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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
5c4cb815-bf72-437b-93fd-d59d05b4c7e3 | learnable-reconstruction-methods-from-rgb | 2106.15944 | null | https://arxiv.org/abs/2106.15944v2 | https://arxiv.org/pdf/2106.15944v2.pdf | A survey on computational spectral reconstruction methods from RGB to hyperspectral imaging | Hyperspectral imaging enables versatile applications due to its competence in capturing abundant spatial and spectral information, which are crucial for identifying substances. However, the devices for acquiring hyperspectral images are expensive and complicated. Therefore, many alternative spectral imaging methods have been proposed by directly reconstructing the hyperspectral information from lower-cost, more available RGB images. We present a thorough investigation of these state-of-the-art spectral reconstruction methods from the widespread RGB images. A systematic study and comparison of more than 25 methods has revealed that most of the data-driven deep learning methods are superior to prior-based methods in terms of reconstruction accuracy and quality despite lower speeds. This comprehensive review can serve as a fruitful reference source for peer researchers, thus further inspiring future development directions in related domains. | ['Yunfeng Nie', 'Felix Heide', 'Qiang Fu', 'Wenqi Ren', 'Runmu Su', 'Jingang Zhang'] | 2021-06-30 | null | null | null | null | ['spectral-reconstruction'] | ['computer-vision'] | [ 8.28603148e-01 -7.89180696e-01 -1.54297128e-01 -1.78287789e-01
-7.52189696e-01 -3.51710707e-01 1.25164896e-01 -4.05591220e-01
-3.01294774e-01 9.22371626e-01 -3.46766382e-01 -2.37151906e-01
-5.30264318e-01 -8.49554241e-01 -4.27578330e-01 -1.42596149e+00
1.86232522e-01 1.08625449e-01 -2.52312481e-01 -5.15873469e-02
1.46848455e-01 7.31130481e-01 -1.77974975e+00 1.68602988e-01
1.06324112e+00 1.18346560e+00 6.61376834e-01 8.71317536e-02
8.91406648e-03 2.56501794e-01 -1.10873692e-01 -9.32226181e-02
4.44662690e-01 -5.68938136e-01 -5.09399176e-01 3.17185521e-01
3.55062544e-01 -5.51482260e-01 -4.67685431e-01 1.57292366e+00
6.26482606e-01 1.65294945e-01 4.66647923e-01 -7.12431192e-01
-9.65412974e-01 1.96745947e-01 -7.81917155e-01 -1.18883548e-03
7.33590424e-02 2.39789173e-01 9.78482366e-01 -8.30847085e-01
7.33762830e-02 5.07867396e-01 6.51526928e-01 2.23399654e-01
-1.03988087e+00 -6.55163050e-01 -1.24785475e-01 5.90968728e-01
-1.27579343e+00 -3.51561576e-01 1.11151981e+00 -2.61030495e-01
6.93198144e-01 2.82459110e-01 9.80412364e-01 8.48579466e-01
-2.50201195e-01 6.60053253e-01 1.59217811e+00 -4.23462749e-01
-2.09237672e-02 -1.08335577e-01 -1.62123561e-01 5.64307570e-01
4.99522120e-01 3.16426963e-01 -5.41424334e-01 1.64478108e-01
8.04225266e-01 3.38816553e-01 -8.13700557e-01 -1.33984998e-01
-8.62442791e-01 6.63812160e-01 6.71774685e-01 3.15881521e-01
-8.04737508e-01 -2.74091989e-01 -1.40423506e-01 3.72460857e-02
5.36000609e-01 3.52956891e-01 -3.82074118e-01 7.31755421e-02
-1.05635154e+00 -3.10387403e-01 1.95361897e-01 6.63748085e-01
9.88529861e-01 4.69926298e-01 3.48480314e-01 9.08543408e-01
4.67765659e-01 8.93843055e-01 2.75800020e-01 -8.27405810e-01
2.98992749e-02 4.69460934e-01 1.19518332e-01 -7.22347558e-01
-3.71032655e-01 -4.91908908e-01 -1.09256065e+00 1.71168655e-01
-1.75973907e-01 -8.67244452e-02 -7.73182631e-01 1.05574954e+00
-1.89388525e-02 -7.66541436e-02 2.54602302e-02 1.39112139e+00
1.10231197e+00 9.07658696e-01 -1.41137779e-01 -5.04923940e-01
1.01122963e+00 -8.18137228e-01 -6.84754789e-01 -2.51559854e-01
-1.41536251e-01 -8.09592783e-01 1.10340166e+00 7.73930550e-01
-8.62720251e-01 -4.40569162e-01 -1.24628758e+00 9.37353075e-02
-3.60268652e-01 5.46157718e-01 1.15646601e+00 8.30788791e-01
-7.33609259e-01 8.23107481e-01 -8.21978629e-01 -2.68963099e-01
7.33436227e-01 1.30274460e-01 -1.57789260e-01 -2.43945614e-01
-9.21368062e-01 5.90069056e-01 3.76228929e-01 4.21648949e-01
-8.54959488e-01 -5.49434960e-01 -5.66502512e-01 -2.14431643e-01
2.66913772e-01 -3.75052452e-01 9.22701716e-01 -8.81073773e-01
-1.84322906e+00 9.05508041e-01 -1.61747739e-01 8.34388137e-02
5.90425655e-02 -5.61400019e-02 -4.97243315e-01 5.82916915e-01
-3.13559771e-01 1.63699344e-01 9.33294475e-01 -1.28713810e+00
-4.04219180e-01 -6.24429643e-01 -1.79771125e-01 3.95377487e-01
-6.25651300e-01 -2.54093230e-01 -2.59428680e-01 -2.04287395e-01
5.83475173e-01 -7.43247151e-01 -5.29617146e-02 1.16436511e-01
-5.33429861e-01 3.38496447e-01 9.23980892e-01 -6.33252144e-01
7.05921948e-01 -2.03574061e+00 1.57560244e-01 -7.76695833e-02
-2.15898231e-02 6.51661575e-01 -1.29396096e-01 5.03740788e-01
-2.66306967e-01 2.22930629e-02 -6.25571966e-01 -1.05161048e-01
-2.43351594e-01 -1.73777148e-01 -3.89580250e-01 9.93505120e-01
-1.62557229e-01 7.91618049e-01 -7.85669148e-01 8.64330903e-02
6.49924338e-01 6.95306003e-01 1.89999253e-01 2.87289202e-01
-6.20406345e-02 7.05660641e-01 -4.81912225e-01 1.09335554e+00
9.15573120e-01 -4.59963500e-01 9.52684041e-03 -7.21845388e-01
-4.61470157e-01 1.38268977e-01 -8.71378243e-01 1.83289123e+00
-2.58724928e-01 5.55159330e-01 4.59048271e-01 -1.08602560e+00
7.73210585e-01 3.31946015e-01 7.66927063e-01 -7.61559248e-01
7.83764198e-02 4.30331290e-01 -3.34866643e-01 -6.50356352e-01
2.20235795e-01 -6.73510194e-01 7.10833728e-01 2.47577935e-01
-2.92077631e-01 -6.13897979e-01 -3.22044231e-02 -4.76592869e-01
2.56452233e-01 2.66405910e-01 2.28679508e-01 3.58125083e-02
5.66697001e-01 2.29212910e-01 2.44436622e-01 4.24874067e-01
-2.90739715e-01 5.07334173e-01 -3.25938433e-01 -5.03455102e-01
-9.04102802e-01 -8.02443087e-01 -5.74249387e-01 6.19814038e-01
3.53496224e-01 1.77046582e-01 -3.45622033e-01 -7.50203338e-03
-5.54820523e-02 2.56173730e-01 -2.15256363e-01 1.76802892e-02
1.05410725e-01 -1.39511335e+00 2.57451296e-01 3.11105907e-01
1.05107248e+00 -8.27992141e-01 -5.93957126e-01 -2.75101624e-02
-3.25960398e-01 -9.37489390e-01 2.29493499e-01 1.74327776e-01
-1.28887546e+00 -1.22164333e+00 -8.38141739e-01 -4.94137257e-01
2.95248836e-01 1.15354526e+00 7.06823707e-01 -1.27830192e-01
-5.65064251e-01 8.83909091e-02 -4.15383101e-01 -6.16028786e-01
1.22563891e-01 -2.10466877e-01 -1.35773465e-01 -1.26777086e-02
5.83269358e-01 -6.63892269e-01 -7.86756814e-01 4.82076295e-02
-1.18285406e+00 7.00450465e-02 7.30786145e-01 7.64027178e-01
9.95462239e-01 6.50469363e-01 2.49423221e-01 -6.96858406e-01
3.16979378e-01 -1.47473037e-01 -9.96588469e-01 1.63055047e-01
-7.59312212e-01 -4.37460244e-01 6.11907423e-01 1.01307280e-01
-1.34284246e+00 1.26723751e-01 -1.50374904e-01 -1.96368158e-01
-3.43446314e-01 9.97447610e-01 -1.16758749e-01 -6.16541088e-01
6.29467309e-01 5.54695725e-01 2.23241020e-02 -4.82468039e-01
7.99281374e-02 8.16530049e-01 3.73209476e-01 -3.54303181e-01
7.53974557e-01 9.12249744e-01 2.16791525e-01 -1.41783571e+00
-1.07868695e+00 -7.85360038e-01 -6.18550181e-01 -3.65351111e-01
6.94517016e-01 -9.36513901e-01 -5.28230429e-01 9.09732521e-01
-7.65367866e-01 -2.70214945e-01 9.07571539e-02 5.75440109e-01
-3.43910992e-01 7.07410455e-01 -7.35521019e-01 -9.27860916e-01
-4.64013070e-01 -1.28087986e+00 1.14105499e+00 4.89220113e-01
6.71599090e-01 -9.26113844e-01 -1.06952228e-01 6.17203116e-01
6.97700918e-01 6.22348227e-02 8.68569136e-01 1.67946428e-01
-8.33662987e-01 -8.94580185e-02 -7.00139046e-01 4.70788598e-01
4.93475497e-01 1.88751385e-01 -1.43027043e+00 -3.88548017e-01
2.37346426e-01 -6.62649751e-01 1.04999912e+00 7.07408249e-01
1.66666150e+00 1.43053651e-01 -1.36962727e-01 1.23641944e+00
2.10148311e+00 4.20616865e-01 7.57029533e-01 4.91429895e-01
7.33088374e-01 4.36801106e-01 5.19740522e-01 4.98500168e-01
-2.99915642e-01 1.95621371e-01 9.55180943e-01 -4.53494072e-01
1.00289956e-01 3.34723115e-01 -6.63881823e-02 4.76929694e-01
-6.65619493e-01 -1.36384577e-01 -6.92674160e-01 2.05127373e-01
-1.47971320e+00 -1.15288043e+00 -4.84076917e-01 2.09337449e+00
7.99582005e-01 -6.36270106e-01 -1.53663322e-01 2.91441500e-01
6.18420482e-01 5.05844295e-01 -9.09227490e-01 5.79432130e-01
-4.10702467e-01 3.29102188e-01 6.76455617e-01 1.40056396e-02
-1.31312597e+00 8.30441713e-01 7.27065706e+00 3.81522685e-01
-1.52495933e+00 -5.79485251e-03 4.01784360e-01 1.74395945e-02
-2.41848603e-01 -3.23912412e-01 -3.02762985e-01 1.79494783e-01
5.86022854e-01 1.44722298e-01 1.07732379e+00 5.62353551e-01
3.33938777e-01 -2.70352215e-01 -5.51742673e-01 1.49899459e+00
-1.69478524e-02 -1.41717517e+00 -1.32306039e-01 1.98904172e-01
9.37764168e-01 4.23905283e-01 3.47371966e-01 -3.15957904e-01
-1.74762234e-01 -8.88943613e-01 2.95327544e-01 6.17595315e-01
1.17685962e+00 -6.34104073e-01 7.88217366e-01 2.50933647e-01
-9.20770049e-01 -3.73665802e-02 -7.64970064e-01 -1.75477758e-01
-4.76898067e-03 1.04066133e+00 -7.92298838e-02 9.05312717e-01
7.97471225e-01 1.08118534e+00 -2.78001260e-02 1.14043701e+00
-4.47272956e-01 7.06289768e-01 -1.18135303e-01 2.39644244e-01
3.41633111e-01 -1.03166366e+00 2.29755864e-01 8.72546911e-01
5.61796665e-01 4.32311922e-01 1.10723823e-01 1.07159305e+00
5.20667247e-02 -6.31418377e-02 -6.48368537e-01 -6.81312442e-01
3.74675423e-01 1.44749534e+00 -5.57517111e-01 -2.09547463e-03
-7.83788979e-01 9.16083276e-01 -1.45025879e-01 6.53186023e-01
-7.28788614e-01 -2.41902441e-01 6.21013463e-01 -1.74413189e-01
1.17294058e-01 -4.54171926e-01 -3.50234270e-01 -1.14554453e+00
-1.90327004e-01 -9.19545352e-01 2.92918026e-01 -1.28937221e+00
-1.42238200e+00 4.34913963e-01 -2.62459457e-01 -1.18855572e+00
5.01750410e-01 -1.28756011e+00 -1.26828194e-01 1.28595257e+00
-2.28494906e+00 -9.57226992e-01 -9.50806856e-01 6.80509627e-01
3.57721865e-01 -2.44077787e-01 1.04325688e+00 1.69739544e-01
-8.49078536e-01 -3.12131703e-01 4.57192481e-01 -2.61299640e-01
4.48974609e-01 -9.38710690e-01 -4.17029619e-01 8.47739160e-01
1.95189621e-02 5.33822834e-01 2.91724712e-01 -3.56938362e-01
-1.96608055e+00 -9.85024393e-01 1.97885279e-02 1.53173879e-01
6.73383594e-01 4.78383631e-01 -8.58079255e-01 3.18716258e-01
3.02739859e-01 -1.10681355e-01 1.16573787e+00 -1.65417343e-01
-8.20663050e-02 -4.22052324e-01 -1.07171929e+00 2.16537863e-01
8.42338026e-01 -8.56257260e-01 -1.12826332e-01 6.28639340e-01
1.82849586e-01 -1.54486462e-01 -8.10766518e-01 5.83675385e-01
3.63345504e-01 -1.40434039e+00 1.00817895e+00 -4.53520231e-02
5.93751252e-01 -3.15949559e-01 -3.25301558e-01 -1.27353668e+00
-6.61252379e-01 -2.17595100e-01 1.35765105e-01 5.74345946e-01
9.45514217e-02 -7.51208067e-01 9.75049615e-01 3.44256490e-01
-4.16125923e-01 -4.24027026e-01 -5.45053601e-01 -6.50621355e-01
-4.66802657e-01 -3.64925295e-01 6.35755181e-01 1.06278229e+00
-3.93890709e-01 7.40901530e-02 -3.31855476e-01 5.91966510e-01
9.37493563e-01 7.11384892e-01 1.51138529e-01 -1.40549099e+00
-9.05365646e-02 -6.31321013e-01 1.35235056e-01 -8.40038240e-01
9.87690166e-02 -6.33664250e-01 4.86013219e-02 -1.84020805e+00
4.24758852e-01 -1.99389607e-01 -3.58395547e-01 4.07445282e-01
-6.15530387e-02 6.46505296e-01 -2.80184567e-01 6.97984338e-01
-2.44885311e-02 7.12712526e-01 1.71603143e+00 -4.43138272e-01
-6.87244311e-02 -6.68274537e-02 -7.58251905e-01 7.11866558e-01
9.37000155e-01 -9.05007422e-02 -5.40037096e-01 -6.32962584e-01
3.58810276e-01 -4.94875014e-02 2.99294770e-01 -9.51796710e-01
8.22970867e-02 -5.77766657e-01 4.64803815e-01 -7.60549545e-01
6.35260820e-01 -1.09796822e+00 2.77668446e-01 2.27635130e-01
2.19106153e-01 -5.50857127e-01 2.03059584e-01 4.63532656e-01
-4.04832453e-01 -4.32165831e-01 8.98154080e-01 -5.74719369e-01
-1.11900806e+00 7.32443452e-01 -2.39332870e-01 -7.30459988e-01
8.41116726e-01 -4.92865384e-01 -3.05133790e-01 -4.00285900e-01
-3.97834718e-01 -3.88861299e-01 6.38372242e-01 -3.81859183e-01
8.17492843e-01 -1.00258315e+00 -4.81263906e-01 3.27458769e-01
8.30169395e-02 1.07159279e-01 4.78181958e-01 7.82015681e-01
-8.45508814e-01 5.60516238e-01 -3.70390922e-01 -6.43788934e-01
-9.58521366e-01 4.39536333e-01 6.50132239e-01 3.34870905e-01
-2.97050804e-01 9.71828878e-01 -1.20962426e-01 -4.06629324e-01
3.47604156e-02 1.03859067e-01 -6.93128556e-02 8.33956292e-05
5.31902432e-01 5.93890607e-01 3.97166461e-01 -4.52382386e-01
-2.11152405e-01 5.69226146e-01 5.43365955e-01 1.19141757e-01
1.86359727e+00 -1.07126087e-02 -5.56664884e-01 3.41453940e-01
8.41038585e-01 -2.14260444e-01 -1.51846099e+00 -2.20808044e-01
-5.68543017e-01 -7.50177205e-01 6.72486424e-01 -8.06110919e-01
-1.37919986e+00 1.12631631e+00 6.55962467e-01 2.96281457e-01
1.74808872e+00 -4.03336525e-01 4.75188434e-01 4.62596923e-01
7.10621953e-01 -9.97872293e-01 8.25134292e-02 3.01934898e-01
7.38287389e-01 -1.46537733e+00 4.75654930e-01 -6.40087664e-01
-2.92970985e-01 1.42442381e+00 2.93144226e-01 4.26669508e-01
5.25114715e-01 -6.57627434e-02 1.27736285e-01 -5.20638525e-01
1.12887196e-01 -4.34982777e-01 2.38346681e-01 9.78634000e-01
7.64796555e-01 1.35282084e-01 -6.19119219e-02 1.62437081e-01
1.50052086e-01 -2.91740187e-02 2.06944466e-01 6.57241404e-01
-6.14400446e-01 -9.06741023e-01 -6.04854107e-01 5.07681549e-01
-1.01937607e-01 -4.50956702e-01 -2.88399905e-01 5.42168319e-01
-6.00971431e-02 1.07198882e+00 -3.69417429e-01 -2.07683519e-01
6.11931011e-02 -1.78711489e-02 7.74905086e-01 -5.07396042e-01
2.16940686e-01 3.00003111e-01 -6.49154410e-02 -4.10273939e-01
-1.16618299e+00 -7.04074085e-01 -9.06803489e-01 -2.12298378e-01
-4.75182295e-01 -6.90620095e-02 9.25253689e-01 9.42595661e-01
1.26010180e-01 4.11437690e-01 6.09176815e-01 -1.34168851e+00
-5.13790190e-01 -8.85004520e-01 -1.18247807e+00 1.60198137e-01
2.31838197e-01 -5.91352522e-01 -2.23909751e-01 3.50016169e-02] | [10.172677993774414, -2.096473455429077] |
474edd34-c408-4493-a0e0-0553aa432651 | channel-importance-matters-in-few-shot-image | 2206.08126 | null | https://arxiv.org/abs/2206.08126v2 | https://arxiv.org/pdf/2206.08126v2.pdf | Channel Importance Matters in Few-Shot Image Classification | Few-Shot Learning (FSL) requires vision models to quickly adapt to brand-new classification tasks with a shift in task distribution. Understanding the difficulties posed by this task distribution shift is central to FSL. In this paper, we show that a simple channel-wise feature transformation may be the key to unraveling this secret from a channel perspective. When facing novel few-shot tasks in the test-time datasets, this transformation can greatly improve the generalization ability of learned image representations, while being agnostic to the choice of training algorithms and datasets. Through an in-depth analysis of this transformation, we find that the difficulty of representation transfer in FSL stems from the severe channel bias problem of image representations: channels may have different importance in different tasks, while convolutional neural networks are likely to be insensitive, or respond incorrectly to such a shift. This points out a core problem of the generalization ability of modern vision systems and needs further attention in the future. Our code is available at https://github.com/Frankluox/Channel_Importance_FSL. | ['Zenglin Xu', 'Jing Xu', 'Xu Luo'] | 2022-06-16 | null | null | null | null | ['few-shot-image-classification'] | ['computer-vision'] | [ 6.90134048e-01 -2.66641587e-01 -6.34579584e-02 -3.89240235e-01
-7.43810594e-01 -8.45241785e-01 6.72751307e-01 -1.91311866e-01
-4.44443703e-01 6.08417928e-01 5.56063801e-02 -3.37187231e-01
-2.97771990e-02 -5.57711184e-01 -9.74626899e-01 -5.98401785e-01
-7.95290992e-02 -9.51758325e-02 3.05900872e-01 -3.61413538e-01
2.92612404e-01 3.83748174e-01 -1.41979265e+00 3.34691137e-01
2.50507534e-01 1.06589329e+00 1.72557458e-01 9.03205633e-01
6.20452911e-02 7.80657530e-01 -4.35302377e-01 -3.00071687e-01
7.08637714e-01 -4.37359750e-01 -7.63993680e-01 -8.89431015e-02
6.68757498e-01 -2.81630456e-01 -6.10644698e-01 1.32060719e+00
5.15286565e-01 3.33502702e-03 7.15736270e-01 -1.35938299e+00
-7.67869473e-01 3.83383900e-01 -4.13966715e-01 5.51274002e-01
-8.06794241e-02 5.06269515e-01 1.07558012e+00 -7.23944902e-01
5.70083380e-01 9.70300615e-01 6.46528244e-01 8.92638743e-01
-1.42099166e+00 -8.44512641e-01 2.27005467e-01 4.56969112e-01
-1.19614434e+00 -7.79640138e-01 6.09882414e-01 -6.65146053e-01
7.43827999e-01 1.12654015e-01 4.84138012e-01 1.38882959e+00
3.74536633e-01 5.44712245e-01 1.27846479e+00 -3.92564386e-01
3.31223637e-01 2.67026246e-01 9.38582420e-02 3.83172214e-01
4.47411656e-01 3.23158443e-01 -8.04125905e-01 -1.19513562e-02
5.47182262e-01 6.56957924e-02 -5.18044889e-01 -6.53994381e-01
-8.63598943e-01 8.85956287e-01 7.23627687e-01 2.21449241e-01
1.27507046e-01 4.72884446e-01 5.11165738e-01 8.56652141e-01
1.34388000e-01 8.73543918e-01 -7.30222404e-01 -2.37980738e-01
-4.68514889e-01 1.14608541e-01 7.42721617e-01 9.77168560e-01
1.01214504e+00 -6.82744533e-02 -8.06666762e-02 5.36586761e-01
-1.74582988e-01 2.88004786e-01 4.51543331e-01 -9.47336078e-01
1.66698575e-01 -5.20456843e-02 -2.14657396e-01 -6.47031367e-01
-2.01385081e-01 -4.64460552e-01 -4.12389874e-01 3.59488517e-01
6.45479620e-01 -2.89435029e-01 -9.55621243e-01 1.96870482e+00
-3.05070221e-01 3.12519595e-02 -1.81424543e-02 7.61630177e-01
4.23839957e-01 4.26813751e-01 -5.10941306e-03 -6.61682263e-02
1.16395974e+00 -4.95194316e-01 -2.51013100e-01 -5.99547029e-01
5.75934589e-01 -7.30529666e-01 1.15378129e+00 2.23116323e-01
-6.51196241e-01 -4.32378650e-01 -1.30647373e+00 1.11042947e-01
-4.87993121e-01 -7.22191453e-01 8.77361715e-01 9.09687579e-01
-8.94084096e-01 4.80006367e-01 -5.29565156e-01 -6.75406575e-01
9.14343774e-01 3.05109978e-01 -2.76497900e-01 -4.41567928e-01
-1.13418436e+00 8.48760247e-01 2.76039690e-01 -3.35297465e-01
-8.22758734e-01 -8.06018233e-01 -7.03326762e-01 2.47745082e-01
4.40898150e-01 -4.60346341e-01 1.52693331e+00 -1.38854098e+00
-1.19230890e+00 5.99533439e-01 9.78235453e-02 -5.84771991e-01
4.06249106e-01 1.26745239e-01 -1.98216781e-01 -3.26601826e-02
1.35537973e-02 7.04158723e-01 1.21620047e+00 -1.27425015e+00
-5.92294753e-01 -4.28262830e-01 1.17486836e-02 4.84360009e-02
-3.72974992e-01 -2.86882550e-01 -2.05638424e-01 -5.80739319e-01
-2.34013662e-01 -1.02002573e+00 -1.40984267e-01 3.21288645e-01
5.74010946e-02 2.90065944e-01 6.96297348e-01 -7.67460614e-02
7.05316901e-01 -2.50462484e+00 -1.74999252e-01 2.25502655e-01
3.76715004e-01 1.43064976e-01 -3.58513802e-01 4.48640674e-01
-1.94349304e-01 -1.30776018e-01 -1.90778673e-01 2.10105032e-01
-2.40482971e-01 1.46832526e-01 -5.69379449e-01 5.01448035e-01
2.43649036e-01 1.09027505e+00 -9.33486342e-01 3.75564396e-02
9.32956263e-02 2.40585700e-01 -4.09027159e-01 -4.42892015e-02
-1.90010369e-01 3.41475964e-01 -1.62095323e-01 6.01002574e-01
6.36489272e-01 -3.91734958e-01 -5.61346412e-02 -1.61219880e-01
4.03816961e-02 -1.40331283e-01 -7.05838025e-01 1.74800265e+00
-3.94129485e-01 1.19211626e+00 -1.36609018e-01 -9.87183213e-01
6.07395113e-01 -6.81634024e-02 2.44103253e-01 -8.96899283e-01
3.66974503e-01 1.20991841e-01 5.13796091e-01 -2.53949106e-01
2.86401302e-01 -4.65127856e-01 -6.86555654e-02 4.84471858e-01
4.75139558e-01 -7.47940242e-02 -1.98938668e-01 2.45874360e-01
1.27562320e+00 -5.25632203e-01 2.88502336e-01 -1.77119836e-01
-2.37376764e-01 2.59644184e-02 4.75818306e-01 1.25870156e+00
-5.47590017e-01 5.05303800e-01 5.49556077e-01 -3.09962153e-01
-1.02938986e+00 -1.12346351e+00 -2.85889894e-01 1.19578087e+00
1.68609843e-02 -2.33416289e-01 -4.30361986e-01 -5.63417792e-01
2.13825628e-01 6.93831801e-01 -8.34299088e-01 -7.42447078e-01
3.77176255e-02 -7.11730480e-01 4.54632580e-01 5.10896742e-01
3.95189583e-01 -5.31260788e-01 -9.09234941e-01 1.45053947e-02
4.40911204e-01 -1.01341021e+00 -2.29916990e-01 7.01212704e-01
-7.25699127e-01 -1.02244496e+00 -8.43028069e-01 -5.68648398e-01
6.88345075e-01 5.68430603e-01 6.44462228e-01 -1.05129063e-01
-5.79638183e-01 7.20698476e-01 -5.47547877e-01 -7.98366904e-01
-2.80745506e-01 5.81869371e-02 -1.27824068e-01 7.61900693e-02
4.48990136e-01 -5.59554160e-01 -7.29352355e-01 1.46414638e-01
-8.42882097e-01 -2.82351673e-01 6.70876741e-01 9.64429557e-01
6.01335615e-02 -4.10700776e-02 4.73182440e-01 -1.29521978e+00
7.23762989e-01 -6.24434292e-01 -4.49417263e-01 3.33642364e-02
-8.22673738e-01 8.99786651e-02 5.96019924e-01 -5.15000820e-01
-6.70633733e-01 2.07721248e-01 2.14356273e-01 -5.22729933e-01
-6.27366602e-02 2.16380492e-01 1.02500021e-02 -6.03096485e-01
1.16916156e+00 3.11799258e-01 1.84854522e-01 -2.11681873e-01
5.79059958e-01 5.66862345e-01 3.90541643e-01 -4.71011490e-01
8.92769158e-01 5.82058668e-01 -1.47662669e-01 -8.06677520e-01
-6.89378977e-01 -3.09798300e-01 -5.19169033e-01 -1.26217082e-01
5.95277011e-01 -9.24288869e-01 -2.43613690e-01 6.04951978e-01
-9.64995623e-01 -5.94369590e-01 -6.16029084e-01 3.19107205e-01
-6.67063117e-01 5.46296686e-02 -3.50085229e-01 -3.40154439e-01
4.27406915e-02 -1.13855088e+00 4.68182802e-01 2.51117736e-01
-1.88315868e-01 -8.30282748e-01 -1.30893782e-01 -2.06757542e-02
7.19876409e-01 -2.50958562e-01 1.16032207e+00 -6.09463096e-01
-6.03472114e-01 -3.11511070e-01 -4.91836965e-01 4.82205033e-01
2.95608222e-01 -3.96481782e-01 -1.57546711e+00 -6.33218050e-01
4.83629517e-02 -4.29288983e-01 1.40824485e+00 4.39006895e-01
1.26567805e+00 -1.28965169e-01 -5.70865124e-02 9.07884777e-01
1.57731771e+00 2.37514302e-01 6.16237879e-01 2.36286670e-01
5.44658005e-01 3.14376593e-01 1.72152624e-01 4.38069940e-01
1.41345531e-01 3.43763947e-01 3.08233857e-01 9.03327465e-02
-2.28020027e-01 -1.54934004e-01 2.93843210e-01 3.41221482e-01
9.85941291e-02 -5.84276989e-02 -9.17770922e-01 3.74816895e-01
-1.62571049e+00 -1.02773690e+00 5.02994120e-01 2.27461863e+00
6.82898700e-01 5.00198305e-01 -2.86592897e-02 -4.57890972e-04
5.54781199e-01 2.99536973e-01 -9.84165549e-01 -5.28285861e-01
-2.25510751e-03 3.07656139e-01 1.07124996e+00 3.41821551e-01
-8.74807239e-01 9.88681555e-01 6.70917225e+00 9.19827580e-01
-1.44228077e+00 1.41355321e-01 6.20285630e-01 -1.03700839e-01
-2.74515063e-01 2.14909688e-01 -6.26558125e-01 4.95098740e-01
8.79407883e-01 -4.88135248e-01 7.08900392e-01 7.23484397e-01
-4.73288357e-01 -8.84459838e-02 -1.35815167e+00 1.18616474e+00
3.01154613e-01 -1.44763327e+00 9.37844217e-02 6.68509677e-02
7.46279776e-01 4.19919252e-01 5.08179963e-01 3.22123796e-01
1.79782063e-01 -1.07946491e+00 5.44865549e-01 3.81495357e-01
1.12229681e+00 -3.68770480e-01 3.74326408e-01 4.19581756e-02
-7.99345672e-01 -4.18072671e-01 -5.34142971e-01 -3.84635299e-01
-4.31889623e-01 2.34535247e-01 -8.89962196e-01 -2.04539597e-01
6.07003689e-01 6.73771918e-01 -8.49385023e-01 1.03345931e+00
-4.12484668e-02 5.67876518e-01 3.20208371e-02 -5.96940555e-02
-5.69540337e-02 5.31967521e-01 4.56127673e-01 9.83932018e-01
2.06492528e-01 -5.98634221e-02 -3.07173938e-01 6.90598965e-01
-2.64858842e-01 -3.01343501e-01 -1.04645371e+00 -2.02500254e-01
5.18582642e-01 9.73775685e-01 -6.42881870e-01 -1.11540385e-01
-8.60468566e-01 1.14485729e+00 2.02740774e-01 6.44249320e-01
-3.40285987e-01 -4.36381966e-01 8.57568204e-01 1.05899088e-01
4.30409759e-01 -2.30426222e-01 -3.80314231e-01 -1.12796223e+00
-2.64484197e-01 -8.56977642e-01 2.59437948e-01 -7.08125353e-01
-1.58150458e+00 2.14044362e-01 -2.06187636e-01 -1.31844747e+00
-2.25391518e-02 -9.24083591e-01 -5.79125524e-01 6.42202675e-01
-1.52201068e+00 -8.45563114e-01 -2.62760222e-01 7.81520784e-01
7.56414711e-01 -4.81801927e-01 8.06511641e-01 1.53115287e-01
-1.45928055e-01 8.76776934e-01 2.22136468e-01 1.06679030e-01
1.09158361e+00 -9.79567230e-01 6.25781596e-01 7.71146536e-01
2.81598210e-01 5.50788701e-01 8.11868370e-01 -2.65695602e-01
-1.69715774e+00 -1.03475845e+00 2.95446396e-01 -5.88075936e-01
8.40446174e-01 -3.88621539e-01 -8.90167117e-01 7.39213645e-01
2.49490645e-02 2.90005416e-01 9.07820880e-01 2.24813879e-01
-9.55987096e-01 -1.54015496e-01 -9.30577338e-01 5.27819216e-01
9.98711467e-01 -9.05578613e-01 -3.55531245e-01 1.43969029e-01
6.16246521e-01 -5.69775105e-02 -3.57362241e-01 5.02948416e-03
7.01967359e-01 -8.50992203e-01 6.70495510e-01 -7.30722308e-01
3.20137173e-01 3.07819787e-02 -5.66111147e-01 -1.54828930e+00
-6.38739049e-01 -4.32751328e-01 1.02957077e-01 6.70282841e-01
5.90384007e-01 -8.39432120e-01 5.54253221e-01 7.36154795e-01
1.12045854e-01 -5.33754230e-01 -9.54730988e-01 -1.00117791e+00
3.39273095e-01 -6.75748587e-01 2.50520378e-01 9.89282668e-01
-7.66219897e-03 4.26426739e-01 -2.85064846e-01 4.35421020e-02
7.08781779e-01 -1.88676789e-01 7.60735273e-01 -1.25560689e+00
-5.21015644e-01 -5.06069422e-01 -8.71144414e-01 -8.27373683e-01
-6.70663789e-02 -1.01066601e+00 3.28258798e-03 -1.01800334e+00
4.25680459e-01 -2.41612434e-01 -6.20109200e-01 4.95803267e-01
2.09075913e-01 2.96962351e-01 5.93604922e-01 2.68259257e-01
-4.28280294e-01 3.49720240e-01 1.02841663e+00 -3.28023791e-01
-2.76143569e-02 9.26983505e-02 -1.18761599e+00 5.91600955e-01
9.81514931e-01 -4.67388928e-01 -5.00550985e-01 -4.83434349e-01
2.91737080e-01 -4.05527204e-01 5.60072184e-01 -1.12384713e+00
3.06331694e-01 -6.34314045e-02 6.47736847e-01 2.04210028e-01
4.29061800e-01 -8.63997281e-01 -2.82077521e-01 6.83986902e-01
-4.99799192e-01 -2.78401047e-01 4.16391224e-01 8.10859203e-01
1.34026840e-01 -1.60428718e-01 1.04675448e+00 -3.65673870e-01
-1.31131828e+00 3.65547627e-01 -3.97729874e-01 2.43340924e-01
1.04888999e+00 -3.76673847e-01 -6.36127830e-01 -4.15421337e-01
-4.96128261e-01 7.57245813e-03 6.24603868e-01 5.37823856e-01
5.83761454e-01 -1.14215994e+00 -3.55621606e-01 6.12844944e-01
6.20386541e-01 -5.40940642e-01 3.11269432e-01 5.86857736e-01
-1.92781687e-01 1.33313298e-01 -6.91441178e-01 -3.52582842e-01
-9.58198309e-01 4.81014013e-01 3.10735911e-01 5.44758797e-01
-4.61173505e-01 1.34600854e+00 3.70681345e-01 1.72347844e-01
2.20072806e-01 -1.43575802e-01 2.87996531e-01 2.16526657e-01
4.97787416e-01 1.30094394e-01 -7.34107420e-02 -2.71614164e-01
-2.89500743e-01 5.13267398e-01 -5.45229614e-01 -6.12968691e-02
1.12841761e+00 -6.95688352e-02 3.73389542e-01 7.47759759e-01
1.25377178e+00 -3.75781834e-01 -1.57137084e+00 -4.44145173e-01
-3.67925614e-01 -8.05845797e-01 1.33632243e-01 -7.30080009e-01
-9.31070268e-01 1.14244187e+00 9.52496767e-01 5.97603321e-02
9.27409410e-01 2.20347956e-01 6.48775935e-01 5.21969557e-01
4.54940110e-01 -1.14944577e+00 9.54441726e-02 6.23608649e-01
5.95048845e-01 -1.48030937e+00 -1.48458064e-01 -1.63800150e-01
-6.05282664e-01 1.05948520e+00 5.50165534e-01 -1.12300545e-01
8.81954610e-01 2.28447273e-01 9.10101905e-02 -1.40357882e-01
-8.59481275e-01 -3.27820390e-01 -1.11626141e-01 9.16502297e-01
1.23942189e-01 7.89596438e-02 1.22922130e-01 2.68299222e-01
-1.22118227e-01 -2.74007465e-03 6.77370965e-01 9.48652327e-01
-6.61292553e-01 -9.50721681e-01 -1.63131267e-01 7.77488589e-01
-6.04483411e-02 -2.55416900e-01 -4.12795454e-01 4.60053384e-01
9.09618214e-02 6.33665204e-01 1.25787005e-01 -6.12651289e-01
1.80750042e-01 2.20386609e-01 7.83963621e-01 -9.42019105e-01
5.91370137e-03 -2.35280916e-01 -3.73199701e-01 -5.02128661e-01
8.06241482e-02 -6.31313384e-01 -9.82938945e-01 -2.54940599e-01
-9.80192348e-02 -3.64715576e-01 5.17356157e-01 7.92648375e-01
3.96268010e-01 3.79407108e-01 7.18629062e-01 -7.55380630e-01
-9.04461503e-01 -6.02668464e-01 -7.95083225e-01 3.02023292e-01
7.02451825e-01 -6.37804568e-01 -5.12002468e-01 9.75646973e-02] | [9.820462226867676, 2.71223521232605] |
b96eff5c-9e77-4611-b0cf-3c0b00c85574 | gamival-video-quality-prediction-on-mobile | 2305.02422 | null | https://arxiv.org/abs/2305.02422v2 | https://arxiv.org/pdf/2305.02422v2.pdf | GAMIVAL: Video Quality Prediction on Mobile Cloud Gaming Content | The mobile cloud gaming industry has been rapidly growing over the last decade. When streaming gaming videos are transmitted to customers' client devices from cloud servers, algorithms that can monitor distorted video quality without having any reference video available are desirable tools. However, creating No-Reference Video Quality Assessment (NR VQA) models that can accurately predict the quality of streaming gaming videos rendered by computer graphics engines is a challenging problem, since gaming content generally differs statistically from naturalistic videos, often lacks detail, and contains many smooth regions. Until recently, the problem has been further complicated by the lack of adequate subjective quality databases of mobile gaming content. We have created a new gaming-specific NR VQA model called the Gaming Video Quality Evaluator (GAMIVAL), which combines and leverages the advantages of spatial and temporal gaming distorted scene statistics models, a neural noise model, and deep semantic features. Using a support vector regression (SVR) as a regressor, GAMIVAL achieves superior performance on the new LIVE-Meta Mobile Cloud Gaming (LIVE-Meta MCG) video quality database. | ['Ioannis Katsavounidis', 'Rahul Gowda', 'Alan C. Bovik', 'Xiaoming Wang', 'Bo Qiu', 'Chase Davis', 'Avinab Saha', 'Yu-Chih Chen'] | 2023-05-03 | null | null | null | null | ['video-quality-assessment', 'video-quality-assessment'] | ['computer-vision', 'time-series'] | [-3.18033136e-02 -8.11075807e-01 7.42831379e-02 -2.93198287e-01
-1.20281792e+00 -4.58256304e-01 2.19596401e-01 -5.56980133e-01
-3.35973613e-02 2.52103746e-01 2.00718999e-01 -2.33225599e-01
2.09015869e-02 -7.77750611e-01 -4.80628610e-01 -3.80470306e-01
-1.49241656e-01 7.09733143e-02 5.07126331e-01 -4.03093725e-01
4.26186889e-01 9.25862789e-02 -1.65116179e+00 7.34622061e-01
7.72732735e-01 1.58933067e+00 3.33035469e-01 1.24962413e+00
1.34036809e-01 1.37307346e+00 -5.76221943e-01 -5.73581994e-01
3.84660274e-01 -3.73587787e-01 -4.92044210e-01 1.27606407e-01
3.57775033e-01 -1.00445271e+00 -8.41647565e-01 1.00183582e+00
4.13167506e-01 1.43759251e-01 3.52899283e-01 -1.44004405e+00
-6.04235232e-01 -1.16518416e-01 -3.41199487e-01 8.51480782e-01
6.57218874e-01 4.01859760e-01 1.25574028e+00 -7.54197955e-01
4.66742009e-01 9.62149203e-01 5.83668947e-01 4.08061147e-01
-7.28600323e-01 -6.84098423e-01 -5.83592169e-02 7.27456927e-01
-1.24043262e+00 -6.33020401e-01 6.44737482e-01 -4.84429359e-01
9.59617138e-01 4.26099330e-01 9.54655170e-01 9.63783026e-01
4.76963699e-01 7.30819404e-01 6.48093402e-01 2.31236264e-01
4.15267557e-01 -1.69332534e-01 -7.16364622e-01 6.10854149e-01
-4.09043223e-01 8.23847800e-02 -7.73057103e-01 8.46493095e-02
1.05149782e+00 3.01960073e-02 -2.94245839e-01 -3.24744672e-01
-4.91115779e-01 6.12398505e-01 4.19856198e-02 -8.35118517e-02
-4.76498693e-01 1.40462548e-01 4.93459195e-01 4.89861608e-01
6.79815829e-01 2.64013529e-01 -2.51093298e-01 -1.05132103e+00
-1.57354534e+00 4.18399870e-01 5.63168943e-01 8.34977508e-01
3.56429309e-01 4.83987838e-01 2.69047350e-01 8.34238470e-01
2.92521596e-01 5.76664865e-01 4.30418432e-01 -1.69959342e+00
5.24180532e-01 2.77327210e-01 4.62504551e-02 -1.41755664e+00
1.03326358e-01 -2.41232008e-01 -4.98013079e-01 3.57690901e-01
2.16225207e-01 2.08993644e-01 -4.32478249e-01 1.24753451e+00
-1.56424835e-01 6.75875127e-01 -2.84548402e-01 1.27327716e+00
1.05120385e+00 8.96455586e-01 -1.72169790e-01 -1.94191225e-02
7.86108553e-01 -6.83898985e-01 -4.12009358e-01 1.64268538e-02
1.73419565e-01 -6.09998703e-01 1.30338895e+00 1.00536942e+00
-1.64091933e+00 -5.49525499e-01 -1.18388426e+00 -5.40846959e-03
1.08166665e-01 -6.22685969e-01 5.02205849e-01 9.76317048e-01
-1.45562172e+00 5.54575205e-01 -8.06342304e-01 2.22436532e-01
6.90364003e-01 1.21556312e-01 -1.79848820e-01 -4.31066841e-01
-1.03769422e+00 5.47094882e-01 -2.79598236e-01 -1.94951952e-01
-1.24742210e+00 -9.84482467e-01 -6.67761028e-01 1.60870522e-01
3.92575890e-01 -5.43605328e-01 1.31603849e+00 -1.58789933e+00
-1.73539472e+00 5.80119848e-01 -4.19269428e-02 1.39060803e-02
5.02365828e-01 -1.36856735e-01 -8.15091372e-01 5.67123294e-01
-3.49777453e-02 1.03535526e-01 1.08591056e+00 -1.05097115e+00
-8.91938984e-01 -4.58554000e-01 3.41542989e-01 3.86078179e-01
2.90054791e-02 3.44816446e-01 -1.03387511e+00 -5.46440244e-01
3.73272523e-02 -4.84100848e-01 3.31298113e-02 -1.27395540e-01
2.23435327e-01 4.22541320e-01 6.94982529e-01 -1.07438123e+00
1.41951728e+00 -2.21135712e+00 1.00402243e-01 1.98470801e-01
5.81748307e-01 3.42376202e-01 -3.46879721e-01 -5.37487380e-02
8.93628597e-02 1.15320235e-01 3.99927944e-01 2.41659004e-02
-2.88569897e-01 7.25979684e-03 -2.20951196e-02 1.98202789e-01
1.33279637e-01 7.90745676e-01 -1.00187552e+00 -4.03147250e-01
3.11171323e-01 4.55891758e-01 -1.21148193e+00 4.44658846e-01
-4.28051464e-02 2.43187144e-01 -2.75323898e-01 9.80141401e-01
7.27862775e-01 -5.39553285e-01 -7.46849477e-02 -1.55524284e-01
1.00580394e-01 2.66390324e-01 -1.24315512e+00 1.79776502e+00
-5.21287739e-01 8.66071701e-01 1.79664403e-01 -5.94339073e-01
4.41210598e-01 2.93352515e-01 9.85181451e-01 -1.39909315e+00
9.15308446e-02 8.36728364e-02 -3.22058141e-01 -8.02027285e-01
9.65482652e-01 -6.24623075e-02 2.65121639e-01 5.74496277e-02
2.54855826e-02 -2.90542722e-01 -1.19923107e-01 3.21450591e-01
1.50434744e+00 1.21023357e-01 -5.03564961e-02 3.71056110e-01
1.74324200e-01 -1.13423876e-01 7.40639806e-01 5.88657498e-01
-7.47822046e-01 1.21883762e+00 5.37395537e-01 -1.90641165e-01
-1.18661702e+00 -1.43474627e+00 2.99590498e-01 1.24686921e+00
3.57414067e-01 -6.45329833e-01 -6.86185300e-01 -6.01281673e-02
-3.09451073e-01 2.44986609e-01 -6.44324794e-02 -4.04265933e-02
-4.56201941e-01 -2.74900883e-01 4.24748033e-01 4.71979648e-01
5.17557502e-01 -7.61896551e-01 -3.83402020e-01 4.13688600e-01
-4.76060748e-01 -1.28788149e+00 -5.50913930e-01 -6.73408806e-01
-7.77453542e-01 -1.13142657e+00 -6.03970706e-01 -2.11225092e-01
-5.38273692e-01 7.61407912e-01 1.74554396e+00 1.16527239e-02
-1.39895976e-01 5.50858617e-01 -4.33851898e-01 4.52198312e-02
-1.91058293e-01 -3.17630768e-01 -1.63598090e-01 -4.03478891e-02
1.58698454e-01 -6.71337903e-01 -7.67791331e-01 4.08219874e-01
-1.01297128e+00 -7.82407150e-02 4.94576134e-02 4.46016282e-01
6.14087999e-01 3.47718030e-01 3.02712023e-01 -6.00456297e-01
4.59174126e-01 -8.47663581e-01 -5.97901404e-01 -1.89057857e-01
-3.18743825e-01 -6.89338923e-01 5.80043018e-01 -8.49682763e-02
-1.01002991e+00 -8.37408423e-01 -4.79538888e-01 -8.21340680e-01
1.54505268e-01 4.75747436e-01 -4.76359427e-01 -4.69968058e-02
5.93539715e-01 1.27957821e-01 -9.75523517e-03 -2.42443979e-01
1.02700137e-01 7.99316883e-01 7.63578534e-01 -6.24133572e-02
3.62898827e-01 4.45424020e-01 -2.20793679e-01 -1.08148098e+00
-1.74804479e-01 -7.06264198e-01 4.67364350e-03 -7.71127522e-01
6.07911706e-01 -1.34866762e+00 -7.20579624e-01 6.54417872e-01
-7.10144758e-01 -3.87029201e-01 -4.67531718e-02 4.10284311e-01
-8.37270439e-01 4.50188905e-01 -9.35165048e-01 -6.82637155e-01
1.39512792e-01 -1.40774393e+00 9.32598412e-01 8.84718150e-02
2.84862407e-02 -7.82709062e-01 -2.42560729e-02 8.64391863e-01
6.34358048e-01 -1.76405255e-02 6.17696464e-01 2.21105695e-01
-1.11323059e+00 -5.63316643e-02 -5.46481967e-01 5.72195232e-01
1.09870784e-01 3.67476702e-01 -9.13883269e-01 -1.55174747e-01
2.69279659e-01 -2.20372155e-01 3.79811466e-01 8.95192564e-01
1.37385368e+00 -7.27439746e-02 4.85578775e-01 1.15724576e+00
1.62476361e+00 4.47283030e-01 1.18797588e+00 4.00962174e-01
6.99239552e-01 1.96474627e-01 6.11984551e-01 6.75577462e-01
6.58592939e-01 8.00677240e-01 7.93291450e-01 1.98028669e-01
3.74021344e-02 -2.19034851e-02 5.23265004e-01 1.04759753e+00
-5.55039287e-01 -3.66560757e-01 -6.29092515e-01 2.21510604e-01
-1.47528815e+00 -1.38188589e+00 3.64046693e-02 1.92365837e+00
3.14719200e-01 1.79293230e-01 4.46787387e-01 2.64262289e-01
2.70389855e-01 3.14394057e-01 -7.78744280e-01 -4.35054749e-01
-1.18872963e-01 2.31591657e-01 3.07829201e-01 2.00990841e-01
-9.00208116e-01 7.26297796e-01 6.50540400e+00 9.40236032e-01
-1.22062814e+00 2.30187640e-01 6.60858870e-01 -6.21843934e-01
-4.44967866e-01 -5.18067360e-01 -2.35646423e-02 7.95624495e-01
1.29034913e+00 1.16274208e-01 9.72612977e-01 1.02053118e+00
8.99966896e-01 -1.52133688e-01 -7.15930998e-01 1.66813171e+00
2.05654219e-01 -1.45036328e+00 -6.19241409e-02 1.39616251e-01
6.21037424e-01 1.77388489e-01 7.03946531e-01 3.43884856e-01
3.31850126e-02 -1.17430580e+00 8.63234758e-01 7.08568513e-01
1.09569502e+00 -7.65164256e-01 6.77365839e-01 4.46558632e-02
-1.17761827e+00 -1.29455775e-01 -3.47374737e-01 -1.44522503e-01
2.05280825e-01 2.04112440e-01 -1.51812762e-01 3.45227391e-01
1.18357265e+00 8.82039011e-01 -5.43996394e-01 1.20631778e+00
5.65569878e-01 7.93262959e-01 3.52499276e-01 4.26330239e-01
1.84483454e-01 -4.57115948e-01 5.28659284e-01 7.62311399e-01
5.08479476e-01 3.17351013e-01 -1.26151711e-01 6.48845196e-01
-6.32536318e-03 1.11494683e-01 -4.83964145e-01 -1.44647956e-01
3.13269608e-02 8.38457882e-01 -3.18000436e-01 -1.60060003e-01
-9.59073067e-01 1.12961400e+00 -2.71794140e-01 3.37944746e-01
-1.01951361e+00 1.91717818e-02 1.46809721e+00 2.52287567e-01
2.70700485e-01 -2.96370566e-01 -2.68309802e-01 -1.42168891e+00
1.56048298e-01 -1.42167974e+00 6.65309951e-02 -1.33577526e+00
-1.14219940e+00 6.69855416e-01 -6.08405352e-01 -1.50708461e+00
-5.49766243e-01 -5.33685088e-01 -2.30983004e-01 8.28491509e-01
-1.39481366e+00 -6.43585801e-01 -5.91796458e-01 9.82805312e-01
9.67926860e-01 -4.93777275e-01 4.20604229e-01 7.28013098e-01
-4.19113666e-01 5.83790481e-01 3.90432775e-01 -6.97293654e-02
3.90574723e-01 -1.03723919e+00 3.01668286e-01 7.19098687e-01
-5.23713836e-03 5.03076315e-02 7.00124681e-01 -4.04730707e-01
-1.62837970e+00 -9.31255460e-01 2.42161259e-01 -5.85718811e-01
5.64213932e-01 -9.53525025e-03 -9.71581638e-01 2.14995369e-01
-3.18114124e-02 -4.11567464e-02 7.53512323e-01 -1.26698866e-01
-4.09809649e-01 -2.57248640e-01 -1.18449783e+00 5.87160647e-01
9.67004955e-01 -9.66313004e-01 -4.14765961e-02 -1.25391573e-01
4.85245347e-01 -4.68227774e-01 -8.01488161e-01 -1.56485271e-02
6.75429583e-01 -1.65122759e+00 1.15184832e+00 -5.19613087e-01
9.19810414e-01 -1.12624496e-01 -7.10289776e-01 -1.16886246e+00
-4.28978920e-01 -4.29713786e-01 -1.70102730e-01 8.34882200e-01
-1.60325237e-03 9.43890586e-02 1.24625826e+00 8.86667132e-01
-8.49512592e-02 -3.84321958e-01 -8.20154488e-01 -7.05566466e-01
-2.72245526e-01 -1.31951988e+00 7.81323433e-01 5.94434857e-01
-2.67346591e-01 -2.15253815e-01 -6.01041973e-01 1.44625694e-01
3.72921765e-01 -4.41121131e-01 6.38197720e-01 -8.17521155e-01
-7.04529166e-01 -5.63509405e-01 -9.80387390e-01 -1.10649419e+00
-1.80550039e-01 -4.77425873e-01 -1.83314115e-01 -1.24252331e+00
2.84375310e-01 -3.37274343e-01 -3.90697956e-01 -5.51705956e-01
-7.70938694e-02 5.59617221e-01 2.74987489e-01 3.34748596e-01
-1.07037044e+00 4.66444939e-01 1.17865944e+00 -2.02782191e-02
-1.91406250e-01 2.67001420e-01 -7.01305926e-01 7.23457932e-01
4.13653672e-01 3.21698897e-02 -5.21305203e-01 -5.87896526e-01
7.24282324e-01 7.98234701e-01 3.33151489e-01 -1.13661194e+00
-9.73493606e-03 -3.82110029e-01 2.34739572e-01 -2.86390901e-01
6.54174507e-01 -6.95834816e-01 3.35356325e-01 -1.33334115e-01
-4.58446741e-02 4.41437423e-01 -1.28638610e-01 5.97620904e-01
-4.48645473e-01 1.26782537e-01 8.14723372e-01 -2.46635020e-01
-1.18789232e+00 6.30274415e-01 -6.65136337e-01 4.47847962e-01
4.34201181e-01 -6.60305262e-01 6.34352863e-02 -9.99345005e-01
-5.34395039e-01 -2.09823862e-01 7.46841133e-01 5.03915668e-01
1.10768402e+00 -1.32903314e+00 -6.28471494e-01 1.80337965e-01
-1.50444405e-02 -4.33590978e-01 6.98748410e-01 3.55100751e-01
-1.06453717e+00 1.61903843e-01 -4.35392648e-01 -6.70631349e-01
-1.23781097e+00 4.25790489e-01 3.70727748e-01 -1.42334218e-04
-3.98085892e-01 8.17806542e-01 1.26678105e-02 1.21155858e-01
-1.79360192e-02 -2.35958174e-01 -2.47473493e-01 -2.54791468e-01
9.77702677e-01 8.90516341e-01 4.19697672e-01 -9.38039720e-01
-5.83325401e-02 1.33195773e-01 2.07852513e-01 -2.62540668e-01
1.42141497e+00 -5.22472441e-01 2.83903509e-01 4.39327985e-01
1.39834714e+00 -8.07719752e-02 -1.66764045e+00 -1.52033865e-01
-2.65023381e-01 -1.27365398e+00 4.87597257e-01 -5.64742744e-01
-1.69474745e+00 7.92216778e-01 9.02694941e-01 3.33675504e-01
1.39766514e+00 -3.47674847e-01 1.02707255e+00 -3.18536460e-01
6.17872000e-01 -1.16443455e+00 5.13238907e-01 5.12096465e-01
6.52204692e-01 -1.45177078e+00 -4.07881051e-01 -3.42392206e-01
-9.05508935e-01 8.25213492e-01 4.27036822e-01 -1.65952340e-01
6.80064380e-01 1.94745168e-01 2.34647393e-01 -6.41614944e-02
-8.97321641e-01 7.04284534e-02 3.40077490e-01 7.97594607e-01
2.98832715e-01 8.04717168e-02 5.05618513e-01 7.21990824e-01
-2.55020231e-01 3.58542919e-01 7.27698207e-01 5.50200582e-01
-2.68028826e-01 -6.91304862e-01 -2.12561712e-01 8.24618042e-01
-8.03658605e-01 -3.57119560e-01 2.89077431e-01 3.59446913e-01
4.33026813e-02 1.43718004e+00 2.56536752e-01 -8.03518534e-01
3.15999776e-01 -4.20150250e-01 4.32032019e-01 -4.00391757e-01
-5.35947919e-01 7.93539360e-02 -1.61806673e-01 -1.41589415e+00
-1.41072825e-01 -7.03192174e-01 -7.53946900e-01 -9.46971953e-01
7.00542778e-02 -2.48518109e-01 7.30520189e-01 9.29255426e-01
1.89352825e-01 6.75486267e-01 7.88264811e-01 -9.65753436e-01
-1.25328228e-01 -3.20828319e-01 -1.07726896e+00 4.82624650e-01
2.84989715e-01 -4.76072580e-01 -2.38208070e-01 5.82095422e-02] | [11.724225997924805, -1.819278597831726] |
b2bea6f5-f978-4327-af9e-24c89b23628d | fd-mar-fourier-dual-domain-network-for-ct | 2207.11678 | null | https://arxiv.org/abs/2207.11678v2 | https://arxiv.org/pdf/2207.11678v2.pdf | Quad-Net: Quad-domain Network for CT Metal Artifact Reduction | Metal implants and other high-density objects in patients introduce severe streaking artifacts in CT images, compromising image quality and diagnostic performance. Although various methods were developed for CT metal artifact reduction over the past decades, including the latest dual-domain deep networks, remaining metal artifacts are still clinically challenging in many cases. Here we extend the state-of-the-art dual-domain deep network approach into a quad-domain counterpart so that all the features in the sinogram, image, and their corresponding Fourier domains are synergized to eliminate metal artifacts optimally without compromising structural subtleties. Our proposed quad-domain network for MAR, referred to as Quad-Net, takes little additional computational cost since the Fourier transform is highly efficient, and works across the four receptive fields to learn both global and local features as well as their relations. Specifically, we first design a Sinogram-Fourier Restoration Network (SFR-Net) in the sinogram domain and its Fourier space to faithfully inpaint metal-corrupted traces. Then, we couple SFR-Net with an Image-Fourier Refinement Network (IFR-Net) which takes both an image and its Fourier spectrum to improve a CT image reconstructed from the SFR-Net output using cross-domain contextual information. Quad-Net is trained on clinical datasets to minimize a composite loss function. Quad-Net does not require precise metal masks, which is of great importance in clinical practice. Our experimental results demonstrate the superiority of Quad-Net over the state-of-the-art MAR methods quantitatively, visually, and statistically. The Quad-Net code is publicly available at https://github.com/longzilicart/Quad-Net. | ['Hongming Shan', 'Ge Wang', 'Meiyun Wang', 'Junping Zhang', 'Chuang Niu', 'Yaping Wu', 'Qi Gao', 'Zilong Li'] | 2022-07-24 | null | null | null | null | ['metal-artifact-reduction'] | ['medical'] | [ 1.83135614e-01 1.00359591e-02 1.39852166e-01 -6.26021475e-02
-1.10100400e+00 -1.17165752e-01 6.14849143e-02 -1.93839446e-01
-1.83684245e-01 6.97923422e-01 3.16574097e-01 -2.08400577e-01
-3.33232284e-01 -8.38413060e-01 -7.85229504e-01 -8.30237210e-01
1.00481503e-01 2.00139731e-01 4.43141222e-01 -1.36030659e-01
-1.70456991e-01 5.06377995e-01 -5.74563265e-01 4.14552987e-01
1.08304930e+00 1.00632334e+00 3.48425120e-01 2.99818158e-01
4.09105659e-01 8.62812400e-01 -3.08859795e-01 -1.57213703e-01
4.17082459e-01 -6.11115038e-01 -6.67866945e-01 -1.88889399e-01
1.35801032e-01 -4.70925719e-01 -8.93448412e-01 1.01832724e+00
7.30628371e-01 -1.23514114e-02 5.45042872e-01 -4.84350592e-01
-9.41531241e-01 4.40462768e-01 -7.55125105e-01 3.42123419e-01
-6.69951513e-02 3.18873614e-01 3.24846238e-01 -9.39378798e-01
5.67020774e-01 7.15848088e-01 1.07864010e+00 5.31032622e-01
-1.24846315e+00 -7.38457978e-01 -3.21479291e-01 -6.61593899e-02
-1.17891383e+00 -8.82308872e-04 1.09333217e+00 -4.13983047e-01
5.38757145e-01 2.11628050e-01 7.69779623e-01 8.72374594e-01
7.17228174e-01 5.87509692e-01 1.11322701e+00 -1.83326334e-01
-3.60366917e-04 -3.64719957e-01 -1.48311093e-01 8.13116491e-01
3.57787907e-01 1.40808210e-01 -2.96455353e-01 -1.93625450e-01
1.46448183e+00 4.62422848e-01 -8.22650135e-01 -2.72676349e-01
-1.17073941e+00 6.39731944e-01 8.61221850e-01 6.72595441e-01
-6.62652731e-01 3.20519596e-01 2.08595291e-01 -1.16117001e-01
5.18426418e-01 5.44045627e-01 -7.59497657e-02 2.56087363e-01
-9.74396646e-01 1.64130062e-01 1.46739244e-01 3.84145945e-01
3.74954045e-01 2.20051870e-01 -3.03052932e-01 9.99360085e-01
2.11676016e-01 3.63886595e-01 8.03115666e-01 -8.94184172e-01
2.49466911e-01 3.71154070e-01 -1.13463826e-01 -9.25264239e-01
-5.53253770e-01 -7.67588794e-01 -1.24874294e+00 2.84127712e-01
2.84863174e-01 -9.95943174e-02 -1.32221961e+00 1.51544309e+00
3.89575601e-01 3.35051179e-01 -2.46389613e-01 1.21016872e+00
9.25908029e-01 4.16825086e-01 -2.65820533e-01 -2.59344459e-01
1.32637656e+00 -8.49577785e-01 -8.04969430e-01 -2.28686884e-01
4.31663156e-01 -8.13012660e-01 9.75421846e-01 4.36446875e-01
-1.53454709e+00 -3.14897478e-01 -1.26622248e+00 -6.92099556e-02
3.38152736e-01 -7.95758143e-02 3.96464497e-01 2.63768256e-01
-9.46081400e-01 9.84042346e-01 -1.22523987e+00 3.50892454e-01
7.63991594e-01 4.55371946e-01 -1.88821524e-01 -3.16780180e-01
-1.21212375e+00 9.44655657e-01 -2.63270497e-01 2.82591224e-01
-9.85368848e-01 -1.26210332e+00 -8.35463166e-01 -2.26690874e-01
2.52878278e-01 -9.73240852e-01 1.37575531e+00 -7.86973715e-01
-1.41747999e+00 5.49740553e-01 1.30801976e-01 -3.25969785e-01
7.55349278e-01 -1.13426000e-01 -4.67351109e-01 4.08788115e-01
2.35789016e-01 9.58282351e-02 8.84097159e-01 -1.33778048e+00
-1.92601718e-02 -2.96306401e-01 -3.11217517e-01 4.34906408e-02
-5.94218113e-02 -1.69826686e-01 -4.56009895e-01 -1.29496753e+00
4.69020933e-01 -7.79996395e-01 -4.74199593e-01 4.42844361e-01
-2.77074307e-01 2.95661360e-01 5.07436812e-01 -1.01409507e+00
1.17324030e+00 -2.17045879e+00 -9.70211998e-02 1.43222302e-01
5.85468888e-01 1.59242094e-01 -4.97707985e-02 -7.37726986e-02
-4.57879812e-01 -1.06663361e-01 -8.14174116e-01 -2.07658321e-01
-6.56110168e-01 1.08299851e-01 1.80078864e-01 8.75968933e-01
1.38331994e-01 9.86098707e-01 -9.24640298e-01 -2.89111525e-01
2.91158825e-01 8.05934966e-01 -6.27869546e-01 9.22481064e-03
1.71606570e-01 7.93329835e-01 -5.42412817e-01 5.55668592e-01
1.02856779e+00 -5.63971519e-01 -6.64819684e-03 -5.37834346e-01
-8.67723003e-02 2.60576785e-01 -7.53196537e-01 2.02181840e+00
-7.07113683e-01 2.68103510e-01 6.95727617e-02 -7.62329996e-01
5.71594954e-01 4.42695171e-01 1.10275912e+00 -1.08570647e+00
2.01541051e-01 6.52129352e-01 1.42133206e-01 -4.46851760e-01
1.06381059e-01 -6.76965475e-01 3.16442966e-01 3.19680899e-01
-1.30325317e-01 -1.15539499e-01 -2.90130734e-01 4.37446460e-02
1.24036336e+00 -8.53304565e-02 5.92586957e-02 -3.09760034e-01
2.90872425e-01 -7.29214773e-02 6.82974517e-01 5.06890833e-01
-6.90505579e-02 1.24681711e+00 1.01014227e-01 -5.76799452e-01
-1.05518329e+00 -1.39789593e+00 -4.43344295e-01 1.68165907e-01
3.18123609e-01 1.85820311e-02 -6.27597630e-01 -5.19082904e-01
-1.77333176e-01 3.84650797e-01 -7.00621486e-01 -3.47045928e-01
-1.07571959e+00 -8.41693938e-01 3.79185349e-01 7.32156098e-01
4.92761046e-01 -9.55346584e-01 -4.90120649e-01 5.22832990e-01
-3.89368623e-01 -6.59161031e-01 -8.55362713e-01 3.17028999e-01
-1.14504087e+00 -1.02490199e+00 -1.21964967e+00 -8.16220224e-01
7.42502809e-01 2.35832319e-01 1.09900296e+00 1.98340401e-01
-5.89507639e-01 -1.03783123e-01 -2.15110153e-01 -5.34078702e-02
-3.95287931e-01 -4.14467126e-01 -3.02188307e-01 -3.65124084e-02
-3.07592392e-01 -9.71981049e-01 -1.21580350e+00 3.02214622e-01
-1.20135570e+00 1.79263428e-01 7.39733338e-01 1.16681230e+00
8.11446071e-01 -3.36475000e-02 4.25544292e-01 -8.41093361e-01
3.82943988e-01 -3.90603572e-01 -2.59265244e-01 -1.87240019e-02
-3.13916862e-01 3.59493913e-03 5.21584451e-01 -2.80128270e-01
-1.06325877e+00 5.97917251e-02 -4.65159118e-01 -6.90029919e-01
3.54821593e-01 4.53770250e-01 2.12239906e-01 -2.58881003e-01
9.30195749e-01 3.52149934e-01 6.62975460e-02 -3.48737717e-01
-7.32394829e-02 1.79565057e-01 7.26754189e-01 -4.31347072e-01
7.60602176e-01 7.76368260e-01 5.94232455e-02 -4.94286388e-01
-9.45717454e-01 -2.77243704e-01 -2.94548929e-01 -2.09195599e-01
9.01084006e-01 -8.18077147e-01 -5.21021843e-01 5.61212957e-01
-1.11790192e+00 -3.50510269e-01 -5.19696414e-01 5.89054286e-01
-4.70567673e-01 4.21997786e-01 -1.02844489e+00 -2.72353768e-01
-6.11621499e-01 -1.44740260e+00 8.42222333e-01 1.80720180e-01
3.56072448e-02 -8.44546676e-01 8.33517984e-02 2.64159977e-01
4.82814878e-01 4.78067189e-01 9.69566762e-01 -6.77053481e-02
-6.54961169e-01 1.02562025e-01 -2.76316375e-01 4.47668076e-01
4.06805485e-01 -6.31621361e-01 -7.42022574e-01 -2.67004490e-01
5.59352279e-01 -1.02750748e-01 9.34717953e-01 1.02143359e+00
1.45142865e+00 -7.76505917e-02 -1.74090236e-01 1.03633082e+00
1.48186827e+00 4.42750275e-01 8.33310843e-01 3.37837785e-01
7.34943807e-01 -6.68360665e-02 2.68275470e-01 3.12598556e-01
-4.56148433e-03 5.95960319e-01 4.88057077e-01 -6.45690024e-01
-6.78024888e-01 -6.80628419e-02 -6.07186146e-02 7.65161574e-01
-1.81406319e-01 1.47858262e-01 -9.51057494e-01 6.19255781e-01
-1.73328733e+00 -6.08406603e-01 -1.86760038e-01 1.98313439e+00
9.64495659e-01 -1.25043914e-01 -2.16171309e-01 -4.49139588e-02
5.67849278e-01 9.87280160e-02 -6.72485292e-01 2.30387047e-01
5.99380657e-02 6.54097378e-01 5.19108534e-01 4.92877513e-01
-9.50303733e-01 1.92386940e-01 5.43378305e+00 9.24077928e-01
-1.27071118e+00 6.53494298e-01 8.32838356e-01 -4.37454619e-02
-5.65829992e-01 -3.53730053e-01 -3.19516398e-02 3.69811505e-01
3.92194271e-01 3.71302217e-02 2.85923839e-01 4.78975177e-01
3.33158374e-01 -5.62414452e-02 -7.62859762e-01 1.00562596e+00
-6.36661500e-02 -1.57935286e+00 -1.91812962e-01 -9.09699053e-02
7.63533115e-01 1.04736641e-01 2.36677930e-01 -6.23613521e-02
1.07339479e-01 -1.07578897e+00 6.58715069e-01 4.72004920e-01
1.05650342e+00 -6.72881126e-01 6.08225703e-01 -1.63334265e-01
-9.51528609e-01 1.74280614e-01 -1.55705705e-01 5.02797067e-01
3.01718235e-01 1.03968251e+00 -4.54443365e-01 8.57072532e-01
8.59405875e-01 8.70127320e-01 -2.66503066e-01 1.17854714e+00
-3.64351086e-02 3.87183964e-01 -1.93296388e-01 6.49563551e-01
1.29009411e-01 -2.39279687e-01 4.81394380e-01 9.46039617e-01
4.87678528e-01 4.20609444e-01 -3.88685502e-02 1.07327628e+00
-2.12957650e-01 -2.17997089e-01 -2.82294691e-01 5.37735164e-01
7.47136027e-02 1.08411527e+00 -8.11638057e-01 -2.62650609e-01
-3.99613053e-01 9.83915806e-01 -1.81696564e-01 3.53205651e-01
-8.91515851e-01 -1.40067592e-01 5.12220740e-01 6.20243311e-01
6.91409186e-02 3.63593139e-02 -6.52682126e-01 -1.06830788e+00
2.80080587e-01 -7.51665771e-01 3.22957247e-01 -7.84544468e-01
-1.47622406e+00 9.56155717e-01 -3.20403934e-01 -1.63557506e+00
2.36820400e-01 -3.54530334e-01 -5.61070442e-01 9.41794813e-01
-1.65422821e+00 -1.04621053e+00 -4.72972482e-01 7.68856466e-01
4.25314963e-01 3.07852507e-01 4.18266833e-01 7.62742519e-01
-1.90146551e-01 5.26857138e-01 2.01733828e-01 3.17175388e-01
7.88214564e-01 -9.61647749e-01 1.22605391e-01 8.10086608e-01
-4.72885400e-01 5.59405088e-01 4.10904050e-01 -9.38468874e-01
-1.26163304e+00 -1.23116124e+00 1.94014877e-01 -6.08166605e-02
5.97525597e-01 4.91457880e-02 -9.88870859e-01 4.99075413e-01
1.22993425e-01 7.37110376e-01 3.29110116e-01 -4.75299209e-01
-1.03753537e-01 -1.82619795e-01 -1.25573909e+00 4.79179651e-01
9.13465977e-01 -4.01390970e-01 -4.55540776e-01 4.53299105e-01
7.97280788e-01 -9.46900666e-01 -1.14971483e+00 6.81616306e-01
4.03450191e-01 -9.92783844e-01 1.18921506e+00 -2.50508264e-02
8.98191571e-01 -3.74597669e-01 1.53803915e-01 -1.38362944e+00
-5.62888026e-01 -4.68517840e-01 1.29146025e-01 6.11209154e-01
3.81327063e-01 -7.05459654e-01 8.34251583e-01 2.70275265e-01
-9.03730631e-01 -1.06672263e+00 -1.18347287e+00 -5.79442680e-01
3.20764869e-01 -4.38198358e-01 1.89448908e-01 1.01658332e+00
-3.80692989e-01 -2.29842469e-01 -3.19178194e-01 3.38438004e-01
6.70519471e-01 6.55097216e-02 2.08326266e-03 -8.28092098e-01
-6.00187540e-01 -2.89761782e-01 -1.07529588e-01 -8.73388946e-01
-4.70686793e-01 -9.04476523e-01 1.76099032e-01 -1.65641463e+00
3.68506104e-01 -5.81307828e-01 -4.82533753e-01 5.39812028e-01
-1.45670608e-01 6.05283797e-01 9.71381813e-02 3.84882718e-01
2.88438667e-02 6.84415281e-01 2.10380411e+00 -2.48787597e-01
-2.32454658e-01 -6.14728667e-02 -6.04942739e-01 8.56023908e-01
4.96739328e-01 -7.46381164e-01 -1.31700978e-01 -5.97676039e-01
-3.82365519e-03 4.97835279e-01 6.04858220e-01 -1.18637097e+00
1.45755872e-01 3.42484862e-02 7.19579220e-01 -4.58417684e-01
3.84663880e-01 -9.32052851e-01 4.50126112e-01 8.28522801e-01
3.81591283e-02 -5.43349944e-02 1.90302074e-01 2.65259236e-01
-4.23685461e-01 -4.66832742e-02 1.27849686e+00 -3.38556319e-01
4.60583605e-02 5.35276175e-01 -1.86799273e-01 2.12456938e-02
7.81025469e-01 -1.61749244e-01 -2.21096262e-01 -1.16549157e-01
-9.09304440e-01 -1.78908840e-01 3.79291981e-01 9.10321027e-02
9.83157158e-01 -1.41756737e+00 -6.18442178e-01 1.96873382e-01
-2.64442563e-01 5.12764573e-01 8.88784707e-01 1.32785082e+00
-9.27848577e-01 2.15444919e-02 -1.42890707e-01 -7.15068877e-01
-7.06114233e-01 4.28218216e-01 9.16114390e-01 -6.80540025e-01
-1.15700758e+00 7.54146934e-01 7.34308362e-01 -2.03304872e-01
-1.61887303e-01 -5.09020865e-01 3.15317273e-01 -5.57149947e-01
4.56673205e-01 6.93957731e-02 4.81377840e-01 -2.28581026e-01
-3.37754816e-01 7.59108782e-01 -2.05351412e-01 4.40255068e-02
1.72510600e+00 1.90455988e-01 -1.12496011e-01 -3.67136262e-02
1.09327471e+00 -8.48026723e-02 -1.36051071e+00 -3.89133930e-01
-4.46708709e-01 -4.88323033e-01 3.82841498e-01 -8.82738054e-01
-1.72517979e+00 6.67239428e-01 8.24398577e-01 -2.98332602e-01
1.54435408e+00 -3.01688891e-02 1.21846938e+00 -4.18463767e-01
1.56432599e-01 -6.20823860e-01 5.31517029e-01 1.38833508e-01
1.11903191e+00 -9.20349717e-01 2.13511005e-01 -5.97488582e-01
-5.37576616e-01 9.96826947e-01 6.09968543e-01 -4.71078604e-01
8.64215672e-01 5.55203617e-01 9.75144878e-02 -4.28546727e-01
-1.09193422e-01 2.74686545e-01 2.78319389e-01 4.54552680e-01
5.42028546e-01 -9.14473087e-02 -3.78282845e-01 8.08974326e-01
8.96818265e-02 1.67940721e-01 5.52015781e-01 1.00039899e+00
-1.18325010e-01 -9.48360026e-01 -5.51196694e-01 5.71498275e-01
-6.45717442e-01 -2.32480943e-01 2.37126455e-01 6.96913540e-01
2.56366789e-01 5.07937968e-01 -2.66676813e-01 -1.89179406e-01
5.32146454e-01 -5.02775609e-01 5.52452862e-01 -5.86967766e-01
-8.89145434e-01 4.75741416e-01 -3.32672238e-01 -6.84743941e-01
-2.91276038e-01 -5.53393185e-01 -1.52608955e+00 -8.63373131e-02
-1.72674641e-01 -1.94757462e-01 4.69944358e-01 7.13053584e-01
1.18605182e-01 1.18458009e+00 5.86579263e-01 -9.73339856e-01
-3.29913288e-01 -8.10068548e-01 -5.90092659e-01 3.36074114e-01
5.72548091e-01 -6.04745090e-01 -7.42413178e-02 -9.48547125e-02] | [13.528361320495605, -2.5510525703430176] |
a5943c2b-c023-4ee3-9a08-c22d9afbc89c | toxicity-detection-for-indic-multilingual | 2201.00598 | null | https://arxiv.org/abs/2201.00598v1 | https://arxiv.org/pdf/2201.00598v1.pdf | Toxicity Detection for Indic Multilingual Social Media Content | Toxic content is one of the most critical issues for social media platforms today. India alone had 518 million social media users in 2020. In order to provide a good experience to content creators and their audience, it is crucial to flag toxic comments and the users who post that. But the big challenge is identifying toxicity in low resource Indic languages because of the presence of multiple representations of the same text. Moreover, the posts/comments on social media do not adhere to a particular format, grammar or sentence structure; this makes the task of abuse detection even more challenging for multilingual social media platforms. This paper describes the system proposed by team 'Moj Masti' using the data provided by ShareChat/Moj in \emph{IIIT-D Multilingual Abusive Comment Identification} challenge. We focus on how we can leverage multilingual transformer based pre-trained and fine-tuned models to approach code-mixed/code-switched classification tasks. Our best performing system was an ensemble of XLM-RoBERTa and MuRIL which achieved a Mean F-1 score of 0.9 on the test data/leaderboard. We also observed an increase in the performance by adding transliterated data. Furthermore, using weak metadata, ensembling and some post-processing techniques boosted the performance of our system, thereby placing us 1st on the leaderboard. | ['Harveen Singh Chadha', 'Devanshu Ramaiya', 'Manan Jhaveri'] | 2022-01-03 | null | null | null | null | ['abuse-detection'] | ['natural-language-processing'] | [-2.08508387e-01 -1.87313706e-01 5.38612083e-02 -8.75206515e-02
-1.17338598e+00 -7.75208712e-01 2.91584194e-01 4.42959696e-01
-5.48053563e-01 5.41917861e-01 2.49221414e-01 -4.66613561e-01
1.79837927e-01 -2.44518280e-01 -4.52968687e-01 -2.18989283e-01
1.19565070e-01 2.21522674e-01 1.22722335e-01 -5.68838716e-01
3.83884281e-01 4.70013097e-02 -1.01760530e+00 7.09906280e-01
1.17471397e+00 5.81148148e-01 3.79661739e-01 6.63945556e-01
-2.28436545e-01 1.25849533e+00 -7.00308800e-01 -1.08314025e+00
1.10736338e-03 -3.00524354e-01 -7.86916733e-01 -1.91050500e-01
6.00512624e-01 -1.27242982e-01 -2.39258222e-02 1.28678250e+00
4.29698050e-01 -4.85946983e-01 6.11159861e-01 -9.79687393e-01
-6.44165397e-01 1.19393015e+00 -7.74711251e-01 5.79226732e-01
6.46720588e-01 -3.62801924e-02 9.53626692e-01 -7.54254878e-01
8.66742790e-01 9.19262767e-01 8.76840055e-01 2.82784194e-01
-8.26777935e-01 -8.48564506e-01 -1.93827808e-01 2.32783422e-01
-1.16749990e+00 -5.57093978e-01 5.22090554e-01 -9.74941671e-01
1.14528286e+00 1.91000640e-01 2.40246922e-01 1.45395243e+00
1.32090166e-01 6.08327687e-01 1.15482128e+00 -5.16216040e-01
-3.34851831e-01 5.60507596e-01 9.60067883e-02 8.15229356e-01
4.34724381e-03 -7.68170238e-01 -6.12285733e-01 -3.78637761e-01
-1.50269076e-01 -2.25307599e-01 -1.84263393e-01 3.29821408e-01
-6.99200392e-01 1.20675826e+00 3.34029421e-02 7.32790828e-01
-7.02616351e-04 -2.33233467e-01 7.63258219e-01 5.11810780e-01
6.40513182e-01 5.37879467e-01 -3.47241998e-01 -5.50365329e-01
-1.09792948e+00 -4.22126763e-02 8.08171690e-01 7.85285532e-01
3.74318838e-01 -6.33110180e-02 2.05152944e-01 1.38849950e+00
3.19560438e-01 5.50089598e-01 6.36806250e-01 -6.22475326e-01
1.05909634e+00 5.30629456e-01 -2.59413272e-01 -1.11486828e+00
-2.78740048e-01 -5.26448309e-01 -3.40160459e-01 -1.76290333e-01
5.40996730e-01 -2.69914299e-01 -4.72681284e-01 1.41523087e+00
-1.98400572e-01 -5.83344400e-01 -2.47225180e-01 4.12215859e-01
3.66921693e-01 5.15275836e-01 3.81012522e-02 -1.95936278e-01
1.18308413e+00 -7.80942976e-01 -7.04289138e-01 -9.04750675e-02
1.12015676e+00 -1.28856027e+00 1.12385988e+00 5.16467869e-01
-8.59764576e-01 -9.11843479e-02 -1.04487801e+00 -4.04081345e-02
-4.00075793e-01 -9.52505246e-02 1.62590817e-01 1.02284181e+00
-6.66689515e-01 4.42522943e-01 -3.97503823e-01 -5.25810421e-01
4.61826473e-01 5.72593026e-02 -5.31818926e-01 -8.13303664e-02
-9.91250575e-01 1.05669129e+00 -8.11486691e-02 -3.63795727e-01
-6.98265433e-01 -5.99447608e-01 -6.73821568e-01 -4.03711319e-01
9.37244669e-02 1.90722123e-01 1.21118653e+00 -1.09209216e+00
-1.03914356e+00 1.13770330e+00 8.44841525e-02 -3.81691813e-01
7.42748260e-01 -5.00744522e-01 -8.60724688e-01 3.39865126e-03
4.39707279e-01 -1.29148718e-02 8.03276420e-01 -7.74819434e-01
-6.26045942e-01 -3.42314303e-01 2.21530143e-02 -2.38332540e-01
-7.99842834e-01 9.84288931e-01 -6.78931922e-02 -6.20607793e-01
-3.66804004e-01 -1.07561731e+00 3.11211020e-01 -5.45717776e-01
-5.89936376e-01 -6.80644363e-02 9.38016176e-01 -1.40874255e+00
1.60902286e+00 -2.30835581e+00 1.31423175e-01 7.21599236e-02
1.56226367e-01 2.96519727e-01 6.65550530e-02 7.23280728e-01
2.22648424e-03 5.70047736e-01 -4.43020284e-01 -6.86580718e-01
-9.23546329e-02 -3.28638852e-02 -3.44954997e-01 6.38942599e-01
-9.33402590e-03 3.37231368e-01 -9.25724804e-01 -4.41051841e-01
-3.10654670e-01 2.58018285e-01 -7.25944638e-01 2.13087946e-01
5.10884598e-02 4.38445687e-01 1.63922440e-02 5.76389849e-01
4.84919608e-01 1.44272983e-01 3.93714346e-02 3.28728169e-01
-4.70711797e-01 5.22020102e-01 -6.96983516e-01 1.66937673e+00
-7.70332634e-01 1.02974999e+00 1.49476439e-01 -2.62454480e-01
6.56768084e-01 3.15203309e-01 2.46304795e-01 -7.99201667e-01
2.39547014e-01 6.79625571e-01 2.72564292e-01 -1.01443899e+00
6.35160863e-01 -1.99919976e-02 -4.70119566e-01 5.65902472e-01
8.99631158e-02 1.43067027e-02 3.47404867e-01 5.00403523e-01
1.22291207e+00 -1.95287302e-01 1.56773701e-01 -2.57884681e-01
5.53465784e-01 -1.11452080e-01 3.86340588e-01 5.42239428e-01
-1.33539155e-01 7.20039308e-01 7.35915363e-01 5.00912927e-02
-1.29036522e+00 -5.89194238e-01 -3.03413868e-01 1.31291711e+00
-6.93086326e-01 -6.40717030e-01 -9.23724711e-01 -9.96810555e-01
-2.02527329e-01 8.34031999e-01 -3.05440575e-01 2.33262092e-01
-6.23807192e-01 -6.69943929e-01 1.01032662e+00 -7.60827810e-02
3.03933144e-01 -7.95920253e-01 -1.46647066e-01 2.12011024e-01
-4.64841723e-01 -1.22861505e+00 -4.37311143e-01 1.45864829e-01
-1.79126859e-01 -1.09158909e+00 -4.22842890e-01 -5.37201703e-01
2.98557431e-01 -1.37151971e-01 6.62727296e-01 1.99327841e-01
-8.08673874e-02 -1.92003250e-01 -8.35592270e-01 -3.76910269e-01
-9.29948568e-01 5.29325843e-01 4.87705208e-02 -3.75393312e-03
2.35154301e-01 -5.37644625e-01 1.01928040e-01 6.15707831e-03
-8.15331638e-01 -1.96502224e-01 2.39841715e-01 3.47980917e-01
-2.98952848e-01 -2.59556234e-01 5.50870061e-01 -1.39978659e+00
5.64000189e-01 -1.04786301e+00 -2.04409927e-01 -1.11432403e-01
-2.61835933e-01 -2.93137461e-01 8.83336365e-01 -3.27289075e-01
-8.69803190e-01 -3.53692323e-01 -6.47507131e-01 -1.99965928e-02
1.38691261e-01 6.02919519e-01 -1.33250803e-01 1.08502686e-01
8.20012510e-01 -2.82359898e-01 -2.84349978e-01 -8.69335592e-01
-7.36128865e-03 1.35626304e+00 1.73211545e-01 -2.51834989e-01
7.99320579e-01 1.05390675e-01 -7.56735504e-01 -1.05949056e+00
-8.94291520e-01 -5.80948234e-01 -6.32065415e-01 -2.01763436e-01
9.29799497e-01 -8.24495077e-01 -3.02569896e-01 9.12708819e-01
-1.35081315e+00 -2.09482923e-01 2.87321836e-01 1.28169224e-01
-7.98945129e-03 5.67067564e-01 -9.97861803e-01 -7.21724331e-01
-2.05874383e-01 -1.20025301e+00 5.33580720e-01 -4.04159814e-01
-5.83220363e-01 -8.32552552e-01 3.33840281e-01 8.67860198e-01
5.54191649e-01 1.93457782e-01 1.12568069e+00 -1.05572605e+00
8.87009725e-02 -1.29840747e-01 -9.83186439e-02 7.94661701e-01
-1.60317823e-01 2.67406642e-01 -1.08331919e+00 -3.33648145e-01
3.59170549e-02 -5.47982514e-01 5.35944283e-01 -3.66685510e-01
7.32895017e-01 -6.11133218e-01 7.14087039e-02 2.79173583e-01
1.26080239e+00 -7.55016357e-02 5.45796990e-01 4.42956299e-01
9.72026110e-01 5.06699562e-01 1.50995925e-01 5.02313495e-01
4.13451403e-01 6.63175285e-01 3.45914274e-01 3.58225077e-01
-1.36143953e-01 -3.55390757e-01 9.99812961e-01 1.56954515e+00
2.90636450e-01 -2.17073515e-01 -1.23366249e+00 5.85122287e-01
-1.51046014e+00 -9.84291792e-01 -7.34570920e-01 1.92107248e+00
9.15543258e-01 2.03453586e-01 2.19855219e-01 3.71504456e-01
5.96025288e-01 -5.18910848e-02 4.94998544e-02 -8.99784744e-01
-2.28249535e-01 7.86467344e-02 4.40385669e-01 5.90140164e-01
-9.62021410e-01 7.05976605e-01 5.34273386e+00 1.00303578e+00
-1.17973113e+00 8.46657813e-01 3.29221606e-01 -9.73204598e-02
-1.61030829e-01 -1.08399048e-01 -8.98907125e-01 9.60334063e-01
1.41729784e+00 1.10132866e-01 4.02273357e-01 7.47553289e-01
8.45842138e-02 -2.07022712e-01 -8.18156779e-01 8.18447292e-01
6.20044827e-01 -8.45102310e-01 -3.56293052e-01 1.89695969e-01
8.68499100e-01 6.02892756e-01 1.35501251e-01 4.41555768e-01
1.92070529e-01 -1.05849361e+00 1.14341748e+00 1.71140894e-01
6.11892164e-01 -6.98569298e-01 8.44629109e-01 6.19006276e-01
-5.06009519e-01 -3.86941016e-01 -1.72058597e-01 -2.24199355e-01
8.88328999e-02 7.28224635e-01 -8.03175688e-01 2.88699925e-01
7.74043798e-01 8.08654368e-01 -1.03811121e+00 9.84110415e-01
-3.36881578e-01 8.39236140e-01 -2.44894892e-01 -1.78514584e-03
1.84608862e-01 5.00565618e-02 7.37441778e-01 1.75643289e+00
3.99627030e-01 -5.77264547e-01 -5.61624132e-02 3.69430363e-01
-3.76834840e-01 5.15692711e-01 -6.63949311e-01 -2.17598125e-01
3.10833901e-01 1.37082052e+00 -3.51089150e-01 9.29719880e-02
-6.99047089e-01 1.01391268e+00 7.39077210e-01 -1.53887048e-01
-8.61470878e-01 -3.32200319e-01 3.97396594e-01 5.66040397e-01
9.10856947e-02 -2.15608940e-01 1.24213293e-01 -1.30145395e+00
9.64839710e-04 -1.33843243e+00 4.92684275e-01 -5.69755256e-01
-1.44224191e+00 8.48707318e-01 -3.18009287e-01 -9.92777646e-01
-2.48428345e-01 -5.96730411e-01 -2.14866936e-01 6.68493569e-01
-1.22531104e+00 -1.21938181e+00 2.03368813e-01 4.08866763e-01
5.31049669e-01 -3.50409299e-01 7.20009506e-01 9.98071790e-01
-7.68177629e-01 8.18773806e-01 2.00852260e-01 4.93345290e-01
9.11951959e-01 -1.02880418e+00 -3.79038006e-02 1.08604980e+00
1.87628463e-01 5.81335545e-01 7.02355444e-01 -6.95629537e-01
-9.26000774e-01 -1.04499280e+00 1.35746074e+00 -8.84194076e-01
1.28904450e+00 -6.81915045e-01 -8.73116672e-01 7.17712522e-01
5.76230347e-01 -5.93646348e-01 1.02143419e+00 1.49462089e-01
-7.08397031e-01 2.50628412e-01 -1.12670660e+00 3.31553549e-01
9.03025925e-01 -1.03872192e+00 -4.36171770e-01 5.16613543e-01
4.24738050e-01 -2.26937905e-01 -8.05118501e-01 -1.39743507e-01
2.44895875e-01 -1.14621317e+00 1.64470702e-01 -4.98210341e-01
8.47885251e-01 -1.48013476e-02 -4.28812087e-01 -1.29085386e+00
-1.36569990e-02 -6.54704154e-01 2.69688368e-01 1.87777781e+00
8.15798879e-01 -3.73392671e-01 7.03050643e-02 4.12632942e-01
-2.59827852e-01 -2.51661658e-01 -1.02726722e+00 -5.59147239e-01
3.48477960e-01 -7.83116460e-01 4.60043773e-02 1.33625722e+00
3.65808725e-01 4.07051265e-01 -3.97016168e-01 -8.81012678e-02
3.21082860e-01 -4.48247164e-01 6.05246723e-01 -8.47861707e-01
-3.73514920e-01 -3.33498210e-01 -3.20937932e-01 -3.00837189e-01
2.62261152e-01 -1.38430989e+00 -3.48448455e-01 -9.88189280e-01
6.34940445e-01 -5.27063727e-01 1.58016086e-01 4.74430650e-01
1.54342994e-01 6.29713178e-01 4.63392586e-01 3.26867431e-01
-5.28930247e-01 -1.45218804e-01 5.72501540e-01 -6.95014969e-02
1.94987625e-01 -2.66941249e-01 -7.55131900e-01 7.98676550e-01
7.03752816e-01 -9.27990496e-01 1.39993951e-01 -6.12087011e-01
1.00755060e+00 -2.64779001e-01 -4.22541648e-02 -1.02588546e+00
-3.17567284e-03 3.23407501e-01 -4.43592481e-02 -2.10150585e-01
8.72832164e-02 -6.03417397e-01 8.84402096e-02 4.47182953e-01
-2.47885361e-01 3.30611497e-01 7.23195001e-02 -6.19850643e-02
-3.08118649e-02 -8.18768263e-01 8.82804573e-01 -7.49970376e-02
-1.89735800e-01 -1.54930174e-01 -9.27593768e-01 4.13824499e-01
8.12615275e-01 2.40425542e-01 -6.07298315e-01 -3.13269347e-01
-4.46352303e-01 -1.65079072e-01 7.94470370e-01 5.73691964e-01
8.84345248e-02 -8.56352627e-01 -1.00203979e+00 -1.49969570e-02
3.24799508e-01 -6.63646460e-01 7.01788887e-02 1.05265737e+00
-6.17792010e-01 2.18553748e-02 -2.45976016e-01 -3.30136210e-01
-1.36990535e+00 1.98687345e-01 2.37208307e-01 -3.49000335e-01
-1.05371848e-01 8.40381086e-01 -5.02946913e-01 -5.30440927e-01
-1.74786046e-01 1.19926840e-01 -3.85920554e-01 5.54373980e-01
5.95470667e-01 5.00707150e-01 4.40224349e-01 -1.07085812e+00
-3.33143771e-01 1.80622727e-01 -3.69630396e-01 -2.11150333e-01
1.52837241e+00 -2.80167628e-02 -5.37940621e-01 6.99786425e-01
1.58813465e+00 9.78797078e-01 -4.51813132e-01 1.15224548e-01
2.33263522e-01 -5.50743043e-01 -1.03043199e-01 -8.65215123e-01
-9.17049706e-01 9.13149655e-01 2.88550884e-01 3.16855520e-01
5.29569089e-01 5.20941354e-02 9.46814895e-01 -1.17254034e-02
3.70391309e-01 -1.20135427e+00 2.36909524e-01 8.80572200e-01
9.88387048e-01 -1.08248711e+00 -1.02447405e-01 -2.75008947e-01
-6.66788280e-01 9.20159400e-01 3.69952649e-01 1.31347433e-01
5.13825238e-01 4.36179370e-01 3.52755964e-01 -8.15246105e-02
-2.78133869e-01 6.01868518e-03 -8.72602388e-02 3.03387880e-01
7.87300169e-01 -2.45246124e-02 -4.45657670e-01 6.97124124e-01
-5.77477694e-01 -5.88788807e-01 9.82513428e-01 7.14400172e-01
-2.62810349e-01 -1.29090869e+00 -3.15164864e-01 5.91070950e-01
-1.28307950e+00 -3.94518465e-01 -5.11608601e-01 6.15614176e-01
3.34924608e-01 1.15940726e+00 -8.62113908e-02 -5.04087627e-01
4.27316539e-02 3.35352451e-01 2.61588186e-01 -9.87634659e-01
-1.31753194e+00 -9.35237259e-02 5.37396789e-01 -2.54163802e-01
-1.12023741e-01 -1.15663922e+00 -1.01842248e+00 -7.06497192e-01
-3.98211658e-01 1.46787852e-01 9.23898578e-01 8.84764433e-01
1.66898817e-01 1.70075044e-01 6.27879918e-01 -2.85876095e-01
-3.04393947e-01 -1.32132304e+00 -3.22980165e-01 7.93492615e-01
2.11254254e-01 -2.63841093e-01 -5.14763832e-01 6.89996853e-02] | [8.975309371948242, 10.588798522949219] |
5e61da09-0368-4a70-a033-a1110c16085a | melm-data-augmentation-with-masked-entity | 2108.13655 | null | https://arxiv.org/abs/2108.13655v2 | https://arxiv.org/pdf/2108.13655v2.pdf | MELM: Data Augmentation with Masked Entity Language Modeling for Low-Resource NER | Data augmentation is an effective solution to data scarcity in low-resource scenarios. However, when applied to token-level tasks such as NER, data augmentation methods often suffer from token-label misalignment, which leads to unsatsifactory performance. In this work, we propose Masked Entity Language Modeling (MELM) as a novel data augmentation framework for low-resource NER. To alleviate the token-label misalignment issue, we explicitly inject NER labels into sentence context, and thus the fine-tuned MELM is able to predict masked entity tokens by explicitly conditioning on their labels. Thereby, MELM generates high-quality augmented data with novel entities, which provides rich entity regularity knowledge and boosts NER performance. When training data from multiple languages are available, we also integrate MELM with code-mixing for further improvement. We demonstrate the effectiveness of MELM on monolingual, cross-lingual and multilingual NER across various low-resource levels. Experimental results show that our MELM presents substantial improvement over the baseline methods. | ['Xin Li', 'Chunyan Miao', 'Luo Si', 'Erik Cambria', 'Lidong Bing', 'Ruidan He', 'Ran Zhou'] | 2021-08-31 | null | https://aclanthology.org/2022.acl-long.160 | https://aclanthology.org/2022.acl-long.160.pdf | acl-2022-5 | ['cross-lingual-ner'] | ['natural-language-processing'] | [-1.34768635e-01 -6.91947639e-02 -3.14856648e-01 -3.85832518e-01
-1.12044287e+00 -5.21638453e-01 3.71902674e-01 4.06364381e-01
-8.97609532e-01 8.74345541e-01 6.15753055e-01 -2.90957868e-01
6.23313665e-01 -5.66675723e-01 -6.24094963e-01 -3.10023248e-01
2.12356791e-01 1.20275654e-01 -4.55539733e-01 -2.09934399e-01
-1.93233654e-01 5.69762439e-02 -1.15263200e+00 2.21909985e-01
1.45157433e+00 3.30905259e-01 2.68825710e-01 2.50820637e-01
-5.11938691e-01 8.56789529e-01 -5.68882644e-01 -6.38337493e-01
2.24611700e-01 -7.10262880e-02 -8.50792766e-01 -2.05390215e-01
4.13281053e-01 -6.80128753e-04 -1.21776342e-01 1.05503249e+00
6.21786535e-01 1.40991613e-01 4.55588400e-01 -1.08197379e+00
-9.02704775e-01 1.14621472e+00 -6.29599750e-01 -8.86860490e-03
-5.55895008e-02 -8.32962431e-03 1.05419588e+00 -1.54901767e+00
4.09256101e-01 1.03744364e+00 9.92956161e-01 6.85739577e-01
-1.27031898e+00 -7.57315099e-01 3.29768598e-01 -8.16293135e-02
-1.46494961e+00 -6.83563828e-01 4.32481885e-01 -1.96513891e-01
1.01789200e+00 1.15444146e-01 -2.05293775e-01 9.52096045e-01
-5.26743114e-01 1.01414359e+00 1.08955586e+00 -7.88421929e-01
-1.17307298e-01 2.25137338e-01 1.70955867e-01 4.15943354e-01
1.09285288e-01 -2.56512821e-01 -5.65078616e-01 -1.03663474e-01
4.09436643e-01 -4.19926383e-02 5.50452759e-03 3.98572326e-01
-1.44582140e+00 5.71036875e-01 3.47214043e-01 4.74120736e-01
-5.09366453e-01 -1.96347743e-01 5.82340121e-01 1.25795543e-01
7.23347545e-01 7.59424329e-01 -9.12667513e-01 -1.89279273e-01
-9.31022167e-01 -1.18374832e-01 4.25775647e-01 1.08608675e+00
8.16769779e-01 3.14258486e-01 -5.12586594e-01 1.26834798e+00
1.24014772e-01 5.08279800e-01 6.88407063e-01 -5.16095340e-01
9.61594999e-01 6.91208184e-01 1.23873219e-01 -5.40314734e-01
-4.70690608e-01 -4.49304491e-01 -1.17689228e+00 -2.18684420e-01
1.38904974e-01 -5.88323772e-01 -9.71892297e-01 2.14864492e+00
3.80404145e-01 2.13318393e-01 4.31290925e-01 6.51060164e-01
8.92754972e-01 6.00886643e-01 5.48687041e-01 -1.99158955e-02
1.52767158e+00 -1.22532523e+00 -1.03508484e+00 -4.20527548e-01
1.14689362e+00 -8.13712239e-01 1.56164098e+00 -3.37593704e-02
-8.03473532e-01 -6.10409558e-01 -6.92684829e-01 -2.54057407e-01
-4.78843659e-01 5.99467337e-01 5.12705922e-01 5.96075833e-01
-6.41354203e-01 2.91058391e-01 -7.82518923e-01 -2.50048339e-02
2.73880303e-01 1.32035315e-01 -5.47304749e-01 -2.98955500e-01
-1.48680079e+00 9.35700953e-01 5.76334059e-01 3.17968607e-01
-4.30905342e-01 -9.72519875e-01 -1.27708137e+00 2.91535296e-02
2.18482673e-01 -1.62992731e-01 1.23198664e+00 -3.03423703e-01
-1.00172520e+00 7.50463307e-01 -3.87220383e-01 -3.28941405e-01
2.35601604e-01 -4.46794450e-01 -7.50712216e-01 -5.84819317e-01
4.39977884e-01 7.28092492e-01 2.20013782e-01 -1.14622319e+00
-5.93944907e-01 1.03671879e-01 -3.82260419e-02 3.73865247e-01
-8.44527662e-01 1.86043933e-01 -3.18480253e-01 -1.03867137e+00
-2.51417339e-01 -6.05681062e-01 -6.26147449e-01 -7.80670047e-01
-5.46417236e-01 -3.26872528e-01 3.10431868e-01 -8.14275265e-01
1.52268267e+00 -2.28196764e+00 -3.63593429e-01 -2.36513227e-01
2.62148511e-02 5.09664953e-01 -4.44756836e-01 3.99764955e-01
-1.46456957e-01 5.30365288e-01 -4.12395447e-01 -1.02546787e+00
1.42031550e-01 3.53717059e-01 -2.23542914e-01 -2.74857618e-02
6.29373312e-01 9.54172909e-01 -9.34824049e-01 -6.38690591e-01
-2.71106753e-02 4.15216446e-01 -5.46888888e-01 3.30833018e-01
-4.90535051e-02 6.52034163e-01 -1.12550020e-01 6.99049234e-01
7.55173266e-01 -1.27032191e-01 2.50090030e-03 -1.57022342e-01
-4.01784033e-01 6.66263163e-01 -1.24486315e+00 1.79387689e+00
-9.88780320e-01 3.40738535e-01 -8.60509425e-02 -4.00126427e-01
8.17279160e-01 5.78091621e-01 2.29903832e-01 -6.02168918e-01
-1.13460839e-01 3.56525153e-01 -2.09445938e-01 -4.13148999e-01
1.08004820e+00 -2.12054446e-01 -4.60853010e-01 3.14581454e-01
9.80481282e-02 5.29404700e-01 3.82037520e-01 1.55091643e-01
9.59482849e-01 -3.13543193e-02 1.12264492e-01 -5.33338450e-02
4.72035736e-01 -4.27799299e-02 1.03369057e+00 6.33547306e-01
9.90412757e-03 4.76475626e-01 -3.07530053e-02 -1.04383938e-01
-1.18652260e+00 -6.32312655e-01 -2.83284694e-01 1.24950814e+00
-3.09281319e-01 -6.56204522e-01 -7.35849798e-01 -8.93702388e-01
-2.19089359e-01 5.63831568e-01 -2.85282165e-01 3.32796276e-02
-6.97076321e-01 -1.30314124e+00 9.10570502e-01 6.46453977e-01
5.81089795e-01 -1.17284882e+00 3.39351475e-01 4.22869831e-01
-7.49699235e-01 -1.41005659e+00 -7.62045622e-01 3.60209674e-01
-5.83255351e-01 -5.00899732e-01 -4.94382620e-01 -1.00600636e+00
8.68346274e-01 8.61719325e-02 1.31434357e+00 1.12663731e-01
-5.72204851e-02 -3.00674647e-01 -5.88985264e-01 -2.05160618e-01
-5.11533201e-01 5.60942590e-01 4.38378602e-01 -9.53780189e-02
6.52804017e-01 -4.31851387e-01 -3.55250478e-01 1.19956724e-01
-8.41709435e-01 -3.33257159e-03 8.22775006e-01 1.03916836e+00
6.82144105e-01 -2.89958958e-02 1.00324464e+00 -1.28671575e+00
4.49959934e-01 -5.84172487e-01 -2.81838924e-01 3.57482255e-01
-5.89363515e-01 2.53432274e-01 9.56276417e-01 -5.38384318e-01
-1.26413202e+00 8.02714378e-02 -4.99688119e-01 1.11911912e-02
-2.49532953e-01 8.37668598e-01 -6.15727305e-01 3.06581676e-01
6.55592501e-01 -3.82508188e-02 -7.05500662e-01 -1.09884202e+00
7.12971985e-01 9.20437753e-01 8.09287131e-01 -9.28104758e-01
8.88591766e-01 8.64666421e-03 -5.12945592e-01 -5.73432803e-01
-1.13073218e+00 -5.08965552e-01 -7.48215258e-01 2.49587923e-01
6.77690387e-01 -1.42400742e+00 -1.78786635e-01 4.90194798e-01
-1.20142663e+00 -2.40350828e-01 -4.82471645e-01 7.00339794e-01
1.54498830e-01 3.28252375e-01 -9.48977709e-01 -7.28730381e-01
-4.85626489e-01 -8.61763835e-01 8.22887719e-01 2.57919341e-01
-3.66586968e-02 -9.44811523e-01 1.04545787e-01 3.10231686e-01
3.29988271e-01 1.23896495e-01 7.81859934e-01 -9.15424943e-01
-1.76585659e-01 -2.84983933e-01 -2.24750996e-01 4.49012935e-01
3.19981068e-01 -2.55879670e-01 -1.11402905e+00 -3.02953541e-01
-2.34086454e-01 -5.25085688e-01 6.39278114e-01 -1.10013500e-01
9.07173514e-01 -5.16912043e-01 3.72290122e-03 5.59182882e-01
1.27589130e+00 -3.77726018e-01 2.93746889e-01 3.89888257e-01
1.15196145e+00 5.65634906e-01 8.58414829e-01 4.62339342e-01
8.64288330e-01 3.97962958e-01 8.83239806e-02 -6.31395280e-01
-2.10958838e-01 -4.94827598e-01 3.92257810e-01 1.59567177e+00
2.24952444e-01 -5.36887422e-02 -1.06268895e+00 8.87234509e-01
-1.71222591e+00 -5.83630919e-01 -4.64644343e-01 2.04381347e+00
1.56005502e+00 -5.02661839e-02 -5.47852255e-02 -1.32094875e-01
8.91438961e-01 -8.09660554e-02 -3.93144041e-01 -1.41745299e-01
-4.23255146e-01 2.41316989e-01 6.63913608e-01 4.43328112e-01
-1.36450458e+00 1.20836079e+00 5.69963741e+00 1.03292608e+00
-7.40302026e-01 5.19097745e-01 4.10159349e-01 2.27070764e-01
-3.42101455e-01 1.14861811e-02 -1.24439299e+00 6.45185828e-01
1.02794147e+00 -1.29340753e-01 2.17576802e-01 7.89200485e-01
1.46960378e-01 3.07121575e-01 -1.00244093e+00 8.40776801e-01
-2.07532614e-01 -9.87937868e-01 -1.95865065e-01 3.43992487e-02
1.02966666e+00 3.84819537e-01 -5.36397584e-02 9.44764555e-01
8.49391758e-01 -1.00901568e+00 5.29856861e-01 8.48303139e-02
9.96662259e-01 -9.01800931e-01 1.03397000e+00 5.88421464e-01
-1.40156627e+00 2.52239943e-01 -4.34072703e-01 5.84118515e-02
3.39304119e-01 8.34665656e-01 -8.87223482e-01 7.04431713e-01
4.63768184e-01 5.88620782e-01 -6.81927204e-01 8.66213739e-01
-4.15819407e-01 7.18229711e-01 -3.13773811e-01 3.83053243e-01
1.12714134e-01 1.26840442e-01 1.87346414e-01 1.70129669e+00
4.98259775e-02 -1.69668138e-01 5.01016736e-01 6.48722172e-01
-7.58368134e-01 4.84168977e-01 -2.75184900e-01 -1.14668325e-01
1.04933810e+00 1.66336620e+00 -3.04626822e-01 -3.35819930e-01
-6.09148085e-01 9.67206895e-01 8.04865599e-01 2.22375691e-01
-6.26419902e-01 -4.45599675e-01 8.23001742e-01 -3.87669742e-01
-1.05296999e-01 -2.77997911e-01 -3.90170932e-01 -1.70391846e+00
1.90677553e-01 -8.95103812e-01 4.36429888e-01 -3.12015653e-01
-1.70905542e+00 7.43572891e-01 -4.98781174e-01 -1.37009275e+00
-1.88297242e-01 -3.02295238e-01 -3.74210596e-01 1.19734359e+00
-1.98295271e+00 -1.43135405e+00 2.01469213e-02 3.96627516e-01
4.46858048e-01 -8.32125247e-02 1.06308496e+00 1.06811607e+00
-8.71986747e-01 1.12031233e+00 2.35608235e-01 7.46815562e-01
1.16506481e+00 -1.52074099e+00 8.89489651e-01 1.27320600e+00
4.18475986e-01 7.04448819e-01 3.17152083e-01 -7.17617631e-01
-7.90528476e-01 -1.68153715e+00 1.46124828e+00 -6.20516479e-01
5.11857331e-01 -6.62528157e-01 -1.18759632e+00 6.67091429e-01
1.92700699e-01 1.83007598e-01 1.13052070e+00 5.64096153e-01
-4.80075270e-01 1.61800325e-01 -1.00248170e+00 8.18297744e-01
6.99193120e-01 -7.82234550e-01 -5.83743095e-01 3.89128447e-01
1.02140296e+00 -4.47898686e-01 -1.17949092e+00 4.89567369e-01
-8.17907378e-02 -1.78252727e-01 6.80142581e-01 -9.03733134e-01
2.74334997e-01 -4.87365693e-01 -2.60183245e-01 -1.54833448e+00
-1.93718627e-01 -4.73413736e-01 -3.95880789e-02 2.21157074e+00
8.32838833e-01 -2.79495120e-01 5.50293624e-01 8.73319685e-01
-3.70905042e-01 -4.28050250e-01 -7.17471361e-01 -9.46035981e-01
3.15364718e-01 -4.35950100e-01 8.69921327e-01 1.62841034e+00
1.36238769e-01 4.87169504e-01 -7.56507099e-01 4.80340719e-01
5.27196646e-01 -2.86928415e-01 5.85690677e-01 -9.25400734e-01
-5.84539324e-02 -4.23603505e-02 1.01668626e-01 -9.40114856e-01
4.63870496e-01 -1.26424837e+00 3.83176655e-01 -1.36016357e+00
1.96642712e-01 -9.04680312e-01 -5.74085891e-01 9.95712817e-01
-8.31646144e-01 3.54719549e-01 1.31230354e-01 2.20872119e-01
-6.00480497e-01 6.93345249e-01 6.79444373e-01 1.42784432e-01
-2.31516227e-01 -3.42491031e-01 -7.73537040e-01 6.94055557e-01
8.92379761e-01 -7.90678620e-01 2.01090544e-01 -8.34476292e-01
2.39466429e-01 -4.37655389e-01 -1.61719948e-01 -8.03019106e-01
2.24690497e-01 4.28801589e-02 2.49624833e-01 -5.68341017e-01
-2.66979784e-02 -7.30887234e-01 -2.70908535e-01 1.44581925e-02
-4.59598571e-01 2.49054059e-01 1.94059029e-01 1.87007234e-01
-4.20395494e-01 -4.26100165e-01 5.28921723e-01 -6.85074031e-02
-6.60077095e-01 4.72434789e-01 -1.94670752e-01 6.32229269e-01
3.39495003e-01 3.69360119e-01 -3.81462127e-01 2.03506172e-01
-6.20362639e-01 6.51068807e-01 4.62955207e-01 5.39315403e-01
1.77517086e-01 -1.83140802e+00 -1.08597159e+00 1.55097038e-01
3.84860426e-01 3.46799850e-01 3.40248287e-01 7.19796419e-01
7.69800320e-02 7.21449405e-02 2.65408933e-01 -2.85595417e-01
-1.05527341e+00 5.83193123e-01 4.60913181e-02 -6.46936834e-01
-3.29290420e-01 8.35131347e-01 -1.38011072e-02 -1.22408354e+00
1.60464764e-01 -2.04608455e-01 -3.19220692e-01 1.11443073e-01
5.22135258e-01 1.82114422e-01 2.96905339e-01 -7.17132270e-01
-3.34415406e-01 2.14114442e-01 -2.86376357e-01 -1.07492559e-01
1.39755583e+00 -3.36462557e-01 -1.99145541e-01 4.55663055e-01
9.86330688e-01 6.31206512e-01 -1.08812344e+00 -8.12745452e-01
4.26263660e-01 -2.63347954e-01 -1.78941846e-01 -8.14206779e-01
-8.11787188e-01 8.44263136e-01 1.80347681e-01 -1.85498044e-01
9.72324848e-01 -3.05886269e-01 9.31427300e-01 3.69872391e-01
2.01478541e-01 -1.19983101e+00 -2.41760090e-01 8.54149640e-01
3.09147358e-01 -1.58813751e+00 -3.78221810e-01 -3.63984436e-01
-8.24571371e-01 6.08644545e-01 9.79893565e-01 2.80876487e-01
4.10400033e-01 4.93461132e-01 3.33038986e-01 3.00436080e-01
-5.78132689e-01 -4.28408504e-01 1.92960083e-01 4.67935115e-01
8.66854489e-01 7.16253743e-02 -2.44685873e-01 9.78005826e-01
-2.14216739e-01 -3.64776313e-01 4.53460723e-01 8.48219872e-01
-1.14753202e-01 -1.65021431e+00 -3.21448624e-01 2.07421914e-01
-8.71743917e-01 -9.56149876e-01 1.12814428e-02 5.99013448e-01
3.44158858e-01 9.58533823e-01 -1.03650704e-01 -2.07241833e-01
5.06987929e-01 3.73001099e-01 -4.73891012e-02 -9.10137057e-01
-8.73470902e-01 -1.46002963e-01 3.01704615e-01 1.16607115e-01
-3.80143464e-01 -5.80750823e-01 -1.46218443e+00 -1.86797217e-01
-4.91602182e-01 6.08602405e-01 6.72187090e-01 8.58850718e-01
5.36032617e-01 5.14237225e-01 8.01627338e-01 -4.14170653e-01
-4.71358210e-01 -1.32450449e+00 -3.02587748e-01 5.82188129e-01
4.31647271e-01 -2.73563206e-01 -2.08576933e-01 1.67878002e-01] | [9.870141983032227, 9.600290298461914] |
e397e449-5f25-4e84-ab5c-a8d5c29994e5 | dort-modeling-dynamic-objects-in-recurrent | 2303.16628 | null | https://arxiv.org/abs/2303.16628v2 | https://arxiv.org/pdf/2303.16628v2.pdf | DORT: Modeling Dynamic Objects in Recurrent for Multi-Camera 3D Object Detection and Tracking | Recent multi-camera 3D object detectors usually leverage temporal information to construct multi-view stereo that alleviates the ill-posed depth estimation. However, they typically assume all the objects are static and directly aggregate features across frames. This work begins with a theoretical and empirical analysis to reveal that ignoring the motion of moving objects can result in serious localization bias. Therefore, we propose to model Dynamic Objects in RecurrenT (DORT) to tackle this problem. In contrast to previous global Bird-Eye-View (BEV) methods, DORT extracts object-wise local volumes for motion estimation that also alleviates the heavy computational burden. By iteratively refining the estimated object motion and location, the preceding features can be precisely aggregated to the current frame to mitigate the aforementioned adverse effects. The simple framework has two significant appealing properties. It is flexible and practical that can be plugged into most camera-based 3D object detectors. As there are predictions of object motion in the loop, it can easily track objects across frames according to their nearest center distances. Without bells and whistles, DORT outperforms all the previous methods on the nuScenes detection and tracking benchmarks with 62.5\% NDS and 57.6\% AMOTA, respectively. The source code will be released. | ['Jiangmiao Pang', 'Dahua Lin', 'Tai Wang', 'Qing Lian'] | 2023-03-29 | null | null | null | null | ['motion-estimation'] | ['computer-vision'] | [-1.66698948e-01 -5.08992910e-01 -1.46770418e-01 -1.23243175e-01
-6.24281228e-01 -8.41070116e-01 5.09574652e-01 -2.61698544e-01
-3.37555677e-01 3.64519566e-01 5.24319634e-02 4.06870767e-02
7.50834271e-02 -5.86168587e-01 -6.18127584e-01 -8.74121487e-01
2.62204230e-01 -2.97840256e-02 9.77054119e-01 2.66898554e-02
4.16046858e-01 4.97694850e-01 -1.65309048e+00 1.07494399e-01
5.66640913e-01 1.15358984e+00 5.41151702e-01 7.60351598e-01
2.64286220e-01 8.16692710e-01 -3.34834933e-01 -2.08235577e-01
4.37480927e-01 6.70450088e-03 -1.88296631e-01 2.36292869e-01
8.18270147e-01 -8.68687093e-01 -4.04377550e-01 1.02456582e+00
5.55697978e-01 1.61922678e-01 4.75157440e-01 -1.12979150e+00
-2.65727252e-01 -1.38808385e-01 -1.18772781e+00 5.47659218e-01
3.96286130e-01 1.80184498e-01 8.37721884e-01 -1.34295428e+00
7.03712940e-01 1.33217871e+00 6.60693109e-01 4.59339410e-01
-9.38525736e-01 -6.28135502e-01 5.83391964e-01 2.77175512e-02
-1.40374446e+00 -5.44867814e-01 6.57667100e-01 -5.48305333e-01
8.42179716e-01 2.49036342e-01 7.57094085e-01 8.40504587e-01
2.64200151e-01 8.99923205e-01 6.65913880e-01 -1.21382453e-01
-7.73817971e-02 -1.85396392e-02 -8.84395372e-03 6.96561337e-01
4.38417584e-01 3.05013120e-01 -6.60034895e-01 -2.38648243e-02
9.63149250e-01 2.38547519e-01 -2.84883022e-01 -7.34632254e-01
-1.49542832e+00 7.64211416e-01 3.02332163e-01 -8.17228407e-02
-2.00381428e-01 3.13635439e-01 2.81611443e-01 -3.00528318e-01
6.33171380e-01 -7.75164366e-02 -2.35558584e-01 -1.99632049e-02
-7.43463278e-01 4.11490321e-01 1.81425318e-01 1.17510164e+00
6.49326384e-01 -5.81321418e-02 -4.13684063e-02 5.39524615e-01
3.68153155e-01 6.32668734e-01 1.13757625e-01 -1.16439116e+00
5.16874790e-01 5.77362299e-01 4.61200356e-01 -1.23397553e+00
-3.42240334e-01 -2.80068278e-01 -5.84086597e-01 3.25933963e-01
4.86124992e-01 -1.89394936e-01 -6.80248260e-01 1.34915459e+00
8.53919804e-01 3.59747469e-01 -1.91030741e-01 1.17540133e+00
9.27741885e-01 5.53331614e-01 -4.08258379e-01 -2.18060523e-01
1.23737073e+00 -1.09114897e+00 -5.40736377e-01 -3.85570049e-01
5.28793037e-01 -8.05330515e-01 6.63929462e-01 2.57293522e-01
-1.16411209e+00 -5.38194120e-01 -8.95771682e-01 -6.47886842e-02
1.31086437e-02 9.39339027e-02 7.26842940e-01 5.85458159e-01
-1.01067996e+00 2.28216574e-01 -8.99378836e-01 -1.95185095e-01
3.88469905e-01 2.50570893e-01 -1.69471204e-01 -2.69545257e-01
-7.31558383e-01 7.31606603e-01 3.03098351e-01 2.74741977e-01
-8.86422455e-01 -7.88574100e-01 -8.43850851e-01 -3.97850782e-01
5.43359458e-01 -8.90597522e-01 1.13916075e+00 -6.25426888e-01
-1.15200865e+00 7.94247448e-01 -5.97857714e-01 -1.48840055e-01
7.37228692e-01 -4.76013422e-01 -1.48872152e-01 2.06096843e-01
2.44419962e-01 7.76211262e-01 7.82420218e-01 -1.20059717e+00
-1.17573440e+00 -4.61196363e-01 1.70191228e-01 4.31235880e-01
-2.30021536e-01 -6.34908527e-02 -1.02181053e+00 -6.08826041e-01
4.06349808e-01 -1.00204813e+00 -2.46378779e-01 3.90397608e-01
-2.27148116e-01 -1.24968551e-01 1.02242267e+00 -2.25929588e-01
1.21194601e+00 -2.17607546e+00 9.71808508e-02 -3.20771009e-01
3.33462149e-01 2.42622811e-02 8.62313714e-03 6.14363588e-02
3.91648829e-01 -3.11791122e-01 2.51642555e-01 -3.13014179e-01
-4.05928910e-01 -4.46519032e-02 -2.90657848e-01 9.37017441e-01
-1.09449126e-01 9.02036905e-01 -1.02491832e+00 -5.26646316e-01
6.99037731e-01 4.43733037e-01 -8.12841952e-01 -8.22549462e-02
-3.01556922e-02 4.15468395e-01 -5.71004748e-01 7.67075717e-01
9.29474711e-01 -3.42219323e-01 -2.56881833e-01 -2.62068510e-01
-4.21388298e-01 1.84787624e-02 -1.45363498e+00 1.63645887e+00
-1.84779733e-01 7.67892241e-01 6.14867918e-02 -5.37050426e-01
7.30684876e-01 5.75908087e-03 7.09664166e-01 -4.80741322e-01
-3.10037140e-04 4.58874516e-02 -3.04228455e-01 -3.15265357e-01
7.41277158e-01 2.28199840e-01 1.55254861e-03 1.24251388e-01
-3.82485867e-01 -9.74145085e-02 3.03860437e-02 1.72720298e-01
1.02513003e+00 1.86109737e-01 3.63271981e-01 -1.50056556e-01
5.10979474e-01 1.88139873e-03 8.09078038e-01 7.60280788e-01
-4.17348981e-01 8.03273976e-01 2.63323225e-02 -5.19342124e-01
-8.72390330e-01 -1.20920455e+00 -1.29654229e-01 9.39981401e-01
7.98707485e-01 -2.96154410e-01 -4.55828339e-01 -4.74752396e-01
8.01080093e-02 2.74315447e-01 -4.52375293e-01 7.63134435e-02
-5.30916810e-01 -5.41488349e-01 9.53268334e-02 6.70294762e-01
4.09022957e-01 -5.12276053e-01 -1.11241579e+00 9.90556851e-02
-2.98647076e-01 -1.35232198e+00 -7.96996951e-01 -1.91615537e-01
-9.87033427e-01 -1.01298189e+00 -8.43334436e-01 -4.80460972e-01
5.02148032e-01 1.15915775e+00 8.32587361e-01 -6.49413913e-02
-2.17114583e-01 4.26244080e-01 -3.33617419e-01 -3.05319697e-01
1.19315512e-01 -1.52850449e-01 2.50352025e-01 -4.25940864e-02
5.03747940e-01 -3.40158641e-01 -1.07525444e+00 6.15937591e-01
-5.48073053e-01 3.16786081e-01 2.72945404e-01 6.41310751e-01
7.25596189e-01 -5.50503246e-02 1.00405537e-01 -4.47735518e-01
-1.70564070e-01 -3.34185272e-01 -1.00036478e+00 -4.63622212e-02
-3.03008914e-01 -3.28304857e-01 1.71819538e-01 -5.37411511e-01
-9.82281685e-01 4.60558861e-01 2.62676388e-01 -8.61185908e-01
7.72175658e-03 -1.76686972e-01 -5.96868852e-03 -2.48507708e-01
2.96331465e-01 2.82404959e-01 -3.55188668e-01 -3.05949509e-01
2.23174229e-01 3.88609827e-01 5.49050689e-01 -1.62577033e-01
7.80258119e-01 1.06354845e+00 -1.11599304e-01 -8.77021253e-01
-1.12812340e+00 -9.35980141e-01 -6.97681546e-01 -5.74478924e-01
8.82523239e-01 -1.30096459e+00 -9.11997736e-01 5.26040316e-01
-1.33343816e+00 -4.60408628e-02 1.14773050e-01 5.87527931e-01
-5.75208724e-01 5.85281312e-01 -4.47422802e-01 -9.72660005e-01
-1.78969547e-01 -1.07945621e+00 1.40728450e+00 2.10846424e-01
-7.48903304e-02 -8.06680202e-01 -1.11024112e-01 2.88772881e-01
7.59647861e-02 2.84329802e-01 2.76946902e-01 8.47111642e-02
-1.06902468e+00 -3.71759199e-02 -3.14464092e-01 8.41091480e-03
1.07277319e-01 5.17124832e-02 -1.08364093e+00 -4.02161926e-01
1.04762532e-01 1.88433960e-01 8.65943074e-01 8.90996516e-01
8.93372178e-01 -2.20737210e-03 -5.56315601e-01 7.57502615e-01
1.41338873e+00 3.05390686e-01 3.45845819e-01 3.46112877e-01
9.08879995e-01 5.05591929e-01 9.75050390e-01 7.63164878e-01
4.78378028e-01 1.01632094e+00 7.24523723e-01 8.60472545e-02
-2.31503341e-02 -2.55413830e-01 6.10800982e-01 6.20846331e-01
-1.44896209e-01 -3.44693214e-01 -7.70532727e-01 6.56406760e-01
-1.78464806e+00 -1.11414421e+00 -4.37715918e-01 2.07712698e+00
2.79740512e-01 2.01770291e-01 1.60028055e-01 -1.46793023e-01
8.68953049e-01 3.59521329e-01 -7.71762371e-01 2.29995668e-01
-1.65021896e-01 -4.63218063e-01 8.69939387e-01 4.47073609e-01
-1.22302473e+00 9.51327980e-01 6.11253929e+00 5.89619815e-01
-8.14882755e-01 1.12243563e-01 3.69970202e-01 -6.07684612e-01
-5.36830090e-02 6.65336847e-03 -1.49981427e+00 4.43605006e-01
3.67455781e-01 1.90516841e-02 2.25762174e-01 8.35217118e-01
5.34620941e-01 -5.99055290e-01 -1.11601245e+00 1.31813574e+00
6.79630041e-02 -1.49541128e+00 -4.47503552e-02 1.30746111e-01
8.53660882e-01 2.79920399e-01 9.58525017e-02 -2.66447783e-01
1.49293259e-01 -5.93656600e-01 1.10681963e+00 4.55928564e-01
4.97773975e-01 -7.16608286e-01 4.82530832e-01 5.76640666e-01
-1.59173441e+00 -2.00685278e-01 -4.91744667e-01 -1.45805582e-01
5.65992236e-01 6.17318094e-01 -4.75674510e-01 4.58775759e-01
9.66162622e-01 1.09626651e+00 -6.07058764e-01 1.12707639e+00
2.12303162e-01 5.09286206e-03 -4.74490076e-01 1.52124986e-02
3.22963417e-01 -1.26598045e-01 8.57786715e-01 9.80061829e-01
4.43103343e-01 1.59794539e-01 3.13721180e-01 6.51733875e-01
3.44840795e-01 -2.06579462e-01 -6.78810418e-01 4.42486852e-01
5.11268556e-01 1.04024434e+00 -8.26370895e-01 -2.72949398e-01
-8.07383180e-01 9.08599794e-01 4.05949578e-02 3.66342843e-01
-9.73648846e-01 -5.32757230e-02 8.11286926e-01 3.18159729e-01
7.13902175e-01 -3.86818826e-01 -1.36358723e-01 -1.35360992e+00
2.60151744e-01 -4.39308614e-01 2.79836953e-01 -8.55419934e-01
-1.03870404e+00 3.20897102e-01 4.75806482e-02 -1.63475192e+00
-1.03944086e-01 -5.35265982e-01 -2.98833340e-01 4.81627613e-01
-1.46985555e+00 -9.29901779e-01 -5.71250975e-01 6.67293608e-01
8.36283684e-01 2.65636146e-01 1.42235160e-01 4.30025786e-01
-5.52211940e-01 3.03548723e-01 1.11349471e-01 7.72529393e-02
6.88907862e-01 -6.91543579e-01 3.99618357e-01 1.04870558e+00
1.02301389e-01 3.47558826e-01 6.21705770e-01 -5.68867922e-01
-1.55401349e+00 -1.13701689e+00 5.30768991e-01 -8.63605201e-01
5.51221073e-01 -4.41231757e-01 -8.22034776e-01 5.60336411e-01
-2.55982220e-01 4.41894412e-01 3.99954356e-02 -3.39739561e-01
-1.13590680e-01 -1.16671413e-01 -6.92856491e-01 5.22744119e-01
1.34508216e+00 -2.64749408e-01 -2.42436692e-01 2.25319013e-01
6.21024311e-01 -8.16324174e-01 -6.60117447e-01 4.48909193e-01
6.20698929e-01 -1.35164320e+00 1.16385937e+00 -4.95309941e-02
1.79363996e-01 -6.52444899e-01 -3.48612309e-01 -6.43562436e-01
-1.12237021e-01 -5.16859472e-01 -4.19464260e-01 1.07477415e+00
-6.01881482e-02 -5.47930062e-01 8.36933374e-01 4.55623537e-01
-8.17815587e-02 -6.08880937e-01 -1.06033850e+00 -7.46155679e-01
-3.86760205e-01 -5.20573735e-01 1.52309895e-01 6.26421928e-01
-5.02050459e-01 2.77085632e-01 -4.90955323e-01 4.51691687e-01
8.64510417e-01 3.65309656e-01 9.86435235e-01 -1.15242910e+00
-1.47520915e-01 -2.62615591e-01 -5.21913528e-01 -1.66437972e+00
-1.55683354e-01 -4.17891681e-01 5.30094057e-02 -1.24237013e+00
3.39972824e-01 -4.10892785e-01 -4.84275669e-02 -7.16788545e-02
-1.88693464e-01 2.72507191e-01 1.99907005e-01 1.84866399e-01
-8.55576038e-01 4.53417391e-01 1.44681525e+00 1.39345348e-01
-2.64254957e-01 1.72053188e-01 -3.52530330e-01 1.03982246e+00
3.74800980e-01 -5.54117680e-01 -3.30740601e-01 -5.25679290e-01
1.22160532e-01 1.78696439e-01 7.81437039e-01 -8.61506224e-01
4.24677193e-01 -1.70188278e-01 5.06416202e-01 -1.32017767e+00
6.23184443e-01 -8.19237590e-01 2.54343629e-01 4.27861720e-01
2.36840159e-01 2.56351709e-01 2.28126720e-01 9.43002820e-01
-1.42102093e-01 -1.75566059e-02 9.40658867e-01 -1.34406149e-01
-1.03261781e+00 5.48593283e-01 -2.38236785e-01 -6.23420672e-03
1.41740036e+00 -6.95891261e-01 -2.22274363e-01 -3.58302146e-01
-2.82914400e-01 3.55106950e-01 8.21434081e-01 5.31092823e-01
6.83754742e-01 -1.35747623e+00 -4.29555684e-01 9.74442735e-02
1.19981460e-01 3.17672312e-01 5.78633904e-01 1.08133280e+00
-4.11824971e-01 6.12945259e-01 1.32611901e-01 -1.22649932e+00
-1.36383092e+00 7.45491207e-01 3.73398602e-01 1.91364020e-01
-9.38916504e-01 7.45188951e-01 7.93700635e-01 1.27214238e-01
2.56993592e-01 -4.33043152e-01 -5.76133616e-02 1.26473039e-01
6.92607701e-01 5.55402339e-01 -2.34120101e-01 -7.32702255e-01
-5.13560414e-01 1.03747988e+00 -1.50397852e-01 1.87999103e-02
1.29944956e+00 -6.15678012e-01 1.63784206e-01 4.72036660e-01
1.08107054e+00 1.23248614e-01 -1.92836821e+00 -1.44982412e-01
-1.79035932e-01 -8.54569316e-01 1.64874464e-01 -3.51289734e-02
-1.14295566e+00 8.53959620e-01 6.46092474e-01 -1.21370688e-01
1.04124165e+00 1.39967799e-01 6.70513272e-01 9.75309983e-02
5.76241910e-01 -8.69790316e-01 1.94547474e-01 4.84486341e-01
5.64567566e-01 -1.28170669e+00 2.42860645e-01 -6.39259458e-01
-6.07607186e-01 9.65938091e-01 8.46459746e-01 -1.32948771e-01
4.38994527e-01 3.25835943e-01 -1.62164196e-01 -1.82780549e-01
-7.88597763e-01 -6.48859963e-02 2.31772646e-01 5.37061989e-01
1.18999131e-01 -4.60137159e-01 1.21954694e-01 1.09413013e-01
2.43725628e-01 -2.11964518e-01 4.68594760e-01 7.39456832e-01
-5.21934390e-01 -5.35905957e-01 -5.04914343e-01 2.77736515e-01
-3.80770206e-01 -1.58811721e-03 7.57413879e-02 9.56696033e-01
2.02068463e-02 8.39060426e-01 2.31871158e-01 -2.71813542e-01
3.98150951e-01 -4.44215149e-01 4.77105141e-01 -5.24933577e-01
-1.82673514e-01 3.70270729e-01 -1.95899904e-01 -8.37467849e-01
-7.89233327e-01 -8.71952474e-01 -9.89681423e-01 -3.91465813e-01
-7.23048985e-01 -4.31205779e-01 2.51491845e-01 6.69437826e-01
3.72073025e-01 2.49525324e-01 5.84418952e-01 -1.38378406e+00
-2.43268192e-01 -4.84921396e-01 -4.41744626e-01 1.07660145e-01
6.65705681e-01 -1.07279098e+00 -4.85624641e-01 -3.72985341e-02] | [8.001287460327148, -1.9384799003601074] |
602cbf6d-182e-4e53-959b-3df8bc5e209f | godp-globally-optimized-dual-pathway-system | 1704.02402 | null | http://arxiv.org/abs/1704.02402v2 | http://arxiv.org/pdf/1704.02402v2.pdf | GoDP: Globally optimized dual pathway system for facial landmark localization in-the-wild | Facial landmark localization is a fundamental module for pose-invariant face
recognition. The most common approach for facial landmark detection is cascaded
regression, which is composed of two steps: feature extraction and facial shape
regression. Recent methods employ deep convolutional networks to extract robust
features for each step, while the whole system could be regarded as a deep
cascaded regression architecture. In this work, instead of employing a deep
regression network, a Globally Optimized Dual-Pathway (GoDP) deep architecture
is proposed to identify the target pixels through solving a cascaded pixel
labeling problem without resorting to high-level inference models or complex
stacked architecture. The proposed end-to-end system relies on distance-aware
softmax functions and dual-pathway proposal-refinement architecture. Results
show that it outperforms the state-of-the-art cascaded regression-based methods
on multiple in-the-wild face alignment databases. The model achieves 1.84
normalized mean error (NME) on the AFLW database, which outperforms 3DDFA by
61.8%. Experiments on face identification demonstrate that GoDP, coupled with
DPM-headhunter, is able to improve rank-1 identification rate by 44.2% compared
to Dlib toolbox on a challenging database. | ['Ioannis A. Kakadiaris', 'Shishir K. Shah', 'Yuhang Wu'] | 2017-04-07 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [ 6.60574883e-02 -7.80257508e-02 -2.01363191e-01 -7.44184136e-01
-1.05054724e+00 -2.25996107e-01 5.15607238e-01 -3.42060596e-01
-7.95047045e-01 2.59895444e-01 -1.83155924e-01 -5.74159622e-02
1.40222982e-01 -3.63713920e-01 -7.89962888e-01 -8.16345870e-01
7.42910877e-02 5.90443790e-01 -1.63221464e-01 -1.34989068e-01
2.16333449e-01 9.16141927e-01 -1.56979394e+00 1.55159563e-01
2.28399500e-01 1.25458026e+00 -3.48946840e-01 3.76156151e-01
3.76175158e-02 3.34144115e-01 -3.36934537e-01 -6.80471241e-01
4.64829445e-01 -6.04036041e-02 -5.09715021e-01 -1.63190156e-01
1.23161674e+00 -2.73674786e-01 7.63834491e-02 9.89847839e-01
9.49479640e-01 -1.16081536e-01 4.77862775e-01 -1.12643194e+00
-3.08429897e-01 2.79623747e-01 -1.11154163e+00 4.83401157e-02
1.40426487e-01 -2.31159274e-02 7.53691554e-01 -1.53087127e+00
3.98353279e-01 1.43858182e+00 8.02416682e-01 8.15177739e-01
-1.44005144e+00 -1.15452397e+00 -7.93654397e-02 2.34041020e-01
-1.90448987e+00 -1.00628984e+00 6.84327185e-01 -2.40497738e-01
9.50393558e-01 1.00604981e-01 2.02790871e-01 9.78638172e-01
-1.64680649e-02 4.97156173e-01 9.57354069e-01 -3.96117628e-01
-1.92258865e-01 -1.82431266e-01 -2.73548454e-01 1.33439457e+00
4.76051569e-02 1.12994865e-01 -7.89524138e-01 -6.92176446e-03
7.00396776e-01 -1.77541807e-01 -3.11728939e-02 -1.59361228e-01
-7.40816116e-01 5.84309876e-01 6.01143658e-01 1.25914998e-02
-4.05073762e-01 1.72659025e-01 2.32238799e-01 -9.73425619e-03
5.87408364e-01 -3.18761989e-02 -5.31407654e-01 4.41821814e-01
-1.41973686e+00 3.19185145e-02 4.36426193e-01 6.18657708e-01
9.28828895e-01 1.36688733e-02 -4.01459336e-01 9.82937038e-01
7.94782996e-01 3.74658465e-01 2.34431699e-01 -5.77998698e-01
3.25252444e-01 7.52747953e-01 -2.78396368e-01 -9.41924036e-01
-7.76150644e-01 -5.46861291e-01 -7.30997980e-01 5.83979309e-01
5.05200267e-01 -1.00730658e-01 -1.26934946e+00 1.76846600e+00
5.34123600e-01 5.45219302e-01 -3.27475131e-01 9.65013623e-01
1.06998301e+00 2.00942412e-01 2.40432009e-01 -6.04527770e-03
1.50907671e+00 -1.15266168e+00 -3.18192184e-01 -9.97298807e-02
2.71378934e-01 -9.39912021e-01 5.71080625e-01 2.95931131e-01
-1.03497183e+00 -7.97698498e-01 -9.24334168e-01 -2.83598423e-01
-2.35434592e-01 7.01555908e-01 3.03303957e-01 8.22752059e-01
-1.35177982e+00 2.87594646e-01 -7.60518491e-01 -2.69881874e-01
9.90537882e-01 9.67430234e-01 -8.45121443e-01 3.63072827e-02
-5.10664344e-01 8.06746840e-01 -4.85543646e-02 7.77925968e-01
-1.10537076e+00 -8.74216735e-01 -8.09421062e-01 -5.97754717e-02
3.23075712e-01 -6.80971682e-01 8.41194153e-01 -8.13442469e-01
-1.68558872e+00 1.32618046e+00 -5.78662038e-01 -2.11516127e-01
5.67850828e-01 -2.93525994e-01 -1.67194068e-01 3.37609053e-02
1.41306315e-02 9.88674521e-01 1.21751726e+00 -9.21253502e-01
-5.47259927e-01 -8.47297490e-01 -4.97325510e-01 -1.20958701e-01
-2.21635863e-01 4.50756669e-01 -1.01143610e+00 -4.33039665e-01
1.22726873e-01 -9.35390651e-01 -1.84620589e-01 1.61715060e-01
-3.39904845e-01 -3.61818254e-01 7.02144444e-01 -7.38277853e-01
8.34897399e-01 -1.88976014e+00 1.82763278e-01 4.03198600e-01
8.42969567e-02 4.49847341e-01 -5.46206832e-01 -2.35897720e-01
-2.55277157e-01 -1.09533645e-01 -2.37128988e-01 -9.47947621e-01
-4.68413047e-02 -3.40180397e-01 4.04205434e-02 8.72278392e-01
5.82345903e-01 9.40666974e-01 -3.47661942e-01 -4.37535971e-01
9.90293771e-02 1.04580808e+00 -7.12693691e-01 5.29913791e-02
-9.62485224e-02 6.99670792e-01 -7.76371434e-02 1.02233028e+00
1.02312827e+00 1.33553848e-01 5.06410897e-02 -6.01752579e-01
-1.58095136e-01 -1.70928121e-01 -1.01967740e+00 2.02867270e+00
-4.49635357e-01 4.65205342e-01 1.78598121e-01 -6.94619536e-01
1.15903628e+00 2.72183269e-01 3.75528425e-01 -7.01315224e-01
3.01327437e-01 1.79653540e-01 -1.40231416e-01 -1.51576936e-01
1.69986906e-03 6.53860569e-02 2.25030079e-01 6.68493798e-03
5.96713662e-01 4.83924985e-01 -7.16212541e-02 -4.48233485e-01
8.19928527e-01 5.04008591e-01 -1.20564606e-02 -2.42692456e-01
1.05758786e+00 -4.69217002e-01 7.94533074e-01 2.47836128e-01
-1.07379824e-01 6.15714133e-01 4.93480891e-01 -5.91443062e-01
-6.10072792e-01 -6.34671807e-01 -2.53275961e-01 1.44714689e+00
-2.26057440e-01 -4.56691056e-01 -9.77264702e-01 -7.78049350e-01
1.50772870e-01 8.91532972e-02 -8.21432292e-01 1.19254395e-01
-6.19198263e-01 -1.05378854e+00 9.29040968e-01 3.26899081e-01
6.21133149e-01 -8.22476387e-01 -2.55576670e-01 -7.75929540e-02
2.62001157e-01 -1.12119555e+00 -4.49838310e-01 -2.02492513e-02
-4.46795702e-01 -1.19254744e+00 -7.31754482e-01 -8.11604500e-01
9.87559378e-01 -5.47063462e-02 7.85905004e-01 1.42765895e-01
-5.99731743e-01 1.67293280e-01 1.51006386e-01 -2.13138565e-01
1.31417215e-01 7.75270984e-02 9.40978974e-02 6.58740342e-01
6.44493520e-01 -3.04621905e-01 -8.80699754e-01 4.75031883e-01
-3.77034396e-01 -2.49360800e-01 8.09877098e-01 9.30194736e-01
8.47675502e-01 -3.67691934e-01 2.55251706e-01 -6.75409496e-01
3.99689525e-02 -6.59041479e-02 -1.04912758e+00 4.61073875e-01
-6.26387715e-01 6.48349449e-02 2.94108391e-01 -1.21878378e-01
-9.54262316e-01 7.32603073e-01 -5.33761203e-01 -5.89993834e-01
-2.05703497e-01 8.04748833e-02 -4.22261953e-01 -7.15416968e-01
4.36477900e-01 9.24976915e-02 9.92177576e-02 -5.93311191e-01
3.22891414e-01 4.31790888e-01 6.74779296e-01 -3.67472082e-01
8.55049431e-01 4.46122706e-01 4.66810286e-01 -8.75021517e-01
-7.64553487e-01 -3.22630316e-01 -1.03340697e+00 -2.24204719e-01
9.19306159e-01 -1.19550061e+00 -1.24220192e+00 6.42558038e-01
-1.09685314e+00 -2.21324638e-01 4.17943865e-01 2.53688246e-01
-2.12870032e-01 8.60113874e-02 -3.51276249e-01 -6.22416675e-01
-5.44624329e-01 -1.40180123e+00 1.48649132e+00 3.85194540e-01
-1.35503681e-02 -5.71228921e-01 -1.58473864e-01 6.74021959e-01
4.46016580e-01 2.67741650e-01 3.67614269e-01 -5.80959022e-01
-4.65853363e-01 -3.95525903e-01 -2.44354695e-01 2.71822691e-01
-1.10284343e-01 1.37319982e-01 -1.40649891e+00 -3.95461708e-01
-3.29233736e-01 -2.79889643e-01 1.12408662e+00 3.61497343e-01
1.28178060e+00 -1.81285560e-01 -1.43647820e-01 1.23068070e+00
1.31252623e+00 -2.28000581e-01 5.31769216e-01 2.39280969e-01
9.11480546e-01 6.92418039e-01 3.91646236e-01 3.19355309e-01
4.00136471e-01 1.02104247e+00 5.92636883e-01 -3.69160265e-01
-2.29294002e-01 -1.85546368e-01 3.21159422e-01 -3.12014893e-02
-2.93078572e-01 3.65960956e-01 -8.78241658e-01 2.11562179e-02
-1.67002857e+00 -8.56468260e-01 5.16434610e-02 2.14087963e+00
6.20815635e-01 -1.89297348e-01 2.02053010e-01 -6.18361719e-02
5.84970415e-01 5.17071858e-02 -4.55735594e-01 -1.28925666e-01
-2.10379422e-01 6.06344402e-01 5.68742454e-01 6.29651546e-01
-1.35928547e+00 1.43111682e+00 5.23468590e+00 8.22416902e-01
-1.52306223e+00 2.21754596e-01 9.14565265e-01 -1.05942205e-01
3.88284773e-01 -3.95200342e-01 -1.47009075e+00 2.38845930e-01
7.78378785e-01 5.38564622e-01 3.54005545e-01 6.75043046e-01
5.85793257e-02 7.20785633e-02 -1.13560426e+00 1.27684617e+00
3.82499903e-01 -1.13583970e+00 -8.15006671e-04 2.48950850e-02
5.79389215e-01 6.56983107e-02 4.14898366e-01 1.15859248e-01
-1.61304146e-01 -1.54293132e+00 5.35339296e-01 5.27063727e-01
8.86963904e-01 -9.86162603e-01 6.90473378e-01 8.33790470e-03
-1.22702920e+00 -4.02350686e-02 -4.08331126e-01 5.30305207e-01
-2.48433560e-01 3.05722326e-01 -9.44429219e-01 3.47664505e-01
8.06440711e-01 5.48483670e-01 -7.48609006e-01 1.02724040e+00
-4.14952487e-01 4.71235842e-01 -3.95023555e-01 4.50625211e-01
6.25595525e-02 -5.65330684e-02 1.56188980e-01 1.36488903e+00
2.74785310e-01 -2.14362755e-01 -9.52805206e-02 6.76794469e-01
-4.28440124e-01 1.78090304e-01 -1.14521325e-01 4.70562488e-01
2.19986588e-01 1.87247276e+00 -5.87221146e-01 2.37973377e-01
-4.26035285e-01 1.12645698e+00 4.52044010e-01 1.95341840e-01
-8.82133603e-01 -5.87579003e-03 8.77267420e-01 -5.85708730e-02
4.92352188e-01 7.12076342e-03 -8.99506081e-03 -5.85905612e-01
-7.73434341e-02 -9.12997425e-01 4.05745983e-01 -2.90658504e-01
-1.01960075e+00 8.40974689e-01 -4.67460275e-01 -8.62909377e-01
-3.84487025e-03 -8.55274379e-01 -4.21596617e-01 1.12939978e+00
-1.86898375e+00 -1.73500288e+00 -4.17326927e-01 8.13945353e-01
3.30730766e-01 -5.51520765e-01 1.01286352e+00 6.02309346e-01
-1.20421743e+00 1.20124066e+00 -4.72919464e-01 4.34403330e-01
1.02480769e+00 -9.21887815e-01 3.58084112e-01 8.89933646e-01
3.65338266e-01 6.40489817e-01 2.17697009e-01 -2.81888336e-01
-1.61651754e+00 -1.41231024e+00 8.33240032e-01 -4.37940985e-01
2.03747153e-01 -4.88275230e-01 -6.86734080e-01 5.91326654e-01
-4.42934707e-02 7.56353080e-01 8.30470741e-01 9.23697874e-02
-7.10192323e-01 -6.40260041e-01 -1.22871494e+00 5.01605928e-01
1.04950571e+00 -5.01032948e-01 4.79684304e-03 9.86524001e-02
5.78080453e-02 -4.78647798e-01 -7.66282320e-01 6.90109253e-01
7.81323493e-01 -8.73918951e-01 1.18703151e+00 -5.54673493e-01
1.05221607e-01 -6.38826609e-01 -9.62995291e-02 -8.69935751e-01
-3.58395725e-01 -7.79666364e-01 -1.18158218e-02 1.38764989e+00
5.84249258e-01 -3.33480090e-01 9.87673402e-01 4.23891604e-01
1.50686413e-01 -8.16725731e-01 -1.25249767e+00 -2.84533560e-01
-2.94934094e-01 -2.27778271e-01 5.24941087e-01 5.54964781e-01
-6.57595873e-01 4.72609580e-01 -3.01285326e-01 5.66720724e-01
8.87925029e-01 -2.05144629e-01 9.31926668e-01 -1.26895928e+00
2.02166885e-01 -7.34946251e-01 -5.97599685e-01 -7.70165980e-01
6.66908562e-01 -1.08104360e+00 -2.72039007e-02 -1.02790725e+00
4.18023542e-02 -1.99247733e-01 -5.54702997e-01 9.20315802e-01
-1.05350077e-01 9.56518948e-01 -8.07587132e-02 -1.20204225e-01
-5.71784079e-01 4.73803222e-01 8.48470271e-01 -2.13530526e-01
-7.05002993e-02 6.32921904e-02 -5.48519790e-01 8.17933083e-01
5.98067224e-01 -3.29676539e-01 3.91307175e-02 -4.69831705e-01
-2.70779356e-02 -3.75985801e-01 5.31068861e-01 -9.24721897e-01
6.48060501e-01 4.06243026e-01 1.00513411e+00 -6.32427037e-01
5.82057416e-01 -8.96572590e-01 -8.49698856e-03 3.92432213e-01
-1.37934893e-01 3.93418074e-02 5.15569627e-01 1.88991457e-01
-1.88907132e-01 2.09306747e-01 1.02645183e+00 2.70682544e-01
-8.18649530e-01 6.68247402e-01 2.02076316e-01 -4.97698039e-01
1.05528176e+00 -1.51523575e-01 -5.15086502e-02 2.03478619e-01
-7.10456908e-01 9.40720588e-02 6.63436800e-02 4.58241045e-01
6.67898595e-01 -1.16227019e+00 -1.13019156e+00 6.07797086e-01
-5.65660074e-02 -5.69930971e-02 1.75155848e-01 1.04326296e+00
-4.61223036e-01 3.41320425e-01 -3.49075317e-01 -8.33678365e-01
-1.77457798e+00 7.62447193e-02 6.96297109e-01 -1.55629991e-02
-2.49262258e-01 1.48076570e+00 -4.23779786e-02 -5.12377739e-01
5.52489161e-01 2.70945638e-01 -3.76482397e-01 5.03888607e-01
6.85313642e-01 2.41004258e-01 4.58573490e-01 -1.04532361e+00
-8.27062070e-01 1.20722342e+00 -1.51223257e-01 1.07882805e-01
1.49860823e+00 2.18404308e-01 -5.54004073e-01 -2.99914867e-01
1.35038900e+00 2.08241493e-02 -1.30263805e+00 -1.84870481e-01
2.21480638e-01 -4.06734020e-01 2.65245795e-01 -8.75301182e-01
-1.55736423e+00 8.21528852e-01 9.57618177e-01 -8.37913573e-01
1.24120784e+00 -2.35742435e-01 2.93046415e-01 3.15052122e-01
2.66429305e-01 -7.04255462e-01 -9.61975083e-02 2.99425572e-01
1.19399524e+00 -1.46799648e+00 -5.98081239e-02 -4.39236581e-01
-2.37587735e-01 1.30695379e+00 7.90492713e-01 -4.23009619e-02
7.88450360e-01 2.88368613e-01 1.86753914e-01 -1.58347338e-01
-5.00280321e-01 -3.82739991e-01 6.99690342e-01 2.66679585e-01
6.70675814e-01 -3.04579288e-01 3.93575169e-02 4.53403503e-01
-8.81060138e-02 -1.27421230e-01 -4.45212841e-01 3.48828435e-01
-7.11796284e-02 -1.21714878e+00 -5.12332976e-01 4.09256108e-02
-7.79595912e-01 -7.05412477e-02 -3.80639225e-01 6.93189442e-01
3.49899918e-01 8.24788213e-01 7.55026937e-02 -4.34472024e-01
2.80554056e-01 2.30513617e-01 8.72186244e-01 -4.57861960e-01
-8.09558511e-01 2.68449664e-01 -2.96017230e-01 -9.00470912e-01
-4.71694708e-01 -5.82225382e-01 -1.06563365e+00 -2.39588484e-01
-1.79354027e-01 -2.28690624e-01 9.19727981e-01 6.62473738e-01
5.13352931e-01 3.56089324e-01 7.09427714e-01 -1.09990859e+00
-2.80167550e-01 -9.97950613e-01 -1.08002715e-01 -8.06935355e-02
4.45761800e-01 -6.77605927e-01 7.78624341e-02 -1.38316363e-01] | [13.50031566619873, 0.3829843997955322] |
ec9a827f-48ca-475e-845d-eaa75ccce8e1 | temporal-shift-module-for-efficient-video | 1811.08383 | null | https://arxiv.org/abs/1811.08383v3 | https://arxiv.org/pdf/1811.08383v3.pdf | TSM: Temporal Shift Module for Efficient Video Understanding | The explosive growth in video streaming gives rise to challenges on performing video understanding at high accuracy and low computation cost. Conventional 2D CNNs are computationally cheap but cannot capture temporal relationships; 3D CNN based methods can achieve good performance but are computationally intensive, making it expensive to deploy. In this paper, we propose a generic and effective Temporal Shift Module (TSM) that enjoys both high efficiency and high performance. Specifically, it can achieve the performance of 3D CNN but maintain 2D CNN's complexity. TSM shifts part of the channels along the temporal dimension; thus facilitate information exchanged among neighboring frames. It can be inserted into 2D CNNs to achieve temporal modeling at zero computation and zero parameters. We also extended TSM to online setting, which enables real-time low-latency online video recognition and video object detection. TSM is accurate and efficient: it ranks the first place on the Something-Something leaderboard upon publication; on Jetson Nano and Galaxy Note8, it achieves a low latency of 13ms and 35ms for online video recognition. The code is available at: https://github.com/mit-han-lab/temporal-shift-module. | ['Song Han', 'Ji Lin', 'Chuang Gan'] | 2018-11-20 | tsm-temporal-shift-module-for-efficient-video | http://openaccess.thecvf.com/content_ICCV_2019/html/Lin_TSM_Temporal_Shift_Module_for_Efficient_Video_Understanding_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Lin_TSM_Temporal_Shift_Module_for_Efficient_Video_Understanding_ICCV_2019_paper.pdf | iccv-2019-10 | ['video-object-tracking'] | ['computer-vision'] | [-3.34382772e-01 -5.40535390e-01 -3.83712709e-01 -2.25370854e-01
-4.92000252e-01 -4.91827190e-01 3.28793377e-01 -2.74328023e-01
-5.13167083e-01 2.68445481e-02 -2.63173312e-01 -5.93405366e-01
1.31997943e-01 -5.25605857e-01 -7.97149539e-01 -4.91900116e-01
-2.45366514e-01 1.43903196e-02 5.91325283e-01 6.08806685e-02
1.66202169e-02 5.75053811e-01 -1.43539703e+00 2.99717844e-01
3.10768094e-02 1.49094951e+00 1.71611905e-01 9.87135708e-01
7.83960968e-02 9.02114868e-01 -2.92433619e-01 -1.54927820e-01
4.97386724e-01 3.36268879e-02 -5.05721331e-01 -1.12871155e-01
4.81096417e-01 -7.07295239e-01 -9.76096809e-01 6.91990912e-01
5.50668240e-01 -6.63011894e-02 1.99905224e-02 -1.36079490e+00
-3.48382920e-01 1.85248673e-01 -3.13245624e-01 7.01564252e-01
9.10807177e-02 2.00912252e-01 7.49933183e-01 -8.89612973e-01
4.20351177e-01 1.07268500e+00 6.51766419e-01 8.17397714e-01
-8.30792546e-01 -8.76205087e-01 2.52066225e-01 4.66833502e-01
-1.26903820e+00 -5.21961987e-01 4.90133613e-01 -4.01100010e-01
1.22176754e+00 1.12371601e-01 1.01290619e+00 9.57302451e-01
8.91988948e-02 9.22114909e-01 5.12240589e-01 -2.97949910e-02
9.05837417e-02 -5.19339383e-01 -4.13413160e-02 7.61730433e-01
-1.23251088e-01 5.88342436e-02 -9.75204647e-01 2.26302981e-01
1.26831055e+00 4.39215988e-01 -2.79997945e-01 1.26505375e-01
-1.44106460e+00 4.21219558e-01 2.04523697e-01 1.60318479e-01
-1.00675814e-01 6.88839197e-01 6.27164781e-01 5.88953018e-01
4.29215848e-01 -2.55054999e-02 -5.33032477e-01 -6.88541651e-01
-9.44941461e-01 3.02201360e-01 5.22809386e-01 1.35417604e+00
4.99368876e-01 1.23047404e-01 1.63611069e-01 4.48659509e-01
-9.10731480e-02 3.14859837e-01 3.47992450e-01 -1.14329410e+00
4.49283928e-01 4.19700116e-01 -6.96397722e-02 -1.09799564e+00
-5.56358278e-01 -3.09194624e-01 -1.01979017e+00 9.90223419e-03
2.55659193e-01 9.85602662e-02 -7.70850599e-01 1.42235267e+00
3.69066536e-01 7.28914976e-01 -1.95702061e-01 1.16869080e+00
9.55102563e-01 7.22354650e-01 -3.62829268e-01 -9.08872709e-02
1.21169353e+00 -1.22237098e+00 -4.84229773e-01 5.21293581e-02
7.95817316e-01 -6.08337879e-01 7.12118447e-01 4.91014332e-01
-1.19647193e+00 -6.27068222e-01 -9.30947542e-01 -2.81870306e-01
-6.38156608e-02 1.92684323e-01 9.04726565e-01 2.87508458e-01
-1.23771584e+00 6.87529922e-01 -1.29199016e+00 -1.82971954e-01
5.16193926e-01 6.17809534e-01 -3.25674683e-01 -6.34148419e-02
-9.72245157e-01 3.62448931e-01 8.61736760e-02 2.75071859e-01
-1.07721865e+00 -8.54701102e-01 -5.66882968e-01 -3.16246822e-02
5.17769039e-01 -5.40711105e-01 1.61504817e+00 -6.52412117e-01
-1.49454701e+00 7.59105980e-01 -3.45817000e-01 -6.60250843e-01
5.35017788e-01 -2.21546128e-01 -3.21177036e-01 3.59408796e-01
-1.94961667e-01 7.32380331e-01 7.80575335e-01 -3.93588930e-01
-6.87687576e-01 -2.40176037e-01 2.35983625e-01 -4.07988504e-02
-4.57775205e-01 5.42859137e-02 -1.20582104e+00 -5.91853857e-01
3.79646093e-01 -1.05331194e+00 -1.47618979e-01 4.71756428e-01
-1.16652615e-01 -1.00676231e-01 1.00011230e+00 -3.63605738e-01
1.12640989e+00 -2.22855973e+00 -1.23292319e-01 -2.13854596e-01
6.62263811e-01 5.30509531e-01 -5.82510196e-02 3.31529886e-01
2.04886999e-02 3.97344725e-03 3.87380093e-01 -3.16101402e-01
-2.03766733e-01 1.01486184e-01 -1.25626504e-01 4.03242588e-01
1.23459175e-01 1.06890905e+00 -7.31980741e-01 -2.12302670e-01
2.46197373e-01 4.95484233e-01 -8.53618920e-01 3.61640565e-03
-1.15794063e-01 4.50481325e-01 -4.76553828e-01 8.81159246e-01
6.68185472e-01 -6.08147919e-01 1.86935335e-01 -3.80094081e-01
-4.85547870e-01 3.01821291e-01 -9.84259546e-01 1.51902354e+00
-2.60020524e-01 1.09993196e+00 1.29207730e-01 -1.12803352e+00
8.39137912e-01 4.11006957e-01 5.66711783e-01 -1.04592156e+00
2.63492048e-01 1.52646005e-01 -2.25726992e-01 -4.61387753e-01
4.57643270e-01 2.92580485e-01 2.73595929e-01 2.64830023e-01
1.30062208e-01 1.52811691e-01 5.31150438e-02 2.27502927e-01
1.21637297e+00 -6.39674291e-02 -1.19674221e-01 -1.60329267e-01
1.16772428e-01 -1.90891623e-01 7.72986889e-01 7.64852107e-01
-3.12588036e-01 5.54245532e-01 5.95889628e-01 -9.15551782e-01
-1.27969730e+00 -7.63923585e-01 2.43872628e-01 8.70663285e-01
3.33673358e-01 -6.64992273e-01 -5.06820858e-01 -2.02896565e-01
-2.88385134e-02 -2.31145248e-01 -3.72249454e-01 -8.33433717e-02
-7.35307992e-01 -3.60595137e-01 6.44740939e-01 7.02868223e-01
7.73063958e-01 -4.13662046e-01 -6.20943666e-01 2.72137582e-01
-1.19166367e-01 -1.57984734e+00 -6.46817744e-01 8.14468414e-03
-1.10105336e+00 -1.09836531e+00 -6.59435391e-01 -8.24174941e-01
3.46000671e-01 8.35993052e-01 8.54211450e-01 2.72068113e-01
-3.01569492e-01 2.48556018e-01 -4.37519968e-01 -3.47483575e-01
1.98373035e-01 3.37757245e-02 2.42590636e-01 -7.04580620e-02
2.37663239e-01 -6.11743689e-01 -8.29442203e-01 5.44004798e-01
-7.27643490e-01 4.00996953e-01 1.20647989e-01 6.83529735e-01
5.80025792e-01 -3.98414098e-02 2.50766337e-01 -2.22847506e-01
4.23267893e-02 -2.55738914e-01 -8.62542868e-01 -2.19406590e-01
-3.26053619e-01 -4.21415389e-01 6.19930208e-01 -7.52754092e-01
-5.50385177e-01 -2.93041840e-02 -2.38571435e-01 -1.08102405e+00
2.40064010e-01 2.39414811e-01 2.45781451e-01 -2.73632020e-01
7.56505057e-02 2.88270891e-01 9.17204916e-02 -5.87490439e-01
-1.18256204e-01 4.23430830e-01 4.66728449e-01 -3.18771869e-01
5.33683121e-01 7.98136771e-01 8.10035169e-02 -8.54536474e-01
-8.06941688e-01 -5.18142939e-01 -4.94735658e-01 -5.98508120e-01
7.39540040e-01 -1.29499984e+00 -1.32020974e+00 8.96948934e-01
-1.28225791e+00 -5.42658508e-01 1.58781394e-01 7.50963509e-01
-3.26146394e-01 2.29359254e-01 -9.34388340e-01 -5.76830208e-01
-2.31750891e-01 -1.16033256e+00 1.13444114e+00 2.31039345e-01
1.39011234e-01 -8.27230692e-01 -5.35626769e-01 3.03371519e-01
4.93755966e-01 -9.76501033e-02 4.96821702e-01 -2.41713494e-01
-1.15222454e+00 -1.95207149e-01 -5.01881003e-01 2.69991070e-01
-2.32507586e-01 1.88875988e-01 -8.27139795e-01 -3.53756487e-01
-3.96301001e-02 -7.55128637e-02 9.92869914e-01 4.85817164e-01
1.40432489e+00 -1.57077968e-01 -2.52023429e-01 8.62893581e-01
1.23791707e+00 5.00107586e-01 4.67580080e-01 2.85781235e-01
6.76916480e-01 1.23420797e-01 5.47933340e-01 5.81261337e-01
4.94650602e-01 9.64409411e-01 5.69241941e-01 -6.89248294e-02
-5.32900766e-02 -1.01213545e-01 3.83527875e-01 1.37232900e+00
-2.96417028e-01 -3.37856412e-01 -9.84580457e-01 3.99065435e-01
-2.08436561e+00 -9.59058940e-01 -2.87421316e-01 2.06877232e+00
4.68929917e-01 4.05860752e-01 1.59995824e-01 -3.66177522e-02
6.99161828e-01 2.30489403e-01 -5.83997130e-01 -1.18657306e-01
-6.30350709e-02 7.85354804e-03 5.97611964e-01 3.66465241e-01
-9.95871663e-01 1.00304508e+00 6.30211592e+00 9.28587317e-01
-1.42359436e+00 3.16815317e-01 5.46113193e-01 -6.21709347e-01
1.78733319e-01 -8.49976689e-02 -1.00859034e+00 4.73663509e-01
9.48674023e-01 -1.00358553e-01 3.47597748e-01 9.06933844e-01
3.88379127e-01 2.37407144e-02 -1.10690308e+00 1.59902644e+00
-2.08722487e-01 -1.93786156e+00 -2.55194791e-02 7.12774470e-02
5.41440248e-01 2.99290001e-01 -8.58637467e-02 3.48127112e-02
-3.50885630e-01 -7.36096859e-01 8.98727655e-01 5.34496665e-01
9.20772672e-01 -6.70906246e-01 6.28610730e-01 3.89286071e-01
-1.48859084e+00 -1.16157196e-01 -4.28297192e-01 -5.12702882e-01
1.66665122e-01 6.92743182e-01 -5.21075547e-01 1.41093910e-01
1.04693043e+00 1.06092918e+00 -1.29434258e-01 9.39106286e-01
2.98536271e-01 6.02791727e-01 -4.18998420e-01 -2.40226448e-01
4.57780659e-01 -5.16509861e-02 4.34076399e-01 1.14233267e+00
5.22154450e-01 2.89287299e-01 1.63536191e-01 3.69528532e-01
-2.52122313e-01 -3.72627974e-01 -5.58325171e-01 -1.42436013e-01
6.02645636e-01 9.36454356e-01 -7.01055229e-01 -4.18715507e-01
-7.20905185e-01 9.97939944e-01 6.73689917e-02 1.79495707e-01
-1.09998679e+00 -1.93783402e-01 9.93104994e-01 1.16644047e-01
4.82438982e-01 -9.06204104e-01 9.81871113e-02 -1.32821989e+00
2.59088069e-01 -6.91091835e-01 7.46286958e-02 -7.04683363e-01
-8.04032207e-01 4.32452649e-01 -4.22443449e-01 -1.57072484e+00
2.08960250e-01 -8.99966538e-01 -3.87877613e-01 2.49630496e-01
-1.34522712e+00 -8.51401389e-01 -6.06867611e-01 6.74243629e-01
7.22232401e-01 -1.05736278e-01 4.40372139e-01 7.97012329e-01
-7.29287803e-01 5.97235501e-01 8.55000839e-02 3.26372415e-01
3.95039827e-01 -6.74069047e-01 8.44686210e-01 7.11910009e-01
-8.43898356e-02 4.51745480e-01 3.60310763e-01 -3.45368475e-01
-2.04030347e+00 -1.20074081e+00 9.00613546e-01 -1.34835750e-01
7.22228646e-01 -6.28243208e-01 -7.32477725e-01 6.20922565e-01
-2.72901714e-01 2.71991819e-01 4.15280968e-01 -2.02297732e-01
-5.24735630e-01 -3.62533063e-01 -5.71071208e-01 6.49556935e-01
1.38492346e+00 -8.20376396e-01 2.08485201e-01 5.17181218e-01
1.13255084e+00 -7.63137758e-01 -8.16660702e-01 2.47225255e-01
6.87513292e-01 -1.10819304e+00 9.42535043e-01 -3.55586857e-01
4.23316002e-01 -3.31252843e-01 -2.35503152e-01 -5.56599259e-01
-3.40612292e-01 -1.03287625e+00 -4.97853845e-01 8.06346595e-01
1.79716125e-01 -6.52557373e-01 1.03758025e+00 4.53738391e-01
-2.36638948e-01 -9.88809764e-01 -1.25562358e+00 -1.14866805e+00
-4.22031879e-01 -1.10218906e+00 4.47609901e-01 7.62750328e-01
-1.83673695e-01 3.66305932e-02 -5.72789848e-01 2.72819638e-01
5.31334996e-01 9.16570872e-02 8.13768268e-01 -8.05846035e-01
-2.14996621e-01 -3.86634976e-01 -8.25179577e-01 -1.91795588e+00
-2.43431985e-01 -5.05344272e-01 -3.72837812e-01 -1.04780400e+00
9.13194269e-02 -3.02378386e-01 -4.76060174e-02 4.19512004e-01
5.69358289e-01 5.47407806e-01 3.58452469e-01 3.14478964e-01
-9.98019218e-01 4.69883174e-01 1.32232749e+00 -8.62857699e-02
2.35130023e-02 -4.08116408e-04 -1.77442610e-01 6.28115833e-01
8.36585402e-01 -3.08686703e-01 -1.59607321e-01 -7.98394799e-01
2.67107159e-01 3.52097988e-01 7.48336911e-01 -1.24069583e+00
6.69528902e-01 9.74778086e-02 1.58347979e-01 -7.74018824e-01
6.44309938e-01 -5.45919657e-01 9.73479599e-02 4.96153533e-01
-6.12673089e-02 3.00607443e-01 2.74215221e-01 4.53616887e-01
-4.37453359e-01 2.10959643e-01 7.27151394e-01 -1.58618614e-01
-9.02035892e-01 9.79902208e-01 -4.85937208e-01 -1.75488576e-01
9.02772069e-01 -3.10699761e-01 -1.96046039e-01 -5.69575608e-01
-6.71381593e-01 2.69251019e-01 3.10784250e-01 5.48981071e-01
7.51292527e-01 -1.46548021e+00 -3.02506238e-01 1.99974388e-01
-2.11516589e-01 5.35384528e-02 6.77778661e-01 1.05506790e+00
-7.31730163e-01 7.57758737e-01 6.78104907e-02 -8.95853877e-01
-1.43455207e+00 2.95679331e-01 4.26759422e-01 1.36977091e-01
-7.44905412e-01 1.03982103e+00 1.70813397e-01 -7.70834684e-02
4.84986573e-01 -4.70289797e-01 2.58248001e-01 -7.85461515e-02
8.79338562e-01 4.63818073e-01 1.64216623e-01 -2.26198435e-01
-4.00761753e-01 6.59974515e-01 -4.31885161e-02 1.19937383e-01
1.40087557e+00 -1.67542621e-01 1.20587600e-02 3.47886741e-01
1.52702665e+00 -5.81225514e-01 -1.54513407e+00 -2.86847353e-01
-3.10489565e-01 -5.21695971e-01 2.52671987e-01 -2.53252555e-02
-1.34796989e+00 1.11872458e+00 4.69783634e-01 2.39294589e-01
1.18270516e+00 8.87114480e-02 1.22805440e+00 4.16953176e-01
5.57148635e-01 -1.01960170e+00 2.55501300e-01 9.96344090e-01
5.36730766e-01 -1.12158012e+00 -2.40655541e-01 -5.05932927e-01
-1.92983240e-01 1.33423507e+00 7.88489699e-01 -5.64729311e-02
8.44386637e-01 5.27728558e-01 -9.15716812e-02 -2.58889943e-01
-1.30038202e+00 7.83393066e-03 1.75456479e-02 4.19553280e-01
2.14655384e-01 -9.04672965e-02 2.07042277e-01 4.07130480e-01
1.15715474e-01 1.55614048e-01 4.18549448e-01 8.84648383e-01
-2.21249402e-01 -7.41020441e-01 9.36678424e-02 3.85333359e-01
-4.31486279e-01 -1.97620735e-01 1.16434656e-01 7.20088184e-01
9.66544077e-02 6.68786764e-01 4.96245831e-01 -8.23697984e-01
2.71098971e-01 -3.64601701e-01 3.84121150e-01 -1.59160018e-01
-2.99029708e-01 1.52044117e-01 6.16877861e-02 -1.13954735e+00
-5.06616414e-01 -6.16702199e-01 -1.21496606e+00 -7.53059983e-01
-2.56184071e-01 -9.40797552e-02 6.91858351e-01 8.09301436e-01
7.65189171e-01 2.37291873e-01 6.67601824e-01 -1.18291664e+00
-1.48357674e-01 -4.78714436e-01 -3.67238939e-01 -1.76656857e-01
5.64173877e-01 -4.80017632e-01 -1.60455659e-01 1.10351287e-01] | [8.975263595581055, 0.14004170894622803] |
77e04cc9-72be-41d4-aa7f-9f8cca0599de | label-information-enhanced-fraud-detection | 2302.10407 | null | https://arxiv.org/abs/2302.10407v1 | https://arxiv.org/pdf/2302.10407v1.pdf | Label Information Enhanced Fraud Detection against Low Homophily in Graphs | Node classification is a substantial problem in graph-based fraud detection. Many existing works adopt Graph Neural Networks (GNNs) to enhance fraud detectors. While promising, currently most GNN-based fraud detectors fail to generalize to the low homophily setting. Besides, label utilization has been proved to be significant factor for node classification problem. But we find they are less effective in fraud detection tasks due to the low homophily in graphs. In this work, we propose GAGA, a novel Group AGgregation enhanced TrAnsformer, to tackle the above challenges. Specifically, the group aggregation provides a portable method to cope with the low homophily issue. Such an aggregation explicitly integrates the label information to generate distinguishable neighborhood information. Along with group aggregation, an attempt towards end-to-end trainable group encoding is proposed which augments the original feature space with the class labels. Meanwhile, we devise two additional learnable encodings to recognize the structural and relational context. Then, we combine the group aggregation and the learnable encodings into a Transformer encoder to capture the semantic information. Experimental results clearly show that GAGA outperforms other competitive graph-based fraud detectors by up to 24.39% on two trending public datasets and a real-world industrial dataset from Anonymous. Even more, the group aggregation is demonstrated to outperform other label utilization methods (e.g., C&S, BoT/UniMP) in the low homophily setting. | ['Junzhou Luo', 'Beilun Wang', 'Jiahui Jin', 'Fang Dong', 'dianhai yu', 'Yu Sun', 'Ziheng Ma', 'Shikun Feng', 'Weibin Li', 'Zhengjie Huang', 'Jinghui Zhang', 'Yuchen Wang'] | 2023-02-21 | null | null | null | null | ['fraud-detection'] | ['miscellaneous'] | [-2.00781114e-02 -1.16838552e-01 -2.85625786e-01 -3.68285209e-01
-2.80063957e-01 -2.22005874e-01 1.97993174e-01 4.75107759e-01
-9.17297974e-02 6.15588725e-01 -1.94619998e-01 -1.98725641e-01
-1.92721531e-01 -1.35748088e+00 -6.22076571e-01 -4.44342017e-01
-3.08739632e-01 3.86916071e-01 3.89432684e-02 -4.58418041e-01
1.13153517e-01 2.47754350e-01 -1.16355181e+00 2.78985083e-01
1.09664035e+00 1.25834763e+00 -2.79731274e-01 5.31194061e-02
-1.31477356e-01 8.75599742e-01 -4.09963459e-01 -9.94989514e-01
3.33359122e-01 -3.28802466e-01 -4.85094517e-01 3.05323792e-03
3.97846758e-01 -4.31191415e-01 -4.05968785e-01 1.19149911e+00
2.74824321e-01 -2.04678729e-01 5.29098094e-01 -1.74887431e+00
-7.70672381e-01 7.95035481e-01 -7.63148069e-01 8.53118002e-02
2.83768862e-01 -1.31164551e-01 1.34219599e+00 -6.87691748e-01
7.65812993e-01 1.28054500e+00 1.32778168e+00 3.00319940e-01
-7.63041139e-01 -1.04357290e+00 2.33062044e-01 2.68809438e-01
-1.22697330e+00 1.47111356e-01 1.20122290e+00 -7.06265792e-02
8.90801013e-01 3.78608778e-02 7.53444672e-01 1.40599740e+00
1.33275166e-01 7.71194816e-01 8.51213396e-01 -1.79327309e-01
1.08047657e-01 -3.73494998e-02 3.41266632e-01 1.39068758e+00
1.06598520e+00 -1.15103245e-01 -7.40513504e-01 -5.15923500e-01
6.47966087e-01 6.39730573e-01 -4.25724462e-02 -5.55597484e-01
-6.93124533e-01 1.08061743e+00 9.83897746e-01 1.88675389e-01
-3.49720001e-01 4.20940369e-01 9.05268550e-01 7.81065762e-01
5.38076222e-01 3.81055027e-01 -1.96176633e-01 3.00560594e-01
-6.37103617e-01 1.99063763e-01 8.55247915e-01 9.54787195e-01
7.84219503e-01 7.26488084e-02 -2.53878143e-02 6.43963516e-01
4.78272051e-01 -2.31478661e-02 4.97630209e-01 -6.00760162e-01
9.23058510e-01 1.34706485e+00 -3.61208230e-01 -1.66471744e+00
-4.96946514e-01 -8.04177344e-01 -1.24407232e+00 -2.85175592e-01
2.99588919e-01 1.63394883e-01 -7.93447673e-01 1.69757700e+00
2.21957698e-01 3.27825397e-01 2.58434415e-02 6.50016069e-01
6.90177858e-01 8.88054892e-02 6.01152666e-02 4.42779660e-02
1.28766477e+00 -9.75194335e-01 -6.26161635e-01 -7.51517341e-02
1.12883210e+00 -1.21771738e-01 8.15497041e-01 2.71009892e-01
-5.87274134e-01 -1.65234789e-01 -1.31345391e+00 -9.52293724e-02
-6.73678100e-01 -5.97634427e-02 1.21056843e+00 9.06591713e-01
-8.14665973e-01 8.51213992e-01 -7.86357224e-01 -5.16704261e-01
7.57404149e-01 3.03199470e-01 -4.98681664e-01 -3.57389212e-01
-1.51108170e+00 4.36294734e-01 4.75168794e-01 1.23507217e-01
-5.49558461e-01 -2.96450168e-01 -1.09649777e+00 1.52003556e-01
4.67076004e-01 -8.93409908e-01 6.52476490e-01 -8.31682503e-01
-8.58074605e-01 7.96534836e-01 2.43211582e-01 -9.91540492e-01
8.19659770e-01 2.39451021e-01 -3.50800216e-01 1.46101117e-01
3.72349203e-01 1.42355800e-01 7.00958610e-01 -1.01316261e+00
-5.70896864e-01 -6.19594932e-01 8.16921592e-02 2.19251871e-01
-9.66426909e-01 -2.17835739e-01 -1.68017551e-01 -8.77589643e-01
3.75838995e-01 -6.48122549e-01 1.02984468e-02 1.91360712e-01
-5.48737168e-01 -4.59435046e-01 1.06609774e+00 -5.53985775e-01
1.35139501e+00 -1.85029638e+00 -1.76616862e-01 6.50030434e-01
5.88566601e-01 1.70550361e-01 2.21301287e-01 6.19194567e-01
8.63222480e-02 1.60180643e-01 -5.01776576e-01 -5.76570272e-01
1.46006256e-01 3.01700264e-01 -1.90148260e-02 5.65266788e-01
2.97866970e-01 8.63605201e-01 -1.12240255e+00 -5.51176965e-01
6.08134549e-03 1.93362936e-01 -6.52433455e-01 -8.86839107e-02
1.22005448e-01 -2.03189895e-01 -5.24039209e-01 1.36824143e+00
7.31719971e-01 -3.84664357e-01 6.05731368e-01 -1.08302638e-01
7.31644869e-01 3.93645018e-02 -1.00488949e+00 1.60345602e+00
-1.26690120e-01 1.83268607e-01 -5.46657257e-02 -1.54725099e+00
1.30562365e+00 7.35365599e-02 5.49695194e-01 -6.46341920e-01
9.60790738e-02 6.53510571e-01 -1.23829730e-02 -2.10508540e-01
4.54971015e-01 -6.64408132e-02 -4.28796619e-01 3.37765962e-01
2.22512901e-01 4.55080420e-01 2.56841868e-01 4.90597248e-01
1.53794694e+00 -1.98002368e-01 2.31557935e-01 -9.69767943e-02
4.03459400e-01 -2.07986906e-01 8.85998607e-01 8.03299069e-01
-5.66404879e-01 4.32459533e-01 7.47519314e-01 -6.73821747e-01
-6.35258794e-01 -7.68553793e-01 1.62031174e-01 8.29404175e-01
4.65140373e-01 -7.63148785e-01 -5.63479662e-01 -1.25934923e+00
6.66636169e-01 1.87687829e-01 -6.91318870e-01 -5.88708818e-01
-6.16499066e-01 -1.14842391e+00 8.72712851e-01 6.05188966e-01
9.91740048e-01 -9.78980780e-01 -1.65621951e-01 2.44099021e-01
-3.49752516e-01 -1.03129637e+00 -2.94360578e-01 2.03397006e-01
-9.58550990e-01 -1.47301435e+00 -2.31452629e-01 -8.60155761e-01
6.02670431e-01 3.76984835e-01 1.09053612e+00 6.59946561e-01
-3.41683567e-01 -1.86926723e-02 -4.36915576e-01 -1.86362267e-01
-1.55131519e-01 2.05274403e-01 1.05868682e-01 8.98774713e-02
5.97112477e-01 -8.52563024e-01 -5.34710228e-01 3.61547112e-01
-7.17673004e-01 -1.53876692e-01 4.26079690e-01 1.13689470e+00
3.45603883e-01 2.80371100e-01 1.00354588e+00 -1.37111068e+00
8.41028452e-01 -9.34941292e-01 -1.83396325e-01 2.62478501e-01
-1.08017898e+00 -1.98336169e-01 8.02638113e-01 -3.03554479e-02
-6.85822606e-01 -2.67118335e-01 7.94911832e-02 -4.76052552e-01
4.27094162e-01 7.78110862e-01 -1.54268384e-01 -3.05424929e-01
6.50877118e-01 -9.52288136e-02 -7.13081509e-02 -3.91008139e-01
-8.33141431e-02 6.04820490e-01 3.39822471e-01 -5.47156036e-01
7.74915278e-01 5.31153619e-01 1.99273929e-01 -2.54601330e-01
-4.53037202e-01 -5.97535908e-01 -3.07236075e-01 1.83418587e-01
3.17049921e-01 -1.01671410e+00 -5.42949200e-01 7.75403321e-01
-8.26024115e-01 8.20874125e-02 -8.63492042e-02 -2.43587177e-02
-4.14239615e-01 8.53113770e-01 -1.09418809e+00 -7.76045978e-01
-7.80071795e-01 -8.29216838e-01 9.72076297e-01 -6.39873222e-02
3.75115648e-02 -1.00871205e+00 -4.40643668e-01 4.55351651e-01
1.57735169e-01 8.21580410e-01 8.55597198e-01 -8.07700217e-01
-5.31213343e-01 -4.18398172e-01 -5.80894291e-01 2.85701811e-01
2.11321503e-01 -3.40699375e-01 -4.92118865e-01 -5.68422019e-01
-1.77410841e-01 -4.50549066e-01 1.28342056e+00 -1.31947249e-01
1.12950301e+00 -3.23613465e-01 -5.19475639e-01 7.99596488e-01
1.54287636e+00 -1.84700452e-02 4.45404768e-01 4.45876092e-01
1.16341460e+00 3.90179634e-01 4.94823039e-01 2.97121674e-01
6.17254198e-01 3.98169726e-01 6.95755243e-01 -3.18330117e-02
-9.64019820e-02 -7.14099348e-01 3.48289400e-01 8.90328348e-01
-6.72969129e-03 -3.93707186e-01 -8.09388816e-01 4.96515334e-01
-2.24952364e+00 -7.41691947e-01 -2.00010955e-01 1.83356357e+00
4.91174489e-01 3.31180274e-01 2.18182817e-01 4.80826527e-01
8.51670742e-01 -2.30702888e-02 -6.58411920e-01 -2.19799682e-01
-3.21408153e-01 -1.26989419e-02 7.47531474e-01 -6.25720695e-02
-1.16154218e+00 8.60558629e-01 4.94647884e+00 7.65257597e-01
-6.03862226e-01 1.76791266e-01 4.59378242e-01 4.40502435e-01
-2.58865833e-01 -9.58706960e-02 -5.11745989e-01 7.30278015e-01
7.12553382e-01 -1.27505511e-01 1.88537985e-01 9.44140732e-01
-4.07232195e-01 3.15671742e-01 -9.49168324e-01 9.99280572e-01
1.23418503e-01 -1.21879375e+00 2.28863716e-01 1.28870726e-01
5.72166979e-01 -9.00238082e-02 -2.22903311e-01 5.89644909e-01
4.06515181e-01 -8.72515261e-01 3.80294085e-01 2.80674666e-01
6.40772879e-01 -8.48117888e-01 1.08737254e+00 2.19154537e-01
-1.29750872e+00 -5.09559631e-01 -4.57225442e-01 -1.66975960e-01
-2.00170860e-01 5.70804477e-01 -7.94750929e-01 1.15054727e+00
7.98476815e-01 1.18550444e+00 -7.87468314e-01 9.33122575e-01
1.84517335e-02 5.90702176e-01 -3.87685716e-01 3.42067070e-02
2.84911543e-01 -2.97874421e-01 3.13341886e-01 1.14670038e+00
4.37698096e-01 -3.37414473e-01 4.27590400e-01 6.83712304e-01
-8.08844149e-01 7.87942111e-02 -9.22554731e-01 -2.66173761e-03
4.73682821e-01 1.23821461e+00 -8.49038005e-01 -3.05156887e-01
-3.11982721e-01 1.12622380e+00 7.06615567e-01 9.60635394e-02
-6.70199394e-01 -6.17466748e-01 4.31120962e-01 1.79220855e-01
1.29737690e-01 1.64005533e-01 -1.69576347e-01 -1.48272872e+00
3.34082991e-01 -8.43409956e-01 9.40879822e-01 -4.03919935e-01
-1.80038941e+00 4.40304488e-01 -3.06845993e-01 -1.14169252e+00
-1.71423778e-01 -7.32877791e-01 -5.89577734e-01 4.58038330e-01
-1.63325596e+00 -1.50486505e+00 -5.08207262e-01 7.25253701e-01
1.70596436e-01 -4.12441581e-01 7.53886640e-01 6.97045326e-01
-7.94933736e-01 1.06727600e+00 7.17091709e-02 5.34221053e-01
6.27713382e-01 -1.42939484e+00 4.45478529e-01 7.88937509e-01
1.19381383e-01 6.01333082e-01 2.75097430e-01 -1.01028109e+00
-1.42106509e+00 -1.33282375e+00 7.45760083e-01 -2.75416136e-01
7.03834832e-01 -6.47342503e-01 -1.08607876e+00 7.86218345e-01
-2.85962313e-01 3.78598899e-01 4.30986464e-01 2.45874211e-01
-7.65200734e-01 -3.50390077e-01 -1.62981248e+00 1.95595220e-01
1.65805101e+00 -7.02458322e-01 -2.43506625e-01 5.32216251e-01
6.21704042e-01 -3.08942676e-01 -9.04808998e-01 5.65804243e-01
3.54520172e-01 -1.24476314e+00 8.38798344e-01 -4.92106736e-01
3.82930458e-01 -1.08715080e-01 -1.24329410e-01 -1.13039279e+00
-2.53491223e-01 -3.69310409e-01 -5.36353707e-01 1.33757579e+00
1.60911388e-03 -1.19354475e+00 1.27162027e+00 9.45134833e-02
-3.39228511e-02 -8.61862361e-01 -9.02987659e-01 -8.53639245e-01
-2.28389487e-01 -1.89785987e-01 6.46100461e-01 1.55855417e+00
1.27540454e-01 1.81102082e-01 -4.48919982e-01 -9.41014662e-02
9.76070166e-01 2.53557503e-01 4.80947852e-01 -1.70831299e+00
-8.63165110e-02 -3.18767250e-01 -7.87792504e-01 -3.07082981e-01
1.83508173e-01 -1.36144352e+00 -5.61390519e-01 -1.49683666e+00
2.13118136e-01 -7.49208808e-01 -8.09591472e-01 6.68168843e-01
-8.76569822e-02 3.67843419e-01 8.01605508e-02 3.68097752e-01
-6.43840253e-01 4.12418544e-01 9.07248616e-01 -3.70235920e-01
8.39959085e-02 1.17908595e-02 -9.56745863e-01 6.00632906e-01
8.62606347e-01 -5.06910443e-01 -4.22507107e-01 -1.11028284e-01
3.58434200e-01 -1.61575377e-01 5.27347505e-01 -9.72923934e-01
1.20002486e-01 2.60130495e-01 3.04305524e-01 -4.23973829e-01
4.42811735e-02 -7.69897938e-01 -1.00232912e-02 6.63654983e-01
-2.56055146e-02 2.86511838e-01 -2.88352102e-01 1.13612151e+00
-2.84516960e-01 -1.63388923e-01 3.15930218e-01 -2.11293265e-01
-6.59723938e-01 5.09018958e-01 2.94050097e-01 -9.34117585e-02
8.08570564e-01 -3.75544578e-01 -5.05690992e-01 -2.02885240e-01
-2.72173375e-01 5.55645704e-01 4.67882603e-01 4.32557702e-01
5.75166047e-01 -1.52837086e+00 -5.86420476e-01 2.91027904e-01
3.28199536e-01 -3.06163039e-02 1.23518519e-01 6.72350824e-01
-5.66075683e-01 8.29961598e-02 -3.65359098e-01 -3.24288130e-01
-1.07359636e+00 4.76609051e-01 3.86195719e-01 -8.91869247e-01
-7.01391518e-01 7.26468742e-01 -1.65667921e-01 -5.87955117e-01
2.80158281e-01 -3.95048112e-01 1.12409971e-03 1.73788685e-02
-5.50390854e-02 4.77998167e-01 2.61015356e-01 -4.49563652e-01
-4.04913038e-01 1.62942797e-01 -2.41988853e-01 7.80687988e-01
1.25054514e+00 1.32617325e-01 -1.79378182e-01 6.95089549e-02
1.28913784e+00 -2.36092478e-01 -7.40522861e-01 -3.98428947e-01
2.87587762e-01 -4.24872428e-01 -2.24447668e-01 -6.40255570e-01
-1.30322170e+00 7.39861369e-01 4.35527414e-01 5.75067580e-01
1.09845829e+00 -4.99475986e-01 1.18037665e+00 6.46137536e-01
7.83445179e-01 -1.11916339e+00 1.27671316e-01 1.39194131e-01
5.91869771e-01 -1.42607951e+00 7.01781139e-02 -7.80870378e-01
-3.92789125e-01 9.86034393e-01 8.15516531e-01 -5.16363263e-01
4.44426596e-01 -1.46361426e-01 -3.62070024e-01 -5.06216526e-01
-4.49091315e-01 3.02752275e-02 -1.65872753e-01 6.22527659e-01
5.83839305e-02 6.93365559e-02 -4.40300852e-01 7.54759312e-01
-3.77495140e-02 6.44983798e-02 4.36046064e-01 7.58419693e-01
-8.05858597e-02 -1.17151511e+00 -3.70671733e-05 9.63368416e-01
-4.43181485e-01 4.97314855e-02 -5.71326256e-01 8.52552176e-01
3.32262784e-01 6.65619850e-01 -1.86448738e-01 -6.22197926e-01
3.27169895e-01 3.91574949e-02 2.09192649e-01 -4.53255981e-01
-9.44594860e-01 -5.12007356e-01 3.65103751e-01 -5.46414375e-01
-2.35980898e-01 -4.71329004e-01 -1.13514292e+00 -5.01107514e-01
-4.61325526e-01 8.30164328e-02 1.92542180e-01 6.43144846e-01
3.52869958e-01 4.80893195e-01 5.22374928e-01 -1.74147516e-01
-6.91031694e-01 -8.77670527e-01 -8.47301185e-01 9.18524921e-01
9.62820351e-02 -9.41656411e-01 -4.53436852e-01 -7.55607188e-01] | [7.100923538208008, 5.965726852416992] |
60be8c6f-4cc9-47de-8871-595912266d69 | pretraining-without-wordpieces-learning-over | 2202.12142 | null | https://arxiv.org/abs/2202.12142v1 | https://arxiv.org/pdf/2202.12142v1.pdf | Pretraining without Wordpieces: Learning Over a Vocabulary of Millions of Words | The standard BERT adopts subword-based tokenization, which may break a word into two or more wordpieces (e.g., converting "lossless" to "loss" and "less"). This will bring inconvenience in following situations: (1) what is the best way to obtain the contextual vector of a word that is divided into multiple wordpieces? (2) how to predict a word via cloze test without knowing the number of wordpieces in advance? In this work, we explore the possibility of developing BERT-style pretrained model over a vocabulary of words instead of wordpieces. We call such word-level BERT model as WordBERT. We train models with different vocabulary sizes, initialization configurations and languages. Results show that, compared to standard wordpiece-based BERT, WordBERT makes significant improvements on cloze test and machine reading comprehension. On many other natural language understanding tasks, including POS tagging, chunking and NER, WordBERT consistently performs better than BERT. Model analysis indicates that the major advantage of WordBERT over BERT lies in the understanding for low-frequency words and rare words. Furthermore, since the pipeline is language-independent, we train WordBERT for Chinese language and obtain significant gains on five natural language understanding datasets. Lastly, the analyse on inference speed illustrates WordBERT has comparable time cost to BERT in natural language understanding tasks. | ['Shuming Shi', 'Yunbo Cao', 'Bing Qin', 'Xiaocheng Feng', 'Shuangzhi Wu', 'Junwei Liao', 'Cong Zhou', 'Duyu Tang', 'Zhangyin Feng'] | 2022-02-24 | null | null | null | null | ['cloze-test'] | ['natural-language-processing'] | [-1.18172124e-01 2.25027487e-01 -2.16814920e-01 -4.87678885e-01
-5.92415094e-01 -6.45676613e-01 2.90167063e-01 4.33092386e-01
-1.07737422e+00 7.53777921e-01 2.95245916e-01 -8.13034475e-01
1.71747789e-01 -8.89759183e-01 -6.98774695e-01 -2.14309171e-01
1.97619051e-01 6.63419485e-01 1.85319617e-01 -5.10630250e-01
4.02132720e-01 3.53593901e-02 -1.05910552e+00 2.91220665e-01
1.11597753e+00 5.61674416e-01 7.55895197e-01 5.07900298e-01
-6.92107260e-01 5.15202045e-01 -9.34277117e-01 -5.08171797e-01
5.12103103e-02 -8.68405178e-02 -1.36880171e+00 -3.00764501e-01
1.61231637e-01 -3.92123699e-01 2.99762696e-01 8.90093684e-01
2.47374088e-01 2.35553697e-01 5.60199738e-01 -6.49342239e-01
-7.60866523e-01 1.23173189e+00 -4.42729175e-01 3.91957253e-01
2.41102695e-01 7.24387988e-02 1.28155720e+00 -8.19379210e-01
1.97417200e-01 1.42406607e+00 6.15073025e-01 5.01703799e-01
-1.05069840e+00 -7.57983267e-01 4.89608824e-01 2.95201182e-01
-1.46052539e+00 -2.23487869e-01 1.47307709e-01 -3.90714675e-01
1.71393085e+00 9.11590084e-02 3.78607869e-01 8.25857639e-01
2.86945164e-01 8.88541639e-01 8.84565175e-01 -8.99163485e-01
-9.65592917e-03 4.59133089e-02 8.38178754e-01 5.33437550e-01
2.73078471e-01 -1.12463273e-01 -6.23181403e-01 2.62852848e-01
4.89015609e-01 -1.60665318e-01 -1.74120247e-01 4.69694853e-01
-1.15547276e+00 1.03756130e+00 1.59845024e-01 3.89433950e-01
-2.32933283e-01 1.89296961e-01 4.97491568e-01 2.57181048e-01
6.49329782e-01 9.33280170e-01 -1.01156366e+00 -5.11875689e-01
-9.27636445e-01 1.36213750e-01 8.64494205e-01 1.07711124e+00
9.93342638e-01 1.39447972e-01 -2.63308614e-01 9.74367082e-01
1.94331855e-01 5.45928538e-01 8.39341640e-01 -4.23782080e-01
6.19617999e-01 2.86880910e-01 -3.57127190e-02 -5.52748024e-01
-4.44459677e-01 -2.30329320e-01 -5.45140684e-01 -3.21503580e-01
4.20193493e-01 -3.63928109e-01 -1.30757797e+00 1.74349320e+00
-2.36008391e-01 9.63308662e-02 1.92065183e-02 5.91304004e-01
1.09974110e+00 7.94079483e-01 5.41505516e-01 -3.54644470e-02
1.79829645e+00 -1.00710893e+00 -8.16780567e-01 -7.64235735e-01
9.06613410e-01 -7.13089049e-01 1.50256693e+00 5.46764374e-01
-9.60447490e-01 -6.43979907e-01 -9.23179388e-01 -3.35225791e-01
-7.79207289e-01 1.54347066e-02 8.13048482e-01 8.34510922e-01
-9.10214067e-01 3.82917255e-01 -8.81283164e-01 -3.19816351e-01
4.93367352e-02 2.89446265e-01 -2.27446541e-01 -1.88045278e-01
-1.62330830e+00 1.24976659e+00 8.31445634e-01 -1.85248330e-01
-6.75196350e-01 -8.28754663e-01 -1.22371364e+00 3.10819268e-01
3.22193742e-01 -4.10109937e-01 1.41761279e+00 -7.18924463e-01
-1.32637417e+00 7.34509230e-01 -6.63531125e-01 -6.12951815e-01
-1.04779720e-01 -6.56991839e-01 -2.49405116e-01 -2.50511140e-01
2.73289293e-01 8.25644314e-01 5.78913450e-01 -7.27520943e-01
-6.61473036e-01 -8.47687479e-03 1.53114513e-01 2.03165278e-01
-3.72883379e-01 1.50157809e-01 -3.81574988e-01 -5.64843357e-01
-1.76527113e-01 -4.52059031e-01 2.11944394e-02 -7.84973562e-01
-2.74928600e-01 -8.08609486e-01 1.39831692e-01 -8.56022239e-01
1.44912839e+00 -2.00290728e+00 -1.37700617e-01 -1.03909820e-01
1.22645684e-01 6.23833776e-01 -3.51706028e-01 5.83621621e-01
2.45369039e-02 7.34059036e-01 5.38733639e-02 -4.36457992e-01
-9.76708159e-02 6.10393822e-01 -4.48505700e-01 -7.40742162e-02
3.32916021e-01 1.09777439e+00 -9.64684129e-01 -3.63266706e-01
1.28705040e-01 1.73854575e-01 -6.42657459e-01 1.55977249e-01
-3.24521065e-01 -4.69114482e-02 -1.46587670e-01 3.54215026e-01
6.84678793e-01 -5.99042736e-02 -8.84447247e-02 7.36268014e-02
-1.77341476e-01 1.00316417e+00 -7.79798687e-01 1.62962103e+00
-9.81338143e-01 5.49262285e-01 -3.33970487e-01 -9.03632462e-01
8.22794139e-01 2.87716001e-01 -3.07720453e-01 -5.56506038e-01
1.15507066e-01 3.66446413e-02 2.04166725e-01 -5.23330986e-01
9.41911995e-01 -4.53987271e-01 -8.28456134e-02 2.91694552e-01
3.56258422e-01 -3.29483300e-01 2.65176237e-01 2.68074423e-01
1.07171547e+00 -1.13135055e-01 5.62012434e-01 -4.26730067e-01
8.21276680e-02 1.63683847e-01 5.35298944e-01 8.30069840e-01
-7.00222850e-02 5.18917143e-01 4.49449956e-01 -7.95019604e-03
-6.91545308e-01 -1.07066143e+00 -1.57886729e-01 1.56487191e+00
1.13338053e-01 -7.49465644e-01 -6.67673469e-01 -6.58208370e-01
-9.96406600e-02 1.37371933e+00 -3.83252203e-01 -7.42845610e-02
-7.13611305e-01 -4.49760020e-01 8.00126910e-01 7.60110378e-01
4.50044572e-01 -1.28811765e+00 -4.34912086e-01 4.13053632e-01
-3.56345862e-01 -1.01509857e+00 -5.58118641e-01 4.67071533e-01
-6.06187165e-01 -6.52475595e-01 -3.45400482e-01 -1.03794885e+00
5.29324770e-01 2.60427684e-01 1.60760367e+00 3.33788872e-01
-3.16699259e-02 8.44832063e-02 -9.43554223e-01 -8.58384430e-01
-7.36765638e-02 4.17013407e-01 -9.76089537e-02 -6.77222431e-01
9.57812607e-01 4.00236100e-02 -3.30870599e-01 1.60748556e-01
-9.52144563e-01 6.64407462e-02 5.19733667e-01 1.14655209e+00
2.14321941e-01 2.57087182e-02 7.16740727e-01 -1.06854200e+00
8.55068326e-01 -4.69954103e-01 -2.54709512e-01 2.75507897e-01
-6.16230547e-01 1.93853095e-01 8.18198919e-01 -4.82956737e-01
-8.93691242e-01 -3.77047062e-01 -6.60811901e-01 1.78830117e-01
-1.97004855e-01 8.20785999e-01 -1.39656276e-01 5.30178308e-01
6.50960207e-01 1.56013384e-01 -8.43386799e-02 -6.44085884e-01
5.62914550e-01 8.01804066e-01 2.82903463e-01 -6.79015279e-01
4.57218349e-01 -4.78392616e-02 -9.41472352e-01 -1.06283021e+00
-1.15300286e+00 -9.24432099e-01 -5.54628313e-01 4.05945092e-01
1.10603309e+00 -9.60130990e-01 -6.54154658e-01 3.11735600e-01
-1.53364420e+00 -4.82950389e-01 -1.69499591e-01 4.34587240e-01
-3.73066217e-02 3.16294640e-01 -6.35495722e-01 -6.55101001e-01
-4.89022732e-01 -1.00319648e+00 1.02548075e+00 2.65946686e-01
-7.00138807e-01 -1.13725960e+00 -1.51946515e-01 2.33875826e-01
4.58656162e-01 -6.43274665e-01 1.19185829e+00 -1.00095642e+00
-4.70629670e-02 4.38080728e-02 -2.29587346e-01 7.49796271e-01
-5.11274114e-03 -4.16186571e-01 -8.04666996e-01 -2.33916938e-01
-1.32869080e-01 -4.38679308e-01 9.80406642e-01 3.06133091e-01
1.07118011e+00 -3.46607655e-01 -1.81487232e-01 5.02189100e-01
1.29231179e+00 2.26546049e-01 7.33339369e-01 3.74603391e-01
7.92125762e-01 5.57341218e-01 6.80454135e-01 1.17236443e-01
5.30490637e-01 2.84433573e-01 8.77873749e-02 -1.19514689e-02
-3.81879285e-02 -4.27099019e-01 5.57042956e-01 1.09533191e+00
2.61819243e-01 -5.01378179e-01 -1.25422204e+00 1.03798950e+00
-1.44271338e+00 -4.77732509e-01 -5.41410036e-02 1.94676244e+00
1.06550658e+00 1.80729836e-01 -2.47630790e-01 -8.66102949e-02
4.69610274e-01 1.54932678e-01 -2.91602194e-01 -1.03703618e+00
-1.11844214e-02 7.40992248e-01 7.18844056e-01 7.89529741e-01
-7.55261004e-01 1.84578180e+00 6.17236280e+00 1.17301810e+00
-1.16174638e+00 3.53503019e-01 4.66459066e-01 2.80209005e-01
-4.13057804e-01 7.04209283e-02 -1.25342178e+00 3.10346395e-01
9.68021691e-01 -2.40976855e-01 3.40274543e-01 6.81352377e-01
1.09007388e-01 -2.80544668e-01 -1.14897048e+00 8.16504836e-01
4.42200042e-02 -9.94559467e-01 2.03016192e-01 -4.11598712e-01
3.13137859e-01 2.48993859e-01 -2.78489411e-01 8.42783332e-01
5.47882378e-01 -1.52900338e+00 6.67967021e-01 -1.73402559e-02
8.44829142e-01 -7.68151462e-01 1.08975899e+00 6.81680977e-01
-1.00883496e+00 3.40540648e-01 -8.23332548e-01 -5.44394970e-01
2.15135813e-01 4.79230911e-01 -1.19884980e+00 3.27360481e-01
4.44657385e-01 4.76060301e-01 -4.42599118e-01 7.68391669e-01
-8.21571052e-01 1.27878058e+00 -2.05765203e-01 -4.01886314e-01
3.44037771e-01 1.14205569e-01 2.15256140e-01 1.69650757e+00
1.79297537e-01 3.46696496e-01 1.78134844e-01 6.76572561e-01
4.26769769e-03 2.74601638e-01 -3.10200721e-01 -9.40353870e-02
7.47755587e-01 1.01946461e+00 -5.22166431e-01 -4.80980903e-01
-4.04365391e-01 9.61094618e-01 6.28127456e-01 3.19228321e-01
-5.27895391e-01 -6.97502017e-01 6.84323013e-01 1.33508086e-01
3.75777870e-01 -5.66361487e-01 -6.52870119e-01 -9.89788711e-01
-2.10786015e-01 -7.68659174e-01 2.26106942e-01 -6.40452266e-01
-1.31011808e+00 6.17243111e-01 1.72824636e-01 -6.71357512e-01
-2.48739347e-01 -8.60972285e-01 -7.25882173e-01 1.19337940e+00
-1.67502987e+00 -8.50964427e-01 -3.86883952e-02 2.14749813e-01
9.81720924e-01 -4.36738366e-03 1.00595558e+00 4.00422402e-02
-4.26771909e-01 8.56887579e-01 -1.89106688e-01 3.78857791e-01
8.01487505e-01 -1.53605735e+00 7.69278049e-01 8.04375768e-01
2.52647161e-01 1.01335323e+00 6.11253917e-01 -5.89337885e-01
-9.12770391e-01 -9.83503997e-01 1.42612875e+00 -4.42323893e-01
7.21975207e-01 -5.42337298e-01 -1.20296586e+00 8.06557477e-01
4.65095371e-01 -5.90118110e-01 6.77044868e-01 7.36933470e-01
-4.13047075e-01 1.29398748e-01 -8.78184259e-01 5.91149986e-01
8.39336336e-01 -4.74728972e-01 -1.23682630e+00 3.60725224e-01
1.25126100e+00 -3.11640322e-01 -5.78400970e-01 1.26811802e-01
2.23935157e-01 -5.16080499e-01 7.29679167e-01 -7.86822140e-01
4.86120850e-01 -8.40804726e-02 1.40103921e-01 -1.97087002e+00
-2.94857562e-01 -4.90148872e-01 3.95656288e-01 1.31410515e+00
7.58243620e-01 -8.13881278e-01 5.76991558e-01 5.96593320e-01
-1.95999607e-01 -6.23813391e-01 -8.34732890e-01 -1.06993294e+00
7.56919444e-01 -8.36587667e-01 7.99559355e-01 8.32102954e-01
3.18214834e-01 6.85634315e-01 -2.25093499e-01 2.92712972e-02
-2.87155043e-02 -3.74511331e-01 5.25755644e-01 -8.80562067e-01
-3.96779418e-01 -2.58313715e-01 -2.40430444e-01 -1.66295326e+00
5.15359402e-01 -1.01770401e+00 5.90559483e-01 -1.59783161e+00
5.89018650e-02 -4.76266116e-01 -2.32709423e-01 9.29405630e-01
-6.89897537e-01 -9.53893065e-02 2.07565010e-01 -1.61439851e-01
-4.25912499e-01 4.44490254e-01 1.03374565e+00 -1.16891682e-01
-1.74447298e-01 -3.46170098e-01 -8.52851272e-01 7.27412283e-01
8.88809025e-01 -4.04933870e-01 -3.01931918e-01 -1.05331492e+00
3.39328349e-01 -2.38916025e-01 -1.18544482e-01 -5.04273891e-01
1.04052722e-01 -4.87348959e-02 -1.92506667e-02 -4.82367426e-01
1.58925310e-01 -4.10685360e-01 -6.56629622e-01 2.96415448e-01
-2.21834049e-01 2.48483285e-01 5.47335446e-01 4.02175300e-02
-3.21924299e-01 -7.65411377e-01 5.34587264e-01 -2.54168987e-01
-9.52445626e-01 -1.21600010e-01 -6.40052736e-01 4.38490599e-01
5.11859775e-01 -2.85362899e-01 -4.62430716e-01 -3.82215470e-01
-3.54243964e-01 4.87667888e-01 -1.26740597e-02 5.49214184e-01
5.14087975e-01 -7.28302538e-01 -8.02606821e-01 2.02182248e-01
2.01358631e-01 3.55750531e-01 -9.85623077e-02 4.28196937e-01
-6.02367759e-01 6.40785992e-01 9.75410417e-02 -3.95542324e-01
-1.19250858e+00 3.19458127e-01 5.32514974e-02 -4.49825287e-01
-3.65933865e-01 1.35758686e+00 3.30947489e-01 -5.53319573e-01
2.05532104e-01 -6.91682696e-01 -3.05093229e-01 1.15305923e-01
7.25763381e-01 1.08851619e-01 3.03063452e-01 -3.79044443e-01
-3.70521575e-01 3.04861933e-01 -6.15361571e-01 -1.19557358e-01
1.22845066e+00 -1.38209254e-01 -1.90998971e-01 5.51512301e-01
1.00952435e+00 1.16127312e-01 -7.52925456e-01 -1.73572764e-01
3.72123957e-01 -2.46530920e-01 -2.16385629e-03 -9.58612323e-01
-5.23578823e-01 1.09571242e+00 1.58066735e-01 1.02002472e-01
7.99201667e-01 8.23647082e-02 1.05757117e+00 4.76112276e-01
5.65656304e-01 -1.17373717e+00 -3.55376482e-01 1.28737724e+00
7.17208266e-01 -1.36894131e+00 -1.84464037e-01 -4.05843616e-01
-6.96978927e-01 8.02341580e-01 6.76401317e-01 9.12905205e-03
5.80816090e-01 2.04765841e-01 3.73863995e-01 -1.64298639e-01
-6.98382735e-01 -4.24601257e-01 2.96978623e-01 5.79405427e-01
8.70198607e-01 4.22872633e-01 -6.84927762e-01 9.59952950e-01
-9.47083175e-01 -3.45662832e-01 4.90554959e-01 6.51360810e-01
-7.30845094e-01 -1.23451710e+00 -3.14210176e-01 5.62985897e-01
-5.86756766e-01 -9.01951849e-01 -2.66365409e-01 8.21163356e-01
1.74666211e-01 1.32845974e+00 1.95798308e-01 -2.51597703e-01
2.92475820e-01 2.79076129e-01 2.43285730e-01 -1.36613953e+00
-8.21002543e-01 -2.69205630e-01 2.69127518e-01 -3.09434861e-01
4.63385768e-02 -2.90798247e-01 -1.64705825e+00 -2.79092818e-01
-6.98822975e-01 3.73020768e-01 6.46120012e-01 1.44898760e+00
6.85648471e-02 4.96167839e-01 2.31688619e-02 -4.98437464e-01
-5.27962327e-01 -1.41115415e+00 -6.42536581e-01 3.76741230e-01
1.67132080e-01 -4.86909896e-01 -2.82526016e-01 -4.28488515e-02] | [10.52548599243164, 9.135225296020508] |
91ebbd5b-b82f-45e7-9427-97100d693323 | correspondence-free-online-human-motion | 2302.00556 | null | https://arxiv.org/abs/2302.00556v1 | https://arxiv.org/pdf/2302.00556v1.pdf | Correspondence-free online human motion retargeting | We present a novel data-driven framework for unsupervised human motion retargeting which animates a target body shape with a source motion. This allows to retarget motions between different characters by animating a target subject with a motion of a source subject. Our method is correspondence-free,~\ie neither spatial correspondences between the source and target shapes nor temporal correspondences between different frames of the source motion are required. Our proposed method directly animates a target shape with arbitrary sequences of humans in motion, possibly captured using 4D acquisition platforms or consumer devices. Our framework takes into account long-term temporal context of $1$ second during retargeting while accounting for surface details. To achieve this, we take inspiration from two lines of existing work: skeletal motion retargeting, which leverages long-term temporal context at the cost of surface detail, and surface-based retargeting, which preserves surface details without considering long-term temporal context. We unify the advantages of these works by combining a learnt skinning field with a skeletal retargeting approach. During inference, our method runs online,~\ie the input can be processed in a serial way, and retargeting is performed in a single forward pass per frame. Experiments show that including long-term temporal context during training improves the method's accuracy both in terms of the retargeted skeletal motion and the detail preservation. Furthermore, our method generalizes well on unobserved motions and body shapes. We demonstrate that the proposed framework achieves state-of-the-art results on two test datasets. | ['Anne-Hélène Olivier', 'Jean-Sébastien Franco', 'Stefanie Wuhrer', 'Rim Rekik', 'Mathieu Marsot'] | 2023-02-01 | null | null | null | null | ['motion-retargeting'] | ['computer-vision'] | [ 3.26653957e-01 9.48043838e-02 -1.99293345e-01 -1.91345252e-02
-5.61696768e-01 -6.31060243e-01 6.75386786e-01 -2.91238546e-01
-5.08149028e-01 4.08597887e-01 2.28423327e-01 1.51725024e-01
2.72899389e-01 -6.98299944e-01 -8.27661574e-01 -5.35539091e-01
-1.53619482e-03 3.46709579e-01 8.06472540e-01 -2.20133394e-01
1.79648250e-02 5.58487654e-01 -1.23934412e+00 1.13698222e-01
4.81324971e-01 4.60487068e-01 4.12052125e-03 1.01142085e+00
2.95268685e-01 6.06465936e-01 -3.18808526e-01 -4.15710598e-01
6.40739679e-01 -5.77894747e-01 -8.51676762e-01 4.04366910e-01
8.26934695e-01 -3.66823584e-01 -6.73162103e-01 7.95441747e-01
5.21947801e-01 5.32040596e-01 4.17751759e-01 -8.96544933e-01
-2.69515842e-01 3.40849519e-01 -8.02203357e-01 1.04622945e-01
5.13307750e-01 5.66768467e-01 8.15887272e-01 -8.39649081e-01
1.10260892e+00 1.25155723e+00 7.65406728e-01 9.06515241e-01
-1.38346076e+00 -4.77183253e-01 3.63634855e-01 -8.46297517e-02
-1.30451643e+00 -4.94169563e-01 9.68613863e-01 -7.28466094e-01
6.99917018e-01 2.56747514e-01 9.07301843e-01 1.08985400e+00
2.59670377e-01 8.38183105e-01 4.94988680e-01 -3.09077263e-01
1.68946281e-01 -6.49413705e-01 -2.53831834e-01 8.44718695e-01
5.94392642e-02 3.02407771e-01 -8.09645236e-01 -6.75987825e-02
1.25118792e+00 -3.71197611e-02 -3.22518796e-01 -6.53877854e-01
-1.77110946e+00 5.40063977e-01 2.47314051e-01 2.78489500e-01
-2.21902102e-01 5.19293249e-01 2.16135815e-01 5.58411144e-02
3.96101177e-01 2.55449265e-01 -2.17167929e-01 -1.08171731e-01
-1.20324862e+00 5.18382609e-01 4.52507615e-01 1.03524697e+00
6.91499233e-01 2.10613072e-01 -4.08126146e-01 3.68194818e-01
1.91807672e-01 5.92939615e-01 5.67538857e-01 -1.34141636e+00
4.19573963e-01 3.20453763e-01 2.11551800e-01 -1.17554426e+00
-3.07820231e-01 -2.31897216e-02 -7.55785465e-01 3.87590051e-01
5.36435723e-01 -1.65586308e-01 -9.81644869e-01 2.00800276e+00
7.17513621e-01 5.16309261e-01 -9.75143686e-02 1.04028893e+00
6.75328135e-01 3.24776471e-01 7.16699213e-02 -1.86404020e-01
1.19637394e+00 -1.25713420e+00 -4.95941162e-01 -2.50432909e-01
4.75862026e-01 -7.35772669e-01 1.11653388e+00 1.52449206e-01
-1.51931334e+00 -7.85085082e-01 -9.55647826e-01 -1.69329539e-01
2.78839171e-01 -1.52393892e-01 2.85211831e-01 3.82484317e-01
-9.47433531e-01 8.57005417e-01 -1.18498945e+00 -5.19861460e-01
1.38314500e-01 3.96896929e-01 -4.83787835e-01 9.66344774e-02
-9.13940787e-01 5.50263405e-01 2.37427041e-01 -1.83875218e-01
-8.18604589e-01 -7.00591207e-01 -9.94685054e-01 -3.39092612e-01
4.81209636e-01 -1.38936317e+00 1.24252045e+00 -9.82097566e-01
-1.89506173e+00 8.31917644e-01 -2.78315306e-01 -2.26372793e-01
9.45114255e-01 -4.58487421e-01 -1.94949597e-01 4.65582818e-01
1.45614773e-01 8.44963789e-01 1.12609518e+00 -9.82090712e-01
-6.06158435e-01 -1.73274070e-01 -5.02363728e-05 4.12654847e-01
-1.76204871e-02 -2.90392250e-01 -9.90987599e-01 -1.40806258e+00
6.35990277e-02 -1.40305734e+00 -3.07958961e-01 3.67639631e-01
-3.86638135e-01 3.03372651e-01 8.29101861e-01 -5.23156881e-01
1.26071048e+00 -2.08684468e+00 7.31188536e-01 7.01801777e-02
2.20729649e-01 8.43010768e-02 -1.80139989e-01 1.56317189e-01
2.18587499e-02 -1.59563094e-01 -6.20880187e-01 -5.37705064e-01
-1.02840073e-01 2.21909776e-01 -9.74429622e-02 5.91184616e-01
-6.39081523e-02 1.18355978e+00 -1.02977955e+00 -5.74826181e-01
2.70562768e-01 4.30222809e-01 -7.28438079e-01 3.22313040e-01
-2.29539603e-01 1.02870691e+00 -3.75505507e-01 4.39518392e-01
2.24668697e-01 -1.36355102e-01 3.67381945e-02 -2.02072576e-01
-1.26023501e-01 -2.05269605e-01 -1.32036269e+00 2.38185573e+00
-6.94175139e-02 3.35094184e-01 -8.95237103e-02 -5.14479458e-01
5.41477621e-01 3.39809090e-01 7.17916191e-01 -2.39062309e-01
-1.24674961e-02 1.33418441e-01 -1.97567016e-01 -5.09673595e-01
5.99475563e-01 -3.25006276e-01 -1.46653429e-01 4.69960511e-01
-2.68849283e-01 -2.34214619e-01 -1.85913220e-02 9.17477682e-02
1.08954537e+00 8.13212693e-01 4.35041696e-01 -2.23877773e-01
4.29011792e-01 1.59586772e-01 7.34930933e-01 4.34606016e-01
-1.43666074e-01 9.40079272e-01 -9.44961831e-02 -5.74564755e-01
-1.06289351e+00 -1.21438801e+00 4.41663384e-01 1.09488773e+00
4.10318226e-01 -2.92795330e-01 -6.68288052e-01 -6.61913395e-01
-1.43132061e-01 3.18870485e-01 -7.57602394e-01 -6.50692880e-02
-1.26618648e+00 -2.90491968e-01 5.55526495e-01 7.39830911e-01
5.46049714e-01 -9.20995891e-01 -1.14284360e+00 3.10211033e-01
-2.99069762e-01 -1.09339881e+00 -1.25860190e+00 -4.72736925e-01
-1.15849268e+00 -1.01392090e+00 -1.06471050e+00 -5.75795710e-01
6.13058746e-01 2.23527446e-01 1.05266929e+00 9.55537036e-02
-2.74972826e-01 5.51238298e-01 -2.47405380e-01 2.34504774e-01
-4.27928448e-01 -1.03200957e-01 1.65598273e-01 1.21605664e-01
-3.75546962e-01 -7.26340175e-01 -7.83873796e-01 4.12422597e-01
-1.07447970e+00 3.04124475e-01 4.23619419e-01 6.31471097e-01
7.43455112e-01 -3.05873275e-01 1.40765607e-02 -8.26858461e-01
-1.62307501e-01 -2.17060730e-01 -1.90288022e-01 2.00592000e-02
-5.23655713e-02 2.11474463e-01 2.72162825e-01 -8.15697968e-01
-9.37941492e-01 2.95537978e-01 -3.28299887e-02 -8.10266137e-01
-8.28911588e-02 6.03784397e-02 -2.32671365e-01 -1.43617943e-01
6.76376939e-01 2.96381605e-03 -1.84316784e-01 -4.71087575e-01
6.94787264e-01 -2.31923923e-01 1.13229322e+00 -6.83831275e-01
1.18164492e+00 8.18867266e-01 2.15406373e-01 -7.01199412e-01
-3.84247839e-01 -2.86441267e-01 -1.12816739e+00 -2.24163517e-01
1.13150549e+00 -7.55511820e-01 -6.08919799e-01 4.89377677e-01
-1.09473372e+00 -6.66128993e-01 -6.38043523e-01 5.44310033e-01
-7.98307955e-01 7.58358777e-01 -6.46957994e-01 -2.18052953e-01
-2.95931518e-01 -1.12124372e+00 1.17881680e+00 7.96046853e-03
-7.39819109e-01 -1.08647454e+00 4.03165281e-01 1.11615524e-01
-7.76996911e-02 7.35079229e-01 6.48529708e-01 -2.40733176e-01
-5.68692267e-01 1.23560213e-01 3.64301413e-01 -2.19225764e-01
3.77479523e-01 -3.18939686e-02 -6.02960229e-01 -4.78168488e-01
-2.90498585e-01 1.57903045e-01 6.03583217e-01 4.14191842e-01
5.30642867e-01 -4.39110070e-01 -4.28286970e-01 7.94666767e-01
1.07527757e+00 1.13585345e-01 5.72512209e-01 3.18674952e-01
1.15976655e+00 6.82705164e-01 5.92538416e-01 4.60611045e-01
3.68342489e-01 1.08111453e+00 2.06263378e-01 -2.30573509e-02
-5.04042983e-01 -3.48378241e-01 4.64838058e-01 7.63266444e-01
-4.50712919e-01 -8.15761611e-02 -8.34857941e-01 8.17902088e-01
-2.09640789e+00 -1.06997859e+00 3.91223878e-02 2.47329950e+00
8.60273421e-01 -3.84878628e-02 4.52107340e-01 4.07679230e-02
7.35694587e-01 2.05545515e-01 -7.39191353e-01 6.04749322e-02
-8.34794715e-03 2.75741339e-01 3.45585912e-01 8.55302751e-01
-1.16331875e+00 1.08515120e+00 6.19946384e+00 4.82744366e-01
-1.17757559e+00 1.28697366e-01 1.63953334e-01 -2.79380500e-01
-3.79406840e-01 -2.57252809e-02 -4.45322216e-01 2.21716732e-01
5.29067814e-01 -7.75308833e-02 1.89770237e-01 6.15552723e-01
2.25779653e-01 1.05755463e-01 -1.33273983e+00 9.29920673e-01
7.76438564e-02 -1.37603414e+00 1.25935748e-01 -1.48762196e-01
8.56687903e-01 -4.19402450e-01 -6.26583621e-02 -8.64041895e-02
2.42022067e-01 -8.18621397e-01 1.14604914e+00 6.63584650e-01
8.96570623e-01 -5.43425262e-01 1.62312552e-01 1.80369779e-01
-1.67401373e+00 3.72149527e-01 2.69856989e-01 -7.15867579e-02
3.88046831e-01 1.55138686e-01 -3.84839237e-01 7.20529377e-01
5.36549687e-01 8.21467757e-01 -3.89762193e-01 6.79503083e-01
-2.13116303e-01 2.66195446e-01 -2.13131547e-01 6.91677928e-01
-9.82323661e-02 -1.33469179e-01 1.01161361e+00 1.16044021e+00
3.24782670e-01 3.72801304e-01 4.05718833e-01 6.11859322e-01
1.75402999e-01 2.04789918e-02 -5.53027689e-01 3.67189944e-01
2.33890399e-01 8.55358124e-01 -7.48465657e-01 -4.70884979e-01
-3.37761581e-01 1.52585173e+00 -4.06016484e-02 4.06323552e-01
-9.01952326e-01 -5.05816713e-02 6.35481656e-01 4.97127444e-01
5.28128326e-01 -4.57556754e-01 -3.37239891e-01 -1.41144454e+00
1.00451849e-01 -6.86769128e-01 5.10252118e-01 -5.59177876e-01
-1.01317167e+00 5.54592371e-01 2.63002545e-01 -1.46495461e+00
-5.16694486e-01 -1.73196997e-02 -5.33253431e-01 5.73463917e-01
-1.03566861e+00 -1.34072864e+00 -2.94401348e-01 9.11042392e-01
8.90591681e-01 1.98832676e-01 4.47066247e-01 1.33123785e-01
-3.52172226e-01 5.55315971e-01 -5.35499096e-01 1.52740642e-01
8.98948908e-01 -1.04949105e+00 7.92623460e-01 1.27766240e+00
1.52047694e-01 4.79427695e-01 8.25586736e-01 -9.61030781e-01
-1.31910038e+00 -1.28992701e+00 6.43840611e-01 -5.90616822e-01
4.05515552e-01 -3.59127577e-03 -8.38831007e-01 9.61950839e-01
-3.33639723e-03 2.97842175e-01 4.16960359e-01 -4.48845506e-01
-2.60435045e-01 2.56168723e-01 -9.64508235e-01 1.01401341e+00
1.42479753e+00 -2.49855071e-01 -8.04405034e-01 -1.50602967e-01
8.00623834e-01 -9.03330207e-01 -8.43097687e-01 5.06630003e-01
7.64122486e-01 -7.92569160e-01 1.14514041e+00 -5.18819630e-01
3.93309116e-01 -6.98826909e-01 -1.93341419e-01 -1.00699580e+00
-4.26115751e-01 -1.10718393e+00 -4.47914720e-01 7.86424100e-01
-9.48682576e-02 -1.95824325e-01 9.97239172e-01 7.52375901e-01
-2.24651113e-01 -5.17673314e-01 -9.28580582e-01 -8.67849469e-01
7.29850531e-02 -4.50707853e-01 4.13362801e-01 1.00318348e+00
-1.64042279e-01 1.40954301e-01 -6.33198559e-01 2.42365703e-01
7.14337766e-01 2.17998222e-01 1.14083779e+00 -9.21992838e-01
-6.48433745e-01 -3.28365564e-01 -4.47902054e-01 -1.40494847e+00
1.01009116e-01 -6.24557614e-01 1.41064405e-01 -1.13160503e+00
6.54348433e-02 -8.91881287e-02 2.58026272e-01 5.68319738e-01
-4.42025155e-01 6.10193491e-01 4.92902607e-01 6.09618068e-01
-2.51023829e-01 6.24820054e-01 1.47795045e+00 -2.54037865e-02
-5.60536921e-01 -1.05208583e-01 -2.10257784e-01 1.01905370e+00
3.20155740e-01 -3.98449808e-01 -4.71215487e-01 -5.68598390e-01
-1.51386276e-01 4.03610170e-01 6.05811059e-01 -1.11119318e+00
1.91732660e-01 -4.62640703e-01 2.42733255e-01 -3.37213069e-01
4.13148791e-01 -6.67655468e-01 6.51555896e-01 5.75748861e-01
-1.84622586e-01 2.90544778e-01 2.15396389e-01 7.88459659e-01
1.71852261e-01 1.86021149e-01 1.01612389e+00 -8.39910135e-02
-8.94887805e-01 5.43759108e-01 -1.91984937e-01 2.11454615e-01
1.11019909e+00 -5.90412676e-01 1.91151753e-01 -2.53632665e-01
-1.13735545e+00 -2.66065568e-01 1.03223860e+00 4.55041051e-01
6.22595727e-01 -1.49849153e+00 -6.19298220e-01 2.24871170e-02
-1.57553591e-02 3.16788077e-01 3.04509073e-01 8.78168464e-01
-6.56777442e-01 1.97986644e-02 -2.12975636e-01 -8.34923804e-01
-1.32172668e+00 5.86055815e-01 2.42298335e-01 -1.88603759e-01
-1.17271411e+00 7.67985523e-01 7.39737689e-01 -4.83479500e-02
-1.32790387e-01 -3.86369109e-01 1.89390749e-01 -3.02809656e-01
4.37589586e-01 5.21584570e-01 -3.05864424e-01 -1.07123351e+00
-3.58016431e-01 1.35386741e+00 2.62231290e-01 -6.17687225e-01
8.98032486e-01 -3.13179612e-01 2.65372306e-01 5.95709562e-01
9.22790170e-01 3.79459500e-01 -1.78879106e+00 -3.82546216e-01
-2.02551857e-01 -6.85036063e-01 -4.03621346e-01 -3.62320572e-01
-1.15721714e+00 5.56300998e-01 2.74952710e-01 -6.50747299e-01
1.10035670e+00 -3.95522639e-02 1.09791851e+00 -6.36377977e-03
5.26655257e-01 -7.28587031e-01 5.29868126e-01 2.37298623e-01
8.06791246e-01 -8.55332911e-01 1.88214362e-01 -5.96042812e-01
-7.70912170e-01 1.01508534e+00 5.27789950e-01 -2.61254281e-01
3.69039506e-01 4.85362001e-02 -1.06551396e-02 -1.25590324e-01
-4.47477669e-01 -1.93901211e-02 6.76103294e-01 5.78585148e-01
1.99776351e-01 -4.58356589e-02 -1.57971382e-01 4.20149744e-01
-1.74891070e-01 -5.87197915e-02 4.15320665e-01 1.21867573e+00
-1.49279445e-01 -1.19232810e+00 -5.61801732e-01 -2.91271955e-01
-2.99704164e-01 4.44339402e-02 -1.84982613e-01 9.91697490e-01
1.84768081e-01 5.90943635e-01 -4.32966650e-02 -4.00897443e-01
4.30523425e-01 -8.59949216e-02 6.86207592e-01 -8.03963780e-01
-4.89369184e-01 3.49582165e-01 -1.74158469e-01 -9.42529261e-01
-7.64664650e-01 -8.11577559e-01 -1.41678321e+00 -3.60146701e-01
2.20099375e-01 -1.71295255e-01 -5.06795421e-02 8.94196033e-01
4.44870472e-01 4.44436669e-01 3.58324260e-01 -1.27902293e+00
6.09131623e-03 -5.51116824e-01 -2.49797404e-01 6.63392842e-01
3.93192947e-01 -7.23464489e-01 -4.84095775e-02 8.14711034e-01] | [7.404534339904785, -0.5631067752838135] |
d9e4607d-9d4c-4fad-9976-5ff6ec8cbdb9 | effects-of-locality-and-rule-language-on | 2302.06967 | null | https://arxiv.org/abs/2302.06967v1 | https://arxiv.org/pdf/2302.06967v1.pdf | Effects of Locality and Rule Language on Explanations for Knowledge Graph Embeddings | Knowledge graphs (KGs) are key tools in many AI-related tasks such as reasoning or question answering. This has, in turn, propelled research in link prediction in KGs, the task of predicting missing relationships from the available knowledge. Solutions based on KG embeddings have shown promising results in this matter. On the downside, these approaches are usually unable to explain their predictions. While some works have proposed to compute post-hoc rule explanations for embedding-based link predictors, these efforts have mostly resorted to rules with unbounded atoms, e.g., bornIn(x,y) => residence(x,y), learned on a global scope, i.e., the entire KG. None of these works has considered the impact of rules with bounded atoms such as nationality(x,England) => speaks(x, English), or the impact of learning from regions of the KG, i.e., local scopes. We therefore study the effects of these factors on the quality of rule-based explanations for embedding-based link predictors. Our results suggest that more specific rules and local scopes can improve the accuracy of the explanations. Moreover, these rules can provide further insights about the inner-workings of KG embeddings for link prediction. | ['Luis Galárraga'] | 2023-02-14 | null | null | null | null | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['graphs', 'methodology'] | [-3.68858606e-01 6.36633039e-01 -8.01219463e-01 -3.85386229e-01
1.23017728e-01 -1.99567318e-01 7.57463813e-01 5.96773565e-01
-8.65842625e-02 7.85894513e-01 6.65221214e-01 -6.76419914e-01
-5.81757247e-01 -1.29590750e+00 -9.41736162e-01 -2.37071648e-01
-1.99715510e-01 4.47644413e-01 3.02464634e-01 -3.22729886e-01
-2.10747644e-02 2.11760566e-01 -1.50729418e+00 8.54742005e-02
1.10808742e+00 4.48391199e-01 -9.38086137e-02 2.99942344e-01
-4.40614164e-01 8.80591035e-01 -3.30477536e-01 -1.00558412e+00
-9.77331400e-02 -3.54240328e-01 -8.08656871e-01 -5.42183936e-01
4.10562485e-01 -1.26058593e-01 -6.57840550e-01 8.97051990e-01
1.18009830e-02 4.13470063e-03 6.95528209e-01 -1.54209065e+00
-1.08936620e+00 1.29751968e+00 -6.50977641e-02 2.53993481e-01
3.80828053e-01 -2.30324298e-01 1.65115213e+00 -7.20925510e-01
8.51446450e-01 1.20599508e+00 8.48882139e-01 3.69722366e-01
-1.27810478e+00 -5.86214006e-01 4.53984916e-01 8.27841103e-01
-1.20709741e+00 -1.96374040e-02 8.95449698e-01 -4.05360639e-01
9.53957260e-01 2.78341472e-01 7.48777568e-01 1.08786428e+00
1.49183363e-01 5.63740909e-01 8.11760187e-01 -5.20342886e-01
1.82995275e-02 3.56005490e-01 4.61062729e-01 7.47420490e-01
8.08613420e-01 7.61383027e-02 -8.60400915e-01 -2.04003200e-01
4.22659814e-01 -1.35211319e-01 -5.16446888e-01 -5.09130061e-01
-1.12409151e+00 1.12360501e+00 7.75345266e-01 4.08539265e-01
-2.07987919e-01 1.23941869e-01 1.79911211e-01 2.38610938e-01
4.45492685e-01 6.95575476e-01 -8.91900182e-01 -1.30914114e-02
-6.27899706e-01 2.62346923e-01 1.10954940e+00 8.61533940e-01
9.41146076e-01 -1.32788926e-01 9.71490890e-02 5.59571385e-01
3.96487355e-01 3.87560017e-02 1.76987603e-01 -5.22692502e-01
5.65932095e-01 9.80470717e-01 5.71187846e-02 -1.34654963e+00
-5.91217160e-01 -3.78795594e-01 -4.24609005e-01 -2.23472521e-01
6.80029154e-01 -1.86588988e-01 -7.00970232e-01 1.81948972e+00
4.11849231e-01 2.54405171e-01 1.44682810e-01 9.22470450e-01
9.61285353e-01 4.22218382e-01 1.39931053e-01 4.78733853e-02
1.10079610e+00 -1.05886734e+00 -6.61045134e-01 -1.18945882e-01
9.47595835e-01 -3.58019710e-01 1.12185681e+00 -3.49338651e-02
-6.70788288e-01 -4.12695110e-01 -7.93603897e-01 -1.80082321e-02
-9.17676508e-01 -1.64698452e-01 1.04462469e+00 4.97482359e-01
-8.37467015e-01 8.18618059e-01 -7.63494134e-01 -5.95149159e-01
1.45950139e-01 3.55562210e-01 -3.85133594e-01 -3.15445185e-01
-1.88843918e+00 1.25864553e+00 4.82711136e-01 -1.01745889e-01
-2.02645108e-01 -9.80456591e-01 -9.13297534e-01 3.04811746e-01
6.12261295e-01 -9.44138765e-01 4.02064592e-01 -7.81104922e-01
-8.21303248e-01 3.41871649e-01 -7.68651292e-02 -5.19961119e-01
2.63749003e-01 -1.81865573e-01 -7.68919170e-01 -1.52482644e-01
1.36157423e-01 5.62256634e-01 4.80095297e-01 -1.13965535e+00
-5.25463223e-01 -4.57371056e-01 4.77333575e-01 1.69425681e-01
-7.13759124e-01 -3.61239105e-01 -2.25808859e-01 -4.23217148e-01
-3.66573744e-02 -7.94294059e-01 -3.10564619e-02 -1.40243858e-01
-5.65843165e-01 -5.76455534e-01 5.14064491e-01 -7.83766925e-01
1.47818875e+00 -1.90437496e+00 3.38673562e-01 3.42558712e-01
2.67283529e-01 1.98171511e-01 4.40168791e-02 6.46864414e-01
-2.70917825e-03 6.01201773e-01 2.61493206e-01 1.13158837e-01
1.15238041e-01 6.28717184e-01 -2.86585599e-01 4.67039533e-02
6.52469173e-02 1.07631433e+00 -1.01450622e+00 -4.99359369e-01
3.94481309e-02 5.58108151e-01 -7.55096912e-01 -2.80937374e-01
-4.09902751e-01 8.08278546e-02 -3.96757036e-01 4.33946967e-01
5.93076609e-02 -9.75362808e-02 3.94413501e-01 -2.33615920e-01
-6.42440766e-02 5.93153179e-01 -8.64947557e-01 1.22975338e+00
-4.54995424e-01 6.35445356e-01 -2.38806516e-01 -9.80178177e-01
8.51594329e-01 2.03160703e-01 1.68895081e-01 -4.07181829e-01
-3.69596690e-01 3.25129956e-01 5.20813465e-01 -4.49142575e-01
5.16433418e-01 -1.03219263e-01 1.84439406e-01 2.28907138e-01
-2.41337568e-01 1.18646100e-01 6.81441575e-02 4.75754797e-01
1.28453350e+00 9.93903279e-02 1.89131245e-01 -1.18285328e-01
2.57933825e-01 2.99192578e-01 7.76523650e-01 5.99126101e-01
-5.04170656e-02 1.06055178e-01 7.53362358e-01 -5.92377424e-01
-9.55341578e-01 -8.26996088e-01 -1.20524436e-01 8.91757548e-01
3.12683344e-01 -9.01439548e-01 -2.19874889e-01 -1.04066420e+00
6.29083633e-01 1.05187809e+00 -8.53856087e-01 -4.90554273e-01
-3.43068838e-01 -3.52666646e-01 4.55199450e-01 7.42994964e-01
1.57501578e-01 -9.80905950e-01 -3.02763999e-01 2.49794185e-01
-1.11185953e-01 -9.36222255e-01 3.28713446e-03 2.19743058e-01
-9.21056271e-01 -1.29057336e+00 -8.36243331e-02 -3.26917291e-01
6.03377402e-01 -1.14590771e-01 1.27902424e+00 5.68286180e-01
-5.70485229e-03 5.18578649e-01 -6.05681419e-01 -3.54379773e-01
5.86118773e-02 2.74022967e-01 3.31645489e-01 -2.35960424e-01
6.05541706e-01 -6.80315256e-01 -3.72751415e-01 3.79475147e-01
-7.10144281e-01 -1.83669969e-01 4.47844386e-01 9.04859483e-01
3.87524456e-01 3.56431752e-01 7.63065338e-01 -1.16474283e+00
7.60201037e-01 -7.02825189e-01 -7.42679909e-02 5.46658576e-01
-1.10412729e+00 3.65026385e-01 7.28707612e-01 -5.35240412e-01
-6.60769999e-01 -4.51231509e-01 2.52722241e-02 -3.73603940e-01
6.84871897e-02 1.12199640e+00 -1.86847746e-01 1.31712586e-01
5.88039815e-01 -1.16127990e-02 -2.77256042e-01 -3.77253979e-01
5.62359989e-01 2.94011503e-01 1.87949017e-01 -6.15032494e-01
9.49846745e-01 4.54555303e-02 -9.42895412e-02 -6.22797608e-01
-7.97226727e-01 -2.17121750e-01 -5.53409338e-01 -2.89851669e-02
8.54284823e-01 -6.56516850e-01 -3.02659482e-01 -3.36389244e-01
-1.01027298e+00 -3.15551549e-01 -2.02275246e-01 6.53297603e-01
-1.66631669e-01 1.35880589e-01 -3.21403235e-01 -4.64967996e-01
1.40579984e-01 -7.21699417e-01 4.19941992e-01 1.63626745e-01
-7.70354331e-01 -1.47568905e+00 -8.17981288e-02 5.28460979e-01
3.42086643e-01 1.37225628e-01 1.75542819e+00 -9.81680930e-01
-5.90576768e-01 -8.83089229e-02 -1.46285772e-01 -4.03303057e-02
1.50337487e-01 1.44804552e-01 -4.44567323e-01 2.42926359e-01
-7.46084869e-01 -8.87921825e-02 9.65128124e-01 9.68561918e-02
8.28540266e-01 -5.82773387e-01 -8.21037292e-01 4.23497409e-01
1.21577895e+00 -5.56842759e-02 5.55502713e-01 4.11171913e-01
8.68535280e-01 7.87510872e-01 7.06190050e-01 2.04560354e-01
8.14771473e-01 7.98211753e-01 5.68007112e-01 1.73208416e-01
-2.68013805e-01 -7.11248100e-01 3.24833214e-01 6.38458610e-01
-2.80753881e-01 -1.36907220e-01 -1.10905349e+00 8.22878420e-01
-1.93914580e+00 -1.00466049e+00 -4.94812012e-01 1.85322821e+00
7.84402072e-01 1.63933069e-01 -1.26284391e-01 2.22571269e-01
5.20370245e-01 2.05213249e-01 -6.18074417e-01 -4.91884559e-01
-4.02653664e-02 -5.24589866e-02 3.48835796e-01 7.59416223e-01
-6.05828166e-01 1.15877724e+00 5.64988422e+00 4.05159742e-01
-7.84753263e-01 5.16483225e-02 9.38969702e-02 1.35029405e-01
-1.07167244e+00 3.63590330e-01 -1.00631285e+00 5.60371041e-01
6.27561390e-01 -1.71940506e-01 4.99622792e-01 8.01315486e-01
-7.96742216e-02 1.29800793e-02 -1.22013986e+00 4.44467783e-01
1.50098100e-01 -1.25832796e+00 2.00327218e-01 2.22258121e-01
7.17632771e-01 -2.31950045e-01 -1.08277798e-01 6.46721244e-01
5.76068521e-01 -1.04366672e+00 5.37854552e-01 5.57845891e-01
3.14137310e-01 -7.46653616e-01 8.35847676e-01 4.05845106e-01
-9.27953959e-01 -1.94053352e-01 -4.66252893e-01 -3.50639910e-01
4.83449474e-02 9.66839433e-01 -8.88721228e-01 8.58668506e-01
6.67054713e-01 7.40783572e-01 -6.67969048e-01 6.58050537e-01
-8.10776651e-01 8.84864688e-01 -2.14151725e-01 -1.51000232e-01
9.90861282e-02 -1.31453916e-01 4.52940911e-01 7.31208861e-01
3.11806947e-01 2.70002276e-01 -3.26278061e-02 9.45916831e-01
-2.13966623e-01 1.31947130e-01 -9.57263231e-01 -2.83063054e-01
6.35986865e-01 8.75710666e-01 -4.65798259e-01 -2.09173381e-01
-7.25484610e-01 4.10740793e-01 6.47259116e-01 4.86972541e-01
-8.58957708e-01 -2.45106712e-01 1.00590968e+00 7.13194489e-01
4.29685652e-01 -1.23791300e-01 -3.05467576e-01 -1.09593105e+00
3.35954353e-02 -5.02720773e-01 5.61639667e-01 -6.86046243e-01
-1.26606560e+00 1.93998396e-01 -3.84622999e-02 -5.71828485e-01
-6.20249175e-02 -4.34006780e-01 -6.80775702e-01 6.71405494e-01
-1.69252443e+00 -1.11930203e+00 1.79820918e-02 5.81841588e-01
1.46570038e-02 -1.02765538e-01 7.96939075e-01 1.09371811e-01
-5.30022383e-01 6.23791456e-01 -8.37780163e-02 9.48724002e-02
6.61411166e-01 -1.18142784e+00 1.88396070e-02 6.61766946e-01
6.39298379e-01 8.83140504e-01 7.24567115e-01 -1.05092204e+00
-1.27990675e+00 -1.00718808e+00 1.66247654e+00 -7.34491467e-01
9.49572563e-01 -6.90625533e-02 -1.11059499e+00 9.22385573e-01
-4.08353992e-02 1.26748249e-01 7.04688728e-01 1.12114096e+00
-6.22544110e-01 -6.89010844e-02 -8.04285884e-01 7.66182959e-01
1.37608194e+00 -4.01239961e-01 -7.86558926e-01 7.10083544e-02
8.08032870e-01 -7.58256018e-03 -1.17904699e+00 5.09694278e-01
5.25097489e-01 -9.63168204e-01 9.74549651e-01 -9.46975827e-01
7.54028618e-01 -2.01681465e-01 -5.69226071e-02 -1.62215078e+00
-4.91622984e-01 5.13047390e-02 -5.08262575e-01 1.20521581e+00
9.05668736e-01 -8.92802954e-01 8.87268007e-01 8.47930193e-01
-2.37673700e-01 -1.09182477e+00 -8.51194680e-01 -8.19895566e-01
1.10872701e-01 -3.56146932e-01 9.67430055e-01 1.43110716e+00
2.78442442e-01 3.71218681e-01 -2.96764165e-01 4.20860559e-01
1.49551615e-01 3.06569695e-01 6.90190315e-01 -1.67102122e+00
-1.79502815e-01 -3.84484231e-01 -7.26059854e-01 -5.40811837e-01
6.46466970e-01 -1.11654305e+00 -5.37684083e-01 -2.17928004e+00
-4.75337021e-02 -7.83712506e-01 -2.16532335e-01 8.25107455e-01
-4.59367990e-01 -1.89117596e-01 1.90448284e-01 1.28906071e-01
-3.88798624e-01 5.18095374e-01 1.15099835e+00 -2.46757418e-01
-1.72347710e-01 -2.52916574e-01 -9.33795154e-01 8.09094071e-01
8.48826468e-01 -5.60593486e-01 -5.79502106e-01 -5.01671433e-01
8.24254513e-01 -8.33617002e-02 3.72333527e-01 -6.48720324e-01
4.37376589e-01 -3.14484090e-01 1.48007244e-01 -2.07559839e-01
2.98561245e-01 -1.06064689e+00 3.12488824e-01 3.26757550e-01
-3.35688025e-01 2.99476478e-02 -4.24314067e-02 8.41758966e-01
-4.19392407e-01 -2.16401756e-01 -1.84935302e-01 3.46721262e-02
-7.51181602e-01 9.26533565e-02 7.98476636e-02 8.03055242e-02
9.69560504e-01 -3.91946137e-01 -4.30799812e-01 -4.18074071e-01
-8.52830112e-01 3.81141722e-01 4.74966705e-01 5.80586970e-01
5.91955781e-01 -1.52569842e+00 -5.14763653e-01 -1.70414764e-02
2.41987213e-01 -1.17262520e-01 -2.37143129e-01 9.71621931e-01
-1.56628221e-01 5.15234768e-01 5.57760969e-02 2.87631340e-02
-1.23749125e+00 5.58218360e-01 2.17133760e-02 -3.94648880e-01
-5.12705743e-01 8.49295676e-01 -2.06285164e-01 -5.40360153e-01
1.41990989e-01 -4.18777645e-01 -3.82929593e-01 3.31657261e-01
-3.04426309e-02 3.92074049e-01 -2.10662901e-01 -4.14420575e-01
-4.51293349e-01 4.36940014e-01 -3.90182622e-02 1.36028066e-01
1.50015497e+00 2.12366983e-01 -1.81396559e-01 5.63221633e-01
7.58365870e-01 2.88696051e-01 -7.21433401e-01 -1.09478951e-01
2.55195946e-01 -5.73407531e-01 -1.41774148e-01 -9.15247738e-01
-1.04952991e+00 6.89198554e-01 -2.75353253e-01 3.64762545e-01
5.11113644e-01 2.92824894e-01 5.23690164e-01 2.95381755e-01
5.27424216e-01 -9.70746994e-01 -2.25969225e-01 4.49858397e-01
7.42143750e-01 -1.13450348e+00 1.42369613e-01 -5.77641428e-01
-6.11176372e-01 9.07462060e-01 6.84414089e-01 3.08784187e-01
7.06746101e-01 -2.46373206e-01 -1.41265437e-01 -2.88506508e-01
-9.27087307e-01 -3.38739872e-01 3.57926697e-01 6.61597490e-01
4.94471520e-01 3.90680999e-01 -5.00548720e-01 7.76302040e-01
-3.62419635e-01 -2.46435285e-01 3.73544097e-01 4.18517649e-01
-2.31358841e-01 -1.30169654e+00 -2.96027005e-01 9.52737749e-01
-1.91679448e-01 -3.40817750e-01 -8.58901083e-01 1.16506076e+00
4.89634305e-01 8.12137842e-01 -1.45891681e-01 -6.17339969e-01
3.81631732e-01 4.97430682e-01 2.74606615e-01 -6.72027111e-01
-5.31991661e-01 -8.44695568e-01 4.53836441e-01 -2.98756361e-01
-3.18973005e-01 -7.97687769e-01 -1.31445730e+00 -6.05480373e-01
-6.57408059e-01 4.35443640e-01 4.03837681e-01 8.74472499e-01
4.46340412e-01 4.45559055e-01 1.07330689e-02 -1.25979811e-01
-2.82040954e-01 -7.91448057e-01 -6.24850452e-01 5.09813666e-01
-6.49470538e-02 -1.03871298e+00 -4.57060605e-01 -3.15651149e-01] | [8.822402000427246, 7.8531718254089355] |
713116d7-6358-483f-9b68-b953759cf664 | sim-trans-structure-information-modeling | 2208.14607 | null | https://arxiv.org/abs/2208.14607v1 | https://arxiv.org/pdf/2208.14607v1.pdf | SIM-Trans: Structure Information Modeling Transformer for Fine-grained Visual Categorization | Fine-grained visual categorization (FGVC) aims at recognizing objects from similar subordinate categories, which is challenging and practical for human's accurate automatic recognition needs. Most FGVC approaches focus on the attention mechanism research for discriminative regions mining while neglecting their interdependencies and composed holistic object structure, which are essential for model's discriminative information localization and understanding ability. To address the above limitations, we propose the Structure Information Modeling Transformer (SIM-Trans) to incorporate object structure information into transformer for enhancing discriminative representation learning to contain both the appearance information and structure information. Specifically, we encode the image into a sequence of patch tokens and build a strong vision transformer framework with two well-designed modules: (i) the structure information learning (SIL) module is proposed to mine the spatial context relation of significant patches within the object extent with the help of the transformer's self-attention weights, which is further injected into the model for importing structure information; (ii) the multi-level feature boosting (MFB) module is introduced to exploit the complementary of multi-level features and contrastive learning among classes to enhance feature robustness for accurate recognition. The proposed two modules are light-weighted and can be plugged into any transformer network and trained end-to-end easily, which only depends on the attention weights that come with the vision transformer itself. Extensive experiments and analyses demonstrate that the proposed SIM-Trans achieves state-of-the-art performance on fine-grained visual categorization benchmarks. The code is available at https://github.com/PKU-ICST-MIPL/SIM-Trans_ACMMM2022. | ['Yuxin Peng', 'Xiangteng He', 'Hongbo Sun'] | 2022-08-31 | null | null | null | null | ['fine-grained-image-classification', 'fine-grained-visual-categorization'] | ['computer-vision', 'computer-vision'] | [ 1.55596789e-02 -3.94988954e-01 -1.10989228e-01 -4.05042380e-01
-4.85596418e-01 -4.42279547e-01 5.41961610e-01 -6.57616481e-02
-2.18550369e-01 7.77508393e-02 1.99557751e-01 -1.34787649e-01
-7.92086869e-02 -7.73054123e-01 -5.90461969e-01 -8.52388501e-01
2.20899671e-01 -4.34176847e-02 4.60759401e-01 -6.42548082e-03
3.93098325e-01 4.84879911e-01 -1.71880019e+00 5.59426785e-01
8.78386736e-01 1.66503179e+00 6.37102246e-01 2.70870447e-01
-2.40059450e-01 9.13674653e-01 -1.45445049e-01 -6.44850135e-02
1.40190050e-01 -9.96655747e-02 -6.88490450e-01 1.58694908e-01
4.13224816e-01 -2.50328124e-01 -2.30209440e-01 1.07741129e+00
3.83867800e-01 1.13931159e-03 7.97938108e-01 -1.15614212e+00
-9.40012872e-01 1.22760661e-01 -9.40775216e-01 5.88974893e-01
-1.37327924e-01 4.27341402e-01 8.96789551e-01 -1.18031526e+00
4.64100689e-02 1.36238527e+00 5.52755773e-01 3.08492720e-01
-1.00501835e+00 -7.90753424e-01 7.46862411e-01 6.71156347e-01
-1.51472175e+00 -4.03349936e-01 9.75333452e-01 -4.73498851e-01
9.04545546e-01 2.19701260e-01 6.00348949e-01 8.48652780e-01
1.24782197e-01 9.40896630e-01 1.10585761e+00 -2.49267131e-01
-2.83196792e-02 3.12072337e-01 5.01059055e-01 8.15454364e-01
3.35916467e-02 1.59948751e-01 -3.31212103e-01 2.76124537e-01
7.40315974e-01 3.74182045e-01 -2.13854477e-01 -4.28713650e-01
-9.67434645e-01 7.82577336e-01 1.08588552e+00 5.23846924e-01
-3.72698963e-01 6.64551631e-02 3.21214914e-01 8.65757465e-02
2.68672317e-01 -9.59764123e-02 -3.26324880e-01 4.11195993e-01
-5.81979871e-01 -1.24642849e-01 7.01038018e-02 9.49800789e-01
9.58274305e-01 -4.89230603e-02 -5.34829617e-01 1.08425593e+00
5.71768165e-01 2.27604464e-01 9.02259827e-01 -5.03441513e-01
3.65294874e-01 1.06324041e+00 -3.19156051e-01 -1.14615345e+00
-1.68917656e-01 -9.22453225e-01 -9.95190918e-01 1.07131943e-01
-1.05174482e-02 3.89260352e-01 -1.03622544e+00 1.66858840e+00
4.60573107e-01 2.33525619e-01 -3.19864511e-01 1.14467645e+00
1.15048528e+00 5.31708837e-01 3.42300683e-01 1.75408162e-02
1.64767969e+00 -1.34956539e+00 -1.25843763e-01 -4.82611597e-01
3.05045843e-01 -7.15606153e-01 1.24486983e+00 7.33727366e-02
-7.01482236e-01 -1.15288234e+00 -1.05140567e+00 -3.26136231e-01
-5.78177214e-01 4.32414651e-01 4.95102733e-01 2.87459552e-01
-7.68431544e-01 1.68139964e-01 -5.00007510e-01 -1.02459520e-01
8.62158477e-01 1.59257546e-01 -3.97630632e-01 -2.28182524e-01
-8.86008322e-01 6.42564774e-01 3.40840429e-01 4.45412010e-01
-1.24516952e+00 -5.91389954e-01 -8.50124180e-01 2.29895562e-01
1.21512629e-01 -6.22800529e-01 8.54455113e-01 -1.23309326e+00
-9.35250223e-01 8.31055760e-01 -2.63973296e-01 2.92826202e-02
2.37191081e-01 6.30502328e-02 -2.37411797e-01 5.10815680e-02
3.43554437e-01 6.80038452e-01 1.05176473e+00 -1.33141363e+00
-8.48479390e-01 -7.15753198e-01 -4.43628291e-03 2.77188808e-01
-4.54217196e-01 -4.78982031e-02 -6.42252266e-01 -7.75416493e-01
2.95199603e-02 -4.96799558e-01 2.70661227e-02 9.37967096e-03
-2.98469454e-01 -5.85888684e-01 1.04699349e+00 -5.85168421e-01
1.05158401e+00 -2.26705050e+00 3.07057379e-03 1.39012989e-02
1.53971553e-01 3.77780885e-01 -3.60617131e-01 1.01096906e-01
-2.16418833e-01 -1.80171028e-01 -1.08347293e-02 -2.09092259e-01
5.27048446e-02 -1.47316888e-01 -3.38475823e-01 4.50341403e-01
4.22469079e-01 1.15483892e+00 -5.51868737e-01 -4.98104274e-01
5.26561737e-01 4.57361758e-01 -2.98639894e-01 3.56142521e-01
-2.30662916e-02 2.28620440e-01 -7.23891914e-01 8.17801535e-01
9.13987398e-01 -2.96666145e-01 -1.68556720e-01 -7.88025439e-01
-1.84785739e-01 -5.92345297e-02 -1.05809081e+00 1.48202038e+00
-4.99858409e-01 1.36317953e-01 1.39912680e-01 -1.14584982e+00
8.68348956e-01 -1.61537513e-01 -7.31073245e-02 -9.40298915e-01
3.60865325e-01 -9.74018723e-02 9.97048914e-02 -4.91046637e-01
-1.11631528e-02 -4.15752083e-02 9.19380167e-04 2.62393862e-01
2.87074357e-01 4.01903600e-01 -1.20229289e-01 9.36303958e-02
7.82266140e-01 1.02471441e-01 2.45968089e-01 -3.95383596e-01
9.73085701e-01 -1.65975824e-01 6.56663060e-01 4.65093881e-01
-4.62552845e-01 3.41722697e-01 -2.63730511e-02 -3.87672603e-01
-6.88476920e-01 -1.04249060e+00 -2.18422711e-01 1.44059348e+00
5.53529739e-01 -1.46804258e-01 -5.08400500e-01 -8.61471534e-01
2.36883387e-01 3.93854141e-01 -1.06507444e+00 -3.75299037e-01
-2.43841290e-01 -5.49590588e-01 2.06684932e-01 9.73010838e-01
8.69646728e-01 -1.24968815e+00 -3.49543571e-01 -3.67146246e-02
-7.89451301e-02 -7.11665273e-01 -6.93169355e-01 3.04639369e-01
-4.81793255e-01 -1.06047022e+00 -5.11427939e-01 -1.09303486e+00
6.83365107e-01 8.09936225e-01 7.53381431e-01 1.88550130e-01
-5.69761753e-01 3.47857803e-01 -3.34283680e-01 -3.63130480e-01
2.68046796e-01 -5.21952054e-03 -1.54663011e-01 4.72031236e-01
5.33997118e-01 -6.00578010e-01 -9.47793067e-01 5.54092586e-01
-7.40109801e-01 1.20878607e-01 9.27492321e-01 1.13370872e+00
8.11559796e-01 4.17875499e-02 6.12881482e-01 -4.50801522e-01
2.97921419e-01 -4.00934905e-01 -4.53325778e-01 3.98782372e-01
-5.41891813e-01 -1.67723790e-01 7.60782659e-01 -3.41975570e-01
-1.18233347e+00 5.00837974e-02 -1.69372246e-01 -7.07984924e-01
-4.14353430e-01 1.74817473e-01 -6.37053370e-01 -2.81104535e-01
2.29087472e-01 7.72764862e-01 -1.66814491e-01 -6.01189792e-01
3.80122125e-01 8.49149585e-01 4.56082016e-01 -4.77480739e-01
7.27233768e-01 3.75415474e-01 -4.33471322e-01 -5.70461154e-01
-1.02172124e+00 -6.22448385e-01 -7.91470468e-01 -1.40609741e-01
8.19966435e-01 -1.09828484e+00 -6.94087505e-01 6.36734545e-01
-7.70877659e-01 -1.57493427e-01 -6.38765246e-02 1.25515014e-01
-3.35498989e-01 3.52799654e-01 -5.59504986e-01 -5.69365382e-01
-5.65629601e-01 -1.16984582e+00 1.22342873e+00 6.04075968e-01
4.06199992e-01 -7.37139821e-01 -3.33307892e-01 5.49886823e-01
4.16095734e-01 -2.96549946e-01 8.86022091e-01 -3.82336378e-01
-6.79277897e-01 1.11948952e-01 -8.84051919e-01 6.05727315e-01
2.29014948e-01 -3.20092469e-01 -1.31756735e+00 -4.44599748e-01
9.92454961e-03 -4.95582312e-01 1.26440215e+00 3.24624360e-01
1.59345531e+00 -3.15897226e-01 -5.14629066e-01 6.88408554e-01
1.62065613e+00 1.66864216e-01 4.99721706e-01 2.54563123e-01
1.04089510e+00 5.56462228e-01 6.67495847e-01 1.69832394e-01
5.20899177e-01 6.71267927e-01 5.50925374e-01 -1.96298048e-01
-3.85364264e-01 -2.48358980e-01 2.05733806e-01 7.08216786e-01
1.09223366e-01 1.85991451e-01 -5.70172966e-01 6.91363990e-01
-1.76118875e+00 -9.02202070e-01 1.74398541e-01 1.77568340e+00
6.38944924e-01 2.28884481e-02 -2.04713568e-02 2.44916342e-02
7.67624438e-01 2.22501487e-01 -7.97718942e-01 -1.52262524e-01
9.40762926e-03 -2.35064421e-02 8.53433684e-02 1.83433980e-01
-1.25334644e+00 9.64024305e-01 4.19885015e+00 1.44458270e+00
-1.18574655e+00 2.31950119e-01 7.81814277e-01 2.41812021e-01
-1.75637975e-01 -2.57532835e-01 -8.36146593e-01 5.80023348e-01
1.94616452e-01 2.00830638e-01 2.94403702e-01 9.75912392e-01
-2.82420740e-02 1.20743319e-01 -7.68021703e-01 9.62843657e-01
4.99937981e-02 -1.13402808e+00 2.93061197e-01 -1.52842224e-01
3.41170073e-01 -6.98284386e-03 9.33989137e-02 4.86696601e-01
-6.94418848e-02 -1.00772834e+00 8.74334872e-01 4.54879493e-01
7.94234335e-01 -7.39777267e-01 7.01606750e-01 4.06440735e-01
-1.75953710e+00 -5.47046423e-01 -6.96540952e-01 8.08630511e-02
-3.04830283e-01 4.66287345e-01 -4.17468309e-01 6.80850983e-01
9.72119987e-01 1.00517797e+00 -8.80191803e-01 9.39470112e-01
-1.17828436e-01 5.50465524e-01 8.27358197e-03 1.29104033e-01
4.20615971e-01 -5.82596138e-02 1.70886442e-01 1.20902467e+00
2.40604132e-02 7.03076497e-02 3.16005588e-01 8.21663260e-01
1.11104287e-01 8.15447643e-02 -2.77096510e-01 2.95568138e-01
4.01168108e-01 1.69818819e+00 -6.73711419e-01 -2.45595336e-01
-5.78571975e-01 9.95301187e-01 6.53250098e-01 2.03850359e-01
-8.59944284e-01 -4.58454847e-01 5.80995142e-01 6.76635578e-02
8.45332980e-01 1.50359228e-01 -2.13886306e-01 -1.15620041e+00
1.87001944e-01 -8.30427825e-01 5.86181343e-01 -7.79311240e-01
-1.70673084e+00 7.75389910e-01 -3.26885849e-01 -1.18997204e+00
3.66405278e-01 -7.59777546e-01 -7.43615925e-01 1.10645533e+00
-1.77746904e+00 -1.80207384e+00 -8.17393541e-01 9.99678314e-01
8.20722759e-01 -1.99834093e-01 6.72409415e-01 3.18151474e-01
-8.54144335e-01 8.41446817e-01 -2.12835267e-01 2.53200829e-01
4.79257226e-01 -9.55592632e-01 2.78648417e-02 9.13469791e-01
-2.06381176e-02 6.68222725e-01 1.44697269e-02 -3.94939184e-01
-1.29332626e+00 -1.53040493e+00 3.98285359e-01 -2.17397422e-01
4.68099415e-01 -4.16140616e-01 -9.50062454e-01 5.50672650e-01
1.57486796e-02 3.28826547e-01 4.43593204e-01 -4.13016267e-02
-8.29405189e-01 -5.87541282e-01 -9.73024726e-01 1.31822050e-01
1.07055485e+00 -6.40297949e-01 -8.21596324e-01 1.04188815e-01
5.07129371e-01 3.18984799e-02 -6.14054501e-01 5.27122378e-01
6.30751729e-01 -9.83510256e-01 1.06832361e+00 -4.17613983e-01
3.07331681e-01 -7.65492380e-01 -3.49509835e-01 -1.05424595e+00
-1.03471088e+00 3.89680952e-01 1.18791714e-01 1.54778314e+00
-1.06492499e-03 -6.21794522e-01 4.13269490e-01 -2.96981074e-03
-2.75092870e-01 -9.74885643e-01 -8.64907384e-01 -6.41919255e-01
-1.21094123e-01 -3.68611246e-01 4.74517912e-01 8.18816602e-01
-3.66027296e-01 6.58191383e-01 -3.58619094e-02 2.18866616e-01
6.94475293e-01 7.04048693e-01 3.98602337e-01 -9.84272480e-01
-2.79859960e-01 -6.05594158e-01 -4.76891667e-01 -1.20232356e+00
7.14566484e-02 -1.00004363e+00 7.96100721e-02 -1.56490147e+00
7.05105305e-01 -7.07986414e-01 -8.40332150e-01 7.02719629e-01
-3.47706258e-01 5.41887224e-01 2.24388301e-01 2.88087398e-01
-7.52830982e-01 8.64026368e-01 1.31699622e+00 -6.30984783e-01
2.19571024e-01 -1.18219294e-01 -1.13911736e+00 6.79134488e-01
4.94244635e-01 -9.84098539e-02 -5.97639441e-01 -4.02660996e-01
-5.72101235e-01 -4.41638410e-01 7.82480776e-01 -1.07255423e+00
2.51952261e-01 -8.60992372e-02 9.53662038e-01 -6.36037469e-01
2.28356138e-01 -8.36004138e-01 -1.47781104e-01 4.17514354e-01
-1.66912451e-01 -1.28876328e-01 2.81838268e-01 5.21393061e-01
-3.60727578e-01 9.38662328e-03 1.03195572e+00 -1.66435376e-01
-1.22162414e+00 6.03096128e-01 -7.31029734e-02 -8.69319662e-02
1.17194498e+00 -1.96338311e-01 -4.80588555e-01 1.08952887e-01
-4.91020381e-01 3.40783834e-01 3.70247036e-01 5.94132721e-01
7.87064672e-01 -1.45410776e+00 -5.61678648e-01 4.63799983e-01
5.12771904e-01 -6.56936094e-02 8.42581511e-01 7.00412095e-01
-1.00719228e-01 4.00426865e-01 -5.51423371e-01 -7.83197761e-01
-1.19308257e+00 9.90736842e-01 4.56686437e-01 -1.47383809e-01
-4.00632799e-01 1.08910334e+00 1.05040061e+00 -2.79001236e-01
2.30088413e-01 -1.85949534e-01 -4.67741370e-01 1.11542098e-01
7.79586613e-01 3.20077501e-02 9.36133694e-03 -8.74039948e-01
-7.78767705e-01 9.04324293e-01 -3.98573279e-01 4.70917016e-01
1.25324416e+00 -3.52495253e-01 -1.69969559e-01 1.23435542e-01
1.26382840e+00 -1.95810184e-01 -1.33123493e+00 -4.43688244e-01
-4.17117983e-01 -4.69276696e-01 2.29883149e-01 -9.69930112e-01
-1.23102891e+00 1.21135008e+00 1.01706338e+00 -6.52858317e-02
1.46692300e+00 1.68694854e-01 6.10913873e-01 -9.52117294e-02
4.04286236e-01 -6.93547964e-01 1.69148490e-01 3.61890465e-01
1.09986401e+00 -1.20655799e+00 -1.71772689e-01 -4.29746300e-01
-6.47494555e-01 8.01777780e-01 1.01427317e+00 -2.57556319e-01
8.52867305e-01 3.80813070e-02 8.99003372e-02 -8.31656083e-02
-7.15274215e-01 -4.61676925e-01 6.56733096e-01 7.24133193e-01
2.02510610e-01 7.28005618e-02 -3.16093192e-02 1.17505193e+00
2.16532543e-01 -2.24662378e-01 -2.68512726e-01 5.62948346e-01
-6.67809963e-01 -7.94902086e-01 -1.71380758e-01 6.02535903e-01
-5.26775531e-02 -2.37437755e-01 -2.34942749e-01 4.50488776e-01
7.20264137e-01 9.23178256e-01 9.91903916e-02 -6.58349454e-01
2.51450479e-01 -2.49072596e-01 4.51511204e-01 -4.43981946e-01
-6.55121922e-01 6.13240115e-02 -3.10739607e-01 -6.64018154e-01
-2.52673745e-01 -4.23000187e-01 -9.14805472e-01 -1.24060828e-02
-2.39036396e-01 1.14927702e-01 1.74063072e-01 8.88050556e-01
5.09253263e-01 7.33857274e-01 8.24414194e-01 -9.46468353e-01
-3.99794519e-01 -1.14399052e+00 -5.62489510e-01 3.41897726e-01
3.08610857e-01 -9.52957988e-01 -1.63005367e-01 -8.70007798e-02] | [9.658470153808594, 2.0048506259918213] |
b5518321-f24b-4f3b-ab05-613309b5a8e4 | fusion-in-t5-unifying-document-ranking | 2305.14685 | null | https://arxiv.org/abs/2305.14685v1 | https://arxiv.org/pdf/2305.14685v1.pdf | Fusion-in-T5: Unifying Document Ranking Signals for Improved Information Retrieval | Common IR pipelines are typically cascade systems that may involve multiple rankers and/or fusion models to integrate different information step-by-step. In this paper, we propose a novel re-ranker named Fusion-in-T5 (FiT5), which integrates document text information, retrieval features, and global document information into a single unified model using templated-based input and global attention. Experiments on passage ranking benchmarks MS MARCO and TREC DL show that FiT5 significantly improves ranking performance over prior pipelines. Analyses find that through global attention, FiT5 is able to jointly utilize the ranking features via gradually attending to related documents, and thus improve the detection of subtle nuances between them. Our code will be open-sourced. | ['Zhenghao Liu', 'Zhiyuan Liu', 'David Jin', 'Chenyan Xiong', 'Chenghao Fan', 'Shi Yu'] | 2023-05-24 | null | null | null | null | ['document-ranking', 'passage-ranking'] | ['natural-language-processing', 'natural-language-processing'] | [-2.03555033e-01 -5.11821806e-01 -1.49941593e-01 -4.67234105e-01
-1.67149866e+00 -9.92298186e-01 1.09830236e+00 4.64208275e-01
-4.57096666e-01 2.97861874e-01 7.49797761e-01 -2.54441023e-01
-5.78527510e-01 -2.14141741e-01 -5.03580868e-01 -1.66169740e-02
1.04909882e-01 6.07273102e-01 5.30761540e-01 -5.88703275e-01
6.71106815e-01 1.93624154e-01 -1.20887423e+00 1.07403255e+00
8.77812624e-01 5.58889031e-01 9.01411325e-02 1.02711141e+00
-1.18516386e-01 7.27752626e-01 -7.71133244e-01 -4.43495899e-01
6.46840334e-02 -9.13879275e-02 -9.99726117e-01 -7.62502730e-01
8.46243680e-01 -5.96129060e-01 -4.82792169e-01 7.88438141e-01
6.17208719e-01 2.85845548e-01 6.07610583e-01 -6.17106795e-01
-8.65879714e-01 8.63169134e-01 -5.26229382e-01 5.00343680e-01
5.09294271e-01 -2.75986493e-01 1.46829259e+00 -1.17125392e+00
3.88936400e-01 1.68674505e+00 3.55592281e-01 2.46200264e-01
-1.15411460e+00 -5.17589390e-01 2.32008219e-01 1.67982161e-01
-1.13701677e+00 -3.75482708e-01 4.00081724e-01 -2.13059843e-01
1.08524275e+00 3.92744273e-01 8.21746662e-02 8.79609942e-01
2.56686330e-01 1.08329380e+00 8.13158989e-01 -4.69176412e-01
-2.36063272e-01 -1.56809032e-01 1.05630732e+00 5.65942585e-01
3.25065374e-01 -4.67585884e-02 -6.82862818e-01 -3.47922891e-01
3.54638934e-01 1.48227468e-01 -1.99731022e-01 2.79032826e-01
-9.76042092e-01 6.37455165e-01 5.44434905e-01 3.29036206e-01
-2.08478317e-01 2.79317290e-01 2.85312474e-01 5.39912224e-01
5.94313681e-01 5.20728707e-01 -5.30933082e-01 1.69724390e-01
-1.05727673e+00 4.91983503e-01 4.43892092e-01 7.03085899e-01
4.82266784e-01 -8.50721180e-01 -1.06532192e+00 9.75947559e-01
6.12110436e-01 3.45171899e-01 6.01419270e-01 -7.60930002e-01
5.60859442e-01 6.81428909e-01 3.51481318e-01 -5.88689983e-01
-4.64984030e-01 -6.04321122e-01 -1.33765832e-01 -1.96901113e-01
3.50132659e-02 1.29620850e-01 -1.20222354e+00 1.43321013e+00
-1.03362560e-01 -2.71200955e-01 2.61912439e-02 8.40099454e-01
1.20762169e+00 5.41657805e-01 2.37017632e-01 3.51658314e-01
1.59607637e+00 -1.25457227e+00 -6.61140084e-01 -1.95093989e-01
6.96161211e-01 -1.08282995e+00 9.32444453e-01 2.20618680e-01
-1.39600849e+00 -6.92419231e-01 -1.06083691e+00 -7.13407040e-01
-4.36568677e-01 3.72665346e-01 5.13053417e-01 3.18680674e-01
-1.40099621e+00 5.42223573e-01 -5.61201811e-01 -3.52211326e-01
1.07871659e-01 4.17077512e-01 -8.92401189e-02 -3.92539769e-01
-1.53600299e+00 9.08649802e-01 1.94263384e-01 1.74298033e-01
-9.06494856e-01 -4.39135820e-01 -4.76449758e-01 4.92029786e-01
2.96558421e-02 -9.28424537e-01 1.67054534e+00 -3.61419737e-01
-1.38664818e+00 3.33820730e-01 -3.75523895e-01 -1.54381856e-01
4.24070954e-01 -9.93425131e-01 -3.57607037e-01 2.94622421e-01
1.30115852e-01 6.01880312e-01 4.98388737e-01 -1.07934749e+00
-5.60994565e-01 -2.91720122e-01 3.27895731e-01 6.40560567e-01
-3.06170553e-01 6.52440012e-01 -9.87223327e-01 -3.89386743e-01
5.45136742e-02 -6.29241168e-01 6.14749379e-02 -6.17565453e-01
-2.19487727e-01 -7.37021625e-01 4.64722067e-01 -6.36700869e-01
1.46553195e+00 -1.95431423e+00 -5.74463606e-02 2.99140632e-01
3.45423669e-01 2.68549144e-01 -6.31715178e-01 4.87583876e-01
1.96823344e-01 4.45094287e-01 4.81787652e-01 -4.88363624e-01
1.48073938e-02 -4.11394298e-01 -3.48899037e-01 -4.64071473e-03
2.67896444e-01 1.17632842e+00 -1.18843448e+00 -5.71125090e-01
-2.03354776e-01 4.98864204e-01 -5.14135897e-01 -1.17483862e-01
-2.00977147e-01 9.66101661e-02 -7.60470510e-01 5.41960955e-01
3.79421294e-01 -4.62716192e-01 8.03673565e-02 -1.62310183e-01
1.60492845e-02 7.32081831e-01 -7.30933070e-01 1.68867040e+00
-2.10364044e-01 7.24919438e-01 -2.07114354e-01 -3.76128793e-01
6.09002233e-01 3.50147009e-01 1.90338254e-01 -7.67971277e-01
1.71275645e-01 1.55185431e-01 -1.37677267e-01 -8.72368142e-02
1.08185351e+00 3.32936496e-01 -2.00515822e-01 4.79883879e-01
2.02267244e-01 2.49639675e-01 4.36267316e-01 9.69150543e-01
1.49642146e+00 -5.38479164e-02 -1.11678526e-01 -8.42165276e-02
4.36049312e-01 -8.54454003e-04 1.55368462e-01 1.53977299e+00
1.02383457e-01 7.41283894e-01 3.64408106e-01 -6.25610724e-02
-7.08730757e-01 -1.01896572e+00 -1.07650958e-01 1.75569904e+00
7.66851157e-02 -7.01473475e-01 -2.62706041e-01 -8.10320973e-01
-3.44728567e-02 4.98902261e-01 -5.95600367e-01 -2.58075297e-01
-4.09794837e-01 -6.69371068e-01 5.16422451e-01 6.36840403e-01
2.60201693e-01 -8.12574923e-01 7.70136416e-02 3.28023165e-01
-2.80034810e-01 -6.94884479e-01 -8.78509045e-01 1.33159593e-01
-8.86944830e-01 -1.00762904e+00 -8.87561381e-01 -4.55580980e-01
3.48113865e-01 8.73815775e-01 1.21600139e+00 4.20087129e-01
2.09909212e-02 6.00214362e-01 -6.31296873e-01 -2.59736210e-01
-1.62120178e-01 2.56646693e-01 -2.26861745e-01 -3.19358081e-01
3.84208590e-01 2.61020631e-01 -7.83035576e-01 3.45254213e-01
-9.05096531e-01 -1.77761748e-01 8.03410232e-01 9.46395278e-01
1.59985512e-01 -2.03585953e-01 4.56785172e-01 -7.58674979e-01
1.22072911e+00 -4.19891298e-01 -3.73460472e-01 6.88235700e-01
-6.23692214e-01 3.36913824e-01 9.71388966e-02 -2.32665539e-01
-1.33106244e+00 -3.94798577e-01 1.84482217e-01 -1.80461362e-01
3.46048534e-01 8.39101553e-01 2.34315604e-01 1.61650434e-01
7.70529926e-01 -1.47635072e-01 -5.53222835e-01 -8.15497875e-01
5.43465614e-01 6.82216942e-01 3.86264563e-01 -7.69689620e-01
6.02915764e-01 2.20734060e-01 -5.34602761e-01 -1.71562091e-01
-1.25814426e+00 -9.76548910e-01 -3.39622766e-01 -1.82201728e-01
6.49129808e-01 -1.19878471e+00 -5.37070990e-01 2.88662076e-01
-1.32005405e+00 -5.35661122e-03 3.47274631e-01 5.89915097e-01
1.56740770e-01 3.38064790e-01 -1.16719067e+00 -3.79191816e-01
-7.40908504e-01 -1.04120541e+00 1.31481099e+00 4.92720813e-01
-6.53687119e-02 -6.78588808e-01 3.67908865e-01 6.99176073e-01
3.85071337e-01 -6.46299243e-01 9.04196262e-01 -1.15491617e+00
-5.33216119e-01 -6.23226345e-01 -5.24751902e-01 1.17584825e-01
-1.37402311e-01 1.80712357e-01 -8.67514014e-01 -4.51445937e-01
-7.43205607e-01 -2.45312259e-01 1.52080166e+00 3.16404700e-01
8.17464292e-01 7.62007432e-03 -3.95625293e-01 -8.33814070e-02
1.12172580e+00 2.53394209e-02 5.67954481e-01 3.27664942e-01
6.31857634e-01 3.64133030e-01 5.69201529e-01 1.39206976e-01
3.58170420e-01 6.28099740e-01 -7.22130165e-02 -5.61242141e-02
-5.55250086e-02 -1.14099912e-01 6.25112891e-01 7.36973822e-01
-3.23103070e-02 -5.89073479e-01 -7.28676617e-01 3.88821959e-01
-1.94131374e+00 -1.01230597e+00 -1.13626651e-01 2.06278992e+00
7.59509504e-01 3.17402273e-01 -2.01483130e-01 -2.91883916e-01
8.86326611e-01 -1.11181110e-01 -3.69029075e-01 -3.07524413e-01
-6.98286742e-02 2.36018211e-01 2.50413001e-01 7.18207955e-01
-1.08038533e+00 1.05884552e+00 7.00288248e+00 7.37902403e-01
-6.81740940e-01 -1.81844328e-02 4.34651494e-01 -3.91744971e-01
-4.61019486e-01 8.36406425e-02 -1.18627846e+00 1.15206130e-01
9.39690113e-01 -2.35132903e-01 2.39733875e-01 5.19317448e-01
1.21621899e-02 1.32011667e-01 -1.03012943e+00 3.92599493e-01
-5.55005483e-03 -1.10154653e+00 4.64091480e-01 -1.08794317e-01
6.49230361e-01 5.01949847e-01 1.87150910e-01 6.90777838e-01
1.03936267e+00 -7.02988982e-01 6.17010772e-01 7.38328636e-01
3.41317773e-01 -4.89637733e-01 1.02430606e+00 -8.05717260e-02
-1.09410679e+00 -2.22190410e-01 -2.14142218e-01 3.16203326e-01
-1.28987879e-01 4.24172103e-01 -7.61591613e-01 7.69688845e-01
8.02178800e-01 5.82362831e-01 -9.94444609e-01 1.33202767e+00
-2.45380878e-01 5.15083075e-01 -1.50518358e-01 -1.43544286e-01
3.16189438e-01 3.57321560e-01 5.71373224e-01 1.41666555e+00
2.54543841e-01 2.03658164e-01 2.21644938e-01 4.71106976e-01
-3.46149623e-01 1.55888781e-01 7.69321248e-02 -7.62163028e-02
3.96539956e-01 1.45432031e+00 -5.32175899e-01 -6.64473414e-01
-4.96223807e-01 8.54885519e-01 5.19325435e-01 6.25859201e-01
-4.94263500e-01 -5.15923321e-01 3.11674714e-01 -2.73370177e-01
1.57982245e-01 -5.66324452e-03 -8.39925110e-02 -1.18034101e+00
-1.24512568e-01 -7.13102579e-01 8.55445504e-01 -1.08240402e+00
-1.31152868e+00 6.32608414e-01 1.36285692e-01 -9.82336879e-01
2.57482864e-02 -5.25577605e-01 -6.50345981e-01 1.02044797e+00
-1.54401600e+00 -9.88383651e-01 -1.23211645e-01 5.40036142e-01
6.40250266e-01 -3.20945643e-02 5.27744293e-01 1.71440423e-01
-4.26252037e-01 6.55882597e-01 5.56297481e-01 2.50562906e-01
1.28785408e+00 -1.36245584e+00 2.75581867e-01 6.44122541e-01
2.44892389e-01 1.37817979e+00 3.08999091e-01 -7.16101110e-01
-1.02983260e+00 -7.36742795e-01 1.11847842e+00 -8.48471045e-01
5.54188132e-01 -1.19658113e-01 -1.07018042e+00 5.50763249e-01
3.24694902e-01 -4.04465139e-01 6.28153503e-01 7.80950069e-01
-7.80563653e-01 -1.23235188e-01 -7.19070017e-01 5.83242655e-01
5.79395652e-01 -8.27527404e-01 -9.94433463e-01 3.18185687e-01
8.92549813e-01 -2.78916359e-01 -7.69057870e-01 4.16759729e-01
7.25814104e-01 -4.71784353e-01 1.12966084e+00 -7.45901644e-01
3.78379285e-01 -3.85408580e-01 -2.41988664e-03 -1.18819129e+00
-8.34342837e-01 -5.32794237e-01 -2.06732348e-01 1.16774178e+00
5.73257565e-01 -2.06725493e-01 2.07843587e-01 5.45237303e-01
-3.26232910e-01 -1.99260309e-01 -5.05310178e-01 -4.76459235e-01
2.44495094e-01 -1.78131849e-01 4.78286952e-01 6.87105894e-01
2.51027286e-01 8.35556090e-01 -9.78262126e-02 1.72379106e-01
3.96185637e-01 -1.01742879e-01 2.74226755e-01 -1.52828932e+00
-1.59760520e-01 -8.24703217e-01 2.16150418e-01 -1.27881062e+00
-1.25581756e-01 -1.07990634e+00 2.31942147e-01 -1.90724254e+00
7.19577670e-01 -1.77283481e-01 -1.19854593e+00 6.07693195e-01
-7.66800284e-01 2.35343084e-01 1.04325101e-01 3.99531424e-01
-1.35965466e+00 3.86305273e-01 1.11685121e+00 -3.15262794e-01
-3.86680305e-01 -1.42255917e-01 -1.13152146e+00 3.15819919e-01
5.73350072e-01 -5.21669447e-01 -1.75869286e-01 -6.48565292e-01
5.49024522e-01 5.95701560e-02 2.14621201e-01 -6.53900802e-01
6.22445822e-01 3.38139713e-01 7.24477470e-01 -1.06419146e+00
8.52280334e-02 -2.76228756e-01 -5.81755161e-01 1.32584944e-01
-1.02572536e+00 2.13277236e-01 2.98775554e-01 7.17783868e-01
-2.29308486e-01 -3.31806391e-01 2.03836471e-01 -1.00517362e-01
-5.75230241e-01 1.15317523e-01 -4.47875768e-01 -1.35204881e-01
-7.83848539e-02 4.41098720e-01 -9.56995368e-01 -4.71079260e-01
-5.08880556e-01 6.83859706e-01 6.19472787e-02 7.83335507e-01
4.70481217e-01 -1.17371047e+00 -1.15941048e+00 -2.83117592e-01
3.01077187e-01 -2.87464023e-01 3.33174735e-01 4.13175136e-01
-1.61396459e-01 9.46025848e-01 2.48874471e-01 -3.71448278e-01
-1.55713046e+00 3.92590582e-01 1.31964758e-01 -9.14181471e-01
-1.88754812e-01 7.66442657e-01 2.44436383e-01 -2.86483556e-01
1.38230816e-01 -2.61158049e-01 -6.03132248e-01 9.86268669e-02
9.19722438e-01 2.66654521e-01 2.80494153e-01 -2.31139317e-01
-2.70283520e-01 5.71381211e-01 -1.00646615e+00 -4.88773614e-01
1.00609100e+00 -1.91260502e-01 -1.09962314e-01 6.83689788e-02
1.05851972e+00 8.48651677e-02 -9.77140844e-01 -7.00653493e-01
3.68328363e-01 -2.27904856e-01 5.21939218e-01 -1.32112277e+00
-6.47872925e-01 6.22917414e-01 5.26359916e-01 -2.84931716e-02
9.62855101e-01 2.25390792e-01 4.27587926e-01 6.45440400e-01
-4.04593088e-02 -1.01650000e+00 4.22183573e-01 9.43271995e-01
9.62282479e-01 -1.17477775e+00 3.07516754e-01 7.26729631e-02
-4.79777277e-01 1.02664578e+00 5.02385378e-01 7.29340538e-02
4.75360543e-01 -3.12785208e-01 -2.15804055e-02 -3.31793100e-01
-1.24790049e+00 -3.61265957e-01 1.17063928e+00 -1.61527991e-01
8.32955301e-01 -2.69966036e-01 -5.62303841e-01 7.89935172e-01
2.47427180e-01 -9.25965011e-02 6.94210827e-02 8.89422476e-01
-6.51780903e-01 -1.26966333e+00 -3.91146928e-01 5.47074556e-01
-7.78653204e-01 -7.41105914e-01 -5.27855814e-01 3.46299469e-01
-4.57889289e-01 1.22284758e+00 4.59529646e-02 -4.18133974e-01
2.40140930e-01 2.11330757e-01 3.46086234e-01 -6.75096929e-01
-1.13896406e+00 5.41389465e-01 2.51521051e-01 -4.88604069e-01
-1.36553735e-01 -6.42840803e-01 -1.05301976e+00 -3.10055781e-02
-5.45074642e-01 3.35132211e-01 6.46673143e-01 8.76823068e-01
6.44348323e-01 7.44173706e-01 4.67085660e-01 -5.43043673e-01
-5.45962989e-01 -1.36543667e+00 -1.85714185e-01 2.62990266e-01
3.12787086e-01 -4.95936513e-01 -2.72901982e-01 -3.37675512e-01] | [11.485848426818848, 7.611485481262207] |
51577bfc-5910-45f0-bbf5-c2f3fcdf684d | non-deterministic-oracles-for-unrestricted | null | null | https://aclanthology.org/W15-2210 | https://aclanthology.org/W15-2210.pdf | Non-Deterministic Oracles for Unrestricted Non-Projective Transition-Based Dependency Parsing | null | ['Anders Bj{\\"o}rkelund', 'Joakim Nivre'] | 2015-07-01 | null | null | null | ws-2015-7 | ['transition-based-dependency-parsing'] | ['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] | [-7.310624122619629, 3.7910807132720947] |
447479e5-6873-4792-a7f6-3da35e6823b5 | improving-robustness-of-facial-landmark | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Zhu_Improving_Robustness_of_Facial_Landmark_Detection_by_Defending_Against_Adversarial_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Zhu_Improving_Robustness_of_Facial_Landmark_Detection_by_Defending_Against_Adversarial_ICCV_2021_paper.pdf | Improving Robustness of Facial Landmark Detection by Defending Against Adversarial Attacks | Many recent developments in facial landmark detection have been driven by stacking model parameters or augmenting annotations. However, three subsequent challenges remain, including 1) an increase in computational overhead, 2) the risk of overfitting caused by increasing model parameters, and 3) the burden of labor-intensive annotation by humans. We argue that exploring the weaknesses of the detector so as to remedy them is a promising method of robust facial landmark detection. To achieve this, we propose a sample-adaptive adversarial training (SAAT) approach to interactively optimize an attacker and a detector, which improves facial landmark detection as a defense against sample-adaptive black-box attacks. By leveraging adversarial attacks, the proposed SAAT exploits adversarial perturbations beyond the handcrafted transformations to improve the detector. Specifically, an attacker generates adversarial perturbations to reflect the weakness of the detector. Then, the detector must improve its robustness to adversarial perturbations to defend against adversarial attacks. Moreover, a sample-adaptive weight is designed to balance the risks and benefits of augmenting adversarial examples to train the detector. We also introduce a masked face alignment dataset, Masked-300W, to evaluate our method. Experiments show that our SAAT performed comparably to existing state-of-the-art methods. The dataset and model are publicly available at https://github.com/zhuccly/SAAT. | ['Songmin Dai', 'Jide Li', 'Xiaoqiang Li', 'Congcong Zhu'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['face-alignment'] | ['computer-vision'] | [ 3.04876000e-01 2.83336371e-01 9.63329524e-02 -3.33152145e-01
-1.00527108e+00 -7.42228925e-01 5.59238851e-01 -5.07250667e-01
-1.77773982e-01 1.42377630e-01 -6.83750659e-02 -1.81463942e-01
3.42909575e-01 -4.31995213e-01 -6.94589078e-01 -8.54046822e-01
-3.13547589e-02 6.50732964e-02 1.49046510e-01 -3.20092827e-01
9.50107649e-02 8.07822824e-01 -1.15110075e+00 8.96890610e-02
6.76081061e-01 9.41333115e-01 -4.27050292e-01 5.22673130e-01
2.33898371e-01 3.68073791e-01 -6.98520541e-01 -8.19118321e-01
8.52781773e-01 -3.65944713e-01 -4.91949081e-01 6.04667142e-02
7.96635091e-01 -5.37076771e-01 -2.64299452e-01 1.34781933e+00
8.63039434e-01 -1.53220549e-01 4.21763390e-01 -1.67586207e+00
-5.77731729e-01 4.43023533e-01 -9.02439415e-01 -8.20648223e-02
2.55694181e-01 5.15190780e-01 7.09750950e-01 -1.18478072e+00
3.07752937e-01 1.34456170e+00 7.63883233e-01 9.54483092e-01
-1.11259925e+00 -1.21842670e+00 2.43434593e-01 -9.34861004e-02
-1.63584983e+00 -8.39763641e-01 1.01976788e+00 -3.72811347e-01
2.17535838e-01 3.12895536e-01 1.98524714e-01 1.33317423e+00
-8.07312280e-02 5.31383693e-01 7.82336771e-01 -3.66812050e-01
8.55136588e-02 1.16671354e-01 -4.08436954e-01 9.61033583e-01
-4.09017354e-02 3.90010267e-01 -3.16925406e-01 -6.41206980e-01
6.65784299e-01 -2.54605055e-01 -1.55766070e-01 -3.05217385e-01
-6.25455797e-01 7.00018585e-01 4.63030547e-01 -2.26320565e-01
-1.30266473e-01 1.25888556e-01 3.93720478e-01 9.95174050e-02
3.31319600e-01 3.98483127e-01 -4.59031940e-01 5.14786780e-01
-6.61924660e-01 8.79211798e-02 3.92484277e-01 8.52848291e-01
5.82411885e-01 3.51218641e-01 -1.66513070e-01 6.87546492e-01
5.34243047e-01 7.37813711e-01 1.73086017e-01 -8.99716198e-01
4.60489213e-01 4.67846543e-01 -1.54187962e-01 -1.06543267e+00
-2.22721323e-01 -1.56827107e-01 -5.73871493e-01 6.66338742e-01
5.12679636e-01 -2.86051065e-01 -1.02681553e+00 2.00660944e+00
6.64719760e-01 4.52254832e-01 -1.04488418e-01 7.71060824e-01
6.35581076e-01 2.05162138e-01 1.30326986e-01 1.07524768e-01
1.31930304e+00 -8.44383597e-01 -5.64804137e-01 -3.26357186e-01
5.21195889e-01 -1.02754378e+00 1.13561952e+00 4.46931049e-02
-9.18018997e-01 -4.40655023e-01 -1.02469194e+00 2.17646614e-01
-1.54970467e-01 3.10152560e-01 2.14399964e-01 1.10877204e+00
-7.23037481e-01 2.64716178e-01 -8.05528820e-01 -1.25775591e-01
8.35874438e-01 6.02717400e-01 -4.47447747e-01 3.17689836e-01
-9.65326428e-01 6.56227767e-01 -5.80133237e-02 2.25135222e-01
-1.03792250e+00 -7.26212442e-01 -9.19085264e-01 -1.90332197e-02
6.03790045e-01 -2.99285412e-01 1.13429821e+00 -1.03765762e+00
-1.68499410e+00 8.14455628e-01 1.66629069e-02 -2.96034068e-01
7.86881328e-01 -3.78231078e-01 -3.43553692e-01 5.76589257e-02
-1.87300965e-01 4.84741092e-01 1.28643155e+00 -1.39084530e+00
-2.23019436e-01 -4.14238632e-01 -8.22196528e-02 -1.53690189e-01
-5.34059703e-01 3.88065308e-01 -4.74618971e-01 -9.94544983e-01
-1.36796862e-01 -1.17165983e+00 -3.08678120e-01 4.09478754e-01
-6.35223925e-01 1.68414280e-01 9.92417216e-01 -4.95989680e-01
1.05202520e+00 -2.43792391e+00 -2.18811035e-01 4.34241682e-01
1.12337828e-01 7.44133532e-01 -5.45524478e-01 1.46050125e-01
-2.97321260e-01 4.05483782e-01 -1.89841047e-01 -5.04330277e-01
-1.35267764e-01 -4.53668684e-02 -6.09664857e-01 6.75902128e-01
5.35759568e-01 7.84953237e-01 -6.24917805e-01 -4.61257011e-01
-3.09973229e-02 4.98199791e-01 -6.98382258e-01 3.71573091e-01
-2.34768558e-02 5.33286154e-01 -4.10340995e-01 1.04398513e+00
1.00354862e+00 2.76074588e-01 -2.94543859e-02 -3.54109913e-01
3.47604632e-01 -1.70076624e-01 -1.31810355e+00 1.05283582e+00
-8.71892497e-02 2.99085140e-01 2.44031385e-01 -6.31066084e-01
9.78719771e-01 2.79781491e-01 2.65713602e-01 -2.70829797e-01
4.36006963e-01 8.92219692e-02 4.01438177e-02 -2.76695251e-01
-5.93540147e-02 1.46412283e-01 1.92832034e-02 3.35676908e-01
-2.33854111e-02 9.66529325e-02 -4.62220520e-01 8.24544653e-02
1.04080117e+00 2.45156005e-01 3.07232976e-01 -8.86214711e-03
8.63357365e-01 -4.35269088e-01 9.40568447e-01 5.34378767e-01
-5.70533991e-01 7.16831923e-01 4.93121684e-01 -4.77155983e-01
-8.34980428e-01 -9.82932568e-01 6.53085709e-02 1.14087498e+00
-5.12280762e-02 -4.32735443e-01 -1.05812728e+00 -1.18136430e+00
1.21573567e-01 2.49782011e-01 -7.62800336e-01 -4.33160633e-01
-7.63080716e-01 -6.50189459e-01 1.17596138e+00 6.00685894e-01
5.66053510e-01 -8.08294594e-01 -1.57475859e-01 -3.21642816e-01
1.24379188e-01 -1.12948072e+00 -9.50816035e-01 -2.48550817e-01
-4.48358059e-01 -1.23445797e+00 -3.23869526e-01 -6.32641912e-01
1.17654324e+00 1.47647783e-01 6.60673618e-01 4.73692328e-01
-3.66124809e-01 2.36786589e-01 -1.47741988e-01 -7.12924719e-01
-6.46998286e-01 -1.78720295e-01 4.09156382e-01 3.63942921e-01
6.00693300e-02 -4.29825306e-01 -5.34842908e-01 7.57122815e-01
-8.46433461e-01 -4.26479071e-01 4.53900933e-01 9.36801791e-01
4.65833515e-01 -1.51224300e-01 4.27604079e-01 -1.02018118e+00
3.06170881e-01 -2.39009202e-01 -9.10742223e-01 2.60650456e-01
-4.89927143e-01 -6.70758635e-02 6.83335781e-01 -9.28794205e-01
-9.28806901e-01 5.60208082e-01 -4.08329487e-01 -6.81233644e-01
7.53776915e-03 -1.03725396e-01 -6.95064425e-01 -7.43154109e-01
8.11259329e-01 -9.57346633e-02 2.53636330e-01 -1.91607520e-01
4.58543211e-01 3.30399781e-01 5.38377643e-01 -6.76207483e-01
1.66124129e+00 5.30697644e-01 1.12826951e-01 -4.39389318e-01
-7.39279926e-01 1.87273063e-02 -5.90104818e-01 -2.23221675e-01
5.91166139e-01 -6.97747529e-01 -8.16535115e-01 6.46302104e-01
-1.01601624e+00 -1.71440050e-01 -8.44375789e-02 -1.22710355e-02
-3.62637907e-01 4.66982841e-01 -3.90760213e-01 -8.86724770e-01
-4.79536891e-01 -1.37858713e+00 9.81133044e-01 3.39879304e-01
-1.20733082e-01 -6.02796018e-01 -1.09164953e-01 3.97053957e-01
3.31544101e-01 6.29384995e-01 4.88353163e-01 -8.87579501e-01
-5.64404488e-01 -5.54077566e-01 5.84838726e-02 5.51179349e-01
2.54649341e-01 3.56191725e-01 -1.24991226e+00 -4.10839349e-01
-3.03693842e-02 -3.74491513e-01 3.90617788e-01 -1.47043556e-01
1.25987875e+00 -8.37245107e-01 -2.03727782e-01 9.98711169e-01
9.19536889e-01 -6.95140939e-03 6.46569133e-01 3.49231452e-01
8.27058375e-01 5.98517239e-01 6.30907655e-01 3.16053331e-01
-1.89597644e-02 7.00264215e-01 7.27047861e-01 -3.36101830e-01
1.41994739e-02 -4.52353716e-01 5.40871561e-01 -9.97279957e-02
-6.34008413e-03 1.18647598e-01 -8.47528338e-01 1.77770540e-01
-1.59991491e+00 -9.86196041e-01 2.76482910e-01 2.15451741e+00
8.56525838e-01 1.37098894e-01 2.70066231e-01 6.70538619e-02
6.72433436e-01 2.48311177e-01 -6.18170321e-01 -1.57200351e-01
-4.27142382e-02 1.59014747e-01 5.18051326e-01 5.63492954e-01
-1.25732481e+00 1.37448978e+00 5.84017181e+00 7.67139673e-01
-1.21816361e+00 2.97689196e-02 5.80045938e-01 -3.33472677e-02
1.00792766e-01 -7.10256514e-04 -9.15924489e-01 4.05797094e-01
5.03004432e-01 -4.66646161e-03 4.53120857e-01 1.15055430e+00
3.43289301e-02 4.67760831e-01 -8.34905744e-01 6.38611913e-01
2.20037907e-01 -1.04636383e+00 1.38307422e-01 2.00567856e-01
4.92095947e-01 -2.74876982e-01 5.15588105e-01 1.37518302e-01
3.52769345e-01 -9.69028652e-01 6.79574192e-01 4.58512306e-02
5.84847391e-01 -7.45248020e-01 6.86110735e-01 9.10825729e-02
-1.16779983e+00 -1.23388328e-01 -2.03554139e-01 4.40273374e-01
-1.22468606e-01 1.44707579e-02 -1.00293684e+00 2.34925821e-01
5.75664043e-01 7.28102773e-02 -7.39230871e-01 6.51405573e-01
-6.43628657e-01 6.51812434e-01 -4.02590752e-01 5.12449443e-01
-6.36893362e-02 9.48328450e-02 6.20251954e-01 1.00372660e+00
7.49390051e-02 5.77687658e-02 1.71281889e-01 7.52789855e-01
-1.91387922e-01 8.59344155e-02 -6.16826534e-01 1.07932359e-01
7.84062028e-01 1.47015882e+00 -3.16064358e-01 1.43988624e-01
-2.99397767e-01 9.40125167e-01 3.15751046e-01 3.14983189e-01
-1.16892576e+00 -1.53341070e-01 1.02620471e+00 2.09859163e-01
9.55500752e-02 -9.61044878e-02 -3.02147299e-01 -8.36630821e-01
1.13165461e-01 -1.32690215e+00 4.80523854e-01 -3.11165482e-01
-1.24190283e+00 8.27113152e-01 -2.92035758e-01 -1.31116629e+00
6.08751364e-03 -4.40001845e-01 -9.63968337e-01 7.86737204e-01
-1.33226228e+00 -1.54443252e+00 -3.24840963e-01 7.88282812e-01
2.17588603e-01 -4.39289004e-01 8.24554145e-01 2.54841566e-01
-1.01258576e+00 1.45354021e+00 -5.71479559e-01 5.40784419e-01
1.08976173e+00 -8.79368782e-01 7.18239963e-01 1.24588549e+00
1.51835769e-01 6.94031477e-01 6.04003370e-01 -5.90260208e-01
-1.19685566e+00 -1.26618075e+00 3.18814486e-01 -7.34707236e-01
7.01610923e-01 -6.35896742e-01 -1.00958657e+00 7.93974221e-01
-1.97148100e-01 6.05008841e-01 8.65522325e-01 -2.67470002e-01
-8.43472660e-01 -2.51454502e-01 -1.52990246e+00 9.42936003e-01
1.07078755e+00 -5.03668606e-01 -2.15550750e-01 2.76919514e-01
5.89134157e-01 -5.91987729e-01 -6.36656284e-01 6.38160408e-01
6.45313442e-01 -7.74708629e-01 1.19880497e+00 -6.36190057e-01
1.24526750e-02 -5.13905644e-01 -4.71608378e-02 -1.05960572e+00
-1.98863432e-01 -1.13456261e+00 -1.64074585e-01 1.56373560e+00
3.57491344e-01 -8.29472721e-01 8.66044760e-01 7.82392204e-01
1.60055116e-01 -8.86810005e-01 -9.11210656e-01 -8.40520740e-01
-3.02045811e-02 -1.46935150e-01 8.37233663e-01 8.96373689e-01
-4.43743974e-01 -2.59988029e-02 -5.69360077e-01 8.83805335e-01
7.00835466e-01 -5.29533982e-01 1.32224810e+00 -8.18010032e-01
-9.77210626e-02 -3.67867380e-01 -4.51719522e-01 -6.18854463e-01
2.64462680e-01 -6.00471318e-01 2.07439754e-02 -5.90780795e-01
-1.21379815e-01 -4.16476220e-01 -1.57888159e-01 9.69841719e-01
-6.35975599e-01 5.61963975e-01 3.40120792e-01 2.50789076e-01
-1.75145447e-01 5.04263341e-01 9.75549936e-01 -7.88729787e-02
-1.93862379e-01 1.70503691e-01 -8.37250710e-01 1.12298501e+00
8.90408874e-01 -6.56428337e-01 -2.14299783e-01 -2.44126707e-01
-1.42074347e-01 -4.64510441e-01 4.56148177e-01 -6.86330378e-01
2.97070503e-01 -2.46165454e-01 4.31301922e-01 -2.69513041e-01
2.65974432e-01 -8.96163583e-01 -9.17537585e-02 4.39417750e-01
-1.59076169e-01 9.64510068e-02 3.56651753e-01 4.17136282e-01
9.03443247e-03 -3.44557315e-03 1.22817743e+00 2.02165425e-01
-3.39124799e-01 5.47409773e-01 7.25441650e-02 -1.16175659e-01
1.28369761e+00 -5.35190888e-02 -4.76167709e-01 -1.48209050e-01
-4.99150246e-01 3.11856598e-01 5.09274840e-01 4.60964411e-01
6.10555291e-01 -1.27107465e+00 -6.81007743e-01 6.44881964e-01
1.49466068e-01 -2.01739848e-01 6.98130578e-02 5.95738947e-01
-3.14690053e-01 -4.00083065e-01 -1.06816821e-01 -3.43337685e-01
-1.81316710e+00 7.00934827e-01 5.94410777e-01 7.14879343e-03
-4.76853758e-01 1.13615799e+00 2.71014720e-01 -5.09236753e-01
5.52163899e-01 2.15143278e-01 5.86800696e-03 -1.04263693e-01
6.95162237e-01 2.34817207e-01 -1.55044258e-01 -7.31789827e-01
-6.08135641e-01 7.61916816e-01 -2.26343587e-01 1.20685659e-01
1.03918850e+00 1.12565411e-02 3.01467329e-02 -3.96404564e-01
9.89291668e-01 5.70067823e-01 -1.47545588e+00 -2.49627799e-01
-4.23544087e-02 -7.11975098e-01 -3.00975055e-01 -6.82064772e-01
-1.29308081e+00 5.41622102e-01 6.58065856e-01 -1.00638591e-01
1.32319140e+00 -2.24446222e-01 7.09508657e-01 2.62926549e-01
1.64416820e-01 -7.62008190e-01 3.31444234e-01 2.61944562e-01
1.13948011e+00 -1.28344059e+00 6.50705397e-02 -7.98625410e-01
-7.39484072e-01 9.90893483e-01 9.67210412e-01 -1.90686986e-01
6.82961941e-01 5.46713769e-01 5.90524316e-01 4.69312221e-02
-2.21912980e-01 -2.22805496e-02 4.28198785e-01 6.33913815e-01
-2.84207482e-02 -3.09272140e-01 1.69956565e-01 5.86539328e-01
-2.70000875e-01 -3.70400906e-01 1.02693960e-01 9.09909189e-01
-6.85373470e-02 -1.34538591e+00 -7.37875044e-01 -1.04988948e-01
-6.91374183e-01 7.70569146e-02 -7.33056247e-01 6.75296307e-01
2.17860177e-01 8.81361246e-01 -5.34397781e-01 -6.32143438e-01
4.48566467e-01 9.79842320e-02 2.95665383e-01 -5.32981336e-01
-6.57315731e-01 1.42050255e-03 -2.05032200e-01 -7.97416031e-01
1.91249717e-02 -4.83526766e-01 -1.14737761e+00 -1.99390218e-01
-4.81303334e-01 -1.68133661e-01 5.12381971e-01 7.01648176e-01
4.63530958e-01 2.72831559e-01 1.08272016e+00 -8.02917480e-01
-1.01860106e+00 -7.90662229e-01 -6.68171421e-02 3.80097717e-01
5.16768277e-01 -6.55275404e-01 -5.99662364e-01 3.71802375e-02] | [5.588517189025879, 7.860280990600586] |
d3c96493-f8a6-4ca2-af33-525d8848f911 | deep-colorization | 1605.00075 | null | http://arxiv.org/abs/1605.00075v1 | http://arxiv.org/pdf/1605.00075v1.pdf | Deep Colorization | This paper investigates into the colorization problem which converts a
grayscale image to a colorful version. This is a very difficult problem and
normally requires manual adjustment to achieve artifact-free quality. For
instance, it normally requires human-labelled color scribbles on the grayscale
target image or a careful selection of colorful reference images (e.g.,
capturing the same scene in the grayscale target image). Unlike the previous
methods, this paper aims at a high-quality fully-automatic colorization method.
With the assumption of a perfect patch matching technique, the use of an
extremely large-scale reference database (that contains sufficient color
images) is the most reliable solution to the colorization problem. However,
patch matching noise will increase with respect to the size of the reference
database in practice. Inspired by the recent success in deep learning
techniques which provide amazing modeling of large-scale data, this paper
re-formulates the colorization problem so that deep learning techniques can be
directly employed. To ensure artifact-free quality, a joint bilateral filtering
based post-processing step is proposed. We further develop an adaptive image
clustering technique to incorporate the global image information. Numerous
experiments demonstrate that our method outperforms the state-of-art algorithms
both in terms of quality and speed. | ['Zezhou Cheng', 'Bin Sheng', 'Qingxiong Yang'] | 2016-04-30 | deep-colorization-1 | http://openaccess.thecvf.com/content_iccv_2015/html/Cheng_Deep_Colorization_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Cheng_Deep_Colorization_ICCV_2015_paper.pdf | iccv-2015-12 | ['patch-matching'] | ['computer-vision'] | [ 2.56597757e-01 -6.81153536e-01 3.39067131e-01 -1.61044285e-01
-7.73461342e-01 -4.74900365e-01 2.22564057e-01 -3.14204425e-01
-5.78287005e-01 5.17116487e-01 -4.72775251e-01 -1.96756810e-01
2.93467846e-02 -9.30349350e-01 -6.46866143e-01 -9.05370414e-01
4.67176706e-01 8.18136334e-02 2.48551294e-01 -2.30940595e-01
3.88221353e-01 6.02398515e-01 -1.47796905e+00 -1.14508212e-01
1.09901321e+00 1.03598607e+00 3.44043732e-01 5.79111218e-01
-3.18044573e-01 2.67637283e-01 -6.57930017e-01 -3.98736447e-01
7.28267670e-01 -7.05704689e-01 -5.76099038e-01 5.82190573e-01
7.41113901e-01 -2.53741115e-01 -5.75315533e-03 1.53773391e+00
4.11389947e-01 3.91081423e-01 3.70252937e-01 -1.11044729e+00
-8.92641723e-01 -1.19176269e-01 -1.08473206e+00 -7.18618929e-02
1.19464202e-02 -3.11894808e-02 6.88354909e-01 -9.65211511e-01
4.45959508e-01 1.03475070e+00 5.03840148e-01 1.89559907e-01
-1.39852560e+00 -7.22530782e-01 -3.05271745e-02 3.26247841e-01
-1.51726747e+00 -1.74396202e-01 1.24245620e+00 -1.38364688e-01
1.60315961e-01 2.06338480e-01 8.49421442e-01 5.42116463e-01
9.48965102e-02 3.74456495e-01 1.49047995e+00 -7.51554012e-01
4.74626988e-01 -5.67708910e-03 -2.37786457e-01 6.05341852e-01
3.88042450e-01 1.18215047e-01 -2.58336544e-01 1.56398118e-01
1.07514000e+00 2.30584696e-01 -3.92337114e-01 -6.49986863e-01
-1.18181074e+00 5.33377290e-01 6.62898540e-01 5.70265472e-01
-3.40018243e-01 9.43069533e-02 -2.93723177e-02 2.31159747e-01
2.58953571e-01 4.21707511e-01 -1.30359843e-01 1.27071992e-01
-1.20974016e+00 -2.02807665e-01 4.71810371e-01 7.74334788e-01
1.22102082e+00 2.34455764e-01 2.26214215e-01 1.02768266e+00
-2.16549989e-02 5.89293361e-01 3.42091978e-01 -1.16389716e+00
1.61607802e-01 6.47270799e-01 4.35860664e-01 -1.38209939e+00
-1.06637709e-01 -3.66737485e-01 -1.35791969e+00 8.57886612e-01
6.81534111e-01 7.44151101e-02 -1.00960124e+00 1.38130426e+00
4.33157951e-01 4.23245355e-02 -1.86912090e-01 1.12099397e+00
2.44472891e-01 7.24128008e-01 -2.97869384e-01 -2.89197743e-01
1.15646160e+00 -9.17849720e-01 -8.13529611e-01 -1.41688347e-01
-1.83242515e-01 -1.04296649e+00 1.33831370e+00 7.89170802e-01
-8.63704264e-01 -8.74428511e-01 -1.32767928e+00 -3.63208130e-02
-5.02713919e-01 8.68095011e-02 4.63627666e-01 8.57905030e-01
-1.18863833e+00 6.32593393e-01 -7.05902338e-01 -4.29179728e-01
1.08344607e-01 9.65129286e-02 -4.40241307e-01 -4.72014874e-01
-9.31872845e-01 7.44473755e-01 2.17068672e-01 2.68423498e-01
-4.07865971e-01 -4.79232192e-01 -4.80208755e-01 -1.08303308e-01
3.71072501e-01 -4.63211715e-01 7.12449133e-01 -1.56636226e+00
-1.68682528e+00 7.98749387e-01 -6.78765401e-02 1.53281868e-01
6.81388617e-01 -1.08340010e-01 -3.67187619e-01 3.93549263e-01
-1.64528370e-01 3.94471794e-01 1.38672817e+00 -1.63036799e+00
-7.30944633e-01 -2.81443834e-01 -1.87804028e-02 1.79734260e-01
-4.92712945e-01 -1.46042138e-01 -1.15160036e+00 -8.99535179e-01
4.03745204e-01 -8.32029819e-01 -3.11268449e-01 2.94797271e-01
-2.75244057e-01 2.94938594e-01 7.36538827e-01 -7.16274083e-01
9.61651802e-01 -2.28095722e+00 -1.70404911e-02 4.85820860e-01
2.01172978e-01 1.85583681e-01 -2.35599980e-01 2.75985181e-01
-2.67122239e-01 -2.11234719e-01 -4.71570700e-01 -2.48134777e-01
-1.59488901e-01 6.21253140e-02 1.06621951e-01 7.99591422e-01
-2.58472059e-02 5.43750763e-01 -7.69931257e-01 -6.71526015e-01
4.21774179e-01 5.96068084e-01 -3.85989487e-01 3.32880944e-01
2.86285400e-01 3.93834621e-01 -1.15652993e-01 5.78385055e-01
1.15582311e+00 -1.66491583e-01 -3.97413336e-02 -6.70450330e-01
-1.57623723e-01 -5.88014066e-01 -1.59405351e+00 1.81686950e+00
-5.82100987e-01 4.85285848e-01 4.57537055e-01 -1.08741653e+00
9.62346494e-01 -6.37038574e-02 6.99425101e-01 -8.08843791e-01
1.67293355e-01 1.61002383e-01 -7.77561143e-02 -4.83752415e-02
4.98539686e-01 -2.35784978e-01 1.51098743e-01 5.57429910e-01
-3.49268049e-01 -5.40239275e-01 1.78666368e-01 3.62143070e-02
6.36169910e-01 1.24842241e-01 1.87317163e-01 -3.23949933e-01
5.84800363e-01 2.17531286e-02 7.69972920e-01 6.26664937e-01
-2.20626861e-01 1.02356672e+00 3.99463065e-02 -3.71481270e-01
-1.11949694e+00 -8.39653254e-01 -1.18571948e-02 9.13187325e-01
5.92932820e-01 5.85019551e-02 -8.70763063e-01 -3.23174268e-01
-2.30732352e-01 3.28577697e-01 -7.19867349e-01 -5.89128360e-02
-6.15551710e-01 -8.19624007e-01 -8.00375491e-02 3.53573442e-01
9.89345729e-01 -7.89557874e-01 -6.22569680e-01 6.11595772e-02
-1.67789876e-01 -8.10200453e-01 -7.19070852e-01 1.77709684e-02
-7.42375195e-01 -1.22987235e+00 -1.10990644e+00 -9.19533670e-01
9.99593973e-01 7.33615518e-01 9.66759920e-01 3.06498706e-01
-3.97661537e-01 4.02284116e-01 -4.89241958e-01 1.13270476e-01
-2.66645163e-01 -4.60482210e-01 -2.33389616e-01 4.73134995e-01
1.46039203e-01 -5.52238107e-01 -9.28787351e-01 3.46501917e-01
-1.20033360e+00 2.08274335e-01 8.87631118e-01 9.30783987e-01
9.04484749e-01 5.73503792e-01 5.12656868e-02 -6.38245344e-01
4.48343098e-01 3.56700212e-01 -8.63314569e-01 5.25496364e-01
-7.37497151e-01 -7.88005814e-02 8.93633544e-01 -3.57436657e-01
-1.16486371e+00 2.62038171e-01 1.33677945e-01 -5.71515977e-01
-1.84494734e-01 1.73227876e-01 -3.06382209e-01 -4.81686771e-01
4.79260832e-01 5.08324146e-01 1.01426952e-01 -5.88949740e-01
6.69611931e-01 4.28805619e-01 8.26385021e-01 -3.35285902e-01
1.33777440e+00 7.34372735e-01 -6.18665852e-02 -7.44191647e-01
-2.65134484e-01 -4.28568363e-01 -9.41308618e-01 -3.50654393e-01
8.24422061e-01 -5.82243383e-01 -4.26068813e-01 7.77079582e-01
-8.83299053e-01 -3.43972802e-01 4.76060472e-02 1.62314996e-01
-4.52185214e-01 7.38111079e-01 -3.79043102e-01 -5.52328169e-01
-2.42627382e-01 -9.99518096e-01 8.89143586e-01 2.52910703e-01
3.47086906e-01 -7.98446774e-01 3.30616720e-02 2.96528209e-02
6.13719881e-01 2.00690404e-01 6.71822429e-01 2.63449937e-01
-6.99238956e-01 -2.94227034e-01 -4.81687218e-01 5.71580231e-01
5.36536694e-01 1.92807063e-01 -7.56878495e-01 -3.60956371e-01
1.17537625e-01 -2.12900899e-02 7.01302648e-01 2.86878079e-01
1.26122642e+00 -3.13403979e-02 -1.11491466e-02 7.81341493e-01
1.91368103e+00 3.79352182e-01 7.03232884e-01 5.74028909e-01
8.26284647e-01 4.93550181e-01 6.29536211e-01 2.85901338e-01
6.88218996e-02 7.56197989e-01 2.68309027e-01 -7.39966989e-01
-4.18374360e-01 6.67155161e-02 -9.32311267e-02 6.50947690e-01
-5.70113286e-02 1.58551514e-01 -5.76473594e-01 3.96860957e-01
-1.51300561e+00 -8.00089121e-01 1.07409284e-01 2.42236233e+00
9.87477660e-01 -2.54205823e-01 -4.48797308e-02 3.80564153e-01
9.30854619e-01 2.21975911e-02 -6.99090540e-01 -8.34741592e-02
-2.50258505e-01 1.40704647e-01 4.99732763e-01 3.35021704e-01
-1.16667306e+00 6.93993151e-01 5.78356981e+00 9.70976472e-01
-1.33487260e+00 -1.05556533e-01 8.04763496e-01 2.97133774e-01
-1.21781707e-01 -2.38666266e-01 -9.17940959e-02 4.99529868e-01
2.73276806e-01 -7.68958852e-02 8.01640332e-01 6.67332113e-01
2.71050394e-01 -4.39914972e-01 -7.51626194e-01 1.45537412e+00
2.40649089e-01 -9.06075239e-01 -9.31184739e-02 -2.00699821e-01
9.01566088e-01 -4.53737587e-01 1.62950903e-01 -2.27047309e-01
2.89095938e-01 -6.19144738e-01 7.55763888e-01 5.03823519e-01
1.02983928e+00 -7.61034191e-01 4.60710198e-01 -1.08841874e-01
-1.24879265e+00 5.38613088e-02 -5.35633981e-01 3.45312059e-01
3.65013666e-02 7.52528608e-01 -1.27972793e-02 7.21570551e-01
1.00384676e+00 6.56075954e-01 -8.49670291e-01 1.28263259e+00
-1.26865879e-01 1.96561545e-01 -3.30736637e-01 3.08643579e-01
4.41177040e-02 -8.32055390e-01 3.25752771e-03 9.61351454e-01
3.41356665e-01 2.17180252e-01 1.20777801e-01 8.52668703e-01
3.40524390e-02 1.33856371e-01 -1.67728141e-01 1.87270954e-01
2.74995476e-01 1.63863337e+00 -1.07949758e+00 -3.11265737e-01
-6.45264626e-01 1.43756568e+00 4.45811413e-02 6.91711664e-01
-6.45201206e-01 -7.49849319e-01 4.13589478e-01 -6.98183700e-02
4.65960592e-01 -3.13383847e-01 -4.27215248e-01 -1.09487081e+00
-1.16232105e-01 -9.54307139e-01 2.33225971e-01 -9.19852853e-01
-1.33903706e+00 6.87829435e-01 -3.07626307e-01 -1.67470050e+00
4.68493886e-02 -5.66077352e-01 -7.68543899e-01 9.18998003e-01
-1.65785933e+00 -1.03215504e+00 -6.75922871e-01 9.86202896e-01
2.80416429e-01 1.55246571e-01 5.41523457e-01 4.81631309e-01
-6.03584349e-01 5.22105336e-01 5.41593909e-01 1.63022548e-01
9.58686769e-01 -1.32167768e+00 1.02796070e-01 1.40962744e+00
7.67349005e-02 5.71256518e-01 7.31971145e-01 -3.12239528e-01
-1.59400666e+00 -1.02776110e+00 1.68351501e-01 2.01596506e-02
3.93457353e-01 -1.33918226e-01 -9.41844583e-01 -2.04460919e-02
4.26700652e-01 2.24030912e-02 3.64784062e-01 -3.26773077e-01
-2.57423490e-01 -6.63997531e-01 -9.66580749e-01 5.68940401e-01
6.43879354e-01 -4.09690857e-01 -4.95207906e-01 1.95177179e-02
2.59823143e-01 -1.91402093e-01 -5.65921485e-01 8.46996978e-02
4.72891062e-01 -1.21555948e+00 1.01688135e+00 2.91886926e-02
1.63503557e-01 -7.79028237e-01 -3.98612392e-05 -1.30024946e+00
-3.89932036e-01 -4.84204441e-01 5.18630445e-01 1.17902827e+00
2.77659297e-02 -3.92671466e-01 7.47341812e-01 8.40516150e-01
1.45270258e-01 -2.23586366e-01 -8.04924905e-01 -8.14865232e-01
2.91239750e-02 -1.11372404e-01 4.45020735e-01 9.22395408e-01
-4.21606809e-01 -1.90256596e-01 -3.84661913e-01 1.58601478e-01
9.69816923e-01 7.69328058e-01 9.05756652e-01 -9.78437722e-01
-2.32832178e-01 -5.93113482e-01 -1.26008928e-01 -7.53672957e-01
-2.31810063e-01 -2.04022974e-01 3.72262776e-01 -1.56864285e+00
1.59828886e-01 -3.87403429e-01 -3.97666782e-01 2.61290073e-01
-4.44684893e-01 8.22118938e-01 1.99326932e-01 2.42970556e-01
-5.47988117e-01 6.14524245e-01 1.43762267e+00 -2.54871547e-01
-2.53435403e-01 -1.16846681e-01 -6.94377899e-01 5.53890049e-01
7.35363722e-01 -8.75248984e-02 -2.25257143e-01 -3.95445555e-01
1.24761218e-03 -2.15553507e-01 1.76059201e-01 -1.08878493e+00
3.11589599e-01 -2.38943115e-01 6.50094986e-01 -4.63194609e-01
3.37162137e-01 -1.14780891e+00 2.39596277e-01 3.54412436e-01
-8.56154691e-03 1.38029724e-01 4.17154059e-02 6.47351563e-01
-4.84582961e-01 8.00807308e-03 1.20721948e+00 -2.87587464e-01
-9.49263573e-01 2.32341170e-01 -2.19033852e-01 -3.27997923e-01
9.85633194e-01 -5.69390118e-01 -5.16595356e-02 -5.34653425e-01
-3.23230237e-01 -2.09698603e-01 9.21649814e-01 1.87722459e-01
6.58835888e-01 -1.37648177e+00 -3.77762139e-01 3.48220706e-01
2.75255572e-02 -1.27736226e-01 3.85102212e-01 7.98081934e-01
-8.30824316e-01 -6.84553152e-03 -4.60118651e-01 -3.78689498e-01
-1.02447498e+00 8.83508623e-01 3.39625776e-01 1.60854295e-01
-7.38591611e-01 6.84099853e-01 1.79855794e-01 2.92601269e-02
2.41635770e-01 -2.00919420e-01 -4.83901128e-02 -6.82959706e-02
6.50550961e-01 2.48486027e-01 1.54852360e-01 -5.62302172e-01
-1.97571769e-01 1.08131158e+00 2.22586021e-01 -2.62348652e-01
1.23100340e+00 -4.65675622e-01 -4.76691276e-01 1.38394758e-01
1.27439570e+00 9.35310945e-02 -1.44283617e+00 -3.23518485e-01
-2.10656479e-01 -9.25265253e-01 2.79886037e-01 -6.27924025e-01
-1.66351724e+00 9.23119426e-01 8.78674150e-01 2.32776091e-01
1.73938012e+00 -4.70968902e-01 6.85838044e-01 2.42589951e-01
3.74109179e-01 -1.34671557e+00 2.24531904e-01 -5.33278398e-02
8.69781375e-01 -1.36837196e+00 1.50615185e-01 -2.61770844e-01
-4.80793267e-01 1.24073911e+00 6.26385629e-01 -2.92940825e-01
5.87119699e-01 -3.81290503e-02 4.48183984e-01 8.88582915e-02
-4.47482169e-02 -2.01022893e-01 4.21864450e-01 6.17431641e-01
1.04884654e-01 -2.64603615e-01 -2.80283689e-01 1.72603324e-01
1.33314217e-02 -2.15875685e-01 5.22637725e-01 6.82163477e-01
-4.40101862e-01 -1.17904842e+00 -1.02187324e+00 9.96681675e-02
-2.28278756e-01 -7.18513578e-02 -2.71001577e-01 6.78109944e-01
1.27371456e-02 1.09645450e+00 5.08952402e-02 -2.49967709e-01
2.63156712e-01 -1.78326577e-01 3.92461896e-01 -1.42703980e-01
-1.75411075e-01 4.82975423e-01 -6.07020378e-01 -8.16257179e-01
-8.12432647e-01 -2.78214097e-01 -8.55457604e-01 -3.57399672e-01
-2.62494922e-01 4.22155112e-02 6.36983514e-01 5.18271685e-01
4.97947074e-02 3.37492079e-01 8.37079823e-01 -1.10230291e+00
-9.99981984e-02 -6.31569564e-01 -1.00731456e+00 8.30049872e-01
2.44044110e-01 -5.19948244e-01 -3.99634957e-01 3.44299763e-01] | [11.010205268859863, -1.614920973777771] |
d5b55c9a-2cf3-48b3-ac85-16e7ee446409 | clinical-text-summarization-with-syntax-based | 2003.00353 | null | https://arxiv.org/abs/2003.00353v1 | https://arxiv.org/pdf/2003.00353v1.pdf | Clinical Text Summarization with Syntax-Based Negation and Semantic Concept Identification | In the era of clinical information explosion, a good strategy for clinical text summarization is helpful to improve the clinical workflow. The ideal summarization strategy can preserve important information in the informative but less organized, ill-structured clinical narrative texts. Instead of using pure statistical learning approaches, which are difficult to interpret and explain, we utilized knowledge of computational linguistics with human experts-curated biomedical knowledge base to achieve the interpretable and meaningful clinical text summarization. Our research objective is to use the biomedical ontology with semantic information, and take the advantage from the language hierarchical structure, the constituency tree, in order to identify the correct clinical concepts and the corresponding negation information, which is critical for summarizing clinical concepts from narrative text. We achieved the clinically acceptable performance for both negation detection and concept identification, and the clinical concepts with common negated patterns can be identified and negated by the proposed method. | ['Yu-An Chung', 'Wei-Hung Weng', 'Schrasing Tong'] | 2020-02-29 | null | null | null | null | ['negation-detection'] | ['natural-language-processing'] | [ 5.03445566e-01 6.67989016e-01 -3.24171394e-01 -8.15195441e-02
-5.49597442e-01 -1.86498582e-01 -9.34037473e-03 1.19582403e+00
-3.52600604e-01 1.21294439e+00 9.56759930e-01 -1.68115929e-01
-5.56512535e-01 -4.84648883e-01 2.99612656e-02 -6.39660895e-01
-1.90684095e-01 4.85348225e-01 1.97105892e-02 -2.65021682e-01
4.52331722e-01 1.84117451e-01 -1.34238148e+00 7.61557698e-01
1.56337643e+00 4.90748495e-01 1.58335999e-01 4.30263668e-01
-6.06806576e-01 1.22181690e+00 -6.00761652e-01 -3.03701609e-01
-5.14428437e-01 -9.75320995e-01 -1.04328156e+00 -1.85138389e-01
-5.02862692e-01 6.30013570e-02 3.13361526e-01 1.33189225e+00
5.31834364e-01 -2.36323863e-01 6.48842096e-01 -7.33315825e-01
-3.33523840e-01 1.01752055e+00 -2.40635157e-01 2.51378268e-01
5.53503990e-01 -2.91980863e-01 8.91318917e-01 -4.11956102e-01
9.46790993e-01 1.19984901e+00 4.60363537e-01 6.63341582e-01
-5.93654335e-01 -3.78454149e-01 5.66361062e-02 2.50935405e-01
-1.13916659e+00 -2.89543271e-01 6.00185871e-01 -2.50403285e-01
9.86726105e-01 4.93378550e-01 9.13504720e-01 7.05200136e-01
7.13479400e-01 6.54006541e-01 6.21267915e-01 -6.05844259e-01
2.58512884e-01 1.21160984e-01 5.36848724e-01 6.97236836e-01
7.37573564e-01 -5.12961745e-01 -4.05848324e-01 -4.33556706e-01
-1.05739214e-01 1.22660495e-01 -5.63923538e-01 4.11211252e-01
-1.08322513e+00 6.82032943e-01 -4.40887138e-02 7.53379941e-01
-5.94413280e-01 -4.85216439e-01 9.51396048e-01 -1.78415671e-01
3.04076284e-01 6.12957060e-01 -4.56469148e-01 -1.70140132e-01
-8.56695414e-01 -6.62023574e-02 9.34818029e-01 7.71469712e-01
5.80521785e-02 -2.02464968e-01 -1.76026046e-01 4.30359751e-01
2.37802267e-01 4.04386073e-01 9.99183178e-01 -7.47629404e-01
3.71216029e-01 1.19941688e+00 -1.55708164e-01 -1.13866317e+00
-6.35093033e-01 -4.13963437e-01 -1.03126907e+00 -6.17238402e-01
-3.07444334e-01 -1.63564101e-01 -7.98044026e-01 1.38131726e+00
2.29719073e-01 -3.25673640e-01 8.95895183e-01 5.69314182e-01
1.31712580e+00 4.45922941e-01 4.80991900e-01 -1.11027551e+00
1.88880789e+00 -4.72284704e-01 -1.67567730e+00 -1.08026832e-01
1.11702192e+00 -8.28010321e-01 4.39320982e-01 4.10282254e-01
-8.78370225e-01 1.75950438e-01 -1.28103709e+00 -1.35296568e-01
-3.16146880e-01 4.24226932e-02 5.54391801e-01 3.41893256e-01
-5.99576831e-01 4.29372877e-01 -8.04007053e-01 -6.84810519e-01
6.58728838e-01 1.51495159e-01 -5.55251539e-01 1.23467430e-01
-1.46583009e+00 1.13082516e+00 1.30265319e+00 2.66766816e-01
-9.42406878e-02 -6.65582418e-01 -8.01588595e-01 2.26960599e-01
5.59741437e-01 -1.00294471e+00 9.94174302e-01 -6.69667482e-01
-9.06853199e-01 7.60788560e-01 -4.70787644e-01 -4.83902484e-01
1.67086020e-01 -3.28342356e-02 -5.26896715e-01 7.60196805e-01
1.09206304e-01 2.71472156e-01 6.46396652e-02 -9.70576763e-01
-7.38965333e-01 -4.68267083e-01 -5.43229699e-01 4.77845162e-01
-5.06906331e-01 -1.58471223e-02 3.18895690e-02 -4.06062543e-01
3.28503340e-01 -2.73235798e-01 -4.92912918e-01 -5.13999045e-01
-6.46809161e-01 -2.17315719e-01 5.59156597e-01 -9.99179959e-01
1.71399236e+00 -1.91373754e+00 -1.23889752e-01 1.54521957e-01
6.43023789e-01 3.68454993e-01 6.31761849e-01 6.83886886e-01
-8.69490281e-02 6.20362878e-01 -3.70945752e-01 2.84586787e-01
-4.62993085e-01 4.10208553e-01 -6.24433644e-02 7.21282959e-02
1.50690287e-01 7.12527156e-01 -1.07811403e+00 -1.40774131e+00
-4.31102663e-02 2.95263156e-02 -5.93901634e-01 4.68061753e-02
-2.20928937e-01 2.95229763e-01 -1.07295167e+00 6.99007750e-01
4.29592669e-01 -2.25976363e-01 6.26402974e-01 -4.47267145e-01
1.69454858e-01 3.00693333e-01 -9.05073524e-01 1.35313761e+00
2.84494013e-01 6.30722865e-02 -1.17417365e-01 -1.03867161e+00
8.14992368e-01 6.81480348e-01 7.83552289e-01 -4.83839840e-01
3.52713287e-01 2.91273266e-01 -1.19319439e-01 -1.21794569e+00
3.38152617e-01 -5.86669266e-01 7.52256578e-03 -7.03061670e-02
-2.46652916e-01 1.14957519e-01 3.27039957e-01 5.41422248e-01
9.38660443e-01 -4.13122088e-01 1.33610427e+00 -4.58337277e-01
8.39660645e-01 6.40345931e-01 9.19723868e-01 3.78654599e-01
-2.62212865e-02 3.10423911e-01 8.87606800e-01 -3.46892089e-01
-5.57197094e-01 -6.12904191e-01 -3.80111665e-01 1.54468283e-01
1.02305233e-01 -7.86159754e-01 -7.31550872e-01 -4.44258124e-01
-4.45947349e-01 9.13395524e-01 -3.33747208e-01 -4.14539695e-01
-5.20624518e-01 -1.11611152e+00 6.09183013e-01 2.74860412e-01
3.53693545e-01 -1.06205583e+00 -6.57016575e-01 5.63249826e-01
-5.73258102e-01 -9.70422685e-01 -1.79210469e-01 2.96938121e-01
-1.13859951e+00 -1.46937370e+00 -2.33525708e-01 -8.07162344e-01
7.95887113e-01 -3.80726039e-01 6.49707973e-01 1.94087192e-01
-3.78294528e-01 4.55052108e-02 -6.49232268e-01 -7.52239406e-01
-1.00091374e+00 3.31323408e-02 -1.54603839e-01 -4.75330889e-01
5.50863206e-01 -3.87506247e-01 -5.08279979e-01 -5.46507835e-01
-1.18576801e+00 3.02377731e-01 7.06350386e-01 9.91736233e-01
5.11435509e-01 1.04088917e-01 9.14370418e-01 -1.22898018e+00
1.10769641e+00 -4.54887211e-01 2.24913180e-01 4.50305790e-01
-9.55577135e-01 2.75155306e-01 7.21116483e-01 5.62136136e-02
-1.21152973e+00 -1.99676007e-01 -3.39265823e-01 3.96217316e-01
3.33478339e-02 1.04197717e+00 -1.63620234e-01 6.55625939e-01
7.77502656e-01 3.92144412e-01 2.87995905e-01 -1.97627738e-01
2.95110464e-01 9.11181509e-01 4.01870608e-01 -2.23046333e-01
-4.52584624e-02 3.64984363e-01 2.06784979e-01 -6.98405266e-01
-8.34738791e-01 -5.86360633e-01 -5.20490408e-01 3.19921635e-02
1.22256839e+00 -6.31343842e-01 -8.61474931e-01 -2.50012755e-01
-1.39434266e+00 8.31332505e-01 -4.37089384e-01 6.85381949e-01
-2.38397285e-01 9.60620463e-01 -5.85770726e-01 -8.42070162e-01
-1.01145899e+00 -7.76028693e-01 7.99281597e-01 4.16472018e-01
-5.17559528e-01 -1.00640798e+00 -3.91374156e-02 1.35839298e-01
-1.88008338e-01 5.26659906e-01 1.69698703e+00 -1.44405770e+00
-2.62874998e-02 -3.79675835e-01 8.27209502e-02 1.72270119e-01
4.49557930e-01 -9.67572406e-02 -5.04493356e-01 1.96184188e-01
4.13955986e-01 -6.81236852e-03 7.46783793e-01 3.93259585e-01
8.17130804e-01 -7.86651433e-01 -7.68905699e-01 1.17519386e-01
1.37965286e+00 6.53283715e-01 6.42787158e-01 1.11122899e-01
4.34531391e-01 7.85915673e-01 6.17826045e-01 4.50486332e-01
2.40815282e-01 -1.08496398e-02 -1.25678228e-02 6.28272519e-02
1.65861964e-01 -2.93868393e-01 -1.39642939e-01 1.38649511e+00
-8.23766924e-03 -2.96397269e-01 -9.77983892e-01 5.98379970e-01
-2.04432917e+00 -8.89190674e-01 -1.27049580e-01 1.56279159e+00
1.44850016e+00 2.05511361e-01 -4.56189722e-01 1.57845929e-01
6.57193124e-01 -3.60087484e-01 -1.51940003e-01 -4.53980416e-01
-2.72633255e-01 -1.48600386e-02 2.25218549e-01 3.32642883e-01
-6.21184766e-01 7.80491233e-01 6.37097931e+00 8.71345043e-01
-7.53875911e-01 -7.20321909e-02 2.77389795e-01 7.42108524e-02
-4.26281303e-01 4.40788008e-02 -7.01730728e-01 3.15936595e-01
8.44383180e-01 -6.57751560e-01 -3.62678558e-01 7.07941532e-01
8.00838768e-01 -4.26312655e-01 -8.94746840e-01 6.67581081e-01
1.43969700e-01 -1.68669188e+00 5.69151461e-01 -1.26290172e-01
5.34809589e-01 -7.56200969e-01 -6.77751780e-01 -1.11395061e-01
3.35284360e-02 -8.13646197e-01 1.90098614e-01 9.66248214e-01
4.14405435e-01 -5.24983764e-01 1.52479935e+00 4.88153726e-01
-9.19980288e-01 -4.26528830e-04 -3.29015046e-01 1.05106361e-01
1.96068540e-01 7.51211822e-01 -1.26499915e+00 1.23160040e+00
3.64636391e-01 9.69550252e-01 -4.25346375e-01 9.66734827e-01
-1.47421300e-01 5.92309475e-01 -5.85349835e-02 -3.15963566e-01
-2.15974245e-02 -2.23865643e-01 6.47541821e-01 1.22528553e+00
2.11872011e-01 7.82210529e-01 2.35648528e-01 4.57751215e-01
6.16332255e-02 9.07905638e-01 -5.47657609e-01 -4.72392261e-01
4.96349990e-01 9.13483083e-01 -1.01305783e+00 -6.58109546e-01
-8.22917968e-02 5.19381285e-01 -2.09944531e-01 -9.31145158e-03
-4.46229547e-01 -4.28471774e-01 5.05467039e-03 1.05520502e-01
-2.20377743e-01 5.62219381e-01 -5.50986528e-01 -1.06711519e+00
1.07842237e-01 -8.82100344e-01 9.33949232e-01 -6.02232039e-01
-1.08378077e+00 6.18476093e-01 2.06026047e-01 -1.27553892e+00
-3.43202740e-01 -3.00385177e-01 -5.25363386e-01 6.05261624e-01
-1.22574675e+00 -1.01471293e+00 -2.42564883e-02 3.21643412e-01
2.98434317e-01 -1.43001720e-01 1.08536518e+00 1.41443074e-01
-5.58141589e-01 1.48629948e-01 3.58109400e-02 -9.84692499e-02
6.62748873e-01 -1.18759418e+00 -7.84655035e-01 5.73605120e-01
-7.37971306e-01 8.70284975e-01 9.50234294e-01 -1.11709726e+00
-8.84244025e-01 -8.00583661e-01 1.49467635e+00 1.36275187e-01
4.38236803e-01 3.50181103e-01 -1.08123052e+00 4.13670838e-01
3.09925050e-01 -7.09717691e-01 1.06958187e+00 -1.82935625e-01
2.99358755e-01 -5.56765832e-02 -1.20563054e+00 7.42319822e-01
8.74450028e-01 -1.26999065e-01 -1.48460615e+00 6.21622026e-01
1.01105511e+00 -2.27190703e-01 -9.32328403e-01 5.01126826e-01
1.88185796e-01 -3.23595524e-01 7.20777810e-01 -9.80194926e-01
5.75554430e-01 -2.81705022e-01 2.17591479e-01 -9.62201774e-01
9.26460177e-02 -4.13852662e-01 2.52315164e-01 1.16601753e+00
6.06130421e-01 -6.30394876e-01 5.41546524e-01 3.56631011e-01
-4.85524952e-01 -8.36258709e-01 -9.23911154e-01 -2.86588501e-02
-1.27235979e-01 2.29660012e-02 2.36941680e-01 9.58919406e-01
1.08352423e+00 4.65494901e-01 -1.18435182e-01 -4.53232713e-02
2.36782268e-01 -3.72931734e-02 -4.52595390e-02 -1.36691093e+00
4.01198030e-01 -3.98416489e-01 -5.45164824e-01 -2.45511010e-01
-5.54088019e-02 -1.12510633e+00 7.88514502e-03 -2.17435861e+00
6.51080489e-01 1.88082740e-01 -6.14725128e-02 6.25946462e-01
-4.20533389e-01 -4.75050867e-01 -2.60464340e-01 4.60419476e-01
-6.70219958e-01 5.66556096e-01 1.30677068e+00 -2.19498366e-01
-3.91505420e-01 -3.16266418e-01 -1.22609258e+00 8.11345041e-01
5.80251098e-01 -9.24126923e-01 -5.45238197e-01 2.50573367e-01
4.44686204e-01 3.76756549e-01 -1.85860142e-01 -6.86879873e-01
6.08917117e-01 -3.47597301e-01 2.09662497e-01 -8.26035976e-01
-3.43351901e-01 -8.57528865e-01 7.21282437e-02 1.09718657e+00
-5.65548837e-01 -2.66456585e-02 3.17289799e-01 4.85477418e-01
-5.97517669e-01 -6.26395941e-01 5.16239345e-01 -3.83881420e-01
-5.59153974e-01 -2.78503835e-01 -6.20178640e-01 3.54875863e-01
1.10540426e+00 -3.26587319e-01 -4.87347007e-01 -6.62670657e-02
-1.04329503e+00 4.89096195e-01 -3.82330567e-02 -2.02824287e-02
9.11462545e-01 -1.05471992e+00 -8.46814394e-01 -2.97954917e-01
1.94759771e-01 5.25901504e-02 5.31401455e-01 1.00607729e+00
-9.97658312e-01 7.06036627e-01 -1.83038652e-01 -4.29802924e-01
-1.40271580e+00 5.76064646e-01 1.69143394e-01 -7.10302770e-01
-8.21436644e-01 9.54005197e-02 -6.42654002e-02 1.07507735e-01
9.64819863e-02 -5.00633597e-01 -9.85068560e-01 4.58041668e-01
7.00497150e-01 9.67110973e-03 3.52731824e-01 -3.81351501e-01
-7.01026618e-01 3.40866983e-01 -3.17220330e-01 1.11418761e-01
1.29309011e+00 -2.50056416e-01 -7.41859734e-01 2.53954470e-01
9.14182365e-01 2.45032117e-01 -4.55726217e-03 -5.46870828e-02
4.76312816e-01 2.00118080e-01 -1.40776951e-02 -9.36578333e-01
-4.43534881e-01 5.25575340e-01 1.66226715e-01 2.50612646e-01
1.29273164e+00 -1.70182083e-02 6.25996828e-01 5.83019912e-01
-8.90105665e-02 -1.10544372e+00 -1.81974441e-01 1.34380311e-01
7.25606084e-01 -8.71152222e-01 4.48565096e-01 -8.17292809e-01
-9.76926982e-01 1.42302430e+00 1.79077953e-01 5.54237127e-01
5.57855070e-01 2.19191283e-01 5.03117107e-02 -5.95651209e-01
-8.56174946e-01 -6.72905967e-02 1.27583072e-01 4.18721110e-01
3.77291501e-01 -2.49322038e-02 -1.23083663e+00 1.28503525e+00
-1.51714668e-01 3.04517776e-01 7.43077457e-01 1.10268009e+00
-7.19053209e-01 -9.04918492e-01 -4.12114263e-01 6.91913784e-01
-8.46658647e-01 -4.59121197e-01 -3.91468763e-01 7.68080890e-01
2.77653962e-01 9.89919066e-01 -4.05536413e-01 -4.10539247e-02
5.27366042e-01 4.42391813e-01 -2.70212591e-02 -7.94297874e-01
-6.33807063e-01 2.01169610e-01 3.77200663e-01 -2.28436157e-01
-8.16010356e-01 -4.91536230e-01 -1.96962464e+00 2.57900953e-02
-3.61285686e-01 8.00800502e-01 2.60788083e-01 1.23631072e+00
3.46230119e-01 8.97845030e-01 3.46742645e-02 3.90507996e-01
-4.93824959e-01 -9.44844842e-01 -3.39507431e-01 4.00003195e-01
2.15358391e-01 -2.63472825e-01 -2.67317891e-01 3.57784092e-01] | [8.442699432373047, 8.61593246459961] |
44d9ddc0-0b22-4a0a-a598-d4c6354bdbe8 | bayesian-optimisation-for-constrained | 2105.13245 | null | https://arxiv.org/abs/2105.13245v1 | https://arxiv.org/pdf/2105.13245v1.pdf | Bayesian Optimisation for Constrained Problems | Many real-world optimisation problems such as hyperparameter tuning in machine learning or simulation-based optimisation can be formulated as expensive-to-evaluate black-box functions. A popular approach to tackle such problems is Bayesian optimisation (BO), which builds a response surface model based on the data collected so far, and uses the mean and uncertainty predicted by the model to decide what information to collect next. In this paper, we propose a novel variant of the well-known Knowledge Gradient acquisition function that allows it to handle constraints. We empirically compare the new algorithm with four other state-of-the-art constrained Bayesian optimisation algorithms and demonstrate its superior performance. We also prove theoretical convergence in the infinite budget limit. | ['Juergen Branke', 'Juan Ungredda'] | 2021-05-27 | null | null | null | null | ['bayesian-optimisation'] | ['methodology'] | [ 3.06216896e-01 9.50715970e-03 -2.59319127e-01 -4.03178573e-01
-7.98156261e-01 -3.71603578e-01 5.75426638e-01 2.11928174e-01
-9.65504467e-01 1.16470611e+00 -1.39506429e-01 -4.22522753e-01
-8.67084861e-01 -5.13009667e-01 -7.43419290e-01 -1.00429165e+00
1.75841209e-02 9.26597118e-01 3.53744239e-01 -1.30658329e-01
6.20315611e-01 2.01026067e-01 -1.32363749e+00 -3.99340451e-01
1.01147056e+00 1.10839212e+00 1.52460665e-01 6.47155464e-01
2.83751905e-01 -4.34060022e-02 -5.18391192e-01 -1.94277540e-01
2.12958530e-01 -3.62760276e-01 -4.79429990e-01 -2.25846380e-01
-5.16650200e-01 2.09617302e-01 3.05067182e-01 8.56903613e-01
8.81321490e-01 6.47622466e-01 6.92849398e-01 -7.17436492e-01
1.50350392e-01 3.28449667e-01 -3.16680044e-01 1.42456457e-01
2.96632886e-01 4.07351315e-01 6.29507422e-01 -3.90287697e-01
2.50789016e-01 1.37306702e+00 4.38816100e-01 4.40779358e-01
-1.54932296e+00 -4.49650645e-01 1.49364382e-01 3.46721292e-01
-1.30799854e+00 -4.33284879e-01 4.60364342e-01 -3.01611125e-01
8.75791192e-01 4.61349905e-01 5.82979858e-01 7.59389699e-01
2.56556273e-01 5.18999755e-01 1.31741583e+00 -7.27636993e-01
8.74480367e-01 8.75893235e-02 -2.24087611e-01 3.57767493e-01
2.02005014e-01 7.34985471e-01 -5.20011127e-01 -4.90328550e-01
3.61022949e-01 -4.91795093e-01 -5.18128633e-01 -6.00516558e-01
-9.41345215e-01 1.24237680e+00 1.64016932e-01 -3.42752427e-01
-5.48790395e-01 2.31128484e-01 8.03440064e-02 7.51106292e-02
5.27963579e-01 6.47250772e-01 -6.68303192e-01 -2.85507232e-01
-7.03559816e-01 3.58709812e-01 1.05061495e+00 3.76318246e-01
3.84544164e-01 -2.17894539e-01 -2.37167150e-01 7.87758112e-01
1.09437525e+00 3.30213428e-01 4.52667922e-01 -6.64932609e-01
1.72735631e-01 2.04582661e-01 5.18002391e-01 -4.55479771e-01
-6.23078763e-01 -5.10233819e-01 -5.40737927e-01 2.07207426e-01
4.52705234e-01 -3.38454306e-01 -8.74549568e-01 1.33469248e+00
8.97080839e-01 1.32853940e-01 -6.01687394e-02 1.17100716e+00
2.41430223e-01 6.60552025e-01 -1.15822837e-01 -5.82653046e-01
9.49502826e-01 -5.65011799e-01 -6.28306627e-01 -2.01264039e-01
2.14653850e-01 -6.44788742e-01 6.26876056e-01 8.72268558e-01
-1.12226665e+00 -9.37806144e-02 -1.17943227e+00 6.30665720e-01
-8.31903815e-02 -4.27291155e-01 6.24740303e-01 1.00584018e+00
-5.97874343e-01 8.71467471e-01 -8.31188679e-01 4.30740565e-02
2.05914184e-01 8.26280177e-01 2.98708826e-01 7.72739649e-02
-1.20505571e+00 1.20172656e+00 8.99744630e-01 3.31277430e-01
-9.78473663e-01 -6.11926019e-01 -6.59124970e-01 -7.75251091e-02
1.06827402e+00 -8.72214496e-01 1.46380615e+00 -6.12999737e-01
-2.43796706e+00 2.67544329e-01 -2.86438223e-02 -4.49719787e-01
5.16454458e-01 -1.52904868e-01 -9.92465541e-02 -3.24493080e-01
-5.99514008e-01 2.14061260e-01 1.06253052e+00 -8.49796474e-01
-5.07554293e-01 -2.77854770e-01 -1.43364623e-01 2.21692383e-01
1.03169747e-01 2.96905249e-01 -4.07109410e-01 -4.91446167e-01
-5.76718487e-02 -9.20896649e-01 -6.71486616e-01 -4.37594384e-01
-3.65874529e-01 -2.47070849e-01 -9.62038562e-02 -3.22344005e-01
1.51384532e+00 -1.62214506e+00 5.19910991e-01 7.38753319e-01
-3.15733582e-01 1.57621190e-01 7.77199939e-02 3.44417185e-01
1.41512975e-01 8.87971371e-02 -5.25843978e-01 -5.78018613e-02
3.38998258e-01 3.77838731e-01 1.22509003e-02 7.75053978e-01
-1.26035914e-01 7.01332331e-01 -9.16894674e-01 -2.99335837e-01
2.03074768e-01 2.49806970e-01 -5.39419234e-01 2.31540367e-01
-5.64872682e-01 6.56627715e-01 -5.35269916e-01 3.54343981e-01
3.43001306e-01 -1.08852852e-02 3.47339272e-01 2.12225929e-01
-8.16145970e-04 1.94459721e-01 -1.50569081e+00 1.47519433e+00
-4.21357602e-01 1.53040573e-01 7.96118677e-02 -1.29304564e+00
7.24470556e-01 1.41654983e-01 2.08864361e-01 -5.20070970e-01
3.51670027e-01 2.03249827e-01 1.50407314e-01 -2.42926747e-01
1.21815287e-01 -2.63172835e-01 -8.28204229e-02 2.58212537e-01
-6.44387975e-02 -5.09019852e-01 2.69153118e-01 -5.34245074e-01
8.50543976e-01 4.44088370e-01 6.96145236e-01 -3.60220373e-01
6.40102804e-01 -1.86586648e-01 7.08365679e-01 9.94815648e-01
2.30363160e-01 2.42068231e-01 3.77291143e-01 -2.41436824e-01
-7.30671465e-01 -6.96372509e-01 -4.92234677e-01 1.13746679e+00
-1.13091059e-01 -1.88832477e-01 -6.71046853e-01 -4.11997348e-01
1.80630282e-01 8.33442330e-01 -4.31515872e-01 -9.40639153e-02
-3.19777548e-01 -1.47501481e+00 -1.68673351e-01 1.80970747e-02
8.58650357e-02 -7.38997340e-01 -7.89211750e-01 5.71781635e-01
2.47088343e-01 -4.88215506e-01 -1.21364944e-01 3.96581411e-01
-9.23677027e-01 -9.02524292e-01 -6.06620729e-01 4.78267670e-02
3.78394008e-01 -6.06233060e-01 1.10477006e+00 -1.57584384e-01
-1.90650672e-01 1.30256742e-01 -2.50123322e-01 -9.51500952e-01
-3.94990176e-01 -8.82152021e-02 1.68127745e-01 9.24547911e-02
1.80959716e-01 -1.11064069e-01 -5.33640265e-01 8.25866103e-01
-8.58294904e-01 -4.01171178e-01 6.62727535e-01 1.17462337e+00
8.78166378e-01 2.74057984e-01 4.22489673e-01 -6.74148262e-01
8.77151668e-01 -4.83108997e-01 -1.47905612e+00 4.44647133e-01
-1.02446294e+00 6.75713181e-01 2.00743005e-01 -6.63830340e-01
-1.03896368e+00 1.30871266e-01 4.92250994e-02 -1.50756016e-01
-1.88822895e-02 8.54522467e-01 -1.72950938e-01 -4.34333146e-01
6.10409915e-01 -1.89262688e-01 -1.40766963e-01 -6.29290640e-01
2.43657470e-01 8.01799774e-01 3.05925697e-01 -9.09659743e-01
5.30572832e-01 1.54228257e-02 3.55357319e-01 -5.18678606e-01
-9.73403811e-01 -5.49474716e-01 -2.40176544e-01 -3.39038759e-01
5.59825659e-01 -3.12633127e-01 -1.15641153e+00 2.29303882e-01
-8.17934036e-01 -5.98214388e-01 -2.49901831e-01 7.60550559e-01
-8.70468974e-01 2.54433811e-01 2.55055189e-01 -1.38722622e+00
-2.04841241e-01 -1.08714736e+00 8.26430619e-01 3.53704959e-01
-1.57788232e-01 -1.12564254e+00 4.68364149e-01 1.06849939e-01
3.71316493e-01 3.68138015e-01 5.33537865e-01 -5.66398859e-01
-3.71688694e-01 -2.52129227e-01 3.55426669e-01 1.46827132e-01
-4.59528625e-01 -1.51954904e-01 -6.34455860e-01 -5.28657317e-01
4.38447863e-01 -1.37079045e-01 9.52618182e-01 8.14275622e-01
1.17267239e+00 -2.95017213e-01 -3.89753342e-01 7.85770178e-01
1.34784043e+00 3.15983951e-01 5.23179531e-01 6.46730244e-01
-4.11081277e-02 5.64278781e-01 9.54997897e-01 7.77626574e-01
-8.17368999e-02 1.05506265e+00 4.97945517e-01 4.19378430e-01
7.05423772e-01 1.58767685e-01 2.27836311e-01 5.01546502e-01
-8.06027651e-02 -3.80163848e-01 -9.11909759e-01 1.81104556e-01
-2.32193875e+00 -4.99017626e-01 1.00734040e-01 2.71061635e+00
1.17593181e+00 4.30727810e-01 6.62492067e-02 -1.36220377e-04
6.53522253e-01 -3.75207901e-01 -8.23433936e-01 -6.36856914e-01
2.08597049e-01 6.43746674e-01 9.29093361e-01 6.79105878e-01
-1.07916105e+00 5.10578871e-01 7.44983006e+00 1.21551883e+00
-5.48539400e-01 1.68651149e-01 3.99407566e-01 -4.19499069e-01
3.20393704e-02 2.10290372e-01 -8.22676539e-01 6.52462602e-01
1.41457224e+00 -1.78887531e-01 8.72750878e-01 4.69259739e-01
4.17904019e-01 -8.55723441e-01 -1.08573735e+00 8.45124960e-01
-1.88516811e-01 -1.00220656e+00 -5.39452732e-01 2.24064544e-01
8.08905780e-01 -4.90924902e-02 -1.09600857e-01 2.08021432e-01
7.31691182e-01 -1.19864571e+00 4.50100571e-01 1.05269468e+00
3.57491076e-01 -9.89769876e-01 9.50055957e-01 6.31628692e-01
-4.54666346e-01 -4.46808815e-01 -4.95536387e-01 -1.40420362e-01
3.71948034e-01 9.50374484e-01 -8.73810709e-01 7.34977961e-01
5.74749172e-01 9.56487581e-02 -2.69689709e-01 1.80433154e+00
-5.49147785e-01 8.42549622e-01 -9.52836156e-01 -5.56141555e-01
2.65790761e-01 -4.83820230e-01 7.82854140e-01 8.99431646e-01
2.21486449e-01 2.91013960e-02 6.53602034e-02 8.10397446e-01
4.31499094e-01 2.45898329e-02 6.36720732e-02 1.27964839e-01
3.05578977e-01 8.69877398e-01 -4.96614784e-01 -3.56112085e-02
8.09233040e-02 3.09597313e-01 -1.17247850e-02 3.01358670e-01
-6.79243922e-01 -3.10651124e-01 3.96429330e-01 -3.32138658e-01
4.34568822e-01 1.33064007e-02 -1.58254579e-01 -7.29848862e-01
-1.31275207e-01 -7.85545111e-01 6.53779149e-01 -5.83012223e-01
-1.15877545e+00 -6.49026930e-02 5.86779714e-01 -7.71691620e-01
-6.52631879e-01 -7.91771173e-01 -3.18437636e-01 1.07185245e+00
-1.46113539e+00 -2.82490611e-01 2.00508788e-01 2.43850008e-01
2.56270647e-01 -1.20571457e-01 6.82306647e-01 -2.68012077e-01
-7.04358160e-01 3.06507409e-01 8.55154097e-01 -5.93924940e-01
5.27247906e-01 -1.36097276e+00 5.55003136e-02 5.11888921e-01
-2.04040930e-01 3.28241110e-01 1.11595893e+00 -5.11006713e-01
-1.44485998e+00 -3.91289055e-01 1.90869644e-01 -4.13600296e-01
6.56727254e-01 -1.45296738e-01 -6.82600200e-01 6.22083098e-02
-1.36060521e-01 -2.47726545e-01 5.31022072e-01 2.53528714e-01
3.94759655e-01 -9.30030495e-02 -1.18954873e+00 3.71629477e-01
4.98555869e-01 1.73370376e-01 -7.12781489e-01 4.17376369e-01
3.61368239e-01 -7.07893312e-01 -1.06505406e+00 6.57702148e-01
4.66240257e-01 -6.54675305e-01 9.29657400e-01 -6.49840772e-01
-2.29425043e-01 -2.06638843e-01 5.89427799e-02 -1.79247451e+00
-8.39855969e-02 -1.31325126e+00 -3.69530618e-01 8.03211749e-01
4.14209515e-01 -1.07831645e+00 5.97750604e-01 7.17755377e-01
1.09935299e-01 -9.78136301e-01 -1.46720004e+00 -1.05260956e+00
-1.97013225e-02 -4.65667129e-01 6.88746274e-01 3.83217931e-01
-5.57531323e-03 1.51774749e-01 -2.21899867e-01 -1.53126316e-02
9.33382034e-01 -1.28584772e-01 7.05839157e-01 -1.45681548e+00
-8.43541443e-01 -6.48890793e-01 -2.88207263e-01 -9.50265229e-01
5.32085225e-02 -4.76450354e-01 4.03776735e-01 -1.11654425e+00
-5.12794293e-02 -4.37323719e-01 -4.22978491e-01 1.94158465e-01
-4.22673672e-01 -2.21828550e-01 -1.97893009e-01 -1.08082592e-01
-5.99083960e-01 6.17208600e-01 9.63329196e-01 2.14773626e-03
-4.31732893e-01 4.75568593e-01 -3.42813581e-01 6.86158299e-01
7.99584627e-01 -8.44697118e-01 -1.35289684e-01 1.04148544e-01
6.98970258e-01 8.13051239e-02 1.43727556e-01 -6.50384784e-01
2.69486964e-01 -5.52916884e-01 2.11320236e-01 -4.41155761e-01
2.12832555e-01 -6.61178887e-01 3.48580509e-01 4.28039163e-01
-1.86043426e-01 -3.23802829e-01 2.30445489e-01 8.40435565e-01
5.65949492e-02 -9.37393963e-01 7.81119466e-01 -3.85662951e-02
-3.18331838e-01 -1.56309247e-01 -2.91618079e-01 7.43093267e-02
1.00316226e+00 5.19816130e-02 -6.28620237e-02 -3.61466318e-01
-7.01189458e-01 6.51404202e-01 1.81904748e-01 3.93915735e-02
4.94199425e-01 -9.19632256e-01 -8.08200955e-01 -1.31646544e-01
-1.23748049e-01 2.05963701e-01 -1.96198523e-01 9.87975657e-01
-2.15704292e-01 6.48311198e-01 3.50805998e-01 -7.19996393e-01
-9.10912991e-01 7.00946689e-01 5.93277276e-01 -5.48488736e-01
-7.05337301e-02 7.93076754e-01 -2.47193232e-01 -4.22968119e-01
2.84506619e-01 -5.59572130e-02 5.91057204e-02 -9.21033788e-03
3.21814597e-01 6.47554040e-01 2.98508942e-01 -4.86668572e-02
-3.81885141e-01 5.40861964e-01 4.12190020e-01 -6.70962036e-01
1.32599306e+00 -2.11836286e-02 -9.34924558e-02 4.60312515e-01
6.54269218e-01 -5.28447807e-01 -1.35295510e+00 -3.05317819e-01
2.94295043e-01 -5.53218842e-01 6.28893018e-01 -1.12322927e+00
-4.93781775e-01 4.30675894e-01 6.55604780e-01 1.06777936e-01
1.10350287e+00 -3.47435832e-01 1.36641115e-01 6.82084322e-01
4.30321783e-01 -1.59989154e+00 -4.24430698e-01 2.59141177e-01
8.91474247e-01 -1.22268164e+00 5.68109453e-01 -5.77160493e-02
-3.54098260e-01 9.22058105e-01 7.67658204e-02 6.62656948e-02
7.76569188e-01 -6.81947693e-02 -3.72349590e-01 -1.93291903e-02
-9.11893845e-01 -3.97935361e-01 6.96258903e-01 5.05612791e-01
-2.17684969e-01 -4.28673252e-03 -6.13332689e-01 3.56334448e-01
-6.40495569e-02 7.78578147e-02 1.75421521e-01 9.94112253e-01
-4.56329226e-01 -1.33793819e+00 -8.72829139e-01 4.70671982e-01
-5.16118705e-01 8.68329257e-02 -1.93583488e-01 4.99160439e-01
-4.05434281e-01 1.21337998e+00 -3.02478969e-01 1.21236481e-01
2.52382159e-01 1.45967200e-01 6.33720875e-01 -3.98987710e-01
-6.43257678e-01 4.62595224e-01 2.80948639e-01 -6.52922630e-01
-5.32382548e-01 -6.97325885e-01 -7.84198761e-01 -4.59263893e-03
-8.98197293e-01 3.93960327e-01 1.29823339e+00 1.16852450e+00
1.09824508e-01 5.87544918e-01 5.80132544e-01 -1.20657408e+00
-1.20064747e+00 -1.01962137e+00 -4.01521474e-01 -9.41494033e-02
1.99232146e-01 -1.11709893e+00 -5.87523043e-01 -5.04634202e-01] | [6.210909843444824, 3.7609386444091797] |
348e8c43-06ee-4110-b3f2-2b74f740e9e1 | perfect-match-improved-cross-modal-embeddings | 1809.08001 | null | http://arxiv.org/abs/1809.08001v2 | http://arxiv.org/pdf/1809.08001v2.pdf | Perfect match: Improved cross-modal embeddings for audio-visual synchronisation | This paper proposes a new strategy for learning powerful cross-modal
embeddings for audio-to-video synchronization. Here, we set up the problem as
one of cross-modal retrieval, where the objective is to find the most relevant
audio segment given a short video clip. The method builds on the recent
advances in learning representations from cross-modal self-supervision.
The main contributions of this paper are as follows: (1) we propose a new
learning strategy where the embeddings are learnt via a multi-way matching
problem, as opposed to a binary classification (matching or non-matching)
problem as proposed by recent papers; (2) we demonstrate that performance of
this method far exceeds the existing baselines on the synchronization task; (3)
we use the learnt embeddings for visual speech recognition in self-supervision,
and show that the performance matches the representations learnt end-to-end in
a fully-supervised manner. | ['Hong-Goo Kang', 'Soo-Whan Chung', 'Joon Son Chung'] | 2018-09-21 | null | null | null | null | ['video-synchronization'] | ['computer-vision'] | [ 1.34849623e-01 -7.94471055e-02 -3.95402849e-01 -2.81871557e-01
-1.62540460e+00 -5.53126037e-01 6.06414735e-01 7.60918036e-02
-3.95332038e-01 9.14886817e-02 7.13281751e-01 3.98891091e-01
-3.16029973e-02 -2.06081849e-02 -7.96032071e-01 -6.57018185e-01
-4.71379340e-01 2.56774426e-01 3.10804751e-02 -5.59939630e-02
8.52273330e-02 4.60528247e-02 -1.71769023e+00 2.47735992e-01
1.64498165e-01 1.08728969e+00 3.75486314e-02 8.63003790e-01
4.92463470e-01 7.44048119e-01 -3.16486299e-01 -1.87401384e-01
9.15985107e-02 -4.20183718e-01 -1.03238213e+00 1.17181279e-01
1.05397403e+00 -2.11409345e-01 -7.04177141e-01 7.09711313e-01
7.93112874e-01 2.16117159e-01 6.71497464e-01 -1.43010557e+00
-7.58948743e-01 4.97462690e-01 -4.94565368e-01 5.21099806e-01
6.27563536e-01 -4.07093138e-01 1.51197433e+00 -1.16550016e+00
5.64510763e-01 1.23397040e+00 6.35231078e-01 5.21647811e-01
-1.31268942e+00 -7.92287350e-01 1.27381802e-01 5.00193655e-01
-1.67279065e+00 -9.93466794e-01 7.88331568e-01 -3.94135505e-01
1.04819489e+00 -1.13312460e-01 5.10439873e-01 1.27054489e+00
-1.69051751e-01 1.08447719e+00 5.85948586e-01 -5.97509921e-01
-1.31661281e-01 -9.46164802e-02 -1.23142652e-01 5.78744590e-01
-5.44474959e-01 1.78338543e-01 -1.20890462e+00 -1.76489145e-01
4.38074529e-01 -2.13536307e-01 -4.57048178e-01 -5.64689517e-01
-1.53698814e+00 8.22707653e-01 2.34698206e-01 8.71029317e-01
3.74109228e-03 3.12415630e-01 8.39507580e-01 6.87922835e-01
6.59971178e-01 2.59212106e-01 -3.41683894e-01 -3.21745366e-01
-1.33803797e+00 1.43505841e-01 6.14808142e-01 8.97812605e-01
6.61722302e-01 1.56992525e-01 -9.33118984e-02 9.01434779e-01
3.11086833e-01 3.72000962e-01 8.66935372e-01 -9.64125454e-01
5.57029366e-01 -2.44614989e-01 -1.35563120e-01 -1.03529882e+00
-1.36910975e-01 -1.27398625e-01 -5.21782219e-01 -2.19408840e-01
-6.24963641e-02 1.76611766e-02 -3.78262103e-01 2.16035104e+00
9.36790183e-02 8.13767314e-01 1.81843519e-01 8.34235370e-01
9.18589354e-01 7.02072084e-01 -2.24568680e-01 -3.61396164e-01
1.15006673e+00 -1.36915433e+00 -9.58979011e-01 8.35787579e-02
3.95856410e-01 -9.46687937e-01 6.73351407e-01 1.68785468e-01
-1.28463316e+00 -9.02280927e-01 -1.13212037e+00 -1.44911900e-01
-1.43358603e-01 1.10481814e-01 3.66852552e-01 5.85708655e-02
-1.56684685e+00 5.08771300e-01 -7.34304428e-01 -5.37929773e-01
-4.36035469e-02 1.26128197e-01 -7.08659232e-01 1.10553026e-01
-1.30657864e+00 6.92719400e-01 2.12763011e-01 -2.99500495e-01
-1.20144784e+00 -7.27000654e-01 -1.26784229e+00 9.25171077e-02
2.68591732e-01 -6.18723989e-01 1.48819089e+00 -1.20386302e+00
-1.34671652e+00 1.28702497e+00 -3.57232600e-01 -5.20368338e-01
8.26922208e-02 -3.07688773e-01 -5.66177189e-01 7.49158084e-01
4.63265330e-01 8.82422447e-01 1.24980104e+00 -1.05430579e+00
-6.04645908e-01 -1.79192528e-01 -2.75175609e-02 2.85451710e-01
-6.16266310e-01 2.86467761e-01 -8.91621947e-01 -1.12096608e+00
-6.06070422e-02 -1.03784084e+00 3.46406937e-01 -8.08320101e-03
-1.68704733e-01 -4.92985278e-01 8.40447903e-01 -4.94367391e-01
1.29326499e+00 -2.55395222e+00 5.58741212e-01 -2.60480583e-01
1.20440990e-01 2.76542734e-03 -3.98406982e-01 8.22025239e-01
-6.17222190e-01 -1.19058162e-01 -1.25304554e-02 -8.45002711e-01
1.75848916e-01 -5.19273169e-02 -7.22378790e-01 5.80468714e-01
2.81456411e-01 6.04603827e-01 -9.15545523e-01 -7.95290112e-01
3.26586701e-02 6.15514517e-01 -4.00243312e-01 6.10744894e-01
3.51690352e-01 1.36158884e-01 -7.98049942e-02 6.39542043e-01
2.48195216e-01 -1.49721682e-01 1.26591727e-01 -4.63848174e-01
-1.05548287e-02 3.78083497e-01 -1.11228228e+00 2.35317683e+00
-6.11982524e-01 1.00434709e+00 2.61601746e-01 -1.33951259e+00
5.46785712e-01 9.99285221e-01 7.24608541e-01 -5.63559651e-01
-2.17908040e-01 -4.08046395e-02 -6.06617868e-01 -6.55524969e-01
3.35326344e-01 -1.84285715e-01 -2.36134186e-01 6.14504457e-01
7.25978971e-01 -1.55722529e-01 1.51221529e-01 4.27999943e-01
9.56071973e-01 2.26681650e-01 1.66431248e-01 -5.73753146e-03
4.10640359e-01 -3.20541531e-01 4.30593610e-01 4.37594980e-01
-2.60441571e-01 8.49205136e-01 3.47689718e-01 -4.68343347e-02
-8.63753676e-01 -1.14344895e+00 -1.43809363e-01 1.34272778e+00
-1.40081972e-01 -9.39966202e-01 -5.51166534e-01 -5.12224436e-01
4.57630157e-02 -1.03916250e-01 -7.10356951e-01 -5.91893643e-02
-4.68769580e-01 3.21000181e-02 7.35745251e-01 7.83839941e-01
8.14962536e-02 -7.78824806e-01 -2.91278780e-01 4.70566303e-02
-5.33388495e-01 -1.37830973e+00 -1.08004856e+00 1.57567456e-01
-7.92813540e-01 -9.46665525e-01 -8.93970847e-01 -1.30234349e+00
1.96000516e-01 7.40656316e-01 1.23812079e+00 -1.78098992e-01
-3.12809825e-01 1.26817501e+00 -4.93948966e-01 1.31437927e-01
4.29053195e-02 -9.21190530e-03 5.04545510e-01 1.95574507e-01
1.61349162e-01 -8.45234752e-01 -4.08418298e-01 2.07971603e-01
-1.00774705e+00 -4.93793249e-01 3.91733617e-01 9.57741499e-01
8.11562777e-01 -3.27869385e-01 7.88964927e-01 -3.10042441e-01
4.80009794e-01 -4.32598054e-01 -2.52882957e-01 1.26852155e-01
-3.55839789e-01 -5.03734462e-02 3.69241267e-01 -5.28486192e-01
-3.99145007e-01 1.94072187e-01 3.32071893e-02 -1.23479116e+00
-8.86770114e-02 4.65647966e-01 1.33278742e-01 -5.18058836e-02
2.57843673e-01 3.10077280e-01 7.56203085e-02 -5.10958016e-01
5.35424113e-01 7.14493394e-01 7.98076093e-01 -4.02655035e-01
9.72043097e-01 4.73708779e-01 -4.04083371e-01 -9.85154092e-01
-8.01790297e-01 -9.59224105e-01 -7.56332815e-01 -1.65891886e-01
1.07350171e+00 -1.64235783e+00 -6.18250012e-01 1.84564758e-02
-1.07344687e+00 -1.18527129e-01 -9.58053023e-02 6.80698574e-01
-9.71137166e-01 7.58519113e-01 -6.66701555e-01 -6.43904030e-01
-1.69423133e-01 -9.26020980e-01 1.65180588e+00 -4.37229574e-02
-3.52166027e-01 -1.07087433e+00 6.31998301e-01 4.90220189e-01
1.58456907e-01 1.23810567e-01 5.17375588e-01 -6.65190279e-01
-4.85551506e-01 -1.64414525e-01 9.53056961e-02 3.19157630e-01
-1.39581626e-02 3.08174808e-02 -1.37705445e+00 -6.88977420e-01
-1.37465581e-01 -1.02606320e+00 1.02375448e+00 2.11511925e-01
9.84795094e-01 2.53712274e-02 -3.05118769e-01 5.74811816e-01
1.21167934e+00 -2.95512080e-01 2.36715958e-01 2.36642048e-01
3.73555750e-01 4.92707074e-01 6.19384289e-01 3.66905898e-01
6.02163374e-01 1.22763491e+00 3.19375575e-01 -1.56596079e-01
-2.02982932e-01 -3.46591532e-01 7.71484196e-01 1.29325974e+00
4.76490200e-01 7.08634779e-02 -5.22574425e-01 1.09068191e+00
-2.12953496e+00 -1.32610869e+00 5.02059460e-01 2.10694766e+00
8.20476830e-01 -3.08526963e-01 4.34414297e-01 2.71691501e-01
5.90979457e-01 5.96352458e-01 -1.03521511e-01 -1.48389891e-01
-1.43601730e-01 3.30631346e-01 -1.70152739e-01 5.23975372e-01
-1.56594968e+00 7.77779341e-01 6.73829317e+00 6.64678514e-01
-1.12585390e+00 4.05990750e-01 -9.98259615e-03 -2.45196074e-01
4.30863462e-02 -1.68664813e-01 -5.51544130e-01 1.30733818e-01
1.04282820e+00 -7.89300203e-02 2.94201434e-01 5.85218787e-01
4.14247066e-03 2.57478565e-01 -1.63294685e+00 1.47686803e+00
7.74883926e-01 -1.30357778e+00 -2.61856318e-01 -2.64539301e-01
5.53966463e-01 1.86954990e-01 2.56238222e-01 4.63388205e-01
-1.04548879e-01 -8.33245933e-01 7.79703140e-01 2.86659658e-01
8.93365443e-01 -5.04700422e-01 3.20854932e-01 -6.09505251e-02
-1.63654673e+00 -2.89961621e-02 -2.26749834e-02 2.84000158e-01
2.35015258e-01 1.73912704e-01 -5.61024964e-01 7.06312597e-01
1.06107855e+00 1.30351889e+00 -4.64864075e-01 1.05074191e+00
-4.45733592e-02 5.88655591e-01 -5.45430966e-02 5.61471403e-01
2.58663595e-01 1.87031612e-01 6.06343567e-01 1.55228531e+00
9.63752493e-02 -5.71336091e-01 3.51482451e-01 3.96340758e-01
-2.78515279e-01 1.02753937e-01 -8.84664357e-01 -2.21076190e-01
4.36377496e-01 1.07706237e+00 -9.13835019e-02 -2.66345114e-01
-6.64460063e-01 1.20023692e+00 3.76444519e-01 4.21651125e-01
-7.42183566e-01 -6.18282080e-01 8.19549561e-01 -2.45967031e-01
8.22348058e-01 -1.43733472e-01 4.72465068e-01 -1.32060087e+00
2.99705863e-01 -9.05428648e-01 6.87997103e-01 -8.98424745e-01
-1.53065121e+00 5.88603735e-01 6.69478029e-02 -1.69237101e+00
-5.81971645e-01 -2.28370100e-01 -5.65188587e-01 4.93954509e-01
-1.88445330e+00 -1.11783910e+00 6.93938732e-02 9.38149035e-01
7.53500521e-01 -2.60196537e-01 1.14301217e+00 7.61101305e-01
-3.05017561e-01 9.00232077e-01 2.29785144e-01 4.28747386e-01
1.35588419e+00 -1.16495240e+00 -3.92564340e-03 6.87048912e-01
8.11911285e-01 5.02683997e-01 4.90175784e-01 -1.66692138e-02
-1.44088435e+00 -8.83723617e-01 1.34122753e+00 -4.13131148e-01
1.08557582e+00 -5.16843736e-01 -6.59745574e-01 8.13184142e-01
7.78612733e-01 4.11303520e-01 1.09232378e+00 2.10878581e-01
-8.64279509e-01 -4.29575294e-01 -6.32548094e-01 1.37695357e-01
7.88735032e-01 -1.38540924e+00 -8.48933756e-01 5.37167788e-01
9.71120238e-01 -3.63768458e-01 -1.01529801e+00 7.13802204e-02
6.04082584e-01 -6.93416178e-01 1.06607115e+00 -8.65449309e-01
2.94864416e-01 -3.42908412e-01 -5.97989500e-01 -1.21054447e+00
-1.82533473e-01 -1.06336057e+00 -1.59648553e-01 1.38345158e+00
2.78589576e-01 -7.91983902e-02 2.88277358e-01 -2.92823255e-01
-2.72776365e-01 -5.50928831e-01 -1.46208024e+00 -8.33085358e-01
-1.06639042e-01 -4.34879333e-01 5.51746078e-02 1.19964814e+00
4.16449696e-01 6.17386341e-01 -6.24278247e-01 2.53866255e-01
5.38705945e-01 5.09416878e-01 7.24892497e-01 -1.06716621e+00
-5.35935044e-01 -2.63144195e-01 -7.02574313e-01 -1.26051295e+00
6.74657226e-01 -9.90600348e-01 6.46745488e-02 -1.08008683e+00
3.01620394e-01 1.00246876e-01 -5.66169918e-01 2.81566083e-01
2.46086821e-01 3.65487367e-01 4.51916099e-01 3.48222584e-01
-1.19539702e+00 6.96692467e-01 6.34101748e-01 -4.14811194e-01
1.45961270e-01 -1.72959208e-01 -5.50641239e-01 2.82266319e-01
1.68373927e-01 -4.93736476e-01 -3.04676533e-01 -6.34628832e-01
2.37296857e-02 5.03970563e-01 3.31411183e-01 -9.57904160e-01
4.31314915e-01 5.11608541e-01 -3.19667719e-02 -5.88798404e-01
6.88000202e-01 -8.80514383e-01 -3.83513004e-01 1.95083003e-02
-6.57449424e-01 4.17716295e-01 1.25587031e-01 9.88324523e-01
-9.42280054e-01 2.03182548e-02 5.54723322e-01 2.96250790e-01
-5.99395931e-01 3.28004599e-01 -3.32323670e-01 3.92025054e-01
9.33116257e-01 5.91276251e-02 4.34911512e-02 -8.22174013e-01
-9.98572707e-01 3.53755921e-01 2.24585220e-01 8.76971304e-01
6.28061891e-01 -1.85405779e+00 -6.52858853e-01 -4.50833887e-02
5.03583610e-01 -4.76425171e-01 2.09858701e-01 1.03836000e+00
1.76910788e-01 5.56806326e-01 1.34636462e-01 -9.04571056e-01
-1.50214159e+00 6.41495466e-01 2.16446966e-01 -5.97433411e-02
-4.25607890e-01 1.05819130e+00 4.55959439e-02 -1.75262898e-01
8.80862236e-01 -5.73659576e-02 -1.95323333e-01 5.23670912e-01
5.94534993e-01 -8.56105424e-03 1.23987086e-01 -1.02556098e+00
-4.88896906e-01 9.67536330e-01 -4.04197127e-02 -2.95699537e-01
1.49870002e+00 -2.04453826e-01 1.22211173e-01 8.64990950e-01
1.82127595e+00 -1.47068286e-02 -1.01242101e+00 -7.04801679e-01
-1.48966208e-01 -4.35635686e-01 2.03186244e-01 -1.84149638e-01
-1.18216622e+00 1.09370792e+00 9.24353719e-01 1.64314583e-01
1.06425941e+00 3.58390331e-01 8.24770689e-01 2.92148888e-01
3.43358107e-02 -1.27461517e+00 6.39662385e-01 3.18663746e-01
9.47263479e-01 -1.30231607e+00 -2.62014091e-01 1.57373846e-01
-7.72345424e-01 1.26910758e+00 2.64851332e-01 -2.57164747e-01
7.96115875e-01 1.65065199e-01 1.23735674e-01 -1.86718807e-01
-1.07485998e+00 -4.90916967e-01 3.40114504e-01 7.55756915e-01
6.84108436e-01 -3.48630905e-01 3.15267771e-01 3.21091443e-01
1.70077056e-01 -1.76115468e-01 1.67137742e-01 8.31376851e-01
-7.98009336e-02 -1.00639820e+00 -2.98284918e-01 -2.38888741e-01
-4.84383166e-01 -1.15792178e-01 -4.63385016e-01 6.96550488e-01
-1.18692167e-01 1.16304910e+00 2.15527669e-01 -5.15578091e-01
3.75779659e-01 2.68736333e-01 4.66742694e-01 -6.17801666e-01
-5.34135461e-01 4.38669324e-01 -6.69409707e-02 -8.17513347e-01
-1.00988209e+00 -7.85674095e-01 -9.10590649e-01 2.28174463e-01
-2.77828038e-01 3.00188154e-01 3.18158239e-01 8.24508905e-01
5.11976779e-01 3.68105143e-01 1.00999737e+00 -1.30370629e+00
-6.16907716e-01 -8.10778022e-01 -4.74835485e-01 5.28290391e-01
9.54106867e-01 -8.31194878e-01 -5.18715024e-01 2.89574772e-01] | [14.661617279052734, 4.891398906707764] |
abf1c3da-be71-494f-b17e-1ace42ea83f1 | a-robust-framework-for-moving-object | 1402.0289 | null | http://arxiv.org/abs/1402.0289v1 | http://arxiv.org/pdf/1402.0289v1.pdf | A Robust Framework for Moving-Object Detection and Vehicular Traffic Density Estimation | Intelligent machines require basic information such as moving-object
detection from videos in order to deduce higher-level semantic information. In
this paper, we propose a methodology that uses a texture measure to detect
moving objects in video. The methodology is computationally inexpensive,
requires minimal parameter fine-tuning and also is resilient to noise,
illumination changes, dynamic background and low frame rate. Experimental
results show that performance of the proposed approach is higher than those of
state-of-the-art approaches. We also present a framework for vehicular traffic
density estimation using the foreground object detection technique and present
a comparison between the foreground object detection-based framework and the
classical density state modelling-based framework for vehicular traffic density
estimation. | ['Pranam Janney', 'Glenn Geers'] | 2014-02-03 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [ 1.07867479e-01 -4.02482688e-01 -1.99788436e-01 -2.78286964e-01
-3.50991607e-01 -6.43389598e-02 8.02731693e-01 4.61407239e-03
-6.09538853e-01 9.30795312e-01 -3.45566958e-01 -4.34289813e-01
-1.15972996e-01 -1.08902574e+00 -5.31540394e-01 -8.16266775e-01
-1.20120972e-01 6.61806941e-01 1.23976219e+00 2.14818478e-01
3.09731990e-01 7.18705952e-01 -1.97533441e+00 1.58593565e-01
6.51778936e-01 1.04404199e+00 5.80785990e-01 1.08437979e+00
-4.83580530e-01 1.22562754e+00 -7.32064664e-01 -1.99790046e-01
-1.89256951e-01 -3.15691680e-01 -6.57186449e-01 4.67874825e-01
3.76702964e-01 -4.85241830e-01 -4.41566765e-01 9.93076384e-01
2.27563865e-02 2.74435699e-01 8.48429143e-01 -1.38157380e+00
-2.03993674e-02 -6.17716908e-02 -5.26472390e-01 9.40876424e-01
1.65670946e-01 -7.17915371e-02 -2.21648961e-02 -6.84627354e-01
5.61204612e-01 1.39721334e+00 4.82016802e-01 4.45798278e-01
-9.42617297e-01 -3.01067203e-01 1.53391987e-01 8.91422868e-01
-1.59408653e+00 -4.36828017e-01 4.19886291e-01 -5.56204498e-01
6.45798326e-01 4.49011266e-01 6.93815649e-01 6.13416910e-01
6.26980960e-02 8.42284441e-01 1.02974701e+00 -4.05531049e-01
4.99159098e-01 2.86830544e-01 1.31018296e-01 9.91739035e-01
6.99679375e-01 -6.01492748e-02 -1.50878012e-01 -4.95424047e-02
8.79288733e-01 -1.50519460e-01 7.36777410e-02 -4.63936985e-01
-9.84840333e-01 6.16123319e-01 -1.77111626e-01 4.41332132e-01
-4.84929413e-01 4.80124503e-01 2.87832767e-01 -3.03303134e-02
6.05723560e-01 -7.39701033e-01 -1.50587887e-01 -4.79878575e-01
-1.38585961e+00 2.28502497e-01 9.48564053e-01 9.92796957e-01
8.11862946e-01 3.31829339e-01 -1.44053459e-01 4.52304751e-01
6.89380407e-01 7.67203391e-01 6.59914762e-02 -1.02192545e+00
1.92455634e-01 2.23939210e-01 2.00916976e-01 -1.03640103e+00
-3.55087444e-02 2.48024151e-01 -4.64923829e-01 4.35344309e-01
5.15513957e-01 9.71477479e-02 -9.54258919e-01 1.13094914e+00
6.19979978e-01 8.60734046e-01 5.43019027e-02 7.41988480e-01
9.16005373e-01 8.54918122e-01 2.29171827e-01 -4.44563389e-01
1.23642385e+00 -6.58535898e-01 -9.08586025e-01 1.23617187e-01
3.48183960e-01 -7.09531665e-01 3.12857658e-01 3.90487343e-01
-9.09320056e-01 -5.91710627e-01 -9.20873344e-01 3.43595207e-01
-4.01640892e-01 -5.57282828e-02 6.11084282e-01 1.41270745e+00
-1.03607619e+00 3.20572257e-01 -8.99712920e-01 -6.59195542e-01
4.33425635e-01 4.19611007e-01 -1.31278053e-01 1.28877610e-02
-7.66318202e-01 1.11399055e+00 5.65847933e-01 -8.16259831e-02
-1.09978628e+00 -3.29804778e-01 -8.23079944e-01 -1.81109577e-01
3.95268679e-01 -6.58135056e-01 1.16092134e+00 -1.00334537e+00
-1.67329657e+00 5.75971425e-01 -6.08522356e-01 -5.92503309e-01
8.01872969e-01 -6.60693226e-03 -5.83247960e-01 5.73903739e-01
-1.74735025e-01 3.82529616e-01 9.03514266e-01 -1.24250805e+00
-1.01164436e+00 5.62827401e-02 -5.47313504e-02 -1.33327186e-01
2.40895212e-01 1.57376587e-01 -6.42624855e-01 -1.93126678e-01
-1.76751494e-01 -5.73981404e-01 -5.47472797e-02 3.67635079e-02
3.08241457e-01 -1.07455134e-01 1.72341204e+00 -4.81684268e-01
1.17924762e+00 -1.76585484e+00 -4.12660986e-01 2.80012280e-01
-7.37485215e-02 6.03604376e-01 3.08658153e-01 4.07028124e-02
5.35837591e-01 -2.87717253e-01 -5.61122894e-02 -5.52141331e-02
-2.06363067e-01 4.41590488e-01 2.96471030e-01 7.48546898e-01
1.74644008e-01 4.85805750e-01 -1.01456165e+00 -1.07405686e+00
1.14876938e+00 8.94986987e-01 -2.58117616e-01 -2.21843779e-01
-6.18867464e-02 2.40445182e-01 -3.89944762e-01 9.19852614e-01
9.82963204e-01 2.46745437e-01 2.54987180e-02 -1.98012441e-01
-2.51936466e-01 -1.79405317e-01 -1.54535723e+00 8.23902428e-01
-3.64476681e-01 1.00716794e+00 2.60535538e-01 -1.20526373e+00
8.87829423e-01 3.56574386e-01 3.62413675e-01 -2.28559747e-01
5.50203562e-01 1.83459938e-01 -2.04026729e-01 -7.16874421e-01
7.32606232e-01 -3.05869848e-01 3.24090183e-01 -1.82903647e-01
2.21250445e-01 1.39692381e-01 6.52286708e-01 1.22597307e-01
9.58075643e-01 8.34631920e-02 3.96977425e-01 -4.99148697e-01
1.04182923e+00 -2.54872918e-01 2.60144651e-01 6.81454301e-01
-6.12990856e-01 1.15781501e-01 4.05827999e-01 -3.88581932e-01
-9.76864874e-01 -1.17391217e+00 -2.03607902e-01 7.91635215e-01
5.46959937e-01 -6.19794689e-02 -1.17787921e+00 -4.92415071e-01
-1.04348466e-01 5.75551450e-01 -5.52535176e-01 1.92012653e-01
-5.69103718e-01 -8.94159615e-01 3.97822827e-01 4.04388547e-01
8.02531958e-01 -6.67943895e-01 -6.85357809e-01 3.52406859e-01
-1.26030996e-01 -1.51525533e+00 1.75408944e-01 -4.05275553e-01
-1.16153574e+00 -1.06111407e+00 -7.75233388e-01 -6.58503115e-01
5.04578650e-01 3.31065953e-01 1.03909123e+00 3.20262820e-01
-3.88401747e-01 7.79919803e-01 -3.75662327e-01 -4.07685280e-01
-6.79479599e-01 -2.65688479e-01 -4.42896746e-02 4.58353192e-01
7.38888800e-01 -1.20046586e-01 -4.08597082e-01 4.54246640e-01
-6.92445576e-01 -1.90243140e-01 5.38285732e-01 2.36860618e-01
4.07645732e-01 4.26598608e-01 2.28927717e-01 -7.08591223e-01
5.01827896e-02 -6.25800729e-01 -9.39405918e-01 -9.73978080e-03
-8.70394781e-02 -2.20669821e-01 4.78214994e-02 -4.78865564e-01
-1.30527639e+00 1.00144796e-01 -1.03215061e-01 -3.87542665e-01
-5.58969140e-01 -1.85383067e-01 -4.33288276e-01 -2.46409968e-01
2.15596676e-01 3.34195644e-01 -3.21398908e-03 -3.65283877e-01
1.81763604e-01 6.54478788e-01 6.16673410e-01 -4.33678389e-01
6.84893906e-01 8.80923331e-01 2.99882412e-01 -1.61123288e+00
-2.69139946e-01 -7.17119336e-01 -7.98159361e-01 -8.81544411e-01
9.61966217e-01 -7.33189344e-01 -7.37460196e-01 6.05183601e-01
-1.17252541e+00 -2.46536747e-01 -4.83960323e-02 7.26016879e-01
-7.40344107e-01 5.98554432e-01 -4.54811543e-01 -1.47635758e+00
1.89224944e-01 -9.95226026e-01 9.75331426e-01 3.82097274e-01
2.46021271e-01 -1.33908451e+00 -1.70776531e-01 3.74509603e-01
5.16509473e-01 2.24100187e-01 2.69241393e-01 -9.36917663e-02
-1.07678568e+00 -1.23894170e-01 -3.46381932e-01 2.24949613e-01
1.92291522e-03 3.86924893e-01 -8.50458324e-01 1.07746869e-01
-9.63998437e-02 5.94625831e-01 8.83517027e-01 6.68471634e-01
7.98975646e-01 -5.26734442e-02 -6.33204341e-01 3.41835171e-01
1.51137221e+00 3.51258218e-01 1.06367564e+00 5.83515227e-01
5.90716839e-01 3.39237481e-01 8.60610008e-01 3.48609954e-01
3.81098092e-01 5.98299623e-01 3.87909383e-01 2.98869330e-02
-3.71533483e-01 2.76432842e-01 2.71913141e-01 5.82620203e-01
-4.28236723e-01 -3.59273374e-01 -7.63558745e-01 6.22523189e-01
-1.97531176e+00 -1.67149723e+00 -7.08413959e-01 1.88761389e+00
1.90706715e-01 3.82531643e-01 5.30543208e-01 5.44373751e-01
1.21421683e+00 -4.36064079e-02 1.27983257e-01 -4.72052187e-01
7.46085271e-02 -3.17174569e-02 8.50071192e-01 7.91897833e-01
-1.42372727e+00 1.05918825e+00 7.59350777e+00 1.00546324e+00
-8.11891496e-01 3.86784881e-01 4.52911288e-01 5.61572649e-02
2.78390288e-01 -1.62433252e-01 -8.53223383e-01 7.19771564e-01
1.29555988e+00 -1.77139789e-01 -5.12838811e-02 8.29827666e-01
3.54069501e-01 -9.24322665e-01 -5.81198275e-01 9.48398471e-01
2.11167216e-01 -1.20563614e+00 1.07673556e-01 1.43448964e-01
4.52063262e-01 -1.53158173e-01 -2.38782838e-01 1.48395434e-01
-2.94527374e-02 -6.02429092e-01 7.88379669e-01 8.41140628e-01
3.13017935e-01 -9.59858537e-01 9.42122996e-01 2.89591312e-01
-1.62314880e+00 1.80049077e-01 -5.61551988e-01 -2.56334990e-01
5.19874036e-01 6.97379291e-01 -8.23962510e-01 2.23800927e-01
6.03455544e-01 3.25394750e-01 -4.33840752e-01 1.44398332e+00
3.41401160e-01 7.05337822e-01 -4.07565624e-01 -3.42964679e-01
3.23274106e-01 -2.57237792e-01 6.25983894e-01 1.82489657e+00
3.58455062e-01 -4.63597253e-02 7.76858479e-02 5.18862307e-01
5.32266736e-01 9.50844213e-03 -5.64148366e-01 3.18736464e-01
2.95033932e-01 1.23627210e+00 -1.42055631e+00 -9.87125993e-01
-6.10938132e-01 8.40837836e-01 -3.28988940e-01 1.49705783e-01
-1.11296940e+00 -1.22626327e-01 5.93816221e-01 5.06576598e-01
8.10670137e-01 -3.51860583e-01 3.01697463e-01 -8.90234768e-01
-3.22057158e-01 -1.77520648e-01 9.81105939e-02 -4.95131373e-01
-7.52828181e-01 2.71822631e-01 6.44791186e-01 -1.08855462e+00
-7.25585371e-02 -9.21146274e-01 -4.91166890e-01 5.13547838e-01
-1.48033404e+00 -9.13580954e-01 -4.11838740e-01 6.72158003e-01
8.19848299e-01 -2.93602377e-01 2.46313781e-01 7.54272699e-01
-2.77041107e-01 -5.24797849e-02 1.02887198e-01 1.59634188e-01
2.62790341e-02 -1.11102259e+00 2.05440849e-01 1.09432089e+00
-4.53031957e-02 2.52164640e-02 9.51153159e-01 -6.85256004e-01
-1.17631471e+00 -1.01397729e+00 7.16125190e-01 -5.35217702e-01
6.23913467e-01 -2.04511568e-01 -7.50890017e-01 3.27163279e-01
-1.54175013e-01 3.24074000e-01 4.81317610e-01 -3.20361793e-01
2.30707422e-01 -7.61501342e-02 -1.36198127e+00 1.68176726e-01
6.88230872e-01 -3.10066164e-01 -6.65681303e-01 -1.12350710e-01
-3.24180839e-03 -1.46621272e-01 -5.60588539e-01 3.06180358e-01
5.74707389e-01 -1.11640573e+00 1.00646567e+00 8.99097100e-02
-4.75928813e-01 -7.46665061e-01 -2.60225594e-01 -5.25120080e-01
-3.23098302e-01 -4.02463287e-01 -6.94576502e-01 1.19236922e+00
-7.23858550e-02 -3.20945799e-01 1.02426612e+00 1.69473767e-01
2.00015053e-01 -1.59649387e-01 -1.18193948e+00 -1.16921020e+00
-4.55915183e-01 -6.61583304e-01 2.35695153e-01 4.66918617e-01
-3.87458205e-01 7.46163577e-02 -3.61851573e-01 1.55073076e-01
8.94830465e-01 -4.85469401e-01 8.57680380e-01 -1.45678830e+00
1.16037384e-01 -3.73718351e-01 -1.36743808e+00 -7.33124137e-01
8.31563026e-02 -2.19770655e-01 -3.25512588e-02 -1.79108334e+00
2.11187094e-01 -1.40239552e-01 -1.76548466e-01 -3.54565203e-01
-8.73195678e-02 6.42675281e-01 2.15776742e-01 -1.83549508e-01
-1.06328702e+00 1.61642298e-01 9.12253499e-01 -2.24198364e-02
2.48844117e-01 8.35240409e-02 1.00326046e-01 1.09268212e+00
7.36932635e-01 -5.64574063e-01 -3.96010607e-01 6.77566975e-02
-4.82701778e-01 -1.45724460e-01 6.95076466e-01 -1.48708856e+00
6.89286813e-02 -2.39550546e-01 4.28280354e-01 -9.22105730e-01
4.73177433e-01 -9.51135218e-01 1.72976017e-01 7.08394885e-01
5.52992463e-01 -1.47701234e-01 2.27369025e-01 8.14109623e-01
-7.08297119e-02 -5.48624754e-01 9.40610528e-01 -2.02233538e-01
-1.35010302e+00 1.21170811e-01 -1.29371715e+00 -2.47538716e-01
1.52676177e+00 -8.59311163e-01 -1.72172561e-01 -3.00222397e-01
-6.64870799e-01 -4.51704621e-01 4.63819474e-01 3.69551212e-01
3.67381334e-01 -1.15350771e+00 -5.64421356e-01 -2.94291321e-02
-2.67313689e-01 -6.23156071e-01 2.00506970e-01 1.01942992e+00
-1.21446371e+00 3.89901370e-01 -2.29784817e-01 -1.00269783e+00
-1.64911282e+00 7.42332637e-01 3.00090641e-01 1.49346501e-01
-5.47988176e-01 5.11668086e-01 3.51524502e-02 2.81814009e-01
3.82103734e-02 -2.94372618e-01 -2.72970647e-01 -6.31650463e-02
9.19370592e-01 9.70112026e-01 -5.14259785e-02 -1.34175706e+00
-6.53254867e-01 5.63759983e-01 3.36165518e-01 -2.45392382e-01
7.29556680e-01 -4.10916746e-01 1.86172023e-01 7.30170846e-01
1.04826772e+00 -2.35295236e-01 -1.20833194e+00 8.96399189e-03
1.85630500e-01 -9.39864337e-01 3.83506030e-01 -2.98220038e-01
-1.04567814e+00 7.58618295e-01 1.01812363e+00 4.46744114e-01
9.11341429e-01 3.43405865e-02 4.45321620e-01 4.23114270e-01
5.69207489e-01 -1.37216282e+00 -1.75815254e-01 4.72823620e-01
1.81384668e-01 -1.14362347e+00 1.97693914e-01 -7.97104537e-01
-3.20002079e-01 1.04989278e+00 4.84946817e-01 -1.20664299e-01
9.33097482e-01 6.92662477e-01 -5.30292979e-03 4.21241000e-02
-7.18669534e-01 -8.72099400e-01 2.81582415e-01 1.10567224e+00
2.80564755e-01 3.98727953e-02 -2.93455005e-01 -2.35382542e-01
2.36438349e-01 2.26704791e-01 5.37697434e-01 1.17462540e+00
-1.12563801e+00 -9.15292382e-01 -8.57414067e-01 3.65728766e-01
-6.61497474e-01 2.58454502e-01 5.10570258e-02 9.99493837e-01
3.42257470e-01 1.14381814e+00 4.59388256e-01 2.79448964e-02
1.27054200e-01 -7.73103982e-02 9.59399045e-01 -1.79271013e-01
-4.40592580e-02 1.11289464e-01 2.84573615e-01 -4.86515284e-01
-1.14766002e+00 -8.57351243e-01 -1.02174842e+00 -7.08489537e-01
-5.55002093e-01 6.37758300e-02 8.27911079e-01 1.07158577e+00
-1.35306939e-01 5.49649954e-01 5.47029227e-02 -1.18123651e+00
4.81147617e-01 -8.43833148e-01 -6.64336979e-01 1.96245477e-01
2.76954591e-01 -1.20672977e+00 -1.93653464e-01 4.49060351e-01] | [8.835467338562012, -0.8569074869155884] |
20d0bedd-993d-4bf3-9e04-5e159a1f9fff | texture-synthesis-via-projection-onto | 2105.10825 | null | https://arxiv.org/abs/2105.10825v1 | https://arxiv.org/pdf/2105.10825v1.pdf | Texture synthesis via projection onto multiscale, multilayer statistics | We provide a new model for texture synthesis based on a multiscale, multilayer feature extractor. Within the model, textures are represented by a set of statistics computed from ReLU wavelet coefficients at different layers, scales and orientations. A new image is synthesized by matching the target statistics via an iterative projection algorithm. We explain the necessity of the different types of pre-defined wavelet filters used in our model and the advantages of multilayer structures for image synthesis. We demonstrate the power of our model by generating samples of high quality textures and providing insights into deep representations for texture images. | ['Matthew Hirn', 'Jieqian He'] | 2021-05-22 | null | null | null | null | ['texture-synthesis'] | ['computer-vision'] | [ 2.84657866e-01 -1.05425701e-01 3.22837234e-02 -3.43661398e-01
-5.18392324e-01 -1.44756868e-01 6.96622193e-01 -3.00052524e-01
6.47668615e-02 5.06880343e-01 5.23300409e-01 4.01823252e-01
-7.06923231e-02 -1.06439781e+00 -6.64913893e-01 -8.87972295e-01
-1.10456593e-01 -1.07863143e-01 2.65032232e-01 -3.45631719e-01
2.17986867e-01 7.13792503e-01 -2.01337743e+00 1.17665577e+00
5.92404790e-02 1.42726862e+00 3.80512804e-01 8.24229062e-01
6.27939124e-03 6.88431859e-01 -4.12539929e-01 -9.66438726e-02
1.18683159e-01 -4.78769869e-01 -5.02206147e-01 6.63948834e-01
6.79272413e-01 -2.56154329e-01 -4.21702206e-01 8.96094382e-01
2.81030148e-01 7.06776083e-02 8.10609639e-01 -5.71714044e-01
-8.16285729e-01 3.91712725e-01 -2.80011296e-01 3.91423926e-02
3.05818290e-01 -2.74169415e-01 1.06993687e+00 -1.33689201e+00
1.02730298e+00 1.57003331e+00 8.09059620e-01 4.99134183e-01
-1.96082532e+00 -2.16930173e-03 -1.43857822e-01 4.47829142e-02
-1.06909251e+00 -7.68280029e-01 8.63482475e-01 -3.26184988e-01
8.93403053e-01 5.34524083e-01 7.36525416e-01 1.39459276e+00
6.43975973e-01 6.95381284e-01 1.75205517e+00 -5.95117509e-01
2.68673807e-01 2.49092907e-01 -1.78608179e-01 8.94561529e-01
-1.66366279e-01 1.60336733e-01 -1.02972198e+00 -1.20676309e-01
1.45420301e+00 -2.91749239e-01 6.61226735e-02 -3.87167275e-01
-1.33834577e+00 5.81659973e-01 1.56655058e-01 3.48525196e-01
-4.48732376e-01 2.19989911e-01 6.83534518e-02 6.10945702e-01
6.95536315e-01 2.42523462e-01 -3.82724017e-01 3.84287268e-01
-9.11019266e-01 2.21477404e-01 6.19189322e-01 7.36554503e-01
8.59697163e-01 3.51541907e-01 -4.17284399e-01 1.01354539e+00
2.53609177e-02 6.68318510e-01 3.64903629e-01 -1.11898327e+00
-3.17235559e-01 -2.83282008e-02 8.78641009e-02 -1.40321052e+00
-5.35553135e-02 -3.43730181e-01 -8.33696544e-01 6.45326197e-01
2.25589991e-01 3.69510323e-01 -9.39297080e-01 1.45571828e+00
4.50834678e-03 4.13109884e-02 -3.98393497e-02 6.20445251e-01
7.70095885e-01 5.62851131e-01 -3.93505216e-01 2.59536028e-01
1.16839254e+00 -6.47693574e-01 -6.72301650e-01 3.83907050e-01
-7.97030628e-02 -9.81613696e-01 1.18712687e+00 3.65941703e-01
-1.23612714e+00 -7.98124075e-01 -1.09237909e+00 -6.30438030e-02
-3.36827129e-01 3.38865578e-01 6.76442742e-01 3.98878634e-01
-1.37467098e+00 9.68244970e-01 -8.46571565e-01 -4.77203310e-01
4.63329107e-01 1.71287909e-01 -4.16567028e-01 1.79120168e-01
-6.51553571e-01 6.88931108e-01 -2.35161617e-01 6.32389495e-03
-8.18830252e-01 -2.93961704e-01 -8.40408742e-01 7.59453997e-02
-4.79054421e-01 -7.35618055e-01 8.85752201e-01 -1.10265887e+00
-1.95278037e+00 9.88495946e-01 -5.03522515e-01 -5.95406815e-02
2.11722657e-01 1.16856433e-02 -2.63398379e-01 4.32467848e-01
-1.66949168e-01 6.23600185e-01 1.37798488e+00 -1.60183561e+00
-6.38115168e-01 5.70874214e-02 -3.16956341e-01 -1.26323551e-01
-4.70751971e-01 5.56226783e-02 -7.86658674e-02 -1.07209802e+00
4.55237418e-01 -5.01824558e-01 -3.44457060e-01 6.76684454e-02
-2.67022699e-01 6.35912299e-01 7.39081383e-01 -6.60390913e-01
7.43352115e-01 -2.25114226e+00 2.63736248e-01 5.15176058e-01
3.86948764e-01 -5.47979414e-01 -5.78943610e-01 3.19903851e-01
1.04343005e-01 -8.74690935e-02 6.10714369e-02 -3.79404992e-01
1.80927175e-03 1.87634021e-01 -6.87181294e-01 2.18966633e-01
4.89277780e-01 6.94491446e-01 -7.39481032e-01 -2.22638384e-01
3.60298663e-01 7.09334791e-01 -6.70470774e-01 1.54949352e-01
-7.70916939e-02 6.13297462e-01 -4.65431929e-01 7.63845205e-01
5.82404971e-01 -8.79302919e-02 1.63204730e-01 -6.86663926e-01
-2.78527141e-01 3.09142739e-01 -1.01896167e+00 1.30005002e+00
-6.15091681e-01 8.00618768e-01 1.69493794e-01 -9.65171158e-01
1.13986838e+00 4.56023544e-01 3.69042248e-01 -5.57714343e-01
6.64397925e-02 3.72607321e-01 -2.55457073e-01 -2.89671928e-01
4.42232639e-01 -1.40676856e-01 9.26077813e-02 3.46375912e-01
5.95059335e-01 -3.90923947e-01 2.58486152e-01 -4.19070035e-01
1.05376887e+00 2.39511296e-01 -1.31199695e-02 -9.05479074e-01
4.64912713e-01 -1.69101745e-01 2.45618716e-01 1.01360857e+00
3.53464663e-01 9.53528166e-01 3.73501092e-01 -1.05133998e+00
-1.48502576e+00 -1.29833019e+00 -4.24292713e-01 9.77537453e-01
-3.87651622e-01 -2.95913696e-01 -5.74904144e-01 -2.66588420e-01
2.57419161e-02 2.93327048e-02 -9.61136401e-01 9.52618942e-02
-4.39290762e-01 -7.52138793e-01 1.76568598e-01 1.66922539e-01
5.44110060e-01 -1.29921913e+00 -6.17323816e-01 2.34819219e-01
4.75771204e-02 -1.24399340e+00 -1.55585736e-01 1.14775598e-01
-1.20965755e+00 -4.69680667e-01 -6.31114244e-01 -1.07327807e+00
9.26813900e-01 2.44889557e-01 1.24316299e+00 -5.58266751e-02
-3.35154742e-01 5.28187692e-01 -1.90067291e-01 4.05708253e-02
-4.36075211e-01 -2.24346027e-01 2.04175100e-01 5.29497862e-01
-1.88418537e-01 -9.80557263e-01 -3.65919024e-01 1.75657630e-01
-9.09049749e-01 5.21044552e-01 5.90174854e-01 1.15534925e+00
8.40323806e-01 7.01208860e-02 2.06480116e-01 -6.96782291e-01
7.85008192e-01 8.27528164e-02 -4.10499185e-01 6.02135919e-02
-2.08881032e-03 2.40349546e-01 4.79084283e-01 -6.85077846e-01
-1.06386578e+00 -1.08356059e-01 -8.45716447e-02 -1.07782587e-01
-2.51881480e-01 4.26070184e-01 1.20245449e-01 -4.59317982e-01
6.15000367e-01 4.54781920e-01 1.21623658e-01 -6.03236794e-01
4.35791403e-01 1.19888574e-01 3.88047546e-01 -1.03620434e+00
6.09155774e-01 1.12477219e+00 8.80197808e-02 -1.21824872e+00
-7.72650003e-01 4.53148425e-01 -8.83406878e-01 -3.08080703e-01
4.57379937e-01 -5.72256207e-01 -5.36007702e-01 7.36688793e-01
-1.09412897e+00 -5.05872488e-01 -7.61898935e-01 2.72053152e-01
-9.05905545e-01 -7.39381760e-02 -9.96533036e-01 -5.23571908e-01
-4.11520377e-02 -1.20710027e+00 1.12486184e+00 -8.83034691e-02
-3.25923860e-01 -1.13516235e+00 8.89793560e-02 -4.30124193e-01
8.84929717e-01 4.12819773e-01 1.15781283e+00 2.89281905e-01
-4.78389025e-01 1.42580476e-02 -2.25818321e-01 4.46959525e-01
4.60268855e-01 2.43039906e-01 -1.02032733e+00 -1.41000509e-01
2.82747090e-01 -4.57795769e-01 1.07204580e+00 6.79542363e-01
1.28316903e+00 -3.60411853e-01 -1.26316443e-01 9.62039351e-01
1.33137810e+00 -2.91375339e-01 5.42493939e-01 4.49404120e-01
4.58456218e-01 7.28672385e-01 -1.75708726e-01 3.71020287e-01
2.40297280e-02 5.29259682e-01 -2.79760689e-01 -4.24707115e-01
-4.73874688e-01 -6.29353598e-02 4.21376348e-01 1.11560237e+00
-3.36049497e-01 3.79387677e-01 -4.33405995e-01 4.29440945e-01
-1.66306019e+00 -9.63333011e-01 3.14259976e-01 1.94613791e+00
8.82208526e-01 2.40656883e-01 -2.33715996e-01 5.24476580e-02
3.93551886e-01 2.93380529e-01 -1.69533640e-02 -5.89452147e-01
-7.13799179e-01 8.69432628e-01 1.47946358e-01 7.13131607e-01
-1.10573983e+00 1.14762890e+00 8.74840641e+00 7.93033063e-01
-1.20952833e+00 -2.85163045e-01 7.74471164e-01 1.25646263e-01
-6.60573721e-01 -7.52514899e-02 -5.68666875e-01 -3.09950532e-03
5.47107279e-01 1.15297318e-01 5.63597739e-01 3.55640411e-01
1.95154414e-01 -3.23265716e-02 -1.01633394e+00 6.63252294e-01
-6.66104555e-02 -1.93762720e+00 6.16956592e-01 -2.23555360e-02
9.77329016e-01 -1.46445826e-01 5.06635606e-01 -1.77549928e-01
4.61370230e-01 -1.04177070e+00 1.08166325e+00 9.80998278e-01
8.92566860e-01 -4.31373000e-01 2.45726258e-01 -3.75645489e-01
-1.12023449e+00 1.88086152e-01 -6.48145199e-01 -2.64069468e-01
-1.88541353e-01 8.38218093e-01 -1.64927259e-01 9.35446993e-02
8.88928652e-01 9.06615138e-01 -3.72018188e-01 5.12605369e-01
-6.04265779e-02 3.70976299e-01 -3.62199545e-01 1.00710347e-01
-7.66300261e-02 -4.44634348e-01 4.32488769e-01 1.42389929e+00
6.36342347e-01 -1.05519183e-01 -8.34247991e-02 1.13838649e+00
2.28845403e-01 -1.75748110e-01 -6.98821902e-01 9.65084583e-02
1.85597241e-01 1.43219936e+00 -9.53265131e-01 -3.74646008e-01
-4.06105369e-01 9.82238293e-01 3.47317815e-01 6.85128748e-01
-1.91777185e-01 -1.84692547e-01 8.01812410e-01 8.13411102e-02
5.77129900e-01 -5.74649870e-01 -6.75634265e-01 -1.29518127e+00
-1.52561575e-01 -1.10764718e+00 -2.82988369e-01 -6.66594386e-01
-1.34475899e+00 8.81571174e-01 -2.27729678e-01 -1.09704828e+00
1.39970064e-01 -9.04249310e-01 -5.01675785e-01 9.44518924e-01
-1.20342290e+00 -1.19939673e+00 -1.32286459e-01 5.40302455e-01
2.97864825e-01 -3.36360067e-01 1.32011783e+00 -4.10303138e-02
-2.13227689e-01 2.50547707e-01 2.12242365e-01 -1.75783321e-01
3.86956453e-01 -9.82888758e-01 7.62235522e-01 4.66261089e-01
2.85222024e-01 5.15220881e-01 9.22637880e-01 -3.58069360e-01
-1.36843741e+00 -6.20235264e-01 7.84666836e-01 -4.62695211e-01
7.91181862e-01 -7.76782930e-01 -6.58670008e-01 5.21629810e-01
1.40306234e-01 2.67290860e-01 5.63111842e-01 -9.22342092e-02
-3.77143800e-01 -1.44005463e-01 -1.08056569e+00 8.06681156e-01
8.80060554e-01 -7.85616755e-01 -3.59750569e-01 6.80037960e-02
1.66497290e-01 -2.12710828e-01 -9.58351076e-01 4.43731934e-01
1.21483731e+00 -1.22701979e+00 1.09017718e+00 -6.45334184e-01
8.55640233e-01 2.90552583e-02 -6.13363624e-01 -1.28737557e+00
-8.02274644e-01 -4.00650501e-01 -5.22109121e-02 7.33692706e-01
4.52570796e-01 -5.04848063e-01 7.54935265e-01 -1.20345801e-01
1.76005512e-01 -8.68527174e-01 -8.48804355e-01 -5.27202368e-01
-1.56743139e-01 -3.31959069e-01 3.73840332e-01 8.25419486e-01
-9.60606262e-02 -5.57439439e-02 -3.88624966e-01 4.39592823e-03
7.29420185e-01 3.81169379e-01 6.82910442e-01 -1.03495610e+00
-3.16583008e-01 -5.60483754e-01 -3.68959367e-01 -9.18450058e-01
-1.36238798e-01 -7.44686663e-01 -7.14053819e-03 -1.29960954e+00
1.74401835e-01 -2.94301242e-01 -4.08616781e-01 2.47679889e-01
3.13681364e-01 7.08428323e-01 -5.30005731e-02 2.85106391e-01
-1.68424863e-02 6.57125354e-01 1.50941241e+00 1.51731279e-02
3.67968194e-02 -3.64135385e-01 -3.55280221e-01 8.90388250e-01
7.33010352e-01 -3.73257965e-01 -4.93257819e-03 -4.37723219e-01
2.14023292e-01 -2.21165150e-01 3.07475269e-01 -1.04128683e+00
-3.79005611e-01 -3.18620056e-01 8.76548588e-01 -8.12272802e-02
5.86767018e-01 -4.60457116e-01 6.20979257e-02 2.73443490e-01
-4.42166448e-01 1.11713380e-01 2.82031685e-01 1.71898216e-01
-5.07593215e-01 2.44354948e-01 9.67728436e-01 -3.37706804e-01
-5.21189392e-01 6.03481010e-02 -9.70771968e-01 -6.45850241e-01
2.81233579e-01 -3.98695350e-01 9.62089282e-03 -3.41268301e-01
-1.25807083e+00 -5.49793124e-01 5.85473418e-01 2.34505102e-01
1.07959306e+00 -1.73230565e+00 -7.75403380e-01 9.24066842e-01
-4.52414639e-02 -6.59668446e-01 -6.68775514e-02 6.59628510e-01
-5.75432479e-01 -9.36591253e-03 -8.76554787e-01 -5.50620615e-01
-1.00634670e+00 -9.87412594e-03 3.82433236e-01 -1.52560368e-01
-8.65264475e-01 9.17093098e-01 3.37649167e-01 -3.41012210e-01
-1.72490895e-01 -4.83079344e-01 -8.62681270e-02 -1.38714060e-01
4.80421901e-01 1.10456385e-01 3.32981832e-02 -6.19301200e-01
4.77565359e-03 8.14728737e-01 1.19805694e-01 -6.75471187e-01
1.48001325e+00 -3.65407765e-02 -5.65109491e-01 7.61137784e-01
8.67163897e-01 3.02650124e-01 -1.41332030e+00 -3.17517936e-01
-8.59528407e-02 -4.76345986e-01 4.79691476e-02 -3.75345200e-01
-1.05354643e+00 6.13417447e-01 4.41122621e-01 1.09490819e-01
1.09830284e+00 -1.64821163e-01 6.94900751e-01 3.33541095e-01
5.69975019e-01 -1.09695292e+00 3.04433733e-01 7.94876039e-01
1.22162080e+00 -7.44135499e-01 5.13532571e-02 -3.07242751e-01
-1.43326208e-01 1.54475617e+00 -1.85227804e-02 -9.25970972e-01
1.03919065e+00 7.15163767e-01 3.01893741e-01 -3.16358119e-01
-8.57607603e-01 -1.78102225e-01 5.27695894e-01 6.43803239e-01
6.67631388e-01 1.25830054e-01 -1.30359665e-01 1.64224714e-01
-3.88773620e-01 -3.12863261e-01 4.34001595e-01 9.30851340e-01
-5.31832039e-01 -1.15788722e+00 -5.01671195e-01 4.69482511e-01
-3.89133573e-01 -1.47847369e-01 -2.83100188e-01 3.11497927e-01
-1.17179705e-03 5.36491573e-01 1.25300840e-01 -7.56537855e-01
1.73580065e-01 -8.52279738e-03 1.09723377e+00 -5.56104660e-01
-1.86368927e-01 4.30229515e-01 5.31815633e-04 -6.30591214e-01
-3.73350203e-01 -6.89096272e-01 -5.49377918e-01 -1.80090338e-01
1.72976449e-01 -2.33614802e-01 5.04842520e-01 7.22566903e-01
2.95779556e-01 6.56705379e-01 8.53783250e-01 -1.52446425e+00
-1.29113108e-01 -1.01031625e+00 -8.70552778e-01 5.13517678e-01
4.47706401e-01 -6.56854331e-01 -4.34366614e-01 5.50002992e-01] | [11.412171363830566, -0.49118563532829285] |
04d01856-16cc-4fda-b659-7638761ec386 | prompting-diffusion-representations-for-cross | 2307.02138 | null | https://arxiv.org/abs/2307.02138v1 | https://arxiv.org/pdf/2307.02138v1.pdf | Prompting Diffusion Representations for Cross-Domain Semantic Segmentation | While originally designed for image generation, diffusion models have recently shown to provide excellent pretrained feature representations for semantic segmentation. Intrigued by this result, we set out to explore how well diffusion-pretrained representations generalize to new domains, a crucial ability for any representation. We find that diffusion-pretraining achieves extraordinary domain generalization results for semantic segmentation, outperforming both supervised and self-supervised backbone networks. Motivated by this, we investigate how to utilize the model's unique ability of taking an input prompt, in order to further enhance its cross-domain performance. We introduce a scene prompt and a prompt randomization strategy to help further disentangle the domain-invariant information when training the segmentation head. Moreover, we propose a simple but highly effective approach for test-time domain adaptation, based on learning a scene prompt on the target domain in an unsupervised manner. Extensive experiments conducted on four synthetic-to-real and clear-to-adverse weather benchmarks demonstrate the effectiveness of our approaches. Without resorting to any complex techniques, such as image translation, augmentation, or rare-class sampling, we set a new state-of-the-art on all benchmarks. Our implementation will be publicly available at \url{https://github.com/ETHRuiGong/PTDiffSeg}. | ['Luc van Gool', 'Julio Delgado Mangas', 'Han Sun', 'Martin Danelljan', 'Rui Gong'] | 2023-07-05 | null | null | null | null | ['image-generation', 'domain-generalization'] | ['computer-vision', 'methodology'] | [ 5.21121860e-01 5.59067801e-02 -2.17224091e-01 -5.42588055e-01
-8.34299386e-01 -8.78964722e-01 7.67949939e-01 -3.49372685e-01
-3.13372701e-01 7.22531915e-01 4.70910482e-02 -4.52452600e-01
6.86526000e-02 -8.74444246e-01 -8.34370375e-01 -6.92184329e-01
-1.44835114e-02 4.93371785e-01 3.61744195e-01 -2.70126283e-01
7.31898099e-02 5.31659484e-01 -1.28829980e+00 1.21396728e-01
1.11988723e+00 7.44615018e-01 3.67405146e-01 5.13822734e-01
-1.67389289e-01 5.26130855e-01 -6.86503887e-01 -1.10934280e-01
4.05848235e-01 -7.68225014e-01 -1.01606560e+00 4.55502033e-01
1.45588949e-01 -3.43072057e-01 -2.77326763e-01 9.12992001e-01
5.39296269e-01 3.72039855e-01 8.88207316e-01 -9.07586753e-01
-9.77797210e-01 4.22559887e-01 -4.73726898e-01 3.10491502e-01
3.00146103e-01 4.90878493e-01 7.56643653e-01 -7.47739136e-01
9.67116296e-01 9.13062692e-01 4.32661504e-01 8.05098176e-01
-1.12729478e+00 -8.03465307e-01 3.64028245e-01 1.76209524e-01
-1.25839698e+00 -4.35929507e-01 9.78588581e-01 -4.35482591e-01
6.85726881e-01 -7.07362220e-02 4.38134789e-01 1.49387991e+00
-3.25363457e-01 9.35605049e-01 1.39456427e+00 -4.54896748e-01
3.47758085e-01 1.74547940e-01 -1.61826089e-01 5.56409240e-01
5.03737479e-03 2.58638352e-01 -3.03875208e-01 5.46213984e-02
9.30871010e-01 -4.07480866e-01 -3.76897603e-01 -4.89708990e-01
-1.15809429e+00 8.93751860e-01 6.69215858e-01 2.53448248e-01
-3.08877975e-01 -1.71731427e-01 2.29211032e-01 3.43426228e-01
7.10746646e-01 5.66889167e-01 -6.42745018e-01 -5.77414297e-02
-7.73873031e-01 2.05172673e-01 6.86686933e-01 9.41835225e-01
8.21799278e-01 3.12531143e-01 -6.85925409e-02 8.35222602e-01
-2.64029726e-02 5.32662928e-01 6.27183616e-01 -1.00659132e+00
3.65156472e-01 3.19109023e-01 -9.22359601e-02 -5.22687674e-01
-2.67389476e-01 -5.03842354e-01 -8.34797263e-01 2.46409401e-02
5.04518211e-01 -4.42873240e-01 -1.40615726e+00 1.82570875e+00
4.14622366e-01 3.68101954e-01 3.64096314e-01 8.16547811e-01
5.39777935e-01 6.33886516e-01 3.32557820e-02 3.69012505e-02
9.55137253e-01 -1.10839760e+00 -2.40376711e-01 -4.47563469e-01
7.21878767e-01 -6.33034170e-01 1.35658860e+00 2.66519427e-01
-7.50354409e-01 -6.44774675e-01 -8.49037647e-01 2.17954695e-01
-5.34532845e-01 -1.90934896e-01 6.63075745e-01 6.56608582e-01
-1.01270938e+00 6.65394425e-01 -7.38175392e-01 -6.35803819e-01
6.84294820e-01 8.63742083e-02 -2.59200096e-01 -2.83416063e-01
-1.17853534e+00 5.97926438e-01 4.56582725e-01 -2.35012978e-01
-1.09816527e+00 -5.26586115e-01 -8.73442769e-01 -3.16871464e-01
3.40399295e-01 -7.87921309e-01 1.36178887e+00 -1.29417515e+00
-1.77092814e+00 8.73895824e-01 -1.30186051e-01 -5.20652354e-01
4.54186887e-01 -4.83715571e-02 -4.12987202e-01 3.86925370e-01
3.36639822e-01 1.04594350e+00 9.83736455e-01 -1.44295204e+00
-3.74029517e-01 -1.65415376e-01 1.49051696e-01 3.65854234e-01
-3.97297829e-01 -1.82967216e-01 -4.35521871e-01 -8.61864090e-01
2.37592012e-02 -1.05901420e+00 -5.15777946e-01 -1.82105988e-01
-4.22978371e-01 2.66770311e-02 6.79190576e-01 -5.55892706e-01
6.98013186e-01 -2.21153426e+00 6.34441748e-02 1.32862210e-01
-8.94617736e-02 3.48560035e-01 -4.18249100e-01 2.58267760e-01
-1.13577813e-01 1.95593864e-01 -7.22335219e-01 -1.93022177e-01
-1.20245703e-01 3.55780393e-01 -4.42267269e-01 3.44944358e-01
4.80803818e-01 9.37620819e-01 -9.64920282e-01 -3.14542741e-01
6.91304877e-02 2.94843346e-01 -6.72258079e-01 2.54834056e-01
-5.12293398e-01 1.02786148e+00 -6.77090406e-01 5.93660474e-01
7.29527891e-01 -4.30854976e-01 -3.61718982e-02 2.61631936e-01
7.11544231e-02 2.53390998e-01 -9.49913502e-01 2.04517341e+00
-4.57374096e-01 4.87299323e-01 -1.42641157e-01 -1.22394443e+00
1.09107602e+00 7.35558569e-02 2.93699682e-01 -9.28465188e-01
1.43353259e-02 2.68271953e-01 -1.53689355e-01 -3.61342758e-01
4.05956656e-01 -7.27457404e-02 -7.51576796e-02 5.11498809e-01
1.86235189e-01 -4.22887951e-01 2.23144040e-01 1.55932337e-01
9.27460790e-01 3.67984265e-01 1.48617566e-01 -2.17292413e-01
3.37817073e-01 3.62900943e-01 4.70349371e-01 7.43483782e-01
-1.74115062e-01 8.70457053e-01 2.62512922e-01 -1.96455315e-01
-8.47890317e-01 -1.19315410e+00 -5.19364402e-02 1.13344800e+00
1.95760176e-01 1.02384210e-01 -8.80123734e-01 -9.68854427e-01
-1.94912761e-01 8.15024018e-01 -6.58095062e-01 -9.86425355e-02
-4.08256769e-01 -8.22189808e-01 6.23465061e-01 6.44734025e-01
8.11340511e-01 -1.13102806e+00 -3.82647038e-01 1.75729692e-01
-1.12274103e-01 -1.20020366e+00 -2.72123039e-01 4.05630976e-01
-9.87805128e-01 -8.26190710e-01 -1.19797349e+00 -8.37570608e-01
6.99613750e-01 3.19591701e-01 1.00563395e+00 -2.31459334e-01
2.57776044e-02 4.96771842e-01 -5.38482189e-01 -1.96730584e-01
-4.03915197e-01 5.23565471e-01 -1.69428423e-01 -1.27736256e-01
2.10364938e-01 -6.66574836e-01 -6.90923512e-01 3.99663299e-01
-1.04541457e+00 -4.50625755e-02 5.66092610e-01 8.57504547e-01
5.88985860e-01 -1.22562237e-01 6.62365913e-01 -1.09932864e+00
6.46504343e-01 -6.09055281e-01 -5.16881824e-01 3.20848264e-02
-4.08991367e-01 7.52120316e-02 6.45361364e-01 -5.65252125e-01
-1.18995106e+00 1.43226951e-01 -3.55480433e-01 -3.81111622e-01
-5.57651281e-01 4.15768892e-01 7.42162066e-03 -1.23768345e-01
1.01645517e+00 5.43534100e-01 -9.23564434e-02 -4.26899910e-01
7.11835265e-01 6.16155684e-01 4.42028910e-01 -9.13718045e-01
1.01930189e+00 5.98769546e-01 -4.80205536e-01 -9.68458354e-01
-8.44261050e-01 -4.43522304e-01 -7.79271781e-01 4.83135022e-02
9.64052796e-01 -1.05975461e+00 8.46159086e-02 5.80144346e-01
-1.04062629e+00 -9.83280301e-01 -4.67162460e-01 1.46684244e-01
-6.31145895e-01 3.68497223e-01 -4.38970894e-01 -3.02213132e-01
-1.07781859e-02 -1.09443986e+00 8.83308768e-01 2.82864153e-01
-3.48813497e-02 -1.24793124e+00 1.65298864e-01 3.86267632e-01
5.27062953e-01 1.45562142e-01 6.21146321e-01 -8.97695124e-01
-6.36984825e-01 2.58783698e-01 -3.63857150e-01 5.89285254e-01
2.16966331e-01 -2.83893555e-01 -1.14526665e+00 -2.66913682e-01
-3.53780910e-02 -4.95990872e-01 9.99520302e-01 2.94672132e-01
1.24370384e+00 -5.53118214e-02 -3.15268546e-01 9.20546770e-01
1.23721540e+00 1.11583307e-01 5.43970823e-01 6.88836277e-01
5.43237507e-01 4.92960334e-01 5.74013710e-01 3.70260060e-01
4.55919564e-01 3.92214984e-01 1.36043951e-01 -4.96642202e-01
-4.37524557e-01 -2.42076203e-01 2.23752737e-01 6.20196700e-01
7.01459646e-02 -3.74415666e-01 -1.09144044e+00 7.49891579e-01
-1.42962575e+00 -6.14012182e-01 1.39903128e-01 2.03620052e+00
9.60257530e-01 2.66183227e-01 7.01162294e-02 -2.10458934e-01
4.66653764e-01 3.33223462e-01 -7.82940388e-01 -3.26812238e-01
-2.15902671e-01 4.27781016e-01 6.16809845e-01 3.65544528e-01
-1.17923546e+00 1.43755293e+00 6.01257849e+00 8.30058217e-01
-1.25550079e+00 1.88755795e-01 7.25403547e-01 3.83366108e-01
-4.74604815e-01 -8.00827667e-02 -7.22521365e-01 2.35977128e-01
7.75914252e-01 -9.22840461e-02 4.93976235e-01 9.63860095e-01
1.99018791e-02 -1.23662082e-02 -7.97260761e-01 5.19392073e-01
6.89420733e-04 -1.27535236e+00 1.03697203e-01 -8.30103680e-02
1.08664358e+00 3.80430639e-01 2.68798798e-01 4.98432666e-01
5.61868966e-01 -8.05995226e-01 4.62764770e-01 1.15861870e-01
8.99382174e-01 -3.53839010e-01 3.23117733e-01 2.87602872e-01
-1.03359461e+00 9.91676748e-02 -3.82825673e-01 -9.68755409e-02
1.31289124e-01 4.40403104e-01 -1.11756539e+00 6.17381394e-01
6.08253956e-01 7.67344713e-01 -7.45152533e-01 8.84881496e-01
-5.32841742e-01 8.95035565e-01 -3.02479416e-01 2.58351713e-01
4.11236852e-01 -8.77698243e-04 3.71254206e-01 1.25681233e+00
3.33895028e-01 7.06638992e-02 2.42145598e-01 9.47799563e-01
-9.90112796e-02 3.04050557e-02 -9.50631440e-01 -1.39264748e-01
3.15625966e-01 8.85820568e-01 -9.59232569e-01 -4.16873246e-01
-3.71165127e-01 1.31516612e+00 2.63300240e-01 8.93855095e-01
-8.94188046e-01 -3.45051408e-01 6.76047325e-01 1.83944926e-02
6.29020333e-01 -4.57553029e-01 -3.74562174e-01 -1.15645456e+00
-1.22188658e-01 -9.57977295e-01 2.02180088e-01 -6.98122978e-01
-1.36488008e+00 7.93754935e-01 1.28612205e-01 -1.18249655e+00
-3.37024331e-01 -4.80662167e-01 -5.88384867e-01 8.82037640e-01
-1.96579182e+00 -1.13905895e+00 -3.11439514e-01 8.81856859e-01
7.87144065e-01 -1.48843914e-01 8.23531926e-01 1.25318989e-01
-4.97379065e-01 5.39636135e-01 2.46079013e-01 1.79724947e-01
8.37811530e-01 -1.11927676e+00 7.49867439e-01 9.61825311e-01
2.38635495e-01 3.79400879e-01 5.71150720e-01 -5.98251641e-01
-8.92487526e-01 -1.26182652e+00 4.64655966e-01 -4.10303026e-01
7.81496584e-01 -3.48191977e-01 -1.15498519e+00 7.92502642e-01
3.17745268e-01 4.92385635e-03 5.96510351e-01 6.59192950e-02
-3.62479240e-01 -6.71520550e-03 -9.69468951e-01 6.64083540e-01
1.24979126e+00 -3.90861928e-01 -6.56989813e-01 5.36755562e-01
1.02208936e+00 -5.90177953e-01 -5.75282574e-01 5.30026317e-01
-2.37257276e-02 -9.57644939e-01 1.03753150e+00 -4.01470602e-01
4.77182657e-01 -4.85114679e-02 -7.14730248e-02 -1.45407593e+00
-2.59772688e-01 -5.28628290e-01 2.74897307e-01 1.25650859e+00
6.14998043e-01 -9.02275920e-01 9.88384843e-01 1.88992336e-01
-2.04670265e-01 -5.68267822e-01 -6.88226998e-01 -1.09562671e+00
4.50813532e-01 -4.01938230e-01 5.94085813e-01 1.15615213e+00
-4.84662920e-01 3.03034395e-01 -6.78113326e-02 3.16147953e-01
4.44834977e-01 1.60539016e-01 9.51224685e-01 -9.23557699e-01
-3.94379139e-01 -3.91716689e-01 -2.02230178e-02 -1.54105675e+00
2.63790876e-01 -8.40386391e-01 1.28657594e-01 -1.47258711e+00
-1.26533642e-01 -7.24709928e-01 -3.28815937e-01 5.49340546e-01
-1.22153834e-01 3.28001291e-01 7.90461004e-02 2.57317960e-01
-5.56891024e-01 6.97173715e-01 1.73592448e+00 -2.25537136e-01
-4.33893144e-01 8.48397166e-02 -8.30714643e-01 5.78492045e-01
1.16617966e+00 -4.73784357e-01 -7.82317460e-01 -7.13234901e-01
-3.15291852e-01 -1.78205207e-01 2.45077714e-01 -1.12309563e+00
2.75078174e-02 -2.35382318e-01 2.99729705e-01 -1.43684804e-01
2.73282886e-01 -4.60975558e-01 -2.91369319e-01 2.13605881e-01
-1.97332636e-01 -1.99986026e-01 4.67452556e-01 5.94251871e-01
-3.35493237e-01 -3.16287398e-01 7.38780439e-01 -2.82468587e-01
-1.20933902e+00 2.93487489e-01 -2.73303002e-01 4.12536383e-01
1.05530894e+00 -3.19025576e-01 -3.62284809e-01 -3.57464015e-01
-7.39641488e-01 1.32310927e-01 7.01174319e-01 4.00667697e-01
4.91463631e-01 -9.66940522e-01 -6.74967766e-01 3.45773101e-01
2.40448534e-01 1.99725673e-01 2.30454475e-01 6.30980194e-01
-4.89486635e-01 1.78344399e-01 -1.70111313e-01 -8.40439439e-01
-6.99453652e-01 5.28996706e-01 2.03535855e-01 -2.27031425e-01
-5.04257202e-01 1.15932405e+00 4.66058910e-01 -7.79496074e-01
1.73277277e-02 -3.28404158e-01 6.37878617e-03 -9.76440683e-02
1.54668719e-01 -1.58343256e-01 -9.28427130e-02 -3.21665019e-01
-1.98812306e-01 6.07200682e-01 -1.77011251e-01 -2.39160478e-01
1.20920122e+00 -1.03370085e-01 3.81535381e-01 2.65428394e-01
1.00671947e+00 -2.48080924e-01 -1.68447351e+00 -2.90934592e-01
-1.01241678e-01 -3.65641207e-01 -1.32829800e-01 -9.48824108e-01
-9.89872575e-01 1.01201165e+00 5.72417736e-01 1.87513307e-01
1.33989692e+00 1.71637580e-01 8.89687896e-01 3.52219999e-01
2.76711583e-01 -9.87685561e-01 3.55054229e-01 6.89533234e-01
7.62524486e-01 -1.40629315e+00 -2.74480045e-01 -4.09471571e-01
-9.12488878e-01 8.81640553e-01 5.71383595e-01 -3.95081975e-02
5.37953913e-01 4.77493107e-02 3.92219752e-01 5.60235372e-03
-3.90542626e-01 -5.08532882e-01 1.55098826e-01 1.09894395e+00
2.40470335e-01 -1.87782012e-02 1.31118238e-01 2.90295124e-01
-1.66391954e-01 1.37824029e-01 3.81297708e-01 1.02151740e+00
-4.74520713e-01 -1.26171553e+00 -1.92799225e-01 1.46248713e-01
-1.28621310e-01 -1.54358342e-01 -4.41180259e-01 9.38908041e-01
4.45833728e-02 6.82607710e-01 -5.73653579e-02 -3.03922266e-01
3.58631223e-01 1.23747766e-01 4.10440028e-01 -9.27254200e-01
-2.42108181e-01 1.49449520e-02 5.39430901e-02 -4.29196924e-01
-4.46589261e-01 -7.18602836e-01 -1.17903733e+00 -5.96883183e-04
-1.26894489e-01 -2.82619265e-03 4.94086593e-01 9.17025506e-01
3.97574455e-01 6.47370994e-01 6.93844974e-01 -9.69613612e-01
-5.98059595e-01 -9.14700925e-01 -3.41657817e-01 4.54529226e-01
2.89643764e-01 -6.37051761e-01 -3.09604734e-01 2.09242448e-01] | [9.753252983093262, 1.3535077571868896] |
6ad64128-6650-4334-bf73-e0cbbf5593b5 | crowdsourcing-pareto-optimal-object-finding | 1409.4161 | null | http://arxiv.org/abs/1409.4161v1 | http://arxiv.org/pdf/1409.4161v1.pdf | Crowdsourcing Pareto-Optimal Object Finding by Pairwise Comparisons | This is the first study on crowdsourcing Pareto-optimal object finding, which
has applications in public opinion collection, group decision making, and
information exploration. Departing from prior studies on crowdsourcing skyline
and ranking queries, it considers the case where objects do not have explicit
attributes and preference relations on objects are strict partial orders. The
partial orders are derived by aggregating crowdsourcers' responses to pairwise
comparison questions. The goal is to find all Pareto-optimal objects by the
fewest possible questions. It employs an iterative question-selection
framework. Guided by the principle of eagerly identifying non-Pareto optimal
objects, the framework only chooses candidate questions which must satisfy
three conditions. This design is both sufficient and efficient, as it is proven
to find a short terminal question sequence. The framework is further steered by
two ideas---macro-ordering and micro-ordering. By different micro-ordering
heuristics, the framework is instantiated into several algorithms with varying
power in pruning questions. Experiment results using both real crowdsourcing
marketplace and simulations exhibited not only orders of magnitude reductions
in questions when compared with a brute-force approach, but also
close-to-optimal performance from the most efficient instantiation. | ['Gensheng Zhang', 'Chengkai Li', 'Naeemul Hassan', 'Gergely V. Zaruba', 'Abolfazl Asudeh'] | 2014-09-15 | null | null | null | null | ['question-selection'] | ['natural-language-processing'] | [-9.88788232e-02 1.36061147e-01 9.23432112e-02 -3.52869868e-01
-1.11400092e+00 -1.18480790e+00 2.52850652e-01 3.21046889e-01
-8.15623462e-01 8.74669075e-01 3.06732982e-01 -1.45626366e-01
-6.34030938e-01 -6.82052314e-01 -4.07225072e-01 -4.60756838e-01
-6.69633821e-02 1.03934133e+00 7.72443414e-01 -5.75493813e-01
5.10158420e-01 3.18825394e-01 -1.88122988e+00 5.39103031e-01
1.08458233e+00 7.25512862e-01 1.02946900e-01 8.64182174e-01
-5.18738091e-01 5.45706153e-01 -7.08563149e-01 -9.57084239e-01
7.31775224e-01 -3.85744780e-01 -1.18890500e+00 3.33543450e-01
1.36856049e-01 -3.09589475e-01 2.40241289e-01 8.82784188e-01
6.82179928e-01 4.31451827e-01 1.64775595e-01 -1.29626095e+00
-7.79754162e-01 2.81988978e-01 -1.66884288e-01 2.28645518e-01
1.05388534e+00 1.61513582e-01 1.60917854e+00 -8.24879527e-01
5.63539922e-01 1.47110987e+00 1.37263060e-01 4.46242124e-01
-9.30924118e-01 -1.46122858e-01 2.86440760e-01 9.83440876e-02
-1.50193870e+00 -1.51957557e-01 6.97921664e-02 -3.58241081e-01
6.35993481e-01 1.08401275e+00 6.67495310e-01 -5.58867827e-02
-2.35448241e-01 9.23671961e-01 1.34266794e+00 -5.14389992e-01
8.07185471e-01 4.21015143e-01 1.08378321e-01 3.43485504e-01
3.66801471e-01 -1.62086397e-01 -8.39959025e-01 -1.05289328e+00
1.92480218e-02 -8.59893337e-02 -1.79130986e-01 -1.18107490e-01
-8.29582453e-01 9.29170787e-01 5.98783009e-02 1.59830637e-02
-4.22426224e-01 -1.25864714e-01 -1.58891067e-01 3.59721184e-01
4.40955162e-01 9.17768538e-01 -4.69846636e-01 7.78679773e-02
-6.59596264e-01 1.13158417e+00 1.24856389e+00 1.14918149e+00
9.51317847e-01 -8.26454341e-01 -7.42551386e-01 6.68922782e-01
3.24440926e-01 4.59991604e-01 2.60627031e-01 -1.28158712e+00
5.10811090e-01 9.16558623e-01 1.11345148e+00 -8.02164555e-01
-2.66886234e-01 9.68135595e-02 3.60793442e-01 -8.59184470e-03
5.92090428e-01 -3.42176884e-01 -4.35624599e-01 1.26978004e+00
8.86040330e-01 -6.39339685e-01 -3.56175542e-01 1.43913794e+00
7.69316912e-01 2.97471255e-01 -2.76347809e-02 -1.59452930e-01
1.82381618e+00 -5.43502271e-01 -5.42500615e-01 1.29258588e-01
5.75111151e-01 -7.79383004e-01 1.22388899e+00 3.96261215e-01
-1.25262582e+00 1.11357020e-02 -5.61298192e-01 -1.44807085e-01
-4.09844249e-01 -3.06223065e-01 4.32417125e-01 1.04118407e+00
-1.23836529e+00 3.24637890e-01 1.08207360e-01 -4.17714357e-01
1.71301499e-01 6.57145977e-01 8.78244489e-02 2.83728391e-02
-1.20840609e+00 5.51500380e-01 -1.23808468e-02 -2.45259985e-01
-4.79245156e-01 -3.57183307e-01 -2.06253827e-01 -8.64762366e-02
8.25657248e-01 -6.97194934e-01 1.52399719e+00 -7.24319398e-01
-1.39511025e+00 7.57275283e-01 -6.91738963e-01 -3.51077318e-01
7.18638361e-01 -1.64423496e-01 -1.54806033e-01 3.37214231e-01
6.46456897e-01 6.41627908e-01 3.35494071e-01 -1.35199547e+00
-1.26454127e+00 -2.35545814e-01 6.56091332e-01 5.71827888e-01
-2.24249065e-01 5.30695081e-01 -5.75243473e-01 -1.27616376e-01
4.60705236e-02 -1.26614094e+00 -6.91614151e-01 -9.36071053e-02
-3.99997592e-01 -7.07537055e-01 3.35576907e-02 -2.69470841e-01
1.16806090e+00 -1.47542036e+00 -5.08299649e-01 3.54744285e-01
3.69026154e-01 -2.23944876e-02 -2.55240351e-01 7.79957652e-01
6.76959991e-01 3.88846427e-01 8.77495483e-02 4.93259765e-02
3.30501229e-01 1.79444432e-01 -3.26305002e-01 4.48936433e-01
1.25558348e-02 1.01488483e+00 -1.21323657e+00 -7.47076392e-01
-3.41882259e-01 -4.90983665e-01 -6.49221182e-01 1.45685598e-01
-6.01732373e-01 4.54027578e-02 -6.33829355e-01 7.72593856e-01
6.03485584e-01 -3.21181446e-01 5.04689068e-02 5.03621757e-01
-1.15625583e-01 2.68519342e-01 -1.60237503e+00 9.66876566e-01
5.97480126e-02 2.29686350e-01 -5.83726075e-03 -1.67683810e-01
7.59390652e-01 1.96638435e-01 6.12484336e-01 -7.66816676e-01
-3.89210545e-02 5.11463642e-01 -2.18339577e-01 -1.08292830e+00
1.03948450e+00 -4.72688228e-02 -1.60682246e-01 8.68329883e-01
-4.61159319e-01 6.25563264e-02 6.17224336e-01 3.93612444e-01
1.07414949e+00 -2.23626450e-01 1.15504049e-01 -5.39796174e-01
5.33575177e-01 6.05210900e-01 3.48318994e-01 1.31949210e+00
-1.24116540e-01 6.47771955e-01 5.17169535e-01 -4.13771152e-01
-6.06785536e-01 -8.69546354e-01 1.91112354e-01 1.86456072e+00
7.01314449e-01 -3.56129438e-01 -8.45276833e-01 -6.96972907e-01
3.19308996e-01 6.02756262e-01 -7.08437562e-01 6.79285526e-01
-3.44595045e-01 -6.69558644e-01 1.17815472e-01 6.28705844e-02
3.72270010e-02 -9.02369082e-01 -9.26077664e-01 7.41709471e-02
-4.35359597e-01 -9.86453652e-01 -7.21745789e-01 -3.22503597e-01
-4.68233109e-01 -1.34165549e+00 -9.38936114e-01 -3.50894451e-01
7.11033642e-01 8.63116324e-01 1.21964133e+00 4.57823396e-01
3.48186567e-02 7.66869068e-01 -6.98701560e-01 -8.75898600e-01
1.15453050e-01 -5.17058596e-02 -1.32175416e-01 2.05597296e-01
9.87420201e-01 2.96005439e-02 -9.92465556e-01 8.62539589e-01
-1.12721241e+00 -7.78965771e-01 2.43813992e-01 1.92605615e-01
4.44950074e-01 -2.20189214e-01 6.60705984e-01 -7.23855972e-01
1.40896273e+00 -7.91876256e-01 -6.97529793e-01 5.95994234e-01
-5.46234548e-01 -5.88877425e-02 3.64615202e-01 -2.13352025e-01
-7.92492390e-01 8.18356946e-02 4.28629369e-01 1.81558177e-01
-5.96935116e-02 -7.45537877e-02 -1.10352896e-02 -1.78487837e-01
1.03240001e+00 -9.71222147e-02 -2.86218822e-01 -3.50052118e-01
4.44175869e-01 7.17729628e-01 5.73068485e-02 -7.36994922e-01
6.10031545e-01 8.90299618e-01 -4.12702560e-01 -7.46726632e-01
-6.46107376e-01 -1.07696855e+00 -3.52182359e-01 -3.22525322e-01
6.95421457e-01 -6.28467143e-01 -1.49448788e+00 -3.27275187e-01
-1.20172036e+00 1.46745995e-01 -5.12707472e-01 8.41911957e-02
-2.33841985e-01 4.72738475e-01 -1.21007368e-01 -1.50627255e+00
-1.00122385e-01 -1.03731060e+00 1.24662435e+00 2.90453702e-01
-4.57210839e-01 -3.77251387e-01 1.45369902e-01 5.77810466e-01
1.98291853e-01 -7.86072612e-02 4.50868249e-01 -1.40390110e+00
-8.80683959e-01 -3.90991777e-01 2.22330600e-01 -3.47089708e-01
-3.16269845e-01 -3.96777660e-01 -7.86457658e-01 -6.83601275e-02
-4.63152044e-02 -2.58876145e-01 4.14800614e-01 5.70199639e-02
6.72261119e-01 -7.13604748e-01 -2.84667790e-01 -2.34673813e-01
1.26287317e+00 1.14916854e-01 2.56117016e-01 5.78998625e-01
5.41843921e-02 1.34268856e+00 9.21564937e-01 8.74033093e-01
8.25452864e-01 6.01644039e-01 3.71119380e-01 3.75807524e-01
5.59681475e-01 -1.23503394e-01 -3.79369035e-02 5.71444742e-02
3.96841690e-02 -6.97808504e-01 -8.84750187e-01 1.02030814e+00
-2.09699678e+00 -8.42952132e-01 -4.09850776e-01 2.32530451e+00
5.51293254e-01 -3.40642393e-01 7.54794240e-01 -8.08628462e-03
8.35777640e-01 -1.69374034e-01 -3.80082935e-01 -4.14292812e-01
-1.62482589e-01 -4.57225442e-02 6.50784135e-01 6.58991158e-01
-6.20878279e-01 7.36462533e-01 6.53478765e+00 7.51777828e-01
-2.05240488e-01 2.69724399e-01 4.25374836e-01 -5.88251531e-01
-9.33273554e-01 3.22984219e-01 -8.99074197e-01 4.95763272e-01
4.73492622e-01 -4.58963037e-01 4.67745155e-01 5.61276197e-01
3.27504009e-01 -7.88846433e-01 -9.29506660e-01 6.36959791e-01
-1.45704359e-01 -1.31692386e+00 -1.03698291e-01 2.51188457e-01
1.03338909e+00 -1.83127373e-01 -3.74882132e-01 -4.78650592e-02
1.08932209e+00 -8.48416746e-01 1.05364847e+00 3.96635115e-01
1.98142260e-01 -6.09648585e-01 6.83184445e-01 7.43263543e-01
-9.19634759e-01 -4.55741376e-01 -4.97599423e-01 -1.49015099e-01
3.02938551e-01 5.98055065e-01 -9.09058869e-01 2.08813548e-01
9.10988271e-01 -5.67194283e-01 -5.25286496e-01 1.42896104e+00
-1.33347020e-01 4.32303309e-01 -3.96257967e-01 -1.09889722e+00
3.15895379e-01 -8.18515867e-02 7.51996160e-01 1.18947542e+00
1.02641851e-01 6.79205418e-01 4.11765128e-01 4.63663965e-01
-5.33652352e-03 5.67966700e-01 -2.92628884e-01 6.03598893e-01
8.03803623e-01 1.11455536e+00 -8.27455342e-01 -3.61658782e-01
-8.31201226e-02 5.26850760e-01 2.97454260e-02 6.13227248e-01
-6.23646975e-01 -3.85871142e-01 3.89211327e-01 5.22679329e-01
2.61265934e-01 -3.21431495e-02 -4.11722660e-01 -5.06542325e-01
5.15891612e-01 -9.61449385e-01 6.80639744e-01 -5.82462251e-01
-1.25500679e+00 5.36550701e-01 2.62308836e-01 -1.13383651e+00
-1.91160992e-01 -3.48214895e-01 -4.62365210e-01 1.05212176e+00
-1.41535962e+00 -3.93716425e-01 -1.85545370e-01 5.61183870e-01
3.41722399e-01 1.67081788e-01 3.36824328e-01 -6.24849163e-02
1.06253333e-01 4.63932425e-01 4.02875207e-02 -3.63201857e-01
3.98173958e-01 -1.24322820e+00 2.68921345e-01 5.76968551e-01
2.40329299e-02 6.81632936e-01 7.90307045e-01 -7.72483587e-01
-1.35746789e+00 -6.98680282e-01 1.26315844e+00 -7.59405077e-01
4.70800221e-01 -5.44145882e-01 -8.14882457e-01 2.95977499e-02
2.26943254e-01 -5.53853624e-02 9.78508830e-01 -1.28626868e-01
1.13668047e-01 5.55363558e-02 -1.42157030e+00 7.39441395e-01
9.94955838e-01 -4.26971197e-01 -4.49289352e-01 8.46845388e-01
7.02193022e-01 -2.89239854e-01 -3.54634941e-01 -5.18494956e-02
5.04950404e-01 -1.06621695e+00 5.28567076e-01 -7.49297976e-01
2.88278377e-03 -5.83256960e-01 -3.57469678e-01 -7.18917668e-01
-3.28486949e-01 -1.19721889e+00 2.28272557e-01 1.11040688e+00
7.12990522e-01 -7.52707899e-01 9.41667497e-01 1.46378934e+00
4.14669275e-01 -1.09134030e+00 -1.02292216e+00 -8.31889272e-01
-1.26925454e-01 -3.00465107e-01 1.07674420e+00 3.67962003e-01
-1.42531633e-01 5.16962409e-02 -3.39247473e-02 2.45004162e-01
3.63771945e-01 2.75710315e-01 8.90823543e-01 -1.03054166e+00
-2.55086511e-01 -3.86846244e-01 1.79724097e-01 -1.30195475e+00
-2.82532424e-01 -4.62135166e-01 3.52073699e-01 -1.78426039e+00
3.19010094e-02 -6.30631804e-01 5.59521727e-02 7.64976293e-02
-5.15560150e-01 -6.64225519e-02 4.60582286e-01 3.98499101e-01
-1.35503948e+00 2.72540510e-01 1.04967427e+00 2.75375068e-01
-4.87481833e-01 3.67449641e-01 -1.17197061e+00 4.90536779e-01
8.15155983e-01 -7.73118377e-01 -4.17079180e-01 -4.69199747e-01
1.01932859e+00 1.16931990e-01 3.14863384e-01 -4.81904387e-01
7.75469065e-01 -5.60448289e-01 -9.57320780e-02 -5.82673967e-01
1.17374003e-01 -5.80757320e-01 -6.03498593e-02 1.47296399e-01
-8.41716528e-01 3.06160927e-01 -2.96308011e-01 9.11009610e-01
9.84271541e-02 -7.59688914e-01 2.04856902e-01 -4.93980974e-01
-5.38380563e-01 1.56751707e-01 -4.90359306e-01 4.85006511e-01
1.18271363e+00 -5.85313976e-01 -3.35172504e-01 -5.61795890e-01
-4.35100973e-01 7.76428938e-01 4.47800606e-01 3.98165792e-01
3.95018131e-01 -1.16455793e+00 -9.67581809e-01 -4.87111390e-01
2.79857695e-01 -5.08848689e-02 -4.45344001e-02 6.27770424e-01
-2.87413806e-01 5.13796031e-01 5.78700900e-01 -4.86355990e-01
-1.06538665e+00 4.79346454e-01 1.55756325e-01 -1.63832188e-01
1.35580048e-01 1.00744939e+00 8.36064760e-03 -4.53005642e-01
7.75417089e-02 4.41091135e-02 -2.55861044e-01 2.31166914e-01
5.89564741e-01 9.06134725e-01 1.39625117e-01 -5.51182032e-01
-5.53306043e-01 3.38345528e-01 1.72898635e-01 -6.99722648e-01
8.45622182e-01 -4.18099552e-01 -3.52820605e-01 -1.41003979e-02
8.08827758e-01 6.33338988e-01 -8.80764782e-01 -3.85997057e-01
5.56623101e-01 -8.85159314e-01 -8.04186106e-01 -6.77302003e-01
-4.31515038e-01 2.00260267e-01 2.58753091e-01 9.64256942e-01
8.97748590e-01 2.45796174e-01 6.74648523e-01 7.45529950e-01
6.99468136e-01 -1.50765622e+00 1.05460897e-01 2.47100040e-01
9.05112505e-01 -1.06719899e+00 9.47305337e-02 -3.19636852e-01
-9.37443912e-01 7.17691004e-01 5.90207219e-01 -1.42291963e-01
4.09968317e-01 -2.23659232e-01 -1.22365234e-02 -4.19530183e-01
-8.65090728e-01 -6.99579895e-01 3.56933087e-01 2.90157288e-01
1.50202140e-01 1.37900263e-01 -7.74064422e-01 7.15363622e-01
-3.94600451e-01 2.97578033e-02 4.68573332e-01 1.12432969e+00
-1.00181270e+00 -8.10119510e-01 -1.00466633e+00 7.10911453e-01
-6.30424917e-01 -8.63194466e-03 -9.20085192e-01 6.04009092e-01
3.12681556e-01 1.82290232e+00 -2.26390883e-01 -2.70330697e-01
6.28289402e-01 -3.18628281e-01 -5.48963137e-02 -5.63446820e-01
-1.18414581e+00 -1.76815584e-01 3.42256933e-01 -5.09545088e-01
-5.08288622e-01 -5.01007736e-01 -1.04705691e+00 -2.04222217e-01
-6.87169611e-01 7.49462605e-01 4.29934174e-01 1.05758750e+00
6.96724832e-01 -3.49953055e-01 8.29263330e-01 -3.49851936e-01
-7.89791405e-01 -4.48364615e-01 -4.93477464e-01 4.62462276e-01
2.38622531e-01 -3.54124784e-01 -2.22361967e-01 -3.22687209e-01] | [9.668664932250977, 4.793181419372559] |
6364e0c7-c253-4d82-9d67-c060720eec03 | multi-value-a-framework-for-cross-dialectal | 2212.08011 | null | https://arxiv.org/abs/2212.08011v3 | https://arxiv.org/pdf/2212.08011v3.pdf | Multi-VALUE: A Framework for Cross-Dialectal English NLP | Dialect differences caused by regional, social, and economic factors cause performance discrepancies for many groups of language technology users. Inclusive and equitable language technology must critically be dialect invariant, meaning that performance remains constant over dialectal shifts. Current systems often fall short of this ideal since they are designed and tested on a single dialect: Standard American English (SAE). We introduce a suite of resources for evaluating and achieving English dialect invariance. The resource is called Multi-VALUE, a controllable rule-based translation system spanning 50 English dialects and 189 unique linguistic features. Multi-VALUE maps SAE to synthetic forms of each dialect. First, we use this system to stress tests question answering, machine translation, and semantic parsing. Stress tests reveal significant performance disparities for leading models on non-standard dialects. Second, we use this system as a data augmentation technique to improve the dialect robustness of existing systems. Finally, we partner with native speakers of Chicano and Indian English to release new gold-standard variants of the popular CoQA task. To execute the transformation code, run model checkpoints, and download both synthetic and gold-standard dialectal benchmark datasets, see http://value-nlp.org. | ['Rahul Gupta', 'Jwala Dhamala', 'Diyi Yang', 'Jingfeng Yang', 'William Held', 'Caleb Ziems'] | 2022-12-15 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [-1.33785427e-01 -3.37510139e-01 -2.96739340e-01 -8.31872165e-01
-1.33578396e+00 -1.21199429e+00 5.47321677e-01 -1.01214588e-01
-4.44873899e-01 6.81563258e-01 4.22290653e-01 -8.35063398e-01
8.30784962e-02 -6.89136028e-01 -7.68395782e-01 1.06046677e-01
4.11553591e-01 7.97886908e-01 -1.35455847e-01 -1.00848722e+00
1.56650499e-01 3.40191066e-01 -1.06771660e+00 4.09526438e-01
1.35582066e+00 3.23873609e-01 -7.17595518e-02 6.87033892e-01
-1.33004859e-01 4.49894130e-01 -5.19623637e-01 -8.29256475e-01
4.92107034e-01 -3.59045774e-01 -9.95239854e-01 -5.52078843e-01
7.66538441e-01 -3.31726335e-02 -7.03152567e-02 1.01404893e+00
6.35111570e-01 -2.90226191e-01 3.59343112e-01 -1.32669795e+00
-1.00987899e+00 1.03723848e+00 -1.98730901e-01 8.03951696e-02
5.43755174e-01 3.23426008e-01 1.09303129e+00 -9.56187963e-01
7.65011251e-01 1.32900000e+00 9.22028184e-01 4.25887942e-01
-1.38018560e+00 -7.89884627e-01 -1.10321060e-01 -1.44631103e-01
-1.39253294e+00 -9.31284130e-01 4.77869630e-01 -3.91171873e-01
1.05012035e+00 5.20877481e-01 4.23188239e-01 1.24877691e+00
2.70825326e-01 4.24617350e-01 1.33149433e+00 -6.13639772e-01
-1.76753893e-01 3.87954116e-01 2.43663386e-01 7.46570706e-01
-6.77797496e-02 -1.88331474e-02 -6.48684919e-01 -3.21455806e-01
1.83641076e-01 -6.94228590e-01 -1.44158751e-01 -9.32503119e-02
-1.59891760e+00 5.76439381e-01 -9.96950343e-02 3.31086665e-01
-1.22084744e-01 -3.23711604e-01 4.57320631e-01 1.11899424e+00
3.44640940e-01 9.40265834e-01 -9.77099121e-01 -6.58118308e-01
-3.98510158e-01 5.28281927e-01 8.93466890e-01 1.10419142e+00
6.99831247e-01 -2.33680625e-02 -1.51311219e-01 1.31870925e+00
-2.22802833e-01 8.79773021e-01 6.97742879e-01 -1.09532201e+00
6.67535722e-01 5.47062755e-01 -4.16477630e-03 -5.14494002e-01
-3.65181834e-01 -9.90408137e-02 -3.96729916e-01 -1.38937742e-01
7.24876463e-01 -2.45867223e-01 -5.84124207e-01 2.09957480e+00
1.22001767e-01 -6.90248072e-01 7.99858272e-02 4.84590650e-01
3.64161104e-01 4.27166373e-01 -8.81900340e-02 1.84285894e-01
1.25126648e+00 -7.46604502e-01 -2.06243590e-01 -2.34156862e-01
9.88873959e-01 -1.22269464e+00 2.14202571e+00 2.08765700e-01
-1.24821913e+00 -5.28528750e-01 -7.30967164e-01 -1.71872094e-01
-5.86002469e-01 -2.28110909e-01 5.44904649e-01 1.14878833e+00
-1.23760664e+00 3.71397175e-02 -4.20491159e-01 -5.80662727e-01
-1.45463124e-01 2.78952330e-01 -3.90256196e-01 -1.37130931e-01
-1.40728307e+00 9.58505869e-01 -8.20971653e-02 -4.68012631e-01
-3.91899258e-01 -1.16662645e+00 -8.33469510e-01 -3.07509005e-01
-1.40675724e-01 -5.31186759e-01 1.58677530e+00 -1.16833091e+00
-1.63979626e+00 1.20965278e+00 -9.77288187e-02 6.00645505e-02
5.64573705e-01 -2.35167086e-01 -9.00194526e-01 -5.07885933e-01
4.51744020e-01 3.98945123e-01 3.07763308e-01 -8.94738972e-01
-6.25374734e-01 -2.90841550e-01 -2.25909445e-02 3.22993279e-01
-4.55725193e-01 4.65874016e-01 -3.50182861e-01 -7.15884089e-01
-9.44597051e-02 -1.16017687e+00 1.83773309e-01 -5.45413554e-01
-4.23998296e-01 1.23407215e-01 8.80160034e-02 -9.66094911e-01
1.10900128e+00 -2.15398836e+00 -5.78192472e-02 3.45513165e-01
-1.06027327e-01 -2.04310238e-01 -4.45042670e-01 2.31122196e-01
-2.23385498e-01 3.08719426e-01 -1.25185907e-01 5.16263656e-02
2.95982659e-01 -1.30990312e-01 -2.40703449e-01 2.18422532e-01
7.15082362e-02 1.13320434e+00 -6.24580204e-01 -2.97418714e-01
-1.02028780e-01 -1.09740116e-01 -9.28301513e-01 2.26476230e-02
-1.20811358e-01 3.12909186e-01 8.36334843e-03 7.55808353e-01
5.45843005e-01 3.54485154e-01 1.84144408e-01 9.54501331e-02
-3.79212499e-01 6.18797481e-01 -7.58439839e-01 1.83787620e+00
-8.02521110e-01 4.01773483e-01 2.38860515e-03 -3.95684332e-01
8.69624078e-01 -1.96791694e-01 3.19963664e-01 -1.17117929e+00
-5.62645271e-02 6.41282737e-01 2.59488851e-01 -2.54512042e-01
8.85929465e-01 -3.64070898e-03 -6.29847527e-01 5.61258197e-01
-2.83434540e-01 -5.91965020e-01 1.08745970e-01 5.33380657e-02
9.73625660e-01 -1.44117609e-01 2.17560250e-02 -8.74080300e-01
2.69963384e-01 2.43252218e-01 6.85073018e-01 6.85058296e-01
-2.65713125e-01 5.53429782e-01 3.88245791e-01 -2.63247371e-01
-1.29323053e+00 -1.63252568e+00 -2.94624329e-01 1.77461672e+00
-1.47726163e-01 -3.25218767e-01 -8.67910385e-01 -6.18380547e-01
7.41960108e-02 1.06941354e+00 -3.87533635e-01 -3.29614848e-01
-6.97775781e-01 -7.62815475e-01 1.17467988e+00 2.40086436e-01
5.11276066e-01 -8.03779662e-01 7.12328106e-02 4.01390530e-02
-5.59351385e-01 -9.26285744e-01 -8.89736354e-01 -2.66584963e-01
-2.41806641e-01 -8.67397189e-01 -4.36918706e-01 -9.90069568e-01
7.03344345e-02 9.74278618e-03 1.65192199e+00 -2.42562994e-01
1.80414155e-01 4.91217047e-01 -1.16675161e-01 -4.83290315e-01
-1.05711889e+00 8.08176756e-01 4.48434144e-01 -2.18274653e-01
5.85932016e-01 -2.98321545e-01 -1.41406581e-01 6.02812350e-01
-5.22531927e-01 5.44983521e-02 2.28946477e-01 6.04442179e-01
3.60974878e-01 -6.66785181e-01 7.33331859e-01 -1.15626287e+00
9.70802963e-01 -5.68046749e-01 -5.41475832e-01 6.86883867e-01
-7.86002636e-01 5.89021593e-02 7.64242947e-01 -3.03043485e-01
-1.08133173e+00 -2.99271613e-01 -2.82399595e-01 3.52013916e-01
1.37559017e-02 4.04480904e-01 -4.08122987e-01 2.41935000e-01
1.15520751e+00 7.41199628e-02 -1.10551283e-01 -4.20368910e-01
5.78938961e-01 1.10049260e+00 8.35617542e-01 -1.13433731e+00
6.99178755e-01 -1.47404056e-02 -7.09636092e-01 -7.92639732e-01
-2.24995926e-01 -2.12158822e-02 -3.95269036e-01 3.19124311e-02
5.70627093e-01 -9.82619941e-01 -3.63917232e-01 8.77617180e-01
-7.83183336e-01 -8.37539256e-01 -2.63171941e-01 3.73054713e-01
-5.47930300e-01 -7.52787888e-02 -6.37484133e-01 5.92523962e-02
-3.32490772e-01 -1.14722633e+00 9.97804344e-01 -8.59119892e-02
-6.64909840e-01 -9.70202982e-01 2.57821232e-01 3.91763479e-01
6.61761880e-01 -1.36166468e-01 1.43169665e+00 -6.74417853e-01
-2.59230193e-02 5.29876612e-02 -2.08270345e-02 3.57011348e-01
3.61565113e-01 2.52993464e-01 -6.94967866e-01 -3.75779331e-01
-3.34963888e-01 -4.52973008e-01 1.44251674e-01 1.39501588e-02
8.25702846e-01 -8.84584151e-03 1.29808158e-01 8.67465675e-01
8.09947491e-01 -4.17182930e-02 3.03663284e-01 4.55349356e-01
7.16402113e-01 5.73911726e-01 4.93204385e-01 -1.36989802e-02
1.00544333e+00 7.03668833e-01 -5.03274918e-01 -1.67932853e-01
-3.76426131e-02 -3.18594187e-01 8.87649655e-01 1.41576982e+00
3.93410772e-01 1.09107666e-01 -1.41667438e+00 6.27709627e-01
-1.25738335e+00 -6.01421237e-01 -1.09632626e-01 2.28763437e+00
1.19877636e+00 3.51100653e-01 2.71131188e-01 -2.97461361e-01
6.12194777e-01 -2.56772369e-01 -6.77814186e-01 -8.66810739e-01
-5.21339357e-01 2.16449231e-01 5.66720605e-01 8.56205225e-01
-5.90351045e-01 1.47044265e+00 6.63524008e+00 5.84422112e-01
-1.31320894e+00 1.59896910e-01 8.11392605e-01 -1.25820309e-01
-8.38524997e-01 -7.62654245e-02 -5.55609345e-01 3.18448842e-01
1.25928485e+00 -7.46348143e-01 9.63937163e-01 6.55294418e-01
1.33143693e-01 3.13254952e-01 -1.13948321e+00 8.67705941e-01
-3.08898278e-02 -9.91811275e-01 -6.73861727e-02 -1.28212214e-01
8.51093233e-01 5.86941659e-01 2.70818412e-01 7.16667950e-01
7.89744496e-01 -9.53048468e-01 1.09715581e+00 3.19680661e-01
1.44323003e+00 -8.28117490e-01 1.52605847e-01 1.02320269e-01
-8.91727030e-01 8.28572642e-03 -1.33568332e-01 -2.88345944e-02
-1.80753604e-01 2.35127538e-01 -6.74492240e-01 2.88496494e-01
8.03165257e-01 1.53840527e-01 -1.04183674e+00 2.48353392e-01
2.33359203e-01 8.12568247e-01 -3.56599927e-01 2.64476597e-01
-8.71509239e-02 -2.19215438e-01 3.61764550e-01 1.24545884e+00
2.33504072e-01 -3.85990500e-01 -5.22811674e-02 5.75274050e-01
-4.10303533e-01 4.68783259e-01 -8.27365696e-01 2.96502803e-02
7.83950210e-01 8.76312673e-01 -2.26254210e-01 -3.13933730e-01
-5.81638396e-01 1.05926669e+00 4.81912881e-01 3.53848517e-01
-8.49592566e-01 -5.41373134e-01 1.17164135e+00 7.87108019e-02
-5.13224900e-01 -3.32065105e-01 -6.55271232e-01 -1.35988462e+00
1.83050349e-01 -1.71639156e+00 3.67805690e-01 -5.30231297e-01
-1.30390835e+00 6.26090467e-01 -1.89766899e-01 -7.81284451e-01
-4.77089524e-01 -5.45142829e-01 -2.39026263e-01 1.21245253e+00
-1.05980957e+00 -1.20242667e+00 -1.69576705e-01 7.64482439e-01
5.71999073e-01 -6.64320827e-01 1.00809741e+00 4.61871147e-01
-4.19077843e-01 1.37840116e+00 4.42224413e-01 2.11213440e-01
1.39771140e+00 -1.28645170e+00 1.30741143e+00 7.56584406e-01
1.36138499e-01 8.80655885e-01 3.93967777e-01 -5.48203647e-01
-1.38786912e+00 -1.03286278e+00 1.27656507e+00 -1.13298917e+00
7.56443858e-01 -8.25320423e-01 -6.63258255e-01 9.21021998e-01
1.55128390e-01 -5.36353052e-01 7.17788517e-01 4.87304807e-01
-6.78737700e-01 -3.25409025e-01 -1.06229687e+00 1.07472384e+00
1.23198771e+00 -9.53302979e-01 -5.01270056e-01 1.72175288e-01
1.08132374e+00 -3.46328735e-01 -1.04012775e+00 2.19023377e-01
8.29473436e-01 -8.37589741e-01 7.29198992e-01 -9.29861307e-01
2.71251410e-01 -8.49052705e-03 -6.71392798e-01 -1.72108448e+00
-2.54974127e-01 -8.33582342e-01 8.58291030e-01 1.51257300e+00
9.56695497e-01 -9.25606608e-01 3.71075183e-01 8.48226547e-01
-1.84607044e-01 -3.61870736e-01 -7.32796252e-01 -8.69108260e-01
9.88111556e-01 -6.41212404e-01 1.14624751e+00 1.55766439e+00
-2.73796078e-02 5.65561056e-01 6.71343431e-02 -1.00703686e-01
1.51814401e-01 2.91472599e-02 1.10431468e+00 -8.26719344e-01
-4.14543033e-01 -6.88408792e-01 1.48796380e-01 -6.65839851e-01
3.14262390e-01 -1.31376874e+00 -1.44247994e-01 -8.87106836e-01
-1.00363217e-01 -8.30365121e-01 -6.69723898e-02 3.73419642e-01
-2.62749881e-01 2.03442171e-01 1.34494632e-01 8.83358791e-02
-9.00876522e-02 3.83387148e-01 8.62782598e-01 -1.46158904e-01
-4.85459864e-01 -9.00665224e-02 -1.19326949e+00 5.34168005e-01
1.14044702e+00 -3.10691714e-01 -2.87954956e-01 -8.91959846e-01
4.18617636e-01 -2.48684719e-01 -1.52102083e-01 -7.74997413e-01
-1.04283251e-01 -4.51325208e-01 1.85319278e-02 2.75837660e-01
-2.48452947e-01 -3.94028872e-01 -8.23664889e-02 2.83186704e-01
-5.01255810e-01 1.01606166e+00 2.99316227e-01 -3.90347183e-01
4.80579212e-02 2.87821651e-01 8.23397458e-01 1.01724274e-01
-6.16669893e-01 8.70643407e-02 -3.92156988e-01 7.80349016e-01
5.89025021e-01 8.93255789e-03 -4.98795837e-01 -2.64482766e-01
-7.62666017e-02 2.89268970e-01 8.37983012e-01 1.09164464e+00
-2.71506961e-02 -1.32280982e+00 -1.29058683e+00 5.33231199e-01
5.66869974e-01 -5.03107190e-01 -6.97031394e-02 4.92092192e-01
-8.06517839e-01 2.15955928e-01 -3.90442133e-01 -2.71852285e-01
-9.63306129e-01 2.63837010e-01 6.31565511e-01 6.33989796e-02
-6.72304854e-02 5.99016368e-01 1.81661509e-02 -1.54632699e+00
-3.64540905e-01 -3.81222337e-01 4.43140626e-01 -2.83370495e-01
1.69587255e-01 3.21592540e-01 3.08340728e-01 -7.43325055e-01
-4.24653083e-01 4.30035293e-01 -1.44413501e-01 -6.26539946e-01
8.52524161e-01 -1.85547635e-01 -2.81934500e-01 7.08617032e-01
1.15947759e+00 7.96051979e-01 -5.17100751e-01 -2.40321249e-01
1.04235522e-01 -2.11857975e-01 -5.25453746e-01 -1.17968702e+00
-5.69417119e-01 4.12340939e-01 6.08942807e-01 6.11145943e-02
1.00226319e+00 -1.87936962e-01 6.69862628e-01 4.82534289e-01
4.02009636e-01 -1.28468609e+00 -5.42326927e-01 1.02560937e+00
8.83788824e-01 -1.16998971e+00 -5.71274400e-01 -1.40444994e-01
-7.27232993e-01 5.36807239e-01 8.40179205e-01 4.27800715e-01
6.94374919e-01 2.64877647e-01 7.86510825e-01 1.92400232e-01
-6.50833845e-01 2.00404182e-01 6.95411861e-02 5.42943060e-01
7.35805809e-01 5.06394565e-01 -2.41298720e-01 7.06084251e-01
-1.26401234e+00 -2.48331115e-01 3.20554942e-01 6.46922469e-01
9.23736244e-02 -1.47212529e+00 -4.10712957e-01 5.33132076e-01
-3.84481490e-01 -5.81302404e-01 -6.53458178e-01 9.44525659e-01
-1.84732676e-03 9.39908504e-01 1.36924103e-01 -8.20125103e-01
7.55255580e-01 4.75634634e-01 3.66526395e-01 -4.65981215e-01
-9.47546124e-01 -6.33018017e-01 3.97756308e-01 -5.48645735e-01
5.06677687e-01 -9.29473519e-01 -1.00852048e+00 -9.50649023e-01
2.95800239e-01 1.22026108e-01 6.84449434e-01 6.94654584e-01
5.04747391e-01 1.53943673e-01 7.92213500e-01 -1.54962251e-03
-7.76944935e-01 -9.53501999e-01 -2.98982561e-01 7.37000048e-01
3.39386426e-02 -5.55922091e-02 8.80820602e-02 -7.74722770e-02] | [11.214483261108398, 10.080942153930664] |
75c9a97a-6ba8-42ef-ab4d-5cacab2b3840 | open-vocabulary-object-detection-using-pseudo | 2303.13040 | null | https://arxiv.org/abs/2303.13040v1 | https://arxiv.org/pdf/2303.13040v1.pdf | Open-Vocabulary Object Detection using Pseudo Caption Labels | Recent open-vocabulary detection methods aim to detect novel objects by distilling knowledge from vision-language models (VLMs) trained on a vast amount of image-text pairs. To improve the effectiveness of these methods, researchers have utilized datasets with a large vocabulary that contains a large number of object classes, under the assumption that such data will enable models to extract comprehensive knowledge on the relationships between various objects and better generalize to unseen object classes. In this study, we argue that more fine-grained labels are necessary to extract richer knowledge about novel objects, including object attributes and relationships, in addition to their names. To address this challenge, we propose a simple and effective method named Pseudo Caption Labeling (PCL), which utilizes an image captioning model to generate captions that describe object instances from diverse perspectives. The resulting pseudo caption labels offer dense samples for knowledge distillation. On the LVIS benchmark, our best model trained on the de-duplicated VisualGenome dataset achieves an AP of 34.5 and an APr of 30.6, comparable to the state-of-the-art performance. PCL's simplicity and flexibility are other notable features, as it is a straightforward pre-processing technique that can be used with any image captioning model without imposing any restrictions on model architecture or training process. | ['Byungseok Roh', 'Wooyoung Kang', 'Won Young Jhoo', 'Han-Cheol Cho'] | 2023-03-23 | null | null | null | null | ['open-vocabulary-object-detection'] | ['computer-vision'] | [ 3.69497627e-01 3.03647459e-01 -2.55672365e-01 -4.29225534e-01
-6.50190830e-01 -5.87965906e-01 8.17774951e-01 -8.44224319e-02
-4.82936561e-01 6.69508636e-01 3.64109054e-02 -2.13354845e-02
4.08428609e-01 -5.83857238e-01 -1.01522124e+00 -4.77529377e-01
4.12139654e-01 5.77742159e-01 3.39128405e-01 -6.06280565e-02
2.78945286e-02 3.08979988e-01 -1.72286391e+00 4.26266134e-01
6.41274691e-01 9.94617760e-01 5.01830399e-01 3.99889857e-01
-4.03408706e-01 5.99640489e-01 -4.17072117e-01 -7.07218349e-01
1.75737947e-01 -4.06764299e-02 -7.73938179e-01 2.31638893e-01
8.33447158e-01 -4.31573331e-01 -4.35157746e-01 1.11130285e+00
1.73949882e-01 -1.10881582e-01 8.08134794e-01 -1.37763798e+00
-1.19399500e+00 5.48772454e-01 -5.78467667e-01 1.65708244e-01
1.33447900e-01 3.68128538e-01 1.11027157e+00 -1.11548626e+00
7.33303607e-01 1.16143286e+00 3.15658242e-01 7.84318328e-01
-1.21769011e+00 -8.26124191e-01 4.17471081e-01 4.46972966e-01
-1.68648601e+00 -3.87898356e-01 7.11760819e-01 -5.76993465e-01
6.99462056e-01 1.84819609e-01 4.20758873e-01 1.39774787e+00
-4.21446353e-01 1.13437271e+00 7.93058991e-01 -5.01699030e-01
-4.73203957e-02 6.75751209e-01 9.35508832e-02 5.72229683e-01
5.93569875e-01 -4.93211336e-02 -2.08193019e-01 1.82428956e-01
5.85430861e-01 8.29658657e-02 -3.33303094e-01 -5.85494101e-01
-1.44341516e+00 7.85124719e-01 7.35355616e-01 2.39092782e-01
-2.47195169e-01 6.50607944e-02 2.46909529e-01 -2.07760379e-01
3.55602801e-01 6.76906526e-01 -4.77340043e-01 4.02904302e-01
-7.74335921e-01 1.13950722e-01 5.48661172e-01 1.38692141e+00
9.34344232e-01 -2.13259727e-01 -5.52809834e-01 7.38331795e-01
4.06666756e-01 8.84456933e-01 3.33617061e-01 -6.30794048e-01
4.45280910e-01 6.80908501e-01 1.81963909e-02 -9.14234996e-01
-3.52684483e-02 -6.07611656e-01 -7.47295141e-01 -2.89184809e-01
2.09674329e-01 2.42781401e-01 -1.17475486e+00 1.84388661e+00
2.00310394e-01 2.50712633e-01 2.69427925e-01 8.48736048e-01
1.17725313e+00 6.98124111e-01 2.73384839e-01 -3.89815611e-03
1.69619632e+00 -1.13841498e+00 -5.91932297e-01 -6.29979670e-01
5.61204672e-01 -6.64325356e-01 1.13046813e+00 -5.67153916e-02
-6.32241905e-01 -6.62236691e-01 -9.25153434e-01 -1.56300560e-01
-6.31013095e-01 2.44968072e-01 6.40372872e-01 3.66214007e-01
-9.63793278e-01 -1.27976716e-01 -4.54545766e-01 -5.81980109e-01
8.48935068e-01 3.25714014e-02 -3.42595339e-01 -3.49724382e-01
-9.65603054e-01 9.66775477e-01 9.01997566e-01 -7.72350356e-02
-1.11235893e+00 -6.23399675e-01 -1.12002015e+00 5.66676520e-02
6.32481515e-01 -8.23092997e-01 1.19853044e+00 -1.07804847e+00
-7.95141459e-01 1.19303823e+00 -1.60093948e-01 -4.63534564e-01
2.75455654e-01 -1.07717544e-01 -3.89394820e-01 2.81098753e-01
3.39883327e-01 1.41128922e+00 1.03484726e+00 -1.55982161e+00
-7.16990769e-01 -1.36216193e-01 2.28904411e-01 2.11195081e-01
-5.26148558e-01 -1.49421349e-01 -8.40959549e-01 -5.83681583e-01
-7.73592144e-02 -1.00824368e+00 -3.78081538e-02 1.60070106e-01
-4.37540919e-01 -3.86577725e-01 8.62795472e-01 -4.36336279e-01
8.97962749e-01 -2.07259274e+00 -2.18503147e-01 -3.29271704e-01
3.02617431e-01 6.97478414e-01 -4.49977309e-01 1.64657176e-01
1.51782572e-01 2.69255172e-02 -2.02569485e-01 -3.29175353e-01
-4.53430302e-02 3.21250111e-01 -5.97026050e-01 1.78025756e-02
6.62104130e-01 1.38982522e+00 -1.03128898e+00 -6.60676837e-01
1.75673187e-01 3.92304659e-01 -4.11131680e-01 2.57866651e-01
-6.23550475e-01 3.35822821e-01 -4.94325668e-01 6.68802857e-01
6.53410852e-01 -6.98575258e-01 -1.68095738e-01 -3.53485078e-01
2.83239365e-01 -1.20474286e-01 -8.63222361e-01 1.53852928e+00
-2.98963934e-01 8.95506203e-01 -3.85837585e-01 -8.80172849e-01
9.75409329e-01 1.55536085e-01 1.07851453e-01 -5.31263053e-01
1.32963911e-01 7.53443548e-03 -2.38616437e-01 -7.29193985e-01
3.73402804e-01 -5.70190465e-03 -2.37163734e-02 2.28349581e-01
1.72421932e-01 -1.34401068e-01 3.72766972e-01 3.73614252e-01
7.52873898e-01 -1.70597993e-02 2.69087762e-01 6.89750388e-02
6.54866636e-01 1.60954863e-01 3.58097464e-01 9.23000872e-01
-2.77994812e-01 6.64266407e-01 8.90170131e-03 -3.72233272e-01
-1.14415967e+00 -1.14630377e+00 -3.01297784e-01 9.94630516e-01
3.17722648e-01 -2.83160657e-01 -5.26721716e-01 -8.63324285e-01
-1.05651475e-01 8.17118585e-01 -6.10839367e-01 -2.92554140e-01
-3.42922837e-01 -4.10985678e-01 4.60727692e-01 5.25057316e-01
6.70312881e-01 -1.23616433e+00 -3.85722548e-01 -8.87935162e-02
-4.91463035e-01 -1.71111369e+00 -4.78781790e-01 -1.42763332e-01
-5.37817061e-01 -1.13151801e+00 -8.07524502e-01 -1.26507914e+00
9.89025533e-01 4.88581538e-01 1.22365344e+00 -7.12175667e-02
-4.69650507e-01 5.04319131e-01 -5.50940752e-01 -7.46590853e-01
-5.05497634e-01 7.85867684e-03 -2.48336159e-02 3.18459958e-01
7.35622048e-01 -1.39768869e-01 -4.61306572e-01 1.24038413e-01
-9.87784326e-01 5.54844856e-01 9.11704957e-01 7.64043212e-01
6.76565766e-01 -3.92190397e-01 5.84982932e-01 -6.40774429e-01
1.68782219e-01 -3.61672044e-01 -4.90152270e-01 5.91816366e-01
-4.81468648e-01 2.32039824e-01 4.40361202e-01 -7.78434813e-01
-9.62308288e-01 2.65524060e-01 1.77298665e-01 -6.70788467e-01
-3.69076639e-01 1.71638057e-01 -1.84650704e-01 -7.86217973e-02
5.61988175e-01 6.80841446e-01 -2.24933341e-01 -3.91008466e-01
8.73051226e-01 8.70078504e-01 7.10799932e-01 -4.20745611e-01
8.89674187e-01 6.49472952e-01 -2.66740710e-01 -7.38976419e-01
-1.49821293e+00 -6.99205041e-01 -6.58550441e-01 -2.31229402e-02
1.00464320e+00 -1.23366058e+00 -3.32384586e-01 2.10957900e-01
-1.42202067e+00 7.12193623e-02 -2.52962440e-01 5.07862270e-01
-5.52850544e-01 2.14034915e-01 -2.63442904e-01 -4.88908052e-01
-3.25255513e-01 -1.05750561e+00 1.24484026e+00 3.31337512e-01
9.69509706e-02 -7.41966844e-01 -3.50145102e-01 6.85861647e-01
2.84689605e-01 3.91776934e-02 7.63291061e-01 -8.69382620e-01
-9.63322878e-01 -7.63441324e-02 -7.28709877e-01 6.00819588e-01
6.52596802e-02 -2.94522464e-01 -1.18494213e+00 -2.96141773e-01
-1.58743247e-01 -4.87021387e-01 1.00674021e+00 7.28960559e-02
1.18959653e+00 -3.20538193e-01 -6.86828732e-01 4.99149084e-01
1.29423213e+00 -1.47055034e-02 5.50296307e-01 2.83835322e-01
9.70986307e-01 5.35680771e-01 6.33537829e-01 1.02337599e-01
6.36168718e-01 6.24690771e-01 5.76780677e-01 -1.68064833e-01
-4.22479004e-01 -4.55569893e-01 1.14701994e-01 5.54677665e-01
3.84397477e-01 -2.65102327e-01 -9.95082200e-01 7.82281876e-01
-1.75255096e+00 -7.11882174e-01 5.42137027e-02 1.83745956e+00
1.04841828e+00 1.04643196e-01 -3.36314321e-01 -3.39128703e-01
9.29037690e-01 2.98391208e-02 -6.57433808e-01 1.00667909e-01
-2.10040167e-01 -1.83309332e-01 5.01061201e-01 7.30174929e-02
-1.24554563e+00 1.36520028e+00 5.94424772e+00 7.92849660e-01
-9.53996599e-01 7.91091472e-02 5.12369156e-01 9.48260073e-03
-1.30412951e-01 -3.48676108e-02 -1.22716463e+00 2.98375815e-01
6.82324767e-01 -1.32936940e-01 8.56807083e-02 9.18646753e-01
-2.64169723e-01 1.29366085e-01 -1.37619591e+00 1.32844841e+00
6.07307613e-01 -1.42456162e+00 6.85564995e-01 -1.30695999e-01
8.41652870e-01 1.54408410e-01 6.90121129e-02 5.39338171e-01
6.30206019e-02 -9.13232386e-01 6.45544827e-01 5.43905556e-01
8.51002932e-01 -2.47099116e-01 6.03268087e-01 4.74161327e-01
-1.03903973e+00 2.44960152e-02 -4.56659853e-01 -2.35670302e-02
2.00960323e-01 3.13762665e-01 -1.25978816e+00 2.52787620e-01
6.70398951e-01 7.66190290e-01 -8.84810805e-01 1.10758293e+00
-5.12243986e-01 5.22382796e-01 -1.79809064e-01 -6.03711195e-02
2.55509943e-01 1.00805245e-01 4.65615809e-01 1.15756047e+00
-1.64733678e-02 -4.81583476e-02 1.96260646e-01 1.07824326e+00
-4.13353145e-01 1.75287314e-02 -6.55959904e-01 -1.88186869e-01
4.33676541e-01 1.25754893e+00 -6.30424738e-01 -6.60300612e-01
-7.33354032e-01 8.55318785e-01 4.32106972e-01 4.93898749e-01
-8.64900112e-01 -2.08207890e-01 6.28729641e-01 7.62576936e-03
6.19105637e-01 -7.69775957e-02 -7.60661438e-02 -1.29911041e+00
2.15196699e-01 -6.94508135e-01 3.34870815e-01 -1.07224667e+00
-1.51331997e+00 6.61435962e-01 8.71213451e-02 -1.16385055e+00
-6.06868230e-02 -7.94914603e-01 2.39752904e-02 5.36447883e-01
-1.93555593e+00 -1.60866332e+00 -4.77565825e-01 6.10859990e-01
7.93032944e-01 -1.41390279e-01 6.99515700e-01 2.91134953e-01
-3.42953891e-01 6.10208750e-01 -6.72443286e-02 4.00120080e-01
7.78884768e-01 -8.45121682e-01 2.50381738e-01 8.49455953e-01
6.92354202e-01 4.80388939e-01 6.48062408e-01 -5.77221334e-01
-1.09848273e+00 -1.56501102e+00 9.12605703e-01 -9.85821068e-01
5.09067714e-01 -6.43726349e-01 -1.07206678e+00 8.42733026e-01
3.06944605e-02 3.38068664e-01 4.99514163e-01 -1.05553798e-01
-7.00176835e-01 -1.20093357e-02 -8.24173450e-01 7.30881393e-01
1.16081715e+00 -6.42189682e-01 -1.04070520e+00 5.50592422e-01
1.13701236e+00 -2.46795148e-01 -4.69378620e-01 5.74479222e-01
1.81790620e-01 -4.23034340e-01 1.19388330e+00 -6.65256917e-01
2.92403162e-01 -4.33997691e-01 -1.85234353e-01 -8.70756924e-01
-2.84159690e-01 -1.43546909e-01 -2.65561134e-01 1.31407821e+00
4.34663117e-01 -3.79537225e-01 5.23954272e-01 6.24814153e-01
-3.52032445e-02 -5.40973842e-01 -5.11825085e-01 -9.63902950e-01
-1.70720398e-01 -5.15680134e-01 5.49141407e-01 9.55913365e-01
-4.38592166e-01 7.34767377e-01 -3.55023682e-01 3.98013622e-01
5.46768188e-01 2.26378888e-01 8.06741238e-01 -1.28297675e+00
4.79109176e-02 -3.13657343e-01 -5.05282283e-01 -1.18872309e+00
4.41665947e-01 -1.13262808e+00 6.58502504e-02 -1.70789170e+00
6.30520284e-01 -4.69546944e-01 -3.23681891e-01 7.36879170e-01
-3.36049408e-01 6.35896742e-01 3.97565573e-01 4.39766347e-01
-1.03992534e+00 7.72124350e-01 1.52842808e+00 -5.67911685e-01
-3.65371294e-02 -2.58604378e-01 -9.05150533e-01 7.70129740e-01
4.88864720e-01 -5.63085973e-01 -4.86898899e-01 -6.42365158e-01
4.17945441e-03 -4.46317554e-01 6.78307176e-01 -8.69020581e-01
1.87205762e-01 -1.75351188e-01 3.08463752e-01 -6.00384057e-01
4.01695102e-01 -8.95213008e-01 -1.79516912e-01 2.73809582e-01
-4.58390146e-01 -4.90993798e-01 2.76296586e-01 8.34739089e-01
-3.00702095e-01 -2.52331614e-01 6.46229029e-01 -1.27942309e-01
-1.22857237e+00 5.46330154e-01 8.28183442e-02 1.80285379e-01
1.42834520e+00 -2.32334480e-01 -4.16139007e-01 -1.48720399e-01
-6.53366864e-01 4.30861831e-01 4.28449422e-01 1.01796854e+00
6.38671279e-01 -1.39166963e+00 -9.01208937e-01 2.59526372e-01
9.53486741e-01 2.43372440e-01 5.14173061e-02 4.36465830e-01
-2.11384282e-01 7.64642656e-01 -8.56338069e-02 -9.45887566e-01
-1.12719977e+00 1.02388346e+00 6.70970455e-02 4.30088192e-02
-6.69548988e-01 9.68421817e-01 7.48966336e-01 -2.26038799e-01
2.42316723e-01 -3.28743309e-01 -3.20227057e-01 1.44283306e-02
7.27068603e-01 -2.54576057e-01 -2.73862243e-01 -8.51124465e-01
-3.77285779e-01 5.42235970e-01 -3.41886550e-01 3.31742465e-01
1.07417703e+00 -4.25456256e-01 7.15939626e-02 2.84486413e-01
1.16396046e+00 -4.78544503e-01 -1.48263001e+00 -8.26335371e-01
4.40166257e-02 -4.48663533e-01 -4.08482663e-02 -9.04311657e-01
-8.87400091e-01 9.51095462e-01 5.63949645e-01 -1.55519351e-01
9.69645798e-01 6.02084756e-01 7.49421418e-01 5.45466721e-01
3.83649915e-01 -8.67570877e-01 3.82139981e-01 5.83774030e-01
9.68370557e-01 -1.80349684e+00 -3.12261462e-01 -5.51044524e-01
-7.91456223e-01 8.68968785e-01 8.63278389e-01 3.21467191e-01
3.16414267e-01 -2.36049369e-01 7.76449740e-02 -1.23660527e-01
-6.64711118e-01 -6.13462448e-01 6.87001884e-01 8.21481407e-01
-4.76778001e-02 2.34270785e-02 5.70363700e-02 6.09308720e-01
1.07679248e-01 1.87417045e-02 3.97919655e-01 6.25003695e-01
-4.90379453e-01 -8.02955508e-01 -2.31423274e-01 4.18248981e-01
-1.45951971e-01 -3.45187634e-01 -1.63447589e-01 6.37073934e-01
2.59497017e-01 8.66785228e-01 8.44512805e-02 -9.04005244e-02
1.71665937e-01 1.30148247e-01 3.53871018e-01 -9.75026846e-01
8.65583420e-02 -4.14106637e-01 -1.76204041e-01 -3.00088972e-01
-5.07927835e-01 -3.67695004e-01 -1.12605631e+00 3.27566236e-01
-4.92188692e-01 -1.18014783e-01 6.86183095e-01 1.13226891e+00
4.82822359e-01 4.99754310e-01 2.74905860e-01 -6.04790032e-01
-4.04445648e-01 -9.08081293e-01 -3.28791887e-01 8.11828017e-01
3.78611296e-01 -7.50411212e-01 -1.67777956e-01 4.72818464e-01] | [10.080842018127441, 1.5968714952468872] |
3a5ec7c8-a3e2-4a8e-9750-03c0620a4942 | self-configuring-nnu-nets-detect-clouds-in | 2210.13659 | null | https://arxiv.org/abs/2210.13659v1 | https://arxiv.org/pdf/2210.13659v1.pdf | Self-Configuring nnU-Nets Detect Clouds in Satellite Images | Cloud detection is a pivotal satellite image pre-processing step that can be performed both on the ground and on board a satellite to tag useful images. In the latter case, it can help to reduce the amount of data to downlink by pruning the cloudy areas, or to make a satellite more autonomous through data-driven acquisition re-scheduling of the cloudy areas. We approach this important task with nnU-Nets, a self-reconfigurable framework able to perform meta-learning of a segmentation network over various datasets. Our experiments, performed over Sentinel-2 and Landsat-8 multispectral images revealed that nnU-Nets deliver state-of-the-art cloud segmentation performance without any manual design. Our approach was ranked within the top 7% best solutions (across 847 participating teams) in the On Cloud N: Cloud Cover Detection Challenge, where we reached the Jaccard index of 0.882 over more than 10k unseen Sentinel-2 image patches (the winners obtained 0.897, whereas the baseline U-Net with the ResNet-34 backbone used as an encoder: 0.817, and the classic Sentinel-2 image thresholding: 0.652). | ['Jakub Nalepa', 'Bertrand Le Saux', 'Nicolas Longépé', 'Michal Kawulok', 'Maciej Ziaja', 'Bartosz Grabowski'] | 2022-10-24 | null | null | null | null | ['cloud-detection'] | ['computer-vision'] | [ 3.70132953e-01 1.21247554e-02 -1.13809772e-01 -2.32513011e-01
-7.01284170e-01 -7.98699021e-01 2.22862706e-01 9.56264958e-02
-7.11248159e-01 4.73618954e-01 -3.31763923e-01 -5.05988955e-01
-2.62535930e-01 -1.03981411e+00 -7.98306227e-01 -8.45355153e-01
-5.28704584e-01 4.78419453e-01 3.11199337e-01 -2.85309553e-01
3.37374285e-02 6.43827558e-01 -1.76606655e+00 2.72106260e-01
6.39956534e-01 1.32803845e+00 4.00009811e-01 7.93025255e-01
-9.27036181e-02 6.52925611e-01 -4.82045054e-01 3.33664306e-02
7.11481512e-01 1.24675766e-01 -8.38240087e-01 2.31374558e-02
7.32553959e-01 -2.46889457e-01 -5.33976108e-02 1.33482933e+00
4.24843132e-01 -2.20157191e-01 8.33932236e-02 -1.03996861e+00
1.97900310e-02 5.96043348e-01 -5.53937852e-01 5.82954705e-01
-5.40406346e-01 1.01255931e-01 1.06766772e+00 -4.46369857e-01
8.06410611e-01 7.99427569e-01 1.03728127e+00 2.83138335e-01
-1.13095891e+00 -6.60997212e-01 3.80136725e-03 1.41715303e-01
-1.51398730e+00 -8.90625194e-02 -7.93009102e-02 -5.23092568e-01
9.46081758e-01 5.83864152e-01 9.65217829e-01 1.30844176e-01
3.52482341e-04 5.48248351e-01 1.29148793e+00 -1.91471934e-01
3.98745626e-01 -3.08578964e-02 6.32602572e-02 3.67565811e-01
3.56687933e-01 -8.55709910e-02 -1.48371652e-01 -7.17283264e-02
5.68782270e-01 1.03958569e-01 -2.47817799e-01 9.30228755e-02
-9.43058968e-01 7.91657507e-01 9.14763391e-01 3.58790427e-01
-6.38243258e-01 2.39665419e-01 3.42235684e-01 5.57062685e-01
8.71935129e-01 6.99298084e-01 -6.37225926e-01 2.30398715e-01
-1.50903475e+00 1.87162966e-01 4.68837738e-01 8.85095537e-01
1.13616168e+00 5.58263101e-02 -1.14768177e-01 3.68376553e-01
-2.25739807e-01 5.80062747e-01 1.22658104e-01 -8.68546784e-01
2.19909087e-01 5.29441833e-01 1.87432632e-01 -5.72068274e-01
-6.26331627e-01 -9.49126422e-01 -1.04808056e+00 3.54240716e-01
-2.05672055e-01 -4.19628829e-01 -1.29823256e+00 1.07750785e+00
2.51048833e-01 2.00606853e-01 -1.30175203e-01 1.05805898e+00
6.38073027e-01 6.07803285e-01 -9.14041921e-02 7.50431791e-02
1.37360835e+00 -9.33713198e-01 -1.78928778e-01 -4.79278922e-01
4.35686737e-01 -4.96038854e-01 5.53029418e-01 3.07692587e-01
-5.29438198e-01 -4.08294499e-01 -9.70878243e-01 4.81739283e-01
-7.33918846e-01 -1.10047422e-02 6.82647526e-01 4.48320448e-01
-1.45457447e+00 8.17201853e-01 -5.79868019e-01 -3.53749543e-01
5.90937972e-01 4.08015698e-01 -1.66291445e-01 -3.00280973e-02
-1.06971121e+00 7.84161508e-01 6.19056463e-01 3.67840111e-01
-9.08326745e-01 -7.37137556e-01 -2.63108164e-01 2.20858127e-01
5.68099856e-01 -4.29035664e-01 9.50955033e-01 -1.41344893e+00
-5.43045580e-01 1.08351564e+00 4.09985602e-01 -1.00513542e+00
4.47688669e-01 -9.83895510e-02 -2.82095432e-01 2.88790017e-01
1.91319972e-01 9.92257178e-01 8.20758879e-01 -1.25717711e+00
-1.26569533e+00 -3.75488669e-01 2.89763901e-02 1.11946106e-01
-1.67935610e-01 2.99963485e-02 -2.47508183e-01 -5.03395796e-01
2.84923613e-01 -1.11628687e+00 -4.57793117e-01 -1.95106417e-01
-1.72770008e-01 6.57816827e-02 8.85450900e-01 -8.10209036e-01
8.94777894e-01 -1.96962059e+00 -1.20082825e-01 4.06376183e-01
3.03430438e-01 6.99701667e-01 -3.89367908e-01 2.18535706e-01
-2.41367698e-01 4.05017793e-01 -4.17546958e-01 -8.26557502e-02
-3.25175703e-01 2.90484697e-01 -3.23850155e-01 4.62815374e-01
2.13562295e-01 6.14653409e-01 -9.81412947e-01 -3.10446590e-01
9.72292293e-03 2.93910354e-01 -1.82010040e-01 1.39383689e-01
-5.36905885e-01 8.35562050e-02 -1.01725705e-01 9.80884969e-01
1.01264632e+00 -9.26795155e-02 2.13043347e-01 8.69498178e-02
-5.03240049e-01 -1.30627677e-01 -1.11168933e+00 1.42373562e+00
-4.01939958e-01 1.06258333e+00 6.93815053e-01 -7.46701419e-01
6.96399510e-01 8.85152221e-02 7.58815646e-01 -7.09700286e-01
-4.41427715e-02 2.67122000e-01 -1.49699748e-01 -3.97025824e-01
9.34261858e-01 8.34552720e-02 1.87913895e-01 2.59349585e-01
-1.92694902e-01 -2.28870720e-01 1.95769086e-01 7.49673042e-03
1.08551955e+00 -3.48437056e-02 -1.89824000e-01 -4.94782358e-01
2.29824215e-01 5.50760806e-01 5.32261908e-01 9.90227222e-01
-1.58158854e-01 6.63415790e-01 4.33972657e-01 -1.00697231e+00
-9.75151181e-01 -4.50254619e-01 -1.76573887e-01 1.21291685e+00
-8.63572583e-02 -2.17304140e-01 -1.08230519e+00 -5.89104295e-01
1.27598256e-01 4.26921695e-01 -7.06575871e-01 5.14324069e-01
-1.60540968e-01 -1.04972816e+00 6.04691148e-01 1.64695024e-01
6.63372219e-01 -1.07196844e+00 -8.66948962e-01 2.48126149e-01
-2.48461321e-01 -1.09324992e+00 1.63472101e-01 6.39382005e-01
-1.00412285e+00 -1.22550106e+00 -4.59657431e-01 -4.67021167e-01
6.08609438e-01 8.60571563e-01 1.39945102e+00 3.84717107e-01
-3.21794033e-01 3.64589132e-02 -5.02817273e-01 -4.46364909e-01
-1.19752631e-01 4.99775648e-01 -1.63403511e-01 -1.10715106e-01
3.08495760e-01 -4.02885199e-01 -5.79054415e-01 1.06945932e-01
-1.28416491e+00 -1.74888611e-01 6.28058195e-01 7.01139867e-01
8.82609367e-01 4.32334810e-01 2.59248354e-02 -7.53575742e-01
4.40500900e-02 -4.51781780e-01 -1.15137291e+00 2.58717269e-01
-6.89367712e-01 -3.81708801e-01 3.22032064e-01 3.21815699e-01
-2.54042178e-01 4.53664839e-01 -2.13375986e-02 -5.28552413e-01
-4.53902446e-02 5.13144076e-01 2.84683228e-01 -5.70953429e-01
5.45487761e-01 3.01262289e-01 -1.37602061e-01 -4.54012573e-01
1.07795104e-01 1.06754196e+00 5.21157026e-01 1.30748767e-02
7.97036052e-01 6.92164063e-01 1.30854044e-02 -1.02830160e+00
-1.10039032e+00 -8.81221950e-01 -7.55429506e-01 -2.96298981e-01
9.30818260e-01 -1.36600828e+00 -4.84769881e-01 5.01003325e-01
-1.08048558e+00 -4.37945127e-01 -2.51231402e-01 -8.44637305e-02
-3.26278843e-02 8.27680156e-02 -9.65677202e-02 -7.49532580e-01
-8.84459674e-01 -1.02207959e+00 1.12642717e+00 2.02830240e-01
4.37578619e-01 -5.36411285e-01 -1.05976008e-01 4.28064734e-01
8.09455514e-01 3.45773220e-01 4.12031710e-01 -4.84982997e-01
-9.23223019e-01 -1.94616809e-01 -6.49609804e-01 7.73645759e-01
-7.57943466e-02 9.75288153e-02 -1.21917939e+00 -6.27025545e-01
-2.69603163e-01 -1.47353068e-01 1.38439465e+00 4.20836836e-01
1.48543549e+00 -1.37671918e-01 3.83248134e-03 9.14788902e-01
1.89188123e+00 -3.24648589e-01 6.91801965e-01 7.77698576e-01
5.86662054e-01 4.37338024e-01 6.16328239e-01 3.61099064e-01
3.16650957e-01 3.16499799e-01 1.38724577e+00 -3.71694088e-01
1.28770024e-01 3.66935611e-01 -4.17502709e-02 4.00421649e-01
-3.88186634e-01 -4.93614599e-02 -1.23955619e+00 6.50091588e-01
-1.71017826e+00 -9.76304948e-01 -4.12659764e-01 2.19562411e+00
4.05675799e-01 2.05262318e-01 -4.59684208e-02 -2.61585981e-01
8.00011396e-01 3.65372866e-01 -4.31481004e-01 -2.23947652e-02
-4.48828638e-01 2.34071881e-01 1.44416785e+00 4.11765218e-01
-1.64995861e+00 1.16450584e+00 6.43744802e+00 8.38642120e-01
-1.34308469e+00 3.84488076e-01 6.71664119e-01 -1.99802667e-01
-7.44255632e-02 7.11455047e-02 -7.47524083e-01 5.55280507e-01
9.01417494e-01 4.02133137e-01 5.27875245e-01 8.70292783e-01
1.00914128e-01 -4.85229135e-01 -2.74328321e-01 8.56786907e-01
-1.35718763e-01 -1.79391623e+00 -1.43964022e-01 2.32047036e-01
9.96069789e-01 1.15511107e+00 -3.49349976e-01 8.92022029e-02
3.12526762e-01 -9.68303144e-01 7.58741736e-01 5.58701277e-01
8.60331774e-01 -6.14791751e-01 1.11041367e+00 3.32842588e-01
-1.04769385e+00 -1.75439060e-01 -6.91777945e-01 1.76498387e-02
-3.14758331e-01 9.20282423e-01 -8.97565901e-01 5.89895964e-01
1.20431209e+00 3.83800387e-01 -6.18020475e-01 1.28787458e+00
1.70409773e-02 7.20345438e-01 -5.96073568e-01 2.53979594e-01
7.62837648e-01 -1.35711819e-01 5.65224826e-01 1.35084260e+00
1.97290421e-01 -4.20477567e-03 2.17070162e-01 3.76521498e-01
-2.33924136e-01 -1.44620836e-01 -3.86724949e-01 2.85202891e-01
3.76608461e-01 1.59653163e+00 -9.89986897e-01 -3.93380284e-01
-4.75104107e-03 6.78197265e-01 -8.29210356e-02 7.68829584e-02
-6.52544618e-01 -3.00542414e-01 7.85171747e-01 1.49885073e-01
6.08592808e-01 -9.95924920e-02 -2.08385170e-01 -9.72928286e-01
-9.21165869e-02 -7.96889782e-01 2.32573956e-01 -9.25673187e-01
-8.13964248e-01 9.80612099e-01 -3.24100554e-01 -1.41065121e+00
1.44887045e-01 -6.82184160e-01 -4.71157521e-01 9.16009247e-01
-2.07081819e+00 -1.09317255e+00 -7.36303806e-01 7.87185729e-02
1.90988123e-01 -1.52787730e-01 6.87260747e-01 3.45543981e-01
-3.10625672e-01 -5.58089353e-02 3.43583703e-01 3.45260352e-02
4.14369613e-01 -1.29858673e+00 4.98842150e-01 1.11076558e+00
-7.19936118e-02 4.08968180e-02 4.98717338e-01 -6.60586715e-01
-1.28748262e+00 -1.60323894e+00 7.60171652e-01 -2.83799693e-02
6.81077123e-01 -1.77566424e-01 -7.83437073e-01 5.03091574e-01
4.00664210e-01 3.52996707e-01 5.53157866e-01 -3.21415737e-02
-1.54821128e-01 -5.71125448e-01 -1.10332882e+00 -7.90624842e-02
7.34568357e-01 -4.02267277e-01 -2.15755664e-02 8.34994197e-01
6.48474038e-01 -5.26619196e-01 -8.45239758e-01 2.40391746e-01
3.57802659e-01 -1.20911467e+00 6.79469347e-01 -6.03177786e-01
3.74963075e-01 -3.96769911e-01 -3.91697615e-01 -1.27096760e+00
-5.11178255e-01 -5.44729054e-01 4.01127487e-01 7.53408730e-01
2.85149187e-01 -3.19118440e-01 9.66831923e-01 -2.97451429e-02
-2.80651480e-01 -4.15131629e-01 -1.12879574e+00 -8.39076936e-01
-2.04356924e-01 -4.00532365e-01 8.71302843e-01 1.00239623e+00
-1.03977668e+00 -1.68817371e-01 -2.37385258e-01 5.05427182e-01
7.46138930e-01 3.48441273e-01 7.56051540e-01 -1.51269388e+00
4.57327180e-02 -4.80853558e-01 -3.93018633e-01 -5.32632709e-01
-8.63754004e-02 -9.14131343e-01 7.45919496e-02 -1.51754487e+00
2.12911591e-02 -7.58455396e-01 -3.22975725e-01 8.24659646e-01
6.49459213e-02 4.79941815e-01 4.68331099e-01 5.11362553e-01
-6.70734406e-01 1.19558178e-01 8.77764523e-01 -4.13605779e-01
-1.80538595e-02 2.75864005e-01 -4.47351515e-01 6.16938591e-01
8.70395362e-01 -7.30796456e-01 2.44578302e-01 -8.39565694e-01
6.48611486e-01 -1.13381043e-01 6.75253212e-01 -1.22106314e+00
3.07227910e-01 -5.00707664e-02 1.85804829e-01 -1.01005602e+00
-8.22184682e-02 -1.13890398e+00 1.36056304e-01 6.55813277e-01
1.81213677e-01 5.40304556e-02 3.29018444e-01 2.46774077e-01
-3.23829710e-01 -3.76740307e-01 8.00483465e-01 -4.50438827e-01
-1.16571271e+00 4.10738796e-01 -3.38996083e-01 -2.26823866e-01
8.38401198e-01 -1.95976347e-01 -3.04348916e-01 -1.33201048e-01
-7.55428791e-01 5.86611271e-01 4.67944235e-01 2.26260439e-01
5.69259599e-02 -7.10479379e-01 -8.58727515e-01 1.58190355e-01
-1.26454383e-02 3.43355417e-01 3.56524050e-01 6.31631017e-01
-1.08383822e+00 2.79214621e-01 -3.35240543e-01 -9.60875273e-01
-1.22114921e+00 -1.07576601e-01 8.58675718e-01 -3.03131521e-01
-4.08483386e-01 9.21370327e-01 -6.47961855e-01 -6.33975148e-01
1.32482335e-01 -2.58922309e-01 -1.68799926e-02 7.14713275e-01
4.94331509e-01 3.65464926e-01 7.93665051e-01 -4.10573035e-01
-4.46712643e-01 2.88259327e-01 2.13230863e-01 1.45779312e-01
1.93002224e+00 7.87712485e-02 -7.32717156e-01 -8.86301473e-02
7.97261596e-01 -3.88224572e-01 -1.28799355e+00 -3.27413708e-01
-5.09881787e-03 -4.06527787e-01 7.12321401e-01 -9.09516931e-01
-1.63420784e+00 6.08116984e-01 1.03235877e+00 6.28681540e-01
1.32406271e+00 -3.85321319e-01 4.83530909e-01 6.37311280e-01
4.52168643e-01 -1.39980602e+00 -6.98535860e-01 7.48065650e-01
6.17698967e-01 -1.42487168e+00 4.06945765e-01 -1.24468520e-01
-3.00757170e-01 1.09971583e+00 4.46636826e-01 -4.60521020e-02
6.26753509e-01 1.88898474e-01 2.35816270e-01 -3.89109254e-01
-5.80205560e-01 -6.71218991e-01 -2.05355510e-01 4.55305338e-01
-4.18503274e-04 4.49485838e-01 2.23574564e-01 1.25583215e-02
-1.46673828e-01 -2.24336367e-02 6.80470824e-01 1.04212606e+00
-8.03731501e-01 -6.80612385e-01 -6.56535745e-01 8.92892957e-01
-4.44129765e-01 -3.79883885e-01 -3.30200285e-01 6.74443841e-01
5.09783387e-01 6.55849993e-01 4.59213018e-01 -5.26047349e-01
1.78685337e-01 -2.28495210e-01 -5.78708872e-02 -5.64864874e-01
-1.17098379e+00 2.74688136e-02 -5.18928133e-02 -6.41970396e-01
-7.77383626e-01 -6.82835340e-01 -7.08422184e-01 -4.02892321e-01
-3.54273349e-01 1.80730104e-01 1.12790191e+00 6.47682548e-01
5.89208663e-01 4.99057472e-01 8.65381718e-01 -1.07370853e+00
-4.66278702e-01 -1.05182791e+00 -7.91986823e-01 -2.36711681e-01
6.33457541e-01 -7.71678016e-02 -3.67010236e-01 -1.83630362e-01] | [9.621465682983398, -1.5378224849700928] |
15addbe8-bddb-444f-9357-996e7703d49f | active-learning-framework-to-automate | 2211.08399 | null | https://arxiv.org/abs/2211.08399v1 | https://arxiv.org/pdf/2211.08399v1.pdf | Active Learning Framework to Automate NetworkTraffic Classification | Recent network traffic classification methods benefitfrom machine learning (ML) technology. However, there aremany challenges due to use of ML, such as: lack of high-qualityannotated datasets, data-drifts and other effects causing aging ofdatasets and ML models, high volumes of network traffic etc. Thispaper argues that it is necessary to augment traditional workflowsof ML training&deployment and adapt Active Learning concepton network traffic analysis. The paper presents a novel ActiveLearning Framework (ALF) to address this topic. ALF providesprepared software components that can be used to deploy an activelearning loop and maintain an ALF instance that continuouslyevolves a dataset and ML model automatically. The resultingsolution is deployable for IP flow-based analysis of high-speed(100 Gb/s) networks, and also supports research experiments ondifferent strategies and methods for annotation, evaluation, datasetoptimization, etc. Finally, the paper lists some research challengesthat emerge from the first experiments with ALF in practice. | ['Tomáš Čejka', 'Dominik Soukup', 'Jaroslav Pešek'] | 2022-10-26 | null | null | null | null | ['traffic-classification'] | ['miscellaneous'] | [-2.90443867e-01 -5.71712628e-02 -7.77822196e-01 -5.68192065e-01
-4.40127850e-01 -5.39962292e-01 4.36360836e-01 2.10242689e-01
-3.49346489e-01 1.25020742e+00 -6.52906477e-01 -6.86221600e-01
-4.54390526e-01 -7.81187236e-01 -4.06996727e-01 -7.04479635e-01
-3.41207802e-01 9.87914443e-01 9.02869403e-01 3.02640468e-01
3.62814933e-01 1.01356590e+00 -1.69811392e+00 4.53624904e-01
6.95679069e-01 8.42262447e-01 -7.11686611e-02 8.97071958e-01
-7.85659134e-01 1.21055460e+00 -1.14309132e+00 -2.07759246e-01
2.14967519e-01 1.19571418e-01 -1.08764458e+00 1.34513542e-01
3.92093480e-01 -7.30144605e-02 8.16446170e-03 3.66461188e-01
4.41030145e-01 -1.63067341e-01 3.31252441e-02 -2.20220447e+00
3.28951418e-01 6.54340506e-01 -3.29213709e-01 9.53269839e-01
-3.00799668e-01 3.74508232e-01 4.20256913e-01 -5.71922958e-01
5.99441588e-01 9.76693153e-01 4.98100400e-01 2.19171360e-01
-1.39181530e+00 -1.28657305e+00 1.50346637e-01 7.35920668e-01
-8.15914392e-01 -1.27460814e+00 5.42882621e-01 -4.12093371e-01
9.38297629e-01 6.90550566e-01 7.08537638e-01 1.14416575e+00
-1.16706692e-01 8.21715415e-01 8.63424301e-01 -6.28764033e-01
4.12240863e-01 7.86214113e-01 5.10593593e-01 5.11172950e-01
5.13650596e-01 -1.56885505e-01 -4.32115495e-01 -3.00119221e-01
6.13441050e-01 -2.55837858e-01 2.51019835e-01 -4.26882178e-01
-6.69810593e-01 2.91562825e-01 -3.01220447e-01 2.32117087e-01
-4.32131559e-01 2.72765607e-02 6.59277201e-01 8.52787673e-01
4.68652666e-01 2.79200256e-01 -7.60110080e-01 -6.50755048e-01
-1.07086492e+00 -1.66972980e-01 1.12005639e+00 1.11199248e+00
1.22849870e+00 4.05522645e-01 2.04110637e-01 6.38065398e-01
5.42040944e-01 3.66501153e-01 -4.03543413e-02 -1.06016982e+00
3.98227304e-01 8.14107478e-01 -2.82277972e-01 -6.15987480e-01
-2.99238622e-01 -5.93334287e-02 -1.49123296e-01 5.65824330e-01
1.75656721e-01 -4.95704383e-01 -5.37450135e-01 9.09552455e-01
3.41158003e-01 6.63262725e-01 -4.54833984e-01 2.03217670e-01
8.11955273e-01 4.88001943e-01 3.35603416e-01 -8.52288008e-01
4.32832897e-01 -7.82548845e-01 -9.29301202e-01 2.54049689e-01
9.17966068e-01 -7.53884017e-01 7.92239904e-01 5.97743630e-01
-9.74217355e-01 -4.38547134e-01 -7.37862051e-01 6.64608300e-01
-6.11378610e-01 -5.97850919e-01 8.24957550e-01 1.13494468e+00
-1.05923581e+00 2.78002948e-01 -8.84465814e-01 -7.53792226e-01
8.88873279e-01 5.99615633e-01 -1.11271314e-01 4.49643612e-01
-6.06499255e-01 5.38749814e-01 5.96788704e-01 -4.05242771e-01
-7.85509884e-01 -1.08968270e+00 -1.04811937e-02 -2.80899316e-01
6.71926796e-01 -3.86789963e-02 1.02278662e+00 -1.08517170e+00
-1.29822242e+00 6.89222395e-01 3.85271519e-01 -4.09209758e-01
2.80792028e-01 -9.11077708e-02 -1.04762256e+00 -3.52887227e-03
-2.99920738e-01 1.11531809e-01 6.08657300e-01 -1.25257969e+00
-6.50843501e-01 -3.22790384e-01 7.72127882e-02 -5.09386063e-01
-5.39751470e-01 3.33691061e-01 -2.66660780e-01 -3.70596290e-01
-3.75917912e-01 -4.57302153e-01 8.99472833e-02 -3.83113436e-02
-3.11018825e-01 -2.91926473e-01 1.88145816e+00 -6.19462915e-02
1.68784535e+00 -1.77473414e+00 -6.07271314e-01 4.67632860e-01
2.82841951e-01 9.80012298e-01 -9.51104537e-02 5.38695395e-01
-2.75217682e-01 4.88581747e-01 4.70691592e-01 -3.71643491e-02
-6.71331212e-02 4.87696886e-01 -2.05412388e-01 4.91492338e-02
9.79098827e-02 5.48347652e-01 -8.04586589e-01 -9.52189684e-01
5.67211449e-01 -8.97485763e-02 -2.98285216e-01 4.36105222e-01
-3.56925070e-01 3.03546131e-01 -1.86228052e-01 9.68573749e-01
8.56304944e-01 -2.54255384e-01 2.83096522e-01 -9.79366452e-02
-4.51006263e-01 1.36935726e-01 -1.33533323e+00 1.13237929e+00
-4.52758133e-01 8.96201551e-01 2.16371477e-01 -1.33014166e+00
9.44087029e-01 5.47781527e-01 1.08668375e+00 -5.82350254e-01
9.17080194e-02 3.46780837e-01 -2.18418509e-01 -1.13255966e+00
-8.47797543e-02 5.90693533e-01 6.53387547e-01 9.44562972e-01
3.58816594e-01 5.09386122e-01 6.97336495e-01 8.15912783e-02
1.41737592e+00 -5.43111041e-02 -3.22810039e-02 -1.99544474e-01
7.73592472e-01 1.15859449e-01 5.38142562e-01 7.94671834e-01
-6.51599884e-01 -4.72320497e-01 5.83924890e-01 -7.24933684e-01
-9.73042190e-01 -1.01337147e+00 -2.21105665e-01 1.40621758e+00
-4.28447634e-01 -4.07975018e-01 -3.39192390e-01 -1.19891167e+00
-2.07810372e-01 4.92305785e-01 7.78365461e-03 2.57420361e-01
-7.99706757e-01 -1.11727798e+00 6.72764182e-01 1.59693047e-01
6.94078088e-01 -9.27630901e-01 -3.70723099e-01 2.49607548e-01
1.24177471e-01 -1.10504222e+00 5.53712130e-01 1.40029415e-01
-1.10112906e+00 -1.46821868e+00 5.01130521e-01 -6.12596154e-01
1.89758584e-01 9.69333167e-04 1.31187093e+00 3.60266775e-01
-6.88836992e-01 3.78921002e-01 -4.26092267e-01 -8.21747541e-01
-4.87266392e-01 4.19677287e-01 -2.32207552e-01 -1.69358522e-01
7.37108350e-01 -9.05855358e-01 -4.81760502e-02 2.58127868e-01
-5.86883605e-01 -4.55255598e-01 4.62228835e-01 2.11000249e-01
1.75857529e-01 4.13786769e-01 8.02084863e-01 -1.37987268e+00
4.02852118e-01 -8.52270782e-01 -9.26263154e-01 4.83551979e-01
-9.95323360e-01 -4.36892301e-01 4.17600304e-01 -3.52250993e-01
-9.59555626e-01 -3.15190434e-01 -9.17052925e-02 -2.15937391e-01
-4.46668833e-01 -2.06582397e-01 -3.53844464e-01 6.94785342e-02
8.40972662e-01 -5.48671663e-01 1.88186184e-01 -3.74141902e-01
-9.80924666e-02 1.18226552e+00 -3.51155728e-01 -5.25324047e-01
7.29131341e-01 3.64321142e-01 -2.32992068e-01 -1.05991924e+00
-5.52241266e-01 -5.83450854e-01 -7.94979036e-01 -5.20834923e-01
2.60068048e-02 -4.30898905e-01 -7.05687881e-01 3.42204779e-01
-9.95582223e-01 -4.66234297e-01 -4.76080239e-01 4.55704242e-01
-3.31875324e-01 6.68059196e-03 -3.66743296e-01 -1.09507263e+00
-4.32119161e-01 -7.79388845e-01 1.39471009e-01 1.35015726e-01
-1.68512374e-01 -1.29052746e+00 4.28990543e-01 4.51437533e-01
7.39833772e-01 -4.07593250e-02 1.19726706e+00 -9.18647230e-01
-9.41057026e-01 -1.41509920e-01 -1.70971081e-01 4.01126683e-01
-8.26292089e-04 8.19873095e-01 -1.26682079e+00 -2.45945811e-01
-3.59863430e-01 -3.21490765e-01 2.59335726e-01 9.19509865e-03
1.54327893e+00 -4.90328163e-01 -4.61812079e-01 3.47150803e-01
1.45220268e+00 5.64327955e-01 9.22072351e-01 4.14026946e-01
2.79635310e-01 5.08752525e-01 3.93255830e-01 4.22327936e-01
-2.40959208e-02 4.37359542e-01 4.86764014e-01 -2.62115508e-01
-3.27724397e-01 5.03113985e-01 3.93661290e-01 8.90387237e-01
2.30476372e-02 -6.92475677e-01 -1.28723943e+00 5.13532348e-02
-1.78029895e+00 -1.42283547e+00 -6.69312537e-01 2.37044239e+00
4.87733185e-01 5.50918519e-01 5.59558034e-01 7.25031257e-01
7.62194395e-01 -5.90720065e-02 -5.82036138e-01 -4.62115467e-01
1.01468302e-01 4.54940200e-01 4.08688962e-01 4.02275831e-01
-1.11210656e+00 8.35924983e-01 6.99083710e+00 6.31788969e-01
-1.39137971e+00 4.07515138e-01 3.85505259e-01 7.58324191e-02
1.98934466e-01 2.53363758e-01 -6.99898303e-01 5.73741317e-01
1.53477216e+00 -1.64183810e-01 2.09666371e-01 8.12242031e-01
5.50830901e-01 1.96088627e-02 -7.35773981e-01 8.93151760e-01
-1.70324743e-01 -1.54671228e+00 8.77576619e-02 -1.05154002e-02
3.39241147e-01 3.72457862e-01 -3.33146691e-01 5.08737385e-01
3.56930763e-01 -5.57222009e-01 -1.01881083e-02 5.58597922e-01
5.51432967e-01 -7.38707483e-01 8.83729815e-01 1.90103814e-01
-7.98874557e-01 -5.01679599e-01 -1.52375013e-01 2.69333184e-01
-3.76940481e-02 8.51482391e-01 -1.39525056e+00 2.89951801e-01
7.80784786e-01 6.43710852e-01 -8.67591202e-01 1.62151480e+00
4.13254887e-01 1.16257489e+00 -1.16343513e-01 2.93956041e-01
-4.36368048e-01 7.39975646e-02 5.35153747e-01 1.26761055e+00
-8.29927027e-02 -4.34196293e-01 4.70728815e-01 4.39719290e-01
1.31351367e-01 3.40451717e-01 -3.08106542e-01 -3.75878930e-01
1.07352197e+00 1.42987704e+00 -8.59318674e-01 -4.25296247e-01
-4.14825946e-01 2.66176224e-01 3.44724692e-02 4.07031029e-01
-8.28386843e-01 -5.58860838e-01 3.81989449e-01 1.02311218e+00
-2.81120628e-01 -8.92340392e-02 -1.53961018e-01 -7.48169363e-01
-2.93690443e-01 -7.99982905e-01 6.70670211e-01 -4.36056584e-01
-1.13499081e+00 4.09943581e-01 1.39786109e-01 -9.28003252e-01
8.30020458e-02 -4.22038764e-01 -7.70174146e-01 6.17176481e-02
-1.39836216e+00 -8.46812248e-01 -7.24829853e-01 5.60644746e-01
7.27425456e-01 -8.69860828e-01 9.47318494e-01 1.01030815e+00
-1.21309733e+00 4.70388800e-01 -1.98102996e-01 1.32337078e-01
8.17227006e-01 -9.87461507e-01 -3.64504382e-02 7.49770880e-01
-7.98567608e-02 2.80928463e-01 3.51805001e-01 -5.97600222e-01
-1.20714486e+00 -1.06668639e+00 6.13974512e-01 -5.57892978e-01
8.96727026e-01 -4.63964760e-01 -7.59912729e-01 7.52827942e-01
1.86805576e-01 2.78538764e-01 1.09140325e+00 2.90281713e-01
-2.17138976e-01 -8.99009168e-01 -1.29512691e+00 1.49543332e-02
9.83605504e-01 -2.36879140e-01 -1.65356603e-03 6.03823066e-01
1.73678622e-01 3.67524207e-01 -8.18565845e-01 3.68899912e-01
3.23483080e-01 -1.07042062e+00 6.47898436e-01 -1.06652594e+00
-6.83927536e-01 1.45075824e-02 2.09403768e-01 -5.87277114e-01
-2.36085400e-01 -9.66134846e-01 -8.50750208e-01 1.53979993e+00
5.83263814e-01 -6.22896910e-01 1.23442924e+00 2.51717806e-01
-1.66557133e-01 -4.02514547e-01 -7.00180292e-01 -8.83858263e-01
-2.88349867e-01 -6.26166880e-01 7.68897533e-01 1.25721991e+00
-1.03587717e-01 4.93225902e-01 3.73611525e-02 -1.48524046e-01
6.14665985e-01 -5.21124780e-01 1.25696301e+00 -1.93345928e+00
1.19464271e-01 -2.75415003e-01 -7.98747003e-01 -4.10303354e-01
1.09014750e-01 -5.35032809e-01 -6.33960545e-01 -1.17689133e+00
-1.07154705e-01 -1.30296874e+00 -4.94879901e-01 6.77547097e-01
5.59201002e-01 1.74000394e-02 -1.79407328e-01 3.21349531e-01
-1.07665038e+00 -2.07992181e-01 6.72830403e-01 2.72782445e-01
-3.21040392e-01 2.60297269e-01 -4.22176532e-02 3.04395825e-01
1.26442444e+00 -6.47285700e-01 -1.03264046e+00 -2.67341286e-01
1.63552701e-01 -3.40288192e-01 3.23635787e-01 -1.12025058e+00
4.16573286e-01 -4.78451610e-01 3.71769518e-01 -8.54070306e-01
3.39230709e-02 -1.17280889e+00 3.97248805e-01 3.12071055e-01
-2.06513684e-02 4.12066787e-01 5.80096953e-02 2.64403969e-01
2.09919631e-01 -1.77614614e-01 9.13761914e-01 -2.23334044e-01
-7.80046582e-01 5.40850878e-01 -6.68812275e-01 2.83801913e-01
1.51517975e+00 -4.59062159e-01 -7.08368242e-01 9.81481597e-02
-7.12619841e-01 -4.29530852e-02 3.66942257e-01 5.18186271e-01
1.37128264e-01 -1.04067075e+00 -5.73082626e-01 5.25270879e-01
-9.68851075e-02 -4.63470250e-01 1.00279134e-02 1.14781332e+00
-8.54429662e-01 4.17734146e-01 -4.32808101e-01 -7.51102567e-01
-1.51999819e+00 4.58883524e-01 3.56008351e-01 2.64608040e-02
-2.61794090e-01 5.44417679e-01 -9.97766435e-01 -5.67384660e-01
6.95392013e-01 4.43974912e-01 -2.86679238e-01 4.70803767e-01
6.15468383e-01 1.14583099e+00 7.25238621e-01 -1.54855907e-01
-4.51127946e-01 -2.37279102e-01 -2.40131989e-01 3.20968181e-01
1.41307557e+00 -1.50773898e-01 -2.60046452e-01 7.52819955e-01
8.88677239e-01 2.91083474e-02 -6.52571380e-01 -9.45704505e-02
6.47571683e-01 -6.63655221e-01 1.23798743e-01 -6.59746885e-01
-1.31316638e+00 6.56446815e-01 1.23518646e+00 6.59782290e-01
9.23419178e-01 -3.75504255e-01 3.14716280e-01 5.79923928e-01
4.51724470e-01 -1.68720615e+00 4.25218761e-01 2.29054824e-01
1.28187791e-01 -1.34396911e+00 1.32134110e-01 -5.22135913e-01
-4.97814007e-02 1.53211761e+00 9.27270293e-01 3.67465556e-01
1.20891106e+00 6.63297653e-01 3.97369176e-01 -3.27832997e-01
-1.41589725e+00 -9.95603874e-02 -5.17336965e-01 7.76547790e-01
5.58127284e-01 -1.40452430e-01 -1.77332654e-03 -2.39861980e-01
1.98337585e-01 1.73068047e-01 5.76537251e-01 1.38716733e+00
-7.03216136e-01 -1.56199479e+00 -1.94363847e-01 6.81980908e-01
-1.19634256e-01 3.36227536e-01 -4.08936054e-01 9.99487698e-01
6.73088491e-01 1.24986827e+00 4.68372107e-01 -4.34853911e-01
2.49403283e-01 1.57952458e-01 2.46627957e-01 -8.19009781e-01
-6.30936265e-01 -3.48300755e-01 5.12552083e-01 -4.98577237e-01
-6.17749274e-01 -5.26984394e-01 -1.02994657e+00 -8.68598521e-01
-6.57977283e-01 4.20757771e-01 6.12803996e-01 6.27093673e-01
4.65004116e-01 2.54149079e-01 8.58940184e-01 2.41834134e-01
2.04221919e-01 -8.09878290e-01 -2.62468874e-01 1.00150242e-01
4.93916720e-02 -5.19847870e-01 -3.96995872e-01 1.19706497e-01] | [5.105746269226074, 7.202227592468262] |
e304622a-5749-4a04-93a0-e4cfb387d051 | trunk-branch-ensemble-convolutional-neural | 1607.05427 | null | http://arxiv.org/abs/1607.05427v2 | http://arxiv.org/pdf/1607.05427v2.pdf | Trunk-Branch Ensemble Convolutional Neural Networks for Video-based Face Recognition | Human faces in surveillance videos often suffer from severe image blur,
dramatic pose variations, and occlusion. In this paper, we propose a
comprehensive framework based on Convolutional Neural Networks (CNN) to
overcome challenges in video-based face recognition (VFR). First, to learn
blur-robust face representations, we artificially blur training data composed
of clear still images to account for a shortfall in real-world video training
data. Using training data composed of both still images and artificially
blurred data, CNN is encouraged to learn blur-insensitive features
automatically. Second, to enhance robustness of CNN features to pose variations
and occlusion, we propose a Trunk-Branch Ensemble CNN model (TBE-CNN), which
extracts complementary information from holistic face images and patches
cropped around facial components. TBE-CNN is an end-to-end model that extracts
features efficiently by sharing the low- and middle-level convolutional layers
between the trunk and branch networks. Third, to further promote the
discriminative power of the representations learnt by TBE-CNN, we propose an
improved triplet loss function. Systematic experiments justify the
effectiveness of the proposed techniques. Most impressively, TBE-CNN achieves
state-of-the-art performance on three popular video face databases: PaSC, COX
Face, and YouTube Faces. With the proposed techniques, we also obtain the first
place in the BTAS 2016 Video Person Recognition Evaluation. | ['DaCheng Tao', 'Changxing Ding'] | 2016-07-19 | null | null | null | null | ['person-recognition'] | ['computer-vision'] | [ 3.69888470e-02 -4.29683238e-01 5.85888922e-02 -6.48484528e-01
-3.65844995e-01 -3.50744337e-01 4.13658589e-01 -1.00298905e+00
-2.10815579e-01 6.79388762e-01 2.35707149e-01 1.72523677e-01
-3.94397490e-02 -4.41916227e-01 -8.72744560e-01 -6.33821726e-01
-2.42728386e-02 -3.73413354e-01 -2.76081055e-01 -1.34687126e-01
-1.88755020e-01 7.37736166e-01 -1.70388329e+00 6.24144077e-01
6.31112158e-01 1.44253981e+00 -2.17915624e-01 4.29000616e-01
3.30040216e-01 1.02106786e+00 -5.46819508e-01 -7.77194381e-01
5.23550928e-01 -8.72326344e-02 -5.03699362e-01 3.80994707e-01
1.09385026e+00 -9.45295870e-01 -1.01527739e+00 1.07939422e+00
4.66475457e-01 -1.59697711e-01 4.45974290e-01 -1.41437447e+00
-1.04560125e+00 5.41059673e-02 -6.92856431e-01 2.89861292e-01
3.51874799e-01 5.35087049e-01 4.41639781e-01 -1.41213846e+00
4.85955328e-01 1.55215001e+00 7.64429390e-01 8.89932573e-01
-8.81075919e-01 -9.37849820e-01 1.51817530e-01 4.53200698e-01
-1.57324350e+00 -1.05797327e+00 7.92517424e-01 -3.46496522e-01
7.25946546e-01 2.12698117e-01 4.36048687e-01 1.40284061e+00
9.05236043e-03 7.72250593e-01 8.05178344e-01 -3.16635706e-02
-2.43344367e-01 -1.49003536e-01 -3.08221579e-02 7.60010600e-01
3.31968725e-01 4.43846792e-01 -6.37066543e-01 2.34146398e-02
7.51617849e-01 5.09058177e-01 -7.93020725e-01 -1.17191344e-01
-7.19810069e-01 3.82263482e-01 6.92104220e-01 1.10121958e-01
-4.52063292e-01 1.69391558e-01 3.18669200e-01 3.03997934e-01
5.60270727e-01 -1.31139368e-01 -4.60852295e-01 1.61026627e-01
-1.12609446e+00 1.74651325e-01 3.04528445e-01 9.98798251e-01
7.11555600e-01 2.36020520e-01 -6.09215379e-01 1.10315466e+00
4.47937399e-01 4.84733135e-01 1.61211878e-01 -8.44385564e-01
4.91518050e-01 3.42562199e-01 9.18357074e-02 -1.03461647e+00
7.29564503e-02 -3.03014636e-01 -1.02791870e+00 1.68800056e-02
2.90525407e-01 -1.71617225e-01 -1.15662682e+00 1.64327562e+00
1.18416034e-01 4.82863247e-01 4.64056432e-02 1.07889831e+00
1.10277891e+00 4.28813279e-01 -1.30717620e-01 -2.51757413e-01
1.37643945e+00 -1.09096813e+00 -8.08564246e-01 5.98279983e-02
7.12448731e-03 -7.08425820e-01 4.62202132e-01 2.66312242e-01
-9.21743095e-01 -8.86563599e-01 -8.83062780e-01 -1.44027948e-01
-1.65261835e-01 5.10490417e-01 3.24025571e-01 7.75509894e-01
-1.20300448e+00 5.33930480e-01 -4.49347734e-01 -1.29730061e-01
1.25867844e+00 3.54749233e-01 -7.76709676e-01 -8.20435822e-01
-8.96637797e-01 5.60314894e-01 1.89773701e-02 8.17002237e-01
-1.38469529e+00 -7.75992155e-01 -9.39312100e-01 1.44236028e-01
2.95220405e-01 -5.92342854e-01 9.58531022e-01 -1.30334711e+00
-1.35028994e+00 6.34893000e-01 -3.41135681e-01 -2.25778162e-01
5.94789445e-01 -5.71021497e-01 -5.68398416e-01 2.59837270e-01
-2.09786803e-01 7.16415703e-01 1.34986019e+00 -1.17432964e+00
-3.01479757e-01 -5.40799201e-01 -1.76558495e-02 -1.79252759e-01
-6.09883428e-01 3.46040279e-01 -6.44079983e-01 -7.43465483e-01
-4.26403105e-01 -6.29184246e-01 3.20598871e-01 2.72372782e-01
-1.37380600e-01 -2.22897872e-01 1.29241717e+00 -9.33595777e-01
8.76822531e-01 -2.39203858e+00 -5.25332540e-02 -1.63988277e-01
3.23211223e-01 7.44783342e-01 -5.83276868e-01 -1.36964694e-01
-3.43950361e-01 -1.46310523e-01 -1.04351357e-01 -4.25594389e-01
-1.99853614e-01 6.61112294e-02 -3.56819868e-01 5.47780514e-01
8.85522366e-01 9.52640474e-01 -6.49182439e-01 -2.16475517e-01
2.08854467e-01 9.68974531e-01 -5.99682570e-01 4.89530951e-01
6.87373877e-02 3.59622717e-01 -2.20197812e-01 1.12293327e+00
1.28755152e+00 9.61556062e-02 -1.88624918e-01 -6.53251946e-01
1.28465146e-01 -2.82348096e-01 -8.10902417e-01 1.44243550e+00
-1.83408156e-01 7.20312119e-01 2.59773999e-01 -7.63854682e-01
7.67278016e-01 4.49252248e-01 3.24409246e-01 -4.90975648e-01
1.72802120e-01 1.27583265e-01 -1.87896565e-01 -6.36652112e-01
1.57895476e-01 2.78716922e-01 5.97292185e-01 -2.68891931e-01
5.08822203e-01 3.70601088e-01 3.54622751e-02 -1.10759698e-01
9.46541488e-01 1.84100360e-01 -1.42215371e-01 -7.29968622e-02
8.15939724e-01 -7.38458574e-01 7.01490998e-01 3.18387300e-01
-5.71054041e-01 8.10427308e-01 3.40871692e-01 -8.62384081e-01
-7.80104458e-01 -8.72518063e-01 -2.03753367e-01 8.09497178e-01
-1.66337699e-01 -4.26259607e-01 -8.11641276e-01 -8.89408171e-01
2.63853241e-02 7.78861865e-02 -9.27692115e-01 -2.10035458e-01
-6.06775701e-01 -5.80766857e-01 7.02958524e-01 5.69394171e-01
9.96778727e-01 -9.37386751e-01 -2.68958747e-01 -1.18984535e-01
-9.29643437e-02 -1.45290434e+00 -1.02401114e+00 -5.05594611e-01
-4.54235613e-01 -1.38242531e+00 -1.00110292e+00 -7.31026828e-01
7.54950047e-01 6.70524180e-01 8.80218625e-01 2.66012609e-01
-5.63446760e-01 5.04767239e-01 -2.90354222e-01 -9.16758999e-02
2.24146247e-01 -5.56982458e-01 1.08552925e-01 6.82847023e-01
4.10249323e-01 -3.09855372e-01 -7.77440250e-01 3.71035963e-01
-9.14670587e-01 -3.02252978e-01 5.30352235e-01 1.20466220e+00
7.75268599e-02 -4.76527885e-02 3.28855097e-01 -2.74243742e-01
3.26028228e-01 -2.98266739e-01 -5.27077377e-01 4.13266450e-01
-2.98107769e-02 -3.84572387e-01 5.49847543e-01 -3.79911393e-01
-1.17581272e+00 2.98417974e-02 -7.69285671e-03 -1.16982639e+00
-2.44516760e-01 1.01474330e-01 -3.95734906e-01 -3.35512578e-01
3.49260241e-01 2.87057251e-01 9.39156935e-02 -3.35220724e-01
3.16222310e-01 9.45612907e-01 6.62048638e-01 -4.52165306e-01
8.08485568e-01 3.79637867e-01 -1.79900914e-01 -1.05351186e+00
-7.61261642e-01 -1.51833221e-01 -4.12261039e-01 -4.13071662e-01
7.12662339e-01 -1.37180984e+00 -8.61860514e-01 9.14758921e-01
-1.49534714e+00 9.88524184e-02 1.74419329e-01 3.43179762e-01
-1.40910655e-01 3.82783413e-01 -7.69290090e-01 -9.31332886e-01
-4.99765605e-01 -1.25361347e+00 1.17903471e+00 4.58506495e-01
4.69429046e-01 -4.47535604e-01 -4.58516359e-01 4.93006408e-01
6.35252595e-01 2.83917099e-01 1.48491591e-01 -1.15742162e-01
-6.93584502e-01 -9.31886584e-02 -6.76706910e-01 8.44490409e-01
4.21003371e-01 2.70914078e-01 -1.45181227e+00 -7.75644064e-01
-2.83763334e-02 -4.53493595e-01 1.17416954e+00 2.83830434e-01
1.54560971e+00 -5.34074843e-01 -1.39637738e-01 9.81244743e-01
1.17813122e+00 3.86234745e-02 9.46208596e-01 -1.42653540e-01
8.83560061e-01 6.01676047e-01 2.74485260e-01 3.28829974e-01
1.46898597e-01 7.47094989e-01 5.06993592e-01 -2.85572652e-02
-4.51495618e-01 -1.02053985e-01 7.22156405e-01 2.43930742e-01
-3.04289281e-01 5.40077947e-02 -3.85131299e-01 4.19153333e-01
-1.52732694e+00 -1.18313599e+00 2.25600228e-01 1.98114204e+00
6.42872751e-01 -3.11160237e-01 7.04546347e-02 -1.94174871e-01
7.82211602e-01 2.13081598e-01 -4.03195411e-01 1.55831680e-01
-3.13189864e-01 2.95475662e-01 1.98592737e-01 9.29150954e-02
-1.28743374e+00 8.18356574e-01 5.59709740e+00 8.01267743e-01
-1.32283640e+00 1.66429114e-02 9.30113792e-01 -3.32953095e-01
3.13692451e-01 -4.88303512e-01 -8.11883569e-01 6.75522149e-01
6.80471659e-01 1.81345999e-01 6.33934557e-01 8.49673271e-01
1.28992228e-02 4.42176998e-01 -1.21268511e+00 1.42637527e+00
3.59415323e-01 -1.27559888e+00 9.11732912e-02 2.26687938e-02
7.51742780e-01 -1.56639591e-01 4.30290073e-01 3.46843809e-01
-1.85701638e-01 -1.46583080e+00 6.95250452e-01 5.88731885e-01
1.16004598e+00 -6.95082188e-01 9.33177412e-01 -2.81392753e-01
-1.25819945e+00 -3.44302118e-01 -6.03480935e-01 1.76915661e-01
-2.68119633e-01 5.06692708e-01 -3.21289331e-01 8.38142574e-01
1.06626010e+00 1.08263516e+00 -7.20112443e-01 1.07007933e+00
8.82002618e-03 3.47588956e-01 4.06004637e-02 4.80143845e-01
9.02524516e-02 1.18630063e-02 3.48075151e-01 1.20361233e+00
3.59632909e-01 2.42112175e-01 -9.95825678e-02 9.76850867e-01
-5.81044436e-01 -3.86068135e-01 -7.07738459e-01 -9.75870062e-03
3.33257943e-01 1.42379487e+00 1.37781501e-02 -2.41094723e-01
-6.47576869e-01 1.08421838e+00 3.86306822e-01 6.20830715e-01
-8.98216128e-01 -2.44813085e-01 1.10153437e+00 -1.81447998e-01
7.30449677e-01 4.93720621e-02 3.43913198e-01 -1.36970723e+00
3.86863798e-01 -1.14675224e+00 1.57371655e-01 -6.36741698e-01
-1.51606643e+00 9.66549277e-01 -4.98762764e-02 -1.15334904e+00
2.42229685e-01 -1.06057239e+00 -5.69359481e-01 8.88282180e-01
-1.72148621e+00 -1.33238423e+00 -7.14993358e-01 9.75502491e-01
7.14333594e-01 -4.81512934e-01 4.15942848e-01 6.36539340e-01
-1.00629199e+00 9.20721591e-01 -1.44593433e-01 6.26311600e-01
8.14358652e-01 -6.28763139e-01 2.31427759e-01 9.95510161e-01
-6.60220981e-02 6.84439898e-01 2.67239779e-01 -4.39192951e-01
-1.61321962e+00 -1.56878769e+00 3.97568285e-01 -4.40503687e-01
8.05560648e-02 -2.57125258e-01 -8.87491584e-01 6.33531034e-01
3.05212289e-01 8.69229496e-01 3.83448422e-01 -2.89039999e-01
-8.89218926e-01 -6.16696000e-01 -1.25267458e+00 2.87556261e-01
1.13118088e+00 -7.01463997e-01 -3.98715854e-01 1.65994465e-01
5.33897936e-01 -2.85849184e-01 -7.07272470e-01 6.96306407e-01
7.79349744e-01 -1.12760592e+00 1.07533658e+00 -6.96829796e-01
5.51292896e-01 -3.68955880e-01 -2.78670549e-01 -1.18323874e+00
-3.19170088e-01 -7.10809588e-01 -3.37415457e-01 1.29777539e+00
-7.06764236e-02 -5.15247464e-01 5.87307990e-01 4.81430322e-01
-6.32406250e-02 -7.60830522e-01 -9.40866411e-01 -7.80423462e-01
-1.37673199e-01 -5.29764257e-02 8.50542068e-01 7.57302761e-01
-6.07605338e-01 -3.76055576e-02 -8.01819742e-01 1.80089012e-01
8.65350962e-01 -3.14394623e-01 6.92194760e-01 -9.94260669e-01
6.48961142e-02 -2.54145324e-01 -4.58746433e-01 -1.03724384e+00
2.16741562e-01 -4.06757265e-01 -2.70639844e-02 -9.12421882e-01
3.67222607e-01 1.61197454e-01 -4.54696059e-01 4.33163643e-01
-2.87444323e-01 6.02170050e-01 4.09232825e-01 2.50127725e-02
-5.46759427e-01 8.77422333e-01 1.34069586e+00 -4.08128947e-01
3.08483392e-01 -2.63367116e-01 -6.28519297e-01 4.84999597e-01
3.13546926e-01 -6.20103851e-02 -1.06091104e-01 -6.92363620e-01
-5.60650110e-01 -6.08255155e-02 8.46623719e-01 -9.01784718e-01
1.26719713e-01 3.28626744e-02 1.10448194e+00 -4.02369678e-01
4.58441198e-01 -8.48181188e-01 1.40734911e-01 3.32714051e-01
-6.39692741e-03 -8.55884999e-02 3.04349989e-01 5.04639983e-01
-4.14389580e-01 2.11396277e-01 1.01018858e+00 1.60061002e-01
-5.47477424e-01 8.41601133e-01 2.04348341e-01 -2.76066303e-01
9.09119964e-01 -2.02669397e-01 -5.97819567e-01 -1.52453333e-01
-3.26873511e-01 7.82161430e-02 2.52522349e-01 7.21699119e-01
1.04546988e+00 -1.62534928e+00 -9.46692169e-01 6.03069425e-01
4.71925288e-02 -9.97110754e-02 5.87995887e-01 7.23304749e-01
-3.32648665e-01 4.75679755e-01 -5.61694860e-01 -4.84831780e-01
-1.36256075e+00 6.95877910e-01 5.43898702e-01 3.08302492e-01
-3.61717790e-01 1.08310795e+00 3.78697932e-01 -3.65336798e-03
3.84922028e-01 -1.22418419e-01 -2.03196898e-01 -1.73959225e-01
1.04243338e+00 2.63469696e-01 8.15358236e-02 -9.78986084e-01
-5.78985512e-01 5.46594739e-01 -3.19558740e-01 5.06527722e-01
1.37686074e+00 5.21242470e-02 -2.93981135e-01 -3.74331683e-01
1.52117407e+00 -2.80942738e-01 -1.78383613e+00 -2.30686799e-01
-4.16694939e-01 -9.77321029e-01 2.82491930e-02 -5.81037939e-01
-1.59395516e+00 8.78513277e-01 8.18514585e-01 -2.23699227e-01
1.40971351e+00 -3.57519418e-01 6.58137679e-01 2.07466334e-01
1.92686543e-01 -6.17655158e-01 3.00187886e-01 3.31841499e-01
1.30383539e+00 -1.31484783e+00 -2.10557371e-01 -5.13103127e-01
-3.47785920e-01 1.41817617e+00 9.22856331e-01 -9.10217762e-02
6.99287355e-01 -6.73467852e-03 -7.17521235e-02 -7.93638453e-02
-8.26288104e-01 -7.04079047e-02 6.61999643e-01 6.98212504e-01
2.75580913e-01 -2.02394396e-01 3.18734348e-01 6.99509323e-01
1.20208740e-01 3.18403095e-01 1.85757697e-01 5.78315973e-01
-1.02657780e-01 -6.80821002e-01 -6.06939435e-01 2.46038213e-01
-5.37263572e-01 -4.40039560e-02 -2.35813528e-01 5.70247769e-01
3.93090248e-01 1.10839832e+00 5.13483398e-02 -4.84678030e-01
4.00562197e-01 -1.33767948e-01 8.20955932e-01 -1.73343316e-01
-3.82632375e-01 -8.51744115e-02 -1.91685244e-01 -8.37010741e-01
-4.10605103e-01 -4.71945941e-01 -2.62644261e-01 -3.48287374e-01
-2.46880040e-01 -1.86818659e-01 2.50522584e-01 9.00847912e-01
5.12897074e-01 6.09385908e-01 8.89830410e-01 -1.16263330e+00
-6.90114439e-01 -1.15207839e+00 -4.50396866e-01 5.14825463e-01
7.92477250e-01 -9.09466863e-01 -2.86237419e-01 2.68656790e-01] | [13.18621826171875, 0.8224337697029114] |
51aaf8f5-a101-41d0-845a-bca497da748a | semi-supervised-lung-nodule-retrieval | 2005.01805 | null | https://arxiv.org/abs/2005.01805v1 | https://arxiv.org/pdf/2005.01805v1.pdf | Semi-supervised lung nodule retrieval | Content based image retrieval (CBIR) provides the clinician with visual information that can support, and hopefully improve, his or her decision making process. Given an input query image, a CBIR system provides as its output a set of images, ranked by similarity to the query image. Retrieved images may come with relevant information, such as biopsy-based malignancy labeling, or categorization. Ground truth on similarity between dataset elements (e.g. between nodules) is not readily available, thus greatly challenging machine learning methods. Such annotations are particularly difficult to obtain, due to the subjective nature of the task, with high inter-observer variability requiring multiple expert annotators. Consequently, past approaches have focused on manual feature extraction, while current approaches use auxiliary tasks, such as a binary classification task (e.g. malignancy), for which ground-true is more readily accessible. However, in a previous study, we have shown that binary auxiliary tasks are inferior to the usage of a rough similarity estimate that are derived from data annotations. The current study suggests a semi-supervised approach that involves two steps: 1) Automatic annotation of a given partially labeled dataset; 2) Learning a semantic similarity metric space based on the predicated annotations. The proposed system is demonstrated in lung nodule retrieval using the LIDC dataset, and shows that it is feasible to learn embedding from predicted ratings. The semi-supervised approach has demonstrated a significantly higher discriminative ability than the fully-unsupervised reference. | ['Hayit Greenspan', 'Mark Loyman'] | 2020-05-04 | null | null | null | null | ['content-based-image-retrieval'] | ['computer-vision'] | [ 5.49966395e-01 1.89036369e-01 -4.46971923e-01 -5.46165586e-01
-1.28618801e+00 -5.59303463e-01 5.53682208e-01 7.86775053e-01
-5.23745477e-01 5.59175670e-01 2.60645807e-01 -2.00062618e-01
-5.38911104e-01 -4.90880787e-01 -7.76974559e-02 -8.55077982e-01
2.68299967e-01 6.10267997e-01 4.29341733e-01 1.02066323e-01
3.32476676e-01 6.63498342e-01 -1.65359497e+00 5.06894886e-01
4.28228199e-01 1.35361552e+00 5.74899852e-01 6.15680218e-01
-2.08847404e-01 9.82921600e-01 -3.50774586e-01 -2.40106031e-01
1.12846948e-01 -3.65258962e-01 -1.01078808e+00 2.22422898e-01
4.91149053e-02 -1.21086158e-01 2.96792369e-02 1.03640950e+00
2.57275432e-01 -5.40306494e-02 1.09485841e+00 -1.02025580e+00
-6.67810678e-01 1.67006537e-01 -2.01588139e-01 4.05959338e-02
3.53391826e-01 -1.89877614e-01 1.10087872e+00 -1.02697110e+00
7.83192873e-01 6.45962954e-01 3.27566534e-01 5.15676975e-01
-1.15434432e+00 -2.33579695e-01 -2.95168817e-01 2.77914435e-01
-1.63625538e+00 -1.66530550e-01 8.05885255e-01 -7.03534484e-01
4.54362482e-01 5.81255674e-01 4.72038567e-01 6.88954592e-01
-1.23661600e-01 4.86393780e-01 1.28550398e+00 -6.22923911e-01
4.15242165e-01 8.99412930e-01 1.01765670e-01 7.03869641e-01
1.93489030e-01 -1.01861708e-01 -2.22755820e-01 -3.02759171e-01
3.82119149e-01 2.28434563e-01 -3.95291686e-01 -7.17937887e-01
-1.06886780e+00 8.73526871e-01 7.57120013e-01 7.51424074e-01
-4.61177707e-01 -1.37927815e-01 3.00562352e-01 9.71228406e-02
4.26117331e-01 6.07417643e-01 3.81687470e-03 2.41615146e-01
-1.19467521e+00 -2.72610217e-01 6.85957313e-01 7.38459587e-01
6.96169496e-01 -5.79170644e-01 -2.38080442e-01 8.43639791e-01
4.77790982e-01 2.07047343e-01 8.00670207e-01 -6.30342782e-01
-6.44886866e-02 7.77258635e-01 1.64164305e-01 -1.10086179e+00
-2.15155765e-01 -7.88977966e-02 -8.02773893e-01 3.51988047e-01
3.69141549e-01 5.58772922e-01 -9.97608960e-01 1.30134654e+00
1.35105237e-01 -4.06267554e-01 1.87465981e-01 1.19277918e+00
1.03384340e+00 1.62086695e-01 1.93422690e-01 -3.57127637e-01
1.49724114e+00 -7.98700631e-01 -7.75599360e-01 1.74847066e-01
6.69816673e-01 -9.42201793e-01 1.04788435e+00 3.00821632e-01
-6.98058486e-01 -3.74466360e-01 -1.04890776e+00 -1.08270217e-02
-6.69332206e-01 4.36806291e-01 4.55097705e-01 6.88088059e-01
-1.29114294e+00 3.90213877e-01 -8.23939085e-01 -7.05281258e-01
2.77167886e-01 3.65778923e-01 -6.70850277e-01 -7.63762519e-02
-8.04324806e-01 1.03006268e+00 4.34422404e-01 -8.17242935e-02
-7.06351340e-01 -3.20573717e-01 -7.22912848e-01 -2.40382738e-02
3.51789862e-01 -5.23067951e-01 1.09566760e+00 -1.21376145e+00
-1.15125871e+00 1.34190011e+00 -1.08779363e-01 -4.20128375e-01
5.10587513e-01 3.18937510e-01 -2.77480066e-01 6.51928782e-01
5.70927449e-02 8.61657977e-01 7.93191731e-01 -1.40198112e+00
-4.32161301e-01 -3.43737543e-01 3.87096480e-02 2.76472032e-01
-6.25596225e-01 -4.64091972e-02 -4.41747665e-01 -4.76684868e-01
1.56327978e-01 -1.02590513e+00 -2.87518829e-01 6.22173309e-01
-3.58428180e-01 -2.81733096e-01 7.43461788e-01 -5.19892991e-01
1.06013989e+00 -2.12721586e+00 -2.45936543e-01 3.72820318e-01
2.94514596e-01 2.66825736e-01 1.30128220e-01 3.65228862e-01
-1.87347099e-01 1.93713546e-01 -1.81356251e-01 -1.22390956e-01
-3.01069796e-01 1.66834936e-01 1.14989899e-01 3.47310483e-01
2.97762394e-01 6.52356505e-01 -8.71930718e-01 -1.14621007e+00
4.84970927e-01 4.18402076e-01 -2.51467347e-01 3.21653038e-01
1.73399419e-01 3.34183037e-01 -5.76503634e-01 7.00852156e-01
1.78369567e-01 -7.04746902e-01 7.84239843e-02 -6.41761482e-01
1.46706119e-01 -2.05800473e-03 -8.97761881e-01 1.67972445e+00
-5.18770158e-01 6.66135848e-01 -3.98208320e-01 -1.10190475e+00
1.02082551e+00 6.23460650e-01 8.66825223e-01 -3.99363488e-01
1.15707085e-01 3.48278672e-01 -9.46115553e-02 -6.08116031e-01
4.09079403e-01 -1.27476394e-01 1.17783219e-01 3.87618721e-01
-7.54906982e-02 -2.94182360e-01 1.35948583e-01 2.70292878e-01
1.18452799e+00 -1.22917458e-01 8.55276942e-01 -2.11127192e-01
7.23732054e-01 4.73913491e-01 9.45531856e-03 5.51449716e-01
-2.96565413e-01 8.93913209e-01 2.87724853e-01 -3.27065408e-01
-1.07270777e+00 -9.99364853e-01 -4.65984643e-01 5.50316870e-01
3.50898832e-01 -3.53270173e-01 -5.27448833e-01 -9.01391685e-01
-1.25701755e-01 4.54113394e-01 -9.47538018e-01 -1.54551312e-01
-3.16805318e-02 -3.65067244e-01 1.29781276e-01 4.39059913e-01
6.71894848e-02 -1.01446819e+00 -7.90997744e-01 7.11615151e-03
-9.07653421e-02 -8.90832603e-01 -3.08510453e-01 2.66671568e-01
-8.60242963e-01 -1.25372016e+00 -8.75444472e-01 -9.12270904e-01
1.14606464e+00 4.02360260e-01 9.30724680e-01 1.62691101e-01
-7.99369574e-01 6.83801949e-01 -5.20759404e-01 -1.34623840e-01
-5.64264953e-01 -1.55007616e-01 -2.08004355e-01 -2.09693480e-02
4.47325110e-01 3.26808766e-02 -9.41444218e-01 4.91112888e-01
-1.08314705e+00 7.08440542e-02 1.03894234e+00 1.00858068e+00
1.07688141e+00 -1.53235927e-01 4.20293897e-01 -1.00425339e+00
4.42671597e-01 -2.65967578e-01 -2.12170660e-01 5.76971710e-01
-9.19817150e-01 7.56991580e-02 3.04458410e-01 -3.59211326e-01
-9.71275806e-01 3.30190897e-01 1.38956860e-01 -5.33187270e-01
-2.49783933e-01 5.81871510e-01 2.57217884e-01 -1.63134083e-01
9.46451068e-01 8.98432732e-02 2.57044137e-01 -2.27926403e-01
2.11945206e-01 1.24269652e+00 3.85382473e-01 -1.68608099e-01
6.13987446e-01 4.72042680e-01 3.68966162e-02 -5.50258279e-01
-9.77171183e-01 -1.19050002e+00 -8.07368577e-01 -2.29180992e-01
1.08311868e+00 -6.95913970e-01 -3.17357957e-01 -4.37499791e-01
-7.95271099e-01 8.37230906e-02 -5.53418875e-01 7.70664573e-01
-6.63678646e-01 2.89381593e-01 -3.52130920e-01 -6.94214284e-01
-5.17619133e-01 -1.19528055e+00 1.17415571e+00 1.12039290e-01
-3.91081065e-01 -9.81473148e-01 3.23823020e-02 6.68904006e-01
4.45487827e-01 -2.67498810e-02 9.19599056e-01 -1.13076758e+00
-3.70525360e-01 -6.39584661e-01 -5.25013685e-01 3.87796938e-01
5.96513212e-01 -2.86157504e-02 -1.04065919e+00 -2.99922764e-01
3.78474146e-02 -5.62769890e-01 6.55186772e-01 6.83205053e-02
1.49233234e+00 -2.30653778e-01 -5.88601768e-01 8.35954174e-02
1.58317137e+00 1.79390341e-01 4.38316494e-01 1.87107921e-01
3.19614708e-01 8.04188490e-01 1.14379656e+00 1.61478087e-01
-8.63156654e-03 7.73443937e-01 4.24945801e-01 -3.37240010e-01
-3.75030905e-01 -4.85455245e-02 -2.78061658e-01 5.90095401e-01
-1.05668046e-01 -1.18385471e-01 -8.39398980e-01 5.08851886e-01
-1.75874674e+00 -6.58143878e-01 6.97776675e-02 2.34589338e+00
9.10966218e-01 -1.75442304e-02 -1.62161171e-01 3.42816502e-01
7.61168540e-01 -2.38898575e-01 -3.47390324e-01 -1.22595653e-01
3.51739287e-01 -5.63584547e-03 3.47682416e-01 1.76737085e-01
-1.08822489e+00 5.83615541e-01 5.50015020e+00 8.94437611e-01
-1.27061081e+00 1.75557762e-01 7.91758418e-01 2.38775447e-01
-1.87858447e-01 -1.30405426e-01 -3.93319607e-01 2.26903200e-01
5.83208978e-01 -2.71215469e-01 4.17386666e-02 1.06387055e+00
1.73300385e-01 -2.98269004e-01 -1.23410952e+00 1.10674119e+00
5.20887256e-01 -1.12571192e+00 1.22436106e-01 1.32824928e-01
5.56489825e-01 -2.50185907e-01 3.82364988e-02 -4.66394387e-02
-1.22208372e-01 -9.31912661e-01 2.59991437e-01 7.12295651e-01
1.09113252e+00 -3.42119306e-01 1.08618283e+00 1.75790057e-01
-1.09210145e+00 1.03517056e-01 -3.53774726e-01 5.20358980e-01
-2.87514448e-01 4.13476199e-01 -1.48979282e+00 5.18669307e-01
5.61891615e-01 4.53053385e-01 -8.25254083e-01 1.18829381e+00
-1.16258271e-01 3.48600119e-01 -3.72871733e-03 -1.52686924e-01
1.87234059e-01 8.10246319e-02 2.80591011e-01 1.16975367e+00
4.09745365e-01 2.41213460e-02 2.15141147e-01 6.18006647e-01
1.08624294e-01 7.25081801e-01 -7.72178054e-01 -1.57698646e-01
2.25964576e-01 1.87463713e+00 -1.15565431e+00 -3.68848622e-01
-4.06553507e-01 9.06607807e-01 1.60500765e-01 5.59776649e-02
-3.86362672e-01 -2.73041070e-01 1.49747133e-01 2.47639462e-01
1.09244265e-01 3.46846491e-01 -1.53026432e-01 -7.65838444e-01
-4.71425280e-02 -6.06245041e-01 6.35942280e-01 -9.75602210e-01
-1.31449687e+00 8.42660010e-01 -1.66704163e-01 -1.78419876e+00
-5.71471870e-01 -6.91373765e-01 -2.48971358e-01 7.15229690e-01
-1.37280750e+00 -1.23241866e+00 -6.48391962e-01 6.48072600e-01
4.99896705e-01 -1.15114741e-01 1.14559579e+00 1.19337663e-01
3.24655809e-02 4.25781757e-01 -2.42734402e-02 8.78553987e-02
1.01682472e+00 -1.45534694e+00 -8.85512710e-01 3.24071348e-01
3.57255995e-01 4.06712651e-01 5.91091871e-01 -3.33110392e-01
-1.03342259e+00 -1.12786961e+00 8.84227931e-01 -3.41073990e-01
7.12091684e-01 9.24279913e-03 -9.07306790e-01 2.40994453e-01
-1.34196565e-01 4.75647062e-01 1.11610007e+00 -2.66737998e-01
-2.68089771e-01 -2.56939828e-01 -1.35189843e+00 3.51666451e-01
5.45274317e-01 -8.32428277e-01 -5.26361346e-01 5.65014780e-01
4.08305913e-01 -1.14742085e-01 -1.07499099e+00 4.40613776e-01
4.71205950e-01 -6.91818833e-01 8.87177110e-01 -5.04771233e-01
4.35236245e-01 -4.33026046e-01 -2.81798005e-01 -9.42926288e-01
-5.22347763e-02 2.11030200e-01 4.39432412e-01 9.89242852e-01
5.14966369e-01 -1.81346640e-01 9.07334924e-01 7.78872073e-01
-3.77309173e-02 -8.24050784e-01 -7.51449287e-01 -6.46679103e-01
-3.99982154e-01 -2.53476977e-01 5.56239299e-02 9.05784249e-01
1.74513727e-01 1.13960477e-02 6.75955787e-02 1.06379591e-01
2.52232999e-01 3.19980711e-01 3.62546325e-01 -1.33846128e+00
-1.88458651e-01 -4.39252794e-01 -9.16187108e-01 -3.35475981e-01
-1.41963303e-01 -1.14462590e+00 2.65581727e-01 -1.74221992e+00
6.81591451e-01 -6.68004334e-01 -7.25567043e-01 6.22291803e-01
-5.48445769e-02 7.35678911e-01 -9.56622437e-02 6.42704248e-01
-7.39037037e-01 1.29981890e-01 1.07934010e+00 -3.33681762e-01
-2.57594194e-02 8.21451098e-02 -6.38488352e-01 7.05912709e-01
5.60173035e-01 -6.02250278e-01 -5.86326182e-01 2.23664165e-01
-1.30136115e-02 3.60358536e-01 3.45457494e-01 -9.14205015e-01
4.01111096e-01 1.77058764e-02 3.74668866e-01 -4.13654983e-01
4.16678876e-01 -1.15343988e+00 1.83911592e-01 6.13529325e-01
-7.31232822e-01 -3.59456390e-01 -2.49071971e-01 7.03256309e-01
-6.89858258e-01 -6.76325381e-01 6.10535622e-01 -1.94724631e-02
-5.34668982e-01 1.62282690e-01 -1.62773058e-01 -3.42828453e-01
1.14397085e+00 -3.86544168e-01 -1.22331813e-01 -3.95047307e-01
-1.07345665e+00 -2.49752894e-01 5.25889158e-01 2.09067702e-01
7.24220395e-01 -1.28827882e+00 -4.20006424e-01 -2.58669168e-01
8.00145566e-01 -1.67827919e-01 -1.43163383e-03 1.02066267e+00
-4.85046446e-01 5.38392663e-01 -4.92471009e-02 -8.53950679e-01
-1.65302920e+00 9.22891319e-01 1.24771893e-01 -4.10782099e-01
-3.60888064e-01 3.07550251e-01 2.10874364e-01 -5.37280925e-02
1.43943131e-01 -1.09529667e-01 -6.64712429e-01 2.93897986e-01
4.53691959e-01 9.23489034e-03 3.70695770e-01 -6.44238055e-01
-3.07999343e-01 5.96355677e-01 -2.64764130e-01 -9.70188826e-02
9.67284143e-01 -8.22625086e-02 5.37943803e-02 4.74238753e-01
1.23224461e+00 -2.75219709e-01 -6.71254277e-01 -4.07071620e-01
2.20554203e-01 -5.37377775e-01 2.99585015e-01 -8.63738000e-01
-9.13167834e-01 8.01692009e-01 1.07422101e+00 4.13530022e-01
1.29332936e+00 3.84645939e-01 1.43635452e-01 6.25978708e-01
4.28957134e-01 -9.09574032e-01 3.14797133e-01 -3.68223935e-01
8.83093715e-01 -1.69945872e+00 2.72768699e-02 -5.40648341e-01
-7.09429264e-01 1.19930053e+00 2.95618922e-01 2.50505000e-01
7.96602309e-01 -6.76220749e-04 3.35462600e-01 -3.25154126e-01
-6.32285416e-01 -4.85064358e-01 6.95669591e-01 4.48948145e-01
6.30857289e-01 1.39090762e-01 -6.47689521e-01 3.16483051e-01
8.16643536e-02 6.36372864e-02 2.79198647e-01 8.96947742e-01
-4.55574125e-01 -1.16919386e+00 -3.19310755e-01 8.13782990e-01
-6.15691841e-01 3.67139243e-02 -4.85532790e-01 7.98091471e-01
-1.02401309e-01 1.00353980e+00 1.44808710e-01 -1.47803038e-01
-9.94024216e-04 -7.10436478e-02 2.27380455e-01 -8.33547950e-01
-3.51950705e-01 2.95915250e-02 -6.94857398e-03 -5.18966734e-01
-7.15785801e-01 -5.03304660e-01 -1.02131796e+00 4.90978122e-01
-5.38003743e-01 4.16557312e-01 1.03979290e+00 7.81840503e-01
1.72353201e-02 3.02233934e-01 7.30911136e-01 -5.11838198e-01
-3.96288127e-01 -8.74692738e-01 -5.21338403e-01 8.47393632e-01
1.02379853e-02 -5.57121098e-01 -4.35066909e-01 3.19177032e-01] | [14.45875358581543, -1.623732328414917] |
afea311f-9b72-4899-b81b-d0837a086eac | boosting-low-data-instance-segmentation-by | 2302.01171 | null | https://arxiv.org/abs/2302.01171v1 | https://arxiv.org/pdf/2302.01171v1.pdf | Boosting Low-Data Instance Segmentation by Unsupervised Pre-training with Saliency Prompt | Recently, inspired by DETR variants, query-based end-to-end instance segmentation (QEIS) methods have outperformed CNN-based models on large-scale datasets. Yet they would lose efficacy when only a small amount of training data is available since it's hard for the crucial queries/kernels to learn localization and shape priors. To this end, this work offers a novel unsupervised pre-training solution for low-data regimes. Inspired by the recent success of the Prompting technique, we introduce a new pre-training method that boosts QEIS models by giving Saliency Prompt for queries/kernels. Our method contains three parts: 1) Saliency Masks Proposal is responsible for generating pseudo masks from unlabeled images based on the saliency mechanism. 2) Prompt-Kernel Matching transfers pseudo masks into prompts and injects the corresponding localization and shape priors to the best-matched kernels. 3) Kernel Supervision is applied to supply supervision at the kernel level for robust learning. From a practical perspective, our pre-training method helps QEIS models achieve a similar convergence speed and comparable performance with CNN-based models in low-data regimes. Experimental results show that our method significantly boosts several QEIS models on three datasets. Code will be made available. | ['Junwei Han', 'Xinggang Wang', 'Chao Zhang', 'Yalun Dai', 'Lechao Cheng', 'Nian Liu', 'Dingwen Zhang', 'Hao Li'] | 2023-02-02 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_Boosting_Low-Data_Instance_Segmentation_by_Unsupervised_Pre-Training_With_Saliency_Prompt_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_Boosting_Low-Data_Instance_Segmentation_by_Unsupervised_Pre-Training_With_Saliency_Prompt_CVPR_2023_paper.pdf | cvpr-2023-1 | ['unsupervised-pre-training'] | ['methodology'] | [ 8.96429345e-02 2.28726700e-01 -5.64505517e-01 -4.31042403e-01
-8.50461841e-01 -3.49977940e-01 3.94084394e-01 9.70264748e-02
-5.58469474e-01 4.16634619e-01 -1.09891579e-01 -7.94721320e-02
-3.80613804e-02 -5.76592863e-01 -9.22810435e-01 -4.29957181e-01
1.84754938e-01 2.33590826e-01 1.03471804e+00 -7.01760426e-02
3.49329472e-01 3.37383002e-01 -1.41313899e+00 3.65931243e-01
1.18914223e+00 1.10420871e+00 5.97313762e-01 5.92995346e-01
-2.08498195e-01 6.94376469e-01 -4.09588635e-01 -2.96536744e-01
2.10208461e-01 -2.20233902e-01 -8.81662786e-01 9.04365908e-03
6.00570440e-01 -2.49269590e-01 -2.03377351e-01 9.46072876e-01
4.57167596e-01 1.11312106e-01 5.42048991e-01 -1.19649327e+00
-7.34287918e-01 3.82898271e-01 -6.18070841e-01 3.98901463e-01
-2.72292495e-01 1.95649475e-01 7.97047913e-01 -1.06662488e+00
5.58985829e-01 8.45813274e-01 7.42022336e-01 6.42074108e-01
-1.07685721e+00 -2.90249825e-01 2.41322070e-01 2.45453238e-01
-1.34665751e+00 -2.40663558e-01 7.96734214e-01 -1.50687918e-01
8.21175635e-01 1.32411793e-01 5.40997386e-01 7.92724371e-01
-7.50547647e-02 1.40289354e+00 9.46669161e-01 -3.97729754e-01
3.06107402e-01 4.43335086e-01 2.93536007e-01 6.27918482e-01
-1.62538633e-01 1.66546822e-01 -6.18583620e-01 1.85122609e-01
1.06611228e+00 9.83005092e-02 -1.45433739e-01 -3.90808403e-01
-1.29283881e+00 7.57121563e-01 9.92793500e-01 3.03933084e-01
-3.20927352e-01 -1.98619626e-02 1.58638090e-01 5.44171408e-02
7.29929388e-01 3.70663822e-01 -6.77184224e-01 1.30089760e-01
-1.37969971e+00 2.43188769e-01 4.82767642e-01 1.16581988e+00
9.16903317e-01 -1.44170627e-01 -4.86284167e-01 8.66075873e-01
1.47970155e-01 2.31646165e-01 4.47633386e-01 -5.63374817e-01
4.22204345e-01 7.38639951e-01 3.57592925e-02 -5.97211123e-01
-3.62316728e-01 -6.87242925e-01 -5.86725771e-01 4.43538651e-02
5.58169246e-01 2.80338302e-02 -1.13236332e+00 1.35063481e+00
5.11609495e-01 6.33313417e-01 -4.25084345e-02 1.21483719e+00
9.35985446e-01 4.56957400e-01 1.13113083e-01 3.95678103e-01
1.43501925e+00 -1.44922602e+00 -4.66688901e-01 -2.96440214e-01
5.74392259e-01 -7.83298850e-01 1.38756669e+00 8.68387613e-03
-1.10778236e+00 -8.21100473e-01 -8.29699755e-01 -1.67281896e-01
-4.71036702e-01 6.84526205e-01 5.79485774e-01 3.95066530e-01
-1.30044711e+00 4.20160234e-01 -8.41835201e-01 -3.86948228e-01
7.12949812e-01 3.34296018e-01 8.05722326e-02 1.51895121e-01
-1.02827537e+00 6.47399724e-01 4.84898359e-01 1.95443600e-01
-8.79999697e-01 -1.00840592e+00 -7.62419701e-01 -1.64372753e-02
6.56750441e-01 -5.33316553e-01 1.36128998e+00 -1.16248000e+00
-1.53748691e+00 8.65957201e-01 -1.79227635e-01 -6.06813133e-01
4.73806143e-01 -3.12604606e-01 1.21335544e-01 3.85306448e-01
3.30671579e-01 1.25685251e+00 1.20269108e+00 -1.12485874e+00
-8.49074483e-01 -2.65023649e-01 -3.32645737e-02 1.91917643e-01
-3.14804196e-01 6.35420457e-02 -8.99869978e-01 -7.61719882e-01
-8.78398046e-02 -7.19099104e-01 -3.42518181e-01 1.05586596e-01
-5.24460733e-01 -3.35632384e-01 9.79935348e-01 -4.97670174e-01
1.02821946e+00 -2.16751981e+00 -1.95126176e-01 -1.38247963e-02
3.38954955e-01 6.28933966e-01 -1.55410975e-01 1.07182488e-01
1.46421930e-02 -1.07601792e-01 -2.78540939e-01 -4.18637246e-01
-1.40808344e-01 3.08776665e-02 -4.87042159e-01 2.01969251e-01
7.19421268e-01 1.42418659e+00 -9.00139332e-01 -6.77733362e-01
4.06994462e-01 3.20425600e-01 -4.52585787e-01 3.89214575e-01
-2.96589762e-01 3.88649553e-01 -5.38152516e-01 8.26179683e-01
6.66285634e-01 -5.31385183e-01 -5.48163772e-01 -3.30043942e-01
-1.28266096e-01 5.39229028e-02 -9.41843033e-01 1.74945414e+00
-2.06725672e-01 4.70062375e-01 8.47346261e-02 -1.10748255e+00
7.50090897e-01 1.39568970e-01 1.69668525e-01 -7.09195077e-01
-3.96464206e-02 2.51250207e-01 -2.54647791e-01 -3.30411643e-01
4.34613764e-01 1.54960483e-01 1.64573193e-01 6.58225715e-02
2.42166698e-01 -1.51421696e-01 8.12735558e-02 5.08482873e-01
8.00372005e-01 4.24651802e-01 -5.50867356e-02 -4.39843476e-01
4.16596144e-01 3.13729703e-01 5.91434836e-01 6.66330934e-01
-5.00478208e-01 9.38036203e-01 1.79901242e-01 -4.80781913e-01
-9.58961785e-01 -1.04829895e+00 -8.90996084e-02 1.24048567e+00
4.05494064e-01 -1.48408100e-01 -1.15578234e+00 -9.69941139e-01
-9.90853831e-02 3.12201321e-01 -7.22760081e-01 -5.56243584e-02
-3.56717914e-01 -6.09781086e-01 5.17529488e-01 9.06211913e-01
8.39457095e-01 -1.22724986e+00 -6.53290868e-01 2.30448306e-01
-7.04633296e-02 -1.19477510e+00 -7.49078870e-01 3.00604463e-01
-1.15383947e+00 -9.90534604e-01 -1.13382673e+00 -1.01721120e+00
8.51078153e-01 6.01743758e-01 9.74123240e-01 2.10432529e-01
-4.53346223e-01 1.30977228e-01 -4.47991937e-01 -5.03299356e-01
1.11082360e-01 4.11283284e-01 -3.04220855e-01 5.07263914e-02
2.22544238e-01 -1.96620345e-01 -7.84193575e-01 4.46173072e-01
-1.00467288e+00 3.06954026e-01 8.41171801e-01 7.37068295e-01
7.35204816e-01 -3.12696904e-01 7.07718730e-01 -8.26051295e-01
3.75232846e-01 -2.43267491e-01 -5.68176270e-01 3.05211216e-01
-5.18486440e-01 -7.97291752e-03 5.12842953e-01 -6.36892259e-01
-1.03184021e+00 2.30416164e-01 -1.67102367e-02 -6.89958155e-01
-4.05111194e-01 2.45849371e-01 1.33634293e-02 -1.09426931e-01
8.73046041e-01 2.71922410e-01 -8.19397569e-02 -5.75781822e-01
5.57700813e-01 5.79005003e-01 6.43377781e-01 -5.43362916e-01
7.86092520e-01 6.09045625e-01 -4.83938992e-01 -7.41916478e-01
-1.18987453e+00 -8.25947762e-01 -7.88587272e-01 -1.28318846e-01
8.91214967e-01 -1.01603317e+00 -3.95952106e-01 5.85542619e-01
-9.42641556e-01 -7.31251955e-01 -4.14366812e-01 3.85003209e-01
-6.19428873e-01 1.57862023e-01 -7.08918869e-01 -5.22752762e-01
-3.93261850e-01 -1.12141621e+00 1.33494508e+00 7.12174118e-01
9.05709639e-02 -1.04631686e+00 -2.01828301e-01 4.26136643e-01
5.88607192e-01 -1.23897120e-01 3.80396068e-01 -6.86030388e-01
-8.40405226e-01 -2.19121724e-02 -7.73293614e-01 2.80375987e-01
7.17233866e-02 -7.16191083e-02 -1.16391110e+00 -1.31929651e-01
-1.03329249e-01 -3.64302367e-01 1.02454662e+00 5.99124372e-01
1.20518875e+00 1.01472698e-02 -5.14100730e-01 7.57211924e-01
1.33386791e+00 -1.49135783e-01 5.25846720e-01 3.88312966e-01
8.40960145e-01 6.47518337e-01 8.25184345e-01 5.16165700e-03
4.88866985e-01 6.26022458e-01 3.42969924e-01 -6.80583715e-01
-3.98452789e-01 -2.72879869e-01 1.43162385e-01 4.94892240e-01
7.65706897e-02 2.40393952e-01 -7.41418481e-01 9.11975861e-01
-2.17492604e+00 -6.56837583e-01 -2.61325657e-01 2.01367664e+00
9.66091096e-01 1.61098108e-01 3.14746380e-01 -2.33191550e-01
6.42112911e-01 -4.33489420e-02 -7.73238361e-01 4.24570516e-02
8.81702378e-02 2.40346447e-01 5.93837440e-01 3.34272057e-01
-1.36666465e+00 1.47475779e+00 5.48626947e+00 1.22350979e+00
-1.29283524e+00 3.02209318e-01 8.81049693e-01 9.43270400e-02
8.07829015e-03 -1.35505630e-03 -1.06400585e+00 5.26705205e-01
7.32422531e-01 3.32919598e-01 1.85566142e-01 1.13286662e+00
1.97081506e-01 -3.37585151e-01 -8.54451358e-01 8.37297440e-01
-1.31977826e-01 -1.57082939e+00 -2.06218287e-01 -1.31014064e-01
7.62756646e-01 3.02993417e-01 1.20431185e-01 3.71262789e-01
1.12785019e-01 -7.42737114e-01 9.20310318e-01 4.70359415e-01
6.69878006e-01 -6.23168468e-01 6.98008358e-01 6.30262792e-01
-1.24513865e+00 6.80965278e-03 -4.79857653e-01 9.28999111e-02
-4.29677479e-02 7.11761475e-01 -1.08482647e+00 5.81199229e-01
7.11456418e-01 6.73099816e-01 -8.97469223e-01 1.25576723e+00
-4.09232557e-01 9.31678116e-01 -3.31577867e-01 3.08534931e-02
5.08687556e-01 -1.35990247e-01 1.18965238e-01 1.35165203e+00
-1.10460766e-01 -4.37669195e-02 2.72665322e-01 1.22015929e+00
1.06287844e-01 8.28643888e-02 -3.57947946e-01 2.20619053e-01
2.42020741e-01 1.48255444e+00 -1.23768306e+00 -5.86355329e-01
-3.48534644e-01 1.21783209e+00 5.53430974e-01 5.75669527e-01
-9.29786921e-01 -5.36347270e-01 3.44172686e-01 3.04483682e-01
5.24711728e-01 -6.20107912e-03 -4.88800675e-01 -1.10997915e+00
6.02169000e-02 -4.51637685e-01 1.88769087e-01 -7.54732132e-01
-1.27402294e+00 5.47487557e-01 3.06585450e-02 -1.03201628e+00
-2.12902445e-02 -5.64064980e-01 -8.71080160e-01 8.49678993e-01
-2.08906913e+00 -1.51556516e+00 -3.08468223e-01 7.10624039e-01
8.14212263e-01 1.52180195e-01 4.67214018e-01 1.54755458e-01
-7.02681839e-01 5.30998945e-01 -1.15586519e-01 3.47217470e-01
7.24946320e-01 -1.45693696e+00 5.60330689e-01 8.83274257e-01
3.04848045e-01 6.29194438e-01 2.32913256e-01 -5.95476091e-01
-9.78297353e-01 -1.26435995e+00 7.28693843e-01 -4.83058840e-01
4.35465336e-01 -3.70323896e-01 -1.05217874e+00 4.38266277e-01
2.12939624e-02 4.30903941e-01 4.00310487e-01 -4.50852476e-02
-9.12964195e-02 -1.20385394e-01 -1.00842237e+00 4.75194514e-01
8.24115157e-01 -5.42616665e-01 -5.12834132e-01 3.76826525e-01
8.49762797e-01 -5.44163406e-01 -5.24362266e-01 4.83238280e-01
4.13435102e-02 -8.78633380e-01 1.04579008e+00 -3.64578098e-01
2.80166656e-01 -5.34393907e-01 2.99235851e-01 -9.83485937e-01
-1.98376209e-01 -5.34456432e-01 -1.17761731e-01 1.14941859e+00
4.22886699e-01 -3.59761894e-01 1.02740908e+00 5.57561219e-01
-3.41948360e-01 -1.18414569e+00 -6.91641867e-01 -7.16808319e-01
-1.73744321e-01 -5.17138004e-01 4.54348862e-01 7.82562077e-01
-2.98454791e-01 2.27621138e-01 -5.93028516e-02 2.10506037e-01
6.30358875e-01 2.26309255e-01 7.90354252e-01 -9.88649428e-01
-1.67785391e-01 -3.99209261e-01 -2.37916008e-01 -1.34745097e+00
-9.29788649e-02 -8.67438555e-01 1.64678499e-01 -1.49962080e+00
1.71446204e-01 -5.97998023e-01 -3.14992160e-01 7.27080047e-01
-6.25957131e-01 4.65143770e-01 1.98556215e-01 4.26105320e-01
-9.02274847e-01 5.35132766e-01 1.38748085e+00 -6.53549377e-03
-4.41337436e-01 3.11291099e-01 -6.15499437e-01 6.97154403e-01
7.58736908e-01 -2.63474852e-01 -5.76372802e-01 -2.47245133e-01
-1.50032103e-01 -2.65396714e-01 7.46664703e-01 -1.10974002e+00
3.92565489e-01 -4.37180996e-02 3.79847676e-01 -6.85252488e-01
3.59386094e-02 -5.26417732e-01 -5.09970009e-01 2.11453393e-01
-2.82129467e-01 -3.50573599e-01 2.56053180e-01 5.42685747e-01
-2.31584400e-01 -1.72190055e-01 8.71602476e-01 -4.58501577e-02
-9.58615005e-01 4.59056824e-01 -1.60610706e-01 2.40270600e-01
9.79042411e-01 -5.64866781e-01 -1.28265340e-02 -5.93524724e-02
-7.67341137e-01 3.75717610e-01 4.68845338e-01 3.92559826e-01
7.36514151e-01 -1.07618845e+00 -4.82880443e-01 3.19953173e-01
2.22177237e-01 5.04808843e-01 2.12468460e-01 1.06006372e+00
-4.29268956e-01 5.14263868e-01 3.24430093e-02 -8.89538646e-01
-1.02473521e+00 6.38277113e-01 2.17859164e-01 -2.45037034e-01
-6.00563645e-01 1.25384331e+00 5.49579442e-01 -5.92441857e-01
3.95701796e-01 -8.42537403e-01 -3.03533766e-02 -1.92390069e-01
5.90153575e-01 1.05693966e-01 9.45439339e-02 -4.02279198e-01
-2.90791661e-01 5.39592862e-01 -4.34624344e-01 -4.45010364e-02
1.24179602e+00 -1.53316511e-02 2.54907072e-01 2.64253080e-01
1.07224882e+00 -2.84666061e-01 -1.85003042e+00 -5.39247572e-01
1.23273797e-01 -4.06450808e-01 1.40336573e-01 -8.73091042e-01
-1.07818484e+00 1.01420534e+00 5.67544103e-01 -9.50843561e-03
1.19525003e+00 2.96290517e-01 1.04583883e+00 6.41513318e-02
3.01893860e-01 -1.28117335e+00 3.30863088e-01 4.41235244e-01
4.57459867e-01 -1.42236447e+00 -3.84930611e-01 -6.45568252e-01
-8.29032123e-01 8.87994766e-01 8.01131666e-01 -2.01932535e-01
6.21674418e-01 1.59146488e-01 2.79986680e-01 -3.05292130e-01
-4.25818950e-01 -5.31719923e-01 6.87105238e-01 6.77447081e-01
2.55379617e-01 -8.49108323e-02 2.93705761e-02 6.62490904e-01
1.00552961e-01 1.49666622e-01 2.29613651e-02 8.35814893e-01
-5.31880915e-01 -9.23649371e-01 -3.20188046e-01 3.54688108e-01
-4.20037925e-01 -2.46135145e-01 -2.72429824e-01 6.91254258e-01
2.29114026e-01 7.54406929e-01 -4.06634398e-02 -3.88586111e-02
3.08956355e-01 -8.39503258e-02 1.86681986e-01 -8.50915790e-01
-7.30461180e-01 2.51090735e-01 -4.73551422e-01 -8.34412932e-01
-3.45614731e-01 -3.10573995e-01 -1.56281209e+00 1.95649132e-01
-6.54298842e-01 1.51406005e-01 6.47276044e-01 1.00022697e+00
6.80472314e-01 6.30784571e-01 3.37029785e-01 -1.05355585e+00
-5.58792949e-01 -9.37383890e-01 -4.56131697e-01 2.81562775e-01
2.53274024e-01 -5.51507950e-01 -3.10519012e-03 2.65164316e-01] | [9.5691499710083, 0.5273327827453613] |
dc9c524e-22fc-46dc-b4de-14e6b6e2c13e | i-code-studio-a-configurable-and-composable | 2305.13738 | null | https://arxiv.org/abs/2305.13738v1 | https://arxiv.org/pdf/2305.13738v1.pdf | i-Code Studio: A Configurable and Composable Framework for Integrative AI | Artificial General Intelligence (AGI) requires comprehensive understanding and generation capabilities for a variety of tasks spanning different modalities and functionalities. Integrative AI is one important direction to approach AGI, through combining multiple models to tackle complex multimodal tasks. However, there is a lack of a flexible and composable platform to facilitate efficient and effective model composition and coordination. In this paper, we propose the i-Code Studio, a configurable and composable framework for Integrative AI. The i-Code Studio orchestrates multiple pre-trained models in a finetuning-free fashion to conduct complex multimodal tasks. Instead of simple model composition, the i-Code Studio provides an integrative, flexible, and composable setting for developers to quickly and easily compose cutting-edge services and technologies tailored to their specific requirements. The i-Code Studio achieves impressive results on a variety of zero-shot multimodal tasks, such as video-to-text retrieval, speech-to-speech translation, and visual question answering. We also demonstrate how to quickly build a multimodal agent based on the i-Code Studio that can communicate and personalize for users. | ['Xuedong Huang', 'Michael Zeng', 'Lu Yuan', 'Takuya Yoshioka', 'Yao Qian', 'Yichong Xu', 'Reid Pryzant', 'ZiYi Yang', 'Chenguang Zhu', 'Mahmoud Khademi', 'Yuwei Fang'] | 2023-05-23 | null | null | null | null | ['speech-to-speech-translation'] | ['speech'] | [-3.86926010e-02 -1.61167219e-01 4.50952128e-02 -4.33667153e-01
-7.51659751e-01 -7.82450259e-01 7.94609070e-01 -2.73649424e-01
6.78484589e-02 -3.54512893e-02 3.34731154e-02 -1.40184432e-01
-1.62780300e-01 -4.16786939e-01 -3.80493492e-01 -2.79055178e-01
1.53880492e-01 7.66916573e-01 -1.41539618e-01 -5.30830681e-01
-8.99726152e-02 6.30454496e-02 -1.91871381e+00 7.16966629e-01
8.94553304e-01 9.58449364e-01 5.21212399e-01 8.95676911e-01
-4.14999574e-01 6.37361228e-01 -5.40350616e-01 -3.99140656e-01
-1.54851213e-01 -3.91375929e-01 -6.94179177e-01 -3.68965343e-02
2.28004813e-01 -2.31166989e-01 7.39280656e-02 6.05933785e-01
5.44643104e-01 -7.71413818e-02 4.03404713e-01 -1.79200542e+00
-5.57935536e-01 7.62323797e-01 2.15279199e-02 -4.16074216e-01
7.56108701e-01 6.22391701e-01 7.87226319e-01 -8.37569892e-01
5.15686095e-01 1.58549941e+00 9.36909765e-02 9.35952008e-01
-8.81612659e-01 -4.25320625e-01 3.80409300e-01 1.48776725e-01
-1.15227127e+00 -7.86815941e-01 5.49307644e-01 -3.14004421e-01
9.07016456e-01 7.17488825e-01 5.37594259e-01 1.26800430e+00
-3.78709257e-01 1.23038864e+00 5.04564762e-01 -4.15617198e-01
1.93510026e-01 1.27095833e-01 2.64069308e-02 7.34202921e-01
-4.61161196e-01 -5.29486299e-01 -9.78578031e-01 -1.64849147e-01
5.50666928e-01 -4.55894135e-02 -1.37738302e-01 -1.96703911e-01
-1.83188415e+00 3.64659250e-01 -6.68006092e-02 2.80431360e-01
-1.32524192e-01 4.10822660e-01 3.37490976e-01 4.05341238e-01
-1.78184822e-01 7.18246162e-01 -5.27973056e-01 -5.43577552e-01
-3.44660133e-01 2.28597611e-01 1.05760467e+00 1.39264762e+00
6.02213264e-01 -7.17387721e-02 -3.00693512e-01 9.72268879e-01
6.02942467e-01 8.10224116e-01 3.63581181e-01 -1.32533205e+00
3.09581906e-01 9.48428333e-01 1.42203540e-01 -6.81500494e-01
-3.68639678e-01 -1.25557154e-01 -5.20439744e-01 1.50020435e-01
2.31064290e-01 -2.79513508e-01 -8.93034875e-01 1.76191735e+00
5.26374757e-01 -3.32762957e-01 7.83675388e-02 1.06425536e+00
1.32946479e+00 7.87969589e-01 2.48978898e-01 3.22609335e-01
1.42231262e+00 -1.44165730e+00 -5.09377062e-01 -3.76414448e-01
4.67090100e-01 -6.54841959e-01 1.58003700e+00 2.44351640e-01
-1.32017708e+00 -3.85505676e-01 -7.99074352e-01 -1.13948271e-01
-5.30852437e-01 9.71905142e-02 6.95204854e-01 1.76526681e-01
-9.69910145e-01 -5.68023995e-02 -6.19031787e-01 -5.40712357e-01
-5.07324422e-03 3.72471511e-01 -4.74745601e-01 -4.14406918e-02
-8.16011071e-01 7.96073020e-01 1.33163944e-01 -7.34001398e-02
-1.05610645e+00 -6.37770534e-01 -9.19490218e-01 1.03013851e-01
5.97569287e-01 -1.37916350e+00 1.56777060e+00 -1.25422251e+00
-1.95171273e+00 7.13664234e-01 -8.48043561e-02 2.07149491e-01
1.42641529e-01 4.72886041e-02 -4.28706706e-01 1.13573909e-01
-2.40296915e-01 9.83194351e-01 1.04219282e+00 -1.31826401e+00
-5.03222346e-01 -3.21464062e-01 6.27657473e-01 5.01486957e-01
-4.93324310e-01 3.49624306e-01 -1.01484609e+00 -2.26857141e-01
-3.59556913e-01 -9.29319799e-01 3.69156078e-02 1.35300741e-01
-2.25060806e-01 -1.10679343e-01 8.51596415e-01 -2.98306555e-01
1.17699182e+00 -2.24964499e+00 7.25593448e-01 3.18342857e-02
2.80379385e-01 2.52168179e-01 -7.68305600e-01 7.88525283e-01
3.98085594e-01 -7.72525603e-03 7.85953458e-03 -4.56453234e-01
7.60659873e-01 2.39975214e-01 1.68673545e-01 -4.24701452e-01
3.75885027e-03 1.15848267e+00 -7.85729229e-01 -5.78022361e-01
1.06021404e-01 4.51292992e-01 -5.53238571e-01 5.89123189e-01
-9.77015018e-01 4.78671640e-01 -4.52109814e-01 1.12280011e+00
1.97326452e-01 -5.28202951e-01 1.47788167e-01 -1.62811801e-01
-5.52169755e-02 -1.44768924e-01 -9.16004300e-01 2.35302758e+00
-7.56576538e-01 4.17095751e-01 7.11287796e-01 -6.22769773e-01
7.67550468e-01 5.01268506e-01 1.78970307e-01 -7.32598722e-01
7.73649886e-02 1.64544120e-01 -1.03227265e-01 -9.51930463e-01
4.23305452e-01 4.47287887e-01 -4.10620898e-01 6.71442986e-01
1.54382005e-01 -2.09349483e-01 5.88003576e-01 3.69752377e-01
1.03809083e+00 2.82329679e-01 -1.00518607e-01 2.44432807e-01
4.04585123e-01 9.45677087e-02 1.59948662e-01 6.89036250e-01
2.86843449e-01 4.44479048e-01 1.84436783e-01 -4.04973328e-01
-7.13428378e-01 -9.21845913e-01 4.47535932e-01 1.89448571e+00
1.13186941e-01 -7.30696678e-01 -8.74453425e-01 -3.28641623e-01
-2.18546733e-01 5.88502407e-01 -2.66593665e-01 -6.66432381e-02
-8.49490985e-02 -4.14681762e-01 4.34860528e-01 3.30914974e-01
4.98665392e-01 -1.14937401e+00 -7.00516701e-01 1.07764162e-01
-5.53398967e-01 -8.99574816e-01 -7.23220408e-01 -3.31544787e-01
-2.27456674e-01 -7.02186048e-01 -5.72153926e-01 -7.86090851e-01
5.54731667e-01 4.81223494e-01 1.28447700e+00 2.01943502e-01
-3.72747093e-01 1.21063781e+00 -5.61090767e-01 -4.50218260e-01
-5.84692180e-01 3.04827183e-01 -4.54013288e-01 8.29381123e-02
-2.36936342e-02 -4.99873638e-01 -6.37551188e-01 6.20635927e-01
-1.14949834e+00 7.39811957e-01 3.27219695e-01 5.46677053e-01
2.85420120e-01 -6.77630842e-01 4.47626859e-01 -2.63203770e-01
8.14119458e-01 -4.08517420e-01 -3.94437402e-01 9.11988139e-01
-7.93594867e-02 -1.52766943e-01 3.80356520e-01 -8.17645490e-01
-1.11273694e+00 -1.54919280e-02 5.98617196e-02 -1.82286292e-01
-3.79293442e-01 8.88959110e-01 -6.00387692e-01 -2.71893777e-02
3.89129341e-01 9.43311080e-02 1.07726149e-01 -4.77127403e-01
8.45297039e-01 8.46945465e-01 5.96142769e-01 -9.18780029e-01
6.76376939e-01 1.13241434e-01 -4.97758389e-01 -5.77970445e-01
-2.83698142e-01 -6.53274953e-02 -1.73632458e-01 -5.41921496e-01
8.72694552e-01 -7.78663635e-01 -1.24470329e+00 5.44171989e-01
-1.21270621e+00 -8.57420087e-01 1.16940007e-01 6.74575120e-02
-5.34788728e-01 8.33333358e-02 -2.15913132e-01 -6.62177444e-01
-6.94554210e-01 -1.38324106e+00 1.33484018e+00 5.18468678e-01
-5.06236196e-01 -7.37777054e-01 -7.93541688e-03 7.95596242e-01
1.00847423e+00 -4.13215272e-02 1.15238857e+00 -3.18865299e-01
-7.33798862e-01 -1.85655162e-01 -1.22177482e-01 -9.55063477e-02
-4.06960882e-02 5.13628602e-01 -7.81866372e-01 -8.93215016e-02
-6.80037439e-01 -5.88670909e-01 1.44163936e-01 -2.13977620e-01
9.63871896e-01 -3.56958926e-01 -2.60157287e-01 5.74189484e-01
9.37991261e-01 4.50902909e-01 4.50103045e-01 3.42765659e-01
7.41027832e-01 5.44509113e-01 5.37807763e-01 5.17749429e-01
9.56950784e-01 9.16350901e-01 6.57514989e-01 -9.70219001e-02
2.29106545e-02 9.86398477e-03 4.29840297e-01 7.86282241e-01
6.33335635e-02 -3.22314471e-01 -1.18480814e+00 5.08538544e-01
-2.21481061e+00 -8.76954198e-01 2.54842609e-01 1.88881958e+00
7.94563711e-01 -5.91520011e-01 2.11173505e-01 -5.00672758e-01
3.67788941e-01 -2.52543241e-01 -5.06549776e-01 -4.31629926e-01
1.67293221e-01 -3.20735723e-01 -5.27948260e-01 2.81213194e-01
-7.17209637e-01 9.61433291e-01 6.42859745e+00 8.07915509e-01
-1.28275514e+00 1.22788921e-01 2.67408401e-01 -4.06753540e-01
-6.63908303e-01 1.63924918e-02 -4.34807569e-01 4.94758874e-01
9.60171580e-01 -3.35061580e-01 9.77378726e-01 8.29773605e-01
5.72254322e-02 1.00486264e-01 -1.23638785e+00 1.17384362e+00
7.82655925e-02 -1.52278399e+00 2.05162361e-01 -2.76730061e-01
5.25943279e-01 -4.21396680e-02 1.96323637e-02 3.95142883e-01
5.95316216e-02 -9.61574614e-01 1.03231180e+00 8.04993629e-01
7.84714043e-01 -5.24938703e-01 1.24005012e-01 3.58515203e-01
-1.01821721e+00 -1.43369555e-01 2.89220065e-01 1.23416744e-01
3.39085877e-01 -1.10782392e-01 -4.57897484e-01 5.37109554e-01
8.45598102e-01 1.98914170e-01 -5.32643855e-01 9.79703605e-01
2.10744113e-01 9.43004712e-02 -2.99284130e-01 -3.00077498e-01
-1.31655350e-01 -1.60449699e-01 5.04231751e-01 1.22465861e+00
4.42000359e-01 7.88415130e-03 1.38896406e-01 7.42340088e-01
6.90774843e-02 1.36545628e-01 -5.61719537e-01 -5.30358315e-01
4.74673569e-01 1.54142082e+00 -1.96611136e-01 -4.84399319e-01
-6.26973569e-01 9.20469582e-01 1.33102521e-01 6.02109909e-01
-9.85556960e-01 -4.97910976e-01 7.54468799e-01 -1.20659731e-01
-5.82873300e-02 -4.87462908e-01 -4.80683111e-02 -1.41947067e+00
-7.54920319e-02 -1.64983022e+00 3.43328893e-01 -1.41991937e+00
-1.10742736e+00 8.21060538e-01 2.91561544e-01 -9.78155017e-01
-5.28689623e-01 -4.69177186e-01 -6.22530937e-01 4.25170481e-01
-9.59456861e-01 -1.80345702e+00 -8.99220884e-01 8.59509170e-01
4.45899338e-01 -5.13640821e-01 1.22676635e+00 3.04228783e-01
-7.88429260e-01 4.91992652e-01 -2.51613110e-01 -2.05499068e-01
7.65343249e-01 -9.06568229e-01 3.04134816e-01 3.87519300e-01
4.00421433e-02 7.47268558e-01 5.93162656e-01 -8.21446031e-02
-2.26878238e+00 -6.21769607e-01 4.94224042e-01 -4.77586538e-01
6.28426850e-01 -4.90531385e-01 -6.20868385e-01 6.89548373e-01
6.90816283e-01 -4.88434047e-01 8.79504502e-01 -7.81897232e-02
-5.32008946e-01 -3.83571088e-01 -8.21482182e-01 1.08904231e+00
1.12861454e+00 -5.69258273e-01 -1.51525229e-01 3.67137700e-01
8.15215230e-01 -3.31519544e-01 -9.72093642e-01 3.15394461e-01
6.51800692e-01 -6.16979897e-01 9.13485110e-01 -4.60342944e-01
4.88930762e-01 -5.40394485e-01 -1.58392414e-01 -1.10505009e+00
5.69791999e-03 -1.14822936e+00 -6.15884475e-02 1.42446494e+00
8.01059544e-01 -5.34337938e-01 2.47454643e-01 1.36904299e+00
-3.81042659e-01 -4.18441981e-01 -4.46654171e-01 -3.64269733e-01
-5.62790811e-01 -6.70161068e-01 1.06390321e+00 7.37309575e-01
5.84510207e-01 5.31603396e-01 -2.55430937e-01 -3.45292911e-02
2.72457540e-01 4.08680588e-01 1.46721900e+00 -9.12730634e-01
-7.31366694e-01 -6.33719981e-01 -3.95882465e-02 -1.14574873e+00
-1.38825059e-01 -7.44211257e-01 -5.05219772e-02 -1.82264388e+00
2.16625795e-01 -2.93945253e-01 1.64028272e-01 9.57038462e-01
1.33705780e-01 5.75761870e-02 6.33093297e-01 3.40927392e-01
-1.17963028e+00 4.70618427e-01 1.36067760e+00 -5.19467890e-01
-3.18680912e-01 -3.17313820e-01 -8.35935175e-01 6.33479536e-01
6.83938563e-01 6.20977618e-02 -6.84911191e-01 -9.99271870e-01
4.93012518e-01 2.32532904e-01 3.99887294e-01 -7.95014262e-01
3.72919589e-01 -5.49683392e-01 -2.06475526e-01 -2.42049918e-01
7.83661783e-01 -7.24776685e-01 6.00163639e-01 -1.55884027e-01
-4.04850900e-01 1.66493654e-01 2.79161066e-01 4.02049944e-02
-2.56576270e-01 -1.83421955e-01 8.55236277e-02 -2.66611993e-01
-9.65567172e-01 2.34813318e-01 -5.26268125e-01 -2.19770655e-01
1.05278802e+00 1.20704710e-01 -7.66935229e-01 -7.87814617e-01
-7.14434206e-01 5.78817189e-01 5.93063891e-01 8.44246864e-01
6.73139453e-01 -1.16323769e+00 -5.45748830e-01 6.71159849e-02
7.07338750e-01 -2.32860193e-01 3.37202221e-01 6.79790556e-01
-3.88251811e-01 3.17132533e-01 -2.49235794e-01 -5.98710597e-01
-1.50278509e+00 4.47299182e-01 3.09016436e-01 1.20092988e-01
-2.64259964e-01 5.99905550e-01 1.98883533e-01 -7.22027004e-01
4.90920126e-01 -1.42831266e-01 9.15560499e-02 -1.20754957e-01
8.85987580e-01 6.56590378e-03 -2.65161753e-01 -3.51033896e-01
-4.29187059e-01 4.59456772e-01 2.22108871e-01 -5.16223669e-01
1.17590964e+00 -2.95765489e-01 -4.79948670e-01 3.57108861e-01
7.91236222e-01 -4.93492872e-01 -1.10138249e+00 3.87280397e-02
-4.23821449e-01 -2.14643791e-01 -4.11568850e-01 -1.35523391e+00
-8.25323105e-01 7.57062614e-01 2.44875818e-01 2.47656509e-01
1.11475480e+00 1.27461672e-01 6.24239922e-01 6.54091537e-01
5.63336551e-01 -9.97093856e-01 3.47925603e-01 3.72204214e-01
1.28560793e+00 -1.06452382e+00 -6.07608974e-01 -2.64018942e-02
-1.06342614e+00 1.07366729e+00 7.33163178e-01 7.78284073e-01
1.65350065e-01 3.21140856e-01 4.98648226e-01 -2.13158742e-01
-1.38870728e+00 -2.58854359e-01 4.43161547e-01 8.52198660e-01
3.81014913e-01 -1.41329532e-02 7.94956833e-02 6.67755961e-01
2.32943282e-01 6.33634776e-02 9.90612283e-02 9.70167041e-01
-2.80313402e-01 -1.36977494e+00 -4.89940792e-01 -9.40701291e-02
2.05938652e-01 3.46413441e-02 -5.77132165e-01 6.38103127e-01
-3.95491123e-02 1.24395168e+00 -1.16538778e-01 -5.24108708e-01
2.46205106e-01 3.02108675e-01 5.15353024e-01 -3.62368584e-01
-7.83526003e-01 -4.92492616e-02 4.12472725e-01 -7.61817932e-01
-3.18547308e-01 -2.77223766e-01 -1.18454587e+00 -3.64284307e-01
1.46977752e-01 -1.29611358e-01 1.20940888e+00 9.58021462e-01
9.56112921e-01 2.98670292e-01 2.11776257e-01 -1.13784730e+00
-5.23784384e-02 -6.58616841e-01 9.25818905e-02 3.67748678e-01
-3.60185243e-02 -2.50314564e-01 -3.90388295e-02 2.78497964e-01] | [10.865217208862305, 1.5214661359786987] |
33d5c348-ceec-4915-9edf-71ee6940e0db | the-ai-mechanic-acoustic-vehicle | 2205.09667 | null | https://arxiv.org/abs/2205.09667v1 | https://arxiv.org/pdf/2205.09667v1.pdf | The AI Mechanic: Acoustic Vehicle Characterization Neural Networks | In a world increasingly dependent on road-based transportation, it is essential to understand vehicles. We introduce the AI mechanic, an acoustic vehicle characterization deep learning system, as an integrated approach using sound captured from mobile devices to enhance transparency and understanding of vehicles and their condition for non-expert users. We develop and implement novel cascading architectures for vehicle understanding, which we define as sequential, conditional, multi-level networks that process raw audio to extract highly-granular insights. To showcase the viability of cascading architectures, we build a multi-task convolutional neural network that predicts and cascades vehicle attributes to enhance fault detection. We train and test these models on a synthesized dataset reflecting more than 40 hours of augmented audio and achieve >92% validation set accuracy on attributes (fuel type, engine configuration, cylinder count and aspiration type). Our cascading architecture additionally achieved 93.6% validation and 86.8% test set accuracy on misfire fault prediction, demonstrating margins of 16.4% / 7.8% and 4.2% / 1.5% improvement over na\"ive and parallel baselines. We explore experimental studies focused on acoustic features, data augmentation, feature fusion, and data reliability. Finally, we conclude with a discussion of broader implications, future directions, and application areas for this work. | ['Joshua E. Siegel', 'Adam M. Terwilliger'] | 2022-05-19 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [ 1.01029783e-01 -5.27401902e-02 1.70494542e-01 -5.22046387e-01
-1.28890765e+00 -4.38165009e-01 4.21936065e-01 -2.58602947e-02
5.36230840e-02 3.26729506e-01 1.37492701e-01 -6.57631814e-01
9.79439020e-02 -8.82356644e-01 -1.15418327e+00 -3.62510145e-01
-2.80355036e-01 8.31201226e-02 1.79637581e-01 -1.46354705e-01
-9.44008157e-02 4.59061563e-01 -1.93016732e+00 7.32905269e-01
5.46171129e-01 1.50871480e+00 -4.47353899e-01 1.05715930e+00
3.65889728e-01 7.45879114e-01 -7.24151731e-01 -2.21009701e-01
6.65471926e-02 2.84288853e-01 -3.82537782e-01 -3.46921980e-01
5.70183158e-01 -4.92848247e-01 -6.00975811e-01 5.76018512e-01
4.88571227e-01 -1.36519521e-01 6.47236228e-01 -1.94776940e+00
-3.28240842e-01 5.87542236e-01 -5.37815392e-02 2.88854390e-01
1.66872174e-01 7.83653557e-01 8.67435873e-01 -9.94223595e-01
-2.45649561e-01 9.81480539e-01 1.25163722e+00 3.92694265e-01
-1.00000060e+00 -1.34291077e+00 -1.73932359e-01 5.90821326e-01
-1.34343517e+00 -1.07382727e+00 7.35468447e-01 -6.17858112e-01
1.40214312e+00 2.32279345e-01 4.55845177e-01 1.35713613e+00
5.84720135e-01 5.99826634e-01 6.12983167e-01 1.17404081e-01
1.92730844e-01 9.43708867e-02 1.58261552e-01 5.45817852e-01
1.81292266e-01 3.51180166e-01 -8.09847474e-01 -1.06818654e-01
1.20425247e-01 -2.37053335e-01 3.29964012e-01 3.58574957e-01
-7.16055512e-01 5.67539692e-01 1.48105487e-01 -2.81326264e-01
-3.20104241e-01 7.47754872e-01 4.82431889e-01 2.21454516e-01
4.20369744e-01 2.75859416e-01 -5.86088836e-01 -3.19405615e-01
-6.65307641e-01 3.21495891e-01 6.24592841e-01 8.12915385e-01
7.30413973e-01 8.70797276e-01 1.80271968e-01 6.90952003e-01
6.72906876e-01 1.16613996e+00 -1.16327532e-01 -1.15102303e+00
4.54008132e-01 2.96188414e-01 -9.49929506e-02 -1.03776729e+00
-8.12703252e-01 -4.10911351e-01 -5.78284323e-01 2.53826827e-01
-2.80590951e-01 -5.21736264e-01 -1.03517926e+00 1.56984508e+00
-7.98597038e-02 7.33936310e-01 2.60621250e-01 4.80434060e-01
1.04870033e+00 8.71636093e-01 3.23346317e-01 2.84042656e-01
1.04069030e+00 -6.15626276e-01 -6.15531564e-01 -3.43557954e-01
5.75272441e-01 -6.03596866e-01 8.77904236e-01 4.64494616e-01
-9.35485482e-01 -8.53649259e-01 -1.34006202e+00 1.98292360e-01
-5.81631720e-01 9.04023573e-02 4.87417400e-01 9.37635303e-01
-9.85872328e-01 3.19720298e-01 -9.41401660e-01 1.43428072e-01
6.48924410e-01 5.44198155e-01 -6.75998107e-02 1.27140179e-01
-1.48441720e+00 6.03233993e-01 -2.80378699e-01 3.06083828e-01
-1.63993549e+00 -1.25529814e+00 -9.30021703e-01 9.30051506e-02
1.04132354e-01 -5.61336458e-01 1.51678324e+00 -3.07422578e-01
-1.33289659e+00 -2.39241705e-03 -1.30174190e-01 -8.34607065e-01
1.16097666e-01 -4.67739165e-01 -1.16099632e+00 -1.93678904e-02
1.21937841e-01 5.46638906e-01 7.19862401e-01 -1.21606719e+00
-1.08290148e+00 2.26841718e-02 -3.07951808e-01 -3.70112985e-01
-1.91100597e-01 -1.28861189e-01 -6.41415967e-03 -3.29459429e-01
-3.33553642e-01 -8.26684117e-01 1.41897485e-01 -3.60618502e-01
-2.13666677e-01 -2.68925935e-01 1.29946935e+00 -8.04426670e-01
1.14894938e+00 -2.19266129e+00 -7.39160299e-01 3.24379295e-01
3.78987342e-01 2.91080236e-01 -1.42772809e-01 3.40723038e-01
-2.70027034e-02 3.50973129e-01 9.80870333e-03 -3.16334367e-01
1.20176002e-02 -1.14251368e-01 -3.95325810e-01 2.88621992e-01
8.70077372e-01 8.55845630e-01 -5.93400836e-01 4.40001488e-02
4.39974338e-01 6.66458309e-01 -7.24181056e-01 -6.36278372e-03
3.16219195e-03 -1.63059114e-04 -3.28494281e-01 9.64937925e-01
4.80925590e-01 2.52944678e-01 -5.37383020e-01 -5.07909775e-01
-8.90665315e-03 5.50769687e-01 -1.09588659e+00 1.04958344e+00
-8.37644398e-01 1.17509186e+00 5.03968783e-02 -7.22418129e-01
8.24652135e-01 3.48263532e-01 5.73839188e-01 -9.45328772e-01
1.47582695e-01 1.79818481e-01 -9.86013412e-02 -8.16958070e-01
4.78866339e-01 7.88604543e-02 -5.55965126e-01 1.72876701e-01
-2.38871351e-02 -2.34487519e-01 -4.01498854e-01 1.11340687e-01
1.42245173e+00 -5.34343302e-01 -5.08054554e-01 -3.32429335e-02
1.75476983e-01 2.78561823e-02 4.36370850e-01 5.65568089e-01
-2.71315962e-01 3.11496526e-01 2.21137270e-01 -3.53335649e-01
-9.68619108e-01 -1.09282577e+00 -1.67823806e-01 1.06037009e+00
-1.06984042e-01 -4.29879963e-01 -4.99201626e-01 -4.09217238e-01
4.88548368e-01 9.69589949e-01 -6.78644419e-01 -8.35179448e-01
-4.81888801e-01 -6.84517860e-01 1.21985555e+00 1.01426804e+00
4.78936255e-01 -5.22509277e-01 -2.93052524e-01 2.53872126e-01
-1.08559415e-01 -1.17945504e+00 1.17295489e-01 1.37048438e-01
-1.18632771e-01 -8.15187931e-01 3.47353578e-01 -5.54377437e-01
-1.74933136e-01 2.67190486e-01 9.65095222e-01 -2.54051876e-03
3.05507611e-03 5.52381694e-01 -8.69956315e-02 -9.38766778e-01
-7.04457462e-01 -2.09133193e-01 6.60655439e-01 7.67406598e-02
3.06385577e-01 -4.89906162e-01 -5.10930777e-01 5.20757973e-01
-4.41845685e-01 -5.62845320e-02 5.64404070e-01 4.15650219e-01
3.28218728e-01 1.79460511e-01 1.19715774e+00 -1.99241519e-01
3.39551419e-01 -9.13554370e-01 -2.73566276e-01 -1.60151720e-01
-8.03100228e-01 -2.64606148e-01 4.40018743e-01 -1.97910108e-02
-9.15655196e-01 6.51022012e-04 -6.01279855e-01 -3.97869796e-01
-4.00592685e-01 3.28742385e-01 -3.42310250e-01 -2.97824871e-02
6.35976136e-01 -1.38358444e-01 3.15848663e-02 -1.22501999e-01
2.84230769e-01 1.23816693e+00 8.76185358e-01 -3.05884272e-01
6.77712858e-01 3.30784738e-01 -1.65845379e-01 -9.09058332e-01
-9.07148123e-02 -1.41023248e-01 -1.10466443e-01 -6.03903234e-01
7.83369184e-01 -1.36283779e+00 -1.01037717e+00 6.71857834e-01
-1.08455741e+00 -6.14073992e-01 -1.18278131e-01 3.85762066e-01
-4.08616275e-01 -2.66695350e-01 -5.12442708e-01 -9.44450259e-01
-2.91464329e-01 -1.23016107e+00 1.14488578e+00 2.39112154e-02
-2.54175842e-01 -4.92360651e-01 -1.91159964e-01 6.90976501e-01
6.60229325e-01 1.16011724e-01 8.18101883e-01 -8.79022539e-01
-5.80260754e-01 -3.07067931e-01 -1.32548183e-01 4.85011458e-01
-3.81754600e-02 3.95982802e-01 -1.44445777e+00 1.29961938e-01
-4.69777435e-01 -2.65195400e-01 8.64644945e-01 3.02432239e-01
1.10051847e+00 -1.84352964e-01 -3.92239571e-01 3.15541953e-01
8.41702163e-01 4.38068122e-01 3.67895812e-01 6.02176972e-02
7.58049071e-01 4.30846125e-01 3.48483354e-01 2.90455550e-01
7.68102825e-01 4.41499054e-01 5.66815019e-01 1.36764899e-01
-6.81117237e-01 -2.51589984e-01 6.60145164e-01 7.86324263e-01
3.77837569e-01 -4.46477830e-01 -1.37194061e+00 6.22332394e-01
-1.39398921e+00 -7.74563670e-01 -3.83140028e-01 1.67788792e+00
3.15002739e-01 5.60419381e-01 8.56353417e-02 5.12992918e-01
4.54747289e-01 -2.70483613e-01 -7.29316711e-01 -5.32944441e-01
8.64911303e-02 -1.46752363e-02 5.29509366e-01 4.68773216e-01
-1.22563553e+00 8.64531755e-01 6.94265032e+00 6.50913417e-01
-1.20132387e+00 7.41840303e-02 8.87795448e-01 -2.30181545e-01
-4.38707799e-01 -6.65980995e-01 -7.75350094e-01 3.98973405e-01
1.88544571e+00 1.28850698e-01 3.25722098e-01 6.63345873e-01
5.18214881e-01 6.36307970e-02 -1.13049459e+00 7.67551899e-01
4.57190871e-02 -1.57755327e+00 -3.50739062e-02 -1.33083239e-01
6.85241282e-01 5.58439016e-01 4.09909010e-01 6.55361116e-01
3.19180846e-01 -1.18692970e+00 9.92572427e-01 5.90181112e-01
9.99034047e-01 -1.03973210e+00 8.71704757e-01 -1.83727339e-01
-1.40278649e+00 -5.02287447e-01 3.44320416e-01 -1.41223475e-01
1.89916492e-01 3.28365833e-01 -1.08713937e+00 3.47560555e-01
1.16997290e+00 4.98062134e-01 -7.69372523e-01 7.41168141e-01
3.03357989e-01 1.25196660e+00 -5.65550864e-01 -5.86798601e-02
-8.46360996e-03 7.10756540e-01 3.72669667e-01 1.43800700e+00
3.13268870e-01 -3.60889211e-02 -1.68427393e-01 5.85158408e-01
-5.83771542e-02 -5.19583166e-01 -7.38255680e-01 3.43344323e-02
8.76412868e-01 1.07956839e+00 -1.05721653e-01 -3.00124258e-01
-5.70443571e-01 1.08747609e-01 -3.72145802e-01 3.92367005e-01
-1.31587684e+00 -4.32074726e-01 1.16352713e+00 1.24899760e-01
1.13220826e-01 -2.09704250e-01 -8.28348219e-01 -3.24159592e-01
-1.63394004e-01 -8.43822360e-01 9.35403034e-02 -8.39842856e-01
-1.09111762e+00 4.69588101e-01 -6.16925396e-02 -1.13412559e+00
-5.27267396e-01 -6.04888916e-01 -6.99339330e-01 4.27908659e-01
-1.35674429e+00 -1.26130271e+00 -3.31889063e-01 3.57469708e-01
7.58885860e-01 -4.79884505e-01 5.28395176e-01 7.12432086e-01
-7.77879238e-01 1.02305460e+00 -2.29359102e-02 1.21435247e-01
5.20255566e-01 -8.01788926e-01 8.14731061e-01 7.93102801e-01
-1.21014416e-01 7.86065757e-02 7.65601993e-01 -7.04906821e-01
-1.83662415e+00 -1.67411184e+00 5.26193023e-01 -9.13701177e-01
9.61083710e-01 -5.17607391e-01 -7.85187900e-01 6.76963449e-01
1.16806393e-02 -7.22122565e-02 8.18816006e-01 -2.24137655e-03
-3.70308667e-01 -8.15969288e-01 -1.10314822e+00 3.25103402e-01
8.46395433e-01 -7.75567234e-01 -1.57622233e-01 -2.52608210e-02
9.43590820e-01 -1.74181193e-01 -9.33602452e-01 8.95258069e-01
8.47162127e-01 -5.94754517e-01 8.86716366e-01 -5.69313407e-01
3.15463632e-01 -3.40304434e-01 -4.26329762e-01 -1.15026355e+00
-1.88552260e-01 -5.75826466e-01 -2.62180507e-01 1.32908881e+00
8.42222333e-01 -5.61906993e-01 5.70122778e-01 7.33692765e-01
-9.39121842e-01 -7.83386230e-01 -1.13353050e+00 -4.75183964e-01
1.20028265e-01 -1.88437676e+00 8.55277598e-01 3.35727751e-01
-2.29685798e-01 3.42537403e-01 -3.03504616e-01 7.95246422e-01
3.06253999e-01 -4.18356746e-01 8.82367253e-01 -1.02132213e+00
1.33732766e-01 -5.03006697e-01 -6.40392661e-01 -5.78160107e-01
9.81608778e-02 -5.59243381e-01 2.99622029e-01 -1.29613137e+00
-2.18299225e-01 -5.86819768e-01 -5.26011586e-01 6.37167811e-01
2.33010516e-01 5.85039318e-01 -1.23691119e-01 -2.02803358e-01
-5.46853185e-01 5.02335548e-01 3.04653496e-01 -5.17298162e-01
-5.24570197e-02 1.97636575e-01 -8.02518547e-01 6.63544416e-01
1.04806566e+00 -2.69594580e-01 -3.10577780e-01 -6.68570399e-01
2.84905642e-01 -1.57351270e-01 8.69129121e-01 -1.53838050e+00
3.97993296e-01 1.55707207e-02 5.13362110e-01 -7.33544052e-01
4.59396005e-01 -7.34778583e-01 2.28560999e-01 2.95073003e-01
-2.46426702e-01 1.55684561e-01 8.66815090e-01 7.03520536e-01
1.08017839e-01 6.46830797e-01 1.92940354e-01 6.76126540e-01
-9.11496162e-01 1.49296999e-01 -9.42801178e-01 -4.09553796e-01
9.51695323e-01 6.33997694e-02 -4.67488438e-01 -4.81447816e-01
-4.48998898e-01 5.70736289e-01 -2.42191181e-01 9.56158400e-01
8.68177533e-01 -1.56892276e+00 -8.43484282e-01 4.97705072e-01
2.26227731e-01 -2.04521909e-01 5.41595757e-01 4.55355823e-01
-7.57820606e-02 5.94247162e-01 2.31020655e-02 -8.37157309e-01
-1.04621351e+00 1.43890262e-01 4.48199242e-01 5.79320014e-01
-1.56793624e-01 8.06991518e-01 -1.00973710e-01 -2.37629339e-01
2.66190648e-01 -7.01233447e-01 -9.30202603e-02 -1.07052743e-01
5.41161835e-01 8.23882580e-01 5.30416548e-01 -7.60616362e-01
-6.46084309e-01 3.00870925e-01 3.31794769e-01 -9.00486857e-02
1.15177202e+00 -1.97173670e-01 5.73555529e-01 5.59578419e-01
1.30900955e+00 -2.15412546e-02 -1.48333621e+00 2.03823894e-01
-3.55038524e-01 2.06971034e-01 3.48087698e-01 -1.21068239e+00
-1.13644195e+00 9.02375042e-01 1.03967905e+00 3.98011297e-01
9.29962873e-01 2.99398363e-01 1.05806816e+00 4.86820877e-01
2.82587893e-02 -1.05896747e+00 -3.64706069e-02 4.20674205e-01
9.49156225e-01 -1.28686571e+00 -4.17492628e-01 -7.56838396e-02
-3.79433036e-01 1.07132792e+00 8.11858356e-01 6.81297481e-02
8.85754526e-01 8.56142402e-01 9.50717852e-02 -2.53138155e-01
-1.09048712e+00 3.79293486e-02 2.72299081e-01 6.56162202e-01
-1.68545291e-01 2.80636549e-01 7.30044961e-01 1.20909119e+00
-5.23591936e-01 -1.59770116e-01 2.64975637e-01 6.41806722e-01
-5.22280216e-01 -2.73634106e-01 -3.31741631e-01 7.85617054e-01
-1.19603045e-01 6.90813884e-02 -1.60153508e-01 6.01243377e-01
4.77218211e-01 1.73347092e+00 3.86045605e-01 -1.48273873e+00
6.50102377e-01 1.85351565e-01 -2.08688885e-01 -2.59606957e-01
-4.14757937e-01 -1.29270658e-01 6.05715275e-01 -6.83643878e-01
3.87133919e-02 -8.89947653e-01 -1.52715862e+00 -5.17692447e-01
-2.03284279e-01 -9.51614752e-02 9.23865616e-01 9.65828300e-01
7.81688690e-01 1.05815625e+00 9.83856559e-01 -7.78594613e-01
-1.46852106e-01 -8.99299264e-01 1.07497685e-02 -2.54095197e-01
8.75847042e-01 -7.07129359e-01 -4.29933369e-01 1.09771959e-01] | [15.103267669677734, 5.219354629516602] |
ae007875-8bf3-48cb-a95a-6f09982847c4 | graph-based-neural-sentence-ordering | 1912.07225 | null | https://arxiv.org/abs/1912.07225v1 | https://arxiv.org/pdf/1912.07225v1.pdf | Graph-based Neural Sentence Ordering | Sentence ordering is to restore the original paragraph from a set of sentences. It involves capturing global dependencies among sentences regardless of their input order. In this paper, we propose a novel and flexible graph-based neural sentence ordering model, which adopts graph recurrent network \cite{Zhang:acl18} to accurately learn semantic representations of the sentences. Instead of assuming connections between all pairs of input sentences, we use entities that are shared among multiple sentences to make more expressive graph representations with less noise. Experimental results show that our proposed model outperforms the existing state-of-the-art systems on several benchmark datasets, demonstrating the effectiveness of our model. We also conduct a thorough analysis on how entities help the performance. | ['Linfeng Song', 'Jiebo Luo', 'Jiali Zeng', 'Yongjing Yin', 'Jinsong Su', 'Chulun Zhou'] | 2019-12-16 | null | null | null | null | ['sentence-ordering'] | ['natural-language-processing'] | [ 2.10291654e-01 4.65750784e-01 -1.16878405e-01 -6.67931557e-01
-1.32516563e-01 -4.16171402e-01 3.34225655e-01 3.89232486e-01
-3.20249856e-01 7.34575212e-01 7.87432790e-01 -4.24077988e-01
-1.93752080e-01 -9.16388810e-01 -9.14927304e-01 7.81988055e-02
8.77642334e-02 3.02123547e-01 4.26953256e-01 -4.08697516e-01
2.35496953e-01 9.66221169e-02 -9.36889231e-01 4.79873389e-01
8.24280083e-01 3.71831089e-01 3.61927241e-01 6.25310838e-01
-4.32348877e-01 1.29376280e+00 -8.47160399e-01 -8.59748721e-01
-1.24980994e-01 -7.09500790e-01 -1.18665981e+00 5.30332364e-02
3.52748156e-01 -3.87123853e-01 -9.63395655e-01 1.28914869e+00
2.08414525e-01 3.37022543e-01 4.50080097e-01 -8.70451629e-01
-1.41109097e+00 1.29154456e+00 -3.89906168e-01 4.26877618e-01
6.12275243e-01 -4.19336230e-01 1.48535800e+00 -5.70561230e-01
7.31006920e-01 1.26343763e+00 5.42790234e-01 3.56691033e-01
-7.61693180e-01 -4.79792356e-01 6.52069926e-01 5.61377823e-01
-1.18704844e+00 -3.18376511e-01 1.08772528e+00 1.01909399e-01
1.46023643e+00 2.61250436e-01 3.98690760e-01 9.03362095e-01
4.28964764e-01 6.84079647e-01 4.11784708e-01 -4.23532188e-01
-8.62278044e-02 -2.69068241e-01 9.71518099e-01 7.58802593e-01
5.64424098e-01 -8.42837453e-01 -4.70882893e-01 2.18595713e-02
3.84030938e-01 1.53904960e-01 -4.72956270e-01 7.90803693e-03
-8.44642401e-01 7.07915187e-01 6.85064197e-01 4.13614511e-01
-2.82508641e-01 2.46469423e-01 6.02900803e-01 4.93401527e-01
5.97960889e-01 2.86390454e-01 -3.27226788e-01 2.84418404e-01
-4.71380144e-01 -9.09210443e-02 8.43403399e-01 1.15727186e+00
5.88247597e-01 -1.96915448e-01 -2.81552017e-01 7.84912765e-01
3.82613182e-01 2.26733517e-02 4.00177538e-01 -4.95752901e-01
1.03169584e+00 9.63104010e-01 -3.20116252e-01 -1.54843497e+00
-5.09055853e-01 -5.08623660e-01 -1.12306595e+00 -1.00355399e+00
-2.58532315e-01 -6.99570924e-02 -7.78509140e-01 1.61972106e+00
-5.99651486e-02 3.77443105e-01 2.81452328e-01 6.07725203e-01
1.38247490e+00 7.01476097e-01 -4.18684930e-02 -2.22912043e-01
1.03636110e+00 -1.05742967e+00 -9.34400558e-01 -4.57496792e-01
7.25246072e-01 -3.82827431e-01 1.02263570e+00 -1.85570225e-01
-8.72612894e-01 -5.44711232e-01 -1.12270689e+00 -3.87226731e-01
-3.53641212e-01 5.69960615e-03 6.72677994e-01 1.81029662e-01
-1.38798225e+00 8.97085607e-01 -5.02649546e-01 -5.75974047e-01
1.82876945e-01 3.65720987e-01 -4.47845757e-01 -1.51036412e-01
-1.77375090e+00 7.34445989e-01 6.96480334e-01 4.69557524e-01
-1.74184904e-01 -3.14267486e-01 -1.14962327e+00 3.40250790e-01
4.03561175e-01 -1.11702454e+00 1.02958941e+00 -5.91145575e-01
-1.31024623e+00 7.16077864e-01 -5.60416877e-01 -6.83428884e-01
3.20249833e-02 -3.27672809e-01 -5.07588983e-01 3.94360125e-01
1.59203768e-01 2.77452677e-01 3.73066962e-01 -9.28050101e-01
-1.30195588e-01 -2.02988699e-01 5.79736531e-01 4.53192383e-01
-5.23795366e-01 1.64801795e-02 -7.69383550e-01 -5.26810944e-01
2.09158748e-01 -6.71229780e-01 -1.96280465e-01 -7.96853483e-01
-9.90667343e-01 -5.96769989e-01 5.07924914e-01 -9.04594123e-01
1.59621954e+00 -1.77937257e+00 2.25883693e-01 -6.09483570e-02
4.47532713e-01 1.32225100e-02 -2.01569751e-01 9.72582877e-01
-1.45237729e-01 5.84201455e-01 -2.45809257e-01 -5.56834936e-01
-1.32966312e-02 2.66346037e-01 -4.87484455e-01 1.27953216e-01
8.05917904e-02 1.20787966e+00 -9.23801482e-01 -5.55200219e-01
-6.70887977e-02 2.94364631e-01 -1.98005781e-01 1.70626789e-01
-3.50456387e-02 5.62164336e-02 -4.18512821e-01 -2.26233490e-02
4.32015926e-01 -6.47496760e-01 5.69885254e-01 -1.87034860e-01
6.66609764e-01 7.39654541e-01 -8.09280813e-01 1.81721258e+00
-1.87829226e-01 4.00452524e-01 -5.73759079e-01 -8.94956052e-01
1.06329286e+00 1.43303022e-01 -5.74882887e-02 -7.11977482e-01
1.38301507e-01 -2.30098888e-01 -2.93038767e-02 -6.17783725e-01
6.94744527e-01 2.09999785e-01 -2.52639771e-01 4.38752264e-01
1.31669119e-01 6.23951964e-02 5.87334335e-01 9.86046791e-01
1.25013840e+00 -2.47157514e-01 3.80484730e-01 -1.24495737e-01
6.38940096e-01 -4.21885580e-01 5.32745600e-01 8.96564960e-01
5.66278249e-02 5.71145713e-01 9.34281886e-01 -3.51596922e-01
-7.85083950e-01 -9.00798976e-01 4.64319378e-01 8.53282869e-01
4.30079341e-01 -8.52985859e-01 -7.08333015e-01 -1.08881211e+00
-1.93775609e-01 1.00540328e+00 -7.07632124e-01 -3.86182755e-01
-7.56512582e-01 -5.69742739e-01 4.98711646e-01 6.47771060e-01
5.76427817e-01 -1.07678318e+00 2.15754658e-01 8.76434445e-02
-4.65382427e-01 -1.45959938e+00 -7.18138576e-01 -2.74044961e-01
-8.10615242e-01 -1.23182452e+00 -1.25721872e-01 -1.12327695e+00
9.25543129e-01 4.38543618e-01 1.28786242e+00 5.11019886e-01
3.83176833e-01 1.76485464e-01 -5.98910034e-01 -7.70969037e-03
-2.61004478e-01 3.94139171e-01 -1.07060514e-01 -1.40037434e-02
3.01594436e-01 -6.41438365e-01 -4.31027591e-01 -2.07556978e-01
-9.88736868e-01 1.44226417e-01 4.37039793e-01 7.29748487e-01
3.35084289e-01 1.31867617e-01 8.35258603e-01 -1.58106732e+00
1.25015283e+00 -5.30293286e-01 6.43240884e-02 6.73647702e-01
-3.49217474e-01 1.72636837e-01 1.08827686e+00 3.61527242e-02
-9.06327128e-01 -3.08169782e-01 -5.48580550e-02 -2.42713735e-01
2.04181671e-01 9.67787623e-01 -1.72212392e-01 4.12575215e-01
1.49709389e-01 3.29034120e-01 -3.45212847e-01 -3.65671307e-01
4.34500694e-01 5.28947294e-01 5.95376194e-01 -1.85983449e-01
5.74667335e-01 1.29541159e-01 -1.19432144e-01 -6.58476174e-01
-1.09728312e+00 -3.14228445e-01 -8.22927773e-01 -1.71693917e-02
6.83927655e-01 -7.85149574e-01 -5.47585905e-01 2.63213456e-01
-1.73361278e+00 1.01200461e-01 8.50441158e-02 1.95662290e-01
-3.40198092e-02 9.22686279e-01 -9.76722956e-01 -3.59584630e-01
-6.51334703e-01 -7.75151014e-01 6.29422605e-01 3.21406811e-01
-3.64778727e-01 -1.32578635e+00 -1.17701560e-01 3.36840630e-01
2.14153584e-02 -3.86190601e-02 1.02260232e+00 -1.01143157e+00
-3.93051803e-01 -1.82513148e-01 -3.93877715e-01 2.54632801e-01
4.97677177e-01 -4.90112975e-02 -4.49134082e-01 -3.22813779e-01
-1.33612946e-01 -8.48710462e-02 1.19173717e+00 1.35905892e-01
1.24418318e+00 -6.15944982e-01 -3.60677242e-01 4.11709934e-01
1.25026107e+00 -3.00354958e-02 6.60937130e-01 1.35565072e-01
1.04759157e+00 6.89255178e-01 1.04666792e-01 1.52242389e-02
8.36533785e-01 9.32359844e-02 4.43096459e-01 2.25544609e-02
-3.12932819e-01 -6.52699232e-01 1.54792592e-01 1.53616142e+00
2.75485307e-01 -7.16809988e-01 -7.10101604e-01 5.44549942e-01
-2.05588531e+00 -8.73197734e-01 -4.12487149e-01 1.67936289e+00
7.30670333e-01 4.74048465e-01 -3.20770681e-01 -1.60020351e-01
1.07995331e+00 4.83215004e-01 -4.15812641e-01 -4.45356548e-01
-3.44545096e-01 3.52398455e-02 3.36663485e-01 8.65723312e-01
-1.02590454e+00 1.14756560e+00 6.42711687e+00 3.18683714e-01
-6.70830607e-01 -1.92181170e-01 6.05105340e-01 2.21405223e-01
-7.37700284e-01 1.50151610e-01 -8.14591527e-01 5.31438828e-01
8.56243014e-01 -7.11287320e-01 2.57736892e-01 3.58801007e-01
1.63504809e-01 2.54520178e-01 -1.00268388e+00 6.63198888e-01
6.82392061e-01 -1.53785479e+00 5.31160176e-01 -6.10894382e-01
5.60727060e-01 -6.54738396e-02 -4.94099051e-01 4.13577944e-01
4.76952821e-01 -8.84711504e-01 4.35058236e-01 7.53740311e-01
3.26022297e-01 -7.13912427e-01 1.02551746e+00 2.77124524e-01
-1.36830604e+00 1.46111667e-01 -6.29900157e-01 -4.30041909e-01
2.07892999e-01 6.61142528e-01 -9.12406802e-01 1.36807430e+00
5.55450141e-01 1.30649281e+00 -1.06529534e+00 6.35375321e-01
-8.13557029e-01 5.63720703e-01 5.68271615e-02 -4.79482293e-01
-7.76256155e-03 -1.59420088e-01 4.17147428e-01 1.40434527e+00
-6.77045807e-02 1.58760265e-01 2.15119600e-01 6.69752717e-01
-7.31190860e-01 1.34488672e-01 -9.49774325e-01 -1.03563726e-01
7.00560868e-01 9.69684005e-01 -8.20044279e-01 -4.50867444e-01
-3.86792421e-01 1.22871983e+00 1.02688384e+00 5.82763553e-01
-7.05875576e-01 -8.64635766e-01 2.57060319e-01 -2.69492716e-01
2.99063414e-01 -3.98816019e-01 -1.40078977e-01 -1.43630636e+00
4.06093925e-01 -3.50618690e-01 6.38999462e-01 -9.08170760e-01
-1.57664645e+00 8.79106224e-01 -2.05814745e-02 -8.26714396e-01
-1.90080985e-01 -2.05080509e-01 -8.97545338e-01 6.78001285e-01
-1.51454699e+00 -1.10932970e+00 -1.29079819e-02 4.56643224e-01
5.78593194e-01 1.13067932e-01 6.07825160e-01 4.81929630e-02
-7.91282535e-01 5.18680632e-01 -1.48413405e-01 7.39701807e-01
3.74656528e-01 -1.10406184e+00 9.69559252e-01 1.25037396e+00
3.55302393e-01 1.08017075e+00 6.36705101e-01 -1.05484164e+00
-1.18260002e+00 -1.30229938e+00 1.55027902e+00 -1.48000821e-01
8.73758137e-01 -4.43527430e-01 -1.07726336e+00 1.16096497e+00
7.14538276e-01 -2.95514613e-01 6.13076806e-01 2.34271526e-01
-3.45173717e-01 -1.33693755e-01 -4.25436318e-01 7.49147415e-01
1.54493141e+00 -7.42355585e-01 -1.20611298e+00 4.62235838e-01
1.50913417e+00 -4.03333634e-01 -5.91205776e-01 4.55507517e-01
-3.84160317e-02 -7.19071567e-01 7.25362360e-01 -1.01825678e+00
5.84410310e-01 -2.32990265e-01 3.68711352e-03 -1.43282139e+00
-6.90738320e-01 -4.59619552e-01 -2.26084962e-01 1.50851011e+00
6.00297630e-01 -8.86178792e-01 7.08253562e-01 4.26513940e-01
-2.93057859e-01 -7.19803095e-01 -6.73565865e-01 -7.55034029e-01
-9.40332264e-02 -7.98586085e-02 7.58825719e-01 1.05475616e+00
2.25727618e-01 1.14957440e+00 -2.70425707e-01 3.37260693e-01
3.73495281e-01 3.07990223e-01 4.51792121e-01 -1.03418982e+00
-3.39173585e-01 -2.50439107e-01 -3.86576056e-01 -1.25267458e+00
8.44990790e-01 -1.21294475e+00 -2.30125368e-01 -2.44895959e+00
4.54618305e-01 3.40702266e-01 -5.26910603e-01 5.96057594e-01
-6.46359324e-01 -1.55797735e-01 2.12926343e-01 1.81792781e-01
-1.04716837e+00 8.26784253e-01 1.20604885e+00 -5.31472623e-01
-2.65624281e-02 -2.78809279e-01 -1.09394133e+00 5.06767392e-01
8.84068012e-01 -5.63406050e-01 -8.77884448e-01 -1.07156539e+00
4.01424915e-01 -4.05835174e-02 4.06937376e-02 -5.76128542e-01
6.27854824e-01 5.24561033e-02 2.03964770e-01 -6.61407173e-01
-5.14492467e-02 -4.47881520e-01 -4.33438234e-02 3.43322963e-01
-6.79051578e-01 3.99288088e-01 -4.00027409e-02 8.22705030e-01
-3.46154332e-01 -2.69597113e-01 1.23018429e-01 -7.37444758e-02
-6.73378408e-01 1.81290910e-01 7.18355179e-02 8.65061209e-02
8.02311838e-01 7.41261914e-02 -7.79722154e-01 -6.77042663e-01
-6.42113626e-01 6.13454282e-01 2.28673369e-01 7.54584551e-01
9.12635028e-01 -1.04878318e+00 -8.26346397e-01 -1.35784358e-01
-1.36046231e-01 7.90832564e-03 3.68572772e-01 4.37033504e-01
-5.26849329e-01 5.70862591e-01 1.83891103e-01 -2.45980486e-01
-1.36749077e+00 5.73895097e-01 2.14185625e-01 -6.05978847e-01
-7.34789610e-01 8.29941571e-01 1.62055627e-01 -6.12767935e-01
8.05841833e-02 -3.02926451e-01 -7.01686740e-01 -2.60448188e-01
3.46589953e-01 -4.81306296e-03 1.34869680e-01 -4.45750445e-01
-5.34428000e-01 3.70365024e-01 -6.27152264e-01 4.41588044e-01
1.38422954e+00 -3.34068865e-01 -6.38331890e-01 5.75551867e-01
1.33707392e+00 -1.56849530e-02 -6.53186381e-01 -3.16224277e-01
1.63133949e-01 -1.64838225e-01 -3.53881955e-01 -4.27169502e-01
-9.35642719e-01 6.24938607e-01 -5.22507966e-01 4.39991325e-01
1.12701178e+00 1.42429128e-01 9.29915547e-01 7.49104440e-01
8.31375793e-02 -7.75594652e-01 2.44267136e-02 9.47332799e-01
1.04293776e+00 -8.65232825e-01 2.03331724e-01 -8.35706770e-01
-7.15703487e-01 1.15417111e+00 7.19815671e-01 -3.36361498e-01
4.46208447e-01 -2.64847666e-01 -2.21616536e-01 -2.39017963e-01
-8.60219598e-01 -1.22020289e-01 2.62658060e-01 2.86051393e-01
4.02251005e-01 -1.66268557e-01 -6.11737609e-01 9.28432643e-01
-3.39656353e-01 -2.00809181e-01 7.92913079e-01 7.92505920e-01
-2.76473939e-01 -1.19427967e+00 3.65729958e-01 5.91360629e-01
-2.25752488e-01 -5.41229486e-01 -1.01679945e+00 6.74853683e-01
-4.63338941e-01 1.23082542e+00 -3.04114856e-02 -6.16804183e-01
6.29936635e-01 4.55169491e-02 2.03126937e-01 -9.02433991e-01
-5.73031902e-01 -4.59169418e-01 4.38509166e-01 -2.86664486e-01
-3.00310075e-01 -3.84829521e-01 -1.62146580e+00 -3.87916088e-01
-2.48254299e-01 4.14157152e-01 1.36901975e-01 1.00605714e+00
7.19135225e-01 1.02525520e+00 7.94995427e-01 -2.55959541e-01
-3.23735088e-01 -1.07229447e+00 -5.50735891e-01 7.09840298e-01
1.23637617e-01 -2.73997456e-01 -3.56005877e-01 -4.26448658e-02] | [11.189460754394531, 8.672701835632324] |
2ea63c69-5223-4e2e-870a-05944a171249 | rgcl-at-semeval-2020-task-6-neural-approaches | 2010.06281 | null | https://arxiv.org/abs/2010.06281v1 | https://arxiv.org/pdf/2010.06281v1.pdf | RGCL at SemEval-2020 Task 6: Neural Approaches to Definition Extraction | This paper presents the RGCL team submission to SemEval 2020 Task 6: DeftEval, subtasks 1 and 2. The system classifies definitions at the sentence and token levels. It utilises state-of-the-art neural network architectures, which have some task-specific adaptations, including an automatically extended training set. Overall, the approach achieves acceptable evaluation scores, while maintaining flexibility in architecture selection. | ['Ruslan Mitkov', 'Constantin Orasan', 'Alistair Plum', 'Tharindu Ranasinghe'] | 2020-10-13 | null | null | null | null | ['definition-extraction'] | ['natural-language-processing'] | [ 1.97356388e-01 3.01427901e-01 -3.23143780e-01 -6.02629125e-01
-4.65688854e-01 -8.46577108e-01 8.77036273e-01 5.68835586e-02
-1.09669614e+00 9.33770359e-01 3.75215411e-01 -7.73154080e-01
-1.49494588e-01 -4.69800055e-01 -1.53345004e-01 1.81898475e-01
1.39302850e-01 7.09360778e-01 -3.19718122e-02 -5.69185793e-01
1.30462393e-01 1.37629166e-01 -1.17828417e+00 7.20558882e-01
5.59343040e-01 1.08763540e+00 2.09617447e-02 9.11298931e-01
-4.66030955e-01 9.84906971e-01 -9.46978152e-01 -7.30157495e-01
-2.96464730e-02 -1.10785114e-02 -1.43328023e+00 -6.84535205e-01
1.04408932e+00 1.83026016e-01 -7.52716139e-02 9.88753378e-01
4.22652245e-01 4.07707483e-01 4.01627600e-01 -6.49606347e-01
-8.95473778e-01 1.38105893e+00 2.98735082e-01 6.27786279e-01
3.23660433e-01 7.45617077e-02 1.46267545e+00 -1.22196543e+00
6.05506539e-01 1.04007137e+00 9.60201323e-01 9.65789795e-01
-1.22504961e+00 -4.83165860e-01 2.10406095e-01 3.19622219e-01
-1.34274089e+00 -6.96416259e-01 3.32564920e-01 -4.01727468e-01
2.50841546e+00 7.04159200e-01 3.76474142e-01 1.18814921e+00
6.06349632e-02 7.49555230e-01 6.20356441e-01 -7.58019269e-01
5.32765687e-03 -1.54349944e-02 6.93261206e-01 2.21975595e-01
1.09518006e-01 -5.25735207e-02 -3.22804272e-01 5.78626804e-02
4.50405777e-01 -8.16011250e-01 -1.08050883e-01 4.69721913e-01
-1.11505198e+00 5.56228280e-01 1.33333996e-01 7.53488481e-01
-5.70772469e-01 3.84998292e-01 1.07317030e+00 5.11007667e-01
5.85766137e-01 1.24367678e+00 -1.26183867e+00 -4.75867867e-01
-1.17513168e+00 5.16080141e-01 9.54805374e-01 1.15531564e+00
2.09519237e-01 5.98813832e-01 -5.95940769e-01 1.09753144e+00
-1.25410315e-02 5.23046404e-02 7.78668463e-01 -8.21246624e-01
8.92671525e-01 3.19673359e-01 -1.09550193e-01 -5.50919712e-01
-6.11337721e-01 -8.16786170e-01 -7.20085680e-01 -3.46371830e-01
-1.53599912e-02 -5.00533462e-01 -1.06851470e+00 1.65003359e+00
-5.14721751e-01 -2.35228986e-01 1.41812921e-01 3.88134658e-01
1.43105626e+00 2.60103822e-01 6.46725953e-01 5.42171411e-02
1.09552479e+00 -9.82919037e-01 -8.71147573e-01 -3.59951615e-01
7.40275562e-01 -4.42617387e-01 6.43403530e-01 4.46499735e-01
-1.47916663e+00 -9.57298815e-01 -1.05263102e+00 2.74191443e-02
-9.13238287e-01 2.51007289e-01 4.14417654e-01 6.47506654e-01
-1.47111332e+00 4.71727163e-01 -4.50610369e-01 -2.63508290e-01
-1.40170529e-01 4.95529532e-01 -1.08958051e-01 4.74391818e-01
-2.02086115e+00 1.58030415e+00 1.43691432e+00 7.98065811e-02
-5.24245858e-01 -8.24782968e-01 -1.09400964e+00 3.59290764e-02
1.16179235e-01 -6.20882571e-01 1.79425001e+00 -7.69711018e-01
-1.24690759e+00 9.65953767e-01 4.99862581e-02 -1.02212405e+00
5.61149083e-02 -4.08074051e-01 -9.11221206e-01 -2.97591150e-01
1.66163463e-02 7.62465179e-01 1.31463572e-01 -5.79652607e-01
-8.77502084e-01 1.69419020e-01 2.83416182e-01 4.57086593e-01
-3.71106207e-01 4.63683724e-01 -1.26602694e-01 -6.51823699e-01
-7.18731284e-01 -1.77181438e-01 -1.92943588e-01 -1.01914096e+00
-5.03845394e-01 -1.03807652e+00 2.55079895e-01 -8.29010606e-01
1.78099322e+00 -1.55280697e+00 2.32737482e-01 -1.41743064e-01
3.02967459e-01 6.40909672e-01 -2.55370677e-01 4.56321090e-01
-3.96435797e-01 4.33020055e-01 8.43853056e-02 -7.08721936e-01
3.21794957e-01 1.42884374e-01 -1.92079082e-01 -2.27852285e-01
2.42010176e-01 1.11954296e+00 -1.07822168e+00 -3.77755612e-01
4.16061282e-01 1.23951115e-01 4.14305516e-02 6.88825920e-02
-3.96601021e-01 -3.90167773e-01 -9.84530598e-02 2.47506961e-01
3.08773160e-01 2.04625860e-01 1.69490397e-01 -1.71159014e-01
-3.63853753e-01 1.01073492e+00 -7.88173020e-01 1.71653867e+00
-5.93324006e-01 8.58427286e-01 -2.84441691e-02 -9.72328961e-01
8.50441873e-01 7.63126075e-01 1.57847613e-01 -4.95720536e-01
1.38409631e-02 2.14021400e-01 -3.67814512e-03 -5.16471505e-01
1.06977415e+00 -5.05094975e-02 -2.90671468e-01 7.21384713e-04
7.33505428e-01 -1.28657818e-01 6.14619732e-01 7.37265721e-02
1.29984832e+00 9.59317237e-02 4.49479103e-01 -4.16152507e-01
6.20923460e-01 -2.06843335e-02 4.78357166e-01 8.74771118e-01
-1.06455289e-01 4.74885643e-01 -8.49376619e-02 -7.41347373e-01
-1.04145169e+00 -1.11861491e+00 -2.93302655e-01 1.34991705e+00
-6.46320403e-01 -7.29981124e-01 -8.63662899e-01 -9.36876297e-01
-1.58275098e-01 1.49294758e+00 -7.95058429e-01 2.24954695e-01
-7.91783631e-01 -4.39674079e-01 1.47152996e+00 8.51939082e-01
4.86809492e-01 -1.74601102e+00 -7.39892542e-01 5.33563673e-01
1.66446455e-02 -1.00633681e+00 -1.96081638e-01 5.68884969e-01
-4.65794325e-01 -8.25067103e-01 -3.39532316e-01 -8.54225755e-01
-7.22900257e-02 -5.72441459e-01 2.10044765e+00 -8.96887258e-02
-6.87237009e-02 1.91222265e-01 -5.27147114e-01 -5.08680940e-01
-2.48486310e-01 8.08291912e-01 1.18438210e-02 -6.72074854e-01
9.38514531e-01 -3.12893540e-02 9.49447881e-03 -3.26387703e-01
-6.25615239e-01 -7.90798590e-02 3.57019484e-01 9.53519344e-01
2.10161716e-01 -1.88422263e-01 8.88851464e-01 -1.17127860e+00
1.31894946e+00 -3.80022436e-01 -1.71557635e-01 3.90000015e-01
-8.82519603e-01 -1.79898202e-01 8.47596526e-01 8.68232746e-04
-1.10361838e+00 -2.06010625e-01 -6.36660695e-01 -4.11916934e-02
-6.15170002e-01 1.10935891e+00 -1.04191057e-01 3.00419450e-01
9.32255566e-01 1.54717803e-01 -4.61411834e-01 -6.11573040e-01
6.51159406e-01 8.12059760e-01 6.95663035e-01 -6.13457680e-01
2.21689463e-01 -6.26276016e-01 -7.43755400e-01 -8.11396956e-01
-1.05308652e+00 -6.40018880e-01 -7.60479212e-01 -2.55027473e-01
8.22603941e-01 -5.83353162e-01 -8.80506486e-02 3.11984479e-01
-1.64862585e+00 -3.82011414e-01 -5.68663001e-01 2.23420203e-01
-2.49901995e-01 -1.60358861e-01 -6.33934617e-01 -4.76519138e-01
-1.01556373e+00 -6.00816488e-01 5.44567823e-01 4.35496867e-02
-8.12005997e-01 -1.59016347e+00 3.66550803e-01 -4.66015190e-02
9.96785104e-01 1.84939638e-01 8.16936076e-01 -1.27199268e+00
4.79807884e-01 -1.36877447e-01 -2.96324398e-02 9.04276252e-01
-1.29698545e-01 -1.26537263e-01 -8.16150904e-01 -1.40369326e-01
-3.65113437e-01 -5.93847334e-01 1.14362514e+00 3.09973955e-01
1.23540735e+00 -4.66594577e-01 -7.46553317e-02 4.57113057e-01
1.24739647e+00 1.42568693e-01 3.72361898e-01 7.32567310e-01
6.14705086e-01 5.09252846e-01 1.89421460e-01 4.19399887e-02
3.83186072e-01 5.24393260e-01 4.05719370e-01 3.83870192e-02
-3.84962142e-01 2.14203313e-01 2.04775140e-01 1.10292125e+00
-1.05217911e-01 -4.35388774e-01 -1.27296984e+00 9.62720037e-01
-1.89032054e+00 -9.93492901e-01 -2.73560286e-01 1.55608571e+00
1.35994661e+00 4.68000919e-01 -7.16978461e-02 -1.16670892e-01
6.19913936e-01 3.71298879e-01 -3.59297514e-01 -1.17195725e+00
-3.05313766e-01 9.44604814e-01 4.29940492e-01 5.13908446e-01
-1.23308849e+00 1.24367976e+00 8.20986176e+00 1.12975883e+00
-7.27869093e-01 2.66651452e-01 2.75006443e-01 -1.92961380e-01
-2.39416763e-01 -3.84818196e-01 -1.21541548e+00 2.29923993e-01
1.58209646e+00 -1.55753851e-01 2.07677245e-01 6.94631279e-01
-2.63599932e-01 6.16329193e-01 -1.19066203e+00 4.80377585e-01
3.74094993e-01 -1.68196404e+00 1.97150826e-01 -5.45040429e-01
4.05491084e-01 7.69965410e-01 -4.54342067e-02 9.76727903e-01
6.63607717e-01 -1.43971670e+00 8.88085485e-01 5.58876634e-01
1.16362512e+00 -8.52864385e-01 1.28223956e+00 1.55996948e-01
-1.04111266e+00 2.45555177e-01 -3.21285784e-01 -4.31603521e-01
2.28898540e-01 2.52252787e-01 -9.48955834e-01 6.56373322e-01
4.81988579e-01 7.41382658e-01 -8.19132566e-01 9.54743862e-01
-6.85336292e-01 9.56662476e-01 -6.20783344e-02 -3.99836272e-01
5.38749158e-01 4.99810845e-01 6.28435850e-01 2.10912681e+00
-4.05262560e-01 -1.13779627e-01 2.58846208e-02 9.10487115e-01
-3.46289873e-01 -1.41831841e-02 -3.70312035e-01 2.38549381e-01
9.12378967e-01 1.14770293e+00 7.66473934e-02 -5.60212076e-01
-1.79022208e-01 5.49390495e-01 6.63280070e-01 3.57420832e-01
-4.84702349e-01 -9.65869248e-01 4.14129227e-01 -3.08107078e-01
1.61861986e-01 -1.41971111e-01 -5.37456930e-01 -1.05757034e+00
-9.27407146e-02 -8.16360831e-01 5.95325351e-01 -4.43613678e-01
-1.49747193e+00 1.15782940e+00 2.54671305e-01 -4.78568017e-01
-6.41997695e-01 -1.02862346e+00 -6.72475696e-01 1.01616287e+00
-1.18542087e+00 -1.37228143e+00 1.96364179e-01 2.24284559e-01
8.17218363e-01 -5.99602699e-01 1.41213226e+00 3.61425698e-01
-3.63879383e-01 1.15637839e+00 6.45478219e-02 6.04647338e-01
4.35066998e-01 -1.84078085e+00 1.13554871e+00 6.80964828e-01
1.95879400e-01 8.62553716e-01 5.80578923e-01 -4.61598188e-01
-6.29794657e-01 -1.41835427e+00 1.52599132e+00 -7.24109828e-01
8.03873897e-01 -2.20255882e-01 -5.15388012e-01 8.73151600e-01
1.00968754e+00 -3.29342663e-01 6.59054995e-01 7.93885589e-01
-2.33466998e-01 1.36429578e-01 -1.01636195e+00 3.00532758e-01
9.22242105e-01 -6.22795939e-01 -1.42821085e+00 1.48285896e-01
9.22113240e-01 -6.48541689e-01 -1.14730108e+00 4.75951105e-01
4.87335980e-01 -3.03831130e-01 8.11864197e-01 -1.08827448e+00
4.42104131e-01 9.08514559e-02 -3.07238512e-02 -1.63765287e+00
-4.96312708e-01 -5.17331541e-01 -4.81903791e-01 1.33618677e+00
1.20252371e+00 -3.96194190e-01 5.38023353e-01 6.74559772e-01
-6.43705964e-01 -8.37104201e-01 -1.46348190e+00 -1.01005816e+00
6.15808487e-01 -8.93581808e-01 5.37157595e-01 1.07309079e+00
2.10659862e-01 8.82182002e-01 -1.87988523e-02 -5.99586308e-01
2.58554041e-01 -6.15947962e-01 1.65355533e-01 -9.40688312e-01
-3.79589647e-01 -1.03217077e+00 -3.37558657e-01 -9.97437954e-01
4.36143100e-01 -1.09147561e+00 -1.35900313e-02 -1.65543568e+00
-3.25775504e-01 -2.33116239e-01 -8.34800482e-01 9.37157094e-01
-9.78986248e-02 1.79419488e-01 7.94823617e-02 -2.10580721e-01
-9.58948791e-01 3.20600927e-01 4.50729460e-01 -3.54397833e-01
1.14148542e-01 -1.37744442e-01 -5.71579039e-01 4.69617724e-01
9.73879516e-01 -4.54206578e-02 -9.49899107e-02 -1.09219360e+00
4.28162843e-01 -6.42891407e-01 -9.86278802e-02 -8.16426575e-01
2.03747228e-01 1.62179843e-02 2.44129732e-01 -8.22223961e-01
1.10737480e-01 -1.89694881e-01 -1.62880644e-01 1.31890923e-01
-9.42816377e-01 4.95621502e-01 6.54285848e-01 -1.77195027e-01
-3.96654367e-01 -5.98854959e-01 4.33113366e-01 -5.01390159e-01
-8.76397491e-01 -6.93782195e-02 -6.18448794e-01 4.15276706e-01
5.68496883e-01 -4.71362732e-02 -4.17172402e-01 1.01876765e-01
-7.25493312e-01 5.15474916e-01 -1.50192752e-01 6.73269868e-01
6.63080156e-01 -1.33161962e+00 -1.25868154e+00 -1.35942176e-01
2.55732089e-01 -3.03103570e-02 -3.14205617e-01 2.98405796e-01
-4.24743742e-01 1.03295028e+00 -8.61434266e-03 -1.19982749e-01
-1.04880810e+00 1.13105334e-01 7.48442054e-01 -7.19044983e-01
-2.53918350e-01 1.39083445e+00 -5.74179530e-01 -7.96579480e-01
3.22850943e-01 -2.34883189e-01 -9.29730237e-01 1.70836791e-01
8.40414286e-01 2.87823826e-01 5.30521274e-01 -4.67715353e-01
-4.89721984e-01 -9.08064917e-02 -4.95520771e-01 -1.96279541e-01
1.39007139e+00 2.80670911e-01 -1.84518963e-01 4.61991310e-01
9.53065276e-01 -5.06116927e-01 -2.73334831e-01 -2.93681234e-01
5.83955526e-01 2.10803106e-01 4.43642110e-01 -1.54816341e+00
-7.89778948e-01 7.23991990e-01 1.65525943e-01 5.37810326e-01
9.37082946e-01 -2.89962977e-01 8.97324860e-01 5.60582578e-01
8.62544924e-02 -1.51867080e+00 -6.59947515e-01 1.51918066e+00
1.04784226e+00 -6.59330726e-01 -2.73314059e-01 2.92889237e-01
-3.03679138e-01 1.34265006e+00 1.01407790e+00 -2.06455514e-01
3.94522101e-01 2.76015103e-01 -9.08480138e-02 -4.81655836e-01
-1.21983063e+00 -6.27490208e-02 5.37780344e-01 5.49727440e-01
1.00654781e+00 3.93431991e-01 -7.18396485e-01 8.20015907e-01
-4.75100458e-01 -2.09265679e-01 1.49032488e-01 6.31883264e-01
-2.11712152e-01 -1.19971550e+00 2.35354453e-01 6.94792211e-01
-9.03218567e-01 -9.54366565e-01 -8.73901725e-01 8.37714911e-01
3.84109706e-01 8.65029693e-01 5.36767207e-02 -6.17578149e-01
7.45017052e-01 2.60267496e-01 3.78680170e-01 -1.23103285e+00
-1.51053488e+00 -5.65421999e-01 1.09164441e+00 -3.30233723e-01
-2.44956583e-01 -6.25831366e-01 -1.10317612e+00 2.41877679e-02
-4.55493450e-01 5.00547171e-01 5.44980705e-01 1.08711660e+00
2.95322448e-01 8.54020059e-01 9.15183034e-03 -6.02473915e-01
-8.25428307e-01 -1.85309851e+00 -1.96199432e-01 2.17173636e-01
1.94838315e-01 -3.26397657e-01 -1.49745986e-01 8.06114152e-02] | [10.135313987731934, 9.346162796020508] |
eadafbb8-8924-466e-bea7-eef5c6cd7690 | hyperspectral-unmixing-using-a-neural-network | null | null | https://ieeexplore.ieee.org/document/8322133 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8322133 | Hyperspectral Unmixing Using a Neural Network Autoencoder | In this paper, we present a deep learning based method for blind hyperspectral unmixing in the form of a neural network autoencoder. We show that the linear mixture model implicitly puts certain architectural constraints on the network, and it effectively performs blind hyperspectral unmixing. Several different architectural configurations of both shallow and deep encoders are evaluated. Also, deep encoders are tested using different activation functions. Furthermore, we investigate the performance of the method using three different objective functions. The proposed method is compared to other benchmark methods using real data and previously established ground truths of several common data sets. Experiments show that the proposed method compares favorably to other commonly used hyperspectral unmixing methods and exhibits robustness to noise. This is especially true when using spectral angle distance as the network's objective function. Finally, results indicate that a deeper and a more sophisticated encoder does not necessarily give better results. | ['Magnus O. Ulfarsson', 'Johannes R. Sveinsson', 'Jakob Sigurdsson', 'Burkni Palsson'] | 2018-03-22 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 2.21260861e-01 -4.18182820e-01 -6.98478520e-02 -1.75115064e-01
-4.46595371e-01 -5.62814057e-01 7.08600581e-01 -4.32970256e-01
-4.36613917e-01 7.51520753e-01 1.51176274e-01 -3.50517601e-01
-3.74469638e-01 -5.79150617e-01 -5.14309168e-01 -1.16088355e+00
5.37975989e-02 3.66096288e-01 -4.95890707e-01 -1.82343021e-01
-2.23377738e-02 8.35311830e-01 -1.63032186e+00 -2.03574039e-02
1.10046124e+00 9.80380118e-01 1.46112636e-01 6.74967706e-01
4.27675903e-01 9.09364641e-01 -8.72715354e-01 -3.32359076e-01
6.63966119e-01 -3.37327242e-01 -6.48639619e-01 3.96009624e-01
9.03874218e-01 -5.96274197e-01 -5.25943160e-01 1.43960226e+00
7.63533890e-01 4.14714515e-01 6.80057049e-01 -9.44897175e-01
-6.98669016e-01 6.43522084e-01 -3.62683922e-01 5.78302257e-02
-2.35905856e-01 4.40968983e-02 9.08537328e-01 -6.07347906e-01
1.97916497e-02 8.54927957e-01 8.68549705e-01 1.79278359e-01
-1.13172615e+00 -6.41529024e-01 -1.95958659e-01 2.19587609e-01
-1.55015767e+00 -5.14120340e-01 7.32618928e-01 -4.76650983e-01
1.01995671e+00 3.98919284e-01 5.60754418e-01 7.83653080e-01
-2.36080930e-01 4.98612016e-01 1.28323424e+00 -6.06024802e-01
2.38137782e-01 7.53503516e-02 6.00498868e-03 7.37029374e-01
6.81084096e-01 3.75666618e-01 -8.53159353e-02 -6.61598891e-02
6.69636905e-01 -9.30265561e-02 -6.55856133e-01 -1.55345246e-01
-1.04867268e+00 8.23548138e-01 6.94668293e-01 5.37211299e-01
-5.27835369e-01 -2.11636834e-02 -1.10007159e-01 2.36021057e-01
5.25953352e-01 6.48048759e-01 -1.82326436e-01 5.12619257e-01
-1.34102345e+00 -1.83273792e-01 7.65874386e-01 5.26108742e-01
9.78322625e-01 6.74312949e-01 1.77018151e-01 1.06050336e+00
4.76027906e-01 5.27477145e-01 7.34408736e-01 -9.36124206e-01
1.84285045e-01 2.68063009e-01 1.22159317e-01 -8.96115661e-01
-4.80056316e-01 -8.48750889e-01 -1.10654414e+00 5.56507289e-01
1.27522752e-01 -4.06715065e-01 -1.17361915e+00 1.36576807e+00
-2.57307559e-01 3.69414359e-01 4.99265581e-01 1.16159856e+00
8.93016398e-01 7.46009350e-01 -2.77051359e-01 -2.10407466e-01
8.62421930e-01 -1.11850119e+00 -8.85655761e-01 -5.15002012e-01
1.89711139e-01 -7.25195289e-01 4.25417602e-01 4.49218035e-01
-8.51998389e-01 -4.89218175e-01 -1.56572330e+00 1.70806825e-01
-5.71479678e-01 7.13846982e-01 6.61643267e-01 9.76549208e-01
-1.19478595e+00 7.83027828e-01 -6.55339777e-01 -3.11610311e-01
2.40710929e-01 6.00640357e-01 -3.13659400e-01 -4.08656485e-02
-1.13866043e+00 9.43701148e-01 8.48461330e-01 4.86486405e-01
-9.10710812e-01 -2.58806109e-01 -8.81251633e-01 1.88914180e-01
-2.19986349e-01 -6.32493913e-01 1.17332387e+00 -1.51718676e+00
-1.77739418e+00 6.90164089e-01 9.16319713e-02 -4.86800432e-01
1.17478959e-01 -2.80098408e-01 -8.81562769e-01 2.45512918e-01
-3.60747635e-01 5.82867205e-01 1.16122806e+00 -1.51173735e+00
-3.87938529e-01 -1.95863396e-01 1.06388517e-01 3.79386574e-01
-6.07012391e-01 -1.54645503e-01 -1.29237715e-02 -7.80049205e-01
2.50698954e-01 -9.02490020e-01 -1.85322434e-01 -4.31340605e-01
-5.78465819e-01 4.14618403e-01 6.99572504e-01 -8.56297374e-01
8.41524899e-01 -2.04420400e+00 2.64405996e-01 3.02513421e-01
1.67527765e-01 5.19421458e-01 -8.47893432e-02 2.53912807e-01
-6.46516681e-01 -8.69860724e-02 -6.57659292e-01 -3.96233022e-01
9.53268185e-02 2.33968139e-01 -1.82404984e-02 8.42954397e-01
3.12998251e-04 5.84686160e-01 -5.64425945e-01 6.41031340e-02
4.38147008e-01 6.08043730e-01 -3.22141618e-01 3.68027925e-01
-7.29794055e-02 2.06845701e-01 4.15885359e-01 7.83947408e-01
9.32161927e-01 -3.46908063e-01 3.10911149e-01 -6.24351442e-01
-1.25934362e-01 2.88683418e-02 -1.13961923e+00 1.56769645e+00
-1.50029168e-01 1.03342068e+00 3.06845903e-01 -1.11721265e+00
6.58558011e-01 6.15682781e-01 3.15136731e-01 -3.54340523e-01
3.28025520e-01 4.61903274e-01 1.48913190e-01 -4.66080964e-01
4.69016612e-01 -4.98019427e-01 8.55343282e-01 2.78344125e-01
4.02314305e-01 -2.44189680e-01 1.73386320e-01 -4.45554763e-01
5.04818141e-01 -1.84040934e-01 1.83743387e-01 -3.70460004e-01
5.49336851e-01 -3.14491764e-02 1.34688422e-01 6.00114226e-01
1.02817686e-02 6.19898856e-01 -3.21057402e-02 -2.74100065e-01
-8.67326319e-01 -6.65243208e-01 -1.89537674e-01 6.71019375e-01
1.01993330e-01 -2.11217217e-02 -6.93374455e-01 -2.51676798e-01
-2.23171279e-01 5.89452386e-01 -4.40475166e-01 -1.03227068e-02
-3.31126228e-02 -1.50608611e+00 7.29126453e-01 3.33944201e-01
7.24686801e-01 -6.33256018e-01 -1.38580158e-01 -1.34473443e-01
-2.01617941e-01 -1.16438711e+00 2.12354604e-02 5.43110132e-01
-1.08444417e+00 -9.93250072e-01 -9.53975320e-01 -8.37008834e-01
6.04533136e-01 4.95622128e-01 9.16616023e-01 -1.78173482e-01
9.56334472e-02 9.52503979e-02 -2.74424344e-01 -2.73023516e-01
-5.59515536e-01 8.18935409e-02 1.75267100e-01 2.83727020e-01
4.54148740e-01 -6.32577002e-01 -4.87627774e-01 2.75926609e-02
-1.38046145e+00 -2.10505500e-01 8.39323461e-01 7.74827898e-01
2.18286633e-01 7.07816899e-01 1.31215200e-01 -4.95909303e-01
6.42896891e-01 -4.01784569e-01 -7.86376119e-01 6.45014122e-02
-7.08802938e-01 1.15566909e-01 3.39291394e-01 -2.38150537e-01
-9.57819462e-01 6.05663061e-02 -1.30286857e-01 -6.62581563e-01
-3.74933422e-01 7.66478181e-01 -3.66380453e-01 -4.15357828e-01
9.00201380e-01 1.99042201e-01 6.29166439e-02 -5.52581012e-01
1.71152398e-01 9.33296680e-01 6.41957223e-01 -6.68408126e-02
9.58026409e-01 5.72743297e-01 -2.31413558e-01 -1.14727259e+00
-6.28269434e-01 -4.66512769e-01 -5.29116929e-01 -6.41092658e-02
6.15968108e-01 -1.28196764e+00 -3.74917358e-01 9.15585995e-01
-9.43708837e-01 -3.85409772e-01 1.84799302e-02 7.55126297e-01
-3.59629244e-01 5.29297292e-01 -6.38380229e-01 -7.69784689e-01
-2.65448868e-01 -1.29934978e+00 8.14896941e-01 1.81539685e-01
2.06651747e-01 -1.23692524e+00 3.75393853e-02 4.66650665e-01
4.85749424e-01 1.68160230e-01 6.69609308e-01 -5.85732281e-01
-3.46343189e-01 -2.93378592e-01 -2.47532725e-01 1.03533649e+00
4.21615601e-01 5.59207574e-02 -1.58473432e+00 -4.99500841e-01
3.66846591e-01 -9.72048715e-02 1.11468172e+00 5.28510988e-01
1.14754605e+00 -5.38708329e-01 -1.51834727e-04 1.05291295e+00
1.82168281e+00 2.54215866e-01 9.25279796e-01 4.85516161e-01
9.13473904e-01 4.54748064e-01 -1.57103226e-01 2.03738630e-01
-1.56865150e-01 1.23945028e-01 9.71425772e-01 -5.14959931e-01
1.17072761e-01 2.31079683e-01 4.25788343e-01 9.37500834e-01
-4.31064695e-01 -5.35362065e-01 -6.93612993e-01 3.01195860e-01
-1.61565816e+00 -1.12342250e+00 -1.42880529e-01 2.18880653e+00
5.77788532e-01 -3.49569142e-01 -1.45405248e-01 5.38136005e-01
7.19937205e-01 4.22806084e-01 -3.78541172e-01 1.18387811e-01
-6.73078179e-01 3.82459402e-01 8.38618636e-01 7.55956173e-01
-1.51976478e+00 7.65424311e-01 7.25642872e+00 4.09226179e-01
-1.47897208e+00 1.34962186e-01 3.52581799e-01 -8.92361905e-03
-1.27667323e-01 -3.73056203e-01 -1.87257200e-01 2.90942520e-01
8.54893804e-01 4.65452939e-01 8.68559778e-01 3.61353159e-01
-5.73717766e-02 5.84712960e-02 -7.59258866e-01 1.21640325e+00
3.59607249e-01 -1.25599253e+00 -3.08503062e-02 3.17904800e-02
9.76857781e-01 5.02108514e-01 2.28700548e-01 -1.06964722e-01
2.38717541e-01 -1.47752416e+00 5.92384756e-01 4.19360280e-01
7.06161976e-01 -6.92867756e-01 1.05936611e+00 1.15475759e-01
-7.14397252e-01 -2.65305907e-01 -4.40609068e-01 -6.77589253e-02
-3.77729654e-01 8.29214990e-01 -5.55729091e-01 7.79831707e-01
4.43323553e-01 7.77438760e-01 -5.30551672e-01 1.13661492e+00
-2.74280012e-01 6.31357431e-01 -4.15864527e-01 2.82729328e-01
3.95003110e-01 -5.00269294e-01 5.27420163e-01 1.36871934e+00
5.06134033e-01 -1.45404890e-01 -2.31517047e-01 7.37798691e-01
-8.89776275e-02 -1.38692051e-01 -6.44628346e-01 -4.74079311e-01
1.63911968e-01 1.27444959e+00 -3.86597097e-01 -4.01899248e-01
-3.37918401e-01 1.04198885e+00 -6.09493256e-02 8.44026983e-01
-7.08228230e-01 -1.44142106e-01 8.55267584e-01 -2.43865028e-01
2.84674257e-01 -1.70135036e-01 -4.11084071e-02 -1.43417454e+00
-3.43412280e-01 -1.23908198e+00 1.88324615e-01 -1.16948521e+00
-1.19902515e+00 8.50371003e-01 -2.32259020e-01 -1.29787803e+00
-4.77523096e-02 -1.25661302e+00 -3.52517426e-01 1.06817174e+00
-1.87821329e+00 -8.73243988e-01 -6.48984373e-01 4.66194689e-01
3.83705422e-02 -6.01231515e-01 1.04759467e+00 5.06137669e-01
-9.22015011e-01 1.97378367e-01 6.21127248e-01 1.00520529e-01
5.52197099e-01 -1.30423343e+00 -1.46978796e-01 1.20892823e+00
3.90949756e-01 4.49054956e-01 7.86675334e-01 -1.92586988e-01
-1.13473284e+00 -1.07877684e+00 2.97244698e-01 -9.21925530e-03
5.24245739e-01 2.44666114e-01 -7.98615336e-01 7.88004577e-01
8.56983364e-01 -2.18910739e-01 1.15706515e+00 -1.52339578e-01
-2.87218004e-01 -1.85228348e-01 -1.01467979e+00 3.35468352e-01
4.56385493e-01 -6.99443161e-01 -4.73721117e-01 5.17563999e-01
3.28028977e-01 -3.01741302e-01 -7.51612365e-01 5.57400584e-01
4.02029157e-01 -1.28169262e+00 9.85237062e-01 -3.71582210e-01
4.70552415e-01 -5.01169026e-01 -5.96994042e-01 -1.86537564e+00
-6.94646358e-01 -5.46044409e-01 -3.19183350e-01 6.08090878e-01
4.90252286e-01 -8.22844446e-01 6.90835178e-01 -5.32804094e-02
-2.58464098e-01 -2.26522356e-01 -5.43753386e-01 -7.41173446e-01
1.95227377e-02 -3.22958261e-01 8.16057622e-01 1.49067652e+00
-4.54168141e-01 2.96914071e-01 -4.97616112e-01 9.35303211e-01
6.10598028e-01 9.68479663e-02 3.02613348e-01 -1.32228398e+00
-3.94215375e-01 -7.70821095e-01 -3.92316818e-01 -6.97661757e-01
4.14946765e-01 -9.24619734e-01 -6.63471147e-02 -1.57023883e+00
-2.32718349e-01 -5.49863167e-02 -5.46159387e-01 5.05782306e-01
-6.86123073e-02 5.59352040e-01 -2.01384008e-01 3.18090260e-01
2.01319128e-01 4.14976418e-01 9.38902497e-01 -6.12703323e-01
-1.19541027e-02 -1.32428870e-01 -7.09532619e-01 7.41990268e-01
1.25154483e+00 -1.54677317e-01 -3.75380307e-01 -8.15194249e-01
1.90243095e-01 -4.08579618e-01 3.88968080e-01 -1.45699811e+00
2.45392755e-01 5.35495803e-02 6.08885825e-01 -2.59597152e-01
6.57625139e-01 -1.10724521e+00 3.84249032e-01 4.48366582e-01
1.79303698e-02 -2.97273517e-01 4.59131062e-01 2.58281320e-01
-4.74847466e-01 -6.10880971e-01 9.65455949e-01 -8.14448204e-03
-7.45359004e-01 2.56679714e-01 -3.97716910e-01 -6.13352299e-01
5.76522231e-01 -2.71750897e-01 -2.08951190e-01 -3.54794890e-01
-7.29356110e-01 -3.02199930e-01 4.63839412e-01 1.09685287e-01
3.66269886e-01 -1.24433291e+00 -7.29312003e-01 4.47356254e-01
3.86503376e-02 -2.04165936e-01 -2.93113440e-02 7.67998338e-01
-1.07691348e+00 4.10113513e-01 -2.42847562e-01 -3.95370692e-01
-1.05433738e+00 4.96384203e-01 1.14492321e+00 1.26638532e-01
-6.98734969e-02 9.57149863e-01 -4.16179560e-02 -5.73352039e-01
2.68317848e-01 -3.61506253e-01 -2.93838292e-01 1.12015255e-01
5.14200389e-01 4.62781727e-01 4.00116205e-01 -7.99289227e-01
-1.14158675e-01 2.15763792e-01 4.74470854e-01 -1.26798004e-01
1.33865297e+00 1.55766070e-01 -5.97625732e-01 2.39575967e-01
1.09473872e+00 1.07482430e-02 -1.03308034e+00 -2.31344119e-01
-3.11723888e-01 -2.83981711e-01 6.20745599e-01 -9.03745234e-01
-1.50453937e+00 1.01549542e+00 1.01867974e+00 4.90667909e-01
1.47867382e+00 -5.49460769e-01 1.63913980e-01 5.26651025e-01
-1.62028834e-01 -9.00584579e-01 -3.26531440e-01 4.96562034e-01
7.70981312e-01 -1.42579007e+00 1.28003344e-01 -1.35317177e-01
-2.62498677e-01 1.17858338e+00 5.08803248e-01 8.76287669e-02
5.86736381e-01 1.70368940e-01 4.10424769e-01 -1.84014112e-01
2.79016364e-02 -4.57221150e-01 4.64935213e-01 6.82309091e-01
5.84501266e-01 2.74321049e-01 2.85864115e-01 1.13789923e-01
-3.37671876e-01 -2.69163907e-01 4.94897515e-01 5.73398709e-01
-5.18885791e-01 -8.05652559e-01 -8.58382583e-01 3.46019536e-01
-3.90346318e-01 -3.58479530e-01 -4.38530356e-01 6.33261502e-01
2.79655010e-01 9.91940618e-01 1.45756630e-02 -3.59429866e-01
-1.46175578e-01 2.79401988e-01 4.76623833e-01 -4.65207934e-01
-3.37466687e-01 1.19556345e-01 -1.75228734e-02 -7.71932974e-02
-1.01685131e+00 -1.95942834e-01 -6.22170508e-01 -4.53486353e-01
-6.69750810e-01 1.62838951e-01 8.41009915e-01 9.45863962e-01
1.13934115e-01 6.26262724e-01 5.44996858e-01 -1.15836811e+00
-5.09389639e-01 -1.37036407e+00 -7.97760367e-01 -3.06050293e-02
8.20310652e-01 -4.33327496e-01 -5.73671639e-01 1.22058578e-01] | [10.068490982055664, -2.0135788917541504] |
b24474d6-b8f0-4981-8be0-05742ba74f26 | towards-sustainable-self-supervised-learning | 2210.11016 | null | https://arxiv.org/abs/2210.11016v1 | https://arxiv.org/pdf/2210.11016v1.pdf | Towards Sustainable Self-supervised Learning | Although increasingly training-expensive, most self-supervised learning (SSL) models have repeatedly been trained from scratch but not fully utilized, since only a few SOTAs are employed for downstream tasks. In this work, we explore a sustainable SSL framework with two major challenges: i) learning a stronger new SSL model based on the existing pretrained SSL model, also called as "base" model, in a cost-friendly manner, ii) allowing the training of the new model to be compatible with various base models. We propose a Target-Enhanced Conditional (TEC) scheme which introduces two components to the existing mask-reconstruction based SSL. Firstly, we propose patch-relation enhanced targets which enhances the target given by base model and encourages the new model to learn semantic-relation knowledge from the base model by using incomplete inputs. This hardening and target-enhancing help the new model surpass the base model, since they enforce additional patch relation modeling to handle incomplete input. Secondly, we introduce a conditional adapter that adaptively adjusts new model prediction to align with the target of different base models. Extensive experimental results show that our TEC scheme can accelerate the learning speed, and also improve SOTA SSL base models, e.g., MAE and iBOT, taking an explorative step towards sustainable SSL. | ['Shuicheng Yan', 'Ming-Ming Cheng', 'Pan Zhou', 'ShangHua Gao'] | 2022-10-20 | null | null | null | null | ['self-supervised-image-classification'] | ['computer-vision'] | [ 3.37679118e-01 4.32936400e-01 -6.10982001e-01 -4.12424773e-01
-6.32564843e-01 -3.04655939e-01 4.77672070e-01 -2.02926517e-01
-3.71167868e-01 8.53696823e-01 3.95676792e-02 -2.71772206e-01
-8.37788656e-02 -7.11719573e-01 -1.01727962e+00 -5.97789049e-01
3.15292895e-01 4.63920414e-01 7.94337571e-01 -3.38120162e-01
1.75897200e-02 4.51585978e-01 -1.53620732e+00 5.49910247e-01
1.27000690e+00 9.04976189e-01 8.97786736e-01 1.68660283e-01
-3.14941347e-01 8.01961541e-01 -2.76508301e-01 -1.31467596e-01
1.87259838e-01 -3.90607357e-01 -8.32729518e-01 6.86723068e-02
1.73813269e-01 1.19456306e-01 -4.62905914e-02 7.36450255e-01
2.39548981e-01 1.67584285e-01 2.82241702e-01 -1.15529120e+00
-4.14808393e-01 9.73541617e-01 -3.93437296e-01 4.17129174e-02
-1.02067605e-01 2.02896208e-01 8.70513916e-01 -1.25700486e+00
6.41081274e-01 1.21961868e+00 7.30251074e-01 6.84896231e-01
-1.15626788e+00 -8.00859451e-01 6.23067498e-01 3.22003365e-01
-1.32111526e+00 -5.12668610e-01 9.00695562e-01 2.98989881e-02
8.72288048e-01 7.99532384e-02 6.99359238e-01 9.35155630e-01
-5.34320101e-02 1.21472740e+00 1.21626878e+00 -6.71779931e-01
7.97797665e-02 4.06115621e-01 -2.65676947e-03 8.18361521e-01
3.80288213e-02 1.19312316e-01 -8.06599915e-01 2.87497848e-01
6.36900127e-01 -1.25643000e-01 -1.92455098e-01 -4.26462650e-01
-1.15056264e+00 4.47976530e-01 5.60798645e-01 4.86510426e-01
-2.15822637e-01 -1.01317219e-01 2.06661578e-02 2.29386851e-01
3.86650681e-01 4.88409370e-01 -1.06073475e+00 2.19599783e-01
-9.44264770e-01 -5.96018583e-02 4.68748569e-01 1.22782028e+00
1.30308127e+00 -2.57236622e-02 -1.03547412e-03 1.08105767e+00
3.82939637e-01 1.52771696e-01 6.66272044e-01 -6.90870047e-01
6.26005948e-01 1.06649041e+00 -3.06533068e-01 -3.95776570e-01
-2.80261964e-01 -9.50832963e-01 -7.77060151e-01 -7.56321400e-02
3.52243036e-02 3.07595301e-02 -1.02115035e+00 1.87653708e+00
4.65056449e-01 6.35996222e-01 1.79112107e-01 3.07897270e-01
7.41742730e-01 8.02540243e-01 1.36684135e-01 -3.67083967e-01
7.02355802e-01 -1.54666746e+00 -5.19178927e-01 -7.26760089e-01
1.27526414e+00 -5.92225969e-01 1.18974507e+00 4.04945374e-01
-1.09029806e+00 -1.07513809e+00 -1.14559197e+00 1.06376536e-01
-3.55352551e-01 2.81705022e-01 5.51697969e-01 2.54444331e-01
-1.06201804e+00 5.59024155e-01 -7.03766227e-01 -1.71906948e-01
5.04133224e-01 2.90881872e-01 -2.44807407e-01 -1.86986417e-01
-1.31412888e+00 1.05669689e+00 9.12889183e-01 2.12617904e-01
-9.61562991e-01 -9.89547968e-01 -9.33818102e-01 9.62053146e-03
6.95199311e-01 -3.78464997e-01 1.15815055e+00 -1.09088433e+00
-1.47440553e+00 5.75688839e-01 -2.57937372e-01 -2.03759089e-01
3.87640953e-01 -1.68623894e-01 -4.88153040e-01 -2.13386878e-01
2.53316969e-01 9.40278649e-01 8.24448109e-01 -1.72754633e+00
-8.89944732e-01 -2.06458703e-01 2.14701034e-02 4.74130541e-01
-5.05155325e-01 -3.13439310e-01 -8.44455540e-01 -7.65330017e-01
4.07693893e-01 -8.49257708e-01 -4.43323582e-01 7.86062423e-03
-3.90576839e-01 -3.08268011e-01 9.60222363e-01 -3.93544197e-01
1.66200316e+00 -2.02114868e+00 1.75219268e-01 2.26858303e-01
4.40890305e-02 7.40962803e-01 -4.96896952e-01 3.76867771e-01
-2.90858805e-01 4.86533977e-02 -4.27076876e-01 -5.50603926e-01
-4.41077709e-01 6.14058435e-01 -1.53871059e-01 -1.75673559e-01
4.22757745e-01 9.13976192e-01 -9.74141836e-01 -7.63968050e-01
9.68548954e-02 -8.02038331e-03 -6.45054698e-01 3.60579193e-01
-4.10625160e-01 5.61979592e-01 -3.42334986e-01 6.45555615e-01
7.95241833e-01 -2.22927883e-01 3.01504374e-01 -3.67087305e-01
-2.09305778e-01 3.42155963e-01 -1.37705898e+00 1.92675686e+00
-8.56390953e-01 8.42143819e-02 -7.22231045e-02 -1.14639938e+00
1.06573558e+00 1.36041299e-01 2.25330770e-01 -5.05772352e-01
-2.83227146e-01 3.05147141e-01 2.30531879e-02 -4.33424234e-01
4.92572151e-02 -1.63787425e-01 1.69802278e-01 2.43190765e-01
3.61079156e-01 -1.18859867e-02 1.52454153e-01 2.67520338e-01
7.79599190e-01 7.07052231e-01 2.62058377e-01 -1.69745117e-01
8.27882111e-01 -3.19506600e-02 9.62513626e-01 7.32265830e-01
-8.50383416e-02 1.61716685e-01 1.79544047e-01 -1.52472049e-01
-5.99772334e-01 -1.05295599e+00 -1.72886238e-01 1.23225713e+00
3.62049550e-01 -5.00104964e-01 -6.40589833e-01 -1.26386523e+00
-1.85987100e-01 9.30759251e-01 -5.01664221e-01 -5.06746233e-01
-7.89245307e-01 -7.26183593e-01 2.43795991e-01 9.15322006e-01
9.15134251e-01 -1.10418344e+00 -6.67909384e-02 3.59878212e-01
-1.14022508e-01 -9.90605116e-01 -3.22428674e-01 5.78179061e-01
-1.20650697e+00 -8.04327846e-01 -2.21679524e-01 -1.08721292e+00
8.77665043e-01 2.80053973e-01 9.24515009e-01 4.07204092e-01
1.80537164e-01 -1.17899001e-01 -4.43320036e-01 -2.76774913e-01
-7.33312190e-01 3.58006358e-01 2.55420413e-02 1.10547081e-01
7.42231235e-02 -5.54464102e-01 -1.63460433e-01 5.43863654e-01
-1.00643218e+00 6.08730257e-01 1.03160119e+00 1.01637220e+00
7.22107649e-01 1.55712277e-01 1.00611532e+00 -1.24132681e+00
8.85109454e-02 -4.63167936e-01 -2.31241539e-01 5.19601643e-01
-1.09615314e+00 2.26600021e-01 7.79694855e-01 -5.99709511e-01
-1.59500659e+00 2.49687731e-01 -3.51224691e-01 -4.78745073e-01
-1.00466572e-01 8.54616284e-01 -6.23895764e-01 -6.44062385e-02
7.08382845e-01 3.46284002e-01 -1.22376442e-01 -7.02495515e-01
3.55891675e-01 5.90211809e-01 4.05347198e-01 -6.55423999e-01
9.99172509e-01 2.31269151e-01 -1.50410637e-01 -3.51693124e-01
-1.43655550e+00 -4.69826043e-01 -1.18024802e+00 -1.58862919e-01
3.94345284e-01 -8.84296238e-01 -3.31338271e-02 6.82738602e-01
-8.59733224e-01 -7.99433529e-01 -3.19040567e-01 2.87672937e-01
-4.15712357e-01 2.66201198e-01 -3.46481204e-01 -4.76920217e-01
7.31152762e-03 -1.07722402e+00 9.21010256e-01 2.20221534e-01
2.98716556e-02 -1.13485348e+00 -7.48748928e-02 3.96488070e-01
2.49989927e-01 -3.80909681e-01 8.93493056e-01 -5.97851813e-01
-5.53621590e-01 -2.80559491e-02 -1.92231894e-01 6.44451022e-01
2.88779825e-01 -2.10911229e-01 -9.64750469e-01 -3.63899887e-01
-1.66388527e-01 -4.12219524e-01 1.04140389e+00 -4.31744605e-02
1.23937130e+00 -2.51443386e-01 -7.05330551e-01 7.37339973e-01
1.36768568e+00 1.56245172e-01 5.43633521e-01 4.07894522e-01
8.72585595e-01 6.06977165e-01 1.01628590e+00 3.43449228e-02
3.64503294e-01 5.73576272e-01 3.09071809e-01 -3.06717038e-01
-4.70670491e-01 -6.78418994e-01 3.96580547e-01 1.11650562e+00
1.39104575e-01 2.41066031e-02 -8.18636179e-01 5.26115656e-01
-2.01117301e+00 -5.25562763e-01 -5.23031801e-02 1.93313479e+00
1.19000041e+00 4.47815567e-01 -2.90646762e-01 1.38079956e-01
5.33853710e-01 2.51747780e-02 -7.25309253e-01 -1.66174248e-02
-1.56859010e-01 3.48382831e-01 2.22034857e-01 5.14532268e-01
-9.66738105e-01 1.55713880e+00 5.45950747e+00 1.23130643e+00
-1.07598829e+00 2.54228920e-01 4.06496137e-01 2.48360872e-01
-3.46904546e-01 3.29174668e-01 -1.22180057e+00 4.22352642e-01
7.19000399e-01 5.07360995e-02 1.80664539e-01 1.06332052e+00
1.67458087e-01 -2.17502058e-01 -1.20030487e+00 5.96737206e-01
3.00806742e-02 -1.47542012e+00 3.25661391e-01 -3.12525451e-01
8.37716937e-01 -2.14246288e-01 -1.22259140e-01 8.33736479e-01
3.02626491e-01 -6.30261421e-01 7.54784524e-01 4.35508549e-01
7.48777270e-01 -6.05776727e-01 5.52426457e-01 8.69809985e-01
-1.44504237e+00 -1.66453972e-01 -1.95409060e-01 2.05360487e-01
2.95872018e-02 4.45399791e-01 -9.21592712e-01 6.30790055e-01
6.40988827e-01 9.26970065e-01 -7.98993647e-01 8.27440441e-01
-7.10395753e-01 8.73353899e-01 -1.37201115e-01 1.80018738e-01
3.58538091e-01 1.28033105e-02 4.17674273e-01 1.16974568e+00
1.28869861e-01 -1.51984960e-01 3.82366896e-01 7.51977324e-01
6.77435426e-03 1.89014841e-02 -3.30306500e-01 3.22209090e-01
6.15806341e-01 1.14777243e+00 -3.96592498e-01 -4.89426464e-01
-4.06362355e-01 8.67768764e-01 6.63653731e-01 3.21936309e-01
-8.00467134e-01 -2.94590481e-02 2.73312271e-01 1.39952719e-01
2.34863743e-01 6.11212011e-03 -3.73065889e-01 -1.19136524e+00
5.33893183e-02 -7.80548871e-01 4.33716953e-01 -8.40410054e-01
-1.19003046e+00 6.12662017e-01 1.23494476e-01 -1.29544973e+00
-5.22956140e-02 -4.60423142e-01 -5.68289816e-01 8.47089827e-01
-1.85616720e+00 -1.61606944e+00 -1.89990088e-01 5.55255711e-01
9.27071393e-01 -2.87138104e-01 6.06945693e-01 2.52360553e-01
-8.78642797e-01 6.63761437e-01 -1.63748428e-01 -2.98606813e-01
6.86454117e-01 -1.06792128e+00 1.42946869e-01 9.44450974e-01
1.44686893e-01 6.01284325e-01 3.07749629e-01 -8.90356898e-01
-1.04804873e+00 -1.56062150e+00 1.06275237e+00 -3.47325057e-01
5.68464816e-01 -2.01227367e-01 -1.17124462e+00 9.18839514e-01
-1.61413193e-01 3.60949710e-02 6.29129291e-01 1.49356842e-01
-2.70886660e-01 -3.88023168e-01 -8.89527857e-01 6.92020118e-01
1.42307699e+00 -2.86948264e-01 -7.48016834e-01 1.71608046e-01
8.12411666e-01 -2.86786348e-01 -5.21667659e-01 8.54239583e-01
1.27834663e-01 -8.16388130e-01 8.98359835e-01 -6.93949461e-01
3.32739204e-01 -4.89297897e-01 3.72969620e-02 -1.57063091e+00
-4.85465676e-01 -3.78635794e-01 -3.37045401e-01 1.33319962e+00
8.04775357e-01 -5.91248214e-01 8.48919630e-01 2.70096630e-01
-7.56315887e-01 -1.33646321e+00 -7.23119438e-01 -9.75957513e-01
9.90071241e-03 -6.57063544e-01 5.66374958e-01 1.16296756e+00
-2.08511472e-01 4.53842193e-01 -3.24917316e-01 2.61526048e-01
4.18759406e-01 1.46646380e-01 7.21339643e-01 -1.10407650e+00
-5.60693204e-01 -2.85807401e-01 2.18019918e-01 -1.41995716e+00
1.79981068e-01 -1.31929195e+00 5.18485196e-02 -1.56802177e+00
2.86967933e-01 -1.12288451e+00 -5.64303160e-01 9.43847895e-01
-5.37976980e-01 1.30441502e-01 8.76441747e-02 3.22401822e-01
-6.60751998e-01 5.26134908e-01 1.51576006e+00 -2.11980771e-02
-4.35036033e-01 1.88613668e-01 -7.75187910e-01 9.34162378e-01
6.70818686e-01 -4.97779608e-01 -7.54486322e-01 -3.47502917e-01
1.19346932e-01 -2.15010598e-01 1.55482709e-01 -1.02277148e+00
3.03469181e-01 -3.08503479e-01 4.57134515e-01 -6.21476531e-01
1.96691677e-01 -7.84810662e-01 -5.19574657e-02 4.20610577e-01
-4.81143087e-01 -3.53835493e-01 2.84298062e-01 5.87498128e-01
-2.16186628e-01 -5.64363718e-01 1.04482591e+00 -1.81553885e-01
-1.07046616e+00 3.59708011e-01 -3.23636718e-02 -8.04832652e-02
1.06069469e+00 -2.81203598e-01 -5.31111956e-02 2.88040899e-02
-9.26874280e-01 5.51868379e-01 1.63817987e-01 4.18310195e-01
6.30603075e-01 -1.28421950e+00 -5.45865536e-01 4.46376115e-01
2.75929421e-01 4.46314126e-01 2.35336974e-01 7.56449938e-01
-1.00308456e-01 2.22337663e-01 2.94499751e-02 -5.61159134e-01
-9.78387415e-01 7.89660871e-01 2.53829837e-01 -6.35881424e-01
-3.72207969e-01 1.13263309e+00 3.57219815e-01 -8.40797424e-01
2.90859997e-01 -2.64768839e-01 -1.65573001e-01 -1.21653743e-01
1.96658209e-01 1.27139822e-01 1.57167599e-01 -3.75432014e-01
-2.04482183e-01 3.93874347e-01 -2.67696083e-01 9.88684446e-02
1.31562936e+00 -1.76594153e-01 -5.50275147e-02 4.67075497e-01
9.63245332e-01 -8.50705206e-02 -1.37116015e+00 -6.83736980e-01
2.44495898e-01 -2.91208297e-01 8.30172524e-02 -1.09221482e+00
-9.56135511e-01 7.44943440e-01 2.46424034e-01 -1.94351986e-01
1.34803009e+00 2.47135505e-01 8.40288103e-01 4.38857406e-01
3.77498448e-01 -1.33372295e+00 2.54933745e-01 4.93060559e-01
9.10816014e-01 -1.04472983e+00 -1.73954323e-01 -7.60832667e-01
-6.49262428e-01 9.09778774e-01 1.18887413e+00 2.54235387e-01
6.98601723e-01 3.05822045e-01 -2.52076890e-02 1.70792088e-01
-9.29793715e-01 -3.53231162e-01 2.86076456e-01 6.03088379e-01
7.63746500e-02 -2.75515378e-01 -1.14282928e-01 7.44521618e-01
1.31253496e-01 1.26856595e-01 -3.21139145e-04 9.62129533e-01
-5.35322428e-01 -1.60234928e+00 -4.30048741e-02 4.94050384e-01
2.30918989e-01 -3.27118158e-01 -1.19564027e-01 7.61306524e-01
6.98435426e-01 7.18752623e-01 -1.14573672e-01 -4.76413667e-01
3.70488971e-01 1.91221237e-01 4.14371520e-01 -1.08280623e+00
-4.40418869e-01 1.04349628e-01 1.73985779e-01 -4.05008823e-01
-4.84680474e-01 -4.99891192e-01 -1.43480194e+00 3.86844903e-01
-5.37860155e-01 7.60827139e-02 3.66629124e-01 1.36483467e+00
3.00243944e-01 5.39249957e-01 8.51585507e-01 -6.84676647e-01
-4.82777387e-01 -1.04220223e+00 -2.58723408e-01 1.55668348e-01
-4.10048384e-03 -8.00881803e-01 -2.96822727e-01 1.78306177e-01] | [9.496569633483887, 3.2047581672668457] |
e3c00993-d1a6-430c-a000-a06f2de51313 | differentiable-supervector-extraction-for | 1812.09484 | null | http://arxiv.org/abs/1812.09484v1 | http://arxiv.org/pdf/1812.09484v1.pdf | Differentiable Supervector Extraction for Encoding Speaker and Phrase Information in Text Dependent Speaker Verification | In this paper, we propose a new differentiable neural network alignment
mechanism for text-dependent speaker verification which uses alignment models
to produce a supervector representation of an utterance. Unlike previous works
with similar approaches, we do not extract the embedding of an utterance from
the mean reduction of the temporal dimension. Our system replaces the mean by a
phrase alignment model to keep the temporal structure of each phrase which is
relevant in this application since the phonetic information is part of the
identity in the verification task. Moreover, we can apply a convolutional
neural network as front-end, and thanks to the alignment process being
differentiable, we can train the whole network to produce a supervector for
each utterance which will be discriminative with respect to the speaker and the
phrase simultaneously. As we show, this choice has the advantage that the
supervector encodes the phrase and speaker information providing good
performance in text-dependent speaker verification tasks. In this work, the
process of verification is performed using a basic similarity metric, due to
simplicity, compared to other more elaborate models that are commonly used. The
new model using alignment to produce supervectors was tested on the
RSR2015-Part I database for text-dependent speaker verification, providing
competitive results compared to similar size networks using the mean to extract
embeddings. | ['Alfonso Ortega', 'Victoria Mingote', 'Eduardo Lleida', 'Antonio Miguel'] | 2018-12-22 | null | null | null | null | ['text-dependent-speaker-verification'] | ['speech'] | [ 1.27670884e-01 1.47923380e-01 1.07535072e-01 -7.66735911e-01
-6.35966539e-01 -4.63124961e-01 7.02777147e-01 5.88042289e-02
-7.72805035e-01 2.09514573e-01 3.13404322e-01 -3.15704077e-01
-1.45705016e-02 -4.27932382e-01 -5.42672276e-01 -9.91006613e-01
-1.35661494e-02 3.17634195e-01 -2.84808189e-01 -3.35634261e-01
1.18637867e-01 6.07179999e-01 -1.49520695e+00 1.40064023e-02
4.59426433e-01 8.72343004e-01 -1.90256834e-02 6.90185308e-01
6.23811930e-02 2.06454366e-01 -7.16593385e-01 -4.45468575e-01
2.68148184e-01 -5.59715092e-01 -7.18069792e-01 -5.47033697e-02
6.28490210e-01 -1.81618541e-01 -2.74419963e-01 8.94998014e-01
7.63319671e-01 2.93211937e-01 7.70910084e-01 -9.75477278e-01
-4.16729659e-01 8.06250095e-01 -2.50582635e-01 9.56262425e-02
2.18863189e-01 -1.86562657e-01 1.14379239e+00 -8.26865792e-01
4.87767965e-01 1.18830144e+00 6.35679305e-01 7.83997655e-01
-1.31388509e+00 -5.68873346e-01 2.29842380e-01 4.69955206e-01
-1.42102015e+00 -8.14132571e-01 1.02897680e+00 -2.15011984e-01
9.52568829e-01 2.00697199e-01 3.47029984e-01 1.15282667e+00
6.77324459e-02 6.06645763e-01 7.28648424e-01 -6.46088481e-01
1.86342046e-01 3.00139159e-01 3.38372618e-01 4.03723210e-01
-5.54526865e-01 1.82260796e-01 -5.95006526e-01 1.16568692e-01
1.74794346e-01 -2.96250582e-01 -3.76195550e-01 -3.69770437e-01
-1.06423116e+00 9.45027053e-01 3.93798888e-01 7.01424301e-01
-3.45443428e-01 -4.33354713e-02 6.76414549e-01 5.10653794e-01
2.93659568e-01 -3.42408754e-02 -3.10675383e-01 -4.46902029e-02
-1.17456996e+00 4.47136909e-02 7.88736105e-01 4.77833033e-01
5.67708373e-01 3.00204039e-01 -2.76084095e-01 8.96049976e-01
5.12680590e-01 5.02643168e-01 9.09497976e-01 -5.29923081e-01
3.59396845e-01 7.91316107e-02 -2.85722703e-01 -7.62431622e-01
-3.18325609e-01 -5.53099632e-01 -8.87147307e-01 2.70435691e-01
4.57086027e-01 -7.16788396e-02 -9.66166139e-01 2.11640239e+00
2.60205656e-01 2.34855697e-01 3.35622489e-01 7.69326568e-01
8.42698276e-01 5.75737894e-01 -1.72278360e-01 -9.90035683e-02
1.46952116e+00 -8.21960211e-01 -8.89806211e-01 -2.62240451e-02
6.53801322e-01 -7.44437456e-01 7.70643115e-01 1.97745532e-01
-9.05438304e-01 -8.97463202e-01 -1.36173534e+00 -5.29625341e-02
-4.64532316e-01 1.78790703e-01 -5.77256642e-02 8.49608302e-01
-1.42536604e+00 5.65768778e-01 -7.24532962e-01 -3.20963472e-01
-1.25531673e-01 5.80821633e-01 -6.11864626e-01 3.06744099e-01
-1.38500059e+00 1.23670721e+00 4.36999559e-01 2.42897406e-01
-6.85566545e-01 -3.75222743e-01 -1.22230959e+00 1.89881533e-01
-3.18363726e-01 -4.28220630e-01 1.20156646e+00 -9.03514147e-01
-1.91522431e+00 7.30236173e-01 -6.17808700e-01 -8.02541196e-01
4.60296631e-01 2.46728271e-01 -3.40888739e-01 -3.66389975e-02
-2.78675258e-01 7.04228938e-01 1.24227726e+00 -7.95346916e-01
-4.74064201e-01 -4.53653932e-01 -1.77424043e-01 9.82490778e-02
-4.46211517e-01 3.92682254e-02 -2.08327964e-01 -4.08838540e-01
2.49337152e-01 -9.19628620e-01 8.62869024e-02 -4.10589397e-01
-2.51294762e-01 -4.22494829e-01 1.09364736e+00 -1.01931620e+00
9.36596215e-01 -2.58837032e+00 4.12781417e-01 2.58488744e-01
-1.96095854e-01 2.52977818e-01 -3.72764051e-01 3.08357775e-01
-3.92650217e-01 -1.17599599e-01 -4.40979332e-01 -1.21152806e+00
2.18111441e-01 2.21764103e-01 -3.04598719e-01 8.29175770e-01
2.88481593e-01 7.66428590e-01 -3.83444101e-01 -3.80501956e-01
2.95294851e-01 1.02422369e+00 -2.77477890e-01 3.43559645e-02
2.61633396e-01 4.44083363e-01 1.05332226e-01 3.62606868e-02
8.05352032e-01 5.88645816e-01 -2.69702375e-02 -7.60076568e-02
-5.09676971e-02 7.20951378e-01 -1.10663760e+00 1.86293912e+00
-8.34848285e-01 8.12078953e-01 3.71468157e-01 -1.16039205e+00
1.24475336e+00 8.51865470e-01 2.54247636e-01 -5.93094885e-01
2.57297605e-01 2.25625306e-01 2.05980226e-01 -9.53251570e-02
3.98525506e-01 -3.90584350e-01 9.14747939e-02 6.34186566e-01
3.47226024e-01 -1.70533374e-01 1.79514319e-01 -9.91170779e-02
5.84664822e-01 1.12711098e-02 5.47285341e-02 -7.14694709e-02
1.00321901e+00 -5.60835660e-01 4.15680468e-01 3.73600781e-01
-2.95849830e-01 6.34911835e-01 1.23324566e-01 -1.93022192e-01
-8.21812272e-01 -8.37175548e-01 -3.67591411e-01 9.12325382e-01
-2.29307398e-01 -2.29237661e-01 -8.65491033e-01 -9.02121544e-01
-1.83964297e-01 8.72980297e-01 -6.79287314e-01 -1.48020178e-01
-9.28025723e-01 -2.56028682e-01 6.51059628e-01 3.73409152e-01
2.43176758e-01 -1.17482436e+00 -1.45442858e-01 2.90693313e-01
-2.56829232e-01 -1.03304505e+00 -7.97263265e-01 6.40237033e-01
-6.89976454e-01 -5.49385130e-01 -7.71508992e-01 -1.08458555e+00
5.25212348e-01 -1.71522781e-01 5.50450742e-01 -1.88363478e-01
1.74007595e-01 3.09322178e-02 -1.60727978e-01 -3.25373113e-01
-8.07867110e-01 1.52542770e-01 1.95940912e-01 3.63627136e-01
5.31800389e-01 -6.89287424e-01 -7.54796639e-02 1.51112348e-01
-7.73248017e-01 -4.49056417e-01 4.04809773e-01 1.03903031e+00
1.86268777e-01 1.37504758e-02 6.95820034e-01 -2.22177282e-01
4.88366723e-01 8.52376409e-03 -6.42260015e-01 7.35314004e-03
-3.89895350e-01 4.21005875e-01 7.65217602e-01 -4.38964963e-01
-9.56647038e-01 3.64108860e-01 -6.38044178e-01 -3.44487906e-01
-4.56046432e-01 3.25176001e-01 -3.42157871e-01 1.13586836e-01
4.64512467e-01 5.06044865e-01 3.59081477e-01 -6.09790623e-01
3.96964699e-01 9.45150018e-01 4.82050776e-01 -6.89473599e-02
8.51681292e-01 3.52149993e-01 -1.02236122e-01 -8.52627099e-01
-4.28092450e-01 -6.53474331e-01 -1.15546906e+00 1.17854312e-01
9.26930428e-01 -5.80140471e-01 -7.96260178e-01 3.67720187e-01
-1.53223741e+00 1.93375885e-01 -3.56283873e-01 8.02142143e-01
-5.08198380e-01 5.76924443e-01 -4.63482648e-01 -8.12320292e-01
-5.47320545e-01 -1.38716578e+00 1.04707599e+00 4.72819582e-02
-2.18070865e-01 -1.09419227e+00 2.25604445e-01 5.20315915e-02
5.67779541e-01 -2.63238549e-01 7.79880702e-01 -1.15602827e+00
-1.10716157e-01 -3.00240427e-01 1.97416410e-01 9.28652167e-01
2.71350354e-01 6.51898608e-03 -1.45885313e+00 -4.16215628e-01
5.90002477e-01 2.05707952e-01 9.40224946e-01 4.06251013e-01
4.99494880e-01 -2.57660031e-01 1.51121803e-03 5.90698600e-01
9.83717799e-01 2.12777913e-01 5.34829497e-01 1.03063516e-01
4.12990510e-01 8.77826691e-01 3.17131311e-01 1.67879220e-02
1.51002347e-01 9.94388282e-01 2.76038796e-01 -1.48343891e-01
-2.20530868e-01 4.32217680e-02 7.85409093e-01 1.03822291e+00
2.67572284e-01 -1.92893259e-02 -6.84243083e-01 6.96039438e-01
-1.61300528e+00 -1.10777569e+00 9.50045064e-02 2.43210673e+00
6.15425229e-01 2.39883680e-02 9.48209837e-02 5.43775737e-01
7.67486989e-01 3.04831117e-01 -8.56088921e-02 -9.01319802e-01
-2.00283632e-01 4.19453025e-01 1.67415500e-01 8.25613856e-01
-1.08983243e+00 6.62905574e-01 5.68189096e+00 6.78738058e-01
-1.65279269e+00 3.03123742e-01 1.19280227e-01 6.95324168e-02
2.43109223e-02 -2.08663553e-01 -9.99159157e-01 2.57982463e-01
1.33987904e+00 -2.01954506e-02 1.46893129e-01 6.68819904e-01
3.98465008e-01 3.77938479e-01 -1.56126046e+00 8.80289853e-01
3.64720881e-01 -7.54284859e-01 -1.75625652e-01 2.41492577e-02
3.45851392e-01 -1.51145263e-02 1.95720479e-01 2.52066553e-01
-1.61443904e-01 -1.03216720e+00 9.00244772e-01 1.73692331e-01
4.31476235e-01 -9.47551429e-01 1.04089439e+00 2.57387727e-01
-1.18523371e+00 2.97123007e-02 -1.97238877e-01 1.28774807e-01
3.59555453e-01 3.59537423e-01 -1.22223115e+00 6.54424429e-01
3.43699872e-01 6.02011085e-01 -2.96211958e-01 8.16395521e-01
-4.35130566e-01 5.21355569e-01 -2.75634587e-01 1.64603099e-01
2.77263850e-01 -1.21452756e-01 7.36143827e-01 1.43918121e+00
2.62137324e-01 -7.60406375e-01 -1.97176307e-01 6.91645563e-01
4.26792763e-02 3.12700301e-01 -5.64406693e-01 1.73287272e-01
2.37941816e-01 1.01442921e+00 -2.80513436e-01 -3.05756986e-01
-9.38170999e-02 1.10402250e+00 5.66955730e-02 2.82434165e-01
-7.42974520e-01 -6.40413404e-01 7.25195825e-01 -3.30692321e-01
7.29763389e-01 -2.46067986e-01 -1.19735755e-01 -8.46050680e-01
2.52771169e-01 -8.52541089e-01 2.07702219e-01 -2.50146806e-01
-1.01049590e+00 1.03120005e+00 -2.16353893e-01 -1.10835063e+00
-8.38054240e-01 -4.70236182e-01 -7.32404649e-01 1.40941668e+00
-1.69191468e+00 -1.24426508e+00 8.39549676e-02 4.88757104e-01
5.46700478e-01 -1.68879539e-01 1.12857842e+00 3.42918307e-01
-4.26077724e-01 8.71489286e-01 -8.85800645e-03 3.85929406e-01
8.62075627e-01 -1.23888087e+00 6.08412445e-01 9.23325837e-01
4.63460326e-01 7.55005419e-01 6.38072908e-01 -1.27532795e-01
-9.88674462e-01 -7.31796861e-01 1.42899656e+00 -2.21042499e-01
4.95398790e-01 -6.48419201e-01 -1.01127875e+00 5.56626439e-01
4.92782146e-01 -1.69289097e-01 5.49425423e-01 1.73489913e-01
-5.30849576e-01 -2.94176072e-01 -1.06686783e+00 2.77816802e-01
6.62411571e-01 -1.00060964e+00 -8.77366841e-01 8.06646422e-03
6.60562575e-01 -1.77763864e-01 -6.87611401e-01 3.52068871e-01
6.13776267e-01 -8.45728576e-01 7.50166953e-01 -3.06016654e-01
-1.22590333e-01 -4.60648775e-01 -8.58630538e-02 -1.47688806e+00
-9.77791920e-02 -4.42890465e-01 1.15859710e-01 1.56247127e+00
6.44270182e-01 -7.22284377e-01 5.58687747e-01 2.65270919e-01
-2.17815578e-01 -2.47746944e-01 -1.49235761e+00 -9.03351903e-01
7.70279318e-02 -4.55414593e-01 7.73338139e-01 7.64892697e-01
-2.72817388e-02 4.71523702e-01 -2.71143734e-01 3.59721780e-01
4.28141773e-01 4.00561243e-02 5.68626583e-01 -1.03914428e+00
-4.30974782e-01 -6.41226411e-01 -7.03397334e-01 -1.02890658e+00
7.18459487e-01 -1.10570073e+00 2.99000919e-01 -1.24552655e+00
-1.50859341e-01 -1.61565110e-01 -5.01722217e-01 4.85061109e-01
7.70723447e-02 1.57114953e-01 2.17968881e-01 -9.07676816e-02
1.47869870e-01 6.84307575e-01 7.49950886e-01 -5.63678145e-01
-3.24285388e-01 3.12046587e-01 -2.61660963e-01 4.01079506e-01
8.11679542e-01 -5.01664758e-01 -1.46544442e-01 -2.94429302e-01
-6.40039027e-01 -1.30639136e-01 1.13083310e-01 -8.72405887e-01
2.81132996e-01 5.67571700e-01 1.01067752e-01 -6.58814669e-01
7.77815402e-01 -1.02961886e+00 -1.76856425e-02 6.10564828e-01
-4.09532785e-01 -1.86406374e-02 2.70491719e-01 7.00231120e-02
-6.58658803e-01 -5.84974825e-01 9.34600770e-01 3.91240418e-01
-4.12223637e-01 -4.36823852e-02 -2.86008388e-01 -6.39409840e-01
6.74941599e-01 -2.04671562e-01 6.93181828e-02 -5.35502911e-01
-8.78519297e-01 -1.53709874e-01 3.01106632e-01 5.49428523e-01
4.45002973e-01 -1.26440406e+00 -1.00775063e+00 4.77198184e-01
-3.61514725e-02 -3.41182470e-01 2.23754853e-01 1.04855907e+00
6.55673295e-02 6.27013087e-01 -1.02744967e-01 -8.66891682e-01
-1.71695447e+00 5.87035894e-01 5.86400688e-01 -1.32269308e-01
-4.37769234e-01 7.33916998e-01 2.06403971e-01 -6.73162758e-01
5.30712008e-01 -3.46262097e-01 -3.48754972e-01 3.63816887e-01
5.98164320e-01 -1.13967866e-01 5.52250087e-01 -1.20558035e+00
-5.45961261e-01 5.95598519e-01 -1.48068652e-01 -5.48495233e-01
1.37406611e+00 -1.43628076e-01 -3.45780551e-02 4.33194190e-01
1.67535353e+00 2.76201695e-01 -8.62866521e-01 -3.30247819e-01
2.24423539e-02 -3.38555872e-02 1.84804767e-01 -4.40767407e-01
-1.04284072e+00 1.14770794e+00 7.55970478e-01 6.03976436e-02
9.94807899e-01 -1.42388597e-01 5.28661430e-01 2.70670384e-01
-1.17540456e-01 -8.71878624e-01 -3.95960420e-01 6.92469239e-01
1.02676964e+00 -1.13523340e+00 -4.25837874e-01 3.34960669e-02
-3.70702356e-01 1.19678998e+00 1.32008448e-01 1.60976693e-01
7.29264319e-01 2.16167897e-01 4.81520683e-01 1.14339165e-01
-3.31427485e-01 -2.19417706e-01 5.21896362e-01 6.68911457e-01
5.33570647e-01 -9.16148648e-02 -3.44550014e-01 1.66599274e-01
-6.20898426e-01 -5.03816068e-01 2.73434967e-01 6.55636549e-01
-9.64057297e-02 -1.54262877e+00 -4.87512827e-01 -2.33349100e-01
-4.34569389e-01 -1.29048064e-01 -4.34714943e-01 5.77249885e-01
2.81520393e-02 1.08371508e+00 1.19408511e-01 -3.68011296e-01
3.17670882e-01 6.20939612e-01 3.13048154e-01 -5.13419032e-01
-8.51518095e-01 3.29968147e-02 3.05894320e-03 -2.67815620e-01
-4.07547534e-01 -6.79016232e-01 -1.23334277e+00 -1.30932601e-02
-5.63238919e-01 4.13334817e-01 1.54977143e+00 1.02497768e+00
2.53864944e-01 4.10969734e-01 8.35190654e-01 -9.06669319e-01
-7.22658932e-01 -1.27182138e+00 -5.35014510e-01 4.83997971e-01
8.21411371e-01 -3.66856188e-01 -5.28420568e-01 1.11952849e-01] | [14.332198143005371, 6.086800575256348] |
87c77338-285c-478d-9e25-c5bad8bdb190 | recouple-event-field-via-probabilistic-bias | 2305.11498 | null | https://arxiv.org/abs/2305.11498v1 | https://arxiv.org/pdf/2305.11498v1.pdf | Recouple Event Field via Probabilistic Bias for Event Extraction | Event Extraction (EE), aiming to identify and classify event triggers and arguments from event mentions, has benefited from pre-trained language models (PLMs). However, existing PLM-based methods ignore the information of trigger/argument fields, which is crucial for understanding event schemas. To this end, we propose a Probabilistic reCoupling model enhanced Event extraction framework (ProCE). Specifically, we first model the syntactic-related event fields as probabilistic biases, to clarify the event fields from ambiguous entanglement. Furthermore, considering multiple occurrences of the same triggers/arguments in EE, we explore probabilistic interaction strategies among multiple fields of the same triggers/arguments, to recouple the corresponding clarified distributions and capture more latent information fields. Experiments on EE datasets demonstrate the effectiveness and generalization of our proposed approach. | ['Yujiu Yang', 'Weigang Guo', 'Qi Ju', 'Weijie Liu', 'Jiayi Li', 'Xuefeng Yang', 'Zhe Zhao', 'Han Guo', 'Taiqiang Wu', 'Xingyu Bai'] | 2023-05-19 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 1.28055066e-01 1.29721999e-01 -2.50278622e-01 -4.61951554e-01
-6.01910591e-01 -7.55902946e-01 1.00485837e+00 8.59663546e-01
-3.58293563e-01 9.04528856e-01 9.14720833e-01 -1.34927258e-01
-4.37751085e-01 -1.07662928e+00 -7.05842316e-01 -7.10716367e-01
-8.09265375e-02 1.64769411e-01 3.34440917e-01 -4.58274549e-03
6.96061999e-02 3.40596974e-01 -1.08424103e+00 4.16151524e-01
7.65038371e-01 7.44861305e-01 7.25290477e-02 -2.58050188e-02
-4.58902121e-01 1.14077866e+00 -4.73630428e-01 -4.44639355e-01
-4.15673107e-01 -2.69197881e-01 -7.76521802e-01 -4.58115608e-01
-5.55279315e-01 -3.25472057e-02 -5.28159618e-01 8.81806791e-01
1.81360379e-01 3.83983813e-02 8.44204605e-01 -1.25516057e+00
-1.90006927e-01 1.28388250e+00 -3.29080433e-01 5.17052233e-01
2.93126255e-01 -1.09750286e-01 1.16046000e+00 -7.63213098e-01
6.85106277e-01 1.16925323e+00 3.30306858e-01 -4.80734855e-02
-9.33111429e-01 -6.59193456e-01 3.39148641e-01 3.49501103e-01
-1.19528437e+00 -2.31594130e-01 1.07387507e+00 -2.52213091e-01
7.55405843e-01 8.62702820e-03 4.21668828e-01 1.59564316e+00
1.42323360e-01 9.70803797e-01 1.00937927e+00 -2.65284240e-01
2.21662596e-01 2.55462043e-02 5.17478347e-01 6.75762370e-02
3.49427998e-01 1.63416043e-01 -8.61374259e-01 -1.74768791e-01
3.35787117e-01 6.17347134e-04 -8.16116575e-03 3.81633779e-03
-1.36699307e+00 5.47585487e-01 1.88955992e-01 4.46280032e-01
-6.86304331e-01 -1.09438047e-01 5.44877708e-01 -4.86754477e-01
2.10304514e-01 9.73254144e-02 -6.03073478e-01 -1.70011491e-01
-6.24631166e-01 2.75386572e-01 5.42076528e-01 9.53958094e-01
6.47203803e-01 -4.05439854e-01 -5.86131930e-01 3.13165992e-01
4.51221943e-01 1.52154833e-01 6.75635114e-02 -4.05370116e-01
7.17153311e-01 8.07458639e-01 3.10074091e-01 -9.95345354e-01
-5.61834872e-01 -6.01230502e-01 -7.22187877e-01 -6.64564669e-01
1.39123842e-01 -1.54307812e-01 -4.70915079e-01 2.06402993e+00
4.91569459e-01 4.08444047e-01 2.83885598e-01 6.70888603e-01
8.72881353e-01 7.39572108e-01 7.75007606e-01 -2.74203032e-01
1.59218693e+00 -2.34041080e-01 -9.31511283e-01 -2.07944825e-01
4.16741639e-01 -4.50756758e-01 6.00954354e-01 -2.42973454e-02
-7.68467128e-01 -8.58057737e-02 -7.47526050e-01 -4.54996638e-02
-2.74882197e-01 3.73915173e-02 9.38117445e-01 1.81043327e-01
1.53421864e-01 2.96934962e-01 -9.66127157e-01 -4.45083976e-02
4.38203484e-01 -9.24583301e-02 -3.51938531e-02 1.72621429e-01
-1.69621420e+00 7.07664192e-01 1.07650316e+00 1.42707318e-01
-9.82102811e-01 -7.64294386e-01 -8.85738850e-01 4.30486411e-01
8.41680944e-01 -3.09118092e-01 9.22020972e-01 -2.71973491e-01
-1.01278579e+00 2.84368157e-01 -4.19942588e-01 -4.85971212e-01
2.85490379e-02 -3.36851686e-01 -6.71312928e-01 1.79643527e-01
2.11972073e-01 2.50353009e-01 3.15964311e-01 -1.05437541e+00
-5.22754252e-01 -3.70528512e-02 1.78931579e-01 -8.04956630e-03
-1.77066162e-01 3.80117476e-01 -1.06322698e-01 -7.27941930e-01
5.20426184e-02 -4.40210670e-01 3.95990647e-02 -7.52134681e-01
-6.88543439e-01 -5.66000819e-01 3.36875021e-01 -4.04420674e-01
1.36628342e+00 -2.31122375e+00 2.48674348e-01 1.56692728e-01
1.53206006e-01 -3.24087441e-01 3.29842865e-01 6.66036963e-01
-6.99465945e-02 -1.42969564e-02 -2.14934424e-01 -1.26393929e-01
3.02590728e-01 4.14728165e-01 -9.46492076e-01 4.34280559e-02
5.79264462e-01 6.86382413e-01 -1.01140368e+00 -8.02223861e-01
2.83451825e-01 2.18063682e-01 -2.77727962e-01 2.11557686e-01
-4.14547622e-01 6.15993559e-01 -7.92813063e-01 5.51953137e-01
7.13897407e-01 -1.91191718e-01 3.99721295e-01 -7.65387774e-01
-3.03035140e-01 9.47978020e-01 -1.40945411e+00 1.43543088e+00
-1.77790537e-01 1.48208290e-02 -3.12285244e-01 -9.33187366e-01
5.98932981e-01 4.80134636e-01 4.91215676e-01 -4.60253268e-01
2.76485026e-01 -3.82445380e-02 -2.54245102e-01 -3.66710514e-01
4.61181879e-01 -2.53343850e-01 -6.02180004e-01 3.48089367e-01
3.15827429e-01 4.64946330e-01 2.69500375e-01 5.86271048e-01
9.64485168e-01 2.93306261e-01 4.34565246e-01 -1.67839706e-01
3.98789436e-01 -1.04133569e-01 1.05477679e+00 7.92381823e-01
-2.10372042e-02 1.99637383e-01 8.45364571e-01 -1.87889248e-01
-5.78772724e-01 -1.42661047e+00 -3.86984020e-01 7.64335752e-01
5.02801001e-01 -9.17555094e-01 -3.11617136e-01 -9.16327894e-01
-4.85318810e-01 1.16512954e+00 -4.44624960e-01 -1.65185124e-01
-6.29977167e-01 -1.21985197e+00 6.76967680e-01 7.29322970e-01
4.64231253e-01 -1.05672228e+00 -4.23386127e-01 6.38016701e-01
-6.95344090e-01 -1.29703176e+00 -4.68295813e-02 4.41050738e-01
-4.29638475e-01 -9.49813485e-01 2.44626962e-02 -1.81112155e-01
2.47773290e-01 -3.20183694e-01 9.73744094e-01 -5.15846550e-01
1.80238545e-01 -9.35368091e-02 -4.85641748e-01 -5.69271088e-01
-1.86659738e-01 9.80283413e-03 -6.59518465e-02 5.38870767e-02
4.69157249e-01 -8.30315351e-01 -4.94176805e-01 1.65963918e-01
-1.11240125e+00 2.60975480e-01 6.04089856e-01 5.07744670e-01
5.14542162e-01 2.14497775e-01 6.79900765e-01 -7.70297885e-01
3.62977892e-01 -9.97114599e-01 -3.61645013e-01 4.20554608e-01
-3.40630174e-01 3.75150651e-01 5.41980028e-01 -5.63149035e-01
-1.68610811e+00 -3.41440111e-01 -1.46844881e-02 1.16023794e-01
-5.03134906e-01 6.87054038e-01 -8.26359808e-01 8.44892919e-01
2.83515692e-01 3.11381280e-01 -1.05288851e+00 -2.80977339e-01
3.84213090e-01 2.92022973e-01 5.40150404e-01 -1.15766335e+00
4.76286083e-01 5.30002236e-01 2.90729646e-02 -1.65807962e-01
-1.13002455e+00 -2.16253325e-01 -3.22980225e-01 -5.02583943e-03
9.33820009e-01 -8.85736525e-01 -6.23268306e-01 2.32212022e-01
-1.39946604e+00 3.52755487e-01 -2.23255113e-01 8.98933232e-01
-1.91538811e-01 3.28153342e-01 -6.94262087e-01 -7.60090709e-01
-5.33319684e-03 -9.76183057e-01 1.10487735e+00 5.55889010e-01
-3.43557715e-01 -7.74125934e-01 1.19917892e-01 -1.22446865e-01
8.05549417e-03 2.79770285e-01 1.18885481e+00 -1.01602054e+00
-6.79114103e-01 8.01876485e-02 -3.41606587e-01 -3.41807812e-01
-1.45774148e-02 -2.66234666e-01 -7.50166059e-01 4.65668380e-01
-4.45235483e-02 2.99728271e-02 6.86693251e-01 5.72736859e-02
9.47408557e-01 -2.84214318e-01 -7.02554822e-01 1.98212087e-01
1.13926065e+00 1.70536175e-01 6.68744147e-01 2.30627462e-01
4.68309253e-01 6.88756824e-01 5.35088897e-01 6.38689637e-01
6.84965074e-01 3.86758119e-01 1.73942387e-01 3.25202376e-01
1.87392101e-01 -7.16096818e-01 4.42626536e-01 6.26112282e-01
1.12109490e-01 -6.05747700e-01 -6.51813447e-01 7.57421196e-01
-1.70501041e+00 -9.46605563e-01 -2.61293858e-01 1.77073896e+00
1.02227342e+00 3.49680007e-01 -2.07940280e-01 1.15173481e-01
8.76083553e-01 3.89841497e-01 -2.37705231e-01 2.47509375e-01
-4.36644346e-01 3.06937724e-01 1.75666600e-01 1.44821078e-01
-9.93647993e-01 8.99931252e-01 4.89882755e+00 1.03603971e+00
-5.72383821e-01 2.13215426e-01 1.39606416e-01 1.09616369e-01
-7.07547426e-01 5.62282205e-01 -9.97179151e-01 6.60578668e-01
7.94077694e-01 -7.03141838e-02 9.33638811e-02 4.86904979e-01
1.46912917e-01 -2.03372166e-01 -1.14139736e+00 4.99653429e-01
-3.35435778e-01 -1.18719912e+00 2.27240592e-01 2.20174603e-02
2.64014095e-01 -2.50486881e-01 -5.18621743e-01 5.02788901e-01
4.39868242e-01 -4.25834745e-01 1.00334108e+00 7.30518222e-01
2.34920364e-02 -5.77333510e-01 8.06899667e-01 2.98660934e-01
-1.27843285e+00 2.12549165e-01 -7.49078318e-02 1.23695537e-01
7.46471107e-01 1.18079495e+00 -5.64167857e-01 1.21494615e+00
5.74934423e-01 6.71539843e-01 -2.55578607e-01 7.84725904e-01
-7.36815035e-01 1.08945310e+00 -6.06274843e-01 -7.17693046e-02
-6.75326884e-02 -1.37234524e-01 7.61453748e-01 1.36624038e+00
2.26627648e-01 5.19756734e-01 -8.47370252e-02 1.15686250e+00
-8.22860822e-02 -6.13348894e-02 -2.86261171e-01 -2.93953747e-01
7.58503318e-01 1.29845870e+00 -9.43492413e-01 -2.82839268e-01
-3.88250858e-01 5.14655948e-01 3.62648934e-01 2.87954807e-01
-1.19986343e+00 -3.69654685e-01 2.53155649e-01 -3.34081918e-01
1.94074467e-01 -2.39715546e-01 -1.54858306e-01 -1.44612491e+00
1.22825228e-01 -3.33323210e-01 6.52939141e-01 -7.30247915e-01
-1.42341363e+00 3.28170687e-01 5.91181695e-01 -9.60842013e-01
-4.63913120e-02 -2.33055294e-01 -7.98310041e-01 5.84182799e-01
-1.50940704e+00 -1.15503001e+00 7.70667270e-02 4.24710751e-01
1.52602866e-01 4.78468597e-01 5.69051206e-01 4.43818450e-01
-7.77016878e-01 2.20292434e-01 -4.08405036e-01 2.28790909e-01
6.31356478e-01 -1.09640360e+00 -1.87142983e-01 9.78333473e-01
2.79676497e-01 8.42285037e-01 7.09798276e-01 -9.13925886e-01
-1.15977228e+00 -1.05848956e+00 1.18846273e+00 -4.31519300e-01
7.64149547e-01 -3.94654006e-01 -9.70046997e-01 9.15574729e-01
2.40697250e-01 -3.09807807e-01 7.74975061e-01 4.42487776e-01
-3.92820448e-01 7.24086538e-03 -8.04246306e-01 6.28947914e-01
1.13972914e+00 -6.20712280e-01 -1.28654826e+00 2.23133966e-01
7.73590326e-01 -2.09667161e-01 -8.27147484e-01 6.88665330e-01
2.50244498e-01 -6.10309243e-01 1.05381715e+00 -5.46901286e-01
4.27267015e-01 -5.02371132e-01 -2.07178414e-01 -9.18261170e-01
-6.97502941e-02 -2.57724911e-01 -2.35631660e-01 2.00603175e+00
3.11080366e-01 -4.00292128e-01 8.01311731e-02 4.87259328e-01
-1.25957549e-01 -2.88620740e-01 -8.66487801e-01 -4.27183837e-01
-2.90758073e-01 -9.14318383e-01 9.43046927e-01 9.64655042e-01
2.18103379e-01 5.14517188e-01 -2.46397197e-01 7.35117316e-01
4.98269796e-01 4.62324560e-01 1.36783928e-01 -1.11730516e+00
-2.36441657e-01 -2.27753639e-01 1.53973669e-01 -6.73321247e-01
3.07778716e-01 -1.00203133e+00 -5.77234812e-02 -1.32410574e+00
4.52976912e-01 -6.12744331e-01 -5.67013144e-01 5.27924120e-01
-4.79657620e-01 -4.32027757e-01 -7.91580379e-02 1.58865437e-01
-8.35813284e-01 8.91425729e-01 5.46231508e-01 1.21536173e-01
6.65471330e-03 -2.47263238e-01 -5.93248665e-01 9.35797334e-01
6.57680333e-01 -9.27591622e-01 -1.71643138e-01 -2.41390839e-01
6.22771263e-01 2.76899010e-01 6.87015891e-01 -7.41938055e-01
2.26234943e-01 -4.28195298e-01 2.27143079e-01 -9.08699989e-01
6.59392029e-02 -6.72540128e-01 3.26611549e-01 -7.36373067e-02
-3.75665963e-01 -2.16034070e-01 1.50683165e-01 7.18451738e-01
-3.50577652e-01 -3.91382039e-01 4.10972461e-02 8.52911472e-02
-4.08666372e-01 1.54427085e-02 -4.59325731e-01 2.51407653e-01
7.83976674e-01 5.47400117e-01 -2.96615452e-01 1.53057069e-01
-7.61981249e-01 1.48727491e-01 -1.29722193e-01 2.53429353e-01
3.15575868e-01 -1.27466774e+00 -5.16934752e-01 -1.67470813e-01
1.61689982e-01 1.74336329e-01 4.68805313e-01 8.96412849e-01
2.02033743e-01 2.44723767e-01 1.58178031e-01 -2.70167470e-01
-6.44957125e-01 6.44441485e-01 -1.26101851e-01 -6.10164464e-01
-4.70557749e-01 4.86790448e-01 3.76683384e-01 -3.01254749e-01
-1.71473891e-01 -5.69710970e-01 -4.01861101e-01 2.25956202e-01
2.80613840e-01 2.18061820e-01 -1.15539223e-01 -3.12801093e-01
-5.21378219e-01 9.04362053e-02 -1.61110032e-02 -3.85108829e-01
1.27863979e+00 -1.05450362e-01 -2.53099293e-01 3.07053298e-01
6.08777821e-01 5.36967039e-01 -1.00681007e+00 -4.28047597e-01
3.91885579e-01 6.86173514e-02 -1.65776275e-02 -8.27682197e-01
-5.38957953e-01 7.83774376e-01 -5.83730452e-02 7.08909184e-02
1.04177880e+00 4.83762592e-01 8.24412823e-01 7.69360317e-03
5.28411031e-01 -6.22989774e-01 -4.42700326e-01 3.92843813e-01
5.31228304e-01 -6.95282817e-01 -2.00790271e-01 -8.44437361e-01
-5.25585055e-01 1.03416765e+00 4.24913675e-01 1.94824547e-01
4.91114318e-01 3.24650377e-01 -6.08303070e-01 -2.50574261e-01
-5.02718210e-01 -1.84798971e-01 1.42169759e-01 2.57018171e-02
1.83199421e-01 1.72579095e-01 -7.58664727e-01 1.56421888e+00
4.72618314e-03 7.68112577e-03 1.85751095e-01 9.26995277e-01
-4.97464798e-02 -1.18910623e+00 -2.90602088e-01 5.97116426e-02
-5.52646220e-01 -3.81205291e-01 -6.10399023e-02 5.70013165e-01
5.28082609e-01 8.94151986e-01 -7.44029135e-02 -1.74283311e-01
9.97191966e-02 4.05311078e-01 3.90526980e-01 -3.34283501e-01
-5.61273515e-01 1.01488322e-01 2.77030259e-01 -2.39394724e-01
-7.04381227e-01 -8.58280182e-01 -1.66820431e+00 2.13648587e-01
-2.66144603e-01 5.20758927e-01 3.53729546e-01 1.36092401e+00
5.41743696e-01 7.94919074e-01 4.33589190e-01 -3.39234084e-01
-1.06798634e-01 -1.06227338e+00 -2.69775212e-01 4.16629463e-01
-1.06832437e-01 -1.05473304e+00 -2.92103916e-01 -7.72145763e-02] | [9.046899795532227, 9.139142036437988] |
74dfd0a9-31a0-4a6f-bccb-269d27fa9449 | clutrr-a-diagnostic-benchmark-for-inductive | 1908.06177 | null | https://arxiv.org/abs/1908.06177v2 | https://arxiv.org/pdf/1908.06177v2.pdf | CLUTRR: A Diagnostic Benchmark for Inductive Reasoning from Text | The recent success of natural language understanding (NLU) systems has been troubled by results highlighting the failure of these models to generalize in a systematic and robust way. In this work, we introduce a diagnostic benchmark suite, named CLUTRR, to clarify some key issues related to the robustness and systematicity of NLU systems. Motivated by classic work on inductive logic programming, CLUTRR requires that an NLU system infer kinship relations between characters in short stories. Successful performance on this task requires both extracting relationships between entities, as well as inferring the logical rules governing these relationships. CLUTRR allows us to precisely measure a model's ability for systematic generalization by evaluating on held-out combinations of logical rules, and it allows us to evaluate a model's robustness by adding curated noise facts. Our empirical results highlight a substantial performance gap between state-of-the-art NLU models (e.g., BERT and MAC) and a graph neural network model that works directly with symbolic inputs---with the graph-based model exhibiting both stronger generalization and greater robustness. | ['William L. Hamilton', 'Shagun Sodhani', 'Koustuv Sinha', 'Jin Dong', 'Joelle Pineau'] | 2019-08-16 | clutrr-a-diagnostic-benchmark-for-inductive-1 | https://aclanthology.org/D19-1458 | https://aclanthology.org/D19-1458.pdf | ijcnlp-2019-11 | ['systematic-generalization'] | ['reasoning'] | [ 2.14870751e-01 6.59069300e-01 -3.80386889e-01 -3.70204687e-01
-2.19059512e-01 -7.10237801e-01 5.95545709e-01 6.05905414e-01
1.66066185e-01 7.16489196e-01 2.81817526e-01 -8.41947675e-01
-4.98430371e-01 -1.38555360e+00 -1.04597366e+00 1.68253854e-01
-4.21861917e-01 6.78440392e-01 2.20197037e-01 -4.51616079e-01
-1.16884997e-02 3.06666285e-01 -1.44647765e+00 2.77972639e-01
1.00926006e+00 7.76376665e-01 -4.54848170e-01 7.16206610e-01
-3.69248204e-02 1.53604650e+00 -4.62250113e-01 -6.01213932e-01
-8.51666555e-02 -5.08242786e-01 -1.09149027e+00 -4.53962028e-01
4.79461402e-01 -3.52390528e-01 -6.68999612e-01 8.01292777e-01
1.77376419e-01 1.31652266e-01 9.77016211e-01 -1.39576757e+00
-1.20276296e+00 1.51058626e+00 -1.08240478e-01 2.41653562e-01
8.74188542e-01 2.12271377e-01 1.65551782e+00 -2.94648647e-01
1.00262010e+00 1.51588917e+00 8.31142724e-01 5.64313948e-01
-1.33783674e+00 -4.37018305e-01 2.33738378e-01 2.07588896e-01
-1.42662156e+00 -2.97760963e-01 4.11067247e-01 -3.82622659e-01
1.43740237e+00 2.86278784e-01 2.51915604e-01 1.04115689e+00
6.13941066e-03 8.35877419e-01 6.49353743e-01 -5.78722179e-01
2.35765457e-01 -2.00831354e-01 7.52999961e-01 9.98840034e-01
5.17024279e-01 2.56159548e-02 -4.95207638e-01 -4.49248403e-02
5.58407128e-01 -6.22506857e-01 -2.99397409e-01 -1.94210541e-02
-1.13933706e+00 6.91717744e-01 5.26320517e-01 3.33531260e-01
1.19868770e-01 3.05364519e-01 4.73146468e-01 3.53360027e-01
3.12946409e-01 1.11502504e+00 -4.06133622e-01 -4.25532088e-02
-5.48927248e-01 4.35206950e-01 1.07326543e+00 1.28249323e+00
4.58984792e-01 -1.03836544e-01 -1.73476174e-01 6.91541851e-01
1.00777000e-01 9.15148631e-02 1.91384599e-01 -6.75248206e-01
4.59874094e-01 9.46881533e-01 -4.59477514e-01 -1.26886129e+00
-6.13067091e-01 -3.37077618e-01 -6.83951497e-01 -1.30978316e-01
4.94915307e-01 -3.32869798e-01 -7.51828551e-01 2.03129911e+00
-1.81701109e-01 2.30306298e-01 4.58881438e-01 3.81929547e-01
1.16644561e+00 5.29056728e-01 5.69108240e-02 2.26920396e-01
9.31552708e-01 -5.07062554e-01 -4.01485920e-01 -5.57275116e-01
1.08944261e+00 1.45202324e-01 1.06653118e+00 2.32447147e-01
-1.01456916e+00 -2.59018838e-01 -1.37769306e+00 -2.54086733e-01
-5.89222133e-01 -2.52022445e-01 1.01666796e+00 4.21628922e-01
-1.04816639e+00 7.79729486e-01 -5.10536373e-01 -5.15702069e-01
4.13844854e-01 1.80229217e-01 -4.90271807e-01 -2.49124542e-01
-1.70892310e+00 1.13423491e+00 1.05687737e+00 -1.72641799e-01
-5.53215325e-01 -8.97379816e-01 -1.19315708e+00 3.96829545e-01
8.87846947e-01 -7.78538942e-01 1.09819090e+00 -3.51175815e-01
-8.57342005e-01 6.01180911e-01 1.66624427e-01 -8.50490808e-01
5.53970754e-01 -7.74637088e-02 -4.61927593e-01 -1.15050025e-01
2.05961205e-02 5.69116592e-01 -4.80436198e-02 -1.25598240e+00
-3.72727334e-01 -1.96696520e-01 5.25608480e-01 -7.65992552e-02
-7.34191760e-02 -2.43116856e-01 -3.24286252e-01 -6.03135526e-01
1.17231049e-01 -6.08922720e-01 1.59175664e-01 -3.87585968e-01
-8.90374422e-01 -5.48858762e-01 5.35440445e-01 -5.23312569e-01
1.72133660e+00 -1.63047707e+00 1.38113841e-01 3.82063985e-01
3.54877174e-01 -3.57959569e-02 -2.48362701e-02 5.07585585e-01
-2.60117412e-01 5.77868164e-01 -2.54719436e-01 1.29466370e-01
2.60352254e-01 4.96970445e-01 -4.35353607e-01 -3.40278223e-02
4.96430695e-01 1.28480256e+00 -1.19212806e+00 -4.48053211e-01
-2.10855350e-01 -8.79618302e-02 -6.05684459e-01 -1.63293734e-01
-7.98812568e-01 -4.16040331e-01 -3.30988437e-01 7.95806468e-01
1.22663707e-01 -3.72132421e-01 4.16119367e-01 1.67041942e-02
4.69643772e-01 5.22388816e-01 -1.01317942e+00 1.28936338e+00
-1.57103017e-01 9.31010962e-01 -7.37887263e-01 -8.60241950e-01
1.08394921e+00 2.45950166e-02 -2.28584528e-01 -4.53757495e-01
1.24290116e-01 -1.84866221e-04 3.06561798e-01 -6.99336827e-01
5.62478781e-01 -2.88925469e-02 -2.19692931e-01 7.17518806e-01
1.69578448e-01 -1.79894522e-01 6.25854790e-01 7.39426494e-01
1.39742219e+00 1.52188554e-01 4.68141764e-01 -3.19452703e-01
1.50400043e-01 2.63596654e-01 3.88536274e-01 1.21927428e+00
3.03793430e-01 3.17733020e-01 1.08439136e+00 -2.84330994e-01
-8.19717526e-01 -1.05761349e+00 -1.56897336e-01 9.33886290e-01
-1.26167104e-01 -6.32517040e-01 -6.31965101e-01 -8.24333251e-01
3.88407975e-01 1.58139694e+00 -1.03717887e+00 -5.71047604e-01
-3.50935072e-01 -7.68015206e-01 1.23071623e+00 9.23660398e-01
3.16128701e-01 -9.67128932e-01 -3.48345965e-01 9.57399979e-02
-1.53398871e-01 -1.13316381e+00 2.30457634e-01 1.05677664e-01
-6.76019490e-01 -1.25508380e+00 2.67058939e-01 -4.47918385e-01
6.33445621e-01 -2.58387178e-01 1.64114797e+00 4.26334411e-01
-1.57635108e-01 3.10559601e-01 -5.50799489e-01 -4.95745301e-01
-1.08504653e+00 2.56081522e-01 1.81004539e-01 -6.47732317e-01
4.81348097e-01 -5.71838677e-01 2.19693944e-01 2.86522180e-01
-1.17397165e+00 1.55663356e-01 3.31966192e-01 6.72845364e-01
8.80163312e-02 2.02003211e-01 8.48991752e-01 -1.21053684e+00
1.09102833e+00 -6.93631947e-01 -2.58949697e-01 8.40561807e-01
-6.46784902e-01 2.77446538e-01 6.81800902e-01 -2.17717409e-01
-7.23978937e-01 -7.22635031e-01 2.19740137e-01 -1.37305921e-02
-2.36891722e-03 1.21020675e+00 -2.42933318e-01 1.60714358e-01
1.26516116e+00 9.75776557e-03 -2.98317343e-01 -3.97340283e-02
5.98555148e-01 4.23703164e-01 8.68623257e-01 -1.11393774e+00
6.53216898e-01 -2.90656518e-02 7.40558580e-02 -6.20219052e-01
-9.74540651e-01 1.45866945e-01 -6.84468269e-01 -2.48322025e-01
4.57278550e-01 -5.37165999e-01 -6.15732431e-01 1.77476421e-01
-1.02855313e+00 -6.56122983e-01 -9.64757577e-02 9.08231512e-02
-4.40115720e-01 2.78717428e-01 -6.34679735e-01 -5.26964486e-01
-1.16230242e-01 -6.99863851e-01 4.30176198e-01 1.51207233e-02
-7.82313943e-01 -1.31140196e+00 -1.50469383e-02 1.41942933e-01
3.61504070e-02 6.04052365e-01 1.65432000e+00 -1.03763413e+00
-3.98308635e-01 -2.83832103e-01 -3.49876553e-01 1.46171153e-01
-6.12932593e-02 4.38397706e-01 -7.87258267e-01 1.73693031e-01
-6.87377393e-01 -6.43111646e-01 6.39997542e-01 1.69711158e-01
9.09940660e-01 -3.59316289e-01 -5.03763378e-01 4.71855223e-01
1.20899618e+00 1.29326686e-01 7.25231349e-01 2.63866484e-01
6.64214253e-01 5.43608189e-01 3.28021079e-01 7.57278204e-02
5.38736343e-01 1.71172172e-01 1.45680726e-01 3.45472872e-01
-2.42214631e-02 -6.32590771e-01 1.16436975e-02 2.93339163e-01
-1.81680322e-01 -5.56557775e-01 -1.51518345e+00 6.05593622e-01
-2.15176702e+00 -7.94327319e-01 3.80033925e-02 1.83637047e+00
9.68944609e-01 6.18668318e-01 -2.31535450e-01 1.65706813e-01
2.92308778e-01 -1.56371146e-02 -6.66850328e-01 -5.65894783e-01
-4.53523099e-01 -9.98359025e-02 3.19527507e-01 5.91419041e-01
-7.33269215e-01 1.14701557e+00 7.36675787e+00 7.25284696e-01
-6.12052202e-01 -5.38862586e-01 7.57697940e-01 1.52292755e-02
-6.05056584e-01 -7.50742704e-02 -6.61475122e-01 -6.53486187e-03
7.56473958e-01 -5.26965559e-01 5.19053221e-01 7.20134377e-01
-2.37053841e-01 -2.64244020e-01 -1.88337648e+00 3.02929014e-01
3.28498900e-01 -1.76902890e+00 5.01629889e-01 -3.38923544e-01
7.21743345e-01 -1.36655897e-01 -3.86116713e-01 5.69934964e-01
9.82717156e-01 -1.46746290e+00 7.37380803e-01 6.16399050e-01
6.51947320e-01 -7.08216548e-01 6.33300006e-01 3.86666715e-01
-8.64927411e-01 -5.54646254e-02 -3.59418482e-01 -4.20761049e-01
-1.41985595e-01 4.13882285e-01 -1.08388531e+00 7.88392961e-01
5.01956642e-01 7.95953155e-01 -9.24033582e-01 3.94207805e-01
-7.63981462e-01 6.52025223e-01 -3.14044029e-01 -3.79076421e-01
2.36070454e-01 1.04924314e-01 4.78057444e-01 1.38704562e+00
-3.53524722e-02 3.40367436e-01 -1.36198506e-01 1.29109156e+00
-3.33570868e-01 -2.16164052e-01 -1.02269971e+00 -3.59158069e-01
6.12560153e-01 7.00392008e-01 -5.99717855e-01 -4.87800807e-01
-1.95439532e-01 4.09608662e-01 7.49218404e-01 4.18712080e-01
-7.45341003e-01 -3.68171096e-01 3.50492895e-01 -6.06064126e-02
-1.61713451e-01 -2.58089542e-01 -6.08667612e-01 -1.11925018e+00
-1.21874832e-01 -8.71928394e-01 7.22361922e-01 -1.18535364e+00
-1.30748188e+00 4.53950316e-01 4.20019060e-01 -6.05730772e-01
-6.05972826e-01 -7.70142853e-01 -6.12524331e-01 6.63565278e-01
-1.10187089e+00 -1.19836783e+00 -5.55374287e-03 3.36343288e-01
4.82197851e-02 -8.86536613e-02 8.48721087e-01 -2.33676612e-01
-7.71205723e-01 7.46380806e-01 -4.02322710e-01 5.99049211e-01
3.88315231e-01 -1.41963816e+00 8.27799559e-01 9.83850360e-01
9.64260772e-02 1.01076496e+00 8.58540177e-01 -8.96663070e-01
-1.30590677e+00 -9.73760724e-01 9.51315105e-01 -8.70815039e-01
1.14759886e+00 -3.01651835e-01 -1.07172894e+00 1.35207200e+00
-9.04602185e-02 -1.24501787e-01 6.07001483e-01 6.14645720e-01
-8.35321665e-01 2.08972454e-01 -9.23418224e-01 1.04656148e+00
1.48973620e+00 -5.91159046e-01 -9.88164186e-01 2.37741470e-01
9.08985376e-01 -4.35009927e-01 -9.69659090e-01 4.82192248e-01
4.97301012e-01 -9.35843945e-01 8.56634140e-01 -1.27738154e+00
1.17547631e+00 -1.75759807e-01 -1.18933342e-01 -1.38880765e+00
-4.02867407e-01 -3.55389625e-01 -3.37684363e-01 1.21985161e+00
1.07744050e+00 -6.10243917e-01 4.35779124e-01 9.96796370e-01
1.27780568e-02 -8.98563921e-01 -4.68912154e-01 -8.15553427e-01
1.41280562e-01 -8.94988537e-01 6.09144747e-01 1.30874801e+00
7.61404872e-01 5.96930146e-01 3.19202952e-02 2.95994401e-01
3.17945957e-01 4.01227623e-02 6.62992239e-01 -1.25948954e+00
-4.12148148e-01 -7.03861415e-01 -5.53660035e-01 -5.36869884e-01
4.87894267e-01 -1.13595998e+00 -1.65702894e-01 -1.95328081e+00
1.07441530e-01 -4.24890816e-01 3.84136438e-02 8.05101454e-01
-3.31410617e-01 -2.51833618e-01 2.07866188e-02 -1.73638552e-01
-4.16665912e-01 1.26297787e-01 8.56388390e-01 -4.26105946e-01
-2.59052038e-01 -4.40155417e-01 -1.04341400e+00 8.07046592e-01
7.01711416e-01 -6.25792071e-02 -7.85349548e-01 -6.24691367e-01
9.90493894e-01 -8.41060504e-02 3.88537467e-01 -9.56364870e-01
3.42482418e-01 -1.27931252e-01 1.62382007e-01 -2.26396337e-01
-1.64393798e-01 -3.40643317e-01 -1.39242604e-01 3.27746749e-01
-6.93022728e-01 1.47334903e-01 4.05899733e-01 4.28003252e-01
-7.48215243e-02 -3.26089114e-01 3.12521130e-01 -1.18654065e-01
-9.39069033e-01 2.24235366e-04 -1.30195975e-01 4.03175294e-01
9.28507328e-01 -1.31074890e-01 -8.56469929e-01 -5.73892593e-01
-6.11307502e-01 5.42610049e-01 3.41288894e-01 5.81574440e-01
5.58110058e-01 -9.72474098e-01 -8.61805618e-01 -1.71442285e-01
4.41884130e-01 1.77033022e-01 3.38227907e-03 4.94022131e-01
-7.23005533e-01 3.53630543e-01 -6.52887300e-03 -2.00374857e-01
-9.83894944e-01 6.68064058e-01 5.27854443e-01 -2.68066674e-01
-5.40073633e-01 1.08877993e+00 -9.28147137e-02 -6.03519499e-01
1.86334074e-01 -8.19220066e-01 -2.54501253e-01 -2.69346647e-02
4.60024238e-01 3.05582017e-01 -8.48123208e-02 -1.41432941e-01
-4.12635982e-01 1.50696158e-01 -8.38332474e-02 1.34084180e-01
1.23639929e+00 2.94876009e-01 -2.94632792e-01 5.00347316e-01
7.40922093e-01 -1.18177176e-01 -7.09858239e-01 -4.13612098e-01
2.11088315e-01 6.42546872e-03 -1.05212256e-01 -1.14037538e+00
-5.60217559e-01 5.30860364e-01 -5.53888381e-01 5.81579924e-01
6.52210176e-01 2.15442389e-01 3.21241736e-01 9.00465190e-01
3.29417437e-01 -8.91573012e-01 -3.02958578e-01 8.55006337e-01
1.00848067e+00 -1.12337601e+00 2.53002226e-01 -5.72926581e-01
-5.49041808e-01 1.19965804e+00 7.41794407e-01 -2.71813814e-02
3.54209989e-01 3.98001790e-01 -2.69783497e-01 -2.92847723e-01
-9.59415317e-01 -9.03331712e-02 3.55890155e-01 5.61042786e-01
3.12004775e-01 2.01640382e-01 7.84579068e-02 6.54932261e-01
-5.64814925e-01 1.37245148e-01 6.80145621e-01 7.31336415e-01
-2.28861257e-01 -6.38756156e-01 -2.14228719e-01 7.00814784e-01
6.83142543e-02 -3.51803988e-01 -9.30970013e-01 1.20136940e+00
-6.43850043e-02 9.71679568e-01 -1.88583229e-02 -6.07966781e-01
3.30585569e-01 2.28870854e-01 5.32662749e-01 -6.07289851e-01
-3.66774172e-01 -8.66756141e-01 8.36804211e-01 -3.22647333e-01
-8.72791857e-02 -5.60861588e-01 -1.30775177e+00 -5.68759382e-01
-6.75412118e-02 -3.29532437e-02 3.21345590e-02 1.12359226e+00
1.31575182e-01 8.76692355e-01 -6.48658276e-02 4.18294184e-02
-2.88322091e-01 -1.01982713e+00 -4.23189908e-01 3.52763891e-01
1.81195547e-03 -5.03701091e-01 -2.21527889e-01 -1.34110600e-01] | [9.237570762634277, 7.54093074798584] |
d78d3b25-f311-42d9-b235-c444cb0be1de | pandas-abusive-comment-detection-in-tamil | null | null | https://aclanthology.org/2022.dravidianlangtech-1.18 | https://aclanthology.org/2022.dravidianlangtech-1.18.pdf | PANDAS@Abusive Comment Detection in Tamil Code-Mixed Data Using Custom Embeddings with LaBSE | Abusive language has lately been prevalent in comments on various social media platforms. The increasing hostility observed on the internet calls for the creation of a system that can identify and flag such acerbic content, to prevent conflict and mental distress. This task becomes more challenging when low-resource languages like Tamil, as well as the often-observed Tamil-English code-mixed text, are involved. The approach used in this paper for the classification model includes different methods of feature extraction and the use of traditional classifiers. We propose a novel method of combining language-agnostic sentence embeddings with the TF-IDF vector representation that uses a curated corpus of words as vocabulary, to create a custom embedding, which is then passed to an SVM classifier. Our experimentation yielded an accuracy of 52% and an F1-score of 0.54. | ['Bharathi B', 'Thenmozhi Durairaj', 'Gayathri G L', 'Divyasri K', 'Krithika Swaminathan'] | null | null | null | null | dravidianlangtech-acl-2022-5 | ['abusive-language'] | ['natural-language-processing'] | [-2.12071508e-01 -6.38776496e-02 -6.18002675e-02 -2.64858425e-01
-3.22540671e-01 -4.27187979e-01 7.27241457e-01 7.29825258e-01
-6.03454232e-01 7.42211759e-01 3.60328645e-01 -4.99289989e-01
6.95628077e-02 -5.16168594e-01 2.37039804e-01 -4.19256836e-01
-4.98221666e-02 8.15174654e-02 3.21863964e-02 -5.81153274e-01
6.87286913e-01 4.68721032e-01 -1.32852006e+00 2.48011246e-01
9.38560784e-01 7.22767115e-01 -1.27984479e-01 6.16964161e-01
-4.56512570e-01 1.01020610e+00 -4.94957715e-01 -7.88187623e-01
-3.93202938e-02 -2.80317754e-01 -9.30265069e-01 -1.08215876e-01
-3.33027616e-02 1.83839183e-02 -3.85700166e-02 1.11387682e+00
2.57284403e-01 -4.46902530e-04 7.11074889e-01 -1.07390487e+00
-4.14312959e-01 4.63474065e-01 -5.24637401e-01 5.48354924e-01
6.11987889e-01 -2.21311972e-01 7.89053321e-01 -9.83816803e-01
5.95725358e-01 1.11109316e+00 7.17485487e-01 2.78376043e-01
-1.07654321e+00 -7.34621525e-01 -2.84736812e-01 9.80008990e-02
-1.32801974e+00 -2.67416000e-01 7.57797480e-01 -9.24372673e-01
9.32646096e-01 8.48941803e-02 7.67184198e-01 1.28054976e+00
4.39570844e-01 9.44493562e-02 1.19906318e+00 -6.45174325e-01
-4.38475423e-02 8.34523976e-01 1.93082571e-01 6.74911082e-01
4.12328154e-01 -5.83440006e-01 -4.21889544e-01 -6.66167319e-01
-4.69126105e-02 1.28795862e-01 3.75365429e-02 1.69728279e-01
-5.49194574e-01 1.45849526e+00 -9.43915769e-02 9.83244717e-01
-3.70846838e-01 -2.73039371e-01 8.28173041e-01 3.64741206e-01
1.02955210e+00 5.61679840e-01 -2.74602681e-01 -4.97551173e-01
-8.41716409e-01 1.56009480e-01 9.72834170e-01 3.10970217e-01
5.27169883e-01 -2.51297027e-01 3.79561663e-01 1.22505546e+00
5.16571999e-01 1.87440887e-01 9.30417299e-01 -1.71973497e-01
4.35610294e-01 1.01214683e+00 -2.02473685e-01 -1.57513142e+00
-3.11654598e-01 -3.20277214e-01 -5.39281309e-01 -8.08295310e-02
2.00253963e-01 -3.59894693e-01 -3.97554398e-01 1.33875477e+00
3.99567842e-01 -2.36684769e-01 -2.90058441e-02 4.97166097e-01
2.52100229e-01 6.13773286e-01 8.03500190e-02 -3.27160001e-01
1.22443140e+00 -5.62434137e-01 -8.44177842e-01 -1.84738964e-01
9.80649471e-01 -1.07416844e+00 9.25646484e-01 3.32754970e-01
-6.40428126e-01 -1.00166211e-02 -1.05792737e+00 4.90529053e-02
-8.44118237e-01 -2.09779426e-01 4.29109246e-01 1.07891798e+00
-7.60662615e-01 5.67083299e-01 -4.25617009e-01 -7.49525666e-01
3.82742912e-01 2.14134902e-01 -6.93314373e-01 1.84545591e-01
-1.05716407e+00 1.24686706e+00 3.49650472e-01 -2.37290978e-01
-1.27785817e-01 -2.72170097e-01 -8.22442353e-01 -2.25727811e-01
7.39821121e-02 7.77656585e-02 7.78542399e-01 -1.49120736e+00
-1.23517883e+00 9.78003502e-01 1.34859100e-01 -1.80607840e-01
4.67036247e-01 -3.31675529e-01 -7.85007536e-01 1.73800319e-01
2.25456730e-01 -2.23254785e-01 9.71561670e-01 -7.93061793e-01
-4.09692168e-01 -4.13918436e-01 6.05300106e-02 -2.17841834e-01
-1.13985002e+00 6.88161850e-01 1.91804826e-01 -5.15223861e-01
-8.24365765e-02 -8.94132495e-01 -9.90130007e-03 -4.13862199e-01
-2.73470074e-01 -3.31419915e-01 8.52336228e-01 -9.58853900e-01
1.69969010e+00 -2.19385052e+00 -3.29149626e-02 3.84653866e-01
3.52082044e-01 5.84955752e-01 2.20758170e-01 8.77577424e-01
-2.47363389e-01 5.23876548e-01 -1.21921629e-01 -3.16522866e-01
-2.60698527e-01 9.69492365e-03 -7.64417797e-02 6.08804166e-01
4.61489826e-01 1.08235180e-01 -9.43652391e-01 -6.46587908e-01
-8.10538307e-02 2.97634155e-01 -6.89854026e-01 2.94307411e-01
3.84153724e-01 -5.81790926e-04 -4.92445678e-01 5.03166914e-01
4.66730237e-01 4.34241779e-02 2.93206066e-01 2.39514396e-01
-4.94514376e-01 3.19058955e-01 -8.59860718e-01 1.07269561e+00
-5.88533401e-01 9.33730304e-01 -1.32660717e-01 -8.22139561e-01
1.09945202e+00 3.03949237e-01 4.28899109e-01 -2.70901203e-01
7.00848699e-01 4.91271257e-01 1.85424298e-01 -9.88944888e-01
5.53058624e-01 -2.72940189e-01 -2.20936358e-01 3.57075959e-01
9.68141928e-02 6.87632114e-02 2.02499658e-01 2.98741281e-01
1.07600212e+00 -3.09618771e-01 6.09849453e-01 -4.44025040e-01
9.28300619e-01 -7.09782317e-02 1.78571150e-01 1.29787683e-01
-3.93755794e-01 1.46082744e-01 8.18720818e-01 -5.32801747e-01
-1.04620993e+00 -3.36917996e-01 -3.02436113e-01 9.02377784e-01
-4.37840372e-01 -6.64780259e-01 -7.18659818e-01 -7.74198651e-01
6.92950860e-02 5.56436002e-01 -6.46725774e-01 -2.27169096e-01
-2.51145691e-01 -7.16192186e-01 5.60946882e-01 -1.32293344e-01
1.87767252e-01 -6.55937552e-01 -4.65741932e-01 3.89285535e-01
2.71211918e-02 -9.26845133e-01 -4.81649190e-02 9.76390094e-02
-6.40354693e-01 -1.00034046e+00 -2.26546586e-01 -4.95993733e-01
4.88924056e-01 -1.06898755e-01 5.69534242e-01 3.18026721e-01
-3.39767337e-01 3.88676897e-02 -7.73934186e-01 -5.27567804e-01
-5.86606503e-01 2.59025753e-01 3.48243326e-01 4.67722625e-01
7.29090512e-01 -4.34871346e-01 -1.00710303e-01 -3.91917974e-01
-9.29090917e-01 -3.16414475e-01 3.05880874e-01 8.58465850e-01
-4.08258826e-01 -3.59400719e-01 8.32941473e-01 -1.11352420e+00
1.09041762e+00 -1.02997553e+00 -1.29607335e-01 -1.53186500e-01
-7.14812636e-01 -2.22504944e-01 7.22656548e-01 -5.77684700e-01
-7.91477203e-01 -4.16697592e-01 -5.20749927e-01 8.53569508e-02
-7.13926479e-02 7.27797210e-01 3.50958318e-01 -6.43449426e-02
7.17494607e-01 -2.03245819e-01 2.18160599e-01 -5.08465707e-01
-1.53756201e-01 1.41653419e+00 -2.94627607e-01 1.03772618e-01
9.36102211e-01 5.85449412e-02 -3.84157002e-01 -1.26423907e+00
-7.36932993e-01 -7.04091787e-01 -5.79432905e-01 -3.00198138e-01
9.63336945e-01 -5.23063600e-01 -4.45746809e-01 2.62364835e-01
-1.35512316e+00 1.83206618e-01 3.31015557e-01 6.12930894e-01
-3.35840061e-02 4.39543307e-01 -4.41972494e-01 -1.15093839e+00
-2.74775952e-01 -7.51598537e-01 2.38880947e-01 3.55569199e-02
-6.85456932e-01 -1.12009096e+00 4.60230380e-01 4.77499366e-01
5.89226305e-01 2.75123149e-01 1.00172698e+00 -1.27965009e+00
3.70400965e-01 -5.65026760e-01 -2.58241951e-01 6.92902088e-01
1.91113725e-01 3.24103326e-01 -1.01395547e+00 -1.48142010e-01
2.34762460e-01 -4.62308198e-01 3.59872669e-01 -4.32900846e-01
8.62927556e-01 -4.48631406e-01 -1.31641030e-01 -1.66235249e-02
1.62397015e+00 2.76466936e-01 4.28045154e-01 4.66675818e-01
5.46129644e-01 5.69258571e-01 5.36544442e-01 8.18783879e-01
9.39353183e-02 4.29404557e-01 3.09346735e-01 2.78517485e-01
5.15983880e-01 -9.23826545e-02 5.51599026e-01 1.10675418e+00
7.66307907e-03 -4.12482433e-02 -1.26484048e+00 6.60847127e-01
-1.53828037e+00 -1.03188205e+00 -2.60040849e-01 1.94558430e+00
8.59359324e-01 1.90970734e-01 1.20906182e-01 5.30997932e-01
4.62357968e-01 1.19229138e-01 2.52111703e-01 -1.29604328e+00
4.68475968e-01 2.30563208e-01 3.04290235e-01 6.58556819e-01
-1.06476355e+00 7.14077175e-01 5.56773758e+00 6.02362573e-01
-1.43956184e+00 5.24174392e-01 4.57975179e-01 2.13437170e-01
-1.03882112e-01 -1.73304722e-01 -5.89046657e-01 6.60434067e-01
1.32992876e+00 -2.04906642e-01 1.27464458e-01 8.55119467e-01
3.31177235e-01 -1.66359216e-01 -5.37339926e-01 9.87161160e-01
6.92717850e-01 -9.40355003e-01 -3.39145213e-01 5.41475862e-02
4.43678349e-01 -1.03962250e-01 -7.87528381e-02 3.03267211e-01
-1.98203087e-01 -9.81326878e-01 4.93327856e-01 2.14271367e-01
4.55008715e-01 -8.70006382e-01 1.11515677e+00 4.62418735e-01
-5.88566780e-01 -3.55348587e-01 -2.37057894e-01 -4.96649474e-01
-1.90319344e-01 7.08455324e-01 -1.06594944e+00 3.12461108e-01
5.37114739e-01 7.74856985e-01 -6.09105825e-01 8.40858579e-01
2.76397094e-02 5.91685474e-01 -2.75237467e-02 -6.51647449e-01
3.36923420e-01 -2.78529316e-01 4.46926892e-01 1.51421726e+00
2.49548525e-01 -2.97645956e-01 -5.76023152e-03 4.29434747e-01
1.11496106e-01 9.01338816e-01 -1.09948254e+00 -5.46907783e-01
1.60911500e-01 1.54791248e+00 -6.58500433e-01 -1.59051880e-01
-6.42752051e-01 9.45197046e-01 6.49372876e-01 -2.26542816e-01
-8.12887132e-01 -6.82127178e-01 6.95882738e-01 2.89193273e-01
-5.95873035e-02 -2.88820863e-01 -2.40568351e-02 -1.28797877e+00
-1.62759066e-01 -9.47716594e-01 1.05047151e-01 -1.35425746e-01
-1.42951059e+00 9.07116771e-01 -2.54994426e-02 -9.75830793e-01
-1.24151222e-01 -6.34966850e-01 -5.72244525e-01 8.09781134e-01
-1.31766152e+00 -8.60492349e-01 -5.32398224e-02 4.24287915e-01
4.24531013e-01 -4.14264739e-01 9.66012239e-01 6.88485384e-01
-8.14807773e-01 4.12352204e-01 4.06047180e-02 2.16883868e-01
7.39097655e-01 -9.84414518e-01 -4.02649850e-01 7.61258423e-01
-2.76412129e-01 7.08596408e-01 9.30748820e-01 -6.05658650e-01
-1.03143358e+00 -8.04677904e-01 1.67561257e+00 -4.16333139e-01
1.32669985e+00 -5.56232274e-01 -7.77194321e-01 4.45572257e-01
3.15675616e-01 -2.33571097e-01 1.18572676e+00 1.15726501e-01
-4.22884911e-01 -4.85714264e-02 -1.40561283e+00 6.14282310e-01
5.33726513e-01 -7.56274462e-01 -8.05519581e-01 4.67355549e-01
3.09964269e-01 3.25696170e-01 -7.65270352e-01 -3.32716972e-01
4.43864763e-01 -1.04749095e+00 3.91569465e-01 -7.57518947e-01
7.04322755e-01 2.85050422e-01 -1.06609061e-01 -1.29286945e+00
-1.72705173e-01 -4.77882594e-01 4.40249175e-01 1.43738008e+00
4.81656373e-01 -9.22429144e-01 3.05796772e-01 4.22426552e-01
1.92492709e-01 -6.18670464e-01 -9.73778605e-01 -4.24888134e-01
-2.12934203e-02 -3.45596194e-01 3.47834527e-02 1.42177796e+00
5.15429080e-01 3.18393111e-01 -3.55118334e-01 -1.36297971e-01
1.11962818e-01 -5.15718699e-01 6.28795922e-01 -1.43078351e+00
5.25880828e-02 -2.97647744e-01 -7.86250532e-01 3.99538726e-02
4.21121776e-01 -9.20009077e-01 -3.82591724e-01 -9.98319507e-01
1.15109324e-01 -3.93928617e-01 -4.79568318e-02 2.99686193e-01
1.32286310e-01 4.13950741e-01 1.30469024e-01 -9.64248851e-02
-1.25764832e-01 3.32970887e-01 6.00012541e-01 4.80625182e-02
-1.37091279e-01 -3.55062932e-01 -7.74821401e-01 9.87789333e-01
9.22099113e-01 -8.70111525e-01 -1.50150344e-01 1.75800040e-01
5.68831205e-01 -3.99093658e-01 -1.93165392e-02 -1.00940132e+00
-2.62133121e-01 -3.70500833e-01 6.95938990e-02 -3.38933915e-02
3.33426803e-01 -1.12693465e+00 -1.92270309e-01 7.95478761e-01
-2.52586097e-01 4.33233768e-01 7.74336886e-03 2.31205165e-01
-3.47183764e-01 -8.52352977e-01 8.38910222e-01 6.72242492e-02
-1.21954016e-01 -3.78995351e-02 -8.18265259e-01 -2.56954748e-02
1.30401134e+00 -9.07851234e-02 -2.01865464e-01 -1.87660083e-01
-3.92686874e-01 -2.68396288e-01 3.84039372e-01 6.02619290e-01
5.37203670e-01 -1.01902890e+00 -6.64347172e-01 2.71423638e-01
1.97842732e-01 -9.89063919e-01 -3.32915634e-02 1.07541001e+00
-8.78951609e-01 -2.62183812e-03 -3.71810019e-01 -1.48770846e-02
-1.61063385e+00 4.61666316e-01 2.31253117e-01 -2.08489671e-01
-2.53665447e-01 4.98642325e-01 -6.61655068e-01 -1.45160690e-01
-2.87687540e-01 1.92105826e-02 -7.72916079e-01 7.34308064e-01
6.29570484e-01 3.89128923e-01 6.22761883e-02 -9.65845525e-01
-5.11079550e-01 1.66402146e-01 -1.98206767e-01 -1.25921056e-01
1.55564785e+00 1.08474277e-01 -7.19330251e-01 8.90796483e-01
1.58246696e+00 6.23853505e-01 -5.58037125e-02 2.45277092e-01
4.06638294e-01 -8.97943139e-01 5.51530197e-02 -3.36303413e-01
-5.61233461e-01 8.07890356e-01 5.69967866e-01 8.35133255e-01
6.75708592e-01 -2.25204960e-01 5.22118866e-01 1.69596151e-01
1.10289879e-01 -1.27266622e+00 2.05008596e-01 8.65899920e-01
8.03586900e-01 -1.35808730e+00 -2.79922877e-02 -2.72978425e-01
-5.28336346e-01 1.34707487e+00 3.04302305e-01 -2.55596757e-01
9.89169419e-01 1.41592562e-01 1.78017706e-01 -1.33181527e-01
-6.31251037e-01 3.58622633e-02 -4.23170254e-02 3.18171978e-01
8.56349111e-01 -2.25553624e-02 -1.18954945e+00 5.76238275e-01
-1.45714462e-01 -1.80463776e-01 9.77500618e-01 1.04264069e+00
-5.26147544e-01 -1.28400385e+00 -2.89770752e-01 6.43780529e-01
-1.20019650e+00 -9.85705201e-03 -5.54827988e-01 6.11842096e-01
4.11529809e-01 1.17011130e+00 5.62209859e-02 -5.83880007e-01
-9.97936949e-02 5.49711287e-01 -1.17730550e-01 -8.13632667e-01
-1.10771942e+00 -3.16723943e-01 4.43253219e-01 -2.00076610e-01
-3.77873421e-01 -6.50524676e-01 -9.21710253e-01 -5.29369295e-01
-4.31568861e-01 2.15997249e-01 9.73020673e-01 1.07258773e+00
8.93964916e-02 7.40060881e-02 9.86441314e-01 -5.48829019e-01
-5.22450686e-01 -1.23539889e+00 -4.94429797e-01 8.20522964e-01
4.07269031e-01 -7.32390404e-01 -6.26250088e-01 -3.72026086e-01] | [8.782768249511719, 10.508416175842285] |
4379fcda-5b4d-46ff-8ef4-e57212904ed7 | autoencoders-for-unsupervised-anomaly-1 | 2104.09051 | null | https://arxiv.org/abs/2104.09051v1 | https://arxiv.org/pdf/2104.09051v1.pdf | Autoencoders for unsupervised anomaly detection in high energy physics | Autoencoders are widely used in machine learning applications, in particular for anomaly detection. Hence, they have been introduced in high energy physics as a promising tool for model-independent new physics searches. We scrutinize the usage of autoencoders for unsupervised anomaly detection based on reconstruction loss to show their capabilities, but also their limitations. As a particle physics benchmark scenario, we study the tagging of top jet images in a background of QCD jet images. Although we reproduce the positive results from the literature, we show that the standard autoencoder setup cannot be considered as a model-independent anomaly tagger by inverting the task: due to the sparsity and the specific structure of the jet images, the autoencoder fails to tag QCD jets if it is trained on top jets even in a semi-supervised setup. Since the same autoencoder architecture can be a good tagger for a specific example of an anomaly and a bad tagger for a different example, we suggest improved performance measures for the task of model-independent anomaly detection. We also improve the capability of the autoencoder to learn non-trivial features of the jet images, such that it is able to achieve both top jet tagging and the inverse task of QCD jet tagging with the same setup. However, we want to stress that a truly model-independent and powerful autoencoder-based unsupervised jet tagger still needs to be developed. | ['Ivan Oleksiyuk', 'Alexander Mück', 'Alessandro Morandini', 'Michael Krämer', 'Thorben Finke'] | 2021-04-19 | null | null | null | null | ['jet-tagging'] | ['graphs'] | [-3.51655841e-01 6.74072504e-02 4.72095758e-01 -3.68865728e-01
-2.02284530e-01 -5.62690377e-01 7.93179333e-01 4.41981882e-01
-5.92363656e-01 2.26678312e-01 -6.15123250e-02 -2.50653327e-01
-1.71051383e-01 -8.45259666e-01 -6.48560762e-01 -9.87066388e-01
-6.90401644e-02 1.29753792e+00 6.61714911e-01 -4.48291928e-01
-3.16739902e-02 8.21582496e-01 -1.48002970e+00 3.48317564e-01
3.74534428e-01 9.77390468e-01 -7.63284713e-02 6.88587070e-01
-3.59789371e-01 3.95921618e-01 -4.82556731e-01 -7.42792368e-01
4.48340058e-01 -5.94951212e-01 -6.89453423e-01 -3.14286172e-01
3.08490187e-01 8.76628458e-02 -4.64515120e-01 1.47592926e+00
4.11990494e-01 2.79538482e-01 7.63293147e-01 -8.81544352e-01
-3.53065223e-01 4.35715497e-01 3.60361308e-01 6.10109568e-01
-2.35656902e-01 -1.21849656e-01 1.09543562e+00 -6.80463731e-01
6.07324719e-01 1.02847075e+00 5.29067099e-01 3.85843813e-01
-1.31205487e+00 -3.01420838e-01 -3.06966245e-01 3.93388659e-01
-7.35408723e-01 -1.09587729e-01 8.13128591e-01 -6.53487146e-01
8.85085344e-01 2.45927364e-01 7.28876531e-01 1.25077403e+00
5.11726081e-01 3.10004771e-01 1.07084000e+00 -6.06482744e-01
4.43608969e-01 2.98410922e-01 4.87728380e-02 4.40147191e-01
1.75858736e-01 6.32297456e-01 -7.04684183e-02 -3.22018087e-01
6.64894879e-01 -1.35082960e-01 3.35642636e-01 -8.31114292e-01
-1.12526572e+00 1.22405922e+00 5.19000530e-01 9.95757282e-01
-2.20525160e-01 -5.03524728e-02 7.65524924e-01 6.76480293e-01
5.28007686e-01 9.95646894e-01 -6.68336749e-01 7.20069557e-02
-7.65872478e-01 3.57476830e-01 8.09815407e-01 3.65219206e-01
4.53603953e-01 5.05771935e-01 -3.85676287e-02 5.81129730e-01
9.81873274e-03 2.07418725e-01 7.98965096e-01 -6.27338767e-01
-2.52761126e-01 4.39147502e-01 -1.43054262e-01 -7.46038675e-01
-5.44283390e-01 -6.91403866e-01 -7.77918279e-01 6.78089023e-01
5.23142099e-01 1.39425531e-01 -8.41876566e-01 1.30371952e+00
2.94130147e-01 9.75638535e-03 2.39099085e-01 9.61790681e-01
5.76894939e-01 3.92334849e-01 8.54214467e-03 1.23219974e-02
1.36366820e+00 -6.88974679e-01 -4.77896005e-01 -1.20268889e-01
6.19203806e-01 -8.95879924e-01 4.54558909e-01 4.54101503e-01
-7.86266446e-01 -6.81910515e-01 -9.65497017e-01 2.03184262e-01
-7.88062096e-01 -3.29906978e-02 8.71303856e-01 3.84145528e-01
-7.19901919e-01 1.10791767e+00 -1.04353619e+00 -4.82672691e-01
-2.08323076e-01 3.82685781e-01 -5.05346358e-01 4.56165910e-01
-1.04707766e+00 8.74586582e-01 9.96760309e-01 -1.63335368e-01
-5.85392118e-01 -3.93845856e-01 -5.78062117e-01 2.32891560e-01
1.50790662e-01 -4.16781068e-01 1.19745934e+00 -1.09119153e+00
-1.27922845e+00 9.52220142e-01 3.81052792e-01 -8.32841039e-01
4.56255794e-01 -5.93112782e-02 -1.03150034e+00 4.82726507e-02
-7.07440749e-02 1.31719902e-01 9.90705729e-01 -1.03691506e+00
-4.78458583e-01 -1.75358325e-01 -1.58774197e-01 -3.87322605e-01
-9.13824961e-02 3.25377345e-01 -9.43254158e-02 -9.14885283e-01
3.95990700e-01 -1.05379641e+00 -2.57681966e-01 -6.23870075e-01
-1.82566822e-01 -1.30541340e-01 7.06950486e-01 -4.93003488e-01
5.26807606e-01 -2.29161787e+00 1.64970249e-01 3.84225786e-01
1.21868432e-01 3.91236484e-01 7.15807080e-02 2.70888299e-01
-5.72718918e-01 -2.33282462e-01 -2.49855146e-01 1.43338763e-03
1.49369955e-01 6.89272285e-01 -2.61581868e-01 3.25514048e-01
2.85844564e-01 6.67237759e-01 -7.88742065e-01 -3.57779302e-02
3.30565661e-01 1.95105880e-01 -7.32322872e-01 2.66071379e-01
-3.34277511e-01 7.61459172e-01 -3.31648856e-01 2.07469285e-01
5.44999957e-01 1.62778303e-01 -2.38728635e-02 -2.32989043e-01
-7.32226148e-02 -2.47463323e-02 -1.11970139e+00 1.30765212e+00
-1.16888091e-01 5.40219128e-01 -7.14390799e-02 -1.53048170e+00
9.96371508e-01 4.45959866e-01 6.23673618e-01 -9.23281193e-01
1.68358743e-01 4.87712592e-01 7.16027975e-01 -3.04964751e-01
4.64046568e-01 -6.26006007e-01 -1.37222677e-01 2.25645840e-01
8.36634099e-01 2.18459308e-01 2.23049283e-01 1.20005332e-01
1.30945671e+00 -1.85688049e-01 -3.10412496e-02 -3.41116309e-01
5.38063765e-01 1.07222600e-02 5.21545827e-01 1.04086792e+00
1.16949536e-01 5.16410351e-01 3.09216082e-01 -9.55554128e-01
-1.59148312e+00 -1.16834116e+00 -4.36340600e-01 1.28179193e+00
-2.00191051e-01 -5.42710125e-01 -2.50628054e-01 -1.03252959e+00
3.36756855e-02 7.34467983e-01 -3.36004049e-01 -3.78639996e-01
-6.05829775e-01 -1.07691848e+00 2.79716909e-01 4.59401697e-01
5.56639694e-02 -1.37810946e+00 -1.14043243e-01 3.31024587e-01
3.67233634e-01 -1.27955115e+00 4.48773503e-01 8.05960357e-01
-7.79628158e-01 -9.58013296e-01 -3.26343775e-01 -4.75360930e-01
5.14048636e-01 -5.27315736e-01 1.39284003e+00 1.31626964e-01
-1.86449260e-01 3.79291058e-01 -5.85716844e-01 -4.40501064e-01
-1.16403508e+00 -2.96562403e-01 6.43358886e-01 -6.50659576e-02
4.96755540e-01 -6.64793611e-01 -2.89641261e-01 1.35261238e-01
-9.92254138e-01 -6.85527086e-01 7.77969956e-01 1.05203795e+00
4.85785156e-01 4.25367415e-01 1.30644456e-01 -7.91695297e-01
2.40490705e-01 -3.45272630e-01 -8.46980810e-01 -2.03310683e-01
-4.84176308e-01 5.02676904e-01 8.91399264e-01 -2.51122624e-01
-8.26893866e-01 -1.35379878e-03 -8.44712973e-01 -5.19058526e-01
-6.29812121e-01 1.50246667e-02 3.81415486e-02 -4.92109478e-01
6.90777540e-01 1.36477500e-01 -2.77024508e-01 -1.10078502e+00
6.49974644e-02 6.71095103e-02 7.08703876e-01 -5.93848646e-01
1.11279809e+00 2.96284109e-01 2.53897369e-01 -5.46650350e-01
-6.95099533e-01 -6.86435580e-01 -9.63906229e-01 1.55137047e-01
1.11207104e+00 -5.25428295e-01 -3.58272016e-01 -1.29844360e-02
-1.14071715e+00 1.31065980e-01 -7.28640079e-01 7.87890553e-01
-4.59007710e-01 5.37482381e-01 -4.37211156e-01 -4.94522363e-01
1.63224265e-02 -1.22826183e+00 7.90589213e-01 -1.30904377e-01
2.36969754e-01 -1.09081709e+00 4.84808385e-01 -1.33506909e-01
3.77813965e-01 -2.80865915e-02 1.33974195e+00 -1.87551033e+00
-1.57142237e-01 -5.45497298e-01 1.62374228e-02 7.68272221e-01
-3.50644648e-01 -3.65337729e-01 -1.01004577e+00 -4.01699781e-01
2.99024910e-01 1.85004547e-02 1.12932944e+00 -8.79534036e-02
1.23008370e+00 -1.99067622e-01 -7.33982846e-02 3.72331262e-01
1.35691619e+00 1.54643938e-01 3.32055986e-01 4.76788312e-01
5.68176806e-01 2.82841980e-01 1.63499966e-01 -2.89951777e-03
-5.10770023e-01 9.83267725e-01 5.18891692e-01 -7.16506764e-02
-2.19868943e-02 6.49387613e-02 3.07066411e-01 1.10766840e+00
-3.02115202e-01 -1.23079859e-01 -7.53521621e-01 3.25824797e-01
-1.72028100e+00 -9.21296775e-01 -2.76835024e-01 2.25479102e+00
2.31976762e-01 5.40085018e-01 1.63247794e-01 2.39263147e-01
4.56813633e-01 -1.34784415e-01 4.20093350e-02 -9.81403470e-01
-8.35859030e-02 6.95637167e-01 2.64767230e-01 2.27015838e-01
-1.39448571e+00 8.26463878e-01 6.21828747e+00 8.07908893e-01
-8.97671700e-01 5.11173666e-01 -2.70006135e-02 2.72468805e-01
8.78872536e-03 1.96414754e-01 -5.10101140e-01 7.16857076e-01
1.02052081e+00 4.83813763e-01 7.98322335e-02 1.14494443e+00
-4.66662288e-01 2.30670139e-01 -1.21233225e+00 6.46678269e-01
-1.01821162e-01 -1.10576248e+00 7.42507502e-02 1.14919707e-01
4.14453179e-01 3.97084743e-01 -3.26279819e-01 7.29600966e-01
2.22214207e-01 -8.02284777e-01 3.74926746e-01 4.28302139e-01
4.77100462e-02 -6.48261607e-01 1.26677597e+00 2.14265898e-01
-6.71846211e-01 2.20881291e-02 -8.05426478e-01 -1.35990158e-02
1.45313814e-02 8.58584940e-01 -7.96081305e-01 7.61893213e-01
8.35090339e-01 4.04854752e-02 -8.02934051e-01 1.25234997e+00
8.20294544e-02 7.03200042e-01 -5.01894057e-01 3.77957374e-01
4.13213134e-01 -3.78259122e-01 1.13840008e+00 1.41134024e+00
4.75962669e-01 -3.49116504e-01 3.21012676e-01 8.00596178e-01
2.73063570e-01 8.66274387e-02 -7.17921257e-01 -2.25132570e-01
-2.27997199e-01 1.36247075e+00 -8.81181717e-01 -2.30513856e-01
-4.73759085e-01 1.14119112e+00 1.00887701e-01 1.04621261e-01
-2.83565044e-01 -5.07165082e-02 6.75251901e-01 -3.21778767e-02
6.94857061e-01 -2.82557487e-01 4.12458658e-01 -1.18293250e+00
-1.64932460e-01 -6.16732836e-01 5.56940675e-01 -5.59734702e-01
-1.62248170e+00 7.77527452e-01 -4.64055724e-02 -1.14097369e+00
-6.83879852e-01 -1.26477718e+00 -8.63709867e-01 4.34311569e-01
-1.14346862e+00 -7.77282119e-01 1.51240081e-01 4.60681081e-01
1.71288073e-01 -8.32022965e-01 1.24859858e+00 1.91418096e-01
-1.87217280e-01 1.95370942e-01 5.84719837e-01 3.35238844e-01
5.66000223e-01 -1.87951124e+00 3.64273518e-01 8.83069515e-01
9.21553433e-01 1.57415524e-01 1.18886948e+00 -6.47948027e-01
-1.15752578e+00 -9.35837686e-01 7.57561564e-01 -5.93181372e-01
9.44388986e-01 -5.91753602e-01 -1.22951615e+00 5.86744785e-01
1.13524325e-01 5.58575332e-01 5.85303545e-01 3.62591207e-01
-1.00245126e-01 9.88480598e-02 -9.38094199e-01 1.60153210e-01
7.09452569e-01 -4.49887514e-01 -9.24621046e-01 6.66736126e-01
4.21518773e-01 -3.57228249e-01 -1.02901387e+00 6.20142102e-01
9.26337764e-02 -1.27297306e+00 1.18001938e+00 -1.00765467e+00
1.63667768e-01 -1.10586770e-01 3.92780937e-02 -1.40344071e+00
-6.90806091e-01 -1.10788383e-01 -2.36275122e-01 1.03821945e+00
6.45390302e-02 -4.52527583e-01 7.58627355e-01 -8.37980397e-03
-1.93755612e-01 -3.91560912e-01 -1.33553195e+00 -1.09770763e+00
2.80924022e-01 -6.46324694e-01 1.75559446e-01 8.57434332e-01
-4.53225076e-01 1.46531418e-01 -2.88658857e-01 4.86258626e-01
3.74752611e-01 6.44530058e-02 4.49622065e-01 -1.72939086e+00
-8.50633383e-01 -4.29021150e-01 -1.17971957e+00 -3.01293880e-01
2.74372339e-01 -1.42871809e+00 -9.83347967e-02 -6.77718103e-01
-7.90534467e-02 -3.86972040e-01 -7.19119966e-01 1.38780951e-01
3.47728819e-01 5.26150882e-01 7.86519572e-02 8.85955095e-02
-4.59875733e-01 3.87080163e-01 6.43477380e-01 -9.97697711e-02
2.46658996e-01 2.19088182e-01 3.86834182e-02 1.16698921e+00
6.67006135e-01 -8.09969187e-01 2.98537552e-01 -1.26745835e-01
5.09536266e-01 -3.62319916e-01 8.00346553e-01 -1.43585587e+00
2.02771034e-02 4.21166450e-01 8.87736797e-01 -2.92283893e-01
3.26756984e-01 -1.08027601e+00 -1.43060148e-01 3.45042199e-01
-8.43287483e-02 1.62552327e-01 2.52952993e-01 6.10781968e-01
-3.43695939e-01 -9.39842284e-01 6.19455099e-01 -5.16315341e-01
-9.57053483e-01 1.52046800e-01 -4.11128998e-01 -4.94873300e-02
5.13164937e-01 2.37267643e-01 1.77432671e-01 -8.73831213e-02
-1.25397623e+00 -2.59970218e-01 4.33594167e-01 4.66689348e-01
2.03106940e-01 -1.49092543e+00 -5.17696857e-01 5.74344158e-01
3.07442605e-01 -4.51282620e-01 5.02428450e-02 6.95990086e-01
-3.00133646e-01 4.50122893e-01 -4.68016177e-01 -6.60220504e-01
-1.05884206e+00 8.45444798e-01 3.69684070e-01 -4.05925572e-01
-1.07999957e+00 5.62959075e-01 2.81941652e-01 -5.62479138e-01
6.88792691e-02 -1.86835065e-01 -1.10158443e-01 -2.49256015e-01
2.41523206e-01 -1.25861794e-01 7.98975229e-01 -5.70183992e-01
-1.55592933e-01 2.97707766e-01 -1.35281786e-01 1.56998709e-01
1.29057419e+00 5.07739484e-01 -1.00035809e-01 4.49700981e-01
8.90124321e-01 2.90666789e-01 -5.56575596e-01 -1.83690906e-01
3.55228037e-01 -1.85767949e-01 2.08099797e-01 -8.47556949e-01
-9.42321539e-01 8.75257969e-01 8.87952626e-01 6.88281059e-01
6.80291533e-01 4.03040588e-01 5.76191425e-01 6.94012344e-01
2.80050095e-02 -1.02165675e+00 4.14756946e-02 7.75628090e-01
8.33336055e-01 -1.39448023e+00 -3.41877490e-01 1.04753986e-01
-4.80870336e-01 1.37800539e+00 3.46746147e-01 -4.16838735e-01
4.96372610e-01 2.33632952e-01 -7.84889087e-02 -5.08808553e-01
-3.55609208e-01 -5.10124743e-01 7.23788083e-01 5.28084159e-01
2.12621123e-01 -1.85577888e-02 -1.20705254e-01 6.49912298e-01
-4.48485911e-01 -6.23553276e-01 1.89403117e-01 5.36461294e-01
-3.87338698e-01 -1.63562512e+00 -3.37866783e-01 6.09766901e-01
-6.34758592e-01 -3.65921594e-02 -4.12352622e-01 9.03570056e-01
4.64952022e-01 2.47111663e-01 3.83653998e-01 -4.06763047e-01
1.21828407e-01 7.97268927e-01 5.34462988e-01 -6.21564090e-01
-6.20876551e-01 3.46365161e-02 -4.16357107e-02 -4.97071564e-01
-2.69917190e-01 -3.71585011e-01 -9.00339901e-01 3.48970741e-02
-2.44348213e-01 6.16154015e-01 7.88323581e-01 1.34514964e+00
1.36647537e-01 7.55122364e-01 2.98707396e-01 -8.53301585e-01
-4.76320028e-01 -9.49510753e-01 -7.41296411e-01 7.53836870e-01
2.27545589e-01 -1.02689850e+00 -4.97354090e-01 -4.91786242e-01] | [15.685892105102539, 2.922760486602783] |
3581a206-18c4-48ac-85f8-98f3398ba532 | tensor-basis-gaussian-process-models-of | 1912.10872 | null | https://arxiv.org/abs/1912.10872v1 | https://arxiv.org/pdf/1912.10872v1.pdf | Tensor Basis Gaussian Process Models of Hyperelastic Materials | In this work, we develop Gaussian process regression (GPR) models of hyperelastic material behavior. First, we consider the direct approach of modeling the components of the Cauchy stress tensor as a function of the components of the Finger stretch tensor in a Gaussian process. We then consider an improvement on this approach that embeds rotational invariance of the stress-stretch constitutive relation in the GPR representation. This approach requires fewer training examples and achieves higher accuracy while maintaining invariance to rotations exactly. Finally, we consider an approach that recovers the strain-energy density function and derives the stress tensor from this potential. Although the error of this model for predicting the stress tensor is higher, the strain-energy density is recovered with high accuracy from limited training data. The approaches presented here are examples of physics-informed machine learning. They go beyond purely data-driven approaches by embedding the physical system constraints directly into the Gaussian process representation of materials models. | ['Reese Jones', 'Laura Swiler', 'Ari Frankel'] | 2019-12-23 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 2.23255381e-01 3.05748731e-01 1.22815460e-01 -2.36581951e-01
-6.19671762e-01 -3.40482980e-01 5.01068890e-01 1.87164154e-02
-2.41904423e-01 3.30951750e-01 -1.21107832e-01 -3.00852284e-02
-4.84645039e-01 -7.18125224e-01 -8.48283648e-01 -1.14456844e+00
1.11849792e-01 8.06126416e-01 3.47906083e-01 -1.24436855e-01
5.10202706e-01 7.80838072e-01 -1.11739862e+00 -3.16926628e-01
5.18368840e-01 8.89796436e-01 -1.65985048e-01 8.00956249e-01
3.76260370e-01 3.72062474e-01 -1.38172314e-01 -2.17033222e-01
4.60192673e-02 8.93210918e-02 -9.24827635e-01 -1.13069318e-01
4.74684387e-02 -2.80337483e-01 -4.39639688e-01 6.67517364e-01
1.60140246e-01 5.55594683e-01 1.19910622e+00 -8.00024152e-01
-7.46771395e-01 2.82919407e-01 -4.64811414e-01 -3.29600990e-01
4.13512945e-01 -6.64179819e-03 9.87875998e-01 -7.95127094e-01
7.70719349e-01 1.21929121e+00 6.26593292e-01 4.73091334e-01
-1.89452648e+00 7.76258707e-02 -2.70863086e-01 -1.77083053e-02
-1.33091831e+00 4.06878709e-04 1.14162421e+00 -7.98905551e-01
9.73496556e-01 1.35917231e-01 3.83828640e-01 1.21153557e+00
7.48102963e-01 5.37172079e-01 1.11706936e+00 -5.05965233e-01
5.05292654e-01 -2.52613097e-01 5.27159452e-01 4.56881553e-01
3.40564966e-01 3.97714317e-01 -2.42000684e-01 -4.55381989e-01
1.08469975e+00 -8.31908360e-02 -5.05841337e-02 -7.49587178e-01
-8.42017114e-01 5.39016306e-01 2.57088989e-02 2.16227531e-01
-5.62049806e-01 6.40934706e-01 6.22703657e-02 -8.98393765e-02
4.81017888e-01 5.65735459e-01 -4.52931255e-01 -2.57191747e-01
-1.01293778e+00 8.04810286e-01 1.14529562e+00 7.56067216e-01
9.13147807e-01 5.86100444e-02 3.38039510e-02 4.77332294e-01
6.85761213e-01 6.96449578e-01 -1.48214623e-02 -1.25129986e+00
1.15142018e-01 1.60593852e-01 1.59176484e-01 -9.32524800e-01
-3.38211238e-01 5.14827408e-02 -3.25745553e-01 2.84553498e-01
5.27065039e-01 -4.03264388e-02 -9.97408092e-01 1.40615857e+00
2.25206256e-01 1.69459701e-01 -1.15881369e-01 6.19252443e-01
1.38144806e-01 5.48394203e-01 2.70756215e-01 -8.85521993e-02
1.19822192e+00 -4.41021532e-01 -5.65623045e-01 2.82421291e-01
2.18862578e-01 -9.22544658e-01 6.47997975e-01 4.16405559e-01
-1.38694835e+00 -4.89875585e-01 -9.49702442e-01 -9.43329260e-02
-3.30803450e-03 4.55590002e-02 4.24138665e-01 4.09244299e-01
-8.06331158e-01 1.35980189e+00 -1.40910363e+00 -4.19589616e-02
-1.91150025e-01 4.29344714e-01 -5.52762374e-02 2.40645990e-01
-7.41297960e-01 1.02055430e+00 7.13604763e-02 2.85043448e-01
-4.81928945e-01 -1.07102251e+00 -5.28059125e-01 3.24802250e-02
2.90037870e-01 -6.02487028e-01 1.26496446e+00 -9.69484895e-02
-1.97757816e+00 4.60292906e-01 -1.85342982e-01 3.72807831e-02
5.59413433e-01 -3.37016165e-01 -7.81854391e-02 4.23970282e-01
-4.29964423e-01 1.77366033e-01 1.05246985e+00 -1.61845982e+00
4.47276503e-01 -2.12705791e-01 -2.96086788e-01 -4.74733084e-01
3.21919560e-01 -8.04645866e-02 3.74182616e-03 -5.58302581e-01
5.99734664e-01 -1.41396546e+00 -5.28168738e-01 -1.90815687e-01
-6.26256943e-01 -1.98846310e-01 5.66245019e-01 -8.75054598e-01
8.76287699e-01 -1.70984745e+00 5.44351637e-01 7.92021692e-01
1.87688380e-01 -1.34759620e-01 1.88557386e-01 8.32562268e-01
-1.67621300e-01 3.00233752e-01 -5.77236712e-01 -3.16543519e-01
1.40016094e-01 1.84917927e-01 -4.51215863e-01 3.87436688e-01
6.35741711e-01 8.84574294e-01 -7.48204410e-01 -1.88794971e-01
3.35275382e-02 7.40226805e-01 -6.12024605e-01 7.86607414e-02
-1.80238172e-01 4.79900986e-01 -7.08048165e-01 4.75766212e-01
8.12883675e-01 1.10961355e-01 1.87821060e-01 -5.65106630e-01
-3.75671983e-02 -1.43542662e-02 -1.16275680e+00 1.23560548e+00
-2.82018006e-01 2.27656513e-01 -7.88755193e-02 -9.82983649e-01
1.22800553e+00 3.70138705e-01 1.11064565e+00 7.66211823e-02
9.54518020e-02 3.81194651e-01 4.72943857e-02 -6.79587066e-01
5.30767262e-01 -5.79165339e-01 -2.92535145e-02 3.28566462e-01
6.89516366e-02 -7.01707423e-01 1.71300583e-02 -2.32275084e-01
9.98346925e-01 1.10138273e+00 -2.16356501e-01 -4.31677580e-01
3.35619926e-01 -2.58742929e-01 2.37891778e-01 2.56885022e-01
2.29461223e-01 7.09530473e-01 6.87371612e-01 -5.86782396e-02
-1.58091891e+00 -1.40298986e+00 -3.91902536e-01 4.07916844e-01
-9.76243690e-02 -2.69881248e-01 -7.44161069e-01 6.03381870e-03
4.09208477e-01 6.69812322e-01 -6.14471376e-01 -2.30898246e-01
-1.10946643e+00 -9.56200659e-01 1.48989514e-01 8.16565573e-01
-7.89832175e-02 -6.90465868e-01 -2.68271983e-01 2.87720829e-01
2.46070549e-01 -1.19096696e+00 -5.62063269e-02 1.23348638e-01
-1.07280672e+00 -8.77328634e-01 -4.76312667e-01 -3.93223196e-01
6.64625943e-01 -4.68621463e-01 7.10756481e-01 -2.61299342e-01
-4.47630942e-01 6.63183093e-01 -5.92729300e-02 -1.26590028e-01
-7.95543909e-01 -1.54445380e-01 6.89989850e-02 -3.97715457e-02
1.40073106e-01 -8.04842949e-01 -5.31204224e-01 1.91160783e-01
-7.68201590e-01 -5.58460951e-01 3.38567078e-01 4.22797590e-01
8.74710977e-01 -1.23383000e-01 2.73398966e-01 -5.34723222e-01
5.64816892e-01 -3.07057053e-01 -4.14429605e-01 -6.25983402e-02
-8.62577379e-01 4.13625360e-01 3.88981164e-01 -6.97825730e-01
-1.20044315e+00 5.99986911e-02 -1.84335947e-01 -5.66093028e-01
-1.06698181e-02 4.41767335e-01 6.16203621e-02 -1.24929607e-01
3.87760788e-01 -8.41702744e-02 1.90556467e-01 -9.08344567e-01
2.29453385e-01 1.02188334e-01 4.32780832e-01 -1.45203829e+00
1.01268315e+00 4.16400880e-01 8.67289305e-01 -8.60464215e-01
-2.50925720e-01 -4.00824130e-01 -9.22110975e-01 -1.78535789e-01
8.20841551e-01 -1.17750481e-01 -1.13490403e+00 4.14453983e-01
-1.07629859e+00 -1.75306544e-01 -6.76838815e-01 6.15688682e-01
-1.09973288e+00 7.40191638e-01 -8.60390842e-01 -1.08264899e+00
-1.35354713e-01 -1.31300604e+00 1.38437009e+00 -2.92815380e-02
-3.48551780e-01 -1.19809818e+00 3.36836398e-01 8.38223621e-02
3.34814936e-01 5.93696654e-01 1.13315320e+00 -5.22604823e-01
-6.52792990e-01 -5.93755364e-01 1.59167156e-01 5.43894351e-01
-1.29120588e-01 7.02173829e-01 -8.75491679e-01 -1.02545992e-02
4.19021577e-01 2.03108177e-01 6.35130584e-01 6.26282275e-01
1.00004160e+00 6.30714968e-02 -3.25594634e-01 3.24867606e-01
1.75332355e+00 5.14500029e-02 6.63921475e-01 -8.53658095e-03
9.31918502e-01 8.94447863e-01 3.01571757e-01 2.42483109e-01
1.66601576e-02 8.12013388e-01 -2.02184934e-02 2.54666597e-01
-3.16371024e-02 -2.35267267e-01 4.03896421e-01 8.25196981e-01
-9.99646723e-01 3.17094713e-01 -9.92155671e-01 8.94544274e-02
-1.80301976e+00 -1.07057035e+00 -5.93168914e-01 2.23912334e+00
7.29655027e-01 8.55320394e-02 5.88957705e-02 6.38359264e-02
3.67493659e-01 -2.12393090e-01 -4.55527008e-01 -5.77214837e-01
2.08489522e-01 5.76867223e-01 6.46934390e-01 8.01473439e-01
-7.42258132e-01 4.76001561e-01 7.90133047e+00 3.63612652e-01
-8.84755671e-01 -2.19141126e-01 3.00783724e-01 3.55851710e-01
-4.58735436e-01 2.78889954e-01 -9.13265824e-01 4.25886989e-01
1.30514240e+00 -1.17093280e-01 3.54382157e-01 7.95414746e-01
4.53084260e-01 -1.43656626e-01 -1.32305253e+00 3.55602115e-01
-2.98231274e-01 -9.53525841e-01 -2.90180653e-01 3.98322344e-01
5.05118668e-01 -4.05673057e-01 -2.22648457e-02 2.37559490e-02
2.74631754e-02 -7.95432746e-01 6.91875339e-01 1.42419636e+00
4.07314390e-01 -4.02797878e-01 5.60836971e-01 7.40206018e-02
-9.07494724e-01 1.08274952e-01 -3.91525269e-01 2.19118685e-01
5.21377206e-01 7.14636207e-01 -7.02284455e-01 5.79339206e-01
3.21201891e-01 4.52729583e-01 -3.07834983e-01 8.25146794e-01
-4.38777357e-02 7.32851744e-01 -5.14893591e-01 2.11922809e-01
-2.81602472e-01 -7.34901786e-01 8.66044998e-01 9.38986123e-01
3.33791852e-01 -2.21763253e-01 1.60306484e-01 1.25909626e+00
3.33541512e-01 -5.39186150e-02 -1.77307546e-01 5.29385433e-02
8.79341736e-02 1.13125086e+00 -5.88257849e-01 8.77914652e-02
-3.62920910e-02 4.24236655e-01 -5.09755351e-02 7.32155144e-01
-6.04764342e-01 4.47080545e-02 5.99811316e-01 5.25663614e-01
2.97637314e-01 -8.17024827e-01 -2.59295046e-01 -9.36944246e-01
2.29147300e-01 -3.58289897e-01 -4.89114791e-01 -7.45989680e-01
-1.52553880e+00 5.84628135e-02 8.05671513e-01 -8.06104779e-01
-6.00453496e-01 -1.16401505e+00 -6.24337494e-01 1.06356466e+00
-1.31330097e+00 -1.23835850e+00 2.28425354e-01 -4.01449949e-02
3.13316152e-04 2.90083885e-01 6.90030873e-01 -1.09587468e-01
-4.71158952e-01 4.78423983e-02 3.16161484e-01 -2.04955235e-01
5.47425032e-01 -1.55871797e+00 4.82294291e-01 3.61335039e-01
-5.67352891e-01 1.11794627e+00 1.29012227e+00 -9.54506636e-01
-1.64070618e+00 -5.05629599e-01 4.06358123e-01 -6.08076930e-01
1.18062615e+00 -2.16369674e-01 -1.19919884e+00 6.53389633e-01
-3.64710778e-01 2.94643361e-03 6.17544472e-01 6.15661889e-02
-1.02895930e-01 1.68944493e-01 -1.00409782e+00 3.79387528e-01
6.20599151e-01 -5.55483997e-01 -5.71538985e-01 2.39495769e-01
5.63189149e-01 -1.74175084e-01 -1.75891876e+00 4.27102327e-01
6.43889487e-01 -2.77444363e-01 1.25381052e+00 -6.98803544e-01
5.42145967e-01 -1.47565231e-01 -1.91984847e-01 -8.74285638e-01
-3.41514498e-01 -9.34069991e-01 -5.18001676e-01 1.07542980e+00
2.36913651e-01 -6.49287224e-01 8.38390112e-01 1.27468193e+00
-1.29603580e-01 -1.12844539e+00 -8.62883091e-01 -1.02446568e+00
6.48243010e-01 -4.14607197e-01 1.75194323e-01 3.96557033e-01
-1.20456643e-01 -7.96614811e-02 -1.20240971e-01 3.33224721e-02
9.40267026e-01 -2.10436225e-01 3.18124503e-01 -1.49395037e+00
-6.41710758e-01 -1.83839977e-01 -4.19687659e-01 -8.84287477e-01
1.88273966e-01 -6.24317408e-01 1.16636783e-01 -1.30188274e+00
6.70574531e-02 -4.40466404e-01 5.86280376e-02 -3.26703601e-02
1.05945289e-03 -2.20937207e-01 -1.55210439e-02 6.24189913e-01
4.21817720e-01 4.25858915e-01 1.44013488e+00 2.38140553e-01
-3.09942901e-01 1.38962686e-01 -1.54985145e-01 8.57236028e-01
8.21213484e-01 -2.92073339e-01 -1.99456066e-01 1.00736782e-01
3.36223245e-01 1.31699085e-01 6.60321236e-01 -7.63397396e-01
8.43672231e-02 -3.15103203e-01 2.78438807e-01 -4.68086571e-01
5.26267886e-01 -7.37967193e-01 3.65281403e-01 2.25852683e-01
-3.23797166e-01 3.87910753e-02 -1.59210324e-01 7.23410428e-01
1.67945594e-01 -7.15033948e-01 6.65550530e-01 1.57592013e-01
4.49067503e-02 1.47211388e-01 -5.56358874e-01 -4.52664524e-01
7.68554509e-01 -4.24200267e-01 -1.26607180e-01 8.26272368e-02
-1.20939624e+00 -5.02978921e-01 6.58084691e-01 3.64050008e-02
4.64885205e-01 -1.04718053e+00 -4.73680884e-01 3.51658277e-02
-5.14621079e-01 -3.06293946e-02 2.47690290e-01 9.35632765e-01
-6.28696859e-01 1.95306525e-01 5.89829311e-02 -5.70164919e-01
-8.36286306e-01 6.51760936e-01 5.64617097e-01 -4.71364051e-01
-4.90777850e-01 3.55055690e-01 -2.72403181e-01 -3.24558616e-01
-5.11895120e-01 -5.90844333e-01 2.33817428e-01 -2.31831685e-01
-7.68820569e-02 8.15926790e-01 9.10582207e-03 -7.79818833e-01
-2.03874633e-01 1.21161115e+00 1.92411125e-01 -3.34211111e-01
1.41963565e+00 7.90868700e-02 -2.16472313e-01 5.81659436e-01
1.29743278e+00 1.74112871e-01 -1.44337285e+00 7.93887079e-02
2.18623206e-01 -9.20478478e-02 2.18589474e-02 -3.84849221e-01
-6.31927133e-01 8.22990954e-01 2.18887150e-01 3.17234516e-01
5.84824920e-01 9.01551098e-02 6.50124848e-01 2.98134148e-01
1.32568836e-01 -1.34752810e+00 2.78030694e-01 2.49715030e-01
9.00904655e-01 -6.85618818e-01 4.39796984e-01 -9.74817455e-01
-1.27365872e-01 1.58209443e+00 2.21409649e-01 -6.49985611e-01
1.18052876e+00 5.17339647e-01 -4.72490788e-01 -2.91854352e-01
-2.06822231e-01 2.82880306e-01 5.49565673e-01 5.96204281e-01
5.54902911e-01 1.28548294e-01 -4.19854045e-01 3.59803408e-01
-2.33949944e-01 -1.16156168e-01 5.15926480e-01 1.18754053e+00
-3.74186248e-01 -1.50120425e+00 -5.01530051e-01 2.74774339e-02
-4.23315227e-01 1.18808568e-01 2.20913757e-02 6.57657027e-01
-3.13640863e-01 7.06257999e-01 -5.83042540e-02 -3.08410764e-01
5.45147061e-01 3.80485982e-01 8.21622789e-01 -4.71543968e-01
-1.36609718e-01 3.47473681e-01 5.76189756e-02 -3.75677913e-01
-4.37936604e-01 -1.09863567e+00 -1.39030397e+00 -3.20112407e-01
-2.09762022e-01 -2.44224995e-01 1.08041477e+00 9.93166685e-01
9.79669690e-02 4.56865311e-01 4.80533004e-01 -1.24520099e+00
-1.11843920e+00 -7.27908671e-01 -6.44129694e-01 4.17290509e-01
6.74899146e-02 -7.80036211e-01 -3.42927396e-01 4.02784109e-01] | [6.419939994812012, 3.453782320022583] |
c9880c94-da30-4858-813f-7010ea0fa588 | open-source-iris-recognition-hardware-and | 2008.08220 | null | https://arxiv.org/abs/2008.08220v1 | https://arxiv.org/pdf/2008.08220v1.pdf | Open Source Iris Recognition Hardware and Software with Presentation Attack Detection | This paper proposes the first known to us open source hardware and software iris recognition system with presentation attack detection (PAD), which can be easily assembled for about 75 USD using Raspberry Pi board and a few peripherals. The primary goal of this work is to offer a low-cost baseline for spoof-resistant iris recognition, which may (a) stimulate research in iris PAD and allow for easy prototyping of secure iris recognition systems, (b) offer a low-cost secure iris recognition alternative to more sophisticated systems, and (c) serve as an educational platform. We propose a lightweight image complexity-guided convolutional network for fast and accurate iris segmentation, domain-specific human-inspired Binarized Statistical Image Features (BSIF) to build an iris template, and to combine 2D (iris texture) and 3D (photometric stereo-based) features for PAD. The proposed iris recognition runs in about 3.2 seconds and the proposed PAD runs in about 4.5 seconds on Raspberry Pi 3B+. The hardware specifications and all source codes of the entire pipeline are made available along with this paper. | ['Zhaoyuan Fang', 'Adam Czajka'] | 2020-08-19 | null | null | null | null | ['iris-segmentation'] | ['medical'] | [ 2.66201347e-01 -3.72020245e-01 -7.85742253e-02 -3.73462170e-01
-1.38261482e-01 -6.03542984e-01 3.71336550e-01 3.75603810e-02
-5.13092875e-01 2.20787451e-01 -1.80400833e-01 -9.14022207e-01
-2.24254936e-01 -6.56596899e-01 -2.88860857e-01 -8.25903773e-01
-5.57818860e-02 1.65317804e-01 -2.05405459e-01 -6.39905110e-02
5.12422860e-01 1.02449429e+00 -1.73174250e+00 -8.94111395e-02
9.56898153e-01 1.17454839e+00 -5.47485888e-01 1.14747369e+00
3.93134892e-01 7.56813169e-01 -5.65465629e-01 -4.65596676e-01
4.97107476e-01 -2.89996803e-01 -8.49333167e-01 9.85494535e-03
7.88167536e-01 -5.89144766e-01 -4.07494456e-02 1.15646708e+00
7.37899959e-01 1.81214772e-02 4.67277952e-02 -7.18703747e-01
-4.67953980e-01 2.70813525e-01 -7.06991911e-01 1.98336720e-01
3.45209062e-01 5.21624029e-01 1.16250344e-01 -6.10339604e-02
1.16809048e-01 9.77229834e-01 5.48312902e-01 5.92413247e-01
-7.18182385e-01 -7.31900692e-01 -8.47929835e-01 1.66265462e-02
-1.26562500e+00 -2.58296847e-01 4.61132735e-01 -2.19727188e-01
9.08833623e-01 6.36603236e-01 7.75494099e-01 5.64531684e-01
2.26906553e-01 3.53906006e-01 1.80648077e+00 -6.13799989e-01
-2.26458117e-01 3.34455699e-01 3.96162748e-01 9.45715308e-01
4.63115603e-01 6.66014850e-01 -3.80193703e-02 -7.10776970e-02
1.05362356e+00 1.03464179e-01 4.43209661e-03 2.64399320e-01
-7.86781490e-01 5.54438233e-01 4.87627536e-01 3.57344419e-01
-2.65983909e-01 -2.51192629e-01 1.35128364e-01 2.19020218e-01
-1.46343959e-02 3.43277067e-01 -2.00284526e-01 -4.73280340e-01
-6.86226845e-01 -1.04810126e-01 8.00161302e-01 4.09085870e-01
5.72453082e-01 8.59962404e-02 3.68457288e-02 5.41094124e-01
3.65476429e-01 8.44713092e-01 7.16009140e-01 -2.19644591e-01
-2.89488405e-01 7.42498755e-01 2.39076495e-01 -9.14853811e-01
-5.43422580e-01 -2.44238511e-01 -6.57592893e-01 5.06225646e-01
7.50688612e-01 -1.76956862e-01 -9.58483398e-01 6.06004953e-01
4.99691248e-01 6.42044544e-01 2.72467639e-02 1.12358236e+00
9.66418207e-01 2.06981972e-01 -3.09023619e-01 3.07837844e-01
1.78649318e+00 -7.35469520e-01 -2.36538619e-01 4.28381681e-01
3.72982800e-01 -1.36691785e+00 8.35849226e-01 4.78850514e-01
-7.01438963e-01 -6.84125245e-01 -1.16206348e+00 -1.38751805e-01
-5.32375097e-01 3.23118299e-01 8.99510682e-01 1.76851511e+00
-1.03546000e+00 5.08984983e-01 -9.49630201e-01 -3.67060661e-01
3.31860632e-01 9.84290838e-01 -1.85511902e-01 3.23846310e-01
-4.63689029e-01 5.35189271e-01 1.22883521e-01 -4.26898897e-02
-4.75336820e-01 -2.63956934e-01 -8.50430965e-01 -1.46820664e-01
-1.19126387e-01 -3.04028779e-01 9.08131778e-01 -1.03363991e+00
-2.12983751e+00 1.35832751e+00 3.77412140e-02 -7.67649949e-01
-5.41517511e-03 -4.33923006e-02 -5.74241638e-01 2.90550828e-01
-6.58888519e-01 2.79221565e-01 1.10478985e+00 -3.11705142e-01
-8.24838817e-01 -5.54303288e-01 7.32636005e-02 -2.39018679e-01
-1.89753488e-01 7.66836166e-01 -1.35955930e-01 -4.52684969e-01
-2.71994144e-01 -9.19121206e-01 -1.06491633e-01 -3.08352143e-01
-3.21522802e-01 -1.95388928e-01 6.63416266e-01 -9.37533259e-01
8.29668820e-01 -2.12030911e+00 -6.93585336e-01 4.35921013e-01
1.58926751e-02 1.23478031e+00 1.18233904e-01 -2.08490491e-01
-1.84429511e-01 -2.16236949e-01 3.21184963e-01 6.74464852e-02
-2.77783811e-01 -9.40248817e-02 -2.45991305e-01 8.32423449e-01
-1.53523553e-02 6.78647161e-01 -5.92196703e-01 -2.65435696e-01
9.62483943e-01 8.07060182e-01 -3.12599510e-01 8.80843401e-02
5.81102610e-01 6.39341176e-01 -2.21885070e-01 1.06873655e+00
1.07633376e+00 4.34204191e-02 -3.77056688e-01 -9.90269706e-02
-4.64298010e-01 3.21644455e-01 -1.41703773e+00 1.01080465e+00
-2.03338444e-01 5.61293900e-01 -3.96280643e-03 -6.15481973e-01
1.16868818e+00 2.67210871e-01 2.41568923e-01 -3.67027760e-01
6.82368159e-01 2.74761379e-01 -3.23521043e-03 -5.29759288e-01
3.03322136e-01 1.56886950e-01 5.63967168e-01 6.71525896e-01
5.54562621e-02 1.69204906e-01 -1.69071808e-01 -4.13893461e-01
4.89724159e-01 8.03775936e-02 4.45746034e-01 -1.86958954e-01
1.14536595e+00 -2.28023663e-01 3.05231839e-01 6.39615655e-01
-5.85180640e-01 1.70560524e-01 -4.22685593e-02 -1.33486784e+00
-6.94311500e-01 -4.98887628e-01 -6.40706062e-01 2.57145166e-01
2.34611705e-01 1.41080871e-01 -9.34005916e-01 -3.98667783e-01
-1.51200488e-01 -8.94221067e-02 -1.81640133e-01 4.82361883e-01
-3.88490558e-01 -7.98388720e-01 8.31520081e-01 7.04959556e-02
8.84556293e-01 -8.57676744e-01 -7.33329177e-01 -2.56332681e-02
3.77545297e-01 -6.38882816e-01 -3.30221891e-01 -3.34858716e-01
-7.26517141e-01 -1.54051936e+00 -4.30025339e-01 -8.69543076e-01
7.92606235e-01 1.92953289e-01 4.99425709e-01 5.13817608e-01
-8.82156909e-01 8.18263367e-02 -1.59347415e-01 -6.72828317e-01
-7.39262300e-03 -4.47804898e-01 3.66733164e-01 5.14835119e-01
1.23726904e+00 -5.46499714e-02 -8.61118555e-01 2.37955660e-01
-6.67556524e-01 -2.90832341e-01 2.15422228e-01 8.51351678e-01
4.20448899e-01 5.31859286e-02 -3.19799304e-01 -4.62846518e-01
2.43064627e-01 3.10267836e-01 -1.38131261e+00 1.09009162e-01
-8.20863783e-01 -2.67901301e-01 3.44911098e-01 -3.39365363e-01
-8.22664320e-01 1.03592120e-01 -3.12877923e-01 -3.14737251e-03
-6.15915835e-01 -2.33765971e-02 5.18291831e-01 -1.23603272e+00
8.02056491e-01 1.00673378e-01 4.42120075e-01 -5.84021389e-01
-5.41067391e-04 1.46610689e+00 7.90405512e-01 -3.22718948e-01
6.45505726e-01 6.35476112e-01 1.93873465e-01 -9.85565543e-01
-3.89094353e-01 -7.26971388e-01 -5.49657166e-01 9.56975818e-02
7.17923462e-01 -8.36995006e-01 -1.65656948e+00 1.18403316e+00
-6.78929627e-01 2.40209419e-03 -4.66240346e-02 6.77130580e-01
-1.63935527e-01 6.54354870e-01 -9.97262418e-01 -9.93867934e-01
-7.99970269e-01 -1.46125877e+00 7.89105237e-01 1.12244821e+00
-5.74378818e-02 -8.42362404e-01 -3.94468792e-02 8.91730607e-01
5.88012218e-01 3.56708139e-01 1.08948052e-01 -4.81497318e-01
-6.52436018e-01 -6.06334567e-01 -4.58910346e-01 4.68397707e-01
2.34508410e-01 4.92004991e-01 -1.21256077e+00 -5.75098932e-01
1.80452451e-01 -2.25402698e-01 5.32745779e-01 6.58659101e-01
1.15530586e+00 -3.51660311e-01 -1.88212752e-01 1.39003837e+00
1.57511199e+00 4.41002131e-01 8.87300730e-01 4.06341523e-01
3.55854541e-01 3.46334308e-01 2.19649121e-01 3.89482170e-01
9.49295983e-02 6.10471785e-01 1.13465391e-01 -6.05719805e-01
-4.47889231e-02 2.65142322e-02 3.54809910e-02 3.50741297e-01
-6.89269483e-01 4.49920416e-01 -8.43370318e-01 3.88478696e-01
-1.10199940e+00 -7.46687651e-01 -1.49161220e-01 2.60606647e+00
9.06651258e-01 -5.35148859e-01 3.22057784e-01 1.16370924e-01
6.01279199e-01 -3.66934061e-01 -1.73423275e-01 -7.96928644e-01
-1.28546609e-02 1.24444461e+00 7.96423852e-01 7.98604906e-01
-1.46020377e+00 9.84456480e-01 5.42616177e+00 7.35197902e-01
-1.68525338e+00 -1.10838465e-01 7.16002047e-01 1.18141204e-01
5.90730906e-01 -5.37957475e-02 -1.06256664e+00 4.01080787e-01
1.18749845e+00 1.21189430e-01 6.13951921e-01 7.97820091e-01
-1.39043123e-01 -7.62300417e-02 -3.93208027e-01 1.28306842e+00
2.22612568e-03 -1.25155449e+00 -4.32317257e-01 3.42416525e-01
5.66550136e-01 6.82694605e-03 3.33428621e-01 -3.25002849e-01
2.29428947e-01 -1.10409188e+00 -3.68397713e-01 3.10467720e-01
1.07400644e+00 -9.56006229e-01 1.14114726e+00 -1.85209826e-01
-8.15817118e-01 -1.35751560e-01 -4.94981378e-01 -3.71521264e-01
-5.15480638e-01 1.60356030e-01 -8.36230099e-01 3.11237842e-01
7.90762305e-01 4.50899214e-01 -6.25299335e-01 1.41037834e+00
-9.72036794e-02 6.67371571e-01 -5.42245984e-01 -2.96594381e-01
2.61413604e-01 -3.42273057e-01 3.45480531e-01 9.66320813e-01
3.08851510e-01 2.85510927e-01 -2.55422205e-01 4.35667485e-01
3.31722170e-01 2.66858548e-01 -2.99186498e-01 -3.22652936e-01
9.97786894e-02 1.35767186e+00 -7.48370349e-01 -2.45373935e-01
-4.41043794e-01 6.97688818e-01 -3.93260837e-01 -8.77544805e-02
-3.76920342e-01 -9.85403717e-01 9.95833933e-01 -1.06313549e-01
1.50503024e-01 3.98681834e-02 -5.09481668e-01 -1.16564167e+00
-4.05002773e-01 -1.20935297e+00 3.25471312e-01 -3.32753509e-01
-8.90355885e-01 5.80136240e-01 -8.81560922e-01 -1.25717425e+00
-6.56961352e-02 -1.13085341e+00 -4.75107342e-01 1.39368737e+00
-1.53806019e+00 -1.26718485e+00 -3.88279498e-01 9.56870198e-01
-9.52379778e-02 -6.96981132e-01 1.23010123e+00 1.03231475e-01
-7.42626309e-01 9.38622057e-01 6.09505512e-02 5.64098120e-01
7.11263776e-01 -1.07195020e+00 4.05594945e-01 1.14814365e+00
3.41154873e-01 1.06537390e+00 2.56350309e-01 -4.29790378e-01
-1.55803490e+00 -4.99297708e-01 1.11310375e+00 -6.25187099e-01
2.88749129e-01 3.25768918e-01 -4.83545333e-01 4.29187089e-01
2.72922099e-01 5.54643720e-02 9.93311465e-01 5.18492833e-02
-2.99347252e-01 -4.72389013e-01 -1.75550675e+00 4.97422218e-01
-1.67824551e-02 -6.06133878e-01 -9.98995602e-02 4.14970666e-01
2.44451076e-01 -1.06070042e+00 -8.97418678e-01 4.95435856e-02
7.50336051e-01 -1.20905244e+00 9.62645531e-01 -4.57636505e-01
-1.17145851e-02 -6.36225939e-01 3.99739087e-01 -4.96665537e-01
-8.26975927e-02 -1.35296714e+00 1.11154489e-01 8.70475948e-01
-7.30521455e-02 -1.17412794e+00 8.65770698e-01 7.24334478e-01
3.59416276e-01 -5.14954209e-01 -9.47514832e-01 -5.21342516e-01
-3.09600443e-01 -2.31750757e-01 1.11845040e+00 7.84640253e-01
2.31080726e-01 -4.44479197e-01 -4.14629996e-01 6.43625319e-01
8.44002426e-01 2.69194514e-01 8.17118883e-01 -1.02575314e+00
-3.03557992e-01 -4.10352051e-01 -1.08641016e+00 -7.64187694e-01
-5.05440533e-01 -4.15829062e-01 -7.03626812e-01 -4.26059365e-01
-2.49917835e-01 -5.54571629e-01 -4.44235295e-01 7.67817616e-01
-8.44072259e-04 6.55718088e-01 -1.71892136e-01 -4.76649329e-02
2.53768206e-01 -5.31620324e-01 9.33792114e-01 -1.30229428e-01
-2.78179586e-01 4.61907566e-01 -6.93769574e-01 4.51988757e-01
7.51882315e-01 3.09020057e-02 8.85622352e-02 4.93568666e-02
-3.24124932e-01 3.51551501e-03 4.52675164e-01 -1.22926581e+00
3.55433196e-01 1.61856655e-02 5.86521685e-01 -1.82107374e-01
-3.66085395e-02 -9.57642555e-01 -4.57388461e-02 9.13963199e-01
2.48014316e-01 -4.94891077e-01 3.10493350e-01 -2.74068266e-01
-2.99341947e-01 -1.44976273e-01 1.03981102e+00 1.91502526e-01
-4.30150777e-01 3.36648971e-01 1.13907401e-02 -8.15105379e-01
1.07085633e+00 -5.89867115e-01 -6.96053505e-01 8.74011144e-02
-2.68260092e-01 -2.22759187e-01 6.60378754e-01 3.54848593e-01
5.87906659e-01 -7.82741487e-01 -4.78313029e-01 1.29516792e+00
-4.68677953e-02 -5.05132973e-01 2.47151718e-01 7.59483695e-01
-1.33788776e+00 9.45874274e-01 -5.55539131e-01 -6.07405663e-01
-1.92964005e+00 5.23393273e-01 5.95275164e-01 5.40650710e-02
-6.57573044e-01 1.29241836e+00 -6.46662116e-01 -2.39911288e-01
4.66730326e-01 -2.51669466e-01 -3.85596573e-01 -5.11830211e-01
1.48681259e+00 2.69777238e-01 1.08050540e-01 -6.79441214e-01
-1.81724280e-01 7.48989224e-01 -2.46584117e-01 3.41016501e-01
1.03295648e+00 1.25573367e-01 -6.07336819e-01 -5.75011015e-01
6.77432060e-01 1.79397073e-02 -9.28052127e-01 -2.50048220e-01
-3.17123741e-01 -1.02524793e+00 2.81756729e-01 -9.63399529e-01
-9.59958494e-01 8.12542439e-01 1.45455468e+00 -1.45473788e-02
1.47306633e+00 -7.85480678e-01 9.13985252e-01 -1.05935009e-02
5.45915782e-01 -7.99860895e-01 -6.34918094e-01 1.88969895e-01
1.01047300e-01 -1.32687879e+00 9.79880020e-02 -1.58477515e-01
-3.33133668e-01 1.41828644e+00 3.02209944e-01 -9.33803171e-02
7.16340423e-01 3.42473954e-01 6.36364877e-01 -1.02169506e-01
5.15809171e-02 -3.54876906e-01 6.08554125e-01 1.12717307e+00
6.52987599e-01 5.21558940e-01 -3.66750687e-01 1.74306825e-01
-4.79572415e-01 3.74988794e-01 5.20817161e-01 5.89878559e-01
-2.32718587e-01 -1.40869355e+00 -7.18720019e-01 4.88410085e-01
-1.01288319e+00 -9.83161405e-02 -9.42380801e-02 9.82835516e-02
4.09305036e-01 1.13442731e+00 7.15026185e-02 -4.58439976e-01
-1.52317047e-01 -1.87909529e-01 4.00161088e-01 -1.23696260e-01
-1.08705950e+00 2.03616098e-01 -1.26884863e-01 -6.61926508e-01
-4.25007045e-01 -3.98266464e-01 -5.59106171e-01 -9.04486239e-01
-2.42399096e-01 -1.63751885e-01 1.04627502e+00 7.67858863e-01
3.96182060e-01 -1.87155947e-01 7.13691831e-01 -5.52677393e-01
-3.30225080e-01 -9.37981784e-01 -8.42807412e-01 7.04255328e-02
7.14511812e-01 -1.39222339e-01 -1.72798365e-01 1.20979458e-01] | [3.7475852966308594, -3.6271414756774902] |
bf1cd79d-a5ae-41cf-be95-f1a28e99b07f | semeval-2015-task-3-answer-selection-in-1 | 1911.11403 | null | https://arxiv.org/abs/1911.11403v1 | https://arxiv.org/pdf/1911.11403v1.pdf | SemEval-2015 Task 3: Answer Selection in Community Question Answering | Community Question Answering (cQA) provides new interesting research directions to the traditional Question Answering (QA) field, e.g., the exploitation of the interaction between users and the structure of related posts. In this context, we organized SemEval-2015 Task 3 on "Answer Selection in cQA", which included two subtasks: (a) classifying answers as "good", "bad", or "potentially relevant" with respect to the question, and (b) answering a YES/NO question with "yes", "no", or "unsure", based on the list of all answers. We set subtask A for Arabic and English on two relatively different cQA domains, i.e., the Qatar Living website for English, and a Quran-related website for Arabic. We used crowdsourcing on Amazon Mechanical Turk to label a large English training dataset, which we released to the research community. Thirteen teams participated in the challenge with a total of 61 submissions: 24 primary and 37 contrastive. The best systems achieved an official score (macro-averaged F1) of 57.19 and 63.7 for the English subtasks A and B, and 78.55 for the Arabic subtask A. | ['Lluís Màrquez', 'Walid Magdy', 'James Glass', 'Preslav Nakov', 'Bilal Randeree', 'Alessandro Moschitti'] | 2019-11-26 | semeval-2015-task-3-answer-selection-in | https://aclanthology.org/S15-2047 | https://aclanthology.org/S15-2047.pdf | semeval-2015-6 | ['answer-selection'] | ['natural-language-processing'] | [-3.22013587e-01 2.39509344e-02 5.37424743e-01 -3.97609919e-01
-1.40071762e+00 -1.16363931e+00 4.92575616e-01 6.11046433e-01
-7.00670302e-01 7.56463349e-01 4.15425658e-01 -5.77169240e-01
1.36547023e-02 -7.60342240e-01 -3.93628597e-01 -4.21751767e-01
3.51153910e-01 8.85731101e-01 3.34545135e-01 -8.92342091e-01
5.08805931e-01 -1.29860193e-01 -1.37462127e+00 7.07092285e-01
1.16431475e+00 1.15107036e+00 -5.46925180e-02 8.06688130e-01
-3.63359272e-01 1.31164455e+00 -9.15497184e-01 -9.60875630e-01
-3.91467333e-01 -5.26712298e-01 -1.76972973e+00 -2.76740015e-01
3.62403601e-01 -5.05554415e-02 6.18102215e-02 7.29574442e-01
5.20547688e-01 2.15780765e-01 2.79596657e-01 -9.12262321e-01
-1.07968605e+00 3.40397447e-01 3.51523794e-03 1.81594387e-01
1.15851724e+00 1.37453705e-01 1.48938441e+00 -1.24113083e+00
6.32882297e-01 1.27320147e+00 3.63158047e-01 5.88447750e-01
-8.15516710e-01 -2.66524732e-01 -2.40191251e-01 4.26875502e-01
-1.21489072e+00 -4.47540879e-01 1.77091181e-01 -6.90132380e-01
9.08219874e-01 5.73261619e-01 4.83165905e-02 5.58904529e-01
-1.69688091e-01 7.68230259e-01 1.35096407e+00 -5.72279990e-01
4.82441396e-01 7.39684179e-02 5.21794319e-01 3.35922897e-01
-4.34427112e-01 -7.05451965e-01 -4.88244772e-01 -5.12525380e-01
-2.88656086e-01 -5.68325877e-01 -2.32535124e-01 4.17731613e-01
-1.05937243e+00 1.19437516e+00 3.85409504e-01 2.61134744e-01
-4.98543829e-01 -3.94083768e-01 3.59869361e-01 7.03272045e-01
3.92044902e-01 8.70024085e-01 -7.68897831e-01 -5.93518950e-02
-2.49562189e-01 8.25818717e-01 1.41873157e+00 6.49012446e-01
9.01497483e-01 -7.12630570e-01 -6.82522058e-01 1.07043254e+00
3.66015732e-01 7.58305967e-01 3.33585918e-01 -1.13719773e+00
7.54822373e-01 8.87601733e-01 8.14240158e-01 -1.07202494e+00
-3.96836817e-01 8.93741995e-02 -3.20648551e-01 -4.54254627e-01
1.00068998e+00 -4.57559079e-01 -3.99584413e-01 1.45251977e+00
4.88848686e-01 -8.41373682e-01 2.18376160e-01 1.01171553e+00
1.50407112e+00 8.02304506e-01 1.54656157e-01 -6.92559779e-02
1.98821950e+00 -1.20655954e+00 -9.86283660e-01 -2.02318266e-01
8.15438926e-01 -1.30596209e+00 1.46358144e+00 1.62211999e-01
-1.16967630e+00 -3.66460681e-01 -4.29910392e-01 -4.94654149e-01
-6.14356816e-01 2.23026261e-01 9.25686285e-02 5.47394037e-01
-1.14238906e+00 -4.69424948e-02 2.61835358e-03 -2.56811202e-01
-1.30347461e-01 8.13319460e-02 -9.68631878e-02 -3.94621998e-01
-1.72440183e+00 9.83195364e-01 -8.53030905e-02 1.65709347e-01
-6.33416951e-01 -2.46058419e-01 -6.92769229e-01 -1.28904626e-01
4.66171384e-01 -4.96473521e-01 1.72061503e+00 -8.05356264e-01
-1.43768275e+00 1.17323303e+00 -5.49455881e-01 -2.53939450e-01
2.44388372e-01 -4.18736666e-01 -5.51488936e-01 3.23421478e-01
6.85651422e-01 4.46506143e-01 3.31473798e-01 -7.99072325e-01
-7.09755361e-01 -6.36097610e-01 7.12070823e-01 3.23429227e-01
-5.32287806e-02 6.74900293e-01 -2.34498158e-01 -2.74420887e-01
7.85116553e-02 -8.95262718e-01 -6.80516735e-02 -5.15283823e-01
-1.67511582e-01 -1.03648150e+00 3.09570968e-01 -1.27196956e+00
1.55362546e+00 -1.72582507e+00 -8.19292441e-02 3.97746153e-02
2.45948613e-01 1.67213142e-01 -2.24758580e-01 9.39528227e-01
3.77003878e-01 1.54957592e-01 -1.47949755e-01 -2.50548050e-02
1.06186472e-01 9.64865237e-02 -3.48081797e-01 -1.24854807e-04
2.98712999e-01 8.88165474e-01 -1.24230278e+00 -4.52050477e-01
-5.79518020e-01 -5.55033609e-02 -3.25660676e-01 4.14263397e-01
-5.05363703e-01 7.01769114e-01 -5.91795862e-01 7.73478210e-01
4.79769409e-01 -4.04287964e-01 -5.05933724e-02 2.24956796e-01
-1.60250217e-01 7.96969175e-01 -9.74526465e-01 1.26977992e+00
-4.98904109e-01 5.54889679e-01 2.71172076e-01 -3.86631608e-01
9.82439578e-01 5.50796986e-01 1.73202567e-02 -1.08666074e+00
-3.01632024e-02 5.95106900e-01 9.95519906e-02 -9.72771406e-01
8.24236989e-01 1.78150788e-01 -3.17773938e-01 6.25765443e-01
-1.43190086e-01 -2.88947314e-01 6.89126074e-01 4.91442859e-01
1.16462159e+00 -2.67360121e-01 9.80865061e-02 -4.99845266e-01
1.19884074e+00 2.67422706e-01 1.49040632e-02 7.81015575e-01
-4.96995330e-01 5.97200751e-01 7.16624200e-01 -4.40478504e-01
-4.82257247e-01 -6.78733170e-01 7.14028478e-02 1.61629868e+00
-4.08799984e-02 -6.49022877e-01 -9.14055705e-01 -6.57275021e-01
-2.91151285e-01 6.11184001e-01 -6.49752498e-01 3.31853479e-01
-7.65212595e-01 -3.28209728e-01 5.59660435e-01 2.00522229e-01
6.82971776e-01 -1.25441015e+00 -2.15648189e-01 2.71289259e-01
-1.22557425e+00 -1.04433811e+00 -5.79460442e-01 -2.87664890e-01
-2.44180024e-01 -1.26214087e+00 -7.50314772e-01 -9.26127672e-01
5.24003282e-02 6.95184469e-02 1.69792604e+00 3.13379526e-01
4.71258610e-01 4.42845345e-01 -8.45004916e-01 -3.60510498e-01
-3.50234210e-01 2.48854861e-01 -4.08205748e-01 1.45716622e-01
5.75913846e-01 2.87737429e-01 -7.19386876e-01 6.58151448e-01
-6.97923124e-01 -5.13692915e-01 -1.69366063e-03 6.25951171e-01
2.25257277e-01 -6.32634044e-01 1.07389927e+00 -7.88124144e-01
1.24878645e+00 -7.98097849e-01 -1.84955180e-01 5.48043668e-01
-3.31728756e-01 -3.44097257e-01 6.11297548e-01 7.15685189e-02
-1.02062607e+00 -3.77705067e-01 -6.30991638e-01 6.28304482e-01
-1.43560220e-03 8.51613283e-01 -6.18936047e-02 9.19512510e-02
1.10581672e+00 -6.92178234e-02 -1.74865410e-01 -3.62968326e-01
4.24985409e-01 9.78139222e-01 4.09956664e-01 -7.75576115e-01
3.87061924e-01 1.77417681e-01 -5.75095713e-01 -6.77569270e-01
-1.21562517e+00 -1.01813090e+00 -3.26724559e-01 -4.13971007e-01
1.13173497e+00 -9.01628137e-01 -1.01201439e+00 5.45719266e-01
-1.40241730e+00 -1.20022044e-01 -4.64664446e-03 6.28563692e-04
-2.24848539e-01 3.61736417e-01 -8.67337406e-01 -8.02079618e-01
-6.77630663e-01 -1.07544518e+00 8.12811673e-01 3.96180421e-01
-5.81335485e-01 -8.45241725e-01 2.67781347e-01 1.23168182e+00
5.96979737e-01 -3.98429595e-02 9.16976511e-01 -9.80493426e-01
5.61416196e-03 -2.79772818e-01 -1.98577017e-01 2.46928185e-01
-1.52811050e-01 -1.81423634e-01 -1.00967038e+00 -1.06011964e-01
2.15786081e-02 -1.01992071e+00 3.99783164e-01 -2.12420315e-01
6.49952829e-01 -4.94928986e-01 3.73890072e-01 -7.56517470e-01
8.85105848e-01 3.33634554e-03 7.27098227e-01 4.65930730e-01
3.76065016e-01 1.10785890e+00 7.68638611e-01 3.70810419e-01
1.19752038e+00 6.21392667e-01 2.55366445e-01 4.29206192e-01
1.69735625e-01 -7.36207664e-02 3.36104333e-01 1.08905971e+00
7.61919767e-02 -2.55285114e-01 -1.28551829e+00 9.62121308e-01
-1.68880117e+00 -5.63210428e-01 -8.66529226e-01 2.02772760e+00
1.11143589e+00 -3.69264185e-01 2.65230626e-01 -2.85988264e-02
5.83128273e-01 -6.96219429e-02 -1.35990143e-01 -5.09439111e-01
-2.82946289e-01 4.89901125e-01 -3.34762990e-01 6.78289771e-01
-1.04439569e+00 8.29528868e-01 5.06675243e+00 7.48538613e-01
-3.42208743e-01 2.90616572e-01 6.30219936e-01 6.99558616e-01
-5.24536729e-01 2.73181915e-01 -6.40947402e-01 4.47096467e-01
1.28921211e+00 5.71917668e-02 2.39610463e-01 4.84897792e-01
5.53385913e-02 -4.33506519e-01 -8.47171545e-01 5.26275873e-01
5.77289835e-02 -1.14676082e+00 1.68790538e-02 -4.97145712e-01
7.83209205e-01 3.48087288e-02 -2.32124954e-01 7.47909427e-01
4.36663091e-01 -1.08931744e+00 8.41953039e-01 4.22122151e-01
3.66967142e-01 -4.34177101e-01 1.13909996e+00 5.82151651e-01
-9.78729308e-01 -5.79098426e-02 -1.16297178e-01 -4.42379653e-01
9.53920707e-02 1.42923459e-01 -3.95249456e-01 4.64083612e-01
1.14222682e+00 1.24484882e-01 -9.43069100e-01 7.76485443e-01
-7.57627070e-01 8.01932752e-01 -8.76120925e-02 -5.85811973e-01
4.76015419e-01 -2.17409834e-01 3.22394103e-01 1.15370011e+00
-2.11446434e-01 3.94273877e-01 8.13753679e-02 4.73558962e-01
-3.83509904e-01 5.68564653e-01 5.90843298e-02 9.45088416e-02
6.24152541e-01 1.22272229e+00 -4.46968466e-01 -3.59654695e-01
-3.18486661e-01 7.04289854e-01 3.10790718e-01 4.38257545e-01
-3.29527199e-01 -7.86601961e-01 -3.31177711e-02 -1.05716988e-01
-9.25771706e-03 1.03046887e-01 -1.61546111e-01 -1.15208495e+00
3.44547927e-01 -1.31025505e+00 8.81066144e-01 -8.80552351e-01
-1.49239886e+00 7.24436820e-01 -4.72789109e-01 -7.93265760e-01
-3.26285511e-01 -5.81620991e-01 -4.73910749e-01 1.31217265e+00
-1.52263069e+00 -9.31239307e-01 -2.89382339e-01 6.66867554e-01
4.94975179e-01 1.40461043e-01 1.01506424e+00 4.61032152e-01
-1.38534814e-01 4.14539337e-01 -1.41665280e-01 3.85431737e-01
9.56276596e-01 -1.35233355e+00 1.52525574e-01 4.81047988e-01
7.12913927e-03 7.65572846e-01 5.97375393e-01 -2.72033125e-01
-1.17948759e+00 -5.46625018e-01 2.05256224e+00 -1.04949260e+00
8.62506926e-01 -1.83814853e-01 -1.19237626e+00 2.96083719e-01
5.60641646e-01 -3.90154094e-01 9.06770349e-01 2.58284330e-01
-3.36204350e-01 1.61738276e-01 -1.17934215e+00 4.43394512e-01
1.47464707e-01 -9.98098552e-01 -7.55477607e-01 6.28100276e-01
8.12487602e-01 -4.02925223e-01 -1.00023627e+00 1.25246018e-01
2.97042370e-01 -6.73266590e-01 7.03816593e-01 -8.14196169e-01
6.25640988e-01 -4.67266381e-01 -4.03933525e-01 -9.88564610e-01
5.40299825e-02 -5.39908409e-01 6.14142092e-03 1.20496142e+00
8.07787120e-01 -4.79315311e-01 2.79296875e-01 7.53042519e-01
-1.30986765e-01 -8.02052498e-01 -9.79958594e-01 -1.17022432e-01
6.47581875e-01 -2.35030755e-01 4.57611054e-01 9.40200508e-01
-8.22065957e-03 8.97018552e-01 1.15673378e-01 5.90684824e-02
-1.55329213e-01 1.64204553e-01 5.84303081e-01 -1.13971174e+00
5.91908023e-02 -2.97180623e-01 3.95733267e-01 -1.04107690e+00
-1.96730234e-02 -7.78793633e-01 1.16572253e-01 -1.74603450e+00
1.02544259e-02 -1.58431023e-01 2.22949296e-01 2.46536836e-01
-5.25085568e-01 2.41913483e-01 5.88332750e-02 4.75030959e-01
-1.16968477e+00 4.12852377e-01 1.17756855e+00 -9.98158455e-02
-5.35796359e-02 2.11742267e-01 -8.32763910e-01 5.56601465e-01
7.08271682e-01 -2.48667106e-01 2.89239381e-02 -7.09065139e-01
1.05145836e+00 2.73692459e-01 2.54601777e-01 -3.22700977e-01
3.98225635e-01 -1.83377527e-02 -1.78191483e-01 -4.50072199e-01
6.51634531e-03 -2.84316123e-01 -7.17604518e-01 2.19324902e-01
-5.75741887e-01 4.83171701e-01 -4.03803736e-02 2.56378233e-01
-5.96143544e-01 -5.81534326e-01 3.47988963e-01 -2.81105399e-01
-5.84841967e-01 -1.41101018e-01 -6.28420234e-01 8.98652554e-01
4.48011875e-01 2.23637834e-01 -6.00489676e-01 -8.22363257e-01
-7.43824005e-01 9.08574760e-01 -1.72095209e-01 4.56062198e-01
4.79338825e-01 -1.22366726e+00 -1.23545110e+00 -4.99805450e-01
6.08331382e-01 -9.20778587e-02 4.46355015e-01 8.22635949e-01
-6.18670285e-01 6.06287360e-01 2.53783107e-01 -3.54611278e-01
-9.34024930e-01 2.63579711e-02 3.34341556e-01 -5.85687757e-01
3.66435289e-01 9.40675974e-01 -4.49495971e-01 -1.03345215e+00
3.44614834e-02 6.64732084e-02 -9.06530380e-01 5.67408323e-01
8.17030251e-01 6.93064272e-01 4.44081485e-01 -8.63177180e-01
-4.08981264e-01 3.56804341e-01 1.52041102e-02 -3.27990532e-01
7.03580618e-01 -4.11916763e-01 -7.97119021e-01 4.79408115e-01
9.86605704e-01 2.07046241e-01 -2.80368239e-01 -6.02871120e-01
6.07893288e-01 -1.30326807e-01 -5.50083458e-01 -1.37693453e+00
-3.69532019e-01 8.96343887e-01 2.41638631e-01 7.89536059e-01
8.48614872e-01 3.20071697e-01 8.93387973e-01 7.90560722e-01
2.85367310e-01 -1.30058682e+00 3.20742220e-01 1.21994007e+00
1.27196538e+00 -1.48705912e+00 -4.72294211e-01 -4.34321053e-02
-8.34480762e-01 1.01896358e+00 5.35773933e-01 1.93110541e-01
3.93962592e-01 -6.12995565e-01 5.70925117e-01 -5.46176195e-01
-6.59363270e-01 -4.48237598e-01 5.25038660e-01 2.44735181e-01
7.67572999e-01 2.08306178e-01 -7.81020463e-01 9.09428477e-01
-4.80943233e-01 -2.08633453e-01 5.41602612e-01 8.92419696e-01
-3.78611833e-01 -7.60392487e-01 -4.27029997e-01 1.82142764e-01
-8.30448985e-01 -8.27252120e-02 -8.22566986e-01 4.80086446e-01
-1.39113188e-01 2.05548811e+00 -2.42596611e-01 -2.33347729e-01
5.80703199e-01 9.39004347e-02 -8.70812684e-02 -6.39730155e-01
-1.33417845e+00 -3.89959782e-01 6.38629973e-01 -2.46272609e-01
-4.60833728e-01 -7.16110766e-01 -1.09285223e+00 -3.07157218e-01
-3.67697388e-01 6.94013059e-01 4.64701444e-01 1.31625807e+00
2.18207568e-01 -1.77306861e-01 5.51856399e-01 1.62641391e-01
-7.60751724e-01 -1.34901333e+00 -9.30306837e-02 6.24356508e-01
3.24292481e-01 -2.06883531e-02 -2.73482740e-01 -2.07508281e-01] | [11.406464576721191, 8.07149600982666] |
1bd9414c-eb0f-4990-b35b-7b84d675b7b7 | counterfactual-analysis-of-the-impact-of-the | 2202.09391 | null | https://arxiv.org/abs/2202.09391v1 | https://arxiv.org/pdf/2202.09391v1.pdf | Counterfactual Analysis of the Impact of the IMF Program on Child Poverty in the Global-South Region using Causal-Graphical Normalizing Flows | This work demonstrates the application of a particular branch of causal inference and deep learning models: \emph{causal-Graphical Normalizing Flows (c-GNFs)}. In a recent contribution, scholars showed that normalizing flows carry certain properties, making them particularly suitable for causal and counterfactual analysis. However, c-GNFs have only been tested in a simulated data setting and no contribution to date have evaluated the application of c-GNFs on large-scale real-world data. Focusing on the \emph{AI for social good}, our study provides a counterfactual analysis of the impact of the International Monetary Fund (IMF) program on child poverty using c-GNFs. The analysis relies on a large-scale real-world observational data: 1,941,734 children under the age of 18, cared for by 567,344 families residing in the 67 countries from the Global-South. While the primary objective of the IMF is to support governments in achieving economic stability, our results find that an IMF program reduces child poverty as a positive side-effect by about 1.2$\pm$0.24 degree (`0' equals no poverty and `7' is maximum poverty). Thus, our article shows how c-GNFs further the use of deep learning and causal inference in AI for social good. It shows how learning algorithms can be used for addressing the untapped potential for a significant social impact through counterfactual inference at population level (ACE), sub-population level (CACE), and individual level (ICE). In contrast to most works that model ACE or CACE but not ICE, c-GNFs enable personalization using \emph{`The First Law of Causal Inference'}. | ['Adel Daoud', 'Jose M. Peña', 'Sourabh Balgi'] | 2022-02-17 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [-4.13847715e-02 4.29121017e-01 -8.76102209e-01 -1.82419807e-01
-1.20182164e-01 -1.04208484e-01 7.83762693e-01 3.13232660e-01
-4.60076362e-01 1.18784165e+00 1.00127053e+00 -9.19484079e-01
-6.04510367e-01 -1.52769899e+00 -1.11882484e+00 -6.38767838e-01
-6.22274995e-01 2.20582634e-01 -5.20290911e-01 -2.12348804e-01
-1.61956679e-02 1.11235611e-01 -1.21340692e+00 1.67856187e-01
1.03596675e+00 5.88061139e-02 -3.28893691e-01 4.24011320e-01
1.59813762e-01 8.18947196e-01 -5.02531826e-01 -6.19461894e-01
8.27788636e-02 -4.87422973e-01 -6.41534030e-01 -8.46165895e-01
2.88343579e-01 -8.82198930e-01 -3.95849317e-01 9.43303227e-01
6.42231226e-01 -2.71076173e-01 8.53137493e-01 -1.48114145e+00
-8.48830223e-01 1.49914432e+00 -5.77962577e-01 3.77635181e-01
5.00064969e-01 1.36371300e-01 8.56075525e-01 -1.28645405e-01
4.12010670e-01 1.81955349e+00 8.64648044e-01 3.12467784e-01
-1.44421792e+00 -1.30987239e+00 3.06058489e-02 -2.82845981e-02
-7.41053343e-01 -8.13731700e-02 5.88662922e-01 -5.92730403e-01
8.97850513e-01 2.03851089e-01 8.69710743e-01 1.26289082e+00
1.92420110e-01 4.48833525e-01 1.11871552e+00 -5.54948568e-01
1.26728848e-01 -6.17857158e-01 6.80893958e-02 3.20159137e-01
6.69414222e-01 7.93847620e-01 -4.88954306e-01 -3.43903840e-01
8.83260369e-01 -1.97150350e-01 -1.62600614e-02 1.24347709e-01
-1.20311677e+00 1.31091583e+00 6.06839240e-01 2.71121562e-01
-4.94782120e-01 6.48968935e-01 3.81257325e-01 3.70784193e-01
6.80507064e-01 7.92367160e-02 -5.15778840e-01 -1.02003192e-04
-8.70693207e-01 7.27145910e-01 6.55946791e-01 1.57337666e-01
3.50033015e-01 2.28731140e-01 -9.92119238e-02 3.98010671e-01
3.35380524e-01 1.18460095e+00 -1.64797038e-01 -1.20786381e+00
7.88313866e-01 4.28289831e-01 1.06081918e-01 -1.10690534e+00
-8.39488089e-01 -1.78581849e-01 -1.12127554e+00 1.15074150e-01
9.37252641e-01 -7.56564498e-01 -4.69214678e-01 2.32164788e+00
1.55260295e-01 2.94452816e-01 -1.07448928e-01 4.97522354e-01
6.81575060e-01 5.60918033e-01 6.72753692e-01 -4.60069120e-01
9.81795907e-01 9.68239605e-02 -7.52241075e-01 1.68833107e-01
8.09990883e-01 -2.61108845e-01 7.33168542e-01 7.82203749e-02
-9.40351427e-01 -1.77652150e-01 -5.62355816e-01 4.56209332e-01
-4.71610725e-01 -4.48378623e-01 1.29781556e+00 9.49639320e-01
-9.93528068e-01 5.91007710e-01 -5.89140594e-01 -2.66202390e-01
7.73490191e-01 4.99467999e-01 -2.75542736e-01 7.86552131e-02
-1.79471421e+00 6.22961700e-01 3.06965351e-01 -3.15611631e-01
-8.40312421e-01 -1.43913972e+00 -8.40728402e-01 2.77925730e-01
-2.52564847e-02 -8.89033914e-01 6.63361371e-01 -1.04414892e+00
-8.10603797e-01 5.66937506e-01 1.23046577e-01 -7.31258869e-01
6.45243704e-01 -2.22860664e-01 -4.44885850e-01 -1.97188228e-01
5.67389607e-01 5.41853905e-01 1.85214952e-01 -9.04325068e-01
-6.90751791e-01 -6.59261465e-01 3.18141758e-01 -2.00223446e-01
-3.47897947e-01 4.70183998e-01 7.91478932e-01 -9.23554838e-01
-3.46072018e-01 -4.15328979e-01 1.56040881e-02 -7.44838893e-01
-3.47335517e-01 -5.14732361e-01 4.41514462e-01 -8.51113439e-01
1.45432150e+00 -1.58088624e+00 -2.73953170e-01 4.42042314e-02
-7.54726399e-03 1.28517300e-01 1.69231027e-01 5.40242791e-01
-6.58632159e-01 5.11410534e-01 -3.98813009e-01 2.56850183e-01
1.64612353e-01 -1.26142818e-02 -3.67582351e-01 5.13458192e-01
1.95696115e-01 8.76305282e-01 -1.19997418e+00 -1.98291883e-01
4.78737742e-01 3.78935009e-01 -8.78528297e-01 -2.61769712e-01
7.29744136e-02 2.07380921e-01 -2.56570298e-02 3.37485999e-01
8.00660133e-01 3.72393399e-01 3.58857274e-01 2.82108039e-01
-5.18539190e-01 3.62197310e-01 -9.33541000e-01 1.18953872e+00
-3.23027760e-01 5.61733067e-01 -9.38592628e-02 -1.42555654e+00
4.67581123e-01 3.95907938e-01 5.22725940e-01 -1.03997242e+00
2.11621244e-02 1.43460885e-01 4.39195096e-01 -5.06842077e-01
-9.75441113e-02 -4.70093548e-01 -1.62825003e-01 7.91252077e-01
-1.36500075e-01 1.07748881e-01 2.95783520e-01 1.76371053e-01
1.06143069e+00 2.29894206e-01 3.42535496e-01 -7.05011785e-01
7.35731050e-02 -2.82779843e-01 6.08153641e-01 8.44295919e-01
-1.19396724e-01 1.31736815e-01 1.11919403e+00 -6.37266040e-01
-9.96234715e-01 -1.31614208e+00 -4.14346278e-01 1.23651135e+00
-5.45341969e-01 3.45187038e-01 -8.95766795e-01 -6.65488303e-01
4.99620050e-01 1.18732846e+00 -9.10684228e-01 -7.41723459e-03
-8.86821747e-01 -1.52565503e+00 9.41600621e-01 4.39258248e-01
5.89846551e-01 -1.13108671e+00 -8.29134464e-01 1.35087341e-01
-1.68116227e-01 -2.77794749e-01 2.15461642e-01 -3.02000254e-01
-8.07289481e-01 -1.12882400e+00 -6.41449988e-01 -1.19523823e-01
1.40741915e-01 -2.64347166e-01 1.30440927e+00 -8.54300037e-02
1.82998210e-01 7.64416996e-03 -9.60676745e-03 -8.31235349e-01
-5.88310301e-01 -1.38984114e-01 3.21486264e-01 -4.80815947e-01
5.23643792e-01 -1.03081334e+00 -6.00043535e-01 -3.14756632e-01
-6.92743242e-01 3.33247110e-02 1.09858744e-01 6.66536450e-01
-2.82032073e-01 1.87339887e-01 1.35386908e+00 -8.99189115e-01
4.46246833e-01 -1.01202166e+00 -5.98035276e-01 -5.51602021e-02
-7.44497716e-01 -2.07951203e-01 2.77812809e-01 -2.57374167e-01
-1.40503001e+00 -6.59822166e-01 -1.29741803e-01 3.70841771e-01
-2.40635797e-01 6.10688508e-01 -3.18076722e-02 9.03467298e-01
6.53750300e-01 -6.13395631e-01 -2.59416282e-01 -1.85107991e-01
3.50119352e-01 4.70562339e-01 6.56189501e-01 -7.27135181e-01
5.37242353e-01 5.85465312e-01 1.81955457e-01 -4.82741416e-01
-4.81783867e-01 1.61942258e-01 -5.87939501e-01 -2.13878497e-01
9.18278694e-01 -1.09346592e+00 -1.17714238e+00 4.41424400e-01
-1.09190297e+00 -7.30923116e-01 -1.87712684e-01 7.93312192e-01
-4.26282823e-01 -2.38119513e-01 -4.43062305e-01 -1.13829136e+00
-3.40207428e-01 -4.82960790e-01 5.14933944e-01 1.29420742e-01
-3.61906081e-01 -1.39047348e+00 2.15033934e-01 1.67071477e-01
2.58070230e-01 9.53805745e-01 1.36508334e+00 -2.13605613e-01
7.38646090e-02 2.59689122e-01 -5.21340251e-01 -9.80508178e-02
1.88602090e-01 2.21525013e-01 -8.29827845e-01 -1.82957888e-01
-3.67895871e-01 9.23446193e-03 8.84475470e-01 1.14354801e+00
7.43597209e-01 -6.45342112e-01 -2.79182553e-01 3.17728430e-01
1.51423395e+00 3.22094113e-01 7.29966462e-01 3.13885450e-01
6.40461445e-01 9.28452194e-01 2.16641158e-01 5.32376766e-01
7.22092867e-01 1.98342741e-01 7.03847885e-01 -2.51320392e-01
-2.53077358e-01 -6.42044663e-01 4.38891321e-01 4.53418344e-02
-4.95218277e-01 -8.26192945e-02 -1.13437843e+00 9.64344382e-01
-1.67350936e+00 -1.43793595e+00 -6.42487943e-01 2.25492787e+00
8.18976402e-01 1.18993714e-01 4.44476336e-01 2.82492936e-01
7.48898804e-01 7.18728155e-02 -2.48678148e-01 -6.62134409e-01
-2.80198753e-01 5.23220003e-01 8.30535948e-01 6.32763743e-01
-1.01045036e+00 5.57608843e-01 6.41744232e+00 4.96236712e-01
-7.76429474e-01 1.52271211e-01 9.06742692e-01 -1.74211502e-01
-7.58468390e-01 1.34952858e-01 -3.82343054e-01 7.45493293e-01
1.18880522e+00 -2.77869016e-01 2.34188959e-01 1.55869111e-01
8.87973189e-01 -2.77859688e-01 -9.54192460e-01 3.15048188e-01
-5.59024990e-01 -1.43300152e+00 1.83983669e-02 3.36570561e-01
1.17329931e+00 -5.78138940e-02 2.08040461e-01 2.75020927e-01
1.06792486e+00 -1.20762610e+00 8.82483602e-01 4.65621084e-01
1.00252414e+00 -1.13480103e+00 9.08319712e-01 2.70851552e-01
-4.62110639e-01 -6.16260529e-01 -1.55386925e-01 -9.44985092e-01
1.07463859e-01 9.67440486e-01 -5.13064086e-01 6.83573484e-01
1.00056541e+00 5.44929743e-01 -8.32857285e-03 3.37876290e-01
-4.31586295e-01 1.19386637e+00 -3.13900173e-01 3.87740582e-01
2.23213404e-01 7.89233856e-03 2.79166788e-01 1.32814264e+00
3.00322562e-01 2.75444567e-01 -6.90211773e-01 1.10786331e+00
-2.25929067e-01 -7.02603757e-02 -8.79197896e-01 1.88412026e-01
4.83643770e-01 3.88343006e-01 -3.37770700e-01 -1.67531744e-01
-4.09918755e-01 -1.07387125e-01 1.24330945e-01 4.28364277e-01
-8.32844853e-01 -3.42063829e-02 6.47625208e-01 3.96679372e-01
-3.21289241e-01 3.50918263e-01 -6.55574262e-01 -8.38996232e-01
-5.71389616e-01 -7.84685850e-01 6.07679129e-01 -3.59196335e-01
-1.05999327e+00 -8.30007732e-01 5.73282957e-01 -3.23992759e-01
-4.01174009e-01 -2.91226476e-01 -8.97606194e-01 9.85953927e-01
-1.26863027e+00 -1.18181860e+00 5.25267959e-01 3.27811271e-01
4.80827019e-02 5.53338192e-02 6.27425134e-01 4.27586675e-01
-6.03429854e-01 3.99874806e-01 1.02336518e-01 2.59952962e-01
3.56250048e-01 -1.33097410e+00 4.52116907e-01 9.81635988e-01
-4.12223816e-01 6.68512762e-01 7.35887587e-01 -1.00156045e+00
-7.07542658e-01 -8.22678387e-01 1.34708393e+00 -3.27327371e-01
7.69202650e-01 -2.57889569e-01 -4.83438402e-01 8.31200600e-01
3.32374305e-01 -3.92352313e-01 5.16009450e-01 5.00179589e-01
-3.62383187e-01 -1.44345134e-01 -1.41827500e+00 5.17163813e-01
1.24142981e+00 -1.39181554e-01 -6.84868455e-01 2.04523742e-01
9.54861581e-01 2.56052017e-01 -1.09207070e+00 6.13837659e-01
8.14437687e-01 -1.34073842e+00 1.10617626e+00 -7.33081520e-01
1.07021379e+00 2.90748388e-01 -1.12030178e-01 -1.30811059e+00
-2.53431976e-01 -4.00016308e-01 4.38182838e-02 1.55656052e+00
1.89901575e-01 -8.37482154e-01 5.25382876e-01 4.94011998e-01
1.62995666e-01 -3.40168834e-01 -1.23131442e+00 -4.10716832e-01
1.03067577e+00 -7.09475279e-01 1.35343385e+00 1.56281793e+00
-7.79902115e-02 -8.79600123e-02 -2.34991744e-01 2.46337458e-01
8.86762917e-01 -2.05519870e-01 6.48457646e-01 -1.30137801e+00
8.08457732e-02 -6.91485703e-01 1.35560825e-01 8.77644420e-02
4.16394114e-01 -7.36772776e-01 -8.29622507e-01 -1.61234188e+00
4.98171836e-01 -6.08865857e-01 -2.44575739e-01 5.87948263e-01
-2.40458399e-01 -2.54142016e-01 2.26880640e-01 -2.98453659e-01
3.99343252e-01 2.01106921e-01 9.39558804e-01 -1.71407953e-01
-1.25239030e-01 3.96934859e-02 -7.49746501e-01 8.81480992e-01
8.11924100e-01 -5.82872570e-01 -2.20260412e-01 -4.86055076e-01
6.54713809e-01 4.03875977e-01 8.45271587e-01 -4.87559408e-01
-2.75661677e-01 -5.86789072e-01 4.70420122e-01 -4.62133110e-01
-5.66639185e-01 -5.17033756e-01 3.72732073e-01 1.03066099e+00
-3.14962268e-01 -3.72469760e-02 2.73774356e-01 9.72682387e-02
1.79282203e-01 1.28981441e-01 3.61357361e-01 -2.43043080e-02
1.87657902e-03 -8.53737369e-02 -1.71059385e-01 5.32675572e-02
6.23244882e-01 2.89195389e-01 -7.76934981e-01 -4.21676129e-01
-2.71134526e-01 2.97896922e-01 1.53921396e-01 2.40819871e-01
1.41189888e-01 -1.37738407e+00 -1.33534586e+00 7.43255857e-03
-4.67942894e-01 -2.92349607e-01 2.03456417e-01 6.57582104e-01
-3.98985356e-01 6.56702995e-01 -1.64854065e-01 -4.60170135e-02
-6.01682961e-01 5.64654827e-01 1.72877118e-01 -3.98105353e-01
-3.88407588e-01 6.56879544e-01 4.35073376e-01 -6.02477491e-01
-2.02106982e-01 -2.43385226e-01 -2.29328096e-01 4.26132500e-01
5.73555827e-01 1.14272308e+00 -5.04380047e-01 -5.59788704e-01
-1.29206106e-01 1.15740813e-01 5.51544547e-01 -2.15521574e-01
1.60611081e+00 -1.57213584e-01 -4.55819756e-01 3.26372057e-01
9.37485993e-01 -9.44977552e-02 -1.20152199e+00 4.31985736e-01
-1.67717695e-01 -3.18682462e-01 -8.46379027e-02 -9.69640136e-01
-1.03564584e+00 7.98570871e-01 6.68208480e-01 4.98593986e-01
1.14239168e+00 -1.52047828e-01 2.44509399e-01 -3.02888274e-01
2.42548600e-01 -8.75132322e-01 -5.48088908e-01 1.07038699e-01
8.25782895e-01 -1.01625466e+00 2.01275036e-01 7.62937311e-03
2.23358110e-01 6.14375532e-01 2.36672848e-01 -2.21934661e-01
6.26989126e-01 1.19310573e-01 -2.95712709e-01 -9.79856327e-02
-5.11962533e-01 4.89916727e-02 -2.80507445e-01 8.83186042e-01
6.40767038e-01 8.32487166e-01 -8.65354598e-01 4.16002661e-01
-7.12133288e-01 1.21089488e-01 5.79157472e-01 2.35958606e-01
-6.51469221e-03 -7.45145559e-01 -8.22080135e-01 7.73833036e-01
-1.02002597e+00 -3.14417273e-01 1.02281503e-01 1.32571876e+00
6.67165995e-01 1.11446202e+00 4.08551514e-01 1.36067897e-01
1.56667888e-01 6.02768771e-02 3.54359388e-01 -1.36193201e-01
-6.28942668e-01 -3.04669261e-01 1.88714176e-01 -3.82628679e-01
-9.09772456e-01 -1.13146150e+00 -1.27109683e+00 -1.09007168e+00
9.07773897e-02 -2.06143752e-01 3.64470214e-01 9.13795829e-01
-2.96013504e-01 4.98725116e-01 5.48453212e-01 -5.79453766e-01
-1.37416750e-01 -1.08403623e+00 -5.84745765e-01 3.42374563e-01
4.12092239e-01 -7.39825964e-01 -6.15964055e-01 -2.53869683e-01] | [8.114216804504395, 5.408175468444824] |
f2abafc3-d788-489b-82c3-902b58667efc | qualitative-analysis-of-a-graph-transformer | 2301.10871 | null | https://arxiv.org/abs/2301.10871v3 | https://arxiv.org/pdf/2301.10871v3.pdf | Qualitative Analysis of a Graph Transformer Approach to Addressing Hate Speech: Adapting to Dynamically Changing Content | Our work advances an approach for predicting hate speech in social media, drawing out the critical need to consider the discussions that follow a post to successfully detect when hateful discourse may arise. Using graph transformer networks, coupled with modelling attention and BERT-level natural language processing, our approach can capture context and anticipate upcoming anti-social behaviour. In this paper, we offer a detailed qualitative analysis of this solution for hate speech detection in social networks, leading to insights into where the method has the most impressive outcomes in comparison with competitors and identifying scenarios where there are challenges to achieving ideal performance. Included is an exploration of the kinds of posts that permeate social media today, including the use of hateful images. This suggests avenues for extending our model to be more comprehensive. A key insight is that the focus on reasoning about the concept of context positions us well to be able to support multi-modal analysis of online posts. We conclude with a reflection on how the problem we are addressing relates especially well to the theme of dynamic change, a critical concern for all AI solutions for social impact. We also comment briefly on how mental health well-being can be advanced with our work, through curated content attuned to the extent of hate in posts. | ['Lukasz Golab', 'Robin Cohen', 'Hong Yi Chen', 'Liam Hebert'] | 2023-01-25 | null | null | null | null | ['hate-speech-detection'] | ['natural-language-processing'] | [ 3.22003603e-01 6.71962559e-01 -1.61932975e-01 8.89900252e-02
-2.54063785e-01 -5.79516172e-01 9.35761094e-01 6.34907186e-01
-2.25305289e-01 3.32968444e-01 1.13248599e+00 -3.89244527e-01
-2.41061822e-01 -4.59931552e-01 -3.91359878e-04 -3.22459131e-01
-1.93492204e-01 5.94372191e-02 -1.49614979e-02 -8.01740825e-01
5.75575888e-01 4.38892126e-01 -1.10482323e+00 4.42974150e-01
5.50373137e-01 1.35668233e-01 -1.97553828e-01 7.10722327e-01
-3.36060710e-02 1.70199001e+00 -1.02579510e+00 -8.56010258e-01
-4.49934751e-01 -4.64721024e-01 -1.19055367e+00 1.09663881e-01
3.17827016e-01 -4.70746040e-01 -2.57492840e-01 1.01601899e+00
5.52009881e-01 -9.46482867e-02 2.83366084e-01 -1.15852869e+00
-6.99887753e-01 7.00663269e-01 -2.97032982e-01 8.80780041e-01
7.29248047e-01 3.01291615e-01 8.24161232e-01 -2.98386902e-01
1.03384268e+00 1.44691300e+00 8.90093267e-01 5.66286683e-01
-9.07345057e-01 -5.37263811e-01 4.87617776e-02 1.90922543e-01
-8.78751755e-01 -7.58777082e-01 8.37857783e-01 -8.66250575e-01
1.06947184e+00 4.41430092e-01 8.94695342e-01 1.48720324e+00
4.70088050e-02 5.28452873e-01 9.70086157e-01 -3.47704142e-01
-1.45913899e-01 2.13590369e-01 1.20003343e-01 6.49557054e-01
6.36808155e-03 -4.67148840e-01 -6.23849571e-01 -7.15383470e-01
9.20098573e-02 8.53626505e-02 -5.32843079e-03 3.88721973e-01
-8.69986713e-01 1.22608876e+00 3.00099790e-01 9.41719174e-01
-2.52775013e-01 -2.61510581e-01 7.98049092e-01 2.20402312e-02
1.10089397e+00 7.46512473e-01 3.00730497e-01 -5.88032544e-01
-1.00171113e+00 2.37475306e-01 9.12292480e-01 1.61706030e-01
2.83615798e-01 -2.56098866e-01 -2.12007418e-01 8.68338287e-01
1.44718990e-01 2.17959270e-01 -2.74152476e-02 -9.41200137e-01
4.19858873e-01 5.43043911e-01 -1.42063990e-01 -1.75651717e+00
-5.63934863e-01 -2.14713842e-01 -4.03370857e-01 -3.51393044e-01
3.57128054e-01 -4.63678211e-01 -3.65523219e-01 1.50516856e+00
1.20768219e-01 9.37573463e-02 -5.64750195e-01 4.40622717e-01
5.38815141e-01 5.28010964e-01 3.49646062e-01 -5.73669195e-01
1.41188204e+00 -3.69390517e-01 -1.02067053e+00 -4.72496986e-01
8.90227079e-01 -7.62168229e-01 7.13817656e-01 2.18321696e-01
-1.06376696e+00 3.82918954e-01 -7.63484716e-01 -1.81709588e-01
-6.73226535e-01 -9.21891093e-01 4.02776212e-01 7.44830906e-01
-1.14344144e+00 6.70841992e-01 -5.33950627e-01 -1.21904516e+00
5.29963851e-01 -9.15218070e-02 -7.64373764e-02 4.21178453e-02
-1.29036009e+00 1.33895135e+00 -5.14774211e-02 7.73884654e-02
-3.69378954e-01 -4.16444182e-01 -7.12613702e-01 -4.28241938e-01
6.04729176e-01 -1.81518212e-01 9.95238006e-01 -1.04465950e+00
-8.14614773e-01 1.10281003e+00 -1.46981299e-01 -3.86986345e-01
3.27084720e-01 1.25707379e-02 -6.38003170e-01 4.43270355e-01
2.18165010e-01 1.52680144e-01 7.24110782e-01 -9.23633039e-01
-2.98688978e-01 -4.64433312e-01 4.65920031e-01 7.64370989e-03
-9.65513706e-01 1.23044598e+00 3.55670005e-01 -5.33379436e-01
-4.99060124e-01 -7.54595101e-01 -2.09593568e-02 -4.44836944e-01
-5.17786562e-01 -3.24561924e-01 1.19557238e+00 -1.07285631e+00
1.94337881e+00 -1.98024249e+00 -1.36991953e-02 4.19383645e-02
8.26802075e-01 3.83591294e-01 3.75118405e-01 1.52409399e+00
1.48972318e-01 7.86537588e-01 -4.41820845e-02 -3.43217313e-01
-8.76056626e-02 -2.95754820e-02 -3.64170581e-01 7.53138661e-01
3.53250742e-01 6.53753519e-01 -1.34885919e+00 -4.20012295e-01
-5.82010066e-03 7.90022612e-01 -4.70567942e-01 -2.29016636e-02
8.97019133e-02 2.40299180e-01 -3.19171965e-01 4.12398517e-01
5.64025640e-02 -2.47694701e-01 3.49837065e-01 3.10038418e-01
-4.81730372e-01 4.83170599e-01 -2.83857971e-01 7.57129729e-01
1.67454436e-01 1.21899271e+00 4.80420113e-01 -5.59199572e-01
6.07629418e-01 2.58256137e-01 3.44333977e-01 -3.78967464e-01
3.46146345e-01 -1.38630927e-01 1.51096895e-01 -8.70222509e-01
6.32635236e-01 -2.93369800e-01 1.88478101e-02 6.42678738e-01
-4.10328895e-01 -7.59912878e-02 -4.19149734e-03 5.72038889e-01
1.42445755e+00 -3.22910935e-01 3.65157783e-01 -3.78264248e-01
3.17358404e-01 1.28094897e-01 1.66127399e-01 5.22391617e-01
-7.54976392e-01 9.06436145e-02 1.15999389e+00 -4.93077606e-01
-1.11887693e+00 -2.67733812e-01 1.44208327e-01 1.32420003e+00
-3.34736973e-01 -8.99043560e-01 -1.05910432e+00 -4.96912301e-01
-3.24725509e-01 8.60037208e-01 -8.99284482e-01 -3.53628583e-02
-6.21938944e-01 -7.50907421e-01 6.72687471e-01 -1.82431206e-01
1.04296066e-01 -1.34756219e+00 -7.56240666e-01 1.28935695e-01
-2.78464645e-01 -9.52914357e-01 -1.42521295e-03 -3.46231967e-01
-3.67559969e-01 -1.27808297e+00 -4.11991477e-01 -4.71673042e-01
3.73521030e-01 3.37702662e-01 5.96015453e-01 6.76045954e-01
-1.99744627e-01 5.71538329e-01 -5.51918745e-01 -4.95605260e-01
-1.00146878e+00 -3.05685820e-03 -1.41120091e-01 -1.04669154e-01
6.25091851e-01 -6.52732790e-01 -4.69217330e-01 -2.25783616e-01
-1.07931817e+00 1.26197129e-01 -1.62011608e-01 3.38688403e-01
-5.51685035e-01 -1.77447405e-02 4.11281109e-01 -1.39016485e+00
1.19305563e+00 -1.39517617e+00 4.33804572e-01 1.40808625e-02
-4.61064368e-01 -7.04164505e-01 5.21541715e-01 -2.18405545e-01
-6.59451008e-01 -7.89596200e-01 2.75203884e-02 -7.39549696e-02
-6.28000945e-02 4.62621540e-01 5.91229379e-01 2.76426286e-01
9.03571963e-01 -3.66472006e-01 3.95545065e-01 -2.60467291e-01
1.93419382e-01 8.93998921e-01 2.95018740e-02 -1.63677916e-01
7.86854208e-01 6.09492302e-01 -3.33605677e-01 -1.44392478e+00
-8.38541150e-01 -5.92220902e-01 -5.04317820e-01 -7.22283781e-01
9.76989329e-01 -4.76661325e-01 -8.41535985e-01 5.84697187e-01
-1.23916876e+00 -5.07098138e-01 1.66640297e-01 -1.62473202e-01
-2.12006971e-01 5.51652491e-01 -1.08430851e+00 -1.37205982e+00
-3.09933066e-01 -6.77196145e-01 6.20443761e-01 -5.92646189e-02
-8.23137701e-01 -1.52928782e+00 1.83936357e-01 8.19752276e-01
6.59602582e-01 8.67579937e-01 8.77704024e-01 -1.04950070e+00
2.42173463e-01 8.15772489e-02 -1.60663202e-01 -5.80067672e-02
1.82201222e-01 3.60493630e-01 -1.04568756e+00 -3.90537143e-01
1.20136306e-01 -4.81190115e-01 1.85510218e-01 -1.57870218e-01
4.83865350e-01 -1.00110030e+00 -2.99006283e-01 -3.24871570e-01
1.22045970e+00 -8.98623560e-03 5.24670839e-01 5.54831207e-01
8.06613088e-01 1.26754415e+00 2.83657253e-01 7.55355656e-01
5.44885397e-01 3.27528656e-01 3.77056777e-01 7.48406351e-02
4.04709466e-02 -4.06864345e-01 5.59911370e-01 8.88595402e-01
-4.25570235e-02 -3.19683492e-01 -1.49872351e+00 7.31574535e-01
-1.82616007e+00 -1.42892206e+00 -3.21152925e-01 1.70425951e+00
4.91019696e-01 2.73322016e-01 8.74812901e-01 6.23049177e-02
9.70062375e-01 8.25019360e-01 -1.15826860e-01 -9.94643033e-01
1.99034944e-01 -9.80512574e-02 2.61671066e-01 8.33717227e-01
-7.67403185e-01 8.58792484e-01 6.87678909e+00 5.01386523e-01
-8.85754764e-01 4.23062176e-01 7.94352233e-01 -2.21008182e-01
-3.18808585e-01 -4.67494968e-03 -2.24199399e-01 6.05377793e-01
1.31141889e+00 -1.13623723e-01 7.62643635e-01 1.45853609e-01
6.21722400e-01 -6.10333607e-02 -6.35048211e-01 5.28371394e-01
6.25735581e-01 -1.31696510e+00 -5.33573270e-01 2.72576004e-01
4.07690674e-01 1.26885295e-01 2.83752903e-02 -5.21180928e-02
2.79689938e-01 -1.10438740e+00 5.61960697e-01 3.30015540e-01
3.81096125e-01 -6.70706093e-01 4.28339899e-01 5.34829378e-01
-4.38176155e-01 -3.73716027e-01 8.34688097e-02 -6.88354492e-01
1.10760845e-01 2.61999249e-01 -1.08933616e+00 -1.28997564e-01
5.74704528e-01 8.84257674e-01 -7.05955982e-01 4.94042188e-01
-2.20155358e-01 8.17373157e-01 -4.99163428e-03 -2.53706366e-01
4.49569166e-01 1.67245209e-01 8.23604584e-01 1.62300086e+00
-1.74835607e-01 3.65993708e-01 -9.55028161e-02 4.77617621e-01
3.44414830e-01 1.54128522e-01 -1.20196259e+00 -7.36865997e-01
6.50110245e-01 1.16131401e+00 -7.12632179e-01 -3.70172113e-02
-4.14882272e-01 6.57540321e-01 6.37229264e-01 4.76418808e-02
-4.06051040e-01 -8.57442394e-02 5.17742872e-01 7.35111237e-01
-3.05588782e-01 -1.35533974e-01 2.44149044e-02 -7.39588976e-01
-4.76023644e-01 -1.18273449e+00 4.35256422e-01 -7.94777572e-01
-1.26187003e+00 4.35634732e-01 2.02775210e-01 -3.02591711e-01
-3.89014065e-01 -2.09281623e-01 -7.96337664e-01 4.35620844e-01
-9.58592653e-01 -1.23237360e+00 9.95124280e-02 2.29553580e-01
2.94151515e-01 3.40825737e-01 6.68835223e-01 1.64233789e-01
-8.51056039e-01 2.31793910e-01 -3.21943313e-01 2.28249550e-01
5.60317814e-01 -8.07424963e-01 4.19966578e-01 9.01554167e-01
-3.85477960e-01 5.59596956e-01 1.13962913e+00 -9.69784975e-01
-1.10121381e+00 -4.99620080e-01 1.25021207e+00 -9.13674653e-01
1.59124887e+00 -4.60172921e-01 -8.42616796e-01 7.63781726e-01
7.58179665e-01 -5.96185803e-01 9.66667950e-01 3.17593634e-01
-2.04447225e-01 6.31327689e-01 -1.34497583e+00 8.13385427e-01
1.38888776e+00 -8.65405142e-01 -7.54601836e-01 7.53965378e-01
6.45665765e-01 -1.43625364e-01 -7.88481236e-01 -3.37528467e-01
2.95557290e-01 -1.24239755e+00 4.46990490e-01 -9.59605575e-01
8.86871457e-01 4.41268265e-01 2.47832552e-01 -9.87036467e-01
-6.31451309e-01 -1.23497057e+00 -7.50843510e-02 1.56952047e+00
-3.30748707e-02 -6.25688672e-01 5.28755665e-01 9.51329470e-01
1.75159529e-01 -6.38070285e-01 -6.42839611e-01 -7.70385265e-02
1.23156823e-01 -3.76149386e-01 2.23885119e-01 1.66937292e+00
6.33259356e-01 3.71915758e-01 -6.30023062e-01 -4.83937524e-02
2.98928171e-01 -7.64130116e-01 4.50263470e-01 -1.20954549e+00
4.01277952e-02 -5.53324938e-01 -5.44342935e-01 -1.30237415e-02
1.51436314e-01 -5.61508954e-01 -4.57316250e-01 -1.57394254e+00
6.37800753e-01 1.42124161e-01 1.48097888e-01 4.22509044e-01
2.49239858e-02 5.40632784e-01 6.80194378e-01 3.23213041e-01
-6.22937977e-01 -2.17267379e-01 1.12573600e+00 1.07694626e-01
-2.31839105e-01 -5.44028342e-01 -1.18612659e+00 9.31858122e-01
7.87365973e-01 -4.50689942e-01 -4.04807925e-01 -8.39056000e-02
8.24414551e-01 -2.07969397e-02 4.12884831e-01 -5.44350922e-01
1.82954177e-01 -4.86198366e-01 -1.62439048e-01 -1.74350336e-01
4.21298862e-01 -5.67777872e-01 7.74413198e-02 4.94421721e-01
-5.21011770e-01 1.19654983e-01 1.83083296e-01 3.40301663e-01
2.16179848e-01 -3.47235978e-01 5.18813252e-01 -1.33413970e-01
-3.27229917e-01 -9.42351576e-03 -9.25048470e-01 4.26229388e-01
1.00882244e+00 -3.15426648e-01 -9.42618191e-01 -9.01348650e-01
-9.28164363e-01 1.32009804e-01 7.50596941e-01 4.61707324e-01
2.88183242e-01 -8.87707651e-01 -8.40730488e-01 -2.09744260e-01
5.77073246e-02 -7.22518146e-01 1.14022739e-01 8.29488099e-01
-3.50019813e-01 5.06777801e-02 -7.20507652e-02 1.24355629e-01
-1.36467886e+00 5.86110651e-01 1.10428281e-01 -7.92473182e-02
-6.01694822e-01 5.48326194e-01 -2.50375330e-01 9.37435776e-02
9.49777383e-03 4.16709930e-01 -6.74905837e-01 5.71633339e-01
9.29626465e-01 6.31077409e-01 -3.80171448e-01 -1.18525040e+00
-5.43748438e-01 -2.28735302e-02 -9.77782160e-02 1.40300900e-01
1.69763970e+00 -3.95717502e-01 -7.58953094e-01 5.68593025e-01
1.31052148e+00 4.85984713e-01 -7.73667157e-01 4.83397618e-02
2.61817187e-01 -5.02618432e-01 -1.60328597e-01 -8.14121306e-01
-4.42826390e-01 7.04729259e-01 -2.60593388e-02 1.16563714e+00
7.29351342e-01 2.62589991e-01 8.38686466e-01 1.71737000e-02
-1.38299027e-02 -1.14731741e+00 2.56925076e-01 5.08457184e-01
9.80477154e-01 -9.34972465e-01 3.47841561e-01 -4.41763192e-01
-5.97338855e-01 1.20966816e+00 3.60038429e-01 -1.58780273e-02
5.91519713e-01 1.17810480e-02 -4.89519611e-02 -9.21941459e-01
-6.02614701e-01 1.48375686e-02 -2.05796346e-01 6.25339627e-01
6.54990613e-01 3.58466208e-02 -4.23867822e-01 -6.86008260e-02
-3.71184528e-01 -4.96483535e-01 8.99254203e-01 8.93067598e-01
-6.13093257e-01 -8.43071401e-01 -5.54845929e-01 5.03201246e-01
-1.04790115e+00 -1.23563938e-01 -1.34984756e+00 8.08852732e-01
1.50736615e-01 1.40816891e+00 -4.01396640e-02 -5.71519673e-01
7.23526254e-02 1.62183210e-01 1.41450211e-01 -9.00428355e-01
-1.08414614e+00 -1.93697616e-01 8.19876373e-01 -3.65837544e-01
-5.26650965e-01 -7.93929577e-01 -6.06260002e-01 -1.15284288e+00
-2.22147748e-01 3.06428280e-02 4.40498710e-01 1.07390761e+00
3.82937074e-01 1.38025165e-01 5.16873062e-01 -6.42176986e-01
-1.16716726e-02 -8.63546491e-01 -2.27797955e-01 4.31489438e-01
6.42477036e-01 -3.48485738e-01 -5.57630956e-01 -4.49677229e-01] | [8.678915977478027, 10.440394401550293] |
2f041f04-aec0-4ad8-8df6-64e9d16f6d9b | hohonet-360-indoor-holistic-understanding | 2011.11498 | null | https://arxiv.org/abs/2011.11498v3 | https://arxiv.org/pdf/2011.11498v3.pdf | HoHoNet: 360 Indoor Holistic Understanding with Latent Horizontal Features | We present HoHoNet, a versatile and efficient framework for holistic understanding of an indoor 360-degree panorama using a Latent Horizontal Feature (LHFeat). The compact LHFeat flattens the features along the vertical direction and has shown success in modeling per-column modality for room layout reconstruction. HoHoNet advances in two important aspects. First, the deep architecture is redesigned to run faster with improved accuracy. Second, we propose a novel horizon-to-dense module, which relaxes the per-column output shape constraint, allowing per-pixel dense prediction from LHFeat. HoHoNet is fast: It runs at 52 FPS and 110 FPS with ResNet-50 and ResNet-34 backbones respectively, for modeling dense modalities from a high-resolution $512 \times 1024$ panorama. HoHoNet is also accurate. On the tasks of layout estimation and semantic segmentation, HoHoNet achieves results on par with current state-of-the-art. On dense depth estimation, HoHoNet outperforms all the prior arts by a large margin. | ['Hwann-Tzong Chen', 'Min Sun', 'Cheng Sun'] | 2020-11-23 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Sun_HoHoNet_360_Indoor_Holistic_Understanding_With_Latent_Horizontal_Features_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Sun_HoHoNet_360_Indoor_Holistic_Understanding_With_Latent_Horizontal_Features_CVPR_2021_paper.pdf | cvpr-2021-1 | ['3d-room-layouts-from-a-single-rgb-panorama'] | ['computer-vision'] | [ 3.55351150e-01 2.20444277e-01 2.25555345e-01 -4.27300513e-01
-7.78311968e-01 -3.39470148e-01 1.29904926e-01 -1.31786108e-01
-3.09290481e-03 2.58177638e-01 4.77299958e-01 -3.04708958e-01
2.69346666e-02 -1.03469753e+00 -9.17522609e-01 -2.97723413e-01
2.69383758e-01 4.33168888e-01 3.30937058e-01 3.11250817e-02
6.95395917e-02 4.45086330e-01 -1.74000299e+00 3.81132126e-01
6.62060678e-01 1.25861990e+00 4.55175519e-01 1.13839412e+00
2.93564856e-01 9.46919799e-01 -5.45637347e-02 -1.01682007e-01
5.25448024e-01 2.39499006e-02 -8.13611567e-01 2.07686260e-01
1.43012273e+00 -7.18187630e-01 -6.20514750e-01 3.06802005e-01
6.97958291e-01 2.02910870e-01 1.99765474e-01 -6.97957516e-01
5.21889143e-02 2.77633011e-01 -5.26903093e-01 -1.73662692e-01
3.22240442e-01 7.47769251e-02 1.14056826e+00 -8.29508543e-01
5.76891184e-01 1.40956271e+00 8.79010558e-01 3.11350673e-01
-8.41784000e-01 -4.69177186e-01 3.94013673e-01 -1.02252038e-02
-1.35515726e+00 -3.81347448e-01 4.28910315e-01 -1.38249889e-01
1.32458234e+00 6.59223318e-01 7.55439937e-01 7.32847154e-01
1.42023414e-01 1.05874670e+00 1.23993838e+00 -3.29264075e-01
1.88951775e-01 -3.71077389e-01 -2.01736540e-02 1.16554034e+00
-2.41262943e-01 -5.28580099e-02 -5.29500842e-01 3.20382982e-01
1.03219640e+00 1.49813026e-01 -3.20743412e-01 -3.39667410e-01
-1.10028601e+00 4.43446338e-01 8.00692737e-01 -1.47734448e-01
1.27917051e-01 4.62031662e-01 6.16059080e-02 -3.74056280e-01
2.38320157e-01 3.86366487e-01 -6.41708314e-01 -1.61618158e-01
-1.25296056e+00 3.83580267e-01 5.13620794e-01 1.17484546e+00
7.50896394e-01 -1.78812504e-01 -1.84567854e-01 7.79953897e-01
2.86235452e-01 7.03151941e-01 -7.27867931e-02 -1.59368920e+00
7.24299729e-01 4.52569395e-01 -6.73376024e-02 -6.74900830e-01
-7.46245921e-01 -4.86264616e-01 -7.63529658e-01 -3.01613547e-02
2.53170311e-01 1.04151160e-01 -1.44319606e+00 1.12887418e+00
3.85277450e-01 5.54027082e-03 -5.00226200e-01 7.46382773e-01
8.77179503e-01 8.62387896e-01 -2.67746598e-01 4.11659479e-01
1.31742990e+00 -1.73582101e+00 -4.11584586e-01 -7.08721399e-01
3.53943288e-01 -7.58835196e-01 1.36579132e+00 7.77484715e-01
-1.12575817e+00 -6.13098264e-01 -1.10517216e+00 -8.42519104e-01
-2.67903924e-01 2.05303520e-01 7.90158272e-01 7.88986981e-01
-1.08286035e+00 4.75472212e-01 -9.89113688e-01 -3.46875817e-01
6.62002921e-01 3.32783788e-01 -1.02954581e-01 -7.50253320e-01
-4.65058923e-01 3.40251833e-01 6.86034635e-02 2.66311467e-01
-8.93015504e-01 -9.18061376e-01 -1.18117654e+00 2.79110491e-01
5.31215370e-01 -9.99890804e-01 1.39280641e+00 -3.36176872e-01
-1.68893588e+00 5.62243283e-01 -2.95746922e-01 -3.63704354e-01
5.61430454e-01 -5.96648276e-01 -1.74683377e-01 1.99672401e-01
-7.52049172e-03 1.07869494e+00 6.36440694e-01 -1.16142952e+00
-6.28009617e-01 -4.51622963e-01 -8.18449631e-02 3.43156487e-01
-1.60672069e-01 -6.44638598e-01 -9.55534637e-01 -3.45986813e-01
5.21480501e-01 -9.40238357e-01 -4.26448405e-01 8.76270160e-02
-8.96757245e-01 3.34194988e-01 8.75963032e-01 -7.57640958e-01
1.25231326e+00 -2.04099584e+00 -2.85570383e-01 2.96620131e-01
3.49411547e-01 -3.89762409e-02 1.38994783e-01 1.77615002e-01
2.28402063e-01 -2.62883127e-01 -4.60750639e-01 -9.90998983e-01
6.00555912e-02 2.68678516e-01 -2.16442063e-01 3.15348625e-01
-4.06275988e-01 8.24360371e-01 -4.86542374e-01 -4.18645769e-01
8.82686853e-01 6.54947877e-01 -8.38797987e-01 1.58899024e-01
-2.79003441e-01 1.39205381e-01 -1.91224411e-01 8.74052703e-01
1.06000519e+00 -3.14897686e-01 1.92312256e-01 -2.25102887e-01
-4.51641142e-01 4.70139503e-01 -1.10686779e+00 2.34281492e+00
-7.42659330e-01 7.72737443e-01 1.81555539e-01 -2.47510731e-01
6.50170624e-01 -2.20822439e-01 4.28847522e-01 -9.20324624e-01
2.46896353e-02 1.94541663e-02 -6.59263432e-01 -3.36848646e-02
9.77077901e-01 3.58889461e-01 -6.65287673e-02 5.30003868e-02
-1.12008177e-01 -6.08091950e-01 -1.45945042e-01 1.35163963e-01
1.20089138e+00 4.17695791e-01 -5.56610934e-02 -2.40847409e-01
8.40357617e-02 -2.41109908e-01 3.10650259e-01 6.91352963e-01
6.45327196e-02 1.28268504e+00 1.15690991e-01 -4.95650470e-01
-1.26528502e+00 -1.21133578e+00 -1.86590388e-01 1.01095378e+00
1.36440784e-01 -9.16638136e-01 -1.07731783e+00 -4.12099600e-01
-1.69293046e-01 6.16487384e-01 -7.41426468e-01 5.89233100e-01
-6.39896214e-01 -5.40569603e-01 2.19216108e-01 1.00243902e+00
1.12473559e+00 -7.99019992e-01 -9.42846596e-01 8.98067653e-02
-2.02878118e-01 -1.56894696e+00 -5.19587636e-01 4.49297160e-01
-9.20081377e-01 -9.05323863e-01 -4.44855213e-01 -5.26934087e-01
5.15079260e-01 4.74613070e-01 1.44859993e+00 -2.14197591e-01
-7.65865326e-01 4.70004022e-01 4.30240370e-02 -2.11175755e-01
5.06797850e-01 4.06302065e-01 -6.75630867e-01 -6.29984915e-01
-1.15458161e-01 -6.13896608e-01 -1.12028587e+00 3.53202909e-01
-7.03769624e-01 6.51218295e-01 5.75419307e-01 2.83901334e-01
9.06115592e-01 1.00873671e-01 -5.37597954e-01 -8.94155860e-01
-3.25702608e-01 5.53159602e-02 -5.69529653e-01 -1.58923015e-01
-6.63949788e-01 -2.20479984e-02 4.44109321e-01 3.22535157e-01
-1.16460800e+00 3.94850671e-01 -5.57030082e-01 -2.47315496e-01
-3.36208045e-01 -2.93052495e-01 -5.74238598e-01 1.44525483e-01
3.15585762e-01 -4.86110114e-02 -6.61686480e-01 -5.86282790e-01
4.46785063e-01 2.61989295e-01 7.66969323e-01 -4.99578148e-01
6.50912583e-01 8.12818110e-01 1.35396466e-01 -1.15757048e+00
-1.30738664e+00 -5.90564668e-01 -6.96615458e-01 -4.43887226e-02
1.17227650e+00 -1.15194118e+00 -7.39546955e-01 6.55683935e-01
-8.63350451e-01 -9.63981569e-01 -4.51220125e-01 3.36963013e-02
-5.69595814e-01 5.30037954e-02 -6.00196183e-01 -5.81117392e-01
-3.51870686e-01 -1.01608312e+00 1.61178029e+00 3.44976395e-01
-1.29094198e-01 -7.83228278e-01 -5.00659645e-02 8.11335683e-01
2.20460519e-01 2.17215091e-01 7.15251625e-01 3.31444770e-01
-1.34312725e+00 8.73837918e-02 -4.58195329e-01 3.21399868e-01
-2.77669072e-01 6.03974685e-02 -1.61827409e+00 -3.91437076e-02
-3.60099286e-01 -1.96940899e-01 1.11144710e+00 9.52776313e-01
1.88299656e+00 -1.20213121e-01 -3.34292412e-01 1.27070701e+00
1.45939934e+00 -3.48562956e-01 1.16072011e+00 6.10856116e-01
1.34464717e+00 4.55450624e-01 4.95570093e-01 4.10190642e-01
9.69982684e-01 6.03138328e-01 8.20858657e-01 -4.89657700e-01
-3.53831470e-01 -5.75094342e-01 1.62951484e-01 5.92279673e-01
2.06359057e-03 -3.54094923e-01 -9.19650495e-01 4.68693048e-01
-1.71828139e+00 -6.12661719e-01 -2.10715830e-01 2.01473403e+00
3.45479369e-01 2.90314734e-01 -5.20633534e-02 1.63514927e-01
1.18338138e-01 6.93789661e-01 -3.79710704e-01 -3.81500721e-01
-1.62276849e-01 5.23423672e-01 9.41812277e-01 8.55244517e-01
-1.26950383e+00 1.04409289e+00 6.00459719e+00 8.61728013e-01
-8.24998200e-01 2.39172187e-02 1.07267928e+00 -2.09267706e-01
-4.47628558e-01 -1.46326169e-01 -1.24619401e+00 -4.37661596e-02
8.53418589e-01 9.06496644e-01 3.86262745e-01 1.13219297e+00
1.72349457e-02 -5.12217760e-01 -9.97468829e-01 1.15694237e+00
2.29182184e-01 -1.48911452e+00 -4.04697001e-01 3.19926590e-01
1.00686276e+00 2.66510606e-01 3.78487021e-01 1.14494346e-01
3.02050203e-01 -1.44106114e+00 8.26461852e-01 2.46637359e-01
9.92980838e-01 -9.59706664e-01 3.94238651e-01 1.96914643e-01
-1.57263064e+00 -1.58935562e-01 -2.62841523e-01 5.15346713e-02
1.45396486e-01 7.10981011e-01 -8.99561882e-01 3.73298407e-01
1.06336296e+00 7.06732690e-01 -7.89393246e-01 7.95445621e-01
-2.40201741e-01 2.79281765e-01 -6.84780657e-01 6.14709437e-01
3.91698718e-01 1.71951875e-01 -4.26558629e-02 1.32472861e+00
3.58998626e-01 -6.86639100e-02 1.86193109e-01 7.12385893e-01
-4.66721058e-02 -2.48254880e-01 -3.50477338e-01 4.14487898e-01
3.25201541e-01 1.39137423e+00 -1.03528774e+00 -2.60858536e-01
-9.30485427e-02 1.29172564e+00 8.30472410e-02 1.43988401e-01
-8.03426206e-01 -1.39949873e-01 4.09671664e-01 7.64255762e-01
6.49247289e-01 -5.03879130e-01 -7.32611775e-01 -8.96797240e-01
-4.34913114e-02 -5.54344118e-01 9.49685797e-02 -9.51604843e-01
-6.66886151e-01 7.30247676e-01 -2.75471449e-01 -8.83480489e-01
1.70310557e-01 -8.92101049e-01 -2.62273461e-01 5.09954512e-01
-1.72075844e+00 -1.64492524e+00 -1.06015265e+00 6.79120243e-01
1.01447487e+00 6.14883542e-01 1.00535834e+00 4.10530299e-01
-5.97302377e-01 4.76652354e-01 -1.28236696e-01 -8.88301879e-02
5.32577395e-01 -1.58085966e+00 9.36637998e-01 8.82433832e-01
-4.74485606e-02 3.62399727e-01 5.08984923e-01 -4.28533792e-01
-1.28064513e+00 -1.30383193e+00 5.74296951e-01 -6.79745376e-01
2.43465845e-02 -8.28385174e-01 -1.97003752e-01 8.42659652e-01
-7.47266933e-02 2.19245598e-01 5.10015726e-01 6.64895400e-02
-3.39643478e-01 -1.99396819e-01 -9.68946636e-01 5.23330808e-01
1.21840298e+00 -5.30952752e-01 1.83603302e-01 3.76517326e-01
7.81210721e-01 -1.05419898e+00 -7.37883329e-01 3.48342657e-01
7.93141663e-01 -1.55762565e+00 1.38434327e+00 5.17046869e-01
7.16164887e-01 -3.12454075e-01 -6.22492373e-01 -7.00643837e-01
-1.98443800e-01 -6.62770391e-01 -3.53839338e-01 6.62929535e-01
1.85846820e-01 -7.28515908e-02 1.23452640e+00 6.27207994e-01
-7.63180792e-01 -8.86390567e-01 -6.70552194e-01 -1.92433253e-01
-3.86538148e-01 -8.51641417e-01 6.25081003e-01 3.88104230e-01
-7.25005805e-01 3.59156907e-01 -3.76886010e-01 2.62746960e-01
6.97276592e-01 7.49788359e-02 1.08013189e+00 -1.07452130e+00
-3.93721431e-01 -9.06083658e-02 3.45292054e-02 -1.91865599e+00
-2.02729821e-01 -2.32146323e-01 2.06637666e-01 -1.91092455e+00
-8.89746472e-02 -4.43985999e-01 2.06906617e-01 3.36024195e-01
8.15142393e-02 5.01903713e-01 1.01900384e-01 -2.41821364e-01
-7.02844143e-01 5.75672507e-01 1.47586370e+00 -5.41400397e-03
-2.45471492e-01 -3.58492397e-02 -4.92425144e-01 1.06594837e+00
4.14453685e-01 3.30526792e-02 -5.12700260e-01 -6.35423005e-01
1.43991441e-01 -3.25310260e-01 3.67147654e-01 -1.48030770e+00
-1.01446494e-01 1.00852862e-01 9.39722359e-01 -1.31971049e+00
8.25432897e-01 -8.79801393e-01 -1.41489416e-01 4.42631170e-02
-1.12188309e-02 9.57561657e-02 4.21036243e-01 4.03563619e-01
1.65193662e-01 3.49550128e-01 7.44295835e-01 -2.81390816e-01
-9.67975974e-01 4.42474782e-01 -1.78847797e-02 -1.26802176e-01
6.65282726e-01 -6.58436179e-01 -4.93625700e-01 -2.28933722e-01
-6.20253265e-01 1.46521986e-01 6.26931965e-01 2.56482601e-01
7.79298246e-01 -9.44844186e-01 -1.07939273e-01 4.69558179e-01
-1.09315105e-01 9.46789265e-01 8.62576008e-01 6.78603232e-01
-1.11126924e+00 6.89217210e-01 1.03754155e-01 -7.94753194e-01
-1.14728522e+00 1.24622092e-01 3.43804121e-01 -5.17189980e-01
-1.08339524e+00 1.23990846e+00 5.36892951e-01 -6.20043814e-01
3.72035205e-01 -8.45336497e-01 2.11309835e-01 -2.53675759e-01
4.27570075e-01 6.74781561e-01 3.62651259e-01 -2.99427897e-01
-2.87263304e-01 8.58482718e-01 1.25378063e-02 -1.81233302e-01
1.51973951e+00 -3.85048598e-01 1.97153725e-02 2.29362860e-01
1.10598147e+00 3.29017580e-01 -1.94661367e+00 9.54294354e-02
-5.34880698e-01 -6.80613935e-01 4.87037688e-01 -9.42681074e-01
-1.07842100e+00 9.90759194e-01 7.77875960e-01 -3.94703001e-01
1.10502541e+00 -3.36339116e-01 1.21089935e+00 2.55555093e-01
1.89773694e-01 -1.33271253e+00 2.07285345e-01 6.70567930e-01
4.39446390e-01 -1.04180360e+00 3.54401797e-01 -7.44650900e-01
-4.11655605e-01 1.14881980e+00 9.08717155e-01 -3.32375579e-02
5.71865857e-01 6.12941384e-01 -1.71289355e-01 -4.59598988e-01
-2.69902408e-01 1.28277257e-01 4.11105692e-01 4.66958970e-01
3.92517984e-01 2.68149644e-01 7.35476792e-01 1.02642640e-01
-7.71830320e-01 -4.17506605e-01 3.08530480e-01 8.35073829e-01
-5.62469900e-01 -7.35624731e-01 -5.51559150e-01 1.73539206e-01
-5.36982715e-02 -3.37422609e-01 1.60864294e-01 7.85557210e-01
3.33124876e-01 7.23985076e-01 3.30338478e-01 -2.86106735e-01
3.73162657e-01 -3.96805137e-01 7.66860783e-01 -6.64496601e-01
-3.38559628e-01 3.34658444e-01 1.74610808e-01 -1.15304720e+00
-2.68670861e-02 -5.75933456e-01 -1.14666712e+00 -5.84072232e-01
3.35729606e-02 -3.74317229e-01 8.39124799e-01 7.52652407e-01
3.10824484e-01 1.07454085e+00 4.41913754e-01 -1.25368977e+00
1.73877969e-01 -9.15161669e-01 -6.74415171e-01 -3.04228663e-01
4.10512060e-01 -2.43614048e-01 -5.99784553e-02 -7.30675012e-02] | [8.724772453308105, -2.839165687561035] |
df1bd839-6093-4bb9-91a7-f43b534c20fc | understanding-the-lexical-simplification | null | null | https://aclanthology.org/C16-1069 | https://aclanthology.org/C16-1069.pdf | Understanding the Lexical Simplification Needs of Non-Native Speakers of English | We report three user studies in which the Lexical Simplification needs of non-native English speakers are investigated. Our analyses feature valuable new insight on the relationship between the non-natives{'} notion of complexity and various morphological, semantic and lexical word properties. Some of our findings contradict long-standing misconceptions about word simplicity. The data produced in our studies consists of 211,564 annotations made by 1,100 volunteers, which we hope will guide forthcoming research on Text Simplification for non-native speakers of English. | ['Lucia Specia', 'Gustavo Paetzold'] | 2016-12-01 | understanding-the-lexical-simplification-1 | https://aclanthology.org/C16-1069 | https://aclanthology.org/C16-1069.pdf | coling-2016-12 | ['misconceptions', 'complex-word-identification'] | ['miscellaneous', 'natural-language-processing'] | [ 5.07815974e-03 4.91910189e-01 -5.07076569e-02 -4.93193150e-01
-3.22227895e-01 -6.10799730e-01 2.98790991e-01 6.69515729e-01
-1.00252914e+00 2.99845070e-01 8.53298843e-01 -7.29245484e-01
1.17576867e-01 -4.50480312e-01 -3.73508781e-02 2.12838426e-01
9.09619927e-02 5.49767315e-02 -8.51334829e-04 -7.54711568e-01
4.94876832e-01 2.06355989e-01 -1.53995275e+00 1.99341714e-01
1.41802955e+00 2.05334015e-02 3.79223406e-01 5.06315768e-01
-1.76327631e-01 3.67680192e-01 -6.25742972e-01 -1.10296047e+00
1.56801306e-02 -1.48033306e-01 -8.78931344e-01 -2.50904918e-01
6.84333682e-01 -3.17660123e-02 -1.60103142e-01 1.41071689e+00
6.34306371e-01 4.53633338e-01 5.53395271e-01 -4.99714851e-01
-1.24864197e+00 1.08027589e+00 -6.35062084e-02 7.37740159e-01
8.51994514e-01 3.23388249e-01 1.20702958e+00 -8.39336574e-01
8.03578436e-01 1.50930989e+00 1.11611772e+00 5.21480381e-01
-1.11945319e+00 -6.74436331e-01 4.90600199e-01 1.90186929e-02
-1.65335321e+00 -1.00181186e+00 2.56352335e-01 -3.82136911e-01
1.58025956e+00 5.90598404e-01 8.46629322e-01 8.23134065e-01
1.67611092e-01 4.86524493e-01 1.01585364e+00 -9.00596857e-01
-3.09693456e-01 4.43395942e-01 4.81268406e-01 6.43697143e-01
7.84291923e-01 -5.79426177e-02 -8.10497224e-01 -1.80745319e-01
4.44215149e-01 -7.23582208e-01 -3.00943881e-01 3.22566092e-01
-8.19096148e-01 8.67283285e-01 -5.55322506e-02 5.56972265e-01
-1.76267967e-01 -4.93633658e-01 6.10875785e-01 5.22293746e-01
4.70439613e-01 8.73443604e-01 -5.85804045e-01 -4.45682287e-01
-5.12639999e-01 3.89559180e-01 1.11500096e+00 1.34575105e+00
2.73072243e-01 1.49674430e-01 1.42966673e-01 1.17213440e+00
4.12538469e-01 6.28400445e-01 8.49953651e-01 -7.63699412e-01
3.55308235e-01 1.90184057e-01 1.47998869e-01 -9.05545652e-01
-6.80110335e-01 -5.87625578e-02 -3.41734499e-01 -1.67408884e-01
3.50797743e-01 -6.71022609e-02 -4.22439456e-01 1.92124641e+00
-1.95609719e-01 -1.02191079e+00 -1.89139292e-01 3.83490175e-01
9.03713465e-01 2.20716998e-01 8.50667238e-01 -4.23882514e-01
1.62235785e+00 -3.93330753e-01 -1.03294766e+00 -3.67215246e-01
8.58172774e-01 -1.30361044e+00 1.92618895e+00 3.00113529e-01
-1.38356113e+00 -6.07926190e-01 -1.01048136e+00 -5.76209307e-01
-6.51143312e-01 -1.83691189e-01 9.04260814e-01 1.46803617e+00
-9.59261894e-01 4.19244856e-01 -5.40255308e-01 -9.23036873e-01
2.28675634e-01 1.39688075e-01 -3.50655019e-01 7.26969615e-02
-1.26377165e+00 1.50091016e+00 2.09197670e-01 -4.20653015e-01
9.12202820e-02 -7.47987986e-01 -1.18735111e+00 -1.99996978e-01
-2.90285200e-02 -7.89470911e-01 1.59150684e+00 -9.29278493e-01
-1.21406031e+00 1.14213169e+00 -3.80652696e-01 5.90802990e-02
1.36195540e-01 -4.81788963e-01 -7.04043806e-01 -5.96170664e-01
4.86633360e-01 4.91141975e-01 3.29410136e-01 -6.75176680e-01
-7.79309690e-01 -2.13292703e-01 1.41894877e-01 6.94203675e-01
-4.89975542e-01 4.00941998e-01 -3.48036475e-02 -1.45145535e+00
-1.42186850e-01 -7.09522724e-01 3.16206776e-02 -2.64571339e-01
-3.89231026e-01 -5.64299881e-01 9.46896449e-02 -1.03427458e+00
1.91968226e+00 -1.97495508e+00 -4.78616133e-02 7.18768835e-02
2.79976457e-01 3.15332443e-01 -1.33720562e-01 7.13965595e-01
-1.54666275e-01 1.04245067e+00 -4.89694215e-02 -3.45588475e-01
2.87269264e-01 -2.67629594e-01 1.56961903e-01 5.07469952e-01
-3.21997166e-01 9.09377038e-01 -8.98679078e-01 -3.25165361e-01
1.17235020e-01 3.19195837e-01 -6.81819856e-01 -1.15234882e-01
1.82273760e-01 -1.72208786e-01 -2.57533994e-02 3.35420191e-01
4.07749116e-01 4.55732912e-01 1.54100999e-01 -1.42163029e-02
-5.24598539e-01 9.45265293e-01 -8.33077967e-01 1.68690228e+00
-4.70687807e-01 7.37536609e-01 -5.32853864e-02 9.02911574e-02
1.09419152e-02 4.19535041e-02 -4.50827122e-01 -8.82612288e-01
3.26727390e-01 1.04550414e-01 6.34346604e-01 -6.48567975e-01
8.25920045e-01 -4.01009828e-01 -2.32751429e-01 7.01682210e-01
-3.22186619e-01 -4.22920108e-01 5.21417558e-01 1.27372354e-01
7.56584167e-01 -1.68081924e-01 9.71832991e-01 -1.08204567e+00
2.91200131e-01 -1.10927507e-01 3.24282646e-01 9.04909849e-01
-2.55109429e-01 -4.70053591e-02 9.90694314e-02 -4.58090395e-01
-1.26411211e+00 -9.29824054e-01 -2.35926002e-01 1.76835370e+00
-1.99852139e-01 -1.05301189e+00 -1.05712652e+00 -1.65128082e-01
-1.68420672e-01 1.35501587e+00 -4.47584212e-01 -1.81710765e-01
-6.16934538e-01 -4.81925309e-01 7.59133697e-01 3.82743657e-01
-1.68504074e-01 -1.04703426e+00 -6.55742884e-01 1.85213968e-01
-3.96630913e-01 -1.04850173e+00 -9.20285225e-01 -2.84749299e-01
-7.04267085e-01 -9.35273528e-01 -3.51117313e-01 -8.77813101e-01
3.72176707e-01 3.82821888e-01 1.35674620e+00 7.71551728e-01
-2.84420580e-01 2.89658338e-01 -5.41822016e-01 -7.88132131e-01
-5.04658878e-01 4.02014017e-01 3.67015511e-01 -1.06683731e+00
7.88617730e-01 -5.22684753e-01 -2.11059123e-01 -1.83897495e-01
-8.41036499e-01 8.65855366e-02 3.86412978e-01 2.92777061e-01
-2.74313856e-02 -6.70614615e-02 3.50653321e-01 -1.48701942e+00
1.42780054e+00 -2.00255617e-01 -5.08023724e-02 3.16672288e-02
-6.59841776e-01 8.56202561e-03 4.92293477e-01 -4.84644413e-01
-1.27905703e+00 -3.42807293e-01 -3.93027842e-01 8.05215240e-01
-1.04136869e-01 5.57316840e-01 -4.05952245e-01 -7.12697059e-02
1.05985248e+00 -2.60531485e-01 -1.15381643e-01 -5.20800352e-01
7.81929493e-01 6.26673639e-01 4.66342449e-01 -6.94285393e-01
8.12240481e-01 -1.59554794e-01 -8.31828475e-01 -1.73564935e+00
-6.85954988e-01 -2.33271807e-01 -6.41585469e-01 2.92166829e-01
5.21764576e-01 -7.63713896e-01 -6.39509797e-01 4.41645741e-01
-1.27045131e+00 -3.10736269e-01 -3.66401255e-01 2.16357857e-01
-1.30854279e-01 7.67997801e-01 -6.25791490e-01 -7.38300264e-01
-4.96813625e-01 -8.59406114e-01 5.73460221e-01 9.89403799e-02
-1.40847516e+00 -1.19616389e+00 -8.56150687e-03 1.45940512e-01
7.10635543e-01 -3.84264410e-01 1.54600823e+00 -7.34597325e-01
1.38183609e-01 -9.47238207e-02 6.19901856e-03 1.77435249e-01
1.20609514e-01 1.13747656e-01 -6.74995780e-01 -4.59613279e-02
1.15666976e-02 -8.19784254e-02 3.26305211e-01 4.03565705e-01
7.60487556e-01 -6.23532712e-01 -6.62867799e-02 3.76163840e-01
1.06098747e+00 1.90167911e-02 5.26874721e-01 1.84866041e-01
7.90823102e-01 6.11199796e-01 3.20293754e-01 4.50774997e-01
8.72811079e-01 3.51951808e-01 -3.98221105e-01 1.43260464e-01
-2.51045406e-01 -2.10294545e-01 1.96906194e-01 1.09694767e+00
-1.59410745e-01 -4.22936648e-01 -9.08150852e-01 4.54665065e-01
-1.19880152e+00 -8.20306361e-01 -2.94722646e-01 2.23660994e+00
1.17129540e+00 3.07105839e-01 2.14097530e-01 2.08750293e-01
6.74101233e-01 2.98071712e-01 -2.78842807e-01 -8.73961687e-01
-2.00314850e-01 4.95004803e-01 3.81517202e-01 1.22045624e+00
-9.47832048e-01 1.52396405e+00 7.49951077e+00 7.17437863e-01
-5.52447617e-01 1.64193764e-01 3.07314366e-01 1.53556287e-01
-8.67777050e-01 -1.69074669e-01 -6.43465400e-01 5.64345457e-02
1.05508423e+00 -9.05726194e-01 6.84840858e-01 4.66198713e-01
3.60745072e-01 -2.08461508e-01 -1.20042145e+00 9.12298441e-01
2.19509661e-01 -5.72623789e-01 2.21772298e-01 -1.84792414e-01
3.43097657e-01 -1.81164682e-01 6.33429289e-02 4.05274421e-01
4.39643204e-01 -1.21777678e+00 1.00906587e+00 4.53546613e-01
8.83586109e-01 -8.68125141e-01 3.12378824e-01 4.69996542e-01
-9.46406066e-01 1.62834018e-01 -1.91482753e-01 -6.75985336e-01
2.14125410e-01 3.50788265e-01 -4.35969770e-01 -3.98017466e-01
5.58267891e-01 2.81018585e-01 -9.63781595e-01 6.14274383e-01
-4.02439296e-01 3.08715135e-01 -3.46006364e-01 -2.65110672e-01
-3.09530169e-01 1.46305799e-01 6.77444935e-01 1.70507455e+00
-1.31813109e-01 5.31694412e-01 -1.91381827e-01 4.29858774e-01
-9.76495445e-03 8.69287908e-01 -6.78412557e-01 -3.04132491e-01
7.39098072e-01 1.01938212e+00 -9.06136215e-01 -4.04473335e-01
-5.42132735e-01 1.07507956e+00 5.35992801e-01 1.88591614e-01
-3.23702097e-01 -7.41526604e-01 9.07299876e-01 3.78532946e-01
-1.80848554e-01 -5.06533265e-01 -9.37768400e-01 -1.28407848e+00
-5.78425527e-02 -1.24781644e+00 2.37170786e-01 -5.77249944e-01
-1.49743426e+00 7.21310496e-01 1.34583041e-01 -2.20656529e-01
-2.13242546e-01 -7.82772243e-01 -3.90897453e-01 1.22832394e+00
-9.98883605e-01 -6.52676761e-01 -7.15030134e-02 1.08053677e-01
9.14986730e-01 4.50192913e-02 1.23906982e+00 1.25222459e-01
-4.80187923e-01 9.69335794e-01 -4.80763316e-01 -1.59955218e-01
6.13144994e-01 -1.27074862e+00 1.28404367e+00 9.71267223e-01
6.92171231e-02 1.30598617e+00 1.00200319e+00 -1.06384444e+00
-8.92131925e-01 -5.71182668e-01 1.84370482e+00 -7.63217092e-01
4.89970744e-01 -2.85375297e-01 -8.96661222e-01 8.84018958e-01
7.01252878e-01 -7.73164034e-01 1.23859346e+00 5.34475863e-01
-2.97720969e-01 5.24471521e-01 -1.15872002e+00 1.06656837e+00
1.72479188e+00 -7.50886321e-01 -1.10993254e+00 2.99122781e-01
7.96737075e-01 -4.22669113e-01 -7.14457095e-01 8.65425467e-02
8.62992346e-01 -7.98589289e-01 7.52206802e-01 -8.32104146e-01
1.22455820e-01 3.41715634e-01 1.60735752e-02 -1.70945013e+00
-6.81467295e-01 -8.69067848e-01 4.60194379e-01 1.00743365e+00
6.38566792e-01 -4.77036744e-01 1.02458537e-01 1.26610374e+00
-3.32987696e-01 -1.90020129e-01 -6.84328973e-01 -4.13876325e-01
4.74227548e-01 -6.91016376e-01 4.69854772e-01 1.06138611e+00
3.27351511e-01 6.99537754e-01 3.35364789e-01 -1.50968313e-01
2.40070522e-01 -8.37267637e-01 3.75444204e-01 -1.46880794e+00
1.62962168e-01 -9.07410622e-01 -7.71930963e-02 -7.83550799e-01
4.46545362e-01 -9.28716779e-01 -2.23118618e-01 -1.25153613e+00
-2.96884757e-02 -2.79927790e-01 3.77165496e-01 3.76076818e-01
-3.75513285e-01 5.21368906e-02 2.27589890e-01 -8.58241320e-02
-3.91132593e-01 4.62210625e-01 8.18435073e-01 3.60435486e-01
-3.99180114e-01 -1.33068502e-01 -1.57753730e+00 1.16977453e+00
7.81306446e-01 -3.59760821e-01 -5.17130673e-01 -8.62309098e-01
5.55282235e-01 -7.09403634e-01 -2.87979603e-01 -7.93824196e-01
1.26575366e-01 -3.46728861e-01 3.24930072e-01 1.28377043e-02
-9.63220373e-03 -4.38132018e-01 -2.76430994e-02 3.19621146e-01
-3.71361077e-01 9.14486945e-01 5.60547709e-01 3.57690267e-02
4.99362618e-01 -6.32443845e-01 7.44053900e-01 -4.51080471e-01
-6.73392892e-01 -1.94424286e-01 -1.00695503e+00 6.88050449e-01
5.48828781e-01 -2.60746062e-01 -1.62711456e-01 -4.75369185e-01
-5.65680861e-01 -2.06564754e-01 6.66052818e-01 5.74138582e-01
3.39517117e-01 -9.50258434e-01 -6.81231439e-01 2.93875396e-01
3.32492501e-01 -5.46704710e-01 -1.17268145e-01 2.58217692e-01
-7.79614747e-01 4.75166529e-01 -1.29555240e-01 3.38175684e-01
-1.20124400e+00 4.44600254e-01 1.11152723e-01 4.14450824e-01
-7.82044530e-01 1.04245365e+00 -1.85800433e-01 -3.45671624e-01
1.78112313e-01 -3.86931866e-01 -1.84965655e-01 3.03589433e-01
7.59725630e-01 6.03069425e-01 1.90601051e-01 -9.84206319e-01
-2.36309141e-01 1.34986177e-01 -5.43532848e-01 -5.29108822e-01
9.36300337e-01 -6.62391186e-01 -2.70402431e-01 6.56307995e-01
8.53505790e-01 7.37513721e-01 -3.19787152e-02 -1.70553759e-01
2.93613672e-01 -5.45914710e-01 -1.64273679e-01 -7.13109672e-01
-3.22662145e-01 5.18912852e-01 2.44651005e-01 2.58614659e-01
7.43243694e-01 -1.59743622e-01 7.47647285e-01 7.54428685e-01
2.27572128e-01 -1.47858822e+00 -6.86636209e-01 9.25880790e-01
9.64829326e-01 -8.91486764e-01 2.01988474e-01 -7.45589375e-01
-7.87582278e-01 9.03540492e-01 7.02728510e-01 1.52149364e-01
7.81671822e-01 -3.17761861e-02 1.45652890e-01 -1.66655838e-01
-5.95094264e-01 -3.27560931e-01 2.57020473e-01 9.61205304e-01
1.45396352e+00 2.68302947e-01 -1.20588470e+00 1.07140887e+00
-1.25655639e+00 -4.28221881e-01 6.08938038e-01 9.27174807e-01
-5.18458843e-01 -1.25893819e+00 -1.30108133e-01 6.02226734e-01
-4.85967427e-01 -1.06092882e+00 -4.97716278e-01 1.03705478e+00
8.60542580e-02 9.47482407e-01 -2.02469841e-01 -1.26912877e-01
8.95104408e-01 4.04224008e-01 5.79689443e-01 -1.15458405e+00
-8.39809179e-01 -2.16730878e-01 5.93759716e-01 -3.16151738e-01
1.12456508e-01 -1.02136314e+00 -1.09903371e+00 -8.42552125e-01
-1.79463819e-01 5.20454124e-02 3.23966593e-01 8.58108819e-01
1.30544007e-01 1.03038952e-01 -1.16429709e-01 -6.01082563e-01
-3.12475115e-01 -1.43491888e+00 -5.29825985e-01 4.95748669e-01
3.38388979e-02 -2.99514472e-01 -2.40328252e-01 2.68434733e-01] | [10.888312339782715, 10.337677955627441] |
87cd585e-980f-4a07-a242-ec3dc8b0303f | an-open-source-gloss-based-baseline-for | 2305.17714 | null | https://arxiv.org/abs/2305.17714v1 | https://arxiv.org/pdf/2305.17714v1.pdf | An Open-Source Gloss-Based Baseline for Spoken to Signed Language Translation | Sign language translation systems are complex and require many components. As a result, it is very hard to compare methods across publications. We present an open-source implementation of a text-to-gloss-to-pose-to-video pipeline approach, demonstrating conversion from German to Swiss German Sign Language, French to French Sign Language of Switzerland, and Italian to Italian Sign Language of Switzerland. We propose three different components for the text-to-gloss translation: a lemmatizer, a rule-based word reordering and dropping component, and a neural machine translation system. Gloss-to-pose conversion occurs using data from a lexicon for three different signed languages, with skeletal poses extracted from videos. To generate a sentence, the text-to-gloss system is first run, and the pose representations of the resulting signs are stitched together. | ['Sarah Ebling', 'Yoav Goldberg', 'Zifan Jiang', 'Anne Göhring', 'Mathias Müller', 'Amit Moryossef'] | 2023-05-28 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 5.69553375e-01 -2.06206739e-01 -4.47344296e-02 -6.79100811e-01
-1.41237676e+00 -9.42147136e-01 5.34937978e-01 -8.28906119e-01
-4.45188910e-01 5.94106853e-01 5.97956419e-01 -1.38721317e-01
1.64913163e-01 -3.18385929e-01 -7.40355074e-01 -4.07323241e-01
4.74755257e-01 8.82707357e-01 3.34104270e-01 -2.50825167e-01
7.05766752e-02 4.53092963e-01 -1.44735098e+00 5.65824151e-01
5.24234414e-01 4.65903968e-01 -9.83646587e-02 9.99181032e-01
1.22405373e-01 4.15683150e-01 -4.09801006e-01 -6.60503387e-01
5.31022668e-01 -1.04508662e+00 -5.03920794e-01 5.19824885e-02
1.21598542e+00 -6.95275068e-01 -2.85726190e-01 6.80447042e-01
7.98541427e-01 -3.16861421e-01 7.68994629e-01 -8.72546077e-01
-4.34284061e-01 2.67114908e-01 -2.14877367e-01 -6.32427990e-01
6.22855902e-01 5.05674779e-01 9.94296968e-01 -8.77848685e-01
1.65192199e+00 1.31297827e+00 5.16795754e-01 8.47729266e-01
-8.86435688e-01 -3.92614275e-01 -2.88830668e-01 -1.11172937e-01
-9.68621671e-01 -3.45671713e-01 3.49941850e-01 -5.32728136e-01
1.21704471e+00 2.11220637e-01 1.40513206e+00 1.19021726e+00
1.88814159e-02 7.50240028e-01 1.16735661e+00 -5.58534861e-01
-1.23473182e-01 -6.24661326e-01 -3.22702736e-01 8.69681180e-01
2.41516396e-01 4.36637521e-01 -1.06299412e+00 1.03389360e-01
7.18733966e-01 -6.52164280e-01 -4.47186977e-02 -4.73228067e-01
-1.30075264e+00 4.30615664e-01 7.52566233e-02 -9.13351104e-02
-3.18782389e-01 4.73030120e-01 5.18575728e-01 4.69251394e-01
-1.08681358e-01 1.84927061e-02 -4.11827981e-01 -4.26733315e-01
-9.53795671e-01 5.03624737e-01 8.82583559e-01 1.06325841e+00
5.08163460e-02 -9.46360603e-02 -3.12373955e-02 6.12711966e-01
5.75751841e-01 1.00610888e+00 3.92155111e-01 -1.02736676e+00
7.34940290e-01 3.05143982e-01 -2.97963113e-01 -4.30155843e-01
-3.13994259e-01 1.44555479e-01 -1.47333607e-01 5.32844067e-01
6.32913828e-01 -1.77927896e-01 -1.45900512e+00 1.53969646e+00
1.85154989e-01 -3.61079305e-01 9.84281860e-03 1.31862104e+00
8.06176960e-01 2.40062952e-01 4.14366536e-02 1.05065130e-01
1.43302298e+00 -9.12485421e-01 -4.51734096e-01 -2.75283843e-01
6.47779167e-01 -1.13850844e+00 9.87425327e-01 4.40086395e-01
-1.45981979e+00 -1.88470498e-01 -8.99632514e-01 -5.37910044e-01
-2.24425897e-01 5.13190985e-01 2.91775763e-01 3.16190034e-01
-9.44797635e-01 3.33104759e-01 -9.17234898e-01 -8.58159840e-01
8.42993855e-02 3.09140891e-01 -6.93778753e-01 -1.88723475e-01
-7.44732440e-01 1.30041981e+00 2.76650935e-01 1.58451632e-01
-4.20703948e-01 -1.65403917e-01 -9.35408235e-01 -6.13866210e-01
-5.89540973e-03 -1.10332036e+00 1.48991227e+00 -1.12090731e+00
-2.08657360e+00 1.42014384e+00 -2.22421318e-01 4.35963608e-02
1.01139164e+00 -3.40805650e-01 -1.11068092e-01 2.22204804e-01
-4.62215394e-02 8.08353186e-01 9.28295851e-01 -9.52083111e-01
-8.10797811e-01 -7.59722367e-02 -4.56937551e-01 4.46580887e-01
6.19326174e-01 5.40368021e-01 -9.34154689e-01 -7.41020679e-01
3.30077291e-01 -1.16973650e+00 2.10959673e-01 4.23043221e-01
-1.47949263e-01 1.63196385e-01 3.86270225e-01 -1.41076291e+00
9.22488213e-01 -1.89656830e+00 6.06636643e-01 4.59866405e-01
-2.32944205e-01 2.40371555e-01 -4.86420214e-01 3.96054149e-01
-6.71625277e-03 -3.72480571e-01 -3.02383721e-01 -2.67582774e-01
8.82436857e-02 4.09613997e-01 1.77285019e-02 4.92707908e-01
4.34771270e-01 1.04303384e+00 -1.04845989e+00 -6.92333221e-01
4.88454312e-01 6.42021775e-01 -7.20469356e-01 -1.87730324e-02
-3.66073668e-01 5.30202925e-01 -5.98947965e-02 9.96604621e-01
2.66130000e-01 4.76225078e-01 3.71711075e-01 -5.00101447e-01
-1.66308448e-01 5.03931165e-01 -9.03428614e-01 1.81002378e+00
-3.46791238e-01 7.69700527e-01 4.27603871e-02 -2.43382618e-01
6.43926501e-01 3.94766927e-01 4.81804371e-01 -4.73752350e-01
4.47765589e-01 1.00635958e+00 4.13733348e-03 -7.58647919e-01
3.47899914e-01 -2.74869174e-01 -2.56996334e-01 3.42500418e-01
3.00953418e-01 -7.41486311e-01 8.66073966e-01 -1.77712724e-01
9.64778960e-01 1.16516471e+00 1.83738515e-01 3.96751285e-01
2.23809958e-01 3.68275076e-01 1.98732883e-01 7.32913539e-02
-5.64685836e-02 1.17614520e+00 3.83181483e-01 -3.09454590e-01
-1.27314341e+00 -1.22480619e+00 3.68802339e-01 7.62179196e-01
-4.20341700e-01 -4.88473713e-01 -9.40703452e-01 -4.46681857e-01
-1.21928915e-01 5.46296597e-01 -3.28003377e-01 -2.34182822e-04
-1.12125432e+00 -2.68833824e-02 9.27054584e-01 4.27906632e-01
2.03649089e-01 -1.13126302e+00 -8.35722804e-01 1.34919643e-01
-3.33577573e-01 -1.04585707e+00 -9.91614103e-01 -2.02017710e-01
-7.72612631e-01 -1.07053530e+00 -1.06530797e+00 -1.25472188e+00
7.56576955e-01 -5.72179198e-01 8.81323397e-01 -9.06413123e-02
-3.82872164e-01 2.08772272e-01 -2.44840696e-01 -2.25668281e-01
-8.93702388e-01 -1.73610881e-01 -6.73485082e-03 -1.53013289e-01
2.35861704e-01 -2.49659687e-01 -3.51303488e-01 2.78998345e-01
-8.42066526e-01 3.58620256e-01 9.27592516e-01 8.11036229e-01
7.09218621e-01 -1.11328709e+00 -3.41937244e-01 -7.16626197e-02
4.48703825e-01 4.02412981e-01 -7.60606349e-01 4.98438954e-01
-1.23867951e-01 2.45427802e-01 2.15689093e-01 -5.20729125e-01
-8.55541110e-01 8.63610148e-01 -3.67162347e-01 -1.90397412e-01
-4.74348525e-03 4.73501414e-01 -3.07075620e-01 6.79498240e-02
6.31841004e-01 3.35136443e-01 2.88194895e-01 -3.00387174e-01
9.24791634e-01 7.38605857e-01 9.90256965e-01 -3.46231073e-01
9.19277072e-01 3.80350292e-01 2.55204085e-03 -7.04916596e-01
-2.34048232e-01 -2.23839164e-01 -1.02546799e+00 -6.58801615e-01
9.84137535e-01 -5.85936069e-01 -3.75030428e-01 8.35780263e-01
-1.38852179e+00 -4.69530076e-01 -5.81472754e-01 7.62674451e-01
-1.08160841e+00 2.55879372e-01 -6.55203104e-01 -1.51116714e-01
-2.79045284e-01 -1.17762470e+00 1.51822197e+00 1.28373289e-02
-5.76709628e-01 -3.77991378e-01 4.45415199e-01 4.65230405e-01
1.76873833e-01 2.37763047e-01 5.58695138e-01 -9.92186666e-02
-4.73831058e-01 -5.04736543e-01 -1.97150916e-01 4.93385106e-01
1.31225914e-01 4.35158253e-01 -3.80371749e-01 -1.62167996e-01
-7.83755898e-01 -5.55279374e-01 7.07012594e-01 4.30083871e-01
-1.30840540e-01 -1.34138986e-01 1.59697235e-02 8.24065983e-01
1.07357240e+00 3.02430261e-02 8.33360255e-01 1.79209128e-01
5.94181538e-01 4.43748653e-01 6.32933795e-01 -1.00296713e-01
4.69333500e-01 1.02829707e+00 -2.28909835e-01 2.02969480e-02
-1.13319111e+00 -6.03872180e-01 1.00733972e+00 1.18903732e+00
-6.75811172e-01 5.09698540e-02 -7.43290007e-01 5.63934147e-01
-1.49246669e+00 -9.89013076e-01 -1.21924438e-01 2.13346386e+00
1.02079463e+00 -1.47264183e-01 2.54922092e-01 -1.00581087e-01
3.29344600e-01 -2.02703819e-01 -2.23696530e-01 -5.36776900e-01
-1.76681623e-01 3.87477219e-01 6.42567098e-01 8.73010278e-01
-8.05492103e-01 1.61900675e+00 6.15264988e+00 4.09037530e-01
-1.60558903e+00 -9.18460861e-02 -2.49785379e-01 -3.44024122e-01
-2.57127769e-02 1.83694601e-01 -6.18230402e-01 -1.21900328e-02
7.00293124e-01 7.52016827e-02 4.90061879e-01 5.06693423e-01
3.87048244e-01 2.03945022e-02 -1.10569251e+00 1.02287912e+00
5.25339842e-01 -9.85565901e-01 3.82378548e-01 -1.63019568e-01
6.26191616e-01 2.69571751e-01 -4.06159341e-01 -4.23229337e-02
4.11213785e-01 -5.55751443e-01 1.27362430e+00 7.35185027e-01
1.51023340e+00 -1.69713005e-01 3.81171256e-01 -3.42977554e-01
-1.29925203e+00 4.13553059e-01 2.86735624e-01 3.76520425e-01
9.83736455e-01 -1.71058014e-01 -8.07612538e-01 4.45523888e-01
3.03612411e-01 5.76190412e-01 -3.97007465e-01 1.14073896e+00
-9.65703666e-01 5.37737370e-01 -6.93521917e-01 -1.69149995e-01
9.03484505e-03 -2.83311367e-01 7.43327558e-01 1.40468919e+00
5.91591358e-01 -1.72136486e-01 -1.02883458e-01 3.87458682e-01
1.40295669e-01 3.14323127e-01 -5.79445899e-01 -1.13834314e-01
-2.43419752e-01 8.02879274e-01 -7.70429850e-01 -4.98600036e-01
-2.60910124e-01 1.67277086e+00 -1.14529796e-01 4.18156147e-01
-7.97310114e-01 -3.98658365e-01 5.93186736e-01 9.35962722e-02
3.33187491e-01 -4.92023796e-01 -1.11412875e-01 -1.29770303e+00
4.05840009e-01 -1.18765342e+00 2.33562097e-01 -1.05128038e+00
-9.53875244e-01 4.22009647e-01 9.44596808e-03 -1.61160398e+00
-7.63744056e-01 -9.26704526e-01 -1.58622950e-01 7.21495271e-01
-1.13267922e+00 -1.74759638e+00 -1.19737461e-01 4.59270000e-01
6.13433659e-01 7.79901689e-04 7.43783593e-01 5.51801920e-01
1.27387762e-01 4.74268198e-01 -9.45292488e-02 5.46965420e-01
1.00746000e+00 -8.23842347e-01 8.67093742e-01 9.42365885e-01
3.25723290e-01 4.55239266e-01 5.23211539e-01 -1.08930957e+00
-1.41641045e+00 -8.62855256e-01 1.48489332e+00 -6.61866605e-01
6.16948962e-01 -1.66941509e-01 2.70849783e-02 6.92061782e-01
1.49507776e-01 -3.45364213e-01 1.36373758e-01 -6.70311928e-01
-3.69936377e-01 1.78495243e-01 -7.99320757e-01 1.03836012e+00
1.39600408e+00 -5.22449076e-01 -9.07191575e-01 4.11725849e-01
7.36417845e-02 -1.02152395e+00 -4.89796728e-01 3.70752588e-02
1.39992416e+00 -4.58198845e-01 8.00125539e-01 -7.68129110e-01
6.09934330e-01 -5.38247466e-01 -1.66481033e-01 -1.25569248e+00
1.91628054e-01 -9.38090026e-01 1.23171814e-01 7.60393560e-01
6.20867968e-01 -4.69015211e-01 5.93273580e-01 2.64264971e-01
-2.11562708e-01 -2.75745094e-01 -1.31688344e+00 -8.28868330e-01
6.47237226e-02 -4.44371492e-01 1.46371033e-02 3.68994087e-01
-8.03041607e-02 2.13442817e-01 -4.48625982e-01 -1.90693229e-01
5.18983960e-01 -1.37388734e-02 1.06921756e+00 -7.76777148e-01
-2.82169223e-01 -6.62220895e-01 -6.45230293e-01 -1.12291777e+00
-2.96840817e-01 -1.02481866e+00 4.64354008e-01 -2.14460015e+00
-2.10713282e-01 3.20861369e-01 4.53909099e-01 7.42737949e-01
3.63384008e-01 5.62026978e-01 4.35322285e-01 2.32653543e-01
-1.08237183e-02 1.54943570e-01 1.34685934e+00 -1.44618750e-01
-3.44755650e-01 -1.53831929e-01 7.15079308e-02 8.52786839e-01
6.56423450e-01 -5.34145057e-01 1.59444481e-01 -6.64824665e-01
2.75505036e-01 -1.96621388e-01 3.35732847e-01 -9.55012083e-01
-6.19249884e-03 -1.00248590e-01 -1.61914468e-01 -6.66641414e-01
5.02306044e-01 -6.10562623e-01 2.17445284e-01 7.52490282e-01
-1.32129982e-01 1.45618021e-01 -2.36821808e-02 -9.63829234e-02
-1.82776228e-01 1.40093267e-01 7.45887935e-01 -1.05576105e-01
-5.19627512e-01 1.92909539e-02 -4.26480025e-01 6.45331815e-02
6.85666621e-01 -5.34624755e-01 -1.62646875e-01 -5.10406673e-01
-6.93555772e-01 -7.94233903e-02 4.90070283e-01 6.72129631e-01
7.15482235e-01 -1.42709005e+00 -1.14833093e+00 4.57836360e-01
2.65947074e-01 -3.56367648e-01 -2.41897181e-01 1.03292358e+00
-1.47029853e+00 3.96034807e-01 -4.37716514e-01 -5.74369848e-01
-1.82496381e+00 -2.60958791e-01 5.05079508e-01 -8.59950930e-02
-6.76365435e-01 6.72162533e-01 -5.63589871e-01 -8.42358947e-01
1.57354996e-01 -8.51100802e-01 4.71588224e-01 -3.13384086e-02
2.43512198e-01 9.31669176e-02 2.50948574e-02 -1.23552918e+00
-3.76154393e-01 1.46308517e+00 4.91395146e-01 -9.25693095e-01
1.14950144e+00 3.86073053e-01 8.94745961e-02 -4.63611744e-02
9.92031276e-01 2.47738644e-01 -1.03894126e+00 1.39535919e-01
-1.03408858e-01 -2.67562032e-01 -4.50772643e-01 -1.12555230e+00
-1.03114653e+00 7.41793573e-01 6.52278543e-01 -9.46786106e-01
1.03476810e+00 1.03759561e-02 9.84378278e-01 5.19048095e-01
5.22476971e-01 -1.45620394e+00 -3.72801453e-01 9.20749068e-01
1.42240620e+00 -7.17107952e-01 1.57169327e-02 -2.68895894e-01
-6.57135785e-01 1.25931156e+00 1.01911709e-01 -3.86116207e-02
4.26785380e-01 4.19756442e-01 8.46393406e-01 -8.99056345e-02
-3.21828961e-01 -4.37951833e-01 7.22672939e-01 7.68854320e-01
5.25488019e-01 -2.70799287e-02 -9.71869826e-01 1.22000547e-02
-6.38336957e-01 5.17847538e-01 1.22187965e-01 1.17935228e+00
3.85259427e-02 -1.65114915e+00 -5.08099318e-01 1.78191453e-01
5.24830297e-02 -1.70465738e-01 -1.01202178e+00 8.23937237e-01
3.47499460e-01 4.11774933e-01 -4.05180186e-01 -3.78503501e-01
8.21123600e-01 4.39742595e-01 1.13499141e+00 -4.29378450e-01
-7.39180267e-01 4.28088188e-01 5.21342039e-01 -8.06494832e-01
-4.70520735e-01 -1.11038649e+00 -1.43444777e+00 8.00706968e-02
-1.51633531e-01 -7.29787171e-01 8.98939013e-01 9.96213794e-01
2.31655583e-01 2.42793500e-01 -1.68540597e-01 -1.18388033e+00
-3.96847785e-01 -9.91038740e-01 -5.73839881e-02 6.47967279e-01
-6.38345676e-03 -4.17512327e-01 -2.28108793e-01 6.46967351e-01] | [9.196379661560059, -6.524487495422363] |
0b7b4f2f-fffb-4b04-90aa-209083c34f51 | fire-burns-sword-cuts-commonsense-inductive | null | null | https://aclanthology.org/2022.acl-short.56 | https://aclanthology.org/2022.acl-short.56.pdf | Fire Burns, Sword Cuts: Commonsense Inductive Bias for Exploration in Text-based Games | Text-based games (TGs) are exciting testbeds for developing deep reinforcement learning techniques due to their partially observed environments and large action spaces. In these games, the agent learns to explore the environment via natural language interactions with the game simulator. A fundamental challenge in TGs is the efficient exploration of the large action space when the agent has not yet acquired enough knowledge about the environment. We propose CommExpl, an exploration technique that injects external commonsense knowledge, via a pretrained language model (LM), into the agent during training when the agent is the most uncertain about its next action. Our method exhibits improvement on the collected game scores during the training in four out of nine games from Jericho. Additionally, the produced trajectory of actions exhibit lower perplexity, when tested with a pretrained LM, indicating better closeness to human language. | ['Reza Haf', 'Shirui Pan', 'Yunqiu Xu', 'Meng Fang', 'Ehsan Shareghi', 'Dongwon Ryu'] | null | null | null | null | acl-2022-5 | ['text-based-games'] | ['playing-games'] | [-2.26759613e-01 3.47391665e-01 2.13389143e-01 1.13969296e-01
-2.77207285e-01 -8.29168320e-01 8.22977006e-01 -2.48129696e-01
-8.06606591e-01 9.81753707e-01 1.64797097e-01 -5.99346220e-01
-3.84586230e-02 -1.00049174e+00 -6.28514767e-01 -3.03985178e-01
-3.42101306e-01 9.30787861e-01 1.90404788e-01 -9.78720307e-01
9.23408717e-02 1.97410539e-01 -1.17887282e+00 -4.07447591e-02
9.23554897e-01 3.49129796e-01 6.31467283e-01 9.87244010e-01
4.94960696e-02 1.48317504e+00 -6.32971585e-01 -1.05316162e-01
5.01375139e-01 -6.45143926e-01 -1.02103174e+00 -1.83627069e-01
-6.20524108e-01 -4.95267421e-01 -4.58372176e-01 1.36573076e+00
3.70272994e-02 5.12051463e-01 2.43376166e-01 -1.19011641e+00
-1.56804144e-01 1.15289021e+00 -1.12129282e-02 1.27121121e-01
5.65492034e-01 9.52995360e-01 7.93693006e-01 -1.18520319e-01
9.49316502e-01 1.40576506e+00 1.75407812e-01 8.03188026e-01
-9.02522266e-01 -6.50894046e-01 5.74969091e-02 2.95304537e-01
-1.10061395e+00 4.75823618e-02 6.09147191e-01 -2.94054657e-01
1.30114639e+00 -6.49812669e-02 1.01128900e+00 1.40549207e+00
2.96125770e-01 9.43940103e-01 1.41110146e+00 -3.89536351e-01
9.02675569e-01 3.02865088e-01 -6.19401753e-01 6.97800100e-01
-1.58068761e-01 8.09336960e-01 -7.81750381e-01 -8.57928023e-02
8.89448762e-01 -5.98581493e-01 2.06759185e-01 -3.78222317e-01
-1.09784114e+00 1.05089998e+00 1.22785382e-01 1.83215022e-01
-6.72962725e-01 2.15261519e-01 4.95561182e-01 6.22964382e-01
-2.37279870e-02 1.18488550e+00 -4.71439570e-01 -1.09173548e+00
-5.16764760e-01 5.30056298e-01 9.59549129e-01 8.94329011e-01
5.09075940e-01 3.58633280e-01 1.50320277e-01 3.11744511e-01
2.79137611e-01 4.56052065e-01 7.76361585e-01 -1.18545043e+00
4.86407399e-01 7.38037705e-01 2.00989738e-01 -6.44547045e-01
-4.11818683e-01 -3.19534183e-01 -1.56038135e-01 7.26732433e-01
3.02426755e-01 -6.73003912e-01 -8.21603954e-01 2.05378151e+00
2.96527863e-01 1.95949987e-01 7.09907293e-01 6.64117336e-01
3.22379023e-01 5.16472936e-01 3.37635696e-01 -9.77350120e-03
9.82235789e-01 -8.00563216e-01 -3.97541434e-01 -9.59247649e-01
7.78317928e-01 5.92878498e-02 1.26355195e+00 5.87516248e-01
-1.05028784e+00 -3.60985160e-01 -1.02789509e+00 5.32989979e-01
-4.88097280e-01 -3.40174377e-01 8.62551093e-01 5.05308986e-01
-1.22480619e+00 6.15288556e-01 -9.37003493e-01 -5.34567237e-01
1.19257204e-01 4.53298956e-01 -2.56651551e-01 2.86976993e-01
-1.71838260e+00 1.06812704e+00 1.09746599e+00 -3.32546890e-01
-1.48101771e+00 -2.40425766e-01 -1.07360840e+00 -5.61095439e-02
7.64391959e-01 -3.25929284e-01 1.55851626e+00 -7.23896921e-01
-2.24930716e+00 6.74873888e-01 5.15666187e-01 -8.92396569e-01
8.77658188e-01 -7.61685520e-02 -8.57281312e-02 4.95652109e-02
7.65185878e-02 7.66274631e-01 5.48585236e-01 -8.16775620e-01
-5.60599744e-01 -9.49510038e-02 6.12848639e-01 6.62268162e-01
5.60670570e-02 -1.16711035e-01 -2.56407589e-01 -3.47361177e-01
-3.65002930e-01 -1.10091877e+00 -5.95753610e-01 -6.59789383e-01
-3.46306711e-01 -1.73986435e-01 2.50231802e-01 -4.09083486e-01
7.61865377e-01 -1.87824965e+00 3.29528213e-01 2.10056126e-01
5.49853109e-02 2.79554754e-01 -3.57482135e-01 7.41248667e-01
1.85922191e-01 1.24934524e-01 1.05895884e-01 -8.69523063e-02
3.16367567e-01 4.02001381e-01 -2.80665427e-01 -9.52041149e-02
1.17688673e-02 1.09784257e+00 -1.35691082e+00 -2.53446847e-01
4.26182091e-01 -2.06543237e-01 -6.74234927e-01 5.31755567e-01
-7.83208728e-01 7.42720664e-01 -6.84977889e-01 3.18776146e-02
2.76148379e-01 2.18738038e-02 4.15567428e-01 8.25316966e-01
1.08906757e-02 3.58579248e-01 -1.08666182e+00 1.89563298e+00
-4.54564065e-01 4.81269747e-01 -2.66884625e-01 -6.35388494e-01
8.07189941e-01 -1.16562052e-02 1.72703624e-01 -1.05053830e+00
3.58754903e-01 -1.76906064e-01 3.60541493e-01 -5.93104184e-01
5.69709539e-01 -3.61197352e-01 -3.32277924e-01 5.96087754e-01
2.20862418e-01 -5.74361145e-01 4.86440778e-01 5.28672755e-01
1.27546918e+00 2.96936184e-01 5.03618598e-01 -1.17484629e-01
3.16756606e-01 3.61857235e-01 3.34517658e-01 1.35708606e+00
-3.10416430e-01 -2.88335592e-01 7.65130579e-01 -2.33581930e-01
-9.33100760e-01 -9.73125577e-01 7.05118775e-01 1.27570271e+00
5.73291443e-02 -6.16268635e-01 -1.03204989e+00 -4.97611642e-01
-5.15945315e-01 1.33731711e+00 -7.58045435e-01 -5.51057458e-01
-4.02108014e-01 -2.64163107e-01 7.68040478e-01 5.67858458e-01
7.17011154e-01 -1.71991515e+00 -1.17078662e+00 2.89373696e-01
-9.82732028e-02 -1.19853723e+00 2.21191496e-02 3.82537484e-01
-6.68484628e-01 -1.03415799e+00 1.07008509e-01 -5.53879380e-01
1.82088524e-01 -5.22493005e-01 9.93713200e-01 -1.12703629e-01
-9.24973190e-02 3.20821702e-01 -3.40826541e-01 -4.72967654e-01
-9.83098507e-01 -6.53372034e-02 2.54962385e-01 -8.35402727e-01
5.12890399e-01 -6.10771537e-01 -7.78623447e-02 -1.05778441e-01
-6.39459431e-01 4.37227875e-01 4.91401941e-01 9.14377689e-01
-4.54834811e-02 4.68982041e-01 1.60728648e-01 -7.43897021e-01
1.20615220e+00 -4.80973244e-01 -9.02052760e-01 -8.85322914e-02
-2.98547536e-01 3.76659960e-01 7.32302606e-01 -5.67941189e-01
-1.12230468e+00 -2.39203528e-01 -6.29305281e-03 -5.15705422e-02
-3.55853945e-01 6.88741684e-01 1.01347275e-01 1.84238672e-01
8.97296071e-01 5.50539613e-01 6.72419891e-02 3.61376181e-02
3.74046147e-01 2.58103520e-01 4.26054329e-01 -8.63979638e-01
8.16660047e-01 3.57481316e-02 -2.47515738e-01 -7.50418723e-01
-3.65551651e-01 -1.10934578e-01 -4.29345042e-01 -2.25276381e-01
7.22021759e-01 -8.86756837e-01 -1.16977561e+00 5.06340384e-01
-9.06401277e-01 -1.24107969e+00 -4.48094249e-01 6.05674326e-01
-1.06682348e+00 3.75862047e-02 -5.73142529e-01 -1.05448365e+00
-3.52472067e-02 -1.26005566e+00 6.62720740e-01 4.83957499e-01
-5.90929031e-01 -1.09763181e+00 4.49334502e-01 3.07997912e-01
3.59321415e-01 1.86710402e-01 9.32757676e-01 -1.03480017e+00
-3.10286701e-01 7.42265731e-02 3.00689250e-01 7.19603822e-02
1.46411953e-03 -4.12109107e-01 -6.71350300e-01 -2.11582243e-01
1.45587698e-03 -8.72629344e-01 2.01885417e-01 -3.05264909e-02
5.99592626e-01 -1.96079612e-01 4.46817540e-02 3.24330837e-01
1.13340700e+00 6.95316374e-01 5.88196516e-01 9.02365088e-01
4.64664213e-02 2.50815451e-01 7.64391661e-01 8.19612503e-01
3.02019835e-01 3.40519041e-01 6.23474181e-01 3.76843065e-01
3.47115010e-01 -7.63451755e-01 9.09301758e-01 4.77246106e-01
4.32595834e-02 -3.79698575e-01 -1.06236112e+00 4.41281378e-01
-1.78015625e+00 -1.05098426e+00 5.81756353e-01 1.89043331e+00
9.89158988e-01 5.34655869e-01 9.66738239e-02 -3.22160035e-01
2.53031313e-01 3.76311168e-02 -1.05967939e+00 -5.90911090e-01
3.29538472e-02 4.32720929e-01 1.79254100e-01 9.47444141e-01
-7.57350445e-01 1.77870190e+00 6.63465881e+00 6.72528863e-01
-8.97033632e-01 -2.59201527e-01 3.16256344e-01 -1.66766644e-01
-9.45947599e-03 -1.12447686e-01 -4.03016806e-01 2.52383232e-01
1.08216000e+00 -4.28213894e-01 1.00945139e+00 1.08457220e+00
4.55105096e-01 -5.10693967e-01 -1.02395880e+00 6.28730774e-01
-2.12187424e-01 -1.16691208e+00 -1.48577064e-01 1.71128258e-01
5.12171268e-01 5.41465044e-01 1.60556510e-01 1.07401216e+00
1.26917720e+00 -1.23505783e+00 7.87441671e-01 2.33388260e-01
4.39001381e-01 -1.03546596e+00 6.50052369e-01 1.01250648e+00
-6.56857908e-01 -2.56696373e-01 -3.30050766e-01 -6.69496953e-01
-1.55922771e-01 -4.78344232e-01 -1.58196723e+00 5.43659478e-02
3.60943317e-01 1.34471163e-01 -5.28629363e-01 5.67272663e-01
-7.63058901e-01 7.18917131e-01 -3.97540420e-01 -5.99072874e-01
8.51765156e-01 -4.24188286e-01 7.76768148e-01 8.42432380e-01
1.12387940e-01 3.51481020e-01 5.48498929e-01 1.20700216e+00
1.44133449e-01 -8.19393247e-02 -8.48382294e-01 -5.33996105e-01
3.62844348e-01 7.98202157e-01 -5.79180300e-01 -3.02918106e-01
2.12782592e-01 9.40089583e-01 4.51933771e-01 3.40089083e-01
-7.61771619e-01 -2.76456416e-01 7.81712592e-01 -3.20976466e-01
-2.25191358e-02 -4.05415148e-01 3.32343616e-02 -9.40458477e-01
-3.47020388e-01 -1.37416530e+00 2.35569358e-01 -8.99969459e-01
-5.43489516e-01 7.53693342e-01 -5.93547933e-02 -7.87967741e-01
-1.04567146e+00 -7.67123520e-01 -7.16751814e-01 7.26635575e-01
-1.02949679e+00 -6.69605136e-01 -6.71149045e-02 6.96312547e-01
7.76175201e-01 -6.39546454e-01 9.71058309e-01 -5.53684890e-01
-2.06221789e-01 2.09343687e-01 -9.59852189e-02 2.34339461e-01
1.09555550e-01 -1.48379254e+00 6.37335241e-01 7.01647699e-01
7.46558383e-02 5.73076606e-01 1.19011116e+00 -8.75000060e-01
-1.38069999e+00 -7.42334545e-01 1.28359109e-01 -3.30887109e-01
1.11435235e+00 -4.78660047e-01 -6.07501924e-01 9.02029812e-01
3.69623721e-01 -5.70544958e-01 4.58995163e-01 5.37079899e-03
8.09681788e-03 5.70061624e-01 -1.16127169e+00 1.10322046e+00
9.41398680e-01 -6.65471554e-01 -9.39073861e-01 4.01000798e-01
8.44843030e-01 -6.73414588e-01 -4.03060824e-01 -1.66660011e-01
7.37304240e-02 -8.47660482e-01 5.62726021e-01 -1.26270986e+00
3.11398327e-01 -1.35188729e-01 -9.14292037e-02 -1.82940519e+00
-9.84612852e-02 -1.07990015e+00 6.79331124e-02 5.70782125e-01
3.81519288e-01 -4.57483917e-01 1.09568238e+00 7.06298709e-01
3.07285190e-01 -3.83640736e-01 -7.83236265e-01 -9.04498279e-01
3.52570504e-01 -8.29246581e-01 5.27230203e-01 6.75465524e-01
7.26704597e-01 3.00840080e-01 -3.85610908e-01 7.20344484e-02
3.75676721e-01 -2.93023229e-01 8.97916973e-01 -8.61303389e-01
-6.71526611e-01 -3.26306880e-01 -3.47163051e-01 -9.91869330e-01
4.11114037e-01 -7.61202216e-01 4.87221211e-01 -1.23886526e+00
7.86945075e-02 -1.30922154e-01 -1.00440770e-01 5.53261578e-01
5.83075732e-03 -3.78048331e-01 3.45603913e-01 -4.19312060e-01
-8.69192302e-01 7.61292040e-01 1.33429539e+00 -1.18591048e-01
-5.36712646e-01 -2.14837223e-01 -6.28640115e-01 1.05564058e+00
1.09066641e+00 -3.77095222e-01 -8.51219118e-01 -2.23882392e-01
5.94318032e-01 2.74236560e-01 9.45503861e-02 -1.16641915e+00
4.88008350e-01 -7.64750361e-01 -4.37617563e-02 -1.44357264e-01
4.50190097e-01 -4.72881973e-01 -7.27589652e-02 7.94549346e-01
-5.72341204e-01 1.09403357e-01 6.40420139e-01 4.20904458e-01
1.45778432e-02 -1.93514064e-01 6.02697432e-01 -6.23195410e-01
-1.27852619e+00 1.13838896e-01 -1.10339451e+00 5.65059185e-01
1.18070340e+00 -1.09085806e-01 7.52509087e-02 -7.97627747e-01
-9.48129594e-01 5.63123345e-01 3.71234864e-01 5.77162445e-01
6.11412585e-01 -1.03009975e+00 -5.63930511e-01 1.31024674e-01
4.07553874e-02 -1.74864486e-01 1.14408486e-01 2.84878433e-01
-6.86875880e-01 3.53123695e-01 -4.20078993e-01 -1.52737498e-01
-9.97835517e-01 4.50254232e-01 6.31921649e-01 -6.53364182e-01
-5.56389928e-01 9.07555223e-01 2.67361909e-01 -8.58374834e-01
1.42671034e-01 -2.96940148e-01 -2.65373915e-01 -4.65799361e-01
6.89582884e-01 6.87512830e-02 -2.66003519e-01 -2.34058112e-01
-2.10086405e-01 -1.58093318e-01 -1.20518282e-01 -7.03957379e-01
1.48049784e+00 7.22011849e-02 2.54321694e-02 3.69339615e-01
6.08846188e-01 -2.17517912e-01 -1.39520109e+00 -2.15114161e-01
1.97644114e-01 -2.26788491e-01 -3.07743605e-02 -1.19899404e+00
-4.93011922e-01 5.49324453e-01 2.79093504e-01 5.81372418e-02
7.15400100e-01 -1.85889930e-01 3.63881916e-01 9.96823132e-01
8.28737259e-01 -1.60613453e+00 4.25042510e-01 1.21827352e+00
7.86645293e-01 -1.14054358e+00 -4.43782091e-01 2.51583487e-01
-1.17570758e+00 1.00166905e+00 9.49231207e-01 -1.23953849e-01
2.68595457e-01 3.68071496e-01 1.99670106e-01 -3.70012611e-01
-1.14413524e+00 -1.83908314e-01 -4.10056323e-01 9.61605489e-01
-2.35459536e-01 2.78741032e-01 3.48720104e-01 5.05089700e-01
-9.07930732e-01 -3.52402404e-02 8.18709075e-01 7.73807466e-01
-5.93885660e-01 -9.79028046e-01 -1.52474061e-01 -1.16314273e-02
8.03542137e-02 -1.83087394e-01 -6.90419495e-01 1.01393580e+00
1.18154593e-01 1.04265773e+00 -2.32096002e-01 -5.19704998e-01
2.11173326e-01 1.44335181e-02 4.28461015e-01 -8.31330419e-01
-5.67894578e-01 -3.51433486e-01 1.79855078e-01 -9.90950584e-01
2.53161222e-01 -7.52241731e-01 -1.68732965e+00 -2.65548140e-01
4.69695573e-04 4.67089206e-01 4.63003397e-01 1.36030257e+00
-6.18997670e-04 5.61897516e-01 3.41868967e-01 -4.72056776e-01
-7.77607560e-01 -9.14384246e-01 -6.07903779e-01 2.73899645e-01
7.97766075e-02 -5.24502039e-01 -9.48161855e-02 -4.66286242e-01] | [3.827516555786133, 1.4408315420150757] |
4d632a9f-df03-4715-9ddc-0837c1f26253 | a-deep-learning-based-and-fully-automated | 2302.10634 | null | https://arxiv.org/abs/2302.10634v1 | https://arxiv.org/pdf/2302.10634v1.pdf | A Deep Learning-Based and Fully Automated Pipeline for Regurgitant Mitral Valve Anatomy Analysis from 3D Echocardiography | 3D transesophageal echocardiography (3DTEE), is the recommended method for diagnosing mitral regurgitation (MR). 3DTEE provides a high-quality 3D image of the mitral valve (MV), allowing for precise segmentation and measurement of the regurgitant valve anatomy. However, manual TEE segmentations are time-consuming and prone to intra-operator variability, affecting the reliability of the measurements. To address this, we developed a fully automated pipeline using a 3D convolutional neural network (CNN) to segment MV substructures (annulus, anterior leaflet, and posterior leaflet) and quantify MV anatomy. The 3D CNN, based on a multi-decoder residual U-Net architecture, was trained and tested on a dataset comprising 100 3DTEE images with corresponding segmentations. Within the pipeline, a custom algorithm refines the CNN-based segmentations and extracts MV models, from which anatomical landmarks and features are quantified. The accuracy of the proposed method was assessed using Dice score and mean surface distance (MSD) against ground truth segmentations, and the extracted anatomical parameters were compared against a semiautomated commercial software TomTec Image Arena. The trained 3D CNN achieved an average Dice score of 0.79 and MSD of 0.47 mm for the combined segmentation of the annulus, anterior and posterior leaflet. The proposed CNN architecture outperformed a baseline residual U-Net architecture in MV substructure segmentation, and the refinement of the predicted annulus segmentation improved MSD by 8.36%. The annular and leaflet linear measurements differed by less than 7.94 mm and 3.67 mm, respectively, compared to the 3D measurements obtained with TomTec Image Arena. The proposed pipeline was faster than the commercial software, with a modeling time of 12.54 s and a quantification time of 54.42 s. | ['Emiliano Votta', 'Alberto Redaelli', 'Eustachio Agricola', 'Francesco Maisano', 'Paolo Denti', 'Giacomo Ingallina', 'Simone Saitta', 'Riccardo Munafò'] | 2023-02-21 | null | null | null | null | ['anatomy'] | ['miscellaneous'] | [-4.19925302e-01 2.00540960e-01 3.45081776e-01 2.30834018e-02
-7.51445472e-01 -1.11423326e+00 -1.23390198e-01 1.14971295e-01
-3.91932577e-01 4.62207854e-01 -2.37884179e-01 -7.75469720e-01
3.04404851e-02 -4.91939217e-01 -3.15522045e-01 -5.03856719e-01
-6.01033688e-01 6.93249285e-01 2.11821899e-01 2.78006643e-01
-5.72899990e-02 9.13466871e-01 -6.09650373e-01 -1.95014402e-01
1.04094827e+00 1.30149209e+00 8.59372988e-02 1.10813212e+00
1.23953857e-01 2.41016418e-01 -8.86873841e-01 4.05276641e-02
5.06813407e-01 -5.00676990e-01 -5.24593413e-01 -6.46628737e-02
7.75995374e-01 -6.56083584e-01 1.70660708e-02 7.66908050e-01
9.82531250e-01 -3.17151457e-01 4.42603469e-01 -2.27956623e-01
-4.50453967e-01 2.77676582e-01 -2.20942989e-01 4.00481045e-01
-3.61253977e-01 3.09684593e-02 5.66164255e-01 -8.28336120e-01
7.12692380e-01 5.78653634e-01 1.18623686e+00 2.56616086e-01
-9.93795156e-01 -2.91826606e-01 -8.36226046e-01 -6.71625972e-01
-1.23448026e+00 1.03950359e-01 6.31499290e-01 -9.35597241e-01
8.25588405e-01 3.04700762e-01 1.17505217e+00 -4.39273156e-02
6.85664237e-01 2.46374875e-01 1.16869116e+00 -6.09435029e-02
-5.97674549e-02 -1.40080631e-01 -2.59856880e-01 1.06618893e+00
4.36297864e-01 1.13550588e-01 4.59716648e-01 -2.01259971e-01
1.42108989e+00 -2.59732515e-01 -5.07718742e-01 -4.41919357e-01
-1.24183810e+00 4.43065137e-01 3.54394585e-01 6.57205999e-01
-3.93821865e-01 -2.42036492e-01 5.97446263e-01 1.18517883e-01
3.62809479e-01 7.99554348e-01 -5.88456035e-01 -2.52664298e-01
-1.31361544e+00 1.22750051e-01 8.57512057e-01 5.85107565e-01
1.16107419e-01 3.20813358e-01 -3.07666898e-01 7.02889025e-01
3.56192946e-01 6.22407377e-01 4.48023230e-01 -1.41876829e+00
1.53968811e-01 6.45471096e-01 8.30564573e-02 -1.16437995e+00
-6.36135459e-01 -9.94887292e-01 -1.18767905e+00 4.86090362e-01
7.36162901e-01 -4.18919504e-01 -1.07834601e+00 1.11593926e+00
2.05324173e-01 -1.14271991e-01 -2.25744799e-01 1.49356735e+00
1.00575233e+00 1.53054878e-01 -2.00361490e-01 -3.53436738e-01
1.09378910e+00 -8.53200197e-01 -6.36009037e-01 1.70995519e-01
1.10168362e+00 -8.06874037e-01 8.10951829e-01 5.93867712e-02
-1.44086230e+00 -7.39216447e-01 -1.20232379e+00 4.53190804e-02
1.88401595e-01 6.59559309e-01 1.36459917e-01 7.79099464e-01
-1.30826437e+00 1.23436284e+00 -1.11964905e+00 2.60087639e-01
6.44243300e-01 3.03391397e-01 -4.80472177e-01 6.18434906e-01
-9.94261086e-01 1.09815907e+00 1.55979171e-01 5.00140488e-01
-8.33320558e-01 -1.00363410e+00 -8.85890484e-01 2.01291233e-01
-6.48821890e-02 -8.02202821e-01 9.50539470e-01 -4.47463632e-01
-1.54785526e+00 1.37634647e+00 3.97943437e-01 -5.53573132e-01
6.62187636e-01 -4.49113660e-02 -2.46676020e-02 3.28459829e-01
-1.80308986e-02 1.41893988e-02 3.39530200e-01 -9.40514982e-01
-2.76477665e-01 -5.31665802e-01 -4.01683927e-01 -5.99121340e-02
2.09689915e-01 -1.57912835e-01 -4.31333482e-01 -7.07313240e-01
4.38867480e-01 -9.69785750e-01 -3.84805024e-01 2.48066127e-01
-1.91811115e-01 4.91465181e-01 4.90117550e-01 -1.49357080e+00
1.24146962e+00 -1.97292626e+00 2.56216079e-02 3.64128798e-01
1.01729298e+00 1.00767362e+00 5.47959566e-01 -3.96089852e-01
1.17572891e-02 3.44603151e-01 -7.35414088e-01 -2.14193568e-01
-5.41872203e-01 -6.32591620e-02 4.99787807e-01 6.36470079e-01
-3.29310924e-01 9.87871587e-01 -6.91301525e-01 -6.50286436e-01
4.05490518e-01 4.85797882e-01 -3.52215528e-01 4.12225217e-01
3.04701656e-01 7.19142437e-01 -1.59156576e-01 6.95653021e-01
7.87324071e-01 -2.79525220e-01 5.60915232e-01 -3.56723785e-01
-2.38306910e-01 -1.25882417e-01 -8.80150497e-01 1.99127507e+00
-4.05707479e-01 6.32876098e-01 2.60035187e-01 -7.60894716e-01
1.40548921e+00 6.47825837e-01 7.83786297e-01 -3.27742308e-01
6.05955422e-01 8.04382622e-01 6.35889590e-01 -4.31722790e-01
-1.45570233e-01 -5.49789518e-02 1.47609860e-01 3.12809497e-01
1.62049100e-01 -2.45485798e-01 -1.18207760e-01 -1.80026039e-01
7.87946105e-01 3.53001177e-01 3.68919015e-01 -6.29065454e-01
5.90510607e-01 3.20823528e-02 9.12034988e-01 6.23418450e-01
-6.17249489e-01 1.10210741e+00 7.03557312e-01 -9.95538116e-01
-1.28872693e+00 -1.18764520e+00 -3.37598830e-01 -4.10633713e-01
4.11674082e-02 -5.13482571e-01 -9.52198744e-01 -6.46386564e-01
2.72641070e-02 -1.28840297e-01 -3.31010371e-01 1.91392988e-01
-8.91607106e-01 -4.71135259e-01 7.88639784e-01 5.69907844e-01
4.95984256e-01 -8.71967912e-01 -1.04467809e+00 4.09644812e-01
5.75235784e-02 -1.08982670e+00 -5.34131408e-01 -3.30957770e-01
-1.53400803e+00 -1.23172605e+00 -1.05676019e+00 -7.76250958e-01
6.72899961e-01 -5.09061515e-01 1.31200135e+00 3.78440678e-01
-5.83762288e-01 -1.01598665e-01 1.35630414e-01 -5.26885800e-02
-6.27702296e-01 -5.26376553e-02 -1.85034238e-02 -3.19051802e-01
-4.22323942e-01 -5.02940238e-01 -9.97306466e-01 3.19909453e-02
-3.84345025e-01 3.88797857e-02 5.26649475e-01 7.71521389e-01
9.87765312e-01 -6.20685995e-01 3.18133920e-01 -8.62634599e-01
6.18215024e-01 -4.84418906e-02 -8.52257133e-01 3.14501636e-02
-9.64728177e-01 -5.24352074e-01 7.06202149e-01 1.41563807e-02
-7.05367327e-01 -2.21569166e-02 -2.06151813e-01 -9.56670880e-01
-5.27206734e-02 4.76947904e-01 2.71486044e-01 -2.27214113e-01
6.10225618e-01 -4.44891714e-02 5.85207582e-01 -4.22336191e-01
1.13004029e-01 6.67147756e-01 7.00662851e-01 -3.46733510e-01
4.72502857e-01 2.43898764e-01 2.69702673e-01 -4.81176764e-01
-5.44207871e-01 -7.94759095e-02 -1.07307255e+00 -4.79656816e-01
1.09956288e+00 -5.35082459e-01 -4.99756157e-01 3.44700277e-01
-1.12294590e+00 -3.25834036e-01 -4.13956165e-01 6.48965716e-01
-6.42423391e-01 6.72001719e-01 -9.05992448e-01 -4.00045574e-01
-1.22742331e+00 -1.37529302e+00 6.65286362e-01 2.03032672e-01
-2.35071450e-01 -1.26506698e+00 -6.81804940e-02 2.62391806e-01
6.13208413e-01 7.97443509e-01 7.32715011e-01 -6.12867296e-01
-3.28276485e-01 -3.98086697e-01 -1.80486739e-01 9.32080388e-01
2.72749305e-01 -4.67806496e-02 -6.72602296e-01 -2.34581783e-01
1.40921369e-01 3.69628668e-01 2.87257165e-01 1.00012577e+00
1.10411561e+00 1.35282636e-01 -1.04650937e-01 1.04631364e+00
1.24658597e+00 3.56510252e-01 2.81647652e-01 -4.76162275e-03
8.22536051e-01 -5.19368192e-03 6.04853451e-01 2.96483219e-01
6.78971410e-02 3.86530995e-01 3.44225377e-01 -6.63288355e-01
-3.61494333e-01 1.34375185e-01 -3.33448499e-01 1.26561201e+00
-5.99833846e-01 4.90773588e-01 -1.17500436e+00 6.18182898e-01
-1.34621847e+00 -3.44097257e-01 -3.38887185e-01 2.39221644e+00
8.71003270e-01 1.78887144e-01 -1.62547156e-01 -9.70003977e-02
4.56721038e-01 9.68235824e-03 -2.06045613e-01 -5.85399926e-01
-4.85386848e-02 3.81179273e-01 5.16290367e-01 6.47260487e-01
-1.10405302e+00 6.72575772e-01 5.80269670e+00 8.72028619e-02
-1.34697032e+00 5.28302453e-02 7.39716709e-01 3.01272303e-01
1.85555264e-01 1.99166331e-02 -3.31332147e-01 3.65053296e-01
7.93684721e-01 -1.85157880e-02 2.81446148e-02 6.00793481e-01
3.77439290e-01 2.50338137e-01 -6.19238257e-01 7.76407063e-01
-3.86293195e-02 -1.80479085e+00 -4.95433599e-01 4.15232070e-02
5.91835916e-01 1.31170908e-02 -3.97781730e-01 -8.48626345e-03
-6.66733384e-01 -9.23829019e-01 2.14299664e-01 8.25762153e-01
1.25520837e+00 -2.83276767e-01 1.20243609e+00 3.00341457e-01
-1.10456538e+00 3.14209461e-01 -4.73816954e-02 2.22882330e-01
2.22794935e-01 6.78643584e-01 -1.08480811e+00 4.78260994e-01
9.19020832e-01 6.83096409e-01 -2.42456034e-01 1.13862336e+00
1.51576698e-01 6.03664815e-01 -2.19092548e-01 5.31156063e-01
4.00626659e-02 -6.52719736e-01 9.87972558e-01 9.64809179e-01
4.85904813e-01 -7.31568709e-02 7.54556581e-02 1.32828677e+00
2.21920130e-03 2.60122567e-01 -3.88091117e-01 7.35241175e-02
4.69661534e-01 1.63334668e+00 -8.76887858e-01 -5.39889336e-01
-2.27314085e-01 6.12863839e-01 -4.70935255e-02 7.87904486e-02
-6.15530133e-01 -4.22928184e-01 3.19799066e-01 3.98426026e-01
3.50696780e-02 -3.82669359e-01 -7.92964041e-01 -8.76264811e-01
2.61383414e-01 -7.55750656e-01 3.42094034e-01 -5.60761392e-01
-7.55692899e-01 8.51787329e-01 -4.52956468e-01 -1.32990837e+00
-8.29136074e-02 -7.80533791e-01 -6.00540102e-01 1.45600557e+00
-1.10711861e+00 -8.12013984e-01 -3.57070446e-01 -2.19779208e-01
2.45432779e-01 -3.77983451e-01 1.09678781e+00 4.67248082e-01
-2.75500298e-01 2.80145377e-01 -1.89664140e-01 5.60016096e-01
5.40324569e-01 -1.83504379e+00 3.65584999e-01 7.99007475e-01
-3.48010510e-01 7.44759142e-01 1.95544675e-01 -8.57956946e-01
-1.11002219e+00 -1.15639842e+00 9.02052224e-01 -3.35424960e-01
1.87820762e-01 1.55185774e-01 -8.64336610e-01 5.67434907e-01
-1.62794143e-01 6.54243469e-01 6.62804186e-01 -4.09596354e-01
2.66280830e-01 3.17111649e-02 -1.17068493e+00 3.97149324e-01
6.40438378e-01 -2.20978841e-01 -4.52233732e-01 -4.49119098e-02
2.96234012e-01 -1.32543671e+00 -2.00280237e+00 7.84918785e-01
8.22207212e-01 -1.02257895e+00 1.05120301e+00 -4.34926808e-01
3.86922807e-01 -5.57924449e-01 4.42474931e-01 -1.03262961e+00
-2.56435424e-01 -6.05902672e-01 -4.02596503e-01 5.96934676e-01
4.76740718e-01 -4.19525683e-01 6.53451324e-01 3.63060027e-01
-8.44651341e-01 -1.25224817e+00 -9.77822483e-01 -2.16337934e-01
3.39428723e-01 -1.26873076e-01 1.04258910e-01 7.95429826e-01
-4.48498547e-01 -1.73513353e-01 -1.02911621e-01 2.58294910e-01
6.18486643e-01 2.80583918e-01 3.62511665e-01 -1.45857465e+00
-8.94736871e-02 -5.34798324e-01 -4.22533691e-01 -7.91932940e-01
-1.52086928e-01 -1.18043005e+00 -2.74769485e-01 -1.63206840e+00
-4.03368801e-01 -5.84231138e-01 -2.09575161e-01 1.34548783e-01
-1.85015842e-01 3.39561552e-01 1.71286047e-01 3.77173871e-01
2.14989316e-02 8.92324001e-02 2.03480101e+00 2.30178222e-01
-4.68363851e-01 1.90250814e-01 -2.27716878e-01 7.91482270e-01
9.55648422e-01 -2.33269557e-01 -8.57154578e-02 -2.95547158e-01
-2.02013105e-01 7.63771176e-01 4.17470694e-01 -1.03560495e+00
-2.65096463e-02 8.35641742e-01 7.46650040e-01 -6.94091499e-01
-7.32957646e-02 -7.70558774e-01 3.45508792e-02 7.56047189e-01
-6.30513579e-02 -6.41994970e-03 2.83139944e-01 -2.41418332e-01
-2.27630958e-01 -2.37883292e-02 9.66849148e-01 -3.78043085e-01
5.32070622e-02 6.71301007e-01 -4.24357057e-01 2.34429643e-01
8.28855157e-01 -4.33632433e-01 2.21301064e-01 5.67863062e-02
-1.25859368e+00 -7.56954029e-02 5.36331721e-02 1.50680110e-01
7.90160179e-01 -1.06533813e+00 -7.14466870e-01 4.36513990e-01
-4.93205905e-01 4.66662437e-01 4.99457031e-01 1.50036848e+00
-1.40693510e+00 3.27735960e-01 -2.93169588e-01 -1.01467264e+00
-1.10650980e+00 1.86509117e-02 1.48351479e+00 -2.63307899e-01
-1.15444958e+00 4.42632705e-01 -4.55199063e-01 -7.02150941e-01
-1.73245929e-02 -6.88389242e-01 -2.90122151e-01 -4.64969516e-01
4.49682735e-02 4.33847368e-01 1.41123399e-01 -4.78265554e-01
-7.47994781e-02 8.89787853e-01 4.68568265e-01 4.01445478e-01
8.65474463e-01 -1.34484738e-01 -4.22669768e-01 1.92112058e-01
9.08280373e-01 1.94499083e-02 -1.00095177e+00 -2.59156767e-02
-1.30836502e-01 -4.21754926e-01 4.60681230e-01 -9.82601881e-01
-1.47528982e+00 9.78324711e-01 1.04099905e+00 1.28164649e-01
1.03212368e+00 -2.92440504e-01 1.04447019e+00 -3.44619066e-01
-1.60427820e-02 -6.11977458e-01 -7.82558262e-01 2.71287322e-01
7.65885770e-01 -1.10295093e+00 -1.28448591e-01 -3.90288264e-01
-3.52039665e-01 1.48917866e+00 5.87086380e-01 -3.11970741e-01
8.52547705e-01 5.25364578e-01 8.11285317e-01 -3.10870200e-01
1.12052917e-01 3.76326084e-01 7.17197120e-01 3.79060745e-01
8.83655131e-01 1.59695700e-01 -6.01422668e-01 9.00827646e-01
-5.90958148e-02 2.09792972e-01 3.57033283e-01 6.70847714e-01
-2.72056729e-01 -6.48473799e-01 -8.18030164e-02 5.83093464e-01
-9.40021813e-01 -1.27334327e-01 5.44685908e-02 7.79213369e-01
4.83132660e-01 2.90456533e-01 1.35226369e-01 -2.48742290e-02
3.81134510e-01 -1.76254317e-01 3.02614123e-01 -4.75740641e-01
-1.04120374e+00 4.40825999e-01 7.09940400e-03 -4.28026527e-01
-1.40879855e-01 -4.26998079e-01 -1.47843242e+00 9.57393423e-02
-1.70270484e-02 2.49086410e-01 6.49937391e-01 8.67199779e-01
5.45330524e-01 7.66230941e-01 5.15088260e-01 -5.66905200e-01
-4.54770178e-01 -1.03449035e+00 -7.68435061e-01 1.42658114e-01
3.04735094e-01 -3.54662150e-01 -2.99524158e-01 1.71269000e-01] | [14.142598152160645, -2.504316568374634] |
a4a1191b-0d6f-4902-b8b5-3049ba81d107 | improving-text-to-sql-semantic-parsing-with | 2209.14415 | null | https://arxiv.org/abs/2209.14415v1 | https://arxiv.org/pdf/2209.14415v1.pdf | Improving Text-to-SQL Semantic Parsing with Fine-grained Query Understanding | Most recent research on Text-to-SQL semantic parsing relies on either parser itself or simple heuristic based approach to understand natural language query (NLQ). When synthesizing a SQL query, there is no explicit semantic information of NLQ available to the parser which leads to undesirable generalization performance. In addition, without lexical-level fine-grained query understanding, linking between query and database can only rely on fuzzy string match which leads to suboptimal performance in real applications. In view of this, in this paper we present a general-purpose, modular neural semantic parsing framework that is based on token-level fine-grained query understanding. Our framework consists of three modules: named entity recognizer (NER), neural entity linker (NEL) and neural semantic parser (NSP). By jointly modeling query and database, NER model analyzes user intents and identifies entities in the query. NEL model links typed entities to schema and cell values in database. Parser model leverages available semantic information and linking results and synthesizes tree-structured SQL queries based on dynamically generated grammar. Experiments on SQUALL, a newly released semantic parsing dataset, show that we can achieve 56.8% execution accuracy on WikiTableQuestions (WTQ) test set, which outperforms the state-of-the-art model by 2.7%. | ['Sudipta Sengupta', 'Bing Xiang', 'Ramesh Nallapati', 'Zhiguo Wang', 'Jiarong Jiang', 'Alexander Hanbo Li', 'Patrick Ng', 'Jun Wang'] | 2022-09-28 | null | null | null | null | ['text-to-sql'] | ['computer-code'] | [ 5.19564040e-02 4.28760916e-01 -4.60551441e-01 -7.92703629e-01
-1.04604876e+00 -1.02290452e+00 6.94396487e-03 4.97884989e-01
-2.80010015e-01 4.88920063e-01 2.20560312e-01 -6.62997305e-01
2.09759563e-01 -1.54740548e+00 -1.25292814e+00 5.03379285e-01
3.42593729e-01 6.64323807e-01 7.06546962e-01 -4.13151860e-01
1.57862797e-01 7.39880130e-02 -1.56559014e+00 9.56120253e-01
8.65916908e-01 1.22068810e+00 6.52793050e-02 6.73230112e-01
-1.14088953e+00 1.05950773e+00 -6.83652222e-01 -5.48641562e-01
-2.51534320e-02 5.60051389e-02 -1.33526766e+00 -8.44752312e-01
2.78144419e-01 -9.46554616e-02 1.79321691e-01 9.27677631e-01
3.56025994e-01 -2.53821969e-01 -1.22458842e-02 -1.01581335e+00
-5.30242324e-01 1.04833806e+00 4.25895005e-02 -6.92737848e-02
5.40585577e-01 -2.02179074e-01 1.45018077e+00 -6.45486712e-01
8.35420489e-01 1.32774222e+00 5.47956288e-01 5.02380967e-01
-1.02468753e+00 -4.89559859e-01 6.85163960e-02 -2.69169718e-01
-1.16802478e+00 -2.16546714e-01 3.43123704e-01 -8.25416371e-02
1.82160246e+00 2.88957149e-01 6.44597933e-02 5.97285688e-01
3.57282683e-02 6.42892838e-01 6.87192202e-01 -3.62458289e-01
4.34974074e-01 2.40443230e-01 6.35710776e-01 9.50594306e-01
8.51091892e-02 -3.03227101e-02 -6.14887238e-01 -3.85237545e-01
5.01312494e-01 -3.52778107e-01 3.11450571e-01 -2.95398653e-01
-6.65303409e-01 9.33275402e-01 4.78940040e-01 -6.32629767e-02
-4.38062638e-01 2.97983050e-01 8.63068342e-01 3.03557366e-01
-1.01944752e-01 6.90185308e-01 -9.38586295e-01 -1.84094772e-01
-5.41612625e-01 2.15836078e-01 1.29466450e+00 1.31560302e+00
9.76802766e-01 -1.95236057e-01 -6.14707917e-02 8.44208837e-01
2.99229920e-01 5.10249615e-01 4.30302978e-01 -1.04126024e+00
7.88371384e-01 1.28724873e+00 -2.72352725e-01 -7.18309283e-01
-3.23726505e-01 -1.97688073e-01 -4.97950055e-03 -2.17512622e-01
2.31102392e-01 8.34531412e-02 -8.72474611e-01 1.83992887e+00
1.90948963e-01 -2.90883213e-01 6.02278650e-01 5.19058406e-01
1.21623218e+00 5.65403342e-01 6.10261023e-01 4.23406154e-01
1.79558730e+00 -6.03941143e-01 -3.84883791e-01 -3.46167713e-01
1.04200828e+00 -4.50223655e-01 1.39911866e+00 7.08786175e-02
-1.13876736e+00 -5.86268187e-01 -9.61567461e-01 -4.44751561e-01
-1.07049811e+00 -2.53428817e-01 8.47727954e-01 7.59538174e-01
-9.67496157e-01 1.02437086e-01 -9.12642241e-01 -8.03501844e-01
2.98935741e-01 4.93024230e-01 -2.19129831e-01 1.36554182e-01
-1.31799293e+00 3.82767975e-01 1.18640184e+00 -5.47636807e-01
-5.08767486e-01 -1.04593110e+00 -1.00773096e+00 2.93681055e-01
9.35487449e-01 -1.03090751e+00 1.50750530e+00 -4.67937589e-01
-1.21246433e+00 9.46628690e-01 -4.87429470e-01 -6.16913557e-01
-9.27767828e-02 -2.41712153e-01 -4.68629509e-01 9.75257829e-02
4.12186801e-01 8.53358388e-01 6.04588799e-02 -1.10650957e+00
-9.15014446e-01 -6.35034084e-01 4.76776302e-01 4.60688360e-02
1.72313228e-01 1.76974848e-01 -9.10383582e-01 -3.69156271e-01
1.11200422e-01 -5.11573553e-01 1.82500705e-01 -4.30860460e-01
-6.12301350e-01 -4.76626247e-01 4.41248387e-01 -5.92152953e-01
1.60919547e+00 -1.91341579e+00 -4.52607453e-01 2.73621112e-01
5.38977571e-02 1.24707408e-02 -2.66594123e-02 6.62542641e-01
-3.09175719e-02 4.86040980e-01 -2.12799639e-01 3.32009435e-01
2.80667216e-01 4.86925095e-01 -7.88583159e-01 -7.80276299e-01
4.46009994e-01 1.41851556e+00 -6.45976067e-01 -7.02829063e-01
-1.66736916e-01 -6.69590682e-02 -1.09376729e+00 4.25870538e-01
-9.93854940e-01 -4.05272305e-01 -9.33618367e-01 1.12963521e+00
5.19833744e-01 -4.11763102e-01 3.23141932e-01 -5.41134775e-01
2.53815472e-01 7.45222449e-01 -1.03872311e+00 2.03547430e+00
-6.73949718e-01 -1.36202902e-01 -2.17562735e-01 -9.36329007e-01
9.61724937e-01 1.27609193e-01 7.36246407e-02 -1.03823853e+00
-4.94834244e-01 4.73848253e-01 -7.10080743e-01 -6.04754865e-01
5.93254089e-01 2.38736402e-02 -8.07320178e-01 3.57691139e-01
1.35972351e-01 1.05188359e-02 4.10085693e-02 4.34869587e-01
1.40762508e+00 4.28845882e-01 3.95151466e-01 -1.04601659e-01
5.00096738e-01 4.75074559e-01 5.23848236e-01 1.01392543e+00
2.39269704e-01 1.69346318e-01 7.78110683e-01 -4.07168537e-01
-8.35510135e-01 -1.28379905e+00 2.60553509e-01 1.71707439e+00
5.39590679e-02 -7.13608205e-01 -1.08065760e+00 -1.05568337e+00
1.48908913e-01 1.05778623e+00 -2.77865171e-01 -1.03406996e-01
-9.52081978e-01 -4.90845412e-01 9.84912157e-01 9.57298517e-01
6.28450155e-01 -1.33204150e+00 -8.66833925e-01 4.87233549e-01
-1.99299395e-01 -1.40583968e+00 -5.73951863e-02 4.23200428e-01
-9.17096078e-01 -1.21544838e+00 4.05358255e-01 -6.45644844e-01
1.89439461e-01 -3.64866853e-01 1.91227317e+00 8.07672516e-02
-1.76972896e-01 4.01242465e-01 -1.95518449e-01 -4.57744062e-01
-6.49535954e-01 5.68931460e-01 -7.71602571e-01 -6.80346131e-01
9.09953952e-01 -1.54221386e-01 -3.05510104e-01 1.94674090e-01
-1.16947389e+00 -8.99584591e-03 4.61352140e-01 6.10010087e-01
8.07101548e-01 -1.47536835e-02 7.35295415e-01 -1.55706835e+00
5.62780917e-01 -5.05940855e-01 -8.02868903e-01 7.65806198e-01
-6.85144603e-01 6.99401796e-01 8.19131792e-01 2.93163896e-01
-1.34719741e+00 1.91799477e-02 -3.40170294e-01 1.66633293e-01
-4.25579935e-01 7.84091711e-01 -4.45005804e-01 4.55039829e-01
7.62457430e-01 1.50305316e-01 -2.08236545e-01 -4.21958923e-01
6.19384527e-01 5.15489519e-01 9.28403139e-01 -1.23028660e+00
4.57446992e-01 2.35700101e-01 -3.02981168e-01 -2.52604455e-01
-8.82091403e-01 -2.73934066e-01 -2.90034682e-01 5.72946906e-01
1.09555840e+00 -8.89103293e-01 -1.03044689e+00 4.29953821e-02
-1.17776179e+00 -2.91949630e-01 -4.10659432e-01 -1.25852361e-01
-6.59255683e-01 -4.50365543e-02 -6.96373343e-01 -4.42850232e-01
-4.32680786e-01 -1.10731351e+00 1.37027383e+00 1.29745200e-01
-3.41324806e-01 -9.62448835e-01 -2.97282845e-01 6.37449443e-01
4.84104246e-01 1.12458058e-01 1.71777797e+00 -1.22443008e+00
-8.92829895e-01 -1.50677124e-02 -5.80359578e-01 9.39892232e-02
-1.56677693e-01 -2.77901739e-01 -8.59821618e-01 3.21038187e-01
-2.45204657e-01 -7.13808239e-01 4.19018775e-01 9.48966760e-03
1.13579237e+00 -2.44912609e-01 -3.66584480e-01 6.20238900e-01
1.87067413e+00 4.39559042e-01 5.00811398e-01 4.52601343e-01
5.53540289e-01 7.46804059e-01 4.98430789e-01 2.08012268e-01
9.62645113e-01 4.51824069e-01 4.25979316e-01 1.68018043e-01
-7.29968697e-02 -8.55770648e-01 5.85407279e-02 2.62739837e-01
6.24918520e-01 -2.89395064e-01 -1.26426935e+00 3.02343190e-01
-1.76596260e+00 -5.06230175e-01 2.50824451e-01 2.07063198e+00
9.67923880e-01 3.26671064e-01 -1.22172900e-01 -4.02569383e-01
4.26264971e-01 -1.51254594e-01 -7.99015343e-01 -7.61305273e-01
-9.03409496e-02 6.97366476e-01 7.76854813e-01 5.51093996e-01
-7.22677290e-01 1.56202018e+00 5.75830793e+00 7.11130440e-01
-9.35220182e-01 -5.82162850e-02 2.89431840e-01 2.51172423e-01
-6.38773382e-01 1.78804219e-01 -1.40043378e+00 2.05355540e-01
1.30392301e+00 -7.63140246e-02 4.72870797e-01 8.33972037e-01
-2.47407198e-01 -1.18488327e-01 -1.28421378e+00 6.81451559e-01
-3.36156756e-01 -1.64511299e+00 4.76076990e-01 -2.69148737e-01
-4.36024182e-02 1.60347983e-01 -3.98670048e-01 9.17935193e-01
5.93977034e-01 -1.07820868e+00 5.56519330e-01 5.32370865e-01
6.95105314e-01 -6.83313787e-01 7.34884262e-01 1.26288146e-01
-1.38215399e+00 -1.44497514e-01 -1.44966811e-01 4.56338435e-01
1.01207905e-02 8.17227140e-02 -1.07131624e+00 7.44569957e-01
8.44769657e-01 2.14312077e-01 -8.13005090e-01 1.02405973e-01
-3.57081816e-02 4.72575307e-01 -5.37861407e-01 7.06264898e-02
2.21741319e-01 3.69210571e-01 4.15579863e-02 1.20288622e+00
6.40167966e-02 3.18967730e-01 2.88372785e-01 1.17644656e+00
-4.31322545e-01 2.21694499e-01 -4.02460545e-01 -3.48588794e-01
8.94003808e-01 7.15499222e-01 -6.29440188e-01 -7.36010134e-01
-6.93941474e-01 8.07255030e-01 4.12151635e-01 4.12904084e-01
-7.65127122e-01 -5.03195226e-01 4.73745048e-01 8.59520957e-02
2.77540475e-01 2.47299284e-01 -4.90865320e-01 -1.07668579e+00
4.45321530e-01 -1.10878277e+00 1.11835849e+00 -8.05517852e-01
-9.51673329e-01 5.37236571e-01 1.30941734e-01 -2.97257930e-01
-5.08081794e-01 -4.75630283e-01 -2.13714093e-01 1.01152837e+00
-1.52691185e+00 -1.24992406e+00 -2.36126736e-01 4.13720101e-01
5.34983456e-01 -1.90238729e-01 1.15266395e+00 3.20688546e-01
-4.00116205e-01 7.83174396e-01 -5.87181568e-01 3.51893604e-01
4.83528674e-01 -1.38276851e+00 8.76394749e-01 4.14029270e-01
-8.95326734e-02 1.03779078e+00 2.71043628e-01 -7.81765401e-01
-1.68872929e+00 -1.19064701e+00 1.14102161e+00 -6.57289624e-01
5.60406625e-01 -4.57436681e-01 -1.29358256e+00 7.94553280e-01
-5.16153984e-02 3.12695689e-02 8.25312257e-01 8.01654309e-02
-8.85442197e-01 -1.36091068e-01 -1.16173446e+00 2.02964440e-01
1.04603338e+00 -9.25200939e-01 -8.14343214e-01 3.92722003e-02
1.43422163e+00 -4.61875945e-01 -1.28756011e+00 6.47180855e-01
6.45630777e-01 -1.00246966e+00 1.18755043e+00 -1.14247620e+00
2.39965841e-01 -2.79428691e-01 -8.43009412e-01 -5.01505792e-01
3.85340840e-01 -1.81064621e-01 -2.34985158e-01 1.32067287e+00
7.61892676e-01 -7.53527761e-01 1.17992282e+00 1.09218931e+00
-2.34536216e-01 -8.36639881e-01 -3.95233780e-01 -7.08586991e-01
5.47438897e-02 -8.96943212e-01 1.05277288e+00 6.76295400e-01
-2.62358725e-01 5.68872035e-01 6.23697639e-01 4.30770576e-01
3.13623011e-01 3.98002207e-01 7.59953976e-01 -1.12887990e+00
-4.35695887e-01 -3.48242790e-01 -1.19858727e-01 -1.02202308e+00
3.50236326e-01 -1.04659700e+00 -2.00886950e-01 -1.50962126e+00
-1.89645849e-02 -6.71596825e-01 -1.78795010e-01 8.12710583e-01
3.63621698e-03 -3.91568214e-01 2.79014604e-03 -4.19913866e-02
-6.05905890e-01 -2.15810426e-02 3.73029292e-01 -1.84300151e-02
-3.58445168e-01 -4.07628775e-01 -1.01429629e+00 4.41390306e-01
6.66457415e-01 -5.61187506e-01 -6.64812744e-01 -6.88258648e-01
6.98461831e-01 4.71950442e-01 2.79166371e-01 -8.16402256e-01
3.70855451e-01 -5.65771898e-03 -1.05978645e-01 -5.55462718e-01
3.08520421e-02 -6.14111006e-01 7.35690147e-02 3.36625874e-01
-7.07275391e-01 2.67610222e-01 4.25875783e-01 3.38977903e-01
-5.83508253e-01 -3.05001646e-01 4.96520191e-01 -5.47258496e-01
-1.11878932e+00 9.66483075e-03 2.68296242e-01 7.42400527e-01
5.59960485e-01 -1.59774348e-01 -5.03168166e-01 -1.19730700e-02
-5.82951427e-01 3.78059924e-01 4.09463078e-01 6.53307736e-01
3.68310124e-01 -8.76104593e-01 -2.44920462e-01 3.47255707e-01
5.10335326e-01 1.90213546e-01 -1.54373320e-02 1.64021373e-01
-7.76608646e-01 9.60372746e-01 1.63461879e-01 -5.36225140e-01
-9.08004165e-01 6.68863714e-01 3.35215062e-01 -5.02409458e-01
-1.37744531e-01 8.54710639e-01 2.87531435e-01 -1.07890713e+00
2.59611905e-01 -7.48117387e-01 6.42070770e-02 -2.08290488e-01
2.36141279e-01 6.84039891e-02 3.08561713e-01 -2.55466104e-01
-5.89561343e-01 5.12289703e-01 -9.08168182e-02 -6.15005242e-03
1.03368640e+00 1.17887892e-01 -2.21654564e-01 1.75461099e-01
1.37941658e+00 -9.70188603e-02 -4.65187013e-01 -4.79325414e-01
6.20746911e-01 3.22719142e-02 -2.62585640e-01 -1.24928522e+00
-8.10028315e-01 7.76392817e-01 4.02314633e-01 2.36793146e-01
1.17958653e+00 2.97486633e-01 1.20977056e+00 9.17142987e-01
5.02636850e-01 -1.07493937e+00 -1.32212132e-01 6.23694181e-01
3.13033998e-01 -1.31640136e+00 -4.82727289e-01 -7.11307526e-01
-4.37425852e-01 1.10907233e+00 9.18775141e-01 2.31542259e-01
3.53273928e-01 3.68187487e-01 1.30276307e-01 -5.71722686e-01
-8.62984896e-01 -5.81204705e-02 1.45823106e-01 3.42090338e-01
5.94083130e-01 -6.53975382e-02 1.18854031e-01 9.74095464e-01
-4.35142338e-01 2.17201635e-01 2.35498529e-02 1.36273372e+00
-5.52052379e-01 -1.33486843e+00 -1.81752164e-02 4.64098841e-01
-7.40835845e-01 -4.72017437e-01 -3.18229258e-01 8.26823533e-01
-3.95072177e-02 8.65218818e-01 1.12559251e-01 -1.93142518e-01
7.59320796e-01 5.91400385e-01 2.39045564e-02 -9.95542824e-01
-1.02416945e+00 -2.77120024e-01 5.16816974e-01 -1.20538008e+00
-3.76275578e-03 -4.41179015e-02 -2.07762337e+00 6.90961704e-02
1.47487134e-01 3.84324551e-01 6.86612427e-01 6.40973151e-01
9.08615291e-01 3.10210615e-01 -3.24196275e-03 6.06174707e-01
-4.25093919e-01 -2.73634046e-01 -1.27716869e-01 3.67919624e-01
-1.79934710e-01 -3.24952528e-02 2.74858713e-01 1.60265878e-01] | [9.970898628234863, 7.800204277038574] |
803a9952-2a9a-4055-9fee-8ab15bb7b6b7 | buffered-streaming-graph-partitioning | 2102.09384 | null | https://arxiv.org/abs/2102.09384v2 | https://arxiv.org/pdf/2102.09384v2.pdf | Buffered Streaming Graph Partitioning | Partitioning graphs into blocks of roughly equal size is widely used when processing large graphs. Currently there is a gap in the space of available partitioning algorithms. On the one hand, there are streaming algorithms that have been adopted to partition massive graph data on small machines. In the streaming model, vertices arrive one at a time including their neighborhood and then have to be assigned directly to a block. These algorithms can partition huge graphs quickly with little memory, but they produce partitions with low solution quality. On the other hand, there are offline (shared-memory) multilevel algorithms that produce partitions with high quality but also need a machine with enough memory. We make a first step to close this gap by presenting an algorithm that computes significantly improved partitions of huge graphs using a single machine with little memory in streaming setting. First, we adopt the buffered streaming model which is a more reasonable approach in practice. In this model, a processing element can store a buffer, or batch, of nodes before making assignment decisions. When our algorithm receives a batch of nodes, we build a model graph that represents the nodes of the batch and the already present partition structure. This model enables us to apply multilevel algorithms and in turn compute much higher quality solutions of huge graphs on cheap machines than previously possible. To partition the model, we develop a multilevel algorithm that optimizes an objective function that has previously shown to be effective for the streaming setting. This also removes the dependency on the number of blocks k from the running time compared to the previous state-of-the-art. Overall, our algorithm computes, on average, 75.9% better solutions than Fennel using a very small buffer size. In addition, for large values of k our algorithm becomes faster than Fennel. | ['Christian Schulz', 'Marcelo Fonseca Faraj'] | 2021-02-18 | null | null | null | null | ['graph-partitioning'] | ['graphs'] | [ 3.80757377e-02 2.67136067e-01 4.91057523e-02 2.68793292e-02
-6.55989766e-01 -5.54625690e-01 7.11954907e-02 8.88177097e-01
-3.11603725e-01 5.02068043e-01 -1.60573050e-01 -2.95013219e-01
-2.60874867e-01 -1.45793402e+00 -8.54310930e-01 -6.53913498e-01
-3.19791824e-01 1.13024914e+00 8.24325979e-01 -1.07542418e-01
1.97709829e-01 4.97516006e-01 -1.75337601e+00 5.83950937e-01
7.61588812e-01 8.79578173e-01 3.38796586e-01 1.06843650e+00
-3.30093682e-01 5.39359510e-01 -4.61346984e-01 -2.99310833e-01
6.35179818e-01 -2.53669620e-01 -1.18909109e+00 3.68832052e-01
8.45050290e-02 1.82365794e-02 -1.81206688e-01 7.22577631e-01
3.45834255e-01 8.95666033e-02 2.70094901e-01 -1.26734531e+00
2.32966006e-01 1.14536846e+00 -9.44989681e-01 1.28284320e-01
3.61550331e-01 -2.08466545e-01 8.62631917e-01 -4.69767570e-01
7.84567654e-01 9.44465697e-01 5.83399534e-01 -7.85312951e-02
-1.59516883e+00 -2.17815250e-01 1.56763166e-01 2.37292290e-01
-1.65299118e+00 -2.31074736e-01 3.06062460e-01 -2.51910299e-01
9.78771329e-01 8.12811494e-01 8.63720894e-01 -1.73615217e-01
-1.65125858e-02 4.58080202e-01 8.19381118e-01 -5.12722611e-01
4.86646295e-01 8.20344612e-02 2.62802899e-01 3.65745515e-01
4.31985766e-01 -5.55532634e-01 -1.90287501e-01 -2.56117254e-01
3.34269077e-01 -6.84966147e-02 -2.67455935e-01 -6.92184985e-01
-1.10674453e+00 7.25189865e-01 2.87087232e-01 4.54731196e-01
-3.94241631e-01 1.26308694e-01 5.55913925e-01 6.31988049e-01
5.24809241e-01 5.90285175e-02 -3.37890655e-01 5.22985831e-02
-1.37867951e+00 1.64081305e-01 1.35061097e+00 8.18332255e-01
1.00827587e+00 -5.90620577e-01 8.84432867e-02 5.64428806e-01
-5.93414977e-02 1.89111829e-01 -9.88789499e-02 -6.12346470e-01
7.18968689e-01 5.99566340e-01 -9.94104668e-02 -1.22153044e+00
-6.15590274e-01 -1.52972311e-01 -1.02070773e+00 9.46922749e-02
5.81355035e-01 2.46353567e-01 -7.05562770e-01 1.32815504e+00
8.90148759e-01 1.00442365e-01 -1.48516938e-01 6.84531569e-01
2.12636337e-01 1.02760577e+00 -4.21907693e-01 -6.69662654e-01
1.31362081e+00 -1.07316005e+00 -2.66197681e-01 2.29773968e-01
9.28336680e-01 -7.78620243e-01 6.83971941e-01 7.31642783e-01
-1.51710105e+00 -3.11106473e-01 -1.01157165e+00 1.17992245e-01
-3.31116110e-01 -5.23316503e-01 6.50054991e-01 9.20518160e-01
-1.67342794e+00 8.93126547e-01 -9.87961709e-01 -3.52467328e-01
3.09866462e-02 6.15829289e-01 -3.40704411e-01 -3.77696574e-01
-7.29256988e-01 4.82215285e-01 5.14847219e-01 9.06148273e-03
-4.92399037e-01 -8.00955772e-01 -6.05025887e-01 5.39862692e-01
7.83664823e-01 -6.71998084e-01 8.43490899e-01 -9.11660254e-01
-1.18928957e+00 6.88226998e-01 -2.46727586e-01 -5.66082776e-01
4.85916525e-01 6.26614094e-01 2.09899202e-01 4.54740375e-01
-2.59215295e-01 3.24169725e-01 6.14698112e-01 -9.93628025e-01
-6.87226474e-01 -4.95509148e-01 4.50935781e-01 1.07323043e-01
-4.71445948e-01 -3.10163442e-02 -6.60221875e-01 -1.21166315e-02
2.87828952e-01 -9.50959206e-01 -7.71010876e-01 -4.89232630e-01
-1.98304430e-01 -3.22358191e-01 1.16755225e-01 -2.30926663e-01
1.56224585e+00 -2.02826953e+00 5.50347686e-01 6.76519513e-01
5.81317544e-01 4.84568216e-02 5.36845475e-02 1.08149087e+00
4.31008823e-02 2.00857908e-01 -1.59595981e-01 -3.99708927e-01
7.98035786e-02 1.42740026e-01 -1.09706976e-01 5.83047926e-01
-5.65219104e-01 4.53248560e-01 -5.77437341e-01 -4.83348072e-01
9.32566822e-02 2.60337945e-02 -9.21071589e-01 -1.66272968e-01
-1.66918233e-01 -2.98807412e-01 9.90839973e-02 2.76821163e-02
1.19679046e+00 -5.32334626e-01 6.18048072e-01 1.19053639e-01
-2.15548143e-01 1.92156911e-01 -1.92793798e+00 1.43320930e+00
-3.57216597e-01 9.54622999e-02 5.56240082e-01 -1.28536355e+00
4.81689066e-01 3.65480274e-01 6.44064486e-01 -3.26065183e-01
-1.81210190e-01 3.37656200e-01 -7.28261918e-02 -1.02099665e-01
6.72547936e-01 -1.48947299e-01 5.33632971e-02 6.65161848e-01
-4.00080979e-01 -3.11510414e-02 8.73687804e-01 6.35117829e-01
1.73498285e+00 -5.73183000e-01 3.71300355e-02 -4.44864094e-01
5.48924863e-01 2.56653100e-01 4.36868578e-01 6.37352288e-01
3.25154454e-01 5.65991938e-01 9.69581485e-01 -6.26175523e-01
-1.10938406e+00 -7.47140110e-01 2.04008222e-01 1.08005309e+00
2.02465162e-01 -1.10027421e+00 -9.10825372e-01 -1.60527736e-01
-1.01818688e-01 2.09743180e-03 -4.32057559e-01 3.01359802e-01
-5.50701797e-01 -5.96410513e-01 -6.76962780e-03 2.57407814e-01
-1.70776680e-01 -7.07263052e-01 -6.01108909e-01 3.97526681e-01
3.53508554e-02 -9.58614826e-01 -2.36557528e-01 1.62173599e-01
-9.69785869e-01 -1.08742714e+00 -4.00230914e-01 -6.81520104e-01
8.78640532e-01 6.55507803e-01 1.20102441e+00 5.25537252e-01
-2.31452182e-01 1.20335028e-01 -5.07247448e-01 3.36634950e-03
-2.60845006e-01 5.45985818e-01 -2.14512259e-01 -5.49129881e-02
-4.19483669e-02 -7.72262454e-01 -4.47223395e-01 2.37645060e-01
-1.30400717e+00 3.51743430e-01 2.55540460e-01 3.35989565e-01
7.36841917e-01 6.76736593e-01 2.50316054e-01 -1.21926367e+00
3.23629826e-01 -6.51199758e-01 -9.27286565e-01 -1.35411816e-02
-4.43796635e-01 -9.59113911e-02 1.13478875e+00 -1.90285861e-01
-4.03062433e-01 1.55874655e-01 -7.81992748e-02 -2.58118123e-01
-5.32957772e-03 7.80033648e-01 -6.57424554e-02 6.66943938e-02
3.56744409e-01 -1.32347032e-01 1.28836129e-02 -3.35499644e-01
3.09861749e-01 4.78869230e-01 -2.77344696e-02 -5.48803031e-01
6.59766138e-01 6.53750837e-01 3.67403001e-01 -7.32072175e-01
-1.43337637e-01 -8.46856296e-01 -6.35546029e-01 -1.57285541e-01
3.62528056e-01 -5.33246338e-01 -8.52388680e-01 3.60320240e-01
-8.16173673e-01 -6.88238263e-01 -3.05785775e-01 6.98841140e-02
-5.10948300e-01 5.75401127e-01 -7.76084721e-01 -6.55205250e-01
-6.94784597e-02 -9.79141891e-01 8.25437963e-01 -1.95731044e-01
-3.64003540e-03 -6.78386927e-01 1.22605637e-01 3.03962082e-01
3.77409220e-01 2.50106514e-01 8.50544930e-01 -4.93733406e-01
-8.14235210e-01 -2.02710003e-01 -1.09912939e-01 -2.18814120e-01
-3.32768321e-01 1.09432079e-01 -3.66097927e-01 -7.27572381e-01
-6.34237006e-02 4.66306619e-02 7.49501109e-01 2.60161787e-01
1.13696015e+00 -2.31595904e-01 -5.66764474e-01 3.56756628e-01
1.73806965e+00 -1.70628622e-01 8.47658932e-01 1.06689453e-01
6.08601153e-01 6.29074037e-01 4.68448848e-01 6.10914767e-01
5.49537718e-01 5.68710744e-01 2.65776068e-01 -8.12774301e-02
1.57527670e-01 1.46809429e-01 2.20174789e-01 1.21373165e+00
-3.58761586e-02 -4.46931332e-01 -9.87506151e-01 7.99971879e-01
-1.98945594e+00 -8.15833867e-01 -6.37051225e-01 2.59817266e+00
5.79345942e-01 3.23494583e-01 4.96579319e-01 7.81593919e-01
7.35028028e-01 9.86406766e-03 1.49981966e-02 -8.61579537e-01
3.40483606e-01 4.32801157e-01 7.80750513e-01 7.03263819e-01
-7.39970803e-01 6.54926419e-01 5.39917898e+00 1.06826413e+00
-9.51407909e-01 1.55426040e-01 6.12289667e-01 -4.95777994e-01
-4.30103481e-01 2.37514228e-01 -5.49630582e-01 4.98755872e-01
1.52106810e+00 -4.35505420e-01 6.50366902e-01 7.30533957e-01
1.76909074e-01 -4.99057800e-01 -1.25514209e+00 8.06737006e-01
-5.95146157e-02 -1.38582873e+00 -2.58669883e-01 6.34034514e-01
8.17768633e-01 -1.43627599e-01 -4.53276068e-01 9.87296253e-02
1.36189818e-01 -9.96960461e-01 5.35240054e-01 1.38124183e-01
4.00797576e-01 -1.13743079e+00 7.71078467e-01 7.39462078e-01
-1.65808141e+00 -3.37980203e-02 -8.41063321e-01 -4.64707255e-01
2.59581745e-01 1.09899330e+00 -6.02981150e-01 8.55908751e-01
6.48350894e-01 1.87850036e-02 -2.65718490e-01 1.21021521e+00
3.78468126e-01 5.75782418e-01 -9.72779453e-01 7.35852718e-02
2.59980589e-01 -4.82636273e-01 2.64823467e-01 9.74565029e-01
3.23863983e-01 3.00206304e-01 6.66339636e-01 2.29399234e-01
-8.68744329e-02 5.44567585e-01 -5.40031433e-01 2.40885347e-01
4.48757231e-01 1.44375396e+00 -1.56559587e+00 -6.06674910e-01
-4.51789916e-01 7.15312839e-01 5.22384107e-01 -5.07370904e-02
-7.97790468e-01 -4.62358087e-01 2.18867064e-01 6.39697134e-01
7.97764838e-01 -3.34152669e-01 -2.90200382e-01 -8.16426814e-01
3.02081347e-01 -9.08528090e-01 6.90368593e-01 -2.51499951e-01
-7.78665125e-01 5.26732504e-01 -2.38201814e-03 -7.50727713e-01
-9.83962119e-02 -2.13859007e-01 -3.01140785e-01 5.45350790e-01
-1.01362646e+00 -6.74916804e-01 -2.20745236e-01 5.73133230e-01
3.72714937e-01 7.09195137e-01 5.26148975e-01 4.84516919e-01
-2.69410133e-01 3.11355740e-01 1.48555577e-01 -3.98425788e-01
3.59116733e-01 -1.30693388e+00 2.90682763e-01 9.54166949e-01
1.39537171e-01 3.58161598e-01 7.30487704e-01 -4.80967969e-01
-1.80709040e+00 -6.34851694e-01 9.65598583e-01 -9.04098153e-02
5.65781295e-01 -5.95454514e-01 -1.00431108e+00 3.88984829e-01
2.23677412e-01 9.08937827e-02 7.03363657e-01 3.03051829e-01
9.06024501e-02 -4.47376490e-01 -1.11739182e+00 2.84368515e-01
9.29591477e-01 -4.86545004e-02 1.88494157e-02 4.74401861e-01
6.27397001e-01 -4.53480095e-01 -1.10567164e+00 8.89387578e-02
2.21309587e-01 -1.11784756e+00 6.75262630e-01 -2.50597686e-01
2.21492425e-01 -5.06743550e-01 7.10703507e-02 -1.15073216e+00
-3.52513671e-01 -9.10486102e-01 -5.56425788e-02 1.12989962e+00
4.84810293e-01 -7.09359050e-01 8.87009144e-01 7.20705688e-01
4.63392250e-02 -1.02985632e+00 -9.64354575e-01 -8.27138722e-01
1.26702473e-01 -3.87312055e-01 7.44675756e-01 6.71552062e-01
4.84701723e-01 2.55421907e-01 -1.31752593e-02 2.86080390e-01
5.18252730e-01 7.49439657e-01 1.08953381e+00 -1.37875462e+00
-4.27703828e-01 -3.81580949e-01 -4.58504051e-01 -1.01246727e+00
-1.92936063e-01 -1.08824754e+00 -2.16132686e-01 -1.87221444e+00
4.23277318e-01 -7.42758751e-01 8.87432229e-03 2.69463539e-01
4.10219096e-03 2.67259449e-01 3.67173284e-01 9.07171220e-02
-9.08492386e-01 -1.72686785e-01 8.91359448e-01 1.29973367e-01
-3.17088425e-01 -5.56292497e-02 -5.06925464e-01 4.66067523e-01
6.88946307e-01 -5.10079086e-01 -4.07489210e-01 -3.37677866e-01
7.58431852e-01 3.29199821e-01 -2.97138065e-01 -1.01309347e+00
6.23578191e-01 -3.33385170e-02 -6.63422644e-02 -7.28805184e-01
1.84782688e-02 -9.60127413e-01 7.66326308e-01 6.26370907e-01
3.23190764e-02 3.22298467e-01 1.16069980e-01 5.72697699e-01
-1.49193347e-01 -3.58870864e-01 6.69981003e-01 -3.23076695e-01
-1.20392807e-01 4.19223636e-01 -5.97221613e-01 -6.37723207e-02
1.36427701e+00 -3.35460633e-01 -1.13900781e-01 -3.19747001e-01
-7.60518789e-01 5.57868659e-01 9.26703393e-01 -1.04329601e-01
1.50304750e-01 -1.06680286e+00 -7.19361067e-01 9.39131677e-02
-3.89320254e-01 3.64030898e-01 4.69705313e-01 1.24204481e+00
-9.97891903e-01 3.06546599e-01 1.01500288e-01 -5.79750717e-01
-1.54943335e+00 1.29970610e+00 -2.94324785e-01 -7.78383970e-01
-6.87576115e-01 6.09640241e-01 1.21196471e-01 -8.45836774e-02
-1.01253152e-01 -4.99490023e-01 1.89258426e-01 3.22544813e-01
6.79488420e-01 7.63131320e-01 4.52189833e-01 -3.87528479e-01
-3.00944895e-01 3.12378377e-01 -4.73209806e-02 -4.77477349e-02
1.50591254e+00 -2.33431399e-01 -7.70641088e-01 3.06746781e-01
9.09457028e-01 3.89507383e-01 -6.86318696e-01 1.00432888e-01
-2.54215933e-02 -6.09449625e-01 -2.45648026e-01 1.62697062e-02
-1.18484890e+00 6.19983077e-01 -4.35823761e-02 1.26220953e+00
1.35509098e+00 2.07612403e-02 9.93589580e-01 1.01277240e-01
7.56005645e-01 -1.14563429e+00 -4.96428967e-01 2.62033522e-01
3.15696925e-01 -5.67667127e-01 1.81678951e-01 -9.82540071e-01
-4.60067242e-02 1.07027054e+00 2.88117349e-01 -1.92280248e-01
5.11985958e-01 7.21980572e-01 -4.37434316e-01 -1.00038022e-01
-1.13104105e+00 -3.06151807e-01 -2.96761751e-01 2.35352278e-01
3.53635810e-02 2.79890805e-01 -8.02627146e-01 3.32087755e-01
-3.08451116e-01 -1.38845533e-01 8.26195121e-01 7.95800328e-01
-7.35365927e-01 -1.44135129e+00 -5.53362548e-01 5.29608428e-01
-5.10040998e-01 -2.79726312e-02 -1.61601648e-01 5.70611715e-01
-1.75118953e-01 1.15994036e+00 2.27240711e-01 -9.95696783e-02
1.85305923e-01 -9.75709632e-02 6.20012403e-01 -7.62443364e-01
-8.14758658e-01 1.25036642e-01 5.93067594e-02 -7.64631152e-01
-1.02627553e-01 -3.90137732e-01 -1.26035845e+00 -9.73231792e-01
-3.86965156e-01 5.27361393e-01 5.00089288e-01 6.88305855e-01
5.36831558e-01 4.19700772e-01 7.27788389e-01 -1.05565286e+00
-2.39552543e-01 -6.81162596e-01 -7.53559113e-01 1.76985547e-01
-1.98480532e-01 -1.38987839e-01 -3.12805921e-01 -1.92781508e-01] | [6.9958930015563965, 5.195828914642334] |
03045dd4-988c-488b-b00b-328349c711fe | dreamix-video-diffusion-models-are-general | 2302.01329 | null | https://arxiv.org/abs/2302.01329v1 | https://arxiv.org/pdf/2302.01329v1.pdf | Dreamix: Video Diffusion Models are General Video Editors | Text-driven image and video diffusion models have recently achieved unprecedented generation realism. While diffusion models have been successfully applied for image editing, very few works have done so for video editing. We present the first diffusion-based method that is able to perform text-based motion and appearance editing of general videos. Our approach uses a video diffusion model to combine, at inference time, the low-resolution spatio-temporal information from the original video with new, high resolution information that it synthesized to align with the guiding text prompt. As obtaining high-fidelity to the original video requires retaining some of its high-resolution information, we add a preliminary stage of finetuning the model on the original video, significantly boosting fidelity. We propose to improve motion editability by a new, mixed objective that jointly finetunes with full temporal attention and with temporal attention masking. We further introduce a new framework for image animation. We first transform the image into a coarse video by simple image processing operations such as replication and perspective geometric projections, and then use our general video editor to animate it. As a further application, we can use our method for subject-driven video generation. Extensive qualitative and numerical experiments showcase the remarkable editing ability of our method and establish its superior performance compared to baseline methods. | ['Yedid Hoshen', 'Yaniv Leviathan', 'Yael Pritch', 'Yossi Matias', 'Alex Rav Acha', 'Dani Valevski', 'Eliahu Horwitz', 'Eyal Molad'] | 2023-02-02 | null | null | null | null | ['text-to-video-editing', 'video-generation', 'subject-driven-video-generation', 'image-to-video', 'image-animation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 5.55657208e-01 -2.91769486e-02 1.66141838e-01 1.39007634e-02
-3.12614560e-01 -5.23403585e-01 8.99397790e-01 -4.05324370e-01
-3.04829925e-01 6.36305630e-01 2.99452007e-01 -1.21191174e-01
1.24264643e-01 -5.68689644e-01 -7.88837731e-01 -5.77573121e-01
6.37023002e-02 1.62368730e-01 4.57190454e-01 -3.53428602e-01
3.66488814e-01 5.47226608e-01 -1.47795498e+00 4.30723548e-01
9.07516658e-01 5.48645973e-01 5.56553960e-01 1.17284918e+00
-2.30172928e-02 1.22513545e+00 -4.42293763e-01 -1.55963898e-01
2.74040669e-01 -7.47253180e-01 -7.34973013e-01 5.42960525e-01
5.93873084e-01 -6.11977756e-01 -5.94127238e-01 9.03626680e-01
3.76531333e-01 4.16039735e-01 4.99924362e-01 -1.03593349e+00
-1.09291422e+00 4.18446153e-01 -8.12995613e-01 2.92826205e-01
7.44675815e-01 3.15935224e-01 5.17191350e-01 -9.68168139e-01
1.27760422e+00 1.24980152e+00 3.51103425e-01 8.20857644e-01
-1.34224284e+00 -2.79525965e-01 3.90841097e-01 1.16167560e-01
-1.15050411e+00 -5.16367674e-01 9.34206843e-01 -5.57895184e-01
7.62478173e-01 3.58514875e-01 9.34017181e-01 1.13641810e+00
2.41856754e-01 5.56277335e-01 1.03440845e+00 -4.82862025e-01
3.96355055e-02 -4.91900742e-02 -5.22819459e-01 7.48700917e-01
-3.73384178e-01 2.48383179e-01 -4.27732408e-01 1.72814041e-01
1.21627843e+00 -2.99802758e-02 -6.19939625e-01 -4.12761807e-01
-1.52517676e+00 5.04421294e-01 1.59487218e-01 4.23799425e-01
-4.81469780e-01 4.06678051e-01 6.34273514e-02 4.12731528e-01
6.76344812e-01 2.80556172e-01 -1.80941331e-03 -1.56822652e-01
-1.33667946e+00 3.56223792e-01 5.22569001e-01 1.02882516e+00
4.81579959e-01 2.91590482e-01 -5.15161037e-01 5.93962729e-01
2.69557629e-02 4.68671262e-01 2.59620428e-01 -1.44398510e+00
1.16120644e-01 8.98166746e-02 2.34887809e-01 -1.17793572e+00
1.11974534e-02 1.32429615e-01 -9.66419697e-01 6.09082580e-01
3.01049888e-01 -7.36034811e-02 -8.79173577e-01 1.72960556e+00
3.39956135e-01 2.58995056e-01 -2.83598155e-01 9.77542520e-01
3.29140335e-01 9.94745851e-01 -5.25031239e-02 -5.38879275e-01
1.02355266e+00 -1.13889372e+00 -9.67377901e-01 2.07901642e-01
2.70019472e-01 -9.77755070e-01 1.07177365e+00 4.29806799e-01
-1.69312549e+00 -6.20429218e-01 -8.91872764e-01 -3.26159596e-01
2.96156537e-02 -8.94516334e-02 2.80988634e-01 2.68440217e-01
-1.42113018e+00 7.53835738e-01 -7.78053164e-01 -2.99536407e-01
1.50339678e-01 1.62002444e-01 -3.30912143e-01 2.41106115e-02
-9.90732372e-01 7.57943094e-01 3.29793505e-02 -1.12615200e-02
-9.30035651e-01 -7.77668059e-01 -6.90085590e-01 -1.35400996e-01
5.09068310e-01 -1.14021373e+00 1.20012069e+00 -1.39422941e+00
-1.93949509e+00 7.31186688e-01 -2.32681796e-01 -2.19773680e-01
9.45823133e-01 -1.65364996e-01 -2.65630811e-01 4.57689524e-01
-1.12444550e-01 1.05356538e+00 1.49205709e+00 -1.38341272e+00
-5.42765200e-01 1.09198473e-01 3.20126653e-01 3.23837012e-01
-3.93053085e-01 4.46277484e-02 -8.04226041e-01 -1.27507710e+00
-2.63254106e-01 -9.22742605e-01 -3.17239434e-01 4.82012153e-01
-2.60526359e-01 3.04805130e-01 1.09612179e+00 -7.39359617e-01
1.40640986e+00 -2.09377861e+00 7.15494573e-01 -1.96008682e-02
4.47994083e-01 2.32279301e-01 -3.39482784e-01 5.62049091e-01
-1.31456271e-01 3.11608631e-02 -4.77250069e-01 -4.60632682e-01
-3.73932719e-01 1.88339800e-02 -3.85939479e-01 2.65617251e-01
2.21565276e-01 9.63850319e-01 -1.05093050e+00 -4.37683225e-01
5.57781577e-01 9.01202142e-01 -9.22473013e-01 1.51207209e-01
-3.82860661e-01 9.14043605e-01 -1.67190343e-01 2.63241947e-01
5.24272740e-01 -1.24099344e-01 -9.74262413e-03 -3.20408523e-01
-3.67051780e-01 -3.43046784e-01 -1.09460354e+00 1.84402370e+00
-4.84447449e-01 8.81303251e-01 3.33328247e-01 -5.87960184e-01
4.99358654e-01 1.61572531e-01 6.61676884e-01 -5.92360258e-01
-6.74238354e-02 -5.37992790e-02 -1.79946348e-01 -6.97039068e-01
9.16920602e-01 -7.60394186e-02 3.99752259e-01 5.58395863e-01
-1.84450373e-01 -4.89545852e-01 2.92922288e-01 6.08458638e-01
8.08628321e-01 5.72240233e-01 -2.03458369e-02 -1.68786943e-01
7.01065361e-01 2.30604187e-02 2.46718470e-02 5.66330969e-01
1.01174876e-01 9.12984192e-01 3.16601396e-01 -2.67285526e-01
-1.47887850e+00 -7.05266118e-01 3.04203272e-01 9.85875905e-01
1.34258166e-01 -4.52032179e-01 -1.07912338e+00 -3.94039631e-01
-3.39212656e-01 6.18816793e-01 -8.35605502e-01 -9.22937617e-02
-8.94104123e-01 -4.00635958e-01 1.81812659e-01 3.48862648e-01
3.24675918e-01 -1.09844840e+00 -4.29107577e-01 3.52181464e-01
-1.42647892e-01 -1.13128710e+00 -1.10711563e+00 -3.69682521e-01
-7.96952844e-01 -7.60418832e-01 -1.33494568e+00 -9.12488639e-01
8.23557317e-01 4.31767911e-01 8.48249018e-01 2.52947688e-01
-2.04993978e-01 6.21093035e-01 -3.06795418e-01 2.48832718e-01
-7.45409489e-01 -3.32238704e-01 3.88982221e-02 3.38639081e-01
-5.41270614e-01 -6.55228317e-01 -6.32313669e-01 2.28444561e-01
-1.35392499e+00 6.27007842e-01 3.01630408e-01 8.11908543e-01
3.92286688e-01 -1.18905596e-01 8.60259123e-03 -8.35634470e-01
6.02290273e-01 -5.24139404e-02 -4.58016276e-01 -2.14240551e-02
-2.47186661e-01 5.23792654e-02 7.77344942e-01 -7.79102743e-01
-1.25818181e+00 1.33369192e-01 -1.62987292e-01 -8.73035550e-01
1.78521127e-01 1.94500819e-01 1.23644948e-01 -3.59748214e-01
4.61923093e-01 2.60830820e-01 1.13323465e-01 -2.67527550e-01
7.86553979e-01 2.18802750e-01 6.83351815e-01 -4.36346442e-01
1.02422976e+00 7.53486395e-01 -1.30041197e-01 -1.03764081e+00
-3.34981292e-01 1.43465176e-01 -7.98169851e-01 -4.22230422e-01
9.76018548e-01 -6.34637713e-01 -6.36861622e-01 6.91575408e-01
-1.32652748e+00 -6.98759794e-01 -5.48688471e-01 2.38639370e-01
-8.39260936e-01 6.95976257e-01 -7.61559129e-01 -4.90866870e-01
5.62826768e-02 -1.23979306e+00 1.03381670e+00 -1.41456217e-01
-1.55915439e-01 -1.15595126e+00 1.65483057e-01 -6.73112720e-02
6.43890738e-01 2.83098370e-01 6.39276743e-01 2.38848493e-01
-8.53266954e-01 2.73886174e-01 -1.71572819e-01 2.05077380e-01
9.77275297e-02 5.68235397e-01 -6.41979396e-01 -2.94482768e-01
7.98522979e-02 1.72762200e-01 8.84506047e-01 5.48385084e-01
1.11948454e+00 -3.09474468e-01 -2.19339311e-01 8.71343732e-01
1.21605611e+00 2.43150711e-01 7.33806670e-01 2.63637871e-01
9.66744006e-01 5.58582127e-01 5.59540510e-01 3.38663042e-01
2.31397137e-01 1.00265706e+00 1.90351620e-01 -2.79823929e-01
-5.99016845e-01 -2.80922711e-01 5.67900181e-01 9.55548048e-01
-4.63910878e-01 -4.18935269e-01 -5.17387271e-01 3.36770326e-01
-1.83827841e+00 -1.26975262e+00 -1.54013634e-01 2.08383512e+00
6.58848345e-01 -5.25291637e-02 9.20726359e-02 2.59741638e-02
8.31829906e-01 3.66361856e-01 -3.98176253e-01 -3.55363190e-01
-1.86621472e-01 -9.90410894e-02 2.05280200e-01 9.54649031e-01
-9.13054407e-01 1.06109309e+00 6.89393425e+00 8.14884365e-01
-1.25214529e+00 7.93005526e-02 5.79195142e-01 -2.97078788e-01
-5.62533915e-01 -5.19628823e-02 -2.43011564e-01 4.09216851e-01
6.06139719e-01 -3.48883480e-01 7.07390010e-01 3.22331220e-01
7.94438541e-01 -8.67094174e-02 -1.08576775e+00 8.85290861e-01
1.69283286e-01 -1.60017705e+00 5.62396049e-01 7.21061975e-02
1.07082355e+00 -5.89242637e-01 2.09448010e-01 -1.43063143e-01
9.28889215e-03 -8.13664317e-01 1.01652598e+00 8.38830352e-01
1.19886303e+00 -5.35073221e-01 2.21568719e-03 1.56171516e-01
-1.24604321e+00 3.52431759e-02 -2.01528162e-01 1.61481332e-02
6.94606185e-01 4.28049594e-01 -7.81803206e-02 3.92077208e-01
3.68901342e-01 1.00837743e+00 -3.14048260e-01 7.21549213e-01
-1.99239757e-02 9.14502665e-02 -2.87964344e-02 2.67845899e-01
2.12076738e-01 -3.85037929e-01 9.16620135e-01 1.19409013e+00
5.23817897e-01 3.95754218e-01 -3.81252263e-03 9.15437579e-01
5.01100952e-03 1.26302809e-01 -7.80021250e-01 -5.37264161e-02
2.74131056e-02 1.12385762e+00 -6.49879992e-01 -7.60471344e-01
-3.58257234e-01 1.69407773e+00 1.65343195e-01 5.35088897e-01
-1.07772553e+00 -2.64709681e-01 4.91510600e-01 2.98454493e-01
3.85988355e-01 -5.00806987e-01 9.81315747e-02 -1.35597706e+00
5.32107335e-03 -8.23806524e-01 -2.12087661e-01 -1.16983986e+00
-8.32724869e-01 7.88504899e-01 -6.28980026e-02 -1.33776569e+00
-3.07734519e-01 -2.66103864e-01 -5.45287788e-01 6.79429352e-01
-1.36564600e+00 -1.05518508e+00 -3.48896772e-01 8.63343537e-01
9.01477396e-01 1.68482840e-01 3.04134279e-01 4.73869085e-01
-3.15737486e-01 2.45507151e-01 8.41327310e-02 -2.69875139e-01
8.67180943e-01 -1.09759235e+00 5.50152957e-01 9.74763989e-01
-6.09637387e-02 4.45680737e-01 7.85065651e-01 -7.62795269e-01
-1.53783393e+00 -1.08887315e+00 7.80144215e-01 -4.53494906e-01
6.63431823e-01 -2.06052035e-01 -9.67580795e-01 6.84797645e-01
5.28541565e-01 -1.74504489e-01 -9.53525007e-02 -7.33205974e-01
1.68096855e-01 1.74256012e-01 -8.52507472e-01 9.71781552e-01
1.28576577e+00 -4.93941724e-01 -4.03057605e-01 2.44545788e-01
9.43908453e-01 -5.84034920e-01 -7.32359350e-01 1.45516232e-01
4.93700504e-01 -1.07348573e+00 9.86240864e-01 -2.97083586e-01
7.47345448e-01 -4.87554431e-01 1.74298003e-01 -1.40654457e+00
-5.58606803e-01 -1.23430872e+00 -2.44064793e-01 1.24105847e+00
4.69683893e-02 -2.40066826e-01 5.29387653e-01 4.91035372e-01
-1.26553938e-01 -4.67509419e-01 -5.60731411e-01 -5.82741797e-01
-6.64220974e-02 -3.80852431e-01 1.75587490e-01 1.04113793e+00
-3.45479734e-02 5.81375286e-02 -1.01096416e+00 -1.20489411e-01
4.50301766e-01 -7.63961226e-02 7.63972998e-01 -6.94044709e-01
-2.66695946e-01 -6.41716003e-01 -1.42512918e-01 -1.44111407e+00
-1.37462988e-01 -5.94465077e-01 -4.47310768e-02 -1.34235573e+00
1.70457959e-01 2.55975388e-02 2.17623070e-01 -1.13687534e-02
-2.66580015e-01 4.49163407e-01 5.31847298e-01 3.26294065e-01
-2.41209552e-01 6.21037245e-01 1.90370178e+00 -1.59174507e-03
-3.83782953e-01 -4.07720298e-01 -3.54368418e-01 5.86754084e-01
3.63642186e-01 -2.19510242e-01 -3.97702396e-01 -6.35961533e-01
8.51149410e-02 3.83342981e-01 3.80527437e-01 -9.39468920e-01
2.96600163e-01 -2.80280918e-01 3.44674885e-01 -1.94784001e-01
3.48207474e-01 -7.54641294e-01 3.99931490e-01 4.56346095e-01
-5.03104806e-01 2.98431218e-01 7.54402429e-02 6.79127514e-01
-1.89756989e-01 5.58545589e-02 9.16785181e-01 -3.72151621e-02
-6.89109921e-01 4.99805719e-01 -7.45873332e-01 -1.18188098e-01
1.13066792e+00 -3.97129238e-01 -1.43281203e-02 -7.30663180e-01
-1.08882391e+00 -7.38170221e-02 1.05830157e+00 5.53707004e-01
6.61806345e-01 -1.33658183e+00 -7.75499880e-01 3.34867597e-01
-3.12116891e-01 -3.91954869e-01 5.10532260e-01 1.00028229e+00
-8.38299870e-01 1.37797296e-01 -4.03137326e-01 -6.19296372e-01
-1.18543434e+00 9.69584167e-01 2.23067433e-01 -2.31983326e-02
-9.59696710e-01 5.76254010e-01 5.17272830e-01 1.62288085e-01
6.93046451e-02 -3.70774388e-01 -1.08391896e-01 9.64757800e-03
7.57001817e-01 3.69535744e-01 -3.20705444e-01 -7.64294684e-01
-2.41150390e-02 1.09546471e+00 4.85763364e-02 -5.42673051e-01
1.12036741e+00 -6.10234499e-01 -2.71514505e-02 3.22094589e-01
9.88658249e-01 3.46490860e-01 -1.73127723e+00 -4.55924962e-03
-4.87176090e-01 -6.59311473e-01 7.85884857e-02 -4.40658718e-01
-1.26863003e+00 9.35739696e-01 2.66810566e-01 2.81785935e-01
1.27484155e+00 -3.91781092e-01 8.11540246e-01 -7.83301443e-02
1.07779354e-01 -8.85597229e-01 2.62370259e-01 4.16840225e-01
1.11918902e+00 -9.57509220e-01 -1.40830547e-01 -5.11194825e-01
-6.85523987e-01 1.16684020e+00 3.98654073e-01 -2.50184178e-01
4.27782595e-01 3.73400092e-01 -5.82580939e-02 -4.84297238e-02
-9.52330828e-01 -1.69929396e-02 4.18643028e-01 6.24637842e-01
4.44511861e-01 -4.56171304e-01 -3.58609080e-01 -1.99872479e-01
1.62289485e-01 3.00319374e-01 8.26503038e-01 7.41183281e-01
-3.16001147e-01 -1.08766174e+00 -4.10573602e-01 6.63313940e-02
-3.04647237e-01 -2.50339508e-01 -2.61959732e-01 7.37410009e-01
-1.81439281e-01 7.18814194e-01 1.02539651e-01 -2.31864333e-01
2.19696909e-01 -2.17630327e-01 7.48723269e-01 -3.14725429e-01
-4.72551286e-01 3.34589392e-01 -9.34516564e-02 -8.38865519e-01
-7.00796485e-01 -6.08558059e-01 -1.00003183e+00 -5.78021467e-01
7.97322839e-02 -1.10277735e-01 4.38010067e-01 5.85666299e-01
3.86120677e-01 8.34559560e-01 7.00693667e-01 -1.59843028e+00
6.03995845e-02 -5.58339655e-01 -3.92077446e-01 5.41202009e-01
6.14589751e-01 -4.98330325e-01 -2.69088537e-01 5.81801295e-01] | [10.915070533752441, -0.7643884420394897] |
58182101-cb26-4feb-9c7e-b44a06c10b1c | clipvg-text-guided-image-manipulation-using | 2212.02122 | null | https://arxiv.org/abs/2212.02122v2 | https://arxiv.org/pdf/2212.02122v2.pdf | CLIPVG: Text-Guided Image Manipulation Using Differentiable Vector Graphics | Considerable progress has recently been made in leveraging CLIP (Contrastive Language-Image Pre-Training) models for text-guided image manipulation. However, all existing works rely on additional generative models to ensure the quality of results, because CLIP alone cannot provide enough guidance information for fine-scale pixel-level changes. In this paper, we introduce CLIPVG, a text-guided image manipulation framework using differentiable vector graphics, which is also the first CLIP-based general image manipulation framework that does not require any additional generative models. We demonstrate that CLIPVG can not only achieve state-of-art performance in both semantic correctness and synthesis quality, but also is flexible enough to support various applications far beyond the capability of all existing methods. | ['Xuning Shao', 'Zhongliang Jing', 'Minzhe Li', 'Weidong Zhang', 'Kang Chen', 'Yiren Song'] | 2022-12-05 | null | null | null | null | ['image-manipulation'] | ['computer-vision'] | [ 5.31291246e-01 -2.12277725e-01 2.74554789e-02 -2.22164497e-01
-5.94918132e-01 -4.95172977e-01 7.97242582e-01 -4.28649634e-01
7.60414079e-02 4.83336419e-01 -4.92651202e-02 -4.37364340e-01
1.47773564e-01 -8.05769324e-01 -7.77184606e-01 -6.15171373e-01
3.94338310e-01 6.03176169e-02 3.20219755e-01 -5.10431170e-01
4.25925821e-01 5.74638009e-01 -1.56930399e+00 1.14966519e-01
1.14492762e+00 7.56003797e-01 5.01947761e-01 8.50147545e-01
-3.12998027e-01 4.74852890e-01 -4.72514153e-01 -2.03380138e-01
3.43511365e-02 -6.64150059e-01 -2.46347308e-01 3.79380971e-01
7.12270617e-01 -5.59791088e-01 -2.43810818e-01 1.06426084e+00
4.91056383e-01 -1.87049229e-02 6.36354744e-01 -1.33858633e+00
-1.17106593e+00 4.04685646e-01 -6.26615226e-01 -2.73059309e-01
2.93530732e-01 5.18246591e-01 5.59914649e-01 -9.56966460e-01
7.09977627e-01 1.29497898e+00 2.99473524e-01 6.40451431e-01
-1.08134103e+00 -5.99050522e-01 4.09753501e-01 -1.97054312e-01
-9.97106671e-01 -2.66487718e-01 1.01813889e+00 -4.14818078e-01
8.98757339e-01 2.27884769e-01 6.40828788e-01 8.77026379e-01
1.77714050e-01 7.17857182e-01 1.06695902e+00 -6.18065953e-01
-3.44328328e-05 -1.29901782e-01 -1.87709004e-01 1.01533997e+00
-6.07403740e-02 2.15994075e-01 -3.96139592e-01 1.38733566e-01
1.50791466e+00 -1.41636088e-01 -4.12433147e-01 -5.63859701e-01
-1.33449566e+00 6.52697384e-01 4.20993984e-01 1.88806146e-01
-2.06983596e-01 7.33864009e-01 1.40616372e-01 1.45789251e-01
3.82300019e-01 4.15339023e-01 -2.29479700e-01 -1.82406962e-01
-1.02440095e+00 3.10231328e-01 3.66781503e-01 1.44866860e+00
6.51319027e-01 7.01613307e-01 -5.48083782e-01 5.18674552e-01
3.65602970e-01 7.22430170e-01 1.91994444e-01 -1.23529816e+00
2.91546106e-01 5.44176698e-01 7.65530095e-02 -1.20415032e+00
-1.27020460e-02 -4.07412529e-01 -8.19888473e-01 6.29420519e-01
1.87484138e-02 -4.95187193e-02 -1.24286532e+00 1.56510532e+00
1.19178526e-01 6.28974214e-02 -3.74573261e-01 7.18947828e-01
9.17055011e-01 8.29931736e-01 -3.26022767e-02 -8.30802694e-02
9.99661684e-01 -1.25378251e+00 -7.65202224e-01 5.25565483e-02
2.39029974e-01 -7.55711555e-01 1.65665841e+00 4.23585713e-01
-1.42516458e+00 -6.46120548e-01 -1.17495370e+00 -2.88001627e-01
-2.66594440e-01 1.75036415e-01 7.98535705e-01 7.91870236e-01
-1.34210157e+00 3.53237897e-01 -8.55528474e-01 -5.35296835e-02
3.99215341e-01 2.17342094e-01 -7.54145393e-03 -2.74964087e-02
-7.17507899e-01 4.96679872e-01 2.38237437e-02 -4.47304323e-02
-8.85841429e-01 -9.33260322e-01 -9.14615571e-01 5.74399084e-02
5.14768898e-01 -9.60288107e-01 1.10264015e+00 -8.27910304e-01
-1.95960224e+00 6.20313704e-01 -1.29884109e-01 -8.03890526e-02
7.85796642e-01 -1.66671053e-01 -4.26014028e-02 6.62654042e-02
-2.25800753e-01 1.04829931e+00 1.10380006e+00 -1.52706766e+00
-3.48419517e-01 -1.42434284e-01 1.67434096e-01 1.35943010e-01
-3.53326410e-01 5.09934267e-03 -1.05320728e+00 -1.06741297e+00
-8.54985788e-02 -8.00236523e-01 -4.32817400e-01 6.15202665e-01
-4.58461225e-01 -6.25435337e-02 1.23680937e+00 -3.71077299e-01
1.22017288e+00 -2.03279829e+00 4.65772636e-02 -3.01727150e-02
1.21992566e-01 4.74697620e-01 -3.81375432e-01 3.76609713e-01
2.27885500e-01 4.38003451e-01 -2.66104251e-01 -3.08710188e-01
1.80035442e-01 4.57239673e-02 -4.48014617e-01 6.27786964e-02
8.96744877e-02 1.22129977e+00 -7.79962659e-01 -4.48369771e-01
6.83915198e-01 7.55302489e-01 -8.05273831e-01 -6.48131371e-02
-7.17808127e-01 4.36676115e-01 -4.50894922e-01 6.59853697e-01
7.67111719e-01 -1.56054989e-01 -1.51791796e-01 -3.20750684e-01
-1.64402291e-01 -2.80020565e-01 -9.58215594e-01 1.93994200e+00
-4.68138099e-01 6.64653778e-01 6.93402812e-02 -6.09426975e-01
8.21268082e-01 1.04158428e-02 3.38416755e-01 -5.89969158e-01
1.23284817e-01 1.85471121e-03 -3.54069829e-01 -2.52776146e-01
6.79308951e-01 3.27874362e-01 -4.92924266e-02 3.06454569e-01
-2.57426262e-01 -7.22557664e-01 1.58188283e-01 1.51690990e-01
5.47646821e-01 5.17136037e-01 -3.33441831e-02 -2.32810125e-01
4.39901233e-01 2.22990494e-02 2.84154505e-01 8.87714803e-01
1.56758830e-01 9.51042831e-01 2.83194184e-01 1.94751933e-01
-1.22872901e+00 -1.11050200e+00 1.23541296e-01 1.00052643e+00
4.37121838e-01 -4.36742038e-01 -8.56730342e-01 -4.58094329e-01
-1.26789644e-01 9.03916419e-01 -3.67058575e-01 9.19559374e-02
-4.98260915e-01 -4.36795443e-01 3.90641719e-01 7.26776361e-01
7.56753087e-01 -9.91796255e-01 -5.07275999e-01 1.51966169e-01
2.15718985e-01 -1.13331366e+00 -8.92191589e-01 -3.40177923e-01
-9.25116479e-01 -8.52309048e-01 -8.14464748e-01 -8.67949426e-01
9.79964256e-01 5.54098725e-01 1.00101423e+00 3.27572167e-01
-4.59199667e-01 6.40392184e-01 -3.00184846e-01 -3.01371008e-01
-4.62108076e-01 -2.53028840e-01 -5.23542821e-01 -3.40784431e-01
-4.34063464e-01 -5.40194511e-01 -5.14141858e-01 3.75622749e-01
-1.19703245e+00 6.01126730e-01 3.72487187e-01 8.86289954e-01
7.18037367e-01 1.36079252e-01 4.34924960e-01 -7.63707817e-01
6.93831265e-01 2.93854773e-01 -7.80642271e-01 2.36779213e-01
-6.42024040e-01 -1.39397662e-02 6.05527818e-01 -4.55530703e-01
-1.22933996e+00 -1.03433244e-01 -1.22829594e-01 -4.99549389e-01
-9.47772861e-02 4.74857569e-01 -2.84414232e-01 -3.02235246e-01
3.86673182e-01 4.30262774e-01 -5.44157065e-02 -3.26179355e-01
8.02487552e-01 3.91607881e-01 8.32952201e-01 -7.15763748e-01
9.18974459e-01 4.16415662e-01 -2.72887610e-02 -8.29398036e-01
-3.49876136e-01 6.18701987e-02 -6.09447598e-01 -5.94204888e-02
8.73339891e-01 -6.93398118e-01 -5.50398827e-01 6.66012824e-01
-1.02048361e+00 -6.86324656e-01 2.97377165e-02 7.84055218e-02
-9.07955348e-01 4.89421248e-01 -6.12746775e-01 -5.55058181e-01
-4.46949452e-01 -1.42811823e+00 1.26224923e+00 1.75344944e-01
3.22800428e-01 -1.01421511e+00 -5.93264759e-01 1.30244270e-01
7.70666957e-01 5.43779850e-01 9.37191427e-01 3.68643075e-01
-9.94212687e-01 -1.96130071e-02 -3.10243994e-01 2.21132755e-01
1.84310600e-01 4.78892118e-01 -4.93086934e-01 -3.74886245e-01
-3.21052998e-01 -1.57935202e-01 7.25726068e-01 4.69391227e-01
1.50824380e+00 -1.37822852e-01 -3.20466518e-01 1.07559931e+00
1.49299216e+00 4.29565191e-01 1.03477240e+00 2.85615325e-01
9.50793803e-01 -5.70762269e-02 6.66164041e-01 4.04606938e-01
2.67122895e-01 9.49474812e-01 4.40066844e-01 -5.14813185e-01
-4.61750150e-01 -5.81302881e-01 3.03049684e-01 4.69714880e-01
-5.01717143e-02 -4.85284507e-01 -5.64614236e-01 3.45756948e-01
-1.88860631e+00 -7.05782413e-01 -1.92470580e-01 1.80487645e+00
7.06703901e-01 -4.41270806e-02 -2.22017720e-01 -1.06784418e-01
4.98125464e-01 2.16478884e-01 -7.85735369e-01 -3.93409669e-01
-1.74278580e-02 3.28783423e-01 3.59308422e-01 3.31299931e-01
-1.02447391e+00 1.35555494e+00 7.38403320e+00 8.77504289e-01
-1.25037575e+00 -1.64788291e-01 4.52427030e-01 1.96094975e-01
-7.82731116e-01 -7.13875070e-02 -6.69597208e-01 2.78248906e-01
1.87365547e-01 -2.50349015e-01 3.19457144e-01 8.56422663e-01
4.16602135e-01 2.45572440e-02 -8.06338787e-01 1.01908672e+00
2.39870086e-01 -1.49076831e+00 4.20835614e-01 -3.17180553e-03
1.03506911e+00 -5.22883177e-01 4.67366338e-01 1.42308399e-01
2.90833622e-01 -1.05479956e+00 6.68906271e-01 5.32016695e-01
1.11179638e+00 -8.75917614e-01 2.20803812e-01 2.83192068e-01
-1.02999794e+00 4.30242926e-01 -2.42077470e-01 3.28736693e-01
3.90185088e-01 4.84321386e-01 -4.83899474e-01 4.30614293e-01
3.59572589e-01 6.23252928e-01 -6.12433076e-01 9.74292696e-01
-3.82726133e-01 5.77183545e-01 -9.47712287e-02 1.85210891e-02
2.43280858e-01 -1.10651001e-01 5.78919411e-01 1.16628265e+00
5.54863870e-01 3.85469720e-02 4.85607952e-01 1.16360950e+00
4.54619601e-02 1.74738824e-01 -5.21417141e-01 -5.36571480e-02
3.23997825e-01 9.73506927e-01 -7.16858327e-01 -4.80410814e-01
-4.02762055e-01 1.37950337e+00 -9.54807550e-02 5.05618453e-01
-1.23763239e+00 -3.55445683e-01 5.45529425e-01 2.18915671e-01
5.65541685e-01 -8.00239742e-01 -4.03809816e-01 -1.12567031e+00
-1.90220267e-01 -8.32640052e-01 -2.84351051e-01 -1.21997046e+00
-8.84914279e-01 3.33988100e-01 1.00905322e-01 -1.11613512e+00
-4.26010430e-01 -6.52425468e-01 -4.71884131e-01 8.60374272e-01
-1.63004589e+00 -1.59960735e+00 -6.33349180e-01 6.90112054e-01
7.08530664e-01 -2.93955356e-02 6.47583067e-01 -3.65665071e-02
-3.99816513e-01 8.77312064e-01 3.03924400e-02 -1.82913467e-01
5.88373899e-01 -1.14696741e+00 5.36849797e-01 9.87230897e-01
-9.69165415e-02 6.83146060e-01 6.84662282e-01 -6.03920162e-01
-1.89347076e+00 -1.17053330e+00 1.36515290e-01 -1.93053424e-01
2.66151130e-01 -2.71533042e-01 -6.34218276e-01 6.14761472e-01
3.50748658e-01 -6.97467625e-02 1.38196215e-01 -4.21081126e-01
-1.87609777e-01 8.42803121e-02 -1.09779704e+00 1.02429342e+00
1.25054491e+00 -4.04571980e-01 -1.81781977e-01 2.17735648e-01
1.00473773e+00 -7.61740744e-01 -6.66229665e-01 5.55134237e-01
2.80173779e-01 -9.50708270e-01 1.10898697e+00 -2.22364753e-01
6.10968530e-01 -6.30277216e-01 -2.00397536e-01 -1.13143575e+00
-3.90605628e-01 -7.07358301e-01 -1.67398546e-02 1.31869864e+00
1.24658391e-01 -3.66462290e-01 8.13099861e-01 5.98119915e-01
-3.83467168e-01 -6.39716923e-01 -2.03777775e-01 -8.90214562e-01
-1.19332355e-02 -7.23309934e-01 7.33916938e-01 7.57547259e-01
-3.81368428e-01 -5.30069135e-02 -5.21519482e-01 1.32037014e-01
5.32572091e-01 3.77562374e-01 1.21668625e+00 -7.10550129e-01
-3.60859424e-01 -7.60403991e-01 -4.89037544e-01 -1.50207686e+00
-1.19945630e-01 -7.13478029e-01 9.32064950e-02 -1.81091285e+00
1.78696275e-01 -3.48877668e-01 1.61437944e-01 6.34428263e-01
-3.21964920e-01 4.71200734e-01 5.02461731e-01 -9.60515463e-04
-2.76117593e-01 6.98036849e-01 1.84651113e+00 -1.77432969e-01
-3.91017258e-01 -3.12108308e-01 -9.40015614e-01 5.85891485e-01
6.72273576e-01 1.79552034e-01 -7.67239928e-01 -6.16946042e-01
-1.82299033e-01 7.07611814e-02 4.65381116e-01 -9.48626459e-01
1.26046926e-01 -5.81665158e-01 2.52837926e-01 -7.71415472e-01
2.49767751e-01 -6.24342859e-01 2.02607021e-01 2.14269012e-01
-3.56280267e-01 8.08136165e-02 3.12547743e-01 4.83684093e-01
-7.41046444e-02 -3.30234058e-02 7.74774969e-01 -7.05746487e-02
-9.45897639e-01 5.31668186e-01 -1.95106611e-01 -3.66931409e-01
1.08026278e+00 -2.47357041e-01 -4.31867272e-01 -5.87626636e-01
-5.24166167e-01 1.67973518e-01 8.94617319e-01 4.74259675e-01
8.48838031e-01 -1.38070750e+00 -6.58641994e-01 3.33099455e-01
6.24314733e-02 1.27132526e-02 1.59863263e-01 4.69400406e-01
-7.13225901e-01 3.64955992e-01 -1.29830569e-01 -7.27624834e-01
-1.33608091e+00 6.37663960e-01 1.36377424e-01 -1.02798827e-01
-8.25154841e-01 6.09176457e-01 5.57132304e-01 -1.61784768e-01
1.85130551e-01 -4.39492792e-01 2.34070584e-01 -5.85378349e-01
5.73991299e-01 7.50197321e-02 -1.51017874e-01 -2.54462570e-01
4.14565615e-02 8.27294827e-01 1.86001156e-02 -2.98270643e-01
1.18397319e+00 -1.38916016e-01 8.75843465e-02 5.80955669e-02
9.73373413e-01 2.11828575e-02 -1.53498805e+00 9.29442719e-02
-7.27954626e-01 -6.60132051e-01 4.04172152e-01 -9.07287300e-01
-1.34447956e+00 1.02944803e+00 4.12802756e-01 -1.89967260e-01
1.28623176e+00 -3.86784256e-01 8.85184586e-01 1.02513440e-01
5.37967205e-01 -9.06738698e-01 4.15505171e-01 5.09334564e-01
1.22487962e+00 -8.69551361e-01 4.91114780e-02 -7.96459913e-01
-6.38203621e-01 1.10889375e+00 6.67104840e-01 -1.53348058e-01
5.97180247e-01 5.56079745e-01 -4.96693291e-02 -1.99175984e-01
-5.27478278e-01 -1.87012389e-01 4.99497086e-01 9.94136453e-01
5.28000057e-01 -1.13521904e-01 -3.97344977e-01 9.69259441e-02
-1.61587924e-01 2.93079466e-01 4.08631802e-01 9.50289905e-01
-5.02538383e-01 -1.17993402e+00 -2.30460495e-01 2.45643929e-01
-4.81463447e-02 -2.71243751e-01 -1.33442730e-01 8.55206668e-01
-8.20129663e-02 7.88405299e-01 -1.14031419e-01 -2.57175833e-01
4.31263864e-01 -3.88861537e-01 7.83418179e-01 -4.85005707e-01
-2.49057665e-01 4.01032209e-01 -2.23101526e-01 -5.80670178e-01
-3.47303540e-01 -2.71482319e-01 -1.35819972e+00 -3.65880609e-01
-3.91796798e-01 -4.09594774e-01 8.59951913e-01 5.68807364e-01
6.07538521e-01 7.28436947e-01 5.52051365e-01 -1.13003087e+00
-1.21485248e-01 -5.34446657e-01 -3.81649524e-01 2.93044150e-01
2.76805293e-02 -6.71533704e-01 -1.21644549e-01 3.55606169e-01] | [11.401976585388184, -0.46513187885284424] |
aeedc158-d8dd-4848-84bb-7468daf20130 | a-comparative-study-of-fingerprint-image | 2111.07432 | null | https://arxiv.org/abs/2111.07432v1 | https://arxiv.org/pdf/2111.07432v1.pdf | A Comparative Study of Fingerprint Image-Quality Estimation Methods | One of the open issues in fingerprint verification is the lack of robustness against image-quality degradation. Poor-quality images result in spurious and missing features, thus degrading the performance of the overall system. Therefore, it is important for a fingerprint recognition system to estimate the quality and validity of the captured fingerprint images. In this work, we review existing approaches for fingerprint image-quality estimation, including the rationale behind the published measures and visual examples showing their behavior under different quality conditions. We have also tested a selection of fingerprint image-quality estimation algorithms. For the experiments, we employ the BioSec multimodal baseline corpus, which includes 19200 fingerprint images from 200 individuals acquired in two sessions with three different sensors. The behavior of the selected quality measures is compared, showing high correlation between them in most cases. The effect of low-quality samples in the verification performance is also studied for a widely available minutiae-based fingerprint matching system. | ['Josef Bigun', 'Klaus Kollreider', 'Hartwig Fronthaler', 'Joaquin Gonzalez-Rodriguez', 'Javier Ortega-Garcia', 'Julian Fierrez', 'Fernando Alonso-Fernandez'] | 2021-11-14 | null | null | null | null | ['image-quality-estimation'] | ['computer-vision'] | [ 6.87630057e-01 -4.04035777e-01 -2.05407515e-01 -6.86457157e-01
-4.52171445e-01 -7.33288765e-01 4.12080407e-01 1.20047629e-01
-3.91037852e-01 6.65337622e-01 -3.83602321e-01 6.78245649e-02
-4.88633662e-01 -8.59191298e-01 -7.60967076e-01 -6.37216330e-01
-2.62431353e-01 2.95122325e-01 -1.36626363e-01 1.83773205e-01
6.65817559e-01 8.15062642e-01 -1.88337195e+00 2.45789737e-01
9.30725694e-01 1.23673499e+00 6.27224892e-02 7.59981215e-01
1.27122819e-01 4.98954626e-03 -8.02469730e-01 -7.71186888e-01
3.25966746e-01 -1.64539948e-01 -3.81197482e-01 -2.21017644e-01
1.11773312e+00 -4.51109648e-01 7.77022392e-02 1.03379762e+00
5.40967345e-01 -5.60317934e-01 6.12935126e-01 -1.03163552e+00
-2.46679902e-01 2.63780355e-01 -2.56570250e-01 -6.31037205e-02
5.99106014e-01 -1.50953427e-01 5.20714283e-01 -6.48730755e-01
8.54591429e-01 1.12543285e+00 9.51480150e-01 2.48604760e-01
-1.60727668e+00 -8.37755501e-01 -7.68690944e-01 3.06650132e-01
-1.40138853e+00 -5.47079384e-01 5.69754183e-01 -3.50224137e-01
1.74636006e-01 5.48825204e-01 3.91864628e-01 1.12657797e+00
3.36831003e-01 1.59673050e-01 1.54922926e+00 -5.98731935e-01
-2.04583108e-02 3.55551660e-01 2.76154310e-01 5.62503219e-01
9.00517106e-01 6.03817701e-01 -4.96925890e-01 -5.21449387e-01
7.67700791e-01 -9.24316570e-02 -2.62831319e-02 -5.83989501e-01
-1.10552251e+00 2.53709018e-01 1.94634408e-01 7.45300472e-01
-2.81713665e-01 6.47364035e-02 6.48688525e-03 5.72363496e-01
-2.62013495e-01 5.25797367e-01 2.35399485e-01 -8.72581825e-02
-1.23327684e+00 3.65032703e-01 6.84930623e-01 7.90381432e-01
8.28616858e-01 -3.82665217e-01 -3.84779811e-01 8.78864288e-01
3.83105099e-01 1.14558160e+00 -1.42607361e-01 -1.02501070e+00
3.13999593e-01 5.18729985e-01 4.68237579e-01 -1.58447170e+00
-2.12281823e-01 -3.61829638e-01 -6.65751994e-01 4.48427647e-01
8.14789295e-01 2.42217124e-01 -9.26717639e-01 1.21778083e+00
-5.22697419e-02 -2.16848001e-01 -3.30329448e-01 6.44359410e-01
4.67477739e-01 8.46343637e-02 -2.20698357e-01 -3.50125879e-03
1.38111031e+00 -1.34708390e-01 -7.50854790e-01 1.10051394e-01
-3.22186381e-01 -1.12305057e+00 6.09254658e-01 6.66299164e-01
-8.04343462e-01 -1.01978469e+00 -1.45228946e+00 3.93096119e-01
-3.30575019e-01 3.80524158e-01 2.40038440e-01 1.54362774e+00
-7.56121814e-01 9.10410285e-01 -3.61759454e-01 -4.98941928e-01
1.19525500e-01 4.43966389e-01 -6.51194692e-01 -2.93123424e-01
-9.83329237e-01 8.29107463e-01 3.81155014e-02 2.81921238e-01
-3.14848065e-01 -3.75767350e-01 -3.80496621e-01 -1.70907691e-01
-2.51547575e-01 -1.84804946e-02 6.65241003e-01 -6.37521684e-01
-1.12217367e+00 1.05816555e+00 -2.36070722e-01 -1.24062642e-01
8.84715199e-01 4.93261255e-02 -8.87307227e-01 2.35025331e-01
8.58949795e-02 2.49483481e-01 9.81099963e-01 -1.44924796e+00
-2.95725495e-01 -6.96695745e-01 -5.40640116e-01 -6.60861969e-01
6.67949989e-02 -1.54196054e-01 -5.02151787e-01 -5.05128920e-01
3.09026718e-01 -9.33298707e-01 1.77111864e-01 8.20611417e-02
-1.64824501e-01 4.51887041e-01 4.33983058e-01 -6.86828673e-01
1.11843562e+00 -2.12267542e+00 -5.44702172e-01 1.05464125e+00
-3.52180123e-01 2.62042761e-01 -2.64961153e-01 5.09607971e-01
2.01941267e-01 4.34973724e-02 -1.00849949e-01 -1.78915709e-02
-5.27158827e-02 1.35777980e-01 -1.24923617e-01 6.13470793e-01
-3.67546864e-02 6.17882729e-01 -5.42887390e-01 -7.34706461e-01
3.76365960e-01 5.70658207e-01 2.55751144e-02 9.55256298e-02
3.96358252e-01 2.96624094e-01 -3.58633637e-01 1.15201283e+00
1.27570653e+00 1.38447046e-01 2.79211491e-01 -5.08426785e-01
-1.50888965e-01 -2.93371916e-01 -1.51478326e+00 1.41694617e+00
-1.48206353e-01 6.47213757e-01 -1.21979322e-02 -5.23766756e-01
1.25323701e+00 2.62462735e-01 5.30031025e-01 -1.28831887e+00
-1.55633286e-01 7.46132910e-01 1.07027553e-02 -5.40035665e-01
6.36791229e-01 1.14379458e-01 2.02770084e-01 3.68634552e-01
1.19583011e-02 2.50324935e-01 4.31356579e-01 -3.39877844e-01
6.77580535e-01 -1.09012341e-02 1.37036378e-02 -3.29389364e-01
7.09016263e-01 -2.98282534e-01 1.61580861e-01 1.27383101e+00
-2.87524492e-01 7.80574381e-01 4.14384902e-01 -2.78381616e-01
-1.16434443e+00 -1.08765280e+00 -8.68366838e-01 3.01627249e-01
4.35409218e-01 -1.24389097e-01 -7.35738933e-01 -1.76106110e-01
6.41265273e-01 3.79940458e-02 -9.39642906e-01 2.73712367e-01
-4.89824235e-01 -5.91280222e-01 1.06795919e+00 -5.71333012e-03
5.33045709e-01 -8.22091222e-01 -7.03815222e-01 2.38804922e-01
-1.92000628e-01 -9.89152551e-01 4.43494469e-02 -4.29362118e-01
-8.88989806e-01 -1.29178035e+00 -1.02607739e+00 -2.05738232e-01
6.21693850e-01 1.09879516e-01 9.53622401e-01 2.84512341e-01
-5.04481852e-01 3.73560697e-01 -1.80441871e-01 -1.95536777e-01
-6.95368946e-01 -9.11815539e-02 -3.90549228e-02 4.43986237e-01
5.99604659e-02 -2.29636252e-01 -7.25670874e-01 9.91953552e-01
-7.16783404e-01 -7.01016188e-01 8.95217478e-01 7.22056031e-01
5.88656306e-01 -5.02662539e-01 2.09538519e-01 -6.03035748e-01
6.51607215e-01 6.87438920e-02 -7.91926444e-01 7.25417137e-01
-1.21270478e+00 2.31295586e-01 1.15539223e-01 -1.48798719e-01
-9.81972754e-01 -1.46056175e-01 9.60543230e-02 2.75754213e-01
-1.60911158e-01 1.13273971e-01 -2.49436200e-01 -9.34109926e-01
9.90121841e-01 2.41938218e-01 2.85378128e-01 -6.27002716e-01
-2.37906098e-01 8.78994226e-01 7.65450239e-01 -6.92492783e-01
6.95353150e-01 5.28009117e-01 2.96046793e-01 -1.02878809e+00
8.96897465e-02 -4.86059219e-01 -2.61744201e-01 -5.90430737e-01
1.99985445e-01 -2.73066491e-01 -1.05114913e+00 8.17136168e-01
-9.69651341e-01 4.23849314e-01 2.20664307e-01 4.42011148e-01
-3.57497692e-01 8.04019928e-01 -2.14807138e-01 -9.85347331e-01
-3.99082273e-01 -1.24970412e+00 8.85383308e-01 4.25152063e-01
6.40849546e-02 -3.70230228e-01 3.38009953e-01 5.76707840e-01
7.41676509e-01 3.69625151e-01 4.42489088e-01 4.31244299e-02
-4.15917307e-01 -6.05903447e-01 -5.27382851e-01 2.96366423e-01
2.14418784e-01 2.30227143e-01 -1.28041852e+00 -2.95941144e-01
-3.10397118e-01 6.97310716e-02 7.21167445e-01 2.97350019e-01
8.43098342e-01 1.46212101e-01 -4.53543961e-01 5.61401546e-01
1.65691292e+00 2.16778561e-01 1.15482509e+00 3.84224057e-01
1.05391175e-01 7.37599909e-01 9.24262166e-01 1.24737620e-01
-4.47053671e-01 1.04629803e+00 2.28228271e-01 -1.50968404e-02
4.90316097e-03 -1.04250252e-01 -2.86692400e-02 -2.76650369e-01
-7.23282993e-01 -1.53242841e-01 -8.01672220e-01 7.86086023e-02
-1.45939302e+00 -1.32057977e+00 -3.43211174e-01 2.60268569e+00
3.53847563e-01 -1.41079769e-01 2.29097575e-01 3.51706058e-01
9.72545445e-01 8.47053230e-02 -2.27540404e-01 -2.57015258e-01
-2.69969225e-01 2.67825752e-01 8.70714009e-01 2.92333156e-01
-1.01148474e+00 5.40135577e-02 7.43983650e+00 4.05534953e-01
-1.22261012e+00 -3.20627332e-01 3.72999698e-01 3.13811392e-01
-1.53999567e-01 -1.75930336e-01 -5.81386209e-01 6.91778719e-01
8.20463300e-01 1.83522269e-01 2.14994222e-01 5.77515304e-01
-3.58234867e-02 -5.24835825e-01 -7.89650738e-01 1.21966910e+00
1.32808283e-01 -1.16316187e+00 -2.09886700e-01 3.31965238e-01
4.99718398e-01 -2.60241389e-01 1.06241114e-01 -5.50805748e-01
-7.85705388e-01 -1.05761445e+00 6.50780737e-01 8.83171797e-01
1.02304077e+00 -6.02986753e-01 9.05551016e-01 -3.11205775e-01
-8.95740688e-01 9.38070789e-02 -4.03670698e-01 4.87372488e-01
-2.51109749e-02 7.54197538e-01 -6.47367656e-01 6.79754019e-01
6.94851339e-01 5.45079336e-02 -7.91184068e-01 1.38965237e+00
1.67275250e-01 2.45904610e-01 -3.92465264e-01 -1.62452430e-01
-3.52624208e-01 -3.06853831e-01 3.31567377e-01 1.38977993e+00
5.60712695e-01 -5.40346861e-01 -8.01902592e-01 1.05960023e+00
2.29753017e-01 1.04016453e-01 -3.83000463e-01 -1.84249058e-01
6.18780434e-01 1.07055569e+00 -6.16185486e-01 -1.30254671e-01
-1.59818247e-01 9.00514007e-01 -4.99805540e-01 2.91845053e-01
-6.28022432e-01 -7.01779127e-01 4.81229573e-01 2.23816738e-01
1.74492389e-01 -1.87580422e-01 -3.87211084e-01 -7.98161089e-01
4.75072026e-01 -9.63094294e-01 4.77468073e-02 -2.83830076e-01
-9.87628162e-01 3.36540401e-01 -1.41083270e-01 -1.26473081e+00
-1.91176593e-01 -6.43285632e-01 -1.22901410e-01 1.07263207e+00
-1.24897838e+00 -1.09317994e+00 -6.90783501e-01 3.49111110e-01
-3.28119755e-01 -2.02114150e-01 8.07478607e-01 7.52324581e-01
-6.64630458e-02 9.70602572e-01 3.05013150e-01 -9.49445292e-02
9.26912010e-01 -9.40912724e-01 5.52426279e-01 8.12094927e-01
1.16522945e-01 9.56128299e-01 7.97487617e-01 -7.01611459e-01
-1.47249210e+00 -2.82691002e-01 9.25476253e-01 -3.37595642e-01
-1.74784884e-02 -2.33965784e-01 -7.75423348e-01 -5.87495007e-02
-1.59217402e-01 -6.36554584e-02 5.65296769e-01 -2.02401634e-02
-4.98667210e-01 -6.17017031e-01 -1.56499887e+00 5.12267016e-02
8.52193177e-01 -5.18825769e-01 -1.57113016e-01 -3.41340512e-01
-6.06950521e-01 -4.90466654e-02 -1.16499078e+00 6.73430026e-01
1.80809975e+00 -1.55210137e+00 1.22001612e+00 1.28611028e-01
-1.51526138e-01 -3.63606900e-01 -8.44216049e-02 -5.01103580e-01
-2.86978006e-01 -1.60190016e-01 2.73858041e-01 1.28208184e+00
4.51480061e-01 -6.74496651e-01 9.18028653e-01 4.74673390e-01
6.36576116e-01 -7.69295916e-02 -8.28689396e-01 -1.05276525e+00
-6.87543511e-01 -2.12368175e-01 6.65586352e-01 5.77456653e-01
-1.97395086e-01 -3.91878188e-01 -4.87009674e-01 8.62466842e-02
1.25803924e+00 4.40172523e-01 9.46253657e-01 -1.38588870e+00
-1.52413145e-01 -2.93291956e-01 -8.31374407e-01 -3.79439741e-01
-6.19743168e-01 -3.10172290e-01 -1.88690841e-01 -1.00212729e+00
2.47376248e-01 -6.62957489e-01 -3.24230373e-01 -7.08899647e-02
1.30171254e-01 5.53283632e-01 1.05385795e-01 3.36022943e-01
9.39423442e-02 -3.66471022e-01 1.10162759e+00 -2.29553014e-01
1.42133549e-01 3.31182361e-01 -1.98095798e-01 -6.61930516e-02
5.74771643e-01 -4.54994738e-01 1.32197682e-02 -3.37617397e-02
2.73054689e-01 1.05582541e-02 5.51757514e-01 -1.40410173e+00
-9.42287617e-04 -3.13432142e-02 7.88852692e-01 -6.28569007e-01
2.07654744e-01 -9.80507970e-01 8.05067241e-01 9.44967747e-01
-1.19238377e-01 -1.77587524e-01 -1.13701865e-01 2.98512578e-01
-4.45119858e-01 -3.95111173e-01 9.27984715e-01 2.13543713e-01
-7.15307474e-01 5.26314043e-02 6.32429719e-02 -6.55810177e-01
5.95236957e-01 -8.88565302e-01 -4.26178217e-01 -8.18976685e-02
-3.52797538e-01 -3.68346512e-01 8.92714858e-01 4.32378650e-01
5.94227076e-01 -1.35293019e+00 -6.55558765e-01 6.64178789e-01
5.97181499e-01 -1.02924156e+00 3.93961698e-01 7.34678388e-01
-9.84159172e-01 4.57742959e-01 -9.42962468e-01 -8.26777339e-01
-1.67523813e+00 3.28800052e-01 4.60325837e-01 7.55808204e-02
2.23507151e-01 3.93491447e-01 -6.23384058e-01 -2.57284105e-01
3.19140255e-01 -1.91132605e-01 -4.73911129e-02 8.08969662e-02
6.71381891e-01 7.53384173e-01 3.43876272e-01 -4.91387486e-01
-6.18711770e-01 9.99072790e-01 2.94012904e-01 -2.89575338e-01
8.12567234e-01 4.77046371e-02 -1.65532604e-01 1.12930499e-01
1.13082945e+00 2.73198366e-01 -6.28127754e-01 1.11735903e-01
2.43088052e-01 -9.58871841e-01 -4.83086288e-01 -8.76747489e-01
-7.77098775e-01 6.56134486e-01 1.51513147e+00 2.98891723e-01
9.68577445e-01 -4.63342100e-01 2.57757336e-01 3.07827353e-01
7.90012777e-01 -1.18132305e+00 -3.47828835e-01 -1.51688650e-01
8.94822478e-01 -1.28792465e+00 2.29305893e-01 -2.46715486e-01
-1.45743750e-02 1.39533901e+00 2.24694498e-02 6.91707432e-02
3.09417486e-01 5.46137616e-02 2.43002132e-01 -1.05342723e-01
2.83320963e-01 2.09326923e-01 5.33922732e-01 8.43342841e-01
5.42855680e-01 1.80221125e-01 -8.28359842e-01 7.51339942e-02
3.24889533e-02 3.43082905e-01 2.20434994e-01 8.97197723e-01
-2.73510039e-01 -1.74018705e+00 -9.32949305e-01 4.19540733e-01
-5.38546562e-01 4.71502453e-01 -6.74683213e-01 7.53644347e-01
4.03244942e-01 1.01428437e+00 -2.90492058e-01 -5.60603321e-01
5.89910030e-01 -2.44931802e-01 9.53735650e-01 2.17385188e-01
-6.39914215e-01 -2.55562246e-01 1.85473680e-01 -9.06900823e-01
-7.33673036e-01 -9.05729711e-01 -5.72469354e-01 -4.49477583e-01
-3.84796441e-01 1.50604948e-01 1.29718244e+00 4.18953300e-01
3.14799786e-01 -2.66049176e-01 6.67608738e-01 -5.22789598e-01
-4.25816178e-01 -8.67388070e-01 -9.78127718e-01 5.03521025e-01
4.46761161e-01 -6.65228546e-01 -1.19591981e-01 -1.45217046e-01] | [12.967262268066406, 0.9896544218063354] |
8501c845-5d1e-4f6d-b28d-cfcc34d80f59 | dnf-decouple-and-feedback-network-for-seeing | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Jin_DNF_Decouple_and_Feedback_Network_for_Seeing_in_the_Dark_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Jin_DNF_Decouple_and_Feedback_Network_for_Seeing_in_the_Dark_CVPR_2023_paper.pdf | DNF: Decouple and Feedback Network for Seeing in the Dark | The exclusive properties of RAW data have shown great potential for low-light image enhancement. Nevertheless, the performance is bottlenecked by the inherent limitations of existing architectures in both single-stage and multi-stage methods. Mixed mapping across two different domains, noise-to-clean and RAW-to-sRGB, misleads the single-stage methods due to the domain ambiguity. The multi-stage methods propagate the information merely through the resulting image of each stage, neglecting the abundant features in the lossy image-level dataflow. In this paper, we probe a generalized solution to these bottlenecks and propose a Decouple aNd Feedback framework, abbreviated as DNF. To mitigate the domain ambiguity, domainspecific subtasks are decoupled, along with fully utilizing the unique properties in RAW and sRGB domains. The feature propagation across stages with a feedback mechanism avoids the information loss caused by image-level dataflow. The two key insights of our method resolve the inherent limitations of RAW data-based low-light image enhancement satisfactorily, empowering our method to outperform the previous state-of-the-art method by a large margin with only 19% parameters, achieving 0.97dB and 1.30dB PSNR improvements on the Sony and Fuji subsets of SID. | ['Chongyi Li', 'Zhi Chai', 'Chun-Le Guo', 'Zhen Li', 'Ling-Hao Han', 'Xin Jin'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['image-enhancement', 'low-light-image-enhancement'] | ['computer-vision', 'computer-vision'] | [ 5.30181527e-01 -2.90642858e-01 2.05768440e-02 -2.71288306e-01
-6.80539429e-01 -2.60329306e-01 2.84441829e-01 -3.59815449e-01
-4.49325383e-01 8.80189061e-01 6.16390035e-02 -1.93851903e-01
1.62310079e-02 -6.98529780e-01 -5.68574190e-01 -9.34033751e-01
3.78082275e-01 -6.56282783e-01 5.90017200e-01 -5.47316134e-01
1.64645299e-01 1.98210642e-01 -1.72386634e+00 5.97012222e-01
1.09177709e+00 1.35209084e+00 5.19800603e-01 6.16950035e-01
-4.67343479e-01 9.45002258e-01 -4.29992259e-01 -3.71286243e-01
5.61914980e-01 -3.96882743e-01 -4.16831583e-01 8.76761153e-02
6.88783348e-01 -7.27583468e-01 -3.81935626e-01 1.40401924e+00
6.76324368e-01 -2.52800971e-01 -1.59446206e-02 -1.14432931e+00
-8.51150751e-01 1.02277614e-01 -1.04444420e+00 2.30729476e-01
4.89380695e-02 4.57355589e-01 8.08945298e-01 -1.05419028e+00
6.01655006e-01 9.55659151e-01 6.25545323e-01 5.31865001e-01
-1.29346001e+00 -8.14533412e-01 1.33969942e-02 2.84675866e-01
-1.23027098e+00 -8.26637864e-01 6.36674225e-01 -3.16694498e-01
7.38716722e-01 1.50503471e-01 3.47008556e-01 6.16410553e-01
1.06427118e-01 5.85428476e-01 1.63034034e+00 -3.58610779e-01
-1.41855419e-01 1.63651317e-01 1.36438727e-01 6.74327374e-01
3.64538670e-01 4.46193784e-01 -8.73150110e-01 4.76293862e-01
7.19768047e-01 -1.38300806e-01 -4.54715133e-01 -8.50613639e-02
-1.03605950e+00 1.71269685e-01 5.12656093e-01 1.97798803e-01
-2.67930657e-01 -2.53466312e-02 2.91178197e-01 5.41900516e-01
3.92167658e-01 1.77073389e-01 -4.27298576e-01 -2.74414252e-02
-1.13934839e+00 -1.19715013e-01 1.88712835e-01 1.19493163e+00
1.06253183e+00 2.18865067e-01 -3.80075157e-01 7.38755345e-01
2.03225210e-01 6.63614273e-01 1.48775533e-01 -9.24353957e-01
6.18312120e-01 4.45459634e-01 1.47898555e-01 -8.54122460e-01
-2.48356447e-01 -6.61828101e-01 -9.99648392e-01 6.68151855e-01
4.70735699e-01 2.24999357e-02 -8.60404134e-01 1.74309886e+00
2.50087231e-01 -1.27124237e-02 4.51227054e-02 1.23148370e+00
9.42514837e-01 6.21613860e-01 -1.13166325e-01 -3.28995168e-01
1.49652028e+00 -1.20313251e+00 -9.54753637e-01 -1.09905049e-01
5.80165982e-02 -1.00073504e+00 1.29114151e+00 5.92041850e-01
-1.44858301e+00 -9.24358189e-01 -1.26714969e+00 -4.54519272e-01
-1.04491614e-01 1.93715766e-01 3.80704254e-01 8.09044361e-01
-1.26382136e+00 4.38075781e-01 -4.25672352e-01 -5.67913800e-02
4.69647825e-01 2.15339661e-01 -2.45639578e-01 -2.85116166e-01
-1.16912532e+00 6.34298861e-01 1.00138500e-01 2.31531560e-01
-6.65463686e-01 -9.76395547e-01 -4.11108404e-01 -8.21045693e-03
5.29111981e-01 -7.25971818e-01 1.05793679e+00 -8.53867769e-01
-1.58690476e+00 6.58576608e-01 -3.03497314e-01 -2.25210011e-01
5.39584816e-01 -1.99483752e-01 -6.93476140e-01 2.20535159e-01
1.08677447e-02 5.32859921e-01 1.13946700e+00 -1.35092223e+00
-1.06794238e+00 -1.57424852e-01 1.54205948e-01 6.82945549e-02
-5.12763560e-01 1.12596013e-01 -8.56135368e-01 -5.72198153e-01
1.80265173e-01 -3.64860505e-01 -2.15679072e-02 4.26638454e-01
-1.89918265e-01 5.25750518e-01 8.44919682e-01 -6.63527846e-01
1.45081556e+00 -2.34673429e+00 -3.68046522e-01 -3.98665845e-01
4.18897033e-01 4.69028980e-01 -3.19966763e-01 1.46883130e-01
2.63841301e-02 -8.13645124e-02 -2.48870119e-01 -3.04955453e-01
-2.06939310e-01 2.81359572e-02 -2.98547566e-01 3.49776030e-01
4.40126628e-01 7.08229721e-01 -8.84073675e-01 -5.17663419e-01
1.92214906e-01 4.45152223e-01 -4.40360636e-01 2.20254898e-01
1.13169879e-01 5.04707754e-01 -8.73378441e-02 8.08424234e-01
1.45510375e+00 -8.04987550e-02 -1.30609106e-02 -8.68708551e-01
-6.20598614e-01 2.46764183e-01 -1.34018934e+00 1.83822405e+00
-5.43366492e-01 5.97530186e-01 5.72589755e-01 -5.54597616e-01
8.84012997e-01 7.58964643e-02 3.09780747e-01 -1.28551865e+00
-2.07889602e-01 4.52210695e-01 -1.53168365e-01 -5.88355601e-01
7.28618562e-01 -3.38508964e-01 2.04901144e-01 2.25917041e-01
1.14804789e-01 1.01586282e-01 2.55008966e-01 1.58346862e-01
1.00692129e+00 2.65394360e-01 1.34660989e-01 -2.32991606e-01
6.28732562e-01 -1.91866308e-01 7.91153133e-01 9.16919589e-01
-5.38160503e-01 8.63489807e-01 2.13706344e-01 -1.50986329e-01
-1.08309066e+00 -1.23923314e+00 -2.25924924e-01 1.03558791e+00
5.53434134e-01 -2.83507943e-01 -5.07461607e-01 -5.30745924e-01
-3.13301891e-01 3.00882667e-01 -2.03067645e-01 -1.22926533e-01
-4.52974766e-01 -9.94301915e-01 3.98316830e-01 3.24626029e-01
1.19541395e+00 -4.96815026e-01 -5.30013740e-01 2.62047142e-01
-2.88795441e-01 -1.48213971e+00 -4.22044605e-01 1.70051321e-01
-8.40995312e-01 -7.94582903e-01 -6.46867752e-01 -4.18879360e-01
5.21018624e-01 6.71828568e-01 1.21442783e+00 -4.38072756e-02
-4.14491862e-01 -1.25566691e-01 -2.59567529e-01 -9.96110961e-02
-3.43906075e-01 -3.21149379e-01 -2.16393813e-01 2.90434778e-01
8.73137787e-02 -5.80049455e-01 -1.14636827e+00 3.30959260e-01
-1.11169600e+00 4.11720663e-01 1.08085668e+00 1.10036719e+00
6.27990246e-01 1.88417330e-01 5.34572124e-01 -7.14432538e-01
3.99292260e-01 -1.81638092e-01 -6.52100027e-01 1.96935937e-01
-1.04986072e+00 -8.85308236e-02 6.38808489e-01 -1.56517223e-01
-1.76518118e+00 -1.77911445e-02 -4.43832390e-02 -1.93146959e-01
-1.11109421e-01 -8.13878179e-02 -4.88043189e-01 -2.39068553e-01
5.56620121e-01 3.73569906e-01 3.54279988e-02 -6.20447516e-01
4.00186926e-01 7.02582121e-01 8.68097723e-01 -3.24371725e-01
9.24038529e-01 8.12694907e-01 1.40750691e-01 -6.28608942e-01
-6.83298469e-01 -5.52111983e-01 -3.47119600e-01 -3.38160843e-01
7.42991328e-01 -1.20919549e+00 -5.61135828e-01 7.83915162e-01
-1.18761194e+00 -1.30302221e-01 -4.77559417e-01 1.88595831e-01
-1.78794786e-01 4.85504776e-01 -8.02298784e-01 -5.53567052e-01
-4.57308054e-01 -1.23231518e+00 8.93356144e-01 3.62622142e-01
4.82519478e-01 -3.80525827e-01 -4.22749996e-01 3.00281376e-01
9.24155712e-01 -1.98265344e-01 7.34012425e-01 2.27829278e-01
-9.99580503e-01 2.67296344e-01 -1.12267315e+00 8.05691302e-01
2.14733869e-01 -1.44589782e-01 -1.41408157e+00 -3.65842789e-01
2.59127468e-01 -1.35656491e-01 1.00177991e+00 2.40487263e-01
1.01708877e+00 4.44163382e-02 1.48855880e-01 9.13629174e-01
1.83977115e+00 1.87578611e-02 9.18495357e-01 4.19307679e-01
5.40527582e-01 4.28861946e-01 7.30223119e-01 4.28148657e-01
2.67323822e-01 6.79059446e-01 5.63064575e-01 -7.12012768e-01
-8.99812579e-01 -6.76378235e-02 4.87474740e-01 6.81044102e-01
-4.89233527e-03 -5.77375777e-02 -3.88524890e-01 4.30450559e-01
-1.45921326e+00 -7.04798579e-01 -4.99207497e-01 2.08864522e+00
1.02497458e+00 1.04457922e-01 -2.00228304e-01 8.83964151e-02
6.50134802e-01 3.74344677e-01 -5.44184864e-01 -1.88381985e-01
-5.25206447e-01 1.33265376e-01 7.36921668e-01 3.56080800e-01
-8.44809175e-01 6.19335055e-01 6.09310150e+00 1.03794265e+00
-1.11778271e+00 3.56596261e-01 5.59407592e-01 -2.42224619e-01
-1.98698834e-01 3.43823619e-02 -9.07903910e-01 5.44066250e-01
5.97378194e-01 1.76642671e-01 4.68306154e-01 3.46999884e-01
4.21251327e-01 -3.42555732e-01 -7.83334851e-01 1.01442552e+00
-7.08350241e-02 -1.24095857e+00 -9.65889469e-02 -7.92589188e-02
8.48662853e-01 3.12889949e-03 2.51285940e-01 2.40436494e-02
-1.16487719e-01 -5.47870398e-01 8.92757177e-01 5.27571380e-01
1.36130834e+00 -2.99539149e-01 4.68782634e-01 3.88983600e-02
-1.13257742e+00 -1.64218307e-01 -3.29365760e-01 -3.94221731e-02
3.31472635e-01 1.10408247e+00 -4.07997459e-01 8.62026930e-01
1.01306474e+00 6.23198867e-01 -5.07610798e-01 9.01887774e-01
-1.99598446e-01 4.54092681e-01 -1.79817840e-01 6.05450869e-01
-1.81405507e-02 -2.35655576e-01 6.22506082e-01 1.35446298e+00
2.83105433e-01 -4.87982109e-02 -2.71423638e-01 1.16817451e+00
-5.95754050e-02 -2.78458416e-01 -2.25058988e-01 3.23108643e-01
2.04636216e-01 1.38603246e+00 -2.26381049e-01 -2.10227996e-01
-8.71726274e-01 1.06766069e+00 -1.90263733e-01 4.95372802e-01
-8.04398954e-01 -6.96245909e-01 7.66726732e-01 3.93006921e-01
3.06935847e-01 -1.01296149e-01 -6.79774821e-01 -1.19420207e+00
3.43884110e-01 -8.09636772e-01 1.09674044e-01 -9.26579297e-01
-1.18269050e+00 5.92401922e-01 -3.17415923e-01 -1.54689968e+00
3.82908583e-01 -5.32540321e-01 -2.38416001e-01 1.25019419e+00
-2.58193326e+00 -1.11706245e+00 -6.71608448e-01 6.87692523e-01
3.67986798e-01 3.44833098e-02 2.95048505e-01 8.81045997e-01
-5.28199553e-01 6.65570498e-01 3.71841371e-01 -1.40445694e-01
1.10719728e+00 -1.06325853e+00 2.52175272e-01 1.38078940e+00
-2.88408726e-01 4.24282819e-01 4.19851393e-01 -3.28065097e-01
-1.52798927e+00 -9.26639736e-01 6.92529500e-01 9.30224881e-02
6.10396028e-01 -3.58215302e-01 -9.16158080e-01 1.35119297e-02
2.37661093e-01 3.12557727e-01 2.93043286e-01 -3.04042101e-01
-4.61490124e-01 -7.84285188e-01 -1.30981946e+00 6.01311266e-01
1.19062459e+00 -5.14691770e-01 -1.76221371e-01 -2.97401100e-01
8.68792951e-01 -3.05322111e-01 -7.16199756e-01 4.47755575e-01
4.48585451e-01 -1.55650640e+00 1.05783927e+00 -5.17254733e-02
7.36157298e-01 -7.63338268e-01 -2.61433929e-01 -9.69387174e-01
-3.17178577e-01 -9.23334301e-01 3.22653800e-02 1.57520998e+00
2.91763306e-01 -6.89667165e-01 5.16119778e-01 4.43796754e-01
-2.25545257e-01 -4.30571914e-01 -7.30040908e-01 -7.51529992e-01
-3.72731358e-01 -4.32677865e-01 7.70080090e-01 5.73146403e-01
-4.08842832e-01 2.03472242e-01 -5.66030025e-01 2.06048027e-01
9.31546092e-01 2.86815584e-01 6.82679772e-01 -7.89012730e-01
-3.60756040e-01 -2.68948555e-01 -1.38791412e-01 -1.56173909e+00
-5.75482190e-01 -5.29990494e-01 2.21542731e-01 -1.30488288e+00
4.14218873e-01 -6.10920668e-01 -5.45613348e-01 3.72374862e-01
-2.15459064e-01 5.82248390e-01 4.32442576e-01 2.24208787e-01
-6.71822190e-01 3.31588030e-01 1.60220170e+00 -5.46694733e-02
-8.69015828e-02 -4.14511532e-01 -1.03904283e+00 4.87462759e-01
4.32897449e-01 -2.76300043e-01 -4.28221375e-01 -9.62656796e-01
1.94642827e-01 -7.03657717e-02 5.05500555e-01 -1.06521738e+00
3.83451521e-01 -1.11616492e-01 2.67415881e-01 -4.33675677e-01
2.59972751e-01 -9.06205952e-01 2.68628690e-02 1.49022326e-01
-1.33480635e-02 -3.78131539e-01 2.44766623e-01 5.96938550e-01
-5.76635778e-01 3.32071967e-02 1.12448978e+00 2.22252216e-02
-1.19972491e+00 2.15833187e-01 -6.13875017e-02 7.54445270e-02
7.44183719e-01 -5.38553536e-01 -7.30981648e-01 -1.72984734e-01
-4.67626452e-01 -8.37631971e-02 5.44980764e-01 1.99765727e-01
6.35589838e-01 -1.14792109e+00 -6.93983674e-01 6.79841042e-01
7.61344656e-02 -5.25645092e-02 7.49494314e-01 1.13754416e+00
-2.54618376e-01 9.42552239e-02 -4.81836379e-01 -6.08297110e-01
-1.23217487e+00 4.31977957e-01 2.89227635e-01 -3.40547353e-01
-5.43511808e-01 7.72452712e-01 4.88879472e-01 1.31138727e-01
-1.94587260e-02 -3.54996800e-01 2.65407026e-01 -3.80262658e-02
6.99216187e-01 5.44749141e-01 3.81929159e-01 -4.27614480e-01
-7.63615668e-02 5.57950139e-01 -1.48288921e-01 4.79063988e-02
1.34736025e+00 -7.91716576e-01 -1.78158253e-01 -5.44704981e-02
1.16742980e+00 9.29289386e-02 -1.58587527e+00 -4.15463060e-01
-4.10721719e-01 -8.88964415e-01 4.72350955e-01 -1.19433010e+00
-1.38159692e+00 1.15973282e+00 9.15491223e-01 4.65694033e-02
2.03023458e+00 -4.24906552e-01 1.09493184e+00 -1.76992819e-01
3.81124794e-01 -1.08410740e+00 1.23976255e-02 2.21221119e-01
5.85031986e-01 -1.38697565e+00 -3.88787314e-02 -6.31583571e-01
-5.23181915e-01 1.02104604e+00 5.75969398e-01 4.08255249e-01
4.60850984e-01 7.41253614e-01 2.11204112e-01 1.00526288e-01
-6.69921100e-01 -3.97632241e-01 5.03513813e-02 7.82655001e-01
3.01203072e-01 -2.75446534e-01 -2.89024562e-01 6.45057797e-01
4.99153227e-01 4.80167508e-01 4.41465169e-01 8.23358595e-01
-5.03019512e-01 -1.11682987e+00 -4.85009909e-01 1.22224078e-01
-6.13751054e-01 -4.76244599e-01 1.82735607e-01 6.13087356e-01
4.26201254e-01 1.13821340e+00 -1.25911564e-01 -4.16158468e-01
5.92451394e-01 -3.15295368e-01 2.70062864e-01 -1.42585248e-01
-5.41309237e-01 1.98578879e-01 3.62933278e-02 -9.38170314e-01
-4.84804899e-01 -3.07367116e-01 -1.05381370e+00 -3.08493078e-01
-4.22943503e-01 -4.17601258e-01 5.54384291e-01 4.10993010e-01
7.57190108e-01 6.57349825e-01 7.23973036e-01 -6.92890346e-01
-5.86631000e-01 -6.50093913e-01 -7.32517064e-01 5.33287585e-01
6.81387067e-01 -4.38239723e-01 -3.15824896e-01 2.43002504e-01] | [10.822454452514648, -2.457315683364868] |
989a4fde-d093-4180-9880-b486ff57b6aa | seasonal-contrast-unsupervised-pre-training | 2103.16607 | null | https://arxiv.org/abs/2103.16607v2 | https://arxiv.org/pdf/2103.16607v2.pdf | Seasonal Contrast: Unsupervised Pre-Training from Uncurated Remote Sensing Data | Remote sensing and automatic earth monitoring are key to solve global-scale challenges such as disaster prevention, land use monitoring, or tackling climate change. Although there exist vast amounts of remote sensing data, most of it remains unlabeled and thus inaccessible for supervised learning algorithms. Transfer learning approaches can reduce the data requirements of deep learning algorithms. However, most of these methods are pre-trained on ImageNet and their generalization to remote sensing imagery is not guaranteed due to the domain gap. In this work, we propose Seasonal Contrast (SeCo), an effective pipeline to leverage unlabeled data for in-domain pre-training of remote sensing representations. The SeCo pipeline is composed of two parts. First, a principled procedure to gather large-scale, unlabeled and uncurated remote sensing datasets containing images from multiple Earth locations at different timestamps. Second, a self-supervised algorithm that takes advantage of time and position invariance to learn transferable representations for remote sensing applications. We empirically show that models trained with SeCo achieve better performance than their ImageNet pre-trained counterparts and state-of-the-art self-supervised learning methods on multiple downstream tasks. The datasets and models in SeCo will be made public to facilitate transfer learning and enable rapid progress in remote sensing applications. | ['Pau Rodriguez', 'David Vazquez', 'Xavier Giro-i-Nieto', 'Alexandre Lacoste', 'Oscar Mañas'] | 2021-03-30 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Manas_Seasonal_Contrast_Unsupervised_Pre-Training_From_Uncurated_Remote_Sensing_Data_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Manas_Seasonal_Contrast_Unsupervised_Pre-Training_From_Uncurated_Remote_Sensing_Data_ICCV_2021_paper.pdf | iccv-2021-1 | ['unsupervised-pre-training'] | ['methodology'] | [ 2.86679685e-01 -3.24158549e-01 -2.96219349e-01 -6.68261945e-01
-9.17507887e-01 -8.16284716e-01 5.93146503e-01 2.11216775e-05
-4.57050800e-01 7.84861684e-01 4.23106831e-03 -7.70231247e-01
-2.32256745e-04 -1.11760008e+00 -7.38237679e-01 -5.78464925e-01
-6.07678235e-01 3.34954917e-01 -2.06290856e-02 -3.50709260e-01
-2.64504701e-01 6.82369769e-01 -1.35728824e+00 8.30028802e-02
8.44656408e-01 7.85597503e-01 4.25952166e-01 5.65205574e-01
1.07040904e-01 5.79819977e-01 -1.73160911e-01 5.84136724e-01
5.81216574e-01 -3.30434710e-01 -9.57061768e-01 1.07352480e-01
5.73671401e-01 -5.93356431e-01 -2.51405180e-01 1.05773413e+00
5.32124281e-01 2.49376744e-01 6.08215511e-01 -9.37361300e-01
-7.51629472e-01 4.21261340e-01 -6.15953267e-01 4.63009447e-01
-5.13977468e-01 1.17821127e-01 7.93847561e-01 -8.18294346e-01
4.70789373e-01 7.73576498e-01 8.18702459e-01 4.52579051e-01
-1.03302872e+00 -7.41768956e-01 2.14605883e-01 -5.66449612e-02
-1.29918981e+00 -5.72609842e-01 3.82249266e-01 -6.20498359e-01
9.82116222e-01 1.80690244e-01 4.59671617e-01 5.71489036e-01
-1.25803396e-01 5.29240191e-01 1.33735120e+00 -2.93118924e-01
2.25821719e-01 -1.29775345e-01 -3.34568769e-02 4.32989806e-01
5.33099547e-02 2.91213393e-01 -3.02303046e-01 4.64740656e-02
8.78329098e-01 4.99731421e-01 -1.61769077e-01 1.99762508e-02
-1.26228225e+00 7.72306085e-01 1.03122044e+00 3.39576691e-01
-5.83263576e-01 1.89587131e-01 1.41023755e-01 5.83281398e-01
1.11360240e+00 2.44863346e-01 -8.72421801e-01 3.29799712e-01
-1.27423978e+00 -2.00067014e-01 2.76709557e-01 7.21250832e-01
1.40727019e+00 3.03843826e-01 5.84030867e-01 6.84280515e-01
2.71482944e-01 8.15564275e-01 5.30669332e-01 -6.09727800e-01
3.66487920e-01 5.17411172e-01 1.23389386e-01 -7.17377067e-01
-4.78754938e-01 -2.40549356e-01 -1.01677048e+00 2.73718119e-01
-1.76275298e-02 -4.73975301e-01 -1.22697139e+00 1.55240369e+00
3.46070230e-01 3.37645829e-01 2.15207487e-01 8.94842565e-01
7.00943053e-01 1.12036884e+00 2.24646196e-01 3.43611926e-01
1.00676239e+00 -9.58841562e-01 -1.81723133e-01 -6.65041208e-01
7.54147828e-01 -1.90339312e-01 8.87263596e-01 -3.34710747e-01
-3.87028039e-01 -5.76556087e-01 -1.00671089e+00 1.81041479e-01
-8.49786818e-01 2.95283169e-01 7.14098394e-01 2.45798111e-01
-1.25829744e+00 7.52659917e-01 -1.12122226e+00 -7.66205430e-01
5.84705174e-01 1.49490997e-01 -3.90024006e-01 -1.40553817e-01
-1.23982906e+00 7.40154743e-01 4.96291339e-01 3.60655606e-01
-1.11182725e+00 -7.70094395e-01 -9.81210709e-01 1.14596583e-01
-2.00259864e-01 -2.18924403e-01 1.03100133e+00 -1.23214149e+00
-1.00031972e+00 1.26291561e+00 1.62491605e-01 -5.06300628e-01
3.69922310e-01 -2.34258622e-01 -5.90730965e-01 2.20235243e-01
5.55546105e-01 9.73559320e-01 9.24391925e-01 -9.91206229e-01
-6.49581611e-01 -5.26224852e-01 -2.03377366e-01 3.17817867e-01
-4.47914958e-01 -1.42344669e-01 3.29342671e-02 -4.89910573e-01
2.98769146e-01 -9.72185969e-01 -4.77271408e-01 2.57151872e-01
9.47733819e-02 1.97714165e-01 1.08053434e+00 -7.82656908e-01
3.98500979e-01 -2.28156686e+00 -4.37870085e-01 2.23864634e-02
-4.45246883e-02 5.42986155e-01 -5.52904665e-01 5.36639333e-01
-1.78607076e-01 2.84489572e-01 -4.11433339e-01 4.05612290e-02
-4.03851330e-01 2.41134167e-01 -7.29391038e-01 8.14716280e-01
4.42417562e-01 9.43428516e-01 -1.15036833e+00 -2.74780303e-01
2.83614039e-01 2.96278715e-01 2.85287630e-02 2.25621000e-01
-4.28773910e-01 9.80784655e-01 -5.84663033e-01 7.24044502e-01
9.02264833e-01 -5.42004883e-01 2.23164096e-01 2.53834367e-01
-2.12772712e-01 3.05039555e-01 -7.62932122e-01 1.61515605e+00
-4.87507433e-01 7.70613313e-01 6.20219344e-03 -1.14505982e+00
1.14382398e+00 2.66651332e-01 5.44559300e-01 -7.75347412e-01
-3.78542900e-01 3.87214482e-01 -4.38873827e-01 -3.84740055e-01
6.27407372e-01 -1.83552802e-01 1.78427532e-01 7.25174308e-01
-1.25078559e-01 -5.92651106e-02 -3.13909203e-01 -1.53455054e-02
7.76341379e-01 5.21258891e-01 1.81978062e-01 -3.96041065e-01
1.42292082e-01 5.22838712e-01 4.44603354e-01 7.38153875e-01
-3.05794865e-01 5.39130390e-01 -4.03725356e-01 -9.20689404e-01
-1.11277580e+00 -8.06824565e-01 -2.90963292e-01 1.53245473e+00
-5.89986378e-03 2.48813286e-01 -3.00739288e-01 -5.34893632e-01
6.21030144e-02 1.52526781e-01 -6.86526060e-01 1.84126452e-01
-3.13355714e-01 -9.70117092e-01 7.07615495e-01 7.20701575e-01
9.70724165e-01 -1.02943337e+00 -6.82537019e-01 2.86278605e-01
-2.52078742e-01 -8.79378259e-01 -6.62037134e-02 2.01266915e-01
-1.34621191e+00 -8.61961186e-01 -7.24604189e-01 -5.74453115e-01
7.13754177e-01 9.92426336e-01 1.00846839e+00 7.83355907e-02
-1.14035256e-01 3.48140329e-01 -5.06875396e-01 -3.55571032e-01
-2.22108766e-01 5.14158487e-01 -5.45106791e-02 1.98854059e-02
4.86905009e-01 -7.90680885e-01 -7.72343993e-01 3.28234047e-01
-1.12884867e+00 -1.41305014e-01 6.46266937e-01 8.13694656e-01
6.23250604e-01 -6.35890588e-02 6.94263220e-01 -8.93840015e-01
-8.73233825e-02 -9.56786752e-01 -8.13702762e-01 2.66776592e-01
-6.46598458e-01 -8.12735707e-02 4.03699189e-01 -2.85192847e-01
-1.09965980e+00 4.26045269e-01 2.60729849e-01 -2.12978825e-01
-4.42035109e-01 1.20395529e+00 3.01734239e-01 2.62420308e-02
1.16814005e+00 4.57959324e-01 -2.59791035e-02 -5.51912189e-01
4.16419178e-01 1.02027786e+00 6.10009253e-01 -3.78302991e-01
1.02391684e+00 9.12162781e-01 -2.84845591e-01 -1.13800347e+00
-1.23521030e+00 -9.19080913e-01 -7.54660726e-01 -1.47170022e-01
6.87611580e-01 -1.65358031e+00 2.20579848e-01 5.75321019e-01
-6.98239625e-01 -9.21678007e-01 -1.76744998e-01 6.21061444e-01
-2.56198972e-01 1.49088517e-01 -2.85155952e-01 -6.23125970e-01
-5.82341850e-01 -3.83629084e-01 1.13342702e+00 1.97186828e-01
2.94515729e-01 -1.08700585e+00 3.69696766e-01 1.67562887e-01
8.23510647e-01 2.83825487e-01 4.05487806e-01 -5.05323708e-01
-6.68786705e-01 -1.12166412e-01 -5.36534727e-01 3.46249133e-01
5.90329230e-01 -1.28012165e-01 -1.23032546e+00 -6.99376881e-01
-2.25505218e-01 -7.24768400e-01 1.25838578e+00 3.69668484e-01
8.53817523e-01 -3.27114046e-01 -3.04512084e-01 8.86215568e-01
1.63282895e+00 -1.88866332e-01 7.42991805e-01 6.43861294e-01
6.15960896e-01 6.56296015e-01 6.18346095e-01 3.11127782e-01
4.27204370e-01 9.05170571e-03 6.12995446e-01 -7.19561279e-01
7.27637559e-02 -1.66166648e-01 3.16102326e-01 4.53384280e-01
-2.46230334e-01 1.54184535e-01 -1.32075691e+00 1.04939318e+00
-1.89232850e+00 -1.25338447e+00 -1.23283066e-01 2.08765960e+00
7.70313025e-01 -5.20301163e-01 -1.65956795e-01 -2.18673557e-01
7.50652611e-01 5.69272518e-01 -8.95962119e-01 4.46681023e-01
-1.78195611e-01 2.24806204e-01 1.09244645e+00 3.08506787e-01
-1.63651013e+00 1.36056733e+00 6.25533247e+00 2.10335985e-01
-1.80791318e+00 3.77887279e-01 6.21292174e-01 3.95941705e-01
-6.08163839e-03 1.73478723e-01 -7.08921313e-01 -5.89718251e-03
1.09118116e+00 2.54975329e-03 3.02751869e-01 1.02841949e+00
3.34664077e-01 -8.89143944e-02 -8.20296109e-01 5.73418796e-01
-3.85820448e-01 -1.46968710e+00 -2.38355979e-01 4.34280373e-02
1.18727887e+00 1.23149312e+00 1.34984583e-01 2.50807792e-01
7.20035136e-01 -9.38982308e-01 4.35388207e-01 2.04857841e-01
1.18289590e+00 -1.25389382e-01 4.48496461e-01 3.75597894e-01
-1.33855808e+00 1.02388270e-01 -6.66713536e-01 -2.30110198e-01
-3.11137557e-01 5.95796347e-01 -8.39395463e-01 6.86493278e-01
1.00365174e+00 1.40553761e+00 -5.16606033e-01 7.78962731e-01
-2.69615769e-01 9.94677007e-01 -4.80178475e-01 4.81197476e-01
6.33413851e-01 -2.48024136e-01 4.41751182e-02 1.00326014e+00
4.02359843e-01 2.73600310e-01 3.41309875e-01 7.66967416e-01
-1.81302130e-01 -2.37529818e-02 -1.05761087e+00 -1.96760610e-01
5.56509256e-01 1.30764973e+00 -6.36041462e-01 -3.70566994e-01
-4.52345222e-01 8.08897853e-01 2.31946185e-01 5.86702228e-01
-5.51899672e-01 -1.39186323e-01 6.08781457e-01 6.74107671e-02
3.57388765e-01 -5.37380576e-01 -1.09793000e-01 -1.50054169e+00
-1.58041313e-01 -6.68560803e-01 4.17769134e-01 -7.70390332e-01
-1.33317113e+00 6.20670915e-01 -1.11063316e-01 -1.72425842e+00
-1.98243320e-01 -3.66563171e-01 -7.43153811e-01 1.03662097e+00
-2.49577093e+00 -1.56769943e+00 -8.06415439e-01 6.98304534e-01
2.62898564e-01 -7.93428719e-02 9.77514923e-01 1.61229044e-01
-2.21088842e-01 -8.95364210e-02 6.72340333e-01 4.05416161e-01
8.57677221e-01 -9.01983798e-01 6.79846346e-01 1.07385349e+00
6.03976771e-02 2.97489196e-01 1.06351279e-01 -5.31017303e-01
-1.16546655e+00 -2.08820796e+00 7.84253359e-01 -8.54260102e-03
8.09708238e-01 7.02984259e-02 -1.12861466e+00 1.27046120e+00
-1.71965674e-01 5.97934902e-01 7.67432570e-01 5.58174103e-02
-6.30244017e-01 -5.01521289e-01 -1.02993250e+00 1.66677445e-01
6.35820389e-01 -8.36639404e-01 -4.76183027e-01 8.06430757e-01
4.65314299e-01 -1.33219615e-01 -7.95820177e-01 2.33086780e-01
1.81444615e-01 -6.29099667e-01 7.09904492e-01 -7.19525516e-01
5.43077171e-01 -4.11552340e-01 -3.71362060e-01 -1.27716064e+00
-3.41930538e-01 -3.53510171e-01 4.59403187e-01 9.89100039e-01
5.80416739e-01 -7.99347758e-01 6.54194117e-01 3.56809735e-01
-7.67563060e-02 2.47761205e-01 -6.15928829e-01 -1.17111993e+00
3.91666412e-01 -4.82488759e-02 6.58057153e-01 1.65343690e+00
-3.78647625e-01 3.27681541e-01 -3.86539966e-01 7.62833238e-01
5.94405353e-01 5.81462562e-01 7.32604504e-01 -1.35072422e+00
-4.37299050e-02 6.35982603e-02 -1.92825690e-01 -1.09390748e+00
8.42634216e-02 -9.29114997e-01 2.66027808e-01 -1.55712557e+00
-3.63280922e-02 -9.73781466e-01 -2.75532454e-01 1.30252266e+00
-1.21316604e-01 3.77623230e-01 -2.67955810e-01 9.23755169e-01
-2.49860957e-01 6.14183486e-01 7.18636572e-01 -5.46683073e-01
-3.88553590e-01 -1.31608590e-01 -2.53327638e-01 4.14816529e-01
1.13986504e+00 -7.88686395e-01 -3.05393517e-01 -1.02941036e+00
1.09256908e-01 -2.04836503e-02 5.21936655e-01 -1.05639410e+00
1.62396114e-02 -5.30219853e-01 2.18848825e-01 -6.80217147e-01
-7.82288089e-02 -7.51226008e-01 2.05588758e-01 4.73938823e-01
-1.23520881e-01 -1.92485914e-01 3.51272523e-01 4.53763336e-01
-3.51270229e-01 2.80395299e-02 9.32347953e-01 -4.38827932e-01
-1.14961255e+00 6.60183489e-01 -4.08003926e-01 -1.99709773e-01
8.45004439e-01 6.73163831e-02 -4.81340408e-01 -3.86786550e-01
-5.94304621e-01 3.34143102e-01 5.09608567e-01 1.73660696e-01
3.36510092e-01 -1.04540563e+00 -1.16985095e+00 1.90496147e-01
3.50885242e-01 3.38706762e-01 2.80917376e-01 5.28768837e-01
-7.47310936e-01 1.25403762e-01 -4.19188023e-01 -7.14047730e-01
-7.82058775e-01 3.24450165e-01 5.51707804e-01 -2.43222296e-01
-6.99583292e-01 4.88205433e-01 -8.55491608e-02 -1.08528602e+00
-3.83679897e-01 -1.10178135e-01 8.05983990e-02 7.61207612e-03
6.34881198e-01 -7.07802400e-02 2.13835374e-01 -6.28211498e-01
-3.28262746e-01 3.68857712e-01 1.83478639e-01 -2.98637033e-01
2.03942919e+00 -2.40090430e-01 -1.44049183e-01 3.80499303e-01
1.05310106e+00 -5.42784572e-01 -1.51827037e+00 -7.15052068e-01
-1.21426992e-01 -2.75859386e-01 3.90140593e-01 -6.64054930e-01
-1.36422861e+00 1.20084095e+00 8.43163013e-01 8.13346505e-02
1.23866975e+00 -8.61464813e-02 5.45592546e-01 9.10396218e-01
3.46868217e-01 -9.33848083e-01 -1.74021468e-01 6.37736440e-01
5.34640968e-01 -1.73802018e+00 5.37818782e-02 7.70299286e-02
-3.33427638e-01 1.05552077e+00 3.92776608e-01 2.60921055e-03
9.11062837e-01 -1.48854032e-01 4.63341355e-01 -2.47533441e-01
-3.69701147e-01 -5.49270809e-01 -3.52765694e-02 9.00067329e-01
3.11635256e-01 1.83437780e-01 2.88492978e-01 -7.18834475e-02
2.88269341e-01 2.46704862e-01 3.51278394e-01 1.27423823e+00
-6.41021967e-01 -8.05418730e-01 -3.83955628e-01 2.80009329e-01
-1.88569695e-01 -4.68818188e-01 -1.06251240e-01 4.68775213e-01
-2.91796952e-01 8.43308747e-01 1.55896500e-01 -2.75488973e-01
-1.98243991e-01 -1.86303079e-01 -2.27464101e-04 -8.37560952e-01
-3.34394366e-01 -1.83833569e-01 -1.36942327e-01 -1.97256550e-01
-1.01195049e+00 -5.98418713e-01 -1.10309565e+00 -4.15956467e-01
-1.33656293e-01 -5.77395689e-03 8.61559212e-01 1.01793706e+00
6.59399390e-01 -1.55318547e-02 1.02936220e+00 -1.35252154e+00
-7.16403484e-01 -1.12404191e+00 -7.58713007e-01 1.76212251e-01
5.27406812e-01 -1.74919799e-01 -1.47004321e-01 4.03721303e-01] | [9.623419761657715, -1.4597957134246826] |
9069ead2-2479-41db-97a4-29aee8ab67f7 | towards-faster-training-of-global-covariance | 1712.01034 | null | http://arxiv.org/abs/1712.01034v2 | http://arxiv.org/pdf/1712.01034v2.pdf | Towards Faster Training of Global Covariance Pooling Networks by Iterative Matrix Square Root Normalization | Global covariance pooling in convolutional neural networks has achieved
impressive improvement over the classical first-order pooling. Recent works
have shown matrix square root normalization plays a central role in achieving
state-of-the-art performance. However, existing methods depend heavily on
eigendecomposition (EIG) or singular value decomposition (SVD), suffering from
inefficient training due to limited support of EIG and SVD on GPU. Towards
addressing this problem, we propose an iterative matrix square root
normalization method for fast end-to-end training of global covariance pooling
networks. At the core of our method is a meta-layer designed with loop-embedded
directed graph structure. The meta-layer consists of three consecutive
nonlinear structured layers, which perform pre-normalization, coupled matrix
iteration and post-compensation, respectively. Our method is much faster than
EIG or SVD based ones, since it involves only matrix multiplications, suitable
for parallel implementation on GPU. Moreover, the proposed network with ResNet
architecture can converge in much less epochs, further accelerating network
training. On large-scale ImageNet, we achieve competitive performance superior
to existing counterparts. By finetuning our models pre-trained on ImageNet, we
establish state-of-the-art results on three challenging fine-grained
benchmarks. The source code and network models will be available at
http://www.peihuali.org/iSQRT-COV | ['Zilin Gao', 'Peihua Li', 'Jiangtao Xie', 'Qilong Wang'] | 2017-12-04 | towards-faster-training-of-global-covariance-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Li_Towards_Faster_Training_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Li_Towards_Faster_Training_CVPR_2018_paper.pdf | cvpr-2018-6 | ['fine-grained-image-recognition'] | ['computer-vision'] | [-1.33469462e-01 -2.68118829e-01 2.35648647e-01 -4.20744270e-01
-2.43111178e-01 -3.38513017e-01 3.46236467e-01 -1.08349144e-01
-9.75183308e-01 2.88206577e-01 -8.77232254e-02 -5.24596870e-01
-9.21585262e-02 -7.30617106e-01 -8.01554680e-01 -8.20259392e-01
-2.82610748e-02 -1.31682575e-01 4.05970871e-01 -2.60382771e-01
-1.99967511e-02 5.11912882e-01 -1.12964213e+00 1.67154014e-01
4.19423252e-01 1.16909480e+00 2.42660627e-01 6.72125578e-01
5.69133461e-02 7.85917342e-01 -2.10190371e-01 -6.04798138e-01
2.05548495e-01 -1.23473451e-01 -3.87892693e-01 -2.11757720e-01
5.55482745e-01 -1.88519120e-01 -6.24857426e-01 1.44238579e+00
6.27847075e-01 2.06374064e-01 4.79664832e-01 -1.05021513e+00
-6.68085396e-01 6.88348472e-01 -6.25032544e-01 2.51294762e-01
-6.06985927e-01 -1.82489119e-02 1.07727444e+00 -1.10571861e+00
3.92712086e-01 9.81283665e-01 8.76423836e-01 4.89476621e-01
-1.27085090e+00 -9.08004701e-01 2.14175776e-01 1.32563516e-01
-1.48697793e+00 -2.55168498e-01 6.75745726e-01 -3.45463276e-01
1.37526989e+00 1.36618549e-02 5.93283951e-01 8.43478084e-01
2.38928989e-01 4.00928319e-01 6.59389973e-01 6.36874363e-02
6.09197207e-02 -2.07946807e-01 2.82104701e-01 1.00998664e+00
3.83954734e-01 -2.23130196e-01 -2.34986708e-01 7.61017129e-02
8.70514631e-01 1.12708338e-01 -7.86603689e-02 -3.72619778e-01
-1.28270578e+00 6.02275372e-01 1.05375862e+00 2.16106638e-01
-4.03981030e-01 4.59188074e-01 6.46478117e-01 2.88603991e-01
3.56718749e-01 3.49205956e-02 -5.84409118e-01 4.88774516e-02
-8.76753151e-01 3.82093005e-02 8.03728580e-01 8.38426650e-01
8.63014638e-01 -5.12135867e-03 -2.25558251e-01 9.33899343e-01
3.87145817e-01 4.64732319e-01 5.81778824e-01 -4.12952662e-01
8.02041650e-01 8.44676971e-01 -5.86241484e-01 -1.19507837e+00
-8.56025636e-01 -8.38114142e-01 -1.70309651e+00 1.29841805e-01
3.18462372e-01 -5.08029103e-01 -9.49074447e-01 1.74085927e+00
5.57135940e-02 3.05209190e-01 8.97897929e-02 9.43283498e-01
8.90300035e-01 3.69215935e-01 -6.89077228e-02 9.11762118e-02
1.58252656e+00 -1.25125062e+00 -4.83548850e-01 -4.54439104e-01
8.67302835e-01 -6.51895285e-01 9.20021832e-01 9.80524346e-02
-7.81401932e-01 -4.23775434e-01 -1.34185231e+00 -3.42440069e-01
-4.05980736e-01 7.11103499e-01 7.16728151e-01 7.19558060e-01
-1.17243433e+00 8.19524229e-01 -1.13857269e+00 3.66936885e-02
7.24886656e-01 5.51971853e-01 -6.73605680e-01 2.04400167e-01
-9.49325979e-01 4.88791972e-01 5.59203207e-01 6.89162970e-01
-4.56828386e-01 -7.91531265e-01 -6.80024207e-01 2.49763101e-01
1.62103623e-01 -6.64002657e-01 9.26534474e-01 -4.98171896e-01
-1.54503822e+00 7.12352633e-01 -7.84946457e-02 -4.97200221e-01
5.58965266e-01 -1.88336939e-01 -7.81253576e-02 -2.38803431e-01
-3.73365462e-01 3.05656731e-01 9.45321441e-01 -4.30554062e-01
-1.86873659e-01 -3.94221216e-01 -1.45050734e-01 -7.95823559e-02
-9.25242960e-01 7.24040419e-02 -7.85156488e-01 -5.99481285e-01
3.91609460e-01 -8.62069488e-01 -5.24579167e-01 -2.86449105e-01
-4.35394078e-01 -1.49208069e-01 5.23119211e-01 -4.42347348e-01
1.36912262e+00 -2.22986150e+00 7.42892250e-02 4.14705396e-01
6.14787519e-01 6.08260512e-01 -3.77851456e-01 1.10512674e-01
-2.63586640e-01 9.89480317e-03 -9.53554809e-02 -6.20693207e-01
9.48429257e-02 -1.37216359e-01 -1.02990000e-02 5.55385649e-01
4.30503339e-01 1.10433137e+00 -5.90500832e-01 -1.49559453e-01
-2.53894422e-02 9.39364612e-01 -8.34214151e-01 -3.94966491e-02
2.04182521e-01 -3.04989815e-02 -2.39488795e-01 7.36342818e-02
1.06249964e+00 -5.49239159e-01 2.14977920e-01 -7.11456180e-01
-1.05056010e-01 1.66961670e-01 -1.37846136e+00 1.67901671e+00
-3.66591483e-01 6.03039861e-01 4.04030494e-02 -9.89449620e-01
1.04278314e+00 -2.63431249e-03 9.71169248e-02 -5.25235832e-01
6.80116177e-01 3.47499460e-01 2.20072865e-01 -1.15646116e-01
4.02511477e-01 2.26522475e-01 3.81547779e-01 2.97928602e-01
3.17371041e-01 2.32197359e-01 5.99455833e-01 1.86023161e-01
1.18743527e+00 -2.20559046e-01 -7.37184053e-03 -4.32915866e-01
8.29917729e-01 -4.36979443e-01 4.50856179e-01 5.08215904e-01
-2.71284580e-03 6.83737755e-01 9.33714926e-01 -4.02572960e-01
-9.21018541e-01 -8.19233835e-01 -6.59482256e-02 1.10874820e+00
-2.50590235e-01 -6.81720436e-01 -1.03006816e+00 -5.04710615e-01
-2.73131907e-01 -1.15101300e-01 -6.56137586e-01 -1.15522578e-01
-6.71381116e-01 -1.29558623e+00 7.14372396e-01 7.58245170e-01
9.06300247e-01 -9.26102996e-01 -3.47115040e-01 3.66415799e-01
2.53411531e-01 -1.33117986e+00 -5.86675465e-01 3.84234428e-01
-1.09156322e+00 -1.04265249e+00 -8.33248496e-01 -1.01637030e+00
1.00500071e+00 3.16741407e-01 8.66386414e-01 1.72724664e-01
-2.88044989e-01 -2.95576245e-01 2.77186045e-03 -1.70251772e-01
1.04966111e-01 5.48035562e-01 -5.25399074e-02 2.13843420e-01
3.95930767e-01 -7.60679066e-01 -8.09141338e-01 3.73582363e-01
-8.51106048e-01 1.65601984e-01 8.62480462e-01 9.01884258e-01
6.61142290e-01 -3.88003774e-02 7.10914508e-02 -9.80538487e-01
8.67990792e-01 -5.58150001e-03 -9.32447910e-01 2.26381376e-01
-3.76439989e-01 3.12817872e-01 8.04888546e-01 -3.30403179e-01
-8.84944975e-01 1.06347956e-01 -4.05104607e-01 -3.85252327e-01
2.98325598e-01 7.00767815e-01 1.72656670e-01 -3.37288737e-01
8.21874321e-01 8.98018107e-02 3.52992341e-02 -5.10013580e-01
3.17901343e-01 2.46095389e-01 4.70549017e-01 -3.04162979e-01
9.12571013e-01 3.91417265e-01 3.36588144e-01 -7.51806438e-01
-7.86874294e-01 -4.20252711e-01 -7.94537127e-01 7.06056729e-02
8.56777191e-01 -1.10464966e+00 -1.07920730e+00 8.34158242e-01
-1.07004237e+00 -5.11804342e-01 2.11402252e-01 7.48412311e-01
-1.23023957e-01 5.79032563e-02 -7.32236087e-01 -3.32734317e-01
-6.91971123e-01 -1.13052690e+00 6.04787886e-01 2.24831328e-01
5.09085953e-01 -1.03146386e+00 -5.38568757e-02 4.35731597e-02
7.14802563e-01 2.64547504e-02 5.99129677e-01 -3.98146540e-01
-5.16874015e-01 -4.63481247e-01 -9.18420613e-01 8.54562998e-01
-2.68411815e-01 -1.91171505e-02 -9.01535809e-01 -5.24418652e-01
-9.77811515e-02 -2.39985168e-01 1.29370534e+00 2.81644285e-01
1.06852901e+00 -1.43205643e-01 -3.14623825e-02 1.16093135e+00
1.58239877e+00 -4.62260664e-01 5.15833676e-01 3.35343063e-01
1.04708910e+00 2.10911319e-01 -1.79368425e-02 2.61613846e-01
4.12152112e-01 1.97978914e-01 3.18028867e-01 -1.70050859e-01
-2.38454029e-01 3.01674027e-02 2.62707055e-01 1.24201429e+00
-4.71673250e-01 1.47309780e-01 -9.71715629e-01 2.09103435e-01
-2.09924960e+00 -6.48540199e-01 -4.03480977e-01 1.87260044e+00
6.06753707e-01 2.57861525e-01 -2.36426488e-01 5.05952239e-02
4.85426933e-01 4.25035924e-01 -5.03728032e-01 -2.09356576e-01
-2.50808328e-01 2.35984042e-01 1.00860226e+00 3.00669223e-01
-1.38757217e+00 1.13644516e+00 4.84132671e+00 7.21650958e-01
-1.18192267e+00 1.90924555e-01 5.83265603e-01 -2.89505105e-02
3.10929716e-01 -3.60776275e-01 -1.02572298e+00 -1.51543930e-01
6.97143555e-01 5.94544895e-02 5.98708987e-01 8.03175688e-01
-3.71209043e-03 3.46233100e-01 -8.24379742e-01 1.03804517e+00
-6.07809126e-02 -1.46178854e+00 -1.36553228e-01 -8.93252268e-02
9.96502876e-01 7.23011255e-01 3.87183949e-03 2.71571666e-01
1.39276788e-01 -7.89540589e-01 4.80009019e-01 2.51000553e-01
6.02726817e-01 -7.70105541e-01 9.80215907e-01 1.67427853e-01
-1.50710118e+00 3.57597694e-02 -9.82059062e-01 -1.47725523e-01
-1.78641990e-01 8.83945465e-01 -3.25644374e-01 3.90692711e-01
8.33218575e-01 8.31637859e-01 -7.76857972e-01 9.49359655e-01
-4.53511178e-01 3.83792728e-01 -3.78467888e-01 -4.03844535e-01
4.51371551e-01 -3.51341099e-01 1.53175846e-01 1.50803661e+00
2.95559853e-01 -1.31648943e-01 -3.41645330e-01 7.00082481e-01
-5.25359154e-01 3.76973122e-01 5.41737163e-03 -2.07271218e-01
-4.85175550e-02 1.72144949e+00 -9.03170943e-01 -1.70627087e-01
-6.43487155e-01 7.33559430e-01 9.43825841e-01 4.29676175e-01
-7.94104218e-01 -9.25777137e-01 9.40749705e-01 -3.83175671e-01
6.27855897e-01 -5.67278743e-01 -3.50327253e-01 -1.44336891e+00
2.61157721e-01 -5.93307376e-01 2.43360903e-02 -2.24979848e-01
-1.16053534e+00 9.24930036e-01 -5.18611789e-01 -8.79368722e-01
2.25609913e-01 -1.20789027e+00 -5.67365944e-01 9.87888455e-01
-1.66256618e+00 -9.00818408e-01 -5.88054419e-01 7.42550850e-01
-1.09056868e-01 -3.18480194e-01 6.72162712e-01 5.73651373e-01
-9.06641483e-01 9.26153541e-01 2.65686303e-01 7.19064951e-01
5.69282889e-01 -1.08946955e+00 8.46157253e-01 1.09450793e+00
8.65190923e-02 8.73831213e-01 2.02555135e-01 -4.11763042e-01
-1.59999537e+00 -1.32528710e+00 8.53712678e-01 6.59081563e-02
1.20553231e+00 -7.25069106e-01 -9.86662209e-01 5.76960087e-01
-1.01186894e-01 4.77870941e-01 3.53615284e-01 1.93564564e-01
-6.78270221e-01 -3.63373458e-01 -5.63957870e-01 7.12274492e-01
1.27044272e+00 -5.06463766e-01 -1.46608844e-01 4.10262764e-01
6.02176547e-01 -6.62092030e-01 -8.31068993e-01 3.38680506e-01
5.41421950e-01 -9.83243465e-01 9.34255123e-01 -6.15648746e-01
4.88192648e-01 -3.42925131e-01 3.11833452e-02 -1.20383203e+00
-5.99213600e-01 -6.20066285e-01 -2.27582343e-02 8.78142416e-01
6.81031585e-01 -9.74147677e-01 1.05176580e+00 4.25800115e-01
-9.46476758e-02 -9.03290689e-01 -6.90052927e-01 -9.34668422e-01
1.14805087e-01 -7.01712072e-01 3.88056785e-01 6.01039648e-01
-2.65424579e-01 3.69398236e-01 -1.34211183e-01 2.98794121e-01
7.59723902e-01 -2.74805874e-01 8.37722898e-01 -1.07401979e+00
-2.73788124e-01 -7.86695659e-01 -6.80786371e-01 -9.98083293e-01
7.28622526e-02 -1.17492962e+00 -1.66675020e-02 -1.48839855e+00
1.31790638e-01 -1.57104090e-01 -5.67386687e-01 7.07947433e-01
-2.37529024e-01 7.97344148e-01 3.44601870e-01 -8.45234748e-03
-4.51883107e-01 4.72779423e-01 1.34329748e+00 -1.13554433e-01
-2.91083246e-01 -1.42798632e-01 -5.51000118e-01 7.39905357e-01
1.09242797e+00 -4.96519536e-01 -3.32818389e-01 -7.96658874e-01
6.50843024e-01 -5.66993594e-01 3.89420867e-01 -1.28762770e+00
7.09744215e-01 4.40908641e-01 2.74241030e-01 -3.03066105e-01
8.44509155e-02 -6.06494188e-01 -1.35048971e-01 5.35051286e-01
-1.41859904e-01 2.48983875e-01 4.85454768e-01 3.30484062e-01
-3.24059010e-01 -2.28192285e-01 7.56064475e-01 2.15195730e-01
-4.53260273e-01 8.25622022e-01 -5.91135360e-02 1.10929094e-01
5.14213681e-01 2.15857580e-01 -4.65133995e-01 2.55007505e-01
-6.58231378e-01 1.35904670e-01 -9.03734639e-02 3.80293131e-01
5.78066468e-01 -1.33996320e+00 -7.36825585e-01 4.55182821e-01
-7.64288381e-02 1.14405684e-01 4.59240019e-01 1.27169216e+00
-8.83861303e-01 4.20509368e-01 -1.55138969e-01 -4.60156739e-01
-1.33150423e+00 2.57740885e-01 3.51707608e-01 -4.93591964e-01
-5.90050161e-01 1.25829816e+00 2.18325377e-01 -4.37007159e-01
3.83889973e-01 -5.59218645e-01 -1.65706083e-01 7.61931539e-02
7.23619640e-01 2.27401510e-01 4.73081321e-01 -4.46319580e-01
-4.41024601e-01 7.84000814e-01 -3.11896682e-01 1.13817677e-01
1.56229734e+00 3.04999769e-01 -5.27567923e-01 2.46837325e-02
1.75429320e+00 -4.08620358e-01 -1.45931947e+00 -5.85328579e-01
-1.63886294e-01 -6.19977899e-02 3.97453517e-01 -2.45362148e-01
-1.57869232e+00 1.16173589e+00 5.59953451e-01 -1.27769858e-02
1.25144303e+00 -3.45590800e-01 5.57347298e-01 1.08761609e+00
5.07295644e-03 -9.64670002e-01 -6.97641894e-02 1.08499396e+00
9.20438826e-01 -1.19203866e+00 9.85092372e-02 -3.13371271e-01
-3.01425546e-01 1.29175127e+00 5.33303916e-01 -5.37555814e-01
1.18564415e+00 4.34206039e-01 1.19686805e-01 -1.21253014e-01
-6.61558330e-01 -1.17980279e-01 5.74299097e-01 1.10563852e-01
6.08696759e-01 4.47157659e-02 -2.89495587e-01 6.97175086e-01
-4.37343270e-01 -5.62313013e-03 2.59626985e-01 5.98617196e-01
-1.81086719e-01 -9.15736616e-01 7.08011687e-02 5.54410160e-01
-4.95043844e-01 -4.99356747e-01 -5.21509629e-03 5.16784370e-01
-2.01479360e-01 5.73835850e-01 1.38698265e-01 -5.33854306e-01
5.63774765e-01 -1.53262436e-01 5.01120090e-01 -3.16856265e-01
-8.21693420e-01 -5.08393310e-02 -2.58780390e-01 -7.74737358e-01
-1.98497176e-01 -3.81288528e-01 -1.24941099e+00 -4.14165020e-01
-3.24122459e-01 -2.19801918e-01 9.49169040e-01 6.28588259e-01
3.68394941e-01 9.34723675e-01 3.07277799e-01 -1.04498863e+00
-5.79236388e-01 -1.01808846e+00 -4.76276726e-01 2.26492897e-01
1.66777939e-01 -2.94129521e-01 -3.57008755e-01 -1.73446789e-01] | [8.997037887573242, 2.302155017852783] |
f499c63d-2d62-4c09-a313-23e5b1dd5261 | tuning-traditional-language-processing | 2305.03737 | null | https://arxiv.org/abs/2305.03737v1 | https://arxiv.org/pdf/2305.03737v1.pdf | Tuning Traditional Language Processing Approaches for Pashto Text Classification | Today text classification becomes critical task for concerned individuals for numerous purposes. Hence, several researches have been conducted to develop automatic text classification for national and international languages. However, the need for an automatic text categorization system for local languages is felt. The main aim of this study is to establish a Pashto automatic text classification system. In order to pursue this work, we built a Pashto corpus which is a collection of Pashto documents due to the unavailability of public datasets of Pashto text documents. Besides, this study compares several models containing both statistical and neural network machine learning techniques including Multilayer Perceptron (MLP), Support Vector Machine (SVM), K Nearest Neighbor (KNN), decision tree, gaussian na\"ive Bayes, multinomial na\"ive Bayes, random forest, and logistic regression to discover the most effective approach. Moreover, this investigation evaluates two different feature extraction methods including unigram, and Time Frequency Inverse Document Frequency (IFIDF). Subsequently, this research obtained average testing accuracy rate 94% using MLP classification algorithm and TFIDF feature extraction method in this context. | ['Nematullah Hassanzada', 'Mohammad Zarif Joya', 'Mursal Dawodi', 'Jawid Ahmad Baktash'] | 2023-05-04 | null | null | null | null | ['text-categorization'] | ['natural-language-processing'] | [ 3.58839668e-02 -2.66255200e-01 -5.17041028e-01 -4.32988614e-01
-9.54021467e-04 -2.93504924e-01 7.66929626e-01 8.82397294e-01
-7.48381793e-01 1.01153982e+00 2.80659467e-01 -6.55292511e-01
-4.85009581e-01 -9.59708929e-01 3.30189645e-01 -5.51979363e-01
2.47927174e-01 5.77531874e-01 -1.34474128e-01 -2.01962620e-01
1.10350192e+00 6.05774701e-01 -1.90443385e+00 4.47693467e-01
8.75146985e-01 9.72547233e-01 8.83315876e-02 7.96696961e-01
-5.72394729e-01 8.74991655e-01 -6.69561505e-01 -2.08864495e-01
-3.66885692e-01 -7.79835284e-02 -1.05293882e+00 -2.45811269e-01
-1.75450236e-01 -1.99983150e-01 -3.61971110e-02 6.76941037e-01
3.65227520e-01 3.53748262e-01 1.38740742e+00 -1.31822181e+00
-6.72283351e-01 7.62114525e-01 -3.77069116e-01 4.17029053e-01
5.05888700e-01 -6.69182360e-01 8.30306768e-01 -7.31590927e-01
2.58907020e-01 1.15612030e+00 6.36424065e-01 -6.97659776e-02
-8.01662385e-01 -7.58744955e-01 -2.43122831e-01 3.09949517e-01
-1.45986426e+00 5.37709258e-02 6.11397922e-01 -6.73171759e-01
1.34497547e+00 3.73975009e-01 2.96802640e-01 7.42948949e-01
7.54188120e-01 4.51993376e-01 1.62139356e+00 -1.12568963e+00
-2.61370442e-03 1.04946387e+00 8.12870085e-01 3.49835068e-01
3.57399523e-01 -3.29828173e-01 -4.12174553e-01 -4.17283565e-01
5.98890744e-02 -1.87435430e-02 1.87679946e-01 5.62685251e-01
-6.78683877e-01 9.39484060e-01 -1.45880386e-01 1.21062148e+00
-2.17449293e-01 -6.45667970e-01 5.91356158e-01 4.36467141e-01
5.14050782e-01 9.95485634e-02 -7.81985164e-01 -3.04440618e-01
-7.36186862e-01 -6.32661656e-02 1.12648249e+00 5.81311166e-01
3.55939269e-01 -9.40837115e-02 1.41038328e-01 1.14544928e+00
6.27320945e-01 4.14408505e-01 1.14247179e+00 -1.28957275e-02
5.14058650e-01 8.57383072e-01 -8.49561617e-02 -1.58047009e+00
-5.56221902e-01 -1.60246566e-01 -9.54800904e-01 -2.10005060e-01
5.97495809e-02 -9.49284881e-02 -4.10018533e-01 7.31963158e-01
1.52194381e-01 -8.75295162e-01 2.18850777e-01 2.43842617e-01
9.24967229e-01 9.84663844e-01 3.78543913e-01 -3.40306193e-01
1.45711052e+00 -6.06153250e-01 -9.23731029e-01 3.69196445e-01
8.80308449e-01 -1.16086340e+00 7.22892761e-01 6.70190215e-01
-3.13037485e-01 -4.86227065e-01 -7.57483840e-01 2.04841316e-01
-1.13406408e+00 5.17123997e-01 7.23303795e-01 1.20164788e+00
-5.60329497e-01 3.64753336e-01 -4.17922497e-01 -7.53782690e-01
4.28767055e-02 3.80107522e-01 -3.61792475e-01 2.41455480e-01
-1.18190026e+00 9.12315786e-01 8.02476704e-01 -2.82438457e-01
2.81024933e-01 1.22618265e-01 -4.76502478e-01 -1.58707406e-02
-3.18184495e-01 -1.30745783e-01 7.27717102e-01 -8.29840302e-01
-1.23085928e+00 6.10430956e-01 -2.63004899e-01 -2.12837026e-01
2.29337625e-02 2.26014495e-01 -9.23142433e-01 1.19168498e-01
2.52473176e-01 8.85809883e-02 5.24686277e-01 -7.81336248e-01
-1.08390987e+00 -6.56579554e-01 -5.59344411e-01 1.79724380e-01
-9.66238618e-01 3.62525761e-01 3.96932632e-01 -7.30343223e-01
2.10826442e-01 -4.79288369e-01 3.83899659e-01 -6.77258968e-01
-3.36542726e-01 -9.00788665e-01 7.39798725e-01 -1.00015521e+00
1.66905117e+00 -1.93393838e+00 -5.90635896e-01 5.04082918e-01
-1.68837503e-01 -9.56706703e-02 5.52931786e-01 9.22942340e-01
1.37952656e-01 2.76539147e-01 1.59032971e-01 2.82743812e-01
-8.10880028e-03 1.85754970e-01 -2.34585896e-01 2.00031325e-01
-4.71124887e-01 1.00941338e-01 -2.49213353e-01 -1.11365569e+00
2.95005828e-01 3.58139634e-01 -1.26483247e-01 -2.75234014e-01
3.00769687e-01 -1.53701782e-01 -7.99429655e-01 8.14158916e-01
3.45574826e-01 5.77519648e-02 5.50949574e-02 -1.30416408e-01
-5.53908706e-01 1.76899195e-01 -1.17218769e+00 5.72951615e-01
-5.29109299e-01 7.68646002e-01 -4.80620205e-01 -1.29234850e+00
1.24007595e+00 5.91163099e-01 4.68012661e-01 -2.03728899e-01
7.44780838e-01 4.52163815e-01 -2.33067095e-01 -7.56678581e-01
7.03922629e-01 3.14169638e-02 1.95145696e-01 4.63680387e-01
7.97876194e-02 3.44774544e-01 2.75919706e-01 -1.34388581e-01
5.23961782e-01 -1.59686878e-01 7.10890532e-01 -4.22207624e-01
8.71219516e-01 3.91503751e-01 -2.69401729e-01 7.38189280e-01
-8.59377533e-02 -8.45693797e-02 2.16366455e-01 -4.60332155e-01
-6.85496867e-01 -3.75042617e-01 -7.21881211e-01 1.25705040e+00
-3.72012883e-01 -3.34118217e-01 -5.29603899e-01 -4.77882385e-01
8.93565416e-02 1.00583160e+00 -3.02405119e-01 2.31846660e-01
-1.30066931e-01 -1.10153604e+00 6.86948180e-01 2.23025352e-01
8.47228885e-01 -9.12587404e-01 -3.61064792e-01 1.73257068e-01
-1.04126111e-01 -5.80080688e-01 -3.72373089e-02 3.88371974e-01
-9.33034837e-01 -8.42302799e-01 -4.32334900e-01 -1.03305554e+00
5.86092174e-01 1.24993622e-01 5.78624070e-01 -1.11683726e-01
-3.62814575e-01 2.46025622e-01 -7.82144725e-01 -5.30549288e-01
-5.22987485e-01 5.89417994e-01 7.50653148e-02 -2.25863948e-01
1.04688692e+00 -3.52180541e-01 -1.46870643e-01 3.53657871e-01
-7.89665759e-01 -3.66631210e-01 7.29012311e-01 6.90812290e-01
1.10456347e-01 9.12375867e-01 7.51628280e-01 -7.45350361e-01
9.92502689e-01 -5.49397647e-01 -2.71706969e-01 2.61511624e-01
-1.25345361e+00 -1.34164765e-01 7.05119312e-01 -3.21242273e-01
-1.12171161e+00 -4.12884355e-01 -3.79989088e-01 7.19716847e-01
-5.59702337e-01 8.09237659e-01 1.59976900e-01 -2.80713812e-02
9.43749785e-01 6.29562497e-01 -1.38148621e-01 -4.71481591e-01
-2.62499839e-01 1.98419595e+00 -1.63547710e-01 -4.81714547e-01
3.52087528e-01 1.13100417e-01 -2.15031445e-01 -1.16325855e+00
-5.80551028e-01 -8.15700889e-01 -7.75274575e-01 -3.62141669e-01
8.55813146e-01 -4.20245260e-01 -9.99592006e-01 5.59766829e-01
-9.25225317e-01 3.56759250e-01 5.60231626e-01 8.08232248e-01
-2.06246618e-02 3.85531217e-01 -4.94698673e-01 -1.14014018e+00
-7.53692150e-01 -6.13683879e-01 5.32553315e-01 1.49014562e-01
-6.40393078e-01 -1.14073420e+00 -2.32234314e-01 6.34140134e-01
3.98954213e-01 -1.63647458e-01 1.19200349e+00 -1.10122633e+00
4.49379832e-01 -6.32805526e-01 -3.22930247e-01 4.46792811e-01
3.75125080e-01 2.05315784e-01 -6.69944882e-01 -8.46014023e-02
3.90849188e-02 -1.41703606e-01 4.91733551e-01 3.15182209e-01
1.02727532e+00 -6.43676817e-01 -4.41388547e-01 -1.80613875e-01
1.37717748e+00 8.37321639e-01 2.10094780e-01 8.97581577e-01
2.25028932e-01 6.97411597e-01 6.70201063e-01 5.97497702e-01
5.41805387e-01 1.76530048e-01 -2.46403560e-01 6.18376732e-01
2.84305662e-01 -7.55149350e-02 2.81162504e-02 9.82161343e-01
-1.82386294e-01 -3.63979459e-01 -1.15526080e+00 1.60475478e-01
-1.32323980e+00 -8.07255507e-01 -4.31799114e-01 1.88071585e+00
7.26402700e-01 2.86498755e-01 1.22973450e-01 1.14428759e+00
6.52774096e-01 -2.71075040e-01 1.87073082e-01 -7.83444643e-01
-4.70706867e-03 5.03495708e-02 5.49309373e-01 1.86265886e-01
-1.18318093e+00 3.64921957e-01 5.46862364e+00 1.08545053e+00
-1.23584616e+00 -1.90772876e-01 6.55022740e-01 5.45393944e-01
1.59207776e-01 -2.69156665e-01 -1.12998104e+00 5.98968625e-01
1.04238915e+00 -2.19199017e-01 1.04491010e-01 9.16054845e-01
7.84105435e-02 -5.94958842e-01 -3.00423145e-01 8.97369802e-01
2.55239606e-01 -8.37980032e-01 2.42189437e-01 1.61405671e-02
3.11203927e-01 -2.67094493e-01 -1.05743289e-01 4.09077197e-01
-1.47434831e-01 -9.63151634e-01 3.07876587e-01 3.29766929e-01
3.34974229e-01 -9.10076857e-01 1.06274939e+00 6.65603161e-01
-9.12881792e-01 -2.01184928e-01 -2.61643916e-01 -2.86444545e-01
-4.44086820e-01 6.14048183e-01 -1.02288091e+00 5.39459825e-01
8.19044530e-01 3.10209453e-01 -6.73494101e-01 7.26294398e-01
3.50102901e-01 7.71424413e-01 -3.06670040e-01 -1.05086422e+00
1.20233402e-01 -9.35425311e-02 2.22429335e-01 1.18765509e+00
3.32706928e-01 2.14812960e-02 6.90137618e-04 -1.11889271e-02
5.13373971e-01 1.08839130e+00 -3.67052972e-01 -1.48512721e-01
4.82017905e-01 1.02091241e+00 -1.27340496e+00 -2.70859301e-01
-4.68790948e-01 5.80432355e-01 -1.89095587e-01 3.40411514e-02
-3.78872901e-01 -1.15913045e+00 -3.09966922e-01 1.27129763e-01
-3.07799101e-01 -1.57234356e-01 -6.74450099e-01 -8.52022648e-01
-3.21016073e-01 -7.13224828e-01 7.06318974e-01 -5.43932617e-01
-1.30021930e+00 6.17004633e-01 5.09355128e-01 -8.73771191e-01
-1.13565676e-01 -9.92591858e-01 -1.24614626e-01 1.04613173e+00
-9.51396763e-01 -1.01664257e+00 -9.16944295e-02 4.02040839e-01
5.78592062e-01 -6.71459079e-01 9.70904052e-01 4.62013900e-01
-5.28252721e-01 3.79384071e-01 5.93965054e-01 1.66064590e-01
5.71690023e-01 -1.14390612e+00 -4.90712643e-01 2.10287765e-01
-2.06289351e-01 5.60675085e-01 6.90043271e-01 -7.07675159e-01
-9.16734636e-01 -7.03794122e-01 1.69322622e+00 -8.83766636e-02
6.85439587e-01 1.77582979e-01 -6.21598899e-01 3.74715269e-01
1.84911236e-01 -6.68607771e-01 1.05481565e+00 8.76836926e-02
4.13096920e-02 -1.72771916e-01 -1.37561548e+00 5.27067840e-01
3.53035182e-02 -3.65277171e-01 -6.91286743e-01 3.54078472e-01
4.80009802e-02 2.27962822e-01 -1.26200938e+00 5.25446050e-02
8.58843088e-01 -7.42553055e-01 6.29291058e-01 -3.38720500e-01
3.15323979e-01 -3.07244062e-02 -2.62244970e-01 -8.18526804e-01
9.44351256e-02 8.79570246e-02 3.30445349e-01 1.61634290e+00
5.60192049e-01 -1.22168326e+00 5.56817651e-01 6.11168981e-01
2.66646475e-01 -6.00842118e-01 -9.11671460e-01 -4.80191886e-01
2.17290029e-01 -4.28722858e-01 2.40501106e-01 1.32358813e+00
3.98405880e-01 5.74928403e-01 -2.52706762e-02 -2.26722538e-01
3.43639374e-01 -2.56421100e-02 3.93907011e-01 -1.50046146e+00
4.37810868e-02 -5.83464086e-01 -4.57392097e-01 -4.71844882e-01
5.07845804e-02 -1.00279260e+00 -5.15313447e-01 -1.51558471e+00
1.46676585e-01 -5.12287676e-01 -2.69728247e-02 4.85537916e-01
1.63721815e-01 -2.01481834e-01 -3.03420305e-01 4.60805386e-01
1.39210537e-01 1.40978619e-01 8.68963480e-01 -1.65309131e-01
-2.73209631e-01 4.45365459e-01 -5.10804594e-01 6.52690887e-01
1.16925859e+00 -7.49378681e-01 -3.66736442e-01 1.25269979e-01
2.85919815e-01 -1.01572245e-01 -2.39784881e-01 -8.91480982e-01
2.99937129e-01 -2.36789122e-01 6.82781041e-01 -9.27473366e-01
-1.58947274e-01 -8.29935789e-01 -2.47016475e-02 6.21244609e-01
-4.38281357e-01 1.21888913e-01 8.89406428e-02 1.83908090e-01
-2.93466479e-01 -9.95481789e-01 4.28354710e-01 5.89213744e-02
-3.89812976e-01 -2.85321116e-01 -1.11615980e+00 -5.46010554e-01
9.16850328e-01 -4.15896624e-01 -2.40787506e-01 -8.55438337e-02
-4.01083112e-01 9.43541527e-02 -1.30597353e-01 2.34856665e-01
2.85704583e-01 -9.31588411e-01 -3.84373575e-01 -9.31337178e-02
8.71872306e-02 -6.91795766e-01 1.18927762e-01 7.06244349e-01
-9.04043138e-01 9.74733472e-01 -3.58520210e-01 -9.73121300e-02
-1.55408502e+00 2.96521157e-01 -1.35854900e-01 -3.31135064e-01
-2.18491524e-01 3.14697087e-01 -8.24984074e-01 -5.56567073e-01
2.32860297e-01 -2.50925750e-01 -1.24503660e+00 4.11117792e-01
3.73011142e-01 8.33160639e-01 1.17649958e-01 -8.54369342e-01
-1.94885820e-01 4.11560386e-01 -8.01752582e-02 -2.05387220e-01
8.14842820e-01 -1.86659649e-01 -5.34488022e-01 6.36407554e-01
1.27775729e+00 -5.22984099e-03 2.75052786e-01 1.12924628e-01
2.99883783e-01 -2.55649000e-01 2.07888991e-01 -9.38778102e-01
-1.23256087e-01 5.63108325e-01 6.40078366e-01 8.75382960e-01
1.02907062e+00 -4.91016746e-01 3.80419552e-01 7.87379324e-01
3.60178977e-01 -1.55729651e+00 -5.37484884e-01 4.70756322e-01
5.66636205e-01 -1.15202999e+00 1.71631426e-01 -4.75029767e-01
-3.07535976e-01 1.61575019e+00 2.61958718e-01 3.93077314e-01
1.45457411e+00 9.51087698e-02 -1.25009581e-01 2.37552643e-01
-5.46133816e-01 2.25689523e-02 2.18396127e-01 4.46823865e-01
9.48709309e-01 -2.87463199e-02 -1.21464133e+00 8.11047792e-01
-6.53766751e-01 2.80832946e-01 3.38670790e-01 1.14738929e+00
-8.35781276e-01 -9.71154869e-01 -7.65758693e-01 1.14388847e+00
-8.98804963e-01 6.36514500e-02 -1.77238092e-01 9.03621018e-01
-1.83338225e-02 1.38516283e+00 -1.62969083e-01 -4.17695552e-01
6.61430787e-03 3.59949559e-01 1.20759159e-01 -2.55087197e-01
-5.74259043e-01 -2.07287967e-01 3.43797237e-01 5.73422194e-01
-5.18660545e-01 -6.26621127e-01 -1.16934049e+00 -4.41423923e-01
-5.88411808e-01 7.19054818e-01 1.24195182e+00 1.17055440e+00
-2.90947139e-01 2.31940925e-01 7.32628226e-01 -3.30424517e-01
-3.33852947e-01 -1.52817774e+00 -7.74023771e-01 -8.68862495e-02
-2.86543876e-01 -5.97304404e-01 -5.42238295e-01 1.78438306e-01] | [10.601959228515625, 7.274409770965576] |
59d547a1-425b-444f-b36a-588ba61cc434 | a-review-of-multi-objective-deep-learning | 2003.12108 | null | http://arxiv.org/abs/2003.12108v1 | http://arxiv.org/pdf/2003.12108v1.pdf | A Review of Multi-Objective Deep Learning Speech Denoising Methods | This paper presents a review of multi-objective deep learning methods that
have been introduced in the literature for speech denoising. After stating an
overview of conventional, single objective deep learning, and hybrid or
combined conventional and deep learning methods, a review of the mathematical
framework of the multi-objective deep learning methods for speech denoising is
provided. A representative method from each speech denoising category, whose
codes are publicly available, is selected and a comparison is carried out by
considering the same public domain dataset and four widely used objective
metrics. The comparison results indicate the effectiveness of the
multi-objective method compared with the other methods, in particular when the
signal-to-noise ratio is low. Possible future improvements that can be achieved
are also mentioned. | [] | 2020-03-26 | null | null | null | null | ['speech-denoising'] | ['speech'] | [-2.52275646e-01 -1.37693033e-01 1.87480077e-01 -4.67208117e-01
-1.29169405e+00 -4.22541723e-02 1.89656034e-01 4.92526144e-02
-6.42323196e-01 8.29148769e-01 4.58455712e-01 1.04195021e-01
-5.39528012e-01 -5.17617702e-01 -2.88377434e-01 -1.23978353e+00
-9.04416516e-02 9.96965766e-02 -2.46901780e-01 -3.64794940e-01
1.45453990e-01 2.94318438e-01 -1.55495906e+00 2.40809351e-01
6.94161594e-01 1.07091618e+00 2.29425922e-01 9.00230527e-01
1.16270548e-02 7.17400134e-01 -1.37124002e+00 -7.31981039e-01
-1.59782276e-01 -2.88764536e-01 -5.36194086e-01 9.02078673e-03
2.73261368e-01 -3.95010948e-01 -5.37141800e-01 1.21519911e+00
1.49266756e+00 6.11393750e-01 6.45964622e-01 -1.02085340e+00
-9.45729315e-01 3.74627531e-01 -8.73059034e-02 7.80090570e-01
7.36862347e-02 -2.16955040e-02 1.04795504e+00 -9.07383323e-01
3.84134310e-03 1.51836693e+00 1.05974281e+00 5.39787114e-01
-9.21119392e-01 -3.11460078e-01 2.03332365e-01 5.90283573e-01
-1.12764204e+00 -6.13784015e-01 8.62459719e-01 -1.01019360e-01
1.26534033e+00 2.13084698e-01 4.12618309e-01 1.22788954e+00
3.55637729e-01 8.76427531e-01 8.85251582e-01 -3.58786553e-01
3.98930639e-01 -2.45225042e-01 1.06763706e-01 6.05846465e-01
4.67404686e-02 4.44093585e-01 -8.59797001e-01 1.20636642e-01
-3.47789787e-02 -6.13390744e-01 -1.03186898e-01 3.38605702e-01
-6.46235704e-01 9.84288156e-01 1.89714134e-01 7.56920516e-01
-4.91344869e-01 2.57633388e-01 8.31939995e-01 6.16098106e-01
1.12988079e+00 -4.40173857e-02 -6.71869338e-01 -2.00034723e-01
-1.34330380e+00 1.91437215e-01 6.91567957e-01 3.21970850e-01
5.67492664e-01 9.69131887e-01 -3.28757763e-01 1.23550391e+00
3.76507550e-01 4.34669137e-01 3.56463164e-01 -1.13551545e+00
3.18973750e-01 -3.09468836e-01 -3.63787502e-01 -1.03946221e+00
-4.14856046e-01 -6.62435055e-01 -1.19753075e+00 4.44142371e-01
-1.71879604e-01 -3.99469435e-01 -1.04201317e+00 1.40783739e+00
-2.93026000e-01 2.04603933e-02 1.89935267e-01 6.93605065e-01
1.59760630e+00 8.45372617e-01 1.17366411e-01 -3.89640182e-01
8.47423971e-01 -1.00906026e+00 -1.49615085e+00 -2.55119801e-01
2.93035861e-02 -9.60424006e-01 6.03757679e-01 7.63899148e-01
-1.18209100e+00 -7.24965870e-01 -1.01773095e+00 -2.13958681e-01
-6.20690644e-01 8.72097313e-02 3.91501337e-01 1.22128105e+00
-1.41883731e+00 6.51199281e-01 -6.05576158e-01 2.78302585e-03
5.31064212e-01 4.22684759e-01 -4.31430228e-02 1.66461036e-01
-1.40635633e+00 1.04764760e+00 1.50926620e-01 3.44979912e-01
-1.37328696e+00 -5.44745445e-01 -7.57741451e-01 1.60149246e-01
2.58627161e-02 -5.70020378e-01 1.57497609e+00 -1.08576119e+00
-1.72757900e+00 6.94581628e-01 -2.02527836e-01 -5.02002537e-01
3.81016046e-01 -3.06862026e-01 -4.83171999e-01 6.04210049e-02
-2.16909617e-01 2.76089579e-01 1.17444122e+00 -1.33271444e+00
-6.07326984e-01 -2.14593142e-01 1.34345025e-01 7.98401535e-02
-4.66801047e-01 4.53773022e-01 -2.49904737e-01 -1.17989969e+00
-1.03612721e-01 1.04831243e-02 -2.25742757e-01 -8.22651088e-02
-2.67907858e-01 -3.30337465e-01 9.38846827e-01 -1.25338852e+00
1.25881815e+00 -1.94383717e+00 4.81320202e-01 -1.49274901e-01
2.05336735e-01 5.64456701e-01 -3.74135226e-01 6.20924354e-01
-2.51638412e-01 2.36566722e-01 -4.26466823e-01 -1.26487541e+00
1.13741808e-01 4.11839664e-01 9.76191014e-02 6.92602515e-01
-1.69529468e-01 5.70719182e-01 -7.41860569e-01 -1.40363008e-01
5.24052143e-01 9.81205165e-01 -2.09423542e-01 1.97267264e-01
2.76410252e-01 3.27254564e-01 -4.57181446e-02 8.16592038e-01
9.76743281e-01 6.11190677e-01 -2.13743448e-01 -3.97920638e-01
-1.25000775e-01 4.90575910e-01 -1.20872962e+00 1.39288521e+00
-8.69313121e-01 1.17085993e+00 7.75442958e-01 -1.60851419e+00
8.29159141e-01 8.61537874e-01 4.15704072e-01 -7.69276023e-01
1.61700755e-01 3.94169271e-01 -1.37269571e-01 -5.25987327e-01
1.24614954e-01 -4.80994493e-01 5.20634294e-01 6.26448393e-02
6.37783706e-01 -3.16489786e-01 3.52852374e-01 -2.09335700e-01
8.06228936e-01 -5.98069131e-01 1.93328574e-01 -3.80943507e-01
1.07535040e+00 -5.41554213e-01 6.88267767e-01 9.53436673e-01
-6.66835010e-01 7.53770709e-01 4.54290003e-01 -4.46669787e-01
-9.37034369e-01 -9.44639444e-01 -5.40946722e-02 1.23374784e+00
-1.62580624e-01 -1.40709862e-01 -8.53531063e-01 -5.35050333e-01
-1.24178298e-01 3.87731224e-01 -5.37696302e-01 2.10523400e-02
-4.84055936e-01 -1.33176732e+00 7.37448931e-01 5.68804801e-01
6.42317593e-01 -1.18440032e+00 9.23448876e-02 1.76162824e-01
-5.23595810e-01 -8.96449268e-01 -1.70224775e-02 4.52721179e-01
-1.01925194e+00 -7.88412035e-01 -9.95469391e-01 -1.10633135e+00
-8.61483216e-02 1.79870814e-01 1.31191778e+00 3.16912442e-01
-7.38189463e-03 6.16501868e-01 -3.76468569e-01 -5.41189611e-01
-6.25776649e-01 -1.45231351e-01 1.44534066e-01 -1.97770268e-01
3.08302104e-01 -5.37056208e-01 -4.42947805e-01 -2.38633305e-01
-7.76879609e-01 -8.97507012e-01 2.43377715e-01 1.09598136e+00
3.73640865e-01 5.33575892e-01 1.00557327e+00 -1.14281043e-01
1.07453668e+00 -2.65678644e-01 -5.00778615e-01 -3.02467346e-02
-3.65230501e-01 -3.25555563e-01 3.98647994e-01 -9.75392237e-02
-1.21801710e+00 -4.73916352e-01 -1.08829069e+00 4.05454226e-02
-8.86196718e-02 7.30929613e-01 -2.73119032e-01 -2.77368277e-01
4.34043318e-01 1.41716093e-01 1.13662146e-01 -8.86184454e-01
1.42726615e-01 7.54072428e-01 2.16796085e-01 -2.16471866e-01
5.51832020e-01 3.72872204e-01 -4.39013481e-01 -1.33838820e+00
-9.03223217e-01 -3.96920830e-01 -3.46399158e-01 -3.80880028e-01
9.70382929e-01 -9.54968989e-01 -4.87363368e-01 1.16370094e+00
-1.54307771e+00 -7.86566436e-02 -2.73104221e-01 5.36025465e-01
-6.74998581e-01 5.43959081e-01 -7.70038843e-01 -1.01493514e+00
-7.13836849e-01 -1.57906258e+00 9.13648844e-01 -3.28899026e-02
3.28372002e-01 -1.54872835e+00 4.88472916e-02 5.66603959e-01
7.33578384e-01 -1.37797832e-01 6.97500646e-01 -6.69373512e-01
-5.86855002e-02 -1.91634655e-01 4.25001122e-02 1.46949387e+00
1.79832280e-01 -1.11723028e-01 -1.42072058e+00 -5.35827339e-01
5.73410273e-01 -2.39411265e-01 1.21558046e+00 1.06377172e+00
1.28735662e+00 -1.82003111e-01 3.44233096e-01 6.02693081e-01
1.25953412e+00 2.40015611e-01 7.93800652e-01 5.40676415e-01
1.65550724e-01 4.75711286e-01 2.15052769e-01 2.78248250e-01
-4.13550362e-02 3.41492712e-01 7.16950834e-01 -4.60532993e-01
-6.07854784e-01 6.82035327e-01 6.42133534e-01 1.54012430e+00
-2.33576164e-01 -7.97539473e-01 -5.97663760e-01 7.98483312e-01
-1.37805104e+00 -1.09078586e+00 1.35346793e-03 1.71867919e+00
5.30622244e-01 -6.09381683e-02 9.14583132e-02 9.28040922e-01
7.90711403e-01 7.15909123e-01 -2.69302011e-01 -7.55493164e-01
-7.07475543e-01 5.92487991e-01 2.85576284e-01 1.01602483e+00
-1.63152277e+00 4.64940012e-01 7.68372154e+00 1.09306693e+00
-8.69296372e-01 6.40913546e-01 7.04297423e-01 3.29237245e-02
1.78800635e-02 -7.22489297e-01 -2.75251716e-01 2.01136872e-01
1.08269882e+00 -1.91034317e-01 4.28695858e-01 5.34494996e-01
7.46961951e-01 1.43789351e-01 -3.96296144e-01 1.04328585e+00
2.42322326e-01 -1.31811750e+00 -1.95266321e-01 -1.80381238e-01
7.47843027e-01 4.05419946e-01 2.51553208e-01 2.66631126e-01
-2.96468083e-02 -9.78315771e-01 6.52943254e-01 4.28584933e-01
4.44498986e-01 -1.11301684e+00 1.29517579e+00 1.47090843e-02
-9.51910257e-01 -4.57964867e-01 -5.34235001e-01 -2.80094873e-02
3.28207523e-01 1.12623870e+00 8.03754106e-02 8.19598913e-01
1.28594530e+00 9.41948473e-01 -1.38339221e-01 1.29247856e+00
-4.78555769e-01 8.86440456e-01 -1.47675620e-02 1.58574790e-01
3.70382071e-01 -1.95742339e-01 1.01515484e+00 1.67537153e+00
4.69266087e-01 -1.76398918e-01 -2.78563976e-01 3.62965047e-01
-8.05620253e-02 4.06161658e-02 -2.75923014e-01 2.34543514e-02
7.42313936e-02 9.32903469e-01 -2.33307406e-01 -2.96775758e-01
-6.18681729e-01 7.08630502e-01 -1.74639761e-01 5.77512264e-01
-6.68224156e-01 -6.61225080e-01 1.20804095e+00 -4.60948586e-01
3.38788033e-01 -3.88476998e-01 -6.20014548e-01 -7.78496861e-01
-4.61005494e-02 -1.30974853e+00 3.08131427e-01 -6.68601513e-01
-1.41095531e+00 7.96594501e-01 -6.21873066e-02 -8.12367976e-01
1.90679669e-01 -8.19311500e-01 -6.22930944e-01 1.04525399e+00
-1.76993549e+00 -7.33556569e-01 -2.63205677e-01 2.57357299e-01
1.13237596e+00 -7.01172233e-01 7.72445321e-01 8.87858570e-01
-7.21016407e-01 6.91150546e-01 7.68301189e-01 -8.77877846e-02
4.19573218e-01 -1.24803793e+00 4.45386559e-01 8.95557344e-01
-5.48829883e-02 1.43277690e-01 8.53072464e-01 -2.90613055e-01
-9.60563362e-01 -8.28673899e-01 1.13736773e+00 1.15213461e-01
5.40464997e-01 -6.43887073e-02 -1.01525331e+00 2.57779151e-01
9.47544694e-01 -3.73005807e-01 5.42541921e-01 -1.27495989e-01
-3.17907184e-02 -4.38095897e-01 -1.43620467e+00 1.66436836e-01
6.29138529e-01 -5.02384424e-01 -5.82528114e-01 4.42046285e-01
5.17033219e-01 -2.29046032e-01 -7.92486429e-01 5.11659682e-01
1.75400123e-01 -1.18001175e+00 1.38858414e+00 -4.94281411e-01
2.59364098e-01 -9.45293754e-02 -4.80150819e-01 -1.81679380e+00
-2.70539016e-01 -6.91723883e-01 -3.56646538e-01 1.11839664e+00
7.63994530e-02 -5.02518415e-01 4.81719553e-01 -4.30903763e-01
-7.34602213e-01 -6.22839808e-01 -1.74379480e+00 -9.79258597e-01
4.06997621e-01 -8.36098969e-01 3.14450145e-01 3.63790005e-01
-8.09163213e-01 1.86344489e-01 -5.99141538e-01 3.43151748e-01
8.28336298e-01 -8.13081861e-01 1.53355137e-01 -1.15690136e+00
2.23490387e-01 -6.21488214e-01 -6.37771249e-01 -8.80076468e-01
3.69662672e-01 -7.03222632e-01 7.72461444e-02 -2.05999136e+00
-2.06967279e-01 2.97852755e-01 -5.99777579e-01 2.22449318e-01
-2.92263944e-02 3.92124861e-01 6.25027493e-02 -3.67337048e-01
-2.17416078e-01 1.00612724e+00 1.15754902e+00 -7.35439360e-01
4.35159430e-02 3.29801679e-01 -5.62393725e-01 8.47816050e-01
1.06082845e+00 -5.64025164e-01 -2.26987213e-01 -1.02590191e+00
-1.44998118e-01 -4.25113618e-01 3.09152275e-01 -1.12571847e+00
1.87859893e-01 3.29585284e-01 7.16380551e-02 -8.16787601e-01
6.90855443e-01 -6.07240140e-01 -4.40148503e-01 4.76196706e-01
-2.36844882e-01 6.62528649e-02 3.40163738e-01 6.06316984e-01
-9.38421369e-01 -6.45183027e-01 1.38057804e+00 -9.06654596e-02
-6.93588138e-01 3.42887901e-02 -9.42880630e-01 -2.15360913e-02
3.93195897e-01 -4.25778627e-02 2.69444380e-02 -8.11620712e-01
-1.12947714e+00 -1.31978959e-01 -6.33615911e-01 4.67340022e-01
9.51334834e-01 -1.17579198e+00 -1.11442268e+00 -1.21251866e-01
-4.02664214e-01 -6.87804878e-01 5.22735357e-01 8.24354589e-01
-4.30300951e-01 3.32657278e-01 1.13186173e-01 -3.77420545e-01
-1.38569772e+00 1.87902123e-01 1.02082539e+00 -8.11284557e-02
-3.08254302e-01 1.18031561e+00 -1.08681023e-01 -6.68150783e-01
9.85112429e-01 -8.32175687e-02 -3.93510640e-01 3.54589880e-01
5.22852838e-01 1.09888327e+00 9.36698377e-01 -6.76157594e-01
-4.35944408e-01 4.00772452e-01 4.07222837e-01 -9.29269940e-02
1.71631742e+00 -3.56077254e-01 -5.16073883e-01 2.91777253e-01
1.40609193e+00 -3.07007372e-01 -8.23473394e-01 -8.39699283e-02
-2.52478868e-01 -1.84680149e-01 8.69675756e-01 -1.01948559e+00
-1.76665711e+00 1.28777468e+00 1.18071067e+00 3.81967336e-01
1.70457947e+00 -2.97676504e-01 9.04827356e-01 4.32028472e-01
-2.34301686e-01 -1.47426808e+00 3.50849986e-01 9.57567811e-01
1.32289290e+00 -1.40670300e+00 1.36292815e-01 -1.61795676e-01
-2.07090154e-01 1.19891691e+00 2.03053877e-01 9.27483570e-03
1.06226742e+00 3.02062869e-01 5.08365870e-01 -1.37950420e-01
-3.50190371e-01 -3.77089828e-01 3.00704926e-01 1.24044645e+00
6.97114706e-01 -3.01999837e-01 -5.35729825e-01 5.14397800e-01
6.67565539e-02 -2.33354256e-01 1.88962549e-01 6.77350342e-01
-5.88923752e-01 -1.19384289e+00 -5.04615486e-01 2.22875312e-01
-9.69707549e-01 -2.84271479e-01 -1.40219748e-01 5.24693608e-01
1.66611448e-01 1.72630453e+00 -1.17240332e-01 -2.13137224e-01
5.96096933e-01 -2.83007741e-01 1.29501924e-01 -3.51568341e-01
-1.02680922e+00 1.10776901e-01 3.65097135e-01 -2.13893846e-01
-8.51904571e-01 -3.80967379e-01 -5.93667448e-01 -4.45253551e-01
-2.99948037e-01 -9.50790418e-04 9.58737731e-01 9.85161066e-01
3.26427743e-02 9.48535860e-01 5.51577866e-01 -1.12196410e+00
-5.22253036e-01 -1.09346807e+00 -4.32199985e-01 -1.75668508e-01
1.05531454e+00 -7.77662218e-01 -9.98250544e-01 2.22942885e-02] | [15.102602005004883, 5.90800142288208] |
f1504c0a-0b49-4941-b7af-a0b012e30571 | pearl-preprocessing-enhanced-adversarial | 2305.15709 | null | https://arxiv.org/abs/2305.15709v1 | https://arxiv.org/pdf/2305.15709v1.pdf | PEARL: Preprocessing Enhanced Adversarial Robust Learning of Image Deraining for Semantic Segmentation | In light of the significant progress made in the development and application of semantic segmentation tasks, there has been increasing attention towards improving the robustness of segmentation models against natural degradation factors (e.g., rain streaks) or artificially attack factors (e.g., adversarial attack). Whereas, most existing methods are designed to address a single degradation factor and are tailored to specific application scenarios. In this work, we present the first attempt to improve the robustness of semantic segmentation tasks by simultaneously handling different types of degradation factors. Specifically, we introduce the Preprocessing Enhanced Adversarial Robust Learning (PEARL) framework based on the analysis of our proposed Naive Adversarial Training (NAT) framework. Our approach effectively handles both rain streaks and adversarial perturbation by transferring the robustness of the segmentation model to the image derain model. Furthermore, as opposed to the commonly used Negative Adversarial Attack (NAA), we design the Auxiliary Mirror Attack (AMA) to introduce positive information prior to the training of the PEARL framework, which improves defense capability and segmentation performance. Our extensive experiments and ablation studies based on different derain methods and segmentation models have demonstrated the significant performance improvement of PEARL with AMA in defense against various adversarial attacks and rain streaks while maintaining high generalization performance across different datasets. | ['Xin Fan', 'Risheng Liu', 'Xinyuan Chu', 'Jiaxin Gao', 'Yaohua Liu', 'Xianghao Jiao'] | 2023-05-25 | null | null | null | null | ['adversarial-attack'] | ['adversarial'] | [ 6.90610945e-01 -7.98885226e-02 3.90178174e-01 -9.13846716e-02
-5.67594528e-01 -1.05748987e+00 6.43541157e-01 -4.52821255e-02
-2.90089041e-01 3.78296167e-01 -1.97682977e-01 -2.69581944e-01
2.23889574e-02 -8.71335566e-01 -9.04792547e-01 -9.33821499e-01
9.13785100e-02 -7.56386071e-02 5.58872521e-01 -4.27049011e-01
3.52122068e-01 7.13873923e-01 -1.40504217e+00 -2.30470095e-02
1.22357702e+00 8.25220406e-01 -1.32965639e-01 8.30016732e-01
4.32488546e-02 4.06466067e-01 -1.09162080e+00 -4.03176576e-01
6.04890883e-01 -2.04018384e-01 -7.40730524e-01 5.53752743e-02
4.57572341e-01 -2.07937330e-01 -1.07461117e-01 1.26609933e+00
5.88586986e-01 2.04108715e-01 4.59199518e-01 -1.44642532e+00
-7.03284442e-01 4.67595309e-01 -7.77407229e-01 1.92411229e-01
1.01639122e-01 3.31384897e-01 3.94402891e-01 -4.31212753e-01
3.24391395e-01 1.43565893e+00 7.08549798e-01 7.71732628e-01
-1.03231788e+00 -7.66832650e-01 3.93318504e-01 -7.43565559e-02
-9.36032355e-01 -3.81546244e-02 7.93284476e-01 -1.68151349e-01
5.35415053e-01 2.66076624e-01 1.39910564e-01 1.43129814e+00
1.86450005e-01 7.89432943e-01 1.32925785e+00 -3.47442031e-01
3.87944788e-01 -2.73441412e-02 1.11232959e-01 4.00153399e-01
1.56494930e-01 1.73562258e-01 -5.14319390e-02 -1.21079631e-01
4.61671174e-01 -3.10865611e-01 -1.91455200e-01 -2.19156325e-01
-5.80290377e-01 7.85691559e-01 5.89441061e-01 6.89514801e-02
-1.87419981e-01 -2.67393719e-02 5.23164093e-01 2.33541131e-01
4.87115264e-01 6.25061035e-01 -2.84976095e-01 3.11441451e-01
-8.32434833e-01 2.91197181e-01 5.26036441e-01 6.40251935e-01
4.51945364e-01 5.46805322e-01 -2.92284429e-01 8.71172011e-01
1.76217958e-01 7.66710222e-01 3.31360072e-01 -7.74828553e-01
3.78275096e-01 4.11320508e-01 -1.26869276e-01 -1.08286357e+00
-3.39825749e-01 -6.00859702e-01 -6.82800412e-01 6.34310842e-01
3.76209795e-01 -2.31978193e-01 -1.53564084e+00 1.96090913e+00
5.62931001e-01 3.70827198e-01 4.53967839e-01 7.22492158e-01
4.32848275e-01 5.47955096e-01 3.72703016e-01 7.56947622e-02
1.03773189e+00 -9.42443907e-01 -5.16600668e-01 -4.32252020e-01
2.30111763e-01 -9.31756318e-01 1.22051537e+00 4.93389636e-01
-7.62985349e-01 -5.62789857e-01 -1.23635888e+00 3.61013293e-01
-6.73599422e-01 -5.22764027e-01 4.98798341e-01 1.24803615e+00
-6.92740262e-01 4.69016999e-01 -8.17866445e-01 -3.27963501e-01
4.93361443e-01 1.92658857e-01 -2.03366816e-01 -1.35125726e-01
-1.39801288e+00 6.89155698e-01 4.47591722e-01 2.69129813e-01
-1.10219610e+00 -7.74403334e-01 -7.28884459e-01 -3.09972465e-01
4.28976148e-01 -3.05602103e-01 8.13260615e-01 -1.21646631e+00
-1.55642152e+00 6.57565415e-01 4.91955757e-01 -7.68938184e-01
7.14306533e-01 -4.86497879e-01 -3.37651014e-01 3.11011642e-01
-1.05244242e-01 5.58916867e-01 1.24273229e+00 -1.59674597e+00
-2.57415473e-01 -3.16483289e-01 4.18209463e-01 6.59337193e-02
-4.75186586e-01 1.77378267e-01 -2.05097303e-01 -1.14953399e+00
1.17910176e-03 -1.16359305e+00 -2.96058953e-01 -1.24451496e-01
-4.89606142e-01 4.01697278e-01 1.30913210e+00 -6.75339580e-01
9.03449714e-01 -2.24761415e+00 1.29489362e-01 2.82663763e-01
-1.89247653e-01 9.75220501e-01 -3.21080327e-01 2.58453190e-01
-2.21670195e-01 3.35080683e-01 -8.87080789e-01 -1.44826949e-01
-8.45601112e-02 4.74908262e-01 -6.53583288e-01 3.98081154e-01
4.24243271e-01 5.89503527e-01 -7.80481577e-01 -2.32589975e-01
3.15541089e-01 4.33137387e-01 -4.73863125e-01 2.58643121e-01
-2.23241702e-01 5.35799086e-01 -5.12783706e-01 9.44048762e-01
1.06598675e+00 4.70741898e-01 -4.67026979e-01 -1.43285453e-01
2.90551037e-01 -4.43441838e-01 -1.15856910e+00 1.27026856e+00
-5.51715791e-01 1.21712729e-01 3.53755772e-01 -9.36035633e-01
8.25030148e-01 1.90889671e-01 9.28408578e-02 -6.40512884e-01
1.49597257e-01 1.31833315e-01 -8.76722783e-02 -3.00186336e-01
2.59503782e-01 -1.77698210e-01 -3.23705494e-01 2.54983246e-01
-1.02521807e-01 -3.07675958e-01 -7.92871118e-02 2.96840101e-01
1.15682960e+00 3.25737536e-01 1.17452163e-02 -8.38733837e-02
7.61056304e-01 -5.55904694e-02 7.19035625e-01 8.55987668e-01
-6.24854863e-01 7.24743605e-01 3.13868612e-01 -6.23267815e-02
-8.29194427e-01 -1.31695342e+00 -6.31932095e-02 9.98585939e-01
3.93531859e-01 -8.16616416e-03 -1.18191528e+00 -1.01291776e+00
-6.74740300e-02 7.19778836e-01 -6.41364217e-01 -5.57991564e-01
-6.72779918e-01 -1.19066954e+00 1.33370650e+00 5.85093617e-01
1.14054894e+00 -1.22729731e+00 -4.70989794e-01 -3.46251465e-02
8.16249475e-02 -1.19703364e+00 -2.20510229e-01 1.55673668e-01
-5.52894771e-01 -1.11998165e+00 -4.82250124e-01 -4.72957253e-01
5.04197478e-01 1.85700506e-01 7.17953980e-01 -1.49139389e-02
-4.83950734e-01 5.45300543e-01 -7.00287819e-01 -6.67912126e-01
-7.04774678e-01 -1.27960116e-01 3.96833494e-02 6.09550886e-02
-1.99052721e-01 -5.66173375e-01 -4.06678706e-01 3.51757765e-01
-1.58634174e+00 -4.37042117e-01 5.36142349e-01 6.40109599e-01
3.98192823e-01 3.58950615e-01 5.88791192e-01 -1.14283502e+00
5.51659405e-01 -5.66180944e-01 -4.95041132e-01 1.88603461e-01
-6.07152581e-01 -1.23824216e-01 9.21713471e-01 -6.42104387e-01
-1.23966336e+00 -6.96133971e-02 -4.28360581e-01 -5.36076427e-01
-5.08885026e-01 1.15022510e-01 -5.10660529e-01 -4.34561461e-01
8.02580953e-01 -1.02279209e-01 -2.48118773e-01 -2.99216419e-01
4.38905209e-01 4.63629633e-01 7.61668384e-01 -7.82503903e-01
1.49033296e+00 5.71808994e-01 1.59956709e-01 -7.73573458e-01
-8.60976100e-01 -6.79703951e-02 -4.21764076e-01 -1.90606534e-01
9.69378352e-01 -5.58134377e-01 -2.45986700e-01 1.27274072e+00
-7.46112645e-01 -3.47243398e-01 -1.07882693e-01 -6.58010989e-02
-3.36795539e-01 6.88519776e-01 -5.27049065e-01 -8.68559599e-01
-5.25175154e-01 -1.15407646e+00 8.52247059e-01 4.17999864e-01
1.96258560e-01 -9.42025065e-01 -1.06071703e-01 6.12427831e-01
6.59747601e-01 1.01609004e+00 9.55481112e-01 -7.21230686e-01
-3.20025951e-01 -2.73171693e-01 -2.34468840e-02 8.67089868e-01
1.78711399e-01 1.41990453e-01 -1.19652843e+00 -4.55585808e-01
3.35476816e-01 -5.18933713e-01 8.39788795e-01 -4.53420803e-02
1.05449831e+00 -2.56657094e-01 1.97987929e-02 7.87865698e-01
1.41053486e+00 2.46670231e-01 7.86770344e-01 6.84472382e-01
8.48480344e-01 6.76196814e-01 8.37044299e-01 2.89875865e-02
-3.74504894e-01 4.26563114e-01 8.86578798e-01 -3.49456787e-01
-1.14158832e-01 -3.38281728e-02 3.70595664e-01 8.39432701e-02
2.39207402e-01 -5.22083223e-01 -8.69673848e-01 3.67686391e-01
-1.61516166e+00 -7.19776273e-01 1.81754697e-02 2.05652833e+00
6.36007845e-01 4.65165526e-01 -1.35838136e-01 4.11261737e-01
7.95890987e-01 4.87621099e-01 -8.19055557e-01 -7.74760067e-01
-3.50188315e-01 4.05391306e-01 6.95746481e-01 5.09090483e-01
-1.41777980e+00 1.28495240e+00 6.18558502e+00 9.49494898e-01
-1.13079071e+00 -4.26210538e-02 3.50082606e-01 9.15489197e-02
-1.23848386e-01 -1.64591134e-01 -3.80931646e-01 3.85064930e-01
6.84339523e-01 2.17753053e-01 4.61848468e-01 7.45777726e-01
9.89203155e-02 1.27259538e-01 -4.45558101e-01 3.44573617e-01
1.13175936e-01 -6.11258566e-01 3.16638380e-01 -1.88410610e-01
7.00134575e-01 -3.09860915e-01 3.60490918e-01 3.75154108e-01
5.19054055e-01 -9.60050642e-01 5.64460695e-01 1.07072346e-01
2.92831898e-01 -8.86811316e-01 7.37680316e-01 2.37375617e-01
-8.82829547e-01 -2.24920556e-01 -5.48111275e-02 2.15617090e-01
-1.03232823e-01 4.38747436e-01 -4.84199703e-01 7.09785700e-01
7.97269344e-01 2.63296396e-01 -8.24746192e-01 6.45235121e-01
-5.22207439e-01 9.85991955e-01 -3.12660336e-01 6.37045562e-01
5.29901624e-01 -3.45357917e-02 8.72914433e-01 1.11678839e+00
-1.74472049e-01 -9.55456048e-02 1.99449837e-01 6.46586299e-01
6.62388802e-02 -6.31417753e-03 -5.81399441e-01 1.45775098e-02
3.47799808e-01 1.16483355e+00 -9.39612091e-01 9.13270786e-02
-1.93297401e-01 1.18182576e+00 2.99612358e-02 3.84647042e-01
-1.15681589e+00 -3.96019310e-01 8.30390930e-01 -2.06700563e-01
2.55497605e-01 -8.28973055e-02 -3.79138499e-01 -7.61528492e-01
5.81034720e-02 -1.28060150e+00 4.57860470e-01 -5.70184290e-01
-1.38771272e+00 6.50027752e-01 7.87815750e-02 -1.05132747e+00
1.04909241e-01 -5.25932431e-01 -1.05096364e+00 8.76069069e-01
-1.43495071e+00 -1.58396173e+00 -2.38394409e-01 7.32725799e-01
6.09921873e-01 -1.62099153e-01 6.57198727e-01 4.43025790e-02
-8.46436441e-01 7.96019614e-01 -6.45713061e-02 -1.69822033e-02
7.61147678e-01 -1.37240958e+00 5.73134184e-01 1.65232956e+00
-2.35393852e-01 6.28291249e-01 1.03344882e+00 -6.89881325e-01
-1.00389004e+00 -1.35606933e+00 -1.66954681e-01 -2.46819451e-01
6.45621002e-01 -4.01139319e-01 -1.23317397e+00 4.83142614e-01
1.50864199e-01 -1.28838748e-01 5.28280437e-01 -3.99053276e-01
-6.74383640e-01 -8.83140862e-02 -1.71443057e+00 7.52326846e-01
7.78637588e-01 -3.48497629e-01 -5.23208380e-01 2.26042107e-01
9.86561656e-01 -4.05396163e-01 -7.16503143e-01 7.50550151e-01
1.76365271e-01 -8.21928382e-01 1.24963355e+00 -4.96708125e-01
1.75842285e-01 -5.49733043e-01 -2.75375664e-01 -1.21409941e+00
-3.14745083e-02 -6.17312133e-01 -3.04404111e-03 1.66171932e+00
7.31835961e-02 -7.50865281e-01 5.77648699e-01 5.15428543e-01
-2.47620791e-01 -5.04505157e-01 -8.39633882e-01 -8.33598495e-01
3.62088799e-01 -3.99715722e-01 4.92572010e-01 8.86808753e-01
-7.52646148e-01 -2.00480193e-01 -4.50436890e-01 8.43343377e-01
6.61184728e-01 -1.52289778e-01 7.78988421e-01 -8.33429813e-01
-4.37487930e-01 -1.53271630e-01 -4.01262939e-01 -2.80017883e-01
2.34853581e-01 -6.26474500e-01 2.24735215e-01 -1.19224727e+00
-2.42288202e-01 -3.46031755e-01 -5.13300896e-01 6.00285828e-01
-7.23136961e-01 6.49487913e-01 3.32510799e-01 1.20404616e-01
-2.23832861e-01 3.81961167e-01 1.05024111e+00 -1.25880703e-01
-1.11839294e-01 2.16525152e-01 -8.17385375e-01 9.51933444e-01
1.03508234e+00 -5.24182260e-01 -6.03382051e-01 -4.17334020e-01
-9.37599242e-02 -5.37359118e-01 4.32303905e-01 -1.15279245e+00
-2.04925805e-01 -7.08275437e-02 1.89561799e-01 -1.98365360e-01
1.10999018e-01 -7.35079527e-01 -2.23504886e-01 5.12463570e-01
-1.52254567e-01 -5.84324747e-02 6.56474829e-01 7.74271846e-01
-1.01762198e-01 -2.08534122e-01 1.25014699e+00 -9.99192987e-03
-7.85867572e-01 1.96630545e-02 -3.48513365e-01 1.05756119e-01
1.26493359e+00 -2.52336890e-01 -5.46282649e-01 9.00864787e-03
-5.84547698e-01 3.35881650e-01 6.50949121e-01 6.38915181e-01
4.66549546e-01 -8.46850872e-01 -5.80589473e-01 2.32187480e-01
-3.81634980e-02 1.07026407e-02 3.28831136e-01 2.88721621e-01
-5.62522411e-01 -1.77892148e-01 -4.25487727e-01 -3.06394875e-01
-1.42784905e+00 8.94731522e-01 4.50895339e-01 -3.00388068e-01
-4.72594708e-01 8.47951412e-01 3.62342596e-01 -5.06406188e-01
3.56664300e-01 8.84425640e-02 -1.49423435e-01 -2.26865411e-01
3.49300057e-01 5.80365360e-01 8.99464712e-02 -5.39466679e-01
-4.51625973e-01 6.14510119e-01 -2.15475634e-01 4.47254106e-02
1.06894362e+00 -3.29061858e-02 2.63615511e-02 8.64019021e-02
7.72009313e-01 1.42486960e-01 -1.36457717e+00 -2.74065472e-02
-1.21580124e-01 -3.20715874e-01 1.28417425e-02 -1.06774569e+00
-1.25545073e+00 7.89107561e-01 8.47123384e-01 3.04198116e-01
1.53708887e+00 -5.24981081e-01 9.46573377e-01 9.31684673e-02
2.21479893e-01 -9.63445544e-01 1.76382348e-01 3.23960483e-01
7.90887475e-01 -1.05449021e+00 -9.72344652e-02 -6.58639312e-01
-7.33839571e-01 7.15353608e-01 7.77526259e-01 -2.95512527e-01
4.17831242e-01 4.64562356e-01 5.42786896e-01 1.25673607e-01
-1.15112729e-01 -1.38404265e-01 1.13020521e-02 9.84828353e-01
-1.11096315e-01 -2.27393702e-01 -2.28180870e-01 3.49859238e-01
1.26800425e-02 -5.14321506e-01 5.63530266e-01 1.16987646e+00
-3.00332129e-01 -1.21272933e+00 -7.83671200e-01 -3.15244459e-02
-7.61275649e-01 1.78495590e-02 -6.50309563e-01 8.18776190e-01
3.70690435e-01 1.28838551e+00 -4.34230328e-01 -3.68982732e-01
4.65496898e-01 4.91332784e-02 1.80705935e-01 -5.23079813e-01
-9.86698747e-01 -1.57838270e-01 -2.79541194e-01 -6.20458961e-01
-4.07939255e-01 -4.69383061e-01 -1.13260651e+00 -1.32581800e-01
-2.18379766e-01 -9.87488553e-02 6.49945736e-01 1.03580141e+00
6.49975911e-02 7.74894178e-01 7.20025957e-01 -6.86440289e-01
-5.86836517e-01 -8.04390430e-01 -4.76111889e-01 7.50964344e-01
2.32058167e-01 -5.98878384e-01 -6.63129866e-01 -8.45260471e-02] | [5.52023983001709, 7.957825660705566] |
6378acc3-477b-4c26-b319-9fcea939f268 | relation-based-generalized-zero-shot | null | null | https://openreview.net/forum?id=BJl8ZlHFwr | https://openreview.net/pdf?id=BJl8ZlHFwr | Relation-based Generalized Zero-shot Classification with the Domain Discriminator on the shared representation | Generalized zero-shot learning (GZSL) is the task of predicting a test image from seen or unseen classes using pre-defined class-attributes and images from the seen classes. Typical ZSL models assign the class corresponding to the most relevant attribute as the predicted label of the test image based on the learned relation between the attribute and the image. However, this relation-based approach presents a difficulty: many of the test images are predicted as biased to the seen domain, i.e., the \emph{domain bias problem}. Recently, many methods have addressed this difficulty using a synthesis-based approach that, however, requires generation of large amounts of high-quality unseen images after training and the additional training of classifier given them. Therefore, for this study, we aim at alleviating this difficulty in the manner of the relation-based approach. First, we consider the requirements for good performance in a ZSL setting and introduce a new model based on a variational autoencoder that learns to embed attributes and images into the shared representation space which satisfies those requirements. Next, we assume that the domain bias problem in GZSL derives from a situation in which embedding of the unseen domain overlaps that of the seen one. We introduce a discriminator that distinguishes domains in a shared space and learns jointly with the above embedding model to prevent this situation. After training, we can obtain prior knowledge from the discriminator of which domain is more likely to be embedded anywhere in the shared space. We propose combination of this knowledge and the relation-based classification on the embedded shared space as a mixture model to compensate class prediction. Experimentally obtained results confirm that the proposed method significantly improves the domain bias problem in relation-based settings and achieves almost equal accuracy to that of high-cost synthesis-based methods.
| ['Yutaka Matsuo', 'Masahiro Suzuki'] | 2019-09-25 | null | null | null | null | ['generalized-zero-shot-learning', 'generalized-zero-shot-learning'] | ['computer-vision', 'methodology'] | [ 6.01302087e-01 4.23995465e-01 -2.04144746e-01 -4.54569250e-01
-7.64096081e-01 -3.56244087e-01 6.31516576e-01 -1.78755894e-01
-1.66475341e-01 7.67056286e-01 -2.27404878e-01 1.74612805e-01
-2.98786610e-01 -1.07131124e+00 -8.77789080e-01 -1.06065774e+00
4.97754991e-01 7.25923121e-01 4.18801367e-01 -1.28874555e-01
9.11417603e-03 3.95937324e-01 -2.14590883e+00 3.15618843e-01
8.32797229e-01 1.15574849e+00 3.87082875e-01 4.13769871e-01
-2.01036185e-01 5.05050659e-01 -6.35969043e-01 -3.63507032e-01
4.97662812e-01 -7.27097750e-01 -5.90731680e-01 3.72572422e-01
5.19549549e-01 -1.66131124e-01 -7.75183290e-02 1.18719435e+00
3.33928078e-01 2.63140470e-01 1.16226065e+00 -1.62116289e+00
-6.26401007e-01 -6.77599292e-03 -1.68709427e-01 -1.94687411e-01
5.25925979e-02 -9.65673104e-02 9.92369950e-01 -9.31717575e-01
8.89791191e-01 9.09855843e-01 2.20831946e-01 8.97345245e-01
-1.40219522e+00 -5.82140267e-01 -5.49995760e-03 2.58178115e-01
-1.48709476e+00 -3.43162954e-01 1.01081431e+00 -4.78608429e-01
5.33594728e-01 1.50419071e-01 3.41354102e-01 1.40830994e+00
-1.41031817e-01 5.83745539e-01 1.17899621e+00 -6.18141651e-01
5.23695886e-01 9.87585723e-01 -1.19037353e-01 3.54286522e-01
8.76671970e-02 1.16536789e-01 -2.42246687e-01 -1.94363284e-03
5.57272971e-01 -9.23678130e-02 -3.42484504e-01 -1.01075757e+00
-8.10418427e-01 9.78925824e-01 3.09800923e-01 3.25529039e-01
-1.44292980e-01 -2.89363414e-01 4.69963765e-03 2.73191929e-01
5.79159141e-01 3.25714290e-01 -4.13104385e-01 5.05699694e-01
-9.83809590e-01 5.38905784e-02 6.91108704e-01 1.04211307e+00
1.04084086e+00 -5.07542007e-02 -3.60842422e-02 9.70665395e-01
2.21372947e-01 4.13903832e-01 5.68489313e-01 -6.48851573e-01
1.76753283e-01 6.86989605e-01 4.69922274e-02 -8.37234020e-01
2.21676528e-01 -5.86118519e-01 -5.46605706e-01 4.83968794e-01
4.64927047e-01 2.22596899e-01 -1.02197134e+00 2.07981968e+00
5.06407797e-01 4.56449836e-01 5.00060320e-01 9.24380362e-01
5.46913087e-01 7.81585336e-01 -7.43466169e-02 -1.76316947e-01
1.08330202e+00 -7.37800360e-01 -5.67187428e-01 -1.72267929e-01
5.15068054e-01 -4.50862169e-01 9.87665057e-01 3.04976583e-01
-6.26770496e-01 -7.08912432e-01 -1.48567641e+00 1.84044302e-01
-6.78445578e-01 1.67467624e-01 -2.89476942e-02 6.89919174e-01
-7.89220870e-01 6.43774867e-01 -3.49024624e-01 -4.73801196e-01
2.62638509e-01 2.68604159e-01 -5.29209793e-01 -2.85544284e-02
-1.23442233e+00 9.51959789e-01 4.23566967e-01 -2.97404885e-01
-9.51695561e-01 -4.38515544e-01 -9.00061667e-01 1.45050272e-01
4.92823094e-01 -3.69588912e-01 8.30459177e-01 -1.42727494e+00
-1.23141325e+00 1.09231520e+00 -1.05218910e-01 -3.61298054e-01
5.68281233e-01 1.63336918e-01 -4.91196692e-01 2.06304803e-01
2.54940569e-01 5.96248150e-01 1.31078613e+00 -1.63065100e+00
-6.10997617e-01 -4.60697472e-01 -3.99531126e-02 7.04886019e-02
-4.02288318e-01 -5.84839761e-01 -9.96195823e-02 -4.25467849e-01
2.45131299e-01 -8.76457810e-01 8.68200660e-02 1.01886287e-01
-2.70234108e-01 -2.85412431e-01 9.69733417e-01 -1.66708678e-01
7.78030038e-01 -2.44753194e+00 8.95990655e-02 2.64073223e-01
9.25551951e-02 2.35451564e-01 -3.18592158e-03 3.24738652e-01
-3.53096813e-01 -1.21533990e-01 -3.88296366e-01 -3.70765001e-01
-2.27302462e-02 4.34721172e-01 -5.59696317e-01 5.34099519e-01
4.28364694e-01 3.08464497e-01 -8.15862298e-01 -6.40397370e-01
2.91686773e-01 4.18043882e-01 -4.27080512e-01 5.30825794e-01
-2.53427386e-01 4.85582829e-01 -2.73461699e-01 2.08320811e-01
7.31909215e-01 -9.65289250e-02 1.42443091e-01 -2.87993640e-01
1.78926691e-01 -1.27961218e-01 -1.41483331e+00 1.30602956e+00
-4.26502317e-01 2.39022121e-01 -1.56816065e-01 -1.32301283e+00
1.27730775e+00 5.07532001e-01 2.78681397e-01 -3.25346738e-01
7.48202354e-02 4.60409135e-01 -1.06362730e-01 -5.15688658e-01
6.84283674e-02 -5.13593256e-01 1.39453575e-01 2.60818809e-01
5.60504079e-01 -6.31447881e-02 -2.75407523e-01 6.82362542e-03
6.43303096e-01 3.05391610e-01 4.30492431e-01 -2.44851753e-01
7.55475402e-01 -1.36925668e-01 8.53492498e-01 6.49816930e-01
-3.13141257e-01 7.47590184e-01 4.56592351e-01 -2.42910594e-01
-1.28318977e+00 -1.18052971e+00 -3.36285591e-01 7.82720864e-01
4.38149512e-01 5.56896180e-02 -7.87440717e-01 -8.58460546e-01
-7.70444870e-02 8.73555720e-01 -7.88202822e-01 -4.87520456e-01
-1.88668758e-01 -3.37089926e-01 2.32705906e-01 2.44286865e-01
4.72473562e-01 -8.98364663e-01 -5.48357427e-01 3.61114033e-02
-3.06011401e-02 -9.90634680e-01 2.53627095e-02 4.05948728e-01
-5.50250232e-01 -1.17449868e+00 -7.97285318e-01 -9.70785618e-01
8.07068288e-01 -1.22896585e-04 8.63457501e-01 -1.97553784e-01
-1.35522544e-01 3.91738206e-01 -3.49276900e-01 -2.44631439e-01
-4.56670165e-01 -2.86405087e-01 7.96799734e-02 7.20268250e-01
4.69638824e-01 -5.16492426e-01 -5.16636670e-01 4.57843393e-01
-9.64911580e-01 -2.27018576e-02 4.54391241e-01 1.08611643e+00
5.70430517e-01 2.23747104e-01 8.66370380e-01 -9.14855301e-01
1.48631066e-01 -5.99668503e-01 -4.84193414e-01 3.17208588e-01
-7.37846434e-01 2.79273808e-01 7.16440499e-01 -6.73597097e-01
-1.10270166e+00 1.68928087e-01 6.89842552e-02 -6.55542135e-01
-4.81401563e-01 -2.32962123e-03 -6.42552435e-01 1.43740520e-01
6.29778206e-01 4.37759429e-01 1.75551862e-01 -3.16721559e-01
1.45949095e-01 9.21067357e-01 4.47453082e-01 -5.53700745e-01
7.58937776e-01 5.67380786e-01 7.02872872e-02 -7.24653482e-01
-8.07257354e-01 -3.02039713e-01 -5.83944678e-01 -2.32704461e-01
1.01022351e+00 -7.17140615e-01 -1.73059106e-01 1.02108419e-01
-9.74165499e-01 -5.18248007e-02 -5.99304497e-01 5.43749273e-01
-8.99142087e-01 6.26442805e-02 -1.23368822e-01 -9.93201733e-01
1.70499757e-01 -1.34084451e+00 9.42710459e-01 -6.44400890e-04
-6.24358691e-02 -8.23342562e-01 4.69139554e-02 1.73674569e-01
1.23358928e-01 2.67654687e-01 1.14693832e+00 -1.22754133e+00
-6.01213038e-01 -3.06729525e-01 -6.82434663e-02 7.97170937e-01
-6.49430379e-02 -2.20130518e-01 -1.34164894e+00 -2.42880329e-01
4.48509425e-01 -3.68900985e-01 7.39760339e-01 1.68408766e-01
1.05113518e+00 5.53261489e-02 -3.50255400e-01 4.57287997e-01
1.70043230e+00 4.77925390e-01 4.92574930e-01 2.19148293e-01
3.57110232e-01 9.84888196e-01 8.00939143e-01 1.76660791e-01
-2.31910031e-02 7.74769783e-01 5.10912418e-01 8.85625854e-02
-1.10562325e-01 -3.39937538e-01 1.98980689e-01 4.01554495e-01
3.39529753e-01 -4.22470182e-01 -5.92869043e-01 7.35708535e-01
-1.80552018e+00 -8.73764396e-01 1.76188901e-01 2.58986044e+00
5.71536660e-01 2.65637096e-02 -1.41055971e-01 3.38032156e-01
9.86570418e-01 -8.25248063e-02 -5.72375000e-01 -2.98497707e-01
-1.98672533e-01 1.90540120e-01 1.73219934e-01 5.25427520e-01
-1.14361691e+00 6.41300559e-01 5.45619822e+00 1.03744006e+00
-1.06042969e+00 1.45138726e-01 5.52755833e-01 1.41499966e-01
-3.60729814e-01 8.20681527e-02 -8.73206258e-01 4.83141541e-01
7.41998255e-01 -1.87701970e-01 2.06458673e-01 1.07515287e+00
-3.19963783e-01 -9.24692079e-02 -1.46504331e+00 9.22992170e-01
3.98872793e-01 -8.65268648e-01 9.10508707e-02 2.43944988e-01
6.22174263e-01 -5.75552762e-01 3.66133839e-01 4.81122017e-01
-1.11740448e-01 -8.58089745e-01 6.15773618e-01 3.95309448e-01
8.13248932e-01 -6.97731912e-01 8.03426743e-01 6.29650295e-01
-8.91875207e-01 -7.73032233e-02 -5.74889183e-01 5.13884285e-03
-2.66914278e-01 4.24888849e-01 -9.01467800e-01 4.48290825e-01
4.44876790e-01 5.65478325e-01 -3.56076002e-01 6.23120546e-01
-2.41274327e-01 3.85823935e-01 3.22963446e-02 2.14020535e-01
1.40900731e-01 -3.43959361e-01 6.99017346e-01 7.34120488e-01
5.98156333e-01 1.40419258e-02 3.35005932e-02 1.20118630e+00
1.67543799e-01 2.98354160e-02 -1.18127429e+00 2.75886089e-01
4.02485430e-01 1.01035976e+00 -5.61835766e-01 -6.50122225e-01
-4.45367247e-01 1.05514491e+00 1.90730840e-01 4.61213261e-01
-8.29859316e-01 -2.57851094e-01 5.06822109e-01 1.09674223e-01
4.69405502e-01 4.37799841e-01 -7.30150416e-02 -1.15883577e+00
1.61577798e-02 -6.97119892e-01 2.53169268e-01 -8.38218749e-01
-1.41835606e+00 9.11295295e-01 1.43073633e-01 -1.71293318e+00
-4.18152779e-01 -6.55640960e-01 -3.12757075e-01 1.04115462e+00
-1.37249255e+00 -1.14030814e+00 -3.14531058e-01 7.46560037e-01
5.09802759e-01 -4.43963438e-01 1.04586470e+00 2.40853146e-01
-2.94150501e-01 5.38734913e-01 1.71089903e-01 -7.39747435e-02
9.71320033e-01 -1.37477767e+00 -3.17122489e-01 6.38725996e-01
1.26466125e-01 2.92000085e-01 9.15301561e-01 -5.11591256e-01
-7.67569244e-01 -1.12306166e+00 9.09253776e-01 -4.67307121e-01
3.53392422e-01 -5.91664910e-01 -1.05294085e+00 6.38666868e-01
1.35070272e-02 2.90336311e-01 7.22001374e-01 -2.27198675e-01
-4.05883402e-01 -2.67887235e-01 -1.49283946e+00 3.25812280e-01
7.74846673e-01 -7.37420797e-01 -8.45633864e-01 1.81677014e-01
4.89528984e-01 1.32878318e-01 -7.03813672e-01 3.59215617e-01
3.22943509e-01 -1.05318785e+00 8.33312809e-01 -4.22213286e-01
5.46183944e-01 -4.07833010e-01 -5.70604444e-01 -1.45874846e+00
-8.07889998e-02 2.04408288e-01 1.34419367e-01 1.36395133e+00
2.54493475e-01 -7.07714140e-01 9.50659215e-01 5.38580775e-01
4.16501127e-02 -7.57770419e-01 -1.00370204e+00 -1.14126420e+00
6.76412582e-02 -1.71931997e-01 5.91927230e-01 1.06978428e+00
-2.46403575e-01 3.59648019e-01 -3.46518844e-01 3.00762564e-01
8.59255791e-01 1.81435823e-01 6.47422791e-01 -1.38612580e+00
-3.60787183e-01 5.21024987e-02 -7.25102007e-01 -6.88203394e-01
4.40018445e-01 -9.26364005e-01 1.33559808e-01 -1.06376803e+00
1.50741994e-01 -5.09853661e-01 -3.60790968e-01 1.74317092e-01
9.03944299e-02 1.72162801e-01 1.57249451e-01 2.48098895e-01
-2.35408530e-01 6.68108404e-01 1.12928593e+00 -2.29757935e-01
3.64932716e-02 -5.03706448e-02 -2.95142800e-01 6.29378080e-01
5.49056768e-01 -5.38853407e-01 -8.05546880e-01 3.61283422e-02
-5.16163558e-02 1.80886403e-01 5.49616456e-01 -1.10090816e+00
7.72245824e-02 -9.54114795e-02 3.99877787e-01 -3.93764406e-01
7.04383850e-01 -1.30310333e+00 1.80632100e-01 1.92471698e-01
-4.24958616e-01 -6.34379983e-01 -3.05973768e-01 7.48906136e-01
-3.59628618e-01 -6.21870697e-01 9.82854307e-01 -3.46180424e-02
-7.67422855e-01 1.40504941e-01 -1.32564202e-01 -1.74978837e-01
1.45216572e+00 -6.60923660e-01 -1.00769795e-01 -4.31086034e-01
-1.09745133e+00 -1.25215545e-01 5.86001992e-01 2.77631700e-01
6.68996036e-01 -1.41926825e+00 -3.94492865e-01 6.50618851e-01
4.66162533e-01 -8.60011727e-02 1.39911458e-01 5.59136927e-01
-1.70859583e-02 1.90740645e-01 -3.28807026e-01 -6.42612576e-01
-1.06874478e+00 9.02781129e-01 3.63442391e-01 -6.87038302e-02
-4.83486146e-01 7.27490187e-01 6.04639769e-01 -4.67418700e-01
1.26615733e-01 -7.36730453e-03 -2.93068856e-01 6.40067682e-02
3.20945621e-01 1.00889638e-01 5.11662774e-02 -7.41780877e-01
-2.62289941e-01 5.23938477e-01 7.80836642e-02 -2.39092663e-01
1.13797832e+00 -7.12082237e-02 1.33171231e-01 7.64895439e-01
1.49931514e+00 -2.19990864e-01 -1.44440329e+00 -3.25305402e-01
-1.27012193e-01 -5.00079274e-01 4.34356891e-02 -6.32609129e-01
-8.07676315e-01 9.97791529e-01 8.06730330e-01 2.54318476e-01
1.07767558e+00 1.43961459e-01 3.52633268e-01 3.61328274e-02
4.31151599e-01 -1.19105625e+00 1.76358029e-01 5.08395918e-02
7.24584222e-01 -1.42541015e+00 -3.48213583e-01 -6.08173549e-01
-7.16422260e-01 9.82753992e-01 7.77817667e-01 -4.29677516e-01
6.26864016e-01 -1.83389813e-01 4.52112686e-03 -7.57279396e-02
-6.85633659e-01 -2.43637353e-01 3.62588942e-01 7.86452591e-01
-6.80948887e-03 4.94721159e-02 -9.46384445e-02 6.41986609e-01
8.53135437e-02 -2.32304752e-01 3.33265990e-01 7.37793148e-01
-5.22785723e-01 -1.06799316e+00 -4.80553061e-01 3.09162945e-01
-3.93390432e-02 1.80098131e-01 -2.61013180e-01 7.82963097e-01
6.11875772e-01 8.47680211e-01 3.22108895e-01 -4.35761303e-01
1.43640682e-01 4.82392192e-01 4.80817914e-01 -7.00986862e-01
-1.66970696e-02 8.25364590e-02 -2.07106203e-01 -3.56106102e-01
-3.35171580e-01 -4.57746148e-01 -9.79004264e-01 2.78623819e-01
-3.86868477e-01 2.70314366e-01 5.30816734e-01 9.83228743e-01
1.88482329e-01 3.92097861e-01 7.73753226e-01 -6.10503316e-01
-8.06754053e-01 -7.71432161e-01 -9.93764877e-01 8.55724275e-01
4.55631971e-01 -1.07577980e+00 -8.27918410e-01 -2.10689045e-02] | [9.921894073486328, 2.7798068523406982] |
00276683-bea0-482b-8c4c-276e5d625c0b | s3cnet-a-sparse-semantic-scene-completion | 2012.09242 | null | https://arxiv.org/abs/2012.09242v1 | https://arxiv.org/pdf/2012.09242v1.pdf | S3CNet: A Sparse Semantic Scene Completion Network for LiDAR Point Clouds | With the increasing reliance of self-driving and similar robotic systems on robust 3D vision, the processing of LiDAR scans with deep convolutional neural networks has become a trend in academia and industry alike. Prior attempts on the challenging Semantic Scene Completion task - which entails the inference of dense 3D structure and associated semantic labels from "sparse" representations - have been, to a degree, successful in small indoor scenes when provided with dense point clouds or dense depth maps often fused with semantic segmentation maps from RGB images. However, the performance of these systems drop drastically when applied to large outdoor scenes characterized by dynamic and exponentially sparser conditions. Likewise, processing of the entire sparse volume becomes infeasible due to memory limitations and workarounds introduce computational inefficiency as practitioners are forced to divide the overall volume into multiple equal segments and infer on each individually, rendering real-time performance impossible. In this work, we formulate a method that subsumes the sparsity of large-scale environments and present S3CNet, a sparse convolution based neural network that predicts the semantically completed scene from a single, unified LiDAR point cloud. We show that our proposed method outperforms all counterparts on the 3D task, achieving state-of-the art results on the SemanticKITTI benchmark. Furthermore, we propose a 2D variant of S3CNet with a multi-view fusion strategy to complement our 3D network, providing robustness to occlusions and extreme sparsity in distant regions. We conduct experiments for the 2D semantic scene completion task and compare the results of our sparse 2D network against several leading LiDAR segmentation models adapted for bird's eye view segmentation on two open-source datasets. | ['Liu Bingbing', 'Xinhai Li', 'Yuan Ren', 'Christopher Agia', 'Ran Cheng'] | 2020-12-16 | null | null | null | null | ['3d-semantic-scene-completion'] | ['computer-vision'] | [ 3.80513966e-01 7.96387941e-02 1.07715711e-01 -6.47608936e-01
-5.99338531e-01 -7.02314019e-01 4.87851918e-01 -1.07897101e-02
-4.24462527e-01 3.01272869e-01 -1.61689833e-01 -1.98764086e-01
-1.00337014e-01 -8.27685893e-01 -9.30252731e-01 -1.64310843e-01
1.53153002e-01 9.75221574e-01 3.93940240e-01 -1.25062570e-01
1.28656358e-01 9.19548750e-01 -2.00539589e+00 -1.05782375e-01
7.40823448e-01 1.25675547e+00 4.66468573e-01 3.86848152e-01
-5.13326943e-01 2.44387403e-01 -7.14390054e-02 -2.90337294e-01
8.62898767e-01 3.97633404e-01 -7.54532874e-01 4.09784526e-01
1.11518514e+00 -3.76321822e-01 -2.99970835e-01 9.98013973e-01
2.91529417e-01 1.63425714e-01 3.82383227e-01 -1.23703980e+00
-1.81308702e-01 3.36282738e-02 -6.54857278e-01 -2.35678077e-01
8.57358575e-02 1.02852724e-01 7.20884860e-01 -1.09532654e+00
6.73874199e-01 1.33519876e+00 1.00361037e+00 1.66570261e-01
-1.45008790e+00 -5.00333011e-01 3.03206861e-01 -1.35342747e-01
-1.47975302e+00 -3.61586750e-01 7.95711815e-01 -4.00589556e-01
1.19987249e+00 2.58727130e-02 5.22459090e-01 7.51742065e-01
-2.70098835e-01 5.46409547e-01 9.49870884e-01 1.26023442e-01
4.29283738e-01 -1.44119918e-01 2.85059903e-02 7.30212212e-01
3.88100237e-01 4.10457961e-02 -3.95444691e-01 1.69665873e-01
7.00599134e-01 3.35721761e-01 1.73540354e-01 -8.49939704e-01
-1.07471871e+00 8.50685179e-01 8.47057521e-01 -2.03932330e-01
-3.93451959e-01 2.82919645e-01 2.64135540e-01 2.72088479e-02
6.93030179e-01 8.80037844e-02 -6.14936531e-01 1.79400474e-01
-1.27958941e+00 4.15298194e-01 8.67601454e-01 1.19894516e+00
1.27580452e+00 -1.21836765e-02 5.01166344e-01 6.88439250e-01
1.87136412e-01 6.73658490e-01 -2.30486184e-01 -1.49460721e+00
5.56920886e-01 6.69303775e-01 -1.59402899e-02 -7.21527278e-01
-5.80429435e-01 -4.74268228e-01 -8.63725126e-01 4.29176658e-01
3.32433015e-01 1.86505780e-01 -1.30762410e+00 1.43045616e+00
5.32701612e-01 2.66484112e-01 3.51401828e-02 9.64157224e-01
8.10411215e-01 3.00452948e-01 -2.62480415e-02 3.67862254e-01
1.09165919e+00 -8.50620151e-01 -6.57280609e-02 -5.63090265e-01
2.92035282e-01 -6.89876556e-01 8.47315311e-01 3.49414021e-01
-9.37397361e-01 -7.04832852e-01 -1.00998640e+00 -6.14938855e-01
-4.72274870e-01 -1.47954091e-01 8.55515540e-01 4.28899080e-01
-1.26651061e+00 6.81909919e-01 -9.41185772e-01 -5.65876007e-01
8.88667881e-01 5.57689488e-01 -5.46908677e-01 -6.51107609e-01
-4.24107403e-01 8.26107979e-01 3.15464586e-01 2.11478263e-01
-8.87990236e-01 -9.99187112e-01 -1.12876737e+00 -7.38392845e-02
4.90359694e-01 -9.28705394e-01 1.09046865e+00 -5.15695870e-01
-1.00023270e+00 1.11117542e+00 -2.18723416e-01 -6.00209117e-01
5.57332337e-01 -4.48618710e-01 1.94343984e-01 7.45033026e-02
4.35810149e-01 1.28394091e+00 7.56214976e-01 -1.46902359e+00
-5.99823534e-01 -7.17406988e-01 1.81688130e-01 3.71952116e-01
2.91305989e-01 -5.68752885e-01 -5.41455090e-01 -1.57609388e-01
6.26600504e-01 -9.41565871e-01 -6.95895553e-01 3.76725256e-01
-4.67158794e-01 -3.78549285e-02 1.15693927e+00 -2.24691674e-01
1.65346727e-01 -2.17376471e+00 1.01851702e-01 8.71451274e-02
2.87391931e-01 4.45125364e-02 -1.94196641e-01 5.88812642e-02
1.00802682e-01 -4.26299646e-02 -7.18362808e-01 -9.97858584e-01
4.88564819e-02 8.17317963e-01 -3.46925944e-01 5.13031542e-01
3.95232737e-01 8.14137399e-01 -7.06729949e-01 -4.27964658e-01
6.94424152e-01 6.66566551e-01 -8.03098083e-01 1.93981454e-01
-3.91030967e-01 4.02521640e-01 -2.91122288e-01 9.59235013e-01
1.08775842e+00 -2.78306156e-01 -1.90893605e-01 -2.68458962e-01
-2.82537967e-01 1.67683363e-01 -1.22035694e+00 2.57717943e+00
-6.05552614e-01 4.46986765e-01 4.97095466e-01 -1.10715485e+00
9.15883005e-01 -1.91889510e-01 6.51324213e-01 -4.89472836e-01
3.65025699e-02 3.03647548e-01 -5.38537323e-01 -2.24992186e-01
6.28976882e-01 -2.98674524e-01 -9.98867825e-02 4.08372432e-02
1.71192408e-01 -1.05185080e+00 2.14259028e-02 1.58499256e-01
1.04102433e+00 6.03483856e-01 -1.16913132e-01 -8.49370658e-02
2.84575433e-01 4.03903306e-01 3.64438057e-01 6.38632178e-01
-4.84763421e-02 9.39991355e-01 1.41938016e-01 -6.06202304e-01
-1.11888587e+00 -1.41265631e+00 -3.06901217e-01 5.29997110e-01
4.40682620e-01 -1.36868551e-01 -4.35790509e-01 -5.19446194e-01
4.78635281e-01 5.95845342e-01 -2.89386153e-01 2.51503944e-01
-5.92559814e-01 -3.35776150e-01 3.19604486e-01 5.18618703e-01
6.18554056e-01 -7.58469880e-01 -8.25526774e-01 2.13346556e-01
1.75712556e-02 -1.82753134e+00 1.50070071e-01 6.28794432e-01
-9.89879012e-01 -1.06856632e+00 -3.60896349e-01 -6.39923275e-01
4.60015088e-01 6.46067977e-01 1.28359723e+00 -1.45933479e-01
-4.84221309e-01 5.65385342e-01 -2.03213152e-02 -4.35333401e-01
1.29509658e-01 2.01760173e-01 1.05559770e-02 -3.36447865e-01
3.51515621e-01 -1.06009662e+00 -6.12443864e-01 1.04718365e-01
-9.64566410e-01 1.46430224e-01 4.52327192e-01 2.96992481e-01
1.00981426e+00 -1.15489289e-01 1.56730935e-01 -8.34653378e-01
-1.42314240e-01 -5.68282068e-01 -9.09444332e-01 -3.48965973e-01
-4.83227998e-01 6.63323328e-02 3.20728511e-01 8.14898834e-02
-8.59272361e-01 6.41002059e-01 -3.87582034e-01 -9.55790401e-01
-6.11971915e-01 1.83332607e-01 -1.60367012e-01 -3.01958948e-01
4.22885507e-01 2.50539631e-02 7.99866095e-02 -6.77540362e-01
7.12882638e-01 3.17759931e-01 8.89108896e-01 -6.03625894e-01
1.10050654e+00 9.72236753e-01 5.16298115e-01 -8.86156499e-01
-1.20052981e+00 -7.00125456e-01 -1.02452147e+00 1.05502531e-01
9.57367480e-01 -1.34823501e+00 -4.01684940e-01 3.01493585e-01
-1.28868091e+00 -3.20908040e-01 -6.25603497e-01 2.53662527e-01
-7.40866184e-01 4.30162609e-01 -1.87235445e-01 -5.37488401e-01
-8.98030698e-02 -1.27942479e+00 1.69769216e+00 5.58059216e-02
7.98278600e-02 -6.40235841e-01 -1.18602380e-01 6.12532139e-01
2.99695343e-01 4.16033477e-01 7.10558951e-01 -3.60535532e-01
-1.05710888e+00 -3.21266018e-02 -5.93167484e-01 5.14202297e-01
-1.03843331e-01 -2.94278651e-01 -1.29400277e+00 3.48432362e-02
-1.98448822e-02 -4.55760986e-01 9.80815709e-01 3.52432847e-01
1.31785321e+00 2.19738305e-01 -2.53755569e-01 1.18136036e+00
1.71452558e+00 -4.13448066e-01 3.02812755e-01 1.02957949e-01
1.07283580e+00 8.29660296e-01 4.95345592e-01 2.99190670e-01
7.08876371e-01 4.24824744e-01 1.14891982e+00 -3.14462006e-01
-3.08392763e-01 -2.33860925e-01 -5.59329689e-02 4.32406664e-01
7.57704526e-02 -7.93979093e-02 -9.41508889e-01 5.43262661e-01
-1.65298605e+00 -5.07241547e-01 -2.83656299e-01 2.05503631e+00
2.81834930e-01 1.81849271e-01 -2.68194318e-01 6.08463585e-02
2.66432732e-01 2.36615136e-01 -7.86743820e-01 -1.86357960e-01
-1.50512263e-01 6.66937828e-01 7.85012901e-01 5.52729666e-01
-1.20385265e+00 1.13417029e+00 5.62285233e+00 5.32731533e-01
-1.02916372e+00 1.53249100e-01 4.38721895e-01 -3.59268337e-01
-3.49797219e-01 1.34913340e-01 -8.49202096e-01 3.34031694e-02
7.60292351e-01 3.42843443e-01 3.99672806e-01 1.00890279e+00
8.67246315e-02 -3.75856340e-01 -1.22061694e+00 1.31595409e+00
-2.69461107e-02 -1.42828143e+00 -4.31418829e-02 2.13716224e-01
7.41209984e-01 8.96721661e-01 -6.60182582e-03 1.70432553e-01
4.42140698e-01 -1.08154285e+00 7.92772532e-01 2.15658173e-01
9.65717316e-01 -6.12803042e-01 4.39249337e-01 4.92368251e-01
-1.30901134e+00 -8.89143199e-02 -5.23064435e-01 -1.45150736e-01
3.36470068e-01 8.80866587e-01 -6.90712690e-01 6.76713765e-01
9.12612438e-01 8.50169241e-01 -5.29116273e-01 9.33210731e-01
4.40093614e-02 9.43740606e-02 -8.95280659e-01 5.67227185e-01
5.30914426e-01 -3.62870663e-01 4.80512530e-01 8.91636312e-01
3.98115963e-01 -3.66767496e-02 4.60498661e-01 1.09843063e+00
-2.14545041e-01 -2.96128064e-01 -9.99102950e-01 2.99481809e-01
3.37300837e-01 1.41142499e+00 -1.02305520e+00 -2.18345255e-01
-4.39782798e-01 9.16020989e-01 3.18744183e-01 2.84983754e-01
-5.35436153e-01 2.40483843e-02 9.87862945e-01 2.46683255e-01
6.13230407e-01 -7.09866047e-01 -7.83782005e-01 -1.05850434e+00
4.27818745e-02 -2.27395371e-01 -3.18621658e-02 -8.82966220e-01
-1.19252276e+00 5.09165168e-01 9.22292843e-02 -1.05566418e+00
-1.35840416e-01 -6.00225449e-01 -2.05191687e-01 8.72513294e-01
-1.99459612e+00 -1.33147550e+00 -6.62429035e-01 7.70235300e-01
7.69513190e-01 2.77395666e-01 6.20236516e-01 4.75712359e-01
-1.42331719e-01 -7.01731965e-02 -3.17978919e-01 -2.00037405e-01
2.89662391e-01 -1.15606713e+00 6.60789847e-01 7.59627283e-01
2.04554379e-01 1.57995790e-01 4.59109604e-01 -6.16183460e-01
-1.47603810e+00 -1.45895576e+00 6.96073949e-01 -6.41258240e-01
4.16652322e-01 -6.52280569e-01 -7.37514794e-01 5.92619181e-01
-3.10957581e-01 4.44421262e-01 2.69665122e-01 1.57445706e-02
-4.33417022e-01 -1.52999073e-01 -1.27678823e+00 1.98415443e-01
1.63574898e+00 -6.27551019e-01 -5.04038095e-01 4.33598518e-01
1.04070914e+00 -5.99171102e-01 -7.43695319e-01 7.55160987e-01
2.67006189e-01 -1.16027427e+00 1.35318148e+00 -1.41176194e-01
4.29593116e-01 -5.01311302e-01 -7.10242689e-01 -8.20128381e-01
1.10700063e-01 -1.17328815e-01 1.06997289e-01 8.35840404e-01
-5.97759932e-02 -3.75026345e-01 1.08728254e+00 4.46985424e-01
-6.73067629e-01 -4.48270082e-01 -1.19768727e+00 -5.88708460e-01
-1.53182745e-01 -9.39938188e-01 5.97968519e-01 6.37514174e-01
-9.91039872e-01 3.42625946e-01 5.05477898e-02 4.13088858e-01
9.56622303e-01 4.11654949e-01 1.05054355e+00 -1.68059123e+00
1.31048590e-01 -2.41687939e-01 -5.92934608e-01 -1.25355422e+00
3.73028934e-01 -1.01681828e+00 2.09696084e-01 -1.59729600e+00
-2.04779163e-01 -7.19916165e-01 1.48686364e-01 4.90384370e-01
3.18901122e-01 6.12178266e-01 2.34572113e-01 2.24226102e-01
-4.83986378e-01 6.35235965e-01 9.97438014e-01 -2.22408965e-01
-9.62907374e-02 -9.51744765e-02 -4.75838810e-01 9.15911019e-01
4.01671261e-01 -3.68566662e-01 -5.30486524e-01 -7.82051206e-01
1.83779180e-01 -8.24905187e-02 8.21202338e-01 -1.28588891e+00
8.80803168e-02 2.01705508e-02 3.69123966e-01 -1.17609048e+00
8.54982495e-01 -1.21388471e+00 1.84845507e-01 2.03526303e-01
1.96354821e-01 -5.22211529e-02 2.41764411e-01 6.33723557e-01
-1.46848083e-01 8.58338699e-02 6.80497766e-01 -3.91958624e-01
-9.96925652e-01 7.71203518e-01 1.20794468e-01 -8.70004445e-02
8.17383945e-01 -6.26359165e-01 6.20996021e-02 1.28746210e-02
-6.88598990e-01 3.53466183e-01 7.55244493e-01 4.77842063e-01
7.32131243e-01 -1.02332103e+00 -4.44623530e-01 4.55128431e-01
5.81126586e-02 1.19248760e+00 3.62136602e-01 6.77857101e-01
-7.98085511e-01 3.60091180e-01 -1.22084841e-01 -1.16767883e+00
-8.91506791e-01 4.28715467e-01 2.52730340e-01 -1.76130272e-02
-9.98237014e-01 9.02362347e-01 4.20284986e-01 -8.31130624e-01
2.56671458e-01 -5.99315941e-01 2.22466439e-01 -8.44672099e-02
-1.01668701e-01 1.33277655e-01 3.59080881e-01 -7.42423475e-01
-3.48421186e-01 8.42186034e-01 2.63149917e-01 1.13918856e-01
1.56819344e+00 -2.52876878e-01 -1.47672623e-01 2.95850545e-01
1.22595251e+00 -3.64709556e-01 -1.53289282e+00 -3.74294996e-01
-1.09371312e-01 -4.93275762e-01 2.36907586e-01 -4.22371417e-01
-1.14490879e+00 1.13491619e+00 6.08613431e-01 -1.02981687e-01
9.13536668e-01 1.87692612e-01 9.12603557e-01 5.63229620e-01
6.01711690e-01 -8.54765534e-01 -2.18685061e-01 7.81588078e-01
5.97899795e-01 -1.49581599e+00 1.61136299e-01 -6.47339344e-01
-2.41918117e-01 9.28131461e-01 5.64850509e-01 -3.95999432e-01
7.95939505e-01 2.97492296e-01 -6.72045201e-02 -5.26369989e-01
-3.18589598e-01 -4.81359035e-01 4.65722531e-02 8.03309381e-01
-2.01546982e-01 4.56570238e-02 4.50260609e-01 1.20579451e-01
-4.57374096e-01 -2.04243027e-02 2.03832269e-01 9.12415266e-01
-4.61091012e-01 -8.98539066e-01 -3.84195596e-01 3.98100972e-01
1.38581591e-03 -5.08207940e-02 -1.67497501e-01 8.13346565e-01
5.32966256e-01 7.48403072e-01 5.45762897e-01 -2.82389879e-01
4.49950427e-01 3.54595259e-02 4.76549923e-01 -8.74953866e-01
-3.11672360e-01 9.58432630e-03 -4.60918657e-02 -1.05574656e+00
-6.45548940e-01 -7.50187337e-01 -1.34739673e+00 -3.03728759e-01
7.00940713e-02 -3.79636317e-01 1.08989942e+00 1.09573686e+00
4.45500731e-01 2.81000972e-01 4.53994006e-01 -1.59624982e+00
-4.41530526e-01 -6.44679844e-01 -5.00840724e-01 2.72043049e-01
4.19524878e-01 -8.62153769e-01 -1.78175777e-01 -1.91902280e-01] | [8.225261688232422, -2.846393346786499] |
10b78fc9-759d-4548-81a4-1f5d44b8cd3f | utilising-knowledge-graph-embeddings-for-data | null | null | https://aclanthology.org/2020.webnlg-1.6 | https://aclanthology.org/2020.webnlg-1.6.pdf | Utilising Knowledge Graph Embeddings for Data-to-Text Generation | Data-to-text generation has recently seen a move away from modular and pipeline architectures towards end-to-end architectures based on neural networks. In this work, we employ knowledge graph embeddings and explore their utility for end-to-end approaches in a data-to-text generation task. Our experiments show that using knowledge graph embeddings can yield an improvement of up to 2 – 3 BLEU points for seen categories on the WebNLG corpus without modifying the underlying neural network architecture. | ['Paul Buitelaar', 'Mihael Arcan', 'Nivranshu Pasricha'] | null | null | null | null | acl-webnlg-inlg-2020-12 | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings', 'data-to-text-generation'] | ['graphs', 'methodology', 'natural-language-processing'] | [ 1.72868118e-01 9.10329044e-01 -2.92132460e-02 -4.44588721e-01
-9.41503346e-01 -6.15536273e-01 1.21896791e+00 3.49144220e-01
-2.96771854e-01 6.75522625e-01 8.48153472e-01 -4.00092989e-01
8.85825884e-03 -1.00169563e+00 -4.40517604e-01 7.31722265e-03
2.41133988e-01 1.07332838e+00 -2.53165532e-02 -4.96219873e-01
1.65167376e-01 3.08726896e-02 -1.16743302e+00 6.08815551e-01
6.17869258e-01 8.24909866e-01 -3.13622862e-01 1.12672365e+00
-4.61979568e-01 6.86797619e-01 -7.00830638e-01 -1.02751350e+00
3.20647098e-02 -4.45645332e-01 -1.18225753e+00 -3.44013125e-01
7.06981957e-01 -2.10490480e-01 -7.02552617e-01 5.47897875e-01
9.00279582e-01 3.99428517e-01 9.07424569e-01 -1.09276807e+00
-1.46575987e+00 1.19191599e+00 1.94992334e-01 8.64471570e-02
5.15651166e-01 2.65076697e-01 1.47143281e+00 -9.21495974e-01
9.10608590e-01 1.34971547e+00 5.72120070e-01 9.43242788e-01
-1.26365900e+00 -2.25368112e-01 -1.67079523e-01 6.84821606e-02
-9.29958940e-01 -5.22695541e-01 5.71044445e-01 -4.05439675e-01
1.74791098e+00 -9.23923701e-02 3.88276875e-01 1.35234952e+00
-8.56643950e-04 8.42297316e-01 5.86320877e-01 -6.14147723e-01
3.40690687e-02 -3.23627234e-01 1.08776055e-01 7.66068280e-01
5.24544477e-01 1.54923558e-01 -7.96833098e-01 1.38119999e-02
3.86291027e-01 -6.33790553e-01 -3.15052867e-02 -5.30295484e-02
-1.23820233e+00 1.26057053e+00 7.61815369e-01 2.59706259e-01
-2.05183432e-01 5.85747659e-01 6.44829392e-01 2.15266302e-01
7.54564524e-01 9.62289691e-01 -3.50011885e-01 -4.17213529e-01
-8.68632615e-01 5.56652129e-01 1.00245786e+00 1.05475712e+00
4.35542792e-01 3.15266341e-01 -8.78248453e-01 7.44745731e-01
3.46472740e-01 2.26108149e-01 8.00587058e-01 -4.22592998e-01
8.32600117e-01 7.17967927e-01 -3.12863663e-02 -5.66755474e-01
-3.87715191e-01 -1.70065805e-01 -6.35770559e-01 7.52637908e-03
3.82756621e-01 -5.40911436e-01 -1.19862616e+00 1.46024489e+00
-1.51753537e-02 -3.36972564e-01 3.71086031e-01 4.84690219e-01
1.17144537e+00 6.11051857e-01 2.08230063e-01 3.83622527e-01
1.21878493e+00 -1.23826015e+00 -7.59476185e-01 -3.88946056e-01
8.03971827e-01 -5.92143059e-01 1.00898433e+00 8.97455141e-02
-1.37999296e+00 -7.16360688e-01 -8.85961652e-01 -6.24104798e-01
-8.97318840e-01 -9.14725289e-02 6.24329686e-01 6.60410225e-01
-1.51200974e+00 7.62254059e-01 -6.40702844e-01 -5.77801585e-01
5.42134285e-01 8.43021199e-02 -2.59828538e-01 -6.57362863e-02
-1.33441126e+00 9.76835907e-01 9.79770780e-01 -1.73128843e-01
-8.13707530e-01 -7.28391886e-01 -9.14002180e-01 1.00177459e-01
-1.03931911e-02 -1.37295198e+00 1.58573616e+00 -4.22622830e-01
-1.61983871e+00 6.33045912e-01 -3.45132947e-02 -7.97032893e-01
4.57270384e-01 -2.52647370e-01 -4.91195679e-01 -1.57357067e-01
-1.46644682e-01 1.11452615e+00 6.12123251e-01 -6.47678256e-01
-6.14764571e-01 -6.64227977e-02 3.17144804e-02 3.21510732e-01
-5.48303187e-01 -1.50906950e-01 -2.04343125e-01 -7.49106765e-01
-5.25942862e-01 -7.26595998e-01 -9.46500301e-02 -3.41828674e-01
-5.48643351e-01 -7.94318020e-01 6.16035700e-01 -8.74182522e-01
1.09310842e+00 -1.51459670e+00 1.96613699e-01 -3.25925767e-01
8.64945799e-02 2.73930520e-01 -5.06721795e-01 9.25184011e-01
1.64266452e-01 3.63567919e-01 1.67470705e-02 -3.84533316e-01
4.45065588e-01 1.06303021e-02 -2.78647900e-01 -3.72874528e-01
5.07462323e-01 1.59817469e+00 -1.23184311e+00 -2.28964254e-01
4.97951061e-02 5.80852866e-01 -5.04948020e-01 2.89847761e-01
-6.26418293e-01 -1.88931674e-01 -1.17500514e-01 1.74053892e-01
3.35895307e-02 -8.32325742e-02 -2.84825303e-02 -2.94083301e-02
2.80543119e-01 7.79154360e-01 -4.38319355e-01 2.00257468e+00
-7.05941081e-01 8.80022585e-01 -5.57675004e-01 -4.28441823e-01
1.02861977e+00 4.84228283e-01 -2.57729471e-01 -5.35861611e-01
1.22748576e-02 1.44993261e-01 1.06904596e-01 -4.38157395e-02
1.02616918e+00 -1.81382686e-01 -3.18264037e-01 3.59690934e-01
6.91401601e-01 -4.29228723e-01 6.20667279e-01 4.29820776e-01
1.37450945e+00 1.58328533e-01 -3.68224550e-03 -1.30168900e-01
-1.26319798e-02 2.63409168e-01 -9.30335894e-02 7.43116498e-01
2.66240180e-01 6.32555544e-01 4.53873992e-01 -4.19138134e-01
-1.25528228e+00 -1.24642015e+00 3.92165601e-01 1.18777430e+00
-6.99817777e-01 -5.46716690e-01 -7.41585314e-01 -1.05965662e+00
1.37298465e-01 1.44946194e+00 -6.26591086e-01 -3.03349346e-01
-3.65862995e-01 -5.79534471e-01 9.74164546e-01 8.65860522e-01
2.30960265e-01 -1.29724705e+00 -1.39646590e-01 5.84288836e-01
-2.92155258e-02 -8.96321297e-01 -4.46369082e-01 1.05677828e-01
-9.72358942e-01 -6.81464136e-01 -7.86428034e-01 -8.15505803e-01
4.74785328e-01 -3.61040682e-01 1.65526855e+00 -4.99142349e-01
-1.86377019e-01 2.09132612e-01 -6.24454319e-01 -6.32479429e-01
-8.38726401e-01 6.66377366e-01 -3.28808218e-01 -4.66554731e-01
4.92679328e-01 -3.42703551e-01 -5.76578438e-01 -3.64134282e-01
-8.95670831e-01 2.95541435e-01 4.66222614e-01 8.33096147e-01
2.09117815e-01 -3.38519096e-01 9.16543424e-01 -8.30617726e-01
1.39804184e+00 -3.02912384e-01 -4.36046749e-01 1.03853919e-01
-9.89986300e-01 5.79351306e-01 7.28896558e-01 7.23754391e-02
-1.05912209e+00 -1.41499102e-01 -2.25778580e-01 -3.34631950e-02
-1.23742752e-01 7.02263832e-01 6.08851686e-02 4.24513429e-01
9.58608449e-01 1.12602204e-01 -2.92933553e-01 -3.97302955e-01
1.33907032e+00 8.52105677e-01 4.44382191e-01 -2.45877504e-01
7.00062454e-01 -8.51865560e-02 -2.93656945e-01 -3.99671316e-01
-9.95189011e-01 -1.23097099e-01 -5.35516441e-01 6.86653331e-02
1.04902279e+00 -1.05032551e+00 -1.94308758e-02 7.63492212e-02
-1.25038505e+00 -6.68441951e-01 -5.86892843e-01 9.00945589e-02
-6.75486624e-01 -5.37171960e-02 -6.62882805e-01 -2.36341089e-01
-1.03831971e+00 -5.91313601e-01 9.49505568e-01 3.26815099e-01
-4.94608551e-01 -1.42467904e+00 3.58570218e-01 5.68195939e-01
6.48936093e-01 7.71919712e-02 1.05512035e+00 -1.05757749e+00
-2.62874573e-01 -4.55130845e-01 -2.81965017e-01 3.19248497e-01
1.04592688e-01 -1.93752110e-01 -9.93914604e-01 -2.38436610e-01
-9.15477574e-01 -4.91023511e-01 9.56379771e-01 -1.02336556e-02
9.12972033e-01 -5.45475125e-01 -1.96614757e-01 3.63221824e-01
1.35565352e+00 -1.62833124e-01 6.80957973e-01 5.10419253e-04
1.07174194e+00 4.07615185e-01 1.02014504e-01 1.11502230e-01
5.87175548e-01 5.26319742e-01 2.17764542e-01 1.95414573e-01
-9.45644021e-01 -7.27260947e-01 4.32727158e-01 7.36389220e-01
1.26957193e-01 -8.80841434e-01 -1.03400993e+00 1.18223250e+00
-1.63399863e+00 -8.90684664e-01 -1.65700540e-01 1.90032601e+00
1.11938655e+00 9.66857970e-02 -5.34490161e-02 -2.91505367e-01
5.05255520e-01 2.53383845e-01 -2.86111802e-01 -1.16192400e+00
1.28623828e-01 5.73185146e-01 3.51249307e-01 4.33417201e-01
-7.87834227e-01 1.30282080e+00 7.17069387e+00 8.95240188e-01
-7.43638754e-01 1.78293630e-01 4.85200912e-01 -3.92018147e-02
-5.79548240e-01 -1.69274807e-02 -1.02386558e+00 3.66979957e-01
1.58964014e+00 -5.61506808e-01 3.99478137e-01 8.22565973e-01
-6.27587959e-02 4.52687353e-01 -1.47755969e+00 6.92976236e-01
1.50558069e-01 -1.64109027e+00 7.49292672e-01 -8.08615237e-02
1.00992525e+00 3.89392734e-01 -2.46795803e-01 6.65442646e-01
1.05424595e+00 -1.32893991e+00 4.41790640e-01 4.26340938e-01
8.81435573e-01 -7.63792932e-01 7.50299454e-01 1.05630113e-02
-8.75498593e-01 2.73222387e-01 -3.76102567e-01 -1.38282359e-01
4.66612846e-01 5.60477674e-01 -1.75064349e+00 7.20454335e-01
1.06815413e-01 3.65238667e-01 -8.53833854e-01 7.95371652e-01
-6.80293858e-01 6.30999923e-01 1.42069329e-02 -4.65106785e-01
5.04987955e-01 4.22654599e-01 2.84661621e-01 1.59567404e+00
5.12500942e-01 -5.90936899e-01 -2.21286967e-01 9.57038701e-01
-7.48267412e-01 1.40727647e-02 -9.55620170e-01 -6.87892258e-01
3.09400976e-01 1.35328531e+00 -3.42456549e-01 -6.25209928e-01
-2.39858493e-01 1.32091665e+00 6.58377826e-01 2.27369457e-01
-3.92937511e-01 -9.98445213e-01 3.98481458e-01 8.65011290e-02
6.38638437e-01 -2.07503498e-01 -2.91131854e-01 -8.32400560e-01
-2.79937238e-01 -5.73427439e-01 4.43326622e-01 -9.89955544e-01
-1.65603328e+00 6.30188227e-01 -6.38808534e-02 -6.66602850e-01
-9.92254615e-01 -7.38425910e-01 -8.51048648e-01 1.09939528e+00
-1.37516463e+00 -1.46659291e+00 -2.56489366e-01 2.28014827e-01
6.82910323e-01 -3.90959918e-01 1.19195616e+00 -1.71177566e-01
-8.09919983e-02 8.69203806e-01 3.15827876e-01 5.34927070e-01
5.04579365e-01 -1.80135250e+00 1.42427731e+00 8.44650626e-01
5.90746820e-01 2.45158672e-01 5.24385691e-01 -7.21847594e-01
-1.22907114e+00 -1.48970473e+00 1.49173188e+00 -8.08287263e-01
8.64373624e-01 -4.33088005e-01 -3.26948911e-01 7.25028276e-01
9.01036978e-01 -1.52593687e-01 6.54899538e-01 4.42489326e-01
-5.98622203e-01 2.35401005e-01 -1.03933942e+00 8.00209880e-01
1.25897789e+00 -4.81325090e-01 -8.58216882e-01 5.18058956e-01
1.04589522e+00 -3.85130942e-01 -1.05570924e+00 4.12166566e-02
1.64712384e-01 -4.02113467e-01 6.90451860e-01 -9.36462402e-01
8.64098847e-01 6.82975305e-03 1.79488823e-01 -2.19339395e+00
-5.13912082e-01 -8.61659586e-01 -4.34679836e-01 1.35022163e+00
8.94918859e-01 -3.09756845e-01 7.98671842e-01 3.87797952e-01
-5.25374889e-01 -4.69650835e-01 -9.81000483e-01 -7.73704171e-01
6.06429100e-01 -2.66335189e-01 5.50029874e-01 6.47931755e-01
1.94423720e-01 9.52140450e-01 7.60661289e-02 -4.31405336e-01
3.19804221e-01 -1.37806311e-01 6.50853157e-01 -1.17778766e+00
-3.09804648e-01 -6.67226493e-01 -3.64935100e-01 -8.63031566e-01
2.47053541e-02 -1.58857250e+00 1.00802220e-01 -2.59486961e+00
-9.81878415e-02 8.26283172e-02 -8.42114240e-02 6.53406620e-01
-3.00430566e-01 1.61379531e-01 2.02194914e-01 -4.43878561e-01
-5.41163266e-01 6.51591659e-01 1.06733775e+00 -3.34442765e-01
4.41669971e-02 -6.25943542e-01 -1.10787809e+00 2.43690372e-01
9.69062448e-01 -3.34108382e-01 -5.89175344e-01 -1.14885986e+00
4.44012344e-01 -2.76174605e-01 1.06215604e-01 -1.14968991e+00
-9.02437121e-02 2.94261307e-01 4.05446559e-01 -3.35108757e-01
1.96222439e-01 -1.28020719e-01 -2.25403696e-01 1.99127674e-01
-6.50091946e-01 1.98190093e-01 3.93965602e-01 6.02288902e-01
-1.99986890e-01 -3.22763503e-01 2.57223189e-01 -1.55908957e-01
-5.80283165e-01 1.12476379e-01 -3.41435045e-01 4.66368198e-01
6.35623455e-01 -8.34095385e-03 -7.64671564e-01 -5.11945605e-01
-5.60410321e-01 2.40726292e-01 1.21335626e-01 8.32429349e-01
5.71335137e-01 -1.55028772e+00 -1.13160384e+00 -1.02216721e-01
4.87072408e-01 9.62754898e-03 -6.95609814e-03 4.03554253e-02
-6.37987614e-01 8.63009810e-01 -4.38814834e-02 -1.65859554e-02
-8.93327951e-01 1.65402308e-01 1.14397027e-01 -6.89473808e-01
-3.56493086e-01 1.08982444e+00 -4.46848899e-01 -6.66155338e-01
-2.81203967e-02 -4.24721003e-01 -7.26577267e-02 2.26157218e-01
5.58299959e-01 4.40402091e-01 2.63027549e-01 -1.79625660e-01
-8.32223669e-02 -4.38005850e-02 -2.47846231e-01 -4.18661326e-01
1.21710205e+00 3.33586633e-01 3.39071304e-01 3.58908683e-01
1.17193615e+00 -1.78825140e-01 -9.07382011e-01 -7.64355529e-03
1.14115357e-01 -1.09897936e-02 3.80843043e-01 -1.46466804e+00
-8.19135249e-01 1.09718108e+00 4.24100071e-01 3.48066688e-01
5.75780571e-01 1.17310777e-01 1.07424235e+00 7.14731693e-01
6.21982589e-02 -1.22426748e+00 1.42335281e-01 8.49957883e-01
1.15755129e+00 -1.06922400e+00 -2.57846743e-01 1.76245272e-01
-7.01663673e-01 1.16329765e+00 5.24371624e-01 -1.17329754e-01
2.02056631e-01 -1.70581996e-01 -7.48253167e-02 -2.57526934e-01
-1.05814886e+00 -3.58846575e-01 5.25562167e-01 1.01844597e+00
7.43589640e-01 2.52703220e-01 -1.79309189e-01 3.88705432e-01
-6.09172165e-01 1.42885923e-01 5.45749784e-01 5.41872025e-01
-3.81326854e-01 -1.40182400e+00 2.52784431e-01 6.69598579e-01
-4.24734324e-01 -5.47872722e-01 -7.22591043e-01 7.56686330e-01
-2.04370171e-01 1.15952754e+00 3.63239944e-01 -3.24904561e-01
4.89501178e-01 6.99676514e-01 4.99499917e-01 -1.02771008e+00
-8.27444851e-01 -4.23442930e-01 8.15944254e-01 -3.24470609e-01
9.38128084e-02 -2.77203709e-01 -1.31276453e+00 -2.95602977e-01
-1.24385759e-01 9.98907760e-02 7.43772626e-01 4.82641190e-01
7.87351906e-01 8.72510970e-01 8.13793540e-02 -6.16022825e-01
-5.62314451e-01 -1.57516623e+00 -2.85370369e-02 4.19299841e-01
-9.83841419e-02 -2.77941748e-02 -1.29607975e-01 1.96803629e-01] | [11.398269653320312, 8.920001029968262] |
d53c4267-1431-4343-8e57-6f1babbd5482 | neural-rgb-d-sensing-depth-and-uncertainty | 1901.02571 | null | http://arxiv.org/abs/1901.02571v1 | http://arxiv.org/pdf/1901.02571v1.pdf | Neural RGB->D Sensing: Depth and Uncertainty from a Video Camera | Depth sensing is crucial for 3D reconstruction and scene understanding.
Active depth sensors provide dense metric measurements, but often suffer from
limitations such as restricted operating ranges, low spatial resolution, sensor
interference, and high power consumption. In this paper, we propose a deep
learning (DL) method to estimate per-pixel depth and its uncertainty
continuously from a monocular video stream, with the goal of effectively
turning an RGB camera into an RGB-D camera. Unlike prior DL-based methods, we
estimate a depth probability distribution for each pixel rather than a single
depth value, leading to an estimate of a 3D depth probability volume for each
input frame. These depth probability volumes are accumulated over time under a
Bayesian filtering framework as more incoming frames are processed
sequentially, which effectively reduces depth uncertainty and improves
accuracy, robustness, and temporal stability. Compared to prior work, the
proposed approach achieves more accurate and stable results, and generalizes
better to new datasets. Experimental results also show the output of our
approach can be directly fed into classical RGB-D based 3D scanning methods for
3D scene reconstruction. | ['Kihwan Kim', 'Chao Liu', 'Jan Kautz', 'Srinivasa Narasimhan', 'Jinwei Gu'] | 2019-01-09 | null | null | null | null | ['3d-scene-reconstruction'] | ['computer-vision'] | [ 4.74203467e-01 -2.00251658e-02 -9.02076624e-03 -5.54722309e-01
-7.40406334e-01 -2.55401552e-01 3.07545751e-01 7.36294836e-02
-6.50275707e-01 5.29102445e-01 2.70721829e-03 -3.97430882e-02
2.18211133e-02 -1.08269143e+00 -7.54943848e-01 -6.94053710e-01
2.56959617e-01 3.83946478e-01 6.97641075e-01 7.20048845e-01
2.63449073e-01 7.77309477e-01 -1.69992900e+00 -8.50895196e-02
4.42876488e-01 1.57608736e+00 5.70961654e-01 6.89457595e-01
-2.36438185e-01 7.82298982e-01 -3.63171697e-01 1.55560568e-01
2.23865390e-01 -1.96948215e-01 -3.37817758e-01 3.20935041e-01
5.02770662e-01 -1.02619517e+00 -6.35718167e-01 9.26702797e-01
5.26111603e-01 -7.58457556e-03 3.80867153e-01 -8.64980280e-01
1.50668517e-01 2.05568559e-02 -6.83005810e-01 -1.54578328e-01
6.91055715e-01 1.07948147e-01 4.47882265e-01 -7.22386897e-01
3.05466026e-01 1.12909949e+00 5.32896221e-01 5.61756611e-01
-8.27391744e-01 -4.52453822e-01 8.34633335e-02 2.30608776e-01
-1.29612792e+00 -4.12482023e-01 8.09632421e-01 -2.98161358e-01
8.14699829e-01 -6.99649155e-02 9.66342986e-01 8.48345816e-01
1.12155564e-01 9.96965647e-01 1.02055609e+00 -2.46897951e-01
6.87452316e-01 -3.74714077e-01 -3.62134367e-01 5.21841288e-01
1.86024293e-01 1.15608707e-01 -8.98090482e-01 2.06899911e-01
1.14516449e+00 3.49786103e-01 -4.72544253e-01 -6.65599704e-01
-1.18112636e+00 3.86810720e-01 4.28567737e-01 -1.73082680e-01
-4.71165448e-01 3.89323443e-01 -5.65665849e-02 -1.33339912e-01
4.07953680e-01 -2.45811656e-01 -3.89449239e-01 -4.01312947e-01
-9.92448151e-01 7.22387359e-02 4.44637656e-01 8.97343576e-01
1.10218954e+00 -2.26252079e-01 2.00886115e-01 4.29823101e-01
7.00619459e-01 9.34176505e-01 1.83386058e-01 -1.52185917e+00
4.48990285e-01 5.53472996e-01 9.26199034e-02 -5.96853256e-01
-3.89835924e-01 9.64174569e-02 -6.26397014e-01 4.76235211e-01
3.84108961e-01 1.33652598e-01 -9.03044403e-01 1.29347444e+00
4.33218300e-01 2.94416964e-01 -1.11203842e-01 1.01199603e+00
7.50517488e-01 5.45097232e-01 -4.15885150e-01 -2.40793899e-01
8.15204859e-01 -3.14140320e-01 -5.26949704e-01 -6.37031853e-01
-4.27585095e-02 -4.97302264e-01 9.02783036e-01 8.76120031e-01
-1.03782964e+00 -2.35565901e-01 -1.10729492e+00 -2.82136112e-01
2.15858802e-01 -2.17372969e-01 4.75206912e-01 6.67377234e-01
-9.48478460e-01 4.75356102e-01 -1.28020990e+00 -1.88363418e-01
6.45974398e-01 1.83493450e-01 -1.53004780e-01 -6.15487218e-01
-6.49979174e-01 5.71525097e-01 3.25611562e-01 -8.33691563e-03
-9.24365520e-01 -5.74494481e-01 -1.02738297e+00 -4.10274386e-01
3.27210099e-01 -6.70147240e-01 1.26535118e+00 -6.57076314e-02
-1.95691013e+00 7.57156491e-01 -4.89873439e-01 -2.60605335e-01
4.83501464e-01 -5.18520296e-01 3.42594028e-01 5.44506371e-01
-7.44247809e-02 6.79495037e-01 7.76211441e-01 -1.30287945e+00
-7.25551486e-01 -8.62036109e-01 1.56053871e-01 2.78272390e-01
-2.13345457e-02 -5.96147120e-01 -7.98651755e-01 1.89290822e-01
1.06660676e+00 -4.54810351e-01 -2.31263041e-01 6.51665092e-01
-1.51819959e-01 2.17748448e-01 6.42632723e-01 -2.25400895e-01
7.41960466e-01 -1.97494721e+00 1.26688212e-01 -8.08851793e-02
3.29549611e-01 -1.59195215e-01 3.44124407e-01 -4.74130549e-02
6.43982708e-01 -3.43511820e-01 -5.00562429e-01 -7.38242745e-01
-3.21091175e-01 4.35476243e-01 -3.56457047e-02 6.98405147e-01
-1.59020975e-01 5.40698171e-01 -1.00089610e+00 -4.90886450e-01
9.43061054e-01 7.52703488e-01 -4.92147684e-01 3.01152885e-01
-3.70034069e-01 6.53077126e-01 -3.77527297e-01 9.43843126e-01
9.81158912e-01 -2.20912732e-02 -7.33301714e-02 -2.81797230e-01
-1.12185627e-01 3.32423717e-01 -1.30581391e+00 2.13243389e+00
-7.20256746e-01 6.61962807e-01 9.70394388e-02 -5.63337505e-01
1.02023315e+00 -3.67998667e-02 5.98513544e-01 -9.21593845e-01
2.51780182e-01 9.26061496e-02 -7.74527013e-01 -4.49025869e-01
4.46926355e-01 -1.68806046e-01 5.41930944e-02 3.79485041e-01
-1.28377423e-01 -9.10512149e-01 -3.59961629e-01 2.06769928e-02
1.25746810e+00 4.90940958e-01 2.25068331e-01 4.15292919e-01
2.88035125e-01 -3.20176631e-01 6.30813599e-01 6.57901406e-01
-1.90412790e-01 8.97218525e-01 1.02853738e-01 -2.66121954e-01
-9.48312044e-01 -1.42080963e+00 -3.47456694e-01 1.01347551e-01
7.13850975e-01 -1.39565885e-01 -3.57293934e-01 -1.94590047e-01
1.03743814e-01 5.09977698e-01 -1.79082543e-01 6.25172704e-02
-4.38036919e-01 -4.40739304e-01 1.77031785e-01 5.14613628e-01
6.93459034e-01 -5.85300148e-01 -1.42991614e+00 2.76025504e-01
-2.94669986e-01 -1.45926285e+00 1.00396737e-01 2.61333197e-01
-1.34970784e+00 -9.98299479e-01 -5.17448366e-01 -5.09351902e-02
3.45585525e-01 3.78814757e-01 9.57683444e-01 -4.67364699e-01
-2.52461016e-01 7.00989723e-01 -2.57511795e-01 -3.63046110e-01
4.01747003e-02 -5.42593062e-01 5.82573339e-02 -1.63459569e-01
5.15930355e-01 -7.45418012e-01 -9.21790957e-01 2.04814225e-01
-8.96017075e-01 6.74254540e-03 4.43033129e-01 3.79420012e-01
1.03209674e+00 1.30260095e-01 -2.16907799e-01 -3.43251497e-01
-1.87210605e-01 -1.64185509e-01 -9.28879499e-01 -4.01647151e-01
-4.89623904e-01 -1.18737891e-01 1.39202252e-01 -1.07027456e-01
-1.14743638e+00 4.50637698e-01 -2.18538657e-01 -6.47324920e-01
-3.76646489e-01 9.06828791e-02 -4.85778749e-01 -4.41774726e-02
4.73031551e-01 2.71507293e-01 1.32317897e-02 -5.31722605e-01
2.23175526e-01 6.44623518e-01 6.89257443e-01 -3.24159801e-01
5.41763186e-01 1.11227167e+00 2.35823020e-01 -9.66767251e-01
-8.54125321e-01 -5.19263685e-01 -9.21060383e-01 -5.46301067e-01
7.11952686e-01 -1.24631727e+00 -8.11647773e-01 8.70069087e-01
-1.14604044e+00 -3.81055921e-01 -3.36543292e-01 8.10207009e-01
-7.05993593e-01 6.75348282e-01 -5.53192437e-01 -1.13332176e+00
1.57942787e-01 -9.86104965e-01 1.51557410e+00 2.88500249e-01
4.74839769e-02 -8.24801803e-01 -4.43562455e-02 4.27908599e-01
-3.51419300e-03 3.64135504e-01 4.52040851e-01 6.90875411e-01
-1.18052518e+00 -2.66201824e-01 -2.72303909e-01 3.53943646e-01
1.17062919e-01 -2.44627595e-01 -1.37284994e+00 4.64947820e-02
3.79880160e-01 -3.44324380e-01 8.49998832e-01 6.79573894e-01
1.26150012e+00 1.51100904e-01 -2.34852806e-01 9.37255442e-01
1.70483410e+00 1.65406436e-01 7.64411390e-01 3.02457541e-01
7.41752267e-01 3.64578396e-01 8.40729892e-01 8.78102720e-01
5.81243157e-01 4.42848682e-01 9.66594040e-01 2.67033160e-01
-8.43544304e-02 -2.19568029e-01 3.40869457e-01 4.80562270e-01
1.38857558e-01 -3.47574502e-01 -8.74620676e-01 4.53093499e-01
-1.67708552e+00 -6.75218761e-01 -1.33620098e-01 2.44763017e+00
6.70890868e-01 2.60237306e-01 -2.12919101e-01 4.78704870e-01
3.29311252e-01 2.39724398e-01 -1.06036425e+00 1.73106611e-01
-1.74075186e-01 2.51800567e-01 8.48609924e-01 6.44431114e-01
-7.27951169e-01 5.64952016e-01 6.46431303e+00 4.17069316e-01
-1.05144215e+00 1.59584098e-02 2.63531208e-01 -5.75263917e-01
-5.52654147e-01 -1.38143152e-01 -7.57158041e-01 3.27586144e-01
5.50914228e-01 2.25817606e-01 2.58902460e-01 7.73254931e-01
3.38186353e-01 -1.12878430e+00 -1.16486907e+00 1.56167042e+00
5.78664690e-02 -1.01824462e+00 -2.32140765e-01 1.96108595e-01
4.68132198e-01 3.34679067e-01 -2.70548016e-01 -4.11877751e-01
1.47112608e-01 -7.31811464e-01 9.80346680e-01 6.21691525e-01
8.35506082e-01 -6.73937798e-01 6.22180462e-01 6.31324768e-01
-9.18970942e-01 -1.39805917e-02 -6.43318951e-01 -3.60544741e-01
5.39888382e-01 1.38509297e+00 -6.21432543e-01 3.60498607e-01
1.08407247e+00 8.76891553e-01 -8.76488611e-02 1.11475003e+00
-5.29837847e-01 2.33400747e-01 -7.88108706e-01 -1.99171919e-02
-1.05120070e-01 -1.18365534e-01 4.54936832e-01 6.01669431e-01
5.03781497e-01 3.63156617e-01 -4.55110259e-02 8.52710426e-01
5.52625656e-02 -5.41118085e-01 -6.81672394e-01 4.62149322e-01
7.56938696e-01 8.85597467e-01 -6.51548922e-01 -1.89829364e-01
-4.53183502e-01 1.08431160e+00 5.14204726e-02 1.44086480e-01
-5.43153465e-01 -1.24754600e-01 5.80226719e-01 3.02704096e-01
2.88389057e-01 -5.86423039e-01 -6.62052095e-01 -1.12439060e+00
2.14164734e-01 -1.56502932e-01 -1.02305198e-02 -1.02018750e+00
-8.87591600e-01 2.04765007e-01 7.61862844e-02 -1.35756266e+00
-1.02113679e-01 -6.58689737e-01 -1.23505659e-01 7.05666244e-01
-1.84514630e+00 -6.10950589e-01 -8.55491996e-01 6.73096359e-01
4.86931890e-01 3.66722733e-01 4.73319203e-01 3.08533281e-01
-1.80502266e-01 -2.08063293e-02 -2.52081230e-02 -2.43431956e-01
4.24474746e-01 -1.06516135e+00 1.84612498e-01 9.74695742e-01
-9.84389484e-02 1.40162716e-02 4.60708052e-01 -5.74896634e-01
-1.70192778e+00 -8.39205861e-01 3.65531862e-01 -5.70205271e-01
2.03385696e-01 -2.91696489e-01 -6.33967280e-01 3.43556195e-01
-4.54159379e-01 8.10459033e-02 3.96072060e-01 -3.77231777e-01
-8.53937864e-02 -3.45676154e-01 -1.44102156e+00 9.92178470e-02
1.36514032e+00 -7.09193945e-01 -3.99027288e-01 -1.15954615e-01
5.39636075e-01 -8.48705649e-01 -7.50883877e-01 6.30614698e-01
7.09599793e-01 -1.58675814e+00 1.01904154e+00 6.33222818e-01
4.12293345e-01 -5.76440632e-01 -5.99362314e-01 -7.62151480e-01
1.71712995e-01 -2.53131181e-01 -5.07888556e-01 7.47391999e-01
-5.52347004e-02 -2.67279595e-01 1.12646616e+00 7.29445755e-01
-1.42354831e-01 -5.41051328e-01 -1.30228698e+00 -5.46861351e-01
-3.93782109e-01 -1.21723533e+00 5.10548115e-01 3.09671342e-01
-3.76422137e-01 -4.19391245e-02 -2.02233851e-01 5.76656997e-01
1.37702763e+00 1.95244014e-01 8.00499856e-01 -1.26863420e+00
-2.71984428e-01 -1.13838933e-01 -6.44648671e-01 -1.90017390e+00
-2.77757287e-01 -1.88135833e-01 4.07220513e-01 -1.77902496e+00
-3.27435695e-02 -4.18399006e-01 5.82944825e-02 8.09438601e-02
1.42978638e-01 5.16209245e-01 -2.22400695e-01 1.32471859e-01
-5.15715420e-01 6.63998961e-01 1.09388542e+00 5.70161119e-02
-1.93427309e-01 4.01261821e-02 -1.00785695e-01 9.64943886e-01
3.77947807e-01 -3.43050331e-01 -4.63528901e-01 -8.39601755e-01
3.82758766e-01 3.48293483e-01 4.12066013e-01 -1.23132408e+00
3.55001956e-01 -1.36720181e-01 7.48810768e-01 -1.08897734e+00
7.62917876e-01 -8.64422560e-01 -6.05831668e-03 3.00513774e-01
3.21128875e-01 -5.22279859e-01 6.63979724e-02 6.85621440e-01
-1.46477684e-01 -1.49730757e-01 9.91267204e-01 -3.89861435e-01
-9.74761605e-01 4.58080947e-01 -3.61427099e-01 -3.57542694e-01
8.34307969e-01 -7.33799160e-01 2.82135874e-01 -5.28246880e-01
-4.79667932e-01 3.74976657e-02 7.81651616e-01 9.11503732e-02
1.17284930e+00 -1.22258687e+00 -2.87023872e-01 3.37773770e-01
1.53174847e-01 9.42556798e-01 4.07669067e-01 4.55954552e-01
-8.10507178e-01 1.95823193e-01 3.66512164e-02 -1.39256942e+00
-8.48789454e-01 1.11939870e-01 3.60869110e-01 4.24377620e-01
-8.98503006e-01 1.07705677e+00 1.81526542e-02 -3.12093318e-01
5.23702800e-01 -6.20651364e-01 2.76769876e-01 -1.02322467e-01
6.47263944e-01 4.14399803e-01 8.11104923e-02 -2.81671882e-01
-4.73267734e-01 1.14498317e+00 5.50768554e-01 -3.72963130e-01
1.33930171e+00 -7.27489650e-01 8.81731212e-02 8.05027664e-01
9.46773291e-01 -1.57950938e-01 -1.92446852e+00 -4.58013684e-01
-3.30032080e-01 -9.92648363e-01 6.35024488e-01 -3.87433708e-01
-1.07872021e+00 1.07349157e+00 6.94088459e-01 -2.14491427e-01
1.21147728e+00 2.80805081e-01 8.65929961e-01 2.16833234e-01
9.32802379e-01 -8.63037407e-01 1.26363054e-01 4.59526122e-01
2.76063770e-01 -1.38956463e+00 3.91589463e-01 -2.79750586e-01
-1.40010893e-01 1.23806131e+00 5.00592649e-01 1.56843260e-01
5.86738527e-01 5.49239397e-01 3.44444551e-02 -3.51157114e-02
-3.70152384e-01 -2.31153578e-01 -2.51058668e-01 8.02852333e-01
2.68645864e-02 -2.12525859e-01 4.16325092e-01 -6.43200576e-02
4.20377441e-02 1.82661980e-01 6.84632599e-01 1.02999127e+00
-7.30365217e-01 -8.99442494e-01 -5.46116412e-01 2.00460821e-01
6.82458058e-02 1.51267797e-01 -2.24014930e-03 4.53110486e-01
1.00608781e-01 9.07132506e-01 3.62753540e-01 -3.15289348e-01
3.07074696e-01 -2.26955101e-01 1.03122818e+00 -6.17067337e-01
4.98433590e-01 1.20256156e-01 -3.34134281e-01 -1.03460789e+00
-7.90360868e-01 -7.99609125e-01 -1.38201654e+00 -2.26280108e-01
-1.44417033e-01 -4.26553011e-01 1.08468711e+00 8.55917275e-01
1.05727360e-01 2.15560853e-01 7.38964140e-01 -1.21782255e+00
2.88349316e-02 -7.23866999e-01 -6.55211329e-01 -1.17341101e-01
6.28122032e-01 -6.51897311e-01 -5.87289929e-01 -3.61054875e-02] | [8.88271427154541, -2.550079107284546] |
bb317c20-f259-4801-add9-6bbd7f94dcd1 | classifying-variable-length-audio-files-with | 1607.02857 | null | http://arxiv.org/abs/1607.02857v1 | http://arxiv.org/pdf/1607.02857v1.pdf | Classifying Variable-Length Audio Files with All-Convolutional Networks and Masked Global Pooling | We trained a deep all-convolutional neural network with masked global pooling
to perform single-label classification for acoustic scene classification and
multi-label classification for domestic audio tagging in the DCASE-2016
contest. Our network achieved an average accuracy of 84.5% on the four-fold
cross-validation for acoustic scene recognition, compared to the provided
baseline of 72.5%, and an average equal error rate of 0.17 for domestic audio
tagging, compared to the baseline of 0.21. The network therefore improves the
baselines by a relative amount of 17% and 19%, respectively. The network only
consists of convolutional layers to extract features from the short-time
Fourier transform and one global pooling layer to combine those features. It
particularly possesses neither fully-connected layers, besides the
fully-connected output layer, nor dropout layers. | ['Alfred Mertins', 'Huy Phan', 'Lars Hertel'] | 2016-07-11 | null | null | null | null | ['audio-tagging'] | ['audio'] | [ 4.15392727e-01 2.59373784e-01 2.43363619e-01 -5.60257673e-01
-1.52252591e+00 -6.05833769e-01 1.49388298e-01 -8.93539935e-02
-7.58406401e-01 9.44388807e-02 4.31082100e-02 3.81915905e-02
4.76319224e-01 -5.14203787e-01 -6.00827098e-01 -8.68447304e-01
-2.79636323e-01 -2.75349855e-01 2.15105727e-01 3.08886051e-01
-4.05006558e-01 2.10499018e-01 -1.57028472e+00 5.80742002e-01
1.40786707e-01 1.88981783e+00 -2.64102906e-01 1.00284016e+00
4.15503234e-01 1.06390822e+00 -8.99341226e-01 -4.48430449e-01
-1.11596540e-01 4.62987386e-02 -9.51616526e-01 -2.87597567e-01
6.86962903e-01 -3.07268411e-01 -6.82012796e-01 8.65326464e-01
1.09517896e+00 3.52353379e-02 2.95200378e-01 -1.17696166e+00
-5.06635308e-01 9.17827010e-01 -3.04465920e-01 2.06832960e-01
1.22085616e-01 -1.70121387e-01 1.27256763e+00 -8.24464083e-01
1.36904284e-01 8.98802638e-01 9.71728265e-01 5.19250929e-01
-9.62695420e-01 -1.14726663e+00 -8.03907365e-02 9.53735113e-02
-1.63470256e+00 -6.69982314e-01 6.63022637e-01 -4.30521011e-01
9.73212421e-01 2.09896222e-01 3.36163044e-01 1.12038422e+00
-1.03780679e-01 7.28703856e-01 1.04388332e+00 -1.59703016e-01
5.44438846e-02 -5.19514494e-02 1.81718051e-01 6.88971639e-01
-4.38854963e-01 -1.48684263e-01 -8.17775369e-01 -1.54932246e-01
2.23268747e-01 -5.82076371e-01 -1.04366481e-01 4.26401615e-01
-1.05835092e+00 4.79435474e-01 4.64970529e-01 3.05993021e-01
-5.69971241e-02 5.48395753e-01 7.17612684e-01 2.23961234e-01
8.11290681e-01 7.06514835e-01 -7.09192336e-01 -2.77996629e-01
-9.73511457e-01 5.98494634e-02 6.05277359e-01 8.01222742e-01
4.35402513e-01 4.84183371e-01 -3.53814244e-01 1.18466353e+00
-4.98050153e-02 6.08531415e-01 3.77158761e-01 -9.94416177e-01
3.47579747e-01 2.77560540e-02 -1.97861388e-01 -6.93171024e-01
-7.66212583e-01 -9.84852076e-01 -7.93913543e-01 -1.94083765e-01
1.61814600e-01 -4.66456681e-01 -1.03085041e+00 1.99016285e+00
-1.30582184e-01 4.02110130e-01 -2.67261732e-03 9.00591433e-01
1.32928360e+00 6.26624763e-01 3.87851238e-01 1.47019520e-01
1.28069723e+00 -1.08196580e+00 -7.44363368e-01 -1.60772443e-01
8.14414144e-01 -9.68969047e-01 9.23788965e-01 2.20245451e-01
-9.81010497e-01 -7.14069009e-01 -1.26018274e+00 2.66197342e-02
-3.89712185e-01 5.61096489e-01 3.15339625e-01 8.15746427e-01
-1.15788329e+00 4.03957397e-01 -4.74104285e-01 3.67211178e-02
6.24990046e-01 4.54146802e-01 -3.86523783e-01 2.40495533e-01
-1.44031560e+00 4.24601823e-01 6.28933385e-02 1.03927411e-01
-1.31701756e+00 -1.00234687e+00 -9.09513056e-01 1.09868877e-01
-2.47116340e-03 -1.02159172e-01 1.65145338e+00 -7.25001931e-01
-1.58925724e+00 1.19445741e+00 2.41510198e-01 -6.22870088e-01
1.95280090e-01 -2.21307248e-01 -8.08973074e-01 2.37051964e-01
-4.09729518e-02 8.43891323e-01 5.62096715e-01 -7.35565245e-01
-7.10887790e-01 2.02674508e-01 1.39236495e-01 -3.48758027e-02
-3.76664668e-01 4.69158828e-01 -2.34032869e-01 -8.85365844e-01
1.01031750e-01 -1.01157355e+00 2.76355632e-02 -3.64792556e-01
-4.18962479e-01 -1.83711216e-01 7.25469530e-01 -9.55193222e-01
1.16653168e+00 -2.73735929e+00 -2.90431201e-01 -1.22110434e-01
8.12921375e-02 1.60751879e-01 -4.55433041e-01 -1.07352711e-01
-3.29410255e-01 3.45534772e-01 -2.24692807e-01 -8.46517324e-01
1.26730829e-01 -3.61361265e-01 -3.92598003e-01 4.49646622e-01
4.84193236e-01 6.74627900e-01 -7.98151314e-01 -5.06267436e-02
-4.45388518e-02 5.29394388e-01 -4.32723284e-01 3.55682731e-01
2.38225400e-01 2.70956099e-01 3.17886658e-03 7.67913520e-01
8.04139912e-01 1.85972184e-01 -1.01367816e-01 -2.29131758e-01
5.34944423e-02 8.32613528e-01 -1.05756712e+00 1.79010880e+00
-7.71393359e-01 8.61044705e-01 4.36193258e-01 -7.43468583e-01
9.31377053e-01 7.32651591e-01 5.71123362e-01 -6.52877271e-01
2.65826762e-01 4.63638723e-01 -6.48848936e-02 -2.98267573e-01
2.58354247e-01 -4.91767339e-02 -7.36225247e-01 1.94291517e-01
5.69906354e-01 -5.47095537e-02 -4.12844241e-01 -2.29221908e-03
1.34734595e+00 -3.77756119e-01 -3.17794323e-01 -2.67565966e-01
5.19695461e-01 -6.39397025e-01 7.58125603e-01 5.76809883e-01
-3.31990361e-01 9.80169594e-01 5.07185519e-01 -4.26191986e-01
-6.89355195e-01 -9.52687442e-01 -2.02086657e-01 1.33952904e+00
-3.03759843e-01 -6.00839853e-01 -8.63637328e-01 -6.04718804e-01
-1.99300468e-01 3.07354093e-01 -5.31555891e-01 -2.69397974e-01
-5.31965256e-01 -6.56846642e-01 1.50768769e+00 7.29031742e-01
9.07339394e-01 -1.27239954e+00 -4.52339888e-01 2.13116512e-01
-4.36983138e-01 -1.60422957e+00 -3.62816989e-01 7.56075919e-01
-5.67873493e-02 -7.27007806e-01 -5.69525063e-01 -9.49853480e-01
-7.12061822e-02 -2.77390629e-01 1.02718127e+00 -2.36301214e-01
-1.83999687e-01 5.90744466e-02 -3.90689194e-01 -6.77312374e-01
-2.10603073e-01 4.50264573e-01 1.13097534e-01 3.40002716e-01
2.44106263e-01 -6.12653017e-01 -4.11337227e-01 3.10806870e-01
-5.24133742e-01 -1.86030105e-01 3.73018831e-01 8.45322013e-01
4.91608202e-01 -1.33880794e-01 8.65094543e-01 -2.39797920e-01
1.44675687e-01 -1.61885709e-01 -3.67505610e-01 -5.97506110e-03
-8.45007657e-04 -4.90440518e-01 3.33224863e-01 -3.27958763e-01
-5.89502335e-01 2.44331092e-01 -8.07830632e-01 -4.76903260e-01
-5.42053342e-01 1.85403407e-01 -1.45531476e-01 -8.77404958e-02
2.18924612e-01 -1.26926497e-01 -4.97370124e-01 -6.79720640e-01
1.34787515e-01 1.14524865e+00 8.01352262e-01 -2.72528529e-01
3.90587807e-01 7.93529488e-03 -2.28539556e-01 -6.43619835e-01
-1.31617200e+00 -5.81712842e-01 -6.17256701e-01 -3.35788161e-01
1.41769612e+00 -1.56183732e+00 -6.49403393e-01 1.21804488e+00
-1.04024374e+00 -5.35044134e-01 -3.35351825e-01 5.23845673e-01
-3.54507267e-01 -2.53641933e-01 -9.72447515e-01 -7.09908187e-01
-3.38906944e-01 -1.10549462e+00 1.21863890e+00 -1.29507212e-02
-9.47895497e-02 -3.52839530e-01 -2.72611946e-01 4.14725840e-01
5.69738805e-01 2.55569398e-01 4.59041655e-01 -7.23832250e-01
-1.59038007e-01 -3.60680401e-01 -1.48144469e-01 7.56996989e-01
-1.25910670e-01 -2.42515385e-01 -2.20069337e+00 -2.72733688e-01
-3.05676013e-01 -8.13540697e-01 1.30491507e+00 4.63769495e-01
1.65835774e+00 -6.01685941e-02 -1.34827625e-02 7.04140365e-01
8.11508298e-01 2.24199593e-01 4.14067924e-01 1.42158806e-01
7.11690307e-01 4.45267498e-01 2.65036851e-01 2.32170507e-01
3.70838016e-01 8.65617096e-01 4.33822989e-01 -2.98271507e-01
-5.34586310e-01 -6.40472174e-02 3.31773549e-01 1.11989570e+00
2.63250500e-01 -7.41534457e-02 -9.86998081e-01 6.20245039e-01
-1.37893021e+00 -6.43477857e-01 7.17422292e-02 1.83753765e+00
8.10720026e-01 8.22673514e-02 8.54182392e-02 5.98848581e-01
6.21810555e-01 4.85639125e-01 -2.11552575e-01 -5.32556474e-01
-3.33365500e-01 6.07099473e-01 4.78583425e-01 2.66994208e-01
-2.01515007e+00 9.30650353e-01 6.67238188e+00 9.74204183e-01
-1.27098989e+00 4.63182211e-01 9.38440621e-01 -2.60995090e-01
3.81685555e-01 -6.58581257e-01 -6.64781034e-01 2.48703495e-01
1.55950499e+00 5.39798558e-01 1.49509162e-01 8.55443478e-01
-3.54095370e-01 2.52331853e-01 -6.09589577e-01 1.03966558e+00
9.68215056e-03 -1.02926433e+00 -5.26417494e-01 -6.47758171e-02
7.37660766e-01 3.06817919e-01 3.22691470e-01 3.98281485e-01
7.64371902e-02 -1.09956443e+00 1.18472159e+00 1.68628946e-01
1.25312722e+00 -1.09251821e+00 1.14513803e+00 -6.23702481e-02
-1.62572265e+00 -3.28181207e-01 -1.68660998e-01 -1.83199987e-01
6.62930086e-02 5.74025750e-01 -5.84407091e-01 4.22263771e-01
1.24037981e+00 4.33105290e-01 -5.88853776e-01 9.81433570e-01
-1.86358556e-01 1.27473211e+00 -4.26010668e-01 2.21791461e-01
3.40900719e-01 6.28864288e-01 2.26864293e-01 1.59295154e+00
2.61186033e-01 -1.49681270e-01 1.03124134e-01 2.28005990e-01
-5.08495152e-01 -6.65814057e-02 -2.40885645e-01 6.89310357e-02
5.03765464e-01 1.42719352e+00 -3.23956341e-01 -1.37549892e-01
-3.06420356e-01 7.18134344e-01 2.42572755e-01 2.85869837e-01
-1.08536661e+00 -9.05583739e-01 8.78846228e-01 -2.81016290e-01
4.85232711e-01 -1.14993438e-01 -3.58202696e-01 -6.75381601e-01
-7.51190707e-02 -6.60910189e-01 3.99500996e-01 -5.10985315e-01
-1.03104556e+00 1.01852179e+00 -3.38487715e-01 -1.00627041e+00
-2.58847214e-02 -4.86658722e-01 -5.33828080e-01 9.64773476e-01
-1.54385209e+00 -1.47653961e+00 -2.60412395e-01 2.77746767e-01
2.48685375e-01 -3.08360457e-01 1.29494190e+00 8.51005256e-01
-6.21567667e-01 1.27698588e+00 -3.40378791e-01 6.64754093e-01
8.29881847e-01 -1.11665297e+00 4.96904463e-01 7.06808627e-01
2.38494575e-01 5.03106080e-02 2.22685933e-01 8.96539837e-02
-5.67886531e-01 -1.47073567e+00 1.21996045e+00 -1.44342259e-01
8.29021573e-01 -9.60517764e-01 -6.33129835e-01 3.50658417e-01
2.06993759e-01 4.26977128e-01 1.03858161e+00 2.82971263e-01
-7.91579008e-01 -3.67000163e-01 -9.05548215e-01 -9.19795968e-03
9.63827670e-01 -1.02400398e+00 -5.31591475e-03 2.35402763e-01
1.18255198e+00 -5.44950485e-01 -1.20020258e+00 6.72514319e-01
6.00036919e-01 -4.27389115e-01 1.02278566e+00 -2.88707793e-01
2.11698771e-01 -2.34273091e-01 -6.86476707e-01 -1.00327170e+00
-4.12999988e-01 -5.69914818e-01 1.87977433e-01 1.61834097e+00
7.78099120e-01 -3.26599419e-01 4.46080506e-01 -2.62166876e-02
-7.50264645e-01 -7.09382355e-01 -1.36075711e+00 -6.69605613e-01
4.73153517e-02 -9.34572995e-01 5.54153919e-01 8.83769989e-01
-3.47914398e-01 1.42249942e-01 -4.83112216e-01 3.75079632e-01
2.96631932e-01 -3.18367392e-01 3.69736284e-01 -1.04527712e+00
-6.03507571e-02 -2.88678378e-01 -7.35853672e-01 -6.74934387e-01
4.94596571e-01 -1.03116608e+00 1.51551634e-01 -1.09447443e+00
-3.54101807e-02 -3.86846513e-01 -1.01905799e+00 8.82473290e-01
2.03979582e-01 1.03573704e+00 1.89836964e-01 -2.48072207e-01
-6.23570979e-01 4.69352096e-01 6.96038604e-01 -5.15844405e-01
2.05863304e-02 1.05987303e-01 -5.48530340e-01 7.28176892e-01
1.00844252e+00 -5.86640656e-01 1.69554874e-01 -4.35700625e-01
-4.26070206e-02 -3.61787915e-01 4.75607783e-01 -1.52730930e+00
7.60218501e-02 5.42360008e-01 2.46326193e-01 -4.81462568e-01
5.29987395e-01 -4.45161134e-01 -1.53100803e-01 2.38570765e-01
-6.95190072e-01 -2.81886309e-01 6.20329738e-01 2.26369739e-01
-7.03301489e-01 1.24200759e-02 8.73333812e-01 2.87552297e-01
-7.32291818e-01 1.40747264e-01 -6.05730176e-01 -2.86700036e-02
5.67380905e-01 4.09930974e-01 -3.29470336e-01 -2.89198577e-01
-9.66072679e-01 2.56510973e-02 -3.63877028e-01 5.38763165e-01
2.34729275e-01 -1.63389492e+00 -6.15792394e-01 8.08183253e-02
1.45610988e-01 -1.81928918e-01 5.28240144e-01 7.19948828e-01
-2.35882550e-01 3.13312799e-01 4.76789549e-02 -6.56543255e-01
-1.40928221e+00 -1.25582576e-01 6.75981402e-01 -1.15715049e-01
-3.25037658e-01 1.48109245e+00 9.08717662e-02 -7.15417087e-01
6.93370819e-01 -4.98906076e-01 -2.07601622e-01 1.89134270e-01
4.98594612e-01 2.60300428e-01 3.34689766e-01 -8.29550326e-01
-6.28394783e-01 6.89422607e-01 3.73651773e-01 -1.18762404e-01
1.30094564e+00 1.92651108e-01 3.95762883e-02 7.23210096e-01
1.66953170e+00 -1.37789384e-01 -1.02403116e+00 -1.64013728e-01
-1.87619686e-01 9.96100083e-02 4.05998409e-01 -1.05029929e+00
-1.24916315e+00 1.19550252e+00 9.39343810e-01 3.20169479e-01
1.33866525e+00 2.82369077e-01 7.37819552e-01 3.16535503e-01
-5.42089939e-02 -1.06222212e+00 6.89883381e-02 7.12812185e-01
9.50743556e-01 -9.15164232e-01 -4.30344850e-01 -3.66573274e-01
-5.52808881e-01 8.80375624e-01 5.57673693e-01 -9.64773074e-02
1.07631290e+00 8.88793409e-01 4.55703497e-01 -5.81910722e-02
-5.95195293e-01 -1.66682824e-01 5.06069362e-01 5.07444978e-01
6.16576612e-01 3.83932352e-01 3.84054869e-01 1.12696183e+00
-5.46834886e-01 -3.80408436e-01 1.12251146e-02 6.25866950e-01
-3.85061085e-01 -6.18324578e-01 1.46903783e-01 1.80928439e-01
-1.03141809e+00 -6.91364035e-02 -4.63582158e-01 3.83484483e-01
3.98470938e-01 1.29038334e+00 4.36516792e-01 -1.02761376e+00
4.82040465e-01 4.12494272e-01 -3.49312811e-03 -4.52287704e-01
-1.07466590e+00 4.26387817e-01 5.09885192e-01 -6.12419605e-01
-6.56965375e-01 -5.46548545e-01 -9.54643548e-01 9.07016098e-02
-4.13412005e-01 -2.98622623e-02 6.75922394e-01 5.35289943e-01
2.48609692e-01 1.01098168e+00 7.70919681e-01 -7.10897744e-01
-2.29584083e-01 -1.34205222e+00 -7.47085452e-01 8.87657031e-02
5.25929689e-01 -5.11762083e-01 -4.67261761e-01 -7.72384927e-02] | [15.211390495300293, 5.099597930908203] |
e5a48aac-09c4-47ae-b9cb-86505704655f | towards-integration-of-discriminability-and | 2304.00824 | null | https://arxiv.org/abs/2304.00824v1 | https://arxiv.org/pdf/2304.00824v1.pdf | Towards Integration of Discriminability and Robustness for Document-Level Relation Extraction | Document-level relation extraction (DocRE) predicts relations for entity pairs that rely on long-range context-dependent reasoning in a document. As a typical multi-label classification problem, DocRE faces the challenge of effectively distinguishing a small set of positive relations from the majority of negative ones. This challenge becomes even more difficult to overcome when there exists a significant number of annotation errors in the dataset. In this work, we aim to achieve better integration of both the discriminability and robustness for the DocRE problem. Specifically, we first design an effective loss function to endow high discriminability to both probabilistic outputs and internal representations. We innovatively customize entropy minimization and supervised contrastive learning for the challenging multi-label and long-tailed learning problems. To ameliorate the impact of label errors, we equipped our method with a novel negative label sampling strategy to strengthen the model robustness. In addition, we introduce two new data regimes to mimic more realistic scenarios with annotation errors and evaluate our sampling strategy. Experimental results verify the effectiveness of each component and show that our method achieves new state-of-the-art results on the DocRED dataset, its recently cleaned version, Re-DocRED, and the proposed data regimes. | ['Lidong Bing', 'Stanley Kok', 'Jia Guo'] | 2023-04-03 | null | null | null | null | ['document-level-relation-extraction'] | ['natural-language-processing'] | [ 2.67688662e-01 1.53401211e-01 -3.93136710e-01 -4.51367706e-01
-1.04698575e+00 -4.93977755e-01 5.62740207e-01 3.75703394e-01
-3.54825646e-01 1.16409481e+00 -1.16045177e-01 -8.79816711e-02
-6.57720566e-01 -8.05655181e-01 -4.39168304e-01 -7.53210902e-01
-2.77841967e-02 6.76852882e-01 4.95737493e-02 -1.15517870e-01
-2.22043231e-01 4.33326364e-01 -1.33322978e+00 2.23472282e-01
1.05873680e+00 1.11914992e+00 -1.47986934e-01 1.25078097e-01
4.85667326e-02 8.08822334e-01 -4.43902284e-01 -7.08086431e-01
2.02511087e-01 2.63277646e-02 -1.04646552e+00 -4.80203629e-02
1.74518123e-01 3.04474950e-01 -1.07309856e-01 1.14083314e+00
6.29556417e-01 2.25636158e-02 1.02450132e+00 -1.39811134e+00
-5.03068745e-01 8.35970819e-01 -7.37074792e-01 5.54855689e-02
1.94665983e-01 -3.25851887e-01 1.40598977e+00 -6.50105655e-01
6.83056951e-01 1.12954581e+00 7.99502254e-01 3.36066961e-01
-1.33751023e+00 -7.95956612e-01 2.92021096e-01 1.41640887e-01
-1.48110402e+00 -2.16565818e-01 7.31928408e-01 -4.54235375e-01
6.02469206e-01 2.43264765e-01 1.23812109e-02 1.18419778e+00
1.05764784e-01 4.81778204e-01 1.24180686e+00 -3.82621258e-01
4.92134877e-02 3.24863106e-01 4.32405293e-01 5.00552475e-01
4.33672905e-01 -1.97242647e-01 -2.13347033e-01 -3.57805640e-01
1.39857277e-01 -4.70108613e-02 -3.93592209e-01 -5.10639846e-01
-9.23922002e-01 5.81250846e-01 2.31628016e-01 3.58249694e-01
-2.29221329e-01 -2.43168786e-01 3.96406472e-01 1.06421217e-01
6.72583461e-01 5.77943563e-01 -8.68588924e-01 2.86481291e-01
-6.02521718e-01 1.47420585e-01 8.47857714e-01 8.46174300e-01
6.72405481e-01 -7.02537358e-01 -5.87022245e-01 1.08410645e+00
2.55668759e-01 1.65359244e-01 2.96176881e-01 -3.72358859e-01
7.74715543e-01 7.52361357e-01 3.68206054e-02 -9.96219397e-01
-6.05924547e-01 -9.77046669e-01 -1.13235104e+00 -3.01759660e-01
3.19828570e-01 -1.78163886e-01 -6.47849739e-01 1.79537523e+00
4.55702901e-01 5.84350862e-02 3.05504680e-01 6.08747244e-01
7.29672432e-01 3.71304959e-01 4.04657722e-01 -4.49576706e-01
1.52821350e+00 -8.53517592e-01 -8.60451639e-01 -2.02476755e-01
9.06479478e-01 -4.11918133e-01 7.83969283e-01 1.78035647e-01
-6.49169803e-01 -1.55875474e-01 -1.13099158e+00 -8.29330683e-02
-5.12651563e-01 4.33111548e-01 7.21215308e-01 4.33220357e-01
-4.62496281e-01 5.94586313e-01 -3.48710179e-01 8.55910499e-03
4.05831337e-01 4.43907589e-01 -5.36547184e-01 -1.37131080e-01
-1.73960268e+00 1.04067719e+00 7.36283839e-01 1.59931511e-01
-2.45300129e-01 -6.73234165e-01 -8.02864790e-01 2.80678809e-01
7.72498906e-01 -5.32333612e-01 8.81256700e-01 -4.20139432e-01
-9.66046095e-01 8.24547410e-01 1.44936025e-01 -3.20823759e-01
5.51956236e-01 -1.02229722e-01 -5.40988564e-01 -1.31339476e-01
1.21072024e-01 1.85441375e-01 3.30438018e-01 -1.42170799e+00
-4.77632493e-01 -4.51341480e-01 -2.15542912e-02 1.16097815e-01
-4.67341930e-01 -1.48702577e-01 -9.12281722e-02 -6.33679032e-01
1.55781403e-01 -9.05587316e-01 -8.63698646e-02 -5.02407253e-01
-6.09937608e-01 -5.44757605e-01 5.67539573e-01 -2.88258672e-01
1.39861286e+00 -1.89488471e+00 -1.56602878e-02 3.22092205e-01
3.21934521e-01 4.64433759e-01 -1.82296634e-02 3.50663424e-01
-2.71988422e-01 2.52323449e-01 -2.98223466e-01 -4.80531394e-01
-6.51100874e-02 2.00544447e-01 -9.74947289e-02 3.39738876e-01
5.19369245e-01 6.78814173e-01 -9.48253453e-01 -7.40444601e-01
-3.58383983e-01 4.07991290e-01 -2.41196066e-01 1.31758749e-01
-1.40500635e-01 2.59246171e-01 -5.62596023e-01 6.94193423e-01
7.83172011e-01 -6.54495001e-01 6.06598735e-01 -4.43936884e-01
3.44327748e-01 3.03364158e-01 -1.33348167e+00 1.14631748e+00
-5.59246957e-01 -7.48083368e-02 -2.26311147e-01 -1.05258131e+00
1.09831965e+00 3.07599753e-01 5.02139986e-01 -3.30217600e-01
1.18334457e-01 2.75842130e-01 -1.08815923e-01 -4.49778438e-01
4.34900254e-01 -3.10666770e-01 -1.37787923e-01 1.93056807e-01
2.53598332e-01 1.14474952e-01 1.97641864e-01 1.87726796e-01
1.02409410e+00 -4.87298034e-02 6.25369906e-01 -2.26118416e-01
6.45335138e-01 -4.33833182e-01 1.01541722e+00 5.33742785e-01
-1.17827170e-01 3.87250066e-01 8.83689761e-01 -7.60613382e-02
-5.08652389e-01 -6.60122275e-01 -4.12115812e-01 1.04565930e+00
1.85746983e-01 -4.92235303e-01 -3.20011884e-01 -1.38218427e+00
1.70384839e-01 6.32794738e-01 -5.48211813e-01 -3.09845865e-01
-3.00225824e-01 -1.38532972e+00 6.10807896e-01 3.48883033e-01
4.97334242e-01 -6.34909153e-01 1.79842785e-01 4.19291221e-02
-3.94801944e-01 -1.26783884e+00 -1.82183489e-01 6.50578797e-01
-3.92764211e-01 -1.18066311e+00 -2.14807004e-01 -6.64681792e-01
5.64153433e-01 -8.33740309e-02 1.18804204e+00 -1.22426309e-01
8.95256475e-02 -1.24176227e-01 -4.12610620e-01 3.02718766e-02
-3.72694910e-01 5.21782577e-01 -1.06225364e-01 1.27621338e-01
4.45613891e-01 -5.62651217e-01 -2.18466058e-01 3.08151037e-01
-8.54733467e-01 -3.60346697e-02 7.10032880e-01 1.13609338e+00
5.61039507e-01 3.01773489e-01 1.06725824e+00 -1.37967288e+00
6.78056180e-01 -8.01119745e-01 -3.54502559e-01 8.33830535e-01
-1.15207434e+00 3.56782258e-01 5.39950371e-01 -4.86826301e-01
-1.17672491e+00 1.16478309e-01 -5.82099482e-02 6.83196262e-02
4.47912551e-02 6.66656256e-01 -5.99502623e-01 2.38520488e-01
5.32839596e-01 -3.21873277e-01 -4.14592832e-01 -5.07008016e-01
4.10643607e-01 9.33129549e-01 3.55409235e-01 -8.32057655e-01
6.44922793e-01 1.61981061e-01 2.90325314e-01 -1.00174129e-01
-1.39350843e+00 -5.43461144e-01 -7.04654098e-01 8.31543505e-02
6.03287280e-01 -9.14268553e-01 -8.32830906e-01 3.21036249e-01
-1.01859736e+00 1.38475657e-01 -1.10032223e-01 3.01043868e-01
-1.65624976e-01 3.44965249e-01 -6.15287244e-01 -7.57705629e-01
-3.88444662e-01 -9.57150698e-01 1.04242682e+00 2.02363610e-01
5.80726266e-02 -1.04443288e+00 1.92855194e-01 3.59855145e-01
-9.19055119e-02 3.09585631e-01 1.05865979e+00 -1.16674852e+00
-2.56441414e-01 -1.93340927e-01 -5.28181493e-01 4.18155491e-01
2.13205889e-01 -2.37871990e-01 -1.08516335e+00 -1.64726198e-01
-2.47484103e-01 -5.89553714e-01 1.20693839e+00 -2.04809397e-01
1.10968697e+00 -1.92774951e-01 -6.23386264e-01 2.31736630e-01
1.40273666e+00 -1.93193629e-02 3.89224410e-01 2.56553292e-01
7.75885940e-01 7.54328370e-01 8.96670461e-01 4.07274127e-01
6.79386258e-01 8.04168344e-01 2.24787682e-01 -4.75744531e-02
1.32213250e-01 -2.30779380e-01 -1.54461950e-01 6.29735053e-01
7.02790841e-02 -5.77595651e-01 -1.00272441e+00 4.49577957e-01
-2.03461623e+00 -6.38382316e-01 -9.02885385e-03 1.96955597e+00
1.33680141e+00 1.94410339e-01 -2.31552243e-01 4.13602322e-01
8.41977954e-01 1.70065984e-01 -2.40421697e-01 6.87360093e-02
-3.51814449e-01 7.08925053e-02 4.27056760e-01 2.69928664e-01
-1.53983366e+00 7.80477345e-01 5.14530182e+00 1.08246815e+00
-8.11684608e-01 2.13097930e-01 8.30415964e-01 2.80716568e-01
-3.88522208e-01 -7.33339414e-02 -1.19108593e+00 3.32151473e-01
8.39095354e-01 1.06449924e-01 -5.90894036e-02 8.30111623e-01
-2.71777451e-01 3.31274569e-02 -1.18478703e+00 7.04731643e-01
-1.37336245e-02 -9.06420946e-01 -3.14633787e-01 1.20012268e-01
8.77235472e-01 -2.94857055e-01 -1.68434769e-01 6.10915422e-01
4.22849238e-01 -9.67929006e-01 3.17880750e-01 4.97128636e-01
7.19067216e-01 -6.32130682e-01 1.07253981e+00 3.87360662e-01
-1.08912826e+00 -2.63576537e-01 -1.09887563e-01 6.79276139e-02
6.83282465e-02 1.22250021e+00 -8.29665244e-01 9.49594915e-01
2.09853262e-01 6.72645152e-01 -8.01444530e-01 7.49197841e-01
-2.58219808e-01 2.71673083e-01 -1.33860394e-01 3.90057452e-02
-1.43006578e-01 -9.63138044e-03 1.60891190e-01 1.23597836e+00
1.12421721e-01 2.19427288e-01 2.79638529e-01 5.39375365e-01
-4.72291321e-01 2.20222816e-01 -4.77155924e-01 7.61839375e-02
7.45145917e-01 1.60281646e+00 -4.89582270e-01 -1.34636760e-01
-2.73517877e-01 5.81533551e-01 9.06729817e-01 2.37623662e-01
-8.26046586e-01 -2.66360283e-01 3.98089826e-01 -1.60296530e-01
1.32915169e-01 6.74496144e-02 -3.35997730e-01 -1.30463982e+00
5.24383858e-02 -5.90348959e-01 5.92374504e-01 -2.55551338e-01
-1.71312809e+00 5.77269316e-01 -5.84999053e-03 -1.08233273e+00
-2.42074460e-01 -3.73554647e-01 -7.07653686e-02 7.22520888e-01
-1.73241270e+00 -1.44621170e+00 -1.43014371e-01 1.63444191e-01
-9.83455107e-02 7.25546032e-02 8.03669393e-01 7.29639828e-01
-8.22422862e-01 9.03283000e-01 8.80884752e-02 1.25625402e-01
9.39117789e-01 -1.34502852e+00 -1.70873210e-01 5.52044630e-01
-8.03369209e-02 4.88202542e-01 3.73010248e-01 -6.62035108e-01
-8.15233290e-01 -1.29113626e+00 1.16736698e+00 -3.38171363e-01
5.19174814e-01 -4.92301673e-01 -9.75552678e-01 6.67398512e-01
-3.01466584e-01 3.93971592e-01 7.39388466e-01 5.26462436e-01
-5.16337812e-01 -3.64349544e-01 -1.35566676e+00 2.52271056e-01
9.67336535e-01 -5.29602945e-01 -5.56612074e-01 4.39700186e-01
6.61628544e-01 -2.95990914e-01 -1.18860281e+00 1.06390500e+00
4.18743134e-01 -5.94156563e-01 9.03764665e-01 -6.31518245e-01
3.95938963e-01 -1.80038035e-01 -1.92801535e-01 -1.30269837e+00
-2.94241309e-01 -2.57614613e-01 -2.19667807e-01 1.91357887e+00
6.42533600e-01 -7.66730070e-01 4.69526201e-01 6.08310461e-01
1.93087533e-01 -1.17305577e+00 -7.47158408e-01 -7.34769940e-01
1.00890905e-01 -6.50952756e-02 6.35070562e-01 1.23670173e+00
1.33506507e-01 6.10658407e-01 -4.34895158e-01 4.06167120e-01
4.34134722e-01 1.72133535e-01 2.18341708e-01 -1.60769510e+00
-2.85465777e-01 -2.73932982e-02 -1.68012470e-01 -7.46862531e-01
5.36578774e-01 -9.87136960e-01 4.47955541e-03 -1.31853831e+00
4.68611628e-01 -1.08945310e+00 -6.64910197e-01 6.48850858e-01
-5.45236290e-01 1.18440330e-01 -9.38489160e-04 1.30925313e-01
-7.34043777e-01 6.74436092e-01 1.06495118e+00 -1.42876104e-01
-7.11975396e-02 2.13110149e-01 -8.84140253e-01 5.02521396e-01
6.48761570e-01 -6.37712538e-01 -4.58147854e-01 1.40899509e-01
5.21958530e-01 5.82890585e-02 1.66379109e-01 -6.52763665e-01
3.39148901e-02 -1.38715774e-01 1.72154456e-01 -5.29218733e-01
1.76521644e-01 -8.07966113e-01 3.44925523e-02 8.02233908e-03
-6.23980105e-01 -3.47133398e-01 -2.16858163e-01 7.02803493e-01
-1.92281052e-01 -2.13679016e-01 7.68968701e-01 1.59221590e-01
-3.01213950e-01 1.71123341e-01 2.46885344e-01 3.69203508e-01
1.07490349e+00 5.23327231e-01 -6.73805118e-01 7.18180314e-02
-7.15991557e-01 5.29269814e-01 6.65136660e-03 4.02902246e-01
1.46056756e-01 -1.33294129e+00 -6.51373267e-01 3.90105369e-03
3.61795276e-01 7.02787787e-02 1.59692898e-01 9.05160546e-01
1.72840580e-01 5.32121956e-01 3.17757577e-01 -2.59052068e-01
-1.23084199e+00 5.67892492e-01 3.22953999e-01 -1.07543075e+00
-1.72228709e-01 7.55880594e-01 2.01913202e-03 -6.50918067e-01
2.01742947e-01 -2.49882981e-01 -6.15209877e-01 3.94609302e-01
1.84358343e-01 2.96366930e-01 3.47997874e-01 -4.51499999e-01
-4.75436151e-01 3.18501383e-01 -3.27924162e-01 3.26119363e-01
1.18173242e+00 -1.44336358e-01 -2.39569634e-01 3.53012294e-01
1.14999139e+00 1.65574968e-01 -1.00091219e+00 -4.18096930e-01
3.16100419e-01 -1.97070777e-01 9.04796273e-03 -1.11216331e+00
-9.42238152e-01 5.78947604e-01 4.36319560e-01 3.27930778e-01
1.02089846e+00 2.41572991e-01 5.59461296e-01 2.85437107e-01
2.96456486e-01 -9.33931530e-01 -1.54801086e-01 5.09070516e-01
5.25116086e-01 -1.39126039e+00 1.90735385e-01 -8.90288174e-01
-5.84164619e-01 8.05151224e-01 6.27932668e-01 3.88481826e-01
7.90573359e-01 3.05636704e-01 -1.16184391e-01 -1.86951578e-01
-8.68676662e-01 -2.92681634e-01 4.38605368e-01 3.19558471e-01
5.44502556e-01 1.96971819e-01 -6.90684855e-01 1.09282196e+00
9.97299850e-02 -1.12093419e-01 2.66402841e-01 5.86995959e-01
5.28023113e-03 -1.43903244e+00 -5.82912192e-02 4.86381263e-01
-5.53761959e-01 -1.34795532e-01 -2.24226400e-01 6.60046041e-01
4.97027874e-01 8.64029408e-01 -3.91760081e-01 -5.17148435e-01
2.11752221e-01 3.51782829e-01 2.43269354e-01 -6.60542905e-01
-5.53666770e-01 -2.30583593e-01 5.30937254e-01 -7.01077729e-02
-6.08910680e-01 -5.76557100e-01 -1.13598132e+00 8.10159370e-02
-8.87323916e-01 2.24647984e-01 4.17663842e-01 1.12774456e+00
4.49205995e-01 6.81984544e-01 6.98667467e-01 -2.31458694e-01
-8.98110390e-01 -1.14054489e+00 -7.53642976e-01 4.93004322e-01
1.90538898e-01 -1.02944446e+00 -3.22503179e-01 -3.68719459e-01] | [9.182272911071777, 8.61283016204834] |
f098a9b5-6d1e-422f-821b-0d5dffff560a | a-survey-on-aspect-based-sentiment-analysis | 2203.01054 | null | https://arxiv.org/abs/2203.01054v2 | https://arxiv.org/pdf/2203.01054v2.pdf | A Survey on Aspect-Based Sentiment Analysis: Tasks, Methods, and Challenges | As an important fine-grained sentiment analysis problem, aspect-based sentiment analysis (ABSA), aiming to analyze and understand people's opinions at the aspect level, has been attracting considerable interest in the last decade. To handle ABSA in different scenarios, various tasks are introduced for analyzing different sentiment elements and their relations, including the aspect term, aspect category, opinion term, and sentiment polarity. Unlike early ABSA works focusing on a single sentiment element, many compound ABSA tasks involving multiple elements have been studied in recent years for capturing more complete aspect-level sentiment information. However, a systematic review of various ABSA tasks and their corresponding solutions is still lacking, which we aim to fill in this survey. More specifically, we provide a new taxonomy for ABSA which organizes existing studies from the axes of concerned sentiment elements, with an emphasis on recent advances of compound ABSA tasks. From the perspective of solutions, we summarize the utilization of pre-trained language models for ABSA, which improved the performance of ABSA to a new stage. Besides, techniques for building more practical ABSA systems in cross-domain/lingual scenarios are discussed. Finally, we review some emerging topics and discuss some open challenges to outlook potential future directions of ABSA. | ['Wai Lam', 'Lidong Bing', 'Yang Deng', 'Xin Li', 'Wenxuan Zhang'] | 2022-03-02 | null | null | null | null | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [-6.29758462e-02 -3.00169975e-01 -3.29777360e-01 -7.87377775e-01
-7.64059544e-01 -7.89102316e-01 6.56216860e-01 6.02278233e-01
-2.30769485e-01 5.15556991e-01 2.73799360e-01 -2.98732489e-01
-7.43941497e-03 -8.65247846e-01 -4.55136001e-02 -5.97813845e-01
3.02856743e-01 3.91698182e-01 -2.54729360e-01 -9.21775281e-01
5.42311370e-01 7.09253922e-02 -1.52307725e+00 7.53950119e-01
7.68811107e-01 1.30780351e+00 -3.38741511e-01 2.18150109e-01
-8.18317533e-01 4.86916006e-01 -9.85129058e-01 -1.13055313e+00
-1.01399019e-01 -1.38023406e-01 -8.56576145e-01 3.08263421e-01
3.38921845e-01 5.25258601e-01 6.97147667e-01 9.94624197e-01
3.80775303e-01 -3.13676894e-02 6.23904824e-01 -1.30727768e+00
-7.12308049e-01 6.55253708e-01 -6.07298434e-01 -1.28643811e-01
6.87141180e-01 -5.36123753e-01 1.60236704e+00 -1.16133797e+00
3.34395498e-01 1.12337017e+00 6.82541192e-01 3.96737784e-01
-4.34530824e-01 -3.16985369e-01 8.16770911e-01 1.04127988e-01
-8.39470804e-01 9.15627927e-02 1.01681161e+00 -3.32817376e-01
1.32079375e+00 4.33237582e-01 1.03022838e+00 8.94784212e-01
5.12638032e-01 1.15606403e+00 1.61539102e+00 -5.11831760e-01
1.78421348e-01 5.23231864e-01 7.88602471e-01 1.70037568e-01
4.35294777e-01 -7.53805280e-01 -9.15079951e-01 -2.68247128e-01
-3.45510125e-01 -2.20281079e-01 2.84045637e-01 -2.76535749e-01
-1.08540916e+00 9.54714894e-01 -3.89149450e-02 4.57978904e-01
-2.00796723e-01 -7.07080424e-01 7.74896979e-01 5.38651586e-01
9.06956673e-01 6.50761545e-01 -1.05296135e+00 -9.89871770e-02
-6.71985507e-01 4.26847398e-01 1.16144896e+00 8.94063234e-01
6.93150640e-01 1.54439226e-01 -4.71022055e-02 1.07924783e+00
4.16410595e-01 8.47212315e-01 5.72140157e-01 -1.86576873e-01
4.58708555e-01 1.14455760e+00 -1.04462802e-01 -1.20137382e+00
-6.48708224e-01 -4.28386867e-01 -6.15445018e-01 -5.89134321e-02
7.83266351e-02 -1.37169868e-01 -6.80677116e-01 1.07444012e+00
2.91327298e-01 -7.39849448e-01 3.67471278e-01 6.33995056e-01
1.02650905e+00 5.73693037e-01 5.83162121e-02 -1.27662405e-01
2.12152600e+00 -1.07432353e+00 -9.63032126e-01 -7.34182715e-01
6.06057167e-01 -1.41810465e+00 1.18139505e+00 5.83660126e-01
-9.24502969e-01 -3.41629148e-01 -9.73154426e-01 -2.61200130e-01
-1.29732656e+00 2.34783083e-01 1.06290174e+00 8.73311460e-01
-9.72447038e-01 -9.76515785e-02 -3.76431227e-01 -4.77844626e-01
1.07730947e-01 3.27483237e-01 -1.66526586e-01 1.25645921e-01
-1.48954415e+00 8.32957864e-01 -1.47313058e-01 2.03907609e-01
-2.91795433e-02 -4.62859660e-01 -1.24380732e+00 -2.95494676e-01
3.04809630e-01 -6.86858773e-01 1.27121186e+00 -1.04613996e+00
-1.50203454e+00 1.17656028e+00 -6.32201076e-01 -1.38473928e-01
-3.15673530e-01 -4.32081074e-01 -9.70268667e-01 -4.47817475e-01
4.09558147e-01 1.31151676e-01 7.87163496e-01 -9.51247811e-01
-9.79554653e-01 -6.04618907e-01 5.23414731e-01 4.49330300e-01
-6.46440029e-01 5.71798921e-01 -4.03720438e-01 -8.26092064e-01
-3.42910171e-01 -8.37267041e-01 -3.19008052e-01 -6.40488982e-01
-3.18464637e-01 -5.41392267e-01 7.41459012e-01 7.98135251e-03
1.56862545e+00 -1.97200131e+00 2.95003094e-02 1.45648718e-01
1.22507829e-02 2.26764679e-01 -1.26408502e-01 7.39706933e-01
-1.15587607e-01 -1.02909803e-01 -2.68700004e-01 -5.50801992e-01
1.38175100e-01 -3.79422307e-03 -6.65228486e-01 -4.69987877e-02
1.16807975e-01 9.25076962e-01 -7.52461791e-01 -3.17713439e-01
1.18070930e-01 5.08592725e-01 -3.39304775e-01 4.47750166e-02
-2.35557407e-01 -6.63033724e-02 -7.71830082e-01 1.13408399e+00
6.07749760e-01 -3.03640425e-01 1.43946990e-01 -2.73664594e-01
-3.94438237e-01 6.61707222e-01 -8.49917054e-01 1.24888349e+00
-8.89629483e-01 4.85197186e-01 7.93024004e-02 -9.91647005e-01
1.13844419e+00 2.31271654e-01 5.68835855e-01 -5.48587441e-01
4.49842393e-01 4.92338538e-01 -1.92098379e-01 -2.27657765e-01
1.15878034e+00 -5.24658084e-01 -8.18967283e-01 4.09629047e-01
-8.24940503e-02 -6.37626827e-01 7.22545624e-01 1.35463327e-01
3.50638479e-01 -7.82059357e-02 7.99342752e-01 -4.77267653e-01
1.22202015e+00 4.72779989e-01 2.42676690e-01 2.45365664e-01
-3.04162800e-01 4.67661470e-01 4.81968343e-01 -7.66802907e-01
-4.36009645e-01 -5.21068513e-01 -2.91618645e-01 1.33347011e+00
1.35471493e-01 -1.04801250e+00 -4.38368648e-01 -9.28298593e-01
-1.88610673e-01 5.70121229e-01 -8.56408119e-01 2.92131543e-01
-3.33903819e-01 -1.38316309e+00 -1.53413177e-01 3.98824036e-01
4.47495520e-01 -1.17274332e+00 6.70029083e-03 -1.33116422e-02
-3.41661155e-01 -1.24036372e+00 -3.19479138e-01 1.42739117e-01
-6.36621952e-01 -9.30940628e-01 -6.61678433e-01 -9.60634232e-01
4.76207942e-01 5.13279796e-01 1.64007771e+00 -3.10063392e-01
2.30326086e-01 4.13131207e-01 -8.66906166e-01 -1.00231051e+00
4.95883077e-02 4.04682398e-01 2.91232169e-02 1.18460178e-01
1.28819895e+00 -5.63559905e-02 -4.57105011e-01 2.72740901e-01
-8.68104219e-01 -3.00908566e-01 5.22418022e-01 4.42224383e-01
7.21857488e-01 1.04974449e-01 6.16940737e-01 -1.35115218e+00
1.10745537e+00 -4.30219352e-01 -2.72528231e-01 1.65100545e-01
-7.43920326e-01 -4.48606759e-01 8.05660486e-01 1.09822594e-01
-1.10907674e+00 -5.15326381e-01 -6.36600018e-01 5.85048974e-01
-5.59329912e-02 9.90053415e-01 -1.25064611e-01 5.62438965e-02
4.00728852e-01 1.73006698e-01 -3.01838338e-01 -2.48923779e-01
3.08712155e-01 9.90230441e-01 -2.66098171e-01 -5.14815092e-01
4.46779221e-01 6.78647876e-01 -2.06254065e-01 -8.93414199e-01
-1.81445909e+00 -9.55869675e-01 -4.96357262e-01 -1.42714798e-01
6.04168296e-01 -1.06480980e+00 -4.16911423e-01 8.16769600e-01
-9.36504543e-01 3.51833671e-01 -3.82001549e-01 2.27971256e-01
-2.36910731e-01 3.09203178e-01 -5.40150046e-01 -5.70412636e-01
-8.64125848e-01 -1.33962286e+00 1.35936439e+00 3.47223818e-01
-7.04396427e-01 -1.45626748e+00 2.60515511e-01 7.62650967e-01
5.98039508e-01 -2.31277853e-01 8.89642239e-01 -8.16712856e-01
9.71108973e-02 -5.37275791e-01 1.53047800e-01 5.12637258e-01
5.58717608e-01 -7.78636569e-03 -8.97457123e-01 -1.89523607e-01
3.60092402e-01 -3.13856721e-01 6.00365996e-01 4.47614133e-01
6.58721447e-01 5.89594729e-02 -1.15298845e-01 2.14998186e-01
1.18363416e+00 2.51552045e-01 1.62810937e-01 9.34173703e-01
6.03196084e-01 8.95762086e-01 1.24914622e+00 2.87936807e-01
8.19037437e-01 4.22689736e-01 1.38457850e-01 -2.08961386e-02
5.61244711e-02 2.92910695e-01 6.54127836e-01 1.78325605e+00
-1.64058223e-01 -3.57784867e-01 -5.97689092e-01 7.92798877e-01
-1.46167374e+00 -3.86621058e-01 -3.26451659e-01 1.56164730e+00
6.36950314e-01 2.45380253e-01 1.27469003e-01 1.13178343e-01
2.97283173e-01 8.00985396e-01 -3.72651607e-01 -1.14739370e+00
-6.23934507e-01 1.38151050e-01 -1.40191063e-01 5.25665522e-01
-1.35094452e+00 1.20247841e+00 6.46008968e+00 8.35276246e-01
-1.10956275e+00 -1.38744175e-01 5.28057098e-01 1.55616969e-01
-7.51174927e-01 -1.08379610e-01 -1.14834309e+00 6.71942532e-02
5.89515805e-01 -2.25360781e-01 -1.47006303e-01 1.34095073e+00
-7.75223821e-02 4.56067249e-02 -5.55172324e-01 6.99297726e-01
6.39012575e-01 -1.04017878e+00 6.63525581e-01 -2.03205556e-01
1.08752680e+00 -1.46172717e-02 4.21770096e-01 5.34620941e-01
1.40078077e-02 -5.97180605e-01 4.09181416e-01 -5.42423092e-02
4.62762594e-01 -8.67367089e-01 1.27763343e+00 -2.74546504e-01
-1.45541179e+00 2.47599483e-01 -2.50029087e-01 -4.74489719e-01
3.07365328e-01 9.28011298e-01 -1.47529736e-01 9.38287497e-01
9.78884041e-01 1.31350780e+00 -5.79836369e-01 4.08610940e-01
-4.04442102e-01 4.46088701e-01 1.24848410e-01 -6.34153426e-01
5.87121487e-01 -7.16373444e-01 5.30058265e-01 1.31097746e+00
2.65518427e-01 -2.28794068e-01 4.23452538e-03 -1.14684608e-02
2.02115208e-01 7.78256178e-01 -5.30327797e-01 -1.63368255e-01
-2.17099518e-01 1.67779315e+00 -7.69060194e-01 -4.18656111e-01
-1.10287380e+00 5.72698355e-01 -1.99682377e-02 2.05824539e-01
-3.31838280e-01 -6.04434252e-01 1.02425122e+00 -1.00531019e-01
1.41088486e-01 -1.24613494e-02 -5.69323182e-01 -1.62166286e+00
1.43293291e-01 -1.44639444e+00 6.32396042e-01 -5.77438295e-01
-1.62809610e+00 1.16386652e+00 -2.48195514e-01 -1.48023319e+00
-1.65036246e-01 -1.19410396e+00 -4.81379867e-01 7.14288294e-01
-1.85153413e+00 -1.33197105e+00 9.42488313e-02 5.39686680e-01
1.00619376e+00 -3.65376592e-01 9.92685080e-01 1.72049880e-01
-4.26186800e-01 3.23333174e-01 -1.29732087e-01 -5.84229343e-02
1.01546299e+00 -1.41468108e+00 6.02420568e-01 6.50207520e-01
1.97133664e-02 1.00152504e+00 7.47110784e-01 -4.90260303e-01
-1.34717655e+00 -7.44843245e-01 1.72581065e+00 -7.50952363e-01
1.23463154e+00 -4.90283906e-01 -5.94683230e-01 7.24569559e-01
6.40133202e-01 -5.06108165e-01 1.28535223e+00 8.48033786e-01
-4.08658803e-01 -3.11987698e-01 -6.76841378e-01 7.45131910e-01
3.73512983e-01 -6.20217323e-01 -8.33297610e-01 2.08371758e-01
4.93371785e-01 -2.08975494e-01 -7.78311610e-01 6.06962681e-01
5.56232214e-01 -9.21245158e-01 7.61337221e-01 -5.66595197e-01
5.40458322e-01 -4.69604939e-01 -2.41196454e-01 -1.53315210e+00
1.13952234e-02 -3.04467559e-01 1.82893276e-01 1.18829453e+00
7.56038189e-01 -6.94258034e-01 6.56233132e-01 3.81296337e-01
-1.28806040e-01 -1.10553002e+00 -5.15401006e-01 -2.71282077e-01
2.44174182e-01 -9.58822250e-01 6.77627742e-01 8.58664036e-01
4.62333232e-01 8.73956740e-01 -3.29879403e-01 -1.56300366e-01
2.15841204e-01 9.67863500e-01 5.63138247e-01 -8.99194539e-01
1.57200381e-01 -6.99715734e-01 -2.37228751e-01 -9.41496968e-01
1.15969613e-01 -7.17648745e-01 -3.49785984e-01 -1.70082998e+00
1.71717629e-01 -1.23990491e-01 -4.61451471e-01 2.29488254e-01
-4.99316067e-01 5.82654119e-01 -1.83521900e-02 -1.10614568e-01
-8.52792978e-01 4.88413870e-01 1.30052471e+00 -3.86974871e-01
-1.96174122e-02 5.39229751e-01 -1.41752386e+00 1.18528271e+00
8.29803050e-01 -2.06745178e-01 -3.79979283e-01 -3.78618419e-01
1.25252783e+00 -6.46420240e-01 -6.06910586e-01 -3.59767139e-01
8.20546746e-02 -1.05826885e-01 -9.90596414e-02 -9.46406484e-01
4.40137058e-01 -7.35285759e-01 -6.53177619e-01 -4.32403199e-03
-6.01576790e-02 3.93336356e-01 2.89273918e-01 1.78118259e-01
-1.11720335e+00 -1.66966379e-01 2.72240818e-01 -2.55245483e-03
-8.96945953e-01 1.74633473e-01 -8.44514847e-01 9.42517966e-02
8.21305990e-01 -3.41666341e-02 -1.38926059e-01 -3.39336157e-01
-4.60415095e-01 3.21107477e-01 1.59714505e-01 7.47343838e-01
3.89656007e-01 -1.03470445e+00 -5.24271071e-01 1.00897439e-01
6.60349727e-01 -1.76254302e-01 4.46837127e-01 6.76069379e-01
-1.87422395e-01 8.75503242e-01 4.54234295e-02 -1.69907331e-01
-1.36214447e+00 6.23506188e-01 -4.56499830e-02 -6.97312236e-01
1.05913602e-01 7.69399822e-01 3.10359478e-01 -8.21286917e-01
-2.06422552e-01 -2.32260704e-01 -1.11870337e+00 7.36753404e-01
6.40208781e-01 3.35524827e-02 4.64552164e-01 -1.02498281e+00
-6.99817538e-01 1.21369600e+00 -4.25092578e-01 1.47887051e-01
1.25581741e+00 -5.75913727e-01 -6.72282815e-01 8.02858770e-01
1.02150249e+00 6.97326660e-01 -1.75656766e-01 -2.21121147e-01
-6.07613707e-03 -1.05402127e-01 -2.03697652e-01 -1.05030417e+00
-1.10214996e+00 8.15628767e-01 -1.41072273e-01 7.44923353e-01
1.41692090e+00 9.20297652e-02 8.13774109e-01 3.20985794e-01
3.83773625e-01 -1.19936299e+00 -1.61692202e-01 1.00393772e+00
8.43159437e-01 -1.49533319e+00 2.61146843e-01 -5.88600218e-01
-1.18099141e+00 1.05996311e+00 3.99772346e-01 4.29894477e-02
1.08457398e+00 3.34723651e-01 8.74708772e-01 -5.13821900e-01
-5.96910834e-01 -2.97819138e-01 5.86325884e-01 3.69697332e-01
8.86967421e-01 1.85605764e-01 -7.52819836e-01 1.00184333e+00
-8.11448097e-01 -4.61844057e-01 4.24885124e-01 9.21308875e-01
-1.48704737e-01 -1.59938800e+00 -2.54654467e-01 4.66997564e-01
-9.45397973e-01 -3.83693933e-01 -6.88460410e-01 7.99694657e-01
-1.54884398e-01 1.33108425e+00 -2.19980478e-01 -2.09755361e-01
6.89990938e-01 -5.85905798e-02 -6.94227964e-02 -8.06864142e-01
-9.21389818e-01 9.59576294e-02 4.24908340e-01 -4.00633991e-01
-1.02782238e+00 -7.46452093e-01 -6.52271986e-01 -2.11770475e-01
-2.16160536e-01 5.72108567e-01 8.66122007e-01 1.24472952e+00
2.25080386e-01 4.21876907e-01 6.15610540e-01 -3.09930831e-01
1.94026202e-01 -8.84974897e-01 -7.78714478e-01 2.09708631e-01
2.95123458e-01 -2.12342054e-01 -4.37671989e-01 -1.59423351e-01] | [11.348712921142578, 6.751372337341309] |
eedddcb8-1c79-4f28-972d-d195d7a70925 | bertic-the-transformer-language-model-for-1 | null | null | https://aclanthology.org/2021.bsnlp-1.5 | https://aclanthology.org/2021.bsnlp-1.5.pdf | BERTić - The Transformer Language Model for Bosnian, Croatian, Montenegrin and Serbian | In this paper we describe a transformer model pre-trained on 8 billion tokens of crawled text from the Croatian, Bosnian, Serbian and Montenegrin web domains. We evaluate the transformer model on the tasks of part-of-speech tagging, named-entity-recognition, geo-location prediction and commonsense causal reasoning, showing improvements on all tasks over state-of-the-art models. For commonsense reasoning evaluation we introduce COPA-HR - a translation of the Choice of Plausible Alternatives (COPA) dataset into Croatian. The BERTić model is made available for free usage and further task-specific fine-tuning through HuggingFace. | ['Davor Lauc', 'Nikola Ljubešić'] | null | null | null | null | eacl-bsnlp-2021-4 | ['commonsense-causal-reasoning'] | ['natural-language-processing'] | [-1.86761335e-01 5.04606724e-01 -3.71372342e-01 -4.49562788e-01
-9.46364701e-01 -6.14634454e-01 1.21620929e+00 3.37776989e-01
-4.92917091e-01 1.15907705e+00 8.51777256e-01 -6.50476813e-01
-3.10909599e-01 -8.73472273e-01 -6.39251411e-01 -1.87584326e-01
1.46743162e-02 1.02357411e+00 2.79361218e-01 -5.94075143e-01
2.45733365e-01 1.52555406e-01 -9.57583964e-01 8.17208111e-01
7.79607058e-01 6.63750887e-01 -1.89783037e-01 4.24535900e-01
-7.24537745e-02 1.22267830e+00 -7.24106908e-01 -1.38872480e+00
-2.24254370e-01 -1.06705941e-01 -1.37316072e+00 -7.86757588e-01
2.42856324e-01 3.14480305e-01 -4.97507811e-01 9.20001864e-01
4.63459939e-01 2.17584014e-01 8.34459007e-01 -1.17783368e+00
-9.88887012e-01 1.60740209e+00 1.74658760e-01 3.65518421e-01
4.79825974e-01 -3.60315353e-01 1.62627494e+00 -6.74923420e-01
1.16499889e+00 1.57878351e+00 8.72967720e-01 4.81442481e-01
-1.02640343e+00 -6.60728574e-01 -2.39827707e-01 8.85620892e-01
-1.25499439e+00 -2.95876294e-01 7.11157858e-01 -1.78534195e-01
1.57491040e+00 2.52301365e-01 3.61544073e-01 1.76984715e+00
2.37731114e-01 8.53460789e-01 1.09450722e+00 -5.92709780e-01
2.12637559e-01 -3.41730654e-01 -8.34123641e-02 5.22654474e-01
3.87885034e-01 -1.69699490e-01 -6.87725186e-01 -5.82205474e-01
2.58508325e-01 -5.83938897e-01 3.06884974e-01 1.78428218e-01
-1.33452475e+00 8.91392291e-01 2.54829407e-01 4.28257436e-01
-6.08666003e-01 5.07438421e-01 5.05193412e-01 -2.99793575e-02
5.86989701e-01 4.81867373e-01 -7.87873447e-01 -5.12544692e-01
-9.27198231e-01 7.20305443e-01 1.14557612e+00 9.32123482e-01
-8.80785808e-02 -3.15902501e-01 -1.00749083e-01 1.14827812e+00
1.78714976e-01 5.74313343e-01 6.22755110e-01 -9.83791888e-01
9.38250721e-01 3.62220734e-01 1.91905230e-01 -7.99544394e-01
-3.69698167e-01 1.25533670e-01 -3.49861920e-01 -2.61607289e-01
6.20184958e-01 -2.46692434e-01 -5.99486113e-01 1.54556084e+00
1.99017897e-01 1.10415228e-01 3.23816836e-01 6.21368706e-01
6.70868933e-01 2.71833509e-01 6.92388892e-01 2.53183901e-01
1.70536840e+00 -5.28453290e-01 -8.41964841e-01 -2.51661509e-01
5.19074380e-01 -7.44277418e-01 9.86800909e-01 2.62071013e-01
-9.24988210e-01 1.52373448e-01 -6.52961671e-01 -4.49162006e-01
-9.21659827e-01 -6.27439097e-02 7.53731191e-01 4.37231869e-01
-3.87838155e-01 6.13889992e-01 -5.92569292e-01 -4.95670468e-01
5.99281013e-01 -4.64980304e-01 -2.75657177e-01 -3.97231728e-02
-2.19545341e+00 1.26679683e+00 6.69409752e-01 -4.08646077e-01
-6.17356777e-01 -7.96006203e-01 -1.01070166e+00 7.65137300e-02
2.79530048e-01 -5.09005785e-01 1.31491876e+00 -1.00136243e-01
-1.23028874e+00 1.23791003e+00 1.28865913e-01 -9.99660790e-01
7.25959659e-01 -6.18854880e-01 -9.51715291e-01 -1.95465490e-01
6.70403183e-01 2.61059254e-01 1.33357540e-01 -6.59502029e-01
-6.48260415e-01 -2.86826253e-01 6.47933111e-02 -8.57483670e-02
1.49222732e-01 4.53811795e-01 1.91995561e-01 -9.50820148e-01
-4.89408106e-01 -8.33152354e-01 -1.02629036e-01 -6.44193113e-01
-7.94265926e-01 -7.97076941e-01 4.44914848e-01 -9.28141117e-01
1.30516660e+00 -1.78301358e+00 -3.75343978e-01 4.72433083e-02
-4.19856548e-01 -4.26527113e-03 -8.31400678e-02 6.32312715e-01
-3.45595300e-01 4.01282728e-01 -3.23640071e-02 -2.54441611e-02
3.96495521e-01 5.04324675e-01 -4.95633960e-01 2.76419759e-01
3.83462578e-01 9.85019565e-01 -1.46397591e+00 -9.45478141e-01
2.83308685e-01 2.11382493e-01 -4.51078147e-01 -3.88110459e-01
-5.50944626e-01 -1.75678909e-01 -5.58405697e-01 6.06158912e-01
2.65979320e-01 6.38723979e-03 4.14579511e-01 1.06222384e-01
-3.21063306e-03 1.46581531e+00 -6.46565378e-01 1.67510045e+00
-6.94213152e-01 6.64326131e-01 -4.46045727e-01 -6.63779736e-01
5.73660076e-01 5.09424448e-01 2.09400505e-01 -7.06511497e-01
6.77218065e-02 3.61009598e-01 -1.42399356e-01 -6.57416224e-01
6.10556126e-01 -3.61028463e-01 -6.60642743e-01 1.18060552e-01
6.22650422e-02 -1.17493495e-01 5.31469584e-01 3.12780261e-01
1.36428046e+00 5.12212038e-01 8.45012963e-01 -3.53546411e-01
4.13632452e-01 5.32384694e-01 6.36697471e-01 5.58402121e-01
1.83036197e-02 3.40254307e-01 6.80851877e-01 -3.36505771e-01
-9.88636136e-01 -1.10231972e+00 -4.64154392e-01 1.10344946e+00
-2.94800639e-01 -7.02438772e-01 -5.82211435e-01 -1.13981116e+00
5.98180480e-02 2.19408441e+00 -8.56003284e-01 4.59668398e-01
-8.12382698e-01 -6.30123019e-01 1.37585270e+00 3.38517338e-01
7.05071032e-01 -1.41692913e+00 -6.72146261e-01 3.68400902e-01
-7.91200936e-01 -1.41347575e+00 2.18758509e-01 1.21248297e-01
-2.97633380e-01 -1.33915424e+00 -2.21389979e-01 -2.39662305e-01
-3.88683349e-01 -5.85660040e-01 1.53821695e+00 -4.68779743e-01
-1.93719700e-01 -1.86738133e-01 -4.53354388e-01 -4.44731683e-01
-5.97052634e-01 4.89344485e-02 -2.81261325e-01 -7.54080176e-01
7.04491854e-01 -7.33802795e-01 -1.06183328e-02 -2.45809063e-01
-3.60679358e-01 -1.15910269e-01 1.50638193e-01 8.80898952e-01
2.85387248e-01 1.37290403e-01 4.21100467e-01 -1.41617239e+00
6.85318649e-01 -8.81992817e-01 -1.00034654e-01 3.81936103e-01
-5.24185717e-01 1.86254829e-01 7.57861137e-01 -4.20422964e-02
-1.76975560e+00 -1.85275465e-01 -3.24782610e-01 4.33635376e-02
-4.59848136e-01 6.56829596e-01 -3.04824859e-01 1.03084850e+00
1.14877486e+00 -8.19573179e-02 -8.62511814e-01 -4.41518396e-01
1.01895225e+00 6.50265038e-01 8.96355987e-01 -9.18518603e-01
5.20878196e-01 3.36348891e-01 7.92918354e-02 -5.40569544e-01
-1.36408865e+00 -3.85636359e-01 -5.76557040e-01 7.21946359e-02
9.16255176e-01 -8.02716970e-01 -3.98570687e-01 5.42969070e-02
-1.74557328e+00 -5.08133173e-01 -3.06871295e-01 2.68764436e-01
-5.69380105e-01 1.15617532e-02 -5.82879543e-01 -7.48527825e-01
-2.35872298e-01 -1.28377959e-01 9.24090803e-01 -5.33511229e-02
-7.32029378e-01 -1.45354247e+00 2.52846956e-01 5.78555226e-01
2.42765788e-02 4.00928706e-01 1.31593454e+00 -1.33632052e+00
3.61891150e-01 -2.31667280e-01 -1.28074810e-01 -3.89852375e-02
-1.13850892e-01 -1.63107648e-01 -6.61051929e-01 8.12333584e-01
-5.14457464e-01 -3.44265074e-01 8.57584476e-01 8.01292900e-03
8.19193602e-01 -7.17233777e-01 -4.63410556e-01 -1.06796604e-02
1.21472621e+00 -9.52921510e-02 9.85582650e-01 7.00446844e-01
2.53012687e-01 4.20179844e-01 7.39098430e-01 5.70043623e-01
8.36101949e-01 4.58718538e-01 2.71372020e-01 4.81633663e-01
-6.16142094e-01 -7.29366302e-01 2.70846516e-01 1.63317159e-01
-3.43020588e-01 -4.58749324e-01 -1.31035221e+00 1.07574654e+00
-1.93624663e+00 -1.74609900e+00 -4.42623258e-01 1.57530570e+00
1.32466781e+00 2.25959256e-01 -2.02193588e-01 1.06802553e-01
7.37707078e-01 2.31190681e-01 4.26111035e-02 -4.97059524e-01
-3.41110110e-01 4.38239515e-01 6.54874384e-01 6.64705575e-01
-1.22501469e+00 1.38092363e+00 5.90573502e+00 1.28998041e+00
-4.16512191e-01 5.85581541e-01 3.50711823e-01 5.24630062e-02
-4.87504631e-01 3.00590634e-01 -7.05723107e-01 5.36203682e-01
1.20736194e+00 -1.96450159e-01 3.93129498e-01 1.10555851e+00
1.93704501e-01 -5.70291504e-02 -9.36709285e-01 6.87934995e-01
6.19030558e-02 -1.68068278e+00 -7.65513554e-02 -1.84861541e-01
6.23534083e-01 4.42536294e-01 -4.54290926e-01 5.16819298e-01
1.13730383e+00 -7.21379757e-01 1.25670385e+00 3.59291524e-01
5.55340052e-01 -5.80908954e-01 7.89718330e-01 1.02046706e-01
-6.02620661e-01 -6.49371818e-02 -1.08418301e-01 -4.17585559e-02
3.29758406e-01 9.90487695e-01 -9.89543140e-01 5.93343318e-01
8.03703070e-01 5.49872279e-01 -6.34777129e-01 5.11550128e-01
-1.27816379e+00 1.28115654e+00 -2.93676823e-01 -3.32129300e-01
2.12050021e-01 3.46678913e-01 8.05850744e-01 1.87583113e+00
1.16283529e-01 4.04848903e-01 -4.05133814e-01 1.16490114e+00
-3.81839752e-01 -1.18157648e-01 -3.17573428e-01 -5.65118939e-02
9.05706882e-01 1.13701010e+00 -3.91916275e-01 -5.27961791e-01
-7.27852434e-02 8.94276261e-01 4.89674479e-01 -8.87115076e-02
-1.08217800e+00 -6.83965981e-01 4.04308170e-01 2.49281153e-01
1.42857760e-01 1.21436991e-01 -5.34922779e-01 -1.17037022e+00
-3.64184767e-01 -4.08739656e-01 9.20404732e-01 -1.09828079e+00
-1.80394399e+00 3.05224955e-01 4.29676235e-01 -3.77442956e-01
-8.67346048e-01 -8.66715133e-01 -7.75173485e-01 5.47930479e-01
-1.40040016e+00 -1.30101430e+00 2.54515350e-01 4.67263639e-01
4.28338736e-01 -6.36678115e-02 1.00557172e+00 1.10355869e-01
-8.42791274e-02 2.33100787e-01 -1.26582578e-01 5.72516024e-01
9.05785501e-01 -1.52827144e+00 8.47194970e-01 7.89844811e-01
2.72481352e-01 5.95985711e-01 1.03670859e+00 -9.82237577e-01
-4.20276850e-01 -1.01280153e+00 2.20433450e+00 -6.69508040e-01
1.38294029e+00 -2.75169045e-01 -4.15294498e-01 7.18482018e-01
4.51767325e-01 -2.16744080e-01 7.56231666e-01 8.28339338e-01
-8.09224486e-01 4.29953605e-01 -1.27897036e+00 7.33384013e-01
1.25007319e+00 -5.09844840e-01 -1.41986322e+00 6.53825700e-01
4.38246280e-01 -2.38141522e-01 -8.79055381e-01 -1.98982298e-01
3.74668539e-01 -7.21732974e-01 8.07760656e-01 -1.04424345e+00
1.03873050e+00 2.44181201e-01 -4.26802427e-01 -1.34305441e+00
-2.84911394e-01 -7.35940814e-01 1.53375834e-01 1.52475739e+00
7.72997916e-01 -5.35347819e-01 3.73357475e-01 6.95724308e-01
-9.19897929e-02 -1.57189682e-01 -1.40583587e+00 -7.77421653e-01
4.21363145e-01 -9.35004950e-01 6.17130756e-01 1.32426369e+00
7.19976485e-01 5.95102489e-01 1.60476148e-01 -1.39217367e-02
6.79121614e-01 5.91528090e-03 7.99859390e-02 -1.24589479e+00
-4.61451411e-02 -3.75889599e-01 -1.87140673e-01 -3.48507941e-01
7.30068028e-01 -9.87854004e-01 -3.65245938e-02 -1.85310709e+00
8.50892439e-02 -6.49961233e-01 -6.42746389e-02 1.09438944e+00
3.37033533e-02 5.36353067e-02 1.18286908e-01 7.78552890e-02
-7.02383399e-01 2.82070220e-01 7.08737493e-01 1.42131122e-02
2.57563978e-01 -2.94840455e-01 -3.49641085e-01 9.12812173e-01
8.80286574e-01 -9.63977933e-01 2.02987697e-02 -3.66926730e-01
7.66520619e-01 1.78028986e-01 7.94112742e-01 -7.13291228e-01
-1.28327720e-02 -5.45005739e-01 1.56331718e-01 -5.08116722e-01
2.44790643e-01 -3.12806189e-01 -1.44778088e-01 3.22454780e-01
-8.30646276e-01 -5.96535616e-02 3.89628336e-02 4.71813083e-01
4.39194739e-02 -5.32055438e-01 4.26360875e-01 -2.76062757e-01
-7.32483447e-01 -5.73591530e-01 -4.91270453e-01 7.61834383e-01
6.37167215e-01 4.25679296e-01 -6.53461874e-01 -3.44728768e-01
-5.59031367e-01 -2.57123411e-01 -6.00140169e-02 5.56038201e-01
1.49470419e-01 -1.41653812e+00 -1.13211977e+00 -6.64490163e-01
1.10227302e-01 -5.27058423e-01 -1.05137102e-01 5.26870847e-01
-5.21904111e-01 7.82942176e-01 -7.56567568e-02 4.06619251e-01
-8.48495901e-01 3.62990022e-01 -5.17126312e-03 -7.05462039e-01
-3.24524820e-01 5.58551490e-01 -5.87911010e-01 -5.95570624e-01
-3.66804123e-01 -1.27000749e-01 -1.90082103e-01 -6.03785105e-02
2.14364737e-01 5.51412880e-01 1.01760663e-01 -5.61490834e-01
-7.82242954e-01 -4.67692018e-01 2.38084659e-01 -5.24380386e-01
1.45540690e+00 5.76411113e-02 -3.23713779e-01 2.97254831e-01
6.05360329e-01 3.72640967e-01 -4.79065061e-01 5.16376682e-02
5.02325594e-01 -8.83682966e-02 -3.48515548e-02 -1.67226315e+00
-3.69526386e-01 6.82570159e-01 -3.34294409e-01 1.78752333e-01
4.45359021e-01 3.13035756e-01 6.30965292e-01 6.10362589e-01
4.25166160e-01 -1.39278984e+00 -4.07583117e-01 9.96015608e-01
9.49178517e-01 -7.00233161e-01 -6.61216080e-02 -4.56768483e-01
-8.93525898e-01 1.06718409e+00 5.54031841e-02 -2.76728332e-01
6.25300407e-01 3.56677979e-01 -1.59159899e-01 -3.06304127e-01
-9.80175912e-01 -2.59074628e-01 9.75348502e-02 7.94934869e-01
8.03450346e-01 4.94407535e-01 -7.86570489e-01 1.08626783e+00
-7.89197564e-01 -6.16018698e-02 4.76912796e-01 4.87822950e-01
7.91879520e-02 -9.39504802e-01 -3.19240093e-01 1.72607794e-01
-9.00209486e-01 -8.40226233e-01 -6.31855130e-01 1.19268465e+00
3.20213199e-01 1.12166643e+00 -7.49089792e-02 -6.27358332e-02
1.92443058e-01 4.50715750e-01 4.75936830e-01 -5.25215149e-01
-7.04509616e-01 -3.09204161e-01 1.23479831e+00 -4.47490752e-01
-3.04473341e-01 -1.08608925e+00 -1.69125617e+00 -4.25862551e-01
-1.32768497e-01 1.67875350e-01 6.19247735e-01 1.09475136e+00
2.19761744e-01 3.86647254e-01 -2.22738534e-01 -3.00141424e-01
-6.53242946e-01 -1.20880246e+00 -3.80040318e-01 6.93000257e-01
-5.18827796e-01 -5.18442452e-01 -2.25041792e-01 2.51082391e-01] | [9.87158489227295, 8.195725440979004] |
541506fd-5ddd-48e8-8d76-9b7832fc551d | learning-action-completeness-from-points-for | 2108.05029 | null | https://arxiv.org/abs/2108.05029v1 | https://arxiv.org/pdf/2108.05029v1.pdf | Learning Action Completeness from Points for Weakly-supervised Temporal Action Localization | We tackle the problem of localizing temporal intervals of actions with only a single frame label for each action instance for training. Owing to label sparsity, existing work fails to learn action completeness, resulting in fragmentary action predictions. In this paper, we propose a novel framework, where dense pseudo-labels are generated to provide completeness guidance for the model. Concretely, we first select pseudo background points to supplement point-level action labels. Then, by taking the points as seeds, we search for the optimal sequence that is likely to contain complete action instances while agreeing with the seeds. To learn completeness from the obtained sequence, we introduce two novel losses that contrast action instances with background ones in terms of action score and feature similarity, respectively. Experimental results demonstrate that our completeness guidance indeed helps the model to locate complete action instances, leading to large performance gains especially under high IoU thresholds. Moreover, we demonstrate the superiority of our method over existing state-of-the-art methods on four benchmarks: THUMOS'14, GTEA, BEOID, and ActivityNet. Notably, our method even performs comparably to recent fully-supervised methods, at the 6 times cheaper annotation cost. Our code is available at https://github.com/Pilhyeon. | ['Hyeran Byun', 'Pilhyeon Lee'] | 2021-08-11 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Lee_Learning_Action_Completeness_From_Points_for_Weakly-Supervised_Temporal_Action_Localization_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Lee_Learning_Action_Completeness_From_Points_for_Weakly-Supervised_Temporal_Action_Localization_ICCV_2021_paper.pdf | iccv-2021-1 | ['weakly-supervised-action-localization', 'weakly-supervised-temporal-action'] | ['computer-vision', 'computer-vision'] | [ 5.18629670e-01 6.98659867e-02 -7.72460938e-01 -2.21150637e-01
-9.36165333e-01 -3.79365474e-01 4.20369834e-01 5.32607660e-02
-3.25158954e-01 8.16279590e-01 4.92640555e-01 1.69883355e-01
1.05774784e-02 -3.49571735e-01 -7.67066598e-01 -6.63806498e-01
-1.36316419e-01 3.77714813e-01 5.44164777e-01 1.72749326e-01
1.96384117e-01 1.26635075e-01 -1.51994753e+00 5.59055507e-01
7.01284409e-01 1.13624620e+00 1.15872048e-01 2.56158024e-01
3.45057905e-01 1.34616065e+00 -4.06090021e-01 -2.91595519e-01
3.93740207e-01 -5.89419901e-01 -9.86867368e-01 4.88311172e-01
2.90204585e-01 -6.33629739e-01 -5.01626968e-01 7.33664751e-01
3.06959867e-01 3.27436239e-01 3.12513918e-01 -1.31538785e+00
-1.77844971e-01 6.14383757e-01 -6.55351996e-01 1.00299940e-01
4.94035870e-01 3.04895282e-01 1.35012424e+00 -8.19874823e-01
7.19725966e-01 9.68348920e-01 5.57940185e-01 6.56815529e-01
-1.05685556e+00 -4.88867879e-01 4.31075156e-01 5.80320954e-01
-1.35907555e+00 -5.68399012e-01 6.33248210e-01 -1.93454191e-01
8.20495129e-01 1.42257765e-01 6.35402620e-01 1.18374789e+00
-1.89218283e-01 1.23947906e+00 9.30488288e-01 -2.20911995e-01
3.16286534e-01 -5.56098282e-01 -1.97321996e-01 7.18690813e-01
-2.28325188e-01 -4.23598811e-02 -7.98700988e-01 -9.34389010e-02
6.58290088e-01 2.49391139e-01 -3.01740438e-01 -3.96448284e-01
-1.45689976e+00 4.36663240e-01 1.95305243e-01 1.72494501e-01
-5.00175953e-01 4.64478344e-01 5.35527408e-01 -1.13437094e-01
4.42581534e-01 1.04220495e-01 -5.30370235e-01 -6.67078853e-01
-7.78193116e-01 1.52771175e-01 4.44604069e-01 1.10704565e+00
6.52095735e-01 -4.25089210e-01 -6.87511861e-01 6.87962830e-01
-2.95951851e-02 5.00103571e-02 2.07032189e-01 -1.43946695e+00
6.44104779e-01 6.37294114e-01 3.80192906e-01 -6.66801751e-01
-2.12284610e-01 -2.60768175e-01 -5.49573362e-01 -1.54672176e-01
4.81144726e-01 1.68765798e-01 -7.85126090e-01 1.74314475e+00
4.53345686e-01 8.80097449e-01 -1.41768202e-01 8.60836208e-01
3.13982397e-01 4.51932579e-01 1.01151370e-01 -3.22751313e-01
1.13217080e+00 -1.38858342e+00 -6.11372232e-01 -2.57400870e-01
1.04943097e+00 -3.93745959e-01 1.14787281e+00 3.54716897e-01
-1.02218688e+00 -5.55283129e-01 -6.01161301e-01 1.49958491e-01
2.00460479e-01 4.97338384e-01 5.72560966e-01 1.61438540e-01
-7.35177636e-01 8.74784768e-01 -1.22162688e+00 -2.35926896e-01
8.86217177e-01 7.81796798e-02 -2.55797476e-01 -1.72841474e-01
-8.92849624e-01 5.51215053e-01 6.16350472e-01 -7.86511078e-02
-1.22311664e+00 -7.01166689e-01 -8.22008550e-01 -4.77914438e-02
1.02695191e+00 -2.61469632e-01 1.48920286e+00 -9.37121212e-01
-1.36985576e+00 6.33258879e-01 -3.29648405e-01 -7.54383385e-01
7.74361491e-01 -4.68912393e-01 -5.90961687e-02 4.05711979e-01
2.70470768e-01 9.82121706e-01 5.09385228e-01 -1.01697624e+00
-1.13001025e+00 1.70308724e-02 4.78630662e-01 3.29525411e-01
-2.33720168e-01 5.05428612e-02 -8.46776366e-01 -5.64487636e-01
7.45938392e-04 -9.76168394e-01 -3.47264171e-01 1.15085058e-01
-4.54627156e-01 -5.13042629e-01 5.17403066e-01 -5.60877979e-01
1.44606888e+00 -2.23969126e+00 6.59836084e-02 -6.75131530e-02
1.52215868e-01 1.63152277e-01 -9.21202898e-02 4.46697474e-01
4.07345742e-02 -6.84300065e-02 -2.71046489e-01 -4.80348706e-01
9.95682627e-02 3.63561511e-01 -7.85961747e-02 4.95248288e-01
1.73190773e-01 9.54466999e-01 -1.17313612e+00 -6.50025845e-01
2.61106908e-01 2.32195318e-01 -6.42670214e-01 8.15228894e-02
-4.70189273e-01 8.10756326e-01 -6.13234639e-01 8.72307122e-01
1.47899568e-01 -5.94322085e-01 2.32882306e-01 -1.12250566e-01
1.65515736e-01 3.90314847e-01 -1.04251659e+00 2.03164768e+00
-1.64704740e-01 2.31575325e-01 -4.97725725e-01 -1.11296690e+00
5.36171794e-01 3.95084381e-01 1.03496802e+00 -5.70026100e-01
-5.18059707e-04 4.50696424e-02 -1.84606314e-01 -4.05962408e-01
1.79059133e-01 1.85449392e-01 3.05060297e-02 3.80980045e-01
-1.20364763e-01 4.82217342e-01 6.00055695e-01 1.76525250e-01
1.46377802e+00 7.73843408e-01 3.87060374e-01 3.07130516e-01
4.88997310e-01 -8.73198360e-02 8.61743987e-01 6.61550581e-01
-4.23297167e-01 6.07411921e-01 6.67107940e-01 -4.86459464e-01
-8.89588714e-01 -6.53796196e-01 1.56398639e-01 1.18764627e+00
3.44383419e-01 -7.13903666e-01 -7.65715003e-01 -1.05601192e+00
-3.00039291e-01 5.51236510e-01 -6.30247951e-01 -7.04179406e-02
-7.41589069e-01 -3.38770181e-01 5.41667938e-01 7.85561204e-01
5.64157784e-01 -1.30752361e+00 -8.31411660e-01 2.69646555e-01
-6.01216495e-01 -1.34776521e+00 -7.09572494e-01 8.53207987e-03
-9.02103364e-01 -1.26567686e+00 -6.42879844e-01 -5.43519914e-01
6.30266190e-01 1.56110495e-01 1.14764738e+00 2.65444666e-01
-1.55866832e-01 2.37310767e-01 -8.22142363e-01 2.29660328e-02
-2.91068316e-01 -9.71916877e-03 -7.37995431e-02 8.06304514e-02
3.29071790e-01 -4.93711859e-01 -8.59865665e-01 6.62724137e-01
-7.43149042e-01 2.58755535e-01 5.76788545e-01 6.40302896e-01
1.02825129e+00 5.14806025e-02 4.39217061e-01 -7.28022754e-01
-7.92110860e-02 -3.70840400e-01 -3.18358481e-01 3.08990687e-01
-3.98227215e-01 -1.44607350e-01 4.57717091e-01 -3.80712837e-01
-9.09163237e-01 4.68313247e-01 4.42820191e-02 -6.68972552e-01
-3.68181914e-01 3.32605511e-01 -1.29432097e-01 2.99917787e-01
5.00088513e-01 3.69288087e-01 -1.76959634e-01 -5.18660367e-01
2.50521183e-01 2.41482139e-01 4.99134779e-01 -6.52814150e-01
4.76286381e-01 6.75561965e-01 -1.59741268e-01 -2.99065024e-01
-1.21539915e+00 -6.42847300e-01 -7.76205659e-01 -3.35129678e-01
7.30375230e-01 -9.92215216e-01 -6.73880041e-01 3.50700319e-01
-9.50306356e-01 -6.52740836e-01 -5.23598671e-01 3.98271710e-01
-1.04814684e+00 6.55197978e-01 -6.16009653e-01 -6.82141781e-01
-2.12902967e-02 -1.00644338e+00 1.31445730e+00 -7.76969269e-02
-2.04975665e-01 -6.49694622e-01 -2.89584231e-02 6.08750165e-01
-2.37326607e-01 3.79912257e-01 3.98072958e-01 -6.02090061e-01
-6.88564181e-01 -1.09335624e-01 3.30722071e-02 3.61080587e-01
2.42456526e-01 -2.89900213e-01 -6.68141842e-01 -2.08575055e-01
-4.56573069e-01 -6.45097852e-01 8.91533673e-01 3.91072214e-01
1.55606055e+00 -3.02401960e-01 -5.12071252e-01 3.70284557e-01
1.08271086e+00 1.03925548e-01 8.32581282e-01 2.56830662e-01
6.48548663e-01 4.11724001e-01 1.36077666e+00 8.93860936e-01
2.67853200e-01 9.80117440e-01 6.10275030e-01 -3.76585079e-03
-1.63645342e-01 -4.59206164e-01 4.62370098e-01 2.83667058e-01
-4.12006050e-01 -3.43479097e-01 -6.51914716e-01 6.09434485e-01
-2.26986980e+00 -1.19394934e+00 8.24191049e-02 2.13639069e+00
1.02560246e+00 2.06061885e-01 4.66315389e-01 1.97256967e-01
6.76477313e-01 3.72588664e-01 -6.85767353e-01 5.00947356e-01
1.47160068e-01 8.20199996e-02 4.73865092e-01 1.57149792e-01
-1.40904653e+00 1.02410150e+00 5.37693167e+00 1.08479798e+00
-4.67545122e-01 2.55015492e-01 8.12942386e-01 -4.20613557e-01
1.82658464e-01 1.98727444e-01 -8.17617416e-01 6.19024515e-01
7.13341117e-01 8.25213566e-02 3.60407025e-01 8.42091501e-01
6.18152320e-01 -1.98438004e-01 -1.38214326e+00 8.28709602e-01
3.30022797e-02 -1.24812698e+00 -1.89741030e-01 -5.18193431e-02
8.02868903e-01 -5.35777956e-02 -2.74619877e-01 2.56091595e-01
2.15018108e-01 -7.41330802e-01 8.26951802e-01 4.36537176e-01
6.37229741e-01 -6.43495500e-01 5.02166271e-01 4.56083506e-01
-1.39116573e+00 -1.95325732e-01 -3.00858706e-01 -1.84511542e-01
4.24435019e-01 2.97453403e-01 -6.01310730e-01 6.21264040e-01
6.30130827e-01 1.24475408e+00 -3.96780401e-01 1.10645926e+00
-4.05911475e-01 7.85153925e-01 -1.76992625e-01 2.37144440e-01
3.88592064e-01 -5.57531305e-02 2.88425922e-01 9.12878513e-01
3.03522140e-01 2.56992221e-01 6.86472952e-01 4.94272619e-01
-1.28954992e-01 6.26545474e-02 -3.48204464e-01 1.58928987e-02
4.68867809e-01 9.96823013e-01 -8.24363708e-01 -5.31475961e-01
-5.30778348e-01 1.15612280e+00 3.40323806e-01 2.39746645e-01
-1.33074129e+00 9.63619351e-02 4.93313670e-01 1.73614815e-01
4.88897532e-01 9.41871628e-02 7.47448727e-02 -1.08914125e+00
3.52364689e-01 -9.11926150e-01 5.72678745e-01 -5.85362196e-01
-8.99537385e-01 2.62512833e-01 7.90726840e-02 -1.71746063e+00
-1.82746157e-01 -1.93956107e-01 -3.99954289e-01 2.27816179e-01
-1.37199926e+00 -1.14383924e+00 -3.80785018e-01 5.25407791e-01
9.68967378e-01 1.37188196e-01 5.14674306e-01 4.02161539e-01
-6.23587847e-01 5.17829418e-01 -2.03225464e-01 1.29892111e-01
6.20862067e-01 -1.04157436e+00 2.53518134e-01 7.40879893e-01
3.96525919e-01 9.34925303e-02 4.75441903e-01 -7.12446213e-01
-9.09619749e-01 -1.26356757e+00 7.50065386e-01 -5.48098505e-01
6.67899668e-01 -8.01627412e-02 -7.51892686e-01 8.79506528e-01
-1.11714087e-01 2.99135417e-01 5.39645612e-01 -1.25417054e-01
-1.55983523e-01 4.28640880e-02 -7.96305001e-01 6.09211922e-01
1.63548815e+00 -1.20032139e-01 -3.39504004e-01 7.54707098e-01
7.27031469e-01 -5.21859705e-01 -9.21771109e-01 5.55125177e-01
4.97148186e-01 -1.07562184e+00 9.00168717e-01 -6.61987126e-01
7.45058239e-01 -4.40289438e-01 -7.44280964e-02 -8.13032866e-01
-2.75077432e-01 -6.27998114e-01 -5.40285289e-01 1.10617495e+00
3.81488949e-01 -1.83745712e-01 1.09336674e+00 3.60048532e-01
-2.91495234e-01 -1.10903454e+00 -9.95592535e-01 -9.98958409e-01
-5.00395834e-01 -4.13811624e-01 4.51965868e-01 6.81571007e-01
2.04953831e-02 1.14558758e-02 -7.69132376e-01 7.64915943e-02
4.29830819e-01 1.23086520e-01 7.44270802e-01 -8.62870097e-01
-5.41906595e-01 -2.60767698e-01 -3.44531089e-01 -1.51664495e+00
3.19521278e-01 -6.70089602e-01 2.60369569e-01 -1.53470433e+00
4.70306635e-01 -4.20775950e-01 -5.68307281e-01 1.02084506e+00
-3.07202280e-01 4.48255241e-01 2.00356677e-01 3.70172024e-01
-1.34648991e+00 6.42911255e-01 1.27377081e+00 4.39151712e-02
-2.32360139e-01 2.19715267e-01 -3.99233878e-01 8.72404456e-01
8.37473273e-01 -5.54238677e-01 -5.40306628e-01 -2.01142326e-01
-1.22657746e-01 1.27001777e-01 5.22056937e-01 -1.23962843e+00
-1.37786055e-02 -4.01861757e-01 1.19381107e-01 -6.77155137e-01
3.96129280e-01 -7.45076120e-01 1.09423466e-01 4.98798549e-01
-5.83202779e-01 -3.82151663e-01 -1.62641376e-01 8.50127697e-01
-1.46955997e-01 -2.16471046e-01 6.82986677e-01 -2.74673942e-02
-9.57658470e-01 6.44857347e-01 1.37982899e-02 1.49018094e-01
1.34669113e+00 -2.20630288e-01 -1.25361323e-01 -3.72523129e-01
-8.41658950e-01 4.28618222e-01 5.26026428e-01 2.98942387e-01
4.28261071e-01 -1.39023995e+00 -6.32108271e-01 -3.35074961e-01
3.31785738e-01 5.77036478e-02 3.46303433e-01 1.28705871e+00
-1.92275971e-01 2.91588664e-01 4.67330627e-02 -7.31284022e-01
-1.24760759e+00 5.45631647e-01 1.16877690e-01 -4.64136451e-01
-1.01168084e+00 7.61939704e-01 2.71885812e-01 2.73232237e-02
6.13519073e-01 -4.14994329e-01 -6.59741983e-02 -1.93220168e-01
5.55533648e-01 5.34132481e-01 -2.28378713e-01 -6.12154305e-01
-4.63383198e-01 3.99844080e-01 1.14064496e-02 7.93190524e-02
1.25183320e+00 2.67617833e-02 2.26554930e-01 2.63834745e-01
8.72891128e-01 -2.92893499e-01 -2.00211453e+00 -4.03953314e-01
1.34182513e-01 -7.56506741e-01 -3.81686509e-01 -7.45042264e-01
-1.14483666e+00 4.72306341e-01 4.07088399e-01 -1.90143257e-01
1.37046695e+00 3.07202458e-01 8.38533938e-01 3.57530296e-01
5.19452274e-01 -1.15338063e+00 4.38086480e-01 2.54051834e-01
5.93822896e-01 -1.18373036e+00 3.23823281e-02 -5.31101763e-01
-8.94508004e-01 6.01596832e-01 7.26900399e-01 3.74069214e-02
2.11507618e-01 1.56396572e-02 -2.74949402e-01 -1.35068133e-01
-8.97126198e-01 -4.19622451e-01 1.16907239e-01 4.45450306e-01
2.54123211e-01 -1.39079811e-02 -5.06146371e-01 3.82259637e-01
3.16764653e-01 3.93739641e-01 2.83667475e-01 1.12150693e+00
-4.74937886e-01 -1.14495754e+00 4.70645390e-02 5.46667874e-01
-5.18123329e-01 -1.22110164e-02 -1.78868487e-01 5.63710988e-01
1.91125110e-01 8.31821859e-01 -4.00728695e-02 -3.09206694e-01
3.93900424e-01 1.38145797e-02 4.07490820e-01 -7.08421707e-01
-1.98422804e-01 2.15106785e-01 2.95709550e-01 -1.15094686e+00
-8.33030045e-01 -7.89244115e-01 -1.45978975e+00 3.43009643e-02
-2.30732054e-01 -4.23014201e-02 -8.29714388e-02 1.12554884e+00
4.50359762e-01 4.54007268e-01 6.27785504e-01 -8.31820667e-01
-5.97434580e-01 -9.61789012e-01 -4.55772817e-01 6.29054964e-01
-3.93829048e-02 -9.39226449e-01 -1.74689218e-01 3.99983853e-01] | [8.46605110168457, 0.5815978646278381] |
0b9d628e-5fa7-4e1f-b268-986dd76a653a | exploring-object-centric-temporal-modeling | 2303.11926 | null | https://arxiv.org/abs/2303.11926v2 | https://arxiv.org/pdf/2303.11926v2.pdf | Exploring Object-Centric Temporal Modeling for Efficient Multi-View 3D Object Detection | In this paper, we propose a long-sequence modeling framework, named StreamPETR, for multi-view 3D object detection. Built upon the sparse query design in the PETR series, we systematically develop an object-centric temporal mechanism. The model is performed in an online manner and the long-term historical information is propagated through object queries frame by frame. Besides, we introduce a motion-aware layer normalization to model the movement of the objects. StreamPETR achieves significant performance improvements only with negligible computation cost, compared to the single-frame baseline. On the standard nuScenes benchmark, it is the first online multi-view method that achieves comparable performance (67.6% NDS & 65.3% AMOTA) with lidar-based methods. The lightweight version realizes 45.0% mAP and 31.7 FPS, outperforming the state-of-the-art method (SOLOFusion) by 2.3% mAP and 1.8x faster FPS. Code has been available at https://github.com/exiawsh/StreamPETR.git. | ['Xiangyu Zhang', 'Ying Li', 'Tiancai Wang', 'Yingfei Liu', 'Shihao Wang'] | 2023-03-21 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [-1.61338910e-01 -4.26092446e-01 -2.64224023e-01 -1.97114274e-01
-1.04562128e+00 -6.81395829e-01 6.98047340e-01 1.05182491e-02
-5.48383474e-01 2.52093405e-01 4.37719189e-03 1.77096222e-02
2.75092125e-01 -7.63463259e-01 -8.45398366e-01 -4.00740385e-01
-3.75087070e-03 5.76177001e-01 1.03712606e+00 -2.69017704e-02
1.59795940e-01 5.87241650e-01 -1.79962623e+00 4.13545221e-01
3.04668725e-01 1.20261300e+00 3.96205962e-01 1.01221442e+00
-6.23237230e-02 7.83326805e-01 -2.17248619e-01 -3.05042922e-01
2.70011812e-01 1.18200101e-01 -6.98894262e-01 -1.05209768e-01
7.83337474e-01 -7.20025241e-01 -6.16158187e-01 5.52312672e-01
6.35837913e-01 4.71373796e-02 2.08774969e-01 -1.15793145e+00
-2.12588742e-01 1.28162861e-01 -9.08923745e-01 4.58055735e-01
4.01124060e-01 2.76219577e-01 1.03400218e+00 -1.50115037e+00
8.42892468e-01 1.27265108e+00 3.90415281e-01 2.49345452e-01
-1.04608440e+00 -5.16217053e-01 3.48964542e-01 3.42004865e-01
-1.56465721e+00 -6.41271293e-01 3.34128052e-01 -3.78303111e-01
9.81403649e-01 2.89276779e-01 7.32658446e-01 7.05683231e-01
1.37352049e-01 1.14276302e+00 6.79209173e-01 7.27750659e-02
-8.88961833e-03 -3.04408491e-01 -1.38015851e-01 6.91588521e-01
-1.80911347e-02 2.05286905e-01 -9.02027786e-01 -1.34919196e-01
6.76248372e-01 1.52778760e-01 -1.37249321e-01 -4.96630162e-01
-1.41951060e+00 4.23653930e-01 4.13763881e-01 -1.84087232e-01
-4.20674592e-01 5.63837945e-01 4.50298280e-01 -1.29579887e-01
7.57638991e-01 -2.91346580e-01 -3.93231004e-01 -4.68405992e-01
-1.02582228e+00 5.12794077e-01 5.07558286e-01 1.31057250e+00
6.76303267e-01 -3.73701677e-02 -1.53500929e-01 5.23859262e-01
3.26332182e-01 8.66008282e-01 -1.53823972e-01 -1.28519189e+00
3.48508388e-01 4.27055866e-01 3.29566270e-01 -6.88849509e-01
-1.27516806e-01 -3.14102441e-01 -3.20785582e-01 1.58168361e-01
1.95635498e-01 2.84207642e-01 -8.22908103e-01 1.17816949e+00
8.68781984e-01 3.47746611e-01 -3.22629303e-01 1.00681388e+00
9.36174691e-01 8.93499196e-01 -1.54285431e-01 -8.87276456e-02
1.22306299e+00 -1.21173775e+00 -4.36734319e-01 -4.56308872e-02
3.67459208e-01 -9.42488015e-01 1.05484653e+00 3.75049621e-01
-1.45290244e+00 -4.91250515e-01 -8.15510631e-01 -2.58429825e-01
-3.75531279e-02 -6.98819309e-02 5.41615129e-01 1.38825923e-01
-1.00030982e+00 3.64154398e-01 -1.26904678e+00 -3.39868486e-01
5.46251416e-01 2.14209095e-01 -8.27992260e-02 -3.13800573e-01
-5.64210713e-01 3.96650791e-01 1.30354360e-01 -1.99978307e-01
-1.29607558e+00 -1.13443637e+00 -6.35810137e-01 -1.27228901e-01
8.07799459e-01 -8.17609966e-01 1.67432618e+00 -2.68570900e-01
-1.25690103e+00 8.52915883e-01 -5.89670956e-01 -3.98509204e-01
8.78969371e-01 -6.69248343e-01 -8.12913030e-02 4.71779734e-01
3.14053625e-01 8.29693556e-01 5.23992896e-01 -1.22690761e+00
-1.01115108e+00 -3.17171901e-01 1.80378303e-01 1.97161689e-01
-1.86044816e-03 2.40321577e-01 -1.28082192e+00 -4.41198885e-01
1.56832933e-01 -1.05172145e+00 -1.38255030e-01 5.56100547e-01
-7.73435086e-02 -1.84445113e-01 1.08566344e+00 -3.23075294e-01
1.18157053e+00 -2.15100622e+00 -2.57717632e-02 -3.75189573e-01
3.08216929e-01 1.59596294e-01 6.23298734e-02 6.52984202e-01
4.71123934e-01 -1.35931909e-01 -9.34462249e-02 -7.77112901e-01
-1.28415346e-01 1.36303693e-01 -2.22785771e-01 5.53132057e-01
-1.05707906e-01 9.92905617e-01 -1.06920552e+00 -5.30538023e-01
3.77983779e-01 4.24538881e-01 -6.70887113e-01 1.48433685e-01
-4.41170603e-01 2.55412549e-01 -3.61892968e-01 1.11476219e+00
8.16318095e-01 -5.53662717e-01 -1.29702166e-01 -3.32984209e-01
-5.25951564e-01 2.82078803e-01 -1.11840451e+00 2.09197140e+00
-3.38472933e-01 5.51872611e-01 8.38499814e-02 -3.00644755e-01
4.30575222e-01 4.52413969e-02 8.92347276e-01 -5.54896355e-01
-8.07541162e-02 1.09077714e-01 -5.48009992e-01 -2.27339178e-01
8.49681914e-01 2.84999400e-01 1.35769412e-01 3.31261963e-01
-1.37029290e-01 -1.34328343e-02 3.35059583e-01 6.23999536e-01
1.16888011e+00 5.73593974e-01 2.09076256e-02 -4.17642519e-02
3.92038584e-01 1.72145426e-01 5.40080667e-01 6.29983962e-01
-2.01247737e-01 8.96232128e-01 1.15874149e-01 -5.29318392e-01
-9.68421996e-01 -1.31625044e+00 3.44484784e-02 1.07179749e+00
5.11305094e-01 -8.43014896e-01 -2.16546282e-01 -4.98740911e-01
1.51687190e-01 3.19357723e-01 -1.74520031e-01 3.36165071e-01
-6.81422770e-01 -3.82554322e-01 2.12930977e-01 7.47928917e-01
5.06725967e-01 -5.26967049e-01 -9.38386679e-01 3.04563046e-01
-2.77437568e-01 -1.55306697e+00 -7.54475117e-01 -4.00335521e-01
-8.49909961e-01 -9.50911522e-01 -5.80665171e-01 -2.69850343e-01
1.73418775e-01 7.92863667e-01 1.25860584e+00 -2.89828163e-02
-2.82776356e-01 7.01505721e-01 -4.59245533e-01 -4.25404608e-01
1.14891946e-01 1.60606261e-02 -3.43574286e-02 -1.17030647e-02
1.38288915e-01 -5.20073652e-01 -1.06440425e+00 5.34671426e-01
-7.76719272e-01 2.70295620e-01 3.79193813e-01 4.79484797e-01
1.21327579e+00 -4.98122692e-01 1.14564627e-01 -5.49218655e-01
-3.26962262e-01 -5.10246098e-01 -9.32757676e-01 -1.23277428e-02
-6.69206440e-01 -3.04654211e-01 1.30529433e-01 -3.67714673e-01
-1.00898731e+00 3.31763208e-01 -9.41923633e-02 -8.38516533e-01
5.55961765e-03 7.83494264e-02 1.37841374e-01 -7.06540048e-02
3.33529383e-01 3.61242175e-01 -2.02110246e-01 -5.82278371e-01
4.64897633e-01 3.03156972e-01 6.61662698e-01 -5.24555922e-01
7.20363021e-01 1.22972250e+00 -1.14867516e-01 -8.15724552e-01
-8.41530144e-01 -9.96058524e-01 -5.58517933e-01 -5.15568972e-01
8.43945146e-01 -1.25913286e+00 -8.30736756e-01 5.41520655e-01
-1.38202012e+00 -3.97962958e-01 -3.40712249e-01 4.30569619e-01
-6.66999340e-01 3.06701928e-01 -6.16759241e-01 -8.84294093e-01
-3.31226587e-01 -1.00465727e+00 1.62499177e+00 -1.08274728e-01
2.21166119e-01 -3.65130931e-01 -1.09264880e-01 4.11687464e-01
2.57070541e-01 2.25048646e-01 4.12209854e-02 -7.29877353e-02
-1.36689234e+00 -6.21300898e-02 -4.72925097e-01 1.94844604e-02
-2.31371909e-01 1.08759671e-01 -1.06866693e+00 -3.63048196e-01
-1.98499128e-01 -1.56819433e-01 8.72172236e-01 4.49860781e-01
9.79418457e-01 2.71680020e-02 -5.16889513e-01 8.68048608e-01
1.59547436e+00 1.94802329e-01 4.00813490e-01 3.30009460e-01
7.34305620e-01 1.78309128e-01 9.55455422e-01 9.03530419e-01
8.32037926e-01 9.29831624e-01 8.21343958e-01 2.08300322e-01
-3.36957544e-01 -3.64833623e-01 3.26199681e-01 7.10215747e-01
-2.56471068e-01 -4.45547253e-01 -1.02916646e+00 6.99678659e-01
-1.99921322e+00 -9.82457876e-01 -5.52131712e-01 2.07669830e+00
2.90388852e-01 2.33041972e-01 3.64676237e-01 -2.30760723e-01
4.18767899e-01 5.22627592e-01 -6.63334131e-01 2.37964198e-01
-2.42903382e-02 -6.99374974e-02 7.44501233e-01 5.01033366e-01
-1.08778405e+00 1.04001081e+00 5.54732656e+00 6.74197674e-01
-9.02362645e-01 4.06710595e-01 2.71132141e-01 -7.78222144e-01
-6.23753443e-02 1.10210486e-01 -1.17884004e+00 3.43560129e-01
9.43956435e-01 -1.94575876e-01 1.69710487e-01 7.84371436e-01
4.09635514e-01 -4.00750160e-01 -1.01309669e+00 1.18065381e+00
1.03212462e-03 -1.65645206e+00 -7.73303434e-02 1.73359856e-01
6.39250100e-01 6.59550071e-01 -1.74903840e-01 -6.11998849e-02
6.10829592e-02 -4.15457815e-01 1.16494691e+00 5.68962038e-01
8.36236179e-01 -6.87181234e-01 3.63153070e-01 3.79857957e-01
-1.82502925e+00 2.13274494e-01 -1.85610995e-01 9.40111205e-02
7.88580239e-01 4.85355556e-01 -6.04883015e-01 8.30067992e-01
1.09790897e+00 9.63092923e-01 -4.60818380e-01 1.18411791e+00
5.52024506e-02 5.44733286e-01 -6.82250917e-01 2.35910505e-01
2.26981759e-01 1.03441894e-01 8.03254664e-01 1.15405190e+00
3.17164809e-01 2.87261397e-01 4.06012326e-01 4.84601617e-01
-3.01579144e-02 -5.81360050e-02 -4.85571593e-01 2.31209904e-01
5.41207373e-01 1.07231820e+00 -5.62286556e-01 -4.29916084e-01
-7.19425321e-01 1.14739180e+00 5.99109679e-02 1.92568958e-01
-1.12545180e+00 2.50591226e-02 6.60514653e-01 5.47351778e-01
7.64601171e-01 -5.37941754e-01 9.04792920e-03 -1.17718017e+00
3.43134135e-01 -3.70905578e-01 3.88085842e-01 -8.85949314e-01
-9.95875001e-01 4.42649156e-01 2.26821363e-01 -1.40988350e+00
-3.17573845e-02 -2.43871734e-01 -2.63700753e-01 5.44409454e-01
-1.65133691e+00 -1.30007660e+00 -5.31356454e-01 5.63528359e-01
8.32225204e-01 2.91220129e-01 2.98760444e-01 6.24941647e-01
-2.37744376e-01 1.43223286e-01 -5.72111346e-02 -2.25925997e-01
6.36518657e-01 -1.01234508e+00 8.03349674e-01 9.52027738e-01
5.74117601e-02 1.65644690e-01 4.89467055e-01 -6.24684632e-01
-1.81719732e+00 -1.19459224e+00 8.20476413e-01 -6.70852184e-01
7.76898563e-01 -4.95285094e-01 -8.20896327e-01 6.34091854e-01
9.19674933e-02 5.42214215e-01 2.36899078e-01 -3.46415699e-01
-2.24037141e-01 -2.57746905e-01 -8.34729910e-01 4.21524823e-01
1.51080120e+00 -3.65410775e-01 -7.46428892e-02 4.00620043e-01
1.07867229e+00 -8.17298532e-01 -9.73367393e-01 4.53036755e-01
8.39942992e-01 -1.09310186e+00 1.26815057e+00 -2.29245871e-01
3.01712453e-01 -6.22498572e-01 -4.76228416e-01 -4.97307509e-01
-2.71668702e-01 -6.41802788e-01 -7.07119763e-01 1.04679394e+00
6.17578626e-02 -3.97443324e-01 1.02913833e+00 2.88475573e-01
-2.49465033e-01 -1.01968110e+00 -9.30407882e-01 -1.03617728e+00
-5.19765615e-01 -7.78339028e-01 3.65919054e-01 3.47783327e-01
-6.81647897e-01 2.35654920e-01 -3.91381085e-01 3.83821130e-01
8.35117459e-01 3.81320924e-01 1.08330917e+00 -8.67661059e-01
-2.65527546e-01 -2.15228125e-01 -3.57889086e-01 -1.58004785e+00
-3.32234353e-01 -6.23203397e-01 -1.60532966e-01 -1.52275622e+00
1.88550398e-01 -3.29603434e-01 -8.15153345e-02 2.37871796e-01
1.16194775e-02 2.25785881e-01 5.60114205e-01 3.24785143e-01
-9.58502829e-01 7.10857213e-01 1.08357954e+00 1.39847830e-01
-6.59439191e-02 4.19430770e-02 -1.15493961e-01 7.17878103e-01
6.65179074e-01 -5.97992241e-01 -3.36812794e-01 -7.75197327e-01
1.58371143e-02 2.16860488e-01 6.29487991e-01 -9.83515918e-01
4.91150439e-01 -1.79910272e-01 2.53008958e-03 -1.50147212e+00
8.99672329e-01 -6.75813138e-01 3.31079870e-01 5.11902869e-01
1.35874897e-01 4.65896040e-01 3.07494074e-01 8.96514356e-01
-1.60362348e-01 2.73203552e-01 7.01609850e-01 -1.45397484e-01
-1.00423908e+00 8.86947095e-01 -6.76504746e-02 1.58019587e-01
1.16588497e+00 -1.14224628e-01 -3.20167989e-01 -2.60412067e-01
-5.11783659e-01 5.12908518e-01 5.94506860e-01 4.84298229e-01
7.52163887e-01 -1.36856389e+00 -8.31205726e-01 -1.73714474e-01
3.24864894e-01 3.06033909e-01 5.62632740e-01 9.66901243e-01
-6.17499769e-01 3.74203563e-01 3.90934080e-01 -1.16503203e+00
-1.38167953e+00 4.31714475e-01 9.80021209e-02 -1.08872809e-01
-8.13770235e-01 8.12276125e-01 3.59784037e-01 3.48074622e-02
1.81905866e-01 -2.45284870e-01 3.91026109e-01 1.05839875e-02
6.59313917e-01 7.69780993e-01 6.92335293e-02 -6.21047139e-01
-5.99223971e-01 6.72288179e-01 -6.38556704e-02 -3.10390830e-01
1.47350991e+00 -1.78634405e-01 1.40460253e-01 5.17659009e-01
1.09739566e+00 1.15741245e-01 -1.64453113e+00 -4.20089662e-01
-1.78881690e-01 -8.49560082e-01 7.69885480e-02 -5.77226818e-01
-9.48488355e-01 7.34876871e-01 6.20900154e-01 -4.57671843e-02
9.74018693e-01 3.27746302e-01 1.05092144e+00 9.33922604e-02
7.31129110e-01 -7.34475017e-01 7.70747960e-02 5.31415999e-01
9.31995451e-01 -1.38020897e+00 2.24643499e-01 -5.75101197e-01
-4.14119482e-01 7.75625587e-01 7.50351012e-01 -7.19159469e-02
6.01705790e-01 4.39786315e-01 -2.62987345e-01 -3.91851574e-01
-1.12622666e+00 -2.98966944e-01 1.78816751e-01 2.56804168e-01
1.13176323e-01 -1.62294522e-01 -5.70457578e-02 2.36220192e-02
2.43005469e-01 1.02059677e-01 2.92125016e-01 1.28755426e+00
-4.05834258e-01 -9.95491087e-01 -2.06706241e-01 2.36703083e-01
-4.14928645e-01 -5.80070429e-02 4.91955653e-02 8.98856580e-01
-1.24969333e-01 6.94324195e-01 6.45771101e-02 -2.97829151e-01
7.08845854e-01 -2.68607110e-01 3.73848975e-01 -6.15753829e-01
-4.12882149e-01 3.69514406e-01 1.05095230e-01 -1.25001752e+00
-7.06619918e-01 -9.75338340e-01 -1.22286785e+00 -5.68171561e-01
-1.74220905e-01 -3.96464944e-01 6.15185916e-01 4.00272071e-01
8.01477492e-01 3.11844796e-01 5.67299068e-01 -1.13713443e+00
-2.47388557e-01 -4.63043034e-01 -1.24983676e-01 -2.82972469e-03
4.41968501e-01 -6.78997159e-01 -1.38262913e-01 1.41204432e-01] | [8.090742111206055, -1.2632662057876587] |
9bc873ea-bae2-413c-8e03-2246bed392c8 | graph-contrastive-pre-training-for-effective | 2108.10821 | null | https://arxiv.org/abs/2108.10821v1 | https://arxiv.org/pdf/2108.10821v1.pdf | Graph Contrastive Pre-training for Effective Theorem Reasoning | Interactive theorem proving is a challenging and tedious process, which requires non-trivial expertise and detailed low-level instructions (or tactics) from human experts. Tactic prediction is a natural way to automate this process. Existing methods show promising results on tactic prediction by learning a deep neural network (DNN) based model from proofs written by human experts. In this paper, we propose NeuroTactic, a novel extension with a special focus on improving the representation learning for theorem proving. NeuroTactic leverages graph neural networks (GNNs) to represent the theorems and premises, and applies graph contrastive learning for pre-training. We demonstrate that the representation learning of theorems is essential to predict tactics. Compared with other methods, NeuroTactic achieves state-of-the-art performance on the CoqGym dataset. | ['Xujie Si', 'Binghong Chen', 'Zhaoyu Li'] | 2021-08-24 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [ 2.97242254e-01 5.57625115e-01 -3.76328081e-01 -1.17638864e-01
-7.82985538e-02 -7.67158985e-01 7.48555541e-01 3.68470438e-02
1.79189697e-01 4.20538664e-01 1.07544139e-02 -1.22112274e+00
-1.70463085e-01 -1.11293340e+00 -1.12653816e+00 2.83661991e-01
-1.74721971e-01 6.20526791e-01 -5.68600604e-03 -4.83955115e-01
4.66661572e-01 1.89863279e-01 -1.21136057e+00 5.23154736e-01
1.28482246e+00 6.22811258e-01 -2.81839818e-01 1.08744514e+00
-2.84618080e-01 2.08552670e+00 -4.83085632e-01 -9.80340600e-01
1.91358000e-01 -6.00191474e-01 -1.38496399e+00 -6.42790258e-01
7.01654315e-01 -7.68980980e-01 -7.06489563e-01 1.21566999e+00
-3.61391425e-01 -3.04568466e-02 6.96710527e-01 -1.83100069e+00
-7.73121297e-01 1.36492991e+00 -2.09126189e-01 1.18410476e-01
4.93340373e-01 4.13502067e-01 1.38525689e+00 -2.30386063e-01
7.81060278e-01 1.21430993e+00 7.16133952e-01 8.14821839e-01
-1.19869208e+00 -8.55638802e-01 1.84408024e-01 9.28773344e-01
-9.07971144e-01 -5.74236773e-02 1.21125233e+00 -5.87199569e-01
1.08429420e+00 1.35247678e-01 7.77971029e-01 1.45852840e+00
3.54581028e-01 1.08933163e+00 7.07060635e-01 -3.43461454e-01
-1.55999422e-01 -3.50033760e-01 4.71953243e-01 1.38368595e+00
3.60912085e-01 2.90952325e-01 -2.93371230e-01 -1.30081907e-01
7.88466871e-01 -1.17876053e-01 -2.19707161e-01 -5.67990243e-01
-1.12456775e+00 1.06125462e+00 5.75937629e-01 2.08332434e-01
-4.96007279e-02 7.00273335e-01 7.64326990e-01 6.31187320e-01
1.28253400e-02 1.20828795e+00 -5.40486634e-01 -2.75076360e-01
-6.76402390e-01 3.59363109e-01 1.11955571e+00 8.80533516e-01
5.05500674e-01 2.22392529e-01 -2.39420488e-01 1.77295960e-03
-8.70599300e-02 2.27257058e-01 -4.66483347e-02 -1.26551080e+00
5.58710754e-01 9.48680580e-01 -1.95694059e-01 -1.15392351e+00
-3.17549974e-01 -5.06408334e-01 -7.49037862e-01 1.45154878e-01
5.34102678e-01 -1.69734195e-01 -7.28467643e-01 1.64341903e+00
-1.36988610e-01 6.32236898e-01 4.66694273e-02 6.67575717e-01
1.07119000e+00 6.92991972e-01 -1.25906944e-01 1.28397584e-01
7.22485423e-01 -1.19601607e+00 -5.22223115e-01 -7.31656775e-02
1.00319135e+00 -5.16040511e-02 1.04755926e+00 8.47313225e-01
-9.62407649e-01 -2.57158220e-01 -1.09516048e+00 -3.28856796e-01
-5.46926081e-01 -2.91428328e-01 1.37931359e+00 2.22912803e-01
-9.61777449e-01 9.26274955e-01 -3.25275362e-01 2.84766585e-01
7.13573396e-01 3.40000778e-01 -1.32033825e-01 -3.04212332e-01
-1.59234917e+00 9.95436490e-01 6.20768309e-01 5.39291576e-02
-1.38536966e+00 -1.02044320e+00 -1.24783170e+00 4.06510055e-01
1.05861437e+00 -8.40529084e-01 1.59718597e+00 -8.34645271e-01
-1.60032678e+00 5.94181657e-01 4.07405078e-01 -8.90483439e-01
5.75506032e-01 -1.17950350e-01 -9.22359601e-02 2.56744832e-01
-7.13629574e-02 5.09023845e-01 8.08290899e-01 -1.11599469e+00
-1.59467220e-01 1.82895154e-01 9.40439522e-01 -2.67845809e-01
-4.44165338e-03 -1.17470935e-01 1.52666524e-01 -2.71519810e-01
-4.13448930e-01 -5.75785518e-01 -8.70695487e-02 -2.22096086e-01
-6.78896785e-01 -7.48451114e-01 7.28465319e-01 -6.70379937e-01
1.20652914e+00 -1.58369315e+00 3.77388507e-01 2.02857792e-01
1.08904278e+00 4.28431451e-01 -3.34230274e-01 3.13393533e-01
-2.54978776e-01 3.55064750e-01 3.83594334e-01 2.84408659e-01
5.21708727e-01 -2.72068661e-02 -4.70104456e-01 6.34254590e-02
2.56676912e-01 1.45219624e+00 -1.31663573e+00 -5.92170000e-01
2.61793107e-01 -1.80531025e-01 -8.59759331e-01 3.81286234e-01
-1.02384424e+00 -7.55361617e-02 -3.62446368e-01 4.42714095e-01
4.98326212e-01 -6.28974438e-01 5.52235723e-01 6.93707094e-02
5.20853460e-01 7.52629459e-01 -5.38628817e-01 1.66687047e+00
-3.91050100e-01 8.82498682e-01 -9.22601670e-02 -1.16203570e+00
5.09934187e-01 7.91979730e-02 3.07277747e-04 -5.34522831e-01
5.56452870e-01 -1.11147799e-01 5.59561789e-01 -4.46882129e-01
2.96783209e-01 6.09417970e-04 1.34321051e-02 7.20119536e-01
2.60231018e-01 -3.51389498e-01 3.25013965e-01 8.18238616e-01
1.45861471e+00 2.70424038e-01 4.02376652e-01 3.68676409e-02
4.25503284e-01 3.63696605e-01 3.79698366e-01 1.01887274e+00
-1.04426824e-01 -3.58784050e-01 1.31452596e+00 -6.70028567e-01
-7.27582753e-01 -6.85000718e-01 8.76119971e-01 1.11615336e+00
-1.06237225e-01 -1.10331655e+00 -7.39906669e-01 -1.52330279e+00
1.43454954e-01 1.12703931e+00 -8.40854049e-01 -6.00092351e-01
-6.17500007e-01 5.37133157e-01 8.67637813e-01 5.39539635e-01
5.68451047e-01 -1.19458508e+00 -2.82763422e-01 9.92513914e-03
-2.46377558e-01 -1.06989264e+00 -6.00440381e-03 1.66867584e-01
-5.77032626e-01 -1.81775522e+00 1.04940481e-01 -6.12349272e-01
7.36889482e-01 3.81991565e-01 1.40104055e+00 7.71045268e-01
6.06112555e-02 3.20632339e-01 -3.62146229e-01 -6.39781237e-01
-7.70685196e-01 -2.58787740e-02 -1.52360231e-01 -6.78364098e-01
2.44923323e-01 -6.78969085e-01 -8.69784281e-02 -1.35254338e-01
-6.02858424e-01 6.48380876e-01 4.79254901e-01 8.81743073e-01
-2.72595227e-01 1.98105112e-01 9.64695588e-02 -1.31914425e+00
9.17686284e-01 -2.06615537e-01 -9.86411214e-01 4.24011976e-01
-7.69119859e-01 2.68753231e-01 1.51978517e+00 -4.89630878e-01
-6.16158426e-01 -4.88857448e-01 4.53461528e-01 -8.77255738e-01
2.03997251e-02 8.77269506e-01 5.42387329e-02 -1.10439070e-01
9.75005269e-01 6.36672005e-02 2.13356111e-02 1.57088280e-01
6.50039375e-01 -6.44795969e-02 5.66111207e-01 -1.07976043e+00
1.40898037e+00 -3.54811281e-01 3.04631412e-01 7.74179250e-02
-1.15522718e+00 2.27715895e-01 -4.51415271e-01 -3.76041353e-01
3.54106307e-01 -3.05455029e-01 -1.50550830e+00 1.12782031e-01
-1.50749576e+00 -1.06464279e+00 1.40372127e-01 1.80674732e-01
-6.47095203e-01 4.14863259e-01 -7.87917554e-01 -5.83496392e-01
-4.96074408e-01 -8.55022669e-01 3.90822530e-01 -2.47392610e-01
-3.33835453e-01 -1.05562067e+00 -1.88953832e-01 6.28891289e-01
1.79654300e-01 4.21358049e-01 1.43235111e+00 -9.41933334e-01
-8.24155867e-01 -5.92253804e-02 -4.68359560e-01 3.92099321e-01
-9.49698091e-02 7.82040786e-03 -6.39133036e-01 1.41158119e-01
-2.99977303e-01 -7.18494117e-01 6.98788345e-01 1.86977208e-01
1.54830623e+00 -7.86563456e-01 -1.56548694e-01 4.72499520e-01
8.29587340e-01 -1.29527882e-01 5.89868665e-01 1.27784848e-01
1.04803002e+00 3.64121437e-01 5.31169772e-01 -3.11324336e-02
4.73499149e-01 -2.00153459e-02 7.24926353e-01 1.51193023e-01
7.31270835e-02 -6.84588850e-01 3.48698497e-01 1.97394609e-01
-2.05029756e-01 -7.29752555e-02 -1.42012370e+00 1.21011965e-01
-1.95667005e+00 -1.40079999e+00 -1.16649561e-01 1.36337578e+00
9.04325306e-01 5.67860246e-01 4.67039347e-02 3.91922921e-01
5.20684183e-01 2.07053766e-01 -5.30079842e-01 -6.87979341e-01
3.60578597e-01 5.54831266e-01 1.30640998e-01 5.88881791e-01
-9.97295141e-01 1.49400818e+00 6.54970455e+00 7.53217518e-01
-1.02794170e+00 -3.76320839e-01 4.49257120e-02 2.24957138e-01
-3.18210870e-01 1.63831174e-01 8.92341733e-02 -6.57588467e-02
9.86249089e-01 -5.45180380e-01 9.99694705e-01 1.10118079e+00
-1.05186850e-01 2.85161495e-01 -1.63787234e+00 6.62284076e-01
1.98966399e-01 -1.96901608e+00 2.95508146e-01 -6.91736639e-02
4.31161255e-01 -3.48335177e-01 -3.06458101e-02 1.09586096e+00
9.70605612e-01 -1.40896571e+00 5.26088476e-01 3.37728590e-01
6.95553720e-01 -8.79308760e-01 7.53607571e-01 5.26154041e-01
-7.02521443e-01 -2.61858970e-01 -1.14011809e-01 -7.80386686e-01
-6.24730051e-01 1.02238990e-01 -1.50789249e+00 7.07873821e-01
8.48910585e-02 1.12368500e+00 -6.91161275e-01 5.01652420e-01
-1.14101398e+00 8.27861190e-01 1.75833926e-01 -5.53530514e-01
3.22951913e-01 2.17594374e-02 3.29705536e-01 9.96446550e-01
-3.38961005e-01 2.79013187e-01 2.46176839e-01 1.44705653e+00
-6.12707138e-01 -4.66732353e-01 -1.06617880e+00 -6.38268828e-01
6.26110584e-02 1.33951509e+00 -3.54954243e-01 -5.54040670e-01
-2.17537105e-01 6.11244023e-01 8.12921703e-01 2.22397730e-01
-1.21460509e+00 -4.77696747e-01 3.34818691e-01 3.12697589e-02
1.76408634e-01 -3.50423217e-01 -3.83117259e-01 -1.34753692e+00
-1.74725175e-01 -1.51499808e+00 4.80444074e-01 -1.10057926e+00
-9.98172522e-01 1.22602634e-01 1.16883591e-01 -1.03033388e+00
-5.91453373e-01 -1.06307483e+00 -9.40386295e-01 5.18577278e-01
-1.36575556e+00 -1.57959867e+00 -2.86703616e-01 6.49445713e-01
1.04376331e-01 -1.96497291e-01 9.35435832e-01 -2.66536862e-01
-4.81382728e-01 6.94913089e-01 -8.34754586e-01 7.97757804e-01
1.32531807e-01 -1.51483297e+00 5.29839993e-01 9.61725235e-01
1.31007269e-01 9.32457983e-01 8.89798284e-01 -6.11935556e-01
-1.99559629e+00 -1.08624351e+00 6.05521500e-01 -6.09390020e-01
1.44148588e+00 -4.45636749e-01 -7.23084807e-01 1.11494505e+00
3.78148228e-01 -2.00424075e-01 4.44782913e-01 7.40775228e-01
-1.00062287e+00 2.30495781e-01 -7.40298808e-01 1.01182616e+00
1.44645524e+00 -9.62219417e-01 -1.10977197e+00 5.89267910e-01
8.17939460e-01 -6.46113217e-01 -5.60108662e-01 2.23268762e-01
3.24828893e-01 -7.98996687e-01 6.39233768e-01 -1.38082659e+00
1.17134571e+00 -4.52438183e-02 3.92086715e-01 -1.33319974e+00
-4.47222739e-01 -1.04235029e+00 -8.59590292e-01 7.43145525e-01
3.63865048e-01 -2.76746601e-01 9.90027010e-01 4.94224459e-01
-2.58532196e-01 -8.54530871e-01 -3.07349294e-01 -8.71213496e-01
2.10325032e-01 -7.24382699e-01 4.82049882e-01 1.32150173e+00
7.49886990e-01 9.70018804e-01 -2.28193074e-01 2.09919233e-02
4.22550887e-01 4.57451224e-01 1.33726168e+00 -1.65090573e+00
-2.19509050e-01 -8.02998900e-01 -2.64385402e-01 -9.11959887e-01
1.31990659e+00 -1.50535166e+00 -2.63917774e-01 -1.86679351e+00
2.75919855e-01 3.45847875e-01 -1.50243759e-01 8.99387836e-01
3.45379375e-02 -5.25642872e-01 1.18217215e-01 -4.02791411e-01
-1.09214509e+00 3.15955490e-01 1.56682193e+00 -7.34280586e-01
1.01730049e-01 -5.41686676e-02 -9.12627876e-01 7.16000915e-01
7.44735897e-01 -1.54148266e-01 -7.07602561e-01 -3.29230309e-01
6.53065205e-01 2.25485653e-01 7.95801938e-01 -8.32561016e-01
4.52367425e-01 -6.62198246e-01 7.29005411e-02 -5.38346291e-01
2.08226452e-03 -6.18532181e-01 -4.34405416e-01 6.22957587e-01
-5.67475557e-01 -4.97413985e-02 2.76172191e-01 3.21180195e-01
-4.59833369e-02 -1.10331491e-01 1.84565768e-01 -2.02329218e-01
-6.44900680e-01 4.44818348e-01 -3.51439655e-01 6.89454824e-02
7.54220545e-01 2.35726982e-02 -8.84529412e-01 -8.73107910e-01
-3.49951476e-01 6.17151260e-01 2.63253540e-01 2.13560104e-01
8.84884179e-01 -1.03534067e+00 -4.55253780e-01 -3.00726086e-01
-2.38973112e-03 3.22898440e-02 -5.38694672e-02 8.15291047e-01
-6.22969031e-01 4.40898299e-01 -3.53723288e-01 -4.88442220e-02
-1.22190559e+00 1.10049748e+00 5.97400546e-01 -7.18317807e-01
-6.07627809e-01 7.71754920e-01 7.39701539e-02 -6.43984258e-01
1.19070418e-01 -6.39639139e-01 -2.16554597e-01 -5.65769851e-01
3.47030729e-01 1.31767273e-01 -1.74625009e-01 3.43063802e-01
-1.70126364e-01 -6.71542510e-02 -5.14523327e-01 3.29153121e-01
1.45653510e+00 9.58969593e-01 -4.01840657e-01 -7.25994110e-02
8.10247421e-01 -2.04340309e-01 -7.78537631e-01 -1.49865746e-01
1.46611586e-01 1.86165422e-02 7.08882883e-02 -1.04277921e+00
-7.15464830e-01 1.00466371e+00 -7.20045388e-01 4.49396998e-01
8.18866611e-01 -1.81117490e-01 4.94420350e-01 1.26835167e+00
4.86486226e-01 -9.70172465e-01 5.21787465e-01 1.30063379e+00
9.50901330e-01 -1.05011165e+00 1.68066740e-01 -6.12022460e-01
-4.51976538e-01 1.44648314e+00 1.12002540e+00 -3.62713814e-01
2.16251552e-01 1.06927276e-01 -2.57655680e-01 -5.14907539e-01
-1.09399211e+00 1.18230030e-01 3.80911916e-01 6.46109939e-01
3.62479150e-01 6.18411936e-02 5.48801236e-02 8.69273424e-01
-5.51432252e-01 1.78859755e-01 7.71847904e-01 9.04179513e-01
-4.04117927e-02 -8.83902311e-01 -1.38202921e-01 6.97738051e-01
-1.11868665e-01 -3.83250743e-01 -1.21499038e+00 1.15905344e+00
-1.53382063e-01 9.72623348e-01 -5.25026977e-01 -9.50903118e-01
3.01638134e-02 3.75575796e-02 9.88874137e-01 -7.16727078e-01
-7.84762442e-01 -8.14706564e-01 5.65635860e-01 -8.16213429e-01
3.36634181e-02 -3.88588607e-02 -1.35989523e+00 -1.04074514e+00
-1.54493853e-01 3.24705362e-01 3.68307868e-04 1.00919294e+00
1.66614100e-01 8.46653700e-01 3.84138227e-01 -5.79666615e-01
-6.05893493e-01 -9.38440621e-01 -2.38169566e-01 3.70434165e-01
1.81778580e-01 -7.29663432e-01 -3.94602954e-01 -3.15466225e-01] | [8.979100227355957, 7.105759620666504] |
d32a3d8e-55bd-493b-a2a8-317cd8457fc8 | partmix-regularization-strategy-to-learn-part | 2304.01537 | null | https://arxiv.org/abs/2304.01537v1 | https://arxiv.org/pdf/2304.01537v1.pdf | PartMix: Regularization Strategy to Learn Part Discovery for Visible-Infrared Person Re-identification | Modern data augmentation using a mixture-based technique can regularize the models from overfitting to the training data in various computer vision applications, but a proper data augmentation technique tailored for the part-based Visible-Infrared person Re-IDentification (VI-ReID) models remains unexplored. In this paper, we present a novel data augmentation technique, dubbed PartMix, that synthesizes the augmented samples by mixing the part descriptors across the modalities to improve the performance of part-based VI-ReID models. Especially, we synthesize the positive and negative samples within the same and across different identities and regularize the backbone model through contrastive learning. In addition, we also present an entropy-based mining strategy to weaken the adverse impact of unreliable positive and negative samples. When incorporated into existing part-based VI-ReID model, PartMix consistently boosts the performance. We conduct experiments to demonstrate the effectiveness of our PartMix over the existing VI-ReID methods and provide ablation studies. | ['Kwanghoon Sohn', 'Seongheon Park', 'Jungin Park', 'Seungryong Kim', 'Minsu Kim'] | 2023-04-04 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Kim_PartMix_Regularization_Strategy_To_Learn_Part_Discovery_for_Visible-Infrared_Person_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Kim_PartMix_Regularization_Strategy_To_Learn_Part_Discovery_for_Visible-Infrared_Person_CVPR_2023_paper.pdf | cvpr-2023-1 | ['person-re-identification'] | ['computer-vision'] | [ 3.58108670e-01 6.14832006e-02 -2.74845690e-01 -5.45329392e-01
-5.29542208e-01 -2.11797252e-01 7.82855332e-01 -2.45785519e-01
-3.66621584e-01 6.36418998e-01 3.23266089e-01 1.13555916e-01
6.83223307e-02 -4.42274421e-01 -6.37717366e-01 -7.26848304e-01
2.68472254e-01 3.15556973e-01 -4.38519865e-01 -3.02349087e-02
-2.45743036e-01 4.00799364e-01 -1.73920178e+00 1.41338304e-01
1.15961218e+00 7.41885424e-01 -4.80464727e-01 2.79002637e-01
-1.50115520e-01 5.04988849e-01 -5.40395796e-01 -7.58941352e-01
8.30978632e-01 -3.86648029e-01 -4.32484537e-01 2.68042590e-02
8.29176366e-01 -3.38729858e-01 -4.48049515e-01 8.90662491e-01
5.87976754e-01 3.23272854e-01 6.40679479e-01 -1.69622993e+00
-7.95943677e-01 4.35115248e-01 -1.10921991e+00 -7.44664073e-02
3.66535097e-01 -2.67591745e-01 3.15245271e-01 -8.91571403e-01
4.42398548e-01 1.29364932e+00 1.15448833e+00 9.28742528e-01
-1.25134695e+00 -1.04637861e+00 3.72235984e-01 2.31455848e-01
-1.58775938e+00 -5.41297555e-01 1.07840240e+00 -2.35111058e-01
4.97186661e-01 4.83728349e-01 5.56821704e-01 1.34439909e+00
-5.60748398e-01 1.03546000e+00 1.32177258e+00 -6.30912900e-01
-1.09193549e-01 4.40585643e-01 2.85941899e-01 4.60543185e-01
5.62449992e-01 3.44777614e-01 -7.15937555e-01 -4.67937648e-01
5.11333942e-01 2.23059520e-01 -5.49141392e-02 -3.53535831e-01
-9.94312882e-01 4.14380461e-01 3.55874538e-01 -1.99348211e-01
-2.21206903e-01 -2.95603365e-01 2.01679245e-01 -4.88472963e-03
7.04557002e-01 2.74728566e-01 -3.77725601e-01 3.57547849e-01
-9.30448413e-01 2.78141052e-01 4.45664465e-01 1.10119069e+00
7.54652798e-01 -1.25931323e-01 -4.60812479e-01 1.09272909e+00
3.43281627e-01 5.85691273e-01 3.79735112e-01 -7.10313976e-01
3.20713282e-01 8.79803479e-01 3.02456599e-03 -4.49414402e-01
-2.68678069e-01 -5.88720322e-01 -1.23385990e+00 1.54485390e-01
4.39434089e-02 -3.38940263e-01 -1.35190606e+00 1.87116873e+00
7.23731637e-01 2.85903960e-01 -1.15221746e-01 7.09806859e-01
1.00840068e+00 1.33595094e-01 4.52555090e-01 -9.59376991e-02
1.27292693e+00 -9.78092849e-01 -7.84146249e-01 -3.81084085e-01
4.90831107e-01 -4.69132096e-01 5.37983119e-01 3.68963517e-02
-6.96914971e-01 -8.50947797e-01 -1.11996734e+00 1.10517800e-01
-2.59806871e-01 3.86037052e-01 7.56895602e-01 9.60513949e-01
-8.33306015e-01 3.26857746e-01 -5.64150751e-01 -2.51638591e-01
6.08899057e-01 5.55598974e-01 -8.97600949e-01 -3.24285626e-01
-1.00228739e+00 7.13889778e-01 3.03018898e-01 8.01038556e-03
-1.68842658e-01 -8.04390550e-01 -1.11750710e+00 -4.03011948e-01
7.31913149e-02 -9.01595592e-01 9.39320922e-01 -9.79068041e-01
-1.15068328e+00 9.94145274e-01 -4.84693110e-01 -2.22067475e-01
6.14554703e-01 -2.19285101e-01 -6.25912130e-01 -2.20118493e-01
1.02358684e-02 9.76257503e-01 1.06436336e+00 -1.71767342e+00
-4.18958634e-01 -5.84142327e-01 -5.21158993e-01 3.83914530e-01
-5.62200189e-01 -8.01137537e-02 -7.07780302e-01 -9.77327943e-01
1.87163457e-01 -1.06793404e+00 -2.07293123e-01 6.02881052e-02
-5.64053535e-01 -1.56712815e-01 7.43806183e-01 -7.61698127e-01
7.71417320e-01 -2.13630724e+00 -2.88215727e-01 3.84186864e-01
1.12417229e-01 5.04407525e-01 -4.38906372e-01 9.29946266e-03
-4.01802182e-01 9.50277671e-02 -3.36125106e-01 -1.01809680e+00
-2.50127852e-01 1.38014272e-01 2.07013428e-01 5.21526873e-01
1.59928471e-01 8.55629265e-01 -4.72729415e-01 -5.52525640e-01
1.91590995e-01 6.11471593e-01 -3.24811369e-01 2.99865156e-01
3.31218362e-01 6.01812243e-01 -1.83415517e-01 8.85829270e-01
1.21130931e+00 7.49910101e-02 -1.05321631e-01 -3.40188146e-01
2.39982530e-01 -3.16069901e-01 -1.15202093e+00 1.57628965e+00
1.08087927e-01 1.16635114e-01 1.04562737e-01 -7.76658058e-01
1.06478858e+00 1.71190545e-01 6.58558249e-01 -8.00110459e-01
1.32461950e-01 -2.21896291e-01 -2.11357117e-01 -3.14008892e-01
7.31786251e-01 -1.40016794e-01 2.96373814e-01 4.22815323e-01
-6.41567707e-02 6.59521401e-01 -2.87364870e-02 1.37509122e-01
5.88719666e-01 4.50641274e-01 7.48938918e-02 3.33742350e-01
5.34895957e-01 -4.61900383e-02 7.72132397e-01 8.20266187e-01
-3.45458061e-01 8.38036001e-01 -3.72069150e-01 -1.53519765e-01
-1.15148163e+00 -8.82591605e-01 -4.19544905e-01 1.01029956e+00
1.88743547e-01 -1.96829796e-01 -6.83652759e-01 -8.81742835e-01
3.16416889e-01 3.78246933e-01 -9.13055360e-01 -3.26195955e-01
-2.45474026e-01 -1.40170181e+00 5.48452020e-01 6.93460226e-01
9.52494264e-01 -5.12354970e-01 5.72252214e-01 -1.48664623e-01
-4.19455200e-01 -1.12184894e+00 -5.04618943e-01 -2.09432378e-01
-7.97690153e-01 -1.04923666e+00 -9.01198804e-01 -6.48942113e-01
1.00122809e+00 6.36338890e-01 7.15028703e-01 8.17978531e-02
-2.01292902e-01 5.03026485e-01 -5.50681114e-01 -9.39027131e-01
-3.30772638e-01 -2.61220708e-02 4.30469453e-01 3.25093150e-01
8.32419395e-01 -4.04644191e-01 -5.29854119e-01 4.27487105e-01
-6.24315560e-01 7.21107945e-02 7.84994841e-01 7.30894983e-01
6.91640854e-01 -3.99000436e-01 6.29917264e-01 -8.74899447e-01
3.99177670e-01 -4.30625647e-01 -1.36110425e-01 3.84295315e-01
-9.69061971e-01 4.18410525e-02 2.98408717e-02 -8.31611693e-01
-1.29506111e+00 4.43552077e-01 -3.38393748e-02 -4.76901203e-01
-1.82521909e-01 1.55601159e-01 -3.95086467e-01 -5.39867103e-01
7.05244482e-01 3.65470201e-01 3.84865761e-01 -7.71678150e-01
3.51670593e-01 9.11224961e-01 9.17953253e-01 -4.85528082e-01
1.18737936e+00 6.37291610e-01 -6.30102977e-02 -7.30814457e-01
-7.98908234e-01 -8.41203272e-01 -7.28880882e-01 -5.67520373e-02
4.90167290e-01 -1.42338264e+00 -4.04272914e-01 7.55744278e-01
-8.21066976e-01 1.27666101e-01 -4.29900497e-01 5.24704993e-01
-8.16652924e-02 7.16325998e-01 -3.49357486e-01 -1.12579143e+00
-7.30333030e-01 -4.25603420e-01 9.70400870e-01 6.41091645e-01
-1.21090770e-01 -4.62871104e-01 3.58131975e-01 8.23213995e-01
1.22332051e-01 2.03871444e-01 2.46345833e-01 -9.10075068e-01
-1.55293211e-01 -8.17105830e-01 -3.26355219e-01 4.36092019e-01
2.85905659e-01 -2.65901566e-01 -1.44570398e+00 -3.57565701e-01
-3.22268367e-01 -9.58066154e-03 9.17584240e-01 1.81247070e-01
1.27778482e+00 -2.54241616e-01 -6.08653069e-01 9.10577238e-01
9.45854545e-01 -1.32447019e-01 6.14111722e-01 5.52075803e-01
9.47493792e-01 6.29527330e-01 4.67076659e-01 4.03596371e-01
4.48580295e-01 6.07473671e-01 1.10303268e-01 -4.28200692e-01
-3.34914774e-01 -4.55643952e-01 1.03025794e-01 3.15153003e-01
-4.15558338e-01 7.13201836e-02 -4.57041472e-01 3.94186765e-01
-1.66501236e+00 -1.12019801e+00 -1.63601786e-01 2.32154012e+00
8.40531409e-01 -3.29802155e-01 3.90154451e-01 2.05044448e-01
9.13849831e-01 -1.75573081e-01 -6.62415802e-01 2.31564686e-01
-4.10474777e-01 1.24623567e-01 5.60388565e-01 -5.04665039e-02
-1.44346058e+00 5.41096747e-01 6.92308378e+00 4.71329659e-01
-6.13665342e-01 1.68675974e-01 5.64778686e-01 -1.15957879e-01
-1.18244857e-01 -3.04566681e-01 -1.04365742e+00 2.92609364e-01
5.69789648e-01 6.89431429e-02 2.69055814e-01 1.00344372e+00
-3.42451558e-02 -1.47513449e-02 -1.00802088e+00 1.30114055e+00
4.75371659e-01 -1.00453210e+00 8.83128569e-02 2.64663160e-01
9.96987104e-01 -1.04827844e-01 7.76358098e-02 4.29830492e-01
7.72560686e-02 -8.86284232e-01 5.73334455e-01 5.51455736e-01
8.50895762e-01 -6.81598961e-01 9.41325068e-01 2.26703644e-01
-1.09756792e+00 -1.60778239e-02 -1.97632417e-01 2.34736919e-01
-3.29974964e-02 3.43071669e-01 -7.09713995e-01 8.95079315e-01
6.50071502e-01 5.20736992e-01 -8.93235862e-01 1.28569281e+00
2.20493693e-02 4.18845177e-01 -3.94292444e-01 4.33370590e-01
-5.42255938e-01 -1.82258725e-01 4.71221924e-01 1.01798630e+00
-1.13809435e-02 1.93632588e-01 1.31663769e-01 7.91326046e-01
-3.38776499e-01 -1.53497517e-01 -3.44012409e-01 2.53593624e-01
6.34966314e-01 1.34974170e+00 -4.07503881e-02 -3.76794815e-01
-5.09207726e-01 1.10368586e+00 1.46713078e-01 5.03148556e-01
-5.75559974e-01 -1.61793530e-01 7.88923919e-01 9.40887257e-02
-1.37943506e-01 -4.08372432e-02 -5.71558058e-01 -1.13247573e+00
2.50003129e-01 -8.37784052e-01 5.25264323e-01 -6.00455284e-01
-1.75528204e+00 4.32283580e-01 2.01433331e-01 -1.27704895e+00
1.74193438e-02 -2.28227302e-01 -5.81823111e-01 1.23642039e+00
-1.59520519e+00 -1.66678333e+00 -7.14633405e-01 7.24257708e-01
2.32474823e-02 -3.29204023e-01 7.63809264e-01 4.58066046e-01
-9.05271292e-01 1.39931488e+00 1.14157274e-01 5.32124758e-01
9.90651250e-01 -9.08599854e-01 5.61870277e-01 1.04764831e+00
1.92569965e-03 7.75349259e-01 4.49729919e-01 -9.72638726e-01
-1.28565478e+00 -1.30357945e+00 4.91227478e-01 -5.07596016e-01
-6.52862042e-02 -2.95428425e-01 -1.03741503e+00 7.03562617e-01
-1.94636717e-01 3.07419389e-01 1.11694026e+00 4.77395505e-01
-6.16725564e-01 -2.45985076e-01 -1.61498177e+00 4.10279781e-01
1.32853448e+00 -4.97802198e-01 -4.73315567e-01 8.29180628e-02
3.31271023e-01 -2.50751942e-01 -7.03158140e-01 6.53717637e-01
6.68513477e-01 -7.32730150e-01 1.26982462e+00 -7.67553627e-01
-2.66799361e-01 -2.87156343e-01 2.56335158e-02 -1.00699306e+00
-5.45099974e-01 -4.02159005e-01 -5.07506073e-01 1.66267169e+00
1.44799680e-01 -6.74185812e-01 1.27110839e+00 1.12563503e+00
1.06825016e-01 -2.38482952e-01 -8.03461671e-01 -8.06140006e-01
-1.98471770e-01 -1.58312663e-01 7.99972236e-01 1.09434545e+00
-2.03085244e-01 -6.09237365e-02 -6.11480474e-01 3.15764368e-01
1.20823479e+00 -2.23730981e-01 9.96522486e-01 -1.48995030e+00
5.25638089e-02 -8.76016542e-02 -3.69485229e-01 -6.60235584e-01
1.77169874e-01 -9.22152936e-01 1.34430351e-02 -1.36608922e+00
7.58118927e-01 -7.35435426e-01 -3.80401373e-01 8.09953570e-01
-6.98939621e-01 5.57727873e-01 1.23476572e-01 3.79527718e-01
-2.66464919e-01 8.49234283e-01 9.87486303e-01 -3.31820250e-01
-4.22052443e-01 2.23342240e-01 -1.12915933e+00 4.17846531e-01
7.77206123e-01 -3.13657552e-01 -3.57011855e-01 -1.92188341e-02
-2.41480485e-01 -6.09114289e-01 3.97619575e-01 -1.05351007e+00
1.52136669e-01 2.38149129e-02 1.10381556e+00 -6.07464194e-01
5.09427547e-01 -7.07665861e-01 2.53325880e-01 6.75546899e-02
-1.89613298e-01 -2.10904136e-01 2.86110342e-01 7.78876781e-01
7.96764940e-02 1.48401961e-01 7.68832028e-01 1.11365236e-01
-6.01857662e-01 6.67405307e-01 9.77239087e-02 -2.76146740e-01
9.66206193e-01 -5.06324053e-01 -3.37733626e-01 -1.10814631e-01
-5.35142899e-01 2.69008279e-01 4.96151984e-01 6.28069162e-01
5.63324928e-01 -1.64425886e+00 -8.33537936e-01 5.73467970e-01
4.69118774e-01 -1.05012760e-01 4.69138235e-01 7.12127924e-01
1.84530109e-01 -2.97622941e-02 -2.09446430e-01 -4.03231323e-01
-1.73741126e+00 7.56802022e-01 3.32065880e-01 1.27977744e-01
-4.58878607e-01 8.92662883e-01 5.83768962e-03 -8.08856428e-01
3.80529642e-01 3.25281411e-01 -3.17754745e-01 -1.01324596e-01
9.49244320e-01 4.91001576e-01 6.73454180e-02 -7.45504081e-01
-4.58833247e-01 4.09130037e-01 -3.67607683e-01 -3.39296609e-02
1.21324837e+00 -1.76127911e-01 -3.38444673e-02 -1.20884426e-01
9.12766755e-01 -1.68539017e-01 -1.29966795e+00 -4.79105055e-01
-3.19377452e-01 -4.04819548e-01 -1.29091233e-01 -9.36394572e-01
-8.87495041e-01 2.13098124e-01 9.79397058e-01 -4.86629128e-01
1.05458653e+00 -1.70903474e-01 6.66497409e-01 2.55608916e-01
2.12579384e-01 -1.03753316e+00 -3.02161306e-01 1.65495165e-02
6.06404185e-01 -1.61225724e+00 3.98679227e-01 -4.96897072e-01
-6.28265142e-01 6.00938797e-01 8.75173569e-01 3.47104728e-01
5.91576278e-01 1.60551697e-01 6.30305111e-02 2.63978302e-01
-1.33586064e-01 -2.00674355e-01 6.86954618e-01 1.15328169e+00
2.71979690e-01 4.73719798e-02 -5.82219660e-02 7.01121509e-01
-3.91405933e-02 1.61952022e-02 6.63616136e-02 9.83833730e-01
-4.88610566e-02 -1.49785972e+00 -8.28353405e-01 6.04325473e-01
-1.79737732e-01 8.01871568e-02 -7.73569345e-01 6.52160645e-01
4.05667365e-01 1.02293253e+00 -3.53489406e-02 -8.11021984e-01
2.35637486e-01 3.72878671e-01 6.92107558e-01 -3.21716934e-01
-5.67327380e-01 -1.95219070e-01 6.67417049e-02 -1.54511690e-01
-8.95001411e-01 -8.57715189e-01 -9.60606456e-01 -5.92981577e-01
-4.52400565e-01 2.57786475e-02 8.18988025e-01 9.96313632e-01
5.76046348e-01 1.68443441e-01 6.89343631e-01 -9.65612710e-01
-5.86943030e-01 -1.29402375e+00 -5.25550246e-01 5.96838236e-01
3.71170074e-01 -7.10960448e-01 -1.85829312e-01 -2.55198963e-02] | [14.74648666381836, 1.0236868858337402] |
bbd1f69c-1246-4990-b14f-b1e5fd135664 | depth-aware-multi-grid-deep-homography | 2107.02524 | null | https://arxiv.org/abs/2107.02524v2 | https://arxiv.org/pdf/2107.02524v2.pdf | Depth-Aware Multi-Grid Deep Homography Estimation with Contextual Correlation | Homography estimation is an important task in computer vision applications, such as image stitching, video stabilization, and camera calibration. Traditional homography estimation methods heavily depend on the quantity and distribution of feature correspondences, leading to poor robustness in low-texture scenes. The learning solutions, on the contrary, try to learn robust deep features but demonstrate unsatisfying performance in the scenes with low overlap rates. In this paper, we address these two problems simultaneously by designing a contextual correlation layer (CCL). The CCL can efficiently capture the long-range correlation within feature maps and can be flexibly used in a learning framework. In addition, considering that a single homography can not represent the complex spatial transformation in depth-varying images with parallax, we propose to predict multi-grid homography from global to local. Moreover, we equip our network with a depth perception capability, by introducing a novel depth-aware shape-preserved loss. Extensive experiments demonstrate the superiority of our method over state-of-the-art solutions in the synthetic benchmark dataset and real-world dataset. The codes and models will be available at https://github.com/nie-lang/Multi-Grid-Deep-Homography. | ['Yao Zhao', 'Shuaicheng Liu', 'Kang Liao', 'Chunyu Lin', 'Lang Nie'] | 2021-07-06 | null | null | null | null | ['image-stitching', 'video-stabilization', 'homography-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 5.30766770e-02 -4.18009013e-01 -1.14660896e-02 -1.73948184e-01
-5.11377156e-01 -3.21454167e-01 4.59373176e-01 -4.03881758e-01
-1.37152761e-01 4.43166316e-01 2.10558530e-02 1.79639995e-01
-1.68515906e-01 -7.73007751e-01 -7.05554962e-01 -9.29923475e-01
4.91158754e-01 2.24332497e-01 2.84836441e-01 -1.92484796e-01
3.49815369e-01 3.63479495e-01 -1.23831642e+00 -6.00439422e-02
8.83853555e-01 1.06079936e+00 3.68375361e-01 4.79276441e-02
3.26621115e-01 4.82836127e-01 3.55441906e-02 -3.87009710e-01
5.14946818e-01 -2.77944535e-01 -1.69652075e-01 3.18764985e-01
7.82816768e-01 -5.25765181e-01 -8.96319628e-01 1.28765571e+00
4.90046263e-01 -8.34074914e-02 3.33657950e-01 -1.14008558e+00
-5.09551704e-01 -1.66418869e-02 -8.92579675e-01 -1.26594096e-01
3.55878830e-01 2.81512529e-01 8.13406527e-01 -9.19459462e-01
6.48107946e-01 1.21961153e+00 5.55738449e-01 1.80544034e-01
-1.10615981e+00 -8.64928663e-01 -3.19370441e-02 3.01262677e-01
-1.42598033e+00 -4.10148650e-01 1.19020426e+00 -4.02892262e-01
4.12334561e-01 -1.10430107e-03 6.14839494e-01 9.19862807e-01
3.07544529e-01 6.32011056e-01 1.11716104e+00 -1.83686391e-01
-2.11271495e-01 -1.04529865e-01 -3.45990360e-01 7.51230717e-01
2.29767665e-01 4.82382447e-01 -6.00424588e-01 1.70329183e-01
1.42362070e+00 3.14874023e-01 -7.70634413e-01 -7.61410356e-01
-1.57856703e+00 6.15920126e-01 6.28805995e-01 2.06588075e-01
-8.34160596e-02 1.46263704e-01 2.11264759e-01 1.85795873e-01
1.46766230e-01 2.13764444e-01 -1.70319617e-01 3.02169863e-02
-6.39549732e-01 1.13729477e-01 3.75604361e-01 1.15926826e+00
1.03914845e+00 -3.11193950e-02 1.25036210e-01 7.35212684e-01
1.74853176e-01 5.79290986e-01 4.62335646e-01 -9.23394382e-01
6.95667386e-01 4.34031099e-01 4.94959205e-02 -1.60508597e+00
-3.55775803e-01 -5.36523342e-01 -1.34331203e+00 1.17283866e-01
2.59842098e-01 9.71542746e-02 -4.16550487e-01 1.49670899e+00
3.86747152e-01 4.45720971e-01 -1.72877103e-01 1.12045610e+00
5.68704426e-01 5.57102501e-01 -7.01188326e-01 -2.71216065e-01
1.09893501e+00 -1.12606978e+00 -7.54812479e-01 -2.68531859e-01
2.85650671e-01 -1.05994749e+00 9.44721341e-01 4.10108417e-01
-8.73812318e-01 -6.70215487e-01 -1.07771659e+00 -3.91372979e-01
7.90524185e-02 2.40827352e-01 6.18677735e-01 2.57569611e-01
-6.80907845e-01 4.61867094e-01 -8.11991334e-01 -4.34372351e-02
1.53818026e-01 2.47031391e-01 -5.16500294e-01 -4.30960447e-01
-9.73957658e-01 5.52550912e-01 4.03766125e-01 2.52137631e-01
-4.71631795e-01 -4.60575759e-01 -8.53390217e-01 9.68781207e-03
4.68967021e-01 -7.93796659e-01 6.98678315e-01 -6.87027991e-01
-1.49148107e+00 6.09025896e-01 1.40617201e-02 5.76053150e-02
8.85466337e-01 -2.02419564e-01 -2.18510300e-01 2.47096822e-01
2.82611921e-02 3.79815072e-01 9.56973493e-01 -1.31115687e+00
-5.46990693e-01 -4.78683025e-01 -8.39203149e-02 3.55199695e-01
-3.08710337e-01 -4.92197096e-01 -7.74925530e-01 -6.40328288e-01
6.49616063e-01 -9.99805808e-01 -1.78145289e-01 2.38544673e-01
-3.79890680e-01 3.26720297e-01 8.33154440e-01 -5.74949622e-01
1.12444079e+00 -2.26827359e+00 3.21654856e-01 -1.49182347e-03
3.44062656e-01 4.68379892e-02 -3.78356352e-02 2.96414465e-01
8.97776708e-02 -4.62928623e-01 -1.09871328e-01 -3.40863377e-01
-1.23933025e-01 6.65681902e-03 -2.66978979e-01 8.25193167e-01
-1.26570627e-01 7.67217934e-01 -8.47141027e-01 -4.49708134e-01
6.52532160e-01 5.96595764e-01 -5.98768711e-01 1.41843855e-01
9.73278657e-02 7.69798696e-01 -5.03172100e-01 5.52017689e-01
1.15462542e+00 -3.55114311e-01 -4.18999344e-02 -6.73092782e-01
-2.24809363e-01 -2.04407200e-01 -1.21797967e+00 1.96785867e+00
-6.20188534e-01 6.37599111e-01 6.17419695e-03 -7.49406278e-01
1.09396708e+00 4.48112711e-02 6.60619795e-01 -8.72149825e-01
2.38076985e-01 5.40733576e-01 -1.06982581e-01 -3.22443187e-01
3.02310169e-01 1.86207831e-01 1.80225000e-01 -1.36602502e-02
-2.24932030e-01 -2.98768699e-01 -1.20530695e-01 -1.17779166e-01
6.87841952e-01 -5.94373718e-02 3.36324126e-01 -3.81906144e-02
8.21758151e-01 -4.79293197e-01 7.74884641e-01 3.28612983e-01
1.40270181e-02 1.05605185e+00 4.17668104e-01 -6.14005268e-01
-1.19607615e+00 -7.18218386e-01 -3.89923662e-01 2.06318364e-01
7.76308596e-01 -3.03459048e-01 -3.71621072e-01 -2.70748764e-01
-2.80129276e-02 1.19167760e-01 -2.76864946e-01 -8.96017551e-02
-6.64730132e-01 -4.93261009e-01 2.12895516e-02 2.49614030e-01
1.06350291e+00 -5.82864106e-01 -5.19569993e-01 1.93395153e-01
-2.86572605e-01 -1.42922163e+00 -7.33287394e-01 -2.15344951e-01
-7.74887919e-01 -1.12375009e+00 -7.59150445e-01 -8.26137841e-01
7.38252103e-01 6.63067698e-01 8.62987757e-01 1.82672128e-01
-7.13029504e-02 9.64529738e-02 -2.77729094e-01 3.24095219e-01
1.26125872e-01 2.65508965e-02 -1.01755830e-02 2.55745620e-01
1.15002140e-01 -8.87924314e-01 -9.52011168e-01 8.14470232e-01
-8.59922647e-01 4.66622144e-01 7.04746306e-01 1.25499022e+00
7.74134874e-01 1.01271302e-01 -1.54832117e-02 -6.30726099e-01
7.63489455e-02 -1.53222948e-01 -1.05150998e+00 2.42512316e-01
-4.55569297e-01 -1.23754852e-01 7.04342544e-01 -4.11929548e-01
-9.02195334e-01 4.16794240e-01 -5.66687472e-02 -8.96001637e-01
1.65425509e-01 2.62088835e-01 -5.49313307e-01 -5.83875716e-01
2.20421240e-01 5.69339454e-01 2.50396878e-02 -4.28410679e-01
8.38633925e-02 2.91962594e-01 6.49464786e-01 -4.28307384e-01
1.13530111e+00 6.76971495e-01 4.25371468e-01 -6.75949097e-01
-6.03381813e-01 -5.37327051e-01 -7.62819290e-01 -9.92769897e-02
5.89907587e-01 -1.17672205e+00 -7.38999248e-01 7.59370029e-01
-1.16874254e+00 -1.15282774e-01 2.45900005e-01 6.46370649e-01
-7.64452279e-01 7.59029508e-01 -3.38817835e-01 -2.43018240e-01
-7.74066895e-02 -1.43060267e+00 1.22699594e+00 2.64882475e-01
4.78853583e-01 -9.98480678e-01 2.26942804e-02 3.72318834e-01
2.27548242e-01 2.90818989e-01 5.37798345e-01 1.60829574e-01
-1.18922520e+00 -1.07573949e-01 -4.92580265e-01 2.84046113e-01
1.39407411e-01 -6.98224902e-02 -7.76872814e-01 -5.05206585e-01
1.52920574e-01 -3.01461369e-01 8.48434448e-01 2.97776908e-01
1.25872457e+00 -2.09250405e-01 -1.38797015e-01 1.43811357e+00
1.65357769e+00 -2.73492504e-02 7.82542408e-01 5.18428683e-01
1.13429677e+00 5.38012028e-01 8.14312160e-01 6.09426320e-01
4.22089458e-01 1.13250959e+00 5.26527762e-01 -1.35566249e-01
-1.77040040e-01 -4.54757810e-01 1.04276687e-01 9.53664601e-01
-6.83169672e-03 -1.17809355e-01 -7.82335222e-01 1.83673248e-01
-2.11174297e+00 -8.11097443e-01 -1.43322363e-01 2.24513054e+00
6.44700110e-01 -5.25441617e-02 -3.94917756e-01 2.59580798e-02
6.03077471e-01 4.31005985e-01 -6.03602946e-01 3.20893884e-01
-4.08983439e-01 -4.87282783e-01 6.73624694e-01 6.74682856e-01
-1.05765676e+00 9.46175992e-01 4.29624414e+00 1.10491395e+00
-1.40927613e+00 -8.27355497e-03 5.47612429e-01 7.59689435e-02
-3.83167267e-01 1.77790076e-01 -8.28016102e-01 7.11973488e-01
-8.05024710e-03 -1.95667058e-01 5.36743760e-01 6.87579274e-01
2.49868277e-02 9.99328587e-03 -1.02456415e+00 1.62412274e+00
2.07695156e-01 -1.36077285e+00 4.49187234e-02 2.51661539e-01
1.02142179e+00 -4.13881503e-02 2.35382363e-01 -1.50325507e-01
-3.08559150e-01 -7.34863877e-01 4.68832821e-01 4.97683614e-01
9.77971613e-01 -5.97373962e-01 7.30665982e-01 2.95741647e-01
-1.36181593e+00 -2.81843934e-02 -7.02538610e-01 1.86299205e-01
2.14838892e-01 6.67689264e-01 -1.54400378e-01 8.13397050e-01
5.21372974e-01 1.22185493e+00 -5.15953183e-01 1.22783887e+00
-1.53485671e-01 -1.56002626e-01 -3.31543028e-01 4.04567301e-01
1.81696087e-01 -5.63516438e-01 4.41003680e-01 6.15386367e-01
6.26523674e-01 2.47112736e-01 2.12052405e-01 6.61678553e-01
7.81466216e-02 8.38115886e-02 -7.05891192e-01 2.84551978e-01
4.53950256e-01 1.14013410e+00 -5.07574797e-01 -2.77663097e-02
-5.76126695e-01 1.07291627e+00 1.53527290e-01 3.66317540e-01
-5.96303761e-01 -2.20747367e-01 5.23179114e-01 2.32716933e-01
1.98674142e-01 -4.26969528e-01 -2.42993563e-01 -1.72673655e+00
2.82808870e-01 -9.30663705e-01 -6.15671836e-02 -6.25854135e-01
-1.19748890e+00 5.85137010e-01 -1.27691731e-01 -1.85856426e+00
-6.73327073e-02 -6.78536475e-01 -5.53350031e-01 6.44364774e-01
-1.78618729e+00 -1.28925300e+00 -9.13160264e-01 9.16380227e-01
5.21169245e-01 -2.69294679e-01 3.14763933e-01 5.30064166e-01
-4.90964681e-01 5.32673120e-01 3.71070176e-01 8.15228447e-02
9.40616131e-01 -8.06513071e-01 2.47426704e-01 7.76306152e-01
-3.48773450e-02 5.24446487e-01 4.43870664e-01 -4.58245665e-01
-1.64519238e+00 -8.09599996e-01 5.21009147e-01 -1.61296785e-01
5.38297713e-01 -2.98642755e-01 -9.69382405e-01 5.06588399e-01
2.24064682e-02 3.27203006e-01 5.92848547e-02 -2.80255914e-01
-3.70205820e-01 -4.01985675e-01 -8.25970173e-01 4.72194582e-01
1.10759890e+00 -5.24805844e-01 8.62196460e-02 3.82504314e-01
5.14778316e-01 -8.17369640e-01 -7.49250889e-01 4.20949847e-01
7.26990283e-01 -1.40867937e+00 1.05300343e+00 2.19790101e-01
6.78527951e-01 -5.16754270e-01 -2.29255319e-01 -1.05661309e+00
-4.54834580e-01 -7.31053770e-01 4.67089638e-02 1.07404423e+00
-2.08231002e-01 -8.17584276e-01 7.72164702e-01 3.08522493e-01
-3.78595404e-02 -7.23149717e-01 -9.89200592e-01 -7.38736331e-01
-1.22217387e-01 9.91409048e-02 6.12836123e-01 1.06282556e+00
-3.37980151e-01 2.04992726e-01 -8.52298677e-01 3.59316289e-01
9.06156123e-01 3.66787940e-01 1.13614357e+00 -1.02471912e+00
-3.74099314e-01 -5.05762458e-01 -7.78225660e-01 -1.49235904e+00
1.03536427e-01 -4.67001379e-01 -1.72215894e-01 -1.04072618e+00
1.50230363e-01 -6.30947590e-01 -5.07969670e-02 1.35540128e-01
-3.65746804e-02 2.52477586e-01 2.59884506e-01 5.28470397e-01
-3.06654960e-01 9.92992043e-01 1.70097864e+00 -6.43215552e-02
-4.72470158e-04 -6.88770935e-02 -2.23341882e-01 6.85861707e-01
6.58008456e-01 -2.12436080e-01 -5.62449217e-01 -8.22616220e-01
1.82716459e-01 3.61883789e-01 4.26219136e-01 -1.19328713e+00
5.57958364e-01 -3.52051824e-01 4.29787457e-01 -6.54017031e-01
4.84195292e-01 -1.10811639e+00 3.95360857e-01 2.81974018e-01
-5.92805333e-02 1.74312532e-01 -1.80370927e-01 5.41302443e-01
-6.27817690e-01 1.86121874e-02 8.99404466e-01 -3.76435928e-02
-6.90369666e-01 9.15619016e-01 5.72954178e-01 -1.65444195e-01
9.40052509e-01 -2.60643929e-01 -4.26669061e-01 -4.00356978e-01
-8.46788958e-02 2.21175894e-01 1.10955274e+00 3.85939360e-01
9.13900077e-01 -1.44249141e+00 -5.57193756e-01 6.20921254e-01
2.06478998e-01 2.71032751e-01 5.95016837e-01 1.01407599e+00
-8.55735421e-01 2.92328477e-01 -5.40143549e-01 -7.68485308e-01
-1.00173664e+00 4.69560862e-01 4.08093244e-01 -9.00410339e-02
-7.99957633e-01 5.23788035e-01 7.79208958e-01 -1.91925034e-01
2.89929837e-01 -8.74962658e-02 5.16350493e-02 -2.83919185e-01
4.31532532e-01 2.05362901e-01 -1.12782620e-01 -8.46743405e-01
-1.99828103e-01 1.29562938e+00 -9.44506899e-02 1.17493719e-01
1.17736077e+00 -3.51785958e-01 -1.11747041e-01 1.50980249e-01
1.52498770e+00 7.57339820e-02 -1.68635845e+00 -4.74491775e-01
-5.31821847e-01 -1.17906237e+00 1.91266894e-01 -1.59544691e-01
-1.49326968e+00 1.09030616e+00 6.30901635e-01 -2.30689779e-01
1.21967804e+00 -3.95666808e-01 9.21816170e-01 3.89079869e-01
4.72156584e-01 -7.50500798e-01 3.51415932e-01 3.95474404e-01
1.06260538e+00 -1.55783403e+00 2.88683504e-01 -6.36606634e-01
-5.88521898e-01 1.39648390e+00 8.61755013e-01 -2.31669724e-01
6.78917646e-01 3.32754739e-02 5.99620193e-02 -8.92285705e-02
-5.96182346e-01 7.51004964e-02 2.93955058e-01 5.81436098e-01
1.88552871e-01 -1.81119397e-01 -1.92520976e-01 2.03953907e-02
-1.37643963e-01 -1.62864476e-01 5.16456246e-01 4.97300535e-01
-1.38662338e-01 -1.19100142e+00 -2.51722872e-01 -1.55911045e-02
5.79182431e-02 2.46656705e-02 -6.55648205e-03 7.93612123e-01
3.79788578e-02 5.82256734e-01 -1.20544463e-01 -5.91592968e-01
3.03060293e-01 -7.97001362e-01 4.53375608e-01 -1.99110955e-01
-8.05198699e-02 4.33644384e-01 -3.78128409e-01 -7.76528001e-01
-3.83164853e-01 -6.29872799e-01 -7.54872978e-01 -5.07351458e-01
-1.46810159e-01 -3.13899070e-01 3.96479189e-01 5.12570739e-01
1.85267478e-01 3.27160172e-02 1.01518524e+00 -9.81535375e-01
-4.71300274e-01 -5.68541348e-01 -6.60352767e-01 4.26378250e-01
5.74421942e-01 -8.06242883e-01 -4.58939642e-01 -1.36376858e-01] | [8.874505996704102, -2.3209385871887207] |
8b32c978-7dd6-4056-aa28-3c4532bb7fcd | evaluating-paraphrastic-robustness-in-textual | 2306.16722 | null | https://arxiv.org/abs/2306.16722v1 | https://arxiv.org/pdf/2306.16722v1.pdf | Evaluating Paraphrastic Robustness in Textual Entailment Models | We present PaRTE, a collection of 1,126 pairs of Recognizing Textual Entailment (RTE) examples to evaluate whether models are robust to paraphrasing. We posit that if RTE models understand language, their predictions should be consistent across inputs that share the same meaning. We use the evaluation set to determine if RTE models' predictions change when examples are paraphrased. In our experiments, contemporary models change their predictions on 8-16\% of paraphrased examples, indicating that there is still room for improvement. | ['Adam Poliak', 'Benjamin Van Durme', 'Shreyashee Sinha', 'Yash Kumar Lal', 'Dhruv Verma'] | 2023-06-29 | null | null | null | null | ['natural-language-inference'] | ['natural-language-processing'] | [ 4.19082701e-01 1.54592857e-01 -3.84727806e-01 -9.88019466e-01
-6.45783305e-01 -1.00199187e+00 6.82731271e-01 2.30369031e-01
-2.87637591e-01 6.30861759e-01 6.82636738e-01 -8.22912872e-01
1.73481211e-01 -5.68448603e-01 -1.10017061e+00 4.58345562e-01
4.10847545e-01 2.39836633e-01 3.01228493e-01 -5.62916338e-01
5.38288295e-01 2.71687508e-01 -1.33080816e+00 1.35482681e+00
7.32480526e-01 4.54412043e-01 -2.80499279e-01 7.36245811e-01
-2.83473134e-01 1.22574115e+00 -7.50879765e-01 -1.01967680e+00
9.86657590e-02 -5.25709271e-01 -1.33845448e+00 -2.49341905e-01
7.17184722e-01 -9.90346000e-02 -4.26618934e-01 1.01431203e+00
-1.98075876e-01 1.64940190e-02 7.55269408e-01 -1.36932611e+00
-9.43824291e-01 1.09441602e+00 -5.21690138e-02 6.15902841e-01
9.76026595e-01 2.20057473e-01 1.47588706e+00 -1.19260919e+00
6.61245525e-01 1.43565190e+00 8.80283475e-01 5.34863532e-01
-1.33000159e+00 -6.38840497e-01 1.10080384e-01 4.58724052e-01
-1.04269946e+00 -6.17362142e-01 4.12238240e-01 -9.56040472e-02
1.82046425e+00 6.36000693e-01 6.30761147e-01 1.17521858e+00
8.36718678e-01 5.65158784e-01 1.11005092e+00 -3.88308525e-01
-9.07395706e-02 2.22421572e-01 5.57647407e-01 3.44586074e-01
5.71286738e-01 6.69692233e-02 -6.70093477e-01 -3.63758624e-01
6.19580224e-02 -1.16431877e-01 -9.14910510e-02 3.72650504e-01
-9.72860754e-01 6.32585764e-01 3.90696675e-01 3.78754795e-01
7.94772804e-02 1.90298110e-01 4.56130207e-01 9.59981084e-01
7.02004284e-02 1.08600748e+00 -5.28508961e-01 -2.07761154e-01
-6.02975130e-01 5.36893427e-01 1.07084060e+00 1.31159413e+00
7.74719000e-01 -2.14839622e-01 6.22185133e-02 8.26298594e-01
2.49998003e-01 4.07953620e-01 8.88964057e-01 -1.04182851e+00
7.03913391e-01 9.32747841e-01 1.91169217e-01 -9.77405906e-01
-5.48126660e-02 -1.31228834e-01 -2.52925366e-01 -5.29484570e-01
1.41449329e-02 5.18115088e-02 -5.09377062e-01 1.64247108e+00
-5.31600475e-01 -2.65792876e-01 5.01256704e-01 2.44308084e-01
7.85834193e-01 7.00772226e-01 1.66838795e-01 -1.24375612e-01
9.61830616e-01 -7.59874284e-01 -4.79498982e-01 -1.11991131e+00
1.02836037e+00 -1.09154463e+00 1.60309386e+00 -5.40201813e-02
-1.08652985e+00 -5.52405953e-01 -1.04389465e+00 -3.30281928e-02
-2.02963501e-01 -4.18797761e-01 2.41860539e-01 3.73480797e-01
-7.85091341e-01 8.58825028e-01 -2.36063793e-01 -4.13542658e-01
1.68811958e-02 8.64464268e-02 -2.64685154e-01 -2.17015818e-01
-1.44487250e+00 1.40493548e+00 6.14071727e-01 -2.22676203e-01
-4.69905972e-01 -6.03396237e-01 -1.04793847e+00 1.00368269e-01
1.95109323e-02 -6.77236974e-01 1.71737731e+00 -1.08008969e+00
-8.80298316e-01 9.30001378e-01 -8.32672119e-01 -6.40588045e-01
2.63855964e-01 -3.18815470e-01 -8.85197639e-01 -2.69913465e-01
1.52187213e-01 4.44664031e-01 5.98829508e-01 -9.60816741e-01
-4.40117717e-01 -5.56819066e-02 2.23012939e-01 9.79335532e-02
-1.00907579e-01 3.03525239e-01 1.54144883e-01 -6.80356145e-01
2.19260558e-01 -9.84382212e-01 1.38577089e-01 -1.94954053e-01
-3.78752321e-01 -2.37122998e-01 4.58781391e-01 -4.17777181e-01
1.51086044e+00 -1.88674366e+00 -1.76086351e-01 1.66646227e-01
-1.04099378e-01 -6.28602579e-02 -2.95775712e-01 7.28267729e-01
-4.73630339e-01 5.55270314e-01 -4.15481001e-01 1.69651844e-02
2.74661630e-02 6.18508101e-01 -1.00465047e+00 -2.31999084e-02
3.60288203e-01 1.45112228e+00 -6.70113504e-01 -3.23240429e-01
-1.22894831e-01 -6.08671486e-01 -6.63507044e-01 1.46889314e-01
-3.51911515e-01 -3.47461104e-01 5.07711098e-02 3.47764522e-01
2.17384443e-01 -4.29344922e-01 1.31155998e-01 -9.05138701e-02
3.19117844e-01 9.26181912e-01 -7.05542803e-01 1.27856219e+00
-5.96422136e-01 9.27652597e-01 -8.29853892e-01 -6.87109828e-01
8.89156103e-01 4.45960350e-02 -5.50974190e-01 -7.01571524e-01
-1.19659707e-01 3.74464482e-01 4.81715709e-01 -6.25046194e-01
8.55394006e-01 -7.21204162e-01 -3.47852379e-01 9.07939851e-01
-4.47911739e-01 -6.84173703e-01 1.59036621e-01 3.54781359e-01
1.15710640e+00 -2.55150974e-01 4.86782432e-01 -1.73955709e-01
4.14784431e-01 5.29075027e-01 4.91314739e-01 1.11388695e+00
-5.60590066e-02 3.78078848e-01 3.93757433e-01 -4.57476079e-01
-1.27263618e+00 -1.05765808e+00 -1.62141085e-01 9.14839149e-01
1.52132571e-01 -7.58398473e-01 -3.22802693e-01 -6.37001395e-01
1.86459303e-01 1.57401490e+00 -5.35479844e-01 -6.47593796e-01
-5.60950875e-01 -3.58665228e-01 9.29184616e-01 8.56137633e-01
1.55794859e-01 -1.01745534e+00 -4.37934309e-01 1.79073468e-01
-3.26296419e-01 -9.33690369e-01 -3.91216308e-01 2.76074201e-01
-1.07470965e+00 -9.97036338e-01 -9.56204161e-02 -6.42722964e-01
5.50976157e-01 2.32170641e-01 1.64642072e+00 2.14930192e-01
2.50263661e-01 2.54434288e-01 -4.65176195e-01 -2.24649638e-01
-1.30131626e+00 -9.05884728e-02 -1.32595167e-01 -8.39176893e-01
7.78863549e-01 -3.29547316e-01 -1.62390526e-02 5.00409544e-01
-8.15763295e-01 8.71119276e-02 4.60163713e-01 8.89547288e-01
2.37571195e-01 -4.45631385e-01 6.34437203e-01 -1.22695780e+00
1.18071806e+00 -5.74104309e-01 -1.14775211e-01 8.01863372e-01
-8.95392120e-01 3.07448864e-01 1.00676894e+00 -4.39742118e-01
-9.46710646e-01 -3.97572428e-01 -1.60583168e-01 -8.80858079e-02
-1.11695282e-01 7.59305477e-01 3.09647143e-01 3.81193578e-01
1.18058240e+00 2.81318247e-01 -2.21148983e-01 -1.02787405e-01
4.17317808e-01 7.97212660e-01 4.15878445e-01 -8.14265728e-01
7.10534751e-01 -1.45454645e-01 -5.00965953e-01 -4.26707685e-01
-9.89771724e-01 -1.88356802e-01 -3.32427025e-02 1.83023885e-01
1.22465640e-01 -7.45151281e-01 -1.89246699e-01 1.27910778e-01
-1.53773284e+00 -2.39469394e-01 -3.67681593e-01 2.61336416e-01
-4.60679650e-01 6.29566252e-01 -8.24049234e-01 -2.83860296e-01
-3.95075768e-01 -9.77342308e-01 5.80856979e-01 -4.36883643e-02
-1.50248849e+00 -7.49458313e-01 1.53081194e-01 1.58083081e-01
2.76739389e-01 -6.05081320e-01 1.61615503e+00 -1.23571026e+00
7.81031474e-02 -6.73524067e-02 -9.79441106e-02 3.70970666e-01
3.38617384e-01 1.43052995e-01 -7.97356725e-01 -1.18330158e-01
1.19014218e-01 -6.57354414e-01 4.49750125e-01 -1.28528669e-01
1.08176291e+00 -6.58851087e-01 -8.64778832e-02 1.89349741e-01
1.07846451e+00 2.12655306e-01 6.56174004e-01 8.04029852e-02
1.20568112e-01 6.87411249e-01 5.15339553e-01 1.59628287e-01
1.38461767e-02 2.64388502e-01 -1.25379890e-01 5.01596630e-01
7.80405626e-02 -7.19424069e-01 4.87944096e-01 9.94397640e-01
6.26756608e-01 -2.73227572e-01 -1.12617469e+00 5.99338055e-01
-1.63335991e+00 -1.23395121e+00 -2.03510165e-01 1.87210560e+00
1.14528549e+00 7.91762471e-01 -2.98964828e-01 8.40115175e-02
5.32978892e-01 1.36320710e-01 -8.03229809e-01 -1.12983072e+00
-1.37356102e-01 3.63864362e-01 2.31216419e-02 7.45353281e-01
-3.30919594e-01 1.03392744e+00 8.61676788e+00 4.04658109e-01
-7.04780936e-01 -4.21408325e-01 5.48346519e-01 8.75709057e-02
-1.10765123e+00 2.58700401e-01 -5.73653400e-01 4.14326251e-01
1.13754451e+00 -8.68142664e-01 3.08989793e-01 8.92680585e-01
-1.43481404e-01 -1.93084963e-02 -1.67845237e+00 4.85616922e-01
2.10735887e-01 -1.42852354e+00 5.72946787e-01 -6.53657258e-01
7.13862062e-01 -1.44650480e-02 -6.34573698e-02 5.95321536e-01
3.73800963e-01 -1.28286242e+00 6.32918596e-01 3.94060761e-01
4.35266376e-01 -6.83036864e-01 7.27065802e-01 5.28109789e-01
-8.81691039e-01 -2.79940944e-02 -6.32424533e-01 -5.59337080e-01
-8.54624435e-03 9.44875777e-02 -1.04578006e+00 1.59519568e-01
3.16359699e-01 9.88008559e-01 -1.02463758e+00 5.28675735e-01
-6.58403933e-01 6.84197009e-01 -1.04806937e-01 -4.30551976e-01
-2.50403918e-02 1.37101188e-01 3.40784758e-01 1.37578154e+00
4.12565619e-02 -1.39467753e-02 -4.56514001e-01 1.23902488e+00
-1.97182730e-01 -1.38147891e-01 -9.87253964e-01 -2.65658468e-01
8.70707393e-01 2.80680209e-01 -2.13058129e-01 -7.92915404e-01
-6.22222364e-01 1.12597191e+00 3.74017984e-01 3.59129816e-01
-8.15539062e-01 -5.20206749e-01 6.64278090e-01 -1.37367278e-01
-1.08084790e-01 2.14814737e-01 -5.35203457e-01 -1.28248465e+00
2.59597957e-01 -1.21486008e+00 2.77850240e-01 -1.43609035e+00
-1.76560509e+00 6.18054152e-01 3.24348658e-01 -1.05969000e+00
-7.75614202e-01 -6.72382295e-01 -8.08651447e-01 8.31695199e-01
-1.01627731e+00 -5.72452843e-01 5.66573814e-02 3.23218256e-01
8.78561199e-01 -1.99846644e-02 8.94254327e-01 -2.14944914e-01
-2.24041060e-01 8.03653240e-01 -3.79961967e-01 8.11374262e-02
8.72031689e-01 -1.05997813e+00 1.12035549e+00 9.78843689e-01
4.37529236e-01 1.48003018e+00 1.17586637e+00 -6.40507102e-01
-1.11885369e+00 -9.22246695e-01 1.49861312e+00 -7.45066285e-01
8.94269049e-01 2.97884464e-01 -1.23053467e+00 1.42226839e+00
3.15063059e-01 -3.23918879e-01 8.83907378e-01 1.48865759e-01
-9.14274812e-01 1.28191292e-01 -8.35564733e-01 1.02179551e+00
1.30132830e+00 -1.21855295e+00 -1.93728185e+00 1.28167465e-01
8.91698301e-01 -3.25684428e-01 -7.16205001e-01 4.42269742e-01
6.51374340e-01 -9.71011281e-01 6.95426047e-01 -1.45816493e+00
1.04557836e+00 -5.35455346e-02 -4.67795968e-01 -1.44022751e+00
-1.08228542e-01 -1.79324895e-01 -4.86035319e-03 8.36783946e-01
1.06417525e+00 -7.30683327e-01 4.18254733e-01 1.19300890e+00
-2.93270737e-01 -6.66186213e-01 -6.86543763e-01 -1.01466584e+00
5.21341026e-01 -7.75266588e-01 6.94855273e-01 9.66212571e-01
8.05546343e-01 7.80038774e-01 2.26170979e-02 -2.47517228e-01
-6.59281462e-02 5.16777992e-01 5.73494017e-01 -6.20391130e-01
-6.32718444e-01 -2.91898876e-01 -1.00517958e-01 -1.40044260e+00
6.01822853e-01 -1.28579378e+00 -9.39744338e-02 -1.40646780e+00
4.73000914e-01 1.07418709e-01 5.08987606e-02 6.62700117e-01
-2.56408572e-01 -4.33055311e-02 1.24280319e-01 2.85095096e-01
-3.21148753e-01 2.90677607e-01 8.57416570e-01 -3.20118904e-01
2.03509942e-01 -2.98467819e-02 -8.48803759e-01 7.11586952e-01
7.36381233e-01 -6.89938962e-01 -6.31396413e-01 -6.56604290e-01
6.30306125e-01 -8.98960382e-02 3.12712640e-01 -7.48473167e-01
1.16742477e-01 -4.82740968e-01 3.64982873e-01 -4.30552393e-01
1.34803399e-01 -8.27178717e-01 3.93437058e-01 9.10130382e-01
-1.11697698e+00 9.54609871e-01 6.50000751e-01 3.61376524e-01
-3.51133645e-01 -8.25146437e-01 6.09652221e-01 -1.24504261e-01
-8.60724211e-01 -5.81106365e-01 -8.19025576e-01 6.58481300e-01
6.05760336e-01 -3.55654180e-01 -6.09190524e-01 -4.13973927e-01
-3.58505398e-01 2.42900409e-04 7.27403283e-01 5.86835861e-01
9.70165551e-01 -1.16138506e+00 -5.81851900e-01 1.42143130e-01
6.59994185e-01 -6.09046996e-01 -3.17165256e-01 1.39175996e-01
-2.81661391e-01 3.90464246e-01 -1.67645946e-01 -3.81508201e-01
-1.30298007e+00 3.62807155e-01 5.96332669e-01 -2.47548651e-02
-2.45560095e-01 7.56579041e-01 -1.44823745e-01 -5.54032147e-01
-1.99389249e-01 -9.88273144e-01 2.48761594e-01 -4.55944091e-01
4.61873144e-01 -2.27858815e-02 1.37531966e-01 -2.59361207e-01
-4.79224801e-01 4.20748651e-01 -5.34708798e-01 -2.04720140e-01
8.32781374e-01 7.93219283e-02 -8.19738358e-02 8.09819281e-01
1.23598254e+00 -6.98504150e-02 -3.77324194e-01 -2.80060977e-01
2.59925365e-01 -6.72685027e-01 -7.07772195e-01 -9.92365122e-01
-1.22553483e-01 6.47341549e-01 -5.05986288e-02 -1.09528072e-01
6.03395879e-01 7.91880786e-02 6.92162454e-01 1.07628238e+00
3.61017913e-01 -7.39630520e-01 1.47654474e-01 1.03143322e+00
1.33633578e+00 -9.90712047e-01 1.17577016e-01 -2.26078033e-01
-8.23362172e-01 1.02344251e+00 8.95027459e-01 -1.38954550e-01
2.24351808e-01 1.19591020e-01 -1.24930486e-01 -6.15991540e-02
-1.32016230e+00 5.41618764e-01 2.67161697e-01 1.21112838e-01
7.74447918e-01 -1.50025994e-01 -4.94624496e-01 5.94549775e-01
-7.35163152e-01 -1.39821723e-01 6.61436319e-01 1.01891506e+00
-5.36512792e-01 -1.01512623e+00 -1.29404783e-01 8.32680523e-01
-2.41145864e-02 -5.11787057e-01 -1.08515394e+00 8.25955391e-01
-4.60448951e-01 1.01540029e+00 1.02861233e-01 -7.07613647e-01
6.17470980e-01 5.10449290e-01 6.39551818e-01 -9.92252588e-01
-8.79864216e-01 -6.87003195e-01 5.60004950e-01 -4.73496169e-01
-1.05807237e-01 -3.86279255e-01 -1.45687866e+00 -7.33332217e-01
-1.43369317e-01 2.06632331e-01 -9.63797122e-02 1.24476576e+00
3.44803840e-01 4.29198667e-02 3.69328201e-01 2.09237665e-01
-1.01522815e+00 -9.79223311e-01 -6.32999018e-02 9.11472619e-01
6.42314404e-02 -7.70413950e-02 -5.84432602e-01 5.96874468e-02] | [11.281116485595703, 9.05488395690918] |
c2e78f74-e089-46b1-a8e8-e21f43f48a62 | emergent-agentic-transformer-from-chain-of | 2305.16554 | null | https://arxiv.org/abs/2305.16554v1 | https://arxiv.org/pdf/2305.16554v1.pdf | Emergent Agentic Transformer from Chain of Hindsight Experience | Large transformer models powered by diverse data and model scale have dominated natural language modeling and computer vision and pushed the frontier of multiple AI areas. In reinforcement learning (RL), despite many efforts into transformer-based policies, a key limitation, however, is that current transformer-based policies cannot learn by directly combining information from multiple sub-optimal trials. In this work, we address this issue using recently proposed chain of hindsight to relabel experience, where we train a transformer on a sequence of trajectory experience ascending sorted according to their total rewards. Our method consists of relabelling target return of each trajectory to the maximum total reward among in sequence of trajectories and training an autoregressive model to predict actions conditioning on past states, actions, rewards, target returns, and task completion tokens, the resulting model, Agentic Transformer (AT), can learn to improve upon itself both at training and test time. As we show on D4RL and ExoRL benchmarks, to the best our knowledge, this is the first time that a simple transformer-based model performs competitively with both temporal-difference and imitation-learning-based approaches, even from sub-optimal data. Our Agentic Transformer also shows a promising scaling trend that bigger models consistently improve results. | ['Pieter Abbeel', 'Hao liu'] | 2023-05-26 | null | null | null | null | ['d4rl'] | ['robots'] | [-9.22743157e-02 -6.97174519e-02 -4.78677005e-01 -1.80163234e-01
-9.28552389e-01 -7.51159012e-01 1.03924656e+00 7.50397816e-02
-9.22171116e-01 9.51444328e-01 2.91372061e-01 -3.52188230e-01
-3.01110744e-01 -5.05897582e-01 -9.19638276e-01 -6.16850197e-01
-2.90941715e-01 9.32541430e-01 2.57916570e-01 -2.79483229e-01
1.87549651e-01 3.82888734e-01 -1.30993021e+00 2.13436991e-01
7.17134893e-01 9.19081748e-01 3.48473459e-01 6.36844635e-01
1.36899561e-01 1.25454247e+00 -2.48463288e-01 -5.71248084e-02
3.29116493e-01 -3.74723822e-01 -8.15630913e-01 -3.11965495e-01
4.94999550e-02 -5.91441035e-01 -6.00801229e-01 6.10891044e-01
4.44311947e-01 2.28866488e-01 6.49219871e-01 -1.40767610e+00
-6.82997882e-01 1.11504674e+00 -1.55564487e-01 1.13753788e-01
2.65132517e-01 5.16212940e-01 1.06365395e+00 -5.64358890e-01
7.16868758e-01 1.30685472e+00 3.28200161e-01 8.17863405e-01
-1.39306498e+00 -3.76188427e-01 5.50610781e-01 3.49322915e-01
-6.09620214e-01 -1.73000917e-01 4.58587170e-01 -3.79730940e-01
1.35864794e+00 -1.30612031e-01 7.92278767e-01 1.52120662e+00
4.63587105e-01 1.35046494e+00 1.49108350e+00 -1.26635849e-01
6.23606622e-01 -1.65110826e-01 -2.56915808e-01 6.38773322e-01
-5.75884938e-01 6.60371482e-01 -5.19895196e-01 -2.78476086e-02
6.55130684e-01 2.82815695e-01 5.76903373e-02 -6.91170335e-01
-1.46381843e+00 6.94589138e-01 4.99088138e-01 2.60421246e-01
-6.02293074e-01 6.37305439e-01 3.63277793e-01 7.34611392e-01
1.51418000e-01 6.46482289e-01 -7.24589348e-01 -6.67751729e-01
-7.05511391e-01 5.53097904e-01 6.16193056e-01 5.96739352e-01
5.37742555e-01 2.53376037e-01 -5.14710546e-01 5.35748661e-01
5.98398931e-02 6.86636686e-01 7.71021307e-01 -1.26569498e+00
3.84135723e-01 4.92085785e-01 2.21051723e-01 -8.13838691e-02
-4.80021209e-01 -3.55766505e-01 -3.26098323e-01 6.05117917e-01
5.28119504e-01 -2.13017106e-01 -1.00347042e+00 2.04688740e+00
-3.74525972e-02 3.40543240e-01 2.35527337e-01 8.44461322e-01
3.96170095e-02 5.95534146e-01 2.21064061e-01 -2.22979873e-01
5.62178791e-01 -9.65005040e-01 -2.99114138e-01 -2.42371440e-01
7.68981278e-01 -3.03456992e-01 1.10680985e+00 6.47733808e-01
-1.10473430e+00 -3.26009721e-01 -6.74846411e-01 3.59414935e-01
-1.14340059e-01 -5.10014258e-02 5.91847420e-01 5.64275905e-02
-1.19760549e+00 1.10280931e+00 -1.14085281e+00 -3.09688747e-01
4.00789946e-01 5.15663505e-01 -2.24929780e-01 9.56407376e-03
-9.31194127e-01 1.09741819e+00 1.72567070e-01 -2.86741525e-01
-1.84667790e+00 -6.90815389e-01 -5.39781630e-01 -1.36889994e-01
6.13554239e-01 -3.86089832e-01 1.79983032e+00 -9.29846168e-01
-1.77067113e+00 4.15594786e-01 8.83174390e-02 -1.16954410e+00
5.85204244e-01 -3.00162971e-01 -2.34645065e-02 -3.20312679e-01
1.41983125e-02 9.74261880e-01 8.59818161e-01 -9.68849421e-01
-8.10188830e-01 -4.24388707e-01 2.46623233e-01 3.39558631e-01
-7.58778602e-02 -2.56025732e-01 -5.05942293e-02 -3.20300788e-01
-4.16818261e-01 -1.18613839e+00 -4.20679122e-01 -2.56874800e-01
9.80023965e-02 -6.76867843e-01 5.29789269e-01 -2.28693381e-01
9.25839067e-01 -1.85641825e+00 6.51043653e-01 -1.43279761e-01
3.49482931e-02 1.50143132e-01 -5.27016103e-01 7.33924329e-01
2.98611641e-01 -1.98867574e-01 -6.53496385e-02 -3.26913118e-01
3.09995025e-01 5.03994346e-01 -7.03233540e-01 3.64946306e-01
9.93693843e-02 1.19966054e+00 -1.35199046e+00 -2.08683401e-01
3.16587478e-01 1.11819565e-01 -6.34531677e-01 3.20945650e-01
-7.07071364e-01 5.80097735e-01 -6.61113322e-01 3.85532141e-01
-2.12651473e-02 5.62258735e-02 1.52491108e-01 3.69061977e-01
-1.68079510e-01 4.05252069e-01 -8.11634660e-01 1.85431945e+00
-4.96668816e-01 4.76323068e-01 -3.57227802e-01 -1.15901124e+00
6.97126508e-01 4.25041556e-01 8.00233901e-01 -1.16197383e+00
-4.29844372e-02 2.92709708e-01 1.51666358e-01 -2.51563191e-01
4.54688966e-01 -1.09238021e-01 -2.68331230e-01 6.56824410e-01
3.21258068e-01 -2.17303917e-01 1.93246409e-01 2.93339252e-01
1.34387457e+00 8.42874527e-01 1.03459999e-01 1.27170935e-01
1.30528569e-01 1.84671566e-01 4.60089654e-01 1.06656671e+00
-3.61839384e-01 1.06209412e-01 7.14414716e-01 -5.04782498e-01
-1.29681015e+00 -1.37588847e+00 3.04162681e-01 1.47792411e+00
-2.05391228e-01 -2.50957459e-01 -4.95404363e-01 -9.29612219e-01
1.55449182e-01 9.82558429e-01 -7.36817300e-01 -4.22048867e-01
-8.01102877e-01 -2.87486881e-01 5.28604865e-01 7.14314938e-01
1.81903228e-01 -1.67872012e+00 -8.21931124e-01 4.99765694e-01
1.92502812e-01 -8.20116878e-01 -5.87522388e-01 5.13408661e-01
-8.88139427e-01 -7.81196773e-01 -7.02278078e-01 -6.11921132e-01
1.99369594e-01 -1.56693846e-01 1.13054562e+00 -2.90891409e-01
6.15001619e-02 7.94925749e-01 -2.22025931e-01 -3.27263921e-01
-4.95025486e-01 -6.16532490e-02 3.33634406e-01 -3.01356196e-01
-7.72968680e-02 -6.50838971e-01 -5.39036930e-01 2.14342982e-01
-5.68960011e-01 7.77820917e-03 7.95016885e-01 8.95952284e-01
5.71791232e-01 -4.34021533e-01 7.52111793e-01 -4.97549415e-01
8.71730566e-01 -4.10658866e-01 -6.99984491e-01 2.83813357e-01
-8.89888585e-01 6.67094767e-01 9.46914256e-01 -7.83749521e-01
-9.47497547e-01 6.30717874e-02 1.72295734e-01 -6.77735746e-01
-9.84194688e-03 3.80970776e-01 4.60710794e-01 3.90554428e-01
6.48177922e-01 5.76821327e-01 2.43779598e-03 -2.71493733e-01
6.65563345e-01 2.07792118e-01 5.33695281e-01 -8.58147442e-01
4.58356678e-01 1.75690234e-01 3.93588096e-02 -1.27840877e-01
-7.04229236e-01 -1.63087994e-01 -3.23240697e-01 -3.38295430e-01
7.39846110e-01 -8.13486099e-01 -1.09840584e+00 5.04576862e-01
-7.44378626e-01 -1.21325970e+00 -7.11920977e-01 6.73816681e-01
-1.22192943e+00 -1.19711079e-01 -6.43776655e-01 -9.29735243e-01
-9.54080224e-02 -1.26321375e+00 9.33542728e-01 9.18446556e-02
4.87381481e-02 -8.92417729e-01 4.59103614e-01 -1.45102754e-01
4.92166162e-01 -8.43738299e-03 7.14790404e-01 -6.37923539e-01
-6.99932992e-01 1.27710536e-01 3.56928468e-01 2.44122937e-01
-1.45019114e-01 -4.76359665e-01 -5.79096854e-01 -5.41632831e-01
-2.94251472e-01 -1.06271160e+00 7.60907173e-01 3.91481966e-01
9.44173217e-01 -1.86211079e-01 -3.40455025e-01 1.75030500e-01
1.28065538e+00 3.74436080e-01 3.26417208e-01 4.63731050e-01
3.26191962e-01 1.76572204e-01 8.92649353e-01 5.67510426e-01
3.83854687e-01 7.40595818e-01 6.37988985e-01 4.45817888e-01
6.39821142e-02 -6.93006873e-01 9.83219445e-01 3.88069957e-01
-4.61494327e-02 -3.20686966e-01 -7.43867517e-01 6.03052676e-01
-2.17502189e+00 -1.21811247e+00 5.85166335e-01 2.29014134e+00
8.14821064e-01 3.66512805e-01 4.16685939e-01 -2.51002789e-01
3.43131050e-02 1.07500732e-01 -1.15143180e+00 -5.11243105e-01
1.40563548e-01 2.37034917e-01 5.47912061e-01 5.52116513e-01
-7.77450025e-01 1.22329938e+00 6.51297617e+00 8.05238128e-01
-1.13918459e+00 -4.14931774e-02 3.64398628e-01 -4.93172675e-01
-3.01874459e-01 7.60991797e-02 -6.66290343e-01 1.64659560e-01
1.04279721e+00 -2.38016099e-01 1.17295682e+00 8.60820532e-01
2.58554846e-01 5.67896478e-02 -1.55257046e+00 7.01435089e-01
-2.29814634e-01 -1.15224540e+00 -1.13359146e-01 3.38008557e-03
7.04468608e-01 6.60290837e-01 3.92652720e-01 1.02151442e+00
1.22104180e+00 -1.06908417e+00 9.63722944e-01 5.38791418e-01
3.67424607e-01 -6.08490407e-01 2.14546904e-01 8.08816433e-01
-1.01155305e+00 -5.69835484e-01 -1.25897527e-01 -1.81134462e-01
1.11856505e-01 -2.28968680e-01 -1.06555295e+00 2.49866828e-01
5.24129510e-01 1.01042414e+00 -2.71286845e-01 7.03435361e-01
-2.22245857e-01 7.93853998e-01 -1.49662092e-01 -3.30390394e-01
7.22185314e-01 -2.60383219e-01 6.09662712e-01 7.36978173e-01
2.23377541e-01 -2.19422162e-01 4.48245734e-01 7.11447656e-01
3.42668891e-02 -2.49152303e-01 -8.07557285e-01 -2.35879228e-01
4.08355176e-01 1.10936069e+00 -2.84652114e-01 -2.88102418e-01
-2.49005288e-01 8.52199018e-01 8.89978409e-01 3.35704207e-01
-8.41316819e-01 3.24859738e-01 5.27582169e-01 -1.15392298e-01
3.73625785e-01 -4.20401901e-01 2.14419737e-01 -1.07340872e+00
-1.20768808e-01 -1.07776034e+00 3.84255528e-01 -7.31598675e-01
-1.00187826e+00 5.61261296e-01 -1.44107686e-02 -1.29206645e+00
-8.19085777e-01 -4.40895498e-01 -4.00788069e-01 4.20067519e-01
-1.22299016e+00 -1.01450324e+00 4.12138104e-01 6.91748321e-01
7.60715306e-01 -5.03908277e-01 7.54067421e-01 -1.12151347e-01
-2.16317192e-01 4.93834436e-01 4.75891352e-01 -1.23698421e-01
6.42287076e-01 -1.66315854e+00 4.13723201e-01 4.06064063e-01
4.02239293e-01 2.19962090e-01 8.02015364e-01 -5.53421140e-01
-1.60342467e+00 -8.46220553e-01 4.87804830e-01 -7.47889280e-01
9.49854672e-01 -2.63686657e-01 -4.73150492e-01 1.09498918e+00
3.09824049e-01 -6.27465025e-02 -1.34665594e-01 9.10904780e-02
-2.46578261e-01 -1.67513341e-01 -8.41878831e-01 9.17510271e-01
1.10381722e+00 -3.66175830e-01 -6.20724201e-01 1.69074282e-01
8.07805717e-01 -4.68277603e-01 -8.87103975e-01 1.86965212e-01
5.58961391e-01 -8.52948368e-01 8.25736403e-01 -9.95869398e-01
3.70218396e-01 2.36678775e-02 -8.88833925e-02 -1.75535929e+00
-8.80518779e-02 -7.72919893e-01 -1.75769061e-01 9.08263564e-01
4.33272600e-01 -6.01651251e-01 7.68089712e-01 3.35905343e-01
-2.21454576e-01 -1.08025789e+00 -1.12222266e+00 -1.00760436e+00
4.17211592e-01 -5.47245562e-01 4.81270224e-01 2.78961807e-01
1.54876828e-01 3.38940740e-01 -7.92032540e-01 -3.57707888e-01
5.60462296e-01 2.29238197e-01 7.63137400e-01 -9.35133994e-01
-6.01854563e-01 -5.56873202e-01 7.51404315e-02 -1.32786274e+00
3.46768200e-01 -1.07152820e+00 1.97482988e-01 -1.39334798e+00
1.87417090e-01 -6.67835414e-01 -5.73391080e-01 6.40810430e-01
1.26918867e-01 -4.99254853e-01 4.92147923e-01 7.25660399e-02
-1.06073940e+00 9.11276281e-01 1.46811628e+00 -1.12003386e-01
-4.02755648e-01 5.33262193e-02 -2.94902712e-01 5.20350397e-01
7.85313547e-01 -7.13770986e-01 -6.47240937e-01 -4.71709311e-01
2.67765135e-01 4.81336623e-01 3.80198866e-01 -9.31341469e-01
2.93519199e-01 -5.37981927e-01 1.94211185e-01 -4.58229333e-01
4.28658158e-01 -7.56245494e-01 -1.30859315e-01 8.33824098e-01
-9.31049645e-01 3.57550412e-01 1.12268880e-01 8.11837614e-01
1.58833385e-01 4.33891900e-02 4.44098920e-01 -1.76214680e-01
-8.25429559e-01 5.92013240e-01 -7.42481709e-01 1.94158047e-01
1.07308602e+00 1.93060562e-01 -1.39125347e-01 -4.44144517e-01
-6.92757845e-01 6.07967198e-01 2.97030181e-01 7.28212237e-01
6.60023391e-01 -1.37665260e+00 -7.74356186e-01 -9.29157883e-02
-2.61249803e-02 -3.20857465e-01 1.48400583e-03 8.25980902e-01
3.01384568e-01 5.03566921e-01 -4.02475119e-01 -6.19012773e-01
-9.47404325e-01 8.45419765e-01 4.46029991e-01 -8.75711918e-01
-5.61737537e-01 3.72464329e-01 3.17429416e-02 -7.19945490e-01
4.04930204e-01 -4.54868734e-01 -1.71006143e-01 -2.84804672e-01
2.39262059e-01 3.56487036e-01 -3.23924601e-01 -1.60341129e-01
-1.67313218e-01 2.52121538e-01 -3.06429178e-01 -6.75236762e-01
1.46202946e+00 1.82871699e-01 3.17329705e-01 6.83459938e-01
7.68918216e-01 -4.21787620e-01 -1.81582677e+00 -4.49964613e-01
5.58679216e-02 -1.23629548e-01 -1.31134301e-01 -1.00472903e+00
-8.00393999e-01 6.75687730e-01 7.41898715e-01 -2.03844886e-02
8.32854271e-01 1.67290494e-01 6.53823435e-01 7.27319777e-01
6.96848512e-01 -1.22220135e+00 5.81026375e-01 7.28569806e-01
9.34604108e-01 -1.19315112e+00 -3.46132368e-01 8.43384802e-01
-1.06585085e+00 6.89010084e-01 7.45092928e-01 -5.06158650e-01
3.19885075e-01 2.40383431e-01 -1.18056647e-01 1.61594942e-01
-1.46903121e+00 -2.78018117e-01 4.70725596e-02 6.90000653e-01
2.02367723e-01 1.69925198e-01 -4.68821749e-02 2.45218322e-01
2.14723982e-02 3.22094262e-01 2.35744163e-01 9.69959021e-01
-6.01084352e-01 -1.44416320e+00 -3.22849005e-02 4.98634845e-01
-3.19025099e-01 5.43583892e-02 -1.36806071e-01 6.98288023e-01
-2.50942379e-01 7.33207405e-01 -3.57014872e-02 -2.69813716e-01
3.56820613e-01 -1.08572848e-01 8.14715147e-01 -2.61377513e-01
-8.05537045e-01 -2.11539477e-01 -5.01536019e-02 -8.47622812e-01
-1.01764940e-01 -8.60748887e-01 -1.47902894e+00 -2.03835800e-01
1.34983659e-01 1.34374753e-01 4.77258980e-01 1.14658904e+00
1.99017942e-01 4.82198179e-01 6.63467467e-01 -9.91154015e-01
-1.40370488e+00 -8.55468035e-01 -4.73467022e-01 3.66165727e-01
4.98268157e-01 -7.28936911e-01 -1.14505850e-01 -2.98994482e-01] | [4.131809234619141, 1.7108391523361206] |
972a94d7-344e-472a-8301-2a5eed6712b4 | combining-deep-and-unsupervised-features-for | null | null | https://link.springer.com/chapter/10.1007%2F978-3-030-68790-8_10 | https://link.springer.com/content/pdf/10.1007%2F978-3-030-68790-8_10.pdf | Combining deep and unsupervised features for multilingual speech emotion recognition | In this paper we present a Convolutional Neural Network for multilingual emotion recognition from spoken sentences. The purpose of this work was to build a model capable of recognising emotions combining textual and acoustic information compatible with multiple languages. The model we derive has an end-to-end deep architecture, hence it takes raw text and audio data and uses convolutional layers to extract a hierarchy of classification features. Moreover, we show how the trained model achieves good performances in different languages thanks to the usage of multilingual unsupervised textual features. As an additional remark, it is worth to mention that our solution does not require text and audio to be word- or phoneme-aligned. The proposed model, PATHOSnet, was trained and evaluated on multiple corpora with different spoken languages (IEMOCAP, EmoFilm, SES and AESI). Before training, we tuned the hyper-parameters solely on the IEMOCAP corpus, which offers realistic audio recording and transcription of sentences with emotional content in English. The final model turned out to provide state-of-the-art performances on some of the selected data sets on the four considered emotions. | ['Roberto Tedesco', 'Licia Sbattella', 'Federico Galati', 'Vincenzo Scotti'] | 2021-01-10 | null | null | null | international-workshop-on-pattern-recognition | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [-2.94110805e-01 7.60628879e-02 4.71698612e-01 -7.03628540e-01
-9.54353988e-01 -3.50529760e-01 4.45732951e-01 3.02478135e-01
-9.10203934e-01 5.47992349e-01 1.08306989e-01 1.05169296e-01
3.03646531e-02 -2.65826225e-01 -4.54468697e-01 -3.70279968e-01
-3.46494138e-01 3.09940577e-01 -3.70010734e-01 -4.19381738e-01
-2.14976192e-01 4.56483155e-01 -1.96183300e+00 5.00500023e-01
5.82312346e-01 1.43885469e+00 -1.75240319e-02 9.56402302e-01
1.33305101e-03 8.41003954e-01 -6.40362501e-01 -6.64953411e-01
-9.90294591e-02 -1.87979713e-01 -7.82172501e-01 -5.14063425e-02
3.83010060e-01 1.91607922e-01 4.96576540e-03 6.81705832e-01
9.38038230e-01 2.00226083e-02 5.16514778e-01 -9.48044360e-01
-1.56144574e-01 8.40923488e-01 1.70314521e-01 1.77634023e-02
4.09227759e-01 -3.67476255e-01 1.18352044e+00 -1.14893520e+00
7.33531654e-01 9.24014568e-01 7.73213089e-01 3.68944198e-01
-8.85539591e-01 -3.82939845e-01 -2.13142008e-01 3.14757138e-01
-1.50539160e+00 -8.70506346e-01 6.77512646e-01 -2.76338100e-01
1.42121315e+00 2.83584744e-01 5.48194408e-01 1.50930440e+00
4.89442907e-02 8.43408227e-01 1.26187277e+00 -7.15172410e-01
2.90134370e-01 6.82210326e-01 4.67260666e-02 3.26102883e-01
-7.07433522e-01 -2.87503093e-01 -7.54750669e-01 -5.59743121e-02
-2.02882156e-01 -7.53865659e-01 -3.13823193e-01 9.36999023e-02
-8.60936642e-01 8.32925558e-01 -4.97831590e-02 9.53503907e-01
-5.31014025e-01 -1.04230739e-01 1.01142693e+00 7.42482305e-01
7.36300945e-01 1.27503380e-01 -7.93706477e-01 -5.50932229e-01
-9.14555788e-01 -2.68069416e-01 1.15816629e+00 4.90726441e-01
5.35845816e-01 3.08398902e-01 3.47635120e-01 1.30777037e+00
1.70464829e-01 5.56340039e-01 6.77613854e-01 -4.51696426e-01
2.84272194e-01 5.15699424e-02 -2.22752914e-01 -9.74219918e-01
-7.93982029e-01 -4.75516737e-01 -8.53406131e-01 -1.13867316e-02
3.18048336e-02 -5.99924564e-01 -3.90877366e-01 1.77006900e+00
-7.03297928e-03 -8.75731111e-02 6.40598238e-01 7.97806621e-01
1.16415417e+00 7.65416980e-01 4.12068740e-02 -1.80360541e-01
1.29304218e+00 -6.85514748e-01 -9.33410823e-01 7.44104665e-03
5.68886817e-01 -9.29637134e-01 1.02264869e+00 8.41107368e-01
-1.05833220e+00 -6.20662451e-01 -1.06010866e+00 1.34853655e-02
-7.93233752e-01 6.31157756e-01 1.96551874e-01 7.56255150e-01
-1.25900471e+00 2.76280999e-01 -4.21497107e-01 -4.66646641e-01
-3.60687256e-01 3.51900280e-01 -6.64174080e-01 5.70543706e-01
-1.72313762e+00 8.93219054e-01 4.98936474e-01 3.65049571e-01
-5.81418872e-01 -1.93091586e-01 -9.62029159e-01 1.82765335e-01
-3.50376256e-02 -6.04875684e-02 1.23973632e+00 -1.54583800e+00
-2.07723570e+00 1.06432760e+00 6.00768579e-03 -5.99615693e-01
4.54495221e-01 -2.73493379e-01 -9.87877786e-01 3.44416648e-01
-4.69165444e-01 6.73408151e-01 8.05652738e-01 -1.07841611e+00
-4.41085875e-01 -5.32455891e-02 -3.18887115e-01 8.45050290e-02
-8.08388948e-01 4.76298243e-01 -2.93634325e-01 -4.28733826e-01
-4.02260512e-01 -7.97591507e-01 7.71609545e-02 -8.19934726e-01
-3.71094704e-01 -7.09541813e-02 4.36667591e-01 -8.42152715e-01
9.57971215e-01 -2.28930426e+00 3.99239272e-01 2.92242199e-01
-2.03529894e-01 6.33556321e-02 -2.41022289e-01 6.52637005e-01
-2.32998729e-01 -1.64221197e-01 -7.50637800e-02 -7.31422603e-01
5.04360914e-01 9.50090587e-02 -2.42630348e-01 3.93943429e-01
2.24893987e-01 5.30993283e-01 -3.66717547e-01 -4.69159752e-01
3.49988699e-01 1.07601249e+00 -3.28130662e-01 1.68751344e-01
-5.47742657e-02 3.61035109e-01 8.42850581e-02 1.96328625e-01
5.16179442e-01 4.91789699e-01 2.09750235e-01 -1.37813807e-01
-3.90059173e-01 2.70281911e-01 -1.21506596e+00 1.78965747e+00
-1.11191332e+00 8.12228322e-01 3.19685608e-01 -8.77604663e-01
1.34515727e+00 9.47216749e-01 5.38949668e-01 -7.75970280e-01
6.99078858e-01 6.66035891e-01 -1.67209014e-01 -6.76310301e-01
7.14254677e-01 -8.05478469e-02 -4.93732721e-01 2.75238723e-01
8.08446348e-01 -1.81479752e-01 2.22943977e-01 -2.62860417e-01
4.92977738e-01 -3.30932021e-01 1.27453774e-01 -2.59340078e-01
9.78925049e-01 -4.72829729e-01 1.24217540e-01 2.57136106e-01
5.83844259e-03 4.07320678e-01 4.88056034e-01 -3.12787831e-01
-8.33610594e-01 -6.10116005e-01 -2.69449621e-01 1.36504424e+00
-6.10636771e-01 -4.49074209e-01 -8.85525167e-01 -3.64833057e-01
-6.21079326e-01 6.32120192e-01 -4.83955950e-01 2.15033934e-01
-2.11652845e-01 -5.62862992e-01 1.05194616e+00 4.31472249e-02
2.01625347e-01 -1.44661212e+00 -6.02907836e-01 4.10057098e-01
-4.52891648e-01 -1.37206709e+00 1.15752399e-01 6.95413828e-01
-1.80820674e-01 -6.63077593e-01 -5.53856015e-01 -9.14916039e-01
-2.05469251e-01 -7.16397643e-01 1.31218398e+00 -4.70042616e-01
-1.82641909e-01 5.63675344e-01 -6.81055784e-01 -5.00589132e-01
-5.48634171e-01 5.50803244e-01 6.00228272e-02 4.70756441e-01
4.33393717e-01 -6.24248624e-01 4.75892983e-02 -1.38566094e-02
-9.73416746e-01 -4.77568418e-01 3.31184655e-01 6.57959282e-01
3.85451436e-01 -3.47802415e-02 7.70033658e-01 -4.12261456e-01
8.41499388e-01 -3.17679435e-01 -2.53668815e-01 6.28891140e-02
6.56338409e-02 -1.96893364e-01 8.49529028e-01 -2.42809266e-01
-9.68038261e-01 1.74801141e-01 -1.07213509e+00 -9.64325219e-02
-6.86086714e-01 7.46915400e-01 -3.84140998e-01 1.14302881e-01
5.06901503e-01 1.55209348e-01 -5.21356165e-01 -5.50198436e-01
4.38464761e-01 1.28183627e+00 5.24376333e-01 -4.11272347e-01
-5.84340654e-02 1.19604096e-01 -5.06641209e-01 -1.34358501e+00
-4.71101344e-01 -3.79116625e-01 -7.40258932e-01 -4.42448854e-01
9.91730630e-01 -1.04456210e+00 -9.17408645e-01 7.69886851e-01
-1.27864766e+00 -2.77959377e-01 -1.49435148e-01 6.81864798e-01
-5.99245191e-01 4.95020598e-02 -8.82259130e-01 -1.02806211e+00
-6.45243406e-01 -9.42101657e-01 1.10140252e+00 -1.27684906e-01
-3.21933776e-01 -1.07936323e+00 2.54913628e-01 -1.05604663e-01
5.06876945e-01 7.80092329e-02 5.03510594e-01 -1.02387881e+00
3.87446284e-01 -3.59662950e-01 2.78197795e-01 8.21472645e-01
-3.08500975e-01 2.42273584e-01 -1.57123339e+00 -1.08790226e-01
1.77297935e-01 -1.04446268e+00 8.09516311e-01 9.14585963e-02
9.07386065e-01 -2.72775651e-03 5.57164073e-01 4.11674708e-01
1.25218487e+00 7.38312230e-02 6.35040283e-01 3.11470896e-01
2.28634030e-01 9.31715429e-01 3.87181252e-01 6.56589925e-01
4.06052381e-01 7.54270017e-01 3.40495050e-01 -2.05239862e-01
4.24991041e-01 3.53320032e-01 6.96505666e-01 1.63501275e+00
2.93657362e-01 -4.46256459e-01 -8.88162553e-01 6.36651635e-01
-1.53208375e+00 -7.72605002e-01 -2.07643971e-01 1.82837868e+00
7.20843852e-01 1.21159572e-02 1.20493777e-01 3.88911396e-01
3.36494923e-01 1.73872739e-01 3.94892655e-02 -1.10066938e+00
-5.47967672e-01 5.77725768e-01 -8.14049989e-02 7.37848818e-01
-1.25292933e+00 8.76157343e-01 5.56748390e+00 7.61345208e-01
-1.56491625e+00 3.05057615e-01 5.61010480e-01 -1.11577995e-01
8.41925070e-02 -7.59710133e-01 -5.63844025e-01 5.92153631e-02
1.70409238e+00 2.36446008e-01 2.54548937e-01 7.72992969e-01
1.81243092e-01 8.92995968e-02 -9.34634209e-01 1.03683078e+00
3.27792197e-01 -6.56330228e-01 -4.68370229e-01 -4.48445469e-01
3.63155633e-01 4.60105985e-01 1.44764215e-01 4.89762574e-01
-4.83592786e-02 -1.11986923e+00 1.00175214e+00 4.07412499e-01
7.67504513e-01 -1.24510694e+00 1.11669374e+00 1.00144826e-01
-1.07899821e+00 1.04602113e-01 -1.29477963e-01 6.70366436e-02
1.86310425e-01 6.25294805e-01 -6.50018752e-01 7.59323299e-01
8.38199198e-01 6.29121006e-01 -4.07527596e-01 5.41026890e-01
-7.40472153e-02 6.09744668e-01 -5.29178798e-01 -6.55795261e-02
4.68538076e-01 5.14521897e-02 2.73605317e-01 1.98657954e+00
5.31864226e-01 -5.62586844e-01 -1.00636780e-01 2.84653485e-01
-1.01606846e-02 8.95700276e-01 -4.64071959e-01 -2.08302826e-01
7.59339929e-02 1.53826773e+00 -4.04627085e-01 -2.58712351e-01
-3.28950584e-01 1.01444101e+00 2.67972142e-01 2.36426786e-01
-7.88602352e-01 -8.10638428e-01 5.15728354e-01 -6.44026220e-01
3.40116322e-01 -3.20327468e-02 1.25279278e-01 -1.15964496e+00
-1.06974579e-02 -1.10742295e+00 1.41878814e-01 -7.11741149e-01
-1.23244596e+00 1.56103814e+00 -4.65098858e-01 -9.15616453e-01
-6.60645187e-01 -7.43628860e-01 -2.72711784e-01 8.61278415e-01
-1.36177266e+00 -9.99489427e-01 -5.84215261e-02 6.80457890e-01
4.65684772e-01 -3.44723582e-01 1.24070132e+00 8.06723833e-01
-3.80430847e-01 6.24480188e-01 1.12197883e-01 9.15372372e-02
8.32131624e-01 -1.19777358e+00 -5.09329584e-06 3.41030091e-01
3.86947483e-01 1.70323774e-01 8.77956271e-01 1.06376201e-01
-1.00693452e+00 -9.26140428e-01 1.49612296e+00 -5.51696867e-02
9.01844978e-01 -7.31590390e-01 -7.82755494e-01 4.76176649e-01
9.67322111e-01 -3.11806500e-01 8.70719492e-01 1.63874000e-01
-3.53595883e-01 -1.01773895e-01 -8.59677732e-01 1.98285639e-01
3.06159884e-01 -9.07664359e-01 -3.77155811e-01 1.74507007e-01
5.72877347e-01 -3.08706850e-01 -1.15138388e+00 2.71954030e-01
5.10692179e-01 -1.36588514e+00 6.04192257e-01 -3.09117526e-01
4.68026638e-01 1.54023260e-01 -4.15288061e-01 -1.63805127e+00
3.43120545e-01 -5.17759919e-01 3.86145383e-01 1.55894470e+00
6.82007909e-01 -6.29374266e-01 5.99627532e-02 -4.00841199e-02
-2.88295388e-01 -4.91632581e-01 -1.15185821e+00 -4.75686640e-01
9.31853056e-02 -1.02174413e+00 3.52521032e-01 9.90476012e-01
2.94088662e-01 5.85057914e-01 -7.07192659e-01 1.79925766e-02
-3.99813466e-02 -1.21941984e-01 5.98970234e-01 -1.12453377e+00
-2.80008405e-01 -4.82061595e-01 -3.81945610e-01 -3.79354626e-01
7.54905581e-01 -8.17298949e-01 1.77985236e-01 -9.52740073e-01
-3.29100221e-01 -6.61232844e-02 -4.21368241e-01 4.20481831e-01
3.88879299e-01 5.22217214e-01 2.05042377e-01 -4.50939864e-01
-6.37462437e-01 7.60705173e-01 4.00146961e-01 -8.27475414e-02
-2.05832049e-01 -2.81068742e-01 -9.69788954e-02 6.77759230e-01
9.99198198e-01 -3.67594570e-01 -6.51013032e-02 -2.66025394e-01
5.68341553e-01 3.37703079e-02 2.02565357e-01 -1.10179520e+00
-2.07682308e-02 5.78396022e-01 9.07486379e-02 -4.96003687e-01
7.09532022e-01 -1.09220886e+00 1.14211127e-01 1.51398450e-01
-5.96228004e-01 -1.63638070e-02 5.56252360e-01 -1.01316519e-01
-8.28344285e-01 -3.78245741e-01 6.85656905e-01 2.87405878e-01
-6.19502664e-01 -1.89545438e-01 -8.69484246e-01 -2.79136132e-02
5.44652343e-01 3.65938514e-01 1.12699389e-01 -7.12885499e-01
-1.03600991e+00 -6.58971667e-02 1.46353662e-01 5.65841794e-01
3.43400478e-01 -1.22548592e+00 -8.96700621e-01 1.42374411e-01
2.45421439e-01 -8.01050663e-01 5.00610650e-01 1.03847766e+00
-4.14312184e-01 6.29355073e-01 -2.50294089e-01 -5.23331761e-01
-1.50824094e+00 2.40617201e-01 5.55996537e-01 -1.81980744e-01
-1.22786559e-01 6.38611794e-01 -3.89610887e-01 -1.01250982e+00
5.33222079e-01 -3.18515599e-01 -4.72597301e-01 6.43658638e-01
3.29171419e-01 9.58551317e-02 5.68834066e-01 -1.13903773e+00
-4.15898681e-01 4.99290735e-01 3.70133400e-01 -7.20248103e-01
1.59705484e+00 -2.53197432e-01 -2.38276422e-01 1.06078398e+00
1.52082193e+00 3.35322917e-01 -6.35367751e-01 -3.03255767e-02
1.02223381e-01 2.97679305e-01 1.72261149e-01 -8.77770066e-01
-1.10070586e+00 1.21785998e+00 7.49183059e-01 3.64681482e-01
1.30141056e+00 -2.78395534e-01 7.24454582e-01 6.95167184e-01
2.23710254e-01 -1.43612337e+00 -1.88533410e-01 9.54880416e-01
1.00302935e+00 -1.17352915e+00 -6.67342067e-01 7.31891841e-02
-9.48341548e-01 1.46116841e+00 2.02233240e-01 1.84095055e-01
7.07134008e-01 5.08369446e-01 6.09936714e-01 -2.18766525e-01
-8.04403067e-01 -3.89578611e-01 3.23089629e-01 2.95581073e-01
8.15407991e-01 7.54829273e-02 -4.23721731e-01 7.04901934e-01
-5.36301494e-01 -8.28959569e-02 4.23134357e-01 4.77503061e-01
-1.87502906e-01 -1.11958933e+00 -3.52047145e-01 -1.52948335e-01
-9.24919009e-01 -1.75324589e-01 -7.26209342e-01 8.53438675e-01
1.52443156e-01 1.19746685e+00 -6.21613935e-02 -5.13643980e-01
4.29222077e-01 6.45377040e-01 1.55637816e-01 -1.55011848e-01
-1.16723263e+00 2.60936528e-01 5.19124568e-01 -4.03736979e-01
-6.83057845e-01 -5.51142454e-01 -1.10299850e+00 3.49521562e-02
-2.13370752e-02 4.26673353e-01 1.19311178e+00 8.90741885e-01
2.63014913e-01 7.08201051e-01 7.47522831e-01 -1.00367796e+00
-8.15792605e-02 -1.29805541e+00 -7.55211234e-01 4.07799006e-01
3.39140654e-01 -1.62517890e-01 -3.14893752e-01 6.14634752e-02] | [13.568646430969238, 5.752387523651123] |
cbcf119f-9318-46fd-b127-f3c9322389d3 | audio-source-separation-with-discriminative | 1412.7022 | null | http://arxiv.org/abs/1412.7022v3 | http://arxiv.org/pdf/1412.7022v3.pdf | Audio Source Separation with Discriminative Scattering Networks | In this report we describe an ongoing line of research for solving
single-channel source separation problems. Many monaural signal decomposition
techniques proposed in the literature operate on a feature space consisting of
a time-frequency representation of the input data. A challenge faced by these
approaches is to effectively exploit the temporal dependencies of the signals
at scales larger than the duration of a time-frame. In this work we propose to
tackle this problem by modeling the signals using a time-frequency
representation with multiple temporal resolutions. The proposed representation
consists of a pyramid of wavelet scattering operators, which generalizes
Constant Q Transforms (CQT) with extra layers of convolution and complex
modulus. We first show that learning standard models with this multi-resolution
setting improves source separation results over fixed-resolution methods. As
study case, we use Non-Negative Matrix Factorizations (NMF) that has been
widely considered in many audio application. Then, we investigate the inclusion
of the proposed multi-resolution setting into a discriminative training regime.
We discuss several alternatives using different deep neural network
architectures. | ['Yann Lecun', 'Joan Bruna', 'Pablo Sprechmann'] | 2014-12-22 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [ 4.67440933e-01 -4.43369180e-01 2.48536497e-01 -2.57460266e-01
-1.29631865e+00 -5.88834107e-01 4.49410051e-01 1.41028631e-02
-5.16078234e-01 5.97629249e-01 2.94196665e-01 1.00022681e-01
-6.62897170e-01 -5.13355136e-01 -5.42973280e-01 -9.48058367e-01
-1.78812519e-01 -8.90958533e-02 1.40809685e-01 -3.47127259e-01
-1.40110694e-03 4.72502202e-01 -1.92774248e+00 5.84697127e-01
6.71335876e-01 1.14980137e+00 1.56299755e-01 9.33554471e-01
2.64662027e-01 4.58635420e-01 -7.28345990e-01 -7.21037679e-04
4.54434872e-01 -4.64949429e-01 -6.37293518e-01 -1.49925668e-02
5.48055053e-01 1.54157564e-01 -1.81171536e-01 8.98532093e-01
6.91038907e-01 5.44175148e-01 4.98519093e-01 -1.04743040e+00
-2.08454698e-01 5.82751215e-01 -5.26938975e-01 7.85577774e-01
2.89018512e-01 -4.18625355e-01 1.08374882e+00 -7.84243047e-01
2.05053285e-01 1.11156774e+00 8.59110177e-01 1.53111517e-01
-1.42601442e+00 -4.59104896e-01 -7.50778615e-02 4.33722585e-01
-1.27393067e+00 -5.86816728e-01 9.42778409e-01 -4.41176146e-01
9.53835845e-01 4.01910245e-01 3.38509172e-01 9.18834388e-01
6.80723488e-02 5.56566358e-01 1.32574785e+00 -6.74493372e-01
1.92722842e-01 -2.57282525e-01 2.53816009e-01 6.75162077e-02
-1.45483658e-01 7.31510669e-02 -9.43339109e-01 -2.82378346e-01
6.98610306e-01 -1.96289137e-01 -5.99588454e-01 -1.17279831e-02
-1.08291566e+00 8.60286593e-01 1.57634974e-01 8.53643179e-01
-5.07846713e-01 6.79926425e-02 4.60912108e-01 4.22270715e-01
5.95815480e-01 4.62174177e-01 -4.51876163e-01 -1.17923848e-01
-1.29857326e+00 2.37701342e-01 5.83375156e-01 3.05511266e-01
5.31884968e-01 3.28705668e-01 -1.38748735e-01 8.83162081e-01
-9.24277306e-02 2.00890437e-01 7.05088019e-01 -8.94684970e-01
2.07676053e-01 -2.30854556e-01 -5.24997450e-02 -7.93288231e-01
-5.15609384e-01 -7.74652541e-01 -8.68218005e-01 8.05462897e-02
4.75839198e-01 -1.41387165e-01 -7.22688377e-01 1.95821285e+00
3.17136854e-01 6.97625637e-01 1.46118477e-01 1.01598537e+00
3.93291235e-01 6.11742795e-01 -1.92227706e-01 -6.15204990e-01
1.37854004e+00 -7.74103522e-01 -9.09079731e-01 2.40302429e-01
7.78294355e-02 -1.06572461e+00 6.34119451e-01 9.34620380e-01
-1.20282412e+00 -8.98669779e-01 -1.01582551e+00 -2.58340184e-02
-2.60831714e-01 2.79635340e-01 5.54257393e-01 7.20275760e-01
-1.24047589e+00 9.31959689e-01 -8.29679072e-01 -1.26141980e-01
-5.45237213e-02 3.15528154e-01 -3.02750498e-01 1.28506422e-01
-1.40225720e+00 5.56649446e-01 6.39341027e-02 1.70848340e-01
-6.45777345e-01 -7.67500520e-01 -6.89469278e-01 1.31122142e-01
1.13671847e-01 -4.94324207e-01 1.36887002e+00 -1.12030482e+00
-1.36837769e+00 4.71263230e-01 -4.32413071e-01 -8.16754282e-01
1.48674488e-01 -4.92132694e-01 -6.82436228e-01 5.06659925e-01
-4.94773909e-02 -7.14617446e-02 1.54887331e+00 -9.12523746e-01
-4.91996586e-01 -3.64600301e-01 2.43499950e-01 -2.12647635e-02
-3.03592026e-01 2.52368808e-01 2.29828671e-01 -1.11085415e+00
2.22029954e-01 -6.86706960e-01 -1.99137136e-01 -4.65731144e-01
6.67460263e-02 7.29846186e-04 5.29621422e-01 -8.65676463e-01
1.25005579e+00 -2.39839554e+00 5.67256212e-01 1.34625332e-02
1.37561053e-01 1.76928043e-01 -2.47129068e-01 6.04436278e-01
-5.07024705e-01 -2.76154786e-01 -2.20649689e-01 -5.34129202e-01
-9.27910805e-02 4.58442196e-02 -6.20835960e-01 6.29677296e-01
2.79979050e-01 2.70855516e-01 -5.72554588e-01 -9.06081572e-02
9.20834094e-02 1.01515424e+00 -4.84195828e-01 1.52445391e-01
2.80571043e-01 6.09318674e-01 -1.29185632e-01 1.51816115e-01
8.44752431e-01 1.14233151e-01 -1.04416534e-02 -5.61984241e-01
-4.47156131e-01 4.12325233e-01 -1.52500820e+00 1.91525114e+00
-7.57527292e-01 6.81525648e-01 5.12616098e-01 -1.39049017e+00
7.24359155e-01 6.95652127e-01 8.06688070e-01 -4.65674549e-01
7.33570084e-02 4.03715461e-01 2.23005220e-01 -3.49981368e-01
4.54427630e-01 -5.84236264e-01 2.39615813e-01 4.04746145e-01
5.69197118e-01 1.46638289e-01 2.10890368e-01 -1.79608598e-01
1.00909090e+00 -1.03974797e-01 2.74178892e-01 -2.17380106e-01
7.73369789e-01 -6.26772404e-01 3.60209525e-01 5.98991036e-01
-3.10241189e-02 7.88964450e-01 9.88475531e-02 -2.58031219e-01
-6.84243858e-01 -9.29283082e-01 -3.57410520e-01 1.28405070e+00
-4.43443537e-01 -6.05587661e-01 -7.05434084e-01 -9.52384248e-02
-2.41797045e-01 2.97491580e-01 -6.33424997e-01 -1.04584128e-01
-7.70070255e-01 -9.13289309e-01 5.76326728e-01 5.58748782e-01
1.53500468e-01 -6.79891169e-01 -8.05947125e-01 3.18040282e-01
-3.91962141e-01 -1.21910620e+00 -1.33495852e-01 7.06840217e-01
-8.23964357e-01 -7.03501463e-01 -7.35775411e-01 -5.50446928e-01
-9.73862037e-02 4.06072617e-01 9.52014267e-01 -5.09906471e-01
-3.05971056e-01 7.82611728e-01 -6.14415884e-01 -2.87716448e-01
-1.38912737e-01 -5.57781793e-02 4.12363142e-01 5.95485151e-01
1.66370749e-01 -1.31806624e+00 -4.73608047e-01 1.34904236e-02
-1.29271054e+00 -3.20824862e-01 5.39683282e-01 6.84102714e-01
5.97474039e-01 4.62916523e-01 7.39140391e-01 -2.81484276e-01
7.42408633e-01 -3.93390954e-01 -2.59173214e-01 -9.66073945e-02
-5.73800020e-02 1.57104090e-01 7.91777968e-01 -7.49400914e-01
-9.74746048e-01 -3.97211947e-02 -2.87699074e-01 -6.35659397e-01
-2.45315239e-01 4.88106638e-01 -1.71678755e-02 -2.58670002e-01
6.97766781e-01 3.18383247e-01 -3.85756254e-01 -9.01588738e-01
2.57633746e-01 5.54281056e-01 4.94742483e-01 -7.81937361e-01
7.91922748e-01 5.91052532e-01 1.00617766e-01 -9.94456053e-01
-8.66372705e-01 -7.79511392e-01 -8.11664522e-01 -7.94094652e-02
8.40654135e-01 -9.71041858e-01 -4.65482444e-01 3.11137199e-01
-1.12838209e+00 3.48868333e-02 -5.51353335e-01 8.50854874e-01
-6.24398768e-01 3.65249187e-01 -5.95363379e-01 -1.07304919e+00
-1.31375566e-01 -8.40935886e-01 1.24447167e+00 1.47081852e-01
-1.86878350e-02 -7.93061316e-01 3.88243109e-01 9.71059054e-02
5.94544768e-01 1.72837690e-01 5.43627322e-01 -5.49898326e-01
-2.73505181e-01 1.37466609e-01 2.33803079e-01 6.34616673e-01
1.92643732e-01 -4.18011338e-01 -1.48123121e+00 -4.89007741e-01
7.17210948e-01 -1.73940137e-01 1.20817471e+00 7.21409500e-01
9.77931738e-01 2.64717415e-02 9.22500119e-02 6.79792941e-01
1.39489520e+00 1.07539468e-01 3.21599901e-01 1.58772185e-01
4.31331784e-01 6.55122757e-01 4.40151274e-01 5.51430285e-01
-8.30938593e-02 7.64628768e-01 1.73518136e-01 -4.52380776e-02
-8.42478052e-02 1.48672402e-01 3.97725433e-01 1.00029325e+00
-4.56400871e-01 1.38555557e-01 -5.84816992e-01 6.26384258e-01
-1.76165855e+00 -1.12422490e+00 -4.46481667e-02 2.24004483e+00
6.42938316e-01 -9.86112058e-02 3.47604603e-01 9.66726601e-01
5.97269416e-01 2.90560216e-01 -2.15589076e-01 -3.34937364e-01
-2.69378155e-01 8.10899854e-01 1.00036189e-01 4.95601237e-01
-1.25915956e+00 1.56249985e-01 5.85043573e+00 1.02112556e+00
-1.46369970e+00 4.16152745e-01 -1.03885457e-02 -1.31779954e-01
-6.18214794e-02 -2.40427762e-01 -4.83214349e-01 8.34947303e-02
1.38975477e+00 -5.74968532e-02 5.22929966e-01 3.61678958e-01
1.54655337e-01 3.24224979e-02 -1.14300001e+00 1.21551943e+00
1.32208079e-01 -8.74611080e-01 -2.79202253e-01 -1.04093611e-01
3.59695315e-01 -1.51385561e-01 3.32885355e-01 2.63051838e-01
-3.86144549e-01 -1.08179450e+00 8.01464617e-01 4.41626251e-01
6.05551302e-01 -7.31068254e-01 4.37704146e-01 2.02961847e-01
-1.60875654e+00 -3.05079609e-01 -3.24162841e-01 -2.23050028e-01
2.77830184e-01 7.81324804e-01 -4.02226269e-01 1.00313878e+00
8.58814120e-01 7.64865398e-01 -2.78625697e-01 1.10764956e+00
1.90525785e-01 8.15173805e-01 -4.44790453e-01 7.57582247e-01
2.80921590e-02 -1.14713676e-01 8.15567672e-01 1.44187677e+00
5.28165221e-01 1.19339973e-01 9.40494984e-02 5.22986650e-01
2.98540026e-01 3.37112360e-02 -2.89981723e-01 -1.95641210e-03
-5.33148237e-02 1.34344578e+00 -6.57105565e-01 6.62668720e-02
-5.24676085e-01 7.10485578e-01 3.60435247e-02 4.08087701e-01
-7.65163481e-01 -2.66896725e-01 7.99003303e-01 9.71415937e-02
6.27697766e-01 -4.39433634e-01 2.24944562e-01 -1.24612880e+00
1.54760942e-01 -8.91442776e-01 4.28196758e-01 -6.38287783e-01
-1.33418167e+00 9.68401551e-01 3.12764086e-02 -1.57202125e+00
-2.89059222e-01 -5.95582306e-01 -3.25582862e-01 1.12029159e+00
-1.77803075e+00 -9.76739466e-01 1.36886433e-01 9.43216205e-01
6.48177803e-01 4.24768031e-02 9.33931351e-01 5.95871210e-01
-1.12950802e-01 3.12823266e-01 1.88300475e-01 -2.03522339e-01
6.51696026e-01 -1.30713212e+00 -3.54752168e-02 9.42943215e-01
5.35476983e-01 7.85721540e-01 8.62717927e-01 -4.67352271e-02
-1.14049125e+00 -7.85004556e-01 6.98823154e-01 -1.29246995e-01
7.71532416e-01 -3.59214097e-01 -9.92870748e-01 4.61488396e-01
3.75704736e-01 2.78927624e-01 1.05568063e+00 1.69790193e-01
-5.20864904e-01 -2.84305811e-01 -8.64008307e-01 -1.62124306e-01
7.17138946e-01 -9.45601165e-01 -7.22921789e-01 8.35910439e-02
6.33941591e-01 -2.49318242e-01 -9.10359561e-01 3.04029286e-01
4.33656096e-01 -1.20009696e+00 1.21833301e+00 -5.76776922e-01
1.40615180e-01 -4.05250669e-01 -5.39362729e-01 -1.44538486e+00
-3.71089637e-01 -9.59313571e-01 -2.08897561e-01 1.12900805e+00
-1.06978953e-01 -4.59571004e-01 1.66112974e-01 -5.70923947e-02
-9.06158537e-02 -5.46186864e-01 -1.28752220e+00 -7.52540767e-01
8.47223587e-03 -7.77744293e-01 2.03721106e-01 8.84827971e-01
-1.57404467e-01 4.82316881e-01 -4.24483091e-01 4.84173596e-01
5.75941980e-01 4.46847051e-01 1.97167054e-01 -1.37559056e+00
-8.53952587e-01 -2.53448278e-01 -4.08049911e-01 -7.67859876e-01
1.46753743e-01 -6.50035977e-01 -1.09220378e-01 -1.07423925e+00
-1.77315012e-01 -1.65763064e-04 -7.15147018e-01 3.44365798e-02
9.98981148e-02 3.24796885e-01 2.60983258e-01 9.51108895e-03
-2.85617113e-01 4.66070443e-01 8.51780474e-01 2.20335517e-02
-1.23601317e-01 2.70303637e-01 -3.81217122e-01 7.26561487e-01
6.92036152e-01 -3.51934314e-01 -5.37974954e-01 -3.61271501e-01
1.92042992e-01 4.45267916e-01 4.51465636e-01 -1.34957337e+00
1.42262295e-01 2.00825304e-01 2.82049507e-01 -3.98135334e-01
7.60299325e-01 -8.84113610e-01 1.71331599e-01 -2.57256534e-02
-4.11759734e-01 2.45860200e-02 4.42553729e-01 6.65779591e-01
-8.01501513e-01 -1.27761558e-01 8.14334691e-01 -2.51729637e-02
-3.07979107e-01 -1.07685558e-01 -4.96058166e-01 -1.21959254e-01
5.01848519e-01 -1.06661739e-02 1.15427986e-01 -4.36658233e-01
-1.12721062e+00 -5.13443112e-01 -2.47818679e-01 3.45537364e-01
5.34769952e-01 -1.28758025e+00 -6.85160935e-01 1.10457398e-01
-2.66239196e-01 -4.17009473e-01 6.22473538e-01 1.25623775e+00
1.84506282e-01 4.11590457e-01 -4.48920697e-01 -5.75210631e-01
-1.31909645e+00 5.52769840e-01 4.89434153e-01 -3.19898367e-01
-4.45344836e-01 8.57403815e-01 3.59355122e-01 3.43016610e-02
8.28243513e-03 -5.11883497e-01 -5.81721902e-01 4.58230644e-01
7.38464296e-01 4.19639736e-01 4.45393354e-01 -1.08739913e+00
-2.82550007e-01 7.63683319e-01 1.81838199e-01 -4.94405448e-01
1.71014810e+00 -9.37594920e-02 -2.26401523e-01 8.13729763e-01
1.46079361e+00 3.71585578e-01 -1.00356638e+00 -3.36515695e-01
-7.34205991e-02 -3.67505193e-01 2.24986702e-01 -4.36456770e-01
-1.05311477e+00 1.19609177e+00 7.89881706e-01 6.11323774e-01
1.77641404e+00 -2.91467279e-01 6.75251365e-01 8.07449892e-02
3.41953427e-01 -7.82167614e-01 5.84575571e-02 3.80787283e-01
9.46693838e-01 -7.61281550e-01 -2.93923348e-01 -2.37921461e-01
-1.36973739e-01 1.44427776e+00 4.19819728e-02 -2.80816942e-01
7.51340866e-01 3.30525249e-01 2.76462138e-02 3.46082486e-02
-5.29878080e-01 -6.13954663e-01 5.67442834e-01 6.50704026e-01
5.87393761e-01 -7.15129599e-02 -3.73516738e-01 1.02143776e+00
-3.19145530e-01 -2.06112325e-01 4.51153189e-01 6.86311722e-01
-3.10818851e-01 -1.32738709e+00 -7.77373254e-01 8.07522982e-02
-8.00595343e-01 -1.65746287e-01 -1.71108190e-02 4.48182672e-01
2.86374152e-01 1.17413986e+00 -1.95105121e-01 -3.28675002e-01
3.45849335e-01 3.57001394e-01 6.91620827e-01 -4.55707192e-01
-7.74751186e-01 7.18613923e-01 -1.76612750e-01 -5.30861855e-01
-1.04814947e+00 -8.35026383e-01 -9.00550723e-01 3.40067714e-01
-1.17639787e-01 3.78069490e-01 6.11792386e-01 7.76170433e-01
1.46193922e-01 9.12478328e-01 5.40152252e-01 -1.16733766e+00
-5.32327533e-01 -1.08477640e+00 -8.64205718e-01 3.98958385e-01
1.03199220e+00 -6.98558807e-01 -5.06716609e-01 3.14037204e-01] | [15.420822143554688, 5.560679912567139] |
98b196fc-1f0d-4541-a276-0150b94cc136 | a-fast-ilp-based-heuristic-for-the-robust | 1704.04640 | null | http://arxiv.org/abs/1704.04640v1 | http://arxiv.org/pdf/1704.04640v1.pdf | A fast ILP-based Heuristic for the robust design of Body Wireless Sensor Networks | We consider the problem of optimally designing a body wireless sensor
network, while taking into account the uncertainty of data generation of
biosensors. Since the related min-max robustness Integer Linear Programming
(ILP) problem can be difficult to solve even for state-of-the-art commercial
optimization solvers, we propose an original heuristic for its solution. The
heuristic combines deterministic and probabilistic variable fixing strategies,
guided by the information coming from strengthened linear relaxations of the
ILP robust model, and includes a very large neighborhood search for reparation
and improvement of generated solutions, formulated as an ILP problem solved
exactly. Computational tests on realistic instances show that our heuristic
finds solutions of much higher quality than a state-of-the-art solver and than
an effective benchmark heuristic. | ["Fabio D'Andreagiovanni", 'Antonella Nardin', 'Enrico Natalizio'] | 2017-04-15 | null | null | null | null | ['robust-design'] | ['miscellaneous'] | [ 6.19185627e-01 9.96393025e-01 -5.54649234e-01 -3.48327532e-02
-9.41441476e-01 -6.11125410e-01 -3.22784364e-01 3.58526915e-01
-3.07246536e-01 1.34607840e+00 -1.40367493e-01 -9.95487943e-02
-8.92054617e-01 -7.66766071e-01 -1.10568273e+00 -8.23513329e-01
-4.26629335e-01 7.36252010e-01 -5.79398312e-02 -2.33646366e-03
1.85638815e-02 1.81497335e-01 -1.08934152e+00 -5.10370024e-02
9.81367648e-01 1.21253598e+00 -6.86999038e-02 3.31095994e-01
8.40012580e-02 4.74369191e-02 -6.78379536e-01 -1.88935697e-01
2.58506536e-01 -1.46883100e-01 -7.69201517e-01 -5.74160405e-02
-4.34404820e-01 1.90120488e-01 3.30403030e-01 9.93098915e-01
6.81842923e-01 -6.03643619e-02 6.09379783e-02 -1.69531155e+00
-3.32028002e-01 1.11089838e+00 -8.11722338e-01 -3.25667977e-01
7.79945314e-01 1.43768061e-02 6.66530490e-01 1.89464986e-01
7.06464469e-01 1.20036578e+00 6.03102922e-01 6.21668875e-01
-1.64628446e+00 -4.69001323e-01 4.57424641e-01 -1.25265151e-01
-1.36112654e+00 -3.84980477e-02 4.85911906e-01 2.19166614e-02
7.25453913e-01 6.95704997e-01 8.15350175e-01 9.74447906e-01
3.95855665e-01 4.94516641e-01 1.10764647e+00 -1.30765289e-01
8.21119130e-01 6.21668175e-02 5.08500356e-03 3.67416590e-01
8.80036592e-01 3.09418142e-01 -3.40676576e-01 -7.04701781e-01
2.02059239e-01 -3.57114077e-01 -3.51266474e-01 -3.62561673e-01
-9.78126287e-01 7.23750651e-01 2.81721443e-01 8.32518190e-02
-7.28982449e-01 3.65441233e-01 8.41298848e-02 -7.08852634e-02
2.04731748e-01 7.68893063e-01 -7.37928808e-01 1.26294553e-01
-9.72553492e-01 6.29287422e-01 1.08781254e+00 8.42543781e-01
1.40460283e-01 -1.02875799e-01 -2.50175267e-01 3.35957915e-01
3.66511941e-01 7.38271356e-01 -1.30078062e-01 -1.01995301e+00
5.32037854e-01 3.54255795e-01 3.87594491e-01 -9.01404798e-01
-9.02608752e-01 -3.70561719e-01 -7.29406893e-01 -3.15453745e-02
3.36583376e-01 -9.10077691e-01 -7.48648465e-01 1.67923903e+00
3.41902852e-01 2.59126574e-01 6.52337000e-02 6.73649669e-01
4.88723665e-01 9.71878350e-01 -1.89237267e-01 -1.20494056e+00
1.05470586e+00 -2.75255352e-01 -1.06604767e+00 -3.10348541e-01
-1.39279673e-02 -1.11831702e-01 -1.02678336e-01 8.26702476e-01
-1.51439607e+00 2.12919548e-01 -1.25653350e+00 6.10042512e-01
-2.15388238e-01 -8.08595181e-01 6.58357561e-01 1.02796280e+00
-9.22220945e-01 4.76886183e-01 -8.02900672e-01 -1.24702364e-01
2.15152562e-01 1.03143704e+00 -2.72316068e-01 -2.45737478e-01
-1.01324320e+00 7.90281773e-01 4.59235847e-01 4.67257887e-01
-6.49887204e-01 -8.65429580e-01 -9.47624147e-01 -2.00649485e-01
9.94843245e-01 -7.63062537e-01 7.94020236e-01 -5.26571572e-01
-1.78969061e+00 2.94449091e-01 -1.99241400e-01 -5.30923486e-01
4.88852918e-01 4.70173895e-01 -1.03144124e-01 -9.44515616e-02
-2.70097643e-01 3.78294140e-01 5.61037779e-01 -1.50246310e+00
-2.80015796e-01 -3.80237639e-01 2.33228326e-01 -3.56984407e-01
1.15856566e-01 4.24298923e-04 -2.36914724e-01 -2.53262103e-01
1.98595703e-01 -9.88934219e-01 -1.09726679e+00 -2.05302835e-01
-8.60449433e-01 1.22333489e-01 2.78375536e-01 -6.96595252e-01
1.33298159e+00 -1.49882972e+00 6.62697017e-01 9.86189902e-01
-3.13815325e-01 -3.09637487e-01 -2.96622932e-01 2.22928330e-01
-9.82882008e-02 3.01482469e-01 -4.72951144e-01 -4.85444665e-02
1.39472291e-01 6.39408767e-01 1.81821182e-01 6.94792449e-01
-2.07419813e-01 8.03957224e-01 -9.51727867e-01 -3.70787501e-01
-9.80780721e-02 2.51796603e-01 -6.18720055e-01 1.01249166e-01
-6.44513011e-01 2.11291894e-01 -7.08766997e-01 7.70788252e-01
9.25550103e-01 7.12153241e-02 8.01771224e-01 -2.59914011e-01
-7.54902586e-02 -5.04426837e-01 -1.80794406e+00 1.53814220e+00
-4.14683580e-01 -1.60294533e-01 4.33235258e-01 -1.27971697e+00
6.40433013e-01 2.58724451e-01 1.12401521e+00 -7.62240946e-01
2.08579913e-01 4.80964035e-02 -3.37113708e-01 -5.53926587e-01
2.27117553e-01 -1.23194218e-01 -4.65476692e-01 3.49287033e-01
-2.04590097e-01 -2.61187732e-01 2.78250307e-01 -2.03287780e-01
1.32346070e+00 2.17183139e-02 2.89126933e-01 -5.91281056e-01
4.07050729e-01 -1.23984061e-01 9.75569487e-01 8.81965518e-01
2.60186493e-01 4.64763135e-01 9.15644109e-01 -3.03438902e-01
-5.19681811e-01 -7.37374365e-01 -3.52471828e-01 5.82851946e-01
2.49426216e-01 -2.27971300e-01 -8.48484755e-01 -3.46138060e-01
3.06166679e-01 7.53152370e-01 -9.01688695e-01 2.03121647e-01
-5.41041911e-01 -1.51064301e+00 1.26810461e-01 1.00116737e-01
-3.85191441e-02 -4.98738825e-01 -6.84018672e-01 6.95672333e-01
-2.74198115e-01 -1.15837479e+00 1.98985428e-01 5.81694484e-01
-7.53043890e-01 -9.35668230e-01 -5.24443567e-01 -1.98131025e-01
8.93037856e-01 -5.66385865e-01 1.00082648e+00 9.41782445e-02
-6.33988619e-01 1.26331478e-01 -9.85895097e-02 -6.03212297e-01
-1.42511368e-01 -6.59156814e-02 -4.85844351e-02 -2.20940620e-01
-4.81071621e-01 -3.92423123e-01 -1.37233049e-01 4.72437471e-01
-1.02243280e+00 -4.77546632e-01 1.22187212e-01 7.08799720e-01
1.00107181e+00 8.02088857e-01 6.32670045e-01 -7.15969384e-01
6.16683185e-01 -7.18509495e-01 -1.18606174e+00 5.82187772e-01
-4.89457905e-01 6.79622591e-01 3.54648799e-01 -5.02050519e-01
-5.64657927e-01 5.73622942e-01 -2.99765747e-02 2.18234748e-01
4.96337295e-01 6.05581224e-01 -5.16561210e-01 -6.00406229e-01
3.89264077e-01 -1.59895971e-01 -1.80242643e-01 -2.17825058e-03
2.97795206e-01 1.73255205e-01 2.17034519e-01 -9.16277468e-01
7.93268561e-01 3.15408826e-01 5.19372761e-01 -4.09371018e-01
-6.14924073e-01 6.26586154e-02 4.91690561e-02 -1.87427282e-01
7.56185114e-01 -4.13984925e-01 -1.31401765e+00 -1.19893495e-02
-1.00237334e+00 -2.04291508e-01 -5.01791477e-01 -1.12480953e-01
-6.44569337e-01 8.31143185e-02 9.98931974e-02 -1.23637795e+00
-9.66018587e-02 -1.27190149e+00 8.51688802e-01 2.44134083e-01
-2.45481670e-01 -7.87577927e-01 1.68038115e-01 4.80096340e-01
3.90158117e-01 1.17256916e+00 5.90359509e-01 -1.49509370e-01
-5.44018865e-01 -1.72730222e-01 1.77101359e-01 -4.03901339e-01
-2.17938989e-01 7.72578269e-02 -6.88345492e-01 -3.28650892e-01
-1.19503085e-02 -1.07792497e-03 4.45182651e-01 7.86591291e-01
1.27867818e+00 -8.96408856e-01 -7.89159358e-01 6.58522487e-01
1.82581842e+00 3.29008698e-02 7.04542875e-01 3.34594071e-01
-4.61490750e-02 5.44564426e-01 7.91440129e-01 9.82549965e-01
4.00091261e-01 7.99324095e-01 1.06415391e+00 1.88257203e-01
7.95239627e-01 1.99837685e-01 9.69908759e-02 -2.86197305e-01
-1.35088459e-01 -5.64506114e-01 -6.17040694e-01 4.50089991e-01
-2.18801522e+00 -7.98643351e-01 1.43420354e-01 2.26310420e+00
9.37509477e-01 1.94356546e-01 3.32364380e-01 4.08661723e-01
6.43423438e-01 -2.77471226e-02 -7.54632473e-01 -7.21891046e-01
-1.97128076e-02 3.46922994e-01 1.20942366e+00 5.52008033e-01
-7.72618771e-01 1.43988252e-01 7.49009657e+00 5.32555938e-01
-6.62169039e-01 -9.42690596e-02 6.61081791e-01 -5.86557269e-01
-7.03955412e-01 -3.52778822e-01 -6.52716398e-01 5.86632609e-01
1.21142757e+00 -4.51697946e-01 8.62483084e-01 4.55983013e-01
3.70084375e-01 -4.24081534e-01 -1.08983684e+00 7.73305357e-01
-3.89613323e-02 -1.47971117e+00 -6.27512872e-01 3.69941711e-01
1.20273721e+00 -4.29895848e-01 3.01999114e-02 -2.48891920e-01
6.11391366e-01 -1.09185028e+00 7.52514184e-01 4.36006010e-01
2.93137282e-01 -1.07155001e+00 8.95150542e-01 1.69537559e-01
-8.32290351e-01 -7.34875560e-01 -3.19222838e-01 1.97807997e-01
4.79899287e-01 1.20475245e+00 -2.55595446e-01 8.83040667e-01
6.83436513e-01 3.09381951e-02 4.03194688e-02 1.46162164e+00
-1.26187250e-01 1.67557821e-01 -9.19088840e-01 -9.79100093e-02
5.02167642e-02 1.36655554e-01 6.28168762e-01 8.52560103e-01
3.54958415e-01 4.55948472e-01 1.86214745e-01 8.95023882e-01
1.31772533e-01 -1.31806806e-01 -2.29621097e-01 8.96033719e-02
4.23170179e-01 1.04070079e+00 -6.20628357e-01 4.11031157e-01
2.89360553e-01 4.59775865e-01 -1.04384467e-01 4.01381075e-01
-1.03983355e+00 -1.03111394e-01 6.18061662e-01 9.89746600e-02
1.40025988e-01 3.32897976e-02 -7.02108979e-01 -5.45418024e-01
2.49554932e-01 -7.99238682e-01 6.30493462e-01 -4.29654479e-01
-9.87768173e-01 3.14349502e-01 2.48287469e-01 -6.24431312e-01
-4.39189970e-01 -5.08389294e-01 -4.17145073e-01 5.10194659e-01
-1.51870668e+00 -6.70232952e-01 -6.53200969e-02 3.78119826e-01
-1.43942535e-01 5.88101506e-01 7.39003360e-01 7.33424649e-02
-8.90665352e-01 7.00825155e-01 9.83228758e-02 -7.74710298e-01
-5.23215742e-04 -9.79026377e-01 -2.43170217e-01 6.60976350e-01
-8.61586630e-01 2.88062900e-01 1.33246219e+00 -6.16109252e-01
-2.21734428e+00 -8.59300196e-01 5.14376938e-01 5.06581925e-02
6.24505103e-01 -2.38935307e-01 -2.39892751e-01 4.00481671e-01
1.15556113e-01 1.71198636e-01 4.87575352e-01 -1.59214567e-02
2.29929700e-01 -3.13812494e-01 -2.06075549e+00 1.58348411e-01
1.00326073e+00 4.19921488e-01 -1.86951265e-01 8.02223444e-01
7.74249494e-01 -7.38313019e-01 -1.08760834e+00 7.15352774e-01
5.42210817e-01 -2.32547015e-01 1.23484778e+00 -2.07209557e-01
5.80099295e-04 -1.85011283e-01 -1.41615584e-01 -1.16148961e+00
-1.39996052e-01 -1.26344681e+00 -2.28895351e-01 1.14300108e+00
6.64012969e-01 -5.34706652e-01 7.34513700e-01 1.17284179e+00
3.38305742e-01 -8.75724554e-01 -1.41734827e+00 -1.04420674e+00
-1.55360863e-01 -2.80077100e-01 8.59068632e-01 5.87391257e-01
5.69261312e-01 -4.67094272e-01 -3.86969775e-01 5.01904428e-01
1.14167225e+00 -3.79887372e-02 1.45801544e-01 -1.23089981e+00
-5.76962948e-01 -1.10756390e-01 -3.88453513e-01 -2.04181045e-01
2.78407544e-01 -3.37266713e-01 5.47541022e-01 -1.73427856e+00
-1.57933142e-02 -6.22727036e-01 -2.68073469e-01 6.98570311e-01
5.89386970e-02 -2.37993989e-02 8.79248604e-02 -9.11187649e-01
-7.99478948e-01 1.30339026e-01 8.50435317e-01 -4.57203627e-01
-5.08405507e-01 2.44402409e-01 -1.04722178e+00 4.03347880e-01
6.41391695e-01 -9.00700629e-01 -1.76631808e-01 -5.55623136e-02
9.70125556e-01 6.66026056e-01 5.37412763e-02 -1.01578856e+00
1.89571962e-01 -6.69677198e-01 -7.00173378e-02 -3.54072809e-01
2.97583908e-01 -1.31522262e+00 8.70090783e-01 7.58107543e-01
-2.34824464e-01 -1.96776256e-01 3.88499200e-01 4.22899693e-01
3.67447883e-01 -5.03139019e-01 7.64664710e-01 -7.16852769e-02
6.97819795e-03 1.32336944e-01 -4.29467082e-01 7.02896640e-02
1.41889191e+00 -1.72248170e-01 -3.93041641e-01 -2.13617042e-01
-8.40965390e-01 7.80525923e-01 1.98711202e-01 1.85662895e-01
4.37765509e-01 -9.48953450e-01 -6.60016418e-01 -1.48323476e-01
-2.22181305e-01 7.73870349e-02 1.55422270e-01 5.69775939e-01
-2.50883520e-01 2.48056278e-01 -1.43408269e-01 -3.75844687e-01
-1.06024981e+00 9.31556165e-01 3.95241141e-01 -6.92694068e-01
-1.11025728e-01 6.93905234e-01 -5.42256713e-01 -3.65530252e-01
4.90862697e-01 -5.17396808e-01 6.15805835e-02 3.80513333e-02
5.38249552e-01 7.71545887e-01 -7.60295764e-02 4.80570011e-02
-6.68120503e-01 6.45267308e-01 5.97442329e-01 -1.58741802e-01
1.70806348e+00 -1.34046659e-01 -4.63799357e-01 -2.45730862e-01
6.21493757e-01 -9.10412297e-02 -8.47782850e-01 1.52921006e-01
-5.65269031e-02 -4.87950556e-02 1.00718915e-01 -1.14503551e+00
-1.48066342e+00 -4.72539738e-02 4.99734998e-01 2.50088274e-01
1.43890202e+00 -2.08420858e-01 5.63695729e-01 4.51944083e-01
9.22217906e-01 -1.22720683e+00 -5.19695282e-01 -2.20097490e-02
9.25156415e-01 -7.57522166e-01 7.46743262e-01 -7.22607851e-01
-1.48783773e-01 8.60720038e-01 2.69590646e-01 7.18161510e-03
7.43297756e-01 1.06030226e+00 -4.47546870e-01 2.41392165e-01
-6.71556056e-01 1.22599728e-01 7.04044923e-02 7.95956790e-01
-2.16263458e-01 2.47059956e-01 -7.07389832e-01 1.06138504e+00
7.99972489e-02 2.78959185e-01 5.36311746e-01 1.02807128e+00
-3.25107038e-01 -1.26595986e+00 -7.69174516e-01 2.29060382e-01
-6.42568290e-01 3.40507895e-01 -1.00903198e-01 4.12961453e-01
4.94190186e-01 1.38864791e+00 -4.51282144e-01 -7.21688271e-02
4.13842827e-01 -4.90865886e-01 7.15793073e-01 -1.91895619e-01
-9.22889173e-01 -1.86494231e-01 3.94551665e-01 -1.12077618e+00
-4.00913686e-01 -6.91826105e-01 -1.42476940e+00 3.12149146e-04
-3.18006605e-01 2.60049403e-01 9.86805379e-01 1.00095499e+00
1.52227312e-01 7.76094794e-01 5.11960745e-01 -9.08683121e-01
-3.44233394e-01 -6.60279319e-02 -5.92801690e-01 -3.35778177e-01
2.12177590e-01 -5.01872718e-01 -1.94564179e-01 -5.80748498e-01] | [5.6157402992248535, 3.4734861850738525] |
3bdc17c0-0e69-4312-9778-f577992cf4e3 | machine-learning-for-clouds-and-climate | null | null | https://www.semanticscholar.org/paper/Machine-Learning-for-Clouds-and-Climate-Beucler-Ebert%E2%80%90Uphoff/ea0ccde42b1f4517e87ee2e3f77fb6c06671666b | https://www.essoar.org/doi/abs/10.1002/essoar.10506925.1 | Machine Learning for Clouds and Climate | Machine learning (ML) algorithms are powerful tools to build models of clouds and climate that are more faithful to the rapidly-increasing volumes of Earth system data than commonly-used semiempirical models. Here, we review ML tools, including interpretable and physics-guided ML, and outline how they can be applied to cloud-related processes in the climate system, including radiation, microphysics, convection, and cloud detection, classification, emulation, and uncertainty quantification. We additionally provide a short guide to get started with ML and survey the frontiers of ML for clouds and climate. | ['P. Gentine', 'M. Pritchard', 'S. Rasp', 'I. Ebert‐Uphoff', 'Tom Beucler'] | 2020-01-01 | null | null | null | open-access-2020-1 | ['cloud-detection'] | ['computer-vision'] | [-3.30943972e-01 -7.04076111e-01 -1.32569626e-01 -3.95627350e-01
-1.10811353e-01 -8.18769991e-01 8.60281527e-01 3.18383425e-01
1.29425660e-01 9.80437338e-01 -3.95330131e-01 -1.04875350e+00
2.75824189e-01 -8.70572686e-01 -1.18908718e-01 -8.81146252e-01
-2.75544107e-01 6.32582843e-01 -2.41138395e-02 -2.11161047e-01
2.15201631e-01 1.13967800e+00 -1.71344256e+00 -1.10165209e-01
1.26359773e+00 1.01179254e+00 4.90670800e-02 9.69541728e-01
-3.41814905e-01 8.29516411e-01 -2.36663818e-01 5.99080622e-01
-1.17799595e-01 -3.09805721e-01 -4.55771744e-01 5.25106005e-02
2.62705773e-01 -1.37290135e-01 -3.76788825e-02 7.72146046e-01
-2.30263397e-02 -3.74203082e-03 1.10174370e+00 -1.16937041e+00
-4.05689269e-01 -2.29610473e-01 -3.56429100e-01 1.81620508e-01
-2.86733508e-01 4.75980729e-01 7.17517316e-01 -1.12022269e+00
3.05031598e-01 1.02138233e+00 5.42283475e-01 1.37515798e-01
-1.09816456e+00 -5.36654770e-01 -1.58227142e-02 2.91412305e-02
-1.26561439e+00 -4.88150924e-01 3.55714470e-01 -1.13098156e+00
1.08572388e+00 7.42125809e-01 1.03688431e+00 1.03956409e-01
3.78999114e-01 4.73325878e-01 1.75764394e+00 -5.81252217e-01
5.35276771e-01 6.61303341e-01 1.22583531e-01 4.06239271e-01
3.34016651e-01 7.72691727e-01 -2.03904778e-01 -5.24186730e-01
4.95413989e-01 -1.51848227e-01 -2.24565208e-01 -3.28830123e-01
-5.51342666e-01 1.21242118e+00 3.78261656e-01 -1.35654435e-02
-2.53820002e-01 1.23200692e-01 -1.86711624e-02 1.44883573e-01
1.06113446e+00 8.14234793e-01 -9.91878569e-01 1.80099204e-01
-1.47958982e+00 8.38579655e-01 8.12994719e-01 7.32538760e-01
9.27368283e-01 6.76367283e-01 3.16214830e-01 4.28886026e-01
9.57851350e-01 1.53422558e+00 -2.69790828e-01 -1.04715919e+00
-2.32857987e-01 2.53182799e-02 4.60584015e-01 -4.85297322e-01
-1.94147155e-01 -4.11022842e-01 -8.46536517e-01 6.70025885e-01
-2.37610966e-01 -1.81919038e-01 -9.35038269e-01 6.84217989e-01
3.40084165e-01 7.05950931e-02 8.95571411e-02 6.49815321e-01
1.00643396e+00 1.09613264e+00 -7.70722236e-03 -6.10988975e-01
1.00863278e+00 -7.74031579e-01 -6.38468981e-01 -5.15416682e-01
3.59809667e-01 -8.51090610e-01 4.49255496e-01 -1.61055073e-01
-6.27062857e-01 -1.57126442e-01 -6.13152981e-01 2.80806333e-01
-6.98971450e-01 -2.77013689e-01 1.21450996e+00 6.59431577e-01
-9.28658724e-01 8.89563560e-01 -9.37674284e-01 -2.67911907e-02
-4.01204042e-02 -3.06808025e-01 3.39174330e-01 3.20176750e-01
-1.10827494e+00 1.33721793e+00 1.79827407e-01 5.42446792e-01
-7.99295962e-01 -8.85504425e-01 -8.05621862e-01 -2.10811183e-01
-1.20465048e-01 -6.13650739e-01 1.43319142e+00 -3.57583284e-01
-1.29826546e+00 9.17840540e-01 -7.18076944e-01 -2.33469442e-01
1.91112056e-01 1.08119316e-01 -4.29059744e-01 -7.41438419e-02
-4.12434727e-01 3.25155199e-01 7.00293064e-01 -1.65811181e+00
-5.14496207e-01 -2.94367701e-01 -6.49193645e-01 1.47075459e-01
6.92818105e-01 3.43652517e-01 4.32728678e-01 -5.93124218e-02
2.30152801e-01 -1.03726554e+00 -3.82315397e-01 3.32865328e-01
-6.95572793e-02 1.66094378e-01 1.07540011e+00 -5.46736419e-01
7.90288448e-01 -1.66798937e+00 -1.33187085e-01 3.43880862e-01
2.93112040e-01 4.55765724e-01 3.79624099e-01 5.13420701e-01
-2.21914407e-02 3.91342610e-01 -8.27212691e-01 -3.72742087e-01
-1.96996033e-01 5.43889344e-01 -7.37156153e-01 6.21996403e-01
4.35152292e-01 8.36274326e-01 -9.85079050e-01 -2.75660425e-01
9.29961443e-01 1.66476190e-01 2.46328171e-02 5.56589425e-01
-7.23613083e-01 7.47393608e-01 -2.47089043e-01 9.54984367e-01
1.35732365e+00 -1.69759076e-02 -1.27584606e-01 7.25888073e-01
-9.27539527e-01 3.70217979e-01 -8.55327785e-01 2.98297107e-01
-5.65465152e-01 1.13677478e+00 5.35716295e-01 -4.13846284e-01
9.94853139e-01 1.62137717e-01 8.73409137e-02 -2.19043136e-01
-3.51496696e-01 5.54670274e-01 -4.31068661e-03 -4.28133696e-01
4.74596262e-01 -4.53782260e-01 7.31555402e-01 3.91452074e-01
-2.74151862e-01 -1.44579601e+00 -4.36056077e-01 5.46684004e-02
1.22360606e-02 1.50718033e-01 4.16939974e-01 -6.95247531e-01
4.72531796e-01 1.51456937e-01 1.94298670e-01 7.04276979e-01
-3.73160951e-02 2.33123049e-01 9.91075709e-02 -5.94438016e-01
-1.03110230e+00 -1.01723838e+00 -8.49537849e-01 7.90929377e-01
-2.21623048e-01 -2.88113296e-01 -4.71211374e-02 9.99535695e-02
5.47776401e-01 8.86614859e-01 -4.11274880e-01 2.56140083e-01
6.64825260e-04 -1.45382762e+00 3.75911929e-02 2.29589298e-01
2.21812636e-01 -1.07239401e+00 -2.18693376e-01 -5.26555628e-02
7.34280795e-02 -1.02089310e+00 6.10939026e-01 3.90420467e-01
-1.26900244e+00 -8.51760745e-01 -6.06995150e-02 -3.41924168e-02
3.60046297e-01 1.88183054e-01 1.54037809e+00 3.98613185e-01
-2.69771010e-01 6.79296777e-02 -1.15654536e-01 -9.36896622e-01
-7.44297087e-01 -3.71473253e-01 1.36826307e-01 -5.43695867e-01
1.73596367e-01 -4.47316498e-01 -4.22023058e-01 -1.02746993e-01
-5.18993199e-01 -4.63525988e-02 3.11614946e-03 3.77805084e-01
4.64575768e-01 -1.30063713e-01 -8.97372663e-02 -8.05934429e-01
2.31086314e-01 -6.18768096e-01 -1.33030033e+00 2.47089639e-01
-1.30537856e+00 -1.09600946e-01 1.58502400e-01 3.37794602e-01
-1.00597477e+00 -1.91737294e-01 -1.25798762e-01 -1.38212934e-01
-4.08083588e-01 5.85453570e-01 3.53970200e-01 -7.21068978e-01
5.77148676e-01 4.91842151e-01 -1.46012872e-01 -5.70827127e-01
5.41369498e-01 8.47127259e-01 2.99487978e-01 -7.44206190e-01
1.15014851e+00 5.03293633e-01 3.20495933e-01 -1.50242150e+00
-6.19019210e-01 -6.14137232e-01 -7.16394246e-01 -4.07759398e-01
4.73831505e-01 -8.70165825e-01 -3.45554054e-01 7.43335247e-01
-1.03268087e+00 -3.40056211e-01 -2.31619880e-01 5.41174829e-01
-2.14472532e-01 2.06079215e-01 -3.28996509e-01 -1.85358977e+00
-4.38791692e-01 -9.61637139e-01 1.03234601e+00 5.28259158e-01
1.57975927e-01 -1.48852324e+00 5.24811089e-01 -1.50964363e-02
8.58292043e-01 5.16371608e-01 1.06186044e+00 1.61469832e-01
-8.23273838e-01 1.24527849e-01 -4.01433468e-01 2.73058236e-01
1.83632478e-01 1.23230922e+00 -1.73801267e+00 -9.99363214e-02
2.24494860e-01 -2.58584052e-01 1.18907428e+00 6.65608168e-01
1.12446547e+00 -3.48250985e-01 -3.10446948e-01 9.64020371e-01
1.52134943e+00 2.06295121e-02 2.22707927e-01 1.43722430e-01
3.84788632e-01 6.42667174e-01 3.59704554e-01 6.50503278e-01
1.02242991e-01 5.07197417e-02 7.05825329e-01 -4.62851375e-01
4.04070377e-01 1.27316892e-01 -1.40051201e-01 8.98426592e-01
-4.79111671e-01 8.32512900e-02 -1.19938600e+00 2.73860693e-01
-1.60661829e+00 -9.14983571e-01 -8.01517308e-01 2.25137568e+00
7.99743891e-01 -3.22571676e-03 -3.58678043e-01 -3.14124584e-01
1.97618455e-01 5.88729739e-01 -5.02804279e-01 -7.84122646e-01
-1.31735817e-01 4.77970093e-01 7.34550655e-01 9.15636659e-01
-1.09948981e+00 1.15692425e+00 8.83937931e+00 2.06752479e-01
-1.42617989e+00 -6.72059041e-03 6.21398926e-01 3.91327232e-01
-7.35460937e-01 3.43632966e-01 -8.01455200e-01 1.48886666e-01
1.30088723e+00 -3.66677269e-02 6.70017362e-01 8.46791565e-01
9.06851649e-01 -5.09608567e-01 -7.29247391e-01 4.61865395e-01
-7.63291717e-01 -1.75901449e+00 -3.03238809e-01 -6.75534382e-02
1.08349156e+00 7.34700084e-01 -1.30257025e-01 3.17009896e-01
3.48125905e-01 -1.08794606e+00 3.72258306e-01 7.79026747e-01
8.13454449e-01 -4.19103861e-01 7.82826722e-01 3.48271191e-01
-1.13294387e+00 4.37394708e-01 -6.49339080e-01 -5.38268268e-01
-2.60987580e-01 1.16291142e+00 -7.73831427e-01 6.13799393e-01
6.72797740e-01 5.40793419e-01 -5.43698370e-01 1.26699579e+00
-1.86897859e-01 1.00262117e+00 -5.19027829e-01 1.35540655e-02
2.61888921e-01 -6.70566440e-01 4.62811977e-01 1.29177237e+00
1.15817124e-02 2.55246341e-01 -2.82016248e-02 1.27887964e+00
4.18462127e-01 -6.30052835e-02 -7.37401903e-01 -2.60333419e-01
6.22880578e-01 1.31449509e+00 -3.13104481e-01 -4.08557236e-01
-1.58144802e-01 2.62504548e-01 -2.12564617e-01 4.65820074e-01
-5.63771129e-01 6.03049882e-02 1.32377970e+00 9.02598351e-02
-9.29544494e-02 -5.23839772e-01 -3.88310969e-01 -1.14090240e+00
-4.70984459e-01 -6.07103586e-01 -2.25778624e-01 -1.19645047e+00
-1.22791457e+00 2.31537238e-01 1.04901843e-01 -8.29011917e-01
-4.71621126e-01 -9.99459207e-01 -1.14221621e+00 1.37612581e+00
-2.25314569e+00 -8.42870772e-01 -4.01497871e-01 -3.48811895e-02
1.22812027e-02 8.74901339e-02 1.34655285e+00 -4.43596065e-01
-2.86236674e-01 -4.14109737e-01 1.29529834e+00 -3.46187055e-01
4.41786289e-01 -1.67853236e+00 7.09196866e-01 5.82178771e-01
-2.71548361e-01 4.68396306e-01 1.05364358e+00 -5.95578134e-01
-1.31591070e+00 -1.33331025e+00 7.99199462e-01 -5.13256729e-01
8.95243108e-01 -3.68825227e-01 -1.06905079e+00 4.53696311e-01
7.08235204e-02 4.55615163e-01 4.59545910e-01 3.33220810e-01
-1.62789747e-01 -1.05759859e-01 -1.12237370e+00 2.22204313e-01
1.72355562e-01 -7.40453422e-01 -4.52184826e-01 9.90046740e-01
5.06331682e-01 -6.03563786e-01 -8.57290506e-01 6.44803345e-01
5.58319867e-01 -7.91638553e-01 5.36659658e-01 -8.79209995e-01
1.78991929e-01 -4.39245850e-01 -6.49202764e-02 -1.34279871e+00
-3.22643429e-01 -5.12991071e-01 -4.41018045e-01 7.65446603e-01
5.16540051e-01 -8.08318734e-01 4.06371623e-01 6.15469456e-01
3.04203425e-02 -6.02581799e-01 -8.09956193e-01 -7.01139092e-01
6.82742476e-01 -4.54178542e-01 4.79787767e-01 1.19166648e+00
-3.12894762e-01 -1.78247362e-01 -2.24631280e-01 6.14720106e-01
9.00668561e-01 8.74521017e-01 4.23605829e-01 -1.76195383e+00
-5.26831336e-02 -4.21015233e-01 1.56468689e-01 -4.14036185e-01
-1.82484686e-02 -5.53741813e-01 3.19608539e-01 -1.31452918e+00
6.36554360e-02 -8.03077996e-01 -5.57632837e-03 8.05206746e-02
-1.73369735e-01 -8.35297480e-02 8.49129409e-02 5.12497783e-01
2.29235306e-01 6.90919220e-01 1.13224530e+00 -3.99897806e-02
-1.88106850e-01 3.45509410e-01 -8.22031870e-02 7.09216833e-01
9.01899040e-01 -5.73689818e-01 7.82860890e-02 -1.11193001e-01
2.56160945e-01 6.15883851e-03 2.74735212e-01 -7.41029561e-01
-3.62353861e-01 -1.03782713e+00 6.15954041e-01 -9.16109741e-01
1.76203862e-01 -5.75262785e-01 -2.96282377e-02 4.16058928e-01
9.83307809e-02 -2.23802552e-01 6.50462508e-01 1.20425351e-01
-1.13983557e-01 -3.71773124e-01 1.13860047e+00 -4.93729591e-01
-5.28510571e-01 4.81271654e-01 -1.05325031e+00 -1.39191985e-01
7.28597403e-01 3.86122525e-01 -5.84767878e-01 -5.44228911e-01
-6.92833841e-01 6.31690502e-01 5.55784762e-01 4.33163624e-03
4.25494254e-01 -8.77649367e-01 -9.27180886e-01 2.49892041e-01
2.22728729e-01 1.09687492e-01 -1.45457044e-01 4.14458156e-01
-1.10222936e+00 9.20550406e-01 2.26353943e-01 -8.06979835e-01
-1.19277751e+00 3.88881922e-01 1.02732694e+00 -1.45987840e-02
-1.40476286e-01 7.22736478e-01 6.59209415e-02 -9.61520910e-01
-3.24040115e-01 -7.36514032e-01 1.43565424e-02 -3.01179051e-01
3.55478317e-01 1.98688924e-01 1.23951234e-01 -3.65704685e-01
-4.88625228e-01 4.74614590e-01 5.04858732e-01 6.39033094e-02
1.14677083e+00 -8.49353299e-02 -7.05960751e-01 1.18268454e+00
4.60417092e-01 -1.67386562e-01 -1.21468985e+00 -1.65709600e-01
-1.16670467e-01 -5.07233620e-01 5.93787193e-01 -7.71568894e-01
-8.44020486e-01 1.28615499e+00 5.90365231e-01 2.98312217e-01
6.61161542e-01 -1.73799843e-01 9.82023999e-02 5.46717346e-01
-6.29054680e-02 -8.49880397e-01 -6.65760815e-01 6.84468031e-01
9.15500224e-01 -1.59562397e+00 5.61010838e-01 -3.52582514e-01
-2.39320368e-01 1.26685381e+00 5.54946005e-01 3.14804763e-01
1.19045770e+00 3.78182441e-01 2.13454038e-01 -1.30902335e-01
-9.38101292e-01 -2.28546903e-01 2.12146714e-01 5.99721968e-01
4.48132843e-01 5.74168682e-01 3.38358223e-01 -5.07990897e-01
-1.02309011e-01 -1.13078043e-01 6.27210796e-01 9.16795731e-01
-9.65321183e-01 -8.39564085e-01 -9.88251507e-01 5.67620695e-01
1.19805314e-01 -6.90899074e-01 -4.61429149e-01 5.18295705e-01
1.25280485e-01 1.05677927e+00 3.18240434e-01 2.14524835e-01
-4.04329658e-01 3.40784341e-01 2.98602730e-01 -7.40054965e-01
-1.66536078e-01 4.18119282e-02 1.15791492e-01 -1.13270409e-01
-4.96123403e-01 -8.85819495e-01 -9.46445048e-01 -7.97268331e-01
-7.27978706e-01 5.23540974e-01 1.02853096e+00 1.03188753e+00
-3.65084894e-02 1.11134045e-01 1.10542488e+00 -1.14389884e+00
-1.43474549e-01 -1.10856247e+00 -1.03019810e+00 -5.63920736e-01
9.42377806e-01 -6.00275457e-01 -8.93350244e-01 -3.35756779e-01] | [6.503835678100586, 3.0646414756774902] |
9f2826f9-b64c-4971-8dee-d891d24836d0 | zero-shot-domain-adaptation-of-anomalous | 2304.02221 | null | https://arxiv.org/abs/2304.02221v1 | https://arxiv.org/pdf/2304.02221v1.pdf | Zero-shot domain adaptation of anomalous samples for semi-supervised anomaly detection | Semi-supervised anomaly detection~(SSAD) is a task where normal data and a limited number of anomalous data are available for training. In practical situations, SSAD methods suffer adapting to domain shifts, since anomalous data are unlikely to be available for the target domain in the training phase. To solve this problem, we propose a domain adaptation method for SSAD where no anomalous data are available for the target domain. First, we introduce a domain-adversarial network to a variational auto-encoder-based SSAD model to obtain domain-invariant latent variables. Since the decoder cannot reconstruct the original data solely from domain-invariant latent variables, we conditioned the decoder on the domain label. To compensate for the missing anomalous data of the target domain, we introduce an importance sampling-based weighted loss function that approximates the ideal loss function. Experimental results indicate that the proposed method helps adapt SSAD models to the target domain when no anomalous data are available for the target domain. | ['Yohei Kawaguchi', 'Takashi Endo', 'Tomoya Nishida'] | 2023-04-05 | null | null | null | null | ['supervised-anomaly-detection', 'semi-supervised-anomaly-detection'] | ['computer-vision', 'computer-vision'] | [ 3.11828345e-01 1.20503977e-01 -1.46611109e-01 -5.55415213e-01
-7.76673615e-01 -3.10364246e-01 3.74269843e-01 -2.62057662e-01
-2.24165499e-01 1.05252981e+00 1.71060916e-02 -1.91010848e-01
3.48909587e-01 -6.44277036e-01 -8.02603722e-01 -6.97341383e-01
1.36576310e-01 5.76081157e-01 2.59780109e-01 -1.11847773e-01
1.09153017e-02 4.08972174e-01 -1.34646189e+00 1.21000476e-01
1.24783134e+00 9.59821820e-01 1.03442393e-01 3.15117002e-01
-3.79235208e-01 5.12326777e-01 -9.75944161e-01 -3.04680113e-02
5.35024345e-01 -6.22062087e-01 -4.48388815e-01 3.04412991e-01
1.04851298e-01 -5.97410023e-01 -5.59820116e-01 1.28048897e+00
3.33905309e-01 2.41160139e-01 9.56056178e-01 -1.66002882e+00
-6.00055218e-01 -4.31011915e-02 -4.70524311e-01 3.96887481e-01
2.49478310e-01 6.65865541e-02 6.79857790e-01 -8.17509949e-01
7.33388901e-01 1.03875268e+00 4.17338878e-01 1.04900634e+00
-1.24038231e+00 -8.09834182e-01 4.69722658e-01 9.65161435e-03
-1.09101331e+00 -4.93008494e-01 1.31954789e+00 -4.36876595e-01
7.28529394e-01 -1.79113626e-01 1.53585345e-01 1.68080771e+00
1.12598613e-01 7.90955842e-01 6.79379821e-01 -2.14658037e-01
6.01839840e-01 2.98440129e-01 -1.98494598e-01 2.40562543e-01
5.18986918e-02 1.53103665e-01 -5.38455129e-01 -4.48227078e-01
6.01921916e-01 1.97101608e-01 -6.15526848e-02 -7.24489510e-01
-9.32116866e-01 7.36434758e-01 -2.43981462e-02 2.68554557e-02
-5.36144614e-01 -3.82925898e-01 5.92799902e-01 8.75387788e-01
8.83702993e-01 3.00567120e-01 -5.55806816e-01 5.24485670e-03
-9.45456922e-01 3.08736324e-01 5.53367436e-01 1.21915758e+00
5.73892474e-01 6.03153110e-01 -6.38923571e-02 7.94910192e-01
3.48889709e-01 7.34384537e-01 5.75296998e-01 -8.45138133e-01
7.65562713e-01 5.49106717e-01 3.79048288e-01 -3.99513125e-01
2.30500385e-01 -2.12075651e-01 -7.47455239e-01 5.11242151e-01
5.73174298e-01 -2.32332721e-01 -9.68538105e-01 1.98904908e+00
3.62665534e-01 4.18155104e-01 4.39294815e-01 7.29442239e-01
1.49268538e-01 7.46960580e-01 -1.37973726e-01 -5.02970755e-01
5.74225545e-01 -5.22897184e-01 -8.22061419e-01 -4.12056863e-01
6.14297330e-01 -3.77014011e-01 1.23818052e+00 3.76018673e-01
-8.30622435e-01 -3.85546088e-01 -1.24053705e+00 2.78075188e-01
-1.37742072e-01 -2.71752924e-01 -8.09536576e-02 2.28414074e-01
-5.09370089e-01 5.41183889e-01 -9.19395268e-01 -3.49910557e-01
5.59857547e-01 8.92500058e-02 -3.70182961e-01 -1.54828578e-01
-1.44459450e+00 6.26726151e-01 5.33963323e-01 -3.90758246e-01
-1.25433362e+00 -5.60588598e-01 -9.16761220e-01 -2.19711065e-01
2.45708779e-01 -2.95634896e-01 1.19299603e+00 -1.22608054e+00
-1.33479786e+00 7.88937867e-01 -3.27625930e-01 -6.33000433e-01
7.45935380e-01 -2.85804123e-01 -8.40592623e-01 1.16118595e-01
2.12793425e-01 -2.32432876e-02 1.49459326e+00 -1.20738530e+00
-5.28069615e-01 -3.15792829e-01 -4.42122072e-01 2.62884438e-01
-3.93851876e-01 -2.57132053e-01 -6.55279830e-02 -8.99840593e-01
1.78086793e-03 -5.91116726e-01 -6.57429844e-02 2.71015882e-01
-4.14767802e-01 -2.96450462e-02 1.32818913e+00 -8.08908522e-01
1.19851494e+00 -2.64038992e+00 5.99153265e-02 3.97068322e-01
9.03587714e-02 3.73570532e-01 -7.18063442e-03 1.20201916e-01
-2.45695069e-01 -3.34727198e-01 -8.18868577e-01 -3.44427019e-01
-8.04715753e-02 6.33022070e-01 -9.14313376e-01 3.77943456e-01
4.51871753e-01 7.52850771e-02 -9.39206660e-01 -2.94577599e-01
1.28829062e-01 1.35785505e-01 -6.68618619e-01 6.22932076e-01
-6.20141745e-01 9.20520008e-01 -7.41197765e-01 6.26013815e-01
8.36872578e-01 1.09743059e-01 -2.29976356e-01 5.16860306e-01
4.19126540e-01 1.96154192e-01 -1.06612182e+00 1.77325594e+00
-2.04685315e-01 4.14943546e-01 -5.42058982e-02 -1.25516808e+00
1.19888699e+00 3.79956365e-01 6.23732328e-01 -6.91112757e-01
-2.34266877e-01 4.44119930e-01 -2.49845341e-01 -4.16232169e-01
2.32583836e-01 -3.94396812e-01 -2.38701805e-01 2.67303199e-01
1.94643348e-01 1.94661304e-01 -4.00828898e-01 1.17419049e-01
1.28726506e+00 2.85582125e-01 2.25552291e-01 1.35394722e-01
6.42240405e-01 7.33622164e-02 1.14873028e+00 5.44391632e-01
-4.92434859e-01 5.35704494e-01 6.99393034e-01 -3.49154621e-01
-1.45604050e+00 -1.48752570e+00 -1.20119527e-01 8.64203334e-01
1.43929422e-02 1.70893833e-01 -6.60497487e-01 -1.11368656e+00
2.83031669e-02 1.27657735e+00 -3.47338796e-01 -7.15224147e-01
-4.25821006e-01 -4.51096684e-01 5.60335040e-01 3.16816390e-01
5.25352895e-01 -1.03300440e+00 -2.03015760e-01 3.33426535e-01
-7.52291456e-02 -9.12222445e-01 -3.80265623e-01 2.42384925e-01
-1.08855820e+00 -7.55550981e-01 -7.34334469e-01 -6.15732491e-01
9.08041060e-01 -2.40831584e-01 1.00857532e+00 -4.53999609e-01
2.86451012e-01 2.38332152e-01 -3.58253270e-01 -4.52745110e-01
-9.58753288e-01 1.20132193e-02 5.63804090e-01 1.53865382e-01
7.21818507e-01 -6.63183570e-01 -2.73757815e-01 2.57751733e-01
-8.49683344e-01 -5.07725954e-01 2.89073795e-01 1.01829457e+00
6.79850757e-01 1.54999316e-01 9.32403803e-01 -1.14012218e+00
5.64286172e-01 -8.09888721e-01 -5.61894119e-01 -5.83060645e-02
-7.80350685e-01 2.72984266e-01 1.10779500e+00 -6.37117207e-01
-1.35815930e+00 -1.70603871e-01 -1.73749447e-01 -1.08050215e+00
-5.20862281e-01 4.87040728e-02 -6.16857469e-01 4.86758083e-01
8.36938620e-01 4.93173927e-01 1.90662086e-01 -7.08071828e-01
-8.79247412e-02 7.86509216e-01 6.30562007e-01 -5.62555134e-01
1.09031749e+00 3.78562093e-01 -2.49258339e-01 -6.72219157e-01
-7.68033087e-01 -4.19562459e-01 -6.70998573e-01 2.06036732e-01
5.86744308e-01 -1.04155600e+00 2.67357498e-01 5.79009950e-01
-1.02593184e+00 -3.02431643e-01 -8.93967152e-01 4.68623817e-01
-6.95886850e-01 5.40198028e-01 -9.51434001e-02 -8.19519103e-01
-2.00288147e-02 -8.48322332e-01 8.65900278e-01 -1.43874004e-01
-1.70979187e-01 -1.01579416e+00 1.50803477e-01 -8.21543410e-02
3.66936699e-02 3.93004268e-01 9.49145675e-01 -1.44671917e+00
-2.31517613e-01 -3.78831595e-01 2.16436729e-01 9.97387528e-01
2.38128394e-01 -4.24763322e-01 -9.76409912e-01 -4.79261577e-01
4.04423714e-01 -1.77920729e-01 6.31330371e-01 1.37129650e-01
1.32975852e+00 -2.11013958e-01 -2.74042457e-01 6.57658339e-01
1.00434995e+00 4.91663069e-01 4.67108846e-01 2.55228668e-01
5.91341674e-01 2.35040843e-01 9.38754499e-01 6.59080625e-01
-4.31939401e-02 4.33970183e-01 3.42098087e-01 1.42440125e-01
1.78095207e-01 -5.60680330e-01 7.10582733e-01 5.80088079e-01
5.25181413e-01 -3.50060612e-01 -8.51691306e-01 7.78292358e-01
-1.66948867e+00 -8.87046754e-01 2.66538441e-01 2.61303830e+00
7.20536768e-01 5.12742937e-01 4.01521437e-02 1.13468915e-01
7.62245059e-01 1.61722839e-01 -1.32522070e+00 -3.50055248e-01
7.30137229e-02 8.79997388e-03 3.28432083e-01 4.09072489e-01
-1.14911759e+00 8.63902271e-01 6.12541389e+00 6.59746349e-01
-9.19915140e-01 6.45489767e-02 5.22837900e-02 -5.99214248e-02
-4.91758853e-01 -9.17531326e-02 -5.88651180e-01 8.76510441e-01
9.61479306e-01 -2.00178444e-01 3.12641680e-01 1.14920926e+00
9.31548253e-02 2.63651729e-01 -1.26200187e+00 7.14933634e-01
-4.08230722e-02 -5.58246195e-01 1.02123991e-02 1.43554315e-01
5.53009808e-01 -1.03704400e-01 4.35989760e-02 5.85567474e-01
1.63712919e-01 -4.54121679e-01 2.40679756e-01 6.10982478e-01
1.11328256e+00 -9.35774744e-01 5.77522874e-01 6.89097762e-01
-6.00572407e-01 -3.30898277e-02 -6.23559415e-01 1.98792256e-02
2.36233808e-02 7.32859910e-01 -1.08499384e+00 2.38609806e-01
4.49253261e-01 8.31571281e-01 -1.34864166e-01 5.94451964e-01
-7.82902837e-02 8.52116346e-01 -2.29949445e-01 6.43530607e-01
9.04904772e-03 -3.25277567e-01 1.31770706e+00 5.76992929e-01
4.97447938e-01 -2.76534885e-01 2.01908678e-01 9.19230282e-01
-1.36649966e-01 -2.37642378e-01 -1.23117423e+00 -1.20840155e-01
6.42188191e-01 2.12498501e-01 4.67416905e-02 -1.79969773e-01
-5.43438256e-01 1.71026206e+00 8.07864666e-02 7.75044441e-01
-6.97852671e-01 -5.26432693e-01 1.05130017e+00 1.87637895e-01
1.57595783e-01 -2.83981301e-02 -2.07287788e-01 -1.48816609e+00
1.84949830e-01 -7.93785632e-01 5.46699703e-01 -6.09233558e-01
-1.72150218e+00 5.02389312e-01 -6.93160146e-02 -1.80545652e+00
-6.67749882e-01 -2.64931679e-01 -6.44751549e-01 1.05509949e+00
-1.39742041e+00 -6.98200464e-01 2.19760656e-01 1.03371966e+00
6.55545056e-01 -8.52281332e-01 9.22521949e-01 2.05478966e-01
-4.72638398e-01 9.42756057e-01 7.10128844e-01 1.97633818e-01
1.05906940e+00 -1.35111022e+00 5.85587740e-01 1.20864582e+00
-3.02893370e-01 3.96311015e-01 9.15256977e-01 -8.80306602e-01
-6.17222309e-01 -1.30490327e+00 7.75774062e-01 -3.96642476e-01
3.29661101e-01 -5.47470927e-01 -1.65221703e+00 9.59973633e-01
-3.05903703e-01 1.88192353e-01 6.94026530e-01 -1.63028702e-01
-3.69291812e-01 -1.98257685e-01 -1.71937692e+00 3.50195467e-01
9.61976051e-01 -7.79291332e-01 -1.10750210e+00 3.03132504e-01
9.45758343e-01 -4.09954131e-01 -6.81461275e-01 1.72411665e-01
-1.11750253e-01 -6.47822917e-01 9.01027679e-01 -8.93062353e-01
1.18884921e-01 -3.73482972e-01 -2.82654852e-01 -1.54271710e+00
1.33516610e-01 -3.23840380e-01 -5.75497925e-01 1.14008379e+00
2.27722749e-01 -8.56852055e-01 8.82994175e-01 6.24799907e-01
-3.09122026e-01 -2.82270670e-01 -1.32104599e+00 -9.85100925e-01
2.43775308e-01 -4.82413888e-01 6.45317018e-01 1.15695965e+00
-2.26089641e-01 9.30118039e-02 -6.79137528e-01 4.63820130e-01
8.69310677e-01 -4.43663336e-02 6.30921483e-01 -1.40660083e+00
-1.52391419e-01 2.39415959e-01 -3.53338182e-01 -1.07495403e+00
5.11152446e-01 -6.72784805e-01 -4.03244561e-03 -9.32129741e-01
-3.59228134e-01 -3.19805682e-01 -5.59144020e-01 3.29344869e-01
-1.08438544e-01 -3.38418424e-01 -2.30623588e-01 3.98927420e-01
-4.32650119e-01 9.66815829e-01 1.12390256e+00 -7.52197171e-04
-4.58924621e-01 4.08607811e-01 -3.94922256e-01 7.82208085e-01
8.03512514e-01 -7.98625231e-01 -8.79475594e-01 -5.27197160e-02
-3.30495894e-01 1.99640006e-01 1.11171134e-01 -1.09975886e+00
-5.57308607e-02 -3.99005681e-01 5.03753245e-01 -5.92127621e-01
2.61106312e-01 -1.22858429e+00 -2.33719796e-01 2.52369702e-01
-3.62752110e-01 -4.12194818e-01 4.12241258e-02 1.04918802e+00
-3.67178202e-01 -2.04723030e-01 9.53338325e-01 1.65094212e-01
-8.31452072e-01 5.20047009e-01 -4.71958578e-01 4.32057142e-01
1.16929901e+00 -4.07893300e-01 1.74331918e-01 -4.12681311e-01
-9.17835891e-01 4.03397828e-01 7.51878142e-01 4.58056718e-01
8.45265031e-01 -1.54706562e+00 -4.96115595e-01 8.63663495e-01
4.27225351e-01 4.33343202e-01 2.14730561e-01 1.50756598e-01
-1.22006156e-01 -8.93105790e-02 -3.56707931e-01 -5.47088385e-01
-7.41149485e-01 8.75461996e-01 3.59768569e-01 -8.85664523e-02
-8.59128833e-01 5.57017922e-01 2.95875072e-01 -6.43802762e-01
3.32242131e-01 -1.24156021e-01 1.24613643e-01 -1.57146275e-01
3.37831557e-01 2.14692369e-01 -1.28225729e-01 -4.93331283e-01
-2.50934690e-01 -1.07027747e-01 -3.12200278e-01 -1.55215234e-01
1.12729299e+00 -3.99357319e-01 4.66856241e-01 8.16535234e-01
1.11220706e+00 -2.74786860e-01 -1.83163500e+00 -8.42533886e-01
-3.16057503e-02 -6.18604898e-01 -2.07982153e-01 -6.28166020e-01
-6.81278169e-01 1.01212561e+00 7.55898952e-01 -4.73492928e-02
1.35036337e+00 -2.23454759e-01 9.53031242e-01 4.08444732e-01
2.20678240e-01 -1.50653028e+00 2.35408410e-01 6.04782999e-01
6.94999576e-01 -1.24429452e+00 -3.42703938e-01 8.27088207e-02
-1.10965085e+00 7.23914683e-01 8.78336191e-01 -2.75731593e-01
6.67432964e-01 -7.08931983e-02 -3.16162547e-03 7.71496966e-02
-7.06184030e-01 2.32570440e-01 3.81197296e-02 1.03348112e+00
-2.11167127e-01 -1.50460646e-01 2.77845949e-01 7.53969014e-01
2.14324877e-01 -1.36696801e-01 4.12092268e-01 8.93064201e-01
-4.13348496e-01 -1.09684873e+00 -4.70479250e-01 6.88919246e-01
-4.29737151e-01 1.35261089e-01 -2.14923531e-01 4.74606484e-01
9.28781033e-02 5.64702034e-01 2.92970836e-01 -2.63146281e-01
5.56135297e-01 8.48495901e-01 -9.17091146e-02 -7.33289838e-01
2.81015903e-01 -2.82142498e-02 -2.38134682e-01 -4.39479023e-01
7.81093910e-02 -8.37741613e-01 -1.33608651e+00 -2.08593339e-01
9.50845554e-02 1.97339371e-01 2.38436848e-01 9.91603076e-01
3.17608833e-01 6.29624903e-01 9.16810989e-01 -2.20052436e-01
-9.45209861e-01 -7.76151657e-01 -1.05159760e+00 8.13379705e-01
8.26164365e-01 -7.61644900e-01 -6.19484603e-01 1.02991097e-01] | [10.313961029052734, 3.1481270790100098] |
1f4bffdb-150d-4462-9531-03541143560e | provably-efficient-representation-learning | 2306.12356 | null | https://arxiv.org/abs/2306.12356v1 | https://arxiv.org/pdf/2306.12356v1.pdf | Provably Efficient Representation Learning with Tractable Planning in Low-Rank POMDP | In this paper, we study representation learning in partially observable Markov Decision Processes (POMDPs), where the agent learns a decoder function that maps a series of high-dimensional raw observations to a compact representation and uses it for more efficient exploration and planning. We focus our attention on the sub-classes of \textit{$\gamma$-observable} and \textit{decodable POMDPs}, for which it has been shown that statistically tractable learning is possible, but there has not been any computationally efficient algorithm. We first present an algorithm for decodable POMDPs that combines maximum likelihood estimation (MLE) and optimism in the face of uncertainty (OFU) to perform representation learning and achieve efficient sample complexity, while only calling supervised learning computational oracles. We then show how to adapt this algorithm to also work in the broader class of $\gamma$-observable POMDPs. | ['Xuezhou Zhang', 'Zhuoran Yang', 'Mengdi Wang', 'Huazheng Wang', 'Zihao Li', 'Jiacheng Guo'] | 2023-06-21 | null | null | null | null | ['efficient-exploration'] | ['methodology'] | [ 4.72659975e-01 8.76561284e-01 -3.10944766e-01 -1.75707236e-01
-1.50270498e+00 -6.12063169e-01 6.18230700e-01 1.62708834e-01
-7.19388902e-01 1.17612672e+00 3.62418801e-01 -6.71552718e-01
-5.09251535e-01 -6.92629576e-01 -6.35153413e-01 -9.50331509e-01
-6.58445776e-01 1.22784340e+00 1.06474943e-02 2.65641063e-01
1.85382724e-01 2.88691700e-01 -9.38151240e-01 -1.33995265e-01
2.24314868e-01 7.77956069e-01 2.01705441e-01 8.82751644e-01
2.52547562e-01 9.61291671e-01 -3.62019479e-01 6.25918955e-02
5.54670632e-01 -5.76870382e-01 -1.10552216e+00 1.80069044e-01
-5.38030446e-01 -6.49720669e-01 -5.74007094e-01 1.11926448e+00
3.89779061e-01 3.76230031e-01 8.69284093e-01 -1.13743579e+00
-1.38055719e-02 7.89614737e-01 -1.66846827e-01 7.22744539e-02
6.11968815e-01 2.73936272e-01 1.08981049e+00 7.47132376e-02
7.03056991e-01 1.67063904e+00 -4.42680083e-02 5.53623796e-01
-1.25071144e+00 -5.39695144e-01 -3.00879944e-02 -4.76006381e-02
-1.03658891e+00 -3.00096929e-01 2.01527983e-01 -7.82687739e-02
1.05193830e+00 1.38762340e-01 6.71743810e-01 1.12080300e+00
5.71394145e-01 1.18853402e+00 1.45697665e+00 -5.13083518e-01
1.08245683e+00 -6.07919991e-01 -1.52670413e-01 7.73180306e-01
3.71532083e-01 6.55713499e-01 -3.60661089e-01 -4.97288197e-01
8.62647355e-01 1.44044189e-02 6.03354629e-03 -6.33627236e-01
-1.15898395e+00 1.11352897e+00 -2.04085395e-01 -3.26519608e-01
-4.41941530e-01 6.01528466e-01 -2.53735594e-02 7.46761739e-01
-2.98244059e-01 5.83592653e-01 -3.30494732e-01 -7.25492299e-01
-4.22062576e-01 5.12253821e-01 1.19888878e+00 9.93217289e-01
7.83680558e-01 2.36953795e-02 -1.07916286e-02 -2.37900510e-01
5.04495203e-01 9.17063594e-01 5.41318543e-02 -1.60400033e+00
8.88189137e-01 3.53451781e-02 7.14811862e-01 -3.10386233e-02
-4.39297467e-01 3.49677801e-01 -4.15002137e-01 5.97460985e-01
3.92069548e-01 -5.42932689e-01 -9.88194823e-01 1.68103087e+00
1.20362304e-01 -2.06594676e-01 6.18571341e-01 6.48298085e-01
-1.56130999e-01 6.76469266e-01 -1.89021692e-01 -5.84257007e-01
8.58917534e-01 -4.19077933e-01 -5.87429166e-01 -3.34634870e-01
4.21402872e-01 2.20428780e-02 5.85126162e-01 5.85575044e-01
-1.18306816e+00 1.49293974e-01 -1.32105339e+00 1.07357278e-01
3.82564247e-01 -6.19417489e-01 1.02961648e+00 8.02600861e-01
-9.33061659e-01 7.80737936e-01 -1.79941916e+00 -1.49118870e-01
4.11403954e-01 8.63033235e-01 -1.14975952e-01 -1.58198074e-01
-8.35935235e-01 9.87717509e-01 9.31636214e-01 -4.97029245e-01
-2.07308054e+00 6.11767471e-01 -8.62924039e-01 4.83612157e-02
1.01450431e+00 -4.97788250e-01 1.72655022e+00 -3.07857215e-01
-1.93120885e+00 1.22894183e-01 5.10424040e-02 -9.00718629e-01
3.62123042e-01 -1.79799721e-02 -5.56502789e-02 3.93532246e-01
-9.96621698e-02 4.63457882e-01 8.45641375e-01 -9.15150762e-01
-6.19569838e-01 -4.28185314e-01 4.92050737e-01 6.48561656e-01
4.28628474e-01 -1.35776088e-01 1.44509211e-01 4.81595695e-02
6.49791479e-01 -1.24685931e+00 -7.20419705e-01 -3.02234679e-01
-2.88593084e-01 -5.94913289e-02 2.91723788e-01 -6.23806790e-02
7.60767758e-01 -1.64268816e+00 3.97610694e-01 3.38822454e-01
6.03930503e-02 -2.27589309e-01 -8.59121084e-02 8.45834315e-01
4.62393939e-01 -8.91647413e-02 -5.28141737e-01 -4.83482778e-01
3.32231969e-01 9.53721583e-01 -5.29695094e-01 7.68076897e-01
-1.50640562e-01 8.36215734e-01 -1.10860074e+00 -4.77816224e-01
1.49176389e-01 -3.24827015e-01 -5.29875398e-01 -3.74759128e-03
-7.17557132e-01 8.83915186e-01 -1.03016746e+00 7.29279637e-01
2.39661559e-01 -8.24754983e-02 5.88840663e-01 1.10554004e+00
1.54671846e-02 6.20895207e-01 -1.52832854e+00 1.86109030e+00
-8.55677649e-02 1.90130070e-01 2.49017224e-01 -7.64858186e-01
5.15164912e-01 5.93460262e-01 4.86070842e-01 -3.96326035e-01
3.74221474e-01 2.74897397e-01 -3.88320424e-02 -1.63117468e-01
4.88605857e-01 -6.90142512e-01 -7.81384587e-01 1.14408791e+00
2.33837247e-01 -5.09534419e-01 -1.94970474e-01 -5.88791259e-02
1.65593600e+00 1.96308538e-01 8.30104113e-01 -3.69722024e-02
-1.10062040e-01 3.93648595e-02 4.80995327e-01 1.55386055e+00
-2.54852951e-01 3.57222587e-01 6.84578598e-01 -2.16705397e-01
-8.71681690e-01 -1.38973057e+00 4.76294979e-02 8.16099107e-01
2.74657875e-01 -4.14918602e-01 -4.89978492e-01 -4.21191961e-01
-2.82752097e-01 1.00963402e+00 -4.36997592e-01 9.70155466e-03
-3.74409467e-01 -6.90122962e-01 5.40772796e-01 2.97414005e-01
4.75878984e-01 -1.22001207e+00 -1.13417590e+00 4.53372270e-01
1.56722561e-01 -4.88219738e-01 2.53712624e-01 8.56294930e-01
-1.15316105e+00 -8.17198813e-01 -2.55972803e-01 -8.00392032e-02
4.76375431e-01 -4.57026996e-02 5.79284370e-01 -6.93793416e-01
4.03720379e-01 7.47625470e-01 -3.29248786e-01 -6.89815760e-01
-5.93533039e-01 1.14309741e-02 3.13632220e-01 -5.67866325e-01
1.93689972e-01 -5.96211374e-01 -3.44466902e-02 -1.69015941e-04
-9.68362868e-01 -8.60956609e-02 6.81229711e-01 7.54361153e-01
7.25411773e-01 3.17099631e-01 -2.47272417e-01 -5.92944443e-01
6.53018236e-01 -4.48447257e-01 -1.00653625e+00 9.59543288e-02
-2.96186924e-01 8.36741745e-01 2.88354099e-01 -1.07356682e-01
-1.02214944e+00 3.43356192e-01 3.67411859e-02 -8.75407904e-02
-1.01499498e-01 4.14161533e-01 -3.12642962e-01 7.39753842e-02
5.88852227e-01 3.51336122e-01 3.27141955e-02 -4.28944200e-01
2.96821177e-01 6.43465877e-01 3.74546170e-01 -9.70034957e-01
5.66373527e-01 8.19722056e-01 4.25596446e-01 -5.00001311e-01
-7.02534676e-01 -2.02146977e-01 -5.23989439e-01 3.77034783e-01
7.18722880e-01 -8.56113672e-01 -1.13985813e+00 -9.38049778e-02
-7.33223617e-01 -6.78881884e-01 -8.14357698e-01 9.69490469e-01
-1.56845057e+00 4.77372229e-01 -4.34740365e-01 -1.52353716e+00
2.14911118e-01 -1.24004757e+00 1.04656756e+00 4.16254066e-02
-9.66219604e-02 -6.92378581e-01 4.78279710e-01 6.07933523e-03
-2.30162680e-01 1.63465321e-01 6.32775784e-01 -8.15397561e-01
-1.09601271e+00 -7.47355074e-02 3.34085941e-01 -1.81640074e-01
-2.78440326e-01 -8.76254618e-01 -7.72667885e-01 -7.59750843e-01
4.67786431e-01 -6.59631431e-01 8.77742529e-01 2.78941005e-01
7.62383640e-01 -7.13132501e-01 -2.84899056e-01 3.81823599e-01
1.31768680e+00 6.72177076e-01 5.78705549e-01 5.43841422e-01
-5.05526736e-02 1.73844680e-01 6.92669868e-01 7.86379695e-01
3.37794363e-01 2.59638757e-01 8.23360205e-01 8.82683098e-01
8.35177422e-01 -5.04094660e-01 7.06766188e-01 2.73003936e-01
-4.07192320e-01 -4.28582788e-01 -7.83281863e-01 3.37106556e-01
-2.13420486e+00 -9.88624394e-01 5.01449525e-01 2.40438199e+00
6.37851655e-01 3.39960009e-01 7.76536241e-02 2.73693781e-02
2.66786218e-01 2.11922884e-01 -8.07005703e-01 -4.63691741e-01
1.82194531e-01 6.35846198e-01 1.15233696e+00 8.87517869e-01
-1.15334308e+00 8.15135241e-01 6.66558599e+00 5.95748127e-01
-2.43679851e-01 4.22431141e-01 5.59422038e-02 -3.12707722e-01
-3.85325819e-01 6.93523169e-01 -7.05456197e-01 1.57809719e-01
1.19124687e+00 -3.28476466e-02 9.01590526e-01 8.86245906e-01
-1.51408076e-01 -6.92047179e-01 -1.22057688e+00 9.64656413e-01
-5.53086460e-01 -1.04565549e+00 -3.00512969e-01 5.88488460e-01
8.47336948e-01 2.55299717e-01 -1.04419820e-01 4.96434867e-01
1.52443230e+00 -1.18579650e+00 7.78847694e-01 2.27056712e-01
6.00480556e-01 -8.09246719e-01 5.69686294e-01 1.02784932e+00
-8.43409121e-01 -6.84678674e-01 -7.02363551e-01 -6.16994858e-01
5.41639924e-01 -1.83880731e-01 -1.06627166e+00 5.60395002e-01
8.89603943e-02 1.64369062e-01 2.04872578e-01 8.67003560e-01
-5.92272639e-01 4.41310883e-01 -7.50173986e-01 -2.31862471e-01
7.12597787e-01 -2.13328436e-01 7.92164087e-01 6.17617249e-01
6.52537465e-01 6.46082759e-01 5.61398745e-01 4.95492339e-01
3.19337040e-01 -6.48355126e-01 -5.47731817e-01 -4.69299972e-01
4.13868606e-01 4.30788964e-01 -7.50170588e-01 -2.82277852e-01
7.63524324e-02 9.33223665e-01 2.30540007e-01 1.34739831e-01
-3.66791815e-01 1.94368765e-01 6.24274909e-01 -3.44484448e-01
4.81406957e-01 -5.79808474e-01 -2.12799180e-02 -1.27520871e+00
-3.81367862e-01 -8.19122791e-01 7.83898294e-01 -3.37894231e-01
-7.58004844e-01 -1.85210127e-02 3.94966692e-01 -1.06780243e+00
-1.03574860e+00 -4.18973953e-01 -1.01313278e-01 4.60625589e-01
-1.02604711e+00 -6.31288707e-01 7.56197095e-01 6.32577300e-01
3.58494818e-01 -2.76620001e-01 1.31390834e+00 -8.56653512e-01
4.83184978e-02 -1.62757665e-01 6.09129250e-01 -2.78672189e-01
-2.62714982e-01 -1.52376020e+00 3.88868690e-01 6.45859838e-01
3.01491708e-01 4.09177631e-01 9.00979817e-01 -8.02604496e-01
-1.96297336e+00 -6.50935531e-01 3.23222935e-01 -6.32264197e-01
6.58745766e-01 -2.69462436e-01 -1.60926521e-01 1.29563987e+00
-1.09099977e-01 -3.84127527e-01 2.89877862e-01 -8.03776644e-03
8.46370310e-02 5.59218109e-01 -1.27560496e+00 5.71008027e-01
1.10716236e+00 -5.47728419e-01 -1.14419198e+00 1.92740396e-01
5.56535542e-01 -7.53940880e-01 -5.60955584e-01 -3.90899666e-02
4.52375859e-01 -7.92198479e-01 8.21796834e-01 -7.89861798e-01
-2.00265959e-01 -2.27820963e-01 -4.86629456e-01 -1.01324832e+00
-1.11275308e-01 -1.38215995e+00 -6.12760603e-01 2.42461145e-01
1.07379377e-01 -6.01130188e-01 8.50088418e-01 7.10282087e-01
9.53381211e-02 -3.26446682e-01 -1.72545445e+00 -9.10016179e-01
3.68685788e-03 -8.18827868e-01 5.47113001e-01 2.02496007e-01
3.97382379e-01 -7.38359056e-03 -5.99097967e-01 2.77794868e-01
8.82604241e-01 2.47528851e-01 7.10760832e-01 -9.34230626e-01
-9.24565256e-01 1.38078168e-01 -3.62357855e-01 -1.47732544e+00
-1.02280639e-01 -7.40874588e-01 3.16108406e-01 -1.44635141e+00
3.14529419e-01 -4.63233948e-01 -2.72034585e-01 5.97531974e-01
4.09944206e-01 -5.43607175e-01 5.56086123e-01 3.88796657e-01
-1.30075204e+00 6.39792442e-01 1.37208366e+00 -6.11608326e-02
-2.68943846e-01 5.61492205e-01 -6.86292350e-01 7.55453467e-01
6.31074071e-01 -9.23032165e-01 -6.10874772e-01 -2.52043039e-01
3.81643742e-01 1.08810079e+00 1.35337934e-01 -1.14664447e+00
2.46257968e-02 -7.22324848e-01 7.41846934e-02 -5.35844445e-01
7.48912513e-01 -6.77068293e-01 2.10937262e-01 7.51625538e-01
-5.89851856e-01 -1.21003062e-01 -2.87884504e-01 1.09165049e+00
3.00534546e-01 -8.08576703e-01 4.25131440e-01 -6.68958724e-01
-5.27601779e-01 3.51263076e-01 -1.00876045e+00 1.07534967e-01
1.00458062e+00 -2.97753159e-02 1.29598662e-01 -1.03239775e+00
-1.09017897e+00 2.48081103e-01 7.08954751e-01 -3.64564121e-01
5.30011773e-01 -8.47335756e-01 -2.69478500e-01 6.04918674e-02
-1.79132700e-01 3.21848035e-01 -9.81458500e-02 3.93876016e-01
-2.69880146e-01 7.22614765e-01 -2.53997058e-01 -2.08451048e-01
-6.48020089e-01 5.49792290e-01 1.24290329e-03 -6.90837264e-01
-7.20774949e-01 5.85709751e-01 -1.91850767e-01 -4.45554197e-01
2.58241236e-01 -5.07571399e-01 2.36469761e-01 -3.82954955e-01
5.92150033e-01 4.99700904e-01 -5.67728221e-01 -2.70024925e-01
1.09514497e-01 -1.16553746e-01 1.03732973e-01 -1.11251962e+00
1.36456823e+00 -4.23765391e-01 2.33777717e-01 5.30843616e-01
7.71091521e-01 -3.38497490e-01 -2.04496932e+00 -2.47379705e-01
3.14184844e-01 -4.05431062e-01 -2.24268645e-01 -3.64513516e-01
-1.51652738e-01 8.71945798e-01 4.06147897e-01 1.52869955e-01
5.42266786e-01 2.53991485e-01 5.87819338e-01 1.36488473e+00
1.44146466e+00 -1.10711932e+00 -7.58737028e-02 9.49318409e-01
2.15508446e-01 -9.29624677e-01 1.53210118e-01 3.20070595e-01
-8.28298748e-01 9.65486228e-01 -2.51719773e-01 -1.44350588e-01
2.74416208e-01 3.34593982e-01 -5.18941045e-01 -2.57780999e-01
-9.23879921e-01 -4.63510811e-01 -5.55527031e-01 8.29917192e-01
-6.99802995e-01 4.71439064e-01 1.23617813e-01 4.94213432e-01
-4.20165598e-01 1.69314444e-01 7.71679521e-01 1.75448799e+00
-9.70259130e-01 -1.19500613e+00 -6.29675865e-01 5.10538042e-01
-3.74684393e-01 2.01827720e-01 -9.14533064e-02 6.19910836e-01
-2.22547144e-01 9.36574221e-01 -7.01749697e-02 -1.07813507e-01
-2.90414065e-01 2.23180652e-01 9.78816330e-01 -8.00921679e-01
1.16399258e-01 1.09316833e-01 2.92875320e-01 -7.39322960e-01
-3.76212656e-01 -1.10982323e+00 -1.58949590e+00 -9.30217579e-02
5.29482253e-02 2.40914583e-01 4.95444119e-01 1.38796270e+00
-2.05160424e-01 -1.36974394e-01 3.94083112e-01 -9.35896873e-01
-1.49943602e+00 -7.71561146e-01 -9.51061666e-01 -3.63078296e-01
4.87478942e-01 -7.86825478e-01 -2.65490174e-01 -4.41501826e-01] | [4.295647144317627, 2.1631722450256348] |
b286b307-d559-4db8-bfd5-cff5e88cc900 | a-permutable-hybrid-network-for-volumetric | 2303.13111 | null | https://arxiv.org/abs/2303.13111v2 | https://arxiv.org/pdf/2303.13111v2.pdf | A Permutable Hybrid Network for Volumetric Medical Image Segmentation | The advent of Vision Transformer (ViT) has brought substantial advancements in 3D volumetric benchmarks, particularly in 3D medical image segmentation. Concurrently, Multi-Layer Perceptron (MLP) networks have regained popularity among researchers due to their comparable results to ViT, albeit with the exclusion of the heavy self-attention module. This paper introduces a permutable hybrid network for volumetric medical image segmentation, named PHNet, which exploits the advantages of convolution neural network (CNN) and MLP. PHNet addresses the intrinsic isotropy problem of 3D volumetric data by utilizing both 2D and 3D CNN to extract local information. Besides, we propose an efficient Multi-Layer Permute Perceptron module, named MLPP, which enhances the original MLP by obtaining long-range dependence while retaining positional information. Extensive experimental results validate that PHNet outperforms the state-of-the-art methods on two public datasets, namely, COVID-19-20 and Synapse. Moreover, the ablation study demonstrates the effectiveness of PHNet in harnessing the strengths of both CNN and MLP. The code will be accessible to the public upon acceptance. | ['Hao Chen', 'Kwang-Ting Cheng', 'Dong Zhang', 'Xiao Fang', 'Yi Lin'] | 2023-03-23 | null | null | null | null | ['volumetric-medical-image-segmentation'] | ['medical'] | [ 2.09447816e-01 3.03364992e-01 -3.25560153e-01 -2.48284966e-01
-5.49003899e-01 -6.71112016e-02 2.97352105e-01 -2.74273027e-02
-6.49518311e-01 6.02440834e-01 2.28667915e-01 -2.91255534e-01
1.70646217e-02 -7.09339559e-01 -8.15181077e-01 -7.78604150e-01
1.14571229e-02 3.75710100e-01 3.65063906e-01 2.55672168e-02
4.99419821e-03 7.23888755e-01 -1.14041650e+00 2.23867923e-01
7.24363625e-01 1.20109010e+00 2.55020410e-01 2.04306558e-01
-2.44063169e-01 8.76021862e-01 -1.56102717e-01 -2.16042861e-01
2.16536403e-01 -2.90476065e-02 -9.26579356e-01 -1.89842582e-01
3.97888780e-01 -3.39069545e-01 -5.77168405e-01 8.13196898e-01
9.26557839e-01 -2.26881519e-01 4.80134577e-01 -9.27609742e-01
-8.01136374e-01 7.07881987e-01 -8.63547921e-01 5.21767139e-01
-2.04380617e-01 3.38799506e-01 9.10173237e-01 -9.13117886e-01
5.38762689e-01 1.04971623e+00 1.07005489e+00 4.43367600e-01
-1.12309682e+00 -3.81715834e-01 -2.87452340e-02 7.42789432e-02
-1.34523046e+00 -3.19846272e-02 1.06076074e+00 -2.41188779e-01
1.27945316e+00 -3.94899808e-02 8.87268901e-01 1.21173906e+00
3.64916831e-01 1.29322672e+00 1.14892757e+00 -7.88666382e-02
-7.98479095e-02 -1.40735328e-01 1.75943658e-01 9.77569282e-01
1.36643872e-01 1.95039570e-01 -2.95689732e-01 1.47461191e-01
1.02469385e+00 1.00419343e-01 -2.31276795e-01 -5.24537265e-01
-1.20660293e+00 7.62497246e-01 9.40952539e-01 2.32466266e-01
-5.91984451e-01 3.57191205e-01 5.92894495e-01 -1.14397548e-01
4.85935718e-01 3.23562205e-01 -3.66585761e-01 -4.29729819e-02
-8.78991902e-01 -4.01347391e-02 4.81984049e-01 6.40989602e-01
4.83199209e-01 -8.54476839e-02 -5.91455579e-01 8.53482008e-01
2.05833450e-01 4.84839708e-01 5.01391411e-01 -6.34751022e-01
4.03347313e-01 7.77602613e-01 -5.36233902e-01 -7.63617337e-01
-8.41723084e-01 -6.68093383e-01 -1.25065804e+00 3.63159887e-02
2.01493353e-01 -7.52260983e-02 -1.29363871e+00 1.39413846e+00
2.66121715e-01 3.16262513e-01 -6.08426146e-02 9.08844829e-01
1.47510242e+00 2.83288777e-01 -1.91779260e-03 2.37745672e-01
1.20419991e+00 -1.19323230e+00 -3.07153463e-01 -2.24030152e-01
2.62273610e-01 -4.22946185e-01 7.88968265e-01 -1.84724666e-02
-1.25617540e+00 -5.82245529e-01 -9.44040418e-01 -2.51981646e-01
-2.16877326e-01 -1.59941629e-01 9.69754934e-01 6.38357997e-01
-1.09237063e+00 6.70832872e-01 -1.03933001e+00 -9.71797332e-02
1.18426454e+00 6.97659016e-01 -7.40743652e-02 2.28136196e-03
-1.01601052e+00 6.60282552e-01 2.70853013e-01 3.47072244e-01
-7.61620879e-01 -1.12847507e+00 -9.52122629e-01 -9.44580436e-02
1.24140970e-01 -9.68493998e-01 9.53481674e-01 -6.94668472e-01
-1.51274347e+00 9.40027535e-01 1.01967521e-01 -5.40243745e-01
7.28457928e-01 7.44404197e-02 9.21541154e-02 4.56141651e-01
1.72297746e-01 1.33500397e+00 6.14315987e-01 -1.00900662e+00
-4.65803772e-01 -4.57196742e-01 -5.04600815e-02 1.60977930e-01
-9.26830024e-02 -3.54000807e-01 -9.18628097e-01 -5.24299741e-01
1.35794610e-01 -8.15573633e-01 -4.21016127e-01 -4.67891358e-02
-8.18372488e-01 -1.31197989e-01 7.12994337e-01 -4.04277623e-01
5.97525537e-01 -1.95345140e+00 1.18457779e-01 2.14886218e-01
6.82431877e-01 2.31336102e-01 -6.50420785e-02 -1.93829745e-01
-2.50357110e-02 -1.00495666e-01 -5.04018247e-01 -6.31012797e-01
-2.41634831e-01 3.74660492e-01 1.50930122e-01 6.53100610e-01
3.82169008e-01 1.61205006e+00 -8.22345316e-01 -8.22985232e-01
3.91582996e-01 8.05364966e-01 -5.24529815e-01 -2.38935053e-01
2.43739411e-02 7.38337338e-01 -5.38626015e-01 9.08647478e-01
8.51818800e-01 -7.10738122e-01 -1.18322082e-01 -3.53160948e-01
-7.22229388e-03 2.08672974e-02 -4.40349400e-01 1.80832827e+00
-2.68069506e-01 4.35676098e-01 -1.41359225e-01 -1.13666964e+00
7.53900647e-01 1.91861495e-01 1.06543171e+00 -1.15054131e+00
3.95347059e-01 9.47563052e-02 -1.86528906e-01 -5.14796317e-01
3.83301824e-01 -1.27996176e-01 7.07347766e-02 1.66405067e-01
1.20060340e-01 1.88374165e-02 4.11894843e-02 -6.64778352e-02
1.06860995e+00 3.45182680e-02 -2.25566998e-01 -2.35722259e-01
4.21743423e-01 -1.83701172e-01 5.85046828e-01 9.45568800e-01
-5.26649475e-01 7.72489786e-01 6.03286922e-01 -4.21126872e-01
-9.12086129e-01 -1.23690498e+00 -4.50696081e-01 5.08817673e-01
4.30206299e-01 1.75679877e-01 -4.84270513e-01 -7.67074645e-01
2.87707567e-01 1.28882453e-01 -9.76948023e-01 5.64019254e-04
-7.07031786e-01 -9.97610986e-01 9.46689129e-01 9.11754966e-01
9.92404997e-01 -1.23629594e+00 -7.04809785e-01 2.14196429e-01
-1.84998408e-01 -1.31923890e+00 -2.53570199e-01 3.41517627e-01
-1.02434182e+00 -9.00206685e-01 -1.06496358e+00 -9.38216865e-01
6.20505154e-01 1.82857856e-01 9.99589622e-01 -5.28941117e-02
-4.54330146e-01 3.98818493e-01 -3.27323787e-02 -3.18801731e-01
-5.07243909e-02 5.47955632e-01 -4.27267015e-01 -2.23511130e-01
1.59266204e-01 -7.91721642e-01 -8.79701436e-01 1.61287487e-02
-7.78364062e-01 3.65286946e-01 9.80408728e-01 8.52462411e-01
9.81574297e-01 -2.69293249e-01 5.17963886e-01 -9.76864576e-01
4.18518126e-01 -3.98952842e-01 -2.03233317e-01 6.24488629e-02
-4.46703136e-01 6.34793118e-02 4.62355494e-01 -8.86263251e-02
-8.38740826e-01 1.18004523e-01 -5.40375769e-01 -6.02652073e-01
2.58322898e-02 5.40795326e-01 1.18911132e-01 -4.14272457e-01
2.90623724e-01 3.51694316e-01 2.80778289e-01 -5.17792583e-01
3.15385729e-01 2.46478423e-01 6.73024714e-01 -3.57452393e-01
4.81740743e-01 9.21587765e-01 1.68724895e-01 -7.07194388e-01
-7.82876432e-01 -4.68293160e-01 -6.18303776e-01 -9.04515013e-02
1.10705769e+00 -7.82946587e-01 -8.67074966e-01 7.98363924e-01
-1.01510906e+00 -3.29549134e-01 -5.62334478e-01 3.68674219e-01
-4.73215342e-01 4.03054237e-01 -1.00323057e+00 -1.78876385e-01
-6.93726242e-01 -1.63302505e+00 1.09949005e+00 4.32959825e-01
1.74479559e-01 -1.20963550e+00 -7.64516592e-02 4.19743955e-01
5.09591222e-01 4.74996597e-01 1.03964317e+00 -4.51595008e-01
-5.99453866e-01 7.77427554e-02 -7.58825719e-01 3.93921643e-01
4.34197336e-02 -3.34403992e-01 -1.15005422e+00 -1.84214607e-01
-3.65691602e-01 -3.53433311e-01 1.28478873e+00 9.61666822e-01
1.40958142e+00 1.36512458e-01 -4.90976721e-01 9.08907592e-01
1.44764042e+00 -2.00669453e-01 5.79395652e-01 2.06690460e-01
1.12654006e+00 8.18574727e-02 -1.79725036e-01 2.27990881e-01
6.15369618e-01 4.00617748e-01 7.13324308e-01 -8.39613438e-01
-3.71356010e-01 -2.40732934e-02 -2.66562223e-01 1.00865364e+00
-2.21700683e-01 1.41108781e-01 -9.70371604e-01 5.97368062e-01
-1.68436861e+00 -6.91950858e-01 4.19638455e-02 1.61886084e+00
7.52988696e-01 3.11912060e-01 1.50188785e-02 -1.62712578e-02
3.55683655e-01 3.46892774e-01 -8.76730502e-01 -2.95982659e-01
-3.38632554e-01 5.55063367e-01 1.01704037e+00 1.35183364e-01
-1.38754487e+00 8.07467461e-01 5.89025307e+00 8.93397093e-01
-1.40424609e+00 1.46805093e-01 9.79583144e-01 -1.94488019e-02
-2.61781126e-01 -5.76740205e-01 -7.16357827e-01 3.49924207e-01
4.43624914e-01 3.73573065e-01 6.06983006e-02 6.82293236e-01
1.24333978e-01 1.56982671e-02 -8.24506164e-01 1.06846774e+00
6.44563581e-04 -1.84546566e+00 -4.28340621e-02 7.84678087e-02
7.89399028e-01 8.50363731e-01 3.70843261e-01 3.26203078e-01
8.06514472e-02 -1.22212899e+00 5.75926781e-01 5.13652980e-01
8.05643141e-01 -8.00297856e-01 9.04115140e-01 -4.58403770e-03
-1.11332417e+00 1.35551468e-01 -2.98347116e-01 3.33478481e-01
1.07885972e-01 7.01012671e-01 -7.34615803e-01 6.71152771e-01
9.33650255e-01 1.00034189e+00 -6.27277553e-01 1.32937074e+00
-9.05889645e-02 4.90506679e-01 -4.06458229e-01 5.30615374e-02
8.04469824e-01 1.05680339e-02 4.65777278e-01 1.18509531e+00
-2.28473023e-01 -3.08850944e-01 1.46824151e-01 1.03296566e+00
-4.05517995e-01 -6.85300007e-02 -3.08254957e-01 2.30096906e-01
-1.22463554e-02 1.25851011e+00 -9.93237495e-01 -1.26753375e-01
-4.58998680e-01 9.10811722e-01 3.11000347e-01 2.88970739e-01
-9.91171598e-01 -3.45228404e-01 4.74638075e-01 -9.58623067e-02
7.37907708e-01 -5.49422167e-02 -6.58610940e-01 -4.91676748e-01
6.41317517e-02 -5.08108974e-01 2.69665182e-01 -4.37848240e-01
-1.32095313e+00 6.78227723e-01 -1.64068237e-01 -8.28554571e-01
3.15963477e-01 -5.88862777e-01 -4.77576405e-01 6.43444359e-01
-2.10240269e+00 -1.39309752e+00 -3.28141242e-01 5.24666905e-01
2.30991736e-01 1.60921052e-01 5.13174593e-01 4.99534339e-01
-7.25772619e-01 6.14988863e-01 9.68043506e-02 5.10846317e-01
2.49083459e-01 -1.20036852e+00 4.65163618e-01 3.31479698e-01
-1.25335291e-01 2.97413796e-01 8.92043337e-02 -4.97424424e-01
-1.51737690e+00 -1.37966859e+00 4.40473855e-01 -1.62750080e-01
4.16175097e-01 -1.08964242e-01 -7.98701286e-01 5.31707346e-01
1.88107923e-01 4.34273630e-01 5.19717753e-01 -8.49458948e-02
-1.62111700e-01 -6.75291643e-02 -1.29516220e+00 5.17051339e-01
1.07302535e+00 -3.87368530e-01 -4.00406420e-01 2.36707166e-01
9.46029544e-01 -8.02721679e-01 -1.10689986e+00 6.64074361e-01
4.16209340e-01 -9.78209257e-01 1.31056094e+00 -2.01720610e-01
6.74411893e-01 -3.86164077e-02 9.62317809e-02 -1.00903094e+00
-4.09891218e-01 -3.54232043e-01 -4.29770350e-02 7.10830033e-01
5.03611743e-01 -7.38928735e-01 1.15210676e+00 1.91961646e-01
-4.70939487e-01 -1.35346007e+00 -1.32014644e+00 -5.93147635e-01
3.29935312e-01 -4.31665242e-01 5.90603232e-01 7.19090104e-01
-3.32481593e-01 4.03002918e-01 -1.22461401e-01 3.99813103e-03
7.44551003e-01 1.62625760e-01 1.97566748e-01 -9.17752922e-01
-2.60516435e-01 -7.22555757e-01 -4.88025099e-01 -1.33854449e+00
7.62650892e-02 -1.35035193e+00 -1.37126967e-01 -1.69248414e+00
3.42616737e-01 -6.17106915e-01 -6.34234250e-01 6.17489696e-01
-1.01670483e-02 7.91730225e-01 4.36263084e-02 1.02504022e-01
-7.56299734e-01 7.49052584e-01 1.89285791e+00 -6.02674782e-01
-3.29299599e-01 -2.92371735e-02 -7.33455181e-01 7.49010861e-01
7.72323966e-01 -3.35080653e-01 -3.13661814e-01 -5.53602219e-01
-1.69652142e-02 -5.09736352e-02 5.37534118e-01 -1.10825539e+00
3.10525179e-01 3.16891372e-01 6.31610155e-01 -9.89836156e-01
4.19694215e-01 -5.93985617e-01 -2.59444624e-01 6.64996028e-01
-5.39689437e-02 -7.04733655e-02 5.60416162e-01 3.81374806e-01
-5.91472723e-02 1.46137267e-01 8.61541092e-01 -1.90052643e-01
-7.63671994e-01 8.88704002e-01 -1.16854750e-01 2.34380692e-01
8.51284206e-01 -5.17546535e-01 -2.98639238e-01 2.39278272e-01
-5.11817694e-01 3.81966949e-01 3.72782014e-02 2.29100794e-01
8.69173169e-01 -1.12218046e+00 -6.49064600e-01 2.00250864e-01
3.03789657e-02 3.58372390e-01 4.26004231e-01 1.42871308e+00
-6.77417338e-01 5.48702538e-01 -2.99354196e-01 -1.12341952e+00
-8.50717604e-01 3.66002589e-01 5.90086222e-01 -5.33239841e-01
-1.29606831e+00 9.85012472e-01 1.12997696e-01 -6.10458553e-01
3.52518320e-01 -7.40055382e-01 -1.57345280e-01 -2.30075672e-01
2.16895655e-01 9.39830914e-02 1.91064224e-01 -4.00139600e-01
-5.81146419e-01 5.36487103e-01 -3.18314075e-01 3.23961407e-01
1.55528605e+00 7.04935044e-02 -3.60038169e-02 1.39045119e-01
1.49344110e+00 -3.65128428e-01 -1.47248828e+00 -4.39481914e-01
-2.55341023e-01 -6.41424768e-03 2.80411959e-01 -7.17753291e-01
-1.70306098e+00 8.68860841e-01 6.05218828e-01 -1.11961856e-01
9.74533498e-01 5.04204705e-02 1.29233027e+00 1.64864913e-01
6.71259761e-02 -8.14890921e-01 4.35408093e-02 4.77191389e-01
4.13134605e-01 -1.33098626e+00 -1.00286407e-02 -3.46283764e-01
-6.12940073e-01 7.80412495e-01 4.73313242e-01 -1.79768369e-01
9.32162464e-01 4.08522755e-01 3.70537452e-02 -5.20636857e-01
-2.31511489e-01 -1.97605118e-01 1.57972857e-01 6.39619827e-01
2.09203258e-01 4.75244150e-02 7.26987571e-02 4.03021157e-01
-2.66818963e-02 1.60129145e-01 7.65368044e-02 8.32202077e-01
-8.61175582e-02 -7.37895250e-01 -2.17417572e-02 6.62199378e-01
-5.66052914e-01 -3.58952671e-01 -4.88131009e-02 9.09328640e-01
1.63785249e-01 3.81645203e-01 2.64048427e-01 -2.08599731e-01
2.39532366e-01 -4.70874280e-01 7.15992272e-01 -2.32070118e-01
-8.28134656e-01 -1.31684855e-01 -2.26649180e-01 -6.48359060e-01
-5.32068670e-01 -3.78135026e-01 -1.46885693e+00 -3.34850550e-01
2.82547045e-02 -1.97661474e-01 6.14881098e-01 1.04033279e+00
4.26015973e-01 1.07330692e+00 4.42783147e-01 -9.79916930e-01
-4.13474768e-01 -6.37371242e-01 -4.76073563e-01 8.77785832e-02
3.21857005e-01 -6.65174365e-01 1.76058069e-01 -3.70283782e-01] | [14.574784278869629, -2.556323289871216] |
e4fd2a34-8fd5-4341-8bcf-0784ee826617 | deep-generative-modeling-of-lidar-data | 1812.01180 | null | https://arxiv.org/abs/1812.01180v4 | https://arxiv.org/pdf/1812.01180v4.pdf | Deep Generative Modeling of LiDAR Data | Building models capable of generating structured output is a key challenge for AI and robotics. While generative models have been explored on many types of data, little work has been done on synthesizing lidar scans, which play a key role in robot mapping and localization. In this work, we show that one can adapt deep generative models for this task by unravelling lidar scans into a 2D point map. Our approach can generate high quality samples, while simultaneously learning a meaningful latent representation of the data. We demonstrate significant improvements against state-of-the-art point cloud generation methods. Furthermore, we propose a novel data representation that augments the 2D signal with absolute positional information. We show that this helps robustness to noisy and imputed input; the learned model can recover the underlying lidar scan from seemingly uninformative data | ['Lucas Caccia', 'Herke van Hoof', 'Joelle Pineau', 'Aaron Courville'] | 2018-12-04 | null | null | null | null | ['point-cloud-generation'] | ['computer-vision'] | [ 4.35765564e-01 4.32516903e-01 -5.45713715e-02 -6.08540297e-01
-1.02045703e+00 -7.34958112e-01 9.22527075e-01 -2.31929362e-01
-8.82759914e-02 7.50181615e-01 2.23750934e-01 -1.93315804e-01
-4.76980209e-03 -1.28253102e+00 -1.21268725e+00 -4.92300570e-01
2.50331044e-01 1.30509615e+00 1.13854900e-01 -1.22928642e-01
1.60734281e-01 8.05896640e-01 -1.61862254e+00 -9.44811851e-02
9.60025668e-01 5.54547966e-01 4.51168150e-01 5.52699804e-01
-2.61950403e-01 2.99439400e-01 -4.81987834e-01 -1.46864459e-01
4.62562829e-01 -3.85521322e-01 -3.86610270e-01 1.61019430e-01
6.75361454e-01 -2.41948918e-01 -3.60664666e-01 8.24810147e-01
2.67739862e-01 -9.14406851e-02 9.73187804e-01 -1.29979563e+00
-7.52581954e-01 5.85946679e-01 -4.26871985e-01 -5.72700143e-01
8.94986317e-02 6.85591400e-02 7.86211967e-01 -9.26854253e-01
7.23070800e-01 1.46486199e+00 7.49254167e-01 4.74031746e-01
-1.80598068e+00 -7.51152635e-01 -1.79759875e-01 -2.46789768e-01
-1.26368177e+00 -2.88562208e-01 8.95793021e-01 -4.63972330e-01
4.89501894e-01 -1.08320534e-01 7.36026287e-01 1.25032187e+00
2.36263573e-01 7.36842394e-01 9.45894361e-01 -2.33681068e-01
2.88185120e-01 2.32459698e-03 -4.72192407e-01 6.68354392e-01
4.39065129e-01 2.21693248e-01 -6.48763716e-01 -1.17843121e-01
1.16469419e+00 2.01686084e-01 1.29286394e-01 -9.30145264e-01
-1.13721013e+00 9.37097192e-01 6.29412353e-01 -1.02592908e-01
-4.61097717e-01 6.81408048e-01 -4.29837137e-01 -5.18425591e-02
4.21074569e-01 5.36864400e-01 -1.13857612e-01 -1.38797089e-01
-1.08774757e+00 5.05326033e-01 7.19327152e-01 1.40332305e+00
1.19026721e+00 1.42757758e-01 2.50807166e-01 6.06569052e-01
4.41205949e-01 1.07026017e+00 9.47453752e-02 -1.15962672e+00
3.35780174e-01 4.69161719e-01 1.36703774e-01 -7.38142431e-01
-8.22282434e-02 -3.98946881e-01 -6.19780481e-01 4.00652915e-01
8.77675936e-02 -4.55792947e-03 -1.30835199e+00 1.67805421e+00
1.08270787e-01 2.11204901e-01 2.35224552e-02 5.21516621e-01
4.05144393e-01 5.17953873e-01 -3.02548558e-01 4.43562150e-01
8.54499519e-01 -5.16897202e-01 -3.25620860e-01 -6.12649560e-01
3.98811027e-02 -5.93981564e-01 8.36386621e-01 2.88313389e-01
-1.08125281e+00 -5.58296204e-01 -1.04350054e+00 -3.91713709e-01
-2.08448529e-01 -7.53594339e-02 7.40232289e-01 4.33466524e-01
-9.58340287e-01 4.30509627e-01 -1.19214451e+00 -2.58695751e-01
5.98600626e-01 3.75184804e-01 -4.43551719e-01 -4.35638338e-01
-5.95600009e-01 9.66573298e-01 2.00711474e-01 -1.48696616e-01
-1.06736696e+00 -7.18550146e-01 -9.98003125e-01 -2.55769163e-01
2.75722355e-01 -1.06182790e+00 1.34106410e+00 -1.85347855e-01
-1.18564713e+00 6.15367472e-01 -4.91413057e-01 -6.05365396e-01
4.94865417e-01 -2.92581379e-01 1.37877807e-01 -1.67849422e-01
3.17441136e-01 1.19380808e+00 8.49949360e-01 -1.79257274e+00
-3.97862345e-01 -5.46562076e-01 -1.96628779e-01 8.33829567e-02
2.86260843e-01 -7.66510725e-01 -1.89539805e-01 -4.47078466e-01
6.26025617e-01 -1.05403316e+00 -3.40834588e-01 1.53933704e-01
-4.93147969e-01 3.79951485e-02 9.49687004e-01 -2.24274650e-01
3.93510967e-01 -1.70847571e+00 1.88146859e-01 3.20471257e-01
2.91006686e-03 -1.79540619e-01 -1.30842730e-01 6.51220739e-01
3.01798731e-01 2.60526896e-01 -6.35417163e-01 -7.25285292e-01
2.71576792e-01 7.00925529e-01 -8.92624617e-01 2.83462614e-01
5.53630531e-01 1.28471363e+00 -1.00631177e+00 -1.63502663e-01
3.27003390e-01 7.30777264e-01 -6.67170584e-01 6.91282451e-02
-5.93471110e-01 6.47961795e-01 -4.04813379e-01 5.32826185e-01
6.40918851e-01 3.66923101e-02 -5.49159758e-02 6.32850379e-02
8.24295506e-02 4.61341113e-01 -1.08266950e+00 2.25497293e+00
-5.58306515e-01 6.46306396e-01 -6.67959601e-02 -6.93319976e-01
1.35529566e+00 -9.17504206e-02 4.40986782e-01 -3.42070788e-01
-8.11529532e-02 2.24668577e-01 -1.81952953e-01 -6.29296973e-02
8.36687386e-01 -4.36277837e-01 -1.82646379e-01 5.95642745e-01
1.00684978e-01 -1.12305033e+00 -2.31036767e-01 1.53977886e-01
1.02536798e+00 7.63931990e-01 -3.77263725e-02 1.11326724e-01
-3.50529343e-01 3.55325341e-01 3.45262438e-01 7.49470174e-01
6.30095363e-01 1.02109647e+00 2.31207356e-01 -2.00574577e-01
-1.38298988e+00 -1.50627959e+00 -7.17372969e-02 2.95541465e-01
2.58950293e-02 -3.34050179e-01 -4.72658187e-01 -3.28097939e-01
4.45587814e-01 1.01684797e+00 -5.58643639e-01 -1.94300726e-01
-5.74764132e-01 -5.38719535e-01 6.27211511e-01 6.95811927e-01
2.22522050e-01 -7.69293368e-01 -6.86577797e-01 2.91326582e-01
-5.31476885e-02 -1.00635827e+00 -2.35616378e-02 2.74734706e-01
-1.09079897e+00 -9.01450992e-01 -2.44266137e-01 -5.15558660e-01
8.01878214e-01 3.15004826e-01 1.24598432e+00 -4.18456435e-01
-2.91309685e-01 3.42307955e-01 -2.82662380e-02 -8.31404328e-01
-6.58732891e-01 2.27792948e-01 9.22352821e-03 -4.05553967e-01
3.88690799e-01 -1.04007506e+00 -1.28821835e-01 1.84171706e-01
-9.88683939e-01 2.67424375e-01 8.51230681e-01 7.18643546e-01
9.94986296e-01 -3.08421522e-01 4.04123425e-01 -9.98070240e-01
5.06205440e-01 -5.48159480e-01 -6.69078112e-01 -2.87750810e-01
-5.50832570e-01 6.56747520e-01 1.76135898e-01 -2.61289835e-01
-7.04069316e-01 5.12333870e-01 -1.90057009e-01 -6.96336567e-01
-3.52506429e-01 3.14001143e-01 -2.28838101e-01 8.79543424e-02
7.43894398e-01 2.21754417e-01 1.31092891e-01 -6.08920872e-01
8.39708447e-01 4.28249359e-01 8.97490323e-01 -8.52609813e-01
1.32001483e+00 7.68877864e-01 5.11337936e-01 -7.93962598e-01
-5.77759266e-01 -1.39752612e-01 -1.00453997e+00 3.18368554e-01
6.37000024e-01 -9.51292217e-01 -1.90272555e-01 3.28205638e-02
-1.25996125e+00 -3.51747692e-01 -7.00404584e-01 3.04537475e-01
-1.01226103e+00 4.58989441e-02 -1.72817092e-02 -8.92658472e-01
9.99367833e-02 -1.15188336e+00 1.65921879e+00 -3.55947204e-02
-2.57091761e-01 -6.75521195e-01 3.67827952e-01 -2.07901262e-02
1.31213918e-01 5.66695154e-01 7.80938923e-01 -2.85990000e-01
-1.29057169e+00 -2.82847255e-01 -3.41937765e-02 8.87041539e-02
2.51614958e-01 -1.57441664e-02 -1.08067155e+00 7.76865408e-02
-1.32996529e-01 -2.03446031e-01 9.19676840e-01 3.75763297e-01
8.80045533e-01 -3.09178941e-02 -5.93635857e-01 8.35708141e-01
1.14755857e+00 2.16352530e-02 6.63672864e-01 7.69519955e-02
8.17574203e-01 5.97516418e-01 4.81731832e-01 1.15455925e-01
5.75715780e-01 5.94869733e-01 8.23724031e-01 2.90079713e-02
-1.85366243e-01 -1.02965569e+00 -2.83957440e-02 3.39743078e-01
1.29594281e-01 -6.66191354e-02 -1.11206472e+00 7.27883637e-01
-1.86881161e+00 -8.78659725e-01 -1.43892914e-01 2.09142256e+00
6.42927885e-01 1.02999799e-01 -1.10658534e-01 1.48692774e-02
3.50825518e-01 -2.74188481e-02 -7.05239832e-01 7.92873576e-02
-1.87661663e-01 6.01668477e-01 6.56940579e-01 5.69328845e-01
-7.38257229e-01 1.08837855e+00 6.75754261e+00 3.10564607e-01
-9.13981080e-01 -8.01226571e-02 3.38341706e-02 -2.97024902e-02
-9.56421196e-01 2.66740769e-01 -7.76047766e-01 2.26215541e-01
7.66888440e-01 -5.95833324e-02 3.14804286e-01 1.01237190e+00
8.74753967e-02 -4.29647118e-02 -1.43682718e+00 9.87572014e-01
1.48423344e-01 -1.45643866e+00 1.38488546e-01 6.03114009e-01
8.35083187e-01 2.53233105e-01 7.31310248e-02 6.90860078e-02
8.78938496e-01 -1.28952360e+00 9.68537509e-01 7.43246853e-01
8.72908890e-01 -7.72057772e-01 2.89132863e-01 7.71161020e-01
-8.67498457e-01 1.96751773e-01 -6.43106937e-01 1.16753288e-01
4.73865747e-01 7.02817619e-01 -1.48182714e+00 5.85227966e-01
3.46417814e-01 7.05158949e-01 -5.05919039e-01 1.00052750e+00
-6.83380842e-01 4.63887304e-01 -7.06815124e-01 2.77702570e-01
2.66873725e-02 -4.04292047e-01 6.51940882e-01 7.42509246e-01
7.32434392e-01 -1.88076138e-01 1.37123272e-01 1.63268864e+00
-2.75166810e-01 -5.37059009e-01 -1.32138634e+00 -8.76226202e-02
7.89099097e-01 1.01593077e+00 -5.42369068e-01 -1.33531913e-02
1.74781606e-01 7.33749866e-01 1.85279176e-01 2.08406925e-01
-5.21155000e-01 -2.26517394e-01 7.86103964e-01 2.42488191e-01
2.71121234e-01 -8.56849134e-01 -6.97189152e-01 -9.88818407e-01
-5.84524348e-02 -3.45380425e-01 -3.47035617e-01 -1.10265946e+00
-1.11792493e+00 3.85269314e-01 1.76228538e-01 -1.12081528e+00
-1.03596485e+00 -5.41377842e-01 -4.13573861e-01 1.22395349e+00
-1.33215165e+00 -1.51193273e+00 -5.12727559e-01 1.30799249e-01
5.66770196e-01 -3.68575566e-02 7.63836563e-01 -2.50186086e-01
2.27891281e-01 2.09634556e-04 1.58818029e-02 5.37870859e-04
5.89051425e-01 -1.41262364e+00 1.10511327e+00 8.82889688e-01
7.41490960e-01 8.21076989e-01 8.22353661e-01 -9.10313249e-01
-1.67960310e+00 -1.14041281e+00 5.03614604e-01 -9.58709180e-01
3.48405719e-01 -6.66212440e-01 -6.97952271e-01 1.00715816e+00
-2.87531137e-01 -1.29050106e-01 3.25036258e-01 -6.71451353e-03
-2.92862594e-01 1.00069389e-01 -1.08172143e+00 3.69575709e-01
1.25586510e+00 -6.57211900e-01 -7.75274873e-01 4.17642035e-02
8.32935750e-01 -6.21907175e-01 -5.12713909e-01 3.35772902e-01
6.47439778e-01 -8.42664957e-01 1.11745000e+00 -4.20691282e-01
4.95557994e-01 -3.95013332e-01 -4.03286934e-01 -1.64423597e+00
-1.52856946e-01 -3.93247217e-01 -1.14549361e-01 1.08165717e+00
2.61717498e-01 -4.96205300e-01 1.21642804e+00 4.38996136e-01
-4.92861450e-01 -4.39398378e-01 -6.96505606e-01 -7.47934878e-01
2.43821546e-01 -7.37310350e-01 1.05284953e+00 5.26908696e-01
-6.55351520e-01 4.35386062e-01 -2.76413411e-01 2.47527391e-01
8.50210726e-01 3.06234717e-01 1.42659712e+00 -1.62836313e+00
-2.40950733e-01 -2.16785520e-01 -5.22751987e-01 -1.27849960e+00
2.67975807e-01 -8.87349308e-01 3.94292474e-01 -2.11232257e+00
-2.41932437e-01 -8.35269630e-01 5.96928775e-01 4.52795982e-01
2.22592086e-01 4.33902174e-01 2.21741363e-01 2.91294217e-01
2.47740045e-01 6.67416096e-01 1.08574045e+00 -8.64442885e-02
-5.48682027e-02 8.77669901e-02 -8.31351221e-01 6.19302928e-01
5.63042045e-01 -6.37973368e-01 -6.08860731e-01 -8.20842683e-01
2.85593420e-01 -2.51802534e-01 6.79307044e-01 -1.07228386e+00
2.00180739e-01 -3.58678579e-01 4.77636129e-01 -1.15357459e+00
8.06798756e-01 -8.70094419e-01 5.64250648e-01 2.02191830e-01
-6.39175475e-02 8.56099576e-02 2.82219406e-02 7.34286606e-01
2.25882996e-02 -2.57859528e-01 4.80236828e-01 -2.15371832e-01
-3.93100321e-01 4.05799031e-01 1.12077668e-01 -1.33492440e-01
6.99571073e-01 -3.07659745e-01 -2.18644917e-01 -5.32434225e-01
-4.51883525e-01 1.79747865e-01 1.04482388e+00 5.03174305e-01
7.49612272e-01 -1.40400922e+00 -6.90445244e-01 5.13512850e-01
1.09207131e-01 9.46895897e-01 -3.33400369e-01 1.71142206e-01
-5.08539259e-01 4.22947168e-01 -1.18100107e-01 -9.90565360e-01
-7.14905441e-01 2.34958231e-01 7.26115704e-02 1.69049442e-01
-6.79415584e-01 6.94064498e-01 1.17151156e-01 -8.48564506e-01
-1.35335848e-01 -5.88916063e-01 5.27592123e-01 -1.28777176e-01
1.38644710e-01 1.54284075e-01 -8.39400440e-02 -7.27240384e-01
-1.84535142e-02 6.45338416e-01 4.83244240e-01 -6.09842837e-01
1.59646666e+00 4.99862656e-02 1.10337064e-01 7.25490510e-01
8.94263387e-01 1.64829478e-01 -1.49735165e+00 -2.02862069e-01
-9.65012461e-02 -5.33429325e-01 -1.80348963e-01 -5.23415923e-01
-5.44694901e-01 1.30004418e+00 1.63653851e-01 1.31410751e-02
4.83194113e-01 2.27956519e-01 6.47373915e-01 5.30779481e-01
9.77138937e-01 -5.54455519e-01 1.68675128e-02 5.20498216e-01
8.82543266e-01 -1.07128596e+00 -6.39478822e-05 -4.46963161e-01
-4.34973896e-01 8.93885076e-01 2.16234624e-01 -4.27569479e-01
3.47413570e-01 5.05937755e-01 -2.50555098e-01 -2.29328245e-01
-5.34466267e-01 -2.70155400e-01 2.59101152e-01 1.15421546e+00
1.73414424e-02 1.30808085e-01 5.34140229e-01 3.49771649e-01
-8.55972469e-01 -5.37799718e-03 5.26034236e-01 1.06280053e+00
-6.81074142e-01 -1.52130389e+00 -5.62697887e-01 4.02687639e-01
1.66048527e-01 -2.45791320e-02 -5.91334164e-01 8.96693051e-01
1.42953306e-01 3.99872780e-01 2.89125592e-01 -4.15915787e-01
3.41209352e-01 2.16367334e-01 7.47539580e-01 -1.03223634e+00
1.43182591e-01 5.73264100e-02 -1.62531346e-01 -4.40206319e-01
-2.34465241e-01 -8.98477733e-01 -1.30008876e+00 -8.46901014e-02
-1.67637784e-02 1.19428843e-01 1.07927799e+00 7.46324062e-01
5.33014119e-01 3.59981954e-01 2.85115153e-01 -1.46218729e+00
-4.97914791e-01 -9.58048999e-01 -4.80458736e-01 1.53917968e-01
4.21906263e-01 -9.81717587e-01 -2.54095942e-02 2.21157804e-01] | [8.782988548278809, -3.6472926139831543] |
a4df0654-d43a-4222-b2d7-56f991c80a7f | worst-case-matters-for-few-shot-recognition | 2203.06574 | null | https://arxiv.org/abs/2203.06574v2 | https://arxiv.org/pdf/2203.06574v2.pdf | Worst Case Matters for Few-Shot Recognition | Few-shot recognition learns a recognition model with very few (e.g., 1 or 5) images per category, and current few-shot learning methods focus on improving the average accuracy over many episodes. We argue that in real-world applications we may often only try one episode instead of many, and hence maximizing the worst-case accuracy is more important than maximizing the average accuracy. We empirically show that a high average accuracy not necessarily means a high worst-case accuracy. Since this objective is not accessible, we propose to reduce the standard deviation and increase the average accuracy simultaneously. In turn, we devise two strategies from the bias-variance tradeoff perspective to implicitly reach this goal: a simple yet effective stability regularization (SR) loss together with model ensemble to reduce variance during fine-tuning, and an adaptability calibration mechanism to reduce the bias. Extensive experiments on benchmark datasets demonstrate the effectiveness of the proposed strategies, which outperforms current state-of-the-art methods with a significant margin in terms of not only average, but also worst-case accuracy. Our code is available at https://github.com/heekhero/ACSR. | ['Jianxin Wu', 'Yun-Hao Cao', 'Minghao Fu'] | 2022-03-13 | null | null | null | null | ['few-shot-image-classification'] | ['computer-vision'] | [ 2.20305502e-01 -2.21008450e-01 -3.26998651e-01 -4.57471102e-01
-9.94724572e-01 -3.04081976e-01 4.87331718e-01 -8.11084509e-02
-4.17920172e-01 6.91610873e-01 -1.38094902e-01 3.76318432e-02
-2.96760261e-01 -6.56521380e-01 -5.97604334e-01 -9.16521132e-01
2.77483881e-01 2.74990618e-01 3.95822197e-01 3.17904353e-02
3.94797832e-01 1.83429673e-01 -1.64418554e+00 -7.60262236e-02
1.09360170e+00 1.28595841e+00 6.17691167e-02 3.98111343e-01
-2.60273740e-02 8.24512362e-01 -4.72321928e-01 -5.27391374e-01
4.25714433e-01 -4.38969165e-01 -3.53280634e-01 3.23005199e-01
2.45915994e-01 -2.88317859e-01 -1.07553199e-01 1.17916036e+00
4.27363366e-01 4.62414682e-01 6.95058227e-01 -1.29179275e+00
-6.25834167e-01 4.96510804e-01 -7.32161641e-01 2.28534907e-01
-7.83726759e-03 5.12768686e-01 1.17010486e+00 -8.86175394e-01
2.28876457e-01 6.52275562e-01 3.38281810e-01 6.28125668e-01
-1.16377401e+00 -7.22317040e-01 3.59402686e-01 2.37881139e-01
-1.55400372e+00 -7.28838742e-01 7.67229915e-01 -3.88913512e-01
5.95500827e-01 2.00472102e-01 3.91856909e-01 8.74602377e-01
4.51816358e-02 6.59933805e-01 9.84194815e-01 -3.46866041e-01
5.99617720e-01 3.30061913e-01 4.63716477e-01 5.48222661e-01
4.14197177e-01 2.32459102e-02 -4.32788163e-01 -1.25925168e-01
5.25904715e-01 2.60079056e-01 -2.96120197e-01 -6.73349977e-01
-8.84165168e-01 9.06758130e-01 3.18355292e-01 2.55876780e-01
-4.50851083e-01 9.37407762e-02 3.31997693e-01 2.33087093e-01
3.63086194e-01 5.16682982e-01 -2.23709077e-01 -3.77246648e-01
-8.46035361e-01 3.01597510e-02 7.16305017e-01 7.66416013e-01
7.50002742e-01 6.41682325e-03 -2.45788693e-01 1.17143774e+00
-2.23239854e-01 1.69306949e-01 4.44243431e-01 -8.61105919e-01
2.97694504e-01 4.19067711e-01 4.20466810e-01 -7.15617597e-01
-3.05100139e-02 -5.94785810e-01 -8.29908669e-01 -3.12455893e-02
4.03613001e-01 -1.03630371e-01 -9.10933733e-01 1.92380154e+00
2.40557671e-01 3.64072055e-01 -1.82313204e-01 1.07500839e+00
3.18837941e-01 5.91191530e-01 8.75702724e-02 -6.49013877e-01
1.10199487e+00 -9.04878259e-01 -6.52627289e-01 -3.57268393e-01
4.46231335e-01 -6.31047130e-01 1.39181983e+00 2.71822065e-01
-8.34847212e-01 -3.18217725e-01 -1.13222194e+00 4.61029917e-01
-9.25998762e-02 1.50613457e-01 5.58993697e-01 6.68583274e-01
-4.72652197e-01 6.02526903e-01 -7.74387300e-01 -3.65874380e-01
4.73028183e-01 1.84884921e-01 -4.91388924e-02 3.84118385e-03
-1.04203272e+00 7.63924897e-01 3.09547961e-01 -9.15624872e-02
-6.05311930e-01 -8.30893636e-01 -5.02958238e-01 2.32299283e-01
8.55373502e-01 -5.30279398e-01 1.40595376e+00 -9.83031571e-01
-1.59171522e+00 5.63434839e-01 -6.37511984e-02 -4.72132713e-01
6.11886084e-01 -3.40425819e-01 -2.43033975e-01 -1.12095781e-01
-2.85037071e-01 3.45523417e-01 7.71852255e-01 -1.12675560e+00
-6.51671231e-01 -2.93261945e-01 2.04278722e-01 2.48762518e-01
-6.78510427e-01 -2.92735010e-01 -4.47812229e-01 -5.10262549e-01
-4.26962078e-02 -9.45630431e-01 -2.45589197e-01 -2.16998532e-02
-1.42894790e-01 -2.10338056e-01 5.05034149e-01 -2.51717031e-01
1.48504972e+00 -2.20030499e+00 -5.31808622e-02 -1.67573255e-03
4.34672683e-02 3.38402033e-01 4.05505635e-02 2.35292256e-01
2.09739134e-01 -3.19934711e-02 -3.04478198e-01 -1.55811952e-02
-8.28071162e-02 4.41214032e-02 -2.47277156e-01 5.94178140e-01
1.61443040e-01 7.35860467e-01 -8.79824221e-01 -2.41423473e-01
2.48243839e-01 1.80058897e-01 -4.57138360e-01 1.82125956e-01
-1.07892953e-01 1.44589111e-01 -3.62277001e-01 5.42200923e-01
5.06389320e-01 -6.67800188e-01 1.01928517e-01 -7.01765046e-02
5.02195284e-02 8.50113332e-02 -1.50209951e+00 1.35214579e+00
-5.29121041e-01 3.17261308e-01 -2.42673144e-01 -1.09399748e+00
8.70505393e-01 4.21943888e-02 4.36791331e-01 -5.47757328e-01
2.68882841e-01 1.02012105e-01 1.29792690e-01 -1.61363363e-01
1.98369756e-01 -2.77150840e-01 7.75915757e-02 4.84053522e-01
-7.01460540e-02 2.01004013e-01 1.84697255e-01 9.57586542e-02
9.23084259e-01 -2.99494535e-01 7.58968592e-01 -2.21106842e-01
2.90399730e-01 -3.07795018e-01 8.45091999e-01 8.63220751e-01
-5.60000479e-01 6.13021791e-01 5.73815346e-01 -1.58365265e-01
-1.05492020e+00 -8.25417161e-01 -5.28516434e-02 1.07887816e+00
4.00077462e-01 -2.40557015e-01 -6.97548628e-01 -5.71639717e-01
-5.03000580e-02 9.37893748e-01 -6.48669600e-01 -4.87710625e-01
-3.06106448e-01 -8.80189538e-01 3.05303279e-02 5.29366195e-01
4.50876802e-01 -7.49364674e-01 -7.64957607e-01 7.31436461e-02
-1.33852344e-02 -9.65832412e-01 -5.61116278e-01 1.23570681e-01
-6.87536597e-01 -8.89649093e-01 -7.81045914e-01 -3.60146016e-01
5.23731589e-01 5.22278726e-01 8.79796803e-01 9.34840068e-02
-1.78183839e-01 9.42851380e-02 -4.70389336e-01 -3.91166687e-01
-8.88566428e-04 1.65446863e-01 2.02879414e-01 2.55154014e-01
4.99105752e-01 -5.38242638e-01 -7.24608839e-01 3.91173035e-01
-7.00126767e-01 -1.22336838e-02 6.95076406e-01 1.14684772e+00
6.17476165e-01 -4.02545705e-02 8.42069328e-01 -8.70074034e-01
3.94581258e-01 -4.23909396e-01 -6.82680070e-01 5.01737297e-01
-9.36742663e-01 -1.51455104e-01 6.82243049e-01 -8.26369703e-01
-9.12127852e-01 -2.36116033e-02 1.08085744e-01 -8.92827690e-01
1.10834725e-02 2.74908751e-01 -6.99555799e-02 3.13448980e-02
7.04422176e-01 3.90753716e-01 5.92259839e-02 -1.81570902e-01
2.73405939e-01 7.15553522e-01 1.52090386e-01 -3.79755467e-01
7.18522489e-01 3.31519723e-01 -2.56177545e-01 -7.97888458e-01
-1.12665570e+00 -6.58237040e-01 -4.40997481e-01 -3.20188940e-01
3.40351671e-01 -7.95088649e-01 -4.85037684e-01 4.78889048e-01
-5.50153375e-01 -3.05922747e-01 -3.35289866e-01 3.14153284e-01
-6.50462747e-01 1.17300831e-01 -2.92869866e-01 -1.09978449e+00
-5.48079967e-01 -1.04382992e+00 7.59091556e-01 4.85843629e-01
-1.32936031e-01 -4.62250203e-01 -1.46536037e-01 2.21627355e-01
4.36445922e-01 -4.68203388e-02 5.59991062e-01 -7.49970853e-01
-5.21470010e-01 -2.92762727e-01 -2.27957636e-01 2.68121392e-01
1.17528185e-01 2.50910502e-02 -8.98027241e-01 -4.14282441e-01
-4.11407053e-02 -2.43910402e-01 9.49799240e-01 4.65643525e-01
1.25062633e+00 -3.32812726e-01 -3.98964956e-02 4.19574648e-01
1.64172113e+00 3.74998897e-01 7.13161290e-01 6.70151040e-02
5.25133312e-01 3.37610573e-01 9.55438912e-01 7.99897850e-01
5.42458817e-02 8.73910725e-01 1.21850498e-01 5.01237869e-01
4.52058353e-02 -7.42063373e-02 2.44856015e-01 7.10826576e-01
4.02147509e-02 -1.92047596e-01 -9.00649786e-01 6.22607887e-01
-2.03615594e+00 -1.08725464e+00 5.11630535e-01 2.58784270e+00
7.51905024e-01 4.15357918e-01 3.35852891e-01 7.58872330e-02
8.91945899e-01 2.29890332e-01 -9.12936687e-01 -1.00000344e-01
1.85723796e-01 -2.14862168e-01 3.87164384e-01 3.91851485e-01
-1.11881483e+00 8.70764017e-01 5.50498533e+00 1.17621195e+00
-1.22442722e+00 1.30552173e-01 8.27238500e-01 -5.59278429e-01
3.05948372e-04 2.48290934e-02 -8.04001808e-01 6.30667269e-01
8.70212615e-01 -5.97515881e-01 5.42290390e-01 1.16594481e+00
6.55703023e-02 -1.22749455e-01 -9.50578392e-01 1.09804595e+00
4.56625223e-02 -1.31635594e+00 -9.33758691e-02 -5.49509265e-02
7.04410315e-01 -1.71764702e-01 -1.48681123e-02 4.42023188e-01
7.09123269e-04 -6.61841691e-01 6.96255207e-01 6.88728511e-01
7.37491667e-01 -1.01425791e+00 6.66974306e-01 6.32864654e-01
-1.13189650e+00 -3.12527388e-01 -5.02122283e-01 2.94354353e-02
-1.10705802e-02 7.99275815e-01 -2.49469563e-01 2.91904151e-01
5.64579844e-01 5.77877343e-01 -3.68262053e-01 1.15911818e+00
-1.50700182e-01 5.71499944e-01 -2.59693474e-01 -4.76336747e-01
6.66582286e-02 -3.28690022e-01 5.20085633e-01 9.47751939e-01
2.40094274e-01 3.96752954e-01 2.32905716e-01 6.52569234e-01
-8.81749615e-02 3.08916986e-01 -4.54150319e-01 -1.16362289e-01
8.65900636e-01 1.12671018e+00 -7.20555246e-01 -4.84540135e-01
-5.08484721e-01 6.62096083e-01 4.61729735e-01 2.61088341e-01
-1.03466702e+00 -5.20563066e-01 7.51201451e-01 6.03352077e-02
5.86132050e-01 1.14657782e-01 -5.56231916e-01 -1.30357420e+00
2.37094253e-01 -9.04525578e-01 5.05907834e-01 -2.87977338e-01
-1.49123228e+00 3.95999640e-01 -2.35017896e-01 -1.58643162e+00
-5.45743778e-02 -2.10755855e-01 -6.65670991e-01 4.91212159e-01
-1.23011363e+00 -7.58613348e-01 -3.40131670e-01 1.52932063e-01
8.03486526e-01 -3.70759033e-02 5.03229499e-01 2.31761783e-01
-9.55939770e-01 9.85395849e-01 2.16474771e-01 -8.35840106e-02
6.69237196e-01 -8.85899544e-01 -5.03208376e-02 8.36151659e-01
6.34188801e-02 6.94073021e-01 9.84699726e-01 -2.55775392e-01
-1.40870762e+00 -9.67814982e-01 3.81974816e-01 -1.88556001e-01
6.66371047e-01 -1.80490479e-01 -1.15078628e+00 3.32418442e-01
-2.17761338e-01 8.66973102e-02 6.64937794e-01 2.59365380e-01
-3.12965870e-01 -3.84560406e-01 -1.10095179e+00 7.94370115e-01
1.05628538e+00 -2.89654791e-01 -3.68581384e-01 2.47003645e-01
5.58739722e-01 -1.53215259e-01 -8.19447458e-01 5.42739391e-01
7.27935314e-01 -9.46596920e-01 7.47968733e-01 -4.98459011e-01
3.86282742e-01 -2.36499161e-01 -4.39499617e-01 -1.33699572e+00
-3.37839067e-01 -3.73328298e-01 -5.48578441e-01 1.21300244e+00
4.40203696e-01 -6.71017051e-01 8.04675400e-01 7.51702845e-01
3.37168217e-01 -1.21322131e+00 -8.95558417e-01 -1.23600841e+00
-2.67745797e-02 -3.80966961e-01 4.61551458e-01 8.16991925e-01
-2.75169853e-02 3.94335300e-01 -7.40858376e-01 6.24539182e-02
8.70959163e-01 4.28072810e-01 8.60144258e-01 -1.05647326e+00
-4.26226616e-01 -5.30665159e-01 -4.88463461e-01 -9.54627335e-01
-5.61527535e-02 -3.19819987e-01 2.19214007e-01 -1.18424821e+00
7.00428486e-01 -3.17909241e-01 -7.02444017e-01 3.27531457e-01
-5.19786477e-01 3.95020507e-02 2.82886833e-01 3.30427021e-01
-8.46647680e-01 7.05918193e-01 8.37209165e-01 6.48738146e-02
-3.61049294e-01 1.18256353e-01 -8.90504658e-01 5.00640035e-01
8.85525703e-01 -3.77850622e-01 -4.62956220e-01 2.00366005e-02
-9.47679505e-02 -4.27740552e-02 9.74414498e-02 -1.18852317e+00
2.83227533e-01 -5.69915891e-01 1.29487246e-01 -2.01832384e-01
3.41692358e-01 -6.01257980e-01 9.99560356e-02 4.56970215e-01
-4.32215989e-01 -6.14568532e-01 -1.78606793e-01 7.70885885e-01
-1.07127979e-01 -2.29664415e-01 1.47650743e+00 7.26179183e-02
-8.90553176e-01 5.01199245e-01 6.22427128e-02 -3.87009718e-02
1.38442445e+00 -1.22180216e-01 -4.34321851e-01 -5.53883791e-01
-4.66918647e-01 2.51427531e-01 6.01557970e-01 5.01859665e-01
4.48784292e-01 -1.34637868e+00 -5.47390223e-01 -6.63981736e-02
5.02351344e-01 -6.02092564e-01 4.71931368e-01 1.06051481e+00
6.53218627e-02 2.78519630e-01 -3.26167271e-02 -5.29842079e-01
-1.21710002e+00 6.90156996e-01 3.57459337e-01 -2.56819576e-01
-5.77850640e-01 8.55046153e-01 2.49636278e-01 -6.08356968e-02
3.79985720e-01 2.96811581e-01 9.81107913e-03 1.80194423e-01
8.37677836e-01 3.73502851e-01 1.48644950e-02 -3.07000846e-01
-4.70875531e-01 5.43921471e-01 -4.46447909e-01 -1.85515489e-02
1.27442229e+00 -2.09118024e-01 4.24433082e-01 7.43576646e-01
1.01289928e+00 -4.52833354e-01 -1.43988502e+00 -4.08628315e-01
-9.03841406e-02 -8.30718100e-01 1.08719110e-01 -6.53082967e-01
-9.79540825e-01 7.08770812e-01 6.44534111e-01 3.06797087e-01
1.37853527e+00 -4.22472544e-02 6.91981435e-01 4.47364777e-01
6.03591084e-01 -1.28675735e+00 1.79970235e-01 3.94124717e-01
7.43517280e-01 -1.58351123e+00 1.50669634e-01 -3.08826685e-01
-1.07271564e+00 7.21159995e-01 8.46264422e-01 -2.35735521e-01
6.68833733e-01 8.38262960e-02 -7.94558823e-02 5.38612381e-02
-1.22873580e+00 -3.07342052e-01 2.78682053e-01 1.28561020e-01
3.33399177e-01 1.77596197e-01 -5.06209552e-01 7.49529064e-01
1.26538321e-01 7.10145012e-02 3.15982848e-01 9.55125570e-01
-8.36805284e-01 -7.11919487e-01 -1.61497131e-01 8.69281471e-01
-3.84079814e-01 1.91327572e-01 -1.60363212e-01 5.94314098e-01
-2.46423170e-01 9.30872619e-01 -7.93647394e-03 -4.60223168e-01
3.73895198e-01 2.03623265e-01 2.84514070e-01 -5.44291437e-01
-9.04246643e-02 7.98142850e-02 -7.28902668e-02 -5.22799015e-01
-2.45622113e-01 -6.99764669e-01 -8.81083608e-01 -5.84388793e-01
-6.52354062e-01 -4.16450165e-02 4.02822942e-01 8.87486458e-01
3.77555758e-01 3.75022054e-01 9.27369595e-01 -4.18015450e-01
-1.05684066e+00 -1.03728676e+00 -6.54852509e-01 3.44413191e-01
1.87477916e-01 -8.47896636e-01 -6.75385952e-01 -1.38113573e-01] | [9.896740913391113, 3.2702767848968506] |
b1853d05-9f4c-4f4f-928f-41ae42d4cb39 | traffic-state-data-imputation-an-efficient | null | null | https://ieeexplore.ieee.org/abstract/document/9582618 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9582618 | Traffic state data imputation: An efficient generating method based on the graph aggregator | Road traffic state estimation is an essential component
of intelligent transportation systems (ITSs). However,
road traffic state data collected by traffic detectors are often
incomplete, which can cause problems across a variety of
transportation applications, such as traffic state prediction and
pattern recognition. We present GA-GAN (Graph Aggregate
Generative Adversarial Network), consisting of graph sample and
aggregate (GraphSAGE) and a generative adversarial network
(GAN), to impute missing road traffic state data. Instead of
using the original road network structure, which presents the
spatial information to process a graph operation, we reconstruct
the road network according to the correlation coefficients of
road historical data. We utilize GraphSAGE to aggregate the
temporal-spatial information from the neighbors of each road
in the reconstructed road network. GAN is used to generate
complete traffic state data from the extracted temporal-spatial
information to achieve traffic state data imputation. To illustrate
the efficient performance of the model, experiments are
conducted on traffic data collected from California and Seattle,
Washington, showing that the proposed model outperforms stateof-the-art
methods. | ['Haijian Li', 'Xuetian Shang', 'Chenchen Wei', 'Hang Peng', 'Dongwei Xu'] | 2022-08-12 | null | null | null | ieee-transactions-on-intelligent-11 | ['traffic-data-imputation'] | ['time-series'] | [ 2.76548773e-01 1.01089664e-01 -3.20024997e-01 -3.80807668e-01
-4.48524684e-01 -2.60644406e-01 5.01103699e-01 -5.36439121e-01
1.48396194e-01 1.06911433e+00 2.52045214e-01 -8.70483160e-01
-2.24907720e-03 -1.44243360e+00 -8.57120097e-01 -7.46760428e-01
2.97573805e-01 4.74984288e-01 4.58586849e-02 -2.37550721e-01
-1.95566878e-01 8.18350554e-01 -1.31757390e+00 -1.88457161e-01
1.40957010e+00 8.55878770e-01 -1.92356899e-01 4.44649577e-01
-3.84623498e-01 9.44602728e-01 -5.15704513e-01 -5.75435340e-01
2.44998023e-01 -4.66089636e-01 -2.06032947e-01 8.49826634e-02
2.42310181e-01 -4.79664952e-01 -1.23638475e+00 1.01650262e+00
1.07031658e-01 3.66572827e-01 6.67907059e-01 -1.81597900e+00
-4.58383799e-01 3.37374121e-01 -4.29053605e-01 -2.28984356e-02
-4.05357871e-03 6.62088215e-01 3.24528456e-01 -9.55877751e-02
4.33760464e-01 1.31566823e+00 4.70293373e-01 5.99123001e-01
-1.07565951e+00 -9.64152813e-01 2.31055737e-01 2.90599853e-01
-1.23606825e+00 -5.07942975e-01 1.11318862e+00 -3.66567671e-01
4.53194320e-01 2.84837246e-01 7.77526438e-01 1.17351222e+00
3.82501811e-01 6.96204782e-01 9.52399015e-01 2.80236512e-01
9.17204916e-02 -2.98251420e-01 4.69869822e-02 7.52010167e-01
1.57481447e-01 3.90899450e-01 -6.08011819e-02 8.31834134e-03
7.45246649e-01 2.40617558e-01 2.19669089e-01 -1.21584773e-01
-8.80269825e-01 4.70523864e-01 7.98576295e-01 -4.52100247e-01
-8.04099560e-01 3.71907532e-01 1.02949783e-01 1.32713705e-01
4.94629622e-01 -5.40388405e-01 1.58462971e-01 -2.04494044e-01
-6.44302785e-01 7.69104362e-02 4.32555765e-01 1.08494377e+00
9.50172782e-01 7.44697511e-01 -3.53141367e-01 5.34834146e-01
2.44704634e-01 1.17392862e+00 -3.40998441e-01 -1.07974637e+00
9.26298618e-01 7.56868601e-01 -4.80840355e-02 -1.12031758e+00
-9.42394957e-02 -1.49858147e-01 -1.58186793e+00 1.90747362e-02
3.58502954e-01 -6.16690934e-01 -1.26799142e+00 1.83410442e+00
3.47444266e-01 8.62195432e-01 3.74489417e-03 4.89082515e-01
6.36365533e-01 9.97094274e-01 2.38958433e-01 -1.70952976e-02
7.00365126e-01 -6.78182721e-01 -8.98001909e-01 -2.12012857e-01
2.89852262e-01 -8.88728201e-02 5.77506006e-01 -2.24646598e-01
-8.38787854e-01 -6.39419436e-01 -5.83142102e-01 2.78655499e-01
-6.69493675e-01 2.79123187e-02 4.93238777e-01 5.20819306e-01
-7.60049284e-01 2.19970137e-01 -8.82660866e-01 -9.23027173e-02
7.42782891e-01 1.52749762e-01 -3.22321355e-01 -3.68254393e-01
-1.45742798e+00 6.82916284e-01 1.14147201e-01 6.46290064e-01
-9.11184788e-01 -6.88214242e-01 -1.09294903e+00 6.79986402e-02
4.14548069e-01 -8.88978839e-01 5.17826378e-01 -3.57993305e-01
-1.25240004e+00 7.11737201e-02 -5.35596132e-01 -4.45333332e-01
6.11073136e-01 4.00590777e-01 -1.04700375e+00 -2.83408016e-01
1.36168897e-01 3.37115467e-01 7.43443191e-01 -1.19596958e+00
-4.92919266e-01 -3.33469063e-01 -3.11906755e-01 -1.39217123e-01
4.14880127e-01 -5.47875643e-01 -4.47325379e-01 -5.03352702e-01
6.26208931e-02 -1.06175470e+00 -4.55762208e-01 -2.33777761e-01
-9.55702603e-01 1.15271948e-01 1.25033593e+00 -9.90278244e-01
1.26750135e+00 -1.92191350e+00 -1.51451319e-01 7.49121845e-01
2.47864619e-01 3.86949420e-01 -3.09644818e-01 3.41431737e-01
-5.15558347e-02 4.82513271e-02 -6.23080254e-01 -1.40047535e-01
3.21479104e-02 6.08551502e-01 -3.62104475e-01 2.46456072e-01
2.17513785e-01 1.36867321e+00 -1.01983535e+00 -2.47378960e-01
5.92940569e-01 4.34613794e-01 -1.18886784e-01 1.87349498e-01
-1.96934298e-01 7.61186957e-01 -8.22216094e-01 4.17332619e-01
8.63423288e-01 1.79453135e-01 -1.20907091e-01 -1.78744525e-01
5.14923446e-02 1.35499462e-01 -1.00242090e+00 1.06091452e+00
-3.82994384e-01 6.28301322e-01 -1.66909203e-01 -9.14988458e-01
1.09136260e+00 -2.75166240e-02 5.85143328e-01 -8.83025348e-01
-1.09842770e-01 -3.00802082e-01 1.68414097e-02 -4.11257654e-01
5.16674161e-01 2.24587426e-01 -3.22984636e-01 4.56638098e-01
-3.96188587e-01 -3.34508643e-02 2.39651725e-01 4.15736556e-01
1.24094343e+00 -1.40954942e-01 -3.17763865e-01 4.39184815e-01
5.06425798e-01 7.18826056e-02 7.08408952e-01 4.61823940e-01
-2.29929492e-01 2.49341026e-01 7.56393790e-01 -5.85934937e-01
-1.03155518e+00 -1.47538745e+00 3.99390101e-01 1.76217303e-01
-2.31355559e-02 1.01568907e-01 -6.35774672e-01 -9.44745958e-01
1.70290262e-01 9.50052917e-01 -6.43945515e-01 -5.87626636e-01
-6.55765235e-01 -5.57067871e-01 6.81433320e-01 5.73000133e-01
9.56245482e-01 -1.05573392e+00 4.84902143e-01 3.50851983e-01
-3.61243129e-01 -1.10885239e+00 -4.74773377e-01 -7.35292971e-01
-5.91254771e-01 -1.07038271e+00 -3.69172186e-01 -2.63756841e-01
9.31805193e-01 2.16149092e-01 8.48491788e-01 -8.23020190e-02
1.49559900e-01 1.72686860e-01 1.48644224e-01 -1.89109981e-01
-8.64241838e-01 2.80207872e-01 -2.02504471e-01 6.02551043e-01
1.71550363e-01 -7.76580453e-01 -3.90846312e-01 4.44819003e-01
-6.05617464e-01 2.80329406e-01 6.55203819e-01 6.45477414e-01
5.91330349e-01 3.64020437e-01 8.19098949e-01 -1.07226431e+00
5.05514205e-01 -8.86026323e-01 -7.27549374e-01 2.94187129e-01
-4.26798373e-01 2.13302784e-02 5.58219850e-01 -1.03016451e-01
-1.23761737e+00 1.29104376e-01 -1.77731991e-01 -5.51710904e-01
-4.21819538e-01 4.93982255e-01 -6.62091434e-01 2.09789097e-01
1.41220018e-01 5.96394718e-01 3.42003912e-01 -1.89773589e-02
4.04983312e-01 5.43070197e-01 8.07503223e-01 -3.34733725e-01
1.08874154e+00 3.17793906e-01 5.43990970e-01 -8.65286291e-01
-5.60194731e-01 -6.92934021e-02 -4.57275093e-01 -6.29539609e-01
8.92365873e-01 -8.17524612e-01 -7.79832661e-01 8.73309314e-01
-1.01675093e+00 -5.22431552e-01 -3.51923138e-01 3.33417952e-01
-5.18535674e-01 1.08707175e-02 -3.27366620e-01 -9.39103305e-01
5.13309985e-02 -9.76368845e-01 7.84250081e-01 2.20295802e-01
4.08828676e-01 -1.06560433e+00 -6.24794997e-02 4.22155619e-01
5.04792750e-01 7.07459152e-01 1.03487122e+00 -1.01012558e-01
-1.06706941e+00 -5.03552735e-01 -3.47222686e-01 3.86694938e-01
4.79506105e-01 2.41817459e-01 -5.73016107e-01 2.48600796e-01
-6.63506150e-01 5.67611933e-01 7.30178952e-01 5.73967636e-01
1.36774480e+00 -6.25320137e-01 -3.88371795e-01 5.38449168e-01
1.01840925e+00 3.89271647e-01 1.33413959e+00 -4.33812112e-01
1.25386429e+00 5.25356352e-01 3.09597373e-01 8.05761144e-02
7.07832277e-01 4.63756382e-01 5.18552423e-01 -2.02240199e-01
-4.68767107e-01 -8.35177481e-01 2.75438458e-01 7.19883025e-01
-1.52726695e-01 -5.78551650e-01 -1.00965166e+00 4.80812758e-01
-1.88020456e+00 -1.42347276e+00 -5.35046220e-01 2.17199230e+00
5.91678545e-02 1.49133489e-01 2.35379159e-01 -3.98376770e-02
1.01271510e+00 4.08830166e-01 -9.73978579e-01 -2.02602848e-01
-8.64162073e-02 -7.70806298e-02 7.81338990e-01 4.86180305e-01
-5.96567452e-01 9.90069211e-01 5.93541098e+00 8.22211981e-01
-8.49381864e-01 -1.26479238e-01 7.33635902e-01 3.73764366e-01
-5.65169334e-01 -2.08264310e-02 -4.59338039e-01 9.27584589e-01
1.16966164e+00 -1.35881305e-01 8.90846014e-01 3.70986789e-01
6.17440999e-01 1.22601360e-01 -5.13004601e-01 6.15698278e-01
-4.41300929e-01 -1.39485359e+00 3.94952089e-01 3.36849064e-01
9.49502230e-01 6.72106892e-02 8.21386874e-02 4.62993234e-01
8.51198733e-01 -8.67415607e-01 9.36076418e-02 1.18014026e+00
8.59668195e-01 -1.02904272e+00 4.79322493e-01 3.35316032e-01
-1.42735291e+00 5.16317599e-02 6.21334948e-02 3.04937571e-01
7.54539430e-01 6.27175033e-01 -6.65134192e-01 8.97068858e-01
1.28515869e-01 1.04478979e+00 -4.25274938e-01 7.72700250e-01
-4.20162022e-01 9.79992032e-01 -2.62113243e-01 2.44136930e-01
1.14366353e-01 -8.65680993e-01 6.54261768e-01 4.67986137e-01
3.67545843e-01 1.16236553e-01 2.66177766e-02 1.04714990e+00
-2.63648301e-01 -6.23125315e-01 -1.07247770e+00 -6.01684023e-03
7.37787783e-01 8.31698418e-01 -8.90450180e-02 -4.67845947e-01
-3.23843688e-01 5.87418437e-01 8.32436234e-03 8.80744457e-01
-1.11146772e+00 -3.08416784e-01 9.36835468e-01 3.36738348e-01
1.77108571e-02 -4.20231253e-01 -1.59649953e-01 -8.50851834e-01
-1.04979575e-01 -3.96672934e-01 7.40433335e-02 -1.06222928e+00
-1.21517634e+00 2.31633797e-01 1.05984285e-01 -1.29259872e+00
-4.06136215e-01 -5.20257875e-02 -1.10039973e+00 1.14995587e+00
-1.42768013e+00 -1.56140769e+00 -5.70160747e-01 7.14951873e-01
1.80805057e-01 -2.79399723e-01 1.77537769e-01 3.31195444e-01
-1.12576854e+00 3.60158324e-01 4.94480878e-02 5.28325081e-01
-4.11238447e-02 -7.69546747e-01 1.20516312e+00 1.20258701e+00
-2.03965411e-01 2.17431724e-01 2.65787184e-01 -1.12895834e+00
-1.59747255e+00 -1.83864164e+00 6.63181484e-01 -2.82916635e-01
6.78245783e-01 -8.95960629e-02 -8.35742414e-01 9.23690915e-01
-2.26010039e-01 3.19521338e-01 1.48556843e-01 -3.77381980e-01
3.18895914e-02 -5.56609094e-01 -1.31741035e+00 8.34416687e-01
1.31939924e+00 -5.42345107e-01 -1.12917498e-01 4.96746562e-02
6.24126136e-01 -3.06458443e-01 -6.21636093e-01 3.77458006e-01
3.68372172e-01 -5.32611072e-01 8.53920400e-01 -6.99116051e-01
1.81921676e-01 -5.19488215e-01 9.76767465e-02 -1.71204150e+00
-2.38856241e-01 -5.13237059e-01 -3.37824017e-01 1.48212898e+00
3.83257717e-01 -1.13416934e+00 9.72517908e-01 9.33992684e-01
-4.16152537e-01 -2.43268520e-01 -1.07835484e+00 -7.61491597e-01
-1.93035215e-01 -5.82880735e-01 1.41251636e+00 5.43900907e-01
-8.68269026e-01 3.37216169e-01 -5.23464620e-01 3.32469553e-01
9.21762049e-01 -1.22760922e-01 1.27840149e+00 -1.08902192e+00
3.62971753e-01 -3.70357215e-01 -5.83914876e-01 -1.03378570e+00
7.33754814e-01 -5.68195462e-01 -4.96193022e-02 -1.78023732e+00
-2.14219987e-01 -7.24701047e-01 -1.80092603e-01 4.69303370e-01
-1.37450814e-01 -1.56107694e-01 -8.08303505e-02 -2.20266327e-01
-9.57617238e-02 9.60311651e-01 1.62638354e+00 -5.00440955e-01
-1.26803875e-01 4.98187989e-01 -3.94880980e-01 2.23175406e-01
8.28118324e-01 -4.22321379e-01 -7.06199706e-01 -2.07861319e-01
-2.24556506e-01 5.32364368e-01 5.83616138e-01 -1.04618204e+00
3.89068127e-01 -5.65005124e-01 1.43734932e-01 -1.10018790e+00
3.62724185e-01 -1.11217034e+00 8.52415144e-01 3.61868888e-01
-5.25965961e-03 1.12114258e-01 2.01579198e-01 9.49563742e-01
-1.44967571e-01 6.46619380e-01 3.22731137e-01 2.76853532e-01
-7.26267278e-01 1.04121828e+00 -3.98884058e-01 -1.09822497e-01
1.08599591e+00 -3.21175307e-01 -5.79517066e-01 -6.91775382e-01
-5.81479669e-01 4.35028672e-01 2.72201568e-01 3.65568280e-01
5.84957242e-01 -1.65752947e+00 -7.17883110e-01 6.48959875e-01
-8.84874910e-02 2.13848457e-01 8.03913653e-01 6.55042708e-01
-1.77197769e-01 1.26492694e-01 -2.13667989e-01 -4.99000877e-01
-5.53226948e-01 5.75883150e-01 2.55144328e-01 -1.50510624e-01
-4.54484165e-01 -1.26117542e-01 -9.48062241e-02 -5.11466146e-01
-3.27574342e-01 -4.05716449e-01 6.14436045e-02 -2.72776723e-01
1.76141635e-01 9.44002807e-01 -1.03839021e-02 -9.11922038e-01
-5.20642428e-03 3.99494857e-01 4.94984090e-01 3.24170850e-02
1.04373360e+00 -3.50530207e-01 1.19643435e-01 1.87099919e-01
1.01981819e+00 -2.11226061e-01 -1.56034124e+00 -1.13319181e-01
-7.36817062e-01 -3.26751918e-01 9.96985137e-02 -6.21062815e-01
-1.80332887e+00 7.34894514e-01 3.67616743e-01 1.64043427e-01
9.84565437e-01 -3.63453150e-01 1.27661288e+00 1.86181217e-01
3.50644022e-01 -6.34287417e-01 -6.54243827e-01 4.61558729e-01
5.57304502e-01 -1.23223829e+00 -4.11604851e-01 -5.19451797e-01
-6.47227705e-01 5.56874633e-01 6.02009714e-01 -1.43671453e-01
8.62353683e-01 6.53082207e-02 -3.05033326e-01 -6.95469677e-02
-5.27141333e-01 -4.70919311e-01 3.09757620e-01 9.68602717e-01
-5.68263054e-01 5.08258283e-01 4.54531789e-01 2.35680103e-01
-1.15919970e-01 5.26039004e-02 3.72608721e-01 6.09167516e-01
2.48439070e-02 -1.02950895e+00 -5.67702986e-02 9.15559292e-01
2.30804369e-01 1.90254241e-01 -1.20668553e-01 7.37635911e-01
-1.64422408e-01 1.03864026e+00 3.43218446e-01 -5.98390520e-01
6.07823193e-01 -8.46003592e-02 -2.11205222e-02 -1.27097428e-01
-4.82890867e-02 -6.94560528e-01 1.87287018e-01 -6.31135881e-01
-1.30629033e-01 -6.71700716e-01 -9.62629676e-01 -1.07733750e+00
1.53619409e-01 9.74567235e-02 7.16907620e-01 9.35119927e-01
7.11649537e-01 8.07498753e-01 1.01799071e+00 -6.03124857e-01
8.14574733e-02 -6.92351341e-01 -5.65972209e-01 5.07937908e-01
5.03620803e-01 -7.65347183e-01 -1.72727123e-01 -3.38252671e-02] | [6.492879390716553, 2.0614700317382812] |
358bddab-f2fc-4e29-8197-0a727e635ab0 | seeing-a-rose-in-five-thousand-ways | 2212.04965 | null | https://arxiv.org/abs/2212.04965v1 | https://arxiv.org/pdf/2212.04965v1.pdf | Seeing a Rose in Five Thousand Ways | What is a rose, visually? A rose comprises its intrinsics, including the distribution of geometry, texture, and material specific to its object category. With knowledge of these intrinsic properties, we may render roses of different sizes and shapes, in different poses, and under different lighting conditions. In this work, we build a generative model that learns to capture such object intrinsics from a single image, such as a photo of a bouquet. Such an image includes multiple instances of an object type. These instances all share the same intrinsics, but appear different due to a combination of variance within these intrinsics and differences in extrinsic factors, such as pose and illumination. Experiments show that our model successfully learns object intrinsics (distribution of geometry, texture, and material) for a wide range of objects, each from a single Internet image. Our method achieves superior results on multiple downstream tasks, including intrinsic image decomposition, shape and image generation, view synthesis, and relighting. | ['Jiajun Wu', 'Noah Snavely', 'Shangzhe Wu', 'Yunzhi Zhang'] | 2022-12-09 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_Seeing_a_Rose_in_Five_Thousand_Ways_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhang_Seeing_a_Rose_in_Five_Thousand_Ways_CVPR_2023_paper.pdf | cvpr-2023-1 | ['intrinsic-image-decomposition'] | ['computer-vision'] | [ 4.04880047e-01 -2.22033963e-01 2.67375708e-01 -2.25460112e-01
-6.06876373e-01 -9.90114450e-01 5.90169489e-01 -1.77916184e-01
4.60296750e-01 3.32636625e-01 1.60615906e-01 4.31062639e-01
1.47386700e-01 -1.03569973e+00 -1.05139184e+00 -7.55747437e-01
3.06681782e-01 6.31829798e-01 9.17939395e-02 -1.59275338e-01
2.73872346e-01 6.73675001e-01 -1.73369193e+00 5.08692682e-01
4.82524455e-01 9.14310038e-01 3.77818197e-01 8.52807641e-01
1.40868932e-01 6.75110161e-01 -7.14848518e-01 -3.63756806e-01
5.15804470e-01 -3.63799930e-01 -5.19882739e-01 6.99846566e-01
9.34593022e-01 -5.37055910e-01 -2.37294897e-01 8.68711531e-01
1.99607253e-01 1.53267622e-01 1.12547600e+00 -1.07103455e+00
-9.11769629e-01 1.40929818e-01 -7.53240824e-01 -5.31115115e-01
4.30701852e-01 5.17229080e-01 1.07937205e+00 -8.68700802e-01
7.29281425e-01 1.22360992e+00 2.70443112e-01 4.19704586e-01
-1.64420307e+00 -3.52508724e-01 5.52607477e-02 -1.97733015e-01
-1.35666645e+00 -3.32109332e-01 8.58739674e-01 -6.54579222e-01
3.51575106e-01 2.19256952e-01 7.46938348e-01 1.11175096e+00
6.82350174e-02 6.06199861e-01 9.56389487e-01 -2.14024276e-01
4.56068635e-01 1.48569763e-01 -5.32181561e-01 5.32555342e-01
2.26780951e-01 -1.11751584e-02 -5.27520776e-01 -6.42869100e-02
1.06413865e+00 -3.54524553e-02 -4.03461099e-01 -8.31518412e-01
-1.36370921e+00 4.46933538e-01 3.42979997e-01 -1.99167073e-01
-1.02241963e-01 1.05340563e-01 -1.55939847e-01 -4.54573445e-02
9.11255628e-02 5.66556931e-01 -5.59953272e-01 1.05048791e-01
-3.15633774e-01 4.30376470e-01 8.62744987e-01 1.53175950e+00
1.07418025e+00 -1.72041892e-03 4.39122692e-03 7.74774730e-01
8.67358670e-02 1.01795030e+00 -1.05676375e-01 -1.31572747e+00
4.45980042e-01 6.07093453e-01 3.17298055e-01 -1.37145782e+00
-1.87063172e-01 -1.85737051e-02 -7.38801956e-01 3.49134505e-01
4.31829482e-01 -3.73146497e-02 -1.07534957e+00 1.71770775e+00
3.58153075e-01 -2.47092545e-01 -2.17434213e-01 8.76429319e-01
1.02508533e+00 6.21000469e-01 -2.77262121e-01 3.94416183e-01
1.42403221e+00 -6.58661485e-01 -1.57669768e-01 -4.16048318e-01
6.17479384e-02 -1.16219139e+00 1.32118821e+00 4.47486460e-01
-1.25460494e+00 -6.33910179e-01 -9.36975777e-01 -3.43184799e-01
-9.77503210e-02 4.91412371e-01 3.78623277e-01 6.52871192e-01
-6.93799138e-01 4.24969673e-01 -4.12514120e-01 -2.25327387e-01
2.06917197e-01 1.73995659e-01 -2.19929904e-01 -2.30873302e-01
-3.04267675e-01 3.67690593e-01 -8.50063004e-03 -7.57026896e-02
-1.10771143e+00 -9.77001846e-01 -7.83836961e-01 1.47460818e-01
4.78352576e-01 -1.03101051e+00 1.05111706e+00 -9.38660562e-01
-1.41944337e+00 8.69960010e-01 -6.90882728e-02 5.11761487e-01
3.50761831e-01 1.10476978e-01 -2.35561915e-02 1.85093641e-01
6.08134158e-02 6.92203224e-01 1.22483814e+00 -1.91631472e+00
-3.02352011e-01 -4.99370247e-01 4.26689565e-01 3.21776241e-01
1.41474679e-01 -3.58741671e-01 -6.01723850e-01 -6.51152492e-01
2.38112748e-01 -1.09155226e+00 -3.21845114e-02 2.78097838e-01
-6.51740313e-01 4.93182629e-01 6.47455573e-01 -4.86002773e-01
4.11791652e-01 -2.24481106e+00 4.63873476e-01 1.98906794e-01
3.24014202e-02 -4.87621605e-01 -3.14244688e-01 1.61406532e-01
2.32586544e-02 2.27188036e-01 -1.38034537e-01 -2.47103050e-01
-1.73273876e-01 1.54288024e-01 -2.12921947e-01 3.00039142e-01
4.32936907e-01 7.18478560e-01 -9.64029193e-01 -2.61716038e-01
3.06652904e-01 6.85403407e-01 -5.90752721e-01 3.65822941e-01
-6.29904270e-01 6.82742655e-01 -4.22829330e-01 6.80141568e-01
8.24147105e-01 -3.18399906e-01 2.96803623e-01 -7.86891282e-01
2.55501628e-01 9.60233063e-03 -1.43604064e+00 1.67051172e+00
-5.27293086e-01 4.39982057e-01 9.04941633e-02 -3.82942468e-01
8.59667361e-01 -2.54000854e-02 5.34873545e-01 -3.58543962e-01
2.14535102e-01 -2.35690307e-02 -1.20370761e-01 -5.57869971e-01
6.25941992e-01 1.46274820e-01 2.06886902e-01 6.09776556e-01
-3.85067500e-02 -7.69221723e-01 1.74778521e-01 4.39938098e-01
9.43657219e-01 2.54525602e-01 -8.06740150e-02 -1.38048247e-01
3.47006321e-02 -2.18352079e-01 3.00977737e-01 4.75467920e-01
4.27474409e-01 1.41187930e+00 5.23552001e-01 -3.93389672e-01
-1.59375012e+00 -1.35273170e+00 -5.42112701e-02 7.53611803e-01
4.13130760e-01 -3.18388045e-01 -5.60088754e-01 -1.65514290e-01
-8.78604501e-02 3.74402225e-01 -7.41892278e-01 -1.04075961e-01
-6.50977969e-01 -6.06517494e-01 -8.29091743e-02 4.85634387e-01
2.15727910e-01 -9.75005031e-01 -6.53031766e-01 -2.60110915e-01
-2.12710679e-01 -1.27963996e+00 -8.81946206e-01 -1.42427132e-01
-6.61179721e-01 -1.21724749e+00 -5.55767477e-01 -5.25869548e-01
1.02312648e+00 6.40097857e-01 1.71362829e+00 -4.11644578e-03
-7.08450079e-01 5.77124834e-01 -2.73534507e-01 -1.81792364e-01
-4.29870695e-01 -6.60725981e-02 -3.77152950e-01 4.43041533e-01
-4.30758685e-01 -6.49432480e-01 -7.93820798e-01 4.03941154e-01
-1.19026279e+00 3.83005291e-01 3.42267752e-01 5.05989730e-01
7.63439834e-01 1.84960946e-01 -2.55737126e-01 -7.01713681e-01
5.96336387e-02 -4.20933366e-01 -4.90846038e-01 4.05919284e-01
-1.18133775e-03 1.32666565e-02 4.40380424e-01 -5.19115508e-01
-1.37833750e+00 2.12720782e-01 6.26805007e-01 -3.19468737e-01
-3.39526385e-01 -2.23229989e-01 -7.43359268e-01 6.02446757e-02
7.61901796e-01 1.07707717e-01 -1.58508644e-01 -4.60423291e-01
5.86303115e-01 1.65884465e-01 3.67919356e-01 -1.07355237e+00
9.10243332e-01 8.04648876e-01 1.22010142e-01 -9.32830989e-01
-8.81939292e-01 -3.22129726e-02 -6.85450733e-01 -1.32617369e-01
7.78312802e-01 -9.39258575e-01 -6.92527711e-01 6.79653227e-01
-1.19133031e+00 -4.71929342e-01 -4.45600957e-01 1.10046618e-01
-7.07549751e-01 -2.41759960e-02 -3.29285651e-01 -5.15796065e-01
6.77373558e-02 -1.25554442e+00 1.48606026e+00 4.20054942e-01
1.75886720e-01 -7.38269925e-01 -2.41940275e-01 4.06556129e-01
1.40935943e-01 3.88667375e-01 1.07744288e+00 3.58507574e-01
-1.28334343e+00 2.39403963e-01 -3.82733971e-01 4.18703169e-01
4.75111604e-01 5.83886683e-01 -7.81457007e-01 -2.96151698e-01
-1.63148537e-01 -3.88730228e-01 3.43650013e-01 6.20329618e-01
1.26110995e+00 -3.76588613e-01 -6.61549019e-03 1.01293361e+00
1.71586537e+00 1.41057089e-01 7.18306482e-01 -3.37312669e-02
1.08812022e+00 7.47673154e-01 2.21131593e-01 5.70684314e-01
4.33171839e-01 8.56252670e-01 6.18509352e-01 -1.97922289e-01
-3.66488993e-01 -2.59310514e-01 2.71553993e-01 2.95543909e-01
-3.63259554e-01 -2.60065287e-01 -6.12184525e-01 4.27409381e-01
-1.35188580e+00 -9.45162237e-01 -1.35699973e-01 2.35979605e+00
7.44802177e-01 -2.99696177e-01 1.56151131e-01 -4.22715306e-01
5.26096344e-01 1.40586838e-01 -8.62407029e-01 -4.72855493e-02
-4.96540397e-01 9.17221159e-02 5.84708869e-01 2.95793504e-01
-6.66567564e-01 5.83014727e-01 7.00777626e+00 5.17119944e-01
-9.67601359e-01 -4.23930675e-01 9.69640672e-01 -2.01002538e-01
-8.08249712e-01 1.23477727e-01 -7.19628870e-01 3.27098280e-01
2.65512079e-01 2.94655338e-02 9.25951779e-01 7.92405427e-01
-1.01565450e-01 -2.01787442e-01 -1.27237940e+00 9.00620818e-01
4.28808868e-01 -1.30153358e+00 2.35981509e-01 8.56671855e-02
1.22569072e+00 -2.46178657e-01 3.28301936e-01 -2.83589244e-01
3.42266917e-01 -1.00919831e+00 9.80496407e-01 4.22069520e-01
9.78297234e-01 -5.30710042e-01 5.02381399e-02 5.74764088e-02
-1.19145298e+00 1.42307401e-01 -1.97668418e-01 3.90875608e-01
-1.69227347e-01 4.37522769e-01 -5.53916991e-01 4.07203406e-01
7.58949399e-01 5.10630548e-01 -5.97648442e-01 5.46664536e-01
-2.54844904e-01 7.48422369e-02 -4.14783448e-01 3.53298157e-01
-5.29293895e-01 -5.13842046e-01 4.17497754e-01 6.13837361e-01
3.80874008e-01 2.51520842e-01 1.19099066e-01 1.46238184e+00
-2.37786889e-01 -2.24683657e-02 -4.61906195e-01 1.55509144e-01
4.79506671e-01 1.34128284e+00 -6.10177040e-01 -2.30451494e-01
-3.57912540e-01 7.90549636e-01 2.74764933e-02 6.31529689e-01
-8.24947417e-01 3.38294767e-02 6.89892054e-01 5.36486566e-01
4.85614628e-01 -2.85367012e-01 -5.10115325e-01 -1.35983133e+00
3.27874660e-01 -1.07814419e+00 -1.45601228e-01 -1.36842728e+00
-1.29346037e+00 4.67049062e-01 2.44007736e-01 -1.41217315e+00
-6.87994361e-02 -7.09017217e-01 -4.36046571e-01 7.47299194e-01
-1.04375947e+00 -1.29204619e+00 -6.82196677e-01 5.92563510e-01
5.54778516e-01 -7.61144087e-02 6.90978527e-01 6.94767684e-02
-2.61240393e-01 1.70907184e-01 4.08061966e-02 8.95912871e-02
6.36064827e-01 -1.18711936e+00 4.06136751e-01 4.79797125e-01
2.95426011e-01 5.41059852e-01 6.43804669e-01 -3.43979776e-01
-1.68722856e+00 -1.07339311e+00 1.00207604e-01 -6.02239907e-01
1.35917112e-01 -4.06597704e-01 -5.79168439e-01 6.55913413e-01
-4.26400155e-02 9.04501602e-03 3.03301454e-01 7.24049807e-02
-7.67965794e-01 -1.40210450e-01 -1.13562942e+00 8.35971534e-01
1.05508423e+00 -3.95865589e-01 4.74716984e-02 5.32958448e-01
4.89222527e-01 -7.60524571e-01 -8.30744863e-01 2.45484352e-01
7.13475347e-01 -1.36382020e+00 1.36637580e+00 -3.17516655e-01
8.93960059e-01 -5.44378698e-01 -3.67042571e-01 -1.41136658e+00
-3.52634847e-01 -4.52844203e-01 -7.56899454e-03 1.11714208e+00
2.28279114e-01 -1.53511003e-01 6.80446982e-01 8.89747918e-01
1.47004128e-01 -6.89816594e-01 -2.94622421e-01 -6.87361062e-01
-1.65890411e-01 -7.56969228e-02 1.09816158e+00 5.60942709e-01
-8.15143704e-01 4.46425050e-01 -4.17276323e-01 2.24884525e-01
7.61831760e-01 6.77501023e-01 1.36781716e+00 -1.12418342e+00
-4.35284615e-01 -1.92633286e-01 -2.44087547e-01 -1.10380781e+00
-3.13487262e-01 -4.82093006e-01 5.29580712e-02 -1.18714440e+00
5.77155232e-01 -7.12276816e-01 4.09298092e-01 3.08820069e-01
-5.31958900e-02 2.63587058e-01 3.95064473e-01 2.61157662e-01
-9.59201381e-02 5.00907958e-01 1.54584527e+00 -2.29693010e-01
-2.82410204e-01 -2.49381408e-01 -7.64975071e-01 7.89565325e-01
6.74234688e-01 -2.95517415e-01 -4.85655010e-01 -1.00107896e+00
3.97428632e-01 7.05704689e-02 4.64780420e-01 -8.14795971e-01
-4.05919433e-01 -6.52540207e-01 6.95394218e-01 -4.44620937e-01
6.83430374e-01 -7.99401879e-01 6.36150062e-01 -1.42051160e-01
-2.17105389e-01 -3.17805558e-02 -9.16837752e-02 5.02504289e-01
2.65619695e-01 1.99655052e-02 8.24015617e-01 -3.76934797e-01
-2.81757832e-01 5.43257952e-01 3.14930439e-01 2.09675193e-01
8.78545940e-01 -4.05304343e-01 -3.66063297e-01 -4.26678061e-01
-4.29800302e-01 -2.30608344e-01 1.12420726e+00 5.95188856e-01
5.58039248e-01 -1.36188328e+00 -7.92966604e-01 5.42055309e-01
2.81000078e-01 5.93267262e-01 4.34155494e-01 1.95943624e-01
-4.55893099e-01 -3.25396746e-01 -3.48418266e-01 -8.65354478e-01
-1.20387411e+00 3.86679024e-01 2.37374380e-01 2.24911988e-01
-6.50671422e-01 6.54646754e-01 9.45008695e-01 -4.11441237e-01
-8.22585523e-02 -3.33298713e-01 3.89087528e-01 -2.12445930e-01
3.90512437e-01 1.26506552e-01 -1.99092589e-02 -7.81284809e-01
-8.93566385e-02 1.03937626e+00 2.29602948e-01 -2.37888936e-02
1.16024029e+00 -6.36385158e-02 2.93859885e-05 4.03471708e-01
1.08057964e+00 2.41152555e-01 -1.63804984e+00 -6.72810301e-02
-8.18314373e-01 -1.09014130e+00 -3.13496470e-01 -6.91759884e-01
-1.57124293e+00 6.46656513e-01 1.62065119e-01 -1.18682915e-02
1.00275016e+00 2.17534095e-01 6.95515215e-01 4.25464883e-02
6.25924706e-01 -9.01511669e-01 4.79910702e-01 2.08001643e-01
1.17647898e+00 -1.17766571e+00 2.12728992e-01 -5.99328279e-01
-6.55049086e-01 1.01424551e+00 6.63111985e-01 -1.80299968e-01
4.41696405e-01 3.55502486e-01 -1.38411233e-02 -2.14882210e-01
-6.90958440e-01 2.62396485e-02 4.55819279e-01 8.47944319e-01
2.42453769e-01 2.73933589e-01 4.24755156e-01 4.87292483e-02
-3.32176745e-01 -4.24750775e-01 5.99601626e-01 6.09372973e-01
-1.25917360e-01 -9.79228675e-01 -5.54166138e-01 2.80713648e-01
-4.15851315e-03 1.27002940e-01 -2.55596042e-01 5.50500631e-01
3.24017614e-01 6.40002608e-01 2.24653319e-01 -2.28246450e-01
4.67043549e-01 -5.88241756e-01 1.03356194e+00 -8.68648529e-01
-1.60725892e-01 5.91212623e-02 -1.98175177e-01 -5.08808911e-01
-4.16849017e-01 -8.06333005e-01 -1.06522369e+00 -2.53771812e-01
-4.52285185e-02 -5.55516481e-01 7.05822527e-01 4.99746472e-01
3.94066364e-01 4.06855315e-01 7.99063385e-01 -1.24965417e+00
-4.44270670e-01 -5.11758626e-01 -8.63764405e-01 9.53210473e-01
3.37166786e-01 -6.31238103e-01 -4.71308082e-01 7.39470959e-01] | [9.260847091674805, -3.1264350414276123] |
26392994-82b2-48d2-bebd-2d747c28a3c8 | machine-learning-with-membership-privacy | 1807.05852 | null | http://arxiv.org/abs/1807.05852v1 | http://arxiv.org/pdf/1807.05852v1.pdf | Machine Learning with Membership Privacy using Adversarial Regularization | Machine learning models leak information about the datasets on which they are
trained. An adversary can build an algorithm to trace the individual members of
a model's training dataset. As a fundamental inference attack, he aims to
distinguish between data points that were part of the model's training set and
any other data points from the same distribution. This is known as the tracing
(and also membership inference) attack. In this paper, we focus on such attacks
against black-box models, where the adversary can only observe the output of
the model, but not its parameters. This is the current setting of machine
learning as a service in the Internet.
We introduce a privacy mechanism to train machine learning models that
provably achieve membership privacy: the model's predictions on its training
data are indistinguishable from its predictions on other data points from the
same distribution. We design a strategic mechanism where the privacy mechanism
anticipates the membership inference attacks. The objective is to train a model
such that not only does it have the minimum prediction error (high utility),
but also it is the most robust model against its corresponding strongest
inference attack (high privacy). We formalize this as a min-max game
optimization problem, and design an adversarial training algorithm that
minimizes the classification loss of the model as well as the maximum gain of
the membership inference attack against it. This strategy, which guarantees
membership privacy (as prediction indistinguishability), acts also as a strong
regularizer and significantly generalizes the model.
We evaluate our privacy mechanism on deep neural networks using different
benchmark datasets. We show that our min-max strategy can mitigate the risk of
membership inference attacks (close to the random guess) with a negligible cost
in terms of the classification error. | ['Amir Houmansadr', 'Milad Nasr', 'Reza Shokri'] | 2018-07-16 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 2.49359176e-01 5.88616848e-01 -2.51946151e-01 -2.62615889e-01
-6.94605708e-01 -1.37396002e+00 3.23318750e-01 -4.64178808e-02
-5.10682166e-01 5.48358560e-01 -6.61952972e-01 -7.04692841e-01
-2.45357230e-01 -1.35277426e+00 -1.22700477e+00 -1.02372992e+00
-1.81463942e-01 6.31018758e-01 -5.66342622e-02 1.37529820e-01
1.65687520e-02 7.10898399e-01 -1.05446494e+00 3.45054358e-01
4.19786662e-01 1.06480312e+00 -3.65234494e-01 7.30028450e-01
4.31020230e-01 6.97815716e-01 -7.82521069e-01 -1.04375005e+00
8.29502165e-01 -3.15788448e-01 -9.27018702e-01 -3.42424780e-01
3.37950528e-01 -5.40213764e-01 -3.70566219e-01 1.53720212e+00
1.24714583e-01 -2.58832097e-01 3.21453601e-01 -1.82865310e+00
-3.93942118e-01 8.77753019e-01 -5.76358624e-02 -3.06424141e-01
1.74016152e-02 2.08798602e-01 1.01622021e+00 1.25737220e-01
5.23893476e-01 7.07243741e-01 4.13439006e-01 9.50674593e-01
-1.51751935e+00 -1.07910788e+00 -2.03977048e-01 -3.17890763e-01
-1.42363572e+00 -3.59308720e-01 5.61202109e-01 -4.22875375e-01
3.25608283e-01 8.43834400e-01 1.47781149e-01 1.33643889e+00
4.99349646e-02 5.32314956e-01 1.19402409e+00 -1.79124430e-01
5.65561354e-01 7.25693345e-01 2.46382058e-01 5.35369158e-01
4.19112533e-01 6.53755605e-01 -1.29772007e-01 -8.60711396e-01
4.28783625e-01 1.17232949e-01 -3.49215031e-01 -5.45699239e-01
-3.57199222e-01 1.01105738e+00 2.40737140e-01 -9.38741416e-02
-9.88340303e-02 1.71904087e-01 1.24294050e-01 8.69814038e-01
-3.89353298e-02 5.20227671e-01 -7.76615202e-01 2.94329196e-01
-3.99802685e-01 4.27347183e-01 1.54208517e+00 8.12555313e-01
8.43109846e-01 -4.21969026e-01 1.22006007e-01 -1.00737456e-02
1.87516093e-01 4.27292764e-01 1.25058696e-01 -1.04637218e+00
4.91528541e-01 2.52064824e-01 1.84209749e-01 -9.14559662e-01
3.41326088e-01 -4.46133554e-01 -7.27515697e-01 7.56267667e-01
6.77969694e-01 -7.12513506e-01 -3.02379102e-01 2.47536969e+00
2.95117021e-01 1.48401797e-01 2.89385408e-01 5.36297739e-01
1.50757536e-01 5.23721039e-01 -4.48940359e-02 -1.25208512e-01
1.10837746e+00 -2.93406844e-01 -1.61543325e-01 -6.94264099e-02
8.95790935e-01 5.58094196e-02 7.56409764e-01 3.44255120e-01
-9.95312989e-01 -3.91282663e-02 -1.13783240e+00 2.81923890e-01
-3.75539362e-01 -3.86856616e-01 6.20887160e-01 1.21435940e+00
-8.70087028e-01 8.06902587e-01 -7.40173161e-01 2.08078191e-01
8.32034528e-01 9.42246437e-01 -8.05587828e-01 2.62403905e-01
-1.35743999e+00 4.19727266e-01 4.88645703e-01 -3.84011149e-01
-1.26505530e+00 -9.50621307e-01 -5.61425269e-01 2.03414500e-01
3.83031815e-01 -7.46045232e-01 1.15903473e+00 -1.48854756e+00
-1.22336042e+00 1.39569116e+00 3.16145867e-01 -1.10957158e+00
1.03323483e+00 3.26156497e-01 -8.46309438e-02 -2.52547145e-01
-4.18250382e-01 1.08462505e-01 9.01691139e-01 -1.32404935e+00
-7.76018679e-01 -7.21252203e-01 7.29776978e-01 -1.39134407e-01
-2.73260832e-01 2.45992164e-03 1.25872701e-01 -1.89630583e-01
-2.77859688e-01 -1.12800837e+00 -3.63443702e-01 2.39285946e-01
-7.70761311e-01 2.20712513e-01 9.68948901e-01 -2.69794673e-01
9.80150580e-01 -2.22075748e+00 -3.97086382e-01 6.41711056e-01
3.10889363e-01 4.36392277e-01 2.34116390e-01 2.71915466e-01
7.76611492e-02 5.79573631e-01 -2.88975149e-01 -5.64370453e-01
3.13993931e-01 4.70056027e-01 -8.05412531e-01 7.58800507e-01
-3.02642941e-01 7.39867806e-01 -3.23814720e-01 1.48595378e-01
-2.70339757e-01 1.15858808e-01 -7.85652220e-01 5.37227273e-01
-5.09662390e-01 4.16447729e-01 -5.10348439e-01 -2.39904243e-02
1.09282410e+00 -2.43325204e-01 3.77024293e-01 3.12539607e-01
4.15529549e-01 2.20060885e-01 -1.17305422e+00 1.04222870e+00
-9.84672382e-02 1.34805247e-01 2.39697412e-01 -7.98362136e-01
6.60593390e-01 4.31202799e-01 3.90934683e-02 -1.73080996e-01
2.04853207e-01 -3.48596238e-02 1.47711888e-01 -1.79259956e-01
-2.20121458e-01 -2.62114167e-01 -2.98816562e-01 8.35624933e-01
-7.06869885e-02 6.19323671e-01 -7.88734496e-01 1.22776337e-01
1.02271140e+00 -4.13771182e-01 1.11981846e-01 -3.60038728e-01
5.07644653e-01 -4.96704936e-01 6.90664768e-01 1.32807803e+00
-1.54348388e-01 1.47767499e-01 1.05936122e+00 -3.14420789e-01
-8.47153604e-01 -1.04083264e+00 2.33252253e-02 9.45651174e-01
3.11054382e-02 -3.29270631e-01 -1.08295858e+00 -1.25186884e+00
3.75205636e-01 6.74478889e-01 -1.03242147e+00 -4.45860982e-01
-1.19444370e-01 -2.72512972e-01 1.01779735e+00 -6.42488599e-02
6.87044322e-01 -7.29991198e-01 -2.44623840e-01 -4.04840648e-01
2.59781420e-01 -7.20215738e-01 -4.74151582e-01 3.10850829e-01
-5.96083164e-01 -1.29011166e+00 1.53129667e-01 -5.64585686e-01
8.72803092e-01 -3.16510469e-01 7.51031220e-01 2.49698535e-01
2.36442387e-01 1.13978788e-01 1.62844092e-01 -6.59495413e-01
-9.00516033e-01 7.79134035e-02 -2.16109082e-02 4.87019479e-01
5.27264118e-01 -6.81153774e-01 -4.77100044e-01 2.49953628e-01
-1.04531574e+00 -2.22383946e-01 -9.79454629e-03 4.01490718e-01
5.58186710e-01 5.09806156e-01 3.15496325e-01 -1.78260958e+00
5.04915118e-01 -6.61689281e-01 -1.14213693e+00 3.64736378e-01
-5.76500058e-01 -2.40226425e-02 1.19317949e+00 -6.19752347e-01
-6.91206455e-01 1.48192495e-01 -1.94887280e-01 -4.93960172e-01
-3.56380194e-01 8.93769935e-02 -8.87016356e-01 -5.14336586e-01
6.41076744e-01 2.77892500e-01 2.06726953e-01 -4.35929418e-01
2.39973560e-01 6.60989285e-01 3.96687776e-01 -9.08898473e-01
1.10483956e+00 5.46035528e-01 5.43301582e-01 -3.72013867e-01
-7.05944121e-01 3.00021589e-01 -3.47254664e-01 2.30429545e-01
5.06198764e-01 -5.18302143e-01 -1.43698967e+00 6.92961633e-01
-9.02409196e-01 -4.58561093e-01 -3.19623619e-01 8.51163790e-02
-5.21604717e-01 2.44315222e-01 -2.83979475e-01 -9.50756967e-01
-3.50712329e-01 -1.10929751e+00 2.52208233e-01 1.66475818e-01
-2.16761399e-02 -1.14324296e+00 -1.53467312e-01 3.72538090e-01
2.62244642e-01 5.82117438e-01 9.50519383e-01 -1.55746579e+00
-8.19715738e-01 -6.40163541e-01 2.96204537e-01 6.69783294e-01
-2.17912048e-01 -4.83434647e-02 -1.29254329e+00 -4.27934021e-01
6.50160789e-01 -2.81170815e-01 4.80520695e-01 -5.19491993e-02
1.69598496e+00 -1.22085869e+00 -5.98399267e-02 1.11881328e+00
1.54106379e+00 2.56339073e-01 5.82346022e-01 1.40057132e-01
4.57796574e-01 4.58444864e-01 8.31087679e-02 4.18260902e-01
-1.40245214e-01 4.92341369e-01 7.17489183e-01 1.52448997e-01
8.81924570e-01 -6.68909132e-01 2.94169843e-01 -2.73535728e-01
3.91216516e-01 -1.26156643e-01 -4.39353555e-01 -3.88567001e-02
-1.66215277e+00 -1.08267879e+00 1.88909799e-01 2.84128618e+00
1.08164048e+00 2.33923018e-01 1.13611467e-01 -3.54224071e-02
2.75359571e-01 -1.19898424e-01 -7.00280249e-01 -8.39420974e-01
-1.32599235e-01 5.13698697e-01 1.00456285e+00 7.22874582e-01
-1.26172042e+00 7.20601797e-01 5.30763292e+00 7.76074827e-01
-9.88328815e-01 2.26958871e-01 1.07361877e+00 -6.19671755e-02
-4.44410741e-01 2.99189985e-01 -7.62403429e-01 6.25664532e-01
1.15515494e+00 -5.58141708e-01 7.97926247e-01 1.26022959e+00
-3.19812566e-01 4.30359691e-01 -1.72058320e+00 5.22741914e-01
-4.06973988e-01 -1.42422163e+00 -1.65484816e-01 5.77393413e-01
5.50861299e-01 -1.42118692e-01 5.63837409e-01 3.39728773e-01
8.22626114e-01 -1.20002365e+00 6.08959198e-01 3.95218104e-01
8.05360675e-01 -1.31367505e+00 4.18097705e-01 1.03903723e+00
-3.75627071e-01 -2.40602046e-01 -3.27605188e-01 -9.33445897e-03
-5.71102858e-01 1.36968851e-01 -4.77146208e-01 2.34488741e-01
2.81505555e-01 -1.63146257e-02 -1.45245418e-01 5.69953978e-01
-3.08778256e-01 7.75056720e-01 -5.59699893e-01 1.97235018e-01
2.30595723e-01 -4.03403103e-01 4.90988582e-01 8.72124672e-01
-1.66892529e-01 3.27598542e-01 2.63486177e-01 1.26825428e+00
-8.65671575e-01 -2.53764182e-01 -8.55912685e-01 8.99250209e-02
6.11729443e-01 8.47199738e-01 1.00850716e-01 -7.31989443e-02
-2.70439163e-02 1.01597404e+00 4.44449097e-01 2.51406491e-01
-7.00841606e-01 -2.91307211e-01 1.35650110e+00 -2.20929063e-03
-1.22180596e-01 3.53299767e-01 -2.51165211e-01 -8.72501910e-01
-4.31966223e-02 -1.02320516e+00 8.02301943e-01 -1.87775776e-01
-1.47740078e+00 2.43195638e-01 -4.56739426e-01 -7.33740866e-01
-6.78492337e-02 -4.71332222e-01 -7.82469988e-01 1.29245758e+00
-1.23016810e+00 -1.05011356e+00 4.54476327e-01 1.05629861e+00
-5.27215838e-01 -1.36176914e-01 1.29313171e+00 -1.55017497e-02
-5.33231199e-01 1.52424872e+00 2.08916485e-01 6.40778542e-01
3.80604677e-02 -1.25385475e+00 2.90926754e-01 9.15341258e-01
2.69782454e-01 1.01022148e+00 4.57374007e-01 -4.13371533e-01
-1.58680749e+00 -1.38619196e+00 7.90916741e-01 -6.36085212e-01
6.14206553e-01 -6.10736549e-01 -9.93929625e-01 1.31277168e+00
-3.26627046e-01 1.50789052e-01 1.03963757e+00 -1.12328477e-01
-7.63470232e-01 -3.86029005e-01 -1.88858080e+00 5.99786580e-01
5.76339424e-01 -9.38776016e-01 -9.13107544e-02 3.51862460e-01
9.05788362e-01 -3.74422640e-01 -7.17287004e-01 -1.96858700e-02
6.47090673e-01 -1.00865126e+00 8.03239346e-01 -1.44412029e+00
5.06414510e-02 -2.53648758e-02 -3.45602959e-01 -7.97093272e-01
-1.92068573e-02 -1.20475173e+00 -5.84059060e-01 1.10701883e+00
5.96632481e-01 -1.14122975e+00 1.35685241e+00 1.16204059e+00
8.88497353e-01 -5.33826530e-01 -1.15069175e+00 -6.68228388e-01
6.74417436e-01 -3.94171387e-01 1.02025986e+00 1.17663538e+00
-2.39943072e-01 -3.71221811e-01 -4.42274719e-01 9.28742170e-01
1.03706348e+00 1.52252123e-01 9.64997649e-01 -1.23907840e+00
-8.91768456e-01 -2.63919801e-01 -3.94953161e-01 -7.61321306e-01
6.11231804e-01 -1.08737838e+00 -4.38814282e-01 -4.28111881e-01
1.43945336e-01 -8.02237630e-01 -4.50479835e-01 5.21736741e-01
1.97218746e-01 -3.34006160e-01 2.90656626e-01 7.85329938e-02
-7.51247406e-02 -1.76104665e-01 7.89845049e-01 4.60517853e-02
-1.03395760e-01 1.00130236e+00 -1.14343572e+00 5.17736554e-01
9.76871371e-01 -9.57012534e-01 -6.70167327e-01 5.13950288e-02
3.31597060e-01 2.71687776e-01 6.31841660e-01 -7.53366351e-01
5.10338724e-01 -2.89804459e-01 4.42993976e-02 3.89631838e-02
1.18838444e-01 -1.36780012e+00 8.69052589e-01 8.34374428e-01
-9.05866504e-01 -4.69454080e-01 -3.33504826e-01 6.18659139e-01
4.12689239e-01 -3.28241140e-01 1.06782722e+00 -1.27595395e-01
8.83945301e-02 9.60715234e-01 -5.63693233e-02 1.34302720e-01
1.40084004e+00 8.89019668e-02 -5.05009711e-01 -4.04985100e-01
-8.99801254e-01 1.10450834e-01 8.66239846e-01 -4.49335948e-02
3.40007663e-01 -8.31539094e-01 -5.14727294e-01 7.07116663e-01
4.52936664e-02 -1.35709524e-01 -8.39592591e-02 3.55300725e-01
-2.90585726e-01 3.33001576e-02 -4.75077219e-02 -3.78358178e-02
-1.57827282e+00 8.66837859e-01 9.12553728e-01 -3.57660919e-01
-3.92079204e-01 9.11910474e-01 3.37206095e-01 -6.70205712e-01
5.88332653e-01 7.01134428e-02 2.87342101e-01 -4.95179981e-01
7.15046704e-01 -2.43398957e-02 -1.20187968e-01 -3.85520935e-01
2.54327839e-04 -1.16666146e-01 -2.58334726e-01 -7.64338076e-02
9.50820684e-01 1.34316131e-01 -2.60459125e-01 -3.15434523e-02
1.69875443e+00 9.84073952e-02 -1.18667459e+00 -4.46216702e-01
-2.02730417e-01 -7.73938835e-01 -1.84222654e-01 -8.46428514e-01
-1.30489933e+00 8.29839826e-01 3.87372464e-01 5.93467593e-01
8.73004436e-01 -1.97970256e-01 5.95907748e-01 4.43362653e-01
7.05242693e-01 -4.52931613e-01 -7.52109587e-01 1.30702019e-01
2.86834598e-01 -8.94445717e-01 -4.38477784e-01 -1.54331610e-01
-6.20966315e-01 7.74063349e-01 3.75747412e-01 -2.45193034e-01
1.00605166e+00 4.05873686e-01 -1.51962340e-01 1.78694800e-01
-8.64093244e-01 4.46665555e-01 -1.97571799e-01 8.45678866e-01
-6.23178005e-01 4.24643546e-01 1.51688114e-01 1.31686294e+00
-2.64280438e-01 -2.07865629e-02 3.97491902e-01 7.43205845e-01
-2.97382414e-01 -1.34234881e+00 -2.27079205e-02 4.35719162e-01
-1.12191832e+00 1.51456982e-01 -6.29681766e-01 4.87004131e-01
2.19030440e-01 7.40585268e-01 -2.12446615e-01 -6.47759557e-01
1.83988273e-01 -5.43387868e-02 6.72817649e-03 -4.11804855e-01
-1.02722204e+00 -7.08166599e-01 -1.70933492e-02 -6.65492594e-01
2.94329971e-01 -3.51578802e-01 -9.83157992e-01 -1.06592393e+00
-2.31714785e-01 4.34785962e-01 5.91015518e-01 7.21630275e-01
3.07636559e-01 -2.22614750e-01 1.50639462e+00 1.54792771e-01
-1.29255891e+00 -2.12822273e-01 -9.58001912e-01 5.82049668e-01
4.45005149e-01 2.52821565e-01 -7.98710644e-01 -1.29683360e-01] | [5.93508243560791, 7.1097187995910645] |
f239f221-3802-4d92-9e62-33551c50e633 | perspective-melanoma-diagnosis-and-monitoring | 1607.06720 | null | http://arxiv.org/abs/1607.06720v2 | http://arxiv.org/pdf/1607.06720v2.pdf | Perspective: Melanoma diagnosis and monitoring: Sunrise for melanoma therapy but early detection remains in the shade | Melanoma is one of the most dangerous forms of cancer. The five-year survival
rate is 98% if it is detected early. However, this rate plummets to 63% for
regional disease and 17% when tumors have metastasized, that is, spread to
distant sites. Furthermore, the incidence of melanoma has been rising by about
3% per year, whereas the incidence of cancers that are more common is
decreasing. A handful of targeted therapies have recently become available that
have finally shown real promise for treatment, but for reasons that remain
unclear only a fraction of patients respond long term. These drugs often
increase survival by only a few months in metastatic patient groups before
relapse occurs. More effective treatment may be possible if a diagnosis can be
made when the tumor burden is still low. Here, an overview of the current
state-of-the-art is provided along with an argument for newer technologies
towards early point-of-care diagnosis of melanoma. | [] | 2016-07-25 | null | null | null | null | ['melanoma-diagnosis'] | ['computer-vision'] | [ 4.86066163e-01 -4.00206372e-02 -5.89047134e-01 3.38484049e-01
-9.94626045e-01 -4.19051737e-01 5.48781574e-01 1.10433483e+00
-8.98419797e-01 1.00592279e+00 4.14066901e-03 -4.76201594e-01
2.44160779e-02 -6.51716948e-01 7.70108402e-02 -1.20596611e+00
2.13721275e-01 4.32333827e-01 3.49049985e-01 -1.22730337e-01
9.02256072e-02 6.96657181e-01 -1.11007023e+00 -9.35272649e-02
1.04507697e+00 6.04396641e-01 4.09087300e-01 7.34777451e-01
-2.45631516e-01 3.04365695e-01 -3.70649159e-01 -9.96383876e-02
-1.90021649e-01 -7.05089509e-01 -7.76472986e-01 9.63166356e-02
-2.83592433e-01 -1.71820462e-01 1.18442416e-01 7.47786283e-01
2.35276967e-01 -4.44667906e-01 6.84724331e-01 -6.25403821e-01
2.62512237e-01 -1.23871364e-01 -8.62760067e-01 -2.82290764e-02
2.03408316e-01 1.81302443e-01 4.21021640e-01 -4.84715700e-01
7.38616467e-01 3.71463209e-01 4.99001533e-01 1.05602884e+00
-1.33989334e+00 -1.64469816e-02 -1.22240193e-01 -1.47964850e-01
-1.34811127e+00 -1.59412637e-01 -1.32071108e-01 -4.08725560e-01
7.25739658e-01 6.93120599e-01 1.09779811e+00 6.24215782e-01
7.98972428e-01 4.79549430e-02 1.26306391e+00 -4.15944099e-01
6.88671470e-02 2.13277742e-01 -2.19651178e-01 4.94606525e-01
4.94545072e-01 -5.92044666e-02 -4.03328896e-01 -5.87388575e-02
5.22129476e-01 9.55205411e-02 -5.45825243e-01 -9.64335129e-02
-1.04189062e+00 7.82291651e-01 2.48997658e-01 8.06037128e-01
-4.22381222e-01 -1.46564454e-01 5.13521433e-01 7.92748556e-02
3.95911098e-01 -6.31715581e-02 -1.83870107e-01 -2.57649779e-01
-7.69666612e-01 -1.18932910e-01 5.25112748e-01 -1.83504209e-01
1.76948413e-01 -4.95819241e-01 1.65586397e-01 7.20021486e-01
3.37147973e-02 4.14559394e-01 5.58301866e-01 -8.01723972e-02
-2.65508473e-01 8.09572518e-01 2.52136827e-01 -4.15534973e-01
-6.73112452e-01 -6.14762068e-01 -9.73382413e-01 4.86809850e-01
9.71341729e-01 -1.21114738e-01 -7.44664788e-01 1.34716225e+00
1.90244779e-01 -3.02941620e-01 6.21286221e-02 8.98033023e-01
3.45632374e-01 4.24650788e-01 4.85221386e-01 -4.01074648e-01
1.50791514e+00 -3.83251965e-01 -3.16874385e-01 -2.39089370e-01
8.28864396e-01 -6.95878744e-01 6.70763254e-01 6.19372666e-01
-8.35958481e-01 2.94655591e-01 -9.54460323e-01 4.15802091e-01
-2.11090952e-01 -9.54999179e-02 6.93897307e-01 1.09217358e+00
-9.94633675e-01 3.02002311e-01 -1.04587209e+00 -1.35933900e+00
2.75615543e-01 3.69211257e-01 -6.66253448e-01 -2.76929468e-01
-8.58750105e-01 1.10769880e+00 8.95891711e-02 -1.66121244e-01
-7.54357040e-01 -6.97130561e-01 -2.95767307e-01 -4.24346119e-01
4.89850119e-02 -1.02560687e+00 7.86534071e-01 -6.51980162e-01
-1.46832669e+00 1.11745071e+00 -5.52891374e-01 -3.18537712e-01
4.68092144e-01 4.30826694e-01 -4.23119962e-01 3.65429133e-01
-1.22035369e-01 3.68837863e-01 4.37569261e-01 -7.02161789e-01
-1.23386264e+00 -3.62556726e-01 -3.17350775e-02 3.42443675e-01
-3.17691594e-01 3.01370710e-01 -1.06020644e-01 3.62960175e-02
-1.98295951e-01 -8.81957591e-01 -6.63223565e-01 3.59754503e-01
-2.45622367e-01 1.08657479e-01 4.59056348e-01 -4.58436847e-01
1.01152146e+00 -1.92710209e+00 1.40338942e-01 1.95455668e-03
1.58152118e-01 1.67086363e-01 1.08570330e-01 7.47075617e-01
1.02970339e-01 4.46823686e-01 -5.46512604e-02 1.98794186e-01
-5.34540772e-01 -6.21726438e-02 3.60455364e-01 7.46907115e-01
3.27520430e-01 6.11658990e-01 -9.75881457e-01 -6.53463528e-02
2.50444889e-01 6.37200892e-01 1.25910446e-01 -5.00412047e-01
2.37913914e-02 7.43145227e-01 -1.84334397e-01 6.79459810e-01
5.03154159e-01 -4.34076101e-01 3.23155731e-01 4.82718349e-01
-3.81996751e-01 2.24255770e-01 -6.20503962e-01 1.36865604e+00
-2.85198241e-01 6.43926024e-01 1.89222582e-02 -4.96750861e-01
4.15014267e-01 4.33127999e-01 4.22014743e-01 -2.91404039e-01
4.85667624e-02 4.98093903e-01 4.76041883e-01 -2.24654064e-01
4.05089073e-02 -8.31843317e-01 1.92914978e-01 -5.05723655e-02
-7.91390002e-01 1.41070589e-01 3.81916434e-01 3.07804160e-02
1.41535306e+00 -3.98642510e-01 7.14715064e-01 -1.46468133e-01
8.47075462e-01 5.73709548e-01 4.68119413e-01 3.53542417e-01
-3.47134136e-02 4.58469599e-01 7.07795620e-01 -1.27640396e-01
-5.21596551e-01 -1.04481435e+00 -6.98497772e-01 2.09282830e-01
9.12170336e-02 -2.58672237e-01 -5.08333802e-01 -4.61322278e-01
-8.35239887e-02 3.54185075e-01 -6.99091375e-01 6.00053258e-02
5.63286208e-02 -1.14204621e+00 3.19721848e-01 -6.28724508e-03
3.94908696e-01 -3.75855155e-02 -5.12945414e-01 4.27567959e-01
1.83512107e-01 -5.64179003e-01 4.23584938e-01 1.18421890e-01
-9.35241401e-01 -1.26253879e+00 -1.43271542e+00 -6.18398964e-01
1.05645216e+00 3.80697310e-01 5.54380298e-01 5.63500345e-01
-6.01078391e-01 7.22877309e-02 -1.19968712e-01 -3.70194912e-01
-5.04121900e-01 -8.76219720e-02 -7.17322826e-02 -8.81831795e-02
3.43578607e-01 -1.10987015e-01 -7.08870530e-01 2.89042562e-01
-5.78974903e-01 -3.74874808e-02 8.48399103e-01 8.24899912e-01
7.10462332e-01 1.01238422e-01 5.67477465e-01 -9.48559940e-01
2.93571383e-01 -3.62454265e-01 -2.81099260e-01 -1.43758640e-01
-4.97834593e-01 -5.55968523e-01 2.85118192e-01 -1.65295243e-01
-9.98095393e-01 -1.22637950e-01 -1.88604161e-01 4.85905081e-01
-6.06451392e-01 8.60173821e-01 3.84262085e-01 -3.70032996e-01
8.40146720e-01 8.06291327e-02 3.87884468e-01 -2.34372672e-02
-4.50437963e-01 6.13254547e-01 2.58711427e-02 2.63506114e-01
7.66975820e-01 8.40365529e-01 6.21558011e-01 -1.39908123e+00
-6.56808197e-01 -5.48481464e-01 -1.29530981e-01 -3.80024880e-01
8.35216284e-01 -7.36167610e-01 -6.57058060e-01 6.59518778e-01
-6.89305782e-01 -4.07116354e-01 -2.14417458e-01 4.56259847e-01
-1.53946400e-01 2.84402132e-01 -6.43482447e-01 -7.54692137e-01
-1.51193887e-01 -9.70577419e-01 7.37083554e-01 4.55686510e-01
-6.21536911e-01 -1.33044732e+00 1.50207490e-01 1.35807440e-01
6.17932558e-01 6.73753977e-01 1.08389354e+00 -1.92793280e-01
-4.78855759e-01 -7.27597654e-01 5.30489720e-02 -2.58582324e-01
6.13268256e-01 2.22072035e-01 -5.70658684e-01 -4.50982362e-01
-4.67013896e-01 -3.28288786e-02 6.91657424e-01 4.55260724e-01
3.39554101e-01 3.15588653e-01 -1.22906744e+00 1.79693446e-01
1.68319082e+00 3.80950898e-01 6.02795839e-01 4.42660093e-01
-1.92039497e-02 5.21112144e-01 7.49492168e-01 -2.71409135e-02
-8.24186951e-02 2.68540621e-01 6.00240707e-01 -1.26481593e-01
-2.97939509e-01 8.18847120e-02 -9.41924453e-02 1.87351286e-01
-3.12931389e-01 -3.58229876e-01 -1.02060080e+00 5.47637939e-01
-9.84287620e-01 -8.13367307e-01 -7.15286314e-01 2.66217518e+00
6.66632771e-01 4.32879001e-01 3.31781745e-01 1.54021844e-01
6.26313150e-01 -2.95336902e-01 -3.17158878e-01 -2.19298601e-01
-1.84273720e-01 -1.11438125e-01 4.29730415e-01 6.95002079e-01
-5.21151602e-01 4.75318402e-01 6.70949650e+00 6.09858274e-01
-1.65350974e+00 -5.98059259e-02 6.69403851e-01 -7.77539983e-02
-7.66513050e-02 5.19949675e-01 -8.74615431e-01 3.19854915e-01
8.31167638e-01 -2.95292199e-01 -4.36458528e-01 2.06687689e-01
3.49068612e-01 -1.15928304e+00 -6.76586866e-01 8.86727512e-01
-7.51890242e-02 -1.36827123e+00 -4.46358830e-01 5.76818883e-01
3.82394195e-01 -1.80110395e-01 6.74849227e-02 -2.51977742e-01
-4.87761855e-01 -9.27952588e-01 -2.01081485e-01 6.92109048e-01
1.24145913e+00 -9.24549162e-01 1.01387084e+00 6.66781962e-01
-5.68453133e-01 1.46237671e-01 -1.03258945e-01 -1.80698216e-01
1.32935137e-01 8.28465044e-01 -1.04868650e+00 4.69768912e-01
1.99663386e-01 2.94525951e-01 -6.17167234e-01 1.67672455e+00
-3.89381677e-01 4.66309339e-01 -5.78991532e-01 -5.26102006e-01
3.77360098e-02 -2.73656603e-02 5.72629154e-01 5.14070034e-01
4.02693361e-01 9.26678702e-02 -3.47439945e-01 1.23314202e-01
5.47035038e-01 4.11700457e-01 -3.69664252e-01 -1.46672398e-01
2.07172960e-01 1.37805903e+00 -9.85951066e-01 1.66550517e-01
-5.46992779e-01 1.17869020e+00 5.10918498e-02 9.37429667e-02
-2.83068538e-01 -4.20273542e-01 3.02851051e-01 6.22824192e-01
-2.72776604e-01 1.97720125e-01 4.49840259e-03 -6.56971514e-01
-3.79122764e-01 -5.35141289e-01 4.30997223e-01 -3.43832314e-01
-7.57426739e-01 4.40280974e-01 -2.50954360e-01 -1.15078759e+00
-4.73960489e-01 -7.18006015e-01 -8.33121955e-01 1.06155097e+00
-1.26596999e+00 -9.31202948e-01 -3.90770316e-01 -1.12902530e-01
3.07966918e-01 1.59069225e-01 1.48895764e+00 -4.59113866e-01
-5.15806258e-01 4.51861441e-01 2.84433454e-01 -3.60224664e-01
8.03828418e-01 -1.02013016e+00 -3.95827919e-01 5.51557481e-01
-7.15565741e-01 4.67363954e-01 8.51325452e-01 -6.44522667e-01
-1.15461791e+00 -4.50533688e-01 1.11980462e+00 3.04688904e-02
8.04639518e-01 -3.63606773e-02 -5.11981130e-01 1.14289083e-01
3.75749499e-01 -3.75190765e-01 1.12512076e+00 2.21775621e-01
1.72967598e-01 -1.31780997e-01 -1.29811466e+00 9.31070447e-01
2.51709670e-01 -1.20777100e-01 2.14519426e-01 5.61034858e-01
-3.21813166e-01 -4.54729944e-01 -9.94647503e-01 1.38809279e-01
5.70488930e-01 -1.05241406e+00 4.12074536e-01 -1.48814678e-01
1.01293534e-01 -4.13178921e-01 3.66158396e-01 -1.41829777e+00
-4.34879124e-01 -6.24049306e-01 6.06045067e-01 1.05129623e+00
7.34012902e-01 -9.23718870e-01 1.16033304e+00 6.13350511e-01
2.82725751e-01 -1.15145719e+00 -1.00900888e+00 -6.62907839e-01
1.43076405e-01 2.92041987e-01 -7.06752986e-02 5.38705945e-01
6.05060816e-01 2.93862075e-01 2.47556299e-01 -1.91734001e-01
4.68217492e-01 -2.70159274e-01 5.24776757e-01 -1.40368748e+00
-6.38049021e-02 -5.42682290e-01 -7.58714795e-01 -4.56006944e-01
-4.44894850e-01 -7.78111756e-01 -7.74660408e-01 -1.86866558e+00
4.34116662e-01 -9.41439420e-02 1.08207107e-01 3.47107559e-01
-7.22718090e-02 2.78233826e-01 -1.76191017e-01 1.71355009e-02
2.12214217e-01 -2.76506841e-01 1.16937900e+00 -7.59069063e-03
-2.07826301e-01 1.89972416e-01 -7.88453400e-01 8.06644201e-01
1.03850305e+00 -3.48438323e-01 -2.60007471e-01 1.60811082e-01
1.91263065e-01 2.64095247e-01 6.21157549e-02 -9.52466011e-01
3.29290330e-01 -4.53384429e-01 4.08981651e-01 -6.40946388e-01
4.32054639e-01 -8.80644023e-01 6.37518406e-01 1.17311645e+00
2.19335526e-01 -3.66548806e-01 1.30938426e-01 6.54686034e-01
-3.49561274e-02 -4.78268087e-01 1.15542936e+00 7.58912414e-02
-7.00466275e-01 2.01205648e-02 -1.34838533e+00 -3.82544965e-01
1.55829108e+00 -5.02367973e-01 -5.39128184e-01 -4.42596048e-01
-9.23716486e-01 -5.23032201e-03 9.77499187e-01 -3.29710394e-01
4.64674562e-01 -8.98319304e-01 -7.01867640e-01 -3.59577090e-01
3.41259748e-01 -1.19999051e-01 5.79947114e-01 1.79761374e+00
-9.10229445e-01 6.14389062e-01 -8.26068670e-02 -6.91073477e-01
-1.60277557e+00 4.49109256e-01 6.03184938e-01 -3.36458564e-01
-6.81551993e-01 1.02412784e+00 2.70963162e-02 4.61144120e-01
-9.57552567e-02 3.95550102e-01 -4.72460352e-02 -2.89769787e-02
8.95024896e-01 3.70665401e-01 1.43079935e-02 -4.70032215e-01
-5.58321476e-01 4.57936108e-01 -6.64418817e-01 -9.21815485e-02
9.83953893e-01 -1.02610171e-01 -4.08526570e-01 4.83908534e-01
6.48515046e-01 3.09130222e-01 -9.30461049e-01 4.63178694e-01
-2.85263509e-01 -5.78649104e-01 5.26541956e-02 -1.09706092e+00
-6.24748588e-01 6.84934556e-01 5.42229056e-01 1.95796594e-01
1.32659614e+00 3.61507200e-02 4.78814483e-01 -3.12504619e-01
5.48148036e-01 -6.74151480e-01 -3.11760515e-01 2.10737556e-01
5.23732007e-01 -7.97976255e-01 1.50524706e-01 -7.76592612e-01
-1.96247265e-01 1.00680172e+00 1.23782940e-01 -9.76701230e-02
4.33112293e-01 4.50853467e-01 3.76762956e-01 7.08854049e-02
-9.00043547e-01 -4.36759740e-01 -2.17735291e-01 8.97240222e-01
6.18854284e-01 1.10406369e-01 -1.05567741e+00 2.54016351e-02
6.55954257e-02 -1.44549951e-01 7.96134830e-01 8.97134244e-01
-7.46079862e-01 -1.67070651e+00 -3.93555760e-01 5.82568228e-01
-8.62900674e-01 6.20885566e-02 -6.78070068e-01 1.38814449e+00
-2.21877873e-01 9.52906787e-01 -5.06162504e-03 1.82933509e-01
1.38407737e-01 2.26528972e-01 6.83630586e-01 -3.87226015e-01
-3.30460280e-01 5.17687321e-01 4.91747797e-01 -1.12599090e-01
-4.96108055e-01 -9.78324294e-01 -1.34073710e+00 -2.21031174e-01
-3.89321923e-01 1.69550106e-01 7.40140319e-01 8.92342687e-01
-2.34633133e-01 3.04153770e-01 2.74300903e-01 -1.41958892e-01
-1.22953758e-01 -5.05748153e-01 -1.16485393e+00 -3.30739617e-01
2.89284945e-01 -4.78162229e-01 -4.65432137e-01 -1.76323593e-01] | [15.174018859863281, -3.0888946056365967] |
070f0b8e-8f6d-4567-a2fb-a918baccf1f4 | l3dor-lifelong-3d-object-recognition | 1912.06135 | null | https://arxiv.org/abs/1912.06135v2 | https://arxiv.org/pdf/1912.06135v2.pdf | L3DOC: Lifelong 3D Object Classification | 3D object classification has been widely-applied into both academic and industrial scenarios. However, most state-of-the-art algorithms are facing with a fixed 3D object classification task set, which cannot well tackle the new coming data with incremental tasks as human ourselves. Meanwhile, the performance of most state-of-the-art lifelong learning models can be deteriorated easily on previously learned classification tasks, due to the existing of unordered, large-scale, and irregular 3D geometry data. To address this challenge, in this paper, we propose a Lifelong 3D Object Classification (i.e., L3DOC) framewor, which can consecutively learn new 3D object classification tasks via imitating 'human learning'. Specifically, the core idea of our proposed L3DOC model is to factorize PointNet in a perspective of lifelong learning, while capturing and storing the shared point-knowledge in a perspective of layer-wise tensor factorization architecture. To further transfer the task-specific knowledge from previous tasks to the new coming classification task, a memory attention mechanism is proposed to connect the current task with relevant previously tasks, which can effectively prevent catastrophic forgetting via soft-transferring previous knowledge. To our best knowledge, this is the first work about using lifelong learning to handle 3D object classification task without model fine-tuning or retraining. Furthermore, our L3DOC model can also be extended to other backbone network (e.g., PointNet++). To the end, comparisons on several point cloud datasets validate that our L3DOC model can reduce averaged 1.68~3.36 times parameters for the overall model, without sacrificing classification accuracy of each task. | ['Yang Cong', 'Gan Sun', 'Yuyang Liu'] | 2019-12-12 | null | null | null | null | ['3d-object-classification', '3d-object-recognition'] | ['computer-vision', 'computer-vision'] | [-3.70214075e-01 -3.20264280e-01 -2.34889984e-03 -2.56560892e-01
7.55422842e-03 -3.72363955e-01 3.04041773e-01 -6.50471300e-02
-2.32325286e-01 3.45517755e-01 -2.00585529e-01 -2.54416406e-01
-2.37767383e-01 -7.92397439e-01 -1.07290363e+00 -6.33341312e-01
-2.34470498e-02 4.80017900e-01 5.53125799e-01 -2.63138413e-01
3.50038022e-01 8.96798313e-01 -1.69912338e+00 2.82453388e-01
9.61946428e-01 1.19237721e+00 5.73316753e-01 2.88730025e-01
-4.19698089e-01 5.71792245e-01 -3.97501409e-01 -3.50490868e-01
1.66126400e-01 5.28037965e-01 -1.13761449e+00 9.15752053e-02
5.59120119e-01 -4.28994089e-01 -4.48858321e-01 7.37035930e-01
4.31281000e-01 2.43557766e-01 2.60464877e-01 -1.35864639e+00
-1.07206917e+00 1.19438194e-01 -3.05661291e-01 3.41228813e-01
-3.96173857e-02 3.53448421e-01 4.16089386e-01 -1.40022326e+00
3.76588494e-01 1.34303021e+00 8.21213484e-01 4.02163148e-01
-9.14716005e-01 -8.43547583e-01 8.20596457e-01 4.28007334e-01
-1.34774506e+00 -5.85584156e-02 9.37865436e-01 -4.88301516e-01
1.18346870e+00 1.19960923e-02 9.50588465e-01 9.42869782e-01
1.96760163e-01 9.11463916e-01 7.24596441e-01 1.34278923e-01
8.32015201e-02 -1.09202132e-01 4.77369905e-01 5.58487773e-01
2.29238525e-01 1.90365259e-02 -6.12705469e-01 5.75638860e-02
8.12331021e-01 6.91255689e-01 -4.78427000e-02 -5.94068170e-01
-1.37852883e+00 2.67024159e-01 9.64875519e-01 2.71646261e-01
-2.89380193e-01 1.74693286e-01 3.84469122e-01 4.63438183e-01
8.14725876e-01 1.71855539e-01 -9.12940800e-01 -4.17165272e-02
-5.54101169e-01 1.84685975e-01 5.25494397e-01 1.34562814e+00
1.11440504e+00 -5.27045801e-02 5.60257994e-02 7.30011702e-01
3.12482715e-01 6.07337713e-01 3.68141949e-01 -6.00897968e-01
4.17479068e-01 9.21766937e-01 -6.69563338e-02 -1.03765523e+00
-5.99632263e-01 -9.15286541e-01 -1.03192663e+00 1.09414816e-01
-5.37924748e-03 2.98014671e-01 -1.08304441e+00 1.48718333e+00
5.81876278e-01 5.68797171e-01 -1.23901509e-01 9.15102541e-01
8.12655568e-01 8.26929271e-01 -2.65564919e-02 -2.17565782e-02
1.12987673e+00 -1.10345221e+00 -1.91098705e-01 -8.93714428e-02
7.01941192e-01 -7.22483218e-01 1.29369557e+00 3.66430819e-01
-8.11890721e-01 -1.19391108e+00 -9.42573011e-01 -4.42668706e-01
-5.55941164e-01 -9.11379829e-02 8.51866782e-01 2.36152604e-01
-7.89836586e-01 5.41009068e-01 -1.06704044e+00 -3.99104029e-01
6.15764022e-01 4.31259245e-01 -4.11807716e-01 -4.39174950e-01
-8.96273911e-01 8.28256011e-01 4.91488039e-01 3.72934759e-01
-9.02050674e-01 -1.21447361e+00 -3.60659987e-01 -1.72448516e-01
4.74328846e-01 -1.03241026e+00 1.17671192e+00 -5.38761079e-01
-1.21740294e+00 6.81000829e-01 -1.29370093e-01 -1.30672082e-01
1.99162498e-01 -6.87270522e-01 -2.54572630e-01 -1.24601848e-01
1.28025100e-01 7.79082119e-01 1.12108457e+00 -1.41160572e+00
-5.78435779e-01 -7.55091190e-01 4.34195697e-01 2.76959240e-01
-4.86178130e-01 -5.84220767e-01 -6.81592286e-01 -7.80281484e-01
5.64538479e-01 -1.01919174e+00 -2.32901983e-02 4.90479797e-01
1.92282751e-01 -6.01842403e-01 1.23646963e+00 -2.33531564e-01
9.13751781e-01 -2.39292598e+00 3.74335498e-01 -1.98435053e-01
3.82943600e-01 3.95901442e-01 -2.10754439e-01 6.29992485e-02
-9.22928974e-02 -7.98434671e-03 1.13360412e-01 -4.44788516e-01
-1.10981151e-01 4.60674733e-01 -5.39124548e-01 2.08774880e-01
2.34047905e-01 1.08814394e+00 -9.25674379e-01 -1.63558915e-01
3.59289318e-01 5.37246704e-01 -6.61966980e-01 1.85796589e-01
-3.97759527e-01 4.04038876e-01 -4.57591534e-01 7.65079260e-01
1.02814436e+00 -5.20039082e-01 -4.56706613e-01 -6.00900590e-01
-7.73658231e-02 7.02258646e-02 -9.30032313e-01 2.46175170e+00
-6.52229369e-01 1.45105133e-02 -1.71677381e-01 -1.08205628e+00
1.02308774e+00 8.30535404e-03 5.76419294e-01 -5.18232942e-01
-1.45719245e-01 2.11943269e-01 -1.77979156e-01 -4.87661123e-01
4.66184467e-01 -3.87000963e-02 -2.38913391e-02 4.02983487e-01
3.38511080e-01 -2.84719802e-02 -2.76009619e-01 1.58350825e-01
7.63916194e-01 3.27855945e-01 -3.45576018e-01 -1.70185357e-01
5.94106019e-01 1.31588608e-01 5.43797195e-01 7.26866007e-01
-6.86944202e-02 3.90414238e-01 -8.03273320e-02 -1.15407050e+00
-8.78363431e-01 -9.58719432e-01 -1.56734779e-01 1.12231946e+00
4.34848994e-01 -2.86995590e-01 -2.31194705e-01 -9.12717462e-01
4.31730211e-01 4.97657120e-01 -4.18228269e-01 -6.02323771e-01
-7.23078787e-01 -6.90527797e-01 1.91212997e-01 5.83813310e-01
7.16807246e-01 -7.84170389e-01 -2.55533338e-01 3.44591260e-01
3.47758755e-02 -9.84674394e-01 -3.14783216e-01 4.73790728e-02
-1.40795684e+00 -9.42414403e-01 -6.68189406e-01 -9.44425404e-01
6.91573024e-01 1.00301075e+00 1.08878851e+00 3.61597210e-01
2.01460887e-02 5.47634065e-01 -6.01633370e-01 -3.34900886e-01
1.45126238e-01 6.87751174e-01 3.95371616e-01 -4.28614998e-03
5.54144740e-01 -9.18812633e-01 -7.04797149e-01 5.13641596e-01
-9.34499562e-01 3.14895838e-01 7.59295762e-01 7.53765404e-01
6.53621614e-01 1.46049961e-01 5.67378819e-01 -6.34391427e-01
5.81062809e-02 -5.44271529e-01 -3.16037059e-01 2.20094904e-01
-6.50364995e-01 7.12704519e-03 5.36307335e-01 -6.79150701e-01
-9.34177101e-01 -1.39472365e-01 -2.16335028e-01 -1.04541099e+00
-1.01225846e-01 4.34609205e-01 -2.82148898e-01 -4.61347044e-01
3.15561831e-01 3.24979931e-01 -5.12199150e-03 -9.44302619e-01
3.38853836e-01 4.52250421e-01 3.05975765e-01 -8.67469728e-01
1.05001283e+00 4.96962190e-01 9.73460451e-02 -4.78038996e-01
-1.28669178e+00 -3.54735106e-01 -1.05436432e+00 -2.17908174e-01
4.91993904e-01 -1.05265462e+00 -8.03300381e-01 9.98501480e-01
-1.44997787e+00 -2.37991989e-01 -3.54148656e-01 3.52860034e-01
-2.96369016e-01 1.98862448e-01 -4.57851410e-01 -3.74430507e-01
-2.38670573e-01 -1.09807014e+00 1.11895335e+00 -5.94340041e-02
4.37108606e-01 -7.61754692e-01 -2.60103524e-01 3.84844124e-01
4.24753129e-01 -2.69929677e-01 1.00840664e+00 -1.47954836e-01
-1.04498708e+00 2.95101228e-04 -4.31430608e-01 5.17998159e-01
1.70797199e-01 -5.60054362e-01 -8.20319712e-01 -5.67065835e-01
2.38828704e-01 -3.36506367e-01 9.59163904e-01 -2.00898841e-01
1.45408702e+00 -2.19977066e-01 -4.20967549e-01 8.83717000e-01
1.09696317e+00 2.51754944e-04 4.57510054e-01 2.06707269e-01
1.22803998e+00 1.83788732e-01 7.73151457e-01 2.06663415e-01
7.33935058e-01 5.48598170e-01 5.87622046e-01 8.41212720e-02
-3.15590024e-01 -3.18520725e-01 7.73776993e-02 1.24232495e+00
-1.85744226e-01 1.31654307e-01 -1.08222258e+00 4.11961347e-01
-1.98043680e+00 -5.37560582e-01 1.25830874e-01 1.75152075e+00
4.81370717e-01 3.92448872e-01 -2.64162093e-01 8.70991684e-03
4.81629282e-01 1.58349797e-01 -9.62412536e-01 2.82620192e-01
1.75338835e-01 9.44526717e-02 1.37210831e-01 1.18212774e-01
-1.09972394e+00 1.03136492e+00 4.82073498e+00 8.10428381e-01
-1.33725822e+00 3.50750417e-01 3.23576391e-01 -1.87584311e-01
-7.64604211e-02 4.79846261e-02 -1.03287053e+00 4.08712268e-01
4.54242587e-01 -4.02707886e-03 2.25870118e-01 1.02726972e+00
-1.33149937e-01 3.58933538e-01 -1.18118000e+00 1.33631611e+00
2.43775994e-01 -1.40249765e+00 6.17235303e-01 -1.23064414e-01
6.40999973e-01 8.55464414e-02 1.58302829e-01 6.77668452e-01
-2.19632596e-01 -6.38144732e-01 7.24381089e-01 9.04392838e-01
6.87399805e-01 -6.69171095e-01 6.30370319e-01 5.92985392e-01
-1.28955412e+00 -1.85513765e-01 -6.37273073e-01 -2.40974039e-01
3.73313800e-02 9.34135497e-01 -3.68190140e-01 6.82183564e-01
1.22160745e+00 1.19685113e+00 -8.31777394e-01 1.07249200e+00
1.02300152e-01 2.38160014e-01 -3.96988660e-01 7.69771710e-02
2.30988443e-01 2.01552138e-01 6.40783489e-01 6.37100697e-01
5.58203042e-01 2.20973745e-01 4.73068058e-01 7.09420204e-01
-1.68934137e-01 -2.32158974e-01 -4.87755686e-01 3.25027466e-01
3.45080465e-01 1.02337039e+00 -3.61550897e-01 -2.63598263e-01
-5.29877365e-01 1.01520634e+00 6.23201668e-01 3.02074552e-01
-5.47497451e-01 -1.50364116e-01 8.18834424e-01 2.67132193e-01
4.59385842e-01 -7.45509803e-01 -3.77005696e-01 -1.47083557e+00
2.99250335e-01 -4.86929148e-01 1.41273499e-01 -1.02888536e+00
-1.72451150e+00 5.83683610e-01 -3.01906262e-02 -1.37015951e+00
4.91127014e-01 -8.47772956e-01 -2.62434453e-01 6.18033767e-01
-1.66017032e+00 -1.52623928e+00 -6.22664094e-01 7.56891012e-01
6.26770437e-01 -2.84819096e-01 6.54905140e-01 5.96605301e-01
-4.12298262e-01 4.12290245e-01 -1.74095139e-01 -1.92198262e-01
6.35256290e-01 -8.63471210e-01 6.20139122e-01 5.32496691e-01
-6.86489865e-02 7.89164066e-01 -3.60203162e-02 -7.67778218e-01
-1.89838660e+00 -1.58303058e+00 5.44841826e-01 -5.86195111e-01
5.49260199e-01 -5.86737275e-01 -1.30685008e+00 6.65996671e-01
-5.18091023e-01 4.55464780e-01 3.79041046e-01 4.45940524e-01
-5.87521076e-01 -6.59706175e-01 -8.04004014e-01 3.41559112e-01
1.80150735e+00 -5.16215980e-01 -6.94621444e-01 5.05842209e-01
1.36727715e+00 -6.05427861e-01 -1.12777185e+00 6.72408521e-01
4.49789375e-01 -5.26728988e-01 1.22895801e+00 -6.75584555e-01
1.51604488e-01 -6.03886068e-01 -1.51335984e-01 -1.15764701e+00
-7.71108687e-01 -2.49320760e-01 -5.42009771e-01 1.10301447e+00
4.75514820e-03 -5.40854216e-01 8.05964828e-01 3.13415647e-01
-6.67696357e-01 -8.52322698e-01 -1.11128056e+00 -8.83581221e-01
2.41054475e-01 -7.70217836e-01 1.03088081e+00 9.04551387e-01
-6.16199434e-01 4.77804601e-01 -3.26583415e-01 4.69521284e-01
5.63721418e-01 4.45804417e-01 9.95273352e-01 -1.54341674e+00
-1.19383886e-01 -1.40912279e-01 -5.81129611e-01 -1.65805364e+00
8.17360282e-02 -1.21391404e+00 -2.82129735e-01 -1.24573326e+00
1.03446938e-01 -1.16238260e+00 -5.21787286e-01 6.94759786e-01
-2.43629321e-01 1.70016333e-01 4.31500345e-01 4.90526408e-01
-7.68459439e-01 9.23191488e-01 1.71018004e+00 -3.57211590e-01
-2.21578836e-01 -1.79391913e-02 -6.43256187e-01 5.31206608e-01
6.28098786e-01 -3.78377497e-01 -6.28916264e-01 -1.13687909e+00
3.53227884e-01 -3.73280317e-01 8.22569311e-01 -1.28206515e+00
5.29007196e-01 -9.27170068e-02 5.08533239e-01 -1.17530596e+00
3.05781007e-01 -1.13593578e+00 2.21132219e-01 2.59765655e-01
3.47882241e-01 7.49337524e-02 4.59224224e-01 6.92016006e-01
-5.60926571e-02 4.85656075e-02 3.63157302e-01 -2.82957405e-01
-1.05547416e+00 1.03323734e+00 2.03835458e-01 -3.01818192e-01
9.75990355e-01 -1.52828366e-01 -3.78250659e-01 4.16720152e-01
-7.73187935e-01 3.79771650e-01 4.81309950e-01 8.37823451e-01
8.94230723e-01 -1.58329690e+00 -5.03582299e-01 5.30265450e-01
2.20858604e-01 7.86678433e-01 7.22132027e-01 6.36069298e-01
-3.43161911e-01 3.02862555e-01 -2.77316749e-01 -1.08641088e+00
-7.02743173e-01 1.05598199e+00 1.79763734e-01 -1.60412967e-01
-8.43401492e-01 7.92080760e-01 2.87239552e-01 -7.93415010e-01
3.21548522e-01 -3.32154155e-01 2.10048586e-01 -6.81248456e-02
4.07442957e-01 3.13209087e-01 3.85147452e-01 -4.25640404e-01
-3.97280455e-01 1.01321697e+00 -5.94875574e-01 5.43259084e-01
1.53419244e+00 -1.40443474e-01 -3.36923927e-01 7.84027040e-01
1.23206985e+00 -7.00187445e-01 -1.21438611e+00 -4.75433707e-01
-3.87105048e-01 -5.79941213e-01 1.01940997e-01 -5.78493476e-01
-1.19703126e+00 1.13463545e+00 7.41321802e-01 -1.42801240e-01
9.74585533e-01 1.35125527e-02 1.08418381e+00 8.20179045e-01
9.01262641e-01 -7.63899863e-01 4.40095007e-01 9.38971341e-01
1.06674874e+00 -1.15305483e+00 -1.32856763e-03 -3.94199818e-01
-1.51973665e-01 1.05231571e+00 1.04058182e+00 -2.76345581e-01
1.19046283e+00 -2.32836798e-01 -2.90328711e-01 -3.22206378e-01
-8.88380826e-01 8.96310508e-02 3.30867380e-01 6.01346791e-01
-7.64252916e-02 -2.68657476e-01 2.71751851e-01 7.35367715e-01
-1.00833520e-01 3.27007622e-01 -6.06266558e-02 1.01935124e+00
-2.96731085e-01 -1.04122543e+00 -1.92004636e-01 5.07662714e-01
2.02196404e-01 1.04931183e-01 1.04463950e-01 5.44479251e-01
7.11911976e-01 4.10339475e-01 2.05445454e-01 -5.93717933e-01
6.76076531e-01 4.94655892e-02 4.73217666e-01 -7.22988963e-01
-4.45826769e-01 -3.14493626e-01 -5.34391522e-01 -5.50373137e-01
-5.65194786e-01 -5.07045448e-01 -1.01291180e+00 -5.30407548e-01
-2.91478217e-01 -5.74001335e-02 5.11786699e-01 8.49800885e-01
7.76846945e-01 6.19836450e-01 4.72001404e-01 -1.14573920e+00
-3.97113234e-01 -9.07928824e-01 -3.06497872e-01 4.55372125e-01
2.99639523e-01 -1.11306345e+00 -1.51502550e-01 -1.16544962e-01] | [7.923289775848389, -3.3732266426086426] |
6d6d18bd-5002-41a7-a85b-58e327381613 | large-scale-multi-granular-concept-extraction | 2208.14139 | null | https://arxiv.org/abs/2208.14139v1 | https://arxiv.org/pdf/2208.14139v1.pdf | Large-scale Multi-granular Concept Extraction Based on Machine Reading Comprehension | The concepts in knowledge graphs (KGs) enable machines to understand natural language, and thus play an indispensable role in many applications. However, existing KGs have the poor coverage of concepts, especially fine-grained concepts. In order to supply existing KGs with more fine-grained and new concepts, we propose a novel concept extraction framework, namely MRC-CE, to extract large-scale multi-granular concepts from the descriptive texts of entities. Specifically, MRC-CE is built with a machine reading comprehension model based on BERT, which can extract more fine-grained concepts with a pointer network. Furthermore, a random forest and rule-based pruning are also adopted to enhance MRC-CE's precision and recall simultaneously. Our experiments evaluated upon multilingual KGs, i.e., English Probase and Chinese CN-DBpedia, justify MRC-CE's superiority over the state-of-the-art extraction models in KG completion. Particularly, after running MRC-CE for each entity in CN-DBpedia, more than 7,053,900 new concepts (instanceOf relations) are supplied into the KG. The code and datasets have been released at https://github.com/fcihraeipnusnacwh/MRC-CE | ['Rui Xie', 'Yanghua Xiao', 'Kaiyan Cao', 'Jingyue Huang', 'Jilun Sun', 'Jiaqing Liang', 'Deqing Yang', 'Siyu Yuan'] | 2022-08-30 | null | null | null | null | ['machine-reading-comprehension'] | ['natural-language-processing'] | [-5.19437730e-01 4.68780339e-01 -3.43330413e-01 -2.35204458e-01
-1.83206245e-01 -4.42289799e-01 3.46273422e-01 6.18628919e-01
-4.77172285e-01 1.09809721e+00 8.49977434e-02 -3.50740880e-01
-3.85263413e-01 -1.53332424e+00 -7.17052102e-01 -1.08726442e-01
-1.60409093e-01 6.49133146e-01 4.50209290e-01 -4.76077318e-01
-2.75963414e-02 7.53288567e-02 -1.60387301e+00 3.51287723e-01
1.49631643e+00 7.79355168e-01 3.20051134e-01 1.02100998e-01
-6.38859272e-01 7.14375019e-01 -5.30308008e-01 -9.29469287e-01
-2.46266231e-01 1.16193101e-01 -1.14287233e+00 -5.35080314e-01
-9.97392386e-02 -1.48428932e-01 -2.64656991e-01 1.11685574e+00
6.45341575e-02 -2.74875034e-02 4.72143590e-01 -1.36326849e+00
-7.71489739e-01 1.15855789e+00 -4.71180737e-01 5.04448339e-02
5.56113541e-01 -4.74157423e-01 1.34410536e+00 -1.04647422e+00
7.60825813e-01 1.25872684e+00 4.69489872e-01 3.01375121e-01
-5.65074861e-01 -9.99968529e-01 3.73481631e-01 6.54629648e-01
-1.72015548e+00 -9.47143063e-02 2.59832978e-01 -6.66953772e-02
1.16009748e+00 1.94818959e-01 8.18595111e-01 5.36263287e-01
-2.93823838e-01 8.61016273e-01 8.86866450e-01 -6.34070694e-01
-8.68045911e-03 7.77166411e-02 3.94659430e-01 7.79593766e-01
1.13365483e+00 -3.18500102e-01 -6.46266580e-01 -3.27026844e-02
6.71766996e-01 -1.62442878e-01 -4.20812160e-01 -7.02168867e-02
-1.10875475e+00 7.19338596e-01 3.95575136e-01 3.86929721e-01
-3.13666195e-01 -3.33477646e-01 2.70896018e-01 9.51345786e-02
2.49242157e-01 6.41294837e-01 -1.07874310e+00 2.57029617e-03
-4.38692331e-01 4.97558951e-01 1.05397630e+00 1.78724122e+00
7.69174099e-01 -6.96568370e-01 1.22333787e-01 7.56019413e-01
3.57474923e-01 5.08991480e-01 2.50801474e-01 -4.85641927e-01
9.28308785e-01 1.13733029e+00 -5.47427535e-02 -1.20538652e+00
-6.33582711e-01 -5.23856819e-01 -8.83114874e-01 -7.42154062e-01
2.33253445e-02 -6.96168169e-02 -8.00706685e-01 1.55489111e+00
5.62634468e-01 -2.57767434e-03 4.09915596e-01 5.56938231e-01
1.48844755e+00 6.17637694e-01 3.28486323e-01 -2.10360125e-01
1.68969452e+00 -7.53797710e-01 -7.33244777e-01 -2.00250119e-01
6.30032778e-01 -1.63500845e-01 8.65772307e-01 3.34603012e-01
-7.34726787e-01 -3.49284858e-01 -1.02682889e+00 -1.82933614e-01
-9.31726336e-01 1.82977878e-02 1.03411245e+00 3.44986677e-01
-3.90628040e-01 2.84672439e-01 -5.68545997e-01 -3.52640271e-01
5.80591619e-01 1.88460827e-01 -4.85478073e-01 -5.87712228e-01
-1.92686975e+00 9.56025064e-01 1.44191456e+00 8.61700475e-02
-2.70504594e-01 -8.74847293e-01 -9.64013159e-01 2.03838691e-01
1.07397783e+00 -9.35920060e-01 9.52258527e-01 -1.44434618e-02
-7.87194729e-01 5.00030398e-01 4.85183001e-02 -2.55261719e-01
-6.07162947e-03 -4.01606023e-01 -9.98942733e-01 2.77924538e-01
4.50355828e-01 5.78664541e-01 1.92200378e-01 -1.10409784e+00
-1.13820136e+00 -2.65676528e-01 5.95998883e-01 5.02663076e-01
-4.28902388e-01 -7.90154561e-02 -7.63887346e-01 -5.01527965e-01
1.91609487e-01 -3.79868418e-01 1.16954260e-01 -7.46290863e-01
-5.50991237e-01 -4.27406311e-01 2.54211843e-01 -7.64309227e-01
1.72670650e+00 -1.58855557e+00 1.16427481e-01 3.28569651e-01
7.04118013e-01 2.68611223e-01 1.93294540e-01 5.63869476e-01
1.54926121e-01 3.48479778e-01 -7.17608631e-02 3.90595198e-01
1.22545607e-01 4.25559103e-01 -5.06482162e-02 -4.75998312e-01
1.13002911e-01 1.02171457e+00 -1.22067964e+00 -7.74479866e-01
-9.74412337e-02 -2.27274701e-01 -4.95759457e-01 -2.49109846e-02
-4.51037169e-01 -5.07394895e-02 -8.28403711e-01 1.05541086e+00
7.49635160e-01 -2.32037738e-01 5.67093253e-01 -2.41178960e-01
1.64965600e-01 2.89940864e-01 -1.25243127e+00 1.68225741e+00
-2.54359663e-01 -5.84612042e-02 -2.65316397e-01 -7.70948648e-01
7.12690711e-01 1.47893518e-01 1.01936281e-01 -5.95001400e-01
-1.59066901e-01 4.57258433e-01 -5.02369814e-02 -5.74015379e-01
7.32627928e-01 -6.33008108e-02 -3.15225184e-01 -8.22612494e-02
3.75422239e-01 -4.47443239e-02 9.05052185e-01 7.06214070e-01
9.70151126e-01 1.87747985e-01 7.13837862e-01 -3.27426583e-01
6.47012711e-01 3.25101554e-01 8.38932037e-01 4.38504964e-01
4.37067121e-01 -8.94160271e-02 5.40012062e-01 -6.51240051e-02
-5.78269780e-01 -7.66822815e-01 -2.28514329e-01 9.11105156e-01
3.03000480e-01 -1.02889752e+00 -5.18008828e-01 -8.86695445e-01
3.33759755e-01 8.20774615e-01 -2.81800628e-01 3.40577937e-03
-2.96499908e-01 -8.59856129e-01 7.58608401e-01 6.30517304e-01
8.65677953e-01 -1.04837406e+00 -1.92469597e-01 4.63436455e-01
-7.43075788e-01 -1.24526787e+00 2.07615256e-01 8.10253471e-02
-5.08738339e-01 -1.52864635e+00 -1.57089695e-01 -7.45988309e-01
6.47146702e-01 -6.43529370e-02 1.54408741e+00 2.76854932e-01
-5.51473871e-02 -3.36233452e-02 -9.23075795e-01 -6.84391022e-01
5.15082851e-02 5.59670866e-01 1.89408343e-02 -8.24033976e-01
1.12334013e+00 -5.86939573e-01 -2.22697914e-01 1.95485830e-01
-8.57117653e-01 4.36249822e-01 7.21531570e-01 6.88247740e-01
5.02817929e-01 7.32865989e-01 9.20913815e-01 -1.26340127e+00
7.18017101e-01 -8.05099130e-01 -4.43218619e-01 6.23764277e-01
-8.80758882e-01 -3.16019729e-02 5.38388431e-01 2.33794451e-02
-1.12824762e+00 -5.24564803e-01 -1.01148166e-01 1.14999495e-01
-1.37537464e-01 1.41980219e+00 -6.75053358e-01 3.11542630e-01
4.13493097e-01 1.86325312e-02 -7.39498973e-01 -5.05906403e-01
5.92930913e-01 7.79779851e-01 5.52343249e-01 -1.18546486e+00
7.67510831e-01 -1.39340907e-01 -1.50980860e-01 -4.99682128e-01
-1.08040142e+00 -6.46464407e-01 -8.00315559e-01 2.03476250e-01
5.24581313e-01 -1.21226060e+00 -5.80795288e-01 5.06609380e-01
-1.02526379e+00 2.21271291e-01 -1.16429657e-01 4.61703300e-01
-1.27886084e-03 1.58983484e-01 -7.09236562e-01 -4.18817431e-01
-4.43625182e-01 -2.41782531e-01 5.35207510e-01 5.96453369e-01
1.90908775e-01 -8.28661859e-01 -2.64922887e-01 4.04894024e-01
3.33325081e-02 1.65467173e-01 1.44102156e+00 -8.80401552e-01
-6.09068751e-01 3.40652280e-02 -4.72418249e-01 9.01507437e-02
1.17495649e-01 -7.94392973e-02 -4.31731433e-01 2.17384333e-03
-6.15937293e-01 -2.10730433e-01 7.52499044e-01 -4.99222130e-02
1.08277690e+00 -3.55357915e-01 -6.41562402e-01 3.23736280e-01
1.47236598e+00 2.09708527e-01 7.39421189e-01 6.01104140e-01
8.36041272e-01 5.14095604e-01 1.24525416e+00 3.34913105e-01
1.05719197e+00 1.25703648e-01 1.24845661e-01 1.69886559e-01
2.38166675e-01 -6.59764767e-01 -2.77550042e-01 1.26192069e+00
-5.50512195e-01 -2.28967354e-01 -1.12558162e+00 7.17898250e-01
-1.80634797e+00 -7.50930965e-01 -2.43125468e-01 1.75193942e+00
1.31847954e+00 3.11602473e-01 -1.85399860e-01 1.73054144e-01
7.12603271e-01 -2.95459181e-01 -4.17870641e-01 1.56092063e-01
-1.64986312e-01 4.78899539e-01 4.01234925e-01 2.03387544e-01
-9.64467227e-01 1.33276069e+00 4.76165342e+00 1.08246326e+00
-2.67947525e-01 -5.08075655e-02 3.30202579e-02 2.25020036e-01
-4.16318059e-01 1.78909332e-01 -1.21094143e+00 2.68479913e-01
5.39476275e-01 -5.72830677e-01 1.81742832e-01 7.53232896e-01
-4.49650615e-01 -1.30691186e-01 -7.72373319e-01 7.34222174e-01
-1.79092988e-01 -1.37062573e+00 3.00799876e-01 -6.66072359e-04
5.19107163e-01 -3.47062737e-01 -7.15791404e-01 8.88062358e-01
5.73344111e-01 -1.00957942e+00 4.71166700e-01 5.18304944e-01
1.07409811e+00 -9.93794680e-01 1.04008842e+00 4.57296997e-01
-1.66213059e+00 -4.98136468e-02 -6.98679090e-01 8.77603889e-02
-6.77224621e-02 8.99842262e-01 -6.48791790e-01 1.56996000e+00
9.25647140e-01 7.22043097e-01 -5.99854350e-01 7.62059689e-01
-9.14619267e-01 4.03757900e-01 -2.31483772e-01 -1.79728165e-01
-2.72210464e-02 -7.43092895e-02 1.32779747e-01 1.35285842e+00
2.48582214e-01 8.21572423e-01 1.35199443e-01 6.88421845e-01
-3.42751771e-01 3.83463919e-01 -2.61462219e-02 -2.53730476e-01
1.15380883e+00 1.38052392e+00 -4.85320389e-01 -6.36807382e-01
-5.94011128e-01 3.55208665e-01 6.84064209e-01 2.30244383e-01
-6.99265718e-01 -1.11112511e+00 3.53615791e-01 1.22536577e-01
3.01880270e-01 -4.06110547e-02 1.01470880e-01 -1.49848485e+00
2.00958356e-01 -9.42669928e-01 8.18813503e-01 -9.00942802e-01
-1.26143360e+00 5.04973412e-01 5.44392943e-01 -8.96764457e-01
-1.35670662e-01 -5.69473803e-01 -4.91913259e-02 8.24373066e-01
-1.68526804e+00 -1.32020557e+00 -4.99354988e-01 7.47919917e-01
5.45853865e-04 -3.78616527e-02 9.02609110e-01 5.35632312e-01
-4.16159511e-01 3.37200403e-01 -3.17723811e-01 3.27375442e-01
5.10367095e-01 -1.43877292e+00 4.77781445e-01 7.75088608e-01
2.00660024e-02 1.11818099e+00 3.30435187e-01 -1.01608884e+00
-1.12395430e+00 -1.11141217e+00 1.13049734e+00 -2.06459388e-01
6.58957779e-01 -2.05007315e-01 -1.09361291e+00 8.40367913e-01
-9.29129273e-02 -1.24295734e-01 7.32960463e-01 7.08533227e-01
-4.83004987e-01 -1.87217802e-01 -1.14611530e+00 4.50562805e-01
1.45452094e+00 -3.27178478e-01 -1.28731370e+00 1.13794371e-01
9.54114616e-01 -6.20808363e-01 -1.29023325e+00 6.61188602e-01
4.16473687e-01 -4.48315799e-01 7.93392241e-01 -6.82669699e-01
5.40583372e-01 -6.26907349e-01 -1.03263788e-01 -1.44614041e+00
-3.61583650e-01 2.06301343e-02 -7.07288563e-01 1.48495758e+00
7.74789572e-01 -7.93812990e-01 5.43070138e-01 4.09772694e-01
-9.19198543e-02 -9.40368474e-01 -5.21412611e-01 -7.23999500e-01
-2.36083791e-02 -3.61727804e-01 1.20666182e+00 1.31889820e+00
5.07245600e-01 5.07687271e-01 5.69903292e-02 3.44885111e-01
3.96449298e-01 3.65109414e-01 5.17035306e-01 -1.57254958e+00
-1.07234873e-01 -5.24408333e-02 -3.54100436e-01 -9.91910636e-01
-1.58230215e-03 -9.26838815e-01 -2.94195265e-01 -2.03971100e+00
3.88163984e-01 -8.56723428e-01 -3.64524335e-01 9.72612083e-01
-4.52526838e-01 -4.15947795e-01 -4.05340232e-02 8.46905187e-02
-6.95257545e-01 4.44267035e-01 1.30407035e+00 4.21652831e-02
-3.88005711e-02 -2.40133092e-01 -1.16591060e+00 6.69326067e-01
6.60589159e-01 -4.61136848e-01 -5.10710835e-01 -3.58832449e-01
7.08087385e-01 -8.41280222e-02 6.62737265e-02 -7.02900767e-01
4.78633612e-01 -3.17021906e-01 3.92310083e-01 -6.95589840e-01
-5.66060953e-02 -5.81057787e-01 8.06543678e-02 2.04936102e-01
6.94460571e-02 -6.23265617e-02 2.62362242e-01 4.92765009e-01
-5.48889279e-01 -4.38734680e-01 1.08155403e-02 -3.89855474e-01
-1.16426969e+00 4.30963457e-01 1.71467334e-01 5.04726052e-01
8.23063374e-01 1.29329130e-01 -7.39350080e-01 1.36729002e-01
-6.63785398e-01 6.77311242e-01 9.49091017e-02 4.29476619e-01
6.59593761e-01 -1.34307134e+00 -7.94999540e-01 -5.37372306e-02
5.38614094e-01 7.78706431e-01 1.99569806e-01 3.46545964e-01
-5.91754198e-01 6.59491241e-01 -1.44388005e-01 4.11660932e-02
-1.06260622e+00 5.88823617e-01 -6.37734085e-02 -5.60279310e-01
-5.65344036e-01 8.40816379e-01 -1.28092289e-01 -6.13901258e-01
-8.15558732e-02 -5.14031410e-01 -6.85301304e-01 1.53145999e-01
3.99407238e-01 1.75664708e-01 2.60416120e-01 -1.69429854e-01
-4.79479730e-01 1.73243135e-01 -2.79771805e-01 2.36982390e-01
1.32175100e+00 -1.30756155e-01 -5.76582789e-01 4.64650057e-02
4.25162643e-01 2.81828135e-01 -4.26209420e-01 -3.87555718e-01
4.07341808e-01 -4.20216084e-01 -2.49763638e-01 -1.16922104e+00
-8.54321361e-01 3.18626702e-01 -2.94599980e-01 8.23549554e-02
1.27439213e+00 1.98374957e-01 7.10032701e-01 6.06371641e-01
9.21837807e-01 -1.06560886e+00 -6.38266325e-01 6.17841899e-01
7.83066988e-01 -8.81348372e-01 3.83261889e-01 -1.17251372e+00
-5.86357951e-01 9.77647603e-01 9.05459106e-01 3.96671414e-01
6.89612925e-01 9.31053683e-02 -4.44790423e-01 -4.36350942e-01
-7.65409648e-01 -4.40689623e-01 3.01771641e-01 6.76474750e-01
2.50748843e-01 3.97716641e-01 -4.91973072e-01 1.39719844e+00
-4.98526782e-01 2.16115385e-01 3.01676393e-01 1.11821032e+00
-4.98235792e-01 -1.13065553e+00 7.71886408e-02 7.77791679e-01
-2.95391589e-01 -6.75894082e-01 -1.60130382e-01 1.18091691e+00
6.12872720e-01 9.83917475e-01 -3.83506387e-01 -4.96867985e-01
5.42892873e-01 -2.64712181e-02 4.96521115e-01 -9.37066615e-01
-3.06988984e-01 -5.30982077e-01 5.35446346e-01 -1.37253240e-01
-5.55425048e-01 -4.42557245e-01 -1.65877318e+00 -6.62055135e-01
-6.56709850e-01 5.87071657e-01 3.18805695e-01 1.11983955e+00
2.72106171e-01 5.75880349e-01 -1.10979574e-02 2.27014124e-01
1.16991671e-02 -1.17293799e+00 -8.08723629e-01 8.99877995e-02
-4.60208923e-01 -8.89459312e-01 2.19861753e-02 -1.52145624e-01] | [9.22546100616455, 8.414558410644531] |
731cbe4d-9b3c-494e-9ada-e05d4f6bc486 | boosting-algorithms-for-uplift-modeling | 1807.07909 | null | http://arxiv.org/abs/1807.07909v1 | http://arxiv.org/pdf/1807.07909v1.pdf | Boosting algorithms for uplift modeling | Uplift modeling is an area of machine learning which aims at predicting the
causal effect of some action on a given individual. The action may be a medical
procedure, marketing campaign, or any other circumstance controlled by the
experimenter. Building an uplift model requires two training sets: the
treatment group, where individuals have been subject to the action, and the
control group, where no action has been performed. An uplift model allows then
to assess the gain resulting from taking the action on a given individual, such
as the increase in probability of patient recovery or of a product being
purchased. This paper describes an adaptation of the well-known boosting
techniques to the uplift modeling case. We formulate three desirable properties
which an uplift boosting algorithm should have. Since all three properties
cannot be satisfied simultaneously, we propose three uplift boosting
algorithms, each satisfying two of them. Experiments demonstrate the usefulness
of the proposed methods, which often dramatically improve performance of the
base models and are thus new and powerful tools for uplift modeling. | ['Michał Sołtys', 'Szymon Jaroszewicz'] | 2018-07-20 | null | null | null | null | ['medical-procedure'] | ['medical'] | [ 5.90405464e-01 5.66758774e-02 -7.63132632e-01 -5.51724076e-01
-2.99865931e-01 -9.19826552e-02 8.80828202e-01 5.37156641e-01
-4.21005100e-01 1.05900729e+00 6.99293315e-02 -5.95402122e-01
-3.48634154e-01 -9.43918705e-01 -9.10988212e-01 -1.16147792e+00
8.53961408e-02 6.34047270e-01 -8.66806731e-02 3.74494493e-02
3.36178958e-01 3.33487242e-01 -1.39992774e+00 1.91451505e-01
7.03543365e-01 9.07117486e-01 -3.98443155e-02 4.84005153e-01
-9.68522280e-02 7.13091910e-01 -4.27181780e-01 -3.57843280e-01
1.46484911e-01 -4.65523899e-01 -6.05602324e-01 1.58157721e-01
6.06587809e-03 -2.25504726e-01 2.93172717e-01 8.11824977e-01
2.86640286e-01 -7.57526830e-02 7.38556504e-01 -1.36035490e+00
-3.00569445e-01 6.25528276e-01 -7.73578823e-01 1.51926413e-01
5.27715504e-01 -2.98374772e-01 1.02942073e+00 -3.82006973e-01
2.83253968e-01 1.10237253e+00 3.85854244e-01 4.88475919e-01
-1.33634603e+00 -7.31316388e-01 2.16957897e-01 7.82254562e-02
-8.71576786e-01 -3.26876283e-01 8.14231515e-01 -5.97495556e-01
3.35561216e-01 5.58422625e-01 7.69213796e-01 7.05876052e-01
2.61867374e-01 9.23265636e-01 1.44992125e+00 -6.48564279e-01
6.15473032e-01 3.41858447e-01 3.33257139e-01 4.17297453e-01
3.06110471e-01 5.08966386e-01 -4.88427043e-01 -5.76432109e-01
4.03834820e-01 2.67598540e-01 -1.37005538e-01 -4.43545461e-01
-9.11826849e-01 1.00570452e+00 5.20394444e-01 3.04551512e-01
-7.85979331e-01 1.47629296e-02 3.21918815e-01 1.32140532e-01
6.98728681e-01 3.12767357e-01 -5.96703827e-01 1.68246731e-01
-7.49818385e-01 5.47703445e-01 3.50743026e-01 3.30358535e-01
6.20494962e-01 -3.55294853e-01 -1.01397254e-01 3.70466560e-01
2.99632382e-02 3.12202990e-01 3.94746780e-01 -3.60859096e-01
1.71277165e-01 7.56316900e-01 4.90631580e-01 -6.75750256e-01
-3.04941863e-01 -4.93516356e-01 -7.35798717e-01 1.63666070e-01
4.75800723e-01 8.80173668e-02 -8.49643528e-01 1.90148628e+00
7.03702271e-01 3.34168524e-01 -2.77726799e-01 4.65926290e-01
4.03867155e-01 6.10722780e-01 4.86643225e-01 -7.46407449e-01
1.10948884e+00 -6.32829547e-01 -9.32016432e-01 -2.03732606e-02
7.45350420e-01 -4.90712672e-01 7.99518824e-01 5.46982050e-01
-9.33449328e-01 -2.69921571e-01 -8.92410040e-01 5.01719892e-01
-1.85622573e-01 -2.26524442e-01 1.01299596e+00 7.98190117e-01
-6.50402129e-01 7.11979389e-01 -6.54470682e-01 -1.12273037e-01
3.18679661e-01 4.82920766e-01 -1.78887829e-01 -1.76846370e-01
-1.18389356e+00 6.98108971e-01 5.15255146e-02 -4.01954167e-02
-4.95348990e-01 -7.22838461e-01 -4.56172854e-01 1.31946191e-01
3.38719785e-01 -8.64712834e-01 1.22650576e+00 -1.09885406e+00
-1.02741241e+00 9.35013473e-01 -2.56232470e-01 -5.79444110e-01
7.18545198e-01 -1.46515012e-01 -1.93357885e-01 -3.25740546e-01
1.82307914e-01 1.40308216e-01 9.80055869e-01 -8.92661154e-01
-8.54274809e-01 -9.68208015e-01 -3.79731804e-02 -1.69856437e-02
-2.53501654e-01 1.25907958e-01 2.71420151e-01 -5.59560537e-01
-1.35023415e-01 -8.65794599e-01 -4.05974448e-01 -2.98803896e-01
-4.19748425e-01 -4.06162262e-01 5.56642890e-01 -4.38970685e-01
1.16669405e+00 -1.90054214e+00 5.10345437e-02 1.94558755e-01
2.00608950e-02 1.76430270e-01 1.42844051e-01 4.13974494e-01
-4.47234124e-01 2.29376853e-01 2.59470697e-02 1.19844772e-01
-2.55386263e-01 2.44977832e-01 -2.45651051e-01 4.36814040e-01
-5.09731378e-03 4.51661795e-01 -6.12453759e-01 -3.83297861e-01
2.32730940e-01 2.54198760e-02 -4.66192722e-01 2.87704498e-01
-1.19043648e-01 5.14836311e-01 -9.18747306e-01 2.20585749e-01
6.29008591e-01 -1.76012129e-01 2.41032213e-01 1.85242802e-01
5.38102314e-02 3.99701178e-01 -9.45371091e-01 7.56148040e-01
-2.97047436e-01 -9.71975103e-02 3.56591418e-02 -1.35343707e+00
8.37859213e-01 4.27860111e-01 7.20031261e-01 -5.06749570e-01
2.26450488e-01 8.29893723e-02 1.15818091e-01 -3.67282450e-01
4.68077436e-02 -7.15835750e-01 2.10534826e-01 4.54588503e-01
-3.47337693e-01 3.07538122e-01 2.41520867e-01 2.36803759e-02
1.01454592e+00 -2.94335783e-01 9.20271397e-01 -3.13508809e-01
4.23427492e-01 -2.04213873e-01 7.70974636e-01 8.79779875e-01
-1.99045479e-01 -3.52285981e-01 4.61370587e-01 -7.23420680e-01
-8.41349602e-01 -7.85885811e-01 -2.97569036e-01 1.31241417e+00
-1.80623919e-01 6.99126348e-02 -3.97144139e-01 -7.90703177e-01
3.24727744e-01 7.02808201e-01 -1.04176521e+00 -9.35594216e-02
-3.58081937e-01 -1.18489313e+00 -2.34410971e-01 5.45071244e-01
3.10174376e-01 -6.95954740e-01 -5.06105006e-01 2.08904758e-01
-1.67138893e-02 -6.02349460e-01 -1.46781623e-01 2.48167679e-01
-1.32667887e+00 -1.03846073e+00 -4.41783130e-01 -5.02139270e-01
6.60465181e-01 9.11680683e-02 1.07174897e+00 2.94209003e-01
-2.67844908e-02 -7.93535560e-02 -3.88457507e-01 -7.96592534e-01
-5.11249304e-01 -3.26150924e-01 1.99562669e-01 3.50730658e-01
4.25448924e-01 -5.50225258e-01 -5.14746845e-01 2.64584631e-01
-6.47133946e-01 2.92562898e-02 4.87632722e-01 9.89920497e-01
3.80801558e-01 2.09770009e-01 1.03593338e+00 -1.20875108e+00
3.86730790e-01 -3.72743040e-01 -4.88701910e-01 2.30801851e-01
-8.34509969e-01 1.15760632e-01 2.89861023e-01 -5.96025348e-01
-1.02581453e+00 3.20698619e-01 -1.81449518e-01 9.58040729e-02
-3.74399731e-03 6.44530058e-01 -2.58111149e-01 2.06792191e-01
6.40083909e-01 8.90511349e-02 -1.12531841e-01 -5.43540597e-01
1.51193127e-01 6.33358896e-01 6.05318397e-02 -6.29594624e-01
5.50432861e-01 3.86879414e-01 2.05961153e-01 -6.48780167e-01
-1.05959022e+00 -4.71327424e-01 -3.37487996e-01 -2.98506856e-01
6.33299708e-01 -4.31866914e-01 -9.00498033e-01 2.32382834e-01
-6.37267172e-01 -9.72620621e-02 -1.75335616e-01 6.04558468e-01
-4.41593915e-01 -2.58010570e-02 -1.30703256e-01 -1.15795910e+00
-1.96796834e-01 -8.42772841e-01 6.69656277e-01 2.23747134e-01
-2.01193273e-01 -1.06010723e+00 5.97440861e-02 5.79006612e-01
9.44940373e-03 3.97598714e-01 1.16316712e+00 -9.71298754e-01
-3.31350863e-01 -7.46261656e-01 2.76548684e-01 2.94892520e-01
3.87952626e-01 -1.79165706e-01 -7.97162533e-01 -4.35048193e-01
2.58963108e-01 -2.50244766e-01 6.80638909e-01 9.70091403e-01
1.17440498e+00 -4.46839899e-01 -5.98676026e-01 -1.75981969e-02
1.11671484e+00 7.19042480e-01 4.05233055e-01 3.08600962e-01
6.05259128e-02 4.87405628e-01 8.47711325e-01 5.38327515e-01
-1.03663847e-01 9.27493989e-01 5.04157543e-01 -3.71177822e-01
3.90719742e-01 -2.79677719e-01 1.09192252e-01 2.61893809e-01
-2.44478181e-01 1.90369785e-02 -5.36433637e-01 4.74435925e-01
-1.72044551e+00 -9.22160745e-01 -3.28279763e-01 2.61307597e+00
9.25824940e-01 2.65624881e-01 4.05587018e-01 4.83455509e-01
8.63151252e-01 -4.24070135e-02 -5.25631011e-01 -5.70253193e-01
2.74507552e-01 1.09627180e-01 3.46368611e-01 5.09190559e-01
-1.12361336e+00 2.83251792e-01 6.63521576e+00 8.54479611e-01
-8.94245088e-01 -1.68969054e-02 9.85598683e-01 1.46356031e-01
-7.19555840e-02 7.45613053e-02 -8.14652324e-01 4.23410684e-01
8.65110517e-01 -5.76362550e-01 1.15310915e-01 1.04991496e+00
4.89423513e-01 -4.40907270e-01 -1.50911582e+00 3.36170793e-01
-3.25874627e-01 -1.12105250e+00 -1.09999701e-01 4.45266306e-01
5.99149942e-01 -6.91827536e-01 -4.49507609e-02 2.91365057e-01
3.28108162e-01 -6.85514987e-01 3.96073163e-01 1.94074616e-01
2.18713015e-01 -6.63532615e-01 7.58360088e-01 8.40928674e-01
-6.12677634e-01 -1.46034881e-01 5.32792835e-03 -4.57348317e-01
5.70330508e-02 8.67691219e-01 -1.20508790e+00 6.31205320e-01
3.75539809e-01 3.47161114e-01 -2.36537114e-01 9.73082125e-01
-3.27791691e-01 9.14087057e-01 -1.46569684e-01 -3.55909020e-01
-5.00213960e-03 -1.90795973e-01 3.80628109e-01 6.55080914e-01
1.46578729e-01 1.51307285e-01 2.80066282e-01 3.88868302e-01
2.08779842e-01 3.52551788e-01 -6.59091413e-01 5.04914075e-02
1.50825426e-01 1.01933038e+00 -6.30642295e-01 -4.94753122e-01
-3.16226125e-01 4.42411453e-01 1.58823060e-03 3.50451134e-02
-6.93718851e-01 1.03193916e-01 4.13477480e-01 3.89439076e-01
8.53866935e-02 5.83722055e-01 -5.94449937e-01 -8.40337217e-01
-8.93435925e-02 -7.13250399e-01 5.77009141e-01 -6.73510373e-01
-1.21478486e+00 1.67397112e-01 5.49463481e-02 -9.67953682e-01
-3.08072656e-01 -5.43725848e-01 -5.16972244e-01 8.36905956e-01
-1.15030265e+00 -1.01020396e+00 2.60608882e-01 5.47076643e-01
3.39596331e-01 4.42989580e-02 7.73440063e-01 3.18003446e-01
-5.38501680e-01 3.52332234e-01 7.72257671e-02 -2.40813985e-01
4.35676038e-01 -1.18522549e+00 -4.18788940e-01 6.71755075e-01
-1.20744623e-01 6.34410858e-01 1.13898051e+00 -7.91292846e-01
-7.54655957e-01 -8.68269086e-01 1.19055486e+00 -3.17696601e-01
4.76543874e-01 -2.88797677e-01 -6.49111807e-01 6.66074634e-01
-2.79181898e-01 -3.33516061e-01 8.28886688e-01 7.14711726e-01
-2.04945982e-01 -4.49887544e-01 -1.24511325e+00 4.00865436e-01
4.69778806e-01 1.22976020e-01 -6.78773046e-01 5.98991513e-01
4.38625485e-01 6.11210102e-03 -7.86637723e-01 5.86368322e-01
6.92402482e-01 -8.91317904e-01 1.01621354e+00 -1.34125042e+00
7.16310263e-01 8.72784704e-02 -1.37700304e-01 -1.28427482e+00
-3.24440449e-01 -2.44266197e-01 2.49961913e-02 1.10249794e+00
4.17906821e-01 -7.02903807e-01 8.00547302e-01 9.02912796e-01
2.08904713e-01 -9.35348511e-01 -9.46003318e-01 -9.29416835e-01
1.11938834e-01 -1.70062855e-01 7.94878185e-01 1.03986800e+00
1.69941679e-01 5.91083169e-01 -8.06240439e-01 -1.28035605e-01
5.32248735e-01 2.94190109e-01 7.74952590e-01 -1.40469575e+00
-3.22175264e-01 -9.00428072e-02 -3.92359167e-01 -6.14951015e-01
-1.01824969e-01 -6.02395952e-01 -1.61140293e-01 -1.39846683e+00
6.60789788e-01 -4.97395188e-01 -4.67580557e-01 6.33078992e-01
-5.79125106e-01 -3.22811365e-01 -8.63558799e-02 5.79207055e-02
3.36730033e-02 4.24727768e-01 1.22123063e+00 6.48693293e-02
-2.69452602e-01 9.05704439e-01 -8.09595764e-01 8.58894169e-01
7.60403752e-01 -6.28307164e-01 -4.44669992e-01 3.31347227e-01
2.01390103e-01 5.84249794e-01 3.31568986e-01 -3.27183992e-01
-2.37467721e-01 -4.23506707e-01 3.56025845e-01 -2.99798280e-01
1.80427477e-01 -9.06706691e-01 3.37431103e-01 1.06344533e+00
-6.26435101e-01 -4.02314439e-02 -1.58701077e-01 7.89830208e-01
-8.50103330e-03 -2.83929378e-01 9.11470652e-01 6.70879632e-02
-1.98569953e-01 8.41902196e-03 -2.06186354e-01 -2.79519051e-01
1.20342386e+00 1.37107102e-02 -1.92489505e-01 -5.17520607e-01
-1.08707559e+00 -6.56423857e-03 3.59527841e-02 1.07186340e-01
2.46456489e-01 -1.32306802e+00 -7.61120677e-01 8.20128694e-02
-5.20149022e-02 -6.38373435e-01 1.84593111e-01 9.48433876e-01
4.41926681e-02 3.66543263e-01 -1.61820203e-01 -4.36440051e-01
-1.65881836e+00 1.07156205e+00 6.01773849e-03 -8.16185594e-01
-3.98078769e-01 7.00796247e-01 6.78310931e-01 -1.38983458e-01
5.71086742e-02 -9.32962168e-03 -2.60979176e-01 -7.20454976e-02
6.10164404e-01 3.32439959e-01 3.14803362e-01 -2.84999698e-01
-3.27687949e-01 1.71627164e-01 -3.43579262e-01 -3.55370566e-02
1.39240575e+00 9.53367352e-02 -1.35409653e-01 4.42635804e-01
7.72229075e-01 -1.58704311e-01 -7.46987760e-01 -2.46014968e-01
-1.00076273e-01 -7.37367988e-01 7.45119825e-02 -1.12423015e+00
-8.58853936e-01 6.43113554e-01 5.77468693e-01 6.50119543e-01
1.55506539e+00 -4.38988023e-02 2.37732947e-01 4.38135751e-02
5.14199853e-01 -7.21052051e-01 -2.09637925e-01 -3.82304370e-01
8.89729679e-01 -1.09234989e+00 2.10610658e-01 -6.18895233e-01
-4.44671184e-01 4.82480079e-01 1.85893968e-01 -3.20404433e-02
6.59408391e-01 -6.43671900e-02 -2.55432576e-01 -5.08861728e-02
-8.92718613e-01 6.69220835e-02 2.00257808e-01 3.88441771e-01
4.36454177e-01 4.35451657e-01 -1.01793861e+00 3.85991424e-01
-5.06018922e-02 4.04961824e-01 3.86930287e-01 8.69072914e-01
-4.31877285e-01 -1.37445307e+00 -5.68285763e-01 8.28917623e-01
-5.56880355e-01 1.71990350e-01 -4.46024507e-01 9.98163581e-01
1.65614486e-01 1.02487278e+00 -1.76702589e-01 -2.39840299e-01
3.14559847e-01 2.78312773e-01 3.85270000e-01 -4.31293398e-01
-4.41315383e-01 2.32309580e-01 2.33745769e-01 -3.81821424e-01
-6.74069703e-01 -1.03949511e+00 -7.21061885e-01 -3.27139229e-01
-5.87844074e-01 3.37201506e-01 6.78882837e-01 1.07898712e+00
-1.77575827e-01 4.30628359e-01 9.64773953e-01 -4.01676178e-01
-7.78809667e-01 -9.63505328e-01 -8.99759054e-01 5.04724681e-01
3.25181395e-01 -9.45307612e-01 -3.36968064e-01 2.41241157e-01] | [8.324538230895996, 5.358263969421387] |
d157818d-9794-4ab2-8bce-7fafe504ac41 | feature-fused-ssd-fast-detection-for-small | 1709.05054 | null | http://arxiv.org/abs/1709.05054v3 | http://arxiv.org/pdf/1709.05054v3.pdf | Feature-Fused SSD: Fast Detection for Small Objects | Small objects detection is a challenging task in computer vision due to its
limited resolution and information. In order to solve this problem, the
majority of existing methods sacrifice speed for improvement in accuracy. In
this paper, we aim to detect small objects at a fast speed, using the best
object detector Single Shot Multibox Detector (SSD) with respect to
accuracy-vs-speed trade-off as base architecture. We propose a multi-level
feature fusion method for introducing contextual information in SSD, in order
to improve the accuracy for small objects. In detailed fusion operation, we
design two feature fusion modules, concatenation module and element-sum module,
different in the way of adding contextual information. Experimental results
show that these two fusion modules obtain higher mAP on PASCALVOC2007 than
baseline SSD by 1.6 and 1.7 points respectively, especially with 2-3 points
improvement on some smallobjects categories. The testing speed of them is 43
and 40 FPS respectively, superior to the state of the art Deconvolutional
single shot detector (DSSD) by 29.4 and 26.4 FPS. Code is available at
https://github.com/wnzhyee/Feature-Fused-SSD. Keywords: small object detection,
feature fusion, real-time, single shot multi-box detector | ['Guimei Cao', 'Xuemei Xie', 'Jinjian Wu', 'Guangming Shi', 'Wenzhe Yang', 'Quan Liao'] | 2017-09-15 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [-6.82245195e-02 -3.94760579e-01 3.59582394e-01 -2.21302509e-01
-7.61207938e-01 -2.86125779e-01 4.27208692e-01 3.25049721e-02
-7.64499187e-01 2.80010253e-01 -1.80982068e-01 1.02615416e-01
1.90124929e-01 -5.19931674e-01 -7.87191868e-01 -7.50764847e-01
3.40811312e-01 7.12286755e-02 1.30227315e+00 -1.93842366e-01
2.41899684e-01 4.29871529e-01 -1.86778748e+00 2.92204201e-01
8.15982223e-01 1.04975772e+00 9.83001471e-01 6.65720642e-01
-4.54257727e-02 4.81885284e-01 -5.14950514e-01 -2.90946335e-01
4.48614895e-01 -7.19044507e-02 -1.92761183e-01 -2.05229353e-02
4.46316063e-01 -7.74413705e-01 -2.77634293e-01 1.19406176e+00
8.13254535e-01 -1.45002127e-01 5.62848628e-01 -1.19764674e+00
-3.76201004e-01 3.04743201e-01 -9.92805898e-01 6.58301473e-01
1.42882764e-01 4.07479137e-01 4.94351029e-01 -1.13837886e+00
3.39549959e-01 1.26322114e+00 6.28988564e-01 5.37482381e-01
-7.81392038e-01 -6.47204101e-01 -1.76158503e-01 4.68341112e-01
-1.43519545e+00 -4.18075740e-01 1.84243768e-01 -4.55053121e-01
9.64417756e-01 5.33421524e-02 1.75979853e-01 6.33447170e-01
1.74710765e-01 8.28399479e-01 9.76219356e-01 -3.21219623e-01
-3.45569886e-02 3.17522049e-01 3.33450705e-01 4.94999021e-01
6.60884678e-01 1.79756671e-01 -2.98421890e-01 3.81230325e-01
7.08523393e-01 2.56672233e-01 -2.35554665e-01 -1.92559063e-01
-1.14039361e+00 6.37474895e-01 5.24997711e-01 2.24516720e-01
-3.30989927e-01 -7.63244703e-02 3.96121383e-01 1.00099362e-01
1.20104834e-01 -2.02205136e-01 -3.38203877e-01 -5.00270799e-02
-5.79730392e-01 3.32815766e-01 3.09259236e-01 1.04240155e+00
4.71980721e-01 3.02368831e-02 -3.72495204e-01 6.78672969e-01
4.71886069e-01 8.77965093e-01 5.77481389e-01 -7.34707475e-01
4.80578631e-01 5.27112126e-01 2.50599980e-01 -6.83730364e-01
-5.14814496e-01 -4.14796591e-01 -5.28368294e-01 5.80383062e-01
3.31375599e-01 -1.30382448e-01 -9.32865381e-01 1.22145975e+00
5.33055604e-01 2.22714677e-01 1.28394276e-01 1.28020012e+00
1.19479406e+00 7.07679033e-01 -4.69611138e-02 -1.32694900e-01
1.69716668e+00 -9.86259043e-01 -5.32880306e-01 -2.18139037e-01
5.97979188e-01 -1.06464875e+00 8.14461708e-01 2.58486092e-01
-9.33153868e-01 -9.58075821e-01 -1.29391944e+00 -3.57022509e-02
-1.31195530e-01 4.52596188e-01 2.55676776e-01 7.48549342e-01
-7.21343219e-01 4.30160642e-01 -7.52901852e-01 -5.89080215e-01
4.29473400e-01 3.53768468e-01 -1.95365787e-01 1.37311788e-02
-7.33516395e-01 9.53770041e-01 7.05690920e-01 -1.66294754e-01
-6.15322769e-01 -7.87530303e-01 -5.17877460e-01 1.48426622e-01
6.38291836e-01 -4.94929045e-01 1.55821991e+00 -4.44654703e-01
-1.36254323e+00 6.88748837e-01 9.41681936e-02 -6.63028479e-01
6.10478640e-01 -4.23218638e-01 -3.20330024e-01 1.10721335e-01
1.31848335e-01 6.89968765e-01 6.59514129e-01 -8.64136457e-01
-1.20699978e+00 -5.58484435e-01 2.93275323e-02 4.16639566e-01
-1.44191474e-01 4.94518876e-01 -5.42365611e-01 -2.26812109e-01
-6.79221153e-02 -7.48033166e-01 -4.66202274e-02 2.89083123e-01
-4.53659669e-02 -4.64587629e-01 1.30640721e+00 -4.20002669e-01
9.58891928e-01 -2.31152630e+00 -3.34663033e-01 -6.13609135e-01
1.67235315e-01 9.04269755e-01 1.83641478e-01 5.87768108e-02
4.39969927e-01 -4.93586898e-01 -1.45589486e-01 -3.96645129e-01
-2.56963700e-01 -2.98515838e-02 1.09673671e-01 5.85491776e-01
2.64786005e-01 7.14907348e-01 -6.24507248e-01 -6.49320006e-01
6.02923930e-01 4.94459867e-01 -2.35247642e-01 1.05954736e-01
1.99426532e-01 9.67473090e-02 -4.30435508e-01 7.87212968e-01
1.19434071e+00 4.99001928e-02 -5.11193156e-01 -3.11845481e-01
-6.15929186e-01 -2.17590541e-01 -1.73390746e+00 1.59473419e+00
-8.72819498e-02 5.60245037e-01 7.01543316e-02 -5.81480205e-01
8.74091804e-01 1.77602381e-01 1.41122177e-01 -6.83952987e-01
5.09782672e-01 4.04722184e-01 1.69824153e-01 -4.83165950e-01
4.40714449e-01 -4.72365953e-02 2.67747909e-01 -6.27631396e-02
1.40746281e-01 3.68704833e-03 4.50695783e-01 1.87968954e-01
8.59391987e-01 2.03831657e-03 3.42169672e-01 -2.01087236e-01
7.52438605e-01 -9.96062756e-02 4.76561159e-01 6.52045250e-01
-4.85807389e-01 1.00553393e+00 1.79339364e-01 -1.79274399e-02
-1.08804047e+00 -9.40521061e-01 -3.83910775e-01 8.28546941e-01
5.39320171e-01 -1.75413206e-01 -6.84781969e-01 -4.34903413e-01
3.41165736e-02 5.04180908e-01 -2.78987586e-01 -4.35357578e-02
-4.37081128e-01 -7.28152096e-01 3.81832212e-01 7.98363030e-01
8.58099639e-01 -7.64836252e-01 -1.33583283e+00 1.13035470e-01
1.67271152e-01 -1.28359187e+00 -4.22790438e-01 -1.45655256e-02
-7.87636638e-01 -1.04374611e+00 -9.49216664e-01 -7.59858310e-01
3.17014128e-01 1.00340116e+00 5.62935829e-01 -1.42201424e-01
-6.36638284e-01 6.85875863e-02 -5.19997597e-01 -8.05235505e-01
-1.01218097e-01 -4.14802611e-01 1.23083189e-01 -1.32645983e-02
4.14954334e-01 -1.87938046e-02 -8.77148390e-01 4.41755325e-01
-8.86343837e-01 9.46075767e-02 1.00938213e+00 4.86364484e-01
3.94202828e-01 -3.77303630e-01 2.39124060e-01 -3.81350398e-01
2.76481241e-01 -2.76281148e-01 -9.39877033e-01 -7.90133700e-02
-4.41166878e-01 -1.12263411e-01 4.06835407e-01 -3.27367008e-01
-1.25100970e+00 3.03441823e-01 -3.26550007e-02 -5.73345959e-01
-1.63198963e-01 -1.44299209e-01 -3.58444750e-02 -2.80231744e-01
8.39058340e-01 3.06770772e-01 2.19507903e-01 -7.50468194e-01
2.78277248e-01 1.02409279e+00 5.45287430e-01 3.53885442e-02
5.02231181e-01 4.80487436e-01 -8.39719698e-02 -7.59872794e-01
-6.37645543e-01 -1.09994102e+00 -6.35731876e-01 -3.02322835e-01
9.72919881e-01 -1.17733610e+00 -6.83653772e-01 8.12452734e-01
-1.23572421e+00 2.03643352e-01 -1.08986780e-01 7.86028504e-01
-3.03674638e-01 4.41758573e-01 -3.80944282e-01 -6.29267812e-01
-6.09098136e-01 -1.10913765e+00 1.18523109e+00 6.94433808e-01
5.35573065e-01 -1.66255161e-01 -2.81888574e-01 2.66158700e-01
3.97547662e-01 1.69014171e-01 -9.73073319e-02 -4.31069076e-01
-6.42379045e-01 -2.55681306e-01 -7.74676323e-01 5.57238817e-01
-1.97916850e-01 -1.02598770e-02 -9.22225595e-01 -2.96554655e-01
2.22437322e-01 -1.46097869e-01 9.85507607e-01 5.69558024e-01
5.14363706e-01 3.71939063e-01 -4.45345491e-01 4.72220868e-01
1.78908122e+00 4.43790972e-01 4.46046561e-01 3.50311786e-01
5.15435874e-01 1.31864980e-01 9.53030109e-01 6.23895407e-01
1.77868694e-01 7.51540542e-01 4.91496831e-01 1.71484157e-01
-6.16910338e-01 3.29212725e-01 4.40213829e-01 3.39403182e-01
-2.47375265e-01 -9.24093053e-02 -9.36749101e-01 3.66960555e-01
-1.89766145e+00 -9.67150688e-01 -7.31450319e-01 2.29524851e+00
2.28026152e-01 5.59196711e-01 3.27722281e-01 9.74633917e-02
9.42425847e-01 -5.97385392e-02 -4.30755883e-01 -9.14398730e-02
-1.61287934e-01 -2.63550311e-01 8.48819733e-01 1.61811918e-01
-1.10238516e+00 7.50541270e-01 4.84840155e+00 8.96235049e-01
-9.88941252e-01 5.19033194e-01 1.39700830e-01 -2.22703829e-01
6.81310236e-01 -2.44169474e-01 -1.44747925e+00 6.43965662e-01
8.51845026e-01 -1.84494570e-01 -7.70163536e-02 9.37477887e-01
1.74308255e-01 -3.77830386e-01 -7.03903198e-01 1.24245822e+00
1.76304579e-01 -1.06429935e+00 -2.43297294e-01 -2.08681837e-01
5.79116046e-01 3.24670672e-01 -1.93414792e-01 3.23255509e-01
-2.03061208e-01 -3.50103229e-01 9.42310750e-01 3.51460069e-01
4.91702318e-01 -6.90738380e-01 9.82371867e-01 6.16798162e-01
-1.47362363e+00 -2.25042269e-01 -6.51808619e-01 -1.36878908e-01
2.86033422e-01 3.21185023e-01 -7.91014552e-01 5.30349374e-01
1.03283131e+00 3.76685560e-01 -7.41037548e-01 1.55719519e+00
5.21689877e-02 2.86200345e-01 -5.13000011e-01 -3.78567964e-01
2.84747660e-01 -1.93442311e-02 7.43571699e-01 1.30596614e+00
4.39804524e-01 3.72965306e-01 7.25527853e-02 6.07310951e-01
2.45756194e-01 2.88724117e-02 -2.97801137e-01 5.00638843e-01
4.40178156e-01 1.34148479e+00 -8.44131827e-01 -5.63083172e-01
-8.04424882e-01 9.75764215e-01 -1.03964753e-01 -1.57234535e-01
-1.28734589e+00 -7.08214164e-01 3.60577583e-01 2.11918786e-01
7.26602554e-01 -3.23891073e-01 -1.62066787e-01 -9.02448118e-01
2.81859189e-01 -3.95090878e-01 2.87468970e-01 -8.78394783e-01
-7.69874394e-01 5.45557022e-01 6.12433441e-02 -1.51284766e+00
1.48874819e-01 -8.87391031e-01 -5.37813962e-01 6.72427177e-01
-1.42257929e+00 -1.18946135e+00 -7.10125566e-01 5.57135582e-01
1.08279586e+00 -8.94355923e-02 1.58463821e-01 5.98952711e-01
-5.34367085e-01 4.84879464e-01 3.94721985e-01 -1.70106038e-01
6.45673573e-01 -1.00706482e+00 2.87213475e-01 1.11744201e+00
-9.13878083e-02 7.38443062e-02 8.02483618e-01 -4.36954647e-01
-1.49766016e+00 -9.59908843e-01 4.32613641e-01 -3.53854358e-01
3.37629616e-01 -3.19871426e-01 -8.94443750e-01 2.82354414e-01
7.27685690e-02 3.22194159e-01 4.33401801e-02 -5.02461314e-01
-9.90340114e-02 -2.85293639e-01 -1.31643534e+00 2.65682876e-01
9.98017073e-01 7.27981403e-02 -6.43944383e-01 2.82114804e-01
9.42296445e-01 -5.91887772e-01 -4.34386134e-01 5.75592935e-01
4.62033421e-01 -1.33959413e+00 1.04965436e+00 3.22523266e-02
3.92618850e-02 -7.69809723e-01 -2.43015841e-01 -7.62527108e-01
-4.38358098e-01 -1.46079555e-01 -1.72657385e-01 1.21773648e+00
9.90096703e-02 -6.73245549e-01 6.21780992e-01 1.88912764e-01
-4.60305065e-01 -4.91733402e-01 -9.41033483e-01 -1.19216418e+00
-4.33311760e-01 -3.36154193e-01 3.13923478e-01 1.28701955e-01
-4.78024185e-01 4.16438043e-01 -2.31197923e-01 3.45905393e-01
8.08185279e-01 -6.07486963e-02 8.00247192e-01 -1.12175572e+00
-3.55510384e-01 -3.98949593e-01 -8.41690421e-01 -9.76243317e-01
-7.83672810e-01 -4.41065282e-01 1.42883331e-01 -1.63090539e+00
4.13982481e-01 5.35928644e-02 -2.99940109e-01 1.14957049e-01
-3.01891059e-01 3.57667506e-01 6.27592921e-01 2.52103865e-01
-8.63892317e-01 4.42043990e-01 1.06257355e+00 1.97404176e-01
-7.41272122e-02 1.73683926e-01 -3.96430671e-01 6.75445318e-01
6.93373561e-01 -5.39633155e-01 -2.36136571e-01 -4.47952986e-01
-5.76917291e-01 -7.82898515e-02 4.25973415e-01 -1.72075617e+00
4.61460471e-01 1.96362913e-01 3.87009501e-01 -1.03474784e+00
5.62384605e-01 -6.69009566e-01 -4.88960296e-02 9.50139523e-01
3.63666922e-01 -2.66083598e-01 4.07652140e-01 6.11799181e-01
-1.32366911e-01 -4.48911905e-01 1.13477886e+00 3.21813812e-03
-1.33973634e+00 1.30140692e-01 -1.57171503e-01 -1.98138192e-01
1.47820401e+00 -5.48226714e-01 -4.07468170e-01 3.32406908e-01
-2.79440343e-01 2.89841563e-01 1.00482665e-01 5.77077746e-01
7.27687478e-01 -1.15264869e+00 -1.00759661e+00 8.01807493e-02
1.94571331e-01 1.15579041e-02 5.91325104e-01 1.12190306e+00
-7.76158869e-01 5.69753289e-01 -5.00205398e-01 -8.79580081e-01
-1.72916865e+00 6.16652429e-01 1.70264721e-01 2.18055755e-01
-6.74770594e-01 1.13637042e+00 3.18415850e-01 1.41358271e-01
2.51977503e-01 -5.62051415e-01 -2.57930100e-01 -1.37624834e-02
9.13640201e-01 6.37674093e-01 1.96611613e-01 -5.77284098e-01
-4.67324674e-01 8.09250593e-01 -1.53375208e-01 2.10748389e-01
1.16272140e+00 -2.16190904e-01 4.07709748e-01 2.39173025e-01
1.01464796e+00 -3.36957186e-01 -1.59387219e+00 -3.07181031e-01
-2.01111525e-01 -7.94221759e-01 1.24369383e-01 -6.16256773e-01
-7.83128679e-01 8.00944149e-01 1.31719565e+00 5.09980991e-02
1.11673832e+00 1.92017749e-01 7.63490200e-01 1.71428725e-01
3.02343786e-01 -8.35934639e-01 5.29904477e-02 3.94999057e-01
8.79200995e-01 -1.67683709e+00 1.08756892e-01 -5.25865495e-01
-7.05682874e-01 1.11785293e+00 8.92816603e-01 -3.47980320e-01
4.42223549e-01 4.63883609e-01 -1.98861197e-01 -5.71561195e-02
-5.22575080e-01 -5.57700694e-01 3.03374052e-01 4.73811805e-01
7.34311044e-02 4.11688238e-02 -6.38967097e-01 6.06869698e-01
2.80914903e-01 -5.95276058e-02 5.54342687e-01 1.12878764e+00
-1.21353805e+00 -6.48208976e-01 -6.41744852e-01 4.66100216e-01
-3.36190403e-01 2.90832911e-02 1.91875562e-01 8.26757610e-01
4.97113347e-01 9.39955473e-01 1.48753256e-01 -3.64568561e-01
7.82482922e-01 -3.63051206e-01 4.64196384e-01 -5.49525797e-01
-4.00053620e-01 2.20756754e-01 -8.46193433e-02 -5.70628524e-01
-2.38423705e-01 -8.56565297e-01 -1.44886720e+00 -1.61393538e-01
-7.50205219e-01 -2.15012580e-01 9.81408119e-01 5.70023477e-01
4.08985019e-01 5.14643610e-01 3.70616823e-01 -8.93481553e-01
-8.43096256e-01 -1.10811687e+00 -6.85471296e-01 2.10015863e-01
2.00777590e-01 -8.96802306e-01 -2.53269821e-01 4.52026986e-02] | [8.642452239990234, -0.5582991242408752] |
e72d11d5-5f3d-4d08-b5bc-1bcc469b1d41 | plmcl-partial-label-momentum-curriculum | 2208.09999 | null | https://arxiv.org/abs/2208.09999v1 | https://arxiv.org/pdf/2208.09999v1.pdf | PLMCL: Partial-Label Momentum Curriculum Learning for Multi-Label Image Classification | Multi-label image classification aims to predict all possible labels in an image. It is usually formulated as a partial-label learning problem, given the fact that it could be expensive in practice to annotate all labels in every training image. Existing works on partial-label learning focus on the case where each training image is annotated with only a subset of its labels. A special case is to annotate only one positive label in each training image. To further relieve the annotation burden and enhance the performance of the classifier, this paper proposes a new partial-label setting in which only a subset of the training images are labeled, each with only one positive label, while the rest of the training images remain unlabeled. To handle this new setting, we propose an end-to-end deep network, PLMCL (Partial Label Momentum Curriculum Learning), that can learn to produce confident pseudo labels for both partially-labeled and unlabeled training images. The novel momentum-based law updates soft pseudo labels on each training image with the consideration of the updating velocity of pseudo labels, which help avoid trapping to low-confidence local minimum, especially at the early stage of training in lack of both observed labels and confidence on pseudo labels. In addition, we present a confidence-aware scheduler to adaptively perform easy-to-hard learning for different labels. Extensive experiments demonstrate that our proposed PLMCL outperforms many state-of-the-art multi-label classification methods under various partial-label settings on three different datasets. | ['Song Wang', 'XiaoFeng Wang', 'Xinyi Wu', 'Zhenyao Wu', 'Xin Zhang', 'Rabab Abdelfattah'] | 2022-08-22 | null | null | null | null | ['multi-label-image-classification', 'partial-label-learning'] | ['computer-vision', 'methodology'] | [ 5.68503737e-01 1.45042285e-01 -5.57948947e-01 -6.35623813e-01
-8.02351654e-01 -6.31277978e-01 1.80003807e-01 2.54036784e-01
-6.67901218e-01 7.81164229e-01 -6.90030038e-01 -8.57061967e-02
1.36833349e-02 -3.68746012e-01 -7.28646398e-01 -1.06485665e+00
4.28877503e-01 6.61626458e-01 1.12660430e-01 5.33982992e-01
7.26658152e-03 1.13523565e-01 -1.55671382e+00 3.91518116e-01
7.94985354e-01 1.21426392e+00 3.98299307e-01 4.32001650e-01
5.36655495e-03 7.95679450e-01 -4.40074176e-01 -3.15209180e-01
2.19781876e-01 -2.86131084e-01 -7.46205330e-01 4.26182389e-01
6.17900133e-01 -8.29711091e-03 4.42074388e-01 1.34534860e+00
2.30708942e-01 -9.11872685e-02 7.26494968e-01 -1.39903343e+00
-5.14234781e-01 5.20366251e-01 -9.86008942e-01 -8.52270573e-02
-2.78031975e-01 -1.40855953e-01 1.04046082e+00 -7.96392500e-01
3.77685636e-01 9.69275594e-01 5.45917392e-01 7.91513681e-01
-1.04979491e+00 -8.91653955e-01 6.22166097e-01 6.46748096e-02
-1.30667198e+00 8.67255107e-02 6.32411301e-01 -3.62252206e-01
2.62071759e-01 1.45091444e-01 2.15586513e-01 7.66122818e-01
6.57349080e-02 9.86032605e-01 1.48858070e+00 -5.90367317e-01
2.23795995e-01 5.14945984e-01 3.56366396e-01 8.35428655e-01
2.19623782e-02 9.48934816e-03 -3.78114820e-01 -2.00803250e-01
2.03378558e-01 3.67209427e-02 -1.35192484e-01 -3.82882714e-01
-9.84056652e-01 7.28075683e-01 1.25156343e-01 3.46955284e-02
-2.16938257e-01 9.71332788e-02 3.15485716e-01 1.19394109e-01
7.51646638e-01 7.98071548e-02 -8.37512255e-01 2.62961596e-01
-8.36750448e-01 -1.89968944e-01 4.66816634e-01 8.17788124e-01
1.05888069e+00 -4.45477247e-01 -2.26368904e-01 1.00387561e+00
3.33422035e-01 2.93730557e-01 4.54408467e-01 -7.42589176e-01
2.82468677e-01 6.86031580e-01 8.43666121e-02 -4.23816621e-01
-5.54817855e-01 -7.30728507e-01 -8.22365582e-01 2.79619604e-01
3.10863495e-01 -2.56515205e-01 -1.07884645e+00 1.87620568e+00
6.27538919e-01 5.50460994e-01 -9.75950211e-02 8.45728874e-01
3.68275940e-01 5.89357316e-01 3.94215375e-01 -6.14386678e-01
1.32096171e+00 -1.38964820e+00 -4.97930825e-01 -3.69964510e-01
7.68082857e-01 -6.79295003e-01 9.52878118e-01 3.87283325e-01
-6.80048943e-01 -7.13144958e-01 -9.81172264e-01 2.99292415e-01
-1.56114772e-01 7.22879827e-01 4.05215055e-01 6.67636216e-01
-7.92629123e-01 4.94857788e-01 -5.87631762e-01 1.57801270e-01
4.15799648e-01 4.45787489e-01 -1.19873263e-01 -2.70653695e-01
-8.56579781e-01 6.68673396e-01 6.99751616e-01 1.29492030e-01
-8.55274379e-01 -5.10202169e-01 -6.65919483e-01 -5.45770628e-03
6.37868166e-01 -1.19241565e-01 1.30759645e+00 -1.29834509e+00
-1.04928195e+00 1.06982243e+00 -1.89578459e-01 -3.60863768e-02
5.56416154e-01 2.12408626e-03 -2.21778467e-01 2.06900332e-02
3.50254804e-01 1.16394961e+00 9.67867553e-01 -1.46587265e+00
-1.22129464e+00 -1.31141111e-01 -1.42479660e-02 4.28116620e-01
-4.32979256e-01 -1.26686022e-01 -4.68055189e-01 -3.87874275e-01
1.17422462e-01 -1.33518755e+00 -2.07767099e-01 1.79647449e-02
-4.29839373e-01 -5.35012782e-01 8.38631809e-01 -4.47156392e-02
8.40710580e-01 -2.17356753e+00 -5.33985011e-02 -5.23630008e-02
-1.08369887e-01 3.38239372e-01 -1.73075408e-01 -2.73078561e-01
-1.69192031e-01 1.04303680e-01 -1.45581171e-01 -9.51873362e-01
-1.55918673e-01 3.94716114e-01 -7.55106658e-02 5.53286910e-01
3.37254584e-01 5.48660398e-01 -9.45286155e-01 -8.49949121e-01
1.18112937e-01 2.95137852e-01 -2.24601716e-01 1.05889589e-01
-4.46205974e-01 8.18858624e-01 -3.44169497e-01 7.85260856e-01
9.23691630e-01 -8.39071512e-01 3.28921914e-01 -5.50704896e-02
6.56484663e-02 -3.49098593e-01 -1.35350978e+00 1.32693267e+00
-5.38885891e-01 8.22613761e-02 1.82192959e-02 -9.37637091e-01
5.69879770e-01 4.97907102e-01 4.95433420e-01 -3.55203122e-01
1.47959605e-01 2.34666988e-01 -3.31315309e-01 -1.93749130e-01
1.74013659e-01 -2.28944525e-01 3.70457023e-02 7.43884087e-01
4.34533134e-02 2.66127974e-01 3.53614211e-01 -1.19070210e-01
5.17401993e-01 1.05831675e-01 1.07953921e-01 5.16720377e-02
6.13201141e-01 -2.78123736e-01 8.64656091e-01 7.05990911e-01
-3.28757197e-01 4.58452135e-01 3.41729045e-01 -3.56814146e-01
-7.77095556e-01 -3.88490736e-01 -5.35116017e-01 1.52047765e+00
4.36818272e-01 2.69199139e-03 -5.97477376e-01 -1.31395006e+00
-1.05883420e-01 4.00046766e-01 -6.57041550e-01 5.24049923e-02
-4.10465300e-01 -1.18483186e+00 1.49102271e-01 2.74869353e-01
4.55575973e-01 -1.18019581e+00 -5.16404629e-01 1.57952666e-01
-1.69256508e-01 -1.11504459e+00 -6.23571873e-01 8.97329152e-01
-6.77338779e-01 -1.22384953e+00 -5.31654656e-01 -1.24244905e+00
1.12472343e+00 2.02557445e-01 7.66393721e-01 1.16382405e-01
-4.45621274e-03 -8.15677922e-03 -3.64528298e-01 -1.05222762e-01
-5.77787697e-01 1.87309161e-01 -1.01784043e-01 4.04456258e-01
2.46340275e-01 -1.61603495e-01 -3.48893285e-01 6.59842372e-01
-7.80468285e-01 2.95435071e-01 5.84041059e-01 1.16092777e+00
1.15207887e+00 3.16948771e-01 7.40506172e-01 -1.21390080e+00
1.93299949e-01 -5.54355681e-01 -6.29297495e-01 7.23590136e-01
-9.75239277e-01 3.36485915e-03 7.50665605e-01 -1.00139737e+00
-8.25728774e-01 6.25203490e-01 1.05294086e-01 -6.05130136e-01
-2.15604693e-01 4.32806879e-01 6.91007227e-02 -2.32679859e-01
2.98735559e-01 3.78278419e-02 -1.53834701e-01 -4.22626078e-01
2.46943936e-01 6.60927892e-01 3.49519908e-01 -5.15916705e-01
4.27748859e-01 3.38510007e-01 1.71487227e-01 -1.32490560e-01
-1.57355702e+00 -7.88544297e-01 -6.70746148e-01 -4.03332710e-01
7.51880467e-01 -1.01284540e+00 -7.57841289e-01 6.14255905e-01
-8.83968532e-01 -3.89833868e-01 -1.36584416e-01 3.76785219e-01
-2.49528214e-01 4.11624223e-01 -7.38454938e-01 -7.67587543e-01
-2.90889382e-01 -1.45983994e+00 1.24687815e+00 5.67225337e-01
2.03476235e-01 -1.04263866e+00 -1.37237549e-01 3.45378876e-01
1.00577874e-02 -4.48100269e-03 8.71040225e-01 -6.58267796e-01
-3.52452576e-01 -1.93208963e-01 -3.08238298e-01 6.46609426e-01
2.83128396e-03 -3.07900041e-01 -1.12918723e+00 -6.06455088e-01
7.50667751e-02 -1.04401517e+00 9.71584916e-01 3.36459100e-01
1.15243697e+00 -6.92511871e-02 -5.12371361e-01 3.52232873e-01
1.40902400e+00 9.60970894e-02 -1.31619245e-01 1.82115391e-01
7.09981322e-01 4.41928506e-01 1.14485562e+00 4.55845475e-01
3.82708192e-01 6.82127774e-01 7.00564265e-01 -2.61461794e-01
-2.62679476e-02 8.51990879e-02 8.95568505e-02 6.12262607e-01
4.95496064e-01 -3.96162510e-01 -4.74117965e-01 1.84148252e-01
-1.91954386e+00 -4.95429873e-01 -1.13263102e-02 2.00860620e+00
1.16926551e+00 1.72186613e-01 -2.05586061e-01 1.11701898e-01
1.14435184e+00 8.03810358e-02 -9.47581768e-01 3.53535861e-02
-1.15904257e-01 5.87216765e-02 6.16827369e-01 3.87649983e-01
-1.73385465e+00 8.17388773e-01 5.18013382e+00 9.29019749e-01
-1.34566855e+00 4.59941357e-01 1.14240742e+00 -4.55567427e-02
1.51746556e-01 -1.50241666e-02 -1.26822376e+00 5.56726635e-01
6.75921440e-01 4.08423901e-01 1.28758103e-01 1.21834576e+00
-2.81493485e-01 -4.67606246e-01 -1.13678753e+00 8.50547791e-01
2.04276040e-01 -9.19356942e-01 -3.97186011e-01 -1.04122721e-01
1.09991896e+00 -2.94925719e-02 2.37447515e-01 4.10750687e-01
2.20544457e-01 -6.28062308e-01 9.39222157e-01 9.75997150e-02
1.12075150e+00 -6.33886218e-01 7.22237408e-01 7.20474780e-01
-1.14183164e+00 -3.06892544e-01 -3.95473152e-01 1.75247952e-01
1.65439263e-01 6.32508159e-01 -7.36183226e-01 2.65072197e-01
4.32318896e-01 6.05505943e-01 -7.67003894e-01 9.80188310e-01
-4.56920624e-01 6.36987746e-01 -3.03968787e-01 6.78419992e-02
3.37840080e-01 -4.25128117e-02 -1.68680608e-01 9.04241085e-01
1.27445281e-01 -1.38681352e-01 8.99468422e-01 3.64044607e-01
-2.46861607e-01 9.07579437e-02 1.39088959e-01 3.81944865e-01
4.39259887e-01 1.56168926e+00 -1.03512168e+00 -3.49484056e-01
-4.66963768e-01 1.01121604e+00 6.03015184e-01 2.04160139e-01
-9.66799319e-01 2.45706245e-01 -1.56138826e-03 -3.74600291e-01
2.72282898e-01 2.72160321e-01 -2.77166933e-01 -9.76003468e-01
1.24274589e-01 -4.81226504e-01 5.99345863e-01 -5.44565082e-01
-1.23958385e+00 4.62944567e-01 -8.25591832e-02 -1.30375671e+00
-5.59158698e-02 -4.69989717e-01 -1.98882550e-01 7.88133979e-01
-1.94409239e+00 -1.34354067e+00 -3.44087660e-01 2.63829559e-01
5.52704990e-01 7.63876811e-02 8.47120643e-01 5.63802779e-01
-6.93945944e-01 7.71168172e-01 2.99947530e-01 -1.96912393e-01
1.06574178e+00 -1.26170003e+00 -2.19424322e-01 5.83491206e-01
2.13060290e-01 2.53452826e-02 2.86979556e-01 -5.46817005e-01
-5.48884392e-01 -1.45619953e+00 8.46920967e-01 -1.21655287e-02
1.95244625e-01 -1.74358204e-01 -8.70567322e-01 6.66019797e-01
-1.19170688e-01 5.99321544e-01 7.55954325e-01 6.50318637e-02
-2.20322892e-01 -1.09311707e-01 -1.09635031e+00 2.19105139e-01
4.88017946e-01 -3.41215432e-01 6.29195496e-02 9.01479602e-01
6.28736496e-01 -5.87389052e-01 -5.44505358e-01 5.60166478e-01
3.29525709e-01 -5.52247584e-01 6.14424348e-01 -1.12348892e-01
1.23728640e-01 -5.59272349e-01 2.14499593e-01 -1.23339963e+00
-2.76957154e-01 7.01787174e-02 9.18394178e-02 1.23020613e+00
6.11872852e-01 -4.21968073e-01 1.00009358e+00 5.02757967e-01
-1.32794738e-01 -1.15433538e+00 -1.01099193e+00 -5.76006770e-01
-1.22293033e-01 -2.53261030e-01 2.32244194e-01 1.11646509e+00
-3.66378039e-01 3.13079536e-01 -7.17035770e-01 3.75385344e-01
6.25485837e-01 3.42949659e-01 1.35109037e-01 -1.28130865e+00
-4.28229243e-01 -8.47910866e-02 2.35854345e-03 -9.36666906e-01
5.69072247e-01 -1.06039822e+00 2.74491280e-01 -1.06134713e+00
7.04033434e-01 -1.17032838e+00 -6.76245928e-01 9.75319684e-01
-5.66560864e-01 6.56725705e-01 8.24008808e-02 3.25762063e-01
-1.13003111e+00 2.44422942e-01 1.34476113e+00 -2.60023743e-01
8.95747333e-04 3.03340584e-01 -4.76304889e-01 5.94546020e-01
7.05213666e-01 -1.00991380e+00 -5.08935571e-01 -2.27270380e-01
2.39070147e-01 4.09440100e-02 4.20563631e-02 -9.59475994e-01
2.91979522e-01 -2.80348122e-01 2.92995363e-01 -6.77289009e-01
2.12876529e-01 -8.56506467e-01 8.08778927e-02 4.17543948e-01
-5.98939836e-01 -3.33480746e-01 -3.35014164e-02 6.52593315e-01
-1.58537596e-01 -8.10380280e-01 1.28318298e+00 -7.98321366e-02
-7.43899345e-01 3.63489956e-01 3.29210050e-02 -3.48172486e-02
1.45155406e+00 5.80469184e-02 -3.38137060e-01 1.03025720e-01
-8.60309005e-01 6.39268637e-01 4.18502361e-01 2.27019355e-01
1.85821190e-01 -1.25231206e+00 -4.41529423e-01 8.73826370e-02
1.57993197e-01 3.47609460e-01 4.37072545e-01 6.37245774e-01
-1.27558857e-01 2.59470522e-01 -1.49402665e-02 -8.15042734e-01
-1.38732874e+00 8.55330884e-01 2.76355088e-01 -6.72318816e-01
-2.45104283e-01 1.11309683e+00 3.17983896e-01 -5.62078953e-01
5.34452438e-01 -1.36412397e-01 -3.32269698e-01 2.38057435e-01
4.59281832e-01 -1.47328675e-02 1.12039689e-02 -6.37937367e-01
-1.82614014e-01 8.58494222e-01 -4.88394856e-01 2.14830413e-01
8.92938256e-01 -1.37713075e-01 -8.47453773e-02 5.94509661e-01
1.28330433e+00 -2.81041801e-01 -1.60960734e+00 -3.04425269e-01
-4.65826578e-02 -2.16692045e-01 -8.69264081e-02 -1.01228464e+00
-1.25872457e+00 7.58833170e-01 9.00869787e-01 -1.21993281e-01
1.25891221e+00 -5.00993207e-02 5.18836498e-01 2.96600848e-01
4.63880479e-01 -1.18208659e+00 3.39879602e-01 4.12766278e-01
2.12820262e-01 -1.79765928e+00 3.60281952e-02 -4.60724026e-01
-8.70096326e-01 9.67573106e-01 8.83111775e-01 3.35537642e-01
7.59449542e-01 1.48143291e-01 2.76824266e-01 1.06971726e-01
-7.94134915e-01 5.12570255e-02 3.12078316e-02 1.36634037e-01
2.02385083e-01 1.55869186e-01 -2.71729678e-01 3.78697515e-01
5.50843120e-01 -7.09219053e-02 3.15446556e-01 9.47872758e-01
-4.76618350e-01 -1.44762862e+00 -2.50068992e-01 4.30534750e-01
-7.42675424e-01 5.92257082e-02 1.41384572e-01 3.98347795e-01
7.01300442e-01 9.69068766e-01 3.53528149e-02 -2.29890361e-01
-1.92103699e-01 2.97430962e-01 2.71722883e-01 -8.98345351e-01
-6.12237155e-01 3.02989632e-01 -1.23902306e-01 -5.54257259e-02
-7.23049402e-01 -5.96789718e-01 -1.34620094e+00 2.10049480e-01
-9.62140977e-01 1.98672175e-01 7.11529613e-01 9.62406516e-01
1.05168700e-01 5.32755136e-01 7.95134485e-01 -7.24247396e-01
-7.27625132e-01 -1.04242849e+00 -8.13035607e-01 4.11940724e-01
2.56003052e-01 -8.14724922e-01 -3.94096017e-01 1.16389923e-01] | [9.53879451751709, 4.0780463218688965] |
e40677e9-41ab-4d2a-aa1c-0b2d66099558 | lit01-196-a-metabolically-stable-apelin-17 | 2211.04115 | null | https://arxiv.org/abs/2211.04115v1 | https://arxiv.org/pdf/2211.04115v1.pdf | LIT01-196, a Metabolically Stable Apelin-17 Analog, Normalizes Blood Pressure in Hypertensive DOCA-Salt Rats via a NO Synthase-dependent Mechanism | Apelin is a neuro-vasoactive peptide that plays a major role in the control of cardiovascular functions and water balance, but has an in-vivo half-life in the minute range, limiting its therapeutic use. We previously developed LIT01-196, a systemically active metabolically stable apelin-17 analog, produced by chemical addition of a fluorocarbon chain to the N-terminal part of apelin-17. LIT01-196 behaves as a potent full agonist for the apelin receptor and has an in vivo half-life in the bloodstream of 28 min after intravenous (i.v.) and 156 min after subcutaneous (s.c.) administrations in conscious normotensive rats. We aimed to investigate the effects of LIT01-196 following systemic administrations on arterial blood pressure, heart rate, fluid balance and electrolytes in conscious normotensive and hypertensive deoxycorticosterone acetate (DOCA)-salt rats. Acute i.v. LIT01-196 administration, in increasing doses, dose-dependently decreases arterial blood pressure with ED 50 values of 9.8 and 3.1 nmol/kg in normotensive and hypertensive rats, respectively. This effect occurs for both via a nitric oxide-dependent mechanism. Moreover, acute s.c. LIT01-196 administration (90 nmol/kg) normalizes arterial blood pressure in conscious hypertensive DOCA-salt rats for more than 7 h. The LIT01-196-induced blood pressure decrease remains unchanged after 4 consecutive daily s.c. administrations of 90 nmol/kg, and does not induce any alteration of plasma sodium and potassium levels and kidney function as shown by the lack of change in plasma creatinine and urea nitrogen levels. Activating the apelin receptor with LIT01-196 may constitute a novel approach for the treatment of hypertension. | ['Catherine Llorens-Cortes', 'Dominique Bonnet', 'Nadia de Mota', 'Lucie Esteoulle', 'Pierre-Emmanuel Girault-Sotias', 'Mathilde Keck', 'Adrien Flahault'] | 2022-11-08 | null | null | null | null | ['kidney-function'] | ['medical'] | [ 1.75171793e-01 -2.85259694e-01 -4.92080271e-01 -2.29219154e-01
1.51686355e-01 -7.24093735e-01 2.29611069e-01 7.82229662e-01
-9.32785809e-01 9.33481455e-01 1.71033040e-01 -4.85933423e-01
2.24631995e-01 -7.27067888e-01 -3.54132771e-01 -4.45998937e-01
-3.38031322e-01 5.42075753e-01 -8.32709018e-03 1.65150344e-01
-3.92127372e-02 9.07519996e-01 -8.42983067e-01 -2.75731325e-01
1.29592681e+00 5.69482028e-01 -4.97545004e-02 6.22794807e-01
-3.29075754e-01 2.14256361e-01 -1.27869517e-01 6.74074739e-02
5.54992974e-01 -1.15766883e+00 -1.85905680e-01 -3.62514317e-01
3.18269968e-01 -4.98249948e-01 1.50279731e-01 8.75564039e-01
5.30493557e-01 -4.79282886e-02 4.50180650e-01 -7.74518907e-01
-7.78328851e-02 6.34737730e-01 -4.23922837e-01 2.12259889e-01
-2.23289669e-01 -1.72800153e-01 1.43950701e-01 -8.17198575e-01
1.52717441e-01 1.00813818e+00 2.90859073e-01 5.81099987e-01
-1.82996142e+00 -7.73521662e-01 -1.05665267e-01 -2.01657519e-01
-1.33866560e+00 -5.88572443e-01 2.08980829e-01 -4.32152838e-01
2.16952875e-01 4.18740958e-01 7.15448260e-01 3.53146106e-01
6.70896351e-01 2.30415583e-01 1.19898057e+00 -4.08170730e-01
4.66901541e-01 1.38849929e-01 -1.16846479e-01 2.31895372e-01
7.72771180e-01 1.67515442e-01 -1.01293109e-01 -3.83293569e-01
1.03488779e+00 -6.95651025e-02 -4.78007644e-01 -6.30283117e-01
-5.92660189e-01 9.31075215e-01 -1.31837115e-01 5.89818716e-01
-4.78134483e-01 -7.99530298e-02 8.40321064e-01 1.91314295e-01
-1.59639582e-01 4.41963941e-01 -4.31353360e-01 -9.75020453e-02
-2.85392165e-01 3.79052073e-01 9.33499098e-01 9.21089500e-02
-8.00567400e-03 2.20073849e-01 1.14259064e-01 9.29785132e-01
4.42449778e-01 7.19977856e-01 6.31404757e-01 -1.00932896e+00
-1.38960090e-02 2.89910167e-01 4.90785509e-01 -2.55852371e-01
-5.35794079e-01 -4.47046787e-01 -7.31980622e-01 4.86020923e-01
1.08704162e+00 -2.18004361e-01 -6.14643753e-01 1.96899581e+00
2.42761597e-01 -4.83107083e-02 1.66258305e-01 1.16908562e+00
4.14702922e-01 3.66735876e-01 1.12637520e+00 -8.09908271e-01
1.49244094e+00 -3.60429287e-01 -5.02821505e-01 -4.89120232e-03
2.58531421e-01 -4.71242845e-01 5.72238982e-01 2.01517805e-01
-9.73630428e-01 -2.39761680e-01 -9.36276197e-01 2.84546822e-01
-1.74141005e-01 -1.23373583e-01 4.37090993e-01 1.12321842e+00
-1.80777177e-01 8.65469754e-01 -9.43387032e-01 -3.48295480e-01
-2.52311546e-02 1.28432766e-01 -9.68821421e-02 -4.58410531e-02
-1.35643291e+00 1.09051549e+00 -2.01564077e-02 -7.19649866e-02
-1.54527530e-01 -1.25467801e+00 -9.72376287e-01 3.12476397e-01
-8.21512640e-02 -1.12262940e+00 8.88127089e-01 1.16411969e-03
-1.78291571e+00 4.86524582e-01 -1.19652212e-01 -8.87002110e-01
5.52966952e-01 1.32202776e-02 -4.42520440e-01 4.72412407e-01
3.28929245e-01 -3.77649185e-03 -1.10902436e-01 -6.58755720e-01
-2.52349108e-01 -6.52149916e-01 -5.70676208e-01 2.97826946e-01
4.41984534e-01 2.95664389e-02 3.94926935e-01 -5.76287568e-01
2.44393960e-01 -6.49443686e-01 -9.81887043e-01 8.61307792e-03
2.58926064e-01 5.21254480e-01 1.52962327e-01 -3.07024211e-01
5.17718315e-01 -1.71734071e+00 -5.08723617e-01 3.94265801e-01
5.42354062e-02 5.28830469e-01 -3.99388038e-02 7.49937356e-01
-1.91587448e-01 4.32824612e-01 -2.40363359e-01 7.65391290e-01
-7.09173307e-02 -4.32802625e-02 -2.01098591e-01 7.26639628e-01
-7.00686097e-01 8.36417973e-01 -1.09894025e+00 -2.02678949e-01
7.92995512e-01 3.41153920e-01 -2.27547646e-01 -7.37920225e-01
-7.64852539e-02 6.15334809e-01 -4.32517409e-01 4.89485443e-01
7.82870770e-01 6.02485538e-01 8.43774080e-01 5.36919609e-02
-5.03091455e-01 -7.58416206e-02 -5.90708017e-01 1.29415870e+00
-2.70380825e-01 8.24195072e-02 -7.01847300e-02 -4.50023979e-01
1.13620746e+00 4.81654227e-01 5.09479702e-01 -1.69570625e+00
2.32846409e-01 4.57762361e-01 3.49196792e-01 -7.70798251e-02
-1.74491376e-01 -1.21178603e+00 3.28313917e-01 1.29949212e-01
-4.86517042e-01 5.21153808e-01 1.00757027e+00 1.14539765e-01
5.97537577e-01 -7.41372444e-03 8.35853040e-01 -8.17148447e-01
1.01214421e+00 -6.02021813e-01 6.75208271e-01 7.57713377e-01
-1.89879999e-01 -1.14639886e-01 8.02696228e-01 -5.34121394e-01
-9.57916200e-01 -1.29253519e+00 -9.72573578e-01 3.24454278e-01
2.97405899e-01 -1.99321404e-01 -1.77786902e-01 5.42680062e-02
8.66553128e-01 9.86750960e-01 -3.53948146e-01 9.23347697e-02
-3.70571136e-01 -6.20127559e-01 6.10947609e-01 1.63855597e-01
4.78649646e-01 -6.39510393e-01 -6.61006689e-01 1.02367997e+00
2.29833007e-01 -4.72708225e-01 5.42284884e-02 9.45022255e-02
-1.19726980e+00 -9.34579432e-01 -1.04115915e+00 -1.80810571e-01
3.94395947e-01 -3.94701421e-01 5.52143872e-01 -2.47105122e-01
-4.43380207e-01 -5.03325462e-01 1.96858168e-01 -5.50732553e-01
-2.81092972e-01 -4.92657304e-01 3.78564417e-01 -2.04802483e-01
4.45179105e-01 -1.37687027e+00 -8.50494146e-01 4.60816413e-01
-1.08539693e-01 -2.73982644e-01 8.50576520e-01 1.48512706e-01
7.40594983e-01 -7.20796525e-01 1.05577409e+00 -1.30259657e+00
4.78803784e-01 -3.48472506e-01 -7.71903157e-01 -5.14798403e-01
-7.22512722e-01 -1.93286519e-02 8.11361492e-01 -2.18409553e-01
-7.66321480e-01 -3.58622074e-01 -5.28791510e-02 2.66570628e-01
8.29844794e-04 8.97387207e-01 9.06815100e-03 8.75087678e-02
1.12847137e+00 2.73408681e-01 3.44922006e-01 -5.94480455e-01
2.85948396e-01 1.29044175e-01 9.52449203e-01 -7.11076975e-01
6.13310397e-01 2.63074517e-01 4.55581307e-01 -1.15875328e+00
-3.80634695e-01 -6.07090354e-01 -2.70363003e-01 2.29312941e-01
4.98731703e-01 -8.06961179e-01 -1.27702785e+00 5.74192926e-02
-3.01897258e-01 -4.91286308e-01 -2.19950527e-01 1.21370041e+00
-3.77197176e-01 4.20006275e-01 -7.29488194e-01 -3.52334648e-01
-6.20156109e-01 -2.73794919e-01 -4.54272002e-01 2.45053932e-01
-5.22609711e-01 -6.11623108e-01 5.78803241e-01 -7.00449347e-02
6.08977020e-01 9.47807968e-01 1.29187000e+00 -5.31415045e-01
-2.36988753e-01 -5.17461479e-01 -1.99737057e-01 2.14178994e-01
1.25465378e-01 -5.00576317e-01 2.03400664e-02 1.04493387e-01
-1.05148613e-01 1.40240684e-01 5.45225143e-01 8.13613534e-01
2.26350993e-01 -3.82819325e-01 -1.97871417e-01 2.00991660e-01
1.67912626e+00 1.07730091e+00 9.25180495e-01 2.80578643e-01
-7.47482553e-02 -1.93691403e-01 8.67223516e-02 4.43726122e-01
-3.53774905e-01 2.30918899e-01 -5.64902881e-03 -1.31846055e-01
1.21595286e-01 -1.21346362e-01 -1.03940919e-01 1.51395753e-01
-3.39608759e-01 3.31963658e-01 -3.94024253e-01 3.17793816e-01
-1.29088497e+00 -1.22933078e+00 -5.89505434e-01 2.51275396e+00
1.34427965e+00 2.02787071e-01 1.33601889e-01 -5.75040758e-01
3.39424998e-01 -3.31306159e-01 -6.01521671e-01 -7.63550043e-01
3.92551929e-01 6.49752736e-01 6.89702332e-01 8.16704571e-01
-4.69232082e-01 3.57969075e-01 5.49400282e+00 -1.87328324e-01
-1.18511546e+00 -5.58331668e-01 1.65119588e-01 9.20194909e-02
4.17300351e-02 5.75212479e-01 -3.23794931e-01 4.38691437e-01
1.10656321e+00 -1.33555126e+00 -9.89390463e-02 9.02838051e-01
6.59028292e-01 -6.80954039e-01 -9.00301933e-01 5.90761900e-01
-4.18893397e-01 -1.60373890e+00 -3.99776608e-01 4.04727459e-01
3.88719857e-01 -1.50003806e-01 -4.04475540e-01 4.35557872e-01
-2.49079302e-01 -5.80964088e-01 2.86271274e-01 5.92108011e-01
7.11853743e-01 -9.86595869e-01 7.39485204e-01 1.85933486e-01
-6.65177584e-01 2.10664362e-01 -3.26809704e-01 -5.65216959e-01
4.60387379e-01 8.49729002e-01 -8.54574561e-01 2.94764806e-02
-2.94560064e-02 -1.65324852e-01 -1.27765253e-01 1.49874818e+00
-3.90902102e-01 3.55932713e-01 -3.08522791e-01 9.80810449e-02
-3.86124142e-02 -5.46964467e-01 2.25608975e-01 5.90312183e-01
-3.95826578e-01 6.57663643e-01 2.03626052e-01 9.95933294e-01
-1.13965526e-01 8.31133127e-01 -2.96215922e-01 1.60167322e-01
1.61703333e-01 9.72143769e-01 -8.24776292e-01 -5.38569510e-01
-2.70993233e-01 3.47012341e-01 -6.36618674e-01 6.24407232e-01
-2.08746701e-01 -8.60247016e-01 5.70263028e-01 4.09622848e-01
3.06864958e-02 -7.33463988e-02 -6.04445159e-01 -3.98777187e-01
-3.30255657e-01 -4.59639460e-01 3.63945693e-01 -3.34088504e-02
-3.07384521e-01 -4.15263735e-02 2.36125931e-01 -7.22526670e-01
-9.35531035e-02 2.26779178e-01 -6.41972601e-01 1.54835272e+00
-1.30036104e+00 -3.71540576e-01 1.39191616e-02 -2.26710930e-01
-3.96685936e-02 5.23806214e-01 1.36567581e+00 1.38362991e-02
-2.80694842e-01 4.42123532e-01 4.86179106e-02 -2.91841608e-02
7.25616038e-01 -1.29378390e+00 -1.71590596e-01 6.85477078e-01
-4.81517076e-01 1.20523882e+00 1.06522727e+00 -7.82952785e-01
-5.33371568e-01 -1.05045962e+00 1.46337700e+00 3.67906243e-01
-1.19417813e-02 -7.34770745e-02 -5.48177004e-01 4.60974663e-01
-1.17244758e-01 -2.22614542e-01 1.24975932e+00 -2.33427241e-01
-1.33389920e-01 -1.63190961e-02 -1.40373421e+00 7.89189100e-01
-8.38125125e-03 -3.56115624e-02 -2.19575763e-01 1.27133399e-01
1.57831296e-01 -4.55827832e-01 -1.37914717e+00 4.03703451e-01
1.24383819e+00 -5.62114358e-01 9.25684154e-01 -5.85202754e-01
2.34728456e-01 -7.43269503e-01 4.83819246e-01 -9.46720958e-01
-2.97954977e-01 -8.11041474e-01 5.13232708e-01 7.67102361e-01
4.36093986e-01 -6.47819996e-01 4.22012359e-01 9.89433408e-01
-1.15420535e-01 -4.03428793e-01 -8.65758598e-01 -8.56120706e-01
2.15945706e-01 3.77382457e-01 -1.03078643e-02 7.74651527e-01
7.92711079e-01 5.39846063e-01 -1.73344508e-01 -1.49881348e-01
7.65211403e-01 4.75783944e-01 7.61606693e-01 -1.29603505e+00
3.36012423e-01 -1.82203189e-01 -4.47361380e-01 -5.80112398e-01
-2.51119882e-01 -9.31654513e-01 -7.73138046e-01 -1.61123729e+00
-1.28422603e-01 -1.26142278e-01 -3.25486809e-02 3.70851517e-01
4.44416136e-01 -4.70941775e-02 1.46323368e-01 -4.14701581e-01
3.24593008e-01 4.67633307e-01 8.04248691e-01 4.25584435e-01
-1.14832962e+00 2.12644294e-01 -8.74743879e-01 4.75729167e-01
1.20633507e+00 -6.24882877e-01 -6.17997169e-01 8.45477045e-01
-2.19635367e-02 5.99093437e-01 -1.81587294e-01 -6.21505558e-01
-1.19225524e-01 -4.10121351e-01 8.11530590e-01 -4.26754028e-01
-1.32292047e-01 -6.00323141e-01 5.28068781e-01 1.25826502e+00
-1.85423702e-01 -7.96583652e-01 2.74450004e-01 9.42054316e-02
2.37382755e-01 -4.84094769e-01 1.28505993e+00 -5.63872278e-01
-4.54929531e-01 -1.28258631e-01 -1.38254428e+00 3.68627459e-02
8.76601994e-01 -3.76476169e-01 -1.40957415e-01 -2.52446353e-01
-1.22509551e+00 4.31941330e-01 4.10433680e-01 6.54306710e-02
1.45893827e-01 -1.10957503e+00 -6.94305003e-01 2.77123123e-01
1.71669692e-01 -5.20250440e-01 8.06863368e-01 1.24789107e+00
-1.03069019e+00 5.98588049e-01 -4.51756984e-01 -3.26155722e-01
-8.90176892e-01 6.83787286e-01 7.16450036e-01 1.59945592e-01
-8.73040020e-01 1.89234748e-01 2.19443604e-01 2.85938472e-01
-2.48646364e-02 -2.87448734e-01 -3.08609486e-01 -3.26395720e-01
1.00501513e+00 4.35688466e-01 -2.48698279e-01 4.85271849e-02
-8.18661600e-02 1.66005731e-01 -2.85926878e-01 9.58600864e-02
8.19056451e-01 -1.90057173e-01 -2.60407031e-01 1.95726231e-01
7.48111844e-01 2.40338892e-01 -4.20887291e-01 -1.34338766e-01
-4.53827143e-01 -5.82580090e-01 -2.59790927e-01 -1.28226912e+00
-6.69708848e-01 1.42965792e-02 6.76776111e-01 -5.18989801e-01
5.35961032e-01 -1.29537299e-01 4.61587042e-01 -5.96012883e-02
1.40860721e-01 -8.68842542e-01 -5.14743328e-01 -2.96588242e-01
9.36631322e-01 1.54943883e-01 3.79572660e-01 -7.28087544e-01
-3.09456646e-01 7.79962242e-01 5.76173067e-01 -9.92668644e-02
3.98216546e-01 1.02572599e-02 4.34789062e-01 -1.14861026e-01
-4.93507892e-01 -4.59603965e-03 -3.27787578e-01 4.38916355e-01
5.60856581e-01 3.05463910e-01 -1.87550032e+00 5.33087969e-01
2.30572205e-02 5.59068501e-01 1.02478552e+00 7.63046741e-01
-8.16105127e-01 -1.17962146e+00 -5.19344071e-03 3.12202692e-01
-6.06573105e-01 1.21642515e-01 1.08152397e-01 8.31065893e-01
-2.24296972e-01 2.69165635e-01 -2.31006250e-01 6.05591416e-01
9.26786304e-01 7.32912779e-01 5.05521595e-01 -2.96205521e-01
2.76616458e-02 6.16308093e-01 6.73354417e-02 -3.05407226e-01
-1.76426083e-01 -7.32626140e-01 -1.62777948e+00 -2.29183510e-01
-5.54802753e-02 5.09689808e-01 7.30491996e-01 5.71194351e-01
-4.64857481e-02 1.19827827e-02 5.82161069e-01 -2.72494014e-02
-8.25560331e-01 -1.07700288e+00 -1.49063122e+00 -4.86364737e-02
-9.90504399e-03 -4.20056939e-01 -3.22778016e-01 5.32855727e-02] | [13.993151664733887, 2.897386074066162] |
a807340c-552d-4638-a340-0b731c77ba98 | steady-state-analysis-of-networked-epidemic | 2305.19023 | null | https://arxiv.org/abs/2305.19023v1 | https://arxiv.org/pdf/2305.19023v1.pdf | Steady-state analysis of networked epidemic models | Compartmental epidemic models with dynamics that evolve over a graph network have gained considerable importance in recent years but analysis of these models is in general difficult due to their complexity. In this paper, we develop two positive feedback frameworks that are applicable to the study of steady-state values in a wide range of compartmental epidemic models, including both group and networked processes. In the case of a group (resp. networked) model, we show that the convergence limit of the susceptible proportion of the population (resp. the susceptible proportion in at least one of the subgroups) is upper bounded by the reciprocal of the basic reproduction number (BRN) of the model. The BRN, when it is greater than unity, thus demonstrates the level of penetration into a subpopulation by the disease. Both non-strict and strict bounds on the convergence limits are derived and shown to correspond to substantially distinct scenarios in the epidemic processes, one in the presence of the endemic state and another without. Formulae for calculating the limits are provided in the latter case. We apply the developed framework to examining various group and networked epidemic models commonly seen in the literature to verify the validity of our conclusions. | ['Lanlan Su', 'Sei Zhen Khong'] | 2023-05-30 | null | null | null | null | ['unity'] | ['computer-vision'] | [ 1.89962953e-01 2.34842598e-01 2.36952409e-01 5.37077665e-01
5.89991212e-01 -6.85551047e-01 6.31820560e-01 3.96568745e-01
-4.35472101e-01 9.35663223e-01 -1.89681917e-01 -4.36077505e-01
-7.05134451e-01 -8.13377023e-01 -3.68393391e-01 -1.18647838e+00
-4.51905340e-01 5.16742229e-01 5.63174665e-01 -6.27026737e-01
-1.48364797e-01 5.40938854e-01 -1.20621729e+00 -7.02898562e-01
8.46646786e-01 6.44610226e-01 -3.14528048e-02 7.56613314e-01
-1.37238324e-01 3.98435414e-01 -9.24977422e-01 -2.19523698e-01
3.38217944e-01 -4.90366250e-01 -1.55029193e-01 -6.29892424e-02
-8.05072546e-01 9.76085514e-02 -1.40483469e-01 9.18162525e-01
3.63227457e-01 -2.68866181e-01 9.55016673e-01 -1.63119686e+00
-1.68562993e-01 3.93192559e-01 -7.59627759e-01 4.30533230e-01
8.65213573e-02 -5.23911193e-02 3.19157541e-01 3.54092084e-02
7.00476885e-01 7.71826267e-01 6.03110015e-01 3.63928288e-01
-1.10117757e+00 -4.91221368e-01 -4.46880274e-02 -4.35611278e-01
-1.41055262e+00 -1.23130634e-01 3.86760354e-01 -6.78138316e-01
4.38918173e-01 3.09420407e-01 9.33965862e-01 5.30156016e-01
5.71277559e-01 -2.69380748e-01 9.80571806e-01 -3.06415439e-01
3.50399464e-01 1.80238232e-01 6.58435524e-02 1.30299851e-01
1.00861621e+00 -8.91162381e-02 1.15392119e-01 -5.65781474e-01
1.00281453e+00 -4.50096689e-02 -3.99831057e-01 -5.68310678e-01
-7.50080228e-01 6.65039778e-01 9.45248157e-02 6.25698090e-01
-6.80667758e-01 -4.13624989e-03 1.88726455e-01 5.71925581e-01
6.13692164e-01 -5.53831160e-02 -1.45936057e-01 1.08578771e-01
-5.55055678e-01 9.51443315e-02 1.26615512e+00 6.60174906e-01
2.18850955e-01 -1.57055289e-01 3.14328909e-01 3.05096537e-01
1.46101534e-01 6.39483690e-01 -1.09656125e-01 -3.87470454e-01
1.57583013e-01 6.96914673e-01 4.67666835e-01 -7.26184726e-01
-4.24397975e-01 -5.63942075e-01 -1.28120434e+00 3.49249691e-02
6.83245599e-01 -4.79183584e-01 -3.66833329e-01 1.90674722e+00
4.21170682e-01 -4.84972335e-02 3.41388375e-01 3.09072047e-01
1.57706678e-01 7.56901324e-01 2.00147003e-01 -9.15373564e-01
9.05353308e-01 3.25018875e-02 -6.65505826e-01 3.71789068e-01
3.46529275e-01 -3.59056890e-01 8.65597576e-02 -5.69035709e-02
-1.03898418e+00 2.99296468e-01 -7.02332616e-01 8.83740366e-01
-3.35652679e-01 -5.42366803e-01 4.48453166e-02 7.11670280e-01
-1.40737271e+00 2.52783269e-01 -6.34141743e-01 -9.29091275e-01
-2.58785278e-01 5.08732438e-01 -3.20026018e-02 1.81478962e-01
-1.22716212e+00 5.39946198e-01 -7.96284676e-02 3.65049005e-01
-2.82363147e-01 -4.81145650e-01 -2.04493985e-01 1.97609007e-01
2.88217813e-01 -7.10944474e-01 5.75972855e-01 -1.06032646e+00
-8.64140928e-01 5.36008179e-01 1.92012936e-01 -4.07395005e-01
1.00794494e+00 5.29722929e-01 -2.83041060e-01 1.44875497e-01
-7.54627213e-02 2.05292366e-02 4.07904595e-01 -1.12784338e+00
-2.48865455e-01 -4.41637248e-01 2.11927578e-01 1.82723731e-01
-2.24009767e-01 2.49831304e-01 -2.51838416e-02 -2.33681053e-01
-2.21103832e-01 -8.74059558e-01 -4.14797425e-01 -5.11457399e-02
-2.27516457e-01 -2.14414403e-01 6.85892045e-01 -1.76053181e-01
1.13920772e+00 -2.04887056e+00 2.87180066e-01 4.39877272e-01
3.18652838e-01 2.29068175e-01 1.64202690e-01 1.13911390e+00
1.48976102e-01 2.01736808e-01 -4.07551944e-01 1.81081116e-01
-2.97669321e-01 1.46973938e-01 -3.92426476e-02 8.87640595e-01
8.23520347e-02 2.12613329e-01 -7.29014039e-01 -2.74466246e-01
-1.77922130e-01 6.85021341e-01 -1.96797833e-01 -7.99211189e-02
2.45705575e-01 3.93655360e-01 -5.00262320e-01 -3.55703086e-02
6.94477618e-01 -1.86342180e-01 5.13880610e-01 4.92132902e-01
-4.37964797e-01 -4.29048330e-01 -1.26083243e+00 8.33738297e-02
-1.73187986e-01 3.38234633e-01 5.61942041e-01 -8.40553939e-01
6.77019238e-01 5.29135942e-01 6.89260066e-01 -5.67653589e-02
3.56626719e-01 1.89974770e-01 5.65750480e-01 -1.71399072e-01
1.77097261e-01 -4.64200199e-01 1.41513944e-01 6.74110293e-01
-3.26575965e-01 3.01400959e-01 6.11813486e-01 3.83386850e-01
1.11580741e+00 -6.31882548e-01 5.43099165e-01 -7.99874723e-01
6.75202072e-01 -2.04556316e-01 3.54813695e-01 4.79161829e-01
-4.02610064e-01 -2.14605927e-01 1.31220734e+00 3.26100379e-01
-1.13855064e+00 -9.70428705e-01 -1.26124844e-01 4.38828498e-01
5.91241121e-01 1.60173830e-02 -8.26620936e-01 2.15406418e-01
1.64578438e-01 1.44834980e-01 -8.11262727e-01 -1.51215091e-01
-3.72047365e-01 -1.14632416e+00 3.60936701e-01 1.02416631e-02
2.83922762e-01 -9.27044749e-01 -7.16130972e-01 1.76389128e-01
2.67116129e-01 -8.12622547e-01 -1.03820637e-01 3.82319242e-02
-7.91125536e-01 -1.43377948e+00 -9.90568399e-01 -5.88851154e-01
7.62791455e-01 1.22220740e-01 5.23237646e-01 3.16294372e-01
-5.09517118e-02 8.17261636e-01 -8.60244706e-02 -2.61846930e-01
-1.01537240e+00 6.05415218e-02 1.15419962e-01 1.01781368e-01
-9.93746594e-02 -3.68951529e-01 -6.58640385e-01 5.91493428e-01
-1.10169351e+00 -2.89938092e-01 3.97279486e-02 8.36288929e-02
1.17391475e-01 1.28400102e-01 9.80266511e-01 -4.26113665e-01
1.12293625e+00 -1.03182232e+00 -8.65545750e-01 2.58065790e-01
-3.64897490e-01 -2.40909770e-01 6.85549915e-01 -5.64219892e-01
-7.27023661e-01 -2.87041932e-01 3.78645986e-01 1.92418486e-01
8.31657127e-02 3.83366138e-01 1.23059258e-01 -1.48376003e-01
1.47092879e-01 3.35778445e-01 3.31725627e-01 -1.40567049e-01
9.15191770e-02 7.52139807e-01 9.11791101e-02 -4.04289961e-02
8.47644866e-01 5.26771486e-01 4.60588247e-01 -1.34414661e+00
3.31212789e-01 -3.82993132e-01 -3.89194787e-01 -5.63896537e-01
5.65661848e-01 -5.40808618e-01 -1.00683689e+00 7.45909750e-01
-8.61453354e-01 -2.83120215e-01 -3.38178933e-01 1.31604925e-01
-5.19133091e-01 2.99917847e-01 -6.82171404e-01 -1.57433009e+00
-2.48641208e-01 -7.01315284e-01 2.58839488e-01 2.55171895e-01
1.87255919e-01 -1.33483708e+00 4.35298562e-01 -5.36730349e-01
7.57035613e-01 5.53932667e-01 1.01961625e+00 -7.61482954e-01
-2.62125850e-01 -4.38577324e-01 -1.53903682e-02 -2.14554891e-01
1.28423557e-01 3.28365922e-01 -1.19617514e-01 -3.56162816e-01
1.87421829e-01 5.03139913e-01 3.76229912e-01 5.64736247e-01
-1.07722498e-01 -3.14028710e-01 -6.75783753e-01 -7.32656121e-02
1.71533835e+00 4.35258657e-01 4.44461703e-01 7.68071190e-02
9.34553891e-03 8.88681471e-01 9.42105651e-02 6.35978043e-01
1.39769735e-02 4.80357796e-01 3.25721830e-01 1.35790497e-01
3.51994157e-01 3.31865430e-01 2.65815139e-01 8.16045523e-01
-3.31684768e-01 -7.63882875e-01 -9.00007069e-01 6.03182852e-01
-1.59100056e+00 -8.35924149e-01 -5.41452467e-01 2.40586066e+00
5.21511614e-01 1.01116583e-01 7.20486581e-01 6.83490708e-02
1.18316698e+00 -9.74551812e-02 -3.38989049e-01 -4.24664736e-01
-2.95373261e-01 -2.75169581e-01 6.96387529e-01 6.15603268e-01
-5.23450434e-01 2.10467547e-01 6.81495094e+00 2.38904342e-01
-9.25048828e-01 8.06358084e-03 3.94955963e-01 1.27859905e-01
-8.97954497e-03 -7.48721957e-02 -4.37332213e-01 6.00167096e-01
1.15534627e+00 -7.76417673e-01 2.78407842e-01 2.40305305e-01
4.53885108e-01 -3.61170799e-01 -2.75746435e-01 1.87543824e-01
-3.89197260e-01 -6.89008653e-01 -2.72098243e-01 5.52287579e-01
5.95377803e-01 -1.88275903e-01 -9.32632312e-02 -2.08205536e-01
3.55448544e-01 -4.67206389e-01 2.23500788e-01 4.99468386e-01
6.96606517e-01 -8.14376295e-01 8.41673493e-01 6.45021915e-01
-1.27968943e+00 -2.43182238e-02 -7.36677051e-02 -1.58716902e-01
5.75891018e-01 4.67581570e-01 -4.19090807e-01 4.24673676e-01
2.26697966e-01 2.98923939e-01 -2.48613417e-01 1.39659274e+00
3.44516754e-01 4.34200555e-01 -7.68838644e-01 -2.67177016e-01
3.38018164e-02 -6.98126197e-01 7.92768419e-01 6.70919776e-01
2.74374187e-01 2.03136399e-01 -3.64550322e-01 6.86997235e-01
2.04313383e-01 2.09470630e-01 -6.47850096e-01 -9.88198966e-02
2.79356331e-01 1.00371480e+00 -1.17569447e+00 -9.98257920e-02
-1.26347169e-01 5.76413393e-01 -8.95522088e-02 5.58303654e-01
-6.36333346e-01 -4.78535324e-01 7.20869958e-01 5.41610241e-01
2.57563472e-01 -1.99839875e-01 2.37533927e-01 -7.20329165e-01
-2.88662672e-01 -1.30085379e-01 5.30359475e-03 -2.20555738e-01
-1.15652096e+00 7.15461850e-01 3.98461968e-01 -8.23371887e-01
-4.55059439e-01 -2.73616612e-01 -7.19790220e-01 9.53713894e-01
-1.01704800e+00 -4.89940375e-01 -1.48138758e-02 5.11445403e-01
-2.36402705e-01 1.75292745e-01 3.58783513e-01 2.30675876e-01
-8.60360742e-01 5.12987599e-02 7.04488039e-01 -2.34321773e-01
1.56922206e-01 -7.70326376e-01 -1.82574987e-02 6.83989704e-01
-7.82127142e-01 7.20989287e-01 1.13438940e+00 -1.02281225e+00
-1.09357345e+00 -6.62224591e-01 8.49928975e-01 7.61767179e-02
1.02147484e+00 -4.91023421e-01 -6.22964978e-01 3.42713594e-01
2.35520408e-01 -2.86036938e-01 4.07621980e-01 -4.98496622e-01
3.08758974e-01 1.74410000e-01 -1.32985795e+00 6.81207120e-01
7.92810619e-01 6.13688678e-02 1.51859909e-01 2.37802371e-01
4.18359101e-01 3.34842235e-01 -8.89147699e-01 1.96750015e-01
4.57515717e-01 -8.93723965e-01 5.32479227e-01 -5.35253167e-01
-1.55047581e-01 -2.30248272e-01 2.83374399e-01 -1.12304378e+00
-8.76200274e-02 -8.01446199e-01 9.56071317e-02 1.16443086e+00
1.60949335e-01 -1.21639884e+00 2.79090494e-01 1.65998802e-01
7.98654497e-01 -8.67921770e-01 -1.22335291e+00 -9.98957694e-01
3.72974932e-01 4.32735413e-01 2.94809222e-01 4.77653772e-01
1.94269016e-01 1.70838878e-01 -8.35804790e-02 6.42229788e-05
5.68438053e-01 -3.67985398e-01 3.91224891e-01 -1.61598468e+00
2.48535462e-02 -3.95776093e-01 -5.74896455e-01 -4.18299496e-01
-2.19261214e-01 -3.30415517e-01 -2.01868236e-01 -1.45819247e+00
9.72032025e-02 -6.63493693e-01 -1.23454429e-01 -3.05890590e-01
1.90398216e-01 4.77902517e-02 1.59154594e-01 5.55898070e-01
-1.02513574e-01 2.38890290e-01 8.54153275e-01 5.00092566e-01
-4.31252211e-01 3.17328423e-01 -4.56964821e-01 5.70205808e-01
8.50255132e-01 -5.62331915e-01 -5.93118608e-01 4.63712901e-01
5.43192148e-01 3.58410895e-01 3.22731316e-01 -7.46491611e-01
2.56949723e-01 -9.14695412e-02 -2.32091516e-01 -3.60778511e-01
1.40124425e-01 -1.01395833e+00 7.03090787e-01 9.37544107e-01
-6.33490682e-02 2.97068417e-01 5.09595610e-02 1.10031080e+00
3.59629363e-01 -2.56004035e-01 6.90160275e-01 1.60843223e-01
2.20731959e-01 7.00738579e-02 -9.18564677e-01 -4.62011434e-03
1.51374447e+00 -3.56881440e-01 -4.11053658e-01 -8.33166540e-01
-7.83372402e-01 2.21619219e-01 6.17733419e-01 -4.12479118e-02
6.41553178e-02 -6.07927203e-01 -4.76647049e-01 -4.91856448e-02
-3.02666157e-01 -7.11600542e-01 2.52687871e-01 1.42574954e+00
-6.18403971e-01 3.46737772e-01 -4.18477476e-01 -2.50617862e-01
-1.19907153e+00 7.41221845e-01 7.35831678e-01 -3.97789866e-01
-1.67975724e-01 -2.32686996e-02 2.48375326e-01 8.55857208e-02
6.58507720e-02 5.99073060e-02 -3.93469721e-01 2.10639507e-01
2.75885880e-01 6.22653186e-01 -4.01237071e-01 -1.03116000e+00
-4.12281662e-01 5.58692694e-01 2.54170716e-01 -2.17843786e-01
1.28533328e+00 -3.23320508e-01 -4.62010086e-01 5.08398294e-01
6.44008458e-01 6.88580275e-02 -1.07279813e+00 1.68932617e-01
-1.40193000e-01 1.93183467e-01 -4.76855069e-01 -4.38427836e-01
-9.71010387e-01 3.17165732e-01 3.60055476e-01 1.18321848e+00
1.10278594e+00 1.14688262e-01 8.63774046e-02 -1.75380901e-01
3.61841619e-01 -4.16689306e-01 -6.06275141e-01 2.55951405e-01
7.01768994e-01 -3.60263705e-01 -1.77403316e-01 -8.39521766e-01
-2.80757129e-01 9.15422380e-01 1.81908011e-01 -5.04671454e-01
1.10181785e+00 5.28354108e-01 -4.42361504e-01 -1.53422337e-02
-7.45131075e-01 -2.54056871e-01 -3.11752409e-01 7.97808170e-01
2.52248198e-01 -5.55867814e-02 -1.02618897e+00 2.05683619e-01
2.51775652e-01 1.70738641e-02 9.65967119e-01 8.37664545e-01
-3.84279966e-01 -8.98732543e-01 -4.12849069e-01 3.46667230e-01
-3.79825622e-01 -6.52053533e-03 -7.25918889e-01 1.31467569e+00
-1.16588183e-01 9.60107207e-01 3.35026562e-01 2.99066097e-01
1.88301116e-01 -1.49712309e-01 3.92503470e-01 -9.68729705e-02
-6.48302257e-01 5.01303263e-02 -1.76092044e-01 3.31012785e-01
-4.49695051e-01 -6.04351819e-01 -9.32343185e-01 -7.70281017e-01
-5.51837265e-01 3.74354005e-01 4.59487796e-01 7.28331804e-01
1.81903243e-01 2.07211182e-01 3.47065508e-01 -2.90338933e-01
-3.12201917e-01 -9.17601764e-01 -1.15495896e+00 -1.31574780e-01
3.63149762e-01 -5.71145952e-01 -8.78874421e-01 -4.03079987e-01] | [5.934700012207031, 4.352358341217041] |
c4f33969-125b-4c37-a369-53bdfe9da13a | contrastive-loss-is-all-you-need-to-recover | 2306.08221 | null | https://arxiv.org/abs/2306.08221v1 | https://arxiv.org/pdf/2306.08221v1.pdf | Contrastive Loss is All You Need to Recover Analogies as Parallel Lines | While static word embedding models are known to represent linguistic analogies as parallel lines in high-dimensional space, the underlying mechanism as to why they result in such geometric structures remains obscure. We find that an elementary contrastive-style method employed over distributional information performs competitively with popular word embedding models on analogy recovery tasks, while achieving dramatic speedups in training time. Further, we demonstrate that a contrastive loss is sufficient to create these parallel structures in word embeddings, and establish a precise relationship between the co-occurrence statistics and the geometric structure of the resulting word embeddings. | ['Nakul Verma', 'Fei-Tzin Lee', 'Narutatsu Ri'] | 2023-06-14 | null | null | null | null | ['word-embeddings'] | ['methodology'] | [-2.17968449e-01 -4.20709029e-02 -4.25459236e-01 -2.28471160e-01
-4.32855785e-01 -6.98072791e-01 9.20918763e-01 6.94873333e-01
-6.70499325e-01 2.63892710e-01 7.66343474e-01 -6.50545597e-01
-2.28072539e-01 -7.97021091e-01 -5.47820807e-01 -4.25058931e-01
-3.18617672e-01 5.08422434e-01 -9.25073698e-02 -4.74193990e-01
6.62008405e-01 6.31955206e-01 -1.11989379e+00 -1.61560461e-01
5.77363372e-01 4.74702001e-01 -7.45251030e-02 7.75573790e-01
-6.18365705e-01 1.12207294e-01 -2.40273878e-01 -5.73619366e-01
2.85594761e-01 -1.68501839e-01 -8.49800885e-01 -3.41089457e-01
7.79507995e-01 -3.39196026e-01 -8.35337758e-01 1.02054965e+00
3.07032973e-01 4.69998896e-01 1.17214882e+00 -8.22303593e-01
-1.63492215e+00 2.26441905e-01 -5.69043279e-01 6.64991319e-01
4.89891559e-01 -1.47766760e-02 1.77056730e+00 -1.18635917e+00
4.46800649e-01 1.39289713e+00 8.70829999e-01 3.61185700e-01
-1.55552447e+00 -2.40428254e-01 9.19801444e-02 1.76287308e-01
-1.40827978e+00 -2.45408028e-01 9.01280284e-01 -3.23174238e-01
1.23507559e+00 4.17651795e-02 7.59750843e-01 1.00586379e+00
4.88361061e-01 3.97644311e-01 7.08588719e-01 -8.17696214e-01
3.43979225e-02 9.49289426e-02 5.62818110e-01 7.98635840e-01
5.99153876e-01 1.35936990e-01 -4.63920385e-01 -5.21161675e-01
7.33205795e-01 2.35511214e-01 -1.94978416e-01 -5.32232106e-01
-6.11969769e-01 1.25601542e+00 6.60503089e-01 4.97327447e-01
-7.96712115e-02 3.15637529e-01 5.15993357e-01 3.41257513e-01
5.93643606e-01 9.08669710e-01 -2.18705133e-01 2.07730368e-01
-2.49488071e-01 5.59157550e-01 4.23221618e-01 6.20773196e-01
7.23002434e-01 -1.44647181e-01 3.54196578e-01 8.38751912e-01
2.09952280e-01 2.47587532e-01 8.37491393e-01 -6.29721463e-01
2.71021605e-01 3.40280026e-01 -6.66957349e-02 -1.29190516e+00
-1.22325718e-01 -2.22783595e-01 -2.71775156e-01 7.46866465e-02
5.16388655e-01 2.09245890e-01 -3.66834551e-01 2.06448531e+00
5.16668670e-02 2.72921830e-01 2.06540927e-01 6.42330587e-01
2.72256821e-01 6.78453803e-01 3.17851722e-01 2.68422008e-01
1.45103920e+00 -6.47483826e-01 -4.72479522e-01 -4.37165767e-01
1.00836313e+00 -7.40894496e-01 1.59206712e+00 -4.76331621e-01
-1.06258070e+00 -4.82544184e-01 -1.30810332e+00 -6.33662283e-01
-6.39756620e-01 -5.97231507e-01 9.10334229e-01 4.40788805e-01
-8.20682287e-01 8.02685261e-01 -5.04006803e-01 -4.48512077e-01
2.52751738e-01 9.49522480e-02 -5.07240772e-01 -4.04021479e-02
-1.26314878e+00 1.27803135e+00 2.35145092e-01 -4.10641611e-01
1.24418288e-01 -1.21265125e+00 -1.06760323e+00 1.18090279e-01
-2.85596162e-01 -6.54293537e-01 1.00142026e+00 -4.13726807e-01
-9.26012933e-01 8.26154649e-01 -2.92524993e-01 -4.66480613e-01
-3.37770469e-02 -1.64179489e-01 -4.61196095e-01 -7.17722550e-02
5.27940169e-02 5.11753559e-01 4.28010672e-01 -9.27086055e-01
-2.29027137e-01 -5.02570093e-01 2.29989320e-01 3.90383154e-01
-9.28744614e-01 -1.58983544e-01 3.04089725e-01 -8.79823506e-01
-4.17535640e-02 -7.78688371e-01 -3.55574042e-01 6.79168820e-01
1.77267492e-01 -7.98603475e-01 5.73322177e-01 -3.23768318e-01
1.09143400e+00 -2.44812274e+00 -1.34056315e-01 2.56461561e-01
3.71997148e-01 1.80902496e-01 -5.23021162e-01 6.61656916e-01
-2.35658318e-01 1.52725220e-01 -1.73535436e-01 -9.85673442e-02
3.38875085e-01 3.90008003e-01 -8.76042902e-01 6.42572880e-01
3.13058019e-01 1.07140756e+00 -1.20474339e+00 -2.95059741e-01
2.15329006e-01 2.87345499e-01 -7.37870693e-01 1.03207827e-01
6.68280497e-02 -4.64814097e-01 -3.39269876e-01 -9.64491665e-02
5.03475308e-01 -8.48039147e-03 3.45168889e-01 -2.51302093e-01
1.89505070e-01 8.32459331e-01 -7.63718724e-01 1.83964527e+00
-6.73171222e-01 9.71654892e-01 -4.67063189e-01 -1.12497163e+00
6.93435490e-01 1.04110852e-01 7.83999562e-02 -5.96666038e-01
2.62685958e-02 3.67933810e-01 2.46049196e-01 -2.55915612e-01
9.45427537e-01 -6.64771795e-01 -7.50426576e-02 8.41727376e-01
-6.84360228e-03 -1.36490762e-01 -2.20737830e-02 2.07779169e-01
1.02032101e+00 -2.45108649e-01 3.25763881e-01 -7.87604570e-01
8.83874968e-02 2.09555551e-02 -5.07380255e-02 4.55009580e-01
-2.49607041e-01 3.02376926e-01 2.96058416e-01 -6.99333429e-01
-1.53099632e+00 -1.63651335e+00 -6.06180012e-01 1.04734778e+00
4.21740919e-01 -4.51246232e-01 -2.96058059e-01 -4.29927409e-01
4.34360892e-01 1.07429492e+00 -8.70447993e-01 -5.75305820e-01
-6.44893587e-01 -6.67936921e-01 6.92724943e-01 9.87930119e-01
-3.03167075e-01 -5.70879877e-01 -2.05282465e-01 2.03518182e-01
4.17828113e-01 -8.71629000e-01 -8.76115739e-01 2.12384462e-01
-1.05768669e+00 -1.05550063e+00 -4.33635622e-01 -1.08334446e+00
6.53445423e-01 4.08499271e-01 1.16359651e+00 -2.04369496e-03
-6.83327436e-01 2.08485425e-01 -5.88780046e-02 -2.97329336e-01
-2.14244723e-01 -1.91841453e-01 4.01447356e-01 -6.74514532e-01
9.59254503e-01 -8.40797305e-01 -5.55468380e-01 -1.79264262e-01
-9.49126422e-01 -6.49206221e-01 1.14692509e-01 9.58374679e-01
2.55183786e-01 -3.96718204e-01 4.20818150e-01 -6.72776997e-01
1.27145815e+00 -5.43909431e-01 -1.93156064e-01 1.71161443e-01
-8.21064413e-01 6.40406132e-01 7.54189849e-01 -7.60518193e-01
-4.86969322e-01 -5.73872626e-01 -3.69248599e-01 -3.01274776e-01
1.07578062e-01 2.57122606e-01 3.34826350e-01 4.72374372e-02
8.36432338e-01 1.00896835e-01 1.46980718e-01 -4.68668193e-01
9.75115180e-01 3.96624953e-01 4.35391843e-01 -8.86391044e-01
9.06482577e-01 3.74721438e-01 1.65372957e-02 -8.55082035e-01
-7.95067012e-01 -3.62152219e-01 -5.50191879e-01 8.66688788e-01
7.65042901e-01 -6.49217308e-01 -2.30342984e-01 -4.17884916e-01
-1.53278434e+00 1.11955762e-01 -6.65036261e-01 6.28870785e-01
-4.01592553e-01 5.25123060e-01 -7.62265384e-01 -5.37305534e-01
-1.82015851e-01 -5.53095460e-01 6.01927400e-01 -1.73558012e-01
-8.86546254e-01 -1.71757150e+00 6.55552447e-01 -3.70372266e-01
3.87682438e-01 -3.27546418e-01 1.85555553e+00 -9.28048313e-01
2.47063860e-02 -4.51719075e-01 -4.14234698e-01 8.55568498e-02
2.33974889e-01 -1.99118569e-01 -7.30679035e-01 -4.20025811e-02
-1.78517714e-01 -1.26938939e-01 5.87516129e-01 7.09640086e-02
9.57665861e-01 -2.33885691e-01 -3.66110355e-01 5.13781130e-01
1.55283344e+00 -1.42744973e-01 3.80317867e-01 1.35661691e-01
6.25425398e-01 4.07863855e-01 -1.57239631e-01 -7.19117280e-03
2.59194747e-02 6.36544406e-01 -8.36260393e-02 -2.54872888e-02
-2.62907714e-01 -7.11023092e-01 9.37222093e-02 1.09841692e+00
2.13420734e-01 5.93724400e-02 -7.40347624e-01 8.98457646e-01
-1.42058122e+00 -9.58331585e-01 -3.98983322e-02 1.95348454e+00
7.43124545e-01 1.69577062e-01 -1.14376150e-01 1.64364561e-01
5.15511513e-01 5.52019536e-01 -5.44839442e-01 -1.01875091e+00
-2.04598695e-01 8.05914402e-01 5.45406520e-01 8.24326754e-01
-5.29408634e-01 9.49633539e-01 7.67702532e+00 8.70754182e-01
-7.97699630e-01 8.53133649e-02 1.32420495e-01 4.01258022e-02
-9.01982248e-01 -9.66180786e-02 -4.15494561e-01 9.89813134e-02
8.30841839e-01 -6.39208317e-01 1.67445928e-01 7.73673534e-01
-2.44318753e-01 5.04795134e-01 -1.53282654e+00 7.72174656e-01
1.07630163e-01 -1.54364789e+00 5.04869998e-01 2.26722844e-02
5.95672607e-01 7.51851313e-03 2.61680782e-01 3.77566367e-02
4.16533768e-01 -1.08737051e+00 3.14675361e-01 -1.23223685e-01
7.99375713e-01 -7.13716805e-01 4.22552764e-01 7.45478570e-02
-1.16745615e+00 -7.18631446e-02 -9.27344143e-01 -4.81915891e-01
1.04728065e-01 2.02826530e-01 -5.50382793e-01 1.25740036e-01
8.40156712e-03 4.11213964e-01 -2.68727869e-01 5.49644589e-01
-2.45958179e-01 1.98024839e-01 -1.64151311e-01 -3.67351234e-01
4.64169770e-01 -3.04067701e-01 4.66230065e-01 1.21934533e+00
2.28980184e-01 2.09950060e-01 -1.79565400e-01 9.77713466e-01
-2.45786786e-01 3.51880163e-01 -1.06936765e+00 -3.53339821e-01
6.81459486e-01 7.58833528e-01 -2.31454834e-01 -2.11473897e-01
-6.30501270e-01 1.01652873e+00 7.71138072e-01 2.48411134e-01
-5.71706474e-01 -7.08764791e-01 1.39932597e+00 3.27242553e-01
2.10842997e-01 -7.39295006e-01 -5.00799060e-01 -9.73142862e-01
3.38910259e-02 -2.78365701e-01 9.77149084e-02 -5.46626329e-01
-2.04076314e+00 5.46853304e-01 -1.64755642e-01 -6.97653353e-01
-1.24102794e-01 -1.22039425e+00 -1.05957317e+00 1.22129190e+00
-1.33749402e+00 -7.46339023e-01 3.99926364e-01 3.76778841e-01
5.95647395e-01 -4.95514721e-02 1.39202499e+00 7.42373541e-02
-3.39631647e-01 8.16906750e-01 3.38044614e-01 2.62146920e-01
5.59485555e-01 -1.40676022e+00 9.64753568e-01 3.74825656e-01
9.07359600e-01 1.11566234e+00 7.96968937e-01 -2.08584979e-01
-1.51429379e+00 -5.99250376e-01 1.21561766e+00 -7.55505383e-01
1.47251451e+00 -2.59584457e-01 -1.14470983e+00 5.27174115e-01
2.85098702e-01 2.20463067e-01 9.78226900e-01 6.11594975e-01
-1.14033759e+00 2.65671700e-01 -8.47759485e-01 9.93949175e-01
1.27920640e+00 -1.10873365e+00 -1.30178642e+00 4.95522797e-01
9.18697298e-01 1.08654253e-01 -7.24920452e-01 -8.39282125e-02
6.54394209e-01 -4.76948082e-01 1.39076650e+00 -1.36191165e+00
7.89194286e-01 1.78685233e-01 -4.65794861e-01 -1.42523444e+00
-4.62156683e-01 -2.95881689e-01 8.22414830e-02 6.60088718e-01
5.20189822e-01 -7.89451003e-01 6.45325243e-01 5.46183109e-01
-1.97814535e-02 -1.02673101e+00 -8.38464677e-01 -1.05478966e+00
9.63687360e-01 -3.27768505e-01 5.67210853e-01 1.14739156e+00
4.82126802e-01 6.75507963e-01 2.69430220e-01 -1.58357739e-01
5.68697989e-01 2.53265072e-02 2.69170076e-01 -1.18437481e+00
-3.03824037e-01 -8.24015617e-01 -6.61645532e-01 -1.37697709e+00
7.77579010e-01 -1.20504439e+00 -3.22717935e-01 -1.27380693e+00
6.23278990e-02 -6.67222261e-01 -5.24416924e-01 -1.42118320e-01
-2.64015287e-01 2.59894520e-01 -7.65242055e-02 1.84042260e-01
-6.41406327e-02 6.68828666e-01 8.27708364e-01 3.80905531e-02
6.96026161e-02 -5.29794693e-01 -9.39904451e-01 7.95665324e-01
6.54048681e-01 -3.87625188e-01 -5.48000157e-01 -9.73184347e-01
2.45812014e-01 -5.50432146e-01 2.43794858e-01 -4.28784281e-01
1.17107272e-01 -1.52695194e-01 2.99222231e-01 -2.64620893e-02
5.05419135e-01 -5.90053082e-01 -6.12791359e-01 5.48172176e-01
-7.56692767e-01 6.93753541e-01 3.48757476e-01 7.72824645e-01
4.82028797e-02 -5.55074394e-01 6.66564822e-01 2.34963655e-01
-6.39761746e-01 2.93226779e-01 -2.83823520e-01 4.71745372e-01
6.98781371e-01 -3.58219117e-01 -1.44492030e-01 -3.27150784e-02
-2.50982106e-01 -3.74588579e-01 5.91196299e-01 5.34830987e-01
3.81216109e-01 -1.72782874e+00 -5.40772080e-01 2.42438689e-01
2.55689412e-01 -6.28871024e-01 -5.62080629e-02 1.37522265e-01
-4.56015736e-01 2.81253010e-01 -1.30263850e-01 2.55440827e-03
-8.62816930e-01 8.13764751e-01 -1.42594296e-02 -6.76624998e-02
-7.98647344e-01 1.05195355e+00 3.55212539e-01 -3.48295242e-01
-3.11239660e-02 -2.86444705e-02 2.99587876e-01 -1.34824380e-01
6.49547935e-01 2.29214504e-01 -2.53157347e-01 -3.69059324e-01
-3.03772151e-01 8.85173619e-01 -2.65514672e-01 -2.15054229e-01
1.30921316e+00 7.67975450e-02 2.32972261e-02 5.48946381e-01
1.81996274e+00 1.63701400e-01 -7.98148930e-01 -5.76324940e-01
1.29953593e-01 -7.28576005e-01 -1.01689599e-01 1.18968748e-01
-3.49388003e-01 1.19686544e+00 4.45274681e-01 3.52791309e-01
4.17198449e-01 2.80502528e-01 1.11096430e+00 6.26325846e-01
1.87330451e-02 -8.54052722e-01 1.40439346e-01 4.49586123e-01
6.41078889e-01 -9.18117642e-01 2.01516762e-01 -3.52911204e-01
-9.28780064e-02 1.14844418e+00 3.15199822e-01 -9.71816599e-01
7.83578932e-01 1.98566183e-01 -1.38344929e-01 -1.27290785e-01
-6.50511086e-01 7.18123093e-02 3.56683344e-01 7.77403772e-01
7.46152103e-01 3.65044773e-02 -4.99556124e-01 2.48305112e-01
-4.28773761e-01 -6.93906963e-01 1.50064364e-01 6.48162305e-01
-5.66937149e-01 -1.40022182e+00 9.18007270e-03 3.20231676e-01
-2.39526927e-01 -5.84079504e-01 -2.30601937e-01 8.22855175e-01
-2.44238660e-01 5.15321612e-01 7.16050148e-01 -1.99491009e-01
2.80823946e-01 2.91544110e-01 9.28371549e-01 -6.72106385e-01
-2.13816404e-01 -6.03149474e-01 -3.21518853e-02 -1.61843374e-01
2.41570234e-01 -3.01705301e-01 -1.28794527e+00 -7.29881465e-01
-2.86826134e-01 3.21804643e-01 6.12980962e-01 1.07308698e+00
2.92313486e-01 1.36462003e-01 5.62703609e-01 -5.92221618e-01
-1.08224678e+00 -8.24416041e-01 -5.94092906e-01 8.30144525e-01
2.26902336e-01 -6.34265125e-01 -5.61130166e-01 -2.89041668e-01] | [10.34595775604248, 8.711685180664062] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.