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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
a3fd1c66-ca89-4cc6-bb56-4c968450660c | learning-robust-feature-representations-for | 2005.12466 | null | https://arxiv.org/abs/2005.12466v1 | https://arxiv.org/pdf/2005.12466v1.pdf | Learning Robust Feature Representations for Scene Text Detection | Scene text detection based on deep neural networks have progressed substantially over the past years. However, previous state-of-the-art methods may still fall short when dealing with challenging public benchmarks because the performances of algorithm are determined by the robust features extraction and components in network architecture. To address this issue, we will present a network architecture derived from the loss to maximize conditional log-likelihood by optimizing the lower bound with a proper approximate posterior that has shown impressive performance in several generative models. In addition, by extending the layer of latent variables to multiple layers, the network is able to learn robust features on scale with no task-specific regularization or data augmentation. We provide a detailed analysis and show the results on three public benchmark datasets to confirm the efficiency and reliability of the proposed algorithm. In experiments, the proposed algorithm significantly outperforms state-of-the-art methods in terms of both recall and precision. Specifically, it achieves an H-mean of 95.12 and 96.78 on ICDAR 2011 and ICDAR 2013, respectively. | ['Taejang Park', 'Sihwan Kim'] | 2020-05-26 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [ 3.14891860e-02 -1.36288688e-01 -7.04362392e-02 -6.99417353e-01
-1.06463695e+00 -1.81041397e-02 6.96771324e-01 -1.46696931e-02
-6.51951253e-01 6.56315863e-01 1.82776637e-02 1.67182133e-01
-1.88807011e-01 -7.39890814e-01 -7.37950027e-01 -8.31488073e-01
1.34582773e-01 3.43599528e-01 1.01752348e-01 2.21268654e-01
8.75532106e-02 2.30821431e-01 -1.42379773e+00 3.23202610e-01
7.64027953e-01 1.24470401e+00 -2.66833859e-03 3.43838096e-01
1.82567928e-02 6.93289995e-01 -6.68241203e-01 -5.73192716e-01
1.48791090e-01 -4.15787548e-02 -3.44286293e-01 3.87568027e-02
3.85233045e-01 -6.00615501e-01 -5.01457870e-01 9.86157954e-01
5.17862737e-01 -6.54795915e-02 7.66770422e-01 -1.09154236e+00
-7.37513185e-01 6.80391312e-01 -6.29452705e-01 -1.43267304e-01
-1.42683208e-01 -1.84522420e-01 1.09182692e+00 -1.06624866e+00
3.40194702e-01 1.27431023e+00 7.47338831e-01 2.58712143e-01
-1.03935373e+00 -9.01147962e-01 2.60527313e-01 6.41961545e-02
-1.67715836e+00 -2.97164738e-01 6.38470411e-01 -4.45077956e-01
9.56783891e-01 -4.51672897e-02 7.63972178e-02 1.30425060e+00
1.60250261e-01 1.13127995e+00 8.20363402e-01 -3.48508030e-01
8.16598088e-02 3.85470986e-01 3.94269794e-01 8.16440344e-01
5.84208190e-01 -2.84996718e-01 -6.36026084e-01 -4.25070673e-02
3.57617885e-01 1.45812601e-01 -8.88212621e-02 -2.24177435e-01
-9.38536704e-01 1.05173004e+00 3.88779461e-01 2.72432059e-01
-3.00278485e-01 5.74174076e-02 4.52674121e-01 -2.52173245e-01
5.82136869e-01 -2.25716624e-02 -3.26935023e-01 1.82325393e-01
-1.31776953e+00 2.24306993e-02 7.16730416e-01 8.61846864e-01
4.94068444e-01 2.54749358e-01 -4.55369145e-01 1.08080184e+00
4.84950364e-01 5.56989670e-01 4.76939321e-01 -3.71651053e-01
5.99098802e-01 6.97865486e-01 -1.77582666e-01 -9.64878917e-01
-4.16564405e-01 -8.37173462e-01 -1.37516809e+00 -2.00885043e-01
1.47565678e-01 -2.09398285e-01 -1.14452541e+00 1.74432671e+00
9.22735184e-02 -1.48862034e-01 6.77448958e-02 5.68809927e-01
9.57626402e-01 8.62830222e-01 1.43318236e-01 -9.17076766e-02
1.21575284e+00 -8.85451436e-01 -9.21203971e-01 -4.06793773e-01
4.72335845e-01 -6.78753018e-01 9.86418188e-01 5.47177553e-01
-6.65292323e-01 -6.34811938e-01 -1.30469573e+00 1.25308827e-01
-3.66064072e-01 9.77140427e-01 5.28269827e-01 6.74722612e-01
-7.31081009e-01 4.02674377e-01 -9.41256583e-01 -4.65329528e-01
5.78457534e-01 3.97140145e-01 -1.97887495e-01 5.85799711e-03
-1.18836904e+00 5.15926778e-01 6.22049928e-01 2.25628212e-01
-9.56203997e-01 -2.57238567e-01 -5.71207583e-01 3.63564938e-01
3.45176905e-01 -6.77127898e-01 1.06929374e+00 -6.38202667e-01
-1.42657113e+00 6.10092223e-01 1.07409663e-01 -7.74494767e-01
7.40967393e-01 -8.40596437e-01 -2.48128027e-01 5.13024954e-03
1.05224559e-02 6.40631914e-01 7.55632818e-01 -1.03873312e+00
-7.30002105e-01 -2.54820287e-01 -2.27475867e-01 -5.12486659e-02
-9.10008729e-01 4.78106886e-02 -8.25713813e-01 -6.72087133e-01
1.63992912e-01 -7.83274770e-01 6.22367300e-03 -9.10859630e-02
-6.97211683e-01 -3.92920464e-01 8.28193367e-01 -6.61558807e-01
1.32507443e+00 -2.25819325e+00 -3.65527183e-01 -3.15957218e-02
9.71057490e-02 2.76095420e-01 1.42390266e-01 3.17526072e-01
2.36112043e-01 9.73209292e-02 -1.87652737e-01 -6.34641528e-01
2.75681138e-01 -5.81485219e-02 -4.10350502e-01 7.64579117e-01
9.97274220e-02 6.55200660e-01 -3.19066554e-01 -5.31761110e-01
1.23815015e-01 8.31893504e-01 -3.75705719e-01 2.62713224e-01
-7.72847086e-02 -2.19001159e-01 -4.79486585e-01 4.55416173e-01
6.45786583e-01 -5.85624814e-01 7.75046498e-02 -1.33046687e-01
1.23255506e-01 2.01146424e-01 -1.23210359e+00 1.49854088e+00
-1.66737303e-01 7.82761812e-01 -1.27656952e-01 -9.24498022e-01
9.77806211e-01 2.97597587e-01 2.80229628e-01 -5.19461632e-01
3.68374676e-01 4.82371487e-02 -3.31867427e-01 -3.10385376e-01
4.44634140e-01 1.91241816e-01 -7.10141659e-02 2.82113969e-01
8.48567188e-02 4.85828936e-01 3.24125588e-01 1.76618099e-01
8.01602066e-01 -6.79137781e-02 1.71078756e-01 -1.14593215e-01
5.57697713e-01 -3.94137651e-01 4.63479966e-01 1.04319239e+00
4.43171524e-03 6.03135407e-01 4.29780036e-01 -4.52682942e-01
-8.84881675e-01 -1.08512378e+00 -3.43931526e-01 1.09703553e+00
-1.76467925e-01 -4.28641498e-01 -8.91702652e-01 -7.29687631e-01
-1.88176885e-01 7.34524190e-01 -6.49398148e-01 -1.35941759e-01
-4.24474657e-01 -1.35388327e+00 7.58489907e-01 7.33074903e-01
9.13487077e-01 -8.16832006e-01 -4.72075641e-01 1.00765198e-01
-2.14032486e-01 -1.44669461e+00 -7.40767643e-02 2.30071783e-01
-8.08058798e-01 -7.33849764e-01 -7.65791953e-01 -6.58020318e-01
5.74330628e-01 1.18140623e-01 9.29093063e-01 -2.07357824e-01
-2.79511362e-01 -3.55602875e-02 -3.23462993e-01 -5.23128152e-01
-2.85859734e-01 4.02320683e-01 3.92166376e-02 1.97261959e-01
4.60585415e-01 -2.49246910e-01 -5.20735323e-01 1.40748754e-01
-1.08372414e+00 9.74455252e-02 8.76447558e-01 8.86100829e-01
5.93631804e-01 2.64163047e-01 4.77214634e-01 -8.53381395e-01
6.39457703e-01 -3.01065385e-01 -7.35485196e-01 3.92335862e-01
-8.04256976e-01 1.93912521e-01 5.72462976e-01 -3.25083375e-01
-1.11718452e+00 2.30505437e-01 -1.12755977e-01 -1.68138742e-01
-2.08159357e-01 5.03758967e-01 -1.00350194e-01 5.46820045e-01
6.71784461e-01 4.08416152e-01 -4.44705546e-01 -6.01596594e-01
2.10933983e-01 1.00367856e+00 4.38914150e-01 -3.23398709e-01
7.13486671e-01 6.04554832e-01 -1.50215566e-01 -7.78085828e-01
-1.23459256e+00 -4.14274007e-01 -4.87274677e-01 1.10790879e-01
9.42417860e-01 -1.14098036e+00 -5.38834453e-01 7.51033068e-01
-1.10508645e+00 8.98414031e-02 1.35101005e-01 5.91695130e-01
-2.58045197e-01 3.52986813e-01 -7.08099663e-01 -9.70460474e-01
-7.66586363e-01 -1.06961334e+00 1.02804101e+00 1.66376442e-01
2.84478981e-02 -6.64206147e-01 -7.81022385e-02 1.59196526e-01
4.25525665e-01 1.92450047e-01 8.65337789e-01 -9.82214510e-01
-5.63900471e-01 -6.96461141e-01 -5.07550120e-01 6.24697268e-01
5.66269346e-02 1.08169027e-01 -1.38284659e+00 -4.43524152e-01
3.08966823e-02 -3.90486568e-01 1.24027336e+00 5.53131342e-01
1.30222762e+00 -3.18724424e-01 -3.32275599e-01 6.86469674e-01
1.43388188e+00 -1.25364244e-01 4.92195725e-01 4.44361955e-01
5.24355948e-01 3.97508472e-01 3.56139630e-01 6.14462316e-01
2.18091786e-01 4.63212013e-01 4.29292947e-01 -9.90003049e-02
4.42792429e-03 -2.69189656e-01 3.53165507e-01 7.03197777e-01
3.43142867e-01 -5.17410338e-01 -9.63275790e-01 3.84816706e-01
-1.93649948e+00 -6.99062884e-01 -3.27048153e-02 2.14887404e+00
6.56093657e-01 4.62728709e-01 -5.69019690e-02 1.25123009e-01
7.74229169e-01 2.33913913e-01 -4.87592548e-01 9.65005606e-02
-2.73577988e-01 -1.74988806e-01 5.47853112e-01 1.38642728e-01
-1.51805174e+00 9.02331114e-01 6.24415255e+00 1.15166557e+00
-1.05290270e+00 1.24551982e-01 8.29900444e-01 -1.23403236e-01
2.83674061e-01 -4.23773319e-01 -1.36501646e+00 2.72880107e-01
8.83883417e-01 5.29698133e-02 -6.60912460e-03 1.15663517e+00
-8.75331089e-02 7.18598068e-02 -1.07796848e+00 1.04023373e+00
2.87258446e-01 -1.20040178e+00 2.02011585e-01 5.18998876e-02
7.64805198e-01 2.91657150e-01 3.60985160e-01 5.47462344e-01
1.38130009e-01 -1.18471968e+00 6.43981218e-01 3.70492339e-01
5.99641681e-01 -8.33792329e-01 1.00517774e+00 6.50037050e-01
-8.90779376e-01 -7.31147155e-02 -5.57313085e-01 2.50653327e-01
-9.63706970e-02 7.31367528e-01 -9.43407059e-01 3.63053948e-01
8.74014318e-01 4.57405716e-01 -6.37973487e-01 1.01425946e+00
-3.61258715e-01 8.49136293e-01 -5.96955538e-01 -2.67170399e-01
3.67616445e-01 1.26486480e-01 1.89874873e-01 1.64987886e+00
3.14646870e-01 -4.06510144e-01 1.36873975e-01 8.98127556e-01
-5.01375139e-01 3.90134752e-01 -3.61102045e-01 -5.56994937e-02
2.74073660e-01 1.29148567e+00 -7.85653353e-01 -4.02407318e-01
-4.21161294e-01 7.36090302e-01 3.80046844e-01 3.73705119e-01
-1.04884887e+00 -4.37653601e-01 1.96577817e-01 -9.28537324e-02
3.95809859e-01 -1.32980675e-01 -3.36827815e-01 -1.11255765e+00
2.56999224e-01 -7.75997758e-01 4.84099448e-01 -4.60528582e-01
-1.32423580e+00 9.08287346e-01 -1.84208937e-02 -1.07841027e+00
-1.57703295e-01 -7.50727654e-01 -2.24294350e-01 6.41939104e-01
-1.36298454e+00 -1.17722404e+00 -3.15087110e-01 3.67172182e-01
6.18296981e-01 -4.23977792e-01 7.19397187e-01 3.39898258e-01
-9.33290482e-01 1.02882493e+00 6.07016504e-01 4.19095099e-01
8.44996989e-01 -9.00752783e-01 3.81167918e-01 1.02861094e+00
3.04454774e-01 5.58354080e-01 6.14795029e-01 -4.76560712e-01
-9.94360328e-01 -1.11627841e+00 7.39692569e-01 -2.19891042e-01
4.29089338e-01 -6.29816651e-01 -8.57191682e-01 5.35010397e-01
1.68608323e-01 -7.81517923e-02 5.00990510e-01 2.53914446e-01
-5.08034170e-01 -2.60385543e-01 -9.36236620e-01 3.54715824e-01
6.07341051e-01 -3.14736158e-01 -3.85208309e-01 4.25140679e-01
5.35740256e-01 -2.92734623e-01 -4.51870024e-01 6.47293389e-01
6.65523410e-01 -9.82072532e-01 7.76823163e-01 -4.19966489e-01
4.27757472e-01 -2.75721960e-02 -5.22809684e-01 -6.30372703e-01
-3.42459679e-01 -2.99115151e-01 -2.28973389e-01 1.48580849e+00
4.72355038e-01 -4.56699878e-01 8.93158674e-01 4.43597198e-01
1.94890261e-01 -9.78653014e-01 -7.61534691e-01 -7.78331339e-01
1.18070550e-01 -5.62939763e-01 3.87098670e-01 5.73488653e-01
-6.34723961e-01 4.58211452e-01 -7.41691530e-01 2.41888419e-01
6.84472919e-01 2.65494809e-02 8.95125806e-01 -1.30058789e+00
-3.19014996e-01 -3.66963863e-01 -1.89063177e-01 -1.21500599e+00
1.28244489e-01 -6.86827660e-01 1.55244142e-01 -1.68676674e+00
6.55941665e-01 -2.27096095e-03 -5.53530276e-01 4.64370996e-01
-2.47087181e-01 1.86542660e-01 1.49383992e-01 2.85299480e-01
-8.41843545e-01 8.38659286e-01 6.74706101e-01 -2.20385954e-01
-8.13379735e-02 -9.67802946e-03 -6.33603692e-01 1.03445625e+00
9.23869848e-01 -8.17243874e-01 -2.50694513e-01 -4.85414922e-01
2.40302160e-01 -4.97342318e-01 1.86449587e-01 -1.09843087e+00
3.11890721e-01 1.54044241e-01 7.01210558e-01 -9.45825875e-01
3.54927242e-01 -7.31041372e-01 -8.43632519e-02 4.71533418e-01
-4.40462559e-01 -2.45163739e-01 7.26749152e-02 6.72289789e-01
-2.90570170e-01 -1.49658024e-01 9.82444763e-01 3.33439171e-01
-3.48325193e-01 2.08611190e-01 -2.63813198e-01 -8.00912231e-02
7.60864258e-01 2.07474921e-02 -4.51227933e-01 -3.88208687e-01
-3.02187204e-01 2.31931135e-02 1.47883845e-02 4.75976884e-01
5.48986733e-01 -1.15547109e+00 -8.77237380e-01 2.70513833e-01
2.17048705e-01 4.18082345e-03 2.15229765e-01 7.08151162e-01
-4.56158668e-01 7.55087912e-01 1.54912800e-01 -7.68664360e-01
-1.14624774e+00 3.61810267e-01 2.31194347e-01 -7.26992011e-01
-6.12055361e-01 8.04364264e-01 5.17355978e-01 -2.25348368e-01
8.21986258e-01 -1.30373240e-01 -1.12883039e-01 -4.28177491e-02
4.43511754e-01 2.92701304e-01 1.76789179e-01 -5.41567862e-01
-2.80580938e-01 3.29373956e-01 -4.10902530e-01 1.62662506e-01
1.48649526e+00 -8.31109192e-03 2.69228131e-01 2.60212183e-01
1.20498443e+00 -2.69462079e-01 -1.24808323e+00 -5.04777849e-01
-1.11831255e-01 -2.11928144e-01 2.57302165e-01 -9.54002261e-01
-1.02024400e+00 1.02317679e+00 8.68601799e-01 1.13412157e-01
1.05858088e+00 -1.11885145e-01 7.92735755e-01 7.79703021e-01
-5.17754480e-02 -1.21002316e+00 1.05093941e-01 6.19192839e-01
7.88256884e-01 -1.26538682e+00 1.65208519e-01 -2.36191854e-01
-3.93219441e-01 9.95006561e-01 6.57911003e-01 -2.19248474e-01
7.08032370e-01 2.95001268e-01 4.50215787e-02 -1.00389645e-01
-6.54406607e-01 7.02235699e-02 4.98957425e-01 1.06108204e-01
6.29314065e-01 -1.23790450e-01 -2.47437239e-01 8.41335416e-01
-1.44238085e-01 -1.42257199e-01 9.28586796e-02 6.13733709e-01
-5.41510463e-01 -7.62834609e-01 -2.62906522e-01 5.22391975e-01
-9.03771698e-01 -1.40778154e-01 -3.14016849e-01 1.03154230e+00
-2.03621134e-01 8.69101405e-01 -2.70913750e-01 -2.92262822e-01
1.71695411e-01 1.54919446e-01 9.46523249e-02 -3.55777591e-01
-2.55239069e-01 2.84700841e-01 -7.58459866e-02 -2.37435997e-01
-3.40166062e-01 -4.22273040e-01 -1.08067787e+00 -1.53151423e-01
-6.14056885e-01 -1.02158137e-01 8.54815066e-01 8.85560095e-01
3.10696423e-01 5.50593257e-01 3.46298665e-01 -4.91053224e-01
-9.65611994e-01 -1.31378806e+00 -5.35183549e-01 2.79893190e-01
1.33042827e-01 -5.84412992e-01 -3.86314690e-01 1.35522291e-01] | [9.632173538208008, 2.955577850341797] |
21377bda-e981-41a8-957f-ec866c7ec5f2 | a-shallow-triple-stream-three-dimensional-cnn | 1902.03634 | null | https://arxiv.org/abs/1902.03634v2 | https://arxiv.org/pdf/1902.03634v2.pdf | Shallow Triple Stream Three-dimensional CNN (STSTNet) for Micro-expression Recognition | In the recent year, state-of-the-art for facial micro-expression recognition have been significantly advanced by deep neural networks. The robustness of deep learning has yielded promising performance beyond that of traditional handcrafted approaches. Most works in literature emphasized on increasing the depth of networks and employing highly complex objective functions to learn more features. In this paper, we design a Shallow Triple Stream Three-dimensional CNN (STSTNet) that is computationally light whilst capable of extracting discriminative high level features and details of micro-expressions. The network learns from three optical flow features (i.e., optical strain, horizontal and vertical optical flow fields) computed based on the onset and apex frames of each video. Our experimental results demonstrate the effectiveness of the proposed STSTNet, which obtained an unweighted average recall rate of 0.7605 and unweighted F1-score of 0.7353 on the composite database consisting of 442 samples from the SMIC, CASME II and SAMM databases. | ['Yen-Chang Huang', 'Huai-Qian Khor', 'Sze-Teng Liong', 'John See', 'Y. S. Gan'] | 2019-02-10 | null | null | null | null | ['micro-expression-recognition'] | ['computer-vision'] | [ 1.74799666e-03 -3.06028187e-01 -1.50663361e-01 -7.33279586e-01
-2.29966626e-01 2.48630494e-02 4.41128314e-01 -4.84756321e-01
-5.52770674e-01 6.33974195e-01 9.11773592e-02 2.41597354e-01
-5.99874035e-02 -4.62129056e-01 -3.06883126e-01 -8.12907875e-01
-3.91308963e-01 -3.22955519e-01 -3.19157809e-01 -3.61365974e-01
2.84739166e-01 9.07564104e-01 -1.86379683e+00 2.46975034e-01
4.50980097e-01 1.71304071e+00 -6.50918663e-01 7.18078732e-01
2.91390456e-02 1.06574047e+00 -5.28078794e-01 -7.04850078e-01
1.39397338e-01 -3.44692051e-01 -6.06422305e-01 1.37585595e-01
7.42706299e-01 -5.52259743e-01 -4.83920962e-01 7.47786880e-01
7.67089784e-01 1.69607937e-01 4.65883851e-01 -1.55455697e+00
-4.47563678e-01 -2.71800756e-01 -7.33819723e-01 4.56629574e-01
1.91323027e-01 1.14718050e-01 8.99331391e-01 -1.09414804e+00
6.36050344e-01 1.26343966e+00 8.43953788e-01 6.25752211e-01
-8.92766893e-01 -1.01142216e+00 -3.33164722e-01 5.25521338e-01
-1.36230874e+00 -6.95073128e-01 9.35365558e-01 -4.50659305e-01
9.32126105e-01 -9.40596387e-02 7.85305619e-01 1.14957607e+00
1.56081036e-01 7.09342480e-01 1.14230490e+00 -2.56625861e-01
2.89954748e-02 1.45818032e-02 -1.62194192e-01 1.18441045e+00
1.85092650e-02 2.00496837e-01 -1.02144372e+00 3.60647030e-02
8.50641489e-01 -2.83293456e-01 -1.86439138e-03 -1.48777485e-01
-6.48709655e-01 8.58109772e-01 2.18761891e-01 3.66552353e-01
-5.62747777e-01 4.49047983e-02 7.25631058e-01 2.67867565e-01
6.66930318e-01 6.29806891e-02 -5.27696908e-01 -6.33237660e-01
-1.01600599e+00 1.82449639e-01 6.55881107e-01 6.85821533e-01
8.52665246e-01 4.40810591e-01 -5.09979948e-02 6.94453835e-01
1.16139591e-01 3.76700848e-01 5.22232890e-01 -1.23982799e+00
1.07243145e-02 6.17068052e-01 -6.60987347e-02 -1.50993788e+00
-4.71089691e-01 -1.93951860e-01 -1.09666467e+00 3.80228490e-01
2.37057164e-01 -5.01228094e-01 -7.09711194e-01 1.81341290e+00
2.34961793e-01 5.03348351e-01 2.69771833e-02 1.01683331e+00
8.28763068e-01 3.78756613e-01 -6.24750853e-02 -3.05900782e-01
9.45399523e-01 -7.99622476e-01 -1.00220549e+00 2.76467592e-01
5.39507985e-01 -5.78703821e-01 7.02124774e-01 5.71768105e-01
-1.03073895e+00 -6.95981264e-01 -1.06691372e+00 -3.90244275e-02
-2.66228318e-01 2.65365481e-01 8.26845646e-01 7.69866049e-01
-1.12353742e+00 6.88006520e-01 -8.36453974e-01 -6.58558980e-02
8.67960691e-01 6.25594735e-01 -7.06469715e-01 2.31695414e-01
-1.10121119e+00 7.10743666e-01 -8.60074759e-02 4.40713137e-01
-7.40559220e-01 -7.13101268e-01 -8.32420826e-01 -8.03450793e-02
-6.31640479e-02 -3.06607306e-01 1.09783673e+00 -1.56317699e+00
-2.04529905e+00 1.06047916e+00 -1.86234742e-01 -2.64240652e-01
3.93825710e-01 -3.65845531e-01 -5.63676059e-01 5.41802466e-01
-2.92278081e-01 6.44694209e-01 9.83530343e-01 -7.66058147e-01
-4.92250293e-01 -5.06962240e-01 -1.89627230e-01 -1.10432021e-01
-6.69680119e-01 3.75557065e-01 -9.90567077e-03 -4.85627472e-01
-3.34197789e-01 -8.16278279e-01 8.30279812e-02 4.88327026e-01
1.68166310e-02 -6.90900981e-02 1.18514705e+00 -4.99858528e-01
1.06398034e+00 -2.18179822e+00 2.85919476e-03 4.95929793e-02
2.81887144e-01 7.93906331e-01 -2.09497705e-01 2.49220319e-02
-1.96390852e-01 5.00402693e-03 9.39179733e-02 -3.97797942e-01
-1.51057571e-01 2.58603811e-01 1.79781452e-01 6.92691982e-01
4.26902890e-01 7.64143646e-01 -8.09568405e-01 -4.99849856e-01
3.00941586e-01 9.52748597e-01 -6.19580507e-01 4.74508345e-01
2.64257967e-01 3.87816876e-01 -3.45642686e-01 7.49384880e-01
7.04671621e-01 -5.19204810e-02 -1.45030215e-01 -2.63786256e-01
-6.91018254e-02 -2.61389315e-01 -8.75423968e-01 1.67061794e+00
-4.29983884e-01 1.17019701e+00 1.16478391e-01 -9.76060510e-01
1.18493187e+00 5.34275234e-01 8.61826658e-01 -8.36667776e-01
6.13133073e-01 1.96880370e-01 -1.23660803e-01 -8.97624612e-01
3.62363130e-01 -2.11487859e-01 3.50600600e-01 1.62137479e-01
5.72528064e-01 2.69072562e-01 -2.07573120e-02 -2.74725348e-01
7.40867496e-01 1.22349694e-01 3.86864573e-01 -9.81494188e-02
9.09244716e-01 -5.80630422e-01 8.01795185e-01 2.14073360e-01
-6.70720816e-01 2.74771273e-01 6.86058700e-01 -7.93419242e-01
-8.65659714e-01 -6.82428598e-01 -1.22846000e-01 1.04740405e+00
-3.08902472e-01 -3.33502948e-01 -8.19612801e-01 -4.10309851e-01
-9.25592408e-02 -4.08280194e-02 -8.43723476e-01 -2.39808947e-01
-6.57395959e-01 -5.34265935e-01 9.55341697e-01 5.65836787e-01
9.59093392e-01 -1.11060429e+00 -7.43435919e-01 3.49498354e-02
-1.92675355e-03 -1.55379224e+00 -1.35549903e-01 -4.16653603e-01
-8.53672385e-01 -9.39688921e-01 -6.08732045e-01 -6.03914499e-01
4.69189256e-01 -9.31531563e-02 8.90209436e-01 1.35993451e-01
-6.00089192e-01 5.72255068e-02 -3.56260568e-01 -3.94497693e-01
7.52706379e-02 -2.13803396e-01 1.84550419e-01 7.89243817e-01
6.51357234e-01 -7.78196454e-01 -6.37952983e-01 1.70866922e-01
-6.86556756e-01 -2.28584513e-01 5.71585774e-01 9.34091032e-01
2.99503028e-01 -3.52173477e-01 4.32648331e-01 -4.92625654e-01
5.08116126e-01 -2.00009003e-01 -4.79121655e-01 -1.88095197e-01
-4.90727931e-01 -2.59419233e-01 5.48472166e-01 -3.63071322e-01
-1.09568620e+00 1.41611937e-02 -2.83932894e-01 -6.54999614e-01
-2.53689080e-01 2.76396185e-01 1.39874056e-01 -5.63150942e-01
5.02447903e-01 8.54475275e-02 3.75995904e-01 -1.15489783e-02
1.19805500e-01 7.43432522e-01 6.39477849e-01 -5.47551870e-01
4.12457526e-01 6.80550873e-01 4.70211297e-01 -1.22539198e+00
-8.40922236e-01 -3.74614179e-01 -9.27075028e-01 -6.33892059e-01
8.15997541e-01 -8.04740250e-01 -1.17844212e+00 1.09829712e+00
-1.03748512e+00 4.61150147e-03 8.88657272e-02 4.10606533e-01
-6.02117419e-01 2.90846288e-01 -6.08410656e-01 -8.78864646e-01
-5.18791020e-01 -1.02026510e+00 1.08204269e+00 4.78796363e-01
-1.26822025e-01 -1.05843747e+00 1.97582077e-02 2.36703113e-01
6.30422294e-01 8.66670489e-01 4.02777880e-01 -2.34100446e-01
-1.70810923e-01 -3.32811892e-01 -3.96866441e-01 6.37412906e-01
2.21258968e-01 3.88235927e-01 -1.34087944e+00 -1.76125556e-01
-9.47879329e-02 -6.22722268e-01 5.71209133e-01 3.78995121e-01
1.33257854e+00 -2.84004837e-01 1.96870565e-01 1.03799391e+00
1.34492826e+00 1.46819353e-01 7.72707283e-01 3.66522133e-01
5.99601328e-01 5.43116331e-01 3.89489830e-01 9.19653296e-01
1.10400595e-01 5.97044945e-01 4.82604474e-01 -2.32680246e-01
1.21999972e-01 1.58188641e-01 2.69954175e-01 7.00764298e-01
-6.74633861e-01 1.36504352e-01 -5.76863527e-01 4.86843824e-01
-1.57906449e+00 -1.08424056e+00 2.08921462e-01 1.67152178e+00
7.40565479e-01 -1.65053189e-01 1.21307597e-01 2.63780683e-01
4.16575879e-01 4.11481082e-01 -4.71562684e-01 -6.69250846e-01
-1.82944775e-01 6.34771347e-01 8.16249400e-02 2.71093637e-01
-1.04614305e+00 8.95712733e-01 6.00613451e+00 3.59683216e-01
-1.63327885e+00 -2.00790808e-01 6.85258269e-01 -2.16990843e-01
4.53296125e-01 -5.87222755e-01 -6.78896427e-01 3.14405501e-01
1.18728924e+00 -1.08319983e-01 1.20808721e-01 7.69053400e-01
4.09286052e-01 6.28060028e-02 -7.62233973e-01 1.27417314e+00
3.29240084e-01 -1.25338638e+00 -1.71421170e-02 3.47735584e-02
7.86181927e-01 -2.45704144e-01 3.06027979e-01 1.98277071e-01
-1.49723783e-01 -1.31566525e+00 2.75140107e-01 6.88397884e-01
9.20108259e-01 -9.65793192e-01 1.02879143e+00 -4.16775458e-02
-1.02404499e+00 -1.75250724e-01 -1.29031345e-01 -2.62874573e-01
-6.88000768e-02 3.52235258e-01 -5.38819671e-01 4.20778513e-01
8.16192567e-01 1.10316420e+00 -2.98386723e-01 7.00897634e-01
-5.39640151e-02 5.62739313e-01 -1.38495550e-01 -1.71939567e-01
4.72834229e-01 -2.63476551e-01 2.71332592e-01 1.40763831e+00
2.37588167e-01 2.11223900e-01 -3.76378715e-01 4.04061884e-01
-1.78992912e-01 1.47681981e-01 -5.34509718e-01 -1.22754000e-01
6.27392381e-02 1.61547470e+00 1.05596948e-02 -1.90078706e-01
-5.32365441e-01 7.56084323e-01 2.96808004e-01 2.50661552e-01
-8.11347783e-01 -5.39573431e-01 1.21749699e+00 -1.92590073e-01
4.41091895e-01 -2.39321619e-01 4.61394042e-02 -1.12485325e+00
9.61119949e-04 -8.62479746e-01 2.50808448e-01 -7.35164344e-01
-1.14969373e+00 8.08051586e-01 -3.18414897e-01 -9.95052516e-01
-3.85455251e-01 -9.95683074e-01 -4.84180719e-01 6.75216079e-01
-1.76904786e+00 -1.00915194e+00 -8.31245661e-01 8.78092945e-01
2.73711205e-01 -4.86879855e-01 1.04990864e+00 5.80933213e-01
-8.05215955e-01 9.45231199e-01 -1.44135088e-01 6.17986739e-01
6.03236258e-01 -8.64588618e-01 -9.48835760e-02 4.99717057e-01
-1.12581328e-01 4.62723643e-01 4.70216811e-01 1.77611604e-01
-1.37316775e+00 -9.84022498e-01 9.35054421e-01 2.05502529e-02
6.31208241e-01 -2.52026916e-01 -6.73173845e-01 5.34631848e-01
2.29113877e-01 7.19630241e-01 7.89679825e-01 -1.47750959e-01
-4.70287949e-01 -5.22592545e-01 -1.33334935e+00 3.44870657e-01
7.92230070e-01 -6.18350267e-01 -1.39900759e-01 -1.70149460e-01
1.02617443e-01 -3.99924368e-01 -1.13946307e+00 5.75657725e-01
1.08890617e+00 -1.40698683e+00 8.61536562e-01 -9.01062965e-01
6.10796511e-01 1.12386249e-01 -2.50689596e-01 -1.06010628e+00
-5.87111041e-02 -8.85166168e-01 -4.10014033e-01 8.96526158e-01
-4.93978597e-02 -4.45976883e-01 9.93982375e-01 2.82781541e-01
9.30260420e-02 -1.18579423e+00 -1.02620912e+00 -5.14249086e-01
-2.02463910e-01 -2.39641517e-01 4.76644069e-01 1.00751233e+00
-2.39361539e-01 1.53800488e-01 -5.90675294e-01 -1.05548941e-01
5.49446464e-01 -1.79125488e-01 8.13562691e-01 -1.30265319e+00
2.36401126e-01 -4.12761301e-01 -1.06938910e+00 -6.61637843e-01
5.96021950e-01 -4.06608850e-01 -2.87548125e-01 -7.71046400e-01
8.64430964e-02 1.57870591e-01 -4.09449369e-01 5.04301667e-01
1.86408997e-01 4.62956041e-01 -8.13222528e-02 -2.77131677e-01
-3.98853809e-01 7.33837724e-01 1.40677786e+00 1.21520519e-01
-1.56140216e-02 -2.11392060e-01 -4.07495201e-01 8.11208308e-01
8.42703462e-01 -7.66577153e-03 -3.15201312e-01 -2.45121986e-01
-1.36223450e-01 1.26981020e-01 4.26619977e-01 -1.00640571e+00
6.52876869e-02 -1.27324358e-01 5.91981709e-01 -2.70354182e-01
6.63675547e-01 -6.12831473e-01 -3.45184773e-01 2.13049948e-01
-2.71814227e-01 1.26116082e-01 3.89456064e-01 1.77228764e-01
-6.30309045e-01 2.14257941e-01 1.09110045e+00 1.20081685e-01
-1.10553408e+00 7.66700387e-01 -3.03314179e-01 7.56901279e-02
9.47970569e-01 -4.05478716e-01 -1.25365853e-01 -4.28742200e-01
-4.94298875e-01 -2.83963561e-01 -4.64210771e-02 4.96657282e-01
7.96785891e-01 -1.50598621e+00 -6.80954099e-01 4.86814559e-01
-3.48676667e-02 -4.07611728e-01 1.77629188e-01 9.47965622e-01
-6.71157181e-01 4.24678564e-01 -9.51369286e-01 -5.09128869e-01
-1.62376928e+00 -1.40175134e-01 7.30883241e-01 6.32994696e-02
-3.93530786e-01 9.84904945e-01 -2.66367704e-01 -1.76079169e-01
2.71589071e-01 8.50549340e-03 -4.15593505e-01 1.21816233e-01
6.11473978e-01 5.22244513e-01 7.92395025e-02 -1.08697367e+00
-3.83664906e-01 7.86251783e-01 -4.37536389e-02 3.67460027e-02
1.65181696e+00 1.77995816e-01 -2.27479920e-01 8.71801302e-02
1.86069095e+00 -4.08425242e-01 -1.41019893e+00 -1.63425013e-01
-1.93919227e-01 -7.26368248e-01 2.48835534e-01 -4.35090452e-01
-1.45079148e+00 1.09192431e+00 8.22878361e-01 -2.55481750e-01
1.51419830e+00 -5.90171993e-01 9.92101371e-01 3.32913160e-01
4.80980836e-02 -1.03479385e+00 4.54375416e-01 5.57630002e-01
6.84004903e-01 -1.25465429e+00 -3.20837535e-02 -1.10718004e-01
-4.93361026e-01 1.45404553e+00 7.57206023e-01 -1.51050478e-01
7.59450793e-01 3.25365454e-01 2.97293901e-01 -3.35058212e-01
-6.82215631e-01 3.73212658e-02 3.05409342e-01 4.91473615e-01
7.36895382e-01 -2.25923643e-01 -1.43665550e-02 2.12020695e-01
-2.46164009e-01 4.52893764e-01 3.63131195e-01 9.38866735e-01
-2.99364150e-01 -6.51965320e-01 2.32670419e-02 2.56730497e-01
-8.50999534e-01 2.98342437e-01 -3.17238480e-01 9.10975993e-01
1.16172306e-01 7.97467470e-01 1.61182404e-01 -5.71411848e-01
3.15379709e-01 -8.87951702e-02 6.07239425e-01 4.94466312e-02
-3.13596338e-01 -4.59146231e-01 1.57698810e-01 -8.89481187e-01
-9.57582474e-01 -6.01651847e-01 -1.05414796e+00 -5.53052485e-01
-4.32709185e-03 -1.28460184e-01 6.98319674e-01 9.69121695e-01
4.99304384e-01 1.89187586e-01 9.82708395e-01 -1.07682049e+00
-2.59277999e-01 -1.04494643e+00 -5.62905014e-01 5.35717010e-01
8.19068730e-01 -8.69195879e-01 -3.60715926e-01 1.46612376e-01] | [13.615650177001953, 1.716415524482727] |
b2581bbc-ef85-4ddd-92fb-6607141b53b8 | a-reverse-jensen-inequality-result-with | 2111.06676 | null | https://arxiv.org/abs/2111.06676v1 | https://arxiv.org/pdf/2111.06676v1.pdf | A Reverse Jensen Inequality Result with Application to Mutual Information Estimation | The Jensen inequality is a widely used tool in a multitude of fields, such as for example information theory and machine learning. It can be also used to derive other standard inequalities such as the inequality of arithmetic and geometric means or the H\"older inequality. In a probabilistic setting, the Jensen inequality describes the relationship between a convex function and the expected value. In this work, we want to look at the probabilistic setting from the reverse direction of the inequality. We show that under minimal constraints and with a proper scaling, the Jensen inequality can be reversed. We believe that the resulting tool can be helpful for many applications and provide a variational estimation of mutual information, where the reverse inequality leads to a new estimator with superior training behavior compared to current estimators. | ['Rafael F. Schaefer', 'Rick Fritschek', 'Benedikt Groß', 'Gerhard Wunder'] | 2021-11-12 | null | null | null | null | ['mutual-information-estimation'] | ['methodology'] | [ 2.04417169e-01 2.63375103e-01 -2.95660853e-01 -3.58078063e-01
-5.42577684e-01 -6.12519622e-01 1.68296516e-01 3.48676026e-01
-4.77373242e-01 9.42452550e-01 -1.71523422e-01 -4.79835868e-01
-5.05210280e-01 -6.47780955e-01 -5.71107805e-01 -1.00781083e+00
-1.03576243e-01 3.15378904e-01 2.79481202e-01 -1.35613665e-01
4.05739397e-01 4.12990719e-01 -1.42586756e+00 -2.80905485e-01
9.10191238e-01 1.21357107e+00 -1.32700086e-01 4.37871069e-01
1.38141001e-02 2.91760415e-01 -2.35189453e-01 -5.54217935e-01
1.88835829e-01 -5.82807302e-01 -8.21866155e-01 -4.21331882e-01
2.26049386e-02 -1.35323197e-01 1.02966584e-01 1.30911863e+00
2.65512556e-01 2.29841217e-01 8.33545685e-01 -1.37319887e+00
-4.12884921e-01 6.73670590e-01 -7.75440633e-01 1.58884436e-01
4.16866630e-01 -7.13839173e-01 1.27383077e+00 -6.91599131e-01
4.06465203e-01 1.02734661e+00 8.56402457e-01 3.64677012e-01
-1.30912316e+00 -4.87791151e-01 -2.07911849e-01 3.40983510e-01
-1.36213493e+00 3.73011902e-02 4.39853400e-01 -4.58667278e-01
3.25770050e-01 3.34864646e-01 3.39533776e-01 5.34437120e-01
3.37439924e-01 8.85327458e-01 1.15136647e+00 -6.31070077e-01
1.21555455e-01 5.28558671e-01 2.55952269e-01 7.11508870e-01
4.80405539e-01 6.19221553e-02 -2.73414850e-01 -1.84657812e-01
4.74734217e-01 -1.37195230e-01 -4.74563509e-01 -5.99565387e-01
-9.77287829e-01 1.18863118e+00 1.35580525e-01 4.16259438e-01
3.82038280e-02 2.13478416e-01 2.14196712e-01 2.52109915e-01
5.30400455e-01 2.70472854e-01 -4.81038183e-01 -1.06122427e-01
-8.59259427e-01 3.00905019e-01 1.25189805e+00 8.66021156e-01
5.46211064e-01 -3.30008149e-01 -1.68854482e-02 5.66399097e-01
7.16282308e-01 5.33391833e-01 -1.99258670e-01 -1.01338613e+00
3.30053300e-01 1.62174299e-01 2.51312822e-01 -8.99388909e-01
-3.58549654e-01 -4.21193570e-01 -7.14461863e-01 7.19662979e-02
8.28213632e-01 -1.26374871e-01 -2.59555042e-01 1.93518555e+00
2.30012909e-01 -1.06678352e-01 -1.99844688e-01 5.73153675e-01
2.09001690e-01 6.47149682e-01 -4.14839625e-01 -4.98596907e-01
1.07795608e+00 -2.65687257e-01 -8.89055073e-01 2.25249201e-01
5.15762806e-01 -6.98130965e-01 5.53369462e-01 5.43965697e-01
-1.06196749e+00 5.11245765e-02 -9.36819315e-01 8.31324086e-02
-2.82010704e-01 -2.99674749e-01 5.38999736e-01 9.78537858e-01
-1.06702065e+00 8.47699106e-01 -7.21769273e-01 -3.26071203e-01
1.65521413e-01 2.43237421e-01 -3.74842942e-01 6.99041225e-03
-1.12447894e+00 1.13495588e+00 3.21784973e-01 2.17960209e-01
-1.48322254e-01 -6.52868390e-01 -9.72743154e-01 1.99986368e-01
3.77119362e-01 -5.56861639e-01 1.05780816e+00 -5.20509839e-01
-1.45849025e+00 5.84501028e-01 -2.14796603e-01 -5.33415914e-01
8.43791425e-01 -3.36259097e-01 1.62122101e-01 -6.02912903e-02
-2.96932869e-02 -1.24887608e-01 5.88190675e-01 -9.18143570e-01
-2.92957217e-01 -6.00674748e-01 -3.19694378e-03 -8.37227255e-02
-7.34304860e-02 -8.05033892e-02 -7.62973353e-02 -4.39906240e-01
2.13388786e-01 -9.07794476e-01 -8.78802389e-02 7.27284551e-02
-4.73941714e-01 -3.86003405e-01 3.59805554e-01 -4.96882975e-01
1.16959476e+00 -2.06429100e+00 2.23767310e-01 5.81951559e-01
8.42986181e-02 -1.89558417e-01 4.36031729e-01 6.30439520e-01
-6.32777363e-02 3.12413961e-01 -6.72724128e-01 -1.66016892e-01
3.35376620e-01 -8.30617175e-03 -9.12499726e-02 1.02016628e+00
-4.58595641e-02 6.11739278e-01 -9.75136042e-01 -5.22828877e-01
1.86758906e-01 4.46395367e-01 -6.45234764e-01 -1.44322515e-02
7.01777115e-02 4.27228063e-01 -1.92422986e-01 -1.61414012e-01
7.34054983e-01 -1.19192794e-01 2.07618013e-01 3.97526734e-02
-1.57172218e-01 2.82957759e-02 -1.39609826e+00 1.30702162e+00
-5.11499286e-01 9.22372997e-01 2.49694899e-01 -1.41512024e+00
6.89112604e-01 1.69100106e-01 5.88592708e-01 -1.13587882e-02
5.09888291e-01 1.37087077e-01 5.24377730e-03 -4.56639260e-01
3.47203135e-01 -7.17419863e-01 -1.78100586e-01 3.44940633e-01
-2.41831094e-01 -2.25158483e-01 2.66428620e-01 5.75562157e-02
7.31080711e-01 2.49703284e-02 5.89515626e-01 -7.76392877e-01
5.98915398e-01 -5.46470106e-01 3.80252361e-01 8.51471364e-01
-5.60421236e-02 3.81629199e-01 7.38392174e-01 2.93818656e-02
-7.61521935e-01 -1.39995170e+00 -7.20833898e-01 8.59039187e-01
1.40870199e-01 -9.66287032e-02 -5.47230661e-01 -3.88402283e-01
1.51569858e-01 6.34378195e-01 -6.99631870e-01 -2.27483869e-01
-6.83914945e-02 -5.80232382e-01 4.95944291e-01 5.02264678e-01
5.10585189e-01 -3.07042897e-01 -3.21614712e-01 -1.29479244e-01
-5.92273809e-02 -8.28525245e-01 -4.59187686e-01 2.32108667e-01
-7.66693234e-01 -9.60526824e-01 -8.67511690e-01 -4.05351579e-01
4.86283273e-01 -1.46973863e-01 8.70165706e-01 -1.23897448e-01
1.61156714e-01 3.05481166e-01 -1.82944521e-01 -4.49394166e-01
-5.30708551e-01 -8.12167898e-02 1.23250671e-01 8.54323208e-02
1.69844210e-01 -4.40680504e-01 -3.54199708e-01 4.88492846e-01
-1.05010426e+00 -3.70993912e-01 3.28996330e-01 8.10666442e-01
3.91278654e-01 2.27730379e-01 5.25491059e-01 -9.40360665e-01
5.70335865e-01 -5.47111154e-01 -9.50990677e-01 2.18819320e-01
-7.31821835e-01 7.08982229e-01 7.25251555e-01 -6.72381893e-02
-9.68720376e-01 -8.64597037e-02 -2.11601794e-01 9.18869004e-02
2.90251851e-01 6.29754961e-01 -3.29698682e-01 -1.03219468e-02
2.12083116e-01 1.37581378e-01 -1.56377748e-01 -4.09796387e-01
4.35841352e-01 6.96219981e-01 3.88397008e-01 -5.11567295e-01
6.66256249e-01 6.14989698e-01 4.34873074e-01 -9.26732242e-01
-1.05415976e+00 -6.53569698e-01 -4.82865959e-01 -2.28586763e-01
7.95490384e-01 -1.06292322e-01 -1.10505843e+00 5.24972454e-02
-1.11397016e+00 -7.43519589e-02 -3.02304506e-01 7.10982263e-01
-7.66849160e-01 5.08831382e-01 -4.73415762e-01 -1.42975402e+00
1.70819283e-01 -9.58592117e-01 8.21700633e-01 4.12021607e-01
-5.39669506e-02 -1.48445523e+00 2.83001930e-01 1.60006598e-01
2.86454648e-01 2.45094344e-01 8.47086549e-01 -6.82788491e-01
-9.66749340e-02 -4.03775126e-01 -3.00005704e-01 6.39276445e-01
-5.79682402e-02 1.04756258e-01 -5.56032956e-01 2.20497847e-02
1.03841200e-01 2.70972401e-01 1.10956740e+00 5.95936835e-01
1.04402292e+00 -2.82752246e-01 -2.47509181e-01 5.26712239e-01
1.50870144e+00 -9.76535231e-02 5.20218909e-01 -1.42531395e-02
3.38537663e-01 7.68562257e-01 4.15365249e-01 6.38288617e-01
2.84690410e-01 6.17921054e-01 2.92144358e-01 3.65047365e-01
5.91595471e-01 -4.98178937e-02 1.91087633e-01 8.05939198e-01
-1.19380675e-01 -5.77824712e-02 -6.53968632e-01 3.50909531e-01
-1.88353741e+00 -1.17927980e+00 -2.79437244e-01 2.79656458e+00
1.00020397e+00 -2.80673672e-02 6.39288425e-02 3.53812367e-01
9.68989491e-01 -1.20514713e-01 -3.74738932e-01 -7.35795259e-01
9.40434262e-02 2.68232137e-01 8.10740709e-01 8.57944906e-01
-8.98231566e-01 3.18307191e-01 7.33644533e+00 7.78609037e-01
-7.18951106e-01 1.86566010e-01 1.71539590e-01 2.58968174e-01
-3.64443988e-01 1.59417093e-02 -6.07006252e-01 5.48801541e-01
8.36729527e-01 -4.85450298e-01 2.91476995e-01 7.58790493e-01
7.31207579e-02 -4.76335436e-01 -1.15049756e+00 9.01015759e-01
-3.27477301e-03 -8.21731567e-01 -6.29233181e-01 4.74495530e-01
7.61155784e-01 -2.65305310e-01 1.02700852e-01 1.79858152e-02
1.35236621e-01 -1.10199142e+00 3.92144322e-01 4.99363035e-01
4.48439538e-01 -8.99695992e-01 1.18223536e+00 4.15515989e-01
-1.20111382e+00 2.68369526e-01 -1.15994155e-01 -2.22326070e-01
2.31867790e-01 1.10912693e+00 -5.58606982e-01 7.85984397e-01
2.37072095e-01 5.23885071e-01 6.15113564e-02 1.29806304e+00
-2.99956232e-01 3.88103932e-01 -6.61308348e-01 -3.53994966e-01
-1.76651791e-01 -6.76820576e-01 6.45663857e-01 1.18194413e+00
2.81406850e-01 -3.21193457e-01 -2.89878398e-01 7.14828908e-01
-1.36289269e-01 1.63943768e-01 -7.25037932e-01 2.82088518e-01
2.76949346e-01 9.27738607e-01 -7.77570128e-01 -1.27159789e-01
-3.43421996e-01 7.91546643e-01 2.75319397e-01 2.20053360e-01
-9.56706047e-01 -6.08517468e-01 5.19236207e-01 -1.15196809e-01
3.86107028e-01 -3.15821916e-01 -4.02382702e-01 -9.98438656e-01
1.32292494e-01 -2.35293865e-01 3.36361676e-01 -2.09322304e-01
-1.28315806e+00 2.24404514e-01 4.73909646e-01 -8.53879690e-01
-3.63321066e-01 -9.20809031e-01 -4.86075610e-01 7.76966810e-01
-1.18220520e+00 -1.60101041e-01 1.34227678e-01 3.23126614e-01
-1.81549832e-01 4.01670247e-01 4.83902693e-01 2.79457569e-01
-6.36468410e-01 5.25546253e-01 6.51408494e-01 -1.24918073e-02
5.29062867e-01 -1.67710662e+00 -3.93211782e-01 8.39228272e-01
2.13321865e-01 5.67161798e-01 1.40376937e+00 -2.25416571e-01
-1.26541984e+00 -5.74571311e-01 7.85825431e-01 -4.65149224e-01
9.02750611e-01 -9.14329588e-02 -6.22204721e-01 5.32289922e-01
-3.64275351e-02 1.40634123e-02 9.14078593e-01 2.60960907e-01
-3.60588431e-01 -8.12752247e-02 -1.29827440e+00 2.73013085e-01
6.21984720e-01 -4.07951474e-01 -5.19413352e-01 3.74962062e-01
2.10195661e-01 -1.55295283e-01 -1.17591536e+00 3.59242946e-01
1.05623829e+00 -1.32920122e+00 7.22635448e-01 -5.77257276e-01
4.16997582e-01 -6.53937161e-02 -4.63798583e-01 -1.29498005e+00
1.08559817e-01 -7.18277991e-01 -2.66712427e-01 1.02353871e+00
6.36684299e-01 -1.07095563e+00 4.70414728e-01 6.45140171e-01
2.61273503e-01 -1.12950540e+00 -1.22152746e+00 -1.20135903e+00
3.58490914e-01 -8.08089375e-01 2.29177386e-01 6.50511503e-01
4.34927493e-01 4.70263138e-02 -3.74323159e-01 2.25469545e-02
7.60226548e-01 -4.28583696e-02 3.04037929e-01 -1.40219927e+00
-5.43919325e-01 -7.36831903e-01 -6.04484379e-01 -1.28419662e+00
2.02956274e-01 -9.15504694e-01 5.21030843e-01 -1.48034465e+00
2.29588091e-01 -4.35974628e-01 -3.78928006e-01 -1.83689147e-01
-1.69490308e-01 2.39084587e-01 6.32857233e-02 -2.82238096e-01
-4.04829979e-01 6.04507327e-01 8.76044989e-01 2.07169756e-01
-8.37019756e-02 4.82835770e-01 -7.92142034e-01 8.90907347e-01
8.08695078e-01 -5.75515866e-01 -4.80202958e-02 1.73852406e-02
7.06077635e-01 -1.11084893e-01 1.94607839e-01 -6.08152807e-01
6.38308795e-03 3.31391692e-02 -4.13321927e-02 -2.44195700e-01
1.91329032e-01 -8.47608328e-01 -9.77426842e-02 5.25341392e-01
-3.29328895e-01 -1.63050130e-01 -1.46807000e-01 7.14259863e-01
-2.89350096e-02 -7.84044087e-01 8.20391417e-01 1.43904001e-01
8.10238123e-02 4.37462702e-02 -2.10425332e-01 3.15187216e-01
9.85664546e-01 1.27475969e-02 3.09880055e-03 -7.43855238e-01
-6.03687227e-01 2.80412465e-01 3.20218384e-01 -6.48638085e-02
4.19003665e-01 -1.13504839e+00 -7.76619673e-01 -2.05543399e-01
-1.63140103e-01 -2.85009086e-01 -4.89110760e-02 1.70393586e+00
-4.27125037e-01 4.59323496e-01 3.35645407e-01 -5.47224224e-01
-1.01440656e+00 3.84431779e-01 1.28596097e-01 -5.03502727e-01
-2.77251393e-01 5.81956625e-01 2.21679732e-01 -1.20531067e-01
1.80282474e-01 -2.65552968e-01 5.27688824e-02 4.27095257e-02
7.09959745e-01 6.62235618e-01 -2.20742941e-01 -5.59583962e-01
-6.69692099e-01 6.22079194e-01 2.50977308e-01 -5.68629265e-01
1.17493415e+00 -3.54710549e-01 -2.04116791e-01 9.62361932e-01
1.57765996e+00 3.92279446e-01 -9.20160890e-01 -3.66975069e-02
2.78089553e-01 -3.24764937e-01 -6.46651611e-02 -3.32854062e-01
-8.53351951e-01 9.29740131e-01 3.25369775e-01 8.93697381e-01
1.05200899e+00 3.75704288e-01 4.96122956e-01 2.03173697e-01
4.76031482e-01 -9.66908634e-01 -5.08535504e-01 3.90112609e-01
6.58173084e-01 -1.25015068e+00 1.87152073e-01 -5.87122023e-01
-3.14114809e-01 1.21539509e+00 -3.05360019e-01 -2.52178937e-01
1.03973937e+00 4.32868004e-01 -4.60698575e-01 2.79383242e-01
-3.52478266e-01 -5.13643086e-01 5.10871649e-01 5.68826556e-01
7.69049942e-01 1.42804414e-01 -6.67889178e-01 4.87242252e-01
-3.00191462e-01 -1.57764390e-01 5.55526912e-01 4.76648092e-01
-3.82272720e-01 -1.01554406e+00 -2.37663671e-01 6.14086032e-01
-6.84965432e-01 -2.40487278e-01 -2.55540013e-01 7.74405658e-01
-8.35536644e-02 9.73344803e-01 -1.11315444e-01 -1.08814947e-01
-5.10089584e-02 1.69070527e-01 7.86910176e-01 -4.49871689e-01
-6.51425719e-02 -2.17641711e-01 -4.76744883e-02 -4.90496010e-01
-6.33957684e-01 -8.38848412e-01 -1.00556242e+00 -5.00580966e-01
-7.62975574e-01 4.18868601e-01 7.28938937e-01 1.42086482e+00
-1.93092749e-01 2.79441684e-01 6.43498421e-01 -4.67367172e-01
-7.78308511e-01 -9.30145383e-01 -9.16546047e-01 -2.94435546e-02
3.58857542e-01 -7.89430141e-01 -8.09513509e-01 -2.77087599e-01] | [7.293030738830566, 4.191829204559326] |
a23cdb44-216f-41b0-9271-ce7323f7fa0e | multimodal-review-generation-with-privacy-and | null | null | https://aclanthology.org/2020.coling-main.37 | https://aclanthology.org/2020.coling-main.37.pdf | Multimodal Review Generation with Privacy and Fairness Awareness | Users express their opinions towards entities (e.g., restaurants) via online reviews which can be in diverse forms such as text, ratings, and images. Modeling reviews are advantageous for user behavior understanding which, in turn, supports various user-oriented tasks such as recommendation, sentiment analysis, and review generation. In this paper, we propose MG-PriFair, a multimodal neural-based framework, which generates personalized reviews with privacy and fairness awareness. Motivated by the fact that reviews might contain personal information and sentiment bias, we propose a novel differentially private (dp)-embedding model for training privacy guaranteed embeddings and an evaluation approach for sentiment fairness in the food-review domain. Experiments on our novel review dataset show that MG-PriFair is capable of generating plausibly long reviews while controlling the amount of exploited user data and using the least sentiment biased word embeddings. To the best of our knowledge, we are the first to bring user privacy and sentiment fairness into the review generation task. The dataset and source codes are available at https://github.com/ReML-AI/MG-PriFair. | ['Lili Jiang', 'Duc-Trong Le', 'Thanh-Son Nguyen', 'Xuan-Son Vu'] | 2020-12-01 | null | null | null | coling-2020-8 | ['review-generation'] | ['natural-language-processing'] | [-1.68687910e-01 3.43079269e-01 -5.95967829e-01 -7.71304071e-01
-4.85926181e-01 -5.74148357e-01 5.87142289e-01 3.79045159e-01
-3.85206491e-01 6.13437176e-01 5.94614327e-01 -1.94644362e-01
2.99794614e-01 -8.91133964e-01 -3.76583397e-01 -3.58414322e-01
4.53633219e-01 -1.03210032e-01 -6.62677586e-01 -4.38148975e-01
1.92102551e-01 -8.55279118e-02 -1.24881959e+00 3.20752710e-01
8.92282903e-01 1.11850321e+00 -3.15106243e-01 4.35286969e-01
1.79751709e-01 5.03106594e-01 -1.62241414e-01 -1.46551800e+00
4.50990409e-01 -2.18467012e-01 -3.38243544e-01 -7.03830943e-02
2.99451739e-01 -6.03592515e-01 -3.03466082e-01 1.09027517e+00
7.00018287e-01 2.99431920e-01 7.92721033e-01 -1.32140768e+00
-1.88234091e+00 8.25540006e-01 -5.97533643e-01 -3.14113587e-01
2.25267410e-01 -1.90752950e-02 1.47703457e+00 -1.25713432e+00
3.62283736e-01 1.14410055e+00 2.43232772e-01 1.05796838e+00
-9.69170272e-01 -6.21049821e-01 2.84014970e-01 -2.22654864e-01
-1.10569251e+00 -6.18526101e-01 5.98716497e-01 -2.23035872e-01
3.29752177e-01 5.17859459e-01 3.08025420e-01 1.29893494e+00
2.49472857e-01 1.06892681e+00 7.18877316e-01 -1.97234809e-01
5.31312883e-01 8.78881156e-01 2.28471309e-01 3.90808433e-01
7.19833314e-01 -1.23946384e-01 -7.36926019e-01 -3.11087072e-01
7.25478977e-02 3.62205535e-01 -2.66692817e-01 -6.60323441e-01
-8.77455115e-01 1.33228207e+00 3.70251954e-01 -3.36577386e-01
-3.19442034e-01 4.70737331e-02 5.85041106e-01 2.42807195e-01
1.03409123e+00 6.19094729e-01 -6.19620442e-01 3.79083246e-01
-3.62153053e-01 2.88013130e-01 8.59268606e-01 9.08769965e-01
6.17730975e-01 -1.92148551e-01 -2.71992415e-01 9.84579027e-01
6.34376347e-01 7.15337574e-01 6.88418627e-01 -6.83144152e-01
3.13256294e-01 5.01022756e-01 3.63595515e-01 -1.40228868e+00
-5.84355146e-02 7.98153803e-02 -8.96359265e-01 -2.08179474e-01
-5.34608439e-02 -5.45680761e-01 -4.03935552e-01 1.76769197e+00
3.77604455e-01 -4.87904847e-01 2.78337866e-01 8.92954886e-01
9.94624019e-01 6.00406408e-01 2.10431010e-01 9.85380262e-02
1.50950170e+00 -1.04619849e+00 -7.63464272e-01 -1.21882729e-01
6.53795838e-01 -6.05169296e-01 1.15487957e+00 1.49829119e-01
-8.69032085e-01 -3.06548417e-01 -9.04623330e-01 -6.10263050e-01
-8.54826927e-01 5.41518986e-01 9.76563871e-01 1.02420890e+00
-9.83263254e-01 3.36780429e-01 -3.12410831e-01 -3.25669825e-01
7.40376592e-01 1.33815229e-01 -5.60042322e-01 -1.76856458e-01
-1.48917341e+00 7.54043043e-01 -3.76353592e-01 2.19092384e-01
-4.06135231e-01 -5.84614694e-01 -1.26721931e+00 6.10558838e-02
7.48504177e-02 -6.90510511e-01 1.22624183e+00 -1.00117517e+00
-1.49818420e+00 8.42022300e-01 -1.66687116e-01 -3.89573604e-01
4.17867124e-01 -3.71335775e-01 -4.41839159e-01 -1.50698781e-01
1.57170892e-01 7.97736049e-01 9.26065862e-01 -1.15577710e+00
-2.34356120e-01 -6.86406910e-01 2.60198891e-01 3.39596033e-01
-8.41498733e-01 -8.25969726e-02 8.05453630e-04 -5.22864759e-01
-7.97988236e-01 -9.84559178e-01 -5.30225456e-01 3.30307335e-01
-4.46206123e-01 -1.70376152e-01 4.61610138e-01 -5.20796180e-01
1.14133024e+00 -2.22734284e+00 -2.47069791e-01 9.04509202e-02
3.04063886e-01 3.95304352e-01 -3.01733583e-01 3.58257681e-01
2.46382914e-02 6.21794045e-01 -7.65956864e-02 -8.56252491e-01
4.92616028e-01 -9.96243209e-02 -4.33029771e-01 5.82123458e-01
1.51382089e-01 1.24375999e+00 -1.00547194e+00 9.50361136e-03
4.56054285e-02 6.60876215e-01 -6.95090234e-01 1.64075240e-01
-1.25444874e-01 2.08368804e-02 -6.64668620e-01 5.35304546e-01
7.41930485e-01 -1.01426683e-01 -1.29043199e-02 -2.19317749e-01
1.93089217e-01 1.22790195e-01 -5.91989696e-01 1.28915322e+00
-5.46829462e-01 4.61348981e-01 6.53739125e-02 -4.74345386e-01
1.03464735e+00 2.77086824e-01 1.97787911e-01 -5.53986192e-01
4.54273462e-01 4.46209759e-02 -5.05607784e-01 -3.27058077e-01
1.32464969e+00 -3.49662416e-02 -4.35772240e-01 8.11302662e-01
-2.80738562e-01 -7.27178976e-02 1.22123569e-01 4.26393330e-01
5.10954559e-01 -2.97014564e-01 4.05563354e-01 -1.31404951e-01
4.74810213e-01 -4.79008883e-01 1.99218884e-01 5.01315713e-01
-5.86000144e-01 7.08684027e-01 4.56456125e-01 -3.99919629e-01
-8.29948306e-01 -8.43005240e-01 -5.07112630e-02 1.34991264e+00
7.60412291e-02 -4.21325207e-01 -3.92477483e-01 -9.11830366e-01
3.71544361e-01 9.56079543e-01 -8.90076160e-01 -4.46054667e-01
1.87418386e-01 -7.34736145e-01 2.19563618e-01 2.98851490e-01
2.77574547e-02 -9.53948140e-01 -8.08673874e-02 -7.80886635e-02
-6.64246082e-02 -7.16031015e-01 -1.14199805e+00 -1.88880146e-01
-3.87971312e-01 -8.42105567e-01 -7.99182653e-01 -4.51444477e-01
1.03881967e+00 4.20769006e-01 1.13138592e+00 -1.50920510e-01
-1.34891599e-01 7.92031467e-01 -5.99789917e-01 -6.77309036e-01
-4.07011108e-03 5.49463592e-02 1.06713563e-01 4.55863327e-01
9.28989530e-01 -1.06519647e-01 -1.07038260e+00 1.51423961e-01
-1.09556282e+00 -3.25487226e-01 2.00942397e-01 7.41212785e-01
4.99857277e-01 -4.09375966e-01 1.07984781e+00 -1.34720647e+00
1.35979366e+00 -9.60683167e-01 -2.25017801e-01 1.29595473e-01
-9.14714873e-01 -1.58758208e-01 6.79873943e-01 -3.68325651e-01
-1.00800192e+00 -1.33892462e-01 2.50087827e-02 -1.09497905e-01
1.81121469e-01 3.39641601e-01 -2.58565694e-01 2.68629491e-01
7.61100531e-01 -3.28615606e-02 2.21091136e-01 -1.12534434e-01
1.01910853e+00 9.12672997e-01 -1.23231143e-01 -2.68845648e-01
3.92105877e-01 4.46810037e-01 -4.83914018e-01 -4.40457344e-01
-9.88836944e-01 -4.58863705e-01 -1.77278802e-01 8.01854301e-03
9.30265963e-01 -1.16309226e+00 -6.57519400e-01 3.22008073e-01
-9.76469755e-01 1.49273559e-01 -4.82948035e-01 2.87157416e-01
-2.29883343e-01 2.66565382e-01 -5.49868822e-01 -9.96745884e-01
-8.06317270e-01 -9.57087338e-01 9.89338994e-01 2.48961776e-01
-4.76547301e-01 -1.01225948e+00 8.05575550e-02 4.95909959e-01
6.48023129e-01 1.38141736e-02 3.98377538e-01 -8.65800738e-01
-3.25871408e-01 -4.77481872e-01 -2.06557140e-01 7.03801870e-01
2.39695042e-01 -5.35284244e-02 -9.99366403e-01 -1.71464369e-01
-2.24262252e-01 -6.84575140e-01 8.91335309e-01 2.31655985e-01
1.17860031e+00 -7.36398995e-01 1.08860292e-01 2.77161747e-01
1.23216867e+00 -2.05683097e-01 6.31420374e-01 -2.01786473e-01
7.69901991e-01 8.07247281e-01 6.14230752e-01 8.70122373e-01
9.06904280e-01 1.67775631e-01 5.63211977e-01 -2.70552766e-02
4.51616079e-01 -5.07301211e-01 4.88893658e-01 6.52442217e-01
5.82726598e-01 -5.93484879e-01 -1.39443874e-01 8.58250737e-01
-1.77573991e+00 -9.38711524e-01 1.01689495e-01 2.06768060e+00
8.50889742e-01 -5.65305412e-01 -1.77574292e-01 -4.59095627e-01
6.82105780e-01 5.27147174e-01 -7.23993123e-01 -1.13013172e+00
4.95102555e-02 5.78373559e-02 7.27280140e-01 4.71349567e-01
-1.13026142e+00 7.35629559e-01 5.06813860e+00 3.36059451e-01
-8.72203290e-01 1.25582978e-01 1.20751715e+00 -3.84515911e-01
-1.12638068e+00 -5.18333256e-01 -8.53309095e-01 4.75028694e-01
8.89237702e-01 -6.54913604e-01 3.36017907e-01 1.19517660e+00
2.54627466e-01 2.37201363e-01 -1.07574689e+00 7.70951986e-01
3.71306479e-01 -1.07853901e+00 2.34861895e-01 1.17437907e-01
1.03845572e+00 -8.90787318e-02 7.80354083e-01 3.28082860e-01
5.24429679e-01 -1.13162661e+00 3.57808113e-01 3.33880603e-01
7.16032267e-01 -9.34518158e-01 7.21670389e-01 -8.69697109e-02
-6.75445139e-01 -8.30700528e-03 -6.88889980e-01 1.25156879e-01
2.51440674e-01 9.67683792e-01 -3.64571691e-01 4.21441942e-01
5.00724733e-01 1.09219897e+00 -4.33322936e-01 3.80897045e-01
-2.93001652e-01 2.03812331e-01 -3.36373411e-02 -5.53773940e-01
-1.31891251e-01 -4.15539682e-01 -3.26600275e-04 8.87946665e-01
4.19619232e-01 4.17088866e-02 -2.98295647e-01 6.97153986e-01
-7.46350110e-01 5.77253819e-01 -9.46304321e-01 -3.86714578e-01
3.03735375e-01 1.58446109e+00 9.88293886e-02 3.80969122e-02
-4.12216455e-01 1.15510809e+00 3.43047500e-01 3.02932858e-01
-5.18329978e-01 -2.02757657e-01 1.11607289e+00 -1.06555708e-01
3.53357852e-01 1.78827628e-01 -3.73627573e-01 -1.54368043e+00
-9.62023251e-03 -8.93987894e-01 1.57176673e-01 -6.76004708e-01
-1.75932980e+00 4.40586448e-01 -4.52225864e-01 -9.33222234e-01
-7.01810643e-02 -7.44921386e-01 -4.91270185e-01 9.11781907e-01
-1.81046462e+00 -1.21471393e+00 2.04843566e-01 5.28042972e-01
2.22192645e-01 -2.02202931e-01 1.14110839e+00 3.35679293e-01
-5.49912155e-01 1.06423759e+00 2.05693647e-01 -1.53464852e-02
9.81506765e-01 -1.19417834e+00 3.90893161e-01 5.88796973e-01
-1.46053061e-01 9.81393814e-01 3.76167238e-01 -4.13493216e-01
-1.38906121e+00 -1.32827866e+00 1.24632597e+00 -6.18060589e-01
6.05801284e-01 -4.44816709e-01 -4.19158280e-01 6.10753357e-01
4.08072263e-01 -4.93510067e-02 1.60241282e+00 1.03283361e-01
-4.45931941e-01 -2.33021840e-01 -1.52004528e+00 1.02103543e+00
5.61192393e-01 -6.36274159e-01 -7.63109475e-02 3.62703949e-01
8.67588758e-01 -1.78569593e-02 -7.91682780e-01 -3.11910249e-02
7.29659855e-01 -7.55212307e-01 7.99755454e-01 -5.81151187e-01
1.09692013e+00 -8.05723071e-02 -2.55222499e-01 -1.64253724e+00
-1.17813259e-01 -4.70887750e-01 -1.69516012e-01 1.35406101e+00
6.89323962e-01 -7.61405349e-01 7.81021714e-01 1.43370891e+00
3.27778637e-01 -9.26319540e-01 -4.87962216e-01 5.43905012e-02
1.06141478e-01 -2.00331643e-01 8.65923882e-01 9.20521200e-01
1.02176487e-01 2.62578189e-01 -9.21440482e-01 6.80698901e-02
4.76509899e-01 -2.96965223e-02 8.79174948e-01 -8.01912129e-01
-4.81291711e-02 -1.41093075e-01 7.21030384e-02 -1.03248835e+00
3.95487309e-01 -9.19612825e-01 -2.17489958e-01 -1.42081118e+00
-2.70683356e-02 -3.12032551e-01 -3.18190604e-01 3.46701115e-01
-3.86321604e-01 2.50289828e-01 1.06405765e-01 -2.80558646e-01
-6.04284883e-01 1.01438856e+00 1.42460704e+00 -7.41900653e-02
9.17549357e-02 2.04639405e-01 -1.93900359e+00 3.12935233e-01
1.38929355e+00 -2.67573595e-01 -7.85043478e-01 -3.69293541e-01
7.54968166e-01 -4.24291998e-01 3.49848382e-02 7.73492157e-02
-6.32451028e-02 -1.23642929e-01 2.91470259e-01 -4.01540339e-01
3.29317033e-01 -8.88711393e-01 -3.72714698e-01 -5.78396805e-02
-8.52642953e-01 1.97482869e-01 -1.12094797e-01 7.70095706e-01
-1.26710653e-01 -2.13600487e-01 4.19326961e-01 -2.72072107e-01
-3.75229269e-01 7.21457958e-01 -2.79820949e-01 -5.03197871e-02
8.86148810e-01 2.81940103e-01 -6.24850452e-01 -9.32093501e-01
-3.97719085e-01 5.91156542e-01 4.98658121e-01 7.66210675e-01
8.61227930e-01 -1.42783737e+00 -7.37875700e-01 2.18117505e-01
6.09532177e-01 -2.81857222e-01 3.47790122e-01 2.24244982e-01
4.81044762e-02 2.65010864e-01 1.55553430e-01 1.86658606e-01
-9.58890736e-01 6.67713702e-01 2.06669290e-02 -2.59437650e-01
1.89941809e-01 1.02853227e+00 3.55324894e-01 -1.07883704e+00
-9.09368992e-02 -4.89070453e-02 -4.04879481e-01 3.53214353e-01
9.27534461e-01 2.12416902e-01 -3.09322745e-01 -6.69164300e-01
-2.05861598e-01 5.54307774e-02 -4.97729927e-01 9.99184549e-02
1.20687234e+00 -5.27427554e-01 -2.19101578e-01 1.53315455e-01
1.57330477e+00 4.12318289e-01 -8.87782037e-01 -3.08719855e-02
-4.87223297e-01 -7.59255707e-01 -9.58051756e-02 -7.22927988e-01
-1.27212632e+00 7.13716567e-01 3.03733170e-01 2.80413479e-01
8.90696406e-01 -2.97572732e-01 1.00705528e+00 1.73058689e-01
2.31229663e-01 -1.04280877e+00 -6.60787895e-02 1.97739512e-01
8.32027376e-01 -1.60499156e+00 1.24636151e-01 -8.06535631e-02
-1.41118348e+00 6.12716198e-01 6.39417231e-01 6.81412127e-03
9.34328139e-01 -2.79088885e-01 4.69316959e-01 -1.14699490e-02
-8.56320500e-01 1.59172192e-01 2.30356961e-01 7.97501862e-01
6.24506891e-01 4.26417917e-01 -6.02145016e-01 9.93468046e-01
-1.64044410e-01 -1.88983958e-02 8.34034503e-01 4.91861701e-01
-9.95014384e-02 -1.37241852e+00 2.14683518e-01 7.65991867e-01
-8.75544667e-01 -4.26188499e-01 -4.86300528e-01 3.43318097e-03
-1.97843581e-01 1.21703172e+00 -2.64158547e-01 -3.03404778e-01
2.66924083e-01 -1.73884705e-01 -2.10973009e-01 -7.45779216e-01
-8.39327097e-01 -6.27973914e-01 3.35923046e-01 -3.92052889e-01
-3.90127629e-01 -4.71578836e-01 -6.42566800e-01 -6.35428667e-01
-2.86447197e-01 1.81765005e-01 7.00643361e-01 2.47918844e-01
8.28546822e-01 -1.09491684e-01 1.06033945e+00 -5.62026560e-01
-7.54799604e-01 -5.98777413e-01 -7.58311808e-01 6.96280658e-01
3.02132905e-01 -1.13044031e-01 -6.48097575e-01 -2.29273707e-01] | [11.334941864013672, 6.741359233856201] |
12b57a11-1fbe-4465-a338-3c5ad5477cf0 | text-compression-aided-transformer-encoding | 2102.05951 | null | https://arxiv.org/abs/2102.05951v1 | https://arxiv.org/pdf/2102.05951v1.pdf | Text Compression-aided Transformer Encoding | Text encoding is one of the most important steps in Natural Language Processing (NLP). It has been done well by the self-attention mechanism in the current state-of-the-art Transformer encoder, which has brought about significant improvements in the performance of many NLP tasks. Though the Transformer encoder may effectively capture general information in its resulting representations, the backbone information, meaning the gist of the input text, is not specifically focused on. In this paper, we propose explicit and implicit text compression approaches to enhance the Transformer encoding and evaluate models using this approach on several typical downstream tasks that rely on the encoding heavily. Our explicit text compression approaches use dedicated models to compress text, while our implicit text compression approach simply adds an additional module to the main model to handle text compression. We propose three ways of integration, namely backbone source-side fusion, target-side fusion, and both-side fusion, to integrate the backbone information into Transformer-based models for various downstream tasks. Our evaluation on benchmark datasets shows that the proposed explicit and implicit text compression approaches improve results in comparison to strong baselines. We therefore conclude, when comparing the encodings to the baseline models, text compression helps the encoders to learn better language representations. | ['Eiichiro Sumita', 'Masao Utiyama', 'Kehai Chen', 'Rui Wang', 'Hai Zhao', 'Zhuosheng Zhang', 'Zuchao Li'] | 2021-02-11 | null | null | null | null | ['text-compression'] | ['natural-language-processing'] | [ 5.46029150e-01 2.51371235e-01 -2.91735798e-01 -4.75928664e-01
-9.48326290e-01 -1.51937276e-01 1.04904187e+00 5.90136826e-01
-6.49594009e-01 5.17397404e-01 1.00133634e+00 -3.87656838e-01
2.25145295e-01 -9.06999588e-01 -8.90603662e-01 -3.99347007e-01
4.01427299e-01 7.59551942e-01 1.96361408e-01 -4.04530525e-01
4.52001452e-01 -9.88977123e-03 -1.47986901e+00 9.14234936e-01
7.88201034e-01 1.07022452e+00 5.71963191e-01 7.07432687e-01
-8.65557253e-01 1.06791818e+00 -4.56171989e-01 -6.08303010e-01
1.17093369e-01 -2.87446588e-01 -1.03747487e+00 -4.69440788e-01
2.71319211e-01 -4.77623582e-01 -6.35826886e-01 8.54958475e-01
4.27422404e-01 1.11821048e-01 8.04172993e-01 -6.84206426e-01
-8.38447750e-01 1.20777667e+00 -5.67122638e-01 1.94786444e-01
2.76415110e-01 -9.59905237e-02 1.10097373e+00 -8.15125108e-01
5.58058262e-01 1.41207969e+00 5.77695727e-01 3.15147877e-01
-9.83744800e-01 -4.98768866e-01 1.51295662e-01 4.51849520e-01
-1.08506238e+00 -6.72484517e-01 5.27125776e-01 -1.69999301e-02
1.70449054e+00 4.37347814e-02 2.97384024e-01 9.73899662e-01
4.33458626e-01 1.22182167e+00 7.34159529e-01 -6.64920628e-01
-5.31764962e-02 7.12920055e-02 2.36014009e-01 5.02693832e-01
1.31557345e-01 -1.68017112e-02 -3.86042893e-01 1.87332183e-01
2.47002929e-01 4.77133840e-02 -1.95037931e-01 1.03645876e-01
-9.30863857e-01 9.53204453e-01 3.85044754e-01 5.10483027e-01
-3.31977487e-01 2.48927847e-01 7.72954226e-01 2.29517311e-01
7.33712375e-01 2.82276511e-01 -4.40406144e-01 -4.20996726e-01
-1.35760057e+00 2.96820402e-01 8.54596972e-01 1.28714621e+00
5.65916717e-01 -2.22166926e-01 -7.16910005e-01 9.57026780e-01
2.86095321e-01 3.05765420e-01 1.02773654e+00 -5.71847618e-01
1.25963354e+00 6.34541214e-01 -4.96425211e-01 -7.22814202e-01
2.66108080e-03 -6.64958060e-01 -1.14256346e+00 -5.90626836e-01
-1.31156549e-01 2.14070648e-01 -1.17858624e+00 1.49574566e+00
-2.05955476e-01 -5.26164286e-02 1.80013448e-01 3.63410681e-01
7.22367287e-01 1.31575036e+00 2.68476188e-01 -6.79748580e-02
1.40399301e+00 -1.33532178e+00 -1.05220926e+00 -4.18903917e-01
1.12197113e+00 -9.07645464e-01 1.10133457e+00 1.18449606e-01
-1.34429908e+00 -5.85544109e-01 -1.13165307e+00 -8.15510213e-01
-6.57683313e-01 1.80942893e-01 3.20960969e-01 2.57953823e-01
-9.63271081e-01 7.65586376e-01 -7.92022228e-01 -2.77555227e-01
4.62519586e-01 9.59227085e-02 -3.34759206e-01 -4.71398771e-01
-1.28680909e+00 1.06292665e+00 7.95580864e-01 -2.91024804e-01
-6.24344587e-01 -7.72023797e-01 -1.02563441e+00 6.52512312e-01
1.40744835e-01 -8.86407912e-01 1.48169458e+00 -5.50139010e-01
-1.34199131e+00 5.00244439e-01 -5.44989526e-01 -9.26160812e-01
1.26613706e-01 -4.59511250e-01 -9.18614957e-03 2.54728734e-01
-1.86743233e-02 9.03798819e-01 7.25537121e-01 -8.78593802e-01
-5.64524710e-01 -2.47808710e-01 -5.61349131e-02 4.01926726e-01
-5.38242340e-01 5.09356037e-02 -7.00963020e-01 -1.01761711e+00
-2.77336746e-01 -4.13165659e-01 -2.08378024e-02 -1.56565160e-01
-2.48307705e-01 -4.08602476e-01 1.01860929e+00 -9.68337595e-01
1.74294972e+00 -2.06046391e+00 2.59005249e-01 -3.09445441e-01
-2.58405264e-02 5.26484847e-01 -3.51384252e-01 8.74865532e-01
1.00183129e-01 2.98497170e-01 -4.69179481e-01 -9.48376179e-01
1.85102761e-01 4.14461851e-01 -7.97200382e-01 -1.47571620e-02
1.80612519e-01 1.14530575e+00 -5.90868592e-01 -6.17965281e-01
2.30400443e-01 6.74251080e-01 -7.92088270e-01 3.78837973e-01
-3.95718575e-01 -2.91548252e-01 -1.77402690e-01 2.62465090e-01
5.38132131e-01 -1.23283781e-01 -1.42508194e-01 -3.24294418e-01
1.20683216e-01 1.16527891e+00 -6.15690887e-01 2.03269267e+00
-6.38426900e-01 6.59046113e-01 -1.19807281e-01 -1.31384516e+00
7.05302715e-01 4.44195986e-01 2.06294790e-01 -9.08776224e-01
2.22675785e-01 1.08478097e-02 -2.51497239e-01 -3.37303251e-01
1.00344288e+00 -1.38561830e-01 2.86805723e-02 7.71663785e-01
4.32864368e-01 -3.34553927e-01 4.19443518e-01 5.30126393e-01
9.98398721e-01 2.54883796e-01 5.20027280e-01 6.14105314e-02
6.21743262e-01 -8.88379011e-03 -1.54210767e-02 6.37119055e-01
1.36079177e-01 5.39649129e-01 3.26122671e-01 -1.69000432e-01
-1.29690850e+00 -4.37785655e-01 -2.30576191e-03 1.26844561e+00
-2.73160636e-01 -1.05572116e+00 -8.16596329e-01 -6.60834610e-01
-9.05560628e-02 9.41162825e-01 -5.47513783e-01 -6.08656824e-01
-7.90014923e-01 -5.43438375e-01 8.92537653e-01 7.86021054e-01
7.24784613e-01 -9.93899286e-01 -2.60047764e-01 3.60700697e-01
-6.60721719e-01 -1.13141775e+00 -5.95169485e-01 4.82542962e-01
-9.78680313e-01 -4.49272484e-01 -6.24976575e-01 -7.36413658e-01
3.79075319e-01 3.69039625e-01 1.09842873e+00 4.57950197e-02
-3.04547168e-04 1.31571919e-01 -8.02618444e-01 -5.58521152e-01
-5.83741665e-01 4.85581011e-01 -5.61710238e-01 -3.15440178e-01
4.94134486e-01 -4.90476459e-01 -2.66838908e-01 -3.15652698e-01
-1.28478491e+00 4.40178066e-01 1.00448108e+00 7.96161234e-01
6.35491908e-01 1.42446190e-01 1.12670869e-01 -9.87356007e-01
8.82195950e-01 -5.16016603e-01 -9.74426866e-02 2.37597406e-01
-7.24891901e-01 6.53905928e-01 8.13835800e-01 -1.39941126e-01
-1.01961935e+00 -2.01644138e-01 -5.13464272e-01 -4.23976660e-01
-8.10524821e-02 9.55239058e-01 -8.77278671e-02 4.32139575e-01
3.41737181e-01 6.68177605e-01 -9.97841582e-02 -8.90428424e-01
5.39150119e-01 8.49452317e-01 4.33771163e-01 -7.07340658e-01
5.23825049e-01 1.17705926e-01 -1.52214602e-01 -6.00972712e-01
-9.39190328e-01 -5.51027358e-01 -6.18830800e-01 3.89553487e-01
6.63190246e-01 -1.01708996e+00 1.87073704e-02 2.00891048e-01
-1.48561299e+00 -2.01636001e-01 -5.51292121e-01 2.59107709e-01
-5.86826563e-01 6.09078526e-01 -8.32027078e-01 -4.53093499e-01
-8.80442083e-01 -1.16660035e+00 1.58196425e+00 -4.61533338e-01
-7.53001273e-02 -9.93864536e-01 3.56575325e-02 3.34214538e-01
7.16724813e-01 -5.33952951e-01 1.30932283e+00 -8.13877344e-01
-3.24222028e-01 -1.67377844e-01 -3.89404982e-01 5.18835843e-01
-1.29193723e-01 -4.62881148e-01 -1.01788509e+00 -2.88418740e-01
2.15200216e-01 -3.88150662e-01 1.44275510e+00 2.14464024e-01
1.45319951e+00 -7.15151310e-01 -2.34132096e-01 8.28603923e-01
1.42605996e+00 -1.75445639e-02 9.84142363e-01 2.53527790e-01
5.39995611e-01 6.60740554e-01 4.84855622e-01 3.05793315e-01
4.90478694e-01 6.21262014e-01 3.66498172e-01 1.21457860e-01
-3.74460131e-01 -7.25321233e-01 7.57386386e-01 1.34734595e+00
2.29707003e-01 -5.93827665e-01 -6.66460037e-01 3.71903569e-01
-1.75981247e+00 -9.90524709e-01 1.89198777e-01 1.71964669e+00
1.14641058e+00 6.01042360e-02 -4.57969308e-01 4.91688937e-01
4.25923973e-01 2.82295167e-01 -2.04566091e-01 -6.56227291e-01
1.07139256e-02 5.31813622e-01 5.47875345e-01 5.19665241e-01
-1.04020059e+00 1.02380717e+00 6.34869862e+00 1.14678025e+00
-9.81486857e-01 1.63648903e-01 4.80946124e-01 -7.61021972e-02
-2.84414321e-01 -1.67614198e-03 -1.18382454e+00 4.19361621e-01
1.24842775e+00 -5.51168561e-01 4.21063304e-01 7.01003373e-01
1.74190588e-02 2.38748938e-01 -1.28670955e+00 7.70549119e-01
1.75070986e-01 -1.46017420e+00 8.16055000e-01 1.84177179e-02
3.16828936e-01 1.00909665e-01 -1.08516537e-01 7.50622213e-01
1.08270347e-01 -1.04617763e+00 7.55786777e-01 2.61072606e-01
9.05330896e-01 -7.08362460e-01 9.37000811e-01 6.55426502e-01
-1.40466475e+00 -2.40093563e-02 -6.38138533e-01 2.41054352e-02
3.32084745e-01 6.97262645e-01 -7.60644794e-01 6.59133434e-01
3.63848209e-01 9.42864776e-01 -6.31398380e-01 5.86682618e-01
-1.90772578e-01 6.33139849e-01 -2.80719548e-01 1.07711561e-01
4.80379343e-01 1.20831989e-01 3.48961711e-01 1.80999660e+00
4.03938383e-01 -2.08765641e-02 -9.70229208e-02 7.77912796e-01
-3.85558069e-01 3.76224995e-01 -6.56412005e-01 -4.70123321e-01
3.13431293e-01 8.81420732e-01 -2.17921034e-01 -8.13136697e-01
-5.09875953e-01 9.12172377e-01 4.40727711e-01 6.62028268e-02
-6.50315642e-01 -5.78442335e-01 2.79701948e-01 7.58279487e-02
5.02137423e-01 -1.44832626e-01 -3.01119953e-01 -1.38364291e+00
1.80357337e-01 -9.39558625e-01 4.56970572e-01 -7.36496449e-01
-9.93949533e-01 5.19317508e-01 2.62495488e-01 -9.99022067e-01
-5.01153052e-01 -4.51643914e-01 -3.34223956e-01 9.36862826e-01
-1.93909895e+00 -1.21551239e+00 4.94971797e-02 7.11199939e-01
1.07833898e+00 -1.53396919e-01 8.82272363e-01 5.04585326e-01
-1.93978861e-01 6.86942160e-01 2.15047617e-02 1.60584405e-01
6.04193628e-01 -1.09711766e+00 4.79879797e-01 9.81278300e-01
2.32231453e-01 7.19519913e-01 4.25176769e-01 -6.95134521e-01
-1.56650758e+00 -1.31662905e+00 1.47071445e+00 -5.53758070e-03
5.09825528e-01 -4.60432440e-01 -1.08934915e+00 9.25135374e-01
5.94583631e-01 -3.30566406e-01 5.03921688e-01 -1.99265376e-01
-4.88182098e-01 6.87684864e-02 -8.54998887e-01 3.66920739e-01
8.75447333e-01 -6.19604111e-01 -1.10172915e+00 2.45877266e-01
1.40304410e+00 -2.82392442e-01 -6.78749800e-01 3.48391861e-01
2.89924115e-01 -6.93526566e-01 9.08895373e-01 -4.19153482e-01
1.03514004e+00 7.32102543e-02 -3.62788320e-01 -1.25466120e+00
-4.06041861e-01 -1.79651618e-01 -5.51341295e-01 1.43465984e+00
2.85582811e-01 -2.99415678e-01 5.35701036e-01 2.01701120e-01
-5.76881886e-01 -7.87414789e-01 -8.70744824e-01 -5.51587820e-01
3.42572391e-01 -6.38543785e-01 7.43715703e-01 6.85393095e-01
1.57494053e-01 7.09610820e-01 -3.08514506e-01 -3.99796546e-01
2.00442761e-01 -2.04741150e-01 6.31602228e-01 -8.59826326e-01
-4.90735888e-01 -7.11538911e-01 -5.15370592e-02 -1.55427933e+00
2.32520267e-01 -1.47028470e+00 -4.24598763e-03 -1.82703412e+00
3.58348310e-01 -2.02515900e-01 -6.23886958e-02 6.37100041e-01
-1.19193554e-01 -2.40375385e-01 3.67689759e-01 2.45560363e-01
-3.23143721e-01 9.67206538e-01 1.12197614e+00 -5.56197286e-01
1.04626417e-01 -3.36150438e-01 -8.99559081e-01 2.64181912e-01
6.17130756e-01 -4.82223630e-01 -5.96730173e-01 -1.09217381e+00
7.47005045e-02 -4.52651121e-02 -6.91053495e-02 -9.86176908e-01
5.99135160e-01 1.68382391e-01 1.82962686e-01 -8.27776492e-01
4.27317113e-01 -9.62794781e-01 -3.27081144e-01 4.73067343e-01
-8.24559927e-01 1.52886555e-01 3.86784136e-01 3.26582998e-01
-7.20456481e-01 -3.99299264e-01 7.71197319e-01 -3.35825920e-01
-2.27748081e-01 3.95877331e-01 -2.72509575e-01 6.77520260e-02
5.41080058e-01 1.21315278e-01 -3.72567892e-01 -4.92475212e-01
-3.05076003e-01 9.45403874e-02 8.48388448e-02 4.88750368e-01
7.95286715e-01 -1.13403666e+00 -8.43146205e-01 4.91011888e-01
5.93712069e-02 8.10671002e-02 -7.60107562e-02 6.78194463e-01
-4.15657848e-01 1.08490157e+00 4.59605604e-02 -3.31027418e-01
-1.14902675e+00 7.30783343e-01 -4.53010686e-02 -1.04658031e+00
-8.19575548e-01 5.16426861e-01 2.30207041e-01 -1.48957774e-01
5.28971970e-01 -7.92057753e-01 -4.42239940e-01 -1.90328322e-02
8.53091478e-01 1.99628904e-01 3.45094979e-01 -5.90192199e-01
1.41989276e-01 4.95287925e-01 -4.98459101e-01 -2.41203249e-01
1.48759472e+00 -2.28793085e-01 -3.55488628e-01 1.79769993e-01
1.53438914e+00 -3.23066592e-01 -5.33000529e-01 -5.54116011e-01
5.25019318e-03 -2.15952441e-01 3.83621156e-01 -7.66133189e-01
-1.02877605e+00 1.54520535e+00 1.03111841e-01 -1.74382836e-01
1.36149228e+00 -2.23729521e-01 1.18790710e+00 6.06903851e-01
1.97183117e-01 -1.05070484e+00 -2.27743015e-01 1.10002732e+00
9.68444169e-01 -8.66431952e-01 1.45255178e-01 -4.03069615e-01
-5.03655076e-01 1.20772207e+00 3.12469304e-01 1.20359614e-01
5.71343839e-01 5.00434279e-01 -5.61371803e-01 3.01494752e-03
-1.35765421e+00 -8.60258788e-02 3.82389963e-01 3.85373056e-01
7.47410774e-01 -2.74836481e-01 -5.14160872e-01 5.48042238e-01
-3.56367916e-01 1.11047462e-01 1.35077745e-01 1.17663670e+00
-7.42338955e-01 -1.48279643e+00 -9.45772454e-02 7.06060350e-01
-6.03332579e-01 -6.54584050e-01 -3.16325217e-01 5.68251550e-01
-5.22978827e-02 6.42049611e-01 1.80697650e-01 -3.52220833e-01
1.63116410e-01 4.06175315e-01 4.37336802e-01 -7.63366759e-01
-8.62414658e-01 4.10252772e-02 1.00984186e-01 -5.77253103e-01
-2.87961304e-01 -4.15193379e-01 -1.20598924e+00 -4.73384380e-01
-4.81588617e-02 2.30292529e-01 7.65511632e-01 1.01583564e+00
5.78208089e-01 5.51814735e-01 1.90255716e-01 -9.22885120e-01
-6.17238641e-01 -1.46494281e+00 -2.24664435e-01 3.57575387e-01
2.72473425e-01 -2.64642596e-01 -2.94107020e-01 2.50462383e-01] | [12.000572204589844, 9.19987678527832] |
968593ed-fd09-4eb9-8131-42bedf3e7794 | isotachophoresis-applied-to-chemical | 1708.08298 | null | http://arxiv.org/abs/1708.08298v1 | http://arxiv.org/pdf/1708.08298v1.pdf | Isotachophoresis applied to chemical reactions | This review discusses research developments and applications of
isotachophoresis (ITP) to the initiation, control, and acceleration of chemical
reactions, emphasizing reactions involving biomolecular reactants such as
nucleic acids, proteins, and live cells. ITP is a versatile technique which
requires no specific geometric design or material, and is compatible with a
wide range of microfluidic and automated platforms. Though ITP has
traditionally been used as a purification and separation technique, recent
years have seen its emergence as a method to automate and speed up chemical
reactions. ITP has been used to demonstrate up to 14,000-fold acceleration of
nucleic acid assays, and has been used to enhance lateral flow and other
immunoassays, and even whole bacterial cell detection assays. We here classify
these studies into two categories: homogeneous (all reactants in solution) and
heterogeneous (at least one reactant immobilized on a solid surface) assay
configurations. For each category, we review and describe physical modeling and
scaling of ITP-aided reaction assays, and elucidate key principles in ITP assay
design. We summarize experimental advances, and identify common threads and
approaches which researchers have used to optimize assay performance. Lastly,
we propose unaddressed challenges and opportunities that could further improve
these applications of ITP. | [] | 2017-08-28 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 4.69031602e-01 -5.79633892e-01 -3.23243588e-02 2.17491657e-01
-2.70337909e-01 -1.14443290e+00 5.49597681e-01 6.12461329e-01
-5.81699848e-01 9.85363364e-01 -1.52038753e-01 -6.25652015e-01
3.39299411e-01 -5.98618925e-01 -3.73638093e-01 -1.16147935e+00
6.52553663e-02 5.34074843e-01 4.09844011e-01 8.10759962e-02
5.81593513e-01 1.25900936e+00 -1.10491514e+00 5.23684584e-02
4.02619928e-01 6.07452989e-01 1.02448173e-01 1.06728339e+00
-5.79096735e-01 3.99800718e-01 -5.62814057e-01 -9.15198550e-02
-2.19442651e-01 -3.43678862e-01 -3.55781466e-01 -6.45358205e-01
-4.97232437e-01 1.55968010e-01 2.19281882e-01 2.31587067e-01
6.60947978e-01 -8.57163519e-02 9.76032615e-01 -9.05125082e-01
-3.68657202e-01 -7.82601088e-02 -5.21829188e-01 4.93430272e-02
4.47192043e-01 2.94986457e-01 2.73749560e-01 -1.23550749e+00
7.14284360e-01 1.31623280e+00 6.18773341e-01 6.62606359e-01
-1.45283914e+00 -6.55332327e-01 -3.31514359e-01 -5.85007370e-01
-9.70556617e-01 -6.90921545e-01 2.49860376e-01 -5.45553565e-01
1.29012716e+00 3.66439253e-01 7.84037590e-01 7.47358620e-01
9.17178988e-01 3.86349797e-01 1.14340341e+00 -1.77960575e-01
4.60505456e-01 1.53074875e-01 2.00209673e-02 3.39093417e-01
7.02907622e-01 -8.38111490e-02 -4.87567455e-01 -3.48712653e-01
7.59157658e-01 1.82117715e-01 5.21487296e-02 -4.26726103e-01
-1.12113738e+00 5.11422753e-01 -3.56721282e-01 4.89317477e-01
-3.82303655e-01 -2.52193883e-02 2.96269834e-01 -4.27312613e-01
1.70010120e-01 1.07947588e+00 -2.35602960e-01 -4.99492913e-01
-2.87898988e-01 2.11298123e-01 1.06657326e+00 4.11302954e-01
4.27660882e-01 -2.58140713e-01 1.52910441e-01 6.67330801e-01
2.79261321e-01 1.04080665e+00 -1.89068422e-01 -7.67994702e-01
2.49734372e-01 4.18063909e-01 1.01741099e+00 -4.75014418e-01
-9.48105693e-01 1.35918632e-01 -3.20202678e-01 1.18019834e-01
6.89878702e-01 -6.25150859e-01 -7.28272378e-01 1.15434539e+00
6.07634604e-01 -1.73219487e-01 3.57526273e-01 8.27896059e-01
5.17452061e-01 9.89760339e-01 2.01138288e-01 -7.42973030e-01
1.38465536e+00 -4.74345326e-01 -7.72723913e-01 6.29015937e-02
5.71819663e-01 -1.16549718e+00 6.56913042e-01 3.46038997e-01
-1.27509534e+00 8.00416619e-02 -1.02116799e+00 -1.87425762e-01
-8.16757917e-01 -3.29857141e-01 6.38062835e-01 1.08797991e+00
-6.81022704e-01 5.21201730e-01 -1.08570659e+00 -7.18978107e-01
-1.44598112e-01 5.03715277e-01 -9.03704613e-02 -5.89240342e-02
-6.82648897e-01 6.83851540e-01 -8.00636783e-02 2.09657386e-01
-3.81420761e-01 -6.81324124e-01 -5.53172827e-01 -1.61064848e-01
-1.20357729e-01 -4.59794790e-01 9.81940329e-01 3.81790400e-01
-2.40501738e+00 7.18794584e-01 -5.58573127e-01 -5.37688434e-02
-9.69411060e-02 -2.62835532e-01 -3.87406796e-01 1.85397610e-01
4.97521050e-02 2.04743460e-01 -2.18494356e-01 -7.95872867e-01
-3.92379761e-01 -1.84305310e-01 -5.84116459e-01 5.49098700e-02
2.08961397e-01 4.30979192e-01 2.53588296e-02 4.77537476e-02
2.86759496e-01 -9.28043067e-01 -2.73065209e-01 -2.97770798e-01
-2.26548940e-01 -2.30636746e-02 8.85066450e-01 1.17383115e-01
6.49255335e-01 -1.91710258e+00 -5.04230894e-02 3.62901032e-01
3.65870073e-02 7.16131806e-01 -5.53609617e-02 1.37170994e+00
4.14306581e-01 1.60478458e-01 8.56992677e-02 8.78262520e-02
-1.08670175e-01 -1.57628924e-01 -1.16658337e-01 6.47370815e-01
2.32357830e-01 7.90517509e-01 -1.21666396e+00 -8.98512527e-02
5.61547101e-01 8.19143653e-01 -1.46696597e-01 1.21761762e-01
1.30584523e-01 7.87830472e-01 -4.55150306e-01 9.09187794e-01
7.81109214e-01 -4.61778492e-01 5.88444769e-01 -1.33603245e-01
-1.03616416e+00 4.32026625e-01 -8.70045185e-01 1.01967371e+00
-5.31416796e-02 3.60257685e-01 2.15889454e-01 -2.60966569e-01
1.19149864e+00 3.84641320e-01 8.30366850e-01 -6.57227218e-01
5.89330234e-02 6.57480478e-01 -1.26350850e-01 -6.06970489e-01
3.49339187e-01 -6.55829906e-01 2.26063728e-01 4.24833953e-01
-4.55138981e-01 -7.88132548e-02 4.52110380e-01 8.64640027e-02
9.70862508e-01 -8.79107043e-02 3.99125487e-01 -5.47073007e-01
6.42175019e-01 5.14737424e-03 2.90551156e-01 8.33975732e-01
-3.66145492e-01 1.86825991e-01 6.48731709e-01 -5.53966045e-01
-1.00075877e+00 -9.07423556e-01 -2.00806782e-01 1.05470097e+00
5.50759017e-01 -3.02918106e-01 -3.23387861e-01 1.83479324e-01
2.58647978e-01 6.84209540e-02 -4.36519295e-01 3.41098249e-01
-4.07068968e-01 -1.12624168e+00 6.42474294e-01 4.94662791e-01
1.07700199e-01 -4.85958964e-01 -4.96442318e-01 7.91513264e-01
1.85973182e-01 -9.86033082e-01 -4.73591983e-02 1.95149273e-01
-5.37922263e-01 -1.40247953e+00 -6.06906831e-01 -3.10621172e-01
2.35329539e-01 5.66586375e-01 3.39986622e-01 -2.81143129e-01
-2.92790055e-01 5.39625227e-01 -6.63876384e-02 -7.05798864e-01
-4.59575474e-01 -7.58327916e-02 2.66462415e-01 -6.62209690e-02
7.34011471e-01 -2.25320294e-01 -1.15235233e+00 3.91592771e-01
-4.13592368e-01 -8.19524229e-02 5.90413570e-01 6.14136100e-01
4.63613868e-01 -9.13950801e-01 5.18433690e-01 -8.65236521e-01
5.30313611e-01 -2.39329487e-01 -7.23939359e-01 2.33487248e-01
-5.51325142e-01 2.08109021e-01 4.67558116e-01 -3.60135823e-01
-1.21312106e+00 -8.77149701e-02 -3.04368883e-01 4.53103602e-01
-9.49807018e-02 3.31670284e-01 -3.84258889e-02 -4.50104892e-01
6.73738241e-01 1.68318972e-01 5.20990908e-01 -1.62699491e-01
-7.62231648e-02 3.59290034e-01 2.91877165e-02 -7.74481595e-01
3.74654710e-01 7.79083014e-01 6.52933300e-01 -1.14823806e+00
-1.26549080e-01 -8.48628581e-01 -4.62158993e-02 -1.53710648e-01
9.25878227e-01 -4.02058601e-01 -1.49871755e+00 6.08202815e-01
-1.07384861e+00 -4.73709345e-01 3.01947802e-01 8.76340330e-01
-1.99467298e-02 6.91074505e-02 -1.24056387e+00 -1.31222057e+00
-3.23938191e-01 -1.40049899e+00 9.24113095e-01 4.56606418e-01
-2.76349843e-01 -9.68379438e-01 5.46298087e-01 8.52286583e-04
7.05715239e-01 6.82252645e-01 5.70696414e-01 -2.57454455e-01
-4.93140280e-01 -2.93843687e-01 -2.13189498e-01 -7.20929503e-01
4.41552520e-01 6.05561376e-01 -8.75498831e-01 -4.25586104e-01
-1.77325830e-01 -5.51403090e-02 5.37155211e-01 5.69664896e-01
1.87692150e-01 1.17181011e-01 -1.13021994e+00 3.81273299e-01
1.32188785e+00 8.95312309e-01 9.28180218e-01 2.21867666e-01
5.24051368e-01 5.80721080e-01 6.48152828e-01 3.25204045e-01
-4.03725713e-01 3.45393121e-01 6.06857873e-02 -1.84984744e-01
3.29784036e-01 -8.16021711e-02 2.35333934e-01 4.37920004e-01
-4.02727574e-01 -3.71340275e-01 -8.46609771e-01 -1.38323888e-01
-1.27262664e+00 -7.40142584e-01 -6.64119065e-01 2.25129342e+00
6.43151224e-01 -2.56616533e-01 4.34804857e-01 -1.77167177e-01
7.33437538e-01 -3.18461984e-01 -6.02857709e-01 -5.60919225e-01
-1.57280028e-01 3.41471463e-01 7.59078324e-01 7.58755922e-01
-7.81489134e-01 7.64595389e-01 8.49637127e+00 1.26883090e-01
-1.55619574e+00 -2.54359603e-01 3.33821267e-01 -1.35328427e-01
2.07196106e-03 3.41004990e-02 -1.23064530e+00 2.17412651e-01
9.03482199e-01 9.92729291e-02 -1.26224142e-02 1.78043768e-01
4.29515421e-01 -4.58908111e-01 -1.06517208e+00 6.06642127e-01
-6.01976871e-01 -1.82394707e+00 -4.97600734e-02 2.58743137e-01
4.76690173e-01 -3.12127054e-01 -7.18804775e-03 -1.69834003e-01
2.52821879e-03 -5.77045918e-01 7.68677518e-02 1.30390361e-01
6.72610104e-01 -5.28139174e-01 7.06552148e-01 -1.07696109e-01
-1.42751133e+00 1.17848873e-01 -2.43160293e-01 -5.12079418e-01
1.57870814e-01 6.65842175e-01 -1.21400714e+00 -8.15278739e-02
3.11251312e-01 2.82267600e-01 2.43383333e-01 7.22712874e-01
2.29613721e-01 3.05146426e-01 -3.72211426e-01 -9.13461566e-01
3.53802532e-01 -7.58238196e-01 2.40825608e-01 1.57471633e+00
1.30909368e-01 6.07987344e-01 -2.80666918e-01 4.84766603e-01
3.22779983e-01 3.59974593e-01 -5.69452584e-01 -4.03366446e-01
8.03949177e-01 1.19610786e+00 -1.42750251e+00 -4.97542620e-01
-4.98629063e-01 3.64602119e-01 -3.42165306e-02 7.94000685e-01
-6.01536334e-01 -7.81416833e-01 1.00440776e+00 2.58430034e-01
-4.98804152e-02 -5.03923833e-01 -1.81790248e-01 -8.27752173e-01
-5.24872303e-01 -1.91693678e-01 1.05848260e-01 -2.49399990e-01
-6.08416319e-01 -5.47750771e-01 -5.08945107e-01 -6.51347756e-01
2.96082348e-01 -1.02265525e+00 -2.13167354e-01 9.40351784e-01
-1.25876069e+00 -4.45413470e-01 -7.62613118e-02 -1.74504668e-02
4.05469127e-02 3.65170687e-01 8.60803783e-01 5.46474867e-02
-7.52373636e-01 -2.14241259e-02 7.76875377e-01 -2.17298165e-01
1.08863044e+00 -7.88324296e-01 2.78746039e-01 7.27455854e-01
-8.01588833e-01 1.21571434e+00 8.15333307e-01 -8.38955462e-01
-2.09912252e+00 -7.30287671e-01 6.64223313e-01 -5.25861621e-01
4.19068962e-01 -6.24126792e-01 -7.20969021e-01 3.25843364e-01
-5.55490479e-02 -4.73869026e-01 1.37821853e+00 6.63573220e-02
-1.90113142e-01 -7.15434849e-02 -1.18278134e+00 7.59335995e-01
6.01687074e-01 -4.12342042e-01 1.30638227e-01 3.30704093e-01
2.37217605e-01 -7.07269967e-01 -1.13405633e+00 1.48588315e-01
1.28976440e+00 -8.44966233e-01 8.76499116e-01 -5.69013238e-01
-1.36456057e-01 -5.97912729e-01 3.33222091e-01 -6.24558032e-01
-6.90063834e-01 -1.07826972e+00 1.13043897e-01 7.55303741e-01
4.73863542e-01 -1.32952988e+00 8.18767488e-01 1.18771887e+00
-1.59587905e-01 -5.09302437e-01 -4.21053559e-01 -7.04020083e-01
1.04871392e-01 2.73497939e-01 5.54221690e-01 6.85113132e-01
9.15309727e-01 3.58557224e-01 -8.41567293e-02 1.80509076e-01
3.56898725e-01 -3.00383326e-02 7.66911566e-01 -9.67972457e-01
4.17432003e-02 -4.46484327e-01 -1.04350947e-01 -9.65585768e-01
-5.72148561e-01 -2.43604183e-01 1.09982463e-02 -1.53160465e+00
-3.11417431e-02 -3.57818693e-01 -3.22233647e-01 -4.85374555e-02
-9.07594413e-02 1.78057224e-01 -1.15960941e-01 2.87903547e-01
-4.69136000e-01 -1.29114881e-01 1.12383163e+00 1.52309284e-01
-7.23368943e-01 -3.82006139e-01 -5.19530177e-01 -7.73182362e-02
8.55736136e-01 -3.85578632e-01 -2.29264855e-01 4.97858040e-02
4.62762654e-01 6.52108192e-02 -1.85709760e-01 -9.04612124e-01
1.56794637e-01 -5.70748985e-01 5.68067610e-01 -4.71102566e-01
3.92717570e-01 -3.53540540e-01 6.64752722e-01 1.06640613e+00
-1.48915444e-02 -2.27841839e-01 3.97619784e-01 3.39784831e-01
2.96057284e-01 2.29613170e-01 9.16247129e-01 -7.57765546e-02
-4.94575463e-02 -7.64342919e-02 -1.36023581e+00 -2.66636521e-01
1.17294931e+00 -4.76795912e-01 -1.00263929e+00 3.14764768e-01
-4.49817538e-01 2.09713191e-01 7.67880499e-01 -2.46223852e-01
3.72499645e-01 -8.99364293e-01 -9.67497230e-02 2.61921525e-01
-4.69831079e-02 -3.11908990e-01 3.31813335e-01 1.19015777e+00
-1.27236474e+00 1.03910792e+00 -1.45721823e-01 -7.75613964e-01
-1.12471068e+00 8.96724105e-01 4.08203542e-01 -5.03353216e-02
-1.16376869e-01 5.60161948e-01 4.23449785e-01 9.72979516e-02
-2.55911797e-01 -3.81955177e-01 4.09481004e-02 -3.27195406e-01
1.00203145e+00 6.92428410e-01 -3.51028070e-02 4.56020758e-02
-6.76516175e-01 7.23401725e-01 -1.86025664e-01 -1.68373495e-01
1.00831091e+00 -2.02746075e-02 -2.14499071e-01 6.43448770e-01
8.93758059e-01 1.75870001e-01 -1.29870450e+00 1.16403989e-01
-1.17630847e-01 -2.91500669e-02 -7.81844079e-01 -3.37351143e-01
-3.15634072e-01 8.93570721e-01 1.77881733e-01 2.60066688e-01
4.91824836e-01 -7.07074031e-02 5.46311498e-01 5.86632252e-01
2.84475565e-01 -1.17726946e+00 2.09205091e-01 3.63750488e-01
5.62290788e-01 -6.97234035e-01 -3.07162106e-03 -6.96791589e-01
-3.97803374e-02 1.29756272e+00 3.23361576e-01 2.00784549e-01
3.45641017e-01 7.71796882e-01 1.11011721e-01 -3.01391035e-01
-6.96912050e-01 1.62140459e-01 -6.06631339e-01 7.64048100e-01
1.07608581e+00 7.31331781e-02 -7.23110974e-01 -1.71149611e-01
5.80302835e-01 2.52056262e-03 6.65855169e-01 1.44677198e+00
-6.70431435e-01 -1.30820119e+00 -7.19017684e-01 2.35616192e-01
-3.93290311e-01 2.02204108e-01 -6.96216762e-01 7.08202004e-01
-5.94774663e-01 1.10637867e+00 -2.15192020e-01 -1.11694574e-01
2.62920946e-01 5.39895594e-01 5.10684192e-01 -3.25767368e-01
-3.32956731e-01 4.27114606e-01 3.25596213e-01 -6.31837785e-01
-6.20374858e-01 -8.40498984e-01 -1.52597106e+00 -6.31015360e-01
-2.36975417e-01 7.10496008e-01 7.75126219e-01 9.00990129e-01
7.18575299e-01 7.73587748e-02 3.40259224e-01 -9.83940959e-01
4.71449584e-01 -5.20211220e-01 -7.92482853e-01 -6.75055161e-02
3.46782088e-01 -7.59039938e-01 -1.95839599e-01 -3.52196783e-01] | [13.797813415527344, -3.1168060302734375] |
5d26ffb8-af31-40c3-b714-46757ca62eae | fine-grained-3d-shape-classification-with | 2005.12541 | null | https://arxiv.org/abs/2005.12541v2 | https://arxiv.org/pdf/2005.12541v2.pdf | Fine-Grained 3D Shape Classification with Hierarchical Part-View Attentions | Fine-grained 3D shape classification is important for shape understanding and analysis, which poses a challenging research problem. However, the studies on the fine-grained 3D shape classification have rarely been explored, due to the lack of fine-grained 3D shape benchmarks. To address this issue, we first introduce a new 3D shape dataset (named FG3D dataset) with fine-grained class labels, which consists of three categories including airplane, car and chair. Each category consists of several subcategories at a fine-grained level. According to our experiments under this fine-grained dataset, we find that state-of-the-art methods are significantly limited by the small variance among subcategories in the same category. To resolve this problem, we further propose a novel fine-grained 3D shape classification method named FG3D-Net to capture the fine-grained local details of 3D shapes from multiple rendered views. Specifically, we first train a Region Proposal Network (RPN) to detect the generally semantic parts inside multiple views under the benchmark of generally semantic part detection. Then, we design a hierarchical part-view attention aggregation module to learn a global shape representation by aggregating generally semantic part features, which preserves the local details of 3D shapes. The part-view attention module hierarchically leverages part-level and view-level attention to increase the discriminability of our features. The part-level attention highlights the important parts in each view while the view-level attention highlights the discriminative views among all the views of the same object. In addition, we integrate a Recurrent Neural Network (RNN) to capture the spatial relationships among sequential views from different viewpoints. Our results under the fine-grained 3D shape dataset show that our method outperforms other state-of-the-art methods. | ['Matthias Zwicker', 'Yu-Shen Liu', 'Xinhai Liu', 'Zhizhong Han'] | 2020-05-26 | null | null | null | null | ['semantic-part-detection', '3d-shape-retrieval'] | ['computer-vision', 'computer-vision'] | [-3.43630642e-01 -1.97423771e-01 -4.60019708e-02 -4.56841856e-01
-6.00390971e-01 -6.13781631e-01 6.26367986e-01 -1.82882205e-01
4.49543029e-01 1.52543671e-02 5.94756782e-01 9.92703959e-02
-9.72743481e-02 -9.62329507e-01 -6.56088114e-01 -7.46485472e-01
3.09990674e-01 5.44744015e-01 4.22188610e-01 -4.72442955e-02
1.68434054e-01 1.12651837e+00 -1.49473238e+00 6.53010011e-01
4.53145742e-01 1.36885047e+00 1.09099098e-01 1.12806559e-01
-3.96918505e-01 1.30702853e-01 -4.47090954e-01 -9.89844576e-02
3.77218813e-01 1.75942201e-02 -5.22702336e-01 5.92710912e-01
6.44143701e-01 -4.73722637e-01 -2.70380408e-01 9.90495563e-01
5.71447611e-01 4.44908589e-02 7.25041866e-01 -1.20322013e+00
-1.07395220e+00 5.10449968e-02 -7.49464095e-01 3.44629399e-02
6.89027011e-02 1.40233800e-01 1.14507890e+00 -1.28792036e+00
5.26922822e-01 1.81786001e+00 3.75275314e-01 4.00194645e-01
-9.64385271e-01 -5.00653207e-01 7.95259833e-01 -2.00115219e-01
-1.33664203e+00 -1.41297877e-01 1.16146815e+00 -4.18593019e-01
9.52090740e-01 3.98768112e-02 6.31278038e-01 9.30671751e-01
1.01664424e-01 8.89829636e-01 1.06516111e+00 1.76747605e-01
-4.26115887e-03 -1.72510564e-01 2.02279305e-03 8.47140074e-01
2.38865659e-01 8.92104581e-02 -1.19093373e-01 -5.27833216e-02
1.22178245e+00 5.81584096e-01 -1.32266745e-01 -8.12226832e-01
-1.11645520e+00 5.90972066e-01 7.29734838e-01 3.68350029e-01
-3.62307131e-01 -1.05186500e-01 2.75363088e-01 -2.42561027e-02
8.93369436e-01 3.37609351e-02 -6.56472385e-01 2.54488766e-01
-4.79811192e-01 2.24424496e-01 5.07439494e-01 1.10860384e+00
6.99041188e-01 2.55944163e-01 -4.22536612e-01 1.00973070e+00
5.26979446e-01 5.08808672e-01 1.29080877e-01 -7.53485799e-01
5.30508757e-01 1.39558125e+00 -2.23775625e-01 -1.01415586e+00
-3.60105544e-01 -5.87613940e-01 -1.07029557e+00 2.84334391e-01
9.86689106e-02 4.31877375e-01 -1.13237190e+00 1.58571076e+00
5.61905205e-01 -1.94789410e-01 -3.34837079e-01 1.07702553e+00
1.37341142e+00 4.71709728e-01 -1.31155372e-01 2.11932659e-01
1.51569533e+00 -1.04770458e+00 -8.89408216e-02 1.11450076e-01
1.12208351e-01 -5.06528378e-01 1.09840918e+00 -3.02015603e-01
-1.08571303e+00 -8.33813369e-01 -7.65585423e-01 -2.20292076e-01
-5.92967987e-01 1.56430170e-01 3.23192567e-01 3.72556806e-01
-8.20830226e-01 2.74997443e-01 -6.07609808e-01 -2.63931304e-01
8.35089564e-01 7.60547072e-02 -5.75458586e-01 -2.12222815e-01
-5.98595560e-01 4.31927979e-01 -6.15155417e-03 1.14567131e-02
-9.24202025e-01 -7.52779961e-01 -1.02601182e+00 2.64724821e-01
4.31605458e-01 -9.44006681e-01 8.51634979e-01 -4.78424370e-01
-1.25580823e+00 1.24990380e+00 -2.24940002e-01 3.25201154e-01
1.34100497e-01 8.83775130e-02 -1.46564439e-01 -3.02880649e-02
3.46981347e-01 6.73286736e-01 9.90788281e-01 -1.62962341e+00
-4.65905845e-01 -8.06075454e-01 2.17735186e-01 2.54403561e-01
1.95905671e-01 -1.95295721e-01 -7.29364038e-01 -9.93194640e-01
6.22325718e-01 -6.68211937e-01 -9.73226875e-03 1.51205719e-01
-4.10551906e-01 -5.75487912e-01 1.03010583e+00 -2.80323535e-01
7.41879404e-01 -2.39295697e+00 1.48763761e-01 -4.28852662e-02
4.07422692e-01 -3.04049645e-02 -3.33009481e-01 1.78356960e-01
-7.82260820e-02 2.64437556e-01 -2.82272045e-02 -3.20074946e-01
2.32566401e-01 2.43073732e-01 -3.99647087e-01 1.61406964e-01
4.21050608e-01 1.32247519e+00 -5.86474299e-01 -5.84200583e-02
1.70707941e-01 4.59382534e-01 -5.15743315e-01 5.05427361e-01
-1.59616545e-01 3.47473890e-01 -8.96264017e-01 1.15520036e+00
9.35774088e-01 -5.56985795e-01 -3.39216292e-01 -7.37373471e-01
2.73646489e-02 1.37434199e-01 -1.03113770e+00 1.58616817e+00
-2.46783078e-01 -1.21055171e-01 1.12597540e-01 -8.72052372e-01
1.23695421e+00 1.57657698e-01 5.13366997e-01 -6.08019531e-01
7.51165673e-02 2.53943563e-01 -1.62190393e-01 -1.74440160e-01
1.85036972e-01 -2.04133615e-01 -1.90720156e-01 6.07003748e-01
6.69063255e-02 -3.80882621e-01 -4.97446418e-01 -1.29044458e-01
7.81888843e-01 3.20709705e-01 4.33150440e-01 -1.13879986e-01
4.83488262e-01 -4.40104306e-01 5.70906997e-01 5.22056937e-01
-3.71864140e-01 9.75959003e-01 2.79822230e-01 -7.67824411e-01
-9.54839766e-01 -1.18266904e+00 -5.13119176e-02 1.10536301e+00
4.46975678e-01 -1.06551565e-01 -3.91012192e-01 -1.21716082e+00
5.46471298e-01 1.57555625e-01 -7.73326874e-01 -5.07271700e-02
-5.09684086e-01 -4.06604469e-01 1.51098922e-01 8.90820384e-01
7.22458422e-01 -1.26048338e+00 -3.37549239e-01 1.47684347e-02
2.39060476e-01 -1.12646806e+00 -8.41114283e-01 8.04254264e-02
-9.65361059e-01 -1.12751770e+00 -8.77635121e-01 -7.43983030e-01
8.34345162e-01 7.07233906e-01 1.31006122e+00 -4.36823033e-02
-4.00444418e-02 5.94114065e-01 -3.55943203e-01 -2.31737763e-01
7.20293969e-02 -1.00195467e-01 -1.06620669e-01 2.94051558e-01
3.02080333e-01 -6.64712429e-01 -5.73896945e-01 5.32994092e-01
-6.55005753e-01 -9.40535590e-02 6.91597462e-01 9.21514690e-01
1.06419039e+00 3.29829566e-02 4.79555100e-01 -7.17654049e-01
2.99992830e-01 -3.06937546e-01 -3.90408278e-01 2.07976073e-01
-7.66549110e-02 4.02607508e-02 5.22884727e-01 -1.70095161e-01
-9.25873876e-01 -1.70859128e-01 -1.95798457e-01 -9.48158383e-01
-7.39740789e-01 -2.82353871e-02 -8.17142844e-01 -6.84774593e-02
-6.45879805e-02 4.57397014e-01 -2.04080701e-01 -8.14577162e-01
3.29021811e-01 5.82230747e-01 1.32414445e-01 -8.09208989e-01
7.97420621e-01 6.00087285e-01 -4.81908247e-02 -5.36673844e-01
-1.15484703e+00 -3.70879352e-01 -8.03310215e-01 -1.74920238e-03
1.05716562e+00 -1.02200043e+00 -7.32822776e-01 5.57142019e-01
-1.08386087e+00 -1.19322864e-02 -4.82944071e-01 4.66751568e-02
-6.29834294e-01 2.43458152e-01 -5.45144737e-01 -5.41783333e-01
-3.45945746e-01 -1.28346348e+00 1.91274297e+00 1.19633205e-01
1.29498035e-01 -7.37125576e-01 -2.78054893e-01 4.46293324e-01
2.15819418e-01 2.67728329e-01 1.33331311e+00 -5.68316579e-01
-8.41096282e-01 1.07952334e-01 -7.51927435e-01 2.30048016e-01
4.44308460e-01 -2.85753965e-01 -9.70886230e-01 -3.31335992e-01
4.58745472e-02 -2.39207819e-01 8.76373351e-01 4.48000669e-01
1.56431353e+00 -1.62362859e-01 -2.99173683e-01 7.24354208e-01
1.13118494e+00 1.91609830e-01 1.27339110e-01 -8.89828429e-02
1.11608791e+00 6.22709990e-01 4.33150619e-01 3.58330905e-01
5.05420327e-01 7.09553301e-01 6.81259751e-01 -1.19243428e-01
-3.58780682e-01 -5.48262298e-01 -6.07759804e-02 7.56379664e-01
-2.19709054e-01 -1.78554237e-01 -4.69256371e-01 5.02130508e-01
-1.61985636e+00 -8.61122906e-01 3.29971313e-01 1.85229850e+00
2.22820148e-01 7.37176090e-02 2.06981301e-01 -2.88004160e-01
7.19680250e-01 6.95209086e-01 -8.13549459e-01 -1.75664410e-01
-1.84505090e-01 -1.07429542e-01 -5.00230864e-02 8.97775888e-02
-1.26280999e+00 8.30011308e-01 5.09164047e+00 1.03659332e+00
-9.93274450e-01 -3.97813618e-02 8.15894186e-01 2.66387522e-01
-6.43139303e-01 -3.53977472e-01 -9.16700363e-01 3.41998994e-01
-7.03680962e-02 4.19886529e-01 1.00816622e-01 9.77784693e-01
3.09146754e-02 4.67620760e-01 -1.07457066e+00 1.13985741e+00
3.06781709e-01 -1.20100009e+00 5.77701271e-01 2.55445272e-01
8.74341965e-01 2.14524586e-02 -8.74686241e-02 3.49366337e-01
7.99892023e-02 -1.03454685e+00 6.78317726e-01 4.98012334e-01
8.46329510e-01 -6.54425144e-01 4.86130953e-01 4.13121432e-01
-1.82567704e+00 -1.97279789e-02 -5.88022828e-01 2.49331787e-01
8.11151415e-02 5.05213320e-01 -2.25789338e-01 7.48476565e-01
7.68574655e-01 1.10955215e+00 -3.38047683e-01 7.84387171e-01
2.85176490e-03 7.45898113e-02 -6.57602027e-02 2.45123401e-01
3.86637270e-01 -2.06632584e-01 5.78647912e-01 8.13695073e-01
4.89041239e-01 4.02583718e-01 4.75089669e-01 1.15836680e+00
-2.91005880e-01 -2.06023902e-01 -8.72279942e-01 1.93283886e-01
2.82382935e-01 1.24252152e+00 -7.67256021e-01 -3.18156809e-01
-7.04830885e-01 9.39178467e-01 4.39399451e-01 3.84636015e-01
-4.35279220e-01 3.08637675e-02 8.94061446e-01 2.18405321e-01
8.22545767e-01 -2.13581532e-01 -3.52289110e-01 -1.33832276e+00
1.68566808e-01 -8.01475108e-01 4.57688063e-01 -9.56578791e-01
-1.84638989e+00 6.89477682e-01 -1.34720266e-01 -1.38118827e+00
3.38530451e-01 -8.48106623e-01 -5.68217516e-01 8.99050057e-01
-1.38582921e+00 -1.55324066e+00 -4.86531109e-01 5.12158990e-01
8.69248033e-01 -2.94993907e-01 7.81194925e-01 7.11905137e-02
-3.16862494e-01 4.86190587e-01 -3.59717965e-01 2.57858574e-01
3.94282460e-01 -1.13989973e+00 6.53888464e-01 3.08455169e-01
2.26579338e-01 4.98327047e-01 -2.68546909e-01 -7.34104276e-01
-1.34390676e+00 -1.38029778e+00 6.86620057e-01 -6.02323174e-01
1.82774723e-01 -4.35028493e-01 -1.04356539e+00 4.62642759e-01
-1.72638103e-01 5.73998511e-01 3.24933946e-01 1.93556902e-04
-8.47845078e-01 -1.51756823e-01 -1.16550410e+00 3.38094592e-01
1.52438200e+00 -7.61848867e-01 -8.36543262e-01 -9.78132933e-02
8.76217902e-01 -4.80562598e-01 -1.05072844e+00 8.40778768e-01
6.02054477e-01 -1.08882189e+00 1.28090084e+00 -6.76587820e-01
3.80203784e-01 -4.47327554e-01 -5.90329707e-01 -1.30235445e+00
-7.91561246e-01 1.91434562e-01 -3.04464847e-01 1.18569541e+00
-5.66212758e-02 -5.60309887e-01 8.55218530e-01 2.66022861e-01
-4.89910722e-01 -1.27902639e+00 -8.48819792e-01 -7.57796109e-01
1.06669635e-01 -2.58212000e-01 1.15374756e+00 6.27447486e-01
-7.86783516e-01 3.91397923e-01 3.10569126e-02 2.30195239e-01
5.00510454e-01 1.25306630e+00 7.16626525e-01 -1.39381552e+00
-1.84786126e-01 -7.90416718e-01 -6.09615266e-01 -1.51575458e+00
2.99005300e-01 -9.64149952e-01 -2.93289989e-01 -1.56733537e+00
5.28876603e-01 -4.55423385e-01 -3.20298970e-01 5.74117780e-01
-1.73888460e-01 4.53059345e-01 3.47624928e-01 2.04409748e-01
-5.65424860e-01 9.15649116e-01 1.96950078e+00 -4.56859201e-01
-1.49883749e-02 1.95196524e-01 -8.19767118e-01 7.67934501e-01
2.82336086e-01 -1.23529583e-01 -3.09247494e-01 -4.70561296e-01
-1.26437172e-01 -5.88192157e-02 6.57036662e-01 -5.13641834e-01
-9.10577029e-02 -1.36217773e-01 7.48507857e-01 -1.27253127e+00
4.56947714e-01 -9.78011310e-01 -1.07993774e-01 9.44378525e-02
3.59931998e-02 -2.75688227e-02 1.08731471e-01 7.07823992e-01
-4.55085605e-01 3.00113559e-01 8.00118625e-01 -5.02825022e-01
-8.08815777e-01 9.79518831e-01 2.02216923e-01 3.11133981e-01
8.86643291e-01 -4.00917947e-01 -1.37206689e-01 -1.31842807e-01
-7.19736457e-01 3.76697183e-02 4.67180520e-01 7.25761175e-01
8.29784393e-01 -1.84002805e+00 -6.23764813e-01 7.40847349e-01
3.72267008e-01 3.79167765e-01 5.99708319e-01 3.86487633e-01
1.32025912e-01 4.48058665e-01 -2.96574444e-01 -9.06138539e-01
-1.10482764e+00 5.65850973e-01 4.99717891e-01 -2.09915280e-01
-7.91107655e-01 8.53351474e-01 1.14202821e+00 -9.86986995e-01
1.00555666e-01 -6.99362040e-01 -3.34103584e-01 -3.03359590e-02
1.88820496e-01 1.24920025e-01 5.12815230e-02 -9.25510108e-01
-4.97145295e-01 1.40453529e+00 9.33184102e-02 5.79947114e-01
1.53370380e+00 -1.87184840e-01 -1.13486916e-01 4.04131204e-01
1.32461655e+00 4.70279679e-02 -1.59409440e+00 -2.90284544e-01
-4.57813531e-01 -5.76061487e-01 -2.76474983e-01 -7.65286088e-01
-1.32475495e+00 1.08503652e+00 3.75705600e-01 1.73132658e-01
1.12149239e+00 3.71090591e-01 6.87330902e-01 7.92270973e-02
5.11485338e-01 -5.41670382e-01 2.16104180e-01 9.64758277e-01
1.25380361e+00 -1.31225002e+00 -1.38149574e-01 -5.36762118e-01
-5.34448326e-01 9.92480338e-01 8.55774224e-01 -2.47635275e-01
9.06560242e-01 -1.14014754e-02 -4.95390855e-02 -5.83726346e-01
-6.13035083e-01 -3.75867188e-01 8.74835193e-01 5.36029339e-01
2.02212781e-01 2.23582700e-01 4.08800185e-01 1.01260245e+00
2.00127572e-01 -5.31118155e-01 -1.42426535e-01 3.93151939e-01
-4.01509166e-01 -9.16799664e-01 -2.06844121e-01 4.75449711e-01
-1.45546183e-01 1.42762318e-01 -5.29084742e-01 7.45184124e-01
2.87379444e-01 4.31357741e-01 2.15772405e-01 -2.14241534e-01
6.54859662e-01 4.80837971e-02 4.23440158e-01 -8.13676178e-01
-5.78845084e-01 2.53627121e-01 -2.94322282e-01 -8.08974028e-01
-3.71128559e-01 -4.26008254e-01 -1.01635361e+00 -7.24599361e-02
1.14275999e-01 -2.59479910e-01 4.57784981e-02 8.86192977e-01
7.81856179e-01 5.79005122e-01 9.24588919e-01 -1.24662828e+00
-4.38749433e-01 -6.16887808e-01 -8.84078920e-01 7.18508005e-01
2.82877445e-01 -1.08963108e+00 -2.49422044e-01 -4.44657326e-01] | [8.108911514282227, -3.7533674240112305] |
ca9cd343-b149-4148-90b4-a9d715bd5fb7 | from-adversarial-arms-race-to-model-centric | 2305.18503 | null | https://arxiv.org/abs/2305.18503v1 | https://arxiv.org/pdf/2305.18503v1.pdf | From Adversarial Arms Race to Model-centric Evaluation: Motivating a Unified Automatic Robustness Evaluation Framework | Textual adversarial attacks can discover models' weaknesses by adding semantic-preserved but misleading perturbations to the inputs. The long-lasting adversarial attack-and-defense arms race in Natural Language Processing (NLP) is algorithm-centric, providing valuable techniques for automatic robustness evaluation. However, the existing practice of robustness evaluation may exhibit issues of incomprehensive evaluation, impractical evaluation protocol, and invalid adversarial samples. In this paper, we aim to set up a unified automatic robustness evaluation framework, shifting towards model-centric evaluation to further exploit the advantages of adversarial attacks. To address the above challenges, we first determine robustness evaluation dimensions based on model capabilities and specify the reasonable algorithm to generate adversarial samples for each dimension. Then we establish the evaluation protocol, including evaluation settings and metrics, under realistic demands. Finally, we use the perturbation degree of adversarial samples to control the sample validity. We implement a toolkit RobTest that realizes our automatic robustness evaluation framework. In our experiments, we conduct a robustness evaluation of RoBERTa models to demonstrate the effectiveness of our evaluation framework, and further show the rationality of each component in the framework. The code will be made public at \url{https://github.com/thunlp/RobTest}. | ['Heng Ji', 'Maosong Sun', 'Zhiyuan Liu', 'Hui Xue', 'Longtao Huang', 'Bo Yuan', 'Ning Shi', 'Hanlu Wu', 'Dehan Kong', 'Lifan Yuan', 'Ganqu Cui', 'Hongcheng Gao', 'Yangyi Chen'] | 2023-05-29 | null | null | null | null | ['adversarial-attack'] | ['adversarial'] | [ 8.36603343e-02 -3.38166565e-01 -4.10610400e-02 -1.65883467e-01
-1.06996393e+00 -1.26567876e+00 8.26625466e-01 -1.61766469e-01
-2.06567466e-01 5.25319278e-01 1.89025775e-01 -5.50062180e-01
-8.49340111e-02 -8.61958683e-01 -5.80167592e-01 -4.54699069e-01
6.88239979e-03 7.38452896e-02 8.46344903e-02 -3.39649022e-01
4.42844421e-01 7.44104564e-01 -1.05926800e+00 2.37847254e-01
7.99489319e-01 6.66274667e-01 -6.34093404e-01 8.86405528e-01
2.72202373e-01 6.63593352e-01 -7.81061947e-01 -7.49110281e-01
6.22843981e-01 -1.91200942e-01 -9.76254702e-01 -5.00439942e-01
1.18550910e-02 -3.86284262e-01 -5.04337847e-01 1.33445036e+00
7.82863975e-01 1.76576167e-01 4.41953063e-01 -1.62366176e+00
-4.95678008e-01 8.96872520e-01 -1.44127324e-01 1.70536220e-01
5.70693195e-01 9.73929763e-01 7.11454511e-01 -4.98609096e-01
4.88185793e-01 1.33342052e+00 4.57859874e-01 8.05458188e-01
-8.94733489e-01 -8.91161561e-01 2.19096556e-01 -1.99705344e-02
-1.34685898e+00 -7.96688318e-01 9.07240152e-01 -2.80388057e-01
5.88871479e-01 8.07335615e-01 -4.85517420e-02 1.66588998e+00
1.86058864e-01 7.36081123e-01 1.02810931e+00 -2.09242418e-01
4.21778262e-01 1.00925481e-02 1.80648446e-01 4.14178342e-01
5.33424392e-02 5.25752425e-01 -1.42021716e-01 -5.33499062e-01
3.44634831e-01 -2.25315884e-01 -2.22268283e-01 -6.22245260e-02
-1.06207883e+00 6.77928209e-01 2.52296448e-01 1.65662378e-01
7.16437446e-03 1.89422220e-01 7.62721360e-01 6.35728836e-01
1.95429340e-01 9.20261145e-01 -5.29774547e-01 -1.01217344e-01
-6.27891064e-01 4.32645202e-01 7.52483308e-01 8.68374765e-01
2.72714883e-01 1.74532756e-01 -5.12518167e-01 5.99546969e-01
1.74312234e-01 5.40980637e-01 4.68927801e-01 -8.92844141e-01
5.07534444e-01 4.24777627e-01 1.04295522e-01 -1.09281385e+00
-3.05310369e-01 -2.86940396e-01 -7.73949325e-01 2.24376068e-01
3.23704869e-01 -2.70629346e-01 -8.41512740e-01 1.91419697e+00
4.03661519e-01 -4.53473963e-02 3.66635650e-01 8.63130510e-01
5.15742898e-01 3.99007946e-01 1.94903508e-01 -1.36140123e-01
1.12396646e+00 -7.51518369e-01 -6.03140712e-01 -4.98784520e-03
5.60574472e-01 -7.54378259e-01 1.56603968e+00 3.43080968e-01
-1.01850021e+00 -3.37939143e-01 -1.11146629e+00 4.27168041e-01
-4.88138348e-01 -3.86785775e-01 3.45443130e-01 9.71856177e-01
-6.63580060e-01 6.52758420e-01 -7.21851647e-01 -4.55840910e-03
2.58516043e-01 1.58326134e-01 -3.44259977e-01 2.17362419e-02
-1.74259925e+00 7.92105258e-01 4.88989681e-01 1.18753135e-01
-1.23327911e+00 -6.58763945e-01 -8.78238380e-01 -3.16335440e-01
5.41049898e-01 -7.64025390e-01 1.23284566e+00 -7.23266661e-01
-1.55459046e+00 5.39272308e-01 3.01005155e-01 -3.20087254e-01
9.56460714e-01 -1.05340309e-01 -7.78452337e-01 -2.43782829e-02
-1.81216553e-01 1.66458979e-01 9.43061352e-01 -1.33710253e+00
-6.70394441e-03 -2.57591426e-01 4.50257003e-01 -1.73342019e-01
-3.23510200e-01 5.56587577e-01 -3.97584498e-01 -1.00601757e+00
-2.65868068e-01 -9.12855983e-01 -5.04447043e-01 -4.14855242e-01
-8.94083142e-01 3.51123989e-01 4.60192978e-01 -4.67310190e-01
1.58572960e+00 -2.24246716e+00 -5.54678626e-02 4.99942660e-01
4.95941974e-02 4.48469400e-01 -4.36533570e-01 6.95159614e-01
-2.41276503e-01 6.76319122e-01 -5.50183058e-01 -1.85102317e-02
3.24133128e-01 -2.30811521e-01 -6.88318253e-01 3.95066887e-01
2.69839257e-01 1.00385022e+00 -8.59485686e-01 -3.67030859e-01
8.34415928e-02 5.63059859e-02 -5.50930917e-01 3.34757239e-01
-2.20829099e-01 3.68958503e-01 -6.80338740e-01 9.57706273e-01
7.45037615e-01 2.22764522e-01 -1.67670585e-02 -7.51861259e-02
2.44471520e-01 -2.28863172e-02 -1.36360121e+00 1.43027842e+00
-3.27746242e-01 -2.61397082e-02 -2.81877648e-02 -4.36457127e-01
6.52580440e-01 2.07503080e-01 -1.02827242e-02 -4.56788242e-01
3.06780040e-01 1.05384506e-01 -7.41878375e-02 -4.56603348e-01
4.65661079e-01 1.98207665e-02 -5.43004334e-01 6.94544375e-01
-3.00069362e-01 -3.29735637e-01 1.85570806e-01 3.26365769e-01
1.43376064e+00 -1.44042581e-01 1.70837045e-01 -1.73173510e-02
7.41800070e-01 -1.44038409e-01 5.22748649e-01 9.17217076e-01
-5.50528169e-01 6.12201691e-01 6.40943348e-01 -4.90212470e-01
-9.61994767e-01 -1.17457438e+00 6.03415724e-03 9.48973894e-01
1.87157858e-02 -5.69505990e-01 -8.79914522e-01 -1.16896224e+00
-1.48198918e-01 9.21934187e-01 -6.13870144e-01 -5.83705485e-01
-5.10339856e-01 -9.44271982e-01 1.32140172e+00 3.24866235e-01
5.39455831e-01 -1.24791634e+00 -1.27118707e-01 -2.11489320e-01
-2.09559396e-01 -1.01271129e+00 -5.30819893e-01 -2.65231550e-01
-3.73218119e-01 -1.25582695e+00 -6.79780692e-02 -1.96904555e-01
6.06447697e-01 -1.21866286e-01 9.52611029e-01 1.75785363e-01
-4.72454354e-02 2.72954375e-01 -4.78022695e-01 -3.28119487e-01
-9.37146902e-01 -8.09300877e-03 3.41964900e-01 -2.79247999e-01
-1.23970032e-01 -5.50476789e-01 -5.08736670e-01 5.42393029e-01
-1.48226511e+00 -3.98099035e-01 3.27573419e-01 5.63383639e-01
3.44950736e-01 2.83716798e-01 6.67681754e-01 -9.09632146e-01
1.21304393e+00 -5.37777126e-01 -6.73963666e-01 4.33043152e-01
-5.19827008e-01 1.08010769e-01 1.04912221e+00 -5.39801657e-01
-5.74286222e-01 -1.74489856e-01 -4.23139066e-01 -5.91800988e-01
-2.92700976e-01 4.44294661e-01 -7.30595887e-01 -1.20402642e-01
1.04835165e+00 1.01358607e-01 -1.31172329e-01 -1.32138476e-01
6.68004572e-01 4.74486172e-01 5.64316750e-01 -9.84768212e-01
1.35143793e+00 1.22896947e-01 -3.12140971e-01 -2.27702543e-01
-4.92805004e-01 5.84487952e-02 -2.13396251e-01 -1.38199642e-01
1.73792273e-01 -5.10619402e-01 -7.81677723e-01 5.02626896e-01
-9.98233616e-01 -4.51680064e-01 -2.50526607e-01 6.78045079e-02
-5.90875030e-01 6.50677860e-01 -6.54036999e-01 -7.67579198e-01
-6.54283822e-01 -1.43170869e+00 7.51919627e-01 -1.59324974e-01
-3.59365731e-01 -7.15531468e-01 2.49250680e-01 2.86533892e-01
4.64072227e-01 5.87791443e-01 7.76794553e-01 -9.78315711e-01
-2.64303982e-01 -5.71987689e-01 8.69566500e-02 4.74656552e-01
-1.24935754e-01 4.48698997e-01 -1.11863828e+00 -3.38401943e-01
2.26572871e-01 -4.30136234e-01 4.91240174e-01 -1.66401312e-01
1.37368560e+00 -7.25444257e-01 1.27936050e-01 7.59553432e-01
1.27224445e+00 -1.34602934e-01 8.59606326e-01 5.77597558e-01
5.40642858e-01 6.25854611e-01 5.30739605e-01 4.08152491e-01
-8.21690187e-02 4.71821010e-01 5.94553947e-01 2.17915967e-01
2.27511749e-01 -4.21681732e-01 6.59092724e-01 4.49455559e-01
1.92482144e-01 -4.52854931e-01 -1.11191106e+00 1.24993429e-01
-1.56333327e+00 -1.05776548e+00 2.57164180e-01 2.19205785e+00
1.02817082e+00 3.86353523e-01 1.03318654e-01 4.75259990e-01
6.45387650e-01 1.86058775e-01 -5.29791892e-01 -6.04846478e-01
-1.70250162e-01 -9.85355973e-02 3.66476804e-01 5.92630386e-01
-1.15300059e+00 1.26167786e+00 6.71145439e+00 1.12443733e+00
-9.21557069e-01 -1.16262771e-01 5.85793793e-01 -1.05324514e-01
-6.38105214e-01 -4.46139127e-02 -3.54769409e-01 4.70649421e-01
1.04215980e+00 -4.82615322e-01 6.06538773e-01 6.97672546e-01
3.46639335e-01 5.41309357e-01 -1.03725910e+00 4.33669478e-01
-2.45768741e-01 -1.04812717e+00 3.46496344e-01 -1.67463005e-01
4.94722098e-01 -3.14303249e-01 3.55992645e-01 3.49794358e-01
6.89936697e-01 -1.10957623e+00 7.36118197e-01 3.86486709e-01
7.81460345e-01 -9.64029610e-01 6.48928463e-01 3.28861475e-01
-7.46328354e-01 -2.38379717e-01 -1.45373121e-01 1.11763164e-01
-1.03913993e-02 4.54938143e-01 -6.28240645e-01 8.26121569e-01
4.64269012e-01 1.65882602e-01 -7.42936969e-01 6.78183913e-01
-4.88834292e-01 6.45256460e-01 -1.22954518e-01 2.33192146e-01
7.28039667e-02 1.47718966e-01 7.99473941e-01 1.19541490e+00
-7.28845298e-02 1.41608357e-01 8.27597082e-02 9.65184748e-01
-2.45662078e-01 1.31718546e-01 -8.12162876e-01 -1.52949750e-01
7.51139700e-01 1.20116735e+00 -5.34339666e-01 -9.35081765e-02
1.30243689e-01 8.13828647e-01 2.41836339e-01 4.94123548e-01
-1.19071269e+00 -4.46232766e-01 7.89421678e-01 -1.81787118e-01
-3.26745957e-01 -1.06891289e-01 -3.61685067e-01 -1.20054257e+00
2.50773311e-01 -1.69288588e+00 5.90652823e-01 -4.10587758e-01
-1.53742373e+00 8.30690563e-01 9.61519629e-02 -1.33616734e+00
-2.87204027e-01 -3.31388324e-01 -8.99485826e-01 7.75455952e-01
-9.89577532e-01 -1.22403741e+00 3.79836820e-02 8.21846843e-01
2.86101878e-01 -3.01032752e-01 9.07420099e-01 1.01395369e-01
-1.08219492e+00 1.30618358e+00 -1.57866925e-01 2.96744466e-01
7.18688726e-01 -8.94787312e-01 8.10453653e-01 1.43577981e+00
-2.80832887e-01 9.85636532e-01 9.01190400e-01 -6.28834546e-01
-1.58586407e+00 -1.23078072e+00 2.96599180e-01 -8.73846650e-01
1.07480240e+00 -4.43426102e-01 -8.94723952e-01 4.68972683e-01
-1.82765901e-01 -6.16296902e-02 7.08446205e-01 -1.05031766e-01
-8.85908723e-01 6.62548542e-02 -1.53296578e+00 1.16275311e+00
9.64172244e-01 -6.67538047e-01 -3.92886370e-01 3.40039611e-01
1.23463726e+00 -3.82317632e-01 -8.94059420e-01 7.86370397e-01
4.13407147e-01 -7.83604085e-01 1.04132986e+00 -9.32666123e-01
4.34344530e-01 -5.27527869e-01 -3.65535378e-01 -1.11927462e+00
-1.18993096e-01 -1.01666307e+00 -7.74334371e-02 1.42223108e+00
6.40667677e-01 -8.49246919e-01 4.03806359e-01 9.46443141e-01
-5.98518550e-02 -6.41421318e-01 -7.12111652e-01 -8.63941014e-01
3.77069801e-01 -9.45333719e-01 1.11133015e+00 1.12909496e+00
5.91957606e-02 -1.67645261e-01 -3.29781920e-01 5.25157392e-01
3.81685793e-01 -1.48773119e-01 9.84267712e-01 -3.66148502e-01
-2.71040052e-01 -6.27041519e-01 -2.76045382e-01 -1.62095144e-01
3.42976451e-01 -7.46425807e-01 -6.24421276e-02 -8.47058654e-01
9.57634449e-02 -3.82175356e-01 -4.14967835e-01 5.92794240e-01
-4.94288802e-01 1.62409902e-01 2.64721841e-01 2.43875012e-01
-6.79066122e-01 4.55076993e-01 1.03211784e+00 -2.63409108e-01
-1.48236096e-01 1.20990649e-01 -1.00590372e+00 6.31175697e-01
1.21699286e+00 -4.89487231e-01 -4.36545551e-01 -3.14218014e-01
2.51040131e-01 -1.70002833e-01 4.00052518e-01 -9.13097084e-01
1.34735197e-01 -5.09761572e-01 4.04908434e-02 -4.99103367e-02
-1.31103337e-01 -5.60206473e-01 2.27733761e-01 5.04501939e-01
-6.28502548e-01 2.35147998e-01 3.92779380e-01 3.04382384e-01
-7.01531917e-02 -1.69640914e-01 7.19366431e-01 5.38251773e-02
-3.25293332e-01 5.06288290e-01 -8.47021863e-02 2.25429818e-01
1.00601459e+00 2.13135898e-01 -6.16096675e-01 -1.85995072e-01
-4.34949845e-01 3.69065315e-01 8.47108245e-01 4.81470585e-01
7.61361897e-01 -1.24967790e+00 -9.54187036e-01 2.62770712e-01
3.35790366e-01 -3.14606309e-01 4.03370410e-01 3.50258291e-01
-4.73087072e-01 -3.44202071e-02 -1.17357135e-01 8.51975530e-02
-9.97912109e-01 1.23233187e+00 4.15562779e-01 -4.82580543e-01
-1.98363677e-01 5.33368528e-01 -9.10109375e-03 -8.36415112e-01
2.85811335e-01 5.14074229e-02 1.00975502e-02 -4.88060802e-01
6.83365047e-01 4.86060172e-01 1.16545089e-01 -3.62788737e-01
-4.25994813e-01 1.67713165e-01 -2.19741818e-02 -3.00066382e-01
7.89443314e-01 1.44765168e-01 -1.06573820e-01 -1.55518297e-02
9.73487675e-01 2.73967505e-01 -7.69215107e-01 1.32841757e-02
-9.87722576e-02 -5.00053883e-01 -3.53085160e-01 -1.00176501e+00
-9.29292202e-01 5.76699317e-01 2.83977598e-01 3.09847385e-01
1.26406300e+00 -3.87513161e-01 5.95169425e-01 4.00508463e-01
2.15422064e-01 -8.76070738e-01 -1.01229928e-01 4.79633629e-01
1.24143279e+00 -1.03585839e+00 -1.31984642e-02 -2.88282067e-01
-7.99318850e-01 7.92507172e-01 6.40570819e-01 5.57950735e-02
4.11569595e-01 4.28608656e-01 2.16732472e-01 6.81464672e-02
-7.60188162e-01 1.78756729e-01 1.39710709e-01 5.66625834e-01
1.06583066e-01 1.33055151e-01 -2.41214052e-01 7.90808797e-01
-5.24196804e-01 -2.80450761e-01 4.74478692e-01 8.45909238e-01
7.24185407e-02 -1.41848540e+00 -6.21995687e-01 -1.48809887e-02
-6.73779666e-01 -2.01413408e-01 -6.30297720e-01 6.53071880e-01
-3.02697182e-01 1.28753328e+00 -7.69023001e-01 -9.56708550e-01
6.81118071e-01 -2.14914046e-02 1.13102987e-01 -1.95802659e-01
-9.79520380e-01 -3.83525252e-01 1.98448062e-01 -8.69484484e-01
1.64186239e-01 -3.48817587e-01 -9.61412787e-01 -7.74969041e-01
-6.58942610e-02 2.15030596e-01 5.49699843e-01 6.89037263e-01
3.20930302e-01 3.66821766e-01 1.08915055e+00 -4.71081465e-01
-1.10829937e+00 -8.05224955e-01 -1.52044058e-01 5.47024488e-01
1.41027167e-01 -2.06497848e-01 -7.16456652e-01 -2.27237150e-01] | [5.9862060546875, 8.04868221282959] |
4568ae36-f749-4be4-8f16-e48cf0ede626 | risk-assessment-of-lymph-node-metastases-in | 2305.10041 | null | https://arxiv.org/abs/2305.10041v1 | https://arxiv.org/pdf/2305.10041v1.pdf | Risk Assessment of Lymph Node Metastases in Endometrial Cancer Patients: A Causal Approach | Assessing the pre-operative risk of lymph node metastases in endometrial cancer patients is a complex and challenging task. In principle, machine learning and deep learning models are flexible and expressive enough to capture the dynamics of clinical risk assessment. However, in this setting we are limited to observational data with quality issues, missing values, small sample size and high dimensionality: we cannot reliably learn such models from limited observational data with these sources of bias. Instead, we choose to learn a causal Bayesian network to mitigate the issues above and to leverage the prior knowledge on endometrial cancer available from clinicians and physicians. We introduce a causal discovery algorithm for causal Bayesian networks based on bootstrap resampling, as opposed to the single imputation used in related works. Moreover, we include a context variable to evaluate whether selection bias results in learning spurious associations. Finally, we discuss the strengths and limitations of our findings in light of the presence of missing data that may be missing-not-at-random, which is common in real-world clinical settings. | ['Fabio Stella', 'Marco Scutari', 'Casper Reijnen', 'Hanny Pijnenborg', 'Peter J. F. Lucas', 'Alice Bernasconi', 'Alessio Zanga'] | 2023-05-17 | null | null | null | null | ['imputation', 'causal-discovery', 'imputation', 'selection-bias', 'imputation'] | ['computer-vision', 'knowledge-base', 'miscellaneous', 'natural-language-processing', 'time-series'] | [ 3.71818304e-01 4.24628109e-01 -8.79194081e-01 -3.86264235e-01
-6.33186400e-01 -1.43055946e-01 3.95241529e-01 4.43908244e-01
-5.89964747e-01 1.10828555e+00 8.04833293e-01 -7.08737969e-01
-7.55955398e-01 -9.70433295e-01 -5.73609710e-01 -6.10191703e-01
-4.04012620e-01 4.53179359e-01 -3.74364913e-01 1.68360397e-01
-1.24349147e-02 1.44231379e-01 -7.21954823e-01 1.35610178e-01
8.98970068e-01 2.46831805e-01 -9.07808319e-02 4.29156244e-01
2.26056337e-01 8.30219984e-01 -1.13690913e-01 -2.50683725e-01
-1.25864252e-01 -4.02450860e-01 -4.58480835e-01 -4.21498954e-01
-4.99973409e-02 -5.40065646e-01 -3.41972530e-01 5.25260627e-01
5.52819967e-01 -3.68124872e-01 9.71077502e-01 -1.07405186e+00
-2.81826973e-01 7.34130204e-01 -5.24785697e-01 1.15067966e-01
6.84281671e-03 2.00832173e-01 6.75410628e-01 -4.32216913e-01
7.73951769e-01 9.08060908e-01 9.94713008e-01 4.11822319e-01
-1.53360510e+00 -8.65725696e-01 -4.23737243e-02 -1.20620504e-01
-9.85450685e-01 -5.45857131e-01 3.14554483e-01 -6.93896055e-01
2.45836377e-01 1.65696457e-01 7.11305082e-01 1.67868054e+00
6.46037936e-01 1.44721478e-01 1.07246447e+00 -3.20410788e-01
3.15872312e-01 -2.53071934e-01 -5.88132143e-02 6.41890585e-01
5.63405693e-01 7.13277698e-01 -5.41340709e-01 -8.11780930e-01
8.86092544e-01 4.72437561e-01 -8.06402415e-02 -4.54726279e-01
-1.34360743e+00 1.09245169e+00 4.08629268e-01 9.44139622e-03
-5.69485188e-01 2.42111549e-01 2.91845053e-01 4.90986463e-03
1.11255229e-01 2.24139363e-01 -4.95798320e-01 -9.53885838e-02
-1.02865160e+00 -1.18910773e-02 7.93941915e-01 3.89839023e-01
2.99628428e-03 -3.56374830e-01 -2.35122383e-01 4.63701755e-01
3.22542876e-01 4.00790513e-01 2.98525065e-01 -1.03480625e+00
1.11287393e-01 3.01729292e-01 2.67200798e-01 -8.02371621e-01
-8.51754546e-01 -4.77477223e-01 -1.37505436e+00 3.49807777e-02
6.19151413e-01 -4.88304853e-01 -8.19396496e-01 2.00023818e+00
2.13759035e-01 2.46222377e-01 -1.97094399e-02 6.34092152e-01
6.62697077e-01 -2.32178882e-01 5.17488599e-01 -5.16016543e-01
1.36876321e+00 -9.37376693e-02 -8.29764545e-01 -2.01921865e-01
7.34454989e-01 -3.51565599e-01 7.00579345e-01 2.91798294e-01
-9.84901607e-01 1.91324964e-01 -6.74276590e-01 1.55044422e-01
-4.17045541e-02 9.21895728e-02 1.04443848e+00 6.74174547e-01
-5.14553368e-01 5.34722686e-01 -1.24963248e+00 -5.70147455e-01
6.93575263e-01 3.38451356e-01 -3.55828077e-01 -1.03918701e-01
-1.31729519e+00 8.34525526e-01 1.81955397e-01 1.76417515e-01
-9.04938877e-01 -1.32742190e+00 -6.44428849e-01 1.38267994e-01
4.82697487e-01 -1.31225944e+00 9.86489415e-01 -6.41375422e-01
-9.36761975e-01 5.23897469e-01 -4.18113440e-01 -3.23838323e-01
7.27531374e-01 2.55148172e-01 -1.05199024e-01 -3.33330393e-01
5.02007939e-02 2.23530099e-01 1.01646647e-01 -8.54737818e-01
-4.74887162e-01 -7.29658067e-01 -4.84775126e-01 3.75330858e-02
-2.22708195e-01 -1.32760152e-01 1.00225486e-01 -6.92390501e-01
1.32437587e-01 -8.87322307e-01 -8.09444547e-01 8.01315065e-03
-1.94660947e-01 2.41053417e-01 7.07221255e-02 -6.26919627e-01
1.15027833e+00 -1.99020445e+00 -8.26848298e-02 1.61351919e-01
3.09056908e-01 -3.72319281e-01 5.77005930e-02 4.72123802e-01
-2.11638689e-01 4.07119095e-01 -2.05828279e-01 1.70002475e-01
-4.49710667e-01 1.91147640e-01 -5.03810979e-02 5.52396417e-01
9.82393101e-02 9.73664701e-01 -9.97198284e-01 -6.16667867e-01
-2.40687616e-02 3.57132494e-01 -5.83885789e-01 -1.79590642e-01
5.01732416e-02 5.11739433e-01 -4.94007021e-01 4.57715094e-01
4.95337963e-01 -4.28099543e-01 8.57015193e-01 1.16115138e-01
9.33105722e-02 4.01363194e-01 -1.03389692e+00 1.59260786e+00
-5.56790173e-01 3.21565419e-01 1.47671476e-01 -8.44635963e-01
4.19574410e-01 4.82700258e-01 6.25611424e-01 -3.08172673e-01
-6.24044612e-02 1.30995899e-01 3.80608022e-01 -5.81818342e-01
-7.08002076e-02 -6.91336930e-01 1.01359680e-01 4.47187543e-01
-3.32957804e-01 2.69861668e-01 -3.22956115e-01 6.68083057e-02
1.56365669e+00 -2.24558994e-01 5.25424778e-01 -2.81062454e-01
-3.02196503e-01 1.21092148e-01 1.17260087e+00 1.11975503e+00
2.70965602e-02 5.78167081e-01 1.12228131e+00 -4.02623951e-01
-7.49980807e-01 -1.25997138e+00 -6.81207418e-01 4.94955927e-01
-4.66804355e-01 -1.14233680e-02 6.91071823e-02 -6.04737282e-01
1.58312127e-01 6.13492787e-01 -9.60482001e-01 -2.36250415e-01
-2.21828729e-01 -1.76333725e+00 5.27337313e-01 6.10363781e-01
-3.10671836e-01 -7.58281767e-01 -4.96104658e-01 4.81365681e-01
-3.66661072e-01 -5.50269842e-01 8.36777836e-02 4.24350500e-01
-1.16966116e+00 -1.44366157e+00 -5.85842609e-01 -2.35190660e-01
6.18162692e-01 -3.26784253e-01 1.20215356e+00 -8.65879003e-03
-4.19381708e-01 -6.14112914e-02 6.89895302e-02 -7.15841472e-01
-5.81684887e-01 -5.87537838e-03 -1.49027869e-01 -3.99892688e-01
3.42044830e-01 -5.57330728e-01 -1.01006186e+00 7.48878345e-02
-7.56015897e-01 -5.39120520e-03 1.06871307e+00 1.18371320e+00
3.14012527e-01 5.84868789e-02 7.92791545e-01 -1.23396587e+00
3.97041619e-01 -9.34133828e-01 -4.45380718e-01 1.50724165e-02
-8.38880181e-01 -4.87960689e-02 1.65011343e-02 -4.59382921e-01
-1.15428615e+00 -1.76138535e-01 2.61664093e-01 2.13290974e-01
-2.58493096e-01 9.46670830e-01 2.26406097e-01 3.06146175e-01
7.40494072e-01 -2.75437236e-01 5.42837977e-01 -1.91402137e-01
-1.93686575e-01 5.64987838e-01 -1.33184157e-02 -3.07428420e-01
2.14921340e-01 6.70247316e-01 5.79764724e-01 -3.65757108e-01
-7.15977907e-01 -7.64595941e-02 -3.14888954e-01 2.85425663e-01
6.26262248e-01 -1.20487106e+00 -1.11686552e+00 1.35168536e-02
-7.89854169e-01 -6.89975023e-01 -1.35968894e-01 9.73829210e-01
-5.15145659e-01 -1.90908164e-02 -7.01624036e-01 -7.25899279e-01
-2.27443546e-01 -1.05931890e+00 6.98894262e-01 -8.93523321e-02
-5.39943337e-01 -1.11448550e+00 2.18550384e-01 1.71186522e-01
4.88706499e-01 6.98948503e-01 1.31418335e+00 -3.98476928e-01
-2.89764255e-01 -2.07507387e-01 -2.30309397e-01 -5.53187847e-01
4.64330345e-01 1.29237115e-01 -4.51214582e-01 -2.30742499e-01
-1.98923454e-01 -2.53235966e-01 8.32857490e-01 1.09492719e+00
1.23524773e+00 -2.80731052e-01 -8.41334581e-01 3.74939084e-01
1.26660275e+00 -8.78432170e-02 4.11863565e-01 6.96555898e-02
2.73285002e-01 8.50865901e-01 3.15378785e-01 5.93327999e-01
7.01888740e-01 6.82771444e-01 3.99324775e-01 -3.02644253e-01
1.15417786e-01 -3.63501132e-01 -2.79553622e-01 -1.92307420e-02
-1.71380267e-01 2.77472241e-03 -1.14848137e+00 5.98696470e-01
-1.91310406e+00 -8.57334256e-01 -4.96778131e-01 2.53789067e+00
1.13458967e+00 1.98060587e-01 1.43618077e-01 -2.94574201e-01
4.76586223e-01 -3.94720018e-01 -3.68534833e-01 -5.43482080e-02
-3.39276567e-02 -1.33336231e-01 8.03930283e-01 3.09508175e-01
-7.01995373e-01 2.11071879e-01 7.24989414e+00 2.93984205e-01
-8.76453519e-01 1.90848812e-01 8.89363945e-01 -3.16367149e-01
-3.37951839e-01 3.58789355e-01 -2.97545880e-01 3.98374438e-01
1.12265003e+00 1.04888596e-01 -3.67215760e-02 2.88004931e-02
8.41956437e-01 -3.32827151e-01 -1.16525614e+00 3.04981381e-01
-4.87161577e-01 -1.35563648e+00 -6.17027700e-01 5.82480252e-01
5.87146699e-01 1.47461981e-01 -5.37413657e-02 -2.09573433e-02
1.01634610e+00 -1.40709507e+00 1.08007424e-01 8.35463881e-01
8.67698431e-01 -3.85908514e-01 1.07884669e+00 3.61330390e-01
-1.41425043e-01 -2.11593926e-01 -2.16528296e-01 -4.12986726e-02
9.71776545e-02 1.19661474e+00 -1.08574021e+00 4.50792402e-01
6.75323963e-01 5.88098586e-01 -2.85351694e-01 9.35696125e-01
-2.79946532e-02 9.51741159e-01 -3.56229722e-01 2.09424496e-01
-1.33267149e-01 9.74780247e-02 5.61399981e-02 7.42785692e-01
5.03798127e-01 1.41870871e-01 -3.14449698e-01 8.05862904e-01
7.27256611e-02 -1.37930036e-01 -6.02867723e-01 8.67789611e-02
4.30807322e-01 1.11410618e+00 -4.07787353e-01 2.22288400e-01
-5.58972597e-01 1.22990295e-01 1.10135570e-01 2.87407011e-01
-3.17398459e-01 3.10192674e-01 2.97370106e-01 5.07391512e-01
-1.74698159e-01 6.31871969e-02 -6.46924794e-01 -9.46961641e-01
-2.55936414e-01 -8.32502842e-01 9.25610542e-01 -3.72519165e-01
-1.35063004e+00 -2.10951656e-01 6.20032176e-02 -7.68953085e-01
-3.73643845e-01 -2.68369853e-01 -4.84892756e-01 9.10472989e-01
-1.31029367e+00 -1.00752831e+00 -1.81678578e-01 2.35586450e-01
-3.98782678e-02 9.81909856e-02 9.62697089e-01 -1.50623456e-01
-5.51897109e-01 4.13524061e-01 3.88064921e-01 1.03849597e-01
1.10053408e+00 -1.17254877e+00 -2.36345708e-01 2.26550907e-01
-4.66674000e-01 7.43582845e-01 6.16797507e-01 -1.12878478e+00
-1.17490768e+00 -8.04002345e-01 1.04161692e+00 -5.58059514e-01
8.13748598e-01 -2.47663260e-01 -7.65147269e-01 7.13945270e-01
-2.17595726e-01 1.19208612e-01 1.15919161e+00 9.39935505e-01
1.49015151e-02 -2.86609493e-02 -1.15478516e+00 7.79491305e-01
8.77386391e-01 9.72203389e-02 -2.64134467e-01 2.43252367e-01
2.45524570e-01 -2.05377415e-01 -1.28265667e+00 7.64860570e-01
8.66192222e-01 -8.83911312e-01 9.85873699e-01 -8.34532619e-01
7.67840981e-01 2.00207666e-01 2.31528252e-01 -1.19139612e+00
-3.82490695e-01 -2.50199199e-01 2.95892507e-01 9.14758444e-01
6.15550101e-01 -4.94971931e-01 1.15367079e+00 9.27935600e-01
3.50983053e-01 -6.38664186e-01 -1.06404567e+00 -2.67648637e-01
5.36475599e-01 -1.75171122e-01 5.45541465e-01 1.04872358e+00
2.56983280e-01 -1.70775354e-01 -2.74461925e-01 1.77914456e-01
8.63901019e-01 1.36239782e-01 3.74525756e-01 -1.45584857e+00
-3.85342062e-01 -2.47678775e-02 4.18364555e-02 5.20971455e-02
1.33553356e-01 -6.54894173e-01 -2.91959524e-01 -1.52015638e+00
8.96602273e-01 -9.38807726e-01 -3.94391358e-01 5.47039807e-01
-3.46449256e-01 -2.21893013e-01 -2.78128594e-01 1.55064449e-01
5.01172282e-02 4.08250213e-01 1.18402183e+00 -3.36030461e-02
-2.43840277e-01 3.39344889e-01 -8.64341199e-01 7.86179245e-01
6.95426106e-01 -9.81214345e-01 -2.88651735e-01 -1.64408624e-01
5.08721530e-01 9.27234888e-01 8.04671049e-01 -2.51994401e-01
1.89244330e-01 -6.55584812e-01 6.69414580e-01 -2.54570335e-01
-8.01081955e-02 -8.72529864e-01 5.34142673e-01 9.98085678e-01
-5.56067109e-01 -2.30913639e-01 8.31954405e-02 8.59621584e-01
3.19695882e-02 -3.39562327e-01 2.25353435e-01 -2.05496758e-01
2.90364176e-02 1.33878380e-01 -6.32852912e-01 -4.09792140e-02
7.05811143e-01 1.94911823e-01 -3.27891141e-01 -3.92075956e-01
-8.03032279e-01 3.08200747e-01 2.84127444e-01 1.45852327e-01
3.09198409e-01 -1.17940736e+00 -1.08655906e+00 -8.69319811e-02
1.53009817e-01 6.69225678e-02 4.84704494e-01 1.23329210e+00
-1.45834610e-01 3.84427667e-01 -4.13535275e-02 -2.18616858e-01
-9.37721848e-01 5.30262470e-01 3.32717210e-01 -5.70157170e-01
-4.70197231e-01 -1.71643123e-02 1.39862999e-01 -5.24281859e-01
8.96185860e-02 6.51053339e-02 -4.05354314e-02 7.23474696e-02
9.45570692e-02 3.36649448e-01 8.98206383e-02 1.35089606e-01
-4.27176207e-01 -1.42979532e-01 4.65371013e-02 -1.74708799e-01
1.44624722e+00 -7.95943961e-02 -1.58977807e-01 4.08988148e-01
6.54741943e-01 -5.03850393e-02 -1.05849648e+00 -1.46656767e-01
-1.98983192e-03 -9.19142365e-02 2.31485412e-01 -9.89544630e-01
-6.15830958e-01 5.92580378e-01 5.93209684e-01 -3.02880079e-01
9.46665466e-01 -1.32849857e-01 6.58805370e-02 1.23544326e-02
1.20119154e-01 -7.08121717e-01 -3.88727665e-01 -1.21635050e-01
7.03676760e-01 -1.56766462e+00 2.91514814e-01 -1.94021210e-01
-2.60578275e-01 8.08412910e-01 2.47149780e-01 2.43176490e-01
8.47385049e-01 5.09798467e-01 -1.37776639e-02 -2.94850022e-02
-1.22224295e+00 2.52526134e-01 -6.67180028e-03 7.20189989e-01
7.82635391e-01 2.65885383e-01 -7.20560968e-01 8.26151431e-01
-1.12323336e-01 6.60541177e-01 9.46210682e-01 7.84621060e-01
2.58248031e-01 -1.18753111e+00 -4.96684164e-01 1.05932546e+00
-9.03471828e-01 -2.08509997e-01 -6.86489940e-02 8.31448197e-01
-1.15319356e-01 9.68552828e-01 1.97378144e-01 2.63258338e-01
1.88551575e-01 -1.64104834e-01 2.25148812e-01 -4.33860868e-01
-2.72799790e-01 1.40422076e-01 1.45636648e-01 -3.01978201e-01
-4.65653300e-01 -1.01687801e+00 -9.02954638e-01 -2.64663815e-01
-1.04549520e-01 -5.84721789e-02 5.66711724e-01 8.73455644e-01
3.87933344e-01 7.05467165e-01 3.51131558e-01 -9.86074805e-02
-6.83489382e-01 -8.35781157e-01 -5.16813278e-01 9.69289988e-02
4.33339268e-01 -5.42246401e-01 -4.52999264e-01 -1.49596170e-01] | [7.987527370452881, 5.509975433349609] |
283cf69f-d3d7-44c9-8e61-6c5f931b267e | automatic-liver-segmentation-using-an | 1707.08037 | null | http://arxiv.org/abs/1707.08037v1 | http://arxiv.org/pdf/1707.08037v1.pdf | Automatic Liver Segmentation Using an Adversarial Image-to-Image Network | Automatic liver segmentation in 3D medical images is essential in many
clinical applications, such as pathological diagnosis of hepatic diseases,
surgical planning, and postoperative assessment. However, it is still a very
challenging task due to the complex background, fuzzy boundary, and various
appearance of liver. In this paper, we propose an automatic and efficient
algorithm to segment liver from 3D CT volumes. A deep image-to-image network
(DI2IN) is first deployed to generate the liver segmentation, employing a
convolutional encoder-decoder architecture combined with multi-level feature
concatenation and deep supervision. Then an adversarial network is utilized
during training process to discriminate the output of DI2IN from ground truth,
which further boosts the performance of DI2IN. The proposed method is trained
on an annotated dataset of 1000 CT volumes with various different scanning
protocols (e.g., contrast and non-contrast, various resolution and position)
and large variations in populations (e.g., ages and pathology). Our approach
outperforms the state-of-the-art solutions in terms of segmentation accuracy
and computing efficiency. | ['S. Kevin Zhou', 'Sasa Grbic', 'Dong Yang', 'Mingqing Chen', 'Dimitris Metaxas', 'Daguang Xu', 'Dorin Comaniciu', 'Bogdan Georgescu'] | 2017-07-25 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [-1.23720609e-01 -9.71135944e-02 2.21895903e-01 -2.54874468e-01
-5.35735607e-01 -6.63135946e-01 3.26352358e-01 1.08830839e-01
-3.57033461e-01 3.99048537e-01 1.98385090e-01 -5.04151762e-01
3.09977740e-01 -7.36641526e-01 -4.79669183e-01 -8.83632720e-01
-3.05868357e-01 5.06811023e-01 3.40412617e-01 1.42061979e-01
-1.12741888e-01 6.95116818e-01 -4.27428633e-01 -1.87683534e-02
9.86973166e-01 1.00061750e+00 9.86882895e-02 6.53820336e-01
-3.96150053e-02 4.33348209e-01 -4.08626795e-01 -3.09699088e-01
6.46858931e-01 -7.71189094e-01 -5.54707408e-01 3.03055853e-01
-7.46826082e-02 -4.79729354e-01 -5.17130554e-01 1.21283984e+00
8.33745122e-01 -3.52149725e-01 6.95013762e-01 -6.08728468e-01
-4.69312966e-01 5.52988410e-01 -5.03504455e-01 4.33492631e-01
-3.54483649e-02 5.40928304e-01 1.24933813e-02 -5.94841123e-01
3.26857865e-01 6.93791449e-01 6.25504017e-01 4.55383301e-01
-1.06863344e+00 -5.44422150e-01 -3.77740562e-01 -2.36501053e-01
-1.28213406e+00 3.31264595e-03 6.79900229e-01 -6.15161717e-01
9.60084572e-02 7.53459856e-02 9.04890299e-01 5.63858032e-01
5.06921887e-01 6.18968546e-01 1.31690609e+00 -1.38848811e-01
-7.59142563e-02 4.68408056e-02 -3.07771683e-01 1.09666455e+00
3.32691878e-01 2.28542939e-01 3.79182577e-01 4.39039283e-02
1.22334099e+00 9.70926210e-02 -6.00120306e-01 -4.58427727e-01
-1.50746095e+00 7.08771348e-01 7.81144381e-01 3.05038035e-01
-5.72091997e-01 -1.43616796e-01 6.07301712e-01 7.40336180e-02
-1.79533120e-02 3.32197100e-01 -2.71974891e-01 5.54520369e-01
-8.98734450e-01 -2.77554572e-01 8.38058114e-01 7.86477089e-01
1.49459913e-01 2.51619935e-01 -2.97418565e-01 3.45245361e-01
3.81419599e-01 2.00388998e-01 1.01149094e+00 -3.10468823e-01
-2.31015459e-02 5.87081492e-01 -1.88522980e-01 -5.52348375e-01
-6.29661202e-01 -6.58962071e-01 -1.49769557e+00 1.66365623e-01
3.85102481e-01 -3.26770723e-01 -1.28223848e+00 1.22351038e+00
4.34025764e-01 2.72634625e-01 8.75868276e-02 1.33128977e+00
1.25203562e+00 2.73793489e-01 3.41238201e-01 -1.70933023e-01
1.62407649e+00 -9.60855186e-01 -3.53524119e-01 5.16887344e-02
3.55312318e-01 -7.58905053e-01 6.94820762e-01 8.53281468e-02
-1.02622211e+00 -2.58129418e-01 -9.29114521e-01 3.36479694e-01
3.27630043e-02 2.98601985e-01 5.40346682e-01 7.02728391e-01
-8.51018429e-01 2.26084486e-01 -1.12211680e+00 -1.17309585e-01
6.63968861e-01 5.74988544e-01 -3.89834553e-01 1.78411882e-02
-1.01691902e+00 8.36607099e-01 5.57892025e-01 1.76395774e-01
-1.21926427e+00 -6.64883316e-01 -9.50045228e-01 -1.45184293e-01
1.30400211e-01 -9.27767158e-01 9.91945565e-01 -9.49571550e-01
-1.55295515e+00 1.06708705e+00 4.54886317e-01 -5.55334747e-01
9.04709458e-01 3.51868927e-01 9.41920839e-03 2.87376553e-01
-1.86258346e-01 3.97999078e-01 6.24087930e-01 -9.42144036e-01
7.87699129e-03 -3.27692926e-01 -3.20861548e-01 3.92665803e-01
4.58548158e-01 8.05301890e-02 -4.10978764e-01 -7.81338096e-01
2.24588081e-01 -9.85151350e-01 -6.15539551e-01 1.32396802e-01
-5.23066759e-01 2.64401644e-01 4.18157160e-01 -1.09572816e+00
6.52240872e-01 -2.10762596e+00 1.41708046e-01 2.33168036e-01
2.48169810e-01 3.25471014e-01 2.63001502e-01 -3.91195446e-01
8.34899470e-02 2.25247547e-01 -4.72943693e-01 1.23677097e-01
-2.53217876e-01 3.95326354e-02 4.83662814e-01 8.29482496e-01
2.21499186e-02 1.12369931e+00 -8.21175694e-01 -9.72014904e-01
4.53758657e-01 4.46266443e-01 -3.65743846e-01 5.47732472e-01
2.66598109e-02 1.20589507e+00 -4.78755832e-01 7.67278075e-01
7.16577530e-01 -3.79810601e-01 1.49239029e-03 -3.73756230e-01
-2.95049231e-02 -2.32241794e-01 -7.07248747e-01 1.89995730e+00
-4.10894901e-01 2.42522746e-01 2.94213235e-01 -9.69718695e-01
6.50291920e-01 7.17707872e-01 5.62134802e-01 -4.01946336e-01
6.04580283e-01 3.80765617e-01 4.31133658e-01 -6.81389332e-01
-2.43486434e-01 -3.51033300e-01 7.57089350e-03 2.85176575e-01
-4.93549556e-03 -4.76791233e-01 9.40295681e-03 -1.92568734e-01
7.98924387e-01 -1.22216105e-01 4.29755569e-01 -4.57446367e-01
8.21068466e-01 3.20573822e-02 4.99501348e-01 3.00031245e-01
-7.26652563e-01 7.34389544e-01 4.31043983e-01 -6.13128662e-01
-1.05359519e+00 -9.89785671e-01 -2.85751820e-01 2.48476818e-01
3.19227666e-01 4.97937381e-01 -8.35698307e-01 -1.01848257e+00
-1.68912172e-01 1.50328830e-01 -5.02382576e-01 1.07877016e-01
-7.79205382e-01 -1.06282616e+00 6.26288533e-01 1.85090959e-01
9.09808218e-01 -9.80424106e-01 -8.75050843e-01 2.47238204e-01
-5.74357575e-03 -1.04137242e+00 -6.63886368e-01 8.84088222e-03
-1.00791526e+00 -1.22266316e+00 -1.09873033e+00 -1.14012575e+00
1.04063880e+00 -7.85752237e-02 1.28446770e+00 2.71961302e-01
-7.12599754e-01 -1.27303898e-01 -7.97626302e-02 -2.20129922e-01
-6.67249739e-01 3.08222361e-02 -3.77955317e-01 -3.02783668e-01
-6.01544194e-02 -3.25262755e-01 -1.27581370e+00 2.60374963e-01
-8.97028625e-01 2.60489792e-01 1.10859263e+00 1.05507290e+00
6.23307884e-01 1.74300075e-02 3.00968528e-01 -8.79609764e-01
4.10471797e-01 -5.18072248e-01 -7.19106793e-01 2.82503486e-01
-3.47672731e-01 -2.26767827e-02 7.28047907e-01 -4.01115984e-01
-8.17576826e-01 3.53354335e-01 -2.41302967e-01 -4.28384900e-01
-3.23218256e-01 4.72144693e-01 5.53280972e-02 -4.96283680e-01
5.39906859e-01 5.68931103e-01 2.57993609e-01 -1.51018411e-01
1.00558680e-02 3.76777828e-01 5.82912207e-01 -1.16748795e-01
6.34695172e-01 1.18895620e-01 2.03939602e-01 -2.42036968e-01
-3.76406282e-01 -7.29028657e-02 -6.90266788e-01 -5.72397048e-03
9.46625888e-01 -9.14066017e-01 -2.44644642e-01 6.93708420e-01
-8.80867124e-01 -3.70119661e-01 -3.74528542e-02 8.03952992e-01
-2.49460518e-01 3.26331526e-01 -1.11678588e+00 -1.66709647e-01
-9.22629774e-01 -1.77612805e+00 5.80088913e-01 5.57775974e-01
4.53553915e-01 -1.08885932e+00 -2.83015847e-01 6.73642606e-02
6.87855780e-01 8.77590954e-01 1.00529277e+00 -8.03314626e-01
-7.62203217e-01 -1.96194157e-01 -4.40326184e-01 3.39817613e-01
2.29397476e-01 -3.71585011e-01 -3.37028325e-01 -3.21897745e-01
1.37704179e-01 -1.68792531e-01 4.45123166e-01 7.60508120e-01
1.19522691e+00 -3.39732856e-01 -9.48201120e-02 1.00790513e+00
1.49597204e+00 2.47728333e-01 3.99919719e-01 1.18257953e-02
5.36557317e-01 1.28129467e-01 1.29776776e-01 2.62524039e-01
2.23148644e-01 1.86879069e-01 5.86522460e-01 -6.00860178e-01
-3.22039515e-01 2.27766275e-01 -7.10411221e-02 8.08410287e-01
1.00940689e-01 1.50829121e-01 -1.08316374e+00 5.83177507e-01
-1.16418886e+00 -5.21406710e-01 2.75640953e-02 2.03161097e+00
1.05341637e+00 -1.03768311e-01 -1.17478281e-01 -2.53233790e-01
7.44725704e-01 -2.10458785e-01 -5.43417692e-01 1.41252372e-02
2.04185337e-01 2.56449282e-01 6.94058120e-01 4.06443089e-01
-1.27643144e+00 5.38030744e-01 5.49436378e+00 4.25382495e-01
-1.63714683e+00 2.87889034e-01 1.04891098e+00 3.74378473e-01
6.72018752e-02 -4.48785961e-01 -6.07983693e-02 7.00120389e-01
3.05402696e-01 -8.66170004e-02 3.66167098e-01 6.13190234e-01
1.13670276e-02 -8.61020535e-02 -8.80935490e-01 8.43625605e-01
2.59559661e-01 -1.05972767e+00 -1.75535932e-01 -1.53010577e-01
7.24030674e-01 1.59425139e-01 -8.86656065e-03 1.22283511e-01
1.96569368e-01 -1.22814012e+00 2.61210799e-01 2.24881336e-01
1.13998079e+00 -5.05555391e-01 1.23094797e+00 3.47153604e-01
-9.05908763e-01 4.08507705e-01 9.90158841e-02 6.56160116e-01
5.33578582e-02 5.32497764e-01 -1.14929104e+00 6.49060011e-01
2.77386487e-01 1.56374440e-01 -4.54774052e-01 1.38923419e+00
-2.84719378e-01 4.73132014e-01 -3.22053492e-01 2.62123495e-01
2.06446156e-01 -3.44128042e-01 3.65650505e-01 1.27036941e+00
4.18358088e-01 4.55984056e-01 4.45585877e-01 6.44725144e-01
-3.72835934e-01 2.47664839e-01 -3.18160564e-01 3.12422246e-01
1.73354730e-01 1.44320488e+00 -1.10516894e+00 -4.56451148e-01
-3.45618784e-01 1.01421249e+00 -2.78303385e-01 6.50336668e-02
-8.83630991e-01 -1.95602313e-01 -6.79270476e-02 1.83067843e-01
-1.16003044e-04 -1.33041471e-01 -2.70343840e-01 -1.32697892e+00
-4.22380745e-01 -8.51520658e-01 4.65178430e-01 -3.82112116e-01
-1.11775100e+00 9.21378493e-01 -3.04344833e-01 -1.19746828e+00
-4.57601547e-02 -4.71449494e-01 -8.35967898e-01 1.04383731e+00
-1.71676123e+00 -1.18570411e+00 -9.05563891e-01 6.46118879e-01
5.20934999e-01 -3.25433761e-02 6.58661425e-01 4.54197466e-01
-4.11749512e-01 4.61472601e-01 -2.75400043e-01 8.02384317e-01
5.35505414e-01 -1.29229486e+00 3.44668590e-02 8.90089035e-01
-3.93257499e-01 1.50175512e-01 3.42978507e-01 -5.19303560e-01
-1.35190189e+00 -1.26652431e+00 2.43086800e-01 1.04508959e-01
2.98512787e-01 1.18601896e-01 -5.78153193e-01 5.92781723e-01
4.21885282e-01 7.44578719e-01 7.49285638e-01 -9.29554701e-01
2.82504976e-01 1.02033146e-01 -1.72765613e+00 4.85691786e-01
4.25018042e-01 8.21876749e-02 -4.23578262e-01 3.27536404e-01
3.97445053e-01 -1.29912257e+00 -1.16846585e+00 5.90379536e-01
3.42023343e-01 -8.57245028e-01 1.02989531e+00 -2.14815885e-01
3.78890425e-01 -4.01367247e-01 3.07019621e-01 -1.33571076e+00
-2.44239599e-01 -4.89786983e-01 2.08324775e-01 4.90033269e-01
1.94688708e-01 -5.74641943e-01 8.07419956e-01 4.57292318e-01
-4.38460857e-01 -8.85185897e-01 -8.08835626e-01 -1.28952384e-01
2.25552738e-01 3.81916434e-01 6.39516652e-01 9.54802096e-01
-4.40780997e-01 -1.10345066e-01 2.04611905e-02 3.51155341e-01
7.91865468e-01 2.80387908e-01 4.42711055e-01 -9.37497377e-01
-8.40030834e-02 -5.04826963e-01 -6.20093286e-01 -8.30529034e-01
-1.90673530e-01 -1.03913486e+00 8.73099044e-02 -1.47679090e+00
2.83653021e-01 -7.79470444e-01 -3.79072070e-01 3.53876173e-01
-3.05521518e-01 4.69383329e-01 7.34574646e-02 2.78821260e-01
-2.38169238e-01 2.74130434e-01 1.84035242e+00 -2.52468467e-01
-1.23250455e-01 1.14870392e-01 -2.92124808e-01 6.52023792e-01
9.25401092e-01 -3.42866719e-01 -6.67875931e-02 -3.67729932e-01
-5.46397150e-01 5.68452120e-01 4.96237397e-01 -9.90073681e-01
3.07262391e-01 1.12471402e-01 9.18829799e-01 -3.43504488e-01
-3.10282052e-01 -9.36070800e-01 2.98108131e-01 1.18034732e+00
-4.59682085e-02 1.28739402e-01 9.86214951e-02 7.72043094e-02
-3.26982379e-01 -2.37673968e-01 1.10790324e+00 -7.79893756e-01
-3.41627300e-01 7.62224734e-01 -1.00051679e-01 2.12196961e-01
1.26140666e+00 6.97428584e-02 8.85842144e-02 1.03475954e-02
-8.30164492e-01 2.99498618e-01 4.34374630e-01 -2.50214309e-01
6.41602039e-01 -1.23247135e+00 -1.12795579e+00 5.03846884e-01
-3.06272835e-01 6.86626375e-01 1.71802148e-01 1.43754780e+00
-1.31911707e+00 1.02656268e-01 -3.12618345e-01 -8.11017096e-01
-1.02658403e+00 4.23190475e-01 8.65835309e-01 -6.27283692e-01
-7.44618416e-01 6.57736778e-01 2.11534321e-01 -3.57666641e-01
6.86867312e-02 -4.30704385e-01 5.93594499e-02 -4.52767968e-01
1.09016471e-01 -2.26004764e-01 1.09719045e-01 -4.89618629e-01
-2.46779591e-01 5.09647548e-01 3.89228240e-02 2.43420660e-01
8.94780695e-01 1.26219049e-01 -1.46685243e-01 -2.75194079e-01
8.27753305e-01 -8.77597034e-02 -1.23315811e+00 -1.92499399e-01
-3.26806813e-01 -5.02095401e-01 1.48843825e-01 -1.05822814e+00
-1.59658301e+00 8.58585298e-01 8.47255111e-01 -1.78839583e-02
1.29410160e+00 -2.18102068e-01 8.65836740e-01 -3.47879618e-01
3.76045167e-01 -1.98428124e-01 -1.23485990e-01 1.50896087e-01
7.25408614e-01 -1.59293830e+00 7.69943818e-02 -3.12570214e-01
-6.92867100e-01 1.10839939e+00 5.68507254e-01 -2.41918400e-01
7.66224027e-01 3.83603126e-01 4.16002959e-01 5.11234999e-02
-9.77558643e-02 6.83746934e-02 1.23568945e-01 3.10339063e-01
4.91875917e-01 2.10807323e-01 -4.34331775e-01 5.43792546e-01
1.08541980e-01 9.56249684e-02 4.46495861e-01 6.77729547e-01
-1.14371821e-01 -7.37275302e-01 -3.30821276e-01 4.68903273e-01
-9.56456244e-01 -1.87080666e-01 2.84694374e-01 8.65155637e-01
1.81394711e-01 3.30209583e-01 -1.83843076e-01 1.72632307e-01
4.26354185e-02 -1.72378451e-01 5.58157504e-01 -4.23961997e-01
-1.03860474e+00 1.69963285e-01 -4.48635876e-01 -1.32908925e-01
-2.18433022e-01 -3.61072809e-01 -1.32388163e+00 -1.27420828e-01
-2.26814955e-01 2.11235687e-01 8.18697691e-01 5.41761220e-01
1.08036861e-01 6.63496792e-01 8.54212165e-01 -7.84634054e-01
-7.27987051e-01 -9.87554789e-01 -2.21119970e-01 6.25459135e-01
3.73685807e-01 -3.26356888e-01 -2.10124269e-01 2.62341291e-01] | [14.512892723083496, -2.6564908027648926] |
f9e57379-8905-4de7-9559-ed0f137dd15b | leveraging-experience-in-lifelong-multi-agent | 2202.04382 | null | https://arxiv.org/abs/2202.04382v3 | https://arxiv.org/pdf/2202.04382v3.pdf | Leveraging Experience in Lifelong Multi-Agent Pathfinding | In Lifelong Multi-Agent Path Finding (L-MAPF) a team of agents performs a stream of tasks consisting of multiple locations to be visited by the agents on a shared graph while avoiding collisions with one another. L-MAPF is typically tackled by partitioning it into multiple consecutive, and hence similar, "one-shot" MAPF queries, as in the Rolling-Horizon Collision Resolution (RHCR) algorithm. Therefore, a solution to one query informs the next query, which leads to similarity with respect to the agents' start and goal positions, and how collisions need to be resolved from one query to the next. Thus, experience from solving one MAPF query can potentially be used to speedup solving the next one. Despite this intuition, current L-MAPF planners solve consecutive MAPF queries from scratch. In this paper, we introduce a new RHCR-inspired approach called exRHCR, which exploits experience in its constituent MAPF queries. In particular, exRHCR employs an extension of Priority-Based Search (PBS), a state-of-the-art MAPF solver. The extension, which we call exPBS, allows to warm-start the search with the priorities between agents used by PBS in the previous MAPF instances. We demonstrate empirically that exRHCR solves L-MAPF instances up to 39% faster than RHCR, and has the potential to increase system throughput for given task streams by increasing the number of agents a planner can cope with for a given time budget. | ['Oren Salzman', 'Kiril Solovey', 'Nitzan Madar'] | 2022-02-09 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [ 6.17342219e-02 4.19565350e-01 -1.74137220e-01 1.41564801e-01
-8.26196790e-01 -7.79009104e-01 5.60083270e-01 5.53835928e-01
-6.10159993e-01 1.03651714e+00 1.22984879e-01 -1.69877350e-01
-8.17319036e-01 -1.03157508e+00 -6.33948982e-01 -4.90026295e-01
-5.30266523e-01 1.20853436e+00 7.83094585e-01 -5.65205991e-01
1.84567273e-01 4.63755727e-01 -1.23942101e+00 9.00524035e-02
5.35950124e-01 4.87271518e-01 8.75279009e-01 7.02602148e-01
1.12031400e-01 7.47543752e-01 -5.22373557e-01 1.29267171e-01
4.81455773e-01 -2.53750622e-01 -1.29415655e+00 -9.62363407e-02
-4.53525782e-01 -3.75550598e-01 -2.41244942e-01 6.56223118e-01
1.03814729e-01 6.44558311e-01 2.38030002e-01 -1.80266440e+00
3.51728797e-01 6.30783916e-01 -4.99107927e-01 7.03917816e-02
8.61392021e-01 3.47258657e-01 9.19180155e-01 -3.14971328e-01
9.79645669e-01 1.22316980e+00 3.99705917e-01 3.75014335e-01
-1.06127441e+00 -1.36135697e-01 5.29740393e-01 3.26472670e-01
-1.30845392e+00 -1.73009843e-01 1.43550113e-01 -1.69135913e-01
1.62972891e+00 3.95971835e-01 6.84933603e-01 3.95249784e-01
4.27879363e-01 4.66919154e-01 9.31182563e-01 -1.97881550e-01
4.86045480e-01 -3.45721334e-01 -1.34434655e-01 5.04616201e-01
2.81922668e-02 2.06377044e-01 -5.64477801e-01 -3.91144246e-01
6.53304935e-01 -3.34453553e-01 -2.38258585e-01 -4.00213987e-01
-1.55023921e+00 8.44715595e-01 3.62432182e-01 8.60811546e-02
-6.94116294e-01 3.21571976e-01 3.77738625e-01 3.90092909e-01
-1.01648979e-01 7.41019726e-01 -2.98619002e-01 -2.94650227e-01
-3.69179398e-01 9.74188626e-01 9.95087981e-01 1.19543588e+00
1.03753138e+00 -5.76393604e-01 9.40482970e-03 2.06611186e-01
1.06229812e-01 4.90442514e-01 -1.67755231e-01 -1.46435571e+00
6.79670930e-01 3.92383873e-01 5.67403674e-01 -9.06057239e-01
-8.90576184e-01 9.50668566e-03 -2.68713295e-01 5.16366541e-01
4.37159956e-01 -1.05574250e-01 -5.55966198e-01 1.68388534e+00
5.81738651e-01 5.78667186e-02 4.57121223e-01 1.09099638e+00
1.13485172e-01 1.04641712e+00 -3.15535396e-01 -5.59767127e-01
1.31294370e+00 -1.28120017e+00 -4.62778687e-01 -4.01111543e-01
6.96218729e-01 -5.98680615e-01 5.00016570e-01 3.99159044e-01
-1.32038558e+00 5.72688952e-02 -8.25248837e-01 3.15775394e-01
-1.60652682e-01 -7.29356587e-01 5.42506635e-01 2.98802890e-02
-1.19098437e+00 4.70587075e-01 -1.04809511e+00 -5.22054434e-01
-1.68548256e-01 6.29852831e-01 -2.28382945e-01 -3.27885419e-01
-1.00448012e+00 1.02878654e+00 5.64126253e-01 -1.93133250e-01
-1.29768825e+00 -4.41704512e-01 -6.00268483e-01 1.14718340e-01
1.50399005e+00 -7.42273688e-01 1.64616334e+00 -1.74428329e-01
-1.36124873e+00 2.28759229e-01 -2.76853949e-01 -4.90377963e-01
3.65005732e-01 7.80705437e-02 4.23996709e-02 8.20426047e-02
7.15515256e-01 6.69123471e-01 3.66101056e-01 -1.22305667e+00
-1.07751012e+00 -1.32241502e-01 5.67127705e-01 7.62546420e-01
5.28717637e-01 4.58973497e-02 -3.86687756e-01 6.17165789e-02
-5.30866310e-02 -1.23462868e+00 -8.61743331e-01 -5.04591167e-01
-3.24019581e-01 -5.14648259e-01 3.97987425e-01 -3.18420604e-02
1.00144541e+00 -1.79778564e+00 7.02753723e-01 4.91415620e-01
2.17618451e-01 -2.22453669e-01 -3.35149437e-01 1.21446097e+00
4.64171827e-01 -2.06314117e-01 -6.28498271e-02 -2.42590964e-01
9.80950221e-02 6.23950958e-01 -1.79952189e-01 4.04739112e-01
-3.30226272e-01 7.12696552e-01 -1.40108693e+00 -3.15279812e-01
-5.97553812e-02 -2.05949426e-01 -5.13832390e-01 2.42399592e-02
-6.11954689e-01 2.56052673e-01 -6.88892722e-01 1.82841480e-01
4.17342246e-01 -5.72986677e-02 4.14044857e-01 3.30682218e-01
-6.26149356e-01 3.73499691e-01 -1.35600245e+00 1.93998563e+00
-2.98046529e-01 2.09986791e-01 3.69121373e-01 -4.56704259e-01
5.60744405e-01 2.09739417e-01 6.51366055e-01 -8.01348448e-01
-1.02907389e-01 3.35121125e-01 -9.74522531e-02 -1.65960029e-01
7.95813918e-01 1.45930737e-01 -4.68405277e-01 9.02778029e-01
-5.33481777e-01 -1.71579793e-01 7.31174886e-01 4.68749911e-01
1.56201041e+00 -1.16084017e-01 4.14345592e-01 -2.32388169e-01
4.24751401e-01 7.33866870e-01 7.35230446e-01 1.10885072e+00
-1.43291667e-01 -4.25937660e-02 4.03965563e-01 -7.32260108e-01
-6.64190769e-01 -9.25333440e-01 5.91252208e-01 1.12287104e+00
8.26708794e-01 -7.16705799e-01 -7.37163484e-01 -5.56808591e-01
-5.53104058e-02 7.05313802e-01 -3.34428132e-01 2.47036234e-01
-1.09696567e+00 -3.25334668e-01 7.98675790e-02 5.65376393e-02
3.44755381e-01 -1.28139186e+00 -1.55129910e+00 7.06629992e-01
-3.29748958e-01 -1.10630131e+00 -6.76138103e-01 3.47353935e-01
-2.00756297e-01 -1.29207575e+00 -3.57114494e-01 -5.82082212e-01
5.91832697e-01 7.50458717e-01 8.79564583e-01 -5.33207692e-02
1.97811984e-02 4.70411181e-01 -6.06596291e-01 -1.73229173e-01
-3.64035845e-01 4.14687455e-01 9.87485871e-02 -4.53225851e-01
-1.71540231e-01 -3.75885218e-01 -3.43085915e-01 5.42118788e-01
-5.02104580e-01 2.74293453e-01 4.14354444e-01 4.40564692e-01
7.47821689e-01 6.66946352e-01 4.80444461e-01 -5.36371648e-01
9.13343132e-01 -5.46723723e-01 -8.19704235e-01 2.51253545e-01
-4.45656240e-01 -8.79486427e-02 6.25986099e-01 -1.59780875e-01
-7.76968718e-01 3.72928269e-02 4.55128729e-01 -1.79713160e-01
-5.81236780e-02 7.79618919e-01 3.33781809e-01 -9.11340714e-02
3.99266750e-01 1.16531596e-01 7.14220554e-02 -8.24826211e-02
3.26825291e-01 1.48059512e-02 5.98133922e-01 -7.29294181e-01
5.78658223e-01 2.19631255e-01 2.70697981e-01 -2.61687398e-01
-4.33158100e-01 -4.91242647e-01 -1.11665323e-01 -3.28601152e-01
7.89277792e-01 -5.07057369e-01 -1.38744748e+00 3.58696491e-01
-1.40955758e+00 -8.74734938e-01 -2.91604519e-01 1.37277484e-01
-9.75832045e-01 2.51858328e-02 -3.89628649e-01 -1.05450416e+00
1.56545863e-01 -1.39726818e+00 9.29474354e-01 1.96404815e-01
-3.10792387e-01 -6.41547322e-01 4.65326905e-01 8.02106857e-02
4.21806842e-01 2.50845760e-01 6.89818323e-01 -6.19416296e-01
-1.05416822e+00 2.89483219e-01 -4.15279977e-02 -9.45696414e-01
-1.43498154e-02 -6.64010048e-01 -1.94020255e-03 -7.33037233e-01
-1.78101078e-01 -1.81267946e-03 2.66786844e-01 2.87168622e-01
1.25323206e-01 -5.33346236e-01 -8.74013782e-01 1.42693613e-02
1.53081596e+00 7.18347251e-01 4.41101074e-01 1.00558639e+00
1.59951657e-01 7.06389129e-01 1.14632201e+00 5.95990598e-01
9.23832655e-01 1.22108793e+00 7.61755764e-01 2.62060553e-01
3.10393453e-01 7.90748671e-02 3.64512116e-01 2.81909972e-01
-1.21051751e-01 -5.30148268e-01 -1.15084696e+00 6.54896975e-01
-2.45305443e+00 -8.57562900e-01 -2.74850857e-02 2.01730394e+00
4.90785480e-01 1.86121061e-01 4.97184098e-01 -3.82326394e-02
6.56809926e-01 -7.83473030e-02 -6.05229497e-01 -5.39852738e-01
2.64584780e-01 -1.99770316e-01 7.19263494e-01 9.68007505e-01
-7.70058393e-01 1.05897045e+00 5.36411619e+00 6.06942654e-01
-5.69535911e-01 1.50005773e-01 -2.60161366e-02 -2.85250485e-01
-4.52026688e-02 2.64297068e-01 -7.31399357e-01 1.94384977e-01
8.69164109e-01 -6.42675340e-01 1.31800044e+00 6.30875587e-01
3.49346370e-01 -7.04849184e-01 -1.15801215e+00 6.34945750e-01
-2.86718816e-01 -1.30363178e+00 -4.25568908e-01 2.78325140e-01
5.40559173e-01 8.85651261e-02 -3.74576241e-01 1.50730848e-01
7.07352757e-01 -8.48754406e-01 8.18827987e-01 1.71998397e-01
1.59847721e-01 -1.12889576e+00 6.57891512e-01 7.05677390e-01
-1.58816147e+00 -3.59141380e-01 -1.92433372e-01 -4.03164119e-01
8.98522556e-01 1.34381512e-03 -1.17877018e+00 1.08011723e+00
6.65736794e-01 1.69369429e-01 2.87217021e-01 1.04412508e+00
-3.47885350e-03 -4.03906405e-01 -5.34715950e-01 1.83310285e-02
7.27897048e-01 -1.24064803e-01 1.05055022e+00 7.12148666e-01
2.17569768e-01 4.80187267e-01 9.72393632e-01 6.34907663e-01
6.42896235e-01 -1.94549978e-01 -5.54015994e-01 3.23067427e-01
8.22655797e-01 1.05451453e+00 -9.81229126e-01 -1.56272173e-01
-1.94372579e-01 7.23746896e-01 3.65931153e-01 3.29993069e-01
-8.24798822e-01 -3.34442437e-01 7.57712960e-01 1.92231014e-01
8.43014792e-02 -3.72474790e-01 1.44953042e-01 -3.82737249e-01
-1.23447187e-01 -8.74727488e-01 4.54525948e-01 -7.28866398e-01
-6.72245681e-01 8.22840929e-01 4.70244974e-01 -9.53463376e-01
-5.71827233e-01 2.90538948e-02 -4.83721107e-01 6.79978609e-01
-1.74558890e+00 -7.11571455e-01 -3.31655800e-01 7.02331126e-01
7.64261425e-01 -4.48345393e-02 8.14076662e-01 -1.12793460e-01
-4.32790518e-01 -9.41836461e-02 -3.86955231e-01 -5.78180432e-01
2.71857172e-01 -1.12654519e+00 5.18254757e-01 6.99486911e-01
-3.83025527e-01 3.08219492e-01 8.14868391e-01 -9.32983220e-01
-1.72545433e+00 -8.82780969e-01 7.67332971e-01 1.24808781e-01
5.63931704e-01 4.62305620e-02 -6.11833811e-01 8.26410651e-01
2.93021768e-01 -4.36734051e-01 6.24975301e-02 -1.91652745e-01
2.49321952e-01 7.52198771e-02 -1.05801678e+00 6.67278171e-01
1.11837971e+00 -2.70340871e-02 -3.70772779e-01 4.89077568e-01
8.73363018e-01 -7.87213326e-01 -6.31568313e-01 2.53338486e-01
1.33658752e-01 -1.01225817e+00 8.15713406e-01 -2.73817360e-01
-1.21861838e-01 -6.95866704e-01 -1.37187690e-01 -1.58727312e+00
-5.33096075e-01 -1.20678580e+00 2.18267471e-01 7.05366015e-01
4.66746718e-01 -8.98453474e-01 7.94262946e-01 5.91987908e-01
-5.46192527e-01 -8.52628350e-01 -1.23798096e+00 -1.07378781e+00
-3.21741700e-01 -2.56774902e-01 9.57461238e-01 5.85652947e-01
4.55270082e-01 2.15179145e-01 -3.68087322e-01 5.46859026e-01
6.20014727e-01 2.63688058e-01 9.75453496e-01 -9.18902636e-01
-5.59500217e-01 -3.46840411e-01 1.72024623e-01 -1.10337341e+00
3.13790329e-03 -7.73491323e-01 4.61657375e-01 -1.90688705e+00
4.78544012e-02 -8.75064313e-01 2.59567320e-01 7.82651961e-01
2.06947148e-01 -4.87783432e-01 7.92302668e-01 3.69940430e-01
-1.08368802e+00 1.92018241e-01 1.45494747e+00 5.37788980e-02
-6.06021345e-01 -1.20640576e-01 -2.07200930e-01 3.67258072e-01
7.37020969e-01 -5.83335280e-01 -6.47139549e-01 -5.00781775e-01
6.23139560e-01 9.14889932e-01 6.64121285e-02 -8.66386116e-01
9.90949988e-01 -7.25393057e-01 -6.84089363e-01 -5.44759810e-01
5.37341058e-01 -9.03774023e-01 8.17153811e-01 7.86891818e-01
-1.24008611e-01 5.78904748e-01 1.83546394e-01 5.47667682e-01
5.68007976e-02 -4.41236377e-01 1.42444223e-01 -4.74367499e-01
-8.80989432e-01 5.13307154e-02 -8.65299165e-01 -2.95087975e-03
1.62412834e+00 -2.48639300e-01 -6.54537082e-01 -5.01661181e-01
-6.61800742e-01 9.43296075e-01 5.00439107e-01 2.66375095e-02
6.16309524e-01 -1.04570401e+00 -5.77294409e-01 -2.00232700e-01
-1.45034432e-01 5.37399709e-01 1.35665968e-01 1.10171342e+00
-4.13231820e-01 5.90070307e-01 -2.43438974e-01 -3.03711832e-01
-1.07567132e+00 8.55261266e-01 2.01380685e-01 -8.72002542e-01
-7.26513326e-01 4.76193041e-01 1.82223722e-01 -1.77284837e-01
1.49911135e-01 -1.82245880e-01 1.93294547e-02 -1.02432154e-01
5.87083340e-01 9.19233918e-01 -1.13903105e-01 -3.08660805e-01
-7.10237801e-01 6.09192371e-01 -2.47924015e-01 -3.53819728e-01
1.31080401e+00 -3.22275490e-01 -1.73628643e-01 -7.10995719e-02
4.50628400e-01 -1.89238384e-01 -1.18509734e+00 -8.97041783e-02
1.23477377e-01 -5.64270437e-01 -1.63433626e-01 -8.70385647e-01
-8.20357680e-01 -2.10269347e-01 -2.42101088e-01 5.96702516e-01
1.03329098e+00 2.25564063e-01 8.92194450e-01 6.25286460e-01
1.31474602e+00 -8.89793396e-01 7.79988617e-02 7.78982043e-01
8.12041759e-01 -5.90487719e-01 -4.49316390e-02 -8.48984480e-01
-7.48986840e-01 9.66162324e-01 5.99350095e-01 -1.00588582e-01
-1.82331473e-01 5.39104760e-01 -3.30970705e-01 -5.08957922e-01
-1.05767131e+00 -3.43752146e-01 -5.64711988e-01 6.58033729e-01
-7.49949694e-01 1.00081451e-01 -2.38519937e-01 1.62646145e-01
-2.62679845e-01 -9.26747918e-02 8.44999552e-01 1.31166112e+00
-7.50225544e-01 -1.24757087e+00 -4.55478698e-01 6.64545819e-02
2.30062053e-01 4.25048679e-01 -2.49123886e-01 8.97537947e-01
-9.37833786e-02 1.15211570e+00 9.82559547e-02 -2.00703382e-01
6.20600164e-01 -4.99812663e-01 5.20362377e-01 -6.67027473e-01
-7.27652073e-01 1.22814983e-01 6.74753428e-01 -1.03749251e+00
-3.33415598e-01 -7.96728432e-01 -1.82289016e+00 -4.22122002e-01
-8.50377902e-02 4.42174584e-01 2.89940804e-01 1.04325795e+00
3.16710204e-01 5.75357258e-01 5.75820208e-01 -1.27296102e+00
-2.75891334e-01 -2.59858489e-01 -2.13065282e-01 -9.41069126e-02
2.15338036e-01 -9.13992107e-01 -4.29701321e-02 -5.98745227e-01] | [4.964629173278809, 1.764493703842163] |
d78d70da-92ce-4554-a832-94742486e7a3 | action-anticipation-with-rbf | 1911.07806 | null | https://arxiv.org/abs/1911.07806v3 | https://arxiv.org/pdf/1911.07806v3.pdf | Action Anticipation with RBF Kernelized Feature Mapping RNN | We introduce a novel Recurrent Neural Network-based algorithm for future video feature generation and action anticipation called feature mapping RNN. Our novel RNN architecture builds upon three effective principles of machine learning, namely parameter sharing, Radial Basis Function kernels and adversarial training. Using only some of the earliest frames of a video, the feature mapping RNN is able to generate future features with a fraction of the parameters needed in traditional RNN. By feeding these future features into a simple multi-layer perceptron facilitated with an RBF kernel layer, we are able to accurately predict the action in the video. In our experiments, we obtain 18% improvement on JHMDB-21 dataset, 6% on UCF101-24 and 13% improvement on UT-Interaction datasets over prior state-of-the-art for action anticipation. | ['Yuge Shi', 'Richard Hartley', 'Basura Fernando'] | 2019-11-18 | action-anticipation-with-rbf-kernelized | http://openaccess.thecvf.com/content_ECCV_2018/html/Yuge_Shi_Action_Anticipation_with_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Yuge_Shi_Action_Anticipation_with_ECCV_2018_paper.pdf | eccv-2018-9 | ['action-anticipation'] | ['computer-vision'] | [ 4.68980938e-01 5.10896683e-01 -3.55406880e-01 -4.31001425e-01
-5.50872862e-01 -1.99588135e-01 6.18724883e-01 -6.18997931e-01
-4.83455420e-01 7.80234993e-01 7.53441691e-01 -6.11106902e-02
1.33819297e-01 -5.77889621e-01 -1.13974404e+00 -3.94456834e-01
-3.49727392e-01 4.67487983e-02 3.69529426e-02 -2.63918400e-01
-7.46077392e-03 3.68179560e-01 -1.52747834e+00 8.12196136e-01
2.03070104e-01 8.77391934e-01 4.79176342e-02 1.18644464e+00
5.00368118e-01 1.60523176e+00 -5.07297277e-01 -2.11752236e-01
3.63388538e-01 -4.93650228e-01 -1.19386220e+00 -1.31744802e-01
1.14693850e-01 -9.02182698e-01 -1.29640734e+00 2.30810940e-01
4.37720299e-01 5.91181874e-01 5.62915206e-01 -1.27160227e+00
-9.40702498e-01 9.21663225e-01 5.38112558e-02 7.55849481e-02
5.80190063e-01 4.97518986e-01 7.30871797e-01 -6.80797398e-01
8.89496326e-01 1.06413162e+00 7.77892232e-01 1.38656247e+00
-9.79354680e-01 -2.59906828e-01 4.43000235e-02 5.60765445e-01
-9.22064364e-01 -5.44672847e-01 6.01978064e-01 -1.92355588e-01
1.70235658e+00 2.81414390e-01 8.47800016e-01 1.79866731e+00
1.36160970e-01 1.22804427e+00 2.74121344e-01 -2.67904341e-01
-7.23235086e-02 -4.53653961e-01 -4.01475459e-01 4.90543276e-01
-6.92337930e-01 3.92095119e-01 -6.00471973e-01 8.40507075e-02
8.97985637e-01 1.60106480e-01 -3.88563871e-01 -9.17065963e-02
-1.52490950e+00 6.17544591e-01 5.26913106e-01 -1.87514052e-02
-5.67190051e-01 8.98121953e-01 8.25838745e-01 3.22545618e-01
4.15670842e-01 5.74875534e-01 -7.35727668e-01 -8.50271583e-01
-5.33238888e-01 3.38470608e-01 3.85973006e-01 8.03183377e-01
1.63317159e-01 3.68674397e-01 -5.41446984e-01 4.32423145e-01
-8.37030448e-03 4.21904474e-01 9.64224219e-01 -1.45499372e+00
5.18670440e-01 2.61489123e-01 3.45930219e-01 -6.94776714e-01
-4.36125338e-01 5.92305511e-02 -6.69252217e-01 1.93721175e-01
3.17191929e-01 -5.97092390e-01 -8.97313118e-01 1.81005895e+00
7.13197440e-02 8.00451517e-01 5.76596379e-01 5.66443086e-01
6.55392587e-01 8.03615987e-01 -4.48667072e-02 5.71385473e-02
5.42145789e-01 -1.29467630e+00 -4.49011981e-01 -1.54331356e-01
7.36437321e-01 -3.46678317e-01 8.13624084e-01 3.28710198e-01
-1.06465411e+00 -9.65310037e-01 -9.35268104e-01 -2.28966936e-01
-6.39068708e-02 2.05112547e-01 1.16805732e+00 7.29715377e-02
-1.34086239e+00 1.27543736e+00 -1.11696577e+00 -1.36074007e-01
4.56357002e-01 4.76074487e-01 -5.96554041e-01 1.65832356e-01
-1.18378019e+00 8.20863128e-01 5.12977123e-01 2.40267694e-01
-1.06112826e+00 -8.19562733e-01 -9.38176930e-01 -4.30688672e-02
1.07745431e-01 -8.77746224e-01 1.46529853e+00 -1.10887039e+00
-1.96557844e+00 3.76081914e-01 -5.00831716e-02 -1.02061224e+00
5.38051724e-01 -8.72080028e-01 -6.42679632e-01 1.87539667e-01
-3.19378465e-01 1.19372571e+00 9.22397673e-01 -3.60875428e-01
-6.43669426e-01 3.31519991e-01 2.02898458e-01 1.78991556e-01
-2.08055779e-01 -4.33841944e-02 -7.79356062e-02 -8.26180160e-01
-4.59489852e-01 -1.32277179e+00 -4.47354317e-01 7.25766793e-02
-3.78505170e-01 -2.35512704e-01 9.11722660e-01 -6.65917754e-01
9.27013874e-01 -2.17403817e+00 3.49982053e-01 -3.51661980e-01
-7.10775331e-02 4.15242195e-01 -5.21555126e-01 3.14190686e-01
-6.93868756e-01 -2.03804541e-02 3.76940906e-01 -4.82127294e-02
-5.65474518e-02 1.25398755e-01 -7.01520741e-01 2.49732003e-01
4.86576438e-01 1.33249271e+00 -1.07202017e+00 2.44889691e-01
4.67905074e-01 8.81194532e-01 -7.61632800e-01 4.75922734e-01
-5.85547090e-01 5.52253187e-01 -3.77943665e-01 3.61657023e-01
-6.11603521e-02 -1.60353661e-01 -9.87774134e-02 -1.79810837e-01
2.93782234e-01 2.38151222e-01 -5.88677466e-01 2.09860134e+00
-4.73251998e-01 9.93863046e-01 -9.66074944e-01 -7.49595344e-01
6.60560727e-01 6.54116333e-01 6.24272346e-01 -3.06625217e-01
-1.31216973e-01 -2.53311843e-01 -1.13652863e-01 -7.49960303e-01
5.31851768e-01 2.50709206e-01 -6.05162717e-02 4.86816049e-01
2.59214282e-01 4.04291451e-01 2.69553568e-02 7.30732605e-02
1.73263133e+00 9.29001868e-01 -4.36103605e-02 5.31935453e-01
2.37721965e-01 -3.84878516e-01 5.07969201e-01 6.21022403e-01
-1.57348961e-01 7.56078959e-01 4.91275281e-01 -9.88830924e-01
-1.13099313e+00 -8.64148498e-01 4.98993337e-01 1.31233001e+00
-6.56348169e-01 -4.17531431e-01 -6.73136353e-01 -8.90143871e-01
-5.08223437e-02 9.86739993e-01 -1.04585087e+00 -6.81918561e-01
-8.79854679e-01 -1.81279764e-01 7.52069950e-01 1.01784384e+00
4.47443396e-01 -1.90356553e+00 -8.70008886e-01 4.25853968e-01
-3.66693288e-01 -1.14737713e+00 -3.41656178e-01 3.61022167e-02
-9.43823218e-01 -9.97641861e-01 -7.08090603e-01 -3.91267806e-01
3.50867450e-01 -2.00643405e-01 1.26073098e+00 -8.45420659e-02
-1.77656859e-01 3.66407722e-01 -6.99999154e-01 -4.20246199e-02
-6.83611691e-01 -1.10034905e-01 2.34081134e-01 -1.38988376e-01
2.83025891e-01 -7.50408471e-01 -7.25616336e-01 -9.87328365e-02
-6.45267367e-01 2.49010518e-01 5.95808744e-01 1.06869686e+00
3.75597328e-01 -3.50637078e-01 5.37289679e-01 -6.80399537e-01
2.05697969e-01 -3.42349410e-01 -6.61434159e-02 1.56159922e-01
-7.77466446e-02 2.65507102e-01 8.80489647e-01 -7.45591104e-01
-1.08090258e+00 4.87622648e-01 -2.24829152e-01 -7.25108147e-01
-2.01857746e-01 1.43276036e-01 8.96047950e-02 1.83867350e-01
9.12704825e-01 3.52500409e-01 -6.51758462e-02 -2.42446214e-01
8.04096997e-01 4.00903136e-01 9.90318954e-01 -1.70095578e-01
4.10371631e-01 3.73453647e-01 -1.07264578e-01 -3.26489359e-01
-8.74588192e-01 6.48617670e-02 -7.41923988e-01 -3.64032537e-01
8.78137767e-01 -9.32061434e-01 -1.12520373e+00 7.10683584e-01
-1.14758372e+00 -7.45031655e-01 -7.51130998e-01 5.36776721e-01
-1.38007772e+00 3.10972519e-02 -7.68006265e-01 -5.11225700e-01
-5.40894687e-01 -8.01541090e-01 8.55777800e-01 2.15969265e-01
-5.01103878e-01 -7.25131512e-01 1.11236334e-01 1.61660865e-01
3.30730021e-01 6.39678597e-01 5.63202977e-01 -6.30269229e-01
-4.42095608e-01 -2.46882260e-01 2.96906173e-01 6.20212138e-01
1.23010144e-01 6.57206625e-02 -8.84983122e-01 -8.47791433e-02
-2.94790238e-01 -7.35633552e-01 9.74271476e-01 5.01994967e-01
1.41790164e+00 -4.26669419e-01 -3.20635915e-01 8.43170464e-01
1.01563621e+00 1.79493204e-01 1.16705835e+00 3.12428862e-01
7.30480254e-01 3.79805490e-02 7.18939424e-01 7.40116358e-01
8.50932375e-02 5.71448624e-01 4.73444641e-01 1.44366905e-01
-2.20186204e-01 -4.42786664e-01 8.11488867e-01 1.92700103e-01
-4.64887142e-01 -3.13586026e-01 -4.51538533e-01 5.55653572e-01
-2.25926232e+00 -1.37884915e+00 4.09394056e-01 1.88212454e+00
7.75268376e-01 1.84762582e-01 2.11559176e-01 -4.14578132e-02
4.59320575e-01 2.25363493e-01 -9.01298642e-01 -4.86129105e-01
1.62633702e-01 2.76689738e-01 2.62077689e-01 9.60327312e-02
-1.40229380e+00 1.21503127e+00 7.15418053e+00 5.65881729e-01
-9.17333961e-01 -1.71412677e-01 6.51778698e-01 -5.71290970e-01
2.86780149e-02 -2.77662426e-01 -4.73096728e-01 4.97274190e-01
1.48047292e+00 6.97618127e-02 7.39872456e-01 9.34359252e-01
2.64943838e-01 2.50141501e-01 -1.57010734e+00 1.03526819e+00
4.43527587e-02 -1.60916269e+00 1.11469887e-01 -4.15904731e-01
5.91884077e-01 3.99986923e-01 1.69005617e-01 6.81939840e-01
5.07990003e-01 -1.40195632e+00 4.31547612e-01 1.22463465e+00
1.06028724e+00 -9.94276106e-01 4.38524723e-01 3.08786351e-02
-8.19422603e-01 -3.34469914e-01 -4.63591248e-01 -3.33151430e-01
3.37091744e-01 7.32814819e-02 -9.21262085e-01 1.33696586e-01
6.03921115e-01 1.15421069e+00 -4.35323685e-01 6.84381127e-01
-4.50662643e-01 5.85602105e-01 -5.51384166e-02 1.28989890e-01
1.66123345e-01 4.11706477e-01 3.32299501e-01 9.77720678e-01
2.44508713e-01 1.70651451e-02 -1.48862258e-01 3.61495227e-01
-2.93193966e-01 -4.42267925e-01 -8.44588399e-01 -2.01628208e-01
1.16456747e-01 1.04649639e+00 -6.31420240e-02 -3.79235178e-01
-3.80704403e-01 1.61966121e+00 5.59059739e-01 4.15073991e-01
-1.12098801e+00 -3.82872075e-01 8.58204782e-01 -2.71394491e-01
6.47956729e-01 2.92138401e-02 4.28806126e-01 -1.09994853e+00
7.29517406e-03 -1.03521323e+00 1.92470565e-01 -1.13210487e+00
-8.25446427e-01 8.64340544e-01 -2.94294089e-01 -1.39111471e+00
-1.31464136e+00 -6.05454385e-01 -5.02738118e-01 5.15737236e-01
-8.55935335e-01 -1.44591796e+00 1.25720888e-01 6.04451537e-01
6.46349251e-01 -5.06461263e-01 1.39235878e+00 -9.60372537e-02
-3.14985722e-01 6.77665770e-01 1.89583451e-01 3.84459734e-01
5.13255477e-01 -1.13889027e+00 1.19458342e+00 6.88881457e-01
3.30277890e-01 1.42935529e-01 6.65106714e-01 -6.14548206e-01
-1.27183354e+00 -1.39744949e+00 8.35892081e-01 -6.52232885e-01
6.37872696e-01 -2.34423608e-01 -5.17165244e-01 1.37568724e+00
4.68648560e-02 4.46428865e-01 6.26824260e-01 4.01257761e-02
-2.80342847e-01 6.35149255e-02 -9.91064131e-01 8.47741187e-01
1.46063519e+00 -5.39115906e-01 -5.07019579e-01 6.94445431e-01
1.13357472e+00 -7.20500708e-01 -9.34752107e-01 5.12878478e-01
8.58852267e-01 -9.35936630e-01 1.27296412e+00 -1.43619013e+00
8.77864599e-01 1.68393746e-01 -4.18131612e-02 -1.46390152e+00
-5.82429588e-01 -1.11291206e+00 -8.40413451e-01 8.78540576e-01
4.95191365e-01 -1.60648599e-01 1.04074478e+00 9.00477886e-01
-2.20583171e-01 -1.12816036e+00 -8.66327465e-01 -6.28042519e-01
-2.67224550e-01 -6.87473595e-01 5.36486089e-01 5.49535692e-01
9.24480483e-02 1.76592767e-01 -1.12798762e+00 -3.35614592e-01
-5.38880900e-02 -2.64580339e-01 8.06291521e-01 -4.66783762e-01
-7.88911819e-01 -6.25406206e-03 -7.60973811e-01 -1.18664455e+00
6.15590632e-01 -6.83122635e-01 -1.49166986e-01 -1.31013381e+00
2.23223761e-01 1.33070707e-01 -5.43071449e-01 9.48809862e-01
-5.62959060e-04 2.83370733e-01 3.47750247e-01 -2.49519348e-01
-8.65234554e-01 8.49433422e-01 1.10736573e+00 -5.74267618e-02
-1.81759700e-01 1.52092636e-01 -4.81444389e-01 8.92280579e-01
6.78810656e-01 -4.00095224e-01 -4.22591895e-01 -5.11579037e-01
1.47158340e-01 3.55579883e-01 4.46786493e-01 -1.24745107e+00
-7.72235915e-02 -1.14793338e-01 1.01303136e+00 -4.12658602e-01
7.89909244e-01 -6.45770311e-01 2.05651686e-01 5.42664945e-01
-6.76125050e-01 1.09492749e-01 2.55529463e-01 7.12018192e-01
1.95450336e-01 -1.41511895e-02 2.22580552e-01 -1.27381980e-01
-1.05506575e+00 2.17698529e-01 -4.49923903e-01 -1.07716009e-01
1.04794216e+00 -7.47666955e-02 -3.17226261e-01 -7.08008468e-01
-1.04628611e+00 -5.47397174e-02 1.45178050e-01 7.78412580e-01
9.89619792e-01 -1.68689370e+00 -9.01997924e-01 7.19860718e-02
-2.10713670e-01 -4.31643009e-01 4.28859115e-01 3.75850022e-01
-4.00084794e-01 3.78350228e-01 -4.95910883e-01 -2.53885627e-01
-1.04635525e+00 7.34333634e-01 3.68618548e-01 -2.74973124e-01
-9.31882441e-01 1.16713130e+00 -2.82024562e-01 -2.56794721e-01
1.04643010e-01 -3.91312301e-01 -2.59547561e-01 -4.88506645e-01
8.37164521e-01 3.60172093e-01 -4.84455526e-01 -5.50599456e-01
-2.73801357e-01 3.35594937e-02 -2.37206310e-01 -1.62725858e-02
1.51897085e+00 2.79495806e-01 3.44087452e-01 3.82509351e-01
1.21273696e+00 -4.72197473e-01 -1.82306325e+00 3.21841724e-02
-3.08988333e-01 -3.58189404e-01 -2.30078712e-01 -9.95758057e-01
-1.00335717e+00 3.29816937e-01 6.69663012e-01 -2.40723297e-01
1.21462750e+00 -1.16948850e-01 1.18909061e+00 9.80350792e-01
4.59231362e-02 -1.08216953e+00 3.14977080e-01 6.96977079e-01
1.17216551e+00 -1.05601823e+00 -2.24316880e-01 1.00422792e-01
-8.29753399e-01 1.30042386e+00 6.75443888e-01 -5.37508905e-01
4.96634245e-01 2.04152048e-01 6.67640101e-03 1.33047700e-01
-1.22382367e+00 1.64120942e-01 3.90972972e-01 6.97610438e-01
4.50197428e-01 -1.28915468e-02 3.44694167e-01 5.44461131e-01
-3.19916219e-01 4.72192138e-01 5.12741446e-01 7.25670636e-01
-5.41110598e-02 -9.98628020e-01 2.24668115e-01 6.25174999e-01
-5.32493114e-01 -2.03157425e-01 -5.48039302e-02 7.59494960e-01
-1.75345108e-01 7.29687810e-01 3.23744744e-01 -8.99839103e-01
2.38303214e-01 2.34256193e-01 6.59589112e-01 -3.94182682e-01
-5.50986350e-01 -3.77817661e-01 3.83960485e-01 -1.23634589e+00
-4.16403860e-01 -6.95693970e-01 -1.51455748e+00 -2.33814865e-01
-7.68197775e-02 -3.58388752e-01 2.42468864e-01 1.05432749e+00
7.43106663e-01 9.15169775e-01 6.94552600e-01 -1.06190908e+00
-6.45413399e-01 -1.28800905e+00 -2.03616813e-01 4.47882652e-01
3.14812571e-01 -4.47125912e-01 2.02981364e-02 4.20348763e-01] | [8.119577407836914, 0.4516393542289734] |
db005b53-cda3-43d6-a330-e532ca274c1c | two-decades-of-colorization-and | 2204.13322 | null | https://arxiv.org/abs/2204.13322v2 | https://arxiv.org/pdf/2204.13322v2.pdf | Two Decades of Colorization and Decolorization for Images and Videos | Colorization is a computer-aided process, which aims to give color to a gray image or video. It can be used to enhance black-and-white images, including black-and-white photos, old-fashioned films, and scientific imaging results. On the contrary, decolorization is to convert a color image or video into a grayscale one. A grayscale image or video refers to an image or video with only brightness information without color information. It is the basis of some downstream image processing applications such as pattern recognition, image segmentation, and image enhancement. Different from image decolorization, video decolorization should not only consider the image contrast preservation in each video frame, but also respect the temporal and spatial consistency between video frames. Researchers were devoted to develop decolorization methods by balancing spatial-temporal consistency and algorithm efficiency. With the prevalance of the digital cameras and mobile phones, image and video colorization and decolorization have been paid more and more attention by researchers. This paper gives an overview of the progress of image and video colorization and decolorization methods in the last two decades. | ['Shiguang Liu'] | 2022-04-28 | null | null | null | null | ['colorization'] | ['computer-vision'] | [ 6.34412289e-01 -6.22499287e-01 -3.15760933e-02 1.61064550e-01
-8.77173431e-03 -4.72916245e-01 2.54177034e-01 -2.60061264e-01
-6.43421173e-01 7.13924348e-01 -2.13308185e-01 -4.31763947e-01
3.14097017e-01 -7.48079658e-01 -2.91764766e-01 -9.91129756e-01
3.95111620e-01 -4.97264057e-01 1.86250165e-01 4.18161601e-02
4.26534325e-01 6.81958795e-01 -1.33653021e+00 1.58347651e-01
9.13265169e-01 7.82555640e-01 1.75199538e-01 8.99683416e-01
-4.44575548e-01 6.46048307e-01 -6.03610218e-01 -2.30313182e-01
1.82362393e-01 -9.50228631e-01 -2.31735528e-01 4.63132113e-01
2.07227245e-01 -4.30304289e-01 -5.60596466e-01 1.66526365e+00
2.11745784e-01 -4.18330692e-02 3.59378070e-01 -1.22588420e+00
-1.21913087e+00 -2.63429638e-02 -1.07672346e+00 2.51281321e-01
1.92797303e-01 2.89876033e-02 1.41419306e-01 -6.88064933e-01
5.67233503e-01 1.08649802e+00 2.37302527e-01 5.32663465e-01
-8.92934203e-01 -6.07032835e-01 -8.38353485e-02 4.40565556e-01
-1.32439744e+00 -6.53228834e-02 8.27613115e-01 -1.73052832e-01
2.07059428e-01 6.72382593e-01 9.26150918e-01 2.27266356e-01
3.86241108e-01 9.81387854e-01 1.26183987e+00 -7.28878081e-01
1.28684059e-01 1.38819143e-01 -1.63059339e-01 6.28974140e-01
5.90660870e-01 9.89101529e-02 -1.66219324e-01 2.88915604e-01
8.78574848e-01 6.59469247e-01 -4.64885890e-01 -3.06863710e-02
-9.96504605e-01 1.61627322e-01 1.50234506e-01 6.45428836e-01
-1.73272505e-01 7.24155232e-02 6.66372571e-03 2.36381769e-01
2.21984565e-01 2.38775790e-01 2.08933845e-01 -3.30782086e-01
-1.12336779e+00 -4.24470335e-01 5.15708387e-01 6.64244831e-01
8.07638645e-01 4.34449852e-01 1.49585065e-02 6.44276261e-01
1.09188259e-02 9.27271724e-01 4.95675772e-01 -1.05277765e+00
2.23870322e-01 6.33370936e-01 2.12189659e-01 -1.03845012e+00
6.37975987e-03 2.47142926e-01 -9.59605515e-01 7.58640230e-01
3.07921529e-01 -2.41802201e-01 -1.03444231e+00 1.08648288e+00
9.77996737e-02 -1.78850237e-02 -4.29784581e-02 1.12183619e+00
6.56534255e-01 1.06438196e+00 -1.42633617e-01 -6.57970309e-01
1.17334116e+00 -8.14677715e-01 -1.12739718e+00 5.53728752e-02
4.82596606e-02 -1.14071703e+00 9.36061680e-01 7.75395811e-01
-1.24266171e+00 -5.59703767e-01 -1.18970990e+00 3.89157273e-02
-4.22208548e-01 2.28307873e-01 2.87163079e-01 1.22783279e+00
-8.40005934e-01 3.91783416e-01 -6.01090133e-01 -3.50040764e-01
8.99911895e-02 6.30275309e-02 -4.18068767e-01 -4.59927768e-01
-7.84994781e-01 6.31031215e-01 2.36326128e-01 3.86660606e-01
-4.18186247e-01 -3.40461940e-01 -3.91267270e-01 4.55188826e-02
1.95908964e-01 -1.28050148e-01 5.74618816e-01 -1.56431365e+00
-1.59401691e+00 9.42079365e-01 -2.55263358e-01 3.81736197e-02
6.15962923e-01 -2.12948456e-01 -9.33754086e-01 5.46417296e-01
-3.38702559e-01 3.08498859e-01 1.09022260e+00 -1.36738563e+00
-8.26258421e-01 -2.05426231e-01 -2.40619078e-01 2.33032048e-01
-2.16590643e-01 1.55921414e-01 -9.24459159e-01 -6.66554034e-01
3.51565003e-01 -6.75182521e-01 1.06907189e-01 9.86861661e-02
-4.73829150e-01 3.65064591e-01 1.37185764e+00 -8.38062108e-01
1.31777644e+00 -2.47307277e+00 -6.86157271e-02 2.22163633e-01
2.75600016e-01 5.25382280e-01 -1.27559602e-01 3.78602117e-01
-1.38257369e-01 1.01464197e-01 -3.09623748e-01 1.73420340e-01
-4.06194240e-01 -1.94335300e-02 4.68480960e-02 8.52589607e-01
-6.22306615e-02 6.24435782e-01 -6.63401842e-01 -7.11967528e-01
4.39320594e-01 4.73352134e-01 -3.79092276e-01 -1.31199986e-01
3.01057547e-01 5.15711308e-01 -4.75107729e-01 8.46970081e-01
1.22115898e+00 2.22977191e-01 1.33754283e-01 -3.10412645e-01
-4.77531165e-01 -8.22726667e-01 -9.85344350e-01 1.15001738e+00
-1.23234078e-01 1.04660976e+00 3.83948714e-01 -8.07353020e-01
8.06784213e-01 9.35526937e-02 7.47696638e-01 -1.03307056e+00
2.30555579e-01 2.52969772e-01 -1.92049250e-01 -7.27318048e-01
9.05061483e-01 -1.98267981e-01 2.98339307e-01 3.74240398e-01
-6.69830322e-01 -2.98452675e-01 4.68962222e-01 2.24933475e-01
4.90992159e-01 9.71802790e-03 -5.38952723e-02 -1.72642708e-01
7.00337172e-01 3.88538577e-02 5.98271847e-01 4.23767954e-01
-1.49227768e-01 7.50038743e-01 2.35172987e-01 -1.66501239e-01
-1.12372315e+00 -7.89497316e-01 3.93044278e-02 7.04582632e-01
9.79891598e-01 6.74975514e-02 -7.95791566e-01 -1.32071525e-01
-1.11288279e-01 4.24007446e-01 -5.16704321e-01 -2.40580574e-01
-6.35873854e-01 -5.25972962e-01 3.21351051e-01 -3.94747034e-03
9.41018701e-01 -9.89749432e-01 -4.93657082e-01 1.10774562e-01
7.79361278e-02 -6.40823841e-01 -8.15015078e-01 -1.37765363e-01
-7.82790720e-01 -1.40245748e+00 -1.28427923e+00 -9.15182173e-01
1.05254960e+00 9.01438057e-01 5.13047814e-01 3.78697753e-01
-6.28105938e-01 4.90106374e-01 -7.52201438e-01 -1.94519069e-02
-4.75220382e-01 -5.38891077e-01 -6.18112125e-02 3.53356630e-01
2.66224414e-01 -3.00711900e-01 -9.13391888e-01 3.69184613e-01
-1.61206770e+00 2.21279621e-01 6.55737102e-01 6.15876436e-01
4.93258983e-01 3.41964513e-01 -1.74597621e-01 -7.89495468e-01
5.94135046e-01 1.88697234e-01 -6.36371434e-01 7.40306497e-01
-5.36180437e-01 -1.50543660e-01 6.02598608e-01 -3.11074257e-01
-1.20345473e+00 -1.53022289e-01 2.16782480e-01 -3.93083543e-01
-2.18639355e-02 2.20289901e-01 -2.72994012e-01 -3.96887094e-01
3.18641543e-01 6.78803086e-01 1.86626986e-01 -4.84698653e-01
6.10490620e-01 7.63758957e-01 8.36896181e-01 -1.75276950e-01
8.07573617e-01 8.61927807e-01 -1.18716545e-01 -1.22989655e+00
1.24900423e-01 -3.97266805e-01 -5.90572357e-01 -4.73290384e-01
1.14814687e+00 -3.13399941e-01 -7.54231036e-01 7.83891380e-01
-9.84463215e-01 -8.07318240e-02 -7.39408955e-02 5.40631890e-01
2.58824062e-02 9.22641337e-01 -6.28364503e-01 -8.07493567e-01
-1.67856663e-01 -1.02884936e+00 5.29354990e-01 8.92740190e-01
3.98663729e-01 -9.69487906e-01 -1.81678340e-01 -3.99108455e-02
4.20696676e-01 1.17638387e-01 8.84300172e-01 3.43998671e-01
-6.38913333e-01 -3.58312756e-01 -4.82138664e-01 5.53555489e-01
4.40671057e-01 4.79971409e-01 -6.10174298e-01 -2.54779488e-01
-3.74500975e-02 1.70754641e-01 9.71974373e-01 5.47474563e-01
9.93715048e-01 -1.75983146e-01 -2.63006032e-01 7.40032077e-01
1.52653515e+00 1.05104744e+00 1.23104715e+00 3.68064612e-01
6.02091134e-01 2.38542810e-01 5.39093673e-01 2.53883570e-01
-3.81503522e-01 3.51721644e-01 2.37840414e-01 -8.93075347e-01
-3.85912806e-01 -2.22972348e-01 1.87062189e-01 5.78136086e-01
-2.60508597e-01 -2.33572155e-01 -4.13803697e-01 1.19899087e-01
-1.25217092e+00 -1.13053238e+00 -2.85765201e-01 2.15479231e+00
5.89938819e-01 -3.50996584e-01 -1.70479357e-01 1.63005248e-01
1.27177238e+00 -6.44833371e-02 -7.42773056e-01 -3.27249914e-01
-5.60827434e-01 -5.47485910e-02 8.34372401e-01 4.51467335e-01
-8.70728314e-01 9.95883346e-01 6.68705511e+00 8.88400614e-01
-1.60424495e+00 -4.28761467e-02 7.67646730e-01 1.34269178e-01
-5.10531247e-01 -7.43253110e-03 -1.37855753e-01 5.96446216e-01
-1.37664378e-01 -2.47779801e-01 7.70675540e-01 5.18993199e-01
4.14988488e-01 -7.61138737e-01 -5.18235862e-01 1.49525523e+00
2.11897552e-01 -9.95295107e-01 3.10009569e-02 -4.94352281e-02
1.03849316e+00 -7.27937937e-01 4.67861682e-01 -2.46694490e-01
-2.65302416e-02 -7.56734252e-01 7.83410192e-01 4.59247708e-01
1.09915614e+00 -4.28333730e-01 3.50942761e-01 -1.92595974e-01
-1.14625859e+00 6.76143020e-02 -4.03894603e-01 1.40070409e-01
3.39395642e-01 4.86062586e-01 3.28296386e-02 3.63692313e-01
5.81291974e-01 7.60845542e-01 -4.15555477e-01 1.42593133e+00
-2.28287593e-01 3.52153689e-01 -3.95908430e-02 -1.23723060e-01
2.86213934e-01 -1.02375102e+00 2.83177435e-01 1.18110359e+00
4.96586770e-01 4.94408309e-01 -4.52966243e-02 8.50891352e-01
1.32174000e-01 -1.52148390e-02 -2.10311860e-01 -3.95129830e-01
3.35118711e-01 1.08421278e+00 -1.00930798e+00 -6.21514738e-01
-5.41713476e-01 1.39363122e+00 -7.59864092e-01 9.86531734e-01
-7.98654079e-01 -9.71619725e-01 3.84712905e-01 4.78399657e-02
9.84499380e-02 -4.19335723e-01 -2.57087737e-01 -1.08971345e+00
-2.21876860e-01 -8.12804639e-01 1.35845914e-01 -9.22982693e-01
-6.95417166e-01 4.18491334e-01 -2.42688969e-01 -1.48515809e+00
4.13471669e-01 -6.94778025e-01 -7.57604063e-01 8.25056911e-01
-1.53848672e+00 -5.81905782e-01 -5.10837793e-01 6.98745251e-01
4.02213216e-01 -1.52049601e-01 1.96168065e-01 3.51620615e-01
-7.48929679e-01 1.24641918e-01 8.91902864e-01 7.46180192e-02
7.93390393e-01 -8.11034143e-01 -1.57632485e-01 1.29444873e+00
-1.54483795e-01 4.41850901e-01 8.68394136e-01 -6.04112804e-01
-1.67993724e+00 -4.46345240e-01 3.38738292e-01 3.84331435e-01
1.83207348e-01 2.48420119e-01 -7.25759566e-01 1.63200498e-02
5.38093925e-01 -3.05131406e-01 3.25410068e-01 -5.97314417e-01
8.01131576e-02 -4.07855302e-01 -1.08066881e+00 6.58107162e-01
5.22854924e-01 -4.49263662e-01 -1.69926584e-01 -2.72832848e-02
1.99777097e-01 -2.64592797e-01 -2.03155115e-01 -1.73663139e-01
7.72502899e-01 -1.16457832e+00 7.35052943e-01 -2.01725036e-01
2.68746018e-01 -7.33844817e-01 1.14874795e-01 -9.35083330e-01
-2.69325346e-01 -6.69381499e-01 7.47737169e-01 9.49021935e-01
1.30913675e-01 -5.32983482e-01 8.18831921e-01 8.44168663e-01
2.32886076e-02 1.11621127e-01 -6.90483272e-01 -6.03784859e-01
-2.17445359e-01 -6.41622469e-02 2.37659216e-01 8.03037286e-01
2.15317652e-01 -4.54724580e-01 -7.04674184e-01 -2.69510560e-02
6.30984128e-01 3.11309159e-01 4.97767687e-01 -6.72080219e-01
-7.06659034e-02 -7.22094119e-01 -2.78713584e-01 -1.04963315e+00
-4.38285917e-01 -4.16568249e-01 -4.07959595e-02 -1.68844247e+00
2.47719407e-01 -1.74684122e-01 -2.81033278e-01 2.32379079e-01
-2.62333661e-01 7.11263478e-01 4.36050028e-01 5.76377749e-01
-5.68435073e-01 4.19229895e-01 1.52954757e+00 -5.16684592e-01
-4.49519962e-01 -2.79773504e-01 -5.53167522e-01 4.87592697e-01
7.24120498e-01 -1.37499884e-01 -1.72386631e-01 -5.31888723e-01
8.81200433e-02 1.30169779e-01 6.22187257e-02 -8.17206502e-01
2.77241021e-01 -4.28344697e-01 5.56419790e-01 -3.15091401e-01
1.69981167e-01 -1.00502062e+00 3.33620816e-01 8.55997205e-01
6.36586100e-02 1.31859511e-01 -1.46048628e-02 5.61424315e-01
-5.20757616e-01 -3.04693162e-01 1.07225645e+00 -2.07044482e-01
-1.21479273e+00 2.22187880e-02 -8.36267710e-01 -3.68101448e-01
1.36177027e+00 -8.93796146e-01 -5.11998117e-01 -5.98961830e-01
-5.54569304e-01 -2.62302041e-01 7.93273389e-01 2.45500639e-01
9.40895200e-01 -1.21279669e+00 -3.94276381e-01 4.68323797e-01
-3.01380575e-01 -7.50860095e-01 7.16487765e-01 9.10007954e-01
-1.50055456e+00 2.04943627e-01 -4.89799142e-01 -2.63398230e-01
-1.56096482e+00 7.46044934e-01 2.57319361e-01 3.73583794e-01
-6.42082512e-01 5.27962446e-01 4.38623697e-01 6.13434494e-01
3.07368208e-02 -3.36036325e-01 -1.24316737e-01 -3.90767008e-01
8.16070974e-01 8.00159812e-01 -4.05940831e-01 -4.94229376e-01
-1.01213336e-01 9.69645143e-01 1.38100281e-01 -2.87733972e-01
7.70805240e-01 -4.45530325e-01 -4.44479734e-01 7.65690953e-02
1.24005425e+00 3.41911644e-01 -1.34283531e+00 1.06483363e-01
-3.52386087e-01 -1.02243721e+00 8.81818086e-02 -6.67311370e-01
-1.54328060e+00 1.02983606e+00 7.55433977e-01 5.30169964e-01
1.38788211e+00 -4.05019522e-01 7.01960087e-01 3.30969766e-02
2.04994202e-01 -1.37656796e+00 2.28909522e-01 -1.06626906e-01
5.06565392e-01 -8.32922637e-01 1.41897157e-01 -2.98128158e-01
-8.67629528e-01 1.35953641e+00 2.47880071e-01 6.77302629e-02
3.76563579e-01 1.99322939e-01 2.91789979e-01 1.66551217e-01
7.19073191e-02 -1.58405215e-01 1.84523463e-01 5.19167125e-01
4.01579320e-01 -9.89909247e-02 -6.69983983e-01 1.14386313e-01
4.50536102e-01 1.11055136e-01 5.30607283e-01 8.74221146e-01
-9.27939475e-01 -1.13740420e+00 -9.56709564e-01 3.09118116e-03
-3.74724269e-01 -1.44012854e-01 -3.48382413e-01 6.50062978e-01
1.29227072e-01 1.18906951e+00 -3.91659625e-02 -2.77493179e-01
1.61760151e-02 -2.92531312e-01 5.01658559e-01 6.99633881e-02
1.05075151e-01 3.57358843e-01 -4.89914864e-01 -3.37527514e-01
-5.65054655e-01 -2.83219755e-01 -1.23726296e+00 -6.65433109e-01
-1.88964233e-01 4.10524338e-01 9.31043208e-01 5.35887778e-01
-2.96261460e-02 4.50111419e-01 6.78145230e-01 -6.75255895e-01
3.12914044e-01 -5.55817246e-01 -1.45983350e+00 6.00430310e-01
2.11216360e-01 -7.93347806e-02 -1.82229802e-01 5.03453255e-01] | [10.845096588134766, -2.4383151531219482] |
d29b0cb3-21d8-4119-86cf-a5c2c292fe42 | twin-two-stage-interest-network-for-lifelong | 2302.02352 | null | https://arxiv.org/abs/2302.02352v2 | https://arxiv.org/pdf/2302.02352v2.pdf | TWIN: TWo-stage Interest Network for Lifelong User Behavior Modeling in CTR Prediction at Kuaishou | Life-long user behavior modeling, i.e., extracting a user's hidden interests from rich historical behaviors in months or even years, plays a central role in modern CTR prediction systems. Conventional algorithms mostly follow two cascading stages: a simple General Search Unit (GSU) for fast and coarse search over tens of thousands of long-term behaviors and an Exact Search Unit (ESU) for effective Target Attention (TA) over the small number of finalists from GSU. Although efficient, existing algorithms mostly suffer from a crucial limitation: the \textit{inconsistent} target-behavior relevance metrics between GSU and ESU. As a result, their GSU usually misses highly relevant behaviors but retrieves ones considered irrelevant by ESU. In such case, the TA in ESU, no matter how attention is allocated, mostly deviates from the real user interests and thus degrades the overall CTR prediction accuracy. To address such inconsistency, we propose \textbf{TWo-stage Interest Network (TWIN)}, where our Consistency-Preserved GSU (CP-GSU) adopts the identical target-behavior relevance metric as the TA in ESU, making the two stages twins. Specifically, to break TA's computational bottleneck and extend it from ESU to GSU, or namely from behavior length $10^2$ to length $10^4-10^5$, we build a novel attention mechanism by behavior feature splitting. For the video inherent features of a behavior, we calculate their linear projection by efficient pre-computing \& caching strategies. And for the user-item cross features, we compress each into a one-dimentional bias term in the attention score calculation to save the computational cost. The consistency between two stages, together with the effective TA-based relevance metric in CP-GSU, contributes to significant performance gain in CTR prediction. | ['Kun Gai', 'Yang song', 'Yanan Niu', 'Dewei Leng', 'Yiqun Hui', 'Jing Lu', 'Lin Guan', 'Xiaoxue Zang', 'Zhiyi Fu', 'Chenbin Zhang', 'Jianxin Chang'] | 2023-02-05 | null | null | null | null | ['click-through-rate-prediction'] | ['miscellaneous'] | [ 3.54190581e-02 -5.85089087e-01 -6.06450915e-01 -2.19841048e-01
-6.22940242e-01 -1.65936261e-01 -5.15004657e-02 -2.67441850e-02
-3.38749468e-01 5.36078632e-01 2.08637878e-01 -3.20033997e-01
-2.70410746e-01 -6.06898069e-01 -4.62063074e-01 -5.80314100e-01
2.05232576e-02 3.13298851e-01 5.59228539e-01 -4.41249639e-01
2.80867070e-01 1.94191203e-01 -1.58342540e+00 1.07694194e-01
8.52737606e-01 1.54649687e+00 3.58989507e-01 3.23397994e-01
-2.12362021e-01 9.20881867e-01 -3.83336753e-01 -3.92636001e-01
2.24597737e-01 -3.66739780e-01 -6.02918863e-01 -4.99273747e-01
-4.40596938e-02 -5.98583639e-01 -6.89520895e-01 7.89233387e-01
4.95454073e-01 4.89036590e-01 3.73427212e-01 -1.29970562e+00
-7.77475059e-01 6.94940746e-01 -8.49208891e-01 6.24803662e-01
3.79483908e-01 9.89159420e-02 1.44314837e+00 -7.61088490e-01
3.15449655e-01 1.06716597e+00 4.86450762e-01 3.75536591e-01
-8.80558550e-01 -9.66848433e-01 6.73554659e-01 5.52333176e-01
-1.47703981e+00 -2.88308650e-01 6.76330268e-01 -1.51563913e-01
9.15413439e-01 6.92438006e-01 5.19958436e-01 9.85601425e-01
8.62829685e-02 9.97543871e-01 4.11729664e-01 1.03557996e-01
1.49777904e-01 5.99985011e-02 8.18711936e-01 2.70529330e-01
7.13437945e-02 -1.92182958e-01 -4.32223439e-01 -2.77244240e-01
5.21038651e-01 5.79249620e-01 -2.90940106e-01 2.58096993e-01
-8.69122565e-01 6.59432352e-01 4.87831026e-01 3.35216761e-01
-2.81958014e-01 8.93627182e-02 3.19558233e-01 4.06520873e-01
3.32832903e-01 1.74367681e-01 -6.79417014e-01 -3.03420931e-01
-7.56985724e-01 -6.03821175e-03 4.97949868e-01 1.25684428e+00
8.13044667e-01 -8.53888243e-02 -6.19915009e-01 9.75259304e-01
1.92546677e-02 5.21266222e-01 7.05247104e-01 -6.11232460e-01
3.29790384e-01 6.75639033e-01 1.06512621e-01 -1.31295741e+00
-3.51149775e-02 -7.45963991e-01 -9.33591604e-01 -6.45537674e-01
7.37072304e-02 5.69268987e-02 -5.43383718e-01 2.02432871e+00
1.05238780e-01 3.76487345e-01 -4.74845409e-01 8.62552047e-01
5.28983295e-01 7.47598588e-01 2.16265004e-02 -7.18821824e-01
1.34160769e+00 -9.66038287e-01 -3.29646111e-01 -2.33782396e-01
4.79480624e-01 -6.12864912e-01 1.33638287e+00 1.41146719e-01
-1.15487587e+00 -6.72649682e-01 -7.05304086e-01 4.83703129e-02
-1.69734076e-01 1.07505862e-02 5.21835685e-01 8.09660852e-02
-9.31226075e-01 4.67326522e-01 -3.30483973e-01 -2.02113509e-01
2.63522506e-01 5.76439559e-01 1.93079621e-01 -1.81248963e-01
-1.54320824e+00 3.51249784e-01 5.22570778e-03 9.41944122e-02
-4.49109614e-01 -5.95932007e-01 -4.82764632e-01 6.30598545e-01
7.79627919e-01 -2.96176761e-01 1.19004214e+00 -9.84804630e-01
-1.06626666e+00 3.12051445e-01 -5.56466579e-01 -2.27451608e-01
3.72460246e-01 -2.34960943e-01 -6.58604443e-01 -2.28437245e-01
1.33548170e-01 1.39925882e-01 8.32067490e-01 -7.87609041e-01
-9.76750612e-01 -3.72657090e-01 -2.67181695e-02 2.40758151e-01
-7.35984981e-01 1.22970954e-01 -1.18904126e+00 -7.79026389e-01
2.68193670e-02 -7.95637131e-01 -1.60934463e-01 -3.83980483e-01
-2.63699651e-01 -5.00667155e-01 7.04793096e-01 -4.85121548e-01
2.32130790e+00 -2.20738268e+00 -7.79280737e-02 4.41600472e-01
5.25399148e-01 8.11416507e-02 -1.35136649e-01 1.24592878e-01
7.34204724e-02 1.10969603e-01 2.88291395e-01 -2.69262612e-01
-2.15080619e-01 6.67454153e-02 -4.21990395e-01 6.64131492e-02
-3.55402827e-01 9.45829034e-01 -8.88537109e-01 -4.19559360e-01
-1.52338594e-01 -3.58465221e-03 -6.34330451e-01 2.52296358e-01
-1.34467721e-01 2.31224939e-01 -1.02843595e+00 6.93463266e-01
6.26680017e-01 -8.27052891e-01 -1.31119881e-02 -2.12415472e-01
-2.37299308e-01 1.78375900e-01 -9.40385818e-01 1.13991213e+00
-4.80163425e-01 2.24303335e-01 -1.02804855e-01 -8.93363476e-01
7.05747247e-01 -2.49502826e-02 9.30882514e-01 -1.07609558e+00
3.32095027e-01 2.46907249e-01 -6.47357851e-02 -2.34696805e-01
5.82331836e-01 4.21080887e-01 -3.65233779e-01 6.24642730e-01
-4.16003823e-01 7.98362911e-01 4.01814803e-02 3.17120314e-01
1.27459586e+00 -3.36378932e-01 3.20600659e-01 -1.18209079e-01
7.01674283e-01 -1.66735098e-01 9.54128444e-01 9.72272754e-01
-3.93448353e-01 4.02660251e-01 4.10838395e-01 -3.85789007e-01
-7.78306484e-01 -6.60772502e-01 1.09806500e-01 1.63322926e+00
6.05070651e-01 -5.06403863e-01 -3.87739778e-01 -7.30902433e-01
-3.37966010e-02 6.27146602e-01 -6.20302200e-01 -5.32324374e-01
-9.10578012e-01 -7.62572348e-01 1.64659128e-01 5.11828482e-01
5.21270633e-01 -1.31406641e+00 -2.71582842e-01 3.92275602e-01
-3.56758446e-01 -7.68791616e-01 -1.20551205e+00 -1.22865051e-01
-4.65805054e-01 -7.98128784e-01 -7.28866756e-01 -3.73112321e-01
4.49340850e-01 8.72747183e-01 1.07230413e+00 3.46177667e-01
-3.99337569e-03 -1.09485239e-02 -6.09456420e-01 -8.81151482e-02
2.86964834e-01 2.50580281e-01 1.86819643e-01 1.86296582e-01
7.47866690e-01 -6.28998578e-01 -1.00835681e+00 8.03693771e-01
-4.79898542e-01 -1.25208393e-01 7.67743587e-01 7.84811258e-01
6.82021141e-01 4.11575325e-02 5.76200306e-01 -5.51371217e-01
5.30463517e-01 -1.04329467e+00 -4.30830628e-01 3.09911996e-01
-9.36551571e-01 -9.99007151e-02 9.61039901e-01 -7.60501504e-01
-7.93710172e-01 -4.16865349e-01 -2.66522557e-01 -8.26815486e-01
3.63307357e-01 5.27512014e-01 -3.11667342e-02 2.98883647e-01
2.88937062e-01 6.14008546e-01 -2.26428568e-01 -5.04597127e-01
2.03311481e-02 7.09957302e-01 2.75061905e-01 -4.01261836e-01
4.60997283e-01 2.86728889e-02 -4.54931974e-01 -4.46728766e-01
-8.86673212e-01 -8.13172042e-01 1.45588964e-02 -9.87455398e-02
5.62629640e-01 -8.33800852e-01 -1.24795473e+00 4.08178300e-01
-7.46881187e-01 -6.45089820e-02 -1.25877202e-01 4.11785781e-01
-1.63103491e-01 3.37618202e-01 -6.83547974e-01 -8.65148127e-01
-6.60071909e-01 -1.12306547e+00 8.28765333e-01 3.61446470e-01
-1.36811599e-01 -4.55245495e-01 -1.77123755e-01 2.32712716e-01
5.75197875e-01 -4.22457725e-01 9.34976995e-01 -7.39910245e-01
-7.01943994e-01 -3.23443592e-01 -5.83473027e-01 1.65908605e-01
-6.36060862e-03 -4.52315956e-01 -6.71076238e-01 -4.43902075e-01
-3.29553969e-02 -4.07493748e-02 7.24877298e-01 4.39071745e-01
1.53716195e+00 -4.02448058e-01 -5.99095702e-01 4.04596359e-01
1.28454089e+00 6.10537171e-01 5.04167199e-01 -3.53923663e-02
6.47084236e-01 1.90278351e-01 9.25763726e-01 6.30545855e-01
3.98628563e-01 8.93616199e-01 1.67359605e-01 1.97890475e-02
2.22969458e-01 -1.38465062e-01 5.49655974e-01 9.90502298e-01
-2.58306533e-01 -5.81389308e-01 -5.30088544e-01 3.39472800e-01
-2.09689641e+00 -1.15999269e+00 3.25501449e-02 2.61770344e+00
5.77133358e-01 3.40191215e-01 1.47080362e-01 -1.32055849e-01
7.27601767e-01 2.42716391e-02 -9.67847884e-01 -1.85120195e-01
1.35225236e-01 -6.57676160e-02 4.92846191e-01 1.82680234e-01
-5.00116408e-01 7.31468201e-01 5.19743347e+00 1.40161812e+00
-1.07605135e+00 3.09133589e-01 9.10180032e-01 -4.40215170e-01
-2.32513070e-01 -8.00801143e-02 -1.18860388e+00 9.33553934e-01
8.20448697e-01 -3.11570615e-01 5.65876305e-01 1.05345154e+00
4.54980791e-01 -4.73758951e-02 -9.48095083e-01 1.14933085e+00
2.85602435e-02 -9.14312243e-01 6.13071881e-02 1.48320526e-01
3.13216656e-01 1.53675331e-02 9.82480273e-02 9.02633190e-01
2.06923380e-01 -5.19729078e-01 6.31889343e-01 4.96374041e-01
8.87044013e-01 -6.86525345e-01 6.84418857e-01 4.52481240e-01
-1.77537537e+00 -5.97755313e-01 -4.44524348e-01 1.98802218e-01
1.98210388e-01 5.12734354e-01 -7.10151419e-02 5.16917348e-01
1.01861560e+00 8.32392633e-01 -4.25546676e-01 6.86423063e-01
5.92729628e-01 6.06154382e-01 -4.80180919e-01 -2.08630890e-01
2.65800148e-01 -2.34341130e-01 6.41407669e-01 7.88080812e-01
7.13169158e-01 4.84661132e-01 3.28083634e-01 5.66040993e-01
-2.10980207e-01 2.91616470e-01 -1.08946048e-01 1.71185717e-01
6.92457497e-01 1.09072959e+00 -3.96788150e-01 -4.62228149e-01
-6.99246228e-01 9.83652472e-01 4.12084162e-01 3.07235271e-01
-1.29018366e+00 -3.99155617e-01 7.56324232e-01 4.13922995e-01
4.67420638e-01 3.81544828e-01 -4.98337634e-02 -1.36977839e+00
1.65452614e-01 -7.22794056e-01 7.00308383e-01 -5.12211621e-01
-1.32357109e+00 7.70458162e-01 -5.87164797e-02 -1.54423809e+00
1.07934251e-01 9.15929489e-03 -8.62124622e-01 8.64431918e-01
-1.28128135e+00 -9.01058495e-01 -4.46282953e-01 8.93949807e-01
7.65298486e-01 -1.06023187e-02 1.89897642e-01 7.16319382e-01
-1.04005110e+00 1.26202822e+00 1.91723466e-01 -2.86533147e-01
6.71975553e-01 -6.81054056e-01 1.88125089e-01 5.94261169e-01
-4.72845852e-01 1.00131166e+00 4.49540704e-01 -7.54338980e-01
-1.23698318e+00 -1.11395109e+00 9.38923895e-01 -2.37108707e-01
5.76602876e-01 -2.57566810e-01 -1.05174685e+00 5.72305620e-01
-3.30786437e-01 -3.48112248e-02 6.34565115e-01 3.05329710e-01
-1.60622865e-01 -4.79482710e-01 -8.08682084e-01 8.59287202e-01
1.49676931e+00 -1.73386797e-01 -2.57573247e-01 1.47513762e-01
8.67373884e-01 6.67272583e-02 -7.80377269e-01 3.01522672e-01
8.77187014e-01 -8.46054375e-01 1.01990676e+00 -5.20871699e-01
1.72997430e-01 -5.91861159e-02 -2.22680449e-01 -5.57227135e-01
-7.83649981e-01 -8.80878270e-01 -5.17915130e-01 9.86563742e-01
3.90402049e-01 -5.86011231e-01 7.58002162e-01 7.49629140e-01
-8.22654292e-02 -1.11174011e+00 -6.36450529e-01 -5.46425343e-01
-2.93264836e-01 -4.18664724e-01 7.76837945e-01 6.68579459e-01
-8.66531953e-02 5.80359399e-01 -9.18871760e-01 -2.19415501e-01
4.06050116e-01 4.55403775e-01 6.16976261e-01 -1.30724621e+00
-5.82506359e-01 -6.66196704e-01 1.85460359e-01 -1.78084850e+00
-2.23882854e-01 -5.51791072e-01 -8.86273608e-02 -9.63861942e-01
5.54921091e-01 -7.66454101e-01 -9.33314383e-01 5.19444823e-01
-4.03293520e-01 3.54116671e-02 6.20659180e-02 8.28445733e-01
-9.65618432e-01 5.24912596e-01 1.23831224e+00 1.05101585e-01
-6.15634143e-01 4.32978898e-01 -8.93034577e-01 4.32479650e-01
6.32666349e-01 -4.83222634e-01 -6.68140709e-01 2.86902282e-02
2.69267321e-01 2.72450566e-01 -1.22840658e-01 -7.97432840e-01
4.06275928e-01 -3.66618574e-01 1.99269027e-01 -8.11464012e-01
2.17846900e-01 -7.37528324e-01 9.34917480e-02 4.01287287e-01
-2.72895008e-01 1.89824268e-01 -1.30770981e-01 8.86919916e-01
-2.19930448e-02 1.67471454e-01 5.85684657e-01 -1.53057752e-02
-8.10507357e-01 9.06254113e-01 -2.14802027e-01 -2.63856277e-02
1.00341141e+00 -4.04545188e-01 -2.29997739e-01 -5.29695630e-01
-7.24182785e-01 6.92786098e-01 1.89888924e-01 4.75313514e-01
4.91470575e-01 -1.38301921e+00 -3.22436690e-01 2.24947199e-01
2.18194097e-01 -3.40815187e-01 7.68105268e-01 1.18007398e+00
2.21720248e-01 5.47538221e-01 9.82054844e-02 -3.28257740e-01
-1.31128311e+00 8.28544319e-01 1.76384021e-02 -6.95363820e-01
-3.41053694e-01 7.71386087e-01 4.32017952e-01 2.00992107e-01
1.37302473e-01 -2.29524314e-01 -4.28583354e-01 7.76447877e-02
6.60844326e-01 4.16346699e-01 -2.35255495e-01 -7.58845031e-01
-2.80033886e-01 5.39750457e-01 -6.11736357e-01 2.43675426e-01
1.05246520e+00 -5.64896822e-01 1.23841710e-01 1.52758896e-01
1.40229487e+00 1.17597230e-01 -1.03409851e+00 -5.24961948e-01
-4.25867796e-01 -5.18197298e-01 4.97672483e-02 -4.60891575e-01
-1.26482105e+00 4.36765075e-01 3.04031610e-01 3.34391654e-01
1.47707808e+00 5.61018325e-02 1.47860909e+00 3.77833933e-01
5.01249135e-01 -1.03374672e+00 1.10986903e-01 6.15463793e-01
6.64532065e-01 -1.13001323e+00 -6.77165687e-02 -2.58006364e-01
-7.23560512e-01 6.49950445e-01 9.55897212e-01 1.94851160e-01
8.21132600e-01 -2.06230640e-01 -4.48012263e-01 -5.68552464e-02
-9.19031322e-01 -2.35343426e-01 3.72523487e-01 -7.71020502e-02
3.04645579e-02 -1.71485424e-01 -5.54860890e-01 1.10030234e+00
1.67206615e-01 6.38417676e-02 -1.14006922e-02 9.24280584e-01
-4.94689226e-01 -9.95694280e-01 1.49669260e-01 9.98756468e-01
-5.79781055e-01 -3.97181153e-01 9.63273123e-02 5.13099909e-01
9.53212604e-02 8.73961329e-01 5.21055870e-02 -9.05728459e-01
4.01189864e-01 -3.50245506e-01 -2.14672491e-01 -3.64484280e-01
-7.31344521e-01 4.45599407e-01 -1.64434135e-01 -1.06863618e+00
1.07802339e-01 -6.39804661e-01 -1.06791472e+00 -7.86685765e-01
-6.12903416e-01 2.81856805e-01 -1.96074709e-01 8.30315828e-01
7.05079734e-01 3.12034845e-01 1.07765090e+00 -3.56800616e-01
-6.12127841e-01 -8.68449628e-01 -5.25597155e-01 4.39924777e-01
1.74196884e-01 -7.77855873e-01 -4.32650834e-01 -3.58483195e-01] | [10.1506986618042, 5.535309791564941] |
3def96db-03fd-4f2c-b9f3-664585ba5fbe | clip-rr-improved-clip-network-for-relation | 2302.06350 | null | https://arxiv.org/abs/2302.06350v2 | https://arxiv.org/pdf/2302.06350v2.pdf | VITR: Augmenting Vision Transformers with Relation-Focused Learning for Cross-Modal Information Retrieval | Relation-focused cross-modal information retrieval focuses on retrieving information based on relations expressed in user queries, and it is particularly important in information retrieval applications and next-generation search engines. While pre-trained networks like Contrastive Language-Image Pre-training (CLIP) have achieved state-of-the-art performance in cross-modal learning tasks, the Vision Transformer (ViT) used in these networks is limited in its ability to focus on image region relations. Specifically, ViT is trained to match images with relevant descriptions at the global level, without considering the alignment between image regions and descriptions. This paper introduces VITR, a novel network that enhances ViT by extracting and reasoning about image region relations based on a Local encoder. VITR comprises two main components: (1) extending the capabilities of ViT-based cross-modal networks to extract and reason with region relations in images; and (2) aggregating the reasoned results with the global knowledge to predict the similarity scores between images and descriptions. Experiments were carried out by applying the proposed network to relation-focused cross-modal information retrieval tasks on the Flickr30K, RefCOCOg, and CLEVR datasets. The results revealed that the proposed VITR network outperformed various other state-of-the-art networks including CLIP, VSE$\infty$, and VSRN++ on both image-to-text and text-to-image cross-modal information retrieval tasks. | ['Georgina Cosma', 'Yan Gong'] | 2023-02-13 | null | null | null | null | ['cross-modal-information-retrieval'] | ['miscellaneous'] | [ 1.72261581e-01 -2.41095766e-01 -3.94638360e-01 -2.33434349e-01
-1.18647468e+00 -4.01044250e-01 9.34071422e-01 7.39063844e-02
-4.00425255e-01 3.13410014e-01 2.48070881e-01 -1.01214414e-02
-5.69271028e-01 -7.07185268e-01 -5.03069937e-01 -5.12162209e-01
2.32110620e-01 5.73374867e-01 3.20301026e-01 -4.04462099e-01
2.14505672e-01 4.66297686e-01 -1.62330329e+00 9.81583774e-01
4.89818275e-01 1.49161255e+00 2.39217982e-01 3.99284571e-01
-1.87540546e-01 1.18350279e+00 -1.70791358e-01 -5.05130410e-01
-4.43357928e-03 -3.54880780e-01 -1.07429230e+00 7.39100529e-03
6.90651536e-01 -1.18140891e-01 -6.51493847e-01 9.85469818e-01
5.24095833e-01 2.17729166e-01 7.37604558e-01 -1.21958888e+00
-1.15846133e+00 4.38074023e-01 -7.10744500e-01 2.74273664e-01
5.32356203e-01 -3.00685525e-01 1.27454042e+00 -1.03015137e+00
9.08118427e-01 1.27341294e+00 2.55717963e-01 2.39868894e-01
-7.20701218e-01 -6.39749110e-01 -7.43919536e-02 7.34423578e-01
-1.85128665e+00 -2.43342206e-01 8.06821287e-01 -2.59994209e-01
1.26631105e+00 2.57017791e-01 2.19467625e-01 7.81170785e-01
2.82273084e-01 1.05258119e+00 9.16000068e-01 -4.04824197e-01
-4.81340617e-01 2.76403636e-01 -7.89007172e-02 6.67405546e-01
-4.98747587e-01 -4.51599173e-02 -6.14544630e-01 1.19986527e-01
5.97690403e-01 1.07872084e-01 -3.65970492e-01 -3.68268400e-01
-1.36584663e+00 6.82338536e-01 1.01649320e+00 7.13289917e-01
-4.72068489e-01 3.83311436e-02 5.53586900e-01 2.88163573e-01
5.28969765e-01 3.84300500e-01 -2.23443061e-01 4.88478154e-01
-9.92730796e-01 1.84774883e-02 3.78469974e-01 1.12306118e+00
8.69963646e-01 -4.84928250e-01 -5.06658435e-01 1.13288760e+00
3.82395804e-01 6.18884325e-01 4.82050270e-01 -6.94698632e-01
6.00605726e-01 8.80695164e-01 -2.65414655e-01 -1.34293842e+00
-1.45284295e-01 -4.71270204e-01 -1.09606957e+00 -4.82132167e-01
-1.27373189e-01 4.90935445e-01 -9.99245346e-01 1.36508334e+00
7.46874809e-02 -2.18935058e-01 3.65212590e-01 1.09974146e+00
1.49370182e+00 7.05436110e-01 6.01284653e-02 -3.43115884e-03
1.53778148e+00 -1.27054191e+00 -8.38471234e-01 -1.67131081e-01
4.66903776e-01 -1.03433669e+00 8.16578865e-01 1.67177767e-02
-1.09394312e+00 -9.21087444e-01 -7.66551912e-01 -4.63865489e-01
-8.65549147e-01 2.37091064e-01 2.93612540e-01 -1.52063057e-01
-1.37723613e+00 8.41580145e-03 -1.31513849e-01 -6.93654954e-01
3.11144680e-01 1.13301806e-01 -6.09457135e-01 -5.19001245e-01
-1.50342727e+00 1.14546895e+00 5.93456686e-01 2.37747505e-01
-7.91068017e-01 -5.17239094e-01 -9.76962328e-01 2.12714419e-01
5.12267351e-01 -6.99329913e-01 7.34180033e-01 -8.26780260e-01
-7.22632945e-01 1.10677648e+00 -1.66470826e-01 -4.09662873e-01
7.91403428e-02 -8.67187753e-02 -6.91407502e-01 8.04051220e-01
3.07271957e-01 1.05651498e+00 6.90628827e-01 -1.49067044e+00
-6.37177229e-01 -3.62267017e-01 1.64029226e-01 4.94740963e-01
-2.24493608e-01 2.25027695e-01 -1.24635243e+00 -3.77678007e-01
9.19306874e-02 -8.73781145e-01 3.29169929e-01 -4.89305332e-02
-3.32956910e-01 -6.40132129e-01 1.16862845e+00 -7.13965654e-01
1.01358545e+00 -2.13042808e+00 1.28067091e-01 1.64786443e-01
5.24092391e-02 4.33453262e-01 -4.52435762e-01 7.08876193e-01
-1.10970922e-01 -6.44542947e-02 1.21050261e-01 -1.95517480e-01
-4.77053300e-02 2.22621426e-01 -3.27669919e-01 2.21933216e-01
1.26558363e-01 1.28443897e+00 -8.76350880e-01 -1.22484672e+00
3.86011362e-01 7.11477757e-01 -1.00084722e-01 2.32445806e-01
-1.83744565e-01 3.57995965e-02 -4.11056936e-01 9.03040469e-01
5.15249610e-01 -4.78053927e-01 -4.12986837e-02 -8.76918018e-01
1.11268923e-01 -1.08564846e-01 -7.23556519e-01 1.87127972e+00
-7.18286395e-01 7.35269606e-01 -2.11104438e-01 -1.05398500e+00
7.15808153e-01 5.36050081e-01 7.39163160e-01 -1.36888158e+00
6.28925264e-02 7.44948536e-02 -4.94941413e-01 -7.41890311e-01
6.71157956e-01 2.04697311e-01 -1.27377421e-01 2.75254905e-01
2.76048630e-01 -9.79262441e-02 2.83651292e-01 5.70992529e-01
7.69433975e-01 2.87322372e-01 1.59770265e-01 3.13575342e-02
1.06701410e+00 1.06582768e-01 2.11562589e-02 6.75413251e-01
-2.40365863e-02 6.99983656e-01 4.40741703e-02 -2.48231843e-01
-7.46175170e-01 -1.03522372e+00 -1.35493875e-01 1.05306137e+00
6.49214208e-01 -3.83081228e-01 -2.38236621e-01 -7.65709639e-01
-1.87810451e-01 6.11606002e-01 -7.85823166e-01 -1.51083350e-01
-2.10599408e-01 -1.51204482e-01 5.89220226e-01 3.46574783e-01
1.18234873e+00 -1.36261928e+00 -6.92943558e-02 -2.09770918e-01
-7.96782494e-01 -1.56353295e+00 -7.36877441e-01 -1.24147661e-01
-2.99157530e-01 -1.20916438e+00 -7.91322231e-01 -1.17417872e+00
5.66778481e-01 5.99124432e-01 1.27981067e+00 1.82623863e-01
-3.33137780e-01 8.73183906e-01 -6.68753684e-01 -1.29896745e-01
-1.10056147e-01 1.63788915e-01 -4.51097757e-01 2.06326589e-01
3.90942633e-01 -2.89448239e-02 -7.75170445e-01 5.71567476e-01
-1.30268860e+00 2.43320558e-02 8.99091661e-01 9.40701842e-01
8.63085330e-01 3.11468095e-01 3.20451707e-01 -6.65364981e-01
4.68414605e-01 -4.27539527e-01 -1.75851658e-01 1.01079452e+00
-6.90470695e-01 1.03116162e-01 2.30759472e-01 -2.36085385e-01
-1.37523818e+00 -8.47170427e-02 6.05556220e-02 -8.03898931e-01
1.86082311e-02 7.44523048e-01 -5.19798808e-02 -1.55108318e-01
4.48793381e-01 4.84459192e-01 -1.78405404e-01 -1.03124768e-01
5.92121363e-01 6.96637809e-01 7.37612069e-01 -4.31130022e-01
6.99125946e-01 4.86209542e-01 2.58505139e-02 -5.92374980e-01
-1.08852243e+00 -1.11418462e+00 -6.79803431e-01 -3.79556209e-01
1.20906162e+00 -1.13179648e+00 -5.74498653e-01 2.00084403e-01
-1.03912556e+00 1.43598452e-01 -5.89427315e-02 3.20830822e-01
-4.36638653e-01 3.40077728e-01 -5.27812243e-01 -4.16239679e-01
-6.54452562e-01 -1.13328195e+00 1.37878299e+00 3.05311054e-01
3.50731403e-01 -1.14136934e+00 -1.64419413e-01 9.45848107e-01
4.90670294e-01 -1.34494543e-01 8.25287580e-01 -6.83067501e-01
-6.47476196e-01 -2.62569666e-01 -8.91167045e-01 2.30100051e-01
2.08123147e-01 -2.93819129e-01 -9.04829800e-01 -1.05761051e-01
-3.67302775e-01 -7.13923156e-01 1.04731965e+00 5.35813756e-02
1.10014915e+00 -2.21189380e-01 -4.30460483e-01 3.54957521e-01
1.62978399e+00 1.04413301e-01 8.51378798e-01 3.74597162e-01
8.57714295e-01 6.43844485e-01 8.50863695e-01 -8.08457732e-02
6.16058052e-01 9.93001938e-01 5.56385875e-01 -4.54752505e-01
-5.45084238e-01 -2.76898921e-01 -1.12638231e-02 6.79364324e-01
1.30834356e-01 -4.33741719e-01 -7.02129304e-01 9.29466307e-01
-2.16930771e+00 -1.12694836e+00 9.44743231e-02 1.74788165e+00
8.23244870e-01 -4.12773669e-01 -4.26356465e-01 -3.26896966e-01
5.38603365e-01 2.42614254e-01 -4.57784325e-01 -9.01751369e-02
-4.66952980e-01 1.10500880e-01 2.75759012e-01 2.96193123e-01
-1.21403182e+00 1.13265479e+00 4.77623272e+00 1.19347954e+00
-9.20290351e-01 1.41381294e-01 4.63167191e-01 2.06438318e-01
-2.26732165e-01 -4.59879190e-02 -6.17999852e-01 -1.42333776e-01
4.35684294e-01 2.05376714e-01 2.38250688e-01 5.87591290e-01
-2.94516116e-01 -2.02998415e-01 -1.03026867e+00 1.30993259e+00
6.91123486e-01 -1.36795521e+00 5.23592710e-01 -1.53153896e-01
8.22785437e-01 1.27862766e-01 1.71709418e-01 4.36857671e-01
-1.67031516e-03 -9.44699049e-01 4.15607899e-01 8.10447514e-01
7.75596797e-01 -6.66392088e-01 1.35509300e+00 1.20063886e-01
-1.51264894e+00 1.70238778e-01 -2.61425525e-01 6.47724569e-01
2.94261537e-02 2.61299282e-01 -9.08427775e-01 1.15880179e+00
9.72029090e-01 9.94261026e-01 -9.35608506e-01 7.33285785e-01
-5.54584712e-02 -2.58725602e-02 -2.33701877e-02 2.44985208e-01
5.09906232e-01 5.37636550e-03 1.57065794e-01 1.42571652e+00
1.77249864e-01 -2.68152636e-02 9.45710242e-02 8.32522869e-01
-1.99841246e-01 1.62357390e-01 -5.75448036e-01 1.10395759e-01
1.57002866e-01 1.51388764e+00 -5.22644877e-01 -4.59002137e-01
-6.20793223e-01 1.10510528e+00 1.36102930e-01 5.39043307e-01
-6.86054945e-01 -4.37217146e-01 2.63027519e-01 -9.83873457e-02
5.48057914e-01 1.99653342e-01 2.05381468e-01 -1.08209467e+00
-1.07895233e-01 -8.08478355e-01 8.29926550e-01 -1.39556265e+00
-1.53039145e+00 8.96882057e-01 3.97777975e-01 -1.26820695e+00
-3.45237106e-01 -3.46436143e-01 5.95619753e-02 1.02320397e+00
-2.10160565e+00 -1.86458480e+00 -4.68664259e-01 1.21541524e+00
5.06238461e-01 -2.06424966e-01 6.25857711e-01 4.95807558e-01
-2.52489269e-01 6.20774686e-01 -2.25466520e-01 3.81249040e-01
9.90085721e-01 -8.67398620e-01 -2.70805478e-01 6.93688035e-01
4.97485816e-01 5.82132041e-01 2.45058432e-01 -3.96931618e-01
-1.37056971e+00 -1.18885803e+00 1.08303881e+00 -1.71190903e-01
5.94311953e-01 1.13558389e-01 -7.60112703e-01 4.84877020e-01
8.52278352e-01 3.10256481e-01 3.59100044e-01 -2.13075489e-01
-6.84071422e-01 -3.77732366e-01 -1.01879346e+00 4.67143148e-01
8.67075264e-01 -1.12046850e+00 -5.06519318e-01 5.68967044e-01
6.67691290e-01 -4.10718530e-01 -1.24058604e+00 6.46621525e-01
5.12690961e-01 -7.90582895e-01 1.37479901e+00 -2.44811937e-01
5.40828586e-01 -3.71117353e-01 -5.48570395e-01 -9.51480150e-01
-3.90644431e-01 -2.24219020e-02 1.64470419e-01 1.46329606e+00
3.78152281e-01 -2.09936008e-01 1.91642940e-01 3.02996010e-01
-8.81378651e-02 -7.33445406e-01 -6.81356549e-01 -4.58023638e-01
-2.55422473e-01 -2.99471945e-01 2.50181258e-01 8.02793443e-01
-2.05946550e-01 5.38114846e-01 -3.03080082e-01 2.51528949e-01
5.38281381e-01 3.70873421e-01 5.14960229e-01 -7.48462081e-01
-2.49357559e-02 -4.10609663e-01 -3.16459239e-01 -9.64763224e-01
3.51756662e-01 -1.04577911e+00 2.11257283e-02 -1.96754026e+00
6.79501891e-01 -3.11275810e-01 -5.73687911e-01 7.11076200e-01
-5.96767217e-02 7.75912523e-01 4.19309556e-01 5.08408427e-01
-1.28444755e+00 4.12003964e-01 1.49449146e+00 -6.41491413e-01
1.23199463e-01 -3.10013235e-01 -4.76717412e-01 3.10976923e-01
4.65355903e-01 -1.49143338e-01 -6.92614377e-01 -3.80457312e-01
4.11798477e-01 3.15592200e-01 5.54641485e-01 -8.49785686e-01
5.96239209e-01 1.11713201e-01 3.84781867e-01 -1.06456506e+00
4.79509979e-01 -1.08577919e+00 1.25905737e-01 4.88886163e-02
-6.91007555e-01 1.73815906e-01 1.45658001e-01 4.45263445e-01
-8.88796031e-01 4.12128586e-03 4.98551458e-01 -3.40002060e-01
-1.16277862e+00 3.36172462e-01 2.82582529e-02 9.67091396e-02
8.58759761e-01 -6.40578344e-02 -6.45940900e-01 -6.39087260e-01
-4.00674731e-01 4.74696815e-01 1.18577272e-01 7.17073798e-01
1.04430485e+00 -1.44851565e+00 -6.33076072e-01 -1.14746481e-01
8.29750538e-01 -3.78055200e-02 5.77637434e-01 7.83702016e-01
-2.88070977e-01 1.02993500e+00 -6.67389482e-02 -8.15180898e-01
-1.53952372e+00 8.89202535e-01 3.30519855e-01 -7.19641864e-01
-4.25353199e-01 7.89088607e-01 3.16303581e-01 -3.80205274e-01
2.53886044e-01 6.15614932e-03 -5.02376556e-01 2.23121747e-01
3.67442429e-01 -2.17237890e-01 2.31314242e-01 -1.11826348e+00
-5.22935748e-01 9.38662767e-01 -3.52236956e-01 -2.98557673e-02
9.87641931e-01 -4.43101525e-01 -4.24169928e-01 1.72017679e-01
1.69363606e+00 -5.19423008e-01 -6.74864233e-01 -9.21246171e-01
-2.76520282e-01 -2.16385692e-01 3.59547734e-01 -1.05216336e+00
-1.42022598e+00 7.14334369e-01 6.37968838e-01 -3.16648185e-02
1.55905545e+00 4.51042473e-01 9.30698097e-01 4.33870912e-01
3.41432005e-01 -9.67899024e-01 4.67447400e-01 3.34877700e-01
1.04418325e+00 -1.48490822e+00 2.87740678e-01 -1.71501458e-01
-8.34873736e-01 1.01111758e+00 6.07171893e-01 1.88785374e-01
8.85006666e-01 -3.73053104e-01 1.48085982e-01 -6.50070488e-01
-7.59964406e-01 -6.16888881e-01 1.07469463e+00 5.09132564e-01
3.16503376e-01 -3.69215012e-01 -5.71527034e-02 -9.21911895e-02
2.05606297e-01 -5.92980795e-02 -1.59023389e-01 7.48404145e-01
5.40123172e-02 -9.32058096e-01 -3.18949908e-01 2.50926316e-01
-4.45722878e-01 -3.81727666e-01 -5.12305439e-01 9.61703360e-01
2.38352612e-01 1.23202682e+00 1.04405530e-01 -4.56672609e-01
2.13602334e-01 -1.95951164e-01 3.77940595e-01 -2.12177962e-01
-7.18678892e-01 4.54164995e-03 -2.38035396e-02 -6.98279858e-01
-1.10885787e+00 -3.36721212e-01 -9.45764661e-01 1.38728172e-01
-4.34634775e-01 9.46806595e-02 5.68099678e-01 1.13875318e+00
4.69179899e-01 6.64556503e-01 4.68334705e-01 -5.92778444e-01
1.23133108e-01 -1.04319048e+00 -3.02474409e-01 7.44330585e-01
2.61552423e-01 -5.64601362e-01 -1.73914954e-01 2.35795286e-02] | [10.850663185119629, 1.229317545890808] |
057ef931-bd32-4887-881c-a4e4b5f15599 | semeval-2010-task-13-tempeval-2 | null | null | https://aclanthology.org/S10-1010/ | https://aclanthology.org/S10-1010.pdf | SemEval-2010 Task 13: TempEval-2 | Tempeval-2 comprises evaluation tasks for time expressions, events and temporal relations, the latter of which was split up in four sub tasks, motivated by the notion that smaller subtasks would make both data preparation and temporal relation extraction easier. Manually annotated data were provided for six languages: Chinese, English, French, Italian, Korean and Spanish. | ['James Pustejovsky', 'Tommaso Caselli', 'Roser Saurí', 'Marc Verhagen'] | 2010-07-01 | null | null | null | proceedings-of-the-5th-international-workshop | ['temporal-relation-extraction', 'temporal-relation-classification', 'temporal-tagging'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-2.53712952e-01 3.01456600e-01 -3.29232514e-01 -5.53961396e-01
-4.27457958e-01 -1.05371058e+00 1.09745324e+00 5.60661435e-01
-7.13672578e-01 1.08144188e+00 7.58508921e-01 -3.46252829e-01
-2.25106135e-01 -4.15926099e-01 1.86266692e-03 -2.13434864e-02
-8.17337990e-01 2.96964347e-01 3.34952831e-01 -1.38961270e-01
1.65748671e-01 4.72863585e-01 -1.08037961e+00 3.59646171e-01
5.09596705e-01 9.26166892e-01 -6.07506514e-01 6.86211586e-01
2.27176204e-01 1.45754743e+00 -3.71534467e-01 -2.97442913e-01
-1.93944633e-01 -2.34461039e-01 -1.32976663e+00 -3.52340162e-01
-6.46202087e-01 2.55391836e-01 -3.93973380e-01 3.07185292e-01
2.36200035e-01 6.21761978e-01 5.97703397e-01 -1.45317376e+00
-3.34901452e-01 9.27374840e-01 -3.10790598e-01 3.97280097e-01
8.21443796e-01 5.30178435e-02 1.19522071e+00 -3.96958262e-01
1.18914211e+00 1.16617525e+00 5.98112762e-01 2.75973856e-01
-1.25023711e+00 -4.39837724e-01 1.88294590e-01 1.86001226e-01
-1.35636640e+00 -4.30065006e-01 5.00233412e-01 -4.88709092e-01
1.79458344e+00 8.30266401e-02 5.49346507e-01 1.37587488e+00
3.59412342e-01 6.96756780e-01 1.08573616e+00 -3.61330926e-01
2.30139315e-01 -4.31856252e-02 4.65373307e-01 1.35731131e-01
-2.83430815e-01 4.35014576e-01 -6.62718534e-01 -6.71327785e-02
4.89700317e-01 -7.33983636e-01 -3.73816378e-02 1.84037045e-01
-1.47702622e+00 4.94878411e-01 -1.36678085e-01 7.29075849e-01
-4.98537570e-01 -4.10331525e-02 1.10637426e+00 5.30852795e-01
5.66616356e-01 4.01711881e-01 -8.70016098e-01 -7.70529568e-01
-4.26726878e-01 5.33831298e-01 8.83490205e-01 9.09251869e-01
2.02084765e-01 -1.85771793e-01 -2.09502310e-01 3.92129004e-01
1.30302280e-01 -9.48301777e-02 3.75728279e-01 -8.04763556e-01
7.77181149e-01 5.26570499e-01 4.57334816e-01 -6.31149530e-01
-9.11282420e-01 3.76884311e-01 -4.76581693e-01 -4.34952468e-01
4.17087257e-01 -5.10632277e-01 -8.45448673e-01 2.20240116e+00
2.57023245e-01 -2.63231900e-02 4.12557185e-01 3.92512262e-01
8.38720977e-01 7.44891942e-01 5.20192981e-01 -8.53441477e-01
1.42020583e+00 -6.99619353e-01 -1.20979011e+00 -6.65410534e-02
8.63227427e-01 -5.17066777e-01 5.60199857e-01 2.19021291e-01
-1.23285413e+00 -4.78182614e-01 -7.62552500e-01 -2.83574730e-01
-7.20452666e-01 -3.51430476e-01 1.00934339e+00 2.28880420e-02
-8.58432531e-01 4.00646389e-01 -1.27506280e+00 -4.96885270e-01
-4.19642553e-02 3.41122270e-01 -4.97457802e-01 7.22944140e-01
-1.69268286e+00 1.00556147e+00 9.56645906e-01 9.94486585e-02
-5.10379016e-01 -2.51837403e-01 -1.28354442e+00 -5.10966122e-01
2.33683243e-01 -1.47623658e-01 1.61314511e+00 -6.33594334e-01
-1.49123526e+00 1.12175870e+00 -1.88752741e-01 -7.03522146e-01
3.26529503e-01 -9.89355892e-02 -1.14670527e+00 -1.79840088e-01
3.09964329e-01 3.79025191e-01 1.25220940e-01 -1.40120536e-01
-6.63882613e-01 -4.66719091e-01 1.42783940e-01 8.00143108e-02
1.75161287e-01 8.94060254e-01 -4.17669058e-01 -7.53539383e-01
-2.96989411e-01 -8.56032550e-01 -3.26230288e-01 -8.14843595e-01
-2.45743260e-01 -9.27594244e-01 5.27006388e-01 -8.42941701e-01
1.79728365e+00 -2.17611361e+00 1.99744463e-01 1.26815304e-01
-1.28149733e-01 -5.30285478e-01 6.41472219e-03 7.26500869e-01
-7.62250841e-01 4.49516550e-02 1.24742694e-01 -7.89107829e-02
1.57059222e-01 2.91877598e-01 -3.35199803e-01 4.07107472e-01
3.89648974e-01 1.06449533e+00 -9.28517282e-01 -6.86840296e-01
1.15337387e-01 3.50361392e-02 -1.55499414e-01 7.66519681e-02
-4.76535738e-01 5.90149879e-01 -4.42646354e-01 6.02572501e-01
-1.61523685e-01 1.49988994e-01 2.29303062e-01 1.53015941e-01
-6.66703999e-01 1.20389640e+00 -1.14606893e+00 1.59286416e+00
-2.26573542e-01 6.83005989e-01 -3.05797249e-01 -5.59525371e-01
5.01416683e-01 8.87645841e-01 1.00407135e+00 -9.72700298e-01
2.59208530e-01 -6.78820461e-02 -1.91030875e-01 -7.48898923e-01
6.18065715e-01 -8.42070431e-02 -7.87733376e-01 6.11008108e-01
-6.18794151e-02 -9.86642018e-02 1.00176907e+00 4.09132317e-02
1.01405036e+00 6.19072914e-01 9.10349429e-01 -8.70515928e-02
3.36153179e-01 9.47212726e-02 9.23469245e-01 -7.01861307e-02
-5.39600849e-01 -1.08452328e-01 8.33561540e-01 -8.68582845e-01
-6.96151733e-01 -9.97060955e-01 -2.67163385e-02 1.22763276e+00
-2.53817856e-01 -1.17760849e+00 -1.67385936e-01 -6.78419590e-01
-4.65444416e-01 8.98515821e-01 -6.98849559e-01 1.81226596e-01
-8.57142568e-01 -5.50411105e-01 7.90218711e-01 8.80617082e-01
4.00914066e-02 -1.41416442e+00 -9.03995454e-01 4.70287919e-01
-7.27899790e-01 -1.52533472e+00 -5.29375851e-01 4.69171345e-01
-8.39212120e-01 -1.07333648e+00 9.01162326e-02 -4.67655331e-01
-8.39668605e-03 -5.58632731e-01 1.35527992e+00 -5.09853005e-01
-8.01212937e-02 2.86354035e-01 -4.90012735e-01 -7.81381965e-01
5.58440981e-04 -2.73300931e-02 1.14003837e-01 -5.92740655e-01
6.02165759e-01 -6.12008870e-01 -1.03809349e-01 2.35824108e-01
-7.17002749e-01 -2.77920417e-03 -8.26602578e-02 3.26793283e-01
4.06810343e-01 1.79632291e-01 1.61183521e-01 -4.28924710e-01
6.20126784e-01 -5.11707842e-01 -4.58420008e-01 2.06826270e-01
-3.24829936e-01 -7.04241991e-02 2.60150909e-01 -5.02177358e-01
-1.03667736e+00 8.85762274e-03 1.41182423e-01 4.44823354e-01
-2.88959652e-01 8.77338469e-01 1.06489301e-01 6.98143303e-01
7.36056566e-01 -1.37316227e-01 -6.72068834e-01 -1.06669106e-01
3.90214771e-01 3.46508265e-01 5.29852331e-01 -9.90248144e-01
3.99953544e-01 7.84735829e-02 -8.75914097e-02 -8.85891795e-01
-6.79474413e-01 -3.33980531e-01 -9.83833849e-01 -3.57613623e-01
1.23852372e+00 -8.09117317e-01 -5.54869413e-01 3.15176427e-01
-1.30348766e+00 -6.45700097e-01 -6.05544686e-01 8.07802320e-01
-6.28151119e-01 5.20632379e-02 -9.22016203e-01 -8.66998374e-01
-2.05233037e-01 -5.54997623e-01 8.27451169e-01 1.34496704e-01
-9.97564852e-01 -1.17414665e+00 3.08115244e-01 -2.32460856e-01
-5.60010448e-02 7.69309759e-01 8.49805057e-01 -7.53729939e-01
1.03090689e-01 -3.33858311e-01 -8.98486525e-02 -1.78893209e-01
-6.01975359e-02 3.32461894e-01 -5.47850609e-01 6.34995922e-02
-9.67693925e-02 -7.90241599e-01 2.56589741e-01 1.97638065e-01
6.06960237e-01 -2.61869192e-01 -3.47380191e-01 1.55351400e-01
9.56281722e-01 7.58016884e-01 6.48615241e-01 2.28999346e-01
1.97155014e-01 7.16967642e-01 8.93085837e-01 5.85168242e-01
8.47257912e-01 7.78263569e-01 -1.22593075e-01 2.65884161e-01
2.42450505e-01 -3.80244963e-02 5.29520810e-01 7.78268814e-01
-7.16140866e-01 -2.25543380e-01 -1.05845881e+00 9.23009753e-01
-1.93406761e+00 -1.06976020e+00 -3.24445039e-01 1.88115728e+00
1.09376323e+00 3.79213661e-01 5.67014873e-01 2.84166425e-01
7.85781443e-02 4.52985615e-01 5.25202192e-02 -7.76130259e-01
-1.05036169e-01 4.34238911e-01 1.35854676e-01 3.54460806e-01
-1.36110640e+00 9.79918361e-01 7.30962896e+00 2.22694382e-01
-1.01444316e+00 -1.06477708e-01 5.37620783e-01 3.72262076e-02
-7.46157542e-02 2.89250493e-01 -4.21193361e-01 1.08521394e-01
1.55716527e+00 -5.50372601e-01 3.58892471e-01 2.84383476e-01
5.49254298e-01 -1.56337485e-01 -1.52634156e+00 6.28727257e-01
-5.86772859e-01 -8.31918538e-01 -5.10002375e-01 -3.60097438e-01
5.21175086e-01 6.44355044e-02 -5.83922684e-01 5.66289604e-01
6.25878453e-01 -8.46109927e-01 1.13633251e+00 5.30944169e-01
4.75920975e-01 -4.33775544e-01 4.65474516e-01 1.21328302e-01
-1.66212106e+00 1.33791879e-01 4.90046114e-01 -4.33493376e-01
6.86080396e-01 2.23249659e-01 -4.36071366e-01 8.58513415e-01
7.94578671e-01 8.89421999e-01 -3.78522843e-01 4.05344427e-01
-5.29024243e-01 6.52019441e-01 -3.80913824e-01 1.91284195e-01
3.87526780e-01 -1.53283745e-01 5.08574486e-01 1.44450510e+00
-8.20127577e-02 3.42859447e-01 2.86131591e-01 5.56709826e-01
4.20126051e-01 1.21754877e-01 -6.83244586e-01 -6.30356073e-01
4.12604600e-01 9.20975387e-01 -9.43704605e-01 -1.31031498e-01
-4.62385029e-01 5.55798709e-01 2.64393806e-01 3.90665531e-01
-1.18426871e+00 -4.72133040e-01 3.02175224e-01 -1.15279533e-01
-1.55358687e-01 -8.12335134e-01 -1.42619655e-01 -1.12909102e+00
9.82076023e-03 -7.33565748e-01 1.21553874e+00 -8.26411724e-01
-9.48678017e-01 7.27820039e-01 4.98822927e-01 -8.48629594e-01
-8.92457545e-01 -4.82360721e-01 -5.58953166e-01 7.51068592e-01
-7.67696857e-01 -1.10521317e+00 6.52292818e-02 6.51776195e-01
4.19464797e-01 3.48604679e-01 9.92271543e-01 3.78619850e-01
-7.26538002e-01 1.83354765e-01 -9.11243796e-01 1.97599202e-01
8.31975162e-01 -1.33526480e+00 6.41616166e-01 9.34351325e-01
3.99480127e-02 5.92735767e-01 6.72930121e-01 -7.52394736e-01
-1.05592620e+00 -8.97456348e-01 1.94232440e+00 -5.66660702e-01
1.23847067e+00 -2.63779372e-01 -3.51461709e-01 1.26461923e+00
4.96294260e-01 -6.54785037e-02 5.24471223e-01 3.76421660e-01
-4.16259915e-01 1.86145484e-01 -6.87619627e-01 6.43402934e-01
1.11286497e+00 -7.62403727e-01 -9.68397737e-01 2.57509202e-01
8.25699508e-01 -7.72527695e-01 -1.38959992e+00 4.96181965e-01
4.01221633e-01 -4.62147146e-01 7.98143923e-01 -1.09057772e+00
4.46518689e-01 -2.24953175e-01 -1.22072123e-01 -8.90855908e-01
-1.76761299e-01 -1.03358841e+00 -3.80235791e-01 1.43565369e+00
7.69012034e-01 -5.38786232e-01 2.67606825e-01 9.31362748e-01
-1.69460073e-01 -2.45604336e-01 -9.38245296e-01 -9.42678094e-01
-2.02881798e-01 -1.10198152e+00 4.99557287e-01 1.19371104e+00
6.75388813e-01 7.90565908e-01 -1.15456551e-01 -2.96933651e-01
3.44223902e-02 6.89962283e-02 4.32816029e-01 -1.33272541e+00
4.75250892e-02 -5.78081131e-01 -1.71413586e-01 -4.96011019e-01
3.21659833e-01 -6.04668319e-01 -2.20287248e-01 -1.44458508e+00
-2.39776820e-01 -2.74554133e-01 3.33006307e-02 9.89302635e-01
1.23996988e-01 -6.85990527e-02 -2.66781777e-01 -6.19794428e-02
-9.79883850e-01 3.80852133e-01 7.74596512e-01 1.29940167e-01
-5.92835903e-01 2.77646296e-02 -4.94693607e-01 6.15309477e-01
5.73896885e-01 -2.41701052e-01 -5.24388194e-01 -3.18727463e-01
5.73363125e-01 5.98602653e-01 -3.37893516e-02 -7.90976167e-01
2.72051364e-01 -6.78562343e-01 -6.50660768e-02 -6.95471644e-01
1.26983061e-01 -6.76219523e-01 4.48394030e-01 1.78405344e-01
-4.91599768e-01 6.20764256e-01 6.50754154e-01 2.23543808e-01
-6.76434696e-01 4.10080910e-01 4.40847129e-01 -2.07524076e-02
-8.35030675e-01 1.43509060e-01 -5.15122771e-01 3.74192774e-01
1.38411796e+00 -1.28383711e-01 -2.69585788e-01 -2.92252213e-01
-1.07626212e+00 4.55205679e-01 -2.73541212e-01 7.39939749e-01
5.70367761e-02 -1.50047541e+00 -5.75373352e-01 -2.79120713e-01
2.71377742e-01 -2.21759826e-02 -9.88869462e-03 1.37494719e+00
-1.24050707e-01 7.77046800e-01 -1.15193948e-01 -1.33079872e-01
-1.22073138e+00 1.01217997e+00 2.87724547e-02 -7.60899365e-01
-4.76112515e-01 6.04832053e-01 -2.87043184e-01 -2.79523760e-01
4.74451214e-01 -7.92049706e-01 -3.80556792e-01 4.22934264e-01
4.23273325e-01 4.78391826e-01 4.48112078e-02 -5.86583316e-01
-6.80947304e-01 1.40382886e-01 5.22869714e-02 -7.03562796e-01
1.49902105e+00 -1.65224329e-01 -7.47649491e-01 8.95157158e-01
8.87100279e-01 -7.32131451e-02 -5.01568317e-01 -2.36324504e-01
9.57956672e-01 4.03502405e-01 -3.08405370e-01 -8.12538624e-01
-2.73272425e-01 1.43742904e-01 -5.33427252e-03 5.29085100e-01
1.35584915e+00 6.83583170e-02 6.94308102e-01 3.63090709e-02
5.54641306e-01 -1.46955502e+00 -3.00922424e-01 1.06693947e+00
8.95393848e-01 -8.03582847e-01 -1.50913581e-01 -2.01674208e-01
-6.62029564e-01 1.10731590e+00 6.57854438e-01 4.40620095e-01
6.64459884e-01 4.64606553e-01 -2.54636258e-01 -4.50302213e-01
-1.22371376e+00 -2.12298810e-01 4.72598374e-01 3.02500933e-01
1.15827632e+00 2.20150799e-01 -9.22481060e-01 1.12113130e+00
-1.85648739e-01 3.83848011e-01 2.13169113e-01 1.37543309e+00
3.33928645e-01 -1.21518075e+00 -1.11044586e-01 1.21142417e-01
-7.36327827e-01 7.21529722e-02 -7.20804334e-01 1.24828231e+00
1.41303331e-01 1.05971944e+00 2.03344509e-01 -4.66348141e-01
6.43229544e-01 1.13255866e-01 4.18311685e-01 -3.37194294e-01
-8.07201266e-01 6.95205256e-02 1.01420975e+00 -9.85241115e-01
-8.03114891e-01 -9.95746434e-01 -1.54443848e+00 -6.82577714e-02
2.47926638e-01 4.47542608e-01 1.68522328e-01 1.11988401e+00
5.72662354e-02 6.63673043e-01 4.94449079e-01 -4.63667363e-01
1.44009039e-01 -1.06741560e+00 -1.82880104e-01 4.05111164e-01
-7.91008323e-02 -3.46694529e-01 -2.04045445e-01 3.01171213e-01] | [9.100245475769043, 9.238439559936523] |
15ac2b7b-f056-4e60-bdba-3352a405c087 | multilegalsbd-a-multilingual-legal-sentence | 2305.01211 | null | https://arxiv.org/abs/2305.01211v1 | https://arxiv.org/pdf/2305.01211v1.pdf | MultiLegalSBD: A Multilingual Legal Sentence Boundary Detection Dataset | Sentence Boundary Detection (SBD) is one of the foundational building blocks of Natural Language Processing (NLP), with incorrectly split sentences heavily influencing the output quality of downstream tasks. It is a challenging task for algorithms, especially in the legal domain, considering the complex and different sentence structures used. In this work, we curated a diverse multilingual legal dataset consisting of over 130'000 annotated sentences in 6 languages. Our experimental results indicate that the performance of existing SBD models is subpar on multilingual legal data. We trained and tested monolingual and multilingual models based on CRF, BiLSTM-CRF, and transformers, demonstrating state-of-the-art performance. We also show that our multilingual models outperform all baselines in the zero-shot setting on a Portuguese test set. To encourage further research and development by the community, we have made our dataset, models, and code publicly available. | ['Joel Niklaus', 'Matthias Stürmer', 'Tobias Brugger'] | 2023-05-02 | null | null | null | null | ['boundary-detection'] | ['computer-vision'] | [ 2.86624655e-02 9.62627605e-02 -2.01068506e-01 -5.02502501e-01
-1.43269730e+00 -8.47078025e-01 4.00182307e-01 1.94389880e-01
-6.31197691e-01 1.01070774e+00 5.78043520e-01 -9.09397066e-01
4.96021777e-01 -3.89160246e-01 -6.25229776e-01 -4.10354286e-02
1.09246708e-01 5.38597584e-01 4.64921921e-01 -5.46953022e-01
3.21539938e-01 1.80741847e-01 -9.34046805e-01 7.26169527e-01
1.27836955e+00 2.83076614e-01 1.03215307e-01 5.56591928e-01
-3.81639570e-01 8.22993994e-01 -7.99625516e-01 -1.05559671e+00
5.61891943e-02 -2.69028068e-01 -1.00445116e+00 -5.00291169e-01
6.11133873e-01 9.40478146e-02 -6.60956800e-02 1.11204624e+00
5.48114896e-01 -2.47126043e-01 3.78835797e-01 -6.37744427e-01
-6.72794700e-01 1.03118527e+00 -3.14782739e-01 8.09511065e-01
5.11406958e-01 2.14046404e-01 1.28953779e+00 -9.51700926e-01
1.13315010e+00 1.40918756e+00 7.68375814e-01 4.56271321e-01
-1.29553688e+00 -4.18104440e-01 2.13782609e-01 2.80739576e-01
-1.09966075e+00 -8.26774657e-01 3.75398070e-01 -4.63074207e-01
1.64173746e+00 -1.96915686e-01 7.54585117e-02 1.08966744e+00
5.09379923e-01 8.25145602e-01 1.10817719e+00 -6.80016279e-01
-1.66527808e-01 -1.88440859e-01 4.41601515e-01 5.89131415e-01
1.31737962e-01 -1.18185930e-01 -4.84711319e-01 -8.59495550e-02
3.23860236e-02 -6.99585617e-01 -1.76284939e-01 3.94393116e-01
-9.72962379e-01 8.18067968e-01 -1.81591690e-01 7.16676533e-01
-7.74548501e-02 -3.08861762e-01 6.22626066e-01 5.45787871e-01
7.23908961e-01 2.75245547e-01 -7.51529455e-01 -5.67367256e-01
-9.17286932e-01 1.90487996e-01 8.77467752e-01 9.49416697e-01
4.44534302e-01 -2.88034350e-01 -4.55889463e-01 9.82184887e-01
1.66445836e-01 5.96534908e-01 2.76341230e-01 -7.51256108e-01
1.25407898e+00 4.47861046e-01 5.15150018e-02 -6.06157899e-01
-3.86289477e-01 -4.95425425e-02 -2.88648307e-01 -1.80867836e-01
8.03612351e-01 -2.54937470e-01 -7.84232259e-01 1.57258987e+00
1.18185870e-01 -3.77919704e-01 2.33129650e-01 4.58584696e-01
7.20358968e-01 6.07223451e-01 2.54517049e-01 -1.47520632e-01
1.22804403e+00 -1.04867244e+00 -7.80598521e-01 -4.96443450e-01
9.19902802e-01 -1.31950247e+00 1.16990864e+00 2.56565958e-01
-1.18627775e+00 -4.58783329e-01 -8.02237451e-01 -6.63898706e-01
-2.65913218e-01 1.37067825e-01 2.89174527e-01 5.34979165e-01
-9.20859754e-01 6.63409650e-01 -7.42060184e-01 -4.65010732e-01
2.98233896e-01 -1.29506186e-01 -2.71816730e-01 -3.43028337e-01
-1.42233264e+00 1.20901644e+00 1.59556881e-01 2.36309931e-01
-5.43241084e-01 -5.62158406e-01 -1.01040626e+00 -1.89000353e-01
1.52873591e-01 -3.36668231e-02 1.78969359e+00 -2.00297937e-01
-1.18261671e+00 1.06624222e+00 -5.49391985e-01 -5.92587352e-01
5.79611897e-01 -4.20032531e-01 -5.83867431e-01 -2.56144479e-02
7.25596368e-01 4.64559883e-01 2.86370009e-01 -7.90884256e-01
-8.20432782e-01 -3.08544397e-01 -1.75830007e-01 7.03978771e-03
2.12597981e-01 6.84122801e-01 -2.38365069e-01 -4.23942685e-01
-2.73889542e-01 -6.21951401e-01 -2.96070844e-01 -7.39349782e-01
-5.06381571e-01 -4.98841226e-01 2.82791853e-01 -1.41925669e+00
1.34038258e+00 -2.15470552e+00 -1.66908309e-01 -2.98893273e-01
-4.01370287e-01 4.27611321e-01 -1.41272679e-01 6.05398834e-01
2.71229088e-01 5.24612606e-01 -4.59051251e-01 -6.05931759e-01
2.98025422e-02 3.24923277e-01 -3.42069834e-01 2.79138058e-01
5.65063715e-01 9.11341846e-01 -9.96479273e-01 -8.02552879e-01
-1.47060364e-01 2.76501387e-01 -3.01450759e-01 -4.91752625e-02
-3.77241164e-01 6.74020708e-01 -1.61201626e-01 6.56975806e-01
4.47451264e-01 1.82571039e-01 2.36032784e-01 3.52776825e-01
-4.64010239e-01 1.10983813e+00 -6.04522169e-01 2.00206709e+00
-4.48413342e-01 7.05405176e-01 5.94635531e-02 -5.03987908e-01
7.90888071e-01 4.65694606e-01 -2.45183319e-01 -1.07165098e+00
-5.33506423e-02 4.73580033e-01 3.01573813e-01 -7.61162877e-01
6.12806916e-01 -1.94584548e-01 -4.99955416e-01 3.49831074e-01
-9.04112011e-02 1.06980121e-02 9.08481836e-01 3.70696664e-01
1.20092487e+00 1.26252234e-01 1.85234398e-01 -2.77333319e-01
5.23404777e-01 4.54602480e-01 9.10302639e-01 7.44436324e-01
-4.11928028e-01 4.63712394e-01 8.11015844e-01 -2.35976562e-01
-1.05827188e+00 -1.07660031e+00 -4.65718240e-01 9.02715862e-01
-3.35626036e-01 -3.01641852e-01 -7.32821763e-01 -9.60466325e-01
-1.31546393e-01 9.24703658e-01 -1.61948383e-01 5.01984119e-01
-1.13531315e+00 -6.22983217e-01 1.02015471e+00 3.40075791e-01
3.15896422e-01 -1.31292820e+00 -6.04601614e-02 5.82784295e-01
-7.00800121e-01 -1.57340443e+00 -5.75889647e-01 -7.93769285e-02
-6.21269941e-01 -1.28381968e+00 -3.30606222e-01 -1.02932286e+00
2.95909375e-01 -1.58246458e-01 1.32083476e+00 -8.50040540e-02
-3.79026793e-02 -1.41679555e-01 -4.12297219e-01 -2.48036757e-01
-8.77007008e-01 4.17237133e-01 -3.33770722e-01 -4.50858206e-01
4.88328099e-01 -2.67741978e-01 -4.30382900e-02 -1.26247346e-01
-5.00952065e-01 -1.10071190e-01 4.70675945e-01 6.87959075e-01
2.99877971e-01 -3.02827477e-01 8.06493759e-01 -1.28695941e+00
9.24210489e-01 -2.51652062e-01 -5.82531035e-01 4.29918677e-01
-3.75502139e-01 1.08313635e-01 7.99291432e-01 1.98579803e-01
-1.53966761e+00 -3.36919844e-01 -6.67868614e-01 4.24716413e-01
-3.05287659e-01 6.12992465e-01 -2.98371494e-01 5.24733007e-01
6.50499940e-01 -1.34285733e-01 -3.64892453e-01 -7.74052262e-01
4.74033087e-01 9.18266892e-01 6.41833901e-01 -5.90567291e-01
4.01474476e-01 2.18113229e-01 -5.78366458e-01 -8.13469410e-01
-1.12441456e+00 -4.36326414e-01 -9.55688775e-01 -2.08390746e-02
9.25192416e-01 -9.28105831e-01 -2.16974407e-01 5.10599077e-01
-1.57920468e+00 -5.00764370e-01 3.18931118e-02 2.41020218e-01
6.70739263e-02 6.60885274e-01 -1.25822914e+00 -5.91784000e-01
-5.42098820e-01 -1.25195098e+00 1.04938769e+00 4.47798632e-02
-3.61416250e-01 -1.00665677e+00 3.00836504e-01 8.37744713e-01
2.44143978e-01 3.66726108e-02 1.21590245e+00 -7.76147187e-01
-1.95892408e-01 1.00523517e-01 -2.03390270e-01 3.23733538e-01
-1.40612409e-01 1.79224491e-01 -7.90450096e-01 -1.28188804e-01
-2.31070265e-01 -3.80196929e-01 9.34325695e-01 1.58991620e-01
1.94769159e-01 -8.56051296e-02 -1.77510038e-01 -4.01495062e-02
1.09851921e+00 5.98243997e-02 5.11814892e-01 4.50927526e-01
2.97806442e-01 7.41668046e-01 7.50778139e-01 -3.06310505e-02
7.04990506e-01 3.01041067e-01 -3.18022482e-02 7.07368553e-02
-2.92877734e-01 -3.71300876e-01 6.64711475e-01 1.07337499e+00
4.49684948e-01 -1.56106383e-01 -1.24984205e+00 9.07667041e-01
-1.74950361e+00 -9.62704241e-01 -6.04169726e-01 1.92007399e+00
1.19528365e+00 5.36661983e-01 -1.44784510e-01 9.25346278e-03
9.47742462e-01 1.15560867e-01 -2.09440753e-01 -9.72493708e-01
-2.28794038e-01 4.78795111e-01 2.45655119e-01 7.91814864e-01
-9.82396841e-01 1.62365162e+00 6.59283590e+00 6.11918151e-01
-8.53919923e-01 3.95545423e-01 5.51519156e-01 2.01299801e-01
-3.43065560e-01 2.77882099e-01 -1.40799201e+00 5.34818292e-01
1.37290263e+00 -1.32771552e-01 2.41810739e-01 5.54799557e-01
5.15529513e-01 -3.61356854e-01 -9.54920650e-01 1.97515279e-01
2.05236562e-02 -1.18564999e+00 -4.31197315e-01 -7.58721530e-02
6.33909047e-01 7.49816239e-01 -4.31374341e-01 7.09115922e-01
8.48210931e-01 -9.36429083e-01 7.36570954e-01 -5.43546528e-02
7.06643820e-01 -4.64731187e-01 8.06771696e-01 4.67500091e-01
-9.72295523e-01 1.18882574e-01 -3.36855978e-01 -2.16218829e-01
8.65180016e-01 7.90997267e-01 -9.15626884e-01 6.07474327e-01
7.49756277e-01 9.02521372e-01 -6.39760673e-01 9.29769099e-01
-8.96828532e-01 9.27003264e-01 -1.24229729e-01 6.50895983e-02
2.45451987e-01 -8.99970531e-02 4.00729090e-01 1.63556921e+00
-5.90908788e-02 -3.48233789e-01 2.02181980e-01 5.81258059e-01
-2.39079446e-01 2.94794410e-01 -5.91326654e-01 9.00695683e-04
5.59339583e-01 1.06891787e+00 -5.48742175e-01 -4.46660817e-01
-6.08248472e-01 6.81278586e-01 7.89500356e-01 3.07441741e-01
-7.43380129e-01 -3.91413391e-01 5.67174792e-01 -1.60139546e-01
3.08888257e-01 -3.87098581e-01 -3.22376668e-01 -1.34325278e+00
3.69544953e-01 -9.30585921e-01 6.06994331e-01 -4.96346444e-01
-1.48498392e+00 7.86667824e-01 -2.13629082e-01 -6.97988510e-01
-2.43974388e-01 -5.53823471e-01 -5.66537857e-01 9.25941825e-01
-2.08769202e+00 -1.10295689e+00 3.96877050e-01 3.78619134e-01
7.53514588e-01 1.35405704e-01 5.85115552e-01 5.39750874e-01
-6.75024033e-01 3.49983215e-01 6.51641861e-02 5.99339545e-01
9.68853951e-01 -1.14042628e+00 9.42838788e-01 1.37249255e+00
2.87983805e-01 5.21202922e-01 5.29402852e-01 -9.28082883e-01
-7.89501429e-01 -1.14846575e+00 1.92964184e+00 -4.41781968e-01
1.07640707e+00 -4.21236545e-01 -8.38148475e-01 9.74558949e-01
6.69836938e-01 -4.56877798e-01 7.31905222e-01 4.17128056e-01
-3.02275687e-01 1.63217813e-01 -9.66757476e-01 5.18016100e-01
1.09514141e+00 -6.13417029e-01 -1.15495133e+00 4.40263927e-01
8.55526030e-01 -5.18721938e-01 -1.04600406e+00 2.69297749e-01
2.14197516e-01 -9.57797766e-01 3.98234814e-01 -5.72838902e-01
4.96878326e-01 -5.69091700e-02 5.72833866e-02 -1.28580141e+00
-3.60224843e-02 -5.20654023e-01 4.86718148e-01 1.58106744e+00
1.04735088e+00 -6.29730463e-01 2.80074865e-01 4.08673197e-01
-6.03958726e-01 -4.14754122e-01 -1.28288424e+00 -7.91574955e-01
6.33314967e-01 -7.67782271e-01 3.40720773e-01 6.26924396e-01
2.86496222e-01 7.93531179e-01 4.07207906e-02 2.04104707e-01
3.64592761e-01 1.16964228e-01 4.50498700e-01 -1.04090822e+00
-9.56751183e-02 -2.66909868e-01 1.28761455e-01 -8.43125463e-01
5.99507034e-01 -1.17233109e+00 1.38008818e-01 -1.79825950e+00
3.03877355e-03 -4.43103820e-01 1.35989189e-01 7.82181919e-01
-2.40132093e-01 6.48681596e-02 1.98132440e-01 1.95701107e-01
-6.30139291e-01 3.23052317e-01 1.13378108e+00 -1.59322634e-01
-1.22349210e-01 -2.28617355e-01 -6.14390552e-01 5.36948085e-01
9.09423172e-01 -7.46113598e-01 1.72766253e-01 -9.22554195e-01
2.15380132e-01 -5.29442951e-02 -2.55845666e-01 -7.24688768e-01
1.20117553e-01 1.02345526e-01 -2.56080762e-03 -7.41122305e-01
1.68433376e-02 -1.17253982e-01 -4.47223365e-01 4.50607330e-01
-3.24908048e-01 3.69490266e-01 2.78424889e-01 1.15249224e-01
-3.28659028e-01 -4.49157476e-01 5.41116059e-01 -2.76270419e-01
-5.60094357e-01 -1.85314879e-01 -6.05407000e-01 8.06083143e-01
7.67001212e-01 2.48022765e-01 -8.53901446e-01 1.05309807e-01
-5.90630054e-01 5.43993115e-01 4.13735539e-01 5.74142694e-01
2.23397359e-01 -7.08446503e-01 -9.93191063e-01 1.74532413e-01
-2.50239875e-02 1.94846883e-01 -5.44273145e-02 8.36328626e-01
-5.96268356e-01 8.11564088e-01 1.15628675e-01 -4.70706016e-01
-1.24498272e+00 1.39079228e-01 4.27148342e-02 -8.48280251e-01
-5.13119221e-01 5.40734172e-01 -5.78486323e-01 -8.43663335e-01
1.21612018e-02 -4.54477072e-01 -2.97052592e-01 7.06454515e-02
4.78720099e-01 2.25916535e-01 3.32350224e-01 -8.31467152e-01
-3.41991723e-01 2.23361433e-01 -3.58617723e-01 -3.12537968e-01
1.20643711e+00 -2.13890467e-02 -4.37979549e-01 3.62386584e-01
8.35479558e-01 2.98833579e-01 -7.80380666e-01 -2.05729157e-01
6.46226108e-01 -1.88297719e-01 -3.64228427e-01 -9.21573102e-01
-4.74137545e-01 1.03035629e+00 -9.62908939e-02 7.12633729e-02
5.48613846e-01 8.34814906e-02 1.07325947e+00 2.08736926e-01
4.58770007e-01 -1.40525222e+00 -3.91554892e-01 1.16889966e+00
5.98768651e-01 -1.30513394e+00 -3.93899441e-01 -4.60001588e-01
-5.72882116e-01 8.86372149e-01 4.51098561e-01 2.18968272e-01
4.27093416e-01 3.96233916e-01 4.63336557e-01 -1.87628414e-03
-8.78129661e-01 -2.43965000e-01 -2.25496098e-01 3.58201265e-01
8.83485019e-01 -1.30750790e-01 -9.76141810e-01 5.53641677e-01
-4.05391723e-01 -6.23807050e-02 5.31751990e-01 9.78319705e-01
-3.69415492e-01 -1.72518599e+00 -2.35139385e-01 2.78893501e-01
-9.27903950e-01 -6.94651783e-01 -3.63916367e-01 9.10709083e-01
5.69828004e-02 1.47091115e+00 -5.97563498e-02 1.35647461e-01
4.22117352e-01 4.25226629e-01 2.98492461e-01 -9.55403864e-01
-9.28162873e-01 -8.56282562e-02 7.66079962e-01 -2.86258489e-01
-2.01798141e-01 -1.14076042e+00 -1.48192441e+00 -1.39116600e-01
-7.57981911e-02 3.89560550e-01 5.77510715e-01 1.28202987e+00
5.02939701e-01 2.98475087e-01 1.77276656e-01 -3.77641976e-01
-5.63897908e-01 -1.25899386e+00 -5.04647568e-02 3.94288510e-01
-3.40958461e-02 -1.42008469e-01 -3.60203832e-01 1.48954690e-01] | [10.496269226074219, 9.724899291992188] |
511ebc1f-cdf1-432b-95c5-66a0899aca5d | multi-level-second-order-few-shot-learning | 2201.05916 | null | https://arxiv.org/abs/2201.05916v1 | https://arxiv.org/pdf/2201.05916v1.pdf | Multi-level Second-order Few-shot Learning | We propose a Multi-level Second-order (MlSo) few-shot learning network for supervised or unsupervised few-shot image classification and few-shot action recognition. We leverage so-called power-normalized second-order base learner streams combined with features that express multiple levels of visual abstraction, and we use self-supervised discriminating mechanisms. As Second-order Pooling (SoP) is popular in image recognition, we employ its basic element-wise variant in our pipeline. The goal of multi-level feature design is to extract feature representations at different layer-wise levels of CNN, realizing several levels of visual abstraction to achieve robust few-shot learning. As SoP can handle convolutional feature maps of varying spatial sizes, we also introduce image inputs at multiple spatial scales into MlSo. To exploit the discriminative information from multi-level and multi-scale features, we develop a Feature Matching (FM) module that reweights their respective branches. We also introduce a self-supervised step, which is a discriminator of the spatial level and the scale of abstraction. Our pipeline is trained in an end-to-end manner. With a simple architecture, we demonstrate respectable results on standard datasets such as Omniglot, mini-ImageNet, tiered-ImageNet, Open MIC, fine-grained datasets such as CUB Birds, Stanford Dogs and Cars, and action recognition datasets such as HMDB51, UCF101, and mini-MIT. | ['Piotr Koniusz', 'Hongdong Li', 'Hongguang Zhang'] | 2022-01-15 | null | null | null | null | ['few-shot-action-recognition', 'unsupervised-few-shot-image-classification'] | ['computer-vision', 'computer-vision'] | [ 1.54131025e-01 -3.05281371e-01 -4.85075057e-01 -4.40752178e-01
-6.53949082e-01 -1.53258303e-02 5.76977432e-01 -6.20479928e-04
-5.44000924e-01 3.95858616e-01 3.60138744e-01 2.76532143e-01
5.75642399e-02 -9.29464877e-01 -6.40213728e-01 -4.58430916e-01
-3.63671571e-01 -1.75238490e-01 7.61725068e-01 -2.56638199e-01
9.84496996e-02 5.02959430e-01 -1.94438410e+00 8.49934936e-01
5.48684239e-01 1.58225954e+00 -2.09102016e-02 8.52261186e-01
2.69768853e-03 1.23012698e+00 -3.48289490e-01 4.82563395e-03
3.34968388e-01 -5.19119084e-01 -5.35246432e-01 3.34723890e-01
4.30828691e-01 -4.74740535e-01 -6.07001841e-01 7.78353989e-01
4.88866359e-01 2.67058253e-01 7.15595841e-01 -1.43962455e+00
-5.73877573e-01 2.24026337e-01 -5.30880928e-01 5.56251526e-01
1.05791241e-01 6.78783774e-01 1.01849341e+00 -1.02740979e+00
5.24600327e-01 1.14946163e+00 4.81516063e-01 5.25699675e-01
-1.03183520e+00 -5.84029555e-01 4.60203327e-02 3.59146744e-01
-1.17256403e+00 -4.62887108e-01 5.95139921e-01 -5.98819315e-01
1.14161229e+00 -9.82709825e-02 5.96561134e-01 1.15161562e+00
3.76328707e-01 9.87511098e-01 1.00822663e+00 -1.99209213e-01
6.31301045e-01 -2.61259049e-01 3.79059613e-01 9.45646405e-01
-8.92017558e-02 9.26124081e-02 -1.03072906e+00 -5.91772422e-02
7.86729097e-01 7.72608757e-01 1.35955483e-01 -4.93859768e-01
-1.10733747e+00 9.40363884e-01 8.65909576e-01 3.20828915e-01
-5.05687118e-01 3.56608957e-01 5.21951258e-01 3.28094572e-01
2.44964674e-01 2.19716921e-01 -2.68398315e-01 -2.21167773e-01
-7.98770308e-01 -5.38502336e-02 4.31544960e-01 8.90793800e-01
1.07128882e+00 1.99327737e-01 -7.73141086e-01 9.02271152e-01
3.07493918e-02 7.38084018e-02 9.43613529e-01 -7.16756105e-01
2.29106784e-01 7.78058052e-01 -3.03677201e-01 -6.37860298e-01
-3.24231446e-01 -2.17377335e-01 -1.02307057e+00 4.60246652e-01
-5.22103012e-02 1.02706492e-01 -1.35835719e+00 1.41918766e+00
1.93100601e-01 5.05575120e-01 7.95795545e-02 9.53935504e-01
8.25392067e-01 3.47084492e-01 3.57381046e-01 1.15372621e-01
1.56851411e+00 -1.22266531e+00 -1.73640475e-01 -1.80692360e-01
6.67361498e-01 1.02182314e-01 1.28117430e+00 1.12813979e-01
-7.83479869e-01 -8.34122658e-01 -1.35921645e+00 -1.09264635e-01
-7.95667648e-01 -1.47499755e-01 7.57291973e-01 1.82531059e-01
-6.78573191e-01 9.54951584e-01 -9.32815790e-01 -3.94374251e-01
1.02311003e+00 -1.04055710e-01 -5.17871082e-01 -1.72175333e-01
-1.14668632e+00 5.17153621e-01 3.97446334e-01 -5.36637306e-01
-1.16650033e+00 -7.80860186e-01 -1.26185358e+00 3.62610489e-01
2.13601887e-01 -4.64780182e-01 8.45186591e-01 -8.71491194e-01
-1.40990925e+00 8.23545992e-01 1.06539279e-01 -7.77776301e-01
2.28774399e-01 2.09394805e-02 -4.66183722e-01 4.49015260e-01
3.65509152e-01 7.20652461e-01 1.09981275e+00 -5.96114933e-01
-8.25316548e-01 -5.49305022e-01 2.30199710e-01 1.15320580e-02
-6.30751073e-01 2.65223905e-02 -2.46590823e-01 -6.59364522e-01
-1.33489877e-01 -5.23773313e-01 -3.27180147e-01 2.30049774e-01
-1.48782775e-01 -1.33178234e-01 7.47180402e-01 -1.06460638e-02
1.03023434e+00 -2.64549208e+00 6.20126612e-02 -2.19377011e-01
4.04783756e-01 1.63870111e-01 -3.72685671e-01 2.61029631e-01
-3.03414259e-02 -1.55480593e-01 -2.76513994e-01 -2.56405264e-01
4.95226830e-02 1.58462673e-01 -1.67138949e-01 3.65272135e-01
7.25282133e-01 1.23598897e+00 -9.61479962e-01 -4.51403856e-01
4.16467667e-01 2.39133537e-01 -5.53875208e-01 3.18625242e-01
4.85451333e-02 2.89560221e-02 -4.59887296e-01 1.02190042e+00
2.72388190e-01 -3.86850446e-01 -4.81249392e-01 -8.41658413e-02
-4.44486700e-02 -2.11459637e-01 -1.07351017e+00 1.88840413e+00
-4.39192027e-01 4.63491470e-01 -4.39358532e-01 -1.01143515e+00
7.58538485e-01 -4.21354957e-02 3.92013937e-01 -9.64059412e-01
9.36869979e-02 -1.55999428e-02 -2.60192156e-01 -4.48608398e-01
1.97675779e-01 -1.33674353e-01 -5.29387534e-01 1.75052971e-01
7.77418017e-01 2.55194515e-01 3.62647206e-01 2.23173529e-01
1.63040543e+00 -7.17398822e-02 6.78545296e-01 -8.19087252e-02
2.51026928e-01 -1.34463713e-01 6.41522110e-01 9.64326918e-01
-5.68315208e-01 8.69744778e-01 5.63432336e-01 -7.44747996e-01
-6.66258812e-01 -1.07363176e+00 -1.43877521e-01 1.74543917e+00
3.26854102e-02 -5.21896303e-01 -4.01842952e-01 -7.13507056e-01
2.97971785e-01 4.35369670e-01 -9.92151320e-01 -5.03625453e-01
-2.35082820e-01 -4.98122752e-01 4.96103376e-01 1.04766405e+00
7.59290040e-01 -1.37961698e+00 -1.43225706e+00 2.93980479e-01
5.69620609e-01 -1.16241860e+00 -3.26369375e-01 6.96518242e-01
-6.01122618e-01 -1.10055113e+00 -7.54251063e-01 -4.48826313e-01
3.04055482e-01 4.37237650e-01 9.99974787e-01 -1.10098965e-01
-8.99329662e-01 3.85704339e-01 -5.20821035e-01 -3.50480407e-01
2.25210264e-01 -3.80239822e-02 1.61922216e-01 4.61838067e-01
4.47856009e-01 -7.90683627e-01 -6.65046334e-01 2.93940514e-01
-1.07432294e+00 -3.02659094e-01 7.68116236e-01 1.10098565e+00
6.61351621e-01 -2.70765811e-01 4.13623780e-01 -6.14889324e-01
3.52307379e-01 -6.09218121e-01 -2.64175266e-01 1.46712556e-01
-1.82626303e-02 3.13976593e-02 9.00933206e-01 -4.78584200e-01
-5.96223593e-01 1.29347295e-01 9.38400999e-02 -8.79274726e-01
-4.59691405e-01 4.72116560e-01 -1.19202413e-01 -1.55835683e-02
9.31057870e-01 4.39060956e-01 -1.82927266e-01 -3.63415837e-01
5.03442407e-01 7.13212073e-01 5.39642155e-01 -1.92500889e-01
6.75160944e-01 5.47348917e-01 -1.57062829e-01 -1.06364465e+00
-9.82161880e-01 -7.21618176e-01 -9.45511103e-01 -1.74781904e-02
1.15962660e+00 -1.16620374e+00 -4.05153364e-01 7.52869010e-01
-8.16275477e-01 -4.76047158e-01 -7.31054783e-01 2.76737422e-01
-7.51109600e-01 -7.93549269e-02 -5.73149085e-01 -5.31863868e-01
-2.78108388e-01 -1.00291908e+00 1.26670599e+00 4.71765965e-01
2.05517747e-02 -4.46426034e-01 1.83595456e-02 4.67911884e-02
4.77188617e-01 5.25565147e-01 6.95797443e-01 -7.37893999e-01
-4.38323200e-01 -2.68356025e-01 -2.87146449e-01 3.72954905e-01
-4.32565585e-02 -2.87389398e-01 -1.03023219e+00 -3.47410858e-01
-2.97913760e-01 -1.18704820e+00 1.52825439e+00 1.17648341e-01
1.39864194e+00 -1.30279005e-01 -6.51509836e-02 1.04843068e+00
1.47181153e+00 -6.81484938e-02 7.06177235e-01 2.42132768e-01
7.79673398e-01 9.98970196e-02 5.65569580e-01 7.33890712e-01
3.58782083e-01 5.41852176e-01 3.99214566e-01 -2.14700066e-02
-1.83966801e-01 -1.56542882e-01 3.77502531e-01 3.71515572e-01
-6.81698043e-03 3.04238021e-01 -7.39569724e-01 4.81737345e-01
-2.02366185e+00 -1.36913419e+00 5.11757672e-01 1.86582482e+00
5.77415645e-01 3.72272879e-01 3.37059706e-01 -2.50907894e-02
2.67771453e-01 6.58207893e-01 -8.83505404e-01 -2.70994663e-01
-4.20717001e-02 4.12111253e-01 3.60409379e-01 -1.26121163e-01
-1.36715007e+00 9.31881070e-01 5.24811745e+00 9.38559294e-01
-1.31637216e+00 1.92944840e-01 5.61365187e-01 -3.43775004e-01
3.49826366e-01 -2.74200112e-01 -6.91321015e-01 5.29401124e-01
8.94515336e-01 3.95713653e-03 2.16708645e-01 1.22906995e+00
-3.71214569e-01 2.94369850e-02 -1.18227232e+00 1.00245917e+00
7.88362920e-02 -1.54563046e+00 -6.88484684e-02 -1.51745379e-01
6.32839024e-01 4.51893747e-01 -1.27242342e-01 9.39170480e-01
2.39295915e-01 -1.05000949e+00 5.71867168e-01 4.71744090e-01
1.16624677e+00 -6.85737491e-01 3.92530888e-01 3.81227642e-01
-1.49405420e+00 -7.05957472e-01 -6.07400179e-01 -1.60376385e-01
-1.75668731e-01 2.04620287e-01 -1.37847796e-01 4.14898962e-01
9.33286011e-01 1.01940274e+00 -6.76193058e-01 9.46797073e-01
-9.02623683e-02 2.85793096e-01 -6.93815947e-02 -1.42373396e-02
6.29446745e-01 4.38346118e-01 1.84533373e-01 1.27964735e+00
1.17802490e-02 3.55652750e-01 3.05582076e-01 6.52422130e-01
-1.51042521e-01 -2.22846076e-01 -6.67384088e-01 -7.14466497e-02
2.50668079e-01 1.35683990e+00 -6.62504971e-01 -6.74077392e-01
-6.26220822e-01 1.28430474e+00 5.58979332e-01 2.50290185e-01
-8.93360913e-01 -8.59797180e-01 1.00208986e+00 6.49251118e-02
7.88665652e-01 -3.07561457e-02 2.53684148e-02 -1.48278749e+00
-1.55947059e-01 -7.51419544e-01 7.18776286e-01 -4.73520666e-01
-1.21718609e+00 6.18043661e-01 -9.94807184e-02 -1.45931768e+00
-1.75516829e-01 -6.94785237e-01 -1.07694983e+00 4.67423767e-01
-1.57777548e+00 -1.23918951e+00 -6.70521379e-01 8.81371439e-01
9.29987311e-01 -5.00181258e-01 9.32756841e-01 6.23209178e-02
-6.55606449e-01 7.15929270e-01 -1.61592796e-01 4.29658443e-01
4.10650373e-01 -1.13351524e+00 3.94884527e-01 6.05152488e-01
3.74926180e-01 2.87645489e-01 1.79753274e-01 -2.77739912e-01
-1.39727271e+00 -1.44011641e+00 3.11731994e-01 -1.70243785e-01
8.34343851e-01 -4.93294924e-01 -7.96452940e-01 4.45697010e-01
-4.71876971e-02 1.03508413e+00 6.95103049e-01 -7.41344541e-02
-7.15896249e-01 -1.82894140e-01 -1.06430352e+00 2.60697097e-01
1.30295479e+00 -7.59755552e-01 -6.23256803e-01 2.25181088e-01
8.25092971e-01 -5.65431379e-02 -7.93276429e-01 4.68651921e-01
6.44699574e-01 -1.18535626e+00 1.02403104e+00 -1.18090534e+00
7.23338902e-01 -2.69868392e-02 -4.22078073e-01 -1.28454256e+00
-7.85390258e-01 -2.51703233e-01 -2.89812058e-01 8.11798930e-01
1.49657475e-02 -3.92913520e-01 7.33854115e-01 1.47501111e-01
-2.22338140e-01 -1.13988888e+00 -9.73038793e-01 -9.51537669e-01
-2.73422867e-01 -4.18801814e-01 4.32405025e-01 8.04358244e-01
1.65395111e-01 4.31193680e-01 -4.25861955e-01 -1.58878729e-01
7.53722966e-01 1.26580134e-01 8.00217271e-01 -1.16234148e+00
-4.66399878e-01 -4.68966991e-01 -1.12914252e+00 -7.42250860e-01
8.17258358e-02 -7.95573413e-01 -2.56472197e-03 -1.17875767e+00
3.52813154e-01 1.41447529e-01 -9.93677974e-01 7.94673264e-01
-9.55689549e-02 5.49343765e-01 2.43835136e-01 1.17356718e-01
-1.01020944e+00 7.94746876e-01 7.96529174e-01 -3.55074853e-01
-8.68214816e-02 -2.96732128e-01 -4.43935484e-01 8.14963102e-01
5.54071665e-01 -4.04628992e-01 -3.04202139e-01 -1.98432684e-01
-6.39987350e-01 1.11472625e-02 7.47510374e-01 -1.57592309e+00
3.72922778e-01 -3.02542210e-01 6.70192063e-01 -3.77978861e-01
3.47297192e-01 -5.59950411e-01 -5.18154740e-01 5.71291566e-01
-5.30918777e-01 -4.40189570e-01 2.56320671e-03 6.53331220e-01
-2.71796733e-01 1.84066698e-01 1.19615483e+00 -3.39664310e-01
-1.42587328e+00 7.57858932e-01 -2.93609910e-02 1.57446444e-01
1.35878205e+00 -3.10467750e-01 -2.98629552e-01 1.17689490e-01
-6.34329379e-01 3.38300556e-01 3.40399027e-01 6.11809731e-01
8.61430526e-01 -1.66528285e+00 -5.76407552e-01 5.25662363e-01
9.36579168e-01 -3.35095435e-01 4.15820420e-01 8.18697989e-01
-1.56491697e-01 1.61503986e-01 -9.12761748e-01 -6.78983152e-01
-8.46915781e-01 5.36630094e-01 1.70765370e-01 -1.41955078e-01
-7.45093703e-01 1.01658940e+00 2.87478894e-01 -5.57700209e-02
3.24752241e-01 -2.45049402e-01 -1.04905076e-01 2.61664629e-01
1.10945046e+00 2.10868984e-01 -1.71974644e-01 -6.23303950e-01
-5.61849415e-01 2.93085337e-01 -8.82198140e-02 2.24739075e-01
1.57250381e+00 2.89602518e-01 3.86305481e-01 7.23900020e-01
1.57591736e+00 -7.58021653e-01 -1.62619293e+00 -3.67141098e-01
-3.27320695e-02 -4.05408740e-01 -6.94148168e-02 -4.03023690e-01
-9.24887657e-01 1.10991228e+00 6.47312999e-01 7.62242228e-02
1.06967044e+00 7.46661425e-02 6.83469236e-01 4.31212038e-01
6.58746004e-01 -1.16027212e+00 5.65543592e-01 6.02398276e-01
7.73973286e-01 -1.50756168e+00 -3.01470667e-01 1.85549468e-01
-8.14040422e-01 9.65226948e-01 9.36080337e-01 -4.31642085e-01
6.32836819e-01 2.84698695e-01 -1.87175825e-01 -3.88353050e-01
-1.02842617e+00 -6.16780579e-01 3.45711082e-01 4.71955478e-01
4.28595617e-02 -6.58084750e-02 6.80764392e-02 9.00243521e-01
3.48671973e-01 2.75162637e-01 1.81887701e-01 1.18791437e+00
-8.96059513e-01 -3.77963305e-01 1.28792554e-01 6.98422074e-01
-2.39784166e-01 -6.25512823e-02 -1.66138470e-01 5.56848288e-01
7.31568784e-02 5.06819963e-01 2.80775428e-01 -7.40678906e-01
4.55074251e-01 8.92837346e-02 2.19913289e-01 -9.01378572e-01
-3.88981253e-01 -5.14165461e-01 -3.00373495e-01 -1.29468441e+00
-1.55747905e-01 -1.93097159e-01 -1.30493879e+00 1.48370758e-01
1.42205998e-01 -2.08308741e-01 2.23513171e-01 9.10880625e-01
5.97018898e-01 6.17860198e-01 8.46748888e-01 -1.18668211e+00
-9.06311452e-01 -9.96907055e-01 -8.21728647e-01 6.96361244e-01
3.93574268e-01 -8.29987168e-01 -2.89274454e-01 -7.41213560e-02] | [9.873152732849121, 2.679337739944458] |
977ef510-c22d-4ea7-8517-2c7f455908e9 | cross-modal-causal-relational-reasoning-for | 2207.12647 | null | https://arxiv.org/abs/2207.12647v8 | https://arxiv.org/pdf/2207.12647v8.pdf | Cross-Modal Causal Relational Reasoning for Event-Level Visual Question Answering | Existing visual question answering methods often suffer from cross-modal spurious correlations and oversimplified event-level reasoning processes that fail to capture event temporality, causality, and dynamics spanning over the video. In this work, to address the task of event-level visual question answering, we propose a framework for cross-modal causal relational reasoning. In particular, a set of causal intervention operations is introduced to discover the underlying causal structures across visual and linguistic modalities. Our framework, named Cross-Modal Causal RelatIonal Reasoning (CMCIR), involves three modules: i) Causality-aware Visual-Linguistic Reasoning (CVLR) module for collaboratively disentangling the visual and linguistic spurious correlations via front-door and back-door causal interventions; ii) Spatial-Temporal Transformer (STT) module for capturing the fine-grained interactions between visual and linguistic semantics; iii) Visual-Linguistic Feature Fusion (VLFF) module for learning the global semantic-aware visual-linguistic representations adaptively. Extensive experiments on four event-level datasets demonstrate the superiority of our CMCIR in discovering visual-linguistic causal structures and achieving robust event-level visual question answering. The datasets, code, and models are available at https://github.com/HCPLab-SYSU/CMCIR. | ['Liang Lin', 'Guanbin Li', 'Yang Liu'] | 2022-07-26 | null | null | null | null | ['relational-reasoning'] | ['natural-language-processing'] | [-2.55219311e-01 -1.70869738e-01 -3.31005365e-01 -4.57266957e-01
-7.36039340e-01 -6.46524727e-01 1.07875502e+00 3.30493689e-01
2.79103518e-01 2.84221858e-01 8.90456200e-01 -4.24307495e-01
-4.52536494e-01 -6.12701774e-01 -8.23210657e-01 -3.58981699e-01
-1.91770434e-01 7.14550242e-02 3.87983948e-01 -2.56198384e-02
1.37486443e-01 2.05726787e-01 -1.75620389e+00 1.15235996e+00
5.30595958e-01 8.28634143e-01 -6.98475316e-02 8.68251920e-01
-3.77105087e-01 2.07772517e+00 -2.09660143e-01 -2.95201302e-01
-3.70268136e-01 -6.89308524e-01 -1.05499506e+00 -9.01806429e-02
4.22658861e-01 -1.42535523e-01 -5.61241090e-01 7.55240083e-01
5.99680878e-02 2.47267976e-01 5.72701573e-01 -1.77060878e+00
-1.07832670e+00 6.35479033e-01 -7.67252028e-01 7.27293730e-01
8.42391491e-01 3.10416073e-01 1.31834757e+00 -9.60189939e-01
7.50446737e-01 1.97642410e+00 2.20652372e-01 2.77369916e-01
-1.08660066e+00 -5.51377714e-01 5.68849027e-01 9.57797647e-01
-1.23300099e+00 -1.16792068e-01 1.04946673e+00 -7.74890363e-01
7.13040292e-01 3.58144939e-01 6.07315063e-01 1.25387788e+00
3.27963494e-02 9.60152030e-01 1.08001018e+00 -2.14708805e-01
-6.24081865e-02 -3.08536977e-01 1.85739100e-01 1.03950369e+00
-3.61082166e-01 2.72673279e-01 -1.13219118e+00 -2.62927234e-01
7.57186174e-01 1.18329935e-01 -1.52507752e-01 -3.00061315e-01
-1.58702052e+00 8.33390951e-01 6.88111663e-01 3.37059021e-01
-4.55765784e-01 6.37908280e-01 4.56253618e-01 1.24998920e-01
9.75317210e-02 -2.28764080e-02 -2.11215779e-01 2.85199761e-01
-4.01254326e-01 2.15560704e-01 2.90243715e-01 8.56872082e-01
5.19090593e-01 -2.34951630e-01 -8.11072230e-01 6.54594004e-01
7.24150121e-01 2.10391566e-01 2.06449162e-02 -1.18025613e+00
3.72786105e-01 8.78067613e-01 -1.57161076e-02 -1.31204975e+00
-4.14413452e-01 4.27268416e-01 -5.97213209e-01 -8.08494836e-02
3.58087778e-01 3.05624813e-01 -7.38190830e-01 1.85989678e+00
6.91879630e-01 3.76830816e-01 -1.43652290e-01 1.12011397e+00
1.32633412e+00 7.80722022e-01 8.20153832e-01 -3.00221264e-01
1.93198323e+00 -7.49314606e-01 -9.30553555e-01 1.18320677e-02
6.40982911e-02 -6.71916485e-01 1.45874274e+00 -1.24328814e-01
-1.16897833e+00 -6.12811863e-01 -6.21708512e-01 -5.64646602e-01
-3.71893436e-01 -2.38026649e-01 6.13993406e-01 -3.85429174e-01
-7.35415637e-01 -2.18292639e-01 -4.97806460e-01 -2.92663127e-01
3.89457405e-01 -3.09718937e-01 -2.77843773e-01 -2.66510755e-01
-1.65196872e+00 5.08265972e-01 2.80136406e-01 5.73581783e-03
-1.24695146e+00 -1.02075958e+00 -7.88830698e-01 -1.56845465e-01
6.69900596e-01 -1.09277821e+00 1.15114605e+00 -8.78508389e-01
-7.91485488e-01 8.56704295e-01 -5.22820830e-01 -4.25210362e-03
4.76088554e-01 -1.01264343e-01 -5.42407990e-01 5.67775369e-01
3.01269352e-01 5.10676861e-01 8.59378040e-01 -1.65901077e+00
-6.39214337e-01 -4.78298753e-01 2.96097100e-01 2.97505289e-01
2.49698639e-01 2.31811672e-01 -7.14668930e-01 -7.22898066e-01
-9.25668627e-02 -1.93400204e-01 2.15118363e-01 1.39754206e-01
-2.75600433e-01 -7.13889539e-01 8.97465110e-01 -6.88672841e-01
1.22303045e+00 -2.19579887e+00 2.83550978e-01 -1.52362511e-01
4.05856192e-01 -3.99057418e-01 -5.04865497e-02 5.44517994e-01
-4.46814358e-01 1.03448972e-01 2.76035607e-01 5.98842511e-03
-4.01829667e-02 2.14654416e-01 -7.03821063e-01 3.11380982e-01
3.46347392e-01 1.35905576e+00 -1.25194585e+00 -1.04134297e+00
4.48743373e-01 6.07790589e-01 -3.00230235e-01 4.36684817e-01
-6.50653780e-01 5.21922946e-01 -5.19481659e-01 8.48973513e-01
5.08336909e-02 -5.92923582e-01 1.65191516e-01 -7.16405511e-01
-8.09478611e-02 2.11994439e-01 -8.81588697e-01 1.52727270e+00
-2.40552202e-01 5.59992313e-01 -3.30354363e-01 -7.81387806e-01
4.90019709e-01 4.71407950e-01 4.59069610e-01 -1.08997679e+00
-4.44140397e-02 -4.64683115e-01 -4.18279886e-01 -1.03451788e+00
9.10895504e-03 -2.47406587e-01 -1.87915429e-01 2.73072302e-01
1.81343719e-01 2.82100469e-01 1.99405029e-01 8.13492715e-01
9.96670663e-01 2.71432161e-01 3.11595291e-01 4.67771515e-02
6.13303602e-01 2.44659424e-01 4.08843458e-01 4.00660753e-01
-4.25186157e-01 3.02409738e-01 9.15134013e-01 -4.09934878e-01
-7.95641840e-01 -1.57455659e+00 2.67785072e-01 1.41494274e+00
5.05753219e-01 -5.78503132e-01 -2.63308167e-01 -5.24013817e-01
1.63360640e-01 9.99799073e-01 -1.04060614e+00 -1.38803586e-01
-4.48949724e-01 -2.70158261e-01 5.22056162e-01 6.79562747e-01
2.45807409e-01 -1.20691407e+00 -3.89896303e-01 -1.36915728e-01
-7.77510762e-01 -1.14971054e+00 -4.61687624e-01 -3.39894772e-01
-4.37290817e-01 -1.55608928e+00 1.72382727e-01 -3.82007509e-01
3.38731527e-01 3.46966594e-01 1.22682500e+00 9.02129151e-03
-5.37711084e-01 9.20449078e-01 -2.59064049e-01 -1.36762246e-01
-1.91981092e-01 -9.50939119e-01 -4.56886977e-01 2.36387312e-01
3.67567420e-01 -2.83095717e-01 -7.06887007e-01 3.21021110e-01
-8.06742251e-01 3.52546930e-01 1.24969043e-01 6.27468705e-01
8.43287230e-01 -1.53021456e-03 4.09600049e-01 -5.06055474e-01
4.83350903e-01 -8.75765622e-01 -3.97606969e-01 6.89517558e-01
-8.89079720e-02 1.05669476e-01 3.61489326e-01 -5.66187739e-01
-1.58016598e+00 -2.87997246e-01 4.45031017e-01 -9.29866672e-01
-2.27993757e-01 4.74859923e-01 -1.92344487e-01 7.53228366e-01
6.50915563e-01 -2.30676681e-02 -3.17996264e-01 -3.01708058e-02
1.07331419e+00 -6.70727417e-02 7.80716062e-01 -6.75863266e-01
6.90242052e-01 9.56592858e-01 -1.23326950e-01 -3.99586052e-01
-1.18113232e+00 -6.40122533e-01 -5.17691135e-01 -7.73218632e-01
1.51867139e+00 -1.15918136e+00 -1.29112136e+00 2.55174227e-02
-1.27556646e+00 -3.01202714e-01 -3.51693571e-01 2.01926023e-01
-6.53947651e-01 2.47633755e-01 -5.98183632e-01 -7.53790021e-01
4.47235443e-02 -6.63454354e-01 1.07965577e+00 1.54548332e-01
-2.94435352e-01 -1.21967924e+00 -1.34697976e-02 9.63726044e-01
-2.14161620e-01 4.81883168e-01 1.24059963e+00 -8.35285932e-02
-8.58459830e-01 3.41577202e-01 -7.23996639e-01 -3.48654002e-01
6.09536357e-02 3.30656797e-01 -8.61496568e-01 3.39591056e-01
-2.48218685e-01 -4.06112343e-01 7.32260704e-01 3.36599231e-01
1.24360693e+00 -2.84081668e-01 -2.35260159e-01 1.84657529e-01
1.24827683e+00 2.08137780e-01 3.42107475e-01 -1.02338372e-02
1.26672435e+00 9.17480111e-01 7.65916765e-01 3.01332235e-01
1.18115664e+00 4.14560556e-01 5.41277707e-01 -7.48200491e-02
-5.73449612e-01 -8.36721301e-01 3.88802618e-01 5.24071753e-01
-1.92339256e-01 -8.06127042e-02 -1.12102234e+00 7.21354842e-01
-2.42313743e+00 -1.57502878e+00 -6.59653187e-01 1.52644265e+00
7.48039663e-01 -4.12795722e-01 1.37075558e-01 -1.03481211e-01
6.99489534e-01 3.36567163e-01 -5.62785506e-01 -9.66384709e-02
-2.66973637e-02 -4.82069612e-01 -1.34595305e-01 5.77926159e-01
-9.70029354e-01 9.34130132e-01 5.28446484e+00 5.71281493e-01
-5.19832909e-01 5.79217017e-01 3.93060476e-01 -1.32733345e-01
-7.59529233e-01 1.94069520e-01 -2.67139226e-01 1.21971436e-01
5.08207440e-01 -3.87950428e-02 4.49002713e-01 2.83559263e-01
4.74058896e-01 -1.71547726e-01 -1.03364336e+00 9.05364990e-01
1.35995448e-01 -1.50336421e+00 3.92288446e-01 -3.93051386e-01
3.82174999e-01 -2.91524529e-01 -3.08054537e-01 7.85126686e-02
7.52848327e-01 -9.46073651e-01 1.16938770e+00 1.00175250e+00
6.69312000e-01 -5.96394181e-01 -2.72159297e-02 5.19060381e-02
-1.73712409e+00 -3.22130203e-01 2.22890034e-01 1.13334112e-01
3.39447975e-01 4.63564157e-01 -2.89664596e-01 7.33704448e-01
1.17267501e+00 8.30049217e-01 -4.59864289e-01 1.92350045e-01
-6.82896972e-01 4.62951362e-01 2.78675139e-01 1.41782060e-01
-4.87542599e-02 2.77940005e-01 4.72544968e-01 1.04418433e+00
-3.09288412e-01 4.68717396e-01 -6.21962734e-02 1.17199981e+00
2.13857368e-01 -1.74322739e-01 -4.44807142e-01 -7.52740502e-02
4.39301282e-01 1.03537989e+00 -5.54004192e-01 -3.86851132e-01
-6.44512832e-01 7.73734689e-01 5.66536248e-01 6.40508175e-01
-1.32464707e+00 3.58692616e-01 6.83255494e-01 3.28423306e-02
1.26736522e-01 -1.57401860e-01 -1.35416597e-01 -1.36394334e+00
-2.01530188e-01 -6.84541643e-01 1.20851707e+00 -1.52676582e+00
-1.67688501e+00 1.46374211e-01 3.71083707e-01 -7.57596672e-01
9.67803970e-02 -2.89526880e-01 -4.60280985e-01 6.74029648e-01
-1.42333114e+00 -1.62458086e+00 -6.43241227e-01 1.38844383e+00
6.83503687e-01 3.38433474e-01 3.61606151e-01 2.41858810e-01
-3.39935482e-01 -7.56947547e-02 -9.32926178e-01 -2.02145837e-02
6.71230197e-01 -1.17107666e+00 -2.34519705e-01 8.72124434e-01
2.74678171e-01 4.91313308e-01 6.06947362e-01 -7.70532906e-01
-1.66602743e+00 -1.20169997e+00 1.04693258e+00 -8.68401706e-01
1.24596775e+00 -3.65007788e-01 -1.01970851e+00 8.59509468e-01
4.18665379e-01 9.17068869e-02 5.84536016e-01 2.05570981e-01
-1.13056719e+00 8.72220565e-03 -6.53281629e-01 8.59362602e-01
1.23103201e+00 -1.14525914e+00 -1.01873732e+00 3.25376540e-01
9.75280344e-01 -2.46747285e-02 -8.50912094e-01 2.68558979e-01
4.32992935e-01 -9.08063591e-01 1.41990030e+00 -6.99924529e-01
8.74591529e-01 -8.32982063e-01 -2.81122684e-01 -6.55946016e-01
-5.44685543e-01 -2.23392993e-01 -4.34935182e-01 1.43182015e+00
7.08393008e-02 -6.44584075e-02 -1.51667118e-01 3.20164502e-01
3.10782522e-01 -4.11122501e-01 -6.81251228e-01 -2.36766174e-01
-2.85588145e-01 -8.70469928e-01 1.73771113e-01 1.34282172e+00
9.21155289e-02 7.43959785e-01 -2.75743455e-01 6.03053570e-01
6.84656501e-01 4.59944218e-01 3.99709165e-01 -9.06520247e-01
-2.26232216e-01 -5.37817121e-01 -2.47224271e-02 -4.56206977e-01
1.41879469e-01 -7.59111226e-01 -1.19141772e-01 -1.91203499e+00
5.34218550e-01 9.28031206e-02 -3.56894761e-01 7.09310055e-01
-4.89530772e-01 -5.47249317e-02 2.52157956e-01 3.20006877e-01
-9.23743069e-01 4.71576542e-01 1.50608492e+00 -1.66488573e-01
1.37941003e-01 -6.80361450e-01 -6.14302039e-01 8.69528294e-01
2.96917826e-01 -2.38037094e-01 -8.59236419e-01 -3.67037147e-01
5.64097524e-01 5.74624538e-01 1.25685561e+00 -2.59042919e-01
3.29175651e-01 -6.23945713e-01 3.20839554e-01 -6.94724917e-01
1.48773506e-01 -6.71087921e-01 2.38282040e-01 1.04703441e-01
-5.54565191e-01 2.18208805e-01 1.16002500e-01 8.94035339e-01
-5.29908419e-01 6.24505103e-01 5.66077888e-01 -1.67394489e-01
-1.02355540e+00 2.98075140e-01 -2.77219206e-01 2.75282353e-01
1.27919161e+00 2.24844828e-01 -8.62167656e-01 -4.10246640e-01
-8.64934504e-01 6.58740342e-01 -3.53447571e-02 9.54726219e-01
6.94977641e-01 -1.55205643e+00 -5.11304200e-01 -2.84935772e-01
5.28757870e-01 -2.48202786e-01 9.21084404e-01 1.03659844e+00
-8.64235088e-02 2.86875069e-01 2.19590530e-01 -5.97743094e-01
-1.29190683e+00 1.12679386e+00 3.50450754e-01 -1.50029242e-01
-5.04000068e-01 9.33392406e-01 7.79494643e-01 -1.25917509e-01
1.35401294e-01 -1.93689957e-01 -3.73364151e-01 4.94174868e-01
5.85983515e-01 2.79095322e-01 -6.04884088e-01 -8.16678345e-01
-6.64728224e-01 4.61451888e-01 4.84282702e-01 -1.85921818e-01
8.90391111e-01 -3.89212310e-01 -4.46136266e-01 1.05232179e+00
9.72656131e-01 -2.45724246e-01 -1.29041779e+00 -2.97725320e-01
-2.78546754e-02 -3.06101561e-01 -6.56588748e-02 -9.42701399e-01
-8.15981567e-01 9.24102664e-01 2.64967442e-01 4.09217417e-01
1.23741603e+00 7.93209255e-01 1.42778277e-01 -4.17689890e-01
8.70383903e-02 -6.31036580e-01 6.13793671e-01 2.01081693e-01
1.43751442e+00 -1.10973012e+00 -5.35830669e-02 -4.92786616e-01
-9.70349729e-01 8.29070747e-01 7.04516470e-01 6.27097040e-02
7.07387984e-01 5.74075803e-02 1.29406169e-01 -9.15551484e-01
-1.49275708e+00 -4.39321667e-01 6.59203231e-01 3.11395943e-01
3.81859541e-01 3.25657904e-01 -9.19043794e-02 4.89910394e-01
3.70643914e-01 -1.02165714e-01 -3.26924883e-02 6.32357538e-01
3.94794904e-02 -5.95146894e-01 -5.26117802e-01 -1.40451267e-01
-1.63660944e-01 -1.14146501e-01 -4.67375904e-01 9.46808457e-01
3.54293525e-01 1.19491982e+00 3.43940228e-01 -2.06721947e-01
4.67870414e-01 1.36668816e-01 5.00785887e-01 -8.52094144e-02
-4.66716141e-01 1.01281099e-01 5.48136607e-03 -1.18247736e+00
-9.55169499e-01 -8.62970829e-01 -1.67052436e+00 -2.11454228e-01
2.56144643e-01 -2.56512880e-01 1.76537663e-01 1.02717233e+00
3.94026846e-01 9.63934064e-01 3.11159521e-01 -2.53589720e-01
3.45061809e-01 -4.72275943e-01 -1.72928035e-01 8.56091499e-01
4.55121666e-01 -9.41949725e-01 -4.16874826e-01 5.68457901e-01] | [10.49211597442627, 1.269737958908081] |
97231c0e-8f77-4791-ad3d-e908a5780921 | text-to-image-editing-by-image-information | 2305.17489 | null | https://arxiv.org/abs/2305.17489v1 | https://arxiv.org/pdf/2305.17489v1.pdf | Text-to-image Editing by Image Information Removal | Diffusion models have demonstrated impressive performance in text-guided image generation. To leverage the knowledge of text-guided image generation models in image editing, current approaches either fine-tune the pretrained models using the input image (e.g., Imagic) or incorporate structure information as additional constraints into the pretrained models (e.g., ControlNet). However, fine-tuning large-scale diffusion models on a single image can lead to severe overfitting issues and lengthy inference time. The information leakage from pretrained models makes it challenging to preserve the text-irrelevant content of the input image while generating new features guided by language descriptions. On the other hand, methods that incorporate structural guidance (e.g., edge maps, semantic maps, keypoints) as additional constraints face limitations in preserving other attributes of the original image, such as colors or textures. A straightforward way to incorporate the original image is to directly use it as an additional control. However, since image editing methods are typically trained on the image reconstruction task, the incorporation can lead to the identical mapping issue, where the model learns to output an image identical to the input, resulting in limited editing capabilities. To address these challenges, we propose a text-to-image editing model with Image Information Removal module (IIR) to selectively erase color-related and texture-related information from the original image, allowing us to better preserve the text-irrelevant content and avoid the identical mapping issue. We evaluate our model on three benchmark datasets: CUB, Outdoor Scenes, and COCO. Our approach achieves the best editability-fidelity trade-off, and our edited images are approximately 35% more preferred by annotators than the prior-arts on COCO. | ['Bryan A. Plummer', 'Jacob Zhiyuan Fang', 'Jian Zheng', 'Zhongping Zhang'] | 2023-05-27 | null | null | null | null | ['image-reconstruction'] | ['computer-vision'] | [ 6.51659429e-01 1.35835305e-01 5.89493141e-02 -3.01161617e-01
-5.28342009e-01 -6.75209761e-01 5.27126491e-01 -2.74125993e-01
-3.76165390e-01 6.92725122e-01 1.62355065e-01 3.73937190e-02
2.60532171e-01 -9.87346113e-01 -1.02003324e+00 -6.86643302e-01
5.67849457e-01 1.86066553e-02 1.06064968e-01 -3.62938613e-01
3.46041471e-01 5.56918681e-01 -1.33872116e+00 3.29947829e-01
1.23079550e+00 7.96939015e-01 5.01278102e-01 5.09773850e-01
-2.30113477e-01 7.64592290e-01 -5.12582839e-01 -4.99521822e-01
4.80865657e-01 -5.01239717e-01 -5.36201596e-01 3.66564244e-01
6.67671800e-01 -5.69531500e-01 -4.30763036e-01 1.37073851e+00
3.17820251e-01 6.91349506e-02 5.63735008e-01 -1.26325285e+00
-1.24736786e+00 3.64714473e-01 -7.81548381e-01 -4.36788470e-01
1.26830727e-01 3.02699029e-01 6.52135968e-01 -1.10076892e+00
1.03884351e+00 1.17846048e+00 3.26723099e-01 5.74408233e-01
-1.45831156e+00 -7.08164692e-01 3.89598906e-01 -9.18866396e-02
-1.37566888e+00 -5.06469488e-01 9.18452859e-01 -3.05984646e-01
5.22575140e-01 2.93049365e-01 5.95699131e-01 1.18672299e+00
2.00168341e-01 6.63984239e-01 1.12901485e+00 -3.80327076e-01
1.17995255e-01 3.10648233e-01 -5.49713910e-01 9.15360987e-01
1.56087145e-01 2.54924625e-01 -4.54109639e-01 2.29059234e-01
1.00289142e+00 1.76414385e-01 -4.29536730e-01 -4.94285196e-01
-1.18929231e+00 8.06060970e-01 6.48425400e-01 -7.76632428e-02
-2.41223782e-01 1.62594810e-01 1.06392078e-01 2.79340088e-01
4.09959644e-01 6.33826971e-01 -3.42908531e-01 1.99591964e-01
-9.58453357e-01 1.76676348e-01 4.17696744e-01 1.26439345e+00
1.13872874e+00 1.17256969e-01 -4.71399188e-01 9.52487469e-01
5.20329028e-02 5.72351933e-01 2.70804912e-01 -1.07838511e+00
5.73336244e-01 5.62120080e-01 5.87365031e-02 -1.17772543e+00
1.32556036e-01 -3.55195343e-01 -1.14770651e+00 4.87832218e-01
2.03743726e-01 -1.70820698e-01 -1.50422907e+00 1.92437673e+00
1.75568312e-01 -2.00841695e-01 -1.07610330e-01 9.96725023e-01
5.47316730e-01 7.41118193e-01 -7.44554102e-02 2.89215520e-02
1.10391951e+00 -1.37935221e+00 -6.68390691e-01 -3.85313034e-01
2.48070195e-01 -8.86884749e-01 1.29800558e+00 1.71836883e-01
-1.10662532e+00 -5.14005184e-01 -1.09854388e+00 -3.79220456e-01
-4.19623196e-01 1.59592777e-01 4.50406671e-01 3.12174320e-01
-1.03426492e+00 4.80588645e-01 -5.86012304e-01 -3.34518969e-01
3.55046749e-01 7.67554939e-02 -4.19180125e-01 -3.66182506e-01
-9.91741240e-01 6.46027565e-01 2.87462562e-01 -7.29527101e-02
-9.61152852e-01 -8.80858183e-01 -9.39475298e-01 3.21431421e-02
6.13783896e-01 -7.46891022e-01 7.80776799e-01 -1.39388239e+00
-1.71479750e+00 6.83388352e-01 -2.09077932e-02 4.24246100e-04
7.72715569e-01 -5.74168488e-02 -1.36424989e-01 6.24779910e-02
2.12291762e-01 1.18478894e+00 1.20411479e+00 -1.50749421e+00
-5.83850026e-01 8.94590691e-02 4.09105152e-01 2.54613698e-01
-4.15012866e-01 -3.19533765e-01 -1.13087082e+00 -1.30640447e+00
-2.42649373e-02 -1.13161325e+00 -3.59380335e-01 5.76905906e-01
-6.52158439e-01 6.30371690e-01 9.35571790e-01 -7.17958570e-01
1.04411066e+00 -2.16918468e+00 3.26914728e-01 3.19904596e-01
1.82890370e-01 8.82633105e-02 -6.63957417e-01 3.25435579e-01
3.35319055e-04 2.80339539e-01 -5.29855549e-01 -3.71452540e-01
-1.73262894e-01 3.05366307e-01 -3.03645313e-01 -2.60210503e-02
3.68149489e-01 1.04267752e+00 -8.67987514e-01 -4.92523581e-01
3.01241755e-01 5.98411858e-01 -8.79368186e-01 2.67340064e-01
-3.56897205e-01 6.46495044e-01 -3.81293952e-01 4.72647280e-01
8.03486228e-01 -2.13698342e-01 -4.69950102e-02 -5.30961812e-01
-2.84612458e-02 -2.27989748e-01 -1.05682743e+00 1.92529869e+00
-7.15128422e-01 6.17357850e-01 1.18863419e-01 -7.33880103e-01
7.22805023e-01 -5.30985817e-02 3.15495253e-01 -8.85613203e-01
-2.02129513e-01 1.27989165e-02 -1.94195345e-01 -1.99938357e-01
6.13285899e-01 1.75954029e-01 1.30676001e-01 5.38077295e-01
-1.15420923e-01 -3.24373573e-01 1.54890284e-01 4.63160545e-01
8.35895479e-01 4.38557476e-01 -1.57590713e-02 -8.06744471e-02
3.34181309e-01 1.21593866e-02 7.08582044e-01 7.96860278e-01
2.74395913e-01 1.15559971e+00 2.51803190e-01 -2.80187011e-01
-1.25363410e+00 -1.00458741e+00 1.69605404e-01 9.18765962e-01
2.64645249e-01 -4.19519126e-01 -1.03692067e+00 -7.40121365e-01
-1.86035916e-01 6.35138333e-01 -7.29112029e-01 -3.54605466e-01
-4.57406014e-01 -6.04232073e-01 3.91713530e-01 4.51451600e-01
8.00824702e-01 -8.88430417e-01 -1.74094915e-01 2.42707983e-01
-3.55775416e-01 -1.12115216e+00 -1.12949765e+00 -8.50219056e-02
-6.98403060e-01 -7.39367425e-01 -9.16586637e-01 -6.45682096e-01
1.37363744e+00 3.28148454e-01 9.96135890e-01 2.81847268e-01
-4.50732589e-01 1.28118619e-01 -3.07384521e-01 -1.06524549e-01
-4.17604685e-01 -6.87861219e-02 -2.92689264e-01 2.58227259e-01
-4.22110498e-01 -4.81245071e-01 -8.38033259e-01 4.74218488e-01
-1.29978704e+00 5.83123684e-01 7.40815282e-01 1.08429873e+00
7.98458934e-01 1.91286709e-02 3.07299554e-01 -1.09829378e+00
4.98238355e-01 -4.21835333e-02 -6.17325306e-01 5.54128230e-01
-7.69405186e-01 3.04644346e-01 8.06093156e-01 -6.65389359e-01
-1.39791381e+00 2.84472734e-01 4.31890264e-02 -5.57221055e-01
2.33643770e-01 3.96575749e-01 -4.09931868e-01 -3.75034750e-01
5.80206871e-01 3.75114918e-01 -8.48268420e-02 -3.40963036e-01
6.77884877e-01 2.27225780e-01 6.04212284e-01 -6.58895254e-01
1.06552422e+00 5.19010901e-01 -2.85421401e-01 -6.60891652e-01
-8.91545415e-01 1.79857954e-01 -5.02577424e-01 1.58470925e-02
8.51631045e-01 -9.47268844e-01 -1.28469422e-01 5.03204823e-01
-1.11301029e+00 -4.32147056e-01 -3.73711497e-01 1.15791559e-01
-4.05686021e-01 3.52684051e-01 -5.58174074e-01 -1.67708889e-01
-3.19285661e-01 -1.39449954e+00 9.99179542e-01 2.41647139e-01
-4.87864576e-03 -8.63042355e-01 -3.63785893e-01 3.06005210e-01
7.06478894e-01 3.27587575e-01 1.06836486e+00 2.41965637e-01
-9.25611556e-01 3.74669060e-02 -5.85091054e-01 5.16165853e-01
3.46982211e-01 1.95312604e-01 -8.24343085e-01 -3.63233060e-01
-2.76248336e-01 -2.18648612e-01 9.00690556e-01 1.86964020e-01
1.42420793e+00 -3.63465220e-01 -1.83265984e-01 9.82286453e-01
1.37500596e+00 7.13512301e-02 8.54808688e-01 1.91806734e-01
9.84864175e-01 4.57045585e-01 5.34041166e-01 2.42252156e-01
3.30627650e-01 6.78518355e-01 1.93938643e-01 -5.22437155e-01
-6.50110066e-01 -6.94723070e-01 4.08966243e-01 5.79501092e-01
-3.66684832e-02 -3.95546228e-01 -5.08566678e-01 3.69120121e-01
-1.79888427e+00 -6.23805165e-01 2.46806771e-01 2.11203933e+00
1.05494809e+00 -4.92634363e-02 -5.36824167e-01 -4.36561465e-01
7.36561120e-01 1.58928260e-01 -7.68230438e-01 -1.05034471e-01
-2.33739883e-01 4.50369716e-02 5.63245237e-01 6.12789154e-01
-8.17471921e-01 1.24563408e+00 5.87781858e+00 9.26650047e-01
-1.34552443e+00 6.04950264e-02 7.77179658e-01 -2.20627055e-01
-5.39216220e-01 1.69756725e-01 -4.34135735e-01 4.85144615e-01
1.65526271e-01 -6.07229732e-02 8.83190572e-01 5.33165038e-01
3.50308716e-01 -1.20924003e-01 -1.03279030e+00 1.08298624e+00
1.62287340e-01 -1.40661645e+00 6.13183320e-01 1.54715210e-01
1.22047448e+00 -2.94103861e-01 2.57502377e-01 1.46484658e-01
4.15326536e-01 -9.63022113e-01 8.97213161e-01 7.23569512e-01
1.24685013e+00 -6.39166176e-01 3.39456558e-01 8.22817609e-02
-9.57041025e-01 1.70120046e-01 -4.92585182e-01 3.47978354e-01
2.17201844e-01 7.48243749e-01 -2.98920155e-01 4.29400116e-01
6.71486199e-01 6.76414549e-01 -6.69225752e-01 6.68354511e-01
-4.16354060e-01 2.34968781e-01 -2.47317746e-01 4.64763284e-01
1.48859665e-01 -4.89866644e-01 3.73056740e-01 9.24292147e-01
4.63881135e-01 8.40186626e-02 1.84097558e-01 1.33430779e+00
-4.61756676e-01 7.73290230e-04 -6.37760162e-01 -1.43027604e-01
3.32250953e-01 1.34650636e+00 -6.47140324e-01 -3.77233356e-01
-4.27962780e-01 1.65968692e+00 4.40206736e-01 7.26860285e-01
-8.71237874e-01 -4.66749758e-01 6.08263314e-01 1.70106634e-01
2.79972792e-01 -9.19051245e-02 -2.55117357e-01 -1.31831563e+00
1.27280563e-01 -1.08206499e+00 -1.34294778e-01 -1.02928305e+00
-1.18619573e+00 5.58411539e-01 -2.05238774e-01 -1.17358351e+00
2.36831065e-02 -3.13613594e-01 -6.40264094e-01 8.91876936e-01
-1.55870974e+00 -1.29712617e+00 -6.02568626e-01 7.09573209e-01
5.06499827e-01 -1.06750205e-01 5.34392536e-01 5.41410148e-01
-4.57410187e-01 8.68410289e-01 4.99949977e-02 1.81782037e-01
1.03501940e+00 -1.13231158e+00 4.49212283e-01 8.97066653e-01
2.98183435e-03 7.62066543e-01 3.43620986e-01 -7.79552698e-01
-1.36451447e+00 -1.50370193e+00 3.25679064e-01 -3.24216306e-01
3.08685213e-01 -4.63226020e-01 -8.60951424e-01 5.69785476e-01
2.64942467e-01 1.17140591e-01 1.69261441e-01 -3.53354096e-01
-3.65767270e-01 -1.79159999e-01 -1.01585984e+00 1.05917501e+00
1.30042303e+00 -6.67079151e-01 2.73072929e-03 2.02796474e-01
8.33530426e-01 -5.54278016e-01 -7.52321661e-01 1.67242020e-01
4.62503135e-01 -7.55813003e-01 9.30611253e-01 -3.16114277e-01
6.33728087e-01 -5.47373831e-01 -1.04937181e-02 -1.59701097e+00
-3.41817856e-01 -6.60140097e-01 2.80960679e-01 1.50710106e+00
6.40580952e-01 -5.28623760e-01 6.09302998e-01 8.28325391e-01
5.30892275e-02 -4.98104244e-01 -4.21143711e-01 -5.43713570e-01
-1.42738735e-02 -1.44686192e-01 7.85660744e-01 1.00125575e+00
-6.18625462e-01 2.13406980e-01 -7.05869079e-01 4.43103835e-02
6.17014587e-01 1.21335723e-01 8.90344739e-01 -7.03128517e-01
-1.56171724e-01 -2.58903980e-01 9.46803857e-03 -1.20418608e+00
-2.34230049e-02 -8.20663691e-01 2.14020655e-01 -1.39913261e+00
3.35224956e-01 -6.87953651e-01 -1.34938881e-01 6.68493629e-01
-2.38307893e-01 5.09785771e-01 5.15170395e-01 1.36621997e-01
-2.80314893e-01 7.58310020e-01 1.85563219e+00 -4.53483880e-01
-1.41878054e-01 -4.82488573e-01 -9.31893110e-01 6.54438734e-01
6.55920744e-01 -4.78197575e-01 -6.78119898e-01 -9.27043915e-01
3.00925434e-01 -1.43889859e-01 3.60830456e-01 -8.06743085e-01
3.40923816e-01 -3.80119354e-01 5.45292854e-01 -1.47617087e-01
2.03520253e-01 -8.60356629e-01 3.46518308e-01 1.19141378e-01
-3.63439649e-01 -1.37243653e-02 7.96834826e-02 6.22406423e-01
-3.43785435e-01 -1.41032875e-01 8.80902827e-01 -1.65665001e-01
-6.35403752e-01 5.63861549e-01 -1.18037589e-01 -1.17946649e-02
9.36157405e-01 -4.03486371e-01 -3.60472769e-01 -6.38713837e-01
-5.38056433e-01 1.74843192e-01 1.04639494e+00 7.02588260e-01
7.51361668e-01 -1.38202083e+00 -6.39546752e-01 5.02658963e-01
2.87506312e-01 1.24189749e-01 2.86155611e-01 5.92703998e-01
-5.15733957e-01 -1.76877156e-03 -3.61468494e-01 -3.65299255e-01
-9.67140138e-01 5.69685340e-01 3.66787851e-01 -1.94836140e-01
-6.87782586e-01 5.63482583e-01 8.63523304e-01 -5.16893923e-01
3.90815511e-02 -3.62856835e-01 3.19434673e-01 -2.38889635e-01
4.07609791e-01 -7.10700303e-02 -1.07863247e-01 -3.75896573e-01
6.15356565e-02 8.28242302e-01 -4.60070133e-01 -2.69275635e-01
1.17112958e+00 -3.35973769e-01 -1.28786296e-01 -2.20665157e-01
1.19647336e+00 2.53202409e-01 -1.76328826e+00 -3.80610615e-01
-4.44796622e-01 -6.55796230e-01 3.03612351e-01 -1.06016684e+00
-1.64287055e+00 8.13847840e-01 5.42900622e-01 -4.03008521e-01
1.29404366e+00 -4.81932461e-01 8.15020561e-01 3.01652730e-01
3.99050564e-01 -1.34648824e+00 3.23810935e-01 3.80330384e-01
1.23917127e+00 -1.28902590e+00 8.39165747e-02 -5.66269696e-01
-8.53632271e-01 8.83324742e-01 7.67830133e-01 7.42384195e-02
4.52683985e-01 2.82196701e-01 1.08175330e-01 -2.32141297e-02
-5.71806967e-01 7.32040331e-02 4.19665992e-01 6.84140205e-01
2.34935388e-01 -8.51689875e-02 -5.77686131e-02 1.71805173e-01
-4.79245335e-02 -1.20237079e-02 5.44423699e-01 8.16727817e-01
2.47021858e-02 -1.14465666e+00 -3.34266722e-01 4.72121328e-01
-2.31463119e-01 -3.53062391e-01 -4.40385133e-01 6.53421819e-01
1.10172510e-01 7.36596942e-01 -1.12193018e-01 -3.07776809e-01
4.24304634e-01 -3.51938695e-01 3.66726428e-01 -6.53259218e-01
-3.19164813e-01 4.87630889e-02 -1.22175880e-01 -7.04308629e-01
-1.96878538e-01 -2.43730992e-01 -1.11554778e+00 -4.13295716e-01
-2.44446889e-01 -2.04583481e-01 5.96268773e-01 6.36908472e-01
6.54803455e-01 6.01109087e-01 4.92588848e-01 -9.03670728e-01
-1.76069230e-01 -6.80861354e-01 -4.51687485e-01 7.43172765e-01
2.69979894e-01 -5.93284547e-01 -1.53698310e-01 4.06302243e-01] | [11.500995635986328, -0.5424659848213196] |
e3103f65-6498-4677-9974-aa099bae53d3 | the-change-that-matters-in-discourse-parsing | null | null | https://openreview.net/forum?id=KkF3IHfdT_z | https://openreview.net/pdf?id=KkF3IHfdT_z | The Change that Matters in Discourse Parsing: Estimating the Impact of Domain Shift on Parser Error | Discourse analysis allows us to attain high-level inferences of a text document beyond the sentence-level. However, currently the performance of discourse models is very low on texts outside of the training distribution's coverage. There is need for a measure that can inform us to what extent our model generalizes from the training to the test sample when these samples may be drawn from distinct distributions. While this can be estimated via distribution shift, we argue that this does not directly correlate with change in the observed error of a classifier (i.e. error-gap). Thus, we propose to use a statistic from the theoretical domain adaptation literature which can be directly tied to error-gap. We study the bias of this statistic as an estimator of error-gap both theoretically and through a large-scale empirical study of over 2400 experiments on 6 discourse datasets from domains including, but not limited to: news, biomedical texts, TED talks, Reddit posts, and fiction. Our results not only motivate our proposal and help us to understand its limitations, but also provide insight on the properties of discourse models and datasets which improve performance in domain adaptation. For instance, we find that non-news datasets are slightly easier to transfer to than news datasets when the training and test sets are very different. We plan to release our code as a Python package to allow practitioners to make more informed model and dataset choices. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['discourse-parsing'] | ['natural-language-processing'] | [ 3.28193218e-01 4.72629189e-01 -5.32360673e-01 -5.89489162e-01
-1.04668629e+00 -7.91687250e-01 8.99107933e-01 3.93676907e-01
-4.97335583e-01 1.17455435e+00 6.85551643e-01 -4.91433173e-01
-1.16866045e-01 -6.22517526e-01 -7.70868719e-01 -5.19160688e-01
1.94729894e-01 6.85160041e-01 3.62370253e-01 -2.36134589e-01
4.19827670e-01 1.61550138e-02 -1.20126402e+00 4.56645310e-01
8.51962686e-01 3.15163285e-01 -8.60151276e-02 6.77645087e-01
-6.10639621e-03 9.16925788e-01 -1.06321561e+00 -4.21226174e-01
-2.18499720e-01 -8.74634564e-01 -1.13770151e+00 -3.48628201e-02
3.05574954e-01 -1.21976241e-01 3.17057259e-02 1.00493407e+00
5.10935128e-01 1.09965816e-01 1.05654740e+00 -8.88591766e-01
-6.35933757e-01 9.81716394e-01 -2.35121474e-01 5.26235521e-01
4.40450251e-01 -6.19173497e-02 8.67024541e-01 -4.29607809e-01
1.11140859e+00 1.30444741e+00 7.29702055e-01 5.02696574e-01
-1.14403498e+00 -2.93596596e-01 6.31429330e-02 1.98280409e-01
-8.34167838e-01 -6.33913636e-01 5.30397594e-01 -6.02908969e-01
5.85473776e-01 3.66384417e-01 3.95788789e-01 1.54494476e+00
1.53839728e-02 6.58510327e-01 1.29989266e+00 -6.04360342e-01
3.71622562e-01 5.94135940e-01 3.51547569e-01 1.69261873e-01
3.23393643e-01 -2.24622384e-01 -5.33157051e-01 -3.35478127e-01
3.16351026e-01 -7.98387766e-01 -6.22067750e-01 -7.55286366e-02
-9.42048073e-01 1.07933557e+00 3.06951869e-02 5.59105456e-01
-5.69542944e-02 -2.04739735e-01 6.21124029e-01 6.01938963e-01
7.37853169e-01 6.42445266e-01 -5.51627576e-01 -4.70961273e-01
-7.05797374e-01 4.72275913e-01 1.21162868e+00 7.37983525e-01
2.84983605e-01 -4.53760475e-01 -1.65043801e-01 9.65193152e-01
3.58378850e-02 2.41465807e-01 6.28643930e-01 -1.02511179e+00
4.95090455e-01 2.40128487e-01 1.75248355e-01 -9.81004059e-01
-5.54934025e-01 -1.48966417e-01 -2.29513109e-01 -2.19873458e-01
1.03375065e+00 -4.60603088e-01 -2.99203038e-01 2.05446720e+00
2.52728999e-01 -3.31460267e-01 1.40044957e-01 7.52851725e-01
7.68323720e-01 4.70047265e-01 -1.13103278e-02 -5.57921529e-01
1.39560521e+00 -5.42604506e-01 -7.55392611e-01 -3.14142734e-01
9.55465913e-01 -7.51636803e-01 1.31792116e+00 2.74422288e-01
-9.82861102e-01 -1.96023375e-01 -1.08667183e+00 -1.27180904e-01
-6.47279471e-02 -1.83622777e-01 3.48458022e-01 8.22663784e-01
-4.88623410e-01 6.70760095e-01 -8.06314766e-01 -6.84768617e-01
3.21754545e-01 -9.54558924e-02 -5.62184528e-02 1.69546545e-01
-1.23553228e+00 1.32863927e+00 4.95526195e-01 -6.00690722e-01
-2.07618102e-01 -7.39913285e-01 -4.64249194e-01 -1.20698757e-01
4.07060862e-01 -5.32601058e-01 1.39920104e+00 -1.22886753e+00
-1.52715707e+00 9.58554447e-01 -1.70576811e-01 -5.56028128e-01
6.88968837e-01 5.49847633e-02 -1.76078290e-01 1.02734650e-02
6.80709258e-02 9.62514728e-02 3.67027521e-01 -1.16705942e+00
-4.63110834e-01 -2.40544572e-01 1.45127550e-01 1.66081011e-01
-3.06848466e-01 8.91070813e-02 3.29206474e-02 -4.80755925e-01
-1.94533348e-01 -8.41755986e-01 2.53399342e-01 -2.26346970e-01
-3.28371078e-01 -2.34100536e-01 5.00651419e-01 -6.88349664e-01
1.29692316e+00 -2.01848245e+00 -5.68070039e-02 -9.73847322e-03
-2.99008433e-02 1.12692855e-01 1.68315351e-01 5.19164383e-01
7.74834976e-02 2.02312648e-01 -3.60909820e-01 -8.21894687e-03
-1.24333374e-01 6.27750233e-02 -2.93126851e-01 6.26131535e-01
1.36655718e-01 6.12237275e-01 -9.37308371e-01 -5.06020367e-01
-1.26027271e-01 9.31443498e-02 -6.05261862e-01 -1.19901478e-01
-3.39773238e-01 4.60894912e-01 -4.48491603e-01 5.63906729e-02
3.02521586e-01 -4.70322520e-01 4.65309560e-01 1.05829582e-01
7.88628832e-02 8.71451676e-01 -8.04279029e-01 1.32106042e+00
-2.00495884e-01 1.15127575e+00 -2.55051702e-01 -1.25829422e+00
8.67491364e-01 1.62012652e-01 8.93513709e-02 -4.57080692e-01
3.02421540e-01 1.18891820e-01 5.11683822e-01 -6.86659515e-01
5.84472358e-01 -5.14846623e-01 -1.51207773e-02 6.22233152e-01
-4.49949466e-02 -7.16609880e-02 2.56407619e-01 -6.78400174e-02
1.09939921e+00 -2.23312214e-01 4.48614061e-01 -5.96778870e-01
1.61737204e-01 2.52242595e-01 2.94123888e-01 8.54811847e-01
-1.84927136e-02 5.36888719e-01 1.02194440e+00 6.57502264e-02
-1.21556306e+00 -9.52749312e-01 -8.26215684e-01 1.18671429e+00
-6.66792914e-02 -2.47079402e-01 -8.22506249e-01 -7.23472297e-01
-5.41091636e-02 1.21237993e+00 -7.40869284e-01 -1.73392877e-01
-6.16744697e-01 -1.18763769e+00 6.73256278e-01 4.11095172e-01
2.76427299e-01 -7.68835545e-01 -6.10738456e-01 8.77048895e-02
-3.29095781e-01 -9.44697618e-01 -3.23984683e-01 1.17856734e-01
-7.84523845e-01 -1.21012080e+00 -5.18914342e-01 -4.53606546e-01
4.01853621e-01 -2.05864593e-01 1.14272892e+00 -5.58869354e-02
4.16920483e-01 3.81296098e-01 -6.07922018e-01 -6.26822531e-01
-1.10871649e+00 4.34191972e-01 -1.88971177e-01 -4.17104989e-01
5.90165436e-01 -4.24206406e-01 -2.64422327e-01 2.82439262e-01
-7.64986098e-01 -1.69001445e-01 1.77743807e-01 1.07160389e+00
3.07577532e-02 -3.34990293e-01 9.71924901e-01 -1.40007114e+00
1.03237844e+00 -7.51925290e-01 -1.60682097e-01 3.07275355e-01
-6.64738595e-01 4.91831265e-02 4.84776676e-01 -7.36692846e-01
-1.13453770e+00 -6.34219885e-01 -1.25889957e-01 3.64881873e-01
-2.44426988e-02 7.97356546e-01 6.76118731e-02 4.92414713e-01
1.26221251e+00 -1.51590317e-01 1.98280916e-01 -3.42423677e-01
1.49237886e-01 1.00632870e+00 3.18427891e-01 -7.51642764e-01
2.78915793e-01 2.39836022e-01 -4.22538072e-01 -1.09655547e+00
-1.02432930e+00 -5.38685210e-02 -3.22106779e-01 -7.11430088e-02
5.52864075e-01 -6.43479764e-01 -5.15531600e-01 -5.51898293e-02
-1.01288581e+00 -7.09271610e-01 -4.75935161e-01 6.06234610e-01
-6.09277487e-01 3.30193192e-01 -5.19422174e-01 -7.99852073e-01
9.05203670e-02 -7.86223590e-01 6.30875647e-01 1.82487994e-01
-9.15567040e-01 -1.49022913e+00 7.30529651e-02 3.06647807e-01
3.23001027e-01 3.58451217e-01 1.10883820e+00 -1.25529087e+00
-2.30787452e-02 -4.66335053e-03 4.01585363e-02 3.32576782e-01
2.32522860e-01 1.27924666e-01 -1.00394917e+00 -3.39249969e-01
2.80541450e-01 -4.63303208e-01 7.27826357e-01 5.88990748e-01
9.05322492e-01 -3.38347971e-01 -3.86642367e-01 4.03199084e-02
9.91142154e-01 1.64294258e-01 6.08669758e-01 6.30082011e-01
1.46790415e-01 7.51381218e-01 6.10224426e-01 2.55182922e-01
3.81903112e-01 6.49291992e-01 -3.09446245e-01 3.24106157e-01
-1.00859478e-01 -1.49835214e-01 4.19232756e-01 8.51480901e-01
-3.64823081e-02 -5.13339937e-01 -1.00082350e+00 5.28759718e-01
-1.64920557e+00 -1.08372259e+00 -1.03963137e-01 2.20270228e+00
1.40293062e+00 4.93577331e-01 3.94622684e-01 4.08166228e-03
6.67946100e-01 8.15015584e-02 -5.01627982e-01 -4.31480378e-01
-2.64400601e-01 3.98746552e-03 3.69947493e-01 7.88109899e-01
-8.06446731e-01 5.45729160e-01 6.89333391e+00 8.47236633e-01
-1.06814051e+00 2.00800389e-01 8.10320854e-01 -2.61313822e-02
-4.86880779e-01 2.41095703e-02 -5.25373459e-01 5.72082520e-01
1.33307624e+00 -6.67859316e-01 8.62133056e-02 7.37073898e-01
3.13259251e-02 -3.57297391e-01 -1.52330720e+00 3.70799184e-01
1.48987204e-01 -1.36371768e+00 -3.38567436e-01 5.86572029e-02
6.92294359e-01 -2.36060638e-02 -1.54416010e-01 3.35317165e-01
3.44646335e-01 -1.05576944e+00 6.46892369e-01 4.00664419e-01
6.62377059e-01 -4.70774293e-01 8.51162314e-01 7.31332123e-01
-8.98518115e-02 5.24822809e-03 -3.11695069e-01 -2.04114333e-01
2.90097035e-02 5.77359259e-01 -1.29744458e+00 2.89620012e-01
3.67777407e-01 4.70359772e-01 -4.03746963e-01 6.26349151e-01
6.66560009e-02 1.01223385e+00 -2.89936125e-01 -4.37372714e-01
-2.13918403e-01 1.04028143e-01 6.24767959e-01 1.39652681e+00
2.09582999e-01 1.29602239e-01 -2.31127679e-01 7.47990608e-01
5.45675680e-03 1.67490870e-01 -5.39791465e-01 -6.49887994e-02
8.70754004e-01 6.15860581e-01 -6.65089726e-01 -3.80332977e-01
-4.97742355e-01 3.84805053e-01 3.31378371e-01 2.17964023e-01
-5.92247546e-01 -2.79107124e-01 4.64686036e-01 2.61991829e-01
6.30913824e-02 -4.91437353e-02 -5.42551219e-01 -1.09654355e+00
9.61102545e-03 -1.17943513e+00 2.78454363e-01 -4.29262131e-01
-1.59906971e+00 3.14078510e-01 5.40840745e-01 -9.39921558e-01
-6.11423016e-01 -6.49029493e-01 -3.28593701e-01 7.39188075e-01
-1.01887703e+00 -4.39090163e-01 3.59094655e-03 6.93848655e-02
5.09724736e-01 -9.88931432e-02 7.38590181e-01 2.83935368e-02
-2.23035350e-01 7.47168779e-01 4.56548691e-01 3.55677128e-01
1.14476776e+00 -1.16966009e+00 2.00165495e-01 3.21227103e-01
-4.81735803e-02 6.45761073e-01 1.29659986e+00 -5.04446208e-01
-8.08728337e-01 -6.47795975e-01 7.09773064e-01 -1.05036020e+00
7.75260985e-01 -9.35842320e-02 -1.18715572e+00 8.62998962e-01
2.79315501e-01 -5.75631618e-01 8.18774462e-01 6.19137704e-01
-1.90044165e-01 2.21982256e-01 -1.13087261e+00 5.01014411e-01
9.60048795e-01 -4.47014362e-01 -9.98390317e-01 4.50183213e-01
5.65098822e-01 -7.14156270e-01 -1.15817463e+00 2.47678399e-01
5.02861619e-01 -1.04084742e+00 6.21281624e-01 -8.04984510e-01
7.22628713e-01 1.59277603e-01 -3.20755005e-01 -1.50879574e+00
-3.46878022e-02 -2.44360819e-01 2.61713024e-02 1.42969441e+00
7.34959304e-01 -8.65551949e-01 5.98964334e-01 7.14687228e-01
-5.85939921e-02 -6.15584493e-01 -1.02268302e+00 -7.71556377e-01
8.73006105e-01 -4.71172065e-01 4.56222802e-01 1.33263767e+00
4.63630974e-01 6.22640252e-01 6.35899082e-02 -1.86562225e-01
2.07413062e-01 -1.50988251e-01 7.51686454e-01 -1.23361981e+00
-5.46994090e-01 -5.05840123e-01 -1.69919536e-01 -1.20329392e+00
2.71977097e-01 -8.49572122e-01 -3.02239414e-02 -1.06848884e+00
4.16878611e-01 -4.05040562e-01 3.13710123e-01 1.05543010e-01
-2.45041087e-01 -2.28235304e-01 1.87484380e-02 1.84753522e-01
-2.18989626e-01 3.99218589e-01 1.29904115e+00 3.03783286e-02
-3.48418444e-01 -4.61006239e-02 -9.78172064e-01 8.67594421e-01
9.56986845e-01 -5.75760186e-01 -5.02735674e-01 -2.67738521e-01
1.98522478e-01 1.99038580e-01 2.31875524e-01 -6.37305081e-01
-8.10458064e-02 -3.73180360e-01 2.92779386e-01 -8.55702907e-02
9.08326879e-02 -2.83033282e-01 -1.01979241e-01 2.44853795e-01
-9.60347831e-01 -1.86858088e-01 1.93917945e-01 4.88874137e-01
3.19566168e-02 -5.89976013e-01 7.16283381e-01 -8.25138465e-02
-4.87473644e-02 -4.42455113e-01 -4.38778818e-01 7.04810560e-01
6.86820507e-01 -2.33661160e-02 -9.70168591e-01 -6.19387269e-01
-6.32890582e-01 4.95176874e-02 6.80727661e-01 3.84356797e-01
-1.15911096e-01 -1.03689516e+00 -1.01961541e+00 -2.91867703e-01
8.42746943e-02 -2.02476993e-01 -1.06523082e-01 1.01040232e+00
-2.96953708e-01 3.95220816e-01 1.56255692e-01 -6.54836297e-01
-1.23922670e+00 2.57996619e-01 2.21496344e-01 -1.50879279e-01
-5.27137756e-01 5.37406445e-01 -4.06677648e-02 -3.89172912e-01
3.53713892e-02 -3.60722780e-01 -1.20854378e-01 2.72451222e-01
5.67848623e-01 3.20307016e-01 -3.78661044e-02 -3.92839372e-01
-2.06676289e-01 1.33136168e-01 -1.63062081e-01 -3.33422333e-01
1.16207254e+00 -2.33081073e-01 6.37663081e-02 1.04520035e+00
1.02567112e+00 2.26773798e-01 -1.08033502e+00 -2.58697122e-01
2.15795889e-01 -4.03213382e-01 -3.37853879e-01 -8.39474857e-01
-3.02458376e-01 5.78415871e-01 2.77555525e-01 7.05601692e-01
8.02611470e-01 4.24092740e-01 2.54540861e-01 4.62926745e-01
2.19468996e-01 -1.31485236e+00 6.81912452e-02 4.90266293e-01
1.10033524e+00 -1.23928607e+00 2.87062049e-01 -2.70513356e-01
-7.83569396e-01 1.10113680e+00 3.38491797e-01 -2.84956060e-02
5.86140752e-01 2.53192216e-01 1.01937771e-01 -2.31429022e-02
-9.66521382e-01 1.73785597e-01 1.11382753e-01 7.55026460e-01
9.34598684e-01 3.55727412e-02 -7.32651889e-01 6.45777941e-01
-7.57900000e-01 -6.68592975e-02 8.56964529e-01 6.08558357e-01
-5.34187675e-01 -1.07464659e+00 -4.21947956e-01 6.96730196e-01
-4.67469662e-01 1.48615584e-01 -5.72172344e-01 1.23491096e+00
-1.83935300e-01 1.02194774e+00 1.60732552e-01 -1.98827893e-01
2.76516795e-01 3.70478064e-01 6.95253015e-01 -6.74800694e-01
-2.62755960e-01 -2.57247716e-01 8.53645325e-01 1.02600783e-01
-8.29407990e-01 -1.16873240e+00 -1.10348678e+00 -6.84102714e-01
-4.63692397e-01 2.98732340e-01 3.00729483e-01 1.19983578e+00
3.06969360e-02 4.52210218e-01 1.99910566e-01 -4.60153431e-01
-7.86670804e-01 -1.42307806e+00 -3.71035099e-01 5.49626350e-01
3.09246361e-01 -6.43076777e-01 -4.88721490e-01 1.67012379e-01] | [10.941591262817383, 9.054040908813477] |
72115840-a083-4d9c-b22a-66e95bd7db58 | deep-state-space-models-for-time-series | null | null | http://papers.nips.cc/paper/8004-deep-state-space-models-for-time-series-forecasting | http://papers.nips.cc/paper/8004-deep-state-space-models-for-time-series-forecasting.pdf | Deep State Space Models for Time Series Forecasting | We present a novel approach to probabilistic time series forecasting that combines state space models with deep learning. By parametrizing a per-time-series linear state space model with a jointly-learned recurrent neural network, our method retains desired properties of state space models such as data efficiency and interpretability, while making use of the ability to learn complex patterns from raw data offered by deep learning approaches. Our method scales gracefully from regimes where little training data is available to regimes where data from millions of time series can be leveraged to learn accurate models. We provide qualitative as well as quantitative results with the proposed method, showing that it compares favorably to the state-of-the-art. | ['Syama Sundar Rangapuram', 'Matthias W. Seeger', 'Jan Gasthaus', 'Tim Januschowski', 'Yuyang Wang', 'Lorenzo Stella'] | 2018-12-01 | null | null | null | neurips-2018-12 | ['probabilistic-time-series-forecasting'] | ['time-series'] | [-1.03218660e-01 -2.90269926e-02 -4.66739476e-01 -4.25735831e-01
-8.80665541e-01 -6.66301250e-01 1.17698526e+00 1.55786425e-02
-7.40885958e-02 6.83311701e-01 4.64344531e-01 -9.29918349e-01
-2.65847117e-01 -6.43347085e-01 -7.94827998e-01 -6.72819912e-01
-6.58785820e-01 4.13167089e-01 -1.39198795e-01 -3.44597548e-01
1.58650801e-02 7.25773036e-01 -1.26920962e+00 6.27018735e-02
4.22437161e-01 1.45363879e+00 -3.09129894e-01 6.55839086e-01
-2.37023741e-01 9.96826589e-01 -3.57078850e-01 1.47739664e-01
3.72574896e-01 7.56390169e-02 -3.44437540e-01 -8.84922370e-02
-1.42589524e-01 -4.63539958e-01 -8.79873037e-01 4.59638685e-01
8.47599134e-02 1.34065285e-01 7.35170484e-01 -1.08692706e+00
-6.09372556e-01 5.51733434e-01 -1.30492121e-01 3.46352309e-01
-1.38333216e-01 3.99048895e-01 9.55780268e-01 -3.56754571e-01
2.44376346e-01 1.26958597e+00 1.04433155e+00 2.16656744e-01
-1.42974019e+00 -5.58691144e-01 4.15149182e-01 -1.83889821e-01
-8.30787182e-01 -5.34641385e-01 1.09009564e+00 -5.36875308e-01
1.20233226e+00 -1.63601618e-02 6.12885356e-01 1.28772187e+00
6.91596746e-01 7.76520669e-01 1.16048336e+00 -1.21926151e-01
4.63263214e-01 -3.30410182e-01 3.35563421e-01 4.81122255e-01
-5.31785674e-02 8.38349700e-01 -2.79449582e-01 -3.86068225e-01
9.19073999e-01 7.59934127e-01 3.53939086e-01 -1.93101496e-01
-1.09881067e+00 7.80673683e-01 3.10500383e-01 2.85736144e-01
-7.43060887e-01 6.60288751e-01 5.56144774e-01 5.55220306e-01
8.70832205e-01 3.68352205e-01 -7.29457617e-01 -3.17563772e-01
-1.14466238e+00 3.08567375e-01 9.04721498e-01 6.66124880e-01
4.12481159e-01 8.03045154e-01 -5.80288805e-02 2.50170797e-01
2.44066522e-01 8.04468870e-01 6.54313385e-01 -9.99002159e-01
1.90738410e-01 2.92064846e-01 4.93634343e-01 -7.08345354e-01
-5.36437571e-01 -5.00042617e-01 -1.14762270e+00 1.46794179e-02
2.90883571e-01 -4.81539816e-01 -1.07117558e+00 1.68639588e+00
-1.92286745e-01 5.01874506e-01 3.26182455e-01 3.39019299e-01
1.60460040e-01 1.28026080e+00 -1.65906250e-01 -5.03272176e-01
7.78809011e-01 -6.35657549e-01 -6.45353556e-01 -1.84656739e-01
3.64152759e-01 -2.52295077e-01 6.11298382e-01 6.38514459e-02
-1.00534129e+00 -4.92971629e-01 -8.47239435e-01 2.19684854e-01
-4.00543034e-01 -1.49511322e-01 8.68893921e-01 3.98747683e-01
-1.20174003e+00 1.03714836e+00 -1.55932224e+00 -1.24716684e-01
1.95184246e-01 3.31694692e-01 1.11907637e-02 4.15382713e-01
-1.28374898e+00 9.77196336e-01 4.40870285e-01 1.90001920e-01
-1.24725688e+00 -6.75054073e-01 -7.09753931e-01 3.35065663e-01
4.01514843e-02 -5.35757482e-01 1.70018959e+00 -6.79186940e-01
-2.03252602e+00 -1.85904186e-02 -2.47014657e-01 -9.95649576e-01
2.30476275e-01 -1.79297552e-01 -7.51042724e-01 -1.66727677e-01
-4.25574660e-01 1.32828847e-01 8.25657606e-01 -8.25451314e-01
-5.02200902e-01 2.94679007e-03 -3.93511914e-02 -4.43323821e-01
-1.90907896e-01 -5.88680245e-02 3.17637742e-01 -6.16513669e-01
1.22750320e-01 -9.43090737e-01 -6.44359469e-01 -1.95722178e-01
-1.91370577e-01 -4.17782754e-01 1.06573987e+00 -7.37596154e-01
1.34291637e+00 -1.75825286e+00 -2.32158720e-01 2.21272796e-01
1.56903356e-01 1.22812919e-01 -1.57601029e-01 1.07462645e+00
-2.75923252e-01 9.98473093e-02 -9.06132460e-02 -5.86301386e-01
2.01573104e-01 4.44146842e-01 -1.15772259e+00 5.48753560e-01
4.67137277e-01 1.15825713e+00 -8.12311709e-01 1.19795658e-01
4.64496970e-01 3.28636646e-01 -9.50308293e-02 3.20262820e-01
-5.47244310e-01 3.33805412e-01 -5.24041355e-01 4.91559684e-01
6.60373718e-02 -8.44271898e-01 2.01240882e-01 2.54581481e-01
-3.04271132e-01 7.42700458e-01 -8.53191674e-01 1.22520649e+00
-5.94848812e-01 6.87468231e-01 -4.32058126e-01 -1.37768555e+00
8.27399194e-01 6.42226875e-01 7.33802199e-01 -4.83681649e-01
-3.14197280e-02 1.76409990e-01 -3.83635104e-01 -1.64335921e-01
1.95233569e-01 -3.74001890e-01 -3.94610733e-01 8.36879373e-01
1.35958344e-01 -1.13095157e-01 -1.80718124e-01 -1.60147876e-01
1.11107445e+00 1.54938623e-01 3.83365422e-01 -2.71759033e-01
-1.47469342e-01 -7.40071852e-03 4.71578807e-01 8.81003022e-01
4.50293496e-02 -1.16546415e-01 5.16891420e-01 -1.04894555e+00
-1.58176708e+00 -1.18287194e+00 -1.07869782e-01 1.06942868e+00
-5.78126192e-01 -3.06976199e-01 -4.37001288e-02 -2.50131339e-01
5.79077229e-02 7.44321525e-01 -7.67093480e-01 -1.33613586e-01
-4.51847792e-01 -8.04669499e-01 3.30564290e-01 8.44525874e-01
3.37672271e-02 -9.95372295e-01 -4.07697886e-01 6.73829675e-01
4.13611829e-01 -6.15944743e-01 -8.78366306e-02 3.53189528e-01
-1.27094543e+00 -3.66258293e-01 -5.26555121e-01 -3.03599715e-01
1.29599869e-01 2.66933697e-03 1.20658624e+00 -5.39009154e-01
2.22170353e-01 2.12786779e-01 1.55750155e-01 -5.39732218e-01
-7.40814626e-01 1.33654371e-01 3.96560699e-01 -1.44404411e-01
9.69584957e-02 -1.05877137e+00 -5.34624934e-01 -1.64036274e-01
-9.88935411e-01 1.14829317e-01 6.08623505e-01 1.04040909e+00
3.90349329e-01 9.99571290e-03 8.81258011e-01 -6.08242869e-01
5.57493806e-01 -7.55462348e-01 -1.02050281e+00 2.02521488e-01
-8.87631834e-01 4.19294715e-01 8.06864679e-01 -5.43081522e-01
-8.04527044e-01 1.46964416e-01 -7.96764251e-03 -7.19558477e-01
-5.33929765e-02 7.72377133e-01 6.30590320e-01 3.04509223e-01
4.68664765e-01 6.89810216e-01 3.37683588e-01 -4.44276810e-01
6.38787687e-01 6.37428403e-01 6.36066020e-01 -5.24351537e-01
6.72002494e-01 5.45965075e-01 8.33090693e-02 -4.73706961e-01
-1.04805899e+00 -2.73779511e-01 -4.72131073e-01 5.32473736e-02
2.90554225e-01 -1.03732741e+00 -7.63920426e-01 5.16962826e-01
-1.04241741e+00 -4.47211057e-01 -5.13723433e-01 3.97612423e-01
-8.70782793e-01 -1.72886893e-01 -1.09515524e+00 -1.33151948e+00
-2.83837736e-01 -5.82540393e-01 1.19025958e+00 -6.85228556e-02
-1.51395917e-01 -1.50054121e+00 4.45958734e-01 -6.42961562e-01
7.61093318e-01 4.91654485e-01 9.36054349e-01 -8.28197479e-01
-4.23464864e-01 -6.89420581e-01 4.86912765e-02 2.71344095e-01
2.37201229e-01 1.10456415e-01 -9.52007711e-01 -3.77144784e-01
9.98077840e-02 -2.06102625e-01 8.75956595e-01 6.52303040e-01
1.21626639e+00 -8.69811356e-01 -3.28658491e-01 3.25106978e-01
1.26435804e+00 2.67414033e-01 3.08546513e-01 7.81159997e-02
5.39352000e-01 3.19622457e-01 -1.04228586e-01 6.06521785e-01
5.25750458e-01 2.66223669e-01 2.34297633e-01 -4.19624224e-02
4.57726926e-01 -5.10237753e-01 4.30081010e-01 1.02888787e+00
-8.84619355e-02 -1.62517160e-01 -1.15108585e+00 6.13467455e-01
-2.20419717e+00 -1.28686309e+00 3.91480356e-01 2.01379538e+00
7.17356205e-01 3.86338949e-01 3.99839818e-01 7.48047307e-02
3.60671997e-01 5.86571395e-01 -8.43370378e-01 -2.41897523e-01
-5.60651906e-02 5.38972840e-02 5.67153394e-01 3.09719771e-01
-1.26804972e+00 7.41144657e-01 8.53760147e+00 5.72585583e-01
-1.50854659e+00 3.34380940e-02 7.95225143e-01 -1.23859875e-01
-4.28756952e-01 -7.18277618e-02 -6.39951229e-01 5.88596344e-01
1.99568939e+00 -4.61699337e-01 5.61294794e-01 7.38032341e-01
5.07439971e-01 5.68164170e-01 -1.23585820e+00 7.30131745e-01
-4.93899018e-01 -1.81604743e+00 -9.21795219e-02 3.50271836e-02
1.01452553e+00 5.04535675e-01 3.69171530e-01 6.56253815e-01
9.68514204e-01 -1.07165205e+00 6.42476380e-01 1.00782299e+00
4.59006369e-01 -5.28935850e-01 5.68064272e-01 6.02442145e-01
-1.07332289e+00 -5.70166290e-01 -1.57344729e-01 -5.87337255e-01
2.74637699e-01 7.85565376e-01 -7.43969619e-01 3.70368093e-01
4.48841572e-01 1.13377702e+00 -1.77356094e-01 7.15610743e-01
1.96594894e-01 1.19519329e+00 -7.46796846e-01 -3.63763757e-02
5.81229389e-01 -5.34994295e-03 4.96324271e-01 9.72216904e-01
4.93755370e-01 -2.39602719e-02 4.49045807e-01 7.92907238e-01
2.08781451e-01 -5.83052814e-01 -9.33264434e-01 -6.55303538e-01
4.33127850e-01 7.66762137e-01 -3.03408355e-01 -6.81955457e-01
-4.51899678e-01 3.98928404e-01 3.55043054e-01 6.45247400e-01
-4.23889220e-01 1.29122183e-01 6.04785979e-01 -2.77776551e-02
4.17410284e-01 -7.40773618e-01 -4.69741315e-01 -1.20282972e+00
-2.27229446e-01 -5.51664472e-01 3.68106335e-01 -6.71865642e-01
-1.81798184e+00 5.83920836e-01 6.15685470e-02 -1.28096032e+00
-1.24311554e+00 -5.97867131e-01 -8.51977408e-01 8.96919847e-01
-1.33201301e+00 -1.35544705e+00 6.74500525e-01 1.97961599e-01
4.00424451e-01 -3.51718664e-01 9.43734169e-01 -3.20148379e-01
-4.72349256e-01 7.61745200e-02 9.10249531e-01 -8.66640210e-02
-6.23233728e-02 -1.20126426e+00 1.06298172e+00 7.19922602e-01
1.45161271e-01 6.04756355e-01 8.90791833e-01 -4.66020226e-01
-1.66188884e+00 -1.20691752e+00 7.36849546e-01 -5.40294349e-01
1.29674149e+00 -2.87426472e-01 -1.03158724e+00 1.08369672e+00
-2.50773877e-02 2.15234935e-01 4.54006672e-01 5.19068122e-01
-4.43770498e-01 -1.80959374e-01 -7.45428562e-01 4.84333932e-01
4.59007770e-01 -9.03728426e-01 -7.87147462e-01 4.08183098e-01
8.75578105e-01 -2.76191980e-01 -9.72768128e-01 2.23571002e-01
7.23027050e-01 -3.27930242e-01 8.72758269e-01 -1.10493612e+00
3.96543860e-01 7.03589525e-03 -8.91362056e-02 -1.53065205e+00
-5.80124974e-01 -1.00484967e+00 -1.02552569e+00 7.60930479e-01
5.29092014e-01 -1.02191782e+00 4.37783152e-01 8.01407099e-01
-9.91761163e-02 -7.51352906e-01 -9.83549058e-01 -9.89173412e-01
4.93324012e-01 -5.29196858e-01 9.81049895e-01 7.91995227e-01
6.86541051e-02 1.09280333e-01 -7.01049745e-01 3.98109376e-01
4.72108901e-01 5.46550274e-01 4.39634979e-01 -1.34180486e+00
-2.77545154e-01 -5.95084846e-01 -1.82783797e-01 -9.97527003e-01
3.57610345e-01 -4.11775053e-01 -9.06815454e-02 -1.20871842e+00
-1.43727779e-01 -4.03004974e-01 -1.02520931e+00 5.84368765e-01
9.78149474e-02 -1.42674625e-01 5.79337664e-02 5.35360396e-01
-6.18158042e-01 7.10025489e-01 7.50195265e-01 -3.03197280e-02
-2.51483351e-01 3.98655027e-01 -5.53333759e-01 6.09544992e-01
8.92719150e-01 -2.90011674e-01 -3.08740497e-01 -2.17336372e-01
1.22961499e-01 8.04361105e-01 5.70195258e-01 -9.74717915e-01
2.61575937e-01 -5.51981807e-01 5.60807705e-01 -8.27562928e-01
4.10965562e-01 -7.37512290e-01 2.99651593e-01 5.23985624e-01
-6.25638485e-01 3.97388548e-01 4.88690704e-01 9.97717440e-01
-1.00721687e-01 5.14878035e-01 4.29787517e-01 9.44893993e-03
-6.44446969e-01 5.30453384e-01 -6.01114631e-01 -3.59834313e-01
6.09080493e-01 2.19127625e-01 -3.05765837e-01 -8.92802238e-01
-7.10269511e-01 1.65665776e-01 9.70767587e-02 5.73102415e-01
1.56945929e-01 -1.40081251e+00 -6.13455713e-01 1.96671009e-01
1.51429074e-02 -4.52386886e-01 2.33572319e-01 4.17915255e-01
4.37091365e-02 9.10548449e-01 -6.15679100e-02 -7.33164787e-01
-9.45232734e-02 8.16610456e-01 4.21806544e-01 -5.77839494e-01
-8.12473118e-01 1.12828031e-01 -2.30884962e-02 -4.95164067e-01
-1.01905316e-02 -7.64352739e-01 1.31013781e-01 -4.01360422e-01
6.09030426e-01 1.42761171e-01 -1.74025044e-01 -1.61229745e-01
5.84149687e-03 -1.12978900e-02 2.92720515e-02 -2.67651796e-01
1.73622143e+00 -9.82305557e-02 1.75082907e-01 1.33044767e+00
1.14093518e+00 -7.01101601e-01 -1.91781855e+00 -4.82491851e-01
1.11462250e-01 7.65362605e-02 3.40877026e-01 -9.00471628e-01
-7.28189290e-01 9.35526669e-01 7.54664600e-01 7.23484695e-01
8.03390563e-01 -6.66872635e-02 9.30390179e-01 7.47915030e-01
3.96690637e-01 -8.70079517e-01 -2.05193639e-01 8.11882615e-01
6.69931948e-01 -1.21298897e+00 -3.14364344e-01 4.48784739e-01
-3.32713991e-01 9.97537494e-01 -1.50793821e-01 -6.54745281e-01
1.17202842e+00 4.32310283e-01 -6.65772567e-03 -6.39017597e-02
-1.54608858e+00 1.43506184e-01 4.85427856e-01 3.34701568e-01
2.20197171e-01 3.48437369e-01 3.43252510e-01 6.63814425e-01
-2.29154542e-01 2.80192226e-01 2.37756968e-01 9.97982740e-01
-3.61234158e-01 -7.24211574e-01 -2.17572600e-01 7.94343054e-01
-4.51013952e-01 -5.58947250e-02 3.04901630e-01 5.17048240e-01
-7.85976887e-01 8.26652229e-01 4.67230886e-01 -2.38828585e-01
-1.74878184e-02 3.35367352e-01 -2.74144374e-02 -3.13825279e-01
-3.67645293e-01 9.31407288e-02 9.61345900e-03 -6.40720129e-01
-2.32408375e-01 -6.75333560e-01 -8.00523579e-01 -6.24154568e-01
2.97041479e-02 -7.26435706e-02 7.04357326e-01 1.23904026e+00
6.27281666e-01 5.95248759e-01 1.05454981e+00 -1.27418387e+00
-1.13243556e+00 -1.02086210e+00 -7.04128146e-01 1.49584832e-02
9.94557023e-01 -5.30807018e-01 -3.68906915e-01 1.51586026e-01] | [6.946745872497559, 3.2085914611816406] |
361f8aba-ba82-4945-b58e-6df39a411008 | novel-hybrid-dnn-approaches-for-speaker | 2112.13353 | null | https://arxiv.org/abs/2112.13353v1 | https://arxiv.org/pdf/2112.13353v1.pdf | Novel Hybrid DNN Approaches for Speaker Verification in Emotional and Stressful Talking Environments | In this work, we conducted an empirical comparative study of the performance of text-independent speaker verification in emotional and stressful environments. This work combined deep models with shallow architecture, which resulted in novel hybrid classifiers. Four distinct hybrid models were utilized: deep neural network-hidden Markov model (DNN-HMM), deep neural network-Gaussian mixture model (DNN-GMM), Gaussian mixture model-deep neural network (GMM-DNN), and hidden Markov model-deep neural network (HMM-DNN). All models were based on novel implemented architecture. The comparative study used three distinct speech datasets: a private Arabic dataset and two public English databases, namely, Speech Under Simulated and Actual Stress (SUSAS) and Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS). The test results of the aforementioned hybrid models demonstrated that the proposed HMM-DNN leveraged the verification performance in emotional and stressful environments. Results also showed that HMM-DNN outperformed all other hybrid models in terms of equal error rate (EER) and area under the curve (AUC) evaluation metrics. The average resulting verification system based on the three datasets yielded EERs of 7.19%, 16.85%, 11.51%, and 11.90% based on HMM-DNN, DNN-HMM, DNN-GMM, and GMM-DNN, respectively. Furthermore, we found that the DNN-GMM model demonstrated the least computational complexity compared to all other hybrid models in both talking environments. Conversely, the HMM-DNN model required the greatest amount of training time. Findings also demonstrated that EER and AUC values depended on the database when comparing average emotional and stressful performances. | ['Kemal Polat', 'Adi Alhudhaif', 'Ashraf Elnagar', 'Nawel Nemmour', 'Ali Bou Nassif', 'Ismail Shahin'] | 2021-12-26 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [-5.42785048e-01 -8.60000700e-02 2.08875865e-01 -4.01238412e-01
-5.89780867e-01 -2.59933084e-01 3.02386642e-01 -3.52872223e-01
-2.59525269e-01 3.90632361e-01 2.11167604e-01 -3.34174842e-01
2.16551170e-01 -2.47086316e-01 -1.46725461e-01 -9.19857442e-01
-1.17221154e-01 2.53054760e-02 -4.44494903e-01 -2.77234316e-01
2.72536953e-03 5.75724304e-01 -1.62870419e+00 -5.41485548e-02
4.78637964e-01 1.11724710e+00 -5.51720411e-02 8.43043923e-01
2.61135876e-01 5.79182565e-01 -9.92146194e-01 -6.42993450e-01
5.18693104e-02 -1.44653663e-01 -3.76297683e-01 -1.99143216e-01
-4.07976024e-02 -4.62546825e-01 -5.52481532e-01 8.81014764e-01
1.33745611e+00 5.11887312e-01 4.87189710e-01 -1.41074073e+00
-6.93929911e-01 5.70447624e-01 -3.52476120e-01 1.24831237e-01
3.53621364e-01 1.55693322e-01 4.53076690e-01 -8.66357267e-01
2.68152654e-01 1.42838311e+00 9.78393734e-01 7.92798579e-01
-9.03901994e-01 -1.01240325e+00 -1.62686631e-01 3.05723429e-01
-1.45602882e+00 -8.54756117e-01 6.00329697e-01 -3.12735558e-01
1.41029668e+00 1.20909907e-01 4.18058634e-01 1.69795477e+00
2.28831008e-01 5.06240308e-01 1.25552869e+00 -5.06506205e-01
3.38853419e-01 5.05785108e-01 3.07797760e-01 2.96689183e-01
-1.32536098e-01 5.69992244e-01 -6.47344470e-01 -1.83065712e-01
2.52338350e-02 -3.98653686e-01 -1.14564970e-01 7.19402969e-01
-4.14601803e-01 7.79833615e-01 -3.00181717e-01 4.78934646e-01
-4.87156779e-01 -4.40294623e-01 8.12286913e-01 4.31886524e-01
3.19591314e-01 -1.90477163e-01 -5.23686349e-01 -6.03631377e-01
-1.09228921e+00 -1.87601238e-01 1.13702810e+00 7.19320655e-01
2.27962404e-01 8.52127790e-01 2.42705643e-01 1.15814161e+00
5.91022849e-01 5.75061798e-01 9.08824086e-01 -7.39455581e-01
1.35434270e-01 -3.16721946e-02 -2.58403569e-01 -1.00037634e+00
-3.68738174e-01 -5.49633026e-01 -1.05951679e+00 8.30664709e-02
-1.12779468e-01 -5.91725111e-01 -9.30403173e-01 1.92628348e+00
2.09511027e-01 -8.58832151e-02 7.59460092e-01 6.81186616e-01
1.23427081e+00 7.97366083e-01 1.86480910e-01 -1.64780036e-01
1.24543703e+00 -7.52009511e-01 -1.26241171e+00 8.43795091e-02
4.78285432e-01 -8.36819530e-01 8.90653014e-01 5.03810346e-01
-1.03004467e+00 -8.52284789e-01 -1.20142698e+00 3.19825441e-01
-4.75876391e-01 6.48558065e-02 1.16240330e-01 1.38749814e+00
-1.53563774e+00 1.51230201e-01 -6.31310821e-01 -5.59566677e-01
-1.06679335e-01 3.92348081e-01 -3.73458266e-01 4.66679335e-01
-1.48690641e+00 9.86378431e-01 3.70329112e-01 4.47405964e-01
-1.33246863e+00 -8.34947973e-02 -9.05546784e-01 2.50853270e-01
-3.56018394e-01 -7.84049109e-02 1.40308225e+00 -1.00182509e+00
-2.14886570e+00 5.97587526e-01 -2.10275307e-01 -4.93742645e-01
6.79682046e-02 -8.68915915e-02 -1.13226199e+00 5.24351820e-02
-5.36316156e-01 7.05376863e-01 5.18305421e-01 -1.37012935e+00
-1.44905627e-01 -2.17787981e-01 -3.13417196e-01 -6.44899486e-03
-4.00942564e-01 4.86701667e-01 8.93128589e-02 -4.32670504e-01
3.03164832e-02 -9.15529251e-01 2.75433838e-01 -7.67606556e-01
-5.81120610e-01 -1.68063879e-01 1.13495314e+00 -1.39260983e+00
1.29422271e+00 -2.51368904e+00 -2.80396044e-01 1.41403437e-01
-1.41813874e-01 4.70908821e-01 -2.53012367e-02 4.67081219e-01
-2.98932314e-01 9.05002728e-02 -1.34163396e-02 -6.56045854e-01
4.79346573e-01 2.19722196e-01 7.37576261e-02 3.62722486e-01
-1.95917919e-01 3.62673730e-01 -2.47988060e-01 -5.48753977e-01
1.52861267e-01 9.06580031e-01 -3.23351324e-01 4.15685952e-01
5.37162125e-01 1.34496346e-01 4.15726840e-01 1.01159656e+00
9.86222088e-01 5.86396933e-01 2.49017775e-01 -4.00665812e-02
-2.93787615e-03 1.07407443e-01 -1.16236413e+00 1.19671524e+00
-5.82809150e-01 1.01931131e+00 3.85339051e-01 -8.19757402e-01
1.33264422e+00 1.09801412e+00 3.21410634e-02 -7.53803015e-01
4.46025789e-01 2.75292903e-01 2.99041361e-01 -7.96820164e-01
5.11802077e-01 -4.76136565e-01 -5.78963235e-02 1.72820747e-01
3.91405195e-01 2.36171022e-01 -4.45060164e-01 -9.08607244e-02
5.86478770e-01 -4.15261835e-01 3.67779918e-02 -4.37420867e-02
5.70277333e-01 -7.79480338e-01 7.99000859e-01 5.49717009e-01
-9.70945895e-01 1.94928586e-01 3.57498527e-01 7.37284720e-02
-8.26567650e-01 -8.93305659e-01 -2.61725426e-01 1.02007079e+00
-3.70449126e-01 -1.24888204e-01 -1.05826175e+00 -1.69345915e-01
-4.83875811e-01 9.91090357e-01 -3.75555426e-01 -2.71194786e-01
-1.14474423e-01 -5.71026027e-01 1.33397841e+00 4.55225825e-01
8.40533435e-01 -1.12965083e+00 -2.87024111e-01 7.84095451e-02
-3.46434623e-01 -1.30319834e+00 -1.22364372e-01 3.51154864e-01
-5.40780962e-01 -3.55949521e-01 -5.89740157e-01 -8.64743412e-01
5.40657900e-02 -3.33308727e-01 7.95109510e-01 -2.75944412e-01
1.49902534e-02 3.61144632e-01 -3.34930241e-01 -5.58251798e-01
-9.20077980e-01 -1.05998844e-01 6.70956314e-01 1.02507845e-01
7.38707602e-01 -6.75140917e-01 -3.49187315e-01 3.27798426e-01
-6.99765563e-01 -4.68736738e-01 5.29274702e-01 9.46386814e-01
-4.58627641e-02 2.08725497e-01 7.36236036e-01 1.06767684e-01
7.80200362e-01 -5.22634983e-01 -2.30760038e-01 4.42297980e-02
-4.66544628e-01 -4.68575925e-01 3.38922054e-01 -5.60995817e-01
-1.43057632e+00 -3.48569721e-01 -8.29318523e-01 -5.72584450e-01
-6.91747367e-01 6.38970077e-01 -6.33234322e-01 2.44958326e-01
3.34883809e-01 3.74647617e-01 3.55771556e-02 -4.29613560e-01
-1.69013634e-01 1.50935650e+00 5.28455436e-01 -3.12712938e-01
1.42271131e-01 -3.27482894e-02 -5.65348446e-01 -1.17949498e+00
-1.83271527e-01 -1.01668745e-01 -5.32945454e-01 -4.14747834e-01
1.08563673e+00 -1.07859194e+00 -9.79701459e-01 1.13561714e+00
-1.08923137e+00 -1.90713540e-01 1.46363214e-01 8.06891322e-01
-1.24288104e-01 5.41710556e-01 -1.00354373e+00 -1.43578088e+00
-6.81064725e-01 -1.35985756e+00 7.40075707e-01 3.65544975e-01
-3.27084571e-01 -9.84741211e-01 8.46589655e-02 5.06569624e-01
6.83189809e-01 2.37667561e-01 7.78952181e-01 -1.31444108e+00
4.69780982e-01 -4.00220603e-01 1.12985305e-01 1.08779418e+00
-1.95510775e-01 4.67407048e-01 -1.75019670e+00 -2.64222056e-01
4.27903712e-01 -4.06440079e-01 1.30246148e-01 5.54924726e-01
7.60143876e-01 -3.48486155e-01 2.81089723e-01 4.87382591e-01
1.06860268e+00 8.79381716e-01 7.74680078e-01 2.64504790e-01
1.68754399e-01 5.68504870e-01 1.95279077e-01 3.47647130e-01
3.86524826e-01 5.99818230e-01 2.75601447e-01 -1.12320364e-01
4.62155268e-02 1.20482050e-01 1.11728764e+00 1.45584369e+00
1.25564635e-01 -4.92498875e-01 -1.10608411e+00 3.85789633e-01
-1.19732249e+00 -1.08788657e+00 6.57831654e-02 1.82911110e+00
4.74362850e-01 9.95423645e-02 2.38916427e-01 6.27067506e-01
9.28960681e-01 -4.09551673e-02 -3.93832713e-01 -1.23822558e+00
-4.08574820e-01 1.79669827e-01 -1.68382257e-01 5.72778225e-01
-9.48427916e-01 6.66995823e-01 6.10678244e+00 7.36869395e-01
-1.32437027e+00 4.53861117e-01 6.43723607e-01 -9.82225016e-02
2.96938330e-01 -3.91995639e-01 -9.35161412e-01 6.30639195e-01
2.04517412e+00 7.88930580e-02 4.24876139e-02 1.04558945e+00
4.03415829e-01 -5.50933741e-02 -8.28181922e-01 1.04443443e+00
1.98707581e-01 -4.29959536e-01 -2.07043543e-01 2.96398193e-01
4.41187322e-01 1.15392350e-01 2.56211132e-01 8.32815647e-01
1.76362932e-01 -9.94691014e-01 7.67338336e-01 3.33543628e-01
6.54944420e-01 -1.02944183e+00 1.40245259e+00 2.45718464e-01
-7.72955477e-01 -1.72541052e-01 -8.43793675e-02 2.85371449e-02
1.97936922e-01 3.52314174e-01 -7.74223089e-01 6.10576808e-01
8.90666485e-01 4.59404550e-02 3.13361287e-02 4.95494336e-01
2.26420894e-01 1.07824624e+00 -2.65045196e-01 3.96035723e-02
2.69359797e-01 4.93506826e-02 5.51277637e-01 1.81287456e+00
5.13103724e-01 -3.77470218e-02 -1.86788663e-01 5.05377531e-01
1.17093615e-01 -8.73292387e-02 -4.83115971e-01 -2.07564980e-02
8.16523790e-01 1.26225078e+00 -3.32181454e-01 -4.86886233e-01
-2.21912667e-01 6.76378548e-01 -2.04320028e-01 4.48202044e-01
-1.32995224e+00 -6.35736644e-01 8.18552673e-01 -5.17372429e-01
1.00034885e-01 -2.24246085e-01 -1.72698215e-01 -8.06976497e-01
-2.03049496e-01 -1.29899228e+00 2.25180626e-01 -8.04353833e-01
-1.11465859e+00 1.19367933e+00 -2.44125854e-02 -9.49078083e-01
-2.51027614e-01 -4.33212548e-01 -6.55043840e-01 1.12212491e+00
-1.19802737e+00 -8.10996413e-01 -4.05235320e-01 5.67958891e-01
4.98713583e-01 -6.46681070e-01 1.07883561e+00 7.27836311e-01
-1.20410764e+00 1.26822245e+00 1.72720701e-01 3.81689459e-01
4.91790175e-01 -7.74111211e-01 5.09385727e-02 9.64311004e-01
-3.49056214e-01 6.17127657e-01 6.69559479e-01 -1.76679730e-01
-1.17287183e+00 -7.31486142e-01 8.61297965e-01 1.24430872e-01
1.72202572e-01 -4.30542082e-01 -8.89040411e-01 5.45967221e-01
8.93250048e-01 -6.39557719e-01 1.24462545e+00 -3.22731324e-02
-3.42905641e-01 3.59250568e-02 -1.57389581e+00 3.83812517e-01
3.48593444e-01 -9.25512433e-01 -4.33303148e-01 -1.16950251e-01
6.23620033e-01 -3.30779910e-01 -1.15873957e+00 3.81530762e-01
7.64196277e-01 -1.38966668e+00 6.18135512e-01 -3.66826087e-01
2.04319045e-01 1.30917341e-01 -9.28781211e-01 -1.16539466e+00
-6.87094778e-02 -6.05035841e-01 -3.28388363e-01 1.71985388e+00
5.08873999e-01 -8.69635165e-01 3.07514846e-01 6.69973552e-01
-5.18817544e-01 -7.26934791e-01 -1.19112670e+00 -9.22528446e-01
-5.51977642e-02 -6.89342976e-01 7.05641448e-01 1.05769396e+00
-2.13956341e-01 3.58616598e-02 -3.04072231e-01 4.54326332e-01
1.98297381e-01 -6.65056467e-01 6.87691927e-01 -7.26492226e-01
-1.46191284e-01 -4.40930277e-01 -7.22102582e-01 -4.61643010e-01
5.88772118e-01 -5.49736559e-01 -1.04748318e-02 -1.18456280e+00
1.09324120e-01 1.30465835e-01 -4.26016510e-01 5.10284245e-01
1.10125124e-01 -1.15018308e-01 7.86117390e-02 -1.88268259e-01
1.29380748e-01 6.80769682e-01 4.06162381e-01 -1.32359862e-01
-2.71261692e-01 -7.73494393e-02 -4.08053756e-01 7.32594013e-01
1.09320438e+00 -3.05589348e-01 -2.14888766e-01 -1.42231137e-01
-5.71326315e-01 2.06307366e-01 2.35421881e-01 -1.23174846e+00
2.19378248e-01 3.39053005e-01 2.98202515e-01 -7.16474533e-01
7.98273146e-01 -6.51126683e-01 3.25496614e-01 4.30638999e-01
-1.30030736e-01 1.98451892e-01 6.48118794e-01 3.47643755e-02
-5.83714783e-01 -8.95404816e-02 1.04980958e+00 4.15007055e-01
-5.44603467e-01 -2.24817917e-01 -8.47137809e-01 -4.94541019e-01
9.72676277e-01 -6.52504563e-01 -1.34386778e-01 -8.34760368e-01
-1.06195998e+00 -2.47851431e-01 -7.93958306e-02 4.09982920e-01
7.21534431e-01 -1.28488696e+00 -6.18130326e-01 3.30084741e-01
-3.59729797e-01 -5.83053827e-01 8.26016843e-01 1.12413847e+00
-3.50191683e-01 5.21683156e-01 -2.56969362e-01 -6.92199767e-01
-1.75472617e+00 2.12836474e-01 7.51480043e-01 -2.99989842e-02
-1.09302893e-01 1.00345862e+00 -1.02820829e-01 -9.44739997e-01
8.71930897e-01 -2.43275031e-01 5.44124730e-02 1.61997795e-01
1.66414708e-01 6.49540186e-01 3.04428846e-01 -1.05356300e+00
-4.99693125e-01 1.31377742e-01 -4.18320149e-02 -5.40910542e-01
1.09995985e+00 -2.28324562e-01 1.06289335e-01 5.10371983e-01
1.34306395e+00 -2.14696378e-01 -6.56464040e-01 1.72081068e-01
-1.07671902e-01 6.48742467e-02 3.57971281e-01 -1.01308453e+00
-1.09057701e+00 1.19380665e+00 1.11228621e+00 2.53389180e-01
1.40615892e+00 -5.88051081e-01 9.34459686e-01 2.64979154e-01
-6.38672560e-02 -1.34407067e+00 -2.03033224e-01 7.47879267e-01
6.47987843e-01 -9.48637605e-01 -5.60851872e-01 3.40828896e-01
-8.93149734e-01 1.09403002e+00 7.72059262e-01 6.30000234e-01
1.07369292e+00 5.08975327e-01 5.82208693e-01 -1.53081551e-01
-7.62318611e-01 2.34404668e-01 -9.80844125e-02 6.81915104e-01
5.72565675e-01 1.35782793e-01 1.57932341e-01 1.03453517e+00
-6.19945824e-01 -1.60027266e-01 3.69623899e-01 9.33295310e-01
-1.05813250e-01 -5.80489039e-01 -5.81406534e-01 -1.05491377e-01
-6.41022623e-01 -8.71477798e-02 -4.24425453e-01 8.47038746e-01
1.15178503e-01 1.54389834e+00 -1.55480299e-02 -1.07651854e+00
2.55319685e-01 5.99735022e-01 -8.23657587e-02 -1.53925829e-03
-9.75467086e-01 1.89769760e-01 2.59079099e-01 -1.49550974e-01
-3.45662981e-01 -6.35942519e-01 -9.71132636e-01 -8.22092533e-01
-6.52148902e-01 3.68196249e-01 1.30070734e+00 7.56172478e-01
5.85117817e-01 4.89661872e-01 8.19827855e-01 -8.82474363e-01
-5.95192373e-01 -1.53084874e+00 -8.17237735e-01 -3.13179605e-02
3.65881950e-01 -4.26712036e-01 -7.20961452e-01 -5.10363057e-02] | [14.338117599487305, 6.016212463378906] |
c6f06ed6-0a30-4fe5-ae59-00a7d18aa968 | multi-view-reasoning-consistent-contrastive | 2210.11694 | null | https://arxiv.org/abs/2210.11694v1 | https://arxiv.org/pdf/2210.11694v1.pdf | Multi-View Reasoning: Consistent Contrastive Learning for Math Word Problem | Math word problem solver requires both precise relation reasoning about quantities in the text and reliable generation for the diverse equation. Current sequence-to-tree or relation extraction methods regard this only from a fixed view, struggling to simultaneously handle complex semantics and diverse equations. However, human solving naturally involves two consistent reasoning views: top-down and bottom-up, just as math equations also can be expressed in multiple equivalent forms: pre-order and post-order. We propose a multi-view consistent contrastive learning for a more complete semantics-to-equation mapping. The entire process is decoupled into two independent but consistent views: top-down decomposition and bottom-up construction, and the two reasoning views are aligned in multi-granularity for consistency, enhancing global generation and precise reasoning. Experiments on multiple datasets across two languages show our approach significantly outperforms the existing baselines, especially on complex problems. We also show after consistent alignment, multi-view can absorb the merits of both views and generate more diverse results consistent with the mathematical laws. | ['Weiming Lu', 'Qingpeng Nong', 'Zeqi Tan', 'Xiaoxia Cheng', 'Yanna Ma', 'Yongliang Shen', 'Wenqi Zhang'] | 2022-10-21 | null | null | null | null | ['math-word-problem-solving', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'reasoning', 'time-series'] | [ 1.36082485e-01 1.18177772e-01 -2.74541527e-01 -5.50191462e-01
-1.09777570e+00 -9.94281709e-01 7.55556405e-01 1.10904731e-01
1.44703493e-01 7.50028610e-01 4.63234723e-01 -3.70155305e-01
-1.74957529e-01 -1.17413199e+00 -6.23676240e-01 -3.74158025e-02
4.90907788e-01 8.65510523e-01 -1.01224393e-01 -5.56939662e-01
2.52958059e-01 9.56153572e-02 -1.28195500e+00 6.37158155e-01
1.18761206e+00 5.28527498e-01 -1.39132380e-01 7.86341786e-01
-6.82028830e-01 1.45168817e+00 -3.58072102e-01 -8.38857830e-01
4.08396691e-01 -4.79026020e-01 -1.00420988e+00 -1.32470414e-01
9.70643580e-01 -4.95235056e-01 1.86490510e-02 1.13103127e+00
1.57096133e-01 4.80095856e-02 5.27043045e-01 -1.36207533e+00
-8.17200005e-01 9.16165471e-01 -5.17367005e-01 -1.03784949e-01
9.70243573e-01 9.61661935e-02 1.49854016e+00 -8.32163930e-01
7.67553151e-01 1.47916234e+00 7.13571131e-01 5.68405509e-01
-1.26219320e+00 -5.65234005e-01 5.73925972e-01 3.50037426e-01
-1.10687530e+00 -3.94130677e-01 4.72269177e-01 -4.60348427e-01
1.47775781e+00 5.00535011e-01 6.82324111e-01 8.29781413e-01
1.25970259e-01 6.54895008e-01 1.17468488e+00 -7.25306422e-02
1.53775364e-01 -1.40370414e-01 4.72036809e-01 9.71164703e-01
5.92787087e-01 -4.52329993e-01 -9.98967886e-01 -6.41472638e-02
7.61002541e-01 -3.41534972e-01 -1.69618428e-01 -3.78316820e-01
-1.39161015e+00 7.10141122e-01 5.59571385e-02 -1.22808777e-02
-2.29699045e-01 2.77952343e-01 3.27042907e-01 3.21713507e-01
2.99357235e-01 9.27735031e-01 -9.24100637e-01 -2.82739192e-01
-8.65332782e-01 9.51351881e-01 1.30940568e+00 1.36957729e+00
5.93048453e-01 -6.48892820e-02 -1.52069345e-01 3.66758645e-01
2.13742837e-01 5.96292317e-01 8.41414332e-02 -1.37930572e+00
1.14838982e+00 7.68592417e-01 1.02248654e-01 -1.00097585e+00
-3.85387331e-01 -3.12501520e-01 -6.54807687e-01 6.16765348e-03
7.54276156e-01 -1.58450782e-01 -6.95921540e-01 1.93076515e+00
4.22823638e-01 -9.20901820e-02 2.63997972e-01 8.96139920e-01
9.23573971e-01 7.92336524e-01 -1.00407869e-01 -1.42463505e-01
1.64212823e+00 -1.26523340e+00 -9.40026402e-01 -5.31660497e-01
8.62146199e-01 -5.80105245e-01 9.80931878e-01 6.38915658e-01
-1.56528258e+00 -2.71001667e-01 -1.02396524e+00 -8.36352289e-01
-3.22170943e-01 -2.39239082e-01 1.04797232e+00 2.70268321e-01
-9.56819713e-01 3.66709769e-01 -7.15698957e-01 1.89237684e-01
7.53143802e-02 9.57579017e-02 -3.34456950e-01 -2.98660457e-01
-1.35105503e+00 1.22821534e+00 2.53733397e-01 -6.48171827e-02
-1.99250370e-01 -1.39336085e+00 -1.25208640e+00 1.26803815e-01
7.56640017e-01 -1.37796998e+00 1.54053557e+00 -4.37582523e-01
-1.35798371e+00 7.97770262e-01 -6.18303537e-01 -2.75762320e-01
5.58835506e-01 -5.38970947e-01 -4.95412052e-02 -1.35600984e-01
4.31331128e-01 3.30740750e-01 4.91925299e-01 -1.19509840e+00
-4.50741857e-01 -4.91591066e-01 5.24651408e-01 5.58984339e-01
1.07139386e-01 -6.03001267e-02 -5.44283807e-01 -6.28859758e-01
5.93922317e-01 -5.01898706e-01 -5.40650785e-02 -5.87363131e-02
-2.88505465e-01 -3.90655309e-01 1.09023079e-01 -9.78473067e-01
1.36529422e+00 -1.39563799e+00 9.22103286e-01 -1.23027891e-01
6.78598762e-01 -3.53031337e-01 1.36976212e-01 5.79901516e-01
-2.91511565e-01 2.01430976e-01 -1.78329363e-01 -2.49150276e-01
4.59812254e-01 2.29717866e-01 -5.24738371e-01 4.25988063e-02
1.74819976e-01 1.26904380e+00 -1.17309701e+00 -6.62309349e-01
5.19429818e-02 -1.68065369e-01 -8.29535306e-01 2.04510227e-01
-6.75737023e-01 1.96744390e-02 -3.75010759e-01 5.33024609e-01
6.56787694e-01 -6.44807398e-01 5.30994058e-01 -3.01163495e-01
1.59633398e-01 8.61936390e-01 -1.51762593e+00 2.08655787e+00
-6.58605576e-01 1.65152654e-01 -1.57977212e-02 -6.89929008e-01
4.32052672e-01 2.48716287e-02 2.44268224e-01 -7.31047690e-01
-2.10581422e-01 8.95029455e-02 -2.91445274e-02 -4.27879155e-01
6.48086190e-01 -3.02900314e-01 -3.42377156e-01 7.01145351e-01
4.98209931e-02 -7.37479985e-01 8.00838649e-01 8.83692801e-01
7.54600406e-01 7.09207356e-01 6.40878141e-01 -1.94454595e-01
5.75385869e-01 1.25324637e-01 7.36820459e-01 6.30208552e-01
4.15171504e-01 4.28274304e-01 7.94095159e-01 -6.72981739e-01
-7.97654569e-01 -1.09343719e+00 3.43698025e-01 9.21150684e-01
2.44342715e-01 -1.25598264e+00 -5.62871754e-01 -6.93990648e-01
5.55987097e-02 1.23780417e+00 -2.98171729e-01 1.47481635e-01
-9.66473043e-01 -3.49165350e-01 4.36641753e-01 7.53692865e-01
4.71478313e-01 -3.96259487e-01 -4.66160953e-01 1.95961401e-01
-5.57321548e-01 -1.51333034e+00 -2.16272250e-01 -2.20054295e-02
-6.24690115e-01 -9.71547425e-01 -1.36666328e-01 -3.85125577e-01
5.00089645e-01 5.90207390e-02 1.71788049e+00 2.90880769e-01
-8.04180652e-02 1.81251734e-01 1.43666444e-02 -3.15822870e-01
-5.27512431e-01 -1.31332189e-01 -1.35989547e-01 -6.00497782e-01
2.16492027e-01 -7.31490433e-01 8.64143372e-02 -4.86047506e-01
-4.89528447e-01 8.89475524e-01 1.16822816e-01 5.20765424e-01
3.01497996e-01 -4.07173745e-02 -6.04040362e-02 -1.25045800e+00
5.46213329e-01 -2.21997887e-01 -6.32081628e-01 7.26497471e-01
-6.64655447e-01 4.86572772e-01 8.96497428e-01 3.36527079e-02
-1.24533510e+00 -1.82600439e-01 1.88189834e-01 -1.15715198e-01
3.20897549e-02 5.52884519e-01 -4.05105114e-01 5.10180116e-01
5.50702333e-01 2.08752096e-01 -4.31819052e-01 -4.44970988e-02
1.06255436e+00 9.47619881e-03 5.37939370e-01 -1.34788990e+00
1.13561356e+00 1.50486469e-01 9.69904661e-02 -2.00917304e-01
-1.33027756e+00 -1.73583701e-01 -6.10678256e-01 4.79301549e-02
7.72506952e-01 -1.20156002e+00 -6.29012465e-01 5.01951516e-01
-1.66315019e+00 -2.68106103e-01 -3.41177523e-01 9.32535306e-02
-5.30181289e-01 3.93395275e-01 -8.57873440e-01 -6.25548422e-01
-3.40056032e-01 -1.05809867e+00 1.43721163e+00 1.60005242e-01
-7.39883661e-01 -1.22694194e+00 -7.39286141e-03 7.76304066e-01
2.88493216e-01 1.11220814e-01 1.27986777e+00 -4.78243768e-01
-9.13444877e-01 2.45312035e-01 -4.70941722e-01 -7.82270283e-02
3.11441541e-01 -1.94816768e-01 -7.27951884e-01 3.79215449e-01
2.50655323e-01 -4.73824292e-01 5.91723144e-01 -1.94928199e-01
7.46222913e-01 -5.56059659e-01 -3.38839972e-03 6.65755391e-01
1.45722616e+00 -3.88573468e-01 4.10190254e-01 -5.91858774e-02
1.08719420e+00 4.60992277e-01 3.17906290e-01 3.46322328e-01
1.19013786e+00 4.99591082e-01 8.96464195e-03 6.13533519e-02
-1.62129477e-01 -4.10899699e-01 2.73877621e-01 1.17789769e+00
-6.35565445e-02 -1.14003502e-01 -1.10925186e+00 2.28985772e-01
-2.05251908e+00 -1.14422846e+00 -3.39251518e-01 1.64999819e+00
1.54473197e+00 3.86951789e-02 -1.25265941e-01 -6.94049969e-02
1.53486028e-01 3.39901239e-01 -4.74501103e-01 -4.99618590e-01
-5.18610189e-03 3.59918505e-01 1.32625192e-01 1.02292979e+00
-7.36900151e-01 1.17062294e+00 6.47154188e+00 7.45641053e-01
-8.21465313e-01 -1.32513493e-01 2.26176500e-01 -2.92157710e-01
-1.06125832e+00 2.54199803e-01 -9.43673372e-01 -7.10498691e-02
4.85021681e-01 -3.32179636e-01 8.51964891e-01 6.89624131e-01
-3.06447387e-01 -9.72526371e-02 -1.51771843e+00 1.14863300e+00
2.89752901e-01 -1.63380098e+00 4.15522337e-01 -2.97189415e-01
8.53886247e-01 -6.65569782e-01 -2.97121435e-01 4.07323539e-01
6.99497998e-01 -1.03733206e+00 1.06804538e+00 5.34764826e-01
6.12712562e-01 -4.09771681e-01 2.92669833e-01 6.03097081e-01
-1.53410673e+00 2.48533070e-01 2.50624627e-01 -5.92712343e-01
4.07143623e-01 5.25929451e-01 -2.24236712e-01 1.20623398e+00
1.69904888e-01 1.07245553e+00 -4.64374363e-01 -7.70429149e-02
-6.63028002e-01 -6.88402774e-03 -1.66752666e-01 -5.11288568e-02
-2.29422580e-02 -5.60080409e-01 3.51446927e-01 9.81671751e-01
-5.75802661e-02 4.84916508e-01 2.23130122e-01 1.32208490e+00
5.08803315e-02 -1.23677433e-01 -4.21774089e-01 -1.85222328e-01
4.37563032e-01 1.19858849e+00 -4.10951465e-01 -8.97538245e-01
-6.53691113e-01 1.08370209e+00 6.79578245e-01 3.96966994e-01
-1.12971795e+00 7.81358331e-02 7.19154596e-01 -3.83728027e-01
8.84617269e-02 -5.92714965e-01 -1.03266418e+00 -1.91614830e+00
2.83036023e-01 -1.45370615e+00 5.23446262e-01 -9.47665274e-01
-1.46517754e+00 5.58332391e-02 4.11290854e-01 -6.32641554e-01
-5.16189337e-01 -7.35305130e-01 -1.51369750e-01 1.07695591e+00
-1.37497783e+00 -1.26485729e+00 -1.95305973e-01 3.75354648e-01
9.44763303e-01 2.18201935e-01 1.04166901e+00 4.07222472e-02
-3.21034074e-01 3.88500333e-01 -8.09327662e-01 5.70604242e-02
5.05433142e-01 -1.81301010e+00 7.49486685e-01 1.14244640e+00
4.12771285e-01 1.20986128e+00 6.18463337e-01 -8.88477027e-01
-1.94749117e+00 -6.53848112e-01 1.35404754e+00 -1.05103743e+00
9.39808726e-01 -5.71068883e-01 -7.35845506e-01 9.01992440e-01
3.58931392e-01 -4.66131449e-01 4.63146180e-01 4.81640220e-01
-1.04481459e+00 1.06718779e-01 -6.79513037e-01 9.46853101e-01
1.25474250e+00 -7.06551909e-01 -9.87489045e-01 6.15040302e-01
8.52749884e-01 -1.26106238e+00 -8.56618702e-01 3.21850330e-01
6.23405635e-01 -1.00508761e+00 1.00939775e+00 -1.07182050e+00
1.15720928e+00 -3.46965134e-01 -3.40867281e-01 -9.83007550e-01
-3.39190096e-01 -8.69936645e-01 -6.91106617e-01 1.14669716e+00
6.26872420e-01 -5.78581393e-01 4.93819743e-01 1.13978326e+00
3.72981057e-02 -9.75740910e-01 -5.17119348e-01 -5.64259410e-01
2.48653069e-01 -7.98878789e-01 8.99853468e-01 1.16488421e+00
1.83193445e-01 9.67148185e-01 -2.18364075e-01 2.33578607e-01
5.51699996e-01 8.08523059e-01 9.97067511e-01 -9.06854451e-01
-7.17033625e-01 -5.17497182e-01 1.30025417e-01 -1.19388568e+00
2.99185812e-01 -1.20011854e+00 -1.44667298e-01 -2.13150978e+00
3.37481886e-01 -1.14204884e-01 3.91705781e-01 4.47536886e-01
-6.03233099e-01 -4.31505471e-01 2.95216203e-01 -1.74559310e-01
-6.85853541e-01 2.30587885e-01 1.45446610e+00 -1.11447223e-01
1.30718678e-01 -5.45632601e-01 -1.14484346e+00 9.14435208e-01
4.87959474e-01 -1.27126262e-01 -6.83418810e-01 -8.31888974e-01
9.25570846e-01 3.04161578e-01 1.36800289e-01 -6.29246235e-01
4.81228292e-01 -6.60897017e-01 1.64773121e-01 -3.89719576e-01
3.95554841e-01 -4.30471718e-01 7.63676837e-02 1.41283825e-01
-4.14371043e-01 4.08852279e-01 1.45729601e-01 1.71454161e-01
1.46438733e-01 7.64848664e-02 2.45499223e-01 -5.16612113e-01
-6.44744992e-01 7.65944347e-02 5.62943853e-02 7.43006229e-01
6.61007643e-01 4.26206514e-02 -5.32657266e-01 -4.38715994e-01
-4.10380036e-01 5.06289005e-01 3.99205834e-01 4.15364832e-01
3.72795552e-01 -1.33740437e+00 -7.48111427e-01 2.47867610e-02
-1.30212307e-01 4.07569200e-01 1.04737155e-01 6.61837876e-01
-7.33844936e-01 4.06366020e-01 2.27261826e-01 -3.68766308e-01
-1.09265816e+00 5.15372396e-01 3.93464267e-01 -8.25806260e-01
-5.82234859e-01 1.00528061e+00 1.72911644e-01 -7.26246476e-01
1.23119242e-01 -7.90370822e-01 1.70064285e-01 -4.86906767e-02
5.84130526e-01 2.48807892e-01 2.41842177e-02 -2.32325599e-01
-4.35762197e-01 7.63982296e-01 -6.56288937e-02 -3.09487998e-01
1.21520829e+00 2.37302501e-02 -7.02083409e-01 6.94635808e-01
6.17481470e-01 2.98933417e-01 -7.08669007e-01 -2.51877308e-01
7.60482019e-03 -2.83015817e-01 -3.78735453e-01 -9.22090530e-01
-7.41985321e-01 7.81736374e-01 -6.66090548e-01 1.37967110e-01
8.60195041e-01 -9.27557200e-02 1.04333198e+00 6.17646992e-01
3.94324690e-01 -8.09611022e-01 4.50883843e-02 8.87658477e-01
9.92421627e-01 -8.45214128e-01 5.23177981e-01 -1.01017404e+00
-7.17422247e-01 1.27388179e+00 9.53159809e-01 4.01140675e-02
4.57110293e-02 8.15999210e-01 -3.03732028e-04 -3.43370527e-01
-1.34681273e+00 5.97754717e-02 4.52407330e-01 2.90524513e-01
6.39474928e-01 8.82636942e-03 -5.80659956e-02 8.21024299e-01
-7.35112369e-01 -1.05194218e-01 4.44838792e-01 1.04465115e+00
3.28711048e-02 -1.35894454e+00 -2.93585807e-01 3.44277710e-01
-5.87091520e-02 -5.99056184e-01 -6.20585322e-01 9.12012339e-01
5.28103039e-02 7.70200133e-01 -2.50488430e-01 -7.48934746e-02
2.15045869e-01 4.08996552e-01 1.20580626e+00 -9.55450356e-01
-6.36281431e-01 -2.86504179e-01 6.13453507e-01 -8.77513766e-01
-2.24401459e-01 -5.60130894e-01 -1.67566240e+00 -4.99626637e-01
-1.91098720e-01 -1.64741531e-01 1.98292881e-01 1.45594585e+00
2.36999616e-01 7.70618081e-01 2.02407241e-02 -3.94890606e-01
-9.24965680e-01 -5.32113254e-01 -2.52782524e-01 5.54875731e-01
1.57945648e-01 -3.05617064e-01 -2.30513752e-01 3.74287546e-01] | [9.715309143066406, 7.459100723266602] |
00694f27-b70b-4349-99ab-8eab1c8b247a | continual-few-shot-intent-detection | null | null | https://aclanthology.org/2022.coling-1.26 | https://aclanthology.org/2022.coling-1.26.pdf | Continual Few-shot Intent Detection | Intent detection is at the core of task-oriented dialogue systems. Existing intent detection systems are typically trained with a large amount of data over a predefined set of intent classes. However, newly emerged intents in multiple domains are commonplace in the real world. And it is time-consuming and impractical for dialogue systems to re-collect enough annotated data and re-train the model. These limitations call for an intent detection system that could continually recognize new intents with very few labeled examples. In this work, we study the Continual Few-shot Intent Detection (CFID) problem and construct a benchmark consisting of nine tasks with multiple domains and imbalanced classes. To address the key challenges of (a) catastrophic forgetting during continuous learning and (b) negative knowledge transfer across tasks, we propose the Prefix-guided Lightweight Encoder (PLE) with three auxiliary strategies, namely Pseudo Samples Replay (PSR), Teacher Knowledge Transfer (TKT) and Dynamic Weighting Replay (DWR). Extensive experiments demonstrate the effectiveness and efficiency of our method in preventing catastrophic forgetting and encouraging positive knowledge transfer across tasks. | ['Yin Zhang', 'Ji Zhang', 'Xing Gao', 'Qianglong Chen', 'Yuchen Zhai', 'Guodun Li'] | null | null | null | null | coling-2022-10 | ['task-oriented-dialogue-systems'] | ['natural-language-processing'] | [ 4.75096196e-01 6.23158738e-02 -1.93463638e-01 -5.52506030e-01
-6.51835203e-01 -3.57842147e-01 6.52985156e-01 1.32519886e-01
-7.33928978e-01 1.03823316e+00 3.64506781e-01 -2.39840686e-01
9.22925174e-02 -3.23910862e-01 -2.41252780e-01 -3.41329932e-01
-2.02776064e-04 5.65041959e-01 3.67324769e-01 -4.11280513e-01
2.73238868e-01 -4.97808158e-02 -1.37980580e+00 2.50106752e-01
1.44005561e+00 6.96264505e-01 1.91581875e-01 7.43235528e-01
-2.37725824e-01 1.20339358e+00 -1.00046635e+00 -6.07322574e-01
-7.65754730e-02 -4.67301488e-01 -1.02379429e+00 1.50081500e-01
1.89720929e-01 -7.83073604e-01 -3.77878755e-01 8.25015664e-01
6.96144283e-01 5.28924346e-01 5.64114571e-01 -1.48597956e+00
-8.05412054e-01 4.21312094e-01 -6.27567828e-01 3.08416307e-01
4.92493421e-01 4.37492222e-01 6.44674480e-01 -1.06380677e+00
3.37417811e-01 1.07065392e+00 7.77277768e-01 1.06590319e+00
-8.88417959e-01 -6.32987559e-01 4.01475400e-01 3.67153108e-01
-8.71978879e-01 -6.00163519e-01 8.10319424e-01 -3.04431587e-01
1.05029213e+00 1.46057740e-01 5.28095365e-01 1.44895995e+00
-1.14689767e-02 1.21766520e+00 7.73639858e-01 -5.22815228e-01
4.61097896e-01 3.94768476e-01 5.07900417e-01 5.44512749e-01
4.42433171e-02 -1.94897294e-01 -1.01996374e+00 -3.47604364e-01
2.17163414e-01 3.35610688e-01 -3.74502957e-01 -1.91912249e-01
-9.83166397e-01 9.91457105e-01 -7.60017037e-02 -1.43279033e-02
-2.67236263e-01 -1.90805271e-01 8.40795994e-01 8.21504056e-01
8.12843144e-01 3.74804109e-01 -6.47637308e-01 -5.39850473e-01
-5.20236611e-01 1.68437764e-01 1.10578859e+00 1.11744595e+00
5.73045194e-01 2.37413086e-02 -5.34466803e-01 1.20755327e+00
-2.10908595e-02 2.91971684e-01 1.12723804e+00 -4.72615659e-01
5.01223803e-01 7.71324217e-01 1.19735532e-01 -3.70625675e-01
-2.18510509e-01 -1.44904628e-01 -8.67636502e-01 -2.08247885e-01
2.05593631e-01 -4.28690523e-01 -9.04404581e-01 1.81324184e+00
4.65500832e-01 -7.01986477e-02 2.70105571e-01 6.28602982e-01
6.19538248e-01 5.64827383e-01 3.72531205e-01 -5.00041068e-01
1.23432589e+00 -1.23557484e+00 -9.96363103e-01 -6.14602387e-01
7.51704931e-01 -5.86704195e-01 1.51309395e+00 3.19571733e-01
-7.34093428e-01 -3.52534950e-01 -8.93919110e-01 -1.48516834e-01
-1.97823316e-01 -8.82700235e-02 6.65703535e-01 4.05367225e-01
-8.11703146e-01 2.36769870e-01 -4.72812235e-01 -2.99565822e-01
2.39721403e-01 2.55821556e-01 -9.22701880e-02 -1.97008833e-01
-1.48585427e+00 8.53631794e-01 3.42890412e-01 -2.45030999e-01
-1.00545788e+00 -9.03526902e-01 -8.09624076e-01 -1.10859983e-01
5.32119811e-01 -5.74425995e-01 1.81434762e+00 -1.08680570e+00
-1.64180863e+00 5.09604096e-01 2.61233132e-02 -4.75143254e-01
7.33517230e-01 -7.70049393e-01 -2.02207431e-01 -1.61758035e-01
1.10896841e-01 6.29467845e-01 9.73225653e-01 -8.46123934e-01
-8.99222374e-01 -2.18602121e-01 -1.08646983e-02 7.70160913e-01
-8.37320447e-01 -1.85300231e-01 -2.23877877e-02 -6.02940679e-01
-5.07965684e-01 -6.77726805e-01 2.03415025e-02 -1.80134580e-01
-2.56462663e-01 -7.48535573e-01 1.02065659e+00 -6.27328277e-01
1.36233294e+00 -2.20458460e+00 1.29664009e-02 -6.77208900e-01
1.89982146e-01 6.59052253e-01 -1.08125322e-01 3.58474195e-01
1.67644024e-01 -3.21552694e-01 -2.86913365e-01 -4.50346619e-01
-1.27541855e-01 2.33339638e-01 -5.11862159e-01 3.46295126e-02
2.23821580e-01 7.21065700e-01 -1.29238987e+00 -3.63948494e-01
1.37932017e-01 1.66661382e-01 -5.11431992e-01 8.86473417e-01
-4.00967002e-01 1.71607479e-01 -2.64684618e-01 4.43218201e-01
3.54748070e-01 -2.58798838e-01 -7.84165575e-04 5.95259257e-02
1.25341341e-01 5.80069244e-01 -8.11642945e-01 1.71010590e+00
-6.02416337e-01 3.82665813e-01 -2.13303253e-01 -7.12535560e-01
8.92685115e-01 4.04401928e-01 7.93217272e-02 -7.19933510e-01
-1.64161190e-01 -2.22928040e-02 -8.79189298e-02 -6.11012042e-01
6.87584102e-01 -4.39252108e-01 -2.20147237e-01 9.76442754e-01
5.05947113e-01 1.25283683e-02 9.18231010e-02 3.32099140e-01
1.35965908e+00 -3.57017905e-01 5.62382162e-01 5.10739051e-02
4.68129516e-01 1.33509830e-01 5.57196736e-01 9.08582151e-01
-6.85742259e-01 2.42426619e-01 2.59793282e-01 -8.36916089e-01
-7.99972534e-01 -6.50355756e-01 2.27437511e-01 1.89475727e+00
-1.45524619e-02 -1.60252407e-01 -4.73789871e-01 -1.23448551e+00
-2.44030654e-02 1.12796295e+00 -5.09833872e-01 -8.27608943e-01
-4.29143459e-01 -7.54069149e-01 5.23586988e-01 3.72430652e-01
7.81288266e-01 -1.41406846e+00 -7.45852768e-01 3.94640684e-01
-3.53293717e-01 -6.35932863e-01 -8.96154106e-01 4.87974256e-01
-6.63724661e-01 -8.81737888e-01 -8.13258529e-01 -7.01878190e-01
5.75614333e-01 6.30040526e-01 1.16747320e+00 7.29251131e-02
-3.25497627e-01 4.95851904e-01 -5.66954255e-01 -6.02659106e-01
-4.35561597e-01 2.42228314e-01 4.30268615e-01 -7.16459900e-02
7.34532058e-01 -2.83206850e-01 -4.87027913e-01 1.62606195e-01
-7.74486780e-01 1.98791012e-01 6.22204721e-01 1.18793154e+00
-1.12081461e-01 -1.65565580e-01 1.05165291e+00 -1.10632098e+00
1.36894679e+00 -5.81960320e-01 -5.80029972e-02 5.65034628e-01
-7.58662105e-01 -8.10309201e-02 7.82294095e-01 -8.96259367e-01
-1.55949295e+00 -1.31325647e-01 -1.43732041e-01 -3.18275869e-01
-1.75340474e-01 2.41621435e-01 2.09105268e-01 2.10483298e-01
7.33292758e-01 6.17100894e-01 6.01855330e-02 -3.76716703e-01
3.47142220e-01 1.06800818e+00 3.25854421e-01 -4.65432733e-01
4.33076113e-01 2.19489876e-02 -1.01622593e+00 -8.55882347e-01
-1.34179962e+00 -6.37516558e-01 -4.65192229e-01 -3.16244721e-01
3.70305061e-01 -1.05943811e+00 -4.18581218e-01 7.37002969e-01
-1.19391286e+00 -6.47400320e-01 -5.75785041e-01 3.05506051e-01
-4.47152108e-01 3.42013270e-01 -7.27337658e-01 -1.11895859e+00
-9.47159827e-01 -8.53786886e-01 7.98846483e-01 4.50041473e-01
-5.63481987e-01 -1.00616217e+00 4.00720894e-01 4.26525682e-01
5.39218724e-01 -5.29618263e-01 8.79901648e-01 -1.23101318e+00
-1.86106737e-03 -7.34216794e-02 -7.59899393e-02 3.51922244e-01
4.09642845e-01 -5.58700621e-01 -1.00053418e+00 -4.45297450e-01
4.62274700e-01 -1.30008435e+00 7.46773124e-01 -2.33635202e-01
7.20550001e-01 -7.08182096e-01 1.36456722e-02 1.27701074e-01
9.20944154e-01 4.72850740e-01 1.78750679e-01 2.07736552e-01
5.29982746e-01 5.86096585e-01 8.99142504e-01 6.88325346e-01
5.19904792e-01 4.27686840e-01 -1.51385695e-01 1.73896521e-01
-6.45261928e-02 -3.94277543e-01 5.57337463e-01 1.05620730e+00
4.42698359e-01 -2.07689688e-01 -9.32390571e-01 6.68440700e-01
-1.91809213e+00 -6.53642595e-01 5.35490334e-01 2.12556720e+00
1.59083211e+00 1.32026449e-01 3.96131054e-02 -5.73782697e-02
6.28743649e-01 6.45493567e-02 -1.00021470e+00 -3.63200694e-01
5.21789730e-01 -2.04617873e-01 7.20564201e-02 4.07516629e-01
-1.02289224e+00 8.57201338e-01 6.16265965e+00 6.55642509e-01
-8.87200594e-01 2.80752033e-01 6.80455565e-01 8.31004605e-03
-1.46516129e-01 -3.49583507e-01 -9.92312133e-01 5.29789627e-01
8.10746133e-01 -3.34244698e-01 2.55628586e-01 1.23415899e+00
-3.81591856e-01 -1.56572640e-01 -1.07076347e+00 7.33045161e-01
3.62996727e-01 -7.01476157e-01 7.48105422e-02 -6.14807427e-01
5.91566980e-01 -7.32668787e-02 -8.45090076e-02 1.01895475e+00
7.72825658e-01 -5.87715924e-01 3.00180584e-01 2.40654081e-01
8.40004921e-01 -5.89246869e-01 6.33421659e-01 7.58091390e-01
-6.53864205e-01 -2.14171737e-01 -4.11571234e-01 -1.23753928e-01
3.57241668e-02 4.79580939e-01 -1.17127478e+00 8.21404681e-02
5.26886880e-01 6.03134692e-01 -5.04586935e-01 7.76259482e-01
-2.70327479e-01 6.43068194e-01 -1.62257925e-01 -3.36528331e-01
3.32967676e-02 1.26333326e-01 4.28202122e-01 1.15035737e+00
-5.43427691e-02 1.15051523e-01 1.80280998e-01 3.43582720e-01
-2.02862769e-01 -1.32505715e-01 -5.95996380e-01 -9.90804285e-03
6.47562206e-01 1.09653866e+00 -3.13289374e-01 -3.61443758e-01
-5.63936591e-01 1.24043334e+00 7.14157581e-01 1.24171793e-01
-7.04583883e-01 -4.73836660e-01 5.62725902e-01 -3.33886117e-01
-1.48145199e-01 9.42447223e-03 -8.84722825e-03 -1.29236794e+00
-1.07630854e-02 -1.06243825e+00 6.86528981e-01 -3.71813148e-01
-1.73298800e+00 3.48308116e-01 -1.34786695e-01 -1.03141928e+00
-2.18630403e-01 -5.77693582e-02 -8.27393651e-01 4.87179190e-01
-1.63961220e+00 -7.27485240e-01 -3.68786514e-01 6.34698808e-01
1.40398264e+00 -3.16042542e-01 7.99651265e-01 2.27120668e-01
-6.81010187e-01 7.69091308e-01 -5.38752750e-02 -4.61871270e-03
1.17320752e+00 -1.35096300e+00 4.69200969e-01 4.75426883e-01
-2.20070675e-01 5.38257122e-01 7.55098224e-01 -8.51787269e-01
-1.43577147e+00 -1.22189844e+00 9.13543224e-01 -3.54016453e-01
5.20374835e-01 -4.96966124e-01 -1.36454296e+00 7.49303520e-01
2.79021442e-01 -4.04497147e-01 8.77161026e-01 2.51849502e-01
-2.87505120e-01 4.60861549e-02 -1.02625751e+00 6.21144533e-01
8.78616393e-01 -3.54431599e-01 -1.10773277e+00 6.09892488e-01
1.13331985e+00 -3.71901661e-01 -4.51324314e-01 3.78503770e-01
2.82609582e-01 -8.31802845e-01 8.13787699e-01 -8.04684401e-01
2.65198201e-01 1.81325808e-01 3.81766617e-01 -1.51927066e+00
-9.06472579e-02 -7.39820242e-01 -5.40678859e-01 1.30080843e+00
1.65428162e-01 -4.68363851e-01 6.50743663e-01 8.85935605e-01
-8.78989100e-02 -6.71930254e-01 -9.23364639e-01 -6.17335320e-01
-1.67524382e-01 -2.41552498e-02 3.12360913e-01 1.04329062e+00
3.90826076e-01 1.04146469e+00 -8.71702909e-01 -3.77514809e-01
4.88690764e-01 -1.03640102e-01 8.31851780e-01 -1.13345206e+00
-1.05922315e-02 1.99252162e-02 2.05436602e-01 -1.15822160e+00
8.74537230e-02 -3.59121650e-01 4.39147681e-01 -1.25429642e+00
3.64524931e-01 -5.00517666e-01 -2.23923177e-01 7.95669496e-01
-7.72482157e-01 -3.21208805e-01 -9.48639587e-02 3.79568815e-01
-1.29949760e+00 1.15443099e+00 9.64183211e-01 -1.52806938e-01
-4.75564241e-01 1.48403257e-01 -6.83476269e-01 6.81879342e-01
8.21940482e-01 -4.93013173e-01 -8.78001034e-01 -1.91111863e-01
7.75184780e-02 6.90092146e-02 -1.16967134e-01 -9.34301496e-01
5.83009303e-01 -2.09159613e-01 2.45102718e-01 -5.14956117e-01
1.53115422e-01 -6.32346332e-01 -4.46485460e-01 6.87579155e-01
-6.84965193e-01 -1.15894273e-01 1.48186177e-01 7.88295329e-01
-1.48439735e-01 -4.62181568e-01 7.51911223e-01 -1.43248662e-01
-9.91815090e-01 1.64350480e-01 -4.19250458e-01 6.05215669e-01
1.10031796e+00 1.16114110e-01 -7.17060030e-01 -4.52112705e-01
-2.39656270e-01 6.79862618e-01 2.25519463e-01 7.20990002e-01
8.04468036e-01 -1.12764704e+00 -5.53182125e-01 4.02349919e-01
5.14962375e-01 1.45389199e-01 4.64411736e-01 5.48664331e-01
-1.06560644e-02 2.87979960e-01 -1.52665660e-01 -3.30128491e-01
-1.15073240e+00 5.15614688e-01 1.33425504e-01 -5.29188931e-01
-7.24063575e-01 1.12150550e+00 1.99461743e-01 -8.20578754e-01
8.15616786e-01 2.25052893e-01 -2.56730914e-01 2.67880738e-01
9.40046251e-01 2.26451829e-01 -5.96685102e-03 5.82614020e-02
-1.27014592e-01 -2.35818475e-01 -8.93675387e-01 1.63761258e-01
1.21272731e+00 -2.31928810e-01 3.14734638e-01 7.95819759e-01
8.01077485e-01 -6.24302506e-01 -1.51836705e+00 -5.89562118e-01
6.00406490e-02 -3.25862914e-01 -3.81939888e-01 -1.01845217e+00
-4.29035962e-01 8.45526934e-01 4.53281850e-01 1.95552677e-01
1.06305957e+00 -2.60155112e-01 1.09166873e+00 7.50229239e-01
3.65988493e-01 -1.36740935e+00 7.66965032e-01 9.04490829e-01
8.14259231e-01 -1.40973878e+00 -2.01075062e-01 3.78401875e-02
-1.25145113e+00 8.31176221e-01 1.17057025e+00 3.66757065e-01
3.28189224e-01 1.49105161e-01 1.31320909e-01 -4.68354635e-02
-1.45091259e+00 6.97388053e-02 -2.38420479e-02 5.46267509e-01
5.10420859e-01 -1.61150604e-01 -1.54554382e-01 6.71891570e-01
2.12888256e-01 1.96282342e-01 5.46043932e-01 1.42097664e+00
-7.38172054e-01 -8.46138597e-01 3.24300118e-02 7.89039016e-01
-1.76309198e-01 -1.83552384e-01 -5.00284255e-01 4.28193182e-01
-3.98010910e-01 8.76539946e-01 -1.79829299e-01 -4.09803867e-01
4.05302465e-01 7.83656836e-01 -7.76928756e-03 -9.74839389e-01
-7.70365953e-01 -4.94379938e-01 -2.06184424e-02 -2.25985989e-01
-5.43565862e-02 -3.58565331e-01 -7.95116782e-01 -1.32706016e-01
-5.36816478e-01 2.72102594e-01 2.73902595e-01 9.54596639e-01
3.94042015e-01 5.61686814e-01 6.68535292e-01 -2.99110383e-01
-1.18939471e+00 -1.33325279e+00 -3.96877587e-01 5.46179891e-01
5.20102024e-01 -7.11427212e-01 -5.87589681e-01 7.87897184e-02] | [12.185718536376953, 7.6198625564575195] |
467c8480-335c-4bbd-a68a-8417cb8a6fb8 | diet-networks-thin-parameters-for-fat | 1611.09340 | null | http://arxiv.org/abs/1611.09340v3 | http://arxiv.org/pdf/1611.09340v3.pdf | Diet Networks: Thin Parameters for Fat Genomics | Learning tasks such as those involving genomic data often poses a serious
challenge: the number of input features can be orders of magnitude larger than
the number of training examples, making it difficult to avoid overfitting, even
when using the known regularization techniques. We focus here on tasks in which
the input is a description of the genetic variation specific to a patient, the
single nucleotide polymorphisms (SNPs), yielding millions of ternary inputs.
Improving the ability of deep learning to handle such datasets could have an
important impact in precision medicine, where high-dimensional data regarding a
particular patient is used to make predictions of interest. Even though the
amount of data for such tasks is increasing, this mismatch between the number
of examples and the number of inputs remains a concern. Naive implementations
of classifier neural networks involve a huge number of free parameters in their
first layer: each input feature is associated with as many parameters as there
are hidden units. We propose a novel neural network parametrization which
considerably reduces the number of free parameters. It is based on the idea
that we can first learn or provide a distributed representation for each input
feature (e.g. for each position in the genome where variations are observed),
and then learn (with another neural network called the parameter prediction
network) how to map a feature's distributed representation to the vector of
parameters specific to that feature in the classifier neural network (the
weights which link the value of the feature to each of the hidden units). We
show experimentally on a population stratification task of interest to medical
studies that the proposed approach can significantly reduce both the number of
parameters and the error rate of the classifier. | ['Marie-Pierre Dubé', 'Marc-André Legault', 'Pierre Luc Carrier', 'Akram Erraqabi', 'Yoshua Bengio', 'Tristan Sylvain', 'Julie G. Hussin', 'Etienne Dejoie', 'Alex Auvolat', 'Adriana Romero'] | 2016-11-28 | null | null | null | null | ['parameter-prediction'] | ['miscellaneous'] | [ 4.21860695e-01 1.98159441e-01 -1.31625757e-02 -6.32431567e-01
-5.65487623e-01 -3.99756074e-01 2.64845073e-01 4.34116870e-01
-6.04997396e-01 7.63855755e-01 -1.32000491e-01 -4.22479123e-01
-3.28603923e-01 -8.85744572e-01 -1.04841411e+00 -1.04101551e+00
-1.85452193e-01 7.17213035e-01 -5.80647364e-02 3.84911858e-02
-5.87137230e-02 4.43572164e-01 -1.51818824e+00 1.85230121e-01
8.64585876e-01 1.11994767e+00 1.91171411e-02 6.73969626e-01
-6.08144104e-02 4.06827442e-02 -7.33810902e-01 -9.54750627e-02
1.56244203e-01 -3.95855337e-01 -4.88063663e-01 -2.35366404e-01
3.03351283e-01 2.57228762e-01 2.20826745e-01 1.07988095e+00
6.42274559e-01 1.87095284e-01 9.22792792e-01 -8.95208836e-01
-3.73909384e-01 5.29146731e-01 -2.04119831e-01 -9.11094546e-02
-2.75032043e-01 1.13233425e-01 9.48951840e-01 -4.55328435e-01
5.11393070e-01 1.04209208e+00 8.33204269e-01 5.25037587e-01
-1.44818056e+00 -2.53535420e-01 -1.13577053e-01 -1.38576597e-01
-1.31660795e+00 -1.60402000e-01 2.78555930e-01 -7.75334418e-01
8.42872441e-01 1.91165045e-01 5.29510856e-01 9.31491792e-01
2.26364732e-01 1.22758396e-01 4.02282059e-01 -3.74730587e-01
7.13591814e-01 2.18718678e-01 3.28924388e-01 6.66685462e-01
4.91787046e-01 4.75359857e-02 -9.50987712e-02 -5.81587315e-01
5.79006314e-01 5.67836650e-02 -1.07310861e-01 -3.22667927e-01
-8.02475631e-01 1.17982340e+00 5.41987121e-01 3.54277283e-01
-3.73100907e-01 2.60238707e-01 4.09333885e-01 1.51588649e-01
3.40370268e-01 6.32055759e-01 -8.12341392e-01 1.75630450e-01
-6.26689553e-01 5.54797873e-02 7.51393080e-01 3.89404356e-01
9.10039246e-01 -2.18578056e-01 -1.26345769e-01 9.82302845e-01
-2.49634739e-02 2.53352940e-01 7.46723592e-01 -3.88064504e-01
3.35389376e-01 7.74790287e-01 -5.80427982e-02 -9.09137487e-01
-8.23515236e-01 -5.26770830e-01 -1.24941683e+00 1.77280113e-01
7.08954871e-01 -5.86317837e-01 -9.40310180e-01 2.05264616e+00
3.63385051e-01 -1.63557772e-02 1.20376265e-02 8.01421344e-01
4.63923126e-01 6.78575218e-01 1.22007802e-01 1.81319762e-03
1.36857319e+00 -3.36331248e-01 -2.21960634e-01 -3.69239569e-01
9.90715981e-01 -2.01461092e-01 8.85647595e-01 8.60201493e-02
-6.12257957e-01 -3.78050953e-01 -9.41346169e-01 -9.76741314e-02
-6.86930537e-01 2.81071037e-01 5.52455664e-01 6.40738487e-01
-7.48207331e-01 7.29967237e-01 -6.90814197e-01 -2.50836730e-01
4.61095184e-01 9.05651867e-01 -4.34547812e-01 6.10599294e-02
-1.31822491e+00 8.75145972e-01 7.16953516e-01 4.10248637e-01
-2.08164096e-01 -8.52152050e-01 -7.13615954e-01 5.20365119e-01
1.86630756e-01 -8.31713438e-01 4.87727225e-01 -1.07969213e+00
-1.15980911e+00 5.72687387e-01 1.17754027e-01 -3.33580524e-01
2.35392585e-01 6.06665500e-02 -1.15223862e-01 -4.06212240e-01
-2.91776389e-01 5.29533744e-01 7.46404409e-01 -5.72670102e-01
-3.38647246e-01 -4.83824074e-01 -1.94061086e-01 -2.35088944e-01
-3.69787782e-01 -2.85707206e-01 -1.58795610e-01 -3.32427144e-01
-8.76270905e-02 -1.10599470e+00 -5.49507439e-01 9.63688269e-02
-4.79170442e-01 -2.35537991e-01 3.78945023e-01 -5.21836996e-01
8.89672935e-01 -2.22407818e+00 3.60009938e-01 5.03463626e-01
6.95565790e-02 5.10665059e-01 -1.98182523e-01 2.49336272e-01
-2.96362877e-01 1.51002482e-01 -4.92097646e-01 2.30040848e-01
-1.34015396e-01 3.09000611e-01 1.03643581e-01 4.93672162e-01
4.92548048e-01 7.23895907e-01 -4.95231390e-01 7.05233663e-02
-2.69634575e-01 8.86188090e-01 -8.58100593e-01 2.56185472e-01
-5.14135659e-01 2.67034709e-01 -6.81607485e-01 7.02880032e-04
5.18397391e-01 -4.81431365e-01 3.67722481e-01 -7.00728297e-02
2.80582696e-01 1.17035218e-01 -1.31833184e+00 1.28580225e+00
-5.15034735e-01 3.26322734e-01 -2.64326781e-01 -1.22766221e+00
1.00030053e+00 1.65223435e-01 1.88715428e-01 -1.56496048e-01
1.14825897e-01 2.28569880e-01 4.19452339e-01 -5.97304583e-01
-1.50204077e-01 -3.20573300e-01 -1.86167583e-01 4.42532301e-01
8.07054341e-02 4.95489627e-01 3.78312059e-02 -2.51509994e-01
1.25771618e+00 -2.82026768e-01 4.98091578e-01 -2.79539376e-01
5.37117302e-01 -3.06094676e-01 8.93868148e-01 6.18143141e-01
3.54202539e-01 3.43591869e-01 1.19123876e+00 -7.72664547e-01
-1.07837522e+00 -5.41291118e-01 -3.87063205e-01 8.19817901e-01
-6.72429979e-01 7.57428957e-03 -7.30795681e-01 -4.60541517e-01
3.33804518e-01 3.93531412e-01 -9.46758628e-01 -2.85011828e-01
-4.77965891e-01 -1.40532398e+00 5.43605864e-01 3.20877880e-01
-7.66178891e-02 -1.09763849e+00 -7.91135788e-01 2.02718765e-01
2.94485301e-01 -6.71083689e-01 -1.64909571e-01 7.61568785e-01
-9.60693061e-01 -1.04336059e+00 -4.31225061e-01 -6.63670659e-01
9.65739667e-01 -6.66947186e-01 9.71762180e-01 1.72801405e-01
-6.76653802e-01 -4.01721209e-01 -8.33341181e-02 -4.98463959e-01
-4.20679539e-01 3.66418809e-01 6.38888637e-03 3.09351474e-01
5.63281953e-01 -4.98582900e-01 -3.91405553e-01 -1.59019604e-01
-1.01344955e+00 -2.74391711e-01 7.85294116e-01 1.25672674e+00
4.30912286e-01 -6.83421418e-02 8.74587655e-01 -1.46385610e+00
5.25921404e-01 -6.12093091e-01 -7.89135337e-01 2.47143269e-01
-1.80314347e-01 6.46244526e-01 9.05994296e-01 -5.12247264e-01
-4.96376604e-01 4.93431836e-01 -2.43819430e-01 1.11164823e-02
-2.34690681e-01 7.84435332e-01 -2.92099804e-01 4.78729084e-02
8.10682893e-01 -1.07311957e-01 1.81902468e-01 -5.50985515e-01
2.27176622e-01 6.37888193e-01 5.48563823e-02 -3.32084477e-01
4.10963893e-01 -1.55693546e-01 4.98514533e-01 -9.04705942e-01
-7.23149538e-01 -1.17604882e-01 -7.89959013e-01 3.27807665e-01
7.78541148e-01 -5.05555451e-01 -9.88609016e-01 1.19865604e-01
-8.49720776e-01 -2.99157590e-01 -2.69026637e-01 5.80512285e-01
-4.54780757e-01 -1.34636790e-01 -3.99665654e-01 -5.19244790e-01
-2.01623678e-01 -1.27247643e+00 7.12694168e-01 1.05061308e-01
-2.21492812e-01 -1.13095653e+00 9.27456990e-02 -1.66752174e-01
4.46154803e-01 3.42889816e-01 1.81922257e+00 -1.08806562e+00
-2.44792238e-01 -5.70557773e-01 -9.51773226e-02 2.05482394e-01
1.99379250e-01 -2.48497978e-01 -9.70529437e-01 -1.94840595e-01
7.65413567e-02 -2.75462031e-01 8.15081000e-01 6.74766243e-01
1.41140091e+00 -3.63503546e-01 -2.39392191e-01 7.06690729e-01
1.50805366e+00 8.79455954e-02 4.55453843e-01 -2.43050978e-02
6.27111197e-01 6.76772535e-01 1.89529806e-01 3.48850429e-01
-1.73167869e-01 8.35573077e-01 2.55000353e-01 -2.35976413e-01
3.57414037e-01 1.02848262e-01 3.50283384e-02 2.80907691e-01
-1.01772882e-02 -2.32689992e-01 -8.16886246e-01 2.50284135e-01
-1.82277083e+00 -6.75088227e-01 8.36389661e-02 2.70801044e+00
9.15139258e-01 -1.75966263e-01 1.49241194e-01 3.30731459e-02
6.40238762e-01 -1.67812735e-01 -8.17387223e-01 -7.52609730e-01
6.54644147e-02 1.52548879e-01 4.09897596e-01 5.23091435e-01
-9.78193700e-01 3.94145012e-01 5.87768650e+00 5.37996769e-01
-1.27365756e+00 -2.71061212e-01 9.94858563e-01 -1.92134947e-01
3.73769663e-02 -3.03863376e-01 -8.97878408e-01 5.49651086e-01
1.43470514e+00 1.07728697e-01 4.44976896e-01 5.65313816e-01
9.97183993e-02 -7.33624548e-02 -1.54935122e+00 6.98949754e-01
-5.14540449e-02 -1.26039708e+00 -4.25318163e-03 2.63353407e-01
3.24220777e-01 -1.80734992e-02 2.08953992e-01 8.47576782e-02
8.36488903e-02 -1.39505851e+00 -1.20774187e-01 5.27954102e-01
6.85533047e-01 -8.56387377e-01 9.48047042e-01 5.92145443e-01
-3.92647147e-01 -2.39413887e-01 -8.96016240e-01 1.09504890e-02
-1.88428178e-01 9.47559714e-01 -1.12167466e+00 1.05070233e-01
4.01518047e-01 2.79376298e-01 -5.49893856e-01 1.04220533e+00
9.30464268e-02 4.56928015e-01 -4.30371642e-01 -3.39529425e-01
1.69910520e-01 -3.23705703e-01 8.53964835e-02 1.18367422e+00
5.29103577e-01 2.21946109e-02 -6.93796389e-03 8.36902738e-01
-1.65366039e-01 2.86442220e-01 -5.81810176e-01 -3.51420492e-02
1.62643507e-01 1.12049890e+00 -4.91574615e-01 -2.06049696e-01
-2.61149704e-01 5.71681499e-01 7.72450209e-01 3.62019420e-01
-3.65573734e-01 -4.65528697e-01 7.77441680e-01 4.69221622e-02
4.26267266e-01 2.98090994e-01 -2.09321082e-01 -8.32628787e-01
1.10235013e-01 -8.54308546e-01 5.08714676e-01 -1.39431924e-01
-1.37096739e+00 4.60133314e-01 -3.35164815e-01 -1.00916672e+00
-6.07854486e-01 -8.90108168e-01 -4.33483303e-01 1.34795201e+00
-1.14647388e+00 -6.56388164e-01 1.94222063e-01 7.12961331e-02
4.27152403e-02 -2.65867949e-01 1.27995872e+00 2.65789360e-01
-8.18641841e-01 6.50644779e-01 4.87967163e-01 6.92714527e-02
6.05464518e-01 -1.13923514e+00 2.31044441e-01 2.48001754e-01
-6.15781620e-02 7.68154323e-01 6.26792848e-01 -3.91316980e-01
-1.38367569e+00 -1.19599223e+00 1.07295489e+00 -1.87612221e-01
4.67892110e-01 -6.52564406e-01 -1.15528762e+00 7.27108061e-01
-4.22321975e-01 4.73587036e-01 9.99273896e-01 4.83243674e-01
-2.80717552e-01 -1.64523378e-01 -1.19936299e+00 3.01595598e-01
4.57387388e-01 -3.06184798e-01 -9.18060243e-02 5.29862583e-01
4.24423605e-01 -2.33191743e-01 -8.84353340e-01 2.92030033e-02
6.43872976e-01 -7.09208131e-01 9.33602989e-01 -1.01686120e+00
3.46960515e-01 -1.48873866e-01 3.27947512e-02 -1.75871027e+00
-4.28444177e-01 -1.45084739e-01 1.96389318e-01 8.30951810e-01
8.61973166e-01 -8.01908731e-01 8.98941696e-01 8.57315540e-01
2.16337517e-01 -9.79167521e-01 -1.04205120e+00 -5.61489880e-01
2.95237452e-01 8.20245370e-02 6.40742481e-01 7.36016452e-01
-2.10469291e-02 5.44602096e-01 -4.10863519e-01 3.29834700e-01
2.61836648e-01 -1.96956955e-02 4.62208509e-01 -1.62705112e+00
-7.15241313e-01 -2.67218649e-01 -8.43144715e-01 -4.05841023e-01
1.46460235e-02 -1.10537922e+00 1.40722647e-01 -9.83180881e-01
1.97258130e-01 -8.50890815e-01 -2.55769104e-01 6.28507972e-01
-3.55950832e-01 -3.82767767e-02 -1.64767668e-01 -3.62710476e-01
1.00568749e-01 1.40655801e-01 7.10948825e-01 -1.21210992e-01
-4.46418196e-01 2.89387554e-01 -6.47856414e-01 6.69947326e-01
7.18072772e-01 -7.49777436e-01 -2.91637361e-01 -3.93234640e-01
4.36743945e-01 2.22555295e-01 2.43414000e-01 -8.89919341e-01
2.49365494e-01 -2.08850913e-02 8.42657745e-01 -8.24209377e-02
3.25720400e-01 -8.42253685e-01 3.85055035e-01 6.75467789e-01
-7.44320810e-01 -1.11681998e-01 1.27596274e-01 5.40294290e-01
-1.07352911e-02 -6.36507511e-01 7.08276510e-01 1.18830940e-02
-4.28183451e-02 2.67455906e-01 -1.54535696e-01 -1.53643519e-01
8.28853250e-01 1.96858257e-01 -1.50496393e-01 -1.21686004e-01
-9.13940310e-01 7.43476525e-02 2.36838311e-01 1.68646812e-01
3.41358811e-01 -1.09662700e+00 -7.69352555e-01 6.31464541e-01
4.52988409e-02 5.27739227e-02 3.22042346e-01 5.78862250e-01
-3.41168910e-01 5.17848015e-01 -2.34065071e-01 -5.26022792e-01
-1.17854619e+00 6.99029326e-01 4.60848302e-01 -1.85367882e-01
-7.91387856e-01 7.09570587e-01 3.97960007e-01 -5.66040993e-01
2.05983907e-01 -3.94258738e-01 -4.01563674e-01 1.48130253e-01
6.51147366e-01 4.96078543e-02 3.52141142e-01 -3.87308985e-01
-3.12554628e-01 6.66017413e-01 -1.77425846e-01 3.42663795e-01
1.68122208e+00 5.77212393e-01 -2.29869798e-01 5.52328944e-01
1.46547937e+00 -4.16820586e-01 -1.11800086e+00 -1.56670600e-01
1.47926375e-01 -1.85667634e-01 -6.37702793e-02 -8.77165079e-01
-1.06473374e+00 1.10906029e+00 8.46699476e-01 8.01034868e-02
9.54876006e-01 -1.17079534e-01 3.33390027e-01 7.39876509e-01
1.06960177e-01 -8.32466781e-01 -5.41084588e-01 3.89557719e-01
6.33050442e-01 -1.21260881e+00 -1.62137091e-01 -3.07249427e-01
-3.55300754e-01 1.24114323e+00 1.73254147e-01 -2.88755685e-01
6.84487998e-01 3.23854506e-01 4.80269901e-02 -1.89334512e-01
-8.79179955e-01 2.29594573e-01 4.65556115e-01 6.00547016e-01
6.83036804e-01 2.37979189e-01 -3.36266547e-01 6.52695060e-01
-8.93664584e-02 -1.14801135e-02 4.67866600e-01 5.24736762e-01
-5.20587385e-01 -1.31483793e+00 -2.45089680e-01 1.02624953e+00
-4.43377078e-01 2.26534978e-02 -1.37518376e-01 4.34414238e-01
2.13115618e-01 4.13826555e-01 2.41005227e-01 -3.35249417e-02
1.78190544e-01 3.51246685e-01 1.95584118e-01 -8.88919055e-01
-7.09601820e-01 -1.04913950e-01 -7.37856254e-02 -3.54088783e-01
4.17508297e-02 -6.04833722e-01 -1.24940288e+00 8.51722434e-02
-2.67768349e-03 1.97610542e-01 8.52652073e-01 9.38065827e-01
4.37554151e-01 5.55732429e-01 3.37256670e-01 -7.27908492e-01
-8.48127365e-01 -8.30646455e-01 -7.95847476e-01 2.10535213e-01
5.31379342e-01 -4.90142435e-01 -2.42428929e-01 -5.45290299e-02] | [8.265966415405273, 4.472285747528076] |
925abaad-c04f-4e7e-a720-ac71d8d6ed18 | full-reconstruction-of-non-stationary-strand | 1610.05057 | null | http://arxiv.org/abs/1610.05057v2 | http://arxiv.org/pdf/1610.05057v2.pdf | Full Reconstruction of Non-Stationary Strand-Symmetric Models on Rooted Phylogenies | Understanding the evolutionary relationship among species is of fundamental
importance to the biological sciences. The location of the root in any
phylogenetic tree is critical as it gives an order to evolutionary events. None
of the popular models of nucleotide evolution used in likelihood or Bayesian
methods are able to infer the location of the root without exogenous
information. It is known that the most general Markov models of nucleotide
substitution can also not identify the location of the root or be fitted to
multiple sequence alignments with less than three sequences. We prove that the
location of the root and the full model can be identified and statistically
consistently estimated for a non-stationary, strand-symmetric substitution
model given a multiple sequence alignment with two or more sequences. We also
generalise earlier work to provide a practical means of overcoming the
computationally intractable problem of labelling hidden states in a
phylogenetic model. | [] | 2016-11-13 | null | null | null | null | ['multiple-sequence-alignment'] | ['medical'] | [ 8.30628574e-01 -2.42883861e-01 -4.29693609e-01 -1.64883405e-01
-2.60807902e-01 -7.43778646e-01 4.32283610e-01 1.68803185e-01
-5.55529237e-01 1.13891566e+00 -2.16268525e-01 -9.78903949e-01
1.30979000e-02 -2.41643772e-01 -5.79114676e-01 -9.68136847e-01
-1.82297274e-01 7.85553336e-01 6.56031132e-01 1.88798662e-02
5.00603259e-01 6.57999694e-01 -1.09771919e+00 -9.73295420e-02
5.60483575e-01 6.17238842e-02 8.47783506e-01 1.12749374e+00
-1.78150594e-01 -7.88535085e-03 -4.37348247e-01 -2.43816629e-01
1.16914958e-01 -9.99038160e-01 -1.10292804e+00 1.21318689e-02
-3.41103673e-02 -2.12329954e-01 1.33344203e-01 9.19997334e-01
3.75230700e-01 -3.55091244e-01 6.61655843e-01 -1.05258846e+00
-2.25082904e-01 4.47002620e-01 -6.13203049e-01 4.75238323e-01
1.92322463e-01 3.93000096e-02 1.02734685e+00 -2.20205545e-01
9.53442097e-01 1.19469082e+00 8.70405734e-01 3.09366584e-01
-1.55178928e+00 -1.72617093e-01 -1.83158636e-01 2.64943421e-01
-1.21335185e+00 -3.38252932e-01 2.48631522e-01 -5.55220425e-01
9.54666615e-01 3.35017085e-01 6.79808021e-01 9.78329360e-01
4.38504130e-01 1.99191451e-01 1.21749485e+00 -6.14018857e-01
3.12000304e-01 -5.71047306e-01 -5.92754148e-02 5.59815407e-01
3.41967732e-01 -1.14717828e-02 -3.79726827e-01 -8.67484331e-01
7.48730600e-01 -2.00552508e-01 -2.54486144e-01 -2.07490072e-01
-1.15892518e+00 8.52265358e-01 -4.85808210e-04 3.17806810e-01
-4.89550143e-01 3.61767471e-01 3.11817884e-01 -1.99515551e-01
8.91767666e-02 3.23811859e-01 -8.12105656e-01 -4.21732575e-01
-8.70297253e-01 -9.88146588e-02 9.64906633e-01 5.85236073e-01
7.11093426e-01 -1.43192843e-01 7.94594705e-01 3.43296111e-01
3.41706902e-01 1.57363161e-01 2.12960079e-01 -1.02507293e+00
-4.07883376e-01 1.63632214e-01 3.11500996e-01 -3.91551912e-01
-2.76277244e-01 -2.06138447e-01 -4.81568485e-01 -9.06416774e-02
6.84693038e-01 1.40885890e-01 -8.89613092e-01 1.94722140e+00
6.44679427e-01 5.65773547e-01 3.52630429e-02 4.26153839e-01
1.26846194e-01 7.56541610e-01 9.77238268e-02 -7.67753720e-01
1.46794391e+00 -3.18986267e-01 -5.05362034e-01 -3.78605664e-01
3.50676090e-01 -1.03531694e+00 3.03668082e-01 2.90629476e-01
-7.33690321e-01 -5.57945343e-03 -9.67361629e-01 7.00146658e-03
5.17044365e-02 -4.14769679e-01 7.92624831e-01 6.71203315e-01
-1.01627851e+00 7.40792453e-01 -1.16435969e+00 -5.93152702e-01
-3.18185836e-01 2.97312140e-01 -2.98408866e-01 2.05160499e-01
-9.94468451e-01 1.25352216e+00 7.25613534e-01 2.66012400e-01
-7.42174089e-01 -2.13815629e-01 -3.90885323e-01 -2.87241340e-01
3.52930576e-01 -6.38369262e-01 1.25177729e+00 -1.14267242e+00
-1.15063870e+00 9.36317980e-01 -8.16920221e-01 -4.12661880e-01
3.98382932e-01 2.23211616e-01 6.38211071e-02 2.18214229e-01
1.31831020e-02 4.07239318e-01 4.23926651e-01 -9.95064497e-01
-6.33329272e-01 -2.85282940e-01 -4.88684624e-01 -2.83190701e-02
5.39180040e-01 2.52016783e-01 -1.74288154e-01 -3.20020229e-01
5.60863495e-01 -1.21272075e+00 -6.06256664e-01 2.29521412e-02
-3.76016915e-01 -9.16586295e-02 4.73297775e-01 -9.87735271e-01
8.81777287e-01 -1.84107208e+00 4.94516730e-01 2.01747585e-02
-8.70334953e-02 1.31029949e-01 -8.14600568e-03 1.06622577e+00
-3.05688471e-01 3.39377493e-01 -7.14064956e-01 2.43213773e-01
-1.95776373e-01 8.13529253e-01 -2.89012611e-01 1.05356932e+00
-3.56511995e-02 4.67078567e-01 -1.01265323e+00 -4.68726933e-01
5.69836535e-02 5.76117814e-01 -3.43698084e-01 7.42932502e-03
-1.51457742e-01 4.28031832e-01 -2.50341266e-01 1.68709084e-01
6.19054556e-01 -3.86852384e-01 1.05996120e+00 3.56656998e-01
-5.09868383e-01 6.55227363e-01 -1.06402588e+00 1.22813785e+00
1.92063436e-01 5.12688398e-01 1.71030536e-01 -7.31378496e-01
7.41636813e-01 5.02046287e-01 4.10993129e-01 2.73433805e-01
-8.01100582e-02 4.16288078e-01 6.81489170e-01 -1.78562909e-01
1.87540337e-01 -6.54132485e-01 1.75750330e-01 3.35258037e-01
-1.89374074e-01 1.77560940e-01 4.59026266e-03 -2.54622605e-02
9.42322075e-01 2.61451870e-01 9.72596526e-01 -4.60844934e-01
2.81442285e-01 4.38266657e-02 1.15177429e+00 7.31875420e-01
-3.11155736e-01 2.82805145e-01 4.86393243e-01 -4.03272182e-01
-1.55905390e+00 -7.18320787e-01 -4.51473534e-01 9.77096260e-01
6.96995761e-03 -3.59312564e-01 -6.56573713e-01 9.37567949e-02
-1.62173942e-01 7.95002580e-01 -4.35885668e-01 -1.04311541e-01
-5.45640171e-01 -1.12621534e+00 6.95172846e-01 8.00060853e-02
-1.39524028e-01 -8.16239059e-01 -1.00866973e+00 4.47062463e-01
-4.94227409e-01 -7.60448277e-01 -2.04171479e-01 6.84570312e-01
-1.00852489e+00 -1.12463307e+00 -4.52145666e-01 -7.60148048e-01
5.48675716e-01 1.51748687e-01 7.49547899e-01 5.26785076e-01
-4.57280278e-01 -1.57562435e-01 -6.98144287e-02 -9.10237655e-02
-8.97824943e-01 -5.42033799e-02 1.34521529e-01 -5.04059553e-01
4.17812288e-01 -6.83181226e-01 -4.77681816e-01 5.72459280e-01
-8.26256514e-01 1.65283665e-01 1.79543018e-01 9.99582827e-01
4.05321419e-01 1.00808665e-02 2.65889406e-01 -6.30214810e-01
8.38864744e-02 -5.31116903e-01 -6.62549555e-01 5.33335626e-01
-1.37800857e-01 1.90094173e-01 3.64321381e-01 -3.16643029e-01
-7.15195715e-01 3.58733535e-01 -4.95882630e-01 2.95056343e-01
-2.75287151e-01 6.79785073e-01 4.27222985e-04 1.61372736e-01
3.56219798e-01 5.45805335e-01 -3.27049531e-02 -5.82839668e-01
1.32582277e-01 5.80379128e-01 4.59346235e-01 -6.88634813e-01
3.29888165e-01 4.64749604e-01 6.35918319e-01 -1.29518855e+00
-1.37240753e-01 -5.44049203e-01 -1.23459399e+00 1.68869480e-01
7.34025002e-01 -4.27734882e-01 -9.78425741e-01 4.23285663e-01
-1.34691870e+00 -1.76759034e-01 3.09904665e-01 4.67509180e-01
-9.03969526e-01 1.20899415e+00 -5.70383251e-01 -1.14863729e+00
1.97773010e-01 -1.09539843e+00 7.75680959e-01 6.30935356e-02
-3.84564310e-01 -9.99204516e-01 5.13757706e-01 -6.30288571e-03
-8.71859025e-03 2.83606559e-01 1.21335888e+00 -6.46830261e-01
-2.89986342e-01 -4.05841880e-02 4.01493549e-01 -3.59318495e-01
3.85361969e-01 8.32530439e-01 -5.81001818e-01 -2.17914373e-01
1.43117130e-01 1.62651137e-01 7.26967394e-01 6.31421387e-01
3.18404794e-01 -2.36186355e-01 -6.33478999e-01 4.76078928e-01
1.28562748e+00 5.52885294e-01 5.57051182e-01 4.80513245e-01
1.72384024e-01 7.80866802e-01 3.41676563e-01 3.05661231e-01
-2.78219245e-02 7.74118125e-01 2.69546628e-01 2.76940048e-01
4.61565197e-01 -2.34700963e-01 1.59106225e-01 7.73282945e-01
-3.42801288e-02 -3.29534292e-01 -1.31185901e+00 6.41986966e-01
-1.87954473e+00 -1.16427839e+00 -4.45798576e-01 2.46318626e+00
1.19535232e+00 -4.57816198e-02 2.31555223e-01 2.90667675e-02
1.08287823e+00 -2.29231954e-01 -8.04832041e-01 -5.43518424e-01
-1.90072939e-01 -5.87917380e-02 7.98336864e-01 8.82188082e-01
-7.14737952e-01 9.56453979e-01 8.57935715e+00 4.75760669e-01
-9.49813366e-01 -1.97819963e-01 2.82521933e-01 3.43630463e-01
-3.27530950e-01 7.73494303e-01 -9.78940010e-01 4.19876963e-01
1.34619081e+00 -2.59632081e-01 2.90761679e-01 5.89632809e-01
3.33998442e-01 -3.53342116e-01 -9.92450118e-01 3.30192715e-01
-3.69600087e-01 -1.02247632e+00 -2.91618586e-01 4.75547105e-01
3.49369943e-01 -2.44822819e-02 -2.95600802e-01 -6.41141593e-01
8.62244487e-01 -8.94598365e-01 3.95921201e-01 1.64902449e-01
3.37841690e-01 -7.83140719e-01 4.32366520e-01 8.71617258e-01
-9.26821053e-01 3.26677412e-01 -8.28380704e-01 -2.69410133e-01
7.67033875e-01 2.36230910e-01 -1.46841574e+00 3.30369830e-01
3.42197597e-01 3.52920711e-01 -2.52665639e-01 1.24410307e+00
-4.38617319e-01 9.23325062e-01 -6.82691097e-01 4.52736802e-02
3.79593104e-01 -4.09002602e-01 6.13544822e-01 1.02608645e+00
3.60367835e-01 8.20146501e-02 2.40070775e-01 2.28751123e-01
6.46585941e-01 -5.72601147e-02 -3.47176403e-01 -3.06696445e-01
5.93258858e-01 6.00253224e-01 -1.28048110e+00 -2.47406974e-01
-5.49741611e-02 1.03952909e+00 3.28788324e-03 1.21711917e-01
-5.94212949e-01 2.80485526e-02 8.65148842e-01 -1.66650712e-01
6.94897652e-01 -5.77274680e-01 1.74985975e-01 -8.84010255e-01
-5.29453635e-01 -8.39207292e-01 4.85845119e-01 -5.84267855e-01
-1.08596933e+00 2.37748578e-01 7.21052811e-02 -4.58166033e-01
-7.01552391e-01 -5.81718564e-01 -1.95304587e-01 1.24437344e+00
-1.14046812e+00 -6.29582644e-01 6.42620683e-01 -7.53518641e-02
4.09420311e-01 4.19941306e-01 9.10577774e-01 -3.42999905e-01
-6.48467422e-01 -5.73550649e-02 5.57338774e-01 -2.85587132e-01
4.30179566e-01 -1.39842653e+00 8.45136583e-01 1.05554402e+00
-8.14296752e-02 1.09631491e+00 1.37177730e+00 -1.18325245e+00
-1.44249523e+00 -4.17711467e-01 1.15421188e+00 -2.73302704e-01
7.42961526e-01 -3.36949348e-01 -1.09879231e+00 1.00503516e+00
9.46152806e-02 -4.67058301e-01 1.04947591e+00 -4.29440886e-02
-2.80197918e-01 7.10579216e-01 -9.98131394e-01 6.42643213e-01
8.89806688e-01 -4.83655989e-01 -7.93855429e-01 2.26017594e-01
5.04539967e-01 9.71154422e-02 -8.33768368e-01 1.37071028e-01
8.74710202e-01 -8.07918549e-01 1.03799045e+00 -9.17918622e-01
5.91835380e-02 -7.21093118e-01 -2.61012971e-01 -8.50277245e-01
-4.74463344e-01 -7.68902838e-01 4.36161011e-01 1.01868856e+00
2.06546292e-01 -7.33653128e-01 7.02629983e-01 2.45954081e-01
5.64174203e-04 -1.17491685e-01 -1.54392576e+00 -8.31148446e-01
1.78107291e-01 9.87248495e-02 2.65280068e-01 9.50861037e-01
-1.30121455e-01 2.76654273e-01 -7.84622908e-01 2.68039018e-01
8.07853758e-01 1.57456338e-01 6.18070364e-01 -1.33834577e+00
-6.59420431e-01 -3.20539176e-01 -5.97432435e-01 -1.06926286e+00
9.57680717e-02 -5.30352414e-01 4.40219104e-01 -1.40833080e+00
4.30957407e-01 -6.61110133e-02 6.79174140e-02 3.03155750e-01
-3.24532807e-01 -4.91297692e-02 -1.53448537e-01 3.60763967e-01
1.82899043e-01 1.08576708e-01 6.58483684e-01 5.18944979e-01
3.69957983e-01 -7.44664744e-02 -2.35132858e-01 8.99893105e-01
6.70598090e-01 -1.03921843e+00 -1.43419415e-01 9.64284465e-02
5.08546054e-01 2.47707546e-01 1.23362862e-01 -1.41950145e-01
8.08472335e-02 -6.51986778e-01 8.08532760e-02 -7.21071005e-01
3.29273760e-01 -8.52433383e-01 1.02499783e+00 1.26460457e+00
-1.78656414e-01 3.99756789e-01 7.48998746e-02 6.48389578e-01
4.73818123e-01 -1.00020504e+00 8.33301365e-01 -3.70196491e-01
-7.12081134e-01 -2.17290491e-01 -9.34227765e-01 -4.70389992e-01
1.05538881e+00 -2.92879611e-01 -3.27442437e-01 -1.41630828e-01
-6.24462306e-01 -6.86234087e-02 1.10852098e+00 -9.15974900e-02
1.80503875e-01 -6.31289124e-01 -6.30093396e-01 -1.91236258e-01
-2.41008684e-01 -5.56113720e-01 7.76361749e-02 6.50765479e-01
-1.25435436e+00 4.97443080e-01 -3.98604780e-01 -7.85971642e-01
-1.82795799e+00 8.73945594e-01 3.34988713e-01 7.69786537e-02
-4.88722473e-01 5.81610143e-01 2.61789002e-02 -3.73044044e-01
-2.50050366e-01 2.95981675e-01 3.14403147e-01 -4.56615463e-02
5.12796640e-01 3.15054983e-01 -6.22916639e-01 -1.22218299e+00
-7.03125298e-01 4.54919815e-01 -1.78615034e-01 -3.57248008e-01
1.37774718e+00 -2.83666402e-01 -5.69428682e-01 9.14468467e-01
7.97050714e-01 -7.70594999e-02 -1.21951306e+00 -5.56958392e-02
4.62365389e-01 -5.22852063e-01 -4.94458735e-01 -3.25687498e-01
-1.05284728e-01 8.85305703e-01 2.13954136e-01 2.20280647e-01
5.51597655e-01 1.02797861e-03 3.22709531e-01 3.79129082e-01
3.59265417e-01 -4.67779160e-01 -6.05995595e-01 3.43032897e-01
3.88147384e-01 -6.79217398e-01 2.31410339e-01 -5.53192973e-01
-1.71081230e-01 1.15850794e+00 -8.01368430e-02 2.66722888e-01
2.16890872e-01 2.81154931e-01 -5.52855693e-02 8.42467099e-02
-1.21307886e+00 -1.17486574e-01 -1.77373558e-01 5.46227813e-01
7.16415107e-01 2.58025508e-02 -8.22138846e-01 -4.07062411e-01
-2.25316092e-01 -2.84559965e-01 7.88768768e-01 1.25568426e+00
-8.24045002e-01 -1.56251574e+00 -4.50115502e-01 -1.73745751e-01
-7.74940670e-01 -2.88532197e-01 -8.37992191e-01 5.86836815e-01
-1.96452096e-01 6.99054360e-01 4.08448540e-02 6.43586963e-02
-3.94992262e-01 5.92017174e-01 5.44865072e-01 -4.94217157e-01
-4.04286414e-01 5.91212749e-01 1.81990847e-01 1.58408791e-01
-5.29433250e-01 -1.05709755e+00 -1.04575717e+00 -6.33909583e-01
-6.68299258e-01 6.46797717e-01 8.77741873e-01 8.71366203e-01
-2.89007332e-02 3.40627432e-02 3.49769741e-01 -5.75541854e-01
-6.99107409e-01 -7.24995494e-01 -4.88059670e-01 -5.49244732e-02
4.01312172e-01 -3.49964321e-01 -5.23332834e-01 3.17959726e-01] | [4.865238189697266, 5.160327434539795] |
039004cb-3ea9-4c75-ad19-61d09a682277 | t-cvae-transformer-based-conditioned | null | null | https://www.ijcai.org/proceedings/2019/727 | https://www.ijcai.org/proceedings/2019/0727.pdf | T-CVAE: Transformer-Based Conditioned Variational Autoencoder for Story Completion | Story completion is a very challenging task of generating the missing plot for an incomplete story, which requires not only understanding but also inference of the given contextual clues. In this paper, we present a novel conditional variational autoencoder based on Transformer for missing plot generation. Our model uses shared attention layers for encoder and decoder, which make the most of the contextual clues, and a latent variable for learning the distribution of coherent story plots. Through drawing samples from the learned distribution, diverse reasonable plots can be generated. Both automatic and manual evaluations show that our model generates better story plots than state-of-the-art models in terms of readability, diversity and coherence. | ['Xiaojun Wan', 'Tianming Wang'] | 2019-07-01 | null | null | null | international-joint-conference-on-artificial-2 | ['story-completion'] | ['natural-language-processing'] | [-3.46250758e-02 4.51313972e-01 -1.23668574e-01 -2.60055751e-01
-8.85357201e-01 -4.29687887e-01 7.80486465e-01 -8.83595049e-02
3.77431780e-01 1.01608479e+00 1.05714464e+00 1.18779860e-01
6.11419678e-02 -9.40208912e-01 -1.03506804e+00 -4.22503710e-01
5.42889059e-01 5.76711059e-01 -1.92387000e-01 -8.44324455e-02
1.10827558e-01 -2.82007992e-01 -1.48755360e+00 7.40585327e-01
1.10664678e+00 4.27578688e-01 6.64094627e-01 1.07930672e+00
-1.74433276e-01 1.38421702e+00 -9.97362614e-01 -4.89385247e-01
-3.93392682e-01 -7.94469059e-01 -5.32865822e-01 4.02705848e-01
3.03797722e-01 -8.84172916e-01 -5.96434951e-01 5.43905318e-01
4.22697186e-01 4.31064703e-02 1.25277972e+00 -1.24576402e+00
-1.35994744e+00 1.22077894e+00 -4.17473704e-01 -3.87669206e-01
5.20241022e-01 3.35324526e-01 1.37823534e+00 -9.34899628e-01
1.01775014e+00 9.75191951e-01 3.63461971e-01 4.47758019e-01
-1.29544413e+00 -4.20103610e-01 3.36331315e-02 2.45154336e-01
-1.28662336e+00 -2.84461141e-01 8.77073050e-01 -7.92886674e-01
7.60808349e-01 7.49500245e-02 9.21975255e-01 1.61321557e+00
2.46081218e-01 1.09145987e+00 4.34880286e-01 -3.33281398e-01
2.79103070e-01 1.87494904e-01 -3.05226654e-01 4.58644539e-01
7.39541873e-02 -3.28057766e-01 -8.31757486e-01 7.55978674e-02
1.08658743e+00 -5.64434230e-02 -3.30142558e-01 -2.06876189e-01
-1.22971630e+00 9.73178566e-01 3.21725607e-01 4.56503071e-02
-6.35844052e-01 3.96673143e-01 6.97604716e-02 -2.84960777e-01
5.21551132e-01 4.36846882e-01 8.26994479e-02 -2.95783520e-01
-1.25779927e+00 7.27552414e-01 6.87971890e-01 1.16642475e+00
3.60724002e-01 3.66023213e-01 -9.34360087e-01 7.59250224e-01
3.74403149e-01 6.82262003e-01 1.14627451e-01 -8.78191650e-01
6.24584854e-01 4.98694420e-01 3.54767561e-01 -8.25680196e-01
7.61299953e-02 -4.95152593e-01 -1.04848540e+00 2.34657347e-01
2.05628440e-01 -3.64167899e-01 -9.76329088e-01 1.76367939e+00
-7.57203996e-02 4.51160483e-02 1.64443236e-02 6.93958879e-01
1.20048392e+00 1.22707999e+00 -3.67127329e-01 3.01631242e-02
1.04923582e+00 -1.05342579e+00 -1.23606670e+00 -1.42408222e-01
-8.42513144e-02 -7.20829427e-01 1.32824016e+00 2.80946732e-01
-1.33999407e+00 -5.29175460e-01 -1.23412859e+00 -4.03042436e-01
1.62703276e-01 6.22013450e-01 3.98856938e-01 1.26948180e-02
-9.25690472e-01 4.54259813e-01 -5.05057633e-01 -5.97740300e-02
5.61823606e-01 -3.71652991e-01 -3.55013721e-02 -1.45311102e-01
-9.51798677e-01 6.65706933e-01 2.03996226e-01 -8.47659782e-02
-1.28835630e+00 -8.39477241e-01 -1.02520633e+00 3.68715376e-01
2.05736935e-01 -9.26442921e-01 1.37225366e+00 -4.18974876e-01
-1.57185233e+00 4.37831253e-01 -3.24928224e-01 -3.01826864e-01
6.99922621e-01 -5.18619120e-01 3.11236810e-02 -8.77359733e-02
2.93279737e-01 8.44494522e-01 9.30540919e-01 -1.30935240e+00
-1.13088891e-01 4.28255618e-01 2.92600435e-03 2.26815164e-01
-7.22540691e-02 -5.04808366e-01 -3.18258345e-01 -7.00430512e-01
-4.54831362e-01 -4.62876678e-01 1.10006388e-02 -5.44199310e-02
-9.44498777e-01 -2.33867705e-01 5.18402576e-01 -1.16718459e+00
1.39979351e+00 -1.93608022e+00 5.12549758e-01 -1.71913937e-01
4.89463806e-01 -5.15713573e-01 -5.27938753e-02 7.54546523e-01
1.87372789e-01 1.22271255e-01 -1.28547177e-01 -4.14319307e-01
3.66390049e-01 -5.46059720e-02 -6.91117764e-01 -1.62214246e-02
2.76334465e-01 9.76578116e-01 -7.30460346e-01 -5.25008261e-01
3.11917156e-01 6.66619539e-01 -6.92341447e-01 6.48129523e-01
-8.97475958e-01 3.90347451e-01 -1.31788120e-01 1.39896706e-01
4.68303382e-01 -5.87404430e-01 -3.12442072e-02 2.34310627e-01
3.21042985e-02 1.91222921e-01 -9.87625003e-01 1.89944947e+00
-3.09197098e-01 1.20925379e+00 -5.52735686e-01 -5.28884567e-02
1.02744174e+00 4.72775728e-01 -1.50758401e-01 -2.95274436e-01
1.52952731e-01 -2.73617119e-01 -4.79778230e-01 -6.80441558e-01
9.00768697e-01 7.75319412e-02 -1.35641664e-01 8.20436954e-01
2.25149527e-01 -3.99118274e-01 2.39947021e-01 6.07658386e-01
8.47066343e-01 3.99262100e-01 1.49401158e-01 1.26119614e-01
-1.38504967e-01 -1.61612302e-03 2.60771751e-01 7.31542110e-01
4.22537863e-01 1.21203566e+00 1.04454899e+00 -1.60127327e-01
-1.84456706e+00 -1.43616927e+00 3.28722388e-01 5.51661372e-01
-1.64422333e-01 -7.21772432e-01 -8.27799320e-01 -1.18535839e-01
-2.76846766e-01 1.48921525e+00 -8.89256358e-01 3.70935835e-02
-1.94223315e-01 -2.45284408e-01 3.23726147e-01 7.87722707e-01
2.96379954e-01 -1.16537452e+00 -6.15217686e-01 1.51415467e-01
-5.68134427e-01 -8.18872273e-01 -5.30366898e-01 -3.45734537e-01
-4.19445366e-01 -8.49489987e-01 -1.03636670e+00 -5.29021502e-01
5.41451633e-01 -4.50339960e-03 1.32602417e+00 -2.45997030e-02
-1.60090655e-01 -1.15446851e-01 -1.83322102e-01 -4.31838602e-01
-5.56172848e-01 9.63411182e-02 -6.03150725e-01 -1.17849134e-01
-3.11451852e-01 -6.18782401e-01 -3.49983603e-01 -2.72949070e-01
-9.20250297e-01 1.01424038e+00 4.68388617e-01 1.01883698e+00
4.77301627e-01 -2.26505119e-02 1.75896063e-01 -7.96820939e-01
9.72786605e-01 -5.71257889e-01 -1.75204933e-01 2.83495605e-01
-1.34496063e-01 3.42601717e-01 6.21930242e-01 -4.60522592e-01
-1.48228979e+00 -3.09558243e-01 3.90535593e-02 -4.21271741e-01
-8.91843736e-02 5.47730088e-01 -3.32971245e-01 1.23163712e+00
6.79251969e-01 2.78719723e-01 -3.29432964e-01 -3.04288805e-01
8.15664053e-01 4.23289478e-01 6.48969233e-01 -3.55529338e-01
6.74418330e-01 2.37988457e-01 -4.78611052e-01 -7.44544148e-01
-8.88957500e-01 3.30949515e-01 -6.60705805e-01 -6.08983278e-01
1.20624411e+00 -1.13493681e+00 -1.71219870e-01 4.22561347e-01
-1.51814878e+00 -7.31811166e-01 -5.94299138e-01 2.63423413e-01
-7.79294312e-01 -3.81678045e-01 -6.27352715e-01 -8.48017693e-01
-1.43597126e-01 -9.54275906e-01 1.11758602e+00 4.66323793e-01
-7.66312718e-01 -8.29440415e-01 3.64158392e-01 2.59570718e-01
1.90044001e-01 7.28363693e-01 1.19632161e+00 -9.27573368e-02
-9.79350209e-01 -1.15477219e-01 -2.07588106e-01 -2.23859787e-01
-1.41683534e-01 5.17117381e-01 -1.08196223e+00 2.01061815e-01
-4.84476686e-01 -5.45539200e-01 8.50179076e-01 9.29943383e-01
8.13502669e-01 -5.04141808e-01 -1.43270537e-01 5.42211354e-01
1.34498715e+00 1.67611856e-02 9.71206069e-01 7.20587745e-02
9.67280686e-01 4.47544783e-01 2.59988546e-01 1.02265751e+00
6.47700846e-01 3.21094811e-01 3.39400172e-01 1.45268915e-02
-3.48087788e-01 -1.00258005e+00 3.43383849e-01 6.76262259e-01
1.28497049e-01 -8.05325508e-01 -6.80443943e-01 8.77643883e-01
-2.04741383e+00 -1.41529274e+00 -1.95840269e-01 1.61480343e+00
8.24601710e-01 1.16308428e-01 -5.21952286e-02 3.66350263e-02
7.17657924e-01 6.59711957e-01 -6.53240800e-01 -3.00387263e-01
-2.91940719e-01 -7.64572322e-02 -1.78063691e-01 6.61647201e-01
-6.67243183e-01 9.06569362e-01 6.71279335e+00 7.74635434e-01
-5.66448271e-01 -3.59219946e-02 7.53737628e-01 -4.90580559e-01
-1.01775062e+00 -1.77864600e-02 -3.22351992e-01 3.09711576e-01
6.18268967e-01 -2.28809923e-01 3.88465375e-01 8.13424408e-01
3.51370960e-01 -1.59539148e-01 -1.18446851e+00 9.43485320e-01
3.49413097e-01 -1.93757153e+00 2.24208742e-01 -7.31192678e-02
1.11597300e+00 -6.60967648e-01 9.56742242e-02 2.32260466e-01
7.74947107e-01 -1.40095901e+00 1.39449942e+00 1.10797548e+00
9.94516730e-01 -9.13096607e-01 5.25115311e-01 3.70916277e-01
-9.87792313e-01 3.73137355e-01 -3.90785068e-01 -2.82299876e-01
3.62140030e-01 6.14441454e-01 -9.45414066e-01 2.20770389e-01
3.01532775e-01 8.91351223e-01 -5.46780229e-01 7.88099468e-01
-9.81307983e-01 8.55405092e-01 1.29545853e-01 -3.41190904e-01
-3.15805748e-02 2.60151755e-02 4.90371406e-01 1.01309586e+00
5.83672822e-01 -1.04265667e-01 -3.46554853e-02 1.74133706e+00
-2.82972008e-01 -5.46447039e-02 -8.17902625e-01 -1.15847655e-01
4.68958676e-01 7.85049438e-01 -4.36537772e-01 -5.14756143e-01
1.01232104e-01 1.19628727e+00 4.77759063e-01 4.88582999e-01
-1.01250005e+00 -3.04172575e-01 3.49230051e-01 3.88066471e-02
4.64500189e-01 -9.17752236e-02 -9.04333532e-01 -1.23336923e+00
1.11829191e-01 -5.54481030e-01 6.01340532e-02 -1.59004211e+00
-9.97556329e-01 5.78127384e-01 8.49840045e-03 -1.18024349e+00
-7.11297095e-01 1.84015751e-01 -1.12859762e+00 9.51830745e-01
-7.86417663e-01 -1.28980780e+00 -4.81307596e-01 2.22203076e-01
8.95673752e-01 -3.54228020e-01 6.19150519e-01 -4.30775076e-01
-4.26329553e-01 3.42471004e-01 2.77134329e-01 1.68919742e-01
5.92024684e-01 -1.43207383e+00 5.91844499e-01 1.02075899e+00
3.58685732e-01 4.47987765e-02 1.15112925e+00 -8.09934139e-01
-9.47591841e-01 -9.09565628e-01 9.76003170e-01 -5.00199020e-01
4.05716985e-01 -6.86308444e-01 -6.95982873e-01 7.12546647e-01
1.01326823e+00 -9.50449705e-01 6.99941278e-01 8.82172212e-02
-2.99761176e-01 3.98338884e-01 -7.53057897e-01 1.01151311e+00
8.47965539e-01 -4.13416982e-01 -5.60959041e-01 2.01381847e-01
1.11905336e+00 -5.47670364e-01 -6.20720565e-01 -4.80450869e-01
5.59416831e-01 -9.91042852e-01 6.04305267e-01 -2.89059818e-01
1.65566695e+00 -1.96084872e-01 4.16831393e-03 -1.64941978e+00
-6.05412185e-01 -6.13462627e-01 -2.70037413e-01 1.51739085e+00
6.88877404e-01 2.92663187e-01 8.47877860e-01 5.27466714e-01
-1.42400578e-01 -6.06965542e-01 -4.02899981e-01 -1.04710504e-01
-2.83809938e-02 -5.97037554e-01 9.23730791e-01 6.13459826e-01
-2.98503344e-03 8.70981455e-01 -1.08905244e+00 -5.14584742e-02
6.59982383e-01 3.02071989e-01 1.04385126e+00 -8.63415182e-01
-4.71770763e-01 -2.87690580e-01 1.26221851e-01 -1.07645130e+00
-5.26480041e-02 -6.49648786e-01 2.59411812e-01 -2.44314718e+00
6.93398178e-01 3.52614731e-01 4.27572042e-01 1.98949158e-01
-3.83459240e-01 -2.49445215e-01 4.05435979e-01 7.89849684e-02
-4.81332451e-01 1.06616509e+00 1.48143375e+00 -4.15242046e-01
-3.74364823e-01 -3.15685034e-01 -7.98276067e-01 4.56681073e-01
6.01748466e-01 -3.83702189e-01 -6.43638134e-01 -7.28216529e-01
6.10704422e-01 5.39334714e-01 4.62180644e-01 -9.49279904e-01
6.30591810e-03 -3.74350309e-01 1.00216043e+00 -1.26200521e+00
4.51399744e-01 -1.76155999e-01 2.94246227e-01 1.25295460e-01
-7.38754034e-01 -4.38628197e-02 -2.76871789e-02 5.66981792e-01
-5.32366615e-03 -8.23350921e-02 3.30394626e-01 -2.27968041e-02
-2.65548587e-01 1.35767475e-01 -5.71285903e-01 9.09549370e-02
8.05483997e-01 -1.39426440e-01 -3.83988708e-01 -1.17077672e+00
-7.12589562e-01 2.86846161e-01 5.09153783e-01 4.23400939e-01
1.02086699e+00 -1.72457814e+00 -1.37113929e+00 -9.81147960e-02
7.54489452e-02 3.66666973e-01 5.39013743e-01 8.61384869e-02
-6.74449384e-01 1.79839581e-01 -3.26232553e-01 -5.30799985e-01
-9.69880044e-01 3.33262146e-01 -2.01386496e-01 -4.96479541e-01
-8.63092124e-01 8.07287872e-01 3.28470141e-01 2.44638044e-02
1.41014442e-01 -4.08203632e-01 -2.54765511e-01 7.30400234e-02
7.61121750e-01 2.26248384e-01 -6.74495339e-01 -2.58764744e-01
2.28885815e-01 1.06657505e-01 1.58796638e-01 -6.30240738e-01
1.67993522e+00 -3.63474190e-02 2.99311399e-01 8.99950504e-01
6.89394593e-01 9.32681654e-03 -1.92967403e+00 7.53970072e-02
-3.93977553e-01 -4.11472559e-01 -1.17676161e-01 -8.00415397e-01
-6.91161275e-01 1.12426007e+00 -9.38547477e-02 -1.15272030e-02
6.68052375e-01 1.07261918e-01 7.53702879e-01 -1.09705642e-01
-7.29809552e-02 -9.67720270e-01 4.38146263e-01 4.70551968e-01
1.47223663e+00 -9.22193050e-01 7.01216282e-03 -9.44265425e-02
-1.30734909e+00 1.07755065e+00 6.52925134e-01 -4.02068973e-01
2.42379487e-01 3.32807392e-01 -1.73305377e-01 -2.09075540e-01
-1.38879192e+00 1.15009487e-01 4.09594119e-01 7.92542517e-01
5.72159827e-01 2.50149846e-01 2.32624963e-01 9.07325029e-01
-6.99202061e-01 -1.26728833e-01 8.83416474e-01 4.62075293e-01
-4.98244911e-01 -4.76127952e-01 -5.05804241e-01 4.75346863e-01
6.49020672e-02 -1.52102545e-01 -6.08766556e-01 6.86550736e-01
-1.58552043e-02 7.78073192e-01 1.97483853e-01 -4.30825561e-01
1.29576579e-01 9.88358930e-02 4.99645531e-01 -5.89709282e-01
-1.39914587e-01 2.75472760e-01 2.33323485e-01 -9.51459780e-02
2.16009393e-01 -6.63809597e-01 -9.87531066e-01 -5.64539433e-01
-1.55172542e-01 -1.56982154e-01 3.19235206e-01 9.67583597e-01
1.02964848e-01 1.28095329e+00 2.99149126e-01 -8.16345215e-01
-1.10707864e-01 -1.41144943e+00 -5.21906555e-01 4.06081021e-01
2.78959453e-01 -4.72906977e-01 9.00088772e-02 4.89563674e-01] | [11.210152626037598, 0.6681820750236511] |
9b68bec6-78be-4c7e-b2db-ac11177bbc22 | vesr-net-the-winning-solution-to-youku-video | 2003.02115 | null | https://arxiv.org/abs/2003.02115v1 | https://arxiv.org/pdf/2003.02115v1.pdf | VESR-Net: The Winning Solution to Youku Video Enhancement and Super-Resolution Challenge | This paper introduces VESR-Net, a method for video enhancement and super-resolution (VESR). We design a separate non-local module to explore the relations among video frames and fuse video frames efficiently, and a channel attention residual block to capture the relations among feature maps for video frame reconstruction in VESR-Net. We conduct experiments to analyze the effectiveness of these designs in VESR-Net, which demonstrates the advantages of VESR-Net over previous state-of-the-art VESR methods. It is worth to mention that among more than thousands of participants for Youku video enhancement and super-resolution (Youku-VESR) challenge, our proposed VESR-Net beat other competitive methods and ranked the first place. | ['Chaowei Shan', 'Zhibo Chen', 'Sen Liu', 'Xu Tan', 'Jiale Chen'] | 2020-03-04 | null | null | null | null | ['video-enhancement'] | ['computer-vision'] | [ 3.09339035e-02 -3.12702179e-01 -1.47683837e-03 -1.97711885e-01
-9.18891191e-01 4.18979377e-02 3.45122427e-01 -6.46757722e-01
-3.97517443e-01 6.34693325e-01 7.03445435e-01 1.22960895e-01
7.14724213e-02 -4.61216986e-01 -8.05105746e-01 -1.36573762e-01
-2.43940860e-01 -3.76243085e-01 5.15977621e-01 -6.36491656e-01
5.07691763e-02 2.96737105e-01 -1.59526229e+00 8.47271264e-01
6.18166924e-01 9.30159926e-01 6.00730896e-01 7.54102588e-01
1.71266988e-01 1.09221733e+00 -2.86115408e-01 -3.88476163e-01
2.52570659e-01 -2.57404894e-01 -7.04918921e-01 3.90814170e-02
6.59774601e-01 -9.07201469e-01 -1.17583323e+00 1.09928668e+00
7.12463200e-01 4.43673879e-01 2.84668773e-01 -7.50853896e-01
-1.18281865e+00 8.65992308e-01 -1.04495561e+00 1.03861094e+00
5.45423448e-01 8.77186134e-02 7.72741497e-01 -1.39919782e+00
9.27728772e-01 1.33631301e+00 6.81063712e-01 6.22624159e-01
-8.62369061e-01 -7.09009647e-01 3.28776807e-01 6.30834699e-01
-1.60022759e+00 -7.45477736e-01 3.60666722e-01 4.46951240e-02
1.11356592e+00 2.02927724e-01 5.79241276e-01 1.15610981e+00
4.22893651e-02 9.29164708e-01 8.68418217e-01 1.86391398e-01
-1.79559797e-01 -2.36324757e-01 5.33986799e-02 5.18028438e-01
-1.81918576e-01 3.35781693e-01 -9.67099428e-01 3.09296101e-01
1.50036728e+00 2.74609644e-02 -6.92453027e-01 2.32609645e-01
-1.30067766e+00 3.62269938e-01 4.47666824e-01 3.62143219e-01
-6.95384860e-01 3.31917912e-01 4.54696298e-01 3.57567638e-01
5.36188483e-01 1.78996339e-01 -2.38636822e-01 -3.76005679e-01
-1.23285794e+00 1.43267825e-01 -2.05050670e-02 1.12432361e+00
2.06408024e-01 4.03577268e-01 -6.32096648e-01 8.70493412e-01
-1.05846591e-01 1.01890184e-01 2.10269630e-01 -1.33616924e+00
5.83212018e-01 -1.79596543e-01 1.22279041e-01 -8.52798045e-01
7.72184059e-02 -2.63280630e-01 -8.99951041e-01 9.11947265e-02
-2.04079583e-01 -8.01718011e-02 -8.22000682e-01 1.51522803e+00
-1.32230846e-02 9.71027970e-01 8.71130526e-02 1.43680799e+00
1.41575015e+00 9.43360567e-01 -6.78883195e-02 -3.27605247e-01
1.39045405e+00 -1.27819526e+00 -9.71019566e-01 1.01296790e-01
-1.67373076e-01 -5.40288627e-01 7.72702694e-01 3.71999502e-01
-1.72556460e+00 -9.99798119e-01 -9.38405573e-01 -5.46260834e-01
5.83856553e-02 6.52278736e-02 6.47154987e-01 -2.33358480e-02
-1.40713298e+00 8.37082148e-01 -6.37418449e-01 6.19890802e-02
6.43848002e-01 1.67342335e-01 -5.29451847e-01 -2.99643606e-01
-1.42068756e+00 6.84423327e-01 2.90946513e-01 1.55023545e-01
-1.15053463e+00 -9.73625779e-01 -9.65254843e-01 3.64463776e-01
4.34023768e-01 -5.67114055e-01 1.09919274e+00 -8.00650656e-01
-1.20784330e+00 5.26964009e-01 -4.23563391e-01 -5.41728735e-01
4.74593729e-01 -5.68016350e-01 -7.75233209e-01 4.50860381e-01
-2.13512376e-01 8.07462275e-01 9.80300903e-01 -1.17337155e+00
-8.31074595e-01 5.40829338e-02 1.84407383e-01 2.61572003e-01
-1.32805705e-01 4.25857544e-01 -1.06358588e+00 -9.70532656e-01
-2.53728569e-01 -1.85986474e-01 -1.63521007e-01 -2.85050631e-01
5.77132441e-02 -2.41078079e-01 9.12665427e-01 -1.20035136e+00
1.51387119e+00 -2.40156245e+00 3.15617323e-01 -1.78790018e-01
8.40954244e-01 4.22857344e-01 -5.76562285e-01 -1.05732694e-01
-5.16135156e-01 1.63390994e-01 2.43233636e-01 -2.15941176e-01
-2.76295811e-01 -2.78988034e-01 -3.83469939e-01 3.07457179e-01
2.51163542e-01 9.85158622e-01 -9.67298985e-01 -4.79090691e-01
4.42211598e-01 1.14009881e+00 -4.56237584e-01 1.87024146e-01
2.29107752e-01 1.71496317e-01 -2.96146125e-01 6.76373363e-01
8.43694866e-01 -2.28513360e-01 -4.57997285e-02 -7.67337322e-01
-2.41059005e-01 1.30238026e-01 -1.14796853e+00 1.71598899e+00
-2.93865353e-01 9.90108788e-01 -9.81956534e-03 -5.08423805e-01
7.15562522e-01 5.60261607e-01 6.39230907e-01 -9.33219433e-01
1.75941661e-01 -5.37742078e-02 -4.01061803e-01 -3.69550377e-01
1.12719548e+00 4.49282080e-01 3.16981554e-01 -4.00475189e-02
3.97171527e-01 7.42625475e-01 1.49524614e-01 6.61931157e-01
1.11373913e+00 3.37243617e-01 2.79829383e-01 -9.30391625e-02
3.94860327e-01 -6.49370193e-01 6.60710573e-01 6.53022468e-01
-3.79543364e-01 9.90732610e-01 1.29973322e-01 -4.69160438e-01
-1.16060305e+00 -1.21502030e+00 4.43223268e-02 1.02938163e+00
5.45510590e-01 -8.10646832e-01 -7.08923221e-01 -4.66352314e-01
-4.15183842e-01 3.39292437e-01 -7.04606712e-01 4.59497571e-02
-6.64179444e-01 -3.10383111e-01 2.47437030e-01 9.76398349e-01
9.12088990e-01 -9.56239998e-01 -3.01455647e-01 1.10825151e-01
-7.42218137e-01 -1.45639718e+00 -9.73021150e-01 -4.01844859e-01
-6.30985677e-01 -8.93829942e-01 -1.07910395e+00 -6.13756061e-01
2.43926898e-01 9.26920831e-01 1.36991799e+00 2.07113698e-01
-1.58272818e-01 4.14931267e-01 -6.04977548e-01 2.76483446e-01
4.54606637e-02 -2.07447931e-01 3.10331527e-02 -7.74795413e-02
2.07413033e-01 -6.32617652e-01 -7.25979447e-01 4.55468923e-01
-7.08347976e-01 1.89101696e-01 4.36075002e-01 7.09793627e-01
1.04148257e+00 2.09530324e-01 5.58008790e-01 -7.43532300e-01
5.50443530e-01 -4.60508734e-01 -3.06341618e-01 4.51719671e-01
-3.29329409e-02 -2.88378417e-01 3.28449428e-01 -5.89570522e-01
-1.29920149e+00 -3.02936226e-01 -3.16256285e-01 -1.06927192e+00
1.44284055e-01 1.42854989e-01 -1.82550356e-01 -1.30646333e-01
3.13795120e-01 2.85240144e-01 -4.60304022e-01 -5.38900971e-01
5.25201499e-01 4.78065789e-01 1.13954043e+00 -2.49454945e-01
7.07632780e-01 3.88080031e-01 -3.67052019e-01 -7.15780973e-01
-6.46718204e-01 -3.38322401e-01 -3.12488765e-01 -3.48595411e-01
1.11967170e+00 -1.56753051e+00 -6.77578151e-01 1.82572320e-01
-1.09034848e+00 -1.86892763e-01 -3.04073721e-01 4.64597851e-01
-4.67096329e-01 4.22259420e-01 -1.01437783e+00 -4.89775002e-01
-4.58849102e-01 -1.27952611e+00 9.47314620e-01 4.44111884e-01
1.53945565e-01 -6.59574270e-01 -3.42269897e-01 4.23655868e-01
8.06939662e-01 9.00479499e-04 -1.37180313e-01 -3.44398506e-02
-9.58952725e-01 3.48301649e-01 -7.75540709e-01 2.17160732e-01
-2.55694985e-02 -3.69696096e-02 -7.77643025e-01 -4.78499174e-01
-2.95132846e-01 -2.26360366e-01 1.52304769e+00 6.99665725e-01
1.60388541e+00 -1.51987765e-02 -6.24565333e-02 1.10483432e+00
1.35002959e+00 1.76373780e-01 1.29397225e+00 1.75607011e-01
8.20183992e-01 -9.72903427e-03 7.06262410e-01 4.23666209e-01
5.31666040e-01 9.01961505e-01 3.87567461e-01 -4.28406328e-01
-5.47797024e-01 -1.97994903e-01 6.34723604e-01 8.26565504e-01
-8.03056538e-01 -7.75898471e-02 -7.08567426e-02 5.70844531e-01
-1.90246701e+00 -1.47356963e+00 8.21232498e-02 1.72477925e+00
4.65183198e-01 -1.29973069e-01 1.66751385e-01 -4.62407172e-01
1.10656190e+00 4.08325434e-01 -5.02789319e-01 -1.42862111e-01
-4.83193636e-01 3.21426064e-01 4.35150504e-01 3.65106523e-01
-1.22727585e+00 1.00829434e+00 6.90917969e+00 1.16240227e+00
-6.85670137e-01 2.79899955e-01 5.52289009e-01 -1.77893326e-01
-2.42963821e-01 -3.52893949e-01 -7.58811355e-01 3.94109905e-01
6.77343309e-01 -6.91521615e-02 8.81062150e-01 6.13747716e-01
2.58360595e-01 2.37370849e-01 -8.79105508e-01 1.49028611e+00
1.41768470e-01 -1.79117179e+00 1.06594414e-01 -1.57584518e-01
7.96511471e-01 2.34718502e-01 2.55656809e-01 4.11843985e-01
2.26376280e-01 -1.11590052e+00 6.18049979e-01 6.60271764e-01
1.21753633e+00 -9.33847129e-01 6.89944327e-01 -4.87030119e-01
-1.86746657e+00 -7.18138665e-02 -7.07185745e-01 3.81606758e-01
5.79164386e-01 3.49821091e-01 8.60657096e-02 8.24671209e-01
1.13993263e+00 1.14062822e+00 -5.41132689e-01 9.37693775e-01
7.40767270e-02 4.20141548e-01 9.24449116e-02 8.21951628e-01
3.61074880e-02 1.58475921e-01 7.52987027e-01 1.46045971e+00
4.38571334e-01 6.16483271e-01 -1.26371354e-01 9.81080115e-01
-4.48673755e-01 -2.73301482e-01 -2.46609181e-01 6.40279381e-04
6.28772795e-01 1.35881507e+00 -3.61992985e-01 -5.93282223e-01
-6.59607112e-01 1.16936076e+00 1.40843764e-01 4.41049963e-01
-1.24640179e+00 -2.90218771e-01 9.15072381e-01 1.57680362e-01
7.55973935e-01 -6.13479689e-02 2.37871021e-01 -1.52864075e+00
-1.24674521e-01 -1.11555696e+00 3.70935678e-01 -1.24078155e+00
-1.21165121e+00 1.06217420e+00 -1.08189255e-01 -1.23530209e+00
1.00845993e-01 -2.14301437e-01 -4.85548258e-01 7.80249059e-01
-1.75869548e+00 -9.16991770e-01 -5.28146327e-01 1.09890127e+00
9.78424191e-01 -3.68541360e-01 2.00117543e-01 6.39425695e-01
-4.89770681e-01 7.21130133e-01 -2.31100753e-01 3.56873721e-01
6.71865165e-01 -9.75726664e-01 8.32570255e-01 1.19431651e+00
4.05039862e-02 5.43986738e-01 4.94911283e-01 -6.47143424e-01
-1.25239301e+00 -1.19734728e+00 2.93607473e-01 -7.53746629e-02
2.88773477e-01 -2.97221541e-02 -1.11065447e+00 6.55176282e-01
4.90158081e-01 4.74364638e-01 2.34753534e-01 -1.14358805e-01
-5.54732859e-01 -4.98287939e-02 -9.81115818e-01 6.76740766e-01
1.54373491e+00 -6.53558314e-01 -4.95533377e-01 -2.52586991e-01
1.25666761e+00 -7.17207849e-01 -1.08990514e+00 3.75996202e-01
4.87246752e-01 -1.13847911e+00 1.50446069e+00 -5.75499892e-01
8.61127257e-01 -4.99602169e-01 -3.86883289e-01 -1.17851818e+00
-8.10588002e-01 -8.16613138e-01 -5.36746860e-01 1.38065541e+00
-1.57573465e-02 -1.55869260e-01 2.80653238e-01 3.43464077e-01
-2.08046556e-01 -5.36844194e-01 -6.85083568e-01 -5.25929213e-01
-4.73343104e-01 -2.62151003e-01 6.47133887e-01 9.67467070e-01
-2.81897902e-01 1.54398188e-01 -1.01526165e+00 2.36472458e-01
7.07536757e-01 -1.82044327e-01 4.19979692e-01 -5.68751097e-01
-5.75456917e-01 -2.34455988e-01 -4.18836981e-01 -1.27701557e+00
-2.48366985e-02 -5.63926578e-01 -2.54183441e-01 -1.60445619e+00
6.70541167e-01 2.29277864e-01 -6.09874547e-01 1.74314976e-01
-5.25785923e-01 5.41123867e-01 6.57418132e-01 -1.83414016e-02
-1.25851667e+00 5.00065863e-01 1.49934125e+00 9.95493084e-02
-2.60862648e-01 -4.87424195e-01 -1.00007534e+00 5.77564836e-01
5.36561489e-01 5.05194813e-02 -2.74855971e-01 -8.04154336e-01
-4.67670895e-02 5.12419879e-01 4.64649200e-01 -8.95653546e-01
1.83171123e-01 -3.18204351e-02 7.15634882e-01 -8.36832345e-01
4.92705286e-01 -5.57730138e-01 2.68992782e-01 -2.47902468e-01
-4.84529793e-01 4.18548614e-01 8.12584832e-02 4.69618052e-01
-3.08213830e-01 3.35339189e-01 9.09875989e-01 -8.54775384e-02
-1.41038394e+00 7.21888483e-01 -1.99631378e-01 1.72563225e-01
7.74432003e-01 -2.65021622e-01 -6.48290753e-01 -5.77460051e-01
-7.30931818e-01 3.01775664e-01 1.65128678e-01 6.47014856e-01
1.44590235e+00 -1.50087523e+00 -1.08507156e+00 -1.23091646e-01
-3.31009477e-01 -1.59240171e-01 1.02035165e+00 7.02045381e-01
-3.19685936e-01 -4.55194153e-03 -5.33378422e-01 -1.19400755e-01
-1.51693070e+00 7.62287617e-01 2.76719898e-01 -2.16484517e-01
-1.17902935e+00 8.80395591e-01 4.20719057e-01 4.74663287e-01
2.31077299e-01 1.73084326e-02 -4.24408108e-01 -3.78470480e-01
1.36648965e+00 6.82204723e-01 -2.80549914e-01 -7.74194419e-01
-2.22003356e-01 3.99516195e-01 -3.71125787e-01 -1.08571490e-02
1.50682366e+00 -5.99464893e-01 1.63986340e-01 -2.09125876e-01
9.85195577e-01 -1.35873035e-01 -1.59302604e+00 -4.35926229e-01
-6.53128803e-01 -9.78462219e-01 4.31796819e-01 -7.18396008e-01
-1.64524388e+00 7.54935503e-01 6.99704409e-01 -4.56810892e-01
1.60398817e+00 -4.07398306e-02 1.14896822e+00 9.65896901e-03
4.06347126e-01 -9.99029338e-01 2.16906995e-01 4.03684914e-01
1.01479244e+00 -1.13164520e+00 4.10237685e-02 -5.69101810e-01
-9.84774113e-01 1.02315986e+00 7.76502848e-01 -3.70780528e-01
5.23486614e-01 4.94019270e-01 -3.04218292e-01 -5.38871214e-02
-9.83293414e-01 -4.03538108e-01 3.82604837e-01 8.84634733e-01
4.72199678e-01 -2.36054093e-01 -1.71159759e-01 8.31898212e-01
2.23412469e-01 4.48192000e-01 5.30790389e-01 3.51552933e-01
-4.31512892e-01 -6.26291156e-01 -2.20431045e-01 1.43189311e-01
-7.32024968e-01 -3.44519764e-01 8.14279243e-02 6.87423348e-01
6.40319660e-02 8.17733765e-01 4.68771197e-02 -7.44489551e-01
3.43167096e-01 -6.02257550e-01 5.27056575e-01 6.71963468e-02
-5.53014457e-01 3.18302065e-01 6.62507042e-02 -1.18399417e+00
-5.60436964e-01 -4.11547691e-01 -1.14083838e+00 -5.72717786e-01
-1.28195956e-01 -6.37343526e-02 6.94037899e-02 6.59198940e-01
5.31300783e-01 1.19317651e+00 5.64242899e-01 -1.20237780e+00
-2.24323153e-01 -6.34275138e-01 -6.81120813e-01 3.55653048e-01
4.07238454e-01 -7.26333380e-01 1.22878812e-01 2.24149525e-01] | [11.094188690185547, -1.9214117527008057] |
24ab638b-bdd3-4d87-92a3-0c969e75763b | c3-cross-instance-guided-contrastive | 2211.07136 | null | https://arxiv.org/abs/2211.07136v2 | https://arxiv.org/pdf/2211.07136v2.pdf | C3: Cross-instance guided Contrastive Clustering | Clustering is the task of gathering similar data samples into clusters without using any predefined labels. It has been widely studied in machine learning literature, and recent advancements in deep learning have revived interest in this field. Contrastive clustering (CC) models are a staple of deep clustering in which positive and negative pairs of each data instance are generated through data augmentation. CC models aim to learn a feature space where instance-level and cluster-level representations of positive pairs are grouped together. Despite improving the SOTA, these algorithms ignore the cross-instance patterns, which carry essential information for improving clustering performance. In this paper, we propose a novel contrastive clustering method, Cross-instance guided Contrastive Clustering (C3), that considers the cross-sample relationships to increase the number of positive pairs. In particular, we define a new loss function that identifies similar instances using the instance-level representation and encourages them to aggregate together. Extensive experimental evaluations show that our proposed method can outperform state-of-the-art algorithms on benchmark computer vision datasets: we improve the clustering accuracy by 6.8%, 2.8%, 4.9%, 1.3% and 0.4% on CIFAR-10, CIFAR-100, ImageNet-10, ImageNet-Dogs, and Tiny-ImageNet, respectively. | ['Narges Armanfard', 'Hadi Hojjati', 'Mohammadreza Sadeghi'] | 2022-11-14 | null | null | null | null | ['image-clustering'] | ['computer-vision'] | [-7.59236962e-02 -2.61155814e-01 -6.05398640e-02 -6.85875893e-01
-6.39733374e-01 -2.90117979e-01 7.00893641e-01 3.82303476e-01
-4.83864307e-01 3.23176712e-01 -2.12988615e-01 2.25087598e-01
-3.01281214e-01 -6.20384395e-01 -5.30648947e-01 -9.88513231e-01
-3.07172328e-01 5.10541618e-01 -1.44631714e-02 2.61840343e-01
1.40890002e-01 1.75076470e-01 -1.50646520e+00 5.63878059e-01
1.13542247e+00 1.09092140e+00 1.12598151e-01 -1.85290016e-02
-2.73141898e-02 8.04093361e-01 -5.39822757e-01 -8.38937312e-02
2.92394370e-01 -3.86767209e-01 -5.30295610e-01 3.95964414e-01
3.48573774e-01 2.42709726e-01 -1.95249498e-01 1.12928486e+00
3.99425536e-01 1.94093183e-01 8.52017224e-01 -1.71422017e+00
-9.00986731e-01 7.44237959e-01 -1.27170444e+00 3.04535627e-02
-5.21050155e-01 1.02529131e-01 7.77759314e-01 -8.36617470e-01
2.23119542e-01 1.25657213e+00 5.19000292e-01 5.79298317e-01
-1.13609540e+00 -1.00394845e+00 3.27225059e-01 4.55799997e-01
-1.58405316e+00 1.40775563e-02 9.33238029e-01 -5.92082560e-01
4.94373173e-01 1.23318002e-01 4.77780849e-01 5.73185682e-01
-3.73205841e-01 8.91105354e-01 1.13466966e+00 -2.08104655e-01
3.28659058e-01 -3.60443816e-02 4.67908978e-01 4.38961715e-01
1.71723232e-01 -1.53850868e-01 -1.08306058e-01 9.72501263e-02
4.04659510e-01 5.49674094e-01 -1.70341432e-01 -5.78931212e-01
-1.12062275e+00 9.58033800e-01 1.12741268e+00 3.80111128e-01
-3.41971487e-01 7.24702850e-02 4.50665861e-01 -1.38747826e-01
4.96724933e-01 3.46494198e-01 -3.03756326e-01 4.23086077e-01
-8.17469180e-01 4.50635813e-02 2.77674615e-01 9.32557762e-01
7.92237818e-01 -3.34487259e-01 -2.29355752e-01 1.18836129e+00
2.56210566e-01 1.84787884e-01 7.43639648e-01 -8.18334818e-01
5.41700363e-01 1.06857622e+00 -2.40899235e-01 -1.34378195e+00
-3.95582914e-01 -8.26227188e-01 -1.51069498e+00 -1.11916944e-01
-2.25644726e-02 -1.20048903e-01 -1.07774150e+00 1.78441036e+00
3.04060489e-01 6.72313392e-01 9.46293399e-02 9.32097077e-01
9.76694167e-01 6.74765825e-01 1.10147692e-01 -1.10925063e-01
1.00519335e+00 -1.08108532e+00 -4.98951435e-01 -1.82294190e-01
4.09147203e-01 -4.96071905e-01 9.15537238e-01 3.19056034e-01
-7.68943191e-01 -7.08997488e-01 -8.95859301e-01 2.26052910e-01
-3.41262877e-01 3.42263669e-01 6.64054930e-01 2.17824206e-01
-8.86016667e-01 5.36053360e-01 -8.42696011e-01 -1.82208009e-02
9.11095083e-01 4.98642534e-01 -2.65910536e-01 -2.34475583e-01
-8.21085691e-01 1.74583375e-01 5.65427780e-01 6.89121932e-02
-7.14883745e-01 -8.13716531e-01 -6.75538242e-01 1.94855317e-01
2.06716999e-01 -3.49190265e-01 7.21639276e-01 -1.04690635e+00
-8.39826643e-01 9.80307400e-01 5.74647374e-02 -5.08722186e-01
3.62931788e-01 -6.75776526e-02 -3.91494572e-01 1.66356727e-01
3.88934374e-01 1.11295676e+00 5.72346628e-01 -1.84428859e+00
-9.05729651e-01 -6.37110293e-01 -4.08036560e-01 8.32684189e-02
-4.05817181e-01 -2.70854533e-01 -7.39781678e-01 -6.52435243e-01
3.25877428e-01 -9.31956947e-01 -4.78425264e-01 -2.65164047e-01
-6.46533787e-01 -6.07531369e-01 9.81615245e-01 -2.18165383e-01
1.10233676e+00 -2.19664550e+00 7.63510447e-03 3.50967139e-01
4.47061002e-01 4.79972422e-01 -1.47646949e-01 1.47566991e-02
-1.52344391e-01 2.66925603e-01 -5.61340749e-01 -5.32383621e-01
-5.24590462e-02 -5.52709699e-02 9.23564360e-02 4.01219279e-01
3.67810071e-01 7.05852509e-01 -9.58705306e-01 -3.97762716e-01
5.12513459e-01 5.52310705e-01 -6.72592700e-01 1.11493737e-01
-1.02322572e-03 4.33425814e-01 -7.98178837e-02 3.09663326e-01
8.08691859e-01 -5.25222659e-01 1.49041578e-01 -2.24937528e-01
1.34937719e-01 -2.12003112e-01 -1.03103817e+00 1.39777684e+00
-1.28375009e-01 4.09549564e-01 -1.80167258e-01 -1.46598411e+00
1.00681591e+00 -4.44028042e-02 6.19227588e-01 -7.76338816e-01
1.14518955e-01 -8.87367949e-02 1.78317979e-01 -3.52690876e-01
9.50456858e-02 3.59933078e-02 -8.71603489e-02 1.59560531e-01
-2.09588066e-01 4.02144343e-01 3.11786324e-01 3.05572957e-01
7.22672582e-01 -5.35075307e-01 -6.42023161e-02 -3.11871618e-01
4.35151339e-01 -1.15790311e-02 8.46090376e-01 5.60930729e-01
-1.53486654e-01 8.12953353e-01 2.87900537e-01 -2.70045787e-01
-7.34796286e-01 -1.01493692e+00 -1.00863189e-01 7.19304740e-01
2.96299219e-01 -3.49354416e-01 -9.47461784e-01 -7.10827291e-01
1.75175015e-02 4.60041910e-01 -8.72768521e-01 -4.41152871e-01
-5.33089280e-01 -1.19613218e+00 1.92271054e-01 6.10849857e-01
9.13872600e-01 -1.08419073e+00 -9.31107476e-02 4.93368395e-02
-3.38581711e-01 -9.06394362e-01 -3.44409734e-01 2.00143874e-01
-7.75373161e-01 -1.31270623e+00 -5.65877855e-01 -1.33529246e+00
1.06861567e+00 5.99899888e-01 1.15394688e+00 2.97793657e-01
-3.60770047e-01 3.99982892e-02 -5.13436437e-01 -3.18132877e-01
5.90491407e-02 -4.58486080e-02 -6.26770556e-02 4.25075263e-01
6.49103284e-01 -5.87410927e-01 -8.47164154e-01 3.39637101e-01
-8.30010176e-01 -4.87154797e-02 6.32398784e-01 8.58684659e-01
8.80117297e-01 3.19683075e-01 8.94670129e-01 -9.92372870e-01
3.74266297e-01 -7.87758708e-01 -3.54306996e-01 1.22172363e-01
-4.50883240e-01 -2.97543049e-01 9.24341261e-01 -3.90559435e-01
-6.71204269e-01 2.09012479e-01 2.56956428e-01 -8.81526291e-01
-2.90016949e-01 5.52491963e-01 -3.80819678e-01 3.55960995e-01
5.21003723e-01 1.24434635e-01 -4.05487120e-02 -4.31564718e-01
4.08249348e-01 7.15823591e-01 8.16315234e-01 -4.31183815e-01
7.01362967e-01 7.57995903e-01 -2.52846360e-01 -5.43808937e-01
-8.30424905e-01 -8.92174661e-01 -7.38011599e-01 -1.29059404e-02
9.32227433e-01 -1.10451221e+00 -6.16835058e-01 5.67413092e-01
-7.25507855e-01 -2.71050990e-01 -5.45939840e-02 4.17555124e-01
-2.76974022e-01 3.88363987e-01 -3.98446113e-01 -4.95982200e-01
-4.06414121e-01 -1.13607252e+00 9.12160039e-01 4.38998491e-01
8.95092413e-02 -8.60769153e-01 -2.88653493e-01 4.11850095e-01
-1.65269915e-02 5.60483158e-01 9.21140134e-01 -7.78525293e-01
-4.67940569e-01 -1.06683031e-01 -4.37929779e-01 4.95189250e-01
3.88960332e-01 1.13599673e-01 -8.89727294e-01 -4.21875089e-01
-1.92033738e-01 -3.19441855e-01 1.24402666e+00 5.06326139e-01
1.87728417e+00 -3.08581471e-01 -5.94496131e-01 5.96509099e-01
1.43393004e+00 4.80806708e-01 4.58772212e-01 8.22731182e-02
1.04257798e+00 7.91172862e-01 5.37087917e-01 4.44903731e-01
4.79329139e-01 4.20161128e-01 5.77800095e-01 -2.87812084e-01
-5.91331273e-02 4.65453155e-02 -2.20224634e-01 7.82794714e-01
2.49124616e-01 5.23855500e-02 -1.06196618e+00 8.97431910e-01
-2.05665088e+00 -1.00183666e+00 -3.05139333e-01 2.04208398e+00
8.14246356e-01 1.52010974e-02 2.41634637e-01 3.55566502e-01
1.21553540e+00 -2.13654056e-01 -5.94321966e-01 4.29089636e-01
7.79877529e-02 1.56363547e-01 7.38524646e-02 1.26926810e-01
-1.58881879e+00 8.12402487e-01 4.17788649e+00 9.88016486e-01
-1.13729203e+00 -7.22441003e-02 1.28566933e+00 -8.69841352e-02
1.57178730e-01 -4.25066948e-01 -5.67018867e-01 9.07617986e-01
4.52895254e-01 -5.21025807e-03 1.46644175e-01 8.78739715e-01
2.30647668e-01 1.87183082e-01 -1.29103076e+00 1.27147651e+00
7.42545426e-02 -1.36137259e+00 1.06646881e-01 1.99034344e-02
1.07366848e+00 -6.00517355e-03 5.70229411e-01 5.40984035e-01
4.42485720e-01 -1.11414838e+00 2.94779569e-01 2.45380923e-01
3.42013270e-01 -1.26623142e+00 8.24837983e-01 3.20488065e-01
-1.15467453e+00 -2.18416229e-01 -4.79063153e-01 2.09173024e-01
-3.40370983e-01 8.48336577e-01 -5.52258790e-01 5.38750470e-01
1.20709538e+00 1.08151281e+00 -7.16925621e-01 1.33984745e+00
1.16376050e-01 7.87498951e-01 -2.97125965e-01 7.16800913e-02
6.22914672e-01 -3.10410798e-01 -1.22286499e-01 1.00200129e+00
-1.53340381e-02 1.79949015e-01 5.11725247e-01 9.31897402e-01
-4.19184834e-01 -4.39667702e-02 -1.54042482e-01 3.03084135e-01
6.76135540e-01 1.41479170e+00 -9.95466053e-01 -4.89074260e-01
-6.15559891e-02 6.70643091e-01 5.32281458e-01 1.44479841e-01
-9.44097221e-01 -3.92194867e-01 6.17252707e-01 -1.15705833e-01
4.13039833e-01 1.77207664e-01 -3.76019567e-01 -7.89229572e-01
4.79834303e-02 -7.63746083e-01 5.35469532e-01 -3.59732419e-01
-1.89633799e+00 5.17692864e-01 -9.99449566e-02 -1.41383982e+00
9.93128344e-02 -1.37443066e-01 -8.08854401e-01 4.57071990e-01
-1.34549487e+00 -1.02659512e+00 -6.63537681e-01 7.01670408e-01
4.34662610e-01 -2.53384590e-01 3.49444389e-01 5.79507172e-01
-7.85253167e-01 7.75271833e-01 4.41274971e-01 6.27657890e-01
6.20671332e-01 -1.28115094e+00 2.98650891e-01 5.81041276e-01
1.94599688e-01 7.75295973e-01 1.38200551e-01 -4.24468577e-01
-6.95972562e-01 -1.88162172e+00 4.77397949e-01 -2.64523625e-01
2.42754996e-01 -3.81380945e-01 -1.04022670e+00 3.73843580e-01
1.60697803e-01 3.72590750e-01 8.91985655e-01 -2.65697148e-02
-4.44591671e-01 -5.00937164e-01 -1.30195367e+00 5.47370315e-01
8.77230942e-01 -1.06158741e-01 -4.18814361e-01 5.28172791e-01
7.58049667e-01 8.34167749e-02 -8.52693617e-01 5.18063247e-01
-3.14458720e-02 -1.04586363e+00 9.07295823e-01 -4.38362420e-01
5.85741580e-01 -6.42358184e-01 -1.03872895e-01 -1.48953962e+00
-5.28674841e-01 -1.70484021e-01 1.71448275e-01 1.44263554e+00
1.23737983e-01 -3.72473389e-01 8.92874956e-01 2.79701173e-01
-4.15697098e-01 -8.95366788e-01 -6.48546159e-01 -7.83323228e-01
3.31844538e-01 -7.68199265e-02 6.07804894e-01 1.56605721e+00
-1.42705470e-01 5.14846861e-01 -7.68650547e-02 1.82821274e-01
8.85373890e-01 1.86813071e-01 6.90915406e-01 -1.48200536e+00
2.72544697e-02 -6.28635108e-01 -4.17765439e-01 -7.15515256e-01
2.37494439e-01 -1.10912597e+00 -5.89370951e-02 -1.58188891e+00
5.15571296e-01 -8.89447749e-01 -5.53060889e-01 5.13931930e-01
-4.91645217e-01 4.63008881e-01 3.79543781e-01 3.62294376e-01
-9.85055149e-01 6.78754628e-01 8.81626546e-01 -4.27160829e-01
-1.88487172e-01 -1.10139504e-01 -8.89469326e-01 7.20235348e-01
1.04093933e+00 -4.78408426e-01 -5.72398782e-01 -4.07000482e-01
-4.18263882e-01 -5.20081997e-01 4.30589020e-01 -1.41967773e+00
4.15988833e-01 9.55668837e-02 6.89239502e-01 -9.42993402e-01
6.56978935e-02 -9.06230032e-01 2.52237869e-03 4.89217788e-01
-5.95403254e-01 -7.73074478e-02 2.75165290e-02 7.83973634e-01
-4.47457284e-01 1.98207766e-01 1.10868108e+00 -1.68980107e-01
-7.68885136e-01 6.24004543e-01 6.92819357e-02 2.29310587e-01
1.43358767e+00 -1.79535523e-02 -2.68356055e-01 -6.54959586e-04
-6.88367248e-01 7.31862485e-01 2.43302345e-01 4.71687347e-01
6.54933572e-01 -1.58771336e+00 -7.80458748e-01 2.56322306e-02
2.93647289e-01 5.64932346e-01 4.34684485e-01 6.53978527e-01
-2.61605352e-01 2.82795042e-01 -1.10449553e-01 -1.14260077e+00
-1.28659320e+00 7.76876390e-01 3.09891820e-01 -1.62017316e-01
-3.96105200e-01 1.00079191e+00 6.48509204e-01 -6.99560642e-01
4.84030247e-01 -2.47294515e-01 -4.25549209e-01 -9.05709434e-03
4.51911658e-01 3.03236127e-01 6.13860749e-02 -5.52891135e-01
-3.92071694e-01 4.87478346e-01 -5.00837266e-01 4.61157888e-01
1.45087755e+00 7.81487972e-02 -2.61574775e-01 3.63396764e-01
1.53377020e+00 -5.72879910e-01 -1.26896048e+00 -2.72252500e-01
1.21479779e-01 -2.53240824e-01 -1.77202061e-01 -8.16143215e-01
-1.71445060e+00 1.04993832e+00 7.37009645e-01 1.48210704e-01
1.20629406e+00 9.24347863e-02 7.33934045e-01 2.63176233e-01
1.42301872e-01 -1.03700984e+00 4.27706808e-01 1.88475937e-01
5.73467433e-01 -1.48123014e+00 -1.82359219e-01 -4.42805886e-01
-6.98716938e-01 4.91613567e-01 1.07833195e+00 -4.28866476e-01
7.91404605e-01 -2.19029300e-02 -3.36662233e-02 -2.79452115e-01
-6.15166903e-01 -2.17134982e-01 2.49774322e-01 7.98744202e-01
4.25443769e-01 3.08649838e-01 -1.13257945e-01 8.07175219e-01
-4.78119738e-02 -3.97302955e-01 1.69999361e-01 6.36761963e-01
-3.60839605e-01 -7.38877416e-01 -3.17087591e-01 7.49217153e-01
-4.05868441e-01 -8.56709480e-02 -5.99416196e-01 8.98643792e-01
3.55286419e-01 1.34879863e+00 5.70931494e-01 -5.07356048e-01
7.62441233e-02 -4.17459935e-01 1.15060829e-01 -6.14356995e-01
-6.75008893e-01 5.30949682e-02 -4.83361840e-01 -2.34023973e-01
-7.31371582e-01 -5.01428127e-01 -1.61273491e+00 -2.11281970e-01
-2.81278312e-01 4.86767858e-01 3.79454613e-01 6.58438683e-01
6.49888813e-01 5.94612956e-01 1.03091955e+00 -7.43239403e-01
-2.67973185e-01 -8.50194752e-01 -5.02319276e-01 8.38588595e-01
1.30950421e-01 -6.67695642e-01 -4.38937366e-01 8.66268203e-02] | [9.222664833068848, 3.3098089694976807] |
8870fc0c-7348-40b2-ae9d-1649f49eb6f9 | meta-learning-for-few-shot-camera-adaptive | 1811.11788 | null | http://arxiv.org/abs/1811.11788v2 | http://arxiv.org/pdf/1811.11788v2.pdf | Formulating Camera-Adaptive Color Constancy as a Few-shot Meta-Learning Problem | Digital camera pipelines employ color constancy methods to estimate an
unknown scene illuminant, in order to re-illuminate images as if they were
acquired under an achromatic light source. Fully-supervised learning approaches
exhibit state-of-the-art estimation accuracy with camera-specific labelled
training imagery. Resulting models typically suffer from domain gaps and fail
to generalise across imaging devices. In this work, we propose a new approach
that affords fast adaptation to previously unseen cameras, and robustness to
changes in capture device by leveraging annotated samples across different
cameras and datasets. We present a general approach that utilizes the concept
of color temperature to frame color constancy as a set of distinct, homogeneous
few-shot regression tasks, each associated with an intuitive physical meaning.
We integrate this novel formulation within a meta-learning framework, enabling
fast generalisation to previously unseen cameras using only handfuls of camera
specific training samples. Consequently, the time spent for data collection and
annotation substantially diminishes in practice whenever a new sensor is used.
To quantify this gain, we evaluate our pipeline on three publicly available
datasets comprising 12 different cameras and diverse scene content. Our
approach delivers competitive results both qualitatively and quantitatively
while requiring a small fraction of the camera-specific samples compared to
standard approaches. | ['Ales Leonardis', 'Sarah Parisot', 'Steven McDonagh', 'Zhenguo Li', 'Xing Zhang', 'Gregory Slabaugh', 'Fengwei Zhou'] | 2018-11-28 | null | null | null | null | ['color-constancy', 'few-shot-camera-adaptive-color-constancy', 'few-shot-camera-adaptive-color-constancy'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 5.54946661e-01 -5.91551781e-01 7.32005537e-02 -4.17914718e-01
-8.25535357e-01 -1.01447082e+00 6.65924668e-01 -8.31572786e-02
-6.77144885e-01 4.55287576e-01 -3.08932304e-01 1.56046957e-01
1.57608375e-01 -1.41350240e-01 -9.22626853e-01 -7.60703564e-01
3.96054298e-01 1.96559563e-01 3.35764706e-01 2.49049991e-01
1.48702905e-01 4.55848038e-01 -1.69320703e+00 2.30576158e-01
3.43059450e-01 1.05166626e+00 3.89750838e-01 1.01532841e+00
1.59927666e-01 8.24571908e-01 -4.96697724e-01 -2.92356223e-01
5.99871039e-01 -2.99873859e-01 -5.65053284e-01 7.25712776e-01
9.41303015e-01 -4.21081066e-01 -8.04182068e-02 7.30161250e-01
2.57390827e-01 1.45130754e-01 5.29067874e-01 -1.10898256e+00
-3.69772077e-01 -2.12904349e-01 -5.34999251e-01 8.53881538e-02
2.60137618e-01 5.45466661e-01 7.56453753e-01 -4.61117029e-01
4.29409981e-01 5.40605068e-01 8.95847321e-01 6.24747336e-01
-1.53486681e+00 -3.02605331e-01 1.97648257e-01 7.00642765e-02
-1.19311666e+00 -6.79760158e-01 7.24705100e-01 -5.37634015e-01
9.12906408e-01 1.31309435e-01 7.57599473e-01 1.17520368e+00
2.86018960e-02 4.66488749e-01 1.43892705e+00 -4.57498848e-01
4.92218077e-01 3.45634460e-01 -2.56633461e-01 5.34759045e-01
2.09290206e-01 1.35737479e-01 -5.25181651e-01 1.83745593e-01
6.47346139e-01 1.13415740e-01 -3.60598624e-01 -6.43250346e-01
-1.19702005e+00 4.02106136e-01 1.28260806e-01 -1.66665941e-01
-5.91118336e-02 2.93893069e-01 2.81941772e-01 1.98230416e-01
5.13935208e-01 4.43772912e-01 -9.15808022e-01 -1.47986278e-01
-1.00688386e+00 -1.58940375e-01 6.88789248e-01 1.09240603e+00
1.24976254e+00 -1.20227762e-01 1.89912483e-01 7.19387650e-01
6.49405569e-02 5.05382359e-01 3.28167558e-01 -1.22685409e+00
1.38628513e-01 3.82566154e-01 3.15055907e-01 -5.75506985e-01
-3.08714509e-01 -2.40773186e-01 -5.44510007e-01 3.20122659e-01
5.48943341e-01 -9.41332728e-02 -1.00219965e+00 1.44742918e+00
3.01514328e-01 6.48928046e-01 2.13564932e-03 9.29223061e-01
3.22866291e-01 5.56241632e-01 1.41869169e-02 -1.73371345e-01
1.13710856e+00 -1.09623742e+00 -1.88576326e-01 -5.69354475e-01
2.34139651e-01 -8.19873333e-01 1.17236793e+00 8.61114025e-01
-8.74395013e-01 -6.83919311e-01 -1.05416918e+00 -1.48901701e-01
-4.55858499e-01 2.17964888e-01 6.00960314e-01 9.59968686e-01
-1.11986291e+00 5.41238785e-01 -7.02967286e-01 -4.67437595e-01
3.70653927e-01 2.43895888e-01 -3.03909004e-01 -4.18035984e-01
-5.59920073e-01 7.80017614e-01 3.73698652e-01 4.77132015e-03
-1.15474224e+00 -9.79357302e-01 -7.12261319e-01 -3.93263131e-01
5.17929077e-01 -6.34019971e-01 1.32615292e+00 -1.30655181e+00
-1.97436762e+00 9.93507802e-01 -4.68949564e-02 -1.10700190e-01
5.85777521e-01 -3.64818335e-01 -2.76397109e-01 2.04160452e-01
-2.10796654e-01 5.65057456e-01 1.30741978e+00 -1.45770550e+00
-6.36102200e-01 -9.74715799e-02 2.35563815e-01 1.19385898e-01
-2.93891728e-01 -1.38488084e-01 -7.93611169e-01 -2.40097910e-01
-2.62924790e-01 -1.11950922e+00 -1.52313650e-01 3.02999735e-01
-8.40447992e-02 3.36420536e-01 6.93290353e-01 -2.30041474e-01
6.46523416e-01 -2.10451937e+00 2.21703380e-01 -1.77227601e-01
-7.14271981e-03 2.03134507e-01 -1.84086204e-01 1.79152980e-01
2.70444620e-03 -3.08921129e-01 -3.59261870e-01 -7.13571608e-01
-2.55412996e-01 3.86151671e-01 -6.98063970e-02 8.08973312e-01
4.06477630e-01 5.70725322e-01 -1.05201304e+00 -3.06205720e-01
8.33852410e-01 5.19296050e-01 -4.37639058e-01 5.19984007e-01
-4.25996721e-01 6.26947880e-01 1.24062747e-01 7.02645779e-01
6.80566609e-01 -1.21375300e-01 4.35098112e-02 -5.28512716e-01
-2.73216724e-01 -2.62238741e-01 -1.22080719e+00 2.23346567e+00
-8.78694773e-01 8.21806312e-01 1.15911007e-01 -7.76612341e-01
7.40067303e-01 4.23905328e-02 5.02174497e-01 -4.10598725e-01
1.91360354e-01 8.48754719e-02 -5.89199781e-01 -6.73338532e-01
5.07456362e-01 -2.25304171e-01 2.89926175e-02 3.74669135e-01
3.00911427e-01 -5.00844836e-01 5.97202107e-02 -1.88347146e-01
9.38109040e-01 6.37375951e-01 2.69310296e-01 -2.03237869e-04
3.94463181e-01 2.86774039e-02 3.75220418e-01 7.72010684e-01
-1.64546743e-01 9.21489239e-01 -9.19257849e-02 -6.07020736e-01
-1.31177533e+00 -1.00112116e+00 -1.19365901e-01 1.21284223e+00
2.13508263e-01 -2.19073758e-01 -7.63563275e-01 -2.55241811e-01
-1.71407193e-01 4.91929561e-01 -5.70369422e-01 -3.62894386e-02
-3.19073439e-01 -7.90094018e-01 3.12496215e-01 3.54626268e-01
4.46632981e-01 -5.35553575e-01 -1.25124955e+00 -1.88552458e-02
2.00403214e-01 -1.53778958e+00 -2.01231658e-01 4.19191360e-01
-6.70697093e-01 -1.22664452e+00 -5.89481175e-01 -2.35080391e-01
5.09308338e-01 5.76078296e-01 1.27898979e+00 -2.04639912e-01
-6.63251877e-01 1.16912007e+00 -4.52976346e-01 -2.83271253e-01
-2.74950475e-01 -1.65809523e-02 -1.83694512e-02 2.99255252e-01
2.72599101e-01 -2.62152016e-01 -7.00812101e-01 1.88085020e-01
-1.15196729e+00 1.05218880e-01 5.90429187e-01 7.80428469e-01
5.19750357e-01 2.16639396e-02 -3.02194417e-01 -8.58315349e-01
-9.63548943e-02 -4.20002609e-01 -9.41240251e-01 3.55057955e-01
-4.96621490e-01 7.26053119e-02 7.93422580e-01 -4.92543995e-01
-1.32801843e+00 7.45886922e-01 5.14391780e-01 -6.39516950e-01
-5.31660795e-01 -5.72026148e-02 -4.48315293e-02 -3.18972141e-01
8.34089518e-01 1.49704427e-01 -2.48539120e-01 -4.39838409e-01
7.34862804e-01 5.04643083e-01 9.20159459e-01 -6.97397649e-01
9.16740656e-01 8.78776550e-01 6.27250820e-02 -8.62955213e-01
-9.83251333e-01 -8.47515583e-01 -1.00820804e+00 -5.40186942e-01
8.34760010e-01 -1.11125779e+00 -5.51261246e-01 6.52725041e-01
-1.06411850e+00 -6.00820422e-01 -2.43449077e-01 3.88717234e-01
-6.23502016e-01 3.66171658e-01 -2.24647880e-01 -8.35611045e-01
-2.18114872e-02 -1.09108388e+00 1.46466351e+00 3.51061881e-01
2.07656324e-01 -1.07935858e+00 3.00482631e-01 3.13662469e-01
3.42150629e-01 4.33127254e-01 2.63869166e-01 1.28225923e-01
-6.47605121e-01 -3.06302905e-01 -3.50234091e-01 5.92278838e-01
3.33347380e-01 1.78940505e-01 -1.55113459e+00 -3.85163426e-01
1.45738106e-02 -4.37532037e-01 8.59206080e-01 1.50035292e-01
1.20585334e+00 1.15309417e-01 2.02192459e-02 9.60974336e-01
2.01752806e+00 -2.75552511e-01 4.54204917e-01 6.07813358e-01
8.95148158e-01 5.50880730e-01 3.09792399e-01 6.20426416e-01
2.53460169e-01 6.81899905e-01 6.45404518e-01 -1.74260885e-01
-1.15236804e-01 9.33052674e-02 5.91413140e-01 2.99407005e-01
-1.32897824e-01 -2.47132584e-01 -6.70022309e-01 5.73751926e-01
-1.61327887e+00 -7.92051792e-01 -1.87294021e-01 2.43558240e+00
9.03083026e-01 -1.34194285e-01 1.21385433e-01 -1.01770058e-01
5.44808447e-01 1.55130625e-01 -7.57954657e-01 -1.86241120e-01
-2.11684853e-01 2.54422665e-01 9.44634616e-01 3.69004428e-01
-1.09830225e+00 7.30150104e-01 6.38627815e+00 2.35194013e-01
-1.31941772e+00 2.10146129e-01 5.83854556e-01 -4.08614248e-01
7.24340379e-02 1.77883431e-01 -4.55164462e-01 3.07624429e-01
1.12590098e+00 3.68537903e-01 9.23348308e-01 8.15307081e-01
1.71355188e-01 -4.14847761e-01 -1.42008948e+00 1.33009696e+00
3.51376474e-01 -1.05411410e+00 -4.51597363e-01 -3.48937660e-01
8.65358531e-01 3.86679351e-01 3.03892009e-02 -1.99773520e-01
1.96471915e-01 -7.25419998e-01 7.91816115e-01 7.27688551e-01
9.95308518e-01 -2.19228625e-01 3.88507664e-01 1.31719708e-01
-1.10497355e+00 -2.19351918e-01 -5.45651793e-01 -1.82623446e-01
7.15076774e-02 4.77746367e-01 -6.86818659e-01 6.08879328e-01
8.86987567e-01 9.50473547e-01 -9.85499024e-01 1.07543039e+00
-1.98016584e-01 4.62202132e-01 -4.01477754e-01 3.21405441e-01
7.80861080e-02 -1.99549019e-01 1.04991771e-01 1.49464893e+00
1.98057130e-01 -2.19216466e-01 5.70593327e-02 6.85215533e-01
1.01232007e-02 -2.82140970e-01 -3.52866352e-01 3.96655083e-01
1.90087929e-01 1.57365501e+00 -6.77686512e-01 -1.61468729e-01
-7.41177082e-01 1.52836573e+00 3.11398566e-01 4.45541441e-01
-8.88190389e-01 -2.91256532e-02 5.14479280e-01 -3.25069465e-02
2.44805142e-01 -2.91964322e-01 -8.23895261e-03 -1.39824986e+00
6.26633912e-02 -4.99286592e-01 1.50201768e-01 -1.11296296e+00
-1.42871404e+00 4.02369022e-01 2.14644685e-01 -1.36828852e+00
-2.00132757e-01 -1.03609538e+00 -3.16714108e-01 6.47163987e-01
-1.95385337e+00 -1.32551861e+00 -8.88839483e-01 7.25393951e-01
6.91364646e-01 4.12481278e-02 7.72580504e-01 2.35442758e-01
-6.57349348e-01 2.24666059e-01 4.34683919e-01 -1.76024586e-01
1.01219559e+00 -1.40629435e+00 1.30807012e-01 9.94237423e-01
3.12874705e-01 3.08471382e-01 7.27944493e-01 5.11863939e-02
-1.73864651e+00 -1.21826386e+00 1.51530271e-02 -7.86069393e-01
6.19361162e-01 -5.75839281e-01 -7.81666517e-01 4.83259976e-01
3.23735714e-01 5.12997627e-01 6.80247068e-01 -4.47526984e-02
-5.84126353e-01 -5.73454618e-01 -9.22104359e-01 2.28244841e-01
7.76746273e-01 -7.19406128e-01 -2.74158031e-01 3.80935013e-01
3.45360368e-01 -4.57729667e-01 -7.15675950e-01 -3.12947948e-03
6.49848759e-01 -1.28235316e+00 8.37373734e-01 -2.39999846e-01
2.94570476e-01 -6.20974302e-01 -2.24174872e-01 -1.14824104e+00
-8.23266879e-02 -7.35285223e-01 -1.70305371e-03 1.20390034e+00
1.12226874e-01 -2.96355903e-01 6.06275201e-01 9.11593199e-01
-1.32169366e-01 -2.63268977e-01 -7.39036560e-01 -7.71687508e-01
-1.54161632e-01 -7.73870230e-01 3.72122794e-01 7.99592018e-01
-4.81531799e-01 1.83591217e-01 -7.32610226e-01 2.38623291e-01
8.62737536e-01 -6.92483783e-02 1.07059360e+00 -1.17781472e+00
-6.81254387e-01 -1.09632231e-01 -4.10686344e-01 -8.66073132e-01
2.19650254e-01 -4.16700125e-01 2.32507184e-01 -1.02747321e+00
4.11957830e-01 -3.86734426e-01 -2.69203544e-01 3.14320862e-01
-1.30429566e-01 6.51846349e-01 1.19538724e-01 3.21369648e-01
-9.11193490e-01 1.70692518e-01 6.25341237e-01 -1.23578869e-01
-1.74196750e-01 -2.25134522e-01 -3.25845957e-01 6.56405330e-01
7.25609243e-01 -3.79462898e-01 -4.89811450e-01 -8.54372919e-01
3.42058957e-01 -4.40624535e-01 7.28643417e-01 -1.25497174e+00
2.54309475e-01 -2.62965143e-01 6.20056272e-01 2.10336689e-02
5.16943455e-01 -1.18687248e+00 4.58871663e-01 2.55082324e-02
-1.17352940e-01 -1.42383039e-01 4.45382774e-01 8.14747214e-01
2.09288418e-01 -3.89064878e-01 9.31325197e-01 -3.62200558e-01
-1.08584106e+00 1.30718350e-01 -1.66463733e-01 -7.99683034e-02
1.02381516e+00 -4.59672689e-01 -2.18198836e-01 -1.31173879e-01
-2.71772563e-01 -2.85232455e-01 1.02806175e+00 3.03541720e-01
2.83381432e-01 -8.33815217e-01 -4.96599257e-01 5.09602353e-02
5.22100091e-01 1.69279829e-01 2.75616169e-01 5.20818293e-01
-6.24879479e-01 1.84143543e-01 -1.43531710e-01 -1.00875688e+00
-9.49927092e-01 6.56232953e-01 5.57952404e-01 1.88567579e-01
-3.55972826e-01 8.16808581e-01 -5.53537458e-02 -2.07176447e-01
-7.79539049e-02 -5.79398394e-01 3.74445319e-01 -8.25938582e-02
3.90509248e-01 1.86154306e-01 2.18115166e-01 -5.61703563e-01
-2.67420948e-01 9.36600804e-01 1.25495315e-01 1.14148483e-03
1.32909393e+00 -3.88045788e-01 2.24245861e-02 7.51774192e-01
1.31512952e+00 -1.25270367e-01 -2.21611357e+00 -2.51983166e-01
-1.49789274e-01 -8.07788432e-01 3.11183393e-01 -9.17130411e-01
-1.11228633e+00 7.18455374e-01 8.52364004e-01 -7.15357140e-02
1.40876257e+00 6.24633431e-02 3.75990272e-01 2.97191024e-01
4.22671556e-01 -1.15472996e+00 1.17051989e-01 7.07891285e-02
3.15805435e-01 -1.68298995e+00 3.73025715e-01 -1.22335702e-01
-6.34588003e-01 1.42508745e+00 3.50208044e-01 -3.36478837e-03
2.64374614e-01 2.25827545e-01 1.35792881e-01 1.57933012e-02
-6.56109989e-01 -2.15752184e-01 2.30404839e-01 8.26276422e-01
3.40925872e-01 -2.60425329e-01 5.32840610e-01 9.15892720e-02
2.07566082e-01 8.86890292e-03 6.63171351e-01 8.94798636e-01
-1.77556485e-01 -1.04776788e+00 -3.58004570e-01 1.51977748e-01
-1.38277739e-01 -7.01811835e-02 -3.20648819e-01 7.14848757e-01
3.52095932e-01 1.05636144e+00 2.15974927e-01 -2.54109293e-01
2.61857331e-01 -1.07242741e-01 8.29709768e-01 -6.73222959e-01
-3.74880552e-01 7.91910589e-02 -2.41863713e-01 -7.29116440e-01
-1.12136400e+00 -9.44798350e-01 -6.23070180e-01 -3.25074121e-02
-4.00747478e-01 -3.80575866e-01 9.44050074e-01 9.07371283e-01
1.49623796e-01 3.60680044e-01 8.18263650e-01 -1.39491534e+00
-1.90156639e-01 -6.92933679e-01 -4.00097072e-01 4.83508945e-01
6.55717850e-01 -6.38882816e-01 -5.40747106e-01 7.92333603e-01] | [10.12796401977539, -2.6218650341033936] |
6c269f65-7566-4c8f-8f12-7782085bf7d3 | meta-optimizing-semantic-evolutionary-search | null | null | https://wiki.opencog.org/w/Meta-Optimizing_Semantic_Evolutionary_Search | http://metacog.org/papers/gecco07b_full.pdf | Meta-Optimizing Semantic Evolutionary Search | I present MOSES (meta-optimizing semantic evolutionary search), a new probabilistic modeling (estimation of distribution) approach to program evolution. Distributions are not estimated over the entire space of programs. Rather, a novel representation-building procedure that exploits domain knowledge is used to dynamically select program subspaces for estimation over. This leads to a system of demes consisting of alternative representations (i.e. program subspaces) that are maintained simultaneously and managed by the overall system. Application of MOSES to solve the artificial ant and hierarchically composed parity-multiplexer problems is described, with results showing superior performance. An analysis of the probabilistic models constructed shows that representation-building allows MOSES to exploit linkages in solving these problems. | ['Moshe Looks'] | 2017-07-11 | null | null | null | association-for-computing-machinery-acm-2017 | ['problem-decomposition'] | ['miscellaneous'] | [ 1.95550650e-01 9.10057649e-02 -5.98329127e-01 -2.38146544e-01
-4.53471601e-01 -6.43523872e-01 3.20027113e-01 -1.33525385e-02
-9.22031924e-02 6.95613980e-01 -7.99981207e-02 -3.96987766e-01
-6.92078710e-01 -8.77912223e-01 -3.88201058e-01 -6.24901056e-01
-1.99581146e-01 9.34537053e-01 4.96744990e-01 -3.82562339e-01
8.33696604e-01 2.80250221e-01 -2.30901432e+00 3.85257185e-01
1.16281044e+00 8.97254050e-02 4.47184205e-01 1.12079358e+00
-6.10424280e-01 7.28393853e-01 -1.07691813e+00 -1.97560906e-01
-2.31157780e-01 -5.06637335e-01 -6.09264672e-01 -7.47952312e-02
-3.25635046e-01 3.99434984e-01 -1.72678441e-01 1.41534042e+00
2.57599354e-01 4.06917669e-02 6.11197054e-01 -1.72423005e+00
-6.91107631e-01 9.46978569e-01 -5.22281647e-01 1.22390166e-01
6.04516029e-01 -1.43852890e-01 8.08139324e-01 -4.39639688e-01
6.92214668e-01 1.47837448e+00 6.74255252e-01 4.98705864e-01
-1.79307532e+00 -2.03639552e-01 -2.18845934e-01 1.67064980e-01
-1.74785876e+00 -7.71530122e-02 5.96064806e-01 -5.18317461e-01
1.45338440e+00 7.12697148e-01 7.09437132e-01 6.89709842e-01
3.99476260e-01 8.59041154e-01 9.62469339e-01 -7.00846970e-01
6.79799736e-01 6.78697348e-01 4.24207211e-01 8.93350363e-01
4.52748299e-01 5.75223684e-01 -7.88937867e-01 -1.03879297e+00
1.70881927e-01 -6.08879685e-01 2.28062496e-02 -1.14474630e+00
-7.56909668e-01 9.79333162e-01 -2.67125309e-01 3.10488015e-01
-4.21554685e-01 2.69672573e-01 3.58844817e-01 2.21998826e-01
-6.35698661e-02 6.56287313e-01 -3.69280130e-01 -5.94665408e-01
-1.05837953e+00 4.43106592e-01 1.29346824e+00 1.25817728e+00
5.89730263e-01 3.65268171e-01 -6.09907433e-02 6.62770152e-01
4.32937413e-01 3.20267260e-01 6.93410456e-01 -9.96145010e-01
1.13503255e-01 1.05010891e+00 -1.66103709e-02 -9.22869980e-01
-1.49308711e-01 -2.06762180e-01 1.75163805e-01 5.26983380e-01
-1.20211571e-01 -2.31470391e-02 -7.16325998e-01 1.66521728e+00
1.11466974e-01 -6.16692483e-01 2.22213849e-01 -8.67683161e-03
1.16313353e-01 7.70381153e-01 3.83136533e-02 -6.58832639e-02
1.09207940e+00 -8.73891592e-01 -5.35847902e-01 2.81712543e-02
7.41744995e-01 -3.33124161e-01 7.45555341e-01 6.16064608e-01
-9.74490047e-01 -3.10634017e-01 -1.29348743e+00 7.24944472e-01
-6.22932136e-01 3.84433083e-02 9.15166199e-01 1.47125566e+00
-1.58945894e+00 5.92945278e-01 -8.83061707e-01 -4.86908585e-01
2.08380073e-01 7.60307074e-01 2.24977046e-01 2.91889787e-01
-4.46426243e-01 9.29281652e-01 1.18093383e+00 -4.37748790e-01
-8.60713720e-01 -3.78584653e-01 -7.56166577e-01 2.96550274e-01
1.66475460e-01 -7.71130562e-01 8.10741961e-01 -1.16625893e+00
-1.70581412e+00 6.72486186e-01 -4.48021352e-01 -1.17773205e-01
4.00304049e-02 3.93193573e-01 -3.69310021e-01 -2.62790591e-01
-1.49568319e-01 3.78774017e-01 6.10220432e-01 -1.60959208e+00
-9.60416377e-01 -4.19816554e-01 -5.76889329e-02 5.96471280e-02
-5.99521399e-01 4.05525684e-01 -1.90999970e-01 -5.54508626e-01
5.12231737e-02 -1.17054319e+00 -4.21563983e-01 -5.46777070e-01
-2.65815288e-01 -2.32134134e-01 5.19165754e-01 -7.17097938e-01
1.93459785e+00 -2.03239775e+00 1.06890488e+00 6.22732341e-01
-1.86263323e-01 -1.73074797e-01 -1.98994175e-01 5.71336567e-01
-2.77474403e-01 2.29369819e-01 -7.35607862e-01 1.62120583e-03
3.87677759e-01 2.50368237e-01 -7.02801570e-02 1.82755336e-01
-9.05489698e-02 5.36222994e-01 -6.74978256e-01 -6.82455599e-01
8.03947672e-02 2.33412180e-02 -9.06338930e-01 -3.20521146e-02
-6.68129206e-01 -3.90049785e-01 -6.51584685e-01 1.05752647e+00
5.70559204e-01 1.63104981e-01 5.85131884e-01 3.93596828e-01
-2.70761341e-01 -3.44563842e-01 -1.15056777e+00 1.87996233e+00
-4.91503865e-01 5.05490422e-01 -1.09213263e-01 -1.03987396e+00
1.24757445e+00 -1.29227847e-01 5.24982691e-01 -1.70167863e-01
-7.16178715e-02 3.08568627e-01 1.16124429e-01 -5.75914264e-01
8.63119543e-01 4.74754542e-01 -5.52701831e-01 6.68361843e-01
3.08301479e-01 -6.95531428e-01 6.08272076e-01 1.29191980e-01
1.26049125e+00 5.29140770e-01 4.09974933e-01 -8.68339121e-01
5.72594762e-01 9.60964084e-01 6.32747352e-01 8.08631003e-01
-1.34808794e-01 8.88277143e-02 8.93829107e-01 -2.02931821e-01
-1.11458969e+00 -1.07988739e+00 1.60773918e-02 1.20741999e+00
-1.12851560e-02 -7.24986017e-01 -8.83710563e-01 -4.91664290e-01
1.37498811e-01 1.68050742e+00 -6.68251932e-01 -3.54791045e-01
-5.85880697e-01 -1.24147880e+00 5.03816724e-01 3.13386589e-01
8.98398310e-02 -7.79503882e-01 -1.10186720e+00 4.11079079e-01
1.91496611e-01 3.79631519e-02 6.44259602e-02 4.40544426e-01
-1.07249701e+00 -9.83373344e-01 -3.56766552e-01 -7.84763396e-01
5.65909445e-01 -2.88281769e-01 1.37917757e+00 1.05442209e-02
-5.55574715e-01 7.99516022e-01 -9.24470127e-02 -2.12305784e-01
-1.08107591e+00 3.49990204e-02 -3.05441797e-01 -6.91158891e-01
4.83233601e-01 -7.65463591e-01 3.60445976e-01 1.41618311e-01
-7.04129338e-01 -4.60502207e-01 4.99067307e-01 1.29027772e+00
2.98664510e-01 9.41216588e-01 3.55251948e-04 -7.79174805e-01
9.14188683e-01 -7.61884987e-01 -1.13467050e+00 1.07556093e+00
-7.19523191e-01 5.97413361e-01 2.74392888e-02 -5.80615282e-01
-1.47524476e+00 2.79207706e-01 4.83941197e-01 -2.20186338e-01
1.19088069e-01 5.98863661e-01 -2.38771647e-01 -3.19569647e-01
8.96806955e-01 6.68457270e-01 2.20997799e-02 -1.44789457e-01
2.67800301e-01 6.30435705e-01 4.11897331e-01 -1.15045989e+00
4.84236568e-01 -1.26938522e-01 -1.34442940e-01 -7.18325675e-01
3.34019572e-01 7.77406320e-02 -2.56206214e-01 -1.22137122e-01
7.22644687e-01 -3.85726064e-01 -6.27800703e-01 1.76351860e-01
-1.07534647e+00 -3.15930210e-02 -4.28197354e-01 1.72259659e-01
-9.44367528e-01 9.63836610e-02 5.84404208e-02 -1.32578671e+00
8.78987834e-02 -1.60013795e+00 7.46715426e-01 5.10188699e-01
-7.24528193e-01 -9.05655205e-01 5.78379929e-01 2.35780459e-02
4.02437359e-01 1.68485537e-01 1.65092230e+00 -6.80935085e-01
-6.03474259e-01 -9.28718150e-02 3.41374040e-01 -1.33831635e-01
-3.30755562e-01 7.01879799e-01 -5.66862583e-01 -1.13719381e-01
3.05368364e-01 1.52077511e-01 3.54095191e-01 2.55948216e-01
1.03549027e+00 -4.41672951e-02 -9.13403213e-01 4.71449733e-01
1.54493856e+00 1.10032678e+00 5.54425955e-01 7.13440239e-01
1.78578928e-01 8.01981926e-01 5.67961514e-01 7.14683115e-01
9.90313813e-02 7.02289820e-01 1.93858281e-01 8.03263009e-01
1.24397196e-01 -2.30069265e-01 5.61747074e-01 6.32830858e-01
2.86975652e-01 -3.39709073e-01 -1.15252280e+00 7.38185167e-01
-1.89274871e+00 -9.65557754e-01 2.98722804e-01 2.09068131e+00
5.60140491e-01 -1.18193440e-01 2.80687779e-01 -2.04173997e-02
1.19793212e+00 -1.94129169e-01 -7.37846792e-01 -8.61425459e-01
-1.67526543e-01 4.02174205e-01 3.80969882e-01 3.84122044e-01
-6.50517583e-01 7.65027583e-01 7.75798368e+00 1.17247570e+00
-2.96185404e-01 -4.51180488e-02 1.40078038e-01 1.40334800e-01
-8.45081210e-01 3.85477006e-01 -5.59134781e-01 6.14320099e-01
1.12453628e+00 -8.88154626e-01 9.05336797e-01 1.36271238e+00
-8.12562287e-01 -6.20756865e-01 -1.02231359e+00 7.97422826e-01
1.98328376e-01 -1.29523933e+00 -3.44075412e-02 3.93153727e-03
8.72333646e-01 -3.28995585e-01 2.45431006e-01 5.25518775e-01
1.09747756e+00 -7.45341897e-01 1.07157159e+00 3.83567661e-01
7.28597343e-02 -1.18195772e+00 3.30022991e-01 3.31694007e-01
-1.01581550e+00 -8.27520788e-01 -2.70333767e-01 6.38981581e-01
-1.89276874e-01 2.30634615e-01 -5.04781783e-01 6.49837315e-01
7.76440024e-01 1.84537575e-01 -6.39896691e-01 1.14164114e+00
-4.84545343e-03 -4.89094704e-02 -2.53161460e-01 -4.42199439e-01
-2.06467941e-01 -1.92733347e-01 1.13788152e+00 9.24704671e-01
8.43382239e-01 -2.94133186e-01 -1.18283167e-01 1.77286506e+00
6.73147440e-01 -3.33839774e-01 -7.19026864e-01 -4.51808929e-01
7.77217150e-01 5.67555487e-01 -7.99736857e-01 -2.43032768e-01
6.63380474e-02 7.44020641e-01 3.22196819e-02 2.95910954e-01
-9.75098312e-01 -7.64745712e-01 5.81889927e-01 -7.24496186e-01
3.69018048e-01 -1.64728403e-01 -5.07604122e-01 -9.03429329e-01
-3.30407739e-01 -1.26333833e+00 5.85804701e-01 -5.39404869e-01
-6.54354334e-01 8.38578343e-01 5.12969851e-01 -5.89758277e-01
-6.40463114e-01 -5.99980056e-01 -6.64664924e-01 9.62358773e-01
-7.02292740e-01 -5.04396141e-01 1.99641615e-01 2.44515687e-01
6.07721031e-01 -8.82830203e-01 1.01792622e+00 -2.03904808e-01
-8.25141549e-01 5.58820963e-01 3.80268872e-01 -8.77942741e-01
-2.16491982e-01 -1.27975225e+00 5.88812493e-02 8.06627035e-01
-2.00539101e-02 7.91187942e-01 9.57283020e-01 -8.29507709e-01
-1.98796248e+00 -4.69621450e-01 5.40398955e-01 -5.88365018e-01
7.20712543e-01 -8.40577185e-02 -6.09804153e-01 3.41703027e-01
2.71753371e-01 -8.58918846e-01 6.17510915e-01 8.91733319e-02
-1.09254606e-01 3.94205689e-01 -1.42581248e+00 5.30638039e-01
8.26180696e-01 -5.98810256e-01 -7.18262255e-01 2.35508103e-02
6.29153848e-01 -1.89650506e-01 -7.52939284e-01 1.88080341e-01
4.39561754e-01 -1.21874940e+00 9.61526453e-01 -4.51445371e-01
2.49665320e-01 -3.54543895e-01 -5.48326015e-01 -1.27607369e+00
-1.13606058e-01 -6.79761291e-01 -2.33804762e-01 1.28724539e+00
4.54091907e-01 -8.97977591e-01 1.00104117e+00 7.99329042e-01
-1.42018676e-01 -2.82242417e-01 -8.09055209e-01 -9.62697148e-01
-1.84483320e-01 -9.78671163e-02 1.01775825e+00 7.11423218e-01
4.84689832e-01 -2.46180907e-01 2.95430720e-01 2.70940900e-01
8.70894432e-01 4.57980424e-01 4.38288540e-01 -1.07769573e+00
-9.10588562e-01 -9.70499456e-01 -4.38150048e-01 -2.84962058e-01
5.98288000e-01 -8.86342824e-01 3.60806510e-02 -8.22996855e-01
3.33707273e-01 -4.94354784e-01 -1.60745561e-01 1.18565165e-01
1.08164281e-01 -8.56552362e-01 -1.71535965e-02 1.51603997e-01
-2.08370164e-01 6.18652165e-01 2.36301050e-01 -3.42526674e-01
-6.22341096e-01 9.17114913e-02 -5.94671428e-01 6.77725255e-01
7.23689795e-01 -9.34199870e-01 -6.98175550e-01 -2.24523231e-01
5.07159233e-01 5.17198265e-01 5.19690395e-04 -1.20569444e+00
3.19764733e-01 -1.64639831e-01 -5.59659628e-03 -7.91450560e-01
1.19687334e-01 -8.27574730e-01 1.00256670e+00 9.22604263e-01
-4.50112462e-01 5.45710444e-01 4.67553735e-01 5.38712502e-01
-3.79157394e-01 -9.81608510e-01 4.86942708e-01 -1.55165002e-01
-1.12252891e+00 -5.00219882e-01 -6.93470418e-01 -4.48453069e-01
1.38847089e+00 -5.11864781e-01 -2.23355934e-01 2.74801850e-01
-8.05428207e-01 1.22036614e-01 8.47203910e-01 2.15910792e-01
4.46420193e-01 -9.79197800e-01 -3.35799068e-01 1.81798831e-01
2.16946378e-01 -5.89138031e-01 1.39944613e-01 1.73471317e-01
-8.02381814e-01 2.83130974e-01 -4.36586708e-01 -6.20109379e-01
-1.25356352e+00 8.60022247e-01 3.44628066e-01 -1.99728489e-01
1.35982350e-01 1.10109842e+00 -2.57401973e-01 -8.35214376e-01
1.69558719e-01 1.67364717e-01 -2.02720970e-01 -5.42695448e-02
1.79984286e-01 7.58419394e-01 -4.25152779e-01 -2.07685933e-01
-5.29256880e-01 3.24255615e-01 3.49764973e-01 -4.21178132e-01
1.55328906e+00 6.74363747e-02 -9.47420359e-01 3.74939561e-01
6.40640736e-01 -2.54513621e-01 -6.29492521e-01 1.84065640e-01
7.11473048e-01 -7.54410923e-01 -1.92388389e-02 -8.46825421e-01
-8.50317657e-01 4.71107155e-01 7.24509478e-01 1.19243212e-01
1.24577010e+00 -3.44381750e-01 -2.00948194e-01 5.10811567e-01
7.78953373e-01 -1.11958969e+00 -2.07353234e-01 2.80345857e-01
5.82094252e-01 -4.46717799e-01 2.83550359e-02 -3.20137084e-01
-5.31348884e-01 1.39085519e+00 6.51934028e-01 4.23311852e-02
4.50960785e-01 7.03568637e-01 -8.04437339e-01 -2.60222018e-01
-8.64757299e-01 2.36334488e-01 -1.33725077e-01 1.01223576e+00
3.22412625e-02 1.62317336e-01 -1.73760235e-01 4.74465698e-01
-8.88339803e-03 -1.00307375e-01 4.70809460e-01 1.45283842e+00
-4.56925124e-01 -1.51547730e+00 -7.80872226e-01 2.21015379e-01
1.69889748e-01 5.44857327e-03 -2.78622746e-01 8.76311064e-01
2.20062673e-01 7.40438342e-01 -1.23537913e-01 -6.53280497e-01
2.62081504e-01 3.04747581e-01 6.37434959e-01 -5.61355472e-01
-6.79361761e-01 -4.08725381e-01 2.98913300e-01 -6.81653440e-01
-1.91460401e-01 -9.17315364e-01 -7.87294745e-01 -2.75620103e-01
-5.07641912e-01 6.17605448e-01 1.07898855e+00 3.13504972e-02
4.80660111e-01 8.86161864e-01 5.74157596e-01 -6.02080762e-01
-8.06591868e-01 -1.89226419e-01 -7.08867490e-01 -4.53450322e-01
-3.06734234e-01 -8.28983247e-01 -2.37023696e-01 -5.39390780e-02] | [8.100727081298828, 7.253157615661621] |
75a41892-fc4f-4ada-89e4-29f487bac339 | time-in-a-box-advancing-knowledge-graph | 2111.06854 | null | https://arxiv.org/abs/2111.06854v1 | https://arxiv.org/pdf/2111.06854v1.pdf | Time in a Box: Advancing Knowledge Graph Completion with Temporal Scopes | Almost all statements in knowledge bases have a temporal scope during which they are valid. Hence, knowledge base completion (KBC) on temporal knowledge bases (TKB), where each statement \textit{may} be associated with a temporal scope, has attracted growing attention. Prior works assume that each statement in a TKB \textit{must} be associated with a temporal scope. This ignores the fact that the scoping information is commonly missing in a KB. Thus prior work is typically incapable of handling generic use cases where a TKB is composed of temporal statements with/without a known temporal scope. In order to address this issue, we establish a new knowledge base embedding framework, called TIME2BOX, that can deal with atemporal and temporal statements of different types simultaneously. Our main insight is that answers to a temporal query always belong to a subset of answers to a time-agnostic counterpart. Put differently, time is a filter that helps pick out answers to be correct during certain periods. We introduce boxes to represent a set of answer entities to a time-agnostic query. The filtering functionality of time is modeled by intersections over these boxes. In addition, we generalize current evaluation protocols on time interval prediction. We describe experiments on two datasets and show that the proposed method outperforms state-of-the-art (SOTA) methods on both link prediction and time prediction. | ['Gengchen Mai', 'Rui Zhu', 'Bo Yan', 'Krzysztof Janowic', 'Ling Cai'] | 2021-11-12 | null | null | null | null | ['knowledge-base-completion', 'knowledge-base-completion'] | ['graphs', 'knowledge-base'] | [-3.48314464e-01 2.43811607e-02 -9.69718099e-01 -5.04168630e-01
-3.52367818e-01 -8.50207746e-01 5.88732302e-01 5.81710279e-01
-1.36492342e-01 1.06575632e+00 2.62706161e-01 -5.62564135e-01
-5.42501926e-01 -1.40116370e+00 -8.62303853e-01 -2.98170388e-01
-3.49990815e-01 5.67500472e-01 8.39152992e-01 -4.55052227e-01
-3.13046992e-01 2.97355056e-01 -1.23070538e+00 2.96520293e-01
5.53314507e-01 1.13135409e+00 -3.60681355e-01 1.80543795e-01
-3.80626738e-01 1.14252651e+00 -7.09487677e-01 -7.09868789e-01
-1.08813301e-01 1.70790106e-01 -1.15305829e+00 -6.60774469e-01
1.42881557e-01 -2.75673062e-01 -7.87215531e-01 6.50675476e-01
1.89897697e-02 1.69668436e-01 4.27444935e-01 -1.83141315e+00
-5.43692172e-01 1.07311213e+00 -1.65349707e-01 6.48310781e-01
6.15181267e-01 -4.22560453e-01 1.36093354e+00 -4.52242881e-01
8.89131248e-01 9.16105449e-01 7.02887952e-01 3.20965439e-01
-1.20958436e+00 -6.37396932e-01 5.10521650e-01 8.36292684e-01
-1.36452210e+00 -2.99804062e-01 7.96664238e-01 -2.47596473e-01
1.12612092e+00 6.35110021e-01 6.41978204e-01 8.48847270e-01
1.19947836e-01 8.38129282e-01 6.59052074e-01 3.20425793e-03
3.50278765e-01 1.60323128e-01 6.11070037e-01 2.85678029e-01
1.83650434e-01 -2.94686668e-02 -8.91812563e-01 -2.35047519e-01
2.76864111e-01 -6.42548956e-04 -5.28497815e-01 -3.70174021e-01
-1.27672827e+00 6.80623055e-01 3.56992722e-01 2.02389166e-01
-9.19010267e-02 2.58311152e-01 6.90162838e-01 6.10570371e-01
2.68465996e-01 2.23361403e-02 -8.06927621e-01 -1.07219353e-01
-7.64545918e-01 3.85908902e-01 1.26696587e+00 1.29240191e+00
7.81093717e-01 -2.36000806e-01 -3.50877762e-01 3.12776029e-01
7.69130290e-02 1.88381895e-01 -1.38382157e-02 -7.02116728e-01
5.45456588e-01 8.38757694e-01 6.25590444e-01 -1.25088930e+00
-3.55573475e-01 -2.36669242e-01 -5.17255306e-01 -4.83776778e-01
3.97395521e-01 3.24730873e-01 -3.56737703e-01 1.96895635e+00
6.22407198e-01 4.06113148e-01 2.13496655e-01 6.46339118e-01
8.59125674e-01 7.20781684e-01 -1.04817852e-01 -5.35048604e-01
1.43340445e+00 -5.86087704e-01 -9.58691657e-01 1.27404124e-01
5.49430728e-01 -2.10254073e-01 6.11137450e-01 1.75215110e-01
-7.39672601e-01 -1.90689862e-01 -1.07192111e+00 5.59198158e-03
-8.28219593e-01 -3.66126567e-01 8.73387158e-01 6.30548179e-01
-8.30100060e-01 3.51582855e-01 -7.80304492e-01 -3.02334696e-01
-1.34904519e-01 2.17864782e-01 -4.02764589e-01 -5.83235659e-02
-2.01438761e+00 8.07619750e-01 6.74878001e-01 -1.37777805e-01
-6.53551877e-01 -9.12282288e-01 -9.83436286e-01 4.03554067e-02
1.00546324e+00 -6.30550206e-01 1.19133949e+00 -2.18892500e-01
-9.09195602e-01 4.81229335e-01 -4.08081293e-01 -9.08062458e-01
3.68010789e-01 -8.11118782e-02 -1.24030912e+00 1.70871779e-01
3.36740196e-01 7.90457502e-02 6.96074605e-01 -8.11629295e-01
-8.20564151e-01 -2.23741785e-01 8.09943318e-01 -3.09712052e-01
-4.70342547e-01 2.48849802e-02 -9.39798057e-01 -6.24436378e-01
2.00420301e-02 -5.51742435e-01 3.13283741e-01 -1.54383302e-01
-5.40123999e-01 -4.47619706e-01 8.51585269e-01 -5.37468612e-01
2.15680528e+00 -1.73857379e+00 5.79799451e-02 2.09498629e-01
1.50948107e-01 -2.58584678e-01 3.62737954e-01 9.55975056e-01
-1.58416137e-01 2.68971413e-01 1.47683457e-01 -6.81010960e-03
3.64334911e-01 5.43185771e-01 -1.04365003e+00 5.03530800e-01
-1.81015432e-01 8.20155919e-01 -1.01703680e+00 -6.28560424e-01
-3.69068570e-02 1.32860526e-01 -3.21842462e-01 -1.85866654e-01
-7.70821631e-01 -6.03977442e-02 -5.20445824e-01 1.01222825e+00
4.68999088e-01 -4.32035416e-01 3.52302670e-01 -6.47957027e-01
-1.25681370e-01 5.39429128e-01 -1.45323956e+00 1.27116311e+00
-4.39948142e-01 1.84070662e-01 -1.94797307e-01 -8.66178513e-01
6.67858303e-01 6.03012562e-01 6.64735794e-01 -4.97352928e-01
-3.71452659e-01 1.12430140e-01 -4.27551568e-01 -3.42408538e-01
6.60034299e-01 9.92976204e-02 -2.98812985e-01 4.50385571e-01
-1.58798024e-01 2.49961212e-01 5.54930627e-01 5.02988338e-01
1.35491872e+00 5.31838201e-02 4.94286001e-01 -2.67599393e-02
6.59188092e-01 2.25387737e-01 1.04435360e+00 6.72843456e-01
-2.67059326e-01 -1.32714182e-01 7.11161971e-01 -5.82524896e-01
-5.72285295e-01 -9.93847489e-01 -8.47591162e-02 1.04114294e+00
3.37697089e-01 -9.91073608e-01 2.35430494e-01 -1.05684030e+00
4.37009662e-01 8.18708897e-01 -7.41521358e-01 -1.76152095e-01
-5.19538879e-01 -2.85485178e-01 5.30129850e-01 8.41822326e-01
3.33459079e-01 -4.61327195e-01 -4.44424897e-01 5.28873503e-01
-4.72953379e-01 -1.04407334e+00 -6.26607478e-01 4.87440266e-02
-6.48388386e-01 -1.44713545e+00 -8.11026171e-02 -3.85102093e-01
2.39398748e-01 7.49212503e-02 1.38615191e+00 -2.14020878e-01
3.79167378e-01 6.93725824e-01 -4.92298514e-01 -3.10637832e-01
-4.78451187e-03 -1.71630606e-01 1.07619934e-01 2.75478940e-02
6.08330548e-01 -6.96375966e-01 -5.66096544e-01 7.05043912e-01
-1.04472721e+00 -1.94212094e-01 -1.69654205e-01 5.78057647e-01
8.19822729e-01 7.36574113e-01 7.86711633e-01 -8.36493611e-01
3.61526906e-01 -8.68648469e-01 -6.53239369e-01 6.65565193e-01
-6.06580734e-01 1.42849982e-01 6.20091736e-01 -6.16492391e-01
-9.69483495e-01 -3.79595369e-01 3.75623524e-01 -3.63003820e-01
3.13658804e-01 1.16867435e+00 -2.14335367e-01 3.22256863e-01
5.22074103e-01 2.65930921e-01 -5.35308540e-01 -1.86331779e-01
4.41719472e-01 1.49578542e-01 5.48515499e-01 -9.61212993e-01
8.68448496e-01 8.65000844e-01 1.01736091e-01 -3.62952560e-01
-7.79113710e-01 -7.57827878e-01 -1.86865553e-01 -2.73493797e-01
3.06785434e-01 -8.62885714e-01 -9.69454288e-01 -2.15238426e-02
-1.15048540e+00 1.31872930e-02 -2.42784798e-01 2.75221616e-01
-5.71926296e-01 3.18607748e-01 -3.71194482e-01 -7.93228984e-01
-1.51479855e-01 -6.41039014e-01 6.34318948e-01 -1.75613895e-01
-5.19945383e-01 -1.08964860e+00 -2.76805516e-02 5.05046919e-02
1.47851378e-01 2.63455659e-01 1.14990044e+00 -8.67300808e-01
-8.59197021e-01 -4.53815311e-01 5.46427369e-02 -1.78738192e-01
2.21392348e-01 -1.56528950e-02 -4.55035150e-01 -1.24215245e-01
-1.85042515e-01 3.70892845e-02 7.22131729e-01 -9.72712934e-02
9.41263437e-01 -6.72491789e-01 -8.25668633e-01 2.78720319e-01
1.41355550e+00 1.90351695e-01 5.38337409e-01 4.01663512e-01
2.09598586e-01 5.24768770e-01 8.97308052e-01 5.22720039e-01
9.29134548e-01 9.71882761e-01 5.00485063e-01 7.11958349e-01
1.43550843e-01 -2.92801112e-01 3.21077764e-01 3.52447957e-01
-4.42477828e-03 -4.10817266e-01 -9.11803544e-01 1.16543293e+00
-1.99312735e+00 -1.35864174e+00 -2.80078530e-01 2.31793737e+00
1.32470655e+00 3.05351347e-01 2.38695651e-01 4.69912261e-01
5.34565151e-01 3.49764913e-01 -4.50103551e-01 4.58437949e-02
-1.25786439e-01 -6.18824828e-03 5.07661283e-01 3.85569721e-01
-1.13219631e+00 6.47353709e-01 5.41527176e+00 7.78402627e-01
-9.65094149e-01 1.30013809e-01 -1.05055176e-01 1.21396452e-01
-7.04783320e-01 2.01847687e-01 -1.14072776e+00 5.91457486e-01
9.62892532e-01 -7.53068924e-01 2.19376355e-01 8.27055514e-01
-8.14167876e-03 4.66140769e-02 -1.59424770e+00 6.58807397e-01
-2.98073024e-01 -1.58654702e+00 -3.30869406e-02 -3.72742027e-01
5.49337029e-01 -4.40040588e-01 -1.09187625e-01 6.86669350e-01
3.28811020e-01 -5.93493640e-01 1.03230822e+00 7.50126779e-01
7.43515193e-01 -5.99285483e-01 6.65782571e-01 2.06432819e-01
-1.75854671e+00 -1.12552509e-01 1.10570276e-02 2.07814991e-01
2.74554729e-01 8.02853346e-01 -8.06676209e-01 1.23733139e+00
8.70326519e-01 8.76975179e-01 -9.98570248e-02 9.95913029e-01
-3.07175159e-01 6.99617028e-01 -5.79502940e-01 4.65774611e-02
-1.51039585e-01 2.50701666e-01 5.75980067e-01 9.70019698e-01
2.74890363e-01 7.01842681e-02 2.68882941e-02 7.88507402e-01
-4.96405773e-02 -2.07411885e-01 -5.33924460e-01 3.82375307e-02
1.08074272e+00 6.78047657e-01 -3.80714655e-01 -4.08156723e-01
-7.95196295e-01 5.92578590e-01 6.55196011e-02 7.26905942e-01
-1.24343848e+00 -3.85014802e-01 7.24529266e-01 9.19975489e-02
4.97817665e-01 -1.70094937e-01 -8.48993752e-03 -1.17484224e+00
4.47133213e-01 -4.96602118e-01 1.11852098e+00 -6.37204647e-01
-1.30189061e+00 1.50703639e-01 4.81029660e-01 -1.30896163e+00
-7.87860230e-02 -2.43228436e-01 -5.25118530e-01 7.78084815e-01
-1.76014304e+00 -1.18058586e+00 -9.81516093e-02 1.00149250e+00
2.59719137e-02 2.81134546e-01 8.71420622e-01 4.96181965e-01
-4.04472142e-01 5.71872413e-01 -2.45154530e-01 1.27461264e-02
6.92741394e-01 -1.33790255e+00 -6.77302014e-03 7.80661762e-01
-1.09535962e-01 1.08938801e+00 9.10181403e-01 -9.41641569e-01
-1.68411136e+00 -1.39480937e+00 1.38677990e+00 -6.73106968e-01
1.11621952e+00 1.17251072e-02 -1.03599405e+00 1.25227630e+00
-1.48136616e-01 1.60873443e-01 6.47922814e-01 4.44676727e-01
-8.56400728e-01 -7.32733905e-01 -9.82047260e-01 5.25272310e-01
8.77088726e-01 -8.42719197e-01 -8.83131623e-01 3.10436666e-01
1.10556245e+00 -4.09335554e-01 -1.26021612e+00 6.12893939e-01
4.02521133e-01 -6.92653954e-01 1.10845244e+00 -7.87737429e-01
-3.15525606e-02 -7.49011219e-01 -3.66769642e-01 -7.70877004e-01
1.34393990e-01 -7.67852724e-01 -1.18628335e+00 1.36600637e+00
3.02983195e-01 -5.75716555e-01 8.25178146e-01 6.77148521e-01
-3.22257541e-03 -6.28359914e-01 -1.25193763e+00 -1.22375131e+00
-3.12351763e-01 -8.15487802e-01 1.12534630e+00 1.14231932e+00
4.09102321e-01 -6.34023771e-02 -3.47950578e-01 5.30647039e-01
3.11717868e-01 5.23835480e-01 5.18079102e-01 -1.35580361e+00
-1.77369341e-01 -1.87143967e-01 -3.17767113e-01 -9.61166799e-01
1.56794846e-01 -8.20222020e-01 -3.90676588e-01 -1.68486547e+00
-2.64974087e-01 -5.91727257e-01 -3.27073961e-01 8.17282617e-01
2.60115236e-01 -3.29560995e-01 -3.97619188e-01 1.20499440e-01
-7.44418383e-01 4.12556440e-01 8.08683038e-01 -5.82831085e-01
-2.17074037e-01 2.67148316e-01 -3.38449091e-01 3.51864010e-01
5.50351560e-01 -4.35562283e-01 -7.98770070e-01 -1.76846996e-01
1.12294805e+00 5.99126101e-01 5.13839185e-01 -8.43170464e-01
6.05345368e-01 -4.34841067e-01 -3.60733598e-01 -9.25505817e-01
4.47755128e-01 -1.18957460e+00 6.88353658e-01 1.32463917e-01
-8.28283057e-02 -1.24606229e-01 3.12107444e-01 1.05286825e+00
-4.90247279e-01 3.34137641e-02 -1.06536232e-01 1.55495748e-01
-1.21491742e+00 4.57331866e-01 -6.72632828e-02 -1.51885390e-01
1.15589213e+00 -1.23449244e-01 -6.97462082e-01 -4.79064882e-01
-7.59126782e-01 6.43966854e-01 1.24379806e-01 3.28237683e-01
6.31166399e-01 -1.56335568e+00 -2.19719425e-01 -3.93027842e-01
6.64307296e-01 -7.16343746e-02 3.18206757e-01 1.19095194e+00
-1.48947388e-02 6.63711190e-01 3.59846950e-01 -1.28393769e-01
-9.86675203e-01 9.90208566e-01 2.86907643e-01 -4.06117380e-01
-4.33943242e-01 8.35224628e-01 -2.75965869e-01 -1.90888852e-01
5.30451357e-01 -6.60577714e-01 -3.51881266e-01 3.59245837e-01
5.33146918e-01 4.70022976e-01 2.23602951e-01 -3.58644933e-01
-7.24913716e-01 1.93741843e-01 -7.89156780e-02 7.15360269e-02
1.11349869e+00 -1.45431042e-01 -3.96011114e-01 6.62551820e-01
8.56408477e-01 3.27197224e-01 -5.70373952e-01 -5.86181819e-01
2.01726303e-01 -2.82482356e-01 -2.88889289e-01 -9.72653866e-01
-8.56224895e-01 2.40521640e-01 -1.06446423e-01 6.00257874e-01
1.21513557e+00 2.37511083e-01 8.51365447e-01 6.07545018e-01
7.47800469e-01 -1.05199480e+00 -2.16577068e-01 5.57046235e-01
8.83360684e-01 -9.76729929e-01 -1.09622248e-01 -5.78754485e-01
-2.20371455e-01 1.11266267e+00 5.60270488e-01 6.02041900e-01
7.64807463e-01 8.04754868e-02 -4.02568638e-01 -2.86031544e-01
-1.09168422e+00 -1.41463637e-01 2.88012266e-01 5.66487312e-01
7.99456164e-02 1.84642136e-01 -4.05901402e-01 1.24179447e+00
-9.87917259e-02 2.69392699e-01 4.41803187e-01 1.14259064e+00
-2.08037257e-01 -1.12076342e+00 -4.21251982e-01 3.30936223e-01
-3.45588267e-01 8.70467201e-02 -9.11066234e-02 9.99323428e-01
3.20915356e-02 1.12748909e+00 -1.67378142e-01 -6.80200815e-01
4.29971665e-01 2.90619582e-01 1.72808588e-01 -4.96060491e-01
-2.44503394e-01 -5.02928376e-01 4.75590080e-01 -8.05092275e-01
-5.64911246e-01 -5.69239438e-01 -1.42041063e+00 -5.47205508e-01
-1.99917078e-01 3.63417089e-01 4.05139551e-02 6.39533043e-01
1.53193861e-01 5.66165984e-01 2.22116143e-01 3.80182922e-01
-2.98767835e-01 -4.19275254e-01 -7.00639844e-01 2.34813750e-01
4.94206607e-01 -8.89994740e-01 -8.61710533e-02 1.14285029e-01] | [8.533204078674316, 7.928412437438965] |
94c44b8e-562d-499d-b19f-f21cdbad3ee6 | quantitative-dynamics-of-design-thinking-and | 2306.10971 | null | https://arxiv.org/abs/2306.10971v1 | https://arxiv.org/pdf/2306.10971v1.pdf | Quantitative dynamics of design thinking and creativity perspectives in company context | This study is intended to provide in-depth insights into how design thinking and creativity issues are understood and possibly evolve in the course of design discussions in a company context. For that purpose, we use the seminar transcripts of the Design Thinking Research Symposium 12 (DTRS12) dataset "Tech-centred Design Thinking: Perspectives from a Rising Asia," which are primarily concerned with how Korean companies implement design thinking and what role designers currently play. We employed a novel method of information processing based on constructed dynamic semantic networks to investigate the seminar discussions according to company representatives and company size. We compared the quantitative dynamics in two seminars: the first involved managerial representatives of four companies, and the second involved specialized designers and management of a design center of single company. On the basis of dynamic semantic networks, we quantified the changes in four semantic measures -- abstraction, polysemy, information content, and pairwise word similarity -- in chronologically reconstructed individual design-thinking processes. Statistical analyses show that design thinking in the seminar with four companies, exhibits significant differences in the dynamics of abstraction, polysemy, and information content, compared to the seminar with the design center of single company. Both the decrease in polysemy and abstraction and the increase in information content in the individual design-thinking processes in the seminar with four companies indicate that design managers are focused on more concrete design issues, with more information and less ambiguous content to the final design product. By contrast, specialized designers manifest more abstract thinking and appear to exhibit a slightly higher level of divergence in their design processes. | ['Danko D. Georgiev', 'Georgi V. Georgiev'] | 2023-06-19 | null | null | null | null | ['word-similarity'] | ['natural-language-processing'] | [-1.03758477e-01 -7.54763633e-02 3.55648547e-02 -5.30410558e-02
4.13141817e-01 -8.08558643e-01 6.11629009e-01 3.18020046e-01
-2.19270796e-01 -3.83609891e-01 1.04268241e+00 -5.99499464e-01
-1.06130815e+00 -7.78987527e-01 2.76954383e-01 -2.90383726e-01
5.98307252e-01 6.56267762e-01 -3.49631011e-01 -4.53535229e-01
1.04459381e+00 4.16687846e-01 -1.35109925e+00 8.52142513e-01
4.98766750e-01 6.28842175e-01 4.99785990e-01 3.35848220e-02
-8.36841583e-01 6.02487504e-01 -5.18463314e-01 -3.33562315e-01
1.08531311e-01 -4.86613959e-01 -9.80650425e-01 5.49805343e-01
-1.36668429e-01 3.03012908e-01 2.81455547e-01 1.18890870e+00
3.16240162e-01 3.25420946e-02 2.99286097e-01 -6.48893893e-01
-1.03908372e+00 1.17492068e+00 -1.73387751e-01 2.27198750e-01
3.90351832e-01 3.16880971e-01 1.29056633e+00 -1.00593019e+00
1.09889638e+00 1.57385206e+00 6.54887080e-01 1.11672416e-01
-1.40168250e+00 -6.05034947e-01 3.56265783e-01 9.14777443e-02
-1.11541033e+00 4.52153990e-03 9.41874683e-01 -1.14649737e+00
7.85919547e-01 5.37109114e-02 1.06134605e+00 8.76678944e-01
1.65315042e-03 -4.41193068e-03 1.22222066e+00 -4.65055436e-01
6.41445100e-01 7.18527734e-01 5.16415715e-01 2.23932378e-02
4.38120812e-01 -4.69887942e-01 -4.29620624e-01 1.26232564e-01
7.65285015e-01 2.01820821e-01 -4.09446321e-02 8.77457485e-03
-1.41003704e+00 8.64370763e-01 1.16105072e-01 1.40761757e+00
-3.45288754e-01 -2.18974605e-01 5.30018210e-01 5.10589719e-01
2.06952497e-01 1.00718391e+00 -2.70687431e-01 -5.78551948e-01
-7.10335195e-01 2.59545594e-01 1.02647328e+00 6.20221198e-01
3.49007100e-01 -3.91147584e-02 2.20514089e-01 5.37912726e-01
4.62688386e-01 -9.77216735e-02 5.73783219e-01 -9.40065622e-01
5.56440473e-01 1.15696442e+00 -2.10270286e-01 -1.50113142e+00
-2.93048114e-01 -6.72381818e-01 -2.04094231e-01 9.84947756e-02
2.57839095e-02 -4.99308594e-02 -4.74038035e-01 1.25125146e+00
-3.75581235e-01 -1.15313506e+00 -1.77635685e-01 9.84247386e-01
5.59903085e-01 3.30984473e-01 2.05552671e-02 -2.54407883e-01
1.70127606e+00 -7.63982952e-01 -1.06202376e+00 -3.19959372e-01
4.70278203e-01 -8.24895561e-01 1.46337879e+00 5.08664191e-01
-1.13827384e+00 -8.44889283e-01 -9.77244377e-01 2.54839510e-01
-4.93471086e-01 -1.34314805e-01 3.86196852e-01 8.78140569e-01
-1.05922091e+00 4.99617279e-01 -2.65655816e-01 -8.86583269e-01
1.38464332e-01 -2.29537696e-01 1.33032829e-01 6.33969307e-02
-8.32962692e-01 7.50522614e-01 3.63118589e-01 4.78360616e-02
3.18864770e-02 -8.96632910e-01 -1.61709324e-01 2.47900024e-01
3.99634659e-01 -6.72196865e-01 9.92162824e-01 -1.27321362e+00
-1.23627603e+00 5.52923679e-01 4.52908665e-01 3.04117262e-01
1.90123662e-01 2.01829150e-01 -8.17768633e-01 -1.27633922e-02
1.16225787e-01 1.75295636e-01 1.63463295e-01 -1.32364070e+00
-4.33810055e-01 -5.48327327e-01 1.07798681e-01 -3.71704367e-03
-6.29048467e-01 3.41464400e-01 1.76159978e-01 -6.51686847e-01
7.66675889e-01 -5.33431590e-01 -1.28806308e-01 -3.12436104e-01
1.45708071e-02 -1.80873856e-01 4.89027768e-01 -4.50355291e-01
1.64303243e+00 -2.32523370e+00 2.78830647e-01 5.19573569e-01
6.37124836e-01 -2.37979814e-01 2.13526294e-01 1.21078515e+00
-2.31781974e-01 6.52280986e-01 3.64166319e-01 1.46241695e-01
3.77061874e-01 7.56498352e-02 -6.92368671e-02 -5.33809066e-02
-2.99380511e-01 6.37075901e-01 -9.38436687e-01 6.89889491e-02
-2.50752926e-01 -1.78776070e-01 -4.66667801e-01 -5.67467868e-01
-1.91984504e-01 2.90604323e-01 -4.35845554e-01 7.02024698e-01
4.12144154e-01 -3.86713803e-01 7.21087396e-01 -2.17422359e-02
-9.64441955e-01 4.28971946e-01 -8.92959297e-01 1.58051586e+00
-5.73319256e-01 1.07855177e+00 -1.61303788e-01 -6.57420576e-01
1.20547581e+00 2.32564747e-01 1.13471478e-01 -9.37021136e-01
2.73577452e-01 1.52126893e-01 7.54684210e-01 -5.90268970e-01
5.80699861e-01 -4.48786169e-01 -1.10598132e-01 6.57278538e-01
-3.52710694e-01 -1.11752480e-01 4.18805957e-01 4.40574214e-02
1.10932040e+00 -6.37810171e-01 -9.86439735e-02 -7.81916082e-01
2.79628873e-01 -4.93356362e-02 6.31807327e-01 2.04303280e-01
1.06480435e-01 1.85442030e-01 9.56813633e-01 -4.69297945e-01
-9.61329281e-01 -9.43176150e-01 1.31560013e-01 6.17483795e-01
2.48949639e-02 -9.23755050e-01 -5.95433056e-01 -1.34848312e-01
-4.68389802e-02 1.25919604e+00 -3.29502285e-01 -1.15404269e-02
-2.49400273e-01 2.33463645e-01 -7.49997497e-02 3.80443454e-01
7.53762543e-01 -9.93227601e-01 -8.04574251e-01 1.44504160e-01
7.39476234e-02 -7.57984638e-01 -2.21493378e-01 -6.66430071e-02
-9.69711602e-01 -5.81391454e-01 -9.70164835e-02 -6.23756886e-01
7.14620173e-01 8.42854828e-02 9.81464386e-01 -8.15287158e-02
-1.93468392e-01 3.99939060e-01 -4.42831427e-01 -3.39428633e-01
-4.63009685e-01 1.12218494e-02 -3.69736046e-01 -3.44226420e-01
5.77267289e-01 -5.73193133e-01 -4.18528557e-01 4.14674550e-01
-9.72703278e-01 7.45558441e-02 4.47175860e-01 2.19709337e-01
-8.22635144e-02 8.71725798e-01 2.58717477e-01 -5.76237082e-01
1.19058323e+00 -4.95460957e-01 -9.83260721e-02 6.24755882e-02
-1.23701704e+00 -2.58743227e-01 3.60729575e-01 -2.37844497e-01
-1.30009699e+00 -8.12326550e-01 4.62555796e-01 1.30266532e-01
5.50069325e-02 9.92922068e-01 -3.98245603e-02 3.14074963e-01
5.01320541e-01 -1.89706132e-01 1.41743273e-01 -5.20740330e-01
8.34921822e-02 6.72268212e-01 1.01090902e-02 -6.25048816e-01
5.85381687e-01 3.35654229e-01 -4.12320852e-01 -7.02772856e-01
-3.00154597e-01 -3.83988947e-01 -4.19143528e-01 -5.80178320e-01
9.69957173e-01 -5.32239318e-01 -9.78444099e-01 1.61172803e-02
-1.01887298e+00 7.76102617e-02 -6.40052140e-01 5.06216824e-01
-4.36930090e-01 -3.42100412e-02 -3.16945225e-01 -6.97653174e-01
3.71924904e-03 -1.20316947e+00 3.69697511e-01 1.21225819e-01
-1.06739640e+00 -1.40901840e+00 -5.49758486e-02 7.39458203e-01
4.51804131e-01 -1.36852749e-02 1.60045493e+00 -6.84861422e-01
-3.94591033e-01 3.08369011e-01 -2.14308679e-01 2.90127665e-01
5.21146595e-01 -6.85158595e-02 -2.11267903e-01 1.82818547e-01
5.65362990e-01 4.64502603e-01 2.48053625e-01 -1.83509558e-01
6.17513597e-01 -2.85677075e-01 1.60121154e-02 -1.53218016e-01
1.72040665e+00 1.00213909e+00 5.58384418e-01 5.78746617e-01
2.71725982e-01 1.28743768e+00 4.07903016e-01 7.37006903e-01
1.73407614e-01 3.99782598e-01 -3.40626568e-01 6.52761281e-01
-7.17475414e-02 -1.54423818e-01 2.25824416e-01 1.32062423e+00
-9.51698720e-02 -1.33333862e-01 -1.48867404e+00 6.94011927e-01
-1.80170810e+00 -8.24046731e-01 -3.29189360e-01 1.53234708e+00
3.02813977e-01 3.65301996e-01 1.29420266e-01 4.39611897e-02
5.29362202e-01 -8.87144953e-02 -6.14674389e-02 -9.34136271e-01
2.01227933e-01 2.34443117e-02 1.05255507e-01 -7.10963383e-02
2.68797904e-01 5.75620055e-01 6.07984543e+00 3.43533665e-01
-7.75205374e-01 3.37085165e-02 2.24948719e-01 -2.33293325e-01
-1.21410227e+00 4.22402442e-01 -3.55786502e-01 6.90486550e-01
8.65730643e-01 -3.37402552e-01 2.83018708e-01 5.99506378e-01
6.76247776e-01 -1.57483980e-01 -1.12654412e+00 8.57452810e-01
-1.44324839e-01 -1.58817816e+00 1.83197066e-01 3.37188244e-01
7.34508812e-01 -8.51590395e-01 2.98509926e-01 -3.13359201e-01
7.00663328e-02 -6.38292134e-01 1.12972045e+00 6.99544907e-01
4.94524427e-02 -6.66994452e-01 6.49737716e-01 -6.31339774e-02
-9.82427776e-01 -7.56862402e-01 1.58240087e-02 -7.60556340e-01
4.22178626e-01 6.91142261e-01 -5.44163287e-01 4.49638993e-01
7.11806476e-01 4.55060035e-01 -1.98391885e-01 2.48082474e-01
7.93117881e-02 1.87305808e-01 4.96428728e-01 -2.78387129e-01
2.42916137e-01 -6.90603733e-01 4.74492252e-01 7.67499030e-01
3.47993851e-01 -1.22556143e-01 -1.91437379e-01 1.64913261e+00
4.28544074e-01 -1.11250661e-01 -2.01143518e-01 -1.11427224e+00
4.46747690e-01 1.04237628e+00 -1.48713982e+00 -6.91418573e-02
-4.00686026e-01 4.57270324e-01 -4.57923114e-01 2.83821702e-01
-4.77611661e-01 -4.84510481e-01 9.72067475e-01 7.14891076e-01
2.40521058e-01 -6.64584100e-01 -1.19044292e+00 -3.62686247e-01
4.40812521e-02 -8.30512285e-01 -2.87926853e-01 -6.55652106e-01
-1.33149374e+00 2.69956470e-01 1.14345647e-01 -6.42623484e-01
3.23360980e-01 -4.36549336e-01 -6.23664439e-01 9.13537502e-01
-2.86611646e-01 -6.67686820e-01 -4.24922734e-01 -1.18744537e-01
8.78138542e-01 -3.68315965e-01 2.38895103e-01 3.14976752e-01
-4.30369079e-01 -8.02183151e-02 -1.77116245e-01 -3.13182026e-01
2.25213364e-01 -9.74841058e-01 5.90135157e-01 2.05471203e-01
-1.99827552e-01 1.10615432e+00 5.65470099e-01 -1.00140738e+00
-1.22025883e+00 -6.91204295e-02 1.28923643e+00 -3.79067928e-01
1.20425546e+00 -3.40930998e-01 -3.79338950e-01 3.26778382e-01
5.65850556e-01 -1.19331551e+00 9.66167867e-01 1.88357741e-01
5.18447608e-02 -3.40881906e-02 -1.03590524e+00 9.27281082e-01
1.58736634e+00 -6.12137258e-01 -9.63807344e-01 3.47875595e-01
1.03894532e+00 5.02444208e-01 -1.00023949e+00 -3.80686730e-01
5.33794522e-01 -1.03820908e+00 6.43461704e-01 -1.35054186e-01
4.11928415e-01 -1.03753582e-01 -1.04167432e-01 -9.08023059e-01
-1.05035210e+00 -5.24105966e-01 1.07246172e+00 1.57543135e+00
1.07132249e-01 -9.40043449e-01 5.14748991e-01 7.01278985e-01
-4.11376238e-01 -3.07070971e-01 -4.30054307e-01 -7.42791355e-01
5.89991547e-03 -2.79756635e-01 6.24365032e-01 8.55873048e-01
6.13925993e-01 3.75065356e-01 8.00820768e-01 -6.49759173e-01
2.36696720e-01 5.18404320e-03 4.58724834e-02 -1.84612918e+00
-8.16987306e-02 -9.03820336e-01 -5.44115007e-01 -3.06172401e-01
-1.93216726e-02 -9.39420879e-01 -5.21724343e-01 -1.89280593e+00
6.26576617e-02 -1.17434993e-01 1.21054366e-01 -1.14040457e-01
9.44405138e-01 -6.53191507e-01 4.49282229e-01 3.73542070e-01
6.95048869e-02 -6.70697354e-03 1.34680331e+00 1.64742783e-01
-5.96135437e-01 -1.71297535e-01 -1.45459783e+00 7.48916507e-01
6.92309558e-01 -1.84950441e-01 -9.67813551e-01 -3.87261957e-01
1.13248849e+00 -2.06803247e-01 2.28833169e-01 -9.36723828e-01
4.66986090e-01 -2.29288474e-01 1.01957403e-01 -4.90597874e-01
-2.56103724e-01 -1.22510993e+00 1.06959474e+00 6.68522656e-01
-4.05912459e-01 6.62701249e-01 1.25508979e-01 8.90543237e-02
-2.28427187e-01 -4.91587460e-01 1.86466455e-01 -4.01685297e-01
-7.29656577e-01 -6.65444672e-01 -9.35153365e-01 -3.27442378e-01
9.80963230e-01 -9.14244354e-01 -2.99207419e-01 -3.94610725e-02
-9.50047791e-01 -3.25713813e-01 6.23683870e-01 6.25294805e-01
5.22958338e-01 -1.27576816e+00 -6.28158525e-02 8.11203495e-02
-1.24676332e-01 -3.51780653e-01 3.25429171e-01 9.36550438e-01
-6.53563082e-01 5.41314721e-01 -4.86591399e-01 7.98539445e-03
-8.24924529e-01 4.20556724e-01 -2.95035690e-01 1.90943256e-02
-4.84895259e-01 6.17550492e-01 3.49436760e-01 -8.25697482e-02
-1.92055434e-01 -7.96830654e-01 -9.76566970e-02 1.07184792e+00
2.43184030e-01 1.11069643e+00 -7.28903338e-02 -3.44123214e-01
-2.72770792e-01 7.71623611e-01 1.03711691e-02 -4.86768782e-01
1.11947286e+00 -3.30876201e-01 -6.85612142e-01 8.97855401e-01
9.95211482e-01 -1.35615185e-01 -3.39819789e-01 -3.58412340e-02
6.81742191e-01 -7.99141765e-01 -7.07948133e-02 -9.89543557e-01
-7.97494411e-01 5.96229970e-01 4.97404784e-01 6.89930797e-01
9.41450477e-01 2.20472232e-01 4.07594979e-01 1.34183422e-01
2.12514400e-01 -1.63027084e+00 6.19722009e-01 4.42753255e-01
1.10402417e+00 -8.71245638e-02 -2.21869811e-01 -4.88734305e-01
-7.00632155e-01 1.37669957e+00 5.55021942e-01 3.86194766e-01
7.60538161e-01 1.71699956e-01 -1.60549849e-01 -8.26907218e-01
-6.00258529e-01 -2.01985706e-02 7.32849464e-02 3.40724677e-01
5.37564397e-01 5.17674573e-02 -6.92208886e-01 8.88879478e-01
-5.26699305e-01 9.13848430e-02 6.41813874e-01 8.52293551e-01
-5.41972458e-01 -1.11097181e+00 -4.00989473e-01 4.12269644e-02
-7.41812959e-02 -1.34252146e-01 -9.91608858e-01 8.58853281e-01
6.00422680e-01 9.10946608e-01 7.20052361e-01 -7.39521265e-01
6.52268708e-01 1.63434342e-01 1.66569978e-01 -7.60308623e-01
-1.17438972e+00 1.53868973e-01 3.18996310e-01 -3.15488636e-01
-3.55391800e-01 -1.05944371e+00 -1.50635743e+00 -6.92747295e-01
-1.26306072e-01 1.87009797e-01 1.09265864e+00 7.90166020e-01
5.13734400e-01 1.12776875e+00 1.84996590e-01 -1.05194166e-01
-3.14507008e-01 -8.02801609e-01 -9.45156097e-01 -2.84795854e-02
-3.01164418e-01 -5.54862916e-01 -4.99013335e-01 -1.69131413e-01] | [9.16716480255127, 6.486791133880615] |
a2982b83-7fff-463f-860a-5130c3591ec7 | mining-false-positive-examples-for-text-based | 2303.08466 | null | https://arxiv.org/abs/2303.08466v1 | https://arxiv.org/pdf/2303.08466v1.pdf | Mining False Positive Examples for Text-Based Person Re-identification | Text-based person re-identification (ReID) aims to identify images of the targeted person from a large-scale person image database according to a given textual description. However, due to significant inter-modal gaps, text-based person ReID remains a challenging problem. Most existing methods generally rely heavily on the similarity contributed by matched word-region pairs, while neglecting mismatched word-region pairs which may play a decisive role. Accordingly, we propose to mine false positive examples (MFPE) via a jointly optimized multi-branch architecture to handle this problem. MFPE contains three branches including a false positive mining (FPM) branch to highlight the role of mismatched word-region pairs. Besides, MFPE delicately designs a cross-relu loss to increase the gap of similarity scores between matched and mismatched word-region pairs. Extensive experiments on CUHK-PEDES demonstrate the superior effectiveness of MFPE. Our code is released at https://github.com/xx-adeline/MFPE. | ['Changxing Ding', 'Zhiyin Shao', 'Wenhao Xu'] | 2023-03-15 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [ 5.81451319e-02 -4.91750807e-01 -6.93758652e-02 -4.80978221e-01
-9.32134569e-01 -3.50693196e-01 6.98765218e-01 4.44269218e-02
-8.44394505e-01 7.30460584e-01 2.62253940e-01 9.93291959e-02
-9.91558097e-03 -5.48460305e-01 -4.37670320e-01 -6.81842327e-01
5.77086620e-02 3.54251832e-01 8.49085525e-02 -1.79982498e-01
2.57273793e-01 2.32979655e-01 -1.39045095e+00 1.16716132e-01
7.99767971e-01 5.89734733e-01 8.59847814e-02 2.97679037e-01
1.42783150e-02 2.23473415e-01 -7.47997165e-01 -1.08248115e+00
5.48243225e-01 -2.98530698e-01 -5.56904554e-01 -4.76289913e-02
6.39163196e-01 -2.56915540e-01 -6.11221433e-01 1.36616564e+00
9.06500101e-01 1.56144679e-01 5.92829168e-01 -1.35737300e+00
-3.77376586e-01 3.97921234e-01 -1.07333529e+00 5.60782552e-01
3.80171418e-01 2.33917713e-01 9.05489445e-01 -1.14496887e+00
4.70480502e-01 1.27815247e+00 7.38894641e-01 7.28874326e-01
-1.11193824e+00 -1.07119966e+00 1.77765921e-01 6.87852204e-01
-1.85371244e+00 -4.74535763e-01 6.56798899e-01 -3.16506147e-01
5.88993788e-01 2.48448730e-01 4.46630865e-01 1.17304182e+00
-9.10883322e-02 8.82473111e-01 8.17509532e-01 -2.17012271e-01
-2.79638559e-01 2.26717845e-01 2.15317756e-01 5.54775834e-01
2.99277425e-01 1.48530424e-01 -7.25729167e-01 -1.36959583e-01
5.06192982e-01 -1.04018729e-02 -3.46400827e-01 -9.66204256e-02
-1.11894023e+00 5.98015785e-01 4.12365079e-01 8.28797594e-02
-1.96760949e-02 -2.47304112e-01 5.44597566e-01 1.04845919e-01
3.24853361e-01 3.19849662e-02 -2.80240215e-02 1.52822837e-01
-9.32972908e-01 4.20965254e-01 1.38129100e-01 1.07434213e+00
6.68698847e-01 -3.99264753e-01 -4.97565359e-01 1.15756929e+00
1.67167578e-02 4.62213248e-01 6.58973336e-01 -2.67162114e-01
7.19378829e-01 5.88366330e-01 8.09111297e-02 -1.14756453e+00
-3.30130219e-01 -6.98380232e-01 -1.05624545e+00 -2.62005985e-01
3.94663751e-01 -1.18735306e-01 -6.23882055e-01 1.79953408e+00
5.54236174e-01 2.34149054e-01 -2.69201726e-01 1.18218279e+00
9.08061743e-01 3.73892784e-01 1.69436082e-01 3.37753072e-02
1.68371069e+00 -9.16748941e-01 -4.00950998e-01 -4.05592203e-01
4.79559481e-01 -5.76665699e-01 6.99451029e-01 -6.84500039e-02
-7.55594611e-01 -6.06538951e-01 -7.93388247e-01 1.61472838e-02
-2.75645912e-01 4.89701509e-01 7.32219145e-02 6.81993961e-01
-7.68382192e-01 1.45242112e-02 -1.75497308e-01 -3.75717491e-01
4.02794778e-01 4.14993107e-01 -6.69635594e-01 -2.41158545e-01
-1.32754397e+00 6.66107893e-01 3.93134117e-01 3.97106141e-01
-4.36003953e-01 -6.41719043e-01 -9.04880106e-01 -1.72904119e-01
3.53819430e-01 -7.01176643e-01 8.70770693e-01 -8.12064588e-01
-7.00158417e-01 1.20139229e+00 -5.00922799e-01 -4.68561947e-01
1.08575153e+00 -6.65411577e-02 -5.35185754e-01 1.94284707e-01
4.99422282e-01 8.72346580e-01 8.33696425e-01 -1.12739837e+00
-8.77062142e-01 -5.05705178e-01 -3.36661160e-01 1.90594196e-01
-4.11182731e-01 2.29189694e-01 -7.92027414e-01 -1.01775551e+00
-1.72912300e-01 -8.38869512e-01 -7.86970556e-02 -3.60898413e-02
-6.25798404e-01 -5.35382211e-01 3.24801624e-01 -7.83173382e-01
1.15151465e+00 -2.13282394e+00 -2.15101153e-01 2.40342706e-01
3.86838287e-01 3.29626769e-01 -3.49114060e-01 2.78873652e-01
-2.62874752e-01 -1.64721563e-01 -1.20382957e-01 -6.53575480e-01
-1.75512448e-01 -3.53484511e-01 -9.54633951e-02 6.73571289e-01
1.83110371e-01 7.82414734e-01 -7.44023621e-01 -7.56806791e-01
1.44471452e-01 3.27346414e-01 -1.73006997e-01 -5.95663954e-03
3.97362411e-01 2.47525200e-01 -3.82136613e-01 6.40010774e-01
8.98274541e-01 -7.30948895e-02 -1.11373514e-01 -3.86200607e-01
-4.46626469e-02 -1.94294453e-01 -1.04776275e+00 1.27547836e+00
8.53290036e-02 3.87656212e-01 -2.05588657e-02 -9.21570420e-01
9.91634667e-01 1.93767773e-03 3.92560273e-01 -8.90190005e-01
2.20196068e-01 4.20443937e-02 -1.94793254e-01 -3.24733466e-01
7.15464830e-01 -2.07535207e-01 -1.05662785e-01 1.88299403e-01
-1.67155877e-01 8.54190350e-01 2.50634193e-01 3.87338072e-01
8.18259478e-01 -2.42695034e-01 2.62560427e-01 -2.32107460e-01
9.01188135e-01 -6.39671311e-02 8.80296171e-01 8.31691623e-01
-7.09686875e-01 8.48356843e-01 1.95006937e-01 -3.83394003e-01
-1.01997626e+00 -8.53951693e-01 -2.41140395e-01 8.24457526e-01
5.96213758e-01 -3.83895367e-01 -6.82360411e-01 -7.71779537e-01
3.41840386e-02 3.51403892e-01 -6.23789251e-01 -1.04762904e-01
-5.99177659e-01 -9.73006010e-01 7.74247408e-01 4.05980855e-01
8.94601882e-01 -7.85949111e-01 -1.14243880e-01 -2.28355937e-02
-5.73141634e-01 -1.12454402e+00 -1.06446719e+00 -4.43975180e-01
-2.06711695e-01 -1.11895561e+00 -1.35621953e+00 -8.97753716e-01
8.83468211e-01 7.31821895e-01 8.14132392e-01 9.95721370e-02
-6.29951537e-01 2.51712441e-01 -3.34033936e-01 -1.91751048e-01
6.98711723e-02 -1.30346483e-02 2.31357932e-01 2.59760112e-01
7.67719686e-01 -1.24634691e-01 -7.95710802e-01 7.44050503e-01
-5.09364963e-01 8.91643688e-02 6.60182714e-01 9.96404290e-01
5.70195973e-01 1.31521866e-01 6.93732858e-01 -4.60777789e-01
3.74323308e-01 -4.08807546e-01 -5.03244877e-01 4.68245298e-01
-4.14022982e-01 -2.73501545e-01 2.64471024e-01 -4.78056192e-01
-9.93764102e-01 -3.73533927e-02 -8.27415437e-02 -2.85540462e-01
-6.98799267e-02 1.20592438e-01 -2.80910790e-01 7.91209787e-02
4.03752655e-01 5.00367582e-01 -8.39723498e-02 -3.17130387e-01
-3.06972209e-02 7.37659335e-01 9.50670362e-01 -4.20233816e-01
9.24060941e-01 4.17208314e-01 -3.11671376e-01 -6.17321789e-01
-5.92728734e-01 -9.84452903e-01 -4.08440113e-01 -3.96827370e-01
6.23362362e-01 -1.23063409e+00 -7.84130156e-01 6.16585195e-01
-1.01135886e+00 2.01930776e-01 1.51681721e-01 4.25696522e-01
5.35039119e-02 6.97180152e-01 -4.64896947e-01 -9.26137805e-01
-6.26352847e-01 -9.58501041e-01 9.75611091e-01 4.56080884e-01
-1.21963806e-01 -4.52335894e-01 -1.11030050e-01 8.22795093e-01
-8.75711441e-03 -1.73285380e-01 4.02037919e-01 -6.51356697e-01
-2.55228132e-01 -7.08199263e-01 -6.80145264e-01 -5.33909583e-03
2.80144322e-03 -5.41243434e-01 -7.88905144e-01 -5.76519728e-01
-4.02501702e-01 -6.29852042e-02 1.02998972e+00 1.71820581e-01
1.09353507e+00 -2.80875325e-01 -5.20353317e-01 4.11484927e-01
1.15091085e+00 -2.60377914e-01 5.11169136e-01 5.72139502e-01
6.44906461e-01 7.22200096e-01 7.03263283e-01 5.88133514e-01
6.18483961e-01 9.16030943e-01 -1.87831316e-02 9.66185033e-02
-1.02965653e-01 -3.82902265e-01 2.35777557e-01 1.56922102e-01
1.60659403e-01 -3.53268057e-01 -9.50670421e-01 7.43560910e-01
-1.81463909e+00 -1.13384426e+00 -1.84707254e-01 2.39825416e+00
7.78930783e-01 1.06990434e-01 4.77204889e-01 3.94518711e-02
1.54090393e+00 2.81440690e-02 -6.01897597e-01 4.54285383e-01
-4.49378520e-01 -4.45226312e-01 6.52710080e-01 2.69266635e-01
-1.20787573e+00 8.45376015e-01 5.03006172e+00 1.25449312e+00
-7.89826870e-01 1.22051239e-01 6.47549629e-01 -2.38498613e-01
-6.12078272e-02 -5.30523419e-01 -1.30412447e+00 8.22901726e-01
3.59272331e-01 -3.29302758e-01 2.02084146e-02 6.01612806e-01
1.57586992e-01 -2.64586419e-01 -8.12863648e-01 1.53991032e+00
3.22703034e-01 -9.37616050e-01 1.20466881e-01 1.71862803e-02
4.56003666e-01 -9.98562798e-02 1.00811623e-01 1.47913963e-01
-5.72796464e-02 -7.12239146e-01 8.75996530e-01 3.40511113e-01
8.56876671e-01 -9.68355656e-01 9.79987741e-01 2.29067698e-01
-1.37411737e+00 -2.25042757e-02 -3.62242430e-01 3.77319485e-01
1.07640021e-01 6.04002535e-01 -6.16483986e-01 6.15180552e-01
7.87602603e-01 6.19179130e-01 -7.20443010e-01 1.33657837e+00
-1.27165660e-01 2.85552025e-01 -1.97264776e-01 1.57735974e-01
-1.27722342e-02 -9.62454304e-02 7.74450898e-01 1.33617890e+00
2.03587636e-01 1.36426594e-02 -5.52502112e-04 7.11955190e-01
-1.17155455e-01 1.65413156e-01 -1.78718433e-01 3.57278287e-01
5.79111814e-01 1.12843513e+00 -5.40292144e-01 -2.91255623e-01
-3.64134401e-01 1.22105324e+00 3.31486076e-01 2.71932483e-01
-8.21684778e-01 -3.56687486e-01 7.07825303e-01 1.47763729e-01
2.61133283e-01 5.30789979e-02 -1.19138263e-01 -1.27383101e+00
4.08270955e-01 -8.09807122e-01 7.70506799e-01 -5.14766335e-01
-1.79359686e+00 5.06008625e-01 -2.36191023e-02 -1.40328455e+00
6.76909685e-02 -2.08356962e-01 -4.68881667e-01 1.04331565e+00
-1.61510718e+00 -1.17625415e+00 -4.64155734e-01 7.30277598e-01
6.29545808e-01 -3.49574149e-01 3.87275636e-01 6.75846338e-01
-1.15009260e+00 1.40470207e+00 -1.03926502e-01 6.99442267e-01
1.07760203e+00 -8.24079990e-01 4.83098119e-01 1.12269258e+00
-2.18204528e-01 6.52281404e-01 7.36053169e-01 -8.35735023e-01
-9.94110644e-01 -1.18326998e+00 1.00386083e+00 -1.80265918e-01
2.91107833e-01 -3.77699047e-01 -7.99268603e-01 3.38027090e-01
-2.58288741e-01 3.79075222e-02 7.33125210e-01 1.77740753e-02
-6.63164973e-01 -2.54966229e-01 -1.22688282e+00 6.73444569e-01
1.16797936e+00 -5.44785500e-01 -3.37005019e-01 1.52918354e-01
1.17338762e-01 -4.65626232e-02 -4.25008386e-01 2.92135954e-01
7.52724469e-01 -9.07398403e-01 1.12141430e+00 -2.73140967e-01
1.80494219e-01 -5.23616552e-01 2.29116455e-01 -9.55077529e-01
-3.00665259e-01 -3.57902974e-01 2.56786138e-01 1.57125664e+00
1.63211718e-01 -7.43848801e-01 8.71236682e-01 7.68786550e-01
2.16001332e-01 -3.20738763e-01 -1.08621085e+00 -9.52940285e-01
-4.56525981e-02 -1.71090111e-01 5.32140970e-01 8.44800651e-01
-6.74000243e-03 2.08727986e-01 -6.61556840e-01 3.13179880e-01
9.87794280e-01 -9.41255838e-02 8.45986724e-01 -9.58146334e-01
-4.45880108e-02 -6.05020046e-01 -6.57874525e-01 -1.06888890e+00
1.67404652e-01 -8.31872642e-01 1.53412625e-01 -1.17132580e+00
8.07983458e-01 -4.83437270e-01 -3.00160348e-01 3.90921831e-01
-6.64001107e-01 5.74932992e-01 1.77056924e-01 4.81467098e-01
-6.94536269e-01 6.93210721e-01 7.58180022e-01 -4.17041123e-01
9.19303522e-02 3.23368609e-02 -6.00755990e-01 3.65863472e-01
9.12875354e-01 -4.79987085e-01 -7.99418613e-02 -3.76620620e-01
-2.43156962e-02 -2.04778150e-01 6.86689556e-01 -9.05360937e-01
6.03798807e-01 1.38419911e-01 7.10633397e-01 -8.83187294e-01
2.61965513e-01 -4.68450189e-01 8.64907503e-02 3.90552014e-01
-2.89966315e-01 1.86498895e-01 6.43390194e-02 7.47124970e-01
-2.35298052e-01 -1.44277811e-01 8.41522515e-01 -3.12902126e-03
-8.73139679e-01 5.14036417e-01 -1.05982929e-01 -3.83322202e-02
1.04213047e+00 -4.29402530e-01 -4.24248427e-01 -1.78122565e-01
-2.66871274e-01 6.14417791e-01 4.24938679e-01 5.01538455e-01
7.56804883e-01 -1.30595648e+00 -1.02888620e+00 2.87479997e-01
5.17131388e-01 -3.13888192e-01 7.70744681e-01 7.91592717e-01
-1.02180406e-01 3.15614820e-01 -1.07802875e-01 -3.21884662e-01
-1.80101466e+00 5.16773164e-01 4.07128274e-01 -6.48281574e-02
-6.76155090e-01 1.16272116e+00 4.17626888e-01 -3.09594423e-01
2.47591034e-01 6.08706653e-01 -1.51007831e-01 1.73114553e-01
1.04473376e+00 4.45585489e-01 -7.23360777e-02 -1.08534086e+00
-5.74843407e-01 5.64626038e-01 -4.72799838e-01 -4.43313420e-02
9.23410237e-01 -3.60901982e-01 1.30264740e-02 -1.74179569e-01
1.04345131e+00 -8.18391815e-02 -1.07623994e+00 -5.76264143e-01
-1.17585234e-01 -7.35988200e-01 -1.62687391e-01 -6.15138888e-01
-9.83938932e-01 5.07339180e-01 8.07593286e-01 -4.39533859e-01
9.84830022e-01 -3.05853691e-03 9.23195183e-01 2.06787586e-01
4.28101480e-01 -1.11117995e+00 3.17754410e-02 1.18134908e-01
7.46479869e-01 -1.50399661e+00 1.10378295e-01 -4.86879826e-01
-5.47231555e-01 7.44236767e-01 6.96131408e-01 2.14248493e-01
3.12328458e-01 -1.71863988e-01 -9.44312885e-02 1.42329857e-01
-1.77099138e-01 -3.49945247e-01 3.98840189e-01 5.64741135e-01
6.73432201e-02 7.36677498e-02 -5.27562499e-01 7.31559515e-01
-3.05109113e-01 -2.47224778e-01 2.29118571e-01 6.53251112e-01
-1.89305380e-01 -1.11186242e+00 -5.74045062e-01 4.71299678e-01
-3.63892764e-01 -1.76280797e-01 -3.53154421e-01 5.60302496e-01
1.49198323e-01 1.03346336e+00 1.70126352e-02 -5.11433363e-01
1.24943666e-01 -1.79123908e-01 3.26985896e-01 -2.32321173e-01
-6.54262424e-01 -9.81263667e-02 -6.77864552e-02 -2.04517215e-01
-2.84333497e-01 -8.64951551e-01 -9.83221889e-01 -6.09816790e-01
-3.40512127e-01 1.41460389e-01 3.33184898e-01 8.20689380e-01
3.20571423e-01 -7.78087601e-02 6.67194307e-01 -4.91725445e-01
-4.12499279e-01 -8.86633217e-01 -5.57825446e-01 4.74452615e-01
3.18557590e-01 -6.66115761e-01 -1.52499244e-01 -1.17274947e-01] | [14.771620750427246, 0.8925073742866516] |
4bd7c213-8938-4ef2-bb5e-a9a93994c753 | an-investigation-of-noise-in-morphological | 2305.16581 | null | https://arxiv.org/abs/2305.16581v1 | https://arxiv.org/pdf/2305.16581v1.pdf | An Investigation of Noise in Morphological Inflection | With a growing focus on morphological inflection systems for languages where high-quality data is scarce, training data noise is a serious but so far largely ignored concern. We aim at closing this gap by investigating the types of noise encountered within a pipeline for truly unsupervised morphological paradigm completion and its impact on morphological inflection systems: First, we propose an error taxonomy and annotation pipeline for inflection training data. Then, we compare the effect of different types of noise on multiple state-of-the-art inflection models. Finally, we propose a novel character-level masked language modeling (CMLM) pretraining objective and explore its impact on the models' resistance to noise. Our experiments show that various architectures are impacted differently by separate types of noise, but encoder-decoders tend to be more robust to noise than models trained with a copy bias. CMLM pretraining helps transformers, but has lower impact on LSTMs. | ['Katharina Kann', 'Miikka Silfverberg', 'Garrett Nicolai', 'Changbing Yang', 'Adam Wiemerslage'] | 2023-05-26 | null | null | null | null | ['morphological-inflection'] | ['natural-language-processing'] | [ 2.37484857e-01 -8.98021385e-02 7.56266490e-02 -3.66129130e-01
-8.76870036e-01 -7.62667596e-01 5.03791749e-01 6.66943610e-01
-8.67760718e-01 2.55996615e-01 5.86792529e-01 -5.53350449e-01
3.11136454e-01 -7.19734728e-01 -8.43225956e-01 -3.25034529e-01
8.79488885e-02 3.89261663e-01 2.10244000e-01 -1.59859046e-01
3.00096571e-01 2.94345707e-01 -1.31093180e+00 6.10535264e-01
8.94698679e-01 4.23733801e-01 2.23565400e-01 3.81801426e-01
-4.29023206e-01 6.06425881e-01 -1.06883836e+00 -5.96816838e-01
1.55400887e-01 -1.89886183e-01 -9.50381041e-01 -2.40885429e-02
7.90844262e-01 6.22561947e-02 8.94394517e-02 1.27723587e+00
5.90122700e-01 1.29595459e-01 5.74877024e-01 -4.12076980e-01
-1.05081391e+00 1.48012376e+00 -3.94682705e-01 4.23122883e-01
-4.94900607e-02 3.18372995e-01 6.56232238e-01 -1.14023900e+00
6.59411907e-01 1.38457227e+00 1.02977812e+00 5.02114117e-01
-1.79074502e+00 -5.65802813e-01 2.82087922e-01 -7.60318041e-02
-1.32117915e+00 -6.54356539e-01 5.38961649e-01 -4.18952316e-01
1.42804122e+00 4.49425094e-02 3.92139167e-01 1.07828557e+00
1.92772731e-01 7.62986302e-01 1.24436831e+00 -6.45941198e-01
1.45587742e-01 1.29024103e-01 5.00197649e-01 4.01236147e-01
4.06807840e-01 2.26389095e-02 -4.62561429e-01 1.06644832e-01
5.55073678e-01 -6.81271374e-01 -1.63857818e-01 2.90375501e-01
-9.67532635e-01 6.99845195e-01 2.95263797e-01 7.56783962e-01
-2.15469360e-01 -5.66507317e-02 5.17638087e-01 4.38148797e-01
6.36146843e-01 7.10895419e-01 -6.32859468e-01 5.36651202e-02
-1.31926215e+00 -4.10146676e-02 4.76674199e-01 7.91851342e-01
6.54995143e-01 6.63767993e-01 -1.34251356e-01 1.16893780e+00
2.54383802e-01 1.88903362e-01 6.13384068e-01 -5.08102715e-01
5.08006454e-01 5.60494483e-01 -3.46066058e-01 -5.66034317e-01
-5.67546844e-01 -7.39895105e-01 -5.98008335e-01 2.73774415e-01
7.18491495e-01 -2.59174109e-02 -1.15270090e+00 2.07293868e+00
-1.19959399e-01 -4.37533945e-01 -6.55691698e-02 6.51131570e-01
5.42538583e-01 4.83191669e-01 6.15936935e-01 -1.23974077e-01
1.29647064e+00 -6.26678765e-01 -6.83674276e-01 -6.49959445e-01
7.89011121e-01 -1.02422071e+00 1.51044452e+00 5.31271636e-01
-1.60051203e+00 -7.00075090e-01 -1.28704119e+00 -3.99082452e-01
-5.99448085e-01 6.59511313e-02 4.29536909e-01 1.00358307e+00
-1.11608255e+00 6.41619205e-01 -8.37665498e-01 -3.56570154e-01
3.34163398e-01 3.17960799e-01 -1.19143315e-01 1.76465943e-01
-1.05335391e+00 1.37667811e+00 5.91396093e-01 4.38088924e-03
-6.57251298e-01 -9.97170210e-01 -8.77206624e-01 -1.08001508e-01
-7.46966079e-02 -4.68535960e-01 1.28158522e+00 -1.06273139e+00
-1.13190353e+00 1.10303581e+00 -2.43182957e-01 -6.34208083e-01
1.76161215e-01 -2.31781453e-01 -5.97378314e-01 -3.87054652e-01
-1.09689072e-01 8.12010229e-01 5.41536808e-01 -1.49796712e+00
-3.70673686e-01 -4.31796879e-01 -3.45491856e-01 -4.86284606e-02
-4.43014592e-01 4.79670644e-01 1.14700392e-01 -1.06865251e+00
1.42769367e-01 -5.34651756e-01 -1.99667394e-01 -5.50217628e-01
-3.25174510e-01 -2.30990440e-01 3.07468712e-01 -8.49153578e-01
1.51479781e+00 -2.05958867e+00 -1.78218875e-02 -1.10844430e-02
-1.74009025e-01 5.29862106e-01 -2.68729031e-01 2.47906193e-01
-1.42359525e-01 5.87168634e-01 -7.34820366e-01 -7.05078542e-01
1.54790908e-01 2.48147443e-01 -3.20645154e-01 1.44283593e-01
5.04365802e-01 8.83377492e-01 -7.21758008e-01 -2.79496253e-01
-4.26697545e-02 3.83596569e-01 -5.58915138e-01 -6.91004023e-02
-3.23416352e-01 3.08090001e-01 7.12886214e-01 5.16047537e-01
9.05315876e-01 5.23952842e-01 4.13602851e-02 -1.45742111e-02
-6.28018379e-01 9.19103801e-01 -1.05750704e+00 1.77114034e+00
-5.57352722e-01 6.00230753e-01 3.78682055e-02 -3.79763156e-01
9.88126040e-01 2.19220191e-01 -3.57214689e-01 -6.06225252e-01
2.08641678e-01 5.61984479e-01 5.67484915e-01 -8.54832530e-02
8.89640629e-01 -5.27212322e-01 -8.72954279e-02 4.45210248e-01
3.34980607e-01 -2.05247849e-01 3.52814049e-01 -1.26354158e-01
9.86265302e-01 -1.11181699e-01 5.38652241e-02 -7.48057485e-01
1.94232315e-01 1.96649566e-01 5.90817690e-01 7.39046633e-01
-8.00283402e-02 6.70191526e-01 1.30839035e-01 -1.86433762e-01
-1.05992401e+00 -1.29123175e+00 -3.83084625e-01 1.13358378e+00
-5.07289708e-01 -3.94974649e-01 -1.04498148e+00 -3.45523685e-01
-2.79397190e-01 1.33153129e+00 -4.45760787e-01 -2.70521849e-01
-9.34114337e-01 -1.25339687e+00 1.07228696e+00 6.14955127e-01
2.74266452e-01 -1.37304294e+00 -3.83683056e-01 3.87158483e-01
-2.23162606e-01 -8.22524846e-01 -6.02566659e-01 5.85692465e-01
-1.13160121e+00 -4.76233006e-01 -3.24218988e-01 -1.06180251e+00
4.17287529e-01 -4.59436364e-02 1.61506367e+00 1.00642323e-01
1.11318249e-02 -7.99949095e-02 -1.06754854e-01 -7.04565048e-01
-9.21704531e-01 4.10432458e-01 1.47237733e-01 -2.77399838e-01
8.06164265e-01 -5.99913716e-01 -2.57490873e-01 -4.23517525e-02
-1.24017251e+00 -4.06280994e-01 5.89500487e-01 4.21682954e-01
5.11701465e-01 5.33640049e-02 6.17549360e-01 -1.07433701e+00
9.46353018e-01 -3.20376217e-01 -3.55085820e-01 -8.38609636e-02
-5.50304174e-01 1.56284422e-01 6.50488853e-01 -7.37823904e-01
-1.16986048e+00 -3.47817913e-02 -5.78912973e-01 3.49045172e-03
-3.22711796e-01 5.06932497e-01 -4.24400449e-01 3.33647311e-01
1.01537383e+00 4.15453278e-02 -4.79971319e-01 -8.22967112e-01
4.42723811e-01 4.35544848e-01 7.14800179e-01 -5.52853167e-01
5.82728684e-01 2.06231937e-01 -5.78116298e-01 -1.08060479e+00
-7.38731325e-01 -1.71610340e-01 -1.00257289e+00 1.43956676e-01
8.08183908e-01 -6.12318695e-01 3.92215624e-02 7.15905488e-01
-1.41014683e+00 -4.78477597e-01 -5.14717937e-01 2.40955368e-01
-2.20332280e-01 3.38442147e-01 -9.83751655e-01 -7.56236672e-01
-5.62581778e-01 -1.17054856e+00 4.65831041e-01 -7.51683414e-02
-5.84745109e-01 -1.11402726e+00 1.04403310e-01 5.11425696e-02
6.09460592e-01 -2.72839189e-01 1.54799259e+00 -6.78685784e-01
-2.16472283e-01 2.85070032e-01 -7.97427818e-02 4.48032677e-01
-4.23950702e-02 2.43809205e-02 -1.17497683e+00 -1.39995858e-01
3.31037134e-01 -1.59744501e-01 1.22855353e+00 4.28512901e-01
6.73409820e-01 -4.43988323e-01 2.17670396e-01 5.75304747e-01
1.39983869e+00 -1.34045020e-01 7.76539028e-01 4.22269762e-01
5.14154136e-01 7.53457308e-01 1.60895243e-01 -1.52001828e-01
2.52988875e-01 3.32267404e-01 9.20633972e-02 -1.15012713e-01
-6.29906118e-01 -3.89593005e-01 7.19190121e-01 1.40705979e+00
3.14856023e-01 -1.93674505e-01 -9.63561773e-01 1.04863882e+00
-1.33332026e+00 -7.22256184e-01 -7.01138735e-01 2.50096035e+00
1.44741011e+00 5.45124590e-01 1.38688847e-01 4.20763314e-01
6.19997323e-01 4.11997773e-02 -1.32168815e-01 -1.01743495e+00
-5.76404810e-01 4.62295622e-01 5.89558303e-01 7.57272840e-01
-9.99268830e-01 1.45910943e+00 7.02288532e+00 8.39572012e-01
-9.32170331e-01 2.55955100e-01 3.34865361e-01 8.78031328e-02
-6.12325966e-01 2.07320657e-02 -8.43138754e-01 3.81263316e-01
1.34382212e+00 2.43813023e-01 2.15024978e-01 4.41571742e-01
3.06429684e-01 -1.96395338e-01 -1.05256748e+00 5.00380814e-01
5.52868508e-02 -9.33550656e-01 2.95777172e-01 -1.60700440e-01
5.27159095e-01 2.64708072e-01 1.59215137e-01 3.80352557e-01
6.05495334e-01 -1.01287746e+00 1.13413727e+00 2.75837541e-01
5.34352005e-01 -8.32319438e-01 4.46024358e-01 3.18616956e-01
-9.41062212e-01 1.27963290e-01 -6.73013270e-01 -1.40991524e-01
2.62488633e-01 8.55614781e-01 -7.81564653e-01 -4.17689234e-02
3.55812609e-01 1.28211519e-02 -9.90541697e-01 1.19451344e+00
-5.42720258e-01 1.19078565e+00 -3.12232912e-01 1.94529593e-01
1.28216863e-01 7.23641068e-02 4.84137088e-01 2.04421711e+00
5.12462817e-02 -2.45550632e-01 3.52064446e-02 1.00751841e+00
-1.90535948e-01 3.97503942e-01 -3.86167079e-01 1.17279962e-01
5.17645717e-01 9.85942960e-01 -8.10358882e-01 -3.65937829e-01
-3.21542948e-01 9.79968250e-01 7.46567488e-01 2.11472839e-01
-5.33377409e-01 -1.28090248e-01 7.33125627e-01 2.03082472e-01
-2.88424063e-02 -5.60312748e-01 -1.01872063e+00 -9.38798249e-01
-1.98669553e-01 -1.08976865e+00 4.27943915e-01 -4.64936942e-01
-1.32725716e+00 7.06703961e-01 -2.65617996e-01 -6.53680503e-01
-7.91819999e-04 -6.81356490e-01 -5.90172529e-01 9.99656975e-01
-1.40729117e+00 -1.18515909e+00 3.88474494e-01 1.60967276e-01
8.53721082e-01 1.01328872e-01 9.63616610e-01 4.55991685e-01
-4.98558700e-01 8.93936992e-01 -1.67817995e-01 3.56240422e-01
7.92009711e-01 -1.56460381e+00 1.05230045e+00 1.31884861e+00
6.25339031e-01 1.02849102e+00 6.98477089e-01 -8.47277641e-01
-9.68808889e-01 -1.28369117e+00 1.37813234e+00 -8.26503575e-01
6.08483553e-01 -6.90250635e-01 -1.42079532e+00 8.18783224e-01
3.72457504e-01 -4.96066332e-01 7.26688623e-01 3.87308300e-01
-7.77890444e-01 4.47774708e-01 -1.01593983e+00 8.74577165e-01
1.12136102e+00 -6.14294350e-01 -1.02662027e+00 -1.56338915e-01
8.43398869e-01 -6.31697923e-02 -6.42119706e-01 2.65266776e-01
1.93070933e-01 -9.17029262e-01 7.23244429e-01 -6.71421289e-01
2.79379725e-01 -2.28285953e-01 -1.63169667e-01 -1.67499256e+00
-6.47329748e-01 -3.04340869e-01 1.51712909e-01 1.58791482e+00
9.16099250e-01 -3.48039895e-01 4.15886939e-01 3.14732879e-01
-6.10173702e-01 -2.58844674e-01 -8.71453583e-01 -8.67610812e-01
9.55040097e-01 -8.78660858e-01 3.47567022e-01 9.61121321e-01
1.33830663e-02 5.30514777e-01 3.99811156e-02 1.78279698e-01
4.57259655e-01 -2.76146352e-01 2.18712285e-01 -9.76525962e-01
-2.53836691e-01 -8.44350398e-01 -9.51119512e-02 -8.42136085e-01
4.45928201e-02 -1.34473848e+00 2.00932086e-01 -1.43824041e+00
-3.46797258e-02 -2.51998991e-01 -1.99702099e-01 5.49819648e-01
-3.12768519e-01 4.93916482e-01 2.30105236e-01 8.88294205e-02
-1.28032580e-01 2.14186683e-01 5.58679402e-01 -1.60195380e-01
-3.75564724e-01 -2.08155081e-01 -6.78211987e-01 9.74033773e-01
9.71484721e-01 -6.09420776e-01 -1.22719496e-01 -1.30546057e+00
1.89440519e-01 -6.21794641e-01 -1.10502824e-01 -1.15213490e+00
2.20370367e-01 2.03032047e-01 4.18353885e-01 -7.00830042e-01
7.84687251e-02 -3.85484248e-01 -2.15511635e-01 4.10726279e-01
-4.09747392e-01 4.56251830e-01 8.27684224e-01 -1.08359277e-01
-1.78022590e-02 -6.42792881e-01 1.07263577e+00 -3.27153355e-01
-4.42695081e-01 -1.46977469e-01 -9.57250237e-01 4.76443321e-01
1.80609107e-01 -2.41631821e-01 -2.49389112e-01 1.88953385e-01
-6.98568642e-01 -2.02485442e-01 5.93705177e-01 6.34571731e-01
3.51793021e-01 -1.05455899e+00 -1.02649832e+00 3.74090791e-01
-6.53277263e-02 -3.01651776e-01 -1.20658912e-01 5.55366576e-01
-1.37197062e-01 1.89706609e-01 5.91045842e-02 -2.20574707e-01
-1.08508754e+00 7.29593754e-01 3.12287986e-01 -2.13867083e-01
-3.52344483e-01 1.18151021e+00 4.38071176e-04 -9.11176562e-01
3.01541060e-01 -6.62961364e-01 3.36337164e-02 3.12327594e-01
4.96723980e-01 3.71903360e-01 6.35172963e-01 -7.55445540e-01
-2.67295212e-01 3.31966907e-01 -2.75736690e-01 -3.99256349e-01
1.03427434e+00 -2.99794339e-02 -2.40218073e-01 8.49349320e-01
8.41867924e-01 3.54955226e-01 -7.87283182e-01 -2.93639809e-01
7.88649261e-01 -1.39966652e-01 -3.51854451e-02 -1.11880064e+00
-6.59996688e-01 1.15828264e+00 3.88175726e-01 3.52696106e-02
1.02026224e+00 -2.30756626e-01 8.78888309e-01 -1.06177377e-02
1.38121083e-01 -1.32743335e+00 -2.65410691e-01 9.81037796e-01
8.41668010e-01 -9.01012301e-01 -3.06830972e-01 -3.53121221e-01
-4.13907677e-01 8.72997105e-01 4.85308021e-01 -4.76693781e-03
4.76525158e-01 8.04720819e-01 4.72750366e-01 8.54502767e-02
-6.39905870e-01 -4.61093098e-01 9.73298177e-02 6.73480272e-01
9.49400544e-01 -3.56216319e-02 -4.87095296e-01 8.62192094e-01
-7.07930565e-01 -3.53100687e-01 5.73980987e-01 6.73734486e-01
-6.55728400e-01 -1.27066982e+00 -4.85762239e-01 2.87701070e-01
-8.19017053e-01 -7.32244313e-01 -6.49648070e-01 5.76723218e-01
4.92375672e-01 1.08978701e+00 2.32755005e-01 -2.05635846e-01
5.23508787e-01 6.25210822e-01 5.20027339e-01 -9.72883105e-01
-1.27856207e+00 -1.63157769e-02 2.04305917e-01 1.27546400e-01
1.31408691e-01 -8.23910177e-01 -1.46573544e+00 -1.72095358e-01
-3.51636261e-01 -3.07984687e-02 6.95678055e-01 7.28297830e-01
1.84861675e-01 5.05492210e-01 -6.66923523e-02 -6.15753949e-01
-6.63069904e-01 -1.34734929e+00 -3.06786895e-01 4.67713386e-01
5.48279956e-02 -1.70101508e-01 -4.21194226e-01 1.45957306e-01] | [10.698877334594727, 9.755813598632812] |
4e90913e-7638-446e-af70-6e2a9ff80bdb | dataset-creation-pipeline-for-camera-based | 2303.01468 | null | https://arxiv.org/abs/2303.01468v1 | https://arxiv.org/pdf/2303.01468v1.pdf | Dataset Creation Pipeline for Camera-Based Heart Rate Estimation | Heart rate is one of the most vital health metrics which can be utilized to investigate and gain intuitions into various human physiological and psychological information. Estimating heart rate without the constraints of contact-based sensors thus presents itself as a very attractive field of research as it enables well-being monitoring in a wider variety of scenarios. Consequently, various techniques for camera-based heart rate estimation have been developed ranging from classical image processing to convoluted deep learning models and architectures. At the heart of such research efforts lies health and visual data acquisition, cleaning, transformation, and annotation. In this paper, we discuss how to prepare data for the task of developing or testing an algorithm or machine learning model for heart rate estimation from images of facial regions. The data prepared is to include camera frames as well as sensor readings from an electrocardiograph sensor. The proposed pipeline is divided into four main steps, namely removal of faulty data, frame and electrocardiograph timestamp de-jittering, signal denoising and filtering, and frame annotation creation. Our main contributions are a novel technique of eliminating jitter from health sensor and camera timestamps and a method to accurately time align both visual frame and electrocardiogram sensor data which is also applicable to other sensor types. | ['Peter Corcoran', 'Joseph Lemley', 'Amr Elrasad', 'Mohamed Moustafa'] | 2023-03-02 | null | null | null | null | ['heart-rate-estimation'] | ['medical'] | [ 5.73602259e-01 -2.23297521e-01 1.28384857e-02 -6.80974305e-01
-3.16832870e-01 -1.14248283e-01 4.98828590e-02 4.56386060e-01
-4.78744298e-01 4.54111487e-01 -2.24630117e-01 1.20124906e-01
1.05122156e-01 -2.87368834e-01 -1.27103865e-01 -6.95455194e-01
8.12022686e-02 -2.01521274e-02 -2.99955755e-01 3.69547546e-01
2.97229856e-01 6.54831052e-01 -1.39490759e+00 -1.66017517e-01
2.47738853e-01 1.26747644e+00 -4.83626693e-01 1.13200891e+00
2.36376941e-01 5.46832442e-01 -9.61280644e-01 -1.18402854e-01
-9.40154865e-03 -7.54271269e-01 -2.55878270e-01 1.37745962e-01
3.55308443e-01 -5.86485863e-01 1.11470491e-01 5.99610031e-01
1.09373212e+00 -3.33963275e-01 3.37124169e-02 -1.44468939e+00
-5.26391268e-02 1.55561715e-01 -5.50810039e-01 5.06588042e-01
5.50607920e-01 2.78222561e-01 7.33814144e-04 -4.30333078e-01
3.11825693e-01 8.38736415e-01 9.58277047e-01 5.44365525e-01
-1.06431091e+00 -6.09356761e-01 -6.96115792e-01 4.94399518e-01
-1.30244195e+00 -7.74976194e-01 1.03271556e+00 -4.27985877e-01
7.83043861e-01 4.02019858e-01 1.03948295e+00 1.16875696e+00
3.24096859e-01 6.85663894e-02 1.12541187e+00 -5.33521891e-01
3.73307586e-01 1.44963831e-01 3.58586073e-01 5.04365802e-01
3.10407609e-01 5.06566279e-02 -8.11797082e-01 -1.03175178e-01
9.04544115e-01 1.33648766e-02 -1.93316117e-01 1.87263519e-01
-1.26459348e+00 3.21446568e-01 -3.19881350e-01 9.28037912e-02
-6.82941973e-01 4.50671822e-01 7.38654733e-01 2.15322524e-01
4.36979651e-01 2.25925893e-02 -3.25837463e-01 -5.44374287e-01
-1.14412332e+00 9.32545029e-03 6.83658481e-01 5.44031441e-01
2.15984479e-01 2.13875145e-01 -2.56239444e-01 3.65885049e-01
2.78935820e-01 6.33424342e-01 4.79131699e-01 -1.26301670e+00
-1.88513637e-01 2.43473306e-01 3.31836380e-02 -1.05320561e+00
-6.61540747e-01 1.67810485e-01 -1.04177535e+00 2.24926457e-01
3.82299036e-01 -4.23180699e-01 -5.96098661e-01 1.42169285e+00
6.91286147e-01 6.32909060e-01 -4.08211946e-01 8.60431850e-01
8.61296535e-01 2.85901457e-01 4.66115251e-02 -8.41775954e-01
1.71197915e+00 3.56707559e-03 -1.20688343e+00 5.08935675e-02
9.00700875e-03 -6.99858487e-01 6.18568897e-01 4.92615610e-01
-1.16519535e+00 -9.41115856e-01 -1.05961442e+00 -1.33917257e-01
-1.02063797e-01 2.57517733e-02 4.57757056e-01 8.93412948e-01
-1.05241334e+00 6.85114622e-01 -1.12200785e+00 -5.38525879e-01
3.99162203e-01 1.80494443e-01 -1.25905469e-01 3.19383293e-01
-1.03814328e+00 7.64263332e-01 5.30858710e-02 4.07971293e-01
-5.92505693e-01 -7.31791794e-01 -8.66684794e-01 -1.94199219e-01
-5.18347956e-02 -7.55025327e-01 1.24978900e+00 -8.23495328e-01
-1.65200412e+00 1.08141315e+00 -4.27515477e-01 -6.01598859e-01
5.90530574e-01 -2.25415200e-01 -4.74356949e-01 3.21608096e-01
-3.35390359e-01 4.50750709e-01 1.46606791e+00 -4.22860146e-01
-3.12487006e-01 -5.30043185e-01 -7.13997185e-01 6.86097294e-02
-2.04707995e-01 2.68193752e-01 -3.88240397e-01 -4.36050355e-01
-2.46765930e-02 -7.01561391e-01 2.73388326e-01 1.91450939e-01
-1.28080413e-01 1.61404703e-02 6.51181579e-01 -9.27528441e-01
1.40522289e+00 -1.96486235e+00 -3.06537330e-01 1.45470485e-01
5.47705829e-01 3.59562397e-01 3.33245397e-01 1.44070879e-01
-3.68706286e-01 -1.26224831e-01 2.05698043e-01 -3.90103787e-01
-2.81557411e-01 2.95170732e-02 1.12311810e-01 8.80682290e-01
9.34052169e-02 7.62293696e-01 -4.52279538e-01 -5.40653825e-01
8.29119802e-01 9.18393612e-01 4.98315878e-02 2.37555340e-01
3.16338450e-01 8.72439861e-01 1.29432693e-01 7.45001674e-01
5.26562512e-01 7.75448233e-02 -3.66345211e-03 -5.83449960e-01
-7.73042291e-02 1.11534558e-02 -1.20978415e+00 1.40238595e+00
-3.17408778e-02 9.48969126e-01 -1.20283887e-01 -9.53952491e-01
1.04895711e+00 6.78912401e-01 1.00283051e+00 -8.00493777e-01
5.44235945e-01 -3.22683066e-01 -2.55471677e-01 -1.07663214e+00
1.03525117e-01 -7.85758048e-02 1.77248046e-01 4.24779058e-01
-2.19350934e-01 2.51259711e-02 -1.58297364e-02 -2.86144763e-01
1.12915087e+00 3.15666460e-02 6.76066339e-01 2.19710991e-01
4.81000811e-01 -4.62124258e-01 6.33580685e-01 3.59798729e-01
-9.00487423e-01 5.13472497e-01 4.36870992e-01 -8.72648120e-01
-1.06796777e+00 -6.94752395e-01 -9.69258398e-02 5.61269999e-01
-1.62827298e-01 -3.47609222e-01 -9.92270172e-01 9.52993259e-02
-1.11615017e-01 5.34162484e-02 -6.33449614e-01 -2.74582535e-01
-6.35827005e-01 -7.04802573e-01 8.54770362e-01 5.34316301e-01
3.72100741e-01 -1.15554667e+00 -1.66665030e+00 2.43855521e-01
-2.89217412e-01 -1.15592742e+00 -1.60443157e-01 4.52062115e-02
-1.02891374e+00 -1.28387427e+00 -4.36039865e-01 -1.20450817e-01
3.58136654e-01 -3.98692070e-03 1.16992891e+00 1.44390892e-02
-9.84140635e-01 7.07988799e-01 -1.35355338e-01 -1.01488817e+00
-2.53213227e-01 -3.75386357e-01 1.86741110e-02 1.21657163e-01
7.38832831e-01 -4.59626377e-01 -8.83234620e-01 -7.36173615e-02
-6.02016270e-01 -6.72239438e-02 1.35222808e-01 2.51084179e-01
6.82026923e-01 -1.75901905e-01 3.88301224e-01 -6.07494652e-01
7.09730744e-01 -7.70368353e-02 -6.46728754e-01 -1.45719843e-02
-7.19740510e-01 -3.50350767e-01 2.31234461e-01 -3.61482590e-01
-6.11342072e-01 2.51275629e-01 -4.36554439e-02 -4.21587467e-01
-5.98776996e-01 9.57848057e-02 3.12029775e-02 1.01456739e-01
7.27817714e-01 -1.39771372e-01 4.71717238e-01 -1.96075916e-01
3.05973645e-02 6.33156538e-01 8.66032064e-01 -3.45119201e-02
2.50663131e-01 5.84844589e-01 1.91634610e-01 -1.25391507e+00
-4.15437162e-01 -6.44314468e-01 -8.00194442e-01 -7.99407244e-01
9.88689542e-01 -7.58368075e-01 -1.09978199e+00 1.02361095e+00
-1.12063801e+00 -3.68783586e-02 -2.66330808e-01 4.64000165e-01
-4.13537443e-01 4.19145375e-01 -6.13938570e-01 -1.02614939e+00
-8.32118154e-01 -5.71271837e-01 9.57866073e-01 5.60024440e-01
-6.44138157e-01 -8.55419278e-01 2.12820888e-01 3.70331824e-01
4.85248119e-01 9.37756300e-01 2.57741779e-01 -6.09063618e-02
-6.94428086e-02 -4.43900943e-01 -1.11678406e-03 3.95211488e-01
2.25802585e-01 3.11831862e-01 -1.32002819e+00 -1.88534588e-01
6.22659326e-01 -1.45662725e-01 3.02567005e-01 8.39133918e-01
1.02856290e+00 2.42464021e-02 -4.76997048e-02 6.88265324e-01
1.26423705e+00 3.13036174e-01 8.86543453e-01 -7.84179941e-03
5.86688519e-01 2.81085908e-01 3.76596510e-01 9.06310797e-01
2.18880385e-01 4.66432542e-01 4.26931560e-01 -4.59735006e-01
1.09970547e-01 5.13702214e-01 4.41792190e-01 4.72668231e-01
-4.14403051e-01 1.76692918e-01 -7.12854445e-01 1.81469709e-01
-1.47334337e+00 -1.22295344e+00 -4.19844717e-01 2.46728563e+00
8.46388817e-01 -1.39353186e-01 4.62109506e-01 7.97414422e-01
9.19244766e-01 -5.34600243e-02 -7.71761894e-01 -6.70046508e-01
3.39576691e-01 5.79324722e-01 5.03748536e-01 1.72701225e-01
-1.17368317e+00 2.07213730e-01 6.22870064e+00 -2.34489784e-01
-1.38977492e+00 -5.07701039e-02 7.75192082e-01 -2.11939275e-01
6.39923215e-01 -3.53878051e-01 -6.60311878e-01 5.58869064e-01
1.57527924e+00 3.37442644e-02 3.55310142e-01 5.49535990e-01
9.53541815e-01 -5.66985726e-01 -1.26029360e+00 1.63905168e+00
3.58701408e-01 -9.12234604e-01 -6.31426811e-01 -2.09363222e-01
2.30631139e-02 -4.43775445e-01 -2.33033642e-01 -2.62950152e-01
-6.70777977e-01 -9.53318894e-01 5.02334058e-01 9.12927806e-01
9.57222223e-01 -5.31060636e-01 7.09522247e-01 -8.40021446e-02
-1.02402365e+00 1.81912735e-01 -6.84924647e-02 -3.56202185e-01
9.44430679e-02 1.02282321e+00 -7.52745211e-01 1.81491107e-01
7.84416676e-01 7.83649147e-01 -6.35168910e-01 1.13087261e+00
1.11306131e-01 5.79129875e-01 -3.52395266e-01 2.82450587e-01
-6.75400317e-01 1.09365009e-01 1.77724108e-01 1.11900425e+00
3.45490754e-01 1.51635870e-01 -1.56789467e-01 5.47751486e-01
2.00817630e-01 -1.98077038e-01 -4.47616011e-01 1.59456402e-01
5.26488602e-01 1.45070410e+00 -8.29891384e-01 -4.50396419e-01
-3.83206993e-01 8.84809077e-01 -2.72659153e-01 6.79361895e-02
-1.31852663e+00 -5.45535862e-01 6.55753314e-01 2.98923016e-01
-3.97322655e-01 -1.49072036e-01 -6.46158993e-01 -7.14593709e-01
1.46451250e-01 -7.73644626e-01 4.20613229e-01 -7.18734682e-01
-8.26660872e-01 2.12479115e-01 6.52402490e-02 -1.02749646e+00
-2.93406397e-01 -1.53124928e-01 -6.61604345e-01 8.00657570e-01
-1.50478053e+00 -6.69453204e-01 -1.29246008e+00 8.12225342e-01
4.03468400e-01 4.12479758e-01 8.09986532e-01 6.07456028e-01
-8.06863904e-01 4.34835017e-01 -4.18989003e-01 2.25928098e-01
1.09575284e+00 -1.02566779e+00 2.20797434e-01 8.56616437e-01
-2.00968429e-01 2.90869832e-01 6.99572146e-01 -4.65162218e-01
-1.47518539e+00 -1.05835950e+00 8.22206438e-01 -4.97059137e-01
8.05185437e-02 -1.17985807e-01 -7.80793250e-01 2.98031300e-01
2.22452343e-01 2.91625470e-01 7.50213683e-01 -4.58079159e-01
2.23251551e-01 -7.22983360e-01 -1.10499299e+00 5.33658685e-03
2.42109463e-01 -4.30958927e-01 -3.00910860e-01 -5.53787872e-02
-5.53053319e-02 -6.36643648e-01 -1.07087588e+00 2.48409852e-01
8.33494246e-01 -1.29174006e+00 8.87331843e-01 5.88078201e-02
-1.52168446e-03 -2.41843000e-01 5.82749903e-01 -6.98493421e-01
3.03121582e-02 -1.03670144e+00 -3.08458954e-01 1.41554391e+00
-3.66799325e-01 -3.89583737e-01 6.73305929e-01 9.92303312e-01
2.12989524e-01 -1.78936362e-01 -7.79864669e-01 -1.88920662e-01
-8.13107491e-01 -6.38197005e-01 7.72791728e-02 9.32192624e-01
-4.76902761e-02 2.80827910e-01 -5.75584531e-01 -5.69711551e-02
8.19175839e-01 -3.14970881e-01 8.57083559e-01 -1.40951073e+00
1.53170645e-01 -1.07441060e-02 -5.61131418e-01 -9.91919711e-02
-4.71471995e-01 -1.55755728e-01 4.92108017e-02 -1.16950238e+00
-4.68859039e-02 2.23566592e-02 -2.98284024e-01 3.40853989e-01
-1.20535515e-01 5.47977507e-01 -9.87078249e-02 2.89191934e-03
-2.90834457e-01 -1.45270929e-01 7.06175387e-01 3.82601142e-01
-4.97014672e-01 9.36197639e-02 -1.82987049e-01 8.19135904e-01
9.50333178e-01 -3.07784468e-01 -5.23968041e-01 -1.28160641e-01
1.69485405e-01 4.26836193e-01 8.61121356e-01 -1.44032073e+00
2.34018207e-01 9.36244950e-02 1.02557349e+00 -4.61624980e-01
2.69906312e-01 -8.27971518e-01 6.62311196e-01 6.05216682e-01
-3.09890896e-01 5.20000577e-01 8.84690434e-02 4.17944640e-01
-5.56197856e-03 2.17252597e-01 1.15207601e+00 -2.56997973e-01
-4.99868631e-01 1.12034805e-01 -4.68565196e-01 -4.14465293e-02
1.23703671e+00 -6.12319469e-01 -2.06130892e-01 -1.98911175e-01
-8.39989007e-01 -1.85759708e-01 1.79972798e-01 2.94577152e-01
7.51335025e-01 -1.05449069e+00 -5.04513264e-01 5.46186805e-01
-9.15047899e-02 -2.27287531e-01 2.88185507e-01 1.21697628e+00
-8.00409496e-01 2.32439771e-01 -6.31165981e-01 -9.39137518e-01
-1.70777547e+00 4.35274839e-01 3.34330440e-01 2.36465737e-01
-7.19099522e-01 4.41437244e-01 -7.38024533e-01 6.77426517e-01
6.22899473e-01 -8.37578833e-01 -2.72156864e-01 3.73204410e-01
8.34527135e-01 7.73650289e-01 3.58282864e-01 -4.50148910e-01
-4.67437059e-01 5.99740207e-01 4.87580925e-01 8.72347131e-03
9.81599748e-01 -6.18481874e-01 -1.60384893e-01 6.80777967e-01
8.55771184e-01 -5.33514977e-01 -1.01600063e+00 2.78777957e-01
-1.47950230e-02 -1.66974783e-01 1.38187155e-01 -5.78860939e-01
-1.11517084e+00 1.13812077e+00 1.24150491e+00 2.97138304e-01
1.58094120e+00 -5.55951774e-01 7.95802951e-01 -3.65070924e-02
-1.52349427e-01 -1.38084221e+00 7.05389455e-02 -8.70438889e-02
3.30108911e-01 -1.10621428e+00 9.64957178e-02 -6.07212596e-02
-5.00285625e-01 1.37946701e+00 3.65498036e-01 9.78970304e-02
7.15493560e-01 5.00870645e-01 3.97902310e-01 -1.17679581e-01
-6.17330492e-01 -1.15734991e-02 -1.02759585e-01 8.72571051e-01
6.74214780e-01 -2.81638324e-01 -3.11913878e-01 9.55939293e-02
1.43530861e-01 8.09220076e-01 6.49313569e-01 9.32022572e-01
-2.56256819e-01 -7.49344409e-01 -6.42657101e-01 2.88790673e-01
-8.21043849e-01 1.30080834e-01 -2.61860967e-01 4.56332296e-01
2.13301659e-01 1.04758108e+00 2.33061183e-02 -1.40410274e-01
2.03851074e-01 4.70856041e-01 4.07901973e-01 -2.20098346e-01
-7.25056469e-01 2.71937311e-01 -1.84712663e-01 -9.37280834e-01
-8.68858397e-01 -6.02735877e-01 -1.11150849e+00 -4.40739632e-01
8.82673189e-02 -3.42315555e-01 1.05029631e+00 8.80779862e-01
4.29638445e-01 7.82224298e-01 4.84451294e-01 -8.71865153e-01
-9.18623060e-02 -8.08525681e-01 -3.61535430e-01 4.87359434e-01
6.14086151e-01 -1.39961541e-01 -8.79579559e-02 9.45757747e-01] | [13.874894142150879, 2.838090419769287] |
01e19d9f-b666-4439-b44d-ac536d7ab409 | tablelab-an-interactive-table-extraction | 2102.08445 | null | https://arxiv.org/abs/2102.08445v1 | https://arxiv.org/pdf/2102.08445v1.pdf | TableLab: An Interactive Table Extraction System with Adaptive Deep Learning | Table extraction from PDF and image documents is a ubiquitous task in the real-world. Perfect extraction quality is difficult to achieve with one single out-of-box model due to (1) the wide variety of table styles, (2) the lack of training data representing this variety and (3) the inherent ambiguity and subjectivity of table definitions between end-users. Meanwhile, building customized models from scratch can be difficult due to the expensive nature of annotating table data. We attempt to solve these challenges with TableLab by providing a system where users and models seamlessly work together to quickly customize high-quality extraction models with a few labelled examples for the user's document collection, which contains pages with tables. Given an input document collection, TableLab first detects tables with similar structures (templates) by clustering embeddings from the extraction model. Document collections often contain tables created with a limited set of templates or similar structures. It then selects a few representative table examples already extracted with a pre-trained base deep learning model. Via an easy-to-use user interface, users provide feedback to these selections without necessarily having to identify every single error. TableLab then applies such feedback to finetune the pre-trained model and returns the results of the finetuned model back to the user. The user can choose to repeat this process iteratively until obtaining a customized model with satisfactory performance. | ['Yunyao Li', 'Douglas Burdick', 'Nancy Xin Ru Wang'] | 2021-02-16 | null | null | null | null | ['table-extraction'] | ['miscellaneous'] | [ 1.03608035e-01 -3.69249433e-02 -1.18671037e-01 -3.25136304e-01
-1.17619824e+00 -1.06104934e+00 3.00127983e-01 5.48263252e-01
-2.84395844e-01 3.78533334e-01 1.35179758e-01 -2.41302982e-01
-9.19943601e-02 -7.57880092e-01 -6.08048439e-01 -1.70789778e-01
1.96952149e-01 9.23676074e-01 2.14070916e-01 -1.22347943e-01
2.92322338e-01 4.47526664e-01 -1.51129889e+00 6.45654082e-01
9.06646192e-01 8.98192048e-01 4.39198881e-01 7.51710594e-01
-6.81849539e-01 4.45120960e-01 -7.74321973e-01 -5.76746702e-01
3.98044974e-01 -2.68622696e-01 -6.68669760e-01 5.03536344e-01
7.42753804e-01 -4.41496879e-01 -1.07238971e-01 1.05437887e+00
2.25339964e-01 -6.35624528e-02 4.00099099e-01 -9.96528089e-01
-7.42720008e-01 8.75294805e-01 -2.82513976e-01 -3.59407850e-02
4.28042144e-01 1.45242855e-01 1.14300728e+00 -1.13412523e+00
8.50714445e-01 1.02794933e+00 5.00231385e-01 4.29895431e-01
-1.37427676e+00 -4.69925791e-01 1.40268713e-01 -1.35764360e-01
-1.40319145e+00 -4.21939254e-01 3.89074922e-01 -4.59901333e-01
9.47927833e-01 2.47210011e-01 5.33245504e-01 8.87055516e-01
-1.72035113e-01 6.78021014e-01 6.00704968e-01 -3.64024639e-01
2.49530405e-01 7.47320175e-01 1.13190316e-01 7.57255554e-01
5.06206393e-01 -7.88953483e-01 -4.68885541e-01 -6.27704337e-03
5.43809414e-01 -5.75955659e-02 -3.53428394e-01 -7.65220284e-01
-1.24682879e+00 4.84630704e-01 3.21392179e-01 2.24341974e-01
-3.94794077e-01 -2.70446002e-01 3.97277027e-01 2.68179864e-01
-1.09833553e-01 9.07214165e-01 -7.46429682e-01 -2.42376879e-01
-1.06857300e+00 4.78996813e-01 9.24005806e-01 1.24284077e+00
7.66392469e-01 -2.87090510e-01 -1.86492484e-02 7.96481311e-01
6.17686138e-02 3.34808737e-01 2.13134676e-01 -5.88055015e-01
7.84709811e-01 1.04629421e+00 3.32871079e-01 -9.34049964e-01
-1.27752870e-01 -3.32319647e-01 -5.03386259e-01 2.96420574e-01
7.29772806e-01 5.42276055e-02 -9.18252349e-01 1.23200822e+00
2.77299076e-01 -8.88639390e-01 -3.47186849e-02 6.02759778e-01
6.72752619e-01 6.79677248e-01 -9.22964066e-02 6.71335608e-02
1.44141090e+00 -6.74166203e-01 -7.90646255e-01 -3.91882956e-01
6.33295238e-01 -9.60408211e-01 1.60846901e+00 5.77051878e-01
-9.02404070e-01 -4.63868707e-01 -1.20359790e+00 -2.97293514e-01
-8.48758042e-01 3.15369040e-01 3.40467900e-01 5.02572596e-01
-7.85655200e-01 6.23895288e-01 -7.15207160e-01 -5.51179588e-01
5.69058776e-01 4.40895796e-01 -4.09998477e-01 -2.49420553e-01
-6.84190691e-01 8.73851657e-01 3.92789066e-01 -1.53028250e-01
-3.41176599e-01 -8.00420821e-01 -9.05146003e-01 2.65944004e-01
8.87062132e-01 -5.49864590e-01 1.33472729e+00 -5.84088862e-01
-8.71101320e-01 5.48275828e-01 -1.68338954e-01 7.75970751e-04
4.82175231e-01 -3.78622383e-01 -2.67321467e-01 -1.74811669e-02
3.39161128e-01 5.81418812e-01 6.59205079e-01 -1.45102632e+00
-5.54166019e-01 -3.97219449e-01 2.31087077e-02 1.98708221e-01
-4.29337353e-01 1.57773569e-02 -9.90254879e-01 -5.13665974e-01
2.23819241e-01 -6.04203939e-01 -1.28297612e-01 1.15180641e-01
-6.02964342e-01 -5.47517203e-02 9.23711836e-01 -7.65485704e-01
1.57090473e+00 -2.03726554e+00 -5.71605191e-02 2.48506114e-01
3.11302274e-01 8.55016559e-02 -6.28445446e-02 5.49743533e-01
1.35840714e-01 3.28220397e-01 -1.88251346e-01 -2.92061955e-01
2.34418318e-01 -3.04393843e-03 -1.24173082e-01 -1.29125923e-01
3.82376492e-01 5.46281159e-01 -9.12765980e-01 -6.07782662e-01
2.10962906e-01 3.56999248e-01 -5.87907851e-01 2.43067801e-01
-6.01369023e-01 -2.86885291e-01 -3.58886123e-01 8.58003318e-01
6.78351700e-01 -4.74261582e-01 4.74366874e-01 -4.91212696e-01
4.73798066e-02 5.49165487e-01 -1.78187704e+00 1.39709413e+00
-4.03005004e-01 4.98120785e-01 -3.66055183e-02 -3.67210448e-01
7.11717665e-01 7.42285177e-02 2.02598974e-01 -3.24070930e-01
5.16937627e-03 3.06506634e-01 3.27963643e-02 -5.90911090e-01
7.54205823e-01 3.42281669e-01 -1.91606835e-01 8.08212399e-01
1.66154101e-01 -3.03768605e-01 7.10836470e-01 5.14640689e-01
1.21720064e+00 6.83441386e-02 3.07546169e-01 3.23396400e-02
2.96647251e-01 1.47094175e-01 2.25953355e-01 8.38237643e-01
4.86425078e-03 7.11352766e-01 5.31169653e-01 -6.03098512e-01
-1.29200017e+00 -1.07897747e+00 -5.12519963e-02 1.10968804e+00
-3.06624562e-01 -9.75816667e-01 -7.16634631e-01 -1.01305556e+00
1.64347231e-01 6.04406655e-01 -5.72875082e-01 2.68982917e-01
-5.28443336e-01 -3.65611047e-01 -7.65294954e-03 7.01641977e-01
3.89578670e-01 -1.04239631e+00 -5.35424531e-01 2.54865974e-01
-1.57119557e-01 -1.09682524e+00 -5.75416446e-01 5.19710779e-01
-6.80270374e-01 -1.01964569e+00 -2.28698239e-01 -7.38384724e-01
9.52936530e-01 5.04743084e-02 1.27277696e+00 1.56674773e-01
-1.77728310e-01 1.36488646e-01 -1.70260832e-01 -3.87251168e-01
-5.54788888e-01 2.84281075e-01 -1.62384093e-01 -2.92938173e-01
6.48156106e-01 -1.14430048e-01 -1.74037158e-01 1.25742152e-01
-1.06804740e+00 6.19148184e-03 5.80122054e-01 6.47932231e-01
6.31271780e-01 4.97206539e-01 2.24537924e-01 -1.05804074e+00
8.01697552e-01 -1.02359690e-01 -8.59970272e-01 4.25122082e-01
-7.51649797e-01 3.41904014e-01 8.16037357e-01 -4.32244211e-01
-6.42086804e-01 2.69431412e-01 2.14060307e-01 -2.81854361e-01
-1.10681422e-01 6.13283455e-01 -7.09468842e-01 4.76055741e-01
7.28540719e-01 4.87194024e-02 -2.31343955e-01 -5.37659645e-01
4.88619000e-01 8.30244303e-01 6.54585719e-01 -5.09855211e-01
1.02262437e+00 -1.22439880e-02 -6.04927599e-01 -3.67071539e-01
-8.32160056e-01 -1.26509801e-01 -9.97678399e-01 -1.06342919e-01
7.76404381e-01 -5.90962231e-01 -2.84845114e-01 2.11876810e-01
-9.95460093e-01 -3.93978000e-01 -1.75787956e-01 1.17570616e-01
-1.12276472e-01 5.67928255e-02 -4.58143175e-01 -6.76413953e-01
-1.99140489e-01 -1.30174959e+00 1.00126600e+00 2.66717374e-01
-6.42182052e-01 -6.76156938e-01 -4.57550377e-01 2.83091247e-01
2.95859039e-01 1.35512710e-01 1.33575976e+00 -8.10143173e-01
-1.07669413e+00 -5.70864677e-01 -1.75475374e-01 1.92934215e-01
4.97524142e-01 4.06682968e-01 -9.26261902e-01 -2.38701142e-02
-5.34939945e-01 -3.23476493e-01 4.84559029e-01 -9.85815525e-02
1.31361210e+00 -3.97847593e-01 -3.26243848e-01 4.13367689e-01
1.43164325e+00 3.39314550e-01 4.39422011e-01 8.01573098e-01
7.40412891e-01 4.91692454e-01 4.96897012e-01 4.66534525e-01
3.78856063e-01 3.96206379e-01 7.51160607e-02 1.16309837e-01
1.36476815e-01 -4.14524019e-01 1.38578536e-02 5.53066194e-01
5.59349477e-01 -3.31515819e-01 -9.63728011e-01 5.14119208e-01
-1.56675875e+00 -7.67485857e-01 2.38586500e-01 2.35540271e+00
1.06094468e+00 7.18266964e-01 1.15776345e-01 3.81614864e-01
4.49961573e-01 -4.48626876e-02 -4.80682164e-01 -4.82675701e-01
5.20066284e-02 1.46367192e-01 4.45211917e-01 2.97988236e-01
-1.15786529e+00 1.08821785e+00 5.86550045e+00 5.34232318e-01
-1.10411334e+00 -5.37436962e-01 5.82217097e-01 -2.36177012e-01
-4.14628357e-01 -6.37242794e-02 -1.04724932e+00 3.57459813e-01
8.01669121e-01 -2.08604708e-01 6.47436738e-01 8.39510441e-01
6.99284580e-03 -2.84535103e-02 -1.50856435e+00 9.31213260e-01
1.14398664e-02 -1.38610327e+00 3.68475258e-01 1.10864714e-01
3.56203705e-01 -4.36078340e-01 2.76471470e-02 3.92508209e-01
3.44985902e-01 -9.81621444e-01 6.89849615e-01 2.36945882e-01
8.24005544e-01 -8.31828117e-01 4.61655110e-01 1.26485080e-01
-9.50029194e-01 -1.19712718e-01 -3.19138139e-01 2.86643445e-01
-1.29904985e-01 4.65167731e-01 -1.17716193e+00 8.04650709e-02
8.43639672e-01 2.67432451e-01 -9.40348804e-01 8.18691790e-01
-1.60655320e-01 3.07394952e-01 -4.68948573e-01 -1.95075333e-01
3.95454429e-02 -1.25211358e-01 2.11269855e-01 1.40127563e+00
3.07916760e-01 -1.85817540e-01 6.71048015e-02 9.94059563e-01
-3.43899757e-01 1.42575711e-01 -6.25069082e-01 -4.95788962e-01
8.18128407e-01 1.51815319e+00 -1.05864561e+00 -5.81105828e-01
-4.21442956e-01 8.35486174e-01 3.04660350e-01 2.32361272e-01
-4.01075512e-01 -8.77904952e-01 5.80456078e-01 3.41513008e-01
6.52846992e-01 -2.85557002e-01 -7.39915013e-01 -9.81776237e-01
4.27808285e-01 -1.40737343e+00 4.19109374e-01 -9.41684842e-01
-8.68853271e-01 8.29275429e-01 -1.22967392e-01 -1.08354485e+00
-3.73824686e-01 -6.26084924e-01 -2.38442719e-01 8.19720507e-01
-8.11654389e-01 -8.07674527e-01 -4.28033948e-01 3.03586930e-01
6.30737960e-01 -1.70649365e-01 8.71595562e-01 2.73901045e-01
-5.25349855e-01 8.62377465e-01 -3.16187777e-02 5.09676933e-01
9.02006507e-01 -1.77892494e+00 6.61542058e-01 9.62418914e-01
2.89467901e-01 9.62328553e-01 7.88033903e-01 -6.96689606e-01
-1.53996015e+00 -9.96168971e-01 1.23010314e+00 -9.98846650e-01
6.80089653e-01 -9.15178120e-01 -1.22165418e+00 7.28173137e-01
2.49064639e-01 -2.26804972e-01 8.11388612e-01 5.74813224e-02
-6.03730142e-01 -2.09733889e-01 -1.19141746e+00 8.92496586e-01
5.60871243e-01 -6.17001772e-01 -4.81184810e-01 3.07348698e-01
6.33809030e-01 -5.36529243e-01 -8.13792825e-01 5.23422025e-02
6.69917166e-01 -6.63981020e-01 5.67493498e-01 -4.63003963e-01
5.84971964e-01 -5.12121499e-01 -2.47194752e-01 -1.20173025e+00
-2.38151163e-01 -5.95774889e-01 -3.20826843e-02 1.45104253e+00
8.98671508e-01 -9.50610861e-02 7.94921875e-01 1.12591791e+00
1.77576289e-01 -8.01412106e-01 -1.78836092e-01 -5.00504911e-01
-1.70507789e-01 -3.87833834e-01 9.78278995e-01 8.03252161e-01
1.23813272e-01 4.80421066e-01 1.19536400e-01 3.19046974e-01
3.37708026e-01 1.75710753e-01 1.19229424e+00 -1.08295095e+00
-2.51091838e-01 -3.63870353e-01 3.22440616e-03 -9.82291758e-01
-5.07814229e-01 -6.95402145e-01 1.03868648e-01 -1.84680760e+00
1.02947302e-01 -5.31182945e-01 -1.07443757e-01 5.17973661e-01
-3.23656291e-01 -1.04440093e-01 4.27759886e-01 1.54920295e-01
-7.34893680e-01 -4.13446352e-02 1.04907143e+00 -1.40280232e-01
-5.11535406e-01 -1.58077955e-01 -1.13371503e+00 5.35367250e-01
6.24012053e-01 -5.75671136e-01 -4.17964339e-01 -5.26548326e-01
3.41957778e-01 -6.16122372e-02 -1.08050659e-01 -1.15348732e+00
1.98577553e-01 -2.70287390e-03 1.06274068e+00 -9.38600242e-01
1.35901168e-01 -7.65268445e-01 -2.53351033e-01 -4.31419909e-02
-4.20031428e-01 4.21582997e-01 3.98778439e-01 1.89294562e-01
1.87361066e-03 -2.91690171e-01 5.63474655e-01 -3.02357793e-01
-4.17998463e-01 1.02546014e-01 -5.26059508e-01 -1.71005304e-04
4.43966120e-01 -4.60010409e-01 -2.42295429e-01 -3.26840222e-01
-6.61695480e-01 2.64317602e-01 7.31832981e-01 4.75669980e-01
4.62750733e-01 -1.12288380e+00 -2.49922946e-01 4.57270592e-01
4.11162138e-01 3.45345914e-01 -2.80494094e-01 1.73196852e-01
-4.77380961e-01 1.37880847e-01 -5.91789298e-02 -5.22704303e-01
-1.20764208e+00 7.63664663e-01 1.87179103e-01 -4.20207500e-01
-5.84117711e-01 7.50577152e-01 -2.17295915e-01 -5.29639959e-01
4.62526798e-01 -6.93105698e-01 -1.38001451e-02 2.45568603e-01
7.06930101e-01 -1.15173668e-01 4.85298038e-01 -6.22561052e-02
-3.73834550e-01 1.27346903e-01 -5.99554479e-01 -2.38739640e-01
1.35884011e+00 -1.16490973e-02 2.20472604e-01 3.87160540e-01
1.09817600e+00 2.67375141e-01 -1.25232255e+00 -2.14488581e-01
2.29930714e-01 -6.29572451e-01 -2.86894828e-01 -1.08204103e+00
-7.71943927e-01 6.42572045e-01 3.78673345e-01 1.73797771e-01
9.90606487e-01 -5.97875863e-02 7.72419095e-01 7.69366145e-01
1.25533342e-01 -1.30718350e+00 1.81705803e-01 4.17344391e-01
8.13018680e-01 -1.19398439e+00 1.30296201e-01 -1.71052471e-01
-6.99004292e-01 1.36740649e+00 8.21785271e-01 1.36145994e-01
4.82040286e-01 6.29814506e-01 4.20685619e-01 -2.20910564e-01
-9.01216447e-01 -6.34669363e-02 2.29584455e-01 6.85431957e-01
5.56729794e-01 -1.63206533e-01 1.58137158e-01 5.90543270e-01
-6.87209547e-01 2.33174115e-02 7.26210117e-01 1.19412863e+00
-5.33708751e-01 -1.34495747e+00 -5.26445985e-01 8.27559948e-01
-3.41187894e-01 -1.39812440e-01 -7.11947262e-01 9.51861799e-01
1.23138547e-01 7.51599133e-01 2.29304180e-01 -3.98910284e-01
5.44282138e-01 2.67381370e-01 3.31348538e-01 -7.78922081e-01
-6.15128100e-01 2.73818195e-01 9.11925584e-02 -3.69188756e-01
3.99741650e-01 -7.36249685e-01 -1.20649898e+00 -1.65164679e-01
-7.25755319e-02 -5.98260340e-05 6.62629247e-01 6.88922167e-01
4.50551510e-01 3.30222040e-01 2.83226818e-01 -4.81894344e-01
-6.51368141e-01 -9.64324713e-01 -3.29492420e-01 4.71120179e-01
2.63736337e-01 -3.74423712e-01 -8.21023658e-02 3.33257347e-01] | [11.690034866333008, 2.918715000152588] |
947e86d3-9016-4233-8de2-4354be2f5622 | dictionary-learning-under-symmetries-via | 2305.19557 | null | https://arxiv.org/abs/2305.19557v1 | https://arxiv.org/pdf/2305.19557v1.pdf | Dictionary Learning under Symmetries via Group Representations | The dictionary learning problem can be viewed as a data-driven process to learn a suitable transformation so that data is sparsely represented directly from example data. In this paper, we examine the problem of learning a dictionary that is invariant under a pre-specified group of transformations. Natural settings include Cryo-EM, multi-object tracking, synchronization, pose estimation, etc. We specifically study this problem under the lens of mathematical representation theory. Leveraging the power of non-abelian Fourier analysis for functions over compact groups, we prescribe an algorithmic recipe for learning dictionaries that obey such invariances. We relate the dictionary learning problem in the physical domain, which is naturally modelled as being infinite dimensional, with the associated computational problem, which is necessarily finite dimensional. We establish that the dictionary learning problem can be effectively understood as an optimization instance over certain matrix orbitopes having a particular block-diagonal structure governed by the irreducible representations of the group of symmetries. This perspective enables us to introduce a band-limiting procedure which obtains dimensionality reduction in applications. We provide guarantees for our computational ansatz to provide a desirable dictionary learning outcome. We apply our paradigm to investigate the dictionary learning problem for the groups SO(2) and SO(3). While the SO(2) orbitope admits an exact spectrahedral description, substantially less is understood about the SO(3) orbitope. We describe a tractable spectrahedral outer approximation of the SO(3) orbitope, and contribute an alternating minimization paradigm to perform optimization in this setting. We provide numerical experiments to highlight the efficacy of our approach in learning SO(3) invariant dictionaries, both on synthetic and on real world data. | ['Brendan K. Y. Tan', 'Zhuohang Feng', 'Yong Sheng Soh', 'Aaron Y. R. Low', 'Subhroshekhar Ghosh'] | 2023-05-31 | null | null | null | null | ['pose-estimation', 'object-tracking', 'multi-object-tracking', 'dimensionality-reduction'] | ['computer-vision', 'computer-vision', 'computer-vision', 'methodology'] | [ 3.50043982e-01 2.37011716e-01 -1.61582083e-01 3.74929160e-02
-4.03407454e-01 -6.00167096e-01 8.29706073e-01 -1.36980057e-01
-2.86287636e-01 4.51090395e-01 2.62673825e-01 -1.74855635e-01
-4.54758912e-01 -6.39064312e-01 -6.31235957e-01 -1.34776688e+00
-4.32112552e-02 6.35774136e-01 -3.34434420e-01 -5.72929919e-01
1.69741735e-01 6.23642325e-01 -1.35654783e+00 -3.20433944e-01
5.62450826e-01 8.25964749e-01 3.60290036e-02 4.45657849e-01
3.44364256e-01 2.51000226e-01 2.19838303e-02 6.69719875e-02
6.85674191e-01 -5.81426859e-01 -9.29884791e-01 5.31665742e-01
4.47203368e-01 3.85918230e-01 -4.12819743e-01 1.11034918e+00
4.85655814e-01 2.49796107e-01 8.56788039e-01 -9.30298209e-01
-5.17047346e-01 3.59147698e-01 -2.24236995e-01 2.39582226e-01
1.29822463e-01 -7.09482580e-02 1.14038992e+00 -8.38649392e-01
8.12530696e-01 8.76909137e-01 3.92747462e-01 3.47797781e-01
-1.78708863e+00 -1.87350437e-01 -3.80307466e-01 -1.28951833e-01
-1.59889662e+00 -4.46428895e-01 9.46630239e-01 -7.52713442e-01
3.89775515e-01 3.85247976e-01 6.78024769e-01 8.09714913e-01
3.13907534e-01 2.44523078e-01 1.16288257e+00 -8.37087691e-01
4.91705239e-01 2.72300523e-02 2.64170140e-01 7.44697034e-01
4.35945481e-01 2.04269201e-01 -5.88257790e-01 -2.83277452e-01
6.84760630e-01 -5.67840859e-02 -4.69666541e-01 -9.82516527e-01
-1.37839282e+00 1.15365946e+00 -2.11131535e-02 4.85610008e-01
-3.28761101e-01 -2.17391878e-01 3.03929448e-01 6.88237071e-01
1.45370409e-01 5.75680315e-01 6.11767881e-02 2.55961865e-01
-6.07125998e-01 2.62292355e-01 1.10638368e+00 9.03357208e-01
8.79961908e-01 2.83584774e-01 3.68140161e-01 4.46219563e-01
1.11823976e-01 5.30514836e-01 2.97002822e-01 -6.97968960e-01
4.84063327e-02 -3.25294212e-02 -9.03128181e-04 -8.71503532e-01
-4.41445917e-01 -6.27151251e-01 -1.11337650e+00 9.23911110e-02
4.23674494e-01 5.50623760e-02 -3.25965017e-01 1.84811652e+00
4.63758320e-01 -9.79577005e-02 -3.63078006e-02 8.98925960e-01
7.52382725e-02 4.87158060e-01 -3.42687964e-01 -6.74909651e-01
1.21546590e+00 -3.32250059e-01 -4.53690439e-01 3.01822722e-01
8.42613995e-01 -5.45583129e-01 9.12541330e-01 6.54268980e-01
-8.54669631e-01 -3.73555183e-01 -1.11621511e+00 1.32493258e-01
-4.78137583e-02 -5.78598715e-02 6.20649219e-01 3.55455071e-01
-9.13578153e-01 7.84525216e-01 -6.67468905e-01 -6.09319270e-01
-1.78264260e-01 4.59993094e-01 -6.01742923e-01 2.47425959e-01
-1.03238761e+00 9.15977001e-01 4.82787251e-01 -5.64843751e-02
-7.69599319e-01 -4.82998520e-01 -7.59464204e-01 -2.69886494e-01
3.00344974e-01 -8.45502794e-01 8.06753755e-01 -8.94353449e-01
-1.33244705e+00 1.18447852e+00 9.83717591e-02 -3.46223474e-01
2.19051480e-01 4.03097212e-01 -3.57144654e-01 1.18749775e-01
-3.40462029e-02 -3.97594385e-02 1.37665009e+00 -1.04397404e+00
-1.54749118e-02 -4.05792326e-01 2.37329483e-01 6.16933070e-02
-3.39173675e-01 -2.05437124e-01 3.15045744e-01 -8.02361608e-01
5.22195756e-01 -1.34785593e+00 -2.32067421e-01 -1.56046450e-01
-2.60832667e-01 8.74206200e-02 7.52487063e-01 -2.40372121e-01
1.21468127e+00 -2.14606643e+00 1.04353023e+00 5.38232863e-01
3.59928817e-01 -1.19520158e-01 2.24478304e-01 8.99874806e-01
-5.74792564e-01 -2.91987002e-01 -2.86387116e-01 -1.49851859e-01
1.02154590e-01 3.51137340e-01 -5.57894289e-01 1.21125996e+00
-7.43975490e-02 4.14166451e-01 -5.06033361e-01 -2.00532421e-01
-2.29043085e-02 3.05355519e-01 -1.03541076e+00 1.03291115e-02
-2.02020362e-01 7.75246680e-01 -6.56250000e-01 4.22288477e-01
4.57345158e-01 -2.59208024e-01 2.28906110e-01 -4.95205641e-01
-3.56003642e-01 8.40279981e-02 -1.63601398e+00 1.78089130e+00
-3.88197899e-01 3.52422148e-01 5.46331167e-01 -1.68440342e+00
8.55123937e-01 1.66802570e-01 7.75561869e-01 -2.89558351e-01
2.34984532e-01 1.40213832e-01 1.91889748e-01 -4.55535352e-01
2.83936858e-01 -6.80647492e-01 -1.95481256e-01 6.72744095e-01
3.12903374e-01 -2.27870286e-01 2.50038151e-02 2.00758368e-01
9.75852489e-01 -3.52784246e-01 5.69720566e-01 -1.04942906e+00
5.87226510e-01 -1.00367531e-01 1.31966844e-01 6.39732063e-01
2.27225780e-01 4.01139349e-01 4.66089517e-01 -6.07273281e-01
-1.23941755e+00 -8.20735216e-01 -6.69162095e-01 6.84101820e-01
-3.95295098e-02 -5.30013025e-01 -6.29827738e-01 5.65622933e-03
1.46318868e-01 1.95147157e-01 -6.58440828e-01 -3.31617683e-01
-5.14558494e-01 -7.87962198e-01 9.85825062e-02 -2.49392882e-01
8.56651887e-02 -7.54166424e-01 -4.32824492e-01 1.39066607e-01
2.53977269e-01 -8.67190599e-01 -5.28478205e-01 5.22706389e-01
-7.63736486e-01 -9.86462593e-01 -3.42700422e-01 -6.85244262e-01
5.17283082e-01 1.70110166e-01 8.57745826e-01 1.72536413e-03
-3.68291825e-01 9.26360369e-01 -6.66294694e-02 -2.23711163e-01
-4.73257005e-01 5.16766906e-02 8.31735373e-01 6.23269677e-01
7.60065913e-02 -1.02793264e+00 -2.25193813e-01 2.32594520e-01
-1.34287965e+00 -4.62499335e-02 1.76052496e-01 9.36899126e-01
7.46805310e-01 8.86376351e-02 3.65300089e-01 -5.91488659e-01
3.53728831e-01 -5.21629512e-01 -7.37268269e-01 3.67816538e-03
-4.34007704e-01 6.59836113e-01 7.06506193e-01 -5.58983564e-01
-5.18978119e-01 3.45204234e-01 1.73877627e-01 -3.62075746e-01
2.44525596e-01 6.69712186e-01 -4.48796242e-01 -7.05177844e-01
7.20971525e-01 5.22865891e-01 2.77653247e-01 -5.51409304e-01
2.52518982e-01 6.26298487e-01 4.43769395e-01 -1.06394839e+00
9.61618066e-01 7.22809911e-01 5.45796931e-01 -1.46025360e+00
-8.49333346e-01 -4.66455579e-01 -8.90750825e-01 -5.53055527e-03
6.64044738e-01 -8.39789271e-01 -7.25791693e-01 1.11243494e-01
-7.96861529e-01 -1.57114357e-01 -5.88822722e-01 5.67408919e-01
-1.12293720e+00 4.10468400e-01 -1.87926650e-01 -5.41252315e-01
-1.22938626e-01 -9.70645905e-01 1.13656926e+00 -4.11143929e-01
-1.69495329e-01 -1.22269452e+00 5.11778593e-01 1.65619060e-01
2.79379547e-01 3.51316005e-01 1.04303110e+00 -5.06639838e-01
-6.38496757e-01 -7.73769291e-03 4.89137918e-01 2.86599100e-01
-1.04076043e-01 -3.71103853e-01 -5.52577078e-01 -6.13243282e-01
7.17007220e-01 -1.38161004e-01 7.13993788e-01 1.59060553e-01
1.00920045e+00 -2.65012503e-01 -9.33249891e-02 9.27882135e-01
1.49200332e+00 -2.52909303e-01 2.71989077e-01 1.61267802e-01
8.11685979e-01 5.36435544e-01 3.70489717e-01 5.83730221e-01
6.90463185e-02 1.08524191e+00 2.71206379e-01 2.93072820e-01
9.02750865e-02 -1.21523999e-02 2.69907296e-01 1.21112227e+00
-1.74205378e-01 2.25891337e-01 -8.86734068e-01 2.72125870e-01
-1.59185684e+00 -1.02829313e+00 -1.25839934e-01 2.39856768e+00
8.17438662e-01 -7.51024857e-02 2.85523385e-01 2.80104756e-01
4.64737982e-01 2.44356200e-01 -3.85816157e-01 -1.95766255e-01
-2.69720584e-01 4.59161073e-01 3.39643210e-01 7.05708146e-01
-1.05130064e+00 5.13131738e-01 6.16245031e+00 7.54866660e-01
-1.30490470e+00 1.79877371e-01 -2.83842571e-02 9.98892114e-02
-4.90261644e-01 2.45699555e-01 -8.12025905e-01 3.98539424e-01
7.90712655e-01 -4.42390829e-01 7.01578915e-01 6.33912206e-01
2.59840816e-01 7.93349966e-02 -1.32456505e+00 1.31677461e+00
5.47718443e-02 -1.47838831e+00 2.35417068e-01 6.52734160e-01
7.17111409e-01 -1.76935971e-01 1.27118692e-01 -1.02022111e-01
-1.24392144e-01 -8.72010350e-01 6.94128990e-01 5.56602180e-01
8.56685817e-01 -5.21831512e-01 -4.31096070e-02 5.37594557e-01
-9.24277246e-01 -5.43795759e-04 -4.18776929e-01 -3.15293819e-01
8.43622163e-02 6.95152104e-01 -1.92856565e-01 5.03337324e-01
8.04532915e-02 1.01090419e+00 -1.67482288e-03 7.08947003e-01
4.74775672e-01 4.97274637e-01 -5.37857711e-01 2.98577756e-01
2.06229597e-01 -1.01780021e+00 8.77630055e-01 8.58163238e-01
4.36936289e-01 7.30210423e-01 2.46757120e-01 6.78219736e-01
-2.17038002e-02 9.52945575e-02 -1.02824342e+00 -5.14807068e-02
-3.93325649e-02 1.22801936e+00 -6.95827901e-01 3.57769616e-02
-3.35186571e-01 5.80989718e-01 1.86947942e-01 2.12533250e-01
-4.08100188e-01 4.06555347e-02 8.72536004e-01 4.47803408e-01
5.62481403e-01 -5.42555869e-01 7.87360594e-02 -1.65928936e+00
5.62015921e-02 -1.09910727e+00 1.78825885e-01 -3.86644453e-01
-1.09623551e+00 3.70761961e-01 1.57633811e-01 -1.42848492e+00
-1.69584960e-01 -8.75926375e-01 -4.01168317e-01 6.14521265e-01
-1.06952846e+00 -9.09582138e-01 7.79985785e-02 9.26705539e-01
2.31764242e-02 -2.27118239e-01 9.93874431e-01 2.67983675e-01
-3.93306553e-01 1.41438469e-01 6.08733594e-01 -3.37696999e-01
2.98446208e-01 -1.30341792e+00 -2.24898294e-01 6.33586168e-01
3.56501877e-01 6.69794679e-01 1.02939057e+00 -2.85151064e-01
-2.18104148e+00 -7.91543067e-01 7.06279874e-01 -5.76811433e-01
1.08375323e+00 -6.23365641e-01 -8.12996030e-01 7.79602945e-01
-2.30786771e-01 4.05449361e-01 6.77412152e-01 -7.51466304e-02
-2.44646356e-01 -5.19606285e-02 -8.96925628e-01 3.52808803e-01
1.40007579e+00 -7.89247453e-01 -5.45124769e-01 8.24785948e-01
3.43205601e-01 -4.10970658e-01 -1.15249658e+00 6.39234856e-03
3.29132169e-01 -6.57685637e-01 1.00586557e+00 -9.27773476e-01
6.21766262e-02 -2.82283902e-01 -4.72891241e-01 -1.30201232e+00
-2.97561646e-01 -1.15478551e+00 -1.02504790e-01 6.88068271e-01
-2.11803839e-01 -6.11876190e-01 4.96497959e-01 4.64782193e-02
-4.69629951e-02 -5.72530687e-01 -1.04309177e+00 -7.47647226e-01
2.27269843e-01 -1.60834149e-01 1.47107869e-01 1.24588728e+00
7.33603835e-02 5.33120096e-01 -4.74817395e-01 2.58099169e-01
9.05909777e-01 2.75947452e-01 6.39122009e-01 -1.35793686e+00
-7.07380056e-01 -9.52666551e-02 -6.50558710e-01 -1.05839694e+00
4.57926691e-01 -1.27195895e+00 -4.02336448e-01 -5.37072361e-01
6.20853081e-02 -6.14169002e-01 1.49718970e-01 -6.60845935e-02
6.00965321e-01 -4.86395136e-02 1.16444632e-01 4.78940964e-01
-4.19829130e-01 5.90338349e-01 1.15934563e+00 -5.12676686e-02
-1.53497458e-02 4.63587865e-02 -7.62021005e-01 5.25196671e-01
5.24212778e-01 -4.59291518e-01 -3.42521101e-01 -1.24849677e-01
5.79789639e-01 2.10679859e-01 4.28643942e-01 -9.65561569e-01
3.16339016e-01 -2.04220176e-01 -2.61482269e-01 -4.05327566e-02
2.02957839e-01 -9.30112898e-01 5.08500218e-01 2.91994780e-01
-2.73695499e-01 -7.21778050e-02 -1.89121142e-01 5.15972078e-01
-2.06042260e-01 -4.62828308e-01 8.67336094e-01 -7.37731680e-02
-3.03656518e-01 4.39691424e-01 -2.96789199e-01 3.03160161e-01
8.89696002e-01 -6.71416745e-02 3.05396020e-01 -4.55418795e-01
-1.18713903e+00 -1.69622377e-01 6.79396927e-01 -5.08476496e-02
1.50506750e-01 -1.50426102e+00 -5.49797773e-01 6.44661665e-01
1.55375302e-01 -1.50103077e-01 -5.06558083e-03 1.00065422e+00
-3.46254796e-01 5.02295673e-01 -2.61565924e-01 -7.39232540e-01
-9.74081397e-01 6.62532687e-01 6.05608940e-01 -1.13865070e-01
-6.06216669e-01 3.37197930e-01 2.22259402e-01 -3.64887923e-01
-1.58920251e-02 -2.89584577e-01 8.22020769e-02 1.45360902e-02
3.68106991e-01 1.93164796e-01 2.73829669e-01 -1.01160610e+00
-5.08283377e-02 8.05802166e-01 1.12695143e-01 -3.55784930e-02
1.53044128e+00 -5.73185235e-02 -4.55363125e-01 2.49030605e-01
1.38365638e+00 1.36836022e-01 -8.40621710e-01 -5.67171931e-01
-1.44670472e-01 -8.35503265e-02 1.52497292e-01 1.08379811e-01
-7.95468450e-01 5.56911826e-01 2.53794700e-01 7.13581979e-01
9.37592447e-01 3.67529899e-01 3.86877626e-01 5.99505126e-01
6.15010977e-01 -7.55605459e-01 1.87108628e-02 6.30012453e-01
1.08433938e+00 -1.03856778e+00 -3.92694883e-02 -2.39185318e-01
-1.86551422e-01 1.27437794e+00 -6.21806942e-02 -5.15792251e-01
1.08325171e+00 7.60412291e-02 -4.78147417e-01 -5.00984013e-01
-4.98148680e-01 -1.74736694e-01 3.93335968e-01 4.72223401e-01
7.04141557e-02 1.47546396e-01 -3.57485145e-01 2.88679868e-01
-5.66087484e-01 -3.97190839e-01 7.54495621e-01 7.69153118e-01
-4.09280658e-01 -1.34112537e+00 -4.71789926e-01 3.56343716e-01
-1.85560450e-01 8.24795961e-02 -1.57867774e-01 8.47601235e-01
9.71350595e-02 3.71040136e-01 -1.09737612e-01 -1.43007517e-01
2.61956781e-01 4.71971594e-02 7.28659630e-01 -8.40903938e-01
-8.18342865e-02 -3.14888693e-02 -1.57997593e-01 -4.77585793e-01
-6.33887649e-01 -1.02945006e+00 -9.41818535e-01 -4.40727472e-01
-1.95283487e-01 4.77066636e-01 4.46065784e-01 1.08248186e+00
1.21233799e-01 4.71718237e-02 6.72337532e-01 -9.68981385e-01
-1.13336861e+00 -5.76012254e-01 -1.20773125e+00 3.98101628e-01
7.25877464e-01 -8.91655684e-01 -6.00653231e-01 2.64008772e-02] | [7.162046909332275, 4.5020575523376465] |
ab057473-cd92-4caa-b9e0-a1c47af47e92 | high-resolution-synthetic-rgb-d-datasets-for | 2305.01732 | null | https://arxiv.org/abs/2305.01732v1 | https://arxiv.org/pdf/2305.01732v1.pdf | High-Resolution Synthetic RGB-D Datasets for Monocular Depth Estimation | Accurate depth maps are essential in various applications, such as autonomous driving, scene reconstruction, point-cloud creation, etc. However, monocular-depth estimation (MDE) algorithms often fail to provide enough texture & sharpness, and also are inconsistent for homogeneous scenes. These algorithms mostly use CNN or vision transformer-based architectures requiring large datasets for supervised training. But, MDE algorithms trained on available depth datasets do not generalize well and hence fail to perform accurately in diverse real-world scenes. Moreover, the ground-truth depth maps are either lower resolution or sparse leading to relatively inconsistent depth maps. In general, acquiring a high-resolution ground truth dataset with pixel-level precision for accurate depth prediction is an expensive, and time-consuming challenge. In this paper, we generate a high-resolution synthetic depth dataset (HRSD) of dimension 1920 X 1080 from Grand Theft Auto (GTA-V), which contains 100,000 color images and corresponding dense ground truth depth maps. The generated datasets are diverse and have scenes from indoors to outdoors, from homogeneous surfaces to textures. For experiments and analysis, we train the DPT algorithm, a state-of-the-art transformer-based MDE algorithm on the proposed synthetic dataset, which significantly increases the accuracy of depth maps on different scenes by 9 %. Since the synthetic datasets are of higher resolution, we propose adding a feature extraction module in the transformer encoder and incorporating an attention-based loss, further improving the accuracy by 15 %. | ['Sunil Jaiswal', 'Philipp Slusallek', 'Klaus Illgner-Fehns', 'Noshaba Cheema', 'Aakash Rajpal'] | 2023-05-02 | null | null | null | null | ['monocular-depth-estimation'] | ['computer-vision'] | [ 3.55430514e-01 -2.25846380e-01 2.30775386e-01 -5.61148763e-01
-8.20730329e-01 -1.35646030e-01 5.28363347e-01 -2.51282096e-01
-1.95206195e-01 8.42010677e-01 -9.92644504e-02 4.08716546e-03
1.12397932e-01 -1.35799778e+00 -9.18698013e-01 -6.29938900e-01
3.61502618e-01 4.81248379e-01 4.62127000e-01 -1.36964679e-01
3.35378975e-01 3.49170685e-01 -1.91917276e+00 2.32460052e-01
1.07383752e+00 1.49146163e+00 5.90327263e-01 4.86698896e-01
-2.41282687e-01 9.56974685e-01 -1.92565903e-01 -3.05233270e-01
5.15646100e-01 -8.03007483e-02 -4.36205685e-01 2.49352157e-02
8.91996801e-01 -9.25111651e-01 -6.38841450e-01 1.06485772e+00
4.25759882e-01 -1.20759690e-02 4.27595913e-01 -1.00523531e+00
-4.42483306e-01 -2.38496512e-01 -6.39477193e-01 -1.99592829e-01
3.14504802e-01 2.53805339e-01 4.91432726e-01 -9.73665237e-01
5.48436046e-01 1.28970551e+00 4.93817568e-01 6.24344885e-01
-9.82365310e-01 -7.67174244e-01 -4.38081883e-02 3.49323422e-01
-1.24377465e+00 -3.83836538e-01 9.16592002e-01 -4.21315432e-01
7.66532362e-01 -2.80788057e-02 7.77656674e-01 1.22220087e+00
2.75694668e-01 5.61174810e-01 1.36281538e+00 3.54031548e-02
3.67687106e-01 -2.08073944e-01 -3.63716990e-01 6.77011669e-01
3.81252140e-01 3.20258677e-01 -6.71127260e-01 4.46979940e-01
1.16422379e+00 1.89832956e-01 -4.22254920e-01 -4.56699044e-01
-1.29859209e+00 6.19787157e-01 6.41269386e-01 -2.28877559e-01
-4.51520979e-01 1.63224891e-01 1.51793167e-01 2.19489753e-01
6.20325744e-01 1.44000128e-01 -3.87698680e-01 -1.77359074e-01
-6.44548595e-01 4.40397471e-01 1.63976043e-01 1.09030187e+00
1.24949646e+00 8.00873786e-02 8.41325819e-02 8.06527078e-01
1.42846823e-01 8.91085029e-01 3.92838687e-01 -1.23445511e+00
7.60982692e-01 5.15411556e-01 2.56751984e-01 -1.20425332e+00
-2.34872282e-01 -2.42908940e-01 -1.23613703e+00 3.40083927e-01
3.32592160e-01 1.56204492e-01 -1.05181503e+00 1.47101355e+00
2.81563073e-01 1.72287330e-01 1.63953111e-01 1.17055511e+00
9.92002130e-01 7.48826385e-01 -4.80724573e-01 4.80402596e-02
8.57516289e-01 -9.33128715e-01 -5.58791041e-01 -5.98077059e-01
2.73829967e-01 -6.12397194e-01 1.22721207e+00 7.01936781e-01
-8.75589490e-01 -6.64926887e-01 -1.11327350e+00 -5.76539993e-01
-8.00613239e-02 4.34282981e-02 8.32993925e-01 8.54995772e-02
-8.73484850e-01 4.16429162e-01 -7.92804360e-01 -1.19821504e-01
4.84472215e-01 -7.04787448e-02 -5.88905752e-01 -8.06999803e-01
-1.01600051e+00 4.78155881e-01 2.35832587e-01 1.99648291e-01
-1.11045253e+00 -7.83915877e-01 -1.12051821e+00 -3.63423437e-01
1.37578040e-01 -7.98548877e-01 9.45338368e-01 -7.43986130e-01
-1.47334957e+00 8.00141573e-01 -1.78410009e-01 -3.31768006e-01
7.45725870e-01 -1.81445956e-01 -1.04376726e-01 1.94083452e-01
4.06247020e-01 9.31848407e-01 7.33266056e-01 -1.33121562e+00
-8.45376968e-01 -6.36862218e-01 1.48947507e-01 2.52020508e-01
1.40869260e-01 -7.99673021e-01 -3.82074922e-01 -2.54410386e-01
6.60861552e-01 -5.63212812e-01 -1.49360418e-01 3.62095237e-01
-3.71423513e-01 4.53048706e-01 7.66660511e-01 -5.91348708e-01
6.00311458e-01 -2.15823054e+00 -9.41905286e-03 -1.93429291e-01
3.24841470e-01 -2.05832511e-01 -8.70039389e-02 3.35099660e-02
2.74776965e-01 -3.17228496e-01 -3.42654735e-01 -5.31492531e-01
-2.98564464e-01 4.21429574e-01 -2.83070862e-01 3.75679284e-01
9.47096050e-02 6.96648121e-01 -9.02058959e-01 -3.61917198e-01
7.30576932e-01 5.62608540e-01 -5.47407985e-01 3.33772451e-01
-4.36052859e-01 8.73343527e-01 -4.16737616e-01 8.18425477e-01
1.12102973e+00 -4.97884080e-02 -3.75309825e-01 -3.05931121e-01
-2.92897671e-01 2.98220038e-01 -1.03691065e+00 2.11374140e+00
-8.28217864e-01 8.00634623e-01 -1.47048071e-01 -6.57537401e-01
1.17300320e+00 -2.47919977e-01 4.72462445e-01 -1.28966677e+00
-2.89184991e-02 4.53796625e-01 -2.75641680e-01 -3.88483614e-01
7.61495888e-01 -1.75166056e-01 8.80513638e-02 -9.50590074e-02
-2.22143471e-01 -7.65445769e-01 -1.30524129e-01 -1.59115821e-01
1.02943599e+00 1.06824569e-01 -2.71300673e-01 1.53112501e-01
3.86868060e-01 1.02935992e-01 7.29621053e-01 3.57772827e-01
5.72703108e-02 1.12801182e+00 2.17790142e-01 -6.54223144e-01
-1.37713468e+00 -1.07971919e+00 -4.00202066e-01 -8.25550966e-03
5.48095465e-01 6.55400306e-02 -5.08184314e-01 -7.16442987e-02
-7.66996443e-02 3.82797420e-01 -5.65592885e-01 -1.37001291e-01
-2.94860035e-01 -6.08345389e-01 1.47410467e-01 4.18213487e-01
1.44664037e+00 -8.34916890e-01 -7.04667389e-01 2.60103464e-01
-5.02452910e-01 -1.54643667e+00 4.97916825e-02 7.58340675e-03
-1.00501442e+00 -1.01627266e+00 -7.43875802e-01 -5.43119073e-01
5.09433687e-01 4.56633419e-01 1.19886744e+00 -3.90181512e-01
-1.76706612e-01 -1.00895114e-01 -3.15622002e-01 -2.37426057e-01
5.82406186e-02 -1.13497481e-01 -1.17478311e-01 9.72221419e-02
8.97391289e-02 -5.74675202e-01 -9.03240800e-01 4.36197728e-01
-8.28873158e-01 5.59987247e-01 4.52922106e-01 1.00888634e+00
1.03152752e+00 2.93285936e-01 1.79916114e-01 -6.82746887e-01
-1.60341665e-01 -1.90127999e-01 -9.09900367e-01 -3.20069343e-01
-4.59294111e-01 -1.45282134e-01 6.63770914e-01 -3.41433361e-02
-1.21068275e+00 2.19620951e-02 -2.88179487e-01 -7.57747710e-01
-3.59973982e-02 1.49164349e-01 -2.52115518e-01 -2.86434770e-01
7.05197573e-01 5.78762889e-01 -1.66182041e-01 -3.96689236e-01
-2.00565141e-02 4.80251193e-01 7.82356739e-01 -4.81270105e-01
7.82702148e-01 8.41276586e-01 2.64762938e-01 -7.16067374e-01
-8.91474724e-01 -1.77759632e-01 -3.05765390e-01 -1.44927874e-01
7.85507441e-01 -1.43953979e+00 -3.86008382e-01 1.09031236e+00
-1.02357030e+00 -6.91870213e-01 -1.00853823e-01 5.21483481e-01
-6.90256476e-01 2.22926378e-01 -5.42528510e-01 -5.22082090e-01
-2.01394126e-01 -1.11014080e+00 1.50570846e+00 2.37284452e-01
5.05352616e-01 -5.70018232e-01 -1.23919941e-01 5.66137731e-01
4.25170690e-01 5.38475633e-01 7.66444325e-01 6.78245485e-01
-1.29327798e+00 2.97974329e-02 -6.37558877e-01 3.96706015e-01
3.37028891e-01 -6.12247102e-02 -1.18602526e+00 -1.26982301e-01
-5.17332368e-03 -5.45562327e-01 9.15613413e-01 5.15088797e-01
1.36310410e+00 -2.11050771e-02 2.23119818e-02 1.13927567e+00
1.82140398e+00 2.04719216e-01 1.04879010e+00 4.51607019e-01
1.09637880e+00 6.11281276e-01 1.10861957e+00 5.66724837e-01
7.44429290e-01 6.62183344e-01 8.24647903e-01 -1.02207646e-01
-1.74393281e-01 -4.55712736e-01 1.37157097e-01 5.89587808e-01
-3.99187170e-02 -1.40358463e-01 -7.79029310e-01 5.55927992e-01
-1.55261183e+00 -7.84560084e-01 -2.87944376e-01 2.34820104e+00
7.38671422e-01 2.89249569e-01 -3.32834214e-01 3.08285803e-01
2.73473024e-01 2.42499813e-01 -8.61210227e-01 2.36764196e-02
-4.64671075e-01 2.09991604e-01 6.90284252e-01 6.37957871e-01
-8.24545562e-01 1.03964806e+00 4.96447134e+00 7.97723353e-01
-1.51203275e+00 -3.11049949e-02 7.56345093e-01 -9.03387889e-02
-5.62951863e-01 -1.93964213e-01 -6.18396997e-01 4.85053778e-01
5.17768264e-01 1.32836461e-01 4.03855711e-01 9.18085754e-01
2.74241447e-01 -5.62692821e-01 -9.28167760e-01 1.63926351e+00
-1.38986066e-01 -1.35874295e+00 2.99423914e-02 -1.00296875e-02
1.10947835e+00 2.81273127e-01 8.65500718e-02 6.49448335e-02
8.74921605e-02 -9.93018925e-01 8.83577645e-01 6.13770306e-01
1.27778268e+00 -7.80016959e-01 7.80118406e-01 3.82787228e-01
-1.12744069e+00 8.39074627e-02 -8.40044975e-01 -1.58108920e-01
8.80381688e-02 1.02066851e+00 -3.23474884e-01 6.13705933e-01
1.00240576e+00 1.19534254e+00 -5.41293442e-01 9.26156223e-01
-1.75142109e-01 -3.48161906e-02 -3.03921640e-01 2.06434354e-01
1.62123993e-01 -2.53417522e-01 1.26776531e-01 3.36845219e-01
5.57215810e-01 1.60121977e-01 -1.51909441e-01 9.58235860e-01
4.62518856e-02 -3.62041503e-01 -9.22132969e-01 3.56742680e-01
3.88136238e-01 1.01051497e+00 -2.74890721e-01 -3.32124919e-01
-3.85751963e-01 1.17828095e+00 1.29578337e-01 3.26157063e-01
-7.87566721e-01 -2.92203784e-01 9.30380166e-01 3.62789720e-01
5.93811311e-02 -3.42085361e-01 -4.95850056e-01 -1.42590082e+00
3.83437008e-01 -6.96167767e-01 -3.84092391e-01 -1.25401890e+00
-1.13045359e+00 7.44745433e-01 -2.65737772e-01 -1.58912766e+00
-2.67742306e-01 -6.41621947e-01 -1.24245420e-01 8.95293772e-01
-1.94458818e+00 -9.02918220e-01 -1.26384568e+00 7.66729414e-01
5.78284204e-01 1.23120099e-01 4.19864029e-01 6.45447969e-01
-4.11916703e-01 2.47392431e-01 1.09015316e-01 -2.16020212e-01
5.73829651e-01 -1.04710209e+00 5.48347890e-01 6.37854040e-01
-3.19764167e-01 3.61729078e-02 5.01348674e-01 -4.20820892e-01
-1.55427063e+00 -1.50322652e+00 4.35270339e-01 -2.16794372e-01
1.32512495e-01 -4.02300745e-01 -8.51999402e-01 3.67043614e-01
-4.30818528e-01 3.58983278e-01 -2.53952235e-01 -5.21912992e-01
-2.05929667e-01 -5.45838118e-01 -1.33141410e+00 4.83899236e-01
1.28171349e+00 -7.28007972e-01 9.79773793e-03 1.94701537e-01
7.76770055e-01 -8.74838769e-01 -6.82494521e-01 7.05267608e-01
4.70074773e-01 -1.67575240e+00 1.08123124e+00 3.54327440e-01
1.15399468e+00 -4.29911077e-01 -5.53544462e-01 -1.12721491e+00
-6.82790130e-02 5.42920269e-02 1.43926501e-01 8.77491891e-01
8.01203698e-02 -7.82419026e-01 9.71772134e-01 3.26704741e-01
-4.45374280e-01 -8.49728644e-01 -9.76158321e-01 -6.47324383e-01
-5.34087680e-02 -6.80432379e-01 8.28921616e-01 8.92682016e-01
-8.24851096e-01 1.83647797e-01 -3.17059338e-01 1.40743077e-01
9.26955402e-01 2.85254478e-01 1.09547007e+00 -1.06326735e+00
-1.20482281e-01 -1.33841574e-01 -7.80511439e-01 -1.47671819e+00
-1.33796215e-01 -2.25456357e-01 1.98983312e-01 -1.75628698e+00
-1.37198478e-01 -8.02474201e-01 3.01816612e-01 2.11751193e-01
1.69913203e-01 5.13393939e-01 -4.40389633e-01 1.43077731e-01
-1.23472221e-01 9.64705586e-01 1.65438128e+00 -3.14868182e-01
8.41484517e-02 -2.32368395e-01 -1.65037528e-01 6.16851509e-01
5.32041430e-01 -6.82861134e-02 -7.33812988e-01 -8.53363514e-01
3.01535130e-01 3.62790823e-01 6.19858861e-01 -1.39208126e+00
-2.97380257e-02 -2.41724461e-01 6.38749242e-01 -8.39532435e-01
8.31618309e-01 -7.35785604e-01 3.34944487e-01 2.54985899e-01
1.96040615e-01 -1.83529392e-01 6.29815459e-02 3.99442345e-01
-6.29719675e-01 3.27985793e-01 9.76713002e-01 -2.39457548e-01
-1.12265658e+00 9.46321666e-01 3.03046376e-01 6.82884231e-02
8.15935671e-01 -5.71178198e-01 -3.41008872e-01 -3.95976305e-01
-9.96352360e-02 -5.35203330e-03 1.06070101e+00 2.59557664e-01
9.59720969e-01 -1.47715223e+00 -6.26947880e-01 5.70730865e-01
5.14068365e-01 1.03066528e+00 6.58967555e-01 3.84983778e-01
-1.11675167e+00 2.70782053e-01 -5.66721082e-01 -9.05304670e-01
-5.73892772e-01 2.25649312e-01 5.76936781e-01 1.70758501e-01
-9.31262910e-01 9.20344651e-01 5.75768232e-01 -3.85114282e-01
-8.73595625e-02 -8.36510181e-01 1.33837566e-01 -3.69099438e-01
3.72689873e-01 1.23293988e-01 2.54232734e-01 -4.34453994e-01
-1.80501014e-01 1.03110445e+00 2.83982366e-01 -1.14629768e-01
1.20350289e+00 -2.65167385e-01 8.18624347e-02 4.06025887e-01
1.14148426e+00 -2.77269721e-01 -1.73495519e+00 -3.08918506e-01
-7.06621408e-01 -1.14745378e+00 3.83271039e-01 -4.37628925e-01
-1.46604919e+00 1.16325319e+00 6.70058846e-01 -2.96514541e-01
1.38164234e+00 -3.42995822e-01 1.10780025e+00 2.85715073e-01
1.13106096e+00 -7.79949486e-01 9.53768864e-02 6.19335949e-01
8.61604452e-01 -1.65459037e+00 -1.37988091e-01 -5.06468415e-01
-3.72738421e-01 9.55002010e-01 9.73451912e-01 -9.90388542e-02
3.44890922e-01 5.43582737e-02 3.20527516e-02 -1.31805196e-01
-6.90122247e-01 -2.78259832e-02 -6.01818822e-02 7.41972804e-01
-3.16516422e-02 2.80076563e-02 2.35250786e-01 8.16593021e-02
-5.19476593e-01 9.93387178e-02 5.84984004e-01 5.46574652e-01
-3.23766142e-01 -7.67081261e-01 -2.93974549e-01 4.17389303e-01
1.26701429e-01 -1.46178782e-01 1.10500358e-01 5.51359296e-01
2.91260779e-01 8.17660093e-01 2.97832787e-01 -6.87583923e-01
3.72830957e-01 -6.37576997e-01 6.96777999e-01 -3.94556254e-01
3.40951383e-01 -3.43325973e-01 3.16609964e-02 -8.76269519e-01
-2.87454367e-01 -5.27077556e-01 -1.10230672e+00 -7.33408451e-01
1.21472612e-01 -4.61783737e-01 7.02187836e-01 7.61635244e-01
1.88138142e-01 5.01178741e-01 8.77196372e-01 -1.01563954e+00
9.47597548e-02 -9.61894989e-01 -7.18458354e-01 2.69748122e-01
5.26524246e-01 -9.42516387e-01 -4.05273825e-01 -2.44183972e-01] | [8.93803596496582, -2.5309174060821533] |
6566d850-0edd-461a-8f40-672ed04cfe63 | iterative-self-transfer-learning-a-general | 2306.08700 | null | https://arxiv.org/abs/2306.08700v1 | https://arxiv.org/pdf/2306.08700v1.pdf | Iterative self-transfer learning: A general methodology for response time-history prediction based on small dataset | There are numerous advantages of deep neural network surrogate modeling for response time-history prediction. However, due to the high cost of refined numerical simulations and actual experiments, the lack of data has become an unavoidable bottleneck in practical applications. An iterative self-transfer learningmethod for training neural networks based on small datasets is proposed in this study. A new mapping-based transfer learning network, named as deep adaptation network with three branches for regression (DAN-TR), is proposed. A general iterative network training strategy is developed by coupling DAN-TR and the pseudo-label strategy, and the establishment of corresponding datasets is also discussed. Finally, a complex component is selected as a case study. The results show that the proposed method can improve the model performance by near an order of magnitude on small datasets without the need of external labeled samples,well behaved pre-trainedmodels, additional artificial labeling, and complex physical/mathematical analysis. | ['Yuli Huang', 'Yifan Fei', 'Xinzheng Lu', 'Yongjia Xu'] | 2023-06-14 | null | null | null | null | ['pseudo-label'] | ['miscellaneous'] | [-1.47079732e-02 -2.88547009e-01 -3.95458072e-01 -2.70341873e-01
-6.65109754e-01 2.02234790e-01 3.45956296e-01 -1.89830706e-01
-2.10435227e-01 1.15767181e+00 -5.08975565e-01 -5.24511755e-01
-3.22301239e-01 -7.09934473e-01 -6.85646117e-01 -1.09308493e+00
7.94010609e-02 4.78880376e-01 8.69955346e-02 -1.21125452e-01
5.47023043e-02 8.10948431e-01 -1.11925077e+00 3.89430597e-02
8.95093560e-01 1.45732284e+00 -8.65075141e-02 -9.01539475e-02
1.64280638e-01 9.94143784e-01 -2.75755137e-01 4.40800905e-01
1.02068589e-03 -6.43814147e-01 -6.48367524e-01 4.70898561e-02
-5.97027779e-01 -2.93975443e-01 -2.88076907e-01 5.57773471e-01
6.88065469e-01 3.94132406e-01 9.14392471e-01 -1.06770134e+00
-3.79392743e-01 3.92747372e-01 -4.44167346e-01 -7.67155141e-02
-2.96468139e-01 2.78518975e-01 1.53315201e-01 -1.05881131e+00
1.46485746e-01 5.77485621e-01 9.43381250e-01 3.44477445e-01
-1.10566318e+00 -9.84288037e-01 -3.66463333e-01 4.59911823e-01
-1.31511855e+00 -2.59432793e-01 1.28863847e+00 -4.21450764e-01
8.28416646e-01 2.02265778e-03 1.95347175e-01 1.05958343e+00
2.53324449e-01 2.68767089e-01 1.54977047e+00 -4.37766045e-01
4.42471027e-01 4.22279149e-01 1.35970175e-01 2.93915600e-01
-3.47034819e-02 6.50372922e-01 6.61040545e-02 1.09581240e-01
9.50438738e-01 -1.23890445e-01 -3.17856409e-02 -2.89199293e-01
-5.97435415e-01 6.70992434e-01 4.24133003e-01 3.18659335e-01
-3.61365914e-01 -1.73145890e-01 6.95174098e-01 3.28839213e-01
7.29804516e-01 4.04436499e-01 -6.59404159e-01 1.00978889e-01
-7.73688793e-01 -1.94564372e-01 5.07064641e-01 6.24181926e-01
7.65106678e-01 8.76944482e-01 2.41662338e-01 9.76033568e-01
-6.14521503e-02 4.37122434e-01 6.06331170e-01 -7.02036619e-01
1.46736592e-01 4.43796128e-01 4.38429475e-01 -7.50566363e-01
-6.42966509e-01 -7.34756947e-01 -1.28427637e+00 3.13429266e-01
9.43007469e-02 -3.83109599e-01 -8.95617664e-01 1.41010368e+00
4.28630412e-01 1.87054560e-01 -1.84841622e-02 9.68132973e-01
7.42246568e-01 9.35908020e-01 8.19850340e-02 -5.85058451e-01
6.21271431e-01 -9.68293190e-01 -6.31677330e-01 1.95373580e-01
6.14526629e-01 -4.30686623e-01 1.15674841e+00 3.61736327e-01
-9.36675191e-01 -9.50065732e-01 -1.15265942e+00 4.36622590e-01
-2.16397330e-01 3.95779938e-01 6.26613200e-01 3.22999746e-01
-5.89544296e-01 1.04177558e+00 -7.26939976e-01 -2.17958868e-01
4.57190201e-02 4.36677545e-01 -9.65845138e-02 1.16239302e-01
-1.50671577e+00 9.76937413e-01 4.76274818e-01 4.05346066e-01
-9.80884731e-01 -6.37439191e-01 -4.49608535e-01 -1.23045161e-01
2.75114384e-02 -2.73789644e-01 1.28164327e+00 -9.19849277e-01
-2.22046566e+00 5.32929957e-01 3.10335726e-01 -7.07184672e-02
4.38429981e-01 -1.05565578e-01 -8.73750627e-01 3.86412814e-02
-3.73757988e-01 2.89457906e-02 8.00265372e-01 -1.23935068e+00
-3.90438735e-02 3.28934528e-02 -1.30426541e-01 -2.35815689e-01
-4.04460818e-01 1.36174649e-01 1.42772228e-01 -6.77820027e-01
8.33639428e-02 -7.16329455e-01 -2.55155534e-01 -3.09244603e-01
-1.69588715e-01 -3.04498315e-01 7.86728621e-01 -7.27490008e-01
1.28479505e+00 -1.84295607e+00 -1.37219429e-01 3.09021801e-01
-5.50569675e-04 5.75823367e-01 -6.43006340e-02 6.99557245e-01
-6.22732341e-01 -2.21498057e-01 -3.75890404e-01 2.26573840e-01
-2.76544958e-01 -1.66673467e-01 -2.60050446e-01 4.93293464e-01
3.67378086e-01 7.02486575e-01 -7.56588697e-01 -2.88704246e-01
3.95510823e-01 3.31639707e-01 8.00367147e-02 4.74990368e-01
-1.91506162e-01 8.73063207e-01 -4.95907992e-01 6.62991166e-01
7.68759191e-01 -7.08449006e-01 2.34852899e-02 -4.27942127e-01
-2.35422209e-01 6.12305626e-02 -9.41898346e-01 1.06829643e+00
-6.86617255e-01 3.11588883e-01 -2.49186661e-02 -1.65508354e+00
1.23897111e+00 5.16397178e-01 8.10197175e-01 -1.02143538e+00
5.54675758e-01 5.37320733e-01 -2.49231011e-02 -7.10925758e-01
-7.90174454e-02 -4.12062794e-01 -7.20824972e-02 6.09776080e-01
-2.22210586e-01 1.95758387e-01 -2.30382994e-01 -5.25438130e-01
6.21049941e-01 4.09965962e-01 6.46873415e-02 -4.63840306e-01
9.02948201e-01 9.61075723e-02 6.16738737e-01 1.82129607e-01
4.94781649e-03 1.10987432e-01 -5.37972115e-02 -8.41717660e-01
-1.23613656e+00 -6.15309000e-01 -3.43495667e-01 6.71522677e-01
1.73330098e-01 3.01903903e-01 -6.51869416e-01 -2.67373938e-02
-3.00416231e-01 5.66002011e-01 -4.37606335e-01 -6.67261600e-01
-8.26792657e-01 -1.11322534e+00 4.39653933e-01 7.96990871e-01
8.01588356e-01 -1.18612731e+00 -2.44979247e-01 4.77431238e-01
1.25704989e-01 -1.04673088e+00 1.68226883e-01 3.35406631e-01
-1.07545722e+00 -9.78767574e-01 -5.42537153e-01 -1.05593193e+00
5.11952877e-01 -1.44913599e-01 7.47602582e-01 1.34900481e-01
-7.62377232e-02 -1.80101961e-01 -1.77216142e-01 -1.88213497e-01
-7.47804582e-01 9.40386355e-02 3.29191566e-01 1.47703039e-02
3.38187933e-01 -8.38684142e-01 -5.37873089e-01 6.82936788e-01
-4.80228782e-01 2.58843333e-01 7.33504534e-01 1.12006533e+00
5.43748796e-01 1.69648677e-01 1.33861935e+00 -6.92633331e-01
5.51138341e-01 -5.93175173e-01 -8.17767322e-01 2.74694949e-01
-1.24107170e+00 3.52151948e-03 1.26570654e+00 -7.22654641e-01
-1.32428968e+00 -2.20474690e-01 9.23561677e-02 -4.25304502e-01
-2.38407150e-01 8.02300990e-01 -1.33773342e-01 -3.09832901e-01
7.81668603e-01 2.52147436e-01 2.05672726e-01 -5.68046927e-01
-3.45596187e-02 8.50828946e-01 2.57929146e-01 -8.96527290e-01
6.94774270e-01 3.85224298e-02 3.15773487e-01 -7.02290118e-01
-6.33050561e-01 2.35581305e-02 -4.81002480e-01 -4.64754850e-01
4.69605237e-01 -6.96748137e-01 -6.42735422e-01 1.03862369e+00
-8.87652814e-01 -8.28838885e-01 -1.29169643e-01 7.73282528e-01
-5.57036519e-01 1.12565309e-01 -1.13255990e+00 -6.90738559e-01
-5.51903963e-01 -9.16256666e-01 3.03716883e-02 1.42941445e-01
8.71501490e-02 -1.09190750e+00 -2.04735637e-01 6.05918877e-02
7.67621696e-01 4.35507983e-01 1.19790864e+00 -5.84054589e-01
-5.22874653e-01 -4.04569894e-01 -3.30166936e-01 6.07004106e-01
2.33600229e-01 -8.19594264e-02 -1.07497299e+00 -4.31565344e-01
5.35489082e-01 -7.37648308e-01 4.03177142e-01 3.09146285e-01
1.53908539e+00 -1.74957085e-02 -3.08798194e-01 5.06528080e-01
1.48285496e+00 7.06720710e-01 6.60916090e-01 2.66993761e-01
5.99225283e-01 4.57463384e-01 4.85515237e-01 4.42825228e-01
-1.31982654e-01 3.31888288e-01 1.99026447e-02 -5.23788333e-01
5.44933975e-02 1.79597791e-02 2.11282924e-01 1.70116270e+00
-3.34284097e-01 -1.73872754e-01 -9.85883415e-01 1.43352717e-01
-1.49006629e+00 -4.87158954e-01 -2.14175254e-01 2.33000493e+00
7.93481946e-01 2.33084977e-01 -1.16341792e-01 3.86142701e-01
8.80028546e-01 -3.97133797e-01 -9.79988456e-01 -3.31799597e-01
-3.77487279e-02 3.83779734e-01 2.79724896e-01 2.62936175e-01
-9.74005044e-01 6.17775261e-01 6.47719812e+00 1.05994749e+00
-1.86278844e+00 1.65222064e-01 9.26329255e-01 4.26489979e-01
1.11499712e-01 -1.00110173e-01 -4.97817099e-01 3.21059406e-01
1.38414323e+00 -1.81735724e-01 4.35896188e-01 7.07098842e-01
4.73477602e-01 1.30016506e-01 -9.18832958e-01 7.82618880e-01
-1.50242731e-01 -1.09892702e+00 -2.77382791e-01 -3.62749726e-01
6.66566014e-01 -2.24979684e-01 -6.67452440e-03 5.70545912e-01
-1.42696965e-03 -7.89366186e-01 2.04471484e-01 5.82257569e-01
1.23436904e+00 -6.70016348e-01 7.45609701e-01 2.91605055e-01
-1.01268625e+00 -2.06770748e-01 -2.54079580e-01 -2.16103151e-01
5.87757453e-02 5.00639617e-01 -4.76789951e-01 7.59545624e-01
4.83869374e-01 7.10644066e-01 -1.11320034e-01 1.02135563e+00
-4.59292084e-02 1.12123692e+00 -2.33139604e-01 -2.11503133e-01
2.48990543e-02 -4.30464387e-01 7.11994246e-02 5.69331586e-01
4.12818223e-01 2.13819742e-01 2.21393302e-01 9.39642429e-01
1.57716051e-01 7.68143013e-02 -1.55951068e-01 -4.99362089e-02
3.21878642e-01 1.22477424e+00 -7.36641943e-01 -3.01255584e-01
-2.36391544e-01 3.29039782e-01 2.64774058e-02 7.37520456e-01
-1.29611731e+00 -5.00794470e-01 -1.09043278e-01 2.66332537e-01
7.67532215e-02 -8.24416801e-02 -2.88670450e-01 -7.06468403e-01
-1.82077825e-01 -6.20003939e-01 7.34861642e-02 -8.35286617e-01
-1.45260096e+00 6.59374297e-01 3.04320961e-01 -1.60666144e+00
-3.87171298e-01 -8.30316484e-01 -6.50961637e-01 9.11569715e-01
-1.63788760e+00 -8.94274056e-01 -5.25987864e-01 2.05396906e-01
2.83443838e-01 -4.40483779e-01 9.63115811e-01 7.16985285e-01
-1.06573796e+00 6.80185854e-01 7.16733634e-01 -1.27703413e-01
5.43156326e-01 -5.19403458e-01 -5.66414408e-02 3.22786003e-01
-9.86196399e-01 2.62070835e-01 5.80562472e-01 -4.75230813e-01
-1.09414446e+00 -1.16172326e+00 3.36361080e-01 2.04086021e-01
7.85296381e-01 -9.80292410e-02 -1.36842418e+00 2.20277831e-01
1.73835918e-01 -2.15375815e-02 5.53563356e-01 -1.87245294e-01
-2.93906406e-02 -5.89925051e-01 -9.32101786e-01 1.85598791e-01
6.10534072e-01 -6.13841236e-01 -2.59381294e-01 5.29084325e-01
6.71296060e-01 -2.59423167e-01 -1.18966925e+00 8.95986319e-01
3.32441717e-01 -5.59886873e-01 7.77079642e-01 -7.94680655e-01
5.37453175e-01 -2.21266314e-01 1.02070428e-01 -1.14831066e+00
-4.66642171e-01 -5.34964144e-01 -2.60708094e-01 1.11495674e+00
3.22080761e-01 -7.79249072e-01 8.83511603e-01 7.83649147e-01
-4.53448623e-01 -9.34509099e-01 -6.88237786e-01 -1.21144903e+00
4.22522068e-01 -6.16890080e-02 5.41274607e-01 1.14688349e+00
-1.37755290e-01 5.07041991e-01 -4.46718425e-01 -4.65020202e-02
4.49258000e-01 3.13838571e-01 4.53869253e-01 -1.21372843e+00
-1.86368436e-01 -1.54873937e-01 7.50733614e-02 -8.23994696e-01
2.77594507e-01 -5.86761296e-01 -8.75039957e-03 -1.17189002e+00
-1.93916425e-01 -8.25734973e-01 -9.12141383e-01 2.77791679e-01
3.33042055e-01 -1.05090514e-01 -6.23930573e-01 2.27657318e-01
-1.82409242e-01 1.05058300e+00 1.40494609e+00 -1.00579530e-01
-6.68743774e-02 1.01663314e-01 4.02385592e-02 4.46422845e-01
1.19044530e+00 -4.33842540e-01 -6.64654553e-01 -5.24349734e-02
-1.91064149e-01 4.46976691e-01 3.09558630e-01 -1.29375660e+00
1.17727920e-01 -3.46022785e-01 4.09353375e-01 -5.22367597e-01
3.66134554e-01 -9.13583696e-01 4.64076042e-01 5.77750146e-01
-2.38634020e-01 6.14743344e-02 3.37381631e-01 2.63850063e-01
-2.22107142e-01 -1.44146249e-01 8.91490757e-01 1.08655304e-01
-5.42940795e-01 2.37342432e-01 -3.40657741e-01 -1.52811229e-01
1.02602518e+00 -3.27588260e-01 -1.65031090e-01 -1.96760535e-01
-5.43283880e-01 -5.62415039e-03 2.58961320e-01 8.70277509e-02
4.60331321e-01 -1.54339278e+00 -4.09973353e-01 2.95244068e-01
-3.27673741e-02 -1.25040621e-01 5.65694809e-01 9.06196892e-01
-4.32002187e-01 5.07213473e-01 -4.44403410e-01 -4.05434519e-01
-4.50146198e-01 8.60208690e-01 6.60519242e-01 -1.82672799e-01
-3.78244162e-01 2.26846993e-01 8.04846436e-02 -6.26621366e-01
-6.63476661e-02 -3.50586236e-01 -1.38050467e-01 -3.66994262e-01
1.99639976e-01 6.30924940e-01 4.24714208e-01 -4.49129432e-01
1.20447157e-02 4.91545886e-01 3.99485976e-02 4.96922463e-01
1.42787147e+00 2.22596809e-01 -2.20835388e-01 6.96310043e-01
1.27039635e+00 -6.29108608e-01 -1.39595139e+00 -3.19232017e-01
-4.57131900e-02 1.30199373e-01 1.17605567e-01 -7.24876881e-01
-1.19328475e+00 1.00671589e+00 7.28320122e-01 -2.21264102e-02
1.33990335e+00 -5.57971537e-01 1.05197954e+00 5.17779291e-01
4.68028188e-01 -1.34088624e+00 3.38419437e-01 4.12769377e-01
8.92510533e-01 -1.11760497e+00 5.80748077e-03 -2.81083941e-01
-1.29523560e-01 1.29460871e+00 1.08189690e+00 -1.46873146e-01
9.65979993e-01 1.20241776e-01 -1.18736140e-02 5.05501665e-02
-5.32488942e-01 5.87833285e-01 4.21520434e-02 2.77710259e-01
3.08152944e-01 -5.61398491e-02 -3.10907662e-01 8.32567096e-01
4.66630012e-01 4.13408905e-01 1.82477832e-01 9.68252838e-01
-9.51408371e-02 -1.05964100e+00 -1.58793151e-01 5.30683339e-01
-7.24851564e-02 1.46942109e-01 2.72605300e-01 1.06147218e+00
-2.02580705e-01 6.60955667e-01 -1.79796755e-01 -8.66473019e-01
2.47374266e-01 2.22542342e-02 2.62483448e-01 -6.34819195e-02
-3.53198260e-01 -4.44165356e-02 4.02144752e-02 -2.72828676e-02
-5.81891894e-01 -1.33279115e-01 -1.42289805e+00 -3.28845829e-01
-6.50260091e-01 2.60594815e-01 3.84280443e-01 9.25428748e-01
4.09614414e-01 4.80669230e-01 1.09232807e+00 -7.97816992e-01
-8.72990072e-01 -1.19310033e+00 -5.45611084e-01 2.74693280e-01
-2.31578778e-02 -1.02425659e+00 -3.96434754e-01 -1.95025317e-02] | [6.5574049949646, 3.2697370052337646] |
e4773412-12c4-49da-a284-052a0c944386 | a-bert-based-dual-embedding-model-for-chinese | 2011.02378 | null | https://arxiv.org/abs/2011.02378v1 | https://arxiv.org/pdf/2011.02378v1.pdf | A BERT-based Dual Embedding Model for Chinese Idiom Prediction | Chinese idioms are special fixed phrases usually derived from ancient stories, whose meanings are oftentimes highly idiomatic and non-compositional. The Chinese idiom prediction task is to select the correct idiom from a set of candidate idioms given a context with a blank. We propose a BERT-based dual embedding model to encode the contextual words as well as to learn dual embeddings of the idioms. Specifically, we first match the embedding of each candidate idiom with the hidden representation corresponding to the blank in the context. We then match the embedding of each candidate idiom with the hidden representations of all the tokens in the context thorough context pooling. We further propose to use two separate idiom embeddings for the two kinds of matching. Experiments on a recently released Chinese idiom cloze test dataset show that our proposed method performs better than the existing state of the art. Ablation experiments also show that both context pooling and dual embedding contribute to the improvement of performance. | ['Jing Jiang', 'Minghuan Tan'] | 2020-11-04 | null | https://aclanthology.org/2020.coling-main.113 | https://aclanthology.org/2020.coling-main.113.pdf | coling-2020-8 | ['cloze-test'] | ['natural-language-processing'] | [ 7.04302862e-02 -1.14908449e-01 -5.44972718e-01 -4.31638718e-01
-5.75939596e-01 -7.66534150e-01 7.48104155e-01 -1.19613759e-01
-2.74816126e-01 2.16333821e-01 8.74533713e-01 -3.97954494e-01
1.72704346e-02 -7.92599022e-01 -1.25510260e-01 -5.71212053e-01
1.27161071e-01 8.01761150e-01 1.93587109e-01 -4.90492791e-01
5.17049015e-01 8.71526718e-04 -1.05300009e+00 1.08030379e+00
6.19201839e-01 6.81921184e-01 3.98898035e-01 -2.71005630e-02
-3.97375017e-01 7.73322880e-01 -4.14830357e-01 -6.14456773e-01
2.49832824e-01 -8.50611210e-01 -9.57428396e-01 -1.64160848e-01
1.06772639e-01 -3.19279991e-02 -2.01174080e-01 9.65112209e-01
3.24089617e-01 -1.43125251e-01 4.06781048e-01 -7.13872612e-01
-1.20630646e+00 1.25332892e+00 -4.84467179e-01 2.05358818e-01
4.09468174e-01 -5.02010524e-01 1.73821294e+00 -1.24875367e+00
1.09914720e+00 9.53210652e-01 4.58527476e-01 7.65192807e-01
-1.01613903e+00 -6.95259392e-01 3.82213593e-01 7.93351889e-01
-1.27655089e+00 -2.09509671e-01 1.33469391e+00 -1.92208216e-01
9.66070175e-01 2.36823127e-01 5.05767584e-01 1.13596284e+00
1.91423133e-01 7.73842812e-01 9.84893322e-01 -5.33301950e-01
-7.62066916e-02 3.56221914e-01 2.51838475e-01 4.95695740e-01
-3.49762529e-01 -1.53408527e-01 -4.83113408e-01 -3.04554462e-01
2.29541957e-01 6.82830885e-02 -1.38256505e-01 -1.37568578e-01
-1.26621664e+00 1.15370107e+00 5.90188801e-01 6.90765083e-01
-1.84119344e-01 -4.74608243e-02 5.23190200e-01 5.11722267e-01
2.44243190e-01 4.34919983e-01 -5.97379625e-01 1.81716010e-01
-5.55445552e-01 5.52076519e-01 7.37199128e-01 8.43912125e-01
5.40805817e-01 -5.39314687e-01 1.53543815e-01 9.11104679e-01
1.08271807e-01 8.72752666e-02 7.36616552e-01 -4.03483957e-01
7.11776912e-01 8.81006420e-01 -9.92653593e-02 -1.16254604e+00
-4.82985415e-02 6.53978586e-02 -3.07286471e-01 -2.14951158e-01
2.38088802e-01 2.10692406e-01 -3.99229378e-01 1.83200681e+00
1.31115615e-01 5.14410250e-02 2.10954487e-01 1.12642336e+00
6.86601102e-01 8.47761095e-01 -3.21118861e-01 1.10419840e-02
1.69947767e+00 -1.08138001e+00 -7.01178789e-01 -3.10345203e-01
4.84761328e-01 -9.93178189e-01 1.19748092e+00 -1.64819565e-02
-9.51621056e-01 -2.64529794e-01 -1.13543987e+00 -4.54732269e-01
-3.58949900e-01 1.31948948e-01 4.98129845e-01 3.29606473e-01
-4.42813545e-01 4.90939260e-01 -7.22290695e-01 1.68031409e-01
2.12982018e-02 7.28586391e-02 -3.05516660e-01 -1.43667623e-01
-1.37566936e+00 1.07691038e+00 6.18687332e-01 -5.13862185e-02
-4.54429239e-01 -5.46932042e-01 -1.00675595e+00 1.43751241e-02
1.83846831e-01 -2.83720106e-01 7.19750106e-01 -1.31582987e+00
-1.40483749e+00 1.24731827e+00 -5.21625817e-01 -2.84331471e-01
1.96872756e-01 -1.47895798e-01 -6.05538785e-01 -7.44269714e-02
3.48417670e-01 4.63637561e-01 6.45136058e-01 -1.09533751e+00
-4.42200571e-01 -2.17202485e-01 1.79151967e-01 2.62264818e-01
-1.68567538e-01 6.49018526e-01 -2.10770771e-01 -8.03055286e-01
8.87297332e-01 -1.06752169e+00 -6.86816573e-02 -3.35240930e-01
-1.86335742e-01 -2.02854276e-01 1.28035140e+00 -8.16401541e-01
1.39332998e+00 -2.46900821e+00 4.99258757e-01 -1.30470023e-01
-6.34058937e-02 -2.64088601e-01 -8.21370557e-02 6.38482213e-01
-2.10047483e-01 1.21718429e-01 -3.56345773e-01 -6.11346513e-02
1.13669038e-01 4.87637281e-01 -7.48857975e-01 2.05086067e-01
3.60823482e-01 1.06103528e+00 -9.39144135e-01 -1.25914469e-01
-1.73498675e-01 2.87711412e-01 -9.37127113e-01 1.02788649e-01
-2.74938315e-01 3.09035331e-01 1.31160254e-02 4.59730238e-01
6.43363893e-01 -5.52034415e-02 1.03700173e+00 -6.80955574e-02
-1.27042457e-01 1.25215638e+00 -1.07852542e+00 1.64940584e+00
-6.64843798e-01 5.07549703e-01 -3.92501563e-01 -7.50043213e-01
9.17689979e-01 4.96781439e-01 -1.95464417e-02 -6.11321390e-01
-7.15409741e-02 5.39061964e-01 5.65473080e-01 -2.76424199e-01
5.73847175e-01 -7.16339290e-01 -5.59988439e-01 8.17783177e-01
-1.31663188e-01 1.41137809e-01 -1.66421846e-01 7.50677893e-03
7.93077469e-01 -1.02330139e-02 5.45236051e-01 -5.81233919e-01
7.89720774e-01 -2.11203679e-01 9.23498690e-01 1.40967980e-01
1.66094646e-01 4.81346369e-01 7.24510968e-01 -9.68368232e-01
-9.88959491e-01 -1.44075203e+00 -2.57817060e-01 1.09435475e+00
4.03875023e-01 -7.36744940e-01 -3.90062481e-01 -9.90815163e-01
-3.09263021e-01 6.73400939e-01 -7.30517805e-01 2.23997355e-01
-1.38811398e+00 -7.65334070e-01 1.76476881e-01 7.65436232e-01
3.07387620e-01 -1.47311878e+00 -7.80154884e-01 2.96663433e-01
-3.73735189e-01 -1.07085693e+00 -5.96996248e-01 2.65253335e-01
-5.59046328e-01 -9.89944279e-01 -3.47286765e-03 -1.46727514e+00
7.22115815e-01 7.37175941e-02 9.28395927e-01 3.26407880e-01
1.64682984e-01 -7.79873192e-01 -6.90789700e-01 1.36106953e-01
-3.15024167e-01 4.48761471e-02 -2.27102190e-02 -3.45343798e-02
7.11046755e-01 -6.61830962e-01 -2.78221786e-01 1.45820007e-01
-8.53336155e-01 3.34711283e-01 1.86698839e-01 1.43479252e+00
4.52536583e-01 -1.06022418e-01 4.85221148e-01 -1.28424060e+00
5.08146584e-01 -6.03783429e-01 -4.06409770e-01 -1.87224336e-02
-3.62805158e-01 2.34959468e-01 8.44972908e-01 -6.51463330e-01
-9.83187139e-01 -2.37342790e-01 -2.45182022e-01 2.27581993e-01
7.98041224e-02 7.81825066e-01 -4.92323875e-01 5.20028055e-01
4.08146858e-01 6.22453205e-02 -4.44107533e-01 -3.91326666e-01
6.46829784e-01 6.38740957e-01 3.56785685e-01 -8.32631409e-01
5.75409412e-01 4.42770243e-01 -4.10496920e-01 -3.87367487e-01
-7.90098429e-01 -2.33869836e-01 -7.91185379e-01 6.25796139e-01
9.07665014e-01 -6.55301273e-01 -7.11619556e-02 -6.17883801e-02
-1.52741814e+00 -9.11911428e-02 7.74230137e-02 1.82071790e-01
-4.74954456e-01 3.68470758e-01 -1.02441537e+00 -4.12711024e-01
-2.05080986e-01 -1.32215941e+00 8.08354199e-01 -1.84212610e-01
-5.92956066e-01 -1.10449100e+00 2.95310408e-01 6.60982653e-02
2.20003918e-01 1.49584070e-01 1.78176558e+00 -7.75344193e-01
-4.94704634e-01 2.90092397e-02 -3.55816007e-01 5.10125421e-02
4.23917264e-01 -4.78137016e-01 -9.99254763e-01 -1.07082807e-01
2.58615315e-01 3.43754217e-02 9.12407100e-01 -1.61359429e-01
5.69062769e-01 -7.82545865e-01 -2.03487694e-01 6.52909935e-01
1.67289162e+00 4.29335117e-01 5.06586254e-01 4.72641557e-01
4.42331821e-01 6.25763714e-01 2.91433424e-01 4.58001681e-02
4.10559326e-01 9.91740286e-01 9.36949253e-02 4.06621486e-01
5.61628304e-02 -3.92710060e-01 7.12243557e-01 1.30748618e+00
4.54779029e-01 2.62347281e-01 -9.07579601e-01 1.11173546e+00
-1.82116318e+00 -1.22491467e+00 -4.59762029e-02 1.71515179e+00
1.08380091e+00 2.00768754e-01 -2.09758237e-01 2.00725928e-01
4.25760537e-01 4.79335636e-01 2.56024092e-01 -7.59813964e-01
-4.06510502e-01 7.23810732e-01 -1.96111202e-01 8.79027069e-01
-1.10422003e+00 1.34016550e+00 5.71330404e+00 3.43584925e-01
-1.09653103e+00 6.18747473e-01 -4.82807569e-02 1.51042134e-01
-8.81721795e-01 6.05635583e-01 -6.28085434e-01 5.82044363e-01
6.61361873e-01 -6.21041767e-02 5.18282950e-01 8.04555595e-01
-3.43081385e-01 1.95067123e-01 -1.24320757e+00 6.23512089e-01
4.57321435e-01 -1.64661264e+00 2.55939960e-01 -1.66008472e-01
1.04938161e+00 -1.98949769e-01 8.00631046e-02 1.80436119e-01
-1.61627665e-01 -8.04887593e-01 8.57481658e-01 -1.67774275e-01
5.80017269e-01 -8.19293439e-01 9.95259821e-01 9.09546018e-02
-1.39207041e+00 -3.30252498e-01 -4.93154436e-01 -5.24040461e-01
3.83420616e-01 1.66235849e-01 -6.64147973e-01 5.83485186e-01
2.69885331e-01 7.46470749e-01 -2.40746498e-01 2.92109609e-01
-8.91342461e-01 3.00712198e-01 1.22034281e-01 5.91308773e-02
4.73493487e-01 -2.95579344e-01 5.75195849e-01 1.39475334e+00
1.04409039e-01 8.36953148e-03 2.03381002e-01 1.24211466e+00
5.72918840e-02 2.93015122e-01 -6.84103787e-01 2.00110018e-01
5.74790657e-01 9.85111713e-01 -7.28491783e-01 -3.33979487e-01
-5.34538507e-01 1.52067947e+00 3.84280592e-01 1.04411818e-01
-7.85002768e-01 -3.91637951e-01 1.04273474e+00 -5.96606433e-02
5.56521356e-01 -1.77120283e-01 -7.12721765e-01 -1.20650494e+00
2.89257139e-01 -7.33302116e-01 5.11410534e-01 -2.70016879e-01
-1.41795158e+00 7.38630116e-01 -3.01940143e-02 -1.19144642e+00
-3.08301330e-01 -8.52104366e-01 -1.02031410e+00 1.00244296e+00
-1.42628229e+00 -1.45493567e+00 3.17706108e-01 1.30511016e-01
7.66496360e-01 -1.96875438e-01 1.28590000e+00 2.74235457e-01
-2.75718659e-01 5.72479665e-01 -3.26049566e-01 5.77295840e-01
3.52638394e-01 -1.20735276e+00 6.73215866e-01 9.30480003e-01
5.62840700e-01 1.01466167e+00 4.01115507e-01 -4.89334106e-01
-8.35878372e-01 -6.43979728e-01 1.71213818e+00 -8.10644209e-01
8.82775009e-01 -6.49475515e-01 -1.04140770e+00 9.45997894e-01
5.56486666e-01 -3.44452292e-01 9.33045447e-01 2.92635888e-01
-9.63245690e-01 1.22948386e-01 -7.89278269e-01 8.53784263e-01
9.19915497e-01 -1.03209829e+00 -1.44936085e+00 -4.84271860e-03
8.23531866e-01 -4.29035783e-01 -3.24341476e-01 3.09716523e-01
7.76978076e-01 -7.46234894e-01 7.01535702e-01 -8.79504383e-01
9.11463678e-01 -4.51737165e-01 -6.76275790e-01 -1.17377341e+00
-3.37239653e-01 -3.07384968e-01 2.68355012e-03 1.11566269e+00
7.52389729e-01 -2.91524202e-01 7.10916519e-01 -5.92173152e-02
-1.20223016e-01 -8.04427445e-01 -1.09765434e+00 -6.54532611e-01
2.95419604e-01 -2.92450547e-01 7.04032362e-01 1.53832877e+00
6.78650677e-01 6.34043753e-01 -2.83186108e-01 1.01757504e-01
1.76474396e-02 8.77985299e-01 3.46939236e-01 -8.63400161e-01
-4.90862489e-01 -6.22064173e-01 -4.33273554e-01 -9.53965485e-01
6.31699920e-01 -1.58362675e+00 -5.92350513e-02 -9.68422413e-01
3.75324726e-01 -5.13644516e-01 -4.36233401e-01 3.20900917e-01
-3.28794509e-01 2.01457456e-01 2.87435263e-01 5.94588034e-02
-9.28738490e-02 3.30628484e-01 6.60443246e-01 -6.32955357e-02
-3.29544455e-01 -4.35768336e-01 -6.78746045e-01 6.54657841e-01
8.02693367e-01 -7.97113359e-01 -2.53961891e-01 -6.96926415e-01
4.37443882e-01 -2.71942943e-01 1.28866047e-01 -3.68549973e-01
-1.25172129e-02 -4.15850252e-01 3.70288342e-02 -5.17284572e-01
2.45337129e-01 -8.92875135e-01 1.90527678e-01 5.09130001e-01
-2.92533129e-01 7.01516390e-01 5.08489311e-02 -4.61701788e-02
-2.39192173e-01 -3.29210043e-01 7.52073765e-01 -2.59630442e-01
-8.20578039e-01 -2.93877095e-01 -3.82035822e-02 -1.45127811e-02
7.46537209e-01 9.15132836e-02 -2.63196021e-01 -6.48126751e-02
-4.54286724e-01 -1.51409015e-01 5.23312747e-01 7.63718784e-01
1.06680870e+00 -1.72498024e+00 -6.38806522e-01 7.01076329e-01
2.87131429e-01 -4.77760226e-01 -1.48710668e-01 9.58056152e-01
-5.13926744e-01 1.59110367e-01 -3.74017477e-01 -9.44649503e-02
-1.14351022e+00 6.46371186e-01 2.03053996e-01 -3.68505031e-01
-8.14907253e-01 8.08685005e-01 2.83972532e-01 -7.59420812e-01
-2.74280161e-01 -4.05218929e-01 1.33537306e-02 6.54635951e-02
8.62248123e-01 -1.37767956e-01 -1.86978623e-01 -9.56412971e-01
-6.81556046e-01 5.50824642e-01 -3.20787817e-01 -5.00024855e-01
1.35577166e+00 9.33368579e-02 -8.83454502e-01 5.91220558e-01
1.16471994e+00 7.33889282e-01 -4.71100301e-01 -2.22618863e-01
7.57314622e-01 -5.99784315e-01 -6.54034257e-01 -6.60019934e-01
-7.62302935e-01 8.16564739e-01 9.67960730e-02 -2.25692749e-01
9.48915064e-01 3.16327721e-01 1.02594817e+00 1.15522392e-01
2.30278239e-01 -8.32253456e-01 2.43282989e-01 9.14263308e-01
8.09010983e-01 -8.45755756e-01 -3.75884622e-01 -3.85905266e-01
-8.44895780e-01 1.22797811e+00 5.15807450e-01 -7.22571075e-01
7.06913471e-01 2.62011439e-01 5.39292812e-01 -2.70567566e-01
-9.32858467e-01 -5.71717834e-03 2.73040891e-01 2.68119484e-01
7.84723461e-01 2.18211770e-01 -9.65281427e-01 1.01742351e+00
-5.02083480e-01 -7.36878097e-01 4.14851189e-01 6.03501558e-01
-1.67253435e-01 -1.64903510e+00 6.63546547e-02 -9.09433737e-02
-4.61445302e-01 -4.81254548e-01 -6.76588714e-01 6.27012670e-01
5.39104521e-01 5.36696315e-01 2.16393635e-01 -5.95834672e-01
1.23686254e-01 5.39445937e-01 3.09464902e-01 -9.41252291e-01
-8.38391304e-01 -5.61501384e-02 1.83554459e-02 -3.55466127e-01
-1.23391144e-01 -6.52917087e-01 -1.51422727e+00 -1.75020754e-01
-2.59991020e-01 9.41010118e-02 2.31861129e-01 1.02122974e+00
4.34798189e-02 -2.68050358e-02 7.41664886e-01 -4.32073921e-01
-2.77460575e-01 -8.57799709e-01 -3.03251117e-01 8.34760189e-01
1.68298796e-01 -5.41752219e-01 -1.43964961e-01 -1.87434509e-01] | [10.840168952941895, 9.444941520690918] |
d8c01fc9-e233-4574-ac64-06644ab51169 | caption-anything-interactive-image | 2305.02677 | null | https://arxiv.org/abs/2305.02677v3 | https://arxiv.org/pdf/2305.02677v3.pdf | Caption Anything: Interactive Image Description with Diverse Multimodal Controls | Controllable image captioning is an emerging multimodal topic that aims to describe the image with natural language following human purpose, $\textit{e.g.}$, looking at the specified regions or telling in a particular text style. State-of-the-art methods are trained on annotated pairs of input controls and output captions. However, the scarcity of such well-annotated multimodal data largely limits their usability and scalability for interactive AI systems. Leveraging unimodal instruction-following foundation models is a promising alternative that benefits from broader sources of data. In this paper, we present Caption AnyThing (CAT), a foundation model augmented image captioning framework supporting a wide range of multimodel controls: 1) visual controls, including points, boxes, and trajectories; 2) language controls, such as sentiment, length, language, and factuality. Powered by Segment Anything Model (SAM) and ChatGPT, we unify the visual and language prompts into a modularized framework, enabling the flexible combination between different controls. Extensive case studies demonstrate the user intention alignment capabilities of our framework, shedding light on effective user interaction modeling in vision-language applications. Our code is publicly available at https://github.com/ttengwang/Caption-Anything. | ['Shanshan Zhao', 'Mingqi Gao', 'Zhe Li', 'Yunlong Tang', 'Hao Zheng', 'Junjie Fei', 'Jinrui Zhang', 'Teng Wang'] | 2023-05-04 | null | null | null | null | ['controllable-image-captioning', 'instruction-following'] | ['computer-vision', 'natural-language-processing'] | [ 2.27490515e-01 2.78641939e-01 -5.93792796e-01 -4.87699926e-01
-8.20374072e-01 -9.29633498e-01 7.86863148e-01 1.60268042e-02
-1.38969675e-01 4.90393162e-01 5.71839571e-01 -5.75032651e-01
3.99710834e-01 -1.11112162e-01 -9.78659749e-01 -2.61364043e-01
5.37055850e-01 4.29356188e-01 -1.20339617e-01 -5.82658827e-01
3.37132752e-01 8.16461444e-02 -1.70627427e+00 7.34139860e-01
6.41147435e-01 7.72605956e-01 3.97580236e-01 9.44797456e-01
-3.34910184e-01 1.24582827e+00 -3.32721084e-01 -3.77593607e-01
-2.58568555e-01 -3.31364632e-01 -5.38875699e-01 9.57884863e-02
8.41459990e-01 -4.19757843e-01 -9.22255218e-02 7.49747872e-01
3.36088479e-01 7.28450716e-02 4.92701560e-01 -1.87157881e+00
-1.26520455e+00 4.39358383e-01 -3.13703746e-01 -4.65697050e-01
9.73949790e-01 7.60804653e-01 9.10846770e-01 -6.37421548e-01
7.12632179e-01 1.30434072e+00 3.52276951e-01 1.10996842e+00
-1.24854290e+00 -4.36956853e-01 5.78792810e-01 -1.00825742e-01
-7.75692284e-01 -3.99637312e-01 5.95371127e-01 -7.12707281e-01
7.42374003e-01 6.93598986e-01 5.80941439e-01 1.71417832e+00
-2.60356590e-02 1.31878769e+00 1.08227372e+00 -6.49275780e-01
-3.81458960e-02 5.55199027e-01 1.17226414e-01 7.97129452e-01
-3.46555293e-01 -7.53294080e-02 -6.70829296e-01 1.83904648e-01
7.11867511e-01 -7.51072392e-02 -5.27552605e-01 -7.05620825e-01
-1.70316195e+00 8.71357858e-01 2.84741282e-01 -1.19792201e-01
-1.11432679e-01 5.72280169e-01 3.50392610e-01 -4.97416742e-02
2.01509744e-01 6.75618470e-01 -3.77385467e-01 -4.36898112e-01
-5.15923977e-01 2.44208708e-01 6.44162178e-01 1.69564772e+00
5.16802549e-01 -8.85062963e-02 -8.96518707e-01 5.79767048e-01
5.69012284e-01 8.49083304e-01 2.41389424e-01 -1.00046194e+00
4.57972765e-01 6.97577596e-01 4.38077182e-01 -6.52601063e-01
-2.01653048e-01 1.41944379e-01 -3.25577110e-01 2.93644130e-01
3.49654585e-01 -2.12760225e-01 -1.22024786e+00 1.88804352e+00
1.32245526e-01 -3.27166528e-01 -1.51554504e-02 9.86063659e-01
1.11623979e+00 8.84771407e-01 6.14231348e-01 3.20748895e-01
1.64866495e+00 -1.22757697e+00 -1.00833929e+00 -6.14203930e-01
6.24196112e-01 -8.23275387e-01 1.85537589e+00 8.56495053e-02
-1.05380583e+00 -5.73147774e-01 -6.53437197e-01 -3.89701039e-01
-6.68218195e-01 3.17729384e-01 4.87778991e-01 3.78856480e-01
-1.38527632e+00 -2.96920240e-01 -6.53765678e-01 -5.71466088e-01
6.43313751e-02 2.35580102e-01 -3.19057018e-01 -7.80007914e-02
-8.81956518e-01 8.52871895e-01 3.96447815e-03 -2.44956687e-01
-7.88176537e-01 -7.83630610e-01 -1.29889059e+00 -1.56444669e-01
3.14258009e-01 -7.51576185e-01 1.87190485e+00 -1.12902522e+00
-1.50546813e+00 1.00036049e+00 -2.78531551e-01 -1.02843717e-01
6.79904878e-01 -3.84460688e-01 -5.40542118e-02 1.53827518e-01
1.11409463e-01 1.58012998e+00 6.67737365e-01 -1.68407631e+00
-4.21710223e-01 -9.53029469e-02 3.04530233e-01 4.96312469e-01
-2.64969677e-01 8.49659070e-02 -7.45611668e-01 -2.52754509e-01
-5.99749207e-01 -1.00919139e+00 -6.66275024e-02 1.94083273e-01
-5.81427932e-01 -2.44527355e-01 7.99637318e-01 -4.14477229e-01
1.34581065e+00 -2.04096103e+00 1.36957586e-01 -2.14085817e-01
3.50625254e-02 -5.80462925e-02 -2.65718222e-01 7.25153208e-01
-3.63250524e-02 3.22122782e-01 -8.24193954e-02 -5.51831245e-01
4.49881315e-01 -2.78206216e-03 -3.66961628e-01 5.38870655e-02
2.12494895e-01 1.26459658e+00 -9.66005385e-01 -7.10242093e-01
6.55605435e-01 7.12325096e-01 -5.72338581e-01 6.81348085e-01
-7.51971722e-01 4.82713580e-01 -4.71587777e-01 7.09165514e-01
1.41240835e-01 -3.76336455e-01 -1.55643180e-01 -3.19733530e-01
-4.50103670e-01 -2.03900591e-01 -6.71245813e-01 1.74896884e+00
-6.26790047e-01 9.62679803e-01 2.21393496e-01 -2.33788639e-01
6.98136568e-01 4.42631066e-01 2.59487092e-01 -6.43668473e-01
5.49215749e-02 -1.14301004e-01 -5.40010750e-01 -1.07641673e+00
6.45022213e-01 4.09396023e-01 -4.69617933e-01 2.02810451e-01
-1.05618700e-01 -5.75025737e-01 2.97849745e-01 4.55881357e-01
6.72708631e-01 6.85769677e-01 2.99483091e-01 -7.16760233e-02
1.86183184e-01 4.30915028e-01 -2.53821403e-01 8.44185114e-01
-4.34137106e-01 9.43736136e-01 5.72003007e-01 -2.24761754e-01
-9.58864808e-01 -9.03888166e-01 7.01680109e-02 1.63110280e+00
5.77611625e-02 -1.56654298e-01 -9.02436078e-01 -5.58350563e-01
-1.40700698e-01 1.28280270e+00 -8.27022910e-01 1.35030791e-01
-3.28837603e-01 1.17472187e-01 2.75229067e-01 4.80444759e-01
2.73441941e-01 -1.42124987e+00 -9.55152333e-01 -1.88944668e-01
-3.72844636e-01 -1.19328249e+00 -1.07028449e+00 -1.98321026e-02
-2.02423617e-01 -7.32345581e-01 -9.28083003e-01 -9.35875058e-01
8.05469334e-01 1.66643426e-01 1.46445096e+00 3.02163232e-02
1.71061233e-01 1.40690172e+00 -4.13926572e-01 -5.86979866e-01
-6.05673790e-01 -2.89131343e-01 -3.97153765e-01 -1.77203238e-01
3.11541468e-01 3.14836025e-01 -5.66219509e-01 2.60077387e-01
-9.47813034e-01 8.49997342e-01 4.98623371e-01 7.39369810e-01
3.27736944e-01 -1.29337287e+00 2.23445177e-01 -7.79059112e-01
9.32877719e-01 -4.64487165e-01 -2.98221111e-01 5.74857533e-01
-3.49389553e-01 1.39329787e-02 2.65033245e-01 -8.57759237e-01
-1.01016057e+00 4.52340156e-01 2.01187983e-01 -5.05997419e-01
-7.70175636e-01 3.50354970e-01 -1.29632354e-01 2.46997714e-01
5.47171593e-01 3.14163983e-01 3.25520039e-01 2.46062890e-01
1.00479698e+00 8.36157918e-01 7.65629649e-01 -6.97776556e-01
3.81289035e-01 1.82721183e-01 -5.53731024e-01 -5.91774702e-01
-7.02025890e-01 -4.77984637e-01 -4.36771363e-01 -7.71785021e-01
1.38991988e+00 -9.24037099e-01 -1.16406989e+00 -7.08798617e-02
-1.32044303e+00 -6.82078898e-01 -1.17767654e-01 2.18642175e-01
-9.76497054e-01 -1.45003423e-01 -4.49961960e-01 -9.69705522e-01
-2.42685258e-01 -1.63022923e+00 1.42782032e+00 3.78054827e-01
-6.38725102e-01 -1.06933379e+00 -9.85335931e-02 6.96710348e-01
4.44131225e-01 4.89378273e-01 7.41888940e-01 -6.55441880e-01
-4.88593072e-01 -3.57395053e-01 -2.17132881e-01 -5.52069284e-02
-1.18812069e-01 1.44328371e-01 -8.86110723e-01 1.14105213e-02
-5.83527744e-01 -8.00257683e-01 1.44256696e-01 3.40277284e-01
9.72135246e-01 -5.01413286e-01 -3.66545707e-01 1.31079212e-01
1.15198517e+00 3.76240641e-01 4.82072771e-01 4.10837680e-01
9.03037608e-01 7.44222939e-01 6.09183311e-01 1.71972990e-01
7.96142340e-01 5.86693943e-01 6.62621140e-01 -3.53771329e-01
-1.44007683e-01 -6.65345430e-01 5.51196933e-01 3.46769035e-01
2.83482730e-01 -6.20863378e-01 -1.07580161e+00 6.13026142e-01
-2.09331465e+00 -9.28571999e-01 -1.16706356e-01 1.86082101e+00
8.48220587e-01 -1.90185368e-01 -9.12195742e-02 -6.75538480e-01
6.39543235e-01 1.80041548e-02 -5.92246294e-01 -7.71410108e-01
2.16312870e-01 -3.39504242e-01 3.25557649e-01 4.72679049e-01
-1.04631448e+00 1.06307518e+00 5.78996420e+00 4.06513125e-01
-9.73441601e-01 -7.35590681e-02 8.17223072e-01 1.21660922e-02
-5.85718691e-01 -2.65936643e-01 -6.90495133e-01 3.23291719e-01
8.50015223e-01 9.07774866e-02 4.41175044e-01 8.63172889e-01
4.97205228e-01 2.01753527e-02 -1.41931510e+00 1.00304973e+00
4.08862561e-01 -1.34430122e+00 3.40811014e-01 -3.19979399e-01
6.17800832e-01 -2.12626576e-01 5.32082379e-01 6.28760934e-01
4.63991463e-01 -1.11395836e+00 1.24222565e+00 5.72196782e-01
1.14775920e+00 -1.92301556e-01 3.27846497e-01 1.71523169e-01
-9.29746330e-01 6.03047088e-02 5.50916374e-01 1.69795170e-01
2.82868803e-01 -4.69600856e-01 -8.74925971e-01 6.63419962e-02
6.50517881e-01 5.41763067e-01 -6.03672266e-01 6.01112425e-01
-2.27799132e-01 3.60119969e-01 9.02920514e-02 -5.47503591e-01
4.25605267e-01 -1.12717152e-01 3.54872167e-01 1.49910057e+00
2.09821880e-01 2.50922501e-01 2.59809345e-01 9.54239249e-01
4.88480367e-02 1.47219777e-01 -8.60139430e-01 -2.57960856e-01
3.01232815e-01 1.20427370e+00 -4.19543356e-01 -3.70745182e-01
-6.85110748e-01 1.03701508e+00 2.02085286e-01 6.93695068e-01
-9.72421110e-01 -1.11670844e-01 6.18063390e-01 2.45919138e-01
-1.51462838e-01 -2.55054116e-01 -4.02368784e-01 -9.93193626e-01
-3.12692344e-01 -1.14372981e+00 2.09285706e-01 -1.63803887e+00
-1.02737963e+00 5.75918853e-01 3.80997270e-01 -1.32382047e+00
-4.00241345e-01 -8.38036716e-01 -6.05610192e-01 5.82938790e-01
-1.20983577e+00 -1.63178289e+00 -5.78504264e-01 5.00852942e-01
1.13564491e+00 1.91967636e-01 7.76026070e-01 -1.26873836e-01
-6.36334717e-01 5.11284709e-01 -5.27356803e-01 -1.09249964e-01
1.03009510e+00 -1.46450424e+00 1.84034988e-01 4.20060486e-01
-1.26693279e-01 6.51327908e-01 1.04194760e+00 -4.83829498e-01
-1.65748298e+00 -8.85590374e-01 6.11048043e-01 -1.05220854e+00
6.24451876e-01 -4.88120347e-01 -6.06821716e-01 1.01583076e+00
1.06223059e+00 -4.65303928e-01 6.85945094e-01 -2.52719820e-01
-3.34442943e-01 3.86336267e-01 -1.00225413e+00 1.21207607e+00
6.85952187e-01 -4.27921951e-01 -4.34870809e-01 5.30063868e-01
1.13273096e+00 -7.53442228e-01 -4.80119050e-01 1.63248658e-01
5.41913629e-01 -8.21487665e-01 9.63226318e-01 -7.00948417e-01
8.14080834e-01 -3.88699740e-01 -1.63724467e-01 -1.20893085e+00
-8.92857313e-02 -7.61704862e-01 3.98959639e-03 1.22487807e+00
7.58278728e-01 -2.48768032e-01 4.78433788e-01 1.43099296e+00
-5.31777382e-01 -6.29176021e-01 -4.29100543e-01 -9.87078622e-02
-1.69270515e-01 -5.96400261e-01 4.77666587e-01 9.22625184e-01
3.73383015e-01 3.42727721e-01 -4.01692867e-01 1.28389493e-01
2.79097110e-01 -2.10232988e-01 9.59055245e-01 -6.58124685e-01
2.04215139e-01 -5.56227267e-01 -3.15786898e-02 -1.30846417e+00
2.23172873e-01 -7.28717268e-01 2.65361428e-01 -1.96194041e+00
1.62150785e-01 6.47859499e-02 1.60019413e-01 9.28796232e-01
-1.45613700e-01 1.30315721e-01 5.67951560e-01 9.84137356e-02
-8.13448310e-01 5.21929264e-01 1.57405221e+00 -3.72131854e-01
-3.48704040e-01 -3.16134781e-01 -7.31181264e-01 5.43554246e-01
8.41078758e-01 2.08296791e-01 -5.65874040e-01 -6.96851075e-01
9.69486609e-02 2.55507916e-01 5.51577806e-01 -8.15129697e-01
3.42726201e-01 -5.91585040e-01 2.63080835e-01 -3.05264205e-01
4.81927812e-01 -9.49762940e-01 -1.80089325e-01 1.79459333e-01
-1.02470672e+00 4.89856839e-01 6.70973778e-01 3.80677670e-01
-7.65252933e-02 6.90552145e-02 3.37695479e-01 -2.12011084e-01
-9.08399403e-01 1.07049197e-01 -5.73826790e-01 -1.93039939e-01
1.29514849e+00 -2.32410744e-01 -6.60956085e-01 -9.62031126e-01
-5.69027305e-01 6.91952467e-01 5.88127494e-01 1.02705944e+00
7.61646390e-01 -1.24126875e+00 -5.02300978e-01 -2.40032449e-02
7.63735712e-01 -2.69968927e-01 2.05022812e-01 4.90261883e-01
-4.13098991e-01 5.65732181e-01 -1.54971868e-01 -9.66521502e-01
-1.18368757e+00 7.01017857e-01 1.45594493e-01 2.83160388e-01
-2.91112453e-01 6.14613235e-01 2.95093834e-01 -6.23642087e-01
3.70162159e-01 -4.52490896e-01 -4.12050009e-01 -1.54629499e-01
4.86819863e-01 -1.73545256e-01 -5.46028674e-01 -7.20771432e-01
-5.93605079e-02 4.35238302e-01 1.24796145e-01 -4.53101039e-01
7.15496063e-01 -3.82503569e-01 1.59649104e-01 6.58672333e-01
1.00112176e+00 -2.05740690e-01 -1.54754293e+00 2.85268307e-01
-2.03331426e-01 -1.25509620e-01 -1.84965596e-01 -1.18887103e+00
-4.73040611e-01 8.48167479e-01 7.24846184e-01 1.92582443e-01
8.99132729e-01 1.65604442e-01 2.97172755e-01 1.79405376e-01
7.40666911e-02 -8.07025671e-01 3.78489822e-01 3.59390974e-01
1.37405968e+00 -1.67906034e+00 -5.87420821e-01 -8.31631124e-02
-1.40608156e+00 1.03306949e+00 1.27479565e+00 5.71947515e-01
7.81629086e-02 1.26392797e-01 7.16470122e-01 -3.32248658e-01
-9.79346395e-01 -2.68506259e-01 5.58088958e-01 7.34570146e-01
9.61541057e-01 2.06301942e-01 1.17518164e-01 3.05855095e-01
-1.77669883e-01 -9.34778452e-02 7.32322693e-01 9.76913691e-01
-2.32938796e-01 -6.64927721e-01 -5.77153623e-01 2.37031996e-01
-2.53351033e-02 -3.14244121e-01 -4.49595243e-01 9.98940647e-01
-2.50973016e-01 1.33584297e+00 1.88320041e-01 -1.34016350e-01
6.11541152e-01 2.45422482e-01 1.37024119e-01 -6.79068267e-01
-7.45732367e-01 -1.23930685e-01 1.93689838e-02 -6.28956974e-01
-2.52908498e-01 -4.65792298e-01 -1.19540632e+00 1.31680205e-01
2.05176473e-01 -5.69065623e-02 9.33108747e-01 5.40209174e-01
4.97559994e-01 4.44917381e-01 1.24507055e-01 -1.04041946e+00
-2.50063032e-01 -9.88088429e-01 4.05906200e-01 6.76239014e-01
5.62895894e-01 -3.39345813e-01 -2.56377816e-01 5.23182869e-01] | [10.939449310302734, 1.562026858329773] |
825ea851-0ea1-426a-b486-cd4af537a7a1 | a-survey-on-graph-classification-and-link | 2307.00865 | null | https://arxiv.org/abs/2307.00865v1 | https://arxiv.org/pdf/2307.00865v1.pdf | A Survey on Graph Classification and Link Prediction based on GNN | Traditional convolutional neural networks are limited to handling Euclidean space data, overlooking the vast realm of real-life scenarios represented as graph data, including transportation networks, social networks, and reference networks. The pivotal step in transferring convolutional neural networks to graph data analysis and processing lies in the construction of graph convolutional operators and graph pooling operators. This comprehensive review article delves into the world of graph convolutional neural networks. Firstly, it elaborates on the fundamentals of graph convolutional neural networks. Subsequently, it elucidates the graph neural network models based on attention mechanisms and autoencoders, summarizing their application in node classification, graph classification, and link prediction along with the associated datasets. | ['Quan Wen', 'Juan Chen', 'Xingyu Liu'] | 2023-07-03 | null | null | null | null | ['node-classification', 'link-prediction', 'graph-classification'] | ['graphs', 'graphs', 'graphs'] | [-1.03549756e-01 5.16230464e-01 -2.17619970e-01 -1.70421749e-01
6.36699498e-01 -2.37334073e-01 3.73669803e-01 3.92829567e-01
-1.89870477e-01 3.06839496e-01 4.69025001e-02 -8.55142534e-01
-3.46855760e-01 -1.53675604e+00 -5.02495050e-01 -2.57816255e-01
-7.05652833e-01 2.19379678e-01 -5.05832620e-02 -3.80438507e-01
-1.85475722e-01 6.85707569e-01 -9.06661212e-01 -2.01770421e-02
6.24291718e-01 8.78289878e-01 -4.09875274e-01 8.58915389e-01
-3.65975231e-01 1.15969610e+00 -6.59341156e-01 -1.04464698e+00
-6.88899159e-02 -1.68793634e-01 -6.40289366e-01 3.76334749e-02
1.90719485e-01 -1.68103978e-01 -1.40997195e+00 1.10032439e+00
2.56188124e-01 3.74034405e-01 5.93922019e-01 -1.59697104e+00
-1.69559467e+00 6.87453210e-01 1.39057189e-01 3.81975085e-01
2.87551999e-01 -1.72487423e-01 1.31651711e+00 -6.21907949e-01
4.67177212e-01 1.08589053e+00 1.02920163e+00 1.41518772e-01
-8.22791696e-01 -2.66194612e-01 1.92322522e-01 4.55892473e-01
-1.43805826e+00 7.52195641e-02 9.66043949e-01 -3.86504859e-01
1.32389331e+00 1.38282878e-02 1.10616565e+00 7.95479059e-01
3.04669946e-01 6.99427962e-01 1.02091096e-01 -1.92962795e-01
-1.07688390e-01 -5.27429283e-01 3.75950634e-01 1.05356610e+00
4.46105063e-01 1.04877455e-02 -1.53019637e-01 2.73486882e-01
9.42393064e-01 3.06462914e-01 3.93258408e-02 -2.52591312e-01
-9.75931883e-01 9.58160639e-01 1.24564230e+00 3.29173654e-01
-5.48386216e-01 5.05824208e-01 5.59604824e-01 5.57052135e-01
5.39050698e-01 2.64283568e-01 -1.20845484e-03 5.33096373e-01
-3.32978010e-01 -9.18924809e-02 5.27121723e-01 1.17024100e+00
5.09804368e-01 6.03381038e-01 -1.08383566e-01 6.78502679e-01
1.98390767e-01 2.48993918e-01 -1.71308354e-01 -3.01090539e-01
4.00357097e-01 1.04937541e+00 -7.04080880e-01 -1.58275497e+00
-6.77059054e-01 -4.05784160e-01 -1.17681968e+00 -2.22022325e-01
-9.39468816e-02 -2.64798433e-01 -7.18569756e-01 1.18337440e+00
-1.52053699e-01 2.13496342e-01 -2.81169955e-02 5.31165600e-01
1.79215407e+00 3.34273189e-01 2.20199242e-01 3.46456826e-01
1.13402212e+00 -9.87709403e-01 -8.23116720e-01 -9.89184603e-02
6.11024916e-01 -3.75040956e-02 4.97403502e-01 -3.65255922e-01
-9.38811004e-01 -6.10169351e-01 -1.00315785e+00 -3.41054678e-01
-1.21706462e+00 -8.98601394e-03 1.44998300e+00 5.57345688e-01
-1.31714749e+00 6.87232733e-01 -6.70608521e-01 -7.85948813e-01
8.44531834e-01 4.23357636e-01 -6.69167578e-01 -4.84353378e-02
-1.53909886e+00 7.25549817e-01 5.00135303e-01 6.87485576e-01
-4.01379675e-01 -2.87095070e-01 -1.43017375e+00 4.29824471e-01
5.47560565e-02 -8.62449229e-01 6.23937249e-01 -6.86596334e-01
-1.20514560e+00 6.18626356e-01 3.65093350e-01 -7.15758801e-01
-7.38547789e-03 2.86336690e-01 -8.58576000e-01 -3.99797969e-02
-3.73028249e-01 4.98780489e-01 4.77387220e-01 -4.61954296e-01
-2.56843888e-03 -4.36846614e-02 3.82335067e-01 7.26911873e-02
-5.19647539e-01 -1.35398611e-01 -3.68545771e-01 -6.74847245e-01
-2.67932415e-02 -6.15549803e-01 -3.34742576e-01 -2.06238270e-01
-6.14921093e-01 -4.05728787e-01 7.28345633e-01 -4.05224204e-01
1.43582928e+00 -2.09157872e+00 1.85037374e-01 3.46734524e-01
9.73686337e-01 2.96870381e-01 -4.31400478e-01 9.74874020e-01
-4.36994284e-01 4.99160402e-02 8.68025199e-02 6.46478236e-02
9.21064019e-02 3.46907705e-01 5.33534354e-03 5.50818264e-01
5.07184863e-01 1.76911056e+00 -1.02104628e+00 -2.60634571e-01
4.84307349e-01 6.21068120e-01 -4.85699207e-01 1.98127665e-02
-1.68045890e-02 -8.49993527e-03 -5.27213037e-01 7.08260775e-01
4.92853880e-01 -6.66405320e-01 2.97241658e-01 -2.94755250e-01
2.71890104e-01 3.74275297e-02 -6.85038447e-01 1.32672250e+00
-1.31469220e-01 1.24169767e+00 -2.43398011e-01 -1.33298981e+00
9.39114273e-01 8.93580765e-02 5.48437297e-01 -5.89532375e-01
4.20922130e-01 -3.31068814e-01 5.76979667e-02 -3.81527364e-01
5.66249967e-01 3.89514983e-01 -1.49983451e-01 2.03693375e-01
4.18091893e-01 9.41067785e-02 3.15515846e-01 4.76926178e-01
1.33059537e+00 -5.38024247e-01 2.13220134e-01 1.06614187e-01
5.10872543e-01 -3.13060462e-01 -1.05723463e-01 4.41289604e-01
-3.13131422e-01 1.89773411e-01 8.60035777e-01 -1.05847180e+00
-7.44229317e-01 -1.07906878e+00 1.85862273e-01 1.00904214e+00
5.61229214e-02 -7.25187004e-01 -4.85641032e-01 -8.09929013e-01
2.28503615e-01 9.32441801e-02 -9.92705047e-01 -5.06315172e-01
-2.81234294e-01 -6.41807079e-01 5.72483420e-01 7.38203228e-01
6.41176045e-01 -1.25338340e+00 1.57699183e-01 9.27338377e-02
2.53265798e-01 -1.20905137e+00 -1.64605439e-01 -3.73969972e-02
-6.65849864e-01 -1.53103995e+00 -3.03839862e-01 -1.10439157e+00
7.88911521e-01 3.32764834e-01 1.39495397e+00 7.11563885e-01
-2.87498653e-01 4.33512568e-01 -2.93903619e-01 -2.68882513e-01
2.75700167e-02 3.35010260e-01 -2.39170194e-01 -8.01178366e-02
5.64268887e-01 -5.43620408e-01 -3.62692714e-01 -1.97222725e-01
-6.17077470e-01 -1.53464720e-01 5.63811660e-01 8.61484706e-01
1.77780375e-01 4.77097988e-01 4.59901661e-01 -9.19222832e-01
1.23772192e+00 -6.07130289e-01 -4.28430200e-01 4.13960963e-01
-4.70093101e-01 -2.61336088e-01 5.45929313e-01 -8.56356323e-02
-3.21536839e-01 -3.84553134e-01 -2.33742803e-01 -3.56064022e-01
7.54556209e-02 9.63362634e-01 -2.62194909e-02 -3.87951314e-01
4.59160715e-01 -4.81460541e-02 1.24571562e-01 9.92424972e-03
6.42955184e-01 2.05447614e-01 3.86285573e-01 5.80265708e-02
7.52831578e-01 2.45500565e-01 3.59732270e-01 -1.10281312e+00
-3.09467822e-01 -5.46240509e-02 -9.13407028e-01 -6.29520237e-01
1.17661464e+00 -5.05683243e-01 -9.89830673e-01 3.89788508e-01
-1.02528858e+00 -4.89260137e-01 -1.55139178e-01 4.35728401e-01
-2.94319779e-01 2.74216175e-01 -1.14237189e+00 -3.54802758e-01
-2.09023431e-01 -9.78335261e-01 7.16699243e-01 9.75626633e-02
1.69964969e-01 -1.76814520e+00 -1.31058350e-01 -4.36987318e-02
6.21690869e-01 4.92226958e-01 1.13963664e+00 -6.98512912e-01
-6.32707894e-01 -7.94469893e-01 -6.93856180e-01 1.66030660e-01
-6.77687377e-02 1.83617800e-01 -5.32843649e-01 -4.90253493e-02
-1.08189297e+00 2.49276720e-02 8.57576132e-01 4.72984552e-01
1.32649052e+00 -1.19705558e-01 -3.66329938e-01 7.68889129e-01
1.15372181e+00 -4.00201976e-02 5.66891015e-01 -2.17232093e-01
1.16930854e+00 5.51792145e-01 -3.27020347e-01 -8.95534530e-02
6.51605725e-01 1.13201827e-01 7.69390583e-01 -4.93225098e-01
-3.44079405e-01 -3.74520063e-01 -1.47200555e-01 1.24410939e+00
-2.42485553e-01 -6.23725414e-01 -8.96512508e-01 3.30430895e-01
-1.90478861e+00 -9.36575532e-01 -3.73285711e-01 1.46892643e+00
-4.16664571e-01 4.20897454e-02 1.24619812e-01 -3.42449695e-02
8.97871554e-01 5.01075149e-01 -3.43558073e-01 -5.35909176e-01
-1.82541743e-01 2.86685884e-01 6.25759125e-01 3.86017591e-01
-1.40007019e+00 1.25979078e+00 7.45014668e+00 1.15952507e-01
-8.75705898e-01 -1.67315140e-01 3.44733238e-01 3.60733956e-01
-2.35317558e-01 -3.45645100e-01 1.64264575e-01 1.19249739e-01
1.05915582e+00 -2.58036554e-02 7.72735178e-01 8.03527176e-01
-3.64838600e-01 6.39013052e-01 -9.90063787e-01 8.64056170e-01
2.11458132e-02 -1.61879742e+00 2.27802247e-01 1.01152793e-01
4.53940183e-01 3.31938028e-01 9.11605433e-02 7.03641117e-01
5.67341685e-01 -1.49334991e+00 1.42052129e-01 6.71758473e-01
7.36699224e-01 -6.98967993e-01 9.52707946e-01 -4.02220428e-01
-1.60508740e+00 -1.26878798e-01 -4.51246917e-01 -3.73128146e-01
6.92400411e-02 3.73670608e-01 -6.86510026e-01 8.25316370e-01
5.85679233e-01 1.45497406e+00 -5.82253993e-01 8.02739561e-01
-4.06769902e-01 3.21419090e-01 1.40509591e-01 -3.09927434e-01
4.54509407e-01 -5.36401749e-01 2.28436112e-01 1.24616301e+00
-8.37734416e-02 1.62025522e-02 1.46615610e-01 8.62718582e-01
-5.69906712e-01 -5.21675590e-03 -1.11805105e+00 -7.46783137e-01
4.04226571e-01 1.10848618e+00 -8.69948685e-01 -1.47516772e-01
-9.48303401e-01 8.51723611e-01 7.12850451e-01 9.21287775e-01
-8.79196942e-01 -1.01944268e+00 7.20161021e-01 -9.06979069e-02
3.47472504e-02 -5.59572577e-01 -1.33528829e-01 -9.49419737e-01
-4.06959265e-01 -2.48233110e-01 7.38223195e-01 -8.00574899e-01
-1.44450235e+00 7.94013321e-01 -1.52224556e-01 -7.65294969e-01
1.10902883e-01 -1.17105591e+00 -1.03628707e+00 6.64710343e-01
-1.32963562e+00 -1.34599686e+00 -7.26520479e-01 6.59816682e-01
-1.49154365e-02 -4.54522014e-01 8.04879189e-01 6.55669272e-01
-8.87983084e-01 5.41856170e-01 -1.35409743e-01 1.01657057e+00
-7.15180859e-02 -1.18138647e+00 1.11205006e+00 7.12243021e-01
2.55538642e-01 6.79293156e-01 -2.34654592e-03 -5.27019322e-01
-1.63483572e+00 -1.37089574e+00 9.40831423e-01 -1.82826832e-01
1.14621544e+00 -5.46283066e-01 -8.40424478e-01 1.06437039e+00
2.44708672e-01 6.24054193e-01 7.17601955e-01 3.80390108e-01
-2.71825403e-01 3.41191590e-02 -8.15218866e-01 7.84120560e-01
1.41688502e+00 -9.49278891e-01 -2.23887134e-02 5.56608975e-01
8.94306600e-01 -3.71620029e-01 -1.22574496e+00 2.32898027e-01
5.47444224e-01 -6.07246399e-01 1.29037404e+00 -1.38394475e+00
2.94534475e-01 1.24466596e-02 1.31849378e-01 -1.40035903e+00
-9.15073931e-01 -3.60138118e-01 -4.11050379e-01 4.74978864e-01
3.93915027e-01 -8.05142999e-01 9.66649294e-01 3.06702316e-01
-2.94479132e-01 -7.71410346e-01 -6.14242435e-01 -4.52494174e-01
2.55448953e-03 -4.74872679e-01 1.03387630e+00 1.26810241e+00
9.06802565e-02 3.69141370e-01 -4.46753018e-02 2.06377715e-01
3.91165733e-01 -1.14783347e-01 6.51005566e-01 -1.29258502e+00
2.02985451e-01 -9.73349929e-01 -1.16592205e+00 -7.45058060e-01
6.12479687e-01 -1.39551306e+00 -5.75159550e-01 -2.10414433e+00
-3.14231396e-01 -2.24483445e-01 -4.84211087e-01 5.50054431e-01
-4.93982360e-02 3.99123758e-01 1.91683197e-04 -3.49635154e-01
-7.36814499e-01 5.90683758e-01 1.10436440e+00 -5.15607834e-01
-1.18711367e-02 3.01926881e-02 -5.36159813e-01 3.16671133e-01
7.88368106e-01 1.16458811e-01 -5.88069856e-01 -7.70154119e-01
6.82909727e-01 -1.67396396e-01 6.09386981e-01 -8.00381184e-01
3.86216223e-01 1.76305115e-01 3.44272941e-01 -4.25174087e-01
-9.37017575e-02 -8.88247252e-01 5.73201589e-02 3.94320995e-01
-3.30349624e-01 5.34493923e-01 7.69874081e-02 7.99817443e-01
-2.56014764e-01 2.79649168e-01 2.13052318e-01 -9.05943438e-02
-1.00694370e+00 9.76710320e-01 -3.76983821e-01 -2.46866941e-01
9.72077310e-01 -3.84211540e-01 -4.51735914e-01 -5.88058829e-01
-1.08452988e+00 4.03977692e-01 -4.71757352e-02 6.74368024e-01
7.32770562e-01 -1.67406332e+00 -5.44887662e-01 4.03977990e-01
2.34031767e-01 -1.90267876e-01 2.71168470e-01 6.69020534e-01
-7.68566549e-01 6.68838501e-01 -2.70406753e-01 -2.07320377e-01
-7.20933259e-01 8.94761086e-01 5.80536187e-01 2.17748666e-03
-8.18591475e-01 8.06370378e-01 -3.97765636e-02 -6.07592702e-01
4.12434578e-01 -5.20813286e-01 -4.18040305e-01 -1.79888234e-01
7.13177100e-02 4.28921312e-01 2.78395027e-01 -5.55413961e-01
-4.26399171e-01 3.09630632e-01 3.89817834e-01 7.98753440e-01
1.30395710e+00 1.77354798e-01 -4.17918295e-01 1.08162738e-01
1.40010202e+00 -5.24110138e-01 -7.32228398e-01 -8.83166268e-02
2.89615188e-02 9.43253655e-03 1.76089615e-01 -1.57324612e-01
-1.55754113e+00 1.02797282e+00 -2.54484173e-02 7.97741830e-01
7.96900511e-01 1.28932193e-01 7.01323509e-01 6.73739791e-01
5.01931943e-02 -7.64746726e-01 1.07691102e-01 8.79772186e-01
7.92050719e-01 -1.20377707e+00 2.41235383e-02 -5.04209757e-01
-3.42251390e-01 1.37343609e+00 4.63867068e-01 -4.67158318e-01
1.29998887e+00 -2.29321513e-02 -4.08936709e-01 -8.82696927e-01
-5.32884955e-01 -4.77045387e-01 6.13088489e-01 9.14401591e-01
5.45299709e-01 4.08553123e-01 2.03285649e-01 4.54240829e-01
-2.13550180e-01 -3.65319490e-01 3.05815399e-01 5.67860842e-01
-9.31846425e-02 -7.49566615e-01 3.58461291e-01 7.37079918e-01
-1.30192339e-01 -3.40618372e-01 -7.67028451e-01 9.06050861e-01
-1.25442520e-01 7.98344254e-01 6.53209984e-01 -6.99067354e-01
4.19195175e-01 -1.65708408e-01 2.75030553e-01 -5.49401104e-01
-5.10716558e-01 -8.24277937e-01 2.22788185e-01 -4.97882664e-01
-1.96880147e-01 -4.66178171e-02 -9.69749689e-01 -8.87902379e-01
-3.74737948e-01 9.20811668e-02 3.90829921e-01 4.72528934e-01
5.32406509e-01 1.22003222e+00 3.38058084e-01 -6.58288181e-01
3.42032462e-01 -8.57593238e-01 -5.97074807e-01 3.45616490e-01
3.09102178e-01 -7.17028797e-01 3.87708768e-02 -4.28004533e-01] | [7.054903984069824, 6.294227123260498] |
84612906-7c57-4b11-8333-01011d6cbf7e | boosting-language-models-reasoning-with-chain | 2306.06427 | null | https://arxiv.org/abs/2306.06427v1 | https://arxiv.org/pdf/2306.06427v1.pdf | Boosting Language Models Reasoning with Chain-of-Knowledge Prompting | Recently, Chain-of-Thought (CoT) prompting has delivered success on complex reasoning tasks, which aims at designing a simple prompt like ``Let's think step by step'' or multiple in-context exemplars with well-designed rationales to elicit Large Language Models (LLMs) to generate intermediate reasoning steps. However, the generated rationales often come with mistakes, making unfactual and unfaithful reasoning chains. To mitigate this brittleness, we propose a novel Chain-of-Knowledge (CoK) prompting, where we aim at eliciting LLMs to generate explicit pieces of knowledge evidence in the form of structure triple. This is inspired by our human behaviors, i.e., we can draw a mind map or knowledge map as the reasoning evidence in the brain before answering a complex question. Benefiting from CoK, we additionally introduce a F^2-Verification method to estimate the reliability of the reasoning chains in terms of factuality and faithfulness. For the unreliable response, the wrong evidence can be indicated to prompt the LLM to rethink. Extensive experiments demonstrate that our method can further improve the performance of commonsense, factual, symbolic, and arithmetic reasoning tasks. | ['Ming Gao', 'Xiang Li', 'Nuo Chen', 'Qiushi Sun', 'Jianing Wang'] | 2023-06-10 | null | null | null | null | ['arithmetic-reasoning'] | ['reasoning'] | [ 1.66652799e-01 6.08411729e-01 1.79193646e-01 -4.68840539e-01
-3.97993475e-01 -4.96907413e-01 6.67068183e-01 1.21649183e-01
-5.86128421e-03 5.99839211e-01 4.86611009e-01 -6.13950431e-01
-2.19520777e-01 -8.41227829e-01 -6.52096450e-01 -2.85891950e-01
5.97424507e-01 2.01700851e-01 1.38687551e-01 -5.25835574e-01
8.14158797e-01 -1.91577096e-02 -1.02375448e+00 7.18947172e-01
1.08946109e+00 7.66596556e-01 6.21846803e-02 3.40038180e-01
-3.76200080e-01 1.88866341e+00 -6.50639653e-01 -8.90748203e-01
-1.54626027e-01 -6.92426622e-01 -1.22738779e+00 -3.18282545e-01
-6.16601482e-02 -4.26034361e-01 1.15788534e-01 1.32616079e+00
1.06854849e-01 4.71910536e-02 4.43848401e-01 -1.33569038e+00
-1.09177005e+00 1.45359123e+00 -2.75496334e-01 1.77140951e-01
8.85393798e-01 5.04876018e-01 7.99279332e-01 -8.86626244e-01
3.03627640e-01 1.59355581e+00 6.26198411e-01 6.81621373e-01
-9.79995728e-01 -6.61762476e-01 3.32487792e-01 6.47131145e-01
-1.38976848e+00 -4.87338483e-01 1.08791876e+00 -3.56185794e-01
8.12038243e-01 2.54296184e-01 6.62138522e-01 1.22013438e+00
2.42948860e-01 7.81672895e-01 1.52196252e+00 -4.53004271e-01
4.41537976e-01 8.12113881e-02 1.68523282e-01 7.78095186e-01
4.09734875e-01 6.14492521e-02 -7.69011378e-01 -1.41527683e-01
9.29663062e-01 -1.32661909e-01 -1.41891927e-01 3.89936328e-01
-1.53706074e+00 6.76056564e-01 4.72427189e-01 3.68319929e-01
-6.28579080e-01 2.38883644e-01 2.19549522e-01 1.12069309e-01
-2.75512874e-01 7.95512915e-01 -2.69258410e-01 -1.96024209e-01
-7.38578796e-01 4.55987751e-01 7.83173680e-01 8.19201648e-01
4.81803834e-01 5.17357998e-02 -5.02804637e-01 5.29364944e-01
4.40375924e-01 6.05228245e-01 6.54406130e-01 -1.30993783e+00
3.70039821e-01 9.02256668e-01 3.57620239e-01 -1.33946466e+00
-3.98601502e-01 -1.05281420e-01 -7.92484581e-01 -5.81464171e-02
3.64279300e-01 9.31714252e-02 -3.62859249e-01 1.86753452e+00
3.50872070e-01 1.01029187e-01 1.87671974e-01 1.13487351e+00
6.47543728e-01 3.31734300e-01 7.35919252e-02 -2.24700168e-01
1.59845090e+00 -7.93900907e-01 -1.02252102e+00 -2.69520611e-01
5.63188612e-01 -4.44093853e-01 1.62397122e+00 6.09237790e-01
-1.04968417e+00 -4.27701116e-01 -9.18189466e-01 -1.07368514e-01
-9.26935822e-02 -2.01502174e-01 8.54729414e-01 2.34009087e-01
-6.62024796e-01 4.42264706e-01 -3.76309574e-01 2.43103638e-01
3.10812205e-01 -2.73261249e-01 1.04240619e-01 -2.56239623e-01
-1.60614598e+00 1.17778826e+00 8.19389343e-01 4.47600394e-01
-8.25891018e-01 -5.73671043e-01 -7.66742527e-01 -1.39209982e-02
9.57613528e-01 -7.62576401e-01 1.40460360e+00 -9.20851529e-01
-1.69490409e+00 6.04983151e-01 7.40811229e-02 -2.31468201e-01
3.84030908e-01 -2.86522031e-01 -5.29616117e-01 2.71634728e-01
1.89715043e-01 3.89720917e-01 7.74989367e-01 -1.05880308e+00
-1.65048568e-03 -1.14543766e-01 5.05500853e-01 -8.46369863e-02
2.29026929e-01 1.11387394e-01 3.25348943e-01 -7.42094994e-01
3.30835760e-01 -6.56205773e-01 -1.59661755e-01 -1.99207932e-01
-7.50838637e-01 -3.82518888e-01 7.95047730e-02 -6.43695474e-01
1.46483338e+00 -1.93485224e+00 -7.69683644e-02 1.70448318e-01
5.38772047e-01 1.31034270e-01 2.27801111e-02 4.88962024e-01
-1.05449241e-02 3.85903060e-01 -2.25119188e-01 5.56752861e-01
4.51396704e-01 2.69173026e-01 -6.98585391e-01 -2.43246794e-01
3.26503634e-01 1.37597895e+00 -1.32721460e+00 -6.96116030e-01
-1.22493140e-01 1.96935292e-02 -6.29157901e-01 5.03380597e-01
-5.24738729e-01 1.83205411e-01 -5.19607484e-01 6.06246650e-01
4.83916461e-01 -7.15691388e-01 1.41868874e-01 -2.34825015e-01
2.02703387e-01 6.11989319e-01 -1.14074039e+00 1.37874305e+00
-2.52943128e-01 5.39054573e-02 -5.34335315e-01 -7.62329578e-01
8.44589591e-01 2.17524588e-01 -5.95933378e-01 -5.97601414e-01
1.30274490e-01 2.33780116e-01 2.73536295e-01 -8.97127450e-01
1.69835880e-01 -6.14935040e-01 -2.04539806e-01 6.28619790e-01
-4.71558094e-01 -5.46462417e-01 -3.12057994e-02 5.86097062e-01
1.02778745e+00 2.77952999e-01 8.07146728e-01 -2.34276861e-01
7.66540766e-01 9.26247165e-02 5.97713947e-01 7.04747975e-01
-9.45512280e-02 4.75429744e-02 8.08829308e-01 -4.73609835e-01
-6.67702436e-01 -9.68351066e-01 3.83485496e-01 8.29391599e-01
1.54372916e-01 -5.27034461e-01 -8.48118961e-01 -5.23359895e-01
-3.12266767e-01 1.55837488e+00 -5.36923945e-01 -5.18087983e-01
-4.46984440e-01 -2.11310133e-01 8.15860450e-01 6.87441230e-01
8.89620423e-01 -1.45761132e+00 -9.34809625e-01 3.18064362e-01
-6.73696458e-01 -1.18947005e+00 -3.54756743e-01 -2.26501450e-01
-7.02986300e-01 -1.26877797e+00 3.35442834e-02 -1.63742647e-01
8.03859651e-01 1.87150225e-01 1.17151141e+00 6.29428566e-01
2.99435407e-01 1.71360672e-01 -4.71998453e-01 -2.30080754e-01
-6.09245002e-01 -7.89002299e-01 1.05017014e-01 -1.91392973e-01
4.72811133e-01 -6.74579203e-01 -5.29715180e-01 2.37891510e-01
-1.01020730e+00 6.12308741e-01 5.96845567e-01 7.32653201e-01
2.37068102e-01 4.27530296e-02 7.51021504e-01 -8.14058363e-01
1.30230558e+00 -3.95021498e-01 -2.28879750e-01 5.36967814e-01
-8.38786781e-01 4.96535599e-01 1.05625701e+00 -6.60358846e-01
-1.25538909e+00 -7.39827037e-01 1.68945640e-01 -2.89135814e-01
-1.56434804e-01 7.72459209e-01 -8.46972913e-02 2.04370543e-01
8.48818719e-01 3.80370438e-01 -3.57782364e-01 -3.77090126e-02
6.82713449e-01 4.60983723e-01 7.41112947e-01 -1.43087363e+00
7.45838344e-01 8.61101896e-02 -1.87562987e-01 4.59436625e-02
-1.37765336e+00 2.98950851e-01 -3.00907433e-01 -3.14115256e-01
5.56311607e-01 -5.27410448e-01 -1.18753469e+00 2.52988189e-01
-1.41068566e+00 -3.33053052e-01 -1.33555800e-01 2.71480501e-01
-6.73516273e-01 5.96681118e-01 -7.16819167e-01 -9.98290658e-01
-4.14998531e-01 -8.75005960e-01 6.65468693e-01 4.14769381e-01
-6.75981998e-01 -6.99120998e-01 -2.86388397e-01 5.30843079e-01
4.12330002e-01 8.69523957e-02 1.35537457e+00 -8.92284453e-01
-4.57375079e-01 1.12484142e-01 -2.07030311e-01 2.72161812e-01
-1.07806027e-01 -1.02466859e-01 -6.71936035e-01 5.60163081e-01
4.64086324e-01 -7.17373550e-01 1.42726138e-01 -3.09928834e-01
1.20727122e+00 -9.61403430e-01 1.40236735e-01 1.55977309e-01
1.01240635e+00 1.22786671e-01 8.58481407e-01 1.61135077e-01
3.45408380e-01 5.74877083e-01 6.09672904e-01 4.66604084e-01
7.76069283e-01 2.25586727e-01 4.34332713e-02 6.36234224e-01
1.04401484e-01 -9.03553009e-01 3.58073473e-01 9.43245649e-01
-2.56849080e-01 3.40379566e-01 -1.07337701e+00 2.17104986e-01
-1.87148714e+00 -1.31468618e+00 -1.89166903e-01 1.75638413e+00
1.48270917e+00 4.12310898e-01 -3.73202682e-01 2.47717664e-01
5.91761649e-01 -7.18735158e-02 -5.18959999e-01 -4.57235783e-01
6.08978979e-02 -5.44200540e-02 -4.65701163e-01 5.25644839e-01
-1.64372716e-02 1.11653280e+00 5.86345387e+00 7.60924935e-01
-8.28848124e-01 -1.31332083e-02 3.37374985e-01 2.74095088e-01
-9.31710124e-01 4.62211519e-01 -4.81029630e-01 5.45980632e-01
7.76667833e-01 -2.23830298e-01 7.96067238e-01 7.89087415e-01
3.02134246e-01 -3.67872804e-01 -1.15891838e+00 8.83763552e-01
1.20814629e-01 -1.26787996e+00 3.21586788e-01 -5.89276135e-01
3.63386184e-01 -8.94468784e-01 -4.24752265e-01 6.06965423e-01
5.59767425e-01 -9.59993303e-01 1.03369915e+00 8.31017315e-01
3.54837865e-01 -3.41984093e-01 5.66898167e-01 9.87708986e-01
-7.83213556e-01 -1.39658883e-01 -2.38971040e-01 -4.88338143e-01
1.69750422e-01 8.19961905e-01 -8.55377614e-01 4.21226740e-01
2.28172123e-01 6.25346676e-02 -5.72750092e-01 3.53614688e-01
-1.17235017e+00 6.15462899e-01 -1.26802802e-01 -3.70764583e-01
-9.21038389e-02 4.11884077e-02 1.73186809e-01 9.12998796e-01
-5.03703915e-02 8.36261451e-01 7.95583874e-02 1.65128422e+00
9.11040306e-02 -1.37097105e-01 -2.91531503e-01 -3.00540626e-01
8.22170317e-01 1.04984808e+00 -7.63823688e-01 -7.85910547e-01
2.10582949e-02 8.00838530e-01 5.56050837e-01 3.13186109e-01
-1.00021386e+00 -2.16219649e-01 -6.46907017e-02 -8.69057029e-02
-2.29819074e-01 -1.13178827e-01 -4.50028241e-01 -1.46749270e+00
9.79912654e-02 -1.02545822e+00 1.92044809e-01 -1.65450478e+00
-1.46384859e+00 3.27467352e-01 3.69243771e-01 -6.82714939e-01
-2.28016287e-01 -4.46625471e-01 -7.73416221e-01 7.43185401e-01
-1.36370385e+00 -9.95301604e-01 -3.71746480e-01 5.28680205e-01
3.72467041e-01 3.01613897e-01 6.77638352e-01 -3.34449530e-01
-3.53991121e-01 3.26487601e-01 -1.14039886e+00 1.15316100e-01
4.10727739e-01 -9.80237782e-01 2.11570859e-01 7.85852790e-01
-4.50377576e-02 1.41129553e+00 8.70764136e-01 -8.77328813e-01
-1.45833302e+00 -4.63506490e-01 1.12547481e+00 -5.85631728e-01
8.71978164e-01 -1.61909778e-02 -1.19573295e+00 7.16512799e-01
2.95835882e-02 -2.07984433e-01 7.66869545e-01 -2.19714753e-02
-9.12643671e-01 2.16795489e-01 -1.15558410e+00 1.00469553e+00
1.30914104e+00 -6.76567435e-01 -1.67376590e+00 4.10020620e-01
1.08703494e+00 -2.55208045e-01 -5.31035900e-01 2.52370954e-01
4.29781973e-01 -9.49690342e-01 7.40246356e-01 -8.92248273e-01
9.77632821e-01 -6.57882690e-01 -2.90109962e-01 -1.05261350e+00
-4.90884125e-01 -6.74632728e-01 -4.19282705e-01 1.15788889e+00
1.60265714e-01 -5.12916923e-01 5.45172170e-02 1.04486740e+00
-6.13216236e-02 -7.35471964e-01 -6.80070639e-01 -5.66594183e-01
-1.72074392e-01 -7.74928033e-01 9.94854271e-01 1.12816298e+00
7.25693345e-01 5.12584150e-01 -2.34265134e-01 1.00704782e-01
4.46737051e-01 3.27031463e-01 5.63466609e-01 -9.41965282e-01
-3.78293902e-01 -3.58898014e-01 2.06077486e-01 -9.65687275e-01
1.77699104e-01 -9.10502374e-01 1.43079773e-01 -1.42029345e+00
3.40132564e-01 -1.14094250e-01 -2.03502983e-01 8.12757611e-01
-6.63569629e-01 -4.99856800e-01 3.10565889e-01 2.43681222e-01
-7.79142618e-01 4.26050514e-01 1.53003395e+00 1.78258568e-02
1.41437680e-01 -5.04766405e-01 -1.30642223e+00 1.01358593e+00
5.89624226e-01 -4.03647542e-01 -6.00333214e-01 -1.10867105e-01
1.13613045e+00 4.75303024e-01 8.53221238e-01 -7.10183024e-01
3.61901432e-01 -8.31013680e-01 1.32571787e-01 -3.30492675e-01
-1.21400379e-01 -6.37558937e-01 1.62474692e-01 5.27741373e-01
-7.22665727e-01 1.37877658e-01 -6.81837499e-02 3.57908070e-01
5.24378717e-02 -4.38227326e-01 4.45574254e-01 -5.21113694e-01
-7.96809852e-01 -4.99562651e-01 -1.45528615e-01 3.62816215e-01
5.77672541e-01 -5.51480521e-03 -7.02827990e-01 -2.61104822e-01
-3.74022305e-01 2.68051565e-01 1.79200217e-01 2.30650634e-01
1.02353263e+00 -1.48935246e+00 -6.08338416e-01 -5.51787624e-03
2.63673812e-01 2.07131561e-02 3.51576358e-01 1.03321421e+00
-2.67073095e-01 3.59094888e-01 -1.52806282e-01 -2.19401523e-01
-4.89777476e-01 9.50478315e-01 2.31304660e-01 -1.74879685e-01
-6.92891777e-01 7.68848240e-01 1.41163599e-02 -2.93982923e-01
-2.22423434e-01 -6.55170977e-01 -2.33660683e-01 -2.87378490e-01
9.55102980e-01 1.10969849e-01 -3.10596228e-01 1.36647999e-01
-4.97315109e-01 2.51795858e-01 4.66443645e-03 -3.24143797e-01
7.92328596e-01 2.76862364e-02 -4.46184427e-01 7.26240396e-01
2.78951436e-01 1.86553374e-01 -9.19818938e-01 -2.41356805e-01
1.87684923e-01 -5.59040248e-01 -4.16183591e-01 -1.40290093e+00
-3.13299268e-01 7.38550484e-01 -4.32896316e-01 1.83722585e-01
9.20974374e-01 5.28203808e-02 6.16611898e-01 8.18938136e-01
7.71835804e-01 -1.03019142e+00 4.40466791e-01 4.88959551e-01
1.29436421e+00 -1.04276299e+00 -2.27118339e-02 -2.51427889e-01
-1.03261721e+00 1.26509595e+00 9.16696608e-01 8.40491429e-02
1.33211464e-01 1.80012316e-01 -1.08061428e-03 -5.29121995e-01
-1.12806749e+00 2.77258843e-01 -1.60653636e-01 3.90365541e-01
3.37136507e-01 1.81913674e-01 -4.78058994e-01 1.39335477e+00
-5.59110880e-01 4.58443850e-01 7.11713016e-01 8.39526534e-01
-6.32396638e-01 -4.73417372e-01 -7.67743528e-01 2.54216254e-01
-6.79966360e-02 -4.99411911e-01 -5.77198088e-01 3.88065815e-01
5.65428361e-02 1.14068961e+00 -7.14451969e-01 -5.39020360e-01
2.62091190e-01 3.65155309e-01 5.48339367e-01 -4.76977080e-01
-6.27628267e-01 -5.07622600e-01 1.75162882e-01 -7.20228374e-01
-3.50296378e-01 -2.08781958e-01 -1.77786267e+00 -6.46551907e-01
-2.72273719e-01 3.62445205e-01 2.36894861e-01 1.44442523e+00
1.78353027e-01 4.71294403e-01 2.90177912e-01 -1.95700750e-01
-1.06154013e+00 -1.01024854e+00 -1.45267546e-01 6.70652151e-01
-2.98291780e-02 -7.02614903e-01 -4.21128601e-01 6.09597415e-02] | [9.601309776306152, 7.513340473175049] |
6036b831-0a6c-4924-b9e8-99f5204257c3 | structure-amplification-on-multi-layer | 2108.00127 | null | https://arxiv.org/abs/2108.00127v1 | https://arxiv.org/pdf/2108.00127v1.pdf | Structure Amplification on Multi-layer Stochastic Block Models | Much of the complexity of social, biological, and engineered systems arises from a network of complex interactions connecting many basic components. Network analysis tools have been successful at uncovering latent structure termed communities in such networks. However, some of the most interesting structure can be difficult to uncover because it is obscured by the more dominant structure. Our previous work proposes a general structure amplification technique called HICODE that uncovers many layers of functional hidden structure in complex networks. HICODE incrementally weakens dominant structure through randomization allowing the hidden functionality to emerge, and uncovers these hidden structure in real-world networks that previous methods rarely uncover. In this work, we conduct a comprehensive and systematic theoretical analysis on the hidden community structure. In what follows, we define multi-layer stochastic block model, and provide theoretical support using the model on why the existence of hidden structure will make the detection of dominant structure harder compared with equivalent random noise. We then provide theoretical proofs that the iterative reducing methods could help promote the uncovering of hidden structure as well as boosting the detection quality of dominant structure. | ['John E. Hopcroft', 'Bart Selman', 'Jialu Bao', 'Kun He', 'Xiaodong Xin'] | 2021-07-31 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 4.80403185e-01 7.57124364e-01 -1.15551151e-01 4.45163250e-01
2.36015752e-01 -9.53464448e-01 1.74615219e-01 -9.11191665e-03
6.56292319e-01 7.21787274e-01 3.88159335e-01 -5.81184566e-01
-4.12361026e-01 -8.47656250e-01 -7.70917237e-01 -9.92192030e-01
-9.19978321e-01 1.14439785e-01 4.48429406e-01 -1.02537856e-01
-9.47082639e-02 8.93927291e-02 -1.26592767e+00 3.94082218e-01
6.03646755e-01 1.16448276e-01 2.43306711e-01 7.19334543e-01
-2.50064641e-01 8.48766088e-01 -4.65309173e-01 -1.12876128e-02
2.07846925e-01 -7.30066419e-01 -5.97319543e-01 2.94578254e-01
-4.00851727e-01 -2.72194352e-02 -1.15522414e-01 1.06319261e+00
1.07304953e-01 -8.72417212e-01 3.87274176e-01 -1.40558887e+00
-5.30659556e-01 1.20761418e+00 -8.31147492e-01 -8.27532336e-02
3.05277348e-01 8.78013372e-02 1.25941586e+00 -7.01771021e-01
9.35419858e-01 1.36235607e+00 7.84495175e-01 4.17027175e-01
-1.65045822e+00 -6.47030473e-01 7.47067854e-02 -4.72944617e-01
-1.30252802e+00 -2.92012125e-01 1.05241430e+00 -7.62622476e-01
5.41788280e-01 3.52552384e-01 9.21534061e-01 1.00930798e+00
2.74846822e-01 8.46467197e-01 1.08618391e+00 -4.92940068e-01
6.58419654e-02 -9.37279463e-02 2.75403917e-01 1.03061593e+00
1.14826202e+00 7.76333595e-03 -4.06329125e-01 -6.85584486e-01
6.71959579e-01 2.60250509e-01 -4.13125187e-01 -5.80541372e-01
-1.11193240e+00 8.36343765e-01 1.92717224e-01 4.47155684e-01
-5.37102632e-02 6.94405437e-02 5.81564307e-02 4.48027849e-01
3.63549531e-01 4.98415470e-01 -2.93730527e-01 2.42881835e-01
-6.91442132e-01 -2.34789580e-01 9.06124353e-01 7.91728079e-01
1.09199905e+00 -2.53301322e-01 2.81290054e-01 1.10726908e-01
4.77458566e-01 1.21024564e-01 -2.55785376e-01 -9.57001150e-01
2.14788139e-01 1.13610470e+00 -1.95894375e-01 -1.30975044e+00
-3.03561926e-01 -6.37375951e-01 -1.30822372e+00 8.80504251e-02
3.29714179e-01 -3.44019413e-01 -5.11343479e-01 1.72372782e+00
2.07420036e-01 1.77837834e-02 -6.77674413e-02 1.66975617e-01
4.23567295e-01 5.49113989e-01 -7.98410416e-01 -4.33952868e-01
1.07819581e+00 -7.22748816e-01 -9.49191332e-01 1.67056352e-01
5.12393117e-01 -4.40793216e-01 7.66606927e-01 1.18479013e-01
-8.01942348e-01 -4.97652106e-02 -1.25597978e+00 5.31878114e-01
-3.47708352e-02 -4.03208464e-01 7.66858697e-01 7.57746637e-01
-1.18740702e+00 5.25112987e-01 -7.97914386e-01 -2.91900307e-01
3.50281835e-01 1.78109050e-01 -2.06446737e-01 -1.01107806e-01
-9.47431982e-01 1.03001697e-02 1.73490375e-01 2.86084175e-01
-1.18832934e+00 -4.25366849e-01 -6.43730283e-01 1.30049989e-01
8.53520989e-01 -7.05699027e-01 4.21735644e-01 -8.40034008e-01
-8.14906180e-01 4.70823318e-01 -4.81149644e-01 -2.99076140e-01
2.11560354e-01 1.58418700e-01 -8.69793072e-03 2.88956702e-01
2.68035531e-01 8.27789456e-02 7.89107919e-01 -1.45291603e+00
4.10152301e-02 -1.40852451e-01 1.79790899e-01 -5.93439400e-01
-4.78468508e-01 -2.39055470e-01 1.51984364e-01 -4.04665440e-01
3.75873983e-01 -9.83978570e-01 -4.32137787e-01 -9.76110846e-02
-1.03597009e+00 5.47742806e-02 7.69149363e-01 -1.63839936e-01
1.34555399e+00 -2.09854722e+00 1.47551626e-01 6.45581067e-01
1.53799391e+00 -9.51713473e-02 -1.26163468e-01 1.15573859e+00
-1.51541427e-01 8.62377644e-01 -2.59915441e-01 7.64878765e-02
-2.98174113e-01 2.74705768e-01 -2.85951823e-01 4.90743667e-01
1.94137394e-01 1.00417042e+00 -8.82478535e-01 -1.81424394e-01
-4.64283288e-01 2.69086629e-01 -6.78468704e-01 -2.12039556e-02
2.51188632e-02 2.63285995e-01 -4.22171086e-01 7.76977956e-01
6.49572372e-01 -1.05800295e+00 8.59582543e-01 2.57116735e-01
-2.29813933e-01 1.32346284e-02 -9.71687973e-01 8.02280426e-01
6.31931901e-01 8.54388714e-01 4.91469353e-01 -9.19559717e-01
6.68667793e-01 3.64847153e-01 5.22614896e-01 2.50379980e-01
-2.92023420e-01 -2.32339412e-01 5.88762224e-01 -3.38772029e-01
-5.93025424e-02 1.32750869e-01 1.42549202e-01 9.84462440e-01
-4.37869132e-01 7.00948358e-01 5.93907274e-02 7.24327266e-01
1.85329688e+00 -6.33771896e-01 2.85659492e-01 -5.75091839e-01
1.38985023e-01 -4.83965725e-02 8.03786814e-01 9.74371493e-01
-2.23546371e-01 -2.10750420e-02 1.30937898e+00 -2.27491230e-01
-1.28940868e+00 -9.09212887e-01 2.98346728e-01 6.88364208e-01
-2.33921632e-02 -9.97238874e-01 -7.46111631e-01 -2.80111521e-01
1.07728355e-01 -3.68155479e-01 -9.53845501e-01 -2.26032525e-01
-9.90488604e-02 -9.11510527e-01 7.41865337e-01 3.57834138e-02
1.89619049e-01 -7.36336350e-01 -7.46617690e-02 5.41891307e-02
-3.84041935e-01 -8.36307883e-01 -3.48506540e-01 1.54388890e-01
-9.10845220e-01 -1.42066598e+00 -4.82574821e-01 -5.67891538e-01
1.14107883e+00 7.82387793e-01 8.64491165e-01 6.80095136e-01
-4.67520773e-01 8.53481889e-02 -2.97578037e-01 -1.75008655e-01
-6.18631542e-01 3.80848609e-02 1.51884347e-01 8.30782354e-02
1.12116985e-01 -8.66431594e-01 -4.08035070e-01 3.16630304e-01
-8.11596990e-01 2.99538821e-01 7.55151093e-01 7.14578032e-01
7.09883403e-03 8.87504816e-01 2.45840982e-01 -9.88664806e-01
8.60717475e-01 -7.76644766e-01 -5.43141723e-01 2.60231465e-01
-7.42182612e-01 2.81769484e-01 2.50835806e-01 -3.12171072e-01
-4.39903945e-01 -5.84871508e-02 4.70896333e-01 -9.59368497e-02
3.16228092e-01 9.31401551e-01 -1.76318854e-01 -1.01575114e-01
3.99613500e-01 2.56230772e-01 2.98768938e-01 -4.53534901e-01
6.91703409e-02 2.42657185e-01 -2.85819322e-01 -4.79861259e-01
1.22523332e+00 8.80132616e-01 1.81135401e-01 -1.09107363e+00
-5.86271584e-01 -3.47190827e-01 -7.07523763e-01 -3.02340984e-01
2.96397775e-01 -7.32070923e-01 -9.22536254e-01 1.60582080e-01
-1.12666202e+00 -2.34528616e-01 -7.76913837e-02 -1.57880425e-01
5.67163788e-02 6.80727780e-01 -7.82897949e-01 -1.23945117e+00
9.52416286e-02 -7.23119080e-01 6.24789596e-01 -5.71994960e-01
-3.57111156e-01 -9.46485817e-01 6.05062723e-01 5.43932803e-02
2.18560398e-01 7.28149176e-01 1.08759356e+00 -1.52351782e-01
-1.16957593e+00 4.29916643e-02 -8.57612491e-02 -1.95977092e-01
3.78913879e-01 4.44812059e-01 -7.20678806e-01 -4.02126789e-01
-2.06192002e-01 2.46892959e-01 9.03255463e-01 4.38089997e-01
4.42902595e-01 -7.28637993e-01 -6.46100998e-01 4.25709784e-01
1.09131587e+00 -1.26713440e-01 5.08141458e-01 -4.71627340e-02
7.29632437e-01 1.04173553e+00 -3.63173455e-01 2.77714670e-01
1.39583414e-02 4.49479669e-02 3.41105729e-01 -3.96975398e-01
3.74977559e-01 -5.16515553e-01 5.80917418e-01 1.32396102e+00
-1.01208940e-01 -4.16915379e-02 -1.04512560e+00 5.05464911e-01
-1.82512784e+00 -1.11972260e+00 -5.29961467e-01 1.76962745e+00
8.84612560e-01 4.33516115e-01 4.01900381e-01 3.67086262e-01
1.17398572e+00 -7.72183612e-02 -4.89074379e-01 2.24939853e-01
-6.69026732e-01 -3.64207566e-01 3.69929850e-01 5.60360730e-01
-5.21927476e-01 5.78195810e-01 7.90375376e+00 3.30660641e-01
-5.90692341e-01 -5.01434319e-02 4.62338030e-01 1.44141212e-01
-6.28945291e-01 4.85005081e-01 -8.03991258e-01 1.85416251e-01
6.48790538e-01 -4.26742971e-01 5.02204061e-01 6.26917064e-01
3.48336548e-01 2.00815678e-01 -8.50735128e-01 4.75125998e-01
-3.87590408e-01 -1.71711338e+00 -4.91016991e-02 1.01954186e+00
1.04804456e+00 -1.23739004e-01 -2.11038083e-01 -3.06948543e-01
8.55165303e-01 -9.61262822e-01 2.40846187e-01 3.31167966e-01
5.19547582e-01 -4.10410225e-01 4.70342070e-01 6.97623909e-01
-1.41698003e+00 -1.87327877e-01 -3.61053884e-01 -5.90173006e-01
-1.18212380e-01 1.31991053e+00 -8.47159147e-01 1.80219337e-01
5.38527787e-01 8.53664815e-01 -4.74915415e-01 6.63228750e-01
-3.23869139e-01 9.23404574e-01 -2.07942739e-01 -1.22984953e-01
5.60737215e-04 -4.82730679e-02 9.31154251e-01 8.40442598e-01
4.54676077e-02 -1.20786525e-01 2.14407459e-01 1.28310466e+00
-2.73666307e-02 -4.11677033e-01 -1.30428779e+00 -8.08722079e-01
4.75069761e-01 1.17121828e+00 -1.19473207e+00 -2.79217958e-01
-2.15480864e-01 5.67231297e-01 2.21869856e-01 4.71175104e-01
-3.34135979e-01 -3.10593337e-01 5.64589977e-01 7.83865273e-01
3.49102050e-01 -5.47347784e-01 -1.19940564e-01 -1.33573961e+00
-4.96375337e-02 -8.86454582e-01 1.37952007e-02 -3.56789678e-01
-1.29941070e+00 4.15541917e-01 -3.14217180e-01 -8.86146307e-01
-2.47881785e-02 -2.59396851e-01 -6.57302916e-01 3.69759470e-01
-9.56725359e-01 -7.52730608e-01 -2.92324796e-02 4.31238472e-01
9.62235220e-03 -5.60963266e-02 6.25260055e-01 -1.31208068e-02
-7.96002328e-01 9.46539789e-02 4.14337963e-01 3.08544368e-01
8.35736319e-02 -1.06780791e+00 4.09329504e-01 9.41036224e-01
-2.74854511e-01 1.35482514e+00 7.53016412e-01 -1.16946650e+00
-1.44491100e+00 -9.64616835e-01 7.92933881e-01 -4.58723068e-01
1.17165291e+00 -1.12705886e+00 -7.77057052e-01 7.33480453e-01
1.89744517e-01 -4.17032361e-01 9.46004748e-01 2.83724993e-01
-4.39748377e-01 3.58992040e-01 -7.67223537e-01 7.15507448e-01
1.36477268e+00 -7.40293682e-01 -2.92685717e-01 1.84538409e-01
1.00533521e+00 7.55745411e-01 -4.51613456e-01 2.91587412e-01
8.07610929e-01 -8.38781893e-01 7.37862110e-01 -5.20502865e-01
4.54962075e-01 -4.96386290e-01 5.86425923e-02 -7.73665309e-01
-8.20001245e-01 -1.45610583e+00 -4.07770365e-01 1.09859776e+00
5.22550762e-01 -7.12580323e-01 1.01432860e+00 -6.95895106e-02
4.42886710e-01 -4.36783195e-01 -5.97411871e-01 -9.54688072e-01
-2.54953682e-01 1.19525380e-01 3.59805793e-01 1.26342344e+00
4.22223598e-01 6.48452401e-01 -5.84820032e-01 1.58088997e-01
1.15936530e+00 5.78675568e-02 9.15487289e-01 -1.62324846e+00
-4.29706335e-01 -3.63003641e-01 -3.28467876e-01 -8.02696645e-01
-1.93336114e-01 -6.97469175e-01 -1.47651985e-01 -1.19314897e+00
8.42035890e-01 -1.63342822e-02 -6.69050738e-02 4.57260251e-01
3.97642441e-02 5.87156741e-03 -1.30507454e-01 5.99715769e-01
-7.41147220e-01 3.40992123e-01 1.36420500e+00 -2.49231100e-01
-1.61713123e-01 -1.50486290e-01 -1.22534895e+00 7.44450033e-01
6.80385172e-01 -7.34777272e-01 -5.33605099e-01 1.91116899e-01
1.02085006e+00 -6.53505698e-02 3.10287476e-01 -5.63939095e-01
2.24543601e-01 6.07752167e-02 8.98534711e-03 -6.64153934e-01
-8.11803043e-02 -6.41152740e-01 5.03315926e-01 1.02683508e+00
-2.02041313e-01 -5.05654467e-03 -1.31398767e-01 9.88772810e-01
1.32222578e-01 1.01919889e-01 2.27263957e-01 -2.75693893e-01
1.64352015e-01 3.48508582e-02 -9.69008744e-01 -1.43541351e-01
8.82448792e-01 -3.95407200e-01 -5.71851075e-01 -6.13973081e-01
-9.38053310e-01 1.03442170e-01 5.16025901e-01 5.10882474e-02
7.14456260e-01 -1.02444613e+00 -7.63610899e-01 3.50701064e-01
-2.94768941e-02 -3.83814901e-01 -9.13763195e-02 9.65897977e-01
-4.58220005e-01 3.76923203e-01 -5.15387543e-02 -7.79186010e-01
-1.59985662e+00 8.24539840e-01 2.98147015e-02 -4.60958362e-01
-6.12177491e-01 5.49170554e-01 7.19879568e-01 -1.00948751e-01
1.77158043e-01 -2.70670921e-01 4.40414548e-02 -9.06691775e-02
7.26608932e-01 2.96990335e-01 -7.02353239e-01 -2.39601970e-01
-2.02826276e-01 4.10616696e-01 -1.41870975e-01 5.85759059e-02
1.33448207e+00 -5.34428954e-01 -1.00340915e+00 7.04731762e-01
8.98586869e-01 4.33549464e-01 -9.70483541e-01 -2.14928120e-01
3.87452304e-01 -1.60411298e-01 -3.63335341e-01 -1.22825891e-01
-7.80314147e-01 8.25997412e-01 -1.25790555e-02 1.09761560e+00
7.00746775e-01 3.97099942e-01 3.89873534e-01 6.80391371e-01
3.79850239e-01 -4.67204392e-01 3.89681309e-01 4.51644331e-01
5.14574766e-01 -6.33604109e-01 2.35948563e-02 -8.78985286e-01
8.22911486e-02 8.42971325e-01 1.70949697e-01 -1.37197331e-01
1.05639338e+00 6.05475008e-01 -5.71615100e-01 -8.93011332e-01
-1.17777860e+00 -2.75246561e-01 -1.96839198e-01 6.10568583e-01
3.13859284e-01 9.82385203e-02 8.29754025e-02 4.47686046e-01
5.75450957e-02 -2.41768807e-01 7.82693267e-01 9.71328080e-01
-8.79472792e-01 -1.04241037e+00 -5.53112030e-01 1.52987301e-01
-5.57208061e-01 -4.40409243e-01 -1.27757132e+00 4.71052736e-01
4.91475835e-02 1.04285729e+00 -1.95523456e-01 -6.91657782e-01
-3.44305694e-01 -1.28744170e-01 1.10492758e-01 -7.19685555e-01
3.01143471e-02 5.54141104e-01 1.47238337e-02 -2.72751629e-01
-4.05765921e-01 -4.58072037e-01 -6.90470695e-01 -7.84752131e-01
-6.16301060e-01 4.59275961e-01 6.36094213e-02 8.21289897e-01
5.24315834e-01 7.42979825e-01 7.37537205e-01 -2.70699084e-01
1.57045096e-01 -7.55639851e-01 -7.99647212e-01 -2.02940553e-01
8.04643095e-01 -5.60819149e-01 -8.31480563e-01 1.96453556e-01] | [6.929213523864746, 5.230286121368408] |
972371ac-4fa0-44d0-a0cd-91d22e872c85 | a-new-ensemble-learning-framework-for-3d | 1812.03945 | null | http://arxiv.org/abs/1812.03945v1 | http://arxiv.org/pdf/1812.03945v1.pdf | A New Ensemble Learning Framework for 3D Biomedical Image Segmentation | 3D image segmentation plays an important role in biomedical image analysis.
Many 2D and 3D deep learning models have achieved state-of-the-art segmentation
performance on 3D biomedical image datasets. Yet, 2D and 3D models have their
own strengths and weaknesses, and by unifying them together, one may be able to
achieve more accurate results. In this paper, we propose a new ensemble
learning framework for 3D biomedical image segmentation that combines the
merits of 2D and 3D models. First, we develop a fully convolutional network
based meta-learner to learn how to improve the results from 2D and 3D models
(base-learners). Then, to minimize over-fitting for our sophisticated
meta-learner, we devise a new training method that uses the results of the
base-learners as multiple versions of "ground truths". Furthermore, since our
new meta-learner training scheme does not depend on manual annotation, it can
utilize abundant unlabeled 3D image data to further improve the model.
Extensive experiments on two public datasets (the HVSMR 2016 Challenge dataset
and the mouse piriform cortex dataset) show that our approach is effective
under fully-supervised, semi-supervised, and transductive settings, and attains
superior performance over state-of-the-art image segmentation methods. | ['Danny Z. Chen', 'Peixian Liang', 'Yizhe Zhang', 'Lin Yang', 'Zhuo Zhao', 'Hao Zheng', 'Chaoli Wang'] | 2018-12-10 | null | null | null | null | ['3d-medical-imaging-segmentation'] | ['medical'] | [ 1.54375374e-01 7.86584243e-02 -2.11784586e-01 -5.04503012e-01
-8.37558389e-01 -3.09215486e-01 2.31691346e-01 -4.61903214e-03
-6.95968151e-01 4.58543479e-01 -1.72179744e-01 -4.82136577e-01
1.72043785e-01 -5.93118966e-01 -6.66539133e-01 -7.93065310e-01
5.27980328e-02 4.03299153e-01 3.21396917e-01 -5.93699664e-02
6.09533954e-03 5.13529539e-01 -9.43745077e-01 1.40050754e-01
1.06249905e+00 9.62448955e-01 2.76914626e-01 4.40652400e-01
-3.97650272e-01 3.31352443e-01 -4.61437166e-01 -5.07519133e-02
3.06177914e-01 -3.66168976e-01 -9.36101556e-01 2.84425080e-01
8.98241550e-02 -3.33574444e-01 -1.06442362e-01 1.04198682e+00
8.34758699e-01 -3.29219043e-01 7.54383564e-01 -8.46451521e-01
-6.47577882e-01 3.56045842e-01 -7.68038809e-01 2.75277644e-01
-2.75079850e-02 5.86576611e-02 5.00425756e-01 -5.72551966e-01
3.87465239e-01 1.07754576e+00 6.39018297e-01 8.19764197e-01
-1.39023411e+00 -6.22554779e-01 4.02871519e-01 -1.80436343e-01
-1.10050535e+00 -4.73090038e-02 8.60459566e-01 -5.60487866e-01
6.59337282e-01 -1.79556921e-01 6.69608235e-01 1.02129984e+00
2.08716884e-01 1.30523324e+00 1.43308485e+00 -2.33079210e-01
1.70914531e-01 -7.95846283e-02 2.75638908e-01 7.86176622e-01
3.01321372e-02 -6.40160292e-02 3.45285125e-02 -2.66574938e-02
1.07147896e+00 -2.46853884e-02 -2.12706253e-01 -4.86162454e-01
-1.19326425e+00 7.10165560e-01 5.41190624e-01 4.41656262e-01
-3.01841587e-01 -2.41894007e-01 2.90320039e-01 -7.73792267e-02
9.66603279e-01 2.63137251e-01 -6.39339626e-01 2.89319128e-01
-8.28095078e-01 1.14731245e-01 3.49109113e-01 5.99001646e-01
6.54998958e-01 -3.77595395e-01 -2.26070538e-01 1.15244544e+00
4.30825680e-01 3.31802845e-01 7.04693615e-01 -6.77103460e-01
2.56069660e-01 9.25715864e-01 -2.06539020e-01 -5.33669710e-01
-8.02765787e-01 -5.57132006e-01 -9.49170053e-01 2.54686505e-01
4.14452314e-01 -1.65197194e-01 -1.64520407e+00 1.57442331e+00
4.87075061e-01 3.49654585e-01 -7.87574574e-02 9.66682792e-01
1.22431505e+00 3.98134381e-01 1.22261107e-01 -1.40593112e-01
1.11899328e+00 -8.03044677e-01 -5.66268325e-01 -2.79592693e-01
8.37912977e-01 -2.92679876e-01 7.39098787e-01 2.57444888e-01
-1.00073934e+00 -6.80335820e-01 -1.10541189e+00 -9.82536450e-02
-3.41923058e-01 8.35927203e-03 6.45283759e-01 4.44072247e-01
-9.13789451e-01 4.06956404e-01 -1.15733111e+00 -4.31852937e-02
1.04521728e+00 5.38720727e-01 -3.79735589e-01 8.80279392e-03
-1.00123847e+00 1.08327484e+00 4.41779584e-01 1.80366740e-01
-8.44069898e-01 -8.69087696e-01 -7.62149274e-01 -4.87779379e-01
2.16985777e-01 -8.24981749e-01 1.10732567e+00 -6.37757719e-01
-1.42085910e+00 1.44817019e+00 3.84761766e-02 -3.73076946e-01
5.33404827e-01 -6.30999729e-02 -9.50648263e-03 2.63107300e-01
7.58925006e-02 9.86791611e-01 5.83608508e-01 -1.48798382e+00
-4.72854316e-01 -8.08027685e-01 -8.49662814e-03 6.67877346e-02
-8.11067894e-02 -3.35501224e-01 -6.96525693e-01 -5.48069775e-01
4.87182736e-01 -8.89790833e-01 -6.11513793e-01 1.20917866e-02
-4.90531892e-01 -3.22153687e-01 7.04340816e-01 -5.25793374e-01
7.21210957e-01 -2.03858972e+00 4.05167758e-01 6.15713075e-02
5.41576743e-01 7.22203910e-01 -1.91035330e-01 -3.55118543e-01
5.31025715e-02 2.56966650e-01 -6.31286979e-01 -3.98202807e-01
-2.65713573e-01 2.29862854e-01 2.71422088e-01 4.06543046e-01
3.02301347e-01 1.11876249e+00 -8.56070757e-01 -7.49517024e-01
3.64515364e-01 5.20600379e-01 -4.42845494e-01 1.67433470e-01
-1.23071007e-01 9.89671469e-01 -7.76006639e-01 6.22943461e-01
7.30315208e-01 -5.08759499e-01 1.66341476e-02 -2.82867372e-01
1.48427874e-01 -6.31017238e-02 -6.64302051e-01 2.08951688e+00
-3.64446729e-01 1.25401407e-01 -1.42151311e-01 -1.61273730e+00
9.57212865e-01 2.84386903e-01 8.41578662e-01 -7.75576890e-01
4.49664444e-01 3.95999551e-01 -1.32633418e-01 -6.53174341e-01
-3.08393836e-01 -3.83075088e-01 4.38412093e-02 3.29174191e-01
4.47146118e-01 -3.60160977e-01 -1.80986468e-02 -1.05403066e-01
7.28545129e-01 4.79020894e-01 1.71241328e-01 -6.05266877e-02
7.05886126e-01 -1.76019877e-01 6.83967710e-01 5.32229662e-01
-5.42883992e-01 5.97166002e-01 5.90324044e-01 -4.51570392e-01
-7.69215107e-01 -1.00106168e+00 -3.50641072e-01 6.06234670e-01
4.29819643e-01 1.32021382e-01 -1.05469203e+00 -1.17112291e+00
4.60507050e-02 3.02560925e-01 -8.07593763e-01 -1.88590705e-01
-5.72095692e-01 -1.29553640e+00 6.71686530e-01 6.77721024e-01
7.55326033e-01 -7.95577824e-01 -4.23698336e-01 2.20160425e-01
-2.56869465e-01 -1.29152608e+00 -2.26343930e-01 3.79328430e-01
-1.30683923e+00 -1.03310502e+00 -1.16641331e+00 -1.02516985e+00
8.45288157e-01 2.88224280e-01 8.15765500e-01 1.82459965e-01
-2.48333618e-01 2.42355064e-01 -2.40308329e-01 -7.15117693e-01
-3.29932272e-01 1.03315704e-01 -1.41461834e-01 -2.54893780e-01
4.01339561e-01 -4.69733149e-01 -5.00908613e-01 2.48214945e-01
-7.01073110e-01 2.50080347e-01 7.15374112e-01 8.87041867e-01
1.17621386e+00 -1.08358264e-01 8.10883760e-01 -1.08777750e+00
4.10938799e-01 -2.86107093e-01 -4.19172198e-01 2.23421782e-01
-5.86664677e-01 -4.89865150e-03 2.84981817e-01 -4.89964217e-01
-1.06568372e+00 2.52779543e-01 -5.59612274e-01 -5.32370389e-01
-3.56278598e-01 4.71068203e-01 -1.60985112e-01 -1.53861105e-01
5.62094986e-01 1.76650941e-01 3.91251057e-01 -6.96078300e-01
3.11227292e-01 7.37200201e-01 4.61818367e-01 -4.92715448e-01
6.13720000e-01 6.20378315e-01 -8.80057141e-02 -6.24405682e-01
-1.38203239e+00 -4.50451106e-01 -1.07072675e+00 -1.20849550e-01
1.28450501e+00 -7.90254116e-01 -4.41820413e-01 9.42922711e-01
-9.59566176e-01 -6.36182666e-01 1.44558130e-02 5.33833683e-01
-6.16223872e-01 4.56239402e-01 -6.40887797e-01 -5.29424310e-01
-4.12480175e-01 -1.52631414e+00 1.28924727e+00 3.04972440e-01
5.38338013e-02 -1.25476682e+00 2.40027998e-02 6.08344734e-01
6.62422413e-03 4.87577528e-01 1.08194435e+00 -8.68307114e-01
-2.45090216e-01 -2.14149252e-01 -2.39930749e-01 6.48136556e-01
3.19400877e-01 -3.97675872e-01 -9.61742818e-01 -1.27245978e-01
5.06019779e-02 -4.11405176e-01 1.12353373e+00 8.06244969e-01
1.50661325e+00 3.08761895e-01 -6.30568802e-01 7.65293777e-01
1.14220643e+00 1.40685946e-01 4.71806496e-01 2.67342895e-01
7.33418643e-01 5.04742265e-01 3.80817711e-01 -5.31261824e-02
5.30901074e-01 4.67232168e-01 4.43954110e-01 -5.32222390e-01
-1.06165394e-01 -1.44249812e-01 -1.79393128e-01 7.83122003e-01
-1.31460235e-01 9.86749604e-02 -8.89820933e-01 3.89243782e-01
-1.83472478e+00 -6.68907225e-01 4.59783040e-02 1.91122925e+00
1.05049598e+00 3.69151473e-01 3.37676704e-01 1.63400128e-01
6.15301073e-01 -4.12605032e-02 -8.53773177e-01 3.97036485e-02
-1.76464580e-02 3.44719499e-01 3.83821785e-01 1.81797996e-01
-1.55083930e+00 8.66893291e-01 6.65853643e+00 6.71158135e-01
-1.25003362e+00 -4.55201976e-03 9.63300049e-01 1.63113981e-01
-1.90867066e-01 -4.75599140e-01 -7.51126170e-01 4.42147255e-01
5.27115881e-01 2.32505277e-01 6.48223311e-02 6.11088276e-01
2.01174587e-01 6.42372221e-02 -1.04322851e+00 1.01914227e+00
1.06367975e-01 -1.41444671e+00 -6.06758185e-02 1.94771469e-01
7.82616198e-01 1.96340561e-01 9.42970160e-03 3.85368615e-01
1.15502492e-01 -1.05348313e+00 4.20545787e-01 4.84609663e-01
5.79957306e-01 -5.49824119e-01 9.58574772e-01 7.06384301e-01
-7.02240705e-01 1.33709475e-01 -2.22387791e-01 4.21640337e-01
2.49017954e-01 7.48253286e-01 -5.72011471e-01 6.01223290e-01
7.14831054e-01 8.22553098e-01 -4.91928965e-01 1.07280576e+00
-2.93673515e-01 3.84879798e-01 -1.36979505e-01 1.01563700e-01
2.92385548e-01 -1.52239695e-01 3.46209347e-01 1.07210612e+00
-2.63079107e-02 3.10987025e-01 3.43137771e-01 9.53753114e-01
-2.56876886e-01 1.76413357e-02 -3.39411587e-01 -2.31538899e-02
6.82012737e-02 1.27491808e+00 -9.90024745e-01 -3.19710016e-01
-4.57734346e-01 8.43158484e-01 3.81214291e-01 3.23357701e-01
-7.81368732e-01 -2.36146331e-01 4.52259004e-01 -2.29939118e-01
1.83938816e-01 -1.58511877e-01 -6.11377954e-01 -1.06024468e+00
-2.85188973e-01 -7.16706693e-01 3.83164763e-01 -4.96401280e-01
-1.44445944e+00 4.90504622e-01 5.61189204e-02 -1.05802131e+00
1.51702911e-01 -8.69204044e-01 -4.06767309e-01 7.93942034e-01
-1.71264303e+00 -1.12640679e+00 -1.73708767e-01 4.20113593e-01
2.96728998e-01 -4.76326831e-02 7.18808949e-01 3.20427150e-01
-5.61256409e-01 4.17309701e-01 -2.60020733e-01 5.07269025e-01
5.72432756e-01 -1.27907073e+00 2.59521127e-01 4.45449144e-01
7.84212425e-02 4.41145480e-01 1.19478889e-01 -4.70578760e-01
-1.10688639e+00 -1.14095032e+00 4.13882941e-01 -3.93035710e-01
1.56480014e-01 -1.30117200e-02 -1.02760005e+00 5.77972591e-01
-3.09196532e-01 3.30566853e-01 9.31915998e-01 1.80481195e-01
-1.08897105e-01 4.67272885e-02 -1.35001647e+00 4.30272937e-01
1.05669749e+00 -3.49020630e-01 -7.89120436e-01 4.03989255e-01
8.58998597e-01 -7.77202249e-01 -1.01652384e+00 7.95954108e-01
4.43784744e-01 -8.26236546e-01 1.13755679e+00 -6.57357454e-01
3.74600053e-01 -2.02707365e-01 1.64417028e-01 -1.42050779e+00
-2.23087177e-01 -1.27336487e-01 -1.42812207e-01 8.59524369e-01
4.26793367e-01 -4.99725103e-01 8.59427452e-01 2.18902498e-01
-5.12824416e-01 -1.25800657e+00 -8.90083134e-01 -6.92768216e-01
5.73335230e-01 -4.40085292e-01 5.57856977e-01 8.40309203e-01
-6.77062571e-02 3.65944743e-01 6.10824116e-03 9.96053219e-03
8.29327643e-01 9.40431282e-02 6.01262808e-01 -1.37702692e+00
4.45884913e-02 -6.06609106e-01 -4.50630844e-01 -1.34505320e+00
4.39384818e-01 -1.35313451e+00 1.76052283e-02 -1.88489604e+00
3.37019473e-01 -5.81497073e-01 -5.64836085e-01 7.14987099e-01
-4.03731465e-01 5.41751564e-01 -6.82048351e-02 1.27254143e-01
-5.26153445e-01 3.85508835e-01 1.90790272e+00 -3.42345834e-01
-4.69050467e-01 1.12238392e-01 -8.26871753e-01 8.91069591e-01
8.20306301e-01 -2.20327750e-01 -3.60030085e-01 -6.23258173e-01
-4.01291460e-01 -4.49803956e-02 4.52782720e-01 -6.78879857e-01
5.75268604e-02 9.49642137e-02 7.27349579e-01 -8.22909236e-01
1.66377023e-01 -5.56295693e-01 -4.46637452e-01 4.39757675e-01
-3.31307352e-01 -7.00296044e-01 2.57450700e-01 4.54187781e-01
-2.52635144e-02 -1.34377897e-01 1.14123440e+00 -3.53189766e-01
-4.71787184e-01 7.81110525e-01 -2.13271946e-01 1.93554446e-01
1.05944943e+00 -3.60614866e-01 -3.95587124e-02 1.19700469e-01
-8.43493760e-01 4.81688470e-01 1.99338615e-01 4.65243995e-01
6.04938030e-01 -1.25575233e+00 -6.32878780e-01 1.64056808e-01
4.92924601e-02 4.77239817e-01 2.56161600e-01 1.02645195e+00
-3.42589974e-01 3.53168458e-01 -1.81273833e-01 -1.07520390e+00
-8.52610409e-01 3.96908522e-01 6.22268617e-01 -2.85489202e-01
-7.08092213e-01 8.91255558e-01 2.55102038e-01 -8.45520318e-01
1.93550736e-01 -4.74966168e-01 -3.24215591e-01 -4.94946428e-02
4.18212712e-01 5.08181425e-03 1.36712283e-01 -4.93141443e-01
-4.58138406e-01 7.44508803e-01 -1.91779062e-01 2.10127130e-01
1.52513075e+00 7.02121109e-03 2.69348845e-02 4.94565010e-01
1.40080273e+00 -6.02358937e-01 -1.33837843e+00 -3.95359278e-01
-6.97931722e-02 -6.12793528e-02 3.50880325e-01 -8.39955986e-01
-1.33427310e+00 1.06111765e+00 6.32242203e-01 -5.88854887e-02
1.19285858e+00 1.83282629e-01 1.06772912e+00 1.04293853e-01
3.41241777e-01 -1.06217790e+00 3.13117579e-02 3.17398667e-01
3.98384750e-01 -1.55092454e+00 4.50122841e-02 -3.33126634e-01
-5.59134841e-01 9.69565988e-01 7.33985662e-01 -1.38644785e-01
8.60750020e-01 2.78958797e-01 3.47374558e-01 -1.12142719e-01
-3.37974042e-01 -3.53032768e-01 4.77203012e-01 7.99396634e-01
4.53441203e-01 -6.53115213e-02 -2.98108369e-01 8.05741847e-01
2.39866570e-01 2.61005789e-01 2.76357215e-03 8.09670508e-01
-4.73416150e-01 -1.25989127e+00 -1.26579404e-01 5.94460785e-01
-5.49691677e-01 2.60747343e-01 -2.73395538e-01 7.73178041e-01
3.53363544e-01 6.36233091e-01 -1.93788335e-02 -4.03652459e-01
3.16943914e-01 5.60526699e-02 8.84076297e-01 -6.55702949e-01
-3.72786492e-01 3.69548410e-01 -3.75189334e-01 -3.51112038e-01
-8.29607904e-01 -3.35533768e-01 -1.66384852e+00 2.27918446e-01
-3.54048222e-01 -5.79693913e-02 6.37708366e-01 1.36236906e+00
2.04999179e-01 6.86993778e-01 5.64339995e-01 -1.00191116e+00
-4.40066218e-01 -9.07992125e-01 -4.86699820e-01 3.13977927e-01
2.80441046e-01 -8.50653470e-01 -2.19596654e-01 1.26102448e-01] | [14.625872611999512, -2.252593755722046] |
02a322b4-0322-47ef-8a68-8766163f2965 | new-methods-metrics-for-lfqa-tasks | 2112.13432 | null | https://arxiv.org/abs/2112.13432v1 | https://arxiv.org/pdf/2112.13432v1.pdf | New Methods & Metrics for LFQA tasks | Long-form question answering (LFQA) tasks require retrieving the documents pertinent to a query, using them to form a paragraph-length answer. Despite considerable progress in LFQA modeling, fundamental issues impede its progress: i) train/validation/test dataset overlap, ii) absence of automatic metrics and iii) generated answers not being "grounded" in retrieved documents. This work addresses every one these critical bottlenecks, contributing natural language inference/generation (NLI/NLG) methods and metrics that make significant strides to their alleviation. | ['Prasanna Kumar', 'Pablo Bertorello', 'Vladimir Blagojevic', 'Suchismit Mahapatra'] | 2021-12-26 | null | null | null | null | ['long-form-question-answering'] | ['natural-language-processing'] | [ 3.02572072e-01 3.52584213e-01 -1.43031821e-01 -4.44941431e-01
-1.73760712e+00 -1.07072437e+00 7.84031332e-01 3.28570902e-01
-3.34093094e-01 1.38653636e+00 5.85567176e-01 -8.40324521e-01
-3.37081999e-01 -8.08170319e-01 -6.49686694e-01 6.68163523e-02
2.64143199e-01 8.53566170e-01 3.05077970e-01 -4.28340822e-01
6.43474698e-01 3.96830887e-01 -1.35448360e+00 7.71221757e-01
1.24614275e+00 7.86437929e-01 -1.26786590e-01 1.11161709e+00
-1.00600564e+00 1.44949508e+00 -9.93833125e-01 -7.62703121e-01
-2.49833927e-01 -7.22862244e-01 -1.53020918e+00 -4.49159294e-01
7.33470440e-01 -3.33381414e-01 1.35344397e-02 5.63739896e-01
2.93142736e-01 1.22222118e-01 8.47858071e-01 -1.16474283e+00
-8.19827437e-01 1.49070412e-01 2.48875413e-02 5.43446839e-01
1.02491069e+00 8.13542530e-02 1.36159813e+00 -1.03040242e+00
7.89708972e-01 1.41588485e+00 3.83489758e-01 6.76799238e-01
-8.64220083e-01 -2.91110277e-01 -2.34305620e-01 1.64231569e-01
-1.08489954e+00 -4.48085308e-01 3.51439059e-01 -3.04929316e-01
1.59521210e+00 5.65171003e-01 3.62903535e-01 8.73868108e-01
4.10515517e-01 1.02560019e+00 8.90073121e-01 -7.04939783e-01
7.43562132e-02 1.93751454e-01 4.23976094e-01 5.42549431e-01
2.51218945e-01 -3.35459501e-01 -5.67763865e-01 -3.61748308e-01
2.99573421e-01 -7.33288646e-01 -1.03840485e-01 2.71507829e-01
-8.73890281e-01 9.87408161e-01 5.85217364e-02 3.26845497e-01
-3.93717557e-01 2.09640861e-02 3.09544742e-01 7.32055843e-01
2.77106136e-01 1.29169929e+00 -6.12523913e-01 -2.43961170e-01
-9.04915273e-01 9.26830590e-01 1.13998890e+00 9.67664659e-01
8.95793259e-01 -2.31460318e-01 -5.53114057e-01 6.03751063e-01
3.31469387e-01 5.26139915e-01 3.19759548e-01 -9.78359878e-01
9.59548295e-01 7.66717255e-01 4.39159125e-01 -1.08999562e+00
9.90773086e-03 -3.53059351e-01 -5.73326871e-02 -6.53042793e-01
4.03481960e-01 -2.60215282e-01 -5.58722377e-01 1.52045655e+00
1.62193984e-01 -4.86030430e-01 2.83615500e-01 5.29390514e-01
1.12058783e+00 9.84053612e-01 4.01403934e-01 -1.72306165e-01
1.29849613e+00 -8.21319103e-01 -9.06088412e-01 -6.43073380e-01
7.86886871e-01 -1.04724872e+00 1.29623520e+00 3.18022996e-01
-1.36625123e+00 -5.30790150e-01 -9.42969799e-01 -6.08021080e-01
-5.45731008e-01 -1.19084530e-01 4.66987520e-01 4.82128352e-01
-1.14273131e+00 5.33467298e-03 -7.73548335e-02 -1.10278502e-01
1.18960582e-01 2.39929467e-01 -1.46811247e-01 -4.88124818e-01
-1.39615142e+00 1.12768531e+00 1.47193640e-01 -4.04463820e-02
-6.71127617e-01 -9.43872571e-01 -6.63199842e-01 -1.40709087e-01
5.14164925e-01 -1.22358263e+00 1.64668429e+00 -4.84744966e-01
-9.58481789e-01 8.10424209e-01 -7.45993614e-01 -4.29669648e-01
1.43648550e-01 -5.89469016e-01 -5.25039852e-01 2.13779017e-01
4.56390828e-01 9.00890052e-01 3.66131246e-01 -1.16984499e+00
-7.32163608e-01 -3.22926670e-01 2.95132190e-01 2.65354931e-01
2.26106253e-02 2.53619641e-01 -1.84881076e-01 -5.01491845e-01
-9.53944772e-03 -4.94160414e-01 -3.15011181e-02 -6.14498138e-01
-2.11036906e-01 -1.07820702e+00 4.80888128e-01 -8.57743561e-01
1.53644204e+00 -1.33594859e+00 -1.88501090e-01 -1.65712222e-01
7.52397552e-02 3.72790486e-01 -4.46459293e-01 1.10147536e+00
4.35840130e-01 5.06711662e-01 1.07021704e-01 8.68504122e-02
1.21224225e-01 4.42111224e-01 -1.06195855e+00 -4.44122344e-01
7.13802040e-01 1.30369866e+00 -1.07675827e+00 -9.55898881e-01
-1.79155007e-01 1.91723853e-02 -5.21557152e-01 4.55316633e-01
-8.34700167e-01 2.32760862e-01 -9.92014110e-01 6.42958581e-01
2.20425874e-01 -8.70453715e-02 -3.41570377e-01 1.16979964e-01
-6.80460455e-03 1.15004647e+00 -8.09689641e-01 1.37306345e+00
-5.53754508e-01 5.44788003e-01 -1.77405447e-01 -4.87437248e-01
9.55080569e-01 4.66723859e-01 1.31615549e-02 -1.01448584e+00
-2.06311017e-01 5.63215733e-01 -3.18720669e-01 -1.05121791e+00
7.29723632e-01 -2.47414559e-01 -2.30913814e-02 6.37757480e-01
9.72966477e-02 -8.00639808e-01 7.24790692e-01 4.92969841e-01
1.26155794e+00 -5.59306405e-02 1.46714881e-01 -3.80839497e-01
7.77882159e-01 5.45413613e-01 2.30779678e-01 1.05722976e+00
8.86300132e-02 4.91445124e-01 4.24748003e-01 -4.72002923e-01
-8.74194741e-01 -9.83665168e-01 7.70460740e-02 1.00801241e+00
-4.04000610e-01 -5.43165267e-01 -6.43818855e-01 -8.14982772e-01
6.96891546e-02 1.15120375e+00 -5.22975266e-01 9.43030044e-02
-8.83131385e-01 -1.22650623e-01 8.16162109e-01 5.70806265e-01
5.56908213e-02 -1.36993468e+00 -5.94762802e-01 5.08856475e-01
-7.70744979e-01 -9.58403707e-01 -6.43204376e-02 -1.31581977e-01
-1.18772101e+00 -1.17059624e+00 -5.16858160e-01 -6.35121882e-01
4.54615533e-01 2.43692949e-01 1.97123945e+00 1.59609452e-01
-1.38628036e-01 6.54508948e-01 -4.25938398e-01 -6.50302589e-01
-6.17368042e-01 2.58679003e-01 -5.01837373e-01 -8.13869774e-01
7.64926612e-01 -1.61221609e-01 -6.15957677e-01 7.32399821e-02
-1.17351377e+00 -3.65502149e-01 8.16298664e-01 5.59550166e-01
5.58663666e-01 -3.89470577e-01 1.36721528e+00 -1.02684689e+00
1.51500762e+00 -6.29915357e-01 -1.87252074e-01 8.05130661e-01
-7.60152340e-01 1.37893796e-01 4.34835643e-01 6.89080581e-02
-1.11742520e+00 -6.40704215e-01 -5.94686627e-01 4.84299123e-01
-1.35749727e-01 8.47017288e-01 8.40846151e-02 4.11786079e-01
1.13757896e+00 6.84757307e-02 -7.18420371e-02 -3.52109492e-01
4.84459728e-01 4.87367421e-01 2.93824047e-01 -8.86469960e-01
5.69766581e-01 -5.12061492e-02 -3.78437668e-01 -7.64460444e-01
-1.32073593e+00 -5.60378909e-01 -3.78512800e-01 -4.41955924e-02
7.54618764e-01 -6.22959077e-01 -2.42060110e-01 -3.53840828e-01
-1.35427415e+00 -8.10337588e-02 -3.80020410e-01 1.92038581e-01
-3.08846384e-01 2.81755298e-01 -5.21990836e-01 -9.49628830e-01
-8.13515663e-01 -6.38280332e-01 8.23430002e-01 2.47417241e-01
-8.74179900e-01 -9.01144624e-01 2.71124005e-01 1.09839046e+00
4.69970405e-01 8.60334784e-02 1.43985772e+00 -5.61398506e-01
-5.02921164e-01 -4.79578912e-01 -1.48621053e-01 4.70614761e-01
-1.66751355e-01 -2.81494763e-02 -7.97326267e-01 3.26361544e-02
1.73002586e-01 -9.41243172e-01 4.31499809e-01 1.37256429e-01
8.20817471e-01 -7.26251423e-01 -2.80795526e-02 -3.58396024e-01
1.21987855e+00 2.81167954e-01 6.62951350e-01 1.93741724e-01
2.02941552e-01 9.42872465e-01 7.02825904e-01 -1.43058315e-01
5.26342034e-01 1.17705755e-01 3.28107591e-04 3.41891617e-01
-4.09725517e-01 -6.23256505e-01 2.70532548e-01 9.88226891e-01
4.70099598e-01 -4.30469841e-01 -1.07537115e+00 9.91203189e-01
-1.50804710e+00 -7.11243570e-01 -4.49297369e-01 1.80645144e+00
1.08662999e+00 9.83672962e-02 -2.47142687e-01 6.49662688e-02
-1.67315930e-01 9.72482190e-02 -4.61980343e-01 -7.17508197e-01
-2.72871315e-01 7.37519145e-01 -3.57323617e-01 8.72900784e-01
-4.38567638e-01 9.51625943e-01 7.31476259e+00 5.98321617e-01
-4.59943116e-01 -1.87688038e-01 4.75175768e-01 1.98741108e-01
-1.05193114e+00 1.02728009e-01 -8.01822722e-01 -2.28624731e-01
1.08402312e+00 -2.62232214e-01 2.39550397e-01 5.07745504e-01
4.06831391e-02 -2.68852353e-01 -1.14963400e+00 3.01692188e-01
2.97223538e-01 -1.31317079e+00 6.96784854e-01 -3.37047994e-01
7.43515193e-01 -1.21883146e-01 -4.78561223e-02 7.29412675e-01
2.11455807e-01 -1.24042070e+00 5.70082486e-01 6.32232308e-01
5.06230652e-01 -7.18541622e-01 8.57099831e-01 4.20090348e-01
-7.47283757e-01 -2.57254243e-02 -5.70912063e-01 -8.68217573e-02
1.75593480e-01 4.73280281e-01 -1.30493176e+00 6.19113922e-01
4.42707747e-01 -5.47604971e-02 -8.22731376e-01 6.06650472e-01
-5.88023484e-01 8.71168256e-01 -7.11930618e-02 -5.45663834e-01
6.53758585e-01 -1.59660071e-01 4.11075443e-01 1.17042232e+00
2.36111298e-01 4.15089875e-01 -2.77408749e-01 9.09985840e-01
5.04080020e-02 4.20508862e-01 -6.28487527e-01 -4.34993804e-01
5.53173363e-01 8.88474345e-01 -1.49674326e-01 -1.94471404e-01
-4.32388097e-01 5.24942458e-01 4.51738179e-01 4.58220243e-01
-1.79399997e-01 -4.35257971e-01 3.15052211e-01 1.11812077e-01
-1.40790462e-01 -1.80547714e-01 -3.55379045e-01 -7.89301932e-01
4.21365589e-01 -1.52232718e+00 8.42140377e-01 -9.56738770e-01
-1.46576095e+00 5.85530937e-01 4.05212305e-02 -5.97998679e-01
-8.32650840e-01 -4.88211721e-01 -5.04708767e-01 1.05104399e+00
-1.60900962e+00 -9.54228103e-01 8.72492939e-02 3.89850944e-01
7.45132923e-01 -2.51255073e-02 1.13755393e+00 3.33863050e-01
2.64288578e-02 4.72443968e-01 -5.02724111e-01 -1.41790316e-01
6.76504791e-01 -1.39889824e+00 2.63987958e-01 6.23747051e-01
3.37145120e-01 1.07084358e+00 7.15311348e-01 -6.14574969e-01
-1.60024607e+00 -6.68732643e-01 1.89035583e+00 -1.11335945e+00
5.95256984e-01 1.25470413e-02 -1.05906045e+00 6.03398681e-01
3.00639302e-01 -8.05223167e-01 1.03374791e+00 3.13986808e-01
-3.98972273e-01 -9.19623375e-02 -9.23014522e-01 6.50365233e-01
5.39951861e-01 -9.10953820e-01 -1.29280639e+00 3.93131196e-01
8.51530373e-01 -2.34565899e-01 -7.46333241e-01 4.96609420e-01
3.81209642e-01 -6.13389134e-01 8.18936408e-01 -1.10673451e+00
6.95989311e-01 -2.05546886e-01 -3.10008731e-02 -8.59291434e-01
1.36556819e-01 -6.58958733e-01 -2.94088274e-01 1.00899935e+00
1.03789127e+00 -4.49198455e-01 6.69825613e-01 9.71022546e-01
-2.76543647e-01 -1.12288249e+00 -9.44794953e-01 -3.12920183e-01
6.91269219e-01 -5.46403706e-01 7.25952148e-01 6.28054082e-01
-4.17189952e-03 9.69799340e-01 1.52079344e-01 -2.33442232e-01
-5.25879934e-02 3.35966051e-02 8.35386276e-01 -1.02325928e+00
1.07065126e-01 -4.40016747e-01 3.79489750e-01 -1.34384596e+00
-6.90491423e-02 -6.53725445e-01 -4.40062471e-02 -2.31155062e+00
4.80533428e-02 -3.77952844e-01 7.38309100e-02 3.50641757e-01
-4.64198917e-01 -1.26099452e-01 -2.12134406e-01 3.87359262e-02
-7.25986540e-01 4.00257319e-01 1.35125411e+00 -5.36730550e-02
1.85045049e-01 7.57668018e-02 -1.12479484e+00 2.77417302e-01
6.02302015e-01 -3.59673053e-01 -8.20810497e-01 -9.52296913e-01
9.87557709e-01 4.49286282e-01 1.42090544e-01 -7.63823688e-01
2.33673647e-01 -1.97487086e-01 2.24129438e-01 -8.59097540e-01
2.41110399e-01 -3.64034206e-01 -3.49936962e-01 3.47686023e-01
-6.87406898e-01 6.65113866e-01 4.35105741e-01 2.86156237e-01
-6.41154349e-01 -7.33653426e-01 1.16938822e-01 -2.77608126e-01
-3.25047702e-01 -5.13379872e-02 -5.18827796e-01 7.22199440e-01
5.33464849e-01 -8.03344399e-02 -6.13958240e-01 -7.59092748e-01
-1.69814795e-01 4.95642930e-01 -9.15191993e-02 5.44698238e-01
7.16222107e-01 -9.64985847e-01 -1.05665600e+00 -2.67704427e-01
3.14578474e-01 4.96181063e-02 1.22557744e-01 3.83245796e-01
-6.73505902e-01 1.28356826e+00 2.66109943e-01 -4.44767848e-02
-9.17628407e-01 2.40491346e-01 1.46302223e-01 -7.11134911e-01
-1.87305421e-01 1.01754951e+00 -2.18237743e-01 -6.65574610e-01
1.62748948e-01 -3.19566578e-01 -3.55260104e-01 2.31350794e-01
7.20506370e-01 3.15567076e-01 4.17091131e-01 -1.85867622e-01
-8.97205696e-02 1.47870854e-01 -2.04065591e-01 -5.30368984e-01
9.69572067e-01 4.89017442e-02 -3.59331757e-01 2.69818515e-01
1.13832188e+00 1.79050311e-01 -4.70853269e-01 -3.20513248e-01
5.25548995e-01 -2.09121451e-01 -1.75556347e-01 -1.22399974e+00
-1.74893409e-01 1.06555521e+00 1.79037661e-03 4.37843382e-01
8.58049095e-01 1.45806596e-01 1.30180061e+00 8.92556012e-01
1.28835350e-01 -1.21761525e+00 2.49724984e-01 9.03702736e-01
1.26919043e+00 -1.08395004e+00 -9.75115076e-02 -1.23850470e-02
-5.51137388e-01 1.14474320e+00 6.71843827e-01 2.03247860e-01
2.65525609e-01 -1.67527452e-01 4.33372080e-01 -6.54356420e-01
-1.08403134e+00 -8.91366825e-02 7.92301953e-01 3.49946022e-01
8.83351207e-01 -2.64802009e-01 -6.92481875e-01 4.81907636e-01
-5.72064459e-01 2.17144024e-02 1.34008199e-01 9.68447983e-01
-5.40492177e-01 -1.17496848e+00 -1.46313041e-01 7.19589055e-01
-6.74964547e-01 -2.06179217e-01 -9.71860230e-01 8.22495043e-01
-3.47193927e-01 1.58973539e+00 -1.97099328e-01 2.74289344e-02
5.08750796e-01 5.68995535e-01 4.85936075e-01 -8.50482047e-01
-8.77190292e-01 -4.98723775e-01 6.03356659e-01 -2.85236329e-01
-1.21315107e-01 -5.50738156e-01 -1.21756470e+00 -1.55560225e-01
-2.96277761e-01 7.84898579e-01 5.49141884e-01 1.20084596e+00
5.43334126e-01 3.21853906e-01 2.06670970e-01 3.64134014e-01
-7.94386208e-01 -1.27066362e+00 1.75753400e-01 2.59835333e-01
2.39848167e-01 1.08116446e-02 -3.25875759e-01 -8.88510123e-02] | [11.356440544128418, 8.075859069824219] |
aa6695e3-e77d-405b-841f-e395c717857f | inexpensive-cost-optimized-measurement | 1705.09879 | null | http://arxiv.org/abs/1705.09879v1 | http://arxiv.org/pdf/1705.09879v1.pdf | Inexpensive Cost-Optimized Measurement Proposal for Sequential Model-Based Diagnosis | In this work we present strategies for (optimal) measurement selection in
model-based sequential diagnosis. In particular, assuming a set of leading
diagnoses being given, we show how queries (sets of measurements) can be
computed and optimized along two dimensions: expected number of queries and
cost per query. By means of a suitable decoupling of two optimizations and a
clever search space reduction the computations are done without any inference
engine calls. For the full search space, we give a method requiring only a
polynomial number of inferences and guaranteeing query properties existing
methods cannot provide. Evaluation results using real-world problems indicate
that the new method computes (virtually) optimal queries instantly
independently of the size and complexity of the considered diagnosis problems. | ['Wolfgang Schmid', 'Konstantin Schekotihin', 'Patrick Rodler'] | 2017-05-28 | null | null | null | null | ['sequential-diagnosis'] | ['medical'] | [ 4.43559945e-01 5.87413609e-01 -6.51868209e-02 -3.36047560e-01
-1.12548053e+00 -4.62382048e-01 2.93072283e-01 6.55783534e-01
-3.71812195e-01 6.44660354e-01 -3.72318923e-01 -3.68210196e-01
-8.86863351e-01 -9.83746529e-01 -3.86786580e-01 -6.18544877e-01
-1.80622488e-01 1.38953185e+00 6.47477984e-01 1.64203793e-02
4.29530926e-02 7.39842653e-01 -1.61835921e+00 -9.98606458e-02
5.96733332e-01 1.08703780e+00 -1.20341033e-01 1.10577631e+00
2.66730785e-01 6.49291217e-01 -6.02125585e-01 -3.29293996e-01
2.27043629e-01 -4.47619110e-01 -1.28050900e+00 4.55890328e-01
2.58953813e-02 -3.40306282e-01 -1.01670109e-01 1.17000735e+00
5.61855614e-01 2.95922309e-02 2.26449803e-01 -1.14409077e+00
3.22126776e-01 8.61217320e-01 3.26518863e-01 2.51116633e-01
6.62863672e-01 -5.00961021e-02 8.55441570e-01 -4.62608010e-01
6.93231881e-01 1.12212110e+00 3.59859675e-01 2.66462386e-01
-1.54708719e+00 -5.87745719e-02 -1.49272174e-01 2.24959359e-01
-1.68498075e+00 -5.42168498e-01 2.00700581e-01 -2.71414459e-01
1.18356955e+00 1.03137064e+00 6.82951093e-01 3.79146606e-01
-4.61422559e-03 2.87820548e-01 9.00400281e-01 -5.59455216e-01
6.96094155e-01 4.06363517e-01 1.91507474e-01 8.63175631e-01
3.38286579e-01 3.06622267e-01 -1.35731339e-01 -6.38194084e-01
3.95960301e-01 1.02679515e-02 -3.58049609e-02 -2.61077106e-01
-1.26729584e+00 6.22395277e-01 -2.55308747e-01 3.05098116e-01
-3.18321347e-01 1.22981355e-01 3.44795793e-01 7.35691249e-01
2.92692304e-01 6.13149285e-01 -6.50725543e-01 1.57671586e-01
-7.64210999e-01 4.52765673e-01 1.26658750e+00 9.94428456e-01
5.49819350e-01 -3.58388752e-01 -2.85991073e-01 3.68034542e-01
-1.89265206e-01 9.40622449e-01 1.31782681e-01 -1.01437688e+00
2.56991744e-01 4.92877305e-01 3.79087448e-01 -9.35945094e-01
-7.16524005e-01 -4.05222446e-01 -5.49970746e-01 -2.22769797e-01
3.71312380e-01 -1.16208412e-01 -3.71577442e-01 1.68892336e+00
6.95106447e-01 8.56184587e-02 -6.82348087e-02 4.51637536e-01
-4.28543352e-02 7.01020733e-02 -6.45004332e-01 -8.04240108e-01
1.35788620e+00 -5.65429926e-01 -4.91909146e-01 2.10977271e-01
9.99260008e-01 -5.56235075e-01 5.66608191e-01 6.54595613e-01
-1.48125851e+00 -1.25392042e-02 -8.33941877e-01 2.29552746e-01
-1.70115441e-01 -9.68072563e-02 5.75395703e-01 6.95084214e-01
-1.29104853e+00 7.22714663e-01 -9.77523804e-01 -4.53598768e-01
-3.72895062e-01 8.12409818e-01 -1.48782670e-01 -1.76219836e-01
-9.03568268e-01 7.92307198e-01 5.12761593e-01 1.16093736e-02
-9.32545602e-01 -3.36807758e-01 -4.76121038e-01 1.70096174e-01
8.86684239e-01 -1.03956485e+00 1.43795967e+00 -1.55784652e-01
-1.42826176e+00 8.19215596e-01 -1.80602923e-01 -6.16153240e-01
5.59691727e-01 2.02283934e-01 -3.03830087e-01 5.63919067e-01
-8.82603973e-02 -7.76397735e-02 5.99526584e-01 -6.46966755e-01
-9.45896149e-01 -4.81177658e-01 5.51785469e-01 -4.00545448e-02
-1.32946193e-01 3.31711173e-02 -4.50465977e-01 1.86979577e-01
6.97239757e-01 -8.98705542e-01 -7.41804600e-01 -1.38862073e-01
-7.73522437e-01 -1.84995160e-01 3.52293104e-02 -4.41345125e-01
1.32889104e+00 -1.66197109e+00 4.98850971e-01 6.97717011e-01
4.28160191e-01 -3.26930821e-01 1.60812065e-01 7.28936911e-01
1.03607692e-01 -7.38525540e-02 -2.88928866e-01 -4.06870306e-01
2.01691806e-01 4.34829712e-01 -9.99381691e-02 6.90196872e-01
-1.74764797e-01 6.21441126e-01 -8.33727419e-01 -8.65929842e-01
2.64563262e-01 -2.77567595e-01 -7.55095005e-01 3.06686878e-01
-4.21843469e-01 1.87788621e-01 -6.21027887e-01 7.75495231e-01
4.19966191e-01 -5.35176456e-01 4.91468400e-01 7.29873031e-02
3.54884446e-01 4.36124802e-01 -1.60893738e+00 1.24660373e+00
-5.07381022e-01 -2.20520064e-01 1.21426500e-01 -1.23048079e+00
4.84577209e-01 5.48530817e-01 6.38332069e-01 -3.15563023e-01
3.18196267e-01 4.09133345e-01 -2.34590515e-01 -7.15653896e-01
7.07181990e-02 -1.13771975e-01 -3.86771828e-01 5.85334301e-01
-6.82015494e-02 -2.07585931e-01 3.16842318e-01 3.16425157e-03
1.53912401e+00 -8.28960121e-01 5.65957665e-01 -3.69786114e-01
6.08435452e-01 2.42448412e-02 5.70227504e-01 1.00125372e+00
3.20948541e-01 1.81766778e-01 6.70679450e-01 -6.18917383e-02
-8.57927442e-01 -7.79129982e-01 -1.07982032e-01 8.21874321e-01
-3.93258072e-02 -4.41102594e-01 -7.64534593e-01 -5.13982952e-01
-4.45065387e-02 6.22024655e-01 -5.72628796e-01 -3.06095909e-02
-4.34066623e-01 -5.56449950e-01 4.25323606e-01 8.65705386e-02
-2.38658279e-01 -5.73067188e-01 -8.63347530e-01 2.36228630e-01
3.73599231e-02 -1.21529138e+00 -1.00821540e-01 5.07823862e-02
-1.16288650e+00 -1.34843397e+00 -2.54373252e-02 -2.78467834e-01
8.69814038e-01 -1.17677823e-01 1.12619150e+00 3.76068473e-01
-1.96232259e-01 4.85686243e-01 -1.01570981e-02 -1.96636692e-01
-5.80307305e-01 3.81705575e-02 9.98776313e-03 -2.00242952e-01
-1.54880807e-02 -6.28191590e-01 -1.58432394e-01 6.29966259e-02
-9.69445586e-01 -1.51748896e-01 5.11949658e-01 9.74022269e-01
8.87798429e-01 4.60401535e-01 2.24049073e-02 -1.33526778e+00
4.34104979e-01 -4.54196692e-01 -1.35018516e+00 4.60434169e-01
-1.09358132e+00 4.24980313e-01 4.44410771e-01 -2.93250889e-01
-4.94121671e-01 2.46408477e-01 -1.15341842e-01 -2.32668072e-01
-1.99260399e-01 6.60421550e-01 1.65799707e-02 -5.21384366e-02
4.84889656e-01 1.43808872e-01 -6.16731234e-02 -4.70862120e-01
1.98143616e-01 3.58648300e-01 3.64350259e-01 -4.99588698e-01
5.88450313e-01 4.50190485e-01 7.59940207e-01 -4.81977701e-01
-6.70576215e-01 -4.52650398e-01 -3.98344785e-01 8.58474076e-02
3.07991594e-01 -2.57632554e-01 -1.14983332e+00 -1.31552771e-01
-8.93500090e-01 6.74504861e-02 -6.66220546e-01 5.18245518e-01
-6.84326470e-01 3.68264705e-01 -5.46081662e-01 -1.15673614e+00
-2.54700035e-01 -1.36800539e+00 1.24068308e+00 -4.09119517e-01
-2.67078280e-01 -9.71708536e-01 1.77303314e-01 8.37796405e-02
1.96090758e-01 2.69723147e-01 1.09729898e+00 -1.01048899e+00
-8.20374727e-01 -7.47053802e-01 -3.00792307e-02 -2.14234307e-01
-1.78303167e-01 -2.88876414e-01 -8.28398168e-01 -3.33729655e-01
3.83226395e-01 5.76974154e-02 2.32452467e-01 1.29863903e-01
1.21261013e+00 -7.30958521e-01 -6.49309039e-01 3.42800170e-01
1.58097243e+00 -5.39305694e-02 1.39532015e-01 -1.00575753e-01
2.28594422e-01 5.77471077e-01 6.96163177e-01 6.51571751e-01
2.11676642e-01 7.71544158e-01 4.55771565e-01 2.67828614e-01
3.32360923e-01 2.00865403e-01 -2.47062892e-01 9.24727738e-01
1.67514458e-01 -1.88962892e-01 -7.46205986e-01 7.34615326e-01
-1.92067075e+00 -6.02102637e-01 2.10579202e-01 2.68003392e+00
9.22693074e-01 2.28547513e-01 2.51726210e-01 6.00457132e-01
3.66344959e-01 -4.66286600e-01 -4.89987403e-01 -3.83938432e-01
1.45439088e-01 4.30408001e-01 7.54804611e-01 1.03265011e+00
-4.96324420e-01 1.87317386e-01 7.66923094e+00 7.52669871e-01
-7.80963957e-01 3.85803759e-01 1.84770435e-01 -3.26829225e-01
-5.09182870e-01 1.32157683e-01 -6.91526890e-01 1.52514935e-01
1.57507992e+00 -3.04091156e-01 5.51558256e-01 9.01440024e-01
-9.12508741e-02 -2.39827722e-01 -1.61380422e+00 9.23661232e-01
-3.54011267e-01 -1.29023015e+00 -4.38743025e-01 1.11572474e-01
4.17003036e-01 -2.64611244e-01 -3.19782555e-01 1.69794615e-02
2.32019678e-01 -7.12256253e-01 5.55852830e-01 6.41618967e-01
6.39830053e-01 -6.14498258e-01 1.02986133e+00 6.59433782e-01
-9.17383969e-01 -4.55290228e-01 -2.49754697e-01 -8.58641639e-02
1.11192115e-01 9.91395295e-01 -1.03826749e+00 8.48956585e-01
3.41440231e-01 -4.70749401e-02 -2.80340642e-01 1.04846191e+00
1.36757493e-02 5.88261008e-01 -9.03363347e-01 -1.58484459e-01
-2.32300535e-01 3.38107720e-02 6.12210989e-01 1.11770082e+00
4.77331847e-01 2.18787283e-01 4.30222839e-01 4.91604567e-01
2.79738128e-01 -7.03091696e-02 -4.02388245e-01 6.26582429e-02
9.50257301e-01 8.04699063e-01 -7.87054658e-01 -6.26471579e-01
6.42752321e-03 6.79332137e-01 2.36476138e-01 5.85295856e-02
-6.14828467e-01 -8.04225132e-02 3.46235245e-01 1.21343024e-01
1.11510701e-01 1.90535635e-01 -2.93231934e-01 -9.08793151e-01
1.18872806e-01 -8.48684371e-01 8.01750958e-01 -2.27399781e-01
-8.89923334e-01 4.02823091e-01 4.60446477e-01 -8.97629619e-01
-8.85010839e-01 -2.32514456e-01 -5.59475347e-02 5.91852427e-01
-9.05558825e-01 -5.29515743e-01 2.16034487e-01 5.15787780e-01
9.19214711e-02 3.09337556e-01 1.23174667e+00 2.75559455e-01
-5.76190591e-01 5.61293006e-01 3.33142690e-02 -6.11406565e-01
1.48606226e-01 -1.56135595e+00 -7.10156783e-02 8.92703354e-01
-1.10622115e-01 4.13849533e-01 1.28340983e+00 -2.85248578e-01
-1.71517980e+00 -7.75504708e-01 1.44640911e+00 -2.75377840e-01
5.75013399e-01 -1.61719173e-01 -5.12636781e-01 7.41781592e-01
-1.99996263e-01 -8.71879458e-02 3.16407710e-01 1.35584205e-01
1.69045001e-01 -5.37840605e-01 -1.51289654e+00 2.86315650e-01
9.18470085e-01 -6.60337329e-01 -2.54844874e-01 1.23588693e+00
8.85944247e-01 -6.94061518e-01 -1.16925097e+00 3.99664670e-01
2.45933294e-01 -8.24953258e-01 9.48097885e-01 -5.56125164e-01
-4.53028440e-01 -1.03329554e-01 -2.53674507e-01 -6.12370193e-01
-1.53192893e-01 -9.81915176e-01 -3.72793913e-01 5.57797909e-01
4.64630753e-01 -1.06183112e+00 6.61429584e-01 7.77013659e-01
1.96342990e-01 -1.14265811e+00 -1.14358175e+00 -7.37259328e-01
-4.73271757e-01 -6.98915780e-01 1.02703357e+00 5.39353907e-01
2.38727912e-01 1.61360040e-01 -1.79912001e-01 8.50444436e-01
7.20306635e-01 3.52068692e-01 4.50571477e-01 -1.32647383e+00
-1.13336968e+00 -3.38608742e-01 -4.28207666e-01 -7.46665239e-01
-3.70647937e-01 -5.43752015e-01 6.91456795e-02 -1.13563001e+00
2.23611191e-01 -5.91070473e-01 -3.04313481e-01 3.53879899e-01
1.67702243e-01 -4.38995928e-01 -2.37867415e-01 -1.52636860e-02
-7.78545856e-01 -1.99809864e-01 7.47057557e-01 1.57421120e-02
6.99102730e-02 5.90044022e-01 -2.51744330e-01 5.19677341e-01
6.46043003e-01 -8.50711524e-01 -7.22248793e-01 -1.85213283e-01
7.26956189e-01 1.02233970e+00 1.97124794e-01 -8.96748722e-01
4.58295614e-01 -3.60882312e-01 -3.93022329e-01 -6.63866162e-01
5.40190578e-01 -8.61327112e-01 5.82619667e-01 9.39858079e-01
-5.17428696e-01 3.05885643e-01 -2.64152497e-01 5.22661567e-01
3.69614772e-02 -7.12062001e-01 4.39474493e-01 7.48569593e-02
-3.57169136e-02 3.36612523e-01 -2.20795855e-01 -2.14777887e-01
1.01410186e+00 1.86437353e-01 1.81580886e-01 -3.69450182e-01
-1.34693158e+00 3.42020482e-01 2.41133481e-01 -4.80093867e-01
5.12444556e-01 -8.66259933e-01 -5.06875575e-01 -2.57880222e-02
1.12322949e-01 4.95817959e-02 1.85467035e-01 1.34618700e+00
-3.52017701e-01 7.81752408e-01 5.01914859e-01 -6.59173787e-01
-1.46433973e+00 9.46144044e-01 4.36486304e-01 -8.22990179e-01
-4.48784649e-01 8.40852678e-01 -3.08627069e-01 -5.78915417e-01
2.74869949e-01 -6.81752264e-01 3.05070370e-01 -2.50187218e-01
5.83462536e-01 8.09977889e-01 3.39997023e-01 2.71323659e-02
-3.22034121e-01 2.17677116e-01 1.33853629e-01 -3.28642279e-01
1.05714941e+00 -4.37225074e-01 -5.08102715e-01 4.13882524e-01
9.47008729e-01 -4.80266437e-02 -4.34562027e-01 -4.59028006e-01
2.83199400e-01 -2.29162499e-01 -1.03139952e-01 -5.79693615e-01
-8.85181189e-01 6.06993377e-01 4.34787899e-01 8.63668740e-01
1.51953340e+00 1.84147790e-01 5.64539433e-01 9.39369559e-01
8.84021819e-01 -8.59603763e-01 -5.77840269e-01 1.61635652e-02
6.27772272e-01 -6.71031892e-01 1.11511558e-01 -6.97683215e-01
1.79993391e-01 1.00575924e+00 -4.47442345e-02 -1.02674261e-01
7.15559244e-01 5.36148310e-01 -4.86509234e-01 -4.10009563e-01
-1.18864179e+00 -2.95959681e-01 -9.31249000e-03 3.66091222e-01
-3.78586560e-01 3.10332090e-01 -5.27271450e-01 4.56907272e-01
-2.60786712e-01 -2.70246137e-02 4.46895093e-01 9.75564063e-01
-5.52415848e-01 -1.25171733e+00 -4.87575710e-01 6.34595037e-01
-4.99256730e-01 1.65719047e-01 -1.17047489e-01 7.01640487e-01
3.81984450e-02 1.13096762e+00 -3.04744840e-01 -3.79999399e-01
3.71374816e-01 -2.13733446e-02 6.87480927e-01 -8.10321450e-01
-3.32362205e-01 7.16417879e-02 5.11449456e-01 -1.02276862e+00
-2.55645394e-01 -6.19590580e-01 -1.18654001e+00 -4.05951887e-01
-6.11160100e-01 3.07250649e-01 3.15063059e-01 1.42209959e+00
2.79857665e-01 5.42914450e-01 5.43843746e-01 -1.71982884e-01
-1.40412617e+00 -8.29144359e-01 -8.16330194e-01 -1.26907855e-01
1.61780760e-01 -3.80087703e-01 -3.32527548e-01 -2.82876611e-01] | [5.461527347564697, 2.7795939445495605] |
621942d2-a948-42f9-9af5-063edad2aefc | auxiliary-learning-induced-graph | null | null | https://openreview.net/forum?id=9QffERDO_rJ | https://openreview.net/pdf?id=9QffERDO_rJ | Auxiliary learning induced graph convolutional networks | In this article, we propose a novel auxiliary learning induced graph convolutional network in a multi-task fashion. Specifically, both the link prediction and pseudo label generation are used as two auxiliary tasks to complement the primary task of node classification. Those two auxiliary tasks are jointly trained with the primary node classification task via the graph meta-learning strategy. The experimental results demonstrate that the proposed method consistently and significantly outperforms the existing methods and achieves state-of-the-art node classification results on the benchmark citation network datasets. | ['Yongjian Wu', 'Feiyue Huang', 'Baochang Zhang', 'Ling Shao', 'Rongrong Ji', 'Taisong Jin', 'Gengchen Duan'] | 2021-05-21 | null | null | null | neurips-2021-12 | ['auxiliary-learning'] | ['methodology'] | [ 2.73369640e-01 5.25142491e-01 -8.04119170e-01 -9.69640017e-02
-4.77643430e-01 -2.30373308e-01 8.88556480e-01 3.87258351e-01
-1.76847294e-01 9.49133992e-01 -1.02482982e-01 -7.21504748e-01
-6.41471818e-02 -8.16863060e-01 -6.69911802e-01 -6.67667449e-01
-3.14359009e-01 5.84118783e-01 2.15815023e-01 -4.13704999e-02
6.17244579e-02 3.07391316e-01 -1.05224633e+00 -7.82497749e-02
9.69534874e-01 8.75695705e-01 -8.14222768e-02 6.19836330e-01
-3.15100729e-01 1.18381512e+00 -2.56765276e-01 -5.74089587e-01
-1.24231532e-01 -4.70550507e-02 -1.05953574e+00 -2.43890047e-01
2.04071417e-01 1.11673633e-02 -8.16408932e-01 9.27823544e-01
1.83646947e-01 8.08206201e-03 6.42120481e-01 -1.65443075e+00
-7.65100598e-01 8.69323134e-01 -6.18626118e-01 2.88650811e-01
7.12827370e-02 -2.94909537e-01 1.38837647e+00 -6.61363184e-01
7.67198086e-01 1.14545083e+00 7.05287993e-01 2.83431500e-01
-1.25028574e+00 -9.87791061e-01 3.10495764e-01 2.73147374e-01
-1.39694655e+00 -1.54555202e-01 1.25310707e+00 -3.20339024e-01
6.28848791e-01 -2.74519593e-01 3.39126080e-01 1.16416037e+00
2.25759566e-01 6.97324872e-01 8.47483814e-01 -1.29953668e-01
-2.81168848e-01 -8.85294303e-02 3.95207167e-01 1.16125596e+00
4.51172173e-01 -1.16269909e-01 -1.84054255e-01 -2.56810933e-01
6.90500081e-01 -2.70457659e-02 -2.25881323e-01 -2.29186267e-01
-1.28158367e+00 9.01534379e-01 7.23932564e-01 2.09460691e-01
-2.54121125e-01 5.23741424e-01 5.94619393e-01 2.96034485e-01
1.15879524e+00 1.34884268e-02 -5.44169307e-01 3.65317702e-01
-6.02922082e-01 -1.26534700e-01 1.01470399e+00 1.06237578e+00
8.78496468e-01 -1.17084824e-01 -5.70616066e-01 5.37089348e-01
7.59813786e-01 1.75234988e-01 -1.88029021e-01 -6.54517472e-01
5.45938134e-01 9.03319478e-01 -5.49535871e-01 -9.14759159e-01
-6.56338573e-01 -1.25672841e+00 -1.16394734e+00 -1.68630064e-01
8.11617300e-02 -1.54625878e-01 -9.94700313e-01 1.78825665e+00
3.17339927e-01 7.59698689e-01 -2.56742742e-02 3.10749620e-01
1.80048728e+00 5.20451844e-01 4.74450171e-01 -2.31136352e-01
1.03775537e+00 -1.67203712e+00 -1.05128574e+00 -2.23449558e-01
1.03443599e+00 -5.01995504e-01 4.84070688e-01 -3.44266057e-01
-6.88591063e-01 -5.69834054e-01 -1.10349131e+00 -7.96286017e-02
-3.86278033e-01 1.85809776e-01 1.30677140e+00 6.26364574e-02
-1.13728499e+00 5.11696577e-01 -4.38829154e-01 -9.63882878e-02
7.84442961e-01 3.01393151e-01 -5.18830180e-01 8.02553073e-02
-1.29018593e+00 5.50307989e-01 5.85512400e-01 -3.74296010e-02
-8.88556480e-01 -7.15392947e-01 -9.05039668e-01 4.64907378e-01
2.75865644e-01 -7.70330906e-01 8.99290204e-01 -4.18488711e-01
-1.10341251e+00 1.13553107e+00 9.13513504e-05 -1.09563857e-01
2.38366783e-01 2.90694028e-01 -4.85355228e-01 2.93827057e-02
3.31437320e-01 6.23462558e-01 6.40699089e-01 -1.46377707e+00
-4.11825895e-01 -1.61463842e-01 -1.60451308e-02 -4.13456969e-02
-4.79808331e-01 -4.42744493e-01 -9.38272715e-01 -9.02017415e-01
-1.79129407e-01 -8.08454037e-01 -1.37857229e-01 3.15002562e-03
-8.54342639e-01 -8.79653931e-01 9.01020765e-01 -4.60695088e-01
1.20434833e+00 -1.66658211e+00 3.14769417e-01 3.57742429e-01
1.00389600e+00 2.32998490e-01 -5.77551603e-01 1.50187761e-01
-2.97966957e-01 2.87073374e-01 2.77227089e-02 -6.81030869e-01
-1.89206168e-01 -6.27735928e-02 2.21178338e-01 2.55511701e-01
6.11483259e-03 1.49565351e+00 -1.08525074e+00 -9.21996057e-01
-2.55700260e-01 3.03825498e-01 4.55187745e-02 3.88330072e-01
-1.43556952e-01 5.91208756e-01 -6.85518920e-01 7.01899946e-01
6.64542794e-01 -1.07968771e+00 3.81058216e-01 -3.26596081e-01
5.56711972e-01 2.33143255e-01 -5.67504644e-01 1.79810131e+00
-1.32534936e-01 4.86416936e-01 3.24253947e-01 -1.40800238e+00
9.77399647e-01 1.71181336e-01 5.36270976e-01 -4.84105974e-01
1.99948654e-01 1.94028988e-01 1.27820432e-01 -1.76950991e-01
-7.51882838e-03 2.93022662e-01 1.13576077e-01 5.64825714e-01
5.42127609e-01 3.61081064e-01 2.74602979e-01 7.30417490e-01
1.31427300e+00 7.59628043e-02 6.27287552e-02 -3.89279425e-01
7.05098867e-01 -3.81325215e-01 4.99585211e-01 6.54331446e-01
-1.65491402e-02 -1.06557071e-01 8.71285677e-01 -3.68414909e-01
-6.37063503e-01 -8.37559998e-01 1.83649912e-01 1.20198810e+00
2.86278486e-01 -4.80122268e-01 -4.22155738e-01 -1.40291560e+00
1.74595237e-01 2.49518380e-01 -9.98487592e-01 -3.30867499e-01
-3.14045131e-01 -6.76784575e-01 7.48191774e-01 5.90490937e-01
6.39278054e-01 -8.44766557e-01 6.14218354e-01 1.54993653e-01
1.38014436e-01 -1.31797040e+00 -4.73013580e-01 3.53724629e-01
-9.32209492e-01 -1.40546799e+00 -6.18213177e-01 -1.23446858e+00
7.99785137e-01 3.88970017e-01 1.39761734e+00 9.23972428e-01
-6.95172027e-02 1.10997878e-01 -3.71175289e-01 -3.97678725e-02
-2.56847888e-01 7.49694109e-01 -4.67554897e-01 -8.44706297e-02
1.04912333e-02 -7.38031924e-01 -4.77048188e-01 -2.31348336e-01
-5.35643578e-01 2.86041051e-01 8.60586584e-01 9.71273601e-01
4.13993776e-01 -1.46704301e-01 9.77321863e-01 -1.48993480e+00
8.98045659e-01 -8.32691371e-01 -3.57812643e-01 5.42128325e-01
-1.14639080e+00 3.20953429e-01 5.27832389e-01 -1.60915598e-01
-8.80020857e-01 -7.81429186e-02 -2.01152340e-02 -1.76808760e-01
3.25252265e-01 9.67636168e-01 -7.25135654e-02 -5.87144315e-01
9.93677005e-02 -1.19859532e-01 8.10201690e-02 -5.37579656e-01
3.85113239e-01 4.23456877e-01 3.85660827e-01 -3.43896359e-01
1.23900080e+00 1.84007779e-01 7.05834150e-01 -6.12600930e-02
-1.14191985e+00 -1.63089007e-01 -8.50815952e-01 -3.41271847e-01
7.46514261e-01 -9.67874646e-01 -6.75870240e-01 5.41033983e-01
-1.33362544e+00 -5.30960679e-01 2.15199918e-01 1.76120698e-01
8.14047977e-02 4.31631893e-01 -8.34796786e-01 -3.46585900e-01
-7.49056637e-01 -1.13365030e+00 9.56433475e-01 1.84931859e-01
5.20671964e-01 -1.70167708e+00 8.77278149e-02 4.14118171e-01
2.76525766e-01 3.29808146e-01 1.26550555e+00 -1.07525098e+00
-7.07286775e-01 -1.26362294e-01 -8.87561858e-01 -1.72616109e-01
3.80819499e-01 -4.22335565e-02 -6.83977365e-01 -4.66184258e-01
-1.08896899e+00 -4.38230097e-01 1.34330273e+00 -1.10918805e-02
1.38062441e+00 -1.33285567e-01 -1.06424677e+00 8.79424691e-01
1.12503684e+00 -1.81131005e-01 3.49672049e-01 1.94859639e-01
1.32375371e+00 4.01857018e-01 3.47064406e-01 -1.19085964e-02
8.61137152e-01 4.57011580e-01 7.01478481e-01 -6.64126873e-01
-5.15402317e-01 -3.13641965e-01 -3.24007452e-01 9.46399212e-01
3.20829488e-02 -4.66437995e-01 -1.00238705e+00 3.65763038e-01
-2.20877838e+00 -6.29273772e-01 -5.52858829e-01 1.54754078e+00
6.13775551e-01 3.47060859e-01 -2.12861657e-01 -2.09460214e-01
9.82719183e-01 5.20588696e-01 -6.22380137e-01 1.75266877e-01
3.36591080e-02 3.00201893e-01 5.20555794e-01 4.21505302e-01
-1.43893754e+00 1.00833237e+00 6.73738289e+00 7.93062031e-01
-6.62107587e-01 3.69456768e-01 7.78124750e-01 6.19984806e-01
-6.22497976e-01 1.33336201e-01 -5.01338363e-01 2.91870713e-01
6.85385466e-01 -1.40719980e-01 2.83805400e-01 6.48601651e-01
-3.52040887e-01 5.87420046e-01 -9.97093856e-01 8.19438457e-01
-8.42108857e-03 -1.69071758e+00 1.94875807e-01 2.33478233e-01
7.30901778e-01 2.49617636e-01 -2.52151370e-01 5.63954532e-01
4.77377832e-01 -1.19800532e+00 1.82589710e-01 5.17995417e-01
8.36606503e-01 -6.22989774e-01 9.88973439e-01 1.29970208e-01
-1.76653552e+00 1.76353306e-01 6.29292056e-02 -1.28403515e-01
-7.50777451e-03 6.56952500e-01 -3.93536240e-01 8.26276183e-01
3.16779345e-01 1.19871962e+00 -9.77822483e-01 8.73576641e-01
-5.09081542e-01 6.68891251e-01 1.00657754e-01 -1.37601167e-01
2.22944647e-01 -5.40489927e-02 4.87711638e-01 1.02764106e+00
-1.04309194e-01 -5.33991307e-02 5.29977262e-01 8.86305213e-01
-1.03188217e+00 -4.31282417e-04 -6.66502714e-01 -2.22565219e-01
7.45950997e-01 1.82077885e+00 -7.72482932e-01 -3.37333471e-01
-6.63910508e-01 7.43310750e-01 9.88649666e-01 4.89336848e-01
-8.04927409e-01 -6.97033465e-01 2.08008349e-01 -3.25955957e-01
1.16912492e-01 3.16863991e-02 -1.39712647e-01 -1.08370352e+00
-2.60740489e-01 -1.19466886e-01 6.06804788e-01 -4.28775370e-01
-1.67896366e+00 5.81105053e-01 -1.39275998e-01 -8.39870691e-01
1.09683342e-01 -5.84782124e-01 -1.26416874e+00 7.63664186e-01
-2.09044003e+00 -1.77450943e+00 -7.07769156e-01 4.06238526e-01
1.28513262e-01 -4.89475876e-01 8.53173435e-01 4.38050717e-01
-8.07828665e-01 9.38993037e-01 5.71346283e-02 5.73404491e-01
6.15013778e-01 -1.24565876e+00 5.45097351e-01 6.81260586e-01
-8.02132413e-02 2.35382378e-01 6.18394651e-02 -9.42246079e-01
-1.28962350e+00 -1.43918586e+00 6.82169139e-01 7.22565502e-02
8.40902984e-01 -4.00636256e-01 -9.59874988e-01 9.45682704e-01
3.00854534e-01 6.16399348e-01 7.19632804e-01 3.34701210e-01
-4.52000111e-01 -2.03123167e-01 -8.36492300e-01 2.07512677e-01
1.36688042e+00 -6.44637465e-01 -2.15238333e-01 5.81919968e-01
1.14330637e+00 -8.35529044e-02 -1.31824207e+00 9.20013070e-01
1.84325561e-01 -1.02622002e-01 1.01365483e+00 -8.49167109e-01
4.08444315e-01 -6.32628351e-02 4.88154650e-01 -1.33438349e+00
-8.80371690e-01 -4.38893765e-01 -8.26851249e-01 1.50876904e+00
7.41377532e-01 -7.60172844e-01 9.61447537e-01 -9.42610800e-02
-1.43750250e-01 -1.03402603e+00 -7.60759652e-01 -5.25163770e-01
-1.58043042e-01 1.45329997e-01 4.13060308e-01 1.22436643e+00
-8.86614323e-02 1.08700418e+00 -1.49587378e-01 -1.65274829e-01
1.19979751e+00 2.68890113e-01 5.68193734e-01 -2.09894609e+00
-6.27782568e-02 -5.82497716e-01 -2.07998350e-01 -1.06525517e+00
1.26096618e+00 -1.57131648e+00 -2.71427304e-01 -1.99675107e+00
5.94902873e-01 -8.65115106e-01 -8.16605806e-01 6.32653534e-01
-7.97434270e-01 2.80542046e-01 -2.18765616e-01 3.52686882e-01
-9.52119827e-01 7.21350968e-01 1.39374435e+00 -6.52822018e-01
1.51326954e-01 -6.32017478e-02 -8.80379021e-01 3.35508823e-01
6.97208345e-01 -4.59604979e-01 -6.18743300e-01 -3.84714216e-01
3.00072998e-01 7.33323917e-02 2.82432199e-01 -4.84366506e-01
2.90235609e-01 1.85005188e-01 3.30765724e-01 -7.26073027e-01
1.05212748e-01 -5.84930658e-01 -1.21993296e-01 5.68256617e-01
-5.51253915e-01 -5.80376014e-02 -1.20653830e-01 1.08927739e+00
1.31277591e-01 5.00577167e-02 3.42923701e-01 8.05140361e-02
-6.04166389e-01 9.22432184e-01 2.64675587e-01 4.03679274e-02
1.03273928e+00 3.95793885e-01 -8.97636533e-01 -2.69129425e-01
-6.39920592e-01 8.58655453e-01 2.43769772e-02 5.63387811e-01
5.51587105e-01 -1.54985833e+00 -9.97514009e-01 -1.26539290e-01
3.81390572e-01 9.95638687e-03 -1.35028258e-01 9.32946384e-01
-2.94351310e-01 3.47647071e-01 -5.07383188e-03 -3.34493250e-01
-1.14775610e+00 7.97247350e-01 8.16329643e-02 -8.78975570e-01
-5.18158019e-01 8.76146197e-01 -7.81639516e-02 -7.26726770e-01
5.64089715e-01 3.20611984e-01 -6.58850968e-01 -1.27061740e-01
-1.27140507e-01 2.92152226e-01 -1.08349331e-01 -5.13727963e-01
-4.45415974e-01 3.28557730e-01 -2.24179685e-01 6.67975724e-01
1.53566825e+00 3.27831022e-02 -7.33463407e-01 2.55507678e-01
1.51429021e+00 -3.87098819e-01 -8.71478200e-01 -5.74665248e-01
1.76434487e-01 1.33602634e-01 3.76032233e-01 -5.88235736e-01
-1.55025649e+00 7.27049053e-01 1.46050766e-01 2.57726103e-01
8.92392993e-01 3.59707832e-01 7.58979797e-01 4.19414103e-01
1.41830698e-01 -7.07598209e-01 2.56461769e-01 4.08283710e-01
3.63581598e-01 -1.34898007e+00 3.32279652e-01 -1.05026317e+00
-7.60778934e-02 1.25274909e+00 1.03187621e+00 -4.88713682e-02
1.00497973e+00 1.86444268e-01 -2.88519472e-01 -6.17805362e-01
-9.59446371e-01 -4.32641119e-01 6.18076026e-01 5.81307054e-01
7.52634704e-01 4.51707607e-03 -5.12936652e-01 4.99956757e-01
5.23287535e-01 -2.75742143e-01 1.26838326e-01 6.91944301e-01
-1.79905906e-01 -1.41848814e+00 3.98310810e-01 8.38968396e-01
-3.32019776e-01 -3.77069682e-01 -6.48528457e-01 7.82673240e-01
-2.73426533e-01 8.16405654e-01 1.19766127e-02 -6.87088251e-01
-1.88490704e-01 -2.67260466e-02 3.11519951e-01 -7.13469744e-01
-5.55447578e-01 -3.99475157e-01 4.79637325e-01 -2.54486859e-01
-5.99894404e-01 4.81320731e-02 -1.20358706e+00 -3.33124250e-01
-5.78264892e-01 5.33402681e-01 3.69944006e-01 1.05766737e+00
4.94319052e-01 1.17176008e+00 7.49079883e-01 -7.98353612e-01
-1.73283085e-01 -1.28577471e+00 -3.28558087e-01 4.40622270e-01
2.96038151e-01 -9.44289744e-01 -3.49774629e-01 -4.40603673e-01] | [7.298776626586914, 6.318594455718994] |
881ae911-f2e4-41ba-abe1-09eee44bb496 | a-study-on-the-impact-of-face-image-quality | 2307.02679 | null | https://arxiv.org/abs/2307.02679v1 | https://arxiv.org/pdf/2307.02679v1.pdf | A Study on the Impact of Face Image Quality on Face Recognition in the Wild | Deep learning has received increasing interests in face recognition recently. Large quantities of deep learning methods have been proposed to handle various problems appeared in face recognition. Quite a lot deep methods claimed that they have gained or even surpassed human-level face verification performance in certain databases. As we know, face image quality poses a great challenge to traditional face recognition methods, e.g. model-driven methods with hand-crafted features. However, a little research focus on the impact of face image quality on deep learning methods, and even human performance. Therefore, we raise a question: Is face image quality still one of the challenges for deep learning based face recognition, especially in unconstrained condition. Based on this, we further investigate this problem on human level. In this paper, we partition face images into three different quality sets to evaluate the performance of deep learning methods on cross-quality face images in the wild, and then design a human face verification experiment on these cross-quality data. The result indicates that quality issue still needs to be studied thoroughly in deep learning, human own better capability in building the relations between different face images with large quality gaps, and saying deep learning method surpasses human-level is too optimistic. | ['Na Zhang'] | 2023-07-05 | null | null | null | null | ['face-image-quality', 'face-recognition', 'face-verification'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-3.11635107e-01 -4.54541355e-01 1.38986841e-01 -8.66302431e-01
-4.38444167e-01 -7.97280520e-02 4.25904304e-01 -7.45520830e-01
-2.60545045e-01 4.19848889e-01 -3.41541618e-02 -1.23122133e-01
-3.43236923e-01 -8.25784385e-01 -5.57509363e-01 -8.16209376e-01
1.23417951e-01 2.06220806e-01 -5.39174378e-01 -4.48223025e-01
1.94952264e-01 7.03734457e-01 -1.63593018e+00 3.62384200e-01
5.09670079e-01 1.25814044e+00 -3.06191027e-01 2.53885239e-01
-1.77907780e-01 6.07306540e-01 -7.82430768e-01 -9.92058992e-01
3.85898918e-01 -4.46106583e-01 -6.97305560e-01 2.06926361e-01
7.10135996e-01 -6.59964502e-01 -3.83186549e-01 1.23267043e+00
9.28105950e-01 -4.61731255e-01 4.43851739e-01 -1.68600798e+00
-1.15640736e+00 2.52584875e-01 -5.82429409e-01 3.85576338e-01
2.21501246e-01 4.37715292e-01 5.32684863e-01 -1.19688392e+00
1.44242302e-01 1.65124166e+00 7.70345986e-01 8.40229988e-01
-8.22848439e-01 -1.11580336e+00 -2.88028598e-01 4.95430321e-01
-1.33081961e+00 -8.40714216e-01 7.57100224e-01 -5.31279802e-01
5.37972808e-01 -4.56795171e-02 2.90941894e-01 1.21211028e+00
2.87566572e-01 4.20704395e-01 1.39751792e+00 -2.29791149e-01
-2.14869946e-01 2.27122992e-01 -1.93595365e-01 7.57332981e-01
2.19123498e-01 6.07401013e-01 -4.99157310e-01 3.32035333e-01
6.49383545e-01 5.58357313e-02 -3.79544169e-01 2.63123035e-01
-7.73227155e-01 8.57876301e-01 4.07813996e-01 5.48263073e-01
-9.99797434e-02 -2.44710043e-01 3.39939535e-01 7.79396415e-01
2.15707839e-01 5.18813990e-02 -5.33523619e-01 5.11431172e-02
-1.00783718e+00 6.49594292e-02 5.64799666e-01 6.11716151e-01
7.26861596e-01 3.47853988e-01 -1.07962050e-01 8.40446055e-01
5.06347358e-01 6.57682598e-01 4.42032069e-01 -6.34976625e-01
2.53400296e-01 6.36059642e-01 -1.37939945e-01 -1.44768131e+00
-2.13822395e-01 -3.89058411e-01 -1.30848491e+00 6.61733270e-01
6.29756689e-01 -3.57377343e-02 -7.70309150e-01 1.60733914e+00
1.90971181e-01 1.00870654e-01 -1.85804233e-01 9.68408227e-01
1.15581930e+00 4.69594121e-01 -2.37755682e-02 -2.64178365e-01
1.50054491e+00 -6.10417306e-01 -9.73940492e-01 2.21447349e-01
1.04880780e-01 -1.06170416e+00 1.05090535e+00 7.10200787e-01
-8.34187090e-01 -1.24257886e+00 -1.05180204e+00 1.51765347e-01
-4.10555869e-01 3.13492388e-01 5.32161534e-01 1.35620737e+00
-1.30346191e+00 7.66224682e-01 -2.35924706e-01 -9.89302695e-02
9.15719450e-01 5.54215550e-01 -8.45858216e-01 -3.71637702e-01
-1.23511398e+00 9.53553259e-01 5.19422591e-02 6.00809693e-01
-1.28402400e+00 -3.39093268e-01 -5.95780730e-01 -5.05403569e-03
2.58451372e-01 -2.26649538e-01 9.76197600e-01 -1.29404604e+00
-1.34838343e+00 1.15382028e+00 8.49407688e-02 -8.00211951e-02
4.63063359e-01 -1.08518355e-01 -1.12831128e+00 -2.62378782e-01
-2.49632671e-01 4.63758439e-01 1.07828236e+00 -1.34636688e+00
-2.63281345e-01 -8.92881513e-01 3.80063020e-02 -3.65763187e-01
-4.57865417e-01 4.27537709e-01 -1.80491269e-01 -3.30314815e-01
-2.27491036e-01 -5.72466314e-01 4.37170297e-01 3.56796324e-01
7.53568709e-02 -5.30552804e-01 8.83272767e-01 -7.39525199e-01
9.99846935e-01 -2.07945108e+00 -3.47201347e-01 -8.16296712e-02
3.95580083e-02 8.07084441e-01 -4.98708725e-01 5.83804063e-02
-1.27267927e-01 3.89732450e-01 1.35031492e-02 -1.14451513e-01
-7.42670032e-04 1.13235183e-01 -1.26406597e-02 5.80388069e-01
6.28494918e-01 9.50228572e-01 -3.95973921e-01 -5.66239774e-01
9.67057645e-02 8.88521552e-01 -4.27562088e-01 4.08956438e-01
2.21842036e-01 4.37282056e-01 -1.58467576e-01 9.35361326e-01
1.29485071e+00 2.86320373e-02 -2.20621228e-01 -5.16851187e-01
2.35198215e-01 -3.55236650e-01 -1.16952276e+00 1.35911524e+00
-2.11260095e-01 6.28886878e-01 1.78659633e-01 -1.04686010e+00
1.11995482e+00 5.25626838e-01 2.45185658e-01 -1.15137851e+00
4.76581007e-01 1.50809780e-01 4.37178135e-01 -7.51745343e-01
1.44120902e-02 -4.60693985e-01 5.46708465e-01 1.76284686e-01
3.45262349e-01 2.16821730e-01 -1.32876426e-01 -3.15411896e-01
5.35259187e-01 -1.09479509e-01 -1.27445370e-01 -2.48707190e-01
8.89708042e-01 -5.84839344e-01 5.00399709e-01 2.14438871e-01
-6.82222605e-01 5.42178571e-01 4.77914095e-01 -6.93052948e-01
-9.04305398e-01 -5.66633701e-01 -3.48127276e-01 8.93438399e-01
-1.10940151e-01 -3.36467661e-02 -1.10350764e+00 -6.95233166e-01
6.46241978e-02 -8.97842571e-02 -9.85070467e-01 -3.30665052e-01
-4.02834952e-01 -7.91017473e-01 6.87756658e-01 2.94068724e-01
1.04503560e+00 -1.39038014e+00 -8.28596279e-02 -9.05105472e-02
2.14010954e-01 -1.00137126e+00 -2.35786632e-01 -4.80399787e-01
-6.63811743e-01 -1.36340404e+00 -7.14870334e-01 -8.85258853e-01
5.27429879e-01 1.29542813e-01 1.26891017e+00 6.12816870e-01
-3.87319326e-01 -1.16591573e-01 -2.76760697e-01 -3.67177099e-01
-4.76561069e-01 -3.56555283e-01 2.67467022e-01 2.26334512e-01
7.10906804e-01 -1.85784906e-01 -6.64756835e-01 7.36599922e-01
-8.37498486e-01 -4.79738146e-01 7.70213783e-01 1.00567031e+00
-5.74880652e-02 6.07263982e-01 8.00946295e-01 -4.38897252e-01
7.41079628e-01 -2.13739410e-01 -5.87427318e-01 3.79300207e-01
-6.89350009e-01 2.80042067e-02 4.12614346e-01 -1.47758707e-01
-1.08555877e+00 -4.24098283e-01 -6.68406963e-01 -3.38791937e-01
-3.13969851e-01 2.88147211e-01 -8.68953168e-01 -3.00248444e-01
6.04771197e-01 1.77785218e-01 3.75139147e-01 -3.76260608e-01
4.98927943e-02 8.84723186e-01 3.25646043e-01 -5.32395363e-01
8.19327652e-01 2.90489018e-01 -1.54122442e-01 -4.98755872e-01
-6.68198109e-01 9.81572270e-02 -5.16255021e-01 -4.81744558e-01
7.23140836e-01 -7.52727568e-01 -1.28116846e+00 9.00970459e-01
-1.25076652e+00 -2.06061434e-02 2.08869100e-01 2.74966091e-01
-5.73757803e-03 2.34440714e-01 -6.12799168e-01 -8.78793776e-01
-3.91382188e-01 -1.45163214e+00 9.82987881e-01 3.20157588e-01
4.89997774e-01 -5.98788500e-01 -1.81955189e-01 6.42058253e-01
6.79440022e-01 1.59626231e-01 5.91207504e-01 -3.23583305e-01
-3.92819852e-01 -1.11045137e-01 -6.38032854e-01 7.86383450e-01
4.21528131e-01 1.40915379e-01 -1.33754861e+00 -5.49962461e-01
3.18460464e-01 -5.08618116e-01 5.70788562e-01 1.58964187e-01
1.44753468e+00 -2.83623695e-01 2.20887005e-01 5.74390590e-01
1.49113977e+00 3.10322046e-01 1.02143502e+00 5.14763594e-02
4.66312736e-01 7.64707863e-01 3.62921953e-01 3.45434576e-01
3.48615134e-03 6.16594672e-01 5.25879920e-01 -1.63469285e-01
-2.33637974e-01 -1.43250495e-01 2.77841955e-01 5.90973377e-01
-1.80476263e-01 -2.06946373e-01 -8.06894004e-01 1.10257179e-01
-1.12671876e+00 -1.20676410e+00 7.05669820e-02 1.84396839e+00
7.71523297e-01 1.57860331e-02 6.93219528e-02 6.46042824e-01
7.63336301e-01 -4.59148958e-02 -4.49807614e-01 -2.38911003e-01
-1.89381406e-01 3.20976198e-01 -7.58789107e-02 7.11558089e-02
-7.68927157e-01 7.58379579e-01 5.71027517e+00 7.98583090e-01
-1.34879005e+00 2.81961530e-01 1.12966228e+00 3.82039011e-01
-9.88426711e-03 -3.93429339e-01 -9.43437099e-01 6.10432982e-01
8.29472125e-01 1.88235715e-01 4.55105454e-01 8.48228395e-01
2.13551432e-01 2.96746731e-01 -1.26795304e+00 1.51318967e+00
3.90290409e-01 -9.37076449e-01 2.54833609e-01 4.27370727e-01
6.07963145e-01 -4.90800381e-01 4.30281222e-01 6.39138341e-01
-8.24532583e-02 -1.88520920e+00 3.27758610e-01 5.60859978e-01
9.10097301e-01 -9.21970308e-01 1.13122463e+00 1.17050111e-01
-9.44283664e-01 -1.50828049e-01 -6.56068861e-01 -4.67960276e-02
-3.08284819e-01 6.20313466e-01 -4.74394768e-01 3.99046183e-01
8.43581736e-01 3.63447011e-01 -7.88410366e-01 8.12371731e-01
1.25554264e-01 4.12228495e-01 2.23990306e-01 2.69709617e-01
1.49367331e-02 1.17775835e-01 -2.32130155e-01 8.15545917e-01
3.95450652e-01 1.62078142e-01 -3.23233962e-01 9.40180838e-01
-5.08078337e-01 1.49741903e-01 -6.44307554e-01 -1.26794994e-01
2.05125108e-01 1.26587415e+00 -3.15758139e-01 -1.29328087e-01
-5.25866449e-01 7.55633831e-01 4.93630730e-02 7.62779787e-02
-9.03250515e-01 -6.05614595e-02 6.44100785e-01 1.73179686e-01
-9.27421823e-02 2.41081454e-02 -2.28898227e-02 -9.83276129e-01
1.00318402e-01 -1.46873021e+00 2.30667457e-01 -6.03986621e-01
-1.39897501e+00 1.00496495e+00 -4.63635653e-01 -1.03928852e+00
1.74436554e-01 -1.17087972e+00 -5.28791070e-01 9.12724912e-01
-1.68022442e+00 -9.46859837e-01 -6.94009364e-01 9.80595767e-01
5.43622613e-01 -7.04963207e-01 5.60586214e-01 9.19809461e-01
-6.04451239e-01 1.04159892e+00 -2.71736920e-01 6.91284657e-01
6.33405805e-01 -6.38461471e-01 2.34590575e-01 5.99896014e-01
2.69833326e-01 4.94491458e-01 3.69848371e-01 -2.58255929e-01
-1.65470529e+00 -7.96420813e-01 5.91533065e-01 -5.82637846e-01
-9.33745354e-02 -1.11644408e-02 -1.06385958e+00 1.16932474e-01
4.49715436e-01 3.09782386e-01 6.24608636e-01 1.89314038e-02
-5.17974973e-01 -6.68321848e-01 -1.57263386e+00 5.36916181e-02
9.45645154e-01 -6.53349936e-01 -3.79358053e-01 1.27568200e-01
2.57919997e-01 1.68895811e-01 -8.66000056e-01 8.23997438e-01
7.27225661e-01 -1.36015809e+00 8.50103736e-01 -6.43563926e-01
5.01949072e-01 -2.01440379e-01 -2.26103932e-01 -1.12229514e+00
-2.36687914e-01 -1.94527507e-02 2.38922492e-01 1.51480782e+00
8.98028091e-02 -3.70784521e-01 8.73257160e-01 5.63422740e-01
2.25809038e-01 -6.87261164e-01 -8.43157291e-01 -7.33773768e-01
3.22085410e-01 -2.33166531e-01 1.12013626e+00 1.08485544e+00
-6.23552620e-01 1.27686933e-01 -5.23962080e-01 2.16761917e-01
6.06877148e-01 -1.73431382e-01 9.01189744e-01 -1.34243000e+00
7.57349283e-02 -7.37440407e-01 -6.32802725e-01 -3.86069298e-01
2.85622954e-01 -5.57708383e-01 -1.08290777e-01 -1.24192238e+00
3.89679670e-01 -2.97018409e-01 -4.49996889e-01 4.51994479e-01
-1.19658329e-01 4.38893884e-01 -2.36348510e-02 -2.19722483e-02
-2.79548496e-01 7.46153533e-01 1.64830995e+00 -5.47337353e-01
4.20525074e-01 -2.29470596e-01 -8.21104527e-01 4.61961806e-01
8.64882052e-01 -2.00418204e-01 -2.82077581e-01 -5.43770015e-01
-3.40991281e-03 6.75669089e-02 3.78787011e-01 -1.31293058e+00
1.05386049e-01 -1.92235515e-01 8.83616924e-01 -2.21175894e-01
2.09448591e-01 -1.00028670e+00 1.58271685e-01 3.85008991e-01
-8.66522789e-02 1.70651674e-01 2.39995614e-01 7.11078122e-02
-5.82349658e-01 -2.55672187e-02 1.12881446e+00 -2.72024184e-01
-8.08612466e-01 8.54031980e-01 3.86736780e-01 1.80515679e-04
7.29597390e-01 -3.66471946e-01 -7.52792656e-02 -2.64443636e-01
-4.91193950e-01 -1.27545267e-01 6.96110278e-02 7.43519902e-01
9.34861779e-01 -1.65027738e+00 -1.14911568e+00 5.43174922e-01
3.96858156e-02 -4.43824440e-01 4.33755547e-01 4.56053853e-01
-3.68920565e-01 2.82919049e-01 -7.32394099e-01 -5.94592750e-01
-1.39552760e+00 6.21719360e-01 6.77659035e-01 2.38585219e-01
3.27645987e-02 1.20746291e+00 4.14456017e-02 -5.07168412e-01
4.15961266e-01 2.43413472e-03 -2.78261691e-01 1.04020923e-01
9.22883809e-01 1.36571810e-01 3.58085752e-01 -1.06666458e+00
-4.20414299e-01 7.13474095e-01 3.65378684e-03 2.36537576e-01
1.13931978e+00 2.32627347e-01 -1.96585193e-01 -1.16273500e-01
1.53299785e+00 -2.57404149e-01 -1.10470045e+00 -9.71779302e-02
-1.47851050e-01 -9.28019464e-01 7.49297142e-02 -9.74664569e-01
-1.85659635e+00 1.29560971e+00 1.40023541e+00 7.76721956e-03
1.14238620e+00 -3.42534870e-01 7.06448972e-01 2.91578352e-01
5.11343658e-01 -9.73933935e-01 5.21647930e-01 1.23440988e-01
1.27551496e+00 -2.01086116e+00 -2.16886759e-01 -1.34134907e-02
-4.22964543e-01 1.17419839e+00 1.00582957e+00 1.88106269e-01
1.03367925e+00 1.75005063e-01 2.49097005e-01 -2.77909189e-01
-4.38194096e-01 -2.63786595e-02 3.00413251e-01 5.73164165e-01
7.45067418e-01 -8.94457027e-02 -2.76027210e-02 5.48725724e-01
-3.24096173e-01 3.93684059e-01 -5.53160310e-02 3.62354904e-01
-4.94154841e-01 -1.25841022e+00 -6.69229507e-01 3.62697721e-01
-5.87914765e-01 8.97240639e-02 -1.10221751e-01 7.86991477e-01
8.16806018e-01 1.24473882e+00 -1.06633224e-01 -7.07569182e-01
4.64716256e-01 -7.24795535e-02 7.45019019e-01 -2.16411635e-01
-3.75977695e-01 -4.99775440e-01 -3.32604438e-01 -4.10614878e-01
-4.04537618e-01 -2.35832497e-01 -6.47867799e-01 -8.32219064e-01
-5.01121283e-01 1.54382437e-01 8.26767743e-01 8.94631207e-01
8.88239443e-02 1.99857667e-01 9.31676269e-01 -5.69076598e-01
-7.10192859e-01 -1.10818219e+00 -5.08352458e-01 6.40241146e-01
3.57626706e-01 -8.00046742e-01 -4.01269048e-01 -9.56623331e-02] | [13.128747940063477, 0.7689566612243652] |
affc66e9-3f95-4531-99b3-1e4fac7c690f | fast-capsnet-for-lung-cancer-screening | 1806.07416 | null | http://arxiv.org/abs/1806.07416v1 | http://arxiv.org/pdf/1806.07416v1.pdf | Fast CapsNet for Lung Cancer Screening | Lung cancer is the leading cause of cancer-related deaths in the past several
years. A major challenge in lung cancer screening is the detection of lung
nodules from computed tomography (CT) scans. State-of-the-art approaches in
automated lung nodule classification use deep convolutional neural networks
(CNNs). However, these networks require a large number of training samples to
generalize well. This paper investigates the use of capsule networks (CapsNets)
as an alternative to CNNs. We show that CapsNets significantly outperforms CNNs
when the number of training samples is small. To increase the computational
efficiency, our paper proposes a consistent dynamic routing mechanism that
results in $3\times$ speedup of CapsNet. Finally, we show that the original
image reconstruction method of CapNets performs poorly on lung nodule data. We
propose an efficient alternative, called convolutional decoder, that yields
lower reconstruction error and higher classification accuracy. | ['Aryan Mobiny', 'Hien Van Nguyen'] | 2018-06-19 | null | null | null | null | ['lung-nodule-classification'] | ['medical'] | [ 7.18269050e-02 2.27420971e-01 -6.46155894e-01 -1.09663531e-01
-7.86718011e-01 -3.48432571e-01 7.65470788e-02 -3.25813711e-01
-3.40935647e-01 6.13238573e-01 -1.72409844e-02 -8.21023226e-01
1.72603354e-01 -1.08860624e+00 -6.99493647e-01 -3.19836348e-01
-3.92710231e-02 3.76266152e-01 7.38449574e-01 3.16301197e-01
-3.78341764e-01 9.11005616e-01 -6.29916608e-01 4.93849963e-01
4.13878292e-01 1.00650799e+00 2.35713691e-01 9.73472416e-01
-2.92661905e-01 1.02895010e+00 -1.02673344e-01 -2.42882818e-01
5.00336945e-01 -6.47781610e-01 -1.05402935e+00 1.05499327e-01
3.35083008e-01 -6.88871861e-01 -7.57069349e-01 8.56556475e-01
3.89747798e-01 -6.50384486e-01 4.51947719e-01 -7.58625507e-01
-1.35426745e-01 5.46358049e-01 -1.86822996e-01 3.03719521e-01
-6.20967627e-01 1.55019248e-02 4.84832346e-01 -8.09293151e-01
4.32496130e-01 7.09127247e-01 1.12548852e+00 9.85597312e-01
-5.65702438e-01 -6.99729443e-01 -3.85910660e-01 -1.78175092e-01
-1.17147446e+00 2.07272526e-02 1.64731219e-01 -9.48642045e-02
6.15346074e-01 4.01147038e-01 9.79032397e-01 7.43832529e-01
4.67559695e-01 7.19167709e-01 5.65431118e-01 -1.43124565e-01
1.75680593e-02 6.13071620e-02 -1.77862853e-01 1.26786029e+00
9.62086499e-01 1.01316482e-01 1.63557366e-01 -2.05958426e-01
1.64515066e+00 5.23194909e-01 -2.44179651e-01 -1.79358721e-01
-1.28588831e+00 8.79217148e-01 1.02906585e+00 6.27220333e-01
-2.79638648e-01 9.34143901e-01 3.53134990e-01 -1.18398249e-01
3.35388742e-02 3.45534950e-01 -4.21304464e-01 3.71107846e-01
-9.70413446e-01 -1.27501085e-01 9.45496619e-01 1.02881491e+00
2.76539654e-01 3.58293541e-02 -5.78837991e-01 4.28774893e-01
1.54293433e-01 3.08129728e-01 7.36115098e-01 -7.35104561e-01
2.57794231e-01 6.65637851e-01 -3.05657655e-01 -3.58966798e-01
-4.79506850e-01 -9.46013033e-01 -1.32801974e+00 -9.82395634e-02
3.29774529e-01 -1.89211205e-01 -1.24521029e+00 1.16590607e+00
-6.97265491e-02 2.63497323e-01 -4.15338092e-02 6.00421548e-01
9.71851587e-01 3.02148402e-01 -2.99643613e-02 1.35321811e-01
1.49468458e+00 -1.37952971e+00 -3.28159392e-01 -2.34687462e-01
8.35974038e-01 -4.73851770e-01 5.24086535e-01 -1.59178957e-01
-1.27111399e+00 -4.36958998e-01 -8.60778570e-01 7.50209615e-02
-5.90374544e-02 7.00848877e-01 7.93998480e-01 9.54133868e-01
-1.09808517e+00 5.23759961e-01 -1.32930803e+00 -6.17077231e-01
7.56545544e-01 9.12234783e-01 -1.59959886e-02 -1.38535887e-01
-6.01162612e-01 5.59221208e-01 3.55049431e-01 -1.10145368e-01
-1.18206167e+00 -8.47156048e-01 -3.84778738e-01 3.33527267e-01
3.95007461e-01 -8.78379881e-01 1.87379611e+00 -8.14146459e-01
-1.21937037e+00 4.86448109e-01 -1.26648679e-01 -6.57848060e-01
5.74326992e-01 8.49131495e-02 6.24103323e-02 4.06734079e-01
-1.37151927e-01 9.51713324e-01 2.99870223e-01 -7.64522791e-01
-7.22247958e-01 2.05376133e-01 -3.83878887e-01 -1.90317377e-01
-5.06379247e-01 -2.59918630e-01 -1.00333810e+00 -4.92915720e-01
2.58906752e-01 -1.15977061e+00 -9.62480068e-01 7.15963602e-01
-5.60514390e-01 2.12220997e-02 8.29690218e-01 -3.25422376e-01
1.12351775e+00 -1.78337717e+00 -5.83407938e-01 4.48066503e-01
5.97331524e-01 4.14329618e-01 1.04115024e-01 -2.14333236e-01
6.51190802e-02 6.02440059e-01 -1.58278391e-01 1.66740760e-01
-6.47956967e-01 4.55490291e-01 3.66838485e-01 3.34250718e-01
3.13468695e-01 1.32393801e+00 -6.54984951e-01 -9.08763230e-01
4.36518714e-02 3.00015777e-01 -5.65710723e-01 6.90360889e-02
-8.23398009e-02 3.49529654e-01 -8.96546841e-01 8.23290408e-01
5.73198020e-01 -1.08798325e+00 3.07332635e-01 -2.80940264e-01
4.72835042e-02 2.57188618e-01 -5.53207219e-01 1.32875133e+00
-2.30053142e-01 4.97287303e-01 -5.04423119e-02 -5.93330979e-01
5.25758445e-01 7.40384459e-01 8.14152896e-01 -1.87645838e-01
3.14258218e-01 6.23625219e-01 3.59836131e-01 -3.42230082e-01
-4.57600597e-03 -1.08378775e-01 3.60825717e-01 2.02033073e-01
-3.15136313e-01 -1.04678154e-01 1.39587790e-01 3.76710147e-02
1.78895557e+00 -6.82631612e-01 5.65467775e-01 -2.68858731e-01
4.23792332e-01 5.18397629e-01 3.25346291e-01 8.15692008e-01
-1.67259529e-01 6.75577462e-01 5.59664071e-01 -6.12048209e-01
-1.25525618e+00 -8.94114316e-01 -1.66558638e-01 1.64750054e-01
-3.02176058e-01 5.98728471e-02 -7.23841131e-01 -9.79502261e-01
-1.11051373e-01 -7.46931881e-02 -5.95139563e-01 2.07785547e-01
-1.05720198e+00 -8.59873176e-01 9.89573896e-01 1.09198487e+00
1.05385804e+00 -7.75059104e-01 -6.89599931e-01 2.88782507e-01
2.40334473e-03 -1.07200158e+00 -3.75243396e-01 3.53879273e-01
-1.57277238e+00 -1.30645251e+00 -1.01188385e+00 -1.22168744e+00
1.15629804e+00 4.79332507e-01 1.11461687e+00 7.26054907e-01
-6.48382187e-01 2.10225031e-01 -3.59922908e-02 -1.74922720e-01
-6.20621622e-01 5.90618849e-01 -5.62660873e-01 -6.04351580e-01
9.76080522e-02 -4.97710370e-02 -7.89513886e-01 1.61407068e-01
-9.11886752e-01 8.04065540e-02 1.39471304e+00 8.25351417e-01
8.76848996e-01 2.42587943e-02 1.45091861e-01 -1.20016801e+00
2.35720471e-01 -5.74469268e-01 -4.40034986e-01 1.33828774e-01
-3.26905191e-01 3.90820950e-02 7.19767988e-01 -3.31226468e-01
-7.03092873e-01 5.89556873e-01 -1.79298818e-01 -4.40844238e-01
1.34392723e-01 3.57872844e-01 5.27943969e-01 -5.56949258e-01
7.41035402e-01 1.82043929e-02 6.41102195e-02 -3.03469896e-01
-1.59328416e-01 3.61554235e-01 3.98882270e-01 5.07556163e-02
9.02778566e-01 7.18228698e-01 4.17866141e-01 -5.73084772e-01
-8.39668334e-01 -5.11811912e-01 -4.50787485e-01 -1.10373735e-01
1.04482567e+00 -1.10521007e+00 -4.25018072e-01 -1.02147616e-01
-1.11600888e+00 -3.17239285e-01 -2.33185381e-01 9.30152833e-01
-1.72718003e-01 1.90565884e-02 -1.05791283e+00 -9.60280448e-02
-5.18402994e-01 -1.13228750e+00 7.21433282e-01 1.73759013e-01
-1.22497473e-02 -8.45193207e-01 -1.61847040e-01 -1.64625019e-01
7.12950289e-01 1.61565199e-01 9.87750351e-01 -6.85161054e-01
-1.35349452e+00 -4.79539007e-01 -7.30474949e-01 2.29515448e-01
2.90032953e-01 -1.03437088e-01 -6.11942649e-01 -4.06082749e-01
2.34907325e-02 -8.85123610e-02 1.04145157e+00 7.59765446e-01
1.87197006e+00 -2.39981771e-01 -9.34727907e-01 7.62789488e-01
1.75127506e+00 2.95354515e-01 8.32433701e-01 8.41419250e-02
6.60640061e-01 -1.64288864e-01 -9.54194218e-02 3.31727229e-03
-2.34641269e-01 -1.78989530e-01 7.36030221e-01 -6.54076934e-01
-6.12085521e-01 -2.31675636e-02 -1.40067220e-01 8.96235883e-01
-5.18504642e-02 -4.03983563e-01 -1.01932013e+00 6.20773435e-01
-1.39134967e+00 -6.07464612e-01 -4.21686351e-01 1.77674043e+00
6.00730479e-01 1.03780881e-01 -3.01882952e-01 -1.26362696e-01
6.30066931e-01 -3.54205191e-01 -5.16394317e-01 1.55709878e-01
4.70875323e-01 6.43837035e-01 1.30221915e+00 1.76478475e-02
-1.12021649e+00 4.35653687e-01 7.12053299e+00 7.58674204e-01
-1.22915959e+00 2.47252703e-01 9.51197028e-01 9.59212109e-02
-6.85050897e-03 -3.27556610e-01 -7.38202453e-01 -4.09669243e-02
8.30166161e-01 1.66501120e-01 -2.64447302e-01 1.19603908e+00
-1.05737016e-01 4.77470160e-02 -1.15471876e+00 5.43516159e-01
-1.20718569e-01 -1.83803070e+00 1.01881407e-01 2.86191612e-01
8.55953991e-01 4.84100014e-01 1.57684591e-02 4.78816293e-02
3.62093657e-01 -1.25635827e+00 2.31941836e-03 1.77922979e-01
1.08112776e+00 -5.22150159e-01 1.07235312e+00 2.24752367e-01
-1.53873825e+00 -4.55725044e-02 -5.93823552e-01 5.50036669e-01
-2.46286497e-01 5.58163285e-01 -1.44478834e+00 2.60273695e-01
6.64391458e-01 4.18464839e-01 -7.81557024e-01 1.57272375e+00
1.64752588e-01 1.04559708e+00 -4.54395473e-01 -3.81004095e-01
4.92273510e-01 5.07675588e-01 5.99180162e-02 1.24281383e+00
7.99504697e-01 2.55460888e-01 2.75959134e-01 8.80040765e-01
-6.45161450e-01 -1.42059222e-01 -4.31013167e-01 -7.70073384e-02
2.50565648e-01 1.25090647e+00 -1.17236245e+00 -4.46948320e-01
-6.40525043e-01 6.97281241e-01 3.28787079e-04 -1.31428555e-01
-9.24197376e-01 4.38922383e-02 -9.47463885e-02 3.38556051e-01
4.16651458e-01 -2.12977882e-02 -3.30982447e-01 -6.32395387e-01
-1.41974539e-01 -4.72142071e-01 1.71285689e-01 -3.42177838e-01
-1.20424211e+00 7.42166400e-01 -3.83554906e-01 -1.52451468e+00
1.38113331e-02 -9.38877165e-01 -7.25860119e-01 3.82482767e-01
-1.70060372e+00 -9.69608188e-01 -7.81266093e-01 5.61446488e-01
5.61762452e-01 -6.91682920e-02 7.06362605e-01 4.14761901e-01
-3.92351717e-01 3.85695368e-01 5.49830645e-02 7.72156119e-01
3.63478184e-01 -1.06933701e+00 5.05833864e-01 6.40092731e-01
-4.09406483e-01 1.05013609e-01 -6.33520111e-02 -8.19635093e-01
-1.21270370e+00 -1.86685669e+00 8.34785938e-01 -7.64766186e-02
2.93216586e-01 1.88280255e-01 -6.32637024e-01 7.62488723e-01
3.98755036e-02 5.70078313e-01 5.57152510e-01 -7.39275932e-01
1.28755778e-01 2.48028841e-02 -1.13572943e+00 5.83816171e-01
7.51690567e-01 -1.52053639e-01 2.22213373e-01 6.18960619e-01
7.65205145e-01 -5.70457935e-01 -7.12704837e-01 6.64873958e-01
3.58028650e-01 -8.41662467e-01 9.58243906e-01 -3.52713197e-01
5.29534280e-01 -2.35059321e-01 2.27750823e-01 -6.00270391e-01
-5.49888968e-01 8.54929350e-03 2.19888121e-01 2.24591140e-02
7.36407280e-01 -3.77065837e-01 1.63536894e+00 3.27774704e-01
-3.62716764e-01 -1.12975216e+00 -7.34350681e-01 -8.90971243e-01
2.48635620e-01 -1.96282476e-01 2.62943089e-01 5.61674833e-01
-5.33376157e-01 -2.91561157e-01 -1.92323718e-02 9.96264592e-02
4.13756818e-01 -4.04865980e-01 3.44167471e-01 -1.10071552e+00
-2.94762462e-01 -3.95108908e-01 -1.83339924e-01 -1.18543768e+00
-5.44254422e-01 -1.02189767e+00 9.72050801e-02 -1.84041786e+00
5.67673206e-01 -4.63553578e-01 -1.83679372e-01 5.08657336e-01
1.39283091e-01 4.31556791e-01 1.21654168e-01 3.15218002e-01
-3.59840274e-01 -1.62386954e-01 1.81109869e+00 -1.55180410e-01
3.19297351e-02 4.12984163e-01 -3.96257430e-01 7.63112664e-01
1.19983673e+00 -8.43249977e-01 -2.05318004e-01 -4.61669594e-01
-8.28761682e-02 4.08575863e-01 5.34841537e-01 -1.46372008e+00
5.02955914e-01 -5.35336025e-02 8.99567485e-01 -9.64029491e-01
1.30712062e-01 -1.10033464e+00 2.87216723e-01 1.55577409e+00
-3.57706934e-01 -4.14389558e-02 1.29678130e-01 5.42094886e-01
-1.61019042e-01 -4.54597712e-01 9.34504092e-01 -9.46520030e-01
-5.52034266e-02 7.97202170e-01 -7.93719113e-01 -3.77112657e-01
1.02122438e+00 -2.79754043e-01 -1.44957706e-01 3.75799984e-02
-5.25542796e-01 1.07713178e-01 1.12995263e-02 -2.14491159e-01
6.08329475e-01 -1.42705131e+00 -7.39821017e-01 -6.18247129e-03
-3.05426002e-01 3.47871453e-01 -1.09277241e-01 9.80788291e-01
-1.35564148e+00 9.44653869e-01 1.44773141e-01 -6.37555659e-01
-1.18595421e+00 2.06952482e-01 1.02670002e+00 -9.86024976e-01
-5.73590398e-01 9.92051840e-01 1.99651659e-01 -2.31839314e-01
4.04911786e-01 -9.16547596e-01 2.20599309e-01 -7.66228914e-01
1.75427347e-01 3.69755924e-01 1.62318572e-01 2.98840523e-01
-1.92249045e-01 1.91607356e-01 -2.16702551e-01 2.30243415e-01
9.75005507e-01 4.76222098e-01 -1.92872986e-01 -2.77818441e-01
1.24349189e+00 -2.78763533e-01 -8.53857517e-01 -1.47504285e-01
-1.85062201e-03 -1.86426565e-01 2.12958515e-01 -7.10907817e-01
-1.58665931e+00 7.51970232e-01 7.77968109e-01 4.20204066e-02
1.01498616e+00 1.54165849e-01 1.13304710e+00 9.17619646e-01
-1.30413321e-03 -5.65895557e-01 4.16070700e-01 2.74618179e-01
4.23085898e-01 -1.11504519e+00 6.84175193e-02 -9.21923459e-01
-1.65347368e-01 1.60414004e+00 8.37000430e-01 -3.56888950e-01
9.57626700e-01 6.06320202e-01 -1.04108572e-01 -3.85496110e-01
-7.36949623e-01 -2.80103117e-01 1.02481797e-01 2.66275108e-01
6.87970757e-01 1.60393670e-01 -1.55545488e-01 4.20425028e-01
1.11910038e-01 3.29591513e-01 5.06526053e-01 1.20918572e+00
-8.13608289e-01 -1.14811254e+00 -4.78599787e-01 1.03988552e+00
-8.27125132e-01 -2.58375555e-01 -4.81156141e-01 1.22931886e+00
5.44304103e-02 4.31926757e-01 2.87233204e-01 2.04870570e-02
2.26345751e-02 -2.74285674e-01 3.37188303e-01 -8.08071256e-01
-9.28401053e-01 2.75529325e-01 -2.80796923e-02 -4.47588503e-01
-5.55180550e-01 -1.68166369e-01 -1.51996386e+00 -3.04411441e-01
-4.33892965e-01 -1.08953163e-01 5.33112645e-01 5.86525023e-01
-2.89209709e-02 1.07624722e+00 4.91858840e-01 -2.04237908e-01
-5.07314563e-01 -7.24329352e-01 -3.49110514e-01 -3.34411919e-01
2.84457356e-01 1.38907760e-01 -8.78913403e-02 -1.63987314e-03] | [15.358817100524902, -2.1904876232147217] |
7488f0ab-663a-4c0c-a149-7700050cdd41 | microbert-effective-training-of-low-resource | 2212.12510 | null | https://arxiv.org/abs/2212.12510v2 | https://arxiv.org/pdf/2212.12510v2.pdf | MicroBERT: Effective Training of Low-resource Monolingual BERTs through Parameter Reduction and Multitask Learning | Transformer language models (TLMs) are critical for most NLP tasks, but they are difficult to create for low-resource languages because of how much pretraining data they require. In this work, we investigate two techniques for training monolingual TLMs in a low-resource setting: greatly reducing TLM size, and complementing the masked language modeling objective with two linguistically rich supervised tasks (part-of-speech tagging and dependency parsing). Results from 7 diverse languages indicate that our model, MicroBERT, is able to produce marked improvements in downstream task evaluations relative to a typical monolingual TLM pretraining approach. Specifically, we find that monolingual MicroBERT models achieve gains of up to 18% for parser LAS and 11% for NER F1 compared to a multilingual baseline, mBERT, while having less than 1% of its parameter count. We conclude reducing TLM parameter count and using labeled data for pretraining low-resource TLMs can yield large quality benefits and in some cases produce models that outperform multilingual approaches. | ['Amir Zeldes', 'Luke Gessler'] | 2022-12-23 | null | null | null | null | ['dependency-parsing', 'part-of-speech-tagging'] | ['natural-language-processing', 'natural-language-processing'] | [-1.79176152e-01 2.81921625e-01 -4.06352878e-01 -5.49815834e-01
-1.70615828e+00 -9.56500649e-01 6.05750561e-01 2.81963408e-01
-7.47651041e-01 8.12390029e-01 3.96399885e-01 -9.32013571e-01
6.13689840e-01 -3.46490860e-01 -8.48141491e-01 -2.45751590e-01
2.68647969e-01 6.95890069e-01 9.49601978e-02 -2.05653012e-01
-1.95529774e-01 2.60832638e-01 -7.58749723e-01 4.24953759e-01
1.11330581e+00 1.95472717e-01 6.02284610e-01 3.49983662e-01
-4.67815757e-01 9.56431627e-01 -5.14224827e-01 -5.47828078e-01
1.20294029e-02 -1.70836061e-01 -9.26185489e-01 -3.88019830e-01
5.72146058e-01 7.66137615e-02 2.42537335e-01 7.71973252e-01
4.06242996e-01 -3.27677950e-02 5.00314295e-01 -6.90100193e-01
-5.21623433e-01 1.38943970e+00 -4.46764052e-01 2.04800591e-01
1.41431317e-01 9.58815217e-02 1.18949056e+00 -1.10354793e+00
6.84362054e-01 1.47900772e+00 8.55939627e-01 4.73050296e-01
-1.54540753e+00 -7.15480566e-01 2.65890867e-01 -3.79994243e-01
-1.20638347e+00 -9.37545121e-01 1.75135136e-01 -3.36791277e-01
1.63127303e+00 -8.77797306e-02 -1.83534250e-01 8.81216347e-01
2.58281112e-01 7.53809929e-01 1.36989152e+00 -8.62841249e-01
-2.55756915e-01 4.09077078e-01 1.16197899e-01 7.08759248e-01
-2.57340781e-02 -1.42360732e-01 -4.71583694e-01 1.11761019e-01
3.59892994e-01 -6.50197029e-01 1.19389698e-01 2.33252451e-01
-1.36124539e+00 8.34492683e-01 1.50497958e-01 6.95374310e-01
-7.71494210e-02 1.13000274e-02 4.93911058e-01 5.20839393e-01
7.57720649e-01 7.52752125e-01 -9.08569574e-01 -5.57308681e-02
-9.14317369e-01 -2.00338900e-01 7.70424306e-01 1.10461760e+00
7.81877875e-01 2.60886610e-01 8.23835433e-02 1.18770552e+00
1.11167744e-01 7.62909830e-01 5.15913188e-01 -8.07597876e-01
1.05712020e+00 3.73350829e-01 -1.67762220e-01 -8.40830207e-02
-4.56498712e-01 -2.61250317e-01 -3.31664324e-01 -2.60398865e-01
5.58888614e-01 -4.69383061e-01 -9.78579342e-01 2.06287694e+00
6.71058446e-02 -3.75968546e-01 4.31629479e-01 2.17170775e-01
6.21917307e-01 9.27186668e-01 6.83949888e-01 -6.04406707e-02
1.33815563e+00 -1.01942706e+00 -5.10574639e-01 -8.47098470e-01
1.39777374e+00 -1.29145646e+00 1.51117373e+00 9.23264921e-02
-1.34977508e+00 -5.60671329e-01 -7.56373405e-01 -3.23883682e-01
-2.97132432e-01 3.55833143e-01 6.04406953e-01 7.01613605e-01
-1.23902571e+00 4.18510586e-01 -1.02978575e+00 -4.22373950e-01
-9.82345939e-02 3.36789817e-01 -5.94903767e-01 -4.45547879e-01
-1.04257619e+00 1.34873879e+00 3.27163249e-01 -1.91806197e-01
-9.20257688e-01 -7.84419775e-01 -1.15382409e+00 9.18038283e-03
7.85511360e-02 -1.66083351e-01 1.63620496e+00 -6.10049725e-01
-1.20398211e+00 1.01099634e+00 -3.23641390e-01 -3.46415669e-01
7.11852685e-02 -3.08603048e-01 -4.17855650e-01 -2.80274332e-01
4.46660846e-01 8.83354902e-01 2.29870856e-01 -9.91265774e-01
-8.20968628e-01 1.14183035e-02 -8.21520537e-02 2.63738483e-01
-4.36388850e-01 6.36588275e-01 -2.85834014e-01 -4.82285351e-01
-1.41596049e-01 -9.51766551e-01 -4.75587904e-01 -9.09882486e-01
-3.24393421e-01 -2.37726912e-01 3.95513594e-01 -1.08142710e+00
1.01941812e+00 -1.87015867e+00 -1.48532003e-01 -2.39968032e-01
-2.42291152e-01 4.25507396e-01 -4.77527678e-01 5.25784850e-01
8.66114348e-02 5.19202471e-01 -1.93855718e-01 -6.98276043e-01
-9.19815376e-02 4.07552034e-01 -1.55478865e-02 8.83784890e-02
3.28801155e-01 8.85312200e-01 -8.20827246e-01 -5.55036306e-01
4.09876034e-02 3.98312181e-01 -4.52743471e-01 1.76311478e-01
-1.69926047e-01 7.18240142e-01 4.54585813e-02 5.50375640e-01
3.25826198e-01 6.92989230e-02 7.04705894e-01 1.06888153e-01
-4.39838052e-01 1.15601861e+00 -6.45990014e-01 1.67690384e+00
-1.26467562e+00 5.64870298e-01 3.36404949e-01 -5.53668618e-01
7.01318681e-01 5.92877805e-01 7.16538876e-02 -7.73322821e-01
-9.42419395e-02 6.58358455e-01 3.30020130e-01 -1.03121117e-01
5.39073825e-01 -3.95357758e-01 -4.60468739e-01 5.30251682e-01
4.74841863e-01 -1.29689887e-01 3.84404272e-01 2.30383158e-01
1.08429658e+00 5.28188124e-02 1.35181323e-01 -6.13512635e-01
3.87278557e-01 2.82010257e-01 8.44510138e-01 4.71059054e-01
2.61994861e-02 8.50065351e-02 1.52140275e-01 -4.26719636e-02
-1.20067906e+00 -1.02161098e+00 -1.75268844e-01 1.56209528e+00
-5.71538866e-01 -4.29229438e-01 -8.88971746e-01 -8.58911753e-01
-2.25455537e-01 1.05256522e+00 1.27952859e-01 1.56040788e-01
-1.22066247e+00 -9.22040343e-01 9.25679862e-01 7.20621526e-01
1.73407838e-01 -1.18663311e+00 1.83307230e-01 4.13857818e-01
-4.27602470e-01 -1.59190178e+00 -6.49706483e-01 5.71436226e-01
-9.92509067e-01 -5.99053383e-01 -4.67759550e-01 -1.35903716e+00
6.33780122e-01 -2.59901471e-02 1.57272327e+00 -2.18292668e-01
2.37240657e-01 -1.73065469e-01 -2.64788210e-01 -2.25259855e-01
-1.11419094e+00 5.98424017e-01 1.09859325e-01 -5.66699684e-01
4.89037693e-01 -5.19364536e-01 1.32142771e-02 1.18765503e-01
-4.77953136e-01 7.29088783e-02 8.52497697e-01 6.97663724e-01
5.18297970e-01 -3.70149255e-01 7.34318912e-01 -1.34876180e+00
4.75172639e-01 -3.40544492e-01 -4.71328259e-01 2.89122581e-01
-8.21302354e-01 2.69195020e-01 8.69740427e-01 -4.31277126e-01
-1.39983535e+00 5.03590964e-02 -4.49381024e-01 3.52136910e-01
-1.21573821e-01 6.20602071e-01 -4.34743285e-01 9.60197598e-02
6.61043286e-01 -1.68435082e-01 -3.46726805e-01 -1.03260481e+00
5.99580407e-01 5.94469965e-01 4.42468792e-01 -9.31255162e-01
7.01669037e-01 -1.38718098e-01 -4.94997144e-01 -5.97977936e-01
-8.59759867e-01 -4.10056859e-01 -8.65602553e-01 3.87683958e-01
7.95192599e-01 -1.27807415e+00 -3.51280719e-02 1.55674726e-01
-1.11226273e+00 -8.52226079e-01 -1.14444047e-01 6.38621390e-01
-1.27820343e-01 1.79791316e-01 -1.21994627e+00 -4.59689319e-01
-3.91931355e-01 -1.31717527e+00 8.69372308e-01 -2.29844227e-01
-1.92075402e-01 -1.37062836e+00 1.76067084e-01 4.86804515e-01
4.27886069e-01 -2.95627058e-01 1.36434138e+00 -8.20503950e-01
-2.88680136e-01 1.50373904e-02 -4.88087088e-02 5.82097054e-01
1.05987027e-01 -4.22162592e-01 -8.94046307e-01 -4.85947520e-01
-2.72586614e-01 -5.15499413e-01 6.24868512e-01 3.37266505e-01
2.30869696e-01 -2.13802710e-01 -3.73449832e-01 4.07664657e-01
1.35174859e+00 9.11749378e-02 1.09131575e-01 1.91501021e-01
8.06460798e-01 6.97514057e-01 5.38997591e-01 -3.84437352e-01
6.93056285e-01 4.62955087e-01 -2.14371264e-01 -3.74761283e-01
-4.10762161e-01 -5.11221945e-01 9.28806424e-01 1.50601220e+00
2.91773170e-01 -2.31575891e-01 -1.29816759e+00 8.75936091e-01
-1.48360121e+00 -3.59244674e-01 -1.64762959e-01 2.19567943e+00
1.22760367e+00 1.48704380e-01 -1.21628068e-01 -3.32493186e-01
6.13659441e-01 -2.40961146e-02 -6.23073392e-02 -7.41461217e-01
-2.41797239e-01 4.77503628e-01 7.27703810e-01 9.73022342e-01
-8.84326041e-01 1.79989469e+00 6.72656727e+00 6.68537855e-01
-9.64255512e-01 6.89005136e-01 5.15195787e-01 2.30369493e-01
-3.69246393e-01 3.92981350e-01 -1.36175203e+00 1.14300676e-01
1.66040778e+00 1.00784734e-01 3.34488481e-01 7.12354898e-01
1.85988307e-01 1.74169503e-02 -1.21806157e+00 4.41752642e-01
-2.04448357e-01 -9.42309797e-01 -3.72056924e-02 4.21739593e-02
7.64082074e-01 6.62375569e-01 -2.18406349e-01 7.27208853e-01
1.05525470e+00 -1.14175427e+00 6.48377061e-01 -1.97056383e-01
9.60222721e-01 -7.60460615e-01 7.25368977e-01 6.84133649e-01
-1.35669196e+00 1.87928036e-01 -5.26789606e-01 2.36349311e-02
4.30485725e-01 4.99102980e-01 -1.22251475e+00 3.27646434e-01
4.32280362e-01 3.28371167e-01 -6.94399416e-01 5.48201859e-01
-5.35066962e-01 1.21039844e+00 -3.75116676e-01 2.77334660e-01
3.90783876e-01 2.07742751e-01 1.92166835e-01 1.72967947e+00
2.51104206e-01 -4.66132969e-01 5.15767872e-01 3.90277326e-01
-3.88273686e-01 5.96907437e-01 -5.92268705e-01 -8.92703012e-02
7.65694559e-01 1.35016942e+00 -4.92307901e-01 -3.97837132e-01
-6.53463960e-01 7.76105821e-01 7.69607306e-01 1.24003045e-01
-5.05074322e-01 -2.12822646e-01 4.07340854e-01 9.84103605e-02
-2.21930305e-03 -4.58443373e-01 -1.61858931e-01 -1.33879375e+00
-5.86432964e-02 -1.18655181e+00 3.22238535e-01 -3.96326870e-01
-1.19936633e+00 9.37137127e-01 -1.22079231e-01 -7.18740165e-01
-6.16912842e-01 -8.13354790e-01 -4.21114981e-01 1.27952683e+00
-1.73336208e+00 -1.66365778e+00 4.50434089e-01 3.44842643e-01
6.84041381e-01 -2.01277390e-01 1.11055577e+00 6.28012300e-01
-6.10427797e-01 8.52069855e-01 9.51518416e-02 4.58269358e-01
1.04489863e+00 -1.45126843e+00 8.65284145e-01 1.20832968e+00
5.36031544e-01 8.52034748e-01 3.20492864e-01 -5.60588837e-01
-1.09805644e+00 -1.21472144e+00 1.79296207e+00 -6.70607746e-01
6.76002741e-01 -7.24023104e-01 -8.23794961e-01 1.36408222e+00
4.40868586e-01 -1.93552166e-01 6.48956716e-01 6.94679439e-01
-4.28014100e-01 2.37975419e-02 -1.02919531e+00 5.06891429e-01
8.66712451e-01 -6.71255946e-01 -6.63537979e-01 4.77528363e-01
8.75996649e-01 -1.90757215e-01 -1.22387695e+00 3.56681794e-01
1.73920885e-01 -4.27407503e-01 7.10342407e-01 -6.71240151e-01
1.59034833e-01 9.04044807e-02 -3.05789828e-01 -1.70431328e+00
-3.97932947e-01 -6.35858834e-01 5.32170057e-01 1.83700025e+00
1.17876089e+00 -6.48342490e-01 4.23554987e-01 3.99380833e-01
-4.43084538e-01 -2.89463103e-01 -7.14968979e-01 -9.77475286e-01
9.33507144e-01 -5.26604652e-01 3.40100616e-01 1.02026069e+00
6.93174452e-02 9.78548467e-01 -1.49614483e-01 2.55123496e-01
5.38633466e-01 -1.98039994e-01 5.51666796e-01 -1.05835307e+00
-2.93529481e-01 -2.35181987e-01 2.20653027e-01 -9.94419754e-01
6.70491517e-01 -1.42175841e+00 3.68072033e-01 -1.45835376e+00
1.92055613e-01 -1.02052057e+00 -1.73865438e-01 8.70257914e-01
-1.90516829e-01 2.18232870e-01 2.35012636e-01 2.39390701e-01
-2.22515702e-01 5.72064556e-02 7.57685959e-01 4.29743268e-02
-2.16537699e-01 -3.00449073e-01 -8.39085519e-01 8.77861559e-01
7.13120639e-01 -8.19302976e-01 -2.53236681e-01 -1.14340472e+00
-4.08985503e-02 1.87894434e-01 -2.77754873e-01 -8.11040461e-01
-4.64632213e-02 -1.13916300e-01 1.08431526e-01 -2.05553517e-01
1.90103754e-01 -3.14826608e-01 -2.45289981e-01 2.51345605e-01
-3.22655410e-01 4.97510135e-01 5.73733151e-01 -1.37404874e-01
-2.78168976e-01 -3.68951440e-01 8.57176781e-01 -3.98328632e-01
-7.49759793e-01 5.55019919e-03 -6.17091417e-01 5.03222466e-01
4.01877046e-01 4.64881510e-01 -3.05939585e-01 -6.78183734e-02
-5.55665672e-01 1.69724196e-01 3.62701237e-01 4.10509408e-01
-1.45862982e-01 -1.01716828e+00 -8.87664199e-01 4.69455868e-02
-3.62662077e-02 -2.09495664e-01 -2.02752665e-01 7.08135724e-01
-3.98236752e-01 8.48664463e-01 6.50095269e-02 -3.77061307e-01
-1.14454842e+00 3.75734806e-01 2.03067183e-01 -9.08484161e-01
-5.00535369e-01 9.62246776e-01 2.74119496e-01 -1.24574912e+00
4.35504727e-02 -4.63636816e-01 1.60537679e-02 -1.63652569e-01
2.61256635e-01 -4.88609727e-03 3.60588908e-01 -8.39204907e-01
-5.28899252e-01 2.91483432e-01 -2.89420605e-01 -5.79725385e-01
1.24274063e+00 -6.37657270e-02 -8.91376510e-02 3.46979916e-01
1.05757487e+00 6.83412731e-01 -8.48160923e-01 -4.09365118e-01
5.10711372e-01 1.18948519e-01 1.39601216e-01 -1.04117274e+00
-8.05575132e-01 1.04211271e+00 1.10714100e-01 -3.79928380e-01
8.02459598e-01 1.74788043e-01 1.16871393e+00 5.91627836e-01
5.57525218e-01 -1.02597034e+00 -4.19339806e-01 7.64454424e-01
3.36715996e-01 -1.25036705e+00 -3.66095662e-01 -3.05318534e-01
-6.12103045e-01 5.42047143e-01 5.79944968e-01 7.18198493e-02
5.77260613e-01 9.04734850e-01 5.60425878e-01 1.66249767e-01
-8.27449977e-01 -1.93587735e-01 1.22042075e-01 4.83453691e-01
1.18397224e+00 3.04969341e-01 -3.62220794e-01 5.58883846e-01
-4.54826921e-01 -6.37079597e-01 2.34657809e-01 8.66115093e-01
-4.34540242e-01 -1.77346098e+00 -2.49130249e-01 9.76641774e-02
-9.48710382e-01 -8.24702382e-01 -1.30686685e-01 1.06565988e+00
4.76835631e-02 1.15752721e+00 -2.59346142e-02 -1.76274091e-01
2.31676906e-01 5.33300042e-01 4.39727753e-01 -1.20828509e+00
-8.61985266e-01 2.74405956e-01 6.64213061e-01 -2.72212029e-01
-1.51668221e-01 -7.59499192e-01 -1.31549394e+00 -2.34976441e-01
-4.18519378e-01 4.26870853e-01 7.04019606e-01 1.07533085e+00
1.63911715e-01 2.96150923e-01 3.57187927e-01 -6.20175779e-01
-5.11666894e-01 -1.40298247e+00 -1.80317000e-01 3.58735509e-02
-2.05790624e-02 -2.64004380e-01 -1.93626329e-01 1.96317703e-01] | [10.604690551757812, 9.945411682128906] |
11201c92-6afa-4fe4-ac45-1593397e74a9 | a-multi-task-architecture-on-relevance-based | 1906.06849 | null | https://arxiv.org/abs/1906.06849v1 | https://arxiv.org/pdf/1906.06849v1.pdf | A Multi-Task Architecture on Relevance-based Neural Query Translation | We describe a multi-task learning approach to train a Neural Machine Translation (NMT) model with a Relevance-based Auxiliary Task (RAT) for search query translation. The translation process for Cross-lingual Information Retrieval (CLIR) task is usually treated as a black box and it is performed as an independent step. However, an NMT model trained on sentence-level parallel data is not aware of the vocabulary distribution of the retrieval corpus. We address this problem with our multi-task learning architecture that achieves 16% improvement over a strong NMT baseline on Italian-English query-document dataset. We show using both quantitative and qualitative analysis that our model generates balanced and precise translations with the regularization effect it achieves from multi-task learning paradigm. | ['James Allan', 'Sheikh Muhammad Sarwar', 'Hamed Bonab'] | 2019-06-17 | a-multi-task-architecture-on-relevance-based-1 | https://aclanthology.org/P19-1639 | https://aclanthology.org/P19-1639.pdf | acl-2019-7 | ['cross-lingual-information-retrieval'] | ['natural-language-processing'] | [ 0.34814474 -0.00827679 -0.8019403 -0.217744 -2.1594117 -0.7046905
0.9279424 0.01156545 -0.77692384 0.84404355 0.5480962 -0.7285159
0.23137264 -0.22573134 -1.1188165 -0.39131266 0.70899814 1.4351768
-0.21988598 -0.7036958 0.0273435 -0.1026575 -0.37028828 0.8080456
0.92471164 0.6026642 0.19831519 0.43166932 -0.14658256 0.4264691
-0.52767193 -0.63316095 0.3066969 -0.4138444 -1.3096021 -0.4728792
0.45122087 -0.0502157 0.02966669 0.65643597 0.79409635 -0.19311592
1.0121075 -0.70927256 -1.0914359 0.7047646 -0.54350483 0.3572229
0.07186382 -0.1589127 1.4213934 -1.1998227 0.8059671 1.2637633
0.29872528 0.49002418 -1.5440058 -0.48714426 -0.17280188 -0.02598725
-1.2358254 -0.4137709 0.4683383 -0.07553633 1.3089889 0.1956464
0.040762 1.6632303 0.7345234 1.1743622 0.9890787 -0.8014517
-0.20746286 0.27473712 -0.15095653 -0.07195687 -0.18752901 0.06998465
-0.56380004 -0.10234519 0.5858136 -0.4031268 0.09696364 -0.19170573
-1.4433652 1.0738618 0.29437196 0.53064793 -0.4108963 0.14889333
0.664145 0.9799711 1.0169255 0.55947065 -0.970003 0.1956177
-0.94702005 0.185742 0.69926184 1.2872819 0.69678277 -0.07996588
-0.75445724 1.0933691 0.2603881 1.1673841 0.7354413 -0.3870782
0.7780179 0.22133599 0.09678749 -0.2811706 -0.0149763 -0.6448823
-0.7236886 -0.5336784 0.17211838 -0.149182 -0.8727967 1.6319984
-0.11934859 -0.70245755 0.28577256 0.6897244 0.59232783 0.7403378
0.2178772 -0.28695416 1.3527433 -1.379714 -0.89458126 -0.6259222
1.0183842 -1.3181235 1.343059 -0.09853493 -1.2843771 -0.50566775
-0.641924 -0.47769094 -0.36806807 0.44228247 0.34133142 0.05741859
-1.2770183 0.04417973 -0.538757 -0.36171007 -0.00955762 0.29324436
-0.25457564 -0.0954117 -1.5151582 1.5117167 0.12875007 -0.13314186
-1.0055134 -0.55368364 -0.6749875 -0.11870275 0.1630069 -0.89668185
1.6315187 -1.2328508 -1.3397232 1.0638458 -0.5283417 -0.3901784
0.3939981 -0.30161914 -0.14896046 -0.07624252 0.5016081 0.60912687
0.95877355 -1.0732892 -0.32319266 -0.37532297 -0.31971925 0.41914135
-0.05610953 0.63419735 -0.52241063 -0.846603 -0.2362942 -1.1333421
-0.14092262 -0.57522506 -0.07312436 -0.5817808 0.4199759 -0.963604
1.0839058 -1.7103524 0.43462828 -0.24551241 -0.28709668 0.2570477
-0.7324256 0.8328641 -0.04278511 0.20501815 0.03015664 -0.63581073
-0.09937622 0.01731233 -0.68177855 0.09906986 0.39832842 1.8503052
-0.83984244 -0.5798068 -0.35019195 0.38815477 -0.13768867 0.1803412
-0.52110857 0.7061032 -0.6667104 0.6245882 0.2332336 -0.418072
0.06681974 -0.11160722 0.19476333 0.82621646 -0.0686693 2.2890017
-0.9610842 0.6006808 -0.09961901 -1.0347204 0.8402107 0.7536357
0.31567344 -1.5147772 0.03306683 0.7105547 0.0352441 -0.26449972
0.5906393 -0.27548772 -0.372164 1.0058266 0.30612645 -0.2258417
-0.08533014 0.15617533 0.7447393 0.49767473 -0.01302241 -0.536043
0.46795052 0.46544397 0.18773358 0.6175446 0.14819477 0.42992952
-0.03605575 -0.57232326 -1.4548228 -0.81039476 -0.13943383 1.5175649
-0.2843629 -0.232748 -0.4229644 -0.7298095 -0.29073092 0.8845528
-0.5198069 -0.23796122 -1.058535 -0.90413094 0.4506844 0.28503323
0.11144357 -1.0807526 0.07496601 0.30894116 -0.83169127 -1.3185304
-1.0938618 0.33348647 -0.9073635 -0.25145793 -0.8951084 -1.1188449
0.34031034 0.18298915 1.7563888 -0.13519074 0.26808715 0.1499772
-0.47915384 -0.48423716 -0.9223615 0.64663815 -0.16246796 -0.31870368
0.69750416 -0.3255005 -0.42387024 0.33777228 -0.73689 0.13055909
1.0352131 1.1518149 0.82180136 -0.80647653 0.8979976 -0.6130109
1.113998 -0.1851002 -0.34534922 0.8916354 -0.76530576 0.5083791
0.36166778 -0.58641666 -0.86562306 -0.3022028 0.02642508 -0.29786962
0.39842927 0.67033255 0.15280959 0.21836817 0.8471646 0.62630886
-0.11127103 -0.4299203 0.41457883 0.96032065 -0.06025122 -0.8476141
0.8582872 -0.038003 -0.1451047 -0.29657796 -1.1074432 -0.4336099
-0.9049285 0.18589327 0.9284405 -1.3546739 -0.16571566 -0.23237836
-1.5258323 -0.5854085 -0.07393724 0.42657718 -0.6589764 -0.1263588
-0.7353796 -0.43812346 -1.0797092 -1.3134705 1.8703055 -0.64914703
-0.11621851 -1.2884433 0.4303998 0.64681 0.7936431 -0.44203869
1.1351632 -1.2026681 -0.63276416 -0.11386143 -0.17282403 0.29786438
0.09483062 -0.6159781 -0.74657375 -0.56598014 0.13756202 -0.89718467
0.8250776 0.2708273 0.44538382 -0.41844693 -0.2444072 0.18264122
1.415594 -0.2828486 0.397558 0.43531448 0.5408021 0.59974116
0.72815585 -0.61090183 0.5106263 1.1326301 -0.15888192 -0.3729829
-0.2679759 -0.34149557 0.44600707 1.2941471 0.2025914 -0.29837593
-1.0920466 0.58572716 -1.9311031 -0.5642258 0.06939743 2.1534195
1.3832241 -0.00915445 -0.0389501 -0.731526 0.2634947 -0.14767253
-0.42002177 -0.70533943 -0.30523795 0.3162075 0.5608196 0.73812044
-0.9049236 1.5410453 6.8798428 1.11517 -1.1879162 0.63332266
0.54516834 0.05689038 -0.61264545 -0.09452176 -0.7702279 -0.06118744
1.3996247 -0.3836509 0.47278333 0.46642053 0.2701895 0.25713196
-1.3597046 0.82721305 0.3021808 -1.1172861 0.580171 0.22620425
0.9067974 0.78865 0.23085694 0.7517645 0.41426852 -1.0930891
0.5427164 0.06041311 1.0503525 -0.44991422 0.8054675 0.6347766
-0.7532646 0.2809993 -0.3670409 0.22935283 0.19632699 0.03361164
-0.97635853 0.81139916 0.32169145 0.55901086 -0.6308001 0.32823277
-0.15520261 0.5352953 0.1408925 0.04131773 0.47550172 -0.45041355
0.3807773 1.2376821 0.2037215 -0.5508537 0.1913931 0.7905046
-0.5158744 0.6511251 -0.7356929 -0.19524735 0.11425302 1.0614297
-0.17207968 -0.35836717 -0.6325098 1.2082354 0.4761938 0.63202745
-0.48576775 0.24648009 0.26880234 -0.18780589 0.03580532 -0.17435955
-0.43351594 -1.488681 0.18603083 -1.3217636 0.12920974 -0.69026995
-1.3260081 0.9370726 -0.03037183 -1.3578541 -0.76163626 -0.27169248
-0.19936512 1.4608899 -1.8506305 -1.8265857 0.78412026 0.499948
1.0601305 -0.37759596 1.2217604 0.42303005 -0.22045712 0.9149416
0.51030034 0.3033916 1.20578 -1.1161132 0.6740797 0.648212
0.3869905 0.7790072 0.56428325 -0.532919 -1.6108719 -1.1444813
1.7128316 -1.101711 0.8637126 -0.6079198 -0.7762709 1.0461556
0.87643385 -0.44473097 0.6830201 0.21169959 -0.43171552 -0.03374099
-0.6031154 0.5441532 0.5907725 -1.0036138 -0.894471 0.93729794
1.284884 -0.2938079 -0.9051835 0.5072539 0.34832332 0.08137908
1.0797151 -1.012143 0.50575346 0.2477813 -0.19502595 -1.4967221
-0.15016209 -0.679553 0.23874533 0.7992139 1.0967959 -0.35611013
0.18018278 0.28533113 -0.13578834 -0.6617339 -1.1064342 -0.89676446
1.0865122 -0.18657424 0.22365412 0.664436 -0.09866232 1.4660815
-0.5150062 -0.41702932 0.4820713 0.2326629 0.6674643 -0.9398592
-0.13249137 -0.4190495 0.6606673 -1.1879336 0.53865355 -1.4253341
-0.01761916 -1.6849554 0.47731772 -0.13752246 -0.18124346 0.25568727
-0.17121613 0.21285659 -0.16185284 0.7425798 -0.6240099 0.7661801
1.5719811 -0.45718816 -0.02351178 0.13348074 -0.70661634 -0.08314329
0.5894479 -0.7638928 -0.28385293 -1.2468781 0.59332776 0.2794149
-0.11500426 -0.02501345 0.15685618 -0.07971042 0.06472037 -0.555634
0.30844167 -0.70005757 -0.25969064 0.4007167 -0.87182677 0.5746295
0.29602033 0.24088004 -0.3278689 0.0732094 0.34848338 -0.2669424
0.14616284 0.36365882 -0.23405963 0.29702497 0.21547839 0.46001205
-0.287776 -0.4069378 -0.39199963 0.42427436 0.21911633 0.81665057
0.24205418 -1.4053217 -1.4836271 -0.12130512 0.48246047 -0.27109036
-0.4807428 0.8828203 0.11050973 1.1727116 0.13436557 -0.64564127
-0.88277715 0.69589156 0.24154596 -0.9326672 -0.22702684 0.5078642
0.02386951 -0.92071646 0.04296081 -0.01293717 -0.05931431 -0.09947536
0.33805567 -0.31621176 0.4834542 -0.81820357 -0.03847983 0.6982862
-0.40472385 -0.6225043 1.2044677 -0.3630264 -0.45939505 0.5066153
1.5223137 -0.24343711 -0.28417066 -0.8237479 0.42125058 -0.07472196
0.02261738 -1.1940926 -0.4518533 0.97395164 0.27343535 -0.3864265
0.8887156 0.23579136 0.94047105 0.766845 0.5870793 -1.1200835
0.23591955 0.8610611 1.2375996 -1.6618283 -0.3818183 0.09724746
-0.79423136 0.8673505 0.33622858 0.09854977 0.42865065 -0.03100761
0.48848906 -0.06415026 -1.221559 -0.06532896 0.79325897 0.36738974
0.808032 -0.1416469 -0.6849713 0.15114987 -0.14581221 -0.16596384
-0.09668945 0.5792635 -0.1003684 -1.6435858 0.02333631 0.10781994
-0.7782356 -0.70498866 -0.4879022 0.46619326 -0.7687136 0.9978443
-0.1794022 -0.21939857 0.16700396 0.45136794 0.44062132 -0.83306444
-0.8431853 0.5774617 0.15297955 -0.43732312 -0.42653626 -0.44004652
-0.621404 0.18668814 -0.33579433 0.48070756 0.8864225 1.2275598
0.38903296 0.36205348 0.6309622 -0.31574377 -1.0881829 -1.5747766
0.20077336 0.26586935 0.35178447 -0.02933016 -0.18361096 -0.06439596] | [11.561169624328613, 10.02457046508789] |
88b74c80-7635-462f-86cf-9245d67923c1 | dreamdiffusion-generating-high-quality-images | 2306.16934 | null | https://arxiv.org/abs/2306.16934v2 | https://arxiv.org/pdf/2306.16934v2.pdf | DreamDiffusion: Generating High-Quality Images from Brain EEG Signals | This paper introduces DreamDiffusion, a novel method for generating high-quality images directly from brain electroencephalogram (EEG) signals, without the need to translate thoughts into text. DreamDiffusion leverages pre-trained text-to-image models and employs temporal masked signal modeling to pre-train the EEG encoder for effective and robust EEG representations. Additionally, the method further leverages the CLIP image encoder to provide extra supervision to better align EEG, text, and image embeddings with limited EEG-image pairs. Overall, the proposed method overcomes the challenges of using EEG signals for image generation, such as noise, limited information, and individual differences, and achieves promising results. Quantitative and qualitative results demonstrate the effectiveness of the proposed method as a significant step towards portable and low-cost ``thoughts-to-image'', with potential applications in neuroscience and computer vision. The code is available here \url{https://github.com/bbaaii/DreamDiffusion}. | ['Yan-Pei Cao', 'Ying Shan', 'Chun Yuan', 'Yixiao Ge', 'Xintao Wang', 'Yunpeng Bai'] | 2023-06-29 | null | null | null | null | ['image-generation', 'eeg', 'eeg'] | ['computer-vision', 'methodology', 'time-series'] | [ 2.53048778e-01 8.75739828e-02 3.06028754e-01 -6.63004756e-01
-7.86666095e-01 -3.40722591e-01 6.75327778e-01 -1.40498698e-01
-4.46752042e-01 8.29436183e-01 4.14542735e-01 -9.80293006e-02
1.34082392e-01 -3.90325904e-01 -8.41879904e-01 -7.77099609e-01
1.02766842e-01 -1.66964963e-01 -5.16472757e-01 1.10301554e-01
1.88837096e-01 4.70185392e-02 -1.12422490e+00 3.26729745e-01
8.97980511e-01 1.11289203e+00 6.52683496e-01 2.97842383e-01
1.73255697e-01 6.22476101e-01 -6.86320186e-01 -4.02200550e-01
-5.74762113e-02 -6.69623554e-01 -3.37045580e-01 1.10661581e-01
9.06559974e-02 -3.35448980e-01 -6.06344402e-01 1.22278750e+00
7.21085966e-01 -6.97229579e-02 5.38915575e-01 -1.13462758e+00
-1.33858800e+00 3.91179174e-01 -6.31236792e-01 6.04759574e-01
3.12329859e-01 3.79001468e-01 4.46789324e-01 -1.20777798e+00
3.05981725e-01 6.02288842e-01 2.46120408e-01 5.95369101e-01
-1.33354557e+00 -1.10117948e+00 -1.31805837e-01 2.89399803e-01
-1.58569944e+00 -7.23480582e-01 8.27989280e-01 -4.64625090e-01
8.67167056e-01 1.26517385e-01 1.03439224e+00 1.69275868e+00
5.51140845e-01 7.33091116e-01 1.60264194e+00 -1.17657587e-01
3.02921295e-01 3.12184155e-01 7.58557096e-02 4.87072140e-01
8.46564304e-03 -8.36284235e-02 -8.82294774e-01 1.52991772e-01
9.11658943e-01 1.76397741e-01 -6.56252146e-01 3.32351953e-01
-1.35078430e+00 3.69534045e-01 6.14573836e-01 4.74226385e-01
-8.16210628e-01 1.17815048e-01 2.08966866e-01 -7.26371072e-03
5.81343234e-01 4.99026477e-01 1.07211307e-01 -2.54315287e-01
-1.28521836e+00 -1.27041996e-01 2.45299950e-01 1.11255932e+00
5.88261187e-01 4.15490270e-01 -3.30066979e-01 7.96388268e-01
-1.91867594e-02 5.49376130e-01 8.21558058e-01 -6.30410969e-01
3.10289383e-01 2.34666020e-01 3.96108590e-02 -7.94224560e-01
-3.23537976e-01 -4.56919402e-01 -8.84929478e-01 -2.13956341e-01
-3.51804346e-01 -1.78366959e-01 -9.26239908e-01 1.76354706e+00
-2.98646212e-01 5.14162660e-01 -1.10785156e-01 1.07972991e+00
9.88741457e-01 6.16994262e-01 -9.87518430e-02 -2.61357158e-01
1.43165982e+00 -8.28750014e-01 -9.53952074e-01 -3.60727996e-01
-9.26604494e-02 -4.72505987e-01 1.34192371e+00 3.08313757e-01
-1.41536772e+00 -5.42276204e-01 -1.27357268e+00 -1.49782658e-01
-1.57424986e-01 4.77933973e-01 5.24793208e-01 3.29419553e-01
-1.36807692e+00 2.72958308e-01 -9.71698940e-01 -1.68406144e-01
8.54860067e-01 3.83541554e-01 -4.07204598e-01 -3.05629894e-02
-9.66316283e-01 7.17008948e-01 1.78504214e-01 2.60686427e-01
-1.03069592e+00 -7.61021554e-01 -6.87414229e-01 6.86203837e-02
-1.78916365e-01 -6.49217963e-01 9.12909687e-01 -9.14781213e-01
-1.56287670e+00 6.98662877e-01 -3.62884820e-01 -4.82099116e-01
1.79213524e-01 -3.62084508e-01 -3.20840299e-01 4.94188219e-01
1.34334266e-01 9.49067354e-01 1.03013349e+00 -9.29257989e-01
5.75068668e-02 -5.41802168e-01 -3.46429914e-01 1.61104158e-01
-6.44775391e-01 3.79842171e-03 -6.29573464e-01 -7.79755712e-01
-1.53353453e-01 -7.13636994e-01 2.00814262e-01 -1.73381478e-01
-3.35804969e-01 1.70505103e-02 2.45773017e-01 -7.06100821e-01
7.26588368e-01 -2.26081491e+00 1.10633008e-01 -1.42085969e-01
3.64776015e-01 -3.34460400e-02 -2.63363481e-01 2.53580898e-01
-2.44320720e-01 -6.13590851e-02 -1.13959104e-01 -6.20674372e-01
-1.38105944e-01 -1.53008699e-01 -3.47187042e-01 6.79503202e-01
4.25286472e-01 1.33181775e+00 -8.65347743e-01 -2.07614824e-01
4.59377557e-01 9.91500020e-01 -2.56981283e-01 4.05039132e-01
2.50641674e-01 9.58317876e-01 -1.38003200e-01 3.03290635e-01
7.48003542e-01 -3.17240328e-01 -3.76330614e-02 -3.96861464e-01
-1.60555542e-01 3.39706421e-01 -4.28663850e-01 2.14599299e+00
-4.23092574e-01 8.59775066e-01 5.14680594e-02 -1.03009951e+00
8.02905321e-01 5.02581656e-01 4.96209711e-01 -1.22910762e+00
5.12392044e-01 -5.45185357e-02 -2.79839009e-01 -5.83152652e-01
1.36807576e-01 -2.52943873e-01 2.73937017e-01 4.56843942e-01
5.28562307e-01 -2.39102170e-01 -9.86534208e-02 1.61410674e-01
1.02767062e+00 1.66006356e-01 -3.32528017e-02 -3.89248192e-01
8.31666365e-02 -4.51235533e-01 2.62346864e-01 4.72475857e-01
8.04856420e-03 8.47324193e-01 3.28096859e-02 -4.62488458e-03
-7.65028536e-01 -1.25501943e+00 -2.76549637e-01 5.72592020e-01
9.68291312e-02 -4.21335310e-01 -9.11246240e-01 -9.51479077e-02
-4.51336205e-01 8.00406873e-01 -8.38771224e-01 -2.35389501e-01
-7.66307041e-02 -9.68161404e-01 4.76280898e-01 5.55031776e-01
4.10397559e-01 -9.32859480e-01 -4.64024723e-01 1.91651165e-01
-4.35592145e-01 -1.12795854e+00 -5.11499405e-01 1.81602120e-01
-7.55958259e-01 -6.05685294e-01 -1.09972930e+00 -8.14910650e-01
9.40870881e-01 3.89313579e-01 6.84239745e-01 -2.10034296e-01
-5.51879644e-01 3.66758317e-01 -2.60897726e-01 -5.61862171e-01
3.22009355e-01 -4.50562984e-01 1.87665030e-01 1.41471758e-01
4.98544782e-01 -1.04952323e+00 -1.17465198e+00 -8.76582339e-02
-8.31697226e-01 5.57561398e-01 8.45105410e-01 9.19111490e-01
6.24859869e-01 -1.49312168e-01 7.95985401e-01 -4.00775611e-01
9.28402543e-01 -6.24285340e-01 -2.87296832e-01 -1.17833339e-01
-6.53417587e-01 -1.55906394e-01 7.10362852e-01 -4.02852148e-01
-1.12685180e+00 -1.90753281e-01 -2.08888486e-01 -6.18376493e-01
-2.45649427e-01 4.20270532e-01 1.52223378e-01 1.07953556e-01
7.36523807e-01 7.64394045e-01 -1.29739895e-01 -3.35675120e-01
3.12874198e-01 7.54760742e-01 8.38335216e-01 -3.70217055e-01
3.21228892e-01 5.31197906e-01 -5.69983006e-01 -7.54528522e-01
-6.26010716e-01 -3.06675225e-01 -5.36706984e-01 -2.70125926e-01
9.25410688e-01 -1.39220059e+00 -4.29903567e-01 2.81505913e-01
-1.06849110e+00 -4.31215435e-01 1.98888630e-02 8.10698688e-01
-5.33051968e-01 8.08007941e-02 -8.96702409e-01 -6.08244538e-01
-8.14230502e-01 -1.10307097e+00 1.11426473e+00 3.12422186e-01
-1.34230480e-01 -6.60034418e-01 -2.63868183e-01 4.12756950e-01
3.36368322e-01 -2.78123897e-02 5.88257670e-01 -1.25637963e-01
-4.91061240e-01 -8.80599320e-02 -3.69015187e-01 4.92649227e-01
3.28847915e-01 -5.71176708e-01 -1.29388392e+00 -4.05979127e-01
4.39796060e-01 -5.24199963e-01 6.25740826e-01 5.08549333e-01
1.37310684e+00 -2.94564754e-01 -7.33115375e-02 8.14378440e-01
1.40063512e+00 2.21182734e-01 8.33107591e-01 -4.85731326e-02
4.71379817e-01 2.25892469e-01 3.46911908e-03 5.81896663e-01
5.38345039e-01 3.52796912e-01 -1.20923147e-02 -2.92326480e-01
-2.94435084e-01 -1.28999159e-01 2.94333726e-01 1.29312444e+00
-7.96865523e-02 3.50964777e-02 -6.61513746e-01 6.68057501e-01
-1.38807857e+00 -8.25646758e-01 1.60377324e-01 2.00560713e+00
9.95359957e-01 -1.70304179e-01 -2.34174266e-01 -1.06114358e-01
4.05635387e-01 -6.74387664e-02 -5.72990596e-01 -7.92708471e-02
-1.86456218e-02 5.96616030e-01 8.48000944e-02 1.24391422e-01
-6.96429670e-01 6.97251558e-01 6.02937555e+00 4.99056429e-01
-1.47061741e+00 6.62223995e-01 4.99362141e-01 -6.75322771e-01
-2.64913917e-01 -4.88364905e-01 -2.04357639e-01 7.88150072e-01
1.25514138e+00 -3.60908210e-01 8.40594709e-01 3.63160402e-01
5.55639088e-01 -8.89100060e-02 -1.11073148e+00 1.50811493e+00
4.47052151e-01 -1.34054327e+00 -1.83051661e-01 -6.68135583e-02
6.06678128e-01 2.98414081e-01 4.13695216e-01 5.62733375e-02
-2.07257122e-01 -1.10316586e+00 9.53610301e-01 6.23056293e-01
1.04878581e+00 -5.69143116e-01 5.28802156e-01 1.87774524e-01
-8.74513149e-01 1.61857847e-02 -4.89410400e-01 -6.69675991e-02
5.56099117e-02 7.16254354e-01 -7.36443937e-01 3.77383262e-01
8.97578418e-01 8.84427130e-01 -5.62696278e-01 8.71684015e-01
-3.19974691e-01 6.35478973e-01 6.96771070e-02 1.10239714e-01
-8.22572112e-02 -2.78482765e-01 2.06632689e-01 1.26730931e+00
6.08978570e-01 4.04371023e-01 -2.71103323e-01 1.46061587e+00
-2.78929412e-01 2.18106508e-02 -5.92181027e-01 -2.31533200e-01
4.95884389e-01 1.38611174e+00 -6.70040488e-01 -3.71851921e-01
-6.30082905e-01 1.45195830e+00 3.29568475e-01 5.69170356e-01
-1.01432645e+00 -5.22578299e-01 3.04921567e-01 -4.19426151e-02
-9.11957920e-02 -2.45846674e-01 -5.45332849e-01 -1.34507906e+00
2.76522219e-01 -6.05869055e-01 -3.00900072e-01 -1.37708449e+00
-1.14531910e+00 9.13592517e-01 -9.00244117e-02 -1.03140295e+00
9.34763402e-02 -3.62515181e-01 -6.64836824e-01 8.89414072e-01
-1.47052526e+00 -1.09368050e+00 -5.85559547e-01 8.91489029e-01
6.28416002e-01 1.15816586e-01 8.94869089e-01 4.80596632e-01
-7.19921291e-01 4.60151672e-01 1.76245004e-01 -1.29430508e-02
6.56105876e-01 -1.06538284e+00 3.77104878e-01 7.72407830e-01
2.53966600e-01 8.49183023e-01 5.35530090e-01 -3.58781129e-01
-1.56130612e+00 -1.21942639e+00 3.63353640e-01 -3.86173010e-01
6.39392078e-01 -8.43594074e-01 -7.28395641e-01 6.40921175e-01
8.04883420e-01 -2.22098172e-01 8.77626777e-01 -1.89914897e-01
-3.28523526e-03 -2.29869351e-01 -8.91116381e-01 6.64316714e-01
8.63882303e-01 -9.12774324e-01 -7.38884747e-01 4.81986761e-01
3.94182205e-01 -2.29249746e-01 -9.60713744e-01 -8.40075910e-02
4.00674045e-01 -8.26628506e-01 8.01836550e-01 1.47687733e-01
3.63947004e-01 1.26038818e-02 7.84937218e-02 -1.63237262e+00
-4.98880923e-01 -6.26120806e-01 9.03987437e-02 1.08412111e+00
3.36310029e-01 -6.27294064e-01 5.63679993e-01 6.84733868e-01
-3.74099642e-01 -7.23130822e-01 -9.04534519e-01 -6.48595750e-01
-7.42524415e-02 -4.36361998e-01 5.95867932e-01 7.11348474e-01
4.55243379e-01 3.84237409e-01 -4.90483105e-01 1.69061199e-01
5.64204574e-01 -7.68629685e-02 2.42958337e-01 -6.80289090e-01
-1.20087892e-01 -1.43290058e-01 -4.60780114e-01 -1.00116396e+00
1.16183847e-01 -1.12862086e+00 2.62613684e-01 -1.59068906e+00
4.59117383e-01 -7.82320499e-02 -6.76572621e-01 5.53235650e-01
-7.26061836e-02 7.38285244e-01 1.81013986e-01 1.60393596e-01
-3.84291291e-01 9.11612988e-01 1.28177440e+00 -1.52420804e-01
-4.05858718e-02 -6.01124287e-01 -9.27284956e-01 2.99925357e-01
1.06392169e+00 -3.63456011e-01 -7.76744545e-01 -7.20263362e-01
-1.54350236e-01 -1.27174426e-02 4.68495548e-01 -1.22426748e+00
3.41759562e-01 2.16239527e-01 7.88780510e-01 -1.97490022e-01
8.29356015e-01 -4.39507395e-01 1.21191256e-01 8.33024010e-02
-1.97703540e-01 2.42050529e-01 3.55257869e-01 4.79775459e-01
-1.92284167e-01 1.16664693e-01 7.14753509e-01 -3.33439648e-01
-3.36053401e-01 3.68080407e-01 -5.22601604e-01 7.78113026e-03
9.26736474e-01 -9.24508348e-02 -5.12845635e-01 -4.33477849e-01
-6.68040037e-01 -1.09073641e-02 3.22826147e-01 5.32815993e-01
1.15067959e+00 -1.36639369e+00 -6.25729203e-01 6.56998515e-01
9.02639925e-02 -3.35029066e-01 4.31286544e-01 1.38820744e+00
-7.61848912e-02 5.65039694e-01 -4.83438343e-01 -6.27878964e-01
-9.38747108e-01 4.44740683e-01 -5.29033318e-03 6.12068951e-01
-1.05312896e+00 1.01126039e+00 5.34538448e-01 1.79889292e-01
1.58111632e-01 -3.54304999e-01 7.88293779e-02 -1.47171512e-01
9.34269488e-01 -7.52258077e-02 2.24075928e-01 -5.20070851e-01
-3.05446714e-01 1.50547832e-01 -1.19601227e-01 -3.74685824e-01
1.63584745e+00 -2.64298767e-01 -7.07423314e-02 4.81825262e-01
1.16281044e+00 -4.15283650e-01 -1.33925903e+00 2.75968369e-02
-4.34335828e-01 -4.96425480e-01 4.39069062e-01 -8.71208429e-01
-1.21296871e+00 1.22069192e+00 7.85238028e-01 -4.09357905e-01
1.32258213e+00 -6.56215921e-02 7.45744407e-01 1.88859552e-01
5.65029323e-01 -9.94697571e-01 4.55113828e-01 -5.90086952e-02
1.26171470e+00 -9.85470533e-01 -1.07497126e-01 1.58293858e-01
-8.30134332e-01 7.90084302e-01 4.57981080e-01 -2.39715502e-01
6.07350826e-01 3.54302138e-01 -2.21679732e-02 -3.94225389e-01
-8.22105587e-01 1.86035112e-01 1.91533566e-01 6.94093645e-01
4.87517536e-01 1.69740111e-01 -1.35505483e-01 8.87094676e-01
-3.52542192e-01 5.35304010e-01 4.76949930e-01 7.17935264e-01
3.66288945e-02 -6.85152829e-01 -1.49570525e-01 8.35161805e-01
-5.66434860e-01 -5.46117127e-01 -1.73962444e-01 2.26671830e-01
3.23449410e-02 9.68196392e-01 5.72045334e-02 -5.86200714e-01
1.31337911e-01 1.80085242e-01 7.97335446e-01 -6.33267641e-01
-3.11980754e-01 3.37360978e-01 -2.42103904e-01 -6.00606620e-01
-3.30003887e-01 -4.13743436e-01 -1.20116711e+00 -7.63447583e-02
-2.38029599e-01 1.35798633e-01 8.28529954e-01 7.79223800e-01
9.22762156e-01 6.98911905e-01 4.58536953e-01 -1.17917132e+00
2.50352807e-02 -1.18803895e+00 -6.77658379e-01 3.55038613e-01
3.81485432e-01 -5.87268651e-01 -1.33206919e-01 3.06141049e-01] | [10.833364486694336, 2.5213890075683594] |
48dd7d9a-7d74-4f1c-8a57-bfbe761070f7 | pose-guided-image-generation-from-misaligned | 2202.00843 | null | https://arxiv.org/abs/2202.00843v1 | https://arxiv.org/pdf/2202.00843v1.pdf | Pose Guided Image Generation from Misaligned Sources via Residual Flow Based Correction | Generating new images with desired properties (e.g. new view/poses) from source images has been enthusiastically pursued recently, due to its wide range of potential applications. One way to ensure high-quality generation is to use multiple sources with complementary information such as different views of the same object. However, as source images are often misaligned due to the large disparities among the camera settings, strong assumptions have been made in the past with respect to the camera(s) or/and the object in interest, limiting the application of such techniques. Therefore, we propose a new general approach which models multiple types of variations among sources, such as view angles, poses, facial expressions, in a unified framework, so that it can be employed on datasets of vastly different nature. We verify our approach on a variety of data including humans bodies, faces, city scenes and 3D objects. Both the qualitative and quantitative results demonstrate the better performance of our method than the state of the art. | ['Kun Zhou', 'Yin Yang', 'Tianjia Shao', 'He Wang', 'Jiawei Lu'] | 2022-02-02 | null | null | null | null | ['pose-guided-image-generation'] | ['computer-vision'] | [ 3.30086827e-01 -1.68719411e-01 9.56085324e-02 -3.95468056e-01
-2.27320313e-01 -6.39991522e-01 7.49093175e-01 -4.38552350e-01
-1.71826761e-02 7.22870946e-01 3.65784615e-01 4.33820873e-01
6.42495081e-02 -6.60393238e-01 -6.15007043e-01 -7.51244664e-01
4.19605613e-01 1.78297997e-01 3.46392632e-01 -2.75346398e-01
2.48948455e-01 7.08652437e-01 -1.92717123e+00 2.63580009e-02
5.40667355e-01 7.96041131e-01 2.07307070e-01 3.72117430e-01
5.13535971e-03 2.42867723e-01 -5.97082734e-01 -7.23969102e-01
5.17637849e-01 -4.75317776e-01 -4.05534387e-01 7.51050055e-01
5.46025455e-01 -3.21538001e-01 -7.92269856e-02 1.21237850e+00
5.46047628e-01 -7.77760670e-02 5.07276058e-01 -1.23087716e+00
-4.24941480e-01 -6.23763055e-02 -8.97827625e-01 -1.93844885e-01
7.73741126e-01 3.13653871e-02 5.50784290e-01 -9.84696209e-01
8.48721147e-01 1.31441069e+00 9.37407464e-02 5.86578608e-01
-1.25206554e+00 -7.69006968e-01 1.17215358e-01 -4.23145108e-02
-1.32075953e+00 -7.39020109e-01 9.29372609e-01 -3.36948633e-01
2.39857942e-01 1.69898912e-01 6.07425690e-01 1.39929628e+00
4.95938547e-02 4.16042626e-01 1.18792129e+00 -5.58540702e-01
2.04793602e-01 5.60455561e-01 -5.57232678e-01 4.12539601e-01
4.21308875e-01 7.26782829e-02 -6.12334967e-01 -1.34001806e-01
9.75865185e-01 -4.51992499e-03 -3.79549801e-01 -9.37255859e-01
-1.27213502e+00 6.37968779e-01 1.48739398e-01 2.02883720e-01
-1.49872601e-01 -2.84366578e-01 -1.40469983e-01 3.25432792e-02
4.10187870e-01 3.07346016e-01 -3.14779103e-01 8.76144022e-02
-5.64206064e-01 4.44871455e-01 6.23267651e-01 1.12893999e+00
6.65730715e-01 -5.24506867e-02 2.38575861e-01 7.92816162e-01
3.34701151e-01 6.55923247e-01 1.94801345e-01 -1.00393605e+00
5.67932844e-01 5.81737339e-01 2.90181696e-01 -1.40069747e+00
-7.64764398e-02 -3.31030130e-01 -9.59188282e-01 4.23956603e-01
5.33508658e-01 -2.79373266e-02 -6.65644884e-01 1.93714082e+00
7.30672061e-01 -1.85606495e-01 4.03431691e-02 9.95833457e-01
6.19390309e-01 3.99351716e-01 -4.27075207e-01 -3.95267785e-01
1.18145847e+00 -5.08682072e-01 -6.93731844e-01 -3.59276891e-01
2.06652824e-02 -1.15181541e+00 6.96421206e-01 5.31975150e-01
-1.08969438e+00 -7.46183157e-01 -7.93643773e-01 1.96688458e-01
-2.68106878e-01 1.36822490e-02 4.39622968e-01 7.95243680e-01
-9.22551751e-01 2.87708461e-01 -4.17414486e-01 -5.90821683e-01
1.93926319e-01 2.23707542e-01 -5.70073366e-01 -2.53920168e-01
-7.49151826e-01 8.20302606e-01 2.38742054e-01 1.52294278e-01
-5.45373678e-01 -3.34424466e-01 -6.68314159e-01 -1.88965604e-01
6.62561476e-01 -7.59123564e-01 9.21401501e-01 -1.16642439e+00
-1.62200880e+00 8.80823195e-01 -4.95462231e-02 2.57308871e-01
6.29233479e-01 -1.82540327e-01 -2.81022847e-01 4.60270345e-02
5.11223031e-03 7.21607804e-01 9.98977184e-01 -1.55569911e+00
-5.10797977e-01 -6.98044658e-01 3.39498669e-01 4.58124846e-01
-3.08206201e-01 1.54518038e-01 -6.91138983e-01 -5.50456703e-01
1.43153146e-01 -1.15638363e+00 -1.49621233e-01 2.91613936e-01
-4.00713742e-01 2.17625663e-01 9.46065962e-01 -4.01392639e-01
7.67790854e-01 -2.29347539e+00 5.37174523e-01 1.37230337e-01
8.85256231e-02 1.67804837e-01 1.93049069e-02 4.72474515e-01
-1.03319891e-01 -7.01288730e-02 2.34521739e-02 -4.10612762e-01
-3.91930014e-01 2.75092870e-01 -6.70786873e-02 4.95486140e-01
4.59426939e-02 4.87520874e-01 -8.46106708e-01 -5.18436611e-01
4.44125354e-01 5.87529957e-01 -4.36770678e-01 3.25429350e-01
-5.52343298e-03 7.85335243e-01 -5.25897682e-01 5.78264832e-01
8.67058277e-01 -1.74021259e-01 1.00864656e-01 -3.60842019e-01
1.45645896e-02 -2.38217473e-01 -1.58787572e+00 1.64487600e+00
-3.39516759e-01 3.77120197e-01 2.11082950e-01 -5.28945565e-01
8.65137935e-01 4.75207716e-01 5.23086786e-01 -4.37481850e-01
9.10447314e-02 4.19952199e-02 -7.59676052e-03 -5.56163192e-01
3.65933120e-01 -1.75542280e-01 2.53758550e-01 1.45438209e-01
-1.80975702e-02 -3.47469896e-01 2.57024884e-01 1.71388984e-01
4.78307217e-01 3.57904345e-01 4.83569533e-01 -1.79814771e-02
4.37107474e-01 -3.16585630e-01 5.52529216e-01 2.23593578e-01
5.32298945e-02 1.01752126e+00 3.41050804e-01 -2.04035372e-01
-1.10494173e+00 -9.08492863e-01 -9.02668759e-02 5.05301595e-01
2.72230953e-01 -2.27834404e-01 -6.20403588e-01 -4.23263848e-01
-1.73182905e-01 2.98758686e-01 -5.61722994e-01 1.70371234e-02
-3.48612517e-01 -6.28827035e-01 1.67453036e-01 4.59482104e-01
5.23848712e-01 -8.35968792e-01 -9.31557178e-01 3.45283449e-02
-2.29087472e-01 -1.38969743e+00 -4.31462258e-01 -4.50933695e-01
-9.40862417e-01 -1.10290146e+00 -8.49321604e-01 -3.30822825e-01
9.90871608e-01 6.82591498e-01 1.04042876e+00 -7.28812292e-02
-1.11702636e-01 4.69620973e-01 -5.05095005e-01 -3.05547833e-01
-4.85841364e-01 -4.02077407e-01 1.50002405e-01 3.96377325e-01
-9.79447663e-02 -5.99291444e-01 -5.24862230e-01 4.98202682e-01
-1.15098917e+00 4.62583959e-01 6.55891895e-01 6.62150443e-01
1.31465152e-01 9.43933874e-02 3.41413349e-01 -7.68255591e-01
2.72411317e-01 -2.73105919e-01 -8.09201360e-01 2.21678853e-01
-3.56882095e-01 -7.49663264e-02 4.26009864e-01 -3.17740142e-01
-1.40652192e+00 1.55395791e-01 1.77606389e-01 -4.48894411e-01
-4.67182249e-01 2.36159980e-01 -6.49245679e-01 -7.61677623e-02
5.53132474e-01 3.42415795e-02 1.35371700e-01 -3.06757003e-01
3.60027432e-01 3.72514963e-01 4.57868695e-01 -4.52968448e-01
1.06952572e+00 6.73922598e-01 1.87718108e-01 -9.15889382e-01
-7.19789267e-01 -2.58587927e-01 -8.49331498e-01 -3.90841514e-01
6.69516444e-01 -7.19013691e-01 -3.20152134e-01 7.11351037e-01
-1.25186408e+00 3.60246956e-01 7.34558403e-02 5.49709022e-01
-3.57674003e-01 3.33697617e-01 -7.87582174e-02 -7.31138229e-01
9.87735167e-02 -1.25459576e+00 1.17922986e+00 4.80546802e-01
-8.31654146e-02 -8.72677028e-01 -9.77486596e-02 4.59634066e-01
2.60189444e-01 5.24535000e-01 6.70562446e-01 -1.29639015e-01
-7.96016991e-01 -7.92910457e-02 -1.93933979e-01 4.16206777e-01
6.42346978e-01 3.32459480e-01 -9.81236398e-01 -3.52119982e-01
1.18501686e-01 -1.44396335e-01 1.63417459e-01 3.08457494e-01
8.17826867e-01 -1.28481090e-01 -4.04169887e-01 4.50292438e-01
1.44549680e+00 2.76739925e-01 6.99075937e-01 9.00970250e-02
6.22555077e-01 8.94877195e-01 7.14666605e-01 4.45475847e-01
7.95167238e-02 1.11336327e+00 5.83151758e-01 -1.87381446e-01
1.61547083e-02 -1.41686574e-01 2.79179990e-01 4.66981739e-01
-4.51807469e-01 -3.35183531e-01 -5.36663830e-01 3.45868289e-01
-1.67233706e+00 -1.02592480e+00 1.10067083e-02 2.59826899e+00
5.60352147e-01 -6.02321923e-02 6.99187368e-02 8.71802494e-02
6.67880714e-01 2.37807065e-01 -5.30361772e-01 -1.00672003e-02
-8.85984600e-02 -2.42224082e-01 1.38629571e-01 9.68156606e-02
-6.98594809e-01 5.42081833e-01 6.09120798e+00 5.56822658e-01
-1.10263240e+00 -2.57216007e-01 6.36234760e-01 -2.44777557e-02
-2.29505315e-01 1.64712351e-02 -8.03888857e-01 3.87417525e-01
2.44731560e-01 -1.03659451e-01 4.26564634e-01 6.59269571e-01
1.51131645e-01 -2.91861981e-01 -1.07535934e+00 1.16570044e+00
4.51781631e-01 -1.01520109e+00 -5.42281196e-02 3.56219560e-01
8.94426584e-01 -4.42339599e-01 1.24574430e-01 -2.66420156e-01
1.03737332e-01 -6.88650727e-01 8.19995821e-01 4.17844325e-01
7.00396597e-01 -6.29970193e-01 5.91906726e-01 4.15518254e-01
-9.13247824e-01 1.88268185e-01 -2.84401149e-01 -6.29945397e-02
1.98304370e-01 7.15931416e-01 -6.58182859e-01 8.16854477e-01
8.25102746e-01 6.07539475e-01 -7.06819177e-01 7.77528405e-01
-2.64356226e-01 2.42027361e-02 -3.93553168e-01 2.20104620e-01
-2.63229012e-01 -4.55541790e-01 5.21357298e-01 5.96064568e-01
5.98865688e-01 8.32083896e-02 6.83031697e-03 8.96556258e-01
8.83221701e-02 1.51526839e-01 -9.71626282e-01 1.53917193e-01
3.78633201e-01 1.44765699e+00 -6.77110136e-01 -1.55579224e-01
-6.87455416e-01 8.12862337e-01 3.64513248e-02 4.21991169e-01
-8.78864586e-01 1.88511848e-01 5.89300096e-01 2.35220358e-01
2.64163852e-01 -2.69162476e-01 5.84483892e-02 -1.37487102e+00
3.32272649e-01 -1.15308774e+00 1.09898984e-01 -1.06050420e+00
-1.16014194e+00 6.75270081e-01 3.76247108e-01 -1.65786803e+00
-3.66829842e-01 -5.64632475e-01 -2.86975980e-01 7.26248682e-01
-1.30054712e+00 -1.13316882e+00 -5.59267402e-01 6.12961531e-01
5.51627576e-01 -1.36155739e-01 6.81300104e-01 2.70525932e-01
-3.69848907e-01 1.17066212e-01 9.07746796e-03 -1.13366298e-01
9.68941867e-01 -8.32744300e-01 8.38401765e-02 9.14142787e-01
3.71515810e-01 5.34942329e-01 8.27667773e-01 -3.57245415e-01
-1.52393937e+00 -7.85516381e-01 5.41237116e-01 -5.61028421e-01
1.75065443e-01 -4.22841251e-01 -6.02489173e-01 4.92630094e-01
3.18245322e-01 -6.22121058e-02 5.04875839e-01 -4.69386689e-02
-1.62137300e-01 -2.72476703e-01 -1.11389852e+00 7.41561413e-01
1.03510988e+00 -5.32525638e-03 -3.56200725e-01 1.24285724e-02
2.58148074e-01 -6.01758718e-01 -5.62225044e-01 4.22125787e-01
6.32353842e-01 -1.65629745e+00 1.02051592e+00 -9.60933268e-02
4.49794322e-01 -4.18421865e-01 -2.60821700e-01 -1.48714352e+00
-1.09403417e-01 -3.74931633e-01 1.38053700e-01 1.58623672e+00
3.57098877e-02 -6.24262631e-01 5.44576883e-01 7.38779604e-01
2.62311071e-01 -5.60587406e-01 -8.73450458e-01 -5.83353400e-01
-3.16692740e-01 -1.17105700e-01 7.16085374e-01 6.88582778e-01
-4.05226350e-01 4.72293407e-01 -6.72725499e-01 2.52642751e-01
6.04521513e-01 3.28209370e-01 1.36883378e+00 -1.35813594e+00
-2.26194158e-01 -2.02698663e-01 -5.22375584e-01 -9.26514030e-01
-1.12125739e-01 -3.02621484e-01 -1.32864892e-01 -1.19468486e+00
3.91951144e-01 -3.39258105e-01 2.02962101e-01 1.98101699e-01
-1.91857040e-01 2.50947565e-01 3.07553619e-01 1.80370107e-01
-1.19533345e-01 5.75543582e-01 1.50795531e+00 2.02413380e-01
-6.45476282e-02 1.29924752e-02 -7.96898127e-01 9.71527338e-01
5.22814274e-01 -3.47114474e-01 -5.51642835e-01 -5.45968115e-01
3.65868032e-01 2.02510849e-01 4.16252315e-01 -1.03847408e+00
-1.14342690e-01 -5.10267079e-01 6.67074740e-01 -4.05477285e-01
6.22513950e-01 -1.13171244e+00 7.27353573e-01 3.28818685e-03
-1.42547395e-02 9.74779725e-02 -6.61668181e-02 5.87811708e-01
-2.31932923e-01 -1.64701059e-01 7.88658500e-01 -3.44727129e-01
-3.62847626e-01 3.14693928e-01 1.74188524e-01 -2.75711507e-01
1.18956447e+00 -5.10259449e-01 -1.39435664e-01 -6.67103171e-01
-3.26731414e-01 -1.96136072e-01 9.45677042e-01 7.72744298e-01
5.87268651e-01 -1.38237262e+00 -6.38208449e-01 4.65745568e-01
3.75829548e-01 1.99582666e-01 7.25736171e-02 7.48533428e-01
-1.38402790e-01 1.59559816e-01 -3.18609476e-01 -7.34550297e-01
-1.57848728e+00 6.34638906e-01 5.86095415e-02 1.90271139e-02
-4.32584673e-01 3.95016402e-01 5.55031955e-01 -2.01318413e-01
-1.02533758e-01 -2.03876764e-01 -1.50077909e-01 -1.74254889e-03
4.17040318e-01 2.46471867e-01 -4.11686227e-02 -1.03675878e+00
-2.00002402e-01 9.48332727e-01 4.74004857e-02 -1.51332468e-01
1.22743189e+00 -4.04364318e-01 2.76965089e-02 2.83351183e-01
9.45749104e-01 3.69330972e-01 -1.36532831e+00 -2.78628498e-01
-4.38389212e-01 -1.15457952e+00 -3.18678081e-01 -3.98986399e-01
-1.40034270e+00 7.50603497e-01 4.14197773e-01 7.85593092e-02
1.20999408e+00 4.09779884e-02 2.94375151e-01 1.19718527e-02
8.91980052e-01 -8.60509515e-01 3.03777725e-01 -4.70868126e-02
1.06239724e+00 -1.36481750e+00 2.03870773e-01 -7.05718160e-01
-6.48218513e-01 1.15640569e+00 6.57210290e-01 1.76131159e-01
3.78310174e-01 1.46098152e-01 2.90312499e-01 -1.14525087e-01
-5.52083433e-01 -2.80632395e-02 2.70186812e-01 6.61643624e-01
3.85361910e-01 -2.36495644e-01 -4.67516743e-02 -3.68574746e-02
-8.96288604e-02 -1.26524009e-02 7.22587645e-01 9.07054365e-01
2.25163754e-02 -1.30191791e+00 -7.71339238e-01 1.40915826e-01
-3.16511840e-01 3.73334348e-01 -3.25404257e-01 8.64208221e-01
2.52871901e-01 1.03789401e+00 -2.08424866e-01 -9.21123698e-02
3.98157269e-01 -1.63510114e-01 7.64030933e-01 -6.67888165e-01
-9.60849375e-02 3.84490013e-01 -2.12204665e-01 -5.24359047e-01
-1.08136511e+00 -8.31913292e-01 -7.44522274e-01 -1.46574914e-01
-6.08359158e-01 -1.44209519e-01 5.28074443e-01 9.04825211e-01
3.76559228e-01 2.00489953e-01 8.29368234e-01 -1.15866959e+00
-3.90363932e-01 -8.20777833e-01 -6.67079747e-01 8.97864938e-01
1.06962301e-01 -1.03852177e+00 -3.92629743e-01 4.49875861e-01] | [9.30114459991455, -2.6328470706939697] |
c7e92911-3937-45c3-b020-6638e9abc5ab | diverse-single-image-generation-with | 2102.04780 | null | https://arxiv.org/abs/2102.04780v4 | https://arxiv.org/pdf/2102.04780v4.pdf | Diverse Single Image Generation with Controllable Global Structure | Image generation from a single image using generative adversarial networks is quite interesting due to the realism of generated images. However, recent approaches need improvement for such realistic and diverse image generation, when the global context of the image is important such as in face, animal, and architectural image generation. This is mainly due to the use of fewer convolutional layers for mainly capturing the patch statistics and, thereby, not being able to capture global statistics very well. We solve this problem by using attention blocks at selected scales and feeding a random Gaussian blurred image to the discriminator for training. Our results are visually better than the state-of-the-art particularly in generating images that require global context. The diversity of our image generation, measured using the average standard deviation of pixels, is also better. | ['Ranga Rodrigo', 'Chamira Edussooriya', 'Sutharsan Mahendren'] | 2021-02-09 | null | null | null | null | ['single-image-generation'] | ['computer-vision'] | [ 2.83774942e-01 2.04694927e-01 4.17081326e-01 -5.19719832e-02
-3.78207803e-01 -4.46462721e-01 7.13350594e-01 -2.13292390e-01
-2.40865126e-01 1.18787754e+00 4.52425815e-02 2.18337655e-01
1.29339948e-01 -1.08933473e+00 -9.07460153e-01 -9.30785179e-01
5.05400002e-02 2.69505531e-01 2.86035687e-01 -3.28732997e-01
7.49595836e-02 7.06808507e-01 -1.63078141e+00 3.04407686e-01
8.82886887e-01 9.23101366e-01 4.62225229e-01 8.07490289e-01
-1.02710523e-01 4.47873145e-01 -1.13660228e+00 -3.62411588e-01
2.73511261e-01 -1.04911292e+00 -3.94095838e-01 1.94766894e-01
6.48640394e-01 -3.17418844e-01 1.11129567e-01 1.04334307e+00
5.82024276e-01 8.02554861e-02 7.72023082e-01 -9.62121725e-01
-8.40805233e-01 3.07329923e-01 -4.40754354e-01 -1.60886034e-01
3.54568720e-01 4.38569665e-01 5.41667342e-01 -4.78909582e-01
8.54583383e-01 1.44011259e+00 3.81534457e-01 7.88034201e-01
-1.56169617e+00 -5.76241851e-01 -1.22409776e-01 -5.46049364e-02
-1.17963338e+00 -1.94766268e-01 7.77725399e-01 -3.63559693e-01
4.03758883e-01 2.82726109e-01 4.99704510e-01 1.35878658e+00
3.46161693e-01 2.23214343e-01 1.31623983e+00 -3.83715183e-01
3.73987257e-01 2.36900449e-01 -7.32528627e-01 3.53889406e-01
2.77436435e-01 3.25044006e-01 -3.86646315e-02 1.25391141e-01
1.16186476e+00 -1.31759733e-01 -2.83653200e-01 -2.23578155e-01
-1.11906445e+00 8.57614756e-01 7.79768288e-01 4.73569483e-01
-6.37632906e-01 4.34586376e-01 -6.76229000e-02 1.88944638e-01
3.26803088e-01 9.19516325e-01 -1.12745993e-01 1.22242719e-01
-1.21307206e+00 3.83227736e-01 6.98663414e-01 7.01427042e-01
8.94169092e-01 4.93783474e-01 -5.28556108e-01 6.78990364e-01
-1.70718938e-01 6.26683712e-01 3.50167900e-01 -9.76452589e-01
-2.86705922e-02 4.97576058e-01 2.12772772e-01 -9.77909863e-01
-1.05933458e-01 -5.65707922e-01 -1.13473916e+00 9.83836055e-01
6.29813135e-01 -3.86570007e-01 -1.29696500e+00 1.65782440e+00
1.44695744e-01 5.69523759e-02 5.39704673e-02 8.75982344e-01
7.70027578e-01 6.56480014e-01 -6.71407953e-02 4.74205390e-02
1.17332757e+00 -8.46336126e-01 -5.38469851e-01 -2.99314678e-01
-1.01843610e-01 -9.88540947e-01 9.76985931e-01 1.56421408e-01
-1.20891702e+00 -8.99616182e-01 -9.46376026e-01 2.32670069e-01
-5.97791731e-01 -7.29760434e-03 3.67363393e-01 5.88659823e-01
-1.19259048e+00 6.00037158e-01 -4.32381928e-01 -2.63984114e-01
5.40390134e-01 2.00114995e-01 -3.84535462e-01 -2.39437129e-02
-1.17144775e+00 9.69069958e-01 3.27377439e-01 -1.72588766e-01
-8.87585342e-01 -6.83936715e-01 -8.64763379e-01 2.23274976e-01
-2.31627692e-02 -7.48321116e-01 5.93795657e-01 -1.26835454e+00
-1.66501391e+00 4.87801164e-01 1.65000170e-01 -5.23696482e-01
7.56520569e-01 9.78619605e-02 -8.30796361e-03 1.65071830e-01
-1.40399009e-01 1.19046760e+00 1.22572780e+00 -1.62843561e+00
-2.28243083e-01 7.82736018e-02 3.14979166e-01 -1.21222347e-01
1.16954312e-01 -2.99998671e-01 -8.66059884e-02 -8.98347378e-01
-3.69265109e-01 -9.62474585e-01 -4.97566074e-01 -3.09457723e-03
-2.61253774e-01 4.29613173e-01 9.80380654e-01 -6.01162493e-01
7.88890660e-01 -2.08846116e+00 9.63164717e-02 -7.95753598e-02
-7.71282688e-02 3.89967859e-01 -4.47320431e-01 3.85204881e-01
-1.25002012e-01 3.55097204e-01 -2.25993276e-01 -3.01714599e-01
-3.37594509e-01 8.95516276e-02 -2.59030759e-01 1.04488179e-01
5.54983675e-01 1.11498678e+00 -7.02014446e-01 -3.00410479e-01
4.06031579e-01 7.65040517e-01 -5.12284994e-01 3.21726948e-01
-4.30872709e-01 9.21852171e-01 -1.22336000e-01 1.65843993e-01
9.53203022e-01 3.04542519e-02 -2.57775396e-01 -2.92617846e-02
1.19858675e-01 -2.60616958e-01 -1.10065806e+00 1.50578940e+00
-7.09258914e-01 7.81362593e-01 1.92709845e-02 -7.38406181e-01
9.39717591e-01 2.30470285e-01 1.29567876e-01 -6.31855726e-01
-1.20658958e-02 1.27699673e-02 1.77432731e-01 -1.52039692e-01
2.83613890e-01 -2.97693133e-01 2.23222170e-02 1.85385406e-01
2.25116615e-04 -7.05932438e-01 2.94289082e-01 -8.76631588e-02
8.21656644e-01 2.76811481e-01 6.23551905e-02 -1.95690677e-01
4.53741252e-01 -2.06565812e-01 2.67024159e-01 7.66566992e-01
2.49601334e-01 1.26643848e+00 6.12431586e-01 -3.75687212e-01
-1.31042027e+00 -1.01030421e+00 1.08566582e-02 4.09799635e-01
1.32447302e-01 1.59818888e-01 -1.00451076e+00 -5.99692523e-01
-2.49712884e-01 7.60655999e-01 -8.49046826e-01 -1.74755260e-01
-5.02624631e-01 -3.70332956e-01 3.91058058e-01 2.77070314e-01
6.26233280e-01 -1.53600633e+00 -7.38647819e-01 2.35838786e-01
6.94536269e-02 -1.14919662e+00 -3.14371765e-01 5.55985346e-02
-8.66409779e-01 -7.13736951e-01 -1.24702990e+00 -4.88662004e-01
8.36548805e-01 -7.01668486e-02 1.25752306e+00 5.13358153e-02
-4.92872328e-01 -1.65663272e-01 -3.61481428e-01 -3.73145521e-01
-8.00557196e-01 1.58599094e-02 -3.95529360e-01 1.73887368e-02
-3.66441399e-01 -7.47540832e-01 -8.63136411e-01 3.40844035e-01
-1.33207095e+00 1.34599864e-01 8.65729749e-01 1.01131165e+00
2.85392523e-01 1.64159015e-01 5.25179386e-01 -8.99388552e-01
5.80755651e-01 -2.48657808e-01 -6.62322700e-01 2.77410187e-02
-2.37627611e-01 2.43429586e-01 1.03711009e+00 -6.32249177e-01
-1.15724897e+00 -1.69138208e-01 -6.47910535e-02 -2.80862451e-01
-5.71991563e-01 -1.48083195e-01 -4.01141569e-02 -1.23884693e-01
9.49087739e-01 2.04676688e-02 1.59967318e-01 -1.39485881e-01
3.44955146e-01 2.50432611e-01 3.11192125e-01 -3.39754611e-01
9.85111713e-01 4.39440876e-01 2.83581018e-01 -8.19954515e-01
-4.33011502e-01 9.91405547e-02 -4.65573013e-01 -1.09289706e-01
9.48757231e-01 -6.66379154e-01 -4.37079966e-01 5.37711263e-01
-1.32005858e+00 -5.01048803e-01 -6.23802483e-01 2.78087795e-01
-6.18062794e-01 -2.34015360e-02 -5.10096908e-01 -6.73074424e-01
-1.70679793e-01 -1.39282441e+00 1.19885707e+00 5.73398590e-01
-1.50115937e-01 -1.02009475e+00 -1.66418016e-01 5.19359559e-02
9.97785270e-01 8.98854792e-01 8.51025641e-01 5.28002456e-02
-7.89832473e-01 -2.56969422e-01 -3.41288447e-01 5.43460310e-01
3.41210365e-01 5.06307483e-02 -1.01446426e+00 -1.53481454e-01
-1.18249789e-01 -2.54544765e-01 8.33137572e-01 5.32493532e-01
1.04545856e+00 -2.77123451e-01 -9.70858485e-02 4.49736208e-01
1.54380274e+00 1.45177200e-01 1.21785843e+00 9.08138609e-05
5.86327314e-01 6.62698925e-01 4.30447370e-01 2.31760129e-01
-2.79962987e-01 8.10251415e-01 6.13567829e-01 -5.89166820e-01
-8.33145499e-01 -4.86928374e-02 2.65860945e-01 -7.15961680e-02
-2.42216513e-01 -5.52501619e-01 -4.63037014e-01 4.82488096e-01
-1.53199089e+00 -1.14462614e+00 9.12485048e-02 2.19086337e+00
7.56666005e-01 2.01353431e-02 -1.07497953e-01 -5.86857758e-02
7.21734881e-01 2.07087293e-01 -3.65024596e-01 -6.55345678e-01
-3.94297987e-01 6.94488227e-01 4.40910816e-01 3.82870018e-01
-7.38748968e-01 9.30194974e-01 6.61766338e+00 1.07286692e+00
-1.45243847e+00 -6.76227957e-02 9.50287998e-01 -1.55737521e-02
-2.56182641e-01 -5.05389161e-02 -5.43953776e-01 8.16981018e-01
6.74794078e-01 2.09272951e-01 4.67929274e-01 6.15411878e-01
2.38341808e-01 -4.41945404e-01 -8.33876491e-01 9.16010916e-01
6.21694401e-02 -1.34219182e+00 2.80580729e-01 3.82553577e-01
1.23222458e+00 -4.69589859e-01 3.09180290e-01 1.12680823e-03
2.26876915e-01 -1.47008657e+00 6.94472909e-01 6.86304450e-01
8.61919403e-01 -7.23478138e-01 7.82298326e-01 3.02645475e-01
-6.21613324e-01 2.67859638e-01 -5.09094834e-01 9.22857970e-02
8.17151293e-02 8.02080750e-01 -6.81508005e-01 3.06398183e-01
5.46363115e-01 2.32958645e-02 -7.99095273e-01 8.98910522e-01
-2.85085171e-01 3.04821938e-01 -2.85365134e-01 -5.48780784e-02
2.82709658e-01 -2.77834177e-01 5.40363789e-01 1.09825563e+00
6.24019444e-01 -3.68544489e-01 -1.18845187e-01 1.20692539e+00
-6.77085519e-02 -1.42239615e-01 -8.53360355e-01 2.62405962e-01
5.30436523e-02 1.34322476e+00 -7.81300247e-01 -2.02481642e-01
-6.76657422e-04 1.26455855e+00 -2.52756197e-02 5.13113618e-01
-8.56527746e-01 -4.89982754e-01 6.02406621e-01 3.82422835e-01
6.38965189e-01 -2.00982526e-01 -2.37961218e-01 -7.74665713e-01
-1.36532560e-01 -8.68534863e-01 -2.57824689e-01 -9.31447744e-01
-1.04066062e+00 9.18497503e-01 -2.87215784e-02 -1.19381809e+00
-6.02350116e-01 -4.54674095e-01 -8.13501060e-01 1.17504215e+00
-1.22426736e+00 -1.28611577e+00 -5.47838211e-01 3.59762162e-01
5.15181482e-01 -5.36304675e-02 8.08026493e-01 7.45950863e-02
-1.87270660e-02 5.78834891e-01 -3.34506556e-02 -1.24811858e-01
8.02712321e-01 -1.30823410e+00 4.67679381e-01 7.78333247e-01
7.33848512e-02 3.27962905e-01 9.16098356e-01 -4.58611727e-01
-8.29530537e-01 -9.91159499e-01 2.71337211e-01 -3.78063023e-01
1.30986601e-01 -4.62376773e-01 -7.60929585e-01 1.12010129e-02
6.81203187e-01 1.18868113e-01 1.85453132e-01 -4.14621711e-01
2.26064324e-02 -1.61368698e-01 -1.49356151e+00 7.15522945e-01
7.98664570e-01 -9.89797488e-02 -6.97124749e-02 -4.59914170e-02
6.25181615e-01 -3.96274447e-01 -6.81358933e-01 3.27872068e-01
4.78200167e-01 -1.28941143e+00 1.00045002e+00 -1.36163056e-01
9.22131419e-01 -4.40858513e-01 2.00414777e-01 -1.79504812e+00
-3.89476568e-01 -4.43433851e-01 2.28956744e-01 1.37057090e+00
5.08448005e-01 -5.93680620e-01 9.34182227e-01 1.74413428e-01
4.57260847e-01 -5.05073190e-01 -6.77745163e-01 -8.05792093e-01
1.23907879e-01 7.26202279e-02 7.23615408e-01 6.50397480e-01
-8.09325516e-01 2.26353914e-01 -5.55649102e-01 -1.03458636e-01
5.02723813e-01 2.06129491e-01 9.58189189e-01 -1.00792265e+00
-3.21125478e-01 -4.79544967e-01 -6.84379160e-01 -6.17336750e-01
-1.11506984e-01 -3.57252091e-01 1.49846077e-01 -1.55428457e+00
-1.06996141e-01 -3.64114106e-01 1.26450092e-01 1.08516023e-01
-4.24178779e-01 8.13210130e-01 3.82534772e-01 -1.12796463e-01
-5.75376190e-02 4.67427194e-01 1.82523143e+00 -1.53200909e-01
-1.30732417e-01 -4.38025743e-02 -5.93817294e-01 5.14661610e-01
8.40500057e-01 -3.24906200e-01 -3.63845289e-01 -2.71539956e-01
2.05309894e-02 -1.50440171e-01 4.74762201e-01 -1.43077266e+00
-1.50800735e-01 -2.00464010e-01 9.18218434e-01 -2.47223914e-01
5.40898621e-01 -7.86811352e-01 6.07011020e-01 4.57575291e-01
-7.73921683e-02 -1.73489079e-01 2.77421832e-01 3.22809368e-01
-4.34148580e-01 -3.49968851e-01 1.28899860e+00 -3.52020949e-01
-4.56240088e-01 1.17857516e-01 -1.29329219e-01 9.59653873e-04
1.06047642e+00 -3.84577155e-01 -3.28760445e-01 -7.98989832e-01
-4.77250785e-01 -2.67448485e-01 8.81678045e-01 3.22085500e-01
5.01123667e-01 -1.35879743e+00 -8.59715521e-01 2.40782112e-01
-9.99850109e-02 1.23631150e-01 3.01642209e-01 3.76494557e-01
-8.56417656e-01 5.88570498e-02 -5.82223356e-01 -5.75086594e-01
-9.36725914e-01 5.89626431e-01 3.22629571e-01 -2.98700333e-01
-3.18768322e-01 6.33145988e-01 6.73669755e-01 4.16394882e-02
-1.66376963e-01 -2.37675115e-01 -1.13927945e-01 -7.61385933e-02
4.35848355e-01 1.17110968e-01 -1.01752140e-01 -5.15991986e-01
-6.59095496e-02 7.38021970e-01 2.34422162e-01 -1.34312496e-01
1.17513669e+00 1.21935695e-01 -2.62837075e-02 -4.71833050e-02
1.02634132e+00 3.34570795e-01 -1.62236667e+00 3.41912806e-01
-5.80662310e-01 -5.53868115e-01 -7.44254813e-02 -8.18833351e-01
-1.32960272e+00 8.86885345e-01 7.79192448e-01 3.41803432e-01
1.19478869e+00 -1.26453206e-01 6.34554207e-01 -1.30612195e-01
2.57631302e-01 -7.32105136e-01 2.47224525e-01 3.09969872e-01
1.28769720e+00 -1.20463586e+00 -3.81497247e-03 -2.40593106e-01
-6.17602110e-01 1.00576854e+00 5.63040137e-01 -5.17508447e-01
3.80494028e-01 4.58642393e-01 1.70565113e-01 1.11034565e-01
-4.44116741e-01 -3.86384398e-01 3.98302972e-01 9.02811110e-01
4.70761716e-01 -5.52974902e-02 -2.19250873e-01 -8.75246525e-02
-2.95794338e-01 -1.93494275e-01 5.70272565e-01 4.68155086e-01
-1.46347016e-01 -1.27999938e+00 -5.67726672e-01 3.17064345e-01
-6.81788385e-01 -1.78382754e-01 -3.67932081e-01 7.71842301e-01
5.04643023e-01 8.31467807e-01 2.35612523e-02 8.64849892e-04
2.28941292e-01 -2.77960390e-01 6.67840958e-01 -6.59183145e-01
-6.07951343e-01 1.00516481e-02 -2.49581173e-01 -4.53158021e-01
-3.20713848e-01 -2.24845290e-01 -7.75431991e-01 -1.92211270e-01
-3.06513071e-01 -2.48528067e-02 8.81845653e-01 5.40305734e-01
3.42809111e-01 8.12188983e-01 6.46414936e-01 -1.14980173e+00
-8.85058120e-02 -1.18844342e+00 -4.18988436e-01 7.26991415e-01
2.59959012e-01 -5.21257997e-01 -4.67511505e-01 1.88606411e-01] | [11.697854042053223, -0.47702357172966003] |
745f536f-9804-4e1d-afb5-5b7be270ba92 | fasterx-real-time-object-detection-based-on | 2209.03157 | null | https://arxiv.org/abs/2209.03157v1 | https://arxiv.org/pdf/2209.03157v1.pdf | FasterX: Real-Time Object Detection Based on Edge GPUs for UAV Applications | Real-time object detection on Unmanned Aerial Vehicles (UAVs) is a challenging issue due to the limited computing resources of edge GPU devices as Internet of Things (IoT) nodes. To solve this problem, in this paper, we propose a novel lightweight deep learning architectures named FasterX based on YOLOX model for real-time object detection on edge GPU. First, we design an effective and lightweight PixSF head to replace the original head of YOLOX to better detect small objects, which can be further embedded in the depthwise separable convolution (DS Conv) to achieve a lighter head. Then, a slimmer structure in the Neck layer termed as SlimFPN is developed to reduce parameters of the network, which is a trade-off between accuracy and speed. Furthermore, we embed attention module in the Head layer to improve the feature extraction effect of the prediction head. Meanwhile, we also improve the label assignment strategy and loss function to alleviate category imbalance and box optimization problems of the UAV dataset. Finally, auxiliary heads are presented for online distillation to improve the ability of position embedding and feature extraction in PixSF head. The performance of our lightweight models are validated experimentally on the NVIDIA Jetson NX and Jetson Nano GPU embedded platforms.Extensive experiments show that FasterX models achieve better trade-off between accuracy and latency on VisDrone2021 dataset compared to state-of-the-art models. | ['JunYi', 'Huan Luo', 'Yiwen Long', 'Rui Hu', 'Xuanlin Min', 'Wei Zhou'] | 2022-09-07 | null | null | null | null | ['real-time-object-detection'] | ['computer-vision'] | [-3.62848252e-01 -4.28441226e-01 1.03982575e-01 -2.56205529e-01
2.83505261e-01 -3.19270998e-01 -4.32796823e-03 -2.30483264e-01
-6.36053205e-01 3.63140851e-01 -3.70010346e-01 -3.38807046e-01
8.18803236e-02 -9.89721835e-01 -6.78924799e-01 -6.52962744e-01
-1.72382742e-02 -1.91480592e-01 6.12778068e-01 1.36092203e-02
-2.31777653e-01 4.86824691e-01 -1.54981983e+00 9.27386340e-03
7.58148193e-01 1.65796959e+00 4.65289026e-01 4.26001072e-01
3.21653932e-01 5.08488536e-01 -3.00153881e-01 -8.95318910e-02
6.19569302e-01 2.21000239e-01 -1.48840636e-01 -7.36908764e-02
2.86843002e-01 -8.18107188e-01 -4.81959373e-01 1.22820413e+00
7.36726463e-01 2.62093470e-02 3.74870837e-01 -1.39767003e+00
-4.21335518e-01 2.21557066e-01 -7.83693016e-01 3.90027255e-01
-4.50511426e-01 4.70601052e-01 6.57311738e-01 -7.90349245e-01
1.44897535e-01 1.06193459e+00 8.34775507e-01 2.32532889e-01
-3.87494147e-01 -1.04007006e+00 3.58643055e-01 4.13252532e-01
-1.31896067e+00 -1.63460150e-01 4.60020632e-01 -2.58056939e-01
1.00450528e+00 1.76013380e-01 8.93954098e-01 6.05937064e-01
4.94155288e-01 7.11203396e-01 4.54479992e-01 1.60380468e-01
1.36463031e-01 2.69938968e-02 1.19015142e-01 1.31734645e+00
5.25670946e-01 2.07631558e-01 -6.87948093e-02 2.10131899e-01
8.28198969e-01 4.26958680e-01 -3.21604282e-01 -8.62452760e-02
-1.00333822e+00 7.04665422e-01 1.22544301e+00 -2.03484863e-01
-5.44474363e-01 4.18691635e-01 7.05758989e-01 -9.54779610e-02
4.41152871e-01 1.67126268e-01 -8.63768280e-01 2.12848514e-01
-5.67476630e-01 -1.38185117e-02 3.86651546e-01 1.23080623e+00
6.48896515e-01 2.05097079e-01 -2.79344618e-01 2.96244591e-01
5.93201637e-01 5.02557814e-01 5.16200185e-01 -4.67937499e-01
5.58188915e-01 1.12591469e+00 -7.47505724e-02 -9.96506095e-01
-6.71393752e-01 -9.07238483e-01 -9.22338188e-01 3.99720132e-01
-2.75045872e-01 -5.15423596e-01 -9.73008871e-01 1.25609016e+00
6.73611939e-01 3.18480700e-01 -9.92254689e-02 1.35998666e+00
1.08960402e+00 8.75543952e-01 1.54763639e-01 2.57832497e-01
1.71118653e+00 -1.46664155e+00 -2.78374076e-01 -3.77173811e-01
1.01361680e+00 -4.65278536e-01 9.99059141e-01 4.92538773e-02
-6.36821032e-01 -7.60673583e-01 -1.34452009e+00 -3.64357233e-01
-4.50574219e-01 9.16288137e-01 9.53171372e-01 4.74115252e-01
-5.45869827e-01 4.78541523e-01 -1.00623667e+00 -4.48413342e-02
8.33868682e-01 6.75209045e-01 8.18476528e-02 8.56183395e-02
-9.44177270e-01 2.20597684e-01 5.69246531e-01 3.77595425e-01
-7.51705706e-01 -8.95183086e-01 -6.89916670e-01 3.35071385e-01
3.46538395e-01 -9.24583495e-01 9.65224326e-01 -8.49576652e-01
-1.45350683e+00 4.38686937e-01 4.46110427e-01 -4.35603857e-01
3.79401535e-01 -1.09481454e-01 -2.65297472e-01 -6.10972829e-02
-9.96150300e-02 1.02354646e+00 8.71033430e-01 -4.72025484e-01
-1.14888000e+00 -5.15735805e-01 2.89040208e-01 2.57916003e-01
-6.84112191e-01 -1.63594335e-01 -4.64162648e-01 -5.52467346e-01
-1.79578960e-01 -1.10151494e+00 -2.67413348e-01 4.81404006e-01
-2.47148097e-01 -4.11220610e-01 1.38249779e+00 -4.48402107e-01
1.16774392e+00 -2.39650130e+00 -7.13607073e-02 -1.70726925e-01
4.50426430e-01 6.71537220e-01 -1.51110068e-01 -3.84108067e-01
2.95145288e-02 -2.80378252e-01 2.03745723e-01 -3.20395857e-01
-8.03106055e-02 -2.64932830e-02 -1.84019089e-01 5.22238195e-01
1.47138759e-01 9.29045022e-01 -6.64337814e-01 -3.41287285e-01
2.70405829e-01 5.45229614e-01 -7.78857350e-01 3.75030339e-02
-7.69489855e-02 4.81122583e-02 -7.52500176e-01 7.73959816e-01
9.90105510e-01 -3.20946962e-01 -5.19285381e-01 -7.03896582e-01
-4.95670915e-01 5.79642728e-02 -9.76942122e-01 1.36885750e+00
-5.49519420e-01 4.11863416e-01 2.76591122e-01 -5.11254728e-01
7.18724847e-01 -1.04240417e-01 9.43892971e-02 -4.43351716e-01
7.91770935e-01 1.06535759e-02 1.44529045e-01 -4.34853315e-01
4.48171377e-01 2.96540856e-01 1.59545496e-01 -1.57428429e-01
-1.10309944e-01 2.67146945e-01 -2.59765629e-02 -7.23630488e-02
9.39583361e-01 -7.82762170e-02 -1.39579147e-01 -5.20363748e-01
4.40101773e-01 8.56172964e-02 7.27196932e-01 1.72180280e-01
-2.24901363e-01 1.55228063e-01 2.21757680e-01 -8.62838268e-01
-6.53194189e-01 -5.07099032e-01 -2.05162674e-01 9.28709030e-01
5.54521859e-01 -5.05617023e-01 -7.90528834e-01 -7.90735424e-01
-5.75813055e-02 2.14725435e-01 -4.46329892e-01 -3.09729666e-01
-3.85939926e-01 -1.11565578e+00 2.24355876e-01 1.01363027e+00
1.23386919e+00 -8.48579764e-01 -1.32001472e+00 1.21430881e-01
4.94479150e-01 -1.24038160e+00 -4.37606275e-01 2.13662490e-01
-8.13802660e-01 -9.82244194e-01 -3.68824095e-01 -8.48563910e-01
8.03729534e-01 5.67179382e-01 5.55399477e-01 2.30605677e-01
-5.28828621e-01 -8.18855464e-02 -3.58288735e-01 -7.82789469e-01
5.47761619e-01 3.77578229e-01 1.80617198e-01 -1.38540387e-01
2.75610268e-01 -2.59940803e-01 -1.21035326e+00 3.14199954e-01
-7.96710253e-01 1.94765255e-01 6.80252075e-01 8.03670108e-01
5.81327617e-01 2.71448642e-01 1.11991219e-01 -3.81129891e-01
2.33068988e-01 -3.67688209e-01 -1.17384088e+00 5.19309472e-03
-5.05689561e-01 -3.03043216e-01 1.03972518e+00 -4.88941997e-01
-6.49399519e-01 3.11743379e-01 -1.21555343e-01 -7.68189251e-01
3.34045142e-01 2.40405813e-01 -2.86022097e-01 -5.61502576e-01
1.35960117e-01 -1.53246909e-01 -2.34077588e-01 -1.83247611e-01
-9.72910002e-02 8.02512109e-01 1.83102846e-01 2.90937889e-02
4.29939240e-01 4.40583527e-01 1.99383438e-01 -7.46005356e-01
-6.95771933e-01 -1.18692696e-01 -1.19844750e-01 -1.63956806e-01
9.28352177e-01 -1.33284533e+00 -1.10892856e+00 5.47573984e-01
-1.18381941e+00 -3.43904883e-01 -1.01663820e-01 6.55295789e-01
6.06429875e-02 -6.48696721e-02 -6.62964344e-01 -4.45014566e-01
-9.49683547e-01 -1.42911601e+00 1.37113988e+00 8.49711120e-01
4.03176218e-01 -6.41322613e-01 -6.00430250e-01 -6.38631806e-02
3.65233243e-01 -4.73284088e-02 5.62666059e-01 -1.27578020e-01
-8.86343420e-01 -3.50065410e-01 -8.89976144e-01 4.34675097e-01
-1.08995579e-01 -1.34161040e-01 -6.93874896e-01 -6.74837589e-01
5.69931976e-02 -1.34268077e-03 1.06355774e+00 3.67771894e-01
1.43408942e+00 -2.22626105e-01 -6.78067327e-01 1.30001760e+00
1.46457171e+00 2.20889613e-01 4.22147870e-01 3.36494446e-01
1.02790260e+00 1.24716811e-01 7.64139473e-01 6.04810238e-01
4.53671247e-01 4.94442731e-01 9.49515641e-01 -3.04966956e-01
-1.66095510e-01 1.74871292e-02 2.52341241e-01 6.80718482e-01
9.39368308e-02 -4.19861704e-01 -5.94030738e-01 2.24331573e-01
-1.72541547e+00 -2.24545464e-01 -2.79069185e-01 1.79427063e+00
1.03469547e-02 1.35819569e-01 -1.80409476e-01 -1.59746662e-01
5.73811829e-01 1.48590118e-01 -8.86862874e-01 -2.04987749e-02
1.39783025e-01 -1.40214831e-01 8.99544954e-01 -2.13027149e-02
-1.44584882e+00 1.01909614e+00 4.28074121e+00 9.56720948e-01
-1.54309201e+00 2.01705277e-01 6.10799253e-01 -2.39137024e-01
3.60040367e-01 -2.27628410e-01 -1.28881884e+00 8.16224575e-01
4.08843189e-01 1.98755220e-01 2.86225110e-01 1.32493389e+00
-2.43180841e-02 1.04408510e-01 -7.24693298e-01 1.20703864e+00
-1.75807729e-01 -1.18521142e+00 -1.23839639e-01 6.34494126e-02
5.03448784e-01 4.61088449e-01 1.19350538e-01 3.21249455e-01
-5.17061260e-03 -5.40728211e-01 6.47056460e-01 7.52079114e-02
8.46778095e-01 -9.68836963e-01 9.67858315e-01 2.66954482e-01
-1.56783521e+00 -5.45916140e-01 -9.33757722e-01 -2.27152944e-01
2.07831524e-02 5.12247026e-01 -4.77523685e-01 3.89646471e-01
1.10353196e+00 7.74012864e-01 -6.20901406e-01 1.15365911e+00
-1.29236430e-01 2.40797862e-01 -6.03279650e-01 -3.43459338e-01
6.29156172e-01 -2.10720822e-01 4.68133271e-01 9.17061329e-01
5.83466291e-01 4.15814668e-01 3.68305445e-01 5.96472085e-01
-3.46810848e-01 -5.50938845e-02 -3.83972049e-01 1.97400376e-01
3.30260873e-01 1.83149743e+00 -8.78115475e-01 -3.51001889e-01
-4.87082511e-01 9.70058322e-01 1.85821638e-01 -6.42018467e-02
-1.18288505e+00 -7.72207141e-01 8.87838781e-01 1.15376845e-01
5.79527736e-01 -2.18931749e-01 -5.44280857e-02 -1.19382453e+00
5.09641245e-02 -4.19113040e-01 1.46899015e-01 -7.86284447e-01
-9.45185542e-01 9.50251758e-01 -5.01047134e-01 -1.44438064e+00
6.60095632e-01 -1.10683215e+00 -7.48031139e-01 5.77260971e-01
-1.61599231e+00 -1.32286859e+00 -9.10689175e-01 3.44248921e-01
5.89763224e-01 -1.36401266e-01 4.63985741e-01 6.91979051e-01
-1.22248232e+00 5.53817987e-01 -2.48567879e-01 6.25365898e-02
3.22502762e-01 -5.56711316e-01 4.74725544e-01 9.29993391e-01
-4.41513419e-01 2.89395332e-01 1.88335866e-01 -7.91356564e-01
-1.60387361e+00 -1.77656448e+00 -8.74195364e-04 1.08276926e-01
5.15287995e-01 -2.53656715e-01 -5.86742103e-01 5.88606477e-01
-4.47665043e-02 5.93616188e-01 3.47272664e-01 -4.57690835e-01
9.45782065e-02 -4.19071943e-01 -1.01912892e+00 5.23144484e-01
1.19156051e+00 8.51255432e-02 -1.38049135e-02 5.33998609e-01
9.97770190e-01 -6.84319079e-01 -5.19387424e-01 7.92234838e-01
4.96486753e-01 -6.84903324e-01 8.47970068e-01 -4.83178437e-01
2.45112777e-01 -6.21943235e-01 -1.73066985e-02 -1.12613475e+00
-5.35876989e-01 -3.12680602e-01 -9.03237760e-02 9.66389835e-01
-9.90123581e-03 -7.70692706e-01 9.33738351e-01 2.65133321e-01
-6.30803883e-01 -1.16469204e+00 -8.05374563e-01 -7.65766442e-01
-5.42688668e-01 -7.52992183e-02 8.41644168e-01 5.03336728e-01
-5.68625271e-01 5.21669269e-01 -9.48904082e-02 5.76774180e-01
3.49375695e-01 7.73607641e-02 6.85550511e-01 -1.22086334e+00
1.38942776e-02 -8.83690342e-02 -6.62602663e-01 -1.36708128e+00
-1.24190144e-01 -6.91131413e-01 -1.33502707e-01 -1.29583001e+00
-1.70845032e-01 -7.31159508e-01 -1.11218043e-01 4.81970608e-01
-2.47564971e-01 4.13712412e-01 2.65416354e-01 -9.27317142e-02
-8.54541540e-01 9.17847097e-01 1.20738685e+00 -1.45402625e-01
-1.70520842e-01 -2.30362061e-02 -3.53185564e-01 1.00434113e+00
6.80881381e-01 -4.86322343e-01 -4.05034184e-01 -1.04045868e+00
2.33500764e-01 -3.36996138e-01 6.34297788e-01 -1.48129880e+00
5.67955613e-01 4.18602705e-01 7.75148809e-01 -6.46238863e-01
3.02524209e-01 -1.08074999e+00 -8.62335265e-02 8.27065945e-01
5.02821863e-01 4.20652539e-01 5.60161293e-01 4.63490725e-01
6.58562854e-02 -1.32753983e-01 7.52927959e-01 1.65776372e-01
-8.50815296e-01 8.70102525e-01 4.02552038e-02 -4.21247900e-01
1.42391682e+00 7.02555524e-03 -5.28096616e-01 3.71441841e-01
-4.32230445e-04 7.36184478e-01 4.33855504e-01 4.08780366e-01
6.41359329e-01 -1.30623043e+00 -3.31410587e-01 5.31219065e-01
-1.15931466e-01 3.33434105e-01 5.79778254e-01 9.19210494e-01
-1.06530464e+00 4.22114611e-01 -3.97022545e-01 -5.10417223e-01
-1.28496480e+00 6.95876718e-01 2.67002285e-01 -8.63593891e-02
-7.44053125e-01 1.21412516e+00 5.54563105e-01 -2.09180832e-01
2.05762625e-01 -6.00949407e-01 -8.78429338e-02 -3.88427563e-02
7.23956048e-01 5.72617054e-01 1.64564282e-01 -3.46633703e-01
-6.26790464e-01 6.90171301e-01 -5.74539937e-02 7.16225028e-01
1.24137044e+00 -3.89783718e-02 -1.79365650e-01 -3.99784237e-01
1.19560158e+00 -2.61958897e-01 -1.54118609e+00 2.68606335e-01
-7.18128085e-01 -2.60819614e-01 7.35472977e-01 -3.95888835e-01
-1.49992454e+00 9.69082594e-01 1.16269636e+00 -1.59075875e-02
1.31338322e+00 -4.19172555e-01 1.28846943e+00 3.99520755e-01
2.82966673e-01 -7.89802551e-01 -9.90386754e-02 5.32389343e-01
5.16551673e-01 -1.26245475e+00 1.20932586e-01 -6.62920296e-01
-2.83658296e-01 1.15823114e+00 1.15479064e+00 -4.10051078e-01
8.16116512e-01 4.56596345e-01 -2.33062580e-01 -1.99940115e-01
-4.77043331e-01 -1.77027479e-01 9.16847661e-02 1.83529556e-01
-1.78012222e-01 3.08735687e-02 -1.65240571e-01 9.95505512e-01
-1.19735278e-01 1.39203310e-01 1.37996465e-01 8.36655974e-01
-3.69005382e-01 -4.63833809e-01 -1.85659959e-03 6.14145219e-01
-3.40815008e-01 -2.71628648e-01 1.73809081e-01 5.86842060e-01
6.47617877e-01 5.78547239e-01 3.36097538e-01 -8.04085672e-01
3.72290403e-01 -5.60932457e-01 1.92140922e-01 -5.25579751e-01
-7.21718311e-01 -1.39989600e-01 -1.76448941e-01 -5.99107623e-01
4.20108624e-02 -8.77541825e-02 -1.24867547e+00 -1.61040992e-01
-7.70453334e-01 -1.12184055e-01 8.20752382e-01 6.30460680e-01
8.90494406e-01 1.05394375e+00 5.25515914e-01 -1.10782957e+00
-3.48857164e-01 -6.65997982e-01 -4.60444331e-01 -4.27179039e-01
1.30401880e-01 -7.50933468e-01 -2.75894910e-01 -2.96953380e-01] | [8.816258430480957, -0.3906027674674988] |
4345cf6e-2400-4923-81ac-6ecc60c4535d | some-strategies-to-capture-karaka-yogyata | 2201.01700 | null | https://arxiv.org/abs/2201.01700v1 | https://arxiv.org/pdf/2201.01700v1.pdf | Some Strategies to Capture Karaka-Yogyata with Special Reference to apadana | In today's digital world language technology has gained importance. Several softwares, have been developed and are available in the field of computational linguistics. Such tools play a crucial role in making classical language texts easily accessible. Some Indian philosophical schools have contributed towards various techniques of verbal cognition to analyze sentences correctly. These theories can be used to build computational tools for word sense disambiguation (WSD). In the absence of WSD, one cannot have proper verbal cognition. These theories considered the concept of 'Yogyat\=a' (congruity or compatibility) as the indispensable cause of verbal cognition. In this work, we come up with some insights on the basis of these theories to create a tool that will capture Yogyat\=a of words. We describe the problem of ambiguity in a text and present a method to resolve it computationally with the help of Yogyat\=a. Here, only two major schools i.e. Ny\=aya and Vy\=akarana are considered. Our paper attempts to show the implication of the creation of our tool in this area. Also, our tool involves the creation of an 'ontological tag-set' as well as strategies to mark up the lexicon. The introductory description of ablation is also covered in this paper. Such strategies and some case studies shall form the core of our paper. | ['Malhar Kulkarni', 'Diptesh Kanojia', 'Swaraja Salaskar'] | 2022-01-05 | null | null | null | null | ['word-sense-disambiguation'] | ['natural-language-processing'] | [-7.20463321e-02 3.14089417e-01 1.88971728e-01 -2.12219834e-01
-6.10992722e-02 -6.25873208e-01 7.13855624e-01 5.10513425e-01
-5.04785240e-01 7.15303600e-01 2.05920741e-01 -6.64475918e-01
-6.34128928e-01 -8.27792168e-01 -5.37572801e-03 -3.84895027e-01
3.36115211e-01 5.09186149e-01 3.92585009e-01 -8.26255679e-01
7.73422658e-01 1.01134986e-01 -1.92523015e+00 2.25025765e-03
9.13491428e-01 5.05661368e-01 8.01642716e-01 1.65388599e-01
-9.57849026e-01 6.52255297e-01 -8.12036932e-01 -3.73006344e-01
-1.60607293e-01 -4.75149721e-01 -1.40219712e+00 -7.93181583e-02
-1.87445730e-01 1.54076844e-01 3.39181662e-01 1.29644978e+00
4.29118872e-01 9.62679684e-02 5.99176943e-01 -9.73868430e-01
-7.02063203e-01 1.01146197e+00 -2.19172798e-02 4.12865520e-01
8.40685666e-01 -5.18925011e-01 6.77993357e-01 -5.65980911e-01
6.98413610e-01 1.43991172e+00 4.69008684e-01 3.03124756e-01
-6.02178395e-01 -5.29630005e-01 -6.81027919e-02 5.80367446e-01
-1.38317692e+00 -1.58998698e-01 7.55157828e-01 -3.17305923e-01
1.06806743e+00 4.08507258e-01 9.42557931e-01 5.76615512e-01
2.08329663e-01 4.48538959e-01 1.41438830e+00 -1.30580676e+00
1.55486509e-01 3.68880242e-01 7.83758700e-01 4.63565648e-01
6.05130136e-01 -4.63769287e-01 -4.36885983e-01 1.06196560e-01
4.53981698e-01 -3.89597893e-01 -3.70734334e-02 3.07344615e-01
-6.49795592e-01 8.79096210e-01 -1.54477462e-01 1.35563862e+00
-3.60062718e-02 -3.97819281e-01 3.07256967e-01 1.96693137e-01
1.92374960e-01 4.66665089e-01 -2.41872489e-01 -1.68307543e-01
-4.66850400e-01 4.15654659e-01 8.57574582e-01 8.20642114e-01
3.25601637e-01 -9.04066768e-03 5.39954722e-01 7.74004042e-01
7.08210349e-01 4.54637915e-01 9.93772089e-01 -6.67921841e-01
4.60064225e-02 9.77405071e-01 -9.47207958e-02 -1.18424678e+00
-4.63654339e-01 5.52863814e-03 -1.23248249e-01 5.49070947e-02
3.16607565e-01 1.51445583e-01 -6.82553947e-01 1.43075001e+00
4.20049191e-01 -1.99479029e-01 4.90851372e-01 6.50106788e-01
1.34446025e+00 4.94496673e-01 3.10514301e-01 -3.87859851e-01
1.93756282e+00 -3.12141299e-01 -1.39454138e+00 -1.18497819e-01
9.08653498e-01 -1.11160815e+00 1.05853808e+00 2.72665977e-01
-1.05321634e+00 -2.07189724e-01 -1.02538073e+00 -1.59509867e-01
-1.12531757e+00 -1.09115265e-01 8.58394921e-01 1.00189626e+00
-1.09512436e+00 3.57823372e-01 -3.80886406e-01 -7.83052564e-01
-2.65413374e-01 2.03763008e-01 -3.28876898e-02 2.18660235e-01
-1.56101966e+00 1.36446679e+00 8.79976749e-01 1.99764714e-01
1.65785596e-01 6.00843169e-02 -8.71502101e-01 -2.90881991e-01
7.44748235e-01 -3.78843099e-01 1.05177844e+00 -1.13098633e+00
-1.15326631e+00 1.39520872e+00 2.82204314e-03 -2.84686476e-01
1.81047797e-01 -1.53070703e-01 -7.89711773e-01 9.58837569e-03
6.27035558e-01 -6.34116456e-02 1.07706882e-01 -1.14142537e+00
-6.90173328e-01 -4.21026200e-01 3.12605172e-01 2.53336549e-01
-1.67059734e-01 6.74681783e-01 -1.97241515e-01 -6.64116263e-01
4.81224328e-01 -6.73150122e-01 1.69715300e-01 -7.90316164e-01
1.20287929e-02 -6.70283914e-01 6.38026357e-01 -6.61409676e-01
1.57498455e+00 -1.86846995e+00 -1.93070129e-01 3.95923495e-01
2.93682009e-01 5.97641230e-01 5.20670295e-01 7.01263785e-01
7.52575696e-03 3.82938981e-01 -9.22088772e-02 2.32076317e-01
2.65626222e-01 6.25289559e-01 -1.69389024e-01 -4.29188199e-02
-3.79231304e-01 4.35553461e-01 -9.34834659e-01 -8.11942041e-01
4.03544277e-01 3.12319726e-01 -5.51580102e-04 -2.41701871e-01
-7.51661658e-02 -1.56229794e-01 -7.56293714e-01 5.90352714e-01
3.78740132e-01 1.82623461e-01 6.95727170e-01 2.00134099e-01
-7.10155427e-01 5.04070103e-01 -1.43359268e+00 1.38103008e+00
-1.27622381e-01 4.49994117e-01 -7.68902451e-02 -9.76654410e-01
9.79382813e-01 6.56711638e-01 1.56556055e-01 -8.35670352e-01
7.15527654e-01 6.66155934e-01 2.19260335e-01 -8.27990413e-01
7.12842762e-01 -2.36401245e-01 -1.25690714e-01 3.86463881e-01
-9.14793164e-02 -3.54589731e-01 4.63330209e-01 1.64734438e-01
4.16923970e-01 2.32861891e-01 9.78572249e-01 -7.34361529e-01
8.88295472e-01 1.52364671e-01 4.03081924e-01 3.92292410e-01
-2.26900503e-02 -1.32577941e-01 2.30887592e-01 -5.07165134e-01
-7.96647251e-01 -5.91420591e-01 -3.79329950e-01 9.72404361e-01
3.48741114e-01 -7.39046216e-01 -9.95473087e-01 -5.71632162e-02
-4.91630882e-01 1.00708938e+00 -1.92581341e-01 2.83004194e-01
-4.43019688e-01 -8.24464798e-01 5.27096927e-01 5.90573400e-02
6.32121980e-01 -1.09250116e+00 -1.06256282e+00 3.50568354e-01
-1.87744111e-01 -9.19279456e-01 5.15204847e-01 1.04067318e-01
-5.78509688e-01 -1.15375137e+00 1.77460089e-02 -1.02140427e+00
2.42926836e-01 3.43622506e-01 9.90359128e-01 8.28239858e-01
-1.29217818e-01 3.39189321e-01 -9.85582530e-01 -9.64943349e-01
-4.28201854e-01 -1.75641626e-01 7.99455494e-02 -7.70388484e-01
9.06511962e-01 -5.49046397e-01 1.16861977e-01 -9.62907970e-02
-1.17878127e+00 -6.02857880e-02 3.59179564e-02 2.30723456e-01
2.00786844e-01 4.15313900e-01 5.96665800e-01 -9.46516395e-01
1.06342936e+00 -3.93794566e-01 -3.49991024e-01 3.20959836e-01
-7.33696163e-01 7.50611201e-02 2.72346228e-01 2.81529073e-02
-1.14464641e+00 -5.62782288e-01 -3.37450713e-01 4.01449054e-01
-2.68203944e-01 7.27416039e-01 -3.51769149e-01 -3.44891250e-02
5.90308368e-01 9.10160765e-02 -8.87687653e-02 -4.97829765e-01
2.08640963e-01 8.64090383e-01 2.94065773e-01 -7.84021378e-01
3.66106778e-01 5.84321767e-02 -1.22989930e-01 -9.62752879e-01
-5.95223546e-01 -4.73463535e-01 -7.05504537e-01 -5.01353383e-01
1.01036251e+00 -5.29791832e-01 -5.66606820e-01 1.16152309e-01
-1.10299480e+00 2.71661401e-01 -1.00546330e-02 5.16979158e-01
-3.07996869e-01 5.22229970e-01 -1.94753870e-01 -1.10383189e+00
-3.34345818e-01 -8.75553429e-01 4.16798592e-01 5.77232897e-01
-5.99657357e-01 -1.33513546e+00 -1.96740642e-01 3.86997044e-01
1.11271322e-01 3.74719650e-02 1.06824124e+00 -1.06671882e+00
-2.16461606e-02 2.14793727e-01 1.84064344e-01 -3.01561877e-02
2.04708919e-01 -8.92476272e-03 -6.07130229e-01 4.35205638e-01
4.32108104e-01 -3.41406390e-02 1.62185460e-01 1.12972684e-01
4.43874866e-01 -1.60801604e-01 -3.22445691e-01 -2.45508999e-02
1.61877394e+00 8.44219208e-01 6.69067264e-01 8.71273398e-01
1.89761356e-01 7.33117163e-01 8.93057764e-01 2.09454387e-01
6.58510089e-01 6.64535463e-01 2.52512321e-02 4.85995203e-01
-1.92869559e-01 1.25760227e-01 5.52924387e-02 1.42719710e+00
-4.50222999e-01 -1.16759494e-01 -1.38544095e+00 4.71287161e-01
-1.76409650e+00 -7.65253067e-01 -5.38176715e-01 1.66814303e+00
8.34428549e-01 2.82197922e-01 -1.20413989e-01 6.53057098e-01
6.19736254e-01 -1.66661337e-01 6.13344491e-01 -8.70905519e-01
-2.51517504e-01 4.92434472e-01 1.03946425e-01 9.31206167e-01
-7.55115867e-01 1.37400305e+00 5.63890457e+00 8.34437907e-01
-9.56007123e-01 1.27533004e-01 -1.81098193e-01 5.57690382e-01
-4.80651796e-01 2.01859683e-01 -7.88488746e-01 5.48824549e-01
7.39923000e-01 -3.87546152e-01 2.31133588e-02 6.59028590e-01
2.52677470e-01 -6.74262226e-01 -3.48815024e-01 8.87728155e-01
5.17166615e-01 -1.06198299e+00 8.04138705e-02 -2.45441109e-01
1.33895516e-01 -8.05216610e-01 -4.90227401e-01 1.51062146e-01
3.60910237e-01 -8.52796733e-01 8.06464136e-01 4.11541760e-01
2.23427460e-01 -7.84151912e-01 1.14401674e+00 3.28936756e-01
-1.21439087e+00 2.70985246e-01 -3.49456429e-01 -7.48211384e-01
1.52159870e-01 1.82684109e-01 -7.94547617e-01 9.13727164e-01
6.49178445e-01 1.96348473e-01 -6.52474701e-01 7.27794170e-01
-3.74105304e-01 2.58080095e-01 -3.76990020e-01 -6.27381623e-01
2.55833000e-01 -3.79627615e-01 5.93422532e-01 9.97926354e-01
3.53883028e-01 6.88899398e-01 1.31233349e-01 6.31637335e-01
5.66704810e-01 6.98872447e-01 -4.65700448e-01 1.32887601e-03
7.41521358e-01 8.17471325e-01 -1.34283364e+00 -4.81127977e-01
-2.80449569e-01 4.22350824e-01 -4.00744230e-02 -1.95447594e-01
-5.49144268e-01 -3.88742685e-01 3.74987990e-01 8.74601454e-02
-2.07125947e-01 -2.69870043e-01 -5.63967526e-01 -8.02045703e-01
-1.81665137e-01 -9.13651943e-01 4.17767465e-01 -8.99821877e-01
-9.09526646e-01 7.11476982e-01 5.40525138e-01 -7.98169553e-01
-3.40742737e-01 -9.52937901e-01 -3.25406134e-01 8.95559788e-01
-9.88033354e-01 -1.03223991e+00 -2.18883064e-02 4.61454540e-01
5.48062623e-01 -1.89163819e-01 1.23948622e+00 1.84420437e-01
-3.16412061e-01 -9.55002233e-02 -2.09589481e-01 4.26481031e-02
4.02267456e-01 -1.21286845e+00 -2.82213092e-01 9.84963894e-01
1.60413146e-01 7.71871269e-01 1.17313468e+00 -8.75274539e-01
-8.57025266e-01 -3.86713184e-02 1.73328400e+00 -3.54536235e-01
7.96618760e-01 1.90986171e-01 -8.76723230e-01 5.96091390e-01
5.76969206e-01 -9.04098868e-01 8.99947226e-01 -1.58636440e-02
1.54025399e-03 1.97178975e-01 -1.05631137e+00 7.90996075e-01
8.68679404e-01 -3.77446890e-01 -1.72245800e+00 2.69997001e-01
5.31857669e-01 -2.54011005e-01 -5.74017286e-01 -2.94667017e-02
2.67839342e-01 -1.02735078e+00 8.88746083e-01 -3.28949243e-01
-1.07769035e-01 -3.47168922e-01 1.10823326e-02 -9.61659312e-01
-1.18277958e-02 -3.94757241e-01 5.94915092e-01 1.36178350e+00
1.22473262e-01 -8.79194558e-01 1.89186066e-01 5.94834507e-01
-3.32601398e-01 -1.33893058e-01 -9.83652115e-01 -3.39391023e-01
2.41336077e-02 -9.62039769e-01 6.03480220e-01 1.31548631e+00
5.56475401e-01 4.21589762e-01 1.26197010e-01 -2.61807349e-02
3.26573640e-01 -8.22308511e-02 1.53125376e-01 -1.61118352e+00
2.41380051e-01 -5.02118826e-01 -6.40367329e-01 -3.01287055e-01
2.72189885e-01 -7.88910806e-01 -2.93981642e-01 -1.80380213e+00
-2.20440507e-01 -3.56873065e-01 2.39959821e-01 5.02057254e-01
-2.63444204e-02 7.92071074e-02 4.16737944e-01 2.29569420e-01
-1.71527803e-01 -8.05100799e-02 1.03590405e+00 4.30975914e-01
-2.63199180e-01 -5.94800591e-01 -9.59568620e-01 1.19105875e+00
1.05871248e+00 -4.75340903e-01 -5.77948332e-01 -3.22999328e-01
7.49950886e-01 -3.22302043e-01 1.41321078e-01 -8.61193657e-01
2.67556340e-01 -3.71714979e-01 -8.28835890e-02 -4.38628167e-01
2.04654843e-01 -9.69813049e-01 2.96622068e-01 3.94000024e-01
8.86628777e-02 4.97666299e-01 1.62276834e-01 -2.08986193e-01
-2.79957414e-01 -1.01704693e+00 3.64987373e-01 -4.52428639e-01
-1.27559936e+00 -8.06806386e-01 -7.99069703e-01 -6.59381552e-03
1.25220633e+00 -6.26852393e-01 -1.49168760e-01 -1.93444446e-01
-7.12246835e-01 1.54562861e-01 2.44193926e-01 2.92503923e-01
4.49075192e-01 -1.05759883e+00 -1.28373682e-01 -8.56742188e-02
-1.80178732e-01 -2.29495615e-01 1.06448762e-01 5.21014571e-01
-1.00909317e+00 6.82504952e-01 -5.95330477e-01 1.79703794e-02
-1.66400409e+00 5.68287432e-01 1.29424363e-01 1.08700544e-01
-5.26152134e-01 5.47348142e-01 -3.42367977e-01 -1.97260976e-01
-1.18360467e-01 -4.02596086e-01 -1.22674620e+00 3.85296941e-01
5.44809461e-01 3.34110439e-01 -5.35744578e-02 -1.16716623e+00
-4.08111125e-01 6.55138016e-01 3.20702255e-01 -3.45211715e-01
9.21652257e-01 -6.55994356e-01 -6.98347449e-01 8.24190974e-01
3.98122877e-01 3.75987440e-01 2.31848016e-01 -4.84006107e-03
4.43635017e-01 -4.47161883e-01 -2.96543855e-02 -8.81246626e-01
-2.17730403e-01 3.68219435e-01 5.40896237e-01 7.95221984e-01
1.02039480e+00 2.07980752e-01 4.50344652e-01 2.91095287e-01
4.48636204e-01 -1.51478803e+00 -5.49327016e-01 9.03603315e-01
5.35821795e-01 -8.93821836e-01 3.49996947e-02 -9.15020645e-01
-4.75113958e-01 1.46317208e+00 3.01897585e-01 1.96867660e-01
6.99308097e-01 2.97916800e-01 3.50319684e-01 -5.03490865e-01
-3.04653347e-01 -7.72873104e-01 1.10149294e-01 6.43393099e-01
8.41401041e-01 2.25086927e-01 -1.70706141e+00 1.00229633e+00
-9.34418857e-01 3.01046297e-03 6.79502964e-01 1.28982055e+00
-1.04564440e+00 -1.54688406e+00 -9.29984033e-01 6.22343607e-02
-6.58959866e-01 -1.72640041e-01 -6.42559409e-01 1.27844036e+00
7.12249756e-01 1.17512894e+00 -9.56156105e-02 -2.18525976e-01
1.42888844e-01 3.87547880e-01 4.83806789e-01 -6.00487709e-01
-7.37519860e-01 1.26165390e-01 4.47528541e-01 -3.19659412e-02
-1.10575628e+00 -5.13518155e-01 -1.65332723e+00 -5.76759577e-01
-1.28615528e-01 6.78584695e-01 6.78169370e-01 1.55601966e+00
-3.92476857e-01 4.18340445e-01 -2.94428617e-01 -3.14351022e-01
-2.89629269e-02 -9.29736257e-01 -5.98799467e-01 3.18265766e-01
-4.17433441e-01 -9.97743607e-01 -2.27594540e-01 2.16718707e-02] | [9.99087905883789, 9.286465644836426] |
17f06fe6-a732-4839-bb72-dd16e4c13ade | adapting-multi-lingual-asr-models-for | 2305.18747 | null | https://arxiv.org/abs/2305.18747v1 | https://arxiv.org/pdf/2305.18747v1.pdf | Adapting Multi-Lingual ASR Models for Handling Multiple Talkers | State-of-the-art large-scale universal speech models (USMs) show a decent automatic speech recognition (ASR) performance across multiple domains and languages. However, it remains a challenge for these models to recognize overlapped speech, which is often seen in meeting conversations. We propose an approach to adapt USMs for multi-talker ASR. We first develop an enhanced version of serialized output training to jointly perform multi-talker ASR and utterance timestamp prediction. That is, we predict the ASR hypotheses for all speakers, count the speakers, and estimate the utterance timestamps at the same time. We further introduce a lightweight adapter module to maintain the multilingual property of the USMs even when we perform the adaptation with only a single language. Experimental results obtained using the AMI and AliMeeting corpora show that our proposed approach effectively transfers the USMs to a strong multilingual multi-talker ASR model with timestamp prediction capability. | ['Michael Zeng', 'Yanmin Qian', 'Takuya Yoshioka', 'Dongmei Wang', 'Naoyuki Kanda', 'Zhuo Chen', 'Yao Qian', 'Chenda Li'] | 2023-05-30 | null | null | null | null | ['automatic-speech-recognition'] | ['speech'] | [-1.77920341e-01 5.66874072e-02 -1.76419108e-03 -6.53687477e-01
-1.37327492e+00 -5.18552959e-01 6.38149083e-01 -1.34504855e-01
-3.35531533e-01 4.38322395e-01 2.58570790e-01 -5.92495501e-01
5.32901645e-01 -2.44911052e-02 -7.01900244e-01 -4.43031818e-01
-9.95648280e-02 7.81668007e-01 2.73151785e-01 -3.56436700e-01
-3.10292572e-01 3.15120488e-01 -1.24402416e+00 4.84054893e-01
8.54401350e-01 7.88528502e-01 4.74912912e-01 1.13006377e+00
-3.39462012e-01 9.05741215e-01 -1.06590426e+00 -4.19171810e-01
-8.07559788e-02 -1.97663292e-01 -7.69583225e-01 1.52735159e-01
9.83237028e-02 -2.49200836e-01 -6.62050009e-01 5.58417857e-01
6.07808232e-01 2.75142819e-01 9.96928066e-02 -1.13115180e+00
-3.01241040e-01 1.22238696e+00 -2.00101957e-01 3.44641775e-01
5.15228629e-01 -1.91051781e-01 8.94895792e-01 -9.06273425e-01
3.10189694e-01 1.44090259e+00 3.35552305e-01 6.77727044e-01
-7.30574310e-01 -6.92006886e-01 4.54279274e-01 3.47987682e-01
-1.65018594e+00 -1.08105731e+00 4.58933115e-01 9.25196614e-03
1.22610569e+00 6.13698959e-01 9.90297198e-02 1.28240204e+00
-3.12721223e-01 9.98464704e-01 7.12611496e-01 -5.06065190e-01
1.63637370e-01 2.04350442e-01 1.57263264e-01 3.52889985e-01
-5.65986633e-01 -4.43135828e-01 -1.04864144e+00 -2.96135277e-01
2.19937563e-01 -2.18039170e-01 -1.48767367e-01 6.60374343e-01
-1.45496202e+00 4.97002363e-01 -4.20836270e-01 4.09036964e-01
-3.65400374e-01 -2.34913677e-01 5.57642937e-01 6.99557900e-01
7.68108368e-01 -1.76258460e-01 -7.19143569e-01 -5.64665020e-01
-1.04197955e+00 -1.38923004e-01 1.07380140e+00 1.19533908e+00
5.95170259e-01 3.15034062e-01 5.85876182e-02 1.31778002e+00
6.04372770e-02 8.31776619e-01 9.58865881e-01 -5.10131836e-01
6.26373827e-01 1.31201655e-01 6.78457096e-02 -1.33708015e-01
-1.10939339e-01 -2.04990491e-01 -6.45610034e-01 -6.72000349e-01
-2.97718048e-02 -1.61431909e-01 -8.37296784e-01 1.60589004e+00
3.59788120e-01 6.19138777e-01 4.88649845e-01 6.82816863e-01
5.01211703e-01 1.10218477e+00 -2.25777298e-01 -6.00353420e-01
1.29233205e+00 -1.26719379e+00 -9.21937764e-01 -3.47955257e-01
9.53886390e-01 -1.09167039e+00 1.00228322e+00 1.68407649e-01
-1.17244673e+00 -4.20198023e-01 -8.84852171e-01 5.95235415e-02
-1.20497145e-01 3.65106761e-01 2.94701159e-01 6.47733033e-01
-9.84660506e-01 1.89878251e-02 -1.02185941e+00 -5.09808958e-01
-7.04352498e-01 2.57446051e-01 -1.39933214e-01 3.94728556e-02
-1.38487458e+00 7.28640676e-01 2.37172008e-01 -2.87821703e-02
-9.81460989e-01 -3.85828823e-01 -7.28642702e-01 -5.24270423e-02
2.59747535e-01 -9.71952379e-02 2.01295877e+00 -9.39481914e-01
-2.08239865e+00 6.76361263e-01 -8.49635959e-01 -7.17208445e-01
3.87141258e-01 -1.39841065e-01 -1.09215415e+00 -2.32084133e-02
-1.72083378e-01 5.94905578e-02 5.68558276e-01 -9.00612414e-01
-9.53963518e-01 -3.14951777e-01 -2.86880702e-01 3.75143737e-01
-4.71596599e-01 5.78262269e-01 -5.64335942e-01 -6.24134958e-01
1.33458391e-01 -1.01699138e+00 -2.63729575e-03 -9.79958713e-01
-2.95832425e-01 -4.60648060e-01 7.46842146e-01 -1.10917687e+00
1.57014179e+00 -2.23861313e+00 -2.56785890e-03 5.49990498e-02
-3.79075080e-01 3.32160324e-01 -1.92291573e-01 6.64854527e-01
5.35356477e-02 -1.76955491e-01 7.55820870e-02 -9.88876760e-01
2.96583567e-02 4.21009094e-01 -6.18654251e-01 2.48803243e-01
-2.98504770e-01 4.36232626e-01 -6.13939166e-01 -3.89141530e-01
1.08553849e-01 2.11398661e-01 -1.22230001e-01 7.68480301e-01
-3.44134837e-01 4.25749958e-01 -2.16479838e-01 5.40358424e-01
4.99465972e-01 -3.21035795e-02 6.02985144e-01 2.07622319e-01
-2.67199010e-01 1.06880498e+00 -1.20428908e+00 1.75573373e+00
-8.40088606e-01 2.54572541e-01 3.70408565e-01 -8.50977123e-01
1.04173493e+00 1.04169345e+00 3.96660626e-01 -5.42219758e-01
-9.73538607e-02 6.53778553e-01 -4.08683941e-02 -2.75645256e-01
6.94771409e-01 2.44505350e-02 -2.15277433e-01 6.13239110e-01
1.00720629e-01 2.51220167e-01 -1.44395441e-01 2.70981371e-01
9.37594533e-01 -4.73695636e-01 3.39573234e-01 2.50705063e-01
8.04686785e-01 -3.95645291e-01 6.01893544e-01 5.71242332e-01
-5.13947718e-02 4.59630907e-01 -1.59819663e-01 -2.38929361e-01
-1.00062442e+00 -8.98304343e-01 3.41903418e-01 1.55403423e+00
-3.37960422e-01 -5.78892350e-01 -7.24580526e-01 -5.73481858e-01
-3.27820301e-01 9.86714363e-01 8.19743052e-02 9.79494527e-02
-9.64536309e-01 -3.66690367e-01 1.01176155e+00 2.38596946e-01
6.43647239e-02 -6.03697419e-01 2.05124691e-01 5.62008083e-01
-7.54998922e-01 -1.70850992e+00 -1.11103284e+00 -1.70438290e-02
-4.71923709e-01 -3.32107753e-01 -7.86868989e-01 -9.75609899e-01
1.15542769e-01 4.59078342e-01 7.45288551e-01 -1.47586718e-01
2.92641491e-01 3.14377874e-01 -6.02912903e-01 -1.16203599e-01
-1.29584813e+00 4.30583596e-01 7.29178727e-01 4.44283366e-01
3.15925270e-01 -5.85559309e-01 -1.44548178e-01 6.28768682e-01
-4.94950175e-01 -3.03246472e-02 4.09592658e-01 5.64046741e-01
2.58384049e-01 -4.27290499e-01 1.07469189e+00 -4.31834072e-01
3.71173173e-01 -4.85297203e-01 -3.29894006e-01 7.84772336e-01
-2.24992901e-01 4.67587188e-02 7.81623960e-01 -8.28678846e-01
-1.19463515e+00 -7.25817382e-02 -5.75425386e-01 -4.67734635e-01
-1.55666962e-01 3.57358128e-01 -3.05602491e-01 3.49723786e-01
2.47713566e-01 7.84352124e-01 -2.12068319e-01 -7.84670889e-01
4.01189119e-01 1.65455258e+00 5.48799574e-01 -6.32572532e-01
5.36405563e-01 6.70242235e-02 -8.74391437e-01 -1.22922373e+00
-4.03151572e-01 -9.93549049e-01 -3.29122305e-01 -1.32061718e-02
4.63219911e-01 -1.28316593e+00 -8.74088168e-01 5.00376701e-01
-1.51044488e+00 -2.67501563e-01 2.44372010e-01 7.64373541e-01
-5.24123669e-01 5.83606124e-01 -8.36470127e-01 -1.23705697e+00
-6.81064308e-01 -1.11378944e+00 1.15551746e+00 -2.37940356e-01
-1.88652277e-01 -8.02679360e-01 7.93107378e-04 4.96709585e-01
2.56168455e-01 -9.37797487e-01 5.65539777e-01 -1.32848167e+00
-4.68693733e-01 -1.98746040e-01 2.96979755e-01 3.79934132e-01
2.45342329e-01 -1.70845464e-01 -1.08279109e+00 -4.49548721e-01
3.05573363e-02 -1.81885660e-01 3.09542596e-01 1.81574538e-01
7.46894240e-01 -7.70167291e-01 -2.60824353e-01 3.33303452e-01
6.28583908e-01 4.73747492e-01 2.97665536e-01 3.66025493e-02
4.93972540e-01 3.89253438e-01 6.48118019e-01 5.91393411e-01
7.68157542e-01 8.82113397e-01 -2.87919044e-01 2.17507154e-01
3.23628709e-02 -2.64034599e-01 1.06306434e+00 2.17692423e+00
2.86757529e-01 -3.86398107e-01 -7.64346242e-01 6.46621406e-01
-1.85531330e+00 -7.36816466e-01 2.40467027e-01 2.52284646e+00
1.03152061e+00 -8.45721141e-02 2.46486291e-01 -2.14804783e-01
9.03483868e-01 1.91084921e-01 -1.95553392e-01 -6.07413590e-01
-2.71604627e-01 -3.33984978e-02 4.87303793e-01 6.64804518e-01
-5.04515171e-01 1.19599855e+00 6.17848253e+00 9.15442348e-01
-1.29968715e+00 5.98105431e-01 4.16408747e-01 -1.61142275e-01
-1.79380968e-01 2.27527916e-02 -1.06751382e+00 3.94200295e-01
1.90744281e+00 -5.00244796e-01 7.03647852e-01 7.88353562e-01
4.62427765e-01 4.81707215e-01 -1.26912236e+00 1.19623685e+00
1.80166662e-01 -1.02022147e+00 5.79762086e-02 -2.67116010e-01
4.31069762e-01 3.86095911e-01 -2.69252300e-01 4.88320768e-01
4.56641734e-01 -6.00813866e-01 9.36089218e-01 -6.59201890e-02
8.84450734e-01 -7.10139811e-01 4.73497987e-01 7.79203117e-01
-1.60663557e+00 -3.93911973e-02 -2.09556937e-01 2.59728432e-01
5.78116298e-01 1.69453532e-01 -1.56322074e+00 8.48525286e-01
3.49382788e-01 2.91720271e-01 -2.06633851e-01 6.12712801e-01
1.72592908e-01 9.27888036e-01 -4.48210567e-01 1.83622725e-02
1.01087071e-01 8.74360129e-02 7.00572789e-01 1.43642426e+00
8.08033466e-01 -5.64354956e-02 4.36463028e-01 9.20082349e-03
-2.27289990e-01 3.92025530e-01 -3.34753692e-01 -1.50300145e-01
1.09795821e+00 8.99167180e-01 -7.33440369e-02 -6.62867248e-01
-5.58082044e-01 1.39903784e+00 3.30410153e-01 4.25483912e-01
-7.95468926e-01 7.69352913e-02 8.65146279e-01 -1.19476296e-01
-1.63510884e-03 -6.50044918e-01 2.59615302e-01 -1.44166851e+00
1.02338426e-01 -1.31450057e+00 5.08189201e-01 -5.85465133e-01
-1.02643681e+00 9.80158627e-01 -4.47399497e-01 -1.01335680e+00
-6.85025096e-01 -1.10225812e-01 -5.38011611e-01 8.36035132e-01
-1.40248048e+00 -1.40240037e+00 3.95140231e-01 7.82940984e-01
1.24521518e+00 -3.81846935e-01 9.43632543e-01 6.49947882e-01
-5.67185283e-01 1.02407348e+00 2.01492667e-01 -4.07917611e-02
9.59489942e-01 -9.65121984e-01 8.02673519e-01 9.08060610e-01
1.74931213e-01 6.48879468e-01 6.71324015e-01 -6.23705745e-01
-1.65967286e+00 -9.90563035e-01 1.39948761e+00 -1.58186674e-01
8.67474079e-01 -7.39916623e-01 -1.03598440e+00 1.32878828e+00
2.81578362e-01 -4.55462962e-01 6.29936516e-01 3.56988996e-01
-3.35608453e-01 -3.89604419e-01 -5.17408371e-01 6.22687221e-01
8.29233468e-01 -9.57055569e-01 -6.19191468e-01 5.18702984e-01
1.41428816e+00 -3.96035999e-01 -8.83579612e-01 -1.64141227e-02
3.72720897e-01 -3.64153415e-01 7.30340421e-01 -4.28278834e-01
-6.24604464e-01 -1.11045495e-01 -5.96680105e-01 -1.40791333e+00
3.06085676e-01 -1.27111757e+00 -2.38954872e-01 1.56459022e+00
4.87624973e-01 -8.66774857e-01 3.15198153e-01 4.53862816e-01
-7.43503928e-01 6.22857660e-02 -1.45377040e+00 -1.11578667e+00
-4.22868788e-01 -6.10633910e-01 8.28071415e-01 8.01210403e-01
2.44835719e-01 4.34022367e-01 -7.11113393e-01 8.20688963e-01
3.42165381e-01 -1.26350164e-01 8.63642752e-01 -7.24407017e-01
-3.57927173e-01 8.43319222e-02 2.28316225e-02 -1.51033759e+00
4.50884789e-01 -8.63020539e-01 3.41016889e-01 -1.02060783e+00
-1.73691899e-01 -4.39969927e-01 -2.20165312e-01 2.69146800e-01
6.48290366e-02 -6.41319573e-01 -1.88392736e-02 3.30037743e-01
-7.16741145e-01 6.80425882e-01 5.52971244e-01 -4.56588157e-02
-4.16384816e-01 4.02995855e-01 -1.82614382e-02 4.50185597e-01
6.41541302e-01 -4.35372382e-01 -2.54463673e-01 -5.51577151e-01
-4.18552667e-01 7.01787889e-01 -1.57357424e-01 -9.88678634e-01
5.42382240e-01 -2.73446351e-01 -4.69925255e-01 -6.84015989e-01
5.98249376e-01 -5.93070805e-01 2.38667369e-01 2.20534220e-01
-3.38990927e-01 2.72871226e-01 4.26428393e-02 3.69951069e-01
-3.70372057e-01 1.14201427e-01 5.92177868e-01 3.66900340e-02
-6.14868701e-01 3.06117386e-01 -6.42484128e-01 -2.31460735e-01
6.76735342e-01 3.53253335e-01 -3.24418455e-01 -6.02988780e-01
-7.03328431e-01 1.37254179e-01 1.47940859e-01 7.26665080e-01
5.62683284e-01 -1.07058680e+00 -8.77943933e-01 2.32313648e-01
3.32064807e-01 -3.92368108e-01 4.96202469e-01 6.72633410e-01
9.17063430e-02 6.12186551e-01 5.49136817e-01 -5.30815721e-01
-1.57452857e+00 4.90702689e-01 2.78690517e-01 -2.59260058e-01
-5.27688026e-01 6.51968002e-01 7.97789544e-02 -6.64965451e-01
4.78617102e-01 -2.72746027e-01 2.67335206e-01 -1.36496425e-01
7.08913088e-01 2.94420749e-01 3.68023127e-01 -9.34208155e-01
-5.14545321e-01 -1.37424236e-02 -2.90798128e-01 -5.83812058e-01
1.00095439e+00 -8.77233028e-01 7.26196915e-02 9.17148352e-01
1.18097162e+00 2.53547877e-01 -7.05506921e-01 -7.16533780e-01
8.85964409e-02 5.05941585e-02 -2.88178474e-01 -5.85285604e-01
-4.72627074e-01 8.39434683e-01 3.79294217e-01 2.68715173e-01
7.61986613e-01 1.45946339e-01 1.54922330e+00 5.15080273e-01
5.39217055e-01 -1.33161950e+00 -5.50341569e-02 8.18196833e-01
7.26221800e-01 -1.10645890e+00 -7.47682869e-01 -7.34276772e-02
-9.77654338e-01 9.06605363e-01 4.93486762e-01 7.65778661e-01
4.84062403e-01 2.95312732e-01 4.55775887e-01 4.20624733e-01
-1.24405324e+00 -5.33367358e-02 -1.69481598e-02 2.44614124e-01
4.79169875e-01 5.88246107e-01 -7.25457519e-02 7.37348557e-01
-3.45478773e-01 -2.04992473e-01 5.61247230e-01 7.07198381e-01
-5.24502933e-01 -1.26046336e+00 -4.33278918e-01 -1.49005026e-01
-5.74780881e-01 -3.64575386e-01 -6.00922555e-02 1.70220092e-01
-6.13880575e-01 1.43272614e+00 -3.85495983e-02 -6.20027184e-01
2.75321186e-01 7.15617657e-01 -1.39184996e-01 -6.81405067e-01
-5.32333851e-01 3.26263845e-01 4.92749870e-01 -2.95372307e-01
-3.07771750e-02 -5.67408502e-01 -1.34283984e+00 -4.82399583e-01
-2.68894792e-01 5.23114860e-01 1.07892096e+00 9.96175766e-01
5.12776077e-01 3.46608132e-01 1.12551725e+00 -4.87415135e-01
-9.30149317e-01 -1.34435678e+00 -6.62395954e-01 1.96763471e-01
5.59857488e-01 -3.13791819e-02 -3.65186095e-01 1.80726856e-01] | [14.547231674194336, 6.522904872894287] |
34881ecc-ce83-45d8-8fa5-458e4c156891 | exploring-attention-mechanisms-for-multimodal | 2306.07115 | null | https://arxiv.org/abs/2306.07115v1 | https://arxiv.org/pdf/2306.07115v1.pdf | Exploring Attention Mechanisms for Multimodal Emotion Recognition in an Emergency Call Center Corpus | The emotion detection technology to enhance human decision-making is an important research issue for real-world applications, but real-life emotion datasets are relatively rare and small. The experiments conducted in this paper use the CEMO, which was collected in a French emergency call center. Two pre-trained models based on speech and text were fine-tuned for speech emotion recognition. Using pre-trained Transformer encoders mitigates our data's limited and sparse nature. This paper explores the different fusion strategies of these modality-specific models. In particular, fusions with and without cross-attention mechanisms were tested to gather the most relevant information from both the speech and text encoders. We show that multimodal fusion brings an absolute gain of 4-9% with respect to either single modality and that the Symmetric multi-headed cross-attention mechanism performed better than late classical fusion approaches. Our experiments also suggest that for the real-life CEMO corpus, the audio component encodes more emotive information than the textual one. | ['Laurence Devillers', 'Lori Lamel', 'Théo Deschamps-Berger'] | 2023-06-12 | null | null | null | null | ['multimodal-emotion-recognition', 'speech-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech', 'speech'] | [ 2.43469626e-01 2.40218360e-02 2.44351611e-01 -5.01711905e-01
-1.08951199e+00 -3.01685967e-02 5.82278609e-01 3.44761848e-01
-6.16782367e-01 8.07714283e-01 5.87703109e-01 9.46075916e-02
-3.21293138e-02 -3.34521502e-01 -1.81991875e-01 -7.07565546e-01
9.75613892e-02 2.50480354e-01 -4.02117610e-01 -5.47241390e-01
-2.82483008e-02 2.71010071e-01 -1.87272501e+00 7.40813911e-01
8.32508266e-01 1.39715827e+00 2.71165937e-01 8.39207828e-01
-1.94423109e-01 1.04790103e+00 -7.71989584e-01 -6.81021392e-01
-2.04355374e-01 -2.95204699e-01 -7.80785680e-01 1.26533657e-01
-6.78010583e-02 -2.79001938e-03 -3.49743396e-01 8.97643745e-01
1.05348074e+00 4.03925925e-01 3.92612785e-01 -1.36766160e+00
-2.12376118e-01 6.39033794e-01 -3.70061457e-01 1.11385107e-01
6.38612926e-01 -2.95406282e-01 8.99429262e-01 -9.32826281e-01
3.62710506e-01 1.21053505e+00 5.19886196e-01 4.69018310e-01
-7.09481537e-01 -5.23710907e-01 -9.53158736e-02 3.79480511e-01
-1.19766891e+00 -8.07698905e-01 9.03487563e-01 -1.31446347e-01
1.32697105e+00 4.99126494e-01 2.22075060e-01 1.45436740e+00
3.85233074e-01 9.22817528e-01 1.33668947e+00 -5.57248950e-01
7.26331607e-04 6.11628234e-01 -1.81652345e-02 4.40288156e-01
-5.27614772e-01 -5.70994131e-02 -8.95511568e-01 -2.94611275e-01
1.03230573e-01 -6.75079674e-02 -4.34842199e-01 3.45906436e-01
-1.10890901e+00 8.59128475e-01 2.78371852e-02 8.01805854e-01
-8.87068868e-01 -2.48794898e-01 7.27687895e-01 5.37720084e-01
7.89454818e-01 3.80167812e-01 -6.30198300e-01 -5.77249110e-01
-9.87408400e-01 -2.72299886e-01 8.13539505e-01 5.72107792e-01
4.50676531e-01 9.19429734e-02 -3.13410312e-01 1.23249078e+00
9.76476967e-02 3.29403907e-01 7.16390312e-01 -6.91990316e-01
5.33886552e-01 3.17827284e-01 -1.08178921e-01 -1.02516663e+00
-5.20384908e-01 -2.92572707e-01 -7.87041247e-01 -7.60852918e-02
-6.01993613e-02 -7.54409134e-01 -7.79058278e-01 1.80784655e+00
1.68788895e-01 -9.09833312e-02 5.42305291e-01 7.94311345e-01
1.07473528e+00 8.62002432e-01 1.97645307e-01 -3.23025823e-01
1.67211282e+00 -1.02157843e+00 -1.25708938e+00 -3.64522338e-01
5.79691052e-01 -1.11238086e+00 6.56742334e-01 4.44463342e-01
-1.21069241e+00 -4.52097774e-01 -8.87531579e-01 7.22970366e-02
-5.68742633e-01 3.43661666e-01 3.91038686e-01 7.05979645e-01
-9.68789816e-01 1.88494191e-01 -4.10493821e-01 -6.06166899e-01
5.34368381e-02 3.70754123e-01 -6.72992527e-01 1.34414390e-01
-1.58447742e+00 1.22718835e+00 4.06549573e-01 1.60609394e-01
-4.36886013e-01 -2.59946108e-01 -8.73979092e-01 3.46322149e-01
4.32260871e-01 -5.11605084e-01 1.28443611e+00 -1.28265083e+00
-1.67413116e+00 6.83156133e-01 -3.37198466e-01 -4.49535251e-01
9.21348035e-02 -2.67147303e-01 -8.81457865e-01 4.27715898e-01
-2.83828139e-01 6.69942677e-01 9.11196411e-01 -1.07377779e+00
-6.92338169e-01 -3.43512535e-01 2.58300193e-02 5.06268561e-01
-5.29357612e-01 5.78926980e-01 -2.00983360e-01 -7.75928676e-01
-4.05430496e-01 -7.46938169e-01 -1.52695496e-02 -6.37077451e-01
-1.96252495e-01 -5.86855225e-02 7.75058746e-01 -8.81904364e-01
1.39139616e+00 -2.29211950e+00 2.39040807e-01 1.09640770e-01
3.12239695e-02 2.30780244e-01 -9.03312042e-02 8.09392452e-01
-3.60005677e-01 -1.21929444e-01 -1.78779125e-01 -6.54291868e-01
1.12920359e-01 1.69702396e-01 -8.42403620e-02 1.07763996e-02
3.92250896e-01 5.95706284e-01 -5.54279804e-01 -6.26089036e-01
3.33002776e-01 9.34217453e-01 -2.25411132e-01 3.82739335e-01
3.34442317e-01 2.61359662e-01 -1.14791617e-01 5.51637948e-01
3.54060888e-01 1.43120021e-01 1.55152887e-01 -3.97968560e-01
8.91392678e-02 1.59434915e-01 -9.22872365e-01 1.68273926e+00
-7.69074082e-01 7.06615210e-01 4.69502330e-01 -1.00977170e+00
9.05835211e-01 9.14797783e-01 6.16347134e-01 -8.02903473e-01
6.94144726e-01 1.60746291e-01 -4.34997166e-03 -7.73919642e-01
8.49502385e-01 -5.89894652e-01 -2.85655916e-01 1.83245763e-01
5.46589434e-01 1.48053942e-02 1.03044689e-01 2.69500434e-01
9.90719020e-01 -4.20463443e-01 3.25357288e-01 2.31971927e-02
6.21396005e-01 -3.74620646e-01 3.18765491e-01 3.57615501e-01
-2.26900339e-01 5.99106431e-01 3.70774031e-01 2.41410837e-01
-4.85586047e-01 -3.16204250e-01 6.93944618e-02 1.28863740e+00
-2.29376316e-01 -5.64242482e-01 -7.65666962e-01 -5.24384618e-01
-4.91543174e-01 8.09010029e-01 -7.44573355e-01 -4.09381539e-01
-9.74501856e-03 -8.87581229e-01 6.72841311e-01 4.72010732e-01
3.46329361e-01 -1.18646371e+00 -8.19715023e-01 2.83073902e-01
-8.59153569e-01 -1.38005567e+00 -3.77185643e-01 5.67996681e-01
-3.18841666e-01 -7.28382528e-01 -8.12532723e-01 -4.21289295e-01
1.85173452e-01 3.25498581e-02 8.94077182e-01 -3.32012832e-01
-2.26667710e-03 7.86231875e-01 -8.45360994e-01 -6.49155557e-01
-3.08059096e-01 -1.14609376e-01 -8.31597745e-02 6.47837222e-01
4.59008008e-01 -2.54246652e-01 -1.62023202e-01 -5.43527007e-02
-9.02382910e-01 -1.16051614e-01 5.98341227e-01 1.01587546e+00
-7.27891624e-02 2.11040810e-01 6.70611441e-01 -4.17897254e-01
1.02262986e+00 -5.95287383e-01 2.70013094e-01 3.46778899e-01
-1.82543829e-01 1.55977486e-02 1.69086933e-01 -3.12843233e-01
-1.63938379e+00 -9.57996771e-02 -4.70148593e-01 -4.02437031e-01
-3.07695031e-01 7.37691939e-01 -2.43374571e-01 1.08741574e-01
2.23719627e-01 -1.35586694e-01 -5.92945144e-02 -2.82069951e-01
6.90658912e-02 1.26747227e+00 4.38342959e-01 -5.29532969e-01
-1.77805871e-01 1.43437922e-01 -5.32065451e-01 -1.05294955e+00
-4.20403093e-01 -5.13169169e-01 -2.57385522e-01 -4.68846589e-01
1.03036976e+00 -8.51844072e-01 -6.97552741e-01 4.35888797e-01
-1.23498726e+00 -6.07022196e-02 -1.62413880e-01 7.90564597e-01
-4.07407165e-01 2.60779709e-01 -7.90862739e-01 -1.24940419e+00
-5.51891387e-01 -1.19853592e+00 1.27687097e+00 1.03081375e-01
-3.73035073e-01 -9.26269710e-01 -9.45650116e-02 6.11515164e-01
3.75557423e-01 -1.17151253e-03 5.96262753e-01 -9.33916867e-01
5.10706425e-01 -2.41710439e-01 -1.50410876e-01 4.46046531e-01
5.16134351e-02 8.50158744e-03 -1.47585595e+00 -5.55607080e-02
4.18751746e-01 -5.88368475e-01 7.92725801e-01 1.13596067e-01
8.46954346e-01 6.87754154e-02 2.21248865e-02 -1.77503645e-01
9.80273306e-01 5.20789385e-01 5.39008617e-01 -1.98204234e-01
3.08543980e-01 1.04219556e+00 6.39918566e-01 7.78015375e-01
4.57362443e-01 5.34876347e-01 2.63502598e-01 -2.13274375e-01
7.95799941e-02 2.52643436e-01 6.17958367e-01 1.29942942e+00
-1.86641701e-02 -6.74139917e-01 -7.83986807e-01 5.25389135e-01
-1.90994859e+00 -1.08436644e+00 1.09221756e-01 2.04713249e+00
6.98176920e-01 -1.78792760e-01 -6.89055324e-02 4.19699997e-01
7.27341831e-01 7.14782923e-02 7.15747327e-02 -7.20240116e-01
-4.19315815e-01 3.78238559e-01 -5.26165143e-02 6.13288999e-01
-1.16602933e+00 4.32822227e-01 6.04145336e+00 9.79587793e-01
-1.33132374e+00 3.51334304e-01 6.19854748e-01 -3.76907051e-01
-5.27498908e-02 -5.16880751e-01 -2.73461163e-01 5.35860538e-01
1.53592038e+00 7.49486359e-03 1.42278850e-01 3.58718187e-01
9.59901288e-02 -3.67234707e-01 -8.01422596e-01 1.32159424e+00
4.29762214e-01 -7.61805117e-01 -4.78426427e-01 -3.13013755e-02
4.23582852e-01 1.18557297e-01 -6.87639117e-02 5.18316269e-01
-1.32852659e-01 -8.73710454e-01 8.19072306e-01 6.73951328e-01
4.52350587e-01 -9.85649467e-01 1.22932649e+00 2.65920192e-01
-1.05847657e+00 -1.74228653e-01 1.46986529e-01 1.08104467e-01
4.69549179e-01 6.18458033e-01 -5.03930449e-01 1.01148939e+00
7.53125727e-01 2.93211251e-01 -2.78229892e-01 6.45426691e-01
1.43835202e-01 5.43199599e-01 -5.20961940e-01 -6.39286116e-02
4.01578665e-01 5.73140606e-02 5.50874293e-01 1.59312534e+00
4.85247582e-01 1.56112164e-01 -4.98551838e-02 8.32237974e-02
-2.68004853e-02 3.86062086e-01 -4.83602673e-01 -2.81017691e-01
4.89498191e-02 1.46961677e+00 -5.67649841e-01 -5.75983763e-01
-5.99537790e-01 1.16562974e+00 8.71255621e-02 3.34506810e-01
-1.06887364e+00 -6.32814467e-01 3.87848884e-01 -5.24302125e-01
2.72763431e-01 2.08140463e-01 -9.41585004e-03 -1.24188340e+00
-8.37439150e-02 -1.03227413e+00 4.28803921e-01 -1.07205212e+00
-1.21301579e+00 1.21215117e+00 4.85461280e-02 -8.11667442e-01
-3.29611868e-01 -4.90476042e-01 -4.13149118e-01 8.29206347e-01
-1.08803046e+00 -1.04264271e+00 -7.70637393e-02 7.07006693e-01
6.00905597e-01 -1.32025793e-01 1.03323233e+00 8.16340268e-01
-6.43187582e-01 4.47556257e-01 -1.84226513e-01 -1.48350686e-01
8.42026591e-01 -1.00147164e+00 -5.08045375e-01 5.79783976e-01
1.45438194e-01 3.29979330e-01 8.51943910e-01 -3.50388169e-01
-1.15433216e+00 -5.16046107e-01 1.32899296e+00 -7.27367327e-02
5.67736506e-01 -1.80353105e-01 -9.82612669e-01 3.86847973e-01
1.13057923e+00 -4.54021424e-01 9.77197409e-01 1.71033248e-01
-1.15129389e-01 3.21161821e-02 -1.23314261e+00 3.58581364e-01
3.95320415e-01 -9.04764235e-01 -7.91066945e-01 -3.58727910e-02
5.03353179e-01 -2.19614372e-01 -1.08486199e+00 4.81302708e-01
4.32665318e-01 -9.39869106e-01 5.21426976e-01 -6.05166078e-01
4.41799432e-01 5.94584271e-02 -6.15669489e-01 -1.52421272e+00
4.28275391e-02 -5.42874873e-01 -1.57513004e-02 1.54132700e+00
4.51361656e-01 -5.89041173e-01 1.71700805e-01 6.15724742e-01
-2.28370547e-01 -7.58821368e-01 -9.69266534e-01 -1.60310909e-01
-3.23858619e-01 -8.21112275e-01 3.99350703e-01 1.05200505e+00
5.28328478e-01 7.92021334e-01 -7.58395016e-01 -2.10057154e-01
-8.51575360e-02 -1.68101236e-01 2.27830365e-01 -8.99576426e-01
-5.09189926e-02 -4.42936152e-01 -2.79453039e-01 -4.08001184e-01
3.20431113e-01 -6.11400604e-01 1.15287498e-01 -1.24249053e+00
2.15086211e-02 1.95425019e-01 -5.49064577e-01 5.94389975e-01
-1.70430616e-01 1.20382428e-01 3.69008809e-01 -5.65607667e-01
-4.88059610e-01 9.60008383e-01 7.05767572e-01 -1.38914585e-01
-2.62370910e-02 -2.55673677e-01 -6.61771357e-01 6.81016743e-01
8.28523695e-01 -3.03179502e-01 -4.65185612e-01 -1.52410343e-01
2.34213859e-01 3.82416904e-01 1.83963463e-01 -8.62096012e-01
4.00819451e-01 2.13490099e-01 5.58213480e-02 -4.85964745e-01
1.06593871e+00 -1.09237707e+00 1.31593361e-01 2.15349272e-02
-3.66340607e-01 8.49168301e-02 5.13054967e-01 2.50342518e-01
-7.67849028e-01 -2.83766538e-01 6.18833959e-01 5.39409295e-02
-6.09896362e-01 -1.64270461e-01 -9.01834965e-01 -1.90210223e-01
8.76104176e-01 1.44456998e-01 2.98529887e-03 -8.76753509e-01
-1.03695631e+00 6.96142539e-02 -2.90827811e-01 5.40689111e-01
6.32694066e-01 -1.40044022e+00 -7.45460868e-01 -3.54493270e-03
9.51000080e-02 -8.17295790e-01 6.95314467e-01 1.27352142e+00
1.48998559e-01 4.42365676e-01 -3.44089776e-01 -1.93955213e-01
-1.69879496e+00 5.01750946e-01 2.75703251e-01 -2.84662157e-01
-7.73963481e-02 7.82166958e-01 -6.90118745e-02 -5.11777818e-01
3.03597450e-01 -1.43150777e-01 -3.62274289e-01 6.96641088e-01
5.88796675e-01 4.34117407e-01 3.72545063e-01 -1.18194580e+00
-5.22070348e-01 1.43515617e-01 1.17161229e-01 -6.35503173e-01
1.23187637e+00 -3.94122332e-01 -1.33585036e-01 6.27031088e-01
1.34304249e+00 -5.74640185e-02 -4.53279883e-01 -2.03371659e-01
-1.79692343e-01 -1.45187438e-01 3.47222090e-01 -1.04579544e+00
-1.13881922e+00 1.21214938e+00 5.75257063e-01 3.51017624e-01
1.66217339e+00 -8.31432268e-02 6.76427245e-01 3.61960739e-01
1.22820571e-01 -1.34555578e+00 -1.59045741e-01 4.72761154e-01
1.14782822e+00 -1.29568994e+00 -4.06641126e-01 -2.48718604e-01
-1.29551661e+00 1.02146149e+00 3.70685488e-01 4.52180207e-01
6.60565913e-01 4.29206222e-01 2.27360785e-01 -2.79730648e-01
-1.12237954e+00 -5.70792139e-01 4.27157998e-01 2.28788391e-01
6.90153062e-01 1.49357766e-02 -2.31717452e-01 9.86255825e-01
-5.06680459e-02 -4.74069491e-02 1.59017548e-01 9.16739702e-01
-1.29894987e-01 -1.01561403e+00 -6.28888786e-01 4.74423110e-01
-8.73356044e-01 -1.53875083e-01 -5.16644061e-01 4.98335540e-01
1.18727230e-01 1.47417462e+00 -7.45885298e-02 -6.84281528e-01
3.16789836e-01 5.38223922e-01 4.11364138e-01 -1.10228702e-01
-1.11363816e+00 4.08993542e-01 6.54188991e-01 -6.07042611e-01
-8.50397408e-01 -6.67055547e-01 -1.07130754e+00 -1.79737523e-01
-3.91898006e-01 4.71056104e-01 8.04093897e-01 1.11633134e+00
5.91127038e-01 9.21912134e-01 5.48040152e-01 -8.59394252e-01
-1.71167597e-01 -1.24672735e+00 -4.09659237e-01 3.75496328e-01
2.55006015e-01 -6.33034408e-01 -2.96488047e-01 -5.17100049e-03] | [13.270339012145996, 5.364918231964111] |
75d697eb-d464-4aef-8417-120575fe1130 | predicting-the-next-action-by-modeling-the | 2209.05044 | null | https://arxiv.org/abs/2209.05044v4 | https://arxiv.org/pdf/2209.05044v4.pdf | Predicting the Next Action by Modeling the Abstract Goal | The problem of anticipating human actions is an inherently uncertain one. However, we can reduce this uncertainty if we have a sense of the goal that the actor is trying to achieve. Here, we present an action anticipation model that leverages goal information for the purpose of reducing the uncertainty in future predictions. Since we do not possess goal information or the observed actions during inference, we resort to visual representation to encapsulate information about both actions and goals. Through this, we derive a novel concept called abstract goal which is conditioned on observed sequences of visual features for action anticipation. We design the abstract goal as a distribution whose parameters are estimated using a variational recurrent network. We sample multiple candidates for the next action and introduce a goal consistency measure to determine the best candidate that follows from the abstract goal. Our method obtains impressive results on the very challenging Epic-Kitchens55 (EK55), EK100, and EGTEA Gaze+ datasets. We obtain absolute improvements of +13.69, +11.24, and +5.19 for Top-1 verb, Top-1 noun, and Top-1 action anticipation accuracy respectively over prior state-of-the-art methods for seen kitchens (S1) of EK55. Similarly, we also obtain significant improvements in the unseen kitchens (S2) set for Top-1 verb (+10.75), noun (+5.84) and action (+2.87) anticipation. Similar trend is observed for EGTEA Gaze+ dataset, where absolute improvement of +9.9, +13.1 and +6.8 is obtained for noun, verb, and action anticipation. It is through the submission of this paper that our method is currently the new state-of-the-art for action anticipation in EK55 and EGTEA Gaze+ https://competitions.codalab.org/competitions/20071#results Code available at https://github.com/debadityaroy/Abstract_Goal | ['Basura Fernando', 'Debaditya Roy'] | 2022-09-12 | null | null | null | null | ['action-anticipation'] | ['computer-vision'] | [ 1.62893429e-01 2.69963771e-01 -9.47664306e-02 -4.63554978e-01
-9.64133739e-01 -4.47445840e-01 7.39166617e-01 -2.98750401e-01
-4.47148502e-01 5.99050641e-01 5.62518120e-01 7.12487176e-02
-9.12116989e-02 -2.90501833e-01 -8.68903339e-01 -5.92330754e-01
1.67667102e-02 2.76883125e-01 -1.26855865e-01 -2.36708567e-01
1.82620555e-01 -1.15818657e-01 -1.62870765e+00 5.13595462e-01
5.03680229e-01 9.95401740e-01 2.14441821e-01 7.64642119e-01
4.25738811e-01 7.74849057e-01 -1.14475802e-01 -4.50592339e-01
2.85906494e-01 -4.85986680e-01 -6.84771836e-01 -9.97614264e-02
3.75252634e-01 -2.56507337e-01 -3.16485941e-01 8.20167661e-01
4.84269321e-01 3.81890744e-01 7.56023884e-01 -1.43198872e+00
-7.13705719e-01 7.75879920e-01 -6.18336797e-01 1.43903466e-02
6.78816974e-01 6.12331450e-01 1.18570828e+00 -7.17384994e-01
7.58184075e-01 1.52678216e+00 3.93904299e-01 9.37581480e-01
-1.03535938e+00 -6.02147400e-01 7.19996393e-01 3.80037546e-01
-1.26975369e+00 -5.56766987e-01 6.89656734e-01 -5.63956797e-01
1.18279910e+00 2.39286244e-01 7.18640804e-01 1.67785454e+00
3.40444565e-01 1.25774503e+00 1.06714034e+00 -3.50981712e-01
1.56539589e-01 -4.86421436e-02 2.29001448e-01 5.20625174e-01
-3.71097624e-01 5.17407775e-01 -9.71587300e-01 2.54882395e-01
2.39282325e-01 -1.08161926e-01 -2.62970835e-01 4.88918647e-02
-1.09519374e+00 6.64278924e-01 3.71982992e-01 -2.81049371e-01
-7.76664019e-01 6.36454701e-01 1.97348207e-01 2.97683179e-02
4.45739239e-01 1.90053344e-01 -4.19233412e-01 -5.91004252e-01
-6.30340874e-01 6.10407054e-01 4.32302177e-01 9.46289659e-01
4.24769640e-01 -3.38132121e-02 -7.19030738e-01 4.42414820e-01
7.48822331e-01 7.22771227e-01 2.37792417e-01 -1.20668638e+00
3.73062283e-01 2.89650679e-01 5.29950380e-01 -6.36868358e-01
-5.76531708e-01 -2.04159155e-01 -4.63894665e-01 3.56902868e-01
5.33573568e-01 -2.23937660e-01 -1.22310138e+00 2.21397114e+00
3.00594091e-01 1.31128907e-01 3.50617766e-01 8.86543572e-01
8.14133167e-01 7.34493434e-01 4.72613096e-01 -3.09135258e-01
1.25948811e+00 -9.04808283e-01 -6.78532958e-01 -5.38273990e-01
5.94509482e-01 -5.16301811e-01 1.26423740e+00 4.49169129e-01
-9.68117893e-01 -5.34393489e-01 -6.42889142e-01 1.86290313e-02
-5.89623041e-02 1.17502041e-01 6.56877816e-01 2.12263510e-01
-1.08615983e+00 5.62128663e-01 -9.51285720e-01 -3.21565777e-01
3.61863732e-01 2.77547628e-01 -1.93077073e-01 7.96607435e-02
-1.25513470e+00 9.90076125e-01 3.79607707e-01 1.24405347e-01
-1.18406832e+00 -6.86389387e-01 -8.51584911e-01 -1.20384492e-01
5.97175300e-01 -6.40846908e-01 1.44044244e+00 -9.38267827e-01
-1.40622628e+00 8.01529348e-01 -4.04683352e-01 -8.34135890e-01
6.75959229e-01 -7.14947820e-01 -2.62815088e-01 -1.88042253e-01
8.87637436e-02 1.16767454e+00 8.80577743e-01 -1.02543437e+00
-7.97383189e-01 -3.35160941e-01 1.47544712e-01 5.63213170e-01
1.96180746e-01 -4.04789522e-02 -3.68814856e-01 -3.43947142e-01
-2.05974117e-01 -1.28261936e+00 -9.91020575e-02 -1.40905917e-01
-5.89700222e-01 -6.55428588e-01 4.76554811e-01 -5.92891812e-01
1.28193867e+00 -2.26629806e+00 1.71522513e-01 -2.48448610e-01
1.02000959e-01 -2.11760700e-02 -1.06317423e-01 2.62281865e-01
-1.41698599e-01 9.87877324e-02 1.11206314e-02 -7.39849865e-01
3.26244354e-01 -2.94078849e-02 -3.85157168e-01 3.06972295e-01
1.79982245e-01 1.03482330e+00 -8.91765058e-01 -3.51531178e-01
4.26217943e-01 4.01640564e-01 -5.20911813e-01 2.54256934e-01
-5.21769106e-01 4.89485413e-01 -5.36993325e-01 5.87932706e-01
2.80638248e-01 -2.37259090e-01 5.36781438e-02 5.45731112e-02
-2.00879246e-01 2.43297607e-01 -1.04757905e+00 1.60209560e+00
-2.71797359e-01 5.70407569e-01 -4.91170138e-01 -4.66953099e-01
6.90759301e-01 3.15089673e-01 3.28657091e-01 -6.65149331e-01
7.67203942e-02 -2.09075660e-01 8.86081904e-02 -5.34867167e-01
5.06316245e-01 -3.47563252e-02 -3.46374214e-01 2.04384387e-01
-1.46606386e-01 4.65070456e-02 1.95990622e-01 2.68906295e-01
9.50340569e-01 7.61605263e-01 3.56460541e-01 1.30535988e-02
1.92279398e-01 -5.40059917e-02 6.47774816e-01 9.10016894e-01
-5.16946912e-01 5.89508712e-01 6.17797852e-01 -3.90179247e-01
-5.64955950e-01 -1.04877055e+00 2.48458624e-01 1.15778112e+00
1.25103503e-01 -5.14615655e-01 -5.21329939e-01 -6.12624526e-01
-1.24713309e-01 1.32380486e+00 -9.43115354e-01 -1.87178120e-01
-5.12009859e-01 -4.12002385e-01 1.82175726e-01 7.77416885e-01
4.24051821e-01 -1.56258309e+00 -8.64099860e-01 4.60932404e-03
-3.49010587e-01 -9.11652684e-01 -5.43387353e-01 -1.09069310e-01
-4.95284885e-01 -9.41511214e-01 -4.06771183e-01 -8.42619762e-02
2.04233900e-01 -2.06023842e-01 1.07110393e+00 -2.05597267e-01
1.57948539e-01 5.38189173e-01 -4.24943686e-01 -7.53983557e-01
-1.40891194e-01 -3.55126977e-01 2.61045426e-01 6.21245913e-02
4.57827628e-01 -3.53906542e-01 -6.85505450e-01 -3.78979482e-02
-2.13188738e-01 4.17809069e-01 4.70987439e-01 7.21746624e-01
6.87026918e-01 -5.38889885e-01 4.19952750e-01 -6.07588589e-01
3.28054309e-01 -3.54811609e-01 -6.07937992e-01 1.89568058e-01
-7.04229653e-01 1.61493063e-01 3.46932530e-01 -3.91807139e-01
-1.34188104e+00 3.73279721e-01 -5.29928766e-02 -5.91363430e-01
-3.36059034e-01 4.34242994e-01 1.12169854e-01 6.27390742e-01
6.19863331e-01 1.14432216e-01 -3.39862347e-01 -2.47007653e-01
4.78790730e-01 4.64473248e-01 4.70786124e-01 -2.46041358e-01
3.54751587e-01 3.76500696e-01 -1.73897743e-01 -5.05425751e-01
-1.08717930e+00 -2.35700056e-01 -2.26303592e-01 -5.03917515e-01
1.05787015e+00 -1.06317544e+00 -1.19678664e+00 5.95834911e-01
-1.18679082e+00 -7.45280981e-01 -2.88472891e-01 6.04411900e-01
-7.90974796e-01 9.42260474e-02 -1.33391842e-01 -1.29145682e+00
-4.25975084e-01 -1.09440351e+00 1.07157910e+00 4.12254423e-01
-7.46420443e-01 -7.53357351e-01 -7.36579448e-02 3.34969461e-01
9.79209668e-04 6.16592169e-01 3.13298076e-01 -4.86494690e-01
-7.13690162e-01 -2.18529832e-02 1.25249594e-01 1.57696187e-01
-1.12781160e-01 1.48143098e-01 -1.01506209e+00 -1.58035040e-01
-3.00230294e-01 -3.88668954e-01 1.02348745e+00 7.92842627e-01
8.61861825e-01 -1.49855629e-01 -4.28460389e-01 3.93637180e-01
1.16534984e+00 4.48827803e-01 6.94477320e-01 1.85100853e-01
6.58078134e-01 6.02725446e-01 1.24078476e+00 7.68607378e-01
5.81395805e-01 7.70627379e-01 7.16182470e-01 4.66291547e-01
-1.01848669e-01 -4.75528091e-01 8.49076331e-01 -3.79227959e-02
-2.94764370e-01 -7.24640131e-01 -9.64230716e-01 8.33278179e-01
-2.20425534e+00 -1.08038580e+00 -1.73357993e-01 2.09553242e+00
6.13108397e-01 2.48873204e-01 1.70210034e-01 -2.40104586e-01
5.65306067e-01 4.34031129e-01 -7.01465368e-01 -3.69976372e-01
2.49206066e-01 -6.91934228e-02 1.41786084e-01 5.93639255e-01
-1.24268484e+00 1.30981588e+00 5.43299770e+00 6.54229403e-01
-8.77176762e-01 -1.02851652e-01 6.42429113e-01 -3.62887532e-01
-9.93656516e-02 3.80539596e-02 -1.19504285e+00 5.51280200e-01
1.11311054e+00 -2.37803087e-01 4.13168877e-01 7.74792492e-01
5.52300990e-01 -4.10625428e-01 -1.38762391e+00 1.13149476e+00
-3.80506963e-02 -1.09979403e+00 -2.34412432e-01 -2.29713857e-01
5.67217708e-01 6.54274672e-02 4.17307615e-01 5.38301885e-01
5.66133320e-01 -1.16402984e+00 1.10287094e+00 9.39934075e-01
5.67027748e-01 -6.76294625e-01 5.12472689e-01 3.67850095e-01
-9.49750304e-01 -1.49003029e-01 1.04580447e-02 -3.67272675e-01
4.18415248e-01 5.89034557e-02 -7.42793262e-01 3.26974690e-01
9.14222240e-01 1.08152771e+00 -3.17539603e-01 5.03130376e-01
-7.63474882e-01 5.52301884e-01 -3.37203354e-01 -2.56264001e-01
3.87913615e-01 1.45274177e-01 7.71146119e-01 8.82134795e-01
2.65240908e-01 3.22250307e-01 6.21240772e-02 7.79644430e-01
-3.23512778e-02 -2.27179989e-01 -4.83442664e-01 -1.16511054e-01
3.27947378e-01 9.58402693e-01 -2.86747932e-01 -2.78792024e-01
-1.40868500e-01 8.54896188e-01 2.77951241e-01 3.56383443e-01
-1.23193347e+00 1.00193836e-01 8.38678241e-01 -2.51627326e-01
3.64592046e-01 1.87280566e-01 -1.26255184e-01 -8.98078918e-01
5.85855432e-02 -8.03330362e-01 6.43858969e-01 -1.17421639e+00
-9.18242216e-01 5.74206650e-01 3.45079660e-01 -1.18853021e+00
-7.93466568e-01 -4.80721295e-01 -5.47650814e-01 8.30531836e-01
-1.10262144e+00 -1.29002059e+00 -2.70058900e-01 4.21513826e-01
7.29431748e-01 5.87693825e-02 6.54839158e-01 -4.02068436e-01
-4.55883414e-01 4.62817222e-01 -2.10544780e-01 -1.95401788e-01
7.23225534e-01 -1.27297997e+00 4.85994846e-01 9.50867891e-01
2.42480546e-01 3.13793570e-01 1.16550922e+00 -7.28518248e-01
-1.10597086e+00 -1.08101892e+00 1.04880595e+00 -8.83281291e-01
4.24158335e-01 -2.75449723e-01 -4.88443017e-01 1.25341392e+00
3.28301966e-01 -1.72188595e-01 3.00228596e-01 2.77641118e-01
-7.25316107e-02 9.28188562e-02 -1.02904367e+00 8.92130077e-01
1.25783432e+00 -2.79956490e-01 -6.78363264e-01 4.82027292e-01
8.30013990e-01 -7.86260366e-01 -6.61281943e-01 2.24246487e-01
7.58801341e-01 -1.12248886e+00 7.85985768e-01 -5.73451877e-01
6.74091280e-01 -1.74817026e-01 -2.12097988e-01 -1.21585631e+00
-2.59297580e-01 -8.14672947e-01 -2.48961836e-01 1.10655475e+00
6.18891597e-01 -4.15207177e-01 7.87923098e-01 8.43979061e-01
-2.86544055e-01 -9.12031412e-01 -7.68578768e-01 -6.26985848e-01
-2.03029945e-01 -8.36267412e-01 3.36902946e-01 4.16920364e-01
-1.76962137e-01 2.75605589e-01 -8.31079304e-01 1.78654417e-01
5.61152220e-01 1.15466736e-01 7.89123535e-01 -7.97405005e-01
-3.33220094e-01 -4.66719232e-02 -8.02987218e-02 -1.07966185e+00
2.59793520e-01 -6.32751584e-01 3.04836720e-01 -1.67437255e+00
1.42226964e-01 -3.00695118e-03 -4.99298453e-01 8.44406188e-01
-1.66284755e-01 8.92859604e-03 5.36254585e-01 4.27322127e-02
-8.20469379e-01 7.23467946e-01 1.13097823e+00 1.50476113e-01
-5.13488472e-01 1.14244677e-01 -7.97200978e-01 8.93746495e-01
9.02623773e-01 -5.94016969e-01 -4.80675161e-01 -3.46126944e-01
2.13815942e-01 1.42671466e-01 6.13474965e-01 -8.17248464e-01
5.44369109e-02 -4.11383212e-01 3.05040300e-01 -7.68133581e-01
7.42651105e-01 -4.06074733e-01 2.24389955e-01 4.77217466e-01
-5.58309376e-01 -3.71638872e-02 2.45247900e-01 8.05833042e-01
1.61594748e-01 4.84265946e-02 6.85592890e-01 -1.35785311e-01
-1.26469934e+00 1.43739566e-01 -2.42116362e-01 2.18285784e-01
1.14249027e+00 -1.49309635e-01 -1.91860810e-01 -7.22460687e-01
-1.18771338e+00 6.21589780e-01 6.54845834e-02 6.85615063e-01
7.64415264e-01 -1.19885528e+00 -8.50247920e-01 -2.33927906e-01
1.36703476e-01 -1.65642828e-01 6.02176309e-01 1.01165783e+00
1.04852095e-01 3.81297022e-01 -2.05424026e-01 -6.31603658e-01
-1.49361193e+00 4.19365019e-01 7.56193176e-02 -4.32548910e-01
-5.05486488e-01 1.19610965e+00 3.41514289e-01 6.13383502e-02
2.58200675e-01 -2.84628421e-01 -2.46240497e-01 3.24142352e-02
4.84225631e-01 3.58125359e-01 -5.35822451e-01 -6.80768430e-01
-7.34354317e-01 3.93110693e-01 -2.92021841e-01 -2.57588148e-01
1.41468954e+00 -1.99439019e-01 4.00773436e-01 5.54777861e-01
7.92116761e-01 -3.75696331e-01 -1.76923323e+00 1.37742385e-01
-7.79286548e-02 -3.68524134e-01 -1.14739791e-01 -1.12821376e+00
-7.05982387e-01 8.19441199e-01 8.35706413e-01 -1.98238343e-01
1.15671921e+00 2.70371795e-01 3.91281426e-01 2.42839471e-01
1.56909809e-01 -1.11125624e+00 9.40845609e-02 6.34715438e-01
1.33636606e+00 -1.52606547e+00 -4.24162596e-02 -1.12784013e-01
-1.25100553e+00 6.63611174e-01 8.16151440e-01 -1.65990740e-01
3.32273334e-01 -1.58080101e-01 3.00962478e-02 -2.85320103e-01
-1.17401934e+00 -3.26809883e-01 3.76140058e-01 5.14876962e-01
3.46029520e-01 2.19637528e-01 -2.64071435e-01 6.39818251e-01
-2.24246427e-01 3.24341655e-02 4.23529476e-01 7.64082313e-01
-1.98076665e-01 -8.18179369e-01 -5.20807132e-02 2.87557602e-01
-4.03530985e-01 -2.71830529e-01 -2.34456643e-01 1.01673508e+00
7.15889409e-02 1.07747138e+00 -1.76757108e-02 -5.48536837e-01
3.71670276e-01 2.75391042e-01 4.08111870e-01 -6.04587793e-01
-4.86281872e-01 6.49417341e-02 3.95184427e-01 -9.92711961e-01
-5.29445529e-01 -9.66983259e-01 -1.41359425e+00 -1.21360376e-01
6.28668396e-03 -1.90225810e-01 3.12831163e-01 9.47250783e-01
4.26642120e-01 4.78617728e-01 2.95987844e-01 -7.54775167e-01
-5.00660896e-01 -9.36246276e-01 -3.20079744e-01 4.42895710e-01
2.35295907e-01 -8.51055026e-01 -4.85814899e-01 1.29371732e-01] | [7.972569465637207, 0.5973602533340454] |
02ef6d85-aac6-40af-8e54-4bafb47fc946 | convolutional-neural-network-hyperparameters | null | null | https://www.semanticscholar.org/paper/Convolutional-Neural-Network-Hyperparameters-for-Vulpe-Grigora%C5%9Fi-Grigore/fe344427eafecc60a1ba29beb87a46e91b7c1420#related-papers | https://ieeexplore.ieee.org/document/9425073 | Convolutional Neural Network Hyperparameters optimization for Facial Emotion Recognition | This paper presents a method of optimizing the hyperparameters of a convolutional neural network in order to increase accuracy in the context of facial emotion recognition. The optimal hyperparameters of the network were determined by generating and training models based on Random Search algorithm applied on a search space defined by discrete values of hyperparameters. The best model resulted was trained and evaluated using FER2013 database, obtaining an accuracy of 72.16%. | ['Ovidiu Grigore', 'Adrian Vulpe-Grigorași'] | 2021-03-25 | null | null | null | 12th-international-symposium-on-advanced | ['facial-emotion-recognition'] | ['computer-vision'] | [ 1.12446569e-01 2.91960686e-01 1.07843064e-01 -7.71536171e-01
-2.75370657e-01 -1.45621762e-01 2.46147424e-01 -4.88296688e-01
-8.85254741e-01 5.89145541e-01 -2.48197153e-01 2.32561320e-01
-2.21382841e-01 -5.03659964e-01 -4.80345428e-01 -7.05003202e-01
-2.70964831e-01 1.06984787e-01 -4.06156927e-01 -2.20825989e-02
3.17072242e-01 6.93139672e-01 -1.73996079e+00 4.65508491e-01
4.24599230e-01 1.43504143e+00 -2.13694423e-01 6.22733891e-01
1.06685810e-01 1.32197455e-01 -9.13194180e-01 -4.95601982e-01
3.94917488e-01 -1.51555404e-01 -7.57497251e-01 1.04445919e-01
1.20251598e-02 2.33105287e-01 6.88802972e-02 1.03766608e+00
5.75240016e-01 2.95635164e-01 6.65251374e-01 -1.23654521e+00
3.35910134e-02 2.90710837e-01 4.55329344e-02 1.75076723e-01
-1.39065251e-01 -2.52478868e-01 6.19243860e-01 -8.59020472e-01
4.38308269e-01 1.11736298e+00 3.42316568e-01 8.67955923e-01
-1.11346817e+00 -1.00967598e+00 -3.35701376e-01 3.31041723e-01
-1.99215984e+00 -5.58817029e-01 7.82247722e-01 -1.45207241e-01
1.36970413e+00 -1.50000472e-02 8.02168548e-01 9.73905146e-01
2.83991516e-01 4.11980212e-01 7.74651766e-01 -6.41178727e-01
4.93617147e-01 9.93499905e-02 -1.17799200e-01 7.53607333e-01
8.67116526e-02 9.09461081e-02 -5.03159225e-01 -4.23442781e-01
5.16883135e-01 -7.57298291e-01 2.85493617e-04 1.55923873e-01
-3.66243035e-01 1.08150220e+00 2.43367448e-01 3.84968668e-01
-9.26412046e-01 5.57180010e-02 6.11764967e-01 1.20306350e-01
1.88672006e-01 8.23645830e-01 -6.85863197e-01 -2.45176461e-02
-8.05155337e-01 -3.14116292e-02 6.98428988e-01 2.53916562e-01
6.62259281e-01 3.30990613e-01 6.78830892e-02 1.11988795e+00
5.68551123e-01 1.76680177e-01 7.30679452e-01 -9.57107663e-01
-8.65635276e-02 6.94042802e-01 -6.34989887e-02 -9.48675990e-01
-5.09486377e-01 -2.67311156e-01 -6.98785603e-01 2.96218246e-01
-1.21068619e-02 -7.50739217e-01 -1.10036051e+00 1.52476084e+00
3.48527044e-01 2.32784763e-01 4.05241072e-01 7.52237499e-01
9.07923996e-01 7.08178282e-01 4.73596543e-01 -2.20744163e-01
1.21234882e+00 -5.05111158e-01 -7.37978339e-01 1.49085283e-01
3.79069567e-01 -4.87234592e-01 4.99564499e-01 8.33783507e-01
-8.54824841e-01 -5.10478377e-01 -1.16074347e+00 7.88457990e-01
-5.69921196e-01 7.84013629e-01 6.24060154e-01 8.73061776e-01
-1.25282955e+00 3.97483557e-01 -5.71911156e-01 -1.11607942e-05
6.38669193e-01 9.27494705e-01 -5.67127228e-01 7.43357182e-01
-1.36128199e+00 8.24873149e-01 1.02719712e+00 6.88156009e-01
-6.33073807e-01 -2.47095317e-01 -5.95685184e-01 1.69526562e-01
-1.89792722e-01 -1.75260365e-01 9.60375965e-01 -1.72866642e+00
-1.98685908e+00 9.26480114e-01 4.43547703e-02 -6.45256519e-01
6.98846504e-02 2.08302140e-02 -6.00089550e-01 2.10891202e-01
-8.62701416e-01 9.44099307e-01 1.17938995e+00 -8.79290581e-01
-1.72884852e-01 -2.08866656e-01 -3.97423059e-01 2.11841360e-01
-5.90600491e-01 1.96735024e-01 -4.03733701e-01 -2.63660520e-01
-8.68363380e-02 -8.82022262e-01 -1.96612194e-01 -5.49923599e-01
-1.34024456e-01 -3.33267927e-01 6.39967918e-01 -4.05190140e-01
1.33038223e+00 -1.91579413e+00 -7.43547305e-02 8.12957466e-01
-3.00666630e-01 8.27737987e-01 -2.11152330e-01 -2.14277610e-01
-5.55723548e-01 1.97785392e-01 1.97280347e-01 3.02799344e-01
-1.59476653e-01 5.59489876e-02 2.55546689e-01 2.41742253e-01
3.82507235e-01 6.94956362e-01 -1.16830677e-01 -4.21725780e-01
1.66617334e-01 1.05494201e+00 -6.16937637e-01 3.95537436e-01
-1.47028774e-01 -1.99900702e-01 -5.37522018e-01 5.40753365e-01
5.83359957e-01 5.30631468e-02 2.52858520e-01 -4.02848452e-01
2.93447882e-01 -5.25503814e-01 -1.24520683e+00 9.62931812e-01
-4.41019475e-01 7.43909836e-01 -9.18404162e-02 -8.11767280e-01
1.51088083e+00 6.20987594e-01 4.87858593e-01 -3.57206583e-01
9.53592181e-01 -7.21429586e-02 -1.17422752e-02 -7.10947394e-01
2.38024563e-01 -9.16481391e-02 2.55422264e-01 1.50285974e-01
5.84152639e-01 2.59871274e-01 -1.74405538e-02 -6.17662847e-01
7.02042401e-01 -1.99999750e-01 1.64203018e-01 -2.63577819e-01
8.18821251e-01 -3.57124805e-01 2.67462701e-01 2.92746335e-01
-3.34584087e-01 3.10257286e-01 5.15786231e-01 -8.26761067e-01
-8.54150057e-01 -2.75103867e-01 -4.36682045e-01 9.36970770e-01
-5.86083353e-01 -1.33356169e-01 -1.38229966e+00 -6.08166575e-01
-4.77814764e-01 4.31464702e-01 -1.06605744e+00 -4.03743804e-01
-4.29979086e-01 -1.13636672e+00 8.41731429e-01 2.78326795e-02
6.92874789e-01 -1.61515498e+00 -1.06395209e+00 -1.37847615e-02
1.45633193e-02 -1.00540507e+00 1.35324478e-01 3.65603954e-01
-8.14268410e-01 -1.13686502e+00 -4.94142324e-01 -7.49926507e-01
8.69634211e-01 -8.58354807e-01 9.62983727e-01 2.49706969e-01
-5.38881302e-01 2.02971348e-03 -1.93017289e-01 -4.86671716e-01
-2.67558366e-01 2.42523238e-01 -1.84399337e-01 3.98340285e-01
5.61569393e-01 -2.55456399e-02 -5.47321975e-01 1.55310825e-01
-1.04037869e+00 -3.23933065e-01 5.92176795e-01 9.74376559e-01
4.40827131e-01 1.84218481e-01 5.41431487e-01 -5.93249559e-01
9.42501307e-01 -3.57237339e-01 -8.34762037e-01 3.89929920e-01
-7.09887862e-01 2.31552929e-01 3.75592232e-01 -6.39794350e-01
-1.08068466e+00 5.23263574e-01 -2.98693597e-01 -6.73588157e-01
-3.74098659e-01 2.74021655e-01 -5.04369438e-02 -4.42186415e-01
8.27933431e-01 -1.31929219e-01 -2.73549818e-02 2.97085084e-02
-6.48952052e-02 5.94546258e-01 2.02463627e-01 -4.19322968e-01
-1.78105280e-01 -2.14875072e-01 5.57632558e-02 -8.63915503e-01
-3.29757959e-01 1.02025624e-02 -4.90234375e-01 -7.56399870e-01
7.52353191e-01 -5.09510696e-01 -1.06941974e+00 6.52323842e-01
-9.77505147e-01 -1.17722847e-01 1.36957794e-01 5.42042494e-01
-3.92157167e-01 -4.35580939e-01 -7.45997578e-02 -8.83641303e-01
-7.89193034e-01 -1.17693448e+00 7.77968466e-01 6.43191755e-01
-3.73253554e-01 -8.54164779e-01 -1.06840558e-01 -4.21923697e-02
6.04944229e-01 2.86156118e-01 6.97602928e-01 -8.96517754e-01
2.28484720e-01 -4.79233831e-01 -2.09310185e-03 5.37351549e-01
-2.49821052e-01 4.31102395e-01 -1.19000924e+00 1.92635611e-01
-2.46022388e-01 -4.79896009e-01 3.95387262e-01 6.27697229e-01
1.53468812e+00 -3.43180507e-01 -1.92601867e-02 8.17178547e-01
1.59238052e+00 4.67458040e-01 9.34650004e-01 5.26474714e-01
-1.46727994e-01 3.37532520e-01 3.65919113e-01 8.12825382e-01
-2.32579842e-01 3.96163493e-01 6.18906200e-01 3.00831676e-01
3.87749046e-01 2.03414217e-01 -2.35073045e-01 -2.62816697e-01
-1.16847120e-01 -2.30500519e-01 -9.90417361e-01 1.70404881e-01
-1.43861401e+00 -7.90297806e-01 6.03945971e-01 1.83898783e+00
7.02325523e-01 -1.05116405e-01 1.82747275e-01 1.68270275e-01
8.59054744e-01 -8.98864642e-02 -3.79609466e-01 -9.58230078e-01
-1.19350694e-01 7.47129083e-01 3.70828956e-01 4.50873464e-01
-1.11767614e+00 1.24028754e+00 7.78467512e+00 6.03304446e-01
-1.60389960e+00 -4.88209486e-01 1.03954160e+00 -2.85232872e-01
3.91888082e-01 -5.78838825e-01 -9.90370691e-01 2.69993007e-01
1.26334441e+00 -2.69669779e-02 4.61093038e-01 8.85873139e-01
1.95056558e-01 -5.19617647e-02 -3.25961232e-01 1.29955065e+00
6.09023608e-02 -1.26201534e+00 1.14364401e-01 -1.79370284e-01
7.51419961e-01 -1.88409284e-01 1.47224039e-01 1.77353472e-01
4.05372940e-02 -1.55114210e+00 3.20666581e-01 8.29794466e-01
3.99908006e-01 -1.23182511e+00 9.63689089e-01 -1.61673039e-01
-5.68269789e-01 -1.95114493e-01 -3.76110077e-01 4.00698215e-01
-5.19080043e-01 1.41602412e-01 -1.10808384e+00 -2.35871926e-01
7.78743267e-01 -8.91623497e-02 -4.51035142e-01 1.12405372e+00
7.71940574e-02 8.01404357e-01 -5.76407909e-01 -6.64114177e-01
3.05318326e-01 -2.96670571e-02 1.38178363e-01 1.41695344e+00
3.06913495e-01 2.93011278e-01 -6.49934351e-01 4.96501118e-01
-2.13049054e-01 4.42997217e-01 -2.72720397e-01 -1.11606680e-01
4.66282666e-01 1.50889158e+00 -7.38198996e-01 3.31868269e-02
3.87349695e-01 5.82210839e-01 3.66342664e-01 5.58778644e-01
-9.28359330e-01 -5.05769908e-01 3.73771369e-01 -2.66002983e-01
4.45969164e-01 9.30932760e-02 -1.42061993e-01 -5.49084723e-01
-1.72236785e-01 -8.42166305e-01 4.97583628e-01 -6.45866275e-01
-6.67592943e-01 1.47500300e+00 1.53000772e-01 -7.30991840e-01
-5.96403122e-01 -8.35723519e-01 -5.44096828e-01 8.87592435e-01
-1.01374984e+00 -6.65766835e-01 -3.30105335e-01 5.01328230e-01
2.36580193e-01 -7.61817753e-01 1.15172315e+00 -8.50867406e-02
-7.09813535e-01 1.01835704e+00 -3.22625339e-01 2.18285564e-02
2.59041548e-01 -5.49620211e-01 -3.25443685e-01 2.37704471e-01
-1.76277831e-01 3.32985193e-01 7.79656112e-01 -9.69035402e-02
-1.08004332e+00 -9.39987004e-01 7.82153845e-01 4.03820187e-01
4.89565618e-02 -4.06754091e-02 -6.89941883e-01 8.60138908e-02
3.36908191e-01 8.66040140e-02 8.62417758e-01 3.50387916e-02
-1.32888824e-01 -4.01494652e-02 -1.62183440e+00 3.67278397e-01
3.36967617e-01 -3.19223255e-02 1.66148975e-01 1.12671349e-02
-1.49216264e-01 -3.32619041e-01 -1.00055325e+00 8.34783196e-01
7.93299437e-01 -6.66895688e-01 7.31118441e-01 -7.35707164e-01
1.07471816e-01 2.01207995e-01 -1.55192077e-01 -1.35389507e+00
-8.82109161e-03 -7.63662636e-01 -5.44650070e-02 7.93524683e-01
7.04450369e-01 -4.85729665e-01 1.09007299e+00 9.30909634e-01
5.42637706e-01 -1.38793015e+00 -9.87611949e-01 -1.28825128e-01
-2.82168478e-01 -3.78231734e-01 7.48168230e-01 5.55557966e-01
-2.72269845e-01 1.09067075e-01 7.08031803e-02 -7.38775507e-02
2.31052265e-01 -5.67469001e-01 2.18533501e-01 -1.04901576e+00
3.70638400e-01 -7.33881891e-01 -6.86765671e-01 5.89388795e-02
7.46990085e-01 -4.84114021e-01 1.03379495e-01 -6.69329464e-01
-1.30573109e-01 -2.95253694e-01 -4.42931443e-01 8.26473117e-01
2.81211793e-01 4.96670634e-01 -1.07973583e-01 -4.26074475e-01
-4.63787764e-01 5.73003769e-01 6.07871711e-01 2.72128373e-01
-3.31231594e-01 1.75018117e-01 -4.93809998e-01 7.28245080e-01
1.25228727e+00 -2.79556423e-01 -3.62456203e-01 -5.06269671e-02
2.60671467e-01 -9.12662819e-02 1.30276918e-01 -9.21056390e-01
2.18226463e-02 -8.52378383e-02 9.46472883e-01 -2.27750257e-01
4.65916365e-01 -8.98857296e-01 3.71698350e-01 7.59932548e-02
-6.64371610e-01 -1.07749917e-01 6.29763126e-01 -2.91778799e-03
-3.24070752e-01 -7.06753194e-01 1.16514397e+00 2.16338128e-01
-8.63935947e-01 2.91505843e-01 -4.40793782e-01 -3.70943874e-01
1.16380727e+00 -2.77314007e-01 3.40168715e-01 -2.64603823e-01
-1.12122762e+00 -1.40905634e-01 -2.13452667e-01 5.47911763e-01
8.69927704e-01 -1.33851206e+00 -4.49079931e-01 7.19179988e-01
-9.20947045e-02 -7.04150677e-01 1.55866235e-01 2.09774682e-03
-5.81179738e-01 4.40164804e-01 -6.12145722e-01 -3.61073405e-01
-1.68569064e+00 -2.43623704e-01 1.22293413e+00 -4.56365570e-02
4.55917902e-02 1.02394235e+00 -5.25891840e-01 -3.25259387e-01
5.30918717e-01 2.65465736e-01 -8.95096123e-01 1.12779059e-01
5.35033405e-01 2.52971560e-01 4.38435197e-01 -8.92064154e-01
-3.82337779e-01 5.02629876e-01 2.81399995e-01 -2.91231722e-01
1.41288233e+00 4.19389158e-01 -5.04684588e-03 -3.38978946e-01
1.64459443e+00 -9.48342919e-01 -1.06810737e+00 1.71680793e-01
-1.21467009e-01 -2.65507013e-01 4.83634859e-01 -1.08121085e+00
-1.33527970e+00 3.09528857e-01 1.12539530e+00 -1.50019303e-01
1.61498249e+00 -3.53684008e-01 2.22280532e-01 6.98104501e-01
-9.16276574e-02 -1.45421493e+00 -6.78048134e-02 7.27192402e-01
9.31343913e-01 -1.20528030e+00 -2.14140460e-01 9.51158851e-02
-8.23241353e-01 1.65159631e+00 7.92734504e-01 -3.82884264e-01
1.22961402e+00 3.62220377e-01 2.42892057e-01 -4.25795525e-01
-1.05401516e+00 1.30793065e-01 6.55065596e-01 4.60121781e-01
5.91855109e-01 -5.55454865e-02 -2.62601912e-01 6.28946662e-01
-2.60928392e-01 3.55385602e-01 -9.01541337e-02 5.87806344e-01
-4.03170317e-01 -7.56106853e-01 -3.24974924e-01 3.51437598e-01
-8.03408265e-01 1.37191385e-01 -4.08676863e-01 6.20450556e-01
2.01048508e-01 8.67891550e-01 1.12535968e-01 -5.85025191e-01
1.80443674e-01 4.32153046e-01 5.06460190e-01 2.53372416e-02
-7.97938883e-01 -1.26157150e-01 7.82861337e-02 -4.24557239e-01
-2.81823158e-01 -4.91196632e-01 -1.35104048e+00 1.79866821e-01
-5.07397354e-01 4.85562593e-01 1.29198384e+00 8.75461876e-01
5.11230648e-01 2.60981798e-01 8.85923147e-01 -7.02377141e-01
-3.11763465e-01 -1.07519770e+00 -3.17071408e-01 1.21840097e-01
-2.20912263e-01 -5.61889470e-01 -3.23916435e-01 -4.53597009e-02] | [13.52357292175293, 1.7722920179367065] |
04f12b82-9601-4602-a8c1-ea7d220c6ff6 | horizonnet-learning-room-layout-with-1d | 1901.03861 | null | http://arxiv.org/abs/1901.03861v2 | http://arxiv.org/pdf/1901.03861v2.pdf | HorizonNet: Learning Room Layout with 1D Representation and Pano Stretch Data Augmentation | We present a new approach to the problem of estimating the 3D room layout
from a single panoramic image. We represent room layout as three 1D vectors
that encode, at each image column, the boundary positions of floor-wall and
ceiling-wall, and the existence of wall-wall boundary. The proposed network,
HorizonNet, trained for predicting 1D layout, outperforms previous
state-of-the-art approaches. The designed post-processing procedure for
recovering 3D room layouts from 1D predictions can automatically infer the room
shape with low computation cost - it takes less than 20ms for a panorama image
while prior works might need dozens of seconds. We also propose Pano Stretch
Data Augmentation, which can diversify panorama data and be applied to other
panorama-related learning tasks. Due to the limited data available for
non-cuboid layout, we relabel 65 general layout from the current dataset for
finetuning. Our approach shows good performance on general layouts by
qualitative results and cross-validation. | ['Hwann-Tzong Chen', 'Chi-Wei Hsiao', 'Min Sun', 'Cheng Sun'] | 2019-01-12 | horizonnet-learning-room-layout-with-1d-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Sun_HorizonNet_Learning_Room_Layout_With_1D_Representation_and_Pano_Stretch_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Sun_HorizonNet_Learning_Room_Layout_With_1D_Representation_and_Pano_Stretch_CVPR_2019_paper.pdf | cvpr-2019-6 | ['3d-room-layouts-from-a-single-rgb-panorama'] | ['computer-vision'] | [ 4.61046815e-01 1.68330848e-01 5.80763482e-02 -4.81795043e-01
-6.65060103e-01 -6.79402053e-01 6.47494912e-01 2.99035132e-01
-2.51319468e-01 3.54391694e-01 5.35897791e-01 -5.43086946e-01
-2.30312437e-01 -8.25783908e-01 -9.96357679e-01 -4.20641035e-01
-1.41417369e-01 4.88795400e-01 -6.75460398e-02 -1.34494543e-01
3.94022942e-01 7.67626524e-01 -1.50313389e+00 3.37356150e-01
6.65438533e-01 1.20181262e+00 6.21658921e-01 1.07017028e+00
-1.07355500e-02 5.85809350e-01 -3.87385488e-01 4.66503978e-01
4.43177670e-01 1.04744025e-02 -7.38674879e-01 4.45443511e-01
1.29539907e+00 -3.58488947e-01 -4.79225397e-01 4.42730725e-01
3.82398218e-01 2.99706638e-01 6.93993628e-01 -7.80218303e-01
-3.69133592e-01 2.01127782e-01 -2.74096459e-01 -1.67707905e-01
4.28598255e-01 -3.13466966e-01 1.01040769e+00 -7.89609015e-01
5.52184582e-01 1.17968452e+00 6.90749288e-01 1.39898703e-01
-1.12285244e+00 -1.01186492e-01 4.69047576e-01 -1.91318512e-01
-1.13208067e+00 -3.11159819e-01 8.52505803e-01 -4.04723972e-01
9.20218945e-01 6.43035173e-01 5.59940517e-01 9.57686305e-01
1.45694926e-01 8.35560381e-01 1.30848742e+00 -6.47998512e-01
2.02055082e-01 -1.20482251e-01 -1.72714084e-01 1.11580265e+00
-1.11128613e-01 -3.29752326e-01 -1.54511988e-01 2.73472309e-01
9.30339396e-01 1.60605386e-01 -3.94526064e-01 -1.09065866e+00
-1.31048274e+00 4.01835829e-01 8.55334640e-01 1.49936259e-01
3.68425697e-02 -4.01012264e-02 1.75740913e-01 4.36868183e-02
6.22597486e-02 7.77855039e-01 -5.44205844e-01 8.05547312e-02
-1.12562478e+00 3.72203708e-01 6.51460052e-01 1.09414232e+00
8.17878127e-01 -3.25209916e-01 -1.16348535e-01 7.83600092e-01
1.07850984e-01 4.89917845e-01 1.28965989e-01 -9.96643364e-01
9.77363348e-01 6.57899320e-01 2.66774446e-01 -8.78542840e-01
-8.79421473e-01 1.39402196e-01 -9.22131181e-01 8.62117931e-02
5.63159645e-01 1.14690810e-01 -1.42824793e+00 1.16577911e+00
1.83284298e-01 -1.75399572e-01 -2.29801953e-01 6.95823073e-01
9.14451063e-01 7.19582021e-01 -6.18116617e-01 2.07290858e-01
1.15695357e+00 -1.54762638e+00 -3.33623171e-01 -6.56521678e-01
3.61194372e-01 -9.24223125e-01 1.43754554e+00 3.76630247e-01
-8.00468862e-01 -7.36572504e-01 -1.24899805e+00 -3.26511651e-01
-5.46791732e-01 5.65189183e-01 5.80864370e-01 6.12893283e-01
-9.72514629e-01 4.69611168e-01 -6.40373826e-01 -5.60631275e-01
2.10527211e-01 3.77986133e-01 -4.98137146e-01 -4.47663844e-01
-3.38339180e-01 8.38121593e-01 1.69173688e-01 3.94737959e-01
-9.14815545e-01 -5.10064781e-01 -1.41336513e+00 1.50769189e-01
4.76600409e-01 -4.36934918e-01 1.12789857e+00 -2.68454105e-01
-1.32107639e+00 7.03155100e-01 -1.72412589e-01 -1.71898961e-01
3.90211225e-01 -3.51237327e-01 -2.27743760e-01 -2.14015424e-01
3.16456780e-02 7.76643693e-01 5.80767691e-01 -1.79916716e+00
-3.81218076e-01 -4.40239340e-01 1.03024893e-01 4.56541508e-01
-9.01602060e-02 -9.05759215e-01 -6.55631900e-01 -2.50680506e-01
6.13366902e-01 -9.79365468e-01 -5.70043683e-01 -1.02537647e-01
-9.28492725e-01 2.81791449e-01 7.73583770e-01 -6.57127857e-01
1.07323360e+00 -1.99146199e+00 -4.84708436e-02 4.97508883e-01
-1.55014679e-01 7.23066973e-03 -1.16942465e-01 3.93087894e-01
7.13560656e-02 -1.69965073e-01 -4.66492116e-01 -5.17592013e-01
7.91510791e-02 1.76951990e-01 -3.83361965e-01 3.12718183e-01
-2.52821237e-01 7.41354287e-01 -6.17429197e-01 -2.90069878e-01
8.20474327e-01 3.48597556e-01 -3.22393030e-01 4.24783200e-01
-1.74260482e-01 3.05856347e-01 -1.86620250e-01 6.35278404e-01
9.10688937e-01 -1.81869715e-01 4.40184712e-01 -1.30811408e-01
-3.88418913e-01 4.50559437e-01 -1.21265686e+00 2.21682048e+00
-1.00218558e+00 8.86971354e-01 -2.53029764e-02 -6.92260623e-01
1.15994179e+00 -1.72039106e-01 1.40841290e-01 -7.70980358e-01
-6.43655434e-02 5.08417375e-02 -5.00057697e-01 -5.08368134e-01
1.02320242e+00 5.17975450e-01 -5.02187729e-01 2.19275504e-01
-2.55143076e-01 -7.01215744e-01 7.16617331e-02 -2.53504962e-01
9.32066023e-01 3.77231896e-01 3.44908088e-01 -2.46574521e-01
4.94938314e-01 -8.41265991e-02 -7.76381642e-02 8.15557182e-01
1.89696968e-01 1.03674233e+00 2.95293987e-01 -9.56906080e-01
-1.36496472e+00 -1.15095401e+00 -3.56814452e-02 1.09047973e+00
2.64819056e-01 -6.00325823e-01 -7.32755840e-01 -5.80593050e-01
-1.97867736e-01 5.34269452e-01 -8.57497036e-01 5.29676199e-01
-9.17094171e-01 -4.27089810e-01 5.04227728e-02 7.63556540e-01
4.69631791e-01 -9.50782716e-01 -9.63572800e-01 -1.29774749e-01
-2.88671315e-01 -1.30206311e+00 -4.95509952e-01 7.00429022e-01
-7.92507529e-01 -1.11336434e+00 -7.12641835e-01 -1.12638044e+00
9.53325152e-01 3.69604111e-01 1.10345769e+00 -1.67220652e-01
-3.79137754e-01 3.76222640e-01 6.87444508e-02 -1.01296492e-01
-1.81819685e-03 2.62265354e-01 -2.82624066e-01 -5.47529280e-01
-2.70445734e-01 -4.71033722e-01 -6.83142364e-01 2.18047932e-01
-5.02790332e-01 4.06184882e-01 5.85684180e-01 5.25125980e-01
8.18194866e-01 -7.66442940e-02 -3.67468327e-01 -8.95811260e-01
3.96586210e-01 1.64329454e-01 -8.26074243e-01 3.91049027e-01
-4.91074115e-01 2.50300676e-01 8.64784479e-01 -7.52913058e-02
-1.01438022e+00 4.66912866e-01 -1.07154667e-01 -2.61854202e-01
-6.85705364e-01 -7.94500783e-02 -3.54323268e-01 1.90021813e-01
5.17238557e-01 -4.75719757e-02 -5.45447350e-01 -5.17026365e-01
5.58362782e-01 3.94429415e-01 6.78215563e-01 -4.89363849e-01
7.50445127e-01 5.58843791e-01 2.06015185e-01 -1.04872537e+00
-1.19943285e+00 -6.25865281e-01 -1.16799235e+00 1.13145858e-02
9.76255894e-01 -8.42162669e-01 -5.97227693e-01 1.16423130e-01
-1.04855072e+00 -8.74589443e-01 -9.40360650e-02 1.33456439e-01
-6.70703948e-01 1.21783867e-01 -2.13190392e-01 -7.97012627e-01
-1.54610321e-01 -9.87121344e-01 1.40934598e+00 1.42250612e-01
-2.10879982e-01 -8.79420042e-01 2.08599865e-01 5.36410332e-01
3.97037491e-02 3.60379457e-01 1.09213316e+00 3.20546292e-02
-1.03328681e+00 -1.46310180e-01 -1.78130418e-01 -7.19317468e-03
1.56990662e-01 -2.71204859e-01 -1.39568663e+00 -2.93581095e-02
-5.76606631e-01 -3.20495665e-01 9.89312887e-01 6.63494945e-01
1.70927060e+00 -4.25590664e-01 -4.73199546e-01 7.23785520e-01
1.45170689e+00 -1.02434218e-01 5.72364569e-01 7.64267325e-01
1.15034509e+00 7.02700496e-01 7.79171586e-01 3.33100289e-01
6.62291646e-01 5.24158478e-01 7.47847021e-01 -2.22106338e-01
6.79820823e-03 -6.88418329e-01 2.77759898e-02 6.51115179e-01
-4.50608619e-02 -5.62984169e-01 -1.01610506e+00 5.77612460e-01
-1.81731200e+00 -6.06981575e-01 4.10492755e-02 2.35586143e+00
1.13807999e-01 2.62618154e-01 -2.04632714e-01 9.65056792e-02
9.98126641e-02 7.99011230e-01 -2.75051743e-01 -7.19747841e-01
6.87605813e-02 1.78817660e-02 8.30043018e-01 8.69431496e-01
-1.50002420e+00 9.25148249e-01 6.50455523e+00 2.89598137e-01
-8.91101480e-01 -4.21202600e-01 8.39063168e-01 8.74827728e-02
-3.29342544e-01 -2.02245519e-01 -7.44855762e-01 -1.19149901e-01
5.97172916e-01 8.42721701e-01 7.39337265e-01 1.14172995e+00
9.87607017e-02 -3.81413847e-01 -1.21118748e+00 1.13058269e+00
3.88057798e-01 -1.54326332e+00 -3.17304671e-01 1.91907898e-01
1.02969503e+00 -4.67756577e-02 6.92640096e-02 1.11351013e-01
1.18738614e-01 -1.30017734e+00 6.44673407e-01 3.68454337e-01
9.04809356e-01 -7.00185776e-01 5.57318449e-01 4.38208848e-01
-1.58202970e+00 -3.94534409e-01 -3.23899984e-01 -1.28958121e-01
6.49711564e-02 1.98851317e-01 -1.26906943e+00 4.41786349e-01
1.08921146e+00 3.31766933e-01 -8.42647195e-01 8.73855591e-01
-3.10685873e-01 -3.76866534e-02 -4.21746612e-01 9.69883725e-02
5.56559801e-01 -1.37268752e-01 -1.89616363e-02 1.13452554e+00
4.57884789e-01 -5.16290724e-01 3.65142167e-01 5.63627243e-01
-4.27843630e-02 -1.99131355e-01 -9.66624081e-01 4.92215008e-01
4.15316314e-01 1.35474265e+00 -1.05906844e+00 -2.14659929e-01
-6.93108812e-02 1.04007053e+00 4.72817928e-01 3.32765996e-01
-4.61929440e-01 -3.95945996e-01 1.39659539e-01 2.30458617e-01
7.28057563e-01 -6.22192681e-01 -6.83756888e-01 -6.78411007e-01
5.13164178e-02 -5.97690403e-01 5.50978957e-03 -8.96053970e-01
-6.48822486e-01 5.72033763e-01 -2.40208451e-02 -9.60047781e-01
-1.05846189e-02 -1.03126657e+00 -6.93778872e-01 3.61558646e-01
-1.45120370e+00 -1.36566615e+00 -8.46593440e-01 2.12587461e-01
7.96733677e-01 2.22914293e-01 1.15862381e+00 -9.23935547e-02
-3.29175025e-01 3.31534505e-01 2.31347814e-01 1.28967658e-01
6.28881156e-01 -1.67078745e+00 7.28061140e-01 7.08136201e-01
3.20092857e-01 3.84133518e-01 7.02569902e-01 -2.50032276e-01
-1.13615191e+00 -1.16310322e+00 1.03286314e+00 -7.31275260e-01
9.89762247e-02 -1.06249034e+00 -5.06621540e-01 7.58675873e-01
1.54830873e-01 -7.08935857e-02 4.02006775e-01 4.70570177e-01
-1.69851169e-01 -1.88513592e-01 -7.36011028e-01 8.09360504e-01
1.14954019e+00 -4.33669329e-01 -3.92997980e-01 3.03099722e-01
5.44361234e-01 -7.96223402e-01 -7.14780509e-01 3.06411892e-01
5.81672549e-01 -1.18391609e+00 1.26292157e+00 1.17353588e-01
5.25381088e-01 -2.52021819e-01 -4.40076739e-01 -1.16718578e+00
-1.44442484e-01 -4.54908639e-01 -9.06957984e-02 7.15042889e-01
4.54823822e-01 2.48943746e-01 9.71784830e-01 4.83751625e-01
-4.68947351e-01 -8.33635926e-01 -5.95742464e-01 -3.29639167e-01
-3.22030842e-01 -4.48077947e-01 7.11681068e-01 4.93403435e-01
-3.47045690e-01 4.74065512e-01 -5.35804868e-01 3.37104201e-01
3.54662031e-01 7.23619461e-01 1.21646965e+00 -1.06581950e+00
8.84500816e-02 -2.47871041e-01 3.96436006e-02 -1.57606781e+00
1.24152623e-01 -6.54524088e-01 5.58770895e-01 -2.09552193e+00
-1.07997477e-01 -6.04521275e-01 2.54509528e-03 3.71894658e-01
2.49413848e-01 1.66849151e-01 9.18203071e-02 -3.68884534e-01
-6.97068572e-01 6.46797478e-01 1.62888348e+00 -2.11404651e-01
-5.10650098e-01 -1.29400538e-02 -1.77613631e-01 9.25857484e-01
7.18307912e-01 -8.32850765e-03 -5.29159606e-01 -5.47152400e-01
9.95345041e-02 -1.09762914e-01 2.30098531e-01 -1.23644328e+00
-5.33406623e-03 -1.06131233e-01 8.92008662e-01 -1.19233143e+00
7.99284518e-01 -9.83801603e-01 -5.58662653e-01 9.73748341e-02
-3.70476991e-01 2.55856991e-01 3.80325645e-01 6.64688528e-01
1.32570118e-01 -4.52214256e-02 2.56993741e-01 -4.32075977e-01
-8.50545764e-01 2.71746904e-01 -4.04245019e-01 -1.92959994e-01
6.64074659e-01 -5.03754735e-01 -1.08508416e-01 -1.97320700e-01
-5.66618502e-01 8.54037106e-02 7.20600486e-01 6.56746924e-01
7.86735654e-01 -1.12599969e+00 1.48326576e-01 6.04359329e-01
1.28969967e-01 7.32336223e-01 3.83628666e-01 3.85890752e-01
-8.48130226e-01 7.84282267e-01 -3.04641426e-01 -7.40446568e-01
-1.34984422e+00 5.75885952e-01 2.62978554e-01 -4.61063862e-01
-8.63403022e-01 7.54067063e-01 4.20102268e-01 -1.23546505e+00
4.28133190e-01 -9.59623814e-01 -2.59655088e-01 -1.35981560e-01
1.93093345e-01 2.88979441e-01 1.15986012e-01 -5.33113122e-01
-2.98618734e-01 8.34447861e-01 1.94908515e-01 -7.78698698e-02
1.46918392e+00 -2.86481410e-01 4.72196080e-02 8.61044526e-01
1.08438909e+00 1.07486397e-01 -1.83751893e+00 1.53261572e-01
-5.81169501e-03 -5.98266661e-01 -1.03660725e-01 -7.87047923e-01
-3.50406915e-01 1.18818212e+00 8.75075340e-01 -2.00068444e-01
7.76969373e-01 -8.72740597e-02 5.22590935e-01 7.99238086e-01
1.97366953e-01 -1.25222683e+00 3.22645247e-01 6.42317295e-01
8.43127728e-01 -1.35793781e+00 2.44684890e-01 -2.20462352e-01
-5.03111064e-01 1.34260464e+00 7.99780488e-01 -2.49295998e-02
5.00008404e-01 1.91641971e-01 1.24651708e-01 -1.32177129e-01
-2.37157792e-01 1.09356619e-01 7.14734137e-01 5.32244802e-01
3.92758936e-01 4.11348879e-01 8.11456084e-01 -8.74793604e-02
-7.02353835e-01 -6.96841121e-01 4.12053436e-01 9.18530464e-01
-5.53972960e-01 -8.63722682e-01 -6.49370909e-01 3.14460874e-01
2.15793908e-01 -7.83331618e-02 -2.13165477e-01 9.65516686e-01
2.78058678e-01 6.62654102e-01 5.38189292e-01 -1.71234682e-01
4.37610924e-01 -2.16825604e-01 8.38342249e-01 -4.64498758e-01
-7.39688054e-02 2.15739146e-01 2.42658690e-01 -5.73067486e-01
-3.34087402e-01 -4.73184764e-01 -9.88380373e-01 -2.44415388e-01
-7.32826162e-03 -1.40831843e-01 7.09517241e-01 8.17701340e-01
9.95317772e-02 6.87394023e-01 4.93488610e-01 -1.44077814e+00
9.86480862e-02 -8.73342872e-01 -3.22558939e-01 9.12897512e-02
6.64129496e-01 -3.62513155e-01 1.15336083e-01 -4.95829694e-02] | [8.712353706359863, -2.823868751525879] |
2141143e-f91c-49c6-971c-9fd8620dcf93 | stdan-deformable-attention-network-for-space | 2203.06841 | null | https://arxiv.org/abs/2203.06841v2 | https://arxiv.org/pdf/2203.06841v2.pdf | STDAN: Deformable Attention Network for Space-Time Video Super-Resolution | The target of space-time video super-resolution (STVSR) is to increase the spatial-temporal resolution of low-resolution (LR) and low frame rate (LFR) videos. Recent approaches based on deep learning have made significant improvements, but most of them only use two adjacent frames, that is, short-term features, to synthesize the missing frame embedding, which cannot fully explore the information flow of consecutive input LR frames. In addition, existing STVSR models hardly exploit the temporal contexts explicitly to assist high-resolution (HR) frame reconstruction. To address these issues, in this paper, we propose a deformable attention network called STDAN for STVSR. First, we devise a long-short term feature interpolation (LSTFI) module, which is capable of excavating abundant content from more neighboring input frames for the interpolation process through a bidirectional RNN structure. Second, we put forward a spatial-temporal deformable feature aggregation (STDFA) module, in which spatial and temporal contexts in dynamic video frames are adaptively captured and aggregated to enhance SR reconstruction. Experimental results on several datasets demonstrate that our approach outperforms state-of-the-art STVSR methods. The code is available at https://github.com/littlewhitesea/STDAN. | ['Qingmin Liao', 'Wenming Yang', 'Yapeng Tian', 'Xiaoyu Xiang', 'Hai Wang'] | 2022-03-14 | null | null | null | null | ['space-time-video-super-resolution', 'video-super-resolution'] | ['computer-vision', 'computer-vision'] | [ 1.64195925e-01 -1.91149175e-01 -3.14683914e-01 -2.99551308e-01
-8.01749170e-01 -9.53377262e-02 4.22437876e-01 -7.58506000e-01
-2.78694272e-01 8.15496981e-01 6.59467816e-01 -2.84911357e-02
5.47031425e-02 -6.85788214e-01 -9.08572912e-01 -6.34082913e-01
2.20638558e-01 -3.39108348e-01 3.87920558e-01 -2.18509808e-01
1.04773030e-01 2.77992070e-01 -1.46522892e+00 6.22709155e-01
7.65048623e-01 8.45848382e-01 7.35933185e-01 5.76287091e-01
-1.43241912e-01 1.06318831e+00 -1.28578901e-01 -2.75596650e-03
1.88495323e-01 -6.04823768e-01 -6.32709980e-01 -1.22321077e-01
2.55238712e-01 -8.39180827e-01 -9.24090385e-01 8.54377449e-01
4.54220295e-01 4.37656909e-01 4.75637754e-03 -7.67576575e-01
-1.16885960e+00 8.70893300e-01 -6.64060295e-01 6.86665595e-01
4.96569365e-01 4.03408468e-01 8.01234126e-01 -1.25295603e+00
8.50301564e-01 1.47447824e+00 3.53347540e-01 7.66645014e-01
-9.83935118e-01 -7.92793930e-01 3.99641275e-01 3.97665799e-01
-1.44285238e+00 -6.67771757e-01 8.58533084e-01 -2.42484182e-01
6.86054826e-01 2.01303869e-01 5.38880527e-01 1.28265250e+00
1.83129348e-02 8.65593791e-01 8.27320158e-01 2.38722470e-02
-2.35920846e-01 -3.59395742e-01 -2.79780805e-01 4.53369260e-01
-2.60579497e-01 2.89325058e-01 -6.45651579e-01 3.75738949e-01
1.56484544e+00 3.12697083e-01 -6.96806729e-01 1.48929015e-01
-1.48992181e+00 6.20110393e-01 5.18059433e-01 6.82213485e-01
-3.48165005e-01 1.49403423e-01 2.01530308e-01 1.58128336e-01
6.30822241e-01 8.14439431e-02 -3.94310832e-01 -8.18025786e-03
-8.68629754e-01 -5.26878331e-03 -4.41025123e-02 8.02499712e-01
7.69811094e-01 3.93187344e-01 -4.93560493e-01 8.39510322e-01
3.25248092e-01 2.73585290e-01 6.45009220e-01 -1.36338401e+00
6.29740000e-01 2.30116308e-01 2.49624208e-01 -9.60259557e-01
-5.99628082e-04 -2.59572238e-01 -9.76447880e-01 -2.34950110e-02
8.72786269e-02 5.04253656e-02 -8.70672584e-01 1.55478835e+00
3.06329608e-01 7.97009230e-01 1.26412004e-01 1.45270932e+00
1.15391338e+00 1.07599926e+00 1.06133791e-02 -5.11443198e-01
1.03179657e+00 -1.06897581e+00 -9.45155621e-01 2.94584744e-02
1.55941099e-01 -6.08504534e-01 1.09472132e+00 1.13704368e-01
-1.46093929e+00 -9.66567874e-01 -8.18966091e-01 -6.16985321e-01
7.74361640e-02 -4.57336707e-03 2.38658473e-01 -1.78364515e-01
-1.07360446e+00 6.76954806e-01 -8.54604959e-01 2.82384336e-01
5.39951622e-01 1.37858197e-01 -3.64434212e-01 -3.33790958e-01
-1.48218215e+00 4.78670955e-01 2.99783915e-01 3.78040284e-01
-1.00930154e+00 -8.00314426e-01 -1.02434003e+00 2.06855070e-02
5.34595072e-01 -6.67203903e-01 9.98967350e-01 -1.11111498e+00
-1.61749470e+00 3.73868674e-01 -4.94864553e-01 -4.03528214e-01
5.13004184e-01 -3.05188417e-01 -5.15212655e-01 2.89825290e-01
-9.01133046e-02 7.10858226e-01 1.06011617e+00 -1.30085921e+00
-7.79917657e-01 -6.64028078e-02 1.42088592e-01 4.26190823e-01
-9.99033079e-02 1.67235330e-01 -5.64713895e-01 -1.07611108e+00
-1.13499742e-02 -3.97521108e-01 -2.40559340e-01 -6.30590320e-03
-5.69674782e-02 -1.03579096e-01 1.04431653e+00 -1.00424075e+00
1.40688992e+00 -2.26566553e+00 5.50653994e-01 -3.86635184e-01
3.06078047e-01 5.42153597e-01 -3.07337493e-01 -8.08430165e-02
-1.34863585e-01 1.40407503e-01 -7.38882050e-02 -3.05657417e-01
-4.90095913e-01 2.62805790e-01 -6.46365285e-01 3.01730275e-01
1.67282656e-01 1.12178874e+00 -9.67820346e-01 -4.71905023e-01
6.15549505e-01 1.12283576e+00 -4.05736566e-01 3.26327980e-01
-1.83748811e-01 9.73678946e-01 -6.22666717e-01 4.38744754e-01
5.51807404e-01 -3.99649084e-01 -1.98731467e-01 -5.53182662e-01
-4.96860176e-01 2.20135868e-01 -9.90607023e-01 1.96017432e+00
-5.19250393e-01 6.58677995e-01 -1.68156102e-01 -6.68451309e-01
8.53202164e-01 2.73225605e-01 5.82654595e-01 -9.08807158e-01
6.13642223e-02 1.24789044e-01 -2.90098399e-01 -5.27456522e-01
6.15994453e-01 2.03893110e-01 3.79983842e-01 3.43752913e-02
-7.71434903e-02 5.53319752e-01 8.78197781e-04 1.04015224e-01
7.18858957e-01 4.62966293e-01 9.82633904e-02 2.04779133e-01
9.09564972e-01 -6.50423408e-01 9.38981414e-01 2.15876117e-01
-1.66388676e-01 1.01161206e+00 1.25953868e-01 -6.54010594e-01
-1.29442370e+00 -9.65948582e-01 -3.13753597e-02 9.16263402e-01
4.82184052e-01 -2.94088274e-01 -4.61868346e-01 -4.15001154e-01
-3.34508568e-01 4.43071216e-01 -5.85866928e-01 -2.06518527e-02
-1.11136711e+00 -2.97827244e-01 1.95113607e-02 6.00174069e-01
6.15898490e-01 -1.28272748e+00 -4.66642767e-01 4.23432052e-01
-6.80139780e-01 -1.30981147e+00 -9.07882094e-01 -5.60251415e-01
-8.55049253e-01 -7.07483947e-01 -1.04998660e+00 -6.91962123e-01
3.94685686e-01 6.68539107e-01 8.22718143e-01 1.46437854e-01
-1.46227583e-01 3.60766985e-02 -6.69075370e-01 3.76313299e-01
-1.61019921e-01 -1.95863426e-01 -4.94844578e-02 2.51091361e-01
3.25102136e-02 -6.94324791e-01 -8.93465638e-01 4.36601281e-01
-9.88150239e-01 4.46690053e-01 5.55578649e-01 9.12800133e-01
1.03917146e+00 -1.19966045e-01 5.97580552e-01 -5.42076886e-01
5.02887219e-02 -4.82901961e-01 -4.23156559e-01 8.59141797e-02
-1.73432350e-01 -5.10200337e-02 7.90872455e-01 -6.02689326e-01
-1.22587693e+00 -1.55701980e-01 -3.63937855e-01 -1.09880531e+00
5.03498055e-02 1.77293316e-01 -1.83423951e-01 1.00828446e-02
4.95039299e-02 6.96041703e-01 -2.33877450e-01 -5.53519368e-01
3.79211068e-01 5.04752278e-01 6.27597511e-01 -2.65445948e-01
8.08718860e-01 5.57238758e-01 -4.02731746e-01 -6.01890802e-01
-9.22990441e-01 -2.03748167e-01 -4.81763095e-01 -2.28726894e-01
1.23082685e+00 -1.28484821e+00 -3.97506416e-01 3.19768131e-01
-1.02612805e+00 -6.50260150e-01 -3.24283183e-01 5.60742557e-01
-6.74802244e-01 4.29690242e-01 -8.31223667e-01 -6.17490053e-01
-3.22090268e-01 -1.20564818e+00 1.21101177e+00 4.63097185e-01
3.43434453e-01 -6.71023369e-01 -2.32292145e-01 4.53801960e-01
4.11089092e-01 1.89764351e-01 2.69689679e-01 1.15412116e-01
-1.06347537e+00 4.88270283e-01 -4.16880339e-01 3.37493777e-01
2.35838786e-01 -4.43618447e-02 -7.58430362e-01 -4.36878830e-01
-8.34955052e-02 -3.54745761e-02 1.14305425e+00 6.52552545e-01
1.65698493e+00 -5.03453195e-01 -8.48859623e-02 1.09306216e+00
1.28515768e+00 2.53925323e-01 1.07104242e+00 2.08564475e-01
1.16139627e+00 2.51997530e-01 9.19750810e-01 3.72444451e-01
5.93426228e-01 8.20172906e-01 3.40998828e-01 -1.82364091e-01
-5.82463920e-01 -4.02099907e-01 6.65986538e-01 8.63599718e-01
-5.49576402e-01 -1.45023420e-01 -2.85176545e-01 4.62187052e-01
-1.92199254e+00 -1.34987700e+00 1.08406752e-01 1.87669742e+00
8.38865042e-01 -9.76011530e-02 -4.75243814e-02 3.35754044e-02
8.96421790e-01 5.85065603e-01 -7.14839160e-01 6.25327677e-02
-2.76132286e-01 -4.77861054e-02 2.32830018e-01 6.79193854e-01
-9.39071059e-01 9.81913149e-01 4.71703625e+00 1.03200746e+00
-1.07132566e+00 2.92277902e-01 8.31348479e-01 -2.33856231e-01
-5.68717718e-01 -2.62133628e-01 -7.58185446e-01 7.76700914e-01
7.54414380e-01 2.25586421e-03 7.08903611e-01 4.88330007e-01
6.06554389e-01 3.24196368e-01 -7.52813816e-01 1.14713132e+00
-8.42675120e-02 -1.62329710e+00 2.13445619e-01 -1.35253251e-01
7.37950623e-01 -3.54125239e-02 2.93780118e-01 1.82793766e-01
9.23981518e-02 -1.02377760e+00 6.44531667e-01 7.77908206e-01
1.29042315e+00 -8.99668872e-01 5.06133556e-01 8.79440978e-02
-1.73442268e+00 -3.55125666e-01 -4.80616510e-01 3.31464946e-01
4.69667137e-01 4.14616019e-01 7.44128823e-02 6.74880207e-01
1.04557490e+00 1.33026135e+00 -2.46085256e-01 6.16861820e-01
-1.87113166e-01 2.81576395e-01 9.24849957e-02 6.91310763e-01
1.36512890e-01 -1.10703573e-01 7.18182683e-01 9.56501186e-01
4.97605979e-01 6.94432557e-01 9.12646651e-02 8.90452683e-01
-1.25485867e-01 -1.53127000e-01 -3.53339404e-01 1.46900728e-01
5.04899681e-01 1.22350669e+00 -3.18214804e-01 -2.70811498e-01
-6.43280923e-01 1.04936075e+00 1.61580682e-01 5.81538498e-01
-1.05416667e+00 -1.04855575e-01 7.40317106e-01 2.09391177e-01
5.88408411e-01 -2.36221865e-01 1.25320509e-01 -1.63556552e+00
8.61502886e-02 -8.57043207e-01 3.72133017e-01 -1.06145763e+00
-1.09105730e+00 7.22194791e-01 -2.54910856e-01 -1.45433855e+00
-8.61394256e-02 4.61942852e-02 -3.00548941e-01 8.77492428e-01
-1.93839490e+00 -1.15086401e+00 -4.30630594e-01 9.28674936e-01
1.20669425e+00 4.76971627e-06 1.95790857e-01 6.11121297e-01
-5.99484921e-01 4.71757531e-01 -1.10063590e-01 1.25420481e-01
6.81445956e-01 -7.30640590e-01 4.12345499e-01 1.06031370e+00
-4.09457907e-02 4.32366222e-01 4.66060251e-01 -6.08210623e-01
-1.41355717e+00 -1.46603489e+00 5.83939373e-01 -2.56673306e-01
4.16055888e-01 -1.20839626e-01 -1.30707800e+00 8.01315784e-01
-6.12083403e-03 5.83534479e-01 1.89827919e-01 -5.87279260e-01
-3.34944814e-01 -1.53088540e-01 -9.77467120e-01 5.49051821e-01
1.29454434e+00 -5.66213906e-01 -4.22489941e-01 -1.76314905e-01
1.44315052e+00 -5.41805089e-01 -9.95980382e-01 5.54245472e-01
5.39557874e-01 -1.04373944e+00 1.28395689e+00 -2.23393351e-01
8.00354302e-01 -6.87994659e-01 -2.85726935e-01 -9.77246761e-01
-5.86983681e-01 -6.64556801e-01 -6.43239558e-01 1.14352870e+00
-8.25024620e-02 -4.66210037e-01 4.09896821e-01 3.64003181e-01
-3.86021823e-01 -8.90762091e-01 -8.67021680e-01 -5.09902060e-01
-2.28009015e-01 -1.40306726e-01 6.81401908e-01 1.03660870e+00
-4.93897349e-01 4.82656881e-02 -8.62287462e-01 1.90073863e-01
5.74583232e-01 2.22748354e-01 4.30389136e-01 -7.95393944e-01
-2.90231586e-01 -2.84333944e-01 -6.82578757e-02 -1.28216827e+00
1.78240731e-01 -5.40152550e-01 -9.85791087e-02 -1.52473569e+00
2.01878428e-01 -2.05310956e-01 -4.57250327e-01 3.54782492e-01
-3.88191581e-01 4.08140093e-01 2.60434538e-01 3.57642859e-01
-7.35356867e-01 8.30858171e-01 1.79060400e+00 2.18043253e-01
-3.58360529e-01 -3.74947786e-01 -5.38253009e-01 6.48963988e-01
6.88467979e-01 -1.13608785e-01 -4.09402251e-01 -7.04966128e-01
-9.83754247e-02 5.19208610e-01 5.64331830e-01 -7.37593830e-01
-2.14917324e-02 -3.74143541e-01 7.54010320e-01 -7.23901629e-01
3.81717294e-01 -6.72494650e-01 2.77537376e-01 2.17104718e-01
-5.33109426e-01 -1.09922634e-02 -1.04027107e-01 6.21966064e-01
-2.57030249e-01 3.26867759e-01 9.79957521e-01 -1.72009856e-01
-1.00074613e+00 8.57803762e-01 -6.45404011e-02 -7.64520094e-02
8.65149796e-01 -1.45903647e-01 -3.75111580e-01 -2.29111835e-01
-6.34435415e-01 1.96615860e-01 4.95542705e-01 8.05127740e-01
1.20153904e+00 -1.50147557e+00 -9.47410226e-01 1.59811795e-01
-3.07561934e-01 3.11809212e-01 9.31416690e-01 7.29204774e-01
-2.50290453e-01 2.65482962e-01 -3.77028644e-01 -4.09420252e-01
-1.13054180e+00 7.46926010e-01 2.71169186e-01 -1.36757180e-01
-1.19979632e+00 8.38874876e-01 5.00342011e-01 1.84981495e-01
3.03524779e-04 -2.14073956e-01 -6.20645583e-01 -1.31352037e-01
1.05550504e+00 3.99620444e-01 -5.90027094e-01 -9.47122574e-01
-2.51371562e-02 8.15812647e-01 -1.46328568e-01 7.51310140e-02
1.45268166e+00 -7.52507150e-01 -7.25535024e-03 4.07730669e-01
1.19350791e+00 -2.20214963e-01 -1.84573603e+00 -4.91180986e-01
-5.48307180e-01 -9.57782924e-01 2.09707469e-01 -3.32204431e-01
-1.53665423e+00 7.78623641e-01 5.31255066e-01 -1.67870164e-01
1.43900776e+00 -5.92042506e-02 1.37543261e+00 -9.03467759e-02
3.81664246e-01 -7.45430827e-01 2.41175517e-01 3.88796777e-01
1.01662314e+00 -1.23379314e+00 -1.74861878e-01 -3.35311055e-01
-7.14450240e-01 1.14678216e+00 7.64500797e-01 -2.12698936e-01
2.82905877e-01 1.34547427e-01 -2.09006339e-01 3.16593081e-01
-9.50262785e-01 -2.35869363e-01 3.48831892e-01 5.52677214e-01
3.84855628e-01 -2.68697202e-01 -2.51767635e-01 6.62858605e-01
1.87881008e-01 2.85580218e-01 6.01586342e-01 3.64223987e-01
-4.40211594e-01 -8.89862180e-01 -8.36356729e-02 1.82770118e-01
-6.15563154e-01 -2.00348601e-01 2.74519324e-01 4.90843296e-01
2.72269547e-01 7.86635756e-01 3.73798683e-02 -5.03685474e-01
1.01645306e-01 -4.65516955e-01 3.25960428e-01 -2.98369706e-01
-2.47908056e-01 3.99627298e-01 -1.89244732e-01 -1.12247288e+00
-6.55071259e-01 -5.94867587e-01 -1.39579856e+00 -2.12439001e-01
4.79215235e-02 -1.25858530e-01 1.63699239e-01 7.31943250e-01
4.06442493e-01 8.67467761e-01 7.98976898e-01 -1.19809556e+00
2.80486457e-02 -6.82424903e-01 -4.25250620e-01 2.28812128e-01
8.67998898e-01 -5.69112718e-01 -2.87490994e-01 3.59741449e-01] | [11.012740135192871, -1.8241175413131714] |
32cf65f0-ee99-4fcb-9093-94aa508f013d | transformer-based-program-synthesis-for-low | 2205.09246 | null | https://arxiv.org/abs/2205.09246v1 | https://arxiv.org/pdf/2205.09246v1.pdf | Transformer-based Program Synthesis for Low-Data Environments | Recent advancements in large pre-trained transformer models (GPT2/3, T5) have found use in program synthesis to generate programs that satisfy a set of input/output examples. However, these models perform poorly on long-horizon and low-data tasks, and often don't seem to understand the semantics of the languages they generate. We investigate an approach that tackles both of these issues, by using attributed context-free-grammars of programming languages to generate programs, and then analyzing generated programs so that they can be annotated with compile and runtime attributes, such as types, so that information about the program can be remembered during long-horizon generation. We firstly find that synthesized datasets can be made efficiently and can provide transformer models with enough data in order to perform well on some synthesis tasks. We also find that giving models access to program attributes is especially effective in low-data environments, and tends improve the quality and reduce errors of transformer-generated programs. | ['Jack Roper'] | 2022-05-18 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 3.10485095e-01 4.54498529e-01 -4.77954447e-01 -4.61221397e-01
-8.28077197e-01 -7.50585854e-01 5.52125812e-01 3.74578387e-01
2.75045693e-01 4.30907309e-01 1.99668422e-01 -9.09567952e-01
3.44476193e-01 -1.40332818e+00 -1.07820654e+00 2.23712906e-01
3.77395167e-03 5.91281831e-01 3.93365175e-01 -4.48966354e-01
2.13211253e-01 -2.44733766e-02 -2.08510160e+00 8.08857858e-01
1.06648922e+00 3.51218343e-01 2.70966858e-01 9.18112695e-01
-6.25034034e-01 1.01559496e+00 -5.68970978e-01 -5.85292757e-01
-4.20674793e-02 -3.98766965e-01 -1.03071761e+00 -1.16538726e-01
2.18391865e-01 -1.64318591e-01 2.08076075e-01 8.41958165e-01
2.15449836e-02 -2.35212579e-01 4.40095544e-01 -1.43335021e+00
-4.83177423e-01 1.17522597e+00 -1.00719757e-01 -1.69584483e-01
5.37647247e-01 3.92336369e-01 1.23294795e+00 -6.51206613e-01
9.14099514e-01 1.29475045e+00 6.66139245e-01 8.53280246e-01
-1.52507973e+00 -3.02564174e-01 -1.55854076e-01 -3.77039373e-01
-7.93768704e-01 -5.39185941e-01 3.08386743e-01 -6.07624054e-01
1.55510330e+00 4.95431453e-01 5.66866338e-01 7.61528015e-01
2.87965983e-01 7.97046840e-01 7.88098097e-01 -7.61375666e-01
-5.60255237e-02 4.55109477e-01 1.74752787e-01 9.70369756e-01
4.94068116e-02 2.11590156e-01 -2.19590768e-01 -4.50827092e-01
2.76264787e-01 -3.57108980e-01 1.30408779e-01 -3.14580053e-01
-1.30896986e+00 8.04744840e-01 4.75393645e-02 3.22334290e-01
4.05042581e-02 3.66254002e-01 6.39900446e-01 6.88918471e-01
2.73770392e-01 8.76846015e-01 -6.91421509e-01 -3.51919919e-01
-9.08054948e-01 7.61649549e-01 1.08436894e+00 1.69204569e+00
9.86595988e-01 2.02519044e-01 -1.60666019e-01 5.13837039e-01
1.07977137e-01 4.77026880e-01 4.24779713e-01 -8.76950204e-01
8.01471055e-01 1.03287542e+00 -1.66568935e-01 -6.12212062e-01
8.65841508e-02 -1.34047464e-01 -1.49791643e-01 2.79175699e-01
3.37878078e-01 8.69961977e-02 -7.63132989e-01 1.73319161e+00
-1.10480428e-01 -5.65651000e-01 3.29098910e-01 1.53466299e-01
6.43259585e-01 8.10008347e-01 8.87821838e-02 7.19862208e-02
1.26046109e+00 -8.12741339e-01 -2.08380997e-01 -7.15218663e-01
1.41479897e+00 -6.34912312e-01 1.47851813e+00 1.75176710e-01
-1.40412128e+00 -6.59114957e-01 -1.01216972e+00 -7.73006901e-02
-3.60062718e-01 1.53677121e-01 9.07789767e-01 8.01275909e-01
-1.29014182e+00 6.81411803e-01 -8.16961586e-01 -8.62125978e-02
2.65907019e-01 3.69632542e-01 -6.80241287e-02 -1.62115186e-01
-7.36871600e-01 8.45324099e-01 5.21568596e-01 -3.30042124e-01
-1.14127231e+00 -9.72575188e-01 -1.17072618e+00 2.72509456e-02
2.58078784e-01 -8.64520967e-01 1.73235643e+00 -1.19365728e+00
-1.00174391e+00 8.66975725e-01 -2.90324271e-01 -3.91477644e-01
1.96411446e-01 1.81902453e-01 -2.60865033e-01 -5.79411983e-01
9.09218416e-02 5.54014802e-01 5.09805143e-01 -1.20826709e+00
-8.95977318e-01 -2.25257531e-01 3.95143837e-01 -1.76001891e-01
-2.92287767e-01 3.28430742e-01 -1.11047067e-01 -2.78389722e-01
-3.30811352e-01 -9.60276306e-01 -2.73154587e-01 -4.17957872e-01
-3.32511663e-01 -1.91377267e-01 7.51196325e-01 -6.77474916e-01
1.34358597e+00 -1.98476923e+00 1.62040859e-01 1.98184729e-01
1.37857676e-01 2.47139096e-01 -1.42107725e-01 4.79657680e-01
-1.09827481e-01 6.43572152e-01 -2.69152880e-01 -4.45339121e-02
4.21107054e-01 3.67547780e-01 -6.41802728e-01 -2.09264666e-01
4.95185673e-01 9.76220548e-01 -8.91060054e-01 -4.15192604e-01
-9.61379036e-02 -1.77899197e-01 -1.05919683e+00 5.94079971e-01
-1.23636460e+00 9.39402953e-02 -5.44713140e-01 4.73159224e-01
6.79546818e-02 -7.90971816e-02 1.85633615e-01 2.46844471e-01
-1.19173616e-01 7.03487515e-01 -9.27978516e-01 1.51285040e+00
-1.13089478e+00 7.10839868e-01 -2.24930391e-01 -6.41254723e-01
1.08374953e+00 3.20715427e-01 -1.95974275e-01 -6.33996964e-01
-3.35448980e-01 5.87946892e-01 7.00313598e-02 -6.83003485e-01
8.13169360e-01 -1.15651548e-01 -5.66707730e-01 7.79906154e-01
-9.91422758e-02 -6.59783900e-01 6.66227221e-01 1.01904631e-01
1.16364324e+00 4.63571042e-01 1.42753527e-01 -2.87011087e-01
4.49353337e-01 6.45246446e-01 5.92972100e-01 6.65582061e-01
8.28371167e-01 3.97480816e-01 9.44244921e-01 -6.14150763e-01
-1.56991613e+00 -3.38998437e-01 3.82423878e-01 1.30007505e+00
-6.57729924e-01 -9.04802024e-01 -8.74158740e-01 -6.76845968e-01
-2.26568639e-01 1.39728725e+00 -3.62812728e-01 -3.11823219e-01
-9.54903483e-01 -5.41549742e-01 7.35690475e-01 7.18205929e-01
-6.49321917e-03 -1.23631132e+00 -1.00744355e+00 3.77803266e-01
-1.29653126e-01 -7.02710450e-01 -2.95904338e-01 3.53041053e-01
-9.92968023e-01 -1.05277872e+00 -1.24295570e-01 -9.93621886e-01
9.22210336e-01 -4.29671764e-01 1.77862811e+00 5.83921432e-01
-1.18059024e-01 5.10585941e-02 -3.42651516e-01 -4.87439215e-01
-1.49775386e+00 2.68429488e-01 -6.29459977e-01 -6.89734697e-01
2.17562318e-01 -5.38519561e-01 3.34843427e-01 1.94872186e-01
-1.05063593e+00 3.05775464e-01 3.63790572e-01 8.08527768e-01
2.89976746e-01 2.87262052e-01 1.79300129e-01 -1.55076110e+00
5.46349168e-01 -2.54533142e-01 -9.66986001e-01 5.37495911e-01
-7.74366975e-01 6.46527946e-01 1.08115470e+00 -2.86247045e-01
-1.08700979e+00 -2.07351804e-01 -2.41903305e-01 2.78898533e-02
4.75301109e-02 6.84007764e-01 -3.31142306e-01 1.40477791e-01
1.08237004e+00 2.32527018e-01 -1.87870726e-01 -1.41054347e-01
3.52656931e-01 4.31277663e-01 3.68145913e-01 -1.51374865e+00
7.89776206e-01 -2.78083175e-01 -1.19555213e-01 -3.43318433e-01
-3.16729575e-01 2.06105277e-01 -3.62363487e-01 4.33090121e-01
2.76603937e-01 -6.73000991e-01 -1.39288887e-01 2.89164364e-01
-1.23836243e+00 -9.66458380e-01 -4.86966908e-01 -1.95315659e-01
-9.19685900e-01 -1.27278436e-02 -4.66352522e-01 -8.00246716e-01
-3.23821336e-01 -1.67478061e+00 1.14859807e+00 -5.22930324e-02
-6.58760309e-01 -1.00049019e+00 2.95804292e-02 -1.63124993e-01
6.15996599e-01 1.39045238e-01 2.08912325e+00 -6.35922730e-01
-8.69945824e-01 -4.46621850e-02 1.41150787e-01 3.34647596e-01
-7.87432864e-02 4.97863650e-01 -7.02257693e-01 -1.20763816e-02
-3.31164241e-01 -3.63238752e-01 1.17048472e-01 -6.60813153e-02
1.19961512e+00 -5.96841514e-01 -4.88462776e-01 5.80291212e-01
1.35167837e+00 2.34070480e-01 8.56026530e-01 1.81485534e-01
6.08705580e-01 7.66000152e-01 5.56397498e-01 7.45648332e-03
5.31056166e-01 6.35094821e-01 1.42715201e-01 1.11687787e-01
-9.64418650e-02 -6.50902152e-01 4.25594062e-01 5.79915702e-01
1.93850935e-01 -3.03331185e-02 -1.25368822e+00 8.87439787e-01
-1.72313750e+00 -7.51714587e-01 -4.12164241e-01 2.28746438e+00
1.22782886e+00 3.40360790e-01 2.53145725e-01 6.44119084e-02
3.26887697e-01 -2.05647215e-01 -1.85193598e-01 -8.84095371e-01
1.62468836e-01 4.98204291e-01 4.16863978e-01 4.23419803e-01
-4.84619051e-01 9.36896443e-01 6.75262880e+00 5.32141089e-01
-9.74015892e-01 -3.97911556e-02 5.16324818e-01 5.42783439e-01
-1.16967332e+00 6.32341266e-01 -9.15432870e-01 2.27561936e-01
1.48976922e+00 -7.06082523e-01 5.46154141e-01 1.22319782e+00
-1.74387380e-01 1.43740103e-02 -1.81906533e+00 2.54672736e-01
-1.58420473e-01 -1.38499749e+00 7.36943707e-02 -5.10515161e-02
6.50368571e-01 -4.43455487e-01 -7.07877725e-02 7.78686166e-01
6.54463351e-01 -1.20242107e+00 9.38565671e-01 2.55956888e-01
8.46098781e-01 -8.32713962e-01 6.34721756e-01 6.02367699e-01
-1.17016566e+00 -1.58099666e-01 -2.75342971e-01 2.04211511e-02
-3.09336364e-01 4.88511443e-01 -1.18481386e+00 3.72070611e-01
2.82298416e-01 5.93986213e-01 -8.81965876e-01 6.61014616e-01
-3.42165291e-01 3.98056000e-01 -3.76791731e-02 -3.07248384e-01
5.18534593e-02 1.51056036e-01 2.53506958e-01 1.20287418e+00
6.74761057e-01 -2.31088385e-01 2.04276666e-01 1.44100106e+00
1.11208878e-01 -5.07624187e-02 -1.05980325e+00 -4.52733040e-01
2.69384056e-01 9.68747139e-01 -3.50680023e-01 -6.45662606e-01
-3.93048257e-01 2.01035202e-01 1.33697391e-01 3.09149716e-02
-6.61415458e-01 -4.23107892e-01 2.63957232e-01 3.30896795e-01
1.61335036e-01 -1.02356754e-01 -2.97799438e-01 -1.09928310e+00
2.79132426e-01 -1.78815317e+00 2.67386407e-01 -9.38246548e-01
-4.89072561e-01 7.85802603e-01 2.92734295e-01 -7.63924479e-01
-8.48158777e-01 -4.95828450e-01 -6.11052632e-01 1.01739097e+00
-1.12460780e+00 -1.20013022e+00 9.83588994e-02 2.61903942e-01
5.30856133e-01 -1.45717740e-01 1.09424388e+00 1.68812335e-01
-1.58508450e-01 6.30315542e-01 -4.96086031e-01 -7.21860155e-02
1.83337390e-01 -1.48852980e+00 1.00388587e+00 9.85825539e-01
-6.21788129e-02 1.04433012e+00 9.29044724e-01 -6.14623666e-01
-1.96309114e+00 -1.33194458e+00 1.21711540e+00 -7.30513275e-01
6.98267937e-01 -3.74586433e-01 -9.15894270e-01 1.09351134e+00
-2.70589236e-02 -1.98051959e-01 3.06734324e-01 2.32119665e-01
-5.84640861e-01 -5.51022515e-02 -8.93962383e-01 5.92991889e-01
1.02230132e+00 -6.72508299e-01 -7.43069291e-01 5.13010263e-01
8.35333228e-01 -7.75038362e-01 -9.69588697e-01 2.43304595e-01
2.21317142e-01 -7.82001972e-01 5.69501936e-01 -9.78793621e-01
9.54958677e-01 -2.07647115e-01 -2.63789624e-01 -1.57163036e+00
1.86072052e-01 -8.06371033e-01 3.72527614e-02 1.44391620e+00
1.13978612e+00 -5.83142698e-01 6.53372228e-01 9.06705320e-01
-4.89300340e-01 -4.63891566e-01 -4.23992537e-02 -6.68127179e-01
1.77758262e-01 -6.46851361e-01 1.22301948e+00 5.25587380e-01
2.69416034e-01 3.85916352e-01 -8.01522806e-02 -1.96153194e-01
2.91624665e-01 6.81275368e-01 1.15379846e+00 -9.27728534e-01
-5.99814534e-01 -4.17893887e-01 -3.82015631e-02 -6.13210559e-01
4.83010501e-01 -1.19136631e+00 2.48146892e-01 -1.33146763e+00
2.63031852e-02 -9.78475273e-01 6.31036758e-01 8.50440741e-01
-3.58364321e-02 -4.53264415e-01 -1.49481550e-01 -3.84829491e-02
-3.16106342e-02 1.07923895e-01 8.59257579e-01 -3.04255635e-01
-1.17255338e-01 2.06652388e-01 -8.04636657e-01 4.07844454e-01
5.21823406e-01 -7.37846375e-01 -6.84448779e-01 -6.61636174e-01
6.70631349e-01 4.93132055e-01 2.50589669e-01 -9.30240691e-01
5.27577549e-02 -4.65538025e-01 -1.80898845e-01 -2.38249853e-01
-2.25539848e-01 -6.74818814e-01 6.36104703e-01 5.05861104e-01
-7.57364154e-01 5.27136683e-01 3.63179356e-01 -2.95417774e-02
-2.31104478e-01 -7.80935585e-01 4.68437970e-01 -6.29296720e-01
-8.17855775e-01 1.95964262e-01 -3.12296987e-01 3.43127131e-01
8.28800321e-01 8.43253359e-03 -4.40097570e-01 -7.72215351e-02
-3.53199661e-01 1.75298676e-01 1.00973070e+00 4.72204417e-01
2.51234502e-01 -1.09429097e+00 -6.19559467e-01 6.00867152e-01
5.04793286e-01 1.81771711e-01 -2.54723012e-01 4.18398708e-01
-6.47761583e-01 4.14112836e-01 -2.29197919e-01 -5.76312542e-01
-1.25308549e+00 7.43062317e-01 2.82831907e-01 -4.92707849e-01
-5.61397672e-01 4.90214765e-01 4.67694327e-02 -9.10061777e-01
-1.95386097e-01 -7.41814971e-01 1.70894742e-01 -3.02047968e-01
4.92592931e-01 -2.84889400e-01 4.44279343e-01 -1.64741129e-01
-2.40971074e-01 2.61844665e-01 1.41454944e-02 -6.05185255e-02
1.56455135e+00 7.34320104e-01 -4.29834187e-01 3.02027494e-01
1.10135329e+00 1.45630002e-01 -7.16656983e-01 -1.96597248e-01
3.32395881e-01 -4.58170086e-01 -3.05101871e-01 -5.12185931e-01
-9.12205756e-01 9.71578598e-01 -2.68728971e-01 4.11174774e-01
1.02184534e+00 -8.24933648e-02 7.10490346e-01 5.40235639e-01
7.69758582e-01 -7.82366812e-01 -6.03823885e-02 6.00052178e-01
5.81863165e-01 -7.49324858e-01 -2.84557313e-01 -4.88106072e-01
-4.65331286e-01 1.34259927e+00 8.41867030e-01 1.17433570e-01
-6.02314584e-02 9.08804297e-01 -2.86852837e-01 -1.31028727e-01
-1.24458826e+00 8.02827924e-02 6.85392916e-02 8.79390836e-01
6.52137518e-01 3.07728350e-02 8.94048661e-02 3.47048283e-01
-6.11468256e-01 1.61577404e-01 8.88238013e-01 1.25160706e+00
-3.35960478e-01 -1.78013575e+00 -2.31049240e-01 7.23768651e-01
-3.61457556e-01 -3.08026850e-01 -2.39101261e-01 9.24631119e-01
-6.02592975e-02 7.13351011e-01 -1.47820756e-01 -3.98951530e-01
5.34598827e-01 3.41970265e-01 6.68892443e-01 -1.35375142e+00
-9.01865780e-01 -3.54169995e-01 8.64560664e-01 -4.37088817e-01
-5.08104973e-02 -5.57055712e-01 -1.05379879e+00 -4.04923707e-01
-1.07530624e-01 5.00325203e-01 5.20797312e-01 8.60748112e-01
3.04963559e-01 6.31050825e-01 3.99655163e-01 -2.92541891e-01
-7.82333374e-01 -4.29379314e-01 -9.00829658e-02 2.45278135e-01
2.14449838e-01 -6.93641976e-02 -5.95128424e-02 4.57683533e-01] | [8.194644927978516, 7.468592166900635] |
c646874c-440b-44b2-9cfa-e5e9a693de70 | grounded-image-captioning-in-top-down-view | 2306.07490 | null | https://arxiv.org/abs/2306.07490v2 | https://arxiv.org/pdf/2306.07490v2.pdf | Top-Down Viewing for Weakly Supervised Grounded Image Captioning | Weakly supervised grounded image captioning (WSGIC) aims to generate the caption and ground (localize) predicted object words in the input image without using bounding box supervision. Recent two-stage solutions mostly apply a bottom-up pipeline: (1) first apply an off-the-shelf object detector to encode the input image into multiple region features; (2) and then leverage a soft-attention mechanism for captioning and grounding. However, object detectors are mainly designed to extract object semantics (i.e., the object category). Besides, they break down the structural images into pieces of individual proposals. As a result, the subsequent grounded captioner is often overfitted to find the correct object words, while overlooking the relation between objects (e.g., what is the person doing?), and selecting incompatible proposal regions for grounding. To address these difficulties, we propose a one-stage weakly supervised grounded captioner that directly takes the RGB image as input to perform captioning and grounding at the top-down image level. In addition, we explicitly inject a relation module into our one-stage framework to encourage the relation understanding through multi-label classification. The relation semantics aid the prediction of relation words in the caption. We observe that the relation words not only assist the grounded captioner in generating a more accurate caption but also improve the grounding performance. We validate the effectiveness of our proposed method on two challenging datasets (Flick30k Entities captioning and MSCOCO captioning). The experimental results demonstrate that our method achieves state-of-the-art grounding performance. | ['Kim-Hui Yap', 'Suchen Wang', 'Chen Cai'] | 2023-06-13 | null | null | null | null | ['image-captioning'] | ['computer-vision'] | [ 3.69578898e-01 7.22593486e-01 -2.37962529e-01 -5.42340994e-01
-1.18394983e+00 -4.48684454e-01 3.49861592e-01 9.35325846e-02
-1.77912846e-01 4.51085925e-01 2.11538300e-01 -9.46119726e-02
5.42766929e-01 -8.62812400e-01 -1.36881888e+00 -5.67817688e-01
5.14367163e-01 7.15238929e-01 3.33613783e-01 -2.50801325e-01
-8.47178772e-02 -1.23401329e-01 -1.50448787e+00 7.18260109e-01
7.69842029e-01 1.23027062e+00 4.84847039e-01 2.15004146e-01
-2.81470269e-01 8.34751129e-01 -4.08175737e-01 -7.30736911e-01
6.06335253e-02 -4.55768734e-01 -7.55575120e-01 2.69602358e-01
5.25340676e-01 -4.21681643e-01 -2.24323094e-01 1.21259606e+00
4.48734425e-02 -1.60719424e-01 3.32188517e-01 -1.51521277e+00
-9.62810814e-01 8.27513397e-01 -5.70422411e-01 -1.61723733e-01
4.18059230e-01 3.22826415e-01 1.31719887e+00 -1.14105654e+00
4.98128057e-01 1.25228989e+00 1.82137311e-01 6.72226489e-01
-9.31587160e-01 -7.86527812e-01 5.51731884e-01 1.08669125e-01
-1.45932662e+00 -1.68805256e-01 9.46500003e-01 -3.42109561e-01
3.87074441e-01 1.65464580e-01 6.39165819e-01 1.30275655e+00
-4.28833276e-01 1.08126998e+00 8.48670661e-01 -2.10507005e-01
2.50885099e-01 2.13185191e-01 2.27501497e-01 8.69881332e-01
3.52479130e-01 -2.32828304e-01 -5.76151311e-01 -4.55220975e-02
5.90117335e-01 5.26947230e-02 -3.46799910e-01 -3.57261926e-01
-1.62917244e+00 6.48124695e-01 1.09097207e+00 -1.61430966e-02
-5.25899768e-01 3.95386785e-01 -3.61427665e-02 -6.08047962e-01
3.28530610e-01 3.31816882e-01 -1.78239837e-01 4.32251871e-01
-9.59774852e-01 3.88285786e-01 2.97571033e-01 1.33145034e+00
1.15111947e+00 -5.60248315e-01 -6.19682729e-01 3.76334459e-01
8.36408556e-01 6.68644369e-01 1.77929565e-01 -4.58512783e-01
1.07227945e+00 1.09950459e+00 3.51062566e-01 -1.06573725e+00
-1.13289103e-01 -3.93135935e-01 -5.37584722e-01 -2.45238155e-01
2.11084664e-01 2.11862147e-01 -1.29985368e+00 1.60929954e+00
4.56872612e-01 4.47765112e-01 7.48412758e-02 1.50388074e+00
1.26176190e+00 8.69206429e-01 5.11932671e-01 2.28736967e-01
1.82242703e+00 -1.35668015e+00 -7.03463912e-01 -7.47497439e-01
6.88752770e-01 -4.49113488e-01 1.23449814e+00 -1.89621463e-01
-8.78011644e-01 -5.82540512e-01 -1.01931906e+00 -2.42584810e-01
-3.21856171e-01 2.61974543e-01 4.47907299e-01 9.01089534e-02
-6.24092758e-01 9.12219435e-02 -6.64091825e-01 -2.29811043e-01
6.77100003e-01 1.92983389e-01 -2.97815412e-01 -2.55756140e-01
-1.29565632e+00 6.78269982e-01 8.04093421e-01 3.65978837e-01
-1.04304707e+00 -4.22349364e-01 -1.11927104e+00 9.62938145e-02
5.80371499e-01 -7.31816471e-01 1.09196460e+00 -1.19509017e+00
-9.68514681e-01 1.14975667e+00 -9.80521217e-02 -2.85711020e-01
3.16031814e-01 -2.51444936e-01 -2.46206447e-01 2.46310666e-01
4.24200118e-01 1.21958685e+00 6.79182708e-01 -1.60144258e+00
-6.63789868e-01 -2.81875879e-01 2.78870612e-01 4.63828981e-01
-1.96980134e-01 -4.77031805e-02 -7.88962483e-01 -5.18681526e-01
3.45056832e-01 -9.59653258e-01 -1.15075149e-01 1.42272785e-01
-8.63941073e-01 -2.57480830e-01 6.26378357e-01 -6.09164417e-01
8.68074775e-01 -2.21587944e+00 9.69293341e-02 -5.82228713e-02
2.50602514e-01 -3.64559740e-02 -1.46404833e-01 -1.88303255e-02
-4.07456495e-02 1.81051984e-01 -3.37892532e-01 -5.66413879e-01
-8.24263170e-02 3.39430630e-01 -5.82295775e-01 3.15883219e-01
6.46180809e-01 1.42086768e+00 -1.11399817e+00 -9.38419163e-01
1.15178891e-01 2.45874465e-01 -5.07745445e-01 6.08828485e-01
-5.97011209e-01 2.53061920e-01 -6.62233353e-01 7.25479305e-01
6.06327951e-01 -6.69036031e-01 -1.85842112e-01 -6.31173193e-01
2.33476669e-01 2.43038341e-01 -9.96533930e-01 1.92954886e+00
-2.09400266e-01 2.95979500e-01 -2.22933665e-01 -8.19183707e-01
9.33206618e-01 1.92610070e-01 1.09699510e-01 -4.95645821e-01
1.68157756e-01 1.77971005e-01 -3.87503982e-01 -6.06331170e-01
4.69033867e-01 -2.29537293e-01 -3.70314002e-01 2.21318662e-01
8.77755210e-02 9.91575122e-02 -2.76069045e-01 4.72775191e-01
6.61526084e-01 5.83540440e-01 -3.99418967e-03 4.35485654e-02
3.61863852e-01 4.14549351e-01 3.74517441e-01 4.68644828e-01
-1.31032348e-01 1.15979803e+00 3.82803053e-01 -3.73859435e-01
-9.44075704e-01 -9.87093329e-01 2.81646490e-01 1.13646960e+00
8.79288077e-01 -2.42430076e-01 -8.35272789e-01 -9.92720485e-01
-2.17881396e-01 7.51373708e-01 -9.52611268e-01 -2.52403796e-01
-5.79776108e-01 -5.04859447e-01 2.78203696e-01 7.51190722e-01
5.52995384e-01 -1.39980793e+00 -4.43561435e-01 7.75667876e-02
-7.22263336e-01 -1.42238653e+00 -4.25918072e-01 1.45985126e-01
-4.48586583e-01 -9.42055404e-01 -7.55771875e-01 -1.03101647e+00
1.10335660e+00 2.00920105e-01 1.16985655e+00 2.62622952e-01
3.23277712e-01 -1.21126659e-02 -5.35072982e-01 -3.48355949e-01
-2.56776959e-01 1.55778199e-01 -2.87583470e-01 4.29911554e-01
3.78625125e-01 1.40916705e-01 -7.31817961e-01 3.72951448e-01
-7.39302337e-01 7.89663374e-01 7.35137820e-01 6.59502745e-01
8.84952724e-01 -5.28380513e-01 3.54634851e-01 -6.82189465e-01
-9.56435688e-03 -6.20947301e-01 -5.18443942e-01 4.36817318e-01
-1.96620241e-01 1.38462454e-01 4.05520201e-01 -4.75958735e-01
-8.88629794e-01 4.91253495e-01 -1.83748640e-02 -6.02183580e-01
-1.58982977e-01 2.85313815e-01 -5.28710008e-01 3.27694237e-01
5.70913434e-01 3.22287172e-01 -3.37827682e-01 -2.17039302e-01
6.36998713e-01 8.40523899e-01 7.88508117e-01 -6.71124160e-01
1.04127824e+00 5.21261990e-01 -3.63762170e-01 -6.54055476e-02
-1.46463645e+00 -4.70113218e-01 -4.04233128e-01 -2.63018012e-01
1.36433017e+00 -1.33304393e+00 -4.93219495e-01 -1.97584368e-02
-1.46465218e+00 -1.87727779e-01 -2.00172454e-01 2.38287777e-01
-5.38540184e-01 -1.44984692e-01 -2.65993625e-01 -6.60430670e-01
-4.32717174e-01 -1.16654611e+00 1.77806175e+00 3.58977586e-01
9.46633369e-02 -3.00303161e-01 -3.41674745e-01 7.51731217e-01
-1.83561996e-01 3.30588341e-01 6.30844533e-01 -6.19289339e-01
-9.80621338e-01 -1.59620345e-01 -6.59023106e-01 -3.35567184e-02
-2.25797549e-01 -3.03708255e-01 -1.16806829e+00 1.06192142e-01
-4.81463462e-01 -4.94498968e-01 6.44242048e-01 -1.12044672e-02
1.14883459e+00 -3.76373231e-01 -5.10002434e-01 5.67719042e-01
1.40929306e+00 -1.78339675e-01 6.25501275e-01 3.87755275e-01
1.21212733e+00 8.57426822e-01 9.98070538e-01 -6.81447387e-02
6.12952232e-01 7.67762303e-01 9.82351601e-01 -4.94154245e-01
-5.97327650e-02 -9.01551187e-01 3.28753352e-01 2.32043207e-01
2.25656405e-01 -5.01298666e-01 -1.03615499e+00 6.70616031e-01
-2.25741768e+00 -6.75831556e-01 -2.58122802e-01 1.73398280e+00
7.52902627e-01 2.04245806e-01 -4.25918698e-02 -2.85405189e-01
1.14878404e+00 2.49421131e-02 -6.11372530e-01 2.30306596e-01
-4.46641631e-02 -3.42009425e-01 3.27405185e-01 2.22899422e-01
-1.05268383e+00 1.29501963e+00 4.38185596e+00 4.86187667e-01
-1.02123404e+00 1.18715346e-01 9.84584987e-01 -5.97317666e-02
-4.87640202e-01 6.48340732e-02 -9.49496686e-01 5.51273942e-01
5.75026631e-01 1.32052705e-01 2.39921585e-02 1.09896183e+00
-1.15926024e-02 1.95533842e-01 -1.27707112e+00 1.26338148e+00
1.76129416e-01 -1.32563078e+00 4.05784547e-01 -2.58434732e-02
5.67014992e-01 -6.01130724e-02 -1.18098356e-01 4.16113406e-01
-6.72209784e-02 -1.04125583e+00 1.42856181e+00 3.18344623e-01
7.24659026e-01 -4.45306271e-01 9.03537631e-01 2.95076013e-01
-1.24102521e+00 4.75702584e-02 -2.92385101e-01 3.90520394e-02
4.15015399e-01 4.27976608e-01 -7.98643291e-01 4.01278794e-01
6.47000849e-01 4.48681951e-01 -5.29164612e-01 8.07344437e-01
-6.41408324e-01 5.65309882e-01 -2.59598106e-01 -4.79668826e-02
6.00323021e-01 9.73613411e-02 1.99326470e-01 9.26234126e-01
2.08345219e-01 4.65475827e-01 2.41517529e-01 1.21919608e+00
-2.97614992e-01 1.45235568e-01 -3.43817800e-01 -6.51057288e-02
2.95237571e-01 1.38857961e+00 -8.54088485e-01 -5.65021932e-01
-4.01401728e-01 1.08856595e+00 3.87763172e-01 4.69516993e-01
-1.26686418e+00 -1.12215415e-01 4.81923491e-01 4.82828051e-01
3.84088188e-01 2.41678044e-01 -3.00152808e-01 -1.33112669e+00
4.07497019e-01 -5.95464110e-01 2.21768215e-01 -1.46818185e+00
-9.74478483e-01 7.55225480e-01 -1.06836140e-01 -1.16111648e+00
1.51139975e-01 -5.89978695e-01 -3.55367899e-01 8.58174264e-01
-1.48392773e+00 -1.66732407e+00 -8.90518069e-01 3.65938753e-01
4.65668350e-01 5.10918796e-01 5.65499663e-01 2.22031608e-01
-7.38070905e-01 3.81600916e-01 -8.05388868e-01 3.89622629e-01
6.09644532e-01 -1.14439535e+00 4.54446197e-01 8.28972459e-01
3.44126374e-01 5.06032467e-01 6.37501121e-01 -7.59346306e-01
-1.01547551e+00 -1.61249244e+00 8.03926170e-01 -6.74422741e-01
4.36481327e-01 -8.43419075e-01 -9.17070568e-01 7.66695917e-01
-5.23645803e-02 5.06527841e-01 3.99378479e-01 -1.61480859e-01
-4.49821204e-01 6.93391487e-02 -1.03547347e+00 6.27061188e-01
1.22776401e+00 -4.66161609e-01 -7.80467629e-01 4.59537923e-01
1.20401776e+00 -5.97998440e-01 -2.55252749e-01 2.56114841e-01
3.75957966e-01 -5.71585774e-01 9.14899468e-01 -5.63247561e-01
8.50987911e-01 -7.29938030e-01 -1.93287462e-01 -8.53069961e-01
-3.16227108e-01 -3.03364545e-01 -1.64192274e-01 1.32209754e+00
6.43680453e-01 -9.30282176e-02 1.00072563e+00 7.63514221e-01
-2.87429094e-01 -9.75823939e-01 -5.26777387e-01 -3.46798748e-01
-4.75702196e-01 -4.42814291e-01 8.89100194e-01 7.82333493e-01
-1.79549456e-01 5.45833230e-01 -2.14424849e-01 7.43018329e-01
6.05533421e-01 4.59987402e-01 6.57159209e-01 -9.25102651e-01
-7.31483996e-02 1.56129867e-01 -5.19752383e-01 -1.30498016e+00
2.21218646e-01 -8.51207733e-01 6.79129362e-01 -1.85707843e+00
4.21821088e-01 -5.46778262e-01 -3.98235261e-01 9.72643614e-01
-7.17086136e-01 6.98615551e-01 2.46623516e-01 3.54824871e-01
-1.07545078e+00 6.27077818e-01 1.29244959e+00 -3.60443354e-01
8.94305035e-02 -4.55107391e-01 -1.04646873e+00 6.30526841e-01
4.85094965e-01 -7.10196078e-01 -3.77744496e-01 -5.82801580e-01
4.09233779e-01 -9.71783847e-02 8.88591349e-01 -8.88407171e-01
5.60581498e-02 -2.75964797e-01 4.13124681e-01 -6.68216586e-01
2.88150191e-01 -8.44827116e-01 -1.53195560e-01 1.87055424e-01
-2.88672984e-01 -1.18263662e-01 2.41521802e-02 6.00934803e-01
-1.89946547e-01 -9.90445316e-02 6.70719385e-01 -1.01339601e-01
-7.94191420e-01 4.14432943e-01 3.51961583e-01 1.99393295e-02
1.32727289e+00 -6.95497543e-02 -3.69610816e-01 -5.02240397e-02
-5.77910066e-01 4.30651933e-01 5.70538163e-01 6.56784892e-01
5.66592336e-01 -1.39060438e+00 -6.66731298e-01 1.41721340e-02
7.93498755e-01 7.09383547e-01 1.32629901e-01 3.86259735e-01
-4.34154809e-01 2.20461264e-01 4.99324873e-02 -8.03310037e-01
-9.73261833e-01 8.18337321e-01 2.72825956e-01 4.33843359e-02
-5.79606652e-01 1.19596589e+00 7.12811053e-01 -4.96203527e-02
1.26497552e-01 -6.00881696e-01 -8.75718333e-03 8.57845917e-02
4.89541084e-01 -4.24838454e-01 -2.05141664e-01 -9.07577872e-01
-6.68784618e-01 5.63172221e-01 -6.67103380e-02 -1.25603169e-01
1.11071157e+00 -5.48160896e-02 2.94989757e-02 1.81331411e-01
1.05467498e+00 -4.55103666e-01 -1.41409481e+00 -9.41107720e-02
-2.49907345e-01 -2.26194680e-01 5.88277392e-02 -7.47491777e-01
-8.84846330e-01 8.18443954e-01 4.26122278e-01 -9.24548060e-02
8.96549582e-01 6.15661144e-01 8.08001459e-01 1.72207326e-01
4.45732981e-01 -6.02225602e-01 3.07730883e-01 -7.46042356e-02
9.63106692e-01 -1.48080993e+00 -2.20249429e-01 -7.39507258e-01
-8.07245553e-01 6.59361959e-01 1.10498285e+00 5.86714223e-02
1.88602712e-02 -8.75581130e-02 2.13472947e-01 -4.20453280e-01
-3.49266648e-01 -4.50257152e-01 3.88041854e-01 4.07669067e-01
1.31928280e-01 5.87343983e-02 1.46190852e-01 7.67252743e-01
-1.26224488e-01 -2.25127295e-01 2.34640807e-01 5.61537504e-01
-4.71698105e-01 -6.04117751e-01 -4.99619752e-01 2.06603974e-01
-7.69853592e-03 -3.26150715e-01 -5.22938490e-01 3.88321459e-01
5.00622511e-01 8.76075447e-01 1.29496425e-01 -3.02278161e-01
3.42657864e-01 3.61861885e-02 1.11547798e-01 -9.89724815e-01
-5.18948436e-01 -2.29639724e-01 -9.52048451e-02 -6.73029721e-01
-3.27856749e-01 -2.60482460e-01 -1.66667831e+00 4.56016600e-01
-5.77512741e-01 3.33162010e-01 6.38280869e-01 1.09718382e+00
3.63123804e-01 5.91160059e-01 2.96697050e-01 -7.69420922e-01
-9.61838886e-02 -9.37317014e-01 -1.33365467e-01 7.36843646e-01
3.25520158e-01 -6.61732674e-01 -2.01879516e-01 3.10681373e-01] | [10.481401443481445, 1.4133226871490479] |
2541dd3a-edab-494d-a3b4-a35593a49172 | simplicial-complex-based-point-correspondence | 2007.02381 | null | https://arxiv.org/abs/2007.02381v3 | https://arxiv.org/pdf/2007.02381v3.pdf | Simplicial Complex based Point Correspondence between Images warped onto Manifolds | Recent increase in the availability of warped images projected onto a manifold (e.g., omnidirectional spherical images), coupled with the success of higher-order assignment methods, has sparked an interest in the search for improved higher-order matching algorithms on warped images due to projection. Although currently, several existing methods "flatten" such 3D images to use planar graph / hypergraph matching methods, they still suffer from severe distortions and other undesired artifacts, which result in inaccurate matching. Alternatively, current planar methods cannot be trivially extended to effectively match points on images warped onto manifolds. Hence, matching on these warped images persists as a formidable challenge. In this paper, we pose the assignment problem as finding a bijective map between two graph induced simplicial complexes, which are higher-order analogues of graphs. We propose a constrained quadratic assignment problem (QAP) that matches each p-skeleton of the simplicial complexes, iterating from the highest to the lowest dimension. The accuracy and robustness of our approach are illustrated on both synthetic and real-world spherical / warped (projected) images with known ground-truth correspondences. We significantly outperform existing state-of-the-art spherical matching methods on a diverse set of datasets. | ['Charu Sharma', 'Manohar Kaul'] | 2020-07-05 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/6515_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123740052.pdf | eccv-2020-8 | ['hypergraph-matching'] | ['graphs'] | [ 6.05399430e-01 1.17298573e-01 1.53612718e-01 -1.29889891e-01
-6.99769497e-01 -8.64933431e-01 6.59714580e-01 -2.77187824e-01
2.80714612e-02 4.01687920e-01 1.30191207e-01 -2.33501121e-02
-4.47503775e-01 -7.76372671e-01 -7.97578573e-01 -5.07220507e-01
4.15492021e-02 6.68430090e-01 1.92799926e-01 -3.96777630e-01
3.40292424e-01 5.64390302e-01 -1.57253182e+00 -1.79834634e-01
9.37997878e-01 4.64098573e-01 4.43536341e-02 6.01659358e-01
2.03077912e-01 -2.13888437e-01 1.70584172e-02 -7.71141648e-01
6.39565289e-01 -2.47645646e-01 -7.32136607e-01 3.87516230e-01
1.21626735e+00 8.19387734e-02 -6.18158579e-01 1.36829090e+00
1.96753964e-01 1.40541032e-01 5.39151549e-01 -1.53511751e+00
-6.93525314e-01 2.84935571e-02 -5.56798577e-01 -2.74761438e-01
6.41842604e-01 -2.89448231e-01 9.60556209e-01 -1.25181437e+00
9.51118290e-01 1.30979466e+00 6.75621986e-01 2.68637687e-01
-1.45316100e+00 -4.00167048e-01 -4.15749490e-01 5.50055355e-02
-1.54689753e+00 -3.32286149e-01 8.92042160e-01 -5.94004273e-01
6.96128786e-01 7.17634201e-01 6.65650725e-01 5.71704388e-01
1.71234399e-01 1.20121032e-01 1.05807257e+00 -3.66945326e-01
-2.23557889e-01 -1.05689876e-01 -2.61198103e-01 8.80164266e-01
4.84530658e-01 3.77598926e-02 -3.77742589e-01 -4.01030660e-01
1.15017474e+00 9.76019278e-02 -5.30766308e-01 -1.22177708e+00
-1.63361335e+00 7.74935246e-01 4.69562680e-01 1.00515366e-01
-2.90781975e-01 -2.37437487e-01 -1.65188193e-01 6.85253143e-02
3.41729820e-01 5.85464478e-01 9.04806182e-02 2.00496078e-01
-6.54841900e-01 3.07663620e-01 6.59523487e-01 1.17182744e+00
1.04833853e+00 -2.24014878e-01 4.06275988e-01 7.28352189e-01
1.58842534e-01 6.39074981e-01 -5.54262809e-02 -8.92020106e-01
5.88169992e-01 9.85161960e-01 3.49651165e-02 -2.01703095e+00
-6.19002044e-01 -2.00929776e-01 -9.26218987e-01 1.04866937e-01
4.60969716e-01 4.83745635e-01 -6.56397879e-01 1.59774256e+00
3.73638034e-01 3.13149542e-01 -4.98222522e-02 1.12687469e+00
5.40753365e-01 3.80769372e-01 -5.96493721e-01 -4.07414548e-02
1.13650179e+00 -7.06281781e-01 -2.88976192e-01 -9.05965269e-03
2.37805709e-01 -1.08974385e+00 8.40730071e-01 5.93969263e-02
-1.19809461e+00 -2.30924606e-01 -1.16438770e+00 -1.40846791e-02
-1.20645419e-01 -2.17755437e-01 2.58568525e-01 4.95474458e-01
-1.12210512e+00 6.99018359e-01 -6.04705334e-01 -6.15547776e-01
-5.78304268e-02 4.25021261e-01 -8.74454200e-01 -2.99458861e-01
-8.37599456e-01 8.79357219e-01 -1.28052205e-01 -2.80019976e-02
-2.84014463e-01 -8.12265754e-01 -1.02347255e+00 -3.30029786e-01
3.55426550e-01 -8.43039989e-01 5.26100934e-01 -5.82681715e-01
-1.22131228e+00 1.36048448e+00 4.06230688e-02 7.99489170e-02
5.36589622e-01 2.76365638e-01 -3.06136966e-01 -1.54788988e-02
8.68510753e-02 5.14347494e-01 8.07113826e-01 -1.45164108e+00
-3.50942403e-01 -7.39497304e-01 2.17555046e-01 5.67250729e-01
-2.89753556e-01 -3.00175726e-01 -4.63778675e-01 -5.17745376e-01
7.26850450e-01 -1.42350876e+00 -1.53382182e-01 1.01853237e-01
-4.66543674e-01 1.26371697e-01 7.37464726e-01 -7.31981039e-01
8.92983556e-01 -2.00738645e+00 8.36636007e-01 4.02513802e-01
3.22213352e-01 -5.48726544e-02 -2.52975047e-01 4.39942509e-01
-2.56026089e-01 -9.45573449e-02 -4.47593987e-01 -1.08442508e-01
-2.08714642e-02 3.37722719e-01 -9.26915035e-02 1.02594709e+00
-5.16285822e-02 7.31356680e-01 -1.12047267e+00 -4.67173249e-01
2.39005253e-01 6.07123017e-01 -4.90207851e-01 3.28428783e-02
3.65304798e-01 5.66501260e-01 -3.29559445e-02 5.76263905e-01
9.78934646e-01 -1.18210092e-01 -1.01593494e-01 -6.28065526e-01
-1.40974239e-01 -1.81487903e-01 -1.38345337e+00 1.86434746e+00
-1.32628039e-01 4.96864140e-01 1.69163734e-01 -1.07866955e+00
1.08018672e+00 9.82720405e-02 8.01635146e-01 -4.53112870e-01
-3.66894752e-02 3.52267385e-01 -2.59713858e-01 -2.65578717e-01
6.65677547e-01 -2.08048169e-02 -4.59059067e-02 8.60314071e-02
-2.25336671e-01 -5.90738297e-01 -7.54109100e-02 3.85523140e-02
8.59391332e-01 6.79121464e-02 1.90698028e-01 -6.11372054e-01
6.19939983e-01 1.64671808e-01 3.50862384e-01 5.76920435e-02
1.50388509e-01 1.06186271e+00 5.01045818e-03 -5.41293979e-01
-1.39968526e+00 -1.24289131e+00 -3.50324810e-01 4.72282708e-01
6.18135095e-01 -1.85686216e-01 -9.19150710e-01 -6.98809922e-02
9.35327485e-02 2.02286601e-01 -4.24008429e-01 -1.53981820e-01
-7.63774216e-01 -6.33683205e-01 3.34428728e-01 5.26954681e-02
2.81791508e-01 -3.53215396e-01 -3.80390793e-01 5.99393621e-02
-4.41464305e-01 -1.28877223e+00 -9.12094712e-01 -7.34489620e-01
-8.75532269e-01 -1.26639152e+00 -8.15283656e-01 -9.61035550e-01
1.10254359e+00 7.89440036e-01 1.03605402e+00 1.81816891e-02
-4.23491359e-01 6.56121135e-01 -4.58236784e-02 5.86780123e-02
-1.85054809e-01 -2.14522243e-01 3.96241486e-01 3.99696618e-01
-5.83033636e-02 -6.85054302e-01 -6.38468981e-01 8.94324541e-01
-9.18537199e-01 3.04472148e-01 2.42410198e-01 6.96164012e-01
7.29232788e-01 -2.35539451e-01 1.67203490e-02 -3.40767562e-01
4.53653485e-01 -3.45674813e-01 -9.44009185e-01 3.13358605e-01
-3.63897532e-01 4.48651016e-02 3.66838813e-01 -2.69376278e-01
-5.27913332e-01 2.92347610e-01 5.01320720e-01 -5.28290749e-01
4.18406069e-01 3.60426575e-01 -1.34529084e-01 -8.29933643e-01
6.81974769e-01 1.79703414e-01 2.67863512e-01 -2.10551724e-01
4.02357101e-01 3.17461610e-01 1.01255953e+00 -3.72816265e-01
1.21817780e+00 7.98043132e-01 6.62353992e-01 -9.61628795e-01
-4.17575806e-01 -5.24237275e-01 -9.19720173e-01 -3.70672226e-01
7.00685024e-01 -5.67534268e-01 -5.21012902e-01 3.60126913e-01
-1.14939451e+00 2.75926054e-01 8.75532329e-02 4.09009993e-01
-9.35857534e-01 8.04679036e-01 -9.46974829e-02 -3.89697909e-01
-1.74646944e-01 -1.22805071e+00 1.14345336e+00 3.89695987e-02
-1.53708190e-01 -1.09635067e+00 3.69170159e-01 5.98224282e-01
2.32639492e-01 7.19841659e-01 8.38939488e-01 -1.06875010e-01
-8.27018857e-01 -3.57387155e-01 -2.60409653e-01 -1.13812163e-01
2.67849624e-01 2.65930407e-03 -4.11942691e-01 -5.02801776e-01
-1.51371300e-01 1.58903152e-01 1.71934873e-01 1.91320270e-01
5.89389384e-01 -4.05055225e-01 -4.42096293e-01 9.38618898e-01
1.46409082e+00 -2.90472806e-01 6.83929920e-01 2.57580042e-01
9.20202732e-01 8.02220106e-01 5.86565614e-01 1.41559035e-01
5.62242448e-01 1.17114520e+00 7.49598086e-01 -2.02684075e-01
-1.73250929e-01 -3.45476627e-01 2.00292878e-02 1.13649821e+00
-4.76703763e-01 1.21860720e-01 -1.02531528e+00 6.90185785e-01
-1.88748479e+00 -7.37133503e-01 -4.21074569e-01 2.77716255e+00
4.91081059e-01 -4.47037399e-01 2.39924192e-02 5.40069453e-02
9.73929644e-01 -1.53044730e-01 -3.68233025e-01 1.26932636e-02
-3.77101123e-01 -2.77369656e-02 6.66904449e-01 7.93513656e-01
-8.98927331e-01 5.66037297e-01 6.10619164e+00 2.98353016e-01
-9.82081234e-01 -1.94796875e-01 5.42649813e-03 3.13510329e-01
-5.20707786e-01 1.14487767e-01 -4.06037927e-01 2.00093955e-01
3.81727815e-01 -5.88026166e-01 8.24077904e-01 5.45258045e-01
-2.58740366e-01 2.63190836e-01 -1.16750765e+00 1.42840385e+00
4.98255372e-01 -1.34038663e+00 2.72861898e-01 2.91594028e-01
1.11043715e+00 -7.46778101e-02 1.60778552e-01 -3.65769267e-01
-3.05287018e-02 -1.13518679e+00 4.01427627e-01 6.06987715e-01
9.65502203e-01 -6.56460285e-01 4.40666556e-01 2.24882677e-01
-1.35196328e+00 4.58496690e-01 -6.07365489e-01 9.01141539e-02
2.04097420e-01 2.16674447e-01 -8.99520278e-01 8.77644777e-01
4.74266529e-01 5.28488994e-01 -3.18265855e-01 1.15157008e+00
3.08062077e-01 -2.28159413e-01 -3.74219835e-01 2.09308147e-01
7.89724812e-02 -9.31857049e-01 9.39314723e-01 7.99901605e-01
6.36820912e-01 4.56582904e-01 7.49618262e-02 6.94891691e-01
2.76107229e-02 2.35630080e-01 -1.00482965e+00 2.71438897e-01
3.32008541e-01 1.53925622e+00 -6.69698358e-01 -4.60418165e-02
-5.90364218e-01 1.00574172e+00 1.42193437e-01 2.41458803e-01
-7.17977166e-01 -1.38838828e-01 9.02279615e-01 4.04953659e-01
-1.32042795e-01 -5.42826533e-01 -2.07541853e-01 -1.30789685e+00
2.10502446e-01 -7.79857397e-01 1.99834779e-01 -7.81355381e-01
-1.20494127e+00 5.66546559e-01 3.27072516e-02 -1.42286468e+00
-5.17460667e-02 -6.36974633e-01 -2.84038544e-01 7.84469545e-01
-1.09635556e+00 -1.13837564e+00 -8.00158620e-01 9.15970743e-01
8.96033496e-02 1.07901700e-01 8.89626622e-01 3.95462096e-01
2.75434218e-02 4.13965881e-01 1.31667748e-01 -1.97399519e-02
5.90420723e-01 -9.79761600e-01 5.32340884e-01 8.56701136e-01
2.99420297e-01 3.95081878e-01 8.18954706e-01 -4.12782758e-01
-1.98416376e+00 -1.01315045e+00 7.53650188e-01 -7.87119269e-01
7.13622570e-01 -3.31469625e-01 -9.91832376e-01 7.25640893e-01
-1.06951758e-01 1.74092561e-01 2.79245079e-01 -4.63152498e-01
-3.77992928e-01 5.52428328e-02 -1.15246940e+00 7.90900946e-01
1.54510021e+00 -4.76547003e-01 -4.50296402e-01 5.03467798e-01
5.03153086e-01 -7.19917595e-01 -1.16464722e+00 6.26224279e-01
5.31756639e-01 -6.87612355e-01 1.38219905e+00 -4.62436914e-01
1.46260351e-01 -4.18281913e-01 -3.82916272e-01 -1.44631255e+00
-4.29613113e-01 -9.63314772e-01 2.90621430e-01 6.83501542e-01
1.04065083e-01 -6.95829988e-01 9.27868307e-01 6.33949399e-01
-1.70247763e-01 -5.34980237e-01 -1.16279078e+00 -9.32635069e-01
-1.85740322e-01 1.66305706e-01 4.90636796e-01 1.31241846e+00
3.27838063e-01 2.01156884e-01 -4.69472021e-01 4.50404108e-01
1.13461685e+00 2.67599821e-01 1.03204894e+00 -1.42555535e+00
1.42750815e-01 -3.68829131e-01 -1.11958420e+00 -9.83632803e-01
2.54766457e-02 -1.17574596e+00 -5.37864715e-02 -1.28489566e+00
2.22771838e-01 -5.51016092e-01 2.82845497e-01 1.60958871e-01
8.88824463e-02 5.62693119e-01 1.25796989e-01 3.30156326e-01
-1.76330984e-01 4.07437116e-01 1.49715066e+00 -6.83249682e-02
5.42519912e-02 -2.42558479e-01 -3.50893080e-01 7.14705467e-01
4.85655248e-01 -2.00296625e-01 -2.91547954e-01 -5.06023884e-01
5.50062537e-01 2.72890955e-01 5.36567330e-01 -8.71944308e-01
5.56572676e-01 -3.75302911e-01 -2.65095443e-01 -4.28880990e-01
5.37367523e-01 -8.70147943e-01 7.72636175e-01 2.65914798e-01
2.17026353e-01 4.84564155e-01 8.50043595e-02 4.61533070e-01
-1.67256862e-01 5.67129776e-02 8.64609063e-01 5.77878207e-02
-3.35170865e-01 6.67439103e-01 4.71803397e-01 2.45518684e-02
1.05616283e+00 -5.95443845e-01 -2.60216892e-01 -3.97691995e-01
-5.03334820e-01 -3.96465659e-02 1.17895031e+00 5.90516090e-01
8.55830550e-01 -1.64191520e+00 -9.91179347e-01 3.38884175e-01
3.75480354e-01 1.88522220e-01 6.93571754e-04 9.43041623e-01
-7.34619915e-01 5.00401318e-01 -5.50698876e-01 -9.23797429e-01
-1.51441479e+00 5.63825965e-01 3.42261195e-01 1.18763283e-01
-6.13938451e-01 5.50490379e-01 3.73809367e-01 -8.50053549e-01
-1.76518872e-01 -5.90450019e-02 8.03703591e-02 -3.04295659e-01
9.05796587e-02 6.95357442e-01 1.99027210e-01 -1.22664380e+00
-3.91146064e-01 1.39540362e+00 4.57131326e-01 -1.59108378e-02
1.26253676e+00 -8.53303149e-02 -4.87999231e-01 -7.64744133e-02
1.44243371e+00 6.95252940e-02 -1.08998060e+00 -3.93294156e-01
-1.71952069e-01 -8.79173756e-01 -2.43980616e-01 -2.38795113e-03
-9.76769209e-01 4.85312432e-01 2.53944367e-01 3.17576557e-01
8.01644981e-01 9.28288177e-02 6.93624973e-01 8.73676389e-02
6.45005047e-01 -5.74129462e-01 9.90243033e-02 2.29095280e-01
1.30086684e+00 -1.10160494e+00 1.03661761e-01 -9.86706257e-01
-1.60746679e-01 1.17816925e+00 3.19151759e-01 -4.05833572e-01
4.48096722e-01 -9.28457975e-02 -1.39072269e-01 -2.54674345e-01
-2.77620517e-02 2.13560566e-01 8.20771098e-01 7.44831860e-01
2.72865929e-02 2.36481413e-01 -3.50082815e-01 -1.21770710e-01
-5.91959715e-01 -4.25650716e-01 7.01114118e-01 5.68482518e-01
-2.67387748e-01 -9.42921638e-01 -8.61598492e-01 1.23228148e-01
2.14618519e-02 2.03468785e-01 -4.99117196e-01 7.18840420e-01
-2.94700086e-01 7.92948723e-01 1.47683978e-01 -4.46203351e-01
7.18014717e-01 -4.55982268e-01 8.53834510e-01 -3.21199000e-01
1.30154654e-01 -3.24158967e-01 -1.24405995e-01 -7.34035134e-01
-5.73882699e-01 -8.59181166e-01 -1.00775170e+00 -5.15168309e-01
-2.42224187e-01 -1.45178765e-01 8.71661663e-01 5.61382115e-01
4.24973547e-01 -1.65900975e-01 5.86245418e-01 -1.25674987e+00
-6.21416748e-01 -3.78414541e-01 -3.66916478e-01 8.68154466e-01
2.02238992e-01 -9.66502666e-01 -4.15941060e-01 5.89172654e-02] | [8.013882637023926, -2.293696403503418] |
bc0f23e3-9f81-430d-b119-be0a69c65b5c | audiopalm-a-large-language-model-that-can | 2306.12925 | null | https://arxiv.org/abs/2306.12925v1 | https://arxiv.org/pdf/2306.12925v1.pdf | AudioPaLM: A Large Language Model That Can Speak and Listen | We introduce AudioPaLM, a large language model for speech understanding and generation. AudioPaLM fuses text-based and speech-based language models, PaLM-2 [Anil et al., 2023] and AudioLM [Borsos et al., 2022], into a unified multimodal architecture that can process and generate text and speech with applications including speech recognition and speech-to-speech translation. AudioPaLM inherits the capability to preserve paralinguistic information such as speaker identity and intonation from AudioLM and the linguistic knowledge present only in text large language models such as PaLM-2. We demonstrate that initializing AudioPaLM with the weights of a text-only large language model improves speech processing, successfully leveraging the larger quantity of text training data used in pretraining to assist with the speech tasks. The resulting model significantly outperforms existing systems for speech translation tasks and has the ability to perform zero-shot speech-to-text translation for many languages for which input/target language combinations were not seen in training. AudioPaLM also demonstrates features of audio language models, such as transferring a voice across languages based on a short spoken prompt. We release examples of our method at https://google-research.github.io/seanet/audiopalm/examples | ['Christian Frank', 'Lukas Zilka', 'Zhishuai Zhang', 'Yu Zhang', 'Neil Zeghidour', 'Vicky Zayats', 'Yongqiang Wang', 'Jiahui Yu', 'Damien Vincent', 'Mihajlo Velimirović', 'Alexandru Tudor', 'Marco Tagliasacchi', 'Ramanovich', 'Michelle Tadmor', 'Matt Sharifi', 'Johan Schalkwyk', 'Tara Sainath', 'Danny Rozenberg', 'James Qin', 'Dirk Padfield', 'Hannah Muckenhirn', 'Eugene Kharitonov', 'Wei Han', 'Dalia El Badawy', 'Peter Chen', 'Félix de Chaumont Quitry', 'Zalán Borsos', 'Ankur Bapna', 'Duc Dung Nguyen', 'Chulayuth Asawaroengchai', 'Paul K. Rubenstein'] | 2023-06-22 | null | null | null | null | ['speech-to-text-translation', 'speech-to-speech-translation'] | ['natural-language-processing', 'speech'] | [ 1.74440518e-01 3.38723689e-01 6.13451190e-02 -3.94079268e-01
-1.44934011e+00 -7.87597656e-01 6.82931602e-01 -2.61789203e-01
-1.87428653e-01 3.65464091e-01 6.68318987e-01 -6.50073886e-01
4.45126563e-01 -3.64885300e-01 -7.10839331e-01 -2.38300711e-01
2.79198468e-01 6.78683281e-01 -1.01061255e-01 -5.09953678e-01
-3.33313346e-01 2.46374141e-02 -1.31568491e+00 7.63568163e-01
5.14131367e-01 5.31775177e-01 5.22893488e-01 1.12933671e+00
-3.98510516e-01 5.53246081e-01 -6.61045909e-01 -3.08488458e-01
-4.80616912e-02 -4.76540893e-01 -8.99573326e-01 -2.06848875e-01
4.09230471e-01 -3.12599450e-01 -3.86371136e-01 5.14333844e-01
8.60015810e-01 1.77267417e-01 4.03743654e-01 -1.18600845e+00
-6.36227429e-01 1.12167513e+00 3.13960202e-02 -9.35501903e-02
6.32195413e-01 2.10206866e-01 9.95388567e-01 -1.42471528e+00
3.18641007e-01 1.75943267e+00 4.52577442e-01 9.06577587e-01
-1.18702543e+00 -6.54717565e-01 -1.81991726e-01 -1.26043499e-01
-1.35806274e+00 -1.43137300e+00 4.51939702e-01 -1.19391672e-01
1.39743567e+00 4.82802719e-01 3.17696720e-01 1.46104443e+00
-1.43286422e-01 1.09014869e+00 7.05562174e-01 -7.15033889e-01
-4.83313464e-02 1.29636467e-01 -3.49928528e-01 4.94151741e-01
-7.30468869e-01 2.49827281e-01 -1.21731925e+00 -1.24761559e-01
5.54572821e-01 -6.54914200e-01 -3.06168735e-01 4.14661765e-01
-1.52887559e+00 6.07810676e-01 -1.61856532e-01 2.49608979e-01
-8.35735351e-02 1.69509202e-01 5.93130231e-01 6.94835007e-01
3.02578509e-01 2.73239940e-01 -4.97206062e-01 -3.75562191e-01
-1.10600340e+00 -1.33189484e-01 8.32582116e-01 1.20475817e+00
4.55761194e-01 6.89745903e-01 -2.54948050e-01 1.25894761e+00
5.94576657e-01 1.08309209e+00 8.96410882e-01 -9.47861671e-01
8.25664937e-01 -2.73686685e-02 -2.07904488e-01 -2.02440262e-01
-1.13297187e-01 -4.26016636e-02 -5.52393913e-01 -2.64199644e-01
1.25501260e-01 -4.00454462e-01 -1.02364874e+00 1.82653153e+00
9.21434760e-02 1.16693534e-01 6.84337914e-01 4.66005176e-01
1.09555423e+00 1.32060587e+00 -5.63880848e-03 -1.33912951e-01
1.24907911e+00 -1.22959197e+00 -7.19216645e-01 -5.76031625e-01
6.88073933e-01 -1.30589807e+00 1.39499605e+00 1.00274131e-01
-1.39732087e+00 -7.11938858e-01 -6.88045323e-01 -2.39294261e-01
-2.80289114e-01 3.00414145e-01 1.93532705e-01 7.66457975e-01
-1.62536132e+00 -9.00818035e-02 -8.46121669e-01 -4.81663913e-01
-1.08926266e-01 2.87614703e-01 -4.06007856e-01 6.75873607e-02
-1.45909119e+00 6.60865307e-01 3.94008219e-01 9.00632739e-02
-1.06474078e+00 -5.48468947e-01 -1.26179147e+00 -8.29696935e-03
6.42535985e-02 -6.86853886e-01 1.91304731e+00 -8.61472607e-01
-1.94509828e+00 6.00782454e-01 -5.46077013e-01 -4.36044723e-01
1.55426532e-01 -9.82190073e-02 -6.86311960e-01 2.00117946e-01
-1.90474857e-02 1.15860796e+00 9.75285530e-01 -9.89939511e-01
-4.29135650e-01 6.84583336e-02 -4.68957126e-01 4.88140702e-01
-4.59438205e-01 4.21270907e-01 -4.94170129e-01 -7.29265869e-01
-1.28331318e-01 -8.88251185e-01 1.74058989e-01 -2.84824461e-01
-5.25899708e-01 -8.91917273e-02 7.80520499e-01 -9.35818195e-01
1.17784560e+00 -2.00874543e+00 7.25109726e-02 6.25570193e-02
-3.76052916e-01 5.34278035e-01 -7.37737238e-01 1.04924595e+00
8.86670947e-02 1.93507031e-01 -1.64697781e-01 -8.65047157e-01
1.33364365e-01 2.72159725e-01 -7.26476729e-01 -1.25265703e-01
1.98835939e-01 1.14065695e+00 -7.39097416e-01 -3.34346265e-01
4.36394453e-01 7.64500022e-01 -3.54893118e-01 2.47454330e-01
-2.23008260e-01 4.33214903e-01 3.10315490e-02 5.23105919e-01
1.62265480e-01 2.34941542e-01 -1.64948963e-03 1.46498293e-01
-2.76129432e-02 9.70331669e-01 -8.38305950e-01 1.82997417e+00
-9.78486359e-01 6.97336674e-01 4.95842189e-01 -3.69197816e-01
8.29057455e-01 1.18894017e+00 1.65392146e-01 -5.46834111e-01
2.73723770e-02 5.20379305e-01 -6.00331798e-02 -2.63209939e-01
6.37514651e-01 -2.26105556e-01 -1.92881614e-01 5.73747158e-01
3.99220467e-01 -7.94492602e-01 2.83114631e-02 4.04482633e-01
7.93303728e-01 -2.45303154e-01 -1.04206726e-02 7.86041915e-02
5.43363094e-01 -2.44742408e-01 -8.99476111e-02 6.63285136e-01
1.21810831e-01 7.53531277e-01 -1.14070788e-01 3.94906074e-01
-9.60168183e-01 -1.22448719e+00 2.36049473e-01 1.53128386e+00
-5.37371397e-01 -7.65104949e-01 -8.12110960e-01 -1.60663173e-01
-2.55738825e-01 1.10335553e+00 -1.13156646e-04 -3.09307486e-01
-4.69821662e-01 -1.19266272e-01 1.35509670e+00 4.51734632e-01
4.31760289e-02 -1.34685349e+00 2.03823969e-01 3.52571636e-01
-5.43628335e-01 -1.25826418e+00 -1.09134853e+00 -7.16989934e-02
-5.96321166e-01 -3.33159864e-01 -8.57497513e-01 -1.22368181e+00
2.30630666e-01 1.49162933e-01 1.04077220e+00 -2.44500428e-01
7.96649307e-02 6.88230097e-01 -3.57746631e-01 -4.32634652e-01
-1.36758828e+00 1.77944601e-01 3.53875756e-01 -6.53912425e-02
7.54953921e-03 -4.58901882e-01 9.01352391e-02 2.72394806e-01
-8.10779989e-01 4.04101461e-01 5.03470480e-01 8.73669147e-01
1.95512161e-01 -4.52294409e-01 9.04866099e-01 -3.50465536e-01
8.07156503e-01 -1.93105370e-01 -2.08683506e-01 1.92277074e-01
-8.39906335e-02 -1.02156229e-01 7.19730377e-01 -5.73204398e-01
-9.92956877e-01 2.88427360e-02 -7.74448216e-01 -3.88993889e-01
-1.09895997e-01 6.45030975e-01 -3.06655943e-01 2.94601053e-01
5.18962562e-01 6.34998620e-01 1.62573799e-01 -5.40821612e-01
6.31569684e-01 1.32626963e+00 7.42512763e-01 -5.82708955e-01
7.13125408e-01 -1.14168733e-01 -7.32095242e-01 -1.21228302e+00
-4.24325585e-01 -5.48909247e-01 -4.71526057e-01 -8.73482972e-02
5.81244826e-01 -1.22827351e+00 -4.57619011e-01 6.99207723e-01
-1.52425206e+00 -6.16448998e-01 -3.66371810e-01 5.45934916e-01
-7.52015293e-01 1.97344050e-01 -8.60974610e-01 -9.69166696e-01
-6.21569574e-01 -1.12129664e+00 1.46436131e+00 -1.99384689e-01
-4.91097838e-01 -1.08108032e+00 -1.42163271e-02 5.30140638e-01
5.85691631e-01 -7.26255774e-01 7.97825098e-01 -7.73347437e-01
-1.93036929e-01 8.17975923e-02 2.69551516e-01 5.92488348e-01
3.23393941e-01 -1.85344636e-01 -1.27096093e+00 -4.98599082e-01
-3.05738717e-01 -6.62728429e-01 6.84793115e-01 1.63227782e-01
4.75894183e-01 -5.58048546e-01 7.46160224e-02 3.72128516e-01
5.69356501e-01 1.57643542e-01 4.07820016e-01 -3.53617489e-01
7.77631521e-01 7.23501384e-01 1.33781880e-01 -3.01612541e-03
5.64166069e-01 7.12078989e-01 -8.14730003e-02 -1.42232671e-01
-6.63183928e-01 -7.06137538e-01 1.27955055e+00 1.72951138e+00
6.47532880e-01 -5.77161551e-01 -1.04252493e+00 7.71287441e-01
-1.56036639e+00 -7.34059393e-01 8.49476084e-02 2.14188743e+00
1.29216778e+00 -1.64900914e-01 -3.67130227e-02 -1.15415432e-01
6.96857750e-01 2.04336300e-01 -2.23141700e-01 -5.76530814e-01
-2.17760965e-01 2.36548364e-01 8.41073915e-02 1.14006209e+00
-7.54776180e-01 1.51163745e+00 6.40138245e+00 1.07841432e+00
-1.22864950e+00 2.45174795e-01 2.01658875e-01 -1.83946505e-01
-6.47324681e-01 -1.19565263e-01 -8.94267738e-01 2.37660304e-01
1.68223357e+00 -4.81897980e-01 7.41453648e-01 1.75149858e-01
5.88421702e-01 4.18088377e-01 -1.32733202e+00 1.07086301e+00
2.38980576e-01 -1.14289129e+00 5.41094184e-01 -3.12851280e-01
4.16761935e-01 4.45946902e-01 1.59746692e-01 4.66008693e-01
4.06348467e-01 -1.21031988e+00 1.05883265e+00 1.14972621e-01
1.25021279e+00 -4.90221113e-01 3.52235287e-01 4.51048553e-01
-1.27592993e+00 1.96375877e-01 5.09789810e-02 2.44808242e-01
6.67445421e-01 -1.29470332e-02 -1.40474749e+00 3.96224260e-01
2.59045213e-01 4.71065402e-01 -3.49769443e-01 5.72064936e-01
-3.06859672e-01 1.25002265e+00 -6.09108746e-01 1.96226284e-01
2.13888034e-01 2.30030984e-01 8.67452502e-01 1.39785182e+00
6.99602246e-01 -2.78000325e-01 3.41360807e-01 6.51995003e-01
-1.96647599e-01 3.82775992e-01 -7.03837693e-01 -5.95821202e-01
8.26898932e-01 8.75041246e-01 -1.28410563e-01 -4.29539382e-01
-3.30683380e-01 1.12628520e+00 -1.83536008e-01 5.53363085e-01
-2.53609926e-01 -2.96031684e-01 5.59662342e-01 -4.28314991e-02
2.72144796e-03 -5.33250988e-01 3.45904939e-02 -9.68013585e-01
-9.84002277e-02 -1.19170904e+00 1.13080926e-01 -1.16607940e+00
-1.10200775e+00 9.88483429e-01 -1.87426224e-01 -1.06731498e+00
-9.17616367e-01 -5.07652223e-01 -5.83299696e-01 1.33332372e+00
-1.26367080e+00 -1.71031451e+00 3.16720992e-01 6.79920673e-01
1.16052747e+00 -5.86384475e-01 1.16868567e+00 2.70071507e-01
-3.09788913e-01 7.56551802e-01 1.83703285e-02 3.49544019e-01
9.84947324e-01 -9.68143046e-01 1.07864332e+00 7.38382041e-01
5.13008356e-01 6.80964231e-01 5.52648187e-01 -5.76387167e-01
-1.45043576e+00 -1.26286232e+00 1.26779878e+00 -5.18085003e-01
7.99628198e-01 -8.03280234e-01 -8.85190666e-01 9.11397934e-01
6.07149661e-01 -4.72200125e-01 7.87835181e-01 -1.48539394e-01
-2.34518185e-01 -2.03923564e-02 -6.85191810e-01 8.59345257e-01
7.48763621e-01 -1.16130650e+00 -5.36881149e-01 2.98466235e-01
1.32871878e+00 -4.62788373e-01 -6.44997835e-01 -5.43579124e-02
4.52458650e-01 -2.97479272e-01 8.27237189e-01 -4.37516749e-01
3.03321471e-03 -1.66247934e-01 -3.94016951e-01 -1.69032800e+00
2.75862455e-01 -1.24454033e+00 8.14793259e-02 1.56567001e+00
8.67067218e-01 -7.63017297e-01 1.43711209e-01 1.62116408e-01
-5.70835292e-01 -1.10170417e-01 -1.11561751e+00 -8.80389929e-01
2.32675746e-01 -8.78459036e-01 5.72690725e-01 6.25776350e-01
1.60648420e-01 7.02786565e-01 -4.43109244e-01 2.90781081e-01
1.11614466e-01 -4.02508378e-01 7.60098040e-01 -6.80504680e-01
-4.13235456e-01 -2.59343594e-01 9.24604461e-02 -1.27566516e+00
5.21637201e-01 -1.32429159e+00 3.78050387e-01 -1.48139203e+00
-3.90464693e-01 -2.71470308e-01 7.01482147e-02 1.03003848e+00
2.84179062e-01 2.48124197e-01 4.54057902e-01 1.29182324e-01
-1.91551477e-01 9.56901670e-01 1.12910092e+00 -2.44347408e-01
-4.68138844e-01 1.22227125e-01 -6.29863501e-01 4.53868419e-01
8.55617046e-01 -2.98477143e-01 -6.44145668e-01 -7.57959664e-01
-1.83775440e-01 5.21005034e-01 2.16009572e-01 -8.21577191e-01
3.91101241e-01 1.15960592e-03 -9.66270491e-02 -4.08931911e-01
8.32281232e-01 -3.76268238e-01 -7.81099722e-02 1.30127855e-02
-5.24888754e-01 1.23898506e-01 7.40820169e-01 -1.38267037e-02
-5.10926306e-01 -7.48810321e-02 4.23596084e-01 -2.84517854e-02
-5.04146874e-01 7.22617730e-02 -9.14673746e-01 7.13074729e-02
3.02337676e-01 3.76919098e-02 -3.90962809e-01 -1.01611555e+00
-7.20002472e-01 3.34310085e-01 1.01859935e-01 9.86419320e-01
7.84776628e-01 -1.35536253e+00 -1.02540648e+00 5.83198607e-01
1.14389896e-01 -1.74159646e-01 3.60909216e-02 7.06122100e-01
-1.34039402e-01 6.11370742e-01 2.16176987e-01 -5.73148251e-01
-1.43310595e+00 3.26776244e-02 3.66474271e-01 2.25096092e-01
-3.06632012e-01 8.45338464e-01 1.38010353e-01 -1.04017365e+00
2.42561713e-01 -2.88431913e-01 2.59899437e-01 -2.10140839e-01
6.21394753e-01 6.99607581e-02 1.98375002e-01 -1.04114902e+00
-3.57928067e-01 1.63170055e-01 -4.52793250e-03 -1.04852545e+00
1.03249669e+00 -5.29688060e-01 -6.39290363e-02 8.48665714e-01
1.00771475e+00 4.45346087e-01 -8.40922356e-01 -3.42564017e-01
-1.82876050e-01 1.36189446e-01 1.19201124e-01 -1.10991752e+00
-6.39864266e-01 1.29263663e+00 1.53750613e-01 -1.16766378e-01
8.45892131e-01 1.07781045e-01 1.29680872e+00 6.15839839e-01
1.81079283e-01 -1.15546942e+00 2.58790776e-02 1.15382230e+00
1.29616642e+00 -1.03807390e+00 -8.44909072e-01 -2.87550837e-01
-8.36038709e-01 1.00664079e+00 4.07209158e-01 7.76336670e-01
4.16483045e-01 5.81132531e-01 6.69881284e-01 3.77611011e-01
-1.10297978e+00 -1.86792672e-01 4.82002050e-01 6.84968412e-01
6.20401025e-01 1.77987069e-01 3.82728517e-01 5.00518858e-01
-7.69840300e-01 -3.91455561e-01 4.33591545e-01 7.43879616e-01
-3.98557484e-01 -1.40736413e+00 -5.55353343e-01 4.02521528e-02
-4.94527072e-02 -7.74626493e-01 -6.06940508e-01 2.59385288e-01
-3.98916423e-01 1.41398787e+00 7.91131034e-02 -3.71073276e-01
2.40594745e-01 6.04114294e-01 1.82520047e-01 -1.08300567e+00
-5.82463920e-01 4.62272614e-01 3.90558749e-01 -2.86381096e-01
9.35431048e-02 -5.36974370e-01 -1.56689835e+00 -1.78157538e-01
-2.04878032e-01 2.37474009e-01 6.75602674e-01 8.22292984e-01
4.00729209e-01 3.82708490e-01 5.00329792e-01 -9.93545711e-01
-5.11084318e-01 -1.23461092e+00 -4.07733619e-01 -1.14192024e-01
7.17735112e-01 8.62217024e-02 -3.21116954e-01 4.14988190e-01] | [14.571606636047363, 6.965937614440918] |
4a209ad3-189e-4c45-b8d5-0e0b1b5400b8 | difficulty-aware-distractor-generation-for | null | null | https://aclanthology.org/U19-1021 | https://aclanthology.org/U19-1021.pdf | Difficulty-aware Distractor Generation for Gap-Fill Items | null | ['John Lee', 'Chak Yan Yeung', 'Benjamin Tsou'] | 2019-04-01 | null | null | null | alta-2019-4 | ['distractor-generation'] | ['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] | [-7.237696170806885, 3.719326972961426] |
9059ca66-8acc-411b-8632-231abcf401d1 | bev-dc-bird-s-eye-view-assisted-training-for | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhou_BEVDC_Birds-Eye_View_Assisted_Training_for_Depth_Completion_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhou_BEVDC_Birds-Eye_View_Assisted_Training_for_Depth_Completion_CVPR_2023_paper.pdf | BEV@DC: Bird's-Eye View Assisted Training for Depth Completion | Depth completion plays a crucial role in autonomous driving, in which cameras and LiDARs are two complementary sensors. Recent approaches attempt to exploit spatial geometric constraints hidden in LiDARs to enhance image-guided depth completion. However, only low efficiency and poor generalization can be achieved. In this paper, we propose BEV@DC, a more efficient and powerful multi-modal training scheme, to boost the performance of image-guided depth completion. In practice, the proposed BEV@DC model comprehensively takes advantage of LiDARs with rich geometric details in training, employing an enhanced depth completion manner in inference, which takes only images (RGB and depth) as input. Specifically, the geometric-aware LiDAR features are projected onto a unified BEV space, combining with RGB features to perform BEV completion. By equipping a newly proposed point-voxel spatial propagation network (PV-SPN), this auxiliary branch introduces strong guidance to the original image branches via 3D dense supervision and feature consistency. As a result, our baseline model demonstrates significant improvements with the sole image inputs. Concretely, it achieves state-of-the-art on several benchmarks, e.g., ranking Top-1 on the challenging KITTI depth completion benchmark. | ['Zhen Li', 'Shuguang Cui', 'Gangming Zhao', 'Jin Huang', 'Yuankai Lin', 'Yinghong Liao', 'Xu Yan', 'Wending Zhou'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['depth-completion'] | ['computer-vision'] | [ 7.83761889e-02 2.67123710e-02 -3.08485270e-01 -6.65876150e-01
-8.49169493e-01 -4.01648283e-01 7.19466209e-01 -9.43085998e-02
-5.40471137e-01 4.17728275e-01 -5.80441058e-02 -3.70917618e-01
1.36852726e-01 -1.03253567e+00 -1.04887676e+00 -5.91883540e-01
4.43560690e-01 4.57461029e-01 4.63050514e-01 -2.27025285e-01
2.80485541e-01 5.06573617e-01 -1.72552097e+00 -9.67635065e-02
1.16997433e+00 1.15018940e+00 4.82897669e-01 3.91871631e-01
-3.09471637e-01 5.27664840e-01 4.14606482e-02 -2.43524760e-01
3.20672065e-01 1.71086594e-01 -4.43619519e-01 3.96807818e-03
8.39936674e-01 -8.82963598e-01 -7.51871228e-01 8.49421620e-01
2.30201572e-01 2.15784952e-01 4.67181474e-01 -1.19606745e+00
-5.03746212e-01 1.65939867e-01 -8.97792518e-01 -9.56989601e-02
1.36260182e-01 4.49210972e-01 8.73033881e-01 -1.20793283e+00
4.48862463e-01 1.26788795e+00 3.48438382e-01 4.94775712e-01
-1.06534314e+00 -8.34787965e-01 5.22016585e-01 2.20292106e-01
-1.33996677e+00 -3.09325665e-01 9.69154894e-01 -2.77382404e-01
1.00895846e+00 -1.28514737e-01 6.16111636e-01 7.82045543e-01
-3.07741016e-02 8.28499675e-01 8.51373553e-01 4.37123701e-03
1.53249949e-01 -1.03680916e-01 1.55429458e-02 9.40680683e-01
1.97264493e-01 3.75427514e-01 -7.73283660e-01 3.58830929e-01
7.67153442e-01 4.43125337e-01 -1.95615798e-01 -4.44981694e-01
-1.09843290e+00 9.43675518e-01 1.04596579e+00 -3.72668833e-01
-2.45418712e-01 4.49861735e-01 1.18835099e-01 -2.40351528e-01
4.82246250e-01 -1.61555484e-02 -2.13338897e-01 5.23621663e-02
-8.00935566e-01 3.57943922e-01 6.34850860e-02 1.13746786e+00
1.40261745e+00 -1.96349710e-01 -1.01678245e-01 4.86073077e-01
5.42099893e-01 8.46127331e-01 -3.93237248e-02 -1.26247525e+00
1.07722521e+00 7.51815856e-01 7.59535516e-03 -7.90101707e-01
-2.48357654e-01 -3.66120875e-01 -8.92667353e-01 3.24643761e-01
1.85216635e-01 2.16538340e-01 -1.06355274e+00 1.56562853e+00
5.51085949e-01 5.04972517e-01 1.49895966e-01 9.95876431e-01
9.52809751e-01 7.28932679e-01 -1.22975394e-01 3.07356834e-01
1.07966483e+00 -1.21278059e+00 -3.03804845e-01 -7.22718120e-01
4.95798200e-01 -3.45657349e-01 1.11967814e+00 3.36205930e-01
-9.01789546e-01 -7.05243528e-01 -1.27218497e+00 -7.96761513e-01
-7.70858824e-02 5.65849133e-02 9.11683500e-01 1.74528956e-01
-9.79721844e-01 5.11735559e-01 -1.09156704e+00 7.52521977e-02
7.23430037e-01 2.42927670e-01 -3.49890530e-01 -8.54886174e-01
-8.75806808e-01 5.37863970e-01 9.08398107e-02 2.66686887e-01
-1.15737081e+00 -8.99603367e-01 -1.33400249e+00 -2.15694025e-01
4.02465820e-01 -9.06181455e-01 1.13633633e+00 1.66329682e-01
-1.34659493e+00 6.88796818e-01 -4.93380129e-01 -3.53493869e-01
5.56701660e-01 -2.80624866e-01 1.61537081e-01 3.51610750e-01
3.56167644e-01 1.25893474e+00 5.76441288e-01 -1.43005419e+00
-9.29542959e-01 -7.29910135e-01 3.22703093e-01 2.63376027e-01
-1.71769023e-01 -8.51447165e-01 -8.76355648e-01 -1.30630299e-01
5.20836353e-01 -7.49705076e-01 -4.25040781e-01 4.35896724e-01
-2.95272589e-01 -3.11441272e-01 7.60524750e-01 -3.93567443e-01
7.07142711e-01 -2.17549253e+00 2.91042060e-01 -2.14078203e-02
5.46303093e-01 -3.81746180e-02 -1.46934569e-01 1.34963796e-01
5.30159354e-01 -1.05456002e-01 -5.34096599e-01 -1.08632123e+00
1.29755303e-01 6.03748620e-01 -4.41593140e-01 4.44456160e-01
3.78206134e-01 1.13598490e+00 -8.97511601e-01 -4.84415054e-01
6.53563499e-01 7.69980907e-01 -8.13838840e-01 2.09378496e-01
-3.10027272e-01 6.85749710e-01 -6.72008455e-01 6.32657349e-01
1.12551546e+00 -7.79691786e-02 -4.49376762e-01 -2.27302909e-01
-2.77382463e-01 3.83047283e-01 -7.12747455e-01 2.43407559e+00
-8.51915717e-01 5.03093719e-01 3.86214070e-02 -7.40768492e-01
9.08125460e-01 -3.29260856e-01 2.17224076e-01 -8.88562560e-01
-9.83671322e-02 3.15000981e-01 -4.31988180e-01 -2.62094051e-01
5.36389828e-01 5.49760796e-02 2.93137580e-02 -3.93347087e-04
-1.26243368e-01 -7.44695485e-01 -1.92692116e-01 4.67935652e-01
9.63580132e-01 4.48493928e-01 -2.02436239e-01 1.59787208e-01
5.59892058e-01 3.67392302e-02 6.27377927e-01 4.02319044e-01
-1.05880097e-01 7.65947580e-01 1.77990004e-01 -7.20886737e-02
-8.02055836e-01 -1.22121739e+00 -2.24330723e-01 5.53320587e-01
7.64824927e-01 -3.08363676e-01 -4.40084547e-01 -5.12757838e-01
2.87509531e-01 5.42252719e-01 -4.22066182e-01 -2.51187086e-01
-5.50815225e-01 -2.29766309e-01 3.92217010e-01 7.26294577e-01
8.23734581e-01 -5.48001587e-01 -5.85078776e-01 1.19514771e-01
-2.31765077e-01 -1.54233909e+00 -3.50780427e-01 1.81395903e-01
-1.21986616e+00 -8.64523411e-01 -5.46136260e-01 -6.34378016e-01
5.20624340e-01 8.59445691e-01 9.80941057e-01 1.36559173e-01
-7.20399618e-02 7.37881064e-02 -2.42165864e-01 -2.48711303e-01
2.63258308e-01 6.60290718e-02 -2.49636009e-01 -2.27925926e-01
2.53182888e-01 -8.43113959e-01 -1.04928696e+00 1.64318308e-01
-7.70446777e-01 2.47778818e-01 7.46817648e-01 8.03412735e-01
8.53559077e-01 -1.35101140e-01 7.08491653e-02 -6.14172697e-01
-2.34492630e-01 -3.40817809e-01 -7.40056098e-01 -2.89249361e-01
-5.81499577e-01 5.40872514e-02 5.62091529e-01 1.68379903e-01
-1.00335956e+00 3.57833445e-01 -4.83957797e-01 -8.52798104e-01
-4.18816619e-02 3.72287720e-01 -4.67759103e-01 -2.67599672e-01
3.24890703e-01 3.39691788e-01 -6.47271276e-02 -5.22264004e-01
5.61540842e-01 4.46116000e-01 8.18099916e-01 -5.58767140e-01
1.10708165e+00 1.03918636e+00 2.76189297e-01 -5.93887389e-01
-9.90085125e-01 -6.76189065e-01 -6.54047668e-01 5.52052818e-02
7.44968057e-01 -1.38957977e+00 -6.38987124e-01 6.04034185e-01
-1.28740478e+00 -4.38699722e-01 -8.03741440e-03 5.72580457e-01
-3.88561070e-01 5.21660507e-01 -4.94473547e-01 -5.95031917e-01
-8.75532478e-02 -1.35657310e+00 1.72912204e+00 1.96959198e-01
6.02157652e-01 -6.76441491e-01 -2.43241861e-01 6.56324863e-01
1.26058627e-02 1.85334563e-01 6.43341005e-01 3.30553025e-01
-1.24989700e+00 -2.96625844e-03 -6.23567462e-01 2.79209614e-01
1.76672675e-02 -2.18308508e-01 -1.33311713e+00 -1.54405728e-01
-1.77894011e-01 -3.32580715e-01 1.40360534e+00 1.89927340e-01
1.33163786e+00 1.92680702e-01 -4.16106313e-01 1.27160621e+00
1.49564540e+00 -2.08118469e-01 6.38414860e-01 1.63048431e-01
1.24101841e+00 7.06731021e-01 9.42130208e-01 3.94725919e-01
9.90624368e-01 5.48266292e-01 1.06635845e+00 -1.71047151e-01
-1.74986988e-01 -6.70522571e-01 2.18100160e-01 5.26001811e-01
3.99487987e-02 -3.29322964e-02 -7.42357850e-01 4.50643092e-01
-1.81550229e+00 -5.09061396e-01 -3.51940036e-01 2.05526114e+00
5.13962150e-01 3.93245548e-01 -2.31988996e-01 5.03803007e-02
3.02213788e-01 4.21625227e-01 -9.47923660e-01 2.42977858e-01
-4.95438613e-02 3.28187317e-01 6.85756087e-01 7.75285602e-01
-9.57422972e-01 1.09281707e+00 4.59209919e+00 7.86079109e-01
-1.01596153e+00 1.99238732e-01 4.50536281e-01 -1.82926893e-01
-6.82902455e-01 3.62583213e-02 -1.01455534e+00 2.60100871e-01
3.70832264e-01 1.69980228e-01 3.64495337e-01 9.00839567e-01
3.35851222e-01 -3.12328279e-01 -1.16800487e+00 1.18288004e+00
-5.88716790e-02 -1.36471069e+00 1.08204208e-01 2.61573374e-01
6.74886107e-01 4.61654872e-01 2.03725025e-01 2.74088591e-01
9.04361308e-02 -8.44243824e-01 7.94901907e-01 2.25276545e-01
1.03004444e+00 -8.61128986e-01 6.45789981e-01 5.42814910e-01
-1.53240120e+00 -3.79046239e-02 -6.62791073e-01 -2.49438241e-01
4.05705750e-01 8.59518647e-01 -2.68546224e-01 8.45813334e-01
6.67859256e-01 1.17734516e+00 -4.53619361e-01 7.65989542e-01
-6.27121747e-01 2.51809210e-01 -4.68116850e-01 3.76625180e-01
5.41994989e-01 -2.73115993e-01 2.37325296e-01 8.07556391e-01
4.54436332e-01 1.28743753e-01 3.44823211e-01 1.00024056e+00
-1.22034997e-01 -2.79954910e-01 -6.84040487e-01 4.11746711e-01
5.33293724e-01 1.37629032e+00 -4.37195957e-01 -2.75922716e-01
-6.16319060e-01 1.00039184e+00 5.29962897e-01 3.11416149e-01
-9.07664359e-01 -2.77410477e-01 1.02400136e+00 1.68335021e-01
3.89595628e-01 -5.92896998e-01 -4.94499594e-01 -1.14056218e+00
3.67445648e-01 -1.73378482e-01 -1.57569647e-02 -7.99028456e-01
-1.02953124e+00 4.04357016e-01 -8.15713704e-02 -1.20349169e+00
3.86686102e-02 -6.50648832e-01 -5.05792022e-01 1.12887907e+00
-2.38865328e+00 -1.38800931e+00 -1.00944602e+00 7.42384851e-01
4.76208210e-01 3.40840459e-01 2.05737934e-01 5.25890648e-01
-5.08568645e-01 3.81983757e-01 -2.83328980e-01 -1.26315221e-01
5.96278608e-01 -1.25391459e+00 6.18758261e-01 8.49107623e-01
-8.51118043e-02 4.36135709e-01 1.37293592e-01 -6.19586587e-01
-1.70818210e+00 -1.59513891e+00 5.08278191e-01 -4.65247989e-01
3.50381136e-01 -4.25558925e-01 -7.91532338e-01 5.07797301e-01
-2.38227442e-01 3.23887765e-01 2.86020972e-02 -2.45690644e-01
-4.98732179e-01 -3.97075593e-01 -8.82434607e-01 4.87776488e-01
1.50085199e+00 -7.27050424e-01 -3.46871674e-01 2.94394642e-01
1.20477581e+00 -7.17931807e-01 -5.19391179e-01 5.66306829e-01
2.42345050e-01 -1.16362667e+00 1.20893955e+00 8.55310820e-03
8.59430671e-01 -5.84578931e-01 -3.69115621e-01 -9.87146080e-01
-2.71620844e-02 -3.29337239e-01 -2.29521021e-01 1.08382607e+00
1.52173951e-01 -6.93902433e-01 1.01698196e+00 3.16672057e-01
-7.52213776e-01 -9.00282204e-01 -9.21700954e-01 -5.79691529e-01
1.46361306e-01 -8.34678411e-01 7.59248435e-01 4.16707724e-01
-5.04882634e-01 4.01038855e-01 -2.30709374e-01 4.39137757e-01
8.89443398e-01 2.31957495e-01 1.06906295e+00 -1.14509964e+00
-1.60294771e-01 -3.08027476e-01 -4.08214092e-01 -1.96378779e+00
1.77606076e-01 -9.63915944e-01 2.99688071e-01 -1.86088538e+00
-1.92184076e-02 -6.91586494e-01 -1.92835815e-02 4.10590529e-01
-3.25798005e-01 4.51820076e-01 1.84012935e-01 2.18546256e-01
-4.53902960e-01 1.03833258e+00 1.63257349e+00 -3.51958960e-01
-8.49865079e-02 -1.36572003e-01 -6.35223269e-01 6.56795323e-01
5.46521783e-01 -3.09187442e-01 -6.57917738e-01 -8.99135590e-01
1.63475618e-01 9.47339311e-02 7.52585053e-01 -1.04420114e+00
3.52238417e-01 -1.72562957e-01 2.08641455e-01 -1.18120289e+00
7.50054181e-01 -6.76452875e-01 -3.21407199e-01 3.42220992e-01
8.43202695e-02 -1.33108124e-01 6.69989288e-02 8.76450360e-01
-3.47757250e-01 2.48394534e-02 6.77175462e-01 -2.12087985e-02
-9.93737876e-01 9.72082198e-01 3.27869743e-01 -5.90245910e-02
8.11889291e-01 -2.55071551e-01 -3.58196288e-01 -1.83884501e-01
-1.17300488e-01 6.66293979e-01 6.85316622e-01 2.90057272e-01
9.79410172e-01 -1.25824928e+00 -5.49459100e-01 2.55799294e-01
3.87676686e-01 1.16927946e+00 5.61335802e-01 9.18829679e-01
-6.10498071e-01 3.80436480e-01 9.07541141e-02 -1.12063754e+00
-7.71283746e-01 5.05537927e-01 9.37287509e-02 -7.15096816e-02
-1.00859320e+00 1.06505311e+00 7.63088524e-01 -5.45310497e-01
2.65846759e-01 -7.01809406e-01 7.73862153e-02 -2.73720145e-01
4.75092411e-01 1.67979077e-01 1.27686024e-01 -4.47400868e-01
-3.81137103e-01 1.02912712e+00 1.73705798e-02 -1.81587562e-01
1.22592449e+00 -3.59039128e-01 6.95584342e-02 2.21689135e-01
1.35361362e+00 -4.55904827e-02 -1.84807837e+00 -3.82066488e-01
-4.82770622e-01 -7.71632075e-01 3.53053927e-01 -3.56429964e-01
-1.34305263e+00 1.49019408e+00 2.61117905e-01 -5.20396590e-01
1.04918849e+00 3.57497185e-02 1.11251760e+00 5.59137821e-01
6.74108863e-01 -6.91288471e-01 1.75674379e-01 5.19145846e-01
6.66193485e-01 -1.64633775e+00 2.37167925e-02 -8.39023709e-01
-4.17704076e-01 8.42448354e-01 8.45738053e-01 -1.95642382e-01
5.16035736e-01 -3.61433290e-02 -2.04471946e-01 -1.44594610e-01
-5.18897712e-01 -3.30214471e-01 1.57447606e-01 7.30644643e-01
-6.84431046e-02 -6.03840090e-02 1.04186319e-01 4.90216494e-01
-1.15592845e-01 -8.56640283e-03 1.99571967e-01 7.82586396e-01
-4.82451230e-01 -9.06769454e-01 -6.38489202e-02 3.37871820e-01
2.80568808e-01 -1.73290029e-01 -7.04300255e-02 7.18853652e-01
3.65013331e-01 9.91443932e-01 1.06782414e-01 -6.45143330e-01
3.89005721e-01 -5.46905398e-01 5.01204252e-01 -7.32531369e-01
-5.14795305e-03 -1.05191410e-01 -2.12433159e-01 -9.85714138e-01
-2.89366394e-01 -5.92296124e-01 -1.59119546e+00 -4.17252749e-01
-6.38787374e-02 -1.49656206e-01 7.69696414e-01 9.73138928e-01
3.75683874e-01 5.39593935e-01 8.22865307e-01 -1.25187564e+00
-4.38316196e-01 -7.18162954e-01 -4.59418356e-01 9.00008678e-02
5.41742921e-01 -1.03314686e+00 -4.49193835e-01 -4.07187343e-01] | [8.56987190246582, -2.666487216949463] |
07c3a4fe-ec6f-481b-8725-55a34f643253 | hardc-a-novel-ecg-based-heartbeat | 2303.06020 | null | https://arxiv.org/abs/2303.06020v1 | https://arxiv.org/pdf/2303.06020v1.pdf | HARDC : A novel ECG-based heartbeat classification method to detect arrhythmia using hierarchical attention based dual structured RNN with dilated CNN | In this paper have developed a novel hybrid hierarchical attention-based bidirectional recurrent neural network with dilated CNN (HARDC) method for arrhythmia classification. This solves problems that arise when traditional dilated convolutional neural network (CNN) models disregard the correlation between contexts and gradient dispersion. The proposed HARDC fully exploits the dilated CNN and bidirectional recurrent neural network unit (BiGRU-BiLSTM) architecture to generate fusion features. As a result of incorporating both local and global feature information and an attention mechanism, the model's performance for prediction is improved.By combining the fusion features with a dilated CNN and a hierarchical attention mechanism, the trained HARDC model showed significantly improved classification results and interpretability of feature extraction on the PhysioNet 2017 challenge dataset. Sequential Z-Score normalization, filtering, denoising, and segmentation are used to prepare the raw data for analysis. CGAN (Conditional Generative Adversarial Network) is then used to generate synthetic signals from the processed data. The experimental results demonstrate that the proposed HARDC model significantly outperforms other existing models, achieving an accuracy of 99.60\%, F1 score of 98.21\%, a precision of 97.66\%, and recall of 99.60\% using MIT-BIH generated ECG. In addition, this approach substantially reduces run time when using dilated CNN compared to normal convolution. Overall, this hybrid model demonstrates an innovative and cost-effective strategy for ECG signal compression and high-performance ECG recognition. Our results indicate that an automated and highly computed method to classify multiple types of arrhythmia signals holds considerable promise. | ['Mohammad Ali Moni', 'Julian M. W. Quinn', 'Pietro Lio', 'Shahadat Uddin', 'Sunjida Sultana', 'Khondokar Fida Hasan', 'Md Shofiqul Islam'] | 2023-03-06 | null | null | null | null | ['heartbeat-classification'] | ['medical'] | [ 3.91031951e-01 -3.79935578e-02 2.85247326e-01 -1.99398622e-01
-1.00613344e+00 -2.51042426e-01 3.39443907e-02 -7.32762506e-03
-2.82175809e-01 7.85372376e-01 1.13830879e-01 -3.07134688e-01
-1.38599232e-01 -6.19226992e-01 -4.32573736e-01 -8.08094859e-01
-1.27402395e-01 -7.93723390e-02 -2.51527637e-01 -1.42211840e-01
-1.96871627e-02 4.28357571e-01 -1.05866361e+00 4.51006740e-01
9.33230698e-01 1.10685003e+00 -3.12381219e-02 9.20705855e-01
4.67042714e-01 5.99575043e-01 -9.95475113e-01 -3.00352663e-01
7.82954767e-02 -9.31204677e-01 -5.51789701e-01 -4.90162045e-01
-1.47607163e-01 -1.13817893e-01 -2.82051656e-02 6.90836668e-01
1.30412138e+00 -2.01592833e-01 4.53498155e-01 -6.68231547e-01
-4.89965022e-01 7.41327882e-01 -3.18414837e-01 3.81465584e-01
1.43175080e-01 2.67401308e-01 4.93711740e-01 -6.78795218e-01
4.09823269e-01 6.09650135e-01 1.20990586e+00 5.04689276e-01
-1.23037863e+00 -7.92579710e-01 -4.77934718e-01 -4.68810648e-02
-1.38791406e+00 -2.53999773e-02 1.00710273e+00 -2.88746327e-01
1.12061489e+00 5.08424699e-01 7.91483045e-01 1.06866956e+00
4.47021246e-01 2.98191160e-01 1.00273764e+00 -4.22363371e-01
-4.22370099e-02 -1.04243673e-01 -6.31034654e-03 5.07715762e-01
-1.27309486e-02 2.13984355e-01 -1.47534102e-01 -1.53314307e-01
5.75760245e-01 1.54651105e-01 -3.94602835e-01 4.92366463e-01
-1.19360924e+00 5.74563622e-01 5.83823502e-01 5.15157104e-01
-6.76853836e-01 2.84403026e-01 8.59308481e-01 2.32738450e-01
3.43579918e-01 7.90717006e-01 -4.23403978e-01 -2.42284000e-01
-9.19105411e-01 2.73962319e-01 3.84358704e-01 4.34270203e-01
1.07488006e-01 7.95070291e-01 -5.61358511e-01 8.44050944e-01
-5.62062636e-02 6.54556155e-01 9.70211387e-01 -7.68176675e-01
2.58208275e-01 6.37947738e-01 -3.32649797e-01 -1.22268701e+00
-5.74565291e-01 -1.25469315e+00 -1.37877929e+00 -4.58004698e-02
-4.29749191e-02 -2.91781604e-01 -8.82075548e-01 1.77045929e+00
-4.79179025e-02 2.41726205e-01 4.61836427e-01 8.11230719e-01
9.34744418e-01 4.53546941e-01 2.87212938e-01 -1.28882319e-01
1.16933620e+00 -4.07664537e-01 -8.83554101e-01 2.67217249e-01
7.51119375e-01 -7.58962274e-01 7.64615893e-01 2.22288996e-01
-1.23529351e+00 -8.58757496e-01 -1.22289348e+00 8.16459507e-02
-9.84250233e-02 4.15893853e-01 2.45115086e-01 7.08083153e-01
-9.22988892e-01 8.34053040e-01 -7.41115391e-01 4.22593988e-02
5.77563345e-01 4.39519465e-01 1.06539205e-01 2.72140443e-01
-1.46792495e+00 5.67375839e-01 2.50461072e-01 5.52778721e-01
-6.53585613e-01 -7.20172048e-01 -7.94991672e-01 2.27128938e-01
-3.85485768e-01 -8.38931680e-01 8.20442080e-01 -8.66010845e-01
-1.41091490e+00 4.85647947e-01 1.74488410e-01 -9.73432302e-01
3.37768257e-01 -2.84403890e-01 -4.34888631e-01 1.02725826e-01
6.73056720e-03 4.03803945e-01 5.81461489e-01 -7.65488446e-01
-1.96099833e-01 -2.82323420e-01 -4.85012203e-01 1.10901035e-01
-3.80964018e-02 -1.52976215e-01 -1.35239825e-01 -1.18532288e+00
4.71467599e-02 -7.67157674e-01 -1.93467677e-01 -5.95925629e-01
-4.48761761e-01 2.38388151e-01 7.50064909e-01 -9.54559505e-01
1.44206071e+00 -2.16879225e+00 1.15155905e-01 3.83908927e-01
2.25771293e-01 7.46655107e-01 -2.06582304e-02 1.40804827e-01
-3.34838092e-01 3.52564186e-01 -3.66101146e-01 -1.87844187e-01
-4.28456694e-01 -1.61178902e-01 -1.54923186e-01 2.83048362e-01
3.10205579e-01 1.33263481e+00 -5.17324567e-01 -1.71642914e-01
2.90381283e-01 9.21001196e-01 -5.74703574e-01 7.68084377e-02
2.23612577e-01 6.15320981e-01 -3.31617147e-01 4.96108443e-01
3.90008837e-01 -6.38016239e-02 1.71068549e-01 -3.23966026e-01
1.19442374e-01 1.74865243e-03 -7.89152384e-01 1.56471062e+00
-4.08750176e-01 4.15675104e-01 -4.34991330e-01 -1.29218173e+00
1.32086921e+00 6.45541251e-01 4.08184499e-01 -8.09476852e-01
5.50046861e-01 3.41655165e-01 3.22719216e-01 -5.03118575e-01
-1.05361164e-01 -2.01005831e-01 -1.37918949e-01 2.20693231e-01
-1.07222617e-01 -7.82882720e-02 -3.48466069e-01 -1.14536420e-01
1.08256841e+00 1.07036062e-01 1.24949999e-01 -1.96196035e-01
5.81205547e-01 -3.45983952e-01 7.70035803e-01 7.37058818e-01
-5.14656194e-02 7.84975231e-01 4.76014197e-01 -6.20728970e-01
-8.50263000e-01 -6.07889950e-01 -1.67470165e-02 2.93476880e-01
-3.92006457e-01 -3.43205363e-01 -8.75941873e-01 -2.05928519e-01
-3.67329419e-01 3.66303504e-01 -6.41756594e-01 -6.03494108e-01
-6.71466410e-01 -8.85371447e-01 1.22769141e+00 7.86128879e-01
7.16534555e-01 -1.39844489e+00 -8.33196223e-01 6.08528137e-01
-4.07978058e-01 -9.56688702e-01 -2.50868797e-01 5.50921708e-02
-1.10781479e+00 -8.50008249e-01 -9.11944628e-01 -8.60753000e-01
2.82344639e-01 -4.90825593e-01 9.52125907e-01 2.03142673e-01
-6.31733239e-01 1.30502938e-03 -3.96328270e-01 -5.89831054e-01
-3.37211132e-01 2.13994712e-01 -2.70248741e-01 1.14601798e-01
-2.15365347e-02 -6.80986822e-01 -9.13523495e-01 -9.10295025e-02
-5.80563486e-01 1.62546769e-01 7.00628757e-01 1.19841099e+00
7.23108232e-01 -2.12046728e-01 1.36202753e+00 -9.59782481e-01
8.33907962e-01 -3.01982641e-01 -2.18906492e-01 -6.45706803e-02
-8.24932039e-01 -2.45948941e-01 8.38544846e-01 -3.09605479e-01
-8.37244272e-01 8.57824311e-02 -5.36581695e-01 -5.83932936e-01
1.59268141e-01 5.00598967e-01 4.34893779e-02 2.09055111e-01
7.10735440e-01 4.04681236e-01 2.78768152e-01 -1.50717124e-01
-2.02112734e-01 5.74364483e-01 7.27822304e-01 -2.64345348e-01
3.56859952e-01 1.13567837e-01 4.88438942e-02 -4.71239418e-01
-4.86912221e-01 2.13886481e-02 -3.25624377e-01 -2.46949196e-02
1.13436711e+00 -8.90733480e-01 -7.32393384e-01 7.07340777e-01
-1.00658274e+00 -3.00909668e-01 -4.52375144e-01 6.19674802e-01
-3.69966686e-01 2.16354057e-01 -8.21996391e-01 -7.50466764e-01
-1.32870710e+00 -1.13868248e+00 8.75918984e-01 4.21449654e-02
-5.32654464e-01 -8.97027969e-01 -1.48893014e-01 3.14777464e-01
8.84115696e-01 1.08742976e+00 9.19594884e-01 -6.67420208e-01
-2.44956613e-01 -5.31063318e-01 3.56063768e-02 7.65886307e-01
3.32404748e-02 -2.69463003e-01 -9.97411013e-01 -1.08324006e-01
5.62560968e-02 -5.48319938e-03 7.14184165e-01 6.20568693e-01
1.39462256e+00 -1.67373508e-01 -1.29610971e-01 8.73402834e-01
1.36293983e+00 5.68528771e-01 1.00689542e+00 -5.17220199e-02
6.42803550e-01 3.99995148e-02 1.43151268e-01 3.34991276e-01
-2.38905177e-02 3.80610645e-01 1.89465851e-01 -7.10632026e-01
-3.27618986e-01 1.52973101e-01 4.29242589e-02 1.03873467e+00
-4.49581444e-01 6.50704056e-02 -8.64693940e-01 4.84463185e-01
-1.48994470e+00 -1.07298708e+00 -2.47666076e-01 2.03012347e+00
8.31106424e-01 2.52202719e-01 -1.79837793e-01 6.48794055e-01
7.19387352e-01 -3.44539016e-01 -5.05081773e-01 -6.28325760e-01
-3.66617143e-01 9.22518194e-01 1.41762555e-01 2.04299644e-01
-9.42977905e-01 4.62998599e-01 5.38229799e+00 6.00986123e-01
-1.48936653e+00 3.51483263e-02 9.31302488e-01 1.26801148e-01
6.63814470e-02 -4.55660194e-01 -3.33798170e-01 5.25478005e-01
1.24986315e+00 -1.49579048e-02 7.26376772e-02 4.75539148e-01
3.82946312e-01 3.89795125e-01 -5.83706796e-01 1.12192762e+00
9.69338194e-02 -1.33677053e+00 -1.40770391e-01 -1.80808783e-01
6.24431252e-01 -1.03766836e-01 2.07469150e-01 4.76348728e-01
-4.22003299e-01 -1.16849291e+00 4.04163986e-01 8.41578901e-01
1.21794772e+00 -9.39210892e-01 1.31104982e+00 5.21957353e-02
-1.11793029e+00 -1.82430178e-01 6.51327968e-02 -2.19743736e-02
-1.12753540e-01 6.42899930e-01 -8.25994253e-01 6.99760318e-01
7.81896114e-01 8.05914998e-01 -5.63396513e-01 7.72878528e-01
2.46599149e-02 9.31462348e-01 -1.23314895e-01 1.44409284e-01
-6.76633343e-02 8.35323185e-02 4.68846202e-01 1.29931617e+00
3.42832923e-01 1.57945991e-01 -1.35473430e-01 8.85589600e-01
-6.27011135e-02 1.22467041e-01 -5.35438001e-01 1.83684871e-01
1.67212263e-01 1.11376965e+00 -5.43550968e-01 -5.25887907e-01
-2.08932105e-02 8.34351242e-01 -1.50128156e-01 3.73611957e-01
-1.18701196e+00 -9.65461254e-01 1.02190234e-01 -1.51018240e-02
6.09796494e-02 4.13179040e-01 -7.19987869e-01 -8.23332012e-01
6.94423839e-02 -9.69283581e-01 2.51211613e-01 -6.29204094e-01
-7.54488766e-01 1.25103724e+00 -4.45019007e-01 -1.34433019e+00
-3.55565906e-01 -7.51629025e-02 -8.23911548e-01 1.26943994e+00
-1.30436373e+00 -1.16200030e+00 -3.48036945e-01 5.11327863e-01
3.98303062e-01 -2.80567467e-01 1.30026829e+00 4.86845583e-01
-6.48811996e-01 7.67349422e-01 -9.70760137e-02 3.42196971e-01
4.50800449e-01 -1.18652749e+00 1.81830645e-01 8.75407755e-01
-4.31119412e-01 7.14994609e-01 3.51945519e-01 -5.42629004e-01
-8.60473990e-01 -1.53950882e+00 8.73600245e-01 -2.91360207e-02
-6.88936487e-02 -1.47155330e-01 -9.05370951e-01 2.96970934e-01
4.94453646e-02 9.09081325e-02 8.09603930e-01 -2.78783739e-01
4.14317707e-03 -4.03341234e-01 -1.01879954e+00 3.12795192e-01
4.99919176e-01 -4.52090502e-01 -4.44462448e-01 1.07584735e-02
5.88090062e-01 -5.10201573e-01 -1.21860170e+00 9.10998642e-01
7.84042239e-01 -8.15771580e-01 9.55428898e-01 -3.78480703e-01
4.69440609e-01 -5.95119819e-02 3.17320637e-02 -1.05329204e+00
-2.17598572e-01 -7.47196674e-01 7.47407153e-02 1.12988710e+00
6.30532384e-01 -8.18861008e-01 4.82329875e-01 2.41174430e-01
-5.65837741e-01 -1.31563163e+00 -6.74354494e-01 -4.09049153e-01
-1.97099783e-02 -4.78842616e-01 4.71514404e-01 7.12675333e-01
-2.50017554e-01 3.31374019e-01 -4.40404147e-01 -2.82678730e-03
3.21867943e-01 8.59495625e-03 2.86342472e-01 -1.06787288e+00
-2.94893712e-01 -2.52777427e-01 -5.67906916e-01 -1.48664832e-01
-1.27979591e-01 -1.10992086e+00 -3.86813402e-01 -1.25522077e+00
-5.54754511e-02 -5.11722565e-01 -6.65237963e-01 5.84238827e-01
-2.31359407e-01 8.35700810e-01 8.19505826e-02 4.20838706e-02
-1.03206620e-01 3.64779562e-01 1.18724835e+00 -4.25891094e-02
-4.82746571e-01 2.73069352e-01 -7.45709777e-01 4.77587253e-01
1.24732125e+00 -3.94263089e-01 -3.30449611e-01 -9.47016478e-02
-6.22688532e-02 3.69618654e-01 5.05865157e-01 -1.31045163e+00
-1.31882310e-01 7.39510477e-01 7.77718723e-01 -4.41213012e-01
2.75849193e-01 -4.96504873e-01 4.37619656e-01 9.60199893e-01
-4.06528413e-01 1.53222710e-01 4.00385082e-01 2.40038902e-01
-3.65957201e-01 3.03246886e-01 8.23581815e-01 1.23029731e-01
6.34636059e-02 8.35814178e-02 -3.02403450e-01 -5.08121774e-02
9.72987711e-01 -2.29177952e-01 6.22823946e-02 -1.64204806e-01
-1.19223046e+00 -1.11026175e-01 -3.65580797e-01 1.67403087e-01
8.84964168e-01 -1.32665479e+00 -8.99334967e-01 5.50854623e-01
-1.57888010e-01 -2.11711619e-02 6.38525605e-01 1.16836107e+00
-6.66840672e-01 3.47306639e-01 -2.74338543e-01 -6.33011103e-01
-1.24298358e+00 1.20740674e-01 6.60381973e-01 -4.72016722e-01
-9.36466157e-01 5.13574004e-01 -1.65455848e-01 -1.29742131e-01
1.03236802e-01 -5.44766843e-01 -2.95132399e-01 -2.15955257e-01
2.27175802e-01 3.39660943e-01 3.20206910e-01 -3.49689156e-01
-2.76766002e-01 6.72602057e-01 2.51560450e-01 2.27499172e-01
1.32377028e+00 1.86251655e-01 4.96335328e-03 1.64826959e-01
1.18124259e+00 -1.83135241e-01 -6.27960920e-01 6.57588467e-02
-4.87107724e-01 1.36930391e-01 -5.32871932e-02 -1.25099206e+00
-1.41844308e+00 1.14952719e+00 1.06784248e+00 -9.33375731e-02
1.69181859e+00 -4.61754233e-01 9.68953311e-01 -5.43288328e-02
-2.53926545e-01 -7.06110537e-01 -1.45028666e-01 2.32218906e-01
1.07876062e+00 -7.26812601e-01 -2.44275957e-01 5.44213876e-03
-6.55993402e-01 1.14759219e+00 3.80339503e-01 -3.13328505e-01
6.74153984e-01 3.57917219e-01 2.63934940e-01 -1.93066508e-01
-5.26988387e-01 1.47395328e-01 2.22236514e-01 4.93792564e-01
6.46136820e-01 1.98545493e-02 -5.21942139e-01 1.21578372e+00
-2.81704724e-01 1.75837934e-01 2.92463571e-01 7.12014794e-01
7.28449970e-02 -8.12656045e-01 -1.01756811e-01 6.33719385e-01
-1.14478159e+00 -2.45022357e-01 7.14468136e-02 7.18632281e-01
4.15171713e-01 6.58971906e-01 -1.48752064e-01 -5.09501278e-01
3.69522929e-01 2.56101549e-01 6.60420880e-02 -3.77368659e-01
-1.16087353e+00 4.58000749e-01 -2.19382569e-01 -3.99850309e-01
-2.27457553e-01 -4.09066588e-01 -1.33703256e+00 2.36402929e-01
-3.39163780e-01 8.92241746e-02 6.03996873e-01 6.39774561e-01
8.29728425e-01 1.39684570e+00 5.59249580e-01 -5.36232710e-01
-4.49312627e-01 -1.26812470e+00 -4.33721662e-01 4.88970697e-01
3.82806331e-01 -2.04105183e-01 -1.79932937e-01 2.91669756e-01] | [14.283663749694824, 3.2619521617889404] |
85dbed2f-baf3-4efb-936a-8f3bf974ea74 | distillpose-lightweight-camera-localization | 2108.03819 | null | https://arxiv.org/abs/2108.03819v1 | https://arxiv.org/pdf/2108.03819v1.pdf | DistillPose: Lightweight Camera Localization Using Auxiliary Learning | We propose a lightweight retrieval-based pipeline to predict 6DOF camera poses from RGB images. Our pipeline uses a convolutional neural network (CNN) to encode a query image as a feature vector. A nearest neighbor lookup finds the pose-wise nearest database image. A siamese convolutional neural network regresses the relative pose from the nearest neighboring database image to the query image. The relative pose is then applied to the nearest neighboring absolute pose to obtain the query image's final absolute pose prediction. Our model is a distilled version of NN-Net that reduces its parameters by 98.87%, information retrieval feature vector size by 87.5%, and inference time by 89.18% without a significant decrease in localization accuracy. | ['Slobodan Ilic', 'Mai Bui', 'Yehya Abouelnaga'] | 2021-08-09 | null | null | null | null | ['camera-localization', 'auxiliary-learning'] | ['computer-vision', 'methodology'] | [ 5.91635667e-02 -2.56396621e-01 -3.99641126e-01 -6.62324727e-01
-1.12206399e+00 -5.56132078e-01 2.65594006e-01 1.15281139e-02
-7.73561239e-01 2.05296323e-01 9.72726122e-02 7.60003105e-02
-6.37594461e-02 -9.70761955e-01 -1.17712843e+00 -3.79798591e-01
1.32405430e-01 4.07341510e-01 3.11628103e-01 1.65677909e-02
4.61166471e-01 1.14074779e+00 -1.35141611e+00 -2.15742551e-02
8.26708749e-02 1.69510698e+00 -9.44437012e-02 8.48697245e-01
3.73261482e-01 7.56044984e-01 -5.31111002e-01 -3.61824483e-02
6.45548761e-01 1.73243031e-01 -7.37675905e-01 -4.10026610e-01
1.13827395e+00 -8.98077726e-01 -1.13873744e+00 7.53723681e-01
5.23522317e-01 4.07111406e-01 3.26318502e-01 -1.06943536e+00
-5.72573066e-01 -1.42265335e-01 -2.99493849e-01 2.82978844e-02
6.55763626e-01 -1.17771216e-02 7.63050914e-01 -9.84148622e-01
8.45150590e-01 8.65128219e-01 8.22160423e-01 2.18870521e-01
-8.79478633e-01 -6.64402664e-01 -2.95042068e-01 1.40191033e-01
-1.95064044e+00 -3.51758152e-01 5.67140043e-01 -4.75379899e-02
1.48765218e+00 1.14258409e-01 8.96077991e-01 4.24863666e-01
4.56134558e-01 2.87782013e-01 3.89245242e-01 2.75737606e-02
1.93079695e-01 -5.70746005e-01 -3.19057763e-01 9.60204482e-01
-1.89839125e-01 1.14432119e-01 -9.56099927e-01 7.54469037e-02
8.55478585e-01 3.92134637e-01 7.56608844e-02 -5.90228379e-01
-1.37244654e+00 6.49490297e-01 1.41839206e+00 -3.25597107e-01
-4.17751849e-01 9.74791765e-01 -5.67491874e-02 9.50435102e-02
1.53445601e-01 4.80999380e-01 -4.68729228e-01 -1.67502061e-01
-9.23486948e-01 3.48277748e-01 7.49994397e-01 1.04707336e+00
1.22832108e+00 -3.21831703e-01 1.45627931e-01 3.43882352e-01
5.97180605e-01 1.07277858e+00 2.84840673e-01 -1.50005579e+00
2.55605102e-01 6.27221584e-01 1.06901236e-01 -1.48599696e+00
-3.99971277e-01 -1.79325059e-01 -2.66918004e-01 -5.56467138e-02
1.82509229e-01 2.63820946e-01 -9.10210431e-01 9.99328494e-01
3.46571892e-01 8.59543607e-02 -1.47923246e-01 1.31267130e+00
7.64837682e-01 6.90411031e-01 -4.24526960e-01 4.60227042e-01
8.17112029e-01 -8.80461276e-01 -2.18478486e-01 -3.83806616e-01
3.44024122e-01 -1.03569353e+00 6.63733244e-01 2.50807285e-01
-8.55833113e-01 -3.61695319e-01 -1.23771584e+00 -6.25530779e-01
-5.13139784e-01 1.74375281e-01 6.19913876e-01 1.09238759e-01
-1.23710012e+00 6.71169698e-01 -9.60204482e-01 -3.11666161e-01
2.26920485e-01 8.28496695e-01 -5.90678096e-01 -3.75720203e-01
-9.32684541e-01 8.71787548e-01 2.99690634e-01 1.85518906e-01
-6.53763950e-01 -6.37075186e-01 -1.10971153e+00 -1.05512023e-01
1.57088473e-01 -5.17042518e-01 1.25794399e+00 -5.43108404e-01
-1.57672846e+00 7.18098342e-01 -4.32665050e-01 -5.51034927e-01
1.47294000e-01 -3.40332299e-01 -2.21966147e-01 2.73555160e-01
2.51614988e-01 1.17048967e+00 7.89658844e-01 -8.12776923e-01
-3.83474737e-01 -6.15038931e-01 -2.42102090e-02 3.65654707e-01
1.13849223e-01 -3.70179921e-01 -1.12826598e+00 -3.23685020e-01
9.47917283e-01 -1.17008746e+00 -1.33394063e-01 5.30327082e-01
-1.83189824e-01 -6.90595657e-02 8.83997500e-01 -3.97859752e-01
7.98903525e-01 -2.21213102e+00 -1.39990896e-01 4.34381098e-01
2.03415230e-01 -3.67614985e-01 -2.25706145e-01 1.85115889e-01
1.16185106e-01 -3.68699223e-01 2.80759633e-01 -2.63586700e-01
-7.08098635e-02 2.67178923e-01 -2.61346549e-01 8.38870585e-01
8.29302222e-02 1.17056072e+00 -9.00050163e-01 -3.31861734e-01
4.83747780e-01 9.05881107e-01 -7.50064254e-01 2.06199691e-01
8.05529580e-02 -1.62064269e-01 -1.58917159e-01 9.17314053e-01
5.52819192e-01 -3.77072915e-02 -3.81581277e-01 -8.43261719e-01
-7.89901824e-04 3.90430421e-01 -1.07448947e+00 2.41450071e+00
-4.59992796e-01 1.02073956e+00 -3.57160121e-01 -2.81641304e-01
1.16288567e+00 -4.07704979e-01 7.30300069e-01 -8.87969851e-01
2.04894289e-01 1.03793822e-01 -2.90074557e-01 -1.14965193e-01
9.04229999e-01 3.80756468e-01 -4.28369373e-01 2.69763291e-01
-2.58592442e-02 -4.59614336e-01 -2.57884949e-01 9.71842259e-02
1.19429064e+00 5.28161883e-01 3.10235936e-03 2.99282342e-01
8.99943635e-02 2.69921064e-01 4.07159358e-01 6.84763730e-01
-2.11539105e-01 9.50198114e-01 2.51708385e-02 -1.16387260e+00
-9.36954439e-01 -1.36883450e+00 2.64906473e-02 1.04866087e+00
5.17796814e-01 -4.57249403e-01 -5.69103837e-01 -4.32449669e-01
3.83500665e-01 1.38459876e-01 -5.39867282e-01 -3.18722099e-01
-5.44132173e-01 8.80273879e-02 3.86697769e-01 6.66519165e-01
7.03232944e-01 -5.98921895e-01 -8.11634481e-01 -1.20259888e-01
5.51292785e-02 -9.36284781e-01 -8.52151453e-01 2.97493190e-01
-7.49929726e-01 -1.02937365e+00 -3.35808158e-01 -7.90169656e-01
7.60091066e-01 2.37627149e-01 9.32821453e-01 -1.96916908e-01
-2.38234639e-01 3.34280014e-01 -1.27262995e-01 -3.70988064e-03
1.64013550e-01 2.40371808e-01 1.77876696e-01 -2.64577121e-01
7.04859078e-01 -1.19688660e-01 -9.77806270e-01 2.89532423e-01
-7.33872831e-01 -4.08794910e-01 6.07372582e-01 6.06476903e-01
1.28875065e+00 -4.13100153e-01 -3.54634643e-01 -1.74420565e-01
1.55653477e-01 -1.18015096e-01 -8.43031645e-01 -3.97826657e-02
-3.76324415e-01 2.33187512e-01 2.35131145e-01 -1.96318626e-01
-3.02361369e-01 1.02948964e+00 -1.96802877e-02 -8.86434555e-01
-5.86303137e-02 3.91038060e-01 4.06985730e-02 -4.81980622e-01
7.11446941e-01 2.33683482e-01 -2.08791252e-03 -1.61908522e-01
4.69104648e-01 6.69793069e-01 1.11826241e+00 -1.06979877e-01
8.09342384e-01 3.78192812e-01 5.52846640e-02 -5.74935198e-01
-9.16696846e-01 -5.01188457e-01 -1.09453511e+00 -1.54077441e-01
9.32906806e-01 -1.12718379e+00 -1.15087998e+00 2.73261756e-01
-1.21441233e+00 -1.75438970e-01 -1.77143708e-01 5.32502770e-01
-3.83940697e-01 -1.31735221e-01 -5.81538677e-01 -9.41386446e-02
-4.18639213e-01 -1.26044261e+00 1.57901001e+00 3.90727997e-01
-2.84641027e-01 -2.99113572e-01 6.14560954e-03 2.53242373e-01
4.04699057e-01 1.41563773e-01 4.59053107e-02 -4.27425325e-01
-1.11454809e+00 -8.49677920e-01 -4.01007503e-01 -3.47165205e-02
-7.59566799e-02 1.05642423e-01 -7.21663594e-01 -2.59087861e-01
-4.34788257e-01 -3.02469969e-01 6.41074955e-01 2.95502216e-01
1.11773884e+00 -7.35045969e-02 -2.85384983e-01 1.21740139e+00
1.56287158e+00 3.57495755e-01 5.28155148e-01 4.23115075e-01
7.04109788e-01 -1.78485945e-01 7.07385182e-01 3.04209560e-01
4.97049451e-01 6.89140677e-01 8.38032722e-01 2.50369962e-02
-6.98750168e-02 -6.81139469e-01 1.11013241e-01 6.89933479e-01
4.24269915e-01 2.05449805e-01 -1.04293060e+00 1.22422546e-01
-1.49742711e+00 -6.64338291e-01 3.52681935e-01 2.23211837e+00
4.14551497e-01 -3.15237828e-02 -3.67935508e-01 -3.50155830e-01
2.83184975e-01 1.41155377e-01 -9.77229536e-01 -2.43432954e-01
1.13145620e-01 2.66606092e-01 1.04714453e+00 7.73728490e-01
-1.30820549e+00 1.21076167e+00 6.64290428e+00 2.68850535e-01
-1.43632531e+00 -3.51767182e-01 4.36535954e-01 -7.01771915e-01
1.45072579e-01 -1.28404140e-01 -8.03115427e-01 1.12746462e-01
9.39488411e-01 2.40013674e-01 8.72071326e-01 1.09505272e+00
-1.34732515e-01 -5.23299456e-01 -1.05361044e+00 1.41638410e+00
5.10302842e-01 -1.50103152e+00 -3.42007796e-03 1.15409032e-01
5.64144790e-01 5.81132412e-01 2.05050260e-02 -5.11194672e-03
6.92439452e-02 -1.07496631e+00 6.65374398e-01 8.90566051e-01
1.01140738e+00 -1.09363842e+00 8.08715463e-01 1.06116319e-02
-1.13407886e+00 4.55458350e-02 -6.49768651e-01 -2.18729034e-01
-2.01245606e-01 2.05094233e-01 -1.11690724e+00 -9.66800675e-02
1.17292047e+00 8.55437577e-01 -7.69329250e-01 8.42653811e-01
-8.45481604e-02 -1.26764640e-01 -8.38118076e-01 -1.42079249e-01
2.31760502e-01 -7.63548836e-02 1.18913546e-01 7.60421395e-01
2.82266498e-01 1.02420815e-03 8.35275054e-02 7.10699499e-01
-3.18233758e-01 -2.21544743e-01 -9.05935407e-01 2.47605041e-01
1.06985390e+00 1.18224955e+00 -7.26298928e-01 -2.36002341e-01
-4.60945442e-02 1.51283216e+00 3.43920946e-01 2.14734077e-01
-6.96720302e-01 -8.73199046e-01 8.47801924e-01 -1.00321032e-01
3.14368695e-01 -4.28831100e-01 -8.50034207e-02 -8.04624975e-01
8.49772617e-03 -4.96242762e-01 -3.07341735e-03 -1.32630622e+00
-8.62460792e-01 5.61131120e-01 -3.76674324e-01 -1.34675145e+00
-4.37068641e-01 -5.54344893e-01 -8.86330605e-02 9.00043309e-01
-1.04133034e+00 -1.11056530e+00 -6.93656504e-01 6.26471460e-01
6.39448911e-02 6.81742504e-02 9.01363671e-01 1.80471644e-01
-4.20816213e-01 6.01223767e-01 9.42254439e-02 6.87349021e-01
9.23211753e-01 -1.03945601e+00 4.50492382e-01 3.97952288e-01
2.07051367e-01 1.02027965e+00 2.08067030e-01 -4.02596772e-01
-2.10694170e+00 -1.07926798e+00 9.91336346e-01 -8.47426414e-01
5.49653113e-01 -3.49421024e-01 -2.89181441e-01 7.90726662e-01
-1.87817946e-01 7.53180027e-01 4.13595796e-01 -2.69183844e-01
-5.95209658e-01 -7.55184054e-01 -1.15870309e+00 4.65738297e-01
8.79206777e-01 -1.14715910e+00 -3.98694187e-01 2.39923954e-01
8.57453048e-01 -1.14188433e+00 -1.42121768e+00 6.97238594e-02
1.11251199e+00 -7.83674359e-01 1.29978871e+00 -6.55406490e-02
3.57168019e-01 -6.15178585e-01 -6.67454183e-01 -8.73403132e-01
-1.38553977e-01 -2.63149440e-01 -5.29991388e-02 4.62364465e-01
2.46576577e-01 -3.60199004e-01 1.24861705e+00 7.91850090e-01
1.03786193e-01 -1.06571019e+00 -1.21073234e+00 -4.19423014e-01
-1.59022123e-01 -5.99285305e-01 6.39394939e-01 5.09739697e-01
-2.86806226e-01 2.28205204e-01 -1.46885309e-03 4.49128896e-01
3.68042469e-01 3.17116112e-01 9.55988646e-01 -7.46334076e-01
1.28936514e-01 -1.37770459e-01 -9.59517896e-01 -1.44544291e+00
-2.35870667e-02 -8.08261275e-01 3.79222631e-01 -1.45522833e+00
-1.28472626e-01 -1.36155441e-01 -2.07507834e-01 5.79554260e-01
3.57291609e-01 9.22354996e-01 2.21519724e-01 1.91233888e-01
-9.27113652e-01 2.49795973e-01 9.27146852e-01 -2.93388098e-01
-2.31843829e-01 -2.32108787e-01 -1.24345377e-01 5.62922657e-01
6.76196277e-01 -5.96208453e-01 -3.70721705e-02 -7.62831509e-01
3.64694983e-01 1.19951196e-01 3.97568643e-01 -1.35219491e+00
7.64070690e-01 1.26519188e-01 1.14416647e+00 -1.04243505e+00
7.56076097e-01 -9.34421480e-01 1.49696007e-01 6.03034556e-01
-3.40422690e-01 5.06081939e-01 6.98300228e-02 3.52407664e-01
-2.28473425e-01 4.51358445e-02 4.33651328e-01 6.89463094e-02
-8.78513694e-01 6.45216048e-01 -8.91880020e-02 -4.02084947e-01
8.68918180e-01 -4.56517071e-01 1.86181553e-02 -6.17684603e-01
-5.20832896e-01 -1.34093180e-01 7.50871599e-01 4.94535685e-01
9.79932904e-01 -1.49740636e+00 1.68061350e-02 4.13184017e-01
1.49881721e-01 3.12976867e-01 -9.01358724e-02 4.63171214e-01
-1.21521044e+00 6.20864272e-01 3.27183455e-02 -8.86254132e-01
-1.06555283e+00 3.96297693e-01 7.96936393e-01 3.59177947e-01
-2.17338517e-01 1.06320584e+00 -6.55995071e-01 -7.78530002e-01
3.34825993e-01 -5.13433039e-01 4.28760052e-01 -2.54076421e-01
4.37404513e-01 4.54449713e-01 1.93573460e-01 -9.51521814e-01
-7.57592618e-01 9.92021620e-01 7.66542405e-02 -7.78484941e-02
1.22057331e+00 4.42580134e-02 -2.05202401e-01 1.05269849e-01
2.11902308e+00 -1.51620418e-01 -1.49527812e+00 -2.10275501e-01
-1.56884119e-01 -6.93120360e-01 5.38150191e-01 -6.63241208e-01
-1.25910711e+00 6.11231923e-01 9.35654223e-01 -6.10498309e-01
9.83616829e-01 8.45357180e-02 8.01638842e-01 1.12749875e+00
5.48777282e-01 -1.03913617e+00 1.18176207e-01 8.00265968e-01
7.92982638e-01 -1.27275038e+00 3.58744621e-01 1.00782868e-02
-1.06906816e-01 1.34304285e+00 7.64861763e-01 -5.91073394e-01
8.56760681e-01 3.04557502e-01 2.56353170e-01 -4.20708656e-01
-2.71728784e-01 1.06958756e-02 5.58589816e-01 4.46825415e-01
2.82279670e-01 -1.68977633e-01 6.27170384e-01 1.76739376e-02
-5.91501653e-01 1.13302236e-02 -6.69710487e-02 1.00057745e+00
-2.59730488e-01 -5.72028458e-01 -4.06106561e-01 3.57859552e-01
-4.87592593e-02 -1.78518951e-01 -5.75572550e-01 6.72206759e-01
1.68667838e-01 5.77952266e-01 8.12904537e-01 -8.21586013e-01
4.87870812e-01 -9.86616686e-02 4.37778503e-01 -3.44496787e-01
-4.06169087e-01 -8.62714835e-03 -4.99380916e-01 -1.43527687e+00
-2.04788059e-01 -3.12213749e-01 -1.42727697e+00 -4.19882715e-01
-1.34480640e-01 -2.92977273e-01 1.01297688e+00 6.35383606e-01
6.78445399e-01 1.27078772e-01 5.74895978e-01 -1.26780427e+00
-1.92186862e-01 -5.97422302e-01 -1.07232630e-01 1.37003466e-01
3.67390752e-01 -2.82902092e-01 -1.74638629e-01 -8.00543278e-02] | [7.698017120361328, -2.17336368560791] |
dde9cdce-dbcb-49d0-8475-07047b4fecd6 | learning-representations-in-model-free | 1810.10096 | null | http://arxiv.org/abs/1810.10096v3 | http://arxiv.org/pdf/1810.10096v3.pdf | Learning Representations in Model-Free Hierarchical Reinforcement Learning | Common approaches to Reinforcement Learning (RL) are seriously challenged by
large-scale applications involving huge state spaces and sparse delayed reward
feedback. Hierarchical Reinforcement Learning (HRL) methods attempt to address
this scalability issue by learning action selection policies at multiple levels
of temporal abstraction. Abstraction can be had by identifying a relatively
small set of states that are likely to be useful as subgoals, in concert with
the learning of corresponding skill policies to achieve those subgoals. Many
approaches to subgoal discovery in HRL depend on the analysis of a model of the
environment, but the need to learn such a model introduces its own problems of
scale. Once subgoals are identified, skills may be learned through intrinsic
motivation, introducing an internal reward signal marking subgoal attainment.
In this paper, we present a novel model-free method for subgoal discovery using
incremental unsupervised learning over a small memory of the most recent
experiences (trajectories) of the agent. When combined with an intrinsic
motivation learning mechanism, this method learns both subgoals and skills,
based on experiences in the environment. Thus, we offer an original approach to
HRL that does not require the acquisition of a model of the environment,
suitable for large-scale applications. We demonstrate the efficiency of our
method on two RL problems with sparse delayed feedback: a variant of the rooms
environment and the first screen of the ATARI 2600 Montezuma's Revenge game. | ['David C. Noelle', 'Jacob Rafati'] | 2018-10-23 | null | https://openreview.net/forum?id=S1gDCiCqtQ | https://openreview.net/pdf?id=S1gDCiCqtQ | null | ['montezumas-revenge'] | ['playing-games'] | [ 1.47488505e-01 1.66499019e-01 -2.82000005e-01 -3.65075730e-02
-6.93939388e-01 -6.86180353e-01 7.56295264e-01 5.88318348e-01
-7.84517348e-01 1.13946915e+00 9.19646323e-02 1.89107787e-02
-3.28461319e-01 -8.28855515e-01 -8.00344884e-01 -6.09549403e-01
-7.27759957e-01 5.61042249e-01 6.66454852e-01 -5.86875141e-01
4.32855278e-01 6.05906963e-01 -1.76650953e+00 1.28227938e-02
6.60017252e-01 6.19354367e-01 5.43910980e-01 8.15380931e-01
1.65762201e-01 1.25172412e+00 -5.10061681e-01 5.99160612e-01
3.40743273e-01 -7.48315156e-01 -1.11093879e+00 9.74883139e-02
-2.08369866e-01 -4.43567991e-01 -1.34063751e-01 8.87210310e-01
2.31296211e-01 6.09722376e-01 2.69071519e-01 -1.10988533e+00
6.98942542e-02 6.80489540e-01 6.08179308e-02 1.49562448e-01
6.15086496e-01 4.74637866e-01 9.65702713e-01 -2.74858207e-01
6.63781345e-01 1.15899229e+00 2.34738708e-01 8.16330016e-01
-1.38740921e+00 -3.21032315e-01 3.89995992e-01 2.80718386e-01
-1.05995989e+00 -2.53917545e-01 4.90106195e-01 -2.84887612e-01
1.11819053e+00 -3.92028876e-02 9.85090911e-01 8.26144099e-01
1.50264502e-01 8.24687481e-01 1.45597291e+00 -5.37204623e-01
9.37136650e-01 2.57287128e-03 -4.25100178e-02 7.44412720e-01
-1.07855290e-01 7.66221583e-01 -7.23587930e-01 -2.71407515e-01
1.02110124e+00 -7.76783610e-03 1.48109719e-01 -7.51055539e-01
-9.43698883e-01 7.07349420e-01 2.17299789e-01 3.81111175e-01
-6.40781045e-01 3.54907125e-01 2.87127048e-01 6.81301892e-01
-2.00716317e-01 8.93270373e-01 -5.59867501e-01 -3.75150770e-01
-5.51063001e-01 5.23584723e-01 6.81284487e-01 7.26013899e-01
1.07279229e+00 8.69359225e-02 -8.63540024e-02 2.63527811e-01
1.47700831e-01 2.44698286e-01 6.49078369e-01 -1.20900118e+00
1.79155350e-01 6.09853685e-01 5.50982773e-01 -2.83484787e-01
-5.25038779e-01 -9.30130780e-02 6.13545291e-02 8.33335936e-01
3.80477667e-01 -2.31536150e-01 -9.59417999e-01 2.01321054e+00
4.16252017e-01 1.42813653e-01 4.70334798e-01 5.74071348e-01
2.74472131e-06 6.27389133e-01 8.76953974e-02 -4.57616568e-01
8.19909155e-01 -7.11730182e-01 -4.11460638e-01 -3.01527977e-01
7.91023672e-01 6.85420260e-02 1.05516398e+00 4.83917445e-01
-1.21742404e+00 -5.79474986e-01 -9.96470571e-01 5.37011027e-01
-2.84598470e-01 -5.30485570e-01 6.11505091e-01 1.72995225e-01
-1.15400660e+00 1.10244823e+00 -1.04023683e+00 -4.39707488e-01
7.67700896e-02 6.60596192e-01 -1.43614054e-01 1.84719428e-01
-1.24341261e+00 1.11813700e+00 7.30715275e-01 -2.42048413e-01
-1.79468513e+00 -1.12559356e-01 -9.38161373e-01 1.17964610e-01
8.64442110e-01 -2.23543625e-02 1.60323238e+00 -1.09204352e+00
-1.74766004e+00 2.82878876e-01 3.10748279e-01 -6.91397130e-01
1.85791492e-01 -2.51604110e-01 -1.61092222e-01 2.89820284e-01
6.38339445e-02 4.46220964e-01 9.10750866e-01 -1.27942348e+00
-1.01899624e+00 -2.29503825e-01 4.90565538e-01 5.46560109e-01
-1.45741388e-01 -8.05235133e-02 3.33973020e-02 -5.95317408e-02
-1.66788220e-01 -8.78938973e-01 -9.18970644e-01 -6.65453851e-01
2.20622003e-01 -3.53436261e-01 3.03203732e-01 -1.67092025e-01
1.19594681e+00 -2.00835943e+00 4.93514478e-01 3.58729482e-01
-1.89421847e-02 1.79179385e-01 -3.91573548e-01 8.07842493e-01
1.55019268e-01 -3.00784647e-01 -8.89131725e-02 1.78696021e-01
-5.09250127e-02 4.44617867e-01 -2.18308091e-01 1.00064203e-01
1.06870160e-01 7.23300338e-01 -1.48971677e+00 -3.75546992e-01
1.30117401e-01 -2.25613803e-01 -6.64128840e-01 5.83918452e-01
-6.40318573e-01 6.28919899e-01 -7.92983651e-01 2.67616689e-01
-1.89702868e-01 2.74723824e-02 6.04787588e-01 6.20332599e-01
-3.34585786e-01 5.12213111e-01 -1.39250243e+00 1.55327034e+00
-2.46056795e-01 -9.22188386e-02 -1.50884353e-02 -1.08823383e+00
8.03087711e-01 4.95263934e-01 7.48364747e-01 -7.62555540e-01
-5.31827398e-02 2.34103635e-01 6.68290481e-02 -5.43207169e-01
3.81160378e-01 -3.59583646e-01 -3.54148686e-01 5.96878290e-01
3.68720502e-01 -4.27002013e-01 5.01613259e-01 2.51377642e-01
1.35098660e+00 6.69642389e-01 6.15523934e-01 -9.60948914e-02
2.93361008e-01 2.68902391e-01 6.72792554e-01 1.11142683e+00
-2.96814680e-01 -1.30725637e-01 5.65017462e-01 -5.45021296e-01
-8.19814205e-01 -1.00672042e+00 3.81557435e-01 1.15503013e+00
7.67929927e-02 -6.71762526e-01 -6.59951031e-01 -8.69339168e-01
-2.72866458e-01 6.39647901e-01 -6.99019194e-01 -3.52750272e-01
-7.91639268e-01 -3.10047656e-01 2.04883143e-01 4.70463514e-01
2.31652841e-01 -1.78204954e+00 -1.45590603e+00 5.54336488e-01
3.10039222e-01 -6.47910595e-01 -1.43403023e-01 8.06827664e-01
-1.16098928e+00 -1.11304998e+00 -2.16314688e-01 -7.22747386e-01
6.41955853e-01 -1.97697982e-01 1.03362250e+00 1.40460417e-01
-1.67757034e-01 8.38533819e-01 -3.23812068e-01 -2.27232501e-01
-5.25232196e-01 -2.09404394e-01 4.85590279e-01 -3.49133790e-01
2.43549034e-01 -6.20871842e-01 -1.81425810e-01 6.81643374e-03
-8.50527883e-01 -1.21088047e-02 5.95479727e-01 9.86570776e-01
5.45561314e-01 3.79811198e-01 7.83659518e-01 -7.46275604e-01
8.51548135e-01 -4.45661545e-01 -1.00486267e+00 1.13983952e-01
-7.20439970e-01 4.68484819e-01 7.65860856e-01 -5.55839717e-01
-9.84158814e-01 3.53748560e-01 1.58393174e-01 -1.60401925e-01
-4.19486284e-01 6.07014775e-01 5.82698323e-02 2.40228370e-01
8.39414537e-01 3.49872738e-01 5.14011383e-02 -2.79098004e-01
2.88275748e-01 1.14417136e-01 2.04038829e-01 -9.71132100e-01
7.34077454e-01 -1.13959387e-01 1.71139941e-01 -5.51977575e-01
-7.26186812e-01 -3.06181788e-01 -6.75113559e-01 -3.23294550e-01
5.95764995e-01 -6.59588754e-01 -9.57356215e-01 1.44044116e-01
-4.63297188e-01 -1.10799670e+00 -9.26031530e-01 5.36252022e-01
-1.26192987e+00 6.55713379e-02 -5.83166063e-01 -1.07105386e+00
3.15460563e-01 -1.00412166e+00 5.10293424e-01 3.77979428e-01
-2.75501728e-01 -7.42419600e-01 4.60370839e-01 -3.27574819e-01
2.36155331e-01 1.19375929e-01 8.02150190e-01 -5.82271993e-01
-6.86903179e-01 1.64866537e-01 5.35813630e-01 1.48338094e-01
1.67407230e-01 -6.61173940e-01 -5.42709768e-01 -6.26454234e-01
1.20165713e-01 -9.65196550e-01 4.59060937e-01 2.14832455e-01
7.92005718e-01 -3.23320419e-01 -1.01188771e-01 -1.35059059e-01
1.47265589e+00 7.06716657e-01 4.22490954e-01 6.02539957e-01
1.71753481e-01 6.49290919e-01 1.11838102e+00 6.65113747e-01
1.76572978e-01 5.71443439e-01 6.36985123e-01 1.24660373e-01
2.43831888e-01 -4.81364965e-01 7.32526302e-01 3.53057325e-01
-3.36355865e-01 5.11078179e-01 -7.28817821e-01 5.60375035e-01
-2.18303347e+00 -1.21698427e+00 6.11449957e-01 2.52996182e+00
1.05793571e+00 3.61675203e-01 5.73856235e-01 5.68719544e-02
2.38301292e-01 -6.99270144e-02 -8.98133159e-01 -3.96059990e-01
4.62029606e-01 2.65004009e-01 2.27923796e-01 8.80064428e-01
-8.43989491e-01 1.24829507e+00 6.91709280e+00 3.87479037e-01
-6.94748640e-01 -1.43591180e-01 2.00544730e-01 -9.88974124e-02
-8.21836665e-02 2.47085705e-01 -8.34809959e-01 4.56160642e-02
1.12684870e+00 -6.17520474e-02 9.14646268e-01 8.53824973e-01
2.53927022e-01 -4.42856610e-01 -1.34576833e+00 3.30073208e-01
-3.58729750e-01 -1.02278316e+00 -3.66281182e-01 3.93248834e-02
6.65863991e-01 -1.02686442e-01 -1.54871926e-01 8.33511114e-01
1.00449049e+00 -9.12757158e-01 5.68518102e-01 4.29361343e-01
4.77414876e-01 -9.22118068e-01 3.42155784e-01 8.11165452e-01
-1.05635893e+00 -6.73168361e-01 -2.14450270e-01 -5.13233304e-01
-1.87629268e-01 -3.53737712e-01 -9.51027036e-01 2.90769845e-01
5.28512299e-01 5.69740951e-01 -3.19715768e-01 7.18254685e-01
-4.73064244e-01 5.15145779e-01 -1.32966384e-01 -4.19854134e-01
5.42160094e-01 -1.54606268e-01 3.77414018e-01 7.09027946e-01
1.50128007e-01 3.68771285e-01 6.92512870e-01 6.15111053e-01
3.77253085e-01 -6.47996813e-02 -8.94037783e-01 -1.13018662e-01
3.04839194e-01 9.43411708e-01 -6.52189672e-01 -4.67323959e-01
-1.10658921e-01 6.85557604e-01 6.66067541e-01 4.37536120e-01
-4.44745153e-01 -8.43386650e-02 5.02294958e-01 -2.40453240e-02
2.47726098e-01 -3.22346926e-01 3.71635824e-01 -9.91388500e-01
-3.99735361e-01 -1.26439297e+00 5.07006407e-01 -5.08807659e-01
-5.69142818e-01 4.35195923e-01 1.78220153e-01 -1.31257689e+00
-8.45250905e-01 -2.68306941e-01 -3.80954266e-01 6.58520520e-01
-1.52920461e+00 -5.62143028e-01 1.47331551e-01 8.88075769e-01
7.05632269e-01 -3.63663584e-01 1.15356421e+00 -4.05529261e-01
-1.17285714e-01 -9.65025499e-02 -3.27922143e-02 -2.21308500e-01
3.15248728e-01 -1.68054116e+00 -1.51689509e-02 7.22120583e-01
1.89084932e-02 5.61959803e-01 8.59632611e-01 -7.56106913e-01
-1.39733171e+00 -6.88760519e-01 5.82735002e-01 -4.67947513e-01
6.86742187e-01 -2.67455041e-01 -7.69670308e-01 8.32958460e-01
7.01291934e-02 -2.52953678e-01 3.12522233e-01 2.44157344e-01
5.74726127e-02 -4.14722301e-02 -1.00013816e+00 8.36045861e-01
8.52190495e-01 -3.89809251e-01 -8.70489120e-01 1.57017246e-01
6.17373526e-01 -2.95701236e-01 -6.98062003e-01 1.44662961e-01
2.90016085e-01 -8.68562281e-01 7.84312725e-01 -1.08617651e+00
2.47718260e-01 -4.31302339e-01 8.87795091e-02 -1.62826955e+00
-4.82868910e-01 -9.20653343e-01 -3.32103491e-01 6.69476688e-01
1.22234851e-01 -3.88130814e-01 7.40925848e-01 5.18988550e-01
-2.15884019e-02 -6.94148660e-01 -7.02568889e-01 -9.84340131e-01
-5.10151275e-02 -2.31272787e-01 5.95154583e-01 5.95692158e-01
6.49748683e-01 4.76405293e-01 -5.96275151e-01 1.24418490e-01
6.72824264e-01 2.54860342e-01 7.46139586e-01 -1.15170634e+00
-6.64296627e-01 -1.22685544e-01 -8.87807682e-02 -9.73176658e-01
9.33324471e-02 -5.73561430e-01 4.94320422e-01 -1.37519300e+00
3.81322913e-02 -4.60616797e-01 -7.29530334e-01 6.98140919e-01
6.97405711e-02 -5.27088761e-01 2.99393058e-01 1.26283273e-01
-1.02110660e+00 5.01358211e-01 1.30858731e+00 6.67948872e-02
-7.94260085e-01 8.94305930e-02 -4.27337587e-01 7.64770091e-01
1.07536101e+00 -5.26647151e-01 -7.39176869e-01 1.15394741e-01
2.38057658e-01 5.50685227e-01 6.05459996e-02 -1.14373100e+00
2.75327295e-01 -7.08388507e-01 1.40685856e-01 -1.76985800e-01
4.33479548e-01 -9.14395511e-01 -8.58196020e-02 9.93269324e-01
-8.46503079e-01 1.42962053e-01 8.86069909e-02 6.30988598e-01
-9.08795968e-02 -5.08288741e-01 6.73561335e-01 -6.12859964e-01
-1.17751360e+00 1.23469427e-01 -9.80190456e-01 -8.24657530e-02
1.18699098e+00 -7.29405209e-02 1.37508184e-01 -4.30270821e-01
-1.11042643e+00 5.85223079e-01 3.24736953e-01 2.92991400e-01
7.75843918e-01 -1.09821498e+00 -3.94673854e-01 2.78850019e-01
6.22338615e-02 -1.37117565e-01 -1.15470074e-01 4.47022021e-01
4.52347957e-02 2.30414808e-01 -6.52581394e-01 -1.09681867e-01
-1.07167482e+00 8.80454659e-01 3.87884051e-01 -7.06634462e-01
-6.45294189e-01 3.83847773e-01 3.28482427e-02 -3.87480557e-01
3.20148051e-01 -2.32646555e-01 -5.88449419e-01 -1.27130538e-01
6.70778692e-01 1.97873384e-01 -3.68848324e-01 -1.14772707e-01
-7.43521079e-02 1.71059668e-01 -2.45509148e-02 -5.61572790e-01
1.44478166e+00 -6.02800995e-02 4.94536906e-02 7.39411116e-01
4.53995496e-01 -2.83584535e-01 -1.77964711e+00 -3.67114574e-01
4.45082337e-01 -3.59812498e-01 -2.63861835e-01 -7.89422750e-01
-3.26308787e-01 4.80838835e-01 5.90331495e-01 4.64624852e-01
1.18357253e+00 -4.94273677e-02 2.90037066e-01 8.03591549e-01
1.01816237e+00 -1.69041133e+00 6.76927090e-01 8.08327019e-01
6.97633564e-01 -1.13569260e+00 -2.35396013e-01 3.73993188e-01
-7.74865270e-01 1.02095699e+00 8.31772089e-01 -2.62943655e-01
3.08946133e-01 1.19752057e-01 -9.44395643e-03 -2.84987807e-01
-1.01318860e+00 -6.16363108e-01 -2.91210949e-01 6.98623478e-01
-5.29303141e-02 1.84962396e-02 -1.70795798e-01 2.08182216e-01
1.86534941e-01 1.09928288e-01 6.19084179e-01 1.47530389e+00
-1.00446296e+00 -1.32626331e+00 -3.30931067e-01 3.11558515e-01
-1.27816409e-01 1.85456142e-01 -1.70933947e-01 6.59951985e-01
-2.69207302e-02 8.28883648e-01 -1.43090814e-01 -2.10658103e-01
3.90485018e-01 1.37761027e-01 7.47345090e-01 -1.09629488e+00
-6.48304641e-01 1.89974844e-01 1.88014418e-01 -1.00821054e+00
-3.35935205e-01 -9.03256297e-01 -1.67161071e+00 2.37964004e-01
5.23277260e-02 5.57039440e-01 2.51636416e-01 1.05428135e+00
2.94399299e-02 6.00395679e-01 8.94315183e-01 -8.61411512e-01
-9.76963997e-01 -6.20227218e-01 -8.50390792e-01 5.32331288e-01
5.20647347e-01 -6.97497606e-01 -7.47702494e-02 -1.19439289e-01] | [4.104305744171143, 1.5860481262207031] |
75b10880-274f-4869-818f-4ffc536f6737 | neural-network-pruning-for-real-time-polyp | 2306.13203 | null | https://arxiv.org/abs/2306.13203v1 | https://arxiv.org/pdf/2306.13203v1.pdf | Neural Network Pruning for Real-time Polyp Segmentation | Computer-assisted treatment has emerged as a viable application of medical imaging, owing to the efficacy of deep learning models. Real-time inference speed remains a key requirement for such applications to help medical personnel. Even though there generally exists a trade-off between performance and model size, impressive efforts have been made to retain near-original performance by compromising model size. Neural network pruning has emerged as an exciting area that aims to eliminate redundant parameters to make the inference faster. In this study, we show an application of neural network pruning in polyp segmentation. We compute the importance score of convolutional filters and remove the filters having the least scores, which to some value of pruning does not degrade the performance. For computing the importance score, we use the Taylor First Order (TaylorFO) approximation of the change in network output for the removal of certain filters. Specifically, we employ a gradient-normalized backpropagation for the computation of the importance score. Through experiments in the polyp datasets, we validate that our approach can significantly reduce the parameter count and FLOPs retaining similar performance. | ['Binod Bhattarai', 'Bibek Panthi', 'Sudarshan Regmi', 'Pranav Poudel', 'Suman Sapkota'] | 2023-06-22 | null | null | null | null | ['network-pruning'] | ['methodology'] | [ 5.15151322e-01 2.60647118e-01 1.07111134e-01 -4.41112995e-01
-5.58807671e-01 -1.76482782e-01 8.22009668e-02 5.36632895e-01
-8.45631480e-01 5.04743516e-01 -1.91038966e-01 -5.26321352e-01
-3.77202153e-01 -6.45137668e-01 -5.64411759e-01 -6.40509963e-01
-1.31105721e-01 9.36035663e-02 3.64176273e-01 8.40608105e-02
4.12876517e-01 6.90493226e-01 -1.20158243e+00 3.08653027e-01
8.36145639e-01 1.06744611e+00 1.46325566e-02 7.84373641e-01
2.59113491e-01 7.44784236e-01 -5.63292682e-01 -2.40086243e-01
2.86890894e-01 -2.74268597e-01 -6.61621928e-01 -2.15732217e-01
2.25431442e-01 -4.56892908e-01 -3.13841343e-01 1.13700187e+00
5.29456198e-01 1.06930025e-01 3.57358396e-01 -6.78713977e-01
1.23170279e-01 5.57236671e-01 -3.84885430e-01 5.86837292e-01
-2.39573330e-01 3.41894589e-02 6.48323357e-01 -4.63961989e-01
3.20480198e-01 8.78844798e-01 8.77100885e-01 3.41771215e-01
-1.12470150e+00 -1.88894272e-01 -1.10711545e-01 3.91489156e-02
-1.37903440e+00 -2.77371794e-01 4.04851109e-01 -9.62897763e-02
1.09116852e+00 4.03484285e-01 7.85246313e-01 1.81408763e-01
5.91561317e-01 5.79833925e-01 6.21619523e-01 -6.87168181e-01
4.87255245e-01 7.20433444e-02 1.90268099e-01 8.69385481e-01
4.24052387e-01 -5.78952096e-02 -3.30996245e-01 -3.06543201e-01
8.60697031e-01 4.64621074e-02 -4.00041044e-01 -1.96853951e-02
-7.81542420e-01 8.04406941e-01 6.66304827e-01 2.82800376e-01
-4.82489377e-01 4.94282186e-01 3.74668509e-01 1.94203500e-02
4.32332069e-01 6.57737434e-01 -6.17640316e-01 -4.49842066e-02
-1.19072938e+00 1.81837574e-01 7.40908504e-01 3.22146177e-01
2.49215722e-01 -5.83250709e-02 -1.75474957e-01 7.06098378e-01
5.85372262e-02 8.78321193e-03 5.39919317e-01 -1.18195057e+00
3.21386941e-02 6.84349656e-01 -2.49764025e-01 -1.00135577e+00
-5.98039567e-01 -8.76505136e-01 -8.55494797e-01 2.82376528e-01
4.90912318e-01 -2.60498852e-01 -9.82264102e-01 1.46121013e+00
2.61322498e-01 9.52327717e-03 -2.09839046e-01 6.39666021e-01
3.70138913e-01 2.90205419e-01 1.05752692e-01 -2.66453147e-01
1.25086832e+00 -7.09178925e-01 -3.85428816e-01 -2.04403624e-01
7.30758905e-01 -6.30442619e-01 6.32458985e-01 6.19168758e-01
-1.32195926e+00 -2.49453187e-01 -1.05740345e+00 -1.98674873e-02
-7.61081055e-02 9.80939344e-02 6.31724954e-01 6.84307754e-01
-9.88315701e-01 1.13559031e+00 -1.23324156e+00 -4.26822156e-02
6.23486340e-01 8.23887825e-01 7.55148306e-02 7.72652104e-02
-7.92741001e-01 9.96089220e-01 4.18797076e-01 3.31802517e-01
-3.51776540e-01 -9.72904563e-01 -6.02511287e-01 4.93572921e-01
2.80721337e-01 -8.12171876e-01 1.34045088e+00 -8.33266437e-01
-1.33217478e+00 4.71951962e-01 1.33160036e-02 -9.07902718e-01
4.28188175e-01 -1.53755471e-01 -1.41443443e-02 3.70735049e-01
-3.59733969e-01 9.29761648e-01 8.04488480e-01 -5.68216741e-01
-7.73791015e-01 -2.61123985e-01 1.43721133e-01 1.94864035e-01
-5.92013299e-01 -1.69017509e-01 -5.60897052e-01 -5.59905112e-01
4.49127436e-01 -1.04582298e+00 -5.81687212e-01 3.59198451e-01
-9.75051373e-02 1.42515019e-01 3.49385738e-01 -7.20677137e-01
1.20557559e+00 -2.09505510e+00 -2.82108486e-01 3.60341996e-01
3.86564195e-01 3.16426158e-01 2.99548984e-01 -1.78102374e-01
-7.44863078e-02 -6.65959716e-02 -4.89208221e-01 -5.62908426e-02
-5.59372425e-01 3.11131448e-01 2.07574740e-01 4.72434133e-01
2.72441655e-01 6.96258128e-01 -6.47832632e-01 -5.50115764e-01
1.32214963e-01 6.92505538e-01 -1.06687355e+00 -2.48745561e-01
1.66373432e-03 3.79011482e-02 -3.49755049e-01 2.70640135e-01
3.67729604e-01 -4.05722648e-01 2.64140487e-01 -3.82578462e-01
-3.34886424e-02 2.80506492e-01 -9.27078664e-01 1.39249921e+00
-4.57423151e-01 6.52491391e-01 1.18105508e-01 -1.04819620e+00
4.97252166e-01 2.88295358e-01 7.13492572e-01 -5.29315829e-01
4.16553348e-01 2.83440173e-01 5.05907297e-01 -2.96718925e-01
3.56923372e-01 -4.35580760e-02 3.76347244e-01 1.39363497e-01
-3.09853226e-01 -1.84541449e-01 3.71679544e-01 6.27247766e-02
1.31675708e+00 -2.07237154e-01 3.65209222e-01 -4.10571665e-01
4.73250002e-01 2.35499308e-01 4.47876990e-01 6.38230145e-01
-1.28991410e-01 5.72325289e-01 4.64905947e-01 -5.74324250e-01
-8.57452631e-01 -7.51171589e-01 -2.02694833e-01 6.63349092e-01
-1.04312606e-01 -1.32865623e-01 -9.22113359e-01 -4.51173067e-01
-1.66981012e-01 5.75360835e-01 -4.85497594e-01 -3.03191215e-01
-9.14913654e-01 -1.09339166e+00 4.64682341e-01 6.04770184e-01
4.63341922e-01 -9.26522732e-01 -1.41335714e+00 3.43709111e-01
4.37117666e-02 -8.56794477e-01 -2.31679708e-01 4.28336591e-01
-1.54363382e+00 -1.05917859e+00 -7.70105302e-01 -6.88786626e-01
1.06558132e+00 -1.35237025e-02 8.98222208e-01 5.12149572e-01
-6.16049886e-01 7.52262548e-02 1.07784234e-01 -4.61881518e-01
-3.72951925e-01 2.60867715e-01 -3.00896943e-01 -5.53623796e-01
9.87219736e-02 -4.62428629e-01 -8.61495256e-01 -1.49979040e-01
-9.15430307e-01 8.13108087e-02 7.80420244e-01 7.97937810e-01
6.62788093e-01 3.84310275e-01 3.49081725e-01 -8.61678600e-01
7.77252555e-01 2.02908479e-02 -6.94491088e-01 4.19867272e-03
-6.97517276e-01 2.45739907e-01 5.92161477e-01 -4.06830996e-01
-8.01645219e-01 2.71145701e-01 -2.26715252e-01 -2.80320406e-01
1.71924055e-01 6.78334951e-01 5.66676080e-01 -3.30316782e-01
6.75614536e-01 -1.60960436e-01 8.32719356e-02 -2.82593608e-01
4.49811183e-02 2.67095923e-01 5.27886689e-01 -1.79528773e-01
2.00433478e-01 5.34948230e-01 2.68713355e-01 -8.06527317e-01
-7.58154094e-01 -3.13008279e-01 -2.98517704e-01 -6.54287487e-02
5.85054457e-01 -4.09603983e-01 -9.13539231e-01 8.91973376e-02
-1.13039911e+00 -7.98479021e-02 -4.83123720e-01 6.57351971e-01
-2.89218128e-01 2.37416357e-01 -7.04117537e-01 -6.13020778e-01
-8.07385921e-01 -1.21814656e+00 5.75527668e-01 3.54182273e-01
-3.42897415e-01 -8.16273153e-01 -3.07288021e-01 -8.88831913e-02
5.08758426e-01 2.07020089e-01 1.17254484e+00 -4.50776905e-01
-3.92735630e-01 -4.53214884e-01 -2.12837294e-01 5.21678507e-01
-6.89128414e-02 -1.28104672e-01 -8.08362126e-01 -1.42995983e-01
3.46232086e-01 2.00327992e-01 8.49386692e-01 8.97150636e-01
1.46346164e+00 -9.27906111e-02 -3.46270293e-01 6.33457899e-01
1.36988950e+00 3.32576394e-01 5.66265881e-01 1.96195811e-01
2.81584591e-01 2.59380400e-01 2.52845258e-01 3.82724404e-01
-2.18108818e-01 2.54620224e-01 3.99148703e-01 -2.68862069e-01
-3.84348817e-02 3.34435791e-01 -1.85482055e-01 8.36028755e-01
-7.48048872e-02 7.02736154e-02 -9.47683632e-01 4.01655883e-01
-1.60995603e+00 -5.54986238e-01 -4.54201885e-02 2.28258944e+00
7.72387505e-01 7.36505270e-01 -2.84675479e-01 4.39005435e-01
3.54434788e-01 -3.75388384e-01 -7.28261650e-01 -5.49558342e-01
4.31107223e-01 6.50032640e-01 8.73951733e-01 5.23180068e-01
-8.88147652e-01 4.58113730e-01 6.45137215e+00 7.06797183e-01
-1.22960114e+00 -4.48554158e-02 7.78419435e-01 -3.46760035e-01
3.13162386e-01 -1.97541326e-01 -6.32884920e-01 2.96524346e-01
9.24020290e-01 5.68763586e-03 2.26610258e-01 9.11796689e-01
2.62881339e-01 -4.72815961e-01 -9.62819636e-01 8.23225200e-01
-1.86606422e-01 -1.34423470e+00 -1.89804256e-01 -1.86055060e-02
4.83935505e-01 -8.34967792e-02 -5.44713587e-02 -5.23196273e-02
-8.22757408e-02 -8.62544239e-01 2.94706017e-01 3.50315332e-01
3.40983063e-01 -8.70444834e-01 8.84456158e-01 3.46155047e-01
-7.24390209e-01 -2.28967667e-01 -3.70990455e-01 -9.52443630e-02
9.25962329e-02 8.74135017e-01 -1.03869283e+00 6.68294728e-02
6.79720700e-01 1.06140934e-01 -4.22821790e-01 1.43682587e+00
4.45656963e-02 5.59734046e-01 -6.86827600e-01 -5.01408763e-02
1.44202158e-01 3.90343666e-02 2.95977920e-01 1.00873375e+00
2.46626005e-01 2.59892166e-01 -1.97767287e-01 5.51324427e-01
1.47175482e-02 -5.26544601e-02 -7.21700350e-03 -1.08952904e-02
2.75758713e-01 1.02383411e+00 -1.16840041e+00 -3.81813347e-01
-1.04350574e-01 8.20646822e-01 2.35507101e-01 2.65280697e-02
-7.99096644e-01 -4.03090775e-01 3.98213714e-01 2.10116580e-01
3.38581055e-01 -2.35184580e-01 -7.33683765e-01 -5.50375879e-01
-1.52002266e-02 -7.88664758e-01 3.12281162e-01 -2.65635461e-01
-5.65033019e-01 5.53085983e-01 -3.51243168e-02 -8.44209492e-01
-2.71877110e-01 -5.97088099e-01 -3.27765673e-01 7.03679085e-01
-1.34311485e+00 -4.67479646e-01 -2.35977784e-01 2.24058211e-01
5.20855784e-01 2.39529595e-01 8.20036769e-01 5.50597906e-01
-3.66576612e-01 5.69108069e-01 -4.39989306e-02 -1.36500239e-01
2.81168461e-01 -8.26961815e-01 1.23950385e-01 8.22488129e-01
-1.93838894e-01 7.78808475e-01 7.84041643e-01 -6.23232543e-01
-1.12105501e+00 -7.16661572e-01 7.23062932e-01 8.04989934e-02
1.76961750e-01 1.66202366e-01 -8.22712183e-01 3.56615365e-01
-4.26764078e-02 6.29195422e-02 3.66957098e-01 -1.80844858e-01
3.22943777e-01 -1.40028641e-01 -1.23906231e+00 6.61231875e-01
6.31445825e-01 3.38864070e-03 -3.66203249e-01 2.88012654e-01
4.94255424e-01 -5.95086455e-01 -8.02294016e-01 6.18265271e-01
6.64159834e-01 -9.75534558e-01 9.58939016e-01 -2.24832267e-01
4.37245071e-01 -1.10331625e-01 2.04190254e-01 -1.06380332e+00
-2.43810281e-01 -3.06199223e-01 -2.02841640e-01 3.87594253e-01
3.96395415e-01 -5.03111720e-01 1.14867020e+00 8.96888077e-01
-1.45845905e-01 -1.25077260e+00 -8.50337148e-01 -4.34681803e-01
-4.49957065e-02 -3.98208231e-01 1.31137341e-01 4.21397746e-01
-1.19061202e-01 5.25762467e-03 7.79521093e-02 1.55445531e-01
4.85101879e-01 -2.34775752e-01 1.94158494e-01 -1.21057498e+00
-5.21717191e-01 -6.03701234e-01 -3.53050917e-01 -9.59592164e-01
-4.40527856e-01 -5.76104164e-01 1.42434046e-01 -1.51498699e+00
-8.82600155e-03 -3.73482674e-01 -3.59881282e-01 4.44785714e-01
-1.60950482e-01 4.49266255e-01 4.87894639e-02 -1.57808959e-02
-3.74108143e-02 -4.63163890e-02 1.14391410e+00 -7.51096532e-02
-4.14370209e-01 1.30499780e-01 -6.01314783e-01 1.13106596e+00
9.52782750e-01 -7.66409457e-01 -4.27958578e-01 -6.19579792e-01
3.47866356e-01 1.42549071e-02 1.78855360e-01 -1.30972862e+00
4.72439617e-01 2.69839853e-01 5.66774666e-01 -6.62228107e-01
4.59369361e-01 -8.38916183e-01 -4.36364226e-02 1.12974870e+00
-4.79468405e-01 -5.33141941e-03 3.88546914e-01 1.72102869e-01
-9.84184891e-02 -6.65060818e-01 1.02866364e+00 -1.84063599e-01
-2.68654197e-01 -5.23594096e-02 -4.14379984e-01 -1.85110778e-01
7.58006394e-01 -3.89394403e-01 8.07859302e-02 -8.58258158e-02
-7.57356822e-01 6.92830384e-02 4.74641174e-02 -1.20964013e-01
6.71340883e-01 -6.51952088e-01 -4.86412197e-01 1.80643320e-01
-5.74453235e-01 2.10277200e-01 2.70349503e-01 9.80144739e-01
-1.05615020e+00 4.03907150e-01 -1.18792370e-01 -5.51450193e-01
-1.50122762e+00 1.83720157e-01 5.63863158e-01 -5.34301221e-01
-7.16142595e-01 9.42813098e-01 -3.79824825e-02 4.25669327e-02
4.29455996e-01 -6.14995718e-01 -6.34972230e-02 -2.37387151e-01
6.46953523e-01 5.80758154e-01 5.97632885e-01 8.67868438e-02
-3.81766349e-01 2.18124792e-01 -2.64649808e-01 -4.66506258e-02
1.48762667e+00 2.26776227e-01 -7.53973201e-02 -1.44346789e-01
1.00785983e+00 -2.77907431e-01 -1.03817403e+00 1.16008937e-01
-4.10602503e-02 -1.70054197e-01 6.72978461e-01 -8.52629542e-01
-1.27838397e+00 8.74656022e-01 8.78925264e-01 1.38342008e-01
1.33349013e+00 -4.81573641e-01 8.97584319e-01 4.68491942e-01
6.54625073e-02 -1.15714574e+00 -2.39215732e-01 3.33686680e-01
4.20320690e-01 -9.27983761e-01 5.87518215e-01 -5.65021396e-01
-6.19255789e-02 1.16516900e+00 4.07936990e-01 -2.89955795e-01
8.55204761e-01 5.32344699e-01 -5.55260032e-02 -1.61988959e-01
-4.66447502e-01 1.55379891e-01 1.69681042e-01 1.50684819e-01
5.20256579e-01 -1.42556638e-01 -7.31626391e-01 2.56733656e-01
-1.91849589e-01 3.60736817e-01 3.94204736e-01 1.15207767e+00
-5.99153459e-01 -8.43084931e-01 -2.57167518e-01 8.61129105e-01
-8.33377898e-01 -3.97220105e-01 9.86289978e-02 7.31033266e-01
3.58142257e-02 7.72216439e-01 2.43570179e-01 1.69813186e-01
2.98871785e-01 -5.87734357e-02 7.25510120e-01 -4.54752028e-01
-9.27483678e-01 8.56470466e-02 -1.29910484e-01 -3.48282278e-01
-1.95203349e-01 -1.80705860e-01 -1.49794376e+00 -2.60799527e-01
-4.67154205e-01 -3.27546075e-02 9.38354909e-01 8.76293659e-01
2.60178268e-01 9.24157798e-01 2.66787652e-02 -6.83260381e-01
-7.23865509e-01 -6.00622714e-01 -2.73391575e-01 4.17832732e-02
1.02167733e-01 -2.72543013e-01 -2.13037565e-01 -1.11969979e-02] | [8.607951164245605, 3.0934510231018066] |
324d258b-1584-47af-bd05-b291db27c58a | temporal-information-extraction-by-predicting | 1808.09401 | null | http://arxiv.org/abs/1808.09401v1 | http://arxiv.org/pdf/1808.09401v1.pdf | Temporal Information Extraction by Predicting Relative Time-lines | The current leading paradigm for temporal information extraction from text
consists of three phases: (1) recognition of events and temporal expressions,
(2) recognition of temporal relations among them, and (3) time-line
construction from the temporal relations. In contrast to the first two phases,
the last phase, time-line construction, received little attention and is the
focus of this work. In this paper, we propose a new method to construct a
linear time-line from a set of (extracted) temporal relations. But more
importantly, we propose a novel paradigm in which we directly predict start and
end-points for events from the text, constituting a time-line without going
through the intermediate step of prediction of temporal relations as in earlier
work. Within this paradigm, we propose two models that predict in linear
complexity, and a new training loss using TimeML-style annotations, yielding
promising results. | ['Marie-Francine Moens', 'Artuur Leeuwenberg'] | 2018-08-28 | temporal-information-extraction-by-predicting-1 | https://aclanthology.org/D18-1155 | https://aclanthology.org/D18-1155.pdf | emnlp-2018-10 | ['temporal-information-extraction'] | ['natural-language-processing'] | [ 4.35568064e-01 2.24305555e-01 -1.77287504e-01 -3.49229813e-01
-6.08742535e-01 -8.23039174e-01 1.20361340e+00 6.38694227e-01
-4.55549777e-01 5.97638130e-01 4.75515053e-02 -4.43637878e-01
-3.76347363e-01 -6.79574251e-01 -3.90505970e-01 -2.87266344e-01
-5.06618977e-01 3.95503968e-01 6.34684980e-01 -1.45563141e-01
2.47811735e-01 7.71486819e-01 -1.48552775e+00 3.00274193e-01
6.07347727e-01 1.16107285e+00 -4.81595606e-01 7.53941119e-01
-2.79912353e-01 1.30429006e+00 -3.71603280e-01 -2.43929073e-01
3.72313932e-02 -6.97782457e-01 -1.16503572e+00 1.15768135e-01
-3.20410520e-01 -2.88626570e-02 -2.34216824e-01 2.93361634e-01
-4.77837548e-02 4.07236069e-01 5.62324107e-01 -1.38883638e+00
1.32518575e-01 7.66119242e-01 -2.87355870e-01 9.51993838e-02
7.03581572e-01 -3.18143010e-01 1.08388829e+00 -5.62870800e-01
8.91557992e-01 8.02496850e-01 7.50207424e-01 2.67728209e-01
-1.28001881e+00 -2.55805194e-01 3.27321589e-01 4.73212600e-01
-1.43523729e+00 -4.71970290e-01 1.02977240e+00 -5.73038578e-01
1.31857288e+00 2.92951703e-01 8.49885702e-01 9.14055109e-01
-1.80324558e-02 7.35337257e-01 9.60357308e-01 -8.20383549e-01
2.91429549e-01 -1.12679288e-01 2.89895594e-01 2.82665074e-01
-6.28749788e-01 1.30555362e-01 -6.25964284e-01 -1.09404564e-01
3.55721295e-01 -3.05015653e-01 1.07311048e-02 -1.91026449e-01
-1.18726456e+00 3.48680884e-01 -1.39975891e-01 7.72658765e-01
-2.80368268e-01 -7.22438246e-02 7.30076194e-01 3.04752916e-01
6.46337748e-01 1.88089818e-01 -5.45326591e-01 -5.00888169e-01
-1.16227543e+00 4.55681264e-01 1.16195309e+00 1.02218354e+00
4.79938179e-01 -4.76411015e-01 -1.63999468e-01 3.43757004e-01
8.90791267e-02 -2.50757039e-01 3.45748544e-01 -5.57506800e-01
5.20214975e-01 7.64013827e-01 4.55702811e-01 -9.56968367e-01
-7.44399488e-01 -1.79114699e-01 -3.74798834e-01 -1.95242330e-01
8.25766683e-01 5.04223350e-03 -5.35615087e-01 1.80576444e+00
5.34013569e-01 4.32218641e-01 4.29528533e-03 4.14241523e-01
2.91058898e-01 7.97156751e-01 5.30145951e-02 -8.39786589e-01
1.29048610e+00 -6.79784238e-01 -9.71735477e-01 -1.47918530e-03
1.05813825e+00 -6.21570528e-01 4.93673205e-01 5.60045540e-01
-1.16206622e+00 -2.08927318e-01 -9.07720268e-01 -2.25384727e-01
-5.61350405e-01 6.18155748e-02 8.68976176e-01 8.11316594e-02
-9.85028267e-01 9.23706055e-01 -9.83957767e-01 -4.79303479e-01
-2.72172004e-01 2.83771902e-01 -2.97986776e-01 5.99370003e-01
-1.23050976e+00 9.91455913e-01 5.23057342e-01 2.45090172e-01
-3.54504675e-01 -5.25087416e-01 -9.11735773e-01 -6.42182678e-02
7.22847581e-01 -2.08007351e-01 1.53876948e+00 -7.83900142e-01
-1.53264737e+00 8.60827208e-01 -4.94876474e-01 -6.70968771e-01
7.82210469e-01 -3.30503702e-01 -5.86491525e-01 3.48816402e-02
-1.32613376e-01 1.63710654e-01 5.87387800e-01 -8.91321123e-01
-7.85152078e-01 -2.49613568e-01 1.60197690e-01 -2.37353712e-01
-8.08031112e-02 5.44104576e-01 -6.07528865e-01 -5.92054725e-01
1.96973637e-01 -8.36867273e-01 -2.37970456e-01 -1.59360901e-01
-4.32690978e-01 -8.21402669e-01 5.42224169e-01 -6.28578961e-01
1.89031339e+00 -1.97647166e+00 3.53082031e-01 1.61950096e-01
3.34149338e-02 4.64408547e-02 2.44316563e-01 8.35965455e-01
-4.86573964e-01 1.69664226e-03 -3.58773977e-01 -4.24968392e-01
2.16913909e-01 1.40924037e-01 -6.99935973e-01 3.81311297e-01
4.69328701e-01 7.01426864e-01 -1.11047816e+00 -6.70770347e-01
2.83594102e-01 2.29206547e-01 -2.72861212e-01 2.30791852e-01
-5.41461289e-01 4.54947829e-01 -3.91942859e-01 2.86855221e-01
1.52851522e-01 -6.54610470e-02 4.77053761e-01 -1.88981801e-01
-6.88200235e-01 7.11529016e-01 -1.15960538e+00 1.63922310e+00
-3.89710069e-01 4.97434646e-01 -5.83689213e-01 -9.37403023e-01
9.22030151e-01 7.66525447e-01 1.10003972e+00 -8.06888342e-01
1.98822573e-01 2.52837032e-01 -2.70953208e-01 -5.63431442e-01
5.63234329e-01 -4.48729992e-02 -3.52965087e-01 6.70407057e-01
1.86823130e-01 -1.42049007e-02 7.99109042e-01 8.09278041e-02
1.20892048e+00 7.42026031e-01 7.92497277e-01 2.79894501e-01
6.68015480e-01 6.21241182e-02 7.38620222e-01 2.90255010e-01
-1.97189793e-01 1.57690585e-01 1.02931345e+00 -6.94366217e-01
-1.09364140e+00 -7.37821043e-01 7.21469000e-02 7.66081572e-01
-2.49982700e-01 -1.06164420e+00 -4.00884360e-01 -9.27348316e-01
-5.09394765e-01 7.79871047e-01 -6.93373203e-01 2.09452897e-01
-1.18484914e+00 -3.94869983e-01 5.78145325e-01 6.18461907e-01
-1.34245187e-01 -9.73479390e-01 -8.15633535e-01 3.86726201e-01
-3.45839441e-01 -1.35459828e+00 -3.90115947e-01 6.02222860e-01
-8.11200559e-01 -1.18923140e+00 -2.00248331e-01 -5.26931167e-01
4.20861691e-01 -3.34257811e-01 1.06631136e+00 -3.60603452e-01
3.78884636e-02 5.37300631e-02 -6.91388547e-01 -4.51464593e-01
-2.40903616e-01 -6.91209920e-03 -2.02053055e-01 2.25514174e-01
2.29766101e-01 -9.00552690e-01 -2.17566043e-01 1.85375512e-01
-9.41241443e-01 3.78507078e-01 1.62742272e-01 3.61183286e-01
5.60062826e-01 1.54058605e-01 3.31631541e-01 -6.92522407e-01
2.48793393e-01 -4.17497814e-01 -8.06784570e-01 3.59582931e-01
-4.85282123e-01 1.19999409e-01 9.68293250e-01 -5.37007630e-01
-1.09866858e+00 4.54186767e-01 -3.92207578e-02 -1.11780122e-01
-2.98510075e-01 7.98290014e-01 1.63993970e-01 4.23832625e-01
6.68061912e-01 2.44316593e-01 -5.85964561e-01 -5.13993382e-01
4.90245223e-01 2.03467935e-01 6.07794344e-01 -7.06239879e-01
8.25795770e-01 6.62101507e-01 3.10513794e-01 -5.04341066e-01
-9.10033166e-01 -5.73856175e-01 -1.24359381e+00 -3.52090955e-01
6.20915890e-01 -4.33411360e-01 -6.31257355e-01 2.20217347e-01
-1.53687775e+00 -3.84779871e-01 -5.80069900e-01 5.67809165e-01
-8.22773576e-01 4.50507194e-01 -5.94594836e-01 -1.11316633e+00
3.64818759e-02 -5.81198156e-01 9.98143911e-01 -9.42127630e-02
-8.17316115e-01 -9.10257101e-01 1.53424665e-01 -9.76029783e-02
-2.92628646e-01 5.79442084e-01 7.33118892e-01 -8.84652793e-01
-4.09993142e-01 -3.95909131e-01 -2.72842627e-02 -7.56237134e-02
5.90523705e-02 2.99507767e-01 -7.73597538e-01 2.20361620e-01
1.50702715e-01 -4.49863635e-02 3.72274369e-01 1.24040604e-01
8.90482426e-01 -3.25497955e-01 -4.36192155e-01 5.27982652e-01
1.12298954e+00 9.00959015e-01 6.32446945e-01 3.12451899e-01
3.24587286e-01 1.09018993e+00 9.05328095e-01 6.02730393e-01
4.80837077e-01 1.15357256e+00 1.42355158e-03 2.47592077e-01
1.77075997e-01 -3.33000183e-01 3.74249429e-01 9.36544657e-01
-3.07218492e-01 -2.80940622e-01 -1.04196274e+00 8.59725416e-01
-2.26649046e+00 -1.05664098e+00 -4.28608716e-01 2.06178856e+00
8.67912054e-01 2.99762458e-01 4.07612950e-01 7.47805774e-01
2.04638287e-01 2.22119316e-01 -1.49798527e-01 -4.25620109e-01
1.60584301e-01 1.87993869e-01 1.20406970e-01 3.97492886e-01
-1.30511403e+00 9.65126455e-01 6.22674417e+00 6.16590440e-01
-1.08222270e+00 -1.10169917e-01 2.45635718e-01 2.55956072e-02
-2.62404382e-02 4.42757756e-01 -8.02725792e-01 2.68282175e-01
1.17762542e+00 -3.56825411e-01 2.03916505e-01 4.49727923e-01
6.47014797e-01 -6.52903095e-02 -1.57587314e+00 7.26486027e-01
-7.43323043e-02 -1.02368772e+00 -3.84807855e-01 -3.06879848e-01
2.66433567e-01 -5.96311331e-01 -6.65509105e-01 1.43259183e-01
2.13722046e-02 -5.84652781e-01 1.25819206e+00 7.13104844e-01
4.01169658e-01 -6.16256058e-01 3.72343183e-01 4.25596207e-01
-1.55295777e+00 4.15389277e-02 5.92793345e-01 -2.47949690e-01
7.05615103e-01 8.18827808e-01 -7.06090391e-01 1.05081785e+00
3.69580597e-01 9.93591130e-01 -2.06082240e-01 9.65888143e-01
-5.15901625e-01 6.81814551e-01 -4.93076921e-01 1.29888222e-01
8.33208561e-02 -1.89989477e-01 5.95524669e-01 1.37357461e+00
3.59639734e-01 5.50759844e-02 1.29593581e-01 7.08623707e-01
3.06800306e-01 1.05819829e-01 -5.65581143e-01 -2.55945474e-01
3.10864687e-01 1.08123374e+00 -1.07790470e+00 -3.07445049e-01
-4.27951276e-01 7.92061448e-01 3.61564308e-01 2.32192904e-01
-1.07798326e+00 -3.86374056e-01 -1.18627260e-02 4.56094071e-02
1.62680790e-01 -6.06587827e-01 -2.61098444e-01 -1.21792436e+00
5.72451472e-01 -5.34388840e-01 5.65932810e-01 -7.00859249e-01
-8.68246973e-01 6.59831643e-01 2.92575032e-01 -1.50078332e+00
-6.94920599e-01 -2.55170017e-01 -4.31042403e-01 6.39458239e-01
-1.22903848e+00 -1.20058060e+00 1.15939923e-01 4.93018389e-01
4.49589372e-01 5.77454448e-01 6.59940064e-01 4.70171064e-01
-4.90159273e-01 1.50918379e-01 -5.35770416e-01 9.84899700e-02
5.49434006e-01 -1.32911253e+00 3.87531608e-01 1.03653944e+00
2.83877790e-01 5.41536033e-01 7.84192979e-01 -5.83984911e-01
-1.12818074e+00 -9.91258740e-01 2.03480530e+00 -3.41815442e-01
1.13724446e+00 -5.02606094e-01 -7.66897559e-01 9.49770451e-01
-6.48973882e-02 -2.44381204e-01 4.13758159e-01 2.83355027e-01
-5.09089470e-01 -2.05933765e-01 -5.63035607e-01 6.93639636e-01
1.16589224e+00 -7.37581372e-01 -7.43746281e-01 3.74722064e-01
7.57934332e-01 -4.86033380e-01 -8.92106473e-01 3.51358145e-01
5.10220289e-01 -7.41983533e-01 7.86252916e-01 -5.85842967e-01
4.56416368e-01 -5.86305559e-01 2.46168286e-01 -8.06635618e-01
5.04592471e-02 -1.19688880e+00 -5.41476011e-01 1.51797652e+00
4.13682252e-01 -2.93907046e-01 5.42738795e-01 5.04829049e-01
-3.32975894e-01 -7.76709318e-01 -9.02611792e-01 -8.40112984e-01
-4.52852368e-01 -1.01952469e+00 4.03076857e-01 1.09418166e+00
4.46714699e-01 3.66957903e-01 -5.12635052e-01 -8.35960880e-02
2.37796694e-01 3.60206604e-01 5.39871991e-01 -1.18678236e+00
-2.88815469e-01 -4.26213086e-01 -1.55741900e-01 -9.76755500e-01
1.92343712e-01 -6.87715292e-01 1.96744964e-01 -1.38449097e+00
-1.74369365e-01 -3.19758087e-01 -1.30427584e-01 6.59784019e-01
2.70620614e-01 -3.22513163e-01 8.03043693e-02 1.23943858e-01
-6.98087752e-01 2.98346996e-01 8.91255796e-01 8.90252665e-02
-4.34982896e-01 1.06130771e-01 -1.54571265e-01 7.10585594e-01
5.26122212e-01 -4.92003471e-01 -5.15166640e-01 1.17525076e-02
5.70633650e-01 5.93971491e-01 3.86970520e-01 -9.91299033e-01
5.42346716e-01 -4.75490391e-01 -1.12323314e-01 -8.37546289e-01
2.71820605e-01 -9.54343855e-01 3.99035394e-01 1.51215717e-01
-4.45210904e-01 1.70600981e-01 1.93235189e-01 4.42450643e-01
-4.07115430e-01 -1.42152086e-01 3.96063834e-01 1.62389755e-01
-7.71206558e-01 -4.91037481e-02 -4.12522376e-01 -3.40622842e-01
1.22568297e+00 -2.39647165e-01 -8.74755532e-02 -5.52449860e-02
-1.26829648e+00 1.28057584e-01 7.25931604e-04 3.89806032e-01
1.70193121e-01 -1.08886039e+00 -3.77047807e-01 -2.65774161e-01
-1.88097265e-02 3.09505984e-02 8.71844739e-02 1.25453234e+00
-2.96140641e-01 7.15922892e-01 2.35209852e-01 -4.23999071e-01
-1.17974114e+00 8.22266459e-01 1.27223432e-01 -8.86125684e-01
-7.16451526e-01 4.00881797e-01 -9.41088125e-02 -2.31325105e-02
2.21097991e-01 -5.96410513e-01 -4.66220826e-01 5.22375643e-01
4.66233313e-01 2.71038383e-01 2.51000136e-01 -6.87355697e-01
-4.29841459e-01 4.18020964e-01 7.69325271e-02 -5.75714707e-01
1.48372865e+00 -1.47394791e-01 -3.37476760e-01 9.84451830e-01
1.00387490e+00 -2.22581495e-02 -1.05886686e+00 -2.56774098e-01
9.96385217e-01 -8.32420439e-02 -4.25465554e-01 -6.52598262e-01
-3.77645910e-01 4.88137066e-01 -1.20358709e-02 6.09049201e-01
1.51304746e+00 2.07854081e-02 8.14538062e-01 -7.15076644e-03
4.37138528e-01 -1.07106781e+00 -1.35301888e-01 8.42757106e-01
7.79454887e-01 -7.09983289e-01 -1.91236556e-01 -7.06380010e-01
-2.38852262e-01 1.22192311e+00 3.00663322e-01 2.09023714e-01
6.75442934e-01 5.43936491e-01 -2.50076979e-01 -1.15716852e-01
-1.22157466e+00 -4.79828060e-01 4.79037523e-01 2.83862174e-01
7.74293244e-01 -1.65700614e-01 -9.70264196e-01 6.65945590e-01
-8.51159990e-02 4.31432456e-01 2.34877527e-01 1.24619782e+00
1.23023286e-01 -1.45174491e+00 -1.42928734e-01 7.67367613e-03
-4.76599872e-01 1.35936379e-01 -5.60637057e-01 7.31605887e-01
3.22270244e-01 1.11837125e+00 6.05650321e-02 -4.52672154e-01
6.28314912e-01 3.66557658e-01 5.50562561e-01 -4.53216940e-01
-6.95861399e-01 2.01451764e-01 6.08754098e-01 -7.73038566e-01
-6.56627417e-01 -9.38722908e-01 -1.27223599e+00 1.19671889e-01
-1.56741992e-01 3.31753731e-01 4.47216809e-01 1.29406178e+00
4.05976363e-02 5.91380835e-01 6.46656454e-01 -7.88038790e-01
6.53839037e-02 -7.69790113e-01 -3.31135094e-01 5.73540092e-01
1.55186549e-01 -5.15719593e-01 -2.67115563e-01 4.83602136e-01] | [9.122038841247559, 9.25343132019043] |
98973367-e4a3-47c3-863f-17266ca77d8e | grounding-counterfactual-explanation-of-image | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Kim_Grounding_Counterfactual_Explanation_of_Image_Classifiers_to_Textual_Concept_Space_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Kim_Grounding_Counterfactual_Explanation_of_Image_Classifiers_to_Textual_Concept_Space_CVPR_2023_paper.pdf | Grounding Counterfactual Explanation of Image Classifiers to Textual Concept Space | Concept-based explanation aims to provide concise and human-understandable explanations of an image classifier. However, existing concept-based explanation methods typically require a significant amount of manually collected concept-annotated images. This is costly and runs the risk of human biases being involved in the explanation. In this paper, we propose counterfactual explanation with text-driven concepts (CounTEX), where the concepts are defined only from text by leveraging a pre-trained multi-modal joint embedding space without additional concept-annotated datasets. A conceptual counterfactual explanation is generated with text-driven concepts. To utilize the text-driven concepts defined in the joint embedding space to interpret target classifier outcome, we present a novel projection scheme for mapping the two spaces with a simple yet effective implementation. We show that CounTEX generates faithful explanations that provide a semantic understanding of model decision rationale robust to human bias. | ['Tara Taghavi', 'Jaeyoung Do', 'Seunghak Yu', 'Sungjin Lee', 'Jinoh Oh', 'Siwon Kim'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['counterfactual-explanation'] | ['miscellaneous'] | [ 5.83783865e-01 7.42244661e-01 -3.16235930e-01 -7.74059892e-01
-6.00220919e-01 -2.63819665e-01 1.01312685e+00 1.67679518e-01
-2.22404629e-01 6.53784752e-01 5.91319442e-01 -4.82016206e-01
-2.23987728e-01 -6.34404302e-01 -7.42691398e-01 -4.63184774e-01
2.84492761e-01 5.74507594e-01 -3.64466131e-01 1.45313218e-01
4.45520371e-01 5.33168353e-02 -1.62700665e+00 4.74588513e-01
1.01070070e+00 7.27327406e-01 9.00931656e-02 4.64496791e-01
-4.14831489e-01 6.94823623e-01 -3.08638781e-01 -8.23169291e-01
1.11652970e-01 -5.69009066e-01 -7.60234594e-01 4.54025954e-01
4.20858651e-01 -3.16418499e-01 7.40015646e-03 9.66679692e-01
-6.43734112e-02 9.57895666e-02 9.86937284e-01 -1.88740778e+00
-1.14382100e+00 7.26939797e-01 -3.70922178e-01 -2.65962929e-01
3.07125479e-01 6.22561388e-02 1.27571940e+00 -1.26477897e+00
5.36384583e-01 1.57496202e+00 3.63680005e-01 1.17935562e+00
-1.47271395e+00 -8.03372502e-01 5.41683853e-01 4.04074311e-01
-8.48392546e-01 1.19396567e-01 9.65726376e-01 -4.76871401e-01
5.91302454e-01 3.67677152e-01 8.75127494e-01 1.31187451e+00
2.50890106e-01 7.77584076e-01 1.09525645e+00 -4.92492259e-01
5.74524820e-01 5.67100346e-01 2.18243450e-01 9.07248199e-01
6.72491133e-01 3.30066651e-01 -7.48961449e-01 -1.93448618e-01
4.73725557e-01 4.55009788e-01 -4.60478216e-01 -1.02349365e+00
-1.34855258e+00 1.51765108e+00 6.65314317e-01 -3.53911147e-02
-5.10779500e-01 4.50804532e-01 2.24050105e-01 -2.27315962e-01
4.76264864e-01 7.95915902e-01 -3.74545127e-01 3.77474248e-01
-7.00922072e-01 2.00985372e-01 4.95120913e-01 1.00533807e+00
7.60208189e-01 1.61412489e-02 -4.14322913e-02 3.99959266e-01
4.82296258e-01 6.15006745e-01 6.95015967e-01 -9.38483536e-01
3.01052630e-01 7.31028914e-01 -4.69358973e-02 -1.02912068e+00
-3.08862422e-02 -2.32453600e-01 -6.20184422e-01 3.23966920e-01
1.11189652e-02 1.41696259e-01 -9.83504653e-01 1.83763456e+00
2.26640284e-01 9.55479443e-02 5.01571417e-01 1.11884201e+00
4.26645756e-01 4.49627191e-01 4.59586680e-01 2.95103546e-02
1.40171003e+00 -1.09951949e+00 -8.20916593e-01 -4.84011173e-01
5.42777777e-01 -2.62694389e-01 1.37992966e+00 -4.31110188e-02
-5.43673396e-01 -2.39074811e-01 -1.31818819e+00 -1.08076371e-01
-4.87343520e-01 1.70297489e-01 1.04791057e+00 5.04544556e-01
-6.46705449e-01 3.70767266e-01 -5.80857575e-01 -4.31143999e-01
8.06649327e-01 2.12375596e-01 -6.19143844e-01 -3.96719486e-01
-8.08123946e-01 1.05716252e+00 4.79160726e-01 -2.19058484e-01
-1.06550002e+00 -7.61400878e-01 -1.26281059e+00 3.57350379e-01
6.02895200e-01 -1.10841000e+00 1.20599127e+00 -1.06292367e+00
-1.15280318e+00 7.38703728e-01 -2.28726298e-01 -6.85388267e-01
4.16289181e-01 -4.29102749e-01 -2.09734514e-01 3.14876586e-01
4.82595831e-01 1.27069414e+00 1.08390713e+00 -1.69816923e+00
-3.39713246e-01 -4.63870108e-01 1.03095837e-01 2.92774409e-01
-3.01677734e-01 -5.30615211e-01 7.34440237e-02 -7.00603068e-01
3.40007752e-01 -9.66188788e-01 -5.03252745e-01 5.27449727e-01
-7.15527177e-01 -1.84976719e-02 9.04199362e-01 -2.21816421e-01
6.67989731e-01 -1.85362041e+00 -1.50770783e-01 -5.52584007e-02
5.24273157e-01 -2.76133180e-01 -1.32850111e-01 3.67954224e-02
-4.00743484e-01 3.79591405e-01 -7.37082899e-01 -2.67856777e-01
1.92181990e-01 5.01054406e-01 -8.30322266e-01 1.87191322e-01
5.28618693e-01 8.65525126e-01 -1.15647364e+00 -5.66680253e-01
5.56868017e-01 4.11355764e-01 -8.43519449e-01 2.91000903e-01
-3.35407406e-01 -7.26160184e-02 -5.90005338e-01 3.59092444e-01
5.43114483e-01 -3.31013054e-01 2.25118086e-01 -2.36498579e-01
3.20679247e-01 1.66044801e-01 -7.26450622e-01 1.69451344e+00
-5.52195907e-01 6.92875922e-01 -7.49899864e-01 -1.00230515e+00
6.98579431e-01 3.99759591e-01 4.57196496e-02 -1.25156596e-01
-2.93199235e-04 1.29830867e-01 -2.20668137e-01 -4.29074168e-01
3.04009199e-01 -8.78778756e-01 1.45708136e-02 7.95670927e-01
1.63117245e-01 -5.13880849e-01 -3.61927062e-01 5.72386265e-01
6.81387544e-01 9.23344046e-02 7.03206360e-01 -2.16825172e-01
3.27757359e-01 3.37384939e-01 4.98149991e-01 7.29133785e-01
-1.02907933e-01 7.67861664e-01 3.87484223e-01 -6.61276400e-01
-8.70899141e-01 -1.15918958e+00 1.33741438e-01 5.08982778e-01
1.44116417e-01 -4.18915719e-01 -7.06887364e-01 -1.29120171e+00
2.74886880e-02 1.51164174e+00 -1.01989067e+00 -4.39827055e-01
9.69366208e-02 -4.81968433e-01 4.50108992e-03 7.60740817e-01
4.41290110e-01 -8.38702917e-01 -9.80522573e-01 -5.31287789e-02
-1.32104456e-01 -8.70565295e-01 -5.54661393e-01 8.55745375e-02
-1.01325107e+00 -1.35088825e+00 -4.84961361e-01 -2.18755886e-01
1.19056666e+00 4.99430984e-01 9.72849727e-01 1.37028605e-01
-2.77983606e-01 6.50359571e-01 -3.63347940e-02 -8.72360587e-01
-1.78703144e-01 -6.73832893e-01 8.55255723e-02 8.18458050e-02
6.36431992e-01 -2.20992431e-01 -5.63515961e-01 -8.17133039e-02
-9.00382817e-01 6.39100432e-01 5.52281976e-01 1.24939823e+00
4.20349658e-01 -2.70268619e-01 4.69752371e-01 -1.12337065e+00
5.83067536e-01 -4.30184543e-01 -3.25689286e-01 2.81308085e-01
-1.11583257e+00 5.62650383e-01 5.90319574e-01 -5.53139985e-01
-1.36305749e+00 2.69043177e-01 3.91086906e-01 -5.45962334e-01
-1.18174084e-01 2.59127557e-01 -2.24196672e-01 4.24902439e-01
7.45929301e-01 1.16727658e-01 -1.99684706e-02 -4.37431075e-02
1.05507183e+00 2.99340874e-01 4.92841840e-01 -4.64247823e-01
8.48797441e-01 9.88607645e-01 7.99864065e-03 -2.75170654e-01
-1.32775831e+00 -2.77292490e-01 -5.00671029e-01 2.97056772e-02
1.08629537e+00 -7.04770923e-01 -3.62344414e-01 -5.47412515e-01
-1.41664875e+00 2.43036777e-01 -4.82671440e-01 8.98461580e-01
-8.35128605e-01 4.50066142e-02 4.98606525e-02 -8.18863451e-01
-1.37977213e-01 -1.00340664e+00 1.03133690e+00 -9.95517615e-03
-6.17614269e-01 -1.08157837e+00 -1.15070537e-01 4.13452119e-01
2.40392238e-02 2.20029324e-01 1.20609653e+00 -8.13320756e-01
-6.77472591e-01 -2.10209489e-01 -4.21814978e-01 1.38074443e-01
7.73835480e-02 -3.52353573e-01 -1.22042143e+00 1.61137372e-01
1.54402584e-01 -3.33618999e-01 9.73049581e-01 4.13813859e-01
1.37421227e+00 -6.44693255e-01 -5.60981512e-01 1.44847497e-01
1.54620469e+00 6.18712381e-02 1.77650705e-01 7.48629123e-02
6.66117966e-01 9.47754860e-01 7.61143684e-01 2.13156134e-01
3.42226744e-01 2.49165922e-01 5.40125251e-01 -2.68457532e-01
1.65759400e-01 -7.58324027e-01 9.92842689e-02 6.45394102e-02
2.19212383e-01 -1.52953759e-01 -7.10107446e-01 7.73721397e-01
-1.93539560e+00 -1.09701693e+00 4.14774893e-03 1.91216266e+00
6.08047307e-01 -9.07467976e-02 -5.29645026e-01 1.58389479e-01
6.38106227e-01 -1.30838245e-01 -7.66297042e-01 -3.89596820e-01
6.91073760e-02 -6.80426359e-02 2.47992158e-01 6.53606176e-01
-6.99093163e-01 9.15292382e-01 6.53858900e+00 3.30068588e-01
-7.81044960e-01 2.48204291e-01 4.31278497e-01 -2.10331380e-01
-1.10332823e+00 5.01655161e-01 -1.65749416e-01 5.58335241e-03
6.72154725e-01 -4.69608277e-01 -1.66527092e-01 1.30822051e+00
2.17280071e-02 -4.05753441e-02 -1.50508201e+00 1.01734495e+00
4.18739110e-01 -1.70401239e+00 8.01625848e-01 1.01649486e-01
6.43921912e-01 -6.78970098e-01 1.31192014e-01 2.17417002e-01
4.93344039e-01 -1.10715485e+00 8.34561348e-01 3.38826209e-01
7.40778089e-01 -5.56276143e-01 7.19738841e-01 1.25699669e-01
-6.66792393e-01 -3.44800144e-01 -5.14082611e-01 -4.78312820e-02
5.83634786e-02 6.23410702e-01 -1.10140121e+00 2.89009720e-01
3.15941632e-01 6.55510962e-01 -3.94101292e-01 3.16327602e-01
-8.06750655e-01 4.41470504e-01 2.04585224e-01 -4.64626327e-02
2.38056168e-01 5.10941222e-02 4.11679626e-01 9.81255651e-01
4.13766176e-01 3.44714314e-01 -5.62412962e-02 1.48008764e+00
1.64798908e-02 -4.48617786e-02 -9.53478515e-01 -1.11317486e-01
2.62026608e-01 1.11000073e+00 -6.84528530e-01 -7.58809805e-01
-3.57254863e-01 1.28774548e+00 2.01047361e-01 4.19449955e-01
-8.44397128e-01 4.37585898e-02 7.09020138e-01 -1.36521652e-01
-4.81483042e-02 2.43885592e-01 -7.19302356e-01 -1.39397871e+00
-1.86576575e-01 -6.11463487e-01 4.59137887e-01 -1.20562530e+00
-1.23855793e+00 4.56141502e-01 2.21203789e-01 -1.10058618e+00
-3.44133437e-01 -7.82098234e-01 -6.95257723e-01 7.22674072e-01
-1.40846002e+00 -1.24513125e+00 -2.98815012e-01 3.78417045e-01
8.30254138e-01 -1.61181942e-01 1.13976252e+00 -4.61668998e-01
-2.07594603e-01 1.99119508e-01 -5.05859792e-01 -3.12644958e-01
4.91390526e-01 -1.63340724e+00 6.81232587e-02 7.88117051e-01
3.80254120e-01 1.09802032e+00 1.11224365e+00 -5.75234532e-01
-8.68958950e-01 -1.06174421e+00 1.09059715e+00 -5.80362678e-01
3.80138725e-01 -3.76367450e-01 -7.55101442e-01 8.66075754e-01
2.19646767e-01 -1.69054493e-01 1.19884419e+00 2.36710101e-01
-8.01064909e-01 1.54955879e-01 -1.27568018e+00 1.01067424e+00
8.81182790e-01 -7.08675921e-01 -1.24402928e+00 3.53396297e-01
9.41730857e-01 2.24450648e-01 -1.23920612e-01 -4.40357625e-02
5.51139235e-01 -8.84208143e-01 9.21016753e-01 -9.87988114e-01
9.68130350e-01 -4.56771761e-01 -2.78864443e-01 -1.54305196e+00
-1.84473172e-01 -5.68266064e-02 9.41393450e-02 7.25704849e-01
6.19600832e-01 -5.47776997e-01 7.84123123e-01 1.03283238e+00
2.14003585e-02 -7.63467848e-01 -6.65308416e-01 -5.63443720e-01
-1.10521920e-01 -7.12764740e-01 8.89273047e-01 1.00982261e+00
3.42618644e-01 5.44430554e-01 -2.40253359e-01 2.04260886e-01
9.18280005e-01 3.76715183e-01 6.01637363e-01 -1.06939447e+00
1.22647308e-01 -2.47931480e-02 -4.38282818e-01 -6.35576725e-01
4.55170453e-01 -8.96709442e-01 1.81520849e-01 -1.66457152e+00
7.58503377e-01 -4.27657105e-02 -5.57320453e-02 6.20652258e-01
-4.28944916e-01 2.86453534e-02 2.83214897e-01 1.04713060e-01
-2.68411011e-01 9.67465639e-01 1.11504686e+00 -2.15404853e-01
3.46230537e-01 -5.07295966e-01 -1.14766324e+00 1.00090599e+00
6.55192077e-01 -8.21427763e-01 -9.53795552e-01 -2.43431836e-01
-1.02817915e-01 -1.49158478e-01 8.38339925e-01 -6.42097116e-01
-3.00994553e-02 -4.07759696e-01 5.10627329e-01 -4.36130643e-01
5.14262438e-01 -1.01924944e+00 -9.82793048e-02 6.16615713e-01
-8.50376308e-01 -1.36891335e-01 5.01553342e-02 1.09120309e+00
-2.45814472e-01 -2.46929318e-01 5.61088562e-01 -1.79995984e-01
-5.57262003e-01 2.02592462e-01 -3.51581603e-01 -4.15055096e-01
1.10372663e+00 -4.19043213e-01 -1.80260003e-01 -5.86071789e-01
-5.58285415e-01 2.36796871e-01 3.54748249e-01 3.90000463e-01
1.16334510e+00 -1.56989074e+00 -4.56381857e-01 2.93441266e-02
5.98730028e-01 -3.83496463e-01 2.76463032e-01 2.65576631e-01
-3.75925265e-02 8.31413925e-01 -3.02679330e-01 -4.31095958e-01
-1.09838855e+00 8.57804537e-01 1.91828042e-01 -4.90765460e-02
-7.88628459e-01 7.82020509e-01 8.50866675e-01 -6.89053416e-01
-1.76985040e-01 -3.70776355e-01 -1.59889609e-01 -2.71809906e-01
5.47584951e-01 -6.28799573e-02 -6.66041076e-01 -2.92470843e-01
-3.46634537e-01 3.33626062e-01 3.42683978e-02 -6.82407022e-01
1.10917699e+00 -1.47167668e-01 3.21004689e-01 4.35978323e-01
1.06455791e+00 -3.03012937e-01 -1.39257789e+00 -4.37437706e-02
2.62064457e-01 -8.02440226e-01 2.30071601e-02 -8.58046710e-01
-6.95153713e-01 1.22235060e+00 4.21946883e-01 -3.81595284e-01
9.64942753e-01 1.05730966e-01 2.92907000e-01 4.55082566e-01
8.34257081e-02 -7.57017493e-01 3.63046169e-01 -1.97855949e-01
1.28091919e+00 -1.59258664e+00 1.08448952e-01 -6.73703432e-01
-1.11062694e+00 1.10380745e+00 7.16982007e-01 1.22081764e-01
3.67613316e-01 -3.72742534e-01 2.03501150e-01 -6.46421731e-01
-1.03338933e+00 -3.59162763e-02 4.50057387e-01 9.07478988e-01
3.75584751e-01 3.62809509e-01 -2.67691135e-01 8.70520532e-01
-3.03458691e-01 -1.61393031e-01 7.31383204e-01 5.86511254e-01
-2.92378545e-01 -7.62563586e-01 -3.62074167e-01 4.90954518e-01
-3.93328667e-02 -3.39398116e-01 -5.39417326e-01 7.43418217e-01
1.31680649e-02 9.42858577e-01 6.07035309e-02 -8.87586847e-02
4.48896661e-02 3.35821271e-01 3.21238190e-01 -9.50168848e-01
7.75986463e-02 -3.91208947e-01 -1.07764192e-01 -7.47215629e-01
-4.52096254e-01 -5.32889724e-01 -1.63909197e+00 1.08308598e-01
-4.83642071e-01 3.23289871e-01 8.16284418e-01 1.16823971e+00
3.02441031e-01 3.50653142e-01 2.50543773e-01 -4.35074270e-01
-6.79225922e-01 -7.99287438e-01 -3.61681789e-01 6.22885704e-01
3.45373183e-01 -8.85297656e-01 -5.77338994e-01 3.35687786e-01] | [8.949581146240234, 5.5840373039245605] |
bedfbd7b-0423-4a86-ab23-4df947d8c8df | offlangone-dravidianlangtech-eacl2021 | null | null | https://aclanthology.org/2021.dravidianlangtech-1.19 | https://aclanthology.org/2021.dravidianlangtech-1.19.pdf | OFFLangOne@DravidianLangTech-EACL2021: Transformers with the Class Balanced Loss for Offensive Language Identification in Dravidian Code-Mixed text. | The intensity of online abuse has increased in recent years. Automated tools are being developed to prevent the use of hate speech and offensive content. Most of the technologies use natural language and machine learning tools to identify offensive text. In a multilingual society, where code-mixing is a norm, the hate content would be delivered in a code-mixed form in social media, which makes the offensive content identification, further challenging. In this work, we participated in the EACL task to detect offensive content in the code-mixed social media scenario. The methodology uses a transformer model with transliteration and class balancing loss for offensive content identification. In this task, our model has been ranked 2nd in Malayalam-English and 4th in Tamil-English code-mixed languages. | ['Radhika Mamidi', 'Suman Dowlagar'] | null | null | null | null | eacl-dravidianlangtech-2021-4 | ['transliteration'] | ['natural-language-processing'] | [-8.88826400e-02 -4.01568320e-03 -1.31760389e-01 1.39794111e-01
-4.54972297e-01 -9.40941632e-01 6.85376346e-01 4.35587764e-01
-2.47239798e-01 5.31944811e-01 2.55549252e-01 -4.21458483e-01
6.13540486e-02 -3.17960799e-01 -2.18475163e-01 -3.78309101e-01
-2.36984696e-02 6.42985478e-02 -1.48015589e-01 -5.11699021e-01
7.42109060e-01 2.59244919e-01 -1.13977444e+00 5.27479053e-01
1.41240418e+00 3.17717969e-01 6.21050633e-02 7.03221381e-01
-4.12432581e-01 1.53159189e+00 -6.75329804e-01 -1.19338143e+00
2.24847406e-01 -4.12925154e-01 -9.97842491e-01 -1.51469335e-01
5.03832459e-01 -2.87978947e-01 -2.72239298e-01 1.72247350e+00
2.73471117e-01 -4.72896367e-01 7.26438046e-01 -1.10585392e+00
-6.09725356e-01 9.08893943e-01 -9.04937208e-01 1.47754982e-01
6.25998020e-01 -2.49663502e-01 6.33104384e-01 -8.43445837e-01
7.53419340e-01 1.18524480e+00 4.97433096e-01 2.86542475e-01
-8.50470066e-01 -9.83060777e-01 -4.22146142e-01 1.22325737e-02
-1.47992849e+00 -6.82286620e-02 8.66581559e-01 -1.10641181e+00
6.39111996e-01 1.82657689e-01 6.54635072e-01 1.19123399e+00
3.71876061e-01 6.36388123e-01 1.22296274e+00 -6.95930839e-01
-2.76959777e-01 6.96929574e-01 -3.26641388e-02 8.68081033e-01
2.09329501e-01 -4.28878784e-01 -6.11836910e-01 -5.16965151e-01
-1.84692517e-01 9.90829021e-02 5.20352051e-02 1.81331426e-01
-4.62887943e-01 1.44462371e+00 1.52480872e-02 8.85481477e-01
4.04160172e-02 -4.37321246e-01 8.60688090e-01 4.69238698e-01
8.65982890e-01 7.60359049e-01 8.07059780e-02 -6.33280933e-01
-1.21021986e+00 2.61381567e-01 8.99931133e-01 8.18468153e-01
4.70236450e-01 -9.78779197e-02 3.42123091e-01 1.42421865e+00
4.34650868e-01 5.76574802e-01 6.55576229e-01 -4.81881171e-01
8.52517605e-01 7.33944356e-01 -3.01130027e-01 -1.51778138e+00
-1.21777527e-01 -4.04047549e-01 -5.77744424e-01 9.40050110e-02
2.77027130e-01 -3.82442445e-01 -6.97383165e-01 1.23703432e+00
2.52935290e-02 -5.89453578e-01 -3.48047882e-01 4.69541281e-01
1.21183179e-01 6.85117662e-01 2.28937253e-01 -5.77757917e-02
1.17595243e+00 -7.32312739e-01 -9.01914120e-01 5.38441055e-02
1.07043850e+00 -1.35474181e+00 8.84481370e-01 6.80610120e-01
-6.98544383e-01 5.68871684e-02 -1.09971368e+00 9.85553302e-03
-8.71928096e-01 -4.26768184e-01 1.58866316e-01 1.42732799e+00
-5.59375346e-01 5.72864413e-01 -1.67444602e-01 -4.70082790e-01
4.57445741e-01 -1.29445307e-02 -5.95501423e-01 -1.36981730e-03
-1.43691814e+00 1.25766289e+00 4.60253716e-01 -3.28632712e-01
-5.55338621e-01 -4.58120733e-01 -8.02814841e-01 -4.99547541e-01
-4.24777791e-02 5.14279842e-01 1.01766562e+00 -1.34795046e+00
-1.03059399e+00 1.33449435e+00 5.43863893e-01 -1.39807239e-01
6.91321671e-01 -5.74771941e-01 -9.14850473e-01 -3.05856466e-02
2.47094214e-01 -5.85069992e-02 1.07835615e+00 -9.85577226e-01
-5.36693692e-01 -4.27719712e-01 7.14868233e-02 -1.05677105e-01
-8.58230889e-01 8.97226810e-01 2.71994263e-01 -8.81899357e-01
-3.59106034e-01 -1.08320928e+00 3.42766672e-01 -4.56803024e-01
-6.39929652e-01 7.30602294e-02 1.21442652e+00 -1.30357528e+00
1.89104593e+00 -2.32137728e+00 5.45935072e-02 4.82460529e-01
4.21425134e-01 5.24697900e-01 4.51176047e-01 8.86975765e-01
9.01478976e-02 5.78942060e-01 -3.36979181e-01 -1.26475692e-01
-1.44083038e-01 -5.95219247e-02 -4.49671626e-01 7.77444243e-01
-6.00630194e-02 4.33995910e-02 -9.73784447e-01 -9.11114752e-01
-1.15469500e-01 2.43625298e-01 -5.99082410e-01 2.08963245e-01
3.91400695e-01 1.33189401e-02 -3.17805380e-01 7.86079705e-01
7.51870930e-01 4.31218654e-01 2.17669178e-02 4.86524880e-01
-5.78140974e-01 4.29794714e-02 -4.94854361e-01 1.26513445e+00
-6.10972583e-01 1.08246624e+00 2.57636815e-01 -1.33976787e-01
8.45146894e-01 1.77069202e-01 2.85205573e-01 -4.70867634e-01
5.39288878e-01 6.40380561e-01 4.64312702e-01 -9.11383331e-01
7.49790013e-01 -3.27752903e-02 -2.95288950e-01 2.30216518e-01
1.49334297e-01 -2.20460445e-01 4.17991310e-01 4.28129822e-01
9.35916960e-01 -9.97768641e-02 4.06788319e-01 -5.26390851e-01
6.63274407e-01 -1.62211820e-01 3.91992442e-02 4.66168642e-01
-4.56357449e-01 4.19520974e-01 8.83349955e-01 -1.76735416e-01
-1.35175824e+00 -6.40878677e-01 -2.61112154e-01 1.34385765e+00
-3.61139596e-01 -6.50354207e-01 -1.02563512e+00 -8.99716675e-01
-6.64945692e-02 7.86392748e-01 -3.64462316e-01 -1.61922857e-01
-3.96151036e-01 -3.99237990e-01 1.05571866e+00 -3.93061042e-01
3.87935668e-01 -8.78718734e-01 -4.15263593e-01 1.94563851e-01
-1.65626034e-01 -8.71119857e-01 -5.16901731e-01 4.35772128e-02
-2.88935363e-01 -1.17593896e+00 -8.35615098e-01 -5.87256789e-01
3.92908931e-01 -1.42666459e-01 4.61765587e-01 2.24471331e-01
-4.83236641e-01 -3.16618234e-01 -8.97909105e-01 -7.23720729e-01
-9.08733189e-01 4.26906317e-01 -9.60488766e-02 2.12397903e-01
5.09267569e-01 -3.56872708e-01 1.65211572e-03 -1.94613144e-01
-1.03416741e+00 -7.64399990e-02 2.46055812e-01 4.26880807e-01
-7.94455051e-01 -2.80180782e-01 1.24881931e-01 -1.47145593e+00
8.26061428e-01 -9.85965967e-01 -2.47192711e-01 -8.21662694e-02
-4.19455469e-01 -2.22921476e-01 1.03840160e+00 -5.88029623e-01
-8.82431567e-01 -1.67935669e-01 -4.15850610e-01 -1.89278170e-01
4.62625399e-02 4.80170339e-01 1.29811257e-01 -2.40061209e-01
8.95255446e-01 1.77581739e-02 -1.05617002e-01 -5.02560258e-01
-3.77717093e-02 1.33958864e+00 -1.03666887e-01 -3.69641066e-01
1.16700780e+00 -1.98328458e-02 -5.75779796e-01 -1.27712727e+00
-8.07931602e-01 -6.85203016e-01 -5.95117807e-01 -7.03679860e-01
1.07567847e+00 -6.44118249e-01 -2.03838006e-01 1.06716073e+00
-1.38614678e+00 1.02939710e-01 7.48022020e-01 3.89210641e-01
-9.95609984e-02 6.38769686e-01 -8.08878481e-01 -1.09141898e+00
-3.41849536e-01 -9.07698214e-01 4.98799086e-01 -7.48127922e-02
-7.27402091e-01 -7.17074275e-01 4.83986676e-01 5.01413703e-01
4.99027908e-01 5.06226897e-01 1.17323339e+00 -1.07713139e+00
4.24807608e-01 -2.08509460e-01 -1.28159314e-01 5.13965845e-01
1.18927248e-01 6.21545076e-01 -8.55989695e-01 -1.62421122e-01
2.93827318e-02 -4.85866517e-01 1.07999608e-01 -4.73969162e-01
7.12202370e-01 -6.72862589e-01 6.45053834e-02 2.07396239e-01
1.29680181e+00 4.18884009e-01 6.24277592e-01 4.21863675e-01
8.52131486e-01 8.42241764e-01 4.69657391e-01 7.13368356e-01
-7.88933113e-02 3.97701055e-01 3.79982084e-01 2.26116315e-01
4.42370474e-01 -3.98189902e-01 5.68232238e-01 1.31868827e+00
-2.28195195e-03 -2.30900738e-02 -1.43343008e+00 5.01035035e-01
-1.39276969e+00 -1.19806433e+00 -4.40233320e-01 1.83478189e+00
1.00311923e+00 1.46686316e-01 3.54981214e-01 3.15264016e-01
9.67193067e-01 2.58511901e-01 2.44932070e-01 -1.14846694e+00
1.39593154e-01 -1.08514056e-01 5.50457418e-01 6.03994012e-01
-1.28134978e+00 8.10882211e-01 5.43497849e+00 9.69823122e-01
-1.07762599e+00 5.24076700e-01 3.66093278e-01 1.86366364e-02
3.42146456e-02 -1.94750279e-01 -4.37517166e-01 1.09619033e+00
9.89657342e-01 -7.75433183e-02 4.20029163e-01 1.00357080e+00
-3.15968283e-02 -7.38702863e-02 -5.06217718e-01 1.18798757e+00
5.90833545e-01 -6.32225811e-01 -2.49011457e-01 2.39255726e-01
7.09622204e-01 -5.36981970e-02 1.27199039e-01 4.10868198e-01
3.05353433e-01 -9.72780883e-01 1.02005208e+00 3.43690217e-02
4.12802398e-01 -1.12791622e+00 8.51286292e-01 5.83392441e-01
-5.48339188e-01 -3.79049838e-01 -1.56660944e-01 -5.19230179e-02
-1.02345057e-01 5.33301771e-01 -5.01241028e-01 -4.58722599e-02
7.22103059e-01 4.61082369e-01 -8.73300791e-01 8.78161967e-01
-1.60563231e-01 6.28483176e-01 5.12366556e-02 -2.83373743e-01
4.53508139e-01 -3.50872695e-01 5.75686336e-01 1.76650465e+00
2.87549436e-01 -3.51134896e-01 -1.54993057e-01 3.83379340e-01
-1.14494830e-01 6.32893443e-01 -1.22240949e+00 -6.61803901e-01
2.73988426e-01 1.24365711e+00 -4.07717437e-01 6.18940741e-02
-4.85038161e-01 9.37069774e-01 5.16463578e-01 -2.77140558e-01
-9.18627262e-01 -9.83983636e-01 4.66689140e-01 3.26516956e-01
-4.70719814e-01 -2.28303626e-01 1.50197625e-01 -9.93336558e-01
-2.08256572e-01 -1.37157404e+00 2.44049624e-01 -4.55649734e-01
-1.35164976e+00 7.44195461e-01 -1.00175254e-02 -1.23422348e+00
-3.45363319e-01 -6.88644767e-01 -2.53633678e-01 5.40397823e-01
-8.50124180e-01 -9.98905599e-01 2.45308131e-01 5.52562118e-01
5.04835665e-01 -5.45617700e-01 5.18902123e-01 8.22688818e-01
-4.77887750e-01 5.98983705e-01 1.00583464e-01 7.58843422e-01
7.20574319e-01 -9.69673753e-01 -1.90006331e-01 9.69154894e-01
-7.32720718e-02 7.56047547e-01 1.00439918e+00 -9.86518264e-01
-8.48522127e-01 -7.38971293e-01 1.20810866e+00 -4.43212152e-01
1.46232355e+00 -8.13628852e-01 -7.27043390e-01 2.93034583e-01
6.21724069e-01 -5.87124169e-01 1.11423516e+00 -2.95045022e-02
-9.09504473e-01 3.43483984e-01 -1.30598426e+00 5.27612686e-01
7.45555460e-01 -1.09063447e+00 -5.83876014e-01 4.99645889e-01
1.98777825e-01 -1.32611021e-01 -7.44604588e-01 -4.34861213e-01
5.90449870e-01 -9.08983648e-01 8.75482410e-02 -6.56577289e-01
1.04819036e+00 6.62710145e-02 -5.01333289e-02 -1.28401041e+00
-1.13424130e-01 -9.09822047e-01 -1.57526915e-03 1.40450537e+00
4.85509276e-01 -3.74644518e-01 3.12808961e-01 4.50713277e-01
1.16904140e-01 -1.00557312e-01 -8.67308855e-01 -4.90026385e-01
4.13828909e-01 -3.39601841e-03 -1.21015862e-01 1.56014585e+00
5.13545394e-01 1.93108901e-01 -7.74516642e-01 -2.99194396e-01
4.81697947e-01 -5.43265164e-01 4.45539027e-01 -1.29667234e+00
1.44495815e-01 -5.08126378e-01 -5.64506233e-01 1.50134405e-02
2.89020061e-01 -9.08266842e-01 -2.67153621e-01 -7.53091872e-01
4.76654321e-01 2.42160354e-02 3.51841956e-01 1.96765319e-01
2.58518845e-01 3.90139490e-01 7.46909738e-01 2.73441166e-01
-3.15392941e-01 1.19160898e-01 6.79611742e-01 -2.49083370e-01
8.75202119e-02 -2.64148355e-01 -4.47348028e-01 9.80884910e-01
8.26621294e-01 -1.04356575e+00 -1.85015023e-01 1.22850440e-01
7.65405715e-01 -3.79161179e-01 -2.13234365e-01 -9.73173857e-01
-1.38411582e-01 -2.07879484e-01 -4.56978492e-02 -2.46427819e-01
9.74573940e-03 -1.11302388e+00 -9.43430811e-02 7.70293236e-01
-3.35201830e-01 2.61315256e-01 -1.50261566e-01 -1.00598782e-02
-2.73025870e-01 -9.02745724e-01 1.03891408e+00 -6.13147169e-02
-1.06465034e-01 -2.27356758e-02 -1.12096739e+00 3.06277454e-01
1.25241649e+00 -1.88911930e-01 -5.14927924e-01 -3.83950263e-01
-2.83574194e-01 -2.31010899e-01 5.52717447e-01 6.47952020e-01
2.93786794e-01 -9.95199680e-01 -9.50389445e-01 -1.42409489e-01
3.74751657e-01 -1.10103905e+00 5.55123538e-02 7.29011416e-01
-1.13646758e+00 -3.90010397e-03 -5.83097637e-01 1.40587196e-01
-1.51184058e+00 4.59313601e-01 1.24147542e-01 -2.54468501e-01
4.15014736e-02 4.13798094e-01 -4.21874821e-01 -4.68184561e-01
-8.71041417e-02 5.72864830e-01 -5.89015782e-01 5.09119511e-01
6.54712856e-01 6.52520537e-01 -2.92405616e-02 -1.20540261e+00
-3.18442911e-01 3.12268138e-01 -4.55270290e-01 1.14362333e-02
1.08739090e+00 1.53700471e-01 -7.11727738e-01 5.30154824e-01
1.82485235e+00 9.23635006e-01 -2.33232811e-01 3.05677831e-01
4.74119067e-01 -8.51519763e-01 -2.08131969e-01 -5.48222125e-01
-7.07549274e-01 6.83966637e-01 3.58157575e-01 8.18604052e-01
6.13805175e-01 -2.42470697e-01 6.10506117e-01 1.20044976e-01
4.21169400e-01 -1.58584940e+00 7.14207739e-02 9.44467068e-01
9.70815182e-01 -9.88953829e-01 -2.57424940e-03 -4.09304380e-01
-7.13637888e-01 1.11764300e+00 5.60805142e-01 -1.29751870e-02
7.98081636e-01 4.96511281e-01 2.72293746e-01 -1.98936164e-01
-4.40529548e-02 1.30685166e-01 9.05854180e-02 5.12945354e-01
9.86340702e-01 -1.53913917e-02 -7.76248455e-01 1.58895567e-01
-6.39130592e-01 -5.57869673e-01 8.56935978e-01 1.03063691e+00
-3.65348816e-01 -1.04183722e+00 -5.00431776e-01 5.55605531e-01
-1.27745700e+00 -2.74046451e-01 -1.11301231e+00 8.43021810e-01
4.26844239e-01 1.20320559e+00 -2.59464920e-01 -6.74330592e-01
-4.45065424e-02 2.54351616e-01 1.00064203e-01 -5.54829895e-01
-1.34588420e+00 -2.74386138e-01 2.86603153e-01 -9.41283628e-02
-1.86511666e-01 -4.77301568e-01 -8.58958662e-01 -8.82577240e-01
-3.67512941e-01 1.73935965e-01 8.89935493e-01 8.03251505e-01
6.48387447e-02 -1.29235953e-01 7.02861547e-01 -3.96812946e-01
-3.38909090e-01 -1.16864479e+00 -7.14026451e-01 7.79077768e-01
3.13889027e-01 -6.11435413e-01 -5.06779909e-01 -1.53064013e-01] | [8.855508804321289, 10.612669944763184] |
4e34c46e-c0b7-4ee2-8c1e-6d4f12a3f7c0 | pick-processing-key-information-extraction | 2004.07464 | null | https://arxiv.org/abs/2004.07464v3 | https://arxiv.org/pdf/2004.07464v3.pdf | PICK: Processing Key Information Extraction from Documents using Improved Graph Learning-Convolutional Networks | Computer vision with state-of-the-art deep learning models has achieved huge success in the field of Optical Character Recognition (OCR) including text detection and recognition tasks recently. However, Key Information Extraction (KIE) from documents as the downstream task of OCR, having a large number of use scenarios in real-world, remains a challenge because documents not only have textual features extracting from OCR systems but also have semantic visual features that are not fully exploited and play a critical role in KIE. Too little work has been devoted to efficiently make full use of both textual and visual features of the documents. In this paper, we introduce PICK, a framework that is effective and robust in handling complex documents layout for KIE by combining graph learning with graph convolution operation, yielding a richer semantic representation containing the textual and visual features and global layout without ambiguity. Extensive experiments on real-world datasets have been conducted to show that our method outperforms baselines methods by significant margins. Our code is available at https://github.com/wenwenyu/PICK-pytorch. | ['Wenwen Yu', 'Xianbiao Qi', 'Rong Xiao', 'Ping Gong', 'Ning Lu'] | 2020-04-16 | null | null | null | null | ['key-information-extraction'] | ['natural-language-processing'] | [ 1.06801257e-01 -6.02624357e-01 -1.48843020e-01 -2.11758256e-01
-4.65629697e-01 -7.93461025e-01 4.99818712e-01 2.77044773e-01
-3.22042257e-01 2.99643725e-01 5.38257435e-02 -3.29322785e-01
3.91135663e-02 -6.68293297e-01 -6.10266268e-01 -6.05880201e-01
1.56483546e-01 2.40943253e-01 2.51068711e-01 -1.77089497e-01
5.63800395e-01 7.31265604e-01 -1.15351844e+00 4.37039644e-01
7.54921973e-01 8.74887526e-01 6.02447629e-01 1.01763201e+00
-4.97465312e-01 8.36919367e-01 -5.65448701e-01 -3.82570386e-01
6.49234653e-02 -2.93552428e-02 -6.49809659e-01 4.69884455e-01
7.12366045e-01 -3.64197403e-01 -8.06434929e-01 1.15919077e+00
5.35640717e-01 5.95110320e-02 6.29303813e-01 -1.19166076e+00
-1.12531865e+00 3.36918890e-01 -8.76183331e-01 7.56015703e-02
2.71799117e-01 4.73828241e-02 1.08835363e+00 -8.23953867e-01
5.88960052e-01 1.23077536e+00 4.03782874e-01 2.53045171e-01
-7.54965365e-01 -4.32483286e-01 2.80202270e-01 5.73979199e-01
-1.23019981e+00 -2.80531138e-01 1.01338792e+00 -3.07539910e-01
9.28624272e-01 3.48021388e-01 4.98556376e-01 9.27488983e-01
1.23765625e-01 1.31706595e+00 7.94414341e-01 -7.02040136e-01
-1.88077152e-01 2.16200575e-02 2.58352071e-01 1.14819622e+00
2.10214838e-01 -6.36091113e-01 -4.70377892e-01 3.58447105e-01
7.66009033e-01 2.54779965e-01 -4.57540512e-01 -1.87485680e-01
-1.03294194e+00 6.45805538e-01 5.24295926e-01 3.36572647e-01
2.44929548e-02 3.60882938e-01 4.97143477e-01 2.38234401e-01
2.51439512e-01 3.04289103e-01 -2.64164329e-01 -5.54387569e-02
-8.07592452e-01 5.54603115e-02 5.46582639e-01 1.28660321e+00
6.81900382e-01 1.08403908e-02 -3.00324798e-01 1.21108353e+00
2.01607466e-01 7.73268521e-01 4.37993705e-01 -4.42172647e-01
7.30295599e-01 9.24937487e-01 -2.72974014e-01 -1.27256215e+00
-2.14299247e-01 -1.55092686e-01 -9.60164785e-01 -7.18721598e-02
3.56469482e-01 1.08972803e-01 -1.21042800e+00 9.93756771e-01
4.92179766e-03 -9.01058316e-02 -1.75739571e-01 8.89775991e-01
1.00631106e+00 7.94352710e-01 -3.67096782e-01 2.97187060e-01
1.48509407e+00 -1.22602856e+00 -7.03715980e-01 -4.82007027e-01
4.87483531e-01 -1.18152893e+00 1.07422090e+00 3.49620789e-01
-6.78226769e-01 -4.75886017e-01 -1.19856930e+00 -6.38554811e-01
-7.02502668e-01 6.31968439e-01 8.85550022e-01 4.33966070e-01
-9.98150289e-01 3.76366258e-01 -5.13387322e-01 -6.03895187e-01
5.71397901e-01 1.55364543e-01 -4.51507092e-01 -5.10531604e-01
-7.82106102e-01 4.04582471e-01 3.19246471e-01 4.76779819e-01
-6.38265312e-01 -5.56188487e-02 -8.80199134e-01 1.31644353e-01
7.43573487e-01 -1.07505955e-01 1.05284345e+00 -6.19774461e-01
-1.32274377e+00 5.80968380e-01 -1.34678081e-01 8.50875303e-02
5.91140032e-01 -4.47030485e-01 -2.38840222e-01 3.45115483e-01
-1.06782690e-01 3.32012773e-01 9.51177716e-01 -1.09792042e+00
-3.49555373e-01 -4.09381509e-01 -8.81476030e-02 1.65908530e-01
-6.74551845e-01 1.96518153e-02 -1.33383250e+00 -7.63273358e-01
-3.08533981e-02 -9.07028139e-01 5.89289367e-02 2.22187370e-01
-9.07616138e-01 -2.31793687e-01 1.00597799e+00 -7.90024698e-01
1.21104407e+00 -1.94003117e+00 -4.22935337e-02 1.71124879e-02
2.35705644e-01 6.18910968e-01 -3.33199739e-01 7.85415411e-01
2.54505187e-01 1.59171626e-01 -1.15535893e-01 -3.65697831e-01
-1.56774111e-02 -1.37704954e-01 -3.73489767e-01 4.35957611e-01
1.75948679e-01 1.11176062e+00 -6.75941765e-01 -6.58541739e-01
4.02776837e-01 4.09826517e-01 -1.02272123e-01 1.49052218e-01
-3.65744144e-01 -2.45714992e-01 -6.40652180e-01 1.12131405e+00
7.64642835e-01 -3.60896081e-01 1.77358761e-01 -1.57383606e-01
-4.77677248e-02 -2.60355234e-01 -1.09495461e+00 1.60115671e+00
-1.80091754e-01 1.21587050e+00 -1.40830763e-02 -8.75701666e-01
8.55429530e-01 -1.30398244e-01 1.03567950e-01 -7.65542924e-01
1.01797521e-01 1.93254836e-02 -3.08335662e-01 -6.48584664e-01
8.36137891e-01 6.64180160e-01 8.33970532e-02 2.35660464e-01
-7.77634084e-02 -8.20938423e-02 4.14484948e-01 4.84136164e-01
1.09464598e+00 -4.00877092e-03 7.23737553e-02 3.96456867e-02
5.80980539e-01 1.19634405e-01 1.52434379e-01 7.92951882e-01
-7.36311749e-02 8.49118292e-01 5.62424004e-01 -1.99349150e-01
-1.14813781e+00 -6.79312944e-01 1.37211815e-01 9.08274114e-01
1.84763864e-01 -4.97315973e-01 -6.19998336e-01 -6.52721226e-01
1.36592135e-01 2.35853598e-01 -4.66268033e-01 1.60634965e-01
-6.38217390e-01 -5.46957374e-01 5.07022679e-01 6.88362420e-01
5.67037940e-01 -1.24056566e+00 -1.39513925e-01 -6.22709803e-02
1.60149001e-02 -1.41108108e+00 -8.26624453e-01 9.38982330e-03
-6.28549814e-01 -1.24165452e+00 -9.73332524e-01 -1.08318746e+00
8.06004763e-01 8.13004315e-01 6.72039449e-01 3.91946107e-01
-9.00552452e-01 3.97895277e-01 -5.65384209e-01 -4.39109415e-01
-9.18349400e-02 1.04238294e-01 -5.30605674e-01 4.24661189e-02
3.96544307e-01 5.23282699e-02 -6.21652007e-01 -1.79134607e-02
-1.12559497e+00 2.05107749e-01 8.41462970e-01 7.04294562e-01
2.62518048e-01 1.45290747e-01 8.72237012e-02 -9.25319314e-01
8.24265301e-01 5.98830953e-02 -6.83911443e-01 5.48957169e-01
-3.66584599e-01 6.98859468e-02 7.15695202e-01 -2.05210075e-01
-8.77881110e-01 -1.08217195e-01 5.64546548e-02 -3.50386947e-01
-2.36307874e-01 5.56924999e-01 -2.15018883e-01 -1.73992943e-02
1.93419337e-01 5.21863699e-01 -1.40722185e-01 -6.89699411e-01
4.02549922e-01 1.07552326e+00 4.61790890e-01 -3.59395474e-01
7.87993848e-01 5.53582072e-01 4.28773016e-02 -1.35980642e+00
-6.00316584e-01 -8.46678436e-01 -8.14026415e-01 -1.67065382e-01
7.60692716e-01 -7.19420195e-01 -6.30164444e-01 1.13396323e+00
-1.26639795e+00 -3.15917909e-01 4.56797957e-01 1.26823023e-01
-1.55481532e-01 8.92489195e-01 -7.91400075e-01 -7.39081204e-01
-5.23902178e-01 -1.17664135e+00 1.20219874e+00 4.26877081e-01
4.37892646e-01 -9.72612798e-01 -2.93829441e-01 4.58493382e-01
2.24465072e-01 1.39081717e-01 1.09049153e+00 -3.12175363e-01
-9.52375650e-01 -5.01387358e-01 -8.04595470e-01 3.97199064e-01
1.89469635e-01 3.61223757e-01 -8.45849156e-01 -4.10185307e-01
-6.37298942e-01 -2.33068451e-01 1.25161564e+00 2.26778507e-01
1.37897539e+00 -4.02039811e-02 -3.77863139e-01 6.42786920e-01
1.82060826e+00 1.39923126e-01 7.61859357e-01 4.46384877e-01
1.34328377e+00 4.46102858e-01 3.88295799e-01 4.15033698e-01
2.93351114e-01 4.86769021e-01 3.28745127e-01 -1.66155845e-01
-4.00058657e-01 -1.60897732e-01 3.19007337e-01 8.40840816e-01
1.25139803e-01 -7.96864808e-01 -1.08989179e+00 4.66255516e-01
-2.03151059e+00 -7.58059680e-01 -3.16480488e-01 1.86961055e+00
6.78531289e-01 1.17154166e-01 -2.03927413e-01 4.68195938e-02
9.91326153e-01 3.90139997e-01 -4.36450839e-01 -4.37292218e-01
-1.76476702e-01 -8.27798173e-02 6.83482945e-01 2.48516351e-01
-1.28011358e+00 1.17462718e+00 5.12280321e+00 8.67544055e-01
-1.10899997e+00 -4.42868471e-01 4.77852255e-01 2.58857369e-01
-5.19575737e-02 -6.96426928e-02 -8.62857223e-01 4.01746511e-01
2.88266838e-01 1.74789146e-01 4.33358043e-01 7.65084445e-01
2.51344815e-02 -1.32191226e-01 -9.32156444e-01 1.12337673e+00
3.13209534e-01 -1.45708084e+00 2.57157713e-01 -1.71832722e-02
5.79725444e-01 -5.70172854e-02 6.83424994e-02 -2.60626897e-03
9.99150500e-02 -9.68346179e-01 6.48990273e-01 4.61881518e-01
1.04733860e+00 -7.09702313e-01 6.26996994e-01 7.66279325e-02
-1.37553620e+00 -2.28495356e-02 -6.68765843e-01 1.74566075e-01
-2.47634187e-01 6.65675104e-01 -9.21009243e-01 4.60848451e-01
5.47964811e-01 8.96525621e-01 -9.89311457e-01 1.28520548e+00
-4.80812073e-01 3.60541433e-01 8.68650824e-02 -5.30482948e-01
4.11605239e-01 -2.23065585e-01 3.50210249e-01 1.62447679e+00
1.49348378e-01 -2.89980263e-01 2.32297510e-01 6.19965851e-01
-4.10570562e-01 2.57965267e-01 -5.01956642e-01 -6.56132102e-01
2.95653313e-01 1.50274408e+00 -9.86585081e-01 -2.01025575e-01
-5.10325730e-01 1.31704283e+00 4.70300049e-01 5.31862319e-01
-6.14730895e-01 -1.13487911e+00 4.19535458e-01 -8.98937881e-02
4.23012257e-01 -7.43259668e-01 -6.30386779e-03 -1.21193314e+00
2.63245314e-01 -6.93702579e-01 1.30214885e-01 -9.31395113e-01
-1.17226684e+00 3.43714744e-01 -3.83941799e-01 -9.22923863e-01
2.13869423e-01 -1.24589729e+00 -6.66637659e-01 8.40287328e-01
-1.76444972e+00 -1.42701364e+00 -7.12931037e-01 5.68168104e-01
9.34720337e-01 -1.89148918e-01 4.63746101e-01 1.11110896e-01
-9.09922779e-01 6.34629905e-01 5.77874899e-01 8.10424387e-01
8.18476975e-01 -1.50288951e+00 6.58520699e-01 9.57965195e-01
3.21493775e-01 5.12522876e-01 3.25380504e-01 -6.81462765e-01
-1.98606086e+00 -1.07913709e+00 5.73095143e-01 -2.58875489e-01
7.18477547e-01 -8.73724699e-01 -7.79254913e-01 4.91386145e-01
3.59676331e-01 -1.68709442e-01 3.36043417e-01 -6.86074719e-02
-4.54940617e-01 1.70763675e-02 -5.42450845e-01 8.39869976e-01
8.39234352e-01 -7.00705588e-01 -1.63241625e-01 5.98560691e-01
6.18916273e-01 -3.70890230e-01 -3.48668486e-01 6.26629777e-03
4.93846506e-01 -7.22514272e-01 7.98286557e-01 -3.52514118e-01
5.81704795e-01 -3.17740440e-01 3.93045768e-02 -1.05994618e+00
-2.80197591e-01 -4.87460285e-01 -1.01213701e-01 1.29729366e+00
2.25842580e-01 -3.49314779e-01 7.07956791e-01 2.01259702e-01
-1.64725244e-01 -6.93722546e-01 -2.76975453e-01 -6.57624424e-01
-1.49067968e-01 -4.31107312e-01 2.60936081e-01 7.18699217e-01
-1.31558046e-01 3.36105287e-01 -4.33301717e-01 8.24253932e-02
6.22072458e-01 3.87443900e-01 7.44386852e-01 -9.89073873e-01
-1.65711507e-01 -6.23811603e-01 -5.98619878e-01 -1.10557985e+00
-8.45660362e-03 -9.60326016e-01 1.28813665e-02 -2.03099537e+00
5.41771948e-01 -9.20450538e-02 -9.98758674e-02 5.38525939e-01
-2.16546178e-01 2.09472969e-01 4.54950064e-01 3.10380191e-01
-7.17736006e-01 3.99223357e-01 1.36734581e+00 -6.49299085e-01
5.27639352e-02 -3.59867901e-01 -6.21287465e-01 5.19219398e-01
8.28141451e-01 -5.19529544e-02 -2.52251804e-01 -7.55510390e-01
7.83437788e-02 -2.50897348e-01 2.44741946e-01 -7.95781612e-01
4.28542316e-01 -1.35984406e-01 7.67336130e-01 -7.52589166e-01
2.40631804e-01 -5.15524864e-01 -6.52186453e-01 1.68458000e-01
-3.08112532e-01 1.61439329e-01 3.00497383e-01 7.33579218e-01
-1.14685580e-01 -4.32937294e-01 5.64245105e-01 -1.99035242e-01
-1.20467961e+00 4.04559433e-01 -2.56411314e-01 -1.28521979e-01
8.03326607e-01 -2.52311051e-01 -7.95627475e-01 -2.46208161e-01
-2.72640198e-01 3.07907790e-01 5.76292515e-01 7.50386894e-01
9.58229363e-01 -1.08913136e+00 -6.18482530e-01 4.92659435e-02
3.45385045e-01 -7.78172985e-02 2.33362585e-01 4.10290241e-01
-1.03742027e+00 8.92832637e-01 -9.49985385e-02 -4.73909497e-01
-1.48002827e+00 6.78276837e-01 -1.13644943e-01 -1.39247254e-01
-7.41011024e-01 8.75000417e-01 8.41389298e-02 -7.43354112e-02
4.72233295e-01 -2.36543909e-01 -1.94133312e-01 9.99692176e-03
6.41028881e-01 2.98052967e-01 1.80911511e-01 -4.74346220e-01
-3.76373231e-01 6.73991561e-01 -5.46896458e-01 1.82843015e-01
1.39478207e+00 -5.71518205e-02 -2.34650046e-01 1.03129677e-01
1.46262360e+00 -1.96606610e-02 -1.14067101e+00 -3.33530188e-01
1.28751576e-01 -7.07272172e-01 1.89116821e-01 -7.09771395e-01
-1.13126171e+00 1.24035394e+00 4.50330406e-01 1.79628134e-01
9.82929647e-01 -1.20713085e-01 9.16060507e-01 7.17111528e-01
1.00531705e-01 -1.38807273e+00 5.42519748e-01 5.52468538e-01
1.02714932e+00 -1.39101434e+00 2.44060993e-01 -4.85229731e-01
-5.58128417e-01 1.53085113e+00 5.93105495e-01 -1.13132253e-01
3.06314260e-01 2.05488786e-01 1.02260657e-01 -2.72187740e-01
-3.90933245e-01 -3.28737378e-01 5.02597511e-01 5.59743702e-01
5.87568879e-01 6.47536665e-02 3.57504189e-02 1.32992193e-01
2.77184755e-01 -3.92457664e-01 7.17900395e-01 1.17389584e+00
-5.98093271e-01 -1.13484180e+00 -2.92478114e-01 5.05925298e-01
-4.18590099e-01 -3.48684013e-01 -6.92976356e-01 7.36221254e-01
-4.15330589e-01 9.20842886e-01 -1.97783411e-01 -2.56740510e-01
1.31998539e-01 -1.02152295e-01 5.44499993e-01 -6.26960397e-01
-9.62496623e-02 2.31801480e-01 -1.42561451e-01 -4.55951422e-01
2.40356326e-02 -5.42790353e-01 -1.28499794e+00 -2.65547752e-01
-4.39777672e-01 -2.40719244e-01 9.50273395e-01 7.27192283e-01
3.04277748e-01 6.38868690e-01 4.70661044e-01 -7.08012819e-01
-3.98129046e-01 -8.68523359e-01 -7.69731522e-01 4.09895241e-01
3.79894406e-01 -3.69270682e-01 -1.21584497e-01 9.27795842e-02] | [11.661229133605957, 2.3574225902557373] |
0b8ce6a9-2a07-423d-a218-8c170920e267 | uturku-drug-named-entity-recognition-and-drug | null | null | https://aclanthology.org/S13-2108 | https://aclanthology.org/S13-2108.pdf | UTurku: Drug Named Entity Recognition and Drug-Drug Interaction Extraction Using SVM Classification and Domain Knowledge | null | ['Jari Bj{\\"o}rne', 'Tapio Salakoski', 'Suwisa Kaewphan'] | 2013-06-01 | null | null | null | semeval-2013-6 | ['drug-drug-interaction-extraction'] | ['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] | [-7.226303577423096, 3.47792911529541] |
1173f79d-bf4f-4e56-bbdd-0b467aa97aeb | a-static-evaluation-of-code-completion-by | 2306.03203 | null | https://arxiv.org/abs/2306.03203v1 | https://arxiv.org/pdf/2306.03203v1.pdf | A Static Evaluation of Code Completion by Large Language Models | Large language models trained on code have shown great potential to increase productivity of software developers. Several execution-based benchmarks have been proposed to evaluate functional correctness of model-generated code on simple programming problems. Nevertheless, it is expensive to perform the same evaluation on complex real-world projects considering the execution cost. On the contrary, static analysis tools such as linters, which can detect errors without running the program, haven't been well explored for evaluating code generation models. In this work, we propose a static evaluation framework to quantify static errors in Python code completions, by leveraging Abstract Syntax Trees. Compared with execution-based evaluation, our method is not only more efficient, but also applicable to code in the wild. For experiments, we collect code context from open source repos to generate one million function bodies using public models. Our static analysis reveals that Undefined Name and Unused Variable are the most common errors among others made by language models. Through extensive studies, we also show the impact of sampling temperature, model size, and context on static errors in code completions. | ['Bing Xiang', 'Dan Roth', 'Sudipta Sengupta', 'Parminder Bhatia', 'Baishakhi Ray', 'Murali Krishna Ramanathan', 'Xiaopeng Li', 'Rob Kwiatkowski', 'Zijian Wang', 'Yuchen Tian', 'Varun Kumar', 'Hantian Ding'] | 2023-06-05 | null | null | null | null | ['code-generation'] | ['computer-code'] | [-1.71804354e-01 -5.13035208e-02 -9.70843211e-02 -4.53794301e-01
-7.17708826e-01 -5.83649814e-01 1.78359419e-01 3.64853829e-01
-1.21866740e-01 4.44627315e-01 -1.62552699e-01 -7.97036469e-01
4.07193005e-01 -8.94464850e-01 -1.00565314e+00 9.27893072e-02
1.07703634e-01 -2.78709590e-01 3.63525540e-01 -9.91534144e-02
5.64806402e-01 -7.24354684e-02 -1.61532187e+00 5.40815413e-01
1.36939073e+00 2.17478633e-01 2.97883838e-01 7.21638381e-01
-5.59570670e-01 1.06038511e+00 -8.71009707e-01 -7.07118452e-01
-1.11000746e-01 -2.51386017e-01 -7.55886972e-01 -5.18481135e-01
2.92460293e-01 -2.45535180e-01 3.23667377e-01 1.06755602e+00
4.60842729e-01 -5.36218166e-01 1.51714072e-01 -1.22220325e+00
-4.53236580e-01 1.14277065e+00 -5.76328158e-01 -4.57922891e-02
5.54360092e-01 2.28336439e-01 9.70256805e-01 -7.87502289e-01
5.21633148e-01 9.70824361e-01 9.14630353e-01 6.02455080e-01
-1.21870530e+00 -5.32565176e-01 -1.69914812e-01 -1.76747933e-01
-1.18869674e+00 -2.46763065e-01 4.11717385e-01 -9.30871546e-01
1.46116245e+00 4.25739229e-01 3.83287787e-01 9.87173617e-01
5.44405341e-01 4.35826421e-01 8.76278043e-01 -7.37977326e-01
2.27126226e-01 3.04991692e-01 3.86546582e-01 8.30130994e-01
4.25756812e-01 -2.24004090e-01 -1.39027044e-01 -6.88646078e-01
-5.64249903e-02 -1.20938681e-02 -1.43825784e-01 -9.18041542e-03
-1.07600653e+00 5.13928056e-01 2.40310691e-02 3.76191944e-01
1.04476243e-01 7.03686893e-01 5.69164336e-01 1.53448045e-01
2.44998321e-01 6.53166354e-01 -7.87574112e-01 -7.51366138e-01
-9.73328412e-01 4.22541857e-01 9.69036400e-01 1.18567550e+00
1.09107387e+00 1.25858739e-01 -2.82771230e-01 7.28940904e-01
4.24962372e-01 4.39689010e-01 6.44955397e-01 -6.26152277e-01
7.77181268e-01 1.45779145e+00 -8.19524601e-02 -9.90591824e-01
-3.14797580e-01 -4.41862524e-01 -8.38171020e-02 4.03811574e-01
2.90960878e-01 -9.63204503e-02 -3.06569248e-01 1.33079219e+00
-9.58873630e-02 -2.76642740e-01 -2.28495479e-01 2.19563574e-01
6.79996789e-01 3.30007762e-01 1.86447650e-02 1.90611333e-01
1.00906670e+00 -9.71838474e-01 -2.80005515e-01 -3.12571257e-01
1.34136248e+00 -9.06701386e-01 1.49331999e+00 2.46142983e-01
-9.15042758e-01 -4.17528331e-01 -9.39635694e-01 5.68686686e-02
-2.33622834e-01 4.85284358e-01 7.64384687e-01 1.15304708e+00
-9.96868968e-01 6.76483870e-01 -1.02556515e+00 -3.84540558e-02
2.76353210e-01 -8.81408108e-04 7.18235178e-03 2.13160545e-01
-5.31578362e-01 6.71231508e-01 1.67708188e-01 -2.02876836e-01
-9.99584079e-01 -1.10006928e+00 -8.34118903e-01 1.13844097e-01
1.23977818e-01 -2.57669806e-01 1.48815429e+00 -9.54811990e-01
-1.09569418e+00 6.84324324e-01 -3.06698114e-01 -1.54447109e-01
4.64438438e-01 -3.26368332e-01 -1.72125205e-01 -8.09526920e-01
1.48043767e-01 -9.04818028e-02 3.40020746e-01 -1.16450727e+00
-4.11515266e-01 -9.76678208e-02 4.74749833e-01 -7.81022966e-01
-4.67346430e-01 5.17291188e-01 -2.59960443e-01 -3.72670382e-01
-3.94204468e-01 -8.23058963e-01 -1.73170298e-01 -2.77143031e-01
-4.20506358e-01 -2.61297803e-02 2.97334075e-01 -9.85468268e-01
1.91618943e+00 -2.16261244e+00 -1.87733158e-01 1.69402331e-01
2.47130036e-01 2.02469230e-01 -8.06351975e-02 4.60941374e-01
-3.39646563e-02 6.71946347e-01 -4.86538857e-01 -9.54539478e-02
2.14139804e-01 -1.73248664e-01 -3.72873664e-01 1.07203871e-01
5.23712076e-02 8.97195339e-01 -7.98452914e-01 -4.04939026e-01
-2.57304788e-01 1.31113261e-01 -1.14071023e+00 1.97864354e-01
-5.71857512e-01 -1.37299255e-01 -3.35601985e-01 8.61012042e-01
5.94323218e-01 -2.60648698e-01 1.74316481e-01 5.21009445e-01
-3.61304730e-01 4.03471768e-01 -8.03721130e-01 1.66818941e+00
-1.02054882e+00 6.57194853e-01 -3.61438990e-01 -3.63934785e-01
1.05860436e+00 1.28791079e-01 -2.27938369e-01 -6.17960215e-01
-3.74630809e-01 5.83911657e-01 2.72413969e-01 -7.33730555e-01
6.50969923e-01 4.88300651e-01 -3.68403941e-01 5.84445536e-01
-3.00810575e-01 -3.86261865e-02 4.86473978e-01 1.03384562e-01
1.60952044e+00 7.95849860e-01 2.47858450e-01 -1.09226026e-01
5.85805595e-01 3.02986354e-02 6.03102505e-01 7.45621562e-01
2.78947860e-01 5.09983301e-01 1.05737448e+00 -3.81711036e-01
-1.07175016e+00 -6.48990214e-01 -4.52117436e-03 1.23035908e+00
-4.07623351e-01 -1.24313426e+00 -1.15067387e+00 -8.63800347e-01
-1.49699777e-01 9.30182219e-01 -3.84763032e-01 -1.39854044e-01
-7.30608225e-01 -8.04589570e-01 9.94601130e-01 5.17819285e-01
2.83093035e-01 -8.67728949e-01 -9.53446865e-01 2.48221263e-01
-1.28590792e-01 -7.70778775e-01 -2.37416759e-01 -3.08791310e-01
-7.62961507e-01 -1.13339698e+00 -3.39974731e-01 -4.62131828e-01
7.11582959e-01 -2.10821018e-01 1.50770211e+00 8.49375546e-01
-6.16917431e-01 1.21156886e-01 -3.86568218e-01 -4.37421650e-01
-1.09642196e+00 1.84225380e-01 -5.91969073e-01 -5.20783901e-01
1.62420943e-01 -5.38617432e-01 -1.95418820e-01 3.16104174e-01
-8.51075649e-01 6.16481528e-02 5.23950696e-01 5.69158077e-01
6.42347857e-02 -3.49119127e-01 2.51357496e-01 -1.16167378e+00
5.67099750e-01 -4.80047762e-01 -1.03499269e+00 5.66596150e-01
-6.59486353e-01 2.95069069e-01 8.52248609e-01 -9.74542052e-02
-1.12109756e+00 -1.96781293e-01 -4.14625943e-01 2.96254963e-01
1.06980555e-01 8.15902412e-01 4.14145226e-03 -1.43452287e-01
1.21014953e+00 2.74817824e-01 -4.01125580e-01 -6.48841023e-01
-1.67463094e-01 6.78598642e-01 6.73881695e-02 -1.00562286e+00
6.91860735e-01 -9.79595701e-04 -3.45532447e-01 -5.43667138e-01
-7.69217312e-02 -3.93684693e-02 -1.92973986e-01 8.03573132e-02
3.85015875e-01 -7.59100616e-01 -5.96305609e-01 4.82268900e-01
-1.38921905e+00 -4.72156078e-01 8.54401439e-02 5.34595996e-02
-3.18638355e-01 5.52027464e-01 -3.19188088e-01 -8.23510706e-01
-3.21245939e-01 -1.57462919e+00 1.03859055e+00 -3.71119529e-02
-4.68198210e-01 -7.29392767e-01 3.37992400e-01 2.10796773e-01
8.27162445e-01 4.07801419e-01 1.41068244e+00 -2.83872962e-01
-7.26887107e-01 -2.40408331e-01 -9.90593880e-02 3.97524655e-01
-1.64627001e-01 7.20843971e-01 -9.18285787e-01 -9.43970829e-02
-3.44826251e-01 5.76576479e-02 2.04839826e-01 -2.42450431e-01
1.31867301e+00 -3.28686863e-01 -3.95748556e-01 5.06009400e-01
1.60830522e+00 -9.37615484e-02 9.68632400e-01 5.04047632e-01
5.75963736e-01 3.94455522e-01 3.94993335e-01 6.60958886e-01
4.06102568e-01 5.80016553e-01 5.07331908e-01 2.83989102e-01
-4.86153141e-02 -2.53938258e-01 7.92382836e-01 7.05832899e-01
-1.44979447e-01 8.48388523e-02 -1.66393530e+00 5.89903057e-01
-1.54322839e+00 -6.27129138e-01 -7.58855164e-01 2.36272311e+00
1.13630617e+00 1.71532825e-01 -1.15133606e-01 8.47567990e-02
3.65887284e-01 -3.31227422e-01 3.45177427e-02 -5.74654877e-01
3.28657866e-01 2.64046341e-01 2.01914772e-01 2.68936902e-01
-4.56080705e-01 5.24435580e-01 5.92027569e+00 5.38190901e-01
-1.18795753e+00 3.28276515e-01 2.19329447e-01 1.49195880e-01
-8.77974749e-01 6.35996640e-01 -1.02645218e+00 7.49676168e-01
1.18847346e+00 -4.55385000e-01 2.83652604e-01 1.49499846e+00
4.24156003e-02 -2.76557565e-01 -1.32845318e+00 5.11750698e-01
-1.01954401e-01 -1.21837068e+00 -2.98638105e-01 -1.31990775e-01
8.30641866e-01 8.74545872e-02 -3.78238440e-01 7.10178316e-01
1.50249511e-01 -9.47249234e-01 1.05293643e+00 5.61647058e-01
5.24757147e-01 -3.98360848e-01 7.87689865e-01 5.32907605e-01
-8.67573142e-01 -1.32326901e-01 -3.33073765e-01 -3.01638097e-01
-4.41007018e-01 8.81913781e-01 -1.06003761e+00 3.72543603e-01
8.94499540e-01 3.44585240e-01 -1.40992343e+00 1.22980416e+00
-1.71239823e-01 8.34300160e-01 2.60622427e-02 -3.59399021e-01
-3.08749348e-01 1.90817833e-01 7.73809031e-02 1.71185148e+00
4.97464359e-01 -7.99510062e-01 -1.17068470e-01 1.47110879e+00
-5.62093593e-02 4.50584829e-01 -5.80583513e-01 -1.82309851e-01
2.01042637e-01 1.12040806e+00 -3.64312768e-01 -1.94834635e-01
-7.62623370e-01 3.97251368e-01 4.27518487e-01 2.17085779e-01
-1.15616858e+00 -7.14284658e-01 4.43171948e-01 3.70767385e-01
-1.58852264e-01 -3.75836194e-01 -4.73707408e-01 -1.17851734e+00
7.46141493e-01 -1.11990297e+00 -1.85635954e-01 -6.76099062e-01
-5.46365798e-01 5.40956020e-01 -1.52421836e-02 -9.70822811e-01
-1.42933160e-01 -3.72376263e-01 -8.61745894e-01 9.97244656e-01
-1.32347870e+00 -7.67985046e-01 -4.20232773e-01 -8.03251266e-02
2.39422381e-01 -1.83932990e-01 8.67895126e-01 6.16626441e-01
-6.48783624e-01 8.71616066e-01 1.05034122e-02 1.19992979e-01
3.09487760e-01 -1.22118092e+00 9.14594173e-01 1.21826768e+00
-3.03630978e-01 1.33531725e+00 6.74690664e-01 -6.61400974e-01
-1.52578223e+00 -1.04186845e+00 1.03024507e+00 -7.24527597e-01
6.35020494e-01 -5.97976685e-01 -1.08535814e+00 7.34565854e-01
-1.06421009e-01 4.09838818e-02 5.49236178e-01 1.08274825e-01
-7.33046651e-01 -4.32147905e-02 -8.49157870e-01 5.96599698e-01
7.69564569e-01 -5.57820320e-01 -1.62811354e-01 3.42781812e-01
7.15739191e-01 -5.06895959e-01 -9.74324703e-01 2.40459055e-01
6.00889862e-01 -1.32871556e+00 4.94702369e-01 -3.49996030e-01
1.11603224e+00 -3.87450218e-01 -2.61577427e-01 -9.30247068e-01
2.23441377e-01 -5.42487681e-01 1.03697725e-01 1.64830065e+00
9.88918304e-01 -7.42988706e-01 4.94337142e-01 8.85725915e-01
-3.03942144e-01 -7.48432219e-01 -3.92984182e-01 -8.24305356e-01
1.49406433e-01 -7.05269277e-01 9.04293597e-01 7.12233067e-01
1.09018043e-01 -1.73851565e-01 -4.13293205e-02 -9.72861424e-02
9.83289704e-02 1.95706859e-01 1.26063943e+00 -8.54642391e-01
-8.43212545e-01 -4.66615498e-01 -3.61284763e-01 -3.83508056e-01
2.50315219e-01 -1.05854976e+00 1.09970108e-01 -1.29478455e+00
4.77417916e-01 -6.26530170e-01 4.08582240e-01 6.91016853e-01
-3.17916632e-01 -2.40609542e-01 -4.13387688e-03 -5.18898480e-03
-4.01497632e-01 1.36381030e-01 5.26465893e-01 -1.08273610e-01
-3.61782126e-02 6.59874380e-02 -4.72609580e-01 8.01140130e-01
7.47723162e-01 -8.27296019e-01 1.53840622e-02 -6.33675992e-01
1.20258462e+00 7.66083375e-02 3.76578122e-01 -1.24756908e+00
-1.41687870e-01 -1.36917472e-01 -2.75689155e-01 -1.32249907e-01
-3.72090310e-01 -5.52861333e-01 4.03307289e-01 6.57584131e-01
-3.66557926e-01 4.15973157e-01 3.84926349e-01 1.61656305e-01
-9.54085961e-02 -9.70905781e-01 3.90157551e-01 -2.40411773e-01
-6.05040371e-01 -1.08626097e-01 -2.54962116e-01 2.16995731e-01
9.50671434e-01 -7.35485135e-03 -5.49253047e-01 3.39101642e-01
-1.46054715e-01 -1.53245926e-01 9.02962983e-01 6.54935837e-01
3.17552119e-01 -7.54034817e-01 -7.15832591e-01 1.41797870e-01
5.38373113e-01 -4.07053560e-01 -6.31752387e-02 7.30682611e-01
-1.05033135e+00 2.30414852e-01 2.31044535e-02 -6.55294120e-01
-1.32356966e+00 4.89276111e-01 2.87136883e-01 -3.02449346e-01
-2.54588753e-01 5.53408146e-01 -7.04369694e-02 -1.00827813e+00
-9.83262658e-02 -8.11622918e-01 4.01828915e-01 -5.51504433e-01
5.26931643e-01 4.93383408e-01 5.91329336e-01 -6.01076186e-02
-4.78873402e-01 4.05429065e-01 1.43613741e-01 2.30511323e-01
1.21746492e+00 4.67881590e-01 -7.52035677e-01 5.27768612e-01
1.09781218e+00 5.41311383e-01 -6.94965303e-01 8.71902853e-02
4.96967345e-01 -6.54956460e-01 -4.98632580e-01 -8.40547621e-01
-7.41042256e-01 1.05151069e+00 3.20926517e-01 3.11514139e-01
6.12237871e-01 -2.77756304e-01 2.65778095e-01 4.78354275e-01
9.20096576e-01 -7.61767626e-01 -1.40221626e-01 5.22875071e-01
8.08034837e-01 -1.05753171e+00 -6.71438575e-02 -4.36282188e-01
-1.27040580e-01 1.25639260e+00 9.64859307e-01 2.37194881e-01
2.31141031e-01 6.14425540e-01 -2.54689723e-01 1.55298516e-01
-8.15092981e-01 4.34031308e-01 -2.19836548e-01 4.08306748e-01
1.13968742e+00 -3.31713147e-02 -5.52359343e-01 6.70910001e-01
-2.31350347e-01 3.63852322e-01 1.11276627e+00 1.19622648e+00
-2.31717288e-01 -1.39777231e+00 -4.51703876e-01 5.45732915e-01
-8.21599662e-01 -3.78672719e-01 -1.70837656e-01 6.34136975e-01
6.78593963e-02 7.92463779e-01 -4.27752107e-01 -3.31837147e-01
4.62009430e-01 2.09215239e-01 4.50958818e-01 -1.06792915e+00
-1.22046638e+00 -7.55939841e-01 2.87775695e-01 -6.76177144e-01
1.29138470e-01 -3.95020515e-01 -1.31708956e+00 -2.91444212e-01
-4.03772920e-01 1.06187044e-02 1.02298415e+00 6.75678194e-01
6.76513731e-01 7.17226088e-01 2.61016995e-01 -3.08169693e-01
-7.14540243e-01 -8.01569223e-01 -4.40676622e-02 2.32992619e-01
7.03295991e-02 -3.06100726e-01 -3.05400014e-01 2.01539934e-01] | [7.791095733642578, 7.772904396057129] |
ec7a6ab2-56b5-48d0-a7e8-a98573e7a486 | evaluation-of-preprocessing-techniques-for-u | 2103.14301 | null | https://arxiv.org/abs/2103.14301v1 | https://arxiv.org/pdf/2103.14301v1.pdf | Evaluation of Preprocessing Techniques for U-Net Based Automated Liver Segmentation | To extract liver from medical images is a challenging task due to similar intensity values of liver with adjacent organs, various contrast levels, various noise associated with medical images and irregular shape of liver. To address these issues, it is important to preprocess the medical images, i.e., computerized tomography (CT) and magnetic resonance imaging (MRI) data prior to liver analysis and quantification. This paper investigates the impact of permutation of various preprocessing techniques for CT images, on the automated liver segmentation using deep learning, i.e., U-Net architecture. The study focuses on Hounsfield Unit (HU) windowing, contrast limited adaptive histogram equalization (CLAHE), z-score normalization, median filtering and Block-Matching and 3D (BM3D) filtering. The segmented results show that combination of three techniques; HU-windowing, median filtering and z-score normalization achieve optimal performance with Dice coefficient of 96.93%, 90.77% and 90.84% for training, validation and testing respectively. | ['Muhammad Salman Khan', 'Kaleem Nawaz Khan', 'Muhammad Islam'] | 2021-03-26 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [-1.41802490e-01 -4.11732852e-01 1.17212735e-01 -4.11438167e-01
-1.77359372e-01 -5.99944293e-01 2.60507941e-01 4.64877814e-01
-5.03582299e-01 6.18176937e-01 3.50802451e-01 -5.00832379e-01
-1.27518788e-01 -7.05769122e-01 -6.95625022e-02 -9.04819071e-01
-7.38531411e-01 2.62153476e-01 2.25478530e-01 4.63744879e-01
3.21485549e-01 9.63305593e-01 -6.74672306e-01 6.29282892e-02
8.81859601e-01 8.63051891e-01 -1.10742837e-01 1.08436930e+00
9.13926512e-02 7.03624129e-01 -3.71378452e-01 -2.97130626e-02
7.21387267e-01 -6.81244254e-01 -7.93152988e-01 2.14590386e-01
-1.79687329e-02 -6.38858140e-01 -2.09798738e-01 1.22560203e+00
8.52329433e-01 2.49271512e-01 1.03440833e+00 -8.14732969e-01
-5.64215720e-01 6.42533302e-01 -9.16242778e-01 9.30719793e-01
-2.00015604e-01 3.01627398e-01 -1.04676738e-01 -6.27504885e-01
3.04063261e-01 6.39360487e-01 7.48610854e-01 1.35169655e-01
-9.05753672e-01 -6.44184291e-01 -1.02428031e+00 6.21638000e-02
-1.37008309e+00 1.62903394e-03 2.34258577e-01 -6.71504855e-01
6.84427619e-01 4.14796472e-01 8.43391120e-01 -3.13943028e-01
8.32716525e-01 3.61881614e-01 1.52042425e+00 -4.69116569e-01
-7.04951435e-02 -1.12376288e-01 -9.65336189e-02 9.11810219e-01
4.69175875e-01 1.56268805e-01 4.62488770e-01 7.49376565e-02
1.13215387e+00 6.17355332e-02 -4.23774540e-01 -3.60034019e-01
-1.55147409e+00 6.26193345e-01 5.28535068e-01 6.27236843e-01
-7.82346308e-01 -1.61341757e-01 7.34502912e-01 2.47506052e-01
-3.14532928e-02 2.19187871e-01 -4.42798257e-01 4.72703397e-01
-1.20073414e+00 -2.18395606e-01 7.02228248e-01 7.28689492e-01
2.18756258e-01 1.90178320e-01 -4.57097232e-01 4.00836021e-01
4.06942129e-01 3.26243192e-01 9.84662831e-01 -5.36577106e-01
-2.85406232e-01 3.58481795e-01 -3.43663126e-01 -6.81510508e-01
-5.39011955e-01 -5.07599294e-01 -1.49105155e+00 3.41456980e-01
6.95051849e-01 -3.38605970e-01 -1.35543263e+00 1.03249598e+00
5.09812355e-01 1.30034745e-01 9.12056863e-03 1.14392984e+00
9.96645510e-01 3.70353192e-01 4.83096451e-01 -3.94664407e-01
1.71316886e+00 -8.72645974e-01 -6.30356431e-01 4.92444009e-01
4.75191236e-01 -1.24261391e+00 5.16893089e-01 1.70098260e-01
-1.26186264e+00 -2.73034513e-01 -1.08182085e+00 3.53943914e-01
-3.47046882e-01 1.32144121e-02 5.54023445e-01 7.67795205e-01
-1.00476944e+00 7.32923031e-01 -9.91073310e-01 -3.10853004e-01
4.28800434e-01 6.63125217e-01 -4.80326861e-01 -1.39925592e-02
-9.00816441e-01 1.13443124e+00 5.86541653e-01 -2.70429522e-01
-7.02520311e-01 -8.48268569e-01 -8.42116356e-01 1.26170486e-01
-1.87895209e-01 -7.34117508e-01 1.03909206e+00 -6.20549202e-01
-1.26354170e+00 9.03131902e-01 3.89716417e-01 -6.13319993e-01
4.74500537e-01 6.60447538e-01 9.27322358e-02 3.40784878e-01
-3.48016590e-01 5.80120802e-01 2.21260652e-01 -8.07171047e-01
-4.58163500e-01 -5.36732852e-01 -6.40582919e-01 3.86072785e-01
4.51437235e-01 4.88412857e-01 8.07947442e-02 -6.77590847e-01
3.87013584e-01 -4.75861967e-01 -3.67710590e-01 -1.85969278e-01
-1.59617364e-01 9.78338942e-02 5.95847309e-01 -1.39313352e+00
9.72397149e-01 -1.75627768e+00 -3.17085624e-01 5.64664543e-01
2.03250542e-01 1.58031404e-01 4.36446428e-01 -4.78393078e-01
-3.23728502e-01 1.89215377e-01 -2.98248857e-01 6.10932291e-01
-2.44413868e-01 -9.17328447e-02 7.87552893e-01 1.12930548e+00
-3.58948261e-01 8.18044960e-01 -8.16700101e-01 -9.87547934e-01
6.67495787e-01 5.11651635e-01 -2.15006456e-01 1.22783817e-01
6.88007236e-01 8.09663177e-01 -7.37614855e-02 8.08824897e-01
9.68032360e-01 -9.68440529e-03 -1.50174974e-02 -6.07882380e-01
-5.18813968e-01 -3.71744603e-01 -1.01392365e+00 1.21685386e+00
-8.00973624e-02 4.05599445e-01 3.08174253e-01 -9.31534290e-01
6.49193048e-01 7.61654973e-01 9.18921888e-01 -8.20170760e-01
7.39850223e-01 2.87015438e-01 6.52647197e-01 -7.26847410e-01
-2.56901890e-01 -3.82596672e-01 4.93456721e-01 5.83610713e-01
1.52477413e-01 -2.09610939e-01 4.36797142e-01 -3.05316985e-01
6.89019382e-01 -5.06538570e-01 8.91997218e-01 -8.68906260e-01
8.80639851e-01 -2.26565897e-02 2.76375294e-01 2.83321530e-01
-9.45717573e-01 4.30629283e-01 2.01926142e-01 -6.60203159e-01
-1.44065905e+00 -8.94983351e-01 -4.00006413e-01 5.61391473e-01
-7.05342069e-02 6.26316607e-01 -1.04133177e+00 -5.41508257e-01
-1.49395153e-01 2.07171872e-01 -4.56165373e-01 1.24355018e-01
-7.29925096e-01 -1.35533023e+00 3.98597598e-01 3.70612025e-01
8.65773857e-01 -8.46131682e-01 -1.20720613e+00 1.36560574e-01
2.26533100e-01 -4.36675459e-01 -6.83207989e-01 1.03835918e-01
-1.29740667e+00 -1.20564246e+00 -1.20554388e+00 -1.06488729e+00
1.01684403e+00 6.26932457e-02 1.29365635e+00 4.20613229e-01
-7.29186118e-01 6.42445385e-02 -1.96939737e-01 -2.50997692e-01
-3.59923273e-01 -3.83077800e-01 -2.78233796e-01 -6.38888955e-01
3.39902461e-01 -4.00914460e-01 -1.22520137e+00 1.20074935e-01
-1.07248425e+00 -7.49187618e-02 1.29832745e+00 8.61405134e-01
6.94209933e-01 3.51648331e-01 7.48769268e-02 -4.80148345e-01
6.40810847e-01 -3.27135503e-01 -5.20552993e-01 3.54803592e-01
-7.49285996e-01 -3.13741624e-01 2.97699422e-01 -3.72139513e-01
-7.12136209e-01 1.13497218e-02 2.26682201e-02 -9.72867459e-02
-3.23785096e-01 4.62713003e-01 3.94022793e-01 -6.11926556e-01
6.04313314e-01 4.36460167e-01 3.38681847e-01 -3.63010308e-03
-7.58418441e-02 5.52060425e-01 5.79477727e-01 5.95128946e-02
4.99389291e-01 3.37897569e-01 1.80805653e-01 -4.82115686e-01
9.65352878e-02 -6.27248526e-01 -1.00800860e+00 -2.49928370e-01
1.25480998e+00 -6.02763593e-01 -4.16576743e-01 4.81877655e-01
-8.08186471e-01 -2.23807424e-01 2.54185982e-02 1.17935085e+00
-8.40014964e-02 6.35692298e-01 -1.00243950e+00 -4.17278856e-01
-1.23047197e+00 -1.46064401e+00 -2.56512258e-02 7.32072234e-01
8.18705708e-02 -1.05094588e+00 -2.41685629e-01 7.92870671e-02
1.06762540e+00 6.91827655e-01 1.09988475e+00 -8.75880063e-01
-3.98837894e-01 -1.54808238e-01 -6.10561788e-01 4.06509757e-01
1.77626461e-01 -1.75463498e-01 -3.03322494e-01 -2.90250272e-01
2.61025190e-01 1.93989411e-01 2.97854632e-01 1.18873107e+00
1.06948876e+00 -4.35563654e-01 2.38000721e-01 8.22756946e-01
1.63853562e+00 8.72632205e-01 5.14743030e-01 2.87591308e-01
6.13490269e-02 2.61144519e-01 1.56281173e-01 5.86578310e-01
8.22175741e-02 -2.64089197e-01 4.10069734e-01 -5.85336268e-01
-4.31810945e-01 6.00063562e-01 -3.69077325e-01 1.07141066e+00
-1.64655864e-01 3.10981125e-01 -1.18842113e+00 5.70540488e-01
-8.90741765e-01 -6.80641115e-01 -1.25249028e-01 2.20408154e+00
8.43566835e-01 -3.06564540e-01 -3.63904350e-02 1.00947216e-01
8.79946232e-01 -5.82898200e-01 -3.77753347e-01 -1.59578115e-01
1.93135127e-01 5.24416625e-01 1.08516848e+00 6.08720243e-01
-1.23739922e+00 1.85179710e-01 5.76113462e+00 4.25501794e-01
-1.31258392e+00 3.13140064e-01 1.20719802e+00 4.89211470e-01
2.59088427e-01 -2.85832673e-01 8.91700760e-02 4.59501207e-01
5.54068267e-01 -4.13574159e-01 5.64528823e-01 4.66969281e-01
3.45468819e-01 -6.24136806e-01 -6.75014853e-01 8.84331226e-01
8.08138698e-02 -9.67464507e-01 -2.09566191e-01 -1.25052452e-01
7.32082844e-01 1.20677315e-01 -1.05007738e-01 1.72657549e-01
1.01165444e-01 -1.18752432e+00 6.99974149e-02 3.39579225e-01
7.54470408e-01 -7.32120514e-01 1.45800042e+00 2.96486448e-02
-9.77107167e-01 2.64536917e-01 -2.76289105e-01 2.42665753e-01
-1.74942702e-01 5.69844007e-01 -1.19274318e+00 3.47875923e-01
6.56877279e-01 -5.98861836e-02 -5.06558597e-01 1.82264626e+00
2.63899118e-01 2.74695396e-01 -3.51202607e-01 1.33741409e-01
2.81620800e-01 -4.74506170e-01 9.14354622e-02 1.54011142e+00
5.04021227e-01 5.74855149e-01 2.29555368e-01 5.59259295e-01
2.59704366e-02 7.25367665e-01 -1.05450042e-01 4.16971475e-01
2.81258970e-01 1.48410249e+00 -1.60140967e+00 -7.82107830e-01
-2.82711595e-01 4.44184363e-01 -7.12774932e-01 3.84718142e-02
-7.04970658e-01 -5.01629889e-01 -2.22666353e-01 6.14694729e-02
-2.66603708e-01 -1.23218656e-01 -7.40274727e-01 -8.21685374e-01
-6.35486901e-01 -7.54576504e-01 6.56899571e-01 -3.70674670e-01
-9.74011540e-01 5.29888332e-01 1.66847840e-01 -8.43673408e-01
1.53679371e-01 -4.30800676e-01 -1.00351822e+00 1.18345892e+00
-1.63229465e+00 -8.62229705e-01 -5.23979306e-01 6.41318798e-01
3.88530642e-01 -3.01534291e-02 3.28352809e-01 6.89180076e-01
-1.15027316e-01 2.39789814e-01 1.36229098e-01 5.71048617e-01
4.66960132e-01 -1.57556975e+00 -4.73083556e-01 9.32970941e-01
-7.09636092e-01 4.50078666e-01 7.21410096e-01 -6.24733508e-01
-1.26458454e+00 -8.93343508e-01 5.14494896e-01 3.60290766e-01
2.10419789e-01 4.46231812e-01 -6.38308585e-01 6.47689044e-01
8.52100492e-01 4.89552081e-01 1.03823721e+00 -9.38472748e-01
4.80924338e-01 2.33270973e-01 -1.98000205e+00 2.43169576e-01
-1.19071759e-01 3.17446023e-01 -5.14834166e-01 2.86778718e-01
-6.06088899e-02 -7.45235145e-01 -1.46227682e+00 6.07414067e-01
6.63036883e-01 -1.00644624e+00 1.16822982e+00 -3.85079324e-01
1.91699818e-01 -3.48194003e-01 2.47259498e-01 -1.30542862e+00
-5.75887144e-01 -4.34729010e-02 1.64091438e-01 5.90764403e-01
2.07270339e-01 -3.23054552e-01 8.32198620e-01 6.36196792e-01
-2.03122318e-01 -8.01882863e-01 -7.45269477e-01 -1.10490136e-01
3.95817429e-01 5.21942854e-01 5.24788737e-01 1.25881696e+00
-1.96236417e-01 -2.20807835e-01 6.31458014e-02 2.58734941e-01
9.88429189e-01 -6.97881281e-02 2.40974426e-01 -1.07296181e+00
2.11866662e-01 -7.18191862e-01 -4.85566556e-01 -2.85230577e-01
-5.56666970e-01 -8.70439291e-01 -2.55945295e-01 -1.69963729e+00
5.05719066e-01 -4.74010676e-01 -3.51033568e-01 2.87885368e-01
4.43042740e-02 4.12331969e-01 1.02075011e-01 2.16426194e-01
-1.30355999e-01 -7.94125497e-02 1.85498917e+00 -1.90867782e-01
-1.80287108e-01 -4.62975875e-02 -1.41895667e-01 5.97778320e-01
1.16118252e+00 -1.69962913e-01 -1.30192652e-01 -3.03105175e-01
-5.71504056e-01 4.30741012e-01 -2.48740334e-02 -1.12942469e+00
6.21698916e-01 -6.34804890e-02 1.33768642e+00 -7.69571662e-01
-6.20498002e-01 -9.55986977e-01 3.13393712e-01 1.30406642e+00
-1.24881402e-01 5.36758184e-01 7.81031176e-02 -2.79155195e-01
-2.06759736e-01 -5.76569319e-01 1.48318219e+00 -8.90720308e-01
-4.80141908e-01 4.00489569e-01 -5.54232121e-01 -3.29616666e-01
1.32555485e+00 -3.89009714e-01 1.88793659e-01 -5.85216507e-02
-8.67056489e-01 1.36039197e-01 1.10473432e-01 -5.47100782e-01
6.00745380e-01 -1.33919382e+00 -9.22151089e-01 4.07486469e-01
-6.00951374e-01 9.54705030e-02 3.43618244e-01 1.75315034e+00
-1.43855131e+00 4.35300052e-01 -7.62424648e-01 -4.64599401e-01
-1.54897714e+00 4.72944081e-01 7.64236450e-01 -5.09275436e-01
-5.20965517e-01 7.48123109e-01 -1.36901572e-01 -3.49183083e-01
4.68316376e-02 -7.42657900e-01 -3.81949902e-01 -2.54431635e-01
4.99958277e-01 5.56218922e-01 2.26074040e-01 -6.48927748e-01
-1.44789696e-01 4.70050961e-01 1.05456337e-01 2.40939870e-01
9.95612979e-01 -1.88294277e-01 -6.49900615e-01 -3.57748181e-01
9.64408100e-01 -2.23342061e-01 -7.52971053e-01 -1.03074744e-01
6.69118240e-02 -5.65952778e-01 5.59689224e-01 -1.18852484e+00
-1.38786197e+00 7.65433967e-01 1.18527985e+00 3.08886766e-01
1.22507846e+00 -4.77396667e-01 7.10170269e-01 -1.94645777e-01
-1.32464662e-01 -7.68903255e-01 -4.51691747e-01 3.57155085e-01
3.88941288e-01 -1.28179789e+00 3.36829752e-01 -1.62193999e-01
-3.06037515e-01 1.41816974e+00 4.52405602e-01 -1.11481465e-01
8.58930707e-01 5.76364696e-01 2.98305660e-01 -7.16111213e-02
-1.85741320e-01 7.80942664e-02 2.17938006e-01 4.25061196e-01
9.24163759e-01 1.45191669e-01 -7.03318238e-01 1.71741635e-01
3.34674530e-02 1.85273305e-01 5.53403437e-01 9.84801471e-01
-8.16789210e-01 -3.46397877e-01 -8.49442601e-01 8.85498047e-01
-1.18069804e+00 -2.37957090e-01 2.66907901e-01 9.61685896e-01
2.04691887e-01 3.13347548e-01 6.67181760e-02 2.36978754e-01
-6.50748238e-03 -2.29254942e-02 6.28544211e-01 -9.32532921e-02
-1.00508475e+00 3.09672952e-01 -7.09160984e-01 1.61043689e-01
-3.95028204e-01 -4.80246753e-01 -1.36390972e+00 -5.79408705e-01
-2.92615235e-01 4.42132056e-01 9.72593606e-01 6.47191226e-01
-1.76086754e-01 5.07143557e-01 3.91015083e-01 -7.41813421e-01
-7.16157734e-01 -1.28543222e+00 -6.80395186e-01 4.12080616e-01
2.46722192e-01 -3.42695177e-01 -3.56938511e-01 3.42206478e-01] | [14.405684471130371, -2.7746798992156982] |
d19b4d76-947c-4ed1-a934-c217dcd75435 | attention-based-relational-graph | null | null | https://aclanthology.org/2021.eacl-main.170 | https://aclanthology.org/2021.eacl-main.170.pdf | Attention-based Relational Graph Convolutional Network for Target-Oriented Opinion Words Extraction | Target-oriented opinion words extraction (TOWE) is a subtask of aspect-based sentiment analysis (ABSA). It aims to extract the corresponding opinion words for a given opinion target in a review sentence. Intuitively, the relation between an opinion target and an opinion word mostly relies on syntactics. In this study, we design a directed syntactic dependency graph based on a dependency tree to establish a path from the target to candidate opinions. Subsequently, we propose a novel attention-based relational graph convolutional neural network (ARGCN) to exploit syntactic information over dependency graphs. Moreover, to explicitly extract the corresponding opinion words toward the given opinion target, we effectively encode target information in our model with the target-aware representation. Empirical results demonstrate that our model significantly outperforms all of the existing models on four benchmark datasets. Extensive analysis also demonstrates the effectiveness of each component of our models. Our code is available at https://github.com/wcwowwwww/towe-eacl. | ['Akiko Aizawa', 'An Wang', 'Junfeng Jiang'] | 2021-04-01 | null | null | null | eacl-2021-2 | ['target-oriented-opinion-words-extraction'] | ['natural-language-processing'] | [-1.75761972e-02 1.56453475e-01 -4.12075251e-01 -7.02201486e-01
-6.52568579e-01 -6.06977582e-01 4.60686445e-01 3.02554011e-01
-6.77789748e-02 3.93416613e-01 5.53790152e-01 -5.83972812e-01
7.59248063e-02 -1.02089310e+00 -4.54416364e-01 -3.09379011e-01
2.53683269e-01 6.10928945e-02 -4.52697538e-02 -5.24494112e-01
4.60169941e-01 7.99611881e-02 -1.06139970e+00 5.28875709e-01
6.38139844e-01 1.05956078e+00 -3.61492902e-01 3.79318625e-01
-5.56926608e-01 1.12339079e+00 -6.23420596e-01 -9.93907034e-01
-2.33057309e-02 -4.92487907e-01 -8.43353570e-01 1.58928856e-01
1.05666079e-01 9.39154532e-03 -1.51770875e-01 1.18576491e+00
1.68389499e-01 -5.64282238e-02 5.76408088e-01 -1.16793871e+00
-1.24282706e+00 9.80095029e-01 -7.51840234e-01 2.23044336e-01
2.67437935e-01 -1.37265608e-01 1.75422525e+00 -1.22722530e+00
5.41074991e-01 1.11231041e+00 3.25428694e-01 2.04809099e-01
-4.60835814e-01 -5.19269586e-01 9.35839117e-01 1.62591666e-01
-8.15456450e-01 -3.34389746e-01 1.11666358e+00 -4.56735194e-02
1.03047740e+00 5.88459633e-02 9.96748447e-01 7.45033383e-01
5.26812911e-01 1.10305977e+00 9.91232038e-01 -2.97243893e-01
2.10585773e-01 6.93762302e-02 1.01038527e+00 8.92363846e-01
5.18843114e-01 -4.23050523e-01 -8.92408550e-01 -1.06888479e-02
-1.21443078e-01 9.97595396e-03 -3.17144156e-01 -2.89625436e-01
-6.37627184e-01 1.00388741e+00 5.06826818e-01 2.20467553e-01
-3.61981332e-01 2.20304966e-01 3.78469765e-01 3.46525371e-01
1.03077853e+00 3.93590778e-01 -9.56958473e-01 2.82032043e-01
-2.05740273e-01 -9.01559070e-02 1.03716612e+00 1.05228591e+00
8.02003622e-01 -6.10606652e-03 -2.98763633e-01 4.82692391e-01
7.91933537e-01 6.25251532e-01 3.08189929e-01 -2.95660347e-01
4.69806165e-01 1.25038707e+00 -2.26413354e-01 -1.30041802e+00
-2.80996054e-01 -4.48753685e-01 -5.66658854e-01 -1.92069024e-01
-2.40059838e-01 -3.86056423e-01 -1.04181063e+00 1.24320590e+00
4.74142283e-01 -1.84305891e-01 2.63982773e-01 7.97606170e-01
1.28510845e+00 6.66216016e-01 2.72938889e-02 -2.35803619e-01
1.52332234e+00 -1.36688995e+00 -8.98357749e-01 -7.37043858e-01
7.41664648e-01 -6.99563324e-01 8.93981278e-01 6.81810081e-02
-6.16476774e-01 -1.94516450e-01 -1.07249320e+00 -1.77865058e-01
-6.65635943e-01 1.03347838e-01 9.41490412e-01 3.84092510e-01
-9.12585020e-01 1.65135011e-01 -5.36912680e-01 -8.54436010e-02
6.12555861e-01 4.24605131e-01 -2.31727675e-01 -1.77627683e-01
-1.29862249e+00 6.28132343e-01 5.55338664e-03 4.66546357e-01
-2.97217935e-01 -5.21987438e-01 -1.28520441e+00 -5.70007227e-02
5.02487481e-01 -7.56805182e-01 1.21058297e+00 -1.38013351e+00
-1.36322439e+00 7.31253326e-01 -6.53547525e-01 -2.24467635e-01
-4.54056352e-01 -4.27965432e-01 -6.06073976e-01 2.86108255e-02
2.27601260e-01 2.33411223e-01 7.71844566e-01 -1.29598176e+00
-5.90811789e-01 -5.49361408e-01 6.76297545e-01 1.60637125e-01
-5.28284550e-01 2.75246918e-01 -8.18692386e-01 -5.41134417e-01
-2.36212686e-01 -6.45341456e-01 -3.85517955e-01 -4.62180406e-01
-6.48574293e-01 -4.72130924e-01 5.60098290e-01 -3.70045483e-01
1.54909647e+00 -1.89581180e+00 1.88328445e-01 1.99547321e-01
5.88333130e-01 1.83988184e-01 -4.04645324e-01 4.91234004e-01
-1.91589490e-01 3.84270966e-01 -3.84826988e-01 -2.22809136e-01
1.87371559e-02 -4.90310192e-02 -5.37471771e-01 1.54840916e-01
4.87735540e-01 1.45077908e+00 -1.03393018e+00 -3.69177759e-01
-2.48623699e-01 4.86439973e-01 -2.96861708e-01 9.89057124e-02
-2.91532010e-01 -1.43481031e-01 -9.33783591e-01 8.29117358e-01
6.48529530e-01 -5.21213353e-01 3.78556788e-01 -4.60503489e-01
2.02615932e-01 6.76191986e-01 -5.45789480e-01 1.34553838e+00
-6.68548882e-01 4.67235535e-01 -4.27100472e-02 -6.91302955e-01
1.04380929e+00 -1.70166977e-02 1.67479903e-01 -6.58944666e-01
4.68370497e-01 6.95057213e-02 -8.46569892e-03 -3.83727759e-01
6.34439707e-01 -8.87299925e-02 -1.76236987e-01 5.63671470e-01
1.33455142e-01 -4.51368056e-02 4.46248770e-01 5.18190444e-01
1.02391541e+00 6.63929060e-02 5.39239645e-01 -2.96004027e-01
7.86006391e-01 -5.03354669e-02 6.82520092e-01 2.90837258e-01
-2.28110095e-03 3.67262036e-01 9.59069848e-01 -6.55398428e-01
-3.14676672e-01 -5.17006636e-01 3.13766003e-01 9.09078956e-01
2.08250016e-01 -1.00337064e+00 -2.07345441e-01 -1.37881339e+00
-1.95604891e-01 7.45353580e-01 -9.83185172e-01 -1.47758320e-01
-5.55365324e-01 -7.82872379e-01 -2.59941190e-01 5.90148270e-01
1.78837746e-01 -1.33973384e+00 -6.35621771e-02 -4.91315238e-02
5.31992875e-02 -9.28422809e-01 -6.71092749e-01 1.33255288e-01
-5.94815314e-01 -1.35628891e+00 -2.66588092e-01 -8.45209718e-01
1.06418872e+00 3.90487671e-01 1.57834852e+00 4.02282596e-01
2.92113632e-01 4.00205761e-01 -8.76464069e-01 -6.81366146e-01
9.67804641e-02 3.01119894e-01 -4.44020212e-01 2.31481582e-01
1.13057196e+00 -3.89758438e-01 -7.55249798e-01 -1.72301739e-01
-8.59069645e-01 -1.40550837e-01 7.01152325e-01 5.61412930e-01
7.82751679e-01 -1.21243067e-01 5.68801284e-01 -1.48775339e+00
1.09833324e+00 -6.04242682e-01 -6.35450840e-01 3.03937405e-01
-7.38667309e-01 3.73193771e-02 5.80467284e-01 4.33623157e-02
-7.50277042e-01 -2.17927128e-01 -1.52946234e-01 -2.89410260e-02
2.30316013e-01 1.17393684e+00 -3.13992321e-01 3.96594346e-01
4.37671058e-02 1.03242341e-02 -4.81998771e-01 -1.53668541e-02
4.44299668e-01 6.09054446e-01 -2.56599873e-01 -2.61962444e-01
6.86847806e-01 4.37480301e-01 -9.04377177e-02 -4.47688669e-01
-1.78598619e+00 -4.42578256e-01 -5.39275825e-01 -1.52852744e-01
6.93994462e-01 -8.39891255e-01 -4.73450750e-01 2.57154107e-01
-1.31893611e+00 1.43026397e-01 -2.91977942e-01 8.67415741e-02
1.61946043e-01 2.47095346e-01 -4.95415866e-01 -8.75109971e-01
-9.87961054e-01 -1.11782205e+00 1.23692417e+00 3.59400719e-01
-1.84160650e-01 -1.25183475e+00 1.15316704e-01 4.30465847e-01
2.43401244e-01 -2.50145923e-02 8.45012784e-01 -9.49199378e-01
-3.59216541e-01 -4.32339609e-01 -1.59249663e-01 4.07015264e-01
4.15120870e-01 1.49284631e-01 -6.94830537e-01 -8.21878612e-02
-9.27744806e-03 -2.97566891e-01 1.18626344e+00 1.70528367e-01
7.90154755e-01 -3.66492659e-01 -1.71624675e-01 3.42562646e-01
1.26981294e+00 4.16561663e-02 3.65373284e-01 3.14184755e-01
9.96462345e-01 7.94902623e-01 6.98870599e-01 3.30995061e-02
7.68982410e-01 1.04453854e-01 4.01737988e-01 -8.36112443e-03
-7.05304295e-02 -1.36706680e-01 6.48109019e-01 1.61999691e+00
1.75387561e-01 -6.19072676e-01 -6.65396690e-01 8.43683541e-01
-1.79415131e+00 -3.61179531e-01 -4.20904309e-01 1.56707287e+00
5.91594577e-01 4.21811223e-01 -4.06294823e-01 -2.63062805e-01
4.86916482e-01 6.91094160e-01 -5.99482417e-01 -6.38639987e-01
-8.59466493e-02 5.38214147e-01 1.36512846e-01 7.37653136e-01
-1.09742868e+00 1.27371955e+00 4.92920208e+00 4.88110185e-01
-8.03062320e-01 2.17412449e-02 7.25505292e-01 1.52202189e-01
-8.91858459e-01 2.10352436e-01 -8.80733073e-01 -2.17308998e-02
8.87565374e-01 -4.65076506e-01 -1.19517013e-01 9.53963697e-01
-9.11833812e-03 2.08530337e-01 -7.63132870e-01 3.16431701e-01
5.29593050e-01 -1.22208607e+00 5.26746571e-01 -1.02201901e-01
7.12783158e-01 -4.67028189e-03 -6.32630195e-04 4.13489670e-01
3.99060845e-01 -7.61917472e-01 3.94519895e-01 4.45643544e-01
3.32920760e-01 -1.10546935e+00 1.16640222e+00 -4.09508467e-01
-1.62901664e+00 1.99948430e-01 -4.03443724e-01 -6.46758974e-02
2.20155120e-01 9.13303375e-01 -1.63981527e-01 8.58995259e-01
5.80921412e-01 1.35156691e+00 -6.76307738e-01 4.32832927e-01
-1.18643010e+00 6.88467979e-01 2.81426311e-01 -5.10102212e-01
2.92790353e-01 -3.20053369e-01 3.68533105e-01 1.06043124e+00
3.33116502e-02 7.61407241e-02 -8.11883435e-02 5.83314061e-01
-5.33430338e-01 4.78274554e-01 -7.45489836e-01 -2.93692946e-01
1.17616303e-01 1.79322577e+00 -7.13504434e-01 -3.11431974e-01
-7.72912681e-01 8.70065928e-01 6.67379379e-01 4.52968776e-01
-6.25888705e-01 -6.13824725e-01 8.03945899e-01 -3.00158352e-01
7.75321841e-01 1.61839707e-03 -1.63665280e-01 -1.46976411e+00
3.29914063e-01 -1.00895250e+00 3.04662466e-01 -8.86455655e-01
-1.54226673e+00 1.09247363e+00 -4.43684995e-01 -1.14884937e+00
1.70108512e-01 -9.23135817e-01 -9.40530539e-01 7.59918213e-01
-1.91456676e+00 -1.49483943e+00 2.20330171e-02 3.52045149e-01
7.42803097e-01 5.40304417e-03 8.28058541e-01 -1.59379631e-01
-7.51825094e-01 3.36952448e-01 -7.11040199e-01 4.60986972e-01
3.83536041e-01 -1.24196422e+00 7.66267002e-01 1.14479446e+00
3.49052995e-01 9.95247364e-01 1.76724002e-01 -7.78499365e-01
-1.69803953e+00 -1.22394216e+00 1.33211458e+00 -6.97574794e-01
1.18690789e+00 -3.06375504e-01 -6.85271919e-01 9.18563545e-01
7.94833481e-01 -2.66323034e-02 1.02369189e+00 4.37436789e-01
-6.99561596e-01 -2.94279248e-01 -2.88432151e-01 7.32283771e-01
7.92158127e-01 -5.19789338e-01 -7.78808415e-01 4.17320371e-01
1.29164660e+00 -7.74667710e-02 -6.19608521e-01 5.26071846e-01
3.81419927e-01 -8.35019171e-01 4.36560631e-01 -9.02673006e-01
9.95700955e-01 -4.97276932e-01 -2.20170300e-02 -1.48906446e+00
-1.36678025e-01 -2.40471035e-01 -5.76600790e-01 1.33756399e+00
1.22930527e+00 -5.42794228e-01 5.42631507e-01 5.65719426e-01
-1.25684500e-01 -1.30212677e+00 -4.21637714e-01 -9.41175297e-02
1.32992351e-03 -5.70776701e-01 7.66597450e-01 7.36927152e-01
1.49706364e-01 1.11459792e+00 -1.43144578e-01 3.74438018e-01
2.91334242e-01 7.32272923e-01 4.59286690e-01 -9.46345389e-01
-1.96988106e-01 -4.44954246e-01 -1.10874780e-01 -1.03912652e+00
5.50878823e-01 -8.75118077e-01 -1.41742975e-01 -2.20816302e+00
2.28669122e-01 1.02938205e-01 -6.78735673e-01 5.19587100e-01
-6.88794613e-01 1.15549192e-01 6.08213916e-02 -1.23212710e-01
-1.03869045e+00 7.59639263e-01 1.41096079e+00 -5.21227658e-01
1.69741660e-02 8.10361430e-02 -1.50010598e+00 8.18183780e-01
9.83148217e-01 -6.09389782e-01 -5.48133731e-01 -6.21674776e-01
1.12935603e+00 -3.94590408e-01 -3.33211839e-01 -5.78224808e-02
3.45909417e-01 1.18792811e-02 3.95125598e-02 -5.70686936e-01
3.28737721e-02 -6.85690582e-01 -7.17406154e-01 2.14229450e-01
-3.07707250e-01 5.38148999e-01 8.44602883e-02 5.90094686e-01
-5.58162391e-01 -1.08093731e-01 1.12088569e-01 -8.48086849e-02
-7.67000079e-01 4.50323761e-01 -2.78373241e-01 1.40697882e-01
7.06734061e-01 5.71385026e-01 -6.17425680e-01 -4.84699488e-01
-2.98574924e-01 4.68059033e-01 9.81029049e-02 7.02648818e-01
1.01900601e+00 -1.15014386e+00 -6.20426595e-01 -4.33982871e-02
5.48848569e-01 3.97220217e-02 -2.79865414e-02 7.52127528e-01
-2.23167539e-01 3.61584753e-01 4.74983364e-01 1.31656379e-01
-1.22043180e+00 6.91736937e-01 2.71832347e-01 -6.19544446e-01
-2.94646263e-01 1.01295865e+00 2.05459669e-01 -6.42188668e-01
-2.21339941e-01 -3.67682070e-01 -8.86071026e-01 1.19731449e-01
6.54455483e-01 -2.32448235e-01 1.13406457e-01 -7.55406141e-01
-6.16498172e-01 6.86138809e-01 -4.78833586e-01 2.14249149e-01
1.35427058e+00 -4.61380482e-02 -9.01985645e-01 3.95023465e-01
1.17991579e+00 5.69733739e-01 -5.76428771e-01 -8.74920711e-02
9.55221206e-02 -8.06150585e-02 1.81000769e-01 -6.36728764e-01
-1.59188569e+00 5.97895622e-01 -2.59780556e-01 2.78037250e-01
1.25638890e+00 1.87990203e-01 7.59036481e-01 5.50235331e-01
6.01657443e-02 -6.73409879e-01 4.77161109e-02 7.42103040e-01
9.64439750e-01 -1.32774687e+00 2.83527255e-01 -6.26215279e-01
-9.06209707e-01 1.06647801e+00 8.29425395e-01 -1.83196202e-01
1.12465155e+00 2.28127345e-01 5.23155630e-01 -8.57202530e-01
-1.02204108e+00 -5.46554625e-01 5.84706545e-01 3.76766115e-01
7.75090277e-01 -7.30709806e-02 -5.39976120e-01 8.51693511e-01
-1.47238865e-01 -3.25552732e-01 4.71458346e-01 1.02731764e+00
-1.87780723e-01 -1.35156190e+00 3.26668322e-01 4.82571334e-01
-5.81499577e-01 -8.25615168e-01 -1.02137804e+00 5.24646878e-01
-1.89509109e-01 1.44594049e+00 -1.36332229e-01 -5.88617921e-01
5.66410661e-01 -6.70415387e-02 -7.01924711e-02 -8.92789245e-01
-9.38434720e-01 -9.69906077e-02 4.25141126e-01 -6.31905258e-01
-8.72396410e-01 -2.10633606e-01 -1.24228013e+00 -2.34616809e-02
-4.08094317e-01 3.96550894e-01 5.23909390e-01 1.02777588e+00
7.43820548e-01 8.92511308e-01 7.15685785e-01 -9.95948315e-02
2.08867267e-01 -8.95158231e-01 -4.02626604e-01 1.50379553e-01
2.68241465e-01 -1.95086837e-01 -4.13184166e-01 -2.18180880e-01] | [11.509889602661133, 6.626028060913086] |
bb97ff3a-ef53-48f0-ab65-3b12e7a93d38 | multi-source-fusion-and-automatic-predictor | 2108.05076 | null | https://arxiv.org/abs/2108.05076v1 | https://arxiv.org/pdf/2108.05076v1.pdf | Multi-Source Fusion and Automatic Predictor Selection for Zero-Shot Video Object Segmentation | Location and appearance are the key cues for video object segmentation. Many sources such as RGB, depth, optical flow and static saliency can provide useful information about the objects. However, existing approaches only utilize the RGB or RGB and optical flow. In this paper, we propose a novel multi-source fusion network for zero-shot video object segmentation. With the help of interoceptive spatial attention module (ISAM), spatial importance of each source is highlighted. Furthermore, we design a feature purification module (FPM) to filter the inter-source incompatible features. By the ISAM and FPM, the multi-source features are effectively fused. In addition, we put forward an automatic predictor selection network (APS) to select the better prediction of either the static saliency predictor or the moving object predictor in order to prevent over-reliance on the failed results caused by low-quality optical flow maps. Extensive experiments on three challenging public benchmarks (i.e. DAVIS$_{16}$, Youtube-Objects and FBMS) show that the proposed model achieves compelling performance against the state-of-the-arts. The source code will be publicly available at \textcolor{red}{\url{https://github.com/Xiaoqi-Zhao-DLUT/Multi-Source-APS-ZVOS}}. | ['Huchuan Lu', 'Lihe Zhang', 'Jiaxing Yang', 'Youwei Pang', 'Xiaoqi Zhao'] | 2021-08-11 | null | null | null | null | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [-5.46218194e-02 -4.04443234e-01 -2.62607306e-01 -2.86430538e-01
-5.25126398e-01 -2.63204247e-01 1.62356626e-02 -1.81213990e-01
-2.47533351e-01 6.90714180e-01 8.22355673e-02 1.50325950e-02
-1.12506784e-01 -5.58191061e-01 -5.55789769e-01 -7.43807018e-01
1.30955845e-01 -2.50352979e-01 7.38466144e-01 -6.38892725e-02
5.09539962e-01 2.02597246e-01 -1.68904853e+00 2.67112404e-01
1.02054715e+00 1.42101753e+00 6.98180258e-01 4.95351553e-01
-1.42715350e-01 9.40551758e-01 -1.75946683e-01 -2.04943255e-01
3.70853782e-01 -3.63989413e-01 -6.16128147e-01 2.42309161e-02
4.34044957e-01 -6.18024230e-01 -5.72503865e-01 1.26696634e+00
4.21134979e-01 2.96743870e-01 7.67754465e-02 -1.62426889e+00
-4.22297299e-01 1.51721776e-01 -7.30280638e-01 9.20209169e-01
3.23348880e-01 5.73184550e-01 8.12726080e-01 -1.03277874e+00
6.12523437e-01 1.15101469e+00 1.17586836e-01 3.86575967e-01
-6.63710833e-01 -8.97014678e-01 3.73736560e-01 6.24456108e-01
-1.34833431e+00 -5.74022353e-01 9.32322204e-01 -3.64758581e-01
3.67518604e-01 3.62027913e-01 6.30604923e-01 7.65548646e-01
3.65710743e-02 1.10915530e+00 9.41713870e-01 1.32749021e-01
2.41041910e-02 7.59677663e-02 1.09822534e-01 1.05122161e+00
1.71682864e-01 4.50444631e-02 -8.55859458e-01 8.57660919e-02
1.01943362e+00 2.91577995e-01 -6.50915623e-01 -1.80069774e-01
-1.28777170e+00 3.53469342e-01 6.85539067e-01 1.21443853e-01
-3.76538754e-01 1.53701872e-01 1.07111037e-01 -8.89788419e-02
3.00150812e-01 9.72966403e-02 -4.16652977e-01 -2.50764817e-01
-7.83118904e-01 3.72429118e-02 2.90682465e-01 1.02258205e+00
9.65175211e-01 3.21343318e-02 -3.55165392e-01 6.80137515e-01
5.31514525e-01 5.48998237e-01 3.11169356e-01 -1.40707612e+00
4.89551157e-01 6.46408975e-01 2.49903515e-01 -1.24648309e+00
-2.02360258e-01 -1.88741833e-01 -5.47632515e-01 7.12892413e-02
3.10562730e-01 -7.32921436e-02 -8.86339188e-01 1.38730025e+00
6.33593321e-01 7.81376839e-01 -2.08798885e-01 1.61973357e+00
1.19104397e+00 6.76824272e-01 7.43508935e-02 -1.52411237e-01
1.04647791e+00 -1.18453276e+00 -5.52907526e-01 -2.30697438e-01
1.44675240e-01 -7.51729667e-01 9.59443808e-01 2.34420165e-01
-1.09831762e+00 -6.63266480e-01 -9.11090553e-01 -1.33619487e-01
-1.17074251e-01 9.40598249e-02 6.15208089e-01 2.15840638e-01
-7.99142540e-01 5.91192126e-01 -8.76265645e-01 -1.37848958e-01
8.32216442e-01 3.50620240e-01 -1.37106359e-01 -2.76525140e-01
-1.04895377e+00 4.17750090e-01 2.28631660e-01 1.39358461e-01
-9.90512490e-01 -6.00668192e-01 -5.68419755e-01 -1.49943009e-01
5.88072479e-01 -4.95681256e-01 9.59792674e-01 -1.21475410e+00
-1.33374906e+00 3.41288447e-01 -4.42250550e-01 3.15398388e-02
3.18702906e-01 -2.68499494e-01 -2.69943118e-01 5.75912833e-01
2.96081722e-01 9.26733613e-01 8.00170958e-01 -1.36299288e+00
-1.11540151e+00 -2.77866274e-01 1.84816465e-01 3.09427202e-01
-2.86862284e-01 1.31896898e-01 -9.05579984e-01 -4.27653879e-01
3.04686993e-01 -6.74034417e-01 -9.55840796e-02 1.92186221e-01
-5.25032341e-01 -6.42253235e-02 9.73877609e-01 -6.88802600e-01
1.25643981e+00 -2.15982556e+00 9.17117745e-02 -4.78569418e-02
3.26298326e-01 4.32189733e-01 -2.24196073e-02 -2.03799829e-01
2.39983857e-01 6.99042249e-03 -1.36627501e-03 -2.39866108e-01
-3.69932741e-01 1.22714136e-02 3.54016684e-02 5.91109633e-01
1.79162726e-01 9.56880569e-01 -1.11129785e+00 -8.39896202e-01
6.26146555e-01 5.00166357e-01 -4.13365424e-01 1.27766579e-01
-5.47528006e-02 6.96973085e-01 -8.75557303e-01 1.07805836e+00
7.00978756e-01 -4.24817085e-01 -4.51828271e-01 -3.03415358e-01
-2.80614257e-01 1.05994008e-01 -1.36204040e+00 1.87701654e+00
3.90393473e-02 4.95737255e-01 8.72525200e-02 -4.78254497e-01
6.40689671e-01 1.36738300e-01 7.69743800e-01 -7.63985157e-01
4.96914834e-01 2.43185505e-01 1.00131501e-02 -5.96429050e-01
3.57862949e-01 5.78159988e-01 3.28496039e-01 1.27814740e-01
2.20574606e-02 4.76723343e-01 3.48031849e-01 3.45067024e-01
8.15463006e-01 2.92452276e-01 -1.33657321e-01 -1.83540851e-01
6.93485975e-01 -8.02303553e-02 1.13088930e+00 4.31729853e-01
-7.98885465e-01 8.51752818e-01 3.79204988e-01 -3.02687317e-01
-6.30354643e-01 -1.03745949e+00 8.37827474e-02 9.50899780e-01
1.04704702e+00 -3.58685970e-01 -4.88412768e-01 -5.58335543e-01
-6.85569793e-02 3.97895634e-01 -5.63699663e-01 -1.02813952e-01
-3.92114609e-01 -3.95472586e-01 8.77205580e-02 4.91142035e-01
7.08780766e-01 -1.14643574e+00 -9.48594868e-01 1.11661747e-01
-5.22113919e-01 -1.11012542e+00 -5.24699032e-01 -2.73930192e-01
-7.09221721e-01 -1.09134102e+00 -7.22732723e-01 -4.92806166e-01
5.64598620e-01 7.84183145e-01 6.59459591e-01 2.53053308e-01
-1.60121039e-01 3.05715948e-01 -4.21783358e-01 -1.60300106e-01
3.41574550e-01 -1.60998657e-01 -1.11808717e-01 3.01512927e-01
2.69687772e-01 -4.57202524e-01 -1.21376491e+00 3.96986514e-01
-6.18896067e-01 3.19504559e-01 3.53390396e-01 4.08115000e-01
7.18728364e-01 -3.55965257e-01 3.47587556e-01 -3.64275604e-01
-7.07254792e-03 -6.25497758e-01 -4.69996125e-01 4.92744222e-02
-2.04865336e-01 -3.06823343e-01 2.77337462e-01 -2.89324909e-01
-1.01039886e+00 1.48049176e-01 1.06618486e-01 -9.96379197e-01
-2.16115817e-01 8.93424377e-02 -3.42167079e-01 -1.21924855e-01
1.82576463e-01 2.87860006e-01 -1.48405179e-01 -4.32871372e-01
4.05737497e-02 5.09905040e-01 4.41539794e-01 -3.19262207e-01
4.91391033e-01 6.94434822e-01 -1.84531927e-01 -5.74787498e-01
-9.05628204e-01 -6.69297457e-01 -4.52705920e-01 -6.38741791e-01
9.37229872e-01 -1.04443705e+00 -7.75228083e-01 4.39153582e-01
-1.12483490e+00 -1.75030187e-01 1.15120728e-02 4.67020929e-01
-4.34708685e-01 3.13780695e-01 -6.25342429e-01 -8.07126701e-01
-2.98096746e-01 -1.39533639e+00 1.00175691e+00 9.63115990e-01
2.41909787e-01 -5.18293917e-01 -4.38176155e-01 5.04257977e-01
3.51132929e-01 1.20807394e-01 1.56509057e-01 -2.53642082e-01
-1.17255700e+00 2.67533362e-01 -5.96673667e-01 1.67530656e-01
2.12514594e-01 2.38141894e-01 -1.01344836e+00 -5.28948680e-02
-1.81773826e-01 2.68409308e-02 1.05261052e+00 6.33214772e-01
1.25964952e+00 -6.53352737e-02 -4.01033044e-01 8.59185517e-01
1.43038726e+00 3.74166727e-01 5.80557108e-01 3.70788574e-01
1.12918866e+00 3.90974522e-01 1.04538631e+00 6.40985429e-01
6.33339703e-01 4.76196289e-01 7.65147984e-01 -8.40999261e-02
-2.87113219e-01 -5.87780364e-02 3.33730638e-01 4.30665791e-01
-2.55065352e-01 -1.86364025e-01 -7.61923611e-01 5.89773059e-01
-2.08485794e+00 -8.99952292e-01 -3.71922553e-01 2.10905528e+00
5.40376067e-01 1.92948535e-01 1.33756651e-02 -8.15934390e-02
7.57214963e-01 3.04924458e-01 -6.78371072e-01 1.52072310e-01
-2.87803799e-01 -1.44939944e-01 5.29392183e-01 3.60766560e-01
-1.23525715e+00 1.00382066e+00 3.91399097e+00 7.73739576e-01
-1.25498068e+00 3.94062430e-01 8.02312255e-01 -4.27848220e-01
-1.81241333e-01 7.18166828e-02 -7.97680676e-01 8.61016691e-01
3.48585606e-01 -4.44208235e-02 4.40476447e-01 6.85685396e-01
4.05781567e-01 -5.10255992e-01 -5.76252520e-01 1.12800479e+00
3.16304974e-02 -1.26940680e+00 -2.88582861e-01 -1.30781248e-01
6.72812164e-01 2.74648666e-01 2.66718082e-02 -1.14115015e-01
-5.17527349e-02 -5.38895071e-01 7.68411517e-01 8.37156117e-01
4.88449097e-01 -6.29142642e-01 4.50158179e-01 2.92636864e-02
-1.46334195e+00 -2.80577272e-01 -2.56881028e-01 1.46587014e-01
2.07183301e-01 5.96424878e-01 -8.33736658e-02 6.48965597e-01
1.13705361e+00 1.07411933e+00 -6.89227581e-01 1.37011063e+00
-1.53303310e-01 3.53132874e-01 -3.45851749e-01 8.73347819e-02
1.74308196e-01 -1.03320278e-01 7.06377327e-01 8.96784544e-01
2.73473293e-01 4.61683363e-01 3.85542959e-01 7.61948109e-01
2.28583172e-01 1.08876221e-01 -1.44389197e-01 1.96099550e-01
4.85265881e-01 1.38096011e+00 -9.53806818e-01 -3.45161289e-01
-5.12025833e-01 9.43816781e-01 6.93540871e-02 5.17018020e-01
-1.09901190e+00 -2.88977534e-01 9.10396934e-01 7.88080320e-02
5.16664147e-01 -1.31162137e-01 -1.69405520e-01 -1.31897724e+00
9.06785801e-02 -3.69602025e-01 2.68929124e-01 -1.01627648e+00
-9.49686706e-01 5.23808956e-01 -2.22767279e-01 -1.40097821e+00
3.28492194e-01 -3.10027063e-01 -5.01570463e-01 8.59890223e-01
-1.70140994e+00 -8.38344514e-01 -5.95909297e-01 8.42288375e-01
6.71626508e-01 1.24725394e-01 1.01332881e-01 5.93408406e-01
-8.16608727e-01 2.20651418e-01 -1.33226395e-01 1.89664274e-01
7.49478519e-01 -8.50982189e-01 -7.20482692e-02 1.09727108e+00
-7.77363032e-02 2.66651720e-01 4.50525463e-01 -7.73300290e-01
-1.37957716e+00 -1.11886013e+00 4.30753291e-01 -2.70956874e-01
3.79752189e-01 -4.02530842e-02 -9.30407465e-01 3.34638208e-01
8.29214081e-02 4.77821380e-01 3.08707505e-01 -5.54474294e-01
4.80124727e-02 -3.13822240e-01 -9.31698084e-01 4.87213343e-01
1.08149564e+00 -1.90209687e-01 -6.45293072e-02 1.59437880e-01
7.97186315e-01 -4.70913589e-01 -6.65609837e-01 4.20800418e-01
4.74270195e-01 -1.35210085e+00 9.87437546e-01 -2.61756241e-01
5.97824693e-01 -7.06151307e-01 -1.84389338e-01 -8.63569558e-01
-3.22613686e-01 -3.35312098e-01 -3.07003140e-01 1.33671367e+00
4.79217172e-02 -4.66269433e-01 6.97088599e-01 6.06895864e-01
-2.54466176e-01 -9.79427397e-01 -8.84721696e-01 -3.07060003e-01
-7.17046142e-01 -4.07600492e-01 3.82456332e-01 7.60922432e-01
-3.36975366e-01 1.87339447e-02 -3.21658075e-01 2.66452312e-01
6.49315715e-01 2.76303470e-01 5.47467351e-01 -1.05074775e+00
-7.69925937e-02 -4.80761826e-01 -4.93981600e-01 -1.11949885e+00
-1.40392885e-01 -5.47947347e-01 1.78605728e-02 -1.52156067e+00
1.25502527e-01 -5.35982847e-01 -7.58542359e-01 4.52509046e-01
-5.09824097e-01 3.27615350e-01 5.85892320e-01 3.21210951e-01
-1.15487194e+00 6.26170814e-01 1.52847850e+00 7.46610165e-02
-3.16564977e-01 -5.39965257e-02 -7.10031569e-01 7.17018902e-01
9.48678672e-01 -4.59438562e-01 -3.29803735e-01 -4.87196088e-01
-2.21030056e-01 2.32494444e-01 6.66060686e-01 -1.10149097e+00
3.63676578e-01 -4.94440973e-01 5.84817350e-01 -6.40079558e-01
4.90828574e-01 -6.18649483e-01 -9.00087506e-02 3.74846250e-01
2.22116467e-02 -1.19676650e-01 1.14440002e-01 6.19382203e-01
-2.33271644e-01 -3.50128300e-02 6.77000165e-01 -2.72948563e-01
-1.22356951e+00 7.64584601e-01 4.99478094e-02 1.18517041e-01
1.08250403e+00 -2.34168112e-01 -5.63464344e-01 -2.15580583e-01
-4.52984065e-01 6.62221074e-01 4.84139413e-01 7.00365901e-01
8.89056146e-01 -1.10997260e+00 -5.04314840e-01 1.57109290e-01
3.54140401e-02 9.82600600e-02 6.39183342e-01 1.30339026e+00
-3.86924267e-01 2.04130009e-01 -4.96484488e-01 -8.05290282e-01
-1.15529871e+00 4.70594704e-01 2.70235717e-01 3.17225695e-01
-3.49951178e-01 1.07167745e+00 2.87090510e-01 2.96511889e-01
1.27624258e-01 -2.39009067e-01 -2.76561648e-01 -4.21686918e-02
6.17470801e-01 5.35894871e-01 -3.16693187e-01 -9.06799555e-01
-6.17503762e-01 5.15633345e-01 8.46018493e-02 9.43182260e-02
1.11740184e+00 -6.35127604e-01 -2.47116983e-02 4.11951184e-01
1.10498869e+00 -1.78409949e-01 -1.78741586e+00 -2.39925310e-01
-3.92009467e-01 -9.94740009e-01 1.89853325e-01 -4.61809456e-01
-1.61976385e+00 9.60069478e-01 8.23475778e-01 -1.17033847e-01
1.37077975e+00 7.48039223e-03 9.19748783e-01 -2.01826617e-01
3.72516423e-01 -8.99475992e-01 1.68490142e-01 3.46095413e-01
5.97632289e-01 -1.52171457e+00 5.06177768e-02 -5.65085888e-01
-8.93609881e-01 7.90445983e-01 1.08865857e+00 -2.27364168e-01
7.57086098e-01 -7.81744868e-02 7.51490593e-02 -2.87468433e-02
-7.25551426e-01 -5.04731834e-01 2.99342930e-01 3.25180888e-01
1.23566948e-01 -1.46543637e-01 -2.46579617e-01 6.20627403e-01
2.49827847e-01 1.87010750e-01 4.01688457e-01 9.77757692e-01
-5.09930789e-01 -7.44109154e-01 -2.95181930e-01 5.66128850e-01
-4.35812205e-01 -9.01417136e-02 1.87260676e-02 3.81175935e-01
4.87383097e-01 1.06278992e+00 4.74158563e-02 -5.76860487e-01
3.02075781e-02 -2.65818626e-01 2.69529641e-01 -3.59442204e-01
-4.80021685e-01 2.51126051e-01 -1.73639745e-01 -1.02549601e+00
-6.70773923e-01 -7.06932783e-01 -1.47446084e+00 -2.01129898e-01
-2.60819882e-01 -1.54550016e-01 3.84701550e-01 7.27566481e-01
5.18395901e-01 5.58873534e-01 6.00648403e-01 -1.04858196e+00
2.26095513e-01 -6.46114647e-01 -4.97159928e-01 3.18495810e-01
3.94979417e-01 -1.01422012e+00 -2.63000280e-01 9.13883820e-02] | [9.376530647277832, -0.37394052743911743] |
2a517d97-51ba-43ad-84d0-52441e5e0856 | physical-activity-recognition-by-utilising | 2201.08688 | null | https://arxiv.org/abs/2201.08688v1 | https://arxiv.org/pdf/2201.08688v1.pdf | Physical Activity Recognition by Utilising Smartphone Sensor Signals | Human physical motion activity identification has many potential applications in various fields, such as medical diagnosis, military sensing, sports analysis, and human-computer security interaction. With the recent advances in smartphones and wearable technologies, it has become common for such devices to have embedded motion sensors that are able to sense even small body movements. This study collected human activity data from 60 participants across two different days for a total of six activities recorded by gyroscope and accelerometer sensors in a modern smartphone. The paper investigates to what extent different activities can be identified by utilising machine learning algorithms using approaches such as majority algorithmic voting. More analyses are also provided that reveal which time and frequency domain based features were best able to identify individuals motion activity types. Overall, the proposed approach achieved a classification accuracy of 98 percent in identifying four different activities: walking, walking upstairs, walking downstairs, and sitting while the subject is calm and doing a typical desk-based activity. | ["Nathan Clarke' Fudong Li", 'Hind Alobaidi', 'Abdulrahman Alruban'] | 2022-01-20 | null | null | null | null | ['computer-security'] | ['miscellaneous'] | [ 5.46867907e-01 -3.50975990e-01 -6.82798386e-01 -1.40986713e-02
-1.85014158e-01 -1.94824085e-01 3.84009928e-01 1.41897753e-01
-3.94978344e-01 7.51905918e-01 4.53714967e-01 -3.06741983e-01
-1.69875354e-01 -8.75927150e-01 7.24289566e-02 -5.04141510e-01
-1.75299868e-01 -1.76211089e-01 2.68635929e-01 5.43813109e-02
2.52023965e-01 3.37614894e-01 -1.68232858e+00 -1.41018555e-01
6.12719595e-01 9.28518891e-01 -2.64391810e-01 7.83884346e-01
4.13471460e-01 5.72023511e-01 -7.49401927e-01 -1.54537736e-02
-8.31503980e-03 -6.82168245e-01 -3.91098112e-01 2.34750092e-01
5.02553135e-02 1.57700293e-02 -1.20156305e-02 3.95808280e-01
5.40289402e-01 4.89011467e-01 2.95495540e-01 -1.10386860e+00
7.78711140e-02 6.66196346e-02 -7.23755956e-01 5.58345854e-01
1.33707094e+00 4.68039960e-02 2.46833503e-01 -4.00939077e-01
1.45025358e-01 6.82353139e-01 1.01253140e+00 9.75211635e-02
-8.02811086e-01 -6.70134068e-01 -4.51663226e-01 4.06254709e-01
-1.44691801e+00 -2.87755072e-01 8.39986086e-01 -5.65379262e-01
1.06400871e+00 5.78207850e-01 1.28990519e+00 1.15043521e+00
5.55969357e-01 3.73614043e-01 1.25224423e+00 -5.01057267e-01
4.44531113e-01 -2.02593997e-01 2.70849496e-01 3.93458575e-01
7.54522920e-01 -2.51657784e-01 -8.97232592e-01 -4.64263082e-01
5.30551970e-01 4.01651829e-01 -5.08918464e-02 -8.32311958e-02
-1.43610048e+00 3.89684767e-01 -1.01538219e-01 5.75871706e-01
-6.10218942e-01 -2.69096177e-02 3.58122885e-01 -6.03884421e-02
3.42150450e-01 9.46421698e-02 -3.07398736e-01 -9.49699104e-01
-1.04838252e+00 2.59559959e-01 8.96886766e-01 2.23977253e-01
3.88550580e-01 8.01417977e-03 1.46107689e-01 5.64228177e-01
2.32489675e-01 5.34328997e-01 9.73616779e-01 -1.01759291e+00
4.85260099e-01 9.52357829e-01 4.40881997e-01 -1.27739298e+00
-8.19762707e-01 3.26725002e-03 -8.55670869e-01 -1.63888074e-02
4.90866542e-01 -4.58858997e-01 -2.75306225e-01 1.18580222e+00
6.51268065e-01 4.86106008e-01 -3.16325933e-01 7.78102815e-01
3.86844367e-01 1.57103777e-01 3.93503726e-01 -3.95937860e-01
1.75808620e+00 -3.25912148e-01 -8.14916313e-01 -3.26527864e-01
6.47528172e-01 -4.78410393e-01 8.87617111e-01 5.17988801e-01
-7.48870671e-01 -7.89211631e-01 -1.47197139e+00 3.95251572e-01
-3.23811948e-01 1.51946336e-01 6.95858836e-01 1.59342062e+00
-3.67103368e-01 7.98381984e-01 -1.32446599e+00 -7.17902422e-01
1.26784742e-01 3.10738236e-01 -3.79297584e-01 4.76296008e-01
-1.06343257e+00 7.47855246e-01 6.95326626e-02 -3.75241078e-02
2.47649364e-02 -1.50698379e-01 -9.59601641e-01 -4.68078017e-01
-6.43306300e-02 -6.21087372e-01 7.70680010e-01 -6.93652809e-01
-1.71970236e+00 8.67878258e-01 -4.46878731e-01 -4.09899354e-01
5.45232952e-01 -4.99697655e-01 -1.02479959e+00 6.17376007e-02
1.54455736e-01 -2.93675244e-01 5.63450992e-01 4.87260818e-02
-6.59631431e-01 -7.68692076e-01 -4.08608794e-01 3.86698067e-01
-3.23755950e-01 -6.08419813e-02 3.35315496e-01 -5.50512493e-01
3.41598958e-01 -1.22851050e+00 1.26968116e-01 -4.08647269e-01
-2.67252684e-01 -1.85889661e-01 7.98191011e-01 -8.19808245e-01
1.61620116e+00 -1.72209215e+00 -1.91911802e-01 4.15339977e-01
-4.74376567e-02 2.40984961e-01 9.70445693e-01 4.21834767e-01
2.04281896e-01 -1.56134218e-01 -6.60324693e-02 1.02887772e-01
-3.09750080e-01 2.99490184e-01 3.12445939e-01 8.43680441e-01
-4.56355065e-01 6.11157358e-01 -7.67355561e-01 -3.67740780e-01
7.32100666e-01 3.38416964e-01 1.46448359e-01 -1.22502752e-01
7.81500518e-01 8.05158794e-01 -4.86416489e-01 8.18150163e-01
1.36823952e-01 -3.01091429e-02 3.05905968e-01 6.45716712e-02
-7.18779340e-02 3.25206637e-01 -1.55933213e+00 1.56319094e+00
-2.96140075e-01 6.40155435e-01 -4.95281428e-01 -1.11664104e+00
8.03320348e-01 4.28130835e-01 8.96268606e-01 -5.14033914e-01
1.09269328e-01 1.32989466e-01 -6.09298721e-02 -8.41856658e-01
4.33975250e-01 1.92136526e-01 -4.01979744e-01 4.31981146e-01
-4.84322846e-01 6.74852848e-01 2.87162364e-01 -4.49827343e-01
1.32822049e+00 3.22410107e-01 1.16049206e+00 -2.10566550e-01
6.32564843e-01 -9.90311429e-02 3.95608187e-01 4.44885164e-01
-6.42022073e-01 1.42196059e-01 -4.51315373e-01 -5.57548702e-01
-3.18227291e-01 -1.01116371e+00 1.76409587e-01 1.18807375e+00
1.83997795e-01 -2.52948165e-01 -6.10667288e-01 -2.32422967e-02
-1.13445982e-01 7.58700222e-02 -4.61039841e-01 -2.55203992e-01
-7.67280579e-01 -9.79507208e-01 8.66976559e-01 7.53226221e-01
1.08646405e+00 -8.06969941e-01 -1.34002602e+00 2.63973862e-01
-4.42939907e-01 -8.68037403e-01 -2.62180120e-01 -1.53002664e-01
-1.25558162e+00 -1.26070011e+00 -7.69084513e-01 -3.11642468e-01
7.46406391e-02 4.61352289e-01 7.88603306e-01 -7.99011812e-02
-2.88764417e-01 4.56430823e-01 -3.49418998e-01 -3.46937984e-01
2.13328943e-01 1.64763689e-01 4.54586297e-01 1.54337823e-01
8.96851242e-01 -7.80714750e-01 -1.03499496e+00 4.82463956e-01
-5.64871021e-02 -3.61331075e-01 2.84352332e-01 1.32776827e-01
2.17994452e-01 1.14679828e-01 2.51101553e-01 -3.73072445e-01
7.63558626e-01 -8.00331414e-01 3.50200623e-01 -5.41365221e-02
-6.52652323e-01 -5.13703346e-01 -3.80478762e-02 -6.14373982e-01
-7.49694049e-01 7.82295763e-02 -9.90289897e-02 5.99882603e-01
-5.26367486e-01 1.28811479e-01 -2.55549490e-01 -6.72705751e-03
1.04771924e+00 2.25935236e-01 -1.37291089e-01 -3.12596589e-01
-2.99097538e-01 9.10187006e-01 5.39348543e-01 -3.27169836e-01
2.81173050e-01 5.61053038e-01 2.37678468e-01 -1.41317511e+00
-3.15600842e-01 -8.80680025e-01 -5.14923036e-01 -6.78926289e-01
1.02864313e+00 -9.12706733e-01 -1.18760848e+00 7.95293212e-01
-3.79645824e-01 -7.06933951e-03 1.78382099e-02 7.98425972e-01
-3.60520273e-01 6.88741386e-01 -2.90596426e-01 -1.29578018e+00
-3.24941307e-01 -4.12322849e-01 9.47741151e-01 6.88732326e-01
-1.37083542e+00 -9.27607358e-01 2.51334786e-01 8.68760347e-01
2.12794602e-01 9.20929372e-01 3.05606425e-02 -1.99613795e-01
3.27626467e-01 -5.23945689e-01 6.87156916e-01 -2.29014248e-01
6.52899265e-01 -5.01553237e-01 -6.42611563e-01 -1.48199439e-01
3.37384641e-02 6.74058422e-02 3.98180522e-02 6.00287437e-01
6.76031470e-01 -2.05302924e-01 -6.43394589e-01 2.92383790e-01
9.11748886e-01 4.26516384e-01 9.95971024e-01 5.06912470e-01
5.62451005e-01 4.79290783e-01 5.20294249e-01 4.99837965e-01
3.06750417e-01 8.69059920e-01 1.15530845e-02 2.97613323e-01
2.19560295e-01 -1.64666548e-01 5.42972267e-01 3.95839721e-01
-8.21775317e-01 3.01288571e-02 -7.92648315e-01 5.07541835e-01
-1.74373007e+00 -1.38283670e+00 -2.99214661e-01 2.36409259e+00
4.18506026e-01 1.73515186e-01 7.85532296e-01 1.08735943e+00
6.86709762e-01 1.91716403e-01 -4.64992493e-01 -4.02817905e-01
2.64250278e-01 2.36947775e-01 6.73345447e-01 1.26756519e-01
-1.13003135e+00 -5.40073626e-02 6.59002733e+00 3.03909391e-01
-8.84174287e-01 1.75420921e-02 4.35971171e-01 -1.75649881e-01
5.21697223e-01 -2.54546911e-01 -5.73352993e-01 9.60491717e-01
1.20548213e+00 1.48140267e-01 -4.82025258e-02 9.03311014e-01
8.01443696e-01 -9.41037357e-01 -5.90625584e-01 1.23982131e+00
-4.73474674e-02 -8.54461908e-01 -5.01231492e-01 2.92074919e-01
5.75367749e-01 -3.99486542e-01 -3.94520551e-01 1.38861372e-03
-2.19169617e-01 -8.03323030e-01 1.93957925e-01 6.20283723e-01
4.19673651e-01 -5.16871035e-01 4.59996283e-01 6.71381116e-01
-1.51834047e+00 -1.12809576e-01 4.02428985e-01 -1.16770005e+00
2.73431927e-01 5.36352038e-01 -5.74428260e-01 3.39993358e-01
8.63079369e-01 8.04295480e-01 -4.11825389e-01 8.18286300e-01
7.82405138e-02 9.03110147e-01 -4.80991155e-01 -2.95809358e-01
-3.89139861e-01 -3.70248944e-01 4.06291485e-01 1.08162618e+00
6.92690432e-01 6.57117516e-02 1.44028559e-01 -7.67779127e-02
5.64683557e-01 4.31368034e-03 -7.92733669e-01 1.75961092e-01
4.55883682e-01 9.96165633e-01 -8.89725745e-01 -2.93951333e-01
-2.90245533e-01 8.57082307e-01 -4.52803433e-01 -7.49624670e-02
-7.21379519e-01 -4.30910408e-01 8.65103662e-01 7.22950161e-01
-4.35988903e-01 -6.35142386e-01 -6.70112789e-01 -9.55378890e-01
2.38096565e-01 -5.16231477e-01 5.58871925e-01 -5.28241754e-01
-6.87877536e-01 -2.19054610e-01 2.61740059e-01 -1.58225012e+00
-6.99409008e-01 -3.66350621e-01 -8.37462723e-01 5.99267781e-01
-4.06819284e-01 -7.19978869e-01 -7.76589990e-01 7.44910181e-01
4.87024516e-01 5.67274652e-02 7.50458181e-01 4.57155019e-01
-4.81373280e-01 2.44182155e-01 -3.60432982e-01 1.00135624e-01
3.44551325e-01 -9.18115437e-01 2.05191344e-01 1.00050223e+00
1.93926126e-01 8.39257240e-01 7.73747027e-01 -7.95636058e-01
-1.17503786e+00 -6.24013543e-01 1.13795340e+00 -4.65718687e-01
2.72812545e-01 1.36157393e-01 -3.10130417e-01 3.97289008e-01
-2.13992447e-01 -3.59425731e-02 1.23013401e+00 5.64513057e-02
5.21152914e-01 -1.72530279e-01 -1.21837735e+00 6.03430152e-01
1.22382939e+00 -3.40013474e-01 -7.39966929e-01 -4.81395312e-02
-4.60887879e-01 -4.47860181e-01 -1.13492644e+00 3.35672170e-01
1.25542641e+00 -1.05133104e+00 1.14406073e+00 -1.58467442e-01
-2.74644673e-01 -3.99758279e-01 1.22712143e-01 -8.76595736e-01
-1.57623053e-01 -7.67494261e-01 -6.29327536e-01 7.45236874e-01
-2.84249485e-01 -7.72881329e-01 1.24308300e+00 7.06822634e-01
3.01767111e-01 -4.14364934e-01 -1.13840675e+00 -6.09233022e-01
-6.99702382e-01 -5.99313259e-01 3.10229808e-01 9.32725906e-01
5.47294736e-01 7.40210861e-02 -6.93103909e-01 -3.15860868e-01
4.48933095e-01 -2.05813691e-01 1.06587970e+00 -1.49789441e+00
-2.54798651e-01 -6.53686449e-02 -1.04673898e+00 -9.23648775e-01
-4.62121397e-01 -2.16446951e-01 -6.30594730e-01 -1.42623842e+00
-1.73983902e-01 2.06723928e-01 -8.35680515e-02 3.15618515e-01
-1.04609787e-01 6.41879857e-01 -4.29682940e-01 1.38555467e-01
-3.57237697e-01 -2.39525229e-01 7.72779107e-01 6.96438551e-02
-6.02826834e-01 7.49114513e-01 -5.35631120e-01 1.00578952e+00
9.02225912e-01 -1.68765262e-01 -5.44388592e-01 1.90704152e-01
3.68017882e-01 2.27743208e-01 4.67255473e-01 -1.67121875e+00
3.34341973e-02 -2.73647070e-01 8.44650507e-01 -2.25842372e-01
4.98106718e-01 -7.29586005e-01 7.54865348e-01 1.02556324e+00
1.37053847e-01 4.08307463e-02 -9.56009179e-02 5.83077967e-01
3.34287994e-02 2.89093047e-01 4.09367144e-01 -1.65070370e-01
-8.85773897e-01 -2.91963845e-01 -9.39377129e-01 -2.56431639e-01
1.22082484e+00 -1.25298035e+00 -8.46729148e-03 -4.25105780e-01
-1.01920605e+00 -2.56675392e-01 2.55846411e-01 5.94377279e-01
2.53656119e-01 -1.50371051e+00 -4.74616364e-02 1.24984905e-01
1.58628106e-01 -7.36204028e-01 3.16040128e-01 1.20807767e+00
-5.39863110e-01 4.02485132e-01 -6.04958534e-01 -6.99059427e-01
-1.52362692e+00 -2.54016221e-01 2.45793477e-01 -2.59350836e-01
-6.11941814e-01 3.56270350e-03 -9.82969046e-01 1.53006136e-01
-1.02822557e-01 -5.02840281e-01 -4.28421676e-01 1.28725976e-01
6.56273305e-01 1.46349919e+00 1.47702262e-01 -9.09991443e-01
-8.15927744e-01 9.04348254e-01 7.97109962e-01 -1.03604093e-01
4.52419311e-01 -4.80498731e-01 5.02112687e-01 7.71575987e-01
7.19682813e-01 1.04330920e-01 -8.42986763e-01 2.40024880e-01
4.12090629e-01 -3.64264101e-01 -4.06541348e-01 -2.80330896e-01
-4.08981800e-01 5.76363802e-01 1.16009176e+00 4.69853222e-01
1.12111306e+00 -5.61200261e-01 1.01015639e+00 3.12768728e-01
7.72587240e-01 -1.36640155e+00 -7.82469511e-02 -9.37666371e-02
3.82574238e-02 -1.14052117e+00 4.00060505e-01 -1.86964795e-01
-5.44098973e-01 7.28821218e-01 2.49982759e-01 -3.15769583e-01
7.13230669e-01 -8.30098242e-03 -2.21843764e-01 1.15391336e-01
-1.27868727e-02 -2.89470196e-01 4.20619071e-01 9.30538237e-01
6.03749216e-01 4.09330636e-01 -8.50507617e-01 3.71153593e-01
-3.03926229e-01 4.37732100e-01 3.48077625e-01 1.36376226e+00
-7.16609836e-01 -8.52355957e-01 -6.63190246e-01 6.38824880e-01
-9.32296216e-01 7.73822963e-01 -3.26381326e-01 9.09951270e-01
3.71869653e-01 1.42394042e+00 -3.86679918e-02 -5.11516273e-01
3.77620578e-01 3.81913185e-01 3.96177948e-01 -3.70921642e-01
-6.48625135e-01 -4.53315705e-01 3.19960952e-01 -8.18973422e-01
-1.06976151e+00 -9.02663946e-01 -1.00123560e+00 -4.45400745e-01
2.78408617e-01 2.52996641e-03 4.80767369e-01 1.12475955e+00
2.38869965e-01 2.81940550e-01 2.35030815e-01 -1.07154715e+00
-2.59775259e-02 -1.20311260e+00 -6.83575630e-01 4.89944458e-01
4.64680754e-02 -8.37842882e-01 -1.09347932e-01 3.35410714e-01] | [7.284724235534668, 0.6203824877738953] |
b013efa1-aabb-44d8-b0e7-bd3f9b8dafbd | a-biomedical-entity-extraction-pipeline-for | 2304.08999 | null | https://arxiv.org/abs/2304.08999v1 | https://arxiv.org/pdf/2304.08999v1.pdf | A Biomedical Entity Extraction Pipeline for Oncology Health Records in Portuguese | Textual health records of cancer patients are usually protracted and highly unstructured, making it very time-consuming for health professionals to get a complete overview of the patient's therapeutic course. As such limitations can lead to suboptimal and/or inefficient treatment procedures, healthcare providers would greatly benefit from a system that effectively summarizes the information of those records. With the advent of deep neural models, this objective has been partially attained for English clinical texts, however, the research community still lacks an effective solution for languages with limited resources. In this paper, we present the approach we developed to extract procedures, drugs, and diseases from oncology health records written in European Portuguese. This project was conducted in collaboration with the Portuguese Institute for Oncology which, besides holding over $10$ years of duly protected medical records, also provided oncologist expertise throughout the development of the project. Since there is no annotated corpus for biomedical entity extraction in Portuguese, we also present the strategy we followed in annotating the corpus for the development of the models. The final models, which combined a neural architecture with entity linking, achieved $F_1$ scores of $88.6$, $95.0$, and $55.8$ per cent in the mention extraction of procedures, drugs, and diseases, respectively. | ['Mário Amorim Lopes', 'Catarina Sousa Santos', 'Alípio Jorge', 'Arian Pasquali', 'Hugo Sousa'] | 2023-04-18 | null | null | null | null | ['entity-linking'] | ['natural-language-processing'] | [ 3.99633311e-02 6.38848782e-01 -2.75561631e-01 -1.58256561e-01
-1.00712132e+00 -4.46116060e-01 2.38238573e-01 1.07468474e+00
-8.48332345e-01 1.21681476e+00 3.44365478e-01 -5.93409657e-01
-2.30837315e-01 -7.70486653e-01 -3.90876234e-01 -3.54671806e-01
2.65422374e-01 6.65187061e-01 -5.13199925e-01 1.19815163e-01
3.90941575e-02 5.79758465e-01 -9.34896648e-01 3.88979644e-01
7.24781990e-01 7.38131583e-01 4.19459529e-02 4.38712984e-01
-3.68968427e-01 1.06142092e+00 -6.04092658e-01 -7.03387201e-01
-1.89100549e-01 -3.85831088e-01 -9.53993857e-01 -2.51994461e-01
-5.19382730e-02 -2.40156934e-01 -3.59531969e-01 8.90486896e-01
5.83264649e-01 -1.22651942e-01 6.48659825e-01 -6.06215894e-01
-3.98787558e-01 8.67082775e-01 -1.46255314e-01 6.91389889e-02
5.08832157e-01 -1.14871584e-01 8.88034642e-01 -5.69198668e-01
1.12512422e+00 3.46716911e-01 8.11848223e-01 7.05252290e-01
-8.21840942e-01 -4.44203347e-01 -3.42664003e-01 -7.62526691e-02
-1.35389698e+00 -7.44871557e-01 4.29115325e-01 -4.31077927e-01
1.42249358e+00 1.42344162e-01 6.74977005e-01 9.65068638e-01
4.73072529e-01 4.70836133e-01 7.29251206e-01 -5.65405190e-01
6.16987608e-02 5.59684575e-01 1.22335114e-01 7.91688502e-01
5.83549559e-01 -2.52830416e-01 -2.31662497e-01 -1.53973803e-01
3.47388357e-01 -2.06894740e-01 -3.17309141e-01 4.56596315e-02
-9.11229372e-01 8.16328764e-01 3.79214108e-01 8.67585123e-01
-6.22990310e-01 -1.65980503e-01 6.94825411e-01 -6.40508672e-03
3.33490402e-01 6.82036161e-01 -4.96051699e-01 -1.43894270e-01
-1.07973802e+00 4.85377945e-02 1.09118068e+00 9.44845378e-01
1.23592746e-02 -2.25520432e-01 1.47623658e-01 5.89419425e-01
2.19191872e-02 8.94268900e-02 5.00641346e-01 -7.80691028e-01
4.57853675e-01 7.85593390e-01 1.47772983e-01 -1.05649316e+00
-8.62213254e-01 -5.03661990e-01 -1.05151749e+00 -4.20421153e-01
5.24559855e-01 -4.76971209e-01 -8.07832479e-01 1.49643505e+00
4.90928479e-02 -6.07124388e-01 3.90576720e-01 3.78331065e-01
1.23770201e+00 2.38549575e-01 5.69731057e-01 -4.66575176e-01
1.57674837e+00 -5.31754375e-01 -1.10289407e+00 -9.58936736e-02
1.05143619e+00 -6.90282881e-01 2.75487781e-01 1.61557853e-01
-1.48785496e+00 -1.59028515e-01 -7.89910138e-01 -1.90865636e-01
-7.07984090e-01 1.47467092e-01 7.09013164e-01 7.15421140e-01
-8.88855934e-01 6.04698122e-01 -1.08460915e+00 -6.00135624e-01
8.33312213e-01 5.77802420e-01 -6.72627211e-01 -7.01106712e-02
-1.13724279e+00 1.30317938e+00 6.97464406e-01 1.09785192e-01
-4.08401549e-01 -7.27992415e-01 -9.46336091e-01 1.54369608e-01
3.52564573e-01 -8.43119144e-01 1.13631415e+00 -5.08010924e-01
-8.33656132e-01 1.08024633e+00 -2.17784159e-02 -5.70356071e-01
3.36383939e-01 1.84275940e-01 -5.56943417e-01 2.03719243e-01
1.25472263e-01 6.31089270e-01 -1.34767786e-01 -5.51645160e-01
-7.73388803e-01 -5.13554454e-01 -1.33579999e-01 1.70411453e-01
-3.64901215e-01 2.13422179e-01 -3.89046997e-01 -3.24226856e-01
-2.92194128e-01 -6.77872121e-01 -5.04020452e-01 -2.01663435e-01
-2.76040614e-01 3.36203836e-02 1.33682787e-01 -1.19070470e+00
1.35102165e+00 -1.92654169e+00 9.67669562e-02 3.40178721e-02
4.33505982e-01 2.62751132e-01 4.13127452e-01 6.72768414e-01
-1.13404401e-01 4.17433649e-01 -4.04331028e-01 -1.75675079e-01
-3.09120864e-02 2.79542208e-01 2.51461387e-01 3.59453678e-01
2.10319594e-01 8.35466921e-01 -8.60717893e-01 -7.78000653e-01
-1.28494743e-02 6.75080895e-01 -4.19939756e-01 -1.42401401e-02
-2.92399665e-03 9.92503837e-02 -4.65888411e-01 7.73593962e-01
1.51447073e-01 -3.91362548e-01 5.15381455e-01 -8.97616223e-02
-3.19887064e-02 3.45270038e-01 -6.89963818e-01 1.96163607e+00
-2.64264703e-01 5.20190597e-01 2.24110648e-01 -9.83857989e-01
4.94596928e-01 8.74561906e-01 8.63349795e-01 -5.93132436e-01
3.60488445e-01 5.34687579e-01 7.24260416e-03 -7.53658295e-01
6.55843139e-01 -3.89031917e-01 -1.69003114e-01 1.06711149e-01
-1.08940564e-02 -4.85805161e-02 5.48780501e-01 2.66851962e-01
1.38601506e+00 6.76007047e-02 7.10185051e-01 -2.52238438e-02
7.37016618e-01 5.11371791e-01 5.10342538e-01 2.26698160e-01
-2.77045637e-01 2.83712387e-01 6.19761348e-01 -5.44944286e-01
-1.07178640e+00 -4.24649805e-01 -5.03395200e-01 2.20734149e-01
-6.23996735e-01 -3.44837606e-01 -7.97898471e-01 -7.59844363e-01
-2.59522259e-01 6.90728247e-01 -5.54437160e-01 9.43770558e-02
-7.44484007e-01 -8.72007191e-01 8.36558938e-01 5.64265430e-01
2.41122857e-01 -1.34854770e+00 -8.61687779e-01 6.44567251e-01
-3.22934598e-01 -1.10921431e+00 -9.51543972e-02 3.97529215e-01
-6.85123920e-01 -1.08316457e+00 -8.37653399e-01 -7.97531724e-01
7.04162061e-01 -7.65123725e-01 1.16658592e+00 2.70852838e-02
-6.25570416e-01 4.96481359e-02 -1.62262604e-01 -6.71240747e-01
-7.36893713e-01 4.49257314e-01 -3.89667004e-01 -7.65670657e-01
6.10427558e-01 4.29741405e-02 -4.22006637e-01 -3.46813142e-01
-1.12388802e+00 -1.06009193e-01 7.15923667e-01 9.16706681e-01
4.47189093e-01 1.25615746e-01 7.85148025e-01 -1.22369504e+00
6.70779169e-01 -5.43426394e-01 -2.93514341e-01 2.01327994e-01
-6.79946542e-01 -1.18580461e-01 5.24584234e-01 1.43396586e-01
-8.84496808e-01 3.34236741e-01 -6.79840267e-01 1.97073668e-01
-4.18325543e-01 1.12751567e+00 -2.44994890e-02 2.60078222e-01
6.50964737e-01 -2.12326318e-01 -5.94410226e-02 -2.28031203e-01
-9.62986983e-03 7.98892081e-01 4.15162623e-01 -1.17845245e-01
9.82168168e-02 1.62328199e-01 -1.07420832e-01 -7.40915835e-01
-5.94356239e-01 -3.39836806e-01 -4.62180763e-01 2.09606603e-01
1.12351274e+00 -7.21184254e-01 -7.33654022e-01 1.84159011e-01
-1.01229417e+00 8.02887678e-02 -3.69510084e-01 5.79998970e-01
-2.42681563e-01 2.46418491e-01 -8.75507593e-01 -5.04059911e-01
-6.52300715e-01 -1.08744287e+00 5.52957892e-01 1.80501461e-01
-7.18146026e-01 -1.15528119e+00 -5.43355308e-02 3.61270040e-01
4.25063044e-01 5.96606612e-01 1.27279198e+00 -8.34889650e-01
-1.22310624e-01 -6.42898321e-01 -2.95259506e-01 -1.97291840e-02
3.37788671e-01 -1.31424427e-01 -7.71064222e-01 -1.06788397e-01
1.49700880e-01 -2.76635587e-01 3.41781974e-01 4.61274117e-01
8.88354540e-01 -2.50576526e-01 -6.59956992e-01 2.90679395e-01
1.56730556e+00 6.14551485e-01 6.39518380e-01 4.00428385e-01
3.78782362e-01 8.18429768e-01 3.06578428e-01 2.51771748e-01
4.00251508e-01 3.27490538e-01 2.09887818e-01 -1.47052318e-01
9.19515714e-02 1.05984226e-01 -1.33395612e-01 6.97673082e-01
-1.24929212e-01 -4.16094154e-01 -1.23482609e+00 8.69871736e-01
-1.34125137e+00 -7.49370456e-01 -5.14446339e-03 1.85207462e+00
1.12608778e+00 2.08868042e-01 -1.65996656e-01 3.77201848e-02
3.66127938e-01 -3.46718580e-01 -3.07290405e-01 -5.23270905e-01
7.32904598e-02 4.30050969e-01 6.19352818e-01 2.54241705e-01
-1.05731440e+00 6.07320189e-01 6.03220701e+00 4.52085644e-01
-8.59388173e-01 -1.09833129e-01 6.99594855e-01 -1.24486648e-01
-7.23572969e-02 -4.24326122e-01 -7.41253853e-01 2.06473008e-01
1.53952408e+00 -2.90510744e-01 -8.16653743e-02 5.97228408e-01
7.01609328e-02 -1.37419105e-02 -1.12573147e+00 5.75669229e-01
-8.51876736e-02 -1.57080317e+00 -3.40448260e-01 3.05545032e-01
4.52828884e-01 -4.67827395e-02 -2.08743304e-01 2.08990216e-01
3.50747824e-01 -1.28813744e+00 2.67315507e-01 5.09656191e-01
7.76453376e-01 -8.17035556e-01 1.32085335e+00 2.85616010e-01
-7.67535627e-01 -7.32438043e-02 -1.45398557e-01 3.30758244e-01
2.78731644e-01 4.61951166e-01 -1.16782033e+00 9.56420243e-01
5.55649579e-01 6.13696575e-01 -2.81377792e-01 7.82350183e-01
6.25531226e-02 1.10267051e-01 -1.14188194e-01 -6.01007044e-02
3.22724700e-01 1.51425630e-01 8.43699742e-03 1.36930931e+00
4.04080331e-01 2.82971382e-01 -9.07235146e-02 5.26425958e-01
-4.41354394e-01 4.87574458e-01 -6.65358901e-01 -6.96809471e-01
2.02437147e-01 1.23788273e+00 -9.28019583e-01 -3.14826041e-01
-5.08004487e-01 5.76192260e-01 3.21340173e-01 -1.16918320e-02
-7.40911186e-01 -6.11849785e-01 8.76515806e-02 6.58286037e-03
4.99055758e-02 1.98323071e-01 -3.40862781e-01 -7.94637144e-01
-1.14447773e-01 -1.15937901e+00 7.84664869e-01 -5.42917728e-01
-9.81637537e-01 8.71823788e-01 -2.64302313e-01 -7.53621221e-01
-5.55001736e-01 -5.55136979e-01 2.15654850e-01 1.06261945e+00
-1.21981657e+00 -1.11585939e+00 5.59455194e-02 3.08598191e-01
2.24468589e-01 -1.19548544e-01 1.35908830e+00 6.96922958e-01
-6.76275790e-01 5.86901784e-01 -4.05550450e-02 4.72783804e-01
6.93533361e-01 -1.10219598e+00 -9.12367478e-02 2.57479578e-01
-2.10264906e-01 8.86711955e-01 4.94427115e-01 -7.75659502e-01
-1.15540493e+00 -8.54401827e-01 1.63905108e+00 -4.57485735e-01
6.45479918e-01 1.92730784e-01 -8.56584966e-01 6.88725293e-01
5.89763939e-01 -5.40464640e-01 1.13054764e+00 -1.95880249e-01
2.63862818e-01 1.67754456e-01 -1.42989624e+00 4.73033220e-01
4.88724589e-01 -5.06834447e-01 -6.54595912e-01 5.58693409e-01
4.67721641e-01 -7.12685525e-01 -1.38812792e+00 3.58129531e-01
4.49099809e-01 -4.27835882e-01 7.83617854e-01 -7.64096200e-01
7.18715727e-01 1.00757219e-01 9.05167833e-02 -8.97606909e-01
-3.97116803e-02 -1.80689350e-01 2.06352681e-01 1.00150120e+00
1.07825327e+00 -4.30714667e-01 8.94353092e-01 1.15976322e+00
-3.79097193e-01 -8.51261199e-01 -8.64013433e-01 -1.49896562e-01
3.59270662e-01 -3.61700296e-01 4.34966654e-01 1.08529687e+00
2.11308882e-01 1.19493775e-01 8.06152970e-02 -9.26339030e-02
1.60439596e-01 -2.45723680e-01 1.62950292e-01 -1.00337458e+00
-2.54122168e-02 -5.41339517e-01 -2.94014394e-01 5.51216602e-02
2.22479701e-02 -9.71692681e-01 -9.57052484e-02 -1.99546075e+00
3.24009836e-01 -2.70789474e-01 -2.97826499e-01 7.45898962e-01
1.29919812e-01 2.95201801e-02 -2.57729828e-01 9.88730118e-02
-8.95922333e-02 -1.44530103e-01 7.85908103e-01 -1.99162275e-01
-1.69565707e-01 -2.33340606e-01 -1.00520289e+00 7.65993714e-01
6.97444022e-01 -7.17686713e-01 -1.53070455e-02 -4.31161910e-01
4.01285768e-01 3.39899331e-01 -1.76775515e-01 -7.87856281e-01
4.66307163e-01 1.32075429e-01 4.58747804e-01 -6.77949905e-01
1.83362037e-01 -1.13632596e+00 5.95181406e-01 8.71195197e-01
-4.33808744e-01 9.39791948e-02 5.64476788e-01 9.65983644e-02
-3.43422204e-01 -5.23862183e-01 4.85668570e-01 -5.69593608e-01
-1.85172141e-01 1.21650770e-02 -5.67656994e-01 -1.41257137e-01
1.02911711e+00 -2.36337036e-02 -2.87120968e-01 -1.34290427e-01
-1.09598982e+00 1.61713868e-01 5.03211558e-01 -1.12590969e-01
4.11715716e-01 -8.34707797e-01 -7.02332258e-01 -1.89928636e-01
-8.84605497e-02 1.73044596e-02 4.54626262e-01 8.39372218e-01
-1.09040391e+00 8.54982018e-01 -2.59790242e-01 -2.41212919e-02
-1.32020509e+00 7.94850349e-01 2.19507113e-01 -8.23760092e-01
-5.54210901e-01 6.33726418e-01 -3.85477543e-01 -2.73212701e-01
2.19742954e-01 -2.66471535e-01 -6.58917725e-01 4.66131628e-01
3.56434584e-01 1.54827371e-01 5.22588551e-01 -6.60089910e-01
-4.02637124e-01 2.12481946e-01 -3.82108569e-01 -1.49645612e-01
1.57372987e+00 2.06414834e-01 -2.65650690e-01 1.94597524e-02
1.04775143e+00 -2.98565011e-02 -2.75962144e-01 6.87640980e-02
3.43322605e-01 1.63643301e-01 1.18290730e-01 -1.12953639e+00
-1.11650693e+00 6.77683711e-01 2.34337613e-01 2.45476708e-01
1.32183886e+00 -1.37111098e-01 7.27320492e-01 4.69534189e-01
8.99307206e-02 -1.02517223e+00 -6.94361985e-01 3.16110432e-01
5.34156978e-01 -9.90376711e-01 3.49164516e-01 -2.53067553e-01
-4.01880234e-01 1.03452361e+00 1.76246867e-01 5.33016860e-01
4.93856192e-01 5.34598231e-01 6.61427826e-02 -5.38284779e-01
-6.09244764e-01 1.15589119e-01 1.16396591e-01 2.33119890e-01
9.39556599e-01 5.91685763e-03 -8.52288663e-01 6.78871274e-01
-5.56483753e-02 4.63011175e-01 5.88148177e-01 1.29388571e+00
5.42712100e-02 -1.24742079e+00 -1.12607613e-01 6.07314646e-01
-1.35217285e+00 -3.00759971e-01 -5.72715640e-01 1.18483019e+00
6.28700331e-02 6.84715688e-01 -1.70526132e-01 9.52971280e-02
5.26605189e-01 3.86298507e-01 2.98819989e-01 -6.25186563e-01
-1.07654858e+00 8.75316709e-02 6.16794705e-01 -1.94512650e-01
-5.69178104e-01 -7.23793149e-01 -1.45872211e+00 -3.24633390e-01
-1.70737952e-01 3.99553567e-01 9.05889809e-01 7.81396449e-01
2.80037194e-01 9.41004157e-01 -1.94761619e-01 -2.38618180e-02
-6.47268966e-02 -9.05818641e-01 -2.96592355e-01 9.55720842e-02
-3.36514935e-02 -1.08749375e-01 3.54359001e-02 1.30400315e-01] | [8.429730415344238, 8.654623031616211] |
e6629033-4610-445f-a5d7-6bdb83a51e23 | global-to-local-neural-networks-for-document | 2009.10359 | null | https://arxiv.org/abs/2009.10359v1 | https://arxiv.org/pdf/2009.10359v1.pdf | Global-to-Local Neural Networks for Document-Level Relation Extraction | Relation extraction (RE) aims to identify the semantic relations between named entities in text. Recent years have witnessed it raised to the document level, which requires complex reasoning with entities and mentions throughout an entire document. In this paper, we propose a novel model to document-level RE, by encoding the document information in terms of entity global and local representations as well as context relation representations. Entity global representations model the semantic information of all entities in the document, entity local representations aggregate the contextual information of multiple mentions of specific entities, and context relation representations encode the topic information of other relations. Experimental results demonstrate that our model achieves superior performance on two public datasets for document-level RE. It is particularly effective in extracting relations between entities of long distance and having multiple mentions. | ['Weijian Sun', 'Difeng Wang', 'Wei Hu', 'Ermei Cao'] | 2020-09-22 | null | https://aclanthology.org/2020.emnlp-main.303 | https://aclanthology.org/2020.emnlp-main.303.pdf | emnlp-2020-11 | ['document-level-relation-extraction'] | ['natural-language-processing'] | [-8.82596150e-02 4.48810637e-01 -4.69136804e-01 -2.71525294e-01
-6.47568703e-01 -6.72151029e-01 7.89444268e-01 9.45803285e-01
-2.92165309e-01 9.07846808e-01 8.22714269e-01 -1.00010864e-01
-4.84662682e-01 -1.38169265e+00 -3.89056355e-01 -1.82145163e-01
-2.03747228e-01 5.15131831e-01 4.68941599e-01 -3.44207048e-01
5.87958544e-02 5.08898556e-01 -1.13555443e+00 5.75602412e-01
8.90150011e-01 9.82733071e-01 -1.65246144e-01 1.75793260e-01
-8.92230093e-01 1.20059919e+00 -1.07030916e+00 -5.58339834e-01
-3.18136483e-01 -1.33434400e-01 -1.47060430e+00 -1.18572585e-01
-3.28773931e-02 3.79232317e-01 -3.54962200e-01 9.91393387e-01
1.17049575e-01 3.19088668e-01 7.72446752e-01 -1.05030441e+00
-1.02149236e+00 9.14255202e-01 -4.55519915e-01 4.07343775e-01
5.63094199e-01 -9.70352173e-01 1.55269659e+00 -9.39606130e-01
1.04439020e+00 1.27969480e+00 4.84449387e-01 2.21236553e-02
-5.84172070e-01 -4.77540612e-01 4.36024368e-01 2.54171461e-01
-1.60054672e+00 -1.51626747e-02 1.75613821e-01 -3.13430160e-01
1.77564251e+00 2.38019690e-01 3.40099744e-02 4.49002028e-01
-1.28343716e-01 4.41749901e-01 3.90734136e-01 -5.15008330e-01
-6.65094256e-02 2.80923843e-02 7.36624062e-01 3.94640595e-01
7.46546268e-01 -7.36411631e-01 -3.96716893e-01 -2.86595374e-01
4.15077269e-01 -4.40897234e-02 -1.98305994e-01 5.43941818e-02
-8.55356812e-01 6.41239941e-01 6.41675651e-01 7.74158835e-01
-4.23915535e-01 -2.19703436e-01 4.90921795e-01 -1.33712769e-01
6.59815431e-01 6.72900438e-01 -8.64732802e-01 3.33599508e-01
-4.52045023e-01 1.28642574e-01 8.91868651e-01 1.56846058e+00
7.35702395e-01 -7.77258933e-01 -2.23411024e-01 8.53618085e-01
1.79377705e-01 2.14859575e-01 5.47281742e-01 -4.02296901e-01
1.06913841e+00 1.44610453e+00 1.06068462e-01 -1.39892232e+00
-5.86255193e-01 -4.30029899e-01 -6.21712327e-01 -8.37360740e-01
-2.47243971e-01 -2.17109174e-01 -7.82994926e-01 1.29243577e+00
5.36627829e-01 1.42510995e-01 6.33270085e-01 2.67378837e-01
1.69542325e+00 7.81512141e-01 4.61268395e-01 -2.53285944e-01
1.85723448e+00 -7.22058952e-01 -1.22196043e+00 -1.95680857e-01
1.07376289e+00 -5.36886156e-01 1.25814900e-01 -5.01609325e-01
-6.54676139e-01 -2.25346074e-01 -9.02886450e-01 -3.47575575e-01
-1.30674362e+00 3.44210684e-01 6.99669600e-01 2.62930602e-01
-4.92178351e-01 3.68602693e-01 -4.76173937e-01 -6.07431233e-01
3.14978570e-01 1.98632941e-01 -5.86633086e-01 5.44787496e-02
-1.72934020e+00 1.04241025e+00 9.74645913e-01 4.95789535e-02
1.50031179e-01 -6.90605819e-01 -1.28648114e+00 5.62549055e-01
8.05663168e-01 -6.38645053e-01 7.16435075e-01 6.66869208e-02
-4.29378778e-01 8.06503296e-01 -5.58181942e-01 -4.15929139e-01
-4.71678823e-01 -4.15812314e-01 -1.22694838e+00 7.73408264e-02
5.29707968e-01 -8.26011524e-02 -7.21774399e-02 -1.05979788e+00
-9.61537957e-01 -5.66416502e-01 2.97033697e-01 2.27415219e-01
-2.70672947e-01 5.51415145e-01 -5.16827703e-01 -5.91509461e-01
4.44668531e-01 -4.73252743e-01 -9.19469967e-02 -1.03007901e+00
-8.30128849e-01 -7.90908635e-01 8.59467030e-01 -5.90532303e-01
1.76194584e+00 -1.88699186e+00 -1.34218425e-01 3.87471735e-01
3.47308040e-01 2.32758299e-01 2.68022418e-01 8.28035593e-01
-2.42902070e-01 6.06411695e-01 5.27752377e-02 2.58046150e-01
-9.33930501e-02 3.76218855e-01 -6.52453184e-01 -2.14502931e-01
4.16266561e-01 1.20135367e+00 -9.61611807e-01 -7.38863826e-01
-3.77195358e-01 4.21294302e-01 -1.49493843e-01 1.97899882e-02
-1.08925402e-01 -1.10624373e-01 -9.59133863e-01 5.41190028e-01
4.47988391e-01 -5.83489180e-01 6.77852154e-01 -4.77042675e-01
1.92397729e-01 9.49515879e-01 -1.29056394e+00 1.19357049e+00
-3.59158248e-01 3.20109606e-01 -5.04628479e-01 -7.12573409e-01
1.13711917e+00 5.66184223e-01 4.40771163e-01 -4.11626428e-01
-1.67494327e-01 7.55674988e-02 -5.78076303e-01 -5.70110321e-01
8.19575727e-01 1.17312461e-01 -3.75130922e-01 8.37640613e-02
1.81101859e-01 7.07288980e-02 6.57777190e-01 6.73172057e-01
1.30402875e+00 1.89325940e-02 8.83102119e-01 -3.96005772e-02
7.65070796e-01 6.39016777e-02 6.29526198e-01 3.80390793e-01
4.66922253e-01 1.19138174e-01 6.39763772e-01 -2.08770528e-01
-4.64300454e-01 -9.24271286e-01 -1.72201589e-01 8.24538231e-01
4.00242031e-01 -1.36609006e+00 -3.09463114e-01 -1.27238047e+00
1.33492341e-02 6.68183982e-01 -6.79249704e-01 9.38343778e-02
-7.77434349e-01 -7.46304572e-01 3.77707720e-01 9.36290681e-01
5.89588642e-01 -7.86522329e-01 -2.56720372e-02 2.82991409e-01
-5.00289500e-01 -1.65987825e+00 4.27699462e-02 2.55742550e-01
-6.92005813e-01 -1.25318825e+00 -2.88490862e-01 -8.71435285e-01
7.85311580e-01 3.89624722e-02 1.26131833e+00 9.85719264e-02
-6.65332153e-02 9.83122811e-02 -6.89103484e-01 -2.61442155e-01
8.69890973e-02 4.06977177e-01 -2.81311601e-01 -3.78974080e-01
7.92920232e-01 -2.70417660e-01 -1.57428637e-01 3.22103649e-01
-7.76871324e-01 -4.72949624e-01 2.61912465e-01 5.60313940e-01
5.96957922e-01 7.10508287e-01 6.49270415e-01 -1.48542702e+00
7.30041504e-01 -9.32060957e-01 -7.98036158e-03 9.20824409e-01
-6.34244084e-01 2.11677521e-01 2.45643005e-01 3.80353183e-02
-1.50493395e+00 -4.26798135e-01 1.32085189e-01 5.17183065e-01
-2.56213009e-01 1.15573537e+00 -5.35583675e-01 7.05002129e-01
6.17249787e-01 -1.37791142e-01 -1.05835760e+00 -5.67059934e-01
5.65191269e-01 7.00208247e-01 5.53871870e-01 -9.16417241e-01
6.55754209e-01 2.59121507e-01 2.14873880e-01 -5.91613293e-01
-1.50812316e+00 -9.17690396e-01 -1.00514174e+00 3.75898093e-01
8.59108984e-01 -1.09723711e+00 -2.45738089e-01 -2.67480463e-01
-1.12578690e+00 6.17854536e-01 -4.84616190e-01 2.88658589e-01
7.01657906e-02 1.66552410e-01 -5.61313212e-01 -4.19049025e-01
-1.85173094e-01 -4.13490057e-01 1.07124317e+00 5.13948262e-01
-5.14106393e-01 -1.17789614e+00 8.86148140e-02 2.35272065e-01
-5.87648265e-02 3.51726979e-01 1.24551427e+00 -1.27030766e+00
-4.15495038e-01 -4.30191517e-01 -6.23211801e-01 -1.33767068e-01
7.15748906e-01 -1.16407141e-01 -5.84281206e-01 2.84138769e-01
-5.40003061e-01 1.57250017e-01 8.43045175e-01 -9.07519683e-02
8.23692560e-01 -5.83079278e-01 -1.01411057e+00 3.17997396e-01
1.21988034e+00 2.99734116e-01 6.27331078e-01 4.86288965e-01
8.59280229e-01 8.16494703e-01 9.12429512e-01 2.92210996e-01
7.02567637e-01 5.69021821e-01 -9.26080719e-02 1.10555589e-01
-8.78894255e-02 -4.04477239e-01 -2.77810574e-01 6.00992680e-01
-3.46440643e-01 -4.96824145e-01 -9.84564126e-01 7.55486727e-01
-1.88042724e+00 -1.13320458e+00 -4.03218061e-01 1.57592440e+00
9.03036118e-01 -4.32714149e-02 -1.98069617e-01 7.27121606e-02
9.32471037e-01 5.77398464e-02 -1.26775175e-01 -2.05755636e-01
-4.03135031e-01 1.84963807e-01 3.96450073e-01 3.09862852e-01
-1.27040923e+00 1.24915695e+00 5.87521410e+00 8.01776469e-01
-2.72124171e-01 -2.21428145e-02 1.75394192e-01 4.48362619e-01
-4.21968997e-01 3.17815125e-01 -1.49539530e+00 8.54217038e-02
8.68886352e-01 -6.13360405e-01 -4.18224454e-01 7.18891621e-01
-5.73433816e-01 -4.11156714e-02 -9.06736732e-01 5.57757318e-01
6.50155321e-02 -1.49072754e+00 3.73210549e-01 6.78728223e-02
8.13528955e-01 -3.96918535e-01 -4.54144508e-01 4.85910952e-01
5.65824449e-01 -1.03070712e+00 1.76638365e-01 2.94692755e-01
5.80038846e-01 -8.94716978e-01 1.12880075e+00 8.33512917e-02
-1.88435590e+00 1.98458228e-02 -3.96855384e-01 1.60584807e-01
1.08130582e-01 7.59401739e-01 -1.01839721e+00 1.33455884e+00
6.92883074e-01 9.54550803e-01 -5.43038070e-01 5.59277892e-01
-7.48897254e-01 3.27081412e-01 -2.05718532e-01 -2.73801107e-03
1.43812090e-01 2.14888770e-02 1.76319286e-01 1.76172435e+00
2.52338856e-01 6.28032148e-01 -3.88406068e-02 4.41875666e-01
-3.55679780e-01 5.39394975e-01 -5.36723673e-01 -1.83330700e-01
7.70880044e-01 1.23247683e+00 -8.35458815e-01 -6.44108355e-01
-4.60686564e-01 6.65083110e-01 6.29865706e-01 5.37927866e-01
-4.77502435e-01 -9.95603442e-01 6.56378090e-01 -9.87250358e-02
4.86913890e-01 -1.70980781e-01 -1.03635751e-01 -1.21815777e+00
2.46702880e-01 -1.44868419e-01 1.14436948e+00 -6.30809605e-01
-1.29758704e+00 7.37892985e-01 1.96887404e-01 -9.83310282e-01
-2.58148253e-01 -5.10394335e-01 -4.04223531e-01 8.27649653e-01
-1.51224530e+00 -1.15832770e+00 4.67028609e-03 4.81986940e-01
1.46113366e-01 -2.30618957e-02 1.05652189e+00 2.79817849e-01
-5.85915685e-01 3.59837711e-01 -1.82135105e-01 7.89047539e-01
4.53378797e-01 -1.31658697e+00 4.97881562e-01 7.03006923e-01
5.12768388e-01 1.34009302e+00 3.04050773e-01 -1.04410827e+00
-6.00820780e-01 -1.12774324e+00 1.94569111e+00 -5.94597757e-01
7.74052262e-01 -4.06059958e-02 -1.02743709e+00 1.11435020e+00
2.48675793e-02 1.71429515e-01 1.05550909e+00 8.97716641e-01
-8.22360575e-01 1.42572835e-01 -1.15031934e+00 3.04947823e-01
1.20960307e+00 -7.69669235e-01 -1.24630499e+00 3.98748279e-01
1.02862954e+00 -3.81583691e-01 -1.30323315e+00 6.64174974e-01
8.81569982e-02 -1.08017601e-01 1.25524449e+00 -1.00312781e+00
4.03315127e-01 -5.26985884e-01 -2.61915505e-01 -1.12847292e+00
-2.73227334e-01 -1.10503897e-01 -7.34066427e-01 2.06124568e+00
8.74687910e-01 -6.55532479e-01 3.75091225e-01 6.33411288e-01
1.46702766e-01 -6.02238357e-01 -6.98323309e-01 -9.46347654e-01
1.22436387e-02 -3.87551636e-02 1.12632608e+00 1.34930277e+00
6.14726126e-01 7.47962058e-01 3.67404222e-01 7.43686140e-01
7.37164989e-02 5.62719524e-01 2.19646230e-01 -1.47407579e+00
2.81623065e-01 -1.24719635e-01 -5.86567760e-01 -9.33826208e-01
4.44239706e-01 -1.17545104e+00 -3.77745986e-01 -2.25095749e+00
2.51662284e-01 -5.99502742e-01 -3.68227661e-01 6.71240389e-01
-4.86353964e-01 -3.33982140e-01 -1.99928775e-01 2.19266802e-01
-8.51094246e-01 3.73107851e-01 9.04525280e-01 -2.52392322e-01
-1.04576960e-01 -1.60747677e-01 -1.04145002e+00 7.76652515e-01
5.60412884e-01 -6.19912624e-01 -3.35574895e-01 -2.98215389e-01
7.27639079e-01 -2.91380793e-01 -1.03246823e-01 -6.85821533e-01
4.45400566e-01 -2.45027095e-01 5.07607043e-01 -6.84943974e-01
5.83964139e-02 -8.79325747e-01 -1.10388190e-01 -1.93266392e-01
-5.57940364e-01 -1.61391512e-01 -9.26697701e-02 4.80927289e-01
-7.06580281e-01 -3.86202127e-01 1.25579074e-01 -1.35267898e-01
-8.27898383e-01 1.01075985e-01 1.07333727e-01 5.81812084e-01
8.55935633e-01 3.12153935e-01 -8.67653549e-01 -1.84882861e-02
-9.67272818e-01 3.49957258e-01 -1.32362008e-01 5.59721231e-01
4.67059910e-01 -1.30361700e+00 -5.19205213e-01 -2.76249617e-01
4.88444418e-01 3.50011051e-01 5.16781956e-02 1.25726923e-01
-2.17256919e-01 8.03423643e-01 3.34000796e-01 7.13061765e-02
-1.47775292e+00 6.51383340e-01 4.57231849e-02 -7.87417352e-01
-6.53010190e-01 1.00954103e+00 1.28079310e-01 -4.95530814e-01
-6.29463326e-03 -4.70971137e-01 -1.02594149e+00 4.41136211e-01
7.46314168e-01 2.60418922e-01 2.32340336e-01 -9.49481726e-01
-7.03667402e-01 6.34456694e-01 -3.46799821e-01 2.50832200e-01
1.25720870e+00 -3.50302577e-01 -5.42936504e-01 2.11200684e-01
1.14198661e+00 5.10475695e-01 -2.29113117e-01 -7.66881347e-01
8.20496202e-01 -2.77631819e-01 -2.27857694e-01 -8.31998706e-01
-8.22977722e-01 3.66540372e-01 -3.38718712e-01 4.32551771e-01
8.65102410e-01 8.33035290e-01 8.09851348e-01 5.97730517e-01
3.95862490e-01 -1.09427714e+00 -3.90006036e-01 9.08589184e-01
7.29676425e-01 -8.63872349e-01 4.98557329e-01 -1.27584398e+00
-5.88751554e-01 1.04939389e+00 6.00376248e-01 2.74341345e-01
7.88025081e-01 3.03640723e-01 -1.69240505e-01 -6.93852365e-01
-6.06079996e-01 -7.57364154e-01 5.83538532e-01 5.72835445e-01
7.17649937e-01 5.76457009e-02 -4.32201356e-01 1.09383631e+00
-1.69641942e-01 -4.34680879e-01 1.40013307e-01 1.18579233e+00
-4.15294826e-01 -1.23131514e+00 8.47309008e-02 3.10562968e-01
-8.19833338e-01 -3.48937929e-01 -6.73865438e-01 6.58354044e-01
3.52918893e-01 1.21421766e+00 1.55558005e-01 -1.61681056e-01
5.65972686e-01 3.29858124e-01 6.50247708e-02 -1.12022018e+00
-7.10288942e-01 -4.74869251e-01 7.07593501e-01 -2.70761579e-01
-5.48256576e-01 -5.36533594e-01 -2.01350498e+00 1.31567135e-01
-5.78951120e-01 7.20355511e-01 5.09803414e-01 1.25649226e+00
6.58849657e-01 8.82095635e-01 4.37982976e-01 2.07686946e-01
1.88385531e-01 -9.59827781e-01 -7.03048170e-01 4.40423042e-01
1.22039288e-01 -6.80217206e-01 -1.93796471e-01 6.36726990e-02] | [9.30034351348877, 8.640356063842773] |
64a734a2-d651-499a-9d8b-14754a831d01 | pose-guided-human-parsing-with-deep-learned | 1508.03881 | null | http://arxiv.org/abs/1508.03881v2 | http://arxiv.org/pdf/1508.03881v2.pdf | Pose-Guided Human Parsing with Deep Learned Features | Parsing human body into semantic regions is crucial to human-centric
analysis. In this paper, we propose a segment-based parsing pipeline that
explores human pose information, i.e. the joint location of a human model,
which improves the part proposal, accelerates the inference and regularizes the
parsing process at the same time. Specifically, we first generate part segment
proposals with respect to human joints predicted by a deep model, then part-
specific ranking models are trained for segment selection using both pose-based
features and deep-learned part potential features. Finally, the best ensemble
of the proposed part segments are inferred though an And-Or Graph.
We evaluate our approach on the popular Penn-Fudan pedestrian parsing
dataset, and demonstrate the effectiveness of using the pose information for
each stage of the parsing pipeline. Finally, we show that our approach yields
superior part segmentation accuracy comparing to the state-of-the-art methods. | ['Jun Zhu', 'Peng Wang', 'Alan Yuille', 'Fangting Xia'] | 2015-08-17 | null | null | null | null | ['human-parsing'] | ['computer-vision'] | [ 9.10065323e-02 5.85376322e-01 -4.05045778e-01 -6.67800605e-01
-1.04223895e+00 -5.89538574e-01 3.76104623e-01 1.07218839e-01
-4.70028311e-01 4.04430211e-01 5.15604496e-01 2.88357705e-01
2.87095755e-01 -7.21990764e-01 -9.24080431e-01 -2.52052337e-01
1.51812643e-01 1.04705024e+00 6.84250951e-01 -1.18408715e-02
1.19003482e-01 5.94476521e-01 -1.49553692e+00 1.60601586e-01
5.32079101e-01 9.51519132e-01 1.40411496e-01 6.48362815e-01
-5.02446033e-02 2.66640902e-01 -4.70321447e-01 -7.13064075e-01
3.07095379e-01 -4.70755726e-01 -1.09466600e+00 3.02305281e-01
6.29517257e-01 -3.54022115e-01 -1.24526275e-02 8.58506858e-01
4.68640894e-01 -3.36796716e-02 3.92741024e-01 -7.76279688e-01
1.60492778e-01 7.11519599e-01 -6.13754451e-01 -2.37516060e-01
4.32679445e-01 2.51704901e-02 1.22858608e+00 -6.77950680e-01
1.00092292e+00 1.57539034e+00 6.25111699e-01 7.31462777e-01
-1.10588086e+00 -2.57862747e-01 3.81866902e-01 -3.81954983e-02
-8.53028953e-01 -1.58968568e-01 9.26562607e-01 -3.57860178e-01
5.46507001e-01 -1.37091383e-01 1.09594202e+00 8.48655641e-01
8.33758190e-02 1.35299528e+00 6.95687532e-01 -2.99638987e-01
2.57279217e-01 -6.72998130e-01 4.48924661e-01 1.28021681e+00
3.46350819e-01 -2.11497635e-01 -4.76891696e-01 -1.48668587e-01
7.03686237e-01 -3.32651019e-01 3.37493867e-01 -7.88591325e-01
-1.13847589e+00 7.65376687e-01 4.80060101e-01 -2.31772110e-01
-5.99222362e-01 3.58159631e-01 4.67229635e-01 -5.43358684e-01
2.45663911e-01 2.55877435e-01 -7.25065172e-01 6.62959367e-02
-1.16467154e+00 6.33853078e-01 7.14377999e-01 8.46039295e-01
7.79153645e-01 -7.06068337e-01 -4.30642754e-01 8.03364992e-01
6.16721809e-01 1.93638936e-01 -1.39716074e-01 -1.45108020e+00
5.16120732e-01 8.04557621e-01 -5.02257869e-02 -2.79872805e-01
-8.18142831e-01 -3.04695249e-01 -1.87109813e-01 4.49272208e-02
6.49799526e-01 -1.67252377e-01 -1.29145515e+00 1.70914948e+00
9.73883331e-01 -3.31461608e-01 -3.76567125e-01 9.77954865e-01
7.15911627e-01 8.79412070e-02 5.79352736e-01 3.62786859e-01
1.99852407e+00 -1.32004666e+00 -3.19888800e-01 -4.62519437e-01
3.50895762e-01 -6.49968505e-01 5.79143941e-01 3.30910951e-01
-1.29567611e+00 -7.92811155e-01 -7.45467961e-01 -3.26203972e-01
5.77993765e-02 5.73709488e-01 6.61004961e-01 5.61871886e-01
-6.68634176e-01 8.57043445e-01 -1.32504499e+00 -5.45479119e-01
6.55891895e-01 3.76952380e-01 -4.94631439e-01 -1.12854309e-01
-7.32333004e-01 7.24167705e-01 5.45584202e-01 9.22906026e-02
-6.93077564e-01 -3.17531049e-01 -1.21410239e+00 -2.47098189e-02
3.36368203e-01 -1.24547923e+00 1.34609842e+00 -5.78415096e-01
-1.61829007e+00 1.04589295e+00 -3.15932304e-01 -4.81162816e-01
7.62696683e-01 -7.12537467e-01 3.09010804e-01 6.54266119e-01
4.62201089e-01 1.45096231e+00 5.96295118e-01 -1.08051038e+00
-7.50410676e-01 -5.44559896e-01 1.33547977e-01 1.86258733e-01
5.59515059e-01 5.84154986e-02 -1.14990997e+00 -3.96385968e-01
5.26165068e-01 -1.27177989e+00 -6.37993336e-01 3.49103391e-01
-9.04774070e-01 -5.08675992e-01 3.23492438e-01 -1.06272113e+00
8.33217204e-01 -1.73925650e+00 3.06604832e-01 3.60210508e-01
1.55704543e-01 4.37847786e-02 -8.96482840e-02 1.66939050e-01
2.61102796e-01 -2.71870270e-02 -3.94833982e-01 -7.53476560e-01
9.32513550e-02 3.13121676e-01 4.50845987e-01 5.80483556e-01
4.25720960e-01 1.15359616e+00 -6.82675064e-01 -1.11532414e+00
3.37647051e-01 2.48401150e-01 -7.57256150e-01 2.04464883e-01
-3.38401437e-01 5.78566313e-01 -8.14425290e-01 8.71259809e-01
5.99655986e-01 -2.05557600e-01 2.74984837e-01 -4.08158243e-01
1.10233119e-02 3.87616724e-01 -9.08142567e-01 2.19117737e+00
2.30906941e-02 -8.32140893e-02 4.53274921e-02 -6.76692843e-01
5.66760063e-01 -5.10063879e-02 4.69823867e-01 -2.90038913e-01
1.76000610e-01 -2.28445008e-02 -3.37504484e-02 -1.61880746e-01
2.78048456e-01 2.41484776e-01 -3.58019203e-01 1.34592369e-01
2.94040740e-01 2.15565078e-02 3.88146967e-01 9.51528102e-02
1.04135048e+00 1.07911444e+00 4.81364131e-01 -3.08417648e-01
3.51194531e-01 2.19994262e-01 7.41552353e-01 6.20959997e-01
-4.62388277e-01 8.33617985e-01 7.67042458e-01 -4.85528469e-01
-9.96453583e-01 -1.12308419e+00 -3.22641246e-02 1.20139205e+00
2.94449151e-01 -4.74358588e-01 -1.37837768e+00 -1.17527354e+00
-5.87420799e-02 5.23269355e-01 -8.34080935e-01 2.04489514e-01
-1.18627739e+00 -4.18266743e-01 4.91151303e-01 9.83309686e-01
4.49193358e-01 -1.14765120e+00 -9.55633342e-01 2.19066083e-01
-3.41090143e-01 -1.12312198e+00 -4.15866941e-01 1.13798909e-01
-1.05314291e+00 -1.32869983e+00 -8.60404432e-01 -7.26081073e-01
7.13823140e-01 -4.00905639e-01 1.33403802e+00 3.97853702e-02
-4.11609590e-01 2.95832157e-01 -3.32026422e-01 -6.46605296e-03
-2.22449169e-01 2.05964774e-01 -4.56691474e-01 -4.23797041e-01
-3.29117551e-02 -1.02939665e-01 -1.01176608e+00 2.89619744e-01
-1.91652477e-01 1.89181104e-01 9.93936479e-01 6.79031312e-01
9.98951495e-01 -5.29961348e-01 1.48627177e-01 -1.05206275e+00
-1.97785273e-01 -5.67530468e-02 -5.77172577e-01 3.47007781e-01
-7.20223933e-02 3.48905265e-01 1.38603762e-01 -6.95755780e-02
-1.08909535e+00 9.59627926e-01 -4.80814397e-01 -1.49216294e-01
-3.66949975e-01 1.94602236e-02 -7.13557959e-01 1.16793871e-01
2.61000574e-01 -1.26375467e-01 -1.34485215e-01 -7.91772664e-01
7.43682146e-01 -8.00631121e-02 6.94335222e-01 -9.28722024e-01
5.99792063e-01 4.88920540e-01 2.85851419e-01 -4.80857998e-01
-9.04987335e-01 -6.91739798e-01 -1.31651473e+00 -2.04065531e-01
1.39775825e+00 -8.98896158e-01 -6.08908653e-01 2.36504421e-01
-1.47367132e+00 -2.67633855e-01 -2.23940626e-01 3.67547989e-01
-6.81694329e-01 6.21053517e-01 -9.20978725e-01 -6.70211732e-01
-4.38868403e-01 -1.24516380e+00 1.90216875e+00 2.88109392e-01
-4.00831997e-01 -3.89153302e-01 7.58021995e-02 7.66494036e-01
-4.28382754e-01 3.77487987e-01 7.31913030e-01 -7.30172873e-01
-7.32394457e-01 -2.05436170e-01 -2.11226106e-01 -3.35853100e-02
-4.28747326e-01 -2.51757018e-02 -7.41025984e-01 1.79645523e-01
-7.74028420e-01 -1.57714665e-01 1.00787354e+00 6.71443045e-01
1.14248288e+00 8.68166983e-03 -7.04443693e-01 4.61722553e-01
9.35555279e-01 -3.63163024e-01 5.20800471e-01 2.75987029e-01
7.97524452e-01 9.17336464e-01 1.02234101e+00 4.57331210e-01
5.96672118e-01 6.27741337e-01 5.29934227e-01 -5.70178069e-02
-4.57209259e-01 -4.99937207e-01 2.99914517e-02 2.45790720e-01
-2.24324763e-01 -2.61182077e-02 -6.96696579e-01 5.02949297e-01
-1.75644720e+00 -6.46576822e-01 -1.84646383e-01 1.90663159e+00
5.37543595e-01 3.47546816e-01 6.32446766e-01 -2.42364645e-01
9.53162611e-01 6.52105436e-02 -4.85294610e-01 4.35059816e-02
2.67293990e-01 5.77153087e-01 5.44393718e-01 3.80920976e-01
-1.64145041e+00 1.40899336e+00 6.37706614e+00 5.06935835e-01
-4.83350039e-01 8.13298300e-02 6.05229080e-01 2.25295618e-01
4.83255908e-02 1.82798862e-01 -9.35372651e-01 8.92589316e-02
4.21531796e-01 6.11596406e-01 -5.86638339e-02 1.24442554e+00
-6.43276274e-02 -3.11408848e-01 -1.18187141e+00 5.73865414e-01
-8.04736093e-02 -9.01326239e-01 -1.19180381e-01 1.70603395e-01
2.10993633e-01 5.27202599e-02 -5.62607467e-01 2.88330644e-01
3.42118114e-01 -4.75709975e-01 1.21843815e+00 4.80140656e-01
1.59564316e-01 -7.87073493e-01 6.68551445e-01 3.01988661e-01
-1.51127589e+00 1.79145485e-01 -3.45595062e-01 2.74970859e-01
5.81651032e-01 5.25306702e-01 -7.75334954e-01 6.39524877e-01
6.08146310e-01 4.45222914e-01 -8.14499319e-01 1.08882105e+00
-7.35906124e-01 7.69042313e-01 -4.09590423e-01 1.45436600e-01
7.54452124e-03 -2.68602255e-03 4.11589772e-01 1.28575325e+00
2.86373347e-02 2.12919384e-01 6.62711561e-01 7.38902867e-01
-5.16692437e-02 6.49785101e-02 4.59166169e-02 2.26575881e-01
3.09338808e-01 1.55167079e+00 -1.28141379e+00 -4.62674141e-01
-1.62566274e-01 1.27217257e+00 5.40037036e-01 -7.59133250e-02
-8.50589216e-01 -6.91926852e-02 5.54930270e-01 1.48122683e-01
6.00444674e-01 -8.12812671e-02 -2.49952629e-01 -9.64712977e-01
1.00127034e-01 -4.21586961e-01 5.98973930e-01 -5.58582842e-01
-1.10143018e+00 3.91643703e-01 1.23078600e-01 -8.77416611e-01
-3.52748603e-01 -5.84953129e-01 -5.23172498e-01 6.00970566e-01
-1.14051759e+00 -1.72484899e+00 -1.12864979e-01 8.07175115e-02
6.37155712e-01 4.98884708e-01 5.88766098e-01 7.86934793e-02
-6.51347637e-01 4.64819103e-01 -8.59072268e-01 6.28252208e-01
4.18718636e-01 -1.42157245e+00 9.34191346e-01 9.15415227e-01
7.26254806e-02 6.89186156e-01 6.86687887e-01 -1.10236359e+00
-1.07678390e+00 -1.15846348e+00 7.75663257e-01 -4.79982376e-01
1.44898325e-01 -3.79690826e-01 -3.99203897e-01 7.29115069e-01
-1.33750916e-01 2.23314583e-01 4.71892238e-01 2.81093031e-01
-2.86614865e-01 2.45620728e-01 -1.19884169e+00 4.64361876e-01
1.38186467e+00 -2.95009394e-03 -7.26637244e-01 1.32597029e-01
6.59929514e-01 -6.04533613e-01 -8.28448355e-01 4.26510990e-01
9.26310956e-01 -9.54445004e-01 1.18967104e+00 -6.43763423e-01
3.83129299e-01 -2.92633504e-01 -8.35653916e-02 -7.59065568e-01
-5.77607632e-01 -1.88375980e-01 -1.12791114e-01 1.15825748e+00
4.19039428e-01 -2.21223477e-02 1.30447888e+00 6.07678354e-01
-3.81185263e-01 -8.20455432e-01 -9.88690197e-01 -4.23398256e-01
-1.47165701e-01 -1.99445099e-01 3.29465330e-01 1.47745326e-01
-2.79242426e-01 2.58926541e-01 -9.93083715e-02 2.95994431e-01
9.10541892e-01 3.18619311e-01 9.19556141e-01 -1.38058293e+00
-5.19424915e-01 -3.51202726e-01 -5.76816320e-01 -1.25145912e+00
3.81133229e-01 -7.03558028e-01 5.13477743e-01 -1.93914855e+00
4.50176299e-01 -1.23953283e-01 3.56661864e-02 8.00266683e-01
-5.25108159e-01 6.45464286e-02 3.76693964e-01 3.83391492e-02
-9.71980810e-01 3.07129264e-01 1.34152865e+00 9.02117789e-02
-2.87534855e-02 2.16509074e-01 -4.08873767e-01 1.06926107e+00
4.90965694e-01 -5.06793082e-01 1.71305284e-01 -8.85996073e-02
-2.82198399e-01 1.82440206e-01 5.24637580e-01 -1.14134622e+00
-5.90273105e-02 1.72722101e-01 7.94837594e-01 -1.16328406e+00
5.29442787e-01 -4.24218923e-01 -1.59350649e-01 8.09131980e-01
-2.91265517e-01 -1.90570399e-01 -1.78163260e-01 6.41262829e-01
1.93437994e-01 -2.10465759e-01 9.22289193e-01 -4.10550803e-01
-6.21582985e-01 3.23890775e-01 5.88767696e-04 7.23435953e-02
7.57962465e-01 -1.23716019e-01 1.15477517e-01 8.90945196e-02
-9.13384438e-01 3.34654450e-01 4.09661025e-01 1.66174710e-01
2.53611892e-01 -9.90785182e-01 -5.92956007e-01 -2.37840459e-01
1.83310732e-01 2.16728747e-01 1.91454887e-01 7.21367955e-01
-7.41809785e-01 1.54317945e-01 -1.90152168e-01 -8.43786955e-01
-1.43918347e+00 3.77644271e-01 1.78049281e-01 -5.49570501e-01
-7.45079517e-01 9.71658289e-01 3.67834777e-01 -5.74739516e-01
-1.47140268e-02 -4.08051014e-01 -2.10416973e-01 -1.17568359e-01
8.93350318e-02 5.02112329e-01 -1.46179959e-01 -8.96860242e-01
-7.02953637e-01 9.09600556e-01 6.49580806e-02 -3.41475219e-01
1.09519899e+00 4.99275215e-02 -3.48269120e-02 -1.40147537e-01
9.84810233e-01 1.92975342e-01 -1.52032149e+00 1.20542638e-01
2.89104968e-01 -3.41727808e-02 -4.13467288e-01 -9.30923402e-01
-1.09626400e+00 7.97762513e-01 4.41634655e-01 -5.30838907e-01
7.86276579e-01 6.05509579e-01 9.90165651e-01 1.95577100e-01
5.02914131e-01 -1.17301095e+00 -8.45925659e-02 2.41542637e-01
4.66862977e-01 -1.08673978e+00 2.32918233e-01 -9.77564871e-01
-6.58296883e-01 1.08795691e+00 5.89945316e-01 -4.23185945e-01
4.51859444e-01 -7.40312040e-02 -6.41423911e-02 -2.27080569e-01
-2.27179781e-01 -6.12488031e-01 6.17060721e-01 4.04484868e-01
5.12094140e-01 2.20998138e-01 -6.81013286e-01 8.62546861e-01
-4.05643314e-01 -2.24427834e-01 -1.30048439e-01 1.03718066e+00
-7.04408228e-01 -1.47131741e+00 -4.37093616e-01 2.69667357e-01
-5.49716294e-01 2.95059443e-01 -5.54231226e-01 9.01519537e-01
4.11035717e-01 6.40882015e-01 -1.51354477e-01 -8.14662874e-02
4.20355678e-01 2.42614791e-01 7.86636293e-01 -7.59292185e-01
-6.38648689e-01 4.08605486e-01 4.35101837e-01 -1.14327085e+00
-2.58048236e-01 -9.95517671e-01 -1.67970705e+00 4.66535509e-01
-3.35152477e-01 1.69324875e-02 7.89515555e-01 1.08057487e+00
3.89585376e-01 4.27853584e-01 4.23194887e-03 -1.34378123e+00
-3.12115639e-01 -8.30221713e-01 -3.44845891e-01 4.81066287e-01
-2.17040285e-01 -8.67765844e-01 2.44945601e-01 6.13550209e-02] | [8.329459190368652, -0.20324167609214783] |
ec48a7c2-1d99-459f-8443-8fd6fcbe4264 | heterogeneous-reconstruction-of-deformable | 2209.15121 | null | https://arxiv.org/abs/2209.15121v1 | https://arxiv.org/pdf/2209.15121v1.pdf | Heterogeneous reconstruction of deformable atomic models in Cryo-EM | Cryogenic electron microscopy (cryo-EM) provides a unique opportunity to study the structural heterogeneity of biomolecules. Being able to explain this heterogeneity with atomic models would help our understanding of their functional mechanisms but the size and ruggedness of the structural space (the space of atomic 3D cartesian coordinates) presents an immense challenge. Here, we describe a heterogeneous reconstruction method based on an atomistic representation whose deformation is reduced to a handful of collective motions through normal mode analysis. Our implementation uses an autoencoder. The encoder jointly estimates the amplitude of motion along the normal modes and the 2D shift between the center of the image and the center of the molecule . The physics-based decoder aggregates a representation of the heterogeneity readily interpretable at the atomic level. We illustrate our method on 3 synthetic datasets corresponding to different distributions along a simulated trajectory of adenylate kinase transitioning from its open to its closed structures. We show for each distribution that our approach is able to recapitulate the intermediate atomic models with atomic-level accuracy. | ['Frédéric Poitevin', 'Daniel Ratner', 'Nina Miolane', 'Gordon Wetzstein', 'Bongjin Koo', 'Axel Levy', 'Julien Martel', 'Ariana Peck', 'Youssef Nashed'] | 2022-09-29 | null | null | null | null | ['cryogenic-electron-microscopy-cryo-em'] | ['computer-vision'] | [ 1.51376292e-01 1.62574738e-01 2.99471974e-01 -9.72858965e-02
-5.25175452e-01 -5.53126752e-01 6.34728730e-01 8.02103952e-02
-4.11645323e-01 9.89210069e-01 1.38442993e-01 -3.21539164e-01
5.44635020e-03 -6.28045022e-01 -1.00517094e+00 -1.33765793e+00
-2.52061725e-01 9.85526979e-01 -1.04857154e-01 -2.51553744e-01
1.61795437e-01 6.76876247e-01 -1.07725346e+00 4.07742560e-01
4.67183381e-01 5.48624933e-01 4.31465924e-01 8.00431609e-01
1.51320770e-01 4.26153660e-01 -2.58683175e-01 -4.86790063e-03
6.54593930e-02 -4.98647749e-01 -9.59421933e-01 -3.35504822e-02
3.01142275e-01 -2.90796645e-02 -2.87791044e-01 8.38594437e-01
5.27861297e-01 2.06740007e-01 8.77199590e-01 -2.88499653e-01
-4.16107506e-01 1.49427444e-01 -9.32040885e-02 9.44322646e-02
3.13347876e-01 2.65712917e-01 9.49367523e-01 -7.05487430e-01
1.37738860e+00 1.07696283e+00 7.58669496e-01 5.49702644e-01
-1.84587216e+00 1.80417225e-01 -2.19356805e-01 1.29777238e-01
-1.22190762e+00 -1.73107311e-01 6.37609303e-01 -7.82181442e-01
1.17347896e+00 -1.00242198e-01 6.66179776e-01 1.09724128e+00
8.11886132e-01 -1.15273654e-01 1.10131192e+00 -3.37955683e-01
5.72585166e-01 -3.36759806e-01 1.10677317e-01 7.47129619e-01
2.22714990e-01 1.06683552e-01 -2.82737643e-01 -4.27475005e-01
8.03065836e-01 1.45620525e-01 -2.98084617e-01 -5.52084923e-01
-1.33073509e+00 6.14538908e-01 4.18810040e-01 2.67926931e-01
-6.50180817e-01 1.04639389e-01 1.06809370e-01 4.08208333e-02
-2.39959434e-02 5.08284271e-01 -4.81821299e-01 -1.38661951e-01
-5.87148845e-01 5.02909899e-01 7.69033849e-01 2.88530767e-01
9.15489256e-01 -3.26577693e-01 6.71653271e-01 1.12840042e-01
2.40337789e-01 8.58802721e-02 4.33770299e-01 -1.20483911e+00
-5.25653102e-02 5.17136693e-01 3.41119587e-01 -6.23453081e-01
-5.30537665e-01 1.17780723e-01 -9.56251979e-01 5.32105386e-01
5.28875709e-01 -2.53421277e-01 -6.42509639e-01 1.64560628e+00
4.60224420e-01 -2.19002753e-01 6.67368919e-02 8.88630271e-01
4.51869547e-01 8.21824789e-01 -1.04983091e-01 -3.22724730e-01
1.22189486e+00 -3.75725806e-01 -5.25821149e-01 4.00573224e-01
4.89551246e-01 -3.76681834e-01 5.30758560e-01 1.95764393e-01
-1.28790474e+00 -5.43324769e-01 -1.06190252e+00 -1.52940691e-01
-2.65973717e-01 -1.93328857e-01 4.60929900e-01 1.52442351e-01
-8.95043492e-01 1.23586822e+00 -1.32333529e+00 -1.96702674e-01
-5.61317690e-02 5.72607100e-01 -9.16977704e-01 5.03415763e-01
-1.06351662e+00 1.03703737e+00 4.60097462e-01 -1.68327317e-01
-6.90609753e-01 -6.34288311e-01 -5.69338322e-01 5.27010560e-02
-2.02846825e-01 -1.08322978e+00 9.00793493e-01 -6.00320339e-01
-1.55301738e+00 7.86895037e-01 -4.38114375e-01 -4.69181031e-01
3.64974618e-01 4.59167302e-01 7.36488178e-02 3.27965617e-01
-3.09969813e-01 7.07702219e-01 4.88962233e-01 -1.33495176e+00
1.36740610e-01 -5.24119496e-01 -1.71569899e-01 -3.13960947e-02
3.79702955e-01 -3.48334014e-01 3.13924909e-01 -2.02235639e-01
5.62967181e-01 -1.11357927e+00 -4.31195974e-01 -2.66261160e-01
-5.03466725e-01 1.87792033e-01 6.76497638e-01 -6.00341856e-01
6.99675977e-01 -1.83142757e+00 1.00051451e+00 1.24918178e-01
6.77356362e-01 -1.35705665e-01 5.28156221e-01 9.57192183e-01
-4.12630349e-01 3.73199508e-02 -3.86641800e-01 -3.46240819e-01
-3.13460715e-02 3.91337387e-02 -6.14590086e-02 7.51245260e-01
6.71991855e-02 7.56758094e-01 -4.91696745e-01 -8.16818699e-02
1.13437667e-01 1.00420988e+00 -5.98739743e-01 2.44903594e-01
-3.14134657e-01 8.32793176e-01 -4.46653008e-01 2.95103729e-01
7.83948839e-01 -5.20176232e-01 6.73267782e-01 -3.86198282e-01
-4.88717973e-01 3.48757803e-01 -9.09141898e-01 1.52428555e+00
-4.31384593e-02 3.77382547e-01 3.11300606e-01 -8.42295587e-01
5.61807156e-01 2.76344955e-01 6.62379503e-01 -3.57921720e-01
-8.50739703e-02 2.55063977e-02 4.01713312e-01 -4.34439033e-01
2.72820503e-01 -7.88088083e-01 1.45248264e-01 6.50917530e-01
1.12182766e-01 -5.82621507e-02 -2.15779990e-01 1.33061856e-01
1.04678917e+00 3.13109994e-01 3.69903266e-01 -5.22201657e-01
3.48846376e-01 5.28600588e-02 1.89287394e-01 4.99567717e-01
-1.91447124e-01 8.06053996e-01 6.70925736e-01 -1.19028080e+00
-1.71997154e+00 -1.01465797e+00 -3.91564757e-01 3.87759060e-01
1.22036755e-01 -4.62640196e-01 -1.17715013e+00 8.25707316e-02
1.96902558e-01 -3.70376594e-02 -1.02576137e+00 -2.77279615e-01
-5.74858308e-01 -1.11072671e+00 5.52668907e-02 -5.33216186e-02
4.93281987e-03 -1.13360357e+00 -9.49426174e-01 2.56681114e-01
-1.15436120e-02 -7.87015557e-01 2.91842204e-02 4.39692914e-01
-1.00031734e+00 -1.05799627e+00 -3.06553483e-01 -2.22328976e-01
5.64993262e-01 -1.55439973e-01 9.99320447e-01 -1.17829278e-01
-6.63729131e-01 3.77587713e-02 1.24423437e-01 5.65327406e-02
-8.44780445e-01 -1.64088115e-01 4.16663647e-01 5.78241274e-02
1.78042218e-01 -1.03516936e+00 -8.31330657e-01 -1.04808390e-01
-8.73658955e-01 7.79115558e-02 8.12834129e-02 7.90095150e-01
9.41615641e-01 -1.55933246e-01 -6.19316921e-02 -6.95011854e-01
4.45068061e-01 -4.65630561e-01 -3.36177319e-01 -1.40401408e-01
-1.60906702e-01 5.32725036e-01 8.10009241e-01 -7.58518577e-02
-9.82973456e-01 3.38283420e-01 -3.85126978e-01 -6.68200329e-02
-5.09197116e-01 3.25325191e-01 -1.88908428e-01 -1.44799501e-01
5.50222754e-01 5.35587311e-01 3.56926888e-01 -6.95237160e-01
1.96373239e-02 2.80604959e-01 4.60191280e-01 -8.89801264e-01
1.88105464e-01 1.02607036e+00 5.26178598e-01 -1.00399673e+00
-1.08351976e-01 -1.34127945e-01 -1.12200844e+00 -2.49276366e-02
9.35248494e-01 -6.48785293e-01 -1.42901182e+00 1.59128964e-01
-1.29837370e+00 -2.86718428e-01 -1.85592294e-01 4.30718601e-01
-9.91520464e-01 6.95863545e-01 -8.03465545e-01 -4.59157914e-01
-9.33685154e-02 -1.61931574e+00 1.41843688e+00 -1.69831797e-01
-3.81456882e-01 -1.04621053e+00 8.75531793e-01 9.51007158e-02
2.61585444e-01 5.35867810e-01 1.38756478e+00 -1.39598593e-01
-6.99318767e-01 -2.12860908e-02 3.58132899e-01 -1.73337802e-01
-1.78461187e-02 3.05013716e-01 -8.38363290e-01 -2.31091872e-01
2.08874807e-01 -1.63712502e-01 1.04102170e+00 7.06071913e-01
8.93447936e-01 -9.10333917e-02 -3.53599727e-01 6.74735308e-01
1.48715198e+00 1.61700666e-01 7.69779503e-01 4.42323893e-01
7.13646293e-01 6.92347944e-01 -1.41238943e-01 6.25956476e-01
1.29626885e-01 9.55161452e-01 5.23104131e-01 2.25395501e-01
2.16399163e-01 7.37506747e-02 3.86581421e-01 6.37074769e-01
-6.50604546e-01 2.69188173e-02 -8.41388881e-01 -6.65587708e-02
-1.55318034e+00 -1.35600305e+00 8.82822424e-02 2.15533590e+00
8.65126610e-01 -1.35962054e-01 1.88192487e-01 -1.07515663e-01
5.39973557e-01 9.50562879e-02 -7.05604196e-01 -5.11971891e-01
-1.49279326e-01 1.80034675e-02 1.20325774e-01 9.38281357e-01
-8.08430254e-01 5.74807882e-01 7.25265360e+00 1.49773017e-01
-1.19100368e+00 -9.10966098e-02 5.60817540e-01 3.27892303e-02
-2.86094040e-01 1.99197888e-01 -6.94896817e-01 8.33613276e-01
1.37858999e+00 9.16697383e-02 4.56421971e-01 5.01086593e-01
3.71634305e-01 -8.22262093e-02 -1.07642448e+00 7.08107114e-01
-5.41813374e-01 -1.97012055e+00 2.05414370e-01 3.65477741e-01
5.01101196e-01 2.66487241e-01 -7.82793611e-02 -3.37065399e-01
3.03403437e-02 -1.20228028e+00 5.27624607e-01 9.75466847e-01
7.54284739e-01 -5.45763969e-01 4.25587803e-01 5.03723145e-01
-7.08958030e-01 2.22097948e-01 -6.68811440e-01 -1.98147088e-01
3.47515196e-01 5.40340304e-01 -6.34105384e-01 8.49381760e-02
4.21231002e-01 4.24313843e-01 5.18684201e-02 8.30308646e-02
6.44020021e-01 1.41315371e-01 -3.29828352e-01 1.77130088e-01
1.13335878e-01 -8.54839444e-01 4.87356663e-01 1.01377594e+00
1.49736866e-01 3.66686225e-01 1.65728386e-02 1.13151884e+00
-1.04351155e-01 -4.70854193e-01 -6.65751994e-01 -1.35472298e-01
1.70236543e-01 1.04243720e+00 -5.91053009e-01 -1.92871347e-01
-3.53272744e-02 8.65090966e-01 4.94375944e-01 2.45690048e-01
-4.78560776e-01 -1.79442927e-01 1.04542911e+00 2.95934200e-01
3.13395739e-01 -6.09613419e-01 1.68264106e-01 -1.19301879e+00
-8.59811977e-02 -7.93772161e-01 -2.09207222e-01 -7.19303846e-01
-7.01584637e-01 5.06519973e-01 -3.30092669e-01 -5.57967246e-01
-3.13070923e-01 -9.07551944e-01 -4.45692331e-01 9.84051526e-01
-9.10899580e-01 -5.54726124e-01 -1.07456952e-01 1.17742077e-01
4.04159613e-02 2.62090061e-02 1.15218258e+00 -1.14135303e-01
-5.24275541e-01 -3.24759707e-02 7.59463072e-01 -4.10908908e-01
3.19528341e-01 -1.54921901e+00 6.80929840e-01 2.01745972e-01
-1.60211921e-01 1.11634243e+00 1.17927396e+00 -6.47402525e-01
-1.80656648e+00 -7.17810214e-01 7.02016532e-01 -8.30718577e-01
4.23533171e-01 -3.43814880e-01 -9.88942683e-01 7.19193876e-01
1.25415295e-01 7.03277960e-02 6.92391276e-01 -3.32862020e-01
-1.08820572e-01 4.35029596e-01 -1.15309834e+00 4.73896742e-01
6.76924229e-01 -8.09215069e-01 -4.72182930e-01 5.29081345e-01
4.73101288e-01 -2.79143423e-01 -1.16781282e+00 1.32509246e-01
8.31298292e-01 -1.54392862e+00 1.03142643e+00 -9.98826921e-01
4.58873391e-01 -3.07547212e-01 -2.57412344e-01 -1.06242085e+00
-4.77165908e-01 -9.25001979e-01 -2.72639990e-01 1.18702844e-01
2.37979546e-01 -4.13098305e-01 9.51903284e-01 5.36438525e-01
-5.55331223e-02 -8.93420160e-01 -1.11667204e+00 -2.76531726e-01
3.14192712e-01 1.56013504e-01 3.95461380e-01 7.45924294e-01
2.71573603e-01 1.85399890e-01 -2.52218340e-02 1.41007379e-01
5.55020392e-01 3.81719351e-01 4.29089129e-01 -1.13082254e+00
-5.96560478e-01 -1.85027551e-02 -5.79153419e-01 -8.68107200e-01
3.25064480e-01 -5.78161240e-01 -2.69099116e-01 -9.84323263e-01
6.63432598e-01 1.36108369e-01 1.07135043e-01 -1.17696360e-01
4.05053109e-01 -5.19289710e-02 -5.78891709e-02 4.82361913e-01
-3.41913462e-01 5.05869031e-01 1.11077857e+00 8.57091323e-02
-7.53391534e-02 -4.24328774e-01 -2.85407394e-01 6.10941410e-01
5.64359963e-01 -3.97332639e-01 1.75273404e-01 -5.12541458e-02
4.58650947e-01 2.99825788e-01 5.44535160e-01 -9.55803037e-01
4.96087410e-02 -2.91313175e-02 3.89165550e-01 -3.84377301e-01
5.67715466e-01 -6.42230868e-01 7.39865839e-01 6.33208156e-01
-4.26933646e-01 3.27516407e-01 -1.23790413e-01 7.84457386e-01
-1.43762976e-01 -5.94502948e-02 1.13040853e+00 -6.31181479e-01
-3.16149294e-02 2.63941288e-01 -6.68386340e-01 -4.53874290e-01
6.76435709e-01 -1.91285059e-01 -2.56630719e-01 -2.56843328e-01
-1.31435442e+00 -5.09334087e-01 1.31222928e+00 -5.69551766e-01
4.31874007e-01 -1.02341211e+00 -2.56319314e-01 4.80456144e-01
-3.28184128e-01 -7.35336766e-02 5.96810520e-01 7.65558124e-01
-1.14170218e+00 5.91441393e-01 -5.58649182e-01 -5.91495633e-01
-1.04726446e+00 5.30987620e-01 9.53673124e-01 -1.90576047e-01
-6.75473154e-01 5.02188802e-01 2.58912534e-01 -4.35408384e-01
-3.85177284e-01 -3.01585615e-01 6.24867901e-02 -2.83470422e-01
5.82175791e-01 3.12395155e-01 2.04566285e-01 -1.03859270e+00
-1.89573511e-01 7.86210060e-01 -1.75261781e-01 4.88742925e-02
1.47725058e+00 -1.06949545e-01 -4.44189310e-01 2.59837240e-01
1.27560878e+00 1.69323981e-02 -1.66967332e+00 2.14366674e-01
-2.11806342e-01 1.37124509e-01 -1.17308490e-01 -3.57480288e-01
-4.41808492e-01 1.13300598e+00 4.08549488e-01 2.76080251e-01
4.58757252e-01 1.26078501e-01 7.35264242e-01 8.92528355e-01
4.32144493e-01 -7.43463516e-01 -3.34008992e-01 4.83101517e-01
5.77959776e-01 -1.13425148e+00 3.47210057e-02 2.74911746e-02
-2.83633113e-01 1.45747304e+00 -1.37347594e-01 -2.56893963e-01
4.55369651e-01 3.45347166e-01 -1.42114550e-01 -4.29623097e-01
-9.96366501e-01 2.74549514e-01 -2.17078865e-01 5.16946077e-01
6.39221489e-01 9.31017771e-02 7.57966712e-02 3.19515765e-01
-1.48296999e-02 -4.45259273e-01 7.59991348e-01 8.81675839e-01
-6.88486636e-01 -1.25299931e+00 -3.29463601e-01 -1.37370229e-01
-5.94464600e-01 8.55785683e-02 -6.56967103e-01 7.34826207e-01
4.29122038e-02 1.25830218e-01 1.14976801e-01 -2.07074940e-01
3.42601091e-02 3.94080967e-01 7.83128798e-01 -4.50195819e-01
-1.61577731e-01 -8.99474397e-02 -3.51378083e-01 -6.80270135e-01
-3.83322686e-01 -8.18431199e-01 -1.57184422e+00 -5.53683460e-01
1.14250429e-01 3.99971932e-01 8.27629864e-01 8.45564544e-01
8.55910540e-01 1.97807640e-01 4.41611081e-01 -1.44042110e+00
-6.96955562e-01 -7.33061850e-01 -8.55502188e-01 6.48643613e-01
8.26768339e-01 -5.16726136e-01 -4.02044535e-01 3.62857699e-01] | [13.258028984069824, -3.068382501602173] |
e166b5f9-8c60-419a-80c3-638fe51800c9 | dime-maximizing-mutual-information-by-a | 2301.08164 | null | https://arxiv.org/abs/2301.08164v2 | https://arxiv.org/pdf/2301.08164v2.pdf | DiME: Maximizing Mutual Information by a Difference of Matrix-Based Entropies | We introduce an information-theoretic quantity with similar properties to mutual information that can be estimated from data without making explicit assumptions on the underlying distribution. This quantity is based on a recently proposed matrix-based entropy that uses the eigenvalues of a normalized Gram matrix to compute an estimate of the eigenvalues of an uncentered covariance operator in a reproducing kernel Hilbert space. We show that a difference of matrix-based entropies (DiME) is well suited for problems involving the maximization of mutual information between random variables. While many methods for such tasks can lead to trivial solutions, DiME naturally penalizes such outcomes. We compare DiME to several baseline estimators of mutual information on a toy Gaussian dataset. We provide examples of use cases for DiME, such as latent factor disentanglement and a multiview representation learning problem where DiME is used to learn a shared representation among views with high mutual information. | ['Luis Gonzalo Sanchez Giraldo', 'Austin J. Brockmeier', 'Jhoan Keider Hoyos Osorio', 'Oscar Skean'] | 2023-01-19 | null | null | null | null | ['mutual-information-estimation'] | ['methodology'] | [ 1.33835509e-01 3.37683439e-01 -3.03993467e-02 -5.09141207e-01
-1.11609948e+00 -6.03132069e-01 6.02989316e-01 -1.34389717e-02
-3.18919808e-01 5.79621136e-01 3.15753907e-01 -2.25468367e-01
-5.15766203e-01 -5.24928808e-01 -3.75793546e-01 -8.40534449e-01
-4.93910998e-01 5.17438054e-01 -5.62551677e-01 2.25193068e-01
3.05588037e-01 2.23776802e-01 -1.28078020e+00 -8.31921771e-02
7.49449313e-01 8.45219016e-01 1.97794405e-03 7.38242209e-01
2.80430526e-01 5.96613705e-01 -2.35937729e-01 -4.53458518e-01
3.72217536e-01 -6.68950379e-01 -8.92265737e-01 1.15960583e-01
2.45923057e-01 -5.42742275e-02 -7.11809099e-02 1.26641774e+00
1.38506621e-01 1.30019099e-01 1.63177502e+00 -1.59171438e+00
-4.14231688e-01 4.66676146e-01 -6.32560313e-01 -1.07450681e-02
5.68061352e-01 -3.17927986e-01 1.69300425e+00 -9.03607070e-01
6.13919675e-01 1.31837952e+00 6.19864345e-01 -3.97266187e-02
-1.90840113e+00 -3.89564097e-01 -5.25180340e-01 -1.26681730e-01
-1.28805840e+00 -2.79218018e-01 4.92744118e-01 -8.53852689e-01
5.15835702e-01 3.79165888e-01 2.86673903e-01 7.26737976e-01
2.13570684e-01 8.27549279e-01 1.20496953e+00 -3.75891834e-01
2.35297322e-01 3.84707808e-01 3.58511470e-02 7.48248816e-01
4.01047081e-01 5.78524470e-02 -2.97355294e-01 -6.17468596e-01
7.60503411e-01 -1.30867884e-01 -3.27933222e-01 -1.12541187e+00
-1.61515355e+00 1.24202573e+00 1.26205673e-02 2.85907835e-01
-1.62516877e-01 -1.25918835e-01 2.70625174e-01 4.68984157e-01
3.33778143e-01 7.01517105e-01 -1.92310706e-01 -1.25390381e-01
-7.70657957e-01 5.94597086e-02 1.28630638e+00 7.21349120e-01
1.18781126e+00 -1.82506725e-01 2.60119706e-01 6.57808185e-01
4.09940779e-01 7.70885408e-01 2.33797550e-01 -1.28614795e+00
3.35265875e-01 2.79706150e-01 4.72402982e-02 -1.16295767e+00
-2.83045769e-01 2.56460793e-02 -8.64372492e-01 2.66022176e-01
5.28766274e-01 -3.21182162e-01 -2.37916440e-01 1.98775303e+00
6.51955679e-02 -2.64183849e-01 2.05686375e-01 7.46715605e-01
2.10815266e-01 3.57952237e-01 -4.66852963e-01 -3.75631928e-01
9.50707912e-01 -4.52252388e-01 -6.77347958e-01 -4.96689975e-03
7.49717653e-01 -6.53120100e-01 4.92863148e-01 2.45509014e-01
-8.66056323e-01 3.21727172e-02 -1.04683530e+00 2.62304068e-01
-2.20818281e-01 2.42179427e-02 7.55248547e-01 7.24484026e-01
-9.60689604e-01 8.30282688e-01 -7.99437702e-01 -2.48986304e-01
-8.17717519e-03 2.35073239e-01 -9.21932697e-01 3.29306543e-01
-8.60171795e-01 1.02221358e+00 3.39941055e-01 -3.28789502e-01
-6.06025219e-01 -4.77363825e-01 -1.12164772e+00 -1.71349645e-02
2.33996302e-01 -6.03041947e-01 9.42606270e-01 -6.81975842e-01
-1.26056004e+00 8.63017321e-01 5.35991602e-02 -1.51465788e-01
2.30001077e-01 -6.91300035e-02 -1.03067368e-01 9.61088881e-05
2.79518276e-01 1.73653960e-01 8.62235010e-01 -9.63879943e-01
8.53219256e-03 -5.16577005e-01 -8.82802084e-02 2.69016683e-01
-4.56253104e-02 -3.08172286e-01 1.24958374e-01 -5.15655935e-01
5.10616243e-01 -1.08294380e+00 -2.30430499e-01 -2.40728594e-02
-7.95286596e-01 5.81153631e-02 4.46742982e-01 -4.81920779e-01
9.68919635e-01 -2.03587317e+00 7.25039363e-01 4.84562248e-01
4.25506264e-01 -5.10543883e-01 -1.12450503e-01 8.32695901e-01
-3.63499135e-01 1.82955667e-01 -5.29000223e-01 -8.60416889e-02
1.28099069e-01 3.04227206e-03 -2.76113544e-02 9.68814969e-01
-1.49026653e-02 4.61813152e-01 -8.09706867e-01 -3.16536784e-01
1.65330887e-01 2.67089516e-01 -5.74156225e-01 3.55685472e-01
3.09576571e-01 2.55219936e-01 -3.01495492e-01 4.17453945e-02
5.76663911e-01 -6.70874715e-01 4.35401738e-01 -3.36758792e-01
2.70702243e-01 6.39125705e-02 -1.54991305e+00 1.75440228e+00
-5.45067608e-01 8.56738985e-01 1.01236217e-01 -9.33965325e-01
6.30062282e-01 2.28105843e-01 6.27531588e-01 1.84835836e-01
1.87639505e-01 2.68389583e-02 -6.56402251e-03 -3.08915108e-01
2.81246930e-01 -4.12294447e-01 -1.82180420e-01 9.73639846e-01
3.76688033e-01 -2.70515889e-01 -6.61909953e-02 6.85770750e-01
1.09599125e+00 -1.59994602e-01 9.25104260e-01 -6.56129241e-01
2.53392398e-01 -7.14452744e-01 2.34867498e-01 5.84293902e-01
-2.91504543e-02 7.49715388e-01 7.62314081e-01 6.82851598e-02
-1.32554555e+00 -1.39999878e+00 -3.66075039e-01 4.51298296e-01
-6.13513254e-02 -7.69854009e-01 -5.79124629e-01 -6.53498709e-01
1.36299908e-01 5.92951357e-01 -6.34325027e-01 4.85927053e-03
3.83176535e-01 -6.71784937e-01 1.72189742e-01 2.02761352e-01
3.98412682e-02 -2.53505975e-01 -3.16468358e-01 -2.21718043e-01
-3.29369545e-01 -8.73836398e-01 -5.76624870e-01 3.20937395e-01
-6.88979924e-01 -1.31111062e+00 -5.26725531e-01 -1.41827777e-01
5.89670420e-01 1.40044421e-01 1.04200566e+00 -6.64162576e-01
-3.22725713e-01 8.96040678e-01 7.86286369e-02 -1.06550656e-01
-6.37685001e-01 -1.46270245e-01 3.69411856e-01 1.49403945e-01
3.88628751e-01 -7.57786691e-01 -4.89443570e-01 5.00860810e-01
-9.30502355e-01 8.74199793e-02 4.86236066e-01 9.48876023e-01
2.31781796e-01 -3.32426399e-01 4.01536554e-01 -8.36358726e-01
6.92464530e-01 -7.53746808e-01 -5.60249686e-01 2.65831888e-01
-6.42400205e-01 8.45925450e-01 1.41878739e-01 -2.81505406e-01
-7.43590832e-01 9.64823738e-02 5.15700817e-01 -3.68188620e-01
1.41556770e-01 6.44919753e-01 -6.38643187e-03 1.57268614e-01
6.65746748e-01 1.34445936e-03 3.81202608e-01 -2.99729228e-01
6.36714220e-01 6.65005386e-01 2.76627034e-01 -4.22269672e-01
6.85458958e-01 4.27617848e-01 3.11364502e-01 -7.63183415e-01
-8.65319729e-01 -7.24555731e-01 -7.17792153e-01 1.11007079e-01
8.84567440e-01 -8.98863077e-01 -1.04660714e+00 -2.91609503e-02
-7.96908557e-01 2.09107459e-01 -2.80415624e-01 9.96093094e-01
-1.18774748e+00 5.43098450e-01 -3.39487582e-01 -1.04745352e+00
1.21902928e-01 -9.58580673e-01 1.12827945e+00 -2.74972469e-01
-4.35278326e-01 -1.44987643e+00 5.83127260e-01 1.36637837e-01
2.88162678e-01 3.86689395e-01 8.58126879e-01 -9.17913020e-01
-2.74499655e-01 -3.32987219e-01 -1.88668326e-01 5.61171532e-01
2.81682730e-01 -2.01251477e-01 -8.06087136e-01 -4.28147107e-01
2.28622392e-01 -4.62537110e-01 8.45572174e-01 3.43092978e-01
6.79393232e-01 -4.63498205e-01 -1.81356221e-01 4.10360932e-01
1.45900285e+00 -2.38940418e-01 4.04155403e-01 -1.31865233e-01
6.60439789e-01 5.98201275e-01 2.23812371e-01 7.25303888e-01
2.86141336e-01 5.82709908e-01 1.64440066e-01 3.27940792e-01
5.65743923e-01 -1.82826281e-01 3.55016261e-01 1.22658360e+00
-8.78511146e-02 6.29875809e-02 -6.97135448e-01 3.44797999e-01
-1.81828439e+00 -1.12165618e+00 1.06330723e-01 2.65502334e+00
6.99383020e-01 -3.60832751e-01 1.13656037e-01 -1.31741121e-01
7.05372214e-01 3.74477923e-01 -4.76538271e-01 -1.61681578e-01
-1.22384466e-01 -2.22951874e-01 5.63964665e-01 7.31444120e-01
-1.23183370e+00 2.52350748e-01 7.52276754e+00 6.78681612e-01
-3.63057077e-01 -1.98355373e-02 3.50379199e-01 1.06673785e-01
-4.61694747e-01 2.83279270e-01 -3.34324032e-01 2.00230986e-01
1.04670596e+00 -3.70376408e-01 5.59836447e-01 9.59518433e-01
-2.74576426e-01 -3.22008908e-01 -1.60308397e+00 1.23535764e+00
1.77617520e-01 -9.24081385e-01 -2.66796827e-01 4.82998937e-01
8.67691398e-01 9.32619497e-02 1.23426877e-01 1.11162951e-02
7.72043109e-01 -1.07874572e+00 1.27795324e-01 6.38687968e-01
7.36169279e-01 -7.76276827e-01 4.48107541e-01 3.40536237e-01
-9.92683411e-01 2.82079905e-01 -4.52824473e-01 4.93758544e-02
3.84355076e-02 8.49506438e-01 -8.71476710e-01 4.97411340e-01
8.44680145e-02 8.88050139e-01 -2.70050347e-01 7.74324358e-01
3.38153429e-02 1.76068857e-01 -4.33033615e-01 1.91582307e-01
-3.03051490e-02 -7.88279414e-01 9.77617681e-01 9.76410151e-01
4.33569968e-01 -2.01591402e-01 3.38262841e-02 9.66596007e-01
1.30345508e-01 2.01763839e-01 -1.22400963e+00 -2.59401441e-01
3.06986868e-01 1.32954407e+00 -4.56501186e-01 -1.89976826e-01
-5.55104554e-01 1.17703009e+00 4.34816390e-01 3.83352429e-01
-4.73833561e-01 -2.67872721e-01 9.04033780e-01 -2.32383937e-01
3.60833764e-01 -3.78550023e-01 1.62913024e-01 -1.73503017e+00
-1.18355490e-01 -8.56961489e-01 3.12871814e-01 -5.40094316e-01
-1.63794315e+00 3.05133730e-01 2.53634572e-01 -1.45474315e+00
-8.77604783e-01 -6.31934524e-01 -2.58487523e-01 9.53470409e-01
-6.44967198e-01 -4.94455516e-01 3.15849006e-01 3.97326052e-01
1.70641840e-02 -2.58621901e-01 1.13912034e+00 -2.70914882e-01
-3.21427733e-01 2.85152465e-01 6.98518157e-01 1.04023628e-02
6.82926655e-01 -1.78807366e+00 9.46594775e-02 4.43604022e-01
5.30717432e-01 6.93351269e-01 9.35135543e-01 -4.11259711e-01
-1.57687342e+00 -5.33874452e-01 5.13445675e-01 -6.79190993e-01
9.82879460e-01 -3.32036167e-01 -5.22480011e-01 1.09829676e+00
1.43610269e-01 6.73425496e-02 1.08088315e+00 4.89484429e-01
-7.34930515e-01 3.50620896e-01 -9.39631879e-01 3.31931770e-01
6.98918879e-01 -9.65576351e-01 -3.01121533e-01 4.14083242e-01
2.60613084e-01 4.60993685e-02 -1.29549015e+00 2.62013853e-01
8.36255074e-01 -1.30478776e+00 8.71556282e-01 -7.04795480e-01
2.52576232e-01 -9.17881131e-02 -7.18625605e-01 -1.50042319e+00
-2.36544937e-01 -7.13182986e-01 6.38971254e-02 8.06504667e-01
5.34837306e-01 -6.60916269e-01 4.99805033e-01 8.85629654e-01
7.62567699e-01 -7.21853495e-01 -7.64512956e-01 -9.99567628e-01
-4.56390232e-02 -3.61220866e-01 2.07154769e-02 1.13200092e+00
4.93036866e-01 6.41276658e-01 -6.88783348e-01 -2.06369231e-03
1.09153414e+00 1.70710638e-01 7.52960563e-01 -1.45453370e+00
-5.99086642e-01 -1.37758777e-01 -9.90051866e-01 -8.37466240e-01
3.15317690e-01 -1.23325324e+00 -1.80593476e-01 -1.12890184e+00
8.04796994e-01 -6.93795308e-02 -1.01844512e-01 -9.54782665e-02
-2.03492846e-02 -1.46798640e-01 3.06112319e-01 2.54377723e-01
-6.34806693e-01 6.61582112e-01 8.88978839e-01 -2.72431993e-03
2.01196726e-02 9.09105837e-02 -5.38887143e-01 1.00357974e+00
4.18166727e-01 -5.65700591e-01 -4.92998153e-01 3.02009553e-01
4.67406809e-01 5.18585980e-01 2.58916795e-01 -7.79162645e-01
4.27715816e-02 -1.47883639e-01 1.55958951e-01 -4.12787795e-01
3.10723722e-01 -7.30826557e-01 4.55856413e-01 1.41965568e-01
-5.85128784e-01 9.99318436e-02 -3.74054283e-01 8.89088154e-01
-2.80549318e-01 -4.83446807e-01 6.67578161e-01 -1.12878859e-01
-1.78596288e-01 1.18010387e-01 -4.44499850e-01 2.58714229e-01
9.08576369e-01 6.52405098e-02 -1.35070294e-01 -1.03119922e+00
-1.02677166e+00 -1.36856362e-01 5.58142424e-01 1.03245834e-02
5.07314622e-01 -1.62845922e+00 -6.84299707e-01 1.25356868e-01
1.98266581e-01 -7.13016033e-01 1.14283361e-01 1.08810437e+00
-6.76022619e-02 3.39046419e-01 2.87747756e-02 -6.14242792e-01
-1.24522364e+00 5.92997432e-01 1.99658364e-01 -4.47901875e-01
-3.30018848e-01 4.52118665e-01 6.25510812e-01 -6.78192556e-01
-8.12230706e-02 -5.73656708e-03 1.12093359e-01 2.68775463e-01
5.57219863e-01 4.90047455e-01 -2.97138393e-01 -8.33864629e-01
-3.44532400e-01 4.19873446e-01 4.82119955e-02 -6.40728593e-01
9.90192354e-01 -4.45994526e-01 -2.62965977e-01 9.55839396e-01
2.00352192e+00 7.68939182e-02 -1.03013062e+00 -3.90306860e-01
-1.27300322e-02 -4.63502705e-01 -4.08534594e-02 -3.64765137e-01
-7.61808395e-01 7.30933547e-01 4.41512764e-01 4.15240198e-01
7.76269257e-01 2.17290193e-01 1.24101810e-01 6.06351316e-01
3.90955240e-01 -7.27275968e-01 7.46155232e-02 2.87626296e-01
8.60277355e-01 -1.59616125e+00 3.07761967e-01 -3.26873749e-01
-8.07669818e-01 1.17638755e+00 -7.14086369e-02 -9.74885747e-02
1.11431861e+00 2.75721312e-01 -3.30054015e-01 -2.50227422e-01
-7.00226605e-01 -1.72488555e-01 7.09889650e-01 5.02769828e-01
4.66667473e-01 3.57460558e-01 -8.48647505e-02 8.42655525e-02
-3.26098233e-01 -5.84871888e-01 6.11130536e-01 5.91239274e-01
-2.97255129e-01 -9.24387157e-01 -1.44112885e-01 7.50892699e-01
-4.41967905e-01 -9.05976743e-02 -2.31170073e-01 5.71043730e-01
-5.96421480e-01 7.36437261e-01 6.22313507e-02 -4.61370111e-01
-3.31506848e-01 1.59170598e-01 5.87145984e-01 -6.03924334e-01
1.97219804e-01 -1.00419633e-01 1.20095514e-01 -6.75235271e-01
-6.79786682e-01 -1.01689196e+00 -6.75467610e-01 -2.27109790e-01
-5.73472142e-01 2.92188048e-01 7.94894218e-01 8.62337530e-01
2.19647631e-01 -1.37337565e-01 8.23127568e-01 -7.79164433e-01
-8.01725268e-01 -9.24055338e-01 -1.21833646e+00 5.46728551e-01
4.41266328e-01 -6.34735048e-01 -8.91760170e-01 -8.96342769e-02] | [7.504778861999512, 4.158319473266602] |
690ebe5f-1925-4c8c-9045-6da7d9d47040 | domain-generalization-in-deep-learning-based | 2201.11620 | null | https://arxiv.org/abs/2201.11620v2 | https://arxiv.org/pdf/2201.11620v2.pdf | Domain generalization in deep learning-based mass detection in mammography: A large-scale multi-center study | Computer-aided detection systems based on deep learning have shown great potential in breast cancer detection. However, the lack of domain generalization of artificial neural networks is an important obstacle to their deployment in changing clinical environments. In this work, we explore the domain generalization of deep learning methods for mass detection in digital mammography and analyze in-depth the sources of domain shift in a large-scale multi-center setting. To this end, we compare the performance of eight state-of-the-art detection methods, including Transformer-based models, trained in a single domain and tested in five unseen domains. Moreover, a single-source mass detection training pipeline is designed to improve the domain generalization without requiring images from the new domain. The results show that our workflow generalizes better than state-of-the-art transfer learning-based approaches in four out of five domains while reducing the domain shift caused by the different acquisition protocols and scanner manufacturers. Subsequently, an extensive analysis is performed to identify the covariate shifts with bigger effects on the detection performance, such as due to differences in patient age, breast density, mass size, and mass malignancy. Ultimately, this comprehensive study provides key insights and best practices for future research on domain generalization in deep learning-based breast cancer detection. | ['Karim Lekadir', 'Laura Igual', 'Oliver Diaz', 'Socayna Jouide', 'Kaisar Kushibar', 'Lidia Garrucho'] | 2022-01-27 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 3.13836277e-01 2.48055264e-01 -3.68853807e-01 -5.08412242e-01
-1.02158189e+00 -1.32510096e-01 2.32427552e-01 4.87961978e-01
-4.84388530e-01 3.46594661e-01 9.89945792e-03 -8.95414591e-01
-2.70147979e-01 -7.86561489e-01 -8.86818588e-01 -7.41373003e-01
-3.36482599e-02 7.62759805e-01 3.64805877e-01 -1.33569613e-01
-2.27700338e-01 5.77296853e-01 -8.78738165e-01 7.57493377e-01
6.22856319e-01 9.61489081e-01 3.89822721e-01 3.32561940e-01
1.11713126e-01 4.97776449e-01 -4.60683316e-01 -3.85022163e-01
1.77824304e-01 -3.55243415e-01 -6.40769124e-01 -7.68855866e-03
5.23868859e-01 -6.81159914e-01 -5.63765347e-01 8.89340162e-01
8.09761643e-01 -5.65542281e-01 8.25214148e-01 -7.55286753e-01
-5.08068800e-01 6.57304466e-01 -3.93217772e-01 5.26656270e-01
-2.23888397e-01 1.98896870e-01 1.86343084e-03 -3.81367356e-01
6.05933070e-01 9.63427842e-01 1.23790467e+00 7.06784964e-01
-1.08249724e+00 -8.90050590e-01 -1.89890131e-01 1.32182479e-01
-9.08468902e-01 -6.32341877e-02 4.36725557e-01 -5.98701119e-01
6.98219478e-01 -9.70397964e-02 5.93061686e-01 1.38330340e+00
7.17780590e-01 6.19120061e-01 8.66609633e-01 -5.84042490e-01
2.71824569e-01 5.38592815e-01 -8.21076334e-02 5.48986018e-01
6.54078782e-01 1.53383598e-01 -1.10298865e-01 -3.15655380e-01
9.02362287e-01 3.16284373e-02 4.00918797e-02 -6.90185010e-01
-9.86602545e-01 1.01180696e+00 6.14841700e-01 6.51014090e-01
-3.35668236e-01 -1.80493429e-01 9.27848876e-01 5.84530123e-02
6.25825524e-01 3.89921814e-01 -5.80660343e-01 3.20514947e-01
-9.99625027e-01 8.22069272e-02 5.84662139e-01 8.42420161e-01
2.22645670e-01 -1.99420124e-01 -2.16374248e-01 7.65985429e-01
8.01720470e-02 4.19156224e-01 8.70585918e-01 -4.14998323e-01
2.81555593e-01 7.67579913e-01 -3.32368463e-01 -6.71665728e-01
-1.13274658e+00 -7.83223331e-01 -8.62775207e-01 -4.70513850e-02
5.80771387e-01 -1.12529352e-01 -1.31594825e+00 1.41053605e+00
2.41225302e-01 -1.41296297e-01 4.86712418e-02 6.93031132e-01
1.02914178e+00 -1.04559787e-01 4.81546909e-01 3.87498140e-02
1.64505935e+00 -5.13679683e-01 -4.70549136e-01 -5.04817307e-01
1.27064312e+00 -4.98251647e-01 6.02607667e-01 2.98321128e-01
-8.69520903e-01 -5.96471965e-01 -1.15035117e+00 6.91813380e-02
-3.72268617e-01 5.10420203e-01 7.41969585e-01 1.07029212e+00
-6.06774628e-01 6.05918229e-01 -1.39103949e+00 -8.29244137e-01
7.10324883e-01 4.01800543e-01 -2.12977380e-01 -2.95190901e-01
-1.19941318e+00 1.14462900e+00 5.81356227e-01 -2.38146529e-01
-9.26812828e-01 -1.35704195e+00 -8.33028674e-01 -2.10607678e-01
1.63674548e-01 -8.26800346e-01 1.65919280e+00 -8.69961381e-01
-8.63555908e-01 1.27547204e+00 2.12359890e-01 -8.24965954e-01
7.72805572e-01 6.25235215e-02 -5.90023279e-01 1.61761537e-01
3.39908540e-01 4.49339658e-01 4.01627064e-01 -9.00317073e-01
-7.03306556e-01 -7.13146329e-01 -4.75879133e-01 -1.20727271e-01
-5.05858421e-01 -6.00229353e-02 -3.29674661e-01 -5.54484189e-01
1.14065699e-01 -1.02306461e+00 -3.30626965e-01 1.84203729e-01
1.72044471e-01 8.63042697e-02 5.80523014e-01 -8.53552997e-01
1.04603159e+00 -2.31685066e+00 -4.31423813e-01 3.20979650e-03
2.92037666e-01 3.10042560e-01 2.11284533e-01 -4.51896079e-02
-2.71247089e-01 -1.07722856e-01 9.34577063e-02 -1.05833806e-01
-4.32354271e-01 -1.08339526e-02 4.27932411e-01 6.31516814e-01
2.89311051e-01 8.57026100e-01 -8.83390129e-01 -5.70183218e-01
3.55183184e-01 2.40191177e-01 -5.04520178e-01 -2.03532577e-01
5.92744760e-02 5.66209495e-01 -4.19573337e-01 6.75732374e-01
9.75268722e-01 -4.23705459e-01 4.64250594e-01 -5.08707881e-01
2.16850758e-01 1.29031688e-01 -7.32118428e-01 1.98224294e+00
-4.73116189e-01 5.48754394e-01 5.46876639e-02 -1.31502044e+00
8.05622756e-01 1.83807939e-01 7.50620425e-01 -9.58712995e-01
3.47947478e-01 5.75731814e-01 4.99459326e-01 -5.90401232e-01
2.28052765e-01 -5.18098354e-01 5.66737242e-02 6.79048384e-03
2.59414196e-01 6.61383644e-02 -1.26322195e-01 -2.40181293e-02
1.20654821e+00 -4.07746166e-01 3.63444358e-01 -5.56986153e-01
-1.63358133e-02 3.92580301e-01 4.71662283e-01 9.76476848e-01
-5.80405235e-01 5.83564699e-01 2.77626157e-01 -5.19042134e-01
-1.07516611e+00 -1.04234099e+00 -7.98036098e-01 1.06267738e+00
-1.13274530e-01 1.69692025e-01 -5.34504652e-01 -7.35190570e-01
2.67153621e-01 5.58051169e-01 -1.13965023e+00 -6.46150231e-01
-5.54213881e-01 -1.19297504e+00 7.46446013e-01 1.07641459e+00
5.06045818e-01 -6.13548398e-01 -9.02091801e-01 3.40856910e-01
-1.19242460e-01 -1.13616490e+00 1.30616814e-01 4.66181487e-01
-1.29832351e+00 -1.17124188e+00 -1.19736636e+00 -9.45631444e-01
4.72115397e-01 -4.13608849e-02 1.14655983e+00 -3.07460338e-01
-7.16748953e-01 2.99904734e-01 -3.66868198e-01 -9.73001659e-01
-1.10167229e+00 3.84726197e-01 -7.32062906e-02 -5.11081278e-01
8.70837331e-01 5.99589795e-02 -8.58351469e-01 2.99157441e-01
-8.99351478e-01 -1.20801970e-01 1.09011841e+00 8.59217525e-01
3.34362268e-01 -2.88327523e-02 5.93992829e-01 -1.14991117e+00
4.05276656e-01 -7.01759756e-01 -1.72730654e-01 2.76710093e-01
-5.55614293e-01 -6.88670427e-02 9.04961899e-02 -7.06914783e-01
-1.27497947e+00 4.59932722e-02 1.28535911e-01 -1.66009322e-01
-2.36931115e-01 4.56553578e-01 2.19820216e-01 -6.40963688e-02
1.42823827e+00 6.97897822e-02 2.26904199e-01 -3.80007416e-01
-2.61819154e-01 6.73441589e-01 4.88801301e-01 -3.35236847e-01
2.59149939e-01 7.01743305e-01 1.47155836e-01 -5.88159561e-01
-5.11172891e-01 -5.99823415e-01 -6.51361585e-01 6.75653443e-02
8.01129103e-01 -1.06242490e+00 1.18708558e-01 5.87625921e-01
-8.10229838e-01 -5.05503476e-01 -2.01542675e-01 7.00359643e-01
-3.17668915e-01 2.75150627e-01 -7.31460869e-01 -3.79585862e-01
-4.69364583e-01 -1.26072586e+00 1.24355531e+00 2.04826340e-01
-1.73808306e-01 -1.09390426e+00 -2.02453136e-01 1.28038943e-01
6.20162964e-01 2.66916722e-01 1.09974182e+00 -9.22588110e-01
1.86223984e-02 -3.06321591e-01 -3.82305652e-01 2.49221250e-01
3.23100060e-01 -4.24370229e-01 -8.20348442e-01 -5.76485455e-01
7.84811601e-02 -1.20853007e-01 9.35269952e-01 1.18880689e+00
1.27555835e+00 5.60374439e-01 -1.05222809e+00 6.91902578e-01
1.12511635e+00 4.76691902e-01 6.09647036e-01 8.58353734e-01
1.34710088e-01 3.38547468e-01 6.44386947e-01 3.06305349e-01
-7.46416971e-02 2.78814107e-01 3.02552015e-01 -8.39196324e-01
-2.66257167e-01 -8.55287090e-02 -2.49365225e-01 -1.80488408e-01
9.35282037e-02 1.88085884e-01 -1.25666678e+00 6.39296055e-01
-1.40465975e+00 -4.17786032e-01 -1.42495800e-02 1.94564247e+00
5.83088815e-01 3.29830706e-01 2.46680290e-01 4.63546328e-02
6.61127269e-01 -5.43656528e-01 -7.69280374e-01 -2.19249517e-01
3.15853395e-02 2.73576945e-01 9.42028284e-01 -2.43224919e-01
-1.30846035e+00 5.39395928e-01 6.81989813e+00 5.83274722e-01
-1.53069544e+00 2.43409231e-01 8.26878607e-01 7.99417421e-02
3.51028860e-01 -5.71970284e-01 -8.04591835e-01 1.76267460e-01
9.34927046e-01 7.95082152e-02 -4.94092554e-01 1.19599128e+00
-3.74848247e-02 -2.19900727e-01 -1.43840396e+00 7.03207433e-01
1.22455293e-02 -1.34799218e+00 -2.49770820e-01 1.14791669e-01
4.85617518e-01 2.49596968e-01 2.63007224e-01 5.63737869e-01
-2.12685719e-01 -7.81852424e-01 2.09727302e-01 5.95139824e-02
1.02673888e+00 -1.62528783e-01 1.08568621e+00 1.90704614e-01
-6.25184655e-01 -1.16602682e-01 -4.17280138e-01 2.65470445e-01
-2.62128204e-01 6.30039394e-01 -1.54922938e+00 5.72391808e-01
1.05537140e+00 2.33440191e-01 -9.16801751e-01 1.04339373e+00
5.24531066e-01 6.88127875e-01 -2.61135817e-01 1.31095633e-01
3.84648629e-02 6.37231946e-01 9.11163315e-02 1.70138407e+00
4.87516731e-01 4.36256826e-03 -9.27701592e-02 6.92668617e-01
8.09400678e-02 -1.26726523e-01 -4.49241281e-01 1.13145292e-01
2.45270029e-01 1.05674267e+00 -7.89252698e-01 -2.65971780e-01
-6.05661333e-01 6.49629951e-01 -2.81960607e-01 1.74708590e-01
-9.38739002e-01 1.06542315e-02 1.73187345e-01 6.83623910e-01
2.07603067e-01 2.62730986e-01 -4.40206945e-01 -7.20444798e-01
-9.85207334e-02 -9.31775868e-01 8.33087742e-01 -3.14206600e-01
-1.43699920e+00 2.50681728e-01 4.11324829e-01 -1.17245281e+00
-1.97940588e-01 -1.20577037e+00 -2.53537774e-01 7.75877237e-01
-1.47267401e+00 -1.25191748e+00 -4.90088791e-01 3.96234632e-01
4.18018103e-01 -3.81208956e-01 9.46709096e-01 5.78409612e-01
-1.50195718e-01 1.07974267e+00 3.82982552e-01 3.47392172e-01
1.07942283e+00 -1.01712942e+00 2.19086438e-01 3.05028677e-01
-6.89993441e-01 3.00730199e-01 4.83315527e-01 -6.90461934e-01
-1.04798424e+00 -1.21666980e+00 2.42098793e-01 -4.15375054e-01
5.32423496e-01 -2.31996864e-01 -1.05649328e+00 7.72390127e-01
-2.79701084e-01 1.17679708e-01 8.86924326e-01 1.72524169e-01
-4.34354879e-02 -1.54757902e-01 -1.52441990e+00 4.89008687e-02
8.61916244e-01 -3.05299349e-02 -4.44557160e-01 4.12293166e-01
4.10350770e-01 -8.80297184e-01 -1.08046997e+00 1.04870009e+00
5.98582268e-01 -9.10452724e-01 9.59585130e-01 -7.49543369e-01
6.21276379e-01 4.22403663e-01 6.52765483e-02 -1.19209278e+00
-6.62034452e-01 2.92439520e-01 2.29263783e-01 7.55629063e-01
5.69964826e-01 -6.87942445e-01 1.31703019e+00 5.92938721e-01
-2.93856233e-01 -6.24268055e-01 -1.04752541e+00 -7.79512405e-01
8.71593714e-01 -3.22064251e-01 3.03383350e-01 8.44964445e-01
-1.59599856e-01 -1.41336501e-01 2.49597311e-01 2.70283282e-01
3.80650133e-01 -2.68757552e-01 5.36320686e-01 -1.17266846e+00
-4.39475417e-01 -3.72748017e-01 -7.24955738e-01 -5.50288856e-01
-3.50939125e-01 -8.00173461e-01 -2.07572043e-01 -1.51918554e+00
6.12061083e-01 -4.53267604e-01 -5.22279859e-01 3.52855951e-01
-1.08717717e-01 -9.92990360e-02 -1.10904276e-01 1.10965734e-02
-2.03321576e-01 1.28445793e-02 1.30096948e+00 -3.99111807e-01
-1.86773732e-01 2.34710410e-01 -7.65681803e-01 7.14338839e-01
7.74758518e-01 -5.58950126e-01 -1.06014989e-01 -5.68536520e-01
-3.32940996e-01 3.16945799e-02 3.92136455e-01 -1.35529459e+00
1.34764478e-01 1.29858389e-01 9.64675725e-01 -6.21191800e-01
1.12525009e-01 -8.13427925e-01 -9.12932009e-02 9.98854160e-01
-1.07890524e-01 -4.48133826e-01 1.10054922e+00 4.49685097e-01
2.67579053e-02 -1.82728261e-01 9.46477413e-01 -2.60375291e-01
-5.80990791e-01 2.96708811e-02 -4.58792508e-01 -7.10165352e-02
1.10078943e+00 -3.58307511e-01 -2.02752978e-01 7.69761875e-02
-9.34031188e-01 5.45700267e-02 1.16817273e-01 4.92834151e-01
1.59142509e-01 -1.14914978e+00 -9.57562566e-01 1.23053677e-01
4.58842099e-01 1.55810490e-01 6.43782258e-01 9.56192613e-01
-5.03478408e-01 6.20859802e-01 -2.42237449e-01 -1.06171501e+00
-1.17503238e+00 5.25987089e-01 9.33274627e-01 -6.57189250e-01
-5.36262572e-01 8.35554361e-01 5.29402792e-01 -5.34037888e-01
3.07345301e-01 -7.26934373e-01 2.65825957e-01 -2.05754623e-01
3.90506715e-01 2.98926294e-01 7.10597932e-01 3.67805809e-02
-5.53279281e-01 1.82654396e-01 -6.28323436e-01 3.46503854e-01
1.14111710e+00 3.22833091e-01 5.68387628e-01 5.02941087e-02
1.11360312e+00 -7.83990920e-01 -9.09889877e-01 -2.29896724e-01
-1.32788017e-01 -1.21529266e-01 2.12180480e-01 -1.12059581e+00
-9.83292818e-01 8.55834246e-01 1.53977060e+00 -1.18734166e-01
1.39810538e+00 4.67303425e-01 5.06807685e-01 1.71227351e-01
2.23200813e-01 -1.06549191e+00 1.24848530e-01 -2.44246200e-02
6.80887163e-01 -1.77690148e+00 1.48279384e-01 -4.48467851e-01
-5.51385105e-01 1.16557288e+00 7.22226858e-01 1.63041189e-01
7.11990118e-01 3.76510501e-01 2.14609727e-01 -3.24645817e-01
-5.78954369e-02 2.36763299e-01 1.11053087e-01 1.04577541e+00
6.48228765e-01 2.22580731e-01 -3.39610547e-01 8.53969812e-01
1.47240823e-02 5.59737086e-01 1.89725325e-01 1.06851339e+00
-2.18171269e-01 -8.82562697e-01 -5.82010627e-01 8.09725761e-01
-6.18331552e-01 2.13980436e-01 -1.89660668e-01 1.27961886e+00
4.21678752e-01 5.98743498e-01 3.50465365e-02 -8.55225921e-02
6.32484019e-01 8.16436112e-03 6.73252404e-01 -8.27518165e-01
-4.36703324e-01 4.15077172e-02 -9.34233591e-02 -1.90635517e-01
-2.29709819e-01 -8.31640244e-01 -1.19388664e+00 1.25000343e-01
-5.09114444e-01 -3.97126496e-01 7.88130939e-01 7.03433514e-01
3.41822773e-01 1.02857161e+00 -9.86838490e-02 -6.53352916e-01
-8.95296097e-01 -1.34017313e+00 -5.83808243e-01 5.59477687e-01
1.81891292e-01 -7.81856239e-01 -2.17088554e-02 -3.29881720e-02] | [15.211512565612793, -2.5045886039733887] |
26e0fb40-b28f-4521-a2b7-34a9355e6c67 | exploring-resiliency-to-natural-image | 2303.09283 | null | https://arxiv.org/abs/2303.09283v1 | https://arxiv.org/pdf/2303.09283v1.pdf | Exploring Resiliency to Natural Image Corruptions in Deep Learning using Design Diversity | In this paper, we investigate the relationship between diversity metrics, accuracy, and resiliency to natural image corruptions of Deep Learning (DL) image classifier ensembles. We investigate the potential of an attribution-based diversity metric to improve the known accuracy-diversity trade-off of the typical prediction-based diversity. Our motivation is based on analytical studies of design diversity that have shown that a reduction of common failure modes is possible if diversity of design choices is achieved. Using ResNet50 as a comparison baseline, we evaluate the resiliency of multiple individual DL model architectures against dataset distribution shifts corresponding to natural image corruptions. We compare ensembles created with diverse model architectures trained either independently or through a Neural Architecture Search technique and evaluate the correlation of prediction-based and attribution-based diversity to the final ensemble accuracy. We evaluate a set of diversity enforcement heuristics based on negative correlation learning to assess the final ensemble resilience to natural image corruptions and inspect the resulting prediction, activation, and attribution diversity. Our key observations are: 1) model architecture is more important for resiliency than model size or model accuracy, 2) attribution-based diversity is less negatively correlated to the ensemble accuracy than prediction-based diversity, 3) a balanced loss function of individual and ensemble accuracy creates more resilient ensembles for image natural corruptions, 4) architecture diversity produces more diversity in all explored diversity metrics: predictions, attributions, and activations. | ['Michael Paulitsch', 'Pablo Munoz', 'Rafael Rosales'] | 2023-03-15 | null | null | null | null | ['architecture-search'] | ['methodology'] | [ 1.43469078e-02 -3.56763095e-01 1.42230913e-01 -2.34782130e-01
-2.93804079e-01 -6.03618741e-01 7.31134772e-01 1.43288717e-01
-4.55941945e-01 5.66291094e-01 1.78890646e-01 -3.58328879e-01
-3.82844567e-01 -6.52512312e-01 -5.81688583e-01 -7.68210232e-01
-2.76199847e-01 -6.48263097e-02 -4.67719845e-02 -3.56958330e-01
5.80910206e-01 7.92374551e-01 -1.81187630e+00 3.00022840e-01
1.02509642e+00 1.10087883e+00 -2.64496654e-01 7.73958981e-01
2.96384782e-01 8.71720195e-01 -1.02806675e+00 -3.01627457e-01
5.33721924e-01 -4.77103472e-01 -5.85726261e-01 -5.95116913e-02
7.24257410e-01 1.15163393e-01 -9.12786871e-02 8.56894135e-01
7.99667835e-01 -1.01486817e-01 9.70358491e-01 -1.52044690e+00
-6.90032721e-01 4.89326298e-01 -3.91568422e-01 5.96787155e-01
7.25884587e-02 8.16748500e-01 6.98608994e-01 -2.83310860e-01
4.21548963e-01 1.34211373e+00 8.86211276e-01 2.69897103e-01
-1.63339090e+00 -7.98539758e-01 6.34425431e-02 1.36823103e-01
-1.30795920e+00 -4.57106948e-01 5.33816755e-01 -8.00267518e-01
1.16813076e+00 2.85368860e-01 3.59114349e-01 1.33585036e+00
7.82825828e-01 -1.11488216e-01 1.39465332e+00 -3.76427233e-01
4.27811384e-01 3.76083821e-01 2.16799915e-01 3.95581692e-01
4.63762701e-01 4.90364522e-01 -3.09345752e-01 -1.28718719e-01
4.03038204e-01 -2.88265765e-01 -1.82263628e-01 -7.45449662e-02
-8.73000622e-01 7.33885765e-01 5.76050580e-01 3.98907840e-01
-4.23951387e-01 3.01704973e-01 5.65887332e-01 9.55963910e-01
2.46249735e-01 1.17121387e+00 -4.12150472e-01 1.35174155e-01
-7.25876451e-01 3.04693222e-01 5.99900186e-01 2.03714088e-01
6.57539606e-01 5.34298539e-01 -2.36489013e-01 7.04137921e-01
-1.37261748e-01 2.50025302e-01 7.07333326e-01 -8.90693069e-01
1.87175393e-01 7.37621248e-01 -2.65195340e-01 -1.29544687e+00
-5.83118677e-01 -9.48430896e-01 -1.08467090e+00 6.68020070e-01
6.76508546e-02 -1.34536937e-01 -7.71747410e-01 1.87302530e+00
-1.63168833e-01 -4.43397343e-01 9.79486182e-02 4.72302020e-01
4.52856272e-01 2.58843869e-01 2.77139425e-01 -1.42358303e-01
9.12404716e-01 -2.92897463e-01 -1.49618015e-01 8.60862136e-02
6.06162012e-01 -5.17231286e-01 1.04909122e+00 4.25186038e-01
-6.89785302e-01 -8.03492904e-01 -1.24455118e+00 5.44842184e-01
-3.45802277e-01 -1.71785727e-01 6.65157065e-02 1.07514048e+00
-1.11000717e+00 7.97549188e-01 -4.41604555e-01 -3.30544055e-01
4.20639515e-01 4.12056178e-01 -1.29896387e-01 3.17234784e-01
-9.07851279e-01 1.23861074e+00 4.20767993e-01 -3.88289571e-01
-8.96866143e-01 -7.17199922e-01 -4.09324557e-01 3.48744988e-02
-2.70207107e-01 -8.81367505e-01 5.37620842e-01 -1.49213517e+00
-1.09296048e+00 5.89242876e-01 3.38824749e-01 -7.09035814e-01
5.97088635e-01 8.15979540e-02 -5.58943391e-01 -4.11134332e-01
-1.09877147e-01 8.46269369e-01 9.37278211e-01 -1.46721935e+00
-2.92833745e-01 -2.88113117e-01 -8.47567543e-02 -8.92915055e-02
-5.75488508e-01 -2.21440479e-01 7.31517017e-01 -6.55068278e-01
-2.67411262e-01 -1.03291094e+00 -1.34240836e-01 -4.10155952e-01
-2.80162811e-01 1.43053338e-01 6.17083788e-01 -4.21010107e-01
1.44996262e+00 -2.09045839e+00 1.58266023e-01 4.10906345e-01
2.54097402e-01 1.86691850e-01 -3.90407562e-01 3.07364255e-01
-3.45374525e-01 4.76185769e-01 -1.17493063e-01 1.72874153e-01
-2.43699491e-01 -3.82629856e-02 -3.92504454e-01 4.22905087e-01
4.48082387e-01 5.10309100e-01 -3.23038280e-01 -1.67653337e-01
1.13198161e-01 4.20554012e-01 -7.30462015e-01 -1.14750244e-01
-9.31813568e-02 5.53837776e-01 1.08400285e-01 5.96354783e-01
4.92864877e-01 -3.38834003e-02 1.34655416e-01 -1.56199977e-01
-2.05313936e-01 -6.87671676e-02 -8.15814674e-01 6.76723838e-01
-6.73884511e-01 7.58798838e-01 -6.15406394e-01 -6.28734231e-01
1.31360555e+00 -1.21434055e-01 1.48442060e-01 -1.00437129e+00
2.13024750e-01 2.90283144e-01 5.65981150e-01 -1.57305792e-01
5.30506432e-01 4.07340452e-02 -2.28644654e-01 5.03774881e-01
1.10248752e-01 1.54291406e-01 1.83074661e-02 -7.82272965e-02
1.30332363e+00 -2.73550719e-01 3.83702219e-01 -6.17792368e-01
3.96368504e-01 -8.25806409e-02 3.72224540e-01 9.57937062e-01
-2.19847575e-01 4.28550303e-01 8.60581636e-01 -7.91788459e-01
-1.47556067e+00 -9.40246820e-01 -2.53401250e-01 8.90689433e-01
-2.99399886e-02 -1.40802905e-01 -6.12700105e-01 -6.33223534e-01
4.75599132e-02 8.87436807e-01 -8.24655473e-01 -6.01856887e-01
-4.66456652e-01 -1.05303848e+00 8.11395407e-01 3.19002837e-01
6.28526330e-01 -9.83504117e-01 -9.52898324e-01 -1.67851746e-01
2.97757000e-01 -6.42584383e-01 -1.91556849e-02 2.77942270e-01
-9.62581336e-01 -9.42024469e-01 -1.79274619e-01 -1.75449654e-01
4.74637836e-01 -1.11647427e-01 1.43443847e+00 5.08693695e-01
-1.89233407e-01 4.44475919e-01 -2.40862414e-01 -4.08988863e-01
-8.33985627e-01 1.77986220e-01 3.86940211e-01 -2.09973469e-01
-1.80244386e-01 -9.01715696e-01 -5.99431455e-01 3.53195667e-01
-9.42879498e-01 -5.22798479e-01 6.71347558e-01 8.60073209e-01
2.80155838e-01 3.03959727e-01 7.01399207e-01 -3.68643880e-01
1.05797148e+00 -5.91110289e-01 -4.91097927e-01 3.17118555e-01
-1.20531583e+00 3.27390134e-01 6.82448983e-01 -6.60697937e-01
-5.61987758e-01 -1.68790340e-01 1.20566539e-01 -5.11311650e-01
-1.99160531e-01 2.88733929e-01 -3.38106789e-02 -3.51451516e-01
1.43309879e+00 8.72546211e-02 1.23900831e-01 -5.99125847e-02
1.91018552e-01 3.10720354e-01 2.42005035e-01 -5.36474764e-01
4.97166276e-01 4.35972586e-02 2.22825080e-01 -8.01438808e-01
-1.25524223e-01 2.94502109e-01 -4.86082256e-01 -4.21282679e-01
4.58261281e-01 -6.88337922e-01 -6.51938856e-01 6.72465920e-01
-1.05735195e+00 -2.48641327e-01 -2.48743013e-01 8.37721974e-02
-3.35766017e-01 1.28824383e-01 -4.48626988e-02 -7.46609986e-01
-4.25295442e-01 -1.38615000e+00 4.37484473e-01 8.57423320e-02
-4.61033255e-01 -9.01398420e-01 1.07873045e-01 -3.49542648e-02
7.91764438e-01 7.48030722e-01 1.21542633e+00 -7.38637030e-01
-3.70434135e-01 -1.26395911e-01 -5.22679910e-02 6.57484531e-01
-1.56648129e-01 2.57608950e-01 -8.69081259e-01 -3.12655061e-01
-5.45050129e-02 -2.23042026e-01 9.54289973e-01 3.88508022e-01
9.78970885e-01 -6.56034470e-01 8.54397193e-02 4.39330071e-01
1.60242999e+00 1.46005630e-01 9.41993535e-01 7.57158339e-01
4.23414588e-01 4.75249231e-01 -7.34494179e-02 4.40627635e-01
8.86227116e-02 6.32380784e-01 5.55234075e-01 1.39161751e-01
-2.05422774e-01 1.52484819e-01 5.48828781e-01 4.15040791e-01
5.22106625e-02 -4.39833045e-01 -1.21937227e+00 3.64939511e-01
-1.31964195e+00 -1.04493248e+00 1.20392153e-02 2.28669786e+00
5.51135361e-01 3.03228498e-01 4.37663674e-01 2.90935814e-01
6.28149271e-01 2.10166089e-02 -5.11489153e-01 -7.53475547e-01
-4.47728515e-01 -5.57338409e-02 6.11127019e-01 3.34570140e-01
-8.17083895e-01 4.18055236e-01 7.08324242e+00 5.61927021e-01
-1.35902393e+00 7.66589015e-04 1.15014935e+00 -4.34076071e-01
-3.23238105e-01 -7.06447661e-02 -6.00592613e-01 7.03438044e-01
1.17430019e+00 1.59243625e-02 2.12766960e-01 8.40880692e-01
1.58859100e-02 3.20236594e-03 -1.05058539e+00 7.93699980e-01
1.00232095e-01 -1.43899822e+00 2.66498119e-01 3.99342090e-01
1.00334740e+00 7.41387457e-02 5.17643631e-01 1.61201179e-01
3.55235994e-01 -1.30209339e+00 8.84117663e-01 8.40044260e-01
4.71365780e-01 -9.40007210e-01 7.54342854e-01 2.47077048e-01
-5.42025805e-01 -7.49109685e-01 -3.90017517e-02 -2.16332689e-01
-4.21445310e-01 6.19039059e-01 -6.74689710e-01 1.21884957e-01
8.04070055e-01 1.19516648e-01 -1.10947299e+00 8.72561574e-01
1.67190716e-01 5.34945071e-01 -1.18257865e-01 9.51720774e-02
7.07300305e-02 2.66906679e-01 6.73396885e-01 1.20221937e+00
2.27358729e-01 -2.63266534e-01 -2.02689424e-01 9.31627035e-01
-1.75216701e-02 -1.90374255e-01 -5.61384857e-01 1.97121888e-01
8.08251381e-01 7.56161153e-01 -5.55174112e-01 -1.86759494e-02
4.03115302e-01 6.47078991e-01 1.71267077e-01 1.57249346e-01
-6.69893861e-01 -4.75121513e-02 8.39115083e-01 1.05309874e-01
1.09128401e-01 -1.23695292e-01 -9.37642157e-01 -6.42538309e-01
-1.34918973e-01 -1.28763413e+00 2.26889297e-01 -5.45996428e-01
-1.36999297e+00 1.03334343e+00 1.28575563e-02 -1.21060407e+00
-2.85188764e-01 -2.80665070e-01 -7.43769884e-01 9.29319322e-01
-1.03560054e+00 -7.87454545e-01 -3.75558168e-01 1.76669225e-01
2.02907443e-01 -7.19563425e-01 6.91959202e-01 -2.62642279e-02
-7.26902604e-01 9.95950878e-01 8.99676904e-02 -2.23445758e-01
3.95306408e-01 -9.18711722e-01 2.19453380e-01 1.00606894e+00
-3.45424823e-02 4.69262749e-01 9.27446425e-01 -6.09691083e-01
-8.50064039e-01 -1.15385437e+00 5.11505544e-01 -8.36457312e-01
3.44412029e-01 1.67244628e-01 -8.14993501e-01 2.30846792e-01
1.61527500e-01 -2.97216058e-01 6.10538125e-01 3.77532430e-02
-8.34817111e-01 -3.36340904e-01 -1.31580102e+00 4.75941271e-01
7.51300335e-01 -3.72207582e-01 -4.65470925e-02 -1.19516052e-01
6.74239278e-01 -2.04894990e-02 -1.10290420e+00 5.84851801e-01
7.45951533e-01 -1.33736038e+00 8.68561924e-01 -4.51691359e-01
6.80143476e-01 -1.91484109e-01 -5.39527714e-01 -1.42093098e+00
-6.23461127e-01 -2.01306343e-01 2.15677604e-01 1.06841362e+00
5.75349331e-01 -6.66428149e-01 2.84999281e-01 4.94687140e-01
-1.61206350e-02 -8.92325699e-01 -1.05362129e+00 -1.03581929e+00
3.76039743e-01 -1.62937939e-01 8.01806748e-01 1.00810647e+00
-6.60348535e-01 2.52913713e-01 -2.39686266e-01 9.43799540e-02
5.23772418e-01 -3.52999687e-01 7.65844584e-01 -1.23475587e+00
-2.39770547e-01 -9.82963622e-01 -9.34903860e-01 1.50208279e-01
1.08080588e-01 -6.57174706e-01 -3.90349090e-01 -8.95138025e-01
1.70327425e-02 -7.53001273e-01 -3.89743119e-01 5.65070510e-01
3.90975783e-03 2.73023814e-01 4.17408854e-01 6.03011906e-01
-3.89160693e-01 3.43068272e-01 4.88286108e-01 -7.52168745e-02
-4.11019593e-01 -3.25204521e-01 -8.83394241e-01 3.92896175e-01
9.23860073e-01 -6.62021041e-01 -3.85066330e-01 -3.91021252e-01
4.20878112e-01 -2.75316060e-01 6.77152336e-01 -1.53525805e+00
-1.25524044e-01 -1.18910365e-01 5.90467632e-01 4.98294011e-02
-1.17121696e-01 -6.37727559e-01 5.36025703e-01 7.59306431e-01
-5.82899451e-01 4.50252473e-01 3.34863633e-01 4.24908549e-01
2.00975537e-01 2.26862300e-02 1.08908927e+00 3.08061410e-02
-6.20024085e-01 -1.95186734e-01 -2.32543245e-01 -1.09251261e-01
8.66269827e-01 -4.96065706e-01 -5.10391355e-01 -1.13794319e-01
-4.42470551e-01 -2.43858665e-01 8.10379744e-01 6.01721823e-01
2.97431260e-01 -1.15707493e+00 -9.99283552e-01 5.08040152e-02
2.25453358e-02 -7.91625202e-01 3.27788889e-01 8.72610211e-01
-4.23833728e-01 1.22439191e-01 -7.41800249e-01 -7.55427420e-01
-1.32354832e+00 3.96446645e-01 7.99919069e-01 -2.09622771e-01
-6.67358488e-02 8.14183116e-01 -1.31019294e-01 -2.64956295e-01
9.35405418e-02 1.17339253e-01 -8.10923576e-02 3.60871665e-02
2.91319609e-01 7.24759936e-01 3.02518219e-01 -6.23703301e-01
-4.77200031e-01 3.63837898e-01 3.78502086e-02 2.21752450e-02
1.19157994e+00 9.19163302e-02 -6.41780943e-02 2.06097096e-01
1.01725364e+00 -4.72265720e-01 -1.13592231e+00 1.92799196e-01
-5.60850911e-02 -3.93382102e-01 2.34030306e-01 -1.21088314e+00
-1.06631672e+00 5.83569348e-01 1.23899126e+00 2.50625521e-01
1.35760665e+00 -3.99813443e-01 4.03604247e-02 1.90528661e-01
2.14737013e-01 -8.03630948e-01 3.49429280e-01 4.59340721e-01
1.15940356e+00 -1.06356454e+00 2.89064884e-01 4.10635591e-01
-8.51514757e-01 9.44512904e-01 7.44656742e-01 -2.91494757e-01
5.69047451e-01 1.95722774e-01 -1.13286562e-01 -1.22090198e-01
-9.99576449e-01 2.92122066e-01 1.66266263e-01 7.43817747e-01
3.30701381e-01 8.36260915e-02 -2.22659647e-01 2.70886749e-01
-4.53057557e-01 -1.79460108e-01 4.63130355e-01 5.67688942e-01
-3.77895236e-01 -9.17120099e-01 -5.97665012e-01 5.97103477e-01
-1.85627237e-01 -1.76407933e-01 -6.93235517e-01 5.98640382e-01
6.86115146e-01 8.97316635e-01 2.40515128e-01 -9.55453634e-01
3.59759122e-01 -2.89292950e-02 4.81145531e-01 -1.06695138e-01
-1.20536983e+00 -6.96149826e-01 1.16371317e-02 -3.88967454e-01
-1.19500987e-01 -6.61898136e-01 -6.22429252e-01 -7.03555703e-01
-4.17399019e-01 -2.99530625e-01 8.29609334e-01 8.18950236e-01
8.14865887e-01 5.82201958e-01 9.39856529e-01 -6.51688218e-01
-7.19588816e-01 -9.47557211e-01 -1.75080955e-01 2.97646016e-01
3.02039295e-01 -6.11948192e-01 -6.15116656e-01 -2.00707048e-01] | [5.565266132354736, 7.7845659255981445] |
f96c570c-356d-49da-97bd-f53c1009c3fa | utilizing-domain-knowledge-in-end-to-end | 1712.00254 | null | http://arxiv.org/abs/1712.00254v1 | http://arxiv.org/pdf/1712.00254v1.pdf | Utilizing Domain Knowledge in End-to-End Audio Processing | End-to-end neural network based approaches to audio modelling are generally
outperformed by models trained on high-level data representations. In this
paper we present preliminary work that shows the feasibility of training the
first layers of a deep convolutional neural network (CNN) model to learn the
commonly-used log-scaled mel-spectrogram transformation. Secondly, we
demonstrate that upon initializing the first layers of an end-to-end CNN
classifier with the learned transformation, convergence and performance on the
ESC-50 environmental sound classification dataset are similar to a CNN-based
model trained on the highly pre-processed log-scaled mel-spectrogram features. | ['Lars Maaløe', 'Hendrik Purwins', 'Tycho Max Sylvester Tax', 'Jose Luis Diez Antich'] | 2017-12-01 | null | null | null | null | ['environmental-sound-classification', 'sound-classification'] | ['audio', 'audio'] | [ 1.52911723e-01 -3.43850516e-02 4.40926373e-01 -5.58147907e-01
-8.71190548e-01 -2.82400012e-01 3.64811778e-01 1.40382512e-03
-6.98799670e-01 1.69441581e-01 3.53791237e-01 -3.70360523e-01
1.43040325e-02 -4.52915281e-01 -6.35380864e-01 -1.95911169e-01
-5.01398146e-01 -2.33147535e-02 -1.44845605e-01 -3.07058185e-01
-1.55796528e-01 3.01132858e-01 -1.72122610e+00 7.24746585e-01
1.79100223e-02 1.41178310e+00 -1.69436499e-01 1.53887284e+00
4.03862000e-01 1.02126956e+00 -7.69608140e-01 -2.69956458e-02
1.59348369e-01 -4.71944094e-01 -8.88248861e-01 -5.80217004e-01
7.63746083e-01 -4.44715321e-01 -4.11916047e-01 7.34828234e-01
8.44154835e-01 4.72469449e-01 5.36527216e-01 -1.05164349e+00
-4.43903983e-01 7.56889284e-01 2.70112455e-01 5.09178400e-01
1.99839026e-01 -1.20887943e-02 1.07641888e+00 -9.58836019e-01
1.10818617e-01 1.16786802e+00 1.41715086e+00 3.68152916e-01
-1.21050620e+00 -6.51572466e-01 -1.48781508e-01 2.29902714e-02
-1.18531823e+00 -8.23126316e-01 7.71952271e-01 -5.15226901e-01
1.63636875e+00 2.64072299e-01 7.38128185e-01 9.43047702e-01
1.86558247e-01 5.63243270e-01 6.10863030e-01 -7.43117690e-01
1.79109275e-01 -2.93970585e-01 -1.26285031e-01 3.87062818e-01
-4.30056572e-01 5.44172227e-01 -6.98418140e-01 -3.80765110e-01
7.14730144e-01 -6.68388903e-01 -1.88852493e-02 3.12931597e-01
-8.98556650e-01 6.48449123e-01 6.35779560e-01 5.39270818e-01
-5.23271859e-01 7.68339694e-01 8.88321519e-01 4.68752593e-01
6.46908820e-01 6.33216739e-01 -7.45156288e-01 -6.08724296e-01
-1.38539636e+00 3.77822340e-01 7.91095316e-01 5.12387455e-01
2.80290842e-01 9.26124692e-01 1.84263840e-01 1.03319561e+00
2.37165213e-01 -4.25182581e-02 7.57302046e-01 -1.07382309e+00
1.25301123e-01 -2.92883664e-01 -1.15970671e-01 -7.04461277e-01
-5.91180682e-01 -9.06353652e-01 -7.37837493e-01 4.41190839e-01
3.85503471e-01 -4.70178515e-01 -9.17962909e-01 1.67341876e+00
-2.58873045e-01 4.72580880e-01 2.36322090e-01 6.51343822e-01
9.74977732e-01 6.86015964e-01 2.40680143e-01 3.45454186e-01
8.70830238e-01 -9.33625937e-01 -6.34171128e-01 -1.57728031e-01
3.89256179e-01 -8.89309645e-01 1.05770063e+00 5.72819412e-01
-1.31835353e+00 -1.35875189e+00 -1.20896721e+00 -1.80154249e-01
-5.98549306e-01 4.34457630e-01 3.47433716e-01 6.20604157e-01
-1.23636627e+00 9.00865674e-01 -8.14319134e-01 5.25557958e-02
2.99133897e-01 5.13100088e-01 -1.75160185e-01 6.60603344e-01
-1.29987967e+00 7.34686852e-01 8.25292051e-01 2.11522251e-01
-1.24541199e+00 -9.73043442e-01 -7.69975841e-01 3.00560534e-01
-4.70335066e-01 -3.64028037e-01 2.08785081e+00 -1.15562499e+00
-1.85104609e+00 3.50607783e-01 1.19480722e-01 -8.49783957e-01
2.84204423e-01 -5.10243535e-01 -6.58729732e-01 6.18584119e-02
-2.67019480e-01 7.62507379e-01 9.97203767e-01 -9.63337004e-01
-9.25562084e-01 2.53508329e-01 -4.49320413e-02 -5.09072542e-02
-3.87821943e-01 2.97895223e-01 3.40028793e-01 -8.58519793e-01
-2.95054615e-01 -6.64354146e-01 -2.89109230e-01 -2.61285663e-01
-1.37831718e-01 1.15372436e-02 6.06582522e-01 -9.41046596e-01
1.18705022e+00 -2.43088937e+00 -2.98697710e-01 -3.07476167e-02
-1.19937479e-01 6.19402289e-01 -4.06895459e-01 3.62692118e-01
-6.88259363e-01 4.53505032e-02 -2.45830342e-02 -5.47428548e-01
1.41589925e-01 -3.04996133e-01 -5.07335782e-01 1.31679043e-01
3.60719532e-01 5.32101810e-01 -9.84397590e-01 2.17045411e-01
2.17891455e-01 7.37869084e-01 -7.35357225e-01 3.95490080e-01
-4.43223938e-02 3.13754827e-01 3.66836458e-01 1.33932278e-01
2.34789655e-01 3.16411018e-01 -1.87155232e-01 -2.92215109e-01
-8.05541649e-02 7.38577962e-01 -8.46341312e-01 1.77981949e+00
-9.97827709e-01 1.25036430e+00 8.21043402e-02 -6.74239039e-01
8.82922232e-01 8.95579636e-01 3.31882656e-01 -2.93337375e-01
3.63264352e-01 4.33025777e-01 4.85315919e-01 -2.72233456e-01
4.76974279e-01 -3.04931313e-01 -8.41175858e-03 4.05884504e-01
7.92658269e-01 -4.36558723e-01 -3.37369680e-01 -3.40959877e-01
7.80995369e-01 2.27677718e-01 1.98552385e-01 -1.20586634e-01
4.04451102e-01 -1.81045458e-01 9.31190625e-02 7.86880195e-01
-1.30460218e-01 8.57758939e-01 -7.67939631e-03 -7.66040742e-01
-1.25760508e+00 -6.30391300e-01 -2.05387890e-01 1.68545437e+00
-1.01682699e+00 -7.74626970e-01 -9.99093413e-01 -1.13154389e-01
-1.52904168e-01 7.84643114e-01 -6.30777895e-01 -2.99956858e-01
-5.90969920e-01 -5.03867626e-01 1.50551236e+00 9.04224157e-01
3.48102033e-01 -1.19501722e+00 -6.99160695e-01 8.71669114e-01
1.46972863e-02 -8.48055899e-01 -2.03103200e-02 9.45503116e-01
-6.97847962e-01 -5.75485229e-01 -5.26267171e-01 -9.01872754e-01
-1.80495277e-01 -4.71116066e-01 1.08706999e+00 -7.30731934e-02
-1.12700239e-01 2.45185271e-01 -3.07270825e-01 -9.98618364e-01
-7.57297456e-01 3.47713768e-01 2.68304318e-01 -1.63822889e-01
3.68024319e-01 -9.61155772e-01 -3.68842572e-01 -3.21358889e-01
-7.89944887e-01 -8.89153555e-02 9.75565165e-02 8.18780422e-01
3.42656285e-01 7.83461183e-02 8.10094893e-01 -2.80028880e-01
9.99494553e-01 3.97024043e-02 -1.17835760e-01 -2.64785051e-01
1.61558595e-02 -2.61568129e-01 9.66724813e-01 -5.98647594e-01
-7.41812944e-01 4.23996449e-01 -8.51845026e-01 -5.18042922e-01
-5.98691404e-01 7.24328101e-01 3.84930730e-01 8.71901810e-02
1.03223550e+00 6.83692023e-02 -1.90908194e-01 -6.75546765e-01
5.08531213e-01 1.16110873e+00 1.04393148e+00 -3.37048918e-01
6.61618531e-01 1.26377761e-01 -2.49508321e-01 -8.58001113e-01
-9.14207578e-01 -1.27207503e-01 -8.16829145e-01 -1.92088723e-01
8.36396456e-01 -1.22311342e+00 -5.01814365e-01 7.34432697e-01
-1.23055911e+00 -6.58214450e-01 -5.48311710e-01 6.24976039e-01
-9.71518219e-01 -3.29966545e-01 -8.10762882e-01 -9.89139616e-01
-3.57433081e-01 -7.04286098e-01 1.02535784e+00 -9.50396061e-02
-6.01485848e-01 -1.17948127e+00 3.58480662e-01 -2.37428114e-01
7.86259115e-01 2.49197826e-01 9.56365287e-01 -9.97408450e-01
2.91615456e-01 -3.55329037e-01 2.18678400e-01 9.36857879e-01
7.90434598e-04 1.59021527e-01 -1.83914459e+00 -2.81937599e-01
7.00993389e-02 -5.32376230e-01 1.01789904e+00 5.76427281e-01
1.26320493e+00 -5.12767397e-02 4.64655787e-01 8.47476065e-01
9.35333133e-01 3.76984537e-01 3.34869742e-01 4.71964985e-01
4.41246748e-01 2.11937279e-01 2.29593769e-01 2.21907571e-01
-1.25412747e-01 4.93729800e-01 3.83586705e-01 -3.64602029e-01
-3.73527497e-01 -4.45792586e-01 2.96230108e-01 1.05909455e+00
-2.08650202e-01 2.35134717e-02 -9.10968840e-01 5.89419603e-01
-1.53402138e+00 -7.89097130e-01 5.26207127e-02 1.63362098e+00
9.65194345e-01 2.84592718e-01 3.16289514e-01 7.07047939e-01
3.49040091e-01 1.36280701e-01 -1.12812087e-01 -7.69366682e-01
2.07097366e-01 8.08599651e-01 3.07984315e-02 4.95455354e-01
-1.49406815e+00 9.76376414e-01 8.06010723e+00 7.73483455e-01
-1.34051394e+00 1.65567130e-01 4.55195040e-01 -1.28151760e-01
4.39637184e-01 -4.91621047e-01 -4.94607359e-01 5.84096201e-02
1.90108430e+00 1.05544835e-01 6.13081932e-01 1.03348458e+00
2.31400654e-01 3.87852430e-01 -1.47773385e+00 9.37109232e-01
-1.77945286e-01 -1.20781374e+00 -1.07119910e-01 -1.57930672e-01
5.23652673e-01 2.13262886e-01 3.14047813e-01 5.61460674e-01
3.88745338e-01 -1.28537941e+00 1.24899185e+00 4.28712130e-01
1.02775633e+00 -1.04859293e+00 8.19229722e-01 9.11555812e-02
-1.39092398e+00 -3.30080748e-01 -5.03604889e-01 -5.15273273e-01
-7.95402974e-02 1.35549575e-01 -1.27158320e+00 3.74990791e-01
7.94117987e-01 5.71280956e-01 -5.37269533e-01 1.19534314e+00
8.07549581e-02 1.24462450e+00 -2.64162183e-01 3.14636230e-01
5.90146959e-01 5.94090641e-01 2.08937034e-01 1.84068859e+00
4.16372508e-01 -2.86359668e-01 -2.61083812e-01 5.89502871e-01
-2.98259892e-02 -9.47716236e-02 -3.87233496e-01 -2.39163727e-01
2.19262078e-01 9.14959610e-01 -3.86948228e-01 -4.23908949e-01
-1.41432479e-01 4.99179363e-01 2.47292981e-01 4.66072619e-01
-5.34983218e-01 -8.80359888e-01 8.36789310e-01 -2.27044314e-01
5.70460439e-01 -5.65592684e-02 -2.89241701e-01 -7.32214868e-01
-4.04386371e-01 -1.02354574e+00 8.73481855e-02 -1.06526589e+00
-1.19255674e+00 1.16382682e+00 -1.97928935e-01 -1.26939201e+00
-8.52091610e-01 -8.01465213e-01 -9.29968357e-01 1.13357568e+00
-1.27178860e+00 -1.07492673e+00 -1.39717042e-01 4.36117858e-01
5.00587046e-01 -7.35237420e-01 1.44452965e+00 2.86194623e-01
-1.11647151e-01 6.13266706e-01 1.88758329e-01 5.99796414e-01
3.71678978e-01 -1.35585868e+00 9.04599667e-01 5.69421828e-01
6.06126308e-01 5.02418637e-01 7.95421660e-01 -1.61865801e-01
-5.76342762e-01 -1.33046341e+00 8.19497705e-01 -3.54886949e-01
5.61334252e-01 -5.18248439e-01 -8.71072054e-01 8.14222336e-01
5.46741068e-01 9.86288711e-02 9.90448236e-01 1.63837269e-01
-3.60468537e-01 -1.71481356e-01 -7.40167499e-01 2.50045687e-01
5.99206746e-01 -1.19396150e+00 -8.58470917e-01 9.53221694e-02
7.24373996e-01 -6.45302713e-01 -1.00363457e+00 3.08516115e-01
7.69470155e-01 -8.03332627e-01 1.14622307e+00 -1.16825330e+00
5.29508591e-01 -6.11279495e-02 -2.85196871e-01 -1.87374270e+00
-4.33871269e-01 -6.56143308e-01 -5.35870902e-02 9.32805479e-01
4.16101426e-01 -8.35853964e-02 4.70109850e-01 1.67986918e-02
-4.58503097e-01 -4.11910266e-01 -1.19020033e+00 -7.54392207e-01
3.05594176e-01 -1.11266291e+00 7.51571298e-01 6.73425496e-01
-9.20442417e-02 2.29295343e-01 -3.24257493e-01 1.14167415e-01
8.55123973e-04 -5.85791945e-01 7.79190898e-01 -1.35675693e+00
-4.22231704e-01 -4.89636004e-01 -7.17970014e-01 -7.43941069e-01
1.32585660e-01 -7.16581464e-01 3.82048458e-01 -1.17478812e+00
-5.08096159e-01 -6.24748506e-02 -8.39420259e-01 5.40822983e-01
7.74566233e-02 6.23824835e-01 3.10935915e-01 -1.78159297e-01
-1.94424674e-01 4.29101318e-01 7.29901850e-01 6.53011799e-02
-4.18999642e-01 1.95184082e-01 -2.77904868e-01 9.91995811e-01
9.07682657e-01 -5.59167743e-01 -3.34899276e-01 -3.75806630e-01
1.62451848e-01 -5.91011569e-02 5.96728683e-01 -1.59175336e+00
8.25371146e-02 3.63214582e-01 6.91606045e-01 -5.72833955e-01
7.71398425e-01 -7.52810419e-01 9.79431570e-02 2.94198543e-01
-9.73506868e-01 -1.68945923e-01 7.42605209e-01 2.64163733e-01
-6.27989471e-01 -4.11860317e-01 7.65256882e-01 -1.16329618e-01
-2.66790181e-01 -1.65429920e-01 -8.80675197e-01 -1.73821464e-01
1.70147985e-01 -1.47495255e-01 6.14806190e-02 -7.51289964e-01
-1.26707292e+00 -6.78513765e-01 -2.75925338e-01 4.43267554e-01
4.66928810e-01 -1.56946504e+00 -8.10955584e-01 4.36460525e-01
-2.76343286e-01 -1.94221556e-01 -7.96068534e-02 3.77645910e-01
-6.78674757e-01 5.25128961e-01 -5.40927351e-01 -4.84577715e-01
-1.18453586e+00 -4.24794927e-02 1.00893772e+00 -1.64442919e-02
-3.24059010e-01 1.19236171e+00 -1.04133055e-01 -6.49850130e-01
6.18458807e-01 -9.18283105e-01 -1.36560917e-01 -1.11595273e-01
7.52509773e-01 1.91374138e-01 4.57293063e-01 -7.18899429e-01
-1.31638363e-01 6.51992187e-02 2.91714728e-01 -6.04640663e-01
1.88418412e+00 4.96304542e-01 2.79316485e-01 7.81031787e-01
1.40337241e+00 -2.73111254e-01 -1.27114165e+00 -1.08849563e-01
-1.36364922e-01 -2.81167720e-02 5.71699262e-01 -9.33681428e-01
-9.45291579e-01 1.44383407e+00 8.26504946e-01 4.93588656e-01
1.12976539e+00 -6.78777993e-01 6.35289550e-01 6.80397868e-01
-9.09260511e-02 -1.16144836e+00 8.32582936e-02 1.03009951e+00
1.13638926e+00 -6.60895169e-01 -3.47691029e-01 8.77055153e-02
-3.45487088e-01 1.68069553e+00 4.13951516e-01 -3.87160897e-01
1.19338751e+00 5.60391486e-01 4.76840794e-01 2.85858158e-02
-8.88311625e-01 -9.13611874e-02 4.25999969e-01 8.85632336e-01
7.70577431e-01 -1.49468258e-01 6.69338644e-01 8.96265388e-01
-1.02685964e+00 1.68663263e-01 3.82390887e-01 8.95241976e-01
-4.67547029e-01 -7.52062142e-01 -3.50998491e-01 2.92690754e-01
-7.90139198e-01 -3.73441637e-01 -4.33031738e-01 6.41689897e-01
3.15112978e-01 1.03262103e+00 3.56703430e-01 -7.83670485e-01
4.30386066e-01 6.65170670e-01 2.35457838e-01 -6.81796134e-01
-1.36607111e+00 3.01484048e-01 3.05821180e-01 -2.18465567e-01
-4.26089495e-01 -4.34511244e-01 -1.03899324e+00 9.77357253e-02
-2.40405008e-01 1.11091904e-01 7.47017503e-01 8.26461136e-01
3.73861194e-02 1.13317144e+00 5.23591578e-01 -1.45135331e+00
-6.82238281e-01 -1.50947165e+00 -5.33956885e-01 9.34726298e-02
8.67887855e-01 -1.18624128e-01 -3.99911255e-01 4.67955679e-01] | [15.321670532226562, 5.37695837020874] |
35c94202-5a60-41df-afd2-9a7cac6f9e30 | expanding-accurate-person-recognition-to-new | 2211.01917 | null | https://arxiv.org/abs/2211.01917v1 | https://arxiv.org/pdf/2211.01917v1.pdf | Expanding Accurate Person Recognition to New Altitudes and Ranges: The BRIAR Dataset | Face recognition technology has advanced significantly in recent years due largely to the availability of large and increasingly complex training datasets for use in deep learning models. These datasets, however, typically comprise images scraped from news sites or social media platforms and, therefore, have limited utility in more advanced security, forensics, and military applications. These applications require lower resolution, longer ranges, and elevated viewpoints. To meet these critical needs, we collected and curated the first and second subsets of a large multi-modal biometric dataset designed for use in the research and development (R&D) of biometric recognition technologies under extremely challenging conditions. Thus far, the dataset includes more than 350,000 still images and over 1,300 hours of video footage of approximately 1,000 subjects. To collect this data, we used Nikon DSLR cameras, a variety of commercial surveillance cameras, specialized long-rage R&D cameras, and Group 1 and Group 2 UAV platforms. The goal is to support the development of algorithms capable of accurately recognizing people at ranges up to 1,000 m and from high angles of elevation. These advances will include improvements to the state of the art in face recognition and will support new research in the area of whole-body recognition using methods based on gait and anthropometry. This paper describes methods used to collect and curate the dataset, and the dataset's characteristics at the current stage. | ['David S. Bolme', 'Hector J. Santos-Villalobos', 'Scott Dolvin', 'Robert Zhang', 'Matt Yohe', 'Leanne Thompson', 'Brandon Stockwell', 'Nisha Srinivas', 'Ian Shelley', 'Christi Johnson', 'Bart Murphy', 'Matt Larson', 'Gavin Jager', 'Jim Goddard', 'Regina Ferrell', 'Andrew Duncan', 'Carl Dukes', 'Nick Burchfield', 'Seth Baird', 'Deniz Aykac', 'Nell Barber', 'Joel Brogan', 'David Cornett III'] | 2022-11-03 | null | null | null | null | ['person-recognition'] | ['computer-vision'] | [ 9.73824784e-02 -5.10641992e-01 -6.56815544e-02 -5.62781036e-01
-3.81714284e-01 -2.98883796e-01 3.88968706e-01 -3.21692854e-01
-5.32921493e-01 3.93016100e-01 4.53334562e-02 5.67509681e-02
-4.47650477e-02 -7.38832951e-01 -2.26494402e-01 -5.57586908e-01
-1.97405398e-01 3.51088196e-01 -1.39249727e-01 -2.06236169e-01
9.90556926e-02 9.92232025e-01 -1.86622167e+00 1.39026627e-01
2.40244135e-01 1.19947028e+00 -3.47086459e-01 5.46235025e-01
4.54005688e-01 1.37393028e-01 -5.07180870e-01 -6.90389693e-01
3.80622834e-01 -1.50856912e-01 -2.72101283e-01 3.11625242e-01
9.95576560e-01 -7.41786063e-01 -2.54074126e-01 4.89084750e-01
8.92399788e-01 1.84696555e-01 3.01838219e-01 -1.27447617e+00
-4.81431842e-01 -1.82077229e-01 -8.99629712e-01 2.87845850e-01
7.10114598e-01 1.08359493e-01 4.99297678e-01 -8.02977324e-01
4.26228851e-01 1.11033213e+00 9.55664515e-01 7.96046913e-01
-7.58943737e-01 -7.17113078e-01 -3.35505366e-01 1.87903419e-01
-1.67217481e+00 -9.88001764e-01 7.33764350e-01 -3.98818940e-01
7.76740491e-01 4.83641438e-02 7.98440993e-01 1.31855500e+00
-3.03207766e-02 1.86775282e-01 8.53278756e-01 -4.20944750e-01
-2.81817261e-02 5.89336306e-02 -3.47248986e-02 1.00721705e+00
6.80510163e-01 7.10733533e-02 -8.95408273e-01 -3.47717881e-01
8.42118740e-01 1.41733795e-01 -1.03214234e-01 -2.09673792e-01
-7.48254001e-01 7.06271887e-01 3.08614429e-02 1.25881955e-01
-3.51206034e-01 -1.80863097e-01 1.36678070e-01 1.26945943e-01
3.85307550e-01 -2.28219450e-01 -1.92313954e-01 -2.96943665e-01
-9.45707798e-01 1.03461273e-01 6.71965122e-01 6.13339067e-01
4.96010929e-01 1.88271508e-01 5.93587160e-01 9.77967083e-01
5.75268090e-01 8.96362722e-01 3.12212229e-01 -6.90535665e-01
5.04686415e-01 6.02618456e-01 -1.37281679e-02 -1.42407882e+00
-6.42069042e-01 9.26143229e-02 -6.66513383e-01 5.75393029e-02
5.47059536e-01 -2.61414438e-01 -7.95809448e-01 1.34135830e+00
6.15814507e-01 -3.44554596e-02 -3.06729257e-01 9.57247555e-01
7.79141843e-01 2.22820312e-01 -1.52584687e-01 2.61878241e-02
1.49284136e+00 -2.04986364e-01 -2.38094509e-01 -2.49954298e-01
1.76718742e-01 -7.45068252e-01 8.15141559e-01 4.27526146e-01
-8.20809305e-01 -6.92727208e-01 -9.04424250e-01 2.94548392e-01
-3.46688688e-01 4.07158673e-01 3.91653150e-01 1.47349679e+00
-9.02906835e-01 3.43835622e-01 -8.41991127e-01 -7.82040775e-01
6.92541778e-01 5.01241922e-01 -7.38104939e-01 -3.81470025e-01
-8.88213634e-01 8.89170289e-01 -5.08717954e-01 4.04436499e-01
-4.84834492e-01 -3.38977069e-01 -1.05876017e+00 -3.89697820e-01
9.52529237e-02 -4.14657295e-01 7.24426627e-01 -6.00505292e-01
-1.26349938e+00 1.27475917e+00 8.14839527e-02 -3.53278928e-02
2.03303471e-01 -4.15793329e-01 -7.31574535e-01 3.78241390e-01
-2.72651445e-02 3.51563156e-01 1.03575623e+00 -6.46803439e-01
-2.87750542e-01 -1.12948465e+00 -2.52089173e-01 1.59642428e-01
-6.03636622e-01 4.61285770e-01 -2.95219451e-01 -4.31950450e-01
-1.77005321e-01 -1.06086814e+00 3.15592170e-01 1.06636599e-01
4.40271907e-02 2.64283270e-01 1.07410491e+00 -9.09198046e-01
6.73823178e-01 -2.07382441e+00 -8.55040699e-02 1.79891095e-01
-9.18631181e-02 5.06680012e-01 6.00243285e-02 2.45853901e-01
1.94489852e-01 4.32467880e-03 8.83921012e-02 -3.20454419e-01
-3.31167281e-01 4.09380049e-02 8.24691877e-02 9.12380457e-01
-6.75038323e-02 3.31906587e-01 -3.34558725e-01 -3.92067850e-01
4.33719009e-01 5.59376121e-01 -3.28365862e-01 1.12457179e-01
4.41378891e-01 3.81212562e-01 -2.98533201e-01 1.21232224e+00
6.60187304e-01 2.65482247e-01 8.99654403e-02 -2.21961185e-01
-7.83258118e-03 -3.39645654e-01 -1.18827844e+00 1.40718281e+00
-1.81021765e-01 6.28612280e-01 3.03617835e-01 -8.43173802e-01
9.86522019e-01 3.23764622e-01 8.18799913e-01 -3.90563071e-01
3.40737015e-01 2.85860174e-03 -1.99219480e-01 -7.81146169e-01
3.66175354e-01 -3.94025594e-01 1.61229387e-01 5.26560068e-01
-5.98358326e-02 1.99004665e-01 9.78885591e-02 -2.93086439e-01
7.76105344e-01 4.84825447e-02 1.93280444e-01 3.36236954e-02
5.53589165e-01 -1.18952103e-01 5.17600477e-01 2.25653291e-01
-5.29873610e-01 5.78206658e-01 -6.14027567e-02 -7.90327966e-01
-8.22608650e-01 -9.18874681e-01 -3.15360606e-01 9.44992244e-01
-1.61929160e-01 -2.58089811e-01 -6.98439300e-01 -2.85111874e-01
2.10849807e-01 -1.99753612e-01 -4.04605180e-01 -1.82622790e-01
-5.81712008e-01 -1.05140233e+00 9.23637927e-01 6.54375434e-01
8.52177382e-01 -7.35048473e-01 -8.95102620e-01 -1.07528389e-01
-8.37199986e-02 -1.26344180e+00 -2.95032829e-01 -7.27644265e-01
-7.59434104e-01 -1.50859821e+00 -8.75334859e-01 -3.59906912e-01
6.39668286e-01 3.94180268e-01 7.95947969e-01 8.30006003e-02
-8.40349138e-01 8.69926870e-01 -3.91504019e-01 -4.38588768e-01
2.36481950e-01 -1.80561647e-01 8.55789781e-01 3.92945945e-01
6.79816663e-01 -3.49551022e-01 -3.80799919e-01 6.67470992e-01
-5.97575009e-01 -5.34108222e-01 4.15844440e-01 6.04169846e-01
2.77417041e-02 -2.78453901e-02 4.20560688e-01 -3.33045244e-01
2.84339845e-01 -3.81955028e-01 -5.66811621e-01 1.16649076e-01
-2.52091110e-01 -5.14997423e-01 3.31864417e-01 -4.81648684e-01
-9.31868613e-01 -1.43027559e-01 -1.57014281e-01 -3.12538445e-01
-3.43280047e-01 3.78684163e-01 -3.02547932e-01 -5.48944831e-01
7.75469005e-01 -1.91848334e-02 3.59649181e-01 -4.29489136e-01
-1.23349123e-01 1.17741311e+00 4.85880435e-01 -6.11200452e-01
7.44034052e-01 6.13266647e-01 1.94825187e-01 -1.67854083e+00
-4.85208958e-01 -3.39260370e-01 -8.14763784e-01 -6.53148293e-01
5.88041663e-01 -8.31464887e-01 -8.26297164e-01 1.05928218e+00
-5.55641949e-01 1.76008306e-02 2.33741730e-01 5.41879237e-01
-6.58388436e-02 4.71244574e-01 -4.54404414e-01 -1.09765339e+00
-4.02351439e-01 -7.76498079e-01 1.11570656e+00 5.04722774e-01
-1.02482028e-01 -7.02033699e-01 1.84687488e-02 1.04119754e+00
4.54088300e-01 4.80967373e-01 3.16854089e-01 -7.84265921e-02
-1.53464064e-01 -6.33108497e-01 -8.41308944e-03 3.08731288e-01
3.25028807e-01 2.47036830e-01 -1.01106679e+00 -5.38891137e-01
-6.74538910e-02 -5.51814616e-01 3.84414077e-01 3.59632194e-01
8.64625871e-01 4.63830829e-02 -1.53149456e-01 7.34771311e-01
1.08693707e+00 2.92155683e-01 6.73951805e-01 1.06838152e-01
5.99433899e-01 8.46342146e-01 4.15474206e-01 6.12217784e-01
4.81420457e-01 8.75281215e-01 1.26420602e-01 2.21932396e-01
5.82284071e-02 3.80511256e-03 4.63971585e-01 3.61144364e-01
-6.69701457e-01 1.10024780e-01 -1.22647321e+00 2.55943179e-01
-1.22056913e+00 -1.14157951e+00 1.41933426e-01 2.41884565e+00
2.79902995e-01 -2.37179980e-01 5.16361713e-01 7.03316554e-02
8.03768635e-01 1.55668959e-01 -5.72109699e-01 -1.56394228e-01
6.59417883e-02 1.54809624e-01 2.49981478e-01 4.18228796e-03
-1.11078238e+00 4.70315576e-01 6.72915745e+00 1.51941091e-01
-1.29310238e+00 -2.10659742e-01 4.83835816e-01 -4.23354894e-01
5.24935961e-01 -6.42411590e-01 -1.15689659e+00 4.61916447e-01
1.14239526e+00 1.87357679e-01 4.63094324e-01 9.21926737e-01
1.76664636e-01 -3.10434788e-01 -7.78361022e-01 1.32426655e+00
7.28657007e-01 -9.34432507e-01 -2.21067846e-01 4.62759644e-01
3.33103269e-01 -1.09026849e-01 2.37333864e-01 5.00874454e-03
-3.07926834e-01 -9.57766354e-01 3.65490496e-01 4.73911673e-01
1.09959280e+00 -7.25612223e-01 7.54810333e-01 1.94867134e-01
-1.24095607e+00 -2.08247781e-01 -4.88772541e-01 -2.07760453e-01
3.02047655e-02 4.41112936e-01 -3.65990043e-01 3.85530174e-01
9.15479004e-01 6.41961217e-01 -5.92608094e-01 6.83355331e-01
2.02383310e-01 3.66938323e-01 -5.30577540e-01 8.61464217e-02
-3.23064744e-01 -2.50867516e-01 2.62063563e-01 7.20649362e-01
4.64787662e-01 4.01309282e-01 -3.11609749e-02 2.15531692e-01
-1.92565024e-01 -1.09847777e-01 -8.67634475e-01 -1.53811514e-01
3.76397073e-01 1.30728924e+00 -5.10950744e-01 -1.01130277e-01
-7.34673738e-01 5.27792156e-01 5.20506464e-02 -5.95093518e-02
-7.93179393e-01 -1.94555730e-01 9.89764750e-01 5.59997797e-01
-5.94549961e-02 -6.71284199e-01 -8.87269005e-02 -1.31706190e+00
2.10738152e-01 -1.04815054e+00 4.87844616e-01 -4.99747664e-01
-1.29169321e+00 5.44650018e-01 2.05606490e-01 -1.04065895e+00
-2.71821171e-01 -9.50179994e-01 -4.70913291e-01 8.46107304e-01
-9.97749567e-01 -1.37401950e+00 -7.59982526e-01 8.13586354e-01
2.84397244e-01 -6.89261913e-01 8.95576119e-01 5.96895516e-01
-9.11617637e-01 5.23914039e-01 -1.04096502e-01 3.49677652e-01
7.78300107e-01 -5.29790759e-01 3.01434040e-01 7.76653171e-01
6.34270906e-02 7.46732593e-01 2.03069568e-01 -7.69079268e-01
-1.83087778e+00 -7.83009112e-01 4.65337217e-01 -7.00905681e-01
2.68922538e-01 -2.24579647e-01 -3.73581052e-01 6.43071294e-01
-3.51447254e-01 9.07839909e-02 1.28967953e+00 3.02471250e-01
-2.15946078e-01 -4.63710666e-01 -1.61203146e+00 2.49977618e-01
8.85112882e-01 -5.46196699e-01 -5.48258960e-01 7.53966570e-02
-4.00897026e-01 -3.24152440e-01 -9.44981217e-01 3.58691037e-01
1.27749729e+00 -9.15923834e-01 1.15441179e+00 -4.58781004e-01
1.75485432e-01 -1.53771052e-02 -3.00477862e-01 -8.97648811e-01
5.38494587e-02 -2.41135046e-01 -1.82728589e-01 1.16019917e+00
-8.14548731e-02 -5.36337614e-01 1.21470916e+00 9.83090043e-01
3.08445245e-01 -5.94772816e-01 -8.86028469e-01 -7.79489160e-01
-2.45977014e-01 -2.32506901e-01 3.78004998e-01 6.91327333e-01
-1.02395937e-01 -2.35146433e-02 -6.80338383e-01 1.89673185e-01
1.04227209e+00 3.07424236e-02 9.33242977e-01 -1.47069561e+00
1.28222629e-01 3.82745303e-02 -9.37149942e-01 -4.19555247e-01
-1.08851627e-01 -3.06896657e-01 -5.38781881e-01 -1.06899512e+00
-1.58689208e-02 -2.19412163e-01 3.20957184e-01 5.28194249e-01
1.18058138e-01 8.45514953e-01 1.05223782e-01 2.22643957e-01
7.14635924e-02 2.81985372e-01 5.90977073e-01 4.89976481e-02
-1.22288065e-02 5.71978614e-02 -6.38836980e-01 7.89316714e-01
8.08956206e-01 1.23044156e-01 -1.52103022e-01 -4.62527096e-01
-1.63872004e-01 -1.42484337e-01 4.49900597e-01 -1.30929828e+00
1.86544761e-01 -2.57384539e-01 9.91503417e-01 -4.15476292e-01
8.17887783e-01 -7.73586154e-01 3.26296151e-01 3.73496920e-01
2.82557398e-01 1.25799388e-01 2.96834469e-01 1.56314671e-01
5.76692708e-02 1.11417912e-01 1.02052450e+00 -2.06881791e-01
-6.95336819e-01 4.59924102e-01 -3.67997408e-01 -2.05449522e-01
1.17108095e+00 -5.38842976e-01 -2.87641883e-01 -2.50520051e-01
-6.34043217e-01 -1.42125160e-01 6.51731431e-01 4.42133963e-01
8.91915619e-01 -1.35878289e+00 -6.37776613e-01 7.79342115e-01
1.62573233e-02 -3.41361642e-01 3.96066964e-01 6.79097176e-01
-7.26717234e-01 3.20166498e-01 -8.53855491e-01 -5.04675448e-01
-1.74404120e+00 5.65487631e-02 2.95910418e-01 3.19489777e-01
-5.87974310e-01 6.21688366e-01 -5.94649553e-01 -2.20726326e-01
1.67684145e-02 3.20314705e-01 -1.74750924e-01 2.70472109e-01
1.05979824e+00 7.32494056e-01 1.79721415e-01 -1.14566195e+00
-6.57530606e-01 8.92169833e-01 1.27835944e-01 -1.40829161e-01
1.34207165e+00 -1.12013951e-01 1.12774760e-01 1.36487901e-01
9.11846876e-01 -1.74089801e-02 -9.79573727e-01 6.91490397e-02
-2.47165799e-01 -9.66183066e-01 -1.22446634e-01 -3.53085876e-01
-1.32071471e+00 8.53646815e-01 7.40851939e-01 -5.82268052e-02
1.15313625e+00 -2.84255326e-01 7.54456341e-01 4.03948307e-01
8.52820218e-01 -1.07256126e+00 7.65269771e-02 1.34100899e-01
6.83429360e-01 -1.20984042e+00 3.31504047e-01 -1.39296785e-01
-4.39224273e-01 1.20168793e+00 6.98635221e-01 2.03930557e-01
5.89080334e-01 3.01499844e-01 2.00916335e-01 -2.76372939e-01
-2.09021956e-01 -5.24668843e-02 2.75647521e-01 1.02921367e+00
5.05586386e-01 -3.09615694e-02 -8.01303014e-02 3.81272495e-01
-2.21980020e-01 9.28425863e-02 3.10288161e-01 1.13615847e+00
-5.05484223e-01 -1.10903871e+00 -9.36940432e-01 7.23968148e-01
-4.73222286e-01 4.60547447e-01 -4.84751135e-01 7.59791195e-01
1.57640383e-01 1.14298749e+00 -9.54504609e-02 -6.81167960e-01
3.04237604e-01 7.13076815e-02 6.92871571e-01 -4.30587769e-01
-2.25175455e-01 -3.37089479e-01 1.27463058e-01 -5.90342402e-01
-5.28488636e-01 -1.06221008e+00 -5.30300379e-01 -8.81722212e-01
-6.07878082e-02 -2.68489122e-01 9.71648932e-01 8.49135518e-01
3.43019038e-01 -9.03230831e-02 6.27257764e-01 -1.30691004e+00
-4.85564113e-01 -9.69810247e-01 -8.28555584e-01 2.68677264e-01
7.68892020e-02 -1.03725004e+00 -1.10247411e-01 2.39132136e-01] | [13.79223918914795, 1.0356782674789429] |
d833255b-edd8-4077-b14e-a11cc4edc8bd | reckon-a-28nm-sub-mm2-task-agnostic-spiking | 2208.09759 | null | https://arxiv.org/abs/2208.09759v1 | https://arxiv.org/pdf/2208.09759v1.pdf | ReckOn: A 28nm Sub-mm2 Task-Agnostic Spiking Recurrent Neural Network Processor Enabling On-Chip Learning over Second-Long Timescales | A robust real-world deployment of autonomous edge devices requires on-chip adaptation to user-, environment- and task-induced variability. Due to on-chip memory constraints, prior learning devices were limited to static stimuli with no temporal contents. We propose a 0.45-mm$^2$ spiking RNN processor enabling task-agnostic online learning over seconds, which we demonstrate for navigation, gesture recognition, and keyword spotting within a 0.8-% memory overhead and a <150-$\mu$W training power budget. | ['Giacomo Indiveri', 'Charlotte Frenkel'] | 2022-08-20 | null | null | null | null | ['gesture-recognition', 'keyword-spotting'] | ['computer-vision', 'speech'] | [ 1.93271086e-01 -1.04178354e-01 -3.49027932e-01 -2.11117297e-01
-5.60040951e-01 -5.02069116e-01 -5.67166461e-03 -1.15631476e-01
-1.12231469e+00 8.48048389e-01 -4.56385702e-01 -4.09628451e-01
1.30009368e-01 -3.94528061e-01 -6.91705823e-01 -4.60437149e-01
-2.55782604e-01 1.06472015e-01 5.08213341e-01 2.92941749e-01
2.94060707e-02 2.37699345e-01 -1.39004290e+00 1.37097940e-01
8.05375203e-02 1.37872899e+00 4.47386175e-01 8.58729184e-01
-6.38360810e-03 2.68111080e-01 -6.33266330e-01 4.08137172e-01
2.83637941e-01 3.72929946e-02 7.06468001e-02 -6.25119507e-01
2.33068123e-01 -4.16125029e-01 -1.14854467e+00 7.12826192e-01
9.92508769e-01 -5.35100438e-02 2.15785116e-01 -8.26461256e-01
-3.41348171e-01 5.86248577e-01 -2.37721056e-01 6.28855407e-01
-1.56942919e-01 5.76214969e-01 2.96401739e-01 -5.14701426e-01
6.76495254e-01 3.40427250e-01 4.74706441e-01 1.05243683e+00
-1.10286081e+00 -1.16516006e+00 2.91058838e-01 -1.63327485e-01
-1.38822734e+00 -1.07057321e+00 4.31238830e-01 2.95904670e-02
1.79573572e+00 -3.51425707e-01 5.07824421e-01 1.42351210e+00
7.07985759e-01 5.78165174e-01 9.79011953e-01 1.41518340e-01
6.96908653e-01 -3.95747662e-01 4.08464521e-01 5.26788354e-01
4.41271275e-01 -3.06894304e-03 -1.09768784e+00 9.88895074e-02
1.07141149e+00 -3.94032001e-02 -6.50549158e-02 1.29826099e-01
-7.55282998e-01 -7.34620467e-02 3.70759636e-01 -4.44907621e-02
-2.52084911e-01 1.25234377e+00 7.14863241e-01 1.03916563e-01
-2.14458302e-01 3.94447654e-01 -7.84273267e-01 -8.50888550e-01
-7.43326843e-01 -3.65729362e-01 6.43041730e-01 1.36052978e+00
2.85756499e-01 7.92996049e-01 1.09953903e-01 1.74397841e-01
2.04193667e-01 1.12828910e+00 7.83628583e-01 -8.49160194e-01
3.08356076e-01 1.28695160e-01 2.83038020e-01 -2.81655878e-01
-9.74572718e-01 -3.61488163e-01 -4.98321801e-01 2.15414047e-01
3.44147146e-01 -4.85013038e-01 -1.50491238e+00 1.89104819e+00
-3.43474060e-01 2.66654909e-01 3.33248898e-02 8.34678113e-01
7.61988223e-01 3.18258673e-01 2.10604534e-01 -1.75530717e-01
1.21478355e+00 -5.62198341e-01 -7.36766160e-01 -9.14932728e-01
1.17003269e-01 -1.25655785e-01 1.09981465e+00 3.84990156e-01
-1.07030904e+00 -3.42673779e-01 -1.45861650e+00 1.10713258e-01
-2.07436696e-01 1.71845362e-01 8.74575198e-01 9.05221820e-01
-1.10036659e+00 3.12214762e-01 -1.49661386e+00 -2.69099802e-01
5.05597413e-01 1.16225946e+00 1.26603350e-01 3.48836333e-01
-6.93345129e-01 2.71315813e-01 7.78393894e-02 -7.67442062e-02
-1.12945318e+00 -4.95662838e-01 -2.68051893e-01 1.25333399e-01
3.18362445e-01 -6.25035286e-01 1.29928267e+00 -4.60440159e-01
-2.05692434e+00 5.39123118e-01 -3.51699650e-01 -7.32146263e-01
-9.47617218e-02 -2.13679045e-01 -7.24717140e-01 2.41681803e-02
-3.01030010e-01 5.73124468e-01 8.34559560e-01 -6.61634564e-01
-2.68963635e-01 -6.19181633e-01 -4.61594820e-01 -8.17314237e-02
-5.52307606e-01 -1.25839517e-01 -6.66655004e-01 -4.04230416e-01
3.81043911e-01 -1.18004501e+00 -2.09777936e-01 4.58898805e-02
3.21603954e-01 4.09279138e-01 1.14776528e+00 -8.47574025e-02
1.11646008e+00 -2.36834073e+00 -5.50294459e-01 2.12376088e-01
2.59419531e-01 1.32704258e-01 -2.24433705e-01 -1.74774647e-01
5.94944000e-01 -7.35677481e-02 2.36658484e-01 -6.99684396e-02
-1.74959019e-01 -7.42250681e-03 -4.08083171e-01 2.67150193e-01
-3.91119570e-01 1.12246513e+00 -6.74677014e-01 1.40678391e-01
-1.50762871e-01 2.17431888e-01 -1.44741371e-01 -1.75388128e-01
-2.86012858e-01 5.40579893e-02 -4.32879895e-01 9.41326797e-01
6.26038313e-02 -3.69905412e-01 5.14092565e-01 -7.53461719e-02
-6.03593662e-02 4.77422059e-01 -8.54408920e-01 2.19656825e+00
-6.45101011e-01 1.08660746e+00 4.68530595e-01 -1.25526175e-01
9.89590645e-01 2.58092582e-01 3.54836404e-01 -1.36656821e+00
7.16391027e-01 6.49270654e-01 2.49907479e-01 -4.75680567e-02
4.57908899e-01 1.86683312e-01 -3.94971102e-01 6.23080790e-01
1.71173841e-01 3.53016555e-01 -3.15122098e-01 1.72245856e-02
1.78283882e+00 -5.10510541e-02 -1.55239269e-01 -6.36257052e-01
-6.43957973e-01 -4.58345324e-01 6.49291217e-01 1.13874269e+00
-6.55309319e-01 6.26986995e-02 -6.98794648e-02 -3.31222057e-01
-7.31249690e-01 -1.33355904e+00 1.01602085e-01 1.23892021e+00
4.26247656e-01 -1.80525199e-01 -5.06319404e-01 -1.88606367e-01
-8.53205398e-02 3.44077379e-01 -1.77520290e-01 -3.79867166e-01
-7.06465840e-01 -7.09352553e-01 9.56344545e-01 1.00463569e+00
6.19815350e-01 -9.12787437e-01 -1.49833286e+00 6.48195207e-01
6.25947177e-01 -1.37706435e+00 -5.17354429e-01 1.21348548e+00
-1.11674762e+00 -4.28220689e-01 -1.73165321e-01 -8.30591500e-01
6.61949992e-01 1.22540191e-01 9.34532225e-01 -5.79313397e-01
-5.62905788e-01 4.22157705e-01 2.58495301e-01 -4.35590237e-01
3.98290485e-01 1.16066530e-01 7.84061015e-01 -5.54218948e-01
6.42351985e-01 -1.03843987e+00 -9.68634605e-01 2.13374510e-01
-2.87461936e-01 -4.44821775e-01 8.85807514e-01 8.84419382e-01
7.88999677e-01 -3.91317904e-01 8.07860374e-01 -6.50686920e-01
4.28520203e-01 -7.19547123e-02 -8.45444024e-01 5.23652732e-02
-7.31255531e-01 4.28431034e-01 7.25967288e-01 -9.38893616e-01
-7.53218114e-01 5.00380695e-01 2.39888191e-01 -5.23869216e-01
1.59354687e-01 6.16024174e-02 -2.16475144e-01 -3.59384477e-01
1.05327106e+00 3.08409184e-01 -3.68567139e-01 1.17603429e-01
9.74789336e-02 5.92675745e-01 7.21199989e-01 -4.53052849e-01
1.72748193e-01 5.82825363e-01 -1.74206883e-01 -7.55068719e-01
1.95851669e-01 -9.00579542e-02 1.08188223e-02 -1.36428162e-01
5.67701697e-01 -1.35671210e+00 -1.25724006e+00 4.46980953e-01
-8.55908990e-01 -1.20719171e+00 5.73769026e-02 5.37692368e-01
-5.31279027e-01 -4.81620848e-01 -7.46744454e-01 -8.68664086e-01
-1.00772727e+00 -9.97081339e-01 8.52912426e-01 7.13578999e-01
-3.93750995e-01 -2.04399362e-01 -3.26686114e-01 -8.15805346e-02
8.88279021e-01 -4.68513995e-01 4.95764464e-01 -2.25904748e-01
-8.77502799e-01 -3.29990059e-01 -3.37873399e-01 -4.39386636e-01
-1.42500311e-01 -5.41832745e-01 -1.21815300e+00 -5.22318661e-01
-1.06405787e-01 -3.97824913e-01 8.84331703e-01 6.24801219e-01
1.20098805e+00 1.46639451e-01 -8.13656330e-01 7.86815941e-01
1.50734162e+00 7.97008157e-01 4.54665124e-01 -9.81336683e-02
3.39264184e-01 -6.30934060e-01 2.06136946e-02 5.23966789e-01
-1.54161081e-01 4.59577948e-01 2.64475524e-01 4.62796718e-01
-1.93648607e-01 -2.53123939e-01 4.84836608e-01 6.69623315e-01
4.01555121e-01 -5.97685218e-01 -9.70341861e-01 5.41749179e-01
-1.67055893e+00 -6.13085210e-01 4.62385029e-01 2.31947064e+00
5.33094287e-01 8.82187366e-01 -2.08280906e-01 -5.51196218e-01
2.64184982e-01 4.32122648e-02 -1.58009171e+00 -3.08822930e-01
-1.08667091e-01 6.12944663e-01 1.36854231e+00 2.09292769e-01
-7.56830692e-01 1.06086147e+00 6.97916842e+00 7.35658228e-01
-1.71444309e+00 6.97747692e-02 5.58806837e-01 -1.05138552e+00
-6.40775710e-02 -3.59747618e-01 -1.10573113e+00 6.41460836e-01
1.41142201e+00 6.37497976e-02 6.16523445e-01 9.45759296e-01
-4.67537604e-02 -2.88964480e-01 -1.00084770e+00 1.62628639e+00
-1.73123226e-01 -1.40904212e+00 -6.46065116e-01 2.58029699e-02
5.38206220e-01 6.85859144e-01 4.58415926e-01 3.41017097e-01
3.52146894e-01 -9.44795072e-01 5.21370292e-01 2.24505618e-01
1.48982835e+00 -6.81630790e-01 2.26465374e-01 1.48654938e-01
-1.22092116e+00 -4.12229866e-01 -1.68248653e-01 -2.37398013e-01
1.45532981e-01 4.13770050e-01 -6.09105945e-01 -5.14934421e-01
9.41350818e-01 1.38136059e-01 -9.70795676e-02 6.10886037e-01
6.24876022e-02 7.54302859e-01 -8.19842517e-01 -6.12992406e-01
-1.03467405e-02 3.85775477e-01 3.36893290e-01 1.21274948e+00
3.56059819e-01 5.39277852e-01 -2.13981017e-01 4.62492377e-01
-4.91247088e-01 -6.61081851e-01 -5.19176364e-01 -2.02317521e-01
1.13050854e+00 1.12787151e+00 -1.25587976e+00 -2.15926338e-02
-1.29992455e-01 1.31351197e+00 1.51769951e-01 5.01676083e-01
-7.80663431e-01 -8.12376440e-01 9.05510902e-01 -1.42693920e-02
2.13844165e-01 -8.22241247e-01 -9.43439186e-01 -7.61997104e-01
1.40244201e-01 -3.73672575e-01 -1.14247426e-01 -5.24563313e-01
-6.54829800e-01 6.20688856e-01 -8.30586493e-01 -7.93879092e-01
-9.14995596e-02 -9.09056246e-01 -4.57768738e-01 2.97595948e-01
-8.52983356e-01 -5.28503537e-01 -3.12696129e-01 2.30805263e-01
3.79091471e-01 -4.22216147e-01 1.00707233e+00 3.52654845e-01
-6.31527245e-01 1.04650414e+00 2.87460357e-01 3.80749889e-02
6.79314673e-01 -6.76105499e-01 9.38085973e-01 7.32626796e-01
-6.73949867e-02 6.02589130e-01 6.86517417e-01 -7.78031886e-01
-2.20156693e+00 -9.04172122e-01 3.20909709e-01 -2.18376532e-01
6.77115023e-01 -1.04083133e+00 -5.62739015e-01 5.32990992e-01
-2.86758039e-02 4.45911437e-01 7.17299342e-01 1.34514883e-01
-4.56249774e-01 -4.56339419e-01 -1.03184199e+00 1.14435065e+00
1.68197989e+00 -7.32040048e-01 2.67212898e-01 -1.87828332e-01
6.30612791e-01 -5.97561181e-01 -3.84561002e-01 4.04397339e-01
9.62151647e-01 -1.45170256e-01 5.42216480e-01 -4.49368984e-01
-6.50690913e-01 -1.78671807e-01 -1.80188656e-01 -6.73012853e-01
-2.67924815e-01 -1.07755780e+00 -4.62324768e-01 4.23387051e-01
6.88900888e-01 -5.96936703e-01 1.45188177e+00 1.02122104e+00
-3.57101023e-01 -6.21040821e-01 -1.28100264e+00 -9.86741662e-01
-5.31121194e-01 -6.37182176e-01 -1.74351241e-02 7.98644647e-02
5.68363011e-01 2.77913868e-01 -2.47391552e-01 6.64312616e-02
3.34977686e-01 -2.32083902e-01 2.90167779e-01 -4.67962563e-01
-7.01349080e-01 -5.46579421e-01 -6.23203695e-01 -1.53712451e+00
-1.01990469e-01 -2.47018918e-01 4.68574047e-01 -1.02872169e+00
-6.47180527e-02 -4.73623276e-01 -6.65136814e-01 4.22372729e-01
1.05944455e-01 2.89660156e-01 -8.26266855e-02 -2.30144590e-01
-1.17149520e+00 3.65514338e-01 2.53352821e-01 -2.78669260e-02
-4.13649499e-01 -3.15831631e-01 -2.25943819e-01 3.66215795e-01
9.17137563e-01 -4.80393827e-01 -6.38906598e-01 -1.09021640e+00
2.97399521e-01 1.27275124e-01 -1.41523927e-01 -1.68816161e+00
8.18096280e-01 -2.99419630e-02 5.87179363e-01 -1.26013666e-01
6.92619801e-01 -6.69523716e-01 -3.07882130e-02 8.38940322e-01
-2.08250970e-01 4.20231819e-02 9.16344702e-01 7.65035033e-01
4.85046118e-01 3.29323828e-01 5.06117284e-01 3.91575620e-02
-9.66019511e-01 2.41247401e-01 -8.33884895e-01 9.88454744e-02
8.07854831e-01 -3.86172384e-01 -5.20204842e-01 -3.89542431e-01
-6.25624001e-01 1.16040707e-01 4.04531240e-01 1.84803382e-01
6.75195098e-01 -8.31146657e-01 7.47380778e-02 3.47305030e-01
-6.99760467e-02 -3.29641193e-01 6.21801853e-01 2.13411316e-01
-3.08233947e-01 6.48732007e-01 -4.05791044e-01 -3.43679965e-01
-1.00714052e+00 3.06496948e-01 3.90755653e-01 3.07535589e-01
-3.33134621e-01 1.23525822e+00 -3.82294118e-01 4.54732269e-01
5.64733565e-01 -1.93158522e-01 6.32403076e-01 -6.42063439e-01
4.74048078e-01 2.48814464e-01 -2.80836727e-02 3.91223162e-01
-5.79211771e-01 3.44797283e-01 -1.41107008e-01 -6.69750214e-01
9.02081192e-01 5.90898991e-02 5.85664570e-01 6.77503705e-01
7.73301125e-01 -2.83564925e-01 -1.92797089e+00 -2.70145506e-01
1.20770648e-01 1.15203880e-01 2.86384791e-01 -9.76770699e-01
-9.32842135e-01 6.01714253e-01 1.18021452e+00 -6.39484942e-01
1.21386135e+00 -1.28771588e-01 1.08548176e+00 9.01860535e-01
1.01112151e+00 -1.56938970e+00 3.03935766e-01 6.96731150e-01
-4.54695933e-02 -9.55035210e-01 -2.32252851e-01 8.28730837e-02
-1.78931236e-01 9.50956047e-01 1.06244838e+00 -9.77554247e-02
7.92577624e-01 1.24341583e+00 9.50372443e-02 4.97458503e-02
-1.10987973e+00 -6.93543404e-02 -1.62410468e-01 7.08062291e-01
1.69296101e-01 2.16836214e-01 -3.76836322e-02 1.11832881e+00
2.52512902e-01 3.46815377e-01 2.38047674e-01 1.16549313e+00
-8.33312213e-01 -7.40526795e-01 3.29847664e-01 7.53680229e-01
-1.66027948e-01 -2.80481488e-01 -1.16036810e-01 2.42877722e-01
-2.95791775e-01 6.73534572e-01 5.59456706e-01 -8.49020660e-01
1.00002751e-01 2.84701377e-01 5.25315166e-01 -4.40352827e-01
-5.66295028e-01 1.80013657e-01 -1.11962937e-01 -7.78172553e-01
5.08885384e-01 -3.75253677e-01 -1.97655356e+00 3.74714695e-02
1.04786634e-01 -5.40789902e-01 1.13509595e+00 6.38130486e-01
1.20788193e+00 7.06890285e-01 7.58016631e-02 -5.67626894e-01
-3.44319433e-01 -5.88617861e-01 -3.64663184e-01 -4.53036249e-01
8.37218910e-02 -2.95093089e-01 -1.04723923e-01 -2.84000814e-01] | [8.264657974243164, 2.5043129920959473] |
6552f579-f1b5-4a30-841e-384555cf397a | neural-pbir-reconstruction-of-shape-material | 2304.13445 | null | https://arxiv.org/abs/2304.13445v1 | https://arxiv.org/pdf/2304.13445v1.pdf | Neural-PBIR Reconstruction of Shape, Material, and Illumination | Reconstructing the shape and spatially varying surface appearances of a physical-world object as well as its surrounding illumination based on 2D images (e.g., photographs) of the object has been a long-standing problem in computer vision and graphics. In this paper, we introduce a robust object reconstruction pipeline combining neural based object reconstruction and physics-based inverse rendering (PBIR). Specifically, our pipeline firstly leverages a neural stage to produce high-quality but potentially imperfect predictions of object shape, reflectance, and illumination. Then, in the later stage, initialized by the neural predictions, we perform PBIR to refine the initial results and obtain the final high-quality reconstruction. Experimental results demonstrate our pipeline significantly outperforms existing reconstruction methods quality-wise and performance-wise. | ['Zhao Dong', 'Shuang Zhao', 'Jia-Bin Huang', 'Carl Marshall', 'Cheng Zhang', 'Kai Yan', 'Zhengqin Li', 'Guangyan Cai', 'Cheng Sun'] | 2023-04-26 | null | null | null | null | ['object-reconstruction', 'inverse-rendering'] | ['computer-vision', 'computer-vision'] | [ 5.72638154e-01 -1.33388162e-01 6.10962629e-01 -5.42259336e-01
-5.89933157e-01 -3.21439743e-01 5.06513655e-01 -2.31256694e-01
2.58168876e-02 4.06208396e-01 6.68500662e-02 -3.14777419e-02
1.32049724e-01 -7.41083026e-01 -9.38340962e-01 -5.10037541e-01
4.23402816e-01 3.74404132e-01 3.52655828e-01 1.72148496e-01
2.91386366e-01 9.58607078e-01 -1.65232933e+00 2.86929667e-01
6.45027041e-01 1.20772028e+00 2.25309312e-01 7.50956178e-01
-1.13388903e-01 6.11750305e-01 -2.31538787e-02 -2.67982095e-01
2.53373295e-01 -1.35582969e-01 -3.72324616e-01 2.62012988e-01
6.47986174e-01 -7.51521587e-01 -2.36234099e-01 8.70993137e-01
1.74815312e-01 2.80459285e-01 5.25942981e-01 -5.40948153e-01
-9.12625134e-01 -2.79843628e-01 -6.88668728e-01 -4.84998524e-01
3.53454769e-01 4.01039898e-01 6.58272207e-01 -1.15926135e+00
4.15512085e-01 1.43115187e+00 7.61515796e-01 3.97670090e-01
-1.55828989e+00 -5.10854363e-01 3.07273239e-01 -1.47874206e-01
-1.24395096e+00 -6.04006171e-01 1.12558353e+00 -2.80272186e-01
6.68681324e-01 2.71283865e-01 7.45059490e-01 5.92757881e-01
6.91280514e-02 3.61377388e-01 1.11258709e+00 -2.17863336e-01
1.32823452e-01 -1.88116297e-01 -5.87444864e-02 7.65030980e-01
-5.99380434e-02 2.70567715e-01 -3.88103902e-01 -2.26390183e-01
1.31178856e+00 3.15408148e-02 -5.92529237e-01 -3.02781433e-01
-1.03471971e+00 1.83751538e-01 6.33560598e-01 -3.56185704e-01
-6.82701468e-01 2.96621382e-01 -2.86825806e-01 -2.40745112e-01
8.10740530e-01 2.92973071e-01 -5.36242127e-01 1.13823302e-01
-6.87168181e-01 2.77202934e-01 5.93988776e-01 4.83758003e-01
9.74410713e-01 3.92817184e-02 -3.56368758e-02 1.01084661e+00
7.89729416e-01 6.45761490e-01 -1.10519163e-01 -1.35376585e+00
-9.83371362e-02 6.15349650e-01 5.32583952e-01 -9.70937073e-01
-2.13655844e-01 -2.17977777e-01 -6.09695554e-01 7.21810758e-01
2.51328856e-01 2.63839185e-01 -1.12768722e+00 1.28026390e+00
8.06573451e-01 6.98901594e-01 -2.32259959e-01 1.19915950e+00
9.85699892e-01 8.14188600e-01 -2.71634430e-01 -5.31095080e-02
1.04521894e+00 -7.72896051e-01 -3.52331281e-01 -2.40063682e-01
-4.16287065e-01 -8.92014027e-01 1.01390553e+00 5.47989190e-01
-1.49487460e+00 -5.83765686e-01 -7.56287277e-01 -4.72040057e-01
4.02067661e-01 7.66003728e-02 4.29646403e-01 1.65297478e-01
-1.10615146e+00 9.01683807e-01 -1.10249650e+00 1.70976803e-01
5.48681319e-01 2.43743122e-01 -1.98510900e-01 -4.18748885e-01
-3.19329441e-01 4.34436440e-01 -7.82293677e-02 3.68505508e-01
-7.72748709e-01 -1.16820967e+00 -7.13634908e-01 -1.55246779e-01
1.03066050e-01 -8.00842822e-01 1.23999631e+00 -1.06562150e+00
-2.06266665e+00 8.22493732e-01 -2.29165971e-01 1.41058192e-01
4.11301494e-01 -1.95204929e-01 4.26497348e-02 2.84732990e-02
-5.28195202e-01 4.69233602e-01 8.56323719e-01 -1.92756057e+00
-2.55530179e-01 -4.82750654e-01 -5.02299219e-02 2.63937563e-01
2.05879942e-01 -4.91078049e-02 -8.73639345e-01 -2.85728693e-01
4.86907899e-01 -6.12587512e-01 -2.45777786e-01 9.69177663e-01
-3.33170652e-01 2.68588960e-01 7.09963858e-01 -9.58698630e-01
3.91658127e-01 -2.24343157e+00 2.32543983e-02 2.26559445e-01
2.13975385e-01 1.50999561e-01 -4.14400667e-01 -1.90881103e-01
-2.73810104e-02 -3.04962516e-01 -2.12054282e-01 -8.11434984e-01
-1.57567576e-01 5.83768636e-02 -3.58218849e-01 6.00636959e-01
3.21927726e-01 9.44630802e-01 -7.32556581e-01 -8.37136284e-02
4.98397827e-01 1.22606766e+00 -6.25101745e-01 6.28783882e-01
-6.33157253e-01 8.95092428e-01 -3.55421662e-01 5.54364920e-01
1.00432181e+00 -4.59647834e-01 -8.48428160e-02 -6.42186522e-01
-1.93167746e-01 1.40097022e-01 -1.03150928e+00 1.75471544e+00
-7.81316936e-01 5.12625933e-01 4.20235455e-01 -4.44445580e-01
9.73440170e-01 8.64196345e-02 5.60198665e-01 -7.51359820e-01
9.82571468e-02 -1.24547705e-02 -2.71111071e-01 -2.42738545e-01
5.01327991e-01 -2.51752973e-01 7.16222107e-01 4.89372492e-01
-4.57634687e-01 -4.72930223e-01 -7.09685624e-01 -2.42235020e-01
8.11518610e-01 8.47203732e-01 -9.16008577e-02 2.85948724e-01
4.26320255e-01 -9.70559716e-02 5.03366470e-01 3.15743178e-01
3.13450992e-01 1.33261371e+00 -2.75007021e-02 -5.39742351e-01
-1.26273930e+00 -1.25688922e+00 -1.50140136e-01 6.21202469e-01
3.94807696e-01 8.94422978e-02 -6.03607774e-01 -8.42115805e-02
1.11719094e-01 6.89315259e-01 -6.27194405e-01 -5.02480678e-02
-6.86529338e-01 -4.41841900e-01 -2.55301952e-01 4.14154530e-01
3.43618721e-01 -1.18036389e+00 -5.35352051e-01 1.90405101e-01
1.35677591e-01 -1.17549670e+00 -3.11291635e-01 -4.74675357e-01
-8.83451521e-01 -8.47707272e-01 -4.76546258e-01 -3.27562988e-01
9.26096022e-01 5.58339655e-01 1.41068757e+00 3.49259228e-01
-4.68631625e-01 4.02709097e-01 3.04900371e-02 -2.42805168e-01
-3.34648758e-01 -8.29582334e-01 -4.06913161e-01 4.90803808e-01
-4.35911447e-01 -9.61124897e-01 -8.77198339e-01 3.93448979e-01
-9.26134646e-01 6.26123726e-01 4.44598824e-01 3.41904908e-01
1.10789144e+00 -1.52277827e-01 -2.10923124e-02 -7.60769904e-01
1.53220043e-01 -2.20010787e-01 -1.02290118e+00 1.24168247e-01
-2.44639158e-01 -8.58494490e-02 5.90179026e-01 -6.49910688e-01
-1.57523561e+00 3.55658710e-01 -3.40688556e-01 -8.64503384e-01
-2.06312776e-01 2.06883132e-01 -3.05740267e-01 -1.53201714e-01
5.87885559e-01 2.13541970e-01 -3.42209376e-02 -7.84119666e-01
2.14959085e-01 3.32228899e-01 8.37207496e-01 -5.54787874e-01
9.36125040e-01 9.02285993e-01 -8.30790680e-03 -7.54752398e-01
-1.04950011e+00 -2.12810352e-01 -4.75013822e-01 -4.23386604e-01
7.00773418e-01 -8.72138262e-01 -7.79850423e-01 6.34853959e-01
-1.41931677e+00 -7.66340971e-01 -3.72016728e-01 2.86976993e-01
-5.15939534e-01 6.61473125e-02 -4.67333168e-01 -9.33989704e-01
-4.55122918e-01 -9.98574018e-01 1.58682954e+00 4.18330103e-01
2.54377633e-01 -7.73819506e-01 7.49538885e-03 4.74930823e-01
3.38561416e-01 3.68604094e-01 8.51710200e-01 4.10471052e-01
-1.20166469e+00 -6.43829107e-02 -7.76662946e-01 3.75762492e-01
2.45963365e-01 2.19871521e-01 -1.27717352e+00 -4.67176810e-02
-7.19080418e-02 -1.16227858e-01 5.09654820e-01 5.87602377e-01
1.57665205e+00 -1.36831716e-01 -1.19490728e-01 1.08036780e+00
1.68298924e+00 -1.15654014e-01 7.16222405e-01 -1.34681597e-01
1.01564169e+00 5.34059286e-01 5.49741387e-01 5.10025799e-01
3.85144621e-01 8.27545762e-01 7.44843543e-01 -3.01714718e-01
-7.18288660e-01 -2.39217728e-01 1.50504097e-01 4.84080702e-01
-1.64822429e-01 -7.25364462e-02 -5.87765574e-01 1.41363978e-01
-1.57864833e+00 -4.25542831e-01 -4.02308911e-01 2.38102150e+00
8.59025717e-01 -2.15146154e-01 -2.52831548e-01 -2.78845608e-01
4.04783994e-01 -1.16924308e-01 -9.95010257e-01 -1.70529887e-01
5.00398315e-02 3.07130426e-01 1.67584151e-01 6.95632637e-01
-4.81673419e-01 8.06339920e-01 5.93998194e+00 3.33010256e-01
-1.37789416e+00 5.95328584e-02 6.54270053e-01 -1.31333023e-01
-7.33235180e-01 -6.84995130e-02 -4.12593186e-01 1.86803013e-01
6.09349370e-01 1.19100437e-01 9.30147529e-01 4.14263695e-01
2.79706955e-01 -1.21495791e-01 -1.01877522e+00 1.03917110e+00
8.78751278e-02 -1.36969829e+00 -1.60915162e-02 5.49632162e-02
7.66712964e-01 2.43372411e-01 -9.88200214e-03 -2.91109085e-01
5.46033859e-01 -9.65709627e-01 9.13082421e-01 9.81151998e-01
8.98188889e-01 -4.25702929e-01 2.08078519e-01 4.02204543e-01
-9.33159113e-01 3.01607251e-01 -2.31167838e-01 1.90033451e-01
3.87704134e-01 7.47396111e-01 -5.57952702e-01 3.62162828e-01
6.71735406e-01 8.04967344e-01 -3.18043083e-01 1.18788695e+00
-2.85856247e-01 5.04487097e-01 -3.93251210e-01 3.90277803e-01
-3.30844700e-01 -4.61861908e-01 4.08785135e-01 7.05932319e-01
1.19400144e-01 6.20292127e-01 2.28362739e-01 1.42077172e+00
-1.59693390e-01 -1.72356650e-01 -2.53245831e-01 3.17198187e-01
6.55105412e-02 1.46300220e+00 -5.30777395e-01 1.32863608e-03
-3.04684669e-01 1.12146795e+00 5.80645740e-01 5.80175400e-01
-7.05972254e-01 2.84181148e-01 7.72617877e-01 5.35560489e-01
2.18953252e-01 -2.87856281e-01 -4.81276780e-01 -1.02611828e+00
1.84699893e-01 -3.33552182e-01 -3.62754911e-01 -1.47060704e+00
-1.34937119e+00 4.43688065e-01 -4.07745153e-01 -9.39345241e-01
2.01413199e-01 -6.64628923e-01 -6.28739297e-01 1.25661218e+00
-1.83414352e+00 -1.37332904e+00 -6.12381399e-01 3.42589945e-01
5.18270135e-01 5.18380225e-01 7.64877915e-01 -3.91142210e-03
-5.12195110e-01 -8.76880810e-02 9.24634114e-02 -1.59416333e-01
2.34851331e-01 -9.24800813e-01 6.13363206e-01 5.62497735e-01
-3.48412767e-02 4.48448032e-01 5.18046677e-01 -5.65290391e-01
-1.77996480e+00 -1.17337692e+00 2.94126514e-02 -3.12265843e-01
2.04970047e-01 -2.69667923e-01 -1.28384984e+00 4.27992553e-01
-2.86372691e-01 6.14747286e-01 1.86018154e-01 -2.27225795e-01
-4.24008608e-01 -1.30306050e-01 -1.23451877e+00 5.59382141e-01
9.39622939e-01 -4.58465040e-01 -1.00136146e-01 2.99315184e-01
6.36907697e-01 -8.97521019e-01 -7.64628410e-01 3.53698879e-01
9.43848073e-01 -1.15319216e+00 1.39076626e+00 -2.08401725e-01
6.56784415e-01 -5.31488657e-01 -1.07541882e-01 -1.11016655e+00
-3.90680492e-01 -4.26194757e-01 -8.14759061e-02 1.04682910e+00
6.90160468e-02 -5.03903210e-01 7.62578607e-01 1.13950336e+00
-2.68419057e-01 -7.92568624e-01 -5.82761526e-01 -2.77486473e-01
-3.34863007e-01 -5.32089114e-01 6.38998747e-01 6.11746311e-01
-1.00495195e+00 1.08501711e-03 -1.33259386e-01 6.21598482e-01
8.91730130e-01 6.48852408e-01 9.83551443e-01 -1.40812814e+00
-4.57598835e-01 -1.99614167e-01 -1.40415542e-02 -1.24805903e+00
3.92778188e-01 -5.51951826e-01 5.65257370e-01 -1.74992824e+00
2.17402056e-01 -8.26503754e-01 -1.16991855e-01 3.21581364e-01
-2.90731877e-01 6.28340542e-01 -1.19220652e-02 2.29837328e-01
-2.98359394e-01 6.16043091e-01 1.55112147e+00 2.31112778e-01
-2.21839175e-01 -3.71751711e-02 -5.57155728e-01 8.34435999e-01
4.31945175e-01 -3.08101773e-01 -9.47458446e-02 -8.65160704e-01
9.14733484e-02 6.65419400e-02 7.34270453e-01 -5.91974258e-01
-1.75362304e-01 -4.35529292e-01 7.22206712e-01 -6.11852825e-01
8.94222558e-01 -9.92763460e-01 5.39140105e-01 1.03454009e-01
-1.68088228e-01 -3.95680845e-01 1.34939477e-01 6.29201353e-01
3.18626791e-01 2.01825239e-02 1.05354428e+00 -1.06808305e-01
-3.09679449e-01 6.68383837e-01 4.28946823e-01 -2.50665694e-01
5.51509857e-01 -4.21320051e-01 -5.17003685e-02 -1.83819249e-01
-2.93749273e-01 -7.73383528e-02 9.61269915e-01 2.36086264e-01
1.08193731e+00 -1.07273042e+00 -8.75230432e-01 4.36662406e-01
-3.75890732e-03 6.36796474e-01 2.71239311e-01 5.07764280e-01
-6.36373043e-01 -3.09561253e-01 6.94613233e-02 -7.62384117e-01
-1.18023670e+00 3.12280685e-01 5.94087005e-01 2.81642169e-01
-1.18351960e+00 8.66676688e-01 5.14614701e-01 -5.69124997e-01
-3.53703974e-03 -3.44876707e-01 1.62784532e-01 -8.55498672e-01
5.84594965e-01 1.22416086e-01 4.02367227e-02 -7.56437540e-01
-7.01035783e-02 8.47325981e-01 1.87071681e-01 -1.44052729e-01
1.78909826e+00 -6.90607354e-02 -3.26675892e-01 3.95470589e-01
1.09202302e+00 -5.86141609e-02 -1.96487343e+00 -4.64425892e-01
-6.81958318e-01 -9.63288128e-01 6.36152387e-01 -9.24124777e-01
-1.51030362e+00 7.68013656e-01 4.57713127e-01 -3.17244172e-01
1.23624110e+00 1.10954687e-03 9.07726228e-01 8.58410634e-03
2.64139265e-01 -5.53117096e-01 -5.26294447e-02 2.03633785e-01
1.17654955e+00 -9.89864767e-01 4.03121948e-01 -7.24994719e-01
-7.49677047e-02 1.05520999e+00 5.45477331e-01 -2.88342863e-01
6.43403709e-01 3.76658946e-01 3.76647823e-02 -1.97880179e-01
-6.71084583e-01 1.84191972e-01 6.79873109e-01 5.18166065e-01
2.63713181e-01 -1.00432940e-01 4.86678004e-01 2.15497792e-01
1.37685984e-01 1.24339134e-01 2.34534830e-01 6.05596960e-01
-1.92949280e-01 -8.39365661e-01 -5.34662247e-01 3.51118475e-01
-9.10451636e-02 -3.03881764e-02 -2.38672942e-01 3.56550813e-02
-4.98289615e-02 5.43668330e-01 2.54466623e-01 2.52618361e-02
4.70795989e-01 -4.63693351e-01 7.54445255e-01 -7.05195069e-01
-4.07793432e-01 9.15463343e-02 -1.48696437e-01 -9.36466277e-01
-4.52435642e-01 -7.21060932e-01 -1.46078622e+00 7.04769511e-03
-3.89178514e-01 -5.64399600e-01 1.19642508e+00 7.25691557e-01
4.70180392e-01 4.57584560e-01 7.58673906e-01 -1.55574083e+00
-1.15164578e-01 -4.79129732e-01 -4.50585246e-01 4.62651044e-01
6.13347948e-01 -5.22092223e-01 -2.29810387e-01 2.14387730e-01] | [9.672821044921875, -3.117753267288208] |
79d566f6-9be4-4730-a539-f018123186d6 | mert-acoustic-music-understanding-model-with | 2306.00107 | null | https://arxiv.org/abs/2306.00107v2 | https://arxiv.org/pdf/2306.00107v2.pdf | MERT: Acoustic Music Understanding Model with Large-Scale Self-supervised Training | Self-supervised learning (SSL) has recently emerged as a promising paradigm for training generalisable models on large-scale data in the fields of vision, text, and speech. Although SSL has been proven effective in speech and audio, its application to music audio has yet to be thoroughly explored. This is primarily due to the distinctive challenges associated with modelling musical knowledge, particularly its tonal and pitched characteristics of music. To address this research gap, we propose an acoustic Music undERstanding model with large-scale self-supervised Training (MERT), which incorporates teacher models to provide pseudo labels in the masked language modelling (MLM) style acoustic pre-training. In our exploration, we identified a superior combination of teacher models, which outperforms conventional speech and audio approaches in terms of performance. This combination includes an acoustic teacher based on Residual Vector Quantization - Variational AutoEncoder (RVQ-VAE) and a musical teacher based on the Constant-Q Transform (CQT). These teachers effectively guide our student model, a BERT-style transformer encoder, to better model music audio. In addition, we introduce an in-batch noise mixture augmentation to enhance the representation robustness. Furthermore, we explore a wide range of settings to overcome the instability in acoustic language model pre-training, which allows our designed paradigm to scale from 95M to 330M parameters. Experimental results indicate that our model can generalise and perform well on 14 music understanding tasks and attains state-of-the-art (SOTA) overall scores. The code and models are online: https://github.com/yizhilll/MERT. | ['Jie Fu', 'Yike Guo', 'Wenhao Huang', 'Yemin Shi', 'Gus Xia', 'Wenhu Chen', 'Ruibo Liu', 'Roger Dannenberg', 'Norbert Gyenge', 'Emmanouil Benetos', 'Anton Ragni', 'Chenghua Lin', 'Hanzhi Yin', 'Xingran Chen', 'Yinghao Ma', 'Ge Zhang', 'Ruibin Yuan', 'Yizhi Li'] | 2023-05-31 | null | null | null | null | ['quantization'] | ['methodology'] | [ 1.94349989e-01 -6.67676106e-02 -1.60397142e-02 -1.95746630e-01
-1.17067850e+00 -5.09345233e-01 4.88896430e-01 -2.64419019e-01
-3.30763817e-01 3.12431790e-02 2.32333601e-01 -2.17237145e-01
-1.61572248e-01 -4.31022882e-01 -7.73855925e-01 -5.76247036e-01
1.55506834e-01 3.90732855e-01 -4.12679352e-02 -1.66954204e-01
-4.32877615e-02 4.28623855e-02 -1.78597617e+00 3.38372171e-01
8.35896313e-01 8.99212360e-01 5.23159385e-01 7.77357042e-01
-1.62740380e-01 7.80374110e-01 -6.35224104e-01 -3.05290878e-01
7.10481554e-02 -4.75700676e-01 -5.83968222e-01 1.52037051e-02
5.33029556e-01 -2.26790249e-01 -2.85434395e-01 9.31422055e-01
8.45124841e-01 3.54593903e-01 5.86618245e-01 -1.16416681e+00
-6.16112113e-01 1.21666098e+00 -3.61014724e-01 -3.14235128e-02
1.88265771e-01 3.95101011e-02 1.19597709e+00 -1.11805034e+00
-1.69880968e-02 1.39292347e+00 7.69111097e-01 6.48693204e-01
-1.30796957e+00 -1.24503267e+00 -2.18872920e-01 4.92975861e-01
-1.63133466e+00 -7.86946297e-01 1.01238155e+00 -3.26065987e-01
8.47952306e-01 2.10363820e-01 6.25102639e-01 1.25977361e+00
-3.31671923e-01 9.86279309e-01 9.96363044e-01 -6.77750707e-01
2.40312040e-01 2.35056549e-01 -4.14092600e-01 3.74496013e-01
-7.05893278e-01 2.07093835e-01 -1.05550003e+00 -1.14570901e-01
7.92410970e-01 -4.75155234e-01 -9.27117616e-02 -2.40138039e-01
-1.01050723e+00 9.21055377e-01 1.03532046e-01 3.10416281e-01
-1.89363152e-01 3.15028667e-01 4.08566773e-01 3.55582952e-01
4.25011218e-01 3.99415523e-01 -5.17271698e-01 -5.10536373e-01
-1.29796398e+00 4.25340459e-02 5.97889662e-01 6.90077603e-01
4.57964212e-01 7.81444013e-01 1.16459886e-02 1.50047553e+00
4.49017018e-01 5.50475955e-01 8.08201373e-01 -1.19908392e+00
2.36315638e-01 8.88248160e-02 -4.82142121e-01 -4.25325394e-01
-1.12439819e-01 -8.03456426e-01 -8.74156058e-01 7.13773891e-02
1.89152323e-02 2.00272068e-01 -7.04007089e-01 1.86886656e+00
2.30785683e-01 8.93000424e-01 2.26290762e-01 7.94298887e-01
1.02798092e+00 7.61794269e-01 -3.14839818e-02 -1.82626173e-01
1.21347618e+00 -1.14022696e+00 -7.29886591e-01 -7.17986077e-02
3.61510217e-01 -1.05427277e+00 1.37775338e+00 7.56386638e-01
-1.14514029e+00 -1.20744014e+00 -8.15114498e-01 -6.90412596e-02
-3.40250172e-02 3.56765509e-01 3.32442015e-01 7.15255320e-01
-1.05288231e+00 6.46791399e-01 -8.16985071e-01 -3.47779468e-02
2.99039394e-01 3.72860044e-01 4.08994928e-02 2.62776196e-01
-1.23319268e+00 5.15346050e-01 5.32588422e-01 -1.36393577e-01
-1.21891713e+00 -9.69571948e-01 -7.62336195e-01 1.70656458e-01
2.88204938e-01 -5.79280674e-01 1.68687510e+00 -8.52737308e-01
-2.06690431e+00 5.09594560e-01 -6.74823895e-02 -7.08704472e-01
1.72428638e-01 -3.02927494e-01 -4.61090714e-01 1.60121456e-01
-1.39545336e-01 1.06452560e+00 1.28228176e+00 -1.00665891e+00
-5.65934896e-01 -3.26959342e-02 -3.21385890e-01 4.63158906e-01
-8.29143822e-01 3.89203914e-02 -5.05645752e-01 -1.12897289e+00
-1.26730110e-02 -1.04953563e+00 -2.67149173e-02 -4.43987399e-01
-1.26399025e-01 -4.69941139e-01 7.80304134e-01 -4.97433871e-01
1.42999303e+00 -2.51327825e+00 6.46915808e-02 -1.33415647e-02
-1.77084997e-01 3.54874760e-01 -3.42958301e-01 4.67613757e-01
-1.42590851e-01 -1.11994594e-01 -1.12921163e-01 -8.46617401e-01
1.98417962e-01 3.18169355e-01 -6.32439256e-01 -7.57203549e-02
1.18795261e-01 8.63749027e-01 -7.04251707e-01 -3.99615556e-01
3.95863175e-01 8.71604741e-01 -7.18746483e-01 2.90524870e-01
-2.30176449e-01 7.32741594e-01 6.71329768e-03 3.90033633e-01
3.42601120e-01 -2.50513423e-02 -1.36258021e-01 -1.51343212e-01
-1.60306171e-02 6.02221310e-01 -1.52478230e+00 2.14821839e+00
-5.80888450e-01 5.21019161e-01 1.98132545e-01 -9.00813401e-01
9.60812151e-01 7.08072543e-01 4.55487609e-01 -3.58032286e-01
-3.01757194e-02 2.34538704e-01 8.04438069e-02 -2.85489231e-01
3.75364661e-01 -2.71960855e-01 2.69675791e-01 2.25015506e-01
5.10243058e-01 -5.48967004e-01 -1.35034591e-01 8.27237219e-02
6.56009674e-01 6.43738806e-02 2.78325845e-02 9.34484303e-02
5.13939857e-01 -3.49733770e-01 4.08591926e-01 6.68136835e-01
4.36428562e-02 5.32358766e-01 -2.18589753e-01 3.06970656e-01
-8.77179503e-01 -1.08052087e+00 -2.51708478e-01 1.61381650e+00
-3.69728625e-01 -8.29918265e-01 -7.74206340e-01 9.00577940e-03
-7.60636702e-02 7.55281210e-01 -1.31410539e-01 -2.28123084e-01
-3.65690351e-01 -2.90332347e-01 9.56783354e-01 5.99581599e-01
1.93337604e-01 -1.25794458e+00 -1.49508104e-01 3.68981034e-01
-1.02287553e-01 -1.04477489e+00 -5.45553207e-01 3.18343401e-01
-8.50757122e-01 -6.09040022e-01 -6.57456636e-01 -8.21348667e-01
-1.89703643e-01 1.44512162e-01 9.11956310e-01 -4.65624452e-01
-1.97026059e-02 6.62726223e-01 -4.14484859e-01 -7.43622541e-01
-8.38202477e-01 1.69470653e-01 5.71588516e-01 9.18480977e-02
1.90840170e-01 -8.48949313e-01 -3.03309441e-01 2.16273159e-01
-9.76946414e-01 5.23506403e-02 5.35898983e-01 8.71518254e-01
7.63160229e-01 4.14015770e-01 7.90345311e-01 -4.83184874e-01
6.25142515e-01 -1.64675772e-01 -4.10394430e-01 -1.78319469e-01
-6.73982382e-01 -2.64824748e-01 6.09841943e-01 -9.52056766e-01
-8.63367975e-01 4.66610678e-02 -5.56846499e-01 -1.06090832e+00
-2.59675473e-01 5.44250786e-01 -1.41788030e-03 3.98976058e-02
5.01332998e-01 3.18427145e-01 1.81330115e-01 -7.60631502e-01
5.89613199e-01 1.03871107e+00 8.57691169e-01 -5.43007791e-01
8.46498787e-01 1.52084842e-01 -3.87838602e-01 -1.21042848e+00
-8.68717551e-01 -5.76712370e-01 -4.16717708e-01 -1.17213011e-01
5.86776316e-01 -1.38827217e+00 -7.21691310e-01 4.85340893e-01
-7.66526341e-01 -5.42167187e-01 -5.59468985e-01 8.39582682e-01
-7.67726064e-01 3.41378242e-01 -8.58790517e-01 -1.17236471e+00
-3.87685478e-01 -1.21944070e+00 1.03495038e+00 3.43021564e-02
-3.80645990e-01 -8.31965268e-01 8.40244591e-02 7.60344744e-01
4.58017379e-01 -5.37541628e-01 7.04408705e-01 -6.65802658e-01
-3.29138279e-01 3.80280852e-01 3.84760439e-01 7.43770242e-01
-1.35623338e-02 -1.68156028e-01 -1.59771180e+00 -5.03441393e-01
8.41768160e-02 -5.31406760e-01 8.44577551e-01 5.42709827e-01
1.30107129e+00 -4.54442911e-02 2.54749596e-01 7.08920896e-01
8.15974295e-01 1.28404304e-01 3.13666761e-01 2.43778691e-01
7.91926682e-01 3.89484406e-01 4.00251597e-01 4.09724176e-01
3.48083228e-01 8.45597088e-01 2.82652169e-01 -6.20210841e-02
-4.66170132e-01 -5.99018812e-01 7.07280874e-01 1.64219785e+00
2.35676646e-01 1.61555201e-01 -7.54583836e-01 4.35869187e-01
-1.56926560e+00 -7.85223842e-01 1.21571921e-01 2.04319501e+00
1.11679602e+00 1.40653670e-01 2.47447506e-01 6.51084542e-01
3.34603578e-01 1.11556947e-01 -6.46262348e-01 -2.45543912e-01
-1.07975863e-01 5.98461092e-01 3.78861353e-02 3.92169505e-01
-1.15112281e+00 1.20544696e+00 5.81750679e+00 1.53077662e+00
-1.15033090e+00 3.42692733e-01 8.22724327e-02 -1.14527732e-01
-1.50673375e-01 -2.04799712e-01 -7.98244894e-01 2.94713438e-01
1.41108823e+00 2.50639766e-01 7.55775213e-01 7.56266534e-01
4.51539129e-01 4.76983219e-01 -1.01479006e+00 1.23865676e+00
4.07244228e-02 -1.04631293e+00 1.21138386e-01 -5.53361811e-02
6.31956875e-01 2.09177136e-01 5.24630606e-01 7.43012130e-01
-3.70582528e-02 -1.03256583e+00 9.05978739e-01 2.14365035e-01
9.86638010e-01 -7.67989159e-01 2.68982470e-01 5.32406688e-01
-1.30349886e+00 -3.02767068e-01 -2.46553749e-01 -5.25908507e-02
-7.80586824e-02 2.27775797e-01 -1.14671791e+00 4.14099842e-01
8.98300231e-01 7.57777274e-01 -3.20072174e-01 8.27735901e-01
-1.88022003e-01 1.39463198e+00 -3.75606209e-01 2.77015477e-01
9.53445882e-02 -1.80640385e-01 6.00122094e-01 1.17851090e+00
2.93034434e-01 -1.19004354e-01 2.89660156e-01 7.48745322e-01
7.96534494e-02 3.27819407e-01 -2.52094418e-01 -2.51254678e-01
6.84240520e-01 9.24444020e-01 -2.94517785e-01 -2.95000106e-01
-2.91082144e-01 7.80322611e-01 1.05746120e-01 4.05229241e-01
-7.05382645e-01 3.76878157e-02 6.63648009e-01 -1.32066840e-02
5.82371712e-01 -2.19978407e-01 -4.37586941e-02 -9.95464742e-01
-2.25036576e-01 -1.19841933e+00 3.08568239e-01 -9.40359831e-01
-1.23445821e+00 5.15679181e-01 -1.58452570e-01 -1.32842076e+00
-4.71255630e-01 -4.19303149e-01 -3.92064601e-01 7.00583220e-01
-1.54362702e+00 -1.28942621e+00 8.14532861e-02 7.02670634e-01
9.39537823e-01 -6.26185060e-01 1.13461423e+00 4.81109887e-01
-3.35015655e-01 8.00484300e-01 2.09484845e-01 -9.48755536e-03
8.72417927e-01 -1.19835663e+00 2.63008833e-01 3.83749038e-01
1.03610516e+00 4.30205524e-01 5.90982854e-01 -2.36745059e-01
-1.37238133e+00 -9.95555937e-01 4.36617911e-01 -3.76016259e-01
8.84524047e-01 -4.24822688e-01 -1.10490775e+00 4.09757465e-01
1.22355923e-01 -3.35071862e-01 1.03854358e+00 1.37656614e-01
-4.76939499e-01 -2.27179408e-01 -4.23071116e-01 4.37839419e-01
7.85328805e-01 -1.10182929e+00 -6.37470782e-01 4.41652797e-02
8.70125473e-01 -4.67108071e-01 -1.04716504e+00 5.18554807e-01
5.66939175e-01 -8.15105379e-01 1.13420260e+00 -2.31775701e-01
1.66849807e-01 -4.29950356e-01 -3.08237463e-01 -1.38749564e+00
-2.31319234e-01 -7.76712656e-01 -3.77690583e-01 1.61708534e+00
2.16463760e-01 -2.38724172e-01 6.82893693e-01 -1.89481974e-01
-3.27073932e-01 -7.25188494e-01 -1.01810479e+00 -8.39085102e-01
5.28206341e-02 -1.19225907e+00 4.36101377e-01 1.01116943e+00
-1.38018340e-01 5.11190474e-01 -4.72467780e-01 2.91579068e-01
5.18951297e-01 -4.12186459e-02 9.30340290e-01 -1.18967557e+00
-7.44997919e-01 -5.91840625e-01 -1.80053011e-01 -1.31394136e+00
4.49509174e-01 -1.11932659e+00 -5.99809177e-02 -1.11713469e+00
4.45194682e-03 -3.92297357e-01 -6.07830346e-01 5.38480759e-01
-3.35881077e-02 4.23884273e-01 3.90133798e-01 2.64848500e-01
-4.70760107e-01 8.70134115e-01 1.01125658e+00 -2.56717503e-01
-4.33427960e-01 2.49978885e-01 -5.33883274e-01 7.96471536e-01
7.90395796e-01 -4.61359411e-01 -6.65891349e-01 -3.14178109e-01
-1.18658394e-01 1.74833700e-01 3.61567885e-01 -1.13697493e+00
3.67597401e-01 2.22235993e-01 1.12396255e-01 -6.61024690e-01
8.84416342e-01 -6.54861748e-01 1.89337835e-01 1.37526721e-01
-5.28790534e-01 -3.33949566e-01 4.31106806e-01 3.83610159e-01
-5.13668418e-01 -2.22976044e-01 6.56176567e-01 2.20283121e-01
-6.06822073e-01 1.38839290e-01 -2.72280723e-01 2.73538735e-02
3.01095366e-01 -1.02884546e-01 2.11500078e-01 -7.66249120e-01
-7.63316751e-01 1.19874462e-01 3.35856490e-02 6.28138006e-01
6.05883420e-01 -1.39178884e+00 -8.88492942e-01 4.46063906e-01
7.27911443e-02 6.59825886e-03 4.44727838e-01 6.72061801e-01
1.45247295e-01 3.52247864e-01 2.97985166e-01 -9.17569041e-01
-1.30152547e+00 3.46716344e-01 1.43657029e-01 9.00155678e-02
-6.67377889e-01 1.07100940e+00 1.84077457e-01 -7.93080509e-01
8.37145567e-01 -3.22376490e-01 -2.83788443e-01 -3.14406790e-02
4.68038529e-01 2.85144717e-01 -6.59658834e-02 -7.63218045e-01
-5.80247007e-02 6.39446378e-01 2.46861994e-01 -3.84652436e-01
1.23770010e+00 -6.35150671e-02 2.10963577e-01 1.02679193e+00
1.00591767e+00 2.84001738e-01 -1.25942051e+00 -3.59318554e-01
-7.20840693e-02 -3.96526866e-02 3.99940342e-01 -6.35381699e-01
-7.63918281e-01 1.06487346e+00 6.80842519e-01 2.06244305e-01
1.13906026e+00 1.24526650e-01 8.53928030e-01 2.83833236e-01
8.08945522e-02 -1.10910296e+00 2.84976482e-01 7.26187289e-01
8.11340809e-01 -1.07187414e+00 -3.83518636e-01 -1.95033282e-01
-7.42261231e-01 9.45942998e-01 3.46920788e-01 1.80317238e-01
7.18842506e-01 3.74353588e-01 5.23925662e-01 1.97365716e-01
-8.38425994e-01 -2.66756535e-01 6.01512194e-01 4.17002350e-01
5.54544330e-01 1.10895574e-01 3.53726208e-01 7.98006415e-01
-1.00575566e+00 -1.94230989e-01 1.78028032e-01 3.29672664e-01
-3.95039111e-01 -1.17762327e+00 -5.84296763e-01 1.49515241e-01
-4.59707856e-01 -3.62680793e-01 -6.50229976e-02 3.10595304e-01
1.64689079e-01 1.25401974e+00 3.84453200e-02 -5.15698791e-01
3.37496310e-01 3.97107035e-01 3.22318137e-01 -7.09198773e-01
-6.07862949e-01 8.63898456e-01 -2.77875036e-01 -2.25802884e-01
-5.35202622e-01 -4.82347071e-01 -1.30162811e+00 2.20475748e-01
-5.42118907e-01 4.43583518e-01 7.97600389e-01 7.21827805e-01
1.50592908e-01 9.15165961e-01 7.07239330e-01 -8.90856743e-01
-8.46822798e-01 -1.33726943e+00 -7.87509501e-01 2.43792504e-01
3.51330072e-01 -6.64987028e-01 -4.34643418e-01 1.30349979e-01] | [15.265071868896484, 5.396842956542969] |
04738f3c-9671-4b93-ae79-2060bb5578b5 | making-the-most-of-text-semantics-to-improve | 2204.09817 | null | https://arxiv.org/abs/2204.09817v4 | https://arxiv.org/pdf/2204.09817v4.pdf | Making the Most of Text Semantics to Improve Biomedical Vision--Language Processing | Multi-modal data abounds in biomedicine, such as radiology images and reports. Interpreting this data at scale is essential for improving clinical care and accelerating clinical research. Biomedical text with its complex semantics poses additional challenges in vision--language modelling compared to the general domain, and previous work has used insufficiently adapted models that lack domain-specific language understanding. In this paper, we show that principled textual semantic modelling can substantially improve contrastive learning in self-supervised vision--language processing. We release a language model that achieves state-of-the-art results in radiology natural language inference through its improved vocabulary and novel language pretraining objective leveraging semantics and discourse characteristics in radiology reports. Further, we propose a self-supervised joint vision--language approach with a focus on better text modelling. It establishes new state of the art results on a wide range of publicly available benchmarks, in part by leveraging our new domain-specific language model. We release a new dataset with locally-aligned phrase grounding annotations by radiologists to facilitate the study of complex semantic modelling in biomedical vision--language processing. A broad evaluation, including on this new dataset, shows that our contrastive learning approach, aided by textual-semantic modelling, outperforms prior methods in segmentation tasks, despite only using a global-alignment objective. | ['Ozan Oktay', 'Hoifung Poon', 'Javier Alvarez-Valle', 'Aditya Nori', 'Tristan Naumann', 'Maria Wetscherek', 'Stephanie Hyland', 'Anton Schwaighofer', 'Daniel C. Castro', 'Shruthi Bannur', 'Naoto Usuyama', 'Benedikt Boecking'] | 2022-04-21 | null | null | null | null | ['pneumonia-detection', 'phrase-grounding'] | ['medical', 'natural-language-processing'] | [ 6.79879665e-01 6.12871945e-01 -4.64380682e-01 -5.80996871e-01
-1.42200315e+00 -1.89109340e-01 6.48153365e-01 5.90667963e-01
-8.87787700e-01 5.08231819e-01 7.48453140e-01 -3.61110061e-01
-8.11308250e-02 -3.14471602e-01 -7.35946298e-01 -4.84979868e-01
1.74384877e-01 1.00854385e+00 2.35644072e-01 -1.50391340e-01
-9.25890356e-02 6.45426065e-02 -8.68046463e-01 8.84483457e-01
5.41812539e-01 5.63939869e-01 7.28689492e-01 8.88897181e-01
-3.44271958e-01 1.08132553e+00 -4.98171747e-02 -3.15126598e-01
-2.69767642e-01 -4.18747872e-01 -1.15119600e+00 3.13857734e-01
5.34683645e-01 5.11915199e-02 -7.93811381e-02 9.74529803e-01
5.94842494e-01 -3.83745879e-01 5.50629556e-01 -5.74494958e-01
-6.84457898e-01 6.52251363e-01 -5.46129584e-01 3.63199592e-01
1.31935209e-01 1.36974916e-01 1.13028586e+00 -6.05735838e-01
9.47059929e-01 1.19436967e+00 9.38216984e-01 8.22202981e-01
-1.41353405e+00 -1.29129171e-01 1.20710686e-01 1.53300881e-01
-1.10401011e+00 -3.33546788e-01 4.40049589e-01 -4.61738497e-01
1.37447512e+00 2.20461875e-01 4.39050168e-01 1.14236808e+00
3.60249639e-01 9.96286750e-01 1.16563475e+00 -6.27893746e-01
-2.67972052e-02 1.22897729e-01 3.12829942e-01 1.10970604e+00
1.97111547e-01 -2.78884441e-01 -5.01209319e-01 -1.14408433e-01
4.87295806e-01 -2.39399403e-01 -5.10878079e-02 -1.96435273e-01
-1.79239058e+00 8.90786588e-01 2.59054452e-01 3.89692098e-01
-3.83993924e-01 3.80145490e-01 6.56915307e-01 -2.61427388e-02
8.04913878e-01 3.37920994e-01 -6.35034323e-01 1.82040855e-01
-1.21121132e+00 -1.39887286e-02 8.82706106e-01 7.12617159e-01
4.18350846e-01 -3.62454116e-01 -2.27772921e-01 1.08717704e+00
5.77750385e-01 6.37714326e-01 7.39841282e-01 -8.22735250e-01
3.07925016e-01 3.38443220e-01 -5.27832329e-01 -6.62886143e-01
-9.57506061e-01 -6.52585328e-01 -8.59515250e-01 -2.88634449e-01
2.44191587e-01 2.01378644e-01 -1.36092782e+00 1.62428212e+00
8.60788822e-02 3.91841620e-01 3.10110807e-01 6.70097053e-01
1.41934931e+00 7.49549121e-02 6.75912857e-01 -3.21913004e-01
2.03219748e+00 -1.11958468e+00 -9.05042768e-01 -6.06244624e-01
1.06479096e+00 -7.94606209e-01 9.55498338e-01 1.60469681e-01
-1.04719651e+00 -2.98784345e-01 -6.83178484e-01 -4.11099702e-01
-1.65225819e-01 -1.75778978e-02 5.07054150e-01 4.09874827e-01
-1.36122811e+00 5.82416914e-03 -1.06199038e+00 -7.55457163e-01
7.95452416e-01 2.28015080e-01 -5.21420479e-01 -2.32960999e-01
-1.07188082e+00 1.30262804e+00 5.23009479e-01 -3.20858657e-01
-6.84986472e-01 -1.08532012e+00 -1.18787539e+00 -3.97613764e-01
5.74272633e-01 -1.54147911e+00 1.47639823e+00 -5.28426945e-01
-9.58641648e-01 1.70129621e+00 -3.33077013e-01 -8.75319779e-01
7.05891252e-01 1.12725817e-01 -1.06315181e-01 5.38056016e-01
5.34520566e-01 1.07623601e+00 6.00916922e-01 -1.30662739e+00
-5.53355992e-01 -5.13678551e-01 -2.86339015e-01 2.18364954e-01
-2.29790404e-01 2.84860982e-03 -6.94015086e-01 -8.68246377e-01
8.51649791e-03 -8.44156861e-01 -7.41885245e-01 1.06593978e-03
-4.33116972e-01 -2.18349651e-01 1.95358619e-01 -7.17592895e-01
1.02347410e+00 -1.81686187e+00 1.49921805e-01 -1.08844541e-01
6.71907365e-01 -5.73351979e-02 -2.73845694e-03 7.00739175e-02
-8.20122212e-02 2.16370329e-01 -5.74601829e-01 -8.08539510e-01
-1.95557699e-01 6.83275104e-01 -7.60308579e-02 3.48970801e-01
3.05798799e-01 1.28412032e+00 -1.01283884e+00 -1.27748930e+00
2.24070311e-01 2.90881574e-01 -6.48832500e-01 -2.27074265e-01
-4.98300284e-01 4.68943864e-01 -5.88064194e-01 5.93133390e-01
1.04134586e-02 -9.36898291e-01 2.05556467e-01 -5.35658956e-01
2.56322712e-01 -3.12897153e-02 -4.89817381e-01 2.34113550e+00
-5.13738692e-01 3.38376731e-01 -2.92232037e-02 -1.37226510e+00
2.80458570e-01 4.24887031e-01 7.85579324e-01 -7.67142832e-01
4.91119139e-02 1.76689088e-01 -2.05263138e-01 -1.10564780e+00
5.88425323e-02 -8.09231222e-01 -3.38370651e-02 2.43252277e-01
2.31493711e-01 -5.05021274e-01 1.75503835e-01 3.36838514e-01
1.09241271e+00 -3.03310864e-02 5.21123230e-01 -3.15848172e-01
6.05660498e-01 6.16506398e-01 1.17463499e-01 1.05853546e+00
-2.10063979e-01 6.75190091e-01 1.53362334e-01 -2.40298375e-01
-9.06548083e-01 -9.37650740e-01 -5.28858364e-01 1.13397276e+00
-2.73731440e-01 -4.75333154e-01 -6.25216484e-01 -7.11360395e-01
-6.22340292e-02 8.16256583e-01 -7.66594172e-01 4.28493582e-02
-5.37033498e-01 -1.11378014e+00 7.87135839e-01 5.97032070e-01
8.44930857e-02 -9.84948933e-01 -5.42114437e-01 3.75855356e-01
-4.15749729e-01 -2.02534556e+00 -2.48857543e-01 1.56822741e-01
-1.04959452e+00 -1.00515771e+00 -9.53350604e-01 -1.11264133e+00
5.65946400e-01 1.16839679e-02 1.46929634e+00 -1.64944995e-02
-7.87672281e-01 1.17979789e+00 -1.79702893e-01 -9.17052209e-01
-8.52143466e-01 -6.68915287e-02 -4.25722003e-01 -4.07568961e-01
3.76308441e-01 1.19783044e-01 -4.36769426e-01 -2.33499423e-01
-1.01469767e+00 5.54338932e-01 8.38727236e-01 1.19012809e+00
9.16933835e-01 -4.36183721e-01 5.22490203e-01 -1.26967609e+00
4.47175294e-01 -5.23829877e-01 2.09466413e-01 4.46591228e-01
-4.99633789e-01 2.01695651e-01 5.88012002e-02 -1.69220902e-02
-9.45663035e-01 -1.35480225e-01 -3.07826161e-01 -5.37676960e-02
-3.77906740e-01 8.17729950e-01 5.29433429e-01 1.59882247e-01
8.28737557e-01 3.62102956e-01 4.66029853e-01 -1.77533224e-01
6.50610149e-01 4.67477471e-01 6.87558770e-01 -4.91665602e-01
3.36968452e-01 1.09096515e+00 2.76743054e-01 -1.11157966e+00
-1.35439098e+00 -8.47796142e-01 -5.82516551e-01 1.34820908e-01
1.45706737e+00 -1.19276762e+00 -1.84918314e-01 1.74875960e-01
-1.11582255e+00 -2.71171778e-01 -4.07078117e-01 6.76353157e-01
-9.21445489e-01 6.41795516e-01 -6.85894907e-01 -3.42780948e-01
-4.24834698e-01 -1.30000806e+00 1.69111145e+00 -3.63528341e-01
-4.54351008e-01 -1.60499656e+00 4.10900414e-02 1.08570576e+00
7.98892304e-02 1.46454886e-01 1.08472002e+00 -8.80909681e-01
-2.35285863e-01 2.39022151e-01 -3.31421763e-01 2.59039223e-01
4.16050777e-02 -5.93832254e-01 -8.36860597e-01 -1.55607954e-01
1.27351522e-01 -5.08908510e-01 1.33009875e+00 7.40958631e-01
1.10636497e+00 1.27853945e-01 -5.85134804e-01 4.60350096e-01
1.32297957e+00 -4.30567920e-01 1.41100556e-01 4.55027580e-01
8.82112145e-01 7.32094347e-01 4.47789431e-01 1.04355209e-01
9.17434633e-01 3.71833593e-01 2.58460343e-01 -6.80943549e-01
-4.05357569e-01 2.74407327e-01 -1.67595282e-01 9.28670585e-01
7.40530714e-02 6.56341091e-02 -1.48001206e+00 7.84609199e-01
-1.94748771e+00 -5.59526324e-01 -1.92673683e-01 1.35941434e+00
1.37020409e+00 2.22777277e-01 -1.47010654e-01 -3.27176064e-01
2.28679553e-01 3.91910970e-02 -4.33447719e-01 9.36179608e-03
-6.78826943e-02 4.45973277e-01 6.56777680e-01 6.76360369e-01
-1.21481955e+00 1.01184762e+00 6.11560297e+00 8.61944616e-01
-9.87076640e-01 6.37846649e-01 6.51499748e-01 -2.44589169e-02
-3.57395321e-01 -4.19859111e-01 -6.97248995e-01 -1.09886713e-02
8.91067803e-01 5.19007184e-02 -2.86172837e-01 4.68049377e-01
4.56549108e-01 -1.18824333e-01 -1.30222440e+00 1.10150588e+00
6.58943832e-01 -1.90906858e+00 2.96859264e-01 -7.92092755e-02
4.79175895e-01 5.20244300e-01 3.99909765e-02 7.98198879e-02
1.94362566e-01 -1.22702885e+00 4.58407313e-01 6.70094311e-01
8.68939757e-01 -2.38978229e-02 8.73173833e-01 3.98053616e-01
-7.05446959e-01 1.84396759e-01 2.91889943e-02 4.82627422e-01
3.17173094e-01 4.92454320e-01 -1.24503636e+00 7.77050436e-01
6.23048007e-01 9.52023149e-01 -5.45590520e-01 5.92740655e-01
8.65958109e-02 6.15643978e-01 -1.00828327e-01 2.22787306e-01
4.87222344e-01 4.69069004e-01 5.96550703e-01 1.91348481e+00
-2.70775229e-01 1.41973913e-01 5.74106038e-01 6.63166642e-01
-3.82899749e-03 4.07737941e-01 -4.21866417e-01 1.83176547e-02
-2.40692317e-01 1.05830288e+00 -9.55942690e-01 -6.39081061e-01
-7.09249556e-01 6.36410594e-01 1.10937528e-01 1.53235003e-01
-6.82929635e-01 5.22467196e-01 1.73047453e-01 1.11113779e-01
7.19401985e-02 -2.24489212e-01 -6.81159019e-01 -1.05263519e+00
-2.78884590e-01 -8.18001091e-01 8.08197439e-01 -7.33792841e-01
-1.61060190e+00 3.76749694e-01 2.67838299e-01 -8.26910019e-01
-2.24015638e-01 -8.89000654e-01 1.86942086e-01 6.66486263e-01
-2.07365417e+00 -1.89026177e+00 -1.66274324e-01 5.63209295e-01
1.01695335e+00 -3.04791242e-01 1.02372873e+00 1.62617177e-01
-3.24078538e-02 3.62197131e-01 -2.64537185e-01 7.60865137e-02
9.91221547e-01 -1.54808009e+00 9.84098911e-02 3.15684468e-01
2.99161613e-01 5.21804154e-01 7.41813600e-01 -6.44339144e-01
-1.22730196e+00 -1.11272120e+00 9.52652454e-01 -9.03576314e-01
1.03223646e+00 5.54254316e-02 -8.20890248e-01 8.45357180e-01
1.41956717e-01 2.00881213e-01 1.17978382e+00 -4.58423607e-02
-1.94744557e-01 3.56845856e-01 -1.02431118e+00 4.73007500e-01
1.08867252e+00 -7.72110879e-01 -1.24631250e+00 8.95877659e-01
1.05105007e+00 -6.15335584e-01 -9.72088039e-01 7.42388666e-01
3.33627820e-01 -3.51847738e-01 1.34758306e+00 -9.60980952e-01
5.60664058e-01 -3.63408029e-02 -2.15812042e-01 -8.34377646e-01
-2.18864724e-01 -1.24797963e-01 1.52340144e-01 4.37477529e-01
4.76323813e-01 -5.12021184e-01 7.42710710e-01 1.31660029e-01
-3.84658754e-01 -9.40970063e-01 -9.35049057e-01 -3.64939809e-01
3.13514918e-01 -9.73780215e-01 -3.91506612e-01 1.07445157e+00
-1.71289921e-01 5.02798438e-01 8.86083916e-02 1.51945978e-01
8.55241954e-01 -2.33809486e-01 2.61783421e-01 -1.04309821e+00
-3.79955173e-01 -6.12242281e-01 -5.01848936e-01 -8.11499178e-01
4.82895643e-01 -1.47720349e+00 -8.85591507e-02 -2.19077110e+00
6.57527387e-01 -1.87744632e-01 -3.11706036e-01 7.29749680e-01
-3.04015547e-01 5.56548536e-01 3.73592228e-02 2.28190184e-01
-8.00588965e-01 4.08370309e-02 1.43495023e+00 -4.88298535e-01
2.68468648e-01 -4.28478718e-01 -8.73753607e-01 1.05325472e+00
3.03851426e-01 -4.41822261e-01 -3.87256145e-01 -5.45727253e-01
1.87091827e-01 -2.31803600e-02 7.40763903e-01 -6.04975522e-01
4.22842741e-01 5.04095703e-02 3.38045917e-02 -5.15286326e-01
1.56299263e-01 -7.43171871e-01 -3.66086334e-01 7.40040421e-01
-7.08753109e-01 -3.66875708e-01 3.87335718e-01 8.95884991e-01
-2.12904632e-01 -1.28486037e-01 6.23190284e-01 -6.47046089e-01
-8.51800203e-01 1.03254378e-01 -2.36092657e-01 5.15963137e-01
8.67930651e-01 -1.78111762e-01 -2.39873827e-01 -5.86850047e-02
-1.51365304e+00 5.62177062e-01 7.32395500e-02 3.93195570e-01
5.82306266e-01 -7.50874639e-01 -1.32507002e+00 -2.07223564e-01
4.87356335e-01 3.01915854e-01 4.28920627e-01 1.33038056e+00
-4.69425440e-01 6.79164827e-01 2.75080085e-01 -1.37694550e+00
-1.53127432e+00 2.49624714e-01 3.66618186e-01 -6.87035084e-01
-7.32728422e-01 8.86850595e-01 4.63786811e-01 -4.48323011e-01
1.42555207e-01 -7.85561025e-01 -2.58446634e-01 9.20553431e-02
3.02654773e-01 -2.58553296e-01 1.99989378e-01 -6.36625469e-01
-5.09285152e-01 7.75795639e-01 -3.38636637e-01 -1.58494651e-01
1.39497411e+00 -1.33869976e-01 -2.80652732e-01 6.06421590e-01
1.05333972e+00 -3.34471732e-01 -5.85180879e-01 -6.66419029e-01
4.60928917e-01 2.36019656e-01 2.48794809e-01 -1.02795219e+00
-5.53533077e-01 8.46514523e-01 4.91947621e-01 -1.15697488e-01
8.26888561e-01 7.08240211e-01 6.96582019e-01 5.16368985e-01
1.56522438e-01 -1.03281617e+00 3.71983469e-01 4.46076632e-01
7.74102926e-01 -1.78564453e+00 1.95211545e-01 -5.40411174e-01
-8.80967796e-01 1.04018581e+00 6.16433136e-02 3.05627316e-01
7.75867403e-01 4.09982651e-01 3.49345446e-01 -7.25429893e-01
-6.60557151e-01 -4.28518146e-01 6.03358448e-01 5.52182496e-01
6.21689439e-01 1.82317287e-01 -2.18529612e-01 5.41348219e-01
-1.66189834e-01 2.49361724e-01 2.19859540e-01 9.36242104e-01
-3.76991272e-01 -1.05780876e+00 -2.78266370e-01 5.25666118e-01
-8.54683816e-01 -5.21054685e-01 1.47614582e-02 7.92152464e-01
1.03911132e-01 8.18704844e-01 -2.44760178e-02 2.92395264e-01
2.24323407e-01 1.11104645e-01 6.95346117e-01 -1.20820498e+00
-3.90384674e-01 3.12284291e-01 3.20304722e-01 -4.55118775e-01
-1.09616256e+00 -5.98477662e-01 -1.79995620e+00 5.12093782e-01
1.36313299e-02 -2.73611844e-01 7.65838623e-01 1.41701806e+00
2.12216303e-01 1.07094085e+00 -2.38273144e-01 -2.87173122e-01
-4.60818768e-01 -7.82591701e-01 -7.96474218e-02 5.24300754e-01
5.24157822e-01 -2.63060033e-01 7.10112378e-02 6.02493525e-01] | [14.91025447845459, -1.7746554613113403] |
ec6e7500-33c5-4749-96c1-2588f19d8149 | meta-auxiliary-learning-for-low-resource | 2206.12774 | null | https://arxiv.org/abs/2206.12774v1 | https://arxiv.org/pdf/2206.12774v1.pdf | Meta Auxiliary Learning for Low-resource Spoken Language Understanding | Spoken language understanding (SLU) treats automatic speech recognition (ASR) and natural language understanding (NLU) as a unified task and usually suffers from data scarcity. We exploit an ASR and NLU joint training method based on meta auxiliary learning to improve the performance of low-resource SLU task by only taking advantage of abundant manual transcriptions of speech data. One obvious advantage of such method is that it provides a flexible framework to implement a low-resource SLU training task without requiring access to any further semantic annotations. In particular, a NLU model is taken as label generation network to predict intent and slot tags from texts; a multi-task network trains ASR task and SLU task synchronously from speech; and the predictions of label generation network are delivered to the multi-task network as semantic targets. The efficiency of the proposed algorithm is demonstrated with experiments on the public CATSLU dataset, which produces more suitable ASR hypotheses for the downstream NLU task. | ['Shilei Zhang', 'Chao Deng', 'Junlan Feng', 'Yingying Gao'] | 2022-06-26 | null | null | null | null | ['auxiliary-learning'] | ['methodology'] | [ 7.45644748e-01 5.72347879e-01 -4.34585452e-01 -6.56740129e-01
-1.22395051e+00 -5.18132091e-01 6.04432046e-01 -2.21310303e-01
-4.58278805e-01 7.68383622e-01 4.73974943e-01 -5.95197022e-01
4.63614792e-01 -5.38778067e-01 -6.62699699e-01 -4.48269933e-01
4.64039147e-01 6.60874367e-01 5.16343266e-02 -2.01468691e-01
-1.88473146e-02 -2.76946425e-01 -1.38069737e+00 4.19702083e-01
1.03531039e+00 1.20319664e+00 8.81125391e-01 4.05462086e-01
-5.00090182e-01 8.91850114e-01 -4.48527753e-01 1.68963149e-01
-1.69741616e-01 -5.65819621e-01 -1.16826141e+00 2.64354080e-01
-2.22150132e-01 -5.94026782e-02 -3.83936465e-02 7.93851852e-01
4.52487260e-01 5.34328938e-01 6.58988595e-01 -9.34200227e-01
-5.42356074e-01 8.59478712e-01 -7.11682215e-02 1.04026392e-01
4.84003901e-01 -2.09182084e-01 1.25321186e+00 -9.58106816e-01
5.54906070e-01 1.35395372e+00 8.04464743e-02 1.07593584e+00
-1.00619161e+00 -6.11360967e-01 3.47446740e-01 1.29794195e-01
-1.37944865e+00 -9.09840107e-01 6.59779429e-01 -3.00093796e-02
1.27298725e+00 7.94280544e-02 -2.60115683e-01 1.28376460e+00
-4.79974151e-01 1.37369037e+00 8.86056066e-01 -6.20274603e-01
3.08687776e-01 1.92536145e-01 4.68089551e-01 6.54089391e-01
-6.29304528e-01 -2.46349752e-01 -5.01653969e-01 1.29162883e-02
3.45810860e-01 -1.49441838e-01 -4.72946256e-01 2.00053126e-01
-1.35692000e+00 8.53244901e-01 2.05881059e-01 4.15836364e-01
-2.25190938e-01 -2.12768704e-01 5.46334267e-01 3.50127101e-01
9.19964850e-01 2.09280565e-01 -8.34123313e-01 -1.82103083e-01
-5.41529655e-01 -5.41512728e-01 6.58909261e-01 1.20242178e+00
1.04853857e+00 1.81675762e-01 -4.49995138e-02 1.54730844e+00
3.89619142e-01 5.54845572e-01 9.87970352e-01 -5.41323543e-01
8.01711440e-01 5.25568485e-01 -9.76800472e-02 -1.56933010e-01
-2.85612404e-01 -2.42441505e-01 -6.17262542e-01 -4.77080673e-01
-1.76292762e-01 -3.06474000e-01 -1.12464786e+00 1.88044941e+00
1.33615583e-01 5.92303872e-01 8.03595960e-01 7.85420120e-01
1.01340353e+00 1.32704747e+00 3.52554142e-01 -3.89165640e-01
1.35240388e+00 -1.41231370e+00 -9.24720705e-01 -6.91643775e-01
1.27478600e+00 -5.61845005e-01 1.14237106e+00 -2.31619682e-02
-6.89779878e-01 -6.65176868e-01 -8.45161796e-01 -1.91294968e-01
-3.29887033e-01 5.83052397e-01 4.58557218e-01 2.63895631e-01
-1.07714891e+00 4.61022928e-02 -5.53994060e-01 -5.06993711e-01
1.52591676e-01 1.75779730e-01 -3.15744996e-01 -1.93312362e-01
-1.50076449e+00 8.84665191e-01 8.73130977e-01 1.72512665e-01
-9.06091571e-01 -3.35068256e-01 -1.23427415e+00 1.74732298e-01
8.33821535e-01 -2.99435914e-01 1.61542702e+00 -1.08405423e+00
-1.88406777e+00 8.30102682e-01 -5.76445937e-01 -4.20940429e-01
9.75941867e-02 -4.78073768e-02 -3.21925044e-01 -1.22022189e-01
1.01640604e-01 6.62635207e-01 5.45742989e-01 -9.72875297e-01
-8.74255776e-01 -3.50042373e-01 -3.53869349e-01 5.83432317e-01
-1.93146080e-01 -8.21761787e-03 -1.04808509e-01 -5.04828811e-01
3.01629573e-01 -6.18463933e-01 -7.87414014e-02 -9.23116744e-01
-2.10657865e-01 -9.74626303e-01 8.87594998e-01 -8.13924015e-01
1.28121805e+00 -1.94046605e+00 2.79301733e-01 -1.74788222e-01
-3.23109031e-01 5.03463805e-01 -4.93186116e-01 3.90948504e-01
-7.25613311e-02 -4.98810336e-02 -2.51908571e-01 -6.79340005e-01
-9.50097367e-02 4.29247379e-01 -5.36127746e-01 -5.67610711e-02
1.65172338e-01 9.37284291e-01 -9.42972422e-01 -3.34041864e-01
3.30543399e-01 -4.41655479e-02 -1.07917346e-01 7.35643625e-01
-6.32636130e-01 5.34924567e-01 -7.38222063e-01 6.55126870e-01
1.86494574e-01 -1.92108497e-01 2.13047266e-01 8.32093358e-02
-1.38046637e-01 9.77210283e-01 -8.31946254e-01 1.94825304e+00
-1.07805383e+00 -3.50974426e-02 1.39402926e-01 -1.46724463e+00
1.17400694e+00 9.12601650e-01 1.29705835e-02 -5.14692962e-01
1.61226526e-01 5.31944454e-01 -1.60362825e-01 -3.68625820e-01
7.65294731e-02 -2.33806208e-01 -3.38728011e-01 7.84719467e-01
5.61616302e-01 3.16319354e-02 -2.83474803e-01 4.33161668e-02
1.10225403e+00 1.51053190e-01 7.02356994e-01 -1.97122976e-01
9.10405934e-01 -1.04028143e-01 4.56600338e-01 6.81871831e-01
-8.77498537e-02 2.81181633e-01 -1.28346667e-01 -2.30057269e-01
-7.74841785e-01 -8.92546475e-01 9.68493968e-02 1.47319448e+00
-9.27671939e-02 -2.45722651e-01 -7.07720459e-01 -1.03122711e+00
-3.10589105e-01 1.25062919e+00 -2.83997118e-01 3.77408639e-02
-4.48620260e-01 -4.58562791e-01 3.13269287e-01 3.89719307e-01
3.12468916e-01 -1.49856198e+00 2.68706381e-01 4.71573353e-01
-5.11756778e-01 -1.53916633e+00 -5.23688197e-01 4.21431333e-01
-6.59707427e-01 -5.84233165e-01 -6.90258324e-01 -1.16017151e+00
6.81850493e-01 4.94020760e-01 7.88661122e-01 -1.24434695e-01
1.56849891e-01 1.15691386e-01 -8.11243355e-01 -1.33368284e-01
-1.08151782e+00 3.13239366e-01 1.84651494e-01 1.92921728e-01
5.09725928e-01 -3.58164251e-01 -1.21539654e-02 5.44524550e-01
-7.62324810e-01 5.62222958e-01 5.31888545e-01 1.05542743e+00
4.62272882e-01 -4.18527305e-01 1.38812351e+00 -9.15101349e-01
3.48980099e-01 -5.51483512e-01 -4.13754553e-01 6.22131526e-01
-3.25688630e-01 1.67637542e-01 7.71994710e-01 -2.77063370e-01
-1.60672295e+00 2.32773528e-01 -3.93516004e-01 -2.17402294e-01
-4.78529960e-01 5.15053809e-01 -6.08147919e-01 4.69841421e-01
2.45708451e-01 6.85703635e-01 -4.07491028e-02 -7.23347664e-01
6.13269329e-01 1.52199483e+00 2.85394907e-01 -5.62876463e-01
3.28113377e-01 -5.70968576e-02 -7.55056143e-01 -1.21213281e+00
-1.46332932e+00 -8.95690918e-01 -7.60888994e-01 -2.30522398e-02
1.01086152e+00 -1.10692048e+00 -4.04400259e-01 3.37061107e-01
-1.39591157e+00 -4.68852997e-01 2.26882666e-01 4.98518735e-01
-6.57085896e-01 4.16809559e-01 -5.12888193e-01 -1.00818145e+00
-4.73565459e-01 -1.14175463e+00 1.24430442e+00 1.00623131e-01
-1.51286066e-01 -1.02525294e+00 -2.68584371e-01 8.86414170e-01
1.24682844e-01 -7.59021938e-01 9.21270132e-01 -1.33645260e+00
-4.52379167e-01 4.59570214e-02 -2.92769581e-01 6.58634365e-01
4.71738875e-01 -8.38914454e-01 -1.21062517e+00 -1.53022066e-01
1.85674371e-03 -9.62958515e-01 8.21173787e-01 1.35450214e-01
9.49874580e-01 -4.02481973e-01 -4.19549912e-01 1.89292520e-01
9.61971581e-01 4.58585650e-01 1.45206720e-01 -1.46706536e-01
6.92715645e-01 8.03760052e-01 7.69736528e-01 1.03159733e-01
3.22713107e-01 6.21757329e-01 -3.85494381e-02 3.94949131e-02
-7.29957502e-03 -3.74916166e-01 5.73012948e-01 1.42143154e+00
4.80554074e-01 -4.59024519e-01 -9.55386221e-01 5.22245705e-01
-2.09018922e+00 -3.95494431e-01 3.90696138e-01 2.10881925e+00
9.31403816e-01 -1.28036857e-01 -4.11555976e-01 -3.98059666e-01
9.43213642e-01 2.96219051e-01 -3.01397353e-01 -1.42558560e-01
6.83859065e-02 -1.47156930e-02 1.08520858e-01 7.38897920e-01
-1.06681526e+00 1.67820573e+00 4.87512445e+00 1.24604940e+00
-8.57542276e-01 5.13522208e-01 7.73837328e-01 4.45203483e-01
-2.55386233e-01 -5.76297976e-02 -1.10057771e+00 3.11712831e-01
1.11798668e+00 -6.80101737e-02 3.35603029e-01 9.31584239e-01
4.08697754e-01 -3.17645147e-02 -1.20316339e+00 8.58022809e-01
2.05453962e-01 -1.10544038e+00 2.90681392e-01 -3.84497911e-01
4.96365011e-01 2.72111088e-01 -4.36097205e-01 5.70325613e-01
4.68362302e-01 -9.14887547e-01 3.51604104e-01 -1.87249538e-02
8.73622298e-01 -6.33851051e-01 9.37125385e-01 8.81216943e-01
-1.07585716e+00 1.21329032e-01 -5.37267923e-01 -8.21298659e-02
4.93211567e-01 1.23170644e-01 -1.40513146e+00 6.04822695e-01
3.77822965e-02 5.90399086e-01 -4.45319340e-02 2.98446834e-01
-5.35590827e-01 9.12951827e-01 -3.43699604e-01 -2.83623397e-01
6.34162247e-01 -3.78672600e-01 4.15341228e-01 1.22410107e+00
4.29064393e-01 3.25422645e-01 6.46303952e-01 5.70601463e-01
-3.93296838e-01 5.36901593e-01 -8.42133343e-01 -3.13428462e-01
7.45822787e-01 1.12463760e+00 -5.83192229e-01 -5.01594067e-01
-6.63722515e-01 1.24562418e+00 5.76690316e-01 3.53053927e-01
-3.21677685e-01 1.37874642e-02 5.33541441e-01 -5.09222925e-01
3.84695493e-02 -8.56602564e-02 8.97923671e-03 -1.50559294e+00
-2.45932028e-01 -6.57859206e-01 4.05627400e-01 -8.65211010e-01
-1.15937769e+00 7.77506828e-01 -3.99627924e-01 -1.03079355e+00
-7.65773296e-01 -4.00853097e-01 -6.19691312e-01 1.06037784e+00
-1.72057211e+00 -1.43486810e+00 2.97490239e-01 3.19323063e-01
1.42455530e+00 -4.97672230e-01 1.12035441e+00 5.63756227e-02
-7.28498518e-01 3.30463678e-01 4.57707755e-02 2.09154502e-01
5.62366545e-01 -1.10548902e+00 4.52117920e-01 7.95730412e-01
2.02178806e-01 2.53931224e-01 4.53057379e-01 -6.47084177e-01
-9.73842144e-01 -1.27710998e+00 1.28348982e+00 -2.50831127e-01
8.98972392e-01 -6.86370373e-01 -9.31624413e-01 9.83592331e-01
1.07791625e-01 -2.34044626e-01 8.14428747e-01 2.22090602e-01
7.86296800e-02 1.67176604e-01 -6.05486572e-01 4.68684196e-01
1.01388550e+00 -7.26090014e-01 -1.05861855e+00 4.79342312e-01
1.25441945e+00 -1.60841003e-01 -5.82954705e-01 3.00266922e-01
-8.97426978e-02 -1.92672670e-01 6.36170983e-01 -7.30618417e-01
1.96366221e-01 -1.86932653e-01 -3.56167644e-01 -1.53612685e+00
1.85285851e-01 -6.44065082e-01 2.89467752e-01 1.56146753e+00
8.58401120e-01 -6.70009911e-01 4.12809968e-01 2.63162106e-01
-6.71673715e-01 -4.05107081e-01 -1.22490489e+00 -5.08217871e-01
-5.12242913e-01 -6.45387292e-01 2.16551766e-01 9.96125519e-01
2.49535352e-01 1.13184786e+00 -4.00739968e-01 2.21308202e-01
4.83023942e-01 8.77826214e-02 4.64209735e-01 -1.12921941e+00
-6.24605194e-02 7.24040493e-02 1.00255467e-01 -1.49076569e+00
1.08232093e+00 -1.23059821e+00 5.19210875e-01 -1.46447051e+00
-8.08860883e-02 -4.97295111e-01 -3.50460708e-01 8.17368627e-01
-2.92538106e-01 -2.34734580e-01 -6.66734949e-03 2.76460648e-01
-7.39994466e-01 1.12937987e+00 1.05485690e+00 -5.78062935e-03
-3.02936822e-01 4.00168359e-01 -4.13339764e-01 7.45266318e-01
4.74708259e-01 -3.48607004e-01 -6.97567582e-01 -3.53802294e-01
-4.64991629e-01 5.85170448e-01 -3.13846171e-01 -6.08475685e-01
6.45691752e-02 8.32339004e-02 -2.21512049e-01 -5.28834999e-01
3.29266638e-01 -6.77407920e-01 -5.85867345e-01 1.42387256e-01
-6.44870400e-01 -5.59467316e-01 -1.30331159e-01 4.71445799e-01
-4.27350968e-01 -6.32437170e-01 7.37144828e-01 -2.67983496e-01
-1.26757956e+00 2.12261513e-01 -5.06982207e-01 -7.87632838e-02
8.24684322e-01 7.24795982e-02 -2.50046581e-01 -5.82964063e-01
-9.40696776e-01 6.07354879e-01 -7.30550364e-02 7.94055521e-01
5.71758568e-01 -1.06159091e+00 -5.23770094e-01 2.79576898e-01
4.17467088e-01 1.78205624e-01 4.31466758e-01 3.71264070e-01
2.05231175e-01 8.63063455e-01 3.34622830e-01 -4.14023668e-01
-1.12809849e+00 4.44687992e-01 1.49272501e-01 -2.11346090e-01
-5.23274302e-01 9.82353806e-01 6.47540927e-01 -1.05191422e+00
2.68960446e-01 -1.56081825e-01 -4.79217231e-01 1.01089180e-01
6.77222371e-01 -5.64287268e-02 6.24988340e-02 -7.65908241e-01
-1.66140154e-01 9.44225788e-02 -7.32702836e-02 -2.27288797e-01
1.24806011e+00 -8.24442685e-01 7.35983253e-02 6.93594396e-01
1.29326868e+00 -4.96645719e-01 -1.09837806e+00 -7.86527455e-01
2.97012568e-01 3.13321382e-01 1.93878591e-01 -8.17051649e-01
-5.10821462e-01 1.01325786e+00 9.97187197e-02 2.60461010e-02
7.37813950e-01 4.44812417e-01 1.19282651e+00 8.09218884e-01
4.98720795e-01 -1.26150882e+00 -1.91014260e-02 8.87247384e-01
6.83650792e-01 -1.39666963e+00 -7.42515802e-01 -6.21345460e-01
-8.45807493e-01 8.93263757e-01 8.70176554e-01 3.23405832e-01
5.07920563e-01 -1.11441389e-01 5.15692294e-01 2.62872726e-01
-9.07860935e-01 -5.25694191e-01 1.99116945e-01 4.94568437e-01
3.38012934e-01 5.60837016e-02 -3.44354004e-01 8.97008002e-01
5.28261298e-03 -1.08155131e-01 3.64371479e-01 5.67945063e-01
-8.48274767e-01 -1.15939450e+00 -2.48446036e-02 2.93073088e-01
-2.85352886e-01 -3.79529983e-01 -6.00869246e-02 2.98144102e-01
-2.73646235e-01 1.06195998e+00 -1.69147719e-02 -2.14275315e-01
3.31710838e-02 7.49660969e-01 -1.84592500e-01 -1.27970505e+00
-1.30266204e-01 4.10533547e-01 5.02648234e-01 -5.17561793e-01
-3.97029102e-01 -4.03176367e-01 -1.36922753e+00 7.31305897e-01
-4.19780761e-01 6.49764061e-01 6.49990380e-01 1.57344413e+00
2.13576466e-01 4.53992814e-01 9.26423788e-01 -6.63629413e-01
-6.58436477e-01 -1.41652274e+00 -5.98567367e-01 2.62631088e-01
1.22218370e-01 -6.23068750e-01 -4.68408078e-01 6.73395246e-02] | [13.798457145690918, 7.073267936706543] |
4017651d-8178-4b65-806c-deaf68b41107 | sagemix-saliency-guided-mixup-for-point | 2210.06944 | null | https://arxiv.org/abs/2210.06944v1 | https://arxiv.org/pdf/2210.06944v1.pdf | SageMix: Saliency-Guided Mixup for Point Clouds | Data augmentation is key to improving the generalization ability of deep learning models. Mixup is a simple and widely-used data augmentation technique that has proven effective in alleviating the problems of overfitting and data scarcity. Also, recent studies of saliency-aware Mixup in the image domain show that preserving discriminative parts is beneficial to improving the generalization performance. However, these Mixup-based data augmentations are underexplored in 3D vision, especially in point clouds. In this paper, we propose SageMix, a saliency-guided Mixup for point clouds to preserve salient local structures. Specifically, we extract salient regions from two point clouds and smoothly combine them into one continuous shape. With a simple sequential sampling by re-weighted saliency scores, SageMix preserves the local structure of salient regions. Extensive experiments demonstrate that the proposed method consistently outperforms existing Mixup methods in various benchmark point cloud datasets. With PointNet++, our method achieves an accuracy gain of 2.6% and 4.0% over standard training in 3D Warehouse dataset (MN40) and ScanObjectNN, respectively. In addition to generalization performance, SageMix improves robustness and uncertainty calibration. Moreover, when adopting our method to various tasks including part segmentation and standard 2D image classification, our method achieves competitive performance. | ['Hyunwoo J. Kim', 'Yunyang Xiong', 'Injae Kim', 'Minkyu Jeon', 'Sanghyeok Lee'] | 2022-10-13 | null | null | null | null | ['3d-point-cloud-classification', '3d-part-segmentation'] | ['computer-vision', 'computer-vision'] | [-7.94013217e-02 4.91438732e-02 -3.24297279e-01 -3.08826834e-01
-6.68945789e-01 -2.36221641e-01 3.60627800e-01 3.04433674e-01
-8.47182944e-02 2.35880449e-01 -1.12975612e-01 -3.25264931e-02
-6.51901681e-03 -6.59682453e-01 -1.10775530e+00 -6.28228724e-01
3.03759843e-01 3.96338373e-01 4.49160546e-01 -8.16467181e-02
2.01065660e-01 8.31954181e-01 -1.62940228e+00 -1.85356781e-01
1.33264506e+00 1.28160489e+00 4.84221578e-01 -7.35242218e-02
-4.66613173e-01 2.82308429e-01 -2.85932630e-01 -2.56928295e-01
5.58421850e-01 2.89801836e-01 -5.44490397e-01 2.21581385e-01
6.66797459e-01 -3.22910577e-01 -3.39988232e-01 1.18295419e+00
3.40969235e-01 7.54671842e-02 3.82625878e-01 -1.57667255e+00
-9.44865286e-01 4.33901787e-01 -1.10722709e+00 2.04928905e-01
-4.06045109e-01 3.32078561e-02 8.37960780e-01 -1.18198538e+00
1.96123615e-01 1.31381738e+00 8.15620363e-01 2.56150872e-01
-1.21821082e+00 -8.78770351e-01 3.39113563e-01 7.60291368e-02
-1.33191013e+00 -2.61083096e-02 1.19527817e+00 -2.49758855e-01
6.75676763e-01 1.88522726e-01 6.78133011e-01 5.18103063e-01
-1.42487884e-01 1.38934743e+00 8.37228835e-01 -5.40802516e-02
1.70710593e-01 3.04836519e-02 2.07688764e-01 4.92617220e-01
3.96476984e-01 -2.50697173e-02 -2.50384927e-01 -5.01470966e-03
1.15449965e+00 6.26761913e-01 -1.47733137e-01 -9.01426315e-01
-1.29323184e+00 5.69486618e-01 1.24883449e+00 7.39564896e-02
-5.47652662e-01 1.98259056e-01 2.52151579e-01 -2.14808181e-01
6.74909115e-01 3.98018479e-01 -4.76152092e-01 3.98071617e-01
-9.76970971e-01 3.80893320e-01 -4.75016236e-02 1.38146222e+00
8.75934780e-01 1.59782484e-01 -1.11456171e-01 1.05328441e+00
2.75789887e-01 7.59278834e-01 2.94791698e-01 -8.88265431e-01
3.82503033e-01 1.10840631e+00 -1.12222759e-02 -1.02196527e+00
-4.61090088e-01 -8.06319416e-01 -9.97184634e-01 2.09761769e-01
-2.83280648e-02 2.82567263e-01 -1.27686346e+00 1.56480610e+00
5.22909284e-01 4.34556335e-01 -1.43910527e-01 1.07165027e+00
9.44850147e-01 5.44170797e-01 1.31822258e-01 1.91534162e-01
1.03525329e+00 -1.11207199e+00 -4.15985972e-01 -4.24807429e-01
1.32100701e-01 -6.36844933e-01 1.20309794e+00 -5.07026352e-03
-1.09813869e+00 -7.04406321e-01 -1.09369576e+00 -2.37478942e-01
-2.23383069e-01 9.82453302e-02 8.48491490e-01 9.97444838e-02
-6.05728924e-01 4.91108954e-01 -8.45016003e-01 -1.05889872e-01
1.06679225e+00 4.23137367e-01 -3.01987231e-01 -2.73956209e-01
-6.78071439e-01 6.39160454e-01 4.48408931e-01 -2.18135759e-01
-7.59362578e-01 -1.13915944e+00 -9.09566820e-01 1.32144511e-01
4.62885857e-01 -6.41245902e-01 1.32820010e+00 -5.97264946e-01
-9.50246930e-01 7.54521191e-01 -8.67671669e-02 -7.12174773e-01
3.45339030e-01 -5.91734648e-01 -6.55393256e-03 1.39670014e-01
4.20702338e-01 1.27398860e+00 8.65470052e-01 -1.51205325e+00
-7.33758688e-01 -5.06605208e-01 -1.27114549e-01 3.09901148e-01
-2.15889215e-01 -3.84850413e-01 -8.91619265e-01 -9.36145723e-01
8.16548467e-01 -8.00873637e-01 -3.24664056e-01 3.24887633e-01
-4.87629116e-01 -2.78008193e-01 9.88613188e-01 -3.69411349e-01
6.47570312e-01 -2.29719090e+00 -5.15617654e-02 -9.48261842e-03
3.79422396e-01 2.80591756e-01 -5.62550314e-02 -2.37026468e-01
-3.72609720e-02 1.54329926e-01 -5.66297650e-01 -4.70395774e-01
9.21194926e-02 1.88682497e-01 -4.19466406e-01 3.39673907e-01
4.79120284e-01 1.23282242e+00 -7.03032255e-01 -5.61578810e-01
6.25990987e-01 3.92523080e-01 -4.75402057e-01 -7.78000429e-02
-3.03265721e-01 1.49514571e-01 -5.03011644e-01 1.12685835e+00
1.24676836e+00 -3.63129705e-01 -6.27173901e-01 -1.17140591e-01
1.52507558e-01 8.25759694e-02 -8.99300218e-01 1.86274862e+00
-2.37686068e-01 5.00966966e-01 -1.32683039e-01 -7.86655128e-01
1.27177882e+00 -2.72100300e-01 5.66900790e-01 -6.08136714e-01
9.17469859e-02 2.18321294e-01 -3.27110231e-01 -8.94765034e-02
6.40705347e-01 -6.08285964e-02 7.92099312e-02 6.23932406e-02
1.05302297e-01 -4.40878630e-01 -2.45472386e-01 1.32242054e-01
5.36895990e-01 1.12111703e-01 3.38990726e-02 -2.94943571e-01
1.46674410e-01 2.59311855e-01 7.84147739e-01 6.06783509e-01
-3.24564993e-01 1.10490620e+00 1.45440459e-01 -2.53883719e-01
-1.11981857e+00 -1.16540134e+00 -3.30572695e-01 7.11275220e-01
6.89845622e-01 6.49343878e-02 -5.50377667e-01 -6.89957857e-01
6.05736196e-01 6.66792810e-01 -5.49819946e-01 -3.47428352e-01
-4.39680725e-01 -6.02513134e-01 2.15397298e-01 9.17935193e-01
1.06038070e+00 -7.89325655e-01 -2.81802475e-01 3.39765251e-02
-4.18298282e-02 -1.15384221e+00 -4.18131202e-01 1.29001766e-01
-1.39461768e+00 -9.87756133e-01 -9.48449552e-01 -9.85273182e-01
8.55262339e-01 1.08711004e+00 1.07319880e+00 6.95655346e-02
4.08164784e-02 1.44023255e-01 -1.96463719e-01 -7.83056617e-01
1.66684240e-01 1.26956046e-01 -5.21881729e-02 -3.15841079e-01
6.65243089e-01 -6.05077267e-01 -5.05897522e-01 3.82411897e-01
-9.11333978e-01 2.55349308e-01 9.05596793e-01 6.40087903e-01
9.77825284e-01 -1.46199569e-01 5.26360512e-01 -4.25891101e-01
2.56854177e-01 -4.83232647e-01 -5.86997867e-01 -2.84212958e-02
-6.52982533e-01 -8.65218192e-02 1.12799123e-01 -3.35255891e-01
-8.10413837e-01 1.80129603e-01 2.11170465e-02 -1.22759748e+00
-1.57581687e-01 2.42396861e-01 -3.18864197e-01 -3.07889909e-01
3.79389107e-01 3.59165072e-01 1.10237606e-01 -7.07636118e-01
3.96940202e-01 4.12585855e-01 8.39003563e-01 -4.42423552e-01
9.97838080e-01 5.50127864e-01 -1.25069532e-03 -5.30710280e-01
-9.36572373e-01 -6.12803996e-01 -6.11942708e-01 7.20785260e-02
5.77485561e-01 -1.20734036e+00 -4.09263968e-01 4.87683862e-01
-9.40233052e-01 8.95498842e-02 -4.72061932e-01 3.27801853e-01
-3.98245394e-01 3.40047389e-01 -3.54544193e-01 -5.92521191e-01
-4.48555380e-01 -1.14451480e+00 1.32488847e+00 5.67965090e-01
3.96381080e-01 -6.59055293e-01 -4.30319816e-01 3.26780409e-01
1.48601040e-01 3.33096027e-01 6.83499515e-01 -6.21868551e-01
-8.79998147e-01 -1.91643208e-01 -5.76698065e-01 6.04695499e-01
1.61137760e-01 -1.91756502e-01 -9.75452602e-01 -6.02921247e-02
6.30076453e-02 -3.90075743e-02 1.01779902e+00 6.57752573e-01
1.44490433e+00 3.27854566e-02 -5.06619573e-01 6.71130419e-01
1.29270208e+00 2.79724561e-02 3.26602638e-01 4.82608080e-01
8.84371579e-01 3.64371538e-01 9.14813697e-01 1.59320131e-01
3.92729282e-01 3.95908833e-01 9.37949896e-01 -3.84879500e-01
-5.45713678e-02 -4.72231030e-01 -2.27288157e-01 3.90088797e-01
2.66613275e-01 3.68819892e-01 -1.08629692e+00 9.33553159e-01
-2.00050235e+00 -4.77412820e-01 -1.92695647e-01 2.02896500e+00
5.77000558e-01 3.55596960e-01 4.32379283e-02 1.16150051e-01
8.78323615e-01 3.64057012e-02 -1.03625977e+00 2.87661463e-01
-3.87676388e-01 -4.99412939e-02 7.52132595e-01 6.60293847e-02
-1.41928494e+00 1.06573296e+00 5.64815760e+00 8.69521141e-01
-1.11894178e+00 5.42384312e-02 6.22435927e-01 -2.83811361e-01
-3.27723324e-01 -2.26957873e-01 -6.79130852e-01 5.03302693e-01
6.05212934e-02 -1.28886029e-01 -2.17464883e-02 1.44881821e+00
-5.47076296e-03 -1.77526008e-02 -8.40388656e-01 1.24902344e+00
3.75396833e-02 -1.60232663e+00 1.22432530e-01 -3.50251654e-03
9.67269778e-01 3.61372679e-01 2.81664550e-01 3.30049664e-01
1.50038227e-01 -7.57314682e-01 8.80854905e-01 3.18616122e-01
2.84585327e-01 -1.01282907e+00 8.64080429e-01 3.37021887e-01
-9.39605594e-01 2.05902872e-03 -7.27277756e-01 2.32787326e-01
1.32817430e-02 9.32987034e-01 -8.51906419e-01 5.11209905e-01
1.03852427e+00 8.88663054e-01 -7.37385690e-01 1.53792274e+00
-1.73529699e-01 3.47886384e-01 -6.10673070e-01 1.73688397e-01
3.53688270e-01 -5.99666834e-02 7.17784703e-01 7.63501823e-01
1.80253953e-01 -1.44147417e-02 1.66948259e-01 1.24401629e+00
-2.88697630e-01 -1.39364362e-01 -5.37271738e-01 1.89144611e-01
7.01799214e-01 1.16572666e+00 -6.46100581e-01 -2.91924268e-01
-1.99955463e-01 6.69927835e-01 9.29001048e-02 2.04404652e-01
-8.04517090e-01 -3.46622407e-01 1.03102148e+00 6.50128489e-03
4.23200905e-01 -3.56425524e-01 -9.73333716e-01 -9.41548586e-01
2.68953890e-01 -4.20927674e-01 -2.77933050e-02 -1.03405392e+00
-1.21195042e+00 3.45684797e-01 1.65919766e-01 -1.46345127e+00
4.12636101e-01 -3.68537784e-01 -5.50331652e-01 8.15406561e-01
-1.77424288e+00 -1.43106067e+00 -6.32097185e-01 3.12164932e-01
7.03962147e-01 -7.80494437e-02 9.41186175e-02 2.40188017e-01
-5.64034283e-01 3.17579538e-01 4.31726640e-03 -9.95570943e-02
3.44612777e-01 -1.28812420e+00 8.16180587e-01 9.49234545e-01
7.33878762e-02 6.44528389e-01 5.40413320e-01 -6.76675677e-01
-1.13411462e+00 -1.36821246e+00 3.51655722e-01 -3.21497113e-01
3.08753610e-01 -3.13563019e-01 -1.33755982e+00 5.49865305e-01
-7.73020685e-02 9.62242410e-02 3.32305253e-01 -9.33839530e-02
-2.49238357e-01 -2.00167403e-01 -1.38856852e+00 5.37370324e-01
1.10325575e+00 -1.87360391e-01 -8.83896530e-01 2.99812734e-01
1.27866030e+00 -6.89058304e-01 -7.49404430e-01 8.43746722e-01
4.89874789e-03 -8.10215890e-01 1.24314034e+00 -4.13446754e-01
3.99650186e-01 -5.46360493e-01 -2.89326370e-01 -1.21834624e+00
-4.55625296e-01 -2.09762722e-01 -2.63992488e-01 1.31131780e+00
1.18506886e-01 -4.84533131e-01 1.01235497e+00 5.97393572e-01
-6.10138535e-01 -8.04864228e-01 -7.84491062e-01 -8.30151558e-01
1.11201197e-01 -5.20181298e-01 1.06211925e+00 8.66670847e-01
-5.33895195e-01 2.02089977e-02 -2.42945589e-02 5.27983248e-01
8.04970980e-01 4.08694386e-01 8.89843524e-01 -1.51649427e+00
4.67920095e-01 -6.07519150e-01 -5.09096920e-01 -1.24540830e+00
3.35294157e-02 -8.38097453e-01 4.74399179e-02 -1.66033578e+00
1.32352695e-01 -7.84030437e-01 -4.78451639e-01 6.02816641e-01
-3.67619723e-01 2.12871134e-01 3.82862538e-01 4.67065871e-01
-4.27045763e-01 9.00035083e-01 1.29955518e+00 -4.31227535e-01
-3.16330135e-01 2.33323462e-02 -1.04403734e+00 7.95454025e-01
1.01281309e+00 -3.35273683e-01 -3.41037512e-01 -6.27056777e-01
-2.36551613e-01 -5.72523892e-01 5.65696716e-01 -1.24183393e+00
1.35536522e-01 -2.13598981e-01 5.12007833e-01 -1.23472738e+00
3.35964620e-01 -9.73186314e-01 -2.10065201e-01 2.13664621e-01
-1.03364028e-02 -1.29506126e-01 6.84618175e-01 6.45500541e-01
-3.07821929e-01 -6.11226074e-02 1.00156748e+00 7.37645244e-03
-9.93130505e-01 6.16005361e-01 4.66398358e-01 -1.43948659e-01
1.18588209e+00 -2.44288877e-01 -2.20003083e-01 9.75055024e-02
-4.39082742e-01 6.06383979e-01 7.18584597e-01 8.29545915e-01
9.05010462e-01 -1.63427758e+00 -4.01479572e-01 2.79481679e-01
2.63714969e-01 9.59612429e-01 3.58225495e-01 7.04255939e-01
-4.53673601e-01 3.92948329e-01 -3.52497011e-01 -1.24903965e+00
-1.06710553e+00 8.05531323e-01 1.83417097e-01 3.58761072e-01
-8.01703095e-01 9.79483187e-01 4.64376062e-01 -5.47927022e-01
3.34799647e-01 -8.87040854e-01 4.89043482e-02 -3.07262719e-01
3.51586521e-01 2.41514757e-01 1.18085496e-01 -4.87973899e-01
-4.36093360e-01 6.67428017e-01 -2.10592687e-01 4.09235477e-01
1.46979094e+00 -6.06920086e-02 -8.44414625e-03 3.75913292e-01
1.05533552e+00 -3.19573402e-01 -1.65900838e+00 -5.68314075e-01
-4.17305641e-02 -6.68502033e-01 3.40576708e-01 -5.52179873e-01
-1.40528595e+00 9.57083404e-01 6.51813209e-01 7.54000172e-02
1.14722037e+00 2.79556990e-01 8.88081193e-01 2.50842512e-01
3.13596219e-01 -8.70207071e-01 2.64640842e-02 2.06171855e-01
1.14573693e+00 -1.69076157e+00 3.98378484e-02 -5.86473465e-01
-7.44432449e-01 5.91389477e-01 9.30418670e-01 -3.94253969e-01
5.90476871e-01 -7.81461895e-02 -1.24955043e-01 -1.72131881e-01
-3.84734064e-01 -3.76141638e-01 4.48924184e-01 6.85761929e-01
-2.48391688e-01 2.73004756e-03 2.06899762e-01 5.35908282e-01
-7.88661093e-02 -1.08401224e-01 1.45578533e-01 1.04182827e+00
-7.02646315e-01 -5.18755615e-01 -5.75662494e-01 5.24895132e-01
-1.19651444e-01 4.45316285e-02 -3.58760715e-01 1.00433338e+00
8.78136978e-02 5.24241984e-01 3.53617668e-01 -4.70311046e-01
4.95055199e-01 -2.06448153e-01 7.30206594e-02 -5.38419425e-01
-2.40614414e-01 1.65953726e-01 -6.09802723e-01 -3.51677537e-01
-3.28528583e-01 -7.55370378e-01 -1.53601873e+00 -2.00735033e-01
-4.27105725e-01 -9.54481249e-04 8.78627419e-01 7.50156581e-01
7.23854184e-01 6.92188025e-01 5.21889031e-01 -1.01757932e+00
-3.63246322e-01 -9.86963332e-01 -4.85118091e-01 3.36538762e-01
5.19043565e-01 -1.11716378e+00 -1.38425231e-01 -2.64187902e-01] | [7.939978122711182, -3.4220130443573] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.