08241118e-01
2.82275289e-01 -1.44203317e+00 6.12830639e-01 4.47546244e-01
-8.60297456e-02 2.33458668e-01 6.40305281e-01 -1.99460283e-01
-2.05432630e+00 -1.34398365e+00 1.28460383e+00 -2.37525731e-01
5.86848915e-01 -5.24481177e-01 -9.72400427e-01 5.71530640e-01
-6.72174916e-02 3.88951033e-01 4.31948245e-01 1.90959238e-02
-3.68071944e-01 -2.41360337e-01 -1.23215199e+00 3.78742009e-01
1.76234436e+00 -2.58179694e-01 -3.77685130e-01 7.63061583e-01
1.36064947e+00 -9.83309805e-01 -1.22039425e+00 4.38507617e-01
3.18899482e-01 -5.13800204e-01 1.22696900e+00 -7.72083938e-01
3.27160835e-01 -4.30808753e-01 -2.63515800e-01 -1.06342423e+00
-4.52972800e-01 -7.73074806e-01 -1.22470304e-01 1.01803100e+00
3.42668205e-01 -4.61313426e-01 1.01854503e+00 4.69666779e-01
-4.78129506e-01 -7.91599512e-01 -1.09619880e+00 -9.28366005e-01
1.95503444e-03 -2.84077495e-01 1.29174376e+00 1.14419580e+00
-1.62424311e-01 3.11271939e-02 -2.46258155e-01 4.49886709e-01
8.24319243e-01 6.03713453e-01 1.00001466e+00 -1.38246334e+00
3.93199362e-02 -6.05863214e-01 -8.30936670e-01 -9.69508767e-01
3.60142142e-01 -1.47360086e+00 -1.15521073e-01 -1.57707489e+00
-5.17449249e-03 -9.09917653e-01 -6.38116375e-02 4.98060048e-01
-5.03546596e-02 1.68917283e-01 2.54301596e-02 -2.46086925e-01
-4.58747745e-01 1.23530827e-01 1.46695113e+00 -4.09525543e-01
9.06750932e-02 -7.23141506e-02 -4.57979620e-01 1.47551015e-01
6.81629837e-01 -6.53955162e-01 -1.41008005e-01 -9.02498305e-01
7.37389326e-01 4.05501835e-02 5.89944422e-01 -9.00568485e-01
6.32161200e-01 -2.72262365e-01 2.66327001e-02 -9.08759832e-01
3.52117419e-01 -1.32863152e+00 7.20043540e-01 3.43240499e-01
-1.29959688e-01 5.45941770e-01 3.18164259e-01 3.14647913e-01
4.83559072e-03 -3.71452160e-02 3.53344142e-01 -1.40541047e-01
-7.42114663e-01 9.81537938e-01 4.95181233e-01 -1.61292069e-02
1.10298777e+00 -4.23937023e-01 -5.53481340e-01 6.74406588e-02
-3.11717033e-01 5.41432142e-01 4.20265615e-01 2.60572493e-01
6.35090351e-01 -1.48295641e+00 -7.95232415e-01 2.19240606e-01
2.11285740e-01 6.45267427e-01 2.88597316e-01 8.05718124e-01
-8.82949710e-01 3.12378496e-01 -5.04355550e-01 -7.65834928e-01
-1.26643419e+00 8.59449208e-01 2.89105535e-01 -2.19457015e-01
-6.99616432e-01 1.00979745e+00 -2.77125061e-01 -7.88815379e-01
6.87458575e-01 -4.90257740e-01 1.56978205e-01 -1.06450722e-01
1.28752366e-01 4.27669555e-01 4.55095559e-01 -3.88640732e-01
-4.37665045e-01 7.99316525e-01 2.71269321e-01 9.47378993e-01
1.49494195e+00 2.84999579e-01 -7.57770658e-01 -1.93255186e-01
1.54686320e+00 -3.80990267e-01 -6.60111547e-01 -4.36048955e-01
-4.89239655e-02 -5.86732566e-01 -1.98888898e-01 -3.69244367e-01
-1.37594688e+00 4.97370541e-01 3.21516901e-01 2.56667793e-01
1.13018823e+00 6.01750277e-02 1.13062716e+00 6.37775123e-01
4.91988927e-01 -7.90641904e-01 -3.24315399e-01 3.15801382e-01
1.06989706e+00 -1.10117841e+00 -2.89055668e-02 -9.90546882e-01
1.00357294e-01 1.24326062e+00 7.87800670e-01 -4.08190727e-01
6.64696813e-01 4.54578847e-01 -4.16648597e-01 -3.19278091e-01
-6.30112112e-01 -3.46724033e-01 5.63521266e-01 6.37167692e-01
-2.59831343e-02 1.80001497e-01 -2.99507588e-01 3.65041256e-01
-4.94118154e-01 -8.26883912e-02 1.41841486e-01 6.76525295e-01
-8.41767043e-02 -1.07278442e+00 -2.34316304e-01 6.71874285e-01
-1.76324442e-01 -1.58262014e-01 -2.65651524e-01 8.26695859e-01
4.31903392e-01 4.33395714e-01 5.56941152e-01 -3.99197429e-01
5.76088727e-01 -4.83241469e-01 6.87306821e-01 -2.53887385e-01
-9.03797388e-01 -6.91433132e-01 9.26454514e-02 -7.05208302e-01
-4.55170840e-01 -2.31033504e-01 -1.22156143e+00 -6.40071034e-01
-4.64129835e-01 -1.03154548e-01 7.30710030e-01 6.41469300e-01
5.25851727e-01 5.79903185e-01 5.08218706e-01 -6.29167557e-01
-6.38007343e-01 -3.76007318e-01 -3.33312362e-01 6.45028532e-01
4.09836084e-01 -5.67248225e-01 -1.58511624e-01 -5.37480652e-01] | [7.093888759613037, 6.334185600280762] |
50f38b65-043f-451c-9161-148255c6ba95 | nethack-is-hard-to-hack | 2305.19240 | null | https://arxiv.org/abs/2305.19240v1 | https://arxiv.org/pdf/2305.19240v1.pdf | NetHack is Hard to Hack | Neural policy learning methods have achieved remarkable results in various control problems, ranging from Atari games to simulated locomotion. However, these methods struggle in long-horizon tasks, especially in open-ended environments with multi-modal observations, such as the popular dungeon-crawler game, NetHack. Intriguingly, the NeurIPS 2021 NetHack Challenge revealed that symbolic agents outperformed neural approaches by over four times in median game score. In this paper, we delve into the reasons behind this performance gap and present an extensive study on neural policy learning for NetHack. To conduct this study, we analyze the winning symbolic agent, extending its codebase to track internal strategy selection in order to generate one of the largest available demonstration datasets. Utilizing this dataset, we examine (i) the advantages of an action hierarchy; (ii) enhancements in neural architecture; and (iii) the integration of reinforcement learning with imitation learning. Our investigations produce a state-of-the-art neural agent that surpasses previous fully neural policies by 127% in offline settings and 25% in online settings on median game score. However, we also demonstrate that mere scaling is insufficient to bridge the performance gap with the best symbolic models or even the top human players. | ['Rob Fergus', 'Lerrel Pinto', 'Ulyana Piterbarg'] | 2023-05-30 | null | null | null | null | ['atari-games', 'nethack'] | ['playing-games', 'playing-games'] | [-1.69598565e-01 1.15742348e-01 -3.99203360e-01 2.93671072e-01
-6.42663360e-01 -6.73485458e-01 7.12955594e-01 -1.82986632e-01
-9.31453705e-01 1.25449502e+00 -7.57971108e-02 -3.91609341e-01
-4.37657773e-01 -4.61196810e-01 -9.13758516e-01 -6.68431520e-01
-6.19820058e-01 6.67693377e-01 4.97456491e-01 -9.02204216e-01
2.87091523e-01 3.18465084e-01 -1.78751123e+00 5.34339026e-02
6.14382148e-01 8.45860004e-01 2.21622378e-01 7.45119512e-01
4.68101799e-01 1.23092008e+00 -5.13925195e-01 -2.08321344e-02
5.65366864e-01 -4.46422815e-01 -6.23415828e-01 -4.55161899e-01
6.14246428e-02 -6.11633241e-01 -4.77673799e-01 8.86040092e-01
6.69079721e-01 4.67774928e-01 2.74633497e-01 -1.53392458e+00
-1.54705554e-01 9.71107900e-01 -2.58570045e-01 2.02973187e-01
1.01726197e-01 7.46816099e-01 9.19598162e-01 -1.13148190e-01
7.84278512e-01 1.14271700e+00 6.02582872e-01 7.71419585e-01
-8.94377828e-01 -6.78384125e-01 6.89686760e-02 2.64964461e-01
-8.89791310e-01 -4.60380822e-01 4.49764371e-01 -1.86672047e-01
1.31932640e+00 -2.08146617e-01 1.05635524e+00 1.55915189e+00
1.39284298e-01 9.21885550e-01 1.23459685e+00 -2.02654585e-01
5.77920556e-01 -3.95573109e-01 -5.10377884e-01 7.73241937e-01
1.30717859e-01 9.13451552e-01 -4.60398108e-01 -1.58468768e-01
8.41018617e-01 -5.68855226e-01 3.79677564e-02 -5.72393179e-01
-1.22089005e+00 8.92904580e-01 2.96336621e-01 1.22236364e-01
-5.44709742e-01 8.10321748e-01 6.60065651e-01 5.69133580e-01
-2.00324446e-01 1.02898955e+00 -5.11296511e-01 -8.54574144e-01
-5.79297245e-01 8.01099360e-01 1.01907575e+00 5.43550134e-01
2.33273104e-01 5.54392397e-01 1.93773851e-01 6.86438322e-01
-4.72894358e-03 3.73072416e-01 7.03694642e-01 -1.67523062e+00
4.64483678e-01 9.06751081e-02 3.31133455e-01 -5.43842673e-01
-6.09043419e-01 -3.17626178e-01 -2.46411502e-01 9.52414095e-01
7.28566825e-01 -5.18697321e-01 -5.93283534e-01 1.93901825e+00
5.00285625e-03 1.84660554e-01 2.65734255e-01 8.37200344e-01
2.83239752e-01 3.99282068e-01 -6.81460053e-02 -1.71569921e-02
9.78041470e-01 -1.05824578e+00 -1.94474965e-01 -4.30551231e-01
6.75947785e-01 7.93543607e-02 9.40719604e-01 5.86427629e-01
-1.25518107e+00 -3.34321022e-01 -1.16958964e+00 5.21401703e-01
-3.47564042e-01 -1.66072503e-01 8.63985658e-01 3.11510205e-01
-9.77810919e-01 9.56429422e-01 -1.22683895e+00 -5.87435246e-01
2.91393191e-01 5.37784100e-01 -1.63550302e-01 4.78171498e-01
-1.33057094e+00 1.31422305e+00 7.54556715e-01 -3.76981676e-01
-1.37462664e+00 -4.14380491e-01 -6.90325320e-01 1.55287847e-01
9.86419797e-01 -3.66565585e-01 1.74112082e+00 -8.35333765e-01
-1.97558928e+00 4.01089430e-01 7.30508864e-01 -1.03720152e+00
8.02161574e-01 -9.56486389e-02 -6.38276786e-02 9.62832719e-02
8.35292414e-03 9.60465789e-01 4.40595359e-01 -1.10075474e+00
-8.21249068e-01 -1.82851627e-02 4.45854306e-01 4.52981621e-01
-4.95324656e-02 -2.21236229e-01 -8.10560808e-02 -4.29947227e-01
-5.81970692e-01 -1.20928884e+00 -4.18247670e-01 -2.50710309e-01
9.12810341e-02 -2.09894583e-01 4.60304976e-01 -3.22119743e-01
7.86558688e-01 -1.93042791e+00 3.90686810e-01 -1.07992016e-01
-3.90431434e-02 3.27642888e-01 -1.26323596e-01 6.26721203e-01
2.20382392e-01 -1.40713781e-01 -1.23294974e-02 -1.15725920e-02
3.44257593e-01 5.93040526e-01 -3.33021939e-01 2.64545381e-01
-1.82251930e-01 1.07884514e+00 -1.18550789e+00 -3.68205607e-01
9.04632285e-02 -1.23980425e-01 -7.51337647e-01 -1.26208767e-01
-5.65900862e-01 3.38435441e-01 -4.69007313e-01 5.74116826e-01
-3.33991915e-01 1.07468009e-01 3.32227349e-01 4.44245279e-01
-3.98652524e-01 1.92438260e-01 -9.63183105e-01 1.79260647e+00
-1.18525840e-01 7.17561126e-01 2.18549684e-01 -1.15072656e+00
5.58713913e-01 2.11026445e-01 7.28770077e-01 -9.37616050e-01
3.58880013e-01 4.50853765e-01 5.77173293e-01 -4.34976190e-01
5.73768795e-01 1.81734171e-02 -2.04993725e-01 1.78668007e-01
1.68761298e-01 -2.26614326e-01 6.50267780e-01 -1.82569012e-01
1.38057125e+00 7.36087620e-01 3.76621276e-01 -2.27639526e-01
-3.64585146e-02 5.83444059e-01 4.49934483e-01 1.28104401e+00
-5.34665525e-01 1.26705676e-01 8.83295774e-01 -3.76146078e-01
-1.12398922e+00 -8.16253483e-01 4.01499778e-01 1.51523781e+00
9.72237512e-02 -2.51888692e-01 -8.42327774e-01 -4.66917872e-01
2.44813994e-01 8.03675532e-01 -8.14119935e-01 -2.24277586e-01
-7.94237077e-01 -5.83809555e-01 1.20621729e+00 6.35470152e-01
6.52970433e-01 -1.59612083e+00 -1.42407465e+00 4.78828073e-01
1.97932348e-01 -9.42312479e-01 8.66350159e-02 5.35665929e-01
-5.77152789e-01 -1.11104739e+00 -5.02197981e-01 -5.90259492e-01
-1.55326098e-01 -2.44711936e-01 8.44506025e-01 -1.45541906e-01
-8.95335618e-03 3.24981868e-01 -2.82180220e-01 -3.71119112e-01
-5.16322255e-01 3.80132496e-01 4.57086265e-01 -9.00610924e-01
-2.34643757e-01 -6.43582702e-01 -3.73141348e-01 3.57607722e-01
-4.66712028e-01 -4.02100235e-02 6.92394793e-01 1.11325836e+00
7.30638579e-02 -3.71365249e-02 7.29704797e-01 -3.15453023e-01
9.42030072e-01 -4.43132222e-01 -1.00033891e+00 4.51667309e-02
-6.33718431e-01 2.32668355e-01 7.35668898e-01 -6.50247395e-01
-7.65099764e-01 -2.25873694e-01 -4.93928641e-02 -3.37497234e-01
9.53451097e-02 6.16010606e-01 5.63357949e-01 -9.66584459e-02
9.31223810e-01 3.31852406e-01 4.74351704e-01 -1.55886069e-01
2.54735917e-01 1.60745427e-01 6.42252386e-01 -1.02271104e+00
4.57520008e-01 2.41453439e-01 1.10303335e-01 -5.78169286e-01
-2.46106163e-01 1.37905985e-01 -2.15527445e-01 -2.76813090e-01
6.72487378e-01 -7.64210045e-01 -1.39933920e+00 4.26322848e-01
-6.36202693e-01 -1.16533804e+00 -4.82382625e-01 5.28913319e-01
-1.30376458e+00 1.89308878e-02 -8.53909016e-01 -7.81645298e-01
1.11089699e-01 -1.27774990e+00 5.60883760e-01 2.99388260e-01
-6.11578859e-02 -7.39618540e-01 4.88698274e-01 6.99548647e-02
5.96951008e-01 2.87444651e-01 7.46778190e-01 -7.62921989e-01
-1.89067915e-01 1.58758193e-01 1.29856959e-01 9.12606120e-02
-3.13888550e-01 -1.72123805e-01 -4.75324035e-01 -5.49366713e-01
-4.86902952e-01 -1.02823508e+00 6.16600454e-01 3.08614969e-01
7.91153491e-01 2.01608315e-02 -1.94827959e-01 4.88719076e-01
1.09840178e+00 6.85947955e-01 3.78241211e-01 1.17184722e+00
2.38146082e-01 3.84976745e-01 5.95143020e-01 4.90637094e-01
2.54572064e-01 8.73539984e-01 9.73286092e-01 3.09311450e-01
5.43731973e-02 -4.68184561e-01 6.58909082e-01 3.30840260e-01
-4.70366776e-01 -1.37243748e-01 -8.55834782e-01 3.20744485e-01
-2.29888034e+00 -1.23915470e+00 6.50961757e-01 1.89224708e+00
6.56736553e-01 6.17018402e-01 5.83316207e-01 -1.76125646e-01
2.89355278e-01 1.10122621e-01 -1.03215313e+00 -4.21416014e-01
-6.00496754e-02 6.45264387e-02 7.09803045e-01 4.06516820e-01
-9.03787792e-01 1.22880745e+00 6.44604254e+00 9.18312192e-01
-1.06572902e+00 -1.02212466e-01 1.54649064e-01 -4.63082045e-01
5.26927531e-01 -1.35868460e-01 -6.31887496e-01 4.77481365e-01
1.01935089e+00 -1.95027292e-01 1.07682788e+00 1.13516223e+00
2.16024935e-01 -1.96377158e-01 -9.33991611e-01 6.87176645e-01
-9.74456519e-02 -1.35686302e+00 -5.22692382e-01 3.38163853e-01
6.33448124e-01 6.37380183e-01 1.23466114e-02 1.10825670e+00
1.01557910e+00 -1.17660975e+00 8.81309330e-01 1.45424679e-01
4.95135337e-01 -7.50744641e-01 6.20654941e-01 5.44654489e-01
-7.53134429e-01 -5.57146370e-01 -1.91407129e-01 -4.99131054e-01
4.26961780e-02 -6.86673522e-01 -7.66421199e-01 3.03503603e-01
6.17731273e-01 4.66598064e-01 -3.14045340e-01 1.02697158e+00
-2.41395742e-01 5.62696934e-01 -5.53714573e-01 -4.94412601e-01
9.61108625e-01 -1.06622130e-01 5.71669042e-01 7.04244733e-01
2.45784864e-01 8.30511227e-02 2.80456364e-01 6.00951254e-01
1.26803145e-01 -4.00609165e-01 -8.43548477e-01 -1.90100491e-01
3.83144438e-01 7.98958242e-01 -6.23430669e-01 -2.52376169e-01
-8.66461247e-02 5.90191424e-01 5.43122888e-01 2.66982496e-01
-1.07956660e+00 -2.53794223e-01 7.88902879e-01 -2.48847872e-01
4.05614763e-01 -4.52462465e-01 8.71624053e-02 -9.32280958e-01
-3.22235405e-01 -1.21555912e+00 1.71565935e-01 -7.80652761e-01
-6.62284493e-01 3.89816463e-01 3.32962483e-01 -1.09142959e+00
-9.47577477e-01 -8.86120856e-01 -4.04459864e-01 2.46555448e-01
-1.13009357e+00 -7.47790337e-01 7.18763322e-02 3.51745814e-01
5.76332390e-01 -6.16525292e-01 6.54940903e-01 2.22970340e-02
-5.64701915e-01 4.06211913e-01 5.17763376e-01 1.50726391e-02
3.14866602e-01 -1.28914678e+00 3.25500220e-01 4.35141116e-01
-8.35298300e-02 3.17250520e-01 8.61927450e-01 -5.20185769e-01
-1.40265870e+00 -4.65254605e-01 -1.20321721e-01 -2.96325207e-01
9.80306864e-01 -1.87302321e-01 -4.84878898e-01 7.10116923e-01
2.90742904e-01 -2.00661048e-01 -1.10985665e-02 -3.89639265e-03
5.67358285e-02 1.90172166e-01 -9.57346499e-01 1.13214421e+00
1.21926677e+00 -1.33099686e-02 -6.54835999e-01 -6.57566041e-02
5.74746251e-01 -7.46649921e-01 -6.79942310e-01 4.11247641e-01
1.02658188e+00 -1.04204130e+00 8.88559341e-01 -1.05155480e+00
7.05267787e-01 -8.35498869e-02 -2.69902021e-01 -1.68330812e+00
-2.40122095e-01 -7.37102151e-01 -1.54097646e-01 6.61049604e-01
3.10412824e-01 -5.76038480e-01 1.03935373e+00 1.61460996e-01
-2.30307400e-01 -1.04724216e+00 -1.14266014e+00 -1.15914321e+00
4.00533766e-01 -3.63998711e-01 2.80930161e-01 7.25976646e-01
3.63074094e-01 -9.05043632e-03 -6.37777507e-01 -3.30181569e-01
5.00799298e-01 -2.58370906e-01 7.84487367e-01 -7.36530364e-01
-7.85741806e-01 -9.08866465e-01 -2.10823983e-01 -9.80261981e-01
4.18189853e-01 -7.26029634e-01 1.91969529e-01 -1.21229970e+00
-1.36687934e-01 -4.01021510e-01 -2.45784208e-01 7.41029382e-01
2.74934441e-01 -3.23038325e-02 4.85678136e-01 1.25843480e-01
-7.10469902e-01 6.38445437e-01 1.06087363e+00 -4.88235168e-02
-2.92815894e-01 -2.91186929e-01 -5.36039352e-01 8.31484854e-01
1.13788807e+00 -3.24976772e-01 -5.18292665e-01 -3.91906857e-01
2.97852337e-01 4.02324319e-01 3.62687469e-01 -1.35489953e+00
3.74303669e-01 -4.42658007e-01 3.40021476e-02 6.05977289e-02
5.73155522e-01 -5.02405643e-01 -1.89867899e-01 1.11543608e+00
-5.26785672e-01 2.86959291e-01 4.83275354e-01 5.42041838e-01
7.74124190e-02 -3.12239349e-01 6.09371305e-01 -3.48948866e-01
-1.13671303e+00 -3.52338366e-02 -8.38931084e-01 2.81979620e-01
1.12701166e+00 -3.02420497e-01 -5.99237442e-01 -4.88828093e-01
-5.90115428e-01 6.15533888e-01 5.02659678e-01 4.99456614e-01
1.04427725e-01 -9.58368540e-01 -5.63868999e-01 -1.26989171e-01
-9.51699764e-02 -4.35230345e-01 -1.37689605e-01 7.47772038e-01
-7.49318182e-01 3.90816748e-01 -8.85902286e-01 -1.88186273e-01
-8.83651435e-01 3.42757821e-01 6.19529963e-01 -5.15242875e-01
-6.76640749e-01 5.02828717e-01 -2.12537587e-01 -5.27367711e-01
3.58816355e-01 -2.81048119e-01 -1.87880710e-01 -1.16870962e-01
3.77528816e-02 5.37296653e-01 -2.42557079e-01 -1.06378086e-01
-2.00550884e-01 1.05152942e-01 1.08768195e-01 -5.25202632e-01
1.33439302e+00 4.61431950e-01 4.48935091e-01 2.96393096e-01
5.67590177e-01 -4.53346133e-01 -1.81173766e+00 2.07304209e-01
6.90360591e-02 4.88882624e-02 -1.29816964e-01 -9.96803343e-01
-7.11822808e-01 3.97714585e-01 4.16532934e-01 2.24643826e-01
6.70515776e-01 -2.80404955e-01 5.53475320e-01 9.67813551e-01
9.05233800e-01 -1.30701649e+00 2.19763085e-01 9.38023925e-01
7.77524650e-01 -1.22575760e+00 -1.15035579e-01 4.68280554e-01
-9.27079380e-01 1.02159154e+00 8.63000751e-01 -4.83532876e-01
6.98494986e-02 2.93234348e-01 -2.27413550e-01 -1.71877012e-01
-1.26973653e+00 -3.90374959e-01 -4.79979664e-01 6.68953478e-01
-1.60068765e-01 -3.55066247e-02 -1.87768459e-01 3.98288518e-01
-5.63369513e-01 1.12810291e-01 5.19636929e-01 1.20900893e+00
-6.30383074e-01 -9.20334995e-01 -1.78271726e-01 2.98923701e-01
-3.83691341e-01 2.34156027e-01 -2.46641740e-01 1.42991018e+00
-1.48120478e-01 6.77840889e-01 -1.58790752e-01 -3.18140775e-01
4.45903718e-01 -6.98932214e-03 7.77930319e-01 -2.59527147e-01
-7.49420464e-01 -2.89643615e-01 4.75047559e-01 -7.71357298e-01
-8.80531296e-02 -7.90372849e-01 -1.34315884e+00 -3.66065711e-01
9.58691612e-02 1.37275368e-01 6.54651463e-01 9.76229966e-01
1.56465203e-01 6.91646099e-01 6.97687790e-02 -1.28630519e+00
-1.06043983e+00 -8.37305546e-01 -4.73265976e-01 1.28987819e-01
1.81031913e-01 -1.07240057e+00 -2.87878722e-01 -3.72246355e-01] | [3.813220500946045, 1.5439258813858032] |
7cf737af-53dd-4691-bab5-8835ccc4a082 | connector-0-5-a-unified-framework-for-graph | 2304.13195 | null | https://arxiv.org/abs/2304.13195v1 | https://arxiv.org/pdf/2304.13195v1.pdf | Connector 0.5: A unified framework for graph representation learning | Graph representation learning models aim to represent the graph structure and its features into low-dimensional vectors in a latent space, which can benefit various downstream tasks, such as node classification and link prediction. Due to its powerful graph data modelling capabilities, various graph embedding models and libraries have been proposed to learn embeddings and help researchers ease conducting experiments. In this paper, we introduce a novel graph representation framework covering various graph embedding models, ranging from shallow to state-of-the-art models, namely Connector. First, we consider graph generation by constructing various types of graphs with different structural relations, including homogeneous, signed, heterogeneous, and knowledge graphs. Second, we introduce various graph representation learning models, ranging from shallow to deep graph embedding models. Finally, we plan to build an efficient open-source framework that can provide deep graph embedding models to represent structural relations in graphs. The framework is available at https://github.com/NSLab-CUK/Connector. | ['O-Joun Lee', 'Van Thuy Hoang', 'Jooho Lee', 'Thanh Sang Nguyen'] | 2023-04-25 | null | null | null | null | ['graph-embedding'] | ['graphs'] | [-4.56557184e-01 4.66879666e-01 -4.54641908e-01 -2.40550280e-01
2.63123941e-02 -4.45208013e-01 5.66940904e-01 5.27972400e-01
3.01069140e-01 3.47635448e-01 3.43776405e-01 -5.76794326e-01
-9.94449556e-02 -1.41190958e+00 -2.05378383e-01 -4.67858166e-01
-5.24233878e-01 4.61942375e-01 4.33214828e-02 -2.21433803e-01
-1.90594003e-01 5.41072071e-01 -1.00797892e+00 -1.31067619e-01
5.30153871e-01 5.03138423e-01 -2.07131490e-01 7.07388461e-01
-2.53225505e-01 8.08864772e-01 -2.16060638e-01 -8.41489851e-01
-1.92483142e-02 -1.42967314e-01 -5.91697872e-01 -2.07961723e-01
-1.23380698e-01 -1.44184291e-01 -1.29611516e+00 1.06648791e+00
5.31359732e-01 -3.52331884e-02 6.70764744e-01 -1.87617719e+00
-1.35315502e+00 7.73548424e-01 -2.40427360e-01 -4.81192097e-02
5.62986791e-01 9.57594961e-02 1.53085625e+00 -5.39612591e-01
7.74706662e-01 1.28583598e+00 6.01455927e-01 3.84775579e-01
-1.12701750e+00 -5.43034256e-01 2.71246016e-01 4.97871041e-01
-1.42018831e+00 -7.72207826e-02 1.15330780e+00 -4.29252207e-01
1.01696515e+00 3.57825965e-01 1.02131510e+00 1.10653174e+00
5.75067699e-02 8.63558412e-01 5.45319080e-01 -3.05410117e-01
-1.24715649e-01 -7.15572685e-02 5.02809644e-01 1.17639589e+00
6.72510266e-01 -1.98537797e-01 -9.67535228e-02 -3.52606475e-01
7.52855182e-01 3.08295071e-01 -3.13880056e-01 -6.93910241e-01
-9.39190090e-01 1.07084882e+00 1.02066863e+00 3.46088618e-01
-2.23304793e-01 4.90983605e-01 4.88972574e-01 3.60029995e-01
5.10618985e-01 2.43069097e-01 -2.11603209e-01 2.00983018e-01
-2.05694705e-01 -1.32368118e-01 8.44253182e-01 1.21457648e+00
8.78100693e-01 8.68874788e-02 -2.10504040e-01 8.09695423e-01
7.85239041e-01 -4.99220900e-02 3.57855469e-01 -2.75957435e-01
5.27086198e-01 1.15706480e+00 -5.44462323e-01 -1.57729208e+00
-4.17071044e-01 -2.35606030e-01 -1.14415050e+00 -4.85504150e-01
-4.17794347e-01 5.02386801e-02 -7.89308965e-01 1.46420014e+00
2.86240965e-01 5.91217756e-01 7.56680667e-02 5.67933083e-01
1.49882054e+00 7.93346763e-01 -1.05559178e-01 1.95655614e-01
1.32336617e+00 -1.22353315e+00 -6.65987432e-01 -7.35113025e-02
1.08198714e+00 -2.83062786e-01 9.35283422e-01 -2.10725129e-01
-7.34519482e-01 -3.42007101e-01 -9.94531393e-01 -2.68003732e-01
-8.18160772e-01 -1.05399437e-01 1.37979138e+00 3.48085433e-01
-1.40180063e+00 4.91081089e-01 -9.07569468e-01 -6.92913234e-01
4.09665763e-01 2.24728838e-01 -8.17308307e-01 -4.15094793e-01
-1.45320427e+00 4.63168800e-01 5.63630283e-01 2.27460250e-01
-6.69380605e-01 -2.71408856e-01 -1.48934889e+00 4.89215046e-01
1.50477067e-01 -8.97521675e-01 5.11238515e-01 -2.05467358e-01
-1.01274383e+00 7.82099843e-01 4.80643734e-02 -1.83844179e-01
-1.04415208e-01 2.28113472e-01 -6.26344264e-01 -8.22574049e-02
-1.44733235e-01 2.66671330e-01 2.65136331e-01 -9.85780120e-01
1.82150736e-01 -3.80582035e-01 3.05583000e-01 2.20052209e-02
-7.82526076e-01 -1.44261837e-01 -7.90922344e-01 -5.85264087e-01
2.45395005e-01 -9.44285929e-01 -4.10509676e-01 1.43040763e-02
-8.86257291e-01 -5.55210650e-01 6.99928939e-01 -5.24452388e-01
1.58493161e+00 -2.12917113e+00 3.93142313e-01 3.08315098e-01
9.27346766e-01 2.63250232e-01 -5.73322833e-01 1.06403244e+00
-4.28144038e-01 3.76727045e-01 1.47517264e-01 -1.70750320e-01
3.08608055e-01 3.37130487e-01 -7.96834081e-02 3.52028400e-01
2.47267812e-01 1.62214828e+00 -1.02892756e+00 -3.52835298e-01
2.60964841e-01 6.06473565e-01 -4.07080024e-01 4.71249104e-01
-7.73842586e-03 -1.29773945e-01 -6.48964822e-01 6.56109035e-01
6.31832778e-01 -6.76993251e-01 7.03679502e-01 -2.16321871e-01
4.44734067e-01 1.94013357e-01 -9.23905253e-01 1.59845126e+00
-3.82642090e-01 5.64831316e-01 -6.66104108e-02 -1.10739553e+00
1.07242918e+00 2.10877120e-01 3.93410921e-01 -2.57124215e-01
1.56901434e-01 -2.37372696e-01 1.33707866e-01 -3.63016456e-01
3.78019392e-01 4.13106620e-01 -1.46590015e-02 5.24246335e-01
2.61214972e-01 1.92350045e-01 2.69615829e-01 8.73517811e-01
1.59272444e+00 -3.38937819e-01 3.20424706e-01 8.89874771e-02
3.52380753e-01 -4.68850613e-01 3.81679565e-01 6.98674619e-02
-1.74835667e-01 2.69736886e-01 1.16173291e+00 -5.71953177e-01
-6.49702609e-01 -1.22508430e+00 1.89935565e-01 6.97903216e-01
1.40078217e-01 -1.19270337e+00 -1.41136840e-01 -6.97031736e-01
2.61733472e-01 1.87120453e-01 -6.54037118e-01 -6.46452904e-01
-2.76263565e-01 -7.52475023e-01 3.43374223e-01 5.55029213e-01
9.96963754e-02 -9.75459754e-01 6.13234043e-01 2.16309518e-01
1.64397568e-01 -9.79653835e-01 -2.93990940e-01 -9.74164829e-02
-8.10578167e-01 -1.31072927e+00 -3.96398783e-01 -1.17222941e+00
1.03636014e+00 3.14250529e-01 1.23822677e+00 6.56016231e-01
-4.29670274e-01 4.88452315e-01 -6.20748699e-01 1.34084493e-01
-1.52615994e-01 3.83907527e-01 -1.79728150e-01 -1.66350096e-01
1.79529563e-01 -8.13704669e-01 -5.64165592e-01 -8.91587734e-02
-8.82984042e-01 1.19873591e-01 4.86058891e-01 8.04682255e-01
4.82298940e-01 1.27186999e-01 3.78577054e-01 -1.17264771e+00
1.14101207e+00 -9.01977777e-01 -3.29209894e-01 3.77336532e-01
-6.79852784e-01 1.65791139e-01 6.97970390e-01 -1.42832190e-01
-2.04698011e-01 -3.52112293e-01 -2.58176416e-01 -5.09251952e-01
1.50100797e-01 9.74953294e-01 -4.37070072e-01 -2.37100129e-03
2.33038232e-01 1.69634134e-01 2.19465699e-02 -3.69719386e-01
9.18527365e-01 4.91842836e-01 -8.42387974e-02 -3.38619947e-01
1.04599404e+00 7.87586048e-02 5.41548915e-02 -6.22756004e-01
-4.42356586e-01 -4.56274033e-01 -6.09181225e-01 -3.86230536e-02
6.02508485e-01 -8.39264989e-01 -4.37181264e-01 2.69952893e-01
-1.11935067e+00 -4.07232672e-01 1.00141205e-01 3.26529175e-01
-2.40147278e-01 6.81918859e-01 -1.11136007e+00 -2.98442394e-01
-4.76160228e-01 -1.01282740e+00 8.45284343e-01 2.05556989e-01
1.68072674e-02 -1.56647968e+00 3.20424497e-01 6.12219498e-02
3.87333363e-01 3.82002234e-01 1.30243897e+00 -6.55453622e-01
-8.43880057e-01 -4.58570957e-01 -5.06733596e-01 1.53520266e-02
2.75496751e-01 2.64098048e-01 -4.42215264e-01 -5.67647457e-01
-1.04904580e+00 -1.39166012e-01 8.73469949e-01 6.60417303e-02
1.32304919e+00 -3.75843376e-01 -8.37602854e-01 1.05200160e+00
1.36641431e+00 -2.42665008e-01 7.73265541e-01 6.45138621e-02
1.19899714e+00 3.41000617e-01 2.46926825e-02 9.52026173e-02
8.49687696e-01 5.32598197e-01 6.55984104e-01 -1.73263088e-01
-1.20719478e-01 -5.11140227e-01 1.38050288e-01 1.45916760e+00
3.46198268e-02 -5.06520092e-01 -8.77557099e-01 5.08016050e-01
-2.07401729e+00 -8.83253098e-01 -4.94655192e-01 1.65415525e+00
3.13910902e-01 1.57539267e-02 -1.01387240e-01 -8.00578445e-02
7.59815276e-01 7.10639596e-01 -3.47479582e-01 -5.56376517e-01
1.55488208e-01 1.75592542e-01 1.56880692e-01 3.98831159e-01
-9.47816730e-01 1.32839334e+00 5.45241880e+00 5.04087269e-01
-8.46967340e-01 4.49438281e-02 3.71028841e-01 4.03182596e-01
-8.30775559e-01 2.25886345e-01 -2.11821109e-01 4.10676509e-01
9.18640316e-01 -6.37032390e-01 5.00631511e-01 9.60155368e-01
-2.82574713e-01 8.32721353e-01 -1.01501608e+00 1.06470394e+00
-7.87009895e-02 -1.45464361e+00 2.99118251e-01 1.62808031e-01
5.20633459e-01 1.85944363e-01 -1.64872646e-01 7.41346657e-01
6.36005282e-01 -1.22962463e+00 -2.20894426e-01 4.79580104e-01
8.32104862e-01 -5.92415750e-01 7.79431641e-01 -5.13439290e-02
-1.96009755e+00 8.99961218e-02 -6.69587135e-01 -9.20496136e-02
1.24499694e-01 6.78616464e-01 -6.41867936e-01 1.11153162e+00
2.93419868e-01 1.27668488e+00 -8.29448581e-01 9.32302892e-01
-7.54584908e-01 5.76461375e-01 1.62595183e-01 -8.73189345e-02
7.55507825e-03 -5.02051830e-01 1.88313320e-01 9.33782458e-01
3.16064715e-01 2.37389673e-02 1.74212694e-01 8.34948003e-01
-4.83266532e-01 2.12411702e-01 -1.09219921e+00 -6.28778160e-01
7.40451396e-01 1.65100288e+00 -5.82370758e-01 -2.28725344e-01
-6.28571391e-01 1.03406990e+00 8.98644745e-01 4.68400985e-01
-8.22978377e-01 -6.92983091e-01 9.09671605e-01 3.13708261e-02
-3.00778467e-02 -4.67144370e-01 4.10117924e-01 -1.60120571e+00
-9.55776274e-02 -4.51975435e-01 6.10200763e-01 -8.15103292e-01
-1.59358859e+00 8.12836051e-01 -1.73748791e-01 -8.69085610e-01
8.30710679e-02 -7.37182856e-01 -1.14884102e+00 7.87327945e-01
-1.51983488e+00 -1.37756693e+00 -6.31344378e-01 5.02906859e-01
-1.74690053e-01 -3.32162470e-01 1.01009142e+00 3.84896070e-01
-8.79715979e-01 6.66372240e-01 4.64526899e-02 7.08580673e-01
3.14244777e-01 -1.26210189e+00 9.39374924e-01 6.02380455e-01
4.75604445e-01 8.79936039e-01 2.17347573e-02 -4.71915841e-01
-1.91023862e+00 -1.44848788e+00 7.88832784e-01 -2.56990761e-01
1.12074053e+00 -6.55549884e-01 -9.48960662e-01 1.18582273e+00
2.44828817e-02 7.21644223e-01 8.68284285e-01 5.55625618e-01
-4.45048541e-01 -4.16126922e-02 -7.14827776e-01 8.17863345e-01
1.35327911e+00 -6.98097527e-01 -9.96793881e-02 5.93618929e-01
1.06169271e+00 -1.37504950e-01 -1.27171767e+00 2.53799349e-01
4.18884665e-01 -6.19915903e-01 1.07817519e+00 -9.94123101e-01
3.57781112e-01 -1.45199209e-01 3.90263312e-02 -1.58499718e+00
-9.70382333e-01 -4.66813982e-01 -6.90493822e-01 1.26543987e+00
3.56676131e-01 -1.02594042e+00 1.06776571e+00 2.92465448e-01
-5.47042675e-02 -1.18999159e+00 -5.66934168e-01 -6.49934232e-01
-1.24275304e-01 -1.93304211e-01 1.02696395e+00 1.35298324e+00
1.03320621e-01 6.85289741e-01 -1.77020699e-01 3.95698816e-01
4.79619592e-01 3.55460346e-01 1.03181696e+00 -1.43771625e+00
-7.32797906e-02 -5.21037757e-01 -1.39572906e+00 -8.95802140e-01
5.80367684e-01 -1.73300600e+00 -8.21308434e-01 -2.40080619e+00
2.41690576e-01 -5.24197876e-01 -5.40354908e-01 6.19004726e-01
-2.55393863e-01 -1.07644394e-01 4.64758500e-02 -8.90031233e-02
-5.68983078e-01 1.00819838e+00 1.20126510e+00 -4.65479493e-01
-2.14637071e-02 -2.55845189e-01 -9.10688698e-01 2.30987042e-01
8.96108568e-01 -2.94572860e-01 -7.55302370e-01 -5.28738797e-01
3.35907340e-01 6.76799566e-02 3.33616197e-01 -6.32170260e-01
1.80746038e-02 4.65668691e-03 2.15750537e-03 -2.91984677e-01
2.30113655e-01 -7.01503098e-01 3.54341388e-01 4.84266490e-01
-1.99746937e-01 2.10295454e-01 -1.27278209e-01 7.37647355e-01
-2.99967587e-01 2.57844999e-02 2.99289882e-01 4.85728309e-02
-7.92172790e-01 1.13875616e+00 2.46527776e-01 -1.16384998e-01
1.21484578e+00 8.65821838e-02 -6.62896812e-01 -5.30324638e-01
-9.47205365e-01 5.44888914e-01 4.34316605e-01 6.20258629e-01
8.82746041e-01 -1.89417005e+00 -7.19029129e-01 2.38929868e-01
6.09049439e-01 -1.31560326e-01 1.31917715e-01 6.41604245e-01
-6.27769172e-01 2.92514592e-01 -2.48008582e-04 -1.19604178e-01
-1.24887884e+00 7.33631134e-01 7.82150179e-02 -4.86606419e-01
-7.70015836e-01 9.71434951e-01 1.76759049e-01 -9.14272487e-01
-7.22574219e-02 -2.95999408e-01 -2.91593254e-01 -2.31083050e-01
1.68905541e-01 1.54078797e-01 -1.41625121e-01 -5.89608908e-01
-4.38416779e-01 3.33711624e-01 -5.02000712e-02 7.56480515e-01
1.33541179e+00 2.97352612e-01 -5.50942421e-01 1.84281185e-01
1.62331104e+00 -9.30233672e-02 -5.98170102e-01 -1.21956319e-01
-6.38618916e-02 -4.40756947e-01 -2.32352912e-02 -1.04370467e-01
-1.39162982e+00 9.48028505e-01 7.29593784e-02 6.44470155e-01
6.89013779e-01 3.14190865e-01 8.65873098e-01 3.61920416e-01
4.45239067e-01 -3.38288128e-01 4.26599346e-02 4.15666878e-01
8.61451447e-01 -1.18883145e+00 1.81351081e-01 -6.49409831e-01
-4.07894373e-01 1.08490264e+00 6.95441544e-01 -3.45202148e-01
1.04564345e+00 -7.98562616e-02 -2.73507386e-01 -6.60425186e-01
-1.08177578e+00 -4.41551387e-01 3.94415885e-01 8.39669347e-01
6.54193401e-01 5.11603892e-01 -1.34357959e-01 6.66718543e-01
-1.52455747e-01 -4.15353805e-01 4.55905676e-01 5.41262805e-01
-5.25024496e-02 -1.55061376e+00 4.45070803e-01 8.98164570e-01
2.20565617e-01 -2.82781452e-01 -6.13843918e-01 7.67371178e-01
-3.72727990e-01 6.99774623e-01 -1.09989747e-01 -7.37091243e-01
3.35396171e-01 -2.47040227e-01 2.25642815e-01 -1.04854691e+00
2.03570761e-02 -5.92895329e-01 1.59460053e-01 -5.45532286e-01
1.58854678e-01 -1.30479470e-01 -1.19566226e+00 -4.57816184e-01
-3.95614475e-01 4.27607507e-01 3.61957312e-01 2.26961315e-01
7.42294371e-01 6.87672079e-01 5.86168647e-01 -6.86048508e-01
1.10284379e-02 -9.22809601e-01 -8.48076344e-01 4.02261794e-01
-1.65224254e-01 -6.59518242e-01 -3.37063164e-01 -5.46940446e-01] | [7.169623851776123, 6.277251720428467] |
ba67ae5e-a1aa-4447-92a1-2595362b7ee8 | understanding-the-role-of-affect-dimensions | 2105.03983 | null | https://arxiv.org/abs/2105.03983v1 | https://arxiv.org/pdf/2105.03983v1.pdf | Understanding the Role of Affect Dimensions in Detecting Emotions from Tweets: A Multi-task Approach | We propose VADEC, a multi-task framework that exploits the correlation between the categorical and dimensional models of emotion representation for better subjectivity analysis. Focusing primarily on the effective detection of emotions from tweets, we jointly train multi-label emotion classification and multi-dimensional emotion regression, thereby utilizing the inter-relatedness between the tasks. Co-training especially helps in improving the performance of the classification task as we outperform the strongest baselines with 3.4%, 11%, and 3.9% gains in Jaccard Accuracy, Macro-F1, and Micro-F1 scores respectively on the AIT dataset. We also achieve state-of-the-art results with 11.3% gains averaged over six different metrics on the SenWave dataset. For the regression task, VADEC, when trained with SenWave, achieves 7.6% and 16.5% gains in Pearson Correlation scores over the current state-of-the-art on the EMOBANK dataset for the Valence (V) and Dominance (D) affect dimensions respectively. We conclude our work with a case study on COVID-19 tweets posted by Indians that further helps in establishing the efficacy of our proposed solution. | ['Niloy Ganguly', 'Soham Dasgupta', 'Sriyash Poddar', 'Atharva Naik', 'Rajdeep Mukherjee'] | 2021-05-09 | null | null | null | null | ['subjectivity-analysis'] | ['natural-language-processing'] | [-2.10917518e-01 -1.03327416e-01 -8.90466571e-02 -6.56705141e-01
-1.12016010e+00 -5.17331481e-01 4.45021421e-01 3.92049462e-01
-5.05024791e-01 5.20690858e-01 3.28291893e-01 2.26512030e-01
5.74279465e-02 -3.57724845e-01 -1.26416340e-01 -4.23043370e-01
-1.38397887e-01 1.37399212e-01 -5.53501964e-01 -4.83455032e-01
-4.61515859e-02 -7.72848651e-02 -1.54743361e+00 5.32903373e-01
7.35297859e-01 1.72387290e+00 -7.16140687e-01 4.72009569e-01
-1.02393918e-01 7.25305080e-01 -5.09482443e-01 -8.13014030e-01
-1.07977711e-01 -2.19460158e-03 -6.26725316e-01 -1.09946720e-01
1.37787133e-01 1.79851905e-01 2.56139368e-01 6.84375465e-01
6.16408050e-01 1.49669304e-01 8.84936512e-01 -1.47013807e+00
-4.43885416e-01 4.94036764e-01 -1.00185096e+00 -3.43350060e-02
2.87941307e-01 -5.41487157e-01 1.32306719e+00 -1.02252328e+00
4.29557711e-01 1.20774579e+00 6.76867247e-01 4.59494978e-01
-9.55522656e-01 -9.30981874e-01 4.85309735e-02 2.49660220e-02
-1.22065020e+00 -4.96598929e-01 7.49429703e-01 -4.10328269e-01
1.06906199e+00 3.86505008e-01 6.84173182e-02 1.29158616e+00
4.84664738e-02 8.18295240e-01 1.53599381e+00 -2.86452919e-01
9.28272530e-02 5.63319623e-01 4.96278107e-01 6.18340373e-01
-4.03948992e-01 -4.10949379e-01 -5.87466180e-01 -2.33761683e-01
-6.38155406e-03 -3.60905141e-01 2.61661559e-01 2.69205600e-01
-6.31804168e-01 1.15507734e+00 1.10146120e-01 3.56299728e-01
-6.35314703e-01 -1.51467979e-01 1.04373229e+00 2.80915052e-01
1.19198096e+00 3.67862940e-01 -7.52591193e-01 -6.13663077e-01
-6.59666598e-01 9.85473394e-02 7.27472365e-01 4.57463324e-01
4.71139073e-01 4.53210399e-02 -2.88248897e-01 1.37032521e+00
1.17827363e-01 5.69099069e-01 4.13921684e-01 -7.35826969e-01
3.32593471e-01 7.23859668e-01 4.60873991e-02 -1.20185244e+00
-9.42638636e-01 -6.86863542e-01 -8.46097291e-01 -2.24723011e-01
2.93886401e-02 -8.13303947e-01 -4.56599295e-01 2.09019232e+00
4.26450431e-01 -1.53958768e-01 2.90024877e-01 8.07444096e-01
1.02937853e+00 6.58166528e-01 4.21444654e-01 -4.77900177e-01
1.67805851e+00 -8.63775790e-01 -9.88712728e-01 -2.63806164e-01
8.20977092e-01 -7.66706169e-01 1.15054238e+00 4.76124346e-01
-9.18930054e-01 -3.98384094e-01 -8.51464629e-01 -2.54633022e-04
-4.81145769e-01 4.00182277e-01 7.77374566e-01 8.92171443e-01
-5.87090254e-01 -8.86812583e-02 -4.17024404e-01 -1.12319380e-01
3.87682080e-01 2.26860777e-01 -3.76367688e-01 3.73604685e-01
-1.34809303e+00 7.87897229e-01 -2.48379514e-01 -1.25985175e-01
-3.80503565e-01 -9.61690247e-01 -6.61985040e-01 -3.18316519e-02
1.29962385e-01 -1.41354412e-01 1.08654034e+00 -1.13365245e+00
-1.58055580e+00 1.05202401e+00 -2.73203641e-01 -2.20357820e-01
2.36798525e-01 -5.17468154e-01 -8.23030293e-01 -4.78915609e-02
1.80185039e-03 4.39724267e-01 4.41850454e-01 -1.02691293e+00
-6.13873541e-01 -5.78189552e-01 -2.17298001e-01 9.36361104e-02
-7.76662171e-01 4.12432522e-01 -2.66243726e-01 -4.15883154e-01
-3.46307486e-01 -1.09778082e+00 1.05944335e-01 -5.79739392e-01
-1.79258674e-01 -5.45187235e-01 8.89181614e-01 -7.36319184e-01
1.38786197e+00 -2.33385801e+00 1.42747074e-01 2.15663463e-01
2.85754085e-01 2.09395736e-01 -1.63594201e-01 1.47615775e-01
-1.57163605e-01 2.18525440e-01 1.40261903e-01 -7.35530078e-01
2.08557382e-01 -2.40981758e-01 -2.67589670e-02 4.80011076e-01
2.82140970e-01 7.92728901e-01 -5.06847143e-01 -2.01532364e-01
-5.69258854e-02 6.96438015e-01 -5.18996477e-01 1.98092654e-01
1.38601631e-01 2.11747125e-01 -3.14361781e-01 4.59329456e-01
5.94047189e-01 -2.49888465e-01 2.25230053e-01 -3.28758955e-01
4.47225794e-02 2.26468712e-01 -8.44280005e-01 1.49457598e+00
-9.90600228e-01 6.26652956e-01 1.29804298e-01 -9.42544878e-01
1.26055956e+00 4.89761680e-01 6.30287230e-01 -1.20388722e+00
4.40626651e-01 7.42691010e-02 -1.46920621e-01 -5.27758718e-01
6.23973966e-01 -3.71848673e-01 -8.19537401e-01 4.97565210e-01
-1.41760884e-02 4.30367708e-01 -7.39526562e-03 3.32787961e-01
8.26339364e-01 -3.07767510e-01 2.71668494e-01 -2.57061422e-01
5.27001023e-01 -4.37585741e-01 5.32114089e-01 3.16625714e-01
-4.07158792e-01 2.11157352e-01 8.65423441e-01 -1.44842193e-01
-6.54671013e-01 -6.89626157e-01 -2.38275856e-01 1.74844730e+00
-3.53051960e-01 -6.18784130e-01 -5.57850778e-01 -8.16651881e-01
-3.52793932e-02 7.87388802e-01 -1.10313952e+00 -5.20993434e-02
-1.26030639e-01 -1.17482567e+00 7.98917413e-01 4.04131860e-01
4.92952168e-01 -6.11394107e-01 -5.00073075e-01 1.03191085e-01
-5.21741748e-01 -1.50533402e+00 1.49924636e-01 2.26408035e-01
-3.20016444e-01 -5.32192469e-01 -2.99579054e-01 -3.11976284e-01
8.11789483e-02 -1.86863676e-01 1.21881437e+00 -2.70604819e-01
6.62736669e-02 2.28827298e-01 -6.47230148e-01 -5.32969117e-01
-1.00369692e-01 2.32832864e-01 1.23038866e-01 3.26963693e-01
5.72048485e-01 -4.79864448e-01 -3.54449183e-01 2.30551094e-01
-6.24373019e-01 -1.86395720e-01 2.37395421e-01 6.51265025e-01
1.98086202e-01 -3.48036021e-01 1.20755363e+00 -9.36924458e-01
7.94362783e-01 -8.76388907e-01 -6.37134984e-02 7.15214610e-02
-5.89668274e-01 -1.97141185e-01 4.23860401e-01 -2.79762059e-01
-1.25647807e+00 -3.54810864e-01 -2.68542200e-01 -1.93088084e-01
-9.93502662e-02 6.07115388e-01 1.92799509e-01 3.25342149e-01
5.82401633e-01 -4.79193747e-01 -2.46747777e-01 -2.74736166e-01
4.29911077e-01 1.20172822e+00 1.31618038e-01 -4.05009687e-01
3.59039567e-02 3.07394832e-01 -1.69352114e-01 -7.23907053e-01
-1.27249956e+00 -6.64089978e-01 -3.40267956e-01 -2.60601938e-01
1.02923608e+00 -1.27890944e+00 -1.04675698e+00 3.29009026e-01
-9.06926334e-01 -9.26412642e-03 3.18486422e-01 1.79571360e-01
-1.66506335e-01 -5.42583354e-02 -7.42048383e-01 -1.17178047e+00
-8.51248503e-01 -9.31522608e-01 1.22403944e+00 5.84564060e-02
-6.73289239e-01 -1.18923473e+00 1.26735836e-01 6.48565710e-01
5.86285889e-01 4.86603856e-01 8.96072149e-01 -1.00738490e+00
7.31320977e-01 -1.12817891e-01 -3.45969200e-01 3.84178162e-01
-8.82821977e-02 -1.17637843e-01 -1.34974194e+00 7.83488341e-03
-1.60739020e-01 -9.70109820e-01 6.56894147e-01 1.46939844e-01
1.06273925e+00 -5.73489033e-02 1.28810760e-02 1.42102942e-01
1.27452934e+00 -5.37565313e-02 3.14422041e-01 2.22677827e-01
5.44118702e-01 8.43879104e-01 7.74889350e-01 1.01211631e+00
8.90908837e-01 7.50690281e-01 3.14345300e-01 -2.36623406e-01
2.36660704e-01 2.85281271e-01 4.33213532e-01 1.09860635e+00
1.00538090e-01 -1.19177118e-01 -8.34133267e-01 4.51166123e-01
-1.90613604e+00 -6.60077870e-01 -3.31168562e-01 1.62675917e+00
8.33961785e-01 -1.22149415e-01 4.33460951e-01 2.24689931e-01
4.26149726e-01 2.95587450e-01 -3.35318625e-01 -1.13207257e+00
-2.94664532e-01 2.72938401e-01 1.06597051e-01 3.19968939e-01
-1.36618912e+00 9.37922001e-01 5.65655470e+00 7.57859588e-01
-1.29807484e+00 4.83958274e-01 1.08980823e+00 -5.66256940e-01
-8.85825455e-02 -6.25390589e-01 -6.83603048e-01 3.20546657e-01
1.40141666e+00 4.29356545e-02 1.52044058e-01 8.45208168e-01
2.58176357e-01 -6.98739812e-02 -7.87350953e-01 1.30144763e+00
3.22669625e-01 -7.99156606e-01 -5.18401384e-01 1.40232667e-02
9.34794068e-01 2.08238646e-01 3.84880513e-01 8.90789509e-01
2.05907188e-02 -1.02571464e+00 7.01421320e-01 2.88050741e-01
6.98471129e-01 -1.21901202e+00 9.58578587e-01 1.16360281e-03
-8.88509154e-01 -7.91052729e-02 -2.47465558e-02 -2.84789532e-01
1.27611414e-01 1.02918065e+00 -4.91086274e-01 6.41019702e-01
7.53678501e-01 6.91489220e-01 -5.03059089e-01 2.33732775e-01
5.55806421e-02 7.36898363e-01 -1.12881877e-01 -2.81929433e-01
3.40937018e-01 -9.43162441e-02 1.48982093e-01 1.73415911e+00
1.18219994e-01 1.31609246e-01 6.23368993e-02 3.64641875e-01
-5.64708173e-01 6.22921646e-01 -1.95348158e-01 -8.36253017e-02
2.26888731e-01 1.87729216e+00 -3.52603465e-01 -4.65299875e-01
-3.72972757e-01 8.57930720e-01 6.70824587e-01 1.60767242e-01
-1.24435854e+00 -5.29066503e-01 9.10918176e-01 -6.32239223e-01
3.79387289e-01 5.86788990e-02 -5.34919560e-01 -9.32938397e-01
-1.58992708e-01 -1.14097106e+00 4.60806221e-01 -4.89750445e-01
-1.48382807e+00 8.59295070e-01 -3.15568119e-01 -6.77437007e-01
-3.54008019e-01 -5.25131524e-01 -3.24797183e-01 7.75583565e-01
-1.34414494e+00 -1.16635275e+00 -2.74888217e-01 3.91951233e-01
3.39078158e-01 -8.16648602e-02 1.16831458e+00 6.14710331e-01
-7.97359645e-01 1.01198316e+00 5.01878187e-03 5.71143776e-02
1.03723669e+00 -1.08461940e+00 -2.29819089e-01 3.84848088e-01
-1.66188270e-01 1.16356336e-01 6.42767429e-01 -2.33357728e-01
-1.25076818e+00 -1.03155947e+00 9.14740026e-01 -4.85395312e-01
7.97295928e-01 -6.77332819e-01 -6.86548889e-01 4.00333881e-01
3.18978190e-01 -1.84437364e-01 1.29145443e+00 9.69758332e-01
-8.08798909e-01 -1.31435856e-01 -1.20421576e+00 3.48627269e-01
5.71518779e-01 -6.79003119e-01 -3.80087122e-02 2.46154338e-01
5.40709198e-01 -1.63167670e-01 -1.40561140e+00 4.34318632e-01
9.48429048e-01 -9.54767287e-01 5.57097375e-01 -7.50039577e-01
7.92941332e-01 3.49108607e-01 -6.09981120e-01 -1.44330585e+00
-3.18550676e-01 -3.35413963e-01 2.17722356e-02 1.62180054e+00
8.09488118e-01 -3.03569049e-01 3.57435197e-01 6.48422003e-01
2.09283590e-01 -1.08472514e+00 -8.33135366e-01 -3.12334657e-01
2.52870321e-01 -9.71138060e-01 3.24107081e-01 1.20271719e+00
3.67772192e-01 9.15519536e-01 -7.29487121e-01 -1.62286967e-01
2.68069267e-01 3.40297073e-02 5.69182158e-01 -1.25312018e+00
1.01444528e-01 -4.01481301e-01 -8.94766599e-02 -3.35898399e-01
6.99808478e-01 -7.60014534e-01 -3.34017694e-01 -1.02586889e+00
3.43367279e-01 -2.84295171e-01 -4.96977866e-01 5.70302248e-01
-2.16556937e-01 5.43855011e-01 2.44004354e-01 -2.36204714e-01
-1.09459889e+00 7.17319191e-01 6.78606272e-01 1.37002930e-01
-3.00722599e-01 -4.07302916e-01 -1.11287296e+00 5.58669984e-01
1.06463122e+00 -4.08322155e-01 -1.69206455e-01 -3.07210565e-01
5.95307767e-01 -1.34784490e-01 -2.03584999e-01 -6.44782662e-01
-2.68185616e-01 2.43063301e-01 3.35367620e-01 -1.74139574e-01
7.20437825e-01 -4.89292175e-01 -3.45003486e-01 -6.21631704e-02
-5.61754227e-01 3.36086750e-01 5.74200392e-01 3.22008431e-01
-1.47581771e-01 2.92106956e-01 6.90606236e-01 3.64525735e-01
-4.14822280e-01 -3.92669663e-02 -3.87032270e-01 3.61333519e-01
8.92116725e-01 4.74599361e-01 -4.14264232e-01 -5.60811460e-01
-6.39198780e-01 2.19407767e-01 -1.90879479e-01 8.28325450e-01
3.05570692e-01 -1.31576550e+00 -9.22200203e-01 -6.44849241e-02
3.39332640e-01 -8.59284222e-01 4.59252506e-01 1.20211446e+00
3.23597372e-01 3.30075771e-01 -1.50760934e-01 -4.15529937e-01
-1.50628996e+00 1.14975214e-01 1.88841924e-01 -6.30255163e-01
2.60812134e-01 7.53131211e-01 7.80056342e-02 -6.08784974e-01
8.12972933e-02 -2.44313665e-02 -4.25383687e-01 6.34793162e-01
5.02478540e-01 6.67009592e-01 3.54470223e-01 -9.64021802e-01
-7.12940395e-01 3.71897072e-01 -2.06393063e-01 -4.25312340e-01
1.39509177e+00 -3.48040432e-01 -1.89328715e-01 8.45150232e-01
1.73784673e+00 2.36615047e-01 -6.02194488e-01 -2.13502869e-01
-7.54681900e-02 -2.33953018e-02 4.73995984e-01 -1.43197608e+00
-1.05355442e+00 8.42174709e-01 7.44518042e-01 3.18097264e-01
1.10265529e+00 1.00791939e-01 7.84445167e-01 1.42788023e-01
-6.05646968e-02 -1.32785189e+00 1.54188514e-01 6.87885880e-01
8.71709466e-01 -1.50586283e+00 -1.77336797e-01 -2.32135788e-01
-1.29520273e+00 8.17890048e-01 4.81783897e-01 1.51869059e-01
7.51606464e-01 2.35634834e-01 5.33409894e-01 -4.78811413e-01
-1.08188176e+00 -1.79023489e-01 5.06164253e-01 7.90943280e-02
9.91035342e-01 3.50383282e-01 -5.52151263e-01 1.10928881e+00
-2.80621737e-01 -4.47764426e-01 1.45391837e-01 4.18334007e-01
-9.26803723e-02 -8.52813244e-01 -2.39136927e-02 4.16730821e-01
-1.06668389e+00 3.69019397e-02 -4.60879445e-01 5.89212775e-01
-9.52873752e-02 1.50944436e+00 1.80035934e-01 -7.31811643e-01
3.68870020e-01 2.87470430e-01 1.20736295e-02 -1.62596017e-01
-1.03170836e+00 2.81504333e-01 8.01806748e-01 -5.32827199e-01
-6.73112333e-01 -6.62653506e-01 -1.21819866e+00 -3.25510770e-01
-1.53813139e-01 3.65101099e-01 1.07537258e+00 9.57841158e-01
7.30099440e-01 7.01460004e-01 9.92279947e-01 -4.23032850e-01
-3.30924273e-01 -1.26920581e+00 -4.95089382e-01 6.96960688e-01
1.15513504e-01 -6.91329956e-01 -4.37691867e-01 -4.42459226e-01] | [13.104931831359863, 5.78048038482666] |
5c9217ba-3625-47ef-b716-f9e09cd469a9 | review-highlights-opinion-mining-on-reviews-a | null | null | https://dl.acm.org/citation.cfm?id=3158385 | http://vixra.org/pdf/1910.0514v1.pdf | Review highlights: opinion mining on reviews: a hybrid model for rule selection in aspect extraction | This paper proposes a methodology to extract key insights from user generated reviews. This work is based on Aspect Based Sentiment Analysis (ABSA) which predicts the sentiment of aspects mentioned in the text documents. The extracted aspects are fine-grained for the presentation form known as Review Highlights.
The syntactic approach for extraction process suffers from the overlapping chunking rules which result in noise extraction. We introduce a hybrid technique which combines machine learning and rule based model. A multi-label classifier identifies the effective rules which efficiently parse aspects and opinions from texts. This selection of rules reduce the amount of noise in extraction tasks.
This is a novel attempt to learn syntactic rule fitness from a corpus using machine learning for accurate aspect extraction. As the model learns the syntactic rule prediction from the corpus, it makes the extraction method domain independent. It also allows studying the quality of syntactic rules in a different corpus. | ['Amit Kushwaha', 'Shubham Chaudhary'] | 2017-10-18 | null | null | null | iml-17-october-1718-2017-liverpool-united | ['aspect-extraction', 'extract-aspect', 'extract-aspect-polarity-tuple'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 2.96262026e-01 5.00574052e-01 -4.02272582e-01 -5.68874061e-01
-6.40626788e-01 -7.53946662e-01 5.31690419e-01 6.13303602e-01
-2.79515266e-01 9.19090807e-01 5.50328135e-01 -4.19849366e-01
-1.26799136e-01 -8.34557116e-01 -2.98941523e-01 -3.14117938e-01
4.40022618e-01 4.49169368e-01 2.14867026e-01 -4.16318893e-01
7.94659436e-01 1.21776752e-01 -1.87607479e+00 1.01072371e+00
7.27400601e-01 7.67554402e-01 1.74320400e-01 8.09757829e-01
-1.01906455e+00 1.06054223e+00 -8.59421372e-01 -6.10187531e-01
7.95232505e-02 -4.82109636e-01 -9.83229578e-01 1.82989687e-01
-7.61708766e-02 3.07181984e-01 9.75305915e-01 8.23814094e-01
1.02276020e-01 -1.68122351e-01 1.10579741e+00 -1.02423489e+00
-6.84181228e-02 8.15926671e-01 -5.35773516e-01 5.66554405e-02
5.77200949e-01 -7.31625140e-01 1.17394853e+00 -7.06967950e-01
7.63346255e-01 9.69378471e-01 6.22086525e-01 2.78531522e-01
-3.80898684e-01 -2.91686803e-01 2.19746232e-01 4.91500832e-03
-6.94849133e-01 -1.29845530e-01 8.03534925e-01 -5.63587368e-01
1.55026245e+00 2.89906442e-01 7.92029321e-01 4.51285839e-01
5.91625869e-01 6.93292022e-01 1.60755372e+00 -9.40147877e-01
4.63314116e-01 8.53009820e-01 9.68543768e-01 4.31493044e-01
7.08200216e-01 -3.61454993e-01 -6.71716988e-01 -3.13255996e-01
-2.90438682e-01 -3.66997570e-01 2.44491979e-01 2.35614970e-01
-4.62496907e-01 9.66018260e-01 -6.67304337e-01 7.21496761e-01
-5.39998949e-01 -5.44294655e-01 7.91249275e-01 4.10551131e-01
5.74570835e-01 5.84241867e-01 -1.32862520e+00 -2.64113516e-01
-9.62526500e-01 2.42719144e-01 1.33642268e+00 8.75639915e-01
8.73918235e-01 -1.13223791e-01 1.09409936e-01 5.92381179e-01
6.03311360e-01 5.55300236e-01 6.59870327e-01 -2.76262045e-01
3.96654636e-01 1.28151464e+00 -3.76135632e-02 -9.91439641e-01
-4.50964630e-01 -1.84946224e-01 7.04076439e-02 2.57436812e-01
1.59126464e-02 -6.00096047e-01 -9.53461528e-01 8.94028127e-01
4.42647070e-01 -7.05278873e-01 3.74359041e-01 1.92899838e-01
1.03295803e+00 6.56791627e-01 3.88555050e-01 -6.04078591e-01
1.79954195e+00 -9.31730628e-01 -9.78833795e-01 -2.41239872e-02
7.17575431e-01 -1.25605476e+00 8.12020838e-01 8.26663196e-01
-6.82949245e-01 -3.79071593e-01 -1.14289331e+00 1.65490717e-01
-1.07592380e+00 4.40032601e-01 7.44882464e-01 7.28114903e-01
-5.71632266e-01 2.56242841e-01 -3.67780566e-01 -4.58018243e-01
6.65613413e-02 5.42447925e-01 -3.45707297e-01 4.07121956e-01
-7.91552603e-01 1.02187884e+00 4.71886933e-01 -4.56798136e-01
1.36906624e-01 -4.22000945e-01 -7.96875358e-01 -4.99117523e-02
4.42956477e-01 -3.59589815e-01 1.20236754e+00 -1.71215153e+00
-1.39465439e+00 8.95295918e-01 -5.41756332e-01 -2.47294322e-01
-3.84162247e-01 -2.73182601e-01 -5.74177504e-01 2.42188409e-01
3.38710189e-01 4.64155227e-02 1.05509472e+00 -1.26425755e+00
-1.17103171e+00 -5.34641862e-01 -4.94170338e-02 1.30365103e-01
-2.25609004e-01 4.76352811e-01 8.71667191e-02 -4.97921735e-01
-3.45843256e-01 -7.28323996e-01 -1.49875298e-01 -1.19588423e+00
-1.68564677e-01 -4.00744855e-01 9.00548816e-01 -5.31248391e-01
1.42367065e+00 -1.66646731e+00 -3.71557981e-01 4.36676800e-01
-2.39867978e-02 2.09170491e-01 2.60247946e-01 6.44387901e-01
1.14088748e-02 4.16275620e-01 -1.38478484e-02 5.04892245e-02
-1.52402133e-01 2.42238447e-01 -3.61055285e-01 -3.59805137e-01
1.03272229e-01 6.09380245e-01 -3.98941338e-01 -8.16404700e-01
-9.78347212e-02 2.71563232e-01 -3.35140765e-01 1.71801522e-01
-3.68406773e-01 -2.18516439e-01 -8.50397170e-01 4.92294222e-01
4.45167810e-01 2.48161450e-01 1.56152472e-01 -3.49486142e-01
-2.86677510e-01 5.69853067e-01 -1.12366700e+00 1.03454709e+00
-6.87220573e-01 3.50739986e-01 -2.24335343e-01 -1.02216482e+00
1.18359172e+00 5.63141704e-01 3.42723340e-01 -3.27858448e-01
4.72361684e-01 3.41421425e-01 -8.07092339e-02 -9.25271392e-01
6.87789381e-01 -5.81618667e-01 -2.95338392e-01 6.36486113e-01
4.11203861e-01 -2.73005068e-01 7.06350446e-01 4.62777317e-02
7.06790447e-01 5.46310067e-01 1.05257416e+00 -4.65897977e-01
9.32887554e-01 6.20243251e-01 6.06450140e-01 4.64371771e-01
9.61965546e-02 1.24350898e-02 7.48837471e-01 -7.18467534e-01
-8.71034503e-01 -1.50678307e-01 -1.86445061e-02 9.63397026e-01
-4.52526122e-01 -9.28083718e-01 -7.37142980e-01 -1.24373698e+00
-3.87198120e-01 6.83975279e-01 -7.28549182e-01 3.70710909e-01
-2.85540938e-01 -8.98762167e-01 -1.26447924e-03 2.36001201e-02
1.59788534e-01 -1.26447201e+00 -7.52271175e-01 2.98913181e-01
1.60757005e-01 -1.10369205e+00 2.55677462e-01 5.96020341e-01
-7.72331536e-01 -1.29383016e+00 1.19292609e-01 -6.74061775e-01
6.51675344e-01 -1.84565768e-01 1.25666106e+00 4.48817015e-02
-5.35755195e-02 2.74808109e-01 -1.07724261e+00 -1.29245353e+00
-6.09224498e-01 4.16213989e-01 -2.29097188e-01 -2.48218387e-01
1.37135136e+00 -4.06383365e-01 1.06993830e-02 -3.54835570e-01
-7.25726306e-01 -1.90207481e-01 8.04910362e-01 3.50854307e-01
6.02707446e-01 4.36570942e-01 7.79867649e-01 -1.69196689e+00
1.12202048e+00 -4.28247571e-01 -2.86817163e-01 4.20878559e-01
-1.05481291e+00 3.68929803e-01 7.75870860e-01 1.04491301e-01
-1.54106712e+00 1.23766735e-01 -1.69575989e-01 7.84224153e-01
-7.08137572e-01 6.99998379e-01 -7.99957067e-02 2.44821772e-01
6.44250154e-01 1.60624459e-02 -3.26619864e-01 -2.96308488e-01
4.94808108e-02 9.35982049e-01 -3.93242538e-01 -4.99629915e-01
3.18825334e-01 1.12153821e-01 -8.55090003e-03 -9.55255210e-01
-1.29814017e+00 -7.10167468e-01 -6.98508561e-01 -3.17970246e-01
8.94018650e-01 -5.88704228e-01 -1.63863808e-01 -2.42547505e-02
-9.80756521e-01 4.64558244e-01 -3.95380914e-01 5.52931428e-01
-3.64499539e-01 1.80072650e-01 -2.47715324e-01 -1.13825691e+00
-9.16504920e-01 -8.20714235e-01 9.88509357e-01 4.86525357e-01
-7.32783794e-01 -8.98193896e-01 3.65863055e-01 4.55695301e-01
9.94428433e-03 8.16451982e-02 1.02766705e+00 -1.28866386e+00
-1.65471081e-02 -3.07226330e-01 3.35659802e-01 3.63186806e-01
2.46298149e-01 3.85886252e-01 -9.47409153e-01 5.53988159e-01
4.05107230e-01 -1.16485193e-01 5.35149932e-01 3.49415094e-01
1.60989195e-01 -5.21213710e-01 -2.57024050e-01 -3.22229192e-02
1.74974430e+00 5.17383039e-01 2.65789539e-01 7.48034000e-01
4.08162206e-01 1.04865658e+00 1.01952696e+00 2.13156059e-01
2.37527147e-01 6.05367348e-02 -3.19306403e-01 4.30933595e-01
2.21765295e-01 -9.65500530e-03 4.29732114e-01 1.22195601e+00
-7.20226318e-02 2.85326578e-02 -7.83752680e-01 6.42308354e-01
-1.63192701e+00 -8.54687214e-01 -4.26347017e-01 1.51952517e+00
9.19225395e-01 5.20529568e-01 1.38181478e-01 5.42801023e-01
2.25373268e-01 -2.41788533e-02 3.26295793e-01 -1.46798968e+00
-1.43350493e-02 7.80169129e-01 3.14717859e-01 7.12345064e-01
-1.06647742e+00 9.40547347e-01 5.58212662e+00 7.28669047e-01
-7.62895823e-01 1.58871502e-01 3.06289077e-01 1.86400294e-01
-5.00530839e-01 2.74419308e-01 -1.38763809e+00 1.00453585e-01
1.22669041e+00 -7.85486475e-02 -3.34817797e-01 1.07411480e+00
3.15387040e-01 -7.30438471e-01 -5.19967914e-01 3.38221759e-01
4.61313188e-01 -1.16769361e+00 4.25701082e-01 -1.06614083e-01
9.43822384e-01 -2.92902708e-01 -6.32461190e-01 2.60510236e-01
1.29372478e-01 -6.90009654e-01 3.59490305e-01 4.79101747e-01
-5.04442304e-02 -1.24945033e+00 1.33814824e+00 3.48000258e-01
-1.02920222e+00 8.43670070e-02 -3.80829215e-01 -3.58368635e-01
1.23021990e-01 9.81346786e-01 -7.69670367e-01 7.00129449e-01
6.28859937e-01 6.09597981e-01 -6.46097362e-01 5.40675700e-01
-4.17747289e-01 6.47118032e-01 -2.01757953e-01 -6.21885836e-01
2.11883217e-01 -3.26768100e-01 4.99135762e-01 1.56143260e+00
1.58126920e-01 7.36511573e-02 -2.61653960e-02 1.57397196e-01
3.98429871e-01 1.03388298e+00 -9.29347813e-01 -7.64621124e-02
-1.22841388e-01 1.62237632e+00 -1.27513647e+00 -4.28525299e-01
-5.98990977e-01 4.19834673e-01 4.94856201e-02 6.34152209e-03
7.73613825e-02 -5.98791540e-01 1.02194890e-01 -8.26990828e-02
6.91141129e-01 1.33076146e-01 -7.78118074e-01 -1.08702874e+00
1.84423715e-01 -1.17361665e+00 3.50645602e-01 -4.26492542e-01
-1.03199399e+00 1.09423089e+00 -2.00059582e-02 -1.14932394e+00
-3.64905328e-01 -8.95042539e-01 -6.64633453e-01 5.49515784e-01
-1.42182529e+00 -1.28280163e+00 2.29367062e-01 1.55285835e-01
8.10515225e-01 -5.82081437e-01 1.03826261e+00 -1.56790525e-01
-1.12024315e-01 4.68472317e-02 -5.04720628e-01 -1.95145041e-01
5.69274187e-01 -1.52351022e+00 -3.26905459e-01 8.63564432e-01
1.90819561e-01 6.30227387e-01 8.81676793e-01 -9.73810732e-01
-9.20781136e-01 -5.50551951e-01 1.72567391e+00 -7.01044381e-01
5.04662395e-01 9.49480198e-03 -5.91604888e-01 2.60080785e-01
7.24322975e-01 -6.63653970e-01 1.60841322e+00 3.31969440e-01
-2.55388796e-01 -1.77357867e-01 -9.48358595e-01 1.38639003e-01
1.86914098e-04 -3.34168524e-01 -1.29471624e+00 -3.70251723e-02
3.75397474e-01 2.94221967e-01 -7.42031276e-01 2.73916870e-01
6.35582805e-01 -9.98789787e-01 1.59215465e-01 -7.48408437e-01
7.17225373e-01 -4.55747187e-01 2.05867030e-02 -1.16697216e+00
2.80764610e-01 -4.13812488e-01 -1.07489906e-01 1.68735039e+00
1.20547438e+00 -1.65445223e-01 6.37945950e-01 6.43199325e-01
3.40532333e-01 -8.82512867e-01 -2.83272713e-01 -1.15594424e-01
-1.71819091e-01 -6.03906512e-01 3.97800863e-01 7.29870915e-01
4.78940547e-01 1.15240347e+00 -4.41986099e-02 -2.72464484e-01
2.92726606e-01 4.25678313e-01 5.27536392e-01 -1.41508377e+00
-1.71540231e-01 -2.05215946e-01 -2.15263799e-01 -1.63610756e-01
2.19961300e-01 -5.21914124e-01 -1.74443021e-01 -1.65241873e+00
1.01410814e-01 -7.45979100e-02 1.14230826e-01 3.31465065e-01
-2.75771230e-01 -2.43574634e-01 5.32200485e-02 -6.60627633e-02
-5.15097558e-01 4.76552621e-02 8.94292116e-01 1.74435437e-01
-5.17021716e-01 3.32644820e-01 -1.16636634e+00 1.14304566e+00
1.07465959e+00 -9.49905455e-01 -7.24820912e-01 3.12170714e-01
1.13712311e+00 -4.16558266e-01 -7.01766014e-01 -6.36921406e-01
-3.02152634e-02 -1.97377220e-01 2.94314444e-01 -7.39317477e-01
-3.27175021e-01 -1.24837053e+00 -3.01553816e-01 1.31110311e-01
-1.94787532e-01 2.53906488e-01 1.75241157e-01 2.86387086e-01
-4.77774203e-01 -1.03224719e+00 2.88048536e-01 -5.40439069e-01
-5.34546614e-01 -2.97524869e-01 -8.29618275e-01 9.78594739e-03
1.00320780e+00 -2.49302864e-01 -5.62171601e-02 -2.91277945e-01
-6.57824159e-01 -1.90366521e-01 8.02089944e-02 3.12558711e-01
2.61768669e-01 -7.21259058e-01 -4.73781437e-01 -1.37010729e-02
1.30526438e-01 -2.56310612e-01 -1.98480502e-01 5.68145156e-01
-5.25839806e-01 5.69606781e-01 -1.38316035e-01 1.15077779e-01
-1.55522406e+00 6.51981175e-01 -1.77832115e-02 -7.39297152e-01
-2.49590248e-01 3.77364308e-01 -3.89565587e-01 -2.52291441e-01
-2.34121934e-01 -3.53971660e-01 -1.37800539e+00 6.30968273e-01
7.01173961e-01 6.82639182e-02 2.85840660e-01 -7.40394890e-01
-2.25131020e-01 9.10157621e-01 -3.03844869e-01 -2.77189791e-01
1.51469946e+00 -2.26038530e-01 -5.22342980e-01 6.96687758e-01
9.46668267e-01 6.96710706e-01 -3.23334605e-01 1.76943764e-01
6.50037825e-01 6.47185668e-02 6.57468243e-03 -1.01653135e+00
-5.84536612e-01 4.86209154e-01 4.12671342e-02 5.60066998e-01
1.13400471e+00 -1.05885155e-01 4.97562647e-01 5.47141194e-01
3.36075909e-02 -1.70113671e+00 -5.45164526e-01 6.11014068e-01
5.42832553e-01 -1.34013975e+00 4.76780087e-01 -5.91794610e-01
-9.26420391e-01 1.43975937e+00 5.70137084e-01 -1.84182554e-01
1.29867923e+00 8.86525989e-01 4.17010367e-01 -6.70628786e-01
-9.99152064e-01 -4.47328299e-01 4.11233634e-01 7.29106665e-01
7.51524150e-01 5.25257848e-02 -1.34441757e+00 1.37546945e+00
-4.91993934e-01 6.49422929e-02 7.19037652e-01 1.23343289e+00
-8.65100920e-01 -1.52645302e+00 -1.92649603e-01 8.32044899e-01
-1.30904853e+00 -2.74943858e-01 -1.00646067e+00 6.03284955e-01
4.92157936e-01 1.25736117e+00 -5.11349618e-01 -2.40187570e-01
2.91603953e-01 5.80478907e-01 1.21477991e-01 -1.00908136e+00
-1.27245426e+00 2.77011544e-01 7.04508185e-01 -2.58964542e-02
-1.22717357e+00 -6.63693070e-01 -1.24613464e+00 2.51669139e-01
-3.87128830e-01 9.64238584e-01 9.25893128e-01 1.55757749e+00
3.68866950e-01 4.10399526e-01 5.50701082e-01 -2.52889376e-02
1.58297822e-01 -1.08751476e+00 -4.56578612e-01 2.02889174e-01
1.69734120e-01 -3.35276693e-01 -2.62982666e-01 4.85138714e-01] | [11.172515869140625, 6.8369011878967285] |
660e5c61-fe7e-4fb2-a1f3-1c330123fe2e | lidar-range-image-compression-with-deep-delta | null | null | https://openreview.net/forum?id=nzqZufLU1v | https://openreview.net/pdf?id=nzqZufLU1v | Lidar Range Image Compression with Deep Delta Encoding | Lidars are widely used in applications such as autonomous driving and augmented reality. However, the large volume of data produced by lidars can lead to high cost in data storage and transmission. Besides rising standards in point cloud compression from the MPEG (G-PCC and V-PCC), recent works also explore using deep networks to improve the compression rates. However, most prior work focus on the generic point cloud representation, neglecting the spatial patterns of the points from lidar range images. In this work, we leverage the range image representation and propose a novel deep delta encoding model to compress lidar data. Our deep model takes in local range image patches and predicts the next pixel value in a raster-scanning manner. The residuals between the prediction and the original value can be entropy encoded to achieve lossless compression under certain quantization rates. Evaluated on the Waymo Open Dataset and KITTI, our method demonstrates significant improvement compared to widely algorithms as well as recent deep methods based on the point cloud representation, in both the point cloud reconstruction quality and the downstream perception model performance. | ['Dragomir Anguelov', 'Yin Zhou', 'Charles R. Qi', 'Xuanyu Zhou'] | 2021-09-29 | null | null | null | null | ['point-cloud-reconstruction'] | ['computer-vision'] | [ 4.40421999e-01 -1.97526321e-01 -2.59355187e-01 -5.24883151e-01
-5.56591630e-01 -1.70924336e-01 3.59583765e-01 1.86311573e-01
-3.08079302e-01 5.55777252e-01 1.22429952e-01 -1.78498358e-01
-9.70157385e-02 -1.41985714e+00 -1.11935401e+00 -3.13631892e-01
-8.24449286e-02 4.62211818e-01 3.24708611e-01 -2.12694749e-01
4.28251296e-01 9.04227972e-01 -1.91465974e+00 1.65344730e-01
9.54826117e-01 1.44901025e+00 4.20682400e-01 4.74362493e-01
-3.31914574e-01 5.88428259e-01 -3.90494734e-01 -4.06482279e-01
6.63390756e-01 2.02253610e-01 -3.91531646e-01 -1.72193542e-01
6.30648434e-01 -1.04560256e+00 -9.62350190e-01 1.01170254e+00
6.02098525e-01 1.78625286e-02 2.04059720e-01 -1.30789554e+00
-5.37597895e-01 2.07808912e-01 -6.41878605e-01 -1.57325044e-01
3.58667225e-02 1.58904389e-01 6.80685818e-01 -8.44466209e-01
6.91394746e-01 1.22117901e+00 7.26797998e-01 -2.92279292e-02
-7.56824136e-01 -9.64353740e-01 -4.04748380e-01 7.87247956e-01
-1.54531193e+00 -4.23963875e-01 5.89351416e-01 3.54552530e-02
1.10030043e+00 1.28356263e-01 8.22977483e-01 4.40265477e-01
2.72288173e-01 2.61809677e-01 6.13725483e-01 1.06525376e-01
3.56327623e-01 -3.36248606e-01 -4.10239309e-01 2.07383141e-01
3.64638716e-01 5.10206640e-01 -8.35871696e-01 2.02649668e-01
6.62215054e-01 1.98440865e-01 -1.64214686e-01 -3.24238062e-01
-9.32546198e-01 7.47882783e-01 8.89507771e-01 -3.02679449e-01
-4.74661916e-01 7.33714998e-01 3.28682572e-01 2.94595927e-01
4.08586204e-01 -3.60074431e-01 -1.64096221e-01 -3.61010283e-01
-1.31598115e+00 3.71630877e-01 3.32456440e-01 1.44459653e+00
1.03835869e+00 -9.61873904e-02 1.42475113e-01 6.92715168e-01
5.24892032e-01 8.08029890e-01 1.31853685e-01 -1.68364215e+00
8.58680427e-01 4.21408534e-01 -1.47187352e-01 -1.00963950e+00
2.88742855e-02 -1.87993065e-01 -1.00176597e+00 4.37797487e-01
-4.25572455e-01 4.14525449e-01 -8.69396210e-01 1.20499074e+00
2.46136382e-01 4.71923321e-01 2.02613205e-01 7.80505478e-01
7.10599780e-01 9.81712997e-01 -3.30725253e-01 9.90890083e-04
9.36050653e-01 -5.29285550e-01 -8.28494251e-01 -1.44593954e-01
5.77858686e-01 -6.21677160e-01 7.34883010e-01 3.19985867e-01
-1.41759086e+00 -6.76109016e-01 -1.40705502e+00 -9.82657135e-01
-2.89272815e-01 -5.15982866e-01 2.43563280e-01 2.72071511e-01
-1.19656789e+00 8.57017219e-01 -8.50573540e-01 1.21620875e-02
6.98276758e-01 3.84290874e-01 -9.31893140e-02 -7.48874962e-01
-1.04871643e+00 7.68091857e-01 3.30374599e-01 -5.38224727e-02
-4.16109920e-01 -1.00542796e+00 -8.20613086e-01 4.41829979e-01
-4.33168858e-02 -6.54978693e-01 1.09321249e+00 3.15806754e-02
-1.13483071e+00 2.50209242e-01 -2.23473459e-01 -1.02777135e+00
4.65194762e-01 -2.06454769e-01 -1.06420696e-01 4.65386599e-01
-2.63312068e-02 1.56376779e+00 4.75849211e-01 -1.17229402e+00
-9.27696347e-01 -3.54430348e-01 -1.24924213e-01 3.75478238e-01
-1.33420927e-02 -6.55386031e-01 -3.98852497e-01 -2.14278668e-01
4.76918846e-01 -8.07125151e-01 -1.15055986e-01 9.58577871e-01
5.98369762e-02 4.58909161e-02 1.42371976e+00 -4.30506825e-01
6.95831001e-01 -2.51848841e+00 -3.28996122e-01 4.22906131e-03
1.00093879e-01 1.70070753e-01 -3.63539942e-02 4.33426380e-01
2.94034958e-01 1.98746920e-01 -5.30854404e-01 -4.92841393e-01
1.53349295e-01 8.50819409e-01 -8.07325542e-01 4.11536187e-01
6.68585002e-02 8.05335402e-01 -5.89128673e-01 -6.45014346e-01
6.64444089e-01 9.86423135e-01 -7.88840830e-01 -1.24991238e-01
-1.65095329e-01 1.06136397e-01 -1.32032726e-02 5.83442688e-01
1.52984405e+00 6.75869286e-02 -2.26359606e-01 -2.83084333e-01
-3.08361560e-01 6.07640386e-01 -7.03235149e-01 1.99322426e+00
-3.93555850e-01 1.02298093e+00 -6.16956614e-02 -5.38849473e-01
8.94458354e-01 7.83738680e-04 7.11482763e-01 -1.04182708e+00
-2.22892329e-01 3.43782365e-01 -2.40232795e-01 -1.17895156e-01
1.30903113e+00 1.25558540e-01 8.52606595e-02 -1.37917355e-01
-3.31241101e-01 -7.66659915e-01 4.25272435e-02 2.17650503e-01
8.68611693e-01 8.15154761e-02 -1.26455456e-01 4.99567926e-01
1.23040624e-01 1.85595199e-01 4.95096236e-01 3.88463229e-01
-1.43312395e-01 7.83177197e-01 9.11262780e-02 -2.08193928e-01
-1.41742706e+00 -1.21983290e+00 -5.03228843e-01 2.51936406e-01
7.31848538e-01 -4.65093791e-01 -2.17046991e-01 4.37599048e-02
3.30051720e-01 9.35024917e-01 1.41420767e-01 -2.84965843e-01
-6.49302006e-01 6.60373867e-02 6.43072426e-01 4.94132251e-01
1.08840644e+00 -6.94012761e-01 -8.42611432e-01 1.24011457e-01
-3.96032959e-01 -1.44507194e+00 -8.72900635e-02 5.40991537e-02
-1.29635274e+00 -5.38544416e-01 -1.79576278e-01 -4.00029063e-01
8.73279646e-02 6.45854235e-01 8.28755498e-01 1.08430926e-02
1.76969320e-02 5.68486191e-02 -3.46093029e-01 -5.07295370e-01
-1.95777252e-01 -1.99310556e-01 -2.76115000e-01 -5.57504952e-01
4.26330686e-01 -8.77380431e-01 -8.14324737e-01 -1.01638883e-01
-1.17166185e+00 8.03212374e-02 7.14820743e-01 2.08931595e-01
1.21450436e+00 2.62841284e-01 6.62302896e-02 -1.99908525e-01
1.49904460e-01 -5.57876110e-01 -6.83078110e-01 -5.07490277e-01
-8.92709434e-01 -1.48425978e-02 2.21762910e-01 2.19360352e-01
-5.25961161e-01 8.00589919e-02 -4.76727098e-01 -8.39227378e-01
-4.19427827e-02 3.93702745e-01 -2.68171337e-02 -3.23015243e-01
-3.00267478e-03 4.59706873e-01 8.12498666e-03 -2.68435359e-01
5.44602990e-01 6.85563505e-01 8.64836574e-01 -3.32645886e-03
8.84114146e-01 8.57064426e-01 5.16221106e-01 -8.75677407e-01
-5.92867211e-02 -2.93787897e-01 -4.64560747e-01 -1.56131953e-01
6.58728302e-01 -1.22349179e+00 -6.87056839e-01 1.09257750e-01
-1.60475242e+00 1.64628997e-01 -7.41480708e-01 5.42052031e-01
-8.42287958e-01 7.42676735e-01 -5.16365945e-01 -4.45284635e-01
-3.36848557e-01 -1.26523316e+00 1.37637949e+00 -1.20379865e-01
3.46734166e-01 -2.07706302e-01 -1.06150798e-01 1.50997281e-01
3.79254282e-01 1.94783017e-01 7.89825559e-01 1.92841291e-01
-1.48183179e+00 -9.20387283e-02 -6.24617100e-01 3.54929000e-01
-2.87429839e-01 -1.83418006e-01 -8.54942083e-01 -1.38124555e-01
1.30538382e-02 -1.63359106e-01 8.52697670e-01 3.03921163e-01
1.50229836e+00 -2.20583409e-01 -1.77791163e-01 1.36802483e+00
1.72968280e+00 1.92287549e-01 1.26328874e+00 1.73733696e-01
6.23865008e-01 3.65051419e-01 7.42359400e-01 4.99778330e-01
6.48497701e-01 5.25292873e-01 1.20883465e+00 3.30800205e-01
-4.08994704e-01 -4.61577296e-01 2.09282458e-01 1.06054866e+00
9.84385833e-02 -3.97534698e-01 -5.98619521e-01 3.94052088e-01
-1.46956301e+00 -9.61337030e-01 -2.53390253e-01 2.19910407e+00
4.44733828e-01 3.37186232e-02 -6.93882644e-01 3.31617981e-01
4.42119658e-01 4.53347832e-01 -8.83717537e-01 -4.90336448e-01
-2.41188601e-01 4.33864027e-01 1.05365133e+00 3.44966233e-01
-4.56915915e-01 6.37621403e-01 5.85681629e+00 8.09423625e-01
-1.26913881e+00 1.70061111e-01 8.69013220e-02 -6.41710460e-01
-3.68622690e-01 -4.04350124e-02 -7.78945088e-01 6.35433793e-01
1.48969638e+00 -3.84239763e-01 3.83563846e-01 6.89417541e-01
3.29722166e-01 -5.03557585e-02 -9.09039021e-01 1.32322121e+00
-1.80625454e-01 -1.67985046e+00 2.50345796e-01 8.13448727e-01
5.09059250e-01 6.84076428e-01 3.01625341e-01 1.22179307e-01
-7.29564652e-02 -7.84563243e-01 9.85231936e-01 2.92523146e-01
1.17299092e+00 -9.54932034e-01 5.15024841e-01 3.49118441e-01
-1.35019934e+00 -2.57859360e-02 -1.06041443e+00 1.52621204e-02
5.47967374e-01 7.28921950e-01 -8.00757766e-01 7.75207877e-01
8.83099675e-01 9.51311946e-01 -2.25138769e-01 1.09224868e+00
1.79843619e-01 3.24270934e-01 -7.64195204e-01 5.81952274e-01
2.19705313e-01 -1.93320900e-01 4.57111567e-01 7.89976001e-01
1.07876492e+00 5.36319353e-02 -2.80864507e-01 8.92490149e-01
-4.25575405e-01 -2.34031305e-01 -1.04711485e+00 2.48972565e-01
8.33327413e-01 6.10210001e-01 -1.99844152e-01 -3.09806973e-01
-4.78087813e-01 7.81103432e-01 8.20378587e-03 6.45850003e-02
-8.26575220e-01 -3.55309606e-01 1.00168562e+00 4.38258380e-01
8.03252161e-01 -6.66494846e-01 -6.96152747e-01 -6.71651423e-01
2.56536365e-01 -2.94305354e-01 -2.47349650e-01 -1.15856087e+00
-6.31355047e-01 2.94558138e-01 1.20833397e-01 -1.87581015e+00
-3.06488603e-01 -1.85838848e-01 -7.45238885e-02 7.56660938e-01
-2.35047770e+00 -8.29630375e-01 -7.04625428e-01 5.12129486e-01
4.74327981e-01 2.88581908e-01 3.63937676e-01 7.51307070e-01
1.83257520e-01 1.95337325e-01 3.11782986e-01 -4.93138164e-01
4.60627288e-01 -6.79329515e-01 7.70310879e-01 8.01291466e-01
-1.01685666e-01 1.72276750e-01 4.58108902e-01 -6.72374487e-01
-1.55589581e+00 -1.48604250e+00 7.66333699e-01 2.36699164e-01
-9.14923660e-03 -3.24198566e-02 -1.05269742e+00 4.53191012e-01
1.74151912e-01 3.22162122e-01 2.90857673e-01 -8.68092954e-01
-2.21668750e-01 -3.62349063e-01 -1.44742918e+00 2.94489235e-01
1.24993670e+00 -5.11854053e-01 -1.66789979e-01 2.36985147e-01
1.27392280e+00 -6.63041294e-01 -1.02485633e+00 6.71563625e-01
4.04470265e-01 -1.29263794e+00 1.20007479e+00 3.08828324e-01
8.63651633e-01 -4.45636094e-01 -7.87304044e-01 -9.20841277e-01
-2.10908011e-01 -9.58883539e-02 -4.75252628e-01 8.67699742e-01
-2.35832974e-01 -4.43395138e-01 8.63975823e-01 2.61467159e-01
-6.10971987e-01 -5.80290854e-01 -1.44275796e+00 -6.51715636e-01
-6.05542660e-02 -8.30382466e-01 1.26235354e+00 4.15054321e-01
-5.78372121e-01 -7.02152774e-02 -2.10131526e-01 5.76119959e-01
8.13313007e-01 9.31972861e-02 7.28705525e-01 -1.18664563e+00
2.96608090e-01 -2.11256281e-01 -9.68456745e-01 -1.58124352e+00
-7.99864084e-02 -9.44753587e-01 1.40503779e-01 -1.70528328e+00
-4.81392831e-01 -6.35274112e-01 3.12817805e-02 3.35930735e-02
4.83285874e-01 4.99307841e-01 4.90455687e-01 5.45097291e-01
-1.84114814e-01 9.33215737e-01 1.22232974e+00 -3.63655627e-01
1.22819036e-01 -3.96019042e-01 -2.96340108e-01 3.60078782e-01
6.56221628e-01 -5.35773456e-01 -5.67663074e-01 -1.02578676e+00
2.74442613e-01 3.01918656e-01 5.71438074e-01 -1.58808637e+00
4.96444196e-01 3.37670208e-03 2.86174715e-01 -1.58252621e+00
1.05117249e+00 -1.26048470e+00 4.09263372e-01 5.08980572e-01
-2.39938172e-03 2.22558916e-01 1.97140723e-01 8.57387185e-01
-4.21353817e-01 -1.48744479e-01 7.79377699e-01 1.51600957e-01
-8.37561846e-01 7.48421848e-01 6.43280372e-02 -5.24017215e-01
8.97375286e-01 -6.03450835e-01 -4.82806772e-01 -6.10672414e-01
-8.63962919e-02 1.21849477e-01 7.96451211e-01 2.46895269e-01
1.11517763e+00 -1.50553191e+00 -5.69760978e-01 4.65042472e-01
6.01079118e-05 7.71575868e-01 3.58858943e-01 4.95477825e-01
-1.18540800e+00 5.17710567e-01 -2.62955487e-01 -9.79824722e-01
-9.52016354e-01 5.06253719e-01 -2.08177492e-02 1.51473030e-01
-9.24645305e-01 4.25917715e-01 -1.87900633e-01 -7.07034245e-02
1.95337638e-01 -7.73736358e-01 3.06813836e-01 -3.47635329e-01
3.79596144e-01 5.51953137e-01 2.65891969e-01 -8.03401172e-01
-4.39069979e-02 6.84750676e-01 2.08743975e-01 -1.50241569e-01
1.26491988e+00 -4.61118132e-01 -2.08381832e-01 1.18955933e-01
1.54112232e+00 -1.90346777e-01 -1.43079674e+00 -8.19323286e-02
-4.09532666e-01 -9.00640547e-01 4.67288852e-01 -1.96178436e-01
-1.27896678e+00 1.19748545e+00 8.99934292e-01 -6.82720393e-02
1.13352501e+00 -3.53249311e-01 1.54326415e+00 4.15693492e-01
6.99857652e-01 -7.28313088e-01 -4.15498972e-01 4.82160389e-01
6.99676991e-01 -1.01137221e+00 3.26577932e-01 -6.36944175e-01
-1.98243141e-01 9.98673499e-01 1.72420561e-01 -9.96611491e-02
6.33071780e-01 3.88332248e-01 -2.92565256e-01 -3.89824286e-02
-8.09943438e-01 -4.78760116e-02 -2.75177062e-01 8.71789098e-01
-1.21927358e-01 -1.12338081e-01 -1.24021634e-01 -3.03766221e-01
-6.52977824e-01 5.26618123e-01 5.95526338e-01 1.16735804e+00
-6.91887200e-01 -1.01884055e+00 -3.57085377e-01 6.06360674e-01
1.17579447e-02 -2.51337171e-01 1.71834975e-01 5.42778671e-01
3.67535979e-01 8.15586209e-01 7.32124567e-01 -5.46895564e-01
2.64054477e-01 -3.49175870e-01 3.24759007e-01 -2.76222885e-01
-7.49505088e-02 -3.12724411e-01 -2.61267275e-01 -9.25971031e-01
-3.47350359e-01 -4.91368622e-01 -1.54545391e+00 -9.79544222e-01
9.80191678e-02 -2.58122474e-01 1.20733440e+00 4.37709183e-01
8.62228513e-01 3.41308951e-01 5.16045630e-01 -1.09841383e+00
-3.84921551e-01 -7.14943230e-01 -5.17711222e-01 1.19161651e-01
5.16092181e-01 -4.31779534e-01 -2.12586373e-01 3.78240505e-03] | [8.297162055969238, -2.9792494773864746] |
57ea74f5-6312-40f6-85f6-bbc97bb1ea65 | self-supervised-cnn-for-unconstrained-3d | 1808.05323 | null | https://arxiv.org/abs/1808.05323v3 | https://arxiv.org/pdf/1808.05323v3.pdf | 3D Face From X: Learning Face Shape from Diverse Sources | We present a novel method to jointly learn a 3D face parametric model and 3D face reconstruction from diverse sources. Previous methods usually learn 3D face modeling from one kind of source, such as scanned data or in-the-wild images. Although 3D scanned data contain accurate geometric information of face shapes, the capture system is expensive and such datasets usually contain a small number of subjects. On the other hand, in-the-wild face images are easily obtained and there are a large number of facial images. However, facial images do not contain explicit geometric information. In this paper, we propose a method to learn a unified face model from diverse sources. Besides scanned face data and face images, we also utilize a large number of RGB-D images captured with an iPhone X to bridge the gap between the two sources. Experimental results demonstrate that with training data from more sources, we can learn a more powerful face model. | ['Juyong Zhang', 'Yudong Guo', 'Lin Cai'] | 2018-08-16 | null | null | null | null | ['3d-face-modeling'] | ['computer-vision'] | [-2.17709422e-01 -4.37743478e-02 -1.36735111e-01 -8.12749445e-01
-7.85383284e-01 -2.81142086e-01 1.12484030e-01 -9.81550336e-01
1.00395024e-01 2.60401785e-01 -2.01144084e-01 2.98069328e-01
4.27977964e-02 -8.89655709e-01 -7.91841865e-01 -5.43521881e-01
3.02367270e-01 6.23606026e-01 -2.26275757e-01 8.39501247e-02
-2.42404610e-01 8.50329220e-01 -1.78591812e+00 -2.73741540e-02
3.59076440e-01 1.33155155e+00 1.03341401e-01 2.89881211e-02
-4.66097891e-01 1.93973139e-01 -3.12733412e-01 -3.80389094e-01
5.83965957e-01 -3.26350600e-01 -1.85992613e-01 8.33636224e-01
6.58087671e-01 -8.72083068e-01 -3.96765649e-01 9.93510902e-01
6.47045135e-01 -2.84990996e-01 4.42289174e-01 -1.51249957e+00
-7.37544239e-01 -2.67757736e-02 -9.56649721e-01 -5.41229069e-01
4.41062570e-01 1.00825556e-01 2.00683862e-01 -1.29787624e+00
6.07464075e-01 1.69704640e+00 7.10251570e-01 8.50457728e-01
-8.83708656e-01 -1.09661412e+00 2.86723822e-02 -1.93865702e-01
-1.72209632e+00 -7.71653056e-01 1.31400573e+00 -2.18093470e-01
-2.64447182e-02 -1.55605478e-02 7.56152928e-01 1.09125555e+00
-5.08601427e-01 7.12789118e-01 1.18403518e+00 -2.24585995e-01
-6.53452575e-02 3.47457319e-01 -4.93032515e-01 9.40085590e-01
1.79556340e-01 -5.92801254e-03 -6.43034935e-01 -2.11607471e-01
1.47610486e+00 4.77023661e-01 -3.88516128e-01 -3.79018992e-01
-6.70861065e-01 5.46878457e-01 2.45963156e-01 6.15990050e-02
-2.85435140e-01 -1.33668289e-01 -1.62075430e-01 1.41970485e-01
5.16447723e-01 -4.47415233e-01 -4.67122167e-01 1.95655987e-01
-1.00001597e+00 3.41295414e-02 6.00341499e-01 1.35003936e+00
1.18705630e+00 8.79883468e-02 5.06431580e-01 9.74841714e-01
8.73540878e-01 1.05178702e+00 2.15800896e-01 -8.84029627e-01
4.43236738e-01 6.80715978e-01 3.80427204e-02 -1.09366930e+00
-4.93796170e-03 1.06802322e-01 -7.67973304e-01 1.39023945e-01
4.01804209e-01 -4.36272733e-02 -7.81466424e-01 1.65191841e+00
6.90415382e-01 4.28550929e-01 -2.92220235e-01 1.08661687e+00
1.01422942e+00 4.83716518e-01 -4.17706192e-01 -4.11908656e-01
1.17408907e+00 -5.49284816e-01 -7.81495214e-01 -1.40978515e-01
7.88363516e-02 -6.45453691e-01 1.01332521e+00 3.38219255e-01
-1.20118725e+00 -7.09390163e-01 -7.44252384e-01 -1.81655645e-01
2.02897424e-03 3.58195394e-01 3.00493866e-01 7.16790795e-01
-9.05703187e-01 2.85737038e-01 -6.61246002e-01 -1.14349626e-01
7.87908971e-01 4.01212871e-01 -7.65154481e-01 -5.37271798e-01
-6.47561073e-01 4.98859823e-01 -1.83603495e-01 2.16263622e-01
-9.30925786e-01 -7.53883779e-01 -8.45105708e-01 -2.72716373e-01
5.45773566e-01 -4.20890272e-01 1.09176874e+00 -9.48608875e-01
-1.50603473e+00 1.15604866e+00 -2.68407583e-01 6.35893524e-01
5.37971318e-01 -6.68912530e-02 -3.86551380e-01 2.50676632e-01
-1.66481033e-01 4.88278151e-01 1.21255255e+00 -1.61757779e+00
-9.07602683e-02 -1.14811397e+00 -1.38096437e-01 -1.48345321e-01
-5.06621242e-01 8.24280381e-02 -9.19832885e-01 -3.40825737e-01
4.32619214e-01 -7.41310894e-01 2.33890668e-01 8.67790759e-01
-1.81883186e-01 -5.69992661e-02 1.38481796e+00 -5.89830339e-01
5.18047035e-01 -2.18111205e+00 -1.18344566e-02 1.67358011e-01
1.66993394e-01 2.09589928e-01 -3.88814479e-01 -1.41928438e-02
4.97205928e-02 2.00182140e-01 -7.95186162e-02 -7.63771892e-01
-6.19047768e-02 3.74748439e-01 -1.93693176e-01 4.84661520e-01
2.31354430e-01 7.09727883e-01 -5.91606975e-01 -8.57911706e-01
3.84874232e-02 9.47365224e-01 -3.26367855e-01 4.39185888e-01
-1.58741012e-01 5.86007535e-01 -8.73497546e-01 1.22602093e+00
1.12027872e+00 -1.62250727e-01 1.06480131e-02 -3.21690977e-01
2.65868753e-01 -3.81730497e-01 -1.35817719e+00 2.03650475e+00
-3.74279857e-01 7.74014220e-02 4.97370034e-01 -7.79403389e-01
1.18500924e+00 5.18151522e-01 6.44619226e-01 -4.64972764e-01
3.39064091e-01 2.96334654e-01 -6.29265010e-01 -7.81362534e-01
-5.97289801e-02 -4.28738713e-01 3.32817614e-01 6.36481762e-01
1.61590934e-01 -3.34419042e-01 -3.02882850e-01 -3.59497488e-01
3.80985081e-01 4.02093083e-01 -4.77781817e-02 1.09064221e-01
4.44601148e-01 -4.20124650e-01 9.07823205e-01 -1.56632233e-02
-3.24494839e-02 8.32395017e-01 1.75763041e-01 -4.90829140e-01
-8.45004559e-01 -8.70434940e-01 -2.81641841e-01 5.20205617e-01
1.32518619e-01 -4.45642829e-01 -7.92062998e-01 -7.48533726e-01
8.11337084e-02 -1.33326009e-01 -4.82602835e-01 1.25752941e-01
-6.21904075e-01 -4.13529009e-01 4.72813427e-01 5.28207302e-01
7.91470826e-01 -6.60149217e-01 -2.26244777e-01 -2.17780903e-01
-1.21035399e-02 -1.13996816e+00 -8.36431265e-01 -8.05857003e-01
-1.13114440e+00 -1.34700954e+00 -7.55860329e-01 -7.92917311e-01
1.13297880e+00 5.36922336e-01 1.01389086e+00 3.54075789e-01
-3.81856322e-01 6.22579336e-01 -1.36740476e-01 -5.76658428e-01
-7.46398568e-02 -6.36531711e-01 4.16954517e-01 3.43814582e-01
3.98449838e-01 -8.04421246e-01 -4.27377760e-01 5.70059896e-01
-8.46935689e-01 -1.60795480e-01 5.26715875e-01 4.92834508e-01
9.59404707e-01 2.40012154e-01 5.22871375e-01 -6.41023993e-01
-2.24494375e-02 -3.11017692e-01 -7.69346535e-01 1.69875890e-01
-1.67244419e-01 -1.48500070e-01 3.96201849e-01 -6.61921918e-01
-1.16700351e+00 3.12255204e-01 -1.84977651e-02 -1.15455616e+00
-2.68260896e-01 2.46344328e-01 -9.95204031e-01 -2.28577211e-01
1.94963023e-01 2.78549433e-01 4.89225358e-01 -9.46879447e-01
1.06733128e-01 7.72756279e-01 4.01363879e-01 -6.77068591e-01
1.14200521e+00 6.75455570e-01 1.24715887e-01 -1.03187680e+00
-6.28332138e-01 1.00690588e-01 -8.10030997e-01 -3.74928564e-01
4.60302979e-01 -1.11396074e+00 -6.74199164e-01 7.04670548e-01
-1.01381683e+00 -7.78838471e-02 -4.65807430e-02 4.94087189e-01
-4.88350511e-01 1.50442749e-01 -3.93666923e-01 -9.46044266e-01
-9.69087109e-02 -1.07705247e+00 1.56311524e+00 3.93699110e-01
4.16970015e-01 -7.64280558e-01 -2.99348027e-01 5.47323167e-01
4.59517874e-02 3.55637431e-01 6.41835272e-01 3.63520756e-02
-8.12071443e-01 -3.18963796e-01 -1.25084445e-01 2.80372292e-01
5.29486418e-01 6.84973076e-02 -1.10535407e+00 -1.71270549e-01
3.79917115e-01 -5.52956581e-01 1.60841659e-01 1.25898600e-01
1.54417908e+00 -2.54412889e-01 -3.04589421e-01 8.34446490e-01
1.34366441e+00 3.62507030e-02 4.78790998e-01 -4.28733110e-01
8.77015710e-01 8.95335913e-01 5.85542619e-01 5.04943728e-01
4.83958155e-01 6.01873517e-01 3.80624533e-01 7.40318298e-02
-2.25959852e-01 -5.66430509e-01 3.26683670e-01 9.41739619e-01
-2.15617031e-01 2.01874226e-01 -6.58640444e-01 1.12022944e-01
-1.30393219e+00 -7.46429443e-01 1.71874434e-01 2.27492857e+00
9.73484755e-01 -5.20513475e-01 1.45532072e-01 6.01885915e-02
7.55954504e-01 -2.47708514e-01 -7.98545539e-01 6.61789477e-01
-3.08501180e-02 7.18711838e-02 -6.56804517e-02 2.01031521e-01
-5.43016195e-01 5.26565731e-01 6.01166248e+00 5.98265886e-01
-1.32366383e+00 1.13438621e-01 5.37391603e-01 -3.45143676e-01
-4.44611877e-01 -2.20681772e-01 -9.37813103e-01 5.79952061e-01
5.68878770e-01 5.64620979e-02 3.00800592e-01 9.90343750e-01
2.44354946e-03 2.02434242e-01 -1.22001290e+00 1.73147666e+00
3.04285496e-01 -9.43645000e-01 1.09791845e-01 4.62934673e-01
5.17867327e-01 -4.56578255e-01 2.47922577e-02 -1.28340637e-02
-2.00293005e-01 -1.19541407e+00 7.41502881e-01 8.33516777e-01
1.19328284e+00 -6.34946346e-01 1.57393575e-01 7.57341444e-01
-1.13020015e+00 2.47027963e-01 -5.39365470e-01 3.17500234e-01
5.45812547e-02 6.66867077e-01 -3.21982950e-01 4.53806728e-01
7.68902957e-01 7.72889316e-01 -4.42062289e-01 7.15957701e-01
-6.84060007e-02 2.58868337e-01 -4.95715886e-01 3.19454134e-01
-4.28232580e-01 -5.00304759e-01 2.02108458e-01 4.01520580e-01
8.83942604e-01 6.28773391e-01 2.15456933e-01 9.98989880e-01
-4.93407488e-01 -6.14969432e-02 -7.88582802e-01 3.42177525e-02
6.80812299e-01 1.22200584e+00 -3.67507160e-01 4.37061079e-02
-7.29321599e-01 6.91951036e-01 8.98190439e-02 1.90886527e-01
-7.41539419e-01 1.06670268e-01 6.58193946e-01 4.86785322e-01
2.36592859e-01 -3.56575608e-01 2.54834771e-01 -1.31594217e+00
2.54061073e-01 -8.10231268e-01 -1.27986923e-01 -1.04493988e+00
-1.50228882e+00 4.83732760e-01 5.59821054e-02 -1.27846110e+00
-7.92221799e-02 -8.74959648e-01 -3.38165939e-01 8.90272439e-01
-1.47663903e+00 -1.41737485e+00 -6.34826183e-01 1.00122571e+00
3.81743222e-01 -2.26762161e-01 7.75741398e-01 5.12502670e-01
-6.17599249e-01 6.26032114e-01 -1.25384361e-01 2.91085064e-01
6.72563195e-01 -5.90002298e-01 -3.50418352e-02 2.70505518e-01
2.92995483e-01 6.51731491e-01 4.84982245e-02 -5.51172674e-01
-2.16467834e+00 -1.01617599e+00 1.94304094e-01 -4.85820562e-01
-1.49689302e-01 -5.35250187e-01 -9.97287810e-01 7.96632707e-01
-3.65748078e-01 5.89899123e-01 9.09662485e-01 -1.46193221e-01
-4.97610867e-01 -3.70108902e-01 -1.54367948e+00 1.90946832e-01
1.42784607e+00 -6.77268028e-01 -3.22746933e-01 1.99256688e-01
3.38954777e-01 -4.96690780e-01 -1.09454775e+00 3.75619024e-01
6.62512064e-01 -8.52765679e-01 9.40337896e-01 -2.83892363e-01
2.33752459e-01 -3.69121730e-01 -3.46256971e-01 -1.11105168e+00
3.46878946e-01 -4.37041759e-01 -3.11364055e-01 1.55039918e+00
-2.06738606e-01 -4.78772253e-01 9.86153066e-01 8.58794510e-01
2.50098497e-01 -7.57918775e-01 -7.21197069e-01 -8.58138502e-01
-3.86933610e-02 -4.05543506e-01 1.24984193e+00 8.72050464e-01
-6.16543710e-01 1.54516608e-01 -5.01286983e-01 1.70708254e-01
1.05609131e+00 4.41370308e-01 1.09269929e+00 -1.65711462e+00
6.82706460e-02 5.97126186e-02 -3.20505202e-01 -1.07968318e+00
5.08586109e-01 -6.45825505e-01 -2.40131766e-01 -1.08448660e+00
2.49161065e-01 -7.04280138e-01 4.41750169e-01 5.21209776e-01
1.40716255e-01 3.77208352e-01 3.54421437e-02 3.43030870e-01
-3.50283161e-02 8.42960119e-01 1.65011871e+00 1.01321042e-01
1.31535500e-01 -7.45702982e-02 -7.88626790e-01 9.26887393e-01
4.77403015e-01 -3.06854546e-01 -6.77858889e-01 -8.14540625e-01
-2.04170063e-01 3.21838111e-01 4.08245981e-01 -5.27332187e-01
1.62346408e-01 -2.80510187e-01 8.80342364e-01 -6.14329398e-01
7.71908462e-01 -1.14751291e+00 4.03373688e-01 -1.43413991e-01
2.94482142e-01 -3.02140415e-01 1.27205819e-01 5.16370595e-01
-1.16220281e-01 -1.30083300e-02 7.72893369e-01 -3.77621740e-01
-3.46150070e-01 1.10980856e+00 4.28617686e-01 -1.17155813e-01
9.58169341e-01 -3.94666880e-01 5.81634343e-02 -3.81728560e-01
-4.93886799e-01 3.61031257e-02 7.30903327e-01 5.27143300e-01
9.32795286e-01 -1.80803311e+00 -6.99342906e-01 8.12371790e-01
-7.91782364e-02 5.56341112e-01 2.73411274e-01 4.02106792e-01
-1.17903382e-01 1.48531571e-01 -3.51003140e-01 -6.99081004e-01
-1.32632184e+00 4.92168933e-01 3.64924192e-01 6.33911550e-01
-4.77270871e-01 6.94269836e-01 1.22987434e-01 -7.54864872e-01
9.86046866e-02 2.80211791e-02 1.99106857e-01 4.54773381e-02
8.67627203e-01 1.82999924e-01 -1.59199491e-01 -1.08714223e+00
-2.60425597e-01 1.21566808e+00 3.42693120e-01 -1.64360087e-03
1.44531929e+00 -1.89237118e-01 -2.33394280e-01 3.54029149e-01
1.46603119e+00 7.66361356e-02 -1.65876734e+00 -6.20363295e-01
-4.93776500e-01 -1.18499613e+00 1.08327076e-01 -1.90297976e-01
-1.83649397e+00 1.07796335e+00 4.68078941e-01 -2.80616105e-01
1.36480033e+00 2.05529019e-01 8.37662280e-01 2.58499503e-01
9.75957334e-01 -8.80676806e-01 3.93784106e-01 -2.78998893e-02
1.13187385e+00 -1.46545231e+00 -1.18811904e-02 -6.98685110e-01
-2.94484109e-01 1.09640324e+00 8.38077903e-01 1.50422603e-01
1.02816033e+00 3.36012244e-01 6.82595745e-02 -3.52160245e-01
-3.58875334e-01 5.39717153e-02 7.89052397e-02 8.35828722e-01
1.36387229e-01 -1.86555728e-01 4.69694585e-01 8.98998260e-01
-1.12493686e-01 2.08845735e-01 2.28994161e-01 9.42463696e-01
-1.29026666e-01 -1.28920519e+00 -8.05803001e-01 2.03701288e-01
-1.99655354e-01 4.56772119e-01 -2.21209928e-01 9.49281156e-01
1.44652396e-01 7.66062081e-01 3.01663987e-02 -3.10203165e-01
4.32435900e-01 2.34058440e-01 1.06794155e+00 -5.73603868e-01
3.07725728e-01 2.29783073e-01 -5.92885911e-01 -5.99322021e-01
-5.99108994e-01 -7.10154891e-01 -1.11564422e+00 -4.14781064e-01
-3.87497663e-01 -1.81037977e-01 7.08394229e-01 7.37921476e-01
4.76922899e-01 -1.06031083e-01 9.52828646e-01 -1.18622053e+00
-4.04685616e-01 -9.07056391e-01 -1.23880470e+00 2.22914040e-01
3.86153460e-01 -9.73990202e-01 -1.42180771e-01 4.01087523e-01] | [13.144681930541992, 0.06212303787469864] |
46473f47-25ba-42a0-81c5-a97f6589bbdb | sls-at-semeval-2016-task-3-neural-based | null | null | https://aclanthology.org/S16-1128 | https://aclanthology.org/S16-1128.pdf | SLS at SemEval-2016 Task 3: Neural-based Approaches for Ranking in Community Question Answering | null | ['Wei-Ning Hsu', 'Tao Lei', 'Scott Cyphers', 'Mitra Mohtarami', 'Jim Glass', 'Kfir Bar', 'Yu Zhang', 'Yonatan Belinkov'] | 2016-06-01 | null | null | null | semeval-2016-6 | ['question-similarity'] | ['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] | [-7.37662410736084, 3.6809213161468506] |
3e585b09-8750-46a0-9e9a-56852fa959b2 | critical-overview-of-privacy-preserving | 2004.09612 | null | https://arxiv.org/abs/2004.09612v6 | https://arxiv.org/pdf/2004.09612v6.pdf | A Critical Overview of Privacy-Preserving Approaches for Collaborative Forecasting | Cooperation between different data owners may lead to an improvement in forecast quality - for instance by benefiting from spatial-temporal dependencies in geographically distributed time series. Due to business competitive factors and personal data protection questions, said data owners might be unwilling to share their data, which increases the interest in collaborative privacy-preserving forecasting. This paper analyses the state-of-the-art and unveils several shortcomings of existing methods in guaranteeing data privacy when employing Vector Autoregressive (VAR) models. The paper also provides mathematical proofs and numerical analysis to evaluate existing privacy-preserving methods, dividing them into three groups: data transformation, secure multi-party computations, and decomposition methods. The analysis shows that state-of-the-art techniques have limitations in preserving data privacy, such as a trade-off between privacy and forecasting accuracy, while the original data in iterative model fitting processes, in which intermediate results are shared, can be inferred after some iterations. | ['Carla Gonçalves', 'Pierre Pinson', 'Ricardo J. Bessa'] | 2020-04-20 | null | null | null | null | ['mathematical-proofs'] | ['miscellaneous'] | [ 4.05261032e-02 1.02706827e-01 4.97573875e-02 -7.12084353e-01
-7.44371116e-01 -1.20394313e+00 5.24891734e-01 3.38907242e-01
-2.46031478e-01 7.65432179e-01 2.95283556e-01 -6.28914654e-01
-2.83874542e-01 -8.83433402e-01 -5.59600174e-01 -1.07234967e+00
-2.44263873e-01 2.46279225e-01 -3.64808947e-01 -7.17973039e-02
1.72579467e-01 6.97450638e-01 -1.11472189e+00 1.84161499e-01
7.13586807e-01 1.17484677e+00 -7.31152236e-01 3.22224766e-01
9.29287672e-02 6.40060008e-01 -4.36855704e-01 -1.05587924e+00
8.31009865e-01 -3.71971913e-02 -5.26999354e-01 -4.09651667e-01
-2.41924241e-01 -5.82586050e-01 -1.58586457e-01 1.17633533e+00
2.31753990e-01 2.80637201e-02 2.35597730e-01 -1.92855644e+00
-8.16236496e-01 5.72212160e-01 -3.67365032e-01 -4.84072357e-01
-1.25376135e-02 -2.46604487e-01 7.38663673e-01 -3.88123721e-01
3.78214926e-01 7.39999533e-01 9.28491890e-01 3.60619307e-01
-1.38952887e+00 -9.03279126e-01 -4.99521755e-02 -1.96919486e-01
-1.40755856e+00 -4.66373533e-01 8.32149982e-01 -5.03092170e-01
5.01391590e-01 1.17880738e+00 3.97434741e-01 6.71955287e-01
4.11164612e-01 6.49142683e-01 1.49490166e+00 6.94746599e-02
4.29878026e-01 5.66837490e-01 1.47802770e-01 -1.06047451e-01
6.26263320e-01 5.18741131e-01 -4.33025628e-01 -1.22200537e+00
4.59326267e-01 3.06608617e-01 -3.65704417e-01 -6.11104071e-01
-6.56752169e-01 9.20191586e-01 5.26164956e-02 7.72156790e-02
-5.36055982e-01 -1.34272829e-01 5.61879337e-01 7.77731955e-01
9.68394279e-01 1.43907040e-01 -7.99470901e-01 2.50911266e-01
-9.46739137e-01 4.57085043e-01 9.24201667e-01 9.12523091e-01
4.36265320e-01 -5.09869568e-02 8.97999946e-03 -3.06193139e-02
4.79954004e-01 5.80170572e-01 4.56112362e-02 -8.23018253e-01
4.66421843e-01 9.02918950e-02 4.94621396e-01 -1.22643960e+00
2.83837639e-04 -1.11019492e-01 -1.25148690e+00 2.21613362e-01
4.67285395e-01 -4.96426672e-01 -5.72889159e-03 1.76099765e+00
4.58394855e-01 -2.83230931e-01 4.02555734e-01 7.85440743e-01
1.13172792e-01 6.29926741e-01 -2.37069521e-02 -5.91312945e-01
1.39157236e+00 -4.63307470e-01 -1.00040627e+00 3.56631994e-01
5.71982205e-01 -5.25105417e-01 3.90639693e-01 2.82920837e-01
-8.90017986e-01 5.91960885e-02 -4.81492370e-01 -4.50431369e-02
-5.95103443e-01 -5.01689613e-01 9.82375920e-01 1.18390572e+00
-8.36813867e-01 3.75315785e-01 -7.96684921e-01 2.44033579e-02
5.92138112e-01 4.22053099e-01 -6.39105380e-01 2.80504435e-01
-1.42493618e+00 3.52535516e-01 -1.48112521e-01 4.99424785e-02
-5.82449019e-01 -1.16622186e+00 -4.91485864e-01 2.68260330e-01
1.46201849e-01 -7.84844100e-01 8.62132847e-01 -8.18031132e-01
-1.21393502e+00 8.02372158e-01 -8.60444307e-02 -8.17035019e-01
1.15502834e+00 1.31072566e-01 -6.64982915e-01 -2.50431567e-01
-2.61484444e-01 -1.02773584e-01 7.98875511e-01 -1.23438513e+00
-7.04476655e-01 -9.33115184e-01 -2.41716698e-01 -2.76590616e-01
-3.45859885e-01 3.84716392e-01 3.23838085e-01 -8.86464894e-01
-9.90287364e-02 -9.29072678e-01 -5.49480379e-01 2.39723191e-01
-2.94923812e-01 1.85844168e-01 8.68252277e-01 -1.00444889e+00
1.12322938e+00 -2.28400707e+00 -4.67914253e-01 6.31953537e-01
1.57187730e-01 1.51863173e-02 4.23704714e-01 6.92487419e-01
1.23797953e-02 4.49021578e-01 -3.30985904e-01 -5.84222794e-01
3.37388039e-01 1.72929779e-01 -1.07475471e+00 1.06065333e+00
-4.70820099e-01 7.02929318e-01 -4.30880159e-01 6.23251945e-02
7.51806498e-02 5.36103845e-01 -1.66933969e-01 6.90534338e-02
1.12418257e-01 7.00743675e-01 -5.36750853e-01 4.74395990e-01
1.49570370e+00 5.03232181e-02 2.90131181e-01 1.52638003e-01
-2.57912666e-01 -1.23874724e-01 -1.06299794e+00 1.17476976e+00
-4.69696075e-02 3.47664982e-01 6.37616634e-01 -7.59821355e-01
9.56379890e-01 6.29876077e-01 6.27881944e-01 -1.23060048e-01
5.87495603e-02 -1.06396468e-03 -5.62228739e-01 -8.83051753e-03
7.10491002e-01 -1.32957315e-02 -2.84103483e-01 6.85140312e-01
-7.69371510e-01 1.16928130e-01 -7.43306220e-01 -1.72015980e-01
6.52066886e-01 -3.77561480e-01 1.39148682e-01 -5.19967616e-01
3.26447278e-01 -9.52009335e-02 6.70698404e-01 5.62528014e-01
-2.01216459e-01 1.87447622e-01 5.97554445e-01 -7.02542782e-01
-9.51790869e-01 -6.49004936e-01 -2.41174743e-01 7.77129889e-01
-3.00341368e-01 -2.58873045e-01 -6.30601883e-01 -5.28713226e-01
5.11557877e-01 1.02552152e+00 -7.31073856e-01 8.52081999e-02
9.99618545e-02 -6.95990562e-01 6.80461109e-01 2.36343428e-01
4.14502740e-01 -2.24079862e-01 -6.96233213e-01 4.52139080e-02
3.61099280e-02 -6.26430452e-01 -4.75162387e-01 -2.03906581e-01
-9.21532929e-01 -7.14323342e-01 -5.31689465e-01 2.09463164e-01
9.62580025e-01 5.09072006e-01 5.41381061e-01 -2.80595690e-01
3.35588306e-01 2.76594996e-01 -3.91390873e-03 -9.26379025e-01
-3.66327733e-01 -2.44276717e-01 6.00332431e-02 6.27747715e-01
4.44156379e-01 -6.96163356e-01 -4.82031196e-01 3.90462041e-01
-1.10849667e+00 -3.38225514e-01 -1.18964754e-01 4.61973101e-01
6.18327916e-01 3.35269213e-01 2.59386510e-01 -1.17977130e+00
7.57592618e-01 -5.51135778e-01 -1.18115342e+00 5.96067131e-01
-1.18010890e+00 -1.57217398e-01 8.26131999e-01 -1.27583414e-01
-1.14236450e+00 1.69638678e-01 3.28005403e-01 -4.80547905e-01
9.06891748e-03 5.04000902e-01 -3.70699316e-01 -3.51711869e-01
5.50584234e-02 3.90612274e-01 4.17448431e-01 -5.71281672e-01
5.17878830e-01 7.50560284e-01 2.60472119e-01 -3.37693244e-01
5.54538786e-01 9.07350719e-01 2.88070291e-01 -2.56745309e-01
-1.40620351e-01 -7.80731663e-02 -1.99340001e-01 2.11524829e-01
5.17987847e-01 -9.64077950e-01 -1.18209267e+00 7.59183824e-01
-1.07355130e+00 3.28204244e-01 -5.80677032e-01 4.86647040e-01
-4.23708767e-01 4.53774095e-01 -4.79386717e-01 -1.29911959e+00
-6.05338871e-01 -7.34437466e-01 5.28679669e-01 -1.99108005e-01
-4.86954898e-02 -1.02959192e+00 4.26735096e-02 4.80014801e-01
7.96509743e-01 6.25606000e-01 6.35598660e-01 -1.14452207e+00
-5.21255195e-01 -8.28810871e-01 1.31830826e-01 2.47899577e-01
-1.45383701e-01 -5.08043766e-02 -1.14891911e+00 -5.32462001e-01
5.91114819e-01 3.90534401e-01 1.67470530e-01 3.00358981e-01
1.34316862e+00 -1.23241937e+00 -1.74577549e-01 8.15749168e-01
1.37487996e+00 9.88985077e-02 7.51715362e-01 -6.72853813e-02
2.67388910e-01 1.03420222e+00 5.37192345e-01 1.02839148e+00
4.45744187e-01 3.31708938e-01 3.96808863e-01 -1.64011121e-01
1.08733451e+00 -4.50805694e-01 1.15615964e-01 3.67346734e-01
-1.18972264e-01 -2.42146775e-01 -5.40111125e-01 4.60113287e-01
-2.12453580e+00 -9.80524063e-01 -3.57169986e-01 2.58786297e+00
6.00470483e-01 -5.88183641e-01 9.72892269e-02 -1.88279569e-01
4.70313221e-01 2.89681911e-01 -6.01857722e-01 -6.05600297e-01
-3.18609625e-01 -2.71054178e-01 1.38736260e+00 3.67912263e-01
-9.65651691e-01 4.07125950e-01 6.52959299e+00 7.50972092e-01
-9.16587353e-01 3.51681083e-01 1.05274332e+00 8.46301988e-02
-1.05292881e+00 4.11700159e-01 -3.80083472e-01 5.59072614e-01
1.03952074e+00 -9.12389100e-01 5.43945372e-01 9.77812827e-01
2.53646314e-01 3.97950381e-01 -8.52758944e-01 7.06975758e-01
-2.59312958e-01 -1.38902581e+00 -3.15466493e-01 6.65836036e-01
7.47386038e-01 -3.40130419e-01 1.80502430e-01 -3.34837526e-01
7.10823119e-01 -8.79581928e-01 5.72607696e-01 8.37129354e-01
3.83307695e-01 -1.10588205e+00 7.96698153e-01 3.47824097e-01
-8.83677721e-01 -5.62334061e-02 -5.66408336e-01 -5.38330898e-03
2.55793691e-01 7.43550062e-01 -1.07865505e-01 1.16224539e+00
7.51394451e-01 1.49447098e-01 3.94393206e-02 5.05811870e-01
2.62490213e-02 5.46664417e-01 -5.24410725e-01 2.18115449e-01
-1.22428000e-01 -7.96067059e-01 5.85153282e-01 6.04839802e-01
5.89216232e-01 4.14192677e-01 -2.91964024e-01 8.92231643e-01
-5.41461222e-02 1.43635958e-01 -8.72385919e-01 -5.74276857e-02
6.83427274e-01 9.85354125e-01 -1.17889546e-01 -1.13510251e-01
-4.82048512e-01 8.20000172e-01 -3.40405971e-01 2.70550817e-01
-5.87422073e-01 -1.84653714e-01 1.01574409e+00 1.61901161e-01
2.64624566e-01 -2.75685698e-01 -7.66946256e-01 -1.09808111e+00
2.59776264e-01 -8.53783190e-01 6.40202343e-01 -4.37184960e-01
-1.54510212e+00 2.75488287e-01 -1.51195198e-01 -1.27416503e+00
-1.71454147e-01 2.99777165e-02 -2.88775563e-01 1.13811779e+00
-1.19917345e+00 -1.49425209e+00 3.63734156e-01 8.56887102e-01
-4.16909724e-01 -3.38647217e-01 1.17570782e+00 1.14780329e-01
-1.71872869e-01 9.57121313e-01 8.75232995e-01 -8.14217627e-02
4.95190382e-01 -6.59017801e-01 4.28587675e-01 1.10049295e+00
-1.03658237e-01 8.56833875e-01 8.36336792e-01 -5.47760427e-01
-1.69245863e+00 -1.10882699e+00 1.47443688e+00 -5.38399160e-01
6.58338308e-01 -5.48589230e-01 -9.51725006e-01 9.28563297e-01
1.85428739e-01 2.07480907e-01 1.18069112e+00 -1.76176392e-02
-5.29839575e-01 -4.78047609e-01 -1.97629797e+00 3.44087839e-01
5.00699341e-01 -6.47213221e-01 -7.23110288e-02 4.02405262e-01
8.12740028e-01 -9.24498215e-02 -1.22038186e+00 -5.76191656e-02
6.72045887e-01 -8.36024106e-01 8.44971359e-01 -1.04378748e+00
-2.02911302e-01 -2.15202674e-01 -4.37755734e-01 -9.56892490e-01
-1.19407140e-01 -1.38383067e+00 6.72315666e-03 1.65301573e+00
3.05972964e-01 -1.50618553e+00 6.62163317e-01 1.96778679e+00
6.61548257e-01 -1.00801148e-01 -1.26246309e+00 -8.27976525e-01
2.59611487e-01 -4.71621156e-01 1.62875307e+00 1.62533092e+00
-9.05849710e-02 -5.90535879e-01 -9.33240116e-01 6.30738318e-01
9.08142567e-01 4.79311317e-01 1.02930796e+00 -1.20462632e+00
9.19696465e-02 8.79827067e-02 -1.72495544e-01 -6.68887615e-01
1.78932026e-01 -5.59963644e-01 -7.04722106e-01 -6.92810833e-01
-1.57957315e-01 -6.36078596e-01 -3.76649052e-01 5.85651636e-01
3.90939385e-01 -4.27661091e-01 3.57408762e-01 5.07738829e-01
8.30561295e-03 4.33280468e-01 6.71168983e-01 9.09711942e-02
-4.67450805e-02 4.95076030e-01 -1.16778362e+00 1.43550768e-01
7.74201572e-01 -7.99098313e-01 -4.10439014e-01 -2.19300494e-01
3.16889822e-01 5.81982374e-01 5.78752041e-01 -2.88873464e-01
5.48624933e-01 -4.57895905e-01 -8.94482359e-02 -6.53950989e-01
2.58875918e-02 -1.75378764e+00 1.26404035e+00 4.45909828e-01
-4.52461332e-01 1.54279545e-01 -7.45661259e-02 7.93197215e-01
-2.18062565e-01 1.12369500e-01 3.44037861e-01 1.83063552e-01
-6.52243793e-02 4.87399936e-01 -3.40055734e-01 -4.75100368e-01
1.30242240e+00 -9.86112356e-02 -3.96908134e-01 -7.61015832e-01
-6.10227823e-01 2.89203316e-01 6.52189314e-01 1.70175672e-01
1.29225254e-01 -1.31353176e+00 -9.07867193e-01 5.00648677e-01
-1.87360212e-01 -2.87643313e-01 5.46867073e-01 8.38503540e-01
-1.59518570e-01 5.60220003e-01 -9.63805243e-02 -6.94154650e-02
-1.43391740e+00 1.07348073e+00 -3.07525452e-02 -3.93386036e-01
-3.30972701e-01 5.73814929e-01 4.22613807e-02 -5.70686162e-01
5.76175936e-02 -2.29439154e-01 4.79030758e-01 1.52846739e-01
5.66810489e-01 6.47051811e-01 4.70064469e-02 -4.54605430e-01
-4.35141891e-01 4.44890819e-02 1.12158604e-01 -1.91330001e-01
1.34191084e+00 -5.28385699e-01 -4.86118942e-01 1.56029267e-02
1.23063660e+00 3.27929854e-01 -1.12007487e+00 -3.09147358e-01
-2.13600710e-01 -1.08341181e+00 7.28321224e-02 -5.48764169e-01
-1.36617982e+00 5.63304484e-01 5.80716550e-01 7.39859879e-01
1.24546170e+00 -4.33892578e-01 8.26515913e-01 -3.72131206e-02
7.49668777e-01 -1.01359975e+00 -1.42909968e+00 -7.92640075e-02
8.98508728e-01 -1.05991614e+00 1.50825948e-01 -4.20149416e-01
-9.94181633e-01 7.19337642e-01 -4.68248039e-01 2.03749239e-01
1.18560541e+00 3.64659816e-01 1.22471310e-01 1.30264267e-01
-7.27976263e-01 6.21124089e-01 -4.07983400e-02 7.57186413e-01
-9.78389606e-02 6.62247598e-01 -5.83115339e-01 1.32460821e+00
-3.01399231e-01 1.45192876e-01 3.49517673e-01 9.88558948e-01
3.51630718e-01 -1.33781564e+00 -6.01092041e-01 3.32350284e-01
-8.59297454e-01 5.29453903e-03 -4.57296133e-01 5.79798877e-01
-2.19274044e-01 1.05736887e+00 9.66618061e-02 -2.25641027e-01
2.41106570e-01 2.94746924e-02 -3.37175965e-01 5.35241604e-01
-1.01421845e+00 -8.03503916e-02 8.28886628e-02 -4.57338452e-01
-4.15061623e-01 -1.16738641e+00 -7.42819071e-01 -1.18670905e+00
-2.69184470e-01 4.51072842e-01 8.60490501e-01 3.64307404e-01
1.03586352e+00 -2.66720325e-01 1.22806144e+00 -5.57359308e-02
-1.05273068e+00 -2.30496928e-01 -1.04504740e+00 3.36307049e-01
3.38847846e-01 2.17472613e-01 -5.04617751e-01 7.19246641e-02] | [5.950911998748779, 6.619917869567871] |
7a7a4a25-7652-4c5c-991c-2ad399f1fba4 | unsupervised-representation-learning-by | 1708.01246 | null | http://arxiv.org/abs/1708.01246v1 | http://arxiv.org/pdf/1708.01246v1.pdf | Unsupervised Representation Learning by Sorting Sequences | We present an unsupervised representation learning approach using videos
without semantic labels. We leverage the temporal coherence as a supervisory
signal by formulating representation learning as a sequence sorting task. We
take temporally shuffled frames (i.e., in non-chronological order) as inputs
and train a convolutional neural network to sort the shuffled sequences.
Similar to comparison-based sorting algorithms, we propose to extract features
from all frame pairs and aggregate them to predict the correct order. As
sorting shuffled image sequence requires an understanding of the statistical
temporal structure of images, training with such a proxy task allows us to
learn rich and generalizable visual representation. We validate the
effectiveness of the learned representation using our method as pre-training on
high-level recognition problems. The experimental results show that our method
compares favorably against state-of-the-art methods on action recognition,
image classification and object detection tasks. | ['Ming-Hsuan Yang', 'Jia-Bin Huang', 'Hsin-Ying Lee', 'Maneesh Singh'] | 2017-08-03 | unsupervised-representation-learning-by-3 | http://openaccess.thecvf.com/content_iccv_2017/html/Lee_Unsupervised_Representation_Learning_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Lee_Unsupervised_Representation_Learning_ICCV_2017_paper.pdf | iccv-2017-10 | ['self-supervised-action-recognition'] | ['computer-vision'] | [ 8.17201197e-01 -2.51370788e-01 -6.23863816e-01 -6.65368378e-01
-3.39085042e-01 -6.03813291e-01 8.72833490e-01 -1.45483598e-01
-5.44160724e-01 5.03640711e-01 5.88789880e-01 -1.37783699e-02
-8.48071426e-02 -4.99952793e-01 -1.01725948e+00 -6.22331202e-01
-4.25186664e-01 1.99261487e-01 3.21691394e-01 8.17779526e-02
5.12954712e-01 2.70815551e-01 -1.93199372e+00 1.19270360e+00
3.60724151e-01 1.27654719e+00 2.69722998e-01 7.85419583e-01
1.40929252e-01 1.64230144e+00 -3.98532957e-01 1.89583585e-01
3.06712657e-01 -7.21285164e-01 -1.21120512e+00 5.28336644e-01
7.06043303e-01 -5.82761765e-01 -7.37332046e-01 7.71680713e-01
-2.08005011e-01 5.29059470e-01 7.10003614e-01 -1.20007932e+00
-6.80990756e-01 4.90651131e-01 -2.83902258e-01 5.60932457e-01
7.08410919e-01 3.50148052e-01 1.11291289e+00 -6.82722330e-01
1.02409792e+00 1.26274741e+00 4.92039233e-01 5.95392942e-01
-1.27000058e+00 -4.94160622e-01 3.70912701e-01 7.51255989e-01
-6.26496017e-01 -5.96129894e-01 6.66136086e-01 -8.60727727e-01
9.54468429e-01 2.39112064e-01 7.56074250e-01 1.50210130e+00
-4.75903302e-02 1.17472661e+00 1.12752509e+00 -3.03872138e-01
2.70318776e-01 -5.42215884e-01 3.37762773e-01 6.16400540e-01
-8.52043256e-02 3.94430935e-01 -9.04445589e-01 3.85731429e-01
7.56653488e-01 3.73198926e-01 -3.35977525e-01 -5.51262319e-01
-1.55049348e+00 4.87231195e-01 6.15623415e-01 2.58394092e-01
-2.94634491e-01 5.60805321e-01 8.21803331e-01 5.09757161e-01
1.45926878e-01 3.48986089e-01 -3.42204869e-01 -3.37465614e-01
-1.07255828e+00 -1.13502415e-02 3.85069132e-01 6.86315179e-01
7.05004334e-01 -4.60591987e-02 -4.97273415e-01 5.64897716e-01
-1.71306029e-01 -9.50257629e-02 8.32574546e-01 -1.41714203e+00
3.50280881e-01 4.73977566e-01 9.11427885e-02 -9.38894331e-01
-2.06963256e-01 1.69342101e-01 -7.19856024e-01 2.47711480e-01
5.95633090e-01 3.82737458e-01 -1.06609690e+00 1.76026666e+00
-2.80726224e-01 6.44727707e-01 -2.31762715e-02 1.10935307e+00
5.59469044e-01 5.72240829e-01 7.21920803e-02 -2.89745539e-01
1.19450903e+00 -8.97599161e-01 -4.65566367e-01 -5.99529967e-02
5.05507410e-01 -2.54829198e-01 1.09152186e+00 2.69064903e-01
-7.70059466e-01 -9.10473168e-01 -9.47345972e-01 3.06094922e-02
-2.43698210e-01 1.29222810e-01 7.30154037e-01 3.98279764e-02
-9.53049183e-01 1.11285579e+00 -1.22014940e+00 -6.61660492e-01
7.18819082e-01 9.93127525e-02 -6.08643949e-01 -1.70855725e-04
-8.93619061e-01 5.90458035e-01 7.00551331e-01 -1.42975017e-01
-1.32369459e+00 -5.37734449e-01 -9.51788962e-01 -1.22769177e-01
3.00431699e-01 -5.29633462e-01 1.26869762e+00 -1.56420290e+00
-1.51263702e+00 1.04705286e+00 -3.11185181e-01 -9.27134991e-01
3.93645734e-01 -2.97631890e-01 -9.04931501e-02 6.67714417e-01
1.64592296e-01 8.68944943e-01 1.03229344e+00 -1.02801967e+00
-7.97538757e-01 -7.23598972e-02 1.57303751e-01 2.70799436e-02
-1.98733494e-01 1.80765172e-03 -2.32004412e-02 -7.04548955e-01
4.32516411e-02 -9.78756607e-01 -1.04356401e-01 3.35079879e-02
-1.53528571e-01 -8.33842754e-02 9.85207379e-01 -5.01059711e-01
9.93310690e-01 -2.17781210e+00 1.64701059e-01 -1.03077531e-01
-5.31629771e-02 2.92049237e-02 -2.41195887e-01 3.66491199e-01
-4.63933289e-01 -2.08449334e-01 -3.65925938e-01 -9.23112631e-02
-8.62980932e-02 4.64167416e-01 -6.09697461e-01 4.70325023e-01
4.50825125e-01 9.80568290e-01 -1.41996598e+00 -2.40101948e-01
3.76229018e-01 4.13622819e-02 -5.26201129e-01 4.41900700e-01
-3.41312021e-01 8.86566997e-01 -1.92340165e-01 4.20358092e-01
6.93525607e-03 -5.09011924e-01 3.73558879e-01 -3.93640846e-01
6.86227232e-02 7.16305554e-01 -7.72659719e-01 1.88465023e+00
-1.98577300e-01 1.23086333e+00 -7.07707584e-01 -1.73265743e+00
6.44841015e-01 1.33979529e-01 8.20146441e-01 -8.08294594e-01
-5.06667653e-04 -1.61273301e-01 -3.00372932e-02 -7.71824121e-01
4.71366882e-01 1.23552576e-01 -1.20822899e-01 6.53029561e-01
3.58148843e-01 3.89373749e-01 4.28157300e-01 2.58159637e-01
1.38019431e+00 6.95493579e-01 2.48992160e-01 -8.48533586e-02
3.50730777e-01 3.20894271e-02 5.86534679e-01 8.28647971e-01
-3.06626946e-01 6.47113681e-01 5.64874530e-01 -9.25649047e-01
-9.47489083e-01 -9.81512547e-01 2.06108302e-01 1.42063916e+00
1.55102387e-01 -4.75634217e-01 -4.40376699e-01 -8.91942382e-01
-2.99504902e-02 3.95224780e-01 -9.86120224e-01 -2.50106007e-01
-8.91600549e-01 -1.79195657e-01 3.36603135e-01 9.38345075e-01
5.81907928e-01 -1.39155912e+00 -1.24474156e+00 2.98762117e-02
-2.55790174e-01 -1.25427222e+00 -4.24560577e-01 3.05284947e-01
-9.70448673e-01 -1.46002066e+00 -3.74283075e-01 -8.47849250e-01
6.48567319e-01 3.38772118e-01 1.27400827e+00 -1.21349506e-01
-3.62248272e-01 7.42886662e-01 -5.72928607e-01 2.69114852e-01
-1.50336549e-01 -4.44853306e-01 4.67344075e-02 3.80122006e-01
3.26175809e-01 -7.07401633e-01 -9.00267899e-01 3.10279936e-01
-8.82322431e-01 1.99485868e-01 3.58409077e-01 9.69779551e-01
5.99712074e-01 -3.07195097e-01 2.11341694e-01 -7.71332800e-01
2.28726819e-01 -2.48481542e-01 -5.20182312e-01 2.56203532e-01
-6.73075989e-02 2.20923766e-01 4.74342853e-01 -6.24846876e-01
-8.48663509e-01 4.81038153e-01 5.95963717e-01 -7.59389699e-01
-3.28342676e-01 2.97206551e-01 1.99044451e-01 3.04930896e-01
5.25739253e-01 5.26324868e-01 -6.47099167e-02 -2.06060410e-01
6.23397648e-01 2.62411654e-01 9.66605663e-01 -5.91235816e-01
3.96392316e-01 7.86424637e-01 -7.75017813e-02 -6.06593728e-01
-8.43015373e-01 -5.93402803e-01 -1.07477355e+00 -4.08290118e-01
1.03501344e+00 -8.42303455e-01 -8.97924721e-01 3.00717145e-01
-1.16809487e+00 -7.37434804e-01 -5.12963235e-01 5.55971801e-01
-1.17759407e+00 4.36116487e-01 -6.80518925e-01 -6.63147628e-01
1.86576024e-01 -9.68218207e-01 1.13637221e+00 -1.63367882e-01
-5.73758423e-01 -6.90019190e-01 1.12007812e-01 1.81760132e-01
-1.63548172e-01 4.88903999e-01 5.72476983e-01 -4.85409319e-01
-9.17347491e-01 2.88539201e-01 -1.95122689e-01 3.79337490e-01
2.42693782e-01 -7.41003230e-02 -8.54256034e-01 -1.53810769e-01
-1.70165226e-02 -6.59031391e-01 1.40843487e+00 3.75937581e-01
1.67760229e+00 -4.38089430e-01 -3.20886016e-01 6.20000303e-01
1.05939555e+00 2.69836545e-01 7.97791719e-01 5.72357237e-01
5.89362085e-01 7.12396204e-01 7.10079491e-01 4.95003700e-01
1.49598703e-01 6.07707202e-01 3.12776119e-01 3.93022060e-01
-3.17903697e-01 -2.89818257e-01 7.04006553e-01 5.58861017e-01
-2.55014032e-01 4.44321893e-02 -7.82583714e-01 5.31015933e-01
-2.30596948e+00 -1.69597101e+00 3.73268455e-01 2.05075336e+00
7.94128180e-01 1.01399934e-02 2.57914633e-01 1.88791797e-01
6.50982380e-01 5.26939631e-01 -4.25356865e-01 -1.69460386e-01
7.81912170e-03 2.55497217e-01 4.83402342e-01 1.70722008e-01
-1.50125217e+00 9.20512915e-01 6.57026482e+00 2.67784923e-01
-1.25848806e+00 -1.26916662e-01 6.04582727e-01 -2.09786873e-02
2.03183606e-01 9.75441337e-02 -1.93699330e-01 6.71155810e-01
8.23115945e-01 2.66440827e-02 5.22557676e-01 5.56349099e-01
2.29036242e-01 -1.17722981e-01 -1.86605430e+00 1.21865904e+00
1.41200408e-01 -1.67078006e+00 2.17105001e-01 -2.43722513e-01
7.63109803e-01 -4.58842255e-02 3.05317007e-02 1.21339351e-01
5.34947813e-01 -1.20151043e+00 9.40006495e-01 7.89041400e-01
5.47922313e-01 -1.60913616e-01 2.80122250e-01 -1.89478807e-02
-1.40833330e+00 -3.21549773e-01 -1.72880426e-01 -5.27217507e-01
8.86373296e-02 1.59626618e-01 -5.52605808e-01 4.91841197e-01
8.92514586e-01 1.70043814e+00 -5.39412320e-01 9.38804686e-01
-3.49598765e-01 6.61110282e-01 -5.07548079e-02 3.36120695e-01
3.22981119e-01 -2.33557269e-01 3.21547657e-01 1.37783158e+00
1.40547872e-01 1.79627433e-01 4.47796673e-01 4.00781393e-01
-1.23467728e-01 -5.33850253e-01 -8.32771361e-01 -2.10866570e-01
1.55211285e-01 8.26397598e-01 -9.13289070e-01 -7.92001069e-01
-4.01850551e-01 1.30909717e+00 3.47487539e-01 4.85964686e-01
-7.31968641e-01 -3.57894935e-02 7.35552192e-01 -1.38751790e-01
6.83371782e-01 -3.44474584e-01 -9.28075612e-02 -1.40044880e+00
1.12351375e-02 -8.18570971e-01 7.20804155e-01 -8.01929832e-01
-1.23593903e+00 3.95294636e-01 8.57046694e-02 -1.78570974e+00
-5.49905717e-01 -7.70725787e-01 -6.52483404e-01 -7.65423700e-02
-1.39467633e+00 -8.37116480e-01 -4.16174084e-01 6.23678744e-01
8.34004104e-01 -1.44841298e-01 5.18372774e-01 9.25065354e-02
-3.38626176e-01 2.50756204e-01 -5.08293845e-02 5.39750159e-01
5.68345726e-01 -1.08172488e+00 1.34379119e-01 8.09171379e-01
5.49312413e-01 5.06321013e-01 5.13940394e-01 -3.03797364e-01
-1.37157309e+00 -1.22389209e+00 4.85633731e-01 -6.37403488e-01
8.97608638e-01 -3.61306161e-01 -9.09179628e-01 9.37327445e-01
2.59849101e-01 4.10419792e-01 5.08054018e-01 -1.96876362e-01
-8.16093385e-01 -1.32453918e-01 -4.93013233e-01 4.85025525e-01
1.66473031e+00 -9.56728518e-01 -1.07610583e+00 3.47057521e-01
4.71015841e-01 -1.36635616e-01 -5.61062336e-01 3.93885195e-01
8.60714078e-01 -1.06374741e+00 1.11062586e+00 -1.11931431e+00
8.26397240e-01 -3.70627522e-01 -3.44427615e-01 -1.19508684e+00
-3.17506343e-01 -5.36508977e-01 -6.72921687e-02 8.78293991e-01
-6.10473305e-02 -2.97599137e-01 8.51529479e-01 2.07655266e-01
-6.08524606e-02 -3.65444392e-01 -7.97102094e-01 -1.10149527e+00
-3.87351304e-01 -4.42912519e-01 2.56748289e-01 9.13940668e-01
1.72035202e-01 1.32193029e-01 -4.63493139e-01 -1.07203931e-01
6.04007006e-01 5.45028985e-01 6.28070712e-01 -1.00642383e+00
-2.95704842e-01 -5.65287054e-01 -9.34995532e-01 -1.24679554e+00
6.42337024e-01 -9.25204992e-01 3.49963367e-01 -1.21035957e+00
2.34852716e-01 -4.29689623e-02 -7.87228286e-01 5.86699307e-01
9.25381631e-02 3.87931496e-01 4.13524926e-01 4.85543549e-01
-1.16088617e+00 5.30447125e-01 9.26790416e-01 -4.88503098e-01
1.02302477e-01 -4.27117527e-01 -1.66978464e-01 6.69833779e-01
5.99091649e-01 -3.62328172e-01 -5.59464157e-01 -4.24040914e-01
-3.03793669e-01 -5.08658811e-02 6.66280866e-01 -1.17503393e+00
1.45923898e-01 -4.17385072e-01 5.21916628e-01 -3.95647824e-01
2.31075972e-01 -7.78175533e-01 -6.06196048e-03 5.36641121e-01
-9.36731040e-01 -1.74210668e-01 -1.24722503e-01 7.17927217e-01
-5.38276494e-01 5.70930764e-02 6.63204074e-01 -2.74179995e-01
-1.48084843e+00 2.84470409e-01 -3.89629334e-01 3.82737033e-02
1.04441857e+00 -5.57408869e-01 -3.89463484e-01 -3.72567117e-01
-7.90651917e-01 8.03197548e-02 4.94300753e-01 6.39376640e-01
7.07018018e-01 -1.43753433e+00 -3.03267866e-01 1.79661244e-01
3.57110918e-01 -5.00421643e-01 -5.08120209e-02 7.74081886e-01
-2.92066783e-01 4.23075199e-01 -7.65508711e-01 -9.92971063e-01
-1.17923832e+00 6.28687203e-01 1.14048280e-01 -1.03097133e-01
-8.07892501e-01 5.22090137e-01 5.56927383e-01 4.26584408e-02
3.95861030e-01 -6.71350300e-01 -2.95997351e-01 2.75149137e-01
6.74767077e-01 1.88430369e-01 -2.65366793e-01 -5.95343769e-01
-5.25045931e-01 5.96207917e-01 -1.47334794e-02 -7.21273646e-02
1.43555582e+00 -3.95476073e-02 -5.39048910e-02 7.30441749e-01
1.48153400e+00 -7.27551818e-01 -1.77192032e+00 -1.95439622e-01
5.07448554e-01 -6.73207283e-01 -4.00292605e-01 -3.25907081e-01
-9.66563284e-01 8.30840826e-01 5.70929289e-01 2.56815068e-02
1.17134881e+00 1.19364150e-01 4.81502652e-01 5.31919718e-01
4.74813342e-01 -9.80079889e-01 8.06032300e-01 7.52377868e-01
8.34635496e-01 -1.28760886e+00 -1.11952938e-01 -8.99970531e-02
-7.88335621e-01 1.27305102e+00 5.19827306e-01 -6.17452085e-01
3.66155356e-01 -1.44035399e-01 -1.16780795e-01 -1.14245154e-01
-9.37624931e-01 -4.94508982e-01 4.13468510e-01 6.96591079e-01
3.92773420e-01 -5.76848872e-02 9.79989208e-03 1.40035018e-01
1.97725687e-02 3.64300102e-01 4.16042298e-01 1.27276957e+00
-3.93566579e-01 -9.86405432e-01 -4.80787978e-02 4.83420640e-01
-7.30259344e-02 1.80883855e-01 -2.27832004e-01 4.85759765e-01
-6.95461361e-03 6.21397138e-01 5.10877967e-01 -5.18156946e-01
7.35989213e-02 2.28251725e-01 6.85360491e-01 -6.32798731e-01
-2.17732504e-01 -3.86314303e-01 2.85933614e-02 -1.26069438e+00
-1.09615672e+00 -6.94929540e-01 -1.15715396e+00 2.43640691e-01
4.98836309e-01 -1.42221615e-01 1.13833331e-01 9.90050793e-01
3.30265939e-01 6.14738047e-01 6.09853923e-01 -1.28275847e+00
-2.56567776e-01 -5.81620634e-01 -1.95658445e-01 1.19058859e+00
6.93837285e-01 -9.24534380e-01 -1.36343181e-01 9.43426788e-01] | [8.561161994934082, 0.7285941243171692] |
1423bed4-95da-48e6-8b0a-f46d669d618a | a-hierarchical-deep-architecture-and-mini | 1806.07987 | null | http://arxiv.org/abs/1806.07987v2 | http://arxiv.org/pdf/1806.07987v2.pdf | A Hierarchical Deep Architecture and Mini-Batch Selection Method For Joint Traffic Sign and Light Detection | Traffic light and sign detectors on autonomous cars are integral for road
scene perception. The literature is abundant with deep learning networks that
detect either lights or signs, not both, which makes them unsuitable for
real-life deployment due to the limited graphics processing unit (GPU) memory
and power available on embedded systems. The root cause of this issue is that
no public dataset contains both traffic light and sign labels, which leads to
difficulties in developing a joint detection framework. We present a deep
hierarchical architecture in conjunction with a mini-batch proposal selection
mechanism that allows a network to detect both traffic lights and signs from
training on separate traffic light and sign datasets. Our method solves the
overlapping issue where instances from one dataset are not labelled in the
other dataset. We are the first to present a network that performs joint
detection on traffic lights and signs. We measure our network on the
Tsinghua-Tencent 100K benchmark for traffic sign detection and the Bosch Small
Traffic Lights benchmark for traffic light detection and show it outperforms
the existing Bosch Small Traffic light state-of-the-art method. We focus on
autonomous car deployment and show our network is more suitable than others
because of its low memory footprint and real-time image processing time.
Qualitative results can be viewed at https://youtu.be/_YmogPzBXOw | ['Steven L. Waslander', 'Oles Andrienko', 'Ali Harakeh', 'Alex D. Pon'] | 2018-06-20 | null | null | null | null | ['traffic-sign-recognition', 'traffic-sign-detection'] | ['computer-vision', 'computer-vision'] | [ 7.57974312e-02 -3.36484581e-01 -1.70795098e-01 -4.89152253e-01
-5.95044374e-01 -4.13411230e-01 5.84546566e-01 -6.31827414e-01
-6.09603226e-01 3.55825037e-01 -5.80147445e-01 -9.58115637e-01
2.19568297e-01 -7.56246567e-01 -6.65362060e-01 -7.52204895e-01
3.55594635e-01 4.01685864e-01 8.99832129e-01 -2.94927984e-01
1.03309285e-02 5.47298551e-01 -2.10145998e+00 7.24644288e-02
5.49814284e-01 1.02322304e+00 2.83293247e-01 9.83276844e-01
1.59558460e-01 7.54363298e-01 -3.67409080e-01 -1.97219253e-01
6.97594166e-01 -1.54379249e-01 -8.13944936e-02 -1.33250669e-01
1.36271226e+00 -6.66429460e-01 -4.13925499e-01 8.29824448e-01
6.14377379e-01 -6.72973394e-02 4.50898170e-01 -1.90127409e+00
-2.14307472e-01 -2.59605706e-01 -5.53063214e-01 2.81069487e-01
-2.27769062e-01 7.75298476e-01 1.03944945e+00 -7.96467781e-01
3.34364891e-01 1.03443563e+00 6.35689914e-01 5.68140030e-01
-7.94217467e-01 -1.08890486e+00 -5.24123088e-02 6.35752797e-01
-1.13031113e+00 -7.83610642e-01 5.91413140e-01 -2.58271188e-01
1.07618415e+00 2.73801535e-01 6.19539440e-01 8.15279484e-01
-3.33207846e-02 1.00079894e+00 1.09827328e+00 -2.82926381e-01
2.80894905e-01 1.49446458e-01 2.43616164e-01 9.12495792e-01
5.70125997e-01 5.09467840e-01 -1.27100244e-01 2.77703673e-01
4.42602128e-01 -1.40669523e-02 3.20174634e-01 -1.77600369e-01
-7.52077937e-01 7.08611667e-01 4.29379582e-01 -9.74291712e-02
-1.52647868e-01 5.94910324e-01 4.28322732e-01 3.59397203e-01
2.42083929e-02 -8.64538252e-02 -3.91469687e-01 -1.15285903e-01
-9.25196588e-01 1.51567027e-01 6.61798298e-01 9.49256539e-01
8.89277697e-01 2.10435435e-01 1.36032075e-01 6.58105969e-01
3.80578101e-01 9.54845011e-01 -1.19194500e-01 -1.04642940e+00
3.12470496e-01 4.50127572e-01 -5.34288213e-02 -6.85614526e-01
-5.02397001e-01 4.95772995e-02 -3.63388062e-01 9.71295238e-01
8.06831539e-01 -3.06183726e-01 -1.25825620e+00 1.30764520e+00
3.99877697e-01 3.81165355e-01 -1.72035903e-01 1.02642453e+00
1.18856108e+00 4.04017866e-01 1.79164365e-01 5.66447079e-01
1.63171732e+00 -1.03268123e+00 -2.80811220e-01 -6.85947359e-01
8.68209064e-01 -7.02732146e-01 1.03836107e+00 1.89969406e-01
-5.94311059e-01 -6.27891481e-01 -1.04188883e+00 -3.27290028e-01
-7.48634338e-01 3.38572145e-01 8.17962170e-01 1.03345823e+00
-1.15159774e+00 -1.05826773e-01 -6.39956713e-01 -5.17397702e-01
7.02509403e-01 3.10675353e-01 -8.90768096e-02 -2.15635017e-01
-1.05858946e+00 1.10940278e+00 -6.88395873e-02 2.59389728e-01
-7.90849805e-01 -4.12586540e-01 -5.34356415e-01 -6.05691373e-02
5.02434134e-01 -4.06500340e-01 1.28598249e+00 -7.31786907e-01
-1.40860903e+00 9.75194693e-01 -2.15512887e-01 -4.68286037e-01
7.38265097e-01 2.58798748e-01 -6.30624413e-01 -1.26621157e-01
6.36101281e-03 9.37459409e-01 8.56022835e-01 -1.14718580e+00
-1.41583753e+00 9.66972299e-03 5.65217361e-02 -1.43526644e-01
2.32364312e-01 1.28145143e-02 -5.42293847e-01 3.21213543e-01
-4.25133765e-01 -1.21214569e+00 -6.30266741e-02 1.78845093e-01
-2.54548073e-01 -5.93358159e-01 1.42406988e+00 -1.46558627e-01
6.57122135e-01 -1.87828684e+00 -1.07847941e+00 2.32125461e-01
2.12693796e-01 8.42386842e-01 -3.84146571e-01 5.20128990e-03
4.53599952e-02 -2.11047292e-01 1.43717527e-01 -3.30518842e-01
1.06558271e-01 5.27001500e-01 -2.75426388e-01 6.52755797e-01
2.74006367e-01 1.04001200e+00 -8.62236440e-01 -4.08682764e-01
6.75257325e-01 5.61100304e-01 -2.57794976e-01 -2.77867407e-01
2.64130738e-02 2.87002269e-02 -1.63747475e-01 9.38010871e-01
9.57536161e-01 1.65448114e-01 -2.53418654e-01 -1.40302598e-01
-4.54755306e-01 3.90498161e-01 -1.23333192e+00 7.01511979e-01
-5.73591948e-01 1.49419820e+00 2.09518850e-01 -7.96402991e-01
9.26963329e-01 8.06212798e-02 4.06568730e-03 -1.22111714e+00
3.75422359e-01 4.42711085e-01 2.31664494e-01 -5.45716703e-01
5.25443375e-01 1.31853253e-01 2.01246485e-01 3.37049246e-01
-4.78439659e-01 -2.64838845e-01 6.58399820e-01 1.80043541e-02
1.24432993e+00 -1.57632545e-01 -2.39199877e-01 6.75663576e-02
4.28556561e-01 2.27011219e-02 4.01842684e-01 8.78125310e-01
-7.12154567e-01 3.21872026e-01 3.52609992e-01 -5.84306180e-01
-1.10163081e+00 -8.30551922e-01 -3.25075239e-01 1.52083015e+00
4.66292985e-02 3.57509479e-02 -3.15498739e-01 -6.73373938e-01
3.29846472e-01 8.23881745e-01 -4.75600630e-01 1.74699455e-01
-7.56013095e-01 -4.86341447e-01 7.49569952e-01 6.20270848e-01
5.27594268e-01 -1.02167320e+00 -1.11773157e+00 -7.74970129e-02
3.49707037e-01 -1.31389058e+00 -1.50144145e-01 4.05882090e-01
-3.10490072e-01 -1.39658511e+00 -3.02872270e-01 -8.55502427e-01
4.47856337e-01 8.39528561e-01 8.70794356e-01 2.28740603e-01
-5.32238185e-01 3.07853401e-01 7.33480603e-03 -8.02532196e-01
-4.81880873e-01 -7.55992457e-02 -1.01644471e-01 -4.77074133e-03
9.20219660e-01 -8.32254514e-02 -7.42615283e-01 6.34375751e-01
-6.73278451e-01 1.37033239e-01 6.24487519e-01 4.62585449e-01
9.91969034e-02 -9.01632160e-02 3.37469995e-01 -6.20313704e-01
3.04452419e-01 -1.86820194e-01 -1.08967006e+00 3.16128274e-03
-5.27773201e-01 -3.79654676e-01 4.74869937e-01 -1.01459533e-01
-7.25392044e-01 3.79874229e-01 -2.52579987e-01 -2.66673476e-01
-6.24230027e-01 -4.02073354e-01 6.43114969e-02 -2.45118380e-01
5.86645544e-01 4.65598656e-03 6.37828410e-02 -5.63976653e-02
3.92848969e-01 8.40757251e-01 5.00639796e-01 -6.46825731e-02
7.89069235e-01 7.21685588e-01 3.96879166e-01 -1.16672182e+00
-5.33197880e-01 -5.92993438e-01 -5.64597726e-01 -4.93849069e-01
7.71110713e-01 -9.06987250e-01 -1.17608190e+00 4.75188732e-01
-9.82350707e-01 -5.12674212e-01 -1.49609685e-01 3.21891397e-01
-3.76221865e-01 1.56084910e-01 -2.76445419e-01 -8.84507120e-01
-2.07896188e-01 -1.29001510e+00 1.20807731e+00 3.18320185e-01
1.05734341e-01 -6.70219302e-01 -1.73507497e-01 5.22966504e-01
6.59054458e-01 -1.66791491e-02 4.07059968e-01 -3.63766968e-01
-7.79147148e-01 -7.67782569e-01 -7.88928509e-01 5.25380015e-01
-5.40950820e-02 4.21800941e-01 -1.26222682e+00 6.82700379e-03
-5.70448577e-01 -2.35829875e-01 1.28915942e+00 4.01816994e-01
5.39867461e-01 1.86876953e-01 -3.85362655e-01 4.09889519e-01
1.23369622e+00 2.50330716e-01 5.97965777e-01 5.28822184e-01
9.11125660e-01 5.14253438e-01 4.10188198e-01 1.58833098e-02
5.95150828e-01 6.69389427e-01 5.79897225e-01 -5.98981142e-01
-5.69212258e-01 5.19843027e-02 3.82453114e-01 1.25103652e-01
4.54694331e-02 -2.52087206e-01 -1.17999446e+00 5.83124340e-01
-1.83040655e+00 -1.13780010e+00 -7.87881553e-01 1.95155084e+00
2.11400196e-01 2.89603204e-01 4.36620325e-01 2.89581686e-01
6.16554737e-01 -7.00812638e-02 -4.85185653e-01 -5.22632778e-01
-6.64663836e-02 1.19300652e-02 1.25992763e+00 4.44951028e-01
-1.16781878e+00 1.00023305e+00 5.71348286e+00 7.25983977e-01
-1.35046577e+00 8.38852748e-02 4.25222009e-01 -2.57286191e-01
2.35987768e-01 -4.15001921e-02 -1.25942719e+00 5.36897898e-01
1.04165149e+00 4.17510062e-01 1.13238521e-01 9.62489247e-01
5.38382769e-01 -4.76913601e-01 -1.11469805e+00 9.34112906e-01
-1.39080346e-01 -1.12583196e+00 -4.41110969e-01 2.23350883e-01
5.16949058e-01 8.69063735e-01 1.12342201e-01 4.50447053e-01
5.20846426e-01 -7.18720138e-01 6.55400932e-01 1.75168842e-01
7.98776329e-01 -5.53959250e-01 6.76362932e-01 2.93152660e-01
-1.34139788e+00 -1.99674621e-01 -3.53587717e-01 -1.00391500e-01
3.37409109e-01 2.88752884e-01 -1.09076333e+00 -1.05591461e-01
5.04899800e-01 5.30182540e-01 -8.53375673e-01 1.27249241e+00
-5.28199494e-01 8.38907182e-01 -8.13585401e-01 -3.45928460e-01
6.23495460e-01 -1.22930810e-01 3.35295618e-01 1.45517087e+00
1.38958856e-01 -3.71183485e-01 2.35329211e-01 6.86861753e-01
4.61921729e-02 -3.15477401e-01 -8.44621718e-01 4.90644097e-01
4.29022133e-01 1.53843915e+00 -9.54367399e-01 -5.51137328e-01
-7.66427338e-01 2.90272236e-01 -2.32915618e-02 2.57048219e-01
-1.09868968e+00 -5.60345292e-01 8.43500733e-01 3.43400419e-01
5.17138302e-01 -2.39837423e-01 -4.45987433e-01 -6.06047034e-01
-1.29662842e-01 -4.83756334e-01 2.26610422e-01 -7.13293076e-01
-8.87478590e-01 2.76881486e-01 -1.30011380e-01 -1.32281566e+00
2.89728167e-03 -1.05472112e+00 -9.03970599e-01 5.27063966e-01
-1.97427571e+00 -1.18067455e+00 -5.56918919e-01 4.61952388e-01
4.05935943e-01 -1.64477870e-01 2.25395039e-01 8.12687337e-01
-8.28976750e-01 5.97877383e-01 9.04161111e-02 2.38630176e-01
7.06881702e-01 -9.97209132e-01 6.66052878e-01 1.01403177e+00
-9.10706352e-03 1.10938028e-01 6.32323623e-01 -3.93901467e-01
-1.39976203e+00 -1.10038149e+00 9.28850830e-01 -7.68450856e-01
6.73727095e-01 -4.95692521e-01 -5.28021753e-01 5.01823664e-01
1.04198316e-02 3.63680035e-01 1.74602598e-01 -1.52530983e-01
-3.31355035e-01 -4.97473180e-01 -9.75639939e-01 8.22291911e-01
8.79295945e-01 -3.04466397e-01 -2.27590501e-01 3.29015762e-01
1.36956954e-02 -2.57283211e-01 1.48176983e-01 1.61208287e-01
7.81934023e-01 -1.05019033e+00 8.18353772e-01 -2.33864397e-01
2.72149257e-02 -5.93724608e-01 -8.47842172e-02 -8.33669424e-01
-1.60531238e-01 -2.90855020e-01 1.57942593e-01 9.88423049e-01
3.80736500e-01 -9.65383708e-01 9.54927206e-01 8.92096758e-01
-2.40539834e-01 -4.65218902e-01 -1.14854598e+00 -8.85646760e-01
-1.99403852e-01 -1.05361843e+00 5.48476338e-01 5.20488322e-01
-6.21600449e-01 4.63686317e-01 -3.65871251e-01 2.43543126e-02
5.78501344e-01 1.78759191e-02 1.31874561e+00 -1.22286463e+00
1.84665456e-01 -8.55330646e-01 -5.94318926e-01 -1.06486201e+00
1.75542217e-02 -7.56312013e-01 3.06515813e-01 -1.58550322e+00
-2.02671930e-01 -4.89629418e-01 -1.03786923e-01 8.27553451e-01
1.05480090e-01 6.11464679e-01 3.46412003e-01 2.62954738e-02
-7.59854138e-01 1.05058089e-01 1.03667796e+00 -6.91573322e-02
-1.31982744e-01 1.07257605e-01 -2.94565797e-01 7.27997839e-01
8.06319773e-01 -3.76124650e-01 -1.14448674e-01 -3.44304323e-01
1.75342992e-01 -6.04344726e-01 7.73653030e-01 -1.20220149e+00
5.55687964e-01 1.03143677e-01 3.01206142e-01 -8.67503107e-01
3.05413634e-01 -1.02678418e+00 -3.08903486e-01 6.21413171e-01
6.30589109e-03 -2.73739755e-01 3.92828017e-01 3.05626869e-01
1.79358661e-01 -5.84465265e-02 1.05927479e+00 1.21851206e-01
-1.27476096e+00 2.02226177e-01 -7.72364199e-01 -1.25330985e-01
1.08586574e+00 -7.14120269e-01 -7.48333335e-01 -3.00483614e-01
1.11198828e-01 4.55975324e-01 3.07176858e-01 6.92197680e-01
3.67514879e-01 -1.24399555e+00 -6.67911232e-01 5.34381568e-01
2.78067291e-01 -8.54673162e-02 1.07881002e-01 9.37410951e-01
-6.85873330e-01 7.30842173e-01 -2.86297411e-01 -7.78698921e-01
-1.58341181e+00 1.60765693e-01 3.50664526e-01 1.94288090e-01
-6.63178384e-01 6.77104592e-01 1.14959136e-01 -4.30712730e-01
2.33499274e-01 -5.16611516e-01 2.77015697e-02 1.42155483e-01
4.59149152e-01 8.29230189e-01 2.63323665e-01 -9.06865358e-01
-6.34559751e-01 6.68012559e-01 6.17518499e-02 2.62919009e-01
1.10651600e+00 -1.14866130e-01 3.47133875e-01 1.03120886e-01
1.24576175e+00 -3.60872000e-01 -1.32974041e+00 -1.47202164e-01
-1.96767047e-01 -4.48225379e-01 4.01665300e-01 -5.70984423e-01
-1.23635709e+00 8.34809244e-01 9.97924566e-01 2.07691699e-01
8.02645981e-01 -3.08386646e-02 9.70618963e-01 7.91070938e-01
1.12611428e-01 -1.43611932e+00 -5.15484691e-01 8.26870501e-01
3.11749786e-01 -1.86633813e+00 -7.00620115e-02 -1.21228017e-01
-5.20295680e-01 1.08981681e+00 7.47368813e-01 -1.91839188e-01
5.51835358e-01 5.43138325e-01 4.15984571e-01 -3.98725361e-01
-8.72683883e-01 -9.18115318e-01 2.76000351e-01 6.22973859e-01
2.07449809e-01 1.97851509e-01 -4.26103286e-02 -3.15108359e-01
-1.06849425e-01 1.68048553e-02 4.54290062e-01 8.83720934e-01
-6.84402883e-01 -8.34853649e-01 -3.28873843e-01 6.75711572e-01
-5.83873950e-02 7.03008845e-02 -3.71926367e-01 9.86727953e-01
5.24241984e-01 1.23965287e+00 3.49395186e-01 -2.41172284e-01
5.60499370e-01 -3.86155508e-02 1.94173381e-01 -2.47577071e-01
-2.85792917e-01 -1.66699678e-01 3.13933879e-01 -6.59859955e-01
-1.19636066e-01 -5.98103225e-01 -1.08896625e+00 -5.13507009e-01
-2.43957683e-01 -4.20103371e-01 9.81292248e-01 7.59965122e-01
2.46912599e-01 2.95556188e-01 5.09338975e-01 -1.25124741e+00
-7.18068406e-02 -5.96462667e-01 -5.03093779e-01 2.38028169e-01
6.95938587e-01 -6.44258082e-01 -4.06872243e-01 -3.40541750e-01] | [8.019814491271973, -0.8958208560943604] |
2da50702-3299-4999-8963-52e88edcbda8 | on-translation-invariance-in-cnns | 2003.07064 | null | https://arxiv.org/abs/2003.07064v2 | https://arxiv.org/pdf/2003.07064v2.pdf | On Translation Invariance in CNNs: Convolutional Layers can Exploit Absolute Spatial Location | In this paper we challenge the common assumption that convolutional layers in modern CNNs are translation invariant. We show that CNNs can and will exploit the absolute spatial location by learning filters that respond exclusively to particular absolute locations by exploiting image boundary effects. Because modern CNNs filters have a huge receptive field, these boundary effects operate even far from the image boundary, allowing the network to exploit absolute spatial location all over the image. We give a simple solution to remove spatial location encoding which improves translation invariance and thus gives a stronger visual inductive bias which particularly benefits small data sets. We broadly demonstrate these benefits on several architectures and various applications such as image classification, patch matching, and two video classification datasets. | ['Jan C. van Gemert', 'Osman Semih Kayhan'] | 2020-03-16 | on-translation-invariance-in-cnns-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Kayhan_On_Translation_Invariance_in_CNNs_Convolutional_Layers_Can_Exploit_Absolute_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Kayhan_On_Translation_Invariance_in_CNNs_Convolutional_Layers_Can_Exploit_Absolute_CVPR_2020_paper.pdf | cvpr-2020-6 | ['small-data', 'patch-matching'] | ['computer-vision', 'computer-vision'] | [ 2.12913498e-01 -5.28653041e-02 -5.18087268e-01 -4.76446480e-01
-1.65716261e-01 -8.29294741e-01 6.50511384e-01 -3.29437435e-01
-6.74653351e-01 4.47034478e-01 1.83348194e-01 -4.08790350e-01
2.22411379e-01 -8.28256428e-01 -1.18596971e+00 -5.32645166e-01
1.07104005e-02 -2.53076226e-01 3.98299992e-01 -2.14346126e-01
2.71164000e-01 8.77724171e-01 -1.43184400e+00 7.30060399e-01
7.38921463e-02 9.79938447e-01 1.02308631e-01 5.99105954e-01
1.75535306e-01 5.40631175e-01 -3.66258472e-01 1.46660149e-01
6.11063540e-01 -1.34763896e-01 -8.44065845e-01 -6.34294450e-02
1.04350162e+00 -3.66591841e-01 -6.56023800e-01 8.96775067e-01
3.35585952e-01 1.31964609e-01 4.49923217e-01 -9.62434947e-01
-1.16605794e+00 2.61729181e-01 -6.11770689e-01 6.05858862e-01
-7.61406049e-02 1.73131734e-01 1.00469029e+00 -1.01272726e+00
8.68665636e-01 1.18808138e+00 9.02076364e-01 5.13174117e-01
-1.34055984e+00 -3.06415856e-01 3.38558018e-01 1.04925232e-02
-1.51919723e+00 -5.49084246e-01 7.15662241e-01 -3.12444568e-01
1.25875461e+00 2.07194105e-01 4.46476191e-01 8.73459220e-01
4.23627824e-01 7.23509073e-01 1.10942221e+00 -4.72462654e-01
-3.41204442e-02 -2.14929059e-01 -1.48982584e-01 6.23180032e-01
-1.87081639e-02 2.79828101e-01 -5.61471283e-01 3.37043822e-01
1.39101946e+00 4.14733231e-01 -4.40898240e-01 -4.42842692e-01
-1.44089091e+00 6.15663052e-01 1.35811305e+00 4.23948109e-01
-3.46376784e-02 5.49687803e-01 1.05062172e-01 6.00295603e-01
1.36457682e-01 5.51522255e-01 -6.09072745e-01 6.27462685e-01
-7.28973746e-01 8.72902721e-02 1.89814433e-01 1.08211243e+00
1.13189280e+00 1.22502865e-02 -9.59824398e-02 5.63179076e-01
2.65740249e-02 4.67559427e-01 5.74006677e-01 -8.61929059e-01
2.75109917e-01 4.44946766e-01 -1.44306794e-01 -1.07996726e+00
-5.84293365e-01 -6.73423827e-01 -9.17061150e-01 3.99525821e-01
4.37419862e-01 7.76213333e-02 -1.34046745e+00 2.02011943e+00
-2.76162416e-01 -1.12203687e-01 -1.54071614e-01 9.47950602e-01
7.74218440e-01 3.40721190e-01 3.23651768e-02 4.28184897e-01
1.29818213e+00 -9.39315438e-01 -3.97992551e-01 -6.65882528e-01
7.95952260e-01 -7.09510207e-01 1.24193943e+00 -1.59122825e-01
-9.44646358e-01 -7.84547627e-01 -1.34159553e+00 -5.25700748e-01
-7.14209139e-01 2.46764451e-01 8.35659325e-01 1.99891925e-01
-1.39816904e+00 6.92321599e-01 -6.10136628e-01 -5.08267760e-01
7.55924463e-01 5.61303139e-01 -6.21174812e-01 -5.93654849e-02
-9.81354415e-01 9.09172237e-01 1.76358402e-01 6.50520325e-02
-4.53177631e-01 -7.75533140e-01 -9.66604531e-01 1.64375588e-01
-2.12416664e-01 -7.37858057e-01 1.16274190e+00 -1.56340432e+00
-9.24340785e-01 1.09361649e+00 -3.76917005e-01 -4.95848358e-01
1.71775892e-01 6.81111589e-02 -1.47490248e-01 6.61088079e-02
8.38951543e-02 1.39296281e+00 9.93903756e-01 -9.49083149e-01
-3.95103246e-01 -4.32177275e-01 3.01494509e-01 1.65900066e-01
-3.41221631e-01 -2.23341525e-01 -1.75477549e-01 -8.38863850e-01
5.29722989e-01 -9.58735824e-01 -4.63661365e-02 4.74635065e-01
-1.50817618e-01 7.60370195e-02 9.77134764e-01 -1.22112341e-01
6.73347831e-01 -2.36739278e+00 8.95480961e-02 1.39266849e-01
1.89864382e-01 -9.10955388e-03 -4.36734140e-01 1.31936327e-01
-5.11828005e-01 1.13086402e-01 -3.67673226e-02 -7.04199672e-02
-2.21317187e-01 3.68139267e-01 -7.02205241e-01 7.07019508e-01
4.44797873e-01 1.52668738e+00 -6.23139262e-01 -8.37552398e-02
2.06124350e-01 5.48050582e-01 -6.72842443e-01 -3.17754865e-01
7.90301152e-03 2.67816871e-01 -1.19056655e-02 5.11650443e-01
8.32634687e-01 -3.75307173e-01 -4.78339158e-02 -3.44239384e-01
-2.21212760e-01 2.05746278e-01 -7.83566654e-01 1.86006927e+00
-4.48753327e-01 1.34032857e+00 -7.40183517e-02 -9.94152784e-01
6.68786108e-01 2.41859723e-02 8.72129947e-02 -1.05173612e+00
1.82973027e-01 1.21665381e-01 1.55223519e-01 -6.99622855e-02
2.25743622e-01 -1.96827650e-01 1.78379223e-01 2.98215002e-01
2.62464792e-01 9.50774997e-02 -2.59370774e-01 -1.21278942e-01
9.15969014e-01 3.06581020e-01 4.78354469e-02 -7.61636615e-01
2.41174102e-01 -5.85806035e-02 3.07297021e-01 8.47819805e-01
-2.59135664e-01 9.36524093e-01 1.54442579e-01 -1.22678161e+00
-1.02477336e+00 -1.10547268e+00 -3.30681890e-01 1.33232510e+00
1.94288656e-01 -1.83259919e-01 -4.28439796e-01 -5.18834531e-01
-1.50803644e-02 -8.32883716e-02 -1.18150127e+00 -2.67585367e-01
-7.35579133e-01 -3.68595600e-01 6.25319421e-01 1.14731276e+00
7.35920548e-01 -1.02355909e+00 -7.41984487e-01 -1.97350889e-01
7.80896768e-02 -1.00609124e+00 -4.94262636e-01 5.72062016e-01
-8.42462897e-01 -7.79987812e-01 -7.94483006e-01 -1.06383932e+00
1.06553292e+00 5.98375082e-01 1.07412195e+00 2.50632167e-01
-3.87680233e-01 1.44799218e-01 -1.99831128e-02 -2.91493118e-01
2.08045810e-01 3.78840894e-01 -1.25259131e-01 -2.37274081e-01
4.86574233e-01 -5.50203383e-01 -7.64503837e-01 4.46571290e-01
-9.49048042e-01 6.32712152e-03 6.69897974e-01 9.09435332e-01
5.25342941e-01 -2.36064225e-01 2.41638675e-01 -6.93699598e-01
3.63074318e-02 -9.13909674e-02 -5.22389233e-01 6.30527884e-02
-1.21168047e-02 2.82125175e-01 5.96873403e-01 -4.86445248e-01
-6.18057966e-01 2.35214457e-01 -5.77108823e-02 -3.83036017e-01
-2.45788544e-01 2.42148727e-01 1.37441874e-01 -6.48565590e-01
1.09228528e+00 1.10617325e-01 -1.77648798e-01 -2.62950540e-01
3.41212034e-01 8.64936039e-02 6.71324790e-01 -3.06273639e-01
8.37617397e-01 9.39199328e-01 2.75263339e-01 -8.44826341e-01
-7.70369411e-01 -2.66739070e-01 -1.03556049e+00 1.55704692e-01
7.61068225e-01 -9.13997889e-01 -6.30062699e-01 2.82921821e-01
-1.25461912e+00 -5.00105798e-01 -3.05082649e-01 3.86379927e-01
-5.44365883e-01 -7.87639916e-02 -5.78788817e-01 2.34217364e-02
9.03553665e-02 -1.04591894e+00 1.11390865e+00 1.09562621e-01
-2.22090781e-01 -1.13178110e+00 -1.63709939e-01 -5.30476809e-01
6.81957960e-01 4.87661771e-02 7.80468166e-01 -8.26661885e-02
-9.04428840e-01 -4.06783558e-02 -6.63049400e-01 1.04517467e-01
6.97189942e-02 -3.27218473e-02 -1.26476216e+00 -4.50046659e-01
-3.11727017e-01 -1.87131017e-01 1.46616638e+00 6.17798030e-01
1.25467122e+00 -8.88248309e-02 -3.88680071e-01 1.18658948e+00
1.55343091e+00 -1.64625242e-01 9.11187947e-01 4.37874615e-01
8.11383605e-01 5.65011322e-01 -7.43870065e-02 -1.81096300e-01
-6.63821474e-02 5.71430981e-01 3.41429442e-01 -6.85773313e-01
-5.30461609e-01 -2.17100874e-01 9.44904014e-02 2.28033498e-01
-1.67888701e-01 -1.37575287e-02 -7.68122494e-01 7.34765232e-01
-1.74758494e+00 -8.96754503e-01 2.30951056e-01 2.03217864e+00
6.20372117e-01 9.55607742e-02 -3.01622480e-01 2.18407474e-02
4.61253941e-01 2.87711471e-01 -3.61267924e-01 -5.70991814e-01
-6.47587061e-01 4.02241230e-01 1.09644806e+00 5.48712671e-01
-1.42240989e+00 9.37893093e-01 8.11287594e+00 3.98119539e-01
-1.63559198e+00 -1.03078753e-01 6.07446194e-01 4.09042910e-02
-3.00219029e-01 -1.27211183e-01 -5.25434315e-01 -7.13282973e-02
3.82574588e-01 3.60689819e-01 3.84469211e-01 7.56100595e-01
-2.34143898e-01 1.54026479e-01 -1.28850722e+00 8.77967775e-01
-1.30668804e-01 -1.69493401e+00 2.06170097e-01 1.53113872e-01
1.00519145e+00 4.50502455e-01 6.06400907e-01 -6.23847507e-02
3.18821929e-02 -1.40374625e+00 7.27453232e-01 2.30006218e-01
1.12289190e+00 -6.24662340e-01 4.53629136e-01 5.91221452e-03
-1.04454184e+00 -1.94562510e-01 -7.64721632e-01 -5.44731081e-01
-3.83502662e-01 2.01689810e-01 -5.18914104e-01 1.24231549e-02
8.88118327e-01 8.92137945e-01 -7.11916924e-01 8.67800593e-01
-2.29885533e-01 1.09953985e-01 -3.05329174e-01 3.25894982e-01
5.24417877e-01 3.58290911e-01 1.42015889e-01 1.31301379e+00
1.79458439e-01 -9.61179361e-02 -2.73335308e-01 8.94230902e-01
-3.24429423e-01 -1.17108911e-01 -1.15276706e+00 4.31965023e-01
2.64757872e-01 9.25566971e-01 -9.49162543e-01 -2.64371365e-01
-7.05766857e-01 1.04465830e+00 4.48906153e-01 8.04128170e-01
-5.36355078e-01 -5.44491649e-01 8.59850347e-01 2.52430826e-01
6.77496910e-01 -3.66398007e-01 -6.03155792e-01 -1.10202897e+00
-1.19754352e-01 -5.08902490e-01 1.06205538e-01 -7.95604885e-01
-7.95133710e-01 4.38576013e-01 -2.03912988e-01 -1.24969947e+00
7.93486834e-02 -1.28047574e+00 -5.93302310e-01 8.61919105e-01
-1.71157110e+00 -1.19811702e+00 -2.04081625e-01 8.15411687e-01
1.69397220e-01 8.09626579e-02 7.95815766e-01 3.02121609e-01
-1.04545787e-01 7.24879444e-01 4.28715236e-02 5.17268121e-01
8.37595701e-01 -1.10110939e+00 7.29817808e-01 8.60399127e-01
4.21290517e-01 1.16688287e+00 3.09413105e-01 -3.96066606e-02
-1.16247296e+00 -9.88391340e-01 9.53624904e-01 -5.10429204e-01
3.11528981e-01 -5.23536742e-01 -8.50504875e-01 1.10013628e+00
3.44198376e-01 5.08663654e-01 3.61108154e-01 2.45451704e-01
-1.03179026e+00 -1.32526562e-01 -9.68922794e-01 7.24595010e-01
1.15143037e+00 -9.29259241e-01 -5.45790017e-01 2.12975711e-01
4.27067339e-01 -3.37650001e-01 -3.78332078e-01 5.76447606e-01
7.48749435e-01 -9.13619220e-01 1.12224269e+00 -7.77176619e-01
3.16866308e-01 -4.60064858e-01 -2.82559246e-01 -1.20669663e+00
-7.01548219e-01 -4.05259103e-01 3.70900124e-01 4.76258099e-01
4.95238751e-01 -7.98857510e-01 5.74720621e-01 3.62614155e-01
-1.23225257e-01 -4.74376529e-01 -9.88191903e-01 -7.48580933e-01
4.02363092e-01 -2.58945823e-01 5.94568491e-01 1.02948344e+00
-6.75912872e-02 2.18663111e-01 -7.67729506e-02 4.41705287e-02
3.22530180e-01 2.67310321e-01 3.75998944e-01 -8.86417866e-01
-6.57129735e-02 -7.59011090e-01 -7.96410978e-01 -1.50551915e+00
3.72986831e-02 -7.59744346e-01 5.05977497e-02 -1.05818164e+00
2.62135249e-02 -2.74060100e-01 -5.41549563e-01 7.68862009e-01
2.20331952e-01 1.12043250e+00 -1.21122666e-01 2.15160072e-01
-3.65645349e-01 -5.74636133e-03 1.54404581e+00 -2.24520147e-01
1.85694352e-01 -2.93884605e-01 -6.26722932e-01 7.38889337e-01
8.00482571e-01 -1.72844812e-01 -4.39367652e-01 -1.07371771e+00
3.77572536e-01 -7.13094473e-01 7.19883084e-01 -9.31670606e-01
2.14327380e-01 -8.15568566e-02 1.07991326e+00 -2.55821675e-01
1.32574782e-01 -8.19002748e-01 -2.77120620e-01 5.01120389e-01
-5.24277627e-01 1.85327098e-01 5.61605215e-01 2.54423082e-01
-2.40893990e-01 9.97088626e-02 9.69886303e-01 -2.93630004e-01
-1.12079012e+00 3.98513466e-01 -3.37520987e-01 -2.19409317e-01
7.03353822e-01 -3.33733529e-01 -4.77972776e-01 -2.53071398e-01
-7.63597131e-01 -2.63310939e-01 8.30694854e-01 5.61310053e-01
5.84640086e-01 -1.63355255e+00 -3.89160931e-01 8.17850351e-01
3.26159328e-01 -1.75236911e-01 1.18714780e-01 7.04432845e-01
-7.55474269e-01 6.27128184e-01 -7.30631769e-01 -8.08224261e-01
-9.68385339e-01 8.72408032e-01 6.14103258e-01 4.95346725e-01
-6.53633475e-01 1.16419375e+00 8.37016344e-01 -1.62347972e-01
1.38416275e-01 -7.87127972e-01 1.15655325e-01 -3.24000537e-01
6.02833688e-01 -2.31478304e-01 1.99957430e-01 -8.81994486e-01
-6.09492838e-01 1.01015651e+00 3.65161747e-02 -8.90779421e-02
1.06266856e+00 -8.94703194e-02 -8.69505852e-02 6.68656006e-02
1.66907203e+00 -2.42832646e-01 -1.39681268e+00 -4.87898916e-01
-3.13473433e-01 -7.11866260e-01 9.96633396e-02 -5.07577240e-01
-1.13164485e+00 1.34839487e+00 5.11218488e-01 3.88032570e-02
1.20875752e+00 3.97668406e-02 3.18714976e-01 5.70500135e-01
1.21877424e-01 -9.00818050e-01 -5.64890681e-03 6.70974910e-01
9.19530749e-01 -1.20008147e+00 -4.98495363e-02 -1.72392398e-01
-2.35543564e-01 1.23855245e+00 6.10474825e-01 -5.75094104e-01
7.98515260e-01 3.40071023e-01 2.68296987e-01 -1.85113966e-01
-6.22007549e-01 -4.22110915e-01 6.21909320e-01 8.51964831e-01
6.96273685e-01 -3.71541120e-02 1.09060667e-01 -1.61268085e-01
1.53843537e-02 -3.74560803e-01 1.15631856e-01 1.06367838e+00
-4.99767005e-01 -8.61181676e-01 -2.54923582e-01 6.04971536e-02
-4.83446807e-01 -1.97283506e-01 -5.03622234e-01 8.12438846e-01
3.78955513e-01 3.81600291e-01 5.48527777e-01 -2.38665298e-01
1.89672500e-01 -1.40856817e-01 8.66214633e-01 -4.66609269e-01
-2.92468846e-01 -4.13102433e-02 -4.74422812e-01 -6.36841059e-01
-5.52477002e-01 -2.77920514e-01 -1.18054283e+00 -4.50756192e-01
-9.25009623e-02 -5.38056076e-01 5.65604806e-01 6.92519903e-01
3.50049376e-01 5.46282530e-01 4.63603109e-01 -9.52319443e-01
-1.48525044e-01 -8.22607934e-01 -2.50172466e-01 4.10082906e-01
9.10525382e-01 -5.78769922e-01 -1.35852277e-01 2.13889971e-01] | [9.255118370056152, 2.157003402709961] |
05317035-3b7d-4971-ae3b-b3975c2ac168 | rps-portfolio-asset-selection-using-graph | 2111.15634 | null | https://arxiv.org/abs/2111.15634v1 | https://arxiv.org/pdf/2111.15634v1.pdf | RPS: Portfolio Asset Selection using Graph based Representation Learning | Portfolio optimization is one of the essential fields of focus in finance. There has been an increasing demand for novel computational methods in this area to compute portfolios with better returns and lower risks in recent years. We present a novel computational method called Representation Portfolio Selection (RPS) by redefining the distance matrix of financial assets using Representation Learning and Clustering algorithms for portfolio selection to increase diversification. RPS proposes a heuristic for getting closer to the optimal subset of assets. Using empirical results in this paper, we demonstrate that widely used portfolio optimization algorithms, such as MVO, CLA, and HRP, can benefit from our asset subset selection. | ['Erfan Loghmani', 'Ali Owfi', 'Parsa Alian', 'Mohammadamin Fazli'] | 2021-11-28 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-2.05266818e-01 -2.98466682e-01 5.36921397e-02 -3.65823299e-01
-5.28145015e-01 -7.13374734e-01 3.86474133e-01 -1.85735822e-01
6.36016801e-02 8.21007907e-01 3.55804622e-01 -3.35394412e-01
-1.14092326e+00 -1.41389203e+00 1.40237466e-01 -6.80387139e-01
-1.83069393e-01 6.82452381e-01 -1.71335340e-02 -3.28049183e-01
9.10546064e-01 9.29767966e-01 -1.53272629e+00 4.00428236e-01
7.58662522e-01 1.20236802e+00 -1.76284447e-01 1.39370635e-01
-3.97786498e-01 9.69930172e-01 -6.13412797e-01 -9.94266808e-01
1.03191817e+00 -5.94223499e-01 -2.80318767e-01 -4.52107489e-02
-1.76359698e-01 2.34025959e-02 -2.41360337e-01 1.02423441e+00
6.16264462e-01 5.02629697e-01 1.00622463e+00 -8.38318825e-01
-5.90226173e-01 1.00044250e+00 -6.99458361e-01 3.70654166e-01
-4.86655273e-02 -2.69720227e-01 1.36103261e+00 -7.94474840e-01
2.34096408e-01 1.13984096e+00 6.01160645e-01 1.16366453e-01
-8.29237998e-01 -6.57311440e-01 -1.25160396e-01 3.85430127e-01
-1.16297221e+00 9.39149484e-02 1.08999825e+00 -4.46703702e-01
8.17881107e-01 6.39178336e-01 6.16093755e-01 3.70106250e-01
2.84622133e-01 4.31854188e-01 1.18988621e+00 -4.61188465e-01
3.73277843e-01 3.80143642e-01 2.75354564e-01 7.29060397e-02
8.94487321e-01 5.84325552e-01 -4.51511234e-01 -4.35520768e-01
6.28540635e-01 3.41762096e-01 -9.94754136e-02 -4.42889452e-01
-6.48895860e-01 1.43292892e+00 -2.50957608e-01 1.20687932e-01
-4.54130739e-01 -4.19700593e-02 1.88479975e-01 5.91557741e-01
4.06719297e-01 1.02840281e+00 -1.66520879e-01 -1.78718641e-01
-9.37487662e-01 4.56585884e-01 1.02101386e+00 6.52394712e-01
3.25857401e-01 2.48046905e-01 -2.15524748e-01 6.87781930e-01
3.51453722e-01 5.09559214e-01 6.08118534e-01 -1.13094413e+00
8.52281928e-01 7.06799686e-01 1.12740129e-01 -1.31132102e+00
-2.25667328e-01 -7.60005653e-01 -6.04137242e-01 7.24790514e-01
-9.69897676e-03 -1.06567569e-01 6.67403564e-02 9.24703240e-01
-2.74180379e-02 2.77102776e-02 2.44791716e-01 4.48767096e-01
2.95232266e-01 5.13739228e-01 -4.90847141e-01 -4.06660527e-01
8.09116244e-01 -8.20449889e-01 -4.10255790e-01 4.64887172e-01
2.69688427e-01 -6.95057631e-01 3.46952528e-01 7.89798081e-01
-1.05526519e+00 -3.09861600e-01 -1.04744530e+00 1.03605652e+00
-3.29919994e-01 -1.49058715e-01 1.04609728e+00 1.35110545e+00
-8.09131563e-01 1.22599947e+00 -3.03947330e-01 2.85235971e-01
5.21232188e-01 4.61549729e-01 1.66666284e-01 3.89093727e-01
-1.07569134e+00 9.64314163e-01 6.59300625e-01 -2.29363009e-01
-2.39612803e-01 -6.66362226e-01 -3.73390317e-01 2.72621751e-01
3.40859056e-01 -3.76345336e-01 6.15770638e-01 -8.34060013e-01
-1.68723881e+00 3.04346621e-01 8.16603780e-01 -7.89839327e-01
6.17288828e-01 -2.83595771e-01 -4.41241354e-01 3.02168876e-02
-5.18872619e-01 -3.09961081e-01 8.64733517e-01 -6.47132337e-01
-7.30247200e-01 -1.68914616e-01 -2.87696868e-01 1.95789322e-01
-6.93598270e-01 2.13452727e-01 4.00432199e-01 -1.35586703e+00
-5.60261942e-02 -5.75893939e-01 -5.19017935e-01 -6.27486527e-01
2.38713603e-02 -2.11409777e-01 2.84462690e-01 -5.90836883e-01
1.60450864e+00 -1.78629684e+00 3.13896894e-01 1.01942205e+00
-1.44698068e-01 1.85913086e-01 5.72140478e-02 6.09743059e-01
-3.59134883e-01 1.35073245e-01 -2.72137463e-01 -6.58940002e-02
2.59025872e-01 4.51263040e-02 -8.74129415e-01 4.43193346e-01
-7.91140571e-02 6.41490340e-01 -5.44473231e-01 -2.36956358e-01
2.83765942e-01 4.26764935e-02 -5.20416558e-01 4.12855029e-01
2.61419024e-02 1.49375066e-01 -5.26849389e-01 8.23543072e-01
6.65961444e-01 -2.69971844e-02 7.84064755e-02 1.34423986e-01
-1.35044619e-01 1.02520660e-01 -1.71165442e+00 7.49503314e-01
-7.50107989e-02 2.64932036e-01 -7.43539810e-01 -1.25104976e+00
1.50373697e+00 1.32217258e-01 7.76009083e-01 -5.65827370e-01
1.27942666e-01 3.62524539e-01 -2.55441014e-02 -1.81839794e-01
5.17069340e-01 -4.15787011e-01 -7.15984777e-02 1.19157815e+00
-3.54577988e-01 1.06445484e-01 2.98406363e-01 -2.40597874e-01
9.41143751e-01 5.91174550e-02 5.09235680e-01 -3.11658829e-01
7.01612830e-01 -2.60660350e-01 6.36274517e-01 6.05536044e-01
2.32522368e-01 6.74417078e-01 5.00049412e-01 -5.07654011e-01
-9.00312901e-01 -1.05483019e+00 -2.89561272e-01 6.91515446e-01
-5.13915658e-01 -5.51481992e-02 -3.15301090e-01 -5.92142403e-01
5.60090780e-01 8.35987508e-01 -5.14637589e-01 -7.84631819e-02
-8.53797436e-01 -1.11699271e+00 7.92318881e-02 6.88525259e-01
3.31443787e-01 -1.37822831e+00 -6.41452491e-01 4.90701467e-01
6.46156371e-01 -9.62723270e-02 -2.11246148e-01 1.16710924e-01
-1.21147585e+00 -1.06140745e+00 -9.47970152e-01 -1.27961501e-01
4.78366703e-01 1.07035033e-01 1.29156590e+00 -7.45381936e-02
-1.93206683e-01 2.88956970e-01 -7.52477288e-01 -4.49682325e-01
-1.20573834e-01 -7.93100148e-03 -7.76035488e-02 2.74031758e-01
7.92469606e-02 -5.57122171e-01 -4.47646141e-01 3.45176339e-01
-5.55654645e-01 -8.01022470e-01 6.39469266e-01 1.00194490e+00
7.06401706e-01 8.22050214e-01 1.02906334e+00 -1.19095576e+00
9.29269433e-01 -8.85946572e-01 -1.10853171e+00 6.93992198e-01
-9.48474586e-01 2.61458099e-01 4.51085269e-01 -2.75519401e-01
-1.09596014e+00 -4.04195935e-01 3.08005929e-01 -4.44386095e-01
6.96811616e-01 7.27934957e-01 -1.60114631e-01 -2.17642426e-01
4.26682115e-01 1.60139740e-01 5.37225744e-03 -6.52611673e-01
1.52589083e-01 4.65074509e-01 9.63162407e-02 -5.36155343e-01
8.07528794e-01 2.91396558e-01 4.37459886e-01 -2.25304112e-01
-7.50360489e-01 -2.16449827e-01 -5.05938768e-01 -2.21171975e-01
2.81265259e-01 -2.40256041e-01 -5.01866043e-01 1.64655522e-01
-4.87868786e-01 2.44565174e-01 -6.23435676e-01 6.64108932e-01
-7.30670154e-01 3.62031698e-01 3.08368467e-02 -1.19106328e+00
-7.26358771e-01 -7.85791039e-01 1.65370405e-01 2.89120287e-01
-3.06940019e-01 -1.28253186e+00 4.83940452e-01 1.69105202e-01
5.55185139e-01 6.11844361e-01 8.15190434e-01 -1.23527515e+00
-6.83975816e-01 -2.57937461e-01 1.21484771e-01 5.48215628e-01
3.50209594e-01 -2.58639567e-02 -4.42001283e-01 -3.71814162e-01
6.25132501e-01 2.99731016e-01 8.31168950e-01 2.01179549e-01
1.23204303e+00 -3.16661447e-01 7.57835358e-02 1.02866185e+00
1.57786608e+00 7.70737171e-01 6.24469995e-01 9.34395730e-01
2.14192048e-01 1.09684956e+00 1.06308460e+00 1.00960481e+00
-3.47653657e-01 4.75844353e-01 1.30041376e-01 6.66194797e-01
3.74040484e-01 2.88184702e-01 2.35045806e-01 6.33031309e-01
-2.66603768e-01 -1.08850375e-01 -8.27632725e-01 1.79969355e-01
-2.08186579e+00 -1.48007488e+00 3.59018236e-01 2.52091789e+00
4.96565074e-01 2.26612240e-01 4.53133434e-01 4.15261716e-01
5.93486607e-01 -1.16031021e-02 -4.17865038e-01 -3.97341251e-01
-3.32824469e-01 6.56989276e-01 6.62560046e-01 1.57115027e-01
-1.15724850e+00 3.88615310e-01 6.84953403e+00 8.18764389e-01
-6.47336304e-01 -4.55367118e-01 8.34436417e-01 -3.50620359e-01
-7.57556140e-01 -2.29305059e-01 -1.04909813e+00 7.93004870e-01
9.39737916e-01 -8.01712513e-01 3.78397226e-01 1.10679638e+00
-1.66989416e-01 4.91522282e-01 -7.66027808e-01 1.04988325e+00
3.18634987e-01 -1.67319679e+00 3.35143685e-01 9.73012969e-02
9.72038686e-01 -7.00652480e-01 4.84065801e-01 2.08879098e-01
4.36914802e-01 -1.28619599e+00 5.82709610e-01 1.26750684e+00
9.92347449e-02 -1.69576931e+00 1.13458526e+00 -1.01736993e-01
-1.23777902e+00 -5.35157740e-01 -7.46868432e-01 2.02788547e-01
1.22046411e-01 4.72326964e-01 -3.09088290e-01 7.67776370e-01
6.76764965e-01 9.93700445e-01 -6.46423876e-01 1.50526321e+00
-5.68660870e-02 4.21502322e-01 -4.00105081e-02 -2.72083461e-01
-1.54560879e-02 -1.02420068e+00 5.13174534e-01 7.56510198e-01
9.11572635e-01 2.22354770e-01 -2.38637298e-01 5.87570667e-01
1.61169112e-01 2.76388317e-01 -7.33136356e-01 -3.06830674e-01
6.07150972e-01 8.96234334e-01 -6.72819316e-01 -1.41348422e-01
-3.66964042e-01 1.77701533e-01 5.72231822e-02 -7.43972957e-02
-7.34356701e-01 -7.20648825e-01 5.45511842e-01 3.72806862e-02
5.89224577e-01 -1.46394104e-01 -7.99201787e-01 -8.41132462e-01
-5.71859293e-02 -1.00593293e+00 8.72399032e-01 2.01896559e-02
-1.70026731e+00 4.03036088e-01 1.33804694e-01 -1.72299528e+00
-3.14991266e-01 -8.20625484e-01 -1.06264806e+00 9.61221576e-01
-1.38179243e+00 -4.43398327e-01 1.78135648e-01 3.68308842e-01
3.30131292e-01 -1.38020623e+00 5.38679600e-01 1.24540895e-01
-4.14627999e-01 6.49762750e-01 6.83664799e-01 -2.28444859e-01
4.10034299e-01 -1.37616360e+00 1.07647769e-01 5.39026737e-01
2.28608280e-01 6.09884262e-01 3.63799125e-01 -7.88515747e-01
-1.10545611e+00 -1.12997127e+00 5.98403931e-01 -4.44075853e-01
8.95714641e-01 3.26709270e-01 -8.27175438e-01 3.55062872e-01
1.13184527e-01 -6.84338927e-01 1.35261154e+00 -5.87078817e-02
-8.15869588e-03 -4.87235725e-01 -1.37375832e+00 5.41165829e-01
6.96130693e-01 -8.44543353e-02 -9.28359628e-01 2.07836241e-01
4.76871014e-01 1.85471430e-01 -1.18121994e+00 3.05482596e-01
4.36019093e-01 -1.29164815e+00 1.52294087e+00 -4.28130448e-01
-6.53214380e-02 1.94645002e-02 -2.91154772e-01 -1.03595853e+00
-5.76236665e-01 -9.00849223e-01 -4.73969668e-01 1.44255245e+00
2.50685215e-01 -1.11123168e+00 1.13341022e+00 4.60774541e-01
2.13295087e-01 -9.98575866e-01 -8.89416754e-01 -1.18624759e+00
3.48449647e-01 -6.43755719e-02 1.29183376e+00 7.21244037e-01
-1.14276856e-01 -2.38882497e-01 -4.62611377e-01 -3.27986628e-01
1.14928079e+00 7.93420672e-01 3.18209827e-01 -1.49186862e+00
-7.16193974e-01 -1.06516171e+00 -3.15372735e-01 -1.39333785e-01
1.04773819e-01 -9.95668471e-01 -7.69296587e-01 -1.21681488e+00
4.25134711e-02 -6.53042734e-01 -9.61905539e-01 -7.66320676e-02
2.80221626e-02 -1.30372837e-01 2.59876341e-01 5.09631157e-01
-3.49579662e-01 5.71069360e-01 8.82691264e-01 -1.11526638e-01
-3.10290217e-01 6.60804391e-01 -1.06840718e+00 5.92536390e-01
1.06677568e+00 -5.83111703e-01 -5.58026135e-01 2.52612859e-01
6.07483625e-01 4.39110883e-02 -4.72034723e-01 -8.47826123e-01
6.54272810e-02 -5.39552748e-01 4.54711974e-01 -9.62985098e-01
4.98612151e-02 -7.74767637e-01 5.99972427e-01 6.95536256e-01
-3.12850028e-01 5.10294437e-01 -2.38616243e-01 5.04413962e-01
-3.57504576e-01 -1.04607308e+00 6.60619855e-01 -1.53475806e-01
-5.19892931e-01 1.92772076e-01 -7.50963017e-02 -6.41996637e-02
1.34806335e+00 -1.48288190e-01 -1.02158524e-01 -7.82661811e-02
-6.28172219e-01 1.68576658e-01 1.14898659e-01 1.41242743e-01
9.45736349e-01 -1.55004692e+00 -9.87647951e-01 5.88576645e-02
-3.68290037e-01 -6.16421878e-01 -1.47803172e-01 3.73660266e-01
-8.65034699e-01 6.50591910e-01 -5.18541515e-01 2.52921045e-01
-1.16514552e+00 5.19460201e-01 3.39756399e-01 -7.60942519e-01
-5.50718069e-01 8.78472447e-01 -3.11711311e-01 -1.90124929e-01
2.52277195e-01 2.58767754e-01 -1.00115812e+00 6.20097399e-01
7.69831538e-01 1.16405571e+00 -8.62448737e-02 -2.64116615e-01
-1.31115809e-01 7.45903790e-01 -4.05030362e-02 -1.40933350e-01
1.80969083e+00 2.80544668e-01 -2.91432083e-01 4.95003201e-02
8.07756484e-01 9.42483395e-02 -1.15325153e+00 1.03737470e-02
7.74194181e-01 -1.11502659e+00 -2.03372985e-01 -3.86900604e-01
-1.34680963e+00 6.14227295e-01 5.03109932e-01 3.66464287e-01
1.16612065e+00 -3.89529824e-01 5.18835962e-01 5.87529242e-01
7.84150541e-01 -1.26484942e+00 4.91509765e-01 3.67758036e-01
1.25110912e+00 -8.15341473e-01 5.43693960e-01 2.96158865e-02
-8.71015847e-01 1.39220977e+00 1.27815291e-01 -7.91397214e-01
1.11425185e+00 6.72620684e-02 -2.92130917e-01 -6.56156242e-02
-6.27165437e-01 4.85605001e-02 4.54533964e-01 6.76235735e-01
2.16586232e-01 1.19685881e-01 -3.31653088e-01 1.03732765e+00
-6.52313471e-01 -7.71150291e-01 3.97774458e-01 1.09005332e+00
-7.27972031e-01 -1.77113497e+00 -7.14208603e-01 8.73671651e-01
-6.47133291e-01 -4.93324697e-02 -3.06589305e-01 6.40917778e-01
-2.86547333e-01 6.66620016e-01 9.15403292e-02 -3.34343076e-01
3.81158948e-01 -1.91734150e-01 4.26305652e-01 -8.40004146e-01
-1.14402413e+00 -6.37603998e-02 -6.59409678e-03 -3.24914366e-01
-2.23693699e-01 -1.18548584e+00 -9.20470536e-01 -2.85928458e-01
-3.05899590e-01 4.31773752e-01 2.87758142e-01 5.09633183e-01
7.64167309e-02 6.62360847e-01 1.34107161e+00 -5.74946702e-01
-1.33799624e+00 -5.15749633e-01 -1.16424668e+00 1.64959952e-01
-4.76186514e-01 -9.74638045e-01 -6.39395773e-01 -4.63792533e-01] | [4.83065128326416, 4.006330490112305] |
6e1c0022-a77e-427b-8b96-bf33e036b8bb | learning-adaptive-dense-event-stereo-from-the | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Cho_Learning_Adaptive_Dense_Event_Stereo_From_the_Image_Domain_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Cho_Learning_Adaptive_Dense_Event_Stereo_From_the_Image_Domain_CVPR_2023_paper.pdf | Learning Adaptive Dense Event Stereo From the Image Domain | Recently, event-based stereo matching has been studied due to its robustness in poor light conditions. However, existing event-based stereo networks suffer severe performance degradation when domains shift. Unsupervised domain adaptation (UDA) aims at resolving this problem without using the target domain ground-truth. However, traditional UDA still needs the input event data with ground-truth in the source domain, which is more challenging and costly to obtain than image data. To tackle this issue, we propose a novel unsupervised domain Adaptive Dense Event Stereo (ADES), which resolves gaps between the different domains and input modalities. The proposed ADES framework adapts event-based stereo networks from abundant image datasets with ground-truth on the source domain to event datasets without ground-truth on the target domain, which is a more practical setup. First, we propose a self-supervision module that trains the network on the target domain through image reconstruction, while an artifact prediction network trained on the source domain assists in removing intermittent artifacts in the reconstructed image. Secondly, we utilize the feature-level normalization scheme to align the extracted features along the epipolar line. Finally, we present the motion-invariant consistency module to impose the consistent output between the perturbed motion. Our experiments demonstrate that our approach achieves remarkable results in the adaptation ability of event-based stereo matching from the image domain. | ['Kuk-Jin Yoon', 'Jegyeong Cho', 'Hoonhee Cho'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['image-reconstruction', 'stereo-matching-1', 'unsupervised-domain-adaptation'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 4.38368887e-01 -2.67774135e-01 1.60596937e-01 -3.70216191e-01
-4.06026661e-01 -1.64332256e-01 5.73072672e-01 -2.76630193e-01
-5.74649811e-01 8.67429078e-01 1.83860779e-01 4.03930902e-01
-1.21260293e-01 -9.14135516e-01 -9.05570090e-01 -8.57492447e-01
6.11497700e-01 1.29588738e-01 6.55095458e-01 -1.19237326e-01
2.03850403e-01 4.12997752e-01 -1.78454232e+00 1.51816532e-01
1.02820861e+00 8.47634494e-01 5.80861986e-01 1.96176842e-01
7.69417509e-02 7.18059659e-01 -7.09138289e-02 -3.42448726e-02
4.83848691e-01 -5.50223470e-01 -5.71842194e-01 4.09578115e-01
6.65165424e-01 -5.22852778e-01 -7.80454993e-01 1.28207362e+00
5.59188366e-01 3.37282121e-01 3.86187792e-01 -1.36872971e+00
-3.75798404e-01 -1.92299619e-01 -4.55009282e-01 2.42520839e-01
4.59826082e-01 3.16513062e-01 5.56397140e-01 -9.20779705e-01
1.23823166e+00 9.67830896e-01 4.19545084e-01 5.59801996e-01
-1.19218957e+00 -7.41903961e-01 -3.21538397e-03 6.89945996e-01
-1.16838145e+00 -4.95857358e-01 1.14725769e+00 -4.85707968e-01
5.71439862e-01 -1.99601263e-01 7.48299241e-01 1.28541565e+00
4.75745760e-02 5.74674964e-01 1.20617521e+00 -3.32460642e-01
3.39849919e-01 -4.77764495e-02 -1.27665922e-01 2.25635141e-01
5.52345291e-02 5.78560412e-01 -9.55790341e-01 1.85529932e-01
1.08180809e+00 2.98718721e-01 -5.69447577e-01 -6.17284536e-01
-1.49253523e+00 4.60008293e-01 3.88310105e-01 2.26889580e-01
-5.70452392e-01 -3.68597478e-01 2.85812795e-01 3.45037192e-01
3.19507718e-01 6.34697899e-02 -2.71075636e-01 2.38559216e-01
-8.68707299e-01 1.14732526e-01 3.50575149e-01 1.08191466e+00
9.67910767e-01 1.40367940e-01 -2.63115093e-02 8.14712286e-01
8.06044117e-02 5.23709655e-01 5.28483689e-01 -1.13969254e+00
4.26877707e-01 5.34729540e-01 3.14614862e-01 -1.02497911e+00
-2.99989879e-01 -2.60554373e-01 -1.15524566e+00 2.96092540e-01
5.45556903e-01 9.32612047e-02 -7.38179922e-01 1.91655242e+00
6.41727805e-01 5.65330923e-01 1.78101391e-01 1.18670583e+00
7.82803595e-01 6.07880235e-01 -1.43468529e-01 -5.94171405e-01
8.40498090e-01 -6.78076804e-01 -6.95993662e-01 -1.52136073e-01
1.32134661e-01 -7.56785333e-01 9.43578184e-01 2.87289798e-01
-9.15030360e-01 -7.60412872e-01 -1.01076114e+00 -1.31031960e-01
-1.52986616e-01 -1.69962943e-01 2.16571525e-01 -8.62620845e-02
-7.73495615e-01 5.09200335e-01 -6.12232447e-01 -7.27750301e-01
6.78252354e-02 1.44276351e-01 -8.20315719e-01 -3.53290886e-01
-1.23218513e+00 6.93560004e-01 5.35604596e-01 -1.14664301e-01
-7.97275543e-01 -6.79481864e-01 -1.12462342e+00 -2.57774293e-01
3.71474028e-01 -8.93894196e-01 9.17619228e-01 -1.14522552e+00
-1.55477357e+00 1.09866548e+00 -2.58451700e-01 -3.34635347e-01
7.06618607e-01 2.02931836e-02 -5.17142355e-01 1.92580029e-01
4.26821470e-01 6.21084929e-01 8.79776239e-01 -1.22822237e+00
-8.51287484e-01 -2.76479363e-01 -6.32542595e-02 3.70873213e-01
-2.40548074e-01 -2.54383594e-01 -4.83802229e-01 -6.45620406e-01
5.17753720e-01 -6.88585460e-01 -2.40869168e-02 2.77452022e-01
-1.39202461e-01 7.27464929e-02 8.84192944e-01 -5.41747034e-01
8.18128705e-01 -2.39024663e+00 2.05490038e-01 -3.67762446e-02
-6.00133510e-03 -8.59785900e-02 -1.60319746e-01 2.03249916e-01
-2.89085537e-01 -7.62317896e-01 -3.36212605e-01 -2.10771695e-01
-3.73488039e-01 4.47678357e-01 -3.68151397e-01 5.49602509e-01
1.08986616e-01 3.13686520e-01 -1.05562603e+00 -8.05748641e-01
6.44288540e-01 2.63657451e-01 -6.41184628e-01 5.29843509e-01
-8.82159472e-02 1.20130122e+00 -2.02722386e-01 4.01249826e-01
1.01540828e+00 -2.31654093e-01 -5.33311740e-02 -5.44551611e-01
-3.78123462e-01 7.33126178e-02 -1.55327845e+00 2.18048334e+00
-3.48932296e-01 5.37353754e-01 -5.01581170e-02 -1.00936103e+00
7.93138385e-01 2.68633336e-01 8.52872849e-01 -1.15198410e+00
-9.42367967e-03 3.82452577e-01 -2.27033198e-01 -6.56382024e-01
3.35747361e-01 -2.89756477e-01 1.58257857e-01 1.59715265e-01
2.10407034e-01 -2.27783129e-01 1.49542212e-01 4.76641804e-02
9.07862723e-01 4.18589771e-01 3.74410957e-01 2.05875356e-02
7.94922233e-01 1.06312260e-01 1.15016866e+00 4.27413344e-01
-3.00424457e-01 1.10866237e+00 -1.73667565e-01 -5.78826070e-01
-1.32383525e+00 -1.17276132e+00 -2.72677273e-01 5.34465671e-01
7.41272986e-01 6.34640306e-02 -4.12780076e-01 -5.56738138e-01
-1.83882117e-01 4.00321394e-01 -3.59264702e-01 -1.70686468e-01
-5.76578081e-01 -4.63295013e-01 -5.91284735e-03 4.03404802e-01
1.29931426e+00 -1.00247979e+00 -4.42939252e-01 3.29729766e-01
-6.63774967e-01 -1.37518954e+00 -5.11467636e-01 -3.55481952e-02
-8.69948447e-01 -1.20181012e+00 -8.39135647e-01 -7.73061633e-01
7.54284203e-01 5.43880880e-01 8.75421286e-01 -3.51385921e-01
-1.47925233e-02 3.87663007e-01 -1.33548334e-01 -4.58825342e-02
-2.61972159e-01 -4.28305149e-01 1.60294309e-01 4.54121590e-01
2.31453270e-01 -9.60107923e-01 -8.04742217e-01 6.27219856e-01
-1.16074061e+00 4.29372340e-01 2.78275341e-01 1.00816417e+00
8.11652184e-01 5.87522872e-02 2.63422400e-01 -4.71549422e-01
-1.75731272e-01 -3.69707823e-01 -7.45439649e-01 1.19759180e-02
-3.64898801e-01 1.49754900e-02 6.60369635e-01 -4.31119412e-01
-1.45207763e+00 4.30515021e-01 -1.03851280e-03 -7.15515614e-01
-2.53892541e-01 9.31295753e-02 -4.26548243e-01 -8.67957473e-02
7.99006343e-01 5.26766837e-01 -1.41555056e-01 -2.79131025e-01
-8.13690200e-02 4.31838870e-01 1.03677309e+00 -4.72101331e-01
9.24897850e-01 9.96123135e-01 1.82957232e-01 -5.76129556e-01
-8.52571130e-01 -7.57033765e-01 -7.76971817e-01 -3.98610443e-01
9.09462094e-01 -1.11451519e+00 -2.29614347e-01 8.21586609e-01
-1.33333230e+00 -2.18804732e-01 -4.73258734e-01 7.53792465e-01
-7.77060091e-01 6.79264605e-01 -2.96004862e-01 -3.08172613e-01
-2.25591511e-02 -9.05362070e-01 1.07533932e+00 2.90884733e-01
9.06897187e-02 -8.22510600e-01 1.95251048e-01 1.70839250e-01
8.88863876e-02 4.01739806e-01 4.62597847e-01 -1.66488931e-01
-9.13150907e-01 1.18528508e-01 -3.84594023e-01 4.65842813e-01
2.42266312e-01 -3.60397071e-01 -8.84066880e-01 -2.11813837e-01
3.17507625e-01 -1.06193125e-01 6.94542766e-01 4.06851888e-01
8.82647693e-01 1.36562139e-01 -1.19988836e-01 9.64147210e-01
1.60881674e+00 2.39548147e-01 7.92326748e-01 7.52599597e-01
6.17136538e-01 6.33641422e-01 9.31901813e-01 5.11306643e-01
3.80095601e-01 8.43445837e-01 4.62456942e-01 -1.48890600e-01
-4.75373834e-01 -3.89986306e-01 3.64294678e-01 6.13771915e-01
-1.64847195e-01 -9.17200446e-02 -6.68143213e-01 7.40882218e-01
-2.10119605e+00 -1.03732991e+00 -1.90049961e-01 2.50258493e+00
8.37368548e-01 -7.66874626e-02 -2.05288857e-01 1.29177377e-01
9.89022195e-01 9.76844206e-02 -7.85242319e-01 3.33937854e-01
-5.58535159e-01 -1.56381965e-01 4.15784299e-01 3.05941135e-01
-1.03867829e+00 7.11559951e-01 5.17202425e+00 6.04250431e-01
-1.30107594e+00 2.69802153e-01 7.86434785e-02 5.19607924e-02
-1.03710398e-01 1.84423983e-01 -5.40886998e-01 6.69471979e-01
2.31180876e-01 -2.84648269e-01 3.18095326e-01 6.91655993e-01
4.79497224e-01 -2.63169616e-01 -1.25737500e+00 1.41594875e+00
9.74140763e-02 -1.22463357e+00 9.77933854e-02 -3.38260084e-02
1.01526845e+00 -8.09356198e-02 -3.50573719e-01 -8.60063583e-02
1.87566608e-01 -2.90102124e-01 6.60034537e-01 6.95844650e-01
9.44304943e-01 -3.25598150e-01 5.82865655e-01 4.96265858e-01
-1.03176653e+00 9.01199728e-02 -4.31392491e-01 -2.41231583e-02
3.43874365e-01 6.78792179e-01 -5.13448060e-01 8.62994850e-01
8.45946729e-01 1.22654617e+00 -2.23766848e-01 1.11622071e+00
-7.20229745e-02 8.67372826e-02 -3.46676201e-01 9.51272607e-01
-1.89020947e-01 -4.48425055e-01 8.58856022e-01 7.61265278e-01
5.73546946e-01 1.25019968e-01 3.40249576e-02 7.73197949e-01
1.44721288e-02 -1.55076742e-01 -9.52746332e-01 5.95569968e-01
3.45131487e-01 9.20124948e-01 -3.13569993e-01 -3.67012650e-01
-5.81893504e-01 1.44010639e+00 5.86252920e-02 4.87391204e-01
-7.27217257e-01 -7.52582997e-02 4.75789338e-01 1.88116908e-01
3.25843468e-02 -1.58652335e-01 -1.21983863e-01 -1.64368081e+00
2.74973124e-01 -7.32655764e-01 6.40031397e-01 -1.29781568e+00
-1.40633810e+00 2.75934815e-01 1.70415379e-02 -1.92659450e+00
-2.48724878e-01 -3.05790275e-01 -4.41705197e-01 7.00339496e-01
-1.77496648e+00 -1.06856894e+00 -9.14690435e-01 1.25543773e+00
5.95634818e-01 -2.94958632e-02 3.35241228e-01 6.49932206e-01
-4.13082033e-01 2.67020226e-01 1.98799819e-01 -9.77050979e-03
1.24871647e+00 -8.67781401e-01 -1.67033836e-01 1.14546454e+00
-1.82917029e-01 1.94614589e-01 6.94191217e-01 -6.17016733e-01
-1.15803552e+00 -1.25230813e+00 8.22789192e-01 -6.02937378e-02
3.11528027e-01 -3.17139737e-02 -1.04739809e+00 6.45353198e-01
6.41037598e-02 3.44425887e-01 5.20849973e-02 -4.03930306e-01
-3.48887265e-01 -5.32949924e-01 -1.20845771e+00 4.94169623e-01
1.41334963e+00 -6.85718238e-01 -7.18142331e-01 2.99241722e-01
5.60328007e-01 -5.74673235e-01 -8.19018066e-01 5.31260848e-01
3.76678228e-01 -1.26936388e+00 1.01141787e+00 -1.97154641e-01
5.62773526e-01 -7.36485422e-01 -2.34998286e-01 -1.19055748e+00
-2.45760247e-01 -3.53019893e-01 1.96595550e-01 1.23196268e+00
-3.18876714e-01 -7.05115259e-01 6.88238502e-01 3.89377862e-01
-2.44341016e-01 1.48301721e-01 -1.08644950e+00 -9.23426569e-01
-3.21471453e-01 -3.19683403e-01 4.62872386e-01 1.16999829e+00
-4.19819683e-01 1.07854337e-01 -4.96022284e-01 4.74248558e-01
9.73782241e-01 3.19354802e-01 9.20085669e-01 -1.22866333e+00
-2.67209888e-01 9.92287323e-02 -5.99253237e-01 -1.06374002e+00
1.40963525e-01 -6.34543180e-01 9.87995639e-02 -1.35730648e+00
3.09006065e-01 -1.22696146e-01 -1.76895380e-01 2.46075630e-01
-4.19631973e-03 2.72307962e-01 1.49385054e-02 5.97562909e-01
-5.29627264e-01 7.63662517e-01 1.54634571e+00 -7.82971531e-02
-2.25513145e-01 -3.31945926e-01 8.45062605e-04 8.09854031e-01
4.90534842e-01 -4.28330809e-01 -4.15828675e-01 -4.28092182e-01
1.13207407e-01 3.22008282e-01 7.27373600e-01 -1.35800922e+00
5.96312046e-01 -3.98045152e-01 2.81355560e-01 -6.38724625e-01
2.60524929e-01 -1.14170909e+00 5.43933332e-01 1.20245762e-01
-1.77067481e-02 -3.58805656e-01 3.82960588e-02 7.23182261e-01
-6.17400229e-01 2.57236976e-02 1.14506578e+00 -6.85598031e-02
-1.13038290e+00 4.70347524e-01 6.25796393e-02 -7.37789785e-03
1.02091956e+00 -6.19098783e-01 -1.04669474e-01 -2.78676212e-01
-6.94507122e-01 1.13244474e-01 8.73599589e-01 2.45070323e-01
6.81609333e-01 -1.55326569e+00 -6.03138089e-01 5.38440526e-01
5.10017574e-01 3.85986567e-01 6.47655785e-01 9.24620688e-01
-3.04639608e-01 3.47743928e-02 -6.73393607e-01 -9.48730469e-01
-1.11101401e+00 5.36402404e-01 4.42407429e-01 -1.21219559e-02
-7.90825665e-01 3.55870932e-01 6.64888382e-01 -3.49410117e-01
5.00924066e-02 -8.27704594e-02 -6.31211251e-02 -6.35975972e-02
2.07959771e-01 3.41554374e-01 3.67560722e-02 -6.77041590e-01
-1.67491347e-01 8.29439640e-01 2.31704935e-01 -1.24284916e-01
1.27863872e+00 -5.61220467e-01 -7.34048262e-02 3.34506720e-01
1.11611664e+00 -2.51002818e-01 -1.69659090e+00 -7.37184405e-01
-2.78672844e-01 -6.74535871e-01 1.80284362e-02 -4.42392945e-01
-1.02345860e+00 7.97573745e-01 6.92972660e-01 -3.05036366e-01
1.58968711e+00 -3.19701135e-01 8.86675358e-01 2.04408646e-01
5.52067816e-01 -1.25552118e+00 1.08781256e-01 5.39415002e-01
9.47573245e-01 -1.57566178e+00 -3.09914146e-02 -4.78740305e-01
-4.38560039e-01 1.09241843e+00 8.36292088e-01 -1.21202901e-01
4.72019076e-01 -2.34841898e-01 -1.40285850e-01 -6.48097694e-02
-4.54491824e-01 -4.03417915e-01 7.23995492e-02 8.02420974e-01
-1.25625953e-01 -5.05534887e-01 -2.99641371e-01 2.76385278e-01
1.59356847e-01 4.22213465e-01 4.33455586e-01 7.92333364e-01
-9.39346179e-02 -1.09299004e+00 -6.01146877e-01 -1.26773551e-01
3.60900499e-02 7.75747523e-02 -4.86387312e-02 8.45707238e-01
4.75783408e-01 7.98096359e-01 1.48770198e-01 -2.79807001e-01
7.18789101e-01 -5.71560450e-02 5.09402454e-01 -4.76431280e-01
-2.78617069e-02 7.56490529e-02 -1.87274262e-01 -8.15609217e-01
-9.53041077e-01 -7.88470864e-01 -1.23744917e+00 -2.08920568e-01
-8.22978914e-02 -1.78369984e-01 3.01758409e-01 8.49287570e-01
4.87899899e-01 1.47994906e-01 8.22379529e-01 -8.75851691e-01
-8.63797069e-02 -6.83476806e-01 -7.39805877e-01 1.02625346e+00
4.00439799e-01 -8.81285727e-01 -5.65446854e-01 5.85108936e-01] | [9.11130428314209, -2.2925126552581787] |
e47a9eee-0fa4-4732-8568-21cdd199f547 | multimorbidity-content-based-medical-image | 2211.12185 | null | https://arxiv.org/abs/2211.12185v1 | https://arxiv.org/pdf/2211.12185v1.pdf | Multimorbidity Content-Based Medical Image Retrieval Using Proxies | Content-based medical image retrieval is an important diagnostic tool that improves the explainability of computer-aided diagnosis systems and provides decision making support to healthcare professionals. Medical imaging data, such as radiology images, are often multimorbidity; a single sample may have more than one pathology present. As such, image retrieval systems for the medical domain must be designed for the multi-label scenario. In this paper, we propose a novel multi-label metric learning method that can be used for both classification and content-based image retrieval. In this way, our model is able to support diagnosis by predicting the presence of diseases and provide evidence for these predictions by returning samples with similar pathological content to the user. In practice, the retrieved images may also be accompanied by pathology reports, further assisting in the diagnostic process. Our method leverages proxy feature vectors, enabling the efficient learning of a robust feature space in which the distance between feature vectors can be used as a measure of the similarity of those samples. Unlike existing proxy-based methods, training samples are able to assign to multiple proxies that span multiple class labels. This multi-label proxy assignment results in a feature space that encodes the complex relationships between diseases present in medical imaging data. Our method outperforms state-of-the-art image retrieval systems and a set of baseline approaches. We demonstrate the efficacy of our approach to both classification and content-based image retrieval on two multimorbidity radiology datasets. | ['ZongYuan Ge', 'Tom Drummond', 'Mehrtash Harandi', 'Benjamin J. Meyer', 'Yunyan Xing'] | 2022-11-22 | null | null | null | null | ['medical-image-retrieval', 'content-based-image-retrieval', 'medical-image-retrieval'] | ['computer-vision', 'computer-vision', 'medical'] | [ 5.21146953e-01 -5.53930402e-02 -6.12170339e-01 -5.70822835e-01
-1.50176561e+00 -5.04913449e-01 4.63422567e-01 8.37629318e-01
-1.29281864e-01 3.39933723e-01 4.82855499e-01 -2.20484763e-01
-6.57394886e-01 -7.17486024e-01 -2.03723252e-01 -7.30131745e-01
-5.70843779e-02 9.20382679e-01 -1.40853941e-01 2.94709265e-01
1.26792118e-01 6.13651454e-01 -1.51931632e+00 5.85202515e-01
3.86450559e-01 1.23689532e+00 2.51037657e-01 7.16721535e-01
1.35780517e-02 8.47899973e-01 -3.38328660e-01 -9.30945203e-02
1.89166635e-01 -3.69470030e-01 -8.82889330e-01 2.49836177e-01
6.16737008e-01 -5.11662006e-01 -2.57354379e-01 8.51211011e-01
4.97877121e-01 -1.46136761e-01 1.13556600e+00 -9.27272081e-01
-5.91916621e-01 2.30894521e-01 -4.71192449e-01 2.04824507e-01
5.84810853e-01 -2.82925576e-01 1.26636434e+00 -8.83451164e-01
9.15992916e-01 1.14157951e+00 3.38278621e-01 4.22041327e-01
-1.05163467e+00 -3.68138760e-01 -2.41100341e-01 4.34842795e-01
-1.20953369e+00 -2.09205687e-01 5.65538526e-01 -6.26911044e-01
3.49720925e-01 6.07610464e-01 6.20413601e-01 6.17585242e-01
2.06766680e-01 7.98034430e-01 9.77054358e-01 -4.45673883e-01
1.08552620e-01 1.19675145e-01 1.46487564e-01 1.08269083e+00
2.40904108e-01 -1.82075441e-01 -3.97719592e-01 -6.34999156e-01
5.11714637e-01 6.34214282e-01 -2.78014570e-01 -4.69371617e-01
-1.31964314e+00 8.52502406e-01 4.15182471e-01 2.07172602e-01
-6.58483088e-01 9.38562080e-02 3.73725444e-01 2.20121861e-01
3.77333969e-01 5.19008279e-01 -1.80994213e-01 3.48553151e-01
-1.07874954e+00 1.94913805e-01 6.20155275e-01 6.07797444e-01
6.40631199e-01 -8.37802947e-01 -5.84805965e-01 9.42456782e-01
3.71441334e-01 5.64150453e-01 8.50412369e-01 -1.03189743e+00
1.13344915e-01 7.16946125e-01 -1.37248024e-01 -1.06940687e+00
-4.42880511e-01 -3.70167285e-01 -5.69107771e-01 -4.72286679e-02
2.87670176e-02 5.47389507e-01 -8.52377534e-01 1.42682076e+00
5.23796976e-01 1.59156486e-01 -2.34221499e-02 9.55603957e-01
7.55761504e-01 1.77280679e-01 -3.22073661e-02 -1.72532156e-01
1.69665515e+00 -7.55416632e-01 -4.69464421e-01 1.61275133e-01
9.47048962e-01 -7.95955181e-01 8.01012635e-01 1.18723564e-01
-9.56538320e-01 -1.33045807e-01 -8.28237832e-01 8.00977647e-02
-1.93780929e-01 9.09273997e-02 4.66819495e-01 3.75162244e-01
-9.07807887e-01 5.93565941e-01 -6.94473565e-01 -4.18896168e-01
4.15696859e-01 1.22195803e-01 -5.99826813e-01 -8.11902285e-01
-7.25840092e-01 8.71857226e-01 6.02279790e-02 -4.53663677e-01
-7.20154881e-01 -9.18993950e-01 -8.14729393e-01 -9.59394947e-02
2.86857426e-01 -6.98115826e-01 1.11024308e+00 -4.04790282e-01
-4.78929192e-01 1.28221416e+00 -1.70989975e-01 -1.89637154e-01
4.96371448e-01 1.47295713e-01 -4.03829604e-01 8.59344900e-01
4.97218639e-01 5.64872921e-01 8.91393125e-01 -1.15419888e+00
-6.57483339e-01 -3.09813768e-01 -1.80878639e-01 2.38888159e-01
-3.69505376e-01 -1.35189295e-01 -4.23304230e-01 -6.41450226e-01
4.46150303e-01 -9.75388944e-01 -8.88544843e-02 7.58168817e-01
-2.61635154e-01 -1.08836174e-01 6.52208447e-01 -3.98070753e-01
1.01741862e+00 -2.19758582e+00 4.36279625e-02 5.43256164e-01
6.04168594e-01 -1.66120514e-01 -1.95740402e-01 1.96378335e-01
-1.03822008e-01 1.75695583e-01 -1.92639321e-01 -3.20990056e-01
-1.99288487e-01 4.17267203e-01 5.68504333e-02 7.14810312e-01
2.48426273e-01 7.73253441e-01 -1.08986485e+00 -1.03338766e+00
1.16508693e-01 1.70541391e-01 -4.20261502e-01 2.72485226e-01
1.72231898e-01 3.36115927e-01 -5.73471606e-01 1.13439512e+00
1.56689554e-01 -8.86101663e-01 2.37668142e-01 -3.99732232e-01
5.50631106e-01 -7.67178042e-03 -8.53104591e-01 1.65104854e+00
-4.02675062e-01 3.62721503e-01 -1.45253837e-01 -1.03987873e+00
4.53896999e-01 5.32169878e-01 1.29997158e+00 -6.26412570e-01
-4.09676097e-02 3.98129731e-01 -9.48473811e-02 -6.48869157e-01
2.01444566e-01 -1.07313260e-01 2.77272724e-02 9.14192796e-01
-3.27653199e-01 -2.80262232e-02 1.13680698e-01 3.64854932e-01
1.48933041e+00 -4.94770914e-01 4.09308970e-01 -8.56751855e-03
4.41577047e-01 4.90118191e-02 3.55252326e-01 9.33169365e-01
-2.20899224e-01 8.73627245e-01 1.74415737e-01 -2.74081796e-01
-1.00663149e+00 -1.01295960e+00 -5.88631988e-01 7.80446529e-01
7.29264840e-02 -3.70594114e-01 -1.22023664e-01 -8.20982814e-01
3.46910864e-01 -7.59763876e-03 -7.86905169e-01 -3.37740868e-01
-3.14061344e-01 -4.88598108e-01 2.49270320e-01 3.58643740e-01
-1.89020008e-01 -7.35377669e-01 -6.68838382e-01 2.18368500e-01
-3.32000464e-01 -8.43349457e-01 -5.91606915e-01 -7.98661485e-02
-9.03580964e-01 -1.61764586e+00 -9.24971819e-01 -6.20122671e-01
8.96598876e-01 4.84028995e-01 1.15288758e+00 7.03587651e-01
-1.03232062e+00 1.08141065e+00 -5.99619269e-01 -4.60735075e-02
-7.02011585e-01 -2.30584368e-01 -1.18806802e-01 4.03434299e-02
1.30323350e-01 -2.08175004e-01 -1.04683948e+00 1.65172055e-01
-1.35021377e+00 -1.34533748e-01 8.75104308e-01 1.12167406e+00
9.55488026e-01 -3.51413816e-01 3.42502087e-01 -1.05670857e+00
5.14384031e-01 -8.30160916e-01 -2.52664997e-03 6.23490334e-01
-1.07835841e+00 2.77402163e-01 7.30272681e-02 -3.72501671e-01
-3.40443522e-01 -1.12074360e-01 2.56538242e-01 -4.84166741e-01
4.15742025e-02 8.89399886e-01 4.87036020e-01 -1.32764310e-01
5.97458184e-01 2.90561058e-02 2.24122465e-01 -4.49026704e-01
3.41828167e-01 7.86280513e-01 4.37353015e-01 -5.17154038e-01
3.59170616e-01 5.70634305e-01 5.34205139e-01 -4.02672559e-01
-8.10002744e-01 -1.01117432e+00 -3.74880195e-01 -2.61272877e-01
5.97148776e-01 -7.91496634e-01 -2.28367165e-01 -1.60895288e-01
-7.53373861e-01 2.89872229e-01 -2.59538412e-01 5.43585420e-01
-5.69479346e-01 3.58196169e-01 -6.00441992e-01 -3.93572330e-01
-3.68179470e-01 -1.34644020e+00 1.45710921e+00 -3.28305304e-01
-5.23238242e-01 -1.00684190e+00 -5.88557795e-02 3.65946680e-01
2.11522281e-01 2.54669964e-01 1.46086013e+00 -7.76560426e-01
-4.88441408e-01 -6.33409441e-01 -4.06527579e-01 -5.92759214e-02
3.87800843e-01 -3.27814251e-01 -6.20712578e-01 -3.86859596e-01
-9.21795815e-02 -4.36236680e-01 8.34946394e-01 3.94968480e-01
1.26793766e+00 -2.19056934e-01 -5.93331516e-01 3.21311891e-01
1.38136995e+00 -9.53639895e-02 2.15988085e-01 2.52377540e-01
5.34991384e-01 6.77346170e-01 8.20147455e-01 5.72407603e-01
4.06598032e-01 6.26468420e-01 4.15713459e-01 -2.59678602e-01
-3.84063572e-01 7.18075037e-02 -3.76212120e-01 6.32216036e-01
4.42288518e-01 -1.90912634e-02 -1.09348702e+00 5.87821782e-01
-1.88934219e+00 -6.47206187e-01 1.13172501e-01 2.13409710e+00
9.27076578e-01 -3.88899058e-01 -1.99734762e-01 1.46174937e-01
5.69304824e-01 -1.72257811e-01 -6.39725804e-01 2.60825694e-01
1.08219109e-01 2.15484351e-01 3.66353095e-01 1.45973518e-01
-1.02246308e+00 1.43974349e-01 6.57790422e+00 4.99056906e-01
-1.16151285e+00 2.12227240e-01 6.38573825e-01 -2.48657748e-01
-4.46378201e-01 -3.43303651e-01 -2.96134025e-01 2.95838445e-01
6.71549499e-01 -1.78603888e-01 1.69530790e-02 7.29063451e-01
-7.27038160e-02 -2.45646507e-01 -1.47688365e+00 1.17951655e+00
3.01781952e-01 -1.40896130e+00 3.95304173e-01 2.94234663e-01
5.36987007e-01 -4.43373583e-02 3.69159937e-01 -2.64888138e-01
-7.44669407e-04 -9.24393415e-01 4.52927381e-01 7.80415714e-01
1.14805794e+00 -2.79592156e-01 6.26292825e-01 -6.63282303e-03
-9.22749937e-01 -1.58106178e-01 -1.37176156e-01 6.48907304e-01
-1.12434849e-01 6.94902182e-01 -1.19071817e+00 5.13948321e-01
4.00393486e-01 8.67266417e-01 -6.41679645e-01 1.19442797e+00
2.41983294e-01 3.09130639e-01 3.19285430e-02 2.80482650e-01
8.97467658e-02 1.95761532e-01 3.91941786e-01 1.06051660e+00
5.13668180e-01 1.89283326e-01 6.07420325e-01 3.75748426e-01
-1.94279119e-01 3.56204182e-01 -6.45590007e-01 3.48982476e-02
6.01931214e-01 1.40837085e+00 -7.63906121e-01 -5.97128391e-01
-3.97988111e-01 8.39226186e-01 1.62448123e-01 7.95366317e-02
-4.11177367e-01 8.90309587e-02 5.43241203e-01 1.75411940e-01
-3.76426033e-03 1.19382188e-01 -3.02911792e-02 -1.02248394e+00
8.14042427e-03 -8.85357022e-01 7.17359304e-01 -6.66569352e-01
-1.54299068e+00 3.63426477e-01 -1.87575981e-01 -1.52117026e+00
-5.74685872e-01 -6.88759625e-01 4.14119847e-02 7.14419961e-01
-1.74276245e+00 -1.13861573e+00 -3.49013835e-01 5.68935573e-01
1.70582443e-01 -2.12007627e-01 1.24684405e+00 4.39403027e-01
-1.07876984e-02 4.98026043e-01 3.76919895e-01 1.07711688e-01
1.15931594e+00 -1.26488817e+00 -4.52286810e-01 2.97433287e-01
1.46848470e-01 5.49856901e-01 3.54888171e-01 -5.55510521e-01
-1.43882799e+00 -1.10450292e+00 5.68543077e-01 -3.50171268e-01
5.53287327e-01 3.52571905e-01 -8.29301178e-01 2.54046887e-01
-4.66438353e-01 4.41010118e-01 1.41232061e+00 -5.91125935e-02
-6.39037430e-01 -1.82377174e-01 -1.41704988e+00 3.19915503e-01
7.11972237e-01 -8.80323827e-01 -3.01898897e-01 6.56278968e-01
3.42393607e-01 -2.17496678e-01 -1.25051844e+00 3.31725091e-01
8.43111634e-01 -4.94779646e-01 1.05199528e+00 -7.94598103e-01
7.73683846e-01 -2.11634375e-02 -5.66290796e-01 -9.82705593e-01
-4.47592318e-01 2.00415313e-01 -1.30699351e-01 5.95656276e-01
2.84373671e-01 -3.90504211e-01 5.39313436e-01 8.71560693e-01
5.73705994e-02 -9.76265132e-01 -8.63984942e-01 -5.37596405e-01
-1.28834724e-01 -3.37177575e-01 4.78609860e-01 9.03450191e-01
-6.53828532e-02 -2.68385857e-01 -6.08043978e-03 1.37396529e-01
7.14167297e-01 3.83064002e-01 1.70941532e-01 -1.43519413e+00
-3.78086507e-01 -4.96627033e-01 -6.97639942e-01 -4.40331578e-01
1.15305185e-01 -1.39512646e+00 -7.30399182e-03 -1.69397020e+00
6.81934297e-01 -9.00520384e-01 -7.32972562e-01 5.91624320e-01
-2.94239521e-01 6.44111395e-01 -2.37983745e-02 8.05834949e-01
-6.02532029e-01 -1.58271924e-01 1.14004886e+00 -4.91745263e-01
3.44762355e-01 -2.23940194e-01 -7.33626902e-01 4.25510913e-01
3.67830306e-01 -8.03204536e-01 -3.30038518e-01 -2.19620571e-01
2.04125956e-01 3.94529790e-01 5.30200958e-01 -7.99773395e-01
2.50347465e-01 -5.02220690e-02 3.89668852e-01 -4.25886482e-01
3.42877328e-01 -9.36959922e-01 2.28099395e-02 7.08872795e-01
-6.52023017e-01 8.35991744e-03 -3.56182963e-01 8.09006214e-01
-3.27704728e-01 -3.75873446e-01 5.94011426e-01 -4.86494452e-01
-4.43849802e-01 6.39925003e-01 -3.10403496e-01 -7.31315091e-02
8.83870006e-01 8.26312080e-02 -2.18816981e-01 -3.31984043e-01
-6.34541273e-01 2.25189418e-01 3.58784705e-01 2.80068040e-01
8.85699034e-01 -1.47307491e+00 -8.59093487e-01 -9.93426368e-02
6.93505943e-01 -2.36120865e-01 9.08458084e-02 9.63100433e-01
-4.38449681e-01 3.06242943e-01 6.93489760e-02 -9.30967987e-01
-1.59865141e+00 4.63247478e-01 1.73082039e-01 -4.47912604e-01
-5.57888687e-01 4.88155693e-01 8.68279189e-02 -1.37178540e-01
6.85370266e-02 -6.19119108e-02 -5.49390018e-02 3.63837838e-01
7.95017719e-01 1.01434581e-01 3.53083968e-01 -5.64180851e-01
-3.17838103e-01 5.31497359e-01 -5.87911487e-01 -2.56610274e-01
1.32941318e+00 -1.11274257e-01 -3.23136210e-01 3.86521995e-01
1.54206014e+00 -2.64375538e-01 -5.56518316e-01 -7.00693071e-01
5.50751910e-02 -5.72222352e-01 1.95482582e-01 -9.12092209e-01
-9.72057581e-01 6.07137024e-01 9.05386448e-01 1.25725612e-01
1.16576135e+00 4.43892479e-01 6.34777844e-01 4.25952613e-01
3.38106424e-01 -7.22090065e-01 4.09987241e-01 -1.39613062e-01
7.77695179e-01 -1.50424862e+00 3.03117752e-01 -2.85229802e-01
-3.76273215e-01 1.22126186e+00 -4.96576317e-02 1.58453733e-01
6.30449533e-01 2.09512357e-02 3.36937666e-01 -6.14021659e-01
-6.46338940e-01 -5.61265275e-02 6.61069691e-01 4.11717951e-01
4.61957514e-01 3.35631758e-01 -4.23816353e-01 1.28131390e-01
2.46132612e-01 -2.29080096e-01 2.59618253e-01 1.21251535e+00
-4.02174890e-01 -1.40246475e+00 -4.96268749e-01 1.08498883e+00
-5.91718554e-01 -1.36186853e-01 -1.11861721e-01 2.33404443e-01
-7.50524774e-02 8.11755657e-01 -8.66932273e-02 -1.71190828e-01
4.98960316e-02 3.79957333e-02 4.62264121e-01 -8.98857892e-01
-2.68830776e-01 -7.24364631e-03 -2.60684520e-01 -6.55097306e-01
-5.54727554e-01 -8.39924395e-01 -1.24828160e+00 2.08961725e-01
-2.17823237e-01 6.66237399e-02 7.58508682e-01 9.56996202e-01
4.03628230e-01 5.16739309e-01 8.00589621e-01 -3.65944713e-01
-5.72481692e-01 -6.24792039e-01 -6.69652879e-01 9.43553865e-01
7.18445063e-01 -7.33347476e-01 7.98888206e-02 6.25086054e-02] | [14.451086044311523, -1.630953073501587] |
c489686d-3825-4f07-8cf8-6cca3c4c434b | smaug-sparse-masked-autoencoder-for-efficient | 2211.11446 | null | https://arxiv.org/abs/2211.11446v3 | https://arxiv.org/pdf/2211.11446v3.pdf | SMAUG: Sparse Masked Autoencoder for Efficient Video-Language Pre-training | Video-language pre-training is crucial for learning powerful multi-modal representation. However, it typically requires a massive amount of computation. In this paper, we develop SMAUG, an efficient pre-training framework for video-language models. The foundation component in SMAUG is masked autoencoders. Different from prior works which only mask textual inputs, our masking strategy considers both visual and textual modalities, providing a better cross-modal alignment and saving more pre-training costs. On top of that, we introduce a space-time token sparsification module, which leverages context information to further select only "important" spatial regions and temporal frames for pre-training. Coupling all these designs allows our method to enjoy both competitive performances on text-to-video retrieval and video question answering tasks, and much less pre-training costs by 1.9X or more. For example, our SMAUG only needs about 50 NVIDIA A6000 GPU hours for pre-training to attain competitive performances on these two video-language tasks across six popular benchmarks. | ['Cihang Xie', 'Alan Yuille', 'Huiyu Wang', 'Chen Wei', 'Yuanze Lin'] | 2022-11-21 | null | null | null | null | ['video-question-answering'] | ['computer-vision'] | [-4.16399427e-02 -4.82476771e-01 -4.69570875e-01 -3.41587216e-01
-1.25814950e+00 -3.98429662e-01 4.88747507e-01 1.34741768e-01
-7.36059129e-01 1.55862868e-01 2.74661392e-01 -5.35248876e-01
4.87705797e-01 -5.63305259e-01 -9.01366353e-01 -5.92324674e-01
1.31909475e-01 -2.16369703e-02 2.96160460e-01 6.40135854e-02
-2.49133911e-02 5.81609979e-02 -1.55398655e+00 9.62425172e-01
3.07294548e-01 1.05802917e+00 4.16478902e-01 8.05183291e-01
-2.94018358e-01 1.16922140e+00 -2.62324542e-01 -2.63083339e-01
1.66589946e-01 -7.69208968e-02 -7.50160813e-01 2.08180740e-01
6.55825913e-01 -6.09826386e-01 -7.71426082e-01 8.04480493e-01
5.81558168e-01 3.33521008e-01 2.19249174e-01 -9.18918252e-01
-4.50333744e-01 7.08881438e-01 -8.56006503e-01 3.28904122e-01
2.41149008e-01 1.97983399e-01 1.21756172e+00 -1.12325966e+00
2.84531206e-01 1.24420226e+00 3.77080530e-01 5.44951916e-01
-8.39552879e-01 -5.08013546e-01 3.79436791e-01 4.27439421e-01
-1.46877909e+00 -6.69962466e-01 4.66658741e-01 -2.72742391e-01
1.16803598e+00 2.86674738e-01 4.98758703e-01 1.07698178e+00
-2.02222075e-02 1.20439923e+00 7.34896958e-01 -3.91095787e-01
6.07291088e-02 -1.13542117e-01 5.66716567e-02 7.43117452e-01
-1.72679007e-01 -3.30220431e-01 -6.33114934e-01 -1.21031389e-01
7.76616216e-01 2.42045373e-01 -7.63976425e-02 -4.46235724e-02
-1.24965620e+00 7.51378238e-01 1.99550986e-01 2.50307262e-01
-2.13894144e-01 4.60992664e-01 8.25611353e-01 2.22535118e-01
3.62676382e-01 -6.76959008e-02 -4.37678427e-01 -2.50384837e-01
-1.11095500e+00 -8.42899010e-02 2.86695451e-01 9.04840171e-01
1.00499821e+00 2.19283521e-01 -3.75941902e-01 9.21464145e-01
1.69274688e-01 4.63191122e-01 5.57528853e-01 -8.31914663e-01
7.08667696e-01 4.19283032e-01 -2.47847497e-01 -9.98977065e-01
-2.19782203e-01 -1.63328320e-01 -9.40081835e-01 -3.95635843e-01
1.70603812e-01 -1.55767858e-01 -1.01147723e+00 1.59027910e+00
3.14443439e-01 5.86156666e-01 -1.18193068e-01 1.04184496e+00
9.27029610e-01 1.06744039e+00 1.38741717e-01 -5.85339963e-02
1.66195524e+00 -1.46808338e+00 -5.78031898e-01 -3.91399801e-01
8.81574810e-01 -8.16486597e-01 1.49185586e+00 2.31083587e-01
-1.26150990e+00 -6.10225737e-01 -9.33813035e-01 -5.84509730e-01
-1.52899101e-01 2.59498060e-01 7.88944185e-01 2.65326828e-01
-1.04416108e+00 2.65535742e-01 -1.04099321e+00 1.02179432e-02
3.04266959e-01 2.33971357e-01 -2.82626748e-01 -3.41073602e-01
-9.91821170e-01 4.07844722e-01 3.62572581e-01 -1.26162842e-01
-9.52457547e-01 -7.76514292e-01 -1.07551944e+00 2.33590245e-01
6.93888247e-01 -6.39314830e-01 1.29027331e+00 -1.09398925e+00
-1.48459780e+00 8.28345895e-01 -5.86855888e-01 -5.52687168e-01
3.54477465e-01 -5.10103881e-01 -3.24063867e-01 5.49060345e-01
-5.84857836e-02 8.70631099e-01 1.25635970e+00 -8.08299065e-01
-5.93150198e-01 -9.83311087e-02 1.45864129e-01 3.82811964e-01
-7.51830339e-01 3.65307152e-01 -1.42903268e+00 -9.75739539e-01
-2.42985010e-01 -8.54971588e-01 -2.64716864e-01 -1.64200678e-01
-1.66658655e-01 -2.35815451e-01 9.49939609e-01 -6.39007866e-01
1.29756558e+00 -2.35436010e+00 1.01091996e-01 7.54492804e-02
3.80102903e-01 3.18892807e-01 -4.79747385e-01 2.40322948e-01
-8.44601393e-02 -5.61525822e-02 1.10460661e-01 -6.55124366e-01
-7.53018260e-02 3.00394416e-01 -4.31742281e-01 4.08611804e-01
1.21488541e-01 1.01988864e+00 -7.13648796e-01 -6.34104311e-01
3.34964812e-01 5.67795098e-01 -8.57710481e-01 2.93956369e-01
-3.38243842e-01 3.33645605e-02 -2.46204451e-01 6.01574242e-01
5.88939965e-01 -6.02669895e-01 2.31592759e-01 -2.51684219e-01
3.33532295e-03 3.35413307e-01 -9.79454160e-01 2.01608300e+00
-7.05393672e-01 8.49067807e-01 2.22424194e-01 -1.10709023e+00
2.92831302e-01 2.99968511e-01 5.74501514e-01 -8.56141269e-01
2.40137547e-01 -1.38086602e-02 -4.78116810e-01 -6.12156749e-01
7.34786987e-01 2.55303681e-01 -5.15827313e-02 4.84955698e-01
4.84343879e-02 3.21725816e-01 2.99300492e-01 4.98142272e-01
9.88684177e-01 -7.34333247e-02 5.16281314e-02 -9.83495340e-02
6.15618646e-01 -2.42559791e-01 4.49699163e-01 6.05880439e-01
1.46795854e-01 5.65100133e-01 3.98437351e-01 -5.67700028e-01
-9.86452997e-01 -7.50817835e-01 3.05790603e-01 1.63468087e+00
1.42547563e-01 -9.32027400e-01 -6.51845694e-01 -5.16979635e-01
-1.74736992e-01 3.10643822e-01 -3.81646067e-01 -1.79308205e-04
-8.43226731e-01 -6.41917825e-01 5.79895735e-01 6.54885054e-01
3.18310529e-01 -7.79293120e-01 -5.01218498e-01 2.72723734e-02
-2.79608667e-01 -1.57133889e+00 -8.23641241e-01 -3.27114053e-02
-6.94482863e-01 -8.23903382e-01 -7.85387158e-01 -9.24328327e-01
6.37769461e-01 8.93687546e-01 1.17719102e+00 3.66401523e-01
-2.67123193e-01 3.20665926e-01 -5.31469047e-01 2.11533234e-02
9.85925272e-02 1.72276735e-01 -6.85587674e-02 1.27090663e-01
3.13015848e-01 -5.14448166e-01 -7.65919209e-01 1.81723580e-01
-1.12109983e+00 3.25057983e-01 6.18355095e-01 8.88877809e-01
7.62011528e-01 -1.67398870e-01 8.25884640e-02 -6.13848567e-01
1.63067400e-01 -4.91270959e-01 -5.89424193e-01 2.48997897e-01
-8.38164017e-02 -1.00906663e-01 8.02393794e-01 -7.11564958e-01
-7.64179468e-01 1.25121260e-02 -2.20878661e-01 -9.07401681e-01
1.36053279e-01 6.02698624e-01 -4.93395561e-03 5.12018725e-02
3.40119779e-01 3.10796857e-01 -1.73873186e-01 -5.24225891e-01
4.26167637e-01 4.90451872e-01 6.50909841e-01 -7.32989490e-01
6.71010196e-01 6.57850146e-01 -4.16157454e-01 -1.10252368e+00
-8.29477966e-01 -7.85623729e-01 -2.98633248e-01 8.92718136e-02
8.85097027e-01 -1.44777381e+00 -7.93824553e-01 3.04220021e-01
-9.76042390e-01 -5.25095403e-01 2.47616507e-02 4.41956550e-01
-1.81509048e-01 7.46256828e-01 -8.88022661e-01 -4.39514726e-01
-5.21531284e-01 -1.23052680e+00 1.19344091e+00 -2.28044577e-02
2.18776807e-01 -7.25991607e-01 -2.60102600e-01 6.52773321e-01
3.83406341e-01 -3.44019949e-01 6.81038320e-01 -3.50565195e-01
-6.73679113e-01 -8.10199752e-02 -5.39533138e-01 3.26295525e-01
-1.01533785e-01 -9.04953703e-02 -9.15410757e-01 -6.89621747e-01
-3.66348065e-02 -5.41600764e-01 1.19758165e+00 3.48639727e-01
1.88057804e+00 -3.79205316e-01 -1.79750904e-01 8.40048075e-01
1.35870337e+00 -1.19997233e-01 6.12977684e-01 2.35854849e-01
1.08940494e+00 2.31834128e-01 5.55290341e-01 5.61116278e-01
6.21119380e-01 8.00377905e-01 1.92485094e-01 -3.86867136e-01
-1.15644366e-01 -1.55691072e-01 6.51413739e-01 1.24255419e+00
2.04168111e-01 -2.62011319e-01 -8.93186152e-01 6.59742117e-01
-1.97660768e+00 -1.03181267e+00 1.47632480e-01 2.06759763e+00
8.50691259e-01 -1.30563736e-01 2.33174637e-01 -1.68546394e-01
5.18045068e-01 4.93472308e-01 -4.56593037e-01 -9.61620137e-02
-7.66877234e-02 3.19138736e-01 4.37143356e-01 5.31913519e-01
-1.31608772e+00 1.13806140e+00 5.66474962e+00 1.19327843e+00
-1.43761587e+00 2.41952956e-01 7.22633243e-01 -6.56080425e-01
-3.38522613e-01 1.63174067e-02 -7.50316262e-01 3.72519732e-01
1.06504786e+00 8.69460925e-02 3.30242664e-01 8.51621509e-01
2.13721082e-01 -1.24353446e-01 -9.02932048e-01 1.41152179e+00
2.81133115e-01 -1.62490034e+00 2.54739702e-01 -2.54433423e-01
6.24980628e-01 3.73442233e-01 2.17879806e-02 4.86377150e-01
7.55794123e-02 -1.02221322e+00 7.35790730e-01 -8.87778327e-02
9.10967350e-01 -7.85199046e-01 4.81131345e-01 1.85278267e-01
-1.39970803e+00 -1.61909144e-02 -4.44272518e-01 6.94501586e-03
2.87791193e-01 6.75703883e-01 -4.53307241e-01 5.12888491e-01
8.82763386e-01 7.02306628e-01 -4.53159869e-01 6.87645376e-01
-7.06986059e-03 6.65065408e-01 -4.36089426e-01 1.52767599e-01
4.91059512e-01 1.10140584e-01 1.41216993e-01 1.60732961e+00
2.67084688e-01 1.09474778e-01 5.14662445e-01 9.71813053e-02
-3.15439939e-01 2.13500872e-01 -2.31575757e-01 -4.98638749e-02
4.03769642e-01 1.25572228e+00 -5.99650741e-01 -6.74400449e-01
-9.04494643e-01 1.17865062e+00 2.79317856e-01 3.79166067e-01
-1.23481178e+00 -2.37561926e-01 6.95243955e-01 5.45163937e-02
5.82695484e-01 -4.86544162e-01 -4.55006352e-03 -1.83270144e+00
1.83814794e-01 -1.16570354e+00 4.99551445e-01 -5.78670800e-01
-9.33608532e-01 5.12076259e-01 -1.27817214e-01 -1.07530797e+00
-1.54445529e-01 -5.90457857e-01 -2.47059539e-01 4.62277681e-01
-1.74306011e+00 -1.36464500e+00 -3.99373978e-01 9.65655267e-01
8.32497478e-01 -1.26844853e-01 5.57085812e-01 8.39380205e-01
-8.35176408e-01 8.75444651e-01 -8.04706067e-02 2.76701838e-01
8.59512210e-01 -9.23330963e-01 4.80351359e-01 9.46201205e-01
4.28471476e-01 6.68227851e-01 3.01102549e-01 -2.71250188e-01
-1.78251922e+00 -1.21569109e+00 6.33907616e-01 -1.79261521e-01
7.78545141e-01 -4.45235044e-01 -9.28950429e-01 8.08821142e-01
4.77944076e-01 3.00168097e-01 7.13618577e-01 7.07371384e-02
-6.99980617e-01 -9.28923041e-02 -3.96169990e-01 8.91570628e-01
8.50621223e-01 -9.75740314e-01 -3.42818707e-01 4.62598830e-01
8.59810889e-01 -6.88648582e-01 -6.94873333e-01 3.11451793e-01
3.76405895e-01 -7.84748018e-01 1.22695374e+00 -5.50452590e-01
5.93339384e-01 -3.57801467e-01 -3.55646253e-01 -7.29107618e-01
-1.60628378e-01 -8.82363617e-01 -4.05329227e-01 1.13743806e+00
2.65621156e-01 -1.07246526e-01 8.42571676e-01 1.72404930e-01
-1.83090985e-01 -9.31910992e-01 -7.60694325e-01 -5.52438498e-01
-4.49363329e-02 -7.91390717e-01 2.86933750e-01 9.96619821e-01
-1.24820963e-01 4.78973597e-01 -7.81898558e-01 9.43767428e-02
2.89208293e-01 1.30900174e-01 9.13360357e-01 -4.40041870e-01
-4.98506516e-01 -3.59448850e-01 -4.51061726e-02 -1.67720151e+00
3.14360142e-01 -7.94118762e-01 -5.46927080e-02 -1.17043984e+00
4.18497473e-01 -1.99947864e-01 -3.29693735e-01 7.88574219e-01
-3.05981696e-01 3.98974448e-01 2.53839612e-01 1.95151493e-01
-1.00795794e+00 5.61507761e-01 9.45439041e-01 -1.55544192e-01
-8.74185562e-02 -4.84438837e-01 -5.10352433e-01 8.47031116e-01
6.32826030e-01 -3.04612815e-01 -4.89420831e-01 -1.14775825e+00
2.63892055e-01 7.20541850e-02 2.88320512e-01 -8.20120037e-01
2.14169040e-01 -2.00693130e-01 2.66885877e-01 -7.62054920e-01
6.52413666e-01 -6.08044386e-01 -2.98521698e-01 1.55168250e-01
-3.28825682e-01 4.23467308e-01 5.33326685e-01 3.76607895e-01
-4.67244565e-01 5.35754971e-02 6.68360889e-01 -1.04720429e-01
-1.00991225e+00 4.98126239e-01 -5.25112212e-01 1.34337440e-01
8.96856368e-01 2.60964762e-02 -2.51839340e-01 -4.40401047e-01
-3.51152599e-01 4.18183893e-01 4.44000900e-01 4.76680726e-01
7.28630424e-01 -1.31006455e+00 -5.38645923e-01 1.12762630e-01
-1.63235143e-02 1.19818613e-01 7.38970637e-01 9.09427643e-01
-6.55153394e-01 5.67845225e-01 1.62216827e-01 -6.95243537e-01
-1.54154849e+00 6.94070637e-01 -5.53041175e-02 -2.90731847e-01
-7.97414362e-01 1.03550804e+00 3.66900891e-01 9.68944281e-02
4.24134225e-01 -3.64201963e-01 3.01411878e-02 1.05600022e-01
9.15166914e-01 5.47112077e-02 5.76870982e-03 -5.75554252e-01
-3.52853060e-01 5.01201868e-01 -4.70492750e-01 -2.09567755e-01
1.24964571e+00 -1.61528990e-01 -1.14189729e-01 2.40307525e-01
1.52832091e+00 2.16432676e-01 -1.21144474e+00 -4.84601796e-01
-1.29373446e-01 -4.64740545e-01 2.94067532e-01 -1.33206591e-01
-1.31712461e+00 1.02962708e+00 2.30716243e-01 -7.29140267e-02
1.29913080e+00 -5.57659976e-02 1.30173850e+00 5.68247259e-01
2.91149057e-02 -1.17357659e+00 4.92519796e-01 6.68842554e-01
6.28966272e-01 -1.21539533e+00 1.15560241e-01 -3.11318725e-01
-7.03341842e-01 8.99150372e-01 6.59769833e-01 -1.57587573e-01
4.54794914e-01 4.02966470e-01 9.73245725e-02 -1.76653732e-02
-1.08228099e+00 -1.04713976e-01 4.96985704e-01 1.09823518e-01
4.39720988e-01 -1.84450045e-01 -1.35593237e-02 3.87122035e-01
1.14656091e-01 -9.06579122e-02 2.30442509e-01 8.72544169e-01
-2.30457738e-01 -1.18815327e+00 -1.74803272e-01 3.06247830e-01
-7.67923534e-01 -4.59429145e-01 1.80304766e-01 6.24629736e-01
-2.04336077e-01 8.28968406e-01 3.05426389e-01 -5.26370585e-01
-1.94111671e-02 -1.58641100e-01 3.40968728e-01 -5.91590166e-01
-5.71300089e-01 4.87150222e-01 -3.27157602e-02 -8.90542507e-01
-5.12563467e-01 -3.93962324e-01 -1.21813643e+00 -5.63169420e-01
-1.64119318e-01 6.27075955e-02 3.83132249e-01 1.02652681e+00
6.86455369e-01 4.01338726e-01 4.97165442e-01 -8.86885464e-01
-2.53808141e-01 -6.24892056e-01 2.92798802e-02 3.23225647e-01
3.65018606e-01 -3.43893021e-01 -1.36708364e-01 1.16478167e-01] | [9.99386215209961, 0.9151372313499451] |
9e73dc83-c1cc-4083-8a4b-a7d59f795cfe | multi-speaker-and-wide-band-simulated | 2211.06750 | null | https://arxiv.org/abs/2211.06750v2 | https://arxiv.org/pdf/2211.06750v2.pdf | Multi-Speaker and Wide-Band Simulated Conversations as Training Data for End-to-End Neural Diarization | End-to-end diarization presents an attractive alternative to standard cascaded diarization systems because a single system can handle all aspects of the task at once. Many flavors of end-to-end models have been proposed but all of them require (so far non-existing) large amounts of annotated data for training. The compromise solution consists in generating synthetic data and the recently proposed simulated conversations (SC) have shown remarkable improvements over the original simulated mixtures (SM). In this work, we create SC with multiple speakers per conversation and show that they allow for substantially better performance than SM, also reducing the dependence on a fine-tuning stage. We also create SC with wide-band public audio sources and present an analysis on several evaluation sets. Together with this publication, we release the recipes for generating such data and models trained on public sets as well as the implementation to efficiently handle multiple speakers per conversation and an auxiliary voice activity detection loss. | ['Lukáš Burget', 'Alicia Lozano-Diez', 'Mireia Diez', 'Federico Landini'] | 2022-11-12 | null | null | null | null | ['activity-detection'] | ['computer-vision'] | [-5.30452058e-02 3.57569128e-01 2.60350615e-01 -6.02162302e-01
-1.53294361e+00 -6.47358000e-01 8.92971218e-01 -4.68787402e-01
-3.60999078e-01 7.02156484e-01 5.62742710e-01 -1.40241936e-01
2.57602721e-01 -1.25145644e-01 -3.61863911e-01 -8.05787861e-01
1.44141288e-02 9.15149152e-01 2.15088278e-01 -3.44820857e-01
-4.01547074e-01 2.19037846e-01 -1.46099222e+00 5.35890400e-01
4.15266693e-01 7.29508877e-01 1.24789126e-01 1.23052621e+00
1.23259924e-01 3.35344851e-01 -1.15931392e+00 -5.75772762e-01
2.14313239e-01 -8.27086210e-01 -4.87443894e-01 2.61703551e-01
5.72041154e-01 -1.96231723e-01 -2.86628813e-01 4.76197422e-01
1.29513216e+00 1.89705983e-01 6.63022518e-01 -1.31985617e+00
-1.07295485e-02 9.75116372e-01 -1.00577451e-01 1.13460913e-01
2.24805817e-01 1.12902820e-01 8.10710788e-01 -9.23094928e-01
3.69144469e-01 1.55660701e+00 5.05208671e-01 9.56285000e-01
-1.24317598e+00 -7.11092830e-01 2.58138124e-02 -1.99641332e-01
-1.08561003e+00 -1.09856629e+00 6.68837726e-01 -1.81034565e-01
9.98749316e-01 4.99799788e-01 5.00280976e-01 1.58189487e+00
-5.88463783e-01 1.06317067e+00 7.04463661e-01 -5.84788382e-01
3.00593883e-01 4.76722181e-01 -3.41649144e-03 7.80965239e-02
-3.70800763e-01 -1.45753352e-02 -4.99535233e-01 -3.29547614e-01
3.09634835e-01 -6.43020391e-01 -1.74380660e-01 -3.91211826e-03
-1.20648098e+00 6.27954483e-01 -1.70489758e-01 1.66458622e-01
-1.14415593e-01 1.61716565e-02 5.32905221e-01 3.38753760e-01
7.88680375e-01 9.81131196e-02 -4.42335784e-01 -3.83055210e-01
-1.29102266e+00 5.42990088e-01 1.15384245e+00 1.06548357e+00
2.37004355e-01 4.40167248e-01 -4.83318359e-01 1.35598993e+00
3.08278650e-01 5.01272619e-01 5.79082966e-01 -1.05164838e+00
6.72298014e-01 -2.84561813e-01 2.04376414e-01 -2.02254221e-01
-2.62704849e-01 -5.58872879e-01 -7.67101347e-01 1.96493149e-01
4.82660085e-01 -6.66896939e-01 -8.53608251e-01 1.85326016e+00
2.86627084e-01 7.53190443e-02 2.09798947e-01 5.61643064e-01
7.07547367e-01 8.11994433e-01 -2.37832025e-01 -2.02500433e-01
1.16378081e+00 -1.22410274e+00 -9.76541817e-01 -2.10488826e-01
1.02426566e-01 -1.13710964e+00 1.00784111e+00 7.94171035e-01
-1.54342949e+00 -6.88224077e-01 -1.02786064e+00 1.71490267e-01
-1.20537892e-01 1.34341866e-01 3.08195561e-01 1.16465235e+00
-1.38386917e+00 4.14724499e-01 -7.51247942e-01 -9.54835936e-02
-7.13213086e-02 5.41023552e-01 6.61846027e-02 3.29791218e-01
-1.26273966e+00 7.73911178e-01 6.02437928e-02 7.22581297e-02
-1.18593097e+00 -4.43171561e-01 -7.81709492e-01 -3.72224413e-02
-4.75065969e-02 -6.35966480e-01 2.01074052e+00 -8.50175679e-01
-2.17778754e+00 7.09096670e-01 -2.86621332e-01 -5.46384275e-01
9.33233678e-01 -4.07152086e-01 -5.83567560e-01 1.18239075e-02
-2.99075603e-01 8.56263757e-01 9.94070649e-01 -1.23476076e+00
-6.22268021e-01 8.05963352e-02 -1.47347808e-01 2.07938984e-01
-2.63654768e-01 3.32714826e-01 -4.78547364e-01 -9.01135683e-01
-2.64945745e-01 -1.02764654e+00 -1.28931314e-01 -4.54566866e-01
-5.70997477e-01 -1.60088181e-01 6.21569514e-01 -6.60349011e-01
1.08868122e+00 -2.18511891e+00 2.22689047e-01 -6.44164830e-02
-2.09262490e-01 4.97575969e-01 -1.43720075e-01 5.05057752e-01
-2.04371542e-01 1.73577052e-02 -2.43136168e-01 -1.39434242e+00
4.50750381e-01 8.61863568e-02 -2.48151854e-01 2.91595727e-01
2.51610205e-02 4.59890455e-01 -5.71107984e-01 -2.46907771e-01
2.97411799e-01 6.42723203e-01 -8.98187995e-01 5.76558053e-01
-3.30151320e-01 5.35426438e-01 4.41818774e-01 1.54121444e-01
5.66402733e-01 4.44358289e-01 5.26638515e-02 8.74723345e-02
-2.10377872e-02 8.09792101e-01 -1.41031480e+00 1.70476699e+00
-7.32737184e-01 7.39374518e-01 6.52292252e-01 -6.20129287e-01
7.51327693e-01 1.11502945e+00 1.91833168e-01 1.94683913e-02
6.91796988e-02 2.92323619e-01 4.02776487e-02 -1.83978096e-01
2.59670079e-01 -5.25810301e-01 -9.94266421e-02 4.77096021e-01
5.14308393e-01 -5.00422299e-01 2.81342775e-01 2.23656327e-01
6.09745324e-01 -4.43290137e-02 -1.69671088e-01 -3.29011157e-02
5.04195929e-01 -3.17959845e-01 3.88870776e-01 7.27265954e-01
-8.39688182e-02 1.00335932e+00 1.78558901e-01 3.12006831e-01
-1.09549868e+00 -1.04309082e+00 1.02137528e-01 1.18112004e+00
-4.07919198e-01 -5.15743852e-01 -1.15218961e+00 -2.51430988e-01
-4.23343211e-01 1.02731752e+00 -1.79739654e-01 4.42856029e-02
-6.76680446e-01 -9.26739573e-01 1.03471160e+00 3.50260615e-01
2.92897165e-01 -9.83956099e-01 2.05445051e-01 4.10478652e-01
-4.41608548e-01 -1.12784648e+00 -6.78409219e-01 3.08048397e-01
-6.11605346e-01 -3.57119918e-01 -1.02896667e+00 -8.07788253e-01
6.57908944e-03 2.75201261e-01 1.15995061e+00 -6.25261545e-01
2.04646625e-02 2.72420794e-01 -1.97012886e-01 -6.50897205e-01
-1.14214146e+00 2.44439289e-01 2.46896997e-01 5.44067472e-02
1.19902425e-01 -7.98359990e-01 -4.56809044e-01 4.07831997e-01
-7.79029131e-01 -4.83245738e-02 3.80223542e-01 6.96185410e-01
1.05898261e-01 -1.69013128e-01 8.66688669e-01 -7.59761214e-01
7.79497504e-01 -3.81141275e-01 -2.81085640e-01 -2.20867202e-01
-2.41715699e-01 9.09324810e-02 4.17318821e-01 -5.45813859e-01
-1.41458809e+00 -8.28176364e-02 -1.03782117e+00 -2.89789855e-01
-4.88216370e-01 -1.38789952e-01 -6.76473379e-01 6.32754743e-01
6.97566032e-01 1.21195167e-02 1.09194256e-02 -9.85804379e-01
6.46775484e-01 1.13861561e+00 5.51814675e-01 -4.60685581e-01
6.79811001e-01 2.40248650e-01 -6.94644392e-01 -9.93181705e-01
-7.57047176e-01 -5.85588276e-01 -5.93420267e-01 1.64165664e-02
6.94946826e-01 -1.25276017e+00 -5.10210574e-01 6.05834782e-01
-1.14505816e+00 -5.64858913e-01 -3.75463784e-01 6.45551205e-01
-7.56997705e-01 2.46451467e-01 -8.39119673e-01 -1.12063396e+00
-4.44862247e-01 -1.05674827e+00 1.12985766e+00 -1.43503860e-01
-4.78132516e-01 -9.86929178e-01 3.70563895e-01 3.85732353e-01
6.32680297e-01 -1.36971146e-01 5.00613809e-01 -8.02466154e-01
4.26305495e-02 -2.70889819e-01 5.01682699e-01 9.04379010e-01
1.21328138e-01 1.05735190e-01 -1.75441563e+00 -3.34452689e-01
1.70706242e-01 -1.57981679e-01 7.25456059e-01 4.20885861e-01
6.80892706e-01 -3.54703069e-01 -9.47448425e-04 4.01035398e-01
6.52884424e-01 1.05685748e-01 4.45057184e-01 -8.97040814e-02
3.54134917e-01 5.60591996e-01 2.60149777e-01 4.34479833e-01
1.99215919e-01 9.28760827e-01 6.52241260e-02 -2.15180382e-01
-5.98425031e-01 -1.14580862e-01 9.73520398e-01 1.14659619e+00
1.76855892e-01 -8.08445096e-01 -3.85941327e-01 5.25349021e-01
-1.46952379e+00 -1.18151379e+00 -1.53261229e-01 2.25582457e+00
1.03019309e+00 3.98630589e-01 8.82650971e-01 5.18646002e-01
7.57004738e-01 6.86012581e-02 -1.01618737e-01 -4.92772758e-01
-2.47612715e-01 3.41990203e-01 -5.11946389e-03 1.08334100e+00
-1.17350757e+00 7.66081333e-01 7.66793871e+00 8.64565611e-01
-8.78858447e-01 4.30330962e-01 4.56841975e-01 -5.27355492e-01
-8.82431716e-02 -3.43471736e-01 -1.14788461e+00 4.00334775e-01
1.58403003e+00 3.95989306e-02 3.89099151e-01 8.24271679e-01
5.11562169e-01 1.97896659e-01 -1.25056767e+00 9.26392376e-01
1.02548465e-01 -8.80334735e-01 -9.35777500e-02 -2.89590266e-02
4.92234141e-01 1.53431490e-01 -1.81263126e-02 4.90569472e-01
5.23556530e-01 -6.92359984e-01 9.90373909e-01 -3.96966226e-02
6.50855601e-01 -7.16724038e-01 5.49082935e-01 4.68549758e-01
-8.54887486e-01 1.25454500e-01 -8.61662552e-02 8.62595066e-02
4.84156430e-01 6.71995103e-01 -1.19565272e+00 4.31969225e-01
2.33836651e-01 1.42579764e-01 -2.17900500e-01 1.00175905e+00
-2.60865688e-01 1.09807563e+00 -6.57628953e-01 1.66702881e-01
2.40560398e-01 6.22466728e-02 8.35890770e-01 1.85889053e+00
3.15139800e-01 -3.90197217e-01 -6.49810359e-02 5.14779806e-01
5.36911152e-02 -1.05799668e-01 -2.98726737e-01 2.47130170e-01
3.48523170e-01 1.06811285e+00 -2.35708117e-01 -5.78025401e-01
-2.58155107e-01 1.20576870e+00 -4.24930342e-02 3.72498989e-01
-8.72634411e-01 -3.18944156e-01 1.05572462e+00 7.15701841e-04
2.58591563e-01 -2.88584232e-01 1.48725241e-01 -1.06105959e+00
-2.44717196e-01 -1.09122252e+00 3.51165801e-01 -4.02166218e-01
-1.22994077e+00 8.60351503e-01 1.77459791e-01 -1.20376885e+00
-9.18343484e-01 -4.79564786e-01 -7.32233763e-01 1.07714212e+00
-1.19162905e+00 -9.05088127e-01 -1.21212341e-01 5.15394926e-01
1.05441260e+00 -1.03911459e-01 1.11456680e+00 6.79519176e-01
-6.28769159e-01 9.17498946e-01 2.87914932e-01 -2.66105235e-02
1.01188576e+00 -1.46721327e+00 8.15120280e-01 6.77236438e-01
3.63963008e-01 2.74720967e-01 1.19507635e+00 -6.96626306e-02
-7.63116598e-01 -9.63997424e-01 8.71763468e-01 -5.45227945e-01
2.47960135e-01 -8.77345622e-01 -5.81799448e-01 6.52943850e-01
6.38321936e-01 -5.31484604e-01 9.53784287e-01 1.17723256e-01
-1.12233628e-02 -8.40141624e-02 -9.58402038e-01 6.75082207e-01
1.00865650e+00 -3.35078299e-01 -6.38429582e-01 3.23921204e-01
7.14298069e-01 -3.83696079e-01 -5.86983383e-01 -5.59088811e-02
4.42375571e-01 -1.23971415e+00 1.02230823e+00 -4.10922259e-01
-6.86398372e-02 -1.03254691e-01 -1.77449390e-01 -1.65466595e+00
-9.81073454e-02 -1.39041078e+00 -1.98910683e-01 1.75181401e+00
6.47630692e-01 -3.66114676e-01 6.02739334e-01 5.43212056e-01
-4.16023284e-01 -2.00224474e-01 -8.96032095e-01 -8.45118165e-01
2.98497081e-01 -6.67289019e-01 3.28848630e-01 3.99617881e-01
3.37712554e-04 7.89330959e-01 -5.93206465e-01 3.11320517e-02
4.94977653e-01 -1.65529355e-01 1.07925379e+00 -9.72016454e-01
-7.77232349e-01 -5.07473290e-01 3.68406177e-02 -1.37122774e+00
3.04138642e-02 -8.22975159e-01 2.68015057e-01 -1.29653049e+00
-3.70908141e-01 -4.06441867e-01 5.72404973e-02 1.52139276e-01
-1.90335676e-01 3.80281180e-01 1.03686586e-01 -1.64173637e-02
-3.07988644e-01 6.52490854e-01 8.99500012e-01 -1.30082056e-01
-3.92676950e-01 5.99590659e-01 -4.90567714e-01 7.26558864e-01
9.32557940e-01 -5.37931621e-01 -5.94587266e-01 -2.68567234e-01
-3.20252061e-01 4.84841391e-02 8.97103176e-02 -1.17978072e+00
-1.45484120e-01 4.40620214e-01 2.22921938e-01 -5.85532427e-01
1.05382526e+00 -3.70551109e-01 1.64247185e-01 1.79647759e-01
-6.22302413e-01 -2.52217352e-01 3.10790122e-01 4.58214760e-01
-4.21937734e-01 -1.99841529e-01 9.85781968e-01 -7.84958825e-02
6.51096329e-02 2.65440047e-02 -8.05419266e-01 1.40234500e-01
6.58293128e-01 1.11680821e-01 2.11452588e-01 -9.98085916e-01
-1.24884510e+00 1.11805778e-02 7.36412033e-02 4.54851210e-01
6.06205277e-02 -1.09777796e+00 -1.02275491e+00 1.84726357e-01
-3.13709289e-01 -2.21482776e-02 4.28939536e-02 5.81329286e-01
-2.31918961e-01 4.45171863e-01 1.56328246e-01 -3.64510477e-01
-1.43195176e+00 4.43966776e-01 5.31015277e-01 -1.03902876e-01
-4.97660458e-01 1.16302526e+00 1.66762009e-01 -4.77785945e-01
7.15083957e-01 -2.76641279e-01 8.03234354e-02 3.09483618e-01
6.96189702e-01 3.75704110e-01 3.34162831e-01 -4.68520641e-01
-2.73070097e-01 -5.13237342e-02 6.59170970e-02 -9.74696577e-01
1.25293815e+00 -3.21132571e-01 3.35662484e-01 7.76193678e-01
1.05457878e+00 4.24644232e-01 -1.35586262e+00 -1.13409914e-01
-1.90524891e-01 -1.71616912e-01 -1.92955837e-01 -7.57546246e-01
-8.08467448e-01 1.27233744e+00 5.70876658e-01 3.65529358e-01
8.27684700e-01 -2.65370263e-03 8.32109749e-01 1.99777469e-01
-1.76357236e-02 -9.51169372e-01 3.84441689e-02 4.14641112e-01
9.38885272e-01 -9.40200031e-01 -4.50218678e-01 -3.94713640e-01
-8.28935444e-01 8.41268778e-01 2.72987276e-01 -4.36401868e-04
7.33071268e-01 6.19224608e-01 4.54731703e-01 2.80521601e-01
-8.68808210e-01 -2.38245696e-01 6.31628111e-02 7.13639557e-01
7.21028149e-01 -3.63672599e-02 5.75999916e-02 6.53419912e-01
-5.63776016e-01 -6.29736781e-01 3.45392019e-01 5.24906576e-01
-2.77754694e-01 -1.52830195e+00 -5.78062952e-01 -4.26674709e-02
-6.99853182e-01 -2.13905469e-01 -5.46991467e-01 4.86376107e-01
-1.04398347e-01 1.47429097e+00 1.29106762e-02 -2.80706465e-01
3.84925365e-01 6.49240553e-01 3.91435087e-01 -7.04475880e-01
-9.89805698e-01 6.11974597e-01 6.45466745e-01 2.06849799e-02
-3.57176870e-01 -7.12056875e-01 -8.79124880e-01 -3.52393627e-01
-4.04256880e-01 4.09538984e-01 8.34719479e-01 8.02838683e-01
3.78035039e-01 7.00962365e-01 7.26718247e-01 -1.59620881e+00
-7.12763309e-01 -1.58941269e+00 -5.90125144e-01 4.47872132e-01
5.65886438e-01 -4.14680153e-01 -7.03298986e-01 2.18683541e-01] | [14.773112297058105, 6.214160919189453] |
cf346c67-59d9-47df-aee9-ea60952a8322 | fir-based-future-trajectory-prediction-in | 2304.05345 | null | https://arxiv.org/abs/2304.05345v1 | https://arxiv.org/pdf/2304.05345v1.pdf | FIR-based Future Trajectory Prediction in Nighttime Autonomous Driving | The performance of the current collision avoidance systems in Autonomous Vehicles (AV) and Advanced Driver Assistance Systems (ADAS) can be drastically affected by low light and adverse weather conditions. Collisions with large animals such as deer in low light cause significant cost and damage every year. In this paper, we propose the first AI-based method for future trajectory prediction of large animals and mitigating the risk of collision with them in low light. In order to minimize false collision warnings, in our multi-step framework, first, the large animal is accurately detected and a preliminary risk level is predicted for it and low-risk animals are discarded. In the next stage, a multi-stream CONV-LSTM-based encoder-decoder framework is designed to predict the future trajectory of the potentially high-risk animals. The proposed model uses camera motion prediction as well as the local and global context of the scene to generate accurate predictions. Furthermore, this paper introduces a new dataset of FIR videos for large animal detection and risk estimation in real nighttime driving scenarios. Our experiments show promising results of the proposed framework in adverse conditions. Our code is available online. | ['Justin Miller', 'Devesh Upadhyay', 'Navid Fallahinia', 'Alireza Rahimpour'] | 2023-03-31 | null | null | null | null | ['motion-prediction', 'trajectory-prediction'] | ['computer-vision', 'computer-vision'] | [ 2.66435921e-01 -3.92464809e-02 1.40356600e-01 -7.02297509e-01
-2.38670543e-01 1.71270333e-02 3.79987240e-01 3.45500082e-01
-9.83883321e-01 6.56920612e-01 -1.82995185e-01 -2.39443988e-01
-1.73575114e-02 -9.33033228e-01 -7.66281724e-01 -5.29349625e-01
-4.65744346e-01 2.06652865e-01 9.02073860e-01 -4.14943516e-01
1.12995617e-01 8.33368242e-01 -1.94020581e+00 1.40806437e-01
6.36071086e-01 8.66827488e-01 7.19524562e-01 9.49725688e-01
3.12651277e-01 9.53853309e-01 -2.86676168e-01 -3.47200543e-01
3.05923671e-01 -1.91859111e-01 -1.35957554e-01 -2.73469508e-01
4.02294576e-01 -9.07229483e-01 -4.22596693e-01 7.81337559e-01
3.52038711e-01 3.01788658e-01 5.34416199e-01 -1.60373867e+00
2.91040003e-01 2.60725319e-01 -5.34873605e-01 5.48844993e-01
-1.67639583e-01 6.56666279e-01 2.72437990e-01 -5.58599293e-01
4.30812776e-01 1.28959012e+00 6.24746144e-01 4.61881548e-01
-7.99846828e-01 -9.56853986e-01 2.26660594e-01 8.44614327e-01
-1.34274042e+00 -5.66335201e-01 4.62268442e-01 -5.61991870e-01
1.02946341e+00 -1.44324556e-01 7.86758602e-01 7.62474477e-01
7.78774083e-01 5.62418461e-01 4.67631370e-01 5.12470938e-02
6.89587146e-02 -1.02898180e-01 5.83471358e-02 7.39710987e-01
2.98451811e-01 7.27873445e-01 -3.52294594e-01 1.74127202e-02
-1.27707034e-01 1.33916318e-01 3.82494450e-01 -1.69968739e-01
-1.10675704e+00 8.17475557e-01 5.84496260e-01 -3.08296174e-01
-6.67677581e-01 3.10873985e-01 4.69371080e-01 5.13287000e-02
3.01165313e-01 -2.13575140e-01 -9.53345560e-03 7.77890384e-02
-8.39474082e-01 5.29624522e-01 3.03423345e-01 1.01534462e+00
8.69618893e-01 7.21852854e-02 -1.89398259e-01 7.15274870e-01
3.94805968e-01 8.28396261e-01 -2.87400037e-01 -9.34888363e-01
4.18926597e-01 1.40022114e-01 2.94589758e-01 -1.23191738e+00
-6.16155446e-01 -1.91988647e-01 -4.99850780e-01 6.98839903e-01
-9.25238803e-02 -3.40197623e-01 -7.64587939e-01 1.57173562e+00
4.82019544e-01 5.84657013e-01 2.04855800e-01 9.07726884e-01
5.59001803e-01 1.04403484e+00 6.02614164e-01 -3.06858212e-01
1.10752320e+00 -9.30538535e-01 -8.26241493e-01 -8.69257569e-01
6.86336935e-01 -5.49433887e-01 1.93817288e-01 2.07794219e-01
-6.34099722e-01 -6.65946186e-01 -1.10170949e+00 1.37917653e-01
-2.14345545e-01 9.22694653e-02 3.37664753e-01 8.03468097e-03
-8.63301396e-01 4.02111679e-01 -8.69243026e-01 -7.32518613e-01
4.41473007e-01 2.67801374e-01 -2.11200327e-01 -1.52533323e-01
-1.18301404e+00 1.33821094e+00 3.98225546e-01 6.35059416e-01
-1.35655296e+00 -3.89812708e-01 -1.04427564e+00 -2.91436046e-01
4.14239734e-01 -2.43150339e-01 1.16330111e+00 -4.91412014e-01
-7.84719944e-01 6.96431100e-01 -2.73880064e-01 -1.13032973e+00
6.05139256e-01 -5.61462641e-01 -4.49916720e-01 4.73070517e-02
2.31842205e-01 1.18638468e+00 6.55420899e-01 -1.14165962e+00
-1.52305877e+00 -2.20803633e-01 -1.31520271e-01 3.91174972e-01
2.49048829e-01 3.00524771e-01 -1.34482250e-01 -3.01585495e-01
-4.56149518e-01 -1.08510649e+00 -6.62571192e-01 4.09890324e-01
1.73719883e-01 -7.95503259e-02 9.19029176e-01 -3.68873030e-01
1.09836650e+00 -2.10764885e+00 -4.06161219e-01 -1.83735058e-01
-3.51863466e-02 4.75988299e-01 -2.14297384e-01 1.54935211e-01
5.06496906e-01 -5.76540947e-01 -2.75031447e-01 -1.30542770e-01
-4.56096292e-01 3.92686874e-01 -1.95990622e-01 6.70945585e-01
4.85268414e-01 4.03058916e-01 -1.01182401e+00 -8.25363517e-01
5.54206312e-01 3.23355794e-01 -4.42446709e-01 5.39302588e-01
8.22729915e-02 2.10952163e-01 -2.93557465e-01 3.79208833e-01
9.79893565e-01 9.05342698e-01 -3.31914335e-01 3.97794461e-03
-7.44365871e-01 -2.06588998e-01 -8.53425264e-01 9.52483356e-01
-3.55832338e-01 8.25774014e-01 1.09672762e-01 -7.40686476e-01
9.15035188e-01 -1.89991683e-01 3.18822384e-01 -6.36834621e-01
1.65576488e-01 7.22414106e-02 1.11327268e-01 -7.91442990e-01
7.97762394e-01 -2.99013127e-02 -1.62091523e-01 -1.12457559e-01
-3.40264052e-01 1.46876901e-01 1.93824306e-01 4.52616997e-02
1.03035152e+00 2.06012323e-01 1.81101993e-01 -6.56854585e-02
5.94860911e-01 4.70852822e-01 9.51639056e-01 7.25203216e-01
-9.29032087e-01 7.19302297e-02 -1.32480711e-02 -6.79005027e-01
-8.12158644e-01 -8.37783754e-01 -1.78114742e-01 9.82376575e-01
6.73843801e-01 9.20202001e-04 -6.15781486e-01 -3.58829200e-01
-2.68053096e-02 1.21961021e+00 -4.09946114e-01 -6.28486454e-01
-9.20365512e-01 -6.31819129e-01 5.15121281e-01 5.09055614e-01
4.91024554e-01 -1.22446823e+00 -1.26700497e+00 4.45478201e-01
-1.39464781e-01 -1.30123305e+00 -8.09214860e-02 2.65262902e-01
-4.39454734e-01 -8.95127892e-01 -1.93356365e-01 -7.97349215e-01
4.57071036e-01 7.33399034e-01 4.90089893e-01 2.19385177e-01
-4.96538788e-01 -2.99889982e-01 -9.37512964e-02 -7.06617177e-01
-4.87779409e-01 -5.44881999e-01 1.74587086e-01 -4.81380746e-02
7.66880512e-01 -3.45160030e-02 -7.17292607e-01 4.33137596e-01
-3.10614914e-01 2.28386387e-01 6.31370485e-01 5.11378765e-01
5.40576577e-01 1.73421428e-01 4.24264640e-01 -3.44199598e-01
3.25454250e-02 -7.99370527e-01 -8.85447145e-01 -1.27359495e-01
-1.52726337e-01 -1.75581485e-01 6.44635737e-01 -3.38657051e-01
-1.24145353e+00 5.45444071e-01 -3.19345802e-01 -2.64962077e-01
-6.61658883e-01 1.57545134e-02 1.45216003e-01 4.20776568e-02
4.02665019e-01 -7.99057633e-02 -2.79262602e-01 -1.99657887e-01
-4.31161001e-02 5.76360047e-01 7.99708545e-01 1.90385476e-01
8.66691470e-01 4.74230856e-01 3.33630234e-01 -1.02474153e+00
-5.69457054e-01 -4.38481867e-01 -6.07869625e-01 -1.03839695e+00
1.15053713e+00 -1.21063876e+00 -7.06582010e-01 6.84800327e-01
-1.29716396e+00 -2.60880649e-01 2.51047522e-01 9.08032596e-01
-5.00567794e-01 2.23223835e-01 -1.44508913e-01 -1.23419845e+00
-1.38311222e-01 -1.11874199e+00 9.32548404e-01 3.76966387e-01
8.06569755e-02 -2.25750789e-01 -1.68274909e-01 2.85010815e-01
2.71184742e-01 3.83150995e-01 3.93387049e-01 -2.74642199e-01
-7.08357394e-01 -1.55738845e-01 -1.78524733e-01 6.12493120e-02
-2.66467005e-01 9.75970328e-02 -6.61579669e-01 -3.57619673e-02
-1.50348976e-01 -1.14369839e-01 1.02173316e+00 5.21156728e-01
6.48877263e-01 -3.86514701e-02 -8.00220072e-01 6.46978915e-01
1.21901536e+00 5.56913376e-01 4.05651242e-01 3.93782973e-01
6.18323565e-01 1.12716722e+00 1.62662077e+00 5.64734697e-01
6.40363693e-01 6.27579331e-01 8.22082877e-01 8.23370591e-02
2.11019441e-02 -2.39162207e-01 6.68376207e-01 1.83568433e-01
2.07017064e-01 -5.04746497e-01 -9.60479856e-01 8.42559695e-01
-2.01490760e+00 -1.36603534e+00 -8.15340519e-01 2.29126191e+00
-1.76147912e-02 4.43349004e-01 1.19508803e-01 -2.39326745e-01
8.56728911e-01 4.43734489e-02 -5.40274620e-01 -5.97556889e-01
1.17069393e-01 -6.18113935e-01 1.16693306e+00 7.24779367e-01
-1.35558772e+00 1.20435739e+00 5.97872353e+00 6.64710939e-01
-8.53386998e-01 -1.20638728e-01 3.52168381e-01 -3.06840658e-01
1.96878001e-01 -1.27629424e-02 -1.25554502e+00 4.03497398e-01
1.15318573e+00 -1.44820943e-01 1.28682628e-02 9.69046831e-01
6.98284090e-01 -7.45694637e-01 -8.43526363e-01 6.04425430e-01
1.58777181e-02 -8.10203791e-01 -2.61102408e-01 -1.80378571e-01
5.83506189e-02 2.92342663e-01 -2.68235207e-01 2.68779844e-01
2.31208324e-01 -5.98622262e-01 9.85547662e-01 6.51272476e-01
3.08002919e-01 -1.25609338e+00 8.48949850e-01 8.02668452e-01
-1.29997170e+00 -3.60134214e-01 -7.34808266e-01 -9.62836519e-02
6.70348465e-01 2.23650634e-01 -7.85602868e-01 6.37581646e-02
8.70131910e-01 7.02063859e-01 -6.60420656e-01 1.28261781e+00
7.09357485e-02 4.55383062e-01 -6.20172739e-01 -2.03263938e-01
5.59521914e-01 -1.92593098e-01 9.32139635e-01 1.33604515e+00
4.17499393e-01 3.72664064e-01 2.05499247e-01 4.83599335e-01
5.63179910e-01 -2.42565900e-01 -1.05174541e+00 5.46001732e-01
1.71226040e-01 1.11499071e+00 -4.69708651e-01 -2.74936169e-01
-3.84736210e-01 6.32420957e-01 2.29034290e-01 -1.14546102e-02
-1.30468738e+00 -2.88345903e-01 9.00976062e-01 1.89628303e-01
2.03132346e-01 -2.88961172e-01 1.09132707e-01 -6.47730649e-01
-1.32782534e-01 -1.32095139e-03 3.06360930e-01 -7.81794071e-01
-6.33165359e-01 8.16457510e-01 3.33313376e-01 -1.50920951e+00
-1.12626530e-01 -2.46068165e-01 -6.69133961e-01 4.98315543e-01
-1.85464144e+00 -1.08586836e+00 -3.68521959e-01 2.49890059e-01
8.84064496e-01 -7.87656307e-02 2.13432863e-01 3.99576098e-01
-5.23283482e-01 2.21932128e-01 -1.24725401e-01 -3.90159428e-01
5.98181248e-01 -5.73583364e-01 4.80291188e-01 1.33664930e+00
-6.51056767e-01 -2.60287493e-01 1.09582281e+00 -1.12703967e+00
-1.15230536e+00 -1.67696702e+00 1.16991484e+00 1.27732242e-02
4.07523870e-01 -4.45129186e-01 -8.74885798e-01 4.48203862e-01
-5.10325795e-03 -3.50820087e-02 2.10136607e-01 -6.54055476e-01
4.65958297e-01 -3.82765114e-01 -1.19981003e+00 7.19377100e-01
1.09825337e+00 1.43199623e-01 -5.04969001e-01 2.75210023e-01
5.64250767e-01 -7.44612813e-02 -1.45920619e-01 7.20241368e-01
5.65979600e-01 -9.71053183e-01 7.69282520e-01 -2.89083391e-01
2.10156336e-01 -6.12959206e-01 -1.49560310e-02 -1.01097822e+00
-5.38473308e-01 -1.60268620e-01 2.20141008e-01 8.77630770e-01
1.47379249e-01 -1.06946245e-01 5.01771510e-01 4.26450163e-01
-5.07048607e-01 -3.17821532e-01 -9.93444264e-01 -6.34340763e-01
-4.07112300e-01 -7.45457232e-01 2.05179006e-01 1.24242194e-02
-4.88938034e-01 8.66872445e-02 -9.91000772e-01 8.07667434e-01
9.59910452e-01 -1.99658662e-01 8.46771479e-01 -1.08496213e+00
4.36386794e-01 1.22776672e-01 -8.70272756e-01 -7.64721215e-01
2.59436309e-01 -4.94819582e-01 8.15434575e-01 -1.51151443e+00
-2.71637682e-02 -1.67374417e-01 -9.41490680e-02 3.11298966e-01
-2.47200385e-01 1.07837163e-01 8.11987743e-02 -1.03991710e-01
-5.86412072e-01 6.05737031e-01 6.98757768e-01 -5.65650761e-02
-5.72333559e-02 1.74629599e-01 9.93237793e-02 9.33754206e-01
7.54250526e-01 -8.95752728e-01 -3.53878558e-01 -4.66400325e-01
-1.25579521e-01 3.11749846e-01 5.83964705e-01 -1.56153893e+00
5.11337996e-01 -6.97213113e-01 1.30538836e-01 -1.26869154e+00
4.82667953e-01 -1.15688145e+00 1.31757706e-01 9.72571075e-01
-2.80420154e-01 -4.57481891e-02 4.90651697e-01 9.46846247e-01
5.55939972e-02 -2.36629173e-01 1.15372431e+00 2.20158562e-01
-1.31672287e+00 4.03095722e-01 -1.34304857e+00 -3.71812552e-01
1.73422098e+00 -2.68361181e-01 -1.54357776e-01 -1.84775010e-01
-3.18882227e-01 1.04123485e+00 1.51487499e-01 9.06350076e-01
8.45749795e-01 -1.03287518e+00 -9.88810301e-01 3.26806784e-01
4.66833860e-01 -2.32912377e-01 6.90863729e-01 6.62578046e-01
-9.05695617e-01 2.42823407e-01 -6.61598325e-01 -5.67765951e-01
-1.58810008e+00 7.47158408e-01 2.89000094e-01 4.03113008e-01
-6.08086050e-01 7.46399701e-01 5.50530255e-01 1.27005428e-01
2.88749993e-01 -1.83631077e-01 -5.80480456e-01 -3.09819970e-02
7.28070498e-01 4.13545519e-01 -2.80294299e-01 -1.40124428e+00
-6.82802558e-01 3.98428649e-01 1.59250945e-02 1.41397744e-01
1.39295685e+00 -5.27740061e-01 3.14909667e-01 2.68818855e-01
8.67699146e-01 -4.03622746e-01 -1.62946129e+00 6.64773285e-02
-2.19466910e-02 -5.28204858e-01 3.63651067e-01 -3.78106415e-01
-8.19037318e-01 1.01892018e+00 1.01842535e+00 -4.36581939e-01
1.01402104e+00 -2.14709625e-01 1.16485035e+00 4.75674212e-01
4.92881209e-01 -1.26950753e+00 -6.67851627e-01 5.37951589e-01
7.16106892e-01 -1.43933940e+00 -1.03332512e-01 -1.64241612e-01
-1.07829654e+00 7.83589065e-01 1.22550786e+00 -1.21439070e-01
5.39540470e-01 4.26550061e-01 9.85924602e-02 -4.15764265e-02
-1.16900504e+00 -5.40905595e-01 -7.16772452e-02 8.13731551e-01
-1.83292001e-01 7.15886578e-02 -2.50412196e-01 1.83852643e-01
-4.41733710e-02 -4.07583900e-02 7.26708949e-01 1.03142202e+00
-1.10572350e+00 -3.94391179e-01 -3.71264338e-01 4.51397449e-01
-2.20597014e-01 1.55735463e-01 -9.08159167e-02 3.14989090e-01
6.11475229e-01 1.23286915e+00 3.50152820e-01 -5.46267271e-01
5.07685125e-01 -2.94160306e-01 -1.90086395e-01 -3.91274124e-01
-4.21802849e-01 -1.98757619e-01 4.38418359e-01 -7.14746892e-01
-4.17557061e-01 -6.83575869e-01 -1.34165323e+00 -3.77206981e-01
-2.75721729e-01 -8.67664218e-02 6.14440501e-01 8.66131365e-01
1.79153636e-01 2.71813482e-01 6.61576688e-01 -9.72771525e-01
5.81449922e-03 -6.20834947e-01 -2.78369069e-01 3.25077325e-02
7.35706508e-01 -8.87452483e-01 -2.51688123e-01 6.84208125e-02] | [7.8504638671875, -1.0479689836502075] |
d5638792-095c-4de7-bd88-c14b913afdb7 | discriminative-localization-in-cnns-for | 1707.01086 | null | http://arxiv.org/abs/1707.01086v2 | http://arxiv.org/pdf/1707.01086v2.pdf | Discriminative Localization in CNNs for Weakly-Supervised Segmentation of Pulmonary Nodules | Automated detection and segmentation of pulmonary nodules on lung computed
tomography (CT) scans can facilitate early lung cancer diagnosis. Existing
supervised approaches for automated nodule segmentation on CT scans require
voxel-based annotations for training, which are labor- and time-consuming to
obtain. In this work, we propose a weakly-supervised method that generates
accurate voxel-level nodule segmentation trained with image-level labels only.
By adapting a convolutional neural network (CNN) trained for image
classification, our proposed method learns discriminative regions from the
activation maps of convolution units at different scales, and identifies the
true nodule location with a novel candidate-screening framework. Experimental
results on the public LIDC-IDRI dataset demonstrate that, our weakly-supervised
nodule segmentation framework achieves competitive performance compared to a
fully-supervised CNN-based segmentation method. | ['Elsa D. Angelini', 'Xinyang Feng', 'Jie Yang', 'Andrew F. Laine'] | 2017-07-04 | null | null | null | null | ['lung-cancer-diagnosis'] | ['medical'] | [ 3.13692451e-01 3.72519106e-01 -7.10652947e-01 -4.87196892e-01
-1.35196900e+00 -4.78285044e-01 9.79133174e-02 -1.57910287e-01
-4.60393906e-01 2.85485238e-01 2.73770336e-02 -6.16974831e-01
3.01919371e-01 -7.19839513e-01 -5.64360082e-01 -6.53194487e-01
1.32118329e-01 1.00568879e+00 7.14860201e-01 5.89611530e-01
-3.74990612e-01 4.32991356e-01 -8.47589731e-01 5.82519531e-01
6.62029862e-01 1.03023624e+00 4.84763145e-01 8.89648438e-01
-3.04165423e-01 9.67353463e-01 2.44775023e-02 -1.62911601e-02
3.31694722e-01 -6.49624586e-01 -1.25203025e+00 5.24980426e-01
3.89422417e-01 -5.18489063e-01 -2.39494175e-01 9.91351962e-01
2.18932070e-02 -6.20147705e-01 7.86194742e-01 -5.32531142e-01
-4.00668681e-01 6.69606626e-01 -3.38430017e-01 3.58894765e-01
-5.12357354e-01 2.67191768e-01 9.34850633e-01 -9.68173742e-01
3.55117828e-01 4.33035344e-01 8.16908181e-01 6.10453904e-01
-9.03905630e-01 -4.98199403e-01 -2.69442737e-01 -3.58511627e-01
-1.44777489e+00 2.81788498e-01 3.33704531e-01 -4.05870467e-01
3.85777533e-01 3.37199062e-01 9.43929374e-01 6.26129627e-01
-5.07797413e-02 8.35916281e-01 9.55324829e-01 -3.64675373e-02
3.15967277e-02 1.42011195e-01 -2.44888559e-01 1.41273940e+00
5.27026415e-01 1.55539336e-02 3.80253047e-01 -2.32466415e-01
1.42389238e+00 3.29811573e-01 -2.51260865e-02 -5.45607746e-01
-1.45102382e+00 7.61353672e-01 1.08057654e+00 6.04851067e-01
-3.51459622e-01 4.54349965e-01 2.84887552e-01 -2.90459126e-01
5.31797826e-01 2.79533327e-01 -1.77020669e-01 3.55842113e-01
-1.26875162e+00 -4.32313621e-01 7.68751144e-01 6.31937027e-01
4.93126482e-01 -2.18331471e-01 -7.86606908e-01 5.53350687e-01
2.55276829e-01 3.96103889e-01 7.50469208e-01 -4.88677680e-01
7.81907961e-02 8.30276966e-01 -1.64032549e-01 2.15496849e-02
-6.62172854e-01 -7.01080918e-01 -8.59626889e-01 -8.48219842e-02
5.62747836e-01 -8.35046321e-02 -1.48357451e+00 9.15881991e-01
3.03879976e-01 3.81034285e-01 -2.76896000e-01 1.03352332e+00
1.19222796e+00 -1.16394991e-02 3.51259828e-01 1.55568376e-01
1.46719491e+00 -1.38159525e+00 -2.23399803e-01 -1.97588459e-01
8.25035691e-01 -5.98496735e-01 1.14766026e+00 -3.23847651e-01
-9.99593019e-01 -2.96334743e-01 -6.20221078e-01 5.56404628e-02
6.96630627e-02 9.30802643e-01 5.11404395e-01 8.13354015e-01
-1.00168288e+00 2.20466420e-01 -1.24638593e+00 -3.90652150e-01
1.07718945e+00 5.85876107e-01 7.76506513e-02 4.91785817e-02
-4.28407758e-01 5.84648192e-01 6.19609594e-01 -9.97103006e-02
-1.33829927e+00 -9.24927235e-01 -3.33791226e-01 1.74040392e-01
6.10397518e-01 -6.94511652e-01 1.77983749e+00 -9.36704099e-01
-1.29827619e+00 1.10166979e+00 -6.48733005e-02 -5.52459061e-01
7.51262844e-01 3.57991546e-01 3.02586377e-01 5.73494077e-01
2.37117410e-01 9.95242476e-01 8.30198944e-01 -1.04873323e+00
-6.55542016e-01 7.30520785e-02 -4.88672674e-01 1.16324008e-01
-2.02911898e-01 -1.62780404e-01 -8.23917568e-01 -4.41947192e-01
3.33864480e-01 -1.05008268e+00 -8.32206368e-01 4.22926068e-01
-7.35553384e-01 -3.27255666e-01 1.06955254e+00 -6.06939077e-01
7.81358600e-01 -1.73879981e+00 -5.07854402e-01 4.68863159e-01
5.17364323e-01 2.12620094e-01 1.22078739e-01 -6.45130515e-01
6.47840053e-02 4.84842002e-01 -5.17863929e-01 -1.80839561e-02
-3.97846401e-01 2.88565129e-01 3.86573374e-01 3.50349337e-01
3.66344035e-01 1.72719145e+00 -7.73181617e-01 -1.23190248e+00
2.66047865e-01 2.15170532e-01 -5.39943457e-01 3.77358824e-01
-4.06645864e-01 7.67438710e-01 -7.98398972e-01 1.09345758e+00
2.22083762e-01 -1.02468669e+00 1.45869821e-01 -2.30323553e-01
9.92196053e-02 7.09507838e-02 -3.12697738e-01 1.46861541e+00
-4.35291320e-01 2.88399607e-01 -9.08341035e-02 -6.43701792e-01
4.08446044e-01 6.90258265e-01 8.81243229e-01 -1.33712247e-01
4.07469243e-01 4.39434469e-01 1.88432604e-01 -6.59981251e-01
-3.15361232e-01 -1.57264560e-01 1.82446726e-02 7.27869511e-01
-6.33215457e-02 -6.01643562e-01 2.19948031e-02 -4.56348620e-02
1.41773987e+00 -2.03262091e-01 3.90008152e-01 -2.15268672e-01
5.61033487e-01 5.76569200e-01 2.01459736e-01 8.78802359e-01
-4.40584183e-01 1.07051623e+00 3.22247475e-01 -4.47750300e-01
-1.16545880e+00 -1.04311991e+00 -3.55265737e-01 8.42934906e-01
-2.08360940e-01 1.51748911e-01 -9.68382061e-01 -1.50596464e+00
-2.36222088e-01 9.01288877e-04 -8.63687932e-01 2.82430708e-01
-8.35639119e-01 -7.21263587e-01 7.48589277e-01 9.02465463e-01
7.26074576e-01 -1.02971768e+00 -5.97783625e-01 1.16894051e-01
-2.27590144e-01 -1.25497186e+00 -6.75324619e-01 5.47089577e-01
-1.18882620e+00 -1.43443108e+00 -1.04324102e+00 -1.36158085e+00
1.38307643e+00 2.97901213e-01 1.24612892e+00 4.87986028e-01
-8.58393908e-01 3.65443885e-01 3.89975011e-02 -2.13898063e-01
-4.15140957e-01 3.51183981e-01 -7.66597092e-01 -3.64311546e-01
2.78502285e-01 2.75485087e-02 -7.44594574e-01 3.81849736e-01
-9.25728738e-01 6.65429235e-02 1.38800275e+00 8.65128338e-01
1.20458889e+00 -4.63017784e-02 2.28978574e-01 -1.51355934e+00
3.12210947e-01 -4.31352258e-01 -4.46268082e-01 2.87692815e-01
-2.75510937e-01 -1.65648803e-01 4.97522593e-01 -2.77500182e-01
-1.17933106e+00 8.42648983e-01 -1.01750046e-01 -3.85598660e-01
-4.15680319e-01 3.85663182e-01 4.60154176e-01 -5.50254762e-01
7.72251189e-01 2.93560475e-01 -9.48486924e-02 -1.68657690e-01
2.56652296e-01 5.33846915e-01 5.93784332e-01 -6.06851652e-02
1.19467330e+00 8.77553403e-01 -8.08712468e-02 -3.81361246e-01
-1.32827020e+00 -9.17406917e-01 -1.04211259e+00 -1.03231847e-01
1.29072201e+00 -8.46140206e-01 -7.09192529e-02 -7.77683556e-02
-7.68490314e-01 -5.64977586e-01 -6.03099763e-01 6.78365886e-01
-4.12002653e-01 2.60674972e-02 -7.78561950e-01 -6.82016239e-02
-4.38123345e-01 -1.10243237e+00 1.24936461e+00 1.91369861e-01
-2.37015076e-02 -8.87297988e-01 7.12238923e-02 5.87362111e-01
5.35533249e-01 -1.05742728e-02 8.35041046e-01 -8.93433869e-01
-1.01578748e+00 -5.02716422e-01 -7.50294924e-01 2.02348888e-01
3.52180302e-01 -1.40161976e-01 -8.73494565e-01 -1.43995509e-01
-2.12275490e-01 -6.12421572e-01 1.09933174e+00 6.81683242e-01
1.97139311e+00 -1.63471073e-01 -6.99385762e-01 8.50850523e-01
1.26922166e+00 -1.86929211e-01 -6.81259995e-03 -7.23598003e-02
1.02644849e+00 1.25749901e-01 2.54304826e-01 8.17761272e-02
-3.58800776e-03 -4.18194383e-02 5.45401692e-01 -9.34098899e-01
-4.95542258e-01 -1.06395505e-01 -3.44136029e-01 2.98128933e-01
-2.46154264e-01 9.83005688e-02 -1.14741218e+00 7.02757120e-01
-1.39051509e+00 -4.15383041e-01 -1.38820827e-01 1.66526806e+00
8.66365314e-01 7.43103772e-02 1.20101653e-01 -3.63690645e-01
7.01252818e-01 -1.66158825e-01 -7.80926406e-01 3.05517375e-01
4.66384977e-01 5.86597919e-01 8.64479899e-01 -1.57801285e-02
-1.65151203e+00 8.07923973e-01 6.73966026e+00 6.36825383e-01
-1.22255015e+00 6.29505038e-01 9.31917965e-01 -1.23924360e-01
-2.24181265e-02 -4.10526156e-01 -3.15654814e-01 -6.29372448e-02
5.25727630e-01 1.23506270e-01 -2.34086081e-01 1.17579770e+00
1.20451219e-01 6.50636479e-02 -1.16193068e+00 5.57628453e-01
-7.41852671e-02 -1.41696453e+00 -2.70914119e-02 1.60948083e-01
1.18757212e+00 6.09316885e-01 2.23038524e-01 1.48065820e-01
4.77605492e-01 -1.26616311e+00 1.02363281e-01 4.44388166e-02
8.88539374e-01 -4.95238692e-01 1.07669318e+00 2.53039539e-01
-1.39236939e+00 1.29680082e-01 -4.41075534e-01 8.10856521e-01
-2.83512563e-01 5.20858407e-01 -1.85826004e+00 1.32493125e-02
7.12488234e-01 4.08431530e-01 -9.79255259e-01 1.38418150e+00
-2.88961351e-01 1.25206661e+00 -2.37895072e-01 -1.01077676e-01
6.49520636e-01 2.05743209e-01 1.43060669e-01 1.26531017e+00
4.05920088e-01 1.91612095e-01 6.92604184e-01 1.25846028e+00
-4.80623841e-01 1.99337170e-01 -1.53064907e-01 -1.80154383e-01
7.81574007e-03 1.78885663e+00 -1.49502659e+00 -4.48871106e-01
-5.30297399e-01 8.07239056e-01 1.82466403e-01 1.55036986e-01
-9.58267570e-01 2.57116616e-01 -3.36924940e-01 3.03214371e-01
5.15968502e-01 2.09936693e-01 -3.69546652e-01 -8.42464209e-01
-3.19522798e-01 -3.06988925e-01 5.34312725e-01 -4.61947978e-01
-1.25706303e+00 7.33415186e-01 -3.41441840e-01 -1.42390990e+00
-1.41265795e-01 -5.42319357e-01 -1.04867017e+00 3.53585511e-01
-1.62308776e+00 -1.48395443e+00 -7.56826937e-01 5.11198819e-01
4.38690096e-01 7.01708719e-02 5.95104694e-01 -5.09631820e-02
-4.61803079e-01 3.82925183e-01 -1.59402058e-01 7.03552365e-01
3.87539327e-01 -1.53655982e+00 6.68421537e-02 5.01737893e-01
1.64021552e-01 -1.86304729e-02 -1.86697483e-01 -7.17810750e-01
-8.49439263e-01 -1.96050346e+00 3.98431063e-01 -4.10276622e-01
5.14103830e-01 7.25873411e-02 -8.48712206e-01 8.00987482e-01
3.30024734e-02 8.07504892e-01 7.87366986e-01 -4.86635387e-01
-6.40139077e-03 2.95440882e-01 -1.35170352e+00 3.71669888e-01
6.94411695e-01 -5.02865911e-01 -2.55792618e-01 9.34936523e-01
5.17173409e-01 -5.50213158e-01 -7.51750410e-01 6.52444005e-01
1.26345053e-01 -6.74100816e-01 1.17769229e+00 -3.79985601e-01
2.28187233e-01 -1.75718978e-01 5.96775115e-01 -7.27929533e-01
-4.40633684e-01 1.56317681e-01 4.10735339e-01 2.83068687e-01
7.73830056e-01 -2.67177701e-01 1.65170681e+00 3.87018025e-01
-2.76866525e-01 -9.10330355e-01 -8.44930410e-01 -4.59436744e-01
2.12382704e-01 -1.18767105e-01 2.02828214e-01 5.05082428e-01
-5.05801022e-01 -3.74212652e-01 3.64450812e-01 1.88310370e-01
7.19353557e-01 1.97809666e-01 2.26905063e-01 -1.15392768e+00
-2.45775491e-01 -6.03163838e-01 -1.30810186e-01 -8.41757894e-01
-5.21048307e-02 -1.41085112e+00 4.96441245e-01 -1.68786502e+00
8.25544000e-01 -3.81332219e-01 -4.28989947e-01 7.70101249e-01
-2.23932117e-01 8.70229959e-01 -3.63193631e-01 5.83143711e-01
-8.23562980e-01 2.07324978e-02 1.82693923e+00 -4.46534425e-01
3.17815058e-02 4.01740611e-01 -3.98379296e-01 1.07524192e+00
9.17361438e-01 -7.54389644e-01 -2.96161801e-01 -1.29902884e-01
-3.77457321e-01 1.81270484e-02 7.22802937e-01 -1.14187872e+00
3.24852735e-01 -9.46955457e-02 8.38373601e-01 -9.49013948e-01
-1.38195036e-02 -8.65193427e-01 -3.84137064e-01 1.11640024e+00
-5.31011760e-01 -7.53144383e-01 -1.00664064e-01 7.12425590e-01
-2.09467798e-01 -3.86670500e-01 8.73242021e-01 -7.91410506e-01
-2.53325135e-01 8.25910866e-01 -5.73347092e-01 -1.08016646e-02
1.20838559e+00 -5.58048636e-02 1.11605674e-01 -8.69190469e-02
-8.21552753e-01 2.42355019e-01 1.23057410e-01 -1.81896016e-01
5.96063673e-01 -1.23176742e+00 -8.67678046e-01 9.00202990e-03
1.76923096e-01 4.58870828e-01 -2.26551265e-01 1.09015489e+00
-7.94854522e-01 8.36540639e-01 2.33707920e-01 -1.09127736e+00
-1.30897629e+00 8.65558460e-02 8.49399924e-01 -7.25880504e-01
-5.71539998e-01 1.21788776e+00 4.16363060e-01 -6.94294512e-01
2.48286560e-01 -9.78778541e-01 3.85207124e-03 -5.39732039e-01
-2.84251906e-02 -1.92276776e-01 9.62430835e-02 -2.96388149e-01
-2.11862937e-01 3.56472045e-01 -1.90647215e-01 1.79061532e-01
8.27043235e-01 3.18187773e-01 -8.35599974e-02 -6.86416179e-02
1.08991849e+00 -3.84738535e-01 -1.14576972e+00 -5.17462492e-01
2.05624789e-01 4.06428287e-03 4.38799381e-01 -8.57306361e-01
-1.44752157e+00 6.34517491e-01 7.61897922e-01 1.65525358e-02
8.45635295e-01 5.86064100e-01 9.48913872e-01 7.93713212e-01
-8.77155289e-02 -6.05649114e-01 4.91089344e-01 1.69404984e-01
4.05473799e-01 -1.75737083e+00 -6.15529530e-02 -9.13328350e-01
-5.00662625e-01 1.25236976e+00 1.05129838e+00 -1.93726659e-01
7.51125753e-01 3.49237084e-01 3.07122320e-01 -3.90017450e-01
-5.09356797e-01 -8.04865122e-01 7.44733274e-01 3.29981685e-01
6.72494292e-01 3.11046660e-01 6.73291609e-02 6.64855123e-01
1.67993456e-01 8.09413940e-03 2.12355152e-01 9.83099818e-01
-9.13186789e-01 -1.05687487e+00 -4.16253656e-01 1.07409275e+00
-5.24627805e-01 -1.16943516e-01 -7.71841109e-01 9.98483002e-01
2.18768105e-01 4.29926991e-01 1.97932556e-01 1.96090922e-01
-1.68266445e-01 3.26593705e-02 2.92299360e-01 -1.27356303e+00
-8.74756217e-01 5.50003469e-01 -1.81793392e-01 -3.99388343e-01
-4.80333686e-01 -4.07827914e-01 -1.70804226e+00 5.40883958e-01
-3.93453985e-01 1.98227331e-01 3.33338082e-01 9.04098630e-01
-2.44458079e-01 8.95010054e-01 6.48923874e-01 -5.92360198e-01
-4.90854114e-01 -9.25822377e-01 -2.49098122e-01 2.01475635e-01
1.25840932e-01 -1.77084342e-01 -2.27900937e-01 1.29832685e-01] | [15.412027359008789, -2.1146183013916016] |
6bce541f-ccf9-4d9d-aab6-1483667f3dec | automated-vulnerability-detection-in-source-1 | 2303.07525 | null | https://arxiv.org/abs/2303.07525v1 | https://arxiv.org/pdf/2303.07525v1.pdf | Automated Vulnerability Detection in Source Code Using Quantum Natural Language Processing | One of the most important challenges in the field of software code audit is the presence of vulnerabilities in software source code. These flaws are highly likely ex-ploited and lead to system compromise, data leakage, or denial of ser-vice. C and C++ open source code are now available in order to create a large-scale, classical machine-learning and quantum machine-learning system for function-level vulnerability identification. We assembled a siz-able dataset of millions of open-source functions that point to poten-tial exploits. We created an efficient and scalable vulnerability detection method based on a deep neural network model Long Short Term Memory (LSTM), and quantum machine learning model Long Short Term Memory (QLSTM), that can learn features extracted from the source codes. The source code is first converted into a minimal intermediate representation to remove the pointless components and shorten the de-pendency. Therefore, We keep the semantic and syntactic information using state of the art word embedding algorithms such as Glove and fastText. The embedded vectors are subsequently fed into the classical and quantum convolutional neural networks to classify the possible vulnerabilities. To measure the performance, we used evaluation metrics such as F1 score, precision, re-call, accuracy, and total execution time. We made a comparison between the results derived from the classical LSTM and quantum LSTM using basic feature representation as well as semantic and syntactic represen-tation. We found that the QLSTM with semantic and syntactic features detects significantly accurate vulnerability and runs faster than its classical counterpart. | ['Zakirul Alam Bhuiya', 'Hossain Shahriar', 'Mst Shapna Akter'] | 2023-03-13 | null | null | null | null | ['vulnerability-detection'] | ['miscellaneous'] | [-1.81386396e-01 -1.38769776e-01 -1.04013175e-01 -4.29720730e-02
-1.01512611e+00 -9.14945662e-01 2.37194791e-01 5.25749385e-01
-3.11535269e-01 2.97349453e-01 -1.37995526e-01 -9.95366037e-01
2.21540287e-01 -1.18886495e+00 -8.60588014e-01 -3.22982252e-01
-3.74233454e-01 -4.23159063e-01 3.33491087e-01 -4.07049149e-01
6.84998751e-01 1.34151906e-01 -1.35324204e+00 4.53483880e-01
6.81569397e-01 8.57885003e-01 -2.02708974e-01 9.36123729e-01
-2.36267477e-01 7.37559915e-01 -7.98048139e-01 -8.27248454e-01
7.06140324e-02 2.08184011e-02 -9.26787376e-01 -9.93945777e-01
-7.42668957e-02 -1.61553398e-01 -5.44381142e-01 1.71820915e+00
4.13688004e-01 -4.22589600e-01 1.69028327e-01 -9.90763307e-01
-1.05081224e+00 7.27746427e-01 -1.85024247e-01 2.92746842e-01
5.00506580e-01 3.47404301e-01 1.03199971e+00 -4.04526949e-01
2.82204390e-01 1.11329281e+00 7.74732053e-01 6.25465930e-01
-8.81273448e-01 -5.49469590e-01 -5.73747575e-01 3.52528691e-01
-1.21506083e+00 -3.08803320e-02 5.71668088e-01 -4.81910139e-01
1.73365188e+00 -5.94187062e-03 2.71048117e-02 1.12943387e+00
1.01342511e+00 -3.52121592e-02 7.40324497e-01 -5.31895578e-01
2.64690906e-01 4.01374698e-01 5.70620775e-01 1.22454762e+00
2.38366038e-01 5.69071472e-01 -1.95081398e-01 -8.72620583e-01
5.14462125e-03 1.82707921e-01 1.91130284e-02 3.68399292e-01
-8.22920144e-01 1.03988373e+00 5.33854544e-01 7.22507298e-01
8.14211145e-02 5.09295464e-01 9.90848184e-01 7.65871286e-01
-8.61104354e-02 6.64681435e-01 -7.12766171e-01 -4.47288960e-01
-6.11411691e-01 -2.67267346e-01 7.98682749e-01 5.41719556e-01
1.01768565e+00 8.26173276e-02 4.88268398e-02 7.15946630e-02
3.99163604e-01 5.75871408e-01 7.49155998e-01 -4.25150067e-01
3.92136723e-01 7.72752464e-01 -4.73342359e-01 -1.11514652e+00
-4.62667979e-02 -1.84952870e-01 -4.43578392e-01 3.90346050e-01
3.94071154e-02 -4.02234755e-02 -5.92886627e-01 1.63962972e+00
-2.37900630e-01 1.75728619e-01 4.74521756e-01 1.24698438e-01
6.04678273e-01 6.40066266e-01 -9.84397233e-02 2.90369034e-01
1.35027850e+00 -4.08010483e-01 -5.42975783e-01 1.05257824e-01
1.35085583e+00 -6.91423833e-01 1.02281141e+00 1.80656731e-01
-4.14481848e-01 -3.12934548e-01 -1.43482077e+00 5.02484962e-02
-1.16294742e+00 -3.21623206e-01 5.76362491e-01 1.45561385e+00
-1.12006259e+00 1.00969744e+00 -6.88033938e-01 3.40948477e-02
2.90967852e-01 5.17807901e-01 -5.75022519e-01 4.22105044e-01
-1.59452069e+00 7.57993042e-01 6.39084339e-01 -2.39913315e-01
-1.25864387e+00 -4.41530496e-01 -1.05799687e+00 3.71316820e-01
8.53546783e-02 -1.16577163e-01 9.47283268e-01 -7.94848025e-01
-1.41525185e+00 7.53529429e-01 1.99409410e-01 -4.70847338e-01
-5.29663444e-01 9.35200304e-02 -8.25971544e-01 -1.93210378e-01
-1.79658517e-01 -1.09450869e-01 8.20975780e-01 -4.41415191e-01
-3.40150476e-01 -3.63432199e-01 3.11335266e-01 -8.51540387e-01
-8.96140039e-01 5.40802121e-01 4.02956665e-01 -1.68633252e-01
-5.15720010e-01 -6.79242432e-01 1.88911825e-01 -5.26113272e-01
-5.73967874e-01 -5.71569093e-02 1.04367220e+00 -9.51163709e-01
1.50658250e+00 -2.41584945e+00 -2.49368802e-01 2.85754770e-01
3.07498217e-01 6.84777677e-01 -2.38035068e-01 4.67128664e-01
-3.61157387e-01 5.95207810e-01 -3.90168339e-01 1.94342583e-01
1.86673939e-01 -5.92150651e-02 -7.49930501e-01 6.07517302e-01
2.18936354e-01 1.15774655e+00 -7.70431399e-01 -7.62860477e-02
-9.90707986e-03 3.43019128e-01 -4.49667066e-01 1.10838629e-01
-3.20635825e-01 -2.44248524e-01 -4.71441627e-01 7.95413554e-01
3.96378577e-01 -1.35432601e-01 -2.38726661e-01 1.59631476e-01
-1.44385472e-01 6.03667140e-01 -6.14138722e-01 1.62753963e+00
-7.49570012e-01 7.70833194e-01 -4.71403301e-01 -6.52010262e-01
9.79115367e-01 4.73370761e-01 -3.11914355e-01 -7.94210613e-01
3.34377110e-01 5.01689434e-01 8.09599683e-02 -6.26113951e-01
3.03801298e-01 1.01957589e-01 -5.28916061e-01 5.20087957e-01
2.31472269e-01 2.60345995e-01 -3.38204622e-01 1.19327798e-01
1.56207240e+00 -6.21437542e-02 1.53808475e-01 -2.44713530e-01
8.44380975e-01 -1.91701502e-01 3.43148887e-01 5.89411855e-01
-3.52431536e-01 3.85743659e-03 8.91687334e-01 -4.96123165e-01
-9.94971454e-01 -8.73708367e-01 -4.71101422e-03 8.57397616e-01
-1.91614807e-01 -7.69773543e-01 -1.12928820e+00 -9.68592227e-01
-2.28595287e-01 8.14184546e-01 -5.53152204e-01 -9.08489406e-01
-4.85136002e-01 -6.53315425e-01 1.35582089e+00 3.87759566e-01
2.80271202e-01 -1.09132338e+00 -6.96691573e-01 1.42715976e-01
3.42869818e-01 -7.12678432e-01 -2.49737859e-01 4.32756960e-01
-6.23951018e-01 -1.23006999e+00 6.77065924e-02 -5.24255335e-01
2.06310719e-01 -4.10104901e-01 8.51817787e-01 2.61594236e-01
-5.20661771e-01 -1.96390659e-01 -4.14879352e-01 5.67997247e-02
-9.88660574e-01 -3.41191655e-03 -1.62389986e-02 -3.00821811e-01
7.35428810e-01 -4.47366863e-01 -2.23602340e-01 -2.02642873e-01
-1.05993724e+00 -9.16419744e-01 5.81177652e-01 8.08990180e-01
5.21056466e-02 2.22720876e-01 4.08160418e-01 -6.20270193e-01
6.84029758e-01 -4.08931881e-01 -8.81622136e-01 3.85543704e-01
-5.86984634e-01 6.16492748e-01 1.11032915e+00 -1.88344643e-01
-6.49089277e-01 -3.62684190e-01 -6.30782843e-01 -2.57722676e-01
4.17735055e-02 6.67687237e-01 -2.10950688e-01 -7.56394625e-01
1.03274763e+00 2.61852831e-01 -2.04909429e-01 -3.01889062e-01
3.64732236e-01 9.18374121e-01 3.23139608e-01 -5.44390440e-01
1.08788180e+00 1.41170412e-01 -9.39207077e-02 -5.63951313e-01
-3.04926872e-01 2.44640484e-01 -4.12462533e-01 3.63693058e-01
9.92985606e-01 -3.04512203e-01 -9.94386494e-01 6.46116555e-01
-1.48373044e+00 5.67365177e-02 -8.65764264e-03 5.84449992e-02
-1.87315434e-01 9.08695519e-01 -9.37771797e-01 -7.43896782e-01
-7.07519948e-01 -1.54253531e+00 8.13395679e-01 9.95901078e-02
1.26070261e-01 -1.00713933e+00 3.82038862e-01 -3.12484622e-01
6.02508664e-01 2.99068689e-01 1.65001619e+00 -8.04603815e-01
-4.54333395e-01 -6.92676365e-01 -2.49677062e-01 7.57003665e-01
-1.09010801e-01 2.62230247e-01 -1.26677942e+00 -4.12149757e-01
2.49316871e-01 -3.70843440e-01 6.49698973e-01 -2.86547095e-01
1.04367244e+00 -3.02604496e-01 -7.08506107e-02 8.14133227e-01
1.71181583e+00 3.17770243e-01 9.19583201e-01 4.30258036e-01
5.70864201e-01 2.93908566e-01 6.93750102e-03 9.50991809e-02
9.57788751e-02 2.46298149e-01 6.70915306e-01 4.88870710e-01
3.33253652e-01 -3.10043156e-01 1.06001687e+00 8.07039022e-01
4.23693120e-01 1.86532855e-01 -1.18279278e+00 3.26755971e-01
-1.32926333e+00 -9.22483981e-01 -1.38062775e-01 2.46613097e+00
6.66983902e-01 3.66426289e-01 -2.39017233e-01 3.75015080e-01
6.48431480e-01 -2.96118166e-02 -1.74581900e-01 -1.16896009e+00
3.32888700e-02 6.77256703e-01 8.03243518e-01 3.59532714e-01
-1.05089474e+00 9.85770345e-01 5.52953529e+00 1.05450225e+00
-1.38943648e+00 5.95560491e-01 2.78128028e-01 4.50406194e-01
-3.43965828e-01 4.30858523e-01 -5.82038164e-01 4.52088982e-01
2.00465417e+00 -2.87331998e-01 4.66885537e-01 9.30922091e-01
-4.72338706e-01 1.91842765e-01 -9.67225850e-01 7.40364969e-01
-6.27936125e-02 -1.09792840e+00 -2.85370111e-01 1.01016864e-01
2.26905420e-01 2.66321719e-01 2.52928972e-01 6.03706658e-01
1.02346018e-01 -1.16253793e+00 6.24245703e-01 3.78735214e-01
9.79417682e-01 -9.24185932e-01 1.02163076e+00 2.39466041e-01
-1.13033652e+00 -6.70357168e-01 -5.69288194e-01 -1.04772598e-01
-2.47964039e-01 3.26142609e-01 -4.24952179e-01 5.83481014e-01
6.92226171e-01 4.81843621e-01 -8.91043186e-01 6.51460111e-01
-3.68563056e-01 5.13505340e-01 -8.34636241e-02 -3.57912749e-01
4.53672379e-01 1.61503866e-01 3.76750886e-01 1.02724385e+00
4.41504598e-01 -5.49864113e-01 -2.58487016e-01 1.39584100e+00
5.06452583e-02 -2.17764005e-01 -8.03731620e-01 -7.57982731e-01
2.33419105e-01 1.10584319e+00 -4.17953730e-01 -3.12156737e-01
-6.80617630e-01 8.37843060e-01 2.73319244e-01 8.83129984e-02
-9.54771996e-01 -1.12208533e+00 5.39869428e-01 -3.39086890e-01
9.94281396e-02 -2.47484654e-01 -3.18233848e-01 -1.30169666e+00
1.55298010e-01 -8.41250062e-01 4.09015745e-01 -1.12227432e-01
-1.06464660e+00 1.02134454e+00 -3.97564173e-01 -8.35437477e-01
-1.53841123e-01 -9.08024848e-01 -9.67618644e-01 1.24584925e+00
-1.52877903e+00 -8.80343735e-01 3.44890177e-01 4.14767385e-01
-2.49730781e-01 -6.08354509e-01 1.40432072e+00 4.42332506e-01
-7.61402190e-01 9.72868025e-01 2.08536282e-01 3.79783928e-01
1.87259600e-01 -1.14574671e+00 9.71276402e-01 1.12460303e+00
-8.05718899e-02 1.10850978e+00 4.41304058e-01 -7.62672305e-01
-1.73386693e+00 -1.06266594e+00 9.09872472e-01 -6.98739886e-01
1.11923492e+00 -6.10586345e-01 -1.05350256e+00 5.20374656e-01
-7.16579482e-02 8.36692378e-02 7.67542005e-01 -2.37729862e-01
-9.84273791e-01 2.26026103e-01 -1.29867506e+00 1.15955904e-01
2.95368373e-01 -1.56257534e+00 -7.83694088e-01 1.81725845e-01
1.14327800e+00 1.05399728e-01 -8.08946192e-01 6.47019818e-02
4.22335654e-01 -1.10659671e+00 8.01600218e-01 -5.55923581e-01
2.94411272e-01 -9.41479057e-02 -4.50869501e-01 -7.51117289e-01
-5.04379347e-02 -7.21586704e-01 -1.96843505e-01 9.93336022e-01
6.45668328e-01 -1.04891634e+00 5.42759717e-01 4.24714535e-01
-1.96987361e-01 -4.20508444e-01 -1.16848898e+00 -9.10126984e-01
5.44682264e-01 -5.94146192e-01 7.01736212e-01 9.18315589e-01
3.68339092e-01 -7.47296065e-02 -4.62396927e-02 3.91362846e-01
6.21408463e-01 -1.46074686e-02 2.12958395e-01 -8.30074370e-01
-5.10045230e-01 -5.18353224e-01 -9.02725756e-01 -1.94360092e-01
6.16580307e-01 -1.14361942e+00 -2.94619113e-01 -6.67783260e-01
1.94161758e-01 2.17948463e-02 -6.67938113e-01 6.61344051e-01
-6.12947308e-02 -6.47331998e-02 -2.55184561e-01 -9.51650962e-02
-2.40758419e-01 4.50520903e-01 2.67136902e-01 -3.97526175e-01
1.65530480e-02 -1.81152597e-01 -4.27513421e-01 5.40380418e-01
7.30322242e-01 -9.55490530e-01 9.11142156e-02 -2.64957517e-01
6.71543956e-01 1.48101449e-01 5.26846588e-01 -1.03611588e+00
5.29436916e-02 2.58012414e-01 -2.75787055e-01 -1.03789188e-01
-6.14675228e-03 -4.27083701e-01 -3.06973100e-01 1.08786762e+00
-2.65550185e-02 3.19987565e-01 3.84570956e-01 3.79238755e-01
-2.20074967e-01 -7.93523550e-01 7.32984841e-01 -1.72071457e-01
-7.53641784e-01 2.51546919e-01 -2.88297355e-01 -2.42374837e-01
1.04709542e+00 -2.21771002e-02 -6.88853920e-01 2.31628850e-01
-3.30899864e-01 -2.87635207e-01 6.33049965e-01 6.14578784e-01
7.74679601e-01 -9.17320073e-01 -2.94820577e-01 5.52201450e-01
2.31797710e-01 -1.03310156e+00 3.38595748e-01 3.51385891e-01
-6.62317276e-01 7.06370831e-01 -2.95310378e-01 -2.45865923e-03
-1.07606173e+00 1.08633101e+00 6.14397287e-01 -9.12022963e-02
-4.01499957e-01 7.71460712e-01 -1.69536889e-01 -7.07937956e-01
-6.82292059e-02 -3.96287739e-01 -4.09437604e-02 -5.64159572e-01
7.64843941e-01 2.68169641e-01 2.42911622e-01 -6.53781414e-01
-6.66306615e-01 4.56844687e-01 1.04035437e-01 9.71812531e-02
1.15798318e+00 5.60107291e-01 -8.05211604e-01 3.30758452e-01
1.78285217e+00 -2.06743300e-01 -2.28156731e-01 -9.58337113e-02
4.59068596e-01 -3.88717353e-01 2.89383322e-01 -7.42103279e-01
-7.26452172e-01 1.52712202e+00 9.94018734e-01 4.67903197e-01
7.48764753e-01 -9.97141525e-02 1.16833341e+00 6.68310106e-01
7.32527137e-01 -5.79387367e-01 -5.71421236e-02 7.62887776e-01
4.98795919e-02 -9.57846165e-01 -6.08515680e-01 6.23240359e-02
1.79557025e-01 1.51310170e+00 4.33935523e-01 -2.46805623e-01
6.78035617e-01 4.19693559e-01 -2.04850972e-01 -4.31603670e-01
-6.03267550e-01 1.16813071e-01 8.92644078e-02 4.76313353e-01
4.49969649e-01 5.16926013e-02 -2.65629590e-02 7.07322657e-01
-1.01408012e-01 -2.80192584e-01 9.42514420e-01 1.14734769e+00
-5.59172988e-01 -1.40911937e+00 -4.54041094e-01 3.44068348e-01
-8.87373209e-01 -5.49019158e-01 -3.50561857e-01 1.84505284e-01
1.35432988e-01 9.48555648e-01 -4.41684842e-01 -1.02421892e+00
6.37768731e-02 2.59567231e-01 1.97190627e-01 -7.74460196e-01
-9.71789598e-01 -6.37623370e-01 -2.61115342e-01 -8.16939116e-01
3.95705551e-01 -7.77872056e-02 -1.37507224e+00 -5.61343431e-01
-5.92385113e-01 2.44206235e-01 1.00498867e+00 9.06536281e-01
6.40460730e-01 7.35285044e-01 6.92484975e-01 -3.46621096e-01
-1.00530815e+00 -7.16618657e-01 -3.60671133e-01 1.84803098e-01
4.99763995e-01 -4.60911900e-01 -6.07318163e-01 -3.49683702e-01] | [7.054675102233887, 7.772696495056152] |
661156fd-6f2b-4a0e-af20-70701f27216f | pathology-aware-generative-adversarial | 2106.01915 | null | https://arxiv.org/abs/2106.01915v1 | https://arxiv.org/pdf/2106.01915v1.pdf | Pathology-Aware Generative Adversarial Networks for Medical Image Augmentation | Convolutional Neural Networks (CNNs) can play a key role in Medical Image Analysis under large-scale annotated datasets. However, preparing such massive dataset is demanding. In this context, Generative Adversarial Networks (GANs) can generate realistic but novel samples, and thus effectively cover the real image distribution. In terms of interpolation, the GAN-based medical image augmentation is reliable because medical modalities can display the human body's strong anatomical consistency at fixed position while clearly reflecting inter-subject variability; thus, we propose to use noise-to-image GANs (e.g., random noise samples to diverse pathological images) for (i) medical Data Augmentation (DA) and (ii) physician training. Regarding the DA, the GAN-generated images can improve Computer-Aided Diagnosis based on supervised learning. For the physician training, the GANs can display novel desired pathological images and help train medical trainees despite infrastructural/legal constraints. This thesis contains four GAN projects aiming to present such novel applications' clinical relevance in collaboration with physicians. Whereas the methods are more generally applicable, this thesis only explores a few oncological applications. | ['Changhee Han'] | 2021-06-03 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 5.84906042e-01 6.12394750e-01 -2.59827197e-01 -1.03417806e-01
-7.19588459e-01 -2.33715296e-01 2.59337872e-01 -1.85323939e-01
-2.51898557e-01 1.02418756e+00 5.37218191e-02 -3.00002962e-01
1.32816032e-01 -1.00474238e+00 -7.19589412e-01 -1.16824484e+00
3.97518188e-01 5.01118004e-01 -2.13911220e-01 -1.86352208e-01
-2.53895581e-01 7.37329721e-01 -1.19411850e+00 3.40789527e-01
1.17614245e+00 8.69857430e-01 1.40406519e-01 5.97313106e-01
-6.20174743e-02 7.77890384e-01 -9.61697102e-01 -5.03833652e-01
1.13019720e-01 -9.40732419e-01 -4.50202107e-01 -7.53487498e-02
1.62919551e-01 -2.58179247e-01 -2.75157392e-01 1.07485390e+00
8.07625651e-01 -1.93161815e-01 5.38689256e-01 -1.11889887e+00
-8.77500296e-01 3.79365474e-01 -4.51795548e-01 -1.26205593e-01
1.42673805e-01 4.62304324e-01 9.22186822e-02 -3.65970254e-01
6.81562662e-01 6.90834522e-01 7.76812077e-01 1.25384307e+00
-1.08386755e+00 -5.85882306e-01 -2.34345436e-01 -7.81663656e-02
-1.03832853e+00 -1.21406071e-01 9.16619599e-01 -4.31237966e-01
4.29861397e-02 7.45105445e-01 9.76296365e-01 1.49425626e+00
5.65794170e-01 7.33276963e-01 1.13041055e+00 -2.12384135e-01
1.47291169e-01 2.60854572e-01 -6.09451532e-01 5.51584125e-01
1.52023897e-01 1.71926335e-01 -1.16882443e-01 -1.46458922e-02
1.19417930e+00 2.96099603e-01 -5.59397459e-01 -4.19449769e-02
-1.25020647e+00 8.03745329e-01 7.96845019e-01 5.64517081e-01
-5.92284441e-01 2.09099084e-01 4.85835135e-01 2.50241496e-02
1.98679984e-01 4.40247267e-01 -6.35936903e-03 5.23165911e-02
-9.00831461e-01 -1.13824029e-02 3.20894688e-01 6.63433492e-01
2.64702052e-01 5.66490114e-01 -4.58806932e-01 6.95389926e-01
5.42262644e-02 5.45902014e-01 9.68451381e-01 -7.74196565e-01
1.04984798e-01 4.95662838e-01 -2.59646147e-01 -8.77397180e-01
-3.54873776e-01 -8.69950235e-01 -1.81864357e+00 3.47501069e-01
3.16465408e-01 -5.14626317e-02 -1.20414090e+00 1.66728222e+00
4.41987485e-01 3.00190728e-02 6.01078719e-02 8.10736358e-01
1.23087490e+00 3.91629487e-01 2.70747125e-01 -3.74241114e-01
1.26376069e+00 -8.25681329e-01 -1.15499520e+00 -1.39062300e-01
5.77696443e-01 -5.18918753e-01 1.20582521e+00 3.06229979e-01
-1.33586156e+00 -6.45003140e-01 -7.47547388e-01 2.67219126e-01
-1.18211687e-01 1.46298841e-01 7.25936353e-01 8.71717036e-01
-9.51956213e-01 3.58953595e-01 -9.43193734e-01 3.16628926e-02
9.16384161e-01 2.23949268e-01 -4.23853904e-01 -2.14730874e-01
-1.21340513e+00 6.87349260e-01 1.90119237e-01 3.87667567e-01
-8.87417972e-01 -1.05205154e+00 -8.96288633e-01 -3.19362402e-01
2.16954444e-02 -1.18437994e+00 8.27981591e-01 -1.37556362e+00
-1.51372111e+00 1.16887867e+00 6.02330118e-02 -3.44009668e-01
9.75872219e-01 2.99556196e-01 -4.56492603e-01 2.31158391e-01
-2.59947423e-02 7.64308870e-01 7.33368635e-01 -1.49402738e+00
-1.54989562e-03 -3.35774809e-01 -3.27325255e-01 8.32896307e-02
-2.92301588e-02 -3.09346676e-01 -1.51520565e-01 -1.20179367e+00
-1.68985911e-02 -8.93936634e-01 -6.37702167e-01 3.27116907e-01
-5.42539299e-01 3.93158138e-01 7.03175187e-01 -9.09846008e-01
9.66368437e-01 -2.12328696e+00 2.04049293e-02 1.58020258e-01
4.15850788e-01 3.55698407e-01 9.70230401e-02 -6.99934065e-02
-2.24149868e-01 3.48165125e-01 -4.35898632e-01 -1.73564032e-01
-4.16373253e-01 2.47104302e-01 2.86527872e-01 4.79136765e-01
-3.07088476e-02 1.57630730e+00 -7.72259533e-01 -6.09334707e-01
3.73438239e-01 5.49171090e-01 -4.46721435e-01 3.37045521e-01
-3.47108506e-02 1.29470301e+00 -4.17624623e-01 7.98336446e-01
7.13976085e-01 -1.41920701e-01 -7.16102496e-02 -3.12206209e-01
6.58767521e-01 -4.79246795e-01 -4.89692301e-01 1.73533535e+00
-5.86305261e-01 3.98981839e-01 2.60829777e-01 -9.77653146e-01
8.50569546e-01 6.09273612e-01 7.13075459e-01 -7.88905263e-01
1.83948100e-01 1.35646656e-01 1.77731365e-01 -5.88392913e-01
6.28109798e-02 -7.11058080e-01 1.23246290e-01 7.31291622e-03
-2.66649842e-01 -1.80381298e-01 -2.76500225e-01 -8.85396600e-02
7.84962535e-01 -2.02804193e-01 9.95608643e-02 -2.53303796e-02
3.90985757e-01 -1.13496773e-01 4.69688743e-01 4.67145085e-01
-1.59607887e-01 9.08198535e-01 3.77201408e-01 -5.99813521e-01
-1.10284805e+00 -1.07970107e+00 -1.43803610e-02 1.80314541e-01
-7.87309557e-02 2.75175869e-01 -9.12473857e-01 -7.95821548e-01
-3.17255735e-01 5.48799098e-01 -1.04838586e+00 -4.37410384e-01
-5.10969222e-01 -7.83694148e-01 6.95495248e-01 6.03224993e-01
5.44994354e-01 -1.43634117e+00 -5.15747726e-01 2.48073474e-01
-2.36445934e-01 -8.79478931e-01 -2.32935384e-01 -6.31431341e-02
-8.99403691e-01 -9.57650304e-01 -1.49194658e+00 -7.95399666e-01
1.13320208e+00 -3.38609517e-01 1.19049847e+00 1.85859382e-01
-5.81054628e-01 3.13700527e-01 -3.48782599e-01 -5.52837312e-01
-1.10035026e+00 -5.02609238e-02 -2.05366492e-01 -1.09901965e-01
-2.29311541e-01 -6.37861490e-01 -8.68861973e-01 3.58736992e-01
-1.28589129e+00 3.22513521e-01 8.08176458e-01 1.30613875e+00
9.78726029e-01 1.75403920e-03 6.12861753e-01 -1.29400551e+00
6.31046474e-01 -2.47486368e-01 -7.61391595e-02 2.77380258e-01
-2.70910084e-01 -4.31853473e-01 8.86337459e-01 -6.03507221e-01
-1.07980633e+00 -5.24740294e-02 -6.30103409e-01 -6.91346407e-01
-4.90274042e-01 4.64415967e-01 -2.70433873e-01 -2.44394466e-01
9.62200582e-01 4.79873836e-01 3.48312140e-01 -4.21718061e-02
7.52962604e-02 2.91642368e-01 6.78020358e-01 -3.68570507e-01
7.51020133e-01 4.58105564e-01 2.06619665e-01 -5.63616514e-01
-4.79321748e-01 2.41354361e-01 -3.71447563e-01 -3.79748046e-01
1.06904447e+00 -7.55455792e-01 -5.60060322e-01 5.65826178e-01
-9.81596589e-01 -5.56835949e-01 -9.98163640e-01 5.51464617e-01
-5.87166488e-01 1.44159406e-01 -7.66043603e-01 -4.91448104e-01
-5.97124279e-01 -1.37019372e+00 1.02173316e+00 4.65229899e-01
-1.66827574e-01 -1.23181748e+00 -1.45235255e-01 4.16119516e-01
6.27940536e-01 1.09249437e+00 8.59752536e-01 -1.73917010e-01
-3.65637571e-01 -4.79242951e-01 1.61316007e-01 5.06982267e-01
4.08793479e-01 -2.19147921e-01 -1.04248202e+00 -3.53846073e-01
2.15411857e-01 -2.27393731e-01 3.81491840e-01 8.52709115e-01
1.74043107e+00 -4.08280581e-01 -2.04103261e-01 8.97046089e-01
1.20827901e+00 4.37322706e-01 1.36895108e+00 4.85762730e-02
8.50539148e-01 5.01443803e-01 4.37384784e-01 2.38748446e-01
-2.22150251e-01 3.80907536e-01 6.15533173e-01 -8.76273215e-01
-4.25893962e-01 -3.46340269e-01 -8.26078430e-02 5.18255949e-01
-3.12993973e-01 -2.59960413e-01 -6.35370135e-01 3.48687112e-01
-1.22717702e+00 -8.80093515e-01 -1.93227947e-01 2.01275921e+00
8.96335125e-01 -3.63032296e-02 -1.53168455e-01 1.99839324e-01
7.30713010e-01 -2.02343702e-01 -7.58321464e-01 -2.32034534e-01
-1.31802186e-01 5.09959579e-01 4.00928110e-01 3.19832824e-02
-6.42099380e-01 4.41787392e-01 6.48559046e+00 1.17550647e+00
-1.41369545e+00 2.00736374e-01 1.24302280e+00 1.20442167e-01
-7.30200529e-01 -6.53431118e-01 -2.14697197e-01 6.80963159e-01
7.06500709e-01 1.09225586e-01 1.24377757e-01 7.76336253e-01
2.55844146e-01 8.85631666e-02 -6.93967879e-01 1.11746716e+00
-1.07885614e-01 -1.62202716e+00 2.11944878e-01 2.66327232e-01
9.05994475e-01 -5.94620347e-01 4.29398328e-01 8.23962763e-02
-5.07270098e-02 -1.53951502e+00 2.86386818e-01 7.00106680e-01
1.50724304e+00 -7.73332059e-01 1.09568810e+00 3.33203793e-01
-6.52682364e-01 3.49736363e-01 -7.42908344e-02 5.84812462e-01
2.65044421e-01 7.96810746e-01 -6.67543948e-01 7.43219376e-01
4.96361822e-01 4.49563384e-01 -4.76260453e-01 8.07434857e-01
-2.82793432e-01 5.71688473e-01 -5.69158904e-02 2.48144060e-01
-1.20849255e-03 -2.09852755e-01 4.09259200e-01 7.99611449e-01
3.99502426e-01 -5.30971307e-03 -1.77498609e-01 1.08268785e+00
-1.36984482e-01 1.03131160e-01 -5.89310944e-01 8.79971683e-02
-3.92659977e-02 1.27975082e+00 -8.17792892e-01 -3.51225674e-01
1.04671791e-01 1.19282413e+00 -3.85359347e-01 3.23881596e-01
-1.00670099e+00 -1.00746922e-01 2.53852159e-01 4.00690496e-01
-3.49060267e-01 3.40385258e-01 -6.07595265e-01 -8.61381173e-01
-3.22070532e-02 -1.18492317e+00 2.86511451e-01 -8.50426435e-01
-1.20943534e+00 8.22606504e-01 -3.88114631e-01 -1.34865201e+00
-2.10999668e-01 -3.91892612e-01 -7.48334467e-01 9.66417611e-01
-1.20942140e+00 -1.54275310e+00 -8.54861081e-01 7.37174213e-01
3.42479408e-01 -1.85027406e-01 1.09293354e+00 4.58965749e-01
-4.16705340e-01 9.70473766e-01 6.31316565e-03 1.48770109e-01
5.62241018e-01 -1.09905076e+00 1.25010669e-01 4.28027779e-01
-2.18778268e-01 4.23815250e-01 5.16230583e-01 -6.64055884e-01
-1.03788996e+00 -1.23103392e+00 8.62603858e-02 -1.23809770e-01
8.05040356e-03 3.73621695e-02 -8.71094227e-01 5.03440619e-01
2.01317593e-01 3.07508796e-01 9.23720300e-01 -5.88825762e-01
3.63927186e-01 -4.25075702e-02 -1.77729678e+00 6.60252512e-01
7.31850982e-01 -1.73735008e-01 6.60379305e-02 5.03276229e-01
5.87580144e-01 -9.76501048e-01 -1.19524348e+00 6.25443399e-01
3.95243138e-01 -7.66477346e-01 9.79577303e-01 -6.19774282e-01
8.44609559e-01 -1.35711536e-01 3.24611694e-01 -1.36251366e+00
-6.18840568e-02 -5.39358139e-01 1.14401303e-01 9.86111999e-01
2.20021978e-01 -6.75986946e-01 1.21164489e+00 6.02117360e-01
-3.15243006e-01 -1.03759396e+00 -9.69789982e-01 -6.38693392e-01
2.25313038e-01 -1.35005340e-01 5.36796153e-01 1.18743789e+00
-4.10807103e-01 -3.21486413e-01 -4.92663920e-01 -2.15867653e-01
4.59223270e-01 -4.05009776e-01 8.49078238e-01 -8.32826376e-01
-2.59323210e-01 -4.83217388e-01 -6.04219317e-01 -6.23919010e-01
-5.59180230e-02 -7.72695899e-01 -2.23488018e-01 -1.64333642e+00
-3.73058813e-03 -6.99228585e-01 -1.08791925e-01 4.50655878e-01
-2.72419542e-01 6.29365921e-01 -9.65953805e-03 1.41126886e-01
1.39585391e-01 5.87364912e-01 2.17206168e+00 -4.11793143e-01
-9.94533859e-03 3.46440792e-01 -8.20040345e-01 4.33131188e-01
8.83579075e-01 -2.65908718e-01 -3.02236527e-01 -1.00853562e-01
-8.00938811e-03 2.95812309e-01 5.41396976e-01 -1.06341362e+00
1.81493610e-02 -1.32925928e-01 7.83600569e-01 -3.59490126e-01
1.24689147e-01 -1.01515973e+00 8.37564290e-01 1.02513134e+00
-1.69881999e-01 -1.32074386e-01 3.55666727e-01 3.14383686e-01
-5.29735863e-01 -1.76940158e-01 1.15818632e+00 -4.97224152e-01
-1.39728218e-01 4.67309862e-01 -3.20575953e-01 1.43950805e-02
1.39539003e+00 -4.62441057e-01 -2.14218497e-02 -6.47926748e-01
-9.30299938e-01 -1.80192068e-01 5.73838055e-01 -1.33924469e-01
7.41671443e-01 -1.59334302e+00 -7.31987417e-01 1.99978292e-01
-1.49667665e-01 5.54990053e-01 1.01281142e+00 1.01424825e+00
-9.17670906e-01 -5.85086606e-02 -4.14999098e-01 -6.49499297e-01
-1.10316849e+00 6.38988495e-01 6.21491790e-01 -6.35134101e-01
-6.31415486e-01 8.32174838e-01 5.24724305e-01 -3.41700196e-01
3.01527809e-02 -1.42776668e-01 -4.85894717e-02 -3.03507835e-01
4.69599783e-01 1.27224308e-02 1.34045973e-01 -2.93028861e-01
-5.85911684e-02 3.20705295e-01 7.97005519e-02 3.41957480e-01
1.12197828e+00 2.47614145e-01 6.07660525e-02 1.22610599e-01
1.00660563e+00 1.58569500e-01 -1.06883645e+00 1.32030219e-01
-8.21945906e-01 -4.60248709e-01 -1.49628207e-01 -9.67688680e-01
-1.80548108e+00 8.78646135e-01 8.91210318e-01 1.64301042e-02
1.53344166e+00 -1.37906745e-01 8.57904911e-01 -3.74884158e-01
1.89434499e-01 -6.07280076e-01 1.33497313e-01 -3.27160090e-01
1.01534855e+00 -1.07046545e+00 -1.46131264e-02 -5.53524494e-01
-9.64645028e-01 1.05328608e+00 6.17952645e-01 1.48633271e-01
2.95581132e-01 3.93783510e-01 3.28117192e-01 -1.33718401e-01
-1.44257573e-02 2.48099118e-01 3.06868345e-01 1.05164909e+00
3.81257713e-01 1.66575640e-01 -1.98946372e-01 5.80105960e-01
-3.43083739e-01 5.07403128e-02 3.83395255e-01 6.64449036e-01
2.44163141e-01 -1.06963074e+00 -4.52918977e-01 6.35512948e-01
-6.32814646e-01 -1.21979125e-01 -1.53192550e-01 8.23381782e-01
3.32971871e-01 3.77351910e-01 -1.91575289e-01 -1.83697194e-01
2.56059349e-01 -2.73060501e-01 5.64135611e-01 -3.54623705e-01
-8.10393751e-01 -3.95297678e-03 -3.30705911e-01 -3.64109635e-01
-4.45421368e-01 -2.50863969e-01 -9.44400728e-01 -3.93744558e-01
-9.25151780e-02 -2.56138798e-02 7.44442582e-01 5.54574311e-01
1.03771232e-01 1.25058448e+00 5.19281209e-01 -6.14705861e-01
-1.47418126e-01 -9.67529416e-01 -7.00274348e-01 6.08813405e-01
3.94156910e-02 -2.77367860e-01 -2.59074718e-01 1.54936582e-01] | [14.090250968933105, -1.994767665863037] |
2ecde487-386e-48c1-9f04-d871d5354d0a | on-the-current-state-of-reproducibility-and | null | null | https://openreview.net/forum?id=f6VUHB7dzXU | https://openreview.net/pdf?id=f6VUHB7dzXU | On the current state of reproducibility and reporting of uncertainty for Aspect-based Sentiment Analysis | For the latter part of the past decade, Aspect-Based Sentiment Analysis has been a field of great interest within Natural Language Processing. Supported by the Semantic Evaluation Conferences in 2014 -- 2016, a variety of methods has been developed competing in improving performances on benchmark data sets. Exploiting the transformer architecture behind BERT, results improved rapidly and efforts in this direction still continue today. Our contribution to this body of research is a holistic comparison of six different architectures which achieved (near) state-of-the-art results at some point in time. We utilize a broad spectrum of five benchmark data sets and introduce a fixed setting with respect to the pre-processing, the train/validation splits, the performance measures and the quantification of uncertainty. Overall, our findings are two-fold: First, we find that the results reported in the scientific articles are hardly reproducible, since in our experiments the observed performance (most of the time) fell short of the reported one. Second, the results are burdened with notable uncertainty (depending on the data splits) which is why a reporting of uncertainty measures is crucial. | ['Anonymous'] | 2021-10-16 | null | null | null | acl-arr-october-2021-10 | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [-4.96714897e-02 2.62816817e-01 -1.22151338e-01 -6.86755955e-01
-8.40384901e-01 -7.37238169e-01 1.02605021e+00 6.88060582e-01
-6.17839217e-01 7.02971160e-01 3.29915702e-01 -9.54444055e-03
-3.11991394e-01 -7.25731432e-01 -4.37809467e-01 -4.76615489e-01
7.74721131e-02 6.31288171e-01 1.03098884e-01 -4.40005690e-01
6.32987380e-01 1.33202493e-01 -1.89388800e+00 3.08756471e-01
7.13272810e-01 1.25680721e+00 -6.43259346e-01 2.79333025e-01
-4.93259132e-01 5.06724179e-01 -6.40195906e-01 -9.77584839e-01
8.49952325e-02 -1.05745278e-01 -9.36390936e-01 -6.84051067e-02
1.27280191e-01 5.12144983e-01 3.32835793e-01 1.02034688e+00
4.11657900e-01 1.18998326e-01 6.23831451e-01 -1.17932749e+00
-3.35759521e-01 8.56032550e-01 -4.94174600e-01 1.17280737e-01
4.46061492e-01 -2.78952777e-01 1.29652429e+00 -1.03453827e+00
6.95440829e-01 1.00997579e+00 7.23607838e-01 2.62492776e-01
-9.19840872e-01 -3.76928717e-01 2.94591278e-01 2.18909249e-01
-1.18990862e+00 -3.07364434e-01 6.93561554e-01 -4.60615814e-01
1.19305623e+00 1.60154760e-01 4.21755135e-01 1.06299186e+00
3.73390675e-01 5.78041255e-01 1.44829500e+00 -5.60982525e-01
6.14541233e-01 6.27929270e-01 5.35283506e-01 1.90376371e-01
4.70550954e-01 -5.89233041e-02 -6.53073549e-01 -1.77048832e-01
-3.13541353e-01 -4.36522812e-01 -6.04208596e-02 -3.40765715e-01
-9.64993894e-01 1.01757181e+00 2.82850936e-02 7.87831247e-01
-1.87619835e-01 -2.74602920e-01 8.23301554e-01 4.71184820e-01
9.37652707e-01 6.35877430e-01 -8.82352948e-01 -4.30119187e-01
-1.24824262e+00 2.64985204e-01 1.12938702e+00 5.71118057e-01
4.04270589e-01 -1.96464330e-01 -5.62012270e-02 5.53278446e-01
2.42543191e-01 2.16216549e-01 5.50691962e-01 -6.09508991e-01
5.91628194e-01 6.74552560e-01 9.53032151e-02 -9.47674632e-01
-6.44544423e-01 -5.36324382e-01 -5.59828460e-01 2.10093945e-01
4.83197123e-01 -1.69431657e-01 -9.50073242e-01 1.50477660e+00
1.52341917e-01 -4.08853710e-01 1.26986325e-01 6.46241426e-01
5.72912574e-01 4.18480754e-01 1.39198020e-01 -1.17168978e-01
1.65027010e+00 -7.94957042e-01 -7.66998708e-01 -3.34914178e-01
3.68426025e-01 -1.00791156e+00 9.93611157e-01 7.95606494e-01
-9.94421959e-01 -2.74027556e-01 -1.24849975e+00 2.53093064e-01
-1.00208974e+00 -8.12774673e-02 7.57727385e-01 8.09922218e-01
-8.89310956e-01 7.60962248e-01 -6.49783075e-01 -5.00586748e-01
2.33446553e-01 8.85633677e-02 -3.90147179e-01 1.85857490e-01
-1.41298795e+00 1.13848150e+00 3.06681126e-01 -5.71826771e-02
-3.28155130e-01 -7.82609880e-01 -7.23774016e-01 2.88020703e-03
4.90571052e-01 -5.76045632e-01 1.11156273e+00 -9.30219173e-01
-1.50609577e+00 1.18260860e+00 1.99534837e-02 -5.16716123e-01
6.77311420e-01 -4.44833249e-01 -5.02870739e-01 -2.79397309e-01
2.21224219e-01 3.11185181e-01 7.03725278e-01 -1.15025520e+00
-6.52057588e-01 -6.52434289e-01 6.18240833e-02 5.54285757e-02
-8.19805115e-02 3.90899062e-01 -3.10049713e-01 -6.54571414e-01
-9.19137970e-02 -9.02727842e-01 -3.17262888e-01 -6.47502780e-01
-2.77631909e-01 -2.89699703e-01 2.99659282e-01 -2.26985902e-01
1.37457359e+00 -2.04193902e+00 1.56775966e-01 4.15959954e-02
-1.27278477e-01 9.14177820e-02 2.92311639e-01 6.45583510e-01
-3.23802471e-01 3.67640793e-01 -6.33227944e-01 -7.45882809e-01
1.82636157e-01 -9.71502513e-02 -5.42971849e-01 3.80581737e-01
2.76317745e-01 6.60152256e-01 -6.97923422e-01 -2.59979874e-01
2.20409915e-01 3.81353140e-01 -2.34737679e-01 -3.01598813e-02
-2.15715408e-01 5.89557402e-02 -4.05750334e-01 7.25288928e-01
5.67292452e-01 -1.28563000e-02 -3.27177532e-02 -1.77829772e-01
-3.17733496e-01 4.79985118e-01 -1.15602994e+00 1.76042950e+00
-4.07683969e-01 4.44775730e-01 -1.43211871e-01 -9.89278257e-01
6.97413087e-01 3.91340971e-01 4.08421040e-01 -5.16964674e-01
3.84786218e-01 2.19315276e-01 4.21260819e-02 -2.12564677e-01
7.42736936e-01 -4.10772830e-01 -3.40736866e-01 5.82777560e-01
3.00154388e-01 -5.64415514e-01 4.95286465e-01 5.25364559e-03
7.43884265e-01 1.00296855e-01 4.89464670e-01 -7.23971069e-01
7.11969852e-01 2.27054000e-01 3.43198329e-01 3.57406646e-01
-3.72046798e-01 8.88750434e-01 7.63709486e-01 -4.05741125e-01
-9.69770968e-01 -8.55382740e-01 -4.53423977e-01 8.50743532e-01
-1.98113024e-01 -7.61098325e-01 -7.83600330e-01 -7.23043799e-01
-1.96667239e-01 9.69797850e-01 -1.03048360e+00 7.48587176e-02
-1.01606280e-01 -1.10210538e+00 4.40387547e-01 4.67592955e-01
2.33647197e-01 -1.05386686e+00 -5.99976897e-01 1.91645473e-01
8.39605648e-03 -1.18722117e+00 2.01237917e-01 4.96437699e-01
-9.24508095e-01 -1.16559172e+00 -4.35650706e-01 -1.18172958e-01
2.26655051e-01 -1.39816150e-01 1.61541212e+00 -4.32573915e-01
3.09386794e-02 2.84938723e-01 -4.66518551e-01 -9.47159529e-01
-1.84680969e-01 3.02157551e-01 -1.20526411e-01 -2.04510018e-01
8.49954903e-01 -3.62883419e-01 -4.65830714e-01 1.47459239e-01
-1.17297363e+00 -5.92533588e-01 4.31921899e-01 6.18878245e-01
4.72044587e-01 1.13892518e-01 4.74870652e-01 -1.16633856e+00
7.89362252e-01 -6.09579325e-01 -5.27778566e-01 1.12560913e-01
-1.25801730e+00 1.15008302e-01 4.80144709e-01 1.41864538e-01
-1.03863287e+00 -5.79361081e-01 -1.63284093e-01 1.61419123e-01
-2.64285982e-01 6.83866322e-01 1.35818049e-01 4.00219768e-01
6.77525103e-01 -3.33914667e-01 -8.13725442e-02 -3.70219409e-01
3.54679674e-01 6.24901950e-01 5.93747757e-02 -5.55477440e-01
3.25626403e-01 7.14311659e-01 -1.95381343e-01 -4.52693760e-01
-1.22606647e+00 -3.80052924e-01 -4.41601068e-01 1.60534377e-03
7.90188849e-01 -7.80427992e-01 -2.21175596e-01 3.90540212e-01
-8.81159544e-01 2.02225402e-01 -6.22172832e-01 4.91820812e-01
-1.63770646e-01 2.97105551e-01 -3.86682063e-01 -7.86860824e-01
-5.20732701e-01 -1.31147707e+00 1.05477405e+00 2.70190746e-01
-5.73440075e-01 -1.09231174e+00 2.68128306e-01 5.50171852e-01
6.95728898e-01 2.45339468e-01 8.56114447e-01 -9.34694469e-01
3.84022035e-02 -4.28598851e-01 -3.63246836e-02 5.15232146e-01
-5.97113967e-02 2.65601099e-01 -1.19100702e+00 1.05568040e-02
3.92695397e-01 -3.09109539e-01 9.08748567e-01 4.02778208e-01
7.73471475e-01 4.07151580e-01 -5.79973385e-02 -2.10614745e-02
1.43347502e+00 -1.48765102e-01 6.08691692e-01 8.31526697e-01
1.15817107e-01 9.80237305e-01 8.76257837e-01 4.10714507e-01
4.23633307e-01 6.18740141e-01 3.60418499e-01 3.72517258e-01
1.97564006e-01 1.64910734e-01 4.95997488e-01 8.39944780e-01
-1.82463527e-01 -1.40046135e-01 -8.73741984e-01 7.36591399e-01
-1.76335907e+00 -7.03314304e-01 5.41913584e-02 2.40117717e+00
6.27822816e-01 5.04505098e-01 1.25098884e-01 4.08983320e-01
4.16428626e-01 4.19033200e-01 -2.83180028e-01 -9.04713452e-01
-3.11632127e-01 3.14320803e-01 3.15476358e-01 3.63076657e-01
-9.74054933e-01 7.63772309e-01 6.88370800e+00 6.41695559e-01
-1.05632007e+00 -4.82658371e-02 8.35304737e-01 -9.44396481e-02
-5.23556471e-01 1.12650715e-01 -6.34057283e-01 3.25544119e-01
1.20215368e+00 -3.18446130e-01 1.54022455e-01 7.65570581e-01
-1.05061442e-01 -3.13235968e-01 -9.21175897e-01 6.09576821e-01
4.01311159e-01 -9.04359579e-01 -2.13010862e-01 -1.37832716e-01
8.03521931e-01 3.19271713e-01 1.42085269e-01 5.36435723e-01
1.65474772e-01 -1.09737718e+00 7.24807918e-01 3.32790643e-01
3.41762125e-01 -8.80607665e-01 1.24990058e+00 2.93349549e-02
-6.84821367e-01 8.09529424e-02 -3.23622614e-01 -1.10714398e-01
1.86787084e-01 1.26769543e+00 -2.52734751e-01 1.03104198e+00
9.56380248e-01 5.18747270e-01 -6.65144742e-01 6.86137736e-01
-2.31497198e-01 6.58773243e-01 -2.41264343e-01 5.37771452e-03
4.32177395e-01 -4.06605899e-01 5.70192158e-01 1.29909372e+00
2.55141377e-01 -3.97271872e-01 -3.62191558e-01 6.31595552e-01
-6.94956556e-02 2.64800489e-01 -5.69557488e-01 -1.25943393e-01
-9.76543874e-03 1.43531489e+00 -8.06920350e-01 -1.35530293e-01
-6.31637156e-01 4.41417843e-01 2.44049355e-01 4.09040898e-02
-7.45705247e-01 -4.30617034e-01 6.64177477e-01 -1.32687554e-01
2.40634561e-01 1.33833155e-01 -7.53169537e-01 -1.23103940e+00
2.88595617e-01 -9.46308434e-01 6.27339602e-01 -5.20015776e-01
-1.40364766e+00 9.40305114e-01 1.66462123e-01 -1.04067469e+00
-4.48499978e-01 -7.60124743e-01 -3.59916717e-01 5.81949711e-01
-1.55852771e+00 -5.40496826e-01 2.05518566e-02 1.13479063e-01
5.00981390e-01 -2.46991664e-01 1.01887882e+00 2.51164079e-01
-1.97515994e-01 3.48084748e-01 4.65560928e-02 -1.56426132e-01
9.78012919e-01 -1.38753021e+00 3.67457151e-01 7.45998442e-01
3.64804983e-01 7.23500013e-01 1.19411623e+00 -2.76552618e-01
-1.07002139e+00 -5.05307674e-01 1.24472487e+00 -8.26916695e-01
1.11316478e+00 -3.22151959e-01 -6.62204266e-01 5.29629290e-01
5.56337297e-01 -1.58159748e-01 8.29423428e-01 5.81427515e-01
-4.62860256e-01 -6.77629411e-02 -1.18568480e+00 3.94019514e-01
5.24407327e-01 -3.79222810e-01 -1.07523417e+00 2.15190127e-02
5.57303965e-01 -3.27188671e-01 -8.74361336e-01 7.11767137e-01
5.52254021e-01 -1.47110140e+00 5.10376513e-01 -6.42850578e-01
6.96421802e-01 -1.78166404e-01 -2.81786531e-01 -1.50764656e+00
5.81843294e-02 -1.68358058e-01 1.20034523e-01 1.59714103e+00
8.01627278e-01 -6.68443978e-01 7.76969612e-01 7.38869011e-01
-2.06654686e-02 -1.07753146e+00 -9.41995978e-01 -4.52538013e-01
4.28525209e-01 -9.72796619e-01 5.77894747e-01 7.46960402e-01
1.37060627e-01 4.87695903e-01 1.18809111e-01 -3.08607697e-01
3.98631364e-01 2.64411628e-01 4.09883946e-01 -1.29003716e+00
6.41574264e-02 -7.02206433e-01 -5.04180014e-01 -2.65637934e-01
5.37233800e-02 -6.09742403e-01 -3.38433743e-01 -1.38021111e+00
1.50027022e-01 -2.54638612e-01 -7.07631469e-01 1.04364149e-01
-1.96724594e-01 1.42225828e-02 6.51138723e-02 -4.37734164e-02
-5.62472582e-01 4.91814584e-01 7.43428051e-01 -1.95004791e-01
1.05075710e-01 1.89778894e-01 -1.17192936e+00 7.77254999e-01
7.39152253e-01 -4.73962367e-01 -3.80390793e-01 -3.24388295e-01
7.73444176e-01 -5.06672204e-01 -7.13149160e-02 -1.00782430e+00
-2.28939690e-02 3.48204046e-01 2.40066610e-02 -4.84608322e-01
1.03173062e-01 -9.24454689e-01 -1.16218030e-01 1.27252743e-01
-2.87205607e-01 3.01333696e-01 4.62200791e-01 5.18147945e-01
-6.73378289e-01 -4.18053865e-01 5.43943107e-01 -6.47058152e-03
-5.34647942e-01 -7.48417303e-02 -1.45574674e-01 3.96648824e-01
1.07244992e+00 6.18238896e-02 -2.83122838e-01 -1.44240618e-01
-5.34669280e-01 -1.47212762e-02 4.13812935e-01 7.44972646e-01
-7.94653893e-02 -8.32249582e-01 -7.71929920e-01 -1.88195214e-01
4.32191759e-01 -1.31231770e-01 2.76825130e-02 8.62846792e-01
-2.07727343e-01 8.44573677e-01 6.46028966e-02 -4.40774411e-01
-8.60871077e-01 4.90110010e-01 2.46063828e-01 -8.20620477e-01
-2.70837963e-01 6.95006311e-01 -3.42944205e-01 -5.52279651e-01
8.27921778e-02 -3.46726924e-01 -4.26347971e-01 6.65253639e-01
4.86703634e-01 3.76822412e-01 6.62595689e-01 -6.05957806e-01
-5.87844014e-01 5.68177819e-01 -2.15771675e-01 -3.18783075e-01
1.48240972e+00 -4.38206829e-02 -1.61716506e-01 9.72281516e-01
9.51363146e-01 1.58423305e-01 -7.10162640e-01 -5.84536716e-02
3.41214329e-01 -2.17558324e-01 8.87627006e-02 -1.09833491e+00
-1.08684373e+00 8.64020884e-01 3.40088606e-01 5.53407490e-01
9.86102164e-01 -8.59058872e-02 3.16042751e-01 2.95822844e-02
4.87646401e-01 -1.59331489e+00 -4.33462441e-01 6.17922306e-01
7.07185507e-01 -1.54585516e+00 4.13041055e-01 -3.00112575e-01
-9.39224124e-01 1.04116368e+00 1.38141736e-01 -8.78147557e-02
7.98233569e-01 3.49881470e-01 2.78474897e-01 -5.74271798e-01
-8.15183580e-01 -1.51378125e-01 3.03354919e-01 2.02351391e-01
9.85650420e-01 -1.00163296e-01 -8.58729839e-01 9.59375441e-01
-3.87656808e-01 -7.26914629e-02 4.36735332e-01 7.78964877e-01
-2.19526038e-01 -1.26228404e+00 -2.29295343e-01 4.12571013e-01
-1.07954431e+00 -1.43245026e-01 -4.86868769e-01 1.07977486e+00
-2.42022917e-01 1.35074711e+00 -1.80710480e-01 -1.81888193e-01
5.88946581e-01 3.76913846e-01 3.06786865e-01 -5.14808595e-01
-9.77291167e-01 -3.92309546e-01 2.20621452e-01 -7.08367527e-01
-6.97944522e-01 -8.83669913e-01 -1.03616726e+00 -1.90112561e-01
-3.06252807e-01 6.41747117e-01 1.16444790e+00 1.18919849e+00
3.87169868e-01 5.69568098e-01 3.38865727e-01 -6.10156000e-01
-7.21961856e-01 -1.04015052e+00 -6.05756462e-01 4.29878861e-01
-1.89068124e-01 -6.31297588e-01 -7.57807314e-01 -2.22547352e-01] | [11.225802421569824, 6.900378704071045] |
25023fbd-afd3-4a9c-94b8-123312e2f185 | r-pred-two-stage-motion-prediction-via-tube | 2211.08609 | null | https://arxiv.org/abs/2211.08609v5 | https://arxiv.org/pdf/2211.08609v5.pdf | Two-Stage Context-Aware model for Predicting Future Motion of Dynamic Agents | Predicting the future motion of dynamic agents is of paramount importance to ensuring safety and assessing risks in motion planning for autonomous robots. In this study, we propose a two-stage motion prediction method, called R-Pred, designed to effectively utilize both scene and interaction context using a cascade of the initial trajectory proposal and trajectory refinement networks. The initial trajectory proposal network produces M trajectory proposals corresponding to the M modes of the future trajectory distribution. The trajectory refinement network enhances each of the M proposals using 1) tube-query scene attention (TQSA) and 2) proposal-level interaction attention (PIA) mechanisms. TQSA uses tube-queries to aggregate local scene context features pooled from proximity around trajectory proposals of interest. PIA further enhances the trajectory proposals by modeling inter-agent interactions using a group of trajectory proposals selected by their distances from neighboring agents. Our experiments conducted on Argoverse and nuScenes datasets demonstrate that the proposed refinement network provides significant performance improvements compared to the single-stage baseline and that R-Pred achieves state-of-the-art performance in some categories of the benchmarks. | ['Jun Won Choi', 'Junyong Yun', 'Jungho Kim', 'Sehwan Choi'] | 2022-11-16 | null | null | null | null | ['motion-planning'] | ['robots'] | [-2.84758210e-01 1.40924782e-01 -3.63901138e-01 -2.22540408e-01
-6.84583843e-01 -2.38120615e-01 1.18134403e+00 1.19087376e-01
-7.70499229e-01 2.96915948e-01 6.62757695e-01 -3.15739572e-01
-8.33984688e-02 -8.41572881e-01 -6.56613469e-01 -7.01644301e-01
-5.82303286e-01 5.37707508e-01 8.53489101e-01 -3.84114385e-01
4.06448007e-01 7.31261492e-01 -1.26692581e+00 2.42488816e-01
7.34047472e-01 6.33425891e-01 5.46183407e-01 7.65328825e-01
1.81142181e-01 1.04668975e+00 -1.71898231e-01 -1.18920252e-01
3.98400962e-01 1.13577314e-01 -9.12349284e-01 -1.54796958e-01
8.23685899e-02 -7.58040607e-01 -8.87023866e-01 7.78400838e-01
7.34004378e-02 8.52893174e-01 6.75226450e-01 -1.54710352e+00
-4.24927860e-01 5.71312129e-01 -4.89031345e-01 4.55234647e-01
2.47042835e-01 8.24056625e-01 8.04695904e-01 -1.03641498e+00
8.53734493e-01 1.60913169e+00 5.49642205e-01 4.15192068e-01
-9.36743021e-01 -3.88538718e-01 6.84027612e-01 7.36291945e-01
-1.29414558e+00 -2.56808698e-01 6.28241241e-01 -4.66761380e-01
1.48374522e+00 -4.89074597e-03 4.07135814e-01 1.00458658e+00
4.99856919e-01 1.04358554e+00 5.66342752e-03 4.13757056e-01
2.67757714e-01 -2.15045899e-01 8.58911574e-02 7.17019558e-01
-2.06937253e-01 4.79925305e-01 -7.55389705e-02 -3.35387230e-01
7.58888483e-01 1.57067850e-01 -6.59409538e-02 -4.07339692e-01
-1.53863931e+00 9.49410617e-01 9.57464457e-01 -1.22733034e-01
-7.50180483e-01 4.94092107e-01 2.67229706e-01 -3.04724604e-01
1.46949202e-01 3.58296424e-01 -2.95168221e-01 -8.80896952e-03
-4.01062369e-01 7.80961990e-01 2.54434258e-01 1.34019363e+00
7.07514167e-01 -5.05632088e-02 -5.71629167e-01 3.49572718e-01
5.99141419e-01 3.35932940e-01 2.15498760e-01 -1.29536974e+00
7.20644414e-01 6.70353353e-01 5.85364580e-01 -1.14296877e+00
-8.15834582e-01 -1.21037655e-01 -4.84021962e-01 2.70692885e-01
2.31921688e-01 -1.63484871e-01 -9.93570805e-01 1.77578056e+00
6.84266925e-01 5.36938131e-01 2.34193429e-01 9.28099334e-01
7.46891975e-01 1.12735438e+00 6.42379940e-01 1.27290383e-01
1.04590309e+00 -1.71160650e+00 -4.14535016e-01 -1.80385932e-01
7.77445674e-01 -2.09599972e-01 6.91166282e-01 -2.32372612e-01
-1.19146466e+00 -8.12998772e-01 -6.48285866e-01 -1.30169719e-01
-2.98573017e-01 1.13508925e-01 4.44049656e-01 -6.33831173e-02
-1.10840786e+00 5.16296268e-01 -1.12191212e+00 -5.62736332e-01
5.34158707e-01 4.55990613e-01 -2.32225999e-01 -2.98097748e-02
-8.49978745e-01 9.56751287e-01 4.28137034e-01 2.08153483e-02
-1.43205738e+00 -6.75531089e-01 -1.12926793e+00 2.34730039e-02
3.92977268e-01 -9.49869156e-01 1.37347686e+00 -1.35576442e-01
-1.34457409e+00 -1.49190484e-03 -3.80664110e-01 -6.74505711e-01
5.28688192e-01 -4.17258203e-01 -8.91528875e-02 2.05551431e-01
3.08014572e-01 1.45664847e+00 4.88474786e-01 -1.25509274e+00
-1.26486695e+00 -1.23401828e-01 1.48363680e-01 5.95684052e-01
2.86236614e-01 -1.75984781e-02 -8.39547694e-01 -5.11455655e-01
-1.62723660e-01 -1.07685173e+00 -1.05618930e+00 -1.11341693e-01
-4.56948996e-01 -5.94822168e-01 1.12925577e+00 -4.14998561e-01
1.15340436e+00 -1.74513197e+00 2.02899888e-01 6.73853233e-02
9.18187797e-02 1.34598285e-01 -6.33022487e-01 6.04327142e-01
3.21082145e-01 -1.34500548e-01 -1.71794385e-01 -6.92789555e-01
2.23568641e-02 9.03094560e-02 -4.34159011e-01 4.43969369e-01
3.15925121e-01 1.16142213e+00 -1.23004830e+00 -3.01506847e-01
4.66913015e-01 3.50353807e-01 -7.52008498e-01 3.32699060e-01
-4.96770650e-01 6.89522028e-01 -9.15960133e-01 3.58439624e-01
4.21544909e-01 -1.74535483e-01 -1.77590698e-01 8.92059729e-02
-4.49308425e-01 3.12338442e-01 -6.35331213e-01 1.67758381e+00
9.78873298e-02 4.47787613e-01 -1.85521662e-01 -2.89231390e-01
5.85272551e-01 8.84426087e-02 6.05198979e-01 -4.46577430e-01
1.10722750e-01 -3.29777837e-01 -1.45221174e-01 -6.75085545e-01
1.17424726e+00 4.95657146e-01 -3.27595711e-01 2.64879048e-01
-3.78315240e-01 1.60243705e-01 -8.32680613e-02 4.30646658e-01
1.32464457e+00 3.15880477e-01 1.76788568e-01 -3.88860516e-02
5.57732761e-01 5.77314436e-01 7.01575160e-01 8.25426400e-01
-7.58901060e-01 2.65061319e-01 1.60562500e-01 -8.08350980e-01
-1.12414873e+00 -1.00018561e+00 5.36250114e-01 1.26105869e+00
7.93622851e-01 -2.80035704e-01 -5.60947239e-01 -9.57326353e-01
-1.05546355e-01 1.06831479e+00 -7.70033777e-01 -1.80954263e-01
-1.01914763e+00 -5.45515656e-01 2.80024737e-01 8.82020175e-01
5.17476678e-01 -1.37872136e+00 -1.01703000e+00 2.94826925e-01
-3.12163830e-01 -1.11894739e+00 -9.24422622e-01 -4.12005693e-01
-5.29765368e-01 -9.00977910e-01 -4.56335843e-01 -6.77378654e-01
7.42977798e-01 7.59631097e-01 6.60989046e-01 -2.70824172e-02
2.11090595e-01 4.07233894e-01 -4.65409160e-01 -2.00180233e-01
-2.45947585e-01 2.02067196e-02 1.26690388e-01 -9.62789059e-02
2.68116027e-01 -1.70972526e-01 -8.48161101e-01 4.85737532e-01
-4.01717126e-01 1.55113593e-01 6.15044773e-01 4.18247908e-01
4.15170878e-01 2.10185722e-03 5.30614138e-01 -4.11164850e-01
3.10702860e-01 -1.13672602e+00 -6.38147950e-01 -3.86859365e-02
-8.85987952e-02 -8.93589631e-02 3.63360852e-01 -4.56957847e-01
-1.18507326e+00 3.52560580e-01 -9.99561325e-02 -5.19016981e-01
-4.03735995e-01 3.95670116e-01 -1.48581669e-01 1.41863987e-01
4.40938205e-01 1.23970192e-02 -2.42730141e-01 -2.82988828e-02
6.65682852e-01 3.14018764e-02 8.14885080e-01 -1.87519953e-01
7.25825012e-01 7.47657597e-01 -1.30628735e-01 -6.64867818e-01
-1.08448952e-01 -7.34530628e-01 -6.35643482e-01 -4.61791933e-01
1.25295317e+00 -7.57199943e-01 -1.01881838e+00 4.31627750e-01
-1.50277007e+00 -6.85396492e-01 -1.55995637e-01 5.53439260e-01
-6.78424239e-01 4.35438275e-01 -5.28044820e-01 -9.87440228e-01
-8.19727033e-02 -1.64047861e+00 1.13550532e+00 2.17585906e-01
-1.60153806e-01 -8.50047052e-01 1.50825128e-01 5.34215085e-02
2.49591678e-01 1.50478348e-01 6.60212874e-01 -6.11806452e-01
-1.29116833e+00 -4.91151102e-02 -1.26397327e-01 -8.13151896e-01
-1.13940723e-01 -2.19250634e-01 -5.65591514e-01 -2.78629333e-01
-3.42669398e-01 8.52105767e-03 1.06739628e+00 7.34961987e-01
8.58924747e-01 -4.57909495e-01 -8.93852711e-01 4.42042679e-01
9.10875082e-01 5.93648791e-01 7.38820910e-01 4.73287821e-01
8.52162242e-01 8.44901502e-01 1.16979229e+00 5.25733411e-01
1.13188684e+00 9.60744083e-01 9.08981383e-01 1.45892560e-01
2.23567992e-01 -4.24268484e-01 6.64971113e-01 6.94895238e-02
-2.89918892e-02 -4.33845699e-01 -1.03316748e+00 9.31391358e-01
-2.46775365e+00 -1.35338831e+00 -3.10007721e-01 1.86534452e+00
-1.09749645e-01 -5.00592738e-02 5.64332366e-01 -6.37525320e-01
7.44931340e-01 3.20955753e-01 -8.90983224e-01 1.73985511e-02
3.02221268e-01 -8.91059160e-01 5.49269259e-01 8.00014436e-01
-1.51535189e+00 1.40071046e+00 5.84947014e+00 6.69681311e-01
-6.23530984e-01 6.65908456e-02 5.50643742e-01 2.64953729e-02
-1.71758160e-01 -8.54377747e-02 -9.71072793e-01 2.52347946e-01
7.30705559e-01 4.71437499e-02 1.71811894e-01 1.05613244e+00
6.53254986e-01 -1.78325847e-01 -1.06375718e+00 4.32869613e-01
-2.69063979e-01 -1.53197026e+00 2.25755855e-01 1.34502798e-01
7.67901957e-01 5.31217575e-01 2.84417868e-01 4.97536778e-01
7.98444390e-01 -9.62277114e-01 1.07750738e+00 5.12160301e-01
-1.07519478e-02 -8.96187127e-01 7.05401957e-01 6.78693235e-01
-1.79594517e+00 -6.31611705e-01 -3.59436303e-01 1.74238354e-01
6.78806841e-01 -4.43649858e-01 -1.23281980e+00 6.26194239e-01
7.24227965e-01 9.01364088e-01 -2.42767558e-01 1.23204362e+00
-5.28817065e-03 1.79674774e-01 -2.44799182e-01 -2.14074161e-02
9.73885238e-01 -1.41536504e-01 1.17102039e+00 1.09815419e+00
2.84049422e-01 4.37978178e-01 6.03707016e-01 9.40638363e-01
3.14459831e-01 -3.44349861e-01 -6.65864348e-01 4.01241034e-01
8.75878215e-01 1.08431458e+00 -6.26096427e-01 -4.04117078e-01
-3.55035186e-01 6.47363603e-01 4.80129778e-01 6.27546191e-01
-1.04609776e+00 3.21806148e-02 1.00587189e+00 -1.31923333e-01
5.56802869e-01 -5.12876928e-01 -3.84791978e-02 -4.57161427e-01
-4.44486529e-01 -2.24079326e-01 5.73402703e-01 -5.88115990e-01
-9.45587158e-01 8.20469379e-01 3.08072478e-01 -1.32457733e+00
-3.09632272e-01 -2.28773966e-01 -1.16587198e+00 6.43983424e-01
-1.47116709e+00 -1.51022279e+00 -2.85991967e-01 5.26756108e-01
1.09365201e+00 -6.87754601e-02 2.92538673e-01 -1.26894578e-01
-6.66718781e-01 2.80279040e-01 -3.55827242e-01 -8.82624462e-02
1.83755964e-01 -1.01026618e+00 1.16080904e+00 8.32896352e-01
-3.14646333e-01 4.23666000e-01 5.68494022e-01 -1.14247000e+00
-1.17131448e+00 -1.94175351e+00 7.19013274e-01 -7.63061404e-01
4.30664062e-01 1.91854686e-01 -8.89546335e-01 9.68104422e-01
1.46738961e-01 -1.58230305e-01 1.97503626e-01 -5.25354087e-01
1.23075902e-01 4.66205090e-01 -1.02773941e+00 1.05770588e+00
1.10694349e+00 -9.72287804e-02 -5.76328933e-01 1.08066469e-01
1.07993078e+00 -3.14921141e-01 -5.25167048e-01 5.15714049e-01
3.97047281e-01 -7.67499506e-01 1.20330620e+00 -6.76699102e-01
1.71232626e-01 -6.38184726e-01 -7.93361515e-02 -1.16034734e+00
-8.23277950e-01 -7.56432712e-01 -1.48046374e-01 6.21203542e-01
5.58930576e-01 -2.29270890e-01 9.94303107e-01 9.52459693e-01
-7.68195629e-01 -8.33500087e-01 -7.12545156e-01 -5.94645858e-01
-5.17403111e-02 -5.64718127e-01 8.54100764e-01 6.06423080e-01
2.06284851e-01 5.36199026e-02 -4.07757849e-01 7.73745179e-01
4.52932149e-01 -1.07725479e-01 1.06092680e+00 -6.02359176e-01
1.24901854e-01 -6.30665064e-01 -2.90297300e-01 -1.75764894e+00
2.75589615e-01 -7.13455141e-01 5.52026689e-01 -1.79365432e+00
1.97731033e-01 -4.78155792e-01 6.87424913e-02 2.05775052e-01
-4.47804838e-01 -2.76433349e-01 3.25989842e-01 4.57699895e-01
-1.04661179e+00 1.15436864e+00 1.26530492e+00 -2.17860326e-01
-6.94222271e-01 1.49345055e-01 -8.09458718e-02 9.68665063e-01
6.43927276e-01 -2.99803734e-01 -6.96656704e-01 -5.47049463e-01
-3.48123908e-01 3.18349630e-01 6.00320935e-01 -1.07201302e+00
8.21234643e-01 -7.16358900e-01 3.13483216e-02 -1.38365233e+00
7.82291174e-01 -7.24633455e-01 6.26291707e-02 6.47108197e-01
-5.27566910e-01 3.99312645e-01 2.13475212e-01 1.11712158e+00
2.48250797e-01 1.99632615e-01 3.80553752e-01 -1.14349328e-01
-1.22489130e+00 8.43634784e-01 -9.69850123e-01 -4.13423508e-01
1.41079354e+00 -2.83626080e-01 -4.55571234e-01 -5.28983235e-01
-5.96840978e-01 1.03396070e+00 4.04073030e-01 6.08550906e-01
1.05546069e+00 -1.45107150e+00 -5.95438600e-01 -2.23508745e-01
1.73563957e-01 3.49388391e-01 4.57158566e-01 8.36650848e-01
-3.65863413e-01 6.18764639e-01 -5.55317029e-02 -6.45394862e-01
-1.10730922e+00 8.21947396e-01 1.63274929e-01 -3.06975752e-01
-8.49670947e-01 1.04944456e+00 8.55562389e-01 -4.64602053e-01
3.95932585e-01 -6.65166736e-01 -7.89608181e-01 -5.95724106e-01
7.14627028e-01 7.74741054e-01 -7.91499078e-01 -1.24945474e+00
-3.05817962e-01 4.30133730e-01 -2.02477232e-01 -3.14520031e-01
1.18718922e+00 -4.73167688e-01 3.07571501e-01 -1.13674529e-01
8.53994906e-01 -4.68422323e-01 -2.00047898e+00 -6.10260591e-02
2.15579212e-01 -4.26286727e-01 -1.14710115e-01 -4.23507929e-01
-8.08151305e-01 5.58892190e-01 4.56922054e-02 -7.10095540e-02
7.37871587e-01 1.98297217e-01 1.09277070e+00 6.29092932e-01
7.35601962e-01 -8.05033982e-01 2.98197657e-01 8.28137517e-01
1.13258255e+00 -1.20422018e+00 -3.30935866e-01 -5.08600295e-01
-1.17216897e+00 6.10893846e-01 1.00985253e+00 -1.83055550e-01
4.70732480e-01 -1.82541519e-01 -2.21503273e-01 -2.52610087e-01
-1.05102336e+00 -3.81718904e-01 3.74950707e-01 7.16596246e-01
-4.39351141e-01 -9.94746536e-02 2.19226524e-01 5.51109612e-01
1.81525409e-01 -5.06143153e-01 2.93864936e-01 8.07577550e-01
-7.50863612e-01 -6.46913230e-01 -2.71444857e-01 1.39338046e-01
1.79897100e-01 2.83915520e-01 -2.46254414e-01 7.72891104e-01
9.83779803e-02 1.28602755e+00 3.18027228e-01 -7.18816996e-01
2.30182335e-01 -4.30376142e-01 -2.45276894e-02 -4.09655601e-01
-7.04142272e-01 2.43115332e-02 2.81534165e-01 -1.06555152e+00
-2.10413426e-01 -8.96557629e-01 -1.75743604e+00 -3.11610460e-01
-9.74546522e-02 -1.73141379e-02 2.12430418e-01 9.67951179e-01
6.90837622e-01 5.28582275e-01 5.41465461e-01 -1.65253353e+00
-2.28664964e-01 -9.18565035e-01 2.62037933e-01 3.03933889e-01
5.26479065e-01 -8.23191762e-01 2.85112292e-01 -1.05114328e-02] | [5.889460563659668, 0.7917446494102478] |
065d5f8f-73c0-4fa6-9d1d-61ef543f8cc4 | visual-forecasting-by-imitating-dynamics-in | 1708.05827 | null | http://arxiv.org/abs/1708.05827v1 | http://arxiv.org/pdf/1708.05827v1.pdf | Visual Forecasting by Imitating Dynamics in Natural Sequences | We introduce a general framework for visual forecasting, which directly
imitates visual sequences without additional supervision. As a result, our
model can be applied at several semantic levels and does not require any domain
knowledge or handcrafted features. We achieve this by formulating visual
forecasting as an inverse reinforcement learning (IRL) problem, and directly
imitate the dynamics in natural sequences from their raw pixel values. The key
challenge is the high-dimensional and continuous state-action space that
prohibits the application of previous IRL algorithms. We address this
computational bottleneck by extending recent progress in model-free imitation
with trainable deep feature representations, which (1) bypasses the exhaustive
state-action pair visits in dynamic programming by using a dual formulation and
(2) avoids explicit state sampling at gradient computation using a deep feature
reparametrization. This allows us to apply IRL at scale and directly imitate
the dynamics in high-dimensional continuous visual sequences from the raw pixel
values. We evaluate our approach at three different level-of-abstraction, from
low level pixels to higher level semantics: future frame generation, action
anticipation, visual story forecasting. At all levels, our approach outperforms
existing methods. | ['Juan Carlos Niebles', 'De-An Huang', 'Kuo-Hao Zeng', 'William B. Shen', 'Min Sun'] | 2017-08-19 | visual-forecasting-by-imitating-dynamics-in-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Zeng_Visual_Forecasting_by_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Zeng_Visual_Forecasting_by_ICCV_2017_paper.pdf | iccv-2017-10 | ['action-anticipation'] | ['computer-vision'] | [ 3.81062239e-01 2.72671372e-01 -2.57022828e-01 -6.36602566e-02
-5.99226713e-01 -6.35209322e-01 1.28269303e+00 -2.67350167e-01
-3.32021475e-01 8.46844554e-01 4.07902688e-01 -1.63407356e-01
2.46659800e-01 -7.27790594e-01 -1.04248464e+00 -6.80223167e-01
1.87880099e-01 3.05495322e-01 1.88558906e-01 -4.81941015e-01
4.32857454e-01 5.74686229e-01 -1.71417630e+00 3.29141140e-01
4.70380425e-01 9.09239471e-01 4.08752829e-01 1.17385125e+00
-7.09316805e-02 1.37070918e+00 -2.00639546e-01 1.43152792e-02
2.70944685e-01 -6.73786640e-01 -8.10161889e-01 4.96514916e-01
1.94260329e-01 -5.77391446e-01 -4.28507894e-01 6.94102645e-01
1.80767655e-01 1.79991975e-01 6.52332842e-01 -1.47986686e+00
-7.24883258e-01 9.46205407e-02 -3.64007145e-01 -1.28484666e-01
5.23966312e-01 6.29095435e-01 8.99088323e-01 -5.81235766e-01
1.06882036e+00 1.39540207e+00 6.33545041e-01 6.94307923e-01
-1.53207982e+00 5.39518297e-02 4.42669272e-01 3.91084731e-01
-8.39896500e-01 -1.85756668e-01 6.96131885e-01 -7.24177659e-01
1.26850927e+00 -3.96072716e-02 9.97927487e-01 1.33612275e+00
6.51534200e-02 1.00267303e+00 1.21788979e+00 -4.18420136e-01
3.12412977e-01 -2.31386870e-01 -3.61860722e-01 8.71884763e-01
-4.79447484e-01 3.70273083e-01 -4.43179905e-01 5.17710038e-02
1.24966347e+00 1.49511084e-01 -1.42764464e-01 -7.68466592e-01
-1.70284665e+00 8.65964651e-01 2.57459670e-01 -4.38285582e-02
-4.73174989e-01 8.68310452e-01 4.51035559e-01 4.33532834e-01
1.88963950e-01 3.16931427e-01 -4.42652434e-01 -4.16527182e-01
-7.63810813e-01 4.56665158e-01 7.21504688e-01 7.45360494e-01
8.19567561e-01 3.34989697e-01 -1.13983475e-01 2.97460765e-01
2.73247421e-01 4.68066156e-01 5.85228324e-01 -1.65267837e+00
1.59804642e-01 1.82509035e-01 4.90258902e-01 -7.36257911e-01
-1.38461426e-01 8.46837088e-02 -6.43869579e-01 7.39104629e-01
4.68066841e-01 -1.68181717e-01 -1.15730965e+00 1.92302334e+00
3.09131563e-01 5.72116137e-01 2.81819373e-01 8.66557539e-01
1.87201306e-01 1.09569228e+00 1.20824371e-02 -2.93299764e-01
1.07957232e+00 -1.13830543e+00 -4.51835841e-01 -1.70712814e-01
6.28012121e-01 -3.26640368e-01 1.17313731e+00 3.28130841e-01
-1.30544877e+00 -5.57378054e-01 -1.00817311e+00 -3.18397433e-01
-1.43702164e-01 -8.35022237e-03 4.21944648e-01 1.68842711e-02
-1.39430785e+00 9.87434208e-01 -1.12962711e+00 -2.05808818e-01
5.75318448e-02 2.33337700e-01 -3.53001922e-01 3.28953087e-01
-1.09521961e+00 1.12447643e+00 2.74442077e-01 3.10557745e-02
-1.12175190e+00 -6.31673634e-01 -1.03658569e+00 -1.53757241e-02
4.16995734e-01 -9.11652148e-01 1.50523520e+00 -1.49126601e+00
-2.15907454e+00 6.18740559e-01 -2.30843410e-01 -6.74672782e-01
9.25411224e-01 -1.62629053e-01 2.48031974e-01 1.22564733e-01
-1.10425532e-01 8.59751046e-01 1.12795329e+00 -1.20990419e+00
-4.10741180e-01 1.09659031e-03 2.59359926e-01 4.73143011e-01
1.55731931e-01 -3.60724032e-01 -1.47414580e-01 -6.62449896e-01
-1.39041856e-01 -1.13220787e+00 -5.06325960e-01 3.94359410e-01
2.86511742e-02 -6.54825568e-02 6.54660583e-01 -5.29074669e-01
6.41746879e-01 -1.80584431e+00 7.62563169e-01 -1.47838756e-01
6.78318962e-02 1.49654090e-01 -2.23280281e-01 6.22248590e-01
3.76058109e-02 -1.50033131e-01 -1.73341051e-01 -4.47304219e-01
1.64890110e-01 5.21638870e-01 -6.72264040e-01 3.04163903e-01
3.34884822e-01 1.28665841e+00 -1.21766520e+00 -4.56014454e-01
3.84807765e-01 5.21556139e-01 -6.75767660e-01 3.03142786e-01
-7.28118241e-01 6.43221140e-01 -4.35842127e-01 8.45303237e-02
1.31195709e-01 -3.71075481e-01 2.70133168e-01 4.57583219e-02
-3.32286805e-01 1.96886316e-01 -1.11878049e+00 2.01269221e+00
-6.82258785e-01 6.63122177e-01 -1.69317141e-01 -1.26074350e+00
8.16683590e-01 2.84917921e-01 3.96469146e-01 -5.19418836e-01
-3.84016484e-01 -3.03449072e-02 -4.01454151e-01 -5.41097701e-01
4.14798439e-01 -2.17861161e-01 -1.12826966e-01 5.65053642e-01
-1.64240804e-02 -5.17087758e-01 -2.22518612e-02 1.52481079e-01
8.04181993e-01 1.08381712e+00 2.78205246e-01 3.18998024e-02
4.35245395e-01 1.84596911e-01 6.33092880e-01 6.99822307e-01
-1.11652818e-02 6.20773554e-01 9.17776823e-01 -7.30823100e-01
-1.49030304e+00 -1.18817449e+00 5.30903816e-01 1.05154264e+00
-8.69657993e-02 -2.26643324e-01 -6.82584107e-01 -5.16073704e-01
-1.20978300e-02 6.38119459e-01 -9.13284242e-01 4.03772220e-02
-8.51731062e-01 -1.33860141e-01 1.41670257e-01 6.70496225e-01
2.32122540e-01 -1.36584127e+00 -1.08503687e+00 3.04054618e-01
-2.11241953e-02 -1.07551432e+00 -4.50340569e-01 -4.90336232e-02
-8.93718660e-01 -5.38560390e-01 -8.99953544e-01 -5.86955488e-01
4.24248248e-01 -6.98988661e-02 1.15098894e+00 -1.50307983e-01
-2.35794246e-01 6.36809230e-01 -4.12915982e-02 -9.54073817e-02
-6.83622181e-01 -2.81200260e-01 -1.48076937e-01 3.87969315e-02
-3.02189201e-01 -7.77720928e-01 -6.89552367e-01 -7.69923106e-02
-7.61419177e-01 7.20009148e-01 3.40628952e-01 1.11029565e+00
6.25033975e-01 -6.86885178e-01 4.77948248e-01 -6.20635629e-01
3.56860429e-01 -2.69753009e-01 -7.73673654e-01 2.63955593e-01
-2.80626118e-01 6.77989185e-01 7.15763211e-01 -6.40498698e-01
-1.09530628e+00 5.65413713e-01 3.96098942e-02 -7.66218543e-01
8.47075880e-03 2.06703439e-01 2.38811746e-01 3.00737888e-01
4.94197637e-01 6.40878856e-01 5.14431298e-01 -9.92380157e-02
8.20188284e-01 2.12183341e-01 6.60054505e-01 -5.62040210e-01
5.65280914e-01 6.73887253e-01 2.48493195e-01 -5.88277638e-01
-6.87371671e-01 1.01849020e-01 -7.91009724e-01 -2.53292829e-01
8.79095376e-01 -8.88962448e-01 -9.79964197e-01 4.67156768e-01
-1.25422060e+00 -8.12374890e-01 -6.75475299e-01 3.01987320e-01
-1.51185000e+00 5.37014425e-01 -7.25306094e-01 -7.54118979e-01
-1.04581518e-02 -1.24715686e+00 1.31721175e+00 -9.94383991e-02
-3.29204798e-01 -9.82810318e-01 2.66592860e-01 -7.69849261e-03
2.56912649e-01 7.47748315e-01 7.79438674e-01 3.17263156e-02
-8.04588437e-01 1.52129248e-01 1.34936310e-02 2.07761407e-01
-2.15604514e-01 1.28818095e-01 -7.06174850e-01 -1.52006522e-01
-7.82107860e-02 -6.23272777e-01 8.23324919e-01 4.53049004e-01
9.03201938e-01 -3.78076822e-01 -8.95394608e-02 4.48924005e-01
1.33798695e+00 -2.03731693e-02 6.20414197e-01 4.31355059e-01
6.87861919e-01 6.30872428e-01 5.07933617e-01 6.68148279e-01
5.61928988e-01 7.91478634e-01 5.81218004e-01 1.15779871e-02
-8.81189927e-02 -6.53033853e-01 6.99162841e-01 4.89576608e-01
-2.39650205e-01 8.87913704e-02 -8.41368020e-01 6.08404696e-01
-2.51275015e+00 -1.15352619e+00 2.14043751e-01 1.93550622e+00
7.24233687e-01 -1.78287551e-01 3.64436537e-01 -6.64927363e-02
4.69530940e-01 3.00804734e-01 -9.95090246e-01 -6.58254206e-01
1.84811413e-01 -2.20983997e-02 2.98918903e-01 7.24852979e-01
-8.45336854e-01 1.12262690e+00 6.37103605e+00 4.80181694e-01
-1.10665929e+00 2.05548704e-02 5.91241300e-01 -1.25893548e-01
-3.29817742e-01 1.86729163e-01 -4.34268147e-01 2.68660754e-01
9.47145998e-01 -1.24321252e-01 7.39403605e-01 8.77909660e-01
3.38517755e-01 3.53747755e-02 -1.31329215e+00 9.78774726e-01
-2.17628345e-01 -1.81147206e+00 7.15379640e-02 -5.46444766e-02
7.50448406e-01 -7.33204409e-02 5.32247648e-02 3.63517851e-01
6.25201166e-01 -9.75915074e-01 1.01543283e+00 7.85383940e-01
7.63512015e-01 -5.88561475e-01 -3.76042873e-02 4.98507053e-01
-1.02657187e+00 -2.16541216e-01 -1.12964928e-01 -4.13738161e-01
5.12708724e-01 -1.60797685e-02 -5.86432159e-01 1.01731479e-01
2.35722959e-01 1.06184006e+00 2.59317327e-02 2.33505517e-01
-2.44453907e-01 1.46601975e-01 -2.75821745e-01 -1.12405114e-01
5.04165411e-01 -2.42216751e-01 3.50203276e-01 1.08432388e+00
3.43294740e-01 -5.78300096e-03 3.02614003e-01 7.86731541e-01
2.91486979e-01 -3.48981231e-01 -6.83910310e-01 -3.32120573e-03
-1.64634287e-02 9.15055335e-01 -4.96574968e-01 -6.37305140e-01
-3.85552257e-01 1.29078400e+00 5.04165292e-01 4.18527186e-01
-9.16812778e-01 4.22836617e-02 8.42924476e-01 -1.27626618e-03
5.24900913e-01 -6.13696754e-01 4.91320975e-02 -1.49935615e+00
-7.50112161e-02 -8.09516251e-01 -6.22450821e-02 -9.40383852e-01
-7.82663643e-01 3.34382892e-01 -7.25530880e-03 -1.18158901e+00
-1.21729076e+00 -4.72132504e-01 -4.51408684e-01 7.04574406e-01
-1.43040359e+00 -1.19538677e+00 -4.56842333e-02 7.66778588e-01
7.98795462e-01 1.35905191e-01 6.97324991e-01 -4.27025199e-01
-1.00792602e-01 8.27049911e-02 1.61837280e-01 -2.93381453e-01
1.47600234e-01 -1.44464338e+00 8.17963362e-01 6.60301089e-01
7.42490888e-02 -1.31002679e-01 8.71387899e-01 -3.42869878e-01
-1.66210067e+00 -8.60310555e-01 6.33838356e-01 -5.04385233e-01
1.02956927e+00 -4.63487059e-01 -6.38855994e-01 8.89219522e-01
9.04416665e-02 2.21029088e-01 -1.95906818e-01 -4.42991108e-01
-2.15307176e-01 3.04559618e-01 -8.35895717e-01 8.48605752e-01
1.16009355e+00 -6.78615808e-01 -4.38845187e-01 4.06434655e-01
7.58882284e-01 -3.82766515e-01 -7.61075318e-01 3.65579762e-02
6.35783076e-01 -8.25028181e-01 1.21582770e+00 -8.64752889e-01
7.62006164e-01 -3.11360449e-01 -1.15538962e-01 -1.19187200e+00
-2.20248207e-01 -1.00064075e+00 -5.01970887e-01 7.19015181e-01
2.03691497e-01 -4.21208799e-01 8.31701159e-01 7.72934139e-01
1.10383920e-01 -8.13059151e-01 -7.52749026e-01 -7.57761240e-01
1.35583475e-01 -3.71683836e-01 3.50758761e-01 5.53036153e-01
1.03598852e-02 9.57320631e-02 -8.34014893e-01 -1.15597241e-01
4.17022854e-01 4.45581496e-01 8.36569607e-01 -7.93812096e-01
-7.70769119e-01 -3.11905235e-01 -4.92634952e-01 -1.42656016e+00
5.27457356e-01 -7.06431866e-01 2.25513607e-01 -1.52405834e+00
-9.41137318e-03 -9.18456391e-02 3.83180417e-02 4.88434255e-01
1.84658349e-01 -9.92196379e-04 7.10193813e-01 2.89823920e-01
-5.45101285e-01 8.70263577e-01 1.45750666e+00 1.18750446e-01
-1.33141428e-01 -2.40153223e-01 -1.75642222e-01 8.67808282e-01
6.16945863e-01 -2.36249283e-01 -5.96720934e-01 -4.60396618e-01
2.44799942e-01 7.24469066e-01 6.57305360e-01 -6.49916470e-01
-2.95424648e-02 -6.25790656e-01 3.15861911e-01 -1.44730791e-01
7.00018466e-01 -4.65495825e-01 1.92510769e-01 7.78049588e-01
-5.92787623e-01 3.85262609e-01 -7.02262595e-02 7.45178044e-01
1.69871710e-02 -2.01496884e-01 6.81017578e-01 -4.45957601e-01
-9.57031131e-01 9.31623280e-02 -6.45864189e-01 -1.39540374e-01
1.15640318e+00 -3.18761557e-01 9.63225029e-03 -7.77617097e-01
-1.06299853e+00 7.62447789e-02 7.78433323e-01 4.08935040e-01
5.49783766e-01 -1.31430185e+00 -6.24830961e-01 9.67083946e-02
-3.47056687e-01 -2.03030661e-01 1.65659755e-01 5.99595129e-01
-5.40202022e-01 1.09015107e-01 -4.55874264e-01 -7.56458163e-01
-9.64226127e-01 7.74414241e-01 3.01343918e-01 -5.16565442e-01
-8.82712901e-01 3.96041363e-01 3.20690215e-01 -2.54122972e-01
-1.06818965e-02 -2.81945974e-01 5.60157374e-03 -8.79338607e-02
4.20624495e-01 3.22145790e-01 -5.23385227e-01 -6.15814865e-01
7.54214749e-02 7.21358716e-01 7.28501007e-02 -5.75044632e-01
1.55398023e+00 -2.12200150e-01 1.85439751e-01 7.61750579e-01
1.26162672e+00 -7.10605025e-01 -2.18863034e+00 -5.52622676e-02
-2.21974794e-02 -2.25728020e-01 -4.03895974e-02 -3.64569783e-01
-7.93785751e-01 1.12446547e+00 3.77613246e-01 7.06015974e-02
9.10577476e-01 -2.24498823e-01 6.96505964e-01 5.65645039e-01
2.18860254e-01 -1.13208926e+00 3.46789747e-01 6.61266208e-01
1.02896285e+00 -1.13104951e+00 -1.86515316e-01 1.96443036e-01
-1.08317304e+00 1.27553415e+00 3.42538446e-01 -6.56123698e-01
3.59307259e-01 3.09140921e-01 -8.85500908e-02 1.69285193e-01
-1.16865849e+00 -2.17001736e-01 -7.02539459e-02 5.33505142e-01
2.45042101e-01 -1.91658884e-01 -4.37208936e-02 -1.73316568e-01
1.02266781e-01 3.56637239e-01 7.52685428e-01 1.13554573e+00
-3.52435023e-01 -1.22703266e+00 -3.88171785e-02 1.16930589e-01
-1.06443278e-01 7.80368149e-02 -3.19866911e-02 8.03003550e-01
-4.02725041e-01 3.70600373e-01 1.43925563e-01 8.82186275e-03
3.19903716e-02 6.55858293e-02 7.06859589e-01 -3.22585881e-01
-2.43515193e-01 -1.09115645e-01 -1.43660501e-01 -8.94381106e-01
-4.38371658e-01 -9.51517522e-01 -1.30281138e+00 -1.82762504e-01
2.23457038e-01 -2.19649494e-01 7.13664234e-01 9.07858729e-01
6.69245481e-01 2.97235012e-01 5.46643019e-01 -1.63480771e+00
-7.50279784e-01 -4.52784836e-01 -1.04814157e-01 6.06047630e-01
7.53434300e-01 -6.45999610e-01 -3.89429666e-02 4.30985928e-01] | [8.316673278808594, 0.1445244699716568] |
a5eff86b-3348-4f08-837e-8db056e138db | revisiting-mid-level-patterns-for-distant | 2008.03128 | null | https://arxiv.org/abs/2008.03128v4 | https://arxiv.org/pdf/2008.03128v4.pdf | Revisiting Mid-Level Patterns for Cross-Domain Few-Shot Recognition | Existing few-shot learning (FSL) methods usually assume base classes and novel classes are from the same domain (in-domain setting). However, in practice, it may be infeasible to collect sufficient training samples for some special domains to construct base classes. To solve this problem, cross-domain FSL (CDFSL) is proposed very recently to transfer knowledge from general-domain base classes to special-domain novel classes. Existing CDFSL works mostly focus on transferring between near domains, while rarely consider transferring between distant domains, which is in practical need as any novel classes could appear in real-world applications, and is even more challenging. In this paper, we study a challenging subset of CDFSL where the novel classes are in distant domains from base classes, by revisiting the mid-level features, which are more transferable yet under-explored in main stream FSL work. To boost the discriminability of mid-level features, we propose a residual-prediction task to encourage mid-level features to learn discriminative information of each sample. Notably, such mechanism also benefits the in-domain FSL and CDFSL in near domains. Therefore, we provide two types of features for both cross- and in-domain FSL respectively, under the same training framework. Experiments under both settings on six public datasets, including two challenging medical datasets, validate the our rationale and demonstrate state-of-the-art performance. Code will be released. | ['José M. F. Moura', 'Shanghang Zhang', 'Yixiong Zou', 'Yonghong Tian', 'JianPeng Yu'] | 2020-08-07 | null | null | null | null | ['cross-domain-few-shot', 'cross-domain-few-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 3.73001873e-01 -3.49722169e-02 -5.28676867e-01 -5.10046005e-01
-8.31874728e-01 -5.04784346e-01 5.05815089e-01 1.65488273e-01
-1.47257254e-01 1.04012656e+00 1.71522200e-01 5.93014807e-02
-4.23345238e-01 -9.29565430e-01 -6.61849022e-01 -7.56900728e-01
-4.69163693e-02 3.01312387e-01 5.22514582e-01 -2.00358465e-01
-1.61085859e-01 1.54708281e-01 -1.41853380e+00 6.04082167e-01
9.92391825e-01 9.28305745e-01 3.08715552e-01 -2.98675615e-02
-1.72768965e-01 4.23051745e-01 -4.68534172e-01 -1.89164475e-01
4.39918101e-01 -4.74816769e-01 -6.66482925e-01 2.15266168e-01
1.91637874e-01 -3.03620428e-01 -3.36878151e-01 9.64015782e-01
5.82084954e-01 2.73267359e-01 7.24808097e-01 -1.25405419e+00
-8.42797458e-01 3.29754859e-01 -4.81126040e-01 2.77945727e-01
3.24823052e-01 1.41452715e-01 8.73376966e-01 -8.45267415e-01
6.58171654e-01 1.15219653e+00 6.77366555e-01 7.76762366e-01
-9.89167213e-01 -8.11723053e-01 3.31444114e-01 2.36562029e-01
-1.22808731e+00 -4.00216669e-01 8.69315684e-01 -4.06536996e-01
4.48287427e-01 3.80440429e-02 2.20055327e-01 1.38038623e+00
3.24532186e-04 1.10837126e+00 1.19865894e+00 -2.03546152e-01
5.37780285e-01 4.51818228e-01 2.24943265e-01 1.35000065e-01
2.28805915e-01 8.46149623e-02 -4.19338554e-01 -3.30313265e-01
5.82507849e-01 5.13740361e-01 -4.93193060e-01 -5.66710293e-01
-1.17161834e+00 9.48095679e-01 3.52756083e-01 4.24005687e-01
-1.79339141e-01 -6.66655004e-01 5.62672675e-01 4.09613311e-01
6.28492177e-01 2.01087683e-01 -7.58917689e-01 -3.44526842e-02
-7.11159289e-01 2.77040273e-01 5.65261245e-01 1.30610132e+00
8.55066419e-01 -4.32273090e-01 -3.57084364e-01 1.11104822e+00
-2.04243138e-01 1.23613119e-01 8.44474554e-01 -4.15580124e-01
5.66539168e-01 6.86473072e-01 -3.42519358e-02 -7.88336158e-01
-2.35747874e-01 -6.24688685e-01 -9.37945843e-01 -3.09513420e-01
2.62774587e-01 -1.97106734e-01 -7.98428178e-01 1.70787036e+00
5.87478220e-01 5.59205353e-01 2.79204190e-01 7.60029614e-01
9.27099168e-01 6.13570273e-01 -1.87914059e-01 -3.62113535e-01
1.27623093e+00 -9.16561365e-01 -5.13133824e-01 -2.76457787e-01
8.00171614e-01 -4.12486762e-01 1.23221433e+00 3.48330200e-01
-5.20040810e-01 -7.14180768e-01 -9.78559911e-01 1.98472589e-01
-4.70644385e-01 -4.55620438e-02 5.40087998e-01 4.46004897e-01
-3.69390935e-01 6.54623926e-01 -4.74980026e-01 -4.77210879e-01
8.40556741e-01 -3.08279600e-02 -4.31933910e-01 -6.03969812e-01
-1.62657952e+00 4.43536222e-01 6.37048006e-01 -4.15308803e-01
-8.21644008e-01 -1.04005706e+00 -8.68552923e-01 -2.33035441e-02
4.91938561e-01 -3.89284223e-01 1.02573943e+00 -1.03148782e+00
-1.08430398e+00 8.43173385e-01 7.36272335e-02 -2.55291373e-01
6.21003687e-01 -1.19027898e-01 -7.58660197e-01 -1.59679856e-02
3.66421998e-01 3.74100447e-01 6.97494626e-01 -1.10410357e+00
-9.32238519e-01 -3.08139324e-01 2.69417137e-01 3.45840901e-02
-7.21565008e-01 -5.10775030e-01 -3.29206169e-01 -7.85769045e-01
-2.42327213e-01 -6.94436073e-01 -1.17251046e-01 3.04026932e-01
-2.44882643e-01 -5.54766119e-01 9.06323671e-01 -3.92305046e-01
1.38477790e+00 -2.49994278e+00 -3.64184082e-01 -1.41591415e-01
2.67322958e-02 6.11314476e-01 -1.83791697e-01 3.24201614e-01
-1.69548243e-01 -4.30689394e-01 -4.97487694e-01 8.44687074e-02
-8.39800611e-02 5.17617762e-01 -4.54318166e-01 4.14976478e-01
5.16490459e-01 5.63639283e-01 -1.26677823e+00 -4.80900913e-01
5.32601885e-02 1.67943269e-01 -6.14866853e-01 1.81758538e-01
-3.24454755e-02 4.94013280e-01 -7.39641666e-01 7.66487896e-01
1.03627217e+00 -2.68612653e-01 1.03152141e-01 -9.85402390e-02
3.24974895e-01 9.42307860e-02 -1.11493826e+00 1.93979442e+00
-4.20882761e-01 1.70061469e-01 -3.98411632e-01 -1.61742675e+00
1.04978812e+00 2.04808310e-01 4.64446396e-01 -7.00826049e-01
-2.37205684e-01 2.14153469e-01 2.86753736e-02 -5.07367909e-01
-7.12583885e-02 -5.68547785e-01 -2.33480945e-01 1.00535743e-01
1.77912936e-01 2.89426416e-01 1.06341019e-01 -5.14242612e-02
1.09626782e+00 1.67797953e-01 7.43889093e-01 -3.77191722e-01
4.95911390e-01 -1.33627132e-01 1.12950015e+00 6.69549108e-01
-6.36064112e-01 8.14214706e-01 5.01976371e-01 -3.52439761e-01
-6.95407987e-01 -1.18911886e+00 -5.37794054e-01 1.07449472e+00
3.68520558e-01 -1.37522042e-01 -3.92737448e-01 -1.31463075e+00
2.43081421e-01 6.65089071e-01 -7.26939917e-01 -5.60666084e-01
-3.35180104e-01 -7.17697501e-01 3.17233294e-01 4.96126771e-01
4.02636319e-01 -9.18437183e-01 -4.75747734e-01 3.38889360e-01
6.76073432e-02 -1.09732556e+00 -3.85012627e-01 2.49729678e-01
-7.91785955e-01 -1.21620488e+00 -1.13651180e+00 -9.32758510e-01
5.82682848e-01 3.38493466e-01 1.06207526e+00 -3.65405262e-01
-3.48885477e-01 -4.06494476e-02 -7.36005723e-01 -3.51258785e-01
-7.37411007e-02 4.16510291e-02 1.13211080e-01 2.03228503e-01
7.97182143e-01 -6.76780581e-01 -6.09376252e-01 4.59560752e-01
-1.03889334e+00 -1.67511776e-01 6.39075696e-01 1.12758422e+00
5.21355093e-01 2.23673671e-01 1.13675570e+00 -1.29252732e+00
3.07767838e-01 -1.13256288e+00 -8.81612301e-02 2.52455145e-01
-2.82775342e-01 -1.59438327e-01 1.12078810e+00 -7.28388309e-01
-9.71782327e-01 -1.77770287e-01 -3.37103877e-04 -4.42620784e-01
-4.68992561e-01 4.41906929e-01 -5.61723292e-01 4.36087668e-01
8.24700713e-01 4.73463178e-01 -1.77766219e-01 -6.30716503e-01
3.40914279e-02 1.04061425e+00 1.64793417e-01 -6.63956940e-01
7.09456742e-01 5.33195078e-01 -3.69923532e-01 -6.43432856e-01
-1.21160042e+00 -7.22391307e-01 -5.15288234e-01 2.86963999e-01
3.88126433e-01 -1.06951475e+00 1.29467919e-01 2.78202951e-01
-7.14586735e-01 -1.85236171e-01 -6.22564018e-01 4.77511048e-01
-4.81176049e-01 4.44205493e-01 -1.53882220e-01 -3.62103671e-01
-1.34552315e-01 -1.04390848e+00 1.21746194e+00 3.29527378e-01
-7.73894861e-02 -1.04068542e+00 1.42428398e-01 1.11518838e-02
1.80784285e-01 3.87043148e-01 9.81052339e-01 -1.11263001e+00
3.11938841e-02 -1.36109576e-01 -3.43435585e-01 4.90010083e-01
7.11713672e-01 -6.40214086e-01 -1.00820780e+00 -5.21218002e-01
-3.80360931e-02 -6.33964837e-01 9.03725624e-01 1.15073286e-01
1.33025026e+00 -4.30390052e-02 -6.69823408e-01 5.15303075e-01
1.29339874e+00 3.04581642e-01 4.24587369e-01 1.54143840e-01
2.88115352e-01 6.55270696e-01 1.13294339e+00 8.01923752e-01
1.76153466e-01 7.36335218e-01 -1.13258831e-01 -1.17898313e-02
-1.67289212e-01 -1.57187268e-01 3.23143810e-01 6.60300851e-01
3.88962388e-01 -2.21477542e-02 -8.65679920e-01 8.82775724e-01
-1.90286672e+00 -7.56938934e-01 3.81669641e-01 2.11787057e+00
1.27934909e+00 2.09808797e-01 1.71520993e-01 -1.11743277e-02
9.33795273e-01 6.01842487e-03 -9.36120749e-01 5.20483106e-02
-1.60814933e-02 2.08630919e-01 8.82430896e-02 -1.01934902e-01
-1.38221085e+00 5.89955688e-01 4.76457262e+00 1.40999866e+00
-1.00953484e+00 3.17614436e-01 6.29838705e-01 -9.52767730e-02
-2.28653222e-01 -6.89369291e-02 -9.50049639e-01 8.05406868e-01
6.13968432e-01 -3.65825087e-01 -8.65267515e-02 1.13192797e+00
-7.47200921e-02 2.33470812e-01 -1.33663297e+00 1.06850171e+00
-3.07506006e-02 -1.02234614e+00 8.34364146e-02 -1.30762652e-01
7.85981297e-01 -1.75922334e-01 -8.64566583e-03 9.56681132e-01
1.00387543e-01 -6.24799430e-01 1.50630876e-01 3.06607366e-01
1.07733655e+00 -8.27630162e-01 6.66995347e-01 6.36440754e-01
-1.15841818e+00 -2.22843379e-01 -8.75902653e-01 4.47934344e-02
-9.64995623e-02 1.07365704e+00 -6.53940916e-01 9.00281489e-01
6.95168138e-01 1.29814529e+00 -4.36143190e-01 1.08846939e+00
2.21237347e-01 5.15951574e-01 -7.52398893e-02 2.45877072e-01
2.93243200e-01 1.28294662e-01 3.01539272e-01 1.17460263e+00
4.58318949e-01 1.11370854e-01 4.97603536e-01 6.53437138e-01
-1.07040890e-02 -1.76560450e-02 -7.48170018e-01 6.69816062e-02
4.87683296e-01 9.94568527e-01 -3.85212272e-01 -5.35180688e-01
-7.31318951e-01 9.24971282e-01 2.80888915e-01 2.68058240e-01
-9.26919937e-01 -7.31438875e-01 1.05707920e+00 1.90530196e-01
3.13338250e-01 3.59473526e-01 1.32830456e-01 -1.53088129e+00
2.56069601e-01 -9.39309955e-01 6.75720334e-01 -9.40650031e-02
-2.11223745e+00 4.31777358e-01 9.17275622e-02 -1.92821968e+00
3.21342684e-02 -6.03183031e-01 -5.26008606e-01 6.08386099e-01
-1.90512300e+00 -1.08459473e+00 -3.04647982e-01 9.24089849e-01
9.21398222e-01 -3.02520901e-01 8.30530703e-01 6.14330530e-01
-3.38921160e-01 8.94632816e-01 6.96245372e-01 1.60480663e-01
1.14775229e+00 -8.93024266e-01 1.62552670e-01 4.20509964e-01
-2.54880607e-01 6.13270819e-01 3.56691360e-01 -6.16939485e-01
-1.11496913e+00 -1.52775896e+00 5.50604463e-01 -1.48710668e-01
5.58419943e-01 -5.24680436e-01 -1.26755989e+00 4.85278189e-01
-2.42686391e-01 5.61382771e-01 9.85630393e-01 -4.78161871e-03
-3.35030377e-01 -2.67516553e-01 -1.37395394e+00 2.33634531e-01
1.30745935e+00 -4.04668093e-01 -8.52211535e-01 5.94926059e-01
7.12256014e-01 -5.10280803e-02 -8.06172967e-01 6.30014837e-01
3.27231824e-01 -8.70088160e-01 9.67033088e-01 -8.25089872e-01
5.62926769e-01 -2.57047147e-01 -3.16006303e-01 -1.43969917e+00
-2.86992997e-01 -2.02934682e-01 -1.87259495e-01 1.39361703e+00
8.83966312e-02 -7.51404226e-01 7.68967211e-01 1.72448784e-01
-2.09984645e-01 -8.85271132e-01 -1.03559482e+00 -1.28504503e+00
3.67194712e-01 -2.09564909e-01 6.38678670e-01 1.40179086e+00
1.00323975e-01 8.85318145e-02 -3.91125053e-01 1.42532196e-02
4.45573360e-01 4.47640300e-01 6.64055943e-01 -1.35647631e+00
-5.31379342e-01 -1.20477661e-01 -5.16489863e-01 -1.10936356e+00
2.48599797e-01 -1.02224779e+00 4.04323488e-02 -1.29606700e+00
4.56547529e-01 -7.96517968e-01 -6.07334435e-01 5.73603809e-01
-5.06013513e-01 -6.20684307e-03 1.23350359e-02 1.22074701e-01
-7.17716038e-01 8.27081442e-01 1.34441996e+00 -2.03938991e-01
-4.17552441e-02 2.17274755e-01 -8.84786308e-01 5.79119384e-01
5.56102574e-01 -4.46310580e-01 -7.94007361e-01 -3.46714854e-02
-4.19175625e-01 -1.67653978e-01 6.02718107e-02 -1.14436376e+00
-9.85980853e-02 -4.41141635e-01 4.33228850e-01 -3.48032624e-01
1.33687645e-01 -8.33846509e-01 -2.06733525e-01 3.96633923e-01
-2.78169960e-01 -6.90349340e-01 2.13762224e-02 8.39907587e-01
-4.50167090e-01 -1.72967151e-01 1.00041115e+00 -9.92497057e-02
-1.10381126e+00 5.75874627e-01 1.91851869e-01 4.65852290e-01
1.45781004e+00 -3.79646242e-01 -1.90949038e-01 -8.38019252e-02
-8.20338547e-01 2.16022789e-01 3.28004330e-01 6.21031642e-01
7.16448903e-01 -1.57159388e+00 -7.07331479e-01 4.14911360e-01
7.08611131e-01 2.71140844e-01 5.12800097e-01 7.06235826e-01
-1.51832020e-02 1.73959658e-01 -2.28259459e-01 -6.24487996e-01
-8.37258101e-01 9.24031079e-01 2.54205670e-02 -4.01420265e-01
-7.77145267e-01 7.81258106e-01 9.12009060e-01 -7.56237090e-01
4.80268933e-02 -1.69984221e-01 -1.68104857e-01 2.74580449e-01
8.09974551e-01 1.33812204e-01 -7.35905051e-05 -2.61271060e-01
-5.05299687e-01 3.89699697e-01 -3.91185194e-01 5.50478995e-01
1.42767048e+00 -6.82361424e-02 3.63149852e-01 5.80666840e-01
1.52109575e+00 -3.48163933e-01 -1.48082852e+00 -8.26871455e-01
7.99582899e-02 -5.80655158e-01 -3.96238774e-01 -7.17181623e-01
-9.39812005e-01 1.06385589e+00 7.46228158e-01 -8.71979073e-02
1.36149907e+00 1.47484258e-01 1.09256864e+00 1.14536859e-01
6.12323642e-01 -1.15172124e+00 1.38272762e-01 4.13378000e-01
5.66149592e-01 -1.42388618e+00 -1.24073267e-01 -4.47353095e-01
-6.88065827e-01 1.07192850e+00 8.32503974e-01 -2.28572905e-01
8.59027088e-01 -3.22673246e-02 -3.12594473e-01 1.76085725e-01
-6.61884069e-01 -1.04527444e-01 1.97528258e-01 8.32828701e-01
3.89304996e-01 1.35046810e-01 -2.96098739e-01 1.17159510e+00
2.29916275e-01 2.89102793e-01 3.01253110e-01 1.25754869e+00
-4.21264976e-01 -1.12260163e+00 -1.37021288e-01 7.28371799e-01
-2.94192851e-01 -3.55405733e-02 2.00715419e-02 7.71902859e-01
6.50624514e-01 7.87717164e-01 1.27004357e-02 -2.40784481e-01
4.73279148e-01 -3.07573681e-03 2.15930864e-01 -9.23529506e-01
-1.64095417e-01 -5.04877679e-02 -1.29771709e-01 -4.25994843e-01
-2.67689317e-01 -6.16007030e-01 -1.09447277e+00 5.18038087e-02
-2.27791637e-01 8.58956948e-02 -1.28517315e-01 9.95232344e-01
5.24327815e-01 5.27648687e-01 7.54796207e-01 -4.31160986e-01
-8.37910235e-01 -8.71073186e-01 -9.74301279e-01 7.45110512e-01
4.83374596e-01 -1.04295695e+00 -2.52855122e-01 -4.28652503e-02] | [10.095932006835938, 3.0800862312316895] |
4ef7753a-ed6f-4664-b08f-a9988176051e | videomae-v2-scaling-video-masked-autoencoders | 2303.16727 | null | https://arxiv.org/abs/2303.16727v2 | https://arxiv.org/pdf/2303.16727v2.pdf | VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking | Scale is the primary factor for building a powerful foundation model that could well generalize to a variety of downstream tasks. However, it is still challenging to train video foundation models with billions of parameters. This paper shows that video masked autoencoder (VideoMAE) is a scalable and general self-supervised pre-trainer for building video foundation models. We scale the VideoMAE in both model and data with a core design. Specifically, we present a dual masking strategy for efficient pre-training, with an encoder operating on a subset of video tokens and a decoder processing another subset of video tokens. Although VideoMAE is very efficient due to high masking ratio in encoder, masking decoder can still further reduce the overall computational cost. This enables the efficient pre-training of billion-level models in video. We also use a progressive training paradigm that involves an initial pre-training on a diverse multi-sourced unlabeled dataset, followed by a post-pre-training on a mixed labeled dataset. Finally, we successfully train a video ViT model with a billion parameters, which achieves a new state-of-the-art performance on the datasets of Kinetics (90.0% on K400 and 89.9% on K600) and Something-Something (68.7% on V1 and 77.0% on V2). In addition, we extensively verify the pre-trained video ViT models on a variety of downstream tasks, demonstrating its effectiveness as a general video representation learner. The code and model is available at \url{https://github.com/OpenGVLab/VideoMAEv2}. | ['Yu Qiao', 'Yali Wang', 'Yi Wang', 'Yinan He', 'Zhan Tong', 'Zhiyu Zhao', 'Bingkun Huang', 'LiMin Wang'] | 2023-03-29 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_VideoMAE_V2_Scaling_Video_Masked_Autoencoders_With_Dual_Masking_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_VideoMAE_V2_Scaling_Video_Masked_Autoencoders_With_Dual_Masking_CVPR_2023_paper.pdf | cvpr-2023-1 | ['action-classification', 'action-recognition-in-videos', 'spatio-temporal-action-localization'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.15675822e-02 -2.71966994e-01 -1.15927361e-01 -2.25432888e-01
-8.52529585e-01 -4.97656912e-01 6.16004944e-01 -5.14021277e-01
-4.55028862e-01 5.11609554e-01 1.21311441e-01 -4.54518139e-01
5.72165728e-01 -4.52575028e-01 -1.31283438e+00 -7.74032116e-01
-5.18488698e-02 9.81193734e-04 4.63408768e-01 -7.82761723e-02
-1.57427981e-01 -5.08653047e-03 -1.72005403e+00 9.48382854e-01
4.08512771e-01 9.07392919e-01 7.16692567e-01 1.08014619e+00
2.03210354e-01 1.27280545e+00 -5.29861331e-01 -3.84283304e-01
1.59709752e-01 -4.59929407e-01 -6.80827498e-01 -1.42740807e-03
3.90929818e-01 -5.61730921e-01 -9.09163415e-01 6.62554204e-01
5.01918554e-01 -5.90044335e-02 5.35456359e-01 -1.36909544e+00
-6.48588896e-01 6.92341805e-01 -3.72847944e-01 4.74732369e-01
2.95803659e-02 5.18734455e-01 8.25556576e-01 -1.17489409e+00
5.01371384e-01 7.84787655e-01 5.84194660e-01 8.94843400e-01
-8.50691855e-01 -7.45466292e-01 1.15743652e-01 3.29906076e-01
-1.42126274e+00 -6.74563169e-01 1.17435247e-01 -6.14525735e-01
1.31178236e+00 2.04131715e-02 6.30870938e-01 1.61058986e+00
1.99815333e-01 8.62830579e-01 1.09618974e+00 -9.46312249e-02
1.41980201e-01 -1.00008370e-02 -4.28078026e-02 7.57942975e-01
-7.73282200e-02 2.37727642e-01 -8.46852839e-01 2.37649143e-01
8.00507307e-01 5.19192638e-03 -4.44155306e-01 1.95133492e-01
-1.29562962e+00 5.83743393e-01 1.85860559e-01 9.32232738e-02
-2.70736426e-01 6.18188024e-01 5.52234888e-01 4.34382707e-01
2.65963078e-01 -1.86628371e-01 -6.25014305e-01 -4.92840588e-01
-1.10165942e+00 -1.55228507e-02 6.43189132e-01 1.22740817e+00
7.48037159e-01 4.45748299e-01 -2.12727904e-01 4.91016954e-01
2.51387805e-01 4.42468852e-01 7.49219954e-01 -1.04539859e+00
6.23080969e-01 1.75931811e-01 -2.69537687e-01 -2.93098032e-01
2.24976204e-02 -3.33904892e-01 -8.64843369e-01 8.31727758e-02
3.23170751e-01 -1.65360808e-01 -1.21965194e+00 1.88370287e+00
1.23895563e-01 7.49317527e-01 2.21711442e-01 7.66146064e-01
1.20743811e+00 1.09856761e+00 1.96295857e-01 -1.62758127e-01
1.36160147e+00 -1.36994052e+00 -3.63308460e-01 -1.78379610e-01
8.77178669e-01 -5.34750223e-01 1.16611862e+00 4.04218882e-01
-1.12812352e+00 -9.31234360e-01 -1.01378417e+00 -1.49982035e-01
-2.12772191e-01 2.49593347e-01 6.64347947e-01 4.61161524e-01
-1.42406416e+00 6.29368722e-01 -9.82132077e-01 -1.31592631e-01
4.17305261e-01 5.02146661e-01 -5.17515361e-01 -1.43503785e-01
-1.18569338e+00 5.18311262e-01 4.36496466e-01 -9.17470157e-02
-1.83045018e+00 -8.38494182e-01 -8.85386169e-01 1.92671046e-01
2.48155802e-01 -7.00028777e-01 1.36657369e+00 -1.09064996e+00
-1.48592770e+00 7.74944365e-01 -1.36817053e-01 -4.94243950e-01
3.83954704e-01 -2.56335974e-01 -3.73537630e-01 3.38783145e-01
-2.83958111e-02 1.02145433e+00 1.00620174e+00 -8.90991330e-01
-6.26618326e-01 9.12749916e-02 1.58498183e-01 2.06286103e-01
-5.22696793e-01 4.59335223e-02 -9.97340739e-01 -6.20023727e-01
-6.01813138e-01 -9.07791436e-01 1.50072947e-01 -4.31898147e-01
1.11440808e-01 -3.72281224e-02 9.12137866e-01 -7.86088228e-01
1.27047145e+00 -2.34064794e+00 2.23702073e-01 -4.28470880e-01
3.56551200e-01 5.05639195e-01 -4.21390027e-01 4.27936226e-01
-2.69674629e-01 2.20398322e-01 -3.08913961e-02 -5.10410309e-01
-3.10530275e-01 8.28040913e-02 -3.07232112e-01 3.26238900e-01
2.57247865e-01 1.09761429e+00 -7.78233349e-01 -2.76469231e-01
1.11793280e-01 6.75262392e-01 -6.61965132e-01 4.57575083e-01
-1.52239472e-01 3.84400576e-01 -2.80338489e-02 6.67906702e-01
3.21672291e-01 -6.22157276e-01 1.16816070e-02 -1.47141337e-01
-1.12918995e-01 1.83162525e-01 -8.89283717e-01 1.90266895e+00
-2.62393296e-01 8.73433650e-01 2.82658190e-02 -1.06316268e+00
3.56846064e-01 7.03210831e-01 4.32885885e-01 -5.89593172e-01
2.18242317e-01 1.25421181e-01 -5.44349253e-02 -5.64919889e-01
3.20144027e-01 1.32077187e-01 9.33215246e-02 4.05233681e-01
5.43239295e-01 1.56023756e-01 4.13911283e-01 6.37061536e-01
1.37389219e+00 4.19492513e-01 -1.06579795e-01 -8.02785754e-02
4.14540201e-01 -1.95228644e-02 4.65113312e-01 6.55497313e-01
-1.85592040e-01 6.06363297e-01 3.72695327e-01 -2.76883245e-01
-1.24627483e+00 -9.34908032e-01 2.23676339e-01 1.27122772e+00
-1.12992875e-01 -8.96044075e-01 -9.05015707e-01 -5.28658330e-01
-1.39417335e-01 3.12997937e-01 -6.32254004e-01 -3.42342615e-01
-4.91973460e-01 -6.35566413e-01 8.28768790e-01 8.47826958e-01
6.49984181e-01 -1.03204620e+00 -3.17624897e-01 9.82058048e-02
-1.94094077e-01 -1.37026465e+00 -4.93389070e-01 3.01979631e-01
-9.21282411e-01 -8.46745312e-01 -7.19276428e-01 -9.73053694e-01
3.38841707e-01 6.81938648e-01 1.09124959e+00 3.77781242e-01
3.56709026e-02 3.43563259e-01 -4.53428835e-01 -1.07403100e-01
-5.59426188e-01 2.12997128e-03 1.91701263e-01 -1.91580638e-01
2.36998826e-01 -5.04012227e-01 -6.39976323e-01 4.16384041e-01
-8.84363353e-01 4.64034528e-01 6.24110341e-01 8.46481025e-01
6.51567638e-01 -1.95005715e-01 4.35297728e-01 -6.99547291e-01
4.99401912e-02 -7.69630551e-01 -5.61259151e-01 1.03310138e-01
-3.91789764e-01 -2.05179900e-01 7.20831454e-01 -6.70185983e-01
-8.72828007e-01 6.97871372e-02 -2.08106667e-01 -8.89121890e-01
1.22154625e-02 5.13328254e-01 -8.16938654e-02 1.03351980e-01
6.03905261e-01 5.07282376e-01 -3.75372823e-03 -2.94092208e-01
4.03897583e-01 7.19560862e-01 6.35142326e-01 -5.18547952e-01
7.03198612e-01 4.05683219e-01 -3.61877680e-01 -8.40694189e-01
-5.50460398e-01 -4.54591632e-01 -3.22721690e-01 -3.43791157e-01
1.03449368e+00 -1.73338890e+00 -6.81495667e-01 7.12753773e-01
-8.65490556e-01 -1.05716169e+00 2.92731654e-02 5.02352178e-01
-6.36275768e-01 3.15448731e-01 -1.03121853e+00 -2.21574441e-01
-3.60123038e-01 -1.34093130e+00 9.78664517e-01 8.81598741e-02
1.32564411e-01 -7.51969695e-01 -4.68629375e-02 5.39708853e-01
3.10871869e-01 -2.26477593e-01 4.16062951e-01 -4.16092098e-01
-7.78540671e-01 1.33308455e-01 -2.07728788e-01 6.22006238e-01
-1.27329722e-01 1.61967084e-01 -1.24331999e+00 -4.84323978e-01
1.16254045e-02 -7.20712960e-01 1.10998237e+00 3.50061208e-01
1.26835191e+00 1.04139093e-02 -2.34838486e-01 1.13061297e+00
1.37743521e+00 6.96076080e-02 8.40198278e-01 2.54686594e-01
9.01773453e-01 5.44321798e-02 3.04846853e-01 3.28512579e-01
4.35457796e-01 3.92282695e-01 3.56946826e-01 8.60025287e-02
-3.31234038e-01 -2.91958988e-01 1.01891100e+00 1.23409176e+00
-3.36938024e-01 -4.53132302e-01 -7.99352884e-01 4.92138594e-01
-1.77067232e+00 -1.19697201e+00 -1.53425604e-01 1.92098188e+00
8.04834366e-01 7.20805675e-02 2.67824143e-01 -3.12470645e-02
6.82753801e-01 3.76387760e-02 -4.44965452e-01 -1.70179844e-01
-1.63894355e-01 1.40003726e-01 5.02699912e-01 2.22863555e-01
-1.11083603e+00 1.35319483e+00 6.38420391e+00 9.23898518e-01
-1.21561587e+00 4.30902690e-01 7.04635262e-01 -3.15213501e-01
-6.36831224e-02 3.58811766e-03 -1.08902800e+00 7.22486079e-01
1.66867983e+00 -9.98066366e-02 6.05074584e-01 8.41387093e-01
2.37514779e-01 1.22680500e-01 -1.14302802e+00 1.07179427e+00
-6.52195290e-02 -1.71608102e+00 9.12770629e-02 8.47152248e-02
7.69929111e-01 5.44886112e-01 -6.33081561e-03 8.24546695e-01
1.76084891e-01 -1.09673047e+00 9.45559680e-01 8.95348862e-02
1.31898665e+00 -3.00574124e-01 4.79220748e-01 4.07919645e-01
-1.47800612e+00 -1.16912246e-01 -5.49101591e-01 -1.70185119e-01
1.15586318e-01 1.96928322e-01 -5.45866609e-01 3.67577106e-01
9.97443855e-01 8.78333926e-01 -4.92752641e-01 7.69645333e-01
-2.72848457e-01 1.05955362e+00 -1.74927607e-01 2.79435128e-01
2.42037207e-01 1.18450738e-01 1.90958485e-03 1.55655766e+00
3.61089379e-01 2.12575138e-01 2.98420917e-02 3.47068101e-01
-3.63786727e-01 -1.00221999e-01 -5.94293475e-01 -2.47212052e-01
3.55974138e-01 1.24647546e+00 -5.23908496e-01 -7.55928338e-01
-9.00651515e-01 1.14516461e+00 5.23767054e-01 5.55536449e-01
-1.41445494e+00 6.99061751e-02 6.13441765e-01 -2.13659517e-02
7.03448474e-01 -3.16148400e-01 3.08517903e-01 -1.53659487e+00
-1.59543559e-01 -1.17937744e+00 4.71922815e-01 -9.63430583e-01
-8.25051188e-01 6.90332949e-01 -3.52554843e-02 -1.23364556e+00
-1.48340091e-01 -8.21033180e-01 -6.37280762e-01 5.14379859e-01
-1.47344100e+00 -1.13281238e+00 -5.26248217e-01 9.47496712e-01
8.98646355e-01 -3.37245673e-01 6.05809331e-01 5.68208694e-01
-8.15455973e-01 6.68771625e-01 1.15987569e-01 2.10653767e-01
6.04523897e-01 -1.04616284e+00 5.78534544e-01 9.90804374e-01
2.33784750e-01 5.41220129e-01 3.06717604e-01 -5.85333288e-01
-1.77112162e+00 -1.26417482e+00 5.19874096e-01 -4.46994215e-01
7.06043124e-01 -5.42714894e-01 -8.39622378e-01 1.07670772e+00
3.77730370e-01 1.69109046e-01 7.04113781e-01 -3.03317338e-01
-4.85527128e-01 -3.52333300e-02 -4.73444372e-01 6.29103959e-01
1.31906867e+00 -6.93412125e-01 -3.05507302e-01 4.29872721e-01
1.02202559e+00 -5.74610710e-01 -9.13217008e-01 2.80140102e-01
4.41343546e-01 -1.04205263e+00 8.79299462e-01 -6.78681910e-01
7.37139046e-01 -3.29889983e-01 -3.32497150e-01 -1.01726305e+00
-4.16989714e-01 -8.30918491e-01 -6.64550364e-01 1.12884104e+00
4.75014001e-01 -2.70233572e-01 8.21575940e-01 2.67733663e-01
-4.51916575e-01 -8.18706989e-01 -6.57786429e-01 -9.15385962e-01
-3.38804871e-02 -7.38577604e-01 2.30816767e-01 6.98889792e-01
-1.05539665e-01 4.94807601e-01 -6.52498662e-01 7.79721364e-02
3.90198857e-01 -3.44451964e-01 9.85227585e-01 -4.55831081e-01
-6.14801228e-01 -1.80237249e-01 -3.39360803e-01 -1.33264601e+00
6.46930709e-02 -1.09225845e+00 -1.25867650e-01 -1.27469933e+00
3.89499754e-01 -8.32754746e-02 -4.44797784e-01 5.52045405e-01
-2.88027346e-01 4.03449327e-01 2.37143546e-01 4.94707316e-01
-6.63319111e-01 5.44525623e-01 1.08290219e+00 -4.78171706e-02
-1.79437865e-02 -2.60944158e-01 -6.18445873e-01 5.04841745e-01
8.35320652e-01 -5.34975827e-01 -6.50106013e-01 -7.93713510e-01
1.34067675e-02 7.43306428e-02 3.09683502e-01 -1.13963711e+00
1.64550900e-01 3.97197939e-02 3.63879710e-01 -4.49191391e-01
4.26593453e-01 -4.39209998e-01 3.09819847e-01 4.92392659e-01
-5.06872032e-03 4.28774625e-01 4.46590185e-01 5.38758993e-01
-1.13463439e-01 -3.33106443e-02 5.71023941e-01 -2.94122338e-01
-1.18297279e+00 6.31036758e-01 -5.37368894e-01 1.96703538e-01
1.04891026e+00 -1.96295366e-01 -5.46284437e-01 -3.90565127e-01
-6.36389494e-01 1.81890592e-01 5.95322490e-01 2.57651955e-01
6.83530629e-01 -1.16958809e+00 -8.17156136e-01 1.80761248e-01
-7.25796074e-03 -2.51102597e-01 4.61138427e-01 8.92796814e-01
-7.32407510e-01 3.15898389e-01 -1.83336258e-01 -8.43327522e-01
-1.22735906e+00 7.88593352e-01 1.67115182e-01 -4.33690213e-02
-7.24411488e-01 1.15796709e+00 5.04201829e-01 1.62693813e-01
2.53688574e-01 -2.61607617e-01 1.28735723e-02 -2.03085110e-01
8.90023291e-01 1.88283935e-01 -8.17716718e-02 -7.19947278e-01
-3.27752501e-01 4.44296539e-01 -1.52849823e-01 -1.47086069e-01
1.37284267e+00 1.12314180e-01 2.21209332e-01 2.57210225e-01
1.38705885e+00 -1.84010416e-01 -1.65753841e+00 -3.02482769e-02
-4.45532590e-01 -1.01098590e-01 3.08126602e-02 -4.19776827e-01
-1.23578703e+00 9.87263501e-01 3.33428234e-01 -1.43446803e-01
1.14435959e+00 4.56655584e-02 8.89667809e-01 2.82205522e-01
2.00918227e-01 -9.11027670e-01 3.24523926e-01 5.91342568e-01
6.58318758e-01 -1.29104495e+00 -1.02829672e-01 -2.48132542e-01
-9.68624175e-01 8.44335437e-01 7.89217472e-01 -1.76274180e-02
6.07459128e-01 4.48595017e-01 1.53539747e-01 -1.92340910e-01
-1.41633534e+00 -1.17180422e-01 7.43264556e-02 3.82987350e-01
5.39140821e-01 -1.47465512e-01 -3.96455340e-02 6.47686660e-01
-5.11113666e-02 2.95663357e-01 5.54643333e-01 8.96957755e-01
-2.66031325e-01 -8.91068876e-01 2.50735655e-02 2.39161521e-01
-5.63652337e-01 -4.61774588e-01 5.46032786e-02 8.33605289e-01
3.97077240e-02 9.14687216e-01 -4.31606881e-02 -9.20443714e-01
-1.09610021e-01 1.76358968e-01 4.81260329e-01 -7.35907197e-01
-7.63817430e-01 2.38205731e-01 1.83330551e-02 -7.06350386e-01
-3.75913799e-01 -4.31276262e-01 -1.20377851e+00 -6.04559064e-01
-1.82136074e-01 4.78619821e-02 5.30179024e-01 7.89788783e-01
5.42075276e-01 5.37386417e-01 4.10248369e-01 -1.00873387e+00
-4.20955151e-01 -9.23303366e-01 -3.06395531e-01 4.55611438e-01
1.22158453e-01 -4.97021675e-01 -5.10094821e-01 5.81901848e-01] | [9.425082206726074, 0.8299190402030945] |
a9014f0b-0208-4b04-802f-a7f3ac73e478 | foundation-models-for-natural-language | 2302.08575 | null | https://arxiv.org/abs/2302.08575v1 | https://arxiv.org/pdf/2302.08575v1.pdf | Foundation Models for Natural Language Processing -- Pre-trained Language Models Integrating Media | This open access book provides a comprehensive overview of the state of the art in research and applications of Foundation Models and is intended for readers familiar with basic Natural Language Processing (NLP) concepts. Over the recent years, a revolutionary new paradigm has been developed for training models for NLP. These models are first pre-trained on large collections of text documents to acquire general syntactic knowledge and semantic information. Then, they are fine-tuned for specific tasks, which they can often solve with superhuman accuracy. When the models are large enough, they can be instructed by prompts to solve new tasks without any fine-tuning. Moreover, they can be applied to a wide range of different media and problem domains, ranging from image and video processing to robot control learning. Because they provide a blueprint for solving many tasks in artificial intelligence, they have been called Foundation Models. After a brief introduction to basic NLP models the main pre-trained language models BERT, GPT and sequence-to-sequence transformer are described, as well as the concepts of self-attention and context-sensitive embedding. Then, different approaches to improving these models are discussed, such as expanding the pre-training criteria, increasing the length of input texts, or including extra knowledge. An overview of the best-performing models for about twenty application areas is then presented, e.g., question answering, translation, story generation, dialog systems, generating images from text, etc. For each application area, the strengths and weaknesses of current models are discussed, and an outlook on further developments is given. In addition, links are provided to freely available program code. A concluding chapter summarizes the economic opportunities, mitigation of risks, and potential developments of AI. | ['Sven Giesselbach', 'Gerhard Paaß'] | 2023-02-16 | null | null | null | null | ['story-generation'] | ['natural-language-processing'] | [ 4.47049975e-01 3.34060669e-01 -3.64586711e-01 -3.01303923e-01
-5.32121360e-01 -5.70818067e-01 7.38079131e-01 5.42242490e-02
-4.44637060e-01 8.54757428e-01 1.05044760e-01 -1.60510793e-01
-8.32634270e-02 -9.70640779e-01 -5.75609028e-01 -5.34763873e-01
2.48477504e-01 6.73347771e-01 -8.92899185e-03 -5.42396367e-01
4.30663139e-01 3.03144306e-01 -1.66324747e+00 4.24058735e-01
9.04911339e-01 8.44978809e-01 6.71583295e-01 5.29740810e-01
-6.98126376e-01 7.40185738e-01 -8.11362565e-01 -6.61805153e-01
-1.80624470e-01 -5.07702410e-01 -1.07930148e+00 4.87727858e-02
-1.34674132e-01 -1.11749008e-01 -1.70694605e-01 9.42514300e-01
4.60312277e-01 1.53313234e-01 8.51648152e-01 -1.22433603e+00
-1.16219544e+00 7.41753757e-01 -1.49501413e-01 1.39611274e-01
7.86062956e-01 -3.03037129e-02 1.10064101e+00 -1.04511201e+00
6.21397078e-01 1.34869170e+00 3.15927535e-01 9.01091933e-01
-8.95381689e-01 -3.94305229e-01 2.55864114e-01 6.87798679e-01
-1.03664744e+00 -8.80760774e-02 7.17528224e-01 -3.72871339e-01
1.29506958e+00 3.40620130e-01 5.03753245e-01 1.41814768e+00
3.42712551e-01 1.13619757e+00 8.28456223e-01 -8.53206575e-01
7.26695806e-02 5.41800380e-01 9.72125456e-02 5.07102609e-01
-1.46487951e-01 7.75669515e-02 -5.43130577e-01 1.49407350e-02
6.17871642e-01 -2.27329180e-01 -2.52760202e-01 -1.46294385e-01
-1.33002973e+00 1.10916531e+00 2.25566849e-01 7.33368576e-01
-2.83477098e-01 -3.16593230e-01 3.94689739e-01 2.96254456e-01
2.59958893e-01 8.14241171e-01 -7.53721058e-01 1.40123218e-01
-5.11412621e-01 4.05016541e-01 8.86556685e-01 1.04206848e+00
4.17884499e-01 1.05703361e-01 -2.59831786e-01 1.05965805e+00
2.17623204e-01 4.13277119e-01 1.02848411e+00 -8.28250051e-01
3.88387024e-01 4.21938688e-01 8.39283690e-02 -9.95711029e-01
-4.42653686e-01 -3.86197492e-02 -7.72928178e-01 -1.44918129e-01
4.16057855e-02 -1.10138752e-01 -8.10686529e-01 1.52164900e+00
1.16207592e-01 -2.92905331e-01 2.68811941e-01 5.89038730e-01
1.16275179e+00 1.24402821e+00 1.76685914e-01 -3.64599109e-01
1.77125466e+00 -1.30156994e+00 -1.02398098e+00 -5.41240931e-01
4.37196702e-01 -6.39378846e-01 1.28828168e+00 2.07355812e-01
-1.12610424e+00 -6.48396671e-01 -7.25709438e-01 -2.97887236e-01
-9.49479342e-01 9.21817347e-02 6.54374957e-01 4.95439589e-01
-7.95955420e-01 4.80542541e-01 -4.11806136e-01 -6.01882339e-01
1.42589733e-01 2.79278725e-01 -7.01417997e-02 -1.31493166e-01
-1.69432116e+00 1.24153638e+00 6.94178402e-01 2.17838096e-03
-6.84213221e-01 -5.27516544e-01 -1.06414115e+00 7.66150281e-02
4.98245031e-01 -9.85772550e-01 1.44031239e+00 -1.15930903e+00
-1.91353166e+00 1.03359163e+00 -1.11886777e-01 -4.93058801e-01
1.77871794e-01 -3.67352277e-01 -5.06456614e-01 1.53078675e-01
1.29213423e-01 8.43168080e-01 7.56003976e-01 -9.96430397e-01
-7.32207835e-01 -4.62803952e-02 2.59685963e-01 4.34871674e-01
-3.58628362e-01 4.92992729e-01 -4.37375426e-01 -8.28794956e-01
-4.98887867e-01 -6.30003810e-01 -3.75907093e-01 -1.73928648e-01
-2.99430907e-01 -4.18900996e-01 6.65268242e-01 -5.61564922e-01
1.23348737e+00 -1.86330497e+00 3.30498308e-01 -2.56313503e-01
-2.32816622e-01 3.96277308e-01 -3.68231565e-01 7.83725202e-01
-1.10602014e-01 1.60451904e-01 -4.02390718e-01 1.75932273e-01
6.60531446e-02 2.88886607e-01 -5.55506110e-01 -1.36295721e-01
3.08487982e-01 1.23694801e+00 -9.72039223e-01 -4.16324705e-01
4.32083040e-01 3.96350473e-01 -1.28176093e-01 3.47939700e-01
-5.91422915e-01 4.03262973e-01 -5.48035979e-01 3.15110147e-01
1.16881892e-01 -4.09436584e-01 -5.80323078e-02 2.91245729e-01
-2.60651577e-02 3.25021416e-01 -7.61197567e-01 1.63978171e+00
-7.13103592e-01 5.56368589e-01 -1.10007890e-01 -1.34670126e+00
9.13632870e-01 6.41240776e-01 1.93786338e-01 -5.25978744e-01
1.16142087e-01 1.00757591e-01 -1.87261596e-01 -8.76855433e-01
4.35226649e-01 -2.33508632e-01 -1.79184750e-01 3.32207352e-01
1.52837083e-01 -4.65014547e-01 5.85380256e-01 -6.29171506e-02
7.14009762e-01 7.54227862e-02 7.52679288e-01 3.30414921e-02
1.10461509e+00 2.32657775e-01 2.02770084e-01 5.36338210e-01
-4.58451733e-02 2.93488115e-01 2.26023301e-01 -6.12096250e-01
-8.24672222e-01 -7.66840279e-01 -9.49940756e-02 1.45154691e+00
-4.61811991e-03 -3.29272479e-01 -8.33662689e-01 -5.89449286e-01
-2.67996579e-01 1.01276600e+00 -5.93706667e-01 -8.75969306e-02
-5.02937317e-01 -1.01455951e+00 2.56550848e-01 5.48973560e-01
7.14444697e-01 -1.77575612e+00 -5.78584313e-01 2.71915287e-01
-4.63577300e-01 -1.20364475e+00 -1.66131705e-01 -4.39247712e-02
-8.75426471e-01 -9.01987672e-01 -1.12712407e+00 -1.36115277e+00
6.21225774e-01 5.37878573e-02 1.18744111e+00 5.75439185e-02
-3.06576639e-01 5.61218500e-01 -5.09365797e-01 -6.40772104e-01
-3.66320223e-01 2.27920320e-02 1.03400424e-02 -2.34342232e-01
5.26277542e-01 -1.49806023e-01 -6.01361506e-02 1.58936158e-01
-9.17776942e-01 1.11565277e-01 7.18234360e-01 1.10724664e+00
5.33627212e-01 7.16086477e-02 7.66652584e-01 -9.53119278e-01
1.01545155e+00 -2.84853846e-01 -4.81926292e-01 5.65790474e-01
-5.30137539e-01 -1.57815348e-02 8.13831270e-01 -4.93424594e-01
-1.19337451e+00 -6.46108389e-02 -4.62004304e-01 7.36193359e-02
-4.53092009e-01 6.55402958e-01 -2.83316970e-01 4.90718186e-02
6.71592474e-01 4.59339738e-01 -1.96865603e-01 -3.80167156e-01
6.94493830e-01 6.80375934e-01 3.59084040e-01 -5.20529628e-01
6.56846166e-01 5.38283922e-02 -3.96227837e-01 -1.05597210e+00
-9.84920442e-01 -1.86566845e-01 -5.75045109e-01 4.06957008e-02
9.60640430e-01 -4.13620800e-01 -4.37897861e-01 3.70675564e-01
-1.39518976e+00 -3.33471596e-01 -3.54295969e-01 3.72223556e-01
-8.52065384e-01 3.51296186e-01 -8.14288914e-01 -5.94761848e-01
-5.09076238e-01 -1.03588450e+00 8.49429607e-01 3.91140193e-01
-2.59193391e-01 -1.31052041e+00 -5.98655492e-02 4.98311460e-01
3.29170674e-01 2.05627084e-03 1.39887774e+00 -8.63842964e-01
-3.09587955e-01 -1.65848300e-01 -1.21871522e-02 3.77016723e-01
7.84751996e-02 -1.84517190e-01 -8.23198199e-01 4.86999638e-02
1.31293833e-01 -5.23990929e-01 5.75557590e-01 3.36849719e-01
1.22988737e+00 -5.50325751e-01 -5.65511465e-01 2.64971167e-01
1.08174384e+00 5.13289332e-01 5.79430759e-01 5.83972931e-01
3.22050869e-01 1.07047081e+00 7.01909900e-01 1.33511528e-01
2.73774356e-01 5.66673577e-01 -5.07622026e-05 1.61768511e-01
7.46223181e-02 -2.00204149e-01 5.58927596e-01 9.57419872e-01
-4.80679348e-02 -5.06785035e-01 -8.64468157e-01 5.12456775e-01
-1.78513122e+00 -9.57556844e-01 2.03075975e-01 1.85078967e+00
8.88610840e-01 -1.13207355e-01 -1.68377683e-01 1.08722314e-01
9.00291860e-01 3.29125524e-02 -5.30829012e-01 -6.50387287e-01
-1.98146150e-01 1.15497366e-01 -2.52545550e-02 5.59005320e-01
-1.12682927e+00 1.34896934e+00 7.39043903e+00 9.77649510e-01
-9.54379976e-01 -1.89906769e-02 6.09177649e-01 2.95046985e-01
-1.37734339e-01 -3.07617068e-01 -7.62538314e-01 2.71941751e-01
1.03531528e+00 -5.76879084e-01 5.19509733e-01 9.91553783e-01
-1.06211249e-02 5.73277250e-02 -1.12046647e+00 1.09016967e+00
3.68600577e-01 -1.42692268e+00 2.34582067e-01 -3.30156446e-01
5.28950095e-01 -1.83023319e-01 1.21201620e-01 6.25142753e-01
1.40460745e-01 -1.09589493e+00 4.40410703e-01 1.57536641e-01
6.24510288e-01 -6.47928238e-01 8.36329222e-01 6.58712029e-01
-8.25455725e-01 -2.55828232e-01 -5.57915390e-01 -1.63158774e-01
3.74338120e-01 3.77345055e-01 -5.26025414e-01 6.02942228e-01
7.42515087e-01 6.00874484e-01 -3.17953378e-01 8.32085609e-01
-8.54689658e-01 3.31895709e-01 -8.82304832e-02 -4.47161943e-01
1.78945750e-01 -4.72104959e-02 3.72698516e-01 1.15605104e+00
2.54889250e-01 3.97896200e-01 -4.39888500e-02 9.09031868e-01
1.15174569e-01 4.85450149e-01 -6.90563262e-01 -3.46625745e-01
1.85752094e-01 1.19973087e+00 -6.05376661e-01 -3.92723411e-01
-7.01708853e-01 1.11385870e+00 2.11073801e-01 4.90482867e-01
-7.68654764e-01 -7.71801353e-01 3.52040678e-01 -1.81245804e-01
7.99600631e-02 6.10429943e-02 -1.50313392e-01 -1.15415490e+00
-2.96596646e-01 -1.03592849e+00 4.88834292e-01 -9.22095299e-01
-1.30584800e+00 8.08176875e-01 2.79458225e-01 -9.19171691e-01
-5.23283839e-01 -1.06719983e+00 -5.12455404e-01 6.77971363e-01
-1.62923694e+00 -8.68693650e-01 4.12112884e-02 4.75024849e-01
9.62894857e-01 -3.91195953e-01 1.20972478e+00 -5.62652014e-02
-5.25557697e-01 3.08259457e-01 1.99830964e-01 2.53771633e-01
5.81784487e-01 -9.75792289e-01 3.54933470e-01 5.06005347e-01
3.63867432e-01 5.04286289e-01 6.60910726e-01 -4.37064081e-01
-1.13636959e+00 -8.93508613e-01 1.39839685e+00 -3.52764368e-01
6.82828248e-01 -2.97642171e-01 -8.75568151e-01 6.57070994e-01
5.88027596e-01 -5.21503687e-01 8.45127881e-01 -4.36486080e-02
-3.81628983e-02 8.56295973e-02 -1.08939147e+00 7.50428677e-01
6.73291802e-01 -2.76234746e-01 -9.73592520e-01 8.21520388e-01
9.32847440e-01 -3.88534069e-01 -6.47448123e-01 1.23574905e-01
2.48868302e-01 -7.11867452e-01 1.01094782e+00 -6.65810287e-01
5.56366265e-01 -4.35612202e-02 1.39115632e-01 -1.19312096e+00
-5.99440098e-01 -7.67970085e-01 -1.56390294e-01 9.83667552e-01
5.91594398e-01 -6.65993392e-01 5.12265623e-01 6.48689091e-01
-1.70630053e-01 -8.40176642e-01 -5.13047278e-01 -7.85121858e-01
2.99574673e-01 -3.57232898e-01 5.84732831e-01 8.21832418e-01
3.75508606e-01 8.90727282e-01 -3.54608864e-01 -1.94794148e-01
1.83739319e-01 1.53366417e-01 4.32064265e-01 -1.08826542e+00
-2.18332693e-01 -5.55811703e-01 -1.11832172e-01 -1.29253125e+00
4.04144913e-01 -8.91513705e-01 2.00159289e-02 -1.81598985e+00
1.13593817e-01 1.99877515e-01 -5.79856411e-02 4.87436354e-01
-1.86169475e-01 -1.09416666e-02 2.40121320e-01 1.33510619e-01
-2.87758827e-01 6.25958920e-01 1.35020041e+00 -2.94495791e-01
-2.00815037e-01 7.68060610e-02 -8.25852454e-01 9.55198586e-01
9.30175841e-01 -2.49402791e-01 -5.32753527e-01 -6.49127960e-01
2.92523772e-01 -1.22182004e-01 6.68775588e-02 -6.51425958e-01
7.23154843e-02 -4.20273572e-01 3.12162280e-01 -4.65303391e-01
4.10402775e-01 -8.35707784e-01 -3.74300033e-01 4.63077515e-01
-6.05148017e-01 1.21232025e-01 2.52339572e-01 4.59764630e-01
-4.43256378e-01 -8.23504448e-01 6.47109151e-01 -5.51974773e-01
-1.05274844e+00 1.57927141e-01 -7.11854935e-01 -1.14243124e-02
1.31333137e+00 -1.10895976e-01 -6.12266622e-02 -5.73183119e-01
-7.42097199e-01 4.40971792e-01 -4.82198084e-03 8.25506985e-01
7.60980725e-01 -1.25372100e+00 -7.25735247e-01 6.10689484e-02
1.89173892e-01 -1.84632406e-01 1.83603972e-01 5.03278852e-01
-4.46948051e-01 9.89076793e-01 -9.02150199e-02 -1.85866326e-01
-1.14548182e+00 1.08377242e+00 1.96782853e-02 -4.57287788e-01
-5.25173843e-01 9.88335609e-01 5.09655058e-01 -5.79845071e-01
3.31332773e-01 -2.88538367e-01 -7.84415185e-01 -1.88373372e-01
9.64252830e-01 2.33399402e-02 -2.12957636e-01 -6.78388298e-01
-3.61125797e-01 7.40325868e-01 -1.29623860e-01 -1.66306779e-01
1.21221101e+00 -2.43514225e-01 -2.98689842e-01 5.65867066e-01
8.76078606e-01 -4.79587048e-01 -5.89965403e-01 -1.92884728e-01
1.81911960e-01 -4.38021459e-02 -3.24437201e-01 -1.02779460e+00
-7.88789630e-01 1.17358923e+00 1.48134395e-01 1.96839705e-01
1.16857028e+00 1.39836341e-01 9.01444256e-01 7.72176325e-01
4.49918270e-01 -1.28312206e+00 2.52498358e-01 9.26339686e-01
1.15969634e+00 -1.11324251e+00 -1.27901077e-01 -4.56436813e-01
-9.34688509e-01 1.34411645e+00 4.98823941e-01 1.15371652e-01
6.04150474e-01 1.66835368e-01 2.80287359e-02 1.26221389e-01
-7.63607681e-01 -7.91239738e-02 2.56471306e-01 1.01750648e+00
5.53127944e-01 -1.95436940e-01 -4.51139599e-01 8.92621934e-01
-4.78286296e-01 -3.22996601e-02 3.51810545e-01 7.76641071e-01
-6.40812874e-01 -1.41522098e+00 -4.38818455e-01 2.41438702e-01
-5.67322016e-01 -2.07827479e-01 -6.45822585e-01 6.97273195e-01
4.00802344e-02 1.06883979e+00 -1.61984265e-01 -1.43857032e-01
2.21066281e-01 3.90431076e-01 4.94511753e-01 -9.06474411e-01
-5.24011433e-01 -1.84261516e-01 1.01685658e-01 -2.69176066e-01
-5.02765954e-01 -4.84140754e-01 -1.25237083e+00 -6.32162392e-02
-3.12283248e-01 2.86260545e-01 5.19530892e-01 9.73816633e-01
2.37021238e-01 4.04268652e-01 2.94627428e-01 -8.57049167e-01
-3.28525126e-01 -9.05694187e-01 -2.60980904e-01 1.46524087e-01
-2.35205330e-02 -4.42579716e-01 -2.01446980e-01 3.76073480e-01] | [11.002580642700195, 8.65248966217041] |
603a8a08-62c2-43b1-be25-96ff815056d4 | evaluation-beyond-task-performance-analyzing | 2211.14673 | null | https://arxiv.org/abs/2211.14673v1 | https://arxiv.org/pdf/2211.14673v1.pdf | Evaluation Beyond Task Performance: Analyzing Concepts in AlphaZero in Hex | AlphaZero, an approach to reinforcement learning that couples neural networks and Monte Carlo tree search (MCTS), has produced state-of-the-art strategies for traditional board games like chess, Go, shogi, and Hex. While researchers and game commentators have suggested that AlphaZero uses concepts that humans consider important, it is unclear how these concepts are captured in the network. We investigate AlphaZero's internal representations in the game of Hex using two evaluation techniques from natural language processing (NLP): model probing and behavioral tests. In doing so, we introduce new evaluation tools to the RL community and illustrate how evaluations other than task performance can be used to provide a more complete picture of a model's strengths and weaknesses. Our analyses in the game of Hex reveal interesting patterns and generate some testable hypotheses about how such models learn in general. For example, we find that MCTS discovers concepts before the neural network learns to encode them. We also find that concepts related to short-term end-game planning are best encoded in the final layers of the model, whereas concepts related to long-term planning are encoded in the middle layers of the model. | ['Michael L. Littman', 'Ellie Pavlick', 'George Konidaris', 'Jessica Zosa Forde', 'Charles Lovering'] | 2022-11-26 | null | null | null | null | ['board-games'] | ['playing-games'] | [-1.39808938e-01 4.68976468e-01 -1.02400951e-01 -3.49285044e-02
-4.85151261e-01 -7.75999963e-01 5.45654237e-01 3.30909133e-01
-7.61915147e-01 7.30578780e-01 5.61235845e-01 -4.84303772e-01
-4.28714484e-01 -1.03057921e+00 -5.21648228e-01 -2.98540980e-01
-3.37366968e-01 7.68667638e-01 3.52126479e-01 -7.21541226e-01
5.59800208e-01 1.85936108e-01 -1.39598227e+00 5.75519741e-01
5.49797654e-01 7.21897006e-01 2.04594269e-01 6.24762833e-01
1.19848527e-01 1.47924924e+00 -5.81817985e-01 -4.34981853e-01
2.69308150e-01 -5.91232419e-01 -1.16858733e+00 -4.81956333e-01
-1.45613626e-02 -5.74877262e-01 -2.82728910e-01 8.96231771e-01
2.49362528e-01 4.63855445e-01 4.86146808e-01 -9.08662021e-01
-3.24699640e-01 1.17922413e+00 -7.20014647e-02 5.77626884e-01
2.83554435e-01 9.24639046e-01 1.19563496e+00 -1.85250580e-01
7.05976844e-01 1.26257885e+00 6.31231487e-01 6.77337468e-01
-1.17585015e+00 -5.48462152e-01 2.73211449e-01 1.53335154e-01
-9.35864508e-01 -3.45867455e-01 4.92504865e-01 -4.14661884e-01
1.30624068e+00 -1.06879801e-01 1.35146761e+00 1.20471179e+00
3.71497512e-01 1.04688442e+00 1.30314004e+00 -3.07499677e-01
6.71995759e-01 -3.92069936e-01 -7.71506652e-02 8.44092429e-01
1.98413238e-01 1.05196297e+00 -8.85295510e-01 -3.44952270e-02
1.18047738e+00 -5.10057986e-01 1.56013697e-01 -4.40410644e-01
-9.18948174e-01 1.15997577e+00 4.62831885e-01 4.81143564e-01
-6.11713052e-01 6.12423420e-01 5.25840163e-01 1.97856843e-01
-5.11298403e-02 1.28533554e+00 -5.72800815e-01 -7.14174926e-01
-8.97483349e-01 6.34741366e-01 7.57905185e-01 5.00949502e-01
6.35864258e-01 3.25467288e-01 -2.01915935e-01 5.38628995e-01
2.81959474e-01 -1.31586879e-01 7.40947366e-01 -1.46513104e+00
5.30015409e-01 2.98017591e-01 4.01386544e-02 -8.41812134e-01
-6.97376311e-01 -7.26786137e-01 -9.38622430e-02 2.68063754e-01
5.01320839e-01 -3.43314558e-01 -5.49991786e-01 1.98215032e+00
-2.20003426e-01 -1.05462372e-01 1.38232797e-01 7.95868099e-01
4.81544316e-01 3.37904334e-01 4.70182866e-01 8.64803866e-02
1.18447149e+00 -9.64985430e-01 -1.87284321e-01 -1.07340097e+00
9.93832886e-01 -1.10182203e-02 1.16849506e+00 5.27192295e-01
-1.51593423e+00 -5.78742325e-01 -1.13907766e+00 5.70185222e-02
-2.62481004e-01 -4.30059880e-01 1.24705815e+00 5.78754663e-01
-1.17882860e+00 9.63807702e-01 -8.16545367e-01 -3.87938082e-01
4.79150325e-01 2.55712658e-01 9.62758437e-02 -4.45190072e-02
-1.41430438e+00 1.29132080e+00 9.81924534e-01 -8.51398408e-02
-1.47278023e+00 -2.57241398e-01 -8.87950122e-01 4.09234762e-01
7.18944490e-01 -7.67659962e-01 1.82007790e+00 -8.51802170e-01
-1.60570645e+00 7.32803464e-01 2.91130006e-01 -5.68511963e-01
2.20288541e-02 1.13398775e-01 2.62854304e-02 1.58839151e-01
9.27637517e-02 7.85171270e-01 2.05783799e-01 -9.83599722e-01
-7.69572794e-01 -3.16071868e-01 6.94589436e-01 5.64449668e-01
2.55304784e-01 6.27216371e-03 -5.50982468e-02 -5.59741199e-01
1.44014120e-01 -6.69251025e-01 -3.94730598e-01 -4.58746314e-01
3.09738759e-02 -2.87833065e-01 -3.69180501e-01 -5.17305017e-01
1.15659416e+00 -1.73781812e+00 8.58308095e-03 2.31963694e-01
3.42551053e-01 -2.74948254e-02 -3.09598148e-01 6.37376010e-01
-9.48579311e-02 2.64572710e-01 3.93112823e-02 2.19434857e-01
2.71711111e-01 2.55541652e-01 -1.95732698e-01 5.53109050e-02
-5.62570691e-02 1.34859514e+00 -1.16935790e+00 -1.51346639e-01
1.62961017e-02 -4.07815039e-01 -9.34343755e-01 -2.11795159e-02
-6.35447681e-01 1.32912144e-01 -3.15054178e-01 3.67424756e-01
-1.19873643e-01 -1.01097919e-01 3.84525001e-01 3.21694523e-01
1.79988090e-02 8.93896043e-01 -8.23984921e-01 1.62676489e+00
-1.57774553e-01 5.56107163e-01 -2.79199302e-01 -9.69459772e-01
5.26239574e-01 1.16552949e-01 -9.08388868e-02 -1.19700921e+00
3.12747806e-01 -2.16933582e-02 7.47099757e-01 -4.32721168e-01
4.38595355e-01 -3.87027383e-01 -2.23102346e-01 6.89638793e-01
3.02361101e-01 -2.78323829e-01 4.54950809e-01 1.71442524e-01
1.23589671e+00 4.76982147e-01 5.90504587e-01 -2.63952821e-01
-2.35293314e-01 5.36383629e-01 5.11364043e-01 1.37236905e+00
-1.98831618e-01 -7.88540486e-03 8.79324794e-01 -5.62430263e-01
-9.59793329e-01 -9.33787823e-01 4.84623671e-01 1.48425782e+00
-3.30355883e-01 -7.76627421e-01 -7.83968985e-01 -4.59760189e-01
-3.90601218e-01 1.13877928e+00 -9.35194731e-01 -6.56444788e-01
-5.44527173e-01 -4.97476310e-01 7.65966833e-01 7.91383862e-01
5.63541114e-01 -1.64380264e+00 -1.16521502e+00 5.29172182e-01
-1.38639718e-01 -6.62545443e-01 1.28875375e-01 5.32948256e-01
-1.19018924e+00 -1.14871657e+00 -7.12521523e-02 -6.13155425e-01
2.12754160e-01 -1.63598329e-01 1.44072771e+00 3.13380778e-01
1.44802243e-01 4.08836007e-01 -2.29488924e-01 -3.04255068e-01
-3.95056099e-01 5.16157672e-02 -1.17622018e-01 -1.01009572e+00
4.15614963e-01 -6.60236895e-01 -4.53209430e-01 1.12725347e-02
-6.64177895e-01 1.67466894e-01 6.73690140e-01 8.96188557e-01
9.31321830e-02 2.39964113e-01 1.96934298e-01 -1.01690495e+00
1.23151803e+00 -5.02649844e-01 -6.21858358e-01 3.18430401e-02
-5.70776820e-01 3.10998291e-01 3.65752786e-01 -2.88781613e-01
-8.40199471e-01 -5.33750474e-01 -1.86613545e-01 1.73543796e-01
-2.79575706e-01 1.14846313e+00 2.21506611e-01 4.04754840e-03
1.26814258e+00 3.01722795e-01 -2.45118678e-01 -8.74293596e-02
3.26012760e-01 -2.70387501e-01 2.92338789e-01 -1.06493318e+00
5.77852905e-01 -5.62751032e-02 -2.94092089e-01 -3.77839923e-01
-9.41302717e-01 4.74745547e-03 -1.17395945e-01 -2.82822818e-01
7.19074905e-01 -7.69163966e-01 -1.24374115e+00 1.56246021e-01
-8.78825247e-01 -1.02752268e+00 -6.79735184e-01 3.98905575e-01
-1.19976950e+00 -6.12654537e-02 -9.51628923e-01 -7.52925336e-01
2.49318004e-01 -1.07570684e+00 5.30400455e-01 2.58762062e-01
-6.93962038e-01 -9.20437455e-01 1.76133484e-01 2.65037030e-01
3.40783268e-01 -2.87514210e-01 1.43271947e+00 -7.99465477e-01
-4.59669203e-01 3.40534121e-01 1.55099586e-01 -2.13298857e-01
-4.75428402e-01 -4.57842499e-01 -7.22788572e-01 -1.06546000e-01
1.61434367e-01 -8.67190361e-01 8.51015091e-01 5.50649941e-01
1.01625717e+00 -4.82269228e-01 -7.75162503e-02 3.92872661e-01
1.36109447e+00 5.25015950e-01 8.50581944e-01 8.12824011e-01
1.36907659e-02 7.65267313e-01 6.21067643e-01 2.89996833e-01
4.83693510e-01 4.44881618e-01 4.99059290e-01 4.87271041e-01
2.69979596e-01 -8.89671326e-01 4.53958750e-01 4.85994041e-01
-1.53462470e-01 -5.58038279e-02 -1.16772175e+00 5.16421199e-01
-1.80353630e+00 -1.25207233e+00 5.59315264e-01 1.96249545e+00
9.63926196e-01 8.06052923e-01 3.19099694e-01 -3.11063714e-02
1.78628772e-01 2.98881799e-01 -6.89993143e-01 -7.20924437e-01
1.58637494e-01 5.13543546e-01 2.71231443e-01 6.24061406e-01
-5.71493864e-01 1.69482780e+00 7.68308306e+00 8.37766886e-01
-6.05373859e-01 -6.12905659e-02 6.81108296e-01 -9.88554806e-02
-2.42176190e-01 3.14289451e-01 -5.95551670e-01 -2.08022054e-02
1.16703093e+00 -2.29652375e-02 8.85650754e-01 8.69622767e-01
1.28012598e-01 -5.36581635e-01 -1.24277556e+00 6.11600220e-01
-1.65027782e-01 -1.57801199e+00 2.77689975e-02 1.91094026e-01
4.89178121e-01 1.28267398e-02 2.84649152e-03 1.10954940e+00
1.19349372e+00 -1.38537061e+00 9.80545044e-01 4.97858703e-01
2.30457917e-01 -7.82826245e-01 4.44008559e-01 8.22908938e-01
-6.52061939e-01 -5.21580279e-01 -5.29909134e-01 -7.45695889e-01
-1.83425352e-01 -1.83282912e-01 -8.97581458e-01 -5.84338009e-02
4.77211386e-01 2.92505413e-01 -7.63566732e-01 1.06303215e+00
-6.97434783e-01 8.37486565e-01 -7.42579177e-02 -3.58305782e-01
7.18811631e-01 -1.89684685e-02 2.65149146e-01 5.59669673e-01
-3.27327922e-02 4.56174314e-01 1.39573678e-01 1.09132445e+00
3.01228493e-01 -1.33599803e-01 -7.07664490e-01 -3.76622885e-01
5.22468328e-01 6.34941757e-01 -8.45073044e-01 -2.25664794e-01
8.80908147e-02 4.70324129e-01 6.63179398e-01 3.48076254e-01
-5.72937787e-01 3.54061536e-02 5.38796425e-01 1.27064228e-01
2.19108775e-01 -4.45859641e-01 -3.57603222e-01 -8.90450358e-01
-4.64103848e-01 -1.40746772e+00 2.74623841e-01 -1.20072150e+00
-7.93047905e-01 3.37714314e-01 1.94332004e-01 -5.77987432e-01
-9.58699465e-01 -6.58739686e-01 -6.41234994e-01 6.67065561e-01
-9.61009562e-01 -6.18620992e-01 1.17227815e-01 5.36324680e-01
5.38846970e-01 -2.11047277e-01 1.02895939e+00 -5.54385543e-01
-1.75100714e-01 3.27056766e-01 -1.84391156e-01 3.37145150e-01
-9.47261043e-03 -1.01203787e+00 8.68892431e-01 5.85613012e-01
3.95272046e-01 7.99020708e-01 7.35559940e-01 -6.27879381e-01
-9.09628987e-01 -2.68034160e-01 4.42520142e-01 -3.54833990e-01
6.14330530e-01 -2.01668113e-01 -4.83742326e-01 9.40996289e-01
2.29278848e-01 -5.07761180e-01 5.93505144e-01 5.75978398e-01
-2.18210205e-01 3.04940313e-01 -9.38928902e-01 9.93968248e-01
1.25992823e+00 -6.05369747e-01 -1.13767314e+00 1.49554655e-01
4.80616212e-01 -3.69370788e-01 -4.16306913e-01 -9.97515097e-02
5.31013489e-01 -1.17239559e+00 7.02046096e-01 -1.24431992e+00
7.02640295e-01 -1.60240419e-02 -9.80667472e-02 -1.88117635e+00
-6.62564933e-01 -4.31451887e-01 -7.34106917e-03 5.64883947e-01
5.19169807e-01 -5.02589405e-01 1.15753734e+00 6.08868062e-01
-1.56208500e-01 -7.16202736e-01 -7.82252610e-01 -5.70469677e-01
4.50401902e-01 -7.90110230e-01 4.71793711e-01 8.00359130e-01
4.49818522e-01 4.04458106e-01 1.40891084e-02 -2.83748478e-01
4.97434527e-01 -2.74662435e-01 4.62655395e-01 -1.35160923e+00
-6.81617439e-01 -6.39220417e-01 -1.51913971e-01 -1.14536881e+00
8.51878300e-02 -8.21973085e-01 1.09778121e-01 -1.69703460e+00
5.76606765e-02 -3.98740172e-01 -1.57868806e-02 6.65265739e-01
2.68279910e-01 -2.11632177e-01 5.14046371e-01 -6.15624972e-02
-6.13472760e-01 4.06782925e-01 1.35050607e+00 4.49815430e-02
-8.84808004e-02 -2.11528808e-01 -1.37125361e+00 1.05581975e+00
1.12506270e+00 -6.24257207e-01 -6.32243395e-01 -5.93072593e-01
1.04603136e+00 3.49997640e-01 2.54363149e-01 -1.37745547e+00
5.13442934e-01 -5.23949325e-01 6.16545200e-01 -1.54849589e-01
3.82151246e-01 -4.36562121e-01 -2.32662544e-01 7.35560060e-01
-7.06577718e-01 3.15015644e-01 6.32413208e-01 2.78106511e-01
-1.33422166e-02 -6.46953702e-01 6.07518256e-01 -9.54511523e-01
-9.05893385e-01 1.05297705e-02 -1.12822080e+00 4.92191762e-01
6.50847018e-01 -5.45655727e-01 -4.25049096e-01 -7.86385775e-01
-1.03340590e+00 3.73108447e-01 1.97408602e-01 2.71369100e-01
6.20254993e-01 -1.14756453e+00 -4.73239839e-01 -2.57987864e-02
1.65602099e-02 -2.34116182e-01 1.46960273e-01 4.54684645e-01
-5.93961418e-01 5.68338752e-01 -5.95856786e-01 3.76678817e-02
-5.52536190e-01 4.21867430e-01 7.41313756e-01 -7.06569135e-01
-2.21890330e-01 9.37399924e-01 2.47090876e-01 -5.80745041e-01
4.68161553e-02 -5.47428071e-01 -3.62116307e-01 2.79290210e-02
3.41141701e-01 -5.13737388e-02 -2.36139178e-01 -1.43458590e-01
-2.06964929e-02 -2.13513933e-02 -1.90072030e-01 -6.39898419e-01
1.45819759e+00 2.49521196e-01 1.02626294e-01 4.89382386e-01
2.90383756e-01 -3.03467214e-01 -1.33128178e+00 -9.17027071e-02
2.95950562e-01 -2.05131888e-01 2.20806748e-02 -1.38430440e+00
-7.93757081e-01 9.07332301e-01 2.40688384e-01 1.13863014e-01
7.49425232e-01 -5.82432561e-02 2.18308926e-01 7.68886387e-01
7.84905851e-01 -1.39476955e+00 3.20146173e-01 1.08384097e+00
8.40272009e-01 -7.11181343e-01 -1.28504202e-01 2.73925394e-01
-8.72133374e-01 1.03399360e+00 9.19825256e-01 -2.60663688e-01
1.92796186e-01 -9.97753963e-02 -3.48570853e-01 -4.58115965e-01
-1.23313832e+00 -3.50637317e-01 -2.41510436e-01 8.24352562e-01
2.61551648e-01 -6.44898713e-02 -1.86908111e-01 9.67281044e-01
-8.50179076e-01 -1.06153283e-02 5.63395023e-01 1.02518070e+00
-7.48765469e-01 -9.00688171e-01 -2.16661796e-01 6.15975499e-01
-2.32788801e-01 -5.49467325e-01 -4.47155178e-01 1.02038789e+00
8.52616206e-02 8.55139613e-01 -3.68186757e-02 -3.82458717e-01
2.90537536e-01 3.38238627e-01 8.98812294e-01 -9.29738522e-01
-1.10223877e+00 -1.55322552e-01 4.20362234e-01 -1.03079498e+00
-8.59013051e-02 -7.44414330e-01 -1.08632302e+00 -5.03117561e-01
3.70199122e-02 4.03780907e-01 2.15959951e-01 1.15108621e+00
-3.78992781e-03 6.02352858e-01 -9.38690901e-02 -6.92276657e-01
-7.42897630e-01 -1.06265962e+00 -6.52656674e-01 -1.21718042e-01
-9.57758129e-02 -8.02586854e-01 -1.27403334e-01 -5.17260909e-01] | [3.830531597137451, 1.3798218965530396] |
ba41fd5d-520c-45c9-86b3-82e73d95da20 | sparcs-recovering-low-rank-and-sparse | null | null | http://papers.nips.cc/paper/4438-sparcs-recovering-low-rank-and-sparse-matrices-from-compressive-measurements | http://papers.nips.cc/paper/4438-sparcs-recovering-low-rank-and-sparse-matrices-from-compressive-measurements.pdf | SpaRCS: Recovering low-rank and sparse matrices from compressive measurements | We consider the problem of recovering a matrix $\mathbf{M}$ that is the sum of a low-rank matrix $\mathbf{L}$ and a sparse matrix $\mathbf{S}$ from a small set of linear measurements of the form $\mathbf{y} = \mathcal{A}(\mathbf{M}) = \mathcal{A}({\bf L}+{\bf S})$. This model subsumes three important classes of signal recovery problems: compressive sensing, affine rank minimization, and robust principal component analysis. We propose a natural optimization problem for signal recovery under this model and develop a new greedy algorithm called SpaRCS to solve it. SpaRCS inherits a number of desirable properties from the state-of-the-art CoSaMP and ADMiRA algorithms, including exponential convergence and efficient implementation. Simulation results with video compressive sensing, hyperspectral imaging, and robust matrix completion data sets demonstrate both the accuracy and efficacy of the algorithm. | ['Andrew E. Waters', 'Richard Baraniuk', 'Aswin C. Sankaranarayanan'] | 2011-12-01 | null | null | null | neurips-2011-12 | ['video-compressive-sensing'] | ['computer-vision'] | [ 8.31546605e-01 -1.00961983e-01 2.75817573e-01 -1.97738111e-01
-1.04122531e+00 -6.91240668e-01 4.52670977e-02 -2.66642034e-01
-1.42092109e-01 5.99764407e-01 1.77048385e-01 -3.98667604e-01
-9.19834077e-01 -5.11394560e-01 -9.37952578e-01 -9.52998996e-01
-4.51026708e-01 1.26036748e-01 -5.12759566e-01 -3.38233322e-01
1.30829558e-01 3.59218061e-01 -1.16900802e+00 -1.69537976e-01
9.76637185e-01 1.28085935e+00 3.45525116e-01 4.04186875e-01
4.70290840e-01 7.97559381e-01 -1.86510161e-01 -1.10054523e-01
8.28769267e-01 -5.33003509e-01 -5.13306618e-01 3.64672899e-01
4.96459395e-01 -4.10438061e-01 -8.64342451e-01 1.57630146e+00
2.93905258e-01 1.98925868e-01 4.69633847e-01 -9.20161545e-01
-5.99840045e-01 6.65844142e-01 -1.02555346e+00 1.65644899e-01
4.85214114e-01 -2.66501158e-01 8.88857603e-01 -1.17955148e+00
4.44509715e-01 8.90279353e-01 7.57406294e-01 -2.69734979e-01
-1.20870233e+00 -8.33346844e-01 -6.51913285e-02 1.41868159e-01
-1.89421237e+00 -7.37088144e-01 5.88392735e-01 -5.66486895e-01
4.90363091e-01 7.23083496e-01 4.56795692e-01 2.35929340e-01
-2.68177569e-01 6.76759243e-01 1.29666412e+00 -5.97253323e-01
3.11025381e-01 -4.44820166e-01 1.86565518e-01 8.80838394e-01
7.05964804e-01 1.81138124e-02 -5.22118747e-01 -3.59202176e-01
8.99852574e-01 7.76669458e-02 -9.61519599e-01 -1.68693691e-01
-1.40912819e+00 8.05854380e-01 1.02075469e-02 1.87610433e-01
-5.89614928e-01 1.03280723e-01 -3.56610209e-01 3.65952611e-01
7.12864101e-02 4.60150778e-01 -3.64183486e-02 4.86260772e-01
-1.41389167e+00 1.03191048e-01 7.57280111e-01 1.19545603e+00
9.54693556e-01 6.86089694e-01 1.52423322e-01 8.76117527e-01
3.59310776e-01 1.37508273e+00 -3.79697606e-02 -1.43403971e+00
5.42930067e-01 -6.53025657e-02 2.60197878e-01 -1.20704937e+00
-2.27938592e-01 -6.78507864e-01 -1.29540920e+00 -1.31794900e-01
2.64220119e-01 -1.52174279e-01 -5.46756387e-01 1.64490008e+00
-7.04802871e-02 4.43520337e-01 2.92366855e-02 8.39505315e-01
4.72080022e-01 9.71952856e-01 -7.66138494e-01 -8.84560585e-01
8.63671184e-01 -4.74161237e-01 -6.72139943e-01 -5.13495088e-01
5.81347421e-02 -1.00406361e+00 4.26930398e-01 7.29938865e-01
-1.39588034e+00 -1.60082743e-01 -1.32559109e+00 3.44994307e-01
4.49282259e-01 3.59160244e-01 5.34861863e-01 6.05938077e-01
-1.07645345e+00 4.79193002e-01 -5.10481358e-01 1.86561942e-01
9.68761370e-02 3.96341920e-01 -5.23627341e-01 -7.40100145e-01
-6.68960690e-01 2.40469784e-01 -2.99414415e-02 4.37963486e-01
-1.03308272e+00 -6.31945848e-01 -8.69279683e-01 -9.07964110e-02
4.75755095e-01 -4.10860807e-01 6.31449401e-01 -6.38182104e-01
-1.16173053e+00 7.46155202e-01 -1.78175956e-01 -2.33656541e-01
8.04265812e-02 -2.65506029e-01 -5.75709283e-01 6.26133800e-01
3.22642982e-01 -1.46928757e-01 1.40397012e+00 -1.11050355e+00
-1.44532546e-01 -6.92588568e-01 -2.31152028e-01 -3.26432809e-02
-1.54794261e-01 -9.59804878e-02 -2.80401170e-01 -1.02130318e+00
1.25698805e+00 -1.17371500e+00 -3.86712492e-01 -2.64164925e-01
-6.26679480e-01 7.35263467e-01 4.70022768e-01 -1.09481514e+00
1.25716937e+00 -2.28398061e+00 5.58753192e-01 8.58546376e-01
2.99868733e-01 1.15168564e-01 -2.68990815e-01 6.74351394e-01
-5.37332416e-01 -3.11890453e-01 -7.93381095e-01 1.48566186e-01
-2.62662798e-01 -5.28536327e-02 -4.51702178e-01 1.16459417e+00
-5.02211094e-01 3.31867933e-01 -7.59529710e-01 1.31271482e-01
4.11580987e-02 2.64140218e-01 -5.36871970e-01 -1.22707691e-02
8.89067054e-02 5.27013898e-01 -4.76972610e-01 8.60638738e-01
1.10747015e+00 -5.44213116e-01 3.38464707e-01 -4.87989575e-01
-2.46796384e-01 -2.73236930e-01 -2.04459834e+00 1.81278527e+00
2.48583168e-01 2.75735438e-01 1.12558270e+00 -1.47756541e+00
7.12852776e-01 1.32913515e-01 1.13973773e+00 -5.09286582e-01
-1.05551137e-02 5.21906137e-01 -2.70069391e-01 -4.31807220e-01
3.48801374e-01 -1.88248158e-01 3.08312595e-01 6.19284987e-01
-2.32988685e-01 -4.66131538e-01 3.13397676e-01 5.91807723e-01
1.36719739e+00 -2.85020262e-01 2.79265285e-01 -6.29090071e-01
6.20115697e-01 -1.65734202e-01 4.25347775e-01 9.06804979e-01
1.70298174e-01 5.16908407e-01 -1.02242149e-01 1.75201148e-01
-8.85154545e-01 -9.37447667e-01 -2.52606332e-01 5.77008963e-01
2.13785872e-01 -5.54397225e-01 -5.90927780e-01 2.58215457e-01
-1.45542413e-01 3.80583733e-01 -8.73704329e-02 8.92954245e-02
-5.23726940e-01 -1.03399134e+00 2.07857922e-01 -2.75737792e-02
5.07299781e-01 -2.92317212e-01 -1.01879664e-01 9.36695263e-02
-5.51562786e-01 -1.21468353e+00 -5.25733054e-01 -2.28186715e-02
-8.79731476e-01 -1.17309868e+00 -6.95872664e-01 -5.21073282e-01
8.52227747e-01 1.03532934e+00 7.83532619e-01 3.68326791e-02
-2.83078581e-01 7.33094156e-01 -3.77819121e-01 -1.58864573e-01
4.07581031e-02 -8.61244977e-01 3.80355358e-01 4.41798061e-01
-2.98691005e-01 -9.05894518e-01 -6.93484783e-01 2.73137599e-01
-1.26042068e+00 -2.01915175e-01 6.34477079e-01 7.78877437e-01
1.11382246e+00 3.20083320e-01 1.54372126e-01 -6.27499640e-01
3.22281927e-01 -3.40520352e-01 -6.80175900e-01 1.04955107e-01
-6.02702260e-01 -7.09115937e-02 4.68413442e-01 -7.22264349e-02
-5.11609852e-01 3.32341880e-01 6.58276528e-02 -6.16363764e-01
5.26508033e-01 8.25778663e-01 8.73801578e-03 -4.03533429e-01
7.66449869e-01 7.36236691e-01 1.15888223e-01 -5.55302918e-01
4.45504785e-01 4.70544219e-01 8.41919661e-01 -6.05102777e-01
1.18831539e+00 8.89777422e-01 3.18140656e-01 -1.43231559e+00
-9.78490889e-01 -6.63857937e-01 -3.30690861e-01 2.57344488e-02
2.58707702e-01 -1.34433711e+00 -6.37433827e-01 2.43032336e-01
-5.81085145e-01 -7.86553845e-02 -4.21921730e-01 9.02599335e-01
-8.18787515e-01 9.39824402e-01 -5.00931144e-01 -7.96518326e-01
-3.92230630e-01 -1.00641334e+00 9.29013968e-01 -2.38186896e-01
2.33233988e-01 -4.27483171e-01 -7.62756243e-02 6.06575489e-01
3.41311544e-01 1.37886360e-01 6.98431671e-01 -3.52386502e-03
-8.27841103e-01 -4.42486286e-01 -2.94549018e-01 5.91398478e-01
-2.12219968e-01 -6.11598849e-01 -5.96143484e-01 -8.88030887e-01
4.56388682e-01 -1.12737700e-01 8.07443202e-01 7.83057630e-01
1.10017300e+00 -6.77527249e-01 -1.27637088e-01 1.11139977e+00
1.65935373e+00 5.53147644e-02 7.31351078e-01 -1.33523613e-01
4.63100702e-01 7.52454400e-02 2.87588447e-01 9.07840371e-01
9.75805242e-03 3.50317508e-01 5.93045413e-01 2.50632167e-01
1.66821092e-01 2.50451148e-01 4.32480067e-01 1.06560409e+00
-2.42650077e-01 8.82790387e-02 -5.91888964e-01 4.01296020e-01
-1.70140588e+00 -1.33330894e+00 -4.42922622e-01 2.43813872e+00
4.29292589e-01 -6.10161066e-01 -3.12281609e-01 5.97198606e-01
6.54434681e-01 4.24649924e-01 -4.11003768e-01 5.20948350e-01
-5.59630454e-01 7.32587337e-01 7.45998085e-01 6.95656002e-01
-8.77402484e-01 4.55973774e-01 6.08367062e+00 8.10158491e-01
-7.47840464e-01 -3.84517945e-02 7.19075873e-02 -1.55540377e-01
-4.77324218e-01 1.38709649e-01 -4.63413507e-01 1.57410607e-01
5.09117901e-01 -2.12560482e-02 1.18005562e+00 4.95869547e-01
1.67522460e-01 -1.15820624e-01 -8.38850200e-01 1.56950879e+00
4.73209232e-01 -1.41064358e+00 -1.58795610e-01 1.07264146e-01
9.52442467e-01 1.98378097e-02 1.60643086e-01 -1.86962470e-01
4.75371599e-01 -9.90046501e-01 8.56794000e-01 4.04606462e-01
1.15797770e+00 -1.42815381e-01 2.09048707e-02 4.03196782e-01
-1.24441707e+00 -3.53351325e-01 -4.27480012e-01 -1.12096161e-01
3.64964694e-01 1.22789109e+00 -1.19026430e-01 9.03134763e-01
6.82678998e-01 9.67364728e-01 6.16932176e-02 9.16102231e-01
1.16710514e-02 5.41280806e-01 -6.20739460e-01 6.10421300e-01
9.93243605e-02 -1.06530905e+00 9.80482936e-01 8.89490426e-01
1.04448628e+00 1.02122617e+00 4.77827162e-01 4.90036815e-01
4.26336080e-02 1.27751663e-01 -5.85003078e-01 -2.00717345e-01
5.98940492e-01 1.05095398e+00 -4.02564615e-01 -1.49255395e-01
-2.89205581e-01 7.72132635e-01 -2.16644436e-01 6.20136976e-01
-4.69492227e-01 -1.43063486e-01 5.77971697e-01 2.54338950e-01
4.34600472e-01 -6.34959161e-01 -2.43923306e-01 -1.37398636e+00
6.93190619e-02 -1.32966554e+00 4.14850980e-01 -8.33665729e-01
-1.04979682e+00 3.52228761e-01 -1.22262791e-01 -1.50303757e+00
2.15847135e-01 -5.95508218e-01 6.82972595e-02 8.14182937e-01
-1.36161923e+00 -6.51327133e-01 -3.91953379e-01 1.18329072e+00
6.84204623e-02 -3.88584793e-01 8.07305217e-01 4.17936921e-01
-4.97025669e-01 8.37877020e-02 6.43958271e-01 -1.74675301e-01
2.33283848e-01 -6.07563019e-01 -4.11761016e-01 1.43671286e+00
1.78106353e-02 7.04649746e-01 7.85539865e-01 -5.48535645e-01
-2.33055067e+00 -9.33824718e-01 2.96856076e-01 1.49041474e-01
7.23419607e-01 1.62971660e-01 -4.46686655e-01 1.11219537e+00
5.41711189e-02 1.85669772e-02 8.76399994e-01 -4.74466980e-01
-5.10052264e-01 -2.62876272e-01 -1.11974192e+00 2.55641550e-01
1.18339252e+00 -5.24769068e-01 -1.25106454e-01 8.47195029e-01
2.15036243e-01 -4.08894867e-01 -9.25502837e-01 5.52310050e-01
3.46330285e-01 -1.03230262e+00 1.55341160e+00 -1.09464929e-01
2.48280704e-01 -4.09644544e-01 -1.09712958e+00 -1.03355110e+00
-5.84716380e-01 -1.20881867e+00 -2.04680234e-01 4.05011714e-01
1.77585274e-01 -3.91190320e-01 5.43910623e-01 3.00884366e-01
-3.55881602e-01 3.52890752e-02 -9.96307790e-01 -7.36711621e-01
-2.51171559e-01 -4.69150335e-01 2.94843584e-01 1.25892246e+00
-8.63384176e-03 1.01639010e-01 -7.33449876e-01 6.83472276e-01
1.21411192e+00 3.41601044e-01 6.08473599e-01 -1.13994598e+00
-7.98141181e-01 -5.59074245e-02 -5.73514253e-02 -1.35498238e+00
-1.39462397e-01 -1.20149541e+00 -3.00181005e-02 -1.45229435e+00
2.24046350e-01 -7.15312839e-01 -8.39601457e-02 2.72114187e-01
2.63061702e-01 3.43104362e-01 2.31683671e-01 4.56114978e-01
-3.20637733e-01 4.03412402e-01 1.17130446e+00 -4.30452108e-01
3.82348225e-02 -1.61637545e-01 -9.86693263e-01 6.84728086e-01
7.60972798e-02 -3.44070792e-01 -4.67411608e-01 -8.19707215e-01
6.46117866e-01 9.03111994e-01 1.90958425e-01 -1.03338289e+00
3.38731408e-01 -2.79109836e-01 1.18871264e-01 -3.76831114e-01
8.46396208e-01 -8.72515082e-01 6.18064106e-01 2.86524624e-01
6.91432059e-02 1.58817116e-02 -1.68882176e-01 7.01755106e-01
-1.46170288e-01 -2.14070722e-01 8.87277782e-01 -2.51534581e-01
-3.69399041e-01 4.99409109e-01 -8.88553634e-02 5.41713275e-02
8.24690700e-01 -1.57599390e-01 -3.92021060e-01 -7.95009971e-01
-7.76854455e-01 -1.31030336e-01 1.21240392e-01 -3.98924977e-01
8.23263824e-01 -1.08819437e+00 -1.18145263e+00 5.24698257e-01
-1.60426617e-01 2.14147102e-02 4.21800226e-01 1.20495033e+00
-5.84978342e-01 2.96781361e-01 1.01542927e-01 -5.95243156e-01
-7.81287730e-01 3.33269805e-01 1.29338220e-01 -2.36239228e-02
-5.22118032e-01 1.14617801e+00 -3.41418423e-02 -1.74981713e-01
-1.07772807e-02 1.76467195e-01 3.16237837e-01 -1.61529511e-01
8.30450416e-01 8.13962162e-01 5.12617491e-02 -1.03673983e+00
-1.97477564e-01 7.91909456e-01 4.11073297e-01 -3.29003513e-01
1.61916566e+00 -3.14644337e-01 -7.81858027e-01 -2.00451493e-01
1.07683504e+00 5.08035064e-01 -1.04122841e+00 -4.57003772e-01
-3.98158550e-01 -8.16296935e-01 1.52091652e-01 -4.73875940e-01
-1.22761631e+00 4.43421572e-01 4.05743867e-01 1.75458208e-01
1.43664873e+00 -3.14873815e-01 5.93903184e-01 6.05980396e-01
6.36195481e-01 -9.85298276e-01 1.97505757e-01 4.99458373e-01
1.16926217e+00 -8.56967032e-01 4.95322466e-01 -7.48762369e-01
-7.33476877e-02 7.65389681e-01 -3.27573925e-01 -3.54723960e-01
9.62442279e-01 1.58861682e-01 -2.30309129e-01 -3.14673245e-01
1.38237923e-01 -8.67746696e-02 1.02542616e-01 2.39532799e-01
1.72908545e-01 2.35641330e-01 -3.20056558e-01 4.42873061e-01
-3.32419723e-01 -9.39430296e-02 8.09282660e-01 1.08837283e+00
-6.31711841e-01 -8.75102818e-01 -9.16947424e-01 6.30665720e-01
-3.00771445e-01 -2.68693775e-01 9.88377407e-02 2.29709059e-01
-1.32286996e-01 1.35379589e+00 -5.77879965e-01 -2.13980198e-01
1.89586326e-01 -1.73939571e-01 6.93728209e-01 -4.69571739e-01
1.76280707e-01 5.40800869e-01 -2.33682021e-01 -5.98569870e-01
-5.33346295e-01 -9.48913693e-01 -1.16976225e+00 -2.62215465e-01
-2.58703083e-01 2.81497598e-01 6.18002534e-01 8.80746782e-01
1.01732008e-01 8.00721124e-02 8.68360698e-01 -6.95733964e-01
-9.82451677e-01 -7.54363120e-01 -1.37185419e+00 3.96271288e-01
4.06696379e-01 -1.99493185e-01 -4.86734241e-01 2.23554850e-01] | [7.235952854156494, 4.4547648429870605] |
1cf724cb-830f-48e1-9208-b183f51bf02e | gender-and-representation-bias-in-gpt-3 | null | null | https://aclanthology.org/2021.nuse-1.5 | https://aclanthology.org/2021.nuse-1.5.pdf | Gender and Representation Bias in GPT-3 Generated Stories | Using topic modeling and lexicon-based word similarity, we find that stories generated by GPT-3 exhibit many known gender stereotypes. Generated stories depict different topics and descriptions depending on GPT-3’s perceived gender of the character in a prompt, with feminine characters more likely to be associated with family and appearance, and described as less powerful than masculine characters, even when associated with high power verbs in a prompt. Our study raises questions on how one can avoid unintended social biases when using large language models for storytelling. | ['David Bamman', 'Li Lucy'] | null | null | null | null | naacl-nuse-2021-6 | ['word-similarity'] | ['natural-language-processing'] | [-1.90710396e-01 5.02486467e-01 -5.27506888e-01 -3.46106648e-01
-1.59353986e-01 -6.01288378e-01 9.61577535e-01 8.87766421e-01
-3.17033619e-01 6.33615792e-01 1.31889164e+00 -1.05030097e-01
-1.15128551e-02 -1.02044141e+00 -2.17301860e-01 -1.78265899e-01
3.87232691e-01 9.49018061e-01 -1.59918964e-01 -6.00414872e-01
5.07291555e-01 -2.79087514e-01 -1.14444864e+00 3.61561269e-01
5.51642179e-01 2.25922525e-01 -5.38648404e-02 3.19104165e-01
-3.10726404e-01 1.01057172e+00 -1.08969700e+00 -1.01946402e+00
-3.61704409e-01 -5.48049808e-01 -5.45023799e-01 7.46564344e-02
6.90006137e-01 -2.96572834e-01 -4.80828673e-01 7.91104734e-01
5.18626213e-01 1.60605639e-01 1.10412621e+00 -1.12423241e+00
-6.43231034e-01 1.50051785e+00 -7.25663543e-01 2.93512464e-01
1.03032756e+00 -2.89260615e-02 9.03723657e-01 -5.83334684e-01
1.30033290e+00 2.03274345e+00 6.71554089e-01 4.70579177e-01
-1.55830562e+00 -1.07530546e+00 1.10216565e-01 -1.39694482e-01
-1.46480167e+00 -4.02770132e-01 7.38187253e-01 -6.79655969e-01
8.21027577e-01 4.29409504e-01 1.17311418e+00 1.79565012e+00
3.48826706e-01 1.01356253e-01 8.42237234e-01 -2.90522748e-03
2.05410391e-01 4.35451537e-01 7.37855434e-02 1.48850605e-01
4.79171962e-01 -5.11625826e-01 -9.83390450e-01 -9.10935283e-01
6.01293325e-01 -4.74969923e-01 3.24528098e-01 1.73835218e-01
-1.34544206e+00 1.46734893e+00 -7.55772293e-02 5.57120323e-01
-2.68999249e-01 3.65393579e-01 7.27064490e-01 -3.27100158e-02
9.22548532e-01 8.89702737e-01 1.92628264e-01 -7.30224669e-01
-7.29916990e-01 9.47263181e-01 8.93623173e-01 1.07954693e+00
1.17854536e-01 -1.06926747e-02 -3.46853673e-01 1.15945852e+00
1.46490589e-01 2.12291107e-01 3.93941432e-01 -8.93085063e-01
4.78204697e-01 4.14731801e-01 -6.45625964e-02 -1.82521129e+00
-3.94719481e-01 -9.47082322e-03 -2.41523817e-01 -4.36249852e-01
3.85837495e-01 -2.64400661e-01 -1.79292932e-01 1.95412290e+00
1.07688069e-01 -2.41792753e-01 -1.00739159e-01 6.19957507e-01
1.14513958e+00 5.97280502e-01 8.92764747e-01 -3.55786473e-01
1.78832841e+00 -1.18944664e-02 -7.80460298e-01 -6.54202878e-01
8.25060666e-01 -1.07625484e+00 1.06788099e+00 -3.21761817e-02
-1.12605572e+00 -1.16630182e-01 -8.72864306e-01 -1.34462401e-01
-8.52866769e-02 -4.78620350e-01 7.45546162e-01 1.00653577e+00
-7.24490404e-01 4.10955966e-01 -9.25553814e-02 -1.34894514e+00
4.89215285e-01 -2.13357940e-01 -2.96680361e-01 2.68128216e-01
-1.07012057e+00 1.00271916e+00 5.55577159e-01 -1.06480300e+00
-4.29483891e-01 -9.63555694e-01 -1.00596714e+00 -1.82059571e-01
2.97937840e-01 -6.69679046e-01 1.24305081e+00 -6.56042993e-01
-9.53681588e-01 1.32064486e+00 -9.65626761e-02 -2.48529911e-01
4.41861421e-01 -9.77187380e-02 -4.54714477e-01 -1.11325020e-02
6.13874614e-01 8.36189210e-01 7.68072128e-01 -1.10546458e+00
-2.93179154e-01 -2.75700897e-01 3.43389153e-01 4.48644578e-01
-6.36466384e-01 9.30483997e-01 4.38315362e-01 -1.19309151e+00
4.99422625e-02 -4.21845257e-01 -4.42497768e-02 -2.56472349e-01
-7.12247908e-01 -5.31261742e-01 6.88771486e-01 -5.44390440e-01
1.35580742e+00 -2.04059267e+00 -2.32137248e-01 -2.33720429e-02
5.98427117e-01 -4.89955246e-01 3.58425915e-01 9.98069406e-01
-5.89615107e-02 5.01212120e-01 5.51230669e-01 9.17928070e-02
1.82416290e-01 -4.15700488e-02 -3.72985840e-01 4.82713282e-01
-1.37783512e-01 4.79758382e-01 -1.15633070e+00 -9.93312597e-01
-1.78506896e-01 2.26693898e-01 -6.19792163e-01 -3.04438978e-01
-1.25887662e-01 4.52934466e-02 -3.90147507e-01 3.86109620e-01
3.63421887e-01 4.23169509e-02 3.61178994e-01 1.57811388e-01
-2.16910630e-01 6.29548311e-01 -7.74600267e-01 9.78969634e-01
-6.84949979e-02 1.02791560e+00 -1.90903351e-01 -3.40318501e-01
1.05428803e+00 2.12982908e-01 -2.78567839e-02 -3.78453135e-01
5.40685833e-01 9.29769129e-02 1.69401169e-01 -6.35477483e-01
1.10643637e+00 -5.94996512e-01 -5.67848146e-01 8.94889891e-01
-1.53534174e-01 -8.11049044e-01 3.27778906e-01 6.77091599e-01
7.07145035e-01 -5.00525534e-01 4.21791673e-01 -7.33329177e-01
-4.87779677e-01 -2.68382742e-03 6.12510324e-01 7.81951904e-01
1.50024548e-01 3.59033495e-01 1.16501749e+00 -3.31267595e-01
-1.39936602e+00 -1.06943321e+00 -3.12221348e-01 1.38656926e+00
1.09142594e-01 -8.96031499e-01 -6.25809848e-01 -1.58303484e-01
1.03435874e-01 1.63584781e+00 -6.28569961e-01 -3.25047016e-01
-3.42850566e-01 -4.45881337e-01 5.99699616e-01 3.18676293e-01
-9.04098079e-02 -5.00129104e-01 -7.33331323e-01 3.36133480e-01
-2.91543752e-01 -9.29653227e-01 -4.89416212e-01 -2.31325477e-01
-4.53604043e-01 -6.79326534e-01 -6.12619102e-01 -5.42649031e-01
5.69158554e-01 -8.70762542e-02 1.22658980e+00 2.34645195e-02
-1.67426556e-01 2.40625903e-01 -4.50021565e-01 -7.35820889e-01
-8.04630935e-01 -1.46965459e-02 1.99556351e-01 -3.90348643e-01
4.90046710e-01 -5.78981340e-01 -6.47606477e-02 1.05236076e-01
-5.82904100e-01 3.21502864e-01 -3.68507713e-01 3.89464557e-01
-4.55536157e-01 -2.37686008e-01 2.62129724e-01 -1.19558144e+00
1.05947733e+00 -9.46552575e-01 4.91548777e-01 -2.26665601e-01
-1.18796170e-01 -5.83701730e-01 1.56729966e-01 -9.78136063e-01
-9.85590994e-01 -9.36382473e-01 9.79450271e-02 1.50627136e-01
5.76330870e-02 1.97107092e-01 1.41115412e-01 4.74495500e-01
9.42486584e-01 -2.39700302e-01 -2.55952090e-01 -5.11055626e-02
2.79913787e-02 4.74041194e-01 1.19550057e-01 -8.05272460e-01
6.75347090e-01 4.65181947e-01 -4.92585272e-01 -1.10905445e+00
-5.20123720e-01 1.86817069e-02 2.10340008e-01 -8.31782341e-01
1.04370546e+00 -8.64223719e-01 -9.08535242e-01 3.55048388e-01
-1.21513355e+00 -1.76218569e-01 -2.83180892e-01 5.94156444e-01
-6.17855906e-01 -1.77196071e-01 -6.35169506e-01 -6.69114470e-01
-1.13988211e-02 -6.09995425e-01 8.17709327e-01 3.01026821e-01
-1.68057084e+00 -1.01436138e+00 3.19293253e-02 3.39127630e-01
2.64500916e-01 3.65096241e-01 1.61264634e+00 -1.18768716e+00
4.83829707e-01 -6.42714230e-03 -2.03465447e-01 -6.23784602e-01
1.81865692e-01 2.70806015e-01 -4.17449027e-01 8.64656195e-02
-3.31568047e-02 -5.75292408e-01 1.31804943e-01 2.36080885e-01
3.60030323e-01 -7.62958765e-01 -6.82158053e-01 9.96109918e-02
1.11264122e+00 1.99175045e-01 4.01842505e-01 3.52782428e-01
4.91568714e-01 9.28474903e-01 6.62212074e-01 9.05980468e-01
5.27568460e-01 6.73958421e-01 -1.30627558e-01 4.07922179e-01
2.71106094e-01 -8.78506541e-01 2.80235142e-01 2.28251815e-01
1.16200067e-01 -3.65252018e-01 -1.11712337e+00 6.86064303e-01
-1.50043857e+00 -1.21708560e+00 -3.72709960e-01 1.55261493e+00
7.54430830e-01 3.27885419e-01 2.90193141e-01 -1.59730658e-01
9.77380574e-01 7.30816483e-01 -2.12932616e-01 -8.65939796e-01
-4.64088768e-01 -1.40104845e-01 1.86302081e-01 2.33581677e-01
-6.72088504e-01 9.74929929e-01 7.86140203e+00 1.24753535e+00
-7.09119797e-01 3.04510862e-01 1.04799485e+00 -4.97372419e-01
-9.12725925e-01 4.57053855e-02 -6.62549317e-01 5.49853921e-01
6.02009296e-01 -1.03514373e+00 -1.06702432e-01 7.21002042e-01
3.53807509e-01 -4.68519270e-01 -1.22310674e+00 8.62455666e-01
3.90545934e-01 -1.28572452e+00 1.79776505e-01 -1.38943717e-01
6.48985088e-01 -7.39005029e-01 6.78082407e-02 1.67006940e-01
4.55451816e-01 -9.79042649e-01 1.49709189e+00 9.92429256e-02
6.16391599e-01 -8.15454662e-01 3.29522878e-01 1.11365981e-01
-5.64981222e-01 -2.48807557e-02 -4.36702967e-01 -7.86624730e-01
4.98461068e-01 2.76476800e-01 -9.17860746e-01 -2.34729722e-01
6.59531474e-01 4.30661678e-01 -4.13506597e-01 4.40280259e-01
7.63887027e-03 6.94613636e-01 -2.76970208e-01 -6.54670000e-01
-7.77861923e-02 -2.89669484e-01 1.23375976e+00 1.16214633e+00
4.20260668e-01 2.22832531e-01 -1.27287898e-02 9.25790846e-01
1.22278824e-01 5.80089450e-01 -9.50121462e-01 -4.11448687e-01
7.99800694e-01 9.32935476e-01 -1.38643360e+00 -3.90341192e-01
-8.14349428e-02 4.53702122e-01 6.19271845e-02 6.18585348e-02
-7.63035059e-01 -2.53196452e-02 9.98574018e-01 9.65796828e-01
-5.10403156e-01 1.61346477e-02 -7.60117233e-01 -5.77042162e-01
-5.75449288e-01 -8.50078404e-01 1.69826850e-01 -1.07731438e+00
-1.23525119e+00 2.68535286e-01 6.89079344e-01 -6.83803320e-01
-3.07178319e-01 6.88757673e-02 -9.62721705e-01 2.67951488e-01
2.71172106e-01 -1.17656732e+00 5.56513108e-02 1.34341300e-01
4.76917803e-01 -8.56964961e-02 5.70982516e-01 4.93665040e-02
-5.78916185e-02 6.61210775e-01 -4.47678328e-01 -7.67051876e-02
8.96775663e-01 -7.17932761e-01 5.11340618e-01 1.03694342e-01
-2.02422321e-01 5.83959460e-01 1.72618771e+00 -1.14771318e+00
-7.65636861e-01 -5.20250916e-01 1.44624698e+00 -4.35763061e-01
1.08888149e+00 -7.44318068e-01 -3.57686132e-01 7.69023657e-01
5.64548194e-01 -1.10198355e+00 9.87043560e-01 3.28214884e-01
-4.41941828e-01 6.98249876e-01 -1.38422585e+00 1.18907022e+00
1.36612749e+00 -3.88208270e-01 -7.64072239e-01 7.00593114e-01
9.70526636e-01 -4.17877167e-01 -7.11617053e-01 -2.27342889e-01
4.64730561e-01 -8.23424399e-01 7.31907547e-01 -7.20013380e-01
8.23295414e-01 6.27503633e-01 -1.04556821e-01 -1.31848371e+00
-3.73599589e-01 -6.12256408e-01 7.21926987e-01 1.90457964e+00
3.51482570e-01 -5.68753362e-01 6.43184423e-01 8.72916460e-01
-5.04271120e-05 -3.59044939e-01 -1.00954306e+00 -5.02610505e-01
4.26480532e-01 -6.29714310e-01 6.12198412e-01 1.37615705e+00
6.81174219e-01 3.08217764e-01 -3.26923311e-01 -7.41177022e-01
2.49020606e-01 -5.57535231e-01 4.63166863e-01 -1.37730014e+00
-8.91918987e-02 -6.79134786e-01 -5.62551618e-01 -2.98960179e-01
1.57032624e-01 -4.89340872e-01 -3.94427925e-01 -1.30799472e+00
9.28568184e-01 -1.97252706e-01 1.00462031e+00 1.43369392e-01
7.62910768e-02 2.84515977e-01 4.34983283e-01 -3.03768069e-01
-2.15820417e-01 3.53140920e-01 1.10219944e+00 -2.06751466e-01
-1.69397444e-01 -4.02697146e-01 -1.32202125e+00 1.22607529e+00
5.47142684e-01 -9.22121823e-01 -3.39401841e-01 -1.79624006e-01
1.02913523e+00 9.09816846e-02 3.79743665e-01 -4.66910452e-01
-4.06394303e-02 -5.55723488e-01 2.51008630e-01 -2.61411905e-01
5.23683548e-01 -3.00862104e-01 5.15258670e-01 3.65500450e-01
-6.35837495e-01 1.79388642e-01 3.12971562e-01 4.31143790e-02
9.03039128e-02 -3.10182631e-01 4.93561804e-01 4.79085818e-02
5.77729493e-02 -3.13606858e-01 -1.37369454e+00 4.01429415e-01
1.37824917e+00 -6.28049314e-01 -5.02488613e-01 -1.12192404e+00
-4.54274416e-01 -9.65393856e-02 9.17936027e-01 7.56604493e-01
3.99612576e-01 -1.58294070e+00 -1.08534443e+00 -3.22757870e-01
3.67591262e-01 -6.80052519e-01 2.78554738e-01 5.51583469e-01
-5.33561170e-01 -6.19335175e-02 -3.01531285e-01 -1.53976888e-03
-1.28364015e+00 2.22368658e-01 -2.78553277e-01 2.05784708e-01
-4.30173635e-01 1.10634112e+00 4.03513908e-01 9.42189898e-03
-2.47438982e-01 1.11658618e-01 -4.08766210e-01 8.39978755e-01
4.73598987e-01 7.21453547e-01 -7.10087657e-01 -1.19881511e+00
-2.61754155e-01 2.81956136e-01 -1.86477035e-01 -3.78281176e-01
1.16950250e+00 5.73280640e-02 -2.69719392e-01 6.95237815e-01
8.76681328e-01 3.09792340e-01 -4.42268163e-01 1.93786129e-01
-1.20616622e-01 -8.89461100e-01 -3.09669405e-01 -3.31802130e-01
-4.07949537e-01 1.26907513e-01 -1.46511376e-01 4.94301319e-01
3.53923559e-01 6.26849830e-01 6.00550652e-01 -1.75683171e-01
5.15173316e-01 -1.33163917e+00 2.07487270e-01 3.94413501e-01
1.26067603e+00 -6.79837108e-01 1.66718274e-01 -6.81379318e-01
-1.00542462e+00 7.93839991e-01 8.53066564e-01 2.23091394e-02
4.26679701e-01 2.77584016e-01 -9.70282406e-02 -4.31050956e-01
-7.48570263e-01 4.48269248e-01 -2.97956556e-01 7.73748696e-01
8.05110276e-01 2.73793161e-01 -1.24106991e+00 9.23639476e-01
-1.10910916e+00 -7.01728344e-01 8.90716732e-01 6.53294027e-01
-1.54912829e-01 -7.34013379e-01 -7.68914938e-01 7.69950688e-01
-7.81285763e-01 1.71367079e-02 -8.52220356e-01 8.22340786e-01
2.08744481e-01 1.08042967e+00 7.39247501e-01 -4.40934032e-01
-6.53044926e-03 -9.07986835e-02 3.82416993e-01 -9.04428661e-01
-8.97664189e-01 6.01847395e-02 9.02085423e-01 -2.61205971e-01
-1.03739724e-01 -1.16267431e+00 -1.09582245e+00 -1.02000499e+00
-7.74978027e-02 -2.69770660e-02 2.83119738e-01 8.65726888e-01
4.92803454e-02 -5.11123659e-03 1.08825855e-01 -6.07194483e-01
6.38070926e-02 -1.31398582e+00 -8.79679859e-01 4.67057496e-01
-5.23307681e-01 -8.09540451e-01 -1.35513857e-01 -2.59300023e-01] | [9.203031539916992, 10.215691566467285] |
17c50e4a-86bf-4d8b-b204-dd94d06196d6 | towards-personalized-preprocessing-pipeline | 2302.14329 | null | https://arxiv.org/abs/2302.14329v1 | https://arxiv.org/pdf/2302.14329v1.pdf | Towards Personalized Preprocessing Pipeline Search | Feature preprocessing, which transforms raw input features into numerical representations, is a crucial step in automated machine learning (AutoML) systems. However, the existing systems often have a very small search space for feature preprocessing with the same preprocessing pipeline applied to all the numerical features. This may result in sub-optimal performance since different datasets often have various feature characteristics, and features within a dataset may also have their own preprocessing preferences. To bridge this gap, we explore personalized preprocessing pipeline search, where the search algorithm is allowed to adopt a different preprocessing pipeline for each feature. This is a challenging task because the search space grows exponentially with more features. To tackle this challenge, we propose ClusterP3S, a novel framework for Personalized Preprocessing Pipeline Search via Clustering. The key idea is to learn feature clusters such that the search space can be significantly reduced by using the same preprocessing pipeline for the features within a cluster. To this end, we propose a hierarchical search strategy to jointly learn the clusters and search for the optimal pipelines, where the upper-level search optimizes the feature clustering to enable better pipelines built upon the clusters, and the lower-level search optimizes the pipeline given a specific cluster assignment. We instantiate this idea with a deep clustering network that is trained with reinforcement learning at the upper level, and random search at the lower level. Experiments on benchmark classification datasets demonstrate the effectiveness of enabling feature-wise preprocessing pipeline search. | ['Xia Hu', 'Qiaoyu Tan', 'Daochen Zha', 'Diego Martinez'] | 2023-02-28 | null | null | null | null | ['automl', 'deep-clustering', 'deep-clustering'] | ['methodology', 'miscellaneous', 'natural-language-processing'] | [-6.25303611e-02 -2.83439755e-01 4.76774946e-02 -5.36237478e-01
-7.89893866e-01 -8.32076311e-01 1.66675851e-01 3.67890388e-01
-2.85960019e-01 -1.53726682e-01 -9.54358280e-02 -1.73551127e-01
-3.94355714e-01 -9.39848602e-01 -5.07743180e-01 -6.60903931e-01
-8.62721913e-03 5.33282876e-01 2.30395347e-01 1.10882320e-01
4.29579258e-01 6.61077619e-01 -1.59984934e+00 5.14155090e-01
9.07726765e-01 1.14625227e+00 2.59955227e-01 4.72277015e-01
-3.66412759e-01 2.25606412e-01 -5.11084199e-01 2.76495535e-02
5.18274367e-01 -2.39356488e-01 -9.39476311e-01 1.40987381e-01
1.07095808e-01 1.57186881e-01 1.04774199e-01 9.89459634e-01
1.96558744e-01 2.15079591e-01 4.06964183e-01 -1.57906830e+00
-2.29670510e-01 8.41482103e-01 -5.26908576e-01 -1.59023345e-01
1.01235032e-01 3.99891287e-01 1.19182587e+00 -7.27878451e-01
2.94673651e-01 1.28270948e+00 6.16870522e-01 4.45909619e-01
-1.37573135e+00 -5.30236244e-01 2.51334965e-01 3.67749304e-01
-1.61468387e+00 -1.09230429e-01 7.03310847e-01 -4.79709625e-01
8.14896345e-01 2.79121906e-01 5.82825899e-01 2.25504622e-01
-9.22324322e-03 7.22537160e-01 5.28002620e-01 -2.87259728e-01
5.35130501e-01 -8.24270248e-02 4.52500999e-01 5.39804459e-01
6.40256181e-02 -4.41393912e-01 -3.13361079e-01 -9.40250754e-02
3.55583489e-01 3.11772853e-01 -3.43449414e-02 -5.59807837e-01
-1.01762903e+00 9.20030236e-01 6.55765235e-01 2.54915357e-01
-3.25449675e-01 4.86245565e-02 4.28842068e-01 4.62451786e-01
-1.77641705e-01 9.32541907e-01 -9.09106612e-01 1.05641171e-01
-1.16748464e+00 1.82488576e-01 8.65957439e-01 7.30168998e-01
1.35407472e+00 -5.71111679e-01 -3.82708818e-01 5.64702570e-01
2.95644730e-01 -1.77994087e-01 6.93440795e-01 -8.12982678e-01
4.17099029e-01 1.33160150e+00 -4.71002236e-02 -8.78456116e-01
-7.01667190e-01 -2.28336021e-01 -8.36443603e-01 2.51584768e-01
4.25799757e-01 -2.56463736e-01 -9.38927114e-01 1.51189208e+00
5.59036732e-01 1.23241104e-01 -5.16431518e-02 6.87113881e-01
4.88327086e-01 5.24819016e-01 -4.48350273e-02 3.75051014e-02
1.45051265e+00 -1.13849449e+00 -2.59613842e-01 -8.92687887e-02
8.82037222e-01 -5.82470894e-01 1.39695656e+00 3.24215770e-01
-7.64174700e-01 -6.22521162e-01 -1.01451170e+00 -5.48445471e-02
-5.87318778e-01 2.67943442e-01 6.52377367e-01 5.29563963e-01
-9.94641960e-01 7.50158787e-01 -1.13953698e+00 -2.56845146e-01
3.89740676e-01 7.97514260e-01 -1.72189146e-01 -8.80809277e-02
-8.54977489e-01 3.96994621e-01 6.25575483e-01 1.86973840e-01
-3.66762400e-01 -9.14174139e-01 -8.06945503e-01 4.54515845e-01
4.49787140e-01 -6.88547254e-01 1.02255023e+00 -9.38151181e-01
-1.40305853e+00 4.06087458e-01 -1.03024103e-01 -2.62009174e-01
2.18315974e-01 -1.87602147e-01 1.51701458e-02 5.04337717e-03
-4.75250743e-02 5.30219674e-01 9.23793972e-01 -8.73436749e-01
-8.33543420e-01 -3.38529468e-01 -1.31867126e-01 1.38064668e-01
-4.68734384e-01 1.14870384e-01 -8.63659978e-01 -2.45288789e-01
1.00918666e-01 -8.44556749e-01 -6.81014717e-01 -2.32161090e-01
-3.46907049e-01 -4.63737577e-01 8.34506750e-01 -2.02592731e-01
1.48052454e+00 -2.34123230e+00 2.75734037e-01 6.35425270e-01
1.80107281e-01 -4.12266143e-03 -2.00028837e-01 1.85015336e-01
-7.49087334e-02 1.49950907e-01 -1.57345369e-01 -3.85977656e-01
1.24583758e-01 1.74150556e-01 -4.26663980e-02 2.72432566e-01
5.99046052e-01 7.97144473e-01 -8.41856301e-01 -4.89893377e-01
8.02316815e-02 1.67936400e-01 -1.07764161e+00 3.75528783e-01
-2.49635577e-01 4.00610149e-01 -5.66003442e-01 5.56585908e-01
6.66560054e-01 -5.25221288e-01 2.03620702e-01 -5.08144617e-01
-3.58304918e-01 -1.20508038e-02 -1.74027574e+00 1.61999977e+00
-3.46051037e-01 1.84475645e-01 1.89085454e-01 -1.16766644e+00
8.67322862e-01 -1.21886261e-01 6.12392187e-01 -2.44742706e-01
6.08687177e-02 1.20545059e-01 1.80919468e-01 -4.25820410e-01
3.53942454e-01 2.78172940e-01 -3.69013906e-01 4.56412554e-01
4.19794302e-03 -2.57583976e-01 3.59197497e-01 1.73231196e-02
1.43884468e+00 -6.79765195e-02 1.69496581e-01 -1.12649381e-01
6.32100523e-01 9.38170329e-02 7.07993507e-01 7.24797666e-01
-1.72227509e-02 4.86525595e-01 6.06186688e-01 -5.72353363e-01
-5.97755909e-01 -6.57148004e-01 1.65980056e-01 1.35962963e+00
6.76539317e-02 -6.71371937e-01 -9.20924485e-01 -7.74360776e-01
3.09504457e-02 2.91415751e-01 -6.73644960e-01 -3.05025518e-01
-6.14270568e-01 -7.68473327e-01 2.25440070e-01 4.50349391e-01
2.91021526e-01 -1.11482930e+00 -1.03482890e+00 1.66110232e-01
1.50791988e-01 -7.94614255e-01 -5.43926418e-01 7.27116048e-01
-6.98754013e-01 -1.23607886e+00 -4.94175241e-04 -9.00460184e-01
9.71502542e-01 6.48861676e-02 1.00870812e+00 4.46135670e-01
-4.42295343e-01 2.55451381e-01 -5.71575403e-01 2.93185469e-02
-1.90311730e-01 5.12286901e-01 -7.12115914e-02 1.76682293e-01
4.73144412e-01 -3.33854437e-01 -7.14959264e-01 1.78754807e-01
-1.05154347e+00 -1.29147366e-01 6.30619586e-01 9.82788801e-01
7.86045074e-01 5.46550453e-01 2.48339981e-01 -7.27974713e-01
4.36582863e-01 -5.14471769e-01 -8.73358548e-01 4.40190077e-01
-4.49108213e-01 4.26040918e-01 1.16409802e+00 -3.18098068e-01
-5.08204043e-01 8.51582289e-01 2.65074316e-02 -6.01256371e-01
-2.80754656e-01 6.88266754e-01 -6.87491357e-01 6.46095648e-02
5.26506364e-01 4.05359566e-02 -8.72247890e-02 -6.69168353e-01
4.87695187e-01 5.09806335e-01 4.40342665e-01 -6.69912875e-01
9.66659486e-01 2.36763552e-01 -5.81494858e-03 -3.71454179e-01
-6.01457596e-01 -6.32685244e-01 -1.08418763e+00 2.28894204e-01
7.66895950e-01 -5.19246995e-01 -8.97925198e-01 3.37683052e-01
-8.57594907e-01 -6.48317337e-01 -3.01864326e-01 1.12246513e-01
-4.03410524e-01 2.46992156e-01 -2.95952171e-01 -2.90837884e-01
-4.25366759e-01 -1.72254634e+00 9.97481167e-01 4.85344112e-01
-4.38155234e-01 -7.03033924e-01 -2.20840693e-01 -1.91050902e-01
2.72154838e-01 2.23804429e-01 1.12208366e+00 -1.02802646e+00
-4.92976636e-01 -2.19837129e-01 -2.19951779e-01 4.88158241e-02
3.79008293e-01 3.40081483e-01 -6.82448745e-01 -3.13800395e-01
-1.97008967e-01 -1.80934191e-01 8.32439005e-01 2.42050380e-01
1.75809956e+00 -2.48615324e-01 -5.61321974e-01 9.77900624e-01
1.31942260e+00 1.07793557e-02 3.02697003e-01 5.54930091e-01
8.10818970e-01 6.46415114e-01 8.00406933e-01 5.78302801e-01
2.74927825e-01 7.08361804e-01 3.01745951e-01 -5.91465496e-02
1.54418603e-01 7.78241605e-02 2.44617447e-01 5.32165468e-01
5.88460088e-01 2.79154062e-01 -1.12283695e+00 3.52045357e-01
-1.94807637e+00 -6.39462352e-01 1.09415479e-01 2.15734220e+00
1.06669950e+00 -6.20187633e-02 2.10142508e-01 3.79678130e-01
6.01869822e-01 -3.16126227e-01 -7.28227019e-01 -2.97002465e-01
3.33733946e-01 3.66125405e-01 1.45698115e-01 2.61707276e-01
-1.21844840e+00 1.05719829e+00 5.56703091e+00 7.93718576e-01
-1.41224146e+00 -3.89741987e-01 5.91131866e-01 -1.19959801e-01
4.91425721e-03 1.49329096e-01 -9.81813252e-01 5.74798822e-01
8.15968990e-01 -1.31791055e-01 6.32525265e-01 9.28358972e-01
1.83284491e-01 -1.67387463e-02 -1.50521314e+00 9.28393960e-01
-3.66403311e-01 -1.21236372e+00 3.45530100e-02 -2.56817013e-01
3.74613404e-01 -4.88423072e-02 -3.55201140e-02 4.17329967e-01
4.12949026e-01 -1.05211139e+00 4.71984744e-01 3.45116168e-01
5.04786849e-01 -9.70518768e-01 3.05095553e-01 3.35794955e-01
-1.54454076e+00 -3.23475927e-01 -3.65623981e-01 1.44151881e-01
-1.41197711e-01 5.33535540e-01 -8.64429235e-01 4.76504326e-01
9.80721831e-01 5.61981201e-01 -8.97195637e-01 1.34014702e+00
-5.55178262e-02 3.55275124e-01 -4.91339535e-01 -1.18503754e-03
1.35011792e-01 -2.31641769e-01 3.13162133e-02 1.39615011e+00
3.09883028e-01 -1.28286839e-01 5.33589065e-01 8.33368838e-01
-6.59280270e-02 2.35334650e-01 -7.72735700e-02 1.19605049e-01
7.88517594e-01 1.75059032e+00 -1.22760856e+00 -2.41045222e-01
-3.73525798e-01 9.45052207e-01 5.72203338e-01 1.64706379e-01
-6.99683130e-01 -8.21748435e-01 7.72987664e-01 -2.19615996e-01
5.36194980e-01 -1.01382293e-01 -4.95968968e-01 -8.97930384e-01
-1.55149877e-01 -8.29809964e-01 7.12760389e-01 -1.24110281e-01
-1.21445501e+00 5.05633950e-01 -1.65686816e-01 -1.20190871e+00
-1.06098853e-01 -6.32009566e-01 -7.52334416e-01 8.50182831e-01
-1.35962069e+00 -1.01052141e+00 -4.11459565e-01 7.54099250e-01
5.28640151e-01 -3.83235030e-02 6.74989641e-01 1.72095582e-01
-9.31004345e-01 9.33176935e-01 1.38927922e-02 9.80651155e-02
7.60719240e-01 -1.42734993e+00 2.47709647e-01 8.01168501e-01
-9.09459591e-02 9.02072787e-01 4.49600965e-01 -4.65292037e-01
-1.77942872e+00 -1.40021682e+00 5.48946738e-01 -3.46338958e-01
8.33607852e-01 -3.72127086e-01 -1.00380969e+00 3.45113605e-01
-8.46854001e-02 2.41277814e-01 1.04790127e+00 2.05102235e-01
-2.41956279e-01 -2.06652045e-01 -1.12689805e+00 5.57309449e-01
6.44068301e-01 -2.32377425e-01 -3.65493625e-01 1.99757189e-01
7.95002937e-01 -2.99668849e-01 -1.01919401e+00 2.90397137e-01
3.63542408e-01 -5.81334472e-01 7.61626780e-01 -5.53945839e-01
2.47766763e-01 -7.74130940e-01 -7.16240928e-02 -1.41252089e+00
-4.76524055e-01 -6.13086641e-01 -1.49243206e-01 1.32391274e+00
4.21010405e-01 -2.99673796e-01 7.62386143e-01 1.00846279e+00
-1.96622357e-01 -9.38602805e-01 -6.02554619e-01 -5.92301309e-01
4.61971872e-02 -4.10571814e-01 1.09803009e+00 9.32829440e-01
4.48478721e-02 2.82896012e-01 1.89878613e-01 4.49415952e-01
2.99744934e-01 5.65633714e-01 1.01276541e+00 -1.40683579e+00
-5.00132263e-01 -8.33016157e-01 -1.93405598e-01 -8.28157425e-01
1.82738855e-01 -1.01024711e+00 3.28819722e-01 -1.40650725e+00
1.88000888e-01 -6.92349434e-01 -3.14398468e-01 1.00593233e+00
-4.99868304e-01 -4.59450632e-02 2.49313191e-01 3.01293552e-01
-8.13586712e-01 2.76643932e-01 8.16318095e-01 -1.49865046e-01
-7.64082253e-01 4.83596213e-02 -8.87392938e-01 6.12902939e-01
7.46882081e-01 -4.09407496e-01 -1.80450290e-01 -3.51144463e-01
2.86705643e-01 -4.19414610e-01 2.36495379e-02 -1.04381227e+00
7.18025684e-01 -2.68103212e-01 5.03463566e-01 -6.64458573e-01
-1.08025715e-01 -1.01094353e+00 -2.54229107e-03 5.38604617e-01
-3.46462667e-01 1.04468212e-01 2.16837674e-01 1.92050204e-01
-2.35541001e-01 -2.28068784e-01 8.04025114e-01 -8.07400607e-03
-7.45441198e-01 3.24322611e-01 -1.68502003e-01 -2.12626293e-01
1.02073014e+00 -2.78268039e-01 2.17246003e-02 3.37558538e-01
-6.76520169e-01 7.70959914e-01 6.27829969e-01 3.09992522e-01
4.06726271e-01 -1.08006585e+00 -3.68700415e-01 4.88157272e-01
1.14592627e-01 6.02667868e-01 2.29635891e-02 7.37036347e-01
-2.80720264e-01 1.31671773e-02 -1.48168206e-01 -5.59390068e-01
-1.25338864e+00 7.97191858e-01 2.35035732e-01 -5.28255463e-01
-2.93624520e-01 8.87915850e-01 7.72929266e-02 -5.61310828e-01
1.58557907e-01 -5.97115993e-01 -4.01910543e-01 3.26291114e-01
6.53453112e-01 3.25378418e-01 2.77780771e-01 -4.24117059e-01
-5.24455428e-01 7.17823207e-01 -3.92535418e-01 3.42404991e-01
1.47775853e+00 -1.87731273e-02 -3.92471105e-01 2.35824138e-01
1.25958300e+00 -1.00783743e-01 -1.40945458e+00 -1.40275702e-01
4.11200583e-01 -3.10944766e-01 -5.57401292e-02 -5.06101966e-01
-1.26888001e+00 6.49866760e-01 4.11122680e-01 3.09760690e-01
1.52140927e+00 3.47329713e-02 4.98279095e-01 5.06989479e-01
1.20949194e-01 -1.20154941e+00 4.47211005e-02 7.80734181e-01
5.82912326e-01 -1.01816058e+00 -1.45756662e-01 -3.36500645e-01
-5.09088755e-01 1.32116699e+00 8.23805809e-01 -2.68805355e-01
7.13734686e-01 3.70079488e-01 -1.01985380e-01 -1.80779368e-01
-6.76718652e-01 -1.77357376e-01 3.14089894e-01 3.36136997e-01
3.58384520e-01 8.33576825e-03 -2.16001784e-03 1.20515764e+00
-3.42143267e-01 -1.48067623e-01 1.57051697e-01 8.07491481e-01
-3.10904533e-01 -1.44761419e+00 -5.13384759e-01 6.49858832e-01
-2.57394403e-01 -1.93115734e-02 -3.26858699e-01 4.18256968e-01
2.87873209e-01 8.84111702e-01 2.00858608e-01 -5.71902394e-01
4.71819878e-01 1.27426520e-01 1.66208968e-01 -7.97176480e-01
-1.09111357e+00 1.39314532e-01 -6.66964889e-01 -8.11092496e-01
-7.57485330e-02 -7.26162136e-01 -1.70029747e+00 5.08428849e-02
-1.72115430e-01 3.40246499e-01 3.90297294e-01 8.94578993e-01
7.41164625e-01 6.52489007e-01 9.00740683e-01 -1.04895341e+00
-4.84470546e-01 -5.93187392e-01 -3.93665344e-01 3.76166254e-01
2.35715672e-01 -5.80418348e-01 -2.38860473e-01 1.52388006e-01] | [9.122529029846191, 3.3256702423095703] |
1476b067-c974-41f5-8ebb-60f3f15556d9 | depth-neus-neural-implicit-surfaces-learning | 2303.17088 | null | https://arxiv.org/abs/2303.17088v1 | https://arxiv.org/pdf/2303.17088v1.pdf | Depth-NeuS: Neural Implicit Surfaces Learning for Multi-view Reconstruction Based on Depth Information Optimization | Recently, methods for neural surface representation and rendering, for example NeuS, have shown that learning neural implicit surfaces through volume rendering is becoming increasingly popular and making good progress. However, these methods still face some challenges. Existing methods lack a direct representation of depth information, which makes object reconstruction unrestricted by geometric features, resulting in poor reconstruction of objects with texture and color features. This is because existing methods only use surface normals to represent implicit surfaces without using depth information. Therefore, these methods cannot model the detailed surface features of objects well. To address this problem, we propose a neural implicit surface learning method called Depth-NeuS based on depth information optimization for multi-view reconstruction. In this paper, we introduce depth loss to explicitly constrain SDF regression and introduce geometric consistency loss to optimize for low-texture areas. Specific experiments show that Depth-NeuS outperforms existing technologies in multiple scenarios and achieves high-quality surface reconstruction in multiple scenarios. | ['Conglin Wang', 'Yichao Gao', 'Yinhe Han', 'Shuai Liang', 'Runnan Chen', 'Cheng Zeng', 'Hanqi Jiang'] | 2023-03-30 | null | null | null | null | ['object-reconstruction'] | ['computer-vision'] | [ 3.04582983e-01 -2.28587762e-02 2.68582068e-02 -4.87727612e-01
-4.65789884e-01 -1.85628165e-03 3.92434835e-01 -1.58180863e-01
4.16743010e-02 6.30909264e-01 -9.87962559e-02 1.15847252e-01
2.72075355e-01 -1.25884497e+00 -8.48479092e-01 -4.01121587e-01
1.68374583e-01 4.70106661e-01 5.11525035e-01 -1.55491009e-01
3.19921941e-01 7.72088766e-01 -1.69602549e+00 6.70277596e-01
8.49539459e-01 1.15842533e+00 2.09786043e-01 2.63552070e-01
-6.51182711e-01 3.53483588e-01 -3.39883476e-01 9.27007478e-03
3.34554046e-01 -1.09607957e-01 -3.89954835e-01 -2.13987738e-01
8.78806651e-01 -8.50136757e-01 -3.40030193e-02 8.48601162e-01
2.45571241e-01 3.33169363e-02 8.10997844e-01 -9.43720281e-01
-4.17538673e-01 -4.28079844e-01 -8.83584321e-01 -4.19940084e-01
1.33798063e-01 -2.74488270e-01 8.30903053e-01 -1.13201475e+00
6.46316588e-01 1.54963708e+00 7.50720620e-01 8.34508181e-01
-1.38116539e+00 -7.39854336e-01 5.31898201e-01 -1.67132244e-01
-1.08263278e+00 -2.48305231e-01 1.10798490e+00 -4.13546085e-01
8.67944896e-01 5.45081854e-01 8.92881870e-01 6.31710410e-01
1.95808083e-01 9.43128407e-01 1.22150850e+00 -2.69806892e-01
2.21437529e-01 -6.49404153e-02 -1.67515699e-03 8.49732578e-01
3.06087673e-01 1.92779422e-01 -7.73915589e-01 -2.01459184e-01
1.66218448e+00 1.07363924e-01 -3.52179945e-01 -7.91934311e-01
-8.18867862e-01 7.09177673e-01 5.11985958e-01 -2.52816528e-01
-8.14319998e-02 4.87705529e-01 2.37239718e-01 1.35188578e-02
1.10247648e+00 1.54214337e-01 -4.57458228e-01 1.34178221e-01
-7.67321885e-01 4.87093478e-01 6.08100593e-01 7.57616937e-01
1.18727994e+00 2.26652980e-01 3.16862911e-01 9.85427797e-01
6.96532726e-01 5.59976757e-01 -1.90191746e-01 -1.44991279e+00
4.17242706e-01 7.33760953e-01 6.98543862e-02 -1.12731647e+00
-2.22156301e-01 -1.24275833e-01 -8.61399829e-01 1.06244516e+00
6.32711276e-02 3.00537109e-01 -1.09200323e+00 1.38902223e+00
5.67325115e-01 3.39309603e-01 -3.66095394e-01 9.76987183e-01
9.19163287e-01 6.27384663e-01 -2.97140241e-01 1.07519120e-01
1.02461600e+00 -7.82667041e-01 -6.60523891e-01 -1.68866560e-01
3.21437448e-01 -4.25803870e-01 1.24333405e+00 6.80776834e-01
-1.16542625e+00 -2.12424397e-01 -1.23927975e+00 -4.15197551e-01
3.74320559e-02 -2.22541526e-01 9.44169223e-01 4.60470825e-01
-9.89844263e-01 8.21561456e-01 -1.14138198e+00 2.73842126e-01
7.39999115e-01 4.68278825e-01 -1.72261540e-02 -1.56201944e-01
-7.62588739e-01 3.69087994e-01 -3.91029209e-01 -3.80568020e-03
-6.58047497e-01 -1.03440607e+00 -9.07866061e-01 -2.30973184e-01
-7.93944485e-03 -9.77721989e-01 1.03238559e+00 -1.11986613e+00
-1.75448203e+00 7.96750247e-01 -2.56533027e-01 1.69970229e-01
5.42651772e-01 -5.11417747e-01 2.42972240e-01 -4.52038180e-03
-1.23207718e-01 5.75984657e-01 7.94510305e-01 -1.90516818e+00
-2.34425455e-01 -6.65894747e-01 1.24893390e-01 4.37654763e-01
-1.45277977e-01 -4.96333361e-01 -4.20594245e-01 -6.64608717e-01
7.29872465e-01 -3.61857891e-01 -4.35171306e-01 9.27367687e-01
-2.76282787e-01 1.55848578e-01 9.06805038e-01 -6.52748406e-01
7.45907187e-01 -2.06990862e+00 -6.85160002e-03 3.30622792e-01
5.54278612e-01 -4.51243855e-02 -1.46254720e-02 -1.11696661e-01
3.35685194e-01 1.62398055e-01 -3.54856223e-01 -7.73086727e-01
-2.47910589e-01 3.69371444e-01 -3.34005386e-01 3.54585439e-01
2.60317307e-02 5.44915199e-01 -6.01637065e-01 -3.18634719e-01
2.97179013e-01 1.05329871e+00 -1.00522661e+00 2.86829978e-01
-6.00920260e-01 6.50134146e-01 -8.04794252e-01 6.56659722e-01
1.08357406e+00 -3.02542537e-01 2.29550693e-02 -2.34369650e-01
-2.07323343e-01 4.14754242e-01 -1.09369528e+00 1.80886889e+00
-5.71451008e-01 5.91027737e-01 4.84242827e-01 -6.81924462e-01
8.77065480e-01 4.90829125e-02 5.85988820e-01 -9.43675995e-01
-1.90056697e-01 2.21055150e-01 -6.46505177e-01 -8.69405419e-02
3.31360340e-01 -4.38229024e-01 6.26904011e-01 8.68335068e-02
-7.32005954e-01 -5.41561902e-01 -6.93947792e-01 -1.38741076e-01
7.54702270e-01 6.44204497e-01 -8.76169652e-02 -1.05893113e-01
7.26995990e-02 -1.72148705e-01 8.33385825e-01 2.21582115e-01
3.74173999e-01 1.08656156e+00 3.75836581e-01 -6.07702732e-01
-9.51542914e-01 -1.31165409e+00 -3.80507052e-01 4.86572832e-01
5.06515384e-01 -3.32214445e-01 -6.99747860e-01 -2.95866430e-01
2.58613169e-01 3.76269519e-01 -6.74566627e-01 2.14808673e-01
-9.24078882e-01 -5.10231853e-01 -1.47921015e-02 5.91784179e-01
3.85625690e-01 -8.63834918e-01 -7.16264129e-01 1.41709462e-01
9.25962180e-02 -7.37517178e-01 -6.04531839e-02 -2.36580074e-01
-1.62701285e+00 -9.10996854e-01 -9.45677578e-01 -3.78880084e-01
8.93062711e-01 3.15033078e-01 1.34329617e+00 3.44817698e-01
-4.45586503e-01 2.60164201e-01 1.71763957e-01 -2.84921467e-01
-4.88103181e-02 -2.17490509e-01 -3.62842947e-01 -5.98610118e-02
-2.60364085e-01 -9.59874809e-01 -9.17836189e-01 3.81334782e-01
-7.04188526e-01 8.07048738e-01 3.02323729e-01 6.17842674e-01
1.17459595e+00 -4.53702658e-01 2.80554295e-01 -1.12557375e+00
-1.57717522e-02 -2.15834185e-01 -8.35168719e-01 -4.87987474e-02
-5.74542105e-01 3.91250178e-02 2.17234462e-01 -2.69271493e-01
-1.21037269e+00 -1.03720702e-01 -3.35796893e-01 -5.89759350e-01
4.19678427e-02 2.76335388e-01 -2.43855670e-01 -3.26224416e-01
4.06857193e-01 3.22422269e-03 3.52659114e-02 -8.95518243e-01
-1.20447561e-01 6.28796965e-02 -1.25377893e-01 -8.33378315e-01
5.16819417e-01 1.05700910e+00 2.98877567e-01 -1.07484901e+00
-7.03095615e-01 -1.56953216e-01 -3.93574953e-01 -3.03799093e-01
6.52025998e-01 -8.55013072e-01 -6.18405938e-01 5.87761581e-01
-1.36677027e+00 -8.17906797e-01 -3.76389205e-01 3.22389454e-01
-7.09723830e-01 3.48902434e-01 -6.62000120e-01 -1.02863944e+00
-1.57257199e-01 -1.11779487e+00 1.33019090e+00 -1.26836687e-01
-7.61935813e-03 -1.02173889e+00 4.19706739e-02 3.19313705e-01
5.01811028e-01 5.65771341e-01 1.09005153e+00 6.23545647e-01
-1.34006381e+00 3.13656777e-01 -5.56007266e-01 3.46575797e-01
1.55602813e-01 -6.49588853e-02 -1.14338017e+00 -1.78982124e-01
1.17658898e-01 -1.68152213e-01 8.45773697e-01 7.06906438e-01
1.64555109e+00 -1.54687494e-01 -4.35228646e-01 1.32691967e+00
1.69792318e+00 1.37929730e-02 8.91978264e-01 3.35645556e-01
1.19402456e+00 7.53798068e-01 4.03629512e-01 3.81283224e-01
4.56362247e-01 7.27099240e-01 7.98927128e-01 -4.58064944e-01
-2.97041327e-01 -2.26614490e-01 1.21107727e-01 7.81301200e-01
-4.24025893e-01 -1.96015323e-03 -7.37097979e-01 9.16926563e-02
-1.53068841e+00 -3.62873703e-01 -4.42892253e-01 2.22632551e+00
6.85272753e-01 3.39174420e-02 -2.87769258e-01 -8.36444274e-02
1.80890709e-01 2.43204996e-01 -7.74022102e-01 -1.72293290e-01
-7.85858855e-02 3.00424844e-01 3.25701416e-01 7.00327694e-01
-6.61589205e-01 7.40903318e-01 6.15892887e+00 7.96664953e-01
-1.22499883e+00 1.71889082e-01 7.26871610e-01 -1.83929339e-01
-1.08249545e+00 -1.52902886e-01 -7.26238728e-01 2.14908943e-02
1.55884206e-01 2.50939637e-01 2.70119965e-01 7.13525951e-01
-5.08693652e-03 -1.37522325e-01 -1.03498054e+00 1.15275002e+00
-8.20685104e-02 -1.62533748e+00 3.37111145e-01 2.43046224e-01
8.70089591e-01 8.31254572e-03 1.53523713e-01 -5.76684177e-02
4.62096371e-02 -1.19288838e+00 8.71850729e-01 6.62814379e-01
1.23916340e+00 -5.41389287e-01 1.39300928e-01 3.69195193e-01
-1.29803240e+00 6.48356140e-01 -3.42722863e-01 -1.07200563e-01
2.21990034e-01 7.19898939e-01 -2.06072494e-01 3.46360147e-01
7.43596613e-01 8.10669243e-01 -6.76701516e-02 9.07368481e-01
1.09154284e-01 4.16227639e-01 -4.68248874e-01 2.35878944e-01
-1.53974667e-01 -3.17339867e-01 4.54675108e-01 7.96648800e-01
2.35570997e-01 2.19208002e-01 1.75049216e-01 1.20649230e+00
-4.44215648e-02 1.76554024e-01 -6.78839922e-01 4.64767814e-01
6.83493167e-02 7.47435093e-01 -6.63265705e-01 -1.11085005e-01
-4.90689486e-01 9.71168935e-01 5.13535380e-01 5.92157960e-01
-5.12487173e-01 1.55579522e-01 9.76640642e-01 8.22327793e-01
-2.07735866e-01 -5.14671743e-01 -9.04859126e-01 -1.18682921e+00
1.83255509e-01 -4.30135846e-01 -2.22497687e-01 -6.66785896e-01
-1.29700553e+00 3.04515839e-01 -3.75614055e-02 -1.08910954e+00
3.83483559e-01 -7.84824133e-01 -3.07421744e-01 1.11543143e+00
-1.81966424e+00 -1.09403646e+00 -3.62585574e-01 3.59157652e-01
6.12171948e-01 2.05160663e-01 7.15655029e-01 3.55510414e-01
-9.49679911e-02 3.63881499e-01 3.10950931e-02 -2.68722117e-01
3.81189048e-01 -1.11091328e+00 6.14012480e-01 1.27624869e-01
-2.61059880e-01 6.14837587e-01 2.51194656e-01 -9.84414220e-01
-1.38836730e+00 -8.10125947e-01 9.63652804e-02 -1.73755020e-01
-1.70730352e-01 -5.54229558e-01 -1.37098944e+00 4.75960016e-01
-3.71389627e-01 2.05602884e-01 4.58805293e-01 1.36786059e-01
-4.55156833e-01 -4.62198853e-02 -1.20137322e+00 5.95695555e-01
1.30217218e+00 -4.63477045e-01 5.50019704e-02 1.17209136e-01
6.03339016e-01 -8.25781584e-01 -5.33355892e-01 8.21365178e-01
8.20322037e-01 -1.29336989e+00 1.28368235e+00 -1.59783170e-01
5.41659415e-01 -1.63606897e-01 -4.03046489e-01 -9.14809644e-01
-1.36970505e-01 -1.44931510e-01 -3.57684433e-01 8.74553204e-01
8.95122364e-02 -7.38847852e-01 1.31164479e+00 7.09998071e-01
-4.49176818e-01 -1.35823047e+00 -9.27016377e-01 -5.70006013e-01
1.73913226e-01 -4.20664281e-01 5.37138045e-01 9.92366731e-01
-8.11933637e-01 -9.09299478e-02 -1.72707915e-01 8.41887519e-02
1.01172507e+00 1.60096422e-01 6.95607483e-01 -1.55078864e+00
-1.69230759e-01 -4.81001824e-01 -2.19093725e-01 -1.67689550e+00
2.00791478e-01 -7.06195652e-01 8.54376554e-02 -1.95441628e+00
-7.35919923e-02 -1.30466020e+00 4.87487540e-02 5.93570210e-02
7.98123851e-02 4.12502229e-01 -2.02916741e-01 2.34909102e-01
-9.94887352e-02 8.90835702e-01 1.83360732e+00 5.46839135e-03
-7.13019297e-02 -1.48943990e-01 -2.71576107e-01 1.14047933e+00
6.64813519e-01 -2.76558548e-01 -4.86471981e-01 -9.27879214e-01
4.88417119e-01 2.06513345e-01 5.26748598e-01 -8.65456402e-01
-1.95034862e-01 -5.27979076e-01 5.94307899e-01 -8.21128488e-01
9.98472691e-01 -7.79454648e-01 1.09935172e-01 2.57541835e-01
2.23671347e-02 -1.87839508e-01 2.32339263e-01 7.38359392e-01
-5.54441921e-02 1.27637431e-01 8.56460154e-01 -1.70792088e-01
-4.90091830e-01 7.67957807e-01 -4.59584594e-02 6.25881031e-02
5.36279202e-01 -5.37475646e-01 1.00494221e-01 -1.83806926e-01
-4.82796490e-01 -6.91553429e-02 9.48743224e-01 8.37686360e-02
1.17722094e+00 -1.39727497e+00 -5.94522536e-01 5.90052664e-01
-2.54495680e-01 8.17260921e-01 4.16959286e-01 2.65654415e-01
-1.03717446e+00 -6.14063367e-02 -9.98400301e-02 -7.85673201e-01
-1.13083935e+00 -1.35054335e-01 5.80584884e-01 -6.04638197e-02
-1.16498339e+00 1.06358838e+00 1.25604749e+00 -5.10038972e-01
1.46061957e-01 -5.59358954e-01 2.06909664e-02 -4.23715025e-01
3.15999866e-01 4.15154964e-01 9.35558975e-02 -3.66442859e-01
-6.79168552e-02 1.08976758e+00 1.42699853e-01 -2.20865175e-01
1.35985041e+00 9.74079594e-02 -1.86636150e-01 8.02009523e-01
1.13385427e+00 1.25317305e-01 -1.69943917e+00 -7.45996386e-02
-3.54355186e-01 -8.83437872e-01 2.68587738e-01 -5.19328475e-01
-1.33978963e+00 1.39609265e+00 5.26121736e-01 -3.15199792e-01
7.44119883e-01 -2.72249490e-01 1.09903169e+00 1.03237152e-01
7.74954915e-01 -8.83190811e-01 2.85400599e-01 5.67525446e-01
1.24685609e+00 -1.01009119e+00 2.99424976e-01 -1.11065304e+00
-6.88932911e-02 1.09416735e+00 9.61829603e-01 -3.55825365e-01
8.96336973e-01 4.88926470e-01 4.58129905e-02 -5.27228296e-01
-4.39655721e-01 4.02083337e-01 4.44768608e-01 5.75104594e-01
5.67394495e-01 9.09952372e-02 1.62674904e-01 1.81469098e-01
8.33453089e-02 -1.72322392e-01 4.66175899e-02 8.53942633e-01
-3.55484724e-01 -1.17156613e+00 -2.91600555e-01 7.23547220e-01
-3.28777373e-01 3.74929048e-02 -5.87931722e-02 6.23573124e-01
-1.09995790e-02 3.67754810e-02 3.17184359e-01 -8.34639147e-02
5.03722012e-01 -4.36142862e-01 9.18969750e-01 -9.68836248e-01
-1.91852957e-01 9.03514847e-02 3.27173881e-02 -8.72758508e-01
-1.78080812e-01 -5.58032990e-01 -1.57068813e+00 -2.11481005e-01
-2.57104009e-01 -2.86077797e-01 9.00995433e-01 5.39093852e-01
2.76047438e-01 6.08625889e-01 3.29490513e-01 -1.48711216e+00
-1.42173856e-01 -2.55374074e-01 -7.05618560e-01 2.27632657e-01
4.43298250e-01 -1.09096873e+00 -3.38299423e-01 -1.89136118e-01] | [9.151484489440918, -3.220156669616699] |
683bfaf6-7e27-4114-b764-50fafc8a0f06 | fast-detection-of-multiple-objects-in-traffic | 1510.03125 | null | http://arxiv.org/abs/1510.03125v1 | http://arxiv.org/pdf/1510.03125v1.pdf | Fast detection of multiple objects in traffic scenes with a common detection framework | Traffic scene perception (TSP) aims to real-time extract accurate on-road
environment information, which in- volves three phases: detection of objects of
interest, recognition of detected objects, and tracking of objects in motion.
Since recognition and tracking often rely on the results from detection, the
ability to detect objects of interest effectively plays a crucial role in TSP.
In this paper, we focus on three important classes of objects: traffic signs,
cars, and cyclists. We propose to detect all the three important objects in a
single learning based detection framework. The proposed framework consists of a
dense feature extractor and detectors of three important classes. Once the
dense features have been extracted, these features are shared with all
detectors. The advantage of using one common framework is that the detection
speed is much faster, since all dense features need only to be evaluated once
in the testing phase. In contrast, most previous works have designed specific
detectors using different features for each of these objects. To enhance the
feature robustness to noises and image deformations, we introduce spatially
pooled features as a part of aggregated channel features. In order to further
improve the generalization performance, we propose an object subcategorization
method as a means of capturing intra-class variation of objects. We
experimentally demonstrate the effectiveness and efficiency of the proposed
framework in three detection applications: traffic sign detection, car
detection, and cyclist detection. The proposed framework achieves the
competitive performance with state-of- the-art approaches on several benchmark
datasets. | ['Anton Van Den Hengel', 'Sakrapee Paisitkriangkrai', 'Chunhua Shen', 'Qichang Hu', 'Fatih Porikli'] | 2015-10-12 | null | null | null | null | ['traffic-sign-detection'] | ['computer-vision'] | [ 1.86711997e-01 -5.63663483e-01 -6.47635311e-02 -3.27918559e-01
-6.29488111e-01 -2.21028507e-01 6.08263254e-01 1.41974585e-03
-7.63347566e-01 4.48362201e-01 -3.11831057e-01 -3.92344445e-02
1.52578838e-02 -9.22396660e-01 -5.84334850e-01 -9.19046819e-01
1.87768504e-01 2.18121618e-01 1.18328714e+00 5.98351471e-02
3.61868709e-01 8.32554817e-01 -1.99184585e+00 2.25131392e-01
8.87967527e-01 1.18600178e+00 4.09629524e-01 4.79707241e-01
-1.02038227e-01 4.50645506e-01 -3.79495054e-01 -8.19870830e-02
2.79101074e-01 -8.67319331e-02 -1.41683519e-01 3.65391910e-01
3.84400666e-01 -3.77881944e-01 -3.09744626e-01 1.06217599e+00
3.22818696e-01 2.27491856e-01 6.18257344e-01 -1.27351725e+00
7.71815330e-02 9.84121859e-02 -7.44622231e-01 4.69327122e-01
5.07548526e-02 3.45094353e-01 7.63600528e-01 -1.06833804e+00
3.14087361e-01 1.11218297e+00 2.96055317e-01 3.26588392e-01
-8.90081406e-01 -8.46368492e-01 3.19362015e-01 8.13615322e-01
-1.52147019e+00 -5.21426976e-01 8.14528525e-01 -4.13075507e-01
3.82968277e-01 3.27806234e-01 5.47480941e-01 5.22349477e-01
2.86940690e-02 9.64650810e-01 9.41309094e-01 -2.81785786e-01
1.56642959e-01 4.13745373e-01 5.01044989e-01 4.90203202e-01
5.18051088e-01 1.46433532e-01 -1.70878425e-01 1.65780321e-01
3.59788120e-01 2.07961991e-01 -3.61299701e-02 -4.65452373e-01
-1.03892243e+00 6.19661868e-01 5.05864799e-01 2.16732427e-01
-4.41803604e-01 4.34153900e-02 3.77407521e-01 -1.39357150e-01
-1.68400966e-02 -4.32486832e-01 -1.17667049e-01 -1.31709948e-01
-5.85470200e-01 2.03824013e-01 3.90537500e-01 8.59033287e-01
8.67395997e-01 -1.26252934e-01 -2.58222938e-01 8.54188263e-01
3.11053216e-01 7.52641797e-01 1.90077052e-01 -4.03190881e-01
6.87298894e-01 8.44820559e-01 1.20400086e-01 -1.12160456e+00
-4.08518076e-01 -4.61637348e-01 -6.23822927e-01 2.45133653e-01
4.31474477e-01 1.71123892e-01 -1.05840993e+00 1.25539255e+00
6.64189875e-01 3.36717188e-01 -1.29314110e-01 1.02838397e+00
8.43189955e-01 5.50502717e-01 8.62600952e-02 -5.78577965e-02
1.53243530e+00 -6.63555801e-01 -4.84982431e-01 -1.45029292e-01
3.67323339e-01 -8.98253858e-01 7.10038543e-01 1.80892631e-01
-6.43277705e-01 -9.83528316e-01 -8.92131805e-01 1.79454371e-01
-5.73392868e-01 5.58315098e-01 3.61137956e-01 7.39485681e-01
-3.85467023e-01 -1.76940542e-02 -6.88236475e-01 -4.12381738e-01
5.29677749e-01 4.09828067e-01 -2.60441840e-01 -2.02187046e-01
-9.47750032e-01 8.72523367e-01 4.98819590e-01 4.41857725e-01
-7.53887594e-01 -2.09731743e-01 -6.42693639e-01 6.50256574e-02
5.67458808e-01 -1.84932664e-01 8.60499859e-01 -6.72019720e-01
-1.13995051e+00 6.51966572e-01 -3.68793100e-01 -3.12383324e-01
6.05185270e-01 -1.72003746e-01 -6.65777802e-01 1.14522010e-01
7.49502257e-02 3.83952290e-01 8.67493093e-01 -1.07059646e+00
-1.36736834e+00 -3.02713037e-01 -4.72021922e-02 -3.20532173e-02
-2.99301594e-01 1.29239365e-01 -7.80988753e-01 -2.88257509e-01
2.30702803e-01 -9.52662706e-01 9.88196805e-02 -4.44998266e-03
-3.17770898e-01 -3.58991742e-01 1.41028953e+00 -4.34492290e-01
1.14445889e+00 -2.17501760e+00 -4.24725950e-01 4.12730277e-01
1.25843018e-01 6.88081145e-01 1.39645776e-02 9.54901278e-02
2.86838621e-01 -4.75886106e-01 2.49759853e-03 1.06488941e-02
-1.12796783e-01 1.24793507e-01 4.11684997e-02 4.81655300e-01
4.28978115e-01 7.23856628e-01 -5.66996634e-01 -7.83067346e-01
7.03290880e-01 4.81581718e-01 -2.97240704e-01 -1.64469570e-01
2.94197083e-01 3.36552650e-01 -9.04173553e-01 7.48038888e-01
1.05754173e+00 1.47328779e-01 -3.23877454e-01 -3.74681115e-01
-3.23514372e-01 8.75075012e-02 -1.67831743e+00 8.18487048e-01
-3.32029879e-01 4.84054863e-01 -9.99600999e-03 -1.24298799e+00
1.04107320e+00 4.59301919e-02 4.31252807e-01 -8.62269163e-01
1.38368875e-01 3.64635170e-01 3.32512766e-01 -6.41623259e-01
3.56860101e-01 1.65660724e-01 3.08991466e-02 1.04635663e-01
-3.08574826e-01 3.62426817e-01 5.48152447e-01 2.35407073e-02
9.17002916e-01 -2.25947291e-01 2.44481802e-01 1.30038679e-01
1.09152329e+00 -7.77339712e-02 7.84032941e-01 6.88626111e-01
-3.85876745e-01 2.18050733e-01 1.67865172e-01 -3.99635464e-01
-7.10281491e-01 -1.10165298e+00 -3.26345950e-01 8.19682181e-01
5.02911448e-01 3.63196172e-02 -3.98301035e-01 -8.19819510e-01
1.41167074e-01 4.09026921e-01 -5.13331354e-01 -1.20767117e-01
-8.40766013e-01 -6.97454870e-01 2.25240409e-01 6.59537196e-01
7.85530984e-01 -1.07651734e+00 -9.72522914e-01 2.43390590e-01
-2.13785879e-02 -1.33026302e+00 -3.68165702e-01 -3.43610793e-02
-6.02575421e-01 -1.27897608e+00 -4.49830115e-01 -7.99557984e-01
6.28846645e-01 7.21321642e-01 4.35108423e-01 1.45107761e-01
-5.76841950e-01 1.57707348e-01 -4.13253844e-01 -4.30108190e-01
-1.90335453e-01 -1.48143381e-01 -9.52791497e-02 8.34913313e-01
4.36051607e-01 -2.31450230e-01 -7.38322854e-01 8.27181399e-01
-6.15431547e-01 -1.74991474e-01 9.20424521e-01 5.88414073e-01
3.82146329e-01 1.92436025e-01 4.94134694e-01 -5.79224765e-01
1.61299914e-01 -8.90242904e-02 -7.09891260e-01 9.97875035e-02
-9.50290784e-02 -1.99683890e-01 3.89432907e-01 -3.73892933e-01
-1.14186907e+00 4.36857790e-01 5.96320592e-02 -1.72814816e-01
-3.61021191e-01 1.11339055e-01 -4.26825404e-01 -2.13991329e-01
3.35958153e-01 4.40681368e-01 -8.99148062e-02 -4.04670775e-01
2.96246678e-01 9.23992872e-01 4.03746426e-01 -3.28624606e-01
1.00160563e+00 5.54812908e-01 1.89352810e-01 -1.04918599e+00
-3.52818370e-01 -1.06017697e+00 -7.80461788e-01 -5.41966915e-01
6.35591745e-01 -7.58940041e-01 -9.04651403e-01 5.43345630e-01
-1.03420711e+00 3.99961591e-01 -1.36058867e-01 7.26632774e-01
-2.37091705e-01 4.61382627e-01 -1.65643916e-01 -1.04204357e+00
-1.01602517e-01 -1.25700438e+00 1.16809404e+00 3.37270945e-01
3.59298676e-01 -5.11167705e-01 -3.88213485e-01 2.93109924e-01
4.18639094e-01 1.20908037e-01 5.05762219e-01 -5.57821929e-01
-1.02289534e+00 -7.41806865e-01 -4.66275156e-01 4.48636979e-01
2.44580165e-01 -3.42841111e-02 -9.51714754e-01 -9.98976529e-02
-2.77327716e-01 1.09601729e-01 1.26441407e+00 3.33489329e-01
8.59576046e-01 1.89398542e-01 -7.53440797e-01 1.92248330e-01
1.16710091e+00 4.79114413e-01 6.24615014e-01 1.40714735e-01
6.87192798e-01 5.80049992e-01 1.00718081e+00 2.92793542e-01
1.56188175e-01 9.16805923e-01 3.08693349e-01 -5.65003380e-02
-1.85396567e-01 1.81514457e-01 3.67834896e-01 4.73178118e-01
-1.61864340e-01 -5.46941757e-02 -7.67561138e-01 5.62133670e-01
-2.00547647e+00 -1.27356136e+00 -4.62446183e-01 2.33938956e+00
2.07334101e-01 5.00465930e-01 3.99099946e-01 4.29366857e-01
9.97559428e-01 -1.29850879e-01 -4.81623203e-01 2.32620779e-02
1.06549755e-01 7.32031763e-02 5.90908289e-01 1.37453243e-01
-1.35446882e+00 7.33729959e-01 5.10894060e+00 9.71625745e-01
-1.20389271e+00 1.08072739e-02 2.33803391e-01 1.86107680e-01
3.90609235e-01 -7.28839785e-02 -1.27387440e+00 6.49875343e-01
4.11475182e-01 1.64475396e-01 -2.14794368e-01 8.18028033e-01
5.12470245e-01 -3.60022753e-01 -8.30895007e-01 8.40537846e-01
-8.12612176e-02 -9.65618968e-01 -1.16037637e-01 5.03270738e-02
2.44143873e-01 -1.48637712e-01 -1.64075121e-02 3.96872938e-01
-2.63883531e-01 -5.32850921e-01 6.89641237e-01 6.10373974e-01
3.42317551e-01 -8.28197479e-01 8.01089644e-01 4.89628226e-01
-1.66537678e+00 -3.88332874e-01 -4.16228443e-01 2.90417392e-02
1.98777646e-01 6.67725861e-01 -7.82660782e-01 5.49741685e-01
4.96993721e-01 6.77178562e-01 -6.76945627e-01 1.61559665e+00
-1.07595071e-01 5.18201947e-01 -4.98474419e-01 -1.38721272e-01
1.23170480e-01 -7.49391178e-03 5.81966460e-01 1.34387064e+00
1.83199137e-01 -6.73099793e-03 5.41655660e-01 7.05196857e-01
3.41721863e-01 2.05390453e-01 -3.60775560e-01 5.46582520e-01
5.41827798e-01 1.42123783e+00 -9.48521376e-01 -4.33731377e-01
-6.11086011e-01 5.98928452e-01 2.11317781e-02 2.20788732e-01
-1.09735250e+00 -6.67053521e-01 4.46631938e-01 3.16207618e-01
9.25337613e-01 -2.05501154e-01 4.66407463e-02 -8.82473648e-01
3.99026275e-01 -5.72960258e-01 3.18597972e-01 -2.81956613e-01
-8.60748410e-01 2.65757322e-01 6.61974177e-02 -1.61903965e+00
7.69728869e-02 -7.07671583e-01 -7.44270861e-01 6.38540566e-01
-1.55491912e+00 -1.37240255e+00 -6.10792518e-01 6.67920649e-01
6.30937994e-01 -3.71345021e-02 1.39934972e-01 6.66315079e-01
-8.02051663e-01 4.80818361e-01 6.58974200e-02 3.01714033e-01
4.65814412e-01 -6.50065601e-01 2.00085729e-01 1.03253114e+00
3.73874456e-02 3.33558559e-01 3.21143925e-01 -6.06276155e-01
-1.22171474e+00 -1.15494251e+00 6.46840334e-01 -1.78166896e-01
3.52773488e-01 -4.18841690e-01 -7.12970316e-01 3.36688370e-01
-4.50062335e-01 4.35738325e-01 3.27227637e-02 -1.66436642e-01
-4.27674726e-02 -6.64954245e-01 -9.22667146e-01 2.30677158e-01
8.98893535e-01 -1.68197632e-01 -4.36427176e-01 4.96169366e-02
8.86131600e-02 -1.40568107e-01 -3.16629380e-01 4.39167708e-01
7.50604630e-01 -8.43127370e-01 1.04951429e+00 -1.41858652e-01
-3.27076703e-01 -8.43567073e-01 -2.56659687e-02 -7.08843946e-01
-2.78618872e-01 -1.41824350e-01 2.60900240e-02 1.22929633e+00
9.41729546e-02 -7.71981955e-01 7.33733952e-01 3.66512120e-01
-1.53197646e-01 -4.39210415e-01 -1.10830116e+00 -9.05008733e-01
-5.58925092e-01 -6.28367424e-01 4.07072455e-01 4.18322325e-01
-3.61300737e-01 2.24026367e-01 -2.45317698e-01 3.02146345e-01
7.82588243e-01 2.39019871e-01 1.07467508e+00 -1.44610775e+00
-4.27345186e-02 -3.55937332e-01 -1.03061783e+00 -1.03280723e+00
-1.59925625e-01 -6.50146186e-01 1.50310293e-01 -1.31839585e+00
3.39275956e-01 -5.80522180e-01 -5.89424670e-01 2.44642884e-01
-3.27142656e-01 2.35470071e-01 2.45880172e-01 1.92547768e-01
-7.48000383e-01 4.57117260e-01 1.19499075e+00 -1.64961815e-01
-1.76222622e-01 3.97711247e-01 -2.68336415e-01 7.33664095e-01
7.01473951e-01 -4.38800782e-01 -2.13530138e-01 -2.83216685e-02
-6.55392408e-01 -3.71712089e-01 8.15117061e-01 -1.39184320e+00
5.07723689e-01 -1.68014675e-01 4.53917533e-01 -1.09787035e+00
5.04760623e-01 -9.12958741e-01 -1.01581715e-01 6.66071117e-01
8.86279494e-02 -3.55900973e-01 2.78162748e-01 6.98182046e-01
-2.53680706e-01 -1.14771798e-01 9.25408661e-01 1.18533894e-01
-1.30903327e+00 1.40567780e-01 -4.05261159e-01 -2.94005424e-01
1.48204875e+00 -5.68471789e-01 -1.77487582e-01 -2.73400787e-02
-5.31586766e-01 2.75751978e-01 -9.62006822e-02 5.92037439e-01
7.79253125e-01 -1.27212298e+00 -7.26426721e-01 3.96289378e-01
3.37130874e-01 -2.59674788e-01 3.33912700e-01 1.05329788e+00
-2.40732953e-01 6.28745675e-01 -2.41739571e-01 -9.36088204e-01
-1.53549790e+00 5.30150115e-01 2.72275031e-01 -2.34285966e-02
-5.49316704e-01 4.44997996e-01 4.22160625e-01 -1.03322431e-01
2.17704311e-01 -5.00802696e-01 -5.34805655e-01 4.64100316e-02
5.35633147e-01 6.25234067e-01 9.18965116e-02 -9.42992747e-01
-6.57758057e-01 1.07355082e+00 -2.33961061e-01 2.22741097e-01
8.47300708e-01 -1.31685719e-01 3.56382102e-01 1.93045110e-01
9.81950402e-01 -6.11718968e-02 -1.12286854e+00 -4.85557050e-01
6.37670457e-02 -6.58381164e-01 7.21546635e-02 -4.52355921e-01
-1.07522261e+00 9.86792266e-01 9.46337283e-01 -5.11739925e-02
1.13073552e+00 -3.75871658e-02 7.48524189e-01 3.45898986e-01
4.34897363e-01 -1.16911221e+00 -4.23912890e-02 2.98253685e-01
5.68013668e-01 -1.31798124e+00 5.12542389e-02 -7.67243028e-01
-4.60719317e-01 1.06488669e+00 7.19956696e-01 -3.17123607e-02
7.05230176e-01 6.80427477e-02 -1.17133558e-01 -1.60614148e-01
-4.00109619e-01 -9.19648528e-01 5.99954724e-01 5.25071561e-01
-3.08406875e-02 1.16699323e-01 -4.05239344e-01 2.03790084e-01
3.97991449e-01 -5.49051613e-02 1.89201519e-01 9.46743727e-01
-1.01416123e+00 -9.89244998e-01 -5.96547842e-01 4.81523871e-01
-1.10770769e-01 3.49166960e-01 -7.48323128e-02 1.10487890e+00
5.16439497e-01 1.05842352e+00 2.87995078e-02 -3.43235195e-01
7.66065598e-01 -1.43051088e-01 2.34714076e-01 -3.87840688e-01
-2.59143025e-01 4.83701825e-02 2.47069001e-02 -6.04939580e-01
-4.64094162e-01 -8.71398509e-01 -1.27234721e+00 9.04895365e-02
-6.65960431e-01 5.10953143e-02 6.55445337e-01 9.53560233e-01
1.90560192e-01 5.34482241e-01 7.65764236e-01 -7.92353153e-01
-4.03612882e-01 -7.70889878e-01 -5.10275364e-01 6.78230822e-01
2.66971856e-01 -1.11934185e+00 -6.94945306e-02 -7.40367025e-02] | [8.051230430603027, -0.8721227049827576] |
378db00b-36de-4c3b-a383-ca0de0f79357 | vgos-voxel-grid-optimization-for-view | 2304.13386 | null | https://arxiv.org/abs/2304.13386v2 | https://arxiv.org/pdf/2304.13386v2.pdf | VGOS: Voxel Grid Optimization for View Synthesis from Sparse Inputs | Neural Radiance Fields (NeRF) has shown great success in novel view synthesis due to its state-of-the-art quality and flexibility. However, NeRF requires dense input views (tens to hundreds) and a long training time (hours to days) for a single scene to generate high-fidelity images. Although using the voxel grids to represent the radiance field can significantly accelerate the optimization process, we observe that for sparse inputs, the voxel grids are more prone to overfitting to the training views and will have holes and floaters, which leads to artifacts. In this paper, we propose VGOS, an approach for fast (3-5 minutes) radiance field reconstruction from sparse inputs (3-10 views) to address these issues. To improve the performance of voxel-based radiance field in sparse input scenarios, we propose two methods: (a) We introduce an incremental voxel training strategy, which prevents overfitting by suppressing the optimization of peripheral voxels in the early stage of reconstruction. (b) We use several regularization techniques to smooth the voxels, which avoids degenerate solutions. Experiments demonstrate that VGOS achieves state-of-the-art performance for sparse inputs with super-fast convergence. Code will be available at https://github.com/SJoJoK/VGOS. | ['Huaizhong Lin', 'Wei Xing', 'Lei Zhao', 'Boyan Ji', 'Guangyuan Li', 'Jiafu Chen', 'Zhanjie Zhang', 'Jiakai Sun'] | 2023-04-26 | null | null | null | null | ['novel-view-synthesis'] | ['computer-vision'] | [ 1.31506398e-01 -2.41392851e-01 3.37852746e-01 -3.30618024e-01
-7.15089798e-01 -3.78999203e-01 2.58888870e-01 -2.64825970e-01
8.01125988e-02 6.71048284e-01 2.42031813e-01 -2.80732393e-01
-6.72239736e-02 -1.06913459e+00 -9.27443326e-01 -6.63716495e-01
2.69819915e-01 1.48319155e-02 1.74423084e-01 -2.61400372e-01
3.87123302e-02 6.45810127e-01 -1.63456774e+00 1.87729821e-01
1.09833360e+00 9.95211899e-01 4.46756572e-01 5.77325106e-01
-1.80446088e-01 7.25997210e-01 -3.38550776e-01 -1.41865179e-01
6.65721595e-01 -3.82557273e-01 -4.83154744e-01 -3.67821194e-02
8.27908754e-01 -4.87017035e-01 -3.67533833e-01 9.43790615e-01
5.19447982e-01 3.42263222e-01 4.51303363e-01 -8.25214624e-01
-3.97261322e-01 1.68549895e-01 -8.64830196e-01 -5.05653769e-02
1.16731972e-01 2.97408551e-01 7.08926201e-01 -1.15144217e+00
5.61862767e-01 9.28475201e-01 8.16743076e-01 3.45829129e-01
-1.25919294e+00 -5.55774987e-01 1.55778915e-01 -2.17037886e-01
-1.50122595e+00 -3.97867262e-01 8.88514817e-01 -3.27029288e-01
8.89747739e-01 4.90647852e-01 7.75695682e-01 7.71145165e-01
2.50669062e-01 3.29129934e-01 1.15799952e+00 -1.99948460e-01
2.58436412e-01 7.48801753e-02 -2.02424213e-01 7.53697217e-01
-3.14509347e-02 2.88490444e-01 -3.99946868e-01 -1.79701939e-01
1.24607015e+00 9.53753293e-02 -6.41300321e-01 -2.95054764e-01
-1.00724363e+00 8.38588059e-01 1.03988898e+00 4.52692918e-02
-4.71096426e-01 3.87259483e-01 2.74208412e-02 9.09059569e-02
5.80856860e-01 4.93072838e-01 -3.50281060e-01 1.86210752e-01
-9.77001488e-01 2.59296417e-01 3.63427639e-01 7.28084266e-01
1.00120425e+00 3.56559485e-01 1.02907158e-01 1.05789006e+00
1.88581750e-01 6.14868283e-01 9.69533175e-02 -1.09181988e+00
5.56682169e-01 3.35418433e-01 1.44809326e-02 -1.07776010e+00
-3.21541220e-01 -6.88735604e-01 -1.14844573e+00 4.69687313e-01
1.92999735e-01 -1.31827399e-01 -1.07853699e+00 1.36196911e+00
6.04127407e-01 2.12844625e-01 -5.30563705e-02 1.18211150e+00
1.02919900e+00 1.07417929e+00 -3.86518151e-01 4.62468006e-02
1.08390784e+00 -1.16553676e+00 -3.21266770e-01 -2.46404991e-01
3.45955461e-01 -8.31984043e-01 1.32353723e+00 4.56916541e-01
-1.16010690e+00 -5.58115244e-01 -8.51344824e-01 -1.81125924e-01
5.32288663e-03 -1.63800180e-01 8.94294262e-01 3.43235523e-01
-8.82234037e-01 5.56030035e-01 -6.99342728e-01 1.31504446e-01
5.34020364e-01 1.00761801e-01 -2.37984940e-01 -4.78565603e-01
-7.65650749e-01 4.51989919e-01 -4.64086123e-02 1.46180600e-01
-8.43949795e-01 -1.20074940e+00 -9.15087402e-01 8.62185359e-02
4.25579071e-01 -9.49644923e-01 8.46941352e-01 -1.03903723e+00
-1.48429823e+00 3.18902612e-01 -1.21506944e-01 1.72581479e-01
4.32499170e-01 -2.40386277e-01 1.39676472e-02 2.21870989e-01
-6.88053370e-02 7.53613472e-01 8.13005447e-01 -1.68109083e+00
-3.01394612e-01 -2.72138655e-01 7.63387382e-02 3.38681281e-01
-1.25137746e-01 -3.02752465e-01 -4.03900802e-01 -8.39858472e-01
5.01792431e-01 -7.14278817e-01 -6.56045079e-01 1.89300388e-01
-1.75380081e-01 4.40177381e-01 5.76596975e-01 -6.07633114e-01
9.29939926e-01 -2.17294741e+00 -3.86876911e-02 3.35574657e-01
2.71131665e-01 -7.93431401e-02 -1.40019551e-01 1.50821835e-01
-4.82639782e-02 -8.43516961e-02 -4.14147168e-01 -3.35839957e-01
-5.10587573e-01 1.21174142e-01 -3.21895897e-01 4.36464399e-01
-1.14607416e-01 7.15284228e-01 -7.39268959e-01 -7.84424767e-02
4.56270248e-01 9.22062635e-01 -8.33372176e-01 1.33792564e-01
-1.92898095e-01 7.70602286e-01 -3.91149074e-01 6.57172918e-01
1.04196846e+00 -4.28612977e-01 -1.80349961e-01 -4.54660177e-01
-2.60739416e-01 1.66756600e-01 -1.38631618e+00 1.62972903e+00
-8.73047948e-01 4.70098883e-01 2.52580017e-01 -5.09264648e-01
8.22305560e-01 1.84174374e-01 6.28677130e-01 -7.73348689e-01
1.70565285e-02 1.91698104e-01 -3.82953823e-01 -1.59057915e-01
4.67492461e-01 -2.21433610e-01 3.27763349e-01 2.12113157e-01
-7.32300282e-02 -4.66160148e-01 -1.38644546e-01 6.84691519e-02
7.20077038e-01 2.00658754e-01 1.39411077e-01 -1.89747572e-01
2.73153722e-01 8.42832625e-02 7.26348162e-01 5.19452691e-01
5.16845047e-01 1.13395584e+00 3.04716881e-02 -5.68635166e-01
-1.11530995e+00 -1.03967416e+00 -3.01821977e-01 6.50190711e-01
1.40387923e-01 -4.57428157e-01 -6.89354420e-01 -3.11834544e-01
-2.15037331e-01 6.02795005e-01 -3.20616663e-01 2.48711422e-01
-7.42779195e-01 -7.95020461e-01 8.33313260e-03 6.39846683e-01
5.35589814e-01 -8.05215716e-01 -7.42636979e-01 1.51687369e-01
-1.22663282e-01 -9.72843826e-01 -3.83054584e-01 2.39428625e-01
-1.23401535e+00 -7.74891913e-01 -9.09376562e-01 -4.42994356e-01
1.09671402e+00 6.73692167e-01 1.28926897e+00 2.17770144e-01
-2.50281155e-01 1.45172149e-01 -1.92515865e-01 -1.64279878e-01
-7.99032077e-02 -3.31806719e-01 -2.20831200e-01 -2.71801949e-01
-4.67638612e-01 -9.27732170e-01 -9.86228824e-01 4.06260908e-01
-9.92147624e-01 5.44358194e-01 4.65733469e-01 9.19373214e-01
1.00123715e+00 7.60309175e-02 7.67125338e-02 -9.36973095e-01
9.03031826e-02 -3.63593310e-01 -1.02320230e+00 -4.71171141e-02
-5.06607950e-01 -6.38793036e-02 1.07420349e+00 -3.17216873e-01
-1.19906521e+00 1.32855892e-01 -3.25918198e-01 -7.13337421e-01
-8.08105320e-02 4.64875728e-01 6.94617778e-02 -5.14975667e-01
8.10015500e-01 2.09504329e-02 -3.03747147e-01 -6.28306806e-01
2.84768790e-01 2.01497540e-01 4.29367810e-01 -5.67371786e-01
8.94750774e-01 6.15308523e-01 1.56400740e-01 -8.79993141e-01
-8.80783200e-01 -4.47706789e-01 -3.12913477e-01 -3.67891490e-01
5.46848476e-01 -1.21095574e+00 -2.52732337e-01 2.97129154e-01
-9.29966629e-01 -6.76839232e-01 -3.67461443e-01 4.53271836e-01
-3.54040205e-01 1.21865541e-01 -5.35818338e-01 -5.34760058e-01
-4.29684997e-01 -1.15494335e+00 9.99924362e-01 4.44931030e-01
1.70791075e-01 -7.73434579e-01 1.32941321e-01 5.27821004e-01
5.14514744e-01 3.14046323e-01 6.84975505e-01 5.22094071e-01
-1.06743157e+00 3.43554199e-01 -2.72158861e-01 3.90674472e-01
6.41073147e-03 -7.82189444e-02 -1.14658296e+00 -3.28085363e-01
1.61319822e-01 -2.61820108e-01 7.84334183e-01 6.85727596e-01
1.42282414e+00 -3.20006639e-01 -1.73282415e-01 1.33888257e+00
2.00296950e+00 3.02685667e-02 9.44514692e-01 1.39018208e-01
1.11661768e+00 6.10561788e-01 5.16105592e-01 4.65322226e-01
2.38823652e-01 6.33183062e-01 5.23979843e-01 -5.13797700e-01
-3.25892240e-01 -2.69224018e-01 1.53289467e-01 8.56075883e-01
-2.95787841e-01 -2.66354114e-01 -6.84903622e-01 4.98526216e-01
-1.61631000e+00 -7.70129085e-01 -3.74525487e-01 2.18817878e+00
5.76801717e-01 -9.70011577e-02 -1.90031528e-01 -1.31923966e-02
1.27771214e-01 4.13780093e-01 -4.03804749e-01 -3.13174248e-01
4.58727824e-03 3.62761378e-01 6.72405541e-01 6.57391548e-01
-6.52360141e-01 7.77516305e-01 6.08072424e+00 7.89215386e-01
-1.36278522e+00 1.67045191e-01 9.31969941e-01 -3.49452645e-01
-8.13995063e-01 3.65677364e-02 -6.84028864e-01 3.13634336e-01
4.52415526e-01 3.77568722e-01 6.97776735e-01 8.41558874e-01
3.40500832e-01 -3.88927639e-01 -6.38632834e-01 1.13516533e+00
6.68660849e-02 -1.55763876e+00 -5.45617100e-03 -2.57999506e-02
1.24454439e+00 3.27018499e-01 -5.41680446e-03 -8.32588673e-02
5.58763705e-02 -1.19604897e+00 6.27634525e-01 4.49987262e-01
9.34234202e-01 -8.15095603e-01 4.64343071e-01 3.58023137e-01
-1.21629846e+00 1.47072658e-01 -6.74722850e-01 1.56183124e-01
3.67811739e-01 1.07618082e+00 -3.30422670e-01 6.96506321e-01
9.57143545e-01 4.34704572e-01 -3.95390928e-01 9.73486364e-01
-1.77053977e-02 4.57339048e-01 -5.42235315e-01 3.16422760e-01
3.24480891e-01 -5.76611519e-01 5.14096856e-01 7.13365912e-01
5.37959397e-01 3.31245273e-01 8.76991525e-02 1.07740259e+00
3.12842801e-03 1.03305928e-01 -6.76221907e-01 5.87121904e-01
5.06446473e-02 1.38434613e+00 -7.00239658e-01 -8.08728039e-02
-5.93176067e-01 8.72008681e-01 2.11507708e-01 5.09841084e-01
-8.94048572e-01 -1.93875045e-01 5.39029479e-01 5.69084108e-01
4.14940119e-01 -1.78537533e-01 -6.09939218e-01 -1.13081384e+00
1.84651837e-01 -8.83916378e-01 6.72334805e-02 -9.87819374e-01
-1.07803357e+00 7.69793451e-01 -1.00573786e-01 -1.29529250e+00
1.34629011e-01 -4.93537664e-01 -6.52397513e-01 9.11316276e-01
-1.62659490e+00 -8.96175325e-01 -8.48228753e-01 5.24352551e-01
5.72461665e-01 1.57291740e-01 6.04416430e-01 4.81936097e-01
-2.88022578e-01 2.99263000e-01 1.82092011e-01 -2.13210717e-01
4.13033575e-01 -9.38709497e-01 3.75519037e-01 7.30371177e-01
2.33391657e-01 3.72667968e-01 5.12617767e-01 -5.41964591e-01
-1.37912393e+00 -1.15174615e+00 3.10485333e-01 -8.22682753e-02
5.72852679e-02 -2.95370519e-01 -1.00281858e+00 5.54663837e-01
-4.61239414e-03 4.74630147e-01 4.87456590e-01 -5.58904037e-02
-1.18082985e-01 -2.95779794e-01 -1.19861543e+00 6.34877682e-01
1.06479287e+00 -2.34619096e-01 -3.75664160e-02 3.06039900e-01
5.77365935e-01 -7.18607783e-01 -9.00504053e-01 6.39724612e-01
3.54172796e-01 -1.33030701e+00 1.21074605e+00 1.45095885e-01
7.05092132e-01 -5.84134936e-01 -2.30743080e-01 -1.47043157e+00
-4.55930352e-01 -5.82771659e-01 -8.54301453e-02 9.06956017e-01
3.82395387e-01 -6.14540219e-01 7.80014634e-01 3.10486287e-01
-2.98473150e-01 -1.14749777e+00 -6.60561025e-01 -5.86898506e-01
-9.30351540e-02 -3.53699803e-01 5.85044086e-01 9.39626932e-01
-6.63748324e-01 3.28061908e-01 -3.40646297e-01 3.39988381e-01
6.59765244e-01 3.74091864e-01 8.63325715e-01 -8.82436454e-01
-5.56775689e-01 -2.07884833e-01 2.48339847e-02 -1.37386179e+00
-3.16877335e-01 -6.64391577e-01 1.45345628e-01 -1.67013395e+00
5.58893681e-02 -8.88086319e-01 5.51361628e-02 2.41853863e-01
-1.98158756e-01 5.33449173e-01 1.30823016e-01 3.14977989e-02
-1.16810776e-01 7.23629177e-01 1.57104814e+00 1.26322553e-01
-3.15402478e-01 -2.09714100e-01 -6.32857382e-01 7.48171449e-01
6.77945614e-01 -4.81540859e-01 -5.09161532e-01 -8.52595747e-01
3.78303409e-01 1.65932253e-01 5.52451849e-01 -1.16109538e+00
-1.03098311e-01 -2.68986315e-01 6.15181625e-01 -7.44138479e-01
5.51460981e-01 -8.45982075e-01 7.23187208e-01 1.57435864e-01
1.78410754e-01 2.19726246e-02 2.53086954e-01 2.35176593e-01
-2.80891865e-01 -2.78242260e-01 1.01680374e+00 -4.25410151e-01
-4.60912377e-01 4.03424501e-01 -1.01204261e-01 -6.15743883e-02
8.06042731e-01 -3.37819248e-01 -2.08753467e-01 -3.33361447e-01
-2.67057568e-01 2.06263468e-01 7.22350061e-01 4.24925890e-03
7.90844142e-01 -1.31739759e+00 -5.96225083e-01 4.25883859e-01
-1.73783496e-01 6.69995785e-01 6.00235879e-01 6.81422770e-01
-9.86440420e-01 3.12775411e-02 -3.48130451e-03 -7.49063373e-01
-1.16058874e+00 3.04065198e-01 4.15048569e-01 -1.69391155e-01
-1.10860586e+00 1.01300609e+00 5.86232781e-01 -5.76042414e-01
6.50362950e-03 -3.34106386e-01 8.61269981e-02 -3.10253054e-01
4.22044158e-01 3.73459995e-01 1.79377079e-01 -4.14483488e-01
-1.02827415e-01 9.13422942e-01 1.52287304e-01 1.36486236e-02
1.60637069e+00 1.20494589e-01 4.69564274e-02 2.08591923e-01
1.18684208e+00 3.59659344e-01 -1.72362137e+00 3.63217369e-02
-8.40943694e-01 -8.84457946e-01 6.30779445e-01 -5.84311545e-01
-1.54198790e+00 7.15954185e-01 4.36124891e-01 -1.47992581e-01
1.28924024e+00 -2.15878591e-01 1.20250237e+00 1.17435548e-02
4.52668190e-01 -8.63877952e-01 1.22764576e-02 4.19059515e-01
9.99690354e-01 -1.03552651e+00 3.95445764e-01 -7.90285110e-01
-5.11931777e-01 9.93620694e-01 6.70104086e-01 -1.67439386e-01
7.03213990e-01 4.95613426e-01 2.08379865e-01 -3.83779436e-01
-5.49391925e-01 2.23924875e-01 3.66262704e-01 3.59708369e-01
3.33152860e-01 -1.74569581e-02 1.22512221e-01 5.43343611e-02
-2.78671443e-01 -2.72173524e-01 3.09326947e-01 8.14695776e-01
-1.77175254e-01 -7.93985605e-01 -6.36480868e-01 5.96972406e-01
-2.92197466e-01 -3.29351664e-01 1.31202444e-01 3.91922563e-01
5.58865182e-02 5.90186536e-01 1.40415072e-01 -6.81490600e-02
5.05806506e-01 -5.24516702e-01 6.88922703e-01 -4.91894513e-01
-7.38411605e-01 3.76957923e-01 -6.03521466e-02 -9.46422696e-01
-3.12074304e-01 -3.73119742e-01 -1.13250697e+00 -4.74942893e-01
-3.80892992e-01 -9.06054974e-02 5.14143407e-01 4.61139202e-01
3.38011682e-01 7.61227906e-01 9.13890123e-01 -1.01035273e+00
-2.50514537e-01 -6.40403748e-01 -4.32118535e-01 1.61349550e-01
2.15861291e-01 -5.69857359e-01 -4.69025075e-01 -1.40526697e-01] | [9.494585990905762, -2.9673984050750732] |
b081d588-133e-40ad-9831-47450d7ba873 | video-p2p-video-editing-with-cross-attention | 2303.04761 | null | https://arxiv.org/abs/2303.04761v1 | https://arxiv.org/pdf/2303.04761v1.pdf | Video-P2P: Video Editing with Cross-attention Control | This paper presents Video-P2P, a novel framework for real-world video editing with cross-attention control. While attention control has proven effective for image editing with pre-trained image generation models, there are currently no large-scale video generation models publicly available. Video-P2P addresses this limitation by adapting an image generation diffusion model to complete various video editing tasks. Specifically, we propose to first tune a Text-to-Set (T2S) model to complete an approximate inversion and then optimize a shared unconditional embedding to achieve accurate video inversion with a small memory cost. For attention control, we introduce a novel decoupled-guidance strategy, which uses different guidance strategies for the source and target prompts. The optimized unconditional embedding for the source prompt improves reconstruction ability, while an initialized unconditional embedding for the target prompt enhances editability. Incorporating the attention maps of these two branches enables detailed editing. These technical designs enable various text-driven editing applications, including word swap, prompt refinement, and attention re-weighting. Video-P2P works well on real-world videos for generating new characters while optimally preserving their original poses and scenes. It significantly outperforms previous approaches. | ['Jiaya Jia', 'Zhe Lin', 'Wenbo Li', 'Yuechen Zhang', 'Shaoteng Liu'] | 2023-03-08 | null | null | null | null | ['video-generation'] | ['computer-vision'] | [ 4.39710855e-01 -5.38958982e-02 -9.03030783e-02 2.45458446e-02
-5.83807349e-01 -3.79422456e-01 6.72203660e-01 -2.52053529e-01
-2.38230407e-01 5.75841486e-01 5.34044564e-01 1.01655819e-01
3.58832628e-01 -4.97210979e-01 -8.70725989e-01 -4.63100582e-01
4.98211056e-01 3.09811294e-01 2.31467694e-01 -3.26454103e-01
5.67400515e-01 2.83356577e-01 -1.35631609e+00 2.87376076e-01
1.13034058e+00 6.76773071e-01 8.53320658e-01 1.06688213e+00
-2.05504388e-01 8.97937179e-01 -5.82239091e-01 -5.12207627e-01
3.41487318e-01 -5.68866372e-01 -4.65942740e-01 1.12238768e-02
5.87555945e-01 -6.85265779e-01 -4.33879733e-01 8.87744546e-01
6.87911987e-01 2.12829307e-01 5.21535039e-01 -1.28878665e+00
-1.36766768e+00 4.01914775e-01 -7.69942999e-01 -1.95688456e-02
6.32709205e-01 3.29564393e-01 8.10291708e-01 -1.09910464e+00
7.88441598e-01 1.25198472e+00 4.78831112e-01 7.32388258e-01
-1.10112917e+00 -7.16900885e-01 2.26540074e-01 3.85908484e-01
-1.34773040e+00 -5.50928533e-01 6.60037220e-01 -4.40203100e-01
9.52776253e-01 3.71616423e-01 9.34765399e-01 1.08814585e+00
2.83315271e-01 9.43885744e-01 5.83014488e-01 -4.70849931e-01
-2.07755074e-01 1.41786054e-01 -4.51857746e-01 6.60817266e-01
-3.24762046e-01 -3.21507044e-02 -8.67664695e-01 2.75624335e-01
1.21167278e+00 -2.72486228e-02 -6.43919289e-01 -1.76831394e-01
-1.60797238e+00 5.78582585e-01 1.03061765e-01 1.30912066e-02
-4.60321903e-01 3.72358173e-01 4.18079525e-01 2.62146443e-01
4.18789744e-01 6.31587684e-01 -1.93148032e-01 -3.70500386e-01
-9.99817550e-01 2.96025962e-01 3.63374591e-01 1.47594833e+00
5.83896220e-01 2.45800465e-01 -9.24356520e-01 9.36234057e-01
1.52682707e-01 3.97667408e-01 3.64853144e-01 -1.43154514e+00
5.81278324e-01 1.89975291e-01 2.06454158e-01 -1.07576764e+00
3.37729037e-01 3.45724635e-02 -7.92168140e-01 7.35686626e-04
-1.77706424e-02 -1.33646533e-01 -8.80828559e-01 1.81776214e+00
1.34118408e-01 2.65055269e-01 -3.24516952e-01 9.19202149e-01
5.58627129e-01 1.06499016e+00 -3.35570201e-02 -1.51579529e-01
1.17775106e+00 -1.49037123e+00 -1.05085015e+00 -3.67720500e-02
7.02983290e-02 -9.81437683e-01 1.35513127e+00 1.84734955e-01
-1.47464812e+00 -7.23723173e-01 -9.31305289e-01 -4.70355928e-01
1.51116118e-01 3.79255533e-01 3.60405356e-01 1.98442191e-01
-1.32178867e+00 4.91380215e-01 -5.38150847e-01 -2.65517771e-01
2.38135636e-01 1.19668677e-01 -3.31513733e-01 -1.03107832e-01
-1.12247550e+00 6.78655326e-01 1.88045710e-01 3.29971127e-02
-1.03361630e+00 -8.71051848e-01 -7.59114623e-01 1.89378917e-01
3.57637584e-01 -8.68868411e-01 1.00017226e+00 -1.14132619e+00
-2.02600265e+00 5.40835917e-01 -2.16191605e-01 -1.65131286e-01
6.20350599e-01 -5.58325946e-01 -1.25268549e-01 2.37341747e-01
2.62323111e-01 1.28192782e+00 1.34240794e+00 -1.09596181e+00
-4.71877307e-01 1.80464476e-01 1.44808799e-01 5.61766922e-01
-7.46719420e-01 1.63096189e-01 -1.29026806e+00 -1.29484630e+00
-3.79014373e-01 -8.41520011e-01 -1.16037063e-01 4.28425997e-01
-3.55606198e-01 2.10967287e-01 9.78073359e-01 -9.44047809e-01
1.43700385e+00 -2.13184047e+00 6.94557786e-01 -1.87408835e-01
2.92542070e-01 3.63047123e-01 -6.08254671e-01 4.42413270e-01
-6.40277788e-02 5.48698604e-02 -6.33379966e-02 -4.53524739e-01
-1.82953939e-01 6.47980720e-02 -3.12462717e-01 -1.43675730e-01
4.74646240e-01 8.65521491e-01 -8.76787484e-01 -6.44013047e-01
3.30359638e-01 6.91762865e-01 -1.25683630e+00 4.58832443e-01
-2.47069702e-01 4.25648302e-01 -1.02428578e-01 4.08020765e-01
4.66251254e-01 -1.47083342e-01 5.51431114e-03 -4.31600481e-01
-1.83889151e-01 -2.08030000e-01 -9.12901700e-01 1.80293143e+00
-5.44761837e-01 9.57689822e-01 1.42314643e-01 -4.63127673e-01
6.98695421e-01 2.60094911e-01 4.11374092e-01 -5.83692968e-01
3.61937992e-02 -1.57579273e-01 -3.46838862e-01 -5.00781894e-01
1.19159853e+00 2.65635997e-01 3.38499814e-01 3.32623571e-01
1.57172740e-01 -4.44585264e-01 5.23274362e-01 6.77283525e-01
8.47121179e-01 4.89950478e-01 -3.44468281e-02 -1.24148205e-01
5.61757386e-01 -3.14150333e-01 6.56646729e-01 6.33001387e-01
-3.86045016e-02 1.17239761e+00 4.83272910e-01 -1.18036315e-01
-1.23403442e+00 -8.95174861e-01 4.75548148e-01 1.27562940e+00
2.37159446e-01 -7.53485799e-01 -9.73524332e-01 -2.96447664e-01
-2.60078460e-01 7.79777706e-01 -5.10314047e-01 -2.49945059e-01
-6.16656542e-01 -2.55515933e-01 2.75070250e-01 6.05017900e-01
5.48213720e-01 -1.15270770e+00 -2.13164896e-01 3.19849104e-01
-6.60606086e-01 -9.51318204e-01 -1.56388223e+00 -4.30894047e-01
-4.32314813e-01 -8.77645850e-01 -1.18857121e+00 -8.79842341e-01
8.51468325e-01 2.91516542e-01 9.58808422e-01 1.13182612e-01
-1.77742347e-01 6.43804848e-01 -4.36647564e-01 -1.80716604e-01
-4.49452460e-01 -1.70210063e-01 -2.05196626e-02 1.62926480e-01
-1.67681739e-01 -2.80484200e-01 -6.09020293e-01 3.27352643e-01
-8.52415800e-01 6.76164627e-01 4.17579979e-01 9.53356683e-01
5.49144864e-01 -5.10833442e-01 2.51266658e-01 -5.13943315e-01
8.06290209e-01 -7.32248276e-02 -3.93392801e-01 4.36481446e-01
-3.06236953e-01 2.66340673e-02 6.40948892e-01 -7.23503172e-01
-1.31304133e+00 -1.78757504e-01 -2.26128753e-02 -8.67241621e-01
4.93330866e-01 9.49156657e-02 -2.15073004e-01 -1.17877722e-02
3.47506881e-01 3.54552209e-01 9.57635641e-02 -1.44290119e-01
5.25579035e-01 4.47920293e-01 6.74000204e-01 -6.86530232e-01
6.74463272e-01 1.72996461e-01 -5.73835671e-01 -8.36339235e-01
-3.95004570e-01 -1.49811413e-02 -4.59512979e-01 -4.26088601e-01
1.03338480e+00 -1.01551032e+00 -4.20679957e-01 7.45494664e-01
-1.46904981e+00 -5.29083967e-01 -4.22894001e-01 3.91538918e-01
-7.51899004e-01 5.16804934e-01 -8.32527518e-01 -2.64575601e-01
-5.01623034e-01 -1.43852067e+00 1.11672354e+00 8.87766704e-02
-2.96577275e-01 -6.06312990e-01 7.05020577e-02 3.52131635e-01
6.55029655e-01 -2.39832208e-01 6.69363260e-01 1.27217844e-01
-8.89748156e-01 1.57113791e-01 -4.61081564e-01 4.69465435e-01
1.12616561e-01 5.31689882e-01 -4.18248743e-01 -3.44131082e-01
-3.35683525e-01 -1.53330043e-01 7.90000796e-01 3.15288514e-01
1.37977898e+00 -3.08751106e-01 -1.99478455e-02 8.40489686e-01
9.59416807e-01 2.51294404e-01 9.50219035e-01 9.68218148e-02
1.02020943e+00 1.57236248e-01 6.29453182e-01 5.82805455e-01
4.50719476e-01 8.67356241e-01 1.32359996e-01 -2.55312938e-02
-5.12891352e-01 -4.89452630e-01 6.65123165e-01 1.18262303e+00
-2.92235583e-01 -5.02300739e-01 -5.91457784e-01 6.32480681e-01
-1.69083893e+00 -1.07356918e+00 2.04884499e-01 2.09601378e+00
1.03104544e+00 -4.91825230e-02 -3.79330546e-01 -3.06077421e-01
9.76104498e-01 2.76291490e-01 -4.14681613e-01 -2.40063369e-01
-7.68101960e-02 5.09524159e-02 2.85280049e-01 6.92025661e-01
-7.61814892e-01 1.09662724e+00 6.49700546e+00 9.66445029e-01
-1.09032667e+00 2.19372839e-01 4.61862653e-01 -4.77573842e-01
-5.58842242e-01 8.06085765e-02 -6.73638582e-01 6.22253954e-01
3.97318870e-01 -3.74778897e-01 6.60746217e-01 4.64054167e-01
4.18662190e-01 1.31686419e-01 -1.06107271e+00 1.15613937e+00
4.98604655e-01 -1.60905099e+00 6.96444213e-01 -2.51650959e-01
9.51457381e-01 -2.36448452e-01 4.54759449e-02 4.19459432e-01
1.50543377e-01 -5.49629450e-01 1.05525720e+00 7.20604539e-01
1.29673195e+00 -5.13139188e-01 1.98930785e-01 -7.52435550e-02
-1.26048994e+00 -6.11370727e-02 -1.99266776e-01 3.22894156e-01
6.03031933e-01 2.64039725e-01 -3.35977763e-01 1.76222667e-01
5.92170596e-01 9.42883432e-01 -3.52734536e-01 7.24644303e-01
-2.58423328e-01 2.67763674e-01 -6.48244917e-02 1.79543480e-01
3.68671352e-03 -4.52462465e-01 7.84490645e-01 1.07137454e+00
6.52701616e-01 1.68175265e-01 -1.79414768e-02 8.02228034e-01
-3.02992433e-01 3.49874273e-02 -3.80036175e-01 -3.52594018e-01
4.86160338e-01 1.14810359e+00 -4.12248820e-01 -5.19689679e-01
-3.36681783e-01 1.62904966e+00 2.25546554e-01 4.76960540e-01
-1.36020696e+00 -5.99707186e-01 7.56276727e-01 2.74990559e-01
3.66348535e-01 -2.83060431e-01 1.93660483e-01 -1.37725937e+00
-8.85816291e-03 -9.64191794e-01 -2.13505015e-01 -1.32345057e+00
-7.93713808e-01 4.76736814e-01 -7.02899322e-02 -1.12553561e+00
-3.61580960e-02 -2.71494210e-01 -7.73269892e-01 7.25552619e-01
-1.29504120e+00 -1.12626481e+00 -5.66954136e-01 7.04570472e-01
9.97742832e-01 -2.87184328e-01 4.29211587e-01 5.59263527e-01
-6.64699614e-01 6.62603378e-01 6.06688745e-02 -5.94757386e-02
1.11257124e+00 -1.05001736e+00 6.12554491e-01 9.31150854e-01
-1.88261524e-01 6.32218599e-01 5.31060696e-01 -7.94989884e-01
-1.37008643e+00 -1.18669534e+00 6.02658212e-01 -2.00828835e-01
3.46972823e-01 -2.56207645e-01 -6.81126714e-01 7.15870202e-01
6.10662103e-01 -3.30232263e-01 3.29928488e-01 -4.48443681e-01
-1.09775081e-01 -1.37741148e-01 -7.05896318e-01 1.16237605e+00
1.21443152e+00 -5.28649867e-01 -2.75008321e-01 4.05242026e-01
9.95031536e-01 -5.65314174e-01 -7.35495090e-01 1.28153563e-01
5.63498855e-01 -7.84414172e-01 9.28882658e-01 -2.53345162e-01
8.87853444e-01 -3.99408638e-01 1.08865078e-03 -1.48102915e+00
-6.18971348e-01 -1.14970171e+00 -1.28752455e-01 1.38576686e+00
2.12285951e-01 -3.16490859e-01 5.87849200e-01 5.22807837e-01
-4.09282118e-01 -5.32332480e-01 -3.96658450e-01 -3.51528525e-01
-3.84122580e-01 -1.74773291e-01 4.98828202e-01 8.30722094e-01
-2.38222942e-01 1.92456305e-01 -9.63210404e-01 -1.82106987e-01
4.00648743e-01 -2.08963171e-01 9.82456982e-01 -5.09227157e-01
-3.75335842e-01 -4.07244653e-01 -1.14784226e-01 -1.57543147e+00
-2.14812998e-02 -6.34436369e-01 6.41920045e-02 -1.57248795e+00
3.47747684e-01 -2.08169043e-01 1.51427299e-01 3.61058921e-01
-4.06356066e-01 2.84200639e-01 6.41319275e-01 2.36500040e-01
-4.92873371e-01 1.03033459e+00 1.65347493e+00 -2.07510173e-01
-2.59571373e-01 -5.51119208e-01 -6.74793959e-01 4.59003747e-01
5.56769788e-01 -2.26884447e-02 -6.05967164e-01 -9.88195300e-01
1.38102233e-01 1.43366754e-01 1.61283284e-01 -1.01598763e+00
3.99210364e-01 -2.54596323e-01 2.92744935e-01 -5.92619658e-01
4.45192099e-01 -5.85755765e-01 3.55934948e-01 1.44058898e-01
-4.32850629e-01 3.96720529e-01 1.00844942e-01 6.17014766e-01
-2.35666543e-01 -9.70491096e-02 7.49497473e-01 -1.07788697e-01
-7.26179004e-01 5.09405315e-01 -4.54230040e-01 7.53582492e-02
1.19543600e+00 -3.03981334e-01 -1.03198543e-01 -7.10116327e-01
-6.16969168e-01 3.95776242e-01 6.41797304e-01 7.60917485e-01
9.16697502e-01 -1.55407906e+00 -8.44792783e-01 4.18882757e-01
-3.54539901e-02 -2.40161970e-01 4.44605917e-01 6.69268966e-01
-9.04534996e-01 1.63714945e-01 -4.80427980e-01 -5.71596205e-01
-1.41927910e+00 4.48699147e-01 1.11895695e-01 -1.05028965e-01
-4.46148008e-01 9.89673495e-01 4.37160373e-01 -2.46272653e-01
1.01907454e-01 -1.69414759e-01 1.91854686e-02 -2.44497252e-03
7.63844669e-01 3.57197404e-01 -3.89188886e-01 -5.08664668e-01
8.97386596e-02 6.54853523e-01 -5.30420959e-01 -1.48058057e-01
1.10435355e+00 -3.73012990e-01 -8.34902301e-02 -9.93746519e-02
9.07934427e-01 6.42372072e-02 -1.68787527e+00 1.87827814e-02
-7.58938849e-01 -8.55360746e-01 4.78812344e-02 -5.40528119e-01
-1.38774478e+00 8.12350750e-01 1.95958599e-01 -3.46980125e-01
1.24791002e+00 -5.55483997e-01 9.34673429e-01 1.76253825e-01
2.88742721e-01 -1.30854237e+00 6.05775356e-01 6.92207158e-01
1.44799733e+00 -8.64648283e-01 -5.45792468e-02 -4.43105757e-01
-1.05284369e+00 8.67673099e-01 9.83473420e-01 4.75332104e-02
1.85391054e-01 2.63010949e-01 -2.65106052e-01 1.61227793e-01
-8.84786725e-01 3.30642641e-01 9.56682190e-02 5.97583234e-01
3.89273435e-01 -2.49758810e-01 -2.21089438e-01 1.17437072e-01
8.77261609e-02 1.84756562e-01 6.44067824e-01 6.59521937e-01
-1.24390356e-01 -1.05710173e+00 -3.93083453e-01 4.21449870e-01
-1.94277212e-01 -4.31583792e-01 -2.18098536e-01 5.03499210e-01
-7.89129138e-02 5.41156054e-01 1.86316893e-01 -3.00289631e-01
2.11994529e-01 -2.28793025e-01 6.44756496e-01 -5.15122712e-01
-5.32427669e-01 1.70509040e-01 -1.26266867e-01 -7.50233114e-01
-3.30462426e-01 -5.25301337e-01 -8.07440996e-01 -4.92816389e-01
-4.59077656e-01 -3.08924671e-02 4.04172689e-01 4.39726502e-01
7.76510954e-01 6.90867424e-01 5.11991739e-01 -1.46903801e+00
-1.97662726e-01 -8.38055551e-01 -2.91225374e-01 4.28845882e-01
1.01046018e-01 -5.98416984e-01 -1.58873629e-02 4.94235575e-01] | [10.892383575439453, -0.5981190204620361] |
8ffdf098-110d-4679-a065-cb149a59b224 | multi-perspective-document-revision | null | null | https://aclanthology.org/2022.coling-1.535 | https://aclanthology.org/2022.coling-1.535.pdf | Multi-Perspective Document Revision | This paper presents a novel multi-perspective document revision task. In conventional studies on document revision, tasks such as grammatical error correction, sentence reordering, and discourse relation classification have been performed individually; however, these tasks simultaneously should be revised to improve the readability and clarity of a whole document. Thus, our study defines multi-perspective document revision as a task that simultaneously revises multiple perspectives. To model the task, we design a novel Japanese multi-perspective document revision dataset that simultaneously handles seven perspectives to improve the readability and clarity of a document. Although a large amount of data that simultaneously handles multiple perspectives is needed to model multi-perspective document revision elaborately, it is difficult to prepare such a large amount of this data. Therefore, our study offers a multi-perspective document revision modeling method that can use a limited amount of matched data (i.e., data for the multi-perspective document revision task) and external partially-matched data (e.g., data for the grammatical error correction task). Experiments using our created dataset demonstrate the effectiveness of using multiple partially-matched datasets to model the multi-perspective document revision task. | ['Ryo Masumura', 'Tomohiro Tanaka', 'Hiroshi Sato', 'Mana Ihori'] | null | null | null | null | coling-2022-10 | ['grammatical-error-correction', 'relation-classification'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.40564495e-01 3.46170276e-01 -1.26598954e-01 -5.74292362e-01
-6.79280937e-01 -2.95979649e-01 6.35693014e-01 4.57775533e-01
-4.87968475e-01 7.28445530e-01 5.76743960e-01 -4.41607624e-01
-4.30711359e-03 -5.15694916e-01 -5.21091521e-01 -1.09995835e-01
8.38423252e-01 4.62296635e-01 2.72545546e-01 -6.74490273e-01
8.19886684e-01 -2.91281372e-01 -1.39921522e+00 5.79229891e-01
1.46917963e+00 2.67658561e-01 8.65978479e-01 6.24530852e-01
-1.18000247e-01 4.37627047e-01 -1.07127297e+00 -5.37315488e-01
-1.22760378e-01 -6.08307064e-01 -1.06121325e+00 3.93300474e-01
4.23501849e-01 -4.77089733e-01 2.95949310e-01 9.65891004e-01
6.04062140e-01 1.64336398e-01 5.52086890e-01 -8.38407278e-01
-1.16814375e+00 9.04758215e-01 -3.55750620e-01 2.29660168e-01
7.28871644e-01 -5.60868442e-01 9.10384536e-01 -1.13446486e+00
7.57890403e-01 1.28718662e+00 5.47812939e-01 4.88921463e-01
-9.15240526e-01 -5.27845562e-01 7.80208886e-01 4.33073565e-02
-7.75091946e-01 -3.12913746e-01 1.02786076e+00 -3.13353449e-01
8.98718059e-01 5.18031895e-01 7.88497269e-01 1.05862224e+00
5.52046895e-01 6.86255693e-01 1.10620904e+00 -6.79144859e-01
-2.47287929e-01 4.10190165e-01 6.71631515e-01 5.28834641e-01
6.70400321e-01 -4.36312377e-01 -4.08054143e-01 1.40088975e-01
3.14700723e-01 -7.31357187e-03 -4.55400586e-01 1.82949193e-02
-1.32625496e+00 5.81695437e-01 -1.91703334e-01 3.21242958e-01
4.28809375e-02 -4.68337566e-01 5.10525465e-01 4.59017277e-01
7.41619587e-01 7.80152559e-01 -6.05331421e-01 -2.55960166e-01
-8.19477677e-01 3.95932496e-01 6.20780706e-01 1.36215305e+00
3.98876876e-01 -2.12162897e-01 -3.73063505e-01 1.00824690e+00
4.63324964e-01 4.97830123e-01 5.72770715e-01 -6.67286992e-01
1.12154448e+00 1.00599504e+00 1.84184417e-01 -1.21489739e+00
-5.87034285e-01 -5.54349184e-01 -8.24604452e-01 -3.92867267e-01
8.64080861e-02 5.92369586e-02 -6.30140245e-01 1.49719918e+00
8.58511031e-02 -9.95065689e-01 1.67765081e-01 6.72252357e-01
1.11028385e+00 5.78674316e-01 -3.99126142e-01 -6.97940886e-01
1.38184953e+00 -1.22583091e+00 -1.01427829e+00 -1.60978600e-01
1.09251082e+00 -9.77134526e-01 1.59775877e+00 5.80280960e-01
-1.19830620e+00 -7.29032636e-01 -1.01548266e+00 -5.68878055e-01
-1.49951786e-01 5.14408052e-01 1.63808137e-01 4.39584821e-01
-6.25257373e-01 1.06950998e-01 -5.72417796e-01 -4.27565843e-01
-2.40723103e-01 -2.23700523e-01 -3.75591159e-01 -1.56247482e-01
-1.24979198e+00 1.16396272e+00 3.99453551e-01 1.75131097e-01
-2.28700444e-01 -4.31484640e-01 -7.03255951e-01 -7.15250522e-02
4.12153155e-01 -8.62591386e-01 1.42884052e+00 -5.84758699e-01
-1.21617734e+00 5.66596448e-01 -3.43548298e-01 3.09930623e-01
3.88377100e-01 -4.05640364e-01 -4.16762918e-01 -3.45083237e-01
3.47346395e-01 2.08325237e-01 6.37874484e-01 -1.56787157e+00
-8.41893554e-01 -6.12946332e-01 3.02799016e-01 5.76990247e-01
-5.18022001e-01 1.94098607e-01 -5.98957360e-01 -1.11768675e+00
3.52283388e-01 -9.43996429e-01 2.04550698e-01 -5.60145378e-01
-5.36896229e-01 -1.76299885e-01 6.02813601e-01 -1.19106340e+00
1.89876246e+00 -1.83770907e+00 6.53251350e-01 -1.68370515e-01
3.93656105e-01 9.20472946e-03 -1.63011745e-01 5.06379962e-01
7.04386309e-02 6.78203285e-01 -3.49191815e-01 -7.87423611e-01
-1.73141584e-01 -6.13221116e-02 -6.32750243e-02 -1.47154212e-01
-2.70061404e-01 5.69577575e-01 -7.44568169e-01 -3.41738909e-01
-5.13607919e-01 -1.21606544e-01 -7.42714942e-01 3.76844928e-02
2.07351311e-03 3.84343624e-01 -2.67691880e-01 6.38334334e-01
6.72136009e-01 -1.01945490e-01 -5.26636131e-02 -2.31118545e-01
-5.06459512e-02 4.93660003e-01 -1.04443908e+00 1.65527785e+00
-4.62623447e-01 4.05766159e-01 -4.14782882e-01 -4.75477159e-01
1.16586494e+00 1.19834870e-01 3.60631980e-02 -7.54473031e-01
-1.20529249e-01 2.15490550e-01 5.05761318e-02 -7.40867138e-01
1.48285174e+00 -3.58690657e-02 -3.13393652e-01 9.91936684e-01
-4.36555535e-01 -2.51416028e-01 6.85489893e-01 5.40062189e-01
5.80030441e-01 -1.56036317e-01 5.67717671e-01 -1.06112741e-01
5.65989494e-01 6.04037531e-02 8.15120339e-01 7.40467370e-01
9.21431184e-02 5.36199689e-01 5.84769964e-01 -1.75743073e-01
-1.02863896e+00 -5.06707191e-01 1.88739840e-02 1.05683792e+00
3.10859412e-01 -1.11982691e+00 -5.08000433e-01 -9.29858088e-01
-1.63143054e-01 1.00202453e+00 -4.49209452e-01 -4.45780873e-01
-5.39991796e-01 -7.71284878e-01 3.85356516e-01 5.28503418e-01
6.19142890e-01 -7.32680321e-01 -3.97849619e-01 1.15506478e-01
-5.65376878e-01 -6.86430931e-01 -8.19445968e-01 -2.17929691e-01
-1.01782346e+00 -1.27240753e+00 -3.71915400e-01 -8.18075418e-01
1.06883216e+00 6.84583187e-01 8.92638743e-01 4.26378101e-01
5.21822214e-01 4.05233830e-01 -7.44796097e-01 -5.69326103e-01
-7.15802491e-01 1.59618095e-01 7.37076178e-02 -7.45342076e-01
4.25146222e-02 8.44115466e-02 3.47332619e-02 2.84310788e-01
-9.42916930e-01 5.90146482e-01 4.75006133e-01 1.13937604e+00
4.71980631e-01 1.42856479e-01 6.54379249e-01 -1.45276797e+00
1.65022242e+00 -1.70769319e-02 -1.66365370e-01 9.27323997e-01
-1.07582378e+00 -1.39760196e-01 3.92366081e-01 -3.51622671e-01
-1.56767571e+00 -8.23965311e-01 3.23305726e-01 5.42031407e-01
3.77663404e-01 1.25496936e+00 -2.19594032e-01 3.83620292e-01
3.90951186e-01 3.07728410e-01 -1.40783027e-01 -6.94056153e-01
2.35891268e-01 9.85704720e-01 1.59046009e-01 -5.86856425e-01
4.90855604e-01 -1.76228404e-01 -5.52497625e-01 -5.42101204e-01
-9.65761244e-01 -3.41553539e-02 -9.36137915e-01 -2.81745583e-01
4.62967277e-01 -7.86908388e-01 -2.18336824e-02 5.62407792e-01
-1.61290324e+00 1.59250349e-01 -2.20164340e-02 4.60988849e-01
-6.56290501e-02 7.56467640e-01 -4.85263199e-01 -5.64117968e-01
-4.33634341e-01 -1.33768535e+00 1.01682293e+00 2.60747690e-02
-6.67023003e-01 -9.53369617e-01 4.41757329e-02 7.74193168e-01
2.56996095e-01 -1.43063590e-01 1.43459809e+00 -5.36307752e-01
-1.37434900e-01 8.53625312e-03 -3.03116702e-02 4.68067050e-01
4.92631435e-01 2.12841377e-01 -2.17257753e-01 -4.36454237e-01
3.34020317e-01 -1.65242583e-01 8.13791037e-01 -6.30931631e-02
1.04130483e+00 -3.58041972e-01 -1.33448094e-01 1.50361195e-01
6.09913945e-01 5.77268124e-01 4.47699517e-01 6.43709958e-01
8.42820644e-01 5.91746986e-01 1.17692077e+00 5.51525772e-01
1.27510893e+00 6.19700491e-01 -8.80230069e-02 2.22911656e-01
-1.59710199e-01 -1.77415252e-01 3.48881781e-01 1.74615574e+00
1.17819645e-02 -7.32266247e-01 -9.35594201e-01 2.40164995e-01
-1.82344365e+00 -6.38005674e-01 -4.33558375e-01 1.84754527e+00
9.78218436e-01 2.33849958e-01 -4.08507526e-01 4.03958827e-01
6.63636565e-01 1.09734528e-01 -2.58089542e-01 -6.75649464e-01
-1.99204773e-01 -6.66584849e-01 -4.08645898e-01 6.03017390e-01
-5.96178710e-01 7.11482346e-01 6.05492115e+00 1.34703532e-01
-8.91546965e-01 4.00106125e-02 1.38089031e-01 9.03316587e-03
-1.00336337e+00 3.25800389e-01 -1.20372880e+00 5.75097501e-01
2.45018706e-01 -3.22889060e-01 1.48622304e-01 4.12008137e-01
4.13233519e-01 -5.11225581e-01 -1.17119169e+00 9.12958980e-01
8.57956409e-01 -9.61280286e-01 7.97716916e-01 -1.93409160e-01
8.06735754e-01 -8.29078376e-01 -1.86839223e-01 5.77769279e-01
-2.27723837e-01 -4.66760814e-01 9.24196064e-01 8.41252029e-01
4.83363152e-01 -5.50207675e-01 8.39054883e-01 7.86859512e-01
-6.95240617e-01 2.47535389e-03 -2.58056343e-01 -1.79472417e-01
4.83515590e-01 6.53759360e-01 -4.46680695e-01 7.57032514e-01
4.14524585e-01 1.09416676e+00 -1.25667059e+00 5.72968841e-01
-2.36872792e-01 8.46113712e-02 3.05286467e-01 6.89160749e-02
-4.78753746e-01 -2.85281867e-01 6.67238593e-01 8.17390919e-01
6.73224330e-01 2.98856497e-02 1.36533424e-01 5.39582193e-01
-2.21649155e-01 9.02394131e-02 -4.13366199e-01 -1.47489592e-01
6.81795061e-01 7.25977659e-01 -1.90826684e-01 -3.44058633e-01
-3.44437182e-01 7.82349765e-01 4.27750409e-01 2.83802301e-01
-6.10543787e-01 -4.64299321e-01 2.30666660e-02 -7.59970490e-03
-1.29994392e-01 -2.66599804e-01 -7.68010855e-01 -1.49505746e+00
4.93333489e-01 -1.14458954e+00 2.55410194e-01 -9.96371746e-01
-9.60403621e-01 5.22372544e-01 1.53897688e-01 -1.29262936e+00
1.34537861e-01 -1.40748639e-02 -4.70629483e-01 7.47493804e-01
-1.42763567e+00 -1.23710811e+00 -4.16945785e-01 2.25864395e-01
8.63217354e-01 -2.29117423e-01 6.75793707e-01 2.19563127e-01
-9.30216551e-01 6.07923865e-01 -6.34831116e-02 -4.06552225e-01
1.23976123e+00 -1.17374718e+00 1.33068666e-01 1.08947563e+00
-2.02127770e-01 1.24511778e+00 5.37571728e-01 -1.13886201e+00
-9.32333052e-01 -8.46859217e-01 1.62699664e+00 -6.00647390e-01
1.47817001e-01 -2.41668344e-01 -1.24286497e+00 1.01731896e+00
1.99246317e-01 -8.39385390e-01 9.18613791e-01 4.61683422e-01
-7.83255696e-03 -1.22521535e-01 -7.14253426e-01 8.13471735e-01
1.03722632e+00 -4.96516973e-01 -1.32571924e+00 3.04168761e-01
8.05750787e-01 -8.30863178e-01 -1.01300299e+00 5.10455310e-01
2.27033362e-01 -6.39761984e-01 2.99908221e-01 -6.75326645e-01
7.49115646e-01 -4.67540026e-01 1.30246645e-02 -1.78509390e+00
-3.51010680e-01 -1.75179824e-01 -2.12875634e-01 1.65937722e+00
6.30713820e-01 -6.88095927e-01 -6.87191337e-02 8.16678762e-01
-6.86652601e-01 -7.12488532e-01 -6.76539183e-01 -4.99626070e-01
-1.20918602e-01 -2.24012583e-01 7.69413650e-01 9.23584878e-01
5.94446898e-01 6.29737318e-01 -4.99024451e-01 -1.94570765e-01
-1.80417880e-01 2.16442123e-01 1.03194630e+00 -1.34908569e+00
4.38537002e-02 -2.54551619e-01 4.28390056e-01 -1.30057931e+00
6.32568598e-02 -9.42209184e-01 -2.63757408e-01 -2.05764079e+00
6.09712541e-01 -4.47167665e-01 1.03323698e-01 2.18046322e-01
-6.10329807e-01 -5.73301375e-01 2.90049672e-01 3.92040402e-01
-5.82742333e-01 8.92349064e-01 1.90163827e+00 -2.20374405e-01
-4.07202542e-01 -8.10921949e-04 -1.52760684e+00 6.11615360e-01
6.69588208e-01 -4.96815473e-01 -6.59468710e-01 -9.74640667e-01
7.12124586e-01 2.37123415e-01 -1.02409847e-01 -4.41245794e-01
4.25884336e-01 -2.36716181e-01 2.24416465e-01 -8.89474094e-01
1.82678938e-01 -4.65779871e-01 -1.15701683e-01 2.35065609e-01
-7.29791343e-01 8.14059258e-01 -6.64852709e-02 6.07686341e-01
-4.02187496e-01 -4.25601035e-01 2.66956091e-01 -2.27729961e-01
-4.46333408e-01 -2.26408824e-01 -5.45215130e-01 1.24008164e-01
9.20652330e-01 -3.46267492e-01 -9.95000362e-01 -2.66605020e-01
-5.68439186e-01 3.45715314e-01 6.56186223e-01 9.13188398e-01
7.87831903e-01 -1.26601624e+00 -6.61579251e-01 1.71419740e-01
4.51873422e-01 3.08800638e-01 5.65550737e-02 7.82235861e-01
-2.79694587e-01 5.19759834e-01 -1.91165745e-01 -3.56734335e-01
-1.65472984e+00 2.36124068e-01 -1.06964894e-01 -3.98133695e-01
-5.97665608e-01 5.43090284e-01 -1.89442843e-01 -8.20216477e-01
1.24373898e-01 -8.55781615e-01 -6.54622316e-01 3.99141937e-01
4.36196655e-01 5.16077995e-01 1.68159455e-01 -4.77861226e-01
-1.08618975e-01 5.84083200e-01 -6.51879966e-01 -7.30365654e-03
1.12499869e+00 -8.86488378e-01 -5.89494646e-01 9.06953990e-01
6.66521370e-01 4.25229162e-01 -5.35477340e-01 -2.76184380e-01
1.32305413e-01 -5.22180974e-01 -1.07472412e-01 -1.05634475e+00
-4.56723332e-01 6.47033155e-01 -1.24245912e-01 -1.12254180e-01
9.84159589e-01 -3.53726327e-01 7.27896214e-01 7.95345604e-01
3.19986761e-01 -1.56909084e+00 3.27607960e-01 9.97313201e-01
1.60518575e+00 -1.36671937e+00 3.64200562e-01 -3.88786435e-01
-9.54201996e-01 1.26481605e+00 1.16813350e+00 6.43801570e-01
4.07979876e-01 -2.25800008e-01 -6.33618385e-02 -2.80230224e-01
-8.85182381e-01 6.50111556e-01 4.35355186e-01 3.89160752e-01
8.06090176e-01 6.71426803e-02 -1.04610038e+00 1.08234394e+00
-4.25094873e-01 -2.38154173e-01 1.18747926e+00 1.22620630e+00
-4.44011509e-01 -1.34486628e+00 -4.00323898e-01 7.83005595e-01
1.81133837e-01 -7.69997612e-02 -6.97343409e-01 7.39074230e-01
-2.19549146e-02 1.08523250e+00 -2.82404244e-01 -4.70721990e-01
7.25627661e-01 -1.32908523e-01 3.60459089e-01 -1.09698808e+00
-7.71009922e-01 -1.89521238e-01 4.80690300e-01 -3.46471295e-02
-4.33503926e-01 -7.15632439e-01 -1.15094495e+00 -3.24241936e-01
-5.65708280e-01 8.76967236e-02 5.25716126e-01 1.21463597e+00
3.13966602e-01 8.93420815e-01 2.89038479e-01 -2.11189270e-01
-6.29081488e-01 -1.43126512e+00 -5.83743095e-01 3.33265752e-01
2.95925774e-02 -6.44092619e-01 -6.64078444e-02 2.54299883e-02] | [12.00517463684082, 9.351727485656738] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.