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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
3f6657a2-204d-43a5-ab8c-3ed26c2c939e | domain-adversarial-neural-networks-to-address | 1707.06183 | null | http://arxiv.org/abs/1707.06183v1 | http://arxiv.org/pdf/1707.06183v1.pdf | Domain-adversarial neural networks to address the appearance variability of histopathology images | Preparing and scanning histopathology slides consists of several steps, each
with a multitude of parameters. The parameters can vary between pathology labs
and within the same lab over time, resulting in significant variability of the
tissue appearance that hampers the generalization of automatic image analysis
methods. Typically, this is addressed with ad-hoc approaches such as staining
normalization that aim to reduce the appearance variability. In this paper, we
propose a systematic solution based on domain-adversarial neural networks. We
hypothesize that removing the domain information from the model representation
leads to better generalization. We tested our hypothesis for the problem of
mitosis detection in breast cancer histopathology images and made a comparative
analysis with two other approaches. We show that combining color augmentation
with domain-adversarial training is a better alternative than standard
approaches to improve the generalization of deep learning methods. | ['Mitko Veta', 'Pim Moeskops', 'Maxime W. Lafarge', 'Josien P. W. Pluim', 'Koen A. J. Eppenhof'] | 2017-07-19 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [ 4.25902843e-01 5.45665137e-02 3.12166065e-01 -4.39633787e-01
-7.42477059e-01 -7.81212151e-01 2.86136717e-01 2.99071997e-01
-6.67461276e-01 7.72169054e-01 -3.39365304e-01 -4.99509454e-01
2.64023364e-01 -5.96055388e-01 -6.72855496e-01 -1.12884688e+00
2.30679765e-01 4.15534556e-01 3.49944443e-01 -1.34088531e-01
1.22277386e-01 9.74193335e-01 -9.49219048e-01 4.47822183e-01
4.97953773e-01 6.02557123e-01 -1.87909067e-01 8.72905374e-01
-3.48745406e-01 4.95877206e-01 -7.70889401e-01 -4.15054232e-01
2.99271375e-01 -4.20253426e-01 -8.87899637e-01 2.11997449e-01
4.26236570e-01 -1.12800397e-01 -5.35938181e-02 1.21282327e+00
4.54455614e-01 -1.51188105e-01 7.77351379e-01 -1.20813775e+00
-4.92308736e-01 2.35947832e-01 -5.70591211e-01 3.09745282e-01
-2.74879396e-01 1.56593442e-01 2.87347347e-01 -3.67349237e-01
7.91536868e-01 9.89272296e-01 8.05845857e-01 9.40461993e-01
-1.71843839e+00 -6.03449345e-01 -9.85468999e-02 1.45226300e-01
-1.42991734e+00 -1.32219627e-01 8.64443958e-01 -5.94421864e-01
2.83503205e-01 1.29525065e-01 6.04629159e-01 1.02503633e+00
2.76328236e-01 3.24721962e-01 1.52002347e+00 -6.35838330e-01
3.58943433e-01 3.10948819e-01 -6.42926544e-02 6.40580654e-01
4.19915974e-01 -1.42306387e-01 3.39685529e-02 -2.10027732e-02
8.85875762e-01 -2.56920028e-02 -6.91626817e-02 -5.79747140e-01
-8.97474051e-01 7.04914331e-01 4.08189744e-01 6.14877224e-01
1.37929618e-02 -3.61799411e-02 4.72326010e-01 1.35785967e-01
2.65812725e-01 6.11865580e-01 -3.64968807e-01 4.14432704e-01
-9.73839939e-01 -5.38422726e-02 5.92294216e-01 3.60216916e-01
9.36999023e-01 -7.53672421e-02 -9.87803340e-02 5.89302123e-01
-2.36961246e-02 2.16402054e-01 5.53305626e-01 -7.51598299e-01
-2.69146204e-01 7.17559278e-01 -1.46805063e-01 -8.61815453e-01
-6.60453439e-01 -5.12641132e-01 -9.03767169e-01 7.69859850e-01
1.04576397e+00 -1.84218913e-01 -1.45662129e+00 1.70301616e+00
4.32565093e-01 -8.08623731e-02 7.98755661e-02 6.48503065e-01
6.08123541e-01 1.63179591e-01 3.47134829e-01 8.68453830e-02
1.16101384e+00 -7.53200412e-01 -7.07513332e-01 -8.61426666e-02
7.38958597e-01 -8.45631421e-01 1.07426393e+00 4.11076874e-01
-8.59588265e-01 -2.69116580e-01 -1.25945616e+00 -4.90553379e-02
-7.80481577e-01 -2.91234683e-02 5.26476085e-01 9.46367264e-01
-1.16675782e+00 6.39609098e-01 -1.04289758e+00 -5.74075520e-01
4.79716748e-01 5.89016736e-01 -6.90215945e-01 8.11186852e-04
-6.65644586e-01 9.17655826e-01 3.63646001e-01 8.10701102e-02
-5.64607203e-01 -7.18563020e-01 -5.47763824e-01 -1.27186075e-01
-2.78658830e-02 -5.99451542e-01 1.07108092e+00 -1.31859756e+00
-1.43468845e+00 1.37351203e+00 4.80729155e-02 -3.96842539e-01
6.56839132e-01 3.99435073e-01 -1.68978959e-01 1.46912396e-01
-3.13369811e-01 7.71137357e-01 7.27915108e-01 -1.41442442e+00
-3.32936734e-01 -4.39205498e-01 -1.28704207e-02 -3.72664005e-01
-4.39551115e-01 -1.80917621e-01 -5.03247261e-01 -5.87783933e-01
-6.53880835e-02 -1.09720528e+00 -4.15101707e-01 3.01015019e-01
-1.95843384e-01 3.08411300e-01 7.72495389e-01 -6.78529143e-01
7.21282184e-01 -2.41036034e+00 -9.81843248e-02 3.01430017e-01
3.43519635e-02 2.85338938e-01 -2.47699127e-01 3.55342627e-02
-3.91819924e-01 2.93309331e-01 -2.59376377e-01 -2.20045447e-01
-2.22668186e-01 2.96430051e-01 2.15256721e-01 7.64273167e-01
3.57470840e-01 7.73810983e-01 -6.00984573e-01 -8.21105480e-01
8.94388556e-02 6.87515557e-01 -3.67495835e-01 4.97706085e-02
-5.18012941e-02 6.81186438e-01 -7.19327629e-02 6.83598101e-01
9.49048758e-01 -3.14922273e-01 3.35457325e-01 -2.87740409e-01
1.80032954e-01 -2.71140754e-01 -9.35102463e-01 1.57301784e+00
-2.82945126e-01 6.67255938e-01 3.31549406e-01 -1.07553875e+00
6.90636873e-01 1.06233105e-01 4.76147503e-01 -4.85587835e-01
3.21475387e-01 2.11755499e-01 3.97759706e-01 -3.51009250e-01
1.00240357e-01 -6.69517398e-01 2.85972446e-01 2.24891141e-01
1.28235310e-01 -1.16786122e-01 2.46412873e-01 -1.44703194e-01
1.12865615e+00 -1.53352574e-01 2.27236897e-01 -2.67022312e-01
6.14241719e-01 2.98426211e-01 5.62867284e-01 3.57141793e-01
-5.89918494e-01 7.55423307e-01 8.38589132e-01 -5.06247938e-01
-1.12828541e+00 -9.61930573e-01 -2.48106286e-01 6.20333135e-01
-1.69277284e-02 2.62249589e-01 -1.05334961e+00 -9.75377023e-01
-3.38440724e-02 3.58525634e-01 -1.19850767e+00 -1.89974144e-01
-4.89598155e-01 -9.72651422e-01 7.59612858e-01 6.82904959e-01
1.25326887e-01 -7.64354944e-01 -4.07425165e-01 1.20943906e-02
1.67534292e-01 -1.10659599e+00 -1.75663441e-01 5.92062116e-01
-1.05011046e+00 -1.15253067e+00 -8.82642865e-01 -8.05067658e-01
1.31693792e+00 8.48339219e-03 1.10651016e+00 3.86438459e-01
-6.45273447e-01 1.69035122e-01 -3.40396971e-01 -6.24210060e-01
-8.84150207e-01 1.83895364e-01 -3.50479513e-01 1.45322420e-02
3.39263290e-01 -3.37988675e-01 -6.58051372e-01 7.29061067e-02
-1.38316679e+00 -1.82112515e-01 9.33973312e-01 1.05963194e+00
7.68863261e-01 1.03620790e-01 1.40486404e-01 -1.24552405e+00
3.38652849e-01 -3.11487131e-02 -5.29757917e-01 2.03906685e-01
-4.75258142e-01 4.30947691e-02 6.77985489e-01 -6.07747376e-01
-8.63719165e-01 4.30154413e-01 -9.74191055e-02 -4.20395643e-01
-4.56376910e-01 2.91547090e-01 -9.50244069e-02 -7.54732192e-01
8.00786495e-01 -7.73537755e-02 3.15716147e-01 -1.83538690e-01
-1.17936336e-01 9.67015401e-02 5.26461482e-01 -2.00032502e-01
1.12213790e+00 7.95067966e-01 3.43678266e-01 -6.05400205e-01
-5.54832101e-01 -2.17607811e-01 -7.33822763e-01 -1.87915549e-01
9.66745257e-01 -4.52952027e-01 -1.69802889e-01 4.95209366e-01
-1.01145196e+00 -6.75905943e-01 -1.21097587e-01 3.17327112e-01
-3.41825336e-01 3.91855866e-01 -6.55243278e-01 -3.41531634e-01
-1.57699168e-01 -1.14053476e+00 6.64512157e-01 3.98925304e-01
-1.14005469e-01 -1.31502366e+00 2.57706195e-01 2.99702018e-01
4.84065682e-01 5.55780232e-01 1.16656208e+00 -7.87218153e-01
-2.81739712e-01 -4.10662889e-01 -2.21524879e-01 5.14837265e-01
3.15918058e-01 3.88754964e-01 -1.17038345e+00 -3.45110625e-01
-1.05817713e-01 -7.19717219e-02 7.40980089e-01 5.20806611e-01
1.44432867e+00 6.26582950e-02 -4.11547154e-01 8.70903313e-01
1.71637022e+00 1.04422249e-01 9.16384876e-01 6.87939584e-01
4.64819580e-01 7.24301755e-01 3.93575042e-01 -1.22455455e-01
-2.42806420e-01 3.23028326e-01 5.32408059e-01 -9.38049495e-01
-2.34577939e-01 2.25687027e-01 -1.40310094e-01 2.03273799e-02
-1.10444367e-01 -9.81436744e-02 -9.98190045e-01 6.38379812e-01
-1.32467139e+00 -6.16274118e-01 7.17600361e-02 1.97372508e+00
9.18073297e-01 2.93113030e-02 2.21364290e-01 2.29379058e-01
7.74308920e-01 -4.68565196e-01 -3.69580269e-01 -4.45853859e-01
-1.93950906e-01 4.77921814e-01 7.77989268e-01 2.80617595e-01
-1.10114634e+00 6.24543011e-01 7.04475927e+00 5.84159434e-01
-1.69013691e+00 -5.30592352e-02 8.82677972e-01 6.78784400e-02
5.82724847e-02 -3.02420378e-01 -3.80250752e-01 2.46124595e-01
6.99439406e-01 4.29486223e-02 7.82280937e-02 6.00676894e-01
-1.04407884e-01 -6.37136176e-02 -1.23785889e+00 6.48298919e-01
8.90988931e-02 -1.24947095e+00 -3.41545455e-02 3.06090564e-01
8.00403476e-01 -3.70831370e-01 2.12763473e-01 -9.99188572e-02
1.12089947e-01 -1.07348657e+00 2.28717566e-01 3.92999649e-01
5.75365543e-01 -7.58842587e-01 1.18118906e+00 -1.24841198e-01
-6.93211317e-01 2.53353715e-01 -1.77899107e-01 3.43179137e-01
-3.87953967e-01 4.46268111e-01 -1.22993445e+00 2.31597051e-01
5.33122063e-01 2.38592904e-02 -8.50819647e-01 1.19446349e+00
4.17620987e-02 3.72022688e-01 2.31450424e-02 1.21551432e-01
1.51975960e-01 -3.80327627e-02 2.08964542e-01 1.50985050e+00
1.60827711e-01 -2.53203511e-01 -5.48064746e-02 7.58319557e-01
-4.02861973e-03 2.83734560e-01 -4.59960103e-01 -1.52772591e-01
-3.78332622e-02 1.54600120e+00 -1.40647674e+00 -1.60127804e-01
-4.73975480e-01 8.98132682e-01 2.81086117e-01 3.09652150e-01
-8.36418271e-01 -2.34532177e-01 5.52858233e-01 2.83164352e-01
3.12456757e-01 -6.35654107e-02 -5.08137584e-01 -7.32164085e-01
-2.27118298e-01 -9.39582944e-01 3.95649672e-01 -4.32130277e-01
-1.42535245e+00 5.62165797e-01 -3.04890484e-01 -1.28104913e+00
1.85025707e-01 -1.05054271e+00 -5.67423344e-01 8.36492062e-01
-1.76850986e+00 -1.27227795e+00 -4.05784518e-01 6.30109549e-01
1.39701471e-01 3.73244029e-03 1.04550493e+00 3.57221425e-01
-6.09746695e-01 8.07670057e-01 1.46445036e-01 2.86559373e-01
1.22228444e+00 -1.53354371e+00 -2.03720093e-01 8.39051068e-01
-2.85717517e-01 6.03070796e-01 9.08037841e-01 -2.55086392e-01
-1.10621977e+00 -1.07602262e+00 4.01467025e-01 -1.94390327e-01
7.07091689e-01 -1.30795047e-01 -1.12153006e+00 5.05932868e-01
3.43000174e-01 2.30224639e-01 1.29260862e+00 -1.84805006e-01
-2.70591915e-01 -2.06639424e-01 -1.49841380e+00 6.07787967e-01
2.16055512e-01 -4.09655333e-01 -1.89899251e-01 2.09489167e-01
1.47481605e-01 -4.32119757e-01 -7.30950832e-01 2.67446339e-01
4.43011343e-01 -8.29462230e-01 8.79214048e-01 -6.77845716e-01
3.93158793e-01 -4.71129864e-01 1.00657500e-01 -1.27789319e+00
-3.33178312e-01 -1.36882022e-01 3.71931523e-01 1.03038442e+00
4.67526674e-01 -5.54654360e-01 1.15639710e+00 6.60617948e-01
4.94368449e-02 -4.98677611e-01 -7.14264452e-01 -7.40424097e-01
7.20040321e-01 1.78569585e-01 3.90593678e-01 9.02501881e-01
-2.20279455e-01 -6.45966083e-02 1.75249025e-01 2.81912535e-01
5.36347270e-01 -1.58100367e-01 8.30873370e-01 -1.25471961e+00
-2.04221368e-01 -5.69764793e-01 -7.46626854e-01 -5.00274003e-02
1.97588935e-01 -7.84617543e-01 1.09566420e-01 -1.26281822e+00
3.50765318e-01 -2.30422080e-01 -5.98811388e-01 6.62063062e-01
-1.00066453e-01 7.38561392e-01 -3.39710452e-02 -1.20762274e-01
-2.07743987e-01 -2.43307188e-01 1.45728087e+00 -5.14396727e-01
-7.00911954e-02 -2.39875346e-01 -6.61262393e-01 6.37961447e-01
8.23848367e-01 -6.29497290e-01 -6.66440800e-02 -2.81398207e-01
6.92748278e-03 -3.75425130e-01 4.54212248e-01 -1.19144392e+00
1.34846568e-01 -1.53028578e-01 8.89704287e-01 -2.75770396e-01
7.24116340e-02 -1.03689814e+00 1.16204485e-01 7.01990604e-01
-1.72758862e-01 -2.16314912e-01 6.89572453e-01 4.53648895e-01
-2.82775372e-01 -2.36642674e-01 1.32775533e+00 -2.39337593e-01
-3.89397860e-01 -1.71978436e-02 -4.42711532e-01 -3.75822276e-01
1.09680724e+00 -4.84623849e-01 -2.98638999e-01 -5.34847341e-02
-9.79617655e-01 -2.01490819e-01 9.37969267e-01 -1.49536416e-01
2.23548323e-01 -9.34701800e-01 -4.86312777e-01 9.30509344e-02
1.18872963e-01 -1.09079786e-01 4.40739006e-01 1.05077600e+00
-9.15314376e-01 6.34211525e-02 -7.21379459e-01 -5.74305296e-01
-1.59642863e+00 6.95147753e-01 7.07036734e-01 -6.90316677e-01
-1.10644259e-01 7.71790326e-01 2.87942708e-01 -2.15618744e-01
1.31404907e-01 -2.97269434e-01 -2.30887994e-01 -1.00499332e-01
3.56706887e-01 1.50856882e-01 3.60577226e-01 -3.29259694e-01
-3.83395165e-01 5.50159156e-01 -4.17818755e-01 1.24680109e-01
1.27256441e+00 1.92970142e-01 -1.84217647e-01 2.50876963e-01
1.19380116e+00 -3.83394444e-03 -1.07415986e+00 1.51615277e-01
-2.05924869e-01 -2.93021977e-01 1.64677352e-01 -9.68063653e-01
-1.34784091e+00 9.26258624e-01 9.55140650e-01 4.02590394e-01
1.38119829e+00 -9.65160578e-02 4.04650033e-01 2.78260916e-01
1.07666485e-01 -1.00008070e+00 5.67827560e-02 2.09784940e-01
5.78116119e-01 -1.41272199e+00 -1.85387004e-02 -5.04595280e-01
-3.97321343e-01 1.42214358e+00 7.04696834e-01 -1.17428064e-01
4.40317184e-01 6.26896679e-01 6.13800645e-01 8.28999036e-04
-3.26809019e-01 -1.61676601e-01 2.13132888e-01 8.76916766e-01
6.92686081e-01 -1.57442629e-01 -3.92417192e-01 4.09868598e-01
2.17921034e-01 1.14285059e-01 6.06936097e-01 1.07357717e+00
-2.65021417e-02 -1.53117788e+00 -6.52061939e-01 1.16576962e-01
-7.72875309e-01 2.69525468e-01 -7.45471954e-01 1.15673602e+00
2.42242172e-01 5.53581774e-01 8.93338174e-02 -2.37013355e-01
2.17237502e-01 2.70763278e-01 7.33282566e-01 -5.76449215e-01
-5.86128712e-01 1.51327938e-01 -3.31693530e-01 -2.54501671e-01
-5.94705343e-01 -6.72320366e-01 -1.41424787e+00 -2.38286719e-01
-1.63934305e-01 -1.81430012e-01 7.65480638e-01 7.39234209e-01
7.31130913e-02 8.38323355e-01 3.99115056e-01 -6.81787491e-01
-2.44504407e-01 -7.86680162e-01 -6.47911370e-01 5.85853219e-01
4.14009601e-01 -6.06863320e-01 -5.73272824e-01 6.01589024e-01] | [15.042521476745605, -3.023608684539795] |
a8cd73bb-17f1-46e6-a311-a18e18a2e727 | cfc-net-a-critical-feature-capturing-network | 2101.06849 | null | https://arxiv.org/abs/2101.06849v2 | https://arxiv.org/pdf/2101.06849v2.pdf | CFC-Net: A Critical Feature Capturing Network for Arbitrary-Oriented Object Detection in Remote Sensing Images | Object detection in optical remote sensing images is an important and challenging task. In recent years, the methods based on convolutional neural networks have made good progress. However, due to the large variation in object scale, aspect ratio, and arbitrary orientation, the detection performance is difficult to be further improved. In this paper, we discuss the role of discriminative features in object detection, and then propose a Critical Feature Capturing Network (CFC-Net) to improve detection accuracy from three aspects: building powerful feature representation, refining preset anchors, and optimizing label assignment. Specifically, we first decouple the classification and regression features, and then construct robust critical features adapted to the respective tasks through the Polarization Attention Module (PAM). With the extracted discriminative regression features, the Rotation Anchor Refinement Module (R-ARM) performs localization refinement on preset horizontal anchors to obtain superior rotation anchors. Next, the Dynamic Anchor Learning (DAL) strategy is given to adaptively select high-quality anchors based on their ability to capture critical features. The proposed framework creates more powerful semantic representations for objects in remote sensing images and achieves high-performance real-time object detection. Experimental results on three remote sensing datasets including HRSC2016, DOTA, and UCAS-AOD show that our method achieves superior detection performance compared with many state-of-the-art approaches. Code and models are available at https://github.com/ming71/CFC-Net. | ['Yunpeng Dong', 'Zhiqiang Zhou', 'Lingjuan Miao', 'Qi Ming'] | 2021-01-18 | null | null | null | null | ['object-detection-in-aerial-images', 'real-time-object-detection'] | ['computer-vision', 'computer-vision'] | [ 3.31450760e-01 -5.52030265e-01 -1.77757874e-01 -3.16450924e-01
-7.93769002e-01 -2.45147452e-01 3.22487533e-01 -6.81560114e-02
-3.13920647e-01 2.81363010e-01 1.83493234e-02 -1.91901848e-01
-4.50482398e-01 -9.10787165e-01 -4.01186883e-01 -1.14042473e+00
-1.18201591e-01 5.86997084e-02 4.29699808e-01 4.54415195e-02
3.34903359e-01 9.21833575e-01 -1.57768023e+00 -1.85092345e-01
9.36059594e-01 1.09158659e+00 5.12228966e-01 3.61395419e-01
2.10882455e-01 2.53340989e-01 -1.25491485e-01 2.62626171e-01
3.66485745e-01 -2.64326185e-02 -4.10427630e-01 8.81847963e-02
2.79481739e-01 -4.79474008e-01 -9.91861001e-02 1.15361547e+00
7.45264709e-01 1.89539433e-01 6.79236054e-01 -9.65250015e-01
-9.05858934e-01 3.15882057e-01 -1.03970003e+00 6.17275536e-01
-1.61966905e-01 1.34273589e-01 1.12930524e+00 -1.31833565e+00
4.87863496e-02 1.11548042e+00 5.49927056e-01 2.60892421e-01
-8.94892812e-01 -1.01402581e+00 4.71030653e-01 2.97999054e-01
-1.73211110e+00 -3.74775797e-01 7.03590751e-01 -4.70230579e-01
5.08208513e-01 1.72894105e-01 4.98172343e-01 5.59132636e-01
-3.20720702e-01 7.63241291e-01 8.92775595e-01 -2.60196149e-01
-8.71570408e-02 -1.61089376e-02 1.80654719e-01 7.27645040e-01
3.42523813e-01 1.29316613e-01 -8.35470855e-02 5.62601835e-02
8.94355953e-01 4.17551279e-01 -4.55607623e-01 -6.09050207e-02
-1.15440714e+00 8.91877472e-01 1.12264919e+00 2.10461766e-01
-5.29107213e-01 4.85292578e-04 -1.16391882e-01 -2.24587694e-01
4.19619679e-01 2.68217981e-01 -4.19660509e-01 6.50960088e-01
-6.41217947e-01 3.05103119e-02 8.83999690e-02 7.21011341e-01
9.93485987e-01 -7.98673406e-02 -4.08641636e-01 9.52149034e-01
6.32558525e-01 9.49157298e-01 1.14563346e-01 -6.99704885e-01
3.64048511e-01 6.94288790e-01 2.52598256e-01 -1.22611535e+00
-6.07294023e-01 -7.57341385e-01 -8.43272090e-01 4.51018624e-02
-9.16358605e-02 -4.76820730e-02 -1.03327560e+00 1.18002713e+00
5.59895039e-01 2.20218182e-01 -1.39169600e-02 1.26132536e+00
1.01325548e+00 9.24870431e-01 1.77761883e-01 5.29906573e-03
1.46811152e+00 -8.14095676e-01 -2.22131133e-01 -3.94630522e-01
2.43902758e-01 -6.45672023e-01 9.81892884e-01 2.60709357e-02
-3.06875378e-01 -7.53280103e-01 -1.12246835e+00 4.55788244e-03
-1.66046366e-01 8.60801101e-01 6.19209349e-01 2.72535026e-01
-6.07898951e-01 6.88738227e-02 -7.55558610e-01 -3.23602617e-01
7.23111331e-01 2.14979827e-01 -8.54127761e-03 -8.27365071e-02
-1.05377364e+00 4.84942138e-01 4.11809176e-01 6.53738916e-01
-1.03921461e+00 -5.65215886e-01 -6.87535346e-01 7.68195167e-02
3.64186823e-01 -2.95961916e-01 8.77425969e-01 -7.79208601e-01
-1.23715305e+00 6.13577604e-01 -3.71088646e-02 -2.69764178e-02
1.92562565e-01 -2.86795974e-01 -4.52385992e-01 2.87091345e-01
3.42776865e-01 8.55333388e-01 9.63922679e-01 -1.21123636e+00
-1.04535377e+00 -4.16412234e-01 1.01635478e-01 2.93641984e-01
-4.08102423e-01 2.07089618e-01 -3.99155378e-01 -5.51806867e-01
6.45940602e-01 -7.75070369e-01 -3.02227557e-01 3.28761995e-01
-3.02682430e-01 -3.58957350e-01 9.57498789e-01 -4.85125810e-01
9.59239841e-01 -2.35205936e+00 -8.78909603e-02 2.71310419e-01
6.76953495e-02 5.05556643e-01 -2.88199425e-01 -8.05681571e-02
2.23879162e-02 5.40558062e-02 -3.05452496e-01 2.29217201e-01
-2.86448628e-01 -6.42319471e-02 -2.99833447e-01 7.49941111e-01
5.88457108e-01 7.39084959e-01 -9.20984864e-01 -5.37324965e-01
4.13042188e-01 7.07584858e-01 -3.61121774e-01 1.77463561e-01
-8.91571771e-03 6.01816416e-01 -1.10255575e+00 1.26024556e+00
9.14327204e-01 -4.54121500e-01 -2.68316120e-01 -5.29110730e-01
-4.23033535e-01 -9.76052061e-02 -1.35325551e+00 1.02911770e+00
-2.65329033e-01 4.79639441e-01 -4.07796949e-02 -1.03796637e+00
1.26249325e+00 -1.55881792e-01 3.88450384e-01 -6.42411530e-01
7.07748160e-02 2.82445133e-01 -2.07649246e-01 -7.22159624e-01
4.36084569e-01 1.04290657e-01 2.22734988e-01 -5.78864254e-02
-3.53694439e-01 2.48732135e-01 -9.19125751e-02 -1.75917566e-01
4.80590671e-01 1.08571887e-01 3.66164684e-01 -1.13047875e-01
7.98725188e-01 -4.71775495e-02 7.27280617e-01 6.88400209e-01
-3.52160871e-01 5.75736821e-01 -1.39240593e-01 -5.97809672e-01
-6.31876826e-01 -7.82781780e-01 -6.25665009e-01 1.28924787e+00
5.86563051e-01 5.67350723e-02 -2.55683213e-01 -5.17457008e-01
2.28695869e-02 2.58443981e-01 -7.17349172e-01 -6.39341995e-02
-5.50305903e-01 -1.18484676e+00 3.59006077e-01 7.04472184e-01
8.80067945e-01 -1.06347620e+00 -6.03291750e-01 2.07045346e-01
-2.94061482e-01 -8.70360494e-01 -2.31860206e-01 -1.14048071e-01
-7.24792659e-01 -9.58839774e-01 -5.96759081e-01 -7.27710843e-01
7.05471933e-01 1.18986511e+00 4.32106107e-01 3.91922921e-01
-4.69970614e-01 -6.53619766e-02 -6.49691582e-01 -4.09143418e-01
3.25654358e-01 2.14720219e-01 -2.32800037e-01 2.95937270e-01
4.29081202e-01 -3.84224951e-01 -1.03254259e+00 5.31289220e-01
-8.24046671e-01 -9.59708914e-02 1.01765931e+00 7.13387549e-01
7.59966016e-01 -1.06801391e-01 4.30616021e-01 -3.54034215e-01
-1.83383226e-01 -5.63959479e-01 -8.73145938e-01 1.43414661e-01
-4.80515599e-01 -2.34952748e-01 2.76363581e-01 -2.09434330e-01
-1.01389277e+00 1.91377625e-01 -1.01292007e-01 -2.00215876e-01
-2.80812114e-01 5.03044665e-01 -2.18598694e-01 -2.39338398e-01
6.92675710e-01 3.04280370e-01 -3.00798833e-01 -6.15577340e-01
1.21057995e-01 1.05587399e+00 2.95751184e-01 -4.04924482e-01
1.05979621e+00 7.03921318e-01 -9.90101844e-02 -1.03149974e+00
-1.28485334e+00 -7.61838555e-01 -5.42900622e-01 -1.70526370e-01
9.10121083e-01 -1.30729616e+00 -4.92627442e-01 5.02699196e-01
-8.17408621e-01 -2.07252614e-02 5.47999889e-02 6.47035956e-01
3.99349071e-02 1.48712650e-01 -2.14254290e-01 -8.72232676e-01
-6.08089209e-01 -1.13015056e+00 1.50261962e+00 8.05987179e-01
6.35838151e-01 -4.37652677e-01 -1.95142776e-01 3.21480542e-01
4.96502817e-01 1.13703668e-01 4.94345605e-01 -3.07397366e-01
-9.82549489e-01 -6.86592758e-02 -9.21581149e-01 2.11235657e-01
2.19906550e-02 1.79198295e-01 -1.03204775e+00 -4.19461280e-01
-4.96899843e-01 -1.69742540e-01 1.29954600e+00 5.60789347e-01
1.34749806e+00 -7.51880705e-02 -6.68598115e-01 1.09762275e+00
1.49853492e+00 2.65624728e-02 4.92304742e-01 5.31296015e-01
9.22291756e-01 4.39137250e-01 9.82967496e-01 4.78781402e-01
4.53904390e-01 5.55229187e-01 8.07314694e-01 -2.35640809e-01
-1.90239493e-02 -5.18411063e-02 2.07296073e-01 2.48090848e-01
-4.50461537e-01 7.58338347e-02 -7.97851324e-01 5.58480740e-01
-1.70017576e+00 -9.72979128e-01 -2.55712271e-01 1.80754685e+00
4.37588006e-01 -1.74551994e-01 -1.87901482e-01 -3.32889557e-02
9.06699955e-01 4.19265807e-01 -5.84086418e-01 6.46406591e-01
-2.19304398e-01 -1.59570426e-01 7.68702686e-01 4.58843231e-01
-1.59009409e+00 9.99486148e-01 4.77095318e+00 8.44845831e-01
-1.29137385e+00 1.23312309e-01 4.34061944e-01 2.05315173e-01
6.37184978e-02 -3.14233042e-02 -1.36451006e+00 2.60864168e-01
1.45848617e-01 4.21768278e-01 1.54989421e-01 9.27528262e-01
3.82609248e-01 1.70040548e-01 -3.27452302e-01 6.98494971e-01
-6.12403527e-02 -1.16478670e+00 2.05004722e-01 -6.28157854e-02
5.97445607e-01 4.58393842e-01 8.01867247e-02 2.48240620e-01
1.10124938e-01 -8.16058517e-01 6.16098523e-01 4.93326873e-01
7.50933528e-01 -5.45473099e-01 7.72227466e-01 2.11764067e-01
-1.67248929e+00 -4.92090821e-01 -8.82751286e-01 1.76343307e-01
-9.80103165e-02 5.75079799e-01 -5.14059961e-01 5.68913341e-01
1.00872242e+00 1.13936448e+00 -7.61787772e-01 1.23629725e+00
-5.82303464e-01 6.09090865e-01 -3.91465992e-01 -2.95853019e-02
3.42943609e-01 -2.09107652e-01 6.16544902e-01 1.13934207e+00
5.13601482e-01 3.99988592e-01 5.26051402e-01 7.54255176e-01
1.38440341e-01 1.57259494e-01 -1.73019201e-01 2.80171037e-01
7.65782952e-01 1.77549541e+00 -7.56736100e-01 1.31843816e-02
-3.52703601e-01 5.34875929e-01 2.96363443e-01 4.13917542e-01
-9.56862867e-01 -3.78876626e-01 5.84099889e-01 9.93071869e-03
6.08555019e-01 -1.62888154e-01 1.45946503e-01 -1.10231149e+00
3.67144570e-02 -5.31992495e-01 5.29058158e-01 -8.38435233e-01
-1.15079916e+00 4.17261034e-01 -9.93145183e-02 -1.43592846e+00
6.84782505e-01 -6.36397541e-01 -6.23703539e-01 8.36459935e-01
-2.24781251e+00 -1.45917535e+00 -9.09321904e-01 4.83673960e-01
4.41211492e-01 -1.51170942e-03 4.83582199e-01 3.86900574e-01
-8.33597839e-01 2.46702969e-01 1.34963602e-01 3.83603275e-01
4.72269475e-01 -7.87363946e-01 7.38081560e-02 1.02681792e+00
-8.11168477e-02 4.25642729e-01 2.28884429e-01 -3.67761523e-01
-1.20207477e+00 -1.63451362e+00 4.56778020e-01 -9.57659036e-02
4.39858288e-01 -2.63122302e-02 -9.83279347e-01 5.81738889e-01
-5.12890995e-01 4.56872046e-01 4.56483006e-01 -2.00877383e-01
-3.72989178e-01 -5.40897846e-01 -7.54052162e-01 2.59471953e-01
1.01307857e+00 -2.82135487e-01 -4.13626999e-01 4.77380216e-01
6.80901647e-01 -1.99375853e-01 -5.21791220e-01 7.36482382e-01
5.67348957e-01 -6.70723855e-01 1.27902126e+00 -1.50948629e-01
2.08058149e-01 -9.38801885e-01 -3.00988823e-01 -8.86923850e-01
-8.05484056e-01 -2.77909264e-02 1.98766246e-01 1.10136974e+00
2.97989398e-01 -7.05134153e-01 3.94785196e-01 -2.84090608e-01
-1.96623400e-01 -5.84688842e-01 -6.96103096e-01 -4.89806771e-01
-3.51986110e-01 -9.07445475e-02 6.75685585e-01 9.11843956e-01
-7.93123662e-01 3.59767169e-01 -2.50830382e-01 1.08922482e+00
6.49269283e-01 6.32610619e-01 6.61942899e-01 -1.49198103e+00
-1.13007642e-01 -4.24614966e-01 -2.55515814e-01 -1.32054961e+00
-1.24983497e-01 -8.77040982e-01 3.27671230e-01 -1.59313262e+00
2.61561781e-01 -8.47768664e-01 -5.00990152e-01 7.51469433e-01
-6.22139812e-01 5.46488464e-01 3.57934982e-02 6.78085148e-01
-6.60951018e-01 6.73765957e-01 1.18974733e+00 -2.88304895e-01
-2.48311594e-01 8.37486237e-02 -8.59659493e-01 6.75746322e-01
9.09428954e-01 -4.75514770e-01 2.99039539e-02 -6.37543261e-01
1.74487978e-01 -3.78900796e-01 7.55629241e-01 -1.10274947e+00
1.57485574e-01 -2.64253467e-01 7.61940718e-01 -6.76052153e-01
5.18306009e-02 -7.32445002e-01 -1.69719219e-01 5.72983861e-01
-6.96018264e-02 -3.95953774e-01 2.19739880e-03 6.90681219e-01
-5.17339930e-02 -1.68173298e-01 1.08377039e+00 8.78346413e-02
-9.29528832e-01 6.56677902e-01 -5.01659624e-02 -2.71875024e-01
9.62391257e-01 -2.24513173e-01 -4.81951684e-01 2.05009565e-01
-5.26069641e-01 3.97048116e-01 4.31419797e-02 4.23520744e-01
6.53940916e-01 -1.17883515e+00 -1.08634949e+00 4.17556047e-01
5.72132468e-01 2.65660971e-01 3.55872810e-01 1.03309262e+00
-5.32955229e-01 1.48547605e-01 -5.48291132e-02 -9.66232777e-01
-9.73379612e-01 2.21864387e-01 5.21031857e-01 1.24688327e-01
-6.90710425e-01 1.00569499e+00 2.09971443e-01 -2.74822086e-01
6.66970387e-02 -2.05467537e-01 -5.77889085e-01 2.41685197e-01
8.02018583e-01 4.58523929e-01 -1.31583959e-01 -8.26231062e-01
-5.80464900e-01 1.11859620e+00 -5.21591343e-02 5.05168140e-01
1.55547810e+00 -3.12325329e-01 -7.64247403e-02 -1.46082789e-01
8.75965238e-01 -1.13941513e-01 -1.51436055e+00 -6.17016494e-01
-3.16261679e-01 -7.01020479e-01 3.80387157e-01 -4.41299707e-01
-1.36646640e+00 9.21125829e-01 9.73574936e-01 6.76974654e-02
1.19604552e+00 9.61248875e-02 4.05064732e-01 5.22064626e-01
1.19718343e-01 -7.49662876e-01 1.09979920e-01 4.98102993e-01
8.15747380e-01 -1.50613987e+00 1.67122394e-01 -4.69042242e-01
-2.40975350e-01 1.02766669e+00 7.30536759e-01 -1.52968377e-01
7.18868554e-01 -1.57816306e-01 2.43926227e-01 -3.55234683e-01
-1.40700964e-02 -5.30838728e-01 3.24132204e-01 6.16803586e-01
1.33195147e-01 2.09728673e-01 -1.72439307e-01 4.97538656e-01
2.96393096e-01 -3.41886878e-01 1.14298746e-01 6.81962013e-01
-1.08370709e+00 -5.14283895e-01 -5.96584618e-01 4.02006030e-01
-2.83890098e-01 -2.05130458e-01 1.98307224e-02 6.25823021e-01
3.20170790e-01 8.89944494e-01 1.50313927e-02 -2.51441061e-01
2.80195028e-01 -4.96826470e-01 3.47809903e-02 -5.83022416e-01
-1.60377458e-01 1.28174618e-01 -3.03839833e-01 -3.51253241e-01
-8.63212705e-01 -6.88812435e-01 -1.35424662e+00 1.59799427e-01
-7.91879773e-01 9.06147286e-02 3.89410287e-01 8.37559044e-01
4.27384347e-01 4.25270349e-01 9.03643668e-01 -1.09046912e+00
-5.67208052e-01 -1.10817337e+00 -5.65421462e-01 -1.05672523e-01
4.54620093e-01 -1.00227225e+00 -4.55163717e-01 -1.26843542e-01] | [9.043428421020508, -0.9068543910980225] |
b7202078-673c-4694-96fa-743bd5cd9907 | moving-poselets-a-discriminative-and | null | null | https://doi.org/10.1109/ICCVW.2015.48 | http://www.vision.jhu.edu/assets/TaoLAP15.pdf | Moving poselets: A discriminative and interpretable skeletal motion representation for action recognition | Given a video or time series of skeleton data, action recognition systems perform classification using cues such as motion, appearance, and pose. For the past decade, actions have been modeled using low-level feature representations such as Bag of Features. More recent work has shown that mid-level representations that model body part movements (e.g., hand moving forward) can be very effective. However, these mid-level features are usually hand-crafted and the dictionary of representative features is learned using ad-hoc heuristics. While automatic feature learning methods such as supervised sparse dictionary learning or neural networks can be applied to learn feature representation and action classifiers jointly, the resulting features are usually uninterpretable. In contrast, our goal is to develop a principled feature learning framework to learn discriminative and interpretable skeletal motion patterns for action recognition. For this purpose, we propose a novel body-part motion based feature called Moving Poselet, which corresponds to a specific body part configuration undergoing a specific movement. We also propose a simple algorithm for jointly learning Moving Poselets and action classifiers. Experiments on MSR Action3D, MSR DailyActivity3D and Berkeley MHAD datasets show that our two-layer model outperforms other two-layer models using hand-crafted features, and achieves results comparable to those of recent multi-layer Hierarchical Recurrent Neural Network (HRNN) models, which use multiple layers of RNN to model the human body hierarchy. | ['René Vidal', 'Lingling Tao'] | 2015-12-07 | null | null | null | 2015-ieee-international-conference-on | ['multimodal-activity-recognition'] | ['computer-vision'] | [ 2.87807465e-01 -1.20530896e-01 -8.71062100e-01 -4.17084813e-01
-6.28169894e-01 1.48758106e-02 6.13359392e-01 -9.49536636e-02
-2.56327450e-01 3.56865168e-01 6.92839980e-01 3.08960021e-01
-9.15947631e-02 -5.51788032e-01 -5.25683224e-01 -7.17221856e-01
-2.82791585e-01 4.29185569e-01 2.67794400e-01 -2.23391742e-01
7.24474015e-03 5.72390497e-01 -1.82753897e+00 4.66108292e-01
1.20132372e-01 1.02796519e+00 -5.35042547e-02 6.15099728e-01
2.36177310e-01 1.06666183e+00 -2.95218498e-01 1.48619533e-01
2.98353374e-01 -6.88505650e-01 -9.24265146e-01 5.61842501e-01
5.02825320e-01 -4.51735139e-01 -6.04264736e-01 4.15268093e-01
5.16345561e-01 4.33362007e-01 7.47541189e-01 -9.76609170e-01
-2.26876527e-01 3.23887080e-01 -5.68051100e-01 2.26284295e-01
6.31538570e-01 6.81228861e-02 1.14715719e+00 -7.47204125e-01
7.41351008e-01 1.40940344e+00 6.54130220e-01 8.02480757e-01
-1.13249671e+00 -2.14141101e-01 2.16293260e-01 5.05739093e-01
-1.18293905e+00 -4.06308472e-01 9.91913438e-01 -4.74460006e-01
1.03582799e+00 1.56959817e-01 1.04549861e+00 1.34083021e+00
5.43089807e-01 1.28690362e+00 6.75879002e-01 -2.90499538e-01
7.94070363e-02 -6.18061066e-01 4.29132208e-02 1.02744257e+00
-3.91543377e-03 8.41679275e-02 -7.97888339e-01 -1.54311091e-01
1.08351648e+00 5.74813664e-01 -1.28476217e-01 -7.77458012e-01
-1.18756986e+00 8.98284495e-01 2.67447084e-01 4.66077775e-01
-7.01078475e-01 4.05732721e-01 5.78033268e-01 1.16791524e-01
3.37395728e-01 3.76413316e-02 -4.84421343e-01 -4.62851405e-01
-1.00986373e+00 2.49532908e-01 5.63150346e-01 5.04236221e-01
6.68258786e-01 2.46765688e-01 -2.02416405e-01 1.01316416e+00
4.37439054e-01 4.07516569e-01 1.10371900e+00 -1.14244521e+00
6.97511956e-02 7.20194101e-01 -2.33737335e-01 -1.24055207e+00
-6.80742741e-01 -3.51576284e-02 -8.68474543e-01 1.16205856e-01
2.06907198e-01 2.50112802e-01 -1.19189656e+00 1.55082095e+00
4.62172568e-01 7.09103942e-02 -6.48249686e-02 1.03521669e+00
6.84349000e-01 3.46289247e-01 1.82059631e-02 -7.98950940e-02
1.09795535e+00 -1.01085496e+00 -5.24614155e-01 -1.14738487e-01
8.41571271e-01 -2.35563546e-01 6.67622209e-01 4.36658800e-01
-9.62292552e-01 -7.73805559e-01 -7.96650052e-01 -7.78321028e-02
-1.27281860e-01 2.08559036e-01 8.06044877e-01 3.18482995e-01
-6.84638023e-01 9.39525068e-01 -1.44043827e+00 -5.54682791e-01
2.98594654e-01 5.08111417e-01 -6.00964308e-01 1.89020131e-02
-9.89809811e-01 8.47106516e-01 3.33753437e-01 2.92415582e-02
-1.15534151e+00 -1.60594851e-01 -1.30896544e+00 -3.42194617e-01
2.97252655e-01 -8.12596679e-01 1.17709589e+00 -1.03613901e+00
-1.73500586e+00 7.50773847e-01 -1.86759695e-01 -6.26068652e-01
1.25489429e-01 -4.30839151e-01 -2.23883390e-01 5.11804342e-01
-8.98366570e-02 6.81368113e-01 1.25874102e+00 -8.19137633e-01
-4.77930695e-01 -5.04051924e-01 -1.75436568e-02 2.24862650e-01
-3.51791322e-01 -3.54720354e-02 -4.76763636e-01 -9.55697834e-01
3.29270393e-01 -1.20968628e+00 -4.14489567e-01 9.37053859e-02
-6.72628582e-02 -3.10307533e-01 7.05618799e-01 -7.03423500e-01
1.16000676e+00 -2.07467437e+00 6.31300926e-01 1.33078054e-01
-4.43212017e-02 8.44877362e-02 -6.84931874e-02 2.93798715e-01
-7.39208758e-02 -3.17497134e-01 -7.87478015e-02 -2.43425414e-01
-2.27429152e-01 8.11727047e-01 1.02948382e-01 7.25631475e-01
6.04098178e-02 7.85837770e-01 -8.14807951e-01 -6.64404213e-01
4.72687244e-01 6.40998006e-01 -7.46547461e-01 1.37852775e-02
-4.91345674e-02 4.02134418e-01 -7.21784115e-01 9.35242116e-01
-3.48784745e-01 -4.03252572e-01 2.81030953e-01 -3.15010786e-01
3.47431958e-01 2.39363223e-01 -1.15167761e+00 2.08066297e+00
-4.58800375e-01 3.47036064e-01 -3.60182047e-01 -1.22205007e+00
8.03829253e-01 5.13609052e-01 1.16016161e+00 -3.98984253e-01
2.35702261e-01 -1.09597579e-01 -8.71064216e-02 -5.78353405e-01
2.90295482e-01 -1.98036581e-01 -1.79392219e-01 3.30101967e-01
2.79059291e-01 2.73716629e-01 1.35501400e-01 -1.01348646e-01
1.35088706e+00 5.49985647e-01 6.47272527e-01 2.63970584e-01
5.97089529e-01 -6.52282164e-02 8.44852090e-01 4.85960990e-01
-1.10963449e-01 8.62158775e-01 -3.84513997e-02 -7.69441187e-01
-6.02217734e-01 -7.36934662e-01 1.72182396e-01 1.32120085e+00
-2.78416909e-02 -7.26571023e-01 -2.82859951e-01 -7.35322535e-01
5.84274642e-02 1.28039151e-01 -8.30694795e-01 -4.09767061e-01
-9.08021152e-01 -4.03053552e-01 3.62594008e-01 1.06691933e+00
3.37047786e-01 -1.21612024e+00 -9.93022203e-01 5.00559926e-01
-1.14938848e-01 -8.66944134e-01 -4.52910036e-01 1.66331708e-01
-1.27751160e+00 -1.15457165e+00 -7.88514435e-01 -4.63319629e-01
6.09273136e-01 2.32998818e-01 8.92256916e-01 7.73022994e-02
-5.37978232e-01 8.64665985e-01 -7.47053206e-01 1.95766259e-02
-1.37235671e-01 -2.14485317e-01 4.39288259e-01 2.30912060e-01
2.56897420e-01 -8.11864376e-01 -6.72948360e-01 3.28718960e-01
-8.39547276e-01 -1.10260502e-01 8.71834219e-01 1.03888130e+00
9.08847451e-01 -1.98438600e-01 1.68366253e-01 -4.27783459e-01
6.56038076e-02 -2.42928118e-01 2.04149961e-01 1.00458644e-01
-2.49974117e-01 1.89244807e-01 3.75390053e-01 -7.85975337e-01
-7.62718618e-01 6.64789200e-01 -7.60252252e-02 -1.00226438e+00
-2.98008323e-01 6.41783834e-01 -1.14265911e-01 1.29333818e-02
5.70798099e-01 5.99642217e-01 1.97004735e-01 -6.65710151e-01
3.13199729e-01 3.58410537e-01 4.86911237e-01 -5.74006021e-01
5.33858240e-01 6.90057456e-01 1.15921281e-01 -9.43546116e-01
-8.99568200e-01 -8.06707084e-01 -1.03861630e+00 -3.77815753e-01
1.00801980e+00 -9.19871747e-01 -4.36715275e-01 4.41338331e-01
-7.35163808e-01 -3.12203437e-01 -4.84526128e-01 8.41844082e-01
-1.17527211e+00 6.93814456e-01 -7.46503174e-01 -6.91254199e-01
-2.20758244e-01 -8.09787691e-01 1.55208027e+00 -1.32092282e-01
-6.73932254e-01 -8.70988250e-01 2.45808318e-01 5.55400729e-01
-1.21490479e-01 7.08481610e-01 6.02684021e-01 -5.74144185e-01
-1.13503911e-01 -3.60996246e-01 6.22047424e-01 3.76102597e-01
5.38495660e-01 -2.53977656e-01 -3.00904930e-01 -3.34864736e-01
-6.28324747e-02 -5.74516714e-01 9.83411968e-01 5.15318394e-01
1.25769997e+00 -3.85581315e-01 -4.44308639e-01 3.99578810e-01
9.05645669e-01 1.19927965e-01 4.98456836e-01 2.39153400e-01
8.39938819e-01 3.98236752e-01 7.77651727e-01 6.52999997e-01
3.16970527e-01 9.41778719e-01 1.73352286e-01 7.66973495e-02
-1.76897660e-01 -3.30588162e-01 6.82107031e-01 7.95768321e-01
-7.08464265e-01 2.59974569e-01 -7.78799176e-01 4.79093790e-01
-2.09399748e+00 -1.27659106e+00 3.05403769e-01 1.85205328e+00
7.51652122e-01 9.35811996e-02 6.84170902e-01 4.14554775e-01
3.37788850e-01 3.97611499e-01 -5.96941888e-01 -1.17553294e-01
1.79796502e-01 1.02425240e-01 2.17549503e-01 4.19570953e-02
-1.24125302e+00 9.48333561e-01 6.22685003e+00 6.04618609e-01
-1.12247491e+00 -8.07265006e-03 -1.27956225e-02 -2.71292567e-01
2.48207435e-01 -3.17580104e-01 -7.32548177e-01 2.12741956e-01
8.66299689e-01 2.11492673e-01 1.98829155e-02 1.02747536e+00
3.20911974e-01 1.63930327e-01 -1.33249342e+00 1.20052314e+00
4.47627515e-01 -1.05479312e+00 2.86785245e-01 1.80524036e-01
3.08878779e-01 -1.04751671e-02 -1.80369854e-01 4.16196704e-01
2.89837718e-01 -8.62898111e-01 4.87226874e-01 6.73361182e-01
4.26511317e-01 -5.00189364e-01 2.64435351e-01 3.63430917e-01
-1.34780395e+00 -2.31293246e-01 -2.27038071e-01 -1.76920056e-01
2.43968397e-01 1.57630563e-01 -4.50922668e-01 5.68600655e-01
6.63531303e-01 1.46420562e+00 -3.83456558e-01 8.05041969e-01
-1.96567982e-01 6.22290313e-01 -3.43645751e-01 1.92627504e-01
2.71279186e-01 1.67307779e-01 4.82254118e-01 8.85091841e-01
1.64732561e-01 3.06896865e-01 5.32229245e-01 2.52746850e-01
3.10625136e-01 8.26506242e-02 -6.77098751e-01 -1.66272521e-01
-3.37203056e-01 1.00149024e+00 -6.54318094e-01 -3.01713705e-01
-6.50822222e-01 1.13343084e+00 9.22731385e-02 9.67882648e-02
-6.42050445e-01 1.58719480e-01 8.30297887e-01 1.09645009e-01
6.31850719e-01 -4.66083616e-01 3.03374112e-01 -1.36507773e+00
-1.40317708e-01 -1.14221454e+00 7.56630123e-01 -3.93177390e-01
-1.12532699e+00 1.76024973e-01 7.98538253e-02 -1.59967983e+00
-7.87338018e-01 -6.39172375e-01 -4.41159040e-01 5.16123325e-02
-8.50575864e-01 -1.28648114e+00 -4.19143541e-03 9.81994450e-01
9.69378889e-01 -2.04709202e-01 9.55385089e-01 1.09999232e-01
-5.28189778e-01 3.73639554e-01 -1.53660744e-01 4.24552023e-01
3.82286042e-01 -1.04671311e+00 -4.06617858e-02 3.73768300e-01
7.50186443e-01 5.62991500e-01 6.42803490e-01 -8.22020531e-01
-1.70811570e+00 -1.00308526e+00 5.66562653e-01 -4.87123340e-01
4.32754517e-01 -3.94176617e-02 -6.58018947e-01 8.17298234e-01
-4.65775222e-01 3.88380975e-01 9.44217622e-01 2.49269247e-01
-2.72749513e-01 2.34525241e-02 -7.71666884e-01 3.43938142e-01
1.29297900e+00 -4.70371693e-01 -1.06951571e+00 3.76238793e-01
1.34277284e-01 -3.72804135e-01 -1.18359065e+00 6.61237240e-01
9.40770507e-01 -6.63564503e-01 1.17592132e+00 -9.69970703e-01
3.81296724e-01 -1.56434700e-01 -3.41871768e-01 -1.07122242e+00
-4.53487396e-01 -4.22798902e-01 -7.48792946e-01 6.58464372e-01
2.51695723e-03 -1.22766653e-02 9.75089729e-01 3.15657645e-01
-3.97688672e-02 -1.12936687e+00 -9.15470481e-01 -1.02129710e+00
-2.92356908e-01 -4.04130846e-01 1.15606964e-01 8.02209139e-01
1.02170259e-01 2.87444800e-01 -9.06532049e-01 -1.73178792e-01
3.64954770e-01 4.55618113e-01 8.48983705e-01 -1.30908692e+00
-5.62637150e-01 -1.72475845e-01 -1.11557722e+00 -1.24420285e+00
3.45390201e-01 -7.47134209e-01 1.07882889e-02 -1.60762227e+00
1.79278046e-01 1.51304767e-01 -4.61488575e-01 9.87674415e-01
2.16810256e-01 1.83307797e-01 7.35967699e-03 3.45595241e-01
-7.70412743e-01 7.91086137e-01 1.23098075e+00 -3.60510737e-01
-3.01196367e-01 1.47739708e-01 -2.76067376e-01 1.11764836e+00
5.42401016e-01 -4.64592099e-01 -3.71730089e-01 -1.16447590e-01
-2.14182466e-01 3.01055253e-01 5.20819366e-01 -1.18244314e+00
1.68296188e-01 -2.68113285e-01 6.39537096e-01 -6.29577100e-01
7.55527854e-01 -7.51531541e-01 8.14216435e-02 6.65803373e-01
-3.90569031e-01 -3.33273470e-01 -2.31150761e-01 8.34347308e-01
-3.70542556e-01 1.57089204e-01 4.00767863e-01 -3.54127854e-01
-1.03481650e+00 5.06659091e-01 -6.27253652e-01 -1.99873105e-01
9.29561973e-01 -6.07760251e-01 4.59515572e-01 -4.72052455e-01
-1.09891915e+00 -5.25026210e-03 2.29091033e-01 8.39355230e-01
8.22245836e-01 -1.62037659e+00 -4.82676953e-01 5.06151887e-03
1.48553893e-01 -2.33115703e-01 1.57907039e-01 9.13818181e-01
-2.18675397e-02 3.86800706e-01 -4.67688143e-01 -7.67157614e-01
-1.48184216e+00 4.33312356e-01 2.89912194e-01 -2.16097385e-01
-1.13787568e+00 5.55401683e-01 -3.46647389e-02 -1.36665598e-01
2.39751250e-01 -4.10557836e-01 -3.47299993e-01 1.46003872e-01
4.59344387e-01 3.77599150e-01 -2.47218683e-01 -1.20161080e+00
-4.82621640e-01 8.43952775e-01 -4.62489538e-02 2.50326574e-01
1.62404203e+00 5.99082448e-02 2.48762950e-01 6.43644392e-01
1.20224607e+00 -5.01758873e-01 -1.21280611e+00 -3.57676566e-01
-5.25716972e-03 -3.55224609e-01 -6.71127986e-04 -3.99070442e-01
-1.24624825e+00 7.48637617e-01 5.52888572e-01 -2.37054408e-01
1.17155480e+00 2.71458298e-01 1.02048910e+00 6.60620511e-01
6.26967728e-01 -1.20936418e+00 5.06862760e-01 3.76144052e-01
9.51533914e-01 -1.12158644e+00 2.10520968e-01 -5.50245717e-02
-6.49861038e-01 1.16058719e+00 5.64788401e-01 -2.14033544e-01
6.90432489e-01 -1.49699867e-01 -9.68403444e-02 -4.31232691e-01
-7.13555217e-01 -4.73121524e-01 6.93494558e-01 3.49630058e-01
4.67794895e-01 -7.35072196e-02 -3.50464761e-01 6.14466965e-01
-9.12004560e-02 1.12565674e-01 1.70533061e-01 1.40186548e+00
-4.76750851e-01 -1.18304420e+00 -1.77455723e-01 6.01185262e-01
-5.75306475e-01 5.13377190e-01 -3.70944649e-01 8.24435294e-01
1.09233037e-01 6.22463107e-01 -1.30993322e-01 -6.10837281e-01
2.85998315e-01 2.71157414e-01 8.12689781e-01 -8.60429049e-01
-4.51762587e-01 1.47490025e-01 6.54284880e-02 -1.13309538e+00
-9.65547740e-01 -9.98520494e-01 -1.39604604e+00 2.86091357e-01
-6.20439239e-02 -2.37213179e-01 1.92015588e-01 1.16188872e+00
1.97937071e-01 3.14962000e-01 4.13902044e-01 -1.22630131e+00
-5.87911844e-01 -9.03714418e-01 -7.72152603e-01 6.88203335e-01
3.28695625e-01 -1.17367446e+00 -1.39430821e-01 4.08386320e-01] | [7.85798454284668, 0.37026363611221313] |
419c3b6c-a838-47d1-b45b-0191707b2687 | recon-relation-extraction-using-knowledge | 2009.08694 | null | https://arxiv.org/abs/2009.08694v2 | https://arxiv.org/pdf/2009.08694v2.pdf | RECON: Relation Extraction using Knowledge Graph Context in a Graph Neural Network | In this paper, we present a novel method named RECON, that automatically identifies relations in a sentence (sentential relation extraction) and aligns to a knowledge graph (KG). RECON uses a graph neural network to learn representations of both the sentence as well as facts stored in a KG, improving the overall extraction quality. These facts, including entity attributes (label, alias, description, instance-of) and factual triples, have not been collectively used in the state of the art methods. We evaluate the effect of various forms of representing the KG context on the performance of RECON. The empirical evaluation on two standard relation extraction datasets shows that RECON significantly outperforms all state of the art methods on NYT Freebase and Wikidata datasets. RECON reports 87.23 F1 score (Vs 82.29 baseline) on Wikidata dataset whereas on NYT Freebase, reported values are 87.5(P@10) and 74.1(P@30) compared to the previous baseline scores of 81.3(P@10) and 63.1(P@30). | ['Manohar Kaul', "Isaiah Onando Mulang'", 'Kuldeep Singh', 'Saeedeh Shekarpour', 'Johannes Hoffart', 'Anson Bastos', 'Abhishek Nadgeri'] | 2020-09-18 | null | null | null | null | ['relationship-extraction-distant-supervised'] | ['natural-language-processing'] | [ 6.15634695e-02 1.07351840e+00 -4.15923804e-01 -4.23333347e-01
-5.68322778e-01 -5.42141259e-01 9.21355367e-01 9.24835980e-01
-4.60697949e-01 1.14010048e+00 3.72540802e-01 -1.80934593e-01
-4.65525210e-01 -1.10855389e+00 -9.88640547e-01 3.51598114e-02
-3.41621161e-01 6.60289884e-01 3.35490644e-01 -3.03980112e-01
-1.27123863e-01 1.61548909e-02 -1.31257331e+00 4.98069108e-01
8.90192270e-01 9.69136119e-01 -3.57058197e-01 4.28836286e-01
-3.33154529e-01 1.35974050e+00 -8.24147165e-01 -1.10147870e+00
4.76738401e-02 6.93032006e-03 -1.27737272e+00 -6.23817861e-01
9.42337871e-01 1.18296780e-01 -4.07769173e-01 9.60306764e-01
2.09119245e-01 -1.08711414e-01 7.57691085e-01 -1.26805329e+00
-8.90736520e-01 1.58039510e+00 -5.22812426e-01 2.01390669e-01
6.67628407e-01 -6.10049784e-01 1.47989786e+00 -6.70332551e-01
1.33372438e+00 9.45089757e-01 7.63478696e-01 8.41971561e-02
-9.36386287e-01 -6.58104300e-01 -1.49262592e-01 4.89536077e-01
-1.39922464e+00 -5.22540748e-01 3.14365059e-01 -2.64240384e-01
1.86560810e+00 2.28443772e-01 4.64524835e-01 6.79843366e-01
4.43908125e-02 4.57520545e-01 8.76175284e-01 -8.39710295e-01
-1.48062021e-01 2.44317457e-01 6.13463938e-01 7.52687335e-01
9.44413066e-01 -3.42730612e-01 -8.35582495e-01 -1.28937945e-01
3.19361925e-01 -7.96939135e-01 -3.50251555e-01 -1.60547465e-01
-1.07703722e+00 4.63898599e-01 5.90014756e-01 3.88284564e-01
-6.02524698e-01 7.68267065e-02 5.09395778e-01 3.54938805e-01
3.88278604e-01 8.19684565e-01 -9.83764529e-01 -7.27846771e-02
-6.71154082e-01 4.47198570e-01 1.28275812e+00 1.31645119e+00
7.45136619e-01 -3.83339524e-01 -2.82822192e-01 6.17978513e-01
-1.52377440e-02 3.28538358e-01 1.62692562e-01 -6.42263234e-01
1.06783533e+00 1.17139065e+00 -5.86264022e-02 -1.31478083e+00
-6.42467022e-01 -4.84219134e-01 -4.53587115e-01 -4.46443886e-01
4.49398637e-01 -1.92611918e-01 -7.63256013e-01 1.62784564e+00
3.25498372e-01 -3.66594754e-02 7.18512177e-01 3.88863713e-01
1.78173196e+00 4.18054968e-01 3.72044444e-01 -2.20671177e-01
1.40533745e+00 -6.99736893e-01 -1.04744411e+00 -2.93965906e-01
1.03160942e+00 -5.86160719e-01 5.16093910e-01 -1.38483150e-02
-9.58754003e-01 -2.30902672e-01 -1.19979405e+00 -2.54618615e-01
-9.82891142e-01 1.70802951e-01 8.26210141e-01 4.44372952e-01
-9.37772214e-01 8.33645940e-01 -4.09826159e-01 -6.25001311e-01
3.31166416e-01 3.61148983e-01 -8.63137722e-01 9.13676694e-02
-1.67892659e+00 1.41976261e+00 1.06785667e+00 -3.62266570e-01
-1.12696409e-01 -7.92460203e-01 -1.17316914e+00 2.04737857e-01
9.77800667e-01 -6.94303095e-01 8.98764670e-01 -1.40474513e-01
-8.93843114e-01 9.92505908e-01 1.48641439e-02 -1.00413752e+00
-2.18872707e-02 -6.06214285e-01 -8.22391093e-01 2.49481961e-01
2.21411079e-01 3.88804644e-01 1.08605132e-01 -1.11611223e+00
-7.04136670e-01 -2.20421597e-01 4.67613935e-01 7.45296553e-02
-2.68952936e-01 6.71161562e-02 -2.24379316e-01 -1.93136081e-01
-6.16485924e-02 -6.34489298e-01 2.91713744e-01 -8.24714303e-01
-9.50105906e-01 -6.34531021e-01 4.64596450e-01 -1.08699226e+00
1.51622272e+00 -1.50469804e+00 -5.06266356e-02 2.26527810e-01
3.76318097e-01 4.28882211e-01 1.37737483e-01 8.80408645e-01
-3.34183961e-01 3.33405226e-01 8.31185579e-02 8.46577659e-02
2.32161954e-02 3.74082297e-01 6.44816086e-03 -7.14293346e-02
3.47146749e-01 1.24710524e+00 -8.86491597e-01 -7.37176120e-01
-2.45198548e-01 2.83358812e-01 -1.43909842e-01 -7.63315260e-02
-2.09744811e-01 -4.57265526e-01 -2.58136541e-01 5.22135496e-01
4.15535033e-01 -3.67553562e-01 7.40234554e-01 -6.53023779e-01
2.20771447e-01 8.33146214e-01 -1.05106544e+00 1.44414604e+00
-1.40164018e-01 6.39295280e-01 -4.97834951e-01 -8.28529418e-01
1.06032300e+00 5.61192513e-01 1.94924667e-01 -5.69196165e-01
4.16024514e-02 1.56758621e-01 1.02069855e-01 -6.52002454e-01
9.11296785e-01 2.53422379e-01 -1.41366780e-01 2.21598059e-01
6.01718187e-01 1.43605277e-01 7.19554245e-01 8.99713695e-01
1.43009043e+00 2.29006499e-01 8.38628531e-01 -2.32267916e-01
4.01556730e-01 2.39774480e-01 4.58933532e-01 5.77208400e-01
3.07668537e-01 1.22203596e-01 1.03982747e+00 -2.80802250e-01
-7.14857161e-01 -6.97990358e-01 3.67983105e-03 6.28187716e-01
-1.24400467e-01 -1.19863081e+00 -6.69971764e-01 -1.31043613e+00
2.83457130e-01 9.39451456e-01 -8.07832181e-01 -2.32050493e-01
-7.33189106e-01 -6.06105983e-01 7.51830935e-01 6.21983945e-01
4.24590051e-01 -1.13027620e+00 -3.20974499e-01 1.33161247e-01
-2.98570246e-01 -1.73597968e+00 1.54224619e-01 2.37108678e-01
-3.53347480e-01 -1.49190998e+00 3.32352996e-01 -4.99021858e-01
3.60031754e-01 -2.88937122e-01 1.81616879e+00 2.30571702e-01
1.25198781e-01 1.23257160e-01 -7.72245824e-01 -4.72637028e-01
-3.55871052e-01 4.75882083e-01 -2.75447313e-02 -4.60470974e-01
7.04327583e-01 -5.29187500e-01 -4.19546776e-02 -2.70431608e-01
-5.81477940e-01 -4.84670810e-02 6.42620802e-01 6.50431156e-01
2.92984456e-01 2.39404172e-01 8.27345371e-01 -1.67263865e+00
6.38462305e-01 -4.73386139e-01 -2.84381360e-01 6.50481701e-01
-1.08744395e+00 1.23623028e-01 3.89874160e-01 -4.06378601e-03
-9.04593766e-01 -3.36413890e-01 2.26309568e-01 1.45046309e-01
-2.03003343e-02 1.20057547e+00 -1.67536721e-01 1.81536108e-01
8.86372924e-01 -2.91039854e-01 -5.08107066e-01 -3.15206975e-01
5.93647063e-01 3.87588352e-01 6.41122341e-01 -5.95796585e-01
8.37942898e-01 -1.73679397e-01 7.00568482e-02 -3.65849078e-01
-1.15525293e+00 -2.80444503e-01 -9.26987231e-01 1.81430861e-01
7.22420454e-01 -8.14362049e-01 -4.99997407e-01 -1.32939629e-02
-1.14781702e+00 -1.08233206e-01 -4.67364728e-01 2.12389410e-01
-1.24611840e-01 1.91541046e-01 -6.12315774e-01 -4.97159719e-01
-7.10117817e-01 -4.44627553e-01 7.23096967e-01 3.49172115e-01
-6.15548849e-01 -9.93361950e-01 -9.41286609e-02 5.15263140e-01
7.67684802e-02 7.50314474e-01 1.10514009e+00 -1.07395411e+00
-4.24765825e-01 -3.89265954e-01 -5.77699661e-01 8.45051464e-03
2.86077797e-01 6.94218576e-02 -7.17241406e-01 9.76467282e-02
-8.46046329e-01 -3.27145994e-01 6.30571008e-01 -1.06469207e-01
4.10295010e-01 -6.98477864e-01 -5.44462264e-01 1.54476181e-01
1.60869753e+00 1.50692478e-01 6.52829647e-01 6.92802012e-01
8.11595857e-01 6.56509995e-01 6.63638651e-01 3.46345119e-02
8.03062022e-01 6.98704004e-01 3.02205324e-01 2.07766756e-01
-4.79768634e-01 -4.81953561e-01 -5.39458506e-02 5.51761866e-01
-3.48186344e-01 -3.65130931e-01 -1.19135427e+00 7.13102043e-01
-1.88361800e+00 -9.32319820e-01 -4.66646880e-01 1.86028850e+00
1.29886723e+00 4.43934947e-01 -9.44086760e-02 2.48943299e-01
5.24900973e-01 -6.74069971e-02 -1.44423230e-03 -3.31614226e-01
-3.53491277e-01 5.60612023e-01 6.16698623e-01 4.05579567e-01
-1.18201780e+00 1.18493772e+00 5.87029696e+00 5.95968485e-01
-5.46151102e-01 -7.15277866e-02 1.16728425e-01 2.71720976e-01
1.41963065e-02 3.03555429e-01 -1.12227166e+00 8.21764171e-02
1.14413559e+00 -5.56946218e-01 1.20795719e-01 6.84837401e-01
-5.47123969e-01 -1.76935524e-01 -1.17971981e+00 6.00099444e-01
6.89506456e-02 -1.62626517e+00 1.13849796e-01 2.40070708e-02
6.77838147e-01 -5.75410388e-02 -5.70746064e-01 6.53125107e-01
6.25571072e-01 -9.97290075e-01 5.41261256e-01 4.43169504e-01
6.42060518e-01 -7.83071458e-01 1.32255220e+00 -9.29877907e-03
-1.10162199e+00 3.27005863e-01 -1.91877052e-01 -2.15514421e-01
5.80117516e-02 8.01956534e-01 -1.42806637e+00 1.44709671e+00
8.76065075e-01 9.52347815e-01 -9.97173905e-01 5.20295382e-01
-7.37950742e-01 5.28354824e-01 -3.07664871e-01 -1.57428622e-01
-7.26273656e-02 2.51126111e-01 4.66415554e-01 1.49928439e+00
-5.43068498e-02 8.57344344e-02 2.63931300e-03 5.16649067e-01
-5.29496670e-01 3.58773947e-01 -6.30925298e-01 -2.76923358e-01
8.60718071e-01 1.40335512e+00 -4.16056037e-01 -6.65474772e-01
-2.82427937e-01 3.16101909e-01 9.29219365e-01 2.47570187e-01
-5.73276103e-01 -7.89974868e-01 3.01432580e-01 1.91603824e-02
4.96916085e-01 1.34227663e-01 -1.71266690e-01 -1.00100231e+00
3.71594071e-01 -6.11894250e-01 8.53758931e-01 -8.52133572e-01
-1.19894588e+00 7.89253175e-01 3.42944771e-01 -6.65381849e-01
-3.79913867e-01 -4.83629644e-01 -2.45967031e-01 6.76554739e-01
-1.47983778e+00 -1.35743332e+00 -2.74412274e-01 2.97164589e-01
-3.47499967e-01 -1.73576251e-01 1.07413936e+00 3.58754545e-01
-6.85500562e-01 8.45959604e-01 -4.77793843e-01 7.54206657e-01
7.17114925e-01 -1.68231583e+00 4.73965287e-01 8.87322664e-01
4.77495760e-01 9.36962426e-01 6.83950126e-01 -1.04332840e+00
-1.18107319e+00 -1.11923409e+00 1.63564456e+00 -8.08526218e-01
9.55505431e-01 -5.46948463e-02 -1.04962134e+00 1.03153181e+00
6.55520439e-01 -1.10427476e-01 6.66051984e-01 7.53237188e-01
-7.23343790e-01 -2.14288849e-02 -1.27753448e+00 2.70294189e-01
1.19290817e+00 -3.99937510e-01 -8.28270316e-01 3.15361321e-01
9.46029723e-01 -6.65202379e-01 -1.63141990e+00 6.22846663e-01
3.84785146e-01 -6.72002673e-01 9.12053525e-01 -8.02441657e-01
5.02327263e-01 -2.53139168e-01 -1.16252370e-01 -1.25018120e+00
-1.75295442e-01 -4.34810340e-01 -9.43534136e-01 1.73820019e+00
1.03828013e+00 -7.48373032e-01 7.73915112e-01 6.34221792e-01
-8.10936540e-02 -8.49693537e-01 -9.00061846e-01 -8.45515072e-01
-4.04610783e-02 3.40775549e-02 8.48475516e-01 1.36912429e+00
5.40582836e-01 8.84748578e-01 -4.63974401e-02 1.74860284e-01
5.28949201e-01 2.55864531e-01 6.53408527e-01 -1.39849746e+00
-1.53250486e-01 -1.39218390e-01 -6.54351354e-01 -2.15100035e-01
2.19044656e-01 -1.17994976e+00 -5.22257984e-01 -2.17093706e+00
2.15896338e-01 -5.71492314e-01 -1.96065858e-01 1.06006980e+00
-3.16869497e-01 -6.54924288e-03 3.22433934e-02 -1.43793477e-02
-5.23552537e-01 9.57386121e-02 8.23842168e-01 -2.26282954e-01
4.58822027e-02 -6.03527546e-01 -1.00767565e+00 4.59337384e-01
8.52979422e-01 -6.63068414e-01 -2.84165680e-01 -3.57996643e-01
6.93603933e-01 -9.39394832e-02 1.12814642e-01 -7.91926265e-01
3.43121409e-01 1.01180911e-01 2.81110078e-01 -5.85651338e-01
2.52413273e-01 -5.81365228e-01 1.21518090e-01 1.22796357e-01
-3.69316548e-01 2.44271439e-02 2.14440823e-01 3.66918236e-01
-2.75045663e-01 -2.75867999e-01 1.41580507e-01 -2.46570017e-02
-7.11414158e-01 -5.68493120e-02 3.30978066e-01 4.31098074e-01
8.14773381e-01 -3.52550745e-02 -1.12056720e+00 -1.56131834e-01
-6.86413884e-01 2.42553204e-01 1.06338628e-01 4.46719587e-01
4.64194417e-01 -1.25980258e+00 -8.29784393e-01 -2.48555526e-01
3.00717235e-01 3.07826381e-02 -2.90090382e-01 7.57382393e-01
-4.32939619e-01 5.37872136e-01 -7.61675090e-03 1.52716581e-02
-1.43294168e+00 4.36604798e-01 1.01769060e-01 -8.74371588e-01
-7.18651474e-01 6.86872840e-01 -6.14433825e-01 -4.19031054e-01
-1.87491179e-02 -3.95804822e-01 -8.72622490e-01 2.27400273e-01
2.12239221e-01 2.61471659e-01 5.52318633e-01 -5.91015041e-01
-5.20311892e-01 3.02291274e-01 -3.60409766e-01 2.90952533e-01
1.55321717e+00 2.32211292e-01 -4.54247206e-01 1.72878519e-01
8.71334612e-01 1.68990299e-01 -2.28205830e-01 -3.85729730e-01
6.48710430e-01 -1.88852355e-01 -2.41382569e-01 -1.29053688e+00
-9.90377545e-01 1.01221010e-01 -1.35669783e-01 6.69555545e-01
6.65395260e-01 3.75237286e-01 7.52807498e-01 6.52634621e-01
4.97635633e-01 -9.15899277e-01 -4.40812767e-01 6.47167683e-01
9.03484643e-01 -1.07053459e+00 5.07386804e-01 -1.21310878e+00
-5.98174751e-01 8.97175610e-01 8.01069498e-01 -8.44126493e-02
5.50325513e-01 2.90078670e-01 -5.68299554e-03 -7.47911036e-01
-8.60093951e-01 -2.96771944e-01 4.99951214e-01 7.48003125e-01
6.32285237e-01 1.08205214e-01 -4.32176352e-01 7.43055761e-01
-6.40652239e-01 -1.66188359e-01 6.74783349e-01 1.02435315e+00
-5.16500250e-02 -1.18431497e+00 1.95811689e-01 8.14489186e-01
-6.51299357e-01 -4.21743155e-01 -7.50072420e-01 1.21682036e+00
1.57060102e-01 1.03596044e+00 -3.17471266e-01 -5.26082575e-01
7.31257200e-01 2.45497286e-01 5.53049207e-01 -8.77861559e-01
-9.20770347e-01 -6.59106731e-01 1.31370211e+00 -3.50823045e-01
-7.30429292e-01 -6.26891434e-01 -1.27631986e+00 -2.22625509e-01
-6.24655247e-01 4.43099499e-01 3.51196527e-01 1.00816965e+00
4.63696599e-01 6.50405586e-01 -2.15011507e-01 9.36990902e-02
1.13066494e-01 -1.24649036e+00 -4.17962462e-01 5.82225859e-01
-2.81972528e-01 -9.31257248e-01 -4.43614386e-02 8.94645378e-02] | [9.327996253967285, 8.559673309326172] |
31daa8d5-a386-44aa-a0d7-ee96abae922f | learning-depth-from-monocular-videos-using | 1712.00175 | null | http://arxiv.org/abs/1712.00175v1 | http://arxiv.org/pdf/1712.00175v1.pdf | Learning Depth from Monocular Videos using Direct Methods | The ability to predict depth from a single image - using recent advances in
CNNs - is of increasing interest to the vision community. Unsupervised
strategies to learning are particularly appealing as they can utilize much
larger and varied monocular video datasets during learning without the need for
ground truth depth or stereo. In previous works, separate pose and depth CNN
predictors had to be determined such that their joint outputs minimized the
photometric error. Inspired by recent advances in direct visual odometry (DVO),
we argue that the depth CNN predictor can be learned without a pose CNN
predictor. Further, we demonstrate empirically that incorporation of a
differentiable implementation of DVO, along with a novel depth normalization
strategy - substantially improves performance over state of the art that use
monocular videos for training. | ['Rui Zhu', 'Chaoyang Wang', 'Jose Miguel Buenaposada', 'Simon Lucey'] | 2017-12-01 | learning-depth-from-monocular-videos-using-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Wang_Learning_Depth_From_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Wang_Learning_Depth_From_CVPR_2018_paper.pdf | cvpr-2018-6 | ['depth-and-camera-motion'] | ['computer-vision'] | [ 3.06735128e-01 2.12132528e-01 -1.91823706e-01 -6.03064299e-01
-4.68841463e-01 -5.87266922e-01 9.04467106e-01 -3.08326036e-01
-4.89875287e-01 8.04540455e-01 2.55944014e-01 -2.96999849e-02
1.88883528e-01 -5.52141845e-01 -1.02411890e+00 -5.31495392e-01
1.63692430e-01 3.65524381e-01 2.15300784e-01 1.30976504e-02
3.80832106e-01 7.33749807e-01 -1.65932381e+00 -1.02432854e-01
2.94662833e-01 1.11101127e+00 3.70601118e-01 7.88084745e-01
2.06553057e-01 9.39289093e-01 -2.30014041e-01 -8.14283267e-02
7.02074051e-01 -2.09587127e-01 -7.73965538e-01 2.45059311e-01
1.17557442e+00 -7.81644762e-01 -7.31841922e-01 7.52317071e-01
4.96069759e-01 1.02423705e-01 6.12499654e-01 -1.05695021e+00
-3.58763903e-01 -1.58960879e-01 -3.53847563e-01 1.65591419e-01
3.52151841e-01 2.38616347e-01 1.04600370e+00 -7.84856021e-01
1.00741434e+00 1.07878911e+00 6.64614201e-01 6.53911114e-01
-1.45430458e+00 -4.54824835e-01 1.28284231e-01 1.86727524e-01
-1.05102122e+00 -4.65903848e-01 7.70718932e-01 -5.48379123e-01
1.51126552e+00 -2.94698238e-01 9.24953341e-01 1.03949809e+00
1.60433099e-01 6.72926545e-01 8.89506698e-01 -4.30858463e-01
6.25435710e-02 -6.98211193e-02 -3.67096514e-01 7.66678989e-01
2.42934033e-01 5.46815455e-01 -7.16232300e-01 3.12875420e-01
1.15838218e+00 2.53304876e-02 -4.16834265e-01 -1.10507023e+00
-9.09752071e-01 7.77212858e-01 6.55080676e-01 -2.84808785e-01
-1.37954012e-01 5.62181771e-01 1.34729311e-01 3.58742326e-01
4.02723700e-01 5.65618217e-01 -6.35959744e-01 -2.73789316e-01
-7.89302230e-01 2.33330578e-01 7.43404090e-01 1.14798880e+00
1.28609240e+00 6.28933758e-02 3.51833314e-01 4.65422720e-01
2.53461808e-01 2.41906524e-01 3.34200293e-01 -1.43075097e+00
4.03678685e-01 5.15325129e-01 3.61481905e-02 -6.66321993e-01
-4.37412858e-01 -2.82305151e-01 -4.69156772e-01 6.49787426e-01
5.29557407e-01 -1.33549586e-01 -1.03354442e+00 1.67789197e+00
2.58507490e-01 2.87699640e-01 1.56276096e-02 9.83027399e-01
4.90301281e-01 1.62015557e-01 -4.18585926e-01 1.90833524e-01
5.89069486e-01 -7.78121829e-01 -1.00267567e-01 -6.10579312e-01
5.23226917e-01 -6.22235775e-01 7.33146667e-01 5.58710217e-01
-9.38867748e-01 -4.62962240e-01 -1.38131237e+00 -6.97487831e-01
-1.09644435e-01 -1.15460999e-01 6.36928082e-01 4.46475834e-01
-1.42048120e+00 8.07524145e-01 -9.56356525e-01 -5.48474967e-01
3.41271013e-01 6.86365128e-01 -5.23979306e-01 -2.70327091e-01
-6.81003749e-01 1.04864883e+00 1.78956285e-01 -2.98335012e-02
-1.22173893e+00 -5.61784267e-01 -1.26297760e+00 -3.72003913e-01
4.88766022e-02 -9.86743808e-01 1.33011365e+00 -1.02141201e+00
-1.61466670e+00 1.04339015e+00 -2.02092350e-01 -5.53325951e-01
6.30827129e-01 -4.52604800e-01 4.49425340e-01 3.54004562e-01
-1.51252940e-01 1.25485361e+00 9.16843593e-01 -1.25208056e+00
-6.29319549e-01 -3.87246639e-01 5.31127453e-01 4.64386880e-01
-1.10378966e-01 -6.13073468e-01 -3.95116657e-01 -2.52412073e-02
4.71367687e-01 -1.06594753e+00 -2.94313461e-01 5.20756304e-01
-1.72214240e-01 2.61565000e-01 7.69867659e-01 -3.41730356e-01
3.44438046e-01 -1.79296172e+00 3.89756203e-01 -5.68615757e-02
3.41912419e-01 -9.39728469e-02 -1.25704017e-02 8.61804262e-02
-5.97779378e-02 -3.24404716e-01 1.49532616e-01 -6.73166931e-01
-1.62303418e-01 5.13660431e-01 -1.81311607e-01 8.66008341e-01
4.14167464e-01 7.82694638e-01 -8.60641479e-01 -2.32427046e-01
5.97560763e-01 7.68384933e-01 -9.91687357e-01 3.40815902e-01
-3.63093525e-01 5.93559563e-01 5.76433018e-02 5.00924885e-01
4.92080092e-01 -3.17632049e-01 1.07334286e-01 -9.14629251e-02
-2.04080239e-01 6.15515888e-01 -9.87473130e-01 1.98297179e+00
-3.97122324e-01 1.04061484e+00 -5.60525581e-02 -8.87297571e-01
9.35103953e-01 1.90515146e-01 5.23421705e-01 -6.33636236e-01
3.38766336e-01 1.11926556e-01 -2.56293535e-01 -2.61972725e-01
4.92290646e-01 -3.15089524e-01 4.36953396e-01 -1.79411676e-02
4.22505170e-01 -6.83918715e-01 -8.80180597e-02 -3.35964113e-02
1.10036719e+00 8.17656577e-01 2.82101929e-01 -8.74413177e-02
3.29955190e-01 2.57477462e-01 5.17394900e-01 3.94909769e-01
-1.81743756e-01 1.00621808e+00 4.50546652e-01 -4.91907030e-01
-1.26688731e+00 -1.02049744e+00 -2.01349333e-01 6.22161448e-01
2.11312592e-01 -7.67312199e-02 -3.94890994e-01 -2.98906207e-01
1.63725764e-01 4.98192422e-02 -5.94865978e-01 8.08772519e-02
-7.32512653e-01 -3.11980605e-01 3.06957453e-01 7.61627376e-01
3.65833491e-01 -6.79311037e-01 -8.52729201e-01 1.69264168e-01
2.15295464e-01 -1.42013931e+00 -1.06288046e-01 6.51519060e-01
-1.14745629e+00 -1.21601713e+00 -7.56652415e-01 -9.67718840e-01
6.10682309e-01 4.24487025e-01 1.13781595e+00 -1.94196269e-01
-2.55482078e-01 7.08059311e-01 -2.04891916e-02 -2.19526410e-01
9.18320101e-03 1.91964954e-01 -5.88575676e-02 -3.74167711e-01
3.90237898e-01 -8.39449227e-01 -7.63045669e-01 1.90528128e-02
-5.34867764e-01 2.66122818e-01 3.82667184e-01 8.02469313e-01
5.84573030e-01 -7.61736631e-01 -1.02252727e-02 -6.85072184e-01
-1.29940063e-01 -2.01181009e-01 -9.84617054e-01 -4.67135131e-01
-6.72068775e-01 1.62219569e-01 5.77561557e-01 -2.71740943e-01
-7.25069880e-01 6.18271649e-01 -9.07022431e-02 -7.02579200e-01
-2.34115243e-01 1.97220147e-01 -6.07794821e-02 -5.22104263e-01
7.85937369e-01 4.51007895e-02 1.30058408e-01 -3.05930912e-01
3.37877989e-01 1.03247859e-01 7.38761485e-01 -4.10665929e-01
8.82605076e-01 8.86860669e-01 4.20093924e-01 -8.10208976e-01
-7.93580055e-01 -5.80912590e-01 -1.07768488e+00 1.51791386e-02
9.13081944e-01 -1.43903911e+00 -7.20172405e-01 3.01576912e-01
-1.25296760e+00 -5.57618499e-01 -1.48639247e-01 6.94479048e-01
-9.69937384e-01 3.17988068e-01 -6.67764127e-01 -5.36956966e-01
2.11519763e-01 -1.04314947e+00 1.11339211e+00 2.13340983e-01
-1.89249873e-01 -1.17388439e+00 2.21793935e-01 9.05526802e-02
1.35952011e-01 3.11836064e-01 4.79327559e-01 -1.56194149e-02
-1.17142391e+00 -1.99825540e-02 -3.79459739e-01 4.61411119e-01
9.84294266e-02 -2.17642933e-01 -1.34928274e+00 -3.77769917e-01
-4.94799167e-02 -5.55726051e-01 8.74594510e-01 4.99872804e-01
7.38262355e-01 1.89252809e-01 -5.97580485e-02 1.21499860e+00
1.82611930e+00 -4.04865108e-02 6.27444506e-01 7.07324386e-01
1.07362652e+00 4.45548356e-01 2.61092097e-01 4.75810200e-01
5.04699230e-01 5.74522793e-01 8.00322890e-01 -7.95359462e-02
-1.95339486e-01 -2.86945432e-01 2.66199559e-01 4.48276013e-01
-1.21190697e-01 1.61195129e-01 -8.70541036e-01 6.96462035e-01
-1.60869479e+00 -6.87503695e-01 1.24511354e-01 2.11462712e+00
7.67570972e-01 2.54171759e-01 -9.01389048e-02 2.19330341e-01
1.19662613e-01 2.16701075e-01 -7.00131178e-01 -2.51162082e-01
-1.76449299e-01 4.21110243e-01 9.36207116e-01 7.53567517e-01
-1.08141243e+00 9.12282526e-01 7.01851225e+00 -2.29040116e-01
-1.46370876e+00 -1.34141728e-01 3.16533476e-01 -4.06848609e-01
-1.92536414e-01 2.31195122e-01 -9.46995974e-01 -2.31531337e-01
7.26592958e-01 1.58731028e-01 4.44575250e-01 1.07689643e+00
1.59208268e-01 -2.95709372e-01 -1.54939497e+00 1.27034438e+00
2.34158918e-01 -1.33665311e+00 -1.75843120e-01 2.55874246e-01
1.11072385e+00 5.23979962e-01 2.34351819e-03 -1.00486711e-01
4.35626894e-01 -1.09549844e+00 6.82344913e-01 3.57459635e-01
6.95705593e-01 -4.69819933e-01 5.17783940e-01 2.73644358e-01
-1.14848530e+00 -2.72455812e-02 -6.25959337e-01 -5.39476573e-01
4.74272706e-02 3.47825468e-01 -9.19995248e-01 3.44597101e-01
5.60728312e-01 1.26666737e+00 -5.51099539e-01 1.05422306e+00
-3.86906713e-01 2.07790598e-01 -3.96727711e-01 2.41072759e-01
2.70275265e-01 2.11585723e-02 1.53275132e-01 9.03919399e-01
1.73283666e-01 -2.06535682e-01 5.77973425e-02 6.68428242e-01
-7.33621866e-02 -2.84418851e-01 -8.50296021e-01 4.08391744e-01
1.55242726e-01 9.28139508e-01 -2.27288827e-01 -1.58518985e-01
-7.01892912e-01 1.01994061e+00 6.58608735e-01 2.74115175e-01
-4.23985124e-01 8.45162868e-02 1.07131612e+00 2.18551934e-01
6.70888603e-01 -7.73902416e-01 -6.27767742e-01 -1.26475620e+00
-4.76563768e-03 -5.56712866e-01 -4.34273034e-02 -9.36920583e-01
-9.52203274e-01 3.80273670e-01 -1.22418568e-01 -1.35596263e+00
-6.52754843e-01 -1.00755131e+00 -3.05168122e-01 9.08049226e-01
-2.04735446e+00 -8.97504210e-01 -6.91798627e-01 6.53346419e-01
2.92775810e-01 9.68128517e-02 5.31735897e-01 1.32271335e-01
-1.29537180e-01 2.65633523e-01 -8.15494508e-02 7.41845369e-02
7.14344919e-01 -1.38540888e+00 5.38980663e-01 7.32135057e-01
1.96239769e-01 4.80215967e-01 7.70723939e-01 -3.63413602e-01
-1.63891995e+00 -8.82684410e-01 8.07487249e-01 -7.36528397e-01
3.88763338e-01 -3.35112065e-01 -5.50812840e-01 9.81397331e-01
7.24932998e-02 3.00817192e-01 8.47643167e-02 -8.67604613e-02
-3.93836737e-01 -9.14308578e-02 -8.83871853e-01 4.49324667e-01
1.24722087e+00 -9.68642771e-01 -4.44119453e-01 5.54405153e-02
5.05067050e-01 -7.61939287e-01 -5.47646999e-01 2.17535779e-01
7.23711848e-01 -1.32248092e+00 1.09297276e+00 -3.03385794e-01
6.18465841e-01 -2.29390770e-01 -4.14566457e-01 -1.09013915e+00
-1.07872508e-01 -3.72862190e-01 -2.37521902e-01 7.27449894e-01
1.37294501e-01 -5.10148823e-01 1.28405857e+00 5.78260481e-01
-1.89883187e-01 -6.45544648e-01 -8.57918680e-01 -7.23853290e-01
4.20943014e-02 -4.20496851e-01 8.79145265e-02 7.01030552e-01
-2.50823677e-01 2.32706070e-01 -4.93740737e-01 2.99457252e-01
5.91952741e-01 -1.62839144e-01 1.18726766e+00 -1.32324040e+00
-4.92837518e-01 -2.08061099e-01 -1.00251615e+00 -1.79432154e+00
1.64395764e-01 -6.85664713e-01 3.03208858e-01 -1.52141464e+00
-1.91533208e-01 -4.11754064e-02 9.22882836e-03 3.75588059e-01
1.93940803e-01 5.45095265e-01 1.15688421e-01 9.97412279e-02
-2.16589570e-01 6.39688194e-01 1.27817214e+00 1.47706732e-01
-6.85394928e-02 -3.93872261e-01 -3.75701040e-01 7.46980071e-01
4.81002450e-01 -4.01885390e-01 -6.72738433e-01 -7.17698038e-01
2.09835961e-01 1.66599452e-02 6.24614418e-01 -1.25937045e+00
2.47820556e-01 -1.21674336e-01 6.91267371e-01 -4.70245570e-01
7.99346328e-01 -7.45167077e-01 -1.33992419e-01 3.77247602e-01
-7.41110593e-02 1.68949649e-01 7.04410672e-02 5.34787059e-01
-1.74750969e-01 -1.27857402e-01 8.39863956e-01 -2.73935109e-01
-1.12122595e+00 3.83414775e-01 -1.59051999e-01 5.68949468e-02
7.08958447e-01 -5.21638453e-01 -1.64300591e-01 -5.37922084e-01
-5.11803806e-01 -9.22747552e-02 9.50222611e-01 2.89581150e-01
6.35454893e-01 -1.15089393e+00 -2.53026813e-01 2.96808839e-01
4.49790843e-02 4.28089589e-01 -3.90404463e-01 5.95530987e-01
-8.80656600e-01 5.30822039e-01 -4.83150870e-01 -9.12897289e-01
-9.32518899e-01 1.33654028e-01 6.44322217e-01 1.60375953e-01
-7.67375648e-01 9.05675411e-01 3.51434469e-01 -4.78771120e-01
3.62944216e-01 -6.09012187e-01 2.23444983e-01 -1.38853297e-01
2.51286149e-01 7.60039017e-02 7.53726736e-02 -5.23497701e-01
-2.64714241e-01 8.03735852e-01 1.04410671e-01 -3.89346510e-01
1.45275998e+00 -3.64353120e-01 2.90982485e-01 4.35185522e-01
1.44829559e+00 -2.98444748e-01 -2.19912767e+00 -1.85711816e-01
1.55660603e-02 -6.88933611e-01 1.66069955e-01 -3.00580025e-01
-9.13496912e-01 1.11531913e+00 4.60519195e-01 -3.87804121e-01
8.17930996e-01 -7.30610788e-02 5.56122422e-01 4.58628803e-01
5.21415353e-01 -8.95341277e-01 2.97229171e-01 9.68282819e-01
5.62614441e-01 -1.47621632e+00 2.25526959e-01 -3.00099999e-01
-2.53932565e-01 1.25870991e+00 8.34535122e-01 -5.76967597e-01
6.68701828e-01 2.63450354e-01 2.96729147e-01 -4.91539799e-02
-7.77826548e-01 -4.72509980e-01 3.19230795e-01 8.13307941e-01
4.90901649e-01 -5.08303881e-01 2.80275822e-01 -1.62540272e-01
-3.45492452e-01 8.82479772e-02 5.56194484e-01 1.08725452e+00
-4.43289012e-01 -1.07919633e+00 -4.05237675e-02 1.61804572e-01
-6.99847937e-03 -8.60183239e-02 -2.88957685e-01 8.40001583e-01
1.78756237e-01 6.43814266e-01 2.37768874e-01 -4.37704086e-01
1.43642798e-01 -1.35571295e-02 9.79216337e-01 -8.60996246e-01
-3.05424929e-01 2.76605114e-02 -4.54403944e-02 -7.32319057e-01
-6.36954069e-01 -7.40100443e-01 -1.11974776e+00 -2.76618958e-01
-5.39898127e-02 -6.03179097e-01 8.11714768e-01 1.23865926e+00
5.54555729e-02 1.46677315e-01 3.88633639e-01 -1.52606249e+00
-3.21733326e-01 -6.61541760e-01 -4.35177207e-01 2.06044957e-01
8.02049637e-01 -7.83449888e-01 -4.26431268e-01 1.67865530e-01] | [8.579565048217773, -2.4307265281677246] |
47c09689-2d42-41f9-a563-159c5719de40 | learning-curves-for-gaussian-process-1 | 2110.12231 | null | https://arxiv.org/abs/2110.12231v2 | https://arxiv.org/pdf/2110.12231v2.pdf | Learning curves for Gaussian process regression with power-law priors and targets | We characterize the power-law asymptotics of learning curves for Gaussian process regression (GPR) under the assumption that the eigenspectrum of the prior and the eigenexpansion coefficients of the target function follow a power law. Under similar assumptions, we leverage the equivalence between GPR and kernel ridge regression (KRR) to show the generalization error of KRR. Infinitely wide neural networks can be related to GPR with respect to the neural network GP kernel and the neural tangent kernel, which in several cases is known to have a power-law spectrum. Hence our methods can be applied to study the generalization error of infinitely wide neural networks. We present toy experiments demonstrating the theory. | ['Guido Montúfar', 'Pradeep Kr. Banerjee', 'Hui Jin'] | 2021-10-23 | learning-curves-for-gaussian-process | https://openreview.net/forum?id=KeI9E-gsoB | https://openreview.net/pdf?id=KeI9E-gsoB | iclr-2022-4 | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [-5.01909666e-02 2.60265261e-01 5.43519296e-02 -9.61558074e-02
-4.40219969e-01 -5.04046321e-01 1.75808430e-01 9.54840519e-03
-4.29132253e-01 5.08869767e-01 -2.95179218e-01 -5.99767506e-01
-4.12494719e-01 -7.05186307e-01 -8.54217410e-01 -1.14387214e+00
-3.56058389e-01 2.08438590e-01 9.32552665e-02 1.51061416e-02
-3.18365805e-02 5.49382389e-01 -9.09070730e-01 -6.06507540e-01
7.95460582e-01 1.03122890e+00 -3.66356909e-01 7.95017362e-01
3.96552801e-01 3.34560305e-01 -2.49417111e-01 -3.50670308e-01
3.35530639e-01 7.35707656e-02 -4.08387959e-01 -3.39402825e-01
2.77632654e-01 6.47157952e-02 -6.55571818e-01 1.38505614e+00
2.29697496e-01 5.84023058e-01 1.28468740e+00 -1.27691698e+00
-9.84997571e-01 3.08728814e-01 -8.95765305e-01 2.47335002e-01
-1.17499076e-01 -3.52845043e-01 7.99927413e-01 -6.08333766e-01
-2.48586759e-02 1.16759658e+00 1.42748213e+00 5.32707393e-01
-1.49067605e+00 -5.85131347e-01 -1.21514551e-01 -4.77116346e-01
-1.56156838e+00 -1.68645248e-01 3.46641809e-01 -4.58567470e-01
6.02872014e-01 -3.63268673e-01 2.91902959e-01 1.33351874e+00
6.21435821e-01 2.77141213e-01 1.03072667e+00 -3.59770328e-01
3.91420275e-01 1.98795259e-01 7.46147990e-01 7.32739747e-01
5.25265872e-01 2.75072068e-01 -1.01200091e-02 -4.44037616e-01
1.29976261e+00 -4.74874377e-02 -5.19033968e-01 -3.72911811e-01
-5.69284618e-01 1.06738985e+00 2.45906234e-01 1.04843453e-01
-4.25905228e-01 6.66434348e-01 -1.05052469e-02 3.20467442e-01
5.04242837e-01 1.94647670e-01 -3.51918757e-01 -1.33238837e-01
-4.07629222e-01 4.95265126e-02 1.26390696e+00 1.12197745e+00
5.25097668e-01 3.44723135e-01 2.73558080e-01 9.23000991e-01
4.43136811e-01 9.01126623e-01 4.02758181e-01 -8.97015154e-01
2.68201411e-01 -2.01800272e-01 -1.86398823e-03 -5.61888516e-01
-5.13259351e-01 -5.56898832e-01 -8.55412245e-01 1.28511786e-01
8.02629352e-01 -5.07428467e-01 -6.59489810e-01 1.83372295e+00
-1.94917277e-01 2.72499084e-01 4.00906473e-01 3.00227404e-01
1.64607629e-01 6.07482612e-01 3.35334539e-01 -4.24469747e-02
1.02184808e+00 -8.47328901e-02 -4.14268553e-01 1.27280682e-01
3.75035614e-01 -2.07076371e-01 1.08927095e+00 1.55867383e-01
-1.00372922e+00 -2.92744935e-01 -1.00904226e+00 2.03188837e-01
-4.13812436e-02 -5.76755218e-02 5.17082751e-01 7.57968545e-01
-1.03075492e+00 1.01187456e+00 -1.08407235e+00 -2.47614145e-01
4.13373619e-01 1.80065840e-01 -3.20491850e-01 1.27551973e-01
-9.48923290e-01 7.39278615e-01 5.60008764e-01 1.69249207e-01
-6.52973056e-01 -1.03358114e+00 -5.98628640e-01 1.97703242e-01
-2.10080985e-02 -4.47527498e-01 1.27993667e+00 -5.16795933e-01
-1.55599642e+00 3.26872110e-01 -6.00147098e-02 -4.27812964e-01
4.06364739e-01 -3.66931021e-01 -2.62331814e-01 2.96113044e-01
-5.79291403e-01 5.22621870e-02 1.19181907e+00 -8.43476593e-01
-4.16871071e-01 -5.42345822e-01 -3.43735069e-01 -1.42163008e-01
-1.13681145e-01 -3.32028627e-01 2.23585770e-01 -5.14733434e-01
1.38121352e-01 -1.07567954e+00 -8.21006373e-02 -2.42491603e-01
-2.64563888e-01 -3.97614330e-01 4.52328652e-01 -5.46604276e-01
8.00196826e-01 -2.69116426e+00 -2.58153796e-01 5.08854985e-01
2.72803605e-01 -4.78492469e-01 -2.52593495e-02 2.21508652e-01
-2.54564166e-01 9.63808596e-02 -2.47366190e-01 -9.87030286e-03
1.34534329e-01 2.00932220e-01 -8.88120770e-01 1.11374807e+00
1.32341132e-01 7.21050441e-01 -7.00042784e-01 3.28352712e-02
-4.98843998e-01 7.29133189e-01 -7.83495083e-02 -2.63003316e-02
5.66667058e-02 -1.02726184e-01 -4.03962314e-01 1.78708136e-01
5.83644867e-01 -2.18840912e-01 -4.67392057e-01 1.20244727e-01
1.38849363e-01 -3.78355891e-01 -8.16010773e-01 7.76767313e-01
-4.04614866e-01 9.04516578e-01 -1.12367585e-01 -8.80828202e-01
9.87905562e-01 4.53349710e-01 1.92642763e-01 1.65321440e-01
3.04386839e-02 1.88735396e-01 7.26753920e-02 -1.01386987e-01
1.59970015e-01 -5.95775247e-01 1.49880484e-01 2.09277228e-01
3.46576691e-01 1.34120136e-01 -3.95442933e-01 -1.16083302e-01
8.85300815e-01 2.62941837e-01 3.14789563e-01 -5.50563276e-01
1.68467835e-02 -4.27748263e-01 1.74391627e-01 1.03242278e+00
-1.14676185e-01 2.63698637e-01 8.47784102e-01 1.88651383e-01
-1.25292242e+00 -1.60662186e+00 -5.34948230e-01 1.08241284e+00
-1.84918895e-01 1.22784056e-01 -7.50214279e-01 -2.49090999e-01
1.89792663e-01 9.77083683e-01 -8.52515280e-01 -4.14259464e-01
-2.23879874e-01 -8.72536004e-01 9.86699998e-01 9.75488961e-01
2.84012109e-01 -4.50533509e-01 -2.40839869e-01 -1.86694413e-01
5.83091915e-01 -1.19196534e+00 -4.35341686e-01 5.87049484e-01
-1.08312893e+00 -8.64766896e-01 -9.04569030e-01 -6.19507194e-01
5.14435291e-01 -8.33249018e-02 6.75991178e-01 -5.75277209e-01
1.23438492e-01 9.51634407e-01 3.33274364e-01 -6.33515298e-01
-5.12128115e-01 -7.13275895e-02 1.73963889e-01 -1.13585927e-01
3.45767647e-01 -9.23659384e-01 -2.94309765e-01 3.60787570e-01
-6.69884264e-01 -5.86530924e-01 5.07253766e-01 5.51919162e-01
5.49872935e-01 4.78197575e-01 5.23993492e-01 -4.81919855e-01
1.15266764e+00 -5.23473203e-01 -1.11277759e+00 3.85989279e-01
-6.05461955e-01 4.14195925e-01 7.39031196e-01 -1.03654587e+00
-1.01116693e+00 -1.87257126e-01 1.12242773e-01 -6.51182294e-01
-3.33772719e-01 3.86655271e-01 1.71378508e-01 -1.64270297e-01
8.36710989e-01 2.09412962e-01 -1.59136638e-01 -2.95994878e-01
3.17252040e-01 3.50107849e-01 8.69607985e-01 -8.58980894e-01
1.02834940e+00 4.49289858e-01 6.30375922e-01 -1.23542953e+00
-8.48303139e-01 -4.03069228e-01 -4.67199296e-01 1.48599163e-01
1.05339694e+00 -6.96739256e-01 -1.19828463e+00 3.08418393e-01
-1.04437268e+00 -4.00604188e-01 -5.32248914e-01 7.30249345e-01
-1.00623369e+00 2.24041834e-01 -8.76729548e-01 -1.42821252e+00
-2.08051696e-01 -4.00906265e-01 6.69376314e-01 3.62736166e-01
-1.38255842e-02 -1.68036973e+00 1.90611660e-01 -6.39917195e-01
4.14928466e-01 1.37006655e-01 1.21121502e+00 -1.23358274e+00
1.83267847e-01 -5.42497754e-01 -1.32609636e-01 6.41388178e-01
-3.56084287e-01 1.94427073e-01 -9.23096955e-01 -1.50559232e-01
5.03699958e-01 3.83726172e-02 8.21647763e-01 8.27319801e-01
8.59443367e-01 -2.53826499e-01 -1.67900205e-01 7.98316121e-01
1.49024355e+00 6.07751422e-02 4.74489838e-01 7.36449212e-02
6.68612897e-01 5.76634347e-01 -2.10301057e-01 5.06767482e-02
-2.59927660e-01 7.19380230e-02 -2.15995952e-01 3.57531577e-01
6.71901762e-01 -4.87235278e-01 5.65985918e-01 5.71715355e-01
-4.60356593e-01 2.43243694e-01 -8.97167981e-01 -7.77696595e-02
-1.69190240e+00 -1.10303009e+00 -1.52703628e-01 2.51313782e+00
6.88709140e-01 2.40229771e-01 1.40285447e-01 -2.47834206e-01
7.39706039e-01 -3.50795299e-01 -6.68311238e-01 -3.73285830e-01
-1.00925028e-01 3.34863126e-01 1.14477003e+00 5.95660985e-01
-1.04266107e+00 4.47614491e-01 7.91563416e+00 8.52581263e-01
-7.79611826e-01 8.96243528e-02 2.80653208e-01 4.17389452e-01
2.06957348e-02 -1.06465049e-01 -1.09968483e+00 1.96256280e-01
1.41164398e+00 -5.27068973e-01 4.41298664e-01 1.24028134e+00
-1.65502846e-01 1.92415968e-01 -1.25960350e+00 9.17277098e-01
-3.61909360e-01 -5.99879384e-01 -3.53879929e-01 4.43351835e-01
5.68674743e-01 1.63844556e-01 5.31465113e-01 6.49773359e-01
7.02185571e-01 -1.22621584e+00 1.88527957e-01 7.35439837e-01
5.83789766e-01 -9.18834448e-01 6.04626894e-01 2.61694938e-01
-9.91361856e-01 -6.73920214e-02 -1.08193064e+00 3.41607839e-01
-2.72224605e-01 5.58613241e-01 -8.72511327e-01 2.16840833e-01
3.20196182e-01 5.09880602e-01 -4.01437670e-01 1.03101945e+00
-1.64217204e-01 8.70233238e-01 -5.86945534e-01 1.74202040e-01
8.61047506e-02 -7.08380044e-01 6.52984262e-01 1.24600732e+00
5.21728992e-01 -9.97917727e-02 -2.32499272e-01 9.88013923e-01
5.24586439e-02 -4.28953953e-02 -5.60211658e-01 -3.74743164e-01
2.72092849e-01 1.02016973e+00 -6.53520584e-01 1.79388449e-01
-4.82418358e-01 6.05802894e-01 2.15040132e-01 1.09082437e+00
-7.14489102e-01 -5.24081349e-01 6.20630085e-01 3.02123073e-02
5.80575645e-01 -4.21060979e-01 -9.52422693e-02 -7.99032927e-01
-1.97866678e-01 -2.28829786e-01 3.17890465e-01 -6.21741414e-01
-1.75667012e+00 3.84162813e-01 2.30647698e-01 -7.31446266e-01
-2.00940087e-01 -1.05311751e+00 -7.35854268e-01 1.11839509e+00
-1.08234024e+00 -7.53868818e-01 2.53450777e-03 6.50974870e-01
-2.23632276e-01 -1.72816187e-01 7.87563026e-01 -5.27841985e-01
-6.00138485e-01 6.27651632e-01 5.33652842e-01 3.26659411e-01
1.15841761e-01 -1.64192581e+00 3.92775476e-01 3.98455709e-01
-4.89829779e-02 1.06222653e+00 9.58060026e-01 -5.22068679e-01
-1.41365969e+00 -1.00151980e+00 -1.95366099e-01 -3.14840704e-01
1.36134887e+00 -1.08468674e-01 -1.38053119e+00 9.04854476e-01
-2.94691294e-01 2.48238057e-01 9.20631349e-01 4.21577573e-01
-6.74684823e-01 6.26130104e-02 -8.06489229e-01 6.56731844e-01
4.34088409e-01 -6.20430171e-01 -6.95574880e-01 1.23509020e-01
8.18375230e-01 -5.26044145e-02 -1.35834146e+00 3.12098086e-01
7.68716455e-01 -4.48483676e-01 7.27528334e-01 -9.91504610e-01
1.47578940e-01 1.10785633e-01 -5.53327441e-01 -1.38451946e+00
-3.44099641e-01 -8.07815075e-01 -3.79603356e-01 8.39123905e-01
4.88121212e-01 -1.20461285e+00 3.38953763e-01 6.21668518e-01
2.00154707e-01 -7.60708511e-01 -1.05009854e+00 -1.29761231e+00
6.88065350e-01 -4.81839389e-01 -3.59402001e-02 5.15084803e-01
1.79024749e-02 4.12015349e-01 4.44698334e-02 5.58124006e-01
8.27380359e-01 -4.61394638e-01 2.08561391e-01 -1.72400510e+00
-7.19833195e-01 -6.92150056e-01 -6.51312709e-01 -9.68011260e-01
4.63221282e-01 -7.46825695e-01 2.71450073e-01 -1.04250193e+00
6.44852519e-02 -2.71961480e-01 -2.86512196e-01 -1.06335647e-01
-1.05281582e-03 -2.89940387e-01 -2.02063531e-01 3.06992948e-01
-1.53559625e-01 5.45710444e-01 7.47126460e-01 4.78351444e-01
-3.88843507e-01 6.23124242e-01 -5.17428994e-01 1.17921996e+00
9.48464155e-01 -5.22254765e-01 -5.98304987e-01 3.32275718e-01
4.53206390e-01 -1.47850156e-01 4.16325778e-01 -1.01177847e+00
2.36437142e-01 3.06803379e-02 4.32562083e-01 -2.12445498e-01
2.72945762e-01 -7.45959580e-01 -1.50979653e-01 1.21193781e-01
-5.25714636e-01 9.67005715e-02 2.07856029e-01 1.16542220e+00
2.86565393e-01 -5.90738356e-01 9.69661772e-01 2.53275514e-01
1.40173882e-01 4.61142212e-01 -4.68425542e-01 3.79956394e-01
7.45495677e-01 -1.99199170e-01 -4.78778541e-01 -6.44044518e-01
-6.83882892e-01 -1.59350798e-01 3.68988365e-01 -2.32304156e-01
6.07928932e-01 -9.63955104e-01 -4.85826164e-01 4.86488827e-02
-1.86346456e-01 -3.18979919e-01 -2.32413672e-02 1.20735204e+00
-1.75041005e-01 1.98072642e-01 1.82341024e-01 -4.75815654e-01
-7.11522222e-01 6.49483800e-01 7.39491403e-01 1.30495816e-01
-7.62320817e-01 5.67699254e-01 4.93998051e-01 -3.49481925e-02
4.57336813e-01 -4.85433847e-01 -5.44815101e-02 -4.39211309e-01
6.43940389e-01 5.88706195e-01 -4.33561325e-01 -1.65594742e-01
1.69742733e-01 6.07996941e-01 5.13585880e-02 -4.17868257e-01
1.23041046e+00 8.98902863e-02 1.05451122e-02 1.05656111e+00
1.37721896e+00 1.22463681e-01 -1.59398437e+00 -2.09330156e-01
1.75158545e-01 1.99361786e-01 -1.31161392e-01 -7.32865334e-02
-6.44829988e-01 9.16568637e-01 5.32833159e-01 4.24260587e-01
7.33065367e-01 1.98006094e-01 1.61744758e-01 7.32368648e-01
-7.26378858e-02 -9.78835523e-01 -1.79828808e-01 6.69540524e-01
7.44782090e-01 -6.29153848e-01 -2.25262985e-01 -3.16026181e-01
-5.44395030e-01 1.51613593e+00 1.09939098e-01 -5.98940253e-01
1.19495237e+00 5.01268446e-01 -3.31837803e-01 2.19003316e-02
-5.93280792e-01 2.26380348e-01 3.76695186e-01 1.00748432e+00
2.98183858e-01 6.15518503e-02 3.84987772e-01 6.61690950e-01
-4.21831727e-01 -2.89912492e-01 4.86694664e-01 4.34413373e-01
-6.63668036e-01 -2.86203742e-01 -2.63045758e-01 3.65945518e-01
-6.84871614e-01 -2.48620018e-01 1.66774198e-01 8.38996530e-01
-7.37596869e-01 5.37557423e-01 2.17990220e-01 8.72515887e-02
3.00011188e-01 5.23497522e-01 7.65456617e-01 -2.26659790e-01
2.36075088e-01 -5.36414534e-02 -3.50494921e-01 -8.17236826e-02
2.43208259e-01 -5.14705658e-01 -1.28011262e+00 -3.88111264e-01
-3.58986765e-01 7.21482858e-02 6.87358022e-01 9.68450904e-01
-2.43426397e-01 1.76643774e-01 3.51703644e-01 -3.11983019e-01
-1.48763609e+00 -1.10061920e+00 -1.22070634e+00 -2.49759611e-02
3.82219970e-01 -5.49705446e-01 -8.22341025e-01 -1.26909181e-01] | [7.377999782562256, 3.8915159702301025] |
bb67ae62-b5f1-446a-b31f-35ac3f5f165d | gcdt-a-chinese-rst-treebank-for-multigenre | 2210.10449 | null | https://arxiv.org/abs/2210.10449v1 | https://arxiv.org/pdf/2210.10449v1.pdf | GCDT: A Chinese RST Treebank for Multigenre and Multilingual Discourse Parsing | A lack of large-scale human-annotated data has hampered the hierarchical discourse parsing of Chinese. In this paper, we present GCDT, the largest hierarchical discourse treebank for Mandarin Chinese in the framework of Rhetorical Structure Theory (RST). GCDT covers over 60K tokens across five genres of freely available text, using the same relation inventory as contemporary RST treebanks for English. We also report on this dataset's parsing experiments, including state-of-the-art (SOTA) scores for Chinese RST parsing and RST parsing on the English GUM dataset, using cross-lingual training in Chinese and English with multilingual embeddings. | ['Amir Zeldes', 'Yang Janet Liu', 'Siyao Peng'] | 2022-10-19 | null | null | null | null | ['discourse-parsing'] | ['natural-language-processing'] | [-1.66487545e-01 9.42902267e-01 -5.23542702e-01 -2.73113221e-01
-1.32278419e+00 -9.51954961e-01 6.24342382e-01 4.15575832e-01
-5.59344709e-01 1.00691307e+00 1.21320736e+00 -9.15884852e-01
1.80436254e-01 -4.88579303e-01 -6.35085166e-01 -3.00001144e-01
-4.19510812e-01 5.74212909e-01 4.22658354e-01 -4.58251119e-01
2.57388026e-01 -3.96420479e-01 -6.02800906e-01 6.94622815e-01
9.20815468e-01 5.26776195e-01 2.90829569e-01 8.44529569e-01
-3.27246726e-01 1.46520519e+00 -9.00448382e-01 -7.76231647e-01
-4.25392777e-01 -3.57023239e-01 -1.60237086e+00 -1.85679078e-01
3.52345675e-01 -2.55915791e-01 -2.16600642e-01 5.06404281e-01
3.66711050e-01 -4.52351153e-01 1.76762879e-01 -4.06726539e-01
-1.04702461e+00 1.60733306e+00 -1.63736075e-01 5.25385320e-01
4.65847284e-01 -5.60966611e-01 1.73654044e+00 -6.75845623e-01
1.47699165e+00 1.61921680e+00 4.60719645e-01 7.68426061e-01
-9.73290205e-01 -2.64644712e-01 1.56922087e-01 1.53818786e-01
-6.46609485e-01 -3.89653802e-01 8.02135468e-01 -2.93655634e-01
1.68909264e+00 -3.17169838e-02 3.89786333e-01 1.17038083e+00
4.84071404e-01 9.89192843e-01 1.23488045e+00 -9.17105913e-01
-6.49364367e-02 -4.42393661e-01 5.37841737e-01 4.74152178e-01
9.84933749e-02 -3.73902559e-01 -6.78415120e-01 -3.42828035e-02
3.35048497e-01 -1.26715767e+00 9.03538167e-02 5.56645334e-01
-1.19334698e+00 1.13084722e+00 -4.52929139e-02 9.06924188e-01
1.13308206e-01 5.24814948e-02 9.69165623e-01 2.06202701e-01
7.59327769e-01 5.27494490e-01 -7.39114523e-01 -4.53371555e-01
-9.13617611e-02 1.98062852e-01 1.09065437e+00 1.00265479e+00
1.90949544e-01 4.23311070e-03 1.08094916e-01 7.91178048e-01
1.04940750e-01 8.47249106e-02 5.03265917e-01 -1.42673290e+00
1.17985189e+00 3.73972267e-01 -3.34484369e-01 -5.86790740e-01
-3.84546220e-01 2.25769415e-01 -2.10021347e-01 -5.55419564e-01
5.58821380e-01 -6.50881708e-01 -1.18698798e-01 1.64690602e+00
3.72964352e-01 -6.07321322e-01 7.39232719e-01 3.26319098e-01
1.28226280e+00 7.60143757e-01 3.69362772e-01 -3.39996636e-01
1.78587484e+00 -9.51776564e-01 -1.12978959e+00 -3.85039687e-01
1.15596831e+00 -9.46823597e-01 8.95660818e-01 1.08476236e-01
-1.14526093e+00 -2.14367792e-01 -1.07507169e+00 -8.57520938e-01
-2.06201091e-01 1.12802029e-01 6.72753632e-01 6.60013258e-01
-8.54219139e-01 2.64225215e-01 -8.89160693e-01 -1.14756957e-01
2.65939951e-01 -3.48688632e-01 -2.30634362e-01 1.04605012e-01
-1.53031957e+00 1.10211325e+00 8.56601536e-01 -3.12410947e-02
-4.62835163e-01 -5.43407977e-01 -1.09496939e+00 -3.00723255e-01
5.62576473e-01 4.27965254e-01 1.66837502e+00 -2.02010065e-01
-1.53295124e+00 1.24012530e+00 -2.10135520e-01 -7.54467189e-01
5.26516959e-02 -9.83830869e-01 -1.25207275e-01 2.21185803e-01
4.87457603e-01 3.40791583e-01 2.31606327e-02 -9.94386256e-01
-6.49579704e-01 -2.41095260e-01 3.03185970e-01 2.35723615e-01
-2.17263713e-01 8.90586734e-01 -1.31575972e-01 -6.06246293e-01
-5.39596900e-02 -6.28459573e-01 -4.55982611e-03 -8.88244510e-01
-5.33182144e-01 -1.17558122e+00 8.54538083e-01 -1.40343881e+00
1.66082871e+00 -1.86150992e+00 1.28333569e-01 -9.09743726e-01
5.03305905e-02 -4.08134572e-02 1.46432325e-01 7.06817150e-01
1.95409223e-01 5.78988850e-01 -1.32046789e-01 -3.57857019e-01
7.47903138e-02 8.39909852e-01 -2.15225279e-01 -5.27852215e-03
7.19095349e-01 1.24195194e+00 -9.84762132e-01 -9.68625307e-01
-2.99144667e-02 9.06895921e-02 -3.65866959e-01 3.08039814e-01
-5.21526814e-01 4.16778892e-01 -6.00506961e-01 5.40634990e-01
1.31268993e-01 7.70479590e-02 1.25495052e+00 2.91838676e-01
-6.40752852e-01 1.43920994e+00 -3.65506947e-01 1.60360181e+00
-2.43829519e-01 9.12784815e-01 1.12497620e-01 -8.19444776e-01
4.87608403e-01 6.36257946e-01 8.27767029e-02 -7.28486598e-01
3.13351244e-01 2.87772447e-01 4.63185847e-01 -5.91977477e-01
6.10306025e-01 1.04862183e-01 -9.34507966e-01 3.94127846e-01
1.29353777e-01 -1.63703635e-01 5.10660589e-01 2.51763523e-01
1.03575623e+00 3.60544890e-01 7.88118958e-01 -8.46039474e-01
3.83217096e-01 5.08517146e-01 7.21363187e-01 1.72030985e-01
2.14030128e-02 1.40621275e-01 1.09111762e+00 -4.63144392e-01
-1.05083239e+00 -7.63727665e-01 -6.56959474e-01 1.52209651e+00
-5.30008137e-01 -9.34229314e-01 -1.13862157e+00 -8.47744048e-01
-4.51237410e-01 9.72224474e-01 -6.87088132e-01 1.14955258e+00
-1.91872239e+00 -9.05611873e-01 1.02701104e+00 5.54123402e-01
6.32095873e-01 -1.38390720e+00 -7.72339046e-01 7.02693284e-01
-4.49968338e-01 -1.86181962e+00 -4.98712473e-02 3.10693502e-01
-5.69882870e-01 -1.60529578e+00 -9.16224867e-02 -1.27623284e+00
-1.42035395e-01 -3.54791164e-01 1.42273080e+00 -6.29231185e-02
1.52963534e-01 -9.62189287e-02 -7.85119057e-01 -5.13912022e-01
-8.55512738e-01 2.87242651e-01 -5.07041633e-01 -1.03124523e+00
3.40267181e-01 5.77374548e-02 3.05152327e-01 -3.30504984e-01
-4.42310363e-01 -3.43365185e-02 -1.15857087e-02 9.53301072e-01
2.62141168e-01 -1.85273401e-03 4.18662697e-01 -1.42960858e+00
7.42610693e-01 -2.65445650e-01 -6.52253807e-01 4.23727483e-01
1.34369329e-01 2.92461757e-02 5.81206799e-01 -1.15132211e-02
-1.66779983e+00 -7.69935250e-01 -5.10340571e-01 8.05755615e-01
9.54378769e-02 6.96952879e-01 -2.31295303e-01 7.33170211e-01
5.17727315e-01 -3.85435611e-01 -7.32859850e-01 -6.41832709e-01
6.32707894e-01 5.85285187e-01 6.00257874e-01 -1.21168888e+00
1.69874728e-01 -1.87540874e-01 -3.52154642e-01 -1.31220293e+00
-1.27014804e+00 4.72771525e-02 -1.05638003e+00 1.55814454e-01
1.70412123e+00 -9.61696267e-01 -5.84122658e-01 2.59691209e-01
-1.77123845e+00 -6.88366652e-01 -4.40360047e-02 8.15421641e-02
-2.14079440e-01 4.66086596e-01 -1.47006500e+00 -6.67654335e-01
-3.26818049e-01 -8.28059793e-01 1.10812283e+00 -8.31428841e-02
-5.17597258e-01 -1.32098353e+00 2.61532992e-01 4.48927432e-01
-1.47918567e-01 5.98412335e-01 1.44841540e+00 -4.89457071e-01
-4.20691371e-01 6.16168678e-01 -2.11841404e-01 3.65758896e-01
1.08483337e-01 -1.37782395e-01 -8.03066075e-01 1.01795532e-01
5.89472018e-02 -8.28629792e-01 4.87371027e-01 3.70482326e-01
3.24562013e-01 -5.29846966e-01 3.73857841e-02 5.84209077e-02
1.31972480e+00 2.49187410e-01 4.59168941e-01 8.22441697e-01
6.52904570e-01 1.08806360e+00 6.37895703e-01 9.03153718e-02
9.35222685e-01 1.82522178e-01 -1.30441740e-01 4.56495225e-01
-4.02816653e-01 -2.91670531e-01 4.41797256e-01 1.71650279e+00
-1.52355015e-01 -1.69856191e-01 -1.42483485e+00 7.32904971e-01
-1.69753182e+00 -4.89114434e-01 -4.43181932e-01 1.21560192e+00
1.50113797e+00 4.90021080e-01 -1.84379578e-01 -8.78843069e-02
4.28540081e-01 6.05907083e-01 2.70004332e-01 -7.35964656e-01
-4.99039590e-01 5.40159464e-01 2.04827011e-01 9.36279655e-01
-1.47157228e+00 1.62076080e+00 6.94565678e+00 6.50926590e-01
-4.21024173e-01 6.85710132e-01 6.93928361e-01 6.74269259e-01
-1.42082080e-01 3.04161221e-01 -1.23442113e+00 7.27987364e-02
1.43171346e+00 -2.46708333e-01 8.01936910e-03 8.09246480e-01
-2.10570335e-01 -1.89031377e-01 -8.10270190e-01 3.81835587e-02
-1.24483973e-01 -1.79591620e+00 -4.65164572e-01 -5.81622273e-02
7.40881920e-01 3.86289746e-01 -4.97869372e-01 3.46402645e-01
9.37913418e-01 -8.60275269e-01 7.47212172e-01 -4.30719376e-01
9.21601713e-01 -4.39437181e-01 9.41913247e-01 3.23144406e-01
-1.28581977e+00 1.04677849e-01 -2.75013596e-01 -2.68415123e-01
5.31303048e-01 1.16465621e-01 -7.71813035e-01 6.89882040e-01
8.06434095e-01 6.30617082e-01 -5.62479377e-01 -4.21755202e-02
-9.04618561e-01 1.28198349e+00 -6.87185526e-02 -3.89265507e-01
4.50691283e-01 5.66868670e-02 5.57201684e-01 1.61256003e+00
-2.68335640e-01 5.52826524e-01 4.95205641e-01 8.55875090e-02
-2.97863305e-01 3.38434190e-01 -3.44256490e-01 -2.96576530e-01
6.71801984e-01 6.95528626e-01 -7.77655959e-01 -4.79749441e-01
-5.17667294e-01 3.26562673e-01 8.90053868e-01 -2.75240671e-02
-4.32069749e-01 -1.75645083e-01 1.90742493e-01 -2.04064488e-01
3.69944602e-01 -8.10851276e-01 -2.93918699e-01 -9.47281420e-01
1.02535067e-02 -1.03224957e+00 5.97413480e-01 -2.76164144e-01
-1.45297194e+00 1.00654519e+00 4.49618429e-01 -5.01025617e-01
-3.54994267e-01 -1.06720543e+00 -5.21270394e-01 5.87285876e-01
-1.65795135e+00 -1.36690962e+00 4.45376873e-01 -1.23735726e-01
9.86652315e-01 7.58253336e-02 1.18141055e+00 -3.59043665e-02
-5.77627659e-01 2.34736174e-01 8.60830694e-02 9.24247384e-01
6.27542853e-01 -1.82033443e+00 9.53079224e-01 6.71411157e-01
-8.70991573e-02 5.20230770e-01 5.56639373e-01 -9.14511323e-01
-1.07651615e+00 -6.46322072e-01 1.79807174e+00 -9.97161508e-01
1.42830217e+00 -6.52340412e-01 -1.11546445e+00 9.83602345e-01
1.17130113e+00 -4.32384372e-01 7.47726679e-01 7.21880674e-01
-1.98056534e-01 5.37345648e-01 -5.61042488e-01 4.15287375e-01
1.05711925e+00 -6.01996958e-01 -1.52457583e+00 3.61885071e-01
1.32075322e+00 -9.17761564e-01 -1.47035718e+00 5.31069934e-02
3.47787976e-01 -5.16757011e-01 4.57236439e-01 -8.02356184e-01
8.97776604e-01 2.86571860e-01 -5.06179690e-01 -7.00358927e-01
-1.05030291e-01 -9.58427727e-01 -1.90722849e-02 1.45984697e+00
4.64074880e-01 -3.74823093e-01 6.74926639e-01 2.55143911e-01
-5.72688878e-01 -6.60374820e-01 -1.53291416e+00 -6.58668578e-01
1.08782887e+00 -5.38439512e-01 2.85396665e-01 9.56134915e-01
4.94146734e-01 1.07253623e+00 6.46020025e-02 -1.98912486e-01
5.45788467e-01 -4.67702523e-02 4.39238757e-01 -1.13527942e+00
-3.30080330e-01 -1.67959392e-01 2.80641854e-01 -9.33468938e-01
6.51106000e-01 -7.44465828e-01 1.68833792e-01 -1.67260110e+00
-1.20527200e-01 -3.40058684e-01 3.13638866e-01 5.18308938e-01
-1.20825790e-01 -3.96790445e-01 3.06378514e-01 3.80071178e-02
-6.13241613e-01 5.22025764e-01 1.05943263e+00 1.04845360e-01
-2.00695947e-01 -8.40996385e-01 -5.96869349e-01 9.09224212e-01
6.80883646e-01 -6.61393225e-01 1.07530460e-01 -9.32096779e-01
-1.38347968e-01 3.60395074e-01 -3.39527667e-01 -3.46766174e-01
-4.14547101e-02 -1.26548177e-02 3.73279713e-02 -8.07785034e-01
3.28759886e-02 -1.14163250e-01 -7.33305812e-01 2.93140769e-01
-4.39375579e-01 4.67570573e-01 4.95117635e-01 9.25240666e-03
-4.13743019e-01 -5.20116568e-01 1.76472247e-01 -3.35798413e-01
-8.28346014e-01 -5.26385188e-01 -7.10462391e-01 9.73368466e-01
6.16195023e-01 2.42045537e-01 -1.06919146e+00 2.37812951e-01
-4.60839927e-01 3.92675936e-01 -2.39745639e-02 4.07445997e-01
1.55542359e-01 -1.07888746e+00 -1.03025711e+00 -4.70892727e-01
-1.18560873e-01 2.03237012e-01 -4.75589521e-02 3.40707958e-01
-6.64476931e-01 8.74894202e-01 1.08207434e-01 -1.77015439e-01
-1.35165501e+00 1.38465032e-01 -2.88113117e-01 -1.00252640e+00
-8.00243378e-01 7.84555137e-01 -5.08839898e-02 -5.27791202e-01
-5.87227084e-02 -4.53665257e-01 -6.91164434e-01 1.97501853e-01
4.70873922e-01 1.60313860e-01 -1.92677811e-01 -9.32895362e-01
-3.86319131e-01 2.23005682e-01 -1.31581083e-01 -3.92803967e-01
1.43269289e+00 -1.95066445e-02 -6.40108645e-01 6.60014093e-01
1.17172599e+00 4.42516714e-01 -9.08541918e-01 -1.19124025e-01
9.96819854e-01 -3.12020420e-03 -2.49030128e-01 -4.68476593e-01
1.20747797e-01 7.77632296e-01 -2.76773751e-01 5.24134517e-01
5.69842875e-01 4.67186481e-01 9.34969068e-01 6.35624826e-01
3.64701450e-01 -1.52769494e+00 6.59846812e-02 1.31036270e+00
9.46439207e-01 -1.05508006e+00 -9.99275967e-03 -7.96889782e-01
-8.73777807e-01 1.25403941e+00 5.90357959e-01 -2.62637198e-01
7.77040124e-01 5.34845650e-01 3.04989785e-01 -3.65582943e-01
-9.47092593e-01 -1.75207295e-02 3.30070034e-02 5.94869435e-01
1.30218101e+00 2.46325627e-01 -8.90576422e-01 8.27657580e-01
-8.51207197e-01 -5.76588452e-01 5.67173719e-01 1.25496459e+00
-5.53604782e-01 -1.49440467e+00 -1.57524809e-01 -1.17678933e-01
-1.09406495e+00 -4.25779879e-01 -5.41002035e-01 1.53297937e+00
7.12523982e-02 1.23722470e+00 8.41709152e-02 1.86694458e-01
1.71272010e-01 9.77926329e-02 6.35913372e-01 -9.84888852e-01
-5.55313289e-01 5.54306924e-01 1.44203699e+00 -2.35453039e-01
-8.55764329e-01 -9.47225869e-01 -1.59580350e+00 -2.07358316e-01
-9.28685963e-02 4.42764670e-01 1.00088388e-01 1.16471696e+00
-2.97499359e-01 6.04882896e-01 3.78660858e-02 -5.93322396e-01
-1.16440006e-01 -1.22861099e+00 -7.26180151e-02 -9.63450521e-02
-6.80775382e-03 -2.50504613e-01 9.65714008e-02 1.60254747e-01] | [10.829485893249512, 9.45113468170166] |
b3251b3a-8ff4-4114-9c12-02cd0e0b2833 | unsupervised-speech-recognition | 2105.11084 | null | https://arxiv.org/abs/2105.11084v3 | https://arxiv.org/pdf/2105.11084v3.pdf | Unsupervised Speech Recognition | Despite rapid progress in the recent past, current speech recognition systems still require labeled training data which limits this technology to a small fraction of the languages spoken around the globe. This paper describes wav2vec-U, short for wav2vec Unsupervised, a method to train speech recognition models without any labeled data. We leverage self-supervised speech representations to segment unlabeled audio and learn a mapping from these representations to phonemes via adversarial training. The right representations are key to the success of our method. Compared to the best previous unsupervised work, wav2vec-U reduces the phoneme error rate on the TIMIT benchmark from 26.1 to 11.3. On the larger English Librispeech benchmark, wav2vec-U achieves a word error rate of 5.9 on test-other, rivaling some of the best published systems trained on 960 hours of labeled data from only two years ago. We also experiment on nine other languages, including low-resource languages such as Kyrgyz, Swahili and Tatar. | ['Michael Auli', 'Alexis Conneau', 'Wei-Ning Hsu', 'Alexei Baevski'] | 2021-05-24 | null | http://proceedings.neurips.cc/paper/2021/hash/ea159dc9788ffac311592613b7f71fbb-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/ea159dc9788ffac311592613b7f71fbb-Paper.pdf | neurips-2021-12 | ['unsupervised-speech-recognition'] | ['speech'] | [ 1.50071606e-01 3.27874541e-01 -1.41739145e-01 -4.13847417e-01
-1.09013712e+00 -7.57685483e-01 6.93045616e-01 -2.83424407e-01
-6.42486274e-01 8.55967760e-01 5.78209043e-01 -8.17150474e-01
6.10143483e-01 -4.91789579e-01 -5.10616541e-01 -4.76894647e-01
7.20349476e-02 6.33135438e-01 -3.30644436e-02 -3.31323922e-01
-1.69524118e-01 2.94764668e-01 -1.22846258e+00 3.63641903e-02
6.30853176e-01 6.93106294e-01 -2.30082180e-02 9.86123562e-01
-1.94305331e-01 7.74961233e-01 -7.60801435e-01 -2.03459665e-01
4.66033705e-02 -3.41298431e-01 -1.14123118e+00 1.68032974e-01
2.88909912e-01 -2.63904035e-01 -7.52506495e-01 7.38882542e-01
7.41642714e-01 4.32469130e-01 7.52747178e-01 -8.08776140e-01
-9.44876075e-01 9.50530946e-01 1.13815203e-01 2.38371417e-01
1.02095686e-01 -5.85645996e-02 1.08280885e+00 -1.19906008e+00
6.41954422e-01 1.07094944e+00 4.50126052e-01 8.45943630e-01
-1.05572021e+00 -7.41589546e-01 -4.77211364e-02 2.06710130e-01
-1.64949954e+00 -9.61427569e-01 5.18108666e-01 -3.60269099e-01
1.44671464e+00 3.29402477e-01 4.83549163e-02 1.42442739e+00
-5.32598019e-01 7.47959673e-01 8.52663398e-01 -8.05503607e-01
2.33181164e-01 1.91862732e-01 1.64954931e-01 4.15010929e-01
-3.07445556e-01 2.29003787e-01 -3.92191678e-01 -5.06671369e-02
4.31930840e-01 -3.32600862e-01 -3.76149267e-01 1.13289401e-01
-1.13645470e+00 1.11005759e+00 8.96157250e-02 3.60394359e-01
4.03256379e-02 -8.21148530e-02 5.85927844e-01 5.89234412e-01
6.92022562e-01 2.74089128e-01 -6.99639678e-01 -4.21150565e-01
-8.59753013e-01 -3.52310300e-01 7.57540107e-01 9.91015613e-01
5.16763151e-01 1.03145730e+00 6.73022494e-02 1.35868657e+00
3.57752681e-01 7.69913197e-01 9.04572189e-01 -4.02711391e-01
6.66267633e-01 1.27903610e-01 -1.23548180e-01 -2.00316921e-01
2.52901297e-02 -2.54989237e-01 -8.11205924e-01 -1.27475690e-02
2.24286035e-01 -4.96385664e-01 -1.47810686e+00 1.42818296e+00
-1.00944206e-01 1.90810263e-01 5.18380046e-01 6.45506263e-01
1.11923432e+00 1.22171021e+00 -5.15383594e-02 -7.73819610e-02
9.07922626e-01 -1.18291652e+00 -7.69217372e-01 -4.07348484e-01
6.78114355e-01 -8.86047959e-01 1.14383197e+00 3.35730612e-01
-8.27098906e-01 -5.29783428e-01 -9.56920266e-01 1.14329815e-01
-8.06299269e-01 -2.07226910e-02 3.09069633e-01 1.19950771e+00
-1.15627086e+00 5.69411777e-02 -6.21048033e-01 -4.73573834e-01
2.54691482e-01 3.38834882e-01 -6.20121956e-01 -1.07195228e-01
-1.21312165e+00 8.15126657e-01 4.02998388e-01 -1.93166807e-01
-1.19798326e+00 -4.00779396e-01 -1.04736483e+00 -1.66571382e-02
6.44753203e-02 2.13481918e-01 1.48174667e+00 -8.79712224e-01
-1.78406811e+00 7.73822844e-01 -1.95204914e-01 -6.40141666e-01
1.88681990e-01 -1.97527841e-01 -1.04911256e+00 -1.92221999e-01
-3.87855142e-01 5.81045985e-01 6.79912031e-01 -1.10440421e+00
-3.06086153e-01 6.71838596e-02 -5.01826286e-01 2.07678527e-02
-6.14817798e-01 1.05074778e-01 -5.78199804e-01 -9.78595614e-01
-1.05261400e-01 -9.26125348e-01 -1.03406243e-01 -7.50904679e-01
-4.86926824e-01 -3.58733177e-01 8.66385043e-01 -8.46081972e-01
1.02776170e+00 -2.29256487e+00 4.32441570e-02 2.58672148e-01
-1.89050689e-01 8.31879258e-01 -3.89370948e-01 6.81916893e-01
-1.70945451e-01 3.09410155e-01 -3.66664082e-01 -5.58047712e-01
-1.78887993e-02 6.03119195e-01 -6.89308465e-01 4.76005733e-01
2.26347357e-01 8.96577477e-01 -8.94098878e-01 6.28032163e-02
2.98888981e-01 6.98683798e-01 -4.26969826e-01 4.28434223e-01
5.59055321e-02 2.24755093e-01 1.07177168e-01 5.11159122e-01
4.63506579e-01 3.31791162e-01 -7.07970886e-03 5.18365324e-01
5.06068021e-02 6.40076756e-01 -7.15856850e-01 1.63932693e+00
-6.46480739e-01 9.61151838e-01 -4.55608778e-02 -1.22693968e+00
1.21692610e+00 7.36629665e-01 6.65174872e-02 -4.47931826e-01
1.55928552e-01 2.50360608e-01 -1.66546404e-01 -1.68478251e-01
6.13634706e-01 -3.52075659e-02 -1.77071795e-01 2.78574258e-01
5.68194449e-01 -2.59323984e-01 -2.98946798e-01 2.61207968e-01
9.96159732e-01 -4.06554669e-01 1.52551487e-01 -6.71348572e-02
4.49324965e-01 -2.08337545e-01 3.06379676e-01 8.26705992e-01
-2.56330878e-01 1.00371647e+00 -1.53280005e-01 -2.41165429e-01
-1.09685957e+00 -1.27078569e+00 -8.43235552e-02 1.26131320e+00
-4.39744532e-01 -4.29907799e-01 -6.71444416e-01 -7.82390356e-01
-2.01153010e-01 1.00382638e+00 -5.20506740e-01 -1.79080054e-01
-3.65042031e-01 -2.39289507e-01 1.13116217e+00 6.04000807e-01
1.76746890e-01 -1.38858986e+00 2.67212778e-01 2.51472771e-01
-1.00220107e-02 -1.27899253e+00 -6.78598404e-01 4.66730297e-01
-5.01866877e-01 -5.13711631e-01 -9.76387203e-01 -1.17030609e+00
4.54702556e-01 5.83179966e-02 1.21505225e+00 -1.94541544e-01
-4.32418697e-02 4.35914129e-01 -8.03491652e-01 -4.51507211e-01
-7.07661450e-01 3.07065308e-01 4.43939418e-01 6.17305934e-03
4.88160729e-01 -1.90502763e-01 1.70093533e-02 1.25484139e-01
-7.05908358e-01 -4.55576599e-01 1.16025664e-01 1.08346403e+00
5.74242175e-01 -2.82238930e-01 8.66752446e-01 -9.71942425e-01
3.40889186e-01 -6.13711357e-01 -3.00543070e-01 1.63742810e-01
-3.87395769e-01 -1.89129740e-01 6.63249910e-01 -5.95680475e-01
-8.32326889e-01 5.34532741e-02 -6.62055552e-01 -4.56814647e-01
-4.26750511e-01 4.65893209e-01 -9.52734202e-02 2.77027935e-01
6.19645774e-01 5.22192836e-01 -1.64402813e-01 -5.08306980e-01
6.90698802e-01 1.32769060e+00 7.30090618e-01 -2.33197678e-02
7.73208261e-01 8.08158144e-02 -9.98949766e-01 -1.52977490e+00
-4.85551238e-01 -5.61216891e-01 -4.33313549e-01 7.82557055e-02
8.49360287e-01 -1.03641033e+00 -1.66698724e-01 3.62280816e-01
-1.06602776e+00 -5.27198553e-01 -3.99414122e-01 5.90961158e-01
-2.94558853e-01 2.29320467e-01 -5.71310878e-01 -9.42203343e-01
-5.51504791e-01 -1.02814400e+00 6.47055626e-01 -2.70604312e-01
-3.06446612e-01 -1.14071143e+00 3.04815203e-01 4.04442370e-01
5.53936422e-01 -4.95009065e-01 5.17961204e-01 -1.22409272e+00
8.81439820e-02 -3.55113894e-01 1.98467851e-01 9.57934201e-01
4.04362887e-01 -2.15674758e-01 -1.47305584e+00 -4.31544751e-01
-4.13506687e-01 -6.37860537e-01 9.18018162e-01 1.38323203e-01
1.11602771e+00 -3.90935361e-01 4.36722599e-02 5.18004775e-01
9.31058645e-01 4.45945978e-01 6.72096848e-01 4.67234440e-02
6.66767299e-01 2.50660539e-01 1.49581060e-01 2.13153929e-01
1.19802617e-01 5.67604125e-01 1.69974163e-01 -1.32691547e-01
-3.14102948e-01 -4.79927868e-01 7.19011307e-01 1.85901451e+00
1.91552430e-01 -5.69394588e-01 -1.18463385e+00 9.13478196e-01
-1.28492486e+00 -8.31421256e-01 3.54245007e-01 2.23318100e+00
8.20547104e-01 4.39107232e-02 -6.31802157e-02 2.59837151e-01
5.09584785e-01 2.77666658e-01 -2.09544554e-01 -7.08872318e-01
-2.93798238e-01 6.75215185e-01 4.86940026e-01 9.68888164e-01
-1.30365288e+00 1.47926486e+00 7.19875479e+00 9.38088834e-01
-1.21379054e+00 3.45938385e-01 5.48378527e-01 9.49687511e-02
-4.67185050e-01 -3.32862049e-01 -6.89067602e-01 1.92702472e-01
1.52579987e+00 3.67725268e-02 6.08757973e-01 9.16451275e-01
-9.98592451e-02 6.66258395e-01 -7.14305460e-01 1.18285346e+00
3.81360799e-01 -1.27146864e+00 2.19582580e-02 -6.60855249e-02
9.11900580e-01 8.10913503e-01 1.84487119e-01 7.58543193e-01
8.03953350e-01 -1.47400582e+00 4.94708240e-01 -2.15625446e-02
1.22002554e+00 -1.03750658e+00 6.99229658e-01 1.79704741e-01
-9.53364789e-01 3.72067094e-01 -5.97672403e-01 8.16722214e-02
2.11291045e-01 2.09278584e-01 -1.47193646e+00 2.53847331e-01
5.25193810e-01 4.41117227e-01 -3.24622720e-01 5.43285549e-01
-2.78889835e-01 1.58129680e+00 -2.76039928e-01 -1.03578739e-01
5.55737317e-01 5.51577052e-03 3.47860605e-01 1.54309058e+00
2.66769081e-01 1.31697413e-02 1.34131342e-01 1.47331610e-01
-4.70618606e-01 4.10668522e-01 -1.04876530e+00 -4.52654034e-01
7.82875180e-01 6.74499512e-01 -3.24739397e-01 -4.80618715e-01
-6.23856664e-01 1.18185043e+00 5.11569262e-01 6.35318995e-01
-4.40460771e-01 -5.95193207e-01 7.91337967e-01 -3.75570744e-01
6.14101350e-01 -5.24919331e-01 -7.00909793e-02 -1.18411493e+00
-4.06080365e-01 -1.01038575e+00 5.75821027e-02 -5.45746684e-01
-1.22396982e+00 1.08930886e+00 -5.53407371e-01 -1.19698429e+00
-6.01210296e-01 -6.32504463e-01 -5.58434308e-01 9.38582122e-01
-1.27398825e+00 -9.49784935e-01 1.90340340e-01 5.83098471e-01
1.02720332e+00 -1.06759250e+00 1.53232622e+00 2.28068709e-01
-5.41859448e-01 1.08396900e+00 5.91644883e-01 6.48132920e-01
6.71070635e-01 -1.11109459e+00 1.11943901e+00 8.59964609e-01
9.47067201e-01 3.17290395e-01 3.78156781e-01 -3.85886818e-01
-1.17992568e+00 -1.19667935e+00 1.26030219e+00 -5.27442515e-01
7.61198401e-01 -7.17678607e-01 -9.30736184e-01 8.88411283e-01
5.42657256e-01 1.10287301e-01 1.01679432e+00 1.91672787e-01
-8.78177822e-01 1.53039902e-01 -8.33904326e-01 5.49094558e-01
7.63826549e-01 -1.09211373e+00 -7.41729498e-01 2.71631867e-01
1.10868657e+00 -3.01689599e-02 -7.54133880e-01 7.06115440e-02
2.39617601e-01 -3.86582613e-01 9.33101416e-01 -1.09830487e+00
1.75066479e-02 1.26095656e-02 -6.11686051e-01 -1.78729391e+00
-9.64612588e-02 -7.22376049e-01 1.48763172e-02 1.54537499e+00
7.31510580e-01 -6.44475937e-01 6.65466607e-01 -5.55232279e-02
-4.50482309e-01 -2.48873562e-01 -1.23653042e+00 -1.09766328e+00
5.65147340e-01 -9.10092115e-01 4.61047143e-01 1.25768673e+00
-1.32490233e-01 4.10661697e-01 -5.12628019e-01 1.34034306e-01
2.75621623e-01 -5.35335898e-01 7.44738996e-01 -8.61229897e-01
-2.24402040e-01 -2.65679926e-01 -5.51547527e-01 -1.06182361e+00
6.65586829e-01 -1.25368536e+00 3.40517938e-01 -1.26518190e+00
-3.94366384e-01 -3.17155629e-01 -4.28708255e-01 6.71995103e-01
2.93407068e-02 4.89472955e-01 -1.78371929e-02 -4.86528873e-02
-1.18558772e-01 7.71819413e-01 5.96414089e-01 -6.80574358e-01
-2.60469377e-01 -2.33182281e-01 -4.09731686e-01 4.42318261e-01
1.07461858e+00 -2.89405853e-01 -5.54243028e-01 -7.49647141e-01
-6.09181583e-01 -2.36219823e-01 -3.52219492e-01 -8.79058838e-01
6.58676699e-02 5.74858189e-02 4.35320660e-03 -3.86468530e-01
5.72085321e-01 -5.28695583e-01 -4.08892453e-01 2.14909777e-01
-4.03283596e-01 -2.14113027e-01 3.50895375e-01 3.40134948e-01
-5.04825056e-01 -2.00085059e-01 6.45941436e-01 1.42826274e-01
-1.08841479e+00 2.39773601e-01 -9.55639958e-01 3.94954741e-01
6.61424458e-01 1.00796536e-01 -2.81630307e-01 -8.10292184e-01
-7.57607043e-01 -6.32603765e-02 1.18516358e-02 7.67436922e-01
8.31567943e-01 -1.37779713e+00 -1.03807569e+00 6.72199070e-01
4.32568997e-01 -3.30213934e-01 -4.39834632e-02 1.61541328e-01
-4.93313670e-01 5.27079463e-01 1.79821700e-01 -3.79336774e-01
-1.16472936e+00 3.77416074e-01 5.74316606e-02 2.59418517e-01
-5.12296617e-01 1.23357499e+00 -1.78005621e-01 -1.03708553e+00
5.31802595e-01 -3.62998508e-02 -2.42549509e-01 -3.55311520e-02
5.98827779e-01 1.35604218e-01 3.29072177e-01 -1.05571413e+00
-4.73732978e-01 7.40114134e-03 -2.43922189e-01 -6.03775799e-01
1.20058239e+00 3.51378061e-02 3.58907431e-01 7.21577704e-01
1.45744145e+00 3.73029560e-01 -7.14064479e-01 -2.67502040e-01
-1.10524662e-01 6.10450283e-04 2.12300628e-01 -5.59519708e-01
-1.06997359e+00 1.05245161e+00 6.03770435e-01 1.62481233e-01
6.59852862e-01 2.03797877e-01 8.31862211e-01 5.95982194e-01
1.95358351e-01 -9.81715858e-01 -1.95955530e-01 1.15511584e+00
8.19154084e-01 -1.24111843e+00 -5.17684221e-01 -1.25850111e-01
-9.27034080e-01 7.10073471e-01 3.14729691e-01 -5.22301868e-02
8.99318278e-01 2.02642679e-01 7.76859701e-01 4.10479873e-01
-5.88061154e-01 -3.46688807e-01 3.92628342e-01 9.68568861e-01
6.58395946e-01 4.51638997e-01 2.17743084e-01 6.27506137e-01
-5.55796027e-01 -5.11473119e-01 5.50686657e-01 7.79701114e-01
-4.41583842e-01 -1.24440515e+00 -3.20869505e-01 3.38929296e-01
-4.25322741e-01 -3.88625741e-01 -5.29436946e-01 4.95814234e-01
-3.47278357e-01 1.37321734e+00 3.48429590e-01 -7.36563087e-01
3.31985056e-01 5.11286557e-01 -1.11946031e-01 -8.73057008e-01
-5.89452326e-01 1.53381199e-01 3.49583685e-01 -1.65736020e-01
-1.08502820e-01 -3.28205734e-01 -1.14565957e+00 -1.30277783e-01
-3.01044017e-01 4.09392774e-01 6.88433051e-01 7.66196430e-01
1.06725290e-01 4.16783601e-01 7.73046374e-01 -7.95204639e-01
-4.91766185e-01 -1.33861983e+00 -6.69861436e-01 4.51892972e-01
4.63760436e-01 -3.93098354e-01 -5.27018726e-01 1.59246922e-01] | [14.377392768859863, 6.698877811431885] |
f33360de-6d62-4aee-a64e-5269f2e80dce | drl-gan-a-hybrid-approach-for-binary-and | 2301.03368 | null | https://arxiv.org/abs/2301.03368v1 | https://arxiv.org/pdf/2301.03368v1.pdf | DRL-GAN: A Hybrid Approach for Binary and Multiclass Network Intrusion Detection | Our increasingly connected world continues to face an ever-growing amount of network-based attacks. Intrusion detection systems (IDS) are an essential security technology for detecting these attacks. Although numerous machine learning-based IDS have been proposed for the detection of malicious network traffic, the majority have difficulty properly detecting and classifying the more uncommon attack types. In this paper, we implement a novel hybrid technique using synthetic data produced by a Generative Adversarial Network (GAN) to use as input for training a Deep Reinforcement Learning (DRL) model. Our GAN model is trained with the NSL-KDD dataset for four attack categories as well as normal network flow. Ultimately, our findings demonstrate that training the DRL on specific synthetic datasets can result in better performance in correctly classifying minority classes over training on the true imbalanced dataset. | ['Anwar Haque', 'Daniel Lizotte', 'Noshin Tasnim', 'Sareh Nejad', 'Muhammad Zakar', 'Chandrika Saha', 'Caroline Strickland'] | 2023-01-05 | null | null | null | null | ['network-intrusion-detection'] | ['miscellaneous'] | [ 2.11484864e-01 -2.46190891e-01 -3.88209522e-01 -3.98642480e-01
-2.74419099e-01 -6.14626229e-01 5.56751192e-01 -4.33788747e-02
-1.38215095e-01 7.32510924e-01 -4.45469648e-01 -9.37874794e-01
3.41439843e-01 -1.24682713e+00 -1.78849712e-01 -2.53736258e-01
-5.22766002e-02 8.15361977e-01 1.38442859e-01 -2.81981081e-01
2.21240491e-01 9.51800466e-01 -8.66052926e-01 4.09638345e-01
6.43916309e-01 9.52761412e-01 -1.17484391e+00 9.84150410e-01
-1.86198309e-01 1.32500887e+00 -1.42395055e+00 -6.00264847e-01
5.84477782e-01 -8.19136083e-01 -6.04464829e-01 -3.45409244e-01
3.14066380e-01 -6.50511801e-01 -6.64232850e-01 8.57295275e-01
4.58451152e-01 -2.33154759e-01 5.67166030e-01 -1.98144233e+00
-2.51719207e-01 4.01985765e-01 -6.37588203e-01 7.19781578e-01
2.88253427e-01 5.08587003e-01 5.01925886e-01 -2.13826954e-01
3.24520767e-01 1.32277715e+00 6.94574654e-01 8.73882115e-01
-1.15138471e+00 -1.26540577e+00 -3.97808760e-01 1.16443299e-01
-8.37707877e-01 -2.13566661e-01 1.01380301e+00 -1.70421526e-01
8.98358703e-01 1.85966417e-01 3.77959371e-01 1.65144515e+00
2.96141654e-01 4.78714317e-01 1.00431335e+00 -1.83104545e-01
3.23434263e-01 2.11085111e-01 1.93644799e-02 2.56614536e-01
5.64234078e-01 5.02091587e-01 6.75536618e-02 -3.66394490e-01
7.31766045e-01 2.35162780e-01 2.18876123e-01 2.15303704e-01
-5.01442313e-01 1.30397010e+00 5.61071694e-01 2.02451542e-01
-4.49359059e-01 1.25122994e-01 8.37861240e-01 8.54054809e-01
4.60882604e-01 7.87865400e-01 -3.68732005e-01 -3.22810888e-01
-8.73293817e-01 -1.37744863e-02 1.10732436e+00 3.36999923e-01
4.85707909e-01 8.94671381e-01 2.26765007e-01 5.50847054e-01
1.79746971e-01 5.84956229e-01 4.22578573e-01 -5.81914485e-01
3.61343563e-01 6.97596431e-01 -3.05500388e-01 -1.10638034e+00
-3.76949877e-01 -5.98012209e-01 -9.60295498e-01 4.91497397e-01
4.82170373e-01 -5.50248682e-01 -9.96220469e-01 1.53260148e+00
1.45460248e-01 4.13127542e-01 2.94972509e-01 2.73175597e-01
3.46405506e-01 6.42733335e-01 3.16101640e-01 2.04159126e-01
7.18644619e-01 -5.99135280e-01 -3.77931833e-01 -2.40097597e-01
5.65804243e-01 -5.83077669e-01 5.69987655e-01 2.79273152e-01
-7.31125593e-01 -4.16117281e-01 -1.07031381e+00 7.62328744e-01
-7.56754994e-01 -5.47259808e-01 5.68603039e-01 1.51818955e+00
-8.02580655e-01 3.51324528e-01 -6.72939181e-01 -2.84213006e-01
9.03505385e-01 3.01952988e-01 -8.11410472e-02 -2.45623905e-02
-1.41584802e+00 7.92416096e-01 3.15335155e-01 -4.27524477e-01
-1.22884393e+00 -6.75333977e-01 -5.73837399e-01 -6.10387512e-02
3.53205279e-02 -1.45873204e-01 1.05183542e+00 -1.27771103e+00
-1.26419806e+00 7.54312158e-01 4.55110639e-01 -9.45082307e-01
5.56243777e-01 2.13462919e-01 -9.47242200e-01 2.72002459e-01
-1.50736108e-01 1.14416182e-01 1.06573594e+00 -1.11257505e+00
-1.91084206e-01 -2.85977423e-01 4.69319746e-02 -3.95661414e-01
-2.95053840e-01 3.26344013e-01 5.94116867e-01 -8.10295403e-01
-3.30568880e-01 -6.41954124e-01 -1.26593873e-01 -3.79841864e-01
-6.02604270e-01 1.15153663e-01 1.75686276e+00 -4.80540425e-01
1.09751678e+00 -1.84168112e+00 -5.63434124e-01 6.47694826e-01
3.16935390e-01 1.08412564e+00 -1.83501527e-01 4.87579107e-01
-4.05436546e-01 5.32803059e-01 -7.22879171e-02 -8.04726779e-02
-2.03413963e-01 2.36728504e-01 -7.29177594e-01 2.11672649e-01
5.78811407e-01 8.75606716e-01 -7.37942219e-01 1.03445657e-01
3.81901681e-01 4.48572427e-01 -5.62389851e-01 6.69187665e-01
-2.53267586e-02 6.61084831e-01 -4.44376051e-01 9.22983587e-01
6.98551536e-01 -1.81741893e-01 4.85176183e-02 1.33072644e-01
7.38602817e-01 2.19969615e-01 -5.99356234e-01 5.28090596e-01
-4.80022907e-01 6.67626143e-01 -2.77302802e-01 -1.21091533e+00
1.26606250e+00 2.12414160e-01 3.26092929e-01 -9.85851467e-01
3.72237474e-01 1.95339128e-01 5.56238770e-01 -2.17181206e-01
-2.53896803e-01 -2.77774543e-01 -9.99270976e-02 9.34213877e-01
7.63485283e-02 6.01154491e-02 -4.10540402e-02 3.17783326e-01
1.83337271e+00 -8.21421444e-01 1.13267355e-01 3.39272112e-01
6.02850318e-01 9.15290788e-03 4.78483051e-01 8.79984975e-01
-6.60523415e-01 -6.93313824e-03 7.90619850e-01 -1.03181732e+00
-1.22888482e+00 -1.29749608e+00 7.57707134e-02 6.15121067e-01
-2.73409873e-01 -2.74362247e-02 -9.05964911e-01 -1.30319166e+00
-9.45392549e-02 8.87106180e-01 -5.69523811e-01 -6.78586662e-01
-6.32998407e-01 -1.14020658e+00 1.34602284e+00 4.88466948e-01
9.58632886e-01 -1.48193800e+00 -3.48192573e-01 2.55266666e-01
3.83543938e-01 -1.30930459e+00 1.32510573e-01 1.17509484e-01
-5.53243160e-01 -1.45497501e+00 -5.51786390e-04 -4.29674715e-01
6.32593036e-01 -2.45378196e-01 1.33594573e+00 3.12500477e-01
-6.30782425e-01 2.29120865e-01 -4.59808260e-01 -7.31902540e-01
-1.16017783e+00 1.65760085e-01 1.44350678e-01 1.07683532e-01
8.47977340e-01 -6.21730089e-01 -4.57269311e-01 3.03302020e-01
-9.79109108e-01 -5.13417125e-01 4.12062913e-01 5.50508916e-01
-2.13815033e-01 3.97980630e-01 1.28641176e+00 -1.45310771e+00
8.72598588e-01 -8.78725946e-01 -3.02445203e-01 -1.85129300e-01
-5.83418131e-01 -4.07424450e-01 1.24643290e+00 -5.33639193e-01
-6.28180683e-01 -7.27465868e-01 -5.35091698e-01 -5.02502739e-01
-4.48618829e-01 -4.69565913e-02 -8.19771737e-02 -2.69784093e-01
9.27485764e-01 5.05536310e-02 2.32547089e-01 1.28614664e-01
-2.94109672e-01 8.10609043e-01 1.58513680e-01 -3.59280229e-01
1.19851613e+00 2.83470631e-01 7.09630921e-02 -5.89084625e-01
-8.14194739e-01 1.18187137e-01 -2.30133504e-01 -2.48426780e-01
5.00693321e-01 -4.45186347e-01 -6.49498522e-01 9.63697374e-01
-8.12406600e-01 -5.47969639e-01 -2.80550271e-01 -6.26785681e-02
-1.46280587e-01 -6.38593733e-02 -9.64696169e-01 -7.60089755e-01
-6.19423449e-01 -1.04951727e+00 3.61287087e-01 3.38614434e-01
-3.13414633e-01 -1.28800082e+00 4.15068746e-01 2.66023129e-01
1.10011780e+00 8.57967377e-01 9.74504709e-01 -1.49831557e+00
-8.71382877e-02 -8.39875758e-01 -3.56515825e-01 9.08514082e-01
3.63496304e-01 3.29006970e-01 -9.65681493e-01 -4.52006698e-01
1.09872773e-01 -5.64690113e-01 3.99263620e-01 4.40352373e-02
1.33735299e+00 -4.87802774e-01 -6.76978976e-02 6.23758912e-01
1.44449997e+00 8.84602308e-01 8.92553270e-01 2.80648023e-01
7.63651788e-01 2.37223595e-01 1.03097849e-01 1.89687654e-01
-1.87409058e-01 1.05318256e-01 6.42345726e-01 -3.08373898e-01
1.09359063e-01 -2.58619994e-01 2.62988865e-01 2.18679771e-01
4.20187891e-01 -6.29399300e-01 -1.39534009e+00 -7.99678862e-02
-1.10560071e+00 -1.22779453e+00 2.49745250e-01 1.93236077e+00
5.55492043e-01 7.29091167e-01 4.52463001e-01 6.57789290e-01
6.78795338e-01 2.58188397e-01 -9.14889932e-01 -1.01169622e+00
-1.46107853e-03 9.73784864e-01 4.48965013e-01 2.06171647e-01
-1.13254845e+00 9.13557351e-01 6.86754799e+00 6.12318277e-01
-1.50158727e+00 -1.19447663e-01 1.23557711e+00 3.14505696e-01
1.33678451e-01 -3.38877052e-01 -4.80420381e-01 6.66127145e-01
1.70015621e+00 -1.43993683e-02 4.28342164e-01 9.28820908e-01
7.27355555e-02 3.74084890e-01 -6.87929749e-01 7.23758996e-01
8.71828496e-02 -1.21758199e+00 4.01714802e-01 1.09099634e-02
6.27568901e-01 -3.20942476e-02 2.30867147e-01 6.25106394e-01
7.13565350e-01 -1.53204989e+00 -2.63689339e-01 -2.11429540e-02
8.90728116e-01 -1.27780175e+00 9.48968112e-01 2.06902549e-01
-5.15633285e-01 -2.64017701e-01 -8.03960338e-02 -6.88554570e-02
-2.76600987e-01 5.22800744e-01 -1.19572639e+00 3.24496776e-02
3.14720005e-01 6.80732787e-01 -6.22660816e-01 6.22596622e-01
-3.69871482e-02 1.27192020e+00 -3.51708271e-02 7.60755390e-02
2.58193880e-01 2.15500090e-02 4.19438571e-01 9.72270668e-01
-1.40590057e-01 -2.47329950e-01 1.44053504e-01 6.34684920e-01
-5.00837982e-01 -2.90274471e-01 -8.91651034e-01 -4.94302809e-01
5.78897536e-01 1.18594468e+00 -8.62774909e-01 -3.59057456e-01
-1.48057535e-01 7.40616143e-01 4.18367907e-02 2.66328245e-01
-1.03173554e+00 -5.76377988e-01 9.38439906e-01 3.33263427e-01
-2.31043905e-01 3.57924812e-02 -3.84413958e-01 -1.01936400e+00
-4.98453140e-01 -1.42550862e+00 5.33899486e-01 -3.55020791e-01
-1.74852693e+00 1.06552076e+00 -4.82234895e-01 -1.18105507e+00
-4.22028810e-01 -7.04135418e-01 -1.16069126e+00 7.56799638e-01
-1.20453143e+00 -9.57696319e-01 -3.48083556e-01 6.61174715e-01
4.44396913e-01 -7.51218021e-01 9.47166800e-01 4.89456862e-01
-9.96903598e-01 9.12251234e-01 -2.52123535e-01 1.04231548e+00
5.49172580e-01 -1.10371196e+00 8.15224051e-01 9.32919919e-01
-8.17662105e-02 1.98709518e-01 4.48256940e-01 -6.92725658e-01
-7.78094709e-01 -1.42992449e+00 7.27384165e-02 -5.93116879e-01
6.79848015e-01 -3.06542605e-01 -9.05133128e-01 9.01868522e-01
4.17488962e-02 2.41354145e-02 9.98961449e-01 -4.15225059e-01
-8.58171999e-01 -1.86446294e-01 -2.02016354e+00 1.66621476e-01
4.65041846e-01 -6.44702077e-01 -2.33342618e-01 1.15567990e-01
5.32282829e-01 -1.89637840e-01 -7.47291565e-01 5.69232404e-01
5.83849400e-02 -1.10565686e+00 1.02289772e+00 -1.22756684e+00
5.50666928e-01 -4.12572362e-02 6.72450438e-02 -1.39565325e+00
1.66097879e-01 -4.09074038e-01 -4.06024188e-01 1.12730706e+00
2.03552082e-01 -1.08749199e+00 1.22815597e+00 2.59663016e-01
4.71162200e-01 -7.93865740e-01 -8.04642081e-01 -5.20277917e-01
1.79174796e-01 -4.23368305e-01 8.11802864e-01 1.38075399e+00
-5.05844653e-01 1.75686926e-01 -3.69596004e-01 2.35525742e-02
8.03702354e-01 -4.58541125e-01 9.17924106e-01 -1.26187789e+00
-1.10266477e-01 -4.34885353e-01 -9.98558998e-01 -6.97160438e-02
2.75919527e-01 -6.70873344e-01 -8.60011756e-01 -7.69896507e-01
-2.07556278e-01 -6.72070622e-01 -3.98043275e-01 5.45783043e-01
1.23731524e-01 7.28264391e-01 -4.61148098e-02 4.64464612e-02
-2.88535595e-01 1.95278466e-01 8.22386861e-01 -1.82993248e-01
1.05567999e-01 4.36251760e-01 -6.94236040e-01 5.78738809e-01
1.31689858e+00 -5.52944601e-01 -4.22780931e-01 2.23753631e-01
8.29670802e-02 -9.82080102e-02 3.03553104e-01 -1.31311274e+00
-1.22441500e-01 -7.76513517e-02 8.42403412e-01 -2.68071502e-01
-4.41316627e-02 -7.81293809e-01 -8.51779580e-02 9.20377612e-01
-2.64960557e-01 4.48280990e-01 3.86238515e-01 3.94983977e-01
-1.98815957e-01 1.60610095e-01 1.19804180e+00 8.31109807e-02
-3.22325557e-01 5.53195596e-01 -4.90352571e-01 5.16666710e-01
1.33600557e+00 -1.44228533e-01 -7.16961741e-01 -6.68746591e-01
-4.49797034e-01 5.58410911e-03 4.13988054e-01 5.42095602e-01
7.13364184e-01 -1.12568676e+00 -6.99054360e-01 8.26554656e-01
-3.56963426e-01 -3.91992331e-01 -8.15604553e-02 1.40749097e-01
-1.04088962e+00 7.93101117e-02 -8.30366731e-01 -2.18597576e-01
-8.91461611e-01 6.08575165e-01 8.52404177e-01 -6.67398572e-01
-2.22243756e-01 5.56208432e-01 -4.95000035e-01 -6.49863720e-01
2.85134222e-02 4.53072578e-01 4.82197590e-02 -2.85770088e-01
7.51998603e-01 5.40458679e-01 4.43621799e-02 -4.65958208e-01
-3.30794692e-01 -2.25405291e-01 -3.64263862e-01 3.46590102e-01
1.16001916e+00 6.20261431e-01 -2.19061542e-02 4.15999582e-03
1.37353981e+00 -2.67134070e-01 -6.05031013e-01 6.87965304e-02
-4.47944850e-02 -5.17359078e-01 -2.06888527e-01 -1.01783478e+00
-1.57174456e+00 9.16231036e-01 6.29282296e-01 9.01387095e-01
1.19933760e+00 -5.84904492e-01 1.00690460e+00 1.02360725e-01
3.35898250e-01 -3.90576243e-01 4.85765874e-01 4.55662042e-01
1.57439068e-01 -1.36085999e+00 -4.53965962e-01 -1.79784596e-01
-4.65255260e-01 1.21819198e+00 1.10219133e+00 -6.26334071e-01
8.25569153e-01 5.71612537e-01 4.24205363e-01 -1.99138924e-01
-6.01811409e-01 3.54018867e-01 -4.30066496e-01 8.68886650e-01
1.91287041e-01 -6.20036293e-03 2.93469578e-01 -1.01683557e-01
-1.75523713e-01 -4.62746322e-01 7.89364338e-01 8.83290827e-01
-8.56693909e-02 -1.40355694e+00 -3.47608685e-01 1.07647693e+00
-7.63637543e-01 3.66423540e-02 -7.37197340e-01 8.23689699e-01
-1.72668323e-01 1.04940903e+00 4.06192601e-01 -8.45489502e-01
1.77303851e-01 1.40996799e-01 -5.71925640e-02 -5.24503112e-01
-9.55576539e-01 -9.10952151e-01 -1.99612066e-01 -6.62229419e-01
-7.08151460e-02 -1.74291506e-01 -9.30176735e-01 -9.11367118e-01
1.60888746e-01 1.52730793e-01 5.30765831e-01 7.13495851e-01
3.35501283e-01 8.37272823e-01 1.23548198e+00 -3.69008780e-01
-6.12334013e-01 -8.60303283e-01 -4.42892671e-01 7.71638751e-01
3.18415374e-01 -4.97788906e-01 -8.41926754e-01 -6.02104604e-01] | [5.441495418548584, 7.4887471199035645] |
a3cca228-01d0-446c-9e52-91b93724f87e | fire-now-fire-later-alarm-based-systems-for | 1905.09568 | null | https://arxiv.org/abs/1905.09568v2 | https://arxiv.org/pdf/1905.09568v2.pdf | Fire Now, Fire Later: Alarm-Based Systems for Prescriptive Process Monitoring | Predictive process monitoring is a family of techniques to analyze events produced during the execution of a business process in order to predict the future state or the final outcome of running process instances. Existing techniques in this field are able to predict, at each step of a process instance, the likelihood that it will lead to an undesired outcome.These techniques, however, focus on generating predictions and do not prescribe when and how process workers should intervene to decrease the cost of undesired outcomes. This paper proposes a framework for prescriptive process monitoring, which extends predictive monitoring with the ability to generate alarms that trigger interventions to prevent an undesired outcome or mitigate its effect. The framework incorporates a parameterized cost model to assess the cost-benefit trade-off of generating alarms. We show how to optimize the generation of alarms given an event log of past process executions and a set of cost model parameters. The proposed approaches are empirically evaluated using a range of real-life event logs. The experimental results show that the net cost of undesired outcomes can be minimized by changing the threshold for generating alarms, as the process instance progresses. Moreover, introducing delays for triggering alarms, instead of triggering them as soon as the probability of an undesired outcome exceeds a threshold, leads to lower net costs. | ['Massimiliano de Leoni', 'Stephan A. Fahrenkrog-Petersen', 'Niek Tax', 'Irene Teinemaa', 'Fabrizio Maria Maggi', 'Matthias Weidlich', 'Marlon Dumas'] | 2019-05-23 | null | null | null | null | ['predictive-process-monitoring'] | ['time-series'] | [ 8.51999938e-01 3.02154541e-01 2.67300755e-01 -3.79600465e-01
-2.24185467e-01 -3.04627270e-01 8.43745589e-01 1.08552992e+00
-2.96531916e-01 4.73173797e-01 -1.88480183e-01 -4.15038854e-01
-4.46267009e-01 -1.20062554e+00 -2.02467471e-01 -3.99352700e-01
-3.58584821e-01 5.69974124e-01 5.82662761e-01 4.18925464e-01
4.24868047e-01 6.66879594e-01 -1.43189168e+00 2.29061738e-01
3.91313702e-01 9.60718274e-01 2.05544569e-02 7.87780881e-01
-2.35553965e-01 9.77901161e-01 -1.04402602e+00 1.99490059e-02
2.62064397e-01 -6.05698645e-01 -4.69697118e-01 2.19130158e-01
-5.43883324e-01 -1.82588875e-01 4.35390681e-01 7.83727884e-01
-5.31121381e-02 6.62377104e-02 6.37764871e-01 -1.46715069e+00
3.03744495e-01 4.65937108e-01 -1.94677487e-01 4.55726504e-01
2.83223838e-01 5.76533675e-01 7.07124233e-01 -1.24951765e-01
1.26182824e-01 1.11839044e+00 1.29521683e-01 8.39203969e-02
-1.28141224e+00 -3.55710179e-01 3.86494815e-01 -5.53802103e-02
-8.95946681e-01 -1.48313344e-01 6.03792906e-01 -4.08465028e-01
1.08005726e+00 3.17519546e-01 4.63257790e-01 6.99798882e-01
7.26101816e-01 9.36027318e-02 8.94237816e-01 -5.41714549e-01
6.87187493e-01 -1.19383242e-02 -7.62518635e-03 1.12797134e-01
5.28326988e-01 1.79166541e-01 -3.36115062e-01 -5.85182965e-01
6.00978434e-01 3.91478717e-01 4.80534583e-02 1.27981290e-01
-9.39440191e-01 3.57080728e-01 -2.72046089e-01 2.28047684e-01
-9.12353039e-01 9.47643518e-02 3.58910263e-01 2.51276553e-01
1.08858965e-01 5.26815355e-01 -3.74509096e-01 -2.36173466e-01
-4.80974853e-01 1.92897990e-01 1.04903018e+00 6.27289534e-01
5.45626998e-01 -2.71597091e-04 -6.22121990e-01 1.25261992e-01
3.05287272e-01 -3.29170888e-03 2.64124304e-01 -7.08272994e-01
3.77677917e-01 9.79561269e-01 6.95070088e-01 -6.55619502e-01
-1.64412290e-01 -6.95305839e-02 -3.26491177e-01 5.13264835e-01
3.29041362e-01 -3.03420335e-01 -6.98639274e-01 1.28913963e+00
1.96280137e-01 3.30937058e-01 -1.97228834e-01 3.80736709e-01
-4.25749600e-01 8.60252857e-01 4.82230633e-01 -8.86366189e-01
1.11468804e+00 -1.81864515e-01 -7.96456277e-01 -1.21855490e-01
2.76239216e-01 -7.60044336e-01 7.28657544e-01 5.29626429e-01
-9.57260132e-01 -4.63260859e-01 -7.56455719e-01 1.18610764e+00
9.48613584e-02 -3.55489343e-01 3.09734225e-01 5.68997562e-01
-4.84620959e-01 8.34320664e-01 -1.31640518e+00 -2.61936694e-01
-1.76786020e-01 3.28208566e-01 2.23622903e-01 3.56283098e-01
-7.49899983e-01 7.29092479e-01 6.82968855e-01 1.61239564e-01
-1.34384298e+00 -5.57809830e-01 -3.16736549e-01 5.48421025e-01
7.35710382e-01 -3.04684162e-01 1.56031942e+00 -7.38621831e-01
-1.24985182e+00 -8.90007988e-02 -2.00151056e-01 -6.35694683e-01
7.67144024e-01 -2.41588220e-01 -7.06904352e-01 -9.56113487e-02
-1.78142801e-01 -2.15293199e-01 7.33869553e-01 -9.61491346e-01
-1.11389601e+00 -2.94850141e-01 8.17094836e-03 -1.50043927e-02
-1.26727089e-01 1.02730617e-01 8.21221098e-02 -1.21983759e-01
-1.34807736e-01 -8.11367989e-01 -7.62031257e-01 -6.45202041e-01
-3.78436863e-01 -1.59671798e-01 7.13531733e-01 -6.02771901e-02
1.24389541e+00 -1.77814412e+00 -3.79498631e-01 4.74613309e-01
-2.81711906e-01 9.12585557e-02 2.67242789e-01 8.14399362e-01
3.99573594e-02 2.63027012e-01 -2.39882350e-01 8.81110653e-02
-3.15091491e-01 2.55802244e-01 -3.28358203e-01 9.90890786e-02
5.68315208e-01 -1.02443844e-01 -7.53579319e-01 -7.33157471e-02
5.92065334e-01 7.71905631e-02 -1.13801949e-01 8.30650270e-01
-4.48173255e-01 5.12696564e-01 -5.79710841e-01 3.23241174e-01
1.36313140e-01 -7.20560849e-02 5.02285361e-01 4.23235536e-01
-4.07684267e-01 4.24789667e-01 -1.47277617e+00 2.77873188e-01
-6.28917873e-01 -3.58847678e-02 -4.43013608e-02 -8.72124970e-01
1.11598933e+00 7.18674600e-01 6.31830573e-01 -3.70011628e-01
-4.14695591e-02 -3.39211412e-02 2.47174036e-02 -4.36084092e-01
1.25742882e-01 -1.73056185e-01 6.94093388e-03 7.27541268e-01
-5.60325980e-01 2.47736394e-01 4.23653156e-01 -2.62171656e-01
1.64341092e+00 1.86551772e-02 6.34428203e-01 1.09303057e-01
7.38377810e-01 -1.11168787e-01 9.10311282e-01 7.23983645e-01
-1.09328024e-01 -1.57782376e-01 8.12357485e-01 -6.76979542e-01
-8.01663756e-01 -1.01203549e+00 3.93689990e-01 8.51760268e-01
6.59378693e-02 -2.66505748e-01 -3.08652997e-01 -5.80213130e-01
-1.12774827e-01 1.14872670e+00 -3.62273216e-01 -1.80216774e-01
-6.58599675e-01 -8.76616895e-01 -1.41210660e-01 2.67164260e-01
2.08656177e-01 -1.38513839e+00 -1.36082959e+00 7.60712326e-01
2.37264410e-01 -8.26132238e-01 2.26797210e-03 2.89843768e-01
-1.11521900e+00 -1.39100122e+00 3.19348216e-01 7.97859114e-03
7.97080457e-01 -2.43722647e-01 9.14393663e-01 6.59030676e-02
-9.40230787e-02 2.87079811e-01 -1.96872532e-01 -7.58836389e-01
-9.56429064e-01 -4.44825888e-01 -2.72587866e-01 2.77127981e-01
3.37024659e-01 -4.47187841e-01 -4.77157712e-01 3.38115066e-01
-1.14652336e+00 -3.27216715e-01 5.11453688e-01 3.27333868e-01
6.35882080e-01 5.29267788e-01 8.91979992e-01 -1.02655149e+00
9.82911766e-01 -4.57312942e-01 -8.50383103e-01 3.98733884e-01
-1.05295002e+00 2.63100654e-01 9.75224197e-01 -4.63465780e-01
-1.37737775e+00 1.12506226e-01 3.39683861e-01 -6.77552149e-02
-6.03399396e-01 3.22057664e-01 -1.54608935e-01 8.11716735e-01
1.91166148e-01 2.34164014e-01 -3.19066346e-01 -2.52462417e-01
-2.74156868e-01 2.72167861e-01 2.41683662e-01 -4.74136382e-01
4.04610187e-01 4.36508417e-01 3.28287601e-01 -3.17513108e-01
-4.45643663e-01 -5.44964969e-01 -2.13237315e-01 -6.01379037e-01
4.37594414e-01 -2.52723068e-01 -1.06209648e+00 -1.00376299e-02
-9.84838545e-01 4.02887277e-02 -2.40573943e-01 2.31053978e-01
-5.69153905e-01 -4.95869247e-03 -3.51684034e-01 -1.36241221e+00
-4.40453947e-01 -8.92930627e-01 4.69115883e-01 3.03460926e-01
-6.95962846e-01 -7.75373578e-01 2.90957615e-02 -1.22231364e-01
3.94478649e-01 5.09673476e-01 9.14063811e-01 -1.14002204e+00
-7.45310247e-01 -5.77318311e-01 4.12836939e-01 1.58686772e-01
5.59671581e-01 3.63792986e-01 -3.99920881e-01 -1.68151140e-01
1.84773937e-01 6.35506094e-01 2.34326050e-02 7.02763274e-02
9.48998868e-01 -6.88585758e-01 -4.04732794e-01 -1.13566510e-01
1.46471548e+00 9.67837453e-01 5.60736656e-01 5.08850813e-01
1.62701979e-01 7.47893989e-01 1.02011395e+00 9.52724457e-01
-1.58186972e-01 4.83924717e-01 7.73113310e-01 4.01462436e-01
6.03831530e-01 -1.42615184e-01 4.35394526e-01 7.06310868e-02
-3.74558032e-01 -2.93150365e-01 -9.42736149e-01 6.71271324e-01
-1.76645505e+00 -1.18477535e+00 -3.66616160e-01 2.78551650e+00
4.65828270e-01 7.65776336e-01 1.71270087e-01 4.65264052e-01
8.15160036e-01 -8.59456062e-02 -3.95821780e-01 -8.28348219e-01
5.00987828e-01 7.65387416e-02 5.10429204e-01 4.51265395e-01
-7.59587646e-01 3.11503500e-01 5.63146448e+00 -2.37564240e-02
-9.81061995e-01 -1.28003672e-01 6.75841272e-01 -1.57518938e-01
-1.39300004e-01 3.42331290e-01 -9.39715207e-01 4.82601970e-01
1.53624392e+00 -5.87369859e-01 1.11093767e-01 8.01000893e-01
9.79685128e-01 -2.45359078e-01 -1.51606691e+00 1.01470038e-01
-5.91543257e-01 -1.00983036e+00 -4.85307304e-03 3.94169986e-02
4.33165371e-01 -5.03685653e-01 -5.37310898e-01 -2.03782436e-03
5.16744435e-01 -7.69300520e-01 6.57982767e-01 8.11546266e-01
-2.03094408e-01 -1.18724966e+00 7.85207570e-01 6.79799139e-01
-1.08060598e+00 -4.26402748e-01 2.20734224e-01 -5.03468812e-01
4.45524663e-01 8.42819631e-01 -1.64816713e+00 3.67964387e-01
4.66754228e-01 -9.28072184e-02 -2.54775047e-01 9.36311901e-01
-5.09689391e-01 7.98711538e-01 -2.94179350e-01 -1.33046865e-01
-1.05939984e-01 -1.82645082e-01 6.82166636e-01 1.11524940e+00
3.70534122e-01 -1.89934343e-01 3.08752924e-01 8.70177329e-01
5.70990205e-01 1.29504753e-02 -4.84426945e-01 1.28891990e-01
6.97790861e-01 1.02229118e+00 -9.01409626e-01 -4.49592978e-01
-1.85307756e-01 6.05254591e-01 -9.35817212e-02 2.15753362e-01
-6.05231106e-01 -1.94875434e-01 3.71502280e-01 6.38565779e-01
1.57519996e-01 -1.30002022e-01 -6.53609216e-01 -2.51822114e-01
1.35501280e-01 -5.49283266e-01 5.32945514e-01 -2.99487531e-01
-1.10852933e+00 3.19252759e-01 4.63464931e-02 -9.64865148e-01
-6.17742896e-01 -6.16006590e-02 -1.23007059e+00 9.66174066e-01
-1.11850619e+00 -3.90203565e-01 -1.33769035e-01 2.04687133e-01
5.45608401e-01 2.94656307e-01 7.51008391e-01 -2.67827719e-01
-5.94953954e-01 -3.03780288e-01 -3.16492528e-01 -3.03306073e-01
4.18282688e-01 -1.16561484e+00 3.96692678e-02 9.85955119e-01
-3.63256037e-01 5.51204085e-01 1.18946886e+00 -1.10182548e+00
-9.91431594e-01 -1.26285636e+00 9.96739626e-01 -1.72854111e-01
6.29119515e-01 6.20355569e-02 -1.06987596e+00 8.29067051e-01
-1.84808989e-04 -2.56935686e-01 4.68503535e-01 -4.61758375e-02
2.98410803e-01 -3.37282240e-01 -1.41522789e+00 5.21571040e-01
4.61576968e-01 -2.13255035e-03 -6.69923902e-01 8.56802706e-03
2.92694867e-01 2.60763973e-01 -1.00357163e+00 5.29630303e-01
2.27814198e-01 -1.12374878e+00 4.72009867e-01 -4.86758262e-01
2.26951003e-01 -4.95354325e-01 3.55777830e-01 -1.23911011e+00
-9.92060676e-02 -6.60525918e-01 -2.29717284e-01 1.30109048e+00
3.69581461e-01 -5.30093253e-01 5.08328915e-01 1.01885545e+00
1.47164151e-01 -6.10962570e-01 -8.50089252e-01 -6.83607280e-01
-8.37007761e-01 -5.56061625e-01 7.76297808e-01 5.67730606e-01
-5.81003539e-03 4.87456135e-02 -6.51705787e-02 6.83327436e-01
5.99483967e-01 9.64564830e-02 7.11999774e-01 -1.12045228e+00
-3.77903968e-01 -2.13558465e-01 -3.53957415e-01 -1.20957211e-01
-5.12175679e-01 -2.56821215e-02 2.12137893e-01 -1.43070710e+00
1.38701510e-03 -3.13913494e-01 -6.29837096e-01 4.57826138e-01
-4.49262708e-01 -7.00848877e-01 3.01845223e-01 3.39260906e-01
-3.76406491e-01 1.35994285e-01 7.26305425e-01 2.42688462e-01
-6.58753991e-01 5.68064570e-01 -3.35896797e-02 8.15637112e-01
1.00929797e+00 -6.65253103e-01 -4.53994155e-01 3.44893634e-01
1.60299867e-01 5.87329865e-01 3.73232275e-01 -9.70240116e-01
2.44359463e-01 -8.19175661e-01 1.82759389e-01 -5.14851034e-01
-1.53831523e-02 -1.18654585e+00 8.70242357e-01 9.63840485e-01
-5.77890337e-01 4.36721683e-01 -4.31259684e-02 1.03222680e+00
-2.28246719e-01 -3.93820643e-01 5.95196724e-01 -9.68127772e-02
-4.43188339e-01 1.54496059e-02 -8.79633725e-01 -4.69778419e-01
1.63728940e+00 -1.20112464e-01 4.15175445e-02 -2.06817836e-01
-8.84968102e-01 2.13428155e-01 9.46282595e-02 3.46589983e-01
2.69374728e-01 -7.04240143e-01 -4.08189327e-01 7.46477470e-02
-1.62910428e-02 -1.87007740e-01 -8.73117372e-02 5.67931771e-01
-2.36230582e-01 3.68551046e-01 -1.15325958e-01 -4.31665689e-01
-1.24374652e+00 5.77051640e-01 3.71161431e-01 -9.04505193e-01
-5.68161428e-01 2.66294748e-01 -1.42259568e-01 2.77237236e-01
6.98801056e-02 -4.42949504e-01 -2.31757730e-01 -1.89711243e-01
6.66790962e-01 5.82662165e-01 2.76249826e-01 4.51375395e-02
-2.08983988e-01 -2.97959208e-01 -1.28411204e-02 -2.50363469e-01
1.28527379e+00 -5.76445907e-02 -1.57240495e-01 7.33679712e-01
1.56178847e-01 -4.78317263e-03 -1.62615323e+00 1.75924487e-02
7.26583183e-01 -5.52860677e-01 -1.85626715e-01 -9.17094529e-01
-6.33283854e-01 3.73690844e-01 3.93011779e-01 8.34808707e-01
1.38345110e+00 -1.01767816e-01 3.30867022e-01 2.66410736e-03
3.96828681e-01 -1.11915016e+00 -5.15864082e-02 3.03326130e-01
6.32083297e-01 -6.93611562e-01 -1.26304522e-01 -5.25792897e-01
-6.91502154e-01 1.08564937e+00 5.88735342e-01 -4.60662181e-03
4.09368157e-01 4.78364646e-01 -1.97883755e-01 -1.61954537e-01
-1.15709519e+00 2.92368650e-01 -4.20292556e-01 4.21405494e-01
3.97853434e-01 4.90720361e-01 -5.28257251e-01 3.69933307e-01
1.65222675e-01 1.93569541e-01 7.96158016e-01 1.44950283e+00
-6.67598426e-01 -1.12404048e+00 -8.78057361e-01 6.46484733e-01
-6.97468102e-01 3.59280705e-01 -2.85045683e-01 5.47574461e-01
1.87178358e-01 1.15638041e+00 3.96360338e-01 5.67404963e-02
8.14292014e-01 3.70955020e-01 -2.04127245e-02 -9.09900367e-01
-6.96741998e-01 1.08723946e-01 4.24446255e-01 -3.72803211e-01
-1.90477505e-01 -7.43540585e-01 -1.46965361e+00 -1.30632266e-01
-4.04532291e-02 4.78271320e-02 5.16359866e-01 1.05666721e+00
2.78884977e-01 9.46606755e-01 7.44218469e-01 -3.19998354e-01
-7.99287140e-01 -1.00014102e+00 -3.56497735e-01 5.50531864e-01
3.28666233e-02 -4.89071846e-01 -4.42517221e-01 1.95351630e-01] | [8.602235794067383, 5.983490467071533] |
f8bfe342-d93f-47bb-8c19-714281a1748a | an-empirical-study-and-improvement-for-speech | 2304.03899 | null | https://arxiv.org/abs/2304.03899v1 | https://arxiv.org/pdf/2304.03899v1.pdf | An Empirical Study and Improvement for Speech Emotion Recognition | Multimodal speech emotion recognition aims to detect speakers' emotions from audio and text. Prior works mainly focus on exploiting advanced networks to model and fuse different modality information to facilitate performance, while neglecting the effect of different fusion strategies on emotion recognition. In this work, we consider a simple yet important problem: how to fuse audio and text modality information is more helpful for this multimodal task. Further, we propose a multimodal emotion recognition model improved by perspective loss. Empirical results show our method obtained new state-of-the-art results on the IEMOCAP dataset. The in-depth analysis explains why the improved model can achieve improvements and outperforms baselines. | ['Xinyu Dai', 'Yizhe Lu', 'Zhen Wu'] | 2023-04-08 | null | null | null | null | ['multimodal-emotion-recognition', 'speech-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech', 'speech'] | [-3.94754261e-02 -1.06236786e-02 1.05983485e-02 -5.52212775e-01
-1.33613980e+00 -2.77811915e-01 4.24377918e-01 -1.72037184e-01
-4.47707504e-01 3.90774280e-01 6.71309352e-01 1.33706644e-01
1.24610052e-01 1.02846920e-01 -3.57497990e-01 -6.81947708e-01
4.55349088e-02 3.19176950e-02 -4.33867931e-01 -4.13201481e-01
-1.82742253e-01 1.86948881e-01 -1.81149960e+00 8.10473502e-01
6.86848760e-01 1.38556695e+00 -1.79569826e-01 8.78675401e-01
-2.16621578e-01 1.11919343e+00 -6.82005227e-01 -8.68600965e-01
-3.61697614e-01 -4.50993419e-01 -7.59456873e-01 -3.35522438e-03
7.37622678e-02 -1.35911033e-01 -2.51508772e-01 8.68412554e-01
9.19192553e-01 4.36839283e-01 7.26715088e-01 -1.63123488e+00
-2.94517040e-01 8.80898833e-01 -3.14042777e-01 -4.05741893e-02
6.72592521e-01 -3.07240754e-01 1.00069189e+00 -9.91338432e-01
2.63791651e-01 1.46969211e+00 4.94103819e-01 7.79858530e-01
-6.34436786e-01 -7.04496503e-01 3.38519812e-01 6.01153374e-01
-1.20454717e+00 -1.00085056e+00 1.00917602e+00 9.74428877e-02
1.11579549e+00 4.56376940e-01 1.56096578e-01 1.75379527e+00
-2.72335321e-01 1.49684525e+00 8.44847858e-01 -4.72664773e-01
-5.29202521e-02 3.72731864e-01 2.87846774e-01 4.17830139e-01
-8.63712490e-01 -6.83278218e-02 -1.22938514e+00 -1.43737793e-01
-1.66395918e-01 -1.03284106e-01 -3.75806689e-01 2.70261258e-01
-1.02572143e+00 7.65482068e-01 -2.03284044e-02 4.31920260e-01
-5.58860898e-01 -1.19966328e-01 6.49741650e-01 6.55390978e-01
6.24277949e-01 -4.91434056e-03 -4.73877937e-01 -7.92939901e-01
-7.83747733e-01 -3.77079993e-01 9.32317734e-01 5.23109913e-01
4.06230450e-01 1.65799037e-01 -1.01841584e-01 1.31523108e+00
6.05394065e-01 4.31189030e-01 3.26128095e-01 -9.39283073e-01
6.04963005e-01 3.71968716e-01 -2.89869756e-01 -6.62153900e-01
-5.11347592e-01 -6.88064322e-02 -7.65138030e-01 -1.67433545e-01
-5.22333421e-02 -5.66386640e-01 -6.66610956e-01 1.81368172e+00
9.21467021e-02 3.00274253e-01 4.49250937e-01 8.89771938e-01
1.33078682e+00 7.65028059e-01 3.21973950e-01 -1.66672602e-01
1.49758244e+00 -1.25171411e+00 -1.21372545e+00 -1.71083272e-01
4.65230733e-01 -8.49763453e-01 8.79360259e-01 5.97746015e-01
-1.22644234e+00 -2.86811501e-01 -7.64486909e-01 6.31890371e-02
-5.87574542e-01 3.18078816e-01 6.70317292e-01 8.88105035e-01
-1.09087849e+00 -3.85039039e-02 -6.55021012e-01 -2.66377836e-01
1.98568374e-01 2.69972801e-01 -5.54816842e-01 1.86458737e-01
-1.53097796e+00 8.84836793e-01 2.01161459e-01 2.44623065e-01
-7.26694524e-01 -4.07990068e-01 -1.05583525e+00 2.32961893e-01
3.47971857e-01 -5.91873765e-01 1.59166169e+00 -1.24260259e+00
-1.99085200e+00 4.91302282e-01 -6.37606740e-01 -3.55618417e-01
5.11666238e-02 -2.53917754e-01 -9.85664129e-01 5.96207440e-01
-6.48035765e-01 9.31260526e-01 8.63349259e-01 -1.21769214e+00
-7.15306222e-01 -2.88507283e-01 -2.57150799e-01 4.53832865e-01
-8.66272092e-01 6.93198562e-01 -6.02822900e-01 -5.28027952e-01
-1.23890631e-01 -6.81461155e-01 1.33506924e-01 -5.68262637e-01
-2.29984760e-01 -4.74551529e-01 7.71114409e-01 -7.89648533e-01
1.28846526e+00 -2.35211968e+00 3.15789789e-01 -2.71026548e-02
8.46332386e-02 7.03009889e-02 -4.47599173e-01 5.72472990e-01
-1.39431477e-01 1.58939824e-01 2.91820932e-02 -1.09827125e+00
5.95313549e-01 -6.53154776e-02 -4.04788345e-01 1.97126847e-02
3.10323507e-01 9.25266922e-01 -2.98404872e-01 -5.51850259e-01
3.27206135e-01 1.22130454e+00 -3.47216606e-01 3.40736330e-01
1.50761038e-01 3.83873761e-01 -1.81545332e-01 9.22525346e-01
5.18035173e-01 1.61164701e-01 -1.20507814e-01 -3.55380416e-01
1.43819228e-01 4.15751398e-01 -9.89851713e-01 1.85058177e+00
-6.23220146e-01 7.68720329e-01 5.85893273e-01 -8.53685915e-01
6.61804080e-01 8.90955627e-01 4.72382575e-01 -6.82499886e-01
6.09138489e-01 -1.32992610e-01 -2.57073790e-01 -6.55178428e-01
6.18281305e-01 -2.18571275e-01 -3.08700383e-01 3.70713204e-01
3.52950037e-01 1.94848210e-01 -2.06523120e-01 1.71443373e-01
7.06259966e-01 -4.09840763e-01 -3.60371433e-02 4.54048574e-01
6.79627776e-01 -6.16839111e-01 2.09054187e-01 5.85605204e-01
-6.50619209e-01 4.60198998e-01 3.80302519e-01 1.22879297e-01
-1.78041294e-01 -7.84764051e-01 1.46971583e-01 1.68821764e+00
-8.82233977e-02 -4.54783231e-01 -5.80329180e-01 -7.99603045e-01
-4.97667044e-01 7.36521661e-01 -6.23920143e-01 -2.18686029e-01
-4.74217050e-02 -6.09585345e-01 1.03185391e+00 5.25614321e-01
3.97799015e-01 -1.07014859e+00 -1.70340165e-01 6.35547191e-02
-9.26212490e-01 -1.34233177e+00 -3.55178654e-01 1.90101326e-01
-3.19935888e-01 -3.90478462e-01 -7.21898139e-01 -7.01293349e-01
-6.41760379e-02 2.14542881e-01 1.06710744e+00 -4.34298426e-01
1.39155507e-01 9.20156121e-01 -7.75374532e-01 -7.16852546e-01
-2.67977715e-01 1.37781665e-01 -1.39327615e-01 4.27504390e-01
6.47737980e-01 -3.97318542e-01 -2.76931882e-01 2.67338932e-01
-9.36732352e-01 -2.67555475e-01 4.85056311e-01 7.20261693e-01
1.67337712e-02 1.95363201e-02 7.90426850e-01 -3.49370427e-02
8.42294395e-01 -4.52386111e-01 4.05905515e-01 4.20085758e-01
-8.41691643e-02 -7.53950253e-02 1.46770537e-01 -6.28795147e-01
-1.49926102e+00 -4.09725029e-03 -7.33708262e-01 -5.30642807e-01
-4.96513009e-01 5.98642826e-01 -5.89412868e-01 7.93303922e-02
8.70769378e-03 4.85384231e-03 -1.24856099e-01 -3.51289898e-01
5.77967227e-01 1.22486818e+00 3.93558323e-01 -5.14182806e-01
-1.17671594e-01 4.61830556e-01 -5.05758405e-01 -9.44157124e-01
-4.84084129e-01 -7.44474947e-01 -1.68368503e-01 -5.02661467e-01
8.34867358e-01 -1.21138990e+00 -1.20835376e+00 4.70805079e-01
-1.15428329e+00 3.12303938e-03 1.27077222e-01 7.81625748e-01
-4.19964612e-01 3.92927349e-01 -9.18340564e-01 -1.38530290e+00
-4.72760350e-01 -1.12795222e+00 1.30354571e+00 1.82861760e-01
-2.07658052e-01 -1.08119011e+00 -4.68103681e-03 6.22445703e-01
5.33915460e-01 -2.40470305e-01 2.54188627e-01 -8.58873129e-01
1.34344578e-01 -1.35861412e-01 -5.12333140e-02 5.05951285e-01
-2.47276485e-01 -4.09223326e-02 -1.59710991e+00 1.48201212e-01
1.65207863e-01 -6.82774723e-01 1.21532500e+00 1.47859201e-01
9.76964533e-01 -1.01910301e-01 -9.13535357e-02 3.20562035e-01
6.38336539e-01 2.17134535e-01 6.67959213e-01 -6.99833632e-02
4.83365357e-01 9.43885684e-01 4.58169907e-01 5.33413589e-01
8.73485506e-01 5.71238816e-01 3.89778554e-01 -1.60573274e-01
5.81112951e-02 5.27137816e-02 9.50438142e-01 1.19180310e+00
2.35185876e-01 -6.90020859e-01 -7.50523627e-01 4.37084436e-01
-1.96282935e+00 -1.03411138e+00 2.64970392e-01 1.49643779e+00
6.03785336e-01 -3.23218822e-01 2.65137404e-01 1.59601599e-01
5.87214828e-01 2.79421151e-01 -1.47263288e-01 -5.75808465e-01
-5.60336173e-01 -4.28095227e-03 -2.96603501e-01 4.68644112e-01
-1.24244690e+00 8.51920903e-01 6.62952948e+00 9.69992220e-01
-1.20015669e+00 2.76882440e-01 5.82283199e-01 -4.85171288e-01
-2.93292671e-01 -5.20873189e-01 -6.42042994e-01 1.70252338e-01
1.47337604e+00 2.07233101e-01 4.55953389e-01 5.05876303e-01
6.69247732e-02 5.75157516e-02 -9.92850661e-01 1.58848536e+00
6.84832752e-01 -6.71990037e-01 -8.47899914e-02 -2.22242445e-01
2.91980982e-01 8.57867301e-02 2.76843995e-01 8.14297259e-01
7.21839396e-03 -8.62111390e-01 5.76553106e-01 7.38170683e-01
2.33247861e-01 -1.01718533e+00 8.93476725e-01 1.03935547e-01
-1.14711416e+00 -1.65242270e-01 2.79591501e-01 7.87534118e-02
4.30281430e-01 3.35788488e-01 -6.04182899e-01 7.33674884e-01
8.82472932e-01 5.36525309e-01 -3.49000663e-01 6.21525288e-01
-6.83088750e-02 8.31379116e-01 -3.51359040e-01 -1.63590357e-01
1.81271955e-01 1.47138938e-01 5.33317804e-01 1.67312109e+00
4.31606233e-01 -7.52829015e-02 -1.32860735e-01 4.03087854e-01
-3.56019646e-01 3.93555701e-01 -4.28380489e-01 -3.56289536e-01
2.93414980e-01 1.38922751e+00 -1.98817775e-01 -4.02901560e-01
-5.67864656e-01 1.01909280e+00 3.25215966e-01 6.34258211e-01
-1.00877690e+00 -3.76722187e-01 8.97965312e-01 -8.42775524e-01
3.56145978e-01 6.48086071e-02 -5.07246517e-02 -1.31849670e+00
-1.28427550e-01 -1.09134650e+00 6.93049014e-01 -9.08481658e-01
-1.45726919e+00 8.37156892e-01 -3.05015832e-01 -9.62034106e-01
-3.94676507e-01 -6.02197170e-01 -5.88830948e-01 6.06394231e-01
-1.48850238e+00 -1.18776464e+00 -7.71472650e-03 8.78063619e-01
5.79285741e-01 -2.25150481e-01 9.19218421e-01 6.54036760e-01
-7.27490485e-01 9.85116601e-01 -2.05593824e-01 3.65708284e-02
1.07226586e+00 -9.96852219e-01 -4.33863729e-01 5.54299593e-01
2.80028194e-01 3.54419708e-01 6.26679838e-01 -9.81924012e-02
-1.38959551e+00 -5.02842426e-01 1.07109487e+00 -5.29255569e-01
6.27827227e-01 -3.15855265e-01 -7.64899015e-01 4.47115928e-01
9.93665040e-01 -6.29814982e-01 1.21622872e+00 5.21789730e-01
-6.32695019e-01 -2.60674860e-02 -9.56353068e-01 6.14839673e-01
4.88181293e-01 -1.02590835e+00 -6.93975925e-01 -2.30814651e-01
8.14378977e-01 -1.52097065e-02 -9.30796206e-01 6.14180982e-01
7.29604542e-01 -1.00256217e+00 9.45944011e-01 -6.55522466e-01
2.85521746e-01 7.78957456e-02 -5.27218342e-01 -1.51563489e+00
2.44131878e-01 -8.13054621e-01 -3.35670412e-01 1.62216377e+00
5.51853299e-01 -5.04997611e-01 3.38519931e-01 7.57446051e-01
-2.11872548e-01 -4.76383448e-01 -1.14798176e+00 -4.06281024e-01
-1.77617386e-01 -1.07177830e+00 6.61551237e-01 1.02963066e+00
6.91056967e-01 6.42406285e-01 -6.54611051e-01 1.77380413e-01
2.55442262e-01 -2.83752024e-01 4.99427557e-01 -8.82739484e-01
9.76927131e-02 -8.97603512e-01 -2.27286786e-01 -8.59780431e-01
7.89786577e-01 -6.15054786e-01 1.31874472e-01 -1.31355834e+00
1.72842130e-01 5.18432140e-01 -7.58097231e-01 5.36815941e-01
-2.89232105e-01 1.84916973e-01 3.79801631e-01 -3.90978456e-01
-1.26539123e+00 1.11927915e+00 7.03507364e-01 -2.24434257e-01
-2.02077612e-01 -1.30145967e-01 -8.93423557e-01 6.78590178e-01
8.75059605e-01 -2.05429286e-01 -2.22628459e-01 -3.18964511e-01
3.11563581e-01 2.49726459e-01 2.58647799e-01 -6.87757254e-01
5.19822359e-01 3.30043644e-01 2.42198482e-02 -8.98079097e-01
1.04352915e+00 -9.67802525e-01 -2.10779890e-01 -3.20770353e-01
-5.35342216e-01 -5.71757704e-02 5.81735015e-01 3.52551550e-01
-8.39932919e-01 -7.33747706e-02 2.50469327e-01 3.60029548e-01
-6.66482091e-01 -2.90985908e-02 -8.14625204e-01 -2.67722040e-01
5.95221817e-01 2.66479731e-01 -2.68279552e-01 -1.06841540e+00
-1.04619467e+00 3.76528531e-01 -2.65074134e-01 8.81288886e-01
7.55240381e-01 -1.44956338e+00 -7.51738369e-01 -3.36499400e-02
2.48994276e-01 -8.91753733e-01 7.68269002e-01 1.14133513e+00
4.25905973e-01 3.34074378e-01 1.37158886e-01 -4.80764270e-01
-1.71698046e+00 3.54171097e-01 4.40062135e-01 -1.11728698e-01
9.67443213e-02 9.62495446e-01 6.14925660e-02 -7.55606890e-01
9.00721133e-01 -1.10161647e-01 -4.49440241e-01 5.84374905e-01
8.62365901e-01 4.68177795e-01 9.20180827e-02 -9.71202850e-01
-5.43192565e-01 2.84164250e-01 7.33927563e-02 -6.91054523e-01
1.20417917e+00 -7.87300527e-01 -3.78065519e-02 7.51891017e-01
1.46896553e+00 -1.49693182e-02 -8.23047340e-01 -3.10678631e-01
-2.84803838e-01 4.37165424e-02 3.29982460e-01 -1.17569637e+00
-1.07017469e+00 1.33404052e+00 8.25761080e-01 2.33565733e-01
1.52435148e+00 2.03222334e-01 7.90456593e-01 5.44318318e-01
-2.11092278e-01 -1.28293824e+00 1.78681746e-01 7.84541547e-01
8.79979908e-01 -1.66179311e+00 -6.88791156e-01 -2.77385890e-01
-1.29417062e+00 1.01680112e+00 4.04616743e-01 7.45880485e-01
7.99728394e-01 3.10402513e-01 5.55862963e-01 -1.73546657e-01
-1.09745789e+00 -5.91865897e-01 6.88628793e-01 3.85828465e-01
5.94478846e-01 5.28182536e-02 1.58893958e-01 1.31028259e+00
-1.11866914e-01 -2.77435899e-01 -3.35461572e-02 6.58276975e-01
-2.44464099e-01 -1.06271708e+00 -5.81532180e-01 7.50616146e-03
-7.73359120e-01 -2.56324142e-01 -8.19627941e-01 4.19965476e-01
-2.06252784e-01 1.77624846e+00 -1.78125218e-01 -6.82800055e-01
3.59155685e-01 6.97062671e-01 2.22025037e-01 1.34809554e-01
-8.37214172e-01 6.11627996e-01 4.47441429e-01 -5.09559393e-01
-8.63613725e-01 -6.21389747e-01 -1.02210391e+00 -8.72382224e-02
-2.87048817e-01 3.23644578e-01 9.64607477e-01 9.59133744e-01
6.89762950e-01 7.47924924e-01 7.13652134e-01 -8.48411143e-01
-1.08151607e-01 -1.20776367e+00 -4.32358503e-01 3.53753924e-01
4.18435782e-01 -3.62645864e-01 -7.67817080e-01 -3.20168287e-02] | [13.25926399230957, 5.406862258911133] |
9d342f6c-aa7e-470c-9ef6-28c1626ba416 | model-invertibility-regularization-sequence | null | null | https://aclanthology.info/papers/N15-1063/n15-1063 | https://www.aclweb.org/anthology/N15-1063 | Model Invertibility Regularization: Sequence Alignment With or Without Parallel Data | null | ['Ashish Vaswani', 'David Chiang', 'Tomer Levinboim'] | 2015-05-01 | null | null | null | hlt-2015-5 | ['decipherment'] | ['natural-language-processing'] | [-2.44508207e-01 3.89024585e-01 -2.65282035e-01 -2.15905145e-01
-8.60921741e-02 -7.76765764e-01 4.48510379e-01 -7.23253429e-01
-5.48377395e-01 1.31954515e+00 3.66348401e-02 -9.49533224e-01
-2.40340635e-01 -1.05564880e+00 -8.44053447e-01 -8.75781775e-01
-7.42435038e-01 6.86515033e-01 1.44298598e-01 -6.52004302e-01
8.47113907e-01 6.29996777e-01 -1.62287033e+00 5.77558100e-01
6.81926727e-01 5.49597681e-01 5.66466339e-02 1.04480565e+00
6.13576770e-02 1.59847176e+00 -6.05450153e-01 -4.56729174e-01
1.43710867e-01 -1.88022718e-01 -5.35770595e-01 -3.36825520e-01
7.91536197e-02 -5.88986814e-01 -3.10633808e-01 6.07356608e-01
1.14270830e+00 -5.53557873e-02 1.07025254e+00 -1.51067197e+00
-7.97060013e-01 5.44234335e-01 2.30399013e-01 1.20632313e-01
6.10446513e-01 -1.79472014e-01 3.51212233e-01 -1.36621380e+00
6.32680655e-01 8.64870071e-01 9.47046995e-01 5.99273622e-01
-1.20647264e+00 -5.60552299e-01 -7.17137158e-01 -2.57835120e-01
-1.59567785e+00 -6.30042374e-01 6.23617955e-02 -3.34735900e-01
1.69142139e+00 6.72587395e-01 1.35647905e+00 1.47341585e+00
1.36649036e+00 4.72517729e-01 1.13904178e+00 -3.16340059e-01
3.65351558e-01 6.43754780e-01 -2.29148138e-02 5.51889002e-01
1.19686353e+00 6.76312149e-01 -5.60917914e-01 -7.92779744e-01
1.11365557e+00 -3.12949389e-01 2.60771513e-01 -7.06989706e-01
-1.06566191e+00 7.88466871e-01 -1.33808568e-01 2.20882341e-01
-3.46203089e-01 -9.92989019e-02 2.24548295e-01 4.12961632e-01
-2.94082850e-01 4.77862865e-01 -8.89795244e-01 -1.87686309e-01
-1.01472771e+00 1.91863477e-01 1.30615628e+00 1.45811689e+00
4.72722203e-03 3.70100170e-01 1.66493103e-01 2.89523482e-01
5.29429734e-01 1.05927479e+00 4.36951250e-01 -1.33365822e+00
-9.55119133e-02 4.20878716e-02 6.12505674e-01 -1.13473701e+00
-7.12930143e-01 -2.41300568e-01 -9.54724610e-01 5.18771529e-01
-1.59270123e-01 -4.59180593e-01 -7.59686172e-01 4.79171664e-01
-1.81984559e-01 -3.19489211e-01 5.45619786e-01 2.69257158e-01
7.28577793e-01 1.45816103e-01 9.25893057e-03 -4.73371416e-01
8.38296890e-01 -1.26899445e+00 -1.11426139e+00 2.33304992e-01
6.82289064e-01 -1.11704338e+00 4.06794518e-01 4.29345578e-01
-1.87200487e+00 -9.69799384e-02 -1.08317137e+00 1.19022481e-01
-7.97166348e-01 -2.57194549e-01 6.66624963e-01 1.49922979e+00
-1.63972652e+00 6.47176981e-01 -6.05021298e-01 5.40335439e-02
-4.73218322e-01 9.07614470e-01 -3.47142309e-01 2.90655226e-01
-1.28751957e+00 9.52449858e-01 -1.04219764e-01 3.03568155e-01
-3.30324143e-01 -1.84501112e-01 -1.03385210e+00 -6.24127567e-01
-2.59417385e-01 -1.07312346e+00 1.36302292e+00 1.09815992e-01
-1.30379498e+00 1.00962746e+00 -4.14315462e-01 -1.14109404e-01
6.61359608e-01 -1.52096570e-01 -8.18174660e-01 1.58321753e-01
2.87593510e-02 3.47163439e-01 5.30214727e-01 -9.69232559e-01
-6.31655395e-01 -1.00153096e-01 -2.54188716e-01 2.48816490e-01
1.43904146e-03 2.05205917e-01 6.38460100e-01 -9.53604504e-02
1.91062525e-01 -8.61961663e-01 -3.21419090e-01 -7.61404037e-01
-1.77353188e-01 -3.84736449e-01 5.59762537e-01 -2.92163283e-01
1.33848476e+00 -1.84365213e+00 -4.54683900e-01 2.94456817e-03
2.53527164e-01 3.89016122e-02 2.33101144e-01 1.23831403e+00
-2.43849114e-01 1.06667292e+00 3.98064762e-01 9.52541176e-03
3.13382268e-01 5.07273376e-01 2.86674555e-02 2.51206756e-01
9.43330750e-02 1.15280068e+00 -1.09214199e+00 -5.20940363e-01
5.57267785e-01 4.60968643e-01 -3.84588838e-01 6.24593735e-01
9.30012882e-01 2.99172718e-02 -8.94021094e-02 1.33241999e+00
1.05377018e+00 -2.90812284e-01 -2.46628653e-02 3.44544262e-01
-6.44541264e-01 6.00222833e-02 -7.41804183e-01 9.49429095e-01
1.31349964e-02 1.99757457e-01 1.74998537e-01 -6.80728197e-01
3.94654661e-01 1.10545933e+00 4.45659161e-01 -9.24297392e-01
-1.69569388e-01 9.00310338e-01 1.03920102e-01 -9.62983370e-01
6.40755475e-01 -1.72503099e-01 7.69825354e-02 1.40852645e-01
-3.61913115e-01 -3.84120375e-01 5.52458428e-02 4.19574350e-01
1.00224769e+00 -1.33707494e-01 5.48685491e-01 -8.72328460e-01
6.88946426e-01 -2.01589778e-01 3.00597936e-01 1.23978651e+00
-2.90862411e-01 9.21478033e-01 3.19022417e-01 -5.37574828e-01
-5.72780192e-01 -1.20098352e+00 -4.80556756e-01 9.86326516e-01
-1.58388391e-01 -3.87671381e-01 -7.10155964e-01 -2.63155133e-01
-1.85071185e-01 4.93483543e-01 -5.93591571e-01 5.23159266e-01
-5.14705896e-01 -6.85349941e-01 7.65277922e-01 3.17972392e-01
6.49624616e-02 -1.49228525e+00 -5.70813596e-01 3.34754556e-01
-1.07132711e-01 -5.39811492e-01 -1.57590687e-01 6.12205923e-01
-1.19366693e+00 -1.01526415e+00 -8.75530541e-01 -1.03603637e+00
7.18889177e-01 2.44572356e-01 1.18463767e+00 2.84042627e-01
-2.87709564e-01 6.49904788e-01 -1.31030381e-02 -2.67939389e-01
-3.16364132e-02 -5.78750819e-02 5.01899779e-01 -1.07851994e+00
6.46393359e-01 -3.92214775e-01 -1.00622165e+00 5.35368681e-01
-3.85821253e-01 -2.04611585e-01 7.85597622e-01 1.04281425e+00
1.42742008e-01 -1.33182362e-01 2.41971418e-01 -1.16735983e+00
1.11498392e+00 -4.85333532e-01 1.30341025e-02 2.62960047e-01
-8.53362501e-01 -4.53584343e-01 2.27881029e-01 -1.33037493e-02
-6.33760452e-01 -2.51423597e-01 -4.14545417e-01 3.22794706e-01
-4.06866640e-01 3.34393084e-02 2.24873364e-01 -5.26000679e-01
5.21005154e-01 1.50258854e-01 -4.07556817e-02 1.21735156e-01
6.30975887e-02 3.75316888e-01 3.67786407e-01 -7.01224983e-01
5.19809723e-01 1.93325520e-01 -3.00995037e-02 -1.28791058e+00
2.66786069e-01 -1.64763972e-01 -6.60750508e-01 -2.69126564e-01
7.75121748e-01 -9.50198710e-01 -6.35705650e-01 5.21708906e-01
-9.54387486e-01 -1.36340424e-01 -5.25335252e-01 6.30965471e-01
-7.02474535e-01 2.94757709e-02 -4.17852998e-01 -1.17976546e+00
-4.78847146e-01 -1.38869667e+00 9.59160328e-01 3.85633826e-01
-2.05811903e-01 -1.19790471e+00 3.38975668e-01 8.99848714e-02
4.29175526e-01 7.39965662e-02 5.03531873e-01 -2.80004859e-01
-2.98304528e-01 -2.85470933e-01 1.59573574e-02 -1.65876746e-01
-3.09365243e-02 4.80770379e-01 -5.48391879e-01 -4.37962860e-01
2.44738594e-01 -4.86500829e-01 -7.51581416e-02 7.11018145e-01
4.09077644e-01 -6.02920949e-01 -7.12544680e-01 5.56552589e-01
1.27709007e+00 4.72497940e-01 5.88340998e-01 6.77552521e-01
9.28869769e-02 5.17871320e-01 4.48484391e-01 2.08113909e-01
1.07506551e-01 1.08483166e-01 1.14217520e-01 -2.17137560e-01
9.45900455e-02 -1.75644696e-01 4.44304377e-01 1.09980536e+00
-8.23837519e-01 -2.15223029e-01 -4.88283783e-01 6.68056428e-01
-1.69156313e+00 -1.31570101e+00 -6.70152605e-01 9.76623476e-01
3.78074080e-01 3.38884830e-01 1.58041969e-01 1.21232189e-01
5.47049284e-01 -4.76693302e-01 -1.27305150e-01 -9.16352808e-01
-3.82233024e-01 4.21136379e-01 7.14838982e-01 6.51479006e-01
-6.10602558e-01 5.08361042e-01 1.25505590e+01 5.49717486e-01
-5.10860123e-02 1.27314463e-01 3.14815581e-01 1.99914098e-01
-2.24020526e-01 1.62632689e-01 -1.04165065e+00 9.26906094e-02
1.64754760e+00 -4.36735392e-01 3.93212438e-01 6.44444764e-01
3.94769877e-01 -1.96429595e-01 -9.78620768e-01 5.21999180e-01
1.12709433e-01 -1.50397491e+00 -3.76861334e-01 9.64888573e-01
5.80093563e-01 -1.81895152e-01 6.81941509e-01 4.33209062e-01
7.39425182e-01 -1.14519978e+00 7.88427055e-01 3.18626672e-01
8.87266219e-01 -8.18826377e-01 1.07315242e+00 2.76328743e-01
-1.04355454e+00 1.15487054e-02 -5.90443075e-01 -9.22581375e-01
1.62244976e-01 2.02188566e-01 -5.61583102e-01 4.06869203e-01
6.19966745e-01 2.15250537e-01 -8.53338003e-01 1.45581341e+00
-4.29231405e-01 4.22687940e-02 -3.00976187e-01 -4.39922899e-01
4.83290881e-01 5.28384298e-02 3.26606274e-01 1.02295005e+00
2.89785862e-01 5.45581102e-01 -2.62208551e-01 2.92286038e-01
6.63503349e-01 5.84383309e-03 -1.38496268e+00 -8.68683681e-02
3.54620427e-01 9.69217539e-01 -3.99494410e-01 -3.56748998e-02
-6.11456513e-01 6.62115753e-01 -1.66942239e-01 5.75146377e-01
-2.92761147e-01 -8.89406264e-01 4.33969021e-01 -2.30117049e-02
-2.88469017e-01 -1.35338321e-01 -3.23372036e-01 -1.02459288e+00
-6.06416404e-01 -3.90122980e-01 7.32179061e-02 -8.63867998e-01
-1.79313409e+00 6.85844719e-01 -3.50327820e-01 -1.03178740e+00
-9.29034173e-01 -9.72984850e-01 -2.90242791e-01 6.67755604e-01
-6.59234166e-01 -1.30476952e+00 1.77773401e-01 4.65253919e-01
1.82401046e-01 -5.88519633e-01 1.09945750e+00 -1.42670095e-01
-3.18711132e-01 1.08882773e+00 3.51253241e-01 -3.05974871e-01
8.32965672e-01 -1.16271269e+00 4.23785180e-01 1.09189451e-02
-6.59259140e-01 1.28460062e+00 6.81860268e-01 -6.81785107e-01
-1.16639841e+00 -1.52893141e-01 1.38110602e+00 -9.15031374e-01
5.27874053e-01 -1.02810413e-01 1.46300867e-01 6.57190859e-01
9.07351732e-01 -5.09461224e-01 1.08935249e+00 -1.13616258e-01
4.93123859e-01 8.23448122e-01 -1.38805819e+00 5.99272728e-01
1.21222687e+00 -4.81854081e-01 -8.27020168e-01 3.28937650e-01
9.94227231e-01 -5.60786903e-01 -1.34066856e+00 6.99696779e-01
1.14132881e+00 -1.04883564e+00 1.54237401e+00 -1.16912270e+00
4.08816934e-01 1.70770600e-01 -2.18455836e-01 -6.52935803e-01
-5.24122715e-01 -8.07741940e-01 -3.97186369e-01 3.40213925e-01
8.64144087e-01 -1.16800272e+00 5.53630650e-01 1.39166510e+00
-3.58625412e-01 -4.51715857e-01 -8.63314152e-01 -1.06149745e+00
-4.66870656e-03 1.81816339e-01 3.85523647e-01 8.18638563e-01
7.09820271e-01 2.02331603e-01 -2.05201268e-01 -4.37674336e-02
6.59316301e-01 -4.87512946e-01 5.82342803e-01 -8.74011815e-01
2.49149442e-01 -3.85976404e-01 -2.94489831e-01 -2.90531307e-01
-3.26795012e-01 -4.87107635e-01 -8.59293997e-01 -1.53084815e+00
8.49664882e-02 1.97367743e-01 -2.33256146e-01 3.44385266e-01
6.77823246e-01 2.45789096e-01 -2.27382466e-01 1.17054820e-01
-1.86927155e-01 -9.06788632e-02 9.18564737e-01 1.10557920e-03
-2.17724726e-01 3.92412305e-01 -1.01716614e+00 9.60534811e-01
7.85404518e-02 -6.09580100e-01 -5.51487744e-01 1.64939865e-01
5.71885824e-01 3.57775450e-01 4.49405432e-01 -9.70509052e-01
9.87445056e-01 -4.32846308e-01 6.12257063e-01 -1.43385470e+00
-2.65485227e-01 -9.00021493e-01 3.61826897e-01 6.29400074e-01
4.08042461e-01 3.53796333e-01 2.26811334e-01 -9.13986787e-02
-2.01378226e-01 -6.44658387e-01 4.43001747e-01 -7.45609999e-01
-4.55086857e-01 -3.90635371e-01 -1.02313221e+00 -9.67486948e-03
7.88939476e-01 -4.61360693e-01 -6.47767961e-01 -1.95199717e-02
-1.44608676e+00 7.30553791e-02 8.33678424e-01 1.34732248e-02
6.98022604e-01 -1.44444084e+00 -1.11748077e-01 6.18802965e-01
-5.42432606e-01 -4.98194307e-01 2.20674574e-01 1.06156552e+00
-1.31528914e+00 1.26915324e+00 -5.21896303e-01 -2.05424115e-01
-8.96585643e-01 7.19095051e-01 3.15837234e-01 1.75098896e-01
-3.93049836e-01 7.00151443e-01 1.12239346e-02 -6.58594191e-01
1.24011397e-01 2.97559589e-01 -4.87763733e-01 -2.52128989e-01
9.22277749e-01 1.17918742e+00 1.32526740e-01 -3.91715407e-01
-5.58191121e-01 4.39446121e-01 2.08604142e-01 -6.12734079e-01
1.38662541e+00 -2.20989168e-01 -7.44191706e-01 5.00318348e-01
7.50003338e-01 -7.33495504e-02 -3.28966863e-02 1.27011871e+00
4.80256975e-03 -5.47239244e-01 -3.11336160e-01 -7.82959402e-01
-3.42039734e-01 5.93214929e-01 6.17733955e-01 4.59080935e-01
9.69997942e-01 -6.47576571e-01 7.70057857e-01 7.64313757e-01
7.83133745e-01 -1.22997761e+00 -6.36951804e-01 3.69483382e-01
9.39836800e-01 -9.30032432e-01 1.30400375e-01 -3.14221263e-01
-5.55534840e-01 1.08689797e+00 5.38173854e-01 -1.62142932e-01
1.17091513e+00 6.96656525e-01 -1.75611794e-01 -4.53250527e-01
-1.04549897e+00 9.83817875e-02 2.10727617e-01 1.32608259e+00
8.26061428e-01 1.41275764e-01 -1.40426433e+00 8.46834600e-01
-6.34404242e-01 4.76996809e-01 9.47268784e-01 1.55018985e+00
-5.12773991e-01 -1.41692734e+00 -4.41162825e-01 2.31142953e-01
-2.82854587e-01 -3.49930972e-01 -8.62966061e-01 1.09929097e+00
1.97083559e-02 1.37330377e+00 -4.48758155e-01 -4.63731140e-01
8.27405751e-01 3.19600284e-01 4.41484600e-01 -1.25142381e-01
-8.94688129e-01 6.48750722e-01 3.84649158e-01 -1.05672014e+00
-7.94483483e-01 -1.27049911e+00 -1.08188879e+00 -1.16740942e+00
-3.07217389e-01 4.65777248e-01 2.31674016e-01 3.19533348e-01
-2.09303293e-02 1.56313553e-03 6.67447627e-01 -1.00111675e+00
-1.97397545e-01 -6.07395887e-01 -1.20775104e+00 -5.56946039e-01
3.94425273e-01 -5.85869551e-01 -1.15223050e+00 -6.24734610e-02] | [-1.5391874313354492, 15.869196891784668] |
3d65789d-7ac6-49f7-a16b-93338e9ba33c | few-shot-incremental-learning-with | 2104.03047 | null | https://arxiv.org/abs/2104.03047v1 | https://arxiv.org/pdf/2104.03047v1.pdf | Few-Shot Incremental Learning with Continually Evolved Classifiers | Few-shot class-incremental learning (FSCIL) aims to design machine learning algorithms that can continually learn new concepts from a few data points, without forgetting knowledge of old classes. The difficulty lies in that limited data from new classes not only lead to significant overfitting issues but also exacerbate the notorious catastrophic forgetting problems. Moreover, as training data come in sequence in FSCIL, the learned classifier can only provide discriminative information in individual sessions, while FSCIL requires all classes to be involved for evaluation. In this paper, we address the FSCIL problem from two aspects. First, we adopt a simple but effective decoupled learning strategy of representations and classifiers that only the classifiers are updated in each incremental session, which avoids knowledge forgetting in the representations. By doing so, we demonstrate that a pre-trained backbone plus a non-parametric class mean classifier can beat state-of-the-art methods. Second, to make the classifiers learned on individual sessions applicable to all classes, we propose a Continually Evolved Classifier (CEC) that employs a graph model to propagate context information between classifiers for adaptation. To enable the learning of CEC, we design a pseudo incremental learning paradigm that episodically constructs a pseudo incremental learning task to optimize the graph parameters by sampling data from the base dataset. Experiments on three popular benchmark datasets, including CIFAR100, miniImageNet, and Caltech-USCD Birds-200-2011 (CUB200), show that our method significantly outperforms the baselines and sets new state-of-the-art results with remarkable advantages. | ['Yinghui Xu', 'Pan Pan', 'Yun Zheng', 'Guosheng Lin', 'Nan Song', 'Chi Zhang'] | 2021-04-07 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zhang_Few-Shot_Incremental_Learning_With_Continually_Evolved_Classifiers_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zhang_Few-Shot_Incremental_Learning_With_Continually_Evolved_Classifiers_CVPR_2021_paper.pdf | cvpr-2021-1 | ['few-shot-class-incremental-learning'] | ['methodology'] | [ 3.69266689e-01 -2.59116255e-02 -1.82043076e-01 -2.26313591e-01
-1.50623530e-01 -3.38370711e-01 4.13739502e-01 2.43893951e-01
-4.38725650e-01 9.58714962e-01 -3.42041016e-01 4.28161509e-02
-2.75651872e-01 -1.01385641e+00 -8.51276577e-01 -8.52973342e-01
-6.23703972e-02 3.99984449e-01 5.84718347e-01 -2.53376514e-01
1.49372015e-02 3.40213716e-01 -1.95075536e+00 2.90051430e-01
1.14617097e+00 1.00210333e+00 3.35886776e-01 1.80715173e-01
-2.04959914e-01 7.99201727e-01 -5.74133575e-01 -3.91455710e-01
-3.19210743e-03 -6.12553895e-01 -6.38609290e-01 4.80216816e-02
2.70754516e-01 5.28752394e-02 -4.82863992e-01 8.63628566e-01
3.54880095e-01 4.18911129e-01 5.64919174e-01 -1.24920118e+00
-9.20334697e-01 7.24637747e-01 -4.32794869e-01 2.06241608e-01
-3.44826132e-02 1.50934935e-01 5.99064231e-01 -1.09235585e+00
7.05350995e-01 8.69578004e-01 8.62762630e-01 9.58826840e-01
-1.04415059e+00 -6.58677578e-01 6.20960057e-01 6.41157806e-01
-1.48318624e+00 -3.01424921e-01 8.23756397e-01 -1.16580322e-01
8.45712185e-01 4.29393351e-02 9.23838258e-01 1.12418711e+00
1.36624783e-01 8.57167125e-01 8.60271633e-01 -4.36120987e-01
6.99393272e-01 1.79217860e-01 4.55115080e-01 8.34835172e-01
4.15380299e-01 1.74110264e-01 -6.75945997e-01 -3.26471366e-02
2.02225998e-01 4.98433530e-01 -3.73527408e-01 -6.26065552e-01
-7.25377381e-01 7.03289449e-01 6.27282739e-01 4.32644188e-01
-2.12156147e-01 -8.82484764e-02 3.76247317e-01 6.59080803e-01
3.08109701e-01 2.07853168e-01 -6.61619306e-01 7.22446591e-02
-6.16302192e-01 -1.02683879e-01 5.21726489e-01 1.04675019e+00
9.79331195e-01 2.00533107e-01 -2.37873629e-01 1.00188994e+00
-2.48264194e-01 3.22841853e-01 1.04444373e+00 -3.88454974e-01
2.72754319e-02 8.39310825e-01 -5.00998139e-01 -7.08196402e-01
-1.99179322e-01 -9.19316590e-01 -9.14047599e-01 -2.80924290e-02
-9.08678025e-02 -6.76366538e-02 -1.14317894e+00 1.91566694e+00
3.22349608e-01 5.75935006e-01 1.43666312e-01 3.65704060e-01
7.26278424e-01 5.40923238e-01 4.12135199e-02 -6.89174771e-01
7.63416171e-01 -9.82202649e-01 -5.75009942e-01 -3.13907951e-01
4.26014483e-01 1.02339432e-01 1.21736109e+00 4.85433280e-01
-6.74343944e-01 -8.51165771e-01 -1.30684781e+00 4.51329619e-01
-6.82291031e-01 -2.97987282e-01 6.87668681e-01 5.01805663e-01
-7.15169370e-01 8.13736141e-01 -7.37878382e-01 -2.66328245e-01
7.40367532e-01 1.57611221e-01 -1.17070183e-01 -3.91582817e-01
-1.24099565e+00 7.54888773e-01 7.50979900e-01 -1.55700088e-01
-1.00936460e+00 -8.84086013e-01 -6.53203666e-01 2.87435859e-01
5.24813533e-01 -5.34540296e-01 1.07785451e+00 -1.27805293e+00
-1.50355422e+00 4.46071386e-01 -4.97676618e-02 -6.50211096e-01
3.12013447e-01 -1.32374257e-01 -4.96918589e-01 -7.64748156e-02
-3.17738473e-01 5.49901962e-01 1.08743191e+00 -1.21472931e+00
-7.75575876e-01 -3.62689644e-01 -1.56695485e-01 2.23957449e-01
-8.72752964e-01 -8.27190518e-01 -4.14008439e-01 -7.04582036e-01
3.90610993e-02 -9.15617585e-01 -3.02879270e-02 -2.06376836e-01
1.84910446e-01 -3.93304408e-01 1.03682041e+00 -7.75121972e-02
1.49021411e+00 -2.25117087e+00 3.18996042e-01 -1.72070310e-01
1.37501240e-01 5.46901464e-01 -2.70468056e-01 7.24488273e-02
-2.69945562e-02 -2.41108268e-01 -4.96175587e-01 -1.04160346e-01
-4.58248556e-01 5.12703776e-01 -3.87048781e-01 4.38618585e-02
8.10427591e-02 1.08820665e+00 -1.20458126e+00 -1.39433414e-01
-1.05734304e-01 3.84976476e-01 -5.53843498e-01 5.57654090e-02
-3.27114671e-01 1.84320018e-01 -4.99566719e-02 6.46378398e-01
5.94983876e-01 -3.48813206e-01 1.53315306e-01 3.78340669e-02
2.58655012e-01 -3.57456028e-01 -9.03931618e-01 1.98358798e+00
-3.96190494e-01 2.75905609e-01 -6.87722802e-01 -1.42917228e+00
8.79275501e-01 -1.64980125e-02 2.32654333e-01 -7.90934443e-01
1.45715000e-02 4.71689925e-02 -6.51399344e-02 -2.90200472e-01
1.04770638e-01 -2.04497185e-02 -8.90171714e-03 2.77295262e-01
6.75603926e-01 1.14720464e-01 2.10670903e-01 2.21601844e-01
1.14234710e+00 -9.42139030e-02 3.17325860e-01 6.91939238e-03
4.77656275e-01 -1.49059609e-01 9.76419091e-01 9.51219916e-01
-1.91579506e-01 3.80214930e-01 1.13396004e-01 -7.66070783e-01
-4.67816204e-01 -1.21786535e+00 -6.31570965e-02 1.20228422e+00
8.76410753e-02 -3.07301074e-01 -5.92474639e-01 -1.17552984e+00
1.43941462e-01 1.04381204e+00 -9.02338028e-01 -1.04412436e+00
-4.43178266e-01 -1.04327321e+00 2.29607373e-02 5.34901679e-01
6.39148891e-01 -1.07200539e+00 -6.79558873e-01 4.87147152e-01
1.97114617e-01 -6.87614799e-01 -3.66573215e-01 4.92255986e-01
-1.27838528e+00 -1.18833745e+00 -6.14751399e-01 -9.19954777e-01
8.94422412e-01 4.10099417e-01 1.02433801e+00 1.64209768e-01
-6.31812274e-01 5.11865914e-01 -6.52088642e-01 -4.73369420e-01
-8.57008770e-02 3.39626610e-01 2.18514159e-01 2.34282494e-01
4.96907413e-01 -8.03085506e-01 -4.48391527e-01 1.86864167e-01
-8.48893940e-01 -5.45534119e-02 7.26186216e-01 1.24805188e+00
7.62685895e-01 2.33655840e-01 1.19463599e+00 -1.31677926e+00
4.57176030e-01 -6.16272211e-01 -3.79555136e-01 6.50144875e-01
-1.13038087e+00 5.66531196e-02 8.91759515e-01 -9.59902585e-01
-1.19077814e+00 1.39675349e-01 2.48692125e-01 -6.21893108e-01
1.98660880e-01 6.52217090e-01 -3.39054950e-02 -1.56454608e-01
9.18864071e-01 6.73748851e-01 -2.65098270e-02 -3.60396892e-01
5.96202552e-01 4.27137554e-01 6.01868451e-01 -4.73556608e-01
7.37298310e-01 2.43674666e-01 -2.66405433e-01 -7.10161150e-01
-1.12836814e+00 -4.97944430e-02 -7.71523356e-01 -1.94802135e-01
2.43479773e-01 -7.91034698e-01 -3.00049901e-01 6.35143876e-01
-6.76874995e-01 -5.11935830e-01 -8.50900769e-01 1.37034506e-01
-3.19465339e-01 -6.35803025e-03 -4.09822464e-01 -5.73180199e-01
-3.56772810e-01 -6.48032665e-01 2.35898644e-01 6.13372445e-01
1.30569816e-01 -8.30345333e-01 1.12791918e-01 -2.57044613e-01
7.09216237e-01 1.24695562e-01 1.11106348e+00 -5.30324101e-01
-4.30983841e-01 -2.55083740e-01 1.16959929e-01 3.78849000e-01
3.59885752e-01 -4.37590837e-01 -1.01174402e+00 -7.75563538e-01
1.21806208e-02 -6.73464954e-01 1.29857290e+00 -5.59072662e-03
1.53441691e+00 -3.22623104e-01 -5.09581923e-01 7.33406365e-01
1.41142285e+00 4.95098889e-01 5.24431944e-01 1.84650086e-02
4.63711739e-01 2.01188013e-01 3.92564118e-01 4.04910654e-01
2.44904280e-01 3.39811414e-01 1.65159672e-01 4.37803566e-01
-5.71095467e-01 -4.78087127e-01 2.44183257e-01 1.14589238e+00
2.64156666e-02 3.85872498e-02 -6.66968763e-01 5.75849533e-01
-1.94338357e+00 -8.90260398e-01 6.81295574e-01 2.37626863e+00
1.13765168e+00 2.00924367e-01 -3.13481301e-01 2.19263792e-01
5.33817887e-01 -1.48235455e-01 -1.16860175e+00 8.03128332e-02
-1.51383728e-01 4.71402645e-01 9.98307765e-02 2.12180048e-01
-9.09716427e-01 1.05481970e+00 5.66906500e+00 8.90939236e-01
-1.22039664e+00 3.17500114e-01 7.33823419e-01 -3.71703655e-01
-1.02407217e-01 -6.37151077e-02 -8.82566094e-01 4.89178091e-01
1.01885283e+00 -5.10259151e-01 6.59259498e-01 1.11923349e+00
-6.59429610e-01 4.69135568e-02 -1.12477422e+00 1.19471610e+00
3.84747922e-01 -1.34325123e+00 2.82498896e-01 -4.95202243e-01
9.88081455e-01 8.71829689e-03 2.12530538e-01 1.03297567e+00
2.19813973e-01 -7.21212387e-01 5.01103699e-01 7.59076655e-01
9.69153583e-01 -8.67585063e-01 4.70377624e-01 5.22789538e-01
-1.12965477e+00 -6.01012588e-01 -6.47049546e-01 1.38258979e-01
-2.16418892e-01 6.44120395e-01 -5.75101495e-01 4.63936657e-01
8.57940495e-01 8.25550914e-01 -1.01146638e+00 1.25300884e+00
-3.04337978e-01 6.85967445e-01 -7.12924674e-02 -1.71454623e-01
-2.17651248e-01 1.95879877e-01 2.23947406e-01 9.27264214e-01
4.09683526e-01 3.46534699e-01 3.21949311e-02 6.39315307e-01
-3.91130120e-01 -9.69220847e-02 -6.64540052e-01 1.23600028e-01
7.49436438e-01 9.45800960e-01 -6.62998140e-01 -3.99568230e-01
-3.16274226e-01 1.31006396e+00 8.17283332e-01 4.77072775e-01
-6.36731446e-01 -5.91710627e-01 2.51869112e-01 -8.25942978e-02
5.38913369e-01 5.93300350e-02 -1.05921980e-02 -1.40813947e+00
-5.37297726e-02 -7.75999904e-01 6.29715323e-01 -3.47469002e-01
-1.45790029e+00 7.92790174e-01 -1.14390813e-01 -1.14139819e+00
-2.05161899e-01 -2.98905998e-01 -6.33293927e-01 1.45750478e-01
-1.59517992e+00 -8.82318318e-01 -6.86460495e-01 7.73822427e-01
7.84508526e-01 -5.82187176e-01 9.05630291e-01 2.73440361e-01
-6.52250469e-01 9.74621594e-01 2.45475337e-01 -9.96942073e-02
5.79775214e-01 -8.90042305e-01 2.01643780e-01 5.25859356e-01
3.28090608e-01 4.81181294e-01 2.82253891e-01 -6.09067559e-01
-1.36717522e+00 -1.35085535e+00 5.34493208e-01 -2.31622934e-01
3.03242981e-01 -4.65847790e-01 -1.29697764e+00 5.84208429e-01
-6.65300786e-02 4.47258621e-01 5.54273427e-01 1.16582885e-01
-4.44548309e-01 -5.39035976e-01 -9.77995098e-01 4.34160054e-01
1.38550019e+00 -2.47545332e-01 -7.43642628e-01 2.85903305e-01
9.79479969e-01 -1.02431394e-01 -2.83379525e-01 5.45293689e-01
4.04291093e-01 -6.91347361e-01 8.24778736e-01 -7.84779668e-01
-1.50698572e-01 -1.56168565e-01 -2.13045068e-02 -1.69141138e+00
-4.54855263e-01 -4.27476496e-01 -7.60468006e-01 1.16688156e+00
4.36620653e-01 -8.65914702e-01 7.62077153e-01 1.77461430e-01
-1.43496484e-01 -9.47053730e-01 -1.00511920e+00 -1.18220210e+00
1.50307685e-01 -1.36576861e-01 6.13756776e-01 1.14203143e+00
-9.11070630e-02 5.11374772e-01 -2.11927846e-01 -2.69258797e-01
7.71227360e-01 7.63010159e-02 3.85560751e-01 -1.62790680e+00
-3.42605412e-01 -2.67923415e-01 -2.63243109e-01 -7.30942488e-01
1.04817793e-01 -1.06825757e+00 -8.04699287e-02 -1.05119073e+00
3.81030828e-01 -6.89425111e-01 -8.61162186e-01 7.66740620e-01
-3.77521992e-01 -1.87064692e-01 1.81809992e-01 2.18395993e-01
-8.97704124e-01 1.03972042e+00 9.19874966e-01 -4.01513577e-01
-4.74746168e-01 2.95324326e-02 -7.72112250e-01 4.82694924e-01
6.35690093e-01 -6.49888039e-01 -1.02318454e+00 -3.77547652e-01
1.32610559e-01 -3.36622149e-01 -4.05594893e-02 -1.45465195e+00
5.94894528e-01 2.15948671e-02 7.06646264e-01 -3.21871221e-01
7.98870325e-02 -7.51826584e-01 1.58741120e-02 7.47472882e-01
-3.57693136e-01 -1.85575873e-01 2.41970152e-01 1.08865738e+00
-5.64150922e-02 -3.52580965e-01 1.02008367e+00 -2.09764093e-01
-9.52680111e-01 7.04873085e-01 -6.44606799e-02 2.02862501e-01
1.37912107e+00 8.08209646e-03 -3.93164217e-01 -1.30263669e-02
-8.34071934e-01 2.64612585e-01 2.98782825e-01 6.56237900e-01
9.84022319e-01 -1.43030751e+00 -5.52601099e-01 4.30331856e-01
3.64263922e-01 -5.04068360e-02 4.10692453e-01 3.95545930e-01
-6.08773436e-03 -8.16038437e-03 -2.31189653e-01 -4.80784506e-01
-8.10232937e-01 1.07673764e+00 2.04228207e-01 -9.12385285e-02
-7.01235414e-01 1.08386981e+00 3.17119174e-02 -4.33939397e-01
5.79407096e-01 -3.10309529e-02 -2.59753913e-01 2.78354347e-01
8.30750287e-01 1.55295700e-01 2.15534642e-01 3.39445136e-02
-1.92511007e-01 3.35446775e-01 -6.16464317e-01 3.22117269e-01
1.51546216e+00 -3.61701250e-02 2.51503021e-01 7.56399155e-01
1.09937739e+00 -6.28715992e-01 -1.44620776e+00 -7.04087138e-01
2.63470747e-02 -2.53796160e-01 -7.38983881e-03 -1.06628489e+00
-1.30696177e+00 8.00792217e-01 1.02484739e+00 -1.21585295e-01
1.33813798e+00 -1.78935066e-01 7.75198281e-01 7.94735014e-01
7.17645645e-01 -1.36336803e+00 4.64226902e-01 6.71146691e-01
8.60875905e-01 -1.12809300e+00 -9.50457305e-02 -5.46014449e-03
-6.50306046e-01 1.14154601e+00 9.15365398e-01 -1.23483371e-02
8.42445970e-01 -7.81409293e-02 -3.92455190e-01 2.71020569e-02
-1.12310815e+00 -9.89587530e-02 1.57563254e-01 7.52501190e-01
3.69944647e-02 -1.78499728e-01 -1.89575419e-01 9.07551765e-01
1.82904691e-01 3.21291834e-01 2.55032241e-01 1.23037457e+00
-6.82570577e-01 -9.70906138e-01 2.02344030e-01 5.89127243e-01
1.32177353e-01 -7.06333295e-02 -2.85380334e-01 6.05848253e-01
3.23589981e-01 5.64317524e-01 7.56777897e-02 -5.62502325e-01
3.97532940e-01 4.13392156e-01 5.81441045e-01 -7.01439738e-01
-2.33152866e-01 -4.46166605e-01 -5.16160846e-01 -3.50292265e-01
-2.43267678e-02 -5.50332189e-01 -1.19679511e+00 -1.18200704e-02
-3.86279166e-01 3.07952642e-01 2.61186033e-01 7.41549253e-01
6.06630445e-01 8.53401482e-01 8.08827877e-01 -5.01237571e-01
-6.86521173e-01 -9.17023063e-01 -5.19526005e-01 3.67639720e-01
1.49759501e-01 -9.55203414e-01 -4.32826519e-01 7.28767067e-02] | [9.835930824279785, 3.384761333465576] |
ab834158-8c1c-4a45-8760-b4c43d7dd9e0 | adversarial-text-generation-via-feature | 1809.06297 | null | https://arxiv.org/abs/1809.06297v2 | https://arxiv.org/pdf/1809.06297v2.pdf | Adversarial Text Generation via Feature-Mover's Distance | Generative adversarial networks (GANs) have achieved significant success in generating real-valued data. However, the discrete nature of text hinders the application of GAN to text-generation tasks. Instead of using the standard GAN objective, we propose to improve text-generation GAN via a novel approach inspired by optimal transport. Specifically, we consider matching the latent feature distributions of real and synthetic sentences using a novel metric, termed the feature-mover's distance (FMD). This formulation leads to a highly discriminative critic and easy-to-optimize objective, overcoming the mode-collapsing and brittle-training problems in existing methods. Extensive experiments are conducted on a variety of tasks to evaluate the proposed model empirically, including unconditional text generation, style transfer from non-parallel text, and unsupervised cipher cracking. The proposed model yields superior performance, demonstrating wide applicability and effectiveness. | ['Lawrence Carin', 'Yizhe Zhang', 'Liqun Chen', 'Zhe Gan', 'Shuyang Dai', 'Haichao Zhang', 'Chenyang Tao', 'Dinghan Shen'] | 2018-09-17 | adversarial-text-generation-via-feature-1 | http://papers.nips.cc/paper/7717-adversarial-text-generation-via-feature-movers-distance | http://papers.nips.cc/paper/7717-adversarial-text-generation-via-feature-movers-distance.pdf | neurips-2018-12 | ['adversarial-text'] | ['adversarial'] | [ 7.78211117e-01 1.00265570e-01 7.13264644e-02 -1.57223180e-01
-1.18525982e+00 -6.40282691e-01 8.54476154e-01 -3.13611448e-01
-1.21432714e-01 1.23303473e+00 1.10014744e-01 -2.33755797e-01
1.78894356e-01 -8.14361393e-01 -6.57938898e-01 -1.04259050e+00
3.66833299e-01 2.82702059e-01 -3.07814062e-01 -3.47046465e-01
2.10732445e-01 2.91975439e-01 -1.06229234e+00 3.60115655e-02
1.20947254e+00 7.92067468e-01 2.27672786e-01 7.21459568e-01
-2.92788465e-02 5.32004297e-01 -1.10754967e+00 -7.84774840e-01
3.52632463e-01 -1.15490925e+00 -5.77780128e-01 1.38668329e-01
6.11516600e-03 -2.43927062e-01 -1.11380361e-01 1.05354810e+00
6.70751870e-01 9.68872234e-02 8.42252135e-01 -1.26159596e+00
-7.98207939e-01 5.44748008e-01 -5.17243028e-01 -2.14790404e-01
4.10493106e-01 2.20945299e-01 1.01377618e+00 -4.88472581e-01
5.07278085e-01 1.11900413e+00 5.55111408e-01 8.28529418e-01
-1.24494231e+00 -5.80958188e-01 -2.06581935e-01 -3.17764074e-01
-1.21785986e+00 -3.83724481e-01 8.76347899e-01 -1.61032587e-01
6.06914878e-01 3.97523701e-01 5.40027320e-01 1.50990653e+00
4.85890776e-01 9.15581286e-01 1.27484822e+00 -4.93546695e-01
3.00631493e-01 1.20910101e-01 -8.24020743e-01 4.79507953e-01
1.57145888e-01 4.03842293e-02 -4.28035706e-01 -4.05878760e-02
6.26637578e-01 -2.25770012e-01 -2.71389276e-01 -9.30905193e-02
-1.21091866e+00 9.36069787e-01 2.73709685e-01 1.13431513e-01
-3.74340743e-01 3.84199679e-01 2.70396382e-01 3.46240044e-01
6.12995267e-01 5.49418569e-01 2.73985323e-02 -4.11796480e-01
-9.87931430e-01 4.91737634e-01 4.82230484e-01 1.12054110e+00
4.74867493e-01 5.32225966e-01 -3.88149053e-01 6.60924494e-01
1.88786790e-01 6.82503760e-01 7.21312284e-01 -5.47969937e-01
8.58450890e-01 2.84594744e-01 -4.73264791e-02 -8.48347187e-01
3.33248004e-02 -4.73776311e-01 -1.25983691e+00 1.23763084e-01
2.44633108e-01 -5.37522256e-01 -9.65128124e-01 1.79745865e+00
8.13822076e-02 -7.45489225e-02 3.84755462e-01 5.54500520e-01
3.52633089e-01 8.47100258e-01 -1.91557974e-01 -2.21315712e-01
7.63748527e-01 -9.13914382e-01 -8.30749512e-01 -1.11091375e-01
4.05197740e-01 -8.78400564e-01 1.14205384e+00 2.42753118e-01
-1.21400344e+00 -4.72764432e-01 -1.18418658e+00 2.14036614e-01
-2.79687852e-01 -6.66362420e-02 4.79428887e-01 8.86431158e-01
-9.33467031e-01 7.21947312e-01 -7.78761446e-01 -1.04567289e-01
5.82135379e-01 3.49256128e-01 -1.30495235e-01 1.15558773e-01
-1.00039625e+00 5.11966109e-01 5.50316632e-01 1.95493296e-01
-9.90772367e-01 -3.25548202e-01 -6.70657873e-01 -9.70407650e-02
1.59398783e-02 -9.87666965e-01 1.03277934e+00 -1.12658823e+00
-2.21759367e+00 3.95382434e-01 -4.59294319e-02 -3.96158636e-01
8.18111658e-01 -2.14072973e-01 -3.38961154e-01 -4.14845906e-02
1.04964701e-02 7.25700557e-01 1.35849190e+00 -1.25662947e+00
-4.06355441e-01 4.29870188e-02 -1.80406054e-03 3.44195783e-01
-5.28980792e-01 -3.18286538e-01 -5.00029512e-03 -1.13906050e+00
-2.12703958e-01 -1.02442825e+00 -1.95970893e-01 -3.58270705e-01
-8.61842752e-01 9.83243063e-02 7.99853563e-01 -6.27744138e-01
9.27487493e-01 -2.00009394e+00 3.84182006e-01 6.10014014e-02
7.32196793e-02 2.06681192e-01 -2.12789357e-01 7.37039983e-01
1.20205097e-01 2.65217632e-01 -6.30689740e-01 -7.33786941e-01
7.54146650e-02 1.63628593e-01 -4.65511024e-01 2.45983705e-01
4.57425058e-01 1.16513956e+00 -8.92674983e-01 -4.71527278e-01
-2.99415402e-02 4.15890813e-01 -4.74298507e-01 3.87414485e-01
-4.35425788e-01 6.99203610e-01 -5.23625016e-01 4.62202519e-01
6.23683035e-01 -9.04141366e-02 -7.86514878e-02 2.42499664e-01
2.85072535e-01 1.80081084e-01 -7.68940985e-01 1.93546116e+00
-4.97679800e-01 6.20269775e-01 -4.13029760e-01 -7.77505934e-01
1.09728348e+00 2.78915823e-01 1.13510370e-01 -6.23688877e-01
-6.51581725e-03 1.37251168e-01 1.26353474e-02 -1.52231604e-01
7.92723596e-01 -3.22017938e-01 -2.86555260e-01 6.01708472e-01
1.32411392e-02 -5.29713035e-01 1.49926409e-01 1.30851537e-01
1.02410030e+00 3.64199191e-01 1.06373347e-01 -1.26304030e-01
3.56017113e-01 -2.22284362e-01 3.24396282e-01 7.03749716e-01
2.92093128e-01 8.99000525e-01 6.69026554e-01 1.93147920e-02
-1.20290780e+00 -1.16838741e+00 1.74063161e-01 6.97858036e-01
9.38286483e-02 -3.53803992e-01 -1.11432242e+00 -9.75661695e-01
-3.65638912e-01 9.27203596e-01 -6.66837931e-01 -3.35489690e-01
-5.70712984e-01 -1.01295412e+00 8.63838613e-01 4.69293028e-01
5.76015234e-01 -1.15313947e+00 -3.44048738e-01 1.85258076e-01
-3.61485690e-01 -9.22681630e-01 -6.09942198e-01 1.03104979e-01
-8.99542928e-01 -3.94754082e-01 -1.00181317e+00 -5.75349987e-01
8.33406866e-01 -2.01851144e-01 8.83387089e-01 -2.38315791e-01
-1.12908386e-01 6.27015829e-02 -5.24029136e-01 -3.10351491e-01
-9.26042855e-01 4.54187125e-01 -1.48694009e-01 2.90407151e-01
-3.31421316e-01 -5.10361791e-01 -6.29298270e-01 1.67481527e-01
-1.17126226e+00 2.20024556e-01 8.83048594e-01 1.30529690e+00
5.24429679e-01 2.10259110e-01 8.85019004e-01 -9.47620153e-01
1.01416624e+00 -2.92796433e-01 -3.61593992e-01 2.68594384e-01
-7.67351985e-01 2.05789849e-01 1.25558543e+00 -4.16643590e-01
-1.25421512e+00 -2.44397521e-01 -2.64563024e-01 -2.02276319e-01
-1.20693348e-01 1.87594429e-01 -3.82439166e-01 -2.17145160e-02
4.51135963e-01 7.59256959e-01 -6.10836111e-02 -6.16053678e-02
4.72665370e-01 6.49130642e-01 5.02518237e-01 -6.39575660e-01
1.34176004e+00 3.00461888e-01 3.00924275e-02 -6.06921315e-01
-5.20112693e-01 3.14644754e-01 -3.92576724e-01 3.80519256e-02
8.61764610e-01 -6.78918064e-01 -4.19884384e-01 6.95250154e-01
-1.05213845e+00 -4.40594107e-01 -6.22776270e-01 2.75719553e-01
-7.91736901e-01 3.75167817e-01 -6.34018779e-01 -7.96231329e-01
-7.45849669e-01 -1.13803208e+00 1.19395649e+00 2.41353109e-01
-1.28735796e-01 -1.17354655e+00 8.42268839e-02 2.67113894e-01
6.09729409e-01 6.90385163e-01 9.54107046e-01 -4.34214383e-01
-4.96348441e-01 -4.22750413e-01 2.87213344e-02 5.82669854e-01
2.96663463e-01 -1.16682634e-01 -8.88898373e-01 -5.88556528e-01
-1.56068932e-02 -5.02841234e-01 5.96700907e-01 3.71503718e-02
1.03976119e+00 -3.98287773e-01 -1.30851954e-01 8.61500680e-01
1.34827828e+00 3.29402685e-01 9.34180081e-01 1.89597502e-01
7.71498322e-01 2.59169638e-01 4.87644166e-01 6.32338643e-01
7.17187524e-02 5.91419578e-01 4.36644793e-01 -2.58368433e-01
-1.09968796e-01 -6.32577717e-01 5.25645077e-01 1.01622427e+00
-1.31298810e-01 -8.02766025e-01 -4.51358616e-01 2.72162139e-01
-1.51898849e+00 -8.26946139e-01 2.03348264e-01 2.06492877e+00
8.90550911e-01 3.78685772e-01 -2.93074995e-02 1.32019177e-01
7.17002153e-01 2.40121052e-01 -5.46231151e-01 -4.95630145e-01
-2.90142626e-01 4.54802275e-01 2.96295434e-01 3.03310066e-01
-8.37253690e-01 9.29552078e-01 6.30488777e+00 1.22934556e+00
-1.18137181e+00 1.72142405e-02 7.86739409e-01 1.24789655e-01
-5.91281950e-01 -5.21709472e-02 -6.84461892e-01 7.21220732e-01
7.52497315e-01 -2.62837797e-01 5.84428787e-01 5.56510866e-01
6.17923625e-02 2.22581074e-01 -1.02048719e+00 7.55630314e-01
1.88706219e-01 -1.22271740e+00 2.08055064e-01 5.54293506e-02
1.08540034e+00 -5.33992529e-01 4.76637751e-01 1.45688996e-01
2.87661254e-01 -1.17342806e+00 7.96255648e-01 4.21935529e-01
1.23050797e+00 -8.62176001e-01 7.31374502e-01 2.44602680e-01
-9.98065591e-01 1.17008619e-01 -2.00046107e-01 2.50093728e-01
2.32521921e-01 4.00843322e-01 -9.42416251e-01 8.05083513e-01
1.44322231e-01 6.18797898e-01 -5.25489748e-01 6.42054439e-01
-2.52862990e-01 6.11435890e-01 -1.16102196e-01 -2.78224349e-01
2.59421438e-01 -2.76836425e-01 6.20621681e-01 1.05205154e+00
5.01202703e-01 -4.34465259e-01 -2.88919806e-01 1.07006872e+00
-2.58658797e-01 -2.80297305e-02 -7.24788547e-01 -3.48259777e-01
1.54205561e-01 1.16454852e+00 -7.85457730e-01 -7.98423290e-02
-1.54880192e-02 1.44000900e+00 3.58714946e-02 4.54652667e-01
-9.62172747e-01 -7.81099319e-01 3.03422987e-01 -1.56628668e-01
3.10261667e-01 -1.51884064e-01 -3.23667347e-01 -1.04320383e+00
2.51129866e-01 -1.07690966e+00 -4.58211191e-02 -5.34982502e-01
-1.37027645e+00 9.02616620e-01 -2.44306345e-02 -1.49432266e+00
-5.89648008e-01 -1.67208195e-01 -8.51287901e-01 9.24899518e-01
-1.34851027e+00 -1.40679741e+00 -3.10868651e-01 6.41994238e-01
6.77654386e-01 -4.72683042e-01 8.97902846e-01 1.04795605e-01
-5.63772917e-01 1.26796436e+00 5.12748301e-01 -6.77484870e-02
5.39537907e-01 -1.38192987e+00 8.60310495e-01 8.99087608e-01
-4.18840349e-02 3.01904052e-01 6.94131255e-01 -6.37129128e-01
-1.49642444e+00 -1.23999846e+00 4.29436922e-01 -2.00765327e-01
4.71461058e-01 -7.08314896e-01 -4.23958123e-01 5.25421023e-01
4.22537804e-01 -4.94500458e-01 5.24580121e-01 -5.18077374e-01
-3.77795398e-02 -1.64485916e-01 -1.23795474e+00 7.37324119e-01
1.04071081e+00 -2.51844347e-01 -2.02958420e-01 4.09293950e-01
7.75488436e-01 -4.45474178e-01 -6.94287121e-01 2.54421711e-01
3.70246977e-01 -8.01064491e-01 6.80537343e-01 -3.91938031e-01
7.89354384e-01 -8.64736214e-02 -1.16259545e-01 -1.63378251e+00
-1.85347758e-02 -1.30082691e+00 -3.37040834e-02 1.47627890e+00
5.43560743e-01 -7.97346652e-01 8.61537457e-01 1.11378767e-01
-2.93310046e-01 -8.04495335e-01 -9.08824384e-01 -8.46139431e-01
3.11891794e-01 -3.18826996e-02 7.50828266e-01 7.20022500e-01
-1.77561462e-01 4.19768929e-01 -7.49070823e-01 -2.37025529e-01
6.84584260e-01 1.50236130e-01 9.09588754e-01 -6.98497653e-01
-7.09247708e-01 -4.18540955e-01 -3.09870780e-01 -1.27541399e+00
2.05599919e-01 -9.80431378e-01 2.95337766e-01 -1.26872492e+00
-4.30471972e-02 -5.61044514e-01 9.74893495e-02 1.99721128e-01
-3.22031647e-01 3.98312330e-01 3.62472713e-01 1.54148772e-01
-4.08721834e-01 1.14459336e+00 1.57761443e+00 -1.98816508e-01
-1.45877883e-01 2.03608602e-01 -6.28767788e-01 3.15196484e-01
9.68783617e-01 -6.34139955e-01 -5.09681702e-01 -3.69427919e-01
1.77241132e-01 2.52000034e-01 1.33651450e-01 -1.07850659e+00
-1.34381562e-01 -1.25216916e-01 4.84030306e-01 -2.85156637e-01
4.50850725e-01 -6.41038895e-01 2.69530684e-01 4.41099107e-01
-5.32885492e-01 2.99985707e-01 -5.35570383e-02 7.38628507e-01
-1.50051504e-01 -2.79608786e-01 7.28153169e-01 9.87534374e-02
2.45963074e-02 2.13968515e-01 -1.90086797e-01 2.50768960e-01
1.09179366e+00 -1.26727134e-01 -2.97710329e-01 -5.74957013e-01
-3.01000088e-01 -2.17084736e-01 5.02019882e-01 3.11081260e-01
7.57799983e-01 -1.50748158e+00 -8.72343957e-01 3.65045875e-01
-9.54148173e-02 4.35659364e-02 8.86660591e-02 6.21696711e-01
-5.92335403e-01 2.04576850e-01 -1.66172042e-01 -5.26857376e-01
-7.82929301e-01 2.56315202e-01 2.30061606e-01 -6.39147401e-01
-6.43120587e-01 7.54587591e-01 1.69673473e-01 -2.75349170e-01
-8.06260258e-02 -4.45480412e-03 2.97061920e-01 -3.58394235e-01
1.41898766e-01 2.83581197e-01 5.19012325e-02 -4.27570999e-01
1.00061446e-01 4.28697854e-01 -1.89220384e-01 -3.24755609e-01
1.03769279e+00 -7.80738797e-03 8.90190452e-02 3.44330132e-01
1.12785065e+00 -5.24171954e-03 -1.40849519e+00 2.17087492e-02
-4.28269893e-01 -4.58513439e-01 -4.05818433e-01 -6.29747987e-01
-1.06075120e+00 9.17365432e-01 5.57836652e-01 4.39910531e-01
1.18187535e+00 -3.79398704e-01 1.28079879e+00 2.17927650e-01
2.58402497e-01 -9.27333355e-01 4.60186213e-01 2.64108300e-01
8.45100582e-01 -1.00874007e+00 -1.54387876e-01 -8.90901163e-02
-7.44860172e-01 1.03001332e+00 5.11901081e-01 -8.81581381e-02
1.83191925e-01 3.69343400e-01 4.26681377e-02 3.80025208e-01
-6.38937950e-01 1.65300936e-01 4.66324016e-02 8.08864474e-01
2.63454080e-01 1.38534456e-01 -2.67680705e-01 1.95481956e-01
-5.54610312e-01 -2.18518078e-01 4.94504392e-01 9.03415978e-01
1.64745763e-01 -1.44232464e+00 -2.10696280e-01 4.93550152e-01
-5.57076633e-01 -2.08458722e-01 -3.09324682e-01 6.32267416e-01
-1.38741016e-01 9.52583730e-01 -1.00869149e-01 -5.93212485e-01
-4.55133580e-02 8.05896055e-03 4.56163526e-01 -3.45814049e-01
-6.17567658e-01 9.05329809e-02 -1.44796103e-01 -4.46404964e-02
-2.80759603e-01 -5.31046987e-01 -9.21212673e-01 -4.14721370e-01
-5.82842946e-01 1.90130576e-01 6.99818492e-01 9.40808058e-01
2.89308488e-01 6.09659672e-01 1.09963608e+00 -8.08478594e-01
-8.34839046e-01 -1.06997263e+00 -5.06814957e-01 5.99031627e-01
2.18326166e-01 -2.74551004e-01 -3.17824841e-01 2.06163168e-01] | [11.78397274017334, 9.322667121887207] |
efc00731-6573-4b94-91ac-28388b757de6 | sematch-semantic-entity-search-from-knowledge | null | null | https://ceur-ws.org/Vol-1556/paper2.pdf | https://ceur-ws.org/Vol-1556/paper2.pdf | Sematch: Semantic Entity Search from Knowledge Graph | As an increasing amount of the knowledge graph is published as Linked Open Data, semantic entity search is required to develop new applications. However, the use of structured query languages such as SPARQL is challenging for non-skilled users who need to master the query language as well as acquiring knowledge of the underlying ontology of Linked Data knowledge bases. In this article, we propose the Sematch framework for entity search in the knowledge graph that combines natural language query processing, entity linking, entity type linking and semantic similarity based query expansion. The system has
been validated in a dataset and a prototype has been developed that translates natural language queries into SPARQL. | ['Ganggao Zhu and Carlos A. Iglesias'] | 2015-06-01 | null | null | null | joint-proceedings-of-the-1st-international | ['entity-linking', 'semantic-textual-similarity'] | ['natural-language-processing', 'natural-language-processing'] | [-4.76099253e-01 5.04639506e-01 -2.38823801e-01 -3.51932853e-01
-3.54208946e-01 -8.45457315e-01 3.78098428e-01 8.86291802e-01
-5.66263437e-01 8.85179579e-01 -2.91841850e-02 -2.23453000e-01
-5.59094548e-01 -1.50542784e+00 -3.82977426e-01 4.74972457e-01
-9.46121141e-02 8.95235717e-01 9.77860093e-01 -6.94307208e-01
-3.95032670e-03 4.92248148e-01 -1.84259868e+00 2.01761842e-01
1.34203625e+00 9.66243923e-01 1.65580720e-01 1.58242181e-01
-9.78263021e-01 7.78369069e-01 -3.81623805e-01 -6.39887154e-01
5.44904508e-02 2.31018998e-02 -1.31786740e+00 -7.84848630e-01
3.79865706e-01 3.07943344e-01 6.36693612e-02 1.31960630e+00
4.21913385e-01 1.56666920e-01 -1.62189215e-01 -1.25762534e+00
-7.13807464e-01 4.96306419e-01 3.64731163e-01 1.49032958e-02
8.88768613e-01 -7.19002128e-01 9.26750422e-01 -7.21458495e-01
1.28630900e+00 1.14142680e+00 3.58181119e-01 1.44749299e-01
-6.83081806e-01 -2.93933660e-01 -5.61774135e-01 5.45537949e-01
-1.74587119e+00 -1.09591089e-01 2.78083503e-01 -2.75691122e-01
1.39529610e+00 3.71619105e-01 6.86354816e-01 -3.70381214e-02
-3.06851447e-01 1.40169114e-01 5.92278242e-01 -7.67868340e-01
3.28850150e-01 6.40266836e-01 3.64367306e-01 8.06802869e-01
7.42566586e-01 -1.51261806e-01 -5.94063461e-01 -2.89110303e-01
4.23380405e-01 -3.34566146e-01 1.29007488e-01 -6.93918824e-01
-7.16593266e-01 5.86961985e-01 6.13589585e-01 6.22660697e-01
-5.92899323e-01 -1.97617024e-01 5.15982568e-01 2.37984240e-01
1.39267281e-01 7.25914240e-01 -4.12876934e-01 2.21490368e-01
-6.14244521e-01 3.27020317e-01 1.40652621e+00 1.49633908e+00
1.29694033e+00 -5.72510242e-01 1.86981052e-01 6.93986535e-01
4.52931285e-01 4.64533180e-01 3.46303910e-01 -9.78973627e-01
2.44494140e-01 1.61071038e+00 2.85239607e-01 -1.01149940e+00
-3.04029346e-01 -1.24191746e-01 -7.83808678e-02 -1.81689575e-01
-2.25021526e-01 3.68882477e-01 -5.72715759e-01 1.23260021e+00
6.54606879e-01 -3.16511750e-01 6.14370584e-01 5.07651925e-01
1.30073738e+00 2.87786126e-01 5.35698056e-01 -4.53479961e-02
1.61743879e+00 -3.92824948e-01 -9.37134266e-01 7.30572641e-03
8.14391851e-01 -6.02236032e-01 7.42885768e-01 -3.82288367e-01
-9.66353297e-01 -3.07265311e-01 -7.43176103e-01 -4.65752184e-01
-1.37961102e+00 -5.65988183e-01 6.66886568e-01 4.75718856e-01
-1.02159512e+00 3.38245854e-02 -4.80989337e-01 -1.16334307e+00
2.50268757e-01 3.55304331e-01 -5.23137927e-01 -2.95079648e-01
-1.79082286e+00 1.14012885e+00 1.52677405e+00 -4.67888504e-01
-6.85794652e-02 -9.96211231e-01 -1.05717015e+00 1.69568554e-01
6.47812247e-01 -9.67026651e-01 8.75033379e-01 -3.94684017e-01
-7.18968272e-01 1.03218830e+00 4.55727801e-03 -5.29274523e-01
-9.01829079e-02 -7.75657371e-02 -1.24809146e+00 1.53450191e-01
5.25941849e-01 3.88452202e-01 -2.09882796e-01 -1.04074430e+00
-1.07719839e+00 -7.11081505e-01 3.15245122e-01 1.85476452e-01
-2.87324816e-01 4.54392940e-01 -7.20748782e-01 -1.56590641e-01
1.35022756e-02 -3.40249449e-01 8.04358795e-02 -1.94647282e-01
2.13528186e-01 -5.20269871e-01 8.06367040e-01 -6.11396432e-01
1.67101526e+00 -1.86906075e+00 -2.47524709e-01 8.07595968e-01
8.85129869e-02 2.19439104e-01 3.34494382e-01 8.93124104e-01
2.07218260e-01 3.30134988e-01 -2.05302052e-03 8.51865232e-01
1.98981136e-01 5.88244438e-01 -1.03577644e-01 -4.71329778e-01
-3.23477119e-01 1.06141174e+00 -9.42197561e-01 -1.01777554e+00
-1.49092227e-02 2.26530269e-01 -5.05667686e-01 1.93830281e-02
-5.83385110e-01 -2.48493612e-01 -7.96981931e-01 9.31250274e-01
3.04755360e-01 -5.17696142e-01 5.51672757e-01 -4.88908976e-01
-2.27754027e-01 6.62114844e-02 -1.24223638e+00 2.08031511e+00
-3.63899648e-01 4.96945530e-02 -2.58105308e-01 -6.65670753e-01
8.80308270e-01 6.39833212e-01 6.35990739e-01 -1.04253244e+00
-3.10888767e-01 7.30253458e-01 -7.35732794e-01 -1.04473114e+00
7.83146203e-01 3.53710651e-02 -2.80408561e-01 1.23226531e-01
2.12843537e-01 1.57616362e-02 7.38392353e-01 5.11130095e-01
1.01066065e+00 2.17918038e-01 6.82645798e-01 -2.93132007e-01
7.40085959e-01 8.54123473e-01 3.04850668e-01 3.00825059e-01
4.59039360e-01 -5.94371736e-01 -2.43860297e-02 -5.45322299e-01
-8.10294271e-01 -1.10587418e+00 -9.11552161e-02 1.04277205e+00
4.29492176e-01 -7.18507707e-01 -5.77588320e-01 -4.75669742e-01
3.07674050e-01 7.82028794e-01 9.06110108e-02 -2.71874368e-02
-4.03169870e-01 6.19063303e-02 5.04952848e-01 3.04145157e-01
5.86786389e-01 -1.17818880e+00 -7.31937349e-01 2.06321269e-01
-6.50709942e-02 -1.45935822e+00 9.16543975e-02 -3.42490345e-01
-5.79751968e-01 -1.34590864e+00 1.79676592e-01 -1.00549650e+00
5.30882478e-01 -2.24415794e-01 1.44473636e+00 1.78593233e-01
-3.51905704e-01 7.46519923e-01 -5.31956434e-01 -6.66644692e-01
-2.63793677e-01 3.46253335e-01 -3.47475827e-01 -5.61019182e-01
9.86375749e-01 -3.04238856e-01 -2.27130771e-01 3.08651447e-01
-1.48111510e+00 -4.57173169e-01 3.13306212e-01 1.21502414e-01
6.15596533e-01 3.30260098e-01 7.39462197e-01 -1.00966859e+00
6.41267300e-01 -6.42453372e-01 -9.55135822e-01 9.66488063e-01
-1.00490761e+00 3.19637120e-01 8.48495401e-04 3.40287477e-01
-1.17194569e+00 6.05863705e-03 7.47083724e-02 1.59953125e-02
-7.52466992e-02 1.33600855e+00 -4.54430401e-01 -1.31432116e-01
6.33607507e-01 -3.44744563e-01 -1.66840136e-01 -7.17045128e-01
6.73878789e-01 5.37460446e-01 7.39100993e-01 -6.20724916e-01
7.10424781e-01 3.16066474e-01 6.36156648e-02 -6.26588821e-01
-6.05228424e-01 -7.52498388e-01 -5.97190022e-01 -1.94773339e-02
8.97475719e-01 -9.47453082e-01 -7.34068871e-01 -3.54393631e-01
-1.07679701e+00 3.23809326e-01 -7.57596970e-01 4.39213872e-01
-3.03995967e-01 1.55859530e-01 1.24964997e-01 -2.42998749e-01
-4.10102427e-01 -3.93415630e-01 6.03313386e-01 3.30775797e-01
-2.25103676e-01 -1.01312768e+00 4.44326699e-01 4.83416796e-01
5.81549942e-01 5.46161234e-02 1.27423024e+00 -1.00803304e+00
-7.67916441e-01 -6.59287453e-01 -2.62381554e-01 -1.95052922e-01
1.73188835e-01 -3.21097642e-01 -2.85492122e-01 4.63523977e-02
-7.28940845e-01 -1.52022421e-01 -1.75556362e-01 -5.52436829e-01
4.56496567e-01 -2.10135788e-01 -5.30214667e-01 2.30968639e-01
1.79125357e+00 2.56236047e-01 6.91830397e-01 7.82731414e-01
3.36682796e-01 6.32228851e-01 6.97897434e-01 1.68605387e-01
7.91880369e-01 6.95770979e-01 1.67834699e-01 2.52409786e-01
1.64871365e-01 -5.79049110e-01 -4.95771587e-01 5.55683017e-01
3.15204598e-02 -2.52586007e-02 -1.25176334e+00 6.04959249e-01
-1.91321278e+00 -9.63565886e-01 5.61581925e-02 2.11935854e+00
9.18873787e-01 -3.99447799e-01 -1.80104539e-01 -2.31930047e-01
7.14200497e-01 -4.11074400e-01 -3.26703906e-01 -1.35324061e-01
-1.48719579e-01 5.82957506e-01 5.07386029e-01 5.83780050e-01
-5.04285395e-01 1.27629530e+00 6.09719419e+00 4.86984611e-01
-7.00177610e-01 1.90132275e-01 -6.71069443e-01 3.07603836e-01
-6.42765164e-01 5.72371602e-01 -9.22426701e-01 1.37924224e-01
8.18618894e-01 -9.62323666e-01 5.10692835e-01 8.66003513e-01
-2.22611964e-01 3.32593992e-02 -8.02659869e-01 8.99496257e-01
-1.31575540e-01 -1.56274247e+00 3.60701919e-01 -6.65894747e-02
3.68979454e-01 -4.88362536e-02 -8.78276706e-01 2.92144001e-01
4.81880635e-01 -7.01022744e-01 4.28268015e-01 9.86982167e-01
6.75779402e-01 -4.70963001e-01 6.69249415e-01 8.89598951e-02
-1.41576242e+00 -2.62523666e-02 -1.70697331e-01 4.16157186e-01
1.60702974e-01 4.14597094e-01 -8.96017611e-01 1.21915865e+00
9.38750267e-01 3.47049743e-01 -6.55266523e-01 1.12221384e+00
-5.00363931e-02 -2.48840839e-01 -5.54985523e-01 1.40331253e-01
-3.36715400e-01 -1.93637922e-01 2.78859556e-01 6.89873576e-01
6.15411162e-01 1.63982064e-01 2.07103252e-01 6.84290051e-01
-2.86717474e-01 6.94399357e-01 -6.33703828e-01 -4.65456843e-01
7.77886093e-01 1.03944504e+00 -3.56876671e-01 -5.77538431e-01
-7.17234135e-01 5.30773222e-01 3.57281387e-01 4.27060664e-01
-2.73430347e-01 -8.52455854e-01 5.19072533e-01 4.22370791e-01
1.57070845e-01 -7.39235431e-02 3.35353792e-01 -9.93422568e-01
3.25546741e-01 -7.48834014e-01 1.10799146e+00 -1.03042936e+00
-9.69259620e-01 4.66462970e-01 2.81586111e-01 -8.36094379e-01
-5.79033792e-01 -2.22425535e-01 2.67758165e-02 8.71787429e-01
-1.54801106e+00 -1.15206587e+00 -5.47422886e-01 8.65587831e-01
-3.36409807e-01 -8.74510705e-02 1.21477437e+00 9.75349724e-01
-9.27964076e-02 -7.30642909e-03 -3.64435792e-01 9.13898796e-02
6.38923883e-01 -9.39705610e-01 -1.78195417e-01 4.96969223e-01
-2.83080004e-02 9.34958160e-01 4.14714098e-01 -9.72649574e-01
-1.57337642e+00 -8.87884974e-01 1.46659625e+00 -2.23213166e-01
6.94949150e-01 1.21305995e-01 -1.15838480e+00 8.51710916e-01
3.25455487e-01 4.87355083e-01 9.90432680e-01 -1.28720865e-01
-4.52017993e-01 -4.81095552e-01 -1.29228759e+00 1.45879546e-02
1.06756270e+00 -7.65687943e-01 -9.80408311e-01 2.16472715e-01
8.08842540e-01 -1.98091969e-01 -1.76819515e+00 6.13862097e-01
5.12135863e-01 -3.57656986e-01 1.01909542e+00 -1.10738957e+00
-5.38205624e-01 -7.83885002e-01 -5.34304321e-01 -8.77296209e-01
1.38135135e-01 -3.14087182e-01 -1.34411409e-01 1.29538691e+00
7.58756757e-01 -7.61019111e-01 6.03338718e-01 1.14121461e+00
8.80478024e-02 -1.05585948e-01 -8.38475287e-01 -9.22983706e-01
-5.23550034e-01 -1.44988939e-01 1.13489938e+00 1.13317943e+00
5.30151248e-01 4.34309959e-01 4.62961763e-01 5.54583549e-01
4.31673676e-01 3.78421426e-01 5.84013522e-01 -1.95420265e+00
3.49655688e-01 -1.95519887e-02 -7.72686958e-01 -2.28157729e-01
6.79569766e-02 -1.18085885e+00 -6.83907270e-01 -2.19409275e+00
-2.49112666e-01 -7.53978252e-01 -1.95977554e-01 5.46641231e-01
2.52614677e-01 -3.06760579e-01 -5.29968739e-02 2.31941730e-01
-9.65940833e-01 1.96774647e-01 8.22168112e-01 6.68029338e-02
-1.42268792e-01 -5.43050885e-01 -6.71755552e-01 3.55001032e-01
6.55644655e-01 -6.06902182e-01 -4.60591733e-01 -4.67528492e-01
1.25990868e+00 4.30105031e-02 3.36872578e-01 -9.77171600e-01
8.37748885e-01 -1.79695427e-01 -2.62825847e-01 -4.31886524e-01
-1.62722468e-02 -1.33992672e+00 8.06249261e-01 2.61867046e-01
-1.48684487e-01 1.01668119e-01 1.20950878e-01 2.07738325e-01
-7.65867352e-01 -3.79796833e-01 3.39758396e-01 -3.81826729e-01
-1.39064693e+00 2.48915732e-01 2.34679013e-01 5.24020970e-01
1.30497670e+00 3.46084386e-02 -4.78620261e-01 2.23913610e-01
-8.42718244e-01 6.05362773e-01 7.53789306e-01 5.50166190e-01
3.79885435e-01 -1.46338153e+00 -1.86101615e-01 5.86682484e-02
6.61470890e-01 -3.32153961e-02 -7.98712000e-02 2.83942819e-01
-8.44453692e-01 8.94562840e-01 -4.18060452e-01 -8.54470581e-02
-1.04636955e+00 8.17632020e-01 2.27608338e-01 -1.62189454e-01
-2.42472798e-01 1.15069985e-01 -6.83209896e-01 -7.28296280e-01
6.52229488e-02 1.23896763e-01 -4.02437150e-01 1.49405211e-01
3.69949996e-01 5.40240824e-01 2.51030505e-01 -6.34357572e-01
-6.53620780e-01 5.38974404e-01 4.21507567e-01 2.14322023e-02
1.18441486e+00 -1.94091141e-01 -8.60649049e-01 1.19001582e-01
1.01459026e+00 6.54292330e-02 2.46582597e-01 -6.39906943e-01
7.88307607e-01 -4.48014349e-01 -1.19889870e-01 -6.94128692e-01
-6.38484836e-01 9.61883813e-02 6.60588980e-01 6.03795290e-01
9.11218643e-01 5.60950518e-01 7.07186818e-01 9.41884756e-01
7.85791516e-01 -1.24637163e+00 -6.48173988e-01 5.00341773e-01
7.47456431e-01 -9.92617190e-01 6.94169626e-02 -6.75572991e-01
-2.81993836e-01 8.38799059e-01 3.88454974e-01 5.38576603e-01
9.32362020e-01 -4.52096239e-02 1.35872766e-01 -1.08965540e+00
-4.47203338e-01 -7.14075923e-01 2.85520613e-01 4.99207616e-01
4.17206854e-01 -2.48705462e-01 -6.70646250e-01 3.08709532e-01
-1.40675306e-01 8.17779660e-01 -2.04799756e-01 1.46385968e+00
-6.21784687e-01 -1.46983778e+00 -9.35432985e-02 2.47552931e-01
-4.83057171e-01 -1.55687153e-01 -3.19874823e-01 7.92627335e-01
2.12363034e-01 6.80159092e-01 1.98755488e-01 1.65378973e-01
8.82276297e-01 4.47252095e-01 2.80163229e-01 -9.01175499e-01
-3.94380927e-01 -7.30533600e-01 3.88224065e-01 -6.44275486e-01
-7.33764172e-01 -1.90541357e-01 -1.87095118e+00 -1.67589802e-02
-2.38844559e-01 9.51069474e-01 9.14216876e-01 7.40724623e-01
8.45619321e-01 1.18593469e-01 -1.56089440e-01 5.19506574e-01
9.30769295e-02 -5.58460712e-01 -4.51081991e-01 5.35613954e-01
-4.14200693e-01 -5.65421999e-01 3.24270189e-01 2.62685239e-01] | [9.261083602905273, 8.049744606018066] |
6993ee16-365a-4fce-be05-58e0c50716fd | vision-language-models-can-identify | 2306.10159 | null | https://arxiv.org/abs/2306.10159v2 | https://arxiv.org/pdf/2306.10159v2.pdf | Vision-Language Models can Identify Distracted Driver Behavior from Naturalistic Videos | Recognizing the activities, causing distraction, in real-world driving scenarios is critical for ensuring the safety and reliability of both drivers and pedestrians on the roadways. Conventional computer vision techniques are typically data-intensive and require a large volume of annotated training data to detect and classify various distracted driving behaviors, thereby limiting their efficiency and scalability. We aim to develop a generalized framework that showcases robust performance with access to limited or no annotated training data. Recently, vision-language models have offered large-scale visual-textual pretraining that can be adapted to task-specific learning like distracted driving activity recognition. Vision-language pretraining models, such as CLIP, have shown significant promise in learning natural language-guided visual representations. This paper proposes a CLIP-based driver activity recognition approach that identifies driver distraction from naturalistic driving images and videos. CLIP's vision embedding offers zero-shot transfer and task-based finetuning, which can classify distracted activities from driving video data. Our results show that this framework offers state-of-the-art performance on zero-shot transfer and video-based CLIP for predicting the driver's state on two public datasets. We propose both frame-based and video-based frameworks developed on top of the CLIP's visual representation for distracted driving detection and classification task and report the results. | ['Mohammed Shaiqur Rahman', 'Soumik Sarkar', 'Anuj Sharma', 'Chinmay Hegde', 'Senem Velipasalar', 'Ameya Joshi', 'Jiyang Wang', 'Jiajing Chen', 'Md Zahid Hasan'] | 2023-06-16 | null | null | null | null | ['activity-recognition'] | ['computer-vision'] | [ 1.74582526e-01 -2.06334114e-01 -5.43795526e-01 -5.71924925e-01
-9.33137059e-01 -2.84787089e-01 8.05029333e-01 -4.28923190e-01
-5.88826060e-01 3.52082610e-01 4.19958174e-01 -5.49693346e-01
2.52207398e-01 -2.57477462e-01 -7.05101550e-01 -5.65380633e-01
2.41754577e-01 1.19469455e-02 3.92540932e-01 -3.00449640e-01
1.66976720e-01 4.40947652e-01 -2.29649138e+00 3.25826526e-01
5.25622427e-01 9.23808217e-01 1.94107339e-01 9.57592785e-01
-1.25050813e-01 1.36546874e+00 -4.87198323e-01 -8.60968307e-02
1.40469104e-01 -1.72847867e-01 -3.06818634e-01 1.97978884e-01
9.50703144e-01 -4.16856289e-01 -1.03399670e+00 6.23716056e-01
3.95981401e-01 2.77082533e-01 4.77704257e-01 -1.77756405e+00
-7.64216661e-01 -3.49592030e-01 -3.53738129e-01 9.59060311e-01
3.35410893e-01 7.96885490e-01 3.78887832e-01 -9.17612374e-01
3.46485347e-01 1.10951340e+00 3.90790641e-01 8.25404227e-01
-8.93151999e-01 -7.01096356e-01 1.55799732e-01 9.77710664e-01
-1.13964736e+00 -8.42188060e-01 7.15045631e-01 -8.27326775e-01
1.45953929e+00 1.48187384e-01 6.49728417e-01 1.33063984e+00
3.31107110e-01 1.03257906e+00 6.33424640e-01 -1.74018651e-01
-1.30498093e-02 2.22539604e-01 4.22496378e-01 5.91988921e-01
-7.13249221e-02 7.18097687e-01 -6.80556476e-01 3.45730305e-01
2.76027042e-02 2.05793619e-01 -7.42704496e-02 -2.52703011e-01
-1.00484574e+00 9.04044271e-01 2.93362617e-01 -3.53513472e-02
-2.26319298e-01 3.47926617e-01 7.72145689e-01 3.26101631e-01
5.56755304e-01 -3.76143567e-02 -1.23957442e-02 -6.71185374e-01
-8.59961092e-01 1.18219562e-01 4.04999614e-01 1.16743350e+00
9.50825810e-01 4.05036241e-01 -8.89922857e-01 8.59699845e-01
-2.06217378e-01 8.00518692e-01 4.93165880e-01 -9.08429742e-01
3.16735327e-01 2.26020724e-01 -2.85476958e-03 -7.86487937e-01
-2.55569369e-01 -6.37082979e-02 -2.98742503e-01 4.11024243e-01
-5.81057929e-03 -3.47495452e-03 -1.19935167e+00 1.59424543e+00
-1.52627021e-01 6.92740679e-01 1.59408692e-02 1.07985842e+00
7.74714768e-01 7.04673469e-01 3.38465124e-01 3.93007062e-02
1.32973564e+00 -1.30871224e+00 -8.59156907e-01 -7.24774182e-01
7.36432433e-01 -4.52494562e-01 1.23012412e+00 1.30180940e-01
-9.70644832e-01 -1.14084458e+00 -9.21571851e-01 -3.23744386e-01
-5.71936727e-01 -9.57690105e-02 4.30874288e-01 6.94789410e-01
-1.00503719e+00 -4.79385942e-01 -6.61505282e-01 -3.22449535e-01
7.68336952e-01 2.22420786e-02 -3.91240239e-01 -5.98636150e-01
-1.09929657e+00 1.27503061e+00 9.95353702e-03 -5.26140817e-02
-1.51613939e+00 -1.01907241e+00 -1.19903481e+00 6.03785478e-02
5.49489737e-01 -5.32811046e-01 1.34538817e+00 -9.70400989e-01
-1.07905543e+00 1.04494953e+00 -4.78527486e-01 -9.49260712e-01
4.17542100e-01 -1.80594698e-01 -8.89527261e-01 5.79704456e-02
2.68226445e-01 1.05491447e+00 1.21134615e+00 -8.10403109e-01
-1.05872321e+00 -4.27720547e-02 8.82137716e-02 -4.11955230e-02
-2.32457444e-01 1.02196649e-01 -6.12928450e-01 -7.00067654e-02
-1.00082541e+00 -9.67097104e-01 1.38662755e-01 3.97125334e-02
1.98507160e-01 -4.23723102e-01 1.42822850e+00 -4.07783538e-01
1.04802084e+00 -2.47218895e+00 -3.81729394e-01 -3.46822321e-01
3.66545886e-01 8.22873533e-01 -5.13862073e-01 8.03322345e-02
1.37407437e-01 -5.22774935e-01 1.08065136e-01 -1.34287447e-01
-9.65996645e-03 5.17728984e-01 -2.20617011e-01 5.70698023e-01
7.33300090e-01 1.19923902e+00 -9.03593719e-01 -3.46151114e-01
8.19077432e-01 6.30636513e-01 -5.48635542e-01 4.86745149e-01
1.53477013e-01 2.54990011e-01 1.52304275e-02 4.84234959e-01
5.89341462e-01 4.84217882e-01 -7.53064811e-01 8.05401360e-04
-4.58896637e-01 -2.38865986e-02 -4.19495285e-01 1.49500549e+00
-7.39628255e-01 1.56544185e+00 -1.27994165e-01 -1.12901151e+00
7.44955778e-01 4.15824838e-02 3.28179359e-01 -1.30082512e+00
5.34735844e-02 -3.37579727e-01 4.22586612e-02 -1.06350255e+00
5.06586015e-01 1.31811559e-01 -2.76820034e-01 -1.84320882e-02
1.60619110e-01 3.82730551e-02 3.81908774e-01 2.06542864e-01
1.39831066e+00 -1.59077004e-01 -3.12040392e-02 4.86038113e-03
8.11080575e-01 2.14084685e-01 2.75424361e-01 7.02246606e-01
-9.08680320e-01 2.97899783e-01 2.31034368e-01 -5.21839082e-01
-8.50380778e-01 -1.10207176e+00 2.18349248e-02 1.48831475e+00
-1.57182831e-02 -1.00636400e-01 -6.61594093e-01 -4.93308455e-01
2.01236248e-01 1.05409122e+00 -8.17444384e-01 -8.46445918e-01
-3.83563131e-01 1.94795337e-02 5.50267518e-01 8.77649009e-01
3.44359964e-01 -1.05185139e+00 -8.51880133e-01 8.87904130e-03
-1.07713126e-01 -1.49513292e+00 -8.60027015e-01 3.54016423e-02
1.48893818e-01 -1.12169170e+00 -3.79362017e-01 -8.32258344e-01
3.93429041e-01 1.16883159e+00 1.06567574e+00 -1.24887988e-01
-7.59128153e-01 1.04605198e+00 1.28882732e-02 -9.84895051e-01
-4.15918320e-01 -2.55281359e-01 5.83339557e-02 3.17293286e-01
1.27914095e+00 -5.78718185e-02 -6.75148547e-01 4.18096900e-01
-7.19086468e-01 -2.86279410e-01 6.64302349e-01 6.79773390e-01
2.74424314e-01 -3.72945279e-01 6.55369639e-01 -3.38810295e-01
7.22459733e-01 -5.69014847e-01 -5.62826455e-01 -1.04952618e-01
-3.80818188e-01 -2.44921729e-01 4.70321774e-01 -5.23388505e-01
-9.49444354e-01 1.70605443e-02 -3.27309221e-02 -1.16989851e+00
-6.28960788e-01 1.23128451e-01 4.11714055e-02 -4.22322042e-02
8.00614893e-01 3.96056324e-01 2.71982521e-01 5.65565713e-02
6.51140332e-01 1.00793397e+00 1.07039690e+00 8.69593993e-02
5.91472685e-01 4.06580746e-01 -2.84865230e-01 -1.20270979e+00
-8.99044275e-01 -1.18783343e+00 -5.83911955e-01 -6.25066638e-01
1.09776187e+00 -1.33860767e+00 -7.41789162e-01 1.92040175e-01
-8.97522330e-01 -3.20832223e-01 -3.65584761e-01 5.11296988e-01
-7.87094295e-01 5.37323691e-02 -8.52056295e-02 -8.14927042e-01
7.95743838e-02 -1.03023934e+00 1.03843379e+00 -1.37449633e-02
4.20236886e-02 -7.16796517e-01 2.01014623e-01 8.17335546e-01
6.15784883e-01 -4.17467169e-02 5.46565294e-01 -3.03968370e-01
-5.10968804e-01 -4.94868070e-01 -4.39021230e-01 7.48465359e-01
7.31054321e-02 -1.88025638e-01 -1.38498378e+00 -2.85032809e-01
-3.01262528e-01 -4.05394763e-01 1.15295863e+00 5.99142790e-01
1.25376642e+00 2.38629952e-02 -4.39170450e-01 4.14655983e-01
9.84346151e-01 2.11686417e-01 1.00311971e+00 -4.50190306e-02
7.41007745e-01 6.25039220e-01 8.09822381e-01 1.77138373e-01
4.66289401e-01 6.39326453e-01 4.00104344e-01 -2.41924092e-01
-6.30812228e-01 -2.62988269e-01 7.57690072e-01 5.23738921e-01
1.13982320e-01 -1.19678967e-01 -9.20305252e-01 9.62495923e-01
-1.80037034e+00 -1.56849563e+00 8.33286792e-02 1.94983697e+00
2.71970838e-01 3.32760483e-01 4.51943189e-01 -1.06741048e-01
4.17514354e-01 1.89141646e-01 -7.81908453e-01 -8.86047125e-01
2.45165005e-01 -1.21761151e-01 6.89098120e-01 5.21985412e-01
-1.02627873e+00 1.05218959e+00 6.53047609e+00 6.72999620e-01
-1.23188210e+00 4.73431706e-01 2.06299707e-01 -6.61677301e-01
1.07434735e-01 -5.79728425e-01 -9.41903889e-01 6.30041182e-01
1.78145158e+00 -4.29566175e-01 3.66854578e-01 1.02750099e+00
7.58184254e-01 -1.95847869e-01 -1.05324554e+00 1.24403286e+00
6.72517121e-01 -1.22941351e+00 -1.89567223e-01 9.13976058e-02
3.63903701e-01 4.85603064e-01 2.24489361e-01 1.05342913e+00
-7.16772377e-02 -1.04899049e+00 6.07986867e-01 6.70076191e-01
9.54073429e-01 -9.18771088e-01 5.49315929e-01 3.93131047e-01
-1.01487780e+00 -5.62833905e-01 -5.08069396e-01 -1.38202354e-01
1.72812954e-01 9.04488564e-02 -9.71034765e-01 -1.46534979e-01
6.60683215e-01 1.02766168e+00 -6.81561947e-01 7.90019155e-01
1.93687946e-01 8.06998372e-01 3.24659705e-01 7.44192004e-02
6.39171422e-01 -6.34059124e-03 4.28428650e-01 1.50032532e+00
5.30495457e-02 -2.46779159e-01 3.32335204e-01 7.46956944e-01
8.80188197e-02 -3.69003624e-01 -1.24434841e+00 3.00713144e-02
1.18163310e-01 1.06295466e+00 1.78263679e-01 -3.63501817e-01
-1.16702783e+00 6.61917329e-01 3.00500274e-01 3.18335384e-01
-1.27583432e+00 -3.10040325e-01 1.48162365e+00 4.72530454e-01
4.34552401e-01 -3.51027817e-01 2.72535145e-01 -7.22004950e-01
-2.04620138e-01 -5.14627337e-01 1.73803598e-01 -8.85209024e-01
-1.07090020e+00 4.63023692e-01 1.67582497e-01 -1.49643040e+00
-4.41864818e-01 -8.70795667e-01 -5.74966550e-01 7.51900733e-01
-2.13462138e+00 -1.18241870e+00 -7.91527390e-01 9.67655778e-01
1.16253889e+00 -6.15482748e-01 4.24902767e-01 5.49944520e-01
-6.39571786e-01 6.68925464e-01 -2.41454795e-01 1.26038134e-01
9.05060291e-01 -8.46267521e-01 2.84967989e-01 1.00522637e+00
-5.46044447e-02 -1.20648377e-01 6.20157361e-01 -1.21635891e-01
-1.62683380e+00 -1.64233077e+00 7.02812016e-01 -8.85840654e-01
5.42842507e-01 -5.29224575e-01 -8.46226156e-01 6.73506320e-01
3.81093651e-01 4.94627088e-01 7.68492460e-01 -1.54773295e-01
-3.93890053e-01 -3.35835278e-01 -8.95374417e-01 5.17135918e-01
1.23480904e+00 -9.81297016e-01 -7.97623992e-01 3.77894938e-01
5.07069051e-01 -4.18351702e-02 -2.57139325e-01 -1.08739100e-01
3.41376007e-01 -8.61186743e-01 1.00373828e+00 -9.56765771e-01
2.34089717e-01 -2.44533002e-01 -6.94298558e-03 -1.30466509e+00
-6.57599866e-01 -4.05540973e-01 -2.49800563e-01 7.93924034e-01
6.07684515e-02 -5.60754120e-01 5.15825868e-01 6.27544463e-01
-7.63302624e-01 -2.15316549e-01 -9.88167524e-01 -9.85834539e-01
-7.03650042e-02 -9.25520957e-01 2.74359167e-01 4.38173413e-01
-1.19471714e-01 4.12275404e-01 -5.29466212e-01 -1.97967272e-02
2.37269729e-01 -3.78499568e-01 1.14422286e+00 -9.41100121e-01
4.11679447e-01 -3.32501590e-01 -1.15270519e+00 -8.89466703e-01
8.47172856e-01 -9.23619390e-01 3.42911839e-01 -1.28269184e+00
8.24073404e-02 2.76376039e-01 -5.30298889e-01 5.28780699e-01
1.66691504e-02 3.38640630e-01 1.04029283e-01 -6.90479726e-02
-7.07794368e-01 4.45523381e-01 1.05903161e+00 -6.23719931e-01
-9.31677073e-02 -1.12691693e-01 -6.42340541e-01 3.03058267e-01
5.63022435e-01 -2.06453636e-01 -8.48775804e-01 -3.76286209e-01
-4.67678100e-01 -1.37424588e-01 7.83904493e-01 -1.25135291e+00
2.92782307e-01 -2.56088018e-01 -1.51435751e-03 -6.08695924e-01
6.33134842e-01 -7.46326029e-01 -4.72744554e-01 4.36120957e-01
-5.09284139e-01 -1.08057626e-01 6.71361208e-01 9.69149768e-01
-4.22378182e-01 3.08647603e-01 8.05391014e-01 1.99617222e-01
-1.53058314e+00 4.54651266e-01 -1.04779255e+00 1.97143629e-01
1.41945028e+00 -4.24700946e-01 -6.55426025e-01 -6.27604604e-01
-3.94060791e-01 3.81670207e-01 -2.92348210e-02 1.26722717e+00
8.55473697e-01 -1.31349826e+00 -9.37251329e-01 6.12169743e-01
9.76948500e-01 -7.18932986e-01 5.29743314e-01 9.47364271e-01
-6.11239523e-02 1.07516170e+00 -6.22425020e-01 -1.00405073e+00
-1.34870148e+00 1.14001679e+00 3.42659056e-01 4.73444104e-01
-6.63762450e-01 5.19159198e-01 2.93954611e-01 1.83029622e-01
1.02975450e-01 -3.86736333e-01 -1.22570336e-01 7.74116861e-03
9.35334623e-01 4.23400372e-01 2.33257294e-01 -1.11886168e+00
-4.29231822e-01 2.27357224e-01 -1.35177076e-01 1.85422346e-01
7.02211499e-01 -3.19348454e-01 7.21892834e-01 3.66923302e-01
1.45176840e+00 -4.76515770e-01 -1.68093252e+00 -1.97591826e-01
-3.17336142e-01 -9.00387585e-01 7.08928764e-01 -5.94197571e-01
-1.02349222e+00 1.33051658e+00 1.13549173e+00 -5.40170446e-02
9.43023503e-01 9.31865796e-02 9.52512026e-01 3.27570081e-01
1.49993554e-01 -1.20093274e+00 1.36167854e-01 3.27115834e-01
8.23277354e-01 -1.69021821e+00 -6.85126901e-01 -1.08255886e-01
-1.06390560e+00 6.79650664e-01 8.45981777e-01 1.24817774e-01
5.77375829e-01 3.20704043e-01 4.00045812e-01 -1.26395226e-01
-1.24808240e+00 -8.26373041e-01 6.31660640e-01 1.22802234e+00
7.21331872e-03 6.62129298e-02 2.58139670e-01 1.76301271e-01
4.50541601e-02 1.74972445e-01 5.32907903e-01 9.22082782e-01
-6.68102443e-01 -4.51307684e-01 -1.37038425e-01 5.98765552e-01
1.12782612e-01 1.58403888e-01 -1.29710302e-01 5.94168425e-01
2.72896886e-01 1.30976915e+00 5.99242926e-01 -2.52885818e-01
7.22072244e-01 3.81295294e-01 3.46184403e-01 -4.93637532e-01
-2.16351926e-01 -4.31499958e-01 1.56430423e-01 -9.10288692e-01
-5.04229605e-01 -7.12732017e-01 -6.87647641e-01 -3.32256526e-01
2.06053689e-01 -2.90604144e-01 5.25867224e-01 8.94859731e-01
7.31027007e-01 7.09340990e-01 6.51458204e-01 -8.97920668e-01
-1.95680842e-01 -8.04456294e-01 -3.49765569e-01 5.99612176e-01
8.55981767e-01 -1.07234132e+00 -3.11076373e-01 2.48642027e-01] | [7.63455867767334, -0.09925885498523712] |
0177a644-0a30-4614-8d08-7fc718d978a5 | unsupervised-detection-of-anomalous-sound | 1810.09133 | null | http://arxiv.org/abs/1810.09133v1 | http://arxiv.org/pdf/1810.09133v1.pdf | Unsupervised Detection of Anomalous Sound based on Deep Learning and the Neyman-Pearson Lemma | This paper proposes a novel optimization principle and its implementation for
unsupervised anomaly detection in sound (ADS) using an autoencoder (AE). The
goal of unsupervised-ADS is to detect unknown anomalous sound without training
data of anomalous sound. Use of an AE as a normal model is a state-of-the-art
technique for unsupervised-ADS. To decrease the false positive rate (FPR), the
AE is trained to minimize the reconstruction error of normal sounds and the
anomaly score is calculated as the reconstruction error of the observed sound.
Unfortunately, since this training procedure does not take into account the
anomaly score for anomalous sounds, the true positive rate (TPR) does not
necessarily increase. In this study, we define an objective function based on
the Neyman-Pearson lemma by considering ADS as a statistical hypothesis test.
The proposed objective function trains the AE to maximize the TPR under an
arbitrary low FPR condition. To calculate the TPR in the objective function, we
consider that the set of anomalous sounds is the complementary set of normal
sounds and simulate anomalous sounds by using a rejection sampling algorithm.
Through experiments using synthetic data, we found that the proposed method
improved the performance measures of ADS under low FPR conditions. In addition,
we confirmed that the proposed method could detect anomalous sounds in real
environments. | ['Hisashi Uematsum Yuta Kawachi', 'Shoichiro Saito', 'Yuma Koizumi', 'Noboru Harada'] | 2018-10-22 | null | null | null | null | ['unsupervised-anomaly-detection-in-sound'] | ['methodology'] | [ 1.94753423e-01 -1.51369765e-01 6.89978421e-01 -1.50449112e-01
-2.02369094e-01 -1.08409725e-01 1.85792018e-02 6.00063019e-02
-4.46342021e-01 4.17970508e-01 -3.43079180e-01 -2.48390600e-01
-2.18169153e-01 -1.00684237e+00 -4.71675456e-01 -9.89246488e-01
-1.36365548e-01 -2.11280286e-02 3.59522969e-01 -7.30810920e-03
3.90888780e-01 5.62781155e-01 -1.96658003e+00 7.36556947e-02
8.80885839e-01 1.24106491e+00 -1.73475668e-02 8.54413331e-01
-1.80173293e-01 5.74249148e-01 -1.26221144e+00 -3.34141739e-02
4.79904681e-01 -7.64482200e-01 -3.67330968e-01 4.38502524e-03
8.64030793e-02 -1.84305459e-01 1.46952331e-01 1.45668209e+00
5.97227037e-01 5.33674896e-01 6.02913499e-01 -1.34045994e+00
-1.86497569e-01 2.92562753e-01 -2.97932267e-01 5.61525524e-01
2.51785755e-01 -5.28551079e-02 8.14253330e-01 -8.44554782e-01
-8.18734318e-02 7.73393810e-01 4.61772352e-01 4.46542621e-01
-8.78721535e-01 -6.80124104e-01 -1.00165717e-01 6.20162264e-02
-1.53255486e+00 -8.69024843e-02 9.89794672e-01 -2.93760896e-01
5.90703487e-01 4.84156251e-01 6.69980109e-01 5.91382682e-01
2.54383236e-01 3.51473987e-01 7.98248112e-01 -7.08844066e-01
5.56353688e-01 2.30556592e-01 -4.38229069e-02 5.14025271e-01
3.39197487e-01 2.96365499e-01 -8.63305181e-02 -3.32498193e-01
4.94918317e-01 -2.00648103e-02 -8.90583545e-02 3.12683254e-01
-6.03257418e-01 7.40274012e-01 2.34980229e-02 6.39805079e-01
-5.87649286e-01 -3.03438932e-01 2.25352496e-01 3.92243892e-01
2.56625772e-01 5.23768365e-01 -2.46519670e-01 2.08003391e-02
-7.45156705e-01 1.04772802e-02 9.83588934e-01 3.54023486e-01
4.26995516e-01 6.69247627e-01 1.22763775e-01 8.75497937e-01
4.01337028e-01 6.70890033e-01 6.35416925e-01 -5.72452784e-01
3.97603437e-02 4.44786102e-01 -5.28136175e-03 -1.22457814e+00
-9.67827216e-02 -8.44049633e-01 -6.70098424e-01 3.04305077e-01
4.17519987e-01 -2.19334632e-01 -5.72690308e-01 1.47512674e+00
3.88922602e-01 4.87392664e-01 2.68852234e-01 8.32154274e-01
2.82579422e-01 8.94764900e-01 -1.89436674e-01 -8.00373673e-01
7.32083976e-01 -3.54114473e-01 -8.79556119e-01 9.08530727e-02
3.49756867e-01 -8.91131043e-01 1.00658274e+00 8.17125201e-01
-6.21327341e-01 -5.48630476e-01 -1.08212924e+00 1.01961327e+00
-1.81976438e-01 8.59196559e-02 1.80667445e-01 7.17166066e-01
-4.37268287e-01 4.20367360e-01 -5.68849385e-01 -2.55810052e-01
-1.03042386e-01 2.05609202e-01 -1.21508418e-02 4.84704882e-01
-1.10112929e+00 3.70105118e-01 3.43256056e-01 2.89124906e-01
-7.70895958e-01 -3.78042310e-01 -4.06458080e-01 7.54557624e-02
2.15405256e-01 -9.92902182e-03 1.16265225e+00 -1.07820356e+00
-1.53304636e+00 3.29704762e-01 1.46579117e-01 -5.21673381e-01
2.54669756e-01 -2.21154109e-01 -1.15789950e+00 1.72028631e-01
-4.94142771e-02 -3.08861315e-01 9.63887513e-01 -1.22050273e+00
-7.21367359e-01 9.58272368e-02 -4.55145627e-01 -2.21921012e-01
-2.32809544e-01 -9.68767405e-02 3.51752073e-01 -6.14150107e-01
4.73392963e-01 -6.41714275e-01 -2.29590945e-02 -4.11446482e-01
-4.49737161e-01 -8.05188268e-02 8.19083691e-01 -3.94736648e-01
1.42431414e+00 -2.72796297e+00 -7.84671307e-01 9.46474731e-01
-1.92862302e-01 3.30847263e-01 8.07803050e-02 2.83774227e-01
-1.38194546e-01 -2.44824380e-01 -5.56313157e-01 3.38869780e-01
-2.88056761e-01 1.81532741e-01 -5.15997112e-01 3.73315275e-01
2.70227313e-01 -2.92548954e-01 -8.46754014e-01 -4.09690946e-01
1.74876526e-01 1.98421657e-01 -7.85456121e-01 5.62530696e-01
-1.10166401e-01 3.15836757e-01 -4.76641595e-01 4.27030623e-01
6.43436551e-01 2.42512032e-01 -2.98248112e-01 -3.78015041e-02
-3.75975847e-01 -2.51144171e-03 -1.73793221e+00 5.89173615e-01
-3.03570211e-01 4.68445569e-01 -1.13657400e-01 -1.36219430e+00
1.53574765e+00 4.60115612e-01 6.46762669e-01 -4.82420087e-01
2.72550702e-01 4.47730660e-01 3.30576509e-01 -7.65926957e-01
1.00362469e-02 -3.13113987e-01 2.89311528e-01 2.89566785e-01
-1.07068822e-01 6.56054961e-03 -1.86709315e-02 -1.89129099e-01
1.08763444e+00 -4.13533926e-01 4.74807382e-01 -1.63400829e-01
9.64369714e-01 -3.05631161e-01 6.66457474e-01 9.15479958e-01
-1.92683786e-02 4.19080287e-01 3.33001465e-01 -3.32726777e-01
-6.44097209e-01 -1.36841655e+00 -3.43912333e-01 7.48457253e-01
-4.72192355e-02 1.40276715e-01 -5.39685249e-01 -6.20116174e-01
-1.77331850e-01 1.09562433e+00 -1.04432367e-01 -3.99088353e-01
-4.58270967e-01 -9.40223038e-01 6.30341232e-01 1.93440970e-02
5.87604046e-01 -1.24426198e+00 -6.63388014e-01 2.87558585e-01
3.04743443e-02 -8.09138894e-01 -6.83537424e-02 3.00887704e-01
-6.11561954e-01 -9.95083451e-01 -2.40004718e-01 -7.07057297e-01
8.75496387e-01 -1.26036078e-01 7.20187426e-01 1.79777786e-01
-3.06645244e-01 3.69986236e-01 -5.18713236e-01 -7.18102634e-01
-8.36786330e-01 -6.48164213e-01 4.08954144e-01 4.53336209e-01
4.10669953e-01 -7.47204304e-01 -4.40790415e-01 3.69951099e-01
-9.36307609e-01 -8.29381883e-01 4.18759942e-01 6.71837926e-01
7.39153028e-01 7.32181549e-01 9.01485384e-01 -5.38114011e-01
7.36783266e-01 -4.42605048e-01 -7.44595051e-01 -8.48859251e-02
-5.16978264e-01 4.45296057e-02 1.02471232e+00 -7.09689796e-01
-9.20887828e-01 -2.08398104e-01 -4.51963842e-01 -3.00061852e-01
-4.52020764e-01 1.93763733e-01 -1.06207386e-01 4.43657339e-02
7.45563090e-01 4.86060619e-01 -8.84298887e-03 -4.76210594e-01
-4.19427782e-01 9.46537137e-01 5.57302654e-01 -2.57983714e-01
7.32740819e-01 7.16866255e-02 2.10644573e-01 -1.22844779e+00
-8.19475472e-01 -4.69893932e-01 -1.51280597e-01 -4.50366348e-01
5.64734995e-01 -3.43211323e-01 -8.15160334e-01 3.01228791e-01
-8.78342390e-01 4.01628822e-01 -4.77646053e-01 1.05157816e+00
-1.30956665e-01 4.56439495e-01 -2.28944048e-02 -1.52951312e+00
-3.86682749e-01 -7.41664171e-01 3.06495637e-01 1.75319493e-01
-9.34191346e-02 -6.08173609e-01 1.65400803e-01 -3.09117705e-01
3.29722077e-01 2.83100754e-01 7.51000464e-01 -1.46347618e+00
-1.44943908e-01 -6.36140108e-01 4.18440878e-01 1.06703806e+00
2.37816349e-01 4.31981564e-01 -9.75091755e-01 1.63317882e-02
6.84705019e-01 2.90628254e-01 3.95901471e-01 2.50090122e-01
1.23693836e+00 -6.16425097e-01 2.41254538e-01 3.09453011e-01
1.36918592e+00 9.02209103e-01 6.47699058e-01 2.99621314e-01
3.13028812e-01 2.54701376e-01 7.39611983e-01 7.37778664e-01
-4.81574029e-01 2.65101045e-01 4.87608820e-01 2.21062854e-01
3.58890742e-01 -2.14805566e-02 4.11049724e-01 1.05768073e+00
9.12135243e-02 -4.65593874e-01 -6.26234174e-01 3.76796842e-01
-1.39714670e+00 -1.28122973e+00 -1.58963144e-01 2.72437716e+00
5.41190445e-01 4.87101108e-01 7.22102225e-02 9.64450359e-01
8.56264174e-01 -1.93402544e-01 -1.77501038e-01 -7.07085013e-01
9.35202837e-02 2.72753984e-01 -8.72088373e-02 3.25088590e-01
-9.26656187e-01 2.16030777e-01 5.79832458e+00 5.23814321e-01
-1.34905756e+00 -2.53209800e-01 7.34238029e-02 4.00812626e-02
1.78973116e-02 -1.94469556e-01 -5.55151641e-01 7.71450996e-01
7.99107492e-01 4.87345792e-02 2.98274517e-01 1.02026904e+00
2.65550286e-01 -1.67809844e-01 -6.44527376e-01 6.69977367e-01
1.54202180e-02 -3.93393755e-01 6.48199767e-02 -1.98588490e-01
5.19709826e-01 -4.65028793e-01 -4.24709618e-02 2.25587189e-01
-3.33174855e-01 -5.29204965e-01 4.21017319e-01 7.89944351e-01
2.21075922e-01 -9.27078187e-01 1.12581825e+00 4.47078943e-01
-8.47489893e-01 -2.19600499e-01 -3.25796127e-01 -1.69633165e-01
-1.61516160e-01 1.03195882e+00 -1.01087213e+00 1.90391988e-01
6.30664706e-01 4.73869145e-02 -1.90451562e-01 1.28757489e+00
9.28565580e-03 9.59900856e-01 -7.74992645e-01 -3.26902717e-01
7.90150389e-02 -2.44229570e-01 1.10822642e+00 1.01864243e+00
7.03453720e-01 1.28084391e-01 -5.17394161e-03 7.80990601e-01
3.82561862e-01 4.26714361e-01 -5.40804327e-01 -4.39780988e-02
7.22498894e-01 7.52070904e-01 -5.96380889e-01 -3.34578529e-02
-2.08302751e-01 5.06316662e-01 -4.19364333e-01 2.99075484e-01
-6.59110904e-01 -7.83386171e-01 1.64822623e-01 6.22397624e-02
2.48845994e-01 1.38542235e-01 -1.33765891e-01 -6.57924533e-01
1.11239307e-01 -7.94974685e-01 5.62972486e-01 -4.03481185e-01
-1.38655126e+00 6.95047081e-01 -2.59134751e-02 -1.79521906e+00
-3.01179618e-01 -3.81789178e-01 -1.11737823e+00 5.46543360e-01
-7.27477610e-01 -1.03121012e-01 -2.10208222e-01 5.97608030e-01
2.73154676e-01 -5.40881157e-01 8.56274664e-01 2.93684512e-01
-5.62097132e-01 6.48546398e-01 9.69944149e-02 1.04015335e-01
3.71394753e-01 -1.03296638e+00 -3.09924960e-01 1.20571291e+00
1.60564601e-01 3.81235987e-01 1.28148985e+00 -6.01662815e-01
-6.71498597e-01 -8.28426719e-01 5.80666363e-01 1.65546447e-01
4.65724200e-01 3.73826250e-02 -1.13043773e+00 1.24937259e-01
-4.01359260e-01 2.82682925e-01 7.41015613e-01 -2.78675646e-01
1.39529079e-01 -4.95597720e-01 -1.44859087e+00 3.57342809e-01
3.81313682e-01 -2.18350813e-01 -7.67672122e-01 1.15590423e-01
4.76201922e-01 4.06695642e-02 -7.56158769e-01 5.49213767e-01
4.26538944e-01 -1.16597891e+00 6.13060832e-01 -1.10269383e-01
6.47240952e-02 -6.32538736e-01 -2.73827195e-01 -1.18820286e+00
-1.51939283e-03 -2.49912962e-01 -2.25325465e-01 1.30682302e+00
3.14366311e-01 -8.96119714e-01 3.89069885e-01 5.62078729e-02
-2.08696246e-01 -7.70745337e-01 -9.87286985e-01 -9.15257752e-01
-6.50815666e-01 -6.40300095e-01 4.77730632e-01 6.61173761e-01
-2.17247605e-01 -1.37219936e-01 -1.08804002e-01 8.03406358e-01
5.80061495e-01 -1.57456785e-01 5.07227421e-01 -1.35869277e+00
-4.17797446e-01 -6.93316385e-02 -7.15647399e-01 -5.15156269e-01
-1.46981791e-01 -4.45505708e-01 3.36794913e-01 -9.67633843e-01
-5.45142174e-01 -2.47753635e-01 -8.01087916e-01 -5.26312133e-03
-8.02298635e-02 3.70490067e-02 -1.80399209e-01 -1.33028105e-02
-6.76527023e-02 5.35896122e-01 9.16836619e-01 1.99676335e-01
-5.31024516e-01 6.97197199e-01 -1.28673613e-01 9.42308664e-01
1.01264262e+00 -9.15161192e-01 -3.96888316e-01 1.34061679e-01
2.59388033e-02 -2.23615747e-02 3.04739147e-01 -1.45121121e+00
1.42930314e-01 -1.71363696e-01 3.36562485e-01 -7.41649151e-01
7.36018736e-03 -1.13803363e+00 -1.91432890e-02 5.30661345e-01
-2.45935872e-01 2.14070175e-02 3.99740599e-02 6.63767159e-01
-3.84693772e-01 -6.14094496e-01 7.10036099e-01 1.38862014e-01
-3.85507524e-01 -2.64094263e-01 -6.33811235e-01 -1.31714240e-01
1.02965081e+00 -2.84173280e-01 8.91721174e-02 -3.50753784e-01
-4.85050023e-01 -2.46019676e-01 -1.99834302e-01 1.04187883e-01
7.78366566e-01 -1.12019253e+00 -5.00342488e-01 7.88773179e-01
5.84971085e-02 -1.66902140e-01 1.62190676e-01 7.74792850e-01
-5.08133054e-01 -9.24159661e-02 3.58175598e-02 -5.69295585e-01
-1.15344095e+00 2.87776768e-01 4.89064068e-01 1.47536978e-01
-4.94731724e-01 7.23004639e-01 -1.09741427e-01 -2.71337688e-01
3.64163846e-01 -1.64897487e-01 -3.78602475e-01 -3.41549397e-01
4.54508424e-01 5.53190887e-01 9.74302217e-02 -3.19356501e-01
-4.10680622e-01 3.54243636e-01 2.54552126e-01 -2.26312727e-01
1.13490820e+00 1.73102632e-01 -1.84548676e-01 6.58180237e-01
1.04651499e+00 5.56698799e-01 -6.26465857e-01 8.66256878e-02
-1.48277625e-01 -4.58499283e-01 1.66228011e-01 -7.00091600e-01
-8.43588412e-01 5.55939436e-01 1.04686701e+00 7.39769042e-01
1.43835986e+00 -4.35384125e-01 6.34811282e-01 6.76071167e-01
2.54590549e-02 -1.15389812e+00 1.50165498e-01 4.82120484e-01
6.18369758e-01 -9.40512240e-01 -1.91738889e-01 -1.29024059e-01
-6.46481693e-01 1.29072690e+00 7.98590541e-01 -3.50141972e-01
9.51633871e-01 3.75823349e-01 8.36690813e-02 -9.09036864e-03
-3.98881972e-01 -3.77022289e-02 3.10427994e-01 3.54307860e-01
1.81955621e-01 -7.08311871e-02 -4.68014956e-01 8.71114612e-01
-5.06726265e-01 -2.33753353e-01 4.81121808e-01 8.94303858e-01
-8.62597048e-01 -4.95585710e-01 -7.11325943e-01 6.23041213e-01
-8.30965996e-01 1.04455024e-01 1.23794172e-02 6.03172183e-01
3.09248179e-01 1.26659155e+00 4.55203772e-01 -6.94802940e-01
5.68823040e-01 2.59041458e-01 -8.75601694e-02 -4.06988978e-01
-3.29834223e-01 1.14827633e-01 -2.37006754e-01 -1.71391368e-01
-1.30960733e-01 -3.95011336e-01 -1.36861265e+00 -8.78756680e-03
-6.25963628e-01 6.41754091e-01 6.19746685e-01 1.06712413e+00
-3.19239236e-02 6.26459539e-01 1.33463538e+00 -4.16463576e-02
-8.06969523e-01 -9.25017595e-01 -8.36825252e-01 2.80099183e-01
4.64852810e-01 -4.63793129e-01 -1.22990978e+00 -1.41196609e-01] | [7.575644493103027, 2.4618382453918457] |
56f7b7a0-8b0d-4959-ad94-1890ac73aff5 | darkness-can-not-drive-out-darkness | null | null | https://aclanthology.org/2022.acl-srw.4 | https://aclanthology.org/2022.acl-srw.4.pdf | Darkness can not drive out darkness: Investigating Bias in Hate SpeechDetection Models | It has become crucial to develop tools for automated hate speech and abuse detection. These tools would help to stop the bullies and the haters and provide a safer environment for individuals especially from marginalized groups to freely express themselves. However, recent research shows that machine learning models are biased and they might make the right decisions for the wrong reasons. In this thesis, I set out to understand the performance of hate speech and abuse detection models and the different biases that could influence them. I show that hate speech and abuse detection models are not only subject to social bias but also to other types of bias that have not been explored before. Finally, I investigate the causal effect of the social and intersectional bias on the performance and unfairness of hate speech detection models. | ['Fatma Elsafoury'] | null | null | null | null | acl-2022-5 | ['abuse-detection'] | ['natural-language-processing'] | [-1.85043767e-01 4.13895547e-02 -1.21361211e-01 -2.23090038e-01
2.86154449e-01 -5.65075338e-01 6.83458984e-01 -2.99800956e-03
-2.74802774e-01 7.34741509e-01 4.85594004e-01 -4.55330729e-01
3.38721760e-02 -4.54188079e-01 -9.78315948e-05 -7.17761040e-01
3.04435164e-01 -9.92954820e-02 -1.67657793e-01 -1.33184448e-01
6.93643153e-01 4.03333068e-01 -1.20230949e+00 -9.94783361e-04
8.24699759e-01 -3.48335989e-02 -2.97898918e-01 6.13578260e-01
1.25892624e-01 1.26128590e+00 -1.19912648e+00 -7.68224478e-01
-1.61106557e-01 -4.33410257e-01 -5.21308422e-01 -1.33895338e-01
5.74962080e-01 -8.86272192e-01 -2.68070549e-01 1.34154797e+00
6.65361643e-01 -3.59645449e-02 7.61855125e-01 -1.40604973e+00
-9.58482563e-01 7.12593198e-01 -4.85591650e-01 3.81221712e-01
4.87564474e-01 3.39754790e-01 3.81172180e-01 -4.23772007e-01
6.72036886e-01 1.72755492e+00 4.75371540e-01 9.74212050e-01
-1.16095972e+00 -1.39407623e+00 -2.89870501e-01 -9.22424868e-02
-1.08175218e+00 -7.86659658e-01 6.56873047e-01 -9.26691353e-01
6.22649372e-01 1.55549139e-01 8.06267381e-01 1.55042303e+00
1.84690997e-01 3.70714247e-01 1.42327094e+00 -1.78796813e-01
1.85519978e-02 6.96043432e-01 4.76194590e-01 5.73728025e-01
7.79109597e-01 2.40178242e-01 -4.33843583e-01 -6.05662882e-01
2.45510295e-01 1.27068637e-02 -1.00886822e-01 1.20483011e-01
-5.94612241e-01 1.49026585e+00 2.98073262e-01 6.89039469e-01
-1.09076068e-01 -1.40242174e-01 4.17430103e-01 1.23693999e-02
6.60896778e-01 8.50223958e-01 1.60464033e-01 -3.72349560e-01
-9.47903812e-01 4.39783543e-01 8.15720975e-01 1.73304919e-02
2.05613106e-01 1.45510539e-01 -1.15120225e-01 7.44801939e-01
4.07712817e-01 7.36950457e-01 1.54948518e-01 -8.46852839e-01
3.17391366e-01 3.71213078e-01 1.25967249e-01 -1.62098598e+00
-2.12338030e-01 -2.74620384e-01 -4.66778815e-01 6.57225132e-01
6.39293969e-01 -5.32416403e-01 -7.60783792e-01 1.61738455e+00
1.82016373e-01 -4.08090800e-01 -4.56289619e-01 9.13233399e-01
3.68788779e-01 6.42668843e-01 5.30799091e-01 -1.90363705e-01
9.00906324e-01 -4.96530801e-01 -1.20389295e+00 -3.90734702e-01
1.17429674e+00 -7.39013314e-01 7.70994067e-01 1.88471735e-01
-8.95910263e-01 1.13218213e-02 -7.21318185e-01 1.17088124e-01
-4.53747422e-01 -4.34650868e-01 6.76403999e-01 1.43153834e+00
-5.77812612e-01 6.30913973e-01 -4.14674699e-01 -4.02092785e-01
6.48703277e-01 6.58577308e-02 -2.38707379e-01 1.16490580e-01
-1.10667574e+00 1.53150713e+00 -8.03068131e-02 1.21349178e-01
-8.18095744e-01 -4.19635683e-01 -6.68058157e-01 -1.38541043e-01
2.87548780e-01 -1.93077117e-01 7.81737387e-01 -1.41753590e+00
-7.75926888e-01 9.15275872e-01 2.87270062e-02 -3.57281953e-01
4.14878756e-01 -3.87336791e-01 -1.97103858e-01 -9.92912352e-02
2.94486552e-01 4.50002074e-01 1.04099381e+00 -1.14989161e+00
-7.00315908e-02 -8.23978662e-01 -1.17678724e-01 -2.07839832e-02
-4.83559132e-01 7.94749677e-01 1.03012252e+00 -7.26351082e-01
-5.35634220e-01 -1.08938503e+00 3.45766008e-01 -2.02656955e-01
-6.06683195e-01 -2.57127821e-01 1.33338809e+00 -8.61207426e-01
1.41981125e+00 -2.26072979e+00 -1.61925480e-01 -1.40606547e-02
3.93940836e-01 9.53826308e-01 4.28602427e-01 4.88454223e-01
6.92453459e-02 8.18305314e-01 1.42266437e-01 -2.82499194e-01
8.82690176e-02 2.88656235e-01 -2.21260935e-01 8.33065152e-01
-6.84148818e-02 3.10363442e-01 -7.38377988e-01 -3.71685594e-01
2.10435554e-01 7.38282144e-01 -4.76983577e-01 3.15787971e-01
4.80815202e-01 1.93474129e-01 -8.58395025e-02 4.42493379e-01
6.43226743e-01 4.63390112e-01 -5.25626838e-02 5.49577773e-01
-1.51258826e-01 6.04380429e-01 -3.56863379e-01 4.00162309e-01
1.21098407e-01 1.16815329e+00 5.23482800e-01 -4.21071261e-01
9.44693267e-01 2.17080027e-01 -2.85916209e-01 -3.22926074e-01
5.07922292e-01 2.67446250e-01 5.95522165e-01 -6.42628789e-01
5.45250237e-01 -5.96232653e-01 8.84298980e-02 6.38420463e-01
-2.73289055e-01 -8.61108899e-02 -2.62749135e-01 2.75794148e-01
7.40000129e-01 -4.37977761e-01 1.12386920e-01 -2.53250629e-01
1.33189827e-01 -3.07880819e-01 6.45347238e-01 6.96846604e-01
-8.16571116e-01 -2.49195263e-01 6.53340340e-01 -2.66469151e-01
-9.44128096e-01 -7.41717637e-01 9.39208195e-02 1.26009393e+00
-1.81366310e-01 -8.68492275e-02 -9.94696140e-01 -7.90936172e-01
8.79954249e-02 1.43654239e+00 -6.65686667e-01 -6.25862479e-01
-1.32123837e-02 -5.13046622e-01 9.40116227e-01 1.07381336e-01
3.12318236e-01 -8.06788564e-01 -6.73393846e-01 -4.66523618e-01
7.18556717e-02 -8.15867007e-01 -1.49903625e-01 -2.67988771e-01
-5.20122647e-01 -9.46886837e-01 -5.57311356e-01 -2.41753355e-01
5.86055636e-01 5.63802123e-01 3.02061260e-01 7.90906489e-01
-6.16729669e-02 -4.13088873e-02 -5.10645568e-01 -8.88442218e-01
-8.52422416e-01 4.77572605e-02 1.45531714e-01 -2.46979803e-01
6.41727865e-01 -2.82474786e-01 -1.48903504e-01 2.86421776e-01
-7.40979552e-01 -9.36410278e-02 6.81664273e-02 4.29637522e-01
-8.36303890e-01 2.97424886e-02 3.79197568e-01 -1.00360572e+00
8.81035984e-01 -6.46955073e-01 -1.04022101e-01 -6.36372417e-02
-6.22563601e-01 -5.41634381e-01 1.85991243e-01 -5.74017465e-01
-1.13133121e+00 -6.25001252e-01 1.11140229e-01 -9.05270651e-02
-4.57665175e-01 -8.68821070e-02 1.06138133e-01 -2.11233795e-01
6.43694937e-01 -3.64192098e-01 3.25752437e-01 -4.74659294e-01
-2.04773962e-01 9.12822843e-01 -4.10374016e-01 -2.30570540e-01
9.55915987e-01 2.69020110e-01 -2.40447491e-01 -1.47747719e+00
-9.36178684e-01 -6.56915158e-02 -3.90375912e-01 -4.58388060e-01
1.11284208e+00 -5.54565012e-01 -3.71417254e-01 7.26151407e-01
-1.24789166e+00 -3.54347289e-01 6.62710965e-01 5.45802355e-01
1.71103418e-01 4.19396698e-01 -6.76912308e-01 -1.53899157e+00
-2.28000790e-01 -9.92277145e-01 4.11243588e-01 3.56817365e-01
-8.77166748e-01 -8.79312813e-01 4.30683531e-02 8.67670953e-01
6.39959395e-01 3.13239336e-01 8.48153591e-01 -1.04461741e+00
6.99647963e-02 8.20797160e-02 -8.40817317e-02 4.74308789e-01
2.08143160e-01 6.76069319e-01 -9.35415506e-01 -1.50999144e-01
2.11680695e-01 -4.12624836e-01 4.86958772e-01 1.50462583e-01
2.67016411e-01 -8.40349019e-01 -3.12675416e-01 1.69605032e-01
8.46556067e-01 4.34483677e-01 5.98254800e-01 3.28676581e-01
6.34358108e-01 1.21573985e+00 4.96595591e-01 2.07371429e-01
-1.03639252e-01 5.47013700e-01 2.59181112e-01 1.15458921e-01
1.72592685e-01 -3.86141002e-01 6.52535915e-01 3.78378570e-01
6.50985539e-02 -7.91952536e-02 -9.64259684e-01 4.74853337e-01
-1.28818333e+00 -1.31243277e+00 -3.61431658e-01 1.97490430e+00
3.84847969e-01 7.24748969e-02 5.28913200e-01 2.77606905e-01
8.85249972e-01 3.96314800e-01 -8.20106268e-03 -1.21888578e+00
-4.05800790e-02 -1.90704897e-01 3.72648895e-01 7.77693152e-01
-8.72696280e-01 1.02593660e+00 7.23300219e+00 4.13488269e-01
-9.83595848e-01 1.59047738e-01 6.12331331e-01 -3.60275745e-01
-2.51674354e-01 2.39231456e-02 -5.69175899e-01 6.66874945e-01
8.14074814e-01 -1.37432128e-01 4.26528156e-01 8.13415647e-01
4.87877280e-01 -4.31856483e-01 -4.93930578e-01 5.63276291e-01
3.08050573e-01 -7.00301826e-01 -3.14891726e-01 4.75793868e-01
2.59692430e-01 -4.23530966e-01 1.46272883e-01 1.70548975e-01
4.06803131e-01 -1.20175195e+00 6.56928360e-01 -4.70706373e-02
-1.60119370e-01 -5.93140006e-01 8.46100986e-01 4.15182710e-01
3.31489965e-02 -2.92537957e-01 -2.36352503e-01 -8.47484350e-01
8.78269598e-02 2.64264464e-01 -1.00291562e+00 -5.92636347e-01
4.68859792e-01 2.74465948e-01 -6.34310544e-01 4.69207615e-01
-5.70768774e-01 9.64017272e-01 3.56420912e-02 -3.98161769e-01
4.44182128e-01 -1.10994928e-01 7.65558302e-01 1.19244528e+00
-1.19747981e-01 -6.50042892e-02 -6.86547101e-01 9.13844228e-01
4.32631582e-01 -7.80661702e-02 -1.35939801e+00 -6.49467349e-01
6.19894743e-01 9.33540821e-01 -5.69837451e-01 3.25825554e-03
-1.74276963e-01 6.65228426e-01 2.67000288e-01 1.85628191e-01
-7.90084004e-01 -3.06318641e-01 1.10365045e+00 5.14900863e-01
-2.84891635e-01 -4.12921727e-01 -1.05272993e-01 -9.13861990e-01
-6.11456096e-01 -1.21021223e+00 1.45097196e-01 -6.98896468e-01
-1.13119543e+00 -1.71290323e-01 5.37106246e-02 -9.88444090e-02
-1.48404375e-01 -4.62457836e-01 -6.29629135e-01 6.33974314e-01
-7.67368078e-01 -9.11794543e-01 3.32859367e-01 1.45668268e-01
1.52248368e-01 -2.52498109e-02 5.91152489e-01 5.88166080e-02
-9.39481497e-01 3.21234077e-01 -3.60822082e-01 6.09910429e-01
6.65612876e-01 -7.04809964e-01 -1.07003659e-01 8.10283244e-01
-3.53185296e-01 9.96610105e-01 9.54085708e-01 -9.61021304e-01
-7.10861862e-01 -3.44975680e-01 1.08671367e+00 -5.06641686e-01
9.45787251e-01 -3.99903446e-01 -9.23140705e-01 6.62854373e-01
2.71988869e-01 -8.64665568e-01 1.11531854e+00 2.62241364e-01
-5.64308465e-01 3.64405274e-01 -1.35173142e+00 6.47389948e-01
8.80252421e-01 -7.59288371e-01 -8.18132401e-01 1.36739030e-01
4.80091721e-01 2.43303686e-01 -4.91782129e-01 -3.20092469e-01
6.55133188e-01 -1.30061150e+00 5.50858259e-01 -8.98344994e-01
5.47146142e-01 2.09002659e-01 9.71241742e-02 -1.13479447e+00
-5.15767634e-01 -5.64082205e-01 2.65148610e-01 1.54560637e+00
2.12887093e-01 -8.35808456e-01 5.75098038e-01 1.14257872e+00
4.51840848e-01 -1.77757069e-01 -8.79399717e-01 -5.85595310e-01
2.98374116e-01 -8.97619054e-02 2.99332201e-01 1.47169030e+00
2.12049216e-01 3.69053721e-01 -8.33117425e-01 1.81586236e-01
8.00364554e-01 -5.40353537e-01 7.96417296e-01 -9.67508256e-01
2.35445246e-01 -4.55912292e-01 -4.97273058e-01 4.48349230e-02
4.52808917e-01 -2.32093036e-01 -3.63627374e-01 -1.04464746e+00
3.28750670e-01 -1.66210875e-01 3.52232039e-01 4.04480934e-01
-2.44454667e-01 4.42608297e-02 8.30103517e-01 2.19375491e-01
-1.65504560e-01 3.11038047e-01 9.51170206e-01 7.41924345e-02
-1.21777862e-01 -1.32568911e-01 -9.38055694e-01 9.95045066e-01
1.02848434e+00 -8.73664975e-01 -2.67808348e-01 -2.73868024e-01
3.01879883e-01 -1.77829698e-01 6.64754868e-01 -7.40368903e-01
-1.77689150e-01 -6.45485222e-01 1.29713088e-01 4.63486426e-02
4.63815957e-01 -7.23292172e-01 -1.28143281e-01 7.82775164e-01
-4.20331687e-01 -1.44220561e-01 -6.57549053e-02 4.94386964e-02
2.44686663e-01 -6.19891703e-01 8.26230586e-01 2.13107578e-02
2.11921260e-01 -2.47946367e-01 -1.14994061e+00 5.65228052e-02
1.19173110e+00 -1.20466441e-01 -8.33938241e-01 -8.53110313e-01
-3.95135552e-01 -4.47520129e-02 7.66175091e-01 4.70244408e-01
3.05449724e-01 -9.85124350e-01 -4.82016265e-01 -2.12474659e-01
-3.08558643e-01 -9.20994461e-01 -6.61374629e-02 8.39332283e-01
-3.13572615e-01 3.24868649e-01 -4.42964137e-01 2.37350225e-01
-1.82545972e+00 6.12680435e-01 3.30609649e-01 2.49620125e-01
-1.04225799e-01 6.62642181e-01 1.31645948e-01 -7.61808604e-02
9.27691013e-02 4.39052284e-01 -2.37702340e-01 2.37521008e-01
8.27420175e-01 9.56371725e-01 -7.96523035e-01 -1.26284778e+00
-4.81192470e-01 1.58374440e-02 -2.05885828e-01 -6.18425058e-03
1.15067208e+00 -6.84614778e-02 -4.48008657e-01 3.15865725e-01
1.04370928e+00 3.07691246e-01 -5.57384372e-01 4.08005536e-01
3.14593725e-02 -1.28482032e+00 3.58295664e-02 -6.85295403e-01
-5.98388731e-01 1.03222835e+00 2.37953886e-01 7.71021962e-01
3.97296846e-01 -1.49413943e-01 4.64936852e-01 5.21874912e-02
2.52175182e-02 -1.25321257e+00 9.07147601e-02 9.18363482e-02
6.60354316e-01 -1.11295331e+00 1.53960168e-01 -5.94161749e-01
-8.03798735e-01 7.97356546e-01 6.46619022e-01 4.74039726e-02
2.95260429e-01 1.32082507e-01 1.23425171e-01 -4.00245011e-01
-5.24230897e-01 2.03863382e-01 -1.77726120e-01 1.20590329e+00
7.21906126e-01 4.59522188e-01 -7.23684728e-01 2.62337983e-01
-2.94769347e-01 -1.84552282e-01 8.24265420e-01 6.84375763e-01
-7.30579138e-01 -9.13840592e-01 -8.74460936e-01 2.77424008e-01
-8.52609634e-01 2.38241225e-01 -1.52233398e+00 9.77403104e-01
3.22215170e-01 1.35315073e+00 -1.69805855e-01 -7.01869309e-01
-1.65868029e-01 1.45070404e-01 1.19457088e-01 -6.56613886e-01
-9.90677476e-01 -3.59557152e-01 7.09675491e-01 -2.17216402e-01
-4.03897405e-01 -6.78444803e-01 -4.67348427e-01 -1.08718491e+00
-6.33860052e-01 5.38103804e-02 6.81805074e-01 9.99866545e-01
1.98987454e-01 -2.55754352e-01 5.00977397e-01 -5.59977472e-01
-5.37623703e-01 -1.15231192e+00 -6.53759360e-01 5.36438465e-01
3.01483661e-01 -8.49840283e-01 -8.30447435e-01 -7.85386324e-01] | [8.699915885925293, 10.49638557434082] |
465cec1b-d8e1-4009-b5b4-c40343065154 | reverse-operation-based-data-augmentation-for | 2010.01556 | null | https://arxiv.org/abs/2010.01556v2 | https://arxiv.org/pdf/2010.01556v2.pdf | Reverse Operation based Data Augmentation for Solving Math Word Problems | Automatically solving math word problems is a critical task in the field of natural language processing. Recent models have reached their performance bottleneck and require more high-quality data for training. We propose a novel data augmentation method that reverses the mathematical logic of math word problems to produce new high-quality math problems and introduce new knowledge points that can benefit learning the mathematical reasoning logic. We apply the augmented data on two SOTA math word problem solving models and compare our results with a strong data augmentation baseline. Experimental results show the effectiveness of our approach. We release our code and data at https://github.com/yiyunya/RODA. | ['Sadao Kurohashi', 'Daisuke Kawahara', 'Fei Cheng', 'Sujian Li', 'Wenyu Guan', 'Qianying Liu'] | 2020-10-04 | null | null | null | null | ['math-word-problem-solving', 'mathematical-reasoning', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'natural-language-processing', 'reasoning', 'time-series'] | [-4.35313769e-02 1.17585570e-01 -1.93881527e-01 -5.65935075e-01
-6.23415589e-01 -6.60223663e-01 3.34088206e-01 4.64968681e-01
-5.47049999e-01 4.85051006e-01 3.27136546e-01 -6.44037306e-01
-2.30247810e-01 -1.33854783e+00 -8.44894767e-01 9.57328454e-02
2.23205552e-01 6.80225194e-01 -8.13408643e-02 -5.45311213e-01
3.03100228e-01 6.94915699e-03 -9.75411713e-01 4.61469084e-01
1.19190133e+00 4.92851585e-01 -3.33583429e-02 6.97468877e-01
-5.42429507e-01 1.07713723e+00 -3.07686239e-01 -6.83133721e-01
4.95684713e-01 -6.07198887e-02 -1.35722077e+00 -5.39279878e-01
8.51683557e-01 -4.03826505e-01 -5.19472122e-01 1.15997446e+00
-1.55175868e-02 4.35658425e-01 2.64269084e-01 -1.20922780e+00
-1.31769145e+00 1.21804667e+00 -6.84913918e-02 2.48407841e-01
3.53193253e-01 2.53311187e-01 1.47216392e+00 -7.76449978e-01
3.71526301e-01 1.26465344e+00 5.96077979e-01 7.88324833e-01
-1.28207529e+00 -7.98759222e-01 2.05030754e-01 6.23862743e-01
-1.17974770e+00 -2.37167642e-01 8.52255881e-01 -2.46882364e-01
1.02878797e+00 1.76352993e-01 5.85777640e-01 6.95286453e-01
-3.08827043e-01 9.81424153e-01 9.48248088e-01 -4.03025150e-01
-1.85508132e-02 -1.93986759e-01 8.59585702e-01 1.12205708e+00
2.81849325e-01 -4.25493330e-01 -2.06227675e-01 8.22595060e-02
7.53593922e-01 -2.20456216e-02 -1.75140068e-01 -1.26112271e-02
-1.14077532e+00 9.42635477e-01 4.96482551e-01 3.52748573e-01
-1.07305340e-01 5.15302718e-01 1.55985624e-01 6.47141218e-01
3.34013253e-01 1.27244711e+00 -8.15528572e-01 -2.20670432e-01
-5.94744325e-01 6.88227117e-01 7.97820807e-01 9.91413772e-01
4.86498833e-01 2.74686068e-02 -3.22035670e-01 6.91422880e-01
5.87893091e-02 3.87316108e-01 2.67662436e-01 -1.07266533e+00
8.76517236e-01 1.05212641e+00 -2.79332787e-01 -9.64933693e-01
-3.52632016e-01 -6.93165362e-02 -5.03685892e-01 -2.77117491e-01
6.56877160e-01 1.48159981e-01 -7.73732543e-01 1.77591872e+00
8.53213817e-02 2.82955229e-01 2.09457427e-01 7.85508573e-01
1.17245436e+00 9.36779320e-01 2.52978772e-01 3.26312244e-01
1.29291177e+00 -1.08328152e+00 -8.21306109e-01 -3.35290551e-01
1.00228763e+00 -3.12628597e-01 1.71175468e+00 4.77336586e-01
-1.42481315e+00 -4.02601004e-01 -7.99462974e-01 -7.18918204e-01
-6.06042206e-01 -1.32349551e-01 1.19115782e+00 4.30797577e-01
-7.22016394e-01 5.49624383e-01 -7.91966259e-01 -2.39463467e-02
7.37960041e-01 2.16734678e-01 -8.56817439e-02 -5.04872918e-01
-1.46610427e+00 9.48267817e-01 6.86931610e-01 -2.06161097e-01
-4.11975324e-01 -1.48658752e+00 -1.23266375e+00 1.35845825e-01
6.32143497e-01 -5.93799651e-01 1.43946397e+00 -3.68707389e-01
-1.35607064e+00 8.10672343e-01 -8.15490857e-02 -4.85472053e-01
2.20244512e-01 -5.07053912e-01 -1.37823224e-01 -9.74889770e-02
-1.66258886e-01 5.64357460e-01 1.55183986e-01 -7.90557861e-01
-2.28800327e-01 -1.78698391e-01 6.30500019e-01 -1.90988332e-02
-5.61270297e-01 -2.37966850e-01 -3.20437819e-01 -8.34163010e-01
4.97698411e-02 -4.15772021e-01 -2.27710158e-01 -2.56326318e-01
-8.44589099e-02 -8.49620759e-01 1.84364785e-02 -7.66730487e-01
1.43172216e+00 -1.86730635e+00 3.36191118e-01 1.77966416e-01
5.39746702e-01 4.50332701e-01 -5.31553864e-01 1.18271112e-01
-1.80991247e-01 3.62261772e-01 -2.29883030e-01 -4.61130403e-02
3.00082982e-01 2.20648780e-01 -6.04828358e-01 -1.40056061e-03
4.62901056e-01 1.56689417e+00 -1.09907436e+00 -2.91427314e-01
1.40313357e-01 -2.20645815e-02 -1.11040330e+00 2.77698278e-01
-8.68117690e-01 -6.33829907e-02 -4.11785692e-01 5.18264294e-01
6.73910558e-01 -4.00891542e-01 6.33369088e-02 -2.53166199e-01
2.78839797e-01 8.72131407e-01 -1.14387596e+00 1.92065930e+00
-3.93763423e-01 3.33078712e-01 -4.56536949e-01 -1.02361512e+00
6.90887749e-01 -1.95420757e-02 5.89006059e-02 -7.49576807e-01
3.52198601e-01 -1.23253368e-01 1.32137537e-01 -5.78054845e-01
5.41777790e-01 -2.33018622e-01 -9.21828151e-02 5.00594854e-01
1.19248316e-01 -8.80190730e-01 5.39967299e-01 5.30802667e-01
1.17426038e+00 2.14755870e-02 1.00810692e-01 -3.11628520e-01
5.43797374e-01 2.80270278e-01 4.76196200e-01 7.49680221e-01
-3.40929888e-02 1.77128509e-01 3.94336432e-01 -5.37588239e-01
-9.44890022e-01 -1.12812304e+00 8.06079507e-02 1.19956589e+00
-2.09034652e-01 -8.90648901e-01 -5.28322697e-01 -6.13023996e-01
1.57402650e-01 1.05442166e+00 -3.75519216e-01 -3.15725416e-01
-8.84577394e-01 -4.16250706e-01 7.46952295e-01 6.18126333e-01
6.35610580e-01 -1.04495513e+00 8.46952423e-02 -5.03028790e-03
-3.96827996e-01 -1.33460224e+00 -9.94011238e-02 -2.26899728e-01
-9.22480822e-01 -1.14523864e+00 -1.17888033e-01 -8.84215415e-01
6.31029546e-01 1.01397716e-01 1.50685346e+00 6.30352199e-01
-3.80050361e-01 5.02888262e-01 -3.68974417e-01 -6.06764495e-01
-3.52842838e-01 1.81195840e-01 1.55187726e-01 -7.61788785e-01
7.18793273e-01 -4.39585626e-01 5.09470608e-03 -5.42015553e-01
-9.53872561e-01 3.61646801e-01 1.59635231e-01 4.80246246e-01
1.93848744e-01 2.38979563e-01 4.28067118e-01 -1.01368618e+00
7.91734159e-01 -4.38502342e-01 -1.02134180e+00 4.04827565e-01
-4.23643917e-01 3.21313709e-01 7.20535398e-01 -3.66001070e-01
-7.09157467e-01 -1.65378734e-01 -1.22844726e-01 -3.23190600e-01
-1.68754384e-01 8.48526955e-01 3.76538858e-02 9.51735973e-02
6.34515882e-01 -5.49326427e-02 -2.66482979e-01 -4.27793413e-01
8.70041013e-01 1.46185428e-01 5.45881629e-01 -1.29697824e+00
1.32144797e+00 5.74246459e-02 -1.24334775e-01 -6.87683761e-01
-1.43621445e+00 -1.47109076e-01 -5.06630242e-01 2.30942607e-01
7.10794747e-01 -8.80334496e-01 -9.94199395e-01 2.14859262e-01
-1.36561584e+00 -7.70662844e-01 -3.99057478e-01 4.16721910e-01
-3.54148805e-01 2.96595931e-01 -7.92104244e-01 -5.70062876e-01
-3.90274018e-01 -6.91450536e-01 4.34843153e-01 2.32859507e-01
-3.71891141e-01 -1.24288416e+00 5.70088886e-02 8.45498741e-01
3.42343956e-01 -1.85801923e-01 1.52715683e+00 -7.59379089e-01
-8.35013986e-01 -5.16463406e-02 -4.38413292e-01 3.50367546e-01
-1.32985324e-01 -2.47107029e-01 -6.95865691e-01 1.71788171e-01
-5.32285273e-02 -8.18421185e-01 9.57637727e-01 7.12904409e-02
1.71701050e+00 -5.16268492e-01 3.19891006e-01 6.55314565e-01
1.39775121e+00 -3.73491347e-01 6.35037005e-01 2.37826303e-01
8.69196534e-01 2.92790323e-01 3.89736772e-01 6.41765893e-02
8.85873914e-01 1.00564480e-01 8.38559419e-02 8.67525041e-02
-2.43333653e-02 -2.94283718e-01 2.55020291e-01 1.02198362e+00
-3.29765342e-02 -1.72066651e-02 -1.54977810e+00 8.01054239e-01
-1.73764205e+00 -8.56296122e-01 -2.53032565e-01 1.58368373e+00
1.56163001e+00 8.77950266e-02 -1.68425754e-01 2.01038063e-01
-5.54741127e-03 -1.18687823e-01 -1.45256862e-01 -5.59822202e-01
1.93302721e-01 1.05026984e+00 2.94557154e-01 9.71930087e-01
-1.10271764e+00 1.68623817e+00 5.99593401e+00 7.01939166e-01
-4.90870476e-01 -8.96722749e-02 2.27397680e-01 -1.70464188e-01
-4.39823538e-01 -2.47490272e-01 -6.00112379e-01 -1.31857514e-01
7.90945828e-01 -2.97935098e-01 9.28200662e-01 6.39222085e-01
-1.69894889e-01 -8.56135320e-03 -1.48323405e+00 9.61331725e-01
5.16570807e-02 -1.66643095e+00 4.15551901e-01 -4.13065702e-01
6.36634707e-01 -1.65009633e-01 9.10482556e-02 7.94997036e-01
8.28142285e-01 -1.43453550e+00 2.64512420e-01 5.24760544e-01
4.13796902e-01 -6.54168665e-01 5.39788485e-01 2.59727091e-01
-9.99697924e-01 -2.64629424e-02 -2.32270703e-01 -7.85760522e-01
-9.75365788e-02 4.23788369e-01 -7.98398972e-01 3.87636721e-01
3.92603129e-01 8.45417976e-01 -9.32636380e-01 5.42765081e-01
-6.91281557e-01 7.57421076e-01 -2.78224796e-01 -2.50064343e-01
2.02471033e-01 -2.00064778e-01 2.59613782e-01 1.17860460e+00
-1.85955375e-01 9.31022048e-01 2.39332855e-01 1.61717224e+00
-4.33789611e-01 -1.64330180e-03 -5.06889343e-01 -7.58568645e-01
3.88090849e-01 1.15551257e+00 -2.39139095e-01 -5.73346674e-01
-4.55056757e-01 6.40646338e-01 7.91095257e-01 2.22321868e-01
-9.13735509e-01 -3.91338885e-01 9.07007575e-01 -3.77948694e-02
-1.96193099e-01 -7.11563468e-01 -7.47528076e-01 -1.57181668e+00
4.22378071e-02 -9.80895519e-01 6.10477507e-01 -6.81380510e-01
-1.66621280e+00 -1.89444229e-01 1.40654698e-01 -3.43688458e-01
3.33421350e-01 -1.14302552e+00 -5.12612104e-01 7.81968594e-01
-1.56572461e+00 -1.08306313e+00 -3.49179238e-01 5.30700207e-01
5.40890932e-01 -2.97710478e-01 1.02987230e+00 2.42948592e-01
-4.74233717e-01 7.88491786e-01 -4.04391527e-01 7.31912971e-01
4.11685854e-01 -1.44726741e+00 7.61659622e-01 7.35461116e-01
2.66108274e-01 1.03269064e+00 4.74835813e-01 -6.74082637e-01
-1.91282785e+00 -1.03313065e+00 1.14670503e+00 -1.10251391e+00
1.10239923e+00 -5.48677206e-01 -1.19690394e+00 1.02421701e+00
1.52344003e-01 -3.15907240e-01 7.88001776e-01 4.86159831e-01
-7.23440170e-01 1.15919746e-01 -1.11912882e+00 8.66743088e-01
1.02792346e+00 -4.32686329e-01 -1.31048894e+00 5.52389205e-01
1.19503582e+00 -5.92023373e-01 -8.06997478e-01 3.57230783e-01
1.39023662e-01 -5.45120537e-02 1.18965220e+00 -1.36590254e+00
1.07202089e+00 -1.03449591e-01 -1.40881017e-01 -1.17396569e+00
-4.50217605e-01 -2.22796023e-01 -2.77424663e-01 1.08158505e+00
4.33786660e-01 -3.79315346e-01 8.40619922e-01 1.08292341e+00
1.53209969e-01 -6.02446139e-01 -4.96315688e-01 -7.65335739e-01
6.75184608e-01 -9.95632589e-01 6.25369191e-01 1.56847703e+00
5.34997523e-01 6.14330828e-01 1.20823190e-01 2.34402180e-01
4.52552229e-01 1.09393723e-01 8.67869258e-01 -1.10406792e+00
-1.58984661e-01 -5.76602280e-01 -2.71395832e-01 -9.15871024e-01
5.88015437e-01 -1.69946420e+00 -3.27288002e-01 -1.86589289e+00
1.54572412e-01 -4.40477699e-01 -2.09773734e-01 1.04634511e+00
-3.34482521e-01 1.42036572e-01 3.71298552e-01 -4.37643826e-01
-7.40794063e-01 5.76462924e-01 1.16470635e+00 -2.61858582e-01
-1.52855173e-01 -5.14611900e-01 -8.78766477e-01 9.02120054e-01
1.23138106e+00 -1.55905798e-01 -3.46181780e-01 -1.05326235e+00
7.94628441e-01 -5.18264174e-01 5.78317642e-01 -9.68998253e-01
2.28457153e-01 -6.61654353e-01 1.85116559e-01 -1.79405928e-01
2.61992604e-01 -7.93003619e-01 -8.31812561e-01 4.05165136e-01
-6.36382282e-01 1.85076222e-01 7.73691118e-01 1.25170752e-01
1.75752953e-01 -2.20201373e-01 6.32031441e-01 -3.43564183e-01
-7.40341306e-01 1.88727245e-01 -7.17951506e-02 6.78240538e-01
7.28081763e-01 4.27112252e-01 -5.86339474e-01 -1.93170115e-01
-4.43366557e-01 6.67534471e-01 -4.42718016e-03 6.25766695e-01
8.13455284e-01 -1.44437087e+00 -8.55048776e-01 1.38294414e-01
7.79264122e-02 2.90430963e-01 -4.69794497e-02 5.15927613e-01
-7.07686722e-01 5.48128843e-01 -3.81469876e-02 3.79768610e-02
-1.17448580e+00 6.15055263e-01 2.43477091e-01 -3.16540211e-01
-5.86147189e-01 1.05726564e+00 -1.57564014e-01 -9.78209794e-01
3.41637462e-01 -1.05397558e+00 -2.15133607e-01 -3.30862254e-01
6.99519038e-01 3.38568628e-01 -1.00947134e-01 1.75588787e-01
-1.40663415e-01 2.93428749e-01 -1.83696941e-01 1.00453489e-01
1.73857486e+00 4.75534618e-01 -5.70895374e-01 5.05277157e-01
8.19354951e-01 -1.20260209e-01 -2.47005299e-01 -7.57127225e-01
1.78049728e-01 -5.77106416e-01 1.71999615e-02 -1.02506018e+00
-8.44871223e-01 9.00536597e-01 1.18434452e-01 -9.71833020e-02
9.03381348e-01 1.01884820e-01 9.26066220e-01 1.09386611e+00
7.50105008e-02 -1.01375592e+00 1.94802806e-01 1.30850065e+00
8.65025938e-01 -1.41154599e+00 2.20804751e-01 -5.74931145e-01
-3.33905369e-01 1.12244713e+00 9.44573462e-01 -3.83927941e-01
5.84702730e-01 2.46796623e-01 -1.05187856e-01 -4.43898916e-01
-7.71847963e-01 -2.34679326e-01 4.85420525e-01 2.90174454e-01
7.05479980e-01 4.12812531e-02 -2.96814531e-01 1.13979375e+00
-7.15402663e-01 3.33055228e-01 4.76016313e-01 8.24471474e-01
-2.37775505e-01 -1.05277240e+00 -1.52449772e-01 3.74932200e-01
-2.01664314e-01 -7.44952321e-01 -4.92472917e-01 6.81963086e-01
-6.89879283e-02 6.92511022e-01 -1.10418331e-02 -2.45825633e-01
4.13351655e-01 6.20334387e-01 8.90640020e-01 -9.62277114e-01
-5.72935224e-01 -7.92870462e-01 -6.28051981e-02 -5.36604822e-01
1.15208942e-02 -2.62609959e-01 -1.76292193e+00 -8.50232720e-01
8.81849229e-02 4.15228605e-02 2.88580537e-01 1.05833626e+00
1.64356619e-01 7.59537518e-01 -1.61946625e-01 -1.65509224e-01
-5.89621484e-01 -7.79471099e-01 -1.98037326e-02 5.19770622e-01
1.96690321e-01 -2.38363430e-01 -1.13861062e-01 1.73438340e-01] | [9.635193824768066, 7.385912895202637] |
1b1e1c3b-5644-4b4d-93a4-c1cae9a96b85 | benchmarking-intent-detection-for-task | 2012.03929 | null | https://arxiv.org/abs/2012.03929v2 | https://arxiv.org/pdf/2012.03929v2.pdf | Benchmarking Commercial Intent Detection Services with Practice-Driven Evaluations | Intent detection is a key component of modern goal-oriented dialog systems that accomplish a user task by predicting the intent of users' text input. There are three primary challenges in designing robust and accurate intent detection models. First, typical intent detection models require a large amount of labeled data to achieve high accuracy. Unfortunately, in practical scenarios it is more common to find small, unbalanced, and noisy datasets. Secondly, even with large training data, the intent detection models can see a different distribution of test data when being deployed in the real world, leading to poor accuracy. Finally, a practical intent detection model must be computationally efficient in both training and single query inference so that it can be used continuously and re-trained frequently. We benchmark intent detection methods on a variety of datasets. Our results show that Watson Assistant's intent detection model outperforms other commercial solutions and is comparable to large pretrained language models while requiring only a fraction of computational resources and training data. Watson Assistant demonstrates a higher degree of robustness when the training and test distributions differ. | ['Mo Yu', 'Saloni Potdar', 'Ladislav Kunc', 'Abhishek Shah', 'Atin Sood', 'Lin Pan', 'Haode Qi'] | 2020-12-07 | null | https://aclanthology.org/2021.naacl-industry.38 | https://aclanthology.org/2021.naacl-industry.38.pdf | naacl-2021-4 | ['goal-oriented-dialog'] | ['natural-language-processing'] | [ 2.25181673e-02 -3.09910148e-01 -3.29497546e-01 -6.57936037e-01
-8.14456105e-01 -8.81542504e-01 5.47466576e-01 2.61551857e-01
-6.61981285e-01 5.03824592e-01 3.62892210e-01 -6.22470200e-01
2.49721974e-01 -4.57849443e-01 1.69350594e-01 8.62439498e-02
1.00291930e-01 9.99430716e-01 2.34506592e-01 -5.90960562e-01
4.38144028e-01 1.13273054e-01 -1.48362970e+00 1.79580688e-01
7.53943205e-01 8.24186921e-01 3.33989620e-01 1.07440495e+00
-2.60858238e-01 1.00960028e+00 -8.35143328e-01 -2.57669806e-01
1.40541404e-01 -1.92433983e-01 -8.08744550e-01 -1.16363041e-01
1.94772482e-01 -7.34082103e-01 -1.25867620e-01 8.42469454e-01
5.26775301e-01 4.94410157e-01 6.49528384e-01 -1.48715520e+00
-3.18973571e-01 1.73359364e-01 -5.26505634e-02 1.76492214e-01
1.06070316e+00 3.41965318e-01 1.23883057e+00 -8.81370544e-01
4.10172403e-01 1.11002934e+00 5.29251456e-01 5.78963220e-01
-1.08321953e+00 -4.67641473e-01 3.20141912e-01 -6.40733689e-02
-1.09399986e+00 -6.48077130e-01 5.14808118e-01 -3.41219366e-01
9.80514646e-01 6.40941560e-01 1.94256961e-01 1.07768118e+00
6.84864894e-02 9.74679291e-01 6.89531684e-01 -2.77946889e-01
4.67556417e-01 4.84777510e-01 6.07745767e-01 6.00950599e-01
-1.27682388e-02 -1.58019915e-01 -5.17379761e-01 -6.45671964e-01
1.46753624e-01 3.43006819e-01 -1.88255489e-01 -7.07946718e-02
-9.78413522e-01 1.02691925e+00 6.75363140e-03 2.96300054e-01
-2.59065747e-01 -2.84309804e-01 5.11397421e-01 5.76797843e-01
3.54613245e-01 6.07699156e-01 -5.42306900e-01 -6.28149688e-01
-6.67940557e-01 4.48255688e-01 1.55560386e+00 1.27513039e+00
6.09642029e-01 -2.60980248e-01 -2.26030052e-01 9.83232796e-01
5.30941784e-01 3.83323461e-01 7.50504434e-01 -6.38807178e-01
5.58505595e-01 9.02080536e-01 4.29903030e-01 -8.53019297e-01
-4.75096524e-01 -9.23475176e-02 -4.36285019e-01 3.18047777e-02
5.18790960e-01 -3.72533977e-01 -8.05170417e-01 1.51828361e+00
5.89985475e-02 -4.26869750e-01 1.00273184e-01 7.78859079e-01
7.93909490e-01 5.38220525e-01 3.66805553e-01 -1.22287065e-01
1.53145170e+00 -6.82713866e-01 -6.99984968e-01 -1.08119726e+00
9.29787993e-01 -1.00270259e+00 1.47763145e+00 2.75936246e-01
-7.71950662e-01 -3.48479569e-01 -8.14178765e-01 -2.36194268e-01
-4.77292478e-01 -4.85760197e-02 9.51977015e-01 7.38973200e-01
-8.82854462e-01 4.11472060e-02 -3.59305531e-01 -6.78092718e-01
-1.18020967e-01 3.74692708e-01 -9.48788747e-02 -1.34833500e-01
-1.13854480e+00 9.21669066e-01 3.23700964e-01 -4.07264680e-01
-5.37455022e-01 -5.68277717e-01 -1.01172888e+00 1.25189915e-01
3.52024615e-01 -3.46797794e-01 2.01729751e+00 -4.05778617e-01
-1.17773068e+00 5.89574575e-01 -4.51584518e-01 -9.81356353e-02
5.13349533e-01 -3.13345313e-01 -3.46005797e-01 -3.03750992e-01
2.11484164e-01 4.86825734e-01 6.59798682e-01 -9.51642990e-01
-8.38666022e-01 -4.13968027e-01 1.20688841e-01 6.49665415e-01
-4.07503933e-01 1.55962020e-01 -6.01505399e-01 -2.46793881e-01
-3.49358879e-02 -7.77337968e-01 -3.11225832e-01 -1.92178175e-01
-2.28899479e-01 -5.79351187e-01 1.21199799e+00 -4.21456695e-01
1.38978505e+00 -2.24487591e+00 -3.82517219e-01 -1.73564218e-02
2.91127741e-01 2.13234276e-01 4.12134491e-02 6.93956494e-01
2.94414699e-01 7.65364021e-02 1.19355440e-01 -5.39879382e-01
3.43222111e-01 2.07193717e-01 -4.81051743e-01 1.08043544e-01
-2.59825587e-01 7.76464105e-01 -9.86759126e-01 -6.01015449e-01
4.15781617e-01 -6.23448454e-02 -7.40132570e-01 6.88189149e-01
-5.23403764e-01 1.45555153e-01 -4.57742095e-01 6.51534438e-01
2.11713299e-01 -4.09852535e-01 1.74950451e-01 1.89744473e-01
-3.31203379e-02 7.53771722e-01 -1.08556449e+00 1.45693564e+00
-7.13521838e-01 6.90323591e-01 1.52718157e-01 -6.41879141e-01
7.53241777e-01 3.53384048e-01 3.48689526e-01 -5.76861024e-01
2.78373271e-01 1.01601072e-01 -6.85905740e-02 -4.66082603e-01
8.48344803e-01 -1.70783296e-01 -5.28428495e-01 1.06433737e+00
-1.14474408e-01 -3.05610657e-01 3.81949276e-01 3.32491904e-01
1.22173059e+00 -7.49268234e-01 4.20696616e-01 6.24118596e-02
2.64385849e-01 3.13283771e-01 4.03344601e-01 1.17330217e+00
-5.15335381e-01 2.85098106e-01 3.72815073e-01 -6.57979429e-01
-6.15907907e-01 -6.37562811e-01 6.18175827e-02 1.95724702e+00
1.66712791e-01 -4.28483188e-01 -3.07574034e-01 -9.25281525e-01
6.90255463e-02 1.24714136e+00 -1.69368029e-01 -1.72183260e-01
-2.72111833e-01 -5.78159869e-01 4.41986620e-01 4.57670063e-01
4.90389973e-01 -1.05246723e+00 -6.55619264e-01 2.49949962e-01
-3.44834179e-01 -9.46383238e-01 -7.73823202e-01 4.10200268e-01
-6.04803264e-01 -1.00501335e+00 -2.12276399e-01 -8.00064385e-01
5.29245913e-01 5.92415571e-01 1.38544714e+00 4.11913276e-01
-1.64181739e-01 6.02630258e-01 -3.34042251e-01 -6.08493268e-01
-5.75608194e-01 1.75175130e-01 2.69642383e-01 -4.91896629e-01
1.04445028e+00 -1.32941082e-01 -4.43398297e-01 5.28069198e-01
-7.97596037e-01 -1.74678594e-01 4.56049472e-01 8.39861214e-01
-3.04840535e-01 1.58647567e-01 4.70779568e-01 -8.39907408e-01
1.26955211e+00 -5.87353110e-01 -3.55474323e-01 3.44818860e-01
-8.25845301e-01 -8.95663202e-02 5.93441248e-01 -6.29061937e-01
-1.05951250e+00 -8.25841054e-02 -3.75929981e-01 7.75498599e-02
-4.76728499e-01 5.22485435e-01 1.19654909e-01 1.60956204e-01
8.91916156e-01 2.75868624e-01 8.85093734e-02 -3.85656238e-01
2.14982063e-01 1.20258451e+00 6.49991557e-02 -3.24527591e-01
5.28709352e-01 2.93407664e-02 -9.24232841e-01 -9.59023595e-01
-8.13316464e-01 -1.01549351e+00 -3.72816682e-01 -1.81368083e-01
4.21561033e-01 -7.32845664e-01 -8.72366190e-01 2.47539923e-01
-1.02411032e+00 -4.53337073e-01 -8.15137550e-02 3.48472297e-01
-2.95657665e-01 2.66720116e-01 -5.65543592e-01 -1.04711926e+00
-5.85628688e-01 -1.19400442e+00 1.14492178e+00 2.41795436e-01
-1.00064564e+00 -1.20531368e+00 1.59575835e-01 5.84984839e-01
5.08713007e-01 -7.88926363e-01 8.38611186e-01 -1.43981552e+00
-2.45432228e-01 -7.27424681e-01 -1.29671499e-01 -4.08727489e-02
5.25738120e-01 -3.46222371e-01 -8.00602973e-01 -4.54444349e-01
2.46728480e-01 -7.74111688e-01 3.23823571e-01 -4.75293845e-02
5.88669896e-01 -4.55136359e-01 -4.25487936e-01 -7.07576470e-03
1.01127410e+00 4.76112157e-01 1.51495799e-01 -1.02966934e-01
2.69525468e-01 4.58761156e-01 8.53308976e-01 4.89131153e-01
4.92144078e-01 5.41081011e-01 -6.44685179e-02 7.02430531e-02
4.19937611e-01 -3.36522311e-01 2.27693155e-01 4.04624641e-01
6.19141519e-01 -4.89813209e-01 -1.06708395e+00 4.52242494e-01
-1.90927863e+00 -9.28650260e-01 3.05095255e-01 2.26558447e+00
8.34935308e-01 2.66329795e-01 2.63393134e-01 -5.62456399e-02
2.67136693e-01 1.28974691e-01 -7.53665984e-01 -3.89978260e-01
4.55299407e-01 -2.99725801e-01 2.41577700e-01 7.33019173e-01
-1.04199421e+00 9.58455145e-01 7.88833904e+00 3.08790684e-01
-1.09778798e+00 -8.15451294e-02 5.73195875e-01 -1.34073813e-02
-2.45045155e-01 -1.69667646e-01 -9.52293456e-01 5.10026097e-01
8.54030550e-01 -2.35354930e-01 1.81721240e-01 1.37567019e+00
2.54020952e-02 -3.45724553e-01 -1.23205864e+00 1.04437113e+00
2.02949926e-01 -8.70510638e-01 -3.66020322e-01 -1.60272419e-01
1.00301869e-01 6.80588335e-02 -1.09173819e-01 9.52566385e-01
9.07901824e-01 -8.98237824e-01 1.53574601e-01 2.89038057e-03
5.13528228e-01 -4.36903059e-01 6.81579947e-01 8.91644120e-01
-8.81056905e-01 -2.74859250e-01 -2.14294255e-01 -2.93157816e-01
1.28836811e-01 2.64582813e-01 -1.40703845e+00 -2.66850740e-01
4.76516873e-01 3.27374160e-01 -4.80873555e-01 6.58539832e-01
1.32842720e-01 4.55670208e-01 -5.37083626e-01 -6.31804705e-01
3.45112644e-02 -2.74999570e-02 4.84941751e-01 1.07002819e+00
-2.91049276e-02 1.55493200e-01 6.58031166e-01 4.95734960e-01
-4.14831899e-02 1.07179090e-01 -1.06909823e+00 -2.78649777e-01
5.80156028e-01 1.19545162e+00 -4.94342446e-01 -3.95154297e-01
-5.91873050e-01 1.01239860e+00 3.98126930e-01 3.22222263e-01
-5.09103537e-01 -1.44562930e-01 9.14960504e-01 -1.56008184e-01
-2.42793187e-01 -3.48403454e-01 -3.14138204e-01 -1.15988457e+00
-6.90877885e-02 -1.13014853e+00 7.93509245e-01 -5.51705122e-01
-1.44713414e+00 4.59713340e-01 -1.31194919e-01 -1.05944371e+00
-7.98026741e-01 -6.91054106e-01 -7.10004687e-01 7.58763850e-01
-1.00067163e+00 -8.19332242e-01 -3.56907785e-01 4.24966156e-01
1.07527840e+00 -2.17793062e-01 1.09499288e+00 2.59151101e-01
-4.89644408e-01 5.69909334e-01 -3.60291988e-01 2.58460641e-01
9.63637114e-01 -1.38258314e+00 6.07019961e-01 5.96822798e-01
6.15407676e-02 1.04672408e+00 9.33861017e-01 -8.60688567e-01
-1.68131089e+00 -8.38470459e-01 1.05327737e+00 -8.46394539e-01
5.51392198e-01 -5.88078260e-01 -9.63588595e-01 1.03885567e+00
7.37439841e-02 -4.13645148e-01 1.02968991e+00 6.77775085e-01
-3.25402856e-01 2.91337669e-01 -1.15356016e+00 9.43221152e-01
8.41649294e-01 -7.17419088e-01 -9.05857682e-01 7.20088601e-01
5.51083028e-01 -5.29513657e-01 -5.01044810e-01 1.43498197e-01
5.49047589e-01 -8.86967301e-01 8.59313607e-01 -8.74232054e-01
-1.78145647e-01 -4.09113839e-02 -8.13016966e-02 -1.02369606e+00
-2.26438299e-01 -6.59625411e-01 -1.38182461e-01 8.73409152e-01
5.76564074e-01 -6.64401114e-01 8.75197172e-01 1.55562830e+00
3.07228982e-01 -4.35319304e-01 -5.37337244e-01 -4.42138314e-01
-3.63880098e-01 -8.21127951e-01 3.76989096e-01 7.86998451e-01
5.97498178e-01 8.20055902e-01 -3.17250609e-01 -5.47147803e-02
3.06632161e-01 1.29190862e-01 1.01725793e+00 -1.34568596e+00
8.46576318e-03 -2.43463516e-01 -1.27671301e-01 -1.55730700e+00
7.21987039e-02 -4.09737587e-01 4.53943312e-01 -1.39519382e+00
-2.85240095e-02 -8.14576209e-01 1.33200318e-01 7.01711714e-01
-5.07233977e-01 -8.39987174e-02 -1.37769923e-01 3.56159747e-01
-8.88260007e-01 4.77091998e-01 5.46215355e-01 -1.59640566e-01
-7.20647871e-01 5.06751120e-01 -9.51541245e-01 7.47784078e-01
8.61087859e-01 -2.66449124e-01 -8.11460733e-01 -1.46427572e-01
1.75672457e-01 8.85914117e-02 3.00100856e-02 -8.47522616e-01
6.16056144e-01 -3.94657880e-01 2.38137573e-01 -8.15007269e-01
4.70523864e-01 -9.35809255e-01 -3.51582587e-01 4.02919769e-01
-5.01346529e-01 2.05079153e-01 7.00117871e-02 5.45481801e-01
-1.15873404e-02 -4.14245248e-01 5.76350570e-01 -2.36153454e-01
-9.30535555e-01 1.31075054e-01 -7.72863090e-01 3.34442347e-01
9.33974981e-01 -1.26326727e-02 -3.53192776e-01 -9.29110765e-01
-1.70524865e-01 4.82496411e-01 5.94629645e-01 8.05256486e-01
6.19814277e-01 -9.10302222e-01 -2.23006964e-01 4.24672157e-01
4.72935319e-01 -1.69616848e-01 -5.93649894e-02 4.65630889e-01
-2.76470631e-01 5.92830598e-01 2.53617406e-01 -6.20089114e-01
-1.37092316e+00 4.38133746e-01 1.07152954e-01 -3.21514755e-01
-4.11129057e-01 7.01877177e-01 -1.50767863e-01 -8.63851428e-01
6.58507645e-01 -2.84473039e-02 -2.03979313e-01 7.98613951e-02
9.92970526e-01 4.61975709e-02 6.36548474e-02 -2.57690042e-01
-4.26671863e-01 -2.30624497e-01 -6.04778171e-01 -3.41999739e-01
8.23363900e-01 6.65522646e-03 1.87879577e-01 5.66179156e-01
1.07550323e+00 -2.34819036e-02 -7.60808647e-01 -3.25336307e-01
6.24344386e-02 -4.48550969e-01 -5.97742051e-02 -6.88928366e-01
-3.88716400e-01 4.86210823e-01 3.85506690e-01 6.44152701e-01
8.57347369e-01 -4.31369171e-02 7.66916513e-01 9.82268751e-01
5.62559724e-01 -1.24251807e+00 1.77710831e-01 8.07558358e-01
7.17444360e-01 -1.82013261e+00 -3.54028344e-02 -1.85492337e-01
-9.60820556e-01 7.47174561e-01 8.46803665e-01 4.55763936e-01
6.58073664e-01 4.24976468e-01 5.57352006e-01 -3.10554564e-01
-9.38596785e-01 -5.49821816e-02 3.14071923e-01 4.21481431e-01
6.56399190e-01 -1.04391113e-01 -2.45318152e-02 3.47163558e-01
-3.09145987e-01 -1.86253652e-01 1.44585848e-01 1.38665020e+00
-5.47007501e-01 -1.15613699e+00 -1.90884516e-01 8.02303910e-01
-1.29660890e-01 -1.84764221e-01 -6.09292388e-01 5.30264795e-01
-6.27125978e-01 1.48306596e+00 -4.02522907e-02 -5.72052300e-01
3.61217052e-01 5.01001298e-01 -3.15585494e-01 -1.08027136e+00
-6.71119273e-01 -4.27891463e-01 2.41890088e-01 -6.18328691e-01
1.86709389e-01 -4.35367614e-01 -1.15769386e+00 -4.53871518e-01
-3.42828453e-01 3.62173110e-01 5.95171809e-01 1.07686245e+00
4.87369895e-01 1.30982948e-02 5.78286886e-01 -4.64726388e-01
-8.51601839e-01 -1.20141172e+00 -2.74542540e-01 5.77996552e-01
3.90247166e-01 -4.66758877e-01 -6.38285697e-01 -2.12975398e-01] | [12.522051811218262, 7.704870700836182] |
8e9f1bee-6169-4f3c-a595-3279b8a902ee | towards-holistic-and-automatic-evaluation-of-1 | null | null | https://aclanthology.org/2020.acl-main.333 | https://aclanthology.org/2020.acl-main.333.pdf | Towards Holistic and Automatic Evaluation of Open-Domain Dialogue Generation | Open-domain dialogue generation has gained increasing attention in Natural Language Processing. Its evaluation requires a holistic means. Human ratings are deemed as the gold standard. As human evaluation is inefficient and costly, an automated substitute is highly desirable. In this paper, we propose holistic evaluation metrics that capture different aspects of open-domain dialogues. Our metrics consist of (1) GPT-2 based context coherence between sentences in a dialogue, (2) GPT-2 based fluency in phrasing, (3) $n$-gram based diversity in responses to augmented queries, and (4) textual-entailment-inference based logical self-consistency. The empirical validity of our metrics is demonstrated by strong correlations with human judgments. We open source the code and relevant materials. | ['Kewei Tu', 'Erik Nijkamp', 'Wenjuan Han', 'Linqi Zhou', 'Bo Pang', 'Yixian Liu'] | 2020-07-01 | null | null | null | acl-2020-6 | ['dialogue-evaluation'] | ['natural-language-processing'] | [-9.83040407e-02 3.41550887e-01 -1.11122854e-01 -6.87074959e-01
-1.02872062e+00 -8.05474997e-01 7.21855640e-01 3.17123622e-01
-3.59911114e-01 1.11565781e+00 7.24703491e-01 -3.66153032e-01
-1.24960192e-01 -7.17047989e-01 3.07021532e-02 7.84341171e-02
6.45725131e-02 5.07764399e-01 8.23246613e-02 -7.60921061e-01
6.62695944e-01 -5.00467606e-02 -1.10285008e+00 4.53120261e-01
1.47526181e+00 1.05847251e+00 -6.70634806e-02 8.58770490e-01
-3.86983246e-01 1.13481081e+00 -6.65884614e-01 -8.03741932e-01
-6.90575615e-02 -9.57124949e-01 -1.52986383e+00 -1.82514623e-01
9.47593972e-02 -2.77127653e-01 1.67858288e-01 8.55637968e-01
4.92370516e-01 2.05287352e-01 8.14470053e-01 -1.05552316e+00
-6.83608592e-01 7.30579793e-01 -7.76868612e-02 -5.65773400e-04
1.07440948e+00 4.23899382e-01 1.49975407e+00 -7.42427707e-01
5.90170026e-01 1.34931505e+00 6.26862764e-01 5.07056236e-01
-1.02825236e+00 -2.11278081e-01 -2.30106592e-01 5.17988801e-02
-1.07752335e+00 -3.33684146e-01 6.61829114e-01 -4.39180613e-01
1.05876946e+00 4.08884078e-01 4.43461090e-01 9.49219942e-01
1.11410163e-01 4.38410431e-01 1.51439857e+00 -6.02606177e-01
2.67153472e-01 5.83945036e-01 2.95812815e-01 6.15827143e-01
-4.59173396e-02 -3.29189241e-01 -5.73728919e-01 -3.10294837e-01
4.88887578e-01 -5.71837783e-01 -2.46321596e-02 3.72056007e-01
-9.70349014e-01 1.16771603e+00 5.70329577e-02 4.05224681e-01
-3.23689103e-01 -3.10805231e-01 4.60109502e-01 6.35280252e-01
3.71923923e-01 1.11392772e+00 -4.82394159e-01 -7.33775377e-01
-6.02816164e-01 6.02573693e-01 1.43327010e+00 7.93360710e-01
6.06017649e-01 -3.23084831e-01 -3.77615273e-01 1.29145837e+00
2.87465423e-01 4.31892604e-01 6.99759364e-01 -1.27538514e+00
4.06359881e-01 8.75628531e-01 2.94825792e-01 -1.07350612e+00
-3.64018470e-01 5.39223552e-02 -3.75518531e-01 -2.75073320e-01
4.24172819e-01 -3.80173355e-01 1.54673681e-01 1.98218143e+00
7.81686530e-02 -9.71902311e-01 1.12061650e-01 7.21515357e-01
8.98247719e-01 6.21860802e-01 -1.32829025e-01 -5.04700482e-01
1.45237982e+00 -7.42107570e-01 -8.15208673e-01 -1.60316542e-01
5.39980352e-01 -9.16861892e-01 1.60921478e+00 1.50212735e-01
-1.29697788e+00 -4.07697707e-01 -8.60728264e-01 -1.32814363e-01
-1.04788840e-01 -3.18358421e-01 4.51537341e-01 7.33443558e-01
-1.00764704e+00 3.25680554e-01 -5.15502430e-02 -6.09991252e-01
-2.64612794e-01 8.87227356e-02 -1.80844277e-01 2.56674349e-01
-1.46917856e+00 1.13771570e+00 1.98874086e-01 -3.37183207e-01
-3.50348443e-01 -2.52962232e-01 -7.17014134e-01 -1.13747768e-01
3.43900532e-01 -6.10276103e-01 1.75008643e+00 -5.43163180e-01
-1.92761230e+00 8.11897457e-01 -1.11901090e-01 -2.32414663e-01
5.11323810e-01 -1.60435840e-01 -3.06653529e-01 2.54744858e-01
3.26591939e-01 4.56176162e-01 4.19849634e-01 -8.93788457e-01
-4.54127967e-01 -1.04128823e-01 4.06777799e-01 5.42754471e-01
-3.43707293e-01 2.67331004e-01 2.32177079e-01 -6.01017118e-01
-1.83644786e-01 -7.55089223e-01 -4.85592410e-02 -3.38571340e-01
-3.93289357e-01 -7.15990961e-01 1.98061347e-01 -8.29285085e-01
1.70565248e+00 -1.74809504e+00 -3.00472051e-01 1.71875656e-02
1.91361591e-01 1.63950607e-01 -1.01021066e-01 9.21995878e-01
3.27634722e-01 3.44617814e-01 -1.49735481e-01 1.47398725e-01
4.01446015e-01 -5.22574224e-02 -2.36345023e-01 -1.74054697e-01
3.28536540e-01 9.17826235e-01 -1.08930075e+00 -8.42649996e-01
-1.41825750e-01 -3.03815842e-01 -7.35622346e-01 5.99696159e-01
-5.63212395e-01 1.86489806e-01 -3.56701940e-01 2.23908588e-01
7.23198131e-02 -4.33104336e-01 1.64781943e-01 1.12446517e-01
-5.75867072e-02 9.89349067e-01 -7.15982497e-01 1.58901572e+00
-4.88054425e-01 4.60725158e-01 -1.84838653e-01 -5.25606215e-01
1.06418359e+00 4.01233584e-01 -8.16310644e-02 -7.67119586e-01
8.90791230e-03 2.71747679e-01 2.76171088e-01 -8.04165959e-01
9.08780038e-01 -2.37424314e-01 -5.81467867e-01 9.84323144e-01
-1.16875626e-01 -6.10215008e-01 5.16673326e-01 4.24506694e-01
1.01100457e+00 -9.84866098e-02 8.14451039e-01 -5.25161445e-01
6.47188604e-01 2.55696714e-01 2.91043282e-01 7.38187969e-01
-2.58159041e-01 2.80633152e-01 8.82153153e-01 -7.06030950e-02
-9.28958416e-01 -9.53427017e-01 3.28461565e-02 1.12577271e+00
-9.86080319e-02 -5.01258969e-01 -9.54087079e-01 -5.23694336e-01
-2.91321814e-01 1.06096292e+00 -3.86643469e-01 -3.41253877e-02
-4.17667985e-01 -2.29174986e-01 9.19917405e-01 3.44374329e-01
6.67207003e-01 -9.68876779e-01 -5.49517691e-01 1.23091511e-01
-9.58659053e-01 -9.88387883e-01 -6.33131385e-01 -1.87030137e-01
-7.18908250e-01 -9.43739057e-01 -3.67888033e-01 -6.40824616e-01
2.11100921e-01 4.16140147e-02 1.36516595e+00 -5.61227500e-02
4.67416756e-02 2.74921596e-01 -6.03995323e-01 -2.32842609e-01
-8.60628188e-01 -1.18713323e-02 5.79421781e-02 -4.57023859e-01
3.75494719e-01 -2.69189477e-01 -5.13042569e-01 4.35787737e-01
-6.47517145e-01 -7.09845871e-02 4.32465196e-01 1.03287423e+00
-7.64460042e-02 -4.50104058e-01 8.55471671e-01 -7.81976283e-01
1.70898318e+00 -2.90010899e-01 -1.99768543e-01 3.93055081e-01
-8.43818843e-01 7.90815651e-02 7.58586586e-01 1.03919543e-01
-1.30031776e+00 -7.72396564e-01 -2.16538161e-01 4.74328756e-01
-2.02815026e-01 8.56548965e-01 1.11864261e-01 2.73205698e-01
1.09759676e+00 -2.51297615e-02 2.36368746e-01 4.24287394e-02
4.51160043e-01 1.02942657e+00 2.89707124e-01 -8.11177909e-01
4.75007713e-01 -2.37463191e-01 -4.99441475e-01 -9.74796474e-01
-8.91907334e-01 -3.26111853e-01 -2.05054149e-01 -2.85220176e-01
8.46329331e-01 -5.36701620e-01 -8.59259367e-01 1.60390958e-01
-1.30792665e+00 -2.69472450e-01 -1.94766790e-01 4.25785661e-01
-6.58486009e-01 7.53567696e-01 -8.12340736e-01 -1.00701725e+00
-5.84517717e-01 -9.68064308e-01 5.75316072e-01 2.06621706e-01
-1.19422972e+00 -9.69791532e-01 3.26048434e-01 6.61129653e-01
4.89697129e-01 8.28350633e-02 1.04229641e+00 -1.10530579e+00
-7.45435879e-02 -2.24119425e-01 -2.23240346e-01 5.31531394e-01
1.54099405e-01 -4.87645119e-02 -6.52509034e-01 3.40829134e-01
2.38590881e-01 -8.88125718e-01 5.32735549e-02 8.45227614e-02
4.79570717e-01 -7.90582001e-01 3.08582306e-01 -2.53888398e-01
1.00767767e+00 1.90797120e-01 4.79609758e-01 2.85505261e-02
6.49935007e-02 1.00126147e+00 8.22767973e-01 6.92650735e-01
5.64497650e-01 5.43621659e-01 -1.83882669e-01 4.67524111e-01
2.20475912e-01 -3.22447300e-01 5.04215300e-01 1.33142865e+00
4.74939272e-02 -2.85779446e-01 -9.40617383e-01 3.56194437e-01
-1.69796753e+00 -1.12496650e+00 -2.34695628e-01 1.90214896e+00
1.23933482e+00 1.96886420e-01 3.36724311e-01 5.31087406e-02
5.67207515e-01 1.84205130e-01 -1.85588613e-01 -1.02585340e+00
-3.22664678e-02 2.76221812e-01 -3.73156816e-01 6.75914884e-01
-4.86249149e-01 8.00599337e-01 6.95830202e+00 6.15098834e-01
-5.54210365e-01 6.32919148e-02 7.02548802e-01 8.76106024e-02
-5.53990722e-01 6.53288811e-02 -4.63448852e-01 4.20345396e-01
8.94596756e-01 -5.70274651e-01 3.32126409e-01 7.04677105e-01
1.85436219e-01 -2.93433160e-01 -1.34838355e+00 6.68093801e-01
1.23046167e-01 -8.62932861e-01 -7.90886730e-02 -2.42521670e-02
6.50735974e-01 -3.85546893e-01 -2.11499929e-01 4.69096929e-01
4.77194071e-01 -9.89621699e-01 5.12827575e-01 3.87979031e-01
7.21763730e-01 -4.73025680e-01 7.22388625e-01 4.92037803e-01
-7.36469626e-01 1.31629184e-01 -4.38136347e-02 -3.49541545e-01
1.27711341e-01 2.97844827e-01 -1.12939394e+00 4.30825949e-01
2.67134279e-01 1.18315332e-01 -3.84936154e-01 4.29617703e-01
-4.09408331e-01 5.20639956e-01 -1.22075863e-01 -8.19157481e-01
2.71810949e-01 -4.26024109e-01 3.97101641e-01 1.27200413e+00
7.47693479e-02 3.53310704e-01 -9.45686642e-03 9.54314590e-01
5.62642375e-03 4.73076612e-01 -6.40016019e-01 -4.76414323e-01
7.46885180e-01 1.12539101e+00 -4.12436843e-01 -3.07534218e-01
-4.29089367e-01 9.02672768e-01 3.46331894e-01 -1.43260295e-02
-6.81983650e-01 -6.61965668e-01 5.87401748e-01 -1.92808792e-01
-1.30195826e-01 -9.76419821e-02 -5.13772607e-01 -9.06254411e-01
1.41123876e-01 -1.22280228e+00 3.18178803e-01 -7.03024030e-01
-1.60027611e+00 8.11369598e-01 -7.88374431e-03 -1.11790156e+00
-8.47723961e-01 -5.24075747e-01 -6.45225942e-01 1.05917168e+00
-9.43266451e-01 -3.73612255e-01 -3.08365494e-01 4.08721328e-01
5.91048539e-01 -1.05115794e-01 1.08237123e+00 1.28810769e-02
-4.39155668e-01 7.69732475e-01 -4.09273833e-01 2.43884206e-01
8.69417071e-01 -1.32905185e+00 2.47080341e-01 5.24098098e-01
-3.05122465e-01 1.08292413e+00 7.90918529e-01 -5.69990039e-01
-1.11080956e+00 -4.83390331e-01 1.37747586e+00 -3.95052433e-01
8.14532638e-01 6.97411001e-02 -7.03776956e-01 1.10290445e-01
6.63509667e-01 -8.62253845e-01 1.20540535e+00 3.09973389e-01
-3.45385998e-01 1.04247211e-02 -1.11928284e+00 6.93766594e-01
7.98128784e-01 -9.32082534e-01 -1.00120676e+00 4.83099878e-01
6.89334393e-01 -2.37394035e-01 -1.12397635e+00 8.07946771e-02
6.13036156e-01 -1.17975688e+00 4.79327947e-01 -5.97546697e-01
9.26106393e-01 9.49685052e-02 -2.83185631e-01 -1.29023302e+00
-2.11006656e-01 -8.01231384e-01 2.26289034e-01 1.03809142e+00
8.08985591e-01 -6.18621945e-01 3.78171384e-01 9.75258470e-01
-1.85470477e-01 -7.38296270e-01 -5.99582434e-01 -6.48903549e-01
2.00522810e-01 -1.66629627e-01 4.31553364e-01 1.02121854e+00
1.07156491e+00 1.00335670e+00 -2.41496012e-01 -5.04187167e-01
1.02763630e-01 1.44449338e-01 8.78620625e-01 -1.15949059e+00
-5.12886822e-01 -6.98678315e-01 -3.73236719e-03 -1.21403635e+00
-2.48606410e-02 -7.11326718e-01 1.65096730e-01 -1.66305578e+00
2.22634777e-01 -1.38180286e-01 4.13985044e-01 8.95479396e-02
-3.11135560e-01 -8.69285464e-02 -2.75051836e-02 9.49781239e-02
-7.72096992e-01 6.22560084e-01 1.03885198e+00 1.64252654e-01
-1.82280377e-01 -3.10303062e-01 -1.02640021e+00 6.41970813e-01
1.07582700e+00 1.64761171e-02 -5.01781702e-01 -2.72592932e-01
5.11968374e-01 4.13097471e-01 -9.97895449e-02 -7.84805775e-01
1.75039675e-02 -4.80679840e-01 -1.16935678e-01 -2.54169166e-01
3.85808498e-01 -3.02154392e-01 -2.17270091e-01 4.21706080e-01
-7.62648284e-01 3.21074963e-01 -3.02211791e-01 2.26466253e-01
-4.36193079e-01 -6.30441248e-01 5.28686345e-01 -4.60552067e-01
-3.95325243e-01 -2.31359452e-01 -5.67813993e-01 7.26530731e-01
7.25402951e-01 -2.65780449e-01 -5.24853170e-01 -8.41908097e-01
-1.09689027e-01 2.67230511e-01 5.31984270e-01 2.82839358e-01
6.14155233e-01 -1.10653806e+00 -7.63602436e-01 -2.89499134e-01
2.91450590e-01 -3.65027368e-01 -1.52786970e-01 7.66732931e-01
-6.50191665e-01 7.60022461e-01 -2.92643011e-02 -3.10165912e-01
-1.09816277e+00 5.32981055e-03 4.57467251e-02 -5.27482331e-01
1.11909717e-01 8.30949843e-01 -5.83680235e-02 -5.28138578e-01
-2.75343321e-02 -2.10062444e-01 -1.84857562e-01 -2.46040262e-02
4.55225050e-01 5.50805449e-01 -8.24041367e-02 -4.73293155e-01
-8.13591182e-02 2.80404221e-02 -1.34746969e-01 -6.90254390e-01
7.00851858e-01 -2.62203455e-01 -4.73948658e-01 7.12727487e-01
1.09242225e+00 1.17467027e-02 -4.46688831e-01 -3.31144810e-01
2.01654240e-01 -3.61547142e-01 -4.99331981e-01 -9.56281543e-01
-4.43434790e-02 5.31791389e-01 -7.47248679e-02 7.60353982e-01
7.74169147e-01 -1.16505206e-01 8.80685687e-01 7.86583543e-01
4.72432941e-01 -1.45537233e+00 4.40730065e-01 1.02786732e+00
9.89749491e-01 -1.27365911e+00 -9.61840302e-02 -3.56272757e-01
-1.06655157e+00 9.65390325e-01 7.99845159e-01 3.06584239e-01
2.06286266e-01 -1.17176034e-01 2.32355624e-01 -1.59847438e-01
-1.05858767e+00 1.07210964e-01 5.50962947e-02 2.22324669e-01
1.10656524e+00 1.23400837e-01 -1.02123845e+00 7.08807111e-01
-7.85836756e-01 -2.02043831e-01 4.68845755e-01 8.99445236e-01
-5.20811439e-01 -1.08402658e+00 -2.37653419e-01 6.95646942e-01
-3.43601286e-01 -1.76229179e-01 -9.05238926e-01 4.58706647e-01
-5.08318305e-01 1.59217811e+00 -3.21703941e-01 -5.16389012e-01
2.64983177e-01 4.00633603e-01 5.14301658e-01 -7.66217828e-01
-8.10064197e-01 -3.05952519e-01 7.29841709e-01 -4.16084379e-01
-3.87419403e-01 -6.50663257e-01 -1.02034914e+00 -4.70842510e-01
-3.80490035e-01 7.17526257e-01 3.84393811e-01 8.74392271e-01
3.59350681e-01 3.48202959e-02 7.58393586e-01 -3.27938870e-02
-1.18048227e+00 -1.44271302e+00 -3.76858801e-01 4.82516646e-01
-1.22150503e-01 -2.92579502e-01 -2.37762138e-01 -5.33179082e-02] | [12.759734153747559, 8.174698829650879] |
65ae247b-2890-43b8-935e-f558ce15d965 | proposalcontrast-unsupervised-pre-training | 2207.12654 | null | https://arxiv.org/abs/2207.12654v2 | https://arxiv.org/pdf/2207.12654v2.pdf | ProposalContrast: Unsupervised Pre-training for LiDAR-based 3D Object Detection | Existing approaches for unsupervised point cloud pre-training are constrained to either scene-level or point/voxel-level instance discrimination. Scene-level methods tend to lose local details that are crucial for recognizing the road objects, while point/voxel-level methods inherently suffer from limited receptive field that is incapable of perceiving large objects or context environments. Considering region-level representations are more suitable for 3D object detection, we devise a new unsupervised point cloud pre-training framework, called ProposalContrast, that learns robust 3D representations by contrasting region proposals. Specifically, with an exhaustive set of region proposals sampled from each point cloud, geometric point relations within each proposal are modeled for creating expressive proposal representations. To better accommodate 3D detection properties, ProposalContrast optimizes with both inter-cluster and inter-proposal separation, i.e., sharpening the discriminativeness of proposal representations across semantic classes and object instances. The generalizability and transferability of ProposalContrast are verified on various 3D detectors (i.e., PV-RCNN, CenterPoint, PointPillars and PointRCNN) and datasets (i.e., KITTI, Waymo and ONCE). | ['Wenguan Wang', 'Jianbing Shen', 'Cheng-Zhong Xu', 'Jin Fang', 'Liangjun Zhang', 'Dingfu Zhou', 'Junbo Yin'] | 2022-07-26 | null | null | null | null | ['point-cloud-pre-training', 'unsupervised-pre-training'] | ['computer-vision', 'methodology'] | [-2.10707113e-02 -2.57368162e-02 -1.96225107e-01 -4.49028403e-01
-4.98708963e-01 -6.19860709e-01 8.28180492e-01 5.76238871e-01
-2.63321549e-01 -3.00613921e-02 -2.69397646e-01 -1.55237436e-01
-2.21203640e-01 -9.07452166e-01 -7.47234285e-01 -4.35376823e-01
-4.59149852e-03 6.00158274e-01 7.56859958e-01 -2.59107724e-02
6.02406442e-01 1.27058017e+00 -1.97848880e+00 2.65457511e-01
7.35417426e-01 9.35202122e-01 4.38844651e-01 1.66733325e-01
-4.20866787e-01 3.71346101e-02 -4.24951941e-01 1.41175777e-01
4.85388637e-01 3.44308674e-01 -3.12680513e-01 2.18677863e-01
8.32119584e-01 -2.21082807e-01 -6.37754351e-02 9.72659767e-01
1.10938944e-01 2.71022379e-01 1.13455582e+00 -1.25803065e+00
-4.96843249e-01 9.92670804e-02 -7.70256519e-01 1.85808003e-01
1.49442792e-01 1.72734290e-01 1.06289315e+00 -1.48062718e+00
4.53543216e-01 1.50252032e+00 5.95216334e-01 1.96236297e-01
-1.37862968e+00 -6.31857574e-01 5.23680747e-01 -6.03100285e-03
-1.80020261e+00 -2.58618653e-01 1.15619683e+00 -5.69800079e-01
1.08608222e+00 3.35989892e-01 5.23874223e-01 9.09693420e-01
1.73631478e-02 8.06715488e-01 1.14661682e+00 -2.48356909e-01
4.13749635e-01 1.92743510e-01 2.35744417e-01 2.16982171e-01
4.16155696e-01 2.33490020e-01 -3.92678171e-01 -5.13070412e-02
1.26160502e+00 4.02743876e-01 6.98059723e-02 -1.00165975e+00
-1.08787131e+00 6.83177650e-01 8.67465317e-01 1.01184078e-01
-3.17515075e-01 -1.12468131e-01 1.85564131e-01 -1.08118646e-01
3.49373817e-01 2.25251004e-01 -1.94871441e-01 5.09198010e-01
-1.00556290e+00 3.87716442e-01 1.24498256e-01 1.35274363e+00
1.18023121e+00 -1.56050650e-02 -2.17613414e-01 7.47354329e-01
5.83857775e-01 5.81961513e-01 1.49840266e-01 -5.97284257e-01
4.32734281e-01 1.07660699e+00 -9.91214216e-02 -1.14467144e+00
-3.79478782e-01 -6.21778011e-01 -6.62746608e-01 5.59294641e-01
1.16611384e-01 8.19032967e-01 -1.25673485e+00 1.22149658e+00
6.03457212e-01 1.53365389e-01 -8.04970935e-02 1.06978261e+00
1.20075619e+00 6.08933628e-01 1.35541767e-01 1.82313472e-01
1.43973923e+00 -6.36018217e-01 1.37901947e-01 -3.35260004e-01
3.50487590e-01 -4.76330698e-01 1.16796589e+00 1.46297827e-01
-1.05175936e+00 -9.11681354e-01 -8.64441693e-01 -3.28666002e-01
-5.75520217e-01 7.59511590e-02 4.96678352e-01 3.53446215e-01
-7.70470798e-01 1.82628542e-01 -6.88107848e-01 -2.84269363e-01
6.29609704e-01 3.11689645e-01 -4.58979994e-01 -6.41680732e-02
-5.36998451e-01 8.16490471e-01 5.61320662e-01 -9.03097093e-02
-9.49757457e-01 -7.86068916e-01 -7.42284894e-01 2.14923933e-01
5.09217381e-02 -5.49945295e-01 7.79568732e-01 -6.23145163e-01
-1.07906854e+00 1.18202579e+00 -4.29138839e-02 -2.60731757e-01
3.57259452e-01 -9.90527961e-03 -2.32914880e-01 2.48884663e-01
3.50908637e-01 1.11613822e+00 1.05563927e+00 -1.75745261e+00
-7.19250023e-01 -6.15940452e-01 1.55144837e-02 4.56425428e-01
6.10180534e-02 -2.26454511e-01 -4.11563635e-01 -5.04723191e-01
9.40191329e-01 -6.45486236e-01 -2.19464436e-01 2.20109373e-01
-4.78871614e-01 -5.75770438e-01 9.67575014e-01 -3.50022106e-04
5.38128018e-01 -2.47532415e+00 -2.59981930e-01 5.30611217e-01
2.16359004e-01 8.95858780e-02 -1.85656145e-01 3.03006709e-01
-1.39200985e-01 1.92851275e-02 -1.98264420e-01 -3.15427482e-01
1.37701944e-01 2.48174056e-01 -3.45979601e-01 5.45442879e-01
5.68705022e-01 7.13890910e-01 -6.98351383e-01 -6.25202894e-01
7.38093376e-01 5.24977446e-01 -7.38277912e-01 -4.53393236e-02
-1.42634094e-01 3.17650437e-01 -7.45580256e-01 9.69035208e-01
9.71907794e-01 -1.35011777e-01 -4.33197141e-01 -2.34911770e-01
-3.50356132e-01 3.02343041e-01 -1.31688702e+00 1.59335303e+00
-1.63506091e-01 4.05635744e-01 -6.78956360e-02 -9.09161627e-01
1.36373854e+00 -6.47812784e-02 2.95033276e-01 -6.31264985e-01
-2.26777226e-01 1.56666502e-01 -2.12913290e-01 -6.72320351e-02
6.94581091e-01 -9.94818434e-02 -7.94256013e-03 3.56608555e-02
3.49000730e-02 -4.48266953e-01 -3.30677003e-01 4.27392460e-02
7.09404945e-01 1.31343424e-01 3.97984833e-01 -4.09609199e-01
3.21083069e-01 1.64040402e-01 4.30004090e-01 9.68735874e-01
-2.35230505e-01 9.69860494e-01 6.09272458e-02 -3.33881199e-01
-9.46129143e-01 -1.77730215e+00 -6.67810023e-01 9.21541274e-01
5.85658133e-01 -1.79114401e-01 -2.42934868e-01 -6.12124741e-01
3.39307755e-01 8.17485452e-01 -4.15211260e-01 -1.06016882e-01
-3.99287760e-01 -1.22672796e-01 2.44204119e-01 7.51300454e-01
2.42989421e-01 -8.76144528e-01 -8.69399011e-01 -6.22079298e-02
2.80850947e-01 -1.04210627e+00 -3.79453376e-02 3.00654501e-01
-1.22593451e+00 -8.50884259e-01 -4.60100293e-01 -7.73975194e-01
1.02661741e+00 8.11481416e-01 1.13918340e+00 -1.07215084e-01
-2.97680825e-01 5.70329666e-01 -4.51354206e-01 -5.29632926e-01
8.21004957e-02 -1.00097187e-01 1.41694278e-01 -1.99109197e-01
4.78451252e-01 -7.00978518e-01 -6.94466650e-01 4.51549500e-01
-6.66218281e-01 -2.91566011e-02 9.90019977e-01 5.17307937e-01
1.10536969e+00 -1.27928138e-01 2.01284379e-01 -5.03784060e-01
9.40523520e-02 -3.65979522e-01 -4.79258746e-01 1.74334161e-02
-1.36948213e-01 -1.89781219e-01 2.39029661e-01 -4.36449975e-01
-8.26463580e-01 1.62251115e-01 8.42557624e-02 -8.75941217e-01
-7.87082434e-01 6.64354339e-02 -2.89039284e-01 -1.05933428e-01
8.66443574e-01 4.35643286e-01 -2.63882130e-01 -4.27538186e-01
5.44676721e-01 3.30534011e-01 4.60210055e-01 -7.26966739e-01
1.17913580e+00 8.11727464e-01 -1.15527557e-02 -1.14458799e+00
-4.94977534e-01 -9.17436063e-01 -9.63460565e-01 -2.12255269e-01
7.79728413e-01 -1.01098335e+00 -4.91206825e-01 -8.68434086e-02
-1.20430815e+00 1.31370962e-01 -6.31418049e-01 4.99749362e-01
-4.15987283e-01 2.47212753e-01 -2.43074805e-01 -8.50459993e-01
-4.36764583e-02 -9.83851671e-01 1.56048059e+00 8.45073834e-02
1.22811552e-02 -6.76368535e-01 -1.78291664e-01 -1.34425005e-02
-2.32185759e-02 2.36998945e-01 9.60005701e-01 -6.11956120e-01
-9.27278757e-01 -3.23952883e-01 -6.05678380e-01 9.89629030e-02
4.10739053e-03 1.10019065e-01 -1.21745360e+00 -2.39664912e-01
-2.98931468e-02 -9.36412886e-02 9.05592382e-01 3.23254973e-01
1.38573992e+00 -4.89469171e-02 -5.65329552e-01 5.12670159e-01
1.35589349e+00 -1.67626902e-01 5.39136887e-01 1.81018829e-01
6.95022941e-01 7.01521814e-01 6.41111970e-01 3.65365177e-01
3.21193844e-01 6.30398452e-01 7.07463264e-01 -2.08225071e-01
-9.11344588e-02 -5.19303679e-01 3.08363773e-02 4.62165713e-01
-3.50829393e-01 1.72322586e-01 -1.00228453e+00 4.54885930e-01
-1.49304473e+00 -7.35478401e-01 -2.53046453e-01 2.19680524e+00
2.27637589e-01 4.12394255e-01 1.44608334e-01 1.68850735e-01
7.30230868e-01 1.10900566e-01 -5.90190232e-01 -2.42976705e-03
-1.35968432e-01 1.69967309e-01 4.24028367e-01 -3.18831950e-03
-1.07525814e+00 1.11537564e+00 5.64455652e+00 9.19268966e-01
-1.02152228e+00 -7.88604617e-02 2.71745354e-01 2.78612196e-01
-3.76935840e-01 1.86810851e-01 -9.05787766e-01 8.64879638e-02
2.25709975e-01 3.36964905e-01 -1.79099232e-01 1.27479255e+00
1.49094462e-01 4.93510552e-02 -1.34195161e+00 1.18074918e+00
-1.52372718e-01 -1.21583927e+00 4.66419935e-01 1.66160226e-01
5.31281888e-01 1.56576455e-01 1.57473043e-01 4.24174547e-01
-3.59929055e-02 -9.41397369e-01 1.07743394e+00 4.15752888e-01
5.61958671e-01 -6.23023808e-01 3.02511007e-01 5.73026538e-01
-1.32038534e+00 1.04663083e-02 -8.87536883e-01 8.50493833e-02
-1.42000362e-01 5.54393053e-01 -8.97931159e-01 6.01161361e-01
8.94079924e-01 7.68416107e-01 -7.22518146e-01 1.16186440e+00
-1.89450696e-01 2.87110120e-01 -5.49332559e-01 1.43126339e-01
4.06224817e-01 -1.75400585e-01 7.82374501e-01 1.23170161e+00
2.36304507e-01 1.14361182e-01 4.14581656e-01 1.39263034e+00
2.77738094e-01 1.56147346e-01 -6.91524327e-01 3.86294037e-01
8.13117623e-01 1.26652181e+00 -1.13179004e+00 -2.38652065e-01
-3.07828277e-01 5.15514076e-01 4.53972608e-01 3.72007221e-01
-6.33219540e-01 -8.42268541e-02 7.37349451e-01 3.78752351e-01
5.28917074e-01 -6.02310181e-01 -5.09210825e-01 -9.02435064e-01
1.77564267e-02 -2.95006305e-01 1.74102336e-01 -8.04505348e-01
-1.37663686e+00 3.27815920e-01 3.93388987e-01 -1.65478969e+00
3.87258202e-01 -8.15899849e-01 -8.56299281e-01 6.72382593e-01
-1.51448143e+00 -1.32378197e+00 -4.48822737e-01 6.36844695e-01
5.50862968e-01 1.35442436e-01 5.51408768e-01 -1.57507509e-02
-1.72091335e-01 3.58860552e-01 -1.72311798e-01 -6.18790723e-02
4.08758342e-01 -1.05961990e+00 3.47335726e-01 5.01578212e-01
4.18008536e-01 8.39436889e-01 3.09913099e-01 -5.02084792e-01
-1.29962099e+00 -1.20454419e+00 4.80781227e-01 -7.64924765e-01
2.41422638e-01 -8.40642631e-01 -1.13550580e+00 3.71354431e-01
-5.91916919e-01 2.54215777e-01 3.33004057e-01 2.73638785e-01
-7.05942154e-01 -1.71971709e-01 -1.19888353e+00 5.65380156e-01
1.31716585e+00 -7.38401175e-01 -8.79764497e-01 2.78747618e-01
5.73213041e-01 -3.58371168e-01 -7.31737077e-01 7.08462000e-01
2.80568868e-01 -1.05748308e+00 1.49661267e+00 -3.37069690e-01
1.33917242e-01 -6.02789819e-01 -3.45366925e-01 -8.23120773e-01
-6.30955637e-01 2.35103965e-01 -7.90613219e-02 1.07645404e+00
1.39375597e-01 -4.98809308e-01 9.60106969e-01 1.20680243e-01
-5.93756378e-01 -4.83741194e-01 -1.05771911e+00 -1.01184344e+00
1.45403147e-01 -8.77386093e-01 6.15033448e-01 9.18447196e-01
-4.50406343e-01 1.38785690e-01 3.30749571e-01 6.81268513e-01
7.52481520e-01 4.24440116e-01 8.94863427e-01 -1.51012373e+00
8.62226859e-02 -8.09605479e-01 -8.74332368e-01 -1.38874471e+00
1.59205105e-02 -1.04083169e+00 1.46285251e-01 -1.61068165e+00
-5.64364791e-02 -9.36754107e-01 -4.05937582e-01 3.74199957e-01
6.33780882e-02 2.11557806e-01 1.71624795e-01 8.13106716e-01
-5.08435667e-01 6.85770810e-01 1.21980441e+00 -2.54067391e-01
-2.95755208e-01 4.94578257e-02 -2.70378143e-01 7.98889756e-01
5.44843435e-01 -2.97830939e-01 -4.79938060e-01 -4.02225196e-01
-2.44466856e-01 -3.72774303e-01 9.28035140e-01 -1.22361696e+00
2.16989160e-01 -2.30040342e-01 6.40833676e-01 -1.38486707e+00
6.22947872e-01 -9.22777236e-01 -2.41532654e-01 1.49611697e-01
-1.96118400e-01 -2.72575915e-01 1.81934074e-01 7.82021344e-01
-1.24743752e-01 -8.51041824e-02 7.51249135e-01 -3.45864654e-01
-1.05310977e+00 4.13132817e-01 -1.39649153e-01 -1.50614977e-01
1.05025542e+00 -9.86615241e-01 -9.94876213e-03 2.54646808e-01
-5.24476171e-01 9.81443077e-02 6.30360782e-01 5.49975812e-01
1.12767851e+00 -1.25824976e+00 -6.00115299e-01 4.71462309e-01
6.96301997e-01 7.26065814e-01 2.95781910e-01 6.45734191e-01
-3.60016078e-01 4.41585898e-01 -2.00406194e-01 -1.47259891e+00
-9.15809810e-01 6.61561847e-01 8.15689638e-02 3.62376571e-01
-1.08734083e+00 8.61914039e-01 9.84344959e-01 -7.01252639e-01
1.02429256e-01 -8.32776964e-01 -1.64748952e-01 -2.22857386e-01
9.29041952e-02 1.73637927e-01 3.16445976e-02 -8.47416282e-01
-6.46049500e-01 9.05396223e-01 -1.40928030e-01 3.00734490e-01
1.01990330e+00 1.72822461e-01 1.70360655e-01 5.44949830e-01
9.19814050e-01 -1.98858842e-01 -1.51054370e+00 -2.16280952e-01
4.91688401e-02 -6.14292681e-01 1.41524300e-02 -2.62359023e-01
-5.90853930e-01 1.05010033e+00 7.63497889e-01 2.68356465e-02
6.92619383e-01 4.23450947e-01 3.17254215e-02 4.60629553e-01
5.94909310e-01 -9.20328856e-01 1.67912766e-01 4.64424491e-01
9.38935220e-01 -1.37539291e+00 1.65983394e-01 -6.99457169e-01
-3.00068557e-01 1.04658628e+00 6.74178600e-01 -4.56988245e-01
5.87412417e-01 -2.49954477e-01 -3.34424824e-01 -5.34871995e-01
-2.19718188e-01 -2.66969413e-01 8.09509397e-01 8.41952026e-01
3.45651917e-02 3.13572466e-01 2.97962159e-01 3.50759298e-01
-1.01007067e-01 -7.72114515e-01 -9.52063203e-02 8.90161574e-01
-8.38473439e-01 -4.84676659e-01 -6.29995465e-01 3.83525699e-01
4.27508414e-01 9.42252800e-02 -4.39010471e-01 1.06416893e+00
2.91633070e-01 5.80924988e-01 5.99375248e-01 -2.27315292e-01
5.25922716e-01 -2.66798317e-01 3.96791220e-01 -9.88685727e-01
-2.45594770e-01 1.10030219e-01 -4.11047995e-01 -4.00102437e-01
-3.89078289e-01 -6.43244684e-01 -1.48139942e+00 1.46232128e-01
-4.52492416e-01 2.80578937e-02 6.84435606e-01 8.57840657e-01
3.81867766e-01 1.68964371e-01 5.98973811e-01 -1.37790477e+00
-3.19144070e-01 -7.15125263e-01 -5.09547412e-01 3.66042703e-01
4.99482416e-02 -1.06133950e+00 -2.82331407e-01 -3.20331365e-01] | [7.821715354919434, -3.016120195388794] |
f20f27b1-9c95-4ec4-8452-f1548ad5ed2f | multi-agent-policy-reciprocity-with | 2304.05632 | null | https://arxiv.org/abs/2304.05632v1 | https://arxiv.org/pdf/2304.05632v1.pdf | Multi-agent Policy Reciprocity with Theoretical Guarantee | Modern multi-agent reinforcement learning (RL) algorithms hold great potential for solving a variety of real-world problems. However, they do not fully exploit cross-agent knowledge to reduce sample complexity and improve performance. Although transfer RL supports knowledge sharing, it is hyperparameter sensitive and complex. To solve this problem, we propose a novel multi-agent policy reciprocity (PR) framework, where each agent can fully exploit cross-agent policies even in mismatched states. We then define an adjacency space for mismatched states and design a plug-and-play module for value iteration, which enables agents to infer more precise returns. To improve the scalability of PR, deep PR is proposed for continuous control tasks. Moreover, theoretical analysis shows that agents can asymptotically reach consensus through individual perceived rewards and converge to an optimal value function, which implies the stability and effectiveness of PR, respectively. Experimental results on discrete and continuous environments demonstrate that PR outperforms various existing RL and transfer RL methods. | ['Jianye Hao', 'Yunfeng Shao', 'Qing Wang', 'Yinchuan Li', 'Haozhi Wang'] | 2023-04-12 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-0.48869777 0.0690933 -0.6805443 0.16511792 -0.63970697 -0.5767802
0.49434492 0.0367951 -0.54817665 1.3963186 0.0832888 -0.09827779
-0.5540369 -0.83538383 -0.5497999 -1.0124712 -0.42604196 0.43205646
0.04733193 -0.34971595 0.04049781 -0.05687886 -0.8676885 -0.43462
1.0061132 0.68246996 0.32984915 0.41747403 0.2677189 0.97570604
-0.60627717 0.06390624 0.3294089 -0.5129238 -0.3866082 -0.14144453
-0.5667309 -0.77646995 -0.33875352 1.1338273 0.7208506 0.35102674
0.35035777 -1.543573 -0.6613787 1.0454359 -0.74581504 -0.10826635
0.30578923 0.45389363 1.1004547 -0.16697226 0.36536428 1.5254096
0.37003812 0.55613065 -1.1117339 -1.0614879 0.6226623 0.05386025
-1.0268157 -0.09121139 0.6215281 0.10237149 0.6754908 -0.11255582
1.0724052 1.0181283 0.04204541 1.2214231 1.2281294 -0.16019943
0.51540655 0.04293868 -0.62906283 0.80409765 0.44777492 0.43579116
-0.3893744 -0.35344502 1.2420211 0.02488207 -0.27795058 -0.6963995
-1.4956458 0.9319512 0.53632486 -0.0141165 -0.80123717 0.52878284
0.28489468 0.7355917 0.08580412 0.4969999 -0.40640956 -0.29135618
0.0096672 0.27662298 0.81576085 0.74378145 0.88096523 0.27615735
-0.15488708 0.6537513 0.47213912 0.7643257 0.42333 -1.3908228
0.5893851 0.527254 0.65910697 -0.7678103 -0.32822087 -0.5513619
-0.8472761 0.39054033 0.26684356 -0.7506025 -0.14634258 2.0768938
0.57875144 0.40448803 0.541756 0.936964 -0.02818814 0.67522335
-0.06488099 -0.65842366 0.75735897 -0.8679487 -0.69358647 -0.2640654
0.5164659 -0.0398065 0.9713434 0.15405916 -1.102624 -0.01613815
-0.8529565 0.82032746 0.1866301 -0.13411503 0.67708206 0.38474506
-0.87542766 0.40680966 -0.8223686 -0.15686843 0.5719387 0.5429881
0.14971788 0.25612524 -1.2462659 0.7268675 0.44853747 -0.05821861
-1.2463338 -0.57418984 -0.46295893 0.08909657 1.0967302 -0.8329734
1.4911464 -1.0249404 -2.3469439 -0.09985279 0.4145981 -0.41734973
0.6573083 -0.19795665 -0.08174178 0.09073985 0.19477554 0.2878981
0.7893175 -1.3487698 -0.960131 -0.09325191 0.48269263 0.7438365
-0.48941442 -0.5594375 0.074567 -0.24366748 -0.56574863 -0.9273422
-0.6170333 -0.24375634 -0.14133708 -0.42333567 0.49007526 0.09037057
0.9528747 -1.9062805 0.2475224 0.43959418 0.25635958 0.29274958
-0.43721837 0.72623235 0.69123137 -0.06188194 0.22365202 0.03837122
0.36630464 0.44400135 -0.3595472 0.4917366 -0.27337605 1.0777277
-1.2811983 -0.26713827 0.17349008 -0.01695632 -0.646753 0.2678783
-0.36762604 0.62325186 -1.3278208 0.18868305 0.37579712 -0.4195446
0.68625057 0.34505752 -0.13802485 -0.14923334 -1.2219934 1.3772483
-0.5281229 0.00941976 0.39770818 -1.1377802 0.84048074 0.3082207
0.7460456 -0.89699256 0.16970125 0.0695698 -0.00646018 -0.30123556
0.32181847 0.04657089 -0.09781008 0.83386976 -0.2905956 0.06693172
-0.09009683 0.13912502 0.9277455 0.14422336 0.62299293 -0.21585052
0.50301003 -0.16419062 0.83985215 1.017327 -0.56604624 -0.5009891
0.733729 -0.13585915 -0.67093986 -0.98261636 0.7279543 1.1903561
0.6881773 -0.11416347 -0.30208737 -0.5327196 0.3554442 0.48871806
-0.56102514 -0.2859856 -0.5038235 -0.57956076 0.21129136 0.16031979
0.87909824 -1.1319093 -0.86708957 0.5459951 -0.12974396 -0.71444243
-0.5173409 -0.36785734 -0.51031625 -1.1764382 -0.7967339 -0.5927411
0.43408987 0.38251817 0.59905285 -0.14215057 0.25464246 0.9145114
-0.23387142 -0.28915352 -0.38017923 0.10711917 0.37794122 -0.0791636
-0.06863289 -0.6083674 -0.94093496 0.49399823 -0.51133543 -0.07835888
0.855101 0.98571175 0.35101995 0.02694831 1.2657739 -0.40741512
1.3072901 -0.58951926 -1.0186914 0.4658864 -0.70165294 0.2164371
0.91492724 -0.78411746 -1.0860976 -0.21625908 0.58415526 -0.37105563
0.42519683 0.46270603 0.12683402 0.02086806 0.3211025 0.38679278
0.6362925 0.06804948 0.5326213 0.5015928 0.10266769 -0.92128676
0.6752281 0.32907385 0.09540194 -0.38440633 -0.6408323 -0.08264592
0.10564092 -0.40027526 0.3502848 -1.140775 -1.8017397 0.2976538
-0.6669111 -0.77148515 -0.33237574 0.6682543 -1.0026777 0.3208767
-0.44616258 -1.2171954 -0.1915491 -0.91187924 0.3983081 0.5783468
0.4424665 -1.1524048 0.43037096 -0.12707397 0.5900993 -0.03872103
0.52735746 -0.20383775 -0.7750473 0.35390818 0.02807102 -0.15991111
0.22791247 -0.44825444 -0.2182145 -0.87138957 -0.41763726 -0.80335176
0.41794798 0.38082582 0.6901884 -0.80142397 -0.46231356 0.01344048
1.3138505 0.35083735 0.20226343 0.5982344 0.19855568 0.35192236
0.7675487 1.1612803 0.8657316 0.38900414 0.6884719 0.19036359
0.4646664 -0.6260463 0.7705483 0.5503398 -0.05790172 -0.15267141
-0.45192626 0.40686846 -2.6017718 -1.1170553 0.6018094 2.2385488
1.0025295 -0.21941458 0.46825713 -0.43559596 0.66990334 0.01278677
-1.3071293 0.06992035 -0.05525273 -0.42722052 0.57173556 0.44647628
-0.5707561 1.0602326 6.1726537 0.793707 -0.8864214 0.04107416
0.37170917 -0.14902696 -0.24026322 -0.05079203 -0.48915482 0.36595488
0.5869027 -0.6080777 1.0651302 0.7654294 0.44883156 -0.20870967
-0.7282757 0.9074368 -0.43258604 -1.2246782 -0.24506463 0.18922271
1.0471401 0.09110543 0.05394747 0.6857858 1.2701733 -0.6083519
0.36419308 0.4834699 0.48560685 -1.1426952 0.38163337 0.5920818
-1.2195176 -0.68803906 -0.56489843 -0.2815976 0.02832991 -0.10905688
-0.6923729 0.59662527 0.38496917 0.794144 0.01960511 0.84563184
-0.3399656 0.23800671 -0.2920403 -0.75698996 0.49782678 -0.5852823
0.5753419 0.4639824 0.30558214 0.09772523 0.74923855 0.73966837
0.05734275 0.01722395 -0.53102463 -0.13417046 0.84519625 1.1682755
-0.33289245 -0.28630432 -0.20661233 0.73049676 0.6571012 0.7652476
-0.7928934 -0.27842307 0.9548824 -0.64951617 0.44021037 -0.24732664
0.3731775 -1.2128967 -0.25420326 -0.9179604 0.41762465 -0.43454695
-1.4222178 -0.02206414 -0.18790308 -1.3145636 -0.5852323 -0.07991085
-0.5312518 0.21965171 -1.817448 -0.9514279 0.02022156 0.87259626
0.17523092 -0.4910442 0.6216282 -0.17592594 -0.6439922 0.51125515
0.72780323 -0.05620457 0.66835153 -1.1020972 -0.40022522 0.27687263
-0.31319806 0.29970205 0.5238742 -0.4530504 -1.9818681 -1.0274858
-0.37752867 0.20883739 1.1371834 0.06077119 -0.42038766 0.6971171
0.34506404 -0.09475564 0.3667645 0.06141457 -0.10767647 -0.37728173
-0.93725264 1.0112253 0.9365084 -0.16975065 -0.2001827 0.19487305
0.69606197 -0.09636597 -0.9101297 -0.04757793 0.44087884 -0.7890267
0.8832861 -0.67126226 0.0171099 -0.24908452 0.09364875 -2.031401
-0.3392925 -1.3405395 -0.34562096 1.0284376 0.18709742 -1.3423147
0.6061546 0.21741608 0.22138204 -0.55716395 -1.0216111 -1.0687306
0.2945829 0.09766635 0.6721824 0.839771 0.46408033 0.45924136
-0.7010015 0.22943984 0.9907977 0.37572247 1.0102421 -0.8325888
-0.5480457 -0.6775399 0.24755388 -1.3083826 0.5236867 -0.3823457
-0.07114233 -1.5630741 0.20037854 -0.90851563 -0.46603483 0.50418264
-0.15326802 -0.33186534 0.35796386 0.4574504 -1.2941118 1.2756466
1.7621847 -0.07201201 -0.5286333 0.19260947 -0.75431526 0.39823204
1.1431527 -0.16435699 -0.7518438 -0.20308393 0.33284262 0.658728
0.19549148 -0.58849543 0.30716658 -0.8397458 -0.06763699 -0.1806896
0.25346646 -0.7427258 -0.06911176 0.9831691 -0.53836083 -0.08169492
-0.23166707 1.1767209 0.1928611 0.13610132 0.56477946 -0.16536486
-0.46640515 0.5724512 -0.4388957 0.2299479 1.4095305 0.23548463
-0.5070302 -0.95648783 -0.24154435 1.1656128 0.150218 0.22769175
0.6040828 -1.4427247 -0.6751595 -0.17036851 -0.14871806 -0.2887082
0.20140417 0.8166175 0.12821123 0.2043833 -0.351813 -0.2704072
-0.71484405 0.5055915 0.46785307 -0.37969643 -0.5647217 0.16335581
0.09092092 -0.51750666 0.20495369 -0.08886614 -0.2052927 -0.12480901
0.54165757 0.47359306 -0.9180038 0.06295365 -0.04560966 0.38329625
0.10543215 -0.38858703 1.3978311 -0.55919063 0.20459644 0.37705433
0.5892783 -0.2406928 -1.9360149 -0.68415827 -0.25043416 -0.4893872
-0.07686177 -0.7766329 -1.052738 0.32211778 0.3979977 0.46116972
0.8638494 -0.1886139 0.58559155 0.8901534 0.7267268 -1.3896977
0.58514196 0.44986308 0.57481575 -1.2345476 -0.00801183 0.16697502
-1.0335234 0.89740115 0.74721116 -0.24861997 0.33624214 -0.02159062
-0.12353154 0.14249782 -1.1534376 -0.3380927 -0.56029296 0.8377363
-0.23532797 0.27993244 -0.16159914 0.25659958 0.23437692 -0.01625208
0.69848716 0.8818708 -0.80467844 -1.1003506 -0.08598655 0.22759826
-0.09764421 0.4787909 -0.01566598 0.8236861 -0.6600513 1.0141733
-0.08061766 -0.03918057 0.04638176 -0.5902868 0.5592013 -0.09317888
-0.461404 0.22384126 -0.13731183 -0.69796693 -0.5715708 -0.44813627
-1.4376801 -0.42785528 -0.30302724 0.40414873 0.2977322 0.8746055
0.59123 0.4911251 1.2103354 -0.47604993 -1.395137 -0.713136
-0.61186165 0.04615393 0.5365917 -0.9199656 -0.1963606 -0.702002 ] | [3.8393571376800537, 2.0903942584991455] |
dc40e20f-8e47-4245-b0a9-a6c608476265 | transferring-neural-speech-waveform | 1910.12381 | null | https://arxiv.org/abs/1910.12381v2 | https://arxiv.org/pdf/1910.12381v2.pdf | Transferring neural speech waveform synthesizers to musical instrument sounds generation | Recent neural waveform synthesizers such as WaveNet, WaveGlow, and the neural-source-filter (NSF) model have shown good performance in speech synthesis despite their different methods of waveform generation. The similarity between speech and music audio synthesis techniques suggests interesting avenues to explore in terms of the best way to apply speech synthesizers in the music domain. This work compares three neural synthesizers used for musical instrument sounds generation under three scenarios: training from scratch on music data, zero-shot learning from the speech domain, and fine-tuning-based adaptation from the speech to the music domain. The results of a large-scale perceptual test demonstrated that the performance of three synthesizers improved when they were pre-trained on speech data and fine-tuned on music data, which indicates the usefulness of knowledge from speech data for music audio generation. Among the synthesizers, WaveGlow showed the best potential in zero-shot learning while NSF performed best in the other scenarios and could generate samples that were perceptually close to natural audio. | ['Yi Zhao', 'Lauri Juvela', 'Junichi Yamagishi', 'Xin Wang'] | 2019-10-27 | null | null | null | null | ['audio-generation'] | ['audio'] | [ 1.55393183e-01 -6.47498518e-02 2.03350365e-01 2.58916020e-02
-9.76495326e-01 -5.29443145e-01 4.49678719e-01 -3.02370071e-01
-4.02791379e-03 7.06415832e-01 6.82827353e-01 -3.26384269e-02
-3.80417049e-01 -8.89242887e-01 -5.39109230e-01 -6.39462888e-01
-5.52742071e-02 3.83374155e-01 2.74971426e-01 -5.54527164e-01
1.81183487e-01 2.09633797e-01 -1.87285423e+00 4.44227248e-01
5.48243046e-01 6.99833512e-01 2.41661102e-01 1.07622099e+00
-3.03515226e-01 4.96892244e-01 -1.13510966e+00 -9.35361013e-02
3.24775428e-01 -8.13122094e-01 -3.45034242e-01 -2.71166027e-01
2.47998267e-01 -2.57841080e-01 -1.72584504e-01 7.50154138e-01
1.16207910e+00 6.49929166e-01 5.96994758e-01 -9.50368643e-01
-5.37495732e-01 1.22148299e+00 1.22888379e-01 5.74265979e-02
6.20727241e-01 2.70652920e-01 1.04312468e+00 -8.64994109e-01
5.18430948e-01 1.31318557e+00 7.55402505e-01 8.00595939e-01
-1.37169015e+00 -1.05145466e+00 -5.46618342e-01 1.96393505e-02
-1.14677715e+00 -8.25377166e-01 7.73914874e-01 -5.26707232e-01
1.06176138e+00 1.28089681e-01 6.47460520e-01 1.17116296e+00
-2.12864175e-01 5.31965315e-01 8.02982330e-01 -6.67451739e-01
5.27271330e-01 2.38225199e-02 -5.69475472e-01 3.83741818e-02
-2.94896662e-01 8.70854616e-01 -1.07808197e+00 -3.14070493e-01
9.79537964e-01 -9.89300847e-01 -2.36117333e-01 -9.36056953e-03
-1.13614416e+00 8.49881530e-01 3.54363889e-01 5.45917630e-01
-4.30313855e-01 2.35857651e-01 4.17156368e-01 6.40449941e-01
2.21651152e-01 1.22238553e+00 -3.11935961e-01 -3.93845707e-01
-1.41352665e+00 5.86075544e-01 9.67030466e-01 6.16031945e-01
2.27365240e-01 1.22053123e+00 -4.17616636e-01 1.31557763e+00
5.22228517e-02 4.89518702e-01 9.03995097e-01 -9.63774323e-01
2.94155955e-01 -3.26478660e-01 2.76868343e-02 -6.03696823e-01
-1.40989646e-01 -5.22814870e-01 -2.74778903e-01 6.59136474e-01
5.52909970e-01 -5.30992806e-01 -1.00854063e+00 1.68753982e+00
-7.93851987e-02 5.49300253e-01 3.36447269e-01 8.10459554e-01
1.01024449e+00 1.11706686e+00 -2.13957801e-01 -1.04852155e-01
7.56203353e-01 -8.86479616e-01 -7.16191947e-01 3.87077779e-02
1.57760695e-01 -9.97837782e-01 1.19712484e+00 6.88732862e-01
-1.35810804e+00 -1.23048568e+00 -1.22244477e+00 4.69410777e-01
-2.33308703e-01 -1.07725151e-01 2.43474111e-01 8.88595283e-01
-1.01173615e+00 9.54578936e-01 -4.04811651e-01 -1.19584873e-01
-3.64766084e-02 2.04881832e-01 1.29064947e-01 5.58018923e-01
-1.41468239e+00 6.29614413e-01 5.66582143e-01 -5.66175103e-01
-1.24348819e+00 -9.78442550e-01 -4.80827779e-01 3.42940122e-01
2.57275067e-02 -5.01308084e-01 1.65197706e+00 -8.17157447e-01
-2.09164786e+00 1.53181642e-01 3.47283095e-01 -7.01899886e-01
4.01147194e-02 -5.53058945e-02 -6.98951423e-01 5.99793866e-02
-1.50561556e-01 6.59164965e-01 1.14735770e+00 -1.08046508e+00
-6.26700640e-01 3.57784301e-01 -4.41860408e-01 9.55370590e-02
-3.05056632e-01 -5.24014607e-02 3.21986705e-01 -1.16396952e+00
-1.78246722e-01 -6.87597573e-01 2.40473244e-02 -3.72849315e-01
-1.85781002e-01 -6.18483387e-02 5.07496238e-01 -5.21535099e-01
1.28581440e+00 -2.30833268e+00 1.82742745e-01 2.10838497e-01
-4.90761966e-01 5.76763034e-01 -5.17222762e-01 7.98697710e-01
-3.15513313e-01 -1.23203382e-01 1.65299311e-01 1.11551248e-01
1.00776322e-01 -1.76167190e-01 -7.82568395e-01 -4.36624318e-01
-5.52904755e-02 6.19709551e-01 -1.04073167e+00 -4.15123962e-02
1.39332205e-01 3.81880403e-01 -7.02856362e-01 4.20803428e-01
-4.95582312e-01 5.89469552e-01 5.04857972e-02 3.01561892e-01
-3.49272080e-02 4.39270020e-01 -3.23934495e-01 -2.75021009e-02
-1.67397961e-01 4.81324553e-01 -1.32060838e+00 1.81880307e+00
-5.83297789e-01 6.90096974e-01 -2.03966405e-02 -4.13232952e-01
1.38116181e+00 1.11430883e+00 2.26056248e-01 -5.73725998e-01
-4.49882559e-02 5.84704995e-01 5.33920825e-01 -6.28758788e-01
4.79421765e-01 -6.85112119e-01 1.05351366e-01 5.88623047e-01
5.22584021e-01 -9.48940337e-01 2.16659129e-01 -2.76580483e-01
8.53922844e-01 1.89320311e-01 9.44378376e-02 9.74715650e-02
2.28503808e-01 -1.57948151e-01 1.66366577e-01 7.80320406e-01
1.65080175e-01 7.34245241e-01 4.31191735e-02 -7.94191360e-02
-1.13257289e+00 -1.42722237e+00 1.74695641e-01 1.51440227e+00
-5.26874125e-01 -3.88821900e-01 -7.91542649e-01 2.09393963e-01
-1.60484180e-01 1.23942709e+00 -1.75434291e-01 -3.78800213e-01
-3.75381738e-01 -1.92876920e-01 9.89720702e-01 4.07507837e-01
-1.26801491e-01 -1.69768441e+00 -4.52670395e-01 7.62432396e-01
-3.52031924e-02 -4.78228569e-01 -3.83920521e-01 2.43147999e-01
-8.53360415e-01 -3.35318595e-01 -1.16190064e+00 -8.11422408e-01
-1.90813646e-01 -2.51771033e-01 7.34541535e-01 -4.40826565e-01
-2.05398694e-01 1.92663640e-01 -4.24997687e-01 -8.79361689e-01
-9.35317576e-01 3.87270078e-02 3.80073965e-01 -1.01721950e-01
-2.23013982e-01 -1.09361756e+00 -1.78866893e-01 1.35902449e-01
-7.33597815e-01 -2.13363677e-01 4.38682735e-01 7.75234282e-01
2.27760762e-01 2.84550458e-01 1.38709319e+00 -4.16549385e-01
1.34338367e+00 -2.58643061e-01 -3.22772801e-01 -7.06850141e-02
-3.93587559e-01 4.55017462e-02 6.33991838e-01 -8.66783023e-01
-1.12869799e+00 -1.75036877e-01 -5.52074492e-01 -5.44830382e-01
-1.37370288e-01 3.80396992e-01 1.40552580e-01 1.47539482e-01
1.29194403e+00 9.66047347e-02 2.14289092e-02 -4.50768620e-01
4.14803654e-01 7.39706159e-01 7.23254740e-01 -7.63630271e-01
8.54243815e-01 -1.55945212e-01 -3.24191779e-01 -1.13634539e+00
-4.56752837e-01 -1.79954052e-01 -2.62414932e-01 -2.37436265e-01
7.23113179e-01 -6.19915605e-01 -1.81679755e-01 3.60805094e-01
-8.69098246e-01 -6.06536508e-01 -9.48948205e-01 7.79341936e-01
-9.83440280e-01 -1.58001453e-01 -4.76472229e-01 -1.06049860e+00
-3.95747036e-01 -9.98307884e-01 7.66654670e-01 3.20786059e-01
-7.61128426e-01 -7.15461016e-01 4.99644130e-01 -3.71874347e-02
8.49926293e-01 -3.57256792e-02 1.08847833e+00 -7.36262679e-01
-2.24777117e-01 5.15647717e-02 3.84668201e-01 4.41901743e-01
3.79307687e-01 5.48001379e-02 -1.30442572e+00 -2.06774935e-01
-1.28122196e-01 -5.69579363e-01 5.94141901e-01 5.30696630e-01
5.86342752e-01 -6.68167546e-02 2.37507507e-01 5.74020922e-01
9.55242038e-01 6.79808557e-01 5.36050498e-01 -6.04607724e-02
2.07578078e-01 4.75842595e-01 3.40503186e-01 4.34255719e-01
-5.31404376e-01 7.78495669e-01 -4.42724228e-02 8.46803933e-02
-6.79623544e-01 -6.02402389e-01 5.44737637e-01 1.09170568e+00
-6.02037646e-02 -1.62893161e-01 -6.08163834e-01 6.07281625e-01
-1.49261451e+00 -1.31228697e+00 3.02080035e-01 2.25129604e+00
1.05997550e+00 2.76401699e-01 5.34499705e-01 6.66856825e-01
5.31960785e-01 1.21771298e-01 -3.52305055e-01 -8.55097830e-01
9.75954905e-02 9.77313399e-01 -1.61861584e-01 3.52702498e-01
-6.55697227e-01 9.96409476e-01 7.26583815e+00 1.05789876e+00
-1.22498655e+00 2.89054308e-02 -9.36997831e-02 -4.58627433e-01
-1.90109551e-01 -2.17120066e-01 -5.58779597e-01 4.31428492e-01
1.47234714e+00 -5.05572498e-01 8.45614076e-01 5.70549846e-01
4.62121367e-01 3.32854807e-01 -9.19087529e-01 7.51763880e-01
-6.62839487e-02 -1.48948729e+00 1.90997750e-01 -3.65016311e-01
1.02868736e+00 -1.78957209e-01 9.50614363e-02 5.77136636e-01
2.75382489e-01 -1.19665003e+00 9.51261818e-01 5.88909745e-01
9.80579615e-01 -8.10332358e-01 2.70343751e-01 3.96176487e-01
-1.04881930e+00 -2.06500232e-01 -2.64618903e-01 -1.79193527e-01
4.39525247e-01 1.82879344e-01 -1.40059865e+00 2.89070666e-01
2.97462732e-01 2.28283733e-01 -6.86032027e-02 1.44205475e+00
-1.37821048e-01 1.08332324e+00 -1.04806170e-01 -1.25785485e-01
1.56666532e-01 1.46633163e-01 9.09616590e-01 1.15714645e+00
8.37383628e-01 -1.62045032e-01 1.46576375e-01 9.27205324e-01
1.99301511e-01 2.09397078e-01 -5.79153776e-01 -4.97637928e-01
6.34336829e-01 8.46714854e-01 -4.46089774e-01 -2.64311969e-01
-2.05242727e-02 4.32114094e-01 -3.92575204e-01 4.52819824e-01
-5.26827574e-01 -7.93284953e-01 4.37023222e-01 2.72599757e-01
4.06129658e-01 -2.94057224e-02 -1.04386039e-01 -4.76525128e-01
-6.58495367e-01 -1.12449861e+00 4.74497862e-02 -1.30680072e+00
-1.15825880e+00 8.36928248e-01 2.05690674e-02 -1.41855800e+00
-8.79215121e-01 -4.12086070e-01 -9.30120230e-01 1.06518352e+00
-8.57305110e-01 -7.74974287e-01 1.44893557e-01 4.53943402e-01
9.13701952e-01 -8.73537302e-01 1.20266700e+00 1.24169260e-01
2.62754112e-01 5.10567605e-01 9.67460405e-03 -8.40604231e-02
7.24209070e-01 -1.13747168e+00 5.37889421e-01 3.45094502e-01
7.42230475e-01 4.30673301e-01 9.22038674e-01 -5.79639554e-01
-8.28027725e-01 -8.25867772e-01 5.44221640e-01 3.46751628e-03
7.49388158e-01 -1.42546549e-01 -1.01009166e+00 3.78006846e-02
3.56543690e-01 -6.62453592e-01 8.33264589e-01 4.85363565e-02
-3.03804487e-01 -1.19548485e-01 -8.80434036e-01 5.91202974e-01
6.49227679e-01 -6.64288461e-01 -1.08168769e+00 7.62268975e-02
8.99727404e-01 -3.20306718e-01 -8.08082879e-01 1.22352615e-01
7.81594872e-01 -8.70571971e-01 1.05922711e+00 -5.73508024e-01
4.57296014e-01 -2.87826002e-01 -2.04888418e-01 -2.06956148e+00
-3.88742417e-01 -9.90755975e-01 1.31201856e-02 1.37735760e+00
7.08933949e-01 -3.35849643e-01 4.37483579e-01 -2.63139129e-01
-3.48355502e-01 -3.01504880e-01 -7.08237588e-01 -1.09028733e+00
1.65686116e-01 -5.98165810e-01 8.82407427e-01 5.12387037e-01
7.05287559e-03 4.70215529e-01 -4.10386026e-01 -1.97932035e-01
2.22751841e-01 -1.52959758e-02 8.86607289e-01 -1.29965997e+00
-9.78827059e-01 -6.13299370e-01 -1.49062961e-01 -4.18666840e-01
-2.51294654e-02 -9.73293722e-01 4.02119786e-01 -1.39476585e+00
-5.73536456e-01 -3.18818837e-01 -3.03141057e-01 2.02163547e-01
1.20351121e-01 1.70715988e-01 5.52771568e-01 -1.03451282e-01
3.23092967e-01 4.27564025e-01 1.43987858e+00 -1.75274700e-01
-6.94425702e-01 5.55496275e-01 -5.46458304e-01 5.19439936e-01
8.30203414e-01 -4.66653407e-01 -9.09672797e-01 -1.53771013e-01
2.23867878e-01 5.16842008e-01 1.26358616e-04 -1.49220848e+00
2.53805339e-01 -1.58441085e-02 2.21539751e-01 -3.61734897e-01
7.23569453e-01 -3.31838638e-01 3.48824948e-01 3.79175216e-01
-6.84646308e-01 -3.21054280e-01 4.91971821e-01 3.33609104e-01
-2.73996264e-01 -5.74456275e-01 8.60359251e-01 -1.49335176e-01
-6.16057277e-01 -1.79605633e-01 -5.63449860e-01 2.27746904e-01
4.85943794e-01 -4.36592579e-01 1.41876176e-01 -8.92155111e-01
-1.15807176e+00 -4.86296952e-01 -1.13171808e-01 5.47003448e-01
6.44489586e-01 -1.39170957e+00 -9.60797966e-01 6.62149370e-01
-1.21422768e-01 -5.63351095e-01 7.98227713e-02 2.13146657e-01
-1.49979532e-01 4.25110191e-01 -5.21765292e-01 -3.85442525e-01
-9.64862108e-01 3.92012239e-01 3.66930366e-01 9.45927203e-02
-4.18491423e-01 1.08212149e+00 -5.94803244e-02 -4.91747171e-01
5.84568620e-01 -3.33624095e-01 -8.09122249e-02 1.95516840e-01
7.02691674e-01 5.24181604e-01 5.10084555e-02 -3.20810616e-01
2.46138573e-01 4.52628314e-01 4.30912971e-01 -9.73070085e-01
1.25120461e+00 4.46975499e-01 3.84918362e-01 1.01836824e+00
5.68728089e-01 2.74068773e-01 -9.74591553e-01 -4.65760380e-02
-3.75635847e-02 -2.91580945e-01 3.78207229e-02 -1.11827254e+00
-7.22215533e-01 9.16513383e-01 3.91066998e-01 4.04061615e-01
1.11222482e+00 -3.01925719e-01 8.47287714e-01 2.36016378e-01
5.61684191e-01 -1.18669569e+00 4.40836728e-01 6.10067010e-01
1.15808785e+00 -4.02638406e-01 -3.43578786e-01 2.17397660e-01
-6.57719672e-01 1.29174221e+00 3.25476050e-01 -4.23586845e-01
7.66286016e-01 4.43900079e-01 1.87009946e-01 1.02654330e-01
-8.51304471e-01 -1.00607798e-01 5.46824038e-01 8.28177571e-01
5.66502154e-01 -2.05113925e-02 1.27930537e-01 5.91710508e-01
-1.06558371e+00 1.66521132e-01 5.12602270e-01 4.13534582e-01
-7.05371857e-01 -1.11867523e+00 -6.97800338e-01 4.33738828e-01
-1.99805990e-01 -1.16368480e-01 -4.30370331e-01 4.99911517e-01
2.16484770e-01 1.20051813e+00 -2.41450053e-02 -4.23589200e-01
4.52398151e-01 5.45184374e-01 5.22717834e-01 -9.70209241e-01
-9.73693490e-01 5.90973020e-01 2.53897488e-01 -1.77147835e-01
-8.51701573e-02 -5.66792905e-01 -1.35736954e+00 1.77222237e-01
-5.42899370e-01 4.75402385e-01 7.30496168e-01 6.70243800e-01
-9.24316701e-03 1.18623853e+00 5.34438491e-01 -1.13682806e+00
-8.29678893e-01 -1.27945077e+00 -8.20756078e-01 1.55305833e-01
3.11756641e-01 -3.75413537e-01 -3.25609684e-01 2.48540103e-01] | [15.657991409301758, 5.901705265045166] |
f2c446ba-b166-42c6-af1a-a4bbb7d521fb | a-method-for-crash-prediction-and-avoidance | 2212.12011 | null | https://arxiv.org/abs/2212.12011v1 | https://arxiv.org/pdf/2212.12011v1.pdf | A Method for Crash Prediction and Avoidance Using Hidden Markov Models | In recent years, automotive technology has made a steady progress. In particular, Advanced Driver Assistance System (ADAS) has enabled many safety features in commercial vehicles, for instance, pedestrian detection, lane keeping assist, emergency automatic braking, etc. Although these features provide drivers with a safer operational environment, crashes still happen occasionally due to the complex road conditions and the unpredictable movement of road users including vehicles, pedestrians, bicyclists, and non-motorized vehicles. In this paper, we aim at predicting the possibilities of crashes between vehicles on highway and implementing an appropriate active safety system to prevent the same. In particular, hidden Markov models are developed for the traffic lanes and speed change of vehicles on highway. Algorithms are developed for the prediction of crash probabilities. Simulation experiments are conducted using Matlab, the results illustrate the effectiveness of the proposed research. | ['Yaobin Chen', 'Brian King', 'Lingxi Li', 'Avinash Prabu'] | 2022-12-22 | null | null | null | null | ['pedestrian-detection'] | ['computer-vision'] | [-3.26568812e-01 -1.17618725e-01 -3.31232607e-01 -3.43660951e-01
-1.09874226e-01 -4.75212671e-02 5.16493857e-01 -1.08943380e-01
-4.62607592e-01 8.58648002e-01 -1.79370970e-01 -8.44547749e-01
-1.67081624e-01 -7.68489301e-01 -3.14760536e-01 -7.46898651e-01
1.25863209e-01 6.64797053e-02 8.14077556e-01 -4.66730118e-01
2.73925960e-01 9.03639853e-01 -2.10249496e+00 -4.13151175e-01
9.61844623e-01 7.04012871e-01 3.81615430e-01 4.87243295e-01
-1.35667503e-01 4.45060372e-01 1.02718463e-02 -2.71045655e-01
-8.89717042e-03 4.02125157e-03 -1.23731131e-02 -1.99935153e-01
-4.38576281e-01 -6.07136309e-01 -5.43523014e-01 7.15189159e-01
-5.27150333e-02 2.67332226e-01 7.36899018e-01 -1.82594645e+00
2.55621880e-01 -4.70849089e-02 -3.90728682e-01 2.68203080e-01
1.42604038e-02 6.87160552e-01 1.38725802e-01 -7.59767056e-01
1.14454173e-01 1.24081147e+00 2.47586250e-01 3.91024768e-01
-7.11145163e-01 -1.00293827e+00 4.05271724e-02 1.12552834e+00
-1.63112652e+00 -5.52172720e-01 5.72116852e-01 -5.42220294e-01
6.88591599e-01 3.80206779e-02 6.77656353e-01 6.67056620e-01
7.69860327e-01 5.37241101e-01 8.58610392e-01 -1.01936065e-01
-3.80860381e-02 4.89216685e-01 7.33469367e-01 2.95476824e-01
5.46835899e-01 4.84841317e-01 1.05094932e-01 2.16534972e-01
-1.78283095e-01 1.49375319e-01 2.42730305e-01 -1.70875877e-01
-6.30585372e-01 7.81093657e-01 -1.59541979e-01 2.97050308e-02
-5.85036874e-01 -3.11967671e-01 3.56215984e-01 -2.39903461e-02
-5.37244342e-02 -7.09335506e-01 -9.35552865e-02 -3.47512573e-01
-3.39826912e-01 5.37616551e-01 4.20653224e-01 1.21591258e+00
1.00243914e+00 1.02807701e-01 2.06488445e-01 3.60055298e-01
4.38622594e-01 8.61599565e-01 -2.01100066e-01 -7.91459620e-01
4.94369864e-01 4.95456189e-01 2.58063912e-01 -1.14205170e+00
-4.90500301e-01 1.28063470e-01 -8.95673990e-01 6.51903272e-01
1.20885991e-01 -2.87416905e-01 -5.46034276e-01 1.18500197e+00
4.03495759e-01 9.73033682e-02 7.76355788e-02 5.14677465e-01
2.87765205e-01 8.97371054e-01 7.07624614e-01 -4.06383038e-01
1.07315648e+00 -4.03373152e-01 -1.24766850e+00 -4.92278971e-02
5.23024380e-01 -7.20535696e-01 5.64805567e-01 2.50827909e-01
-8.47671092e-01 -8.86885166e-01 -8.82104099e-01 3.85358930e-01
-6.50095820e-01 -1.03604354e-01 8.38582069e-02 6.25927866e-01
-5.32117844e-01 4.27661724e-02 -7.06789672e-01 -1.88991830e-01
1.15527034e-01 3.04999113e-01 -3.11371922e-01 -1.37679011e-01
-1.50321400e+00 1.41949797e+00 3.03434849e-01 3.59055042e-01
-8.35194886e-01 -3.30264211e-01 -7.91337550e-01 -1.70556605e-01
4.76946384e-01 -2.57067829e-01 9.15623724e-01 -4.66489553e-01
-1.08783531e+00 2.77249426e-01 -4.33042198e-01 -4.94893134e-01
6.09507322e-01 -1.80200249e-01 -1.09350204e+00 -3.03455234e-01
4.50792685e-02 1.83572873e-01 3.09815466e-01 -9.38407838e-01
-1.28263128e+00 -2.48621985e-01 -2.55838245e-01 3.42252036e-03
2.06871316e-01 2.17340991e-01 -3.42054404e-02 2.52101004e-01
-5.63123763e-01 -1.17506313e+00 -3.09889853e-01 -2.85321921e-01
-5.11680484e-01 -5.78821957e-01 1.44438279e+00 -6.07917368e-01
1.41449881e+00 -2.56509757e+00 -3.84188473e-01 3.42486888e-01
-1.41470596e-01 4.58719850e-01 4.21795398e-01 4.54670250e-01
1.36267707e-01 -2.40747124e-01 1.06049232e-01 -8.56204517e-03
-2.41364643e-01 4.09270972e-01 -9.21952128e-02 2.95539051e-01
1.74282625e-01 2.81276733e-01 -4.75110322e-01 -6.11973524e-01
8.83560598e-01 3.38597655e-01 3.81607329e-03 3.22418094e-01
3.41384679e-01 3.33180249e-01 -4.91083473e-01 2.91244626e-01
9.10874009e-01 8.28245282e-01 -3.58666599e-01 2.23481178e-01
-1.09076297e+00 -5.14172725e-02 -1.21517277e+00 2.43517250e-01
-4.20104623e-01 8.20529521e-01 3.09342388e-02 -9.62448895e-01
9.15166855e-01 2.99804598e-01 1.58784464e-01 -8.71105433e-01
2.31302947e-01 2.25155786e-01 1.87050954e-01 -9.30243075e-01
5.54440439e-01 1.35707200e-01 1.20896660e-01 -1.56234145e-01
-6.38132870e-01 3.25853884e-01 2.59055287e-01 -1.37385121e-02
5.97247481e-01 -3.60330671e-01 2.26141900e-01 -8.54813457e-02
1.00744665e+00 2.08853692e-01 5.58363795e-01 1.88760370e-01
-6.32163942e-01 -3.00076544e-01 2.68409282e-01 -3.28177780e-01
-1.01084197e+00 -1.12827659e+00 -1.91163823e-01 3.51996809e-01
6.28175437e-01 8.03603232e-02 -5.50452054e-01 -1.27197087e-01
7.43915737e-02 1.30654681e+00 -5.90904430e-02 -5.44596791e-01
-7.52551138e-01 -1.20435193e-01 6.97264224e-02 3.52613688e-01
6.66631937e-01 -7.74293244e-01 -5.09501100e-01 3.96349192e-01
1.70264989e-02 -1.24530292e+00 4.84446762e-03 -3.99708658e-01
-4.13465142e-01 -1.28131974e+00 -2.40052298e-01 -8.54834676e-01
4.35629189e-01 9.31412518e-01 2.62347966e-01 3.60030681e-01
-2.45125592e-01 1.13039352e-02 3.18235811e-03 -7.55907178e-01
-7.75858402e-01 -4.74807881e-02 1.88523814e-01 1.74308404e-01
6.82425499e-01 -6.07438348e-02 -6.19468033e-01 7.83202231e-01
-4.27422136e-01 2.44086102e-01 4.99513507e-01 3.26309383e-01
2.13723898e-01 6.80028558e-01 9.10663426e-01 -4.04058099e-01
4.53913450e-01 -8.05782497e-01 -8.04176748e-01 -7.75840804e-02
-8.12941968e-01 -3.71928841e-01 7.24813700e-01 8.12995732e-02
-1.40127766e+00 9.42005441e-02 -4.80612189e-01 2.23648742e-01
-1.01366067e+00 1.30667379e-02 -6.15848362e-01 1.21042348e-01
1.24774054e-01 2.55640417e-01 5.35202265e-01 -1.55935153e-01
1.04298830e-01 1.06710684e+00 4.47499633e-01 2.25578621e-01
9.44276154e-01 1.35611445e-01 2.51120031e-01 -1.16232622e+00
1.32344365e-01 -5.23267627e-01 -5.10177612e-01 -9.84721601e-01
1.01311147e+00 -8.82642329e-01 -1.11352932e+00 7.57732213e-01
-1.21403444e+00 1.12496413e-01 3.65886897e-01 8.76202166e-01
-3.99217665e-01 3.27366978e-01 -2.18749687e-01 -1.35281050e+00
3.15164834e-01 -1.37646627e+00 3.16009909e-01 5.36363900e-01
-7.64147863e-02 -7.42009103e-01 -1.97348684e-01 2.95069903e-01
3.57138664e-01 2.19382197e-01 1.13991594e+00 -1.96699709e-01
-9.45913315e-01 -4.57807839e-01 -1.99638724e-01 5.01634181e-01
1.80842400e-01 5.41187286e-01 -4.65850651e-01 1.86805308e-01
-4.82385606e-01 4.49213922e-01 2.10204810e-01 4.70671177e-01
8.34669173e-01 -5.65566234e-02 -7.63897479e-01 -6.77013919e-02
1.11008191e+00 1.05473995e+00 1.04822028e+00 2.87505239e-01
3.79436702e-01 1.08766866e+00 1.32742965e+00 2.02967957e-01
6.49324179e-01 6.97890520e-01 6.78831220e-01 -6.06306642e-02
1.56920910e-01 2.95699690e-03 2.71569490e-01 5.29635966e-01
-2.84747481e-01 2.11400408e-02 -8.86656642e-01 5.38796067e-01
-1.82129550e+00 -1.21874571e+00 -1.05163622e+00 2.36473799e+00
2.69283861e-01 5.48986495e-01 3.93653661e-01 5.96345842e-01
1.09256423e+00 -4.30467218e-01 -2.59380817e-01 -6.15792215e-01
1.57151967e-01 -4.66067761e-01 7.93321729e-01 7.15857387e-01
-8.53810549e-01 6.75872982e-01 5.99486589e+00 7.29406655e-01
-8.41539860e-01 -4.42413837e-02 5.39302170e-01 3.50986362e-01
-6.67097420e-02 8.44159126e-02 -1.21393692e+00 9.92576301e-01
1.16015255e+00 -5.05352557e-01 9.59087387e-02 8.56037676e-01
1.06028056e+00 -5.32096803e-01 -3.48695576e-01 6.39888048e-01
-5.05346715e-01 -9.43954527e-01 -2.40293264e-01 -1.31697124e-02
1.81155413e-01 -5.37641704e-01 -1.23639852e-01 4.42546070e-01
-1.42885372e-01 -5.40773988e-01 5.29798746e-01 8.70172858e-01
3.09437364e-01 -1.52265203e+00 9.32517946e-01 7.50364602e-01
-1.16671193e+00 -3.06427747e-01 -9.06749105e-04 -2.23871082e-01
8.10540974e-01 3.98056775e-01 -7.16920555e-01 4.07431751e-01
4.10221845e-01 3.35537642e-01 -3.55778337e-01 1.20401156e+00
-1.01791784e-01 5.21376073e-01 -1.07586287e-01 -4.29388940e-01
2.58461833e-01 -4.61925298e-01 6.21948421e-01 9.93649483e-01
2.76641726e-01 1.37031913e-01 -1.03003040e-01 3.82856399e-01
8.68010879e-01 1.13737516e-01 -1.15180051e+00 5.20221949e-01
5.40704012e-01 1.06133246e+00 -3.02433342e-01 -3.32292676e-01
-5.66919267e-01 6.68851808e-02 -4.19042796e-01 4.68168020e-01
-1.24746668e+00 -7.24884808e-01 1.17746997e+00 8.18433285e-01
-2.10401475e-01 -6.90277398e-01 -1.82957411e-01 -5.22872396e-02
-2.22000182e-01 -1.36435732e-01 -1.38457060e-01 -5.87392151e-01
-5.17756343e-01 1.32459417e-01 5.22183001e-01 -1.23079717e+00
-2.67622441e-01 -6.02721035e-01 -9.02393162e-01 9.61137116e-01
-1.59732962e+00 -6.09580696e-01 -3.30865115e-01 7.41479456e-01
7.22802401e-01 -4.61010128e-01 1.06338702e-01 7.80244172e-01
-8.65250707e-01 2.67392546e-01 1.12925582e-01 -2.39914298e-01
3.68818730e-01 -3.78179133e-01 3.00176263e-01 9.79032040e-01
-1.10748839e+00 3.18398595e-01 1.09805179e+00 -7.38302469e-01
-1.37766659e+00 -1.02230549e+00 1.04988658e+00 1.86295629e-01
5.27144253e-01 -4.56970297e-02 -6.91631019e-01 3.08165014e-01
1.48642361e-01 -6.10199690e-01 2.74641663e-01 -2.20119417e-01
5.48363924e-01 -3.71262252e-01 -8.72888327e-01 9.04610813e-01
6.87375426e-01 -2.35942647e-01 -3.46194923e-01 9.14974511e-02
3.78944844e-01 1.21976063e-01 -8.15247893e-02 4.11481798e-01
5.62491417e-01 -9.66166615e-01 6.87714577e-01 -5.60912490e-01
-1.84823930e-01 -4.76380557e-01 3.14745098e-01 -1.02615845e+00
-4.09746587e-01 -2.81743377e-01 1.85547143e-01 1.08548462e+00
2.77327865e-01 -9.50538337e-01 4.44349498e-01 1.01401544e+00
-5.05586803e-01 -4.18874562e-01 -1.17048299e+00 -8.58393073e-01
-3.30426991e-01 -7.02611387e-01 4.70564961e-01 1.95448160e-01
-1.79094151e-01 1.67698577e-01 -5.25176764e-01 3.72826099e-01
7.02777565e-01 -6.03899181e-01 9.91661727e-01 -1.18915081e+00
4.66079235e-01 -4.19692189e-01 -7.54514813e-01 -8.24529827e-01
4.09368873e-01 -1.11593343e-01 3.28261405e-01 -1.33538067e+00
-1.71268880e-01 -4.68534678e-01 6.40767887e-02 7.44819120e-02
-2.11330831e-01 -3.92395496e-01 -1.56092569e-01 -1.95526227e-01
-1.31473213e-01 4.32452291e-01 9.08046603e-01 -1.00951605e-01
-9.79218483e-02 5.96320271e-01 -1.54961839e-01 6.41915023e-01
1.03091586e+00 -3.67469281e-01 -5.20263255e-01 2.15218395e-01
-2.45855093e-01 3.07945102e-01 4.60881680e-01 -1.00875270e+00
4.13358897e-01 -5.69569290e-01 -2.25853816e-01 -1.06938362e+00
3.15103948e-01 -1.36511648e+00 5.96510530e-01 8.26873541e-01
1.91292077e-01 1.44126520e-01 2.87803352e-01 7.01041400e-01
-1.49327502e-01 -6.22158647e-02 9.12280142e-01 6.44073427e-01
-1.13250995e+00 3.01488817e-01 -1.46412528e+00 -6.23754680e-01
2.03508592e+00 -5.72459042e-01 -2.77968794e-01 -3.69123936e-01
-5.25483251e-01 5.31577110e-01 -6.71392726e-03 7.56875038e-01
7.22346365e-01 -1.20848393e+00 -4.71308202e-01 4.22312140e-01
5.45202121e-02 -4.21645731e-01 8.13036621e-01 1.02126932e+00
-5.07501543e-01 6.04147911e-01 -5.71843028e-01 -2.99228877e-01
-1.62519777e+00 7.48369098e-01 4.83829938e-02 4.32786852e-01
-4.46354926e-01 -2.19863564e-01 1.16177276e-01 3.43590558e-01
7.27130994e-02 4.43412550e-03 -5.80455661e-01 -1.49219751e-01
6.02214098e-01 1.19321513e+00 -7.73215145e-02 -1.20578539e+00
-4.16803092e-01 3.69226903e-01 4.55265455e-02 2.41188943e-01
5.29155552e-01 -6.79871559e-01 4.23916459e-01 5.57366788e-01
8.57870281e-01 -1.84123307e-01 -1.03800201e+00 2.74781376e-01
4.48812358e-03 -5.29869497e-01 -9.70480032e-04 -2.32615829e-01
-7.83564270e-01 1.00633824e+00 7.44293630e-01 4.78169084e-01
7.69473255e-01 -3.52680892e-01 1.10583436e+00 1.73157096e-01
6.79478884e-01 -1.34467936e+00 -8.21680009e-01 4.03729528e-01
4.00913090e-01 -1.19853973e+00 -2.86841959e-01 -5.85862458e-01
-6.77272141e-01 1.10545850e+00 7.09021568e-01 1.56924754e-01
1.04432619e+00 2.83145517e-01 1.20023359e-03 3.28456402e-01
-7.44086444e-01 -3.96431893e-01 -1.21121891e-01 7.23160803e-01
-5.01517318e-02 2.72852480e-01 -7.21197367e-01 2.79577553e-01
1.53658435e-01 -4.12322730e-02 5.59507132e-01 7.22545147e-01
-9.32530105e-01 -8.92911851e-01 -4.14163113e-01 4.09528673e-01
2.57857814e-02 5.03777027e-01 2.63937891e-01 1.18754375e+00
3.43424141e-01 1.45369804e+00 1.30378321e-01 -5.08097231e-01
7.59425581e-01 1.09785311e-01 -3.30556452e-01 -6.72371835e-02
1.34424284e-01 -6.38082683e-01 3.30849618e-01 -2.88882345e-01
-6.28799945e-02 -8.94362867e-01 -1.42901874e+00 -7.92286098e-01
-2.29968384e-01 1.70287400e-01 9.48759139e-01 1.25969803e+00
3.40489984e-01 3.18028450e-01 1.07739258e+00 -5.33692539e-01
-6.15660474e-02 -6.51870787e-01 -5.94477236e-01 7.91530386e-02
2.79839575e-01 -1.23392451e+00 -4.62042332e-01 -2.88370758e-01] | [5.78560733795166, 1.1666598320007324] |
a94ffa13-a1df-40e7-baf2-6adaa29faa6f | region-based-evidential-deep-learning-to | 2208.06038 | null | https://arxiv.org/abs/2208.06038v1 | https://arxiv.org/pdf/2208.06038v1.pdf | Region-Based Evidential Deep Learning to Quantify Uncertainty and Improve Robustness of Brain Tumor Segmentation | Despite recent advances in the accuracy of brain tumor segmentation, the results still suffer from low reliability and robustness. Uncertainty estimation is an efficient solution to this problem, as it provides a measure of confidence in the segmentation results. The current uncertainty estimation methods based on quantile regression, Bayesian neural network, ensemble, and Monte Carlo dropout are limited by their high computational cost and inconsistency. In order to overcome these challenges, Evidential Deep Learning (EDL) was developed in recent work but primarily for natural image classification. In this paper, we proposed a region-based EDL segmentation framework that can generate reliable uncertainty maps and robust segmentation results. We used the Theory of Evidence to interpret the output of a neural network as evidence values gathered from input features. Following Subjective Logic, evidence was parameterized as a Dirichlet distribution, and predicted probabilities were treated as subjective opinions. To evaluate the performance of our model on segmentation and uncertainty estimation, we conducted quantitative and qualitative experiments on the BraTS 2020 dataset. The results demonstrated the top performance of the proposed method in quantifying segmentation uncertainty and robustly segmenting tumors. Furthermore, our proposed new framework maintained the advantages of low computational cost and easy implementation and showed the potential for clinical application. | ['Guang Yang', 'Javier Del Ser', 'Yang Nan', 'Hao Li'] | 2022-08-11 | null | null | null | null | ['brain-tumor-segmentation'] | ['medical'] | [-1.45618454e-01 3.31400424e-01 -3.53204429e-01 -7.15848207e-01
-1.17167568e+00 -1.36351183e-01 4.58074808e-01 4.26513106e-01
-4.96391922e-01 1.10496998e+00 1.88448802e-01 -1.58719867e-01
-4.99175638e-01 -8.27068746e-01 -4.59935665e-01 -9.25724268e-01
8.10285360e-02 6.04580045e-01 3.57973367e-01 4.98762250e-01
3.35651159e-01 3.99240971e-01 -1.25307870e+00 1.53999895e-01
1.39107788e+00 1.43334484e+00 2.42539681e-02 1.27094164e-01
-3.80265474e-01 6.53367817e-01 -7.51189172e-01 -1.78557098e-01
-1.41694382e-01 -8.45710486e-02 -6.57615542e-01 -2.73839049e-02
-4.08563495e-01 -4.70380098e-01 1.43640846e-01 1.23516679e+00
4.78421301e-01 -1.23412140e-01 1.15501404e+00 -1.28908181e+00
-4.25599903e-01 9.77291584e-01 -5.24743736e-01 -1.56101212e-02
6.92953914e-02 -4.06947033e-03 4.86713916e-01 -6.55481696e-01
2.80599535e-01 9.41779971e-01 6.85576737e-01 3.16836476e-01
-7.76609063e-01 -5.54823041e-01 3.94127779e-02 2.48409938e-02
-1.43696582e+00 -1.53765440e-01 4.90271807e-01 -6.82707846e-01
4.89902347e-01 -7.22035989e-02 7.11813152e-01 8.81821334e-01
8.68263960e-01 8.32508743e-01 1.42006433e+00 -2.73325592e-01
7.42930532e-01 2.85970867e-01 2.42294043e-01 6.65709436e-01
4.82581496e-01 8.95694047e-02 -2.42314354e-01 -1.41959399e-01
5.88272989e-01 -6.62154630e-02 -1.44329384e-01 -7.95222819e-02
-8.95794213e-01 9.52128887e-01 5.95120430e-01 1.46896660e-01
-3.94536734e-01 2.33725041e-01 4.17511165e-01 -5.10514677e-01
7.08869100e-01 -4.79502045e-02 -1.17977023e-01 -3.12514300e-03
-1.30362368e+00 -7.90364444e-02 7.26366043e-01 5.99025905e-01
1.65771782e-01 -2.26489320e-01 -6.19413793e-01 6.38597369e-01
7.60853291e-01 3.99338156e-01 6.10805333e-01 -8.75128984e-01
1.56539027e-02 3.28703880e-01 9.23304930e-02 -9.44717467e-01
-5.99336863e-01 -4.30117518e-01 -9.65638101e-01 2.97296375e-01
3.72526377e-01 -2.41853416e-01 -1.40680182e+00 1.50754738e+00
1.01472802e-01 1.37941003e-01 -7.32380077e-02 6.97995663e-01
9.71989691e-01 4.33786839e-01 3.86237413e-01 -2.43746221e-01
1.13794696e+00 -5.75285375e-01 -9.81957316e-01 2.23510548e-01
3.09082031e-01 -4.94820774e-01 4.36150640e-01 4.75924224e-01
-8.79690886e-01 -2.31638432e-01 -1.01480067e+00 5.69230795e-01
-3.25432926e-01 1.98920853e-02 5.94143093e-01 1.02239037e+00
-9.32978213e-01 5.06847441e-01 -1.11221600e+00 -4.80917282e-03
9.01506424e-01 2.03101322e-01 -1.46229025e-02 2.38142163e-02
-1.32708681e+00 9.61647809e-01 8.94127071e-01 4.03143734e-01
-9.77014422e-01 -4.34979171e-01 -9.13108408e-01 -9.59572103e-03
1.62019208e-01 -4.54648107e-01 1.21343219e+00 -4.96655762e-01
-1.64378774e+00 5.14189839e-01 2.42801026e-01 -6.51018381e-01
8.06275785e-01 -5.85388541e-02 -2.33691663e-01 1.54676959e-01
6.97941259e-02 7.73217618e-01 5.81826270e-01 -1.30222189e+00
-5.49153745e-01 -3.08813363e-01 -4.96998250e-01 2.59007979e-02
5.94849139e-02 -2.38394350e-01 -2.88944036e-01 -3.97784472e-01
5.76537490e-01 -5.60685575e-01 -3.65130156e-01 -1.23809390e-01
-4.40678924e-01 -2.88888335e-01 2.39549205e-01 -7.26559103e-01
1.09860981e+00 -1.78003788e+00 -1.81904286e-01 6.33900046e-01
1.91976458e-01 -4.57343198e-02 5.68892300e-01 -9.83723104e-02
3.38534623e-01 4.35937434e-01 -8.39883089e-01 -5.73978536e-02
8.60030428e-02 3.38722408e-01 5.53739890e-02 5.01336455e-01
1.97638154e-01 6.98151767e-01 -8.57659817e-01 -1.11655664e+00
3.20315272e-01 4.49200660e-01 -1.90948650e-01 3.16279121e-02
-3.82864147e-01 4.37800795e-01 -5.66622674e-01 8.32139850e-01
7.86375284e-01 -5.71250468e-02 -2.93993980e-01 -2.13003770e-01
4.05949913e-02 -3.78269225e-01 -1.21309114e+00 1.60297918e+00
-2.28596076e-01 3.00919175e-01 -1.46945328e-01 -1.01379478e+00
9.60962236e-01 3.48802805e-01 6.51839435e-01 -2.62121379e-01
6.09458089e-01 5.03548086e-01 -1.46696582e-01 -6.98267937e-01
2.01313078e-01 -2.36477375e-01 -1.90822154e-01 2.17805713e-01
-5.55870263e-03 -7.72702038e-01 -9.77764726e-02 1.69543419e-02
7.03082681e-01 2.59063751e-01 5.43994009e-01 -2.94140339e-01
3.68845314e-01 1.02471814e-01 6.66895688e-01 7.40399539e-01
-6.47565484e-01 6.12223804e-01 6.66126311e-01 8.31023380e-02
-5.86072803e-01 -1.30143094e+00 -9.07887280e-01 2.80312210e-01
1.56001836e-01 1.64513618e-01 -1.05965292e+00 -6.71105444e-01
-2.80120578e-02 9.57629681e-01 -4.92271185e-01 -8.93978216e-03
5.04398644e-02 -1.01056278e+00 5.07566392e-01 5.54786086e-01
8.40134025e-01 -9.35621142e-01 -5.76615393e-01 3.41198266e-01
-1.81250423e-01 -9.65049028e-01 6.76785856e-02 2.12173119e-01
-1.02791119e+00 -8.97932172e-01 -9.15123463e-01 -3.90911072e-01
7.90897071e-01 -6.42397523e-01 9.39746082e-01 -2.81977206e-01
-3.82633686e-01 2.62333453e-01 -1.71635792e-01 -6.86016619e-01
-3.61189097e-01 -2.09435329e-01 -8.24120194e-02 -3.56173903e-01
2.30173215e-01 -1.11335769e-01 -7.43675232e-01 3.11738700e-01
-1.19189250e+00 -2.10832179e-01 7.60665894e-01 9.46455121e-01
8.84517431e-01 6.23127460e-01 8.70998979e-01 -9.18052971e-01
9.59227383e-01 -6.31249964e-01 -6.45839751e-01 2.51175284e-01
-8.82708251e-01 1.46267503e-01 3.57806608e-02 1.57248259e-01
-1.47040081e+00 -5.12522794e-02 -3.92225921e-01 1.80495363e-02
-2.52496928e-01 9.68913794e-01 9.56257284e-02 1.27834246e-01
7.43129313e-01 -2.09391356e-01 2.27565885e-01 -2.05803551e-02
2.90040135e-01 9.73009586e-01 5.00486493e-01 -8.38157713e-01
3.04018818e-02 3.33647788e-01 -1.81048606e-02 -2.86298096e-01
-8.86618257e-01 -1.10889576e-01 -4.66614157e-01 -5.53824961e-01
1.05148983e+00 -7.14507699e-01 -4.94886011e-01 8.10907125e-01
-1.00219238e+00 -3.35617363e-02 -2.60646194e-01 8.02951217e-01
-5.65836251e-01 3.91100585e-01 -4.51256037e-01 -1.03805554e+00
-4.54374433e-01 -1.57078862e+00 9.29406047e-01 4.85267311e-01
-1.05802976e-01 -9.98647571e-01 -1.77808806e-01 3.44359845e-01
5.24819136e-01 6.29950166e-01 8.39860141e-01 -6.73901320e-01
-4.17729169e-01 -4.28814083e-01 -4.89515334e-01 5.47110021e-01
7.74017125e-02 2.65972704e-01 -9.93284047e-01 6.77480921e-02
1.83888733e-01 -2.66111672e-01 9.25477266e-01 1.00361896e+00
1.61000776e+00 1.27602786e-01 -5.59452891e-01 2.05020964e-01
1.54560935e+00 1.60214365e-01 5.23527145e-01 1.23533539e-01
1.26264855e-01 5.28407931e-01 6.58329189e-01 5.77089429e-01
1.54194325e-01 1.92699357e-04 6.66851461e-01 1.22220263e-01
2.74616014e-02 -2.38443930e-02 -2.88621131e-02 6.78378940e-01
1.76881477e-02 -2.51814783e-01 -1.30993629e+00 5.54001868e-01
-1.98288381e+00 -6.41983271e-01 7.88448304e-02 2.11102986e+00
9.65079308e-01 7.21059561e-01 -1.51386112e-01 6.85703680e-02
8.54251087e-01 -3.02323133e-01 -5.76836586e-01 -2.76858866e-01
2.14101583e-01 1.03784211e-01 6.08405769e-01 3.61095846e-01
-9.35128927e-01 5.95635056e-01 6.54929733e+00 1.30061364e+00
-7.63891280e-01 5.69906682e-02 1.11391628e+00 2.12292850e-01
-3.78476024e-01 -3.85890305e-01 -6.13754034e-01 6.73065126e-01
8.71348143e-01 3.79833430e-02 -2.81201452e-01 7.70321727e-01
2.12801591e-01 -8.34346652e-01 -8.14872444e-01 7.94027507e-01
-9.49186981e-02 -1.17783749e+00 -2.74451673e-01 -8.00138339e-02
1.02015221e+00 -5.20892106e-02 1.30305380e-01 1.68050051e-01
5.02626956e-01 -1.45068991e+00 5.41658044e-01 1.15824735e+00
4.80019361e-01 -8.50451648e-01 1.49666500e+00 4.83232737e-01
-5.49093127e-01 3.76847275e-02 -3.17099899e-01 2.90677905e-01
1.61116645e-01 1.28552508e+00 -1.19411576e+00 5.63526928e-01
5.81692159e-01 3.09734583e-01 -2.41941929e-01 1.37692571e+00
-3.64883006e-01 7.12299049e-01 -6.14236116e-01 -1.82765916e-01
2.79346377e-01 -8.25767070e-02 2.02053726e-01 1.19650149e+00
4.87558544e-01 -2.76679192e-02 1.16499640e-01 1.18208981e+00
2.02592611e-01 8.37062225e-02 -3.33226681e-01 9.03790593e-02
4.62858081e-01 1.12259364e+00 -1.31894255e+00 -2.00911686e-01
-2.24515293e-02 3.49813104e-01 2.90454440e-02 1.50815859e-01
-1.08739185e+00 -1.76152915e-01 -5.93432039e-02 -3.13078821e-01
-1.99933514e-01 -6.01602253e-03 -8.30522656e-01 -6.52159274e-01
-2.16485783e-01 -4.03010070e-01 4.09982234e-01 -7.34442949e-01
-1.46885765e+00 6.98932230e-01 4.57064509e-01 -9.49503541e-01
-1.96122676e-01 -7.73526013e-01 -4.96947289e-01 8.61854076e-01
-1.27201271e+00 -8.71988595e-01 -3.63385797e-01 3.58920544e-01
3.93887430e-01 -1.91173896e-01 7.91890562e-01 -1.54118747e-01
-6.64141595e-01 2.79829234e-01 2.67314613e-01 -9.97036509e-03
5.13332427e-01 -1.38095284e+00 -5.21228075e-01 6.46205366e-01
-3.94222975e-01 1.50896773e-01 7.90129244e-01 -8.65984082e-01
-4.30535018e-01 -7.95548916e-01 2.10924745e-01 -2.23273411e-01
5.74918926e-01 1.95885301e-01 -7.30937839e-01 1.90508559e-01
8.87751430e-02 1.33646801e-01 7.32315719e-01 -2.56475806e-01
3.42598200e-01 -1.68347117e-02 -1.76353943e+00 3.31752688e-01
3.23824644e-01 -1.58838391e-01 -7.53847837e-01 2.32613206e-01
5.37200630e-01 -6.25840127e-01 -1.22395384e+00 7.63497293e-01
5.26035964e-01 -1.01797664e+00 5.68499684e-01 -2.45577823e-02
4.45850849e-01 -2.19704911e-01 -2.21531704e-01 -1.44934213e+00
3.68678421e-02 1.00390486e-01 9.75366831e-02 1.18018794e+00
6.04598820e-01 -6.12930655e-01 7.74837077e-01 1.13087022e+00
-2.91730314e-01 -1.06932116e+00 -1.14712977e+00 -3.69300395e-01
2.31610358e-01 -7.59908080e-01 6.76951766e-01 3.64178836e-01
-3.53289284e-02 -4.34562773e-01 1.44591391e-01 2.43371531e-01
8.65812421e-01 -3.39646131e-01 -2.54246034e-02 -1.24908721e+00
1.29314974e-01 -6.07948720e-01 -5.84362209e-01 -2.26965100e-01
2.12127894e-01 -7.73978710e-01 5.69003940e-01 -2.00833321e+00
2.63212591e-01 -4.58107531e-01 -6.20252252e-01 1.73742563e-01
-7.64220059e-02 -7.53443688e-02 -3.30490112e-01 1.53005272e-01
-3.05996537e-01 4.84401047e-01 1.19110811e+00 -2.34051943e-01
2.53033247e-02 2.73173749e-01 -3.99635255e-01 1.04016459e+00
9.99016762e-01 -5.84492981e-01 -5.41548252e-01 -1.23408057e-01
1.57638788e-01 3.05867642e-01 1.81353658e-01 -1.08801258e+00
4.86887753e-01 -4.51360121e-02 6.07000232e-01 -8.82412791e-01
4.27177772e-02 -9.80659068e-01 5.99732436e-02 4.18784142e-01
-2.35929698e-01 -5.74093163e-01 2.34172538e-01 7.33013570e-01
-3.67116243e-01 -7.38779604e-01 8.59728873e-01 -1.10029943e-01
-6.34685278e-01 2.21842170e-01 -4.42071497e-01 -1.56849340e-01
1.26199269e+00 -2.33622029e-01 -1.28323317e-01 -1.50888547e-01
-9.15750802e-01 4.05934721e-01 -6.06347695e-02 -2.58750975e-01
8.51260900e-01 -1.25629127e+00 -7.82559872e-01 -7.50784725e-02
7.42035285e-02 4.89734173e-01 2.87172228e-01 8.74740720e-01
-7.43441761e-01 4.27720040e-01 -2.75018185e-01 -1.09691882e+00
-5.18921971e-01 1.11810446e-01 5.15931904e-01 -3.27160090e-01
-1.89891756e-01 9.46919322e-01 -2.86922425e-01 -3.28551501e-01
5.11952281e-01 -6.91191733e-01 -4.44043100e-01 1.07662141e-01
1.02635689e-01 4.79657888e-01 1.21630862e-01 -2.67868161e-01
-3.60560030e-01 3.91182959e-01 2.03510970e-02 -4.16514635e-01
1.04793787e+00 -4.13018316e-02 -2.14886218e-01 4.18509126e-01
7.06023157e-01 -3.74864995e-01 -1.21061444e+00 -6.53507113e-02
1.76125735e-01 -1.15021668e-01 6.20353758e-01 -1.18569088e+00
-1.02115715e+00 6.65244639e-01 8.31169367e-01 2.27082357e-01
1.09273005e+00 -6.52275831e-02 4.69342381e-01 1.02904938e-01
4.52789873e-01 -1.39595461e+00 -1.92087963e-01 1.23939306e-01
7.04685688e-01 -1.73689318e+00 1.37789756e-01 -2.89034069e-01
-7.20960498e-01 1.18264937e+00 4.67111319e-01 -4.53153066e-03
1.20711243e+00 4.66953427e-01 2.73165274e-02 -2.49778181e-01
-2.33620167e-01 2.43618619e-02 4.06356901e-01 5.10911822e-01
4.81150091e-01 4.25439656e-01 -4.81834263e-01 1.00864148e+00
-3.42010632e-02 5.17673314e-01 2.54721850e-01 8.27109694e-01
-7.53408134e-01 -5.90579927e-01 -5.08169413e-01 7.91387200e-01
-6.39293075e-01 1.68602526e-01 2.42442086e-01 6.83180869e-01
2.16713011e-01 1.02043629e+00 -1.48964003e-01 1.84520660e-03
-2.88093891e-02 5.90101518e-02 3.20591837e-01 -5.02971590e-01
-4.55971844e-02 6.67273328e-02 -1.11364335e-01 -3.37644279e-01
-5.39261222e-01 -4.66974497e-01 -1.58868134e+00 1.79291785e-01
-5.23833454e-01 2.74224848e-01 1.15586793e+00 1.13881636e+00
-6.65178746e-02 8.43115687e-01 3.55114549e-01 -5.39681554e-01
-8.52356076e-01 -1.05098188e+00 -6.98536754e-01 -1.83467043e-03
-7.34108984e-02 -8.38331819e-01 -4.90777373e-01 -2.95398414e-01] | [14.365873336791992, -2.0784010887145996] |
952292b1-6dbf-4a17-95df-105eee97db43 | a-probabilistic-relaxation-of-the-two-stage | 2306.00892 | null | https://arxiv.org/abs/2306.00892v1 | https://arxiv.org/pdf/2306.00892v1.pdf | A Probabilistic Relaxation of the Two-Stage Object Pose Estimation Paradigm | Existing object pose estimation methods commonly require a one-to-one point matching step that forces them to be separated into two consecutive stages: visual correspondence detection (e.g., by matching feature descriptors as part of a perception front-end) followed by geometric alignment (e.g., by optimizing a robust estimation objective for pointcloud registration or perspective-n-point). Instead, we propose a matching-free probabilistic formulation with two main benefits: i) it enables unified and concurrent optimization of both visual correspondence and geometric alignment, and ii) it can represent different plausible modes of the entire distribution of likely poses. This in turn allows for a more graceful treatment of geometric perception scenarios where establishing one-to-one matches between points is conceptually ill-defined, such as textureless, symmetrical and/or occluded objects and scenes where the correct pose is uncertain or there are multiple equally valid solutions. | ['Onur Beker'] | 2023-06-01 | null | null | null | null | ['pose-estimation'] | ['computer-vision'] | [ 2.52507716e-01 -1.02357924e-01 8.10328573e-02 -4.29097950e-01
-1.07006526e+00 -9.72469151e-01 8.49956989e-01 4.72393125e-01
-2.11859956e-01 2.45724633e-01 -9.41905752e-02 -1.87963471e-01
-3.68486315e-01 -6.39964104e-01 -6.96053028e-01 -5.04156888e-01
2.50988424e-01 1.11401320e+00 5.18985093e-01 4.99508902e-02
5.06847262e-01 9.11299109e-01 -1.74077070e+00 -1.38694346e-01
4.70217079e-01 1.12118351e+00 2.81009048e-01 6.09313905e-01
9.65886638e-02 5.92180528e-02 -1.85146049e-01 -4.35381114e-01
3.42183828e-01 4.76288274e-02 -6.24775290e-01 3.46663743e-01
8.70614767e-01 -2.91953474e-01 2.02470526e-01 1.12962174e+00
1.82482183e-01 2.36365676e-01 7.59098291e-01 -1.28220797e+00
-1.92707330e-01 -2.28781968e-01 -8.74217153e-01 -3.75683069e-01
7.68450022e-01 5.53151108e-02 1.13626254e+00 -1.16711652e+00
6.29031420e-01 1.48001277e+00 6.15740299e-01 -1.91103071e-01
-1.58480465e+00 -1.82902709e-01 2.39414006e-01 1.00689337e-01
-1.67993712e+00 -6.04677260e-01 7.41545856e-01 -6.75146282e-01
6.54838562e-01 4.85046476e-01 5.10824800e-01 6.88818872e-01
1.42980423e-02 3.08038920e-01 6.96332157e-01 -3.25741351e-01
3.40125322e-01 -9.34922621e-02 -2.05162957e-01 3.25032681e-01
3.69274646e-01 4.94052321e-02 -5.98456025e-01 -5.82837999e-01
8.47180247e-01 -1.51488185e-03 -5.59374653e-02 -1.06457555e+00
-1.28389370e+00 5.05874574e-01 2.87905991e-01 -1.81861550e-01
-4.93595421e-01 -2.39894744e-02 -1.06222823e-01 -9.32320580e-02
1.48686603e-01 4.17726904e-01 -3.21005404e-01 2.36049108e-02
-8.63903522e-01 7.89382458e-01 5.35835803e-01 1.20837450e+00
1.02831435e+00 -4.05542284e-01 2.48419091e-01 4.58888382e-01
5.50646067e-01 5.44791818e-01 -2.51938492e-01 -1.08454955e+00
5.06937146e-01 3.02277863e-01 6.90551817e-01 -1.22395778e+00
-2.77948052e-01 -1.08533971e-01 -3.14284444e-01 4.99943316e-01
6.16669476e-01 3.94739330e-01 -6.85824335e-01 1.60086012e+00
6.36743188e-01 -9.21941400e-02 -2.99921781e-01 1.01132107e+00
3.90631109e-01 2.86342889e-01 -5.14482819e-02 -7.52415061e-02
1.57214868e+00 -4.11047280e-01 -2.09989339e-01 -4.46992785e-01
-1.66167635e-02 -1.30008602e+00 7.87152827e-01 2.17537567e-01
-1.29016161e+00 -4.07057315e-01 -8.78025234e-01 -3.77766579e-01
-2.14565799e-01 -2.06580594e-01 3.78735632e-01 3.13290894e-01
-9.06603456e-01 3.70606303e-01 -8.80779743e-01 -3.81573707e-01
9.97726172e-02 4.53772575e-01 -4.25431132e-01 -1.25649035e-01
-4.17436421e-01 9.93406892e-01 3.39098364e-01 2.85246111e-02
-4.49620456e-01 -5.24234176e-01 -8.90276790e-01 4.27195877e-02
5.07772803e-01 -1.00929499e+00 1.21326447e+00 -8.44346404e-01
-1.05874515e+00 9.51646805e-01 -5.73149562e-01 1.74295142e-01
7.27404892e-01 -2.67808825e-01 1.73943356e-01 1.34408459e-01
1.91214710e-01 7.19706476e-01 7.82072961e-01 -1.72200596e+00
-3.78329277e-01 -6.95467651e-01 8.75094011e-02 6.88758075e-01
4.25469160e-01 2.06679970e-01 -5.80663502e-01 -4.18167621e-01
9.54995513e-01 -9.07096028e-01 -1.05769426e-01 7.10584283e-01
-4.79483634e-01 -7.96743184e-02 7.10329294e-01 -5.36827683e-01
4.52772915e-01 -2.07659340e+00 1.50513172e-01 5.20489514e-01
1.64225668e-01 -2.34350801e-01 -3.59003246e-02 5.48643053e-01
-6.31336197e-02 -3.42306606e-02 -3.20148724e-03 -7.56571293e-01
9.38419327e-02 4.40197811e-02 -2.86086380e-01 8.82815003e-01
3.77533942e-01 6.41660154e-01 -8.68912101e-01 -4.68606651e-01
4.57464844e-01 5.28036594e-01 -4.30011243e-01 2.15323299e-01
-3.69463950e-01 5.73786378e-01 -4.46518034e-01 8.74821782e-01
9.86597121e-01 -8.09270442e-02 7.28608966e-02 -3.94351512e-01
-4.44798291e-01 1.64862126e-01 -1.73941040e+00 1.58121645e+00
-6.59667328e-02 3.34097534e-01 1.25172764e-01 -4.21630889e-01
8.26780379e-01 1.01670533e-01 5.09790719e-01 -2.15142831e-01
9.83632058e-02 8.77168924e-02 -3.46151024e-01 -1.70535952e-01
7.22980142e-01 -1.44897744e-01 1.57928452e-01 2.64598548e-01
-2.22172543e-01 -5.59542835e-01 -1.95674837e-01 -8.22743252e-02
7.56413281e-01 4.82576936e-01 6.45566046e-01 -2.02682719e-01
1.23736992e-01 4.53364439e-02 6.48898900e-01 6.90387666e-01
-9.16031841e-03 1.18477166e+00 2.38076746e-01 -2.29217619e-01
-1.20272386e+00 -1.53216088e+00 -3.96097064e-01 5.82020402e-01
6.47248745e-01 -3.29929739e-01 -3.04579496e-01 -7.92846754e-02
1.60429761e-01 6.50507927e-01 -3.41498464e-01 3.65831144e-02
-3.45604777e-01 -1.41576141e-01 -2.57890046e-01 5.38925052e-01
-6.12389147e-02 -6.10540986e-01 -7.13171244e-01 2.43906695e-02
-3.11754525e-01 -1.03594911e+00 -5.06623566e-01 1.51887208e-01
-7.18910694e-01 -1.06503379e+00 -4.92612928e-01 -4.12972480e-01
8.44751418e-01 6.82558298e-01 1.20055187e+00 -5.87128438e-02
-9.08646882e-02 4.32094395e-01 -1.93004280e-01 -4.76290733e-01
2.48998273e-02 -5.52039504e-01 8.45577791e-02 2.51704734e-02
2.84318328e-01 -5.85010946e-01 -6.27278507e-01 6.89478993e-01
-4.50921714e-01 5.44321239e-02 2.75887668e-01 3.02553296e-01
9.61990833e-01 -2.05308542e-01 -9.84545797e-02 -2.33489946e-01
1.61015049e-01 -3.84131461e-01 -9.11284268e-01 4.79168564e-01
-2.37171873e-01 -1.14100173e-01 1.68982357e-01 -5.65683365e-01
-7.87187994e-01 5.96143782e-01 1.18286215e-01 -7.19768047e-01
-3.07196945e-01 1.99559778e-01 -3.78750563e-01 -2.49256387e-01
6.43198609e-01 7.35656917e-02 -6.22884231e-03 -4.61126834e-01
5.55526733e-01 3.76860827e-01 5.68955898e-01 -7.83952057e-01
1.12148130e+00 7.03652263e-01 8.20411369e-02 -8.26771557e-01
-5.05354524e-01 -6.87745392e-01 -1.05677485e+00 -3.01317990e-01
9.77984250e-01 -8.42871964e-01 -8.02883327e-01 3.14436167e-01
-1.50659990e+00 2.07442775e-01 -2.07766533e-01 6.32187068e-01
-8.41553986e-01 5.45878947e-01 -8.18065777e-02 -9.37897861e-01
2.63391227e-01 -1.34133387e+00 1.53767812e+00 1.48547247e-01
-6.00432217e-01 -6.72482133e-01 -1.12096518e-01 2.75372177e-01
1.24553159e-01 3.89167041e-01 7.64289021e-01 -3.89355630e-01
-1.06350708e+00 -3.33861679e-01 -3.24788004e-01 -2.34911233e-01
1.20997369e-01 3.61256093e-01 -1.14385259e+00 -3.71154785e-01
6.71549290e-02 -2.32493043e-01 2.79533803e-01 3.89197350e-01
7.99458027e-01 -1.10316329e-01 -3.74234527e-01 5.78563452e-01
1.26155007e+00 1.43400133e-01 5.60579062e-01 1.88378721e-01
6.30991220e-01 9.65196609e-01 8.18201005e-01 4.37102854e-01
6.30535901e-01 1.12959409e+00 8.54376078e-01 -6.31169155e-02
1.31680444e-01 -3.10371399e-01 -1.90734431e-01 3.25741112e-01
4.81824204e-02 -5.17242253e-02 -1.06267798e+00 3.42756182e-01
-1.91587138e+00 -7.94899166e-01 -1.30874857e-01 2.82959843e+00
4.29407120e-01 1.14312649e-01 2.14766338e-01 -6.65865690e-02
7.73494542e-01 4.92160209e-02 -7.24212408e-01 -1.40468711e-02
6.09398596e-02 -1.68792844e-01 1.79223686e-01 7.20344543e-01
-9.27172065e-01 6.63639724e-01 5.95666552e+00 6.54011309e-01
-9.23967719e-01 -2.10303769e-01 3.52920711e-01 9.70253944e-02
-4.88409668e-01 5.20276010e-01 -6.60822332e-01 2.12658927e-01
9.44192931e-02 9.39721465e-02 3.48742664e-01 7.58152187e-01
4.76956628e-02 -5.30697525e-01 -1.32823133e+00 1.16837084e+00
-3.20450701e-02 -1.18569577e+00 2.44170912e-02 3.13665181e-01
2.85906583e-01 -2.28787675e-01 -8.49526525e-02 -3.73751938e-01
5.75486980e-02 -7.30913222e-01 1.42222416e+00 6.63368702e-01
6.95889831e-01 -6.69457376e-01 2.42056385e-01 5.13619363e-01
-1.39858544e+00 2.74923831e-01 -3.35216135e-01 4.23363149e-02
3.55597615e-01 5.12448192e-01 -5.47625005e-01 7.16832519e-01
6.11468792e-01 2.22609609e-01 -3.87096316e-01 1.32869112e+00
-2.09403806e-03 -2.14608282e-01 -6.74964726e-01 3.05901557e-01
-1.37230650e-01 -4.44660515e-01 8.33006918e-01 6.84368551e-01
3.16788912e-01 -2.88411975e-02 4.93492872e-01 9.97127116e-01
4.49009269e-01 -5.81839085e-02 -5.86444855e-01 5.41697860e-01
8.62615526e-01 1.13729227e+00 -9.78005290e-01 1.13708258e-01
-3.86150599e-01 6.83815479e-01 2.97234356e-01 3.00863415e-01
-5.86464882e-01 -4.20001671e-02 7.49413669e-01 3.84754151e-01
2.38941088e-01 -5.95046997e-01 -5.21182120e-01 -1.00905859e+00
3.58632118e-01 -5.36674678e-01 1.91807430e-02 -1.20323336e+00
-1.26945937e+00 4.82437581e-01 4.11415309e-01 -1.44675457e+00
-3.18015754e-01 -5.15572190e-01 -6.67379797e-01 1.06601989e+00
-1.28970754e+00 -1.39585674e+00 -3.52914780e-01 4.82103020e-01
2.38684714e-01 5.39608896e-01 7.10543752e-01 -3.92075442e-02
-5.51536456e-02 2.42490441e-01 -3.00716907e-01 -3.67372513e-01
6.62437141e-01 -1.17630982e+00 3.82251531e-01 9.09588873e-01
2.01276839e-01 7.95207202e-01 9.01187718e-01 -6.78658783e-01
-1.37512112e+00 -7.18631089e-01 1.06246412e+00 -6.81785703e-01
4.76885796e-01 -5.74972272e-01 -8.58771384e-01 6.01395488e-01
-2.68304646e-01 -6.57857880e-02 2.43094757e-01 1.92534864e-01
-4.29400325e-01 8.81601349e-02 -1.23806107e+00 7.33416855e-01
8.83307755e-01 -6.05749190e-01 -4.48683321e-01 4.56371516e-01
3.39470178e-01 -7.17015147e-01 -6.66933894e-01 4.55883443e-01
6.56833231e-01 -1.01059520e+00 1.28335655e+00 -2.45440945e-01
-5.97889423e-02 -6.22965753e-01 -5.47153771e-01 -7.80945182e-01
-3.82462531e-01 -7.04701543e-01 -4.60675871e-03 1.22744095e+00
1.79018870e-01 -4.65724379e-01 4.86657172e-01 9.04483557e-01
-1.16571691e-02 -6.69729590e-01 -1.08952403e+00 -7.90234685e-01
-2.83506483e-01 -5.05014479e-01 5.70236623e-01 8.11266363e-01
-4.80913818e-01 9.70776379e-02 -4.53906171e-02 5.44038177e-01
8.29309106e-01 3.73261213e-01 9.22137737e-01 -1.34447575e+00
-2.65274554e-01 -4.62445974e-01 -5.54605484e-01 -1.36950994e+00
-1.48619503e-01 -4.70150411e-01 3.22446197e-01 -1.34589994e+00
1.81267783e-01 -7.28759348e-01 2.50998795e-01 3.10642242e-01
-5.46436608e-02 -2.91310344e-02 1.20845549e-01 5.15074790e-01
-4.44173068e-01 4.01574314e-01 8.93840611e-01 2.56750911e-01
-4.09091935e-02 2.83266604e-01 -6.14508092e-01 9.90022242e-01
5.12821436e-01 -4.22920555e-01 -3.47934693e-01 -4.41750199e-01
3.60240281e-01 2.66344249e-01 8.96074474e-01 -8.99758637e-01
2.66208887e-01 -5.84729016e-01 3.75520796e-01 -8.41913223e-01
9.09063101e-01 -9.61969554e-01 3.89496028e-01 -8.49787220e-02
3.94879021e-02 4.59171742e-01 -2.75022537e-02 5.39485335e-01
4.44368795e-02 -3.15599948e-01 7.64989197e-01 -5.77758290e-02
-4.45962518e-01 3.31977069e-01 -1.92315001e-02 -1.80197999e-01
1.11160731e+00 -7.72996485e-01 -1.42950758e-01 -4.42329079e-01
-7.01801240e-01 3.40162255e-02 9.95654702e-01 4.31773275e-01
6.52554572e-01 -1.31913269e+00 -5.62662065e-01 1.56306356e-01
3.90492320e-01 4.05344665e-01 4.47682180e-02 8.04393589e-01
-3.79708260e-01 2.87888516e-02 1.96112141e-01 -1.15009594e+00
-1.37364864e+00 4.74818110e-01 3.54314685e-01 2.56651640e-01
-4.26574260e-01 9.02878821e-01 3.78115207e-01 -5.08407295e-01
2.59165585e-01 -2.03054011e-01 2.63007164e-01 -8.03480968e-02
2.69708335e-01 2.96860665e-01 1.47971973e-01 -1.11107349e+00
-7.58715689e-01 1.04788888e+00 7.47331232e-02 -3.02639514e-01
1.08606982e+00 -3.57479423e-01 -4.94988598e-02 5.15792847e-01
8.48475516e-01 2.46022165e-01 -1.62864387e+00 -2.98665464e-01
-8.94074235e-03 -8.62515330e-01 -2.19664082e-01 -5.15946567e-01
-4.57525164e-01 8.02261591e-01 4.59693998e-01 5.81982099e-02
7.30514050e-01 3.66738111e-01 3.05258125e-01 2.01305553e-01
6.17456377e-01 -6.85638249e-01 -2.55878991e-03 3.71952474e-01
1.23552871e+00 -1.17973435e+00 3.19709539e-01 -7.82608390e-01
-4.23649162e-01 1.15656078e+00 5.09465098e-01 1.06006309e-01
5.86713850e-01 1.36719063e-01 6.69112653e-02 -3.47707927e-01
-4.14487511e-01 -1.50537983e-01 8.06259811e-01 7.11286783e-01
2.25842014e-01 -4.91344817e-02 1.45834550e-01 -8.21386557e-03
-2.89495647e-01 -5.46348929e-01 1.94535088e-02 1.04669452e+00
-5.06280422e-01 -9.69569385e-01 -6.16947174e-01 1.81684107e-01
1.71352342e-01 4.31217253e-02 -3.43913138e-01 7.47126698e-01
1.18544526e-01 9.63245034e-01 2.81874001e-01 -1.87355325e-01
4.60517466e-01 -1.93027243e-01 7.77188659e-01 -6.90945983e-01
-2.72707850e-01 3.61959696e-01 -9.23989341e-02 -7.53350317e-01
-3.27439010e-01 -1.14178050e+00 -9.23933685e-01 -1.98760182e-01
-5.86448610e-01 -3.13149035e-01 7.92859674e-01 9.89116490e-01
4.00075585e-01 -1.27807990e-01 3.65632832e-01 -1.53295743e+00
-6.50922060e-01 -4.73669589e-01 -3.19480479e-01 5.09293199e-01
3.85679662e-01 -9.33489263e-01 -2.32232407e-01 -9.26849544e-02] | [7.454356670379639, -2.711169958114624] |
39b6900e-a264-4bf0-b11f-7d22f63d5574 | expert-sample-consensus-applied-to-camera-re | 1908.02484 | null | https://arxiv.org/abs/1908.02484v1 | https://arxiv.org/pdf/1908.02484v1.pdf | Expert Sample Consensus Applied to Camera Re-Localization | Fitting model parameters to a set of noisy data points is a common problem in computer vision. In this work, we fit the 6D camera pose to a set of noisy correspondences between the 2D input image and a known 3D environment. We estimate these correspondences from the image using a neural network. Since the correspondences often contain outliers, we utilize a robust estimator such as Random Sample Consensus (RANSAC) or Differentiable RANSAC (DSAC) to fit the pose parameters. When the problem domain, e.g. the space of all 2D-3D correspondences, is large or ambiguous, a single network does not cover the domain well. Mixture of Experts (MoE) is a popular strategy to divide a problem domain among an ensemble of specialized networks, so called experts, where a gating network decides which expert is responsible for a given input. In this work, we introduce Expert Sample Consensus (ESAC), which integrates DSAC in a MoE. Our main technical contribution is an efficient method to train ESAC jointly and end-to-end. We demonstrate experimentally that ESAC handles two real-world problems better than competing methods, i.e. scalability and ambiguity. We apply ESAC to fitting simple geometric models to synthetic images, and to camera re-localization for difficult, real datasets. | ['Eric Brachmann', 'Carsten Rother'] | 2019-08-07 | expert-sample-consensus-applied-to-camera-re-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Brachmann_Expert_Sample_Consensus_Applied_to_Camera_Re-Localization_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Brachmann_Expert_Sample_Consensus_Applied_to_Camera_Re-Localization_ICCV_2019_paper.pdf | iccv-2019-10 | ['camera-localization'] | ['computer-vision'] | [-5.20163402e-02 -3.38862538e-01 1.72443613e-01 -4.04943943e-01
-1.15204310e+00 -8.46850157e-01 2.41731450e-01 -3.51535022e-01
-5.22336602e-01 3.18440497e-01 -2.05092192e-01 -3.84171270e-02
-7.77876601e-02 -1.59031600e-01 -1.09353483e+00 -5.57467759e-01
4.68212157e-01 9.95547771e-01 1.62200540e-01 1.32155925e-01
4.16180223e-01 8.35920274e-01 -1.24344909e+00 -2.09863320e-01
1.01319027e+00 9.00937855e-01 4.40917611e-01 5.96610487e-01
1.70058608e-01 6.76569566e-02 -4.06857640e-01 -2.84692943e-01
6.27072871e-01 -4.00647558e-02 -5.20095050e-01 5.67294478e-01
9.13656712e-01 -2.91501909e-01 1.85639426e-01 1.40574539e+00
5.10878205e-01 2.64684677e-01 5.77896297e-01 -1.29837382e+00
-2.76636243e-01 9.28497408e-03 -7.25082278e-01 -2.42679745e-01
3.75643790e-01 3.91127095e-02 6.63465142e-01 -1.21004677e+00
6.39547288e-01 1.15319502e+00 1.05207682e+00 4.38505083e-01
-1.36479938e+00 -4.10998464e-01 2.96948969e-01 -9.70375091e-02
-1.50711453e+00 -4.05441195e-01 8.39180708e-01 -5.18647373e-01
5.76879978e-01 4.57748771e-02 3.97555828e-01 1.09060442e+00
1.63052753e-01 6.32392585e-01 8.53154123e-01 -3.20160836e-01
3.63741368e-01 -8.26965719e-02 -9.95237008e-02 5.18290341e-01
2.54186809e-01 7.74498358e-02 -3.23852837e-01 -4.66260672e-01
1.13934803e+00 2.47149020e-01 -4.31718916e-01 -9.11383808e-01
-1.29035461e+00 7.08760977e-01 4.71283108e-01 6.40333220e-02
-5.51408350e-01 2.01064944e-01 -8.16848036e-03 1.83389857e-01
1.66903332e-01 7.32490599e-01 -4.69926387e-01 2.47131526e-01
-8.86273861e-01 1.28932759e-01 7.37397134e-01 1.05510974e+00
8.44654858e-01 -1.47578895e-01 5.36188304e-01 9.13437486e-01
5.34078360e-01 6.09985828e-01 4.31664973e-01 -1.41802835e+00
4.58825886e-01 3.32647055e-01 5.40757239e-01 -1.10149074e+00
-3.14845055e-01 -4.88749444e-01 -9.27173972e-01 4.05539662e-01
5.03282249e-01 -1.22545183e-01 -1.00955355e+00 1.58306670e+00
5.10057747e-01 4.44801360e-01 -5.98666817e-02 1.46324670e+00
3.27179581e-01 4.64316458e-01 -5.27418077e-01 -1.87417325e-02
8.71899009e-01 -9.20801580e-01 -2.53707707e-01 -7.04786599e-01
1.42389461e-01 -1.12563348e+00 4.84919488e-01 7.23039567e-01
-1.01538360e+00 -9.23456371e-01 -8.94418597e-01 4.05350141e-02
-1.80411920e-01 5.16949534e-01 5.32485135e-02 1.09637387e-01
-1.15627801e+00 6.30460799e-01 -8.44565153e-01 -4.69192237e-01
2.49612983e-02 4.22924697e-01 -7.34346569e-01 -1.48888499e-01
-6.23788238e-01 9.31122422e-01 3.32509905e-01 4.94920313e-01
-8.21282744e-01 -2.39865407e-01 -9.54913378e-01 -2.46121824e-01
8.14558446e-01 -9.78911221e-01 1.31751823e+00 -1.14099872e+00
-1.52991760e+00 8.14888537e-01 -8.36526826e-02 -2.75861919e-01
6.71673417e-01 -6.96875691e-01 8.79696682e-02 -1.74169242e-01
1.69395104e-01 6.79893672e-01 1.29640043e+00 -1.71754003e+00
-5.34399033e-01 -5.86105227e-01 -2.55307466e-01 2.64145792e-01
3.81368458e-01 -1.71384618e-01 -8.68187726e-01 -4.40498620e-01
8.93938243e-01 -1.17939019e+00 -6.22671843e-01 1.71197340e-01
-4.68631774e-01 -2.41377968e-02 5.73505580e-01 -5.33951759e-01
6.17631018e-01 -1.95610380e+00 5.03525436e-01 5.57370722e-01
2.27678835e-01 1.23491950e-01 -2.21286580e-01 2.53297478e-01
-1.59942508e-01 -3.04020584e-01 -3.59920830e-01 -7.32324064e-01
-3.52651323e-03 4.26460505e-01 -3.55860651e-01 8.42453003e-01
1.03517286e-02 6.58305228e-01 -8.27361941e-01 -1.88121974e-01
4.52413678e-01 4.06857610e-01 -4.32750612e-01 3.64394665e-01
-1.94926724e-01 6.09563470e-01 -4.26220119e-01 4.19780195e-01
1.01758683e+00 -2.94889688e-01 -1.64978415e-01 -4.63227004e-01
-1.19124770e-01 -2.47202426e-01 -1.92508972e+00 2.00785017e+00
-4.61874634e-01 3.27372551e-01 3.87398869e-01 -8.07002425e-01
1.20649540e+00 8.78812820e-02 4.83528763e-01 1.62314437e-02
2.37797081e-01 5.48189402e-01 -2.36593395e-01 -3.75666499e-01
4.75997537e-01 1.38754681e-01 3.13051417e-02 2.71323860e-01
3.33849967e-01 -5.75532496e-01 -1.23735927e-01 -1.77235201e-01
6.80474997e-01 3.99488747e-01 5.66932619e-01 3.52931470e-02
4.83606964e-01 -3.05821989e-02 8.01035941e-01 8.16019893e-01
-1.15232281e-01 1.25587499e+00 5.00096194e-02 -6.83559418e-01
-1.16832972e+00 -9.15146768e-01 2.30931938e-01 5.46516955e-01
3.92827421e-01 -6.88644946e-02 -6.35352910e-01 -7.64336288e-01
7.50155523e-02 4.06867981e-01 -2.38685712e-01 3.57215889e-02
-6.43524289e-01 -2.83286124e-01 -8.81770104e-02 6.50961459e-01
3.97455245e-01 -5.93888164e-01 -6.19696081e-01 2.50683784e-01
-8.36332440e-02 -1.22241974e+00 -8.62138212e-01 2.72027194e-01
-8.80405903e-01 -1.15180480e+00 -7.10676014e-01 -5.85267484e-01
9.83859539e-01 5.08651733e-01 1.10600471e+00 -8.52217972e-02
8.30101371e-02 6.06861532e-01 -2.06428736e-01 -1.85074046e-01
-2.90300310e-01 -2.74882704e-01 3.58370483e-01 2.09893137e-01
1.32621631e-01 -4.00460094e-01 -3.24371010e-01 6.93833888e-01
-6.95859194e-01 -1.63079873e-01 6.26858234e-01 9.58817303e-01
9.98850882e-01 -9.56032798e-02 -6.79322407e-02 -6.03545368e-01
4.07104313e-01 -2.17739299e-01 -1.19399679e+00 1.17974751e-01
-1.63379028e-01 9.80085805e-02 7.21412957e-01 -5.47892511e-01
-7.58001089e-01 8.92151654e-01 -3.41781117e-02 -1.27403259e+00
-5.12394726e-01 4.58222628e-01 -1.19162403e-01 -2.26080313e-01
7.39840865e-01 -7.19569344e-03 7.22440258e-02 -6.00753188e-01
3.45647931e-01 5.70747316e-01 1.04444540e+00 -6.94491148e-01
9.77321327e-01 5.20015955e-01 -6.58093840e-02 -5.33478498e-01
-9.43487167e-01 -9.84404922e-01 -1.05673242e+00 -7.32579455e-02
7.94246256e-01 -9.83179629e-01 -4.92619962e-01 4.75601912e-01
-1.60495746e+00 -3.63327824e-02 -1.30157143e-01 7.91420817e-01
-7.90463507e-01 4.73548383e-01 -1.48801669e-01 -7.41004944e-01
-1.19610287e-01 -1.44027209e+00 1.49434483e+00 2.14466557e-01
-1.51414961e-01 -8.42211366e-01 1.92266718e-01 2.14597434e-01
2.01960113e-02 2.62571186e-01 2.83252131e-02 -6.94242239e-01
-6.24830246e-01 -4.55578357e-01 -1.03788763e-01 4.61027265e-01
-1.38917193e-02 6.47813901e-02 -8.85566175e-01 -5.04422784e-01
2.30773032e-01 -2.82674015e-01 6.79771483e-01 5.00632584e-01
1.15866411e+00 9.12131444e-02 -4.04526532e-01 7.66937256e-01
1.42247200e+00 1.13582671e-01 2.27174416e-01 2.92613000e-01
6.89125776e-01 4.20831084e-01 5.81267476e-01 2.56685108e-01
1.67218775e-01 6.81826055e-01 6.36638582e-01 -2.24587303e-02
4.06871915e-01 -3.06528956e-01 1.67360947e-01 7.32491076e-01
8.55669007e-03 -2.23637283e-01 -9.27045882e-01 3.69341284e-01
-2.13774753e+00 -5.44243634e-01 8.25813264e-02 2.36307144e+00
3.43784004e-01 1.30696282e-01 -1.16956621e-01 -2.98867375e-01
9.13246214e-01 -2.01816648e-01 -7.54788756e-01 8.82418081e-02
4.98880707e-02 -2.76044458e-01 6.52891457e-01 6.63762152e-01
-1.24921381e+00 6.77937090e-01 5.27909327e+00 5.31136632e-01
-1.08457148e+00 -1.21833056e-01 3.30723822e-01 4.34243917e-01
1.67606354e-01 1.02618009e-01 -8.44311953e-01 3.61770928e-01
3.75067919e-01 3.29891741e-01 5.42095304e-01 1.12521935e+00
1.82928070e-02 -1.30107313e-01 -1.44929719e+00 1.48273122e+00
3.98321390e-01 -1.15711725e+00 -1.98104143e-01 -1.11681774e-01
8.82637680e-01 3.16853523e-01 -1.17885478e-01 -1.36512727e-01
6.14194393e-01 -9.26482260e-01 6.77199304e-01 7.60688365e-01
5.87888360e-01 -4.46119279e-01 1.01732481e+00 8.85613143e-01
-9.40948486e-01 -3.35241780e-02 -5.70923209e-01 3.86356622e-01
2.48723820e-01 4.50910985e-01 -8.59181464e-01 7.14815080e-01
8.09185743e-01 7.96743631e-01 -2.64461458e-01 1.56853569e+00
-2.07781851e-01 -3.17098945e-03 -8.09623539e-01 3.38586509e-01
3.04888010e-01 -5.43812692e-01 7.14839339e-01 7.25352347e-01
6.77646816e-01 4.80480567e-02 5.05784571e-01 9.28664684e-01
3.19621302e-02 -3.54265600e-01 -7.01292455e-01 4.48924214e-01
5.79376042e-01 1.31266010e+00 -6.73821270e-01 -6.00970946e-02
-2.65696347e-01 1.20032680e+00 3.59860510e-01 3.81116927e-01
-4.36253339e-01 -6.29534721e-02 4.03569043e-01 -2.90191829e-01
4.27613676e-01 -3.79485101e-01 -1.80707742e-02 -1.19315696e+00
1.32108167e-01 -1.03452611e+00 2.68108577e-01 -1.23152268e+00
-1.69864595e+00 6.57276928e-01 -1.42211914e-02 -1.68433285e+00
-2.54180610e-01 -8.08523834e-01 -5.60466766e-01 9.33166265e-01
-1.23897052e+00 -9.34467793e-01 -5.62235653e-01 3.82343143e-01
7.06599116e-01 -6.47383928e-02 5.34554541e-01 4.48855348e-02
-3.06903958e-01 9.27868187e-02 1.45499483e-01 1.32596985e-01
9.08187866e-01 -1.38294530e+00 5.32537818e-01 7.66484737e-01
4.20935035e-01 6.98647618e-01 7.59784579e-01 -4.84273911e-01
-1.41486895e+00 -9.72670496e-01 6.09713197e-01 -6.50486946e-01
4.25768673e-01 -3.27041686e-01 -9.57506657e-01 8.59274983e-01
-3.11058369e-02 4.85544980e-01 7.73751363e-02 -2.99090624e-01
-2.56494284e-01 -1.26349404e-01 -1.10784662e+00 4.62435663e-01
8.85239661e-01 -4.11510348e-01 -6.34724796e-01 2.48805448e-01
5.09851575e-01 -9.68120575e-01 -6.71194077e-01 2.89462954e-01
2.52652109e-01 -1.05521894e+00 1.00504112e+00 -4.12759423e-01
-4.22847494e-02 -8.61872256e-01 -3.40640128e-01 -1.82042170e+00
-2.48878255e-01 -8.62387180e-01 1.25025183e-01 8.47978532e-01
1.39828652e-01 -5.52583694e-01 7.49090433e-01 3.67449582e-01
-2.30626032e-01 -4.28855598e-01 -1.06379533e+00 -6.90153360e-01
-1.15684085e-01 -5.08252859e-01 5.90782166e-01 8.18034470e-01
-8.20840955e-01 3.69102746e-01 -4.16688144e-01 7.11751819e-01
7.99569130e-01 2.26681411e-01 1.16514230e+00 -1.47370923e+00
-3.53273511e-01 -1.85854912e-01 -3.45773697e-01 -1.54430115e+00
1.67144567e-01 -4.54873353e-01 5.37627757e-01 -1.35430503e+00
-9.03186500e-02 -4.01531070e-01 1.39964372e-01 1.04078181e-01
-2.51816660e-02 -9.73123759e-02 2.64937222e-01 2.76869595e-01
-7.36325562e-01 3.77007842e-01 1.21051550e+00 3.69138010e-02
-3.34058195e-01 3.97632539e-01 -2.31940567e-01 1.14661241e+00
5.67887366e-01 -3.83376449e-01 -2.31364310e-01 -7.34554887e-01
5.62743805e-02 2.28361145e-01 5.00985026e-01 -1.04999053e+00
6.86682105e-01 -1.61840647e-01 6.51859879e-01 -8.67633998e-01
4.68712658e-01 -1.49110603e+00 3.80392551e-01 8.84427354e-02
-6.12950735e-02 3.50678444e-01 -6.57753274e-02 6.17720187e-01
-2.78005540e-01 -4.92418438e-01 7.58065879e-01 -3.39009911e-01
-7.67617643e-01 4.24925059e-01 1.76003918e-01 -1.41352981e-01
8.82320642e-01 -3.68838757e-01 -1.76831689e-02 -4.57970202e-01
-7.63877153e-01 4.55287695e-01 6.91211462e-01 3.07001501e-01
9.58874702e-01 -1.29934120e+00 -6.11855745e-01 4.90155220e-01
1.71402231e-01 7.66882122e-01 -4.47243489e-02 7.79684305e-01
-4.94650245e-01 6.33755624e-02 3.46638784e-02 -1.13865983e+00
-1.11850715e+00 5.39170742e-01 7.73360372e-01 2.40310077e-02
-4.15001035e-01 8.68575990e-01 -6.18866794e-02 -1.01951766e+00
4.20900643e-01 -4.49785501e-01 9.12317634e-02 -2.74952620e-01
1.71385035e-01 2.70554334e-01 4.92987223e-02 -7.78808534e-01
-2.02414632e-01 1.25476050e+00 1.19620614e-01 1.63155235e-02
1.33381701e+00 -1.96370512e-01 -7.64626637e-02 3.46997380e-01
1.20428932e+00 -7.82108009e-02 -1.70072997e+00 -4.02131498e-01
7.36269280e-02 -6.15839362e-01 1.75031647e-02 -5.25254011e-01
-1.02757370e+00 6.64448559e-01 5.00709236e-01 -5.65307736e-02
9.39582407e-01 4.56592664e-02 5.70938051e-01 7.30298340e-01
4.58247602e-01 -1.19824958e+00 1.61874369e-01 5.65795898e-01
1.14504910e+00 -1.49192202e+00 -8.79526138e-02 -2.47216821e-01
-7.16541708e-01 1.34792173e+00 8.00923944e-01 -5.34426212e-01
5.23418784e-01 2.02423424e-01 2.89590091e-01 -6.34372188e-03
-4.63217139e-01 -1.30328670e-01 4.49302703e-01 4.81375843e-01
-2.63437539e-01 -2.79623508e-01 3.72159958e-01 3.30951214e-01
-1.23459324e-01 4.70600575e-02 2.85977006e-01 8.03531408e-01
-5.27404904e-01 -8.35511386e-01 -9.87799287e-01 7.80953839e-02
4.16700318e-02 4.09652323e-01 -4.19615597e-01 7.14709282e-01
3.63004237e-01 6.74937189e-01 2.59143561e-01 -2.71145523e-01
5.17789483e-01 -1.42691180e-01 4.54000264e-01 -4.81370270e-01
-2.98820049e-01 5.98565936e-01 -3.61192822e-01 -7.04375565e-01
-5.22329330e-01 -6.49137497e-01 -1.02511477e+00 1.06375717e-01
-5.67899585e-01 1.19261548e-01 7.51881182e-01 9.44451630e-01
2.72473663e-01 -6.14495389e-02 7.13832974e-01 -1.16953766e+00
-9.42913055e-01 -7.15252876e-01 -3.36043507e-01 4.12339985e-01
7.52734840e-01 -6.87568605e-01 -5.03580451e-01 1.30990759e-01] | [7.834887504577637, -2.3239803314208984] |
118c81e7-a178-4f56-92df-1216c51118f6 | style-invariant-cardiac-image-segmentation | 2009.12193 | null | https://arxiv.org/abs/2009.12193v1 | https://arxiv.org/pdf/2009.12193v1.pdf | Style-invariant Cardiac Image Segmentation with Test-time Augmentation | Deep models often suffer from severe performance drop due to the appearance shift in the real clinical setting. Most of the existing learning-based methods rely on images from multiple sites/vendors or even corresponding labels. However, collecting enough unknown data to robustly model segmentation cannot always hold since the complex appearance shift caused by imaging factors in daily application. In this paper, we propose a novel style-invariant method for cardiac image segmentation. Based on the zero-shot style transfer to remove appearance shift and test-time augmentation to explore diverse underlying anatomy, our proposed method is effective in combating the appearance shift. Our contribution is three-fold. First, inspired by the spirit of universal style transfer, we develop a zero-shot stylization for content images to generate stylized images that appearance similarity to the style images. Second, we build up a robust cardiac segmentation model based on the U-Net structure. Our framework mainly consists of two networks during testing: the ST network for removing appearance shift and the segmentation network. Third, we investigate test-time augmentation to explore transformed versions of the stylized image for prediction and the results are merged. Notably, our proposed framework is fully test-time adaptation. Experiment results demonstrate that our methods are promising and generic for generalizing deep segmentation models. | ['Yuxin Zou', 'Xin Yang', 'Mingyuan Luo', 'Xiaoqiong Huang', 'Zhendong Liu', 'Zejian Chen', 'Wufeng Xue', 'Dong Ni'] | 2020-09-24 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [ 4.63697225e-01 1.72744185e-01 -9.83033255e-02 -4.82481271e-01
-7.04693675e-01 -5.17709076e-01 2.03250833e-02 -5.35954654e-01
-1.49762735e-01 6.25717223e-01 -1.24064930e-01 -3.41246784e-01
8.20207745e-02 -5.76082706e-01 -8.39612663e-01 -5.38756311e-01
4.39318448e-01 3.66133034e-01 4.25706744e-01 -1.67013466e-01
-1.29499912e-01 3.38476390e-01 -7.88429558e-01 2.14171469e-01
1.21673226e+00 8.76327097e-01 1.47093087e-01 4.96629268e-01
-1.56787947e-01 4.31763232e-01 -5.84926903e-01 -5.18579125e-01
5.37529111e-01 -7.83001542e-01 -8.42089355e-01 5.43066859e-01
4.00892526e-01 -6.17862463e-01 -1.35910124e-01 1.06884146e+00
8.20365727e-01 -2.29031849e-03 4.50015992e-01 -1.08392203e+00
-8.66360128e-01 5.65768242e-01 -7.08109677e-01 2.14419797e-01
-8.49707052e-02 3.44187707e-01 3.60534489e-01 -6.62302792e-01
8.15822601e-01 1.05863774e+00 7.16909111e-01 9.05582547e-01
-1.19692910e+00 -6.45334423e-01 2.73311377e-01 -8.13411251e-02
-1.02534759e+00 -8.15360695e-02 1.08965778e+00 -4.87247020e-01
2.23864708e-02 1.84330627e-01 6.80405915e-01 1.39591157e+00
2.33027697e-01 9.91073906e-01 1.32159138e+00 -2.93860644e-01
1.91527302e-03 2.04884812e-01 4.80157435e-02 8.20551991e-01
1.59561634e-01 -1.14734948e-01 7.40933716e-02 9.56123546e-02
1.21143866e+00 2.20910266e-01 -4.14697319e-01 -7.61783600e-01
-1.24584353e+00 4.90936220e-01 5.82930982e-01 2.85056978e-01
-9.11059529e-02 -8.83046314e-02 5.54528654e-01 2.13779256e-01
4.74308640e-01 3.41731578e-01 -4.67817545e-01 1.34336710e-01
-1.09973443e+00 -1.74034953e-01 5.13379991e-01 1.17951226e+00
3.95182401e-01 2.81285316e-01 -7.25754142e-01 1.01724744e+00
2.44514085e-02 3.09761941e-01 7.00870812e-01 -8.21500301e-01
2.15971678e-01 4.85409856e-01 -2.90020764e-01 -6.54643834e-01
-4.15219188e-01 -8.08007598e-01 -1.08985460e+00 8.36873576e-02
4.14126635e-01 -4.01032627e-01 -1.34409821e+00 1.75935137e+00
2.96814531e-01 6.43968403e-01 -2.55933762e-01 9.67426419e-01
9.59502161e-01 6.34956136e-02 9.17252153e-02 -1.44972771e-01
1.15433705e+00 -1.29374909e+00 -7.63821959e-01 -8.07996467e-03
6.88323975e-01 -5.89917958e-01 1.39612770e+00 1.99944764e-01
-1.13688242e+00 -8.04830790e-01 -1.18866563e+00 1.94604665e-01
-4.80884351e-02 -9.28502157e-02 3.88629407e-01 9.27650034e-01
-1.01270771e+00 7.13662207e-01 -8.87479305e-01 -1.10117190e-01
7.77661443e-01 2.50542849e-01 -4.64737080e-02 -1.56510368e-01
-9.73256171e-01 5.52851915e-01 2.88902849e-01 2.86775291e-01
-7.87097991e-01 -9.69694734e-01 -8.20961475e-01 -1.96893096e-01
5.68453789e-01 -9.46871042e-01 1.15433097e+00 -1.37717688e+00
-1.70242822e+00 9.59464848e-01 1.61213949e-01 -1.55270621e-01
1.04188013e+00 -1.02639422e-02 -3.22615981e-01 1.63287506e-01
2.26328686e-01 6.29038513e-01 8.16878140e-01 -1.44609249e+00
-2.03470200e-01 -2.44178325e-01 -1.53842464e-01 1.13826007e-01
-2.99924850e-01 -1.50188431e-01 -8.84785116e-01 -1.16609681e+00
8.07678476e-02 -1.10469401e+00 -5.11548638e-01 2.80427456e-01
-5.05834162e-01 4.77515012e-01 7.33614862e-01 -7.61463284e-01
1.06555963e+00 -2.10996604e+00 1.97285488e-02 2.77018517e-01
3.84808242e-01 3.29470366e-01 -2.59074479e-01 -2.31566876e-01
-2.21531913e-01 2.91510731e-01 -6.52042031e-01 -2.42343798e-01
-2.57890105e-01 3.01645070e-01 -1.18118405e-01 2.72543222e-01
2.78832048e-01 1.22540867e+00 -9.46128190e-01 -9.19754148e-01
1.71481207e-01 3.22891712e-01 -7.65229762e-01 1.65759295e-01
-3.38374637e-02 1.29486632e+00 -6.43650413e-01 7.92134404e-01
8.45367193e-01 -3.18824798e-01 1.36940807e-01 -2.79069901e-01
3.50845069e-01 -4.60078210e-01 -9.18853104e-01 2.25571704e+00
-3.11744899e-01 -4.64897975e-02 -5.29419407e-02 -9.46247756e-01
7.25955784e-01 3.51742655e-01 6.74845159e-01 -6.42791033e-01
3.94378483e-01 3.22629064e-01 1.58710226e-01 -4.82328176e-01
-6.06043078e-02 -2.83828735e-01 8.30815956e-02 2.91860044e-01
1.89519703e-01 -2.15832755e-01 9.91872773e-02 1.60396591e-01
7.76969910e-01 4.60467756e-01 -1.40296727e-01 -2.75894612e-01
5.44123292e-01 -3.07491899e-01 9.60272551e-01 6.66893482e-01
-6.92183316e-01 1.04951692e+00 5.50194740e-01 -4.70212102e-01
-9.54681158e-01 -1.09570277e+00 -2.70056099e-01 7.69236147e-01
3.31335783e-01 -4.23776135e-02 -1.16472876e+00 -1.02321494e+00
-4.10902321e-01 3.70441020e-01 -8.58888865e-01 -3.48733306e-01
-6.65414989e-01 -8.97996962e-01 5.41734576e-01 7.40099132e-01
6.76015735e-01 -1.16197288e+00 -3.21883559e-01 2.43770793e-01
-1.70891672e-01 -1.13126266e+00 -8.86202693e-01 -1.81860596e-01
-1.21834445e+00 -9.27227437e-01 -1.26784360e+00 -8.66394937e-01
8.76368105e-01 1.15218632e-01 1.07853520e+00 1.90821290e-01
-4.31152046e-01 4.21234906e-01 -4.02618289e-01 -4.70131427e-01
-4.91423547e-01 -2.27402486e-02 -2.44360223e-01 2.33814880e-01
-2.27121443e-01 -6.36409342e-01 -9.80899572e-01 4.49992269e-01
-1.03334892e+00 2.38812655e-01 7.81704545e-01 1.00699937e+00
9.15356159e-01 -5.51041842e-01 5.94145477e-01 -1.55139887e+00
5.22412062e-01 -2.33179733e-01 -1.50898278e-01 5.23027837e-01
-5.84666491e-01 -6.91087618e-02 5.35044909e-01 -6.52843058e-01
-1.17471111e+00 6.10345788e-02 -1.12689614e-01 -8.25923204e-01
-1.28068119e-01 2.34786630e-01 -7.30120018e-02 -3.21320593e-01
5.55472672e-01 2.04418436e-01 1.67454466e-01 -3.90226573e-01
4.07417357e-01 1.98783100e-01 6.01739645e-01 -8.00030351e-01
8.51632595e-01 5.57532668e-01 -1.01563923e-01 -3.82167429e-01
-1.01637018e+00 -1.08585283e-01 -1.09003520e+00 -3.18451345e-01
1.00803685e+00 -6.66421473e-01 -4.19080891e-02 5.25139570e-01
-9.62611377e-01 -5.62342405e-01 -4.30082411e-01 2.58547783e-01
-6.60943151e-01 6.88574612e-01 -8.42757761e-01 -2.74439842e-01
-6.39652491e-01 -1.63638771e+00 9.49963391e-01 2.46803090e-01
7.46792853e-02 -1.05934358e+00 8.20059702e-02 3.37199777e-01
3.60166728e-01 6.29393041e-01 8.52748096e-01 -5.40380597e-01
-4.53892231e-01 4.43481952e-02 -3.16461891e-01 5.28745055e-01
3.21736068e-01 -1.83854222e-01 -9.04545426e-01 -4.06791985e-01
1.89652964e-01 -2.14890897e-01 6.07800901e-01 6.95092678e-01
1.65470219e+00 2.02550322e-01 -1.77389294e-01 1.21792758e+00
1.19731963e+00 1.62425220e-01 6.01941228e-01 2.33232960e-01
1.09219205e+00 4.82900470e-01 4.07567590e-01 1.20797969e-01
8.20105299e-02 4.62443441e-01 2.08560213e-01 -8.68340969e-01
-3.62525851e-01 -5.34744449e-02 -2.71961819e-02 1.05656958e+00
-1.39854580e-01 1.16954207e-01 -8.31815600e-01 4.32219654e-01
-1.72638011e+00 -5.46413422e-01 -9.27835796e-03 2.05845213e+00
8.86398256e-01 2.17530191e-01 9.69407409e-02 -3.86428177e-01
7.65846670e-01 -1.29747152e-01 -8.42764556e-01 -2.00847343e-01
1.27519742e-02 4.21115667e-01 4.11326140e-01 2.88122557e-02
-1.01718855e+00 8.71296585e-01 6.08392048e+00 8.25061023e-01
-1.22606993e+00 4.37400162e-01 1.13318825e+00 2.08345335e-02
-1.69794321e-01 -3.27347338e-01 -2.62824267e-01 4.97759402e-01
5.78355432e-01 -2.57742405e-02 2.63305921e-02 6.99890018e-01
1.36924058e-01 4.71370906e-01 -9.40436900e-01 9.80626941e-01
2.89800555e-01 -1.12124097e+00 2.41318107e-01 -6.89907596e-02
9.73389864e-01 -2.70780057e-01 2.72435874e-01 4.46099609e-01
5.61430044e-02 -8.70964408e-01 5.63944638e-01 5.15907168e-01
1.23104072e+00 -5.08849800e-01 7.45886803e-01 5.25193252e-02
-9.39711511e-01 1.72108576e-01 -2.16583222e-01 5.41236818e-01
1.86946765e-01 3.49116117e-01 -7.29481578e-01 8.31740141e-01
5.55838585e-01 7.33835042e-01 -7.53015101e-01 1.11397719e+00
-3.15598142e-03 6.11593366e-01 1.28013864e-01 5.43366313e-01
3.08288276e-01 -4.02130306e-01 4.06197757e-01 1.07737041e+00
2.71092474e-01 9.20402408e-02 4.77133721e-01 9.97061074e-01
-1.35305002e-01 3.47894907e-01 -5.32423079e-01 4.68408346e-01
4.43046317e-02 1.41412270e+00 -1.11328995e+00 -5.86255491e-01
-5.91429949e-01 1.28640926e+00 1.43606430e-02 4.68473226e-01
-1.05216312e+00 -2.78146602e-02 3.23959514e-02 2.38663733e-01
8.55544135e-02 2.21829399e-01 -4.83897805e-01 -1.30347204e+00
5.43180108e-02 -9.44439590e-01 3.67219716e-01 -5.98750710e-01
-1.42149508e+00 8.80145967e-01 -7.37989619e-02 -1.64364564e+00
1.52133629e-01 -4.73149151e-01 -8.56044531e-01 8.75309706e-01
-1.46016812e+00 -1.44062436e+00 -3.75869513e-01 7.38607645e-01
7.85716832e-01 -8.83469433e-02 6.36731386e-01 5.08461356e-01
-8.19947720e-01 9.56937075e-01 -4.28108424e-02 3.94983858e-01
1.09843266e+00 -1.32438099e+00 3.12865973e-01 9.00964797e-01
-1.55293718e-01 6.60996854e-01 2.74054378e-01 -8.30803931e-01
-9.05321538e-01 -1.22287416e+00 -1.56978950e-01 -4.75089043e-01
6.01441622e-01 -2.45134369e-01 -1.10110664e+00 7.56436110e-01
1.10726014e-01 3.76071990e-01 7.77336121e-01 -2.38915741e-01
-1.89498290e-02 -6.61926568e-02 -1.12315845e+00 8.12194824e-01
1.02255929e+00 -1.14214897e-01 -5.02623975e-01 2.58003086e-01
9.06845391e-01 -7.92945623e-01 -9.57084239e-01 6.49491549e-01
4.22187686e-01 -6.68042123e-01 8.97812128e-01 -6.96832776e-01
5.43656111e-01 -5.65009750e-02 3.43280882e-01 -1.18388355e+00
-2.88819969e-01 -5.41556180e-01 2.19596058e-01 1.18851912e+00
4.56699312e-01 -4.87650454e-01 8.99085760e-01 7.58618534e-01
-5.84128916e-01 -8.38641167e-01 -7.76793301e-01 -6.59673095e-01
4.04844165e-01 -2.24330410e-01 3.97697300e-01 1.20388663e+00
-4.65033859e-01 1.90261930e-01 -4.49054509e-01 -9.38693509e-02
7.61784315e-01 1.11823650e-02 7.12911606e-01 -1.14149582e+00
-5.01711190e-01 -3.90773088e-01 -1.17819898e-01 -8.11972678e-01
-4.90516946e-02 -9.81995404e-01 3.28686759e-02 -1.31152606e+00
3.97482246e-01 -4.88712430e-01 -7.28645861e-01 3.61348748e-01
-4.34507787e-01 4.48693067e-01 2.20290899e-01 1.79275841e-01
-5.14557838e-01 5.27016699e-01 2.27779937e+00 -1.83508381e-01
-2.33596861e-01 1.01288985e-02 -7.43566692e-01 7.36226797e-01
7.40345180e-01 -2.97019571e-01 -7.33245909e-01 -4.83358711e-01
-2.00746626e-01 1.51002094e-01 2.90866286e-01 -8.96399736e-01
-1.72932342e-01 2.04818934e-01 4.80273038e-01 -3.31454635e-01
-8.64875615e-02 -7.97188103e-01 -1.20398715e-01 6.03846669e-01
-1.96074754e-01 -7.25738555e-02 3.82374972e-01 6.32959902e-01
-1.20987318e-01 -2.18080074e-01 1.03958213e+00 -3.08290571e-01
-5.42828500e-01 7.79059350e-01 1.48064524e-01 3.32529545e-01
9.64339674e-01 -5.06226718e-01 3.28522958e-02 -9.36066583e-02
-1.02517617e+00 3.93797696e-01 3.72578055e-01 3.48746449e-01
4.36384022e-01 -1.31250620e+00 -7.39589334e-01 2.12233573e-01
-3.34913284e-02 2.47706547e-01 7.26258993e-01 1.23931193e+00
-7.28147984e-01 -3.34372111e-02 -4.95099694e-01 -6.73444450e-01
-9.55609083e-01 7.15188026e-01 4.97552007e-01 -5.18529236e-01
-9.42616522e-01 8.09993804e-01 9.24987376e-01 -6.10102773e-01
-2.00390872e-02 -4.05343324e-01 -2.81039253e-02 -3.49393368e-01
2.79728383e-01 -9.40427333e-02 3.66750220e-03 -3.96385938e-01
-1.07064277e-01 7.57997394e-01 -3.35199684e-01 -3.43556032e-02
1.22648358e+00 -1.27186507e-01 6.26673922e-02 3.99659514e-01
8.88173163e-01 -2.85168797e-01 -1.42718434e+00 -2.97368914e-01
-2.84435749e-01 -1.84300080e-01 -9.74936113e-02 -9.36592877e-01
-1.46994388e+00 9.59892213e-01 8.88352990e-01 -2.55816996e-01
1.21417880e+00 -4.41986501e-01 1.03675056e+00 -1.69945866e-01
2.24000379e-01 -1.06501174e+00 1.56877831e-01 1.04392648e-01
7.59591997e-01 -1.42386830e+00 -3.20949495e-01 -6.33596838e-01
-9.20670688e-01 9.81733739e-01 9.09761548e-01 -1.45642057e-01
6.22720599e-01 1.04728684e-01 4.47237909e-01 -1.27586216e-01
-1.22107379e-01 4.63575907e-02 5.24025559e-01 6.87186182e-01
4.31450218e-01 -1.45753637e-01 -3.71192932e-01 8.68014574e-01
8.84147510e-02 1.13972604e-01 4.67982620e-01 7.82878935e-01
7.40414485e-02 -1.21036029e+00 -3.96384329e-01 4.19708192e-01
-5.86487770e-01 -1.34523347e-01 -1.85866043e-01 8.14540863e-01
1.77445441e-01 3.97529364e-01 -1.60694674e-01 -2.64950395e-01
5.12813568e-01 1.02212176e-01 6.13237560e-01 -8.39966953e-01
-5.33956110e-01 5.14460683e-01 -5.31177461e-01 -4.78388608e-01
-3.44254434e-01 -3.77028316e-01 -1.11200786e+00 1.17729425e-01
-1.75749689e-01 -2.91779935e-01 1.71298131e-01 8.60402465e-01
1.79670423e-01 1.14970124e+00 5.74051321e-01 -7.04966009e-01
-4.22312111e-01 -9.50900555e-01 -6.22764409e-01 8.29419911e-01
3.79038192e-02 -6.81856990e-01 1.50505766e-01 3.49149644e-01] | [14.52486801147461, -2.1699790954589844] |
d18131c6-5ba1-4869-81b0-fe8d54fd56c3 | deep-semi-supervised-metric-learning-with-1 | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhuang_Deep_Semi-Supervised_Metric_Learning_With_Mixed_Label_Propagation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhuang_Deep_Semi-Supervised_Metric_Learning_With_Mixed_Label_Propagation_CVPR_2023_paper.pdf | Deep Semi-Supervised Metric Learning With Mixed Label Propagation | Metric learning requires the identification of far-apart similar pairs and close dissimilar pairs during training, and this is difficult to achieve with unlabeled data because pairs are typically assumed to be similar if they are close. We present a novel metric learning method which circumvents this issue by identifying hard negative pairs as those which obtain dissimilar labels via label propagation (LP), when the edge linking the pair of data is removed in the affinity matrix. In so doing, the negative pairs can be identified despite their proximity, and we are able to utilize this information to significantly improve LP's ability to identify far-apart positive pairs and close negative pairs. This results in a considerable improvement in semi-supervised metric learning performance as evidenced by recall, precision and Normalized Mutual Information (NMI) performance metrics on Content-based Information Retrieval (CBIR) applications. | ['Pierre Moulin', 'Furen Zhuang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [ 9.83686745e-02 -1.33077174e-01 -2.40920946e-01 -7.18257904e-01
-8.94460440e-01 -8.65361333e-01 5.18436670e-01 7.41146028e-01
-5.34079492e-01 7.99432814e-01 -2.29484349e-01 1.00680636e-02
-6.17654443e-01 -7.88599312e-01 -1.40674040e-01 -6.68127179e-01
-1.95716083e-01 9.36271727e-01 2.76355714e-01 -2.45278068e-02
4.54302460e-01 5.74760914e-01 -1.50708234e+00 1.55735835e-01
8.77694547e-01 7.51034081e-01 -5.26397936e-02 3.47253650e-01
-2.17991784e-01 6.20183110e-01 -3.30532491e-01 -4.95183200e-01
3.77168000e-01 -4.99087006e-01 -9.68928397e-01 -3.92247677e-01
3.49284887e-01 5.94044812e-02 4.68853787e-02 1.10068476e+00
2.42442608e-01 2.70076334e-01 1.17105353e+00 -1.36974967e+00
-4.38998848e-01 2.61388987e-01 -7.54589975e-01 4.65195514e-02
3.53139400e-01 -6.43782914e-01 1.36448467e+00 -1.05249298e+00
5.34828484e-01 9.45716441e-01 8.31431568e-01 3.88182700e-01
-1.29721868e+00 -6.70717955e-01 -1.84267342e-01 3.95135105e-01
-1.87083769e+00 -1.36964425e-01 7.25138664e-01 -3.44995916e-01
6.57692432e-01 2.82306880e-01 2.71819741e-01 3.36190522e-01
-2.09280565e-01 5.12565374e-01 9.64086235e-01 -7.62774050e-01
2.45001554e-01 4.43769932e-01 4.35841620e-01 5.97392797e-01
1.28725380e-01 6.99209943e-02 -2.96629906e-01 -2.78254390e-01
1.87956065e-01 2.42415015e-02 -1.48554891e-01 -8.20689619e-01
-1.13227904e+00 9.68557298e-01 5.69364488e-01 5.82850933e-01
-2.46716961e-02 -3.91023725e-01 2.21810848e-01 6.43238902e-01
3.66717994e-01 7.98815846e-01 -3.76069456e-01 2.82505974e-02
-7.90762246e-01 2.50451919e-02 9.99418437e-01 7.98471749e-01
1.15477407e+00 -8.86620879e-01 -4.18670699e-02 1.15720463e+00
3.11760694e-01 2.62819588e-01 6.60648584e-01 -8.19368422e-01
3.02516490e-01 9.02550220e-01 2.48635009e-01 -1.25506282e+00
-4.56382245e-01 -1.52240366e-01 -6.61119819e-01 1.37269512e-01
4.52220500e-01 6.01916909e-02 -6.83692634e-01 1.83428848e+00
3.42079580e-01 2.91058458e-02 9.39944237e-02 9.17283297e-01
6.48525596e-01 2.83999979e-01 -2.35426221e-02 -2.68684357e-01
8.52820933e-01 -8.96769345e-01 -3.79363567e-01 -6.20636120e-02
1.18767405e+00 -7.59473979e-01 8.30722153e-01 1.73610345e-01
-7.22525716e-01 -3.46395642e-01 -1.38324893e+00 7.83937797e-02
-6.33786976e-01 -1.90855965e-01 5.82623184e-01 5.13373733e-01
-8.75123084e-01 8.27296495e-01 -5.79887509e-01 -4.42970097e-01
2.03260526e-01 6.59316957e-01 -6.92084968e-01 -3.95763457e-01
-1.22868884e+00 1.01127362e+00 3.30537289e-01 -2.00654835e-01
-2.21501708e-01 -5.27700603e-01 -7.61906207e-01 5.95695525e-02
-2.20719222e-02 -3.87937605e-01 9.34340358e-01 -8.55704963e-01
-7.41408825e-01 1.22038150e+00 -1.53534725e-01 -1.34997651e-01
4.85035539e-01 3.30099702e-01 -5.82237124e-01 8.08060467e-02
1.99837863e-01 6.88311517e-01 3.83322805e-01 -1.28741837e+00
-8.22642148e-01 -5.66963553e-01 -2.24195272e-01 5.60083330e-01
-7.91388988e-01 3.14098410e-02 -4.07620579e-01 -3.96334797e-01
6.00612104e-01 -1.02574682e+00 -1.62828118e-02 1.91512704e-01
-1.22452237e-01 -5.00342727e-01 8.42580199e-01 -1.81122541e-01
1.19999397e+00 -2.13342476e+00 -1.09579913e-01 8.39733243e-01
3.70510429e-01 4.53045994e-01 -3.20207387e-01 4.46737081e-01
-9.21886191e-02 8.94071683e-02 -2.54754186e-01 -2.27895066e-01
-1.08270749e-01 3.65688443e-01 2.69958347e-01 5.01329839e-01
1.82799995e-01 4.80435699e-01 -1.32551444e+00 -6.70309663e-01
1.48270264e-01 2.68732637e-01 -5.68409748e-02 2.56306261e-01
3.59409750e-01 2.29684860e-01 -3.85527939e-01 4.22660649e-01
6.25668764e-01 -3.06396991e-01 2.85783529e-01 -1.27725750e-01
4.87279147e-01 2.37048101e-02 -1.23676932e+00 1.31484330e+00
-4.49415267e-01 6.70825899e-01 -3.77823591e-01 -1.22594869e+00
1.24282300e+00 2.21277773e-01 6.69158101e-01 -9.49000239e-01
-4.35363837e-02 3.53267103e-01 -3.19783203e-02 -3.20459843e-01
3.05087745e-01 -3.14267188e-01 1.88535541e-01 7.92625904e-01
-2.56359249e-01 -2.40435265e-02 3.72635603e-01 2.55264729e-01
1.00257885e+00 -4.61751550e-01 2.67332911e-01 -2.58983374e-01
7.58056760e-01 -9.63826850e-02 5.47974467e-01 6.49919212e-01
-3.23047161e-01 8.59887123e-01 1.72516361e-01 -1.28627434e-01
-8.95069718e-01 -1.48722935e+00 -3.44525486e-01 9.30198193e-01
5.52788615e-01 -3.87447417e-01 -3.94244254e-01 -1.16156840e+00
2.52401322e-01 3.93095285e-01 -6.83089674e-01 -4.91881460e-01
-3.98029923e-01 -4.96356636e-01 2.65224785e-01 5.21407425e-01
-9.72258858e-03 -6.98696256e-01 3.44400853e-01 -8.95052962e-03
-1.71659634e-01 -6.10994935e-01 -4.03834283e-01 5.87088168e-01
-8.13224733e-01 -1.46171641e+00 -6.16645515e-01 -1.27542174e+00
9.84446764e-01 4.05520409e-01 1.05340743e+00 2.58614480e-01
-9.08159763e-02 5.82640478e-03 -3.92782986e-01 -1.17343599e-02
-4.59786773e-01 -3.79390619e-03 2.37195566e-01 -9.68309417e-02
9.96365964e-01 -3.98316413e-01 -4.26790237e-01 9.20251787e-01
-6.49599910e-01 -4.19137716e-01 3.36356521e-01 1.11705077e+00
7.13929534e-01 2.07964540e-01 8.55309010e-01 -9.61851418e-01
5.66938162e-01 -5.73160648e-01 -3.43642205e-01 6.33087277e-01
-1.13793516e+00 1.43180907e-01 5.71793973e-01 -4.74257916e-01
-7.09520400e-01 -1.00012809e-01 2.14480966e-01 -1.32040739e-01
-6.55113161e-02 4.98158157e-01 1.99657790e-02 -3.92598659e-01
1.02495682e+00 -2.79538751e-01 4.29459512e-02 -4.71482128e-01
1.46463901e-01 1.23140717e+00 2.86799282e-01 -4.13682699e-01
6.99939549e-01 3.92939776e-01 -2.64436584e-02 -2.88477272e-01
-1.02259445e+00 -1.03061128e+00 -8.76192510e-01 1.83577687e-02
3.82892758e-01 -5.53087056e-01 -4.59874988e-01 1.80354550e-01
-6.20476007e-01 1.29684478e-01 -1.81475416e-01 8.78155053e-01
-2.80712545e-01 5.65946102e-01 -5.00379562e-01 -4.75916445e-01
-1.89746678e-01 -7.74737358e-01 5.90430915e-01 2.30091229e-01
-6.41975164e-01 -1.33692515e+00 2.88190633e-01 3.29070538e-01
1.20829567e-01 -9.75542143e-02 1.23021626e+00 -1.28834677e+00
1.62314177e-01 -7.01647401e-01 -5.25000334e-01 6.04827821e-01
6.07801795e-01 -1.88479066e-01 -6.86274171e-01 -5.02439320e-01
-1.60911724e-01 -3.34384888e-01 5.88272452e-01 2.43372768e-02
6.97502255e-01 1.84312031e-01 -7.56772101e-01 1.83603182e-01
1.26844621e+00 2.03297570e-01 1.98263049e-01 1.73699632e-01
7.14008212e-01 8.52963209e-01 9.20121610e-01 2.15116292e-01
1.94308460e-01 6.75959587e-01 3.08375247e-02 -1.95329294e-01
-2.87931990e-02 -1.90378949e-01 -2.13107020e-01 7.04058230e-01
4.10618246e-01 -2.49385953e-01 -9.57551837e-01 6.22971475e-01
-1.86036599e+00 -7.22624302e-01 -3.47139359e-01 2.62278223e+00
9.17319834e-01 -4.63194884e-02 -1.44673049e-01 3.80678028e-01
1.09239864e+00 -3.74698967e-01 -4.72337842e-01 -3.56629044e-01
-1.22768553e-02 9.48068872e-02 2.51795352e-01 4.62576509e-01
-1.02839303e+00 6.38501227e-01 6.82398033e+00 8.30572367e-01
-7.71538615e-01 -1.36658803e-01 5.41337490e-01 -5.07019982e-02
-3.02055269e-01 -6.35176003e-02 -5.26747406e-01 3.96050185e-01
7.40909755e-01 -1.75290719e-01 2.56705314e-01 5.40084600e-01
-5.29617146e-02 -1.34085238e-01 -1.62418175e+00 1.11128533e+00
6.43127337e-02 -8.09128284e-01 -1.80924520e-01 -1.88807115e-01
9.27062690e-01 1.20959342e-01 -2.61671245e-02 7.24835470e-02
4.57824439e-01 -9.45322514e-01 4.04883437e-02 3.57879341e-01
6.90595627e-01 -1.01963186e+00 9.09746885e-01 3.60510498e-01
-9.48866189e-01 -1.67807154e-02 -4.41372752e-01 -3.96401733e-02
-1.43639997e-01 8.78016531e-01 -9.18856025e-01 3.34469527e-01
6.45314753e-01 7.75805056e-01 -5.51137984e-01 1.18781054e+00
-8.45285207e-02 2.84722447e-01 -3.81702662e-01 3.71165238e-02
2.12165445e-01 -4.67647791e-01 1.40510619e-01 9.14988577e-01
3.00329775e-01 -2.18042001e-01 9.68959630e-02 4.38836545e-01
-2.05887422e-01 4.68042850e-01 -5.46801507e-01 1.34476975e-01
6.67564094e-01 1.14150727e+00 -7.75281549e-01 -3.04948330e-01
-4.53459054e-01 9.87207532e-01 6.39017880e-01 1.40833199e-01
-5.45101345e-01 -7.63247073e-01 6.15509093e-01 -1.85253397e-01
7.36379325e-02 1.60359263e-01 -2.10082129e-01 -8.43573034e-01
1.79869980e-01 -4.28913713e-01 8.45309854e-01 -5.06349921e-01
-1.98331773e+00 3.94623876e-01 -3.23920339e-01 -1.43688214e+00
-2.04635367e-01 -3.54650497e-01 -2.04482540e-01 8.87898266e-01
-1.47060835e+00 -7.97743738e-01 -2.53095716e-01 6.06264353e-01
-3.28528374e-01 4.14325967e-02 1.02680910e+00 5.25933087e-01
-1.07793018e-01 9.59237516e-01 6.91248417e-01 1.92322075e-01
1.23428249e+00 -1.45961893e+00 -5.04168384e-02 3.48161548e-01
4.68249202e-01 5.94501317e-01 3.01576287e-01 -3.68606389e-01
-5.83036602e-01 -9.10910547e-01 1.34189880e+00 -4.91246104e-01
5.59144676e-01 -1.94957763e-01 -9.49295521e-01 3.01082581e-01
-4.52849627e-01 1.69210464e-01 1.24999797e+00 3.70711893e-01
-7.58542955e-01 -2.04320312e-01 -1.50617552e+00 4.07654166e-01
9.33639228e-01 -7.90917397e-01 -5.81079066e-01 6.95126235e-01
1.86489597e-01 2.47854576e-01 -1.10571134e+00 4.19987917e-01
5.18152177e-01 -9.00218725e-01 9.11743283e-01 -5.42544007e-01
-1.22610582e-02 -5.21086454e-01 -2.02105016e-01 -1.28141880e+00
-3.22329670e-01 -1.83398843e-01 1.57843366e-01 1.36221218e+00
7.39392042e-01 -6.65951192e-01 9.60878968e-01 7.38748968e-01
1.24285340e-01 -5.76609075e-01 -7.14393198e-01 -1.16452086e+00
1.46691933e-01 -9.00001905e-04 5.16699314e-01 1.33749139e+00
6.14923060e-01 4.18936014e-01 -1.06346264e-01 -4.55133505e-02
4.40409184e-01 2.49447078e-01 2.36008972e-01 -1.76694882e+00
-1.01951160e-01 -3.78331333e-01 -7.97434688e-01 -8.41692030e-01
3.27892900e-01 -1.17857552e+00 2.40713492e-01 -1.31623411e+00
4.71396178e-01 -1.13267374e+00 -7.70452559e-01 4.83139127e-01
-2.33992115e-01 8.01835895e-01 -2.77958483e-01 5.27040720e-01
-6.33969486e-01 2.36091256e-01 9.00056839e-01 -1.71692356e-01
-2.24299759e-01 3.13861147e-02 -6.89089179e-01 5.36911488e-01
6.62056804e-01 -8.35491836e-01 -5.37461579e-01 -2.48070166e-01
1.97080195e-01 -2.00323954e-01 -2.52528608e-01 -9.38595533e-01
2.49222711e-01 6.28028587e-02 4.25168604e-01 -2.39386410e-01
1.11729153e-01 -1.05453885e+00 -3.43149416e-02 2.51098126e-01
-6.64868534e-01 1.65625867e-02 -3.38065982e-01 5.99407613e-01
-4.66568172e-01 -6.05110288e-01 8.45524371e-01 2.73106545e-01
-6.42006636e-01 2.25261658e-01 4.11987863e-02 -1.56316254e-02
1.26923430e+00 -4.04993355e-01 -1.65416822e-01 -2.03036219e-01
-8.13169837e-01 3.42517227e-01 8.35894346e-01 4.04284865e-01
3.84809583e-01 -1.42404878e+00 -5.95866561e-01 1.88870504e-01
8.20359290e-01 -2.37155482e-01 -1.11845352e-01 8.02717924e-01
-3.92184287e-01 2.04842150e-01 3.21918987e-02 -6.32269144e-01
-1.57132876e+00 6.73146427e-01 3.15949112e-01 -3.15286815e-01
-2.66193509e-01 9.04304981e-01 -5.38034663e-02 -7.21293807e-01
3.15885961e-01 2.52355456e-01 -2.50283241e-01 1.31651565e-01
5.12004495e-01 3.47413093e-01 4.74427819e-01 -6.54557645e-01
-4.41799402e-01 6.60086930e-01 -5.12523055e-01 7.57074654e-02
9.98436451e-01 -2.55312324e-01 -2.14897469e-01 3.71701300e-01
1.94038367e+00 -2.06801966e-01 -7.67890513e-01 -5.93192577e-01
4.01251078e-01 -7.40144014e-01 -1.34610876e-01 -8.47296298e-01
-9.05059993e-01 6.36619151e-01 8.48978639e-01 2.83929408e-01
9.26087141e-01 2.39923269e-01 7.58662403e-01 6.24783099e-01
2.38011554e-01 -1.20187581e+00 5.92804234e-03 3.35163414e-01
1.71108350e-01 -1.73493767e+00 -5.96938729e-02 -4.36697066e-01
-2.88686454e-01 9.31050301e-01 4.57578301e-01 7.91017041e-02
8.72519433e-01 3.68261337e-02 3.30789685e-01 -2.80478090e-01
-3.01997691e-01 -1.04270980e-01 4.81933236e-01 8.31689119e-01
5.95539987e-01 5.18829226e-02 -6.85460925e-01 -2.80724186e-03
9.57700536e-02 -3.99711698e-01 5.05120046e-02 1.02624524e+00
-2.79445052e-01 -1.41165137e+00 -7.12274984e-02 7.14029551e-01
-1.37395233e-01 -6.88939318e-02 -6.55898392e-01 5.57761192e-01
-2.74368703e-01 1.36552286e+00 2.66558498e-01 -4.57458407e-01
-7.54893497e-02 1.98821351e-03 5.11254191e-01 -7.06979454e-01
-4.28069979e-01 -4.22879070e-01 -4.90686037e-02 -2.34642118e-01
-3.91626477e-01 -3.95962954e-01 -1.49265993e+00 -2.17195094e-01
-7.86735475e-01 6.78041339e-01 6.59836709e-01 1.05921948e+00
2.97818840e-01 -1.26690090e-01 1.09282410e+00 -3.10275346e-01
-6.99294329e-01 -7.85038173e-01 -8.96053135e-01 1.02741778e+00
4.38897423e-02 -7.50943840e-01 -6.25991464e-01 -1.70654923e-01] | [9.489611625671387, 3.6531786918640137] |
0ee126a4-b46c-4737-9abd-7c076a0a419b | face-alignment-across-large-poses-a-3d | 1511.07212 | null | http://arxiv.org/abs/1511.07212v1 | http://arxiv.org/pdf/1511.07212v1.pdf | Face Alignment Across Large Poses: A 3D Solution | Face alignment, which fits a face model to an image and extracts the semantic
meanings of facial pixels, has been an important topic in CV community.
However, most algorithms are designed for faces in small to medium poses (below
45 degree), lacking the ability to align faces in large poses up to 90 degree.
The challenges are three-fold: Firstly, the commonly used landmark-based face
model assumes that all the landmarks are visible and is therefore not suitable
for profile views. Secondly, the face appearance varies more dramatically
across large poses, ranging from frontal view to profile view. Thirdly,
labelling landmarks in large poses is extremely challenging since the invisible
landmarks have to be guessed. In this paper, we propose a solution to the three
problems in an new alignment framework, called 3D Dense Face Alignment (3DDFA),
in which a dense 3D face model is fitted to the image via convolutional neutral
network (CNN). We also propose a method to synthesize large-scale training
samples in profile views to solve the third problem of data labelling.
Experiments on the challenging AFLW database show that our approach achieves
significant improvements over state-of-the-art methods. | ['Hailin Shi', 'Zhen Lei', 'Xiangyu Zhu', 'Xiaoming Liu', 'Stan Z. Li'] | 2015-11-23 | face-alignment-across-large-poses-a-3d-1 | http://openaccess.thecvf.com/content_cvpr_2016/html/Zhu_Face_Alignment_Across_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Zhu_Face_Alignment_Across_CVPR_2016_paper.pdf | cvpr-2016-6 | ['head-pose-estimation'] | ['computer-vision'] | [-3.00342869e-02 1.38374001e-01 -6.12911843e-02 -7.39627600e-01
-4.82927412e-01 -4.49505419e-01 3.81913722e-01 -7.66710103e-01
1.32139856e-02 2.30449915e-01 8.33353624e-02 1.05786264e-01
2.74590671e-01 -5.04943073e-01 -6.76831186e-01 -4.25479263e-01
2.98032492e-01 6.00614190e-01 -9.10552815e-02 -1.19870007e-01
2.94661634e-02 9.64406729e-01 -1.51030624e+00 -1.10817194e-01
2.51823425e-01 1.02777767e+00 2.53258063e-03 1.37990311e-01
-2.62595415e-01 1.95060506e-01 -4.19639289e-01 -6.62773192e-01
7.56078601e-01 -3.29428732e-01 -7.52242565e-01 5.95736086e-01
1.21468747e+00 -4.04391587e-01 4.10732022e-03 1.27492929e+00
6.16727173e-01 -2.37412021e-01 3.84311765e-01 -1.54767966e+00
-3.99917901e-01 -9.77928340e-02 -1.06788361e+00 -2.10086286e-01
3.94892544e-01 -2.97304094e-01 6.02742910e-01 -1.32424700e+00
6.77155793e-01 1.62480271e+00 7.87760556e-01 8.71878386e-01
-1.08407938e+00 -8.88044655e-01 2.08417356e-01 -3.50005962e-02
-1.67197359e+00 -9.68273282e-01 1.00334144e+00 -4.72811788e-01
4.44132745e-01 9.83358026e-02 6.92494690e-01 8.57870340e-01
-1.38901711e-01 2.79691696e-01 1.10757279e+00 -4.01247680e-01
-1.30144786e-02 -2.45758165e-02 -4.58485603e-01 8.58506143e-01
7.67688379e-02 -1.82600707e-01 -4.09302860e-01 -3.96696776e-02
1.06379712e+00 1.76532865e-01 -1.32820159e-01 -6.31386042e-01
-8.65521669e-01 6.24302447e-01 3.72522414e-01 4.33792882e-02
-2.11120009e-01 -6.15068302e-02 4.39261496e-02 1.07657477e-01
6.77435517e-01 1.18896373e-01 -4.99872983e-01 4.24210846e-01
-1.00874233e+00 2.48340040e-01 6.46888375e-01 1.26398039e+00
1.02644455e+00 -1.22046182e-02 1.86329171e-01 9.85080361e-01
6.52157426e-01 6.17983341e-01 -1.84291694e-02 -1.02913451e+00
3.06277066e-01 6.06083095e-01 -6.21435903e-02 -1.22612393e+00
-5.24929166e-01 -2.09134802e-01 -9.06006098e-01 2.70112783e-01
3.95113856e-01 2.50659417e-02 -1.00513232e+00 1.82917011e+00
8.41646969e-01 3.91273558e-01 -2.87408143e-01 9.16634679e-01
1.09563315e+00 3.17671180e-01 -2.32705191e-01 -2.34786093e-01
1.58967781e+00 -1.00840950e+00 -8.81297052e-01 -6.35100424e-01
3.68580371e-02 -1.07661259e+00 7.88280368e-01 7.67683983e-03
-1.17813146e+00 -7.31866002e-01 -8.63324165e-01 -2.13755175e-01
-1.17584005e-01 1.68976024e-01 4.22071934e-01 7.39322186e-01
-1.44104588e+00 2.53332462e-02 -5.03495097e-01 -3.91732782e-01
7.62435675e-01 5.95849812e-01 -8.93517077e-01 -3.23283583e-01
-7.53031969e-01 8.85662079e-01 -1.90558329e-01 6.70373142e-01
-8.48680556e-01 -4.45219755e-01 -1.07682097e+00 -8.81798267e-02
4.46817994e-01 -5.24414241e-01 1.02136815e+00 -1.10593545e+00
-1.54471505e+00 1.27856743e+00 -6.59226060e-01 3.14530313e-01
5.26254535e-01 -1.75910398e-01 -3.51893306e-01 -9.27694514e-02
1.81022808e-01 9.46732938e-01 1.13024735e+00 -1.42100000e+00
-2.17620224e-01 -8.20215642e-01 4.45532687e-02 1.58206210e-01
-1.41627744e-01 3.35813135e-01 -9.82153356e-01 -3.48336518e-01
3.78444105e-01 -1.09185660e+00 -1.68411106e-01 3.47739667e-01
-2.23329842e-01 -2.00679541e-01 1.10900950e+00 -7.62377918e-01
8.16247702e-01 -1.97289908e+00 1.35449022e-01 3.08434010e-01
2.62825847e-01 3.04489464e-01 -1.70495436e-01 -3.33836605e-03
-2.45055512e-01 -2.24403553e-02 2.24962160e-02 -7.58895159e-01
-1.09692454e-01 2.02393860e-01 -1.19468812e-02 6.70584917e-01
2.42287368e-01 8.65075231e-01 -5.81249654e-01 -6.92451060e-01
1.37300283e-01 7.06509233e-01 -5.97161472e-01 3.94324511e-01
7.97069371e-02 5.64085126e-01 -2.29344577e-01 9.13248181e-01
1.30476594e+00 -2.25531161e-01 2.14606076e-01 -5.55793345e-01
-2.13039126e-02 -2.29019016e-01 -1.30736804e+00 1.84857953e+00
-4.13122714e-01 3.04507732e-01 3.15564036e-01 -7.04087734e-01
1.12064195e+00 4.81755525e-01 5.11947751e-01 -4.19907987e-01
1.12335667e-01 9.83935148e-02 -2.42515475e-01 -4.16488379e-01
7.52018988e-02 -3.14127803e-02 2.67267793e-01 3.13955873e-01
2.21503958e-01 -2.52337873e-01 -8.47715065e-02 -1.57999396e-01
4.86366183e-01 3.83634746e-01 3.48109663e-01 -2.03325167e-01
7.33000219e-01 -5.36400974e-01 8.54436934e-01 -1.20177038e-01
-2.85949647e-01 8.98066819e-01 4.34183091e-01 -8.29237938e-01
-9.16802108e-01 -7.67267048e-01 -1.01455301e-01 6.43444002e-01
1.59462512e-01 -5.41253209e-01 -1.14642560e+00 -8.38194251e-01
-2.26629883e-01 -1.75695002e-01 -6.69169366e-01 2.22149849e-01
-7.40750194e-01 -4.25816745e-01 2.23799542e-01 4.60445821e-01
6.50214076e-01 -9.19756770e-01 -2.46770293e-01 -2.54072458e-01
-3.17512423e-01 -1.45590627e+00 -8.27379227e-01 -3.76687706e-01
-5.54286838e-01 -1.15926158e+00 -5.89245081e-01 -1.03696549e+00
1.33757126e+00 3.88396770e-01 1.15306795e+00 2.58927882e-01
-2.77250618e-01 2.16723964e-01 6.06809370e-02 -4.90698397e-01
7.65375327e-03 -2.41957054e-01 1.33271515e-01 3.60103488e-01
5.21075785e-01 -2.88288653e-01 -6.08724713e-01 6.95558786e-01
-5.29094994e-01 1.01748044e-02 4.87055898e-01 7.09811687e-01
7.40192771e-01 -2.16822907e-01 4.08692122e-01 -8.87263775e-01
-6.89947605e-02 -1.43476248e-01 -6.89474344e-01 3.21906418e-01
-2.30610088e-01 -3.41851920e-01 3.23408753e-01 -2.11106822e-01
-9.35008705e-01 5.78996241e-01 -4.57814783e-01 -6.37214720e-01
-2.03018248e-01 -2.66576916e-01 -8.82601023e-01 -5.23377657e-01
2.62378663e-01 -1.94226682e-01 3.50615919e-01 -4.05965567e-01
3.41023177e-01 5.84651649e-01 4.50208992e-01 -5.25769234e-01
1.06043839e+00 7.43614495e-01 2.32160255e-01 -8.27658117e-01
-1.14801705e+00 -3.92699540e-01 -1.23294175e+00 -4.08877850e-01
9.10083413e-01 -1.04047823e+00 -6.63312674e-01 4.89317924e-01
-1.35821307e+00 3.02964766e-02 -2.14162059e-02 1.48975447e-01
-6.29663467e-01 3.02758098e-01 -1.73816100e-01 -5.71946681e-01
-3.51633966e-01 -1.46687257e+00 1.47669518e+00 3.79341930e-01
-2.11233005e-01 -8.45983088e-01 -1.45932421e-01 5.03486216e-01
3.27266902e-01 2.88501263e-01 5.00769258e-01 -1.85466588e-01
-4.69366401e-01 -6.02520071e-02 -2.83414185e-01 3.60110521e-01
3.49985629e-01 -1.66838174e-03 -1.31862438e+00 -3.21233988e-01
1.63879186e-01 -3.97271872e-01 1.18904002e-01 3.83811533e-01
1.29141819e+00 -1.18212909e-01 -3.28966558e-01 9.27708089e-01
1.13886809e+00 -3.47373565e-03 7.05297470e-01 -2.43076291e-02
8.99714947e-01 8.31158817e-01 6.45608008e-01 1.19866699e-01
5.09441733e-01 1.13639605e+00 5.64740181e-01 -4.75735366e-01
-2.96783566e-01 -2.53897250e-01 1.40426770e-01 6.06901646e-01
-2.34266415e-01 2.43733570e-01 -7.30075359e-01 2.80756414e-01
-1.49587011e+00 -8.22816491e-01 4.67048101e-02 2.07337213e+00
7.58285880e-01 -3.66375268e-01 3.85101028e-02 -9.45535079e-02
8.10933650e-01 1.32135928e-01 -3.20079625e-01 -6.53753430e-02
4.57338914e-02 2.19595164e-01 1.17059931e-01 4.70041007e-01
-1.18998909e+00 1.10983956e+00 6.36933851e+00 5.41422009e-01
-1.10495341e+00 2.14161158e-01 7.91339338e-01 1.86117217e-01
1.34818787e-02 2.71897181e-03 -1.26624537e+00 2.91376323e-01
4.10559505e-01 3.94615322e-01 3.28224272e-01 9.40622449e-01
-5.86801097e-02 2.59422243e-01 -1.24368513e+00 1.44422710e+00
5.68091154e-01 -8.71811748e-01 -8.09678808e-02 2.60293186e-01
8.51116538e-01 -3.61532480e-01 1.14920124e-01 -6.99352622e-02
-6.44511133e-02 -1.31448257e+00 7.01253533e-01 3.65476787e-01
1.08625340e+00 -7.43048906e-01 6.56432331e-01 1.13205001e-01
-1.31813025e+00 3.67505103e-01 -7.02505529e-01 2.50674069e-01
2.42415860e-01 4.20119077e-01 -7.84099340e-01 3.51822406e-01
8.34597290e-01 5.93345702e-01 -5.71345627e-01 7.70591378e-01
-2.14724153e-01 -3.26999626e-03 -3.15191478e-01 6.08204007e-01
-1.20703325e-01 -3.16582680e-01 4.47347574e-02 8.13976109e-01
5.27114332e-01 -4.96972874e-02 2.22329408e-01 7.43957281e-01
-3.28936756e-01 4.20522988e-02 -7.33965397e-01 3.72824371e-01
5.19747376e-01 1.68223226e+00 -7.72480428e-01 5.70700131e-02
-6.68026686e-01 1.06846941e+00 4.46656555e-01 2.30188265e-01
-6.82724535e-01 2.08862036e-01 7.92789340e-01 3.03721756e-01
1.64733306e-01 -5.37883043e-02 8.72162133e-02 -9.80993688e-01
1.98732600e-01 -9.47183728e-01 9.99731049e-02 -6.77465916e-01
-1.29656875e+00 8.07647586e-01 -1.45965353e-01 -1.17461729e+00
-4.69942875e-02 -6.18734479e-01 -3.64679128e-01 1.00544322e+00
-1.66737902e+00 -1.55784380e+00 -6.58558547e-01 8.02376509e-01
5.71795166e-01 -1.13767117e-01 9.56172228e-01 5.05590379e-01
-5.25783777e-01 6.56088531e-01 -5.20139813e-01 3.94298077e-01
9.22312140e-01 -8.66177738e-01 6.94918871e-01 6.70109868e-01
3.31342161e-01 7.09598958e-01 2.31792063e-01 -5.12672842e-01
-1.53901315e+00 -1.25946701e+00 1.05382431e+00 -6.76792026e-01
-9.76352021e-03 -6.29789054e-01 -6.88063323e-01 9.03140128e-01
1.03640251e-01 7.38506079e-01 6.87119603e-01 4.66596894e-02
-5.00523567e-01 -3.33105475e-01 -1.33725762e+00 5.07621229e-01
1.34836316e+00 -4.34606224e-01 -1.51948079e-01 3.50914538e-01
2.19673023e-01 -7.54381120e-01 -7.95306087e-01 4.40616131e-01
7.00413167e-01 -7.85385251e-01 1.21614301e+00 -3.53893220e-01
1.46119803e-01 -5.72318196e-01 -1.75399750e-01 -1.27481723e+00
-1.94662124e-01 -6.10184968e-01 6.92762015e-03 1.41458499e+00
2.39326507e-02 -5.51304936e-01 8.08110833e-01 4.62463140e-01
6.56178296e-02 -7.56009638e-01 -1.02833128e+00 -3.98720056e-01
-3.40542048e-01 -8.57410133e-02 9.86861706e-01 1.04638875e+00
-5.77865243e-01 3.99144113e-01 -5.74509740e-01 3.62032205e-01
6.67513907e-01 1.07309990e-01 1.21245015e+00 -1.49108660e+00
2.48981491e-01 -1.12823039e-01 -7.01429129e-01 -1.05395412e+00
5.16426802e-01 -7.64420807e-01 7.59548694e-03 -1.36592734e+00
1.96174666e-01 -5.27460515e-01 2.81130970e-01 5.28238058e-01
-6.47432506e-02 7.30790436e-01 1.34687856e-01 3.85775641e-02
-3.61910343e-01 3.91247094e-01 1.41968250e+00 4.99488972e-03
3.12883496e-01 -6.24279156e-02 -6.97925210e-01 1.02597368e+00
4.48918372e-01 -2.70150542e-01 -4.61503565e-01 -6.15219712e-01
1.27033100e-01 -2.46681899e-01 2.58574426e-01 -6.51710451e-01
1.43528908e-01 -1.35235384e-01 8.52826476e-01 -6.98080242e-01
6.20789587e-01 -1.26842391e+00 4.60341871e-01 -1.47502711e-02
2.07264110e-01 3.55374157e-01 9.03121457e-02 2.26037621e-01
-2.75913000e-01 -1.03551142e-01 1.01458180e+00 -2.36098349e-01
-5.62894404e-01 8.21269333e-01 4.88123804e-01 2.28962563e-02
1.28689051e+00 -2.91960239e-01 -6.95427209e-02 -2.59836704e-01
-6.18336737e-01 2.28111595e-01 6.35108292e-01 6.73473418e-01
7.90538907e-01 -1.66112101e+00 -7.56009102e-01 8.24779451e-01
5.28636165e-02 4.10709053e-01 3.31389546e-01 7.62291014e-01
-6.30260527e-01 3.02311867e-01 -3.07921171e-01 -8.64836276e-01
-1.59046829e+00 4.55164790e-01 4.91935760e-01 2.30136544e-01
-5.48663795e-01 1.02789557e+00 5.35311401e-01 -7.22149253e-01
2.22258404e-01 1.99694216e-01 -2.75435925e-01 1.12322450e-01
6.85522676e-01 1.91511940e-02 2.29840383e-01 -1.30448604e+00
-4.22859818e-01 1.40631771e+00 -4.72637527e-02 2.72883415e-01
1.28433657e+00 -2.35024631e-01 -4.47241813e-01 -1.15314253e-01
1.26960540e+00 1.68218240e-01 -1.35534608e+00 -3.32018256e-01
-2.70335972e-01 -9.59914684e-01 -1.61620542e-01 -2.87565529e-01
-1.64484274e+00 1.04915595e+00 6.00342095e-01 -3.42229784e-01
1.13592947e+00 1.50540173e-02 5.49335718e-01 2.79560573e-02
5.29929936e-01 -9.15784776e-01 4.41752449e-02 3.09543192e-01
1.03854561e+00 -1.25508344e+00 2.28809025e-02 -8.10133934e-01
-4.84292179e-01 1.18500113e+00 9.82371926e-01 2.29294822e-01
8.34887385e-01 4.20748949e-01 3.72391552e-01 -5.32168210e-01
-2.51897007e-01 -4.64132763e-02 4.45258766e-01 6.10932708e-01
4.29107547e-01 -2.23662540e-01 1.89138159e-01 5.90194575e-02
-2.52166778e-01 -2.00037554e-01 1.19343109e-01 6.74772739e-01
-6.81323633e-02 -1.25381625e+00 -5.80662906e-01 5.00175953e-02
-4.80210394e-01 2.19697446e-01 -5.10360181e-01 7.31334448e-01
4.18937653e-01 8.02630007e-01 2.12896764e-01 -8.03886428e-02
3.96165520e-01 2.01196328e-01 7.18757927e-01 -7.33335197e-01
-1.73753917e-01 3.45935762e-01 -2.56798565e-01 -7.77594566e-01
-7.74704039e-01 -4.95100051e-01 -9.54924822e-01 -2.49457732e-01
-4.02315617e-01 -8.50136802e-02 8.06184947e-01 8.44961286e-01
3.58295172e-01 1.88813642e-01 8.03288043e-01 -1.20940804e+00
-2.81660110e-01 -9.36938584e-01 -4.72781539e-01 3.55152011e-01
2.86127776e-01 -9.67872918e-01 -1.55604631e-01 1.66619509e-01] | [13.28525447845459, 0.22136206924915314] |
bb72b8d7-99f8-413f-a283-a84888be8862 | data-free-quantization-with-accurate | 2204.04215 | null | https://arxiv.org/abs/2204.04215v2 | https://arxiv.org/pdf/2204.04215v2.pdf | Data-Free Quantization with Accurate Activation Clipping and Adaptive Batch Normalization | Data-free quantization is a task that compresses the neural network to low bit-width without access to original training data. Most existing data-free quantization methods cause severe performance degradation due to inaccurate activation clipping range and quantization error, especially for low bit-width. In this paper, we present a simple yet effective data-free quantization method with accurate activation clipping and adaptive batch normalization. Accurate activation clipping (AAC) improves the model accuracy by exploiting accurate activation information from the full-precision model. Adaptive batch normalization firstly proposes to address the quantization error from distribution changes by updating the batch normalization layer adaptively. Extensive experiments demonstrate that the proposed data-free quantization method can yield surprisingly performance, achieving 64.33% top-1 accuracy of ResNet18 on ImageNet dataset, with 3.7% absolute improvement outperforming the existing state-of-the-art methods. | ['Hong Zhou', 'Weijia Wu', 'Luoming Zhang', 'Yefei He'] | 2022-04-08 | null | null | null | null | ['data-free-quantization', 'data-free-quantization'] | ['computer-vision', 'methodology'] | [ 2.68767595e-01 -4.93045390e-01 -3.85471672e-01 -5.63606143e-01
-7.00877070e-01 -1.90325454e-01 2.45384499e-01 1.74241647e-01
-1.06240177e+00 7.37365961e-01 5.61060756e-02 -2.90529788e-01
1.71613172e-01 -6.54946208e-01 -7.29696929e-01 -7.82045245e-01
1.51394621e-01 -9.89652276e-02 4.46476638e-01 -1.41924933e-01
4.41892624e-01 3.46996248e-01 -1.64632237e+00 3.94792706e-01
8.39468598e-01 1.56168711e+00 1.72690615e-01 6.92444146e-01
-1.85934007e-01 6.79733157e-01 -1.04044819e+00 -2.33577073e-01
5.23138821e-01 1.87302213e-02 -3.95370305e-01 -4.38137472e-01
6.77680314e-01 -7.42347777e-01 -5.24181247e-01 1.49007702e+00
8.12843025e-01 9.45082903e-02 3.99649084e-01 -8.74363065e-01
-6.64133310e-01 8.00337017e-01 -4.92806703e-01 7.16841459e-01
-4.74409729e-01 5.67291817e-03 4.75410372e-01 -9.44545448e-01
8.28509182e-02 1.02600479e+00 7.03905642e-01 5.05380988e-01
-1.19451165e+00 -1.32426667e+00 2.22334247e-02 4.42882597e-01
-1.99544692e+00 -6.34898782e-01 4.72290725e-01 -2.03780621e-01
1.30187011e+00 -2.05454808e-02 5.52341640e-01 3.20167542e-01
2.43947297e-01 4.57147181e-01 6.65312827e-01 -3.32234412e-01
6.51652873e-01 -3.71922165e-01 1.87386155e-01 3.15065175e-01
3.24533373e-01 -6.26748651e-02 -6.73312187e-01 1.49353489e-01
1.00326622e+00 2.29405742e-02 -3.83777589e-01 2.98064768e-01
-9.39156234e-01 4.47437465e-01 7.38914967e-01 2.01245934e-01
-6.16164088e-01 6.16095841e-01 6.87644005e-01 3.43063623e-01
2.51944214e-01 8.83295685e-02 -5.24658382e-01 -4.28406268e-01
-1.36224949e+00 1.59378156e-01 2.12329417e-01 9.80130017e-01
6.69993281e-01 8.09625208e-01 -4.34526533e-01 9.22182560e-01
2.05943480e-01 3.84085625e-01 1.13293564e+00 -8.57820034e-01
6.37298524e-01 4.94949520e-01 -2.22612008e-01 -8.18039179e-01
-1.81183025e-01 -5.97440064e-01 -1.37102604e+00 2.33671397e-01
1.50750980e-01 -9.37647820e-02 -1.18387210e+00 1.71404147e+00
-6.51169568e-03 2.56675929e-01 9.18116868e-02 8.25493217e-01
7.51019537e-01 8.52429926e-01 3.13761503e-01 -1.89204127e-01
1.23154414e+00 -7.58283794e-01 -1.18211293e+00 -2.51744390e-02
6.01065159e-01 -5.71892619e-01 1.09737790e+00 7.39459157e-01
-1.26881027e+00 -8.39760900e-01 -1.65748739e+00 -2.30422184e-01
-3.70118052e-01 3.27779531e-01 3.50043058e-01 9.87293601e-01
-1.29965401e+00 6.00074768e-01 -8.55565071e-01 3.66702467e-01
6.82236671e-01 8.46705079e-01 -5.75482473e-02 -1.08968556e-01
-1.48072040e+00 4.61987793e-01 8.39629769e-01 -1.53117299e-01
-3.62823159e-01 -9.69174206e-01 -6.93693459e-01 4.04445112e-01
-7.44829103e-02 -9.23933368e-03 1.36297059e+00 -1.10508680e+00
-1.63554323e+00 4.62326929e-02 -1.28739804e-01 -9.69784021e-01
1.64904654e-01 -1.54371649e-01 -4.46832567e-01 1.43661782e-01
-4.67536211e-01 1.09795773e+00 1.08222640e+00 -5.24310589e-01
-5.16601741e-01 -3.81606132e-01 -5.00526965e-01 1.82671785e-01
-1.07284260e+00 -3.52947891e-01 -6.57771647e-01 -1.10201919e+00
3.58324111e-01 -2.09269524e-01 -1.53740019e-01 1.76839847e-02
1.58651039e-01 -1.48392573e-01 7.23714292e-01 -6.82473838e-01
1.92352808e+00 -2.26199293e+00 -6.51153505e-01 1.15217432e-01
7.17504472e-02 7.59185255e-01 -2.92970128e-02 -2.85348743e-01
-6.60208836e-02 2.81941384e-01 -2.70271420e-01 -3.92575562e-01
1.65082663e-02 2.86642879e-01 -3.54762167e-01 4.20078814e-01
2.93006927e-01 8.51941764e-01 -6.83657110e-01 -5.08194029e-01
2.77342886e-01 8.14724863e-01 -7.93860435e-01 -1.31842285e-01
1.56638518e-01 -1.58947796e-01 1.45310447e-01 5.34910440e-01
1.05825841e+00 -4.51587290e-02 2.54153437e-03 -6.28583372e-01
-6.45920113e-02 4.24667358e-01 -1.22684348e+00 1.74422812e+00
-1.96705267e-01 6.47026420e-01 3.10009308e-02 -7.97877192e-01
8.52098405e-01 3.09328318e-01 2.15967998e-01 -1.08479178e+00
3.08313191e-01 1.27587229e-01 -5.43895848e-02 1.77820727e-01
8.55793536e-01 1.03075273e-01 2.43149832e-01 -7.82968029e-02
3.02470982e-01 1.55773805e-02 8.21985602e-02 -1.24379717e-01
8.36200237e-01 -5.02010643e-01 2.65842259e-01 -3.12556177e-01
4.59242195e-01 -4.72608477e-01 6.82044029e-01 6.65122807e-01
-5.16801298e-01 5.67260385e-01 -2.57482156e-02 -2.88286358e-01
-1.07336390e+00 -7.49221206e-01 -5.29723704e-01 1.23054528e+00
-1.24853775e-02 -7.20825553e-01 -1.00602853e+00 -1.88550591e-01
-1.81931779e-01 5.34805357e-01 -2.62691140e-01 -3.45644414e-01
-5.92599511e-01 -9.91507411e-01 9.72760439e-01 7.74788797e-01
1.14722538e+00 -3.89605045e-01 -4.36030567e-01 3.96368414e-01
9.14655849e-02 -1.05612707e+00 -5.49499750e-01 4.29221809e-01
-1.28875899e+00 -3.13037843e-01 -8.10423255e-01 -7.81752348e-01
6.36367798e-01 -9.73833576e-02 7.76633382e-01 1.35897413e-01
-1.42418337e-03 -3.82241100e-01 -1.51571929e-01 -3.84618998e-01
-2.34442562e-01 3.14216644e-01 2.33897880e-01 -1.67075798e-01
5.46039283e-01 -4.66009051e-01 -8.71382236e-01 1.29664123e-01
-1.22732592e+00 -1.46527126e-01 6.95181906e-01 7.74487734e-01
1.00086629e+00 3.86066973e-01 8.08338344e-01 -4.56998855e-01
6.64781868e-01 -1.65423691e-01 -8.06071162e-01 -1.62490115e-01
-1.06865215e+00 2.71284223e-01 8.77986968e-01 -6.02382123e-01
-7.35070407e-01 5.30054457e-02 -5.23935556e-01 -6.66434348e-01
8.32802206e-02 2.66091317e-01 -8.89958963e-02 -3.27208281e-01
8.17663789e-01 2.91842103e-01 -2.01753631e-01 -5.10801971e-01
2.68205732e-01 8.57254803e-01 9.73772883e-01 -2.49936670e-01
2.73280174e-01 1.48738772e-01 -2.09299952e-01 -7.27884829e-01
-1.63842529e-01 -4.18478012e-01 -5.92591286e-01 7.85253718e-02
6.30385697e-01 -1.11598027e+00 -4.49601322e-01 7.27304816e-01
-1.03439462e+00 -4.02977884e-01 -3.19463164e-01 5.21896541e-01
-4.44227085e-02 2.87538439e-01 -7.22606480e-01 -5.79254568e-01
-8.15367103e-01 -1.25143778e+00 6.50847614e-01 1.77577034e-01
1.85514510e-01 -6.35302305e-01 -4.08893526e-01 -1.91449881e-01
8.87264669e-01 -2.62470186e-01 6.99155092e-01 -5.37805736e-01
-3.08521509e-01 -2.64122009e-01 -3.64818037e-01 8.39564741e-01
1.60718672e-02 -1.86543837e-01 -1.18781233e+00 -4.33118939e-01
1.28754124e-01 -2.50559330e-01 1.04622090e+00 5.25368094e-01
1.84294796e+00 -5.96333325e-01 1.88695937e-01 1.02576160e+00
1.48963499e+00 5.98988011e-02 9.04049695e-01 6.20396212e-02
5.44896126e-01 -4.39299226e-01 3.97981346e-01 6.94285929e-01
6.22089766e-02 5.11086941e-01 4.69485223e-01 3.03731263e-02
-4.29039568e-01 -8.00996274e-02 2.49586135e-01 1.14113247e+00
1.73859835e-01 -9.89983237e-05 -7.86408663e-01 3.86242449e-01
-1.42392004e+00 -6.99760973e-01 1.69069678e-01 2.41089129e+00
1.44884181e+00 5.55997431e-01 -3.60349983e-01 7.21010864e-01
7.15466321e-01 3.12268883e-02 -7.21299827e-01 -4.55026388e-01
-1.38296843e-01 5.96668839e-01 1.36864519e+00 3.87374192e-01
-1.15059137e+00 9.34046566e-01 6.61862659e+00 1.42523456e+00
-1.45507705e+00 3.09377819e-01 7.92939663e-01 -3.60832393e-01
1.78714812e-01 -6.65305972e-01 -1.26407337e+00 6.55272603e-01
1.44462419e+00 3.97617333e-02 3.76814425e-01 9.40032363e-01
1.17094275e-02 5.57307042e-02 -7.89082766e-01 1.45734835e+00
-3.36809047e-02 -1.47443652e+00 3.29513341e-01 -9.31876823e-02
1.05361295e+00 2.67281055e-01 1.94452211e-01 3.86584163e-01
-5.14935106e-02 -1.07350063e+00 8.31399500e-01 2.07101613e-01
1.44960332e+00 -9.46400642e-01 9.18452263e-01 1.88811943e-01
-1.36204123e+00 -2.27412373e-01 -7.17012644e-01 -2.16728792e-01
-2.27318317e-01 7.54323244e-01 -6.97172344e-01 -7.70403594e-02
6.99452162e-01 4.94768620e-01 -6.05509877e-01 1.04072106e+00
-7.97916353e-02 1.04208338e+00 -5.68333089e-01 -5.14863096e-02
2.05057323e-01 2.51615375e-01 1.06184948e-02 1.44694734e+00
3.33312571e-01 3.75123978e-01 -2.57496595e-01 5.80180943e-01
-5.11420667e-01 6.97308555e-02 7.09827021e-02 1.78669125e-01
1.04764163e+00 8.80312026e-01 -5.28829932e-01 -7.04285324e-01
-8.41905475e-02 9.27343965e-01 1.52612820e-01 3.99044901e-01
-7.50914693e-01 -7.55658567e-01 7.33838320e-01 1.60794705e-02
4.90808070e-01 -2.69266248e-01 -7.13619053e-01 -7.44116306e-01
-8.70341584e-02 -6.26172900e-01 1.59998402e-01 -4.20173943e-01
-7.59295225e-01 4.78083909e-01 -1.17781654e-01 -9.61130500e-01
-1.53532639e-01 -5.99667132e-01 -2.34783813e-01 8.71247709e-01
-1.84430718e+00 -5.14577448e-01 -3.29573393e-01 5.72170198e-01
6.85690880e-01 -1.57313213e-01 6.76363528e-01 7.72267759e-01
-5.47697186e-01 1.39543056e+00 4.19395953e-01 1.76832959e-01
7.33793378e-01 -1.06341267e+00 5.60212731e-01 9.37734067e-01
-3.14257890e-01 6.14291847e-01 2.88625389e-01 -4.24270123e-01
-1.28091121e+00 -1.45485139e+00 6.69199705e-01 3.28069210e-01
2.25068018e-01 -3.12330484e-01 -1.27352786e+00 1.25548407e-01
-1.10515393e-01 3.18879366e-01 5.64480066e-01 -7.86739230e-01
-3.05853486e-01 -6.22095108e-01 -1.33270979e+00 3.92846733e-01
5.69545269e-01 -4.29610282e-01 -2.94162631e-01 -5.49210049e-02
8.44863713e-01 -5.15311122e-01 -1.03269148e+00 5.33202171e-01
4.14664149e-01 -6.43459797e-01 1.10966980e+00 -7.82669187e-02
2.57467907e-02 -4.01921540e-01 -5.53253829e-01 -9.46222425e-01
-4.13060695e-01 -5.66163778e-01 -4.12461609e-01 1.07960343e+00
8.39603171e-02 -6.46250844e-01 7.92909563e-01 6.46258712e-01
-2.80935317e-01 -6.55411124e-01 -1.31727803e+00 -6.77654684e-01
1.61524981e-01 -7.97191858e-01 9.16322947e-01 7.44904101e-01
-4.23197076e-02 -1.94758534e-01 9.05579403e-02 1.41726300e-01
5.39951324e-01 -9.02297199e-01 2.34169841e-01 -8.16159904e-01
6.55930042e-02 -6.81528211e-01 -6.60179079e-01 -1.49650264e+00
-1.30857289e-01 -6.10053122e-01 1.20348006e-01 -1.09558237e+00
-4.62868661e-01 -5.18675745e-01 -7.43014514e-01 6.69606745e-01
-1.46695137e-01 8.87793422e-01 5.17246090e-02 3.44522625e-01
-4.98986453e-01 6.83289349e-01 1.01683664e+00 -2.86505163e-01
-1.87587425e-01 -4.92432803e-01 -3.69714111e-01 6.03965998e-01
1.14487231e+00 -4.36945111e-01 -6.58547342e-01 -7.33889163e-01
2.32181791e-02 -5.29961765e-01 8.09187293e-02 -1.66115224e+00
5.74025154e-01 1.48735657e-01 7.33255744e-01 -7.28510737e-01
2.91608274e-01 -7.00862825e-01 -2.30324969e-01 5.45991004e-01
-4.47850496e-01 3.27239931e-01 6.60527647e-01 3.96630496e-01
-3.03057581e-01 -3.77160341e-01 1.08798492e+00 2.48680577e-01
-9.83451784e-01 2.87197560e-01 -4.02447224e-01 -3.26622911e-02
5.27361751e-01 -3.39983135e-01 -3.26477617e-01 -7.44812638e-02
-2.88091660e-01 -9.45965350e-02 2.60740936e-01 5.31247333e-02
8.09123874e-01 -1.58589005e+00 -5.61735213e-01 8.08714449e-01
-2.37400487e-01 2.84734815e-01 1.85956866e-01 4.67374563e-01
-7.29974270e-01 5.56155682e-01 -2.51033425e-01 -5.45736372e-01
-1.10061276e+00 2.44445711e-01 4.12084281e-01 7.43589103e-02
-2.69757777e-01 1.12830853e+00 -2.36034840e-01 1.60839528e-01
7.53721774e-01 -1.03478062e+00 -1.62756205e-01 -2.13125452e-01
1.08341134e+00 4.38186049e-01 6.53540850e-01 -4.76164371e-01
-1.52540177e-01 3.60036463e-01 -1.34398341e-01 -1.97116718e-01
8.42501938e-01 -3.79011892e-02 9.20125097e-02 1.60976157e-01
1.32719111e+00 -6.42292619e-01 -1.41389835e+00 -2.69404858e-01
-4.17722642e-01 -3.99446249e-01 9.08808529e-01 -7.22179711e-01
-1.27267504e+00 1.23049235e+00 1.19579673e+00 -1.74843296e-01
1.54782391e+00 -8.66212666e-01 9.58914757e-01 4.85686481e-01
1.75699472e-01 -1.40953577e+00 -2.07707733e-01 7.24699914e-01
5.98855555e-01 -9.25415874e-01 1.21225730e-01 3.51272970e-02
-7.83781037e-02 1.07058465e+00 7.30435491e-01 -5.97052872e-02
8.54070306e-01 5.40591598e-01 -4.91193943e-02 3.64335209e-01
-7.70477951e-01 3.40434462e-01 2.80710220e-01 5.83014548e-01
3.23644042e-01 -1.64819404e-01 -3.23052555e-01 8.64541173e-01
-3.98825198e-01 2.90911704e-01 1.68321595e-01 9.30249453e-01
-7.93746412e-01 -7.72457361e-01 -4.79878992e-01 6.11522496e-01
-6.90862894e-01 -6.30437732e-01 1.79500967e-01 1.64696828e-01
2.75687814e-01 7.17342198e-01 6.53605044e-01 -5.43037713e-01
1.86153054e-01 1.83610931e-01 2.73080677e-01 -1.44263819e-01
-7.08352447e-01 7.01756403e-02 -4.67171818e-01 -4.82402951e-01
-2.25379560e-02 -1.27478614e-01 -1.68784714e+00 -4.61391598e-01
-5.20316541e-01 -4.06251363e-02 1.19887209e+00 5.18596590e-01
5.90336740e-01 8.59732747e-01 2.48047903e-01 -9.00759459e-01
-9.54984963e-01 -1.07093155e+00 -4.05841351e-01 1.54202446e-01
4.90651011e-01 -3.56238902e-01 -2.46579170e-01 2.83673644e-01] | [8.665858268737793, 3.030188798904419] |
9b954e6c-fb74-44e0-aa50-ebb3e71b6d1d | m3p-learning-universal-representations-via | 2006.02635 | null | https://arxiv.org/abs/2006.02635v4 | https://arxiv.org/pdf/2006.02635v4.pdf | M3P: Learning Universal Representations via Multitask Multilingual Multimodal Pre-training | We present M3P, a Multitask Multilingual Multimodal Pre-trained model that combines multilingual pre-training and multimodal pre-training into a unified framework via multitask pre-training. Our goal is to learn universal representations that can map objects occurred in different modalities or texts expressed in different languages into a common semantic space. In addition, to explicitly encourage fine-grained alignment between images and non-English languages, we also propose Multimodal Code-switched Training (MCT) to combine monolingual pre-training and multimodal pre-training via a code-switch strategy. Experiments are performed on the multilingual image retrieval task across two benchmark datasets, including MSCOCO and Multi30K. M3P can achieve comparable results for English and new state-of-the-art results for non-English languages. | ['Dongdong Zhang', 'Jianfeng Gao', 'Taroon Bharti', 'Lin Su', 'Minheng Ni', 'Edward Cui', 'Nan Duan', 'Lijuan Wang', 'Haoyang Huang'] | 2020-06-04 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Ni_M3P_Learning_Universal_Representations_via_Multitask_Multilingual_Multimodal_Pre-Training_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Ni_M3P_Learning_Universal_Representations_via_Multitask_Multilingual_Multimodal_Pre-Training_CVPR_2021_paper.pdf | cvpr-2021-1 | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 1.11204907e-01 -4.47915554e-01 -3.18307996e-01 -5.46024680e-01
-1.55291235e+00 -7.62249351e-01 9.83929634e-01 2.96352189e-02
-7.61514485e-01 4.51323807e-01 1.71279237e-01 -3.04677129e-01
2.86124617e-01 -2.16918349e-01 -9.74051476e-01 -2.50169307e-01
3.01317275e-01 6.06374681e-01 -7.07004145e-02 -2.00398549e-01
-1.27228007e-01 -1.85811639e-01 -1.32880986e+00 1.22263110e+00
1.03580713e+00 7.08117962e-01 6.80663228e-01 6.01204038e-01
-2.98898697e-01 6.77937150e-01 -2.79576909e-02 -8.75873387e-01
-9.58437398e-02 -1.99209809e-01 -1.10638559e+00 6.83319345e-02
8.14115047e-01 1.92714348e-01 -2.51563601e-02 1.09464943e+00
4.27333444e-01 -1.16144426e-01 8.91292512e-01 -1.16502392e+00
-1.22081673e+00 3.86399984e-01 -7.01989353e-01 -2.62115300e-01
5.42759001e-01 -1.69902936e-01 8.98911357e-01 -1.35958993e+00
7.97031820e-01 1.49503613e+00 4.00867820e-01 5.51766336e-01
-1.26299608e+00 -6.49092495e-01 1.41985342e-02 4.02523518e-01
-1.61894488e+00 -4.11766797e-01 2.64168739e-01 -5.58908582e-01
1.00090849e+00 1.63321346e-01 -3.65435742e-02 1.27791595e+00
2.40353674e-01 9.66703475e-01 1.34141481e+00 -6.21293068e-01
-5.03186285e-01 4.23889577e-01 -3.17898601e-01 9.53856111e-01
-3.93775165e-01 -2.40089551e-01 -6.69633150e-01 -2.06675883e-02
4.11414266e-01 -1.83093771e-01 -1.09705053e-01 -5.06768525e-01
-1.84293008e+00 8.53027463e-01 3.43427539e-01 2.72429407e-01
5.41411527e-02 6.74815401e-02 7.29838967e-01 5.56762815e-01
2.16369301e-01 1.08475551e-01 -6.84613645e-01 2.82495588e-01
-6.45575941e-01 -2.29922116e-01 3.25222969e-01 1.27601683e+00
1.09769261e+00 -3.74887258e-01 -2.89657921e-01 1.29145467e+00
5.28456032e-01 1.04219723e+00 5.62842429e-01 -7.22814739e-01
9.85406518e-01 4.69418436e-01 -2.51582026e-01 -4.19863999e-01
-1.30670995e-01 -1.51336566e-01 -6.61707819e-01 -2.85280228e-01
1.77947357e-01 -2.18161568e-02 -9.73494947e-01 1.75951338e+00
-1.20886669e-01 -1.77449867e-01 5.93838573e-01 7.83387423e-01
1.15130210e+00 7.52007484e-01 5.05120873e-01 2.29867503e-01
1.52824426e+00 -1.43264294e+00 -5.95866323e-01 -1.79140687e-01
8.76826823e-01 -1.21530759e+00 1.45582032e+00 7.93943256e-02
-7.28295088e-01 -8.09121966e-01 -6.64128304e-01 -5.30125558e-01
-9.32220101e-01 6.51138604e-01 4.98208582e-01 3.43442887e-01
-1.00764453e+00 -3.76138896e-01 -4.51522738e-01 -6.04350209e-01
1.16045773e-01 -1.66738052e-02 -7.28234112e-01 -6.01378739e-01
-1.14915693e+00 1.16613173e+00 6.30658507e-01 -2.47884259e-01
-1.09547269e+00 -5.37127793e-01 -1.15014446e+00 -2.70036876e-01
3.33636463e-01 -6.71704829e-01 9.74463463e-01 -1.37334979e+00
-1.20542908e+00 1.57280481e+00 -2.68405348e-01 2.14703958e-02
2.80004948e-01 -2.37520710e-01 -7.77384281e-01 2.20213935e-01
4.01652515e-01 1.37282395e+00 7.15198576e-01 -1.56913102e+00
-5.46506107e-01 4.22823951e-02 7.47821108e-02 6.76796198e-01
-4.09983963e-01 3.28689724e-01 -1.19499362e+00 -5.72529018e-01
-2.76693135e-01 -1.27442908e+00 8.00208896e-02 -2.87903488e-01
-4.67620075e-01 -2.66036153e-01 5.57021618e-01 -7.10846186e-01
7.34137177e-01 -2.03984928e+00 6.56208575e-01 5.44243455e-02
-2.26577520e-01 -1.25748187e-01 -8.93923223e-01 5.04526496e-01
-2.54526258e-01 -9.00118724e-02 -2.50106722e-01 -6.34418070e-01
1.12931393e-01 4.04940903e-01 -1.77222699e-01 1.06811315e-01
3.58501554e-01 1.16982722e+00 -8.52799296e-01 -7.95587540e-01
4.53525186e-02 5.52869856e-01 -4.07318860e-01 1.59654677e-01
-8.79892185e-02 7.54449904e-01 -3.49201933e-02 9.82112050e-01
5.24659574e-01 -4.45841372e-01 3.16756010e-01 -4.96064931e-01
-6.27515912e-02 -2.47964889e-01 -7.86289573e-01 2.59486961e+00
-9.69747543e-01 6.52533770e-01 8.31604004e-04 -9.07506406e-01
3.27416629e-01 4.59764689e-01 1.99757159e-01 -1.11665535e+00
-1.76178694e-01 4.04252380e-01 -4.61893588e-01 -6.77617848e-01
7.37465203e-01 8.15464184e-02 -5.77923417e-01 1.82594225e-01
6.46120548e-01 -6.00039214e-02 2.93638945e-01 4.09059644e-01
2.72268802e-01 2.25372404e-01 1.24950685e-01 -4.46910501e-01
8.74885023e-01 -1.64857939e-01 8.50436464e-02 6.99320793e-01
2.14361921e-01 5.08032739e-01 1.06215253e-01 -1.21872298e-01
-8.26487064e-01 -1.22296107e+00 -2.27076143e-01 1.89337599e+00
2.34861806e-01 -4.81888831e-01 -3.35448444e-01 -8.67504060e-01
-4.76956218e-02 3.70485008e-01 -6.25561059e-01 1.51473299e-01
-2.18244612e-01 -6.09546483e-01 6.77571416e-01 2.18129218e-01
5.64638555e-01 -8.30246091e-01 -1.37585178e-01 -9.73025113e-02
-7.90681958e-01 -1.62970710e+00 -8.46904933e-01 2.18958750e-01
-4.06224012e-01 -9.66660917e-01 -1.11195803e+00 -1.26369727e+00
8.17433298e-01 2.76425779e-01 1.38302958e+00 -7.36210793e-02
-7.62464628e-02 1.00655198e+00 -4.96508598e-01 -1.23376869e-01
-5.95220327e-01 3.92947614e-01 -2.06377357e-01 1.76083431e-01
1.51354074e-01 1.60319865e-01 -4.82663140e-02 3.87706101e-01
-1.06219149e+00 3.68804723e-01 8.04307759e-01 8.78504515e-01
6.60215974e-01 -5.42479813e-01 4.13552850e-01 -7.52068102e-01
3.49198312e-01 -5.44848382e-01 -4.58065987e-01 1.10745215e+00
-2.20570877e-01 1.60142496e-01 1.39196992e-01 -7.19291151e-01
-1.14221060e+00 2.20137715e-01 1.82093084e-01 -4.32110697e-01
-8.07789788e-02 9.78921413e-01 -2.33429000e-01 -2.63750792e-01
3.57263416e-01 4.25026447e-01 -4.48797464e-01 -4.45172399e-01
9.52449977e-01 6.32275939e-01 7.86126912e-01 -1.10644031e+00
6.19681358e-01 2.63600707e-01 -4.01825219e-01 -6.80563033e-01
-7.52970815e-01 -4.64635074e-01 -8.71756196e-01 -3.01921129e-01
1.21593392e+00 -1.61157393e+00 -3.20492983e-01 4.53291327e-01
-1.26688564e+00 -3.89854252e-01 3.01628262e-01 5.18772244e-01
-3.76631707e-01 2.99218833e-01 -6.01910293e-01 -2.48414472e-01
-3.30159850e-02 -1.50531089e+00 1.42968321e+00 2.81540994e-02
1.91935703e-01 -1.15025151e+00 1.23412147e-01 7.39905775e-01
2.36660391e-01 -2.14227036e-01 1.04243505e+00 -5.64268887e-01
-4.88382131e-01 1.80097088e-01 -6.25004709e-01 2.57498920e-01
9.07045156e-02 -1.26319855e-01 -8.73127520e-01 -5.65253317e-01
-7.46434927e-01 -1.22748971e+00 1.09380877e+00 -1.43107906e-01
9.98543680e-01 1.92096848e-02 -2.51990438e-01 5.61314642e-01
1.52902853e+00 -2.19862789e-01 5.62957644e-01 3.71289819e-01
8.72718751e-01 6.80388749e-01 4.02518898e-01 -1.97212636e-01
1.14089000e+00 8.76434326e-01 1.73140764e-01 -4.29018885e-01
-2.11870417e-01 -1.50611877e-01 4.98085529e-01 1.44113553e+00
1.33942410e-01 4.29340713e-02 -1.16963422e+00 7.33282208e-01
-1.80879903e+00 -5.64638793e-01 -9.52699110e-02 2.08491611e+00
1.16150761e+00 -4.83038247e-01 -1.54041529e-01 -8.23843360e-01
5.70570052e-01 -1.32693619e-01 -2.60647207e-01 -2.00648248e-01
-6.88479781e-01 1.90333687e-02 4.79686350e-01 5.97773433e-01
-1.49095416e+00 1.27379751e+00 6.00213909e+00 1.15301168e+00
-1.28521228e+00 7.87045836e-01 3.29661280e-01 1.58137813e-01
-5.25555134e-01 -7.58120492e-02 -6.37416363e-01 2.28147075e-01
8.39627326e-01 4.54408042e-02 4.40540761e-01 4.85607117e-01
-6.27947450e-01 -8.90061483e-02 -1.14323258e+00 1.39948142e+00
4.11569089e-01 -1.18441665e+00 4.77210224e-01 -1.27252251e-01
1.23969018e+00 6.68867052e-01 3.36079001e-01 8.26003551e-01
5.03633618e-01 -9.55215335e-01 9.43080485e-01 6.31623447e-01
1.27288353e+00 -5.34450114e-01 5.24320662e-01 3.04326601e-02
-1.45532203e+00 1.16012134e-01 -1.24189839e-01 7.20907390e-01
1.09538637e-01 1.37011781e-01 -3.40056092e-01 1.06693745e+00
8.59445632e-01 7.47421086e-01 -1.03092694e+00 6.84204638e-01
-1.44106507e-01 -2.65751723e-02 1.21881194e-01 4.35293525e-01
3.27242732e-01 1.06474198e-01 1.60231233e-01 1.84607577e+00
3.56513500e-01 -6.07414007e-01 5.77478647e-01 5.87544560e-01
-3.74314427e-01 3.85092348e-01 -6.47226810e-01 -5.31982481e-02
7.63906986e-02 1.32892311e+00 -2.89883137e-01 -5.85443258e-01
-1.07524049e+00 1.39480495e+00 4.65937197e-01 7.77869225e-01
-8.71997237e-01 -1.23406798e-01 3.43218058e-01 -5.20523727e-01
1.58906579e-01 -3.26923162e-01 2.35571817e-01 -1.74089539e+00
-3.31158251e-01 -1.13131785e+00 6.80645823e-01 -1.05811584e+00
-1.49062061e+00 6.65957749e-01 8.51367638e-02 -1.26605856e+00
-1.49494618e-01 -6.12241805e-01 1.42247960e-01 9.69202042e-01
-2.01559806e+00 -1.92308605e+00 1.08170360e-02 1.18699932e+00
6.98992848e-01 -5.07160425e-01 1.07236695e+00 7.75791168e-01
-2.68130720e-01 7.83243537e-01 2.37355784e-01 3.21207285e-01
1.36232138e+00 -1.04434180e+00 -2.17009246e-01 4.72093523e-01
4.70080286e-01 6.08914435e-01 1.13950714e-01 -4.92587149e-01
-1.37451434e+00 -1.21179879e+00 1.03354645e+00 -6.65032983e-01
7.41145909e-01 -4.39264089e-01 -8.65657270e-01 8.27005625e-01
1.01261020e+00 -1.32933393e-01 8.76584351e-01 3.50136876e-01
-1.01633632e+00 4.13386375e-02 -5.50999999e-01 4.78816360e-01
5.61517537e-01 -1.45304191e+00 -5.05952954e-01 5.41740000e-01
6.08169317e-01 -4.98502433e-01 -1.07134163e+00 4.89333093e-01
5.25607228e-01 -3.27034056e-01 1.13344944e+00 -7.34933376e-01
5.77120066e-01 -1.89705104e-01 -7.54521489e-01 -1.42666149e+00
-1.70203485e-02 -1.91159979e-01 4.21303391e-01 1.20747662e+00
6.92728043e-01 -3.30973089e-01 -4.61424366e-02 1.22204937e-01
-1.15345001e-01 -2.84822583e-01 -1.11349714e+00 -8.37346852e-01
5.83171248e-01 -4.42449391e-01 6.15103617e-02 1.58437383e+00
-1.27363028e-02 5.65289378e-01 -6.17847502e-01 3.04784149e-01
5.25009990e-01 5.84422126e-02 7.84979045e-01 -8.30731034e-01
-4.91200648e-02 -4.22195256e-01 1.37459189e-01 -1.06151640e+00
7.95555890e-01 -1.61915481e+00 4.63303663e-02 -1.31321418e+00
8.69158566e-01 -1.16536044e-01 -4.54706222e-01 8.84363711e-01
-2.72561520e-01 4.77042913e-01 3.23640615e-01 3.62229049e-01
-1.25165665e+00 6.38453066e-01 1.25648439e+00 -5.49048245e-01
2.47324809e-01 -8.40849876e-01 -3.66218418e-01 4.65309292e-01
2.97542691e-01 -4.88276541e-01 -4.38918471e-01 -1.18601763e+00
3.95602405e-01 4.02007513e-02 3.94587487e-01 -7.62758195e-01
1.47127137e-01 -4.68603782e-02 2.51495719e-01 -6.34007514e-01
3.67465168e-01 -8.63934219e-01 -4.71512899e-02 6.70281872e-02
-6.42823160e-01 3.74427289e-01 5.53666949e-01 4.72070217e-01
-5.49505711e-01 -2.13717427e-02 5.92117012e-01 -1.15318991e-01
-1.12889254e+00 -1.63810682e-02 -3.88944685e-01 1.77252635e-01
6.42506838e-01 3.36591244e-01 -5.98737597e-01 -2.97135174e-01
-8.74606788e-01 7.22233236e-01 3.57799292e-01 1.01345026e+00
4.91301924e-01 -1.81009305e+00 -9.33265626e-01 -3.04092616e-02
9.67560172e-01 -7.84781277e-01 3.99393022e-01 9.24627900e-01
-1.58124939e-01 8.10007989e-01 -2.58711398e-01 -9.41789269e-01
-1.45677948e+00 3.85517508e-01 2.89668232e-01 -3.55151057e-01
8.25991333e-02 7.70350337e-01 3.25364769e-01 -1.39013517e+00
1.04049981e-01 -5.52568622e-02 -1.27367049e-01 2.11161062e-01
2.58014560e-01 -3.11298758e-01 4.75124083e-02 -1.19336867e+00
-5.23448110e-01 8.15049529e-01 -7.14584962e-02 -5.35912097e-01
8.94322276e-01 -4.64504898e-01 -3.86389613e-01 7.53188014e-01
1.67590499e+00 -3.26105356e-02 -7.80555308e-01 -6.56831145e-01
1.42728880e-01 -1.69381961e-01 -1.83180794e-01 -1.18600774e+00
-1.07643878e+00 9.55827653e-01 8.13093722e-01 -4.72001612e-01
1.06073773e+00 5.11942446e-01 4.86676306e-01 8.61666977e-01
3.93818200e-01 -1.26007867e+00 4.76804703e-01 9.15073574e-01
1.10203338e+00 -1.67296183e+00 -3.08695138e-01 -1.52724311e-01
-1.13372254e+00 9.67249751e-01 6.22728169e-01 5.70396960e-01
5.04881680e-01 -2.45594099e-01 4.71574038e-01 8.68227426e-03
-7.38960981e-01 -4.62717682e-01 9.33161616e-01 5.27090549e-01
7.68621385e-01 1.03370860e-01 6.35017753e-02 3.90492141e-01
3.50188822e-01 -1.28306419e-01 1.06750894e-02 9.59747493e-01
1.00149550e-01 -1.37991166e+00 -3.54109734e-01 -9.39904824e-02
-3.44868422e-01 -6.06485248e-01 -2.71951586e-01 7.39269733e-01
3.47531915e-01 8.19061816e-01 -9.98759419e-02 -3.35850626e-01
1.77968130e-01 3.55549484e-01 6.52679563e-01 -6.52921379e-01
-4.75014150e-01 3.52384180e-01 1.42774686e-01 -5.09952068e-01
-8.84505033e-01 -5.06996930e-01 -9.05555844e-01 1.24513082e-01
-1.45757273e-02 -1.19805131e-02 8.45834970e-01 9.94613945e-01
4.28265750e-01 4.35830593e-01 1.58215702e-01 -7.33038187e-01
9.06887054e-02 -8.68401527e-01 -4.31426130e-02 6.60450161e-01
1.69413522e-01 -6.05298400e-01 7.82801732e-02 4.83571738e-01] | [11.172224998474121, 1.5543960332870483] |
669bbf4d-b218-4f28-8f5a-a479fcef29df | aspect-based-sentiment-analysis-using-local | 2207.02424 | null | https://arxiv.org/abs/2207.02424v2 | https://arxiv.org/pdf/2207.02424v2.pdf | Aspect-Based Sentiment Analysis using Local Context Focus Mechanism with DeBERTa | Text sentiment analysis, also known as opinion mining, is research on the calculation of people's views, evaluations, attitude and emotions expressed by entities. Text sentiment analysis can be divided into text-level sentiment analysis, sen-tence-level sentiment analysis and aspect-level sentiment analysis. Aspect-Based Sentiment Analysis (ABSA) is a fine-grained task in the field of sentiment analysis, which aims to predict the polarity of aspects. The research of pre-training neural model has significantly improved the performance of many natural language processing tasks. In recent years, pre training model (PTM) has been applied in ABSA. Therefore, there has been a question, which is whether PTMs contain sufficient syntactic information for ABSA. In this paper, we explored the recent DeBERTa model (Decoding-enhanced BERT with disentangled attention) to solve Aspect-Based Sentiment Analysis problem. DeBERTa is a kind of neural language model based on transformer, which uses self-supervised learning to pre-train on a large number of original text corpora. Based on the Local Context Focus (LCF) mechanism, by integrating DeBERTa model, we purpose a multi-task learning model for aspect-based sentiment analysis. The experiments result on the most commonly used the laptop and restaurant datasets of SemEval-2014 and the ACL twitter dataset show that LCF mechanism with DeBERTa has significant improvement. | ['Zhe Xue', 'Zeli Guan', 'Ang Li', 'Junping Du', 'Tianyu Zhao'] | 2022-07-06 | null | null | null | null | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 3.38334665e-02 -1.57928318e-01 -2.80235648e-01 -7.62057722e-01
-4.63037819e-01 -3.25721681e-01 7.48071671e-01 4.13567156e-01
-4.14716989e-01 4.41508085e-01 7.77091801e-01 -4.94678974e-01
1.58193126e-01 -8.76367390e-01 -4.45598096e-01 -5.38027883e-01
2.85413295e-01 2.92520970e-01 -2.00523317e-01 -9.11243498e-01
4.97779816e-01 -4.24488157e-01 -1.35969436e+00 8.36533487e-01
5.92351913e-01 1.15292192e+00 -2.88764298e-01 4.35148835e-01
-7.39759207e-01 8.41539383e-01 -5.46373427e-01 -8.71938348e-01
-1.47106797e-01 -3.44973177e-01 -9.10993397e-01 -1.19146593e-01
-1.86109811e-01 3.09393346e-01 6.23805583e-01 9.43604589e-01
5.36897779e-01 7.37725571e-03 6.51902735e-01 -1.43883514e+00
-8.05548668e-01 8.71972203e-01 -7.87214041e-01 1.85027093e-01
3.80582243e-01 -4.96180564e-01 1.36305404e+00 -9.51451898e-01
1.83123142e-01 1.49386537e+00 7.01908588e-01 2.43186235e-01
-3.98935080e-01 -6.75680697e-01 6.47774756e-01 1.92997217e-01
-7.42234826e-01 3.79223861e-02 9.34066713e-01 -2.86041409e-01
1.42776930e+00 1.51890069e-01 8.94726634e-01 1.04019737e+00
7.80218124e-01 1.08359313e+00 1.67327034e+00 -4.18086231e-01
1.36658385e-01 5.39441884e-01 9.55925643e-01 3.32781881e-01
2.54018605e-01 -4.85513121e-01 -8.78593564e-01 -1.51178688e-01
-2.78016478e-01 -1.37600213e-01 2.54517406e-01 2.16706768e-01
-1.06186938e+00 1.01766205e+00 8.91333818e-02 3.49801719e-01
-3.04441631e-01 -3.01268369e-01 8.86335015e-01 5.65564334e-01
1.10062528e+00 3.74355078e-01 -1.28035808e+00 -1.46863088e-01
-5.36475241e-01 2.43317068e-01 1.17637384e+00 6.18745029e-01
6.30396605e-01 7.26874322e-02 -1.39373258e-01 6.95545137e-01
6.95734441e-01 7.73542345e-01 9.95357931e-01 8.23607482e-03
5.29850483e-01 1.15671635e+00 -3.17514002e-01 -1.28025377e+00
-6.27937615e-01 -3.56687039e-01 -8.18251967e-01 -2.36920547e-02
-1.91086844e-01 -6.42250896e-01 -7.43986428e-01 1.58680534e+00
4.76947516e-01 -3.33897322e-01 3.44627827e-01 7.29109526e-01
1.18369472e+00 9.08095181e-01 2.47621849e-01 -1.42195314e-01
1.96992087e+00 -1.00694406e+00 -1.00430429e+00 -4.78622377e-01
7.69430697e-01 -9.61440325e-01 1.07099140e+00 6.75424457e-01
-4.88497674e-01 -4.25546587e-01 -1.06282043e+00 -1.39018387e-01
-1.32536089e+00 7.74886310e-02 9.60516393e-01 8.97432089e-01
-8.40288639e-01 -2.38219108e-02 -3.30297202e-01 -2.76060522e-01
4.05846715e-01 3.67550969e-01 -4.37178552e-01 2.74682164e-01
-1.60396361e+00 7.41797447e-01 2.21125394e-01 3.25810820e-01
-2.96678394e-01 -2.82952219e-01 -1.05193889e+00 -9.32932198e-02
4.03539121e-01 -6.13777816e-01 1.05518377e+00 -1.47956359e+00
-1.39881277e+00 8.98619533e-01 -5.92510700e-01 -9.67182443e-02
-3.77468705e-01 -4.26044017e-01 -6.42888129e-01 -5.93465328e-01
4.95981604e-01 1.26646176e-01 8.13775003e-01 -9.05524313e-01
-6.76061809e-01 -7.91293681e-01 1.80962116e-01 4.96773750e-01
-5.89903772e-01 5.10149777e-01 -4.30794172e-02 -5.86252153e-01
-2.85725057e-01 -8.28489602e-01 -3.17225456e-01 -8.64273667e-01
-3.91884059e-01 -6.56257153e-01 9.21643972e-01 -4.45639312e-01
1.21156704e+00 -1.86775541e+00 -9.37496945e-02 1.09968558e-01
6.34906217e-02 2.13051304e-01 -9.66477543e-02 3.88511419e-01
-2.90653348e-01 3.29159498e-01 -1.70808118e-02 -4.83614922e-01
2.31324825e-02 -1.33649737e-01 -5.14600635e-01 2.21015345e-02
1.05472945e-01 9.79637504e-01 -5.16653419e-01 -4.75290596e-01
-1.82033867e-01 2.92141616e-01 -5.60306132e-01 1.70319527e-01
-3.64813656e-01 1.95037246e-01 -8.52489531e-01 7.31856585e-01
6.13321185e-01 -2.27492571e-01 -1.05114050e-01 -4.19757456e-01
-2.23093525e-01 5.63276172e-01 -9.95177150e-01 1.30345809e+00
-6.73271537e-01 5.34595251e-01 -1.28143609e-01 -1.07199550e+00
9.44112480e-01 3.85759711e-01 2.53940344e-01 -8.25732887e-01
7.65325427e-01 2.01379582e-02 -5.95157109e-02 -6.59023583e-01
8.33829582e-01 -5.72106302e-01 -6.97799027e-01 7.63966620e-01
2.66294507e-03 -1.50137752e-01 2.38530561e-01 1.70482099e-01
4.77134258e-01 2.31041938e-01 5.99095345e-01 -4.90252376e-01
1.00726640e+00 2.37080872e-01 4.72169548e-01 1.82157040e-01
-6.63039535e-02 2.99104065e-01 7.88057566e-01 -7.06488669e-01
-5.84122062e-01 -1.69181690e-01 -8.03750530e-02 1.57270789e+00
8.63974914e-02 -8.98871064e-01 -5.81452072e-01 -8.46716523e-01
-3.42719257e-01 6.48749709e-01 -9.62660313e-01 -5.05209640e-02
-2.51186132e-01 -1.40521848e+00 4.96805571e-02 4.72537667e-01
7.64344811e-01 -1.29874575e+00 -3.36019918e-02 3.06514781e-02
-1.98528722e-01 -1.11186588e+00 -2.66436398e-01 3.33443254e-01
-5.22446990e-01 -8.62593353e-01 -2.42710397e-01 -7.99871743e-01
4.94540483e-01 2.25311384e-01 1.19349492e+00 -3.72620910e-01
3.27676475e-01 1.56767815e-01 -8.06957603e-01 -1.19999766e+00
7.17974529e-02 4.81912166e-01 -1.67541206e-02 3.44511032e-01
1.14849305e+00 -4.46840286e-01 -3.49020898e-01 1.66967556e-01
-7.80487537e-01 6.56124949e-02 6.64912701e-01 5.49724460e-01
5.36896408e-01 5.73817044e-02 8.14839005e-01 -1.46174765e+00
1.27738321e+00 -7.66615689e-01 -1.55505538e-01 -2.41341814e-02
-8.66010189e-01 -4.39476036e-02 8.30820501e-01 -1.47750378e-01
-1.23920226e+00 -6.40639544e-01 -4.21452701e-01 3.02523881e-01
-7.94599950e-02 1.14404702e+00 -3.09094667e-01 4.48546410e-01
2.57997900e-01 3.04478914e-01 -2.55172282e-01 -1.24046169e-01
2.56902903e-01 1.10953712e+00 -2.81614929e-01 -4.09592986e-01
2.02759713e-01 3.87674183e-01 -2.29776800e-01 -7.01061249e-01
-1.59248698e+00 -7.12351978e-01 -3.43422890e-01 -2.50912994e-01
1.06549609e+00 -1.17527461e+00 -8.31173122e-01 6.69145107e-01
-1.00567484e+00 9.13693905e-02 6.72650412e-02 4.08953071e-01
-1.34121537e-01 1.96403742e-01 -3.46902966e-01 -8.38772416e-01
-9.65827167e-01 -1.11359870e+00 1.29698086e+00 4.26698267e-01
-4.33006287e-01 -1.18459761e+00 2.62695909e-01 8.30957830e-01
5.45444191e-01 2.56373957e-02 9.53876436e-01 -1.09141326e+00
3.71935889e-02 -2.80340940e-01 3.67736965e-02 4.70131785e-01
1.15330704e-01 -3.97694290e-01 -1.02280116e+00 2.51653735e-02
3.65429401e-01 -5.75967610e-01 7.87561119e-01 2.63375282e-01
7.55979657e-01 -3.93179893e-01 -1.62527990e-02 2.88316458e-01
1.23215997e+00 8.40093717e-02 4.14690167e-01 7.47842848e-01
8.48367333e-01 6.51509345e-01 8.54009807e-01 3.03072602e-01
1.01856625e+00 5.00315800e-02 2.05804080e-01 -2.10909061e-02
2.38526657e-01 -1.29340902e-01 7.54461706e-01 1.49218285e+00
-1.52485698e-01 -2.67823011e-01 -6.19084418e-01 6.17884278e-01
-1.87431991e+00 -6.95785761e-01 -4.92849380e-01 1.47580409e+00
7.30050504e-01 3.87230843e-01 -6.38299510e-02 2.54765183e-01
2.81932950e-01 6.85252547e-01 -3.59315574e-01 -1.13452959e+00
-4.09488440e-01 4.94848154e-02 9.71657336e-02 3.88735503e-01
-1.32744288e+00 1.04153049e+00 5.20702076e+00 8.75667930e-01
-1.18924809e+00 3.15044254e-01 8.75872612e-01 3.87814194e-01
-6.30319118e-01 9.31999832e-03 -1.16265178e+00 3.53309870e-01
9.48005021e-01 -2.99630016e-01 -1.33194044e-01 1.06295633e+00
2.69503295e-01 -1.38650209e-01 -6.17550313e-01 6.20362043e-01
6.99887216e-01 -8.62747371e-01 3.64364207e-01 -1.27735794e-01
8.89847815e-01 1.00739248e-01 1.86864570e-01 8.76272559e-01
1.40640631e-01 -8.52632642e-01 4.78186011e-01 2.24385723e-01
3.06812704e-01 -8.05726349e-01 1.46984875e+00 2.97967136e-01
-1.19129312e+00 1.29749939e-01 -2.76814193e-01 -4.39923912e-01
1.62795857e-01 7.38218844e-01 -3.74591589e-01 6.22804523e-01
9.32385743e-01 1.21776736e+00 -5.95001817e-01 3.15348327e-01
-3.59407395e-01 8.09973478e-01 5.83577808e-03 -7.86069751e-01
4.04839963e-01 -5.35017490e-01 3.72691303e-01 1.23664284e+00
-5.55598596e-03 2.38587391e-02 -1.37730334e-02 2.98564702e-01
-4.20282111e-02 7.05292404e-01 -5.94072580e-01 -1.05971299e-01
-4.57882732e-01 1.69124234e+00 -7.32658923e-01 -4.37461346e-01
-6.67558193e-01 5.29740810e-01 7.20053911e-02 9.68599841e-02
-4.84287351e-01 -3.45928639e-01 3.91490996e-01 -1.10881396e-01
3.40535313e-01 1.28014341e-01 -5.75903952e-01 -1.45417047e+00
-6.79276437e-02 -1.28520334e+00 4.30386513e-01 -8.61219823e-01
-1.43314564e+00 9.70436394e-01 -1.66215360e-01 -1.00573266e+00
-8.35184157e-02 -1.03663778e+00 -9.54051077e-01 7.56077886e-01
-1.84607232e+00 -1.59883046e+00 7.04589263e-02 7.22597241e-01
8.16253245e-01 -3.67471069e-01 8.62897217e-01 1.80740550e-01
-5.50575674e-01 3.26292783e-01 -4.04931158e-01 1.57498553e-01
6.86321437e-01 -1.31859517e+00 1.73622355e-01 5.69155991e-01
-2.04497516e-01 6.74124062e-01 6.95445180e-01 -6.25190437e-01
-1.49310732e+00 -9.84035909e-01 1.48195720e+00 -7.09260643e-01
7.31702864e-01 -4.25244957e-01 -5.77594578e-01 8.28391075e-01
6.76404476e-01 -5.43953776e-01 1.37185347e+00 8.63442659e-01
-3.23284894e-01 -4.02391523e-01 -7.27329612e-01 5.52948654e-01
2.93123752e-01 -4.30239230e-01 -1.04557240e+00 2.36101255e-01
8.36117864e-01 1.52889704e-02 -6.41695678e-01 4.48350221e-01
5.63375533e-01 -9.08990860e-01 5.66376209e-01 -7.83675015e-01
7.64171362e-01 -3.19538593e-01 -2.83678710e-01 -1.38595092e+00
1.11879207e-01 -7.04342499e-02 2.46874109e-01 1.44412267e+00
8.66389513e-01 -7.82689929e-01 5.53520739e-01 5.10573387e-01
4.89513353e-02 -1.09891534e+00 -5.55757225e-01 1.53750181e-01
1.48548782e-01 -8.21394980e-01 7.21994340e-01 1.02328515e+00
2.94143111e-01 1.44577444e+00 -3.01339060e-01 -1.56752348e-01
1.19289078e-01 5.26507556e-01 6.97034776e-01 -1.09503126e+00
-3.46048772e-02 -4.90365624e-01 -7.02328533e-02 -7.34341443e-01
3.55722249e-01 -8.86592448e-01 -3.51625383e-01 -1.63243592e+00
2.90051848e-01 1.73062347e-02 -3.90917450e-01 5.58476985e-01
-4.86454219e-01 6.13246262e-02 4.29990608e-03 -1.75634727e-01
-1.03376031e+00 7.66242385e-01 1.22629297e+00 -1.91013396e-01
-1.36906561e-02 3.45359564e-01 -1.42089581e+00 9.28913891e-01
9.04236913e-01 -5.11407256e-01 -4.09855098e-01 -2.67551929e-01
1.46179497e+00 -3.21101129e-01 -4.08590972e-01 -3.54169846e-01
1.47609249e-01 -6.10242672e-02 1.63898960e-01 -1.00158525e+00
3.36177021e-01 -7.44977951e-01 -6.39095843e-01 1.16574287e-01
-3.36856157e-01 4.12779480e-01 1.85204208e-01 3.96835804e-01
-7.00170815e-01 -9.64472517e-02 3.84083688e-02 -2.17991650e-01
-7.49173224e-01 2.24247709e-01 -7.51763284e-01 1.03090763e-01
7.11738586e-01 8.47843587e-02 -2.53307551e-01 -3.42597693e-01
-2.53608197e-01 4.04244304e-01 -3.30411762e-01 6.77744925e-01
5.12808144e-01 -1.12525189e+00 -6.93580627e-01 2.45205596e-01
3.27575058e-01 5.26661947e-02 2.99237967e-01 7.73873687e-01
-3.42071876e-02 5.23497105e-01 1.44890636e-01 -2.50566989e-01
-1.20862830e+00 4.53376651e-01 5.03879525e-02 -7.08636224e-01
1.28173649e-01 6.89088821e-01 1.88524410e-01 -1.02492750e+00
-3.65428060e-01 -3.71119827e-01 -1.27332222e+00 5.74155927e-01
5.23151755e-01 -6.00065589e-02 1.36297375e-01 -9.52759206e-01
-4.09095705e-01 9.52224016e-01 -2.54038841e-01 -1.30678415e-01
1.42928183e+00 -2.63003320e-01 -6.58839047e-01 8.69504273e-01
1.28066969e+00 2.54937202e-01 -8.57977644e-02 -1.01110429e-01
-1.82360649e-01 4.38287407e-02 2.82000571e-01 -9.49830711e-01
-9.78016794e-01 8.99586201e-01 1.29409686e-01 4.68116730e-01
1.06223786e+00 -1.52466774e-01 8.42342019e-01 5.34511685e-01
2.62603819e-01 -1.29689229e+00 -4.22632834e-03 1.05352485e+00
7.31513977e-01 -1.66920114e+00 4.87726851e-04 -1.02901600e-01
-1.14845335e+00 8.87778759e-01 7.86086142e-01 -1.31118655e-01
1.30234993e+00 3.03023875e-01 4.00548220e-01 -5.40184677e-01
-1.00180995e+00 -1.84767693e-01 3.08539093e-01 2.49502555e-01
7.69155204e-01 1.08957574e-01 -7.26490617e-01 1.55310702e+00
-7.58557558e-01 -9.59595442e-02 4.17810261e-01 8.93542886e-01
-3.52507651e-01 -9.69460368e-01 -3.27625036e-01 5.34968436e-01
-9.08040702e-01 -4.69933838e-01 -4.26307380e-01 4.51139182e-01
2.68931419e-01 1.30220628e+00 -2.59788424e-01 -5.14570117e-01
4.09018278e-01 6.35112375e-02 -3.39987159e-01 -5.39245784e-01
-1.09922957e+00 -5.25737042e-03 4.33662176e-01 -3.78589243e-01
-1.03670669e+00 -3.86949927e-01 -9.42555904e-01 -1.02250651e-01
-4.79720175e-01 5.40038347e-01 9.90434825e-01 1.43310177e+00
3.87575030e-01 6.53036475e-01 8.13387573e-01 -3.10315251e-01
7.33848140e-02 -1.18323052e+00 -5.06790340e-01 3.23957920e-01
2.01235965e-01 -1.54824942e-01 -3.94211590e-01 -1.45766228e-01] | [11.351303100585938, 6.765265941619873] |
7bcbe56b-2b5e-4da0-8a97-904994e2bf05 | accurate-point-cloud-registration-with-robust | 2111.00648 | null | https://arxiv.org/abs/2111.00648v1 | https://arxiv.org/pdf/2111.00648v1.pdf | Accurate Point Cloud Registration with Robust Optimal Transport | This work investigates the use of robust optimal transport (OT) for shape matching. Specifically, we show that recent OT solvers improve both optimization-based and deep learning methods for point cloud registration, boosting accuracy at an affordable computational cost. This manuscript starts with a practical overview of modern OT theory. We then provide solutions to the main difficulties in using this framework for shape matching. Finally, we showcase the performance of transport-enhanced registration models on a wide range of challenging tasks: rigid registration for partial shapes; scene flow estimation on the Kitti dataset; and nonparametric registration of lung vascular trees between inspiration and expiration. Our OT-based methods achieve state-of-the-art results on Kitti and for the challenging lung registration task, both in terms of accuracy and scalability. We also release PVT1010, a new public dataset of 1,010 pairs of lung vascular trees with densely sampled points. This dataset provides a challenging use case for point cloud registration algorithms with highly complex shapes and deformations. Our work demonstrates that robust OT enables fast pre-alignment and fine-tuning for a wide range of registration models, thereby providing a new key method for the computer vision toolbox. Our code and dataset are available online at: https://github.com/uncbiag/robot. | ['Marc Niethammer', 'Raul San Jose Estepar', 'Ruben San Jose Estepar', 'Ariel Hernán Curiale', 'Peirong Liu', 'Jean Feydy', 'Zhengyang Shen'] | 2021-11-01 | null | http://proceedings.neurips.cc/paper/2021/hash/2b0f658cbffd284984fb11d90254081f-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/2b0f658cbffd284984fb11d90254081f-Paper.pdf | neurips-2021-12 | ['scene-flow-estimation'] | ['computer-vision'] | [-2.81098008e-01 -3.19626093e-01 -2.25478992e-01 3.24191432e-03
-1.24131799e+00 -6.19555473e-01 4.42652732e-01 5.79002798e-02
-1.57859564e-01 2.76912391e-01 -1.97502568e-01 -1.84512705e-01
-1.59203038e-01 -5.69031060e-01 -5.68238199e-01 -7.11389482e-01
-9.63845924e-02 1.27487886e+00 2.58550763e-01 -4.12824415e-02
-9.79433954e-02 9.69030559e-01 -1.06865990e+00 -4.66755824e-03
7.03947842e-01 9.27975953e-01 -6.15239069e-02 8.18977654e-01
-1.21251166e-01 1.36833251e-01 1.26663428e-02 -3.87028992e-01
4.96315569e-01 3.14800702e-02 -8.93465757e-01 -2.10951194e-01
1.09636974e+00 -1.30075425e-01 -4.18113053e-01 5.08643448e-01
9.81607616e-01 2.02276126e-01 7.35707164e-01 -1.22831917e+00
-3.26108098e-01 2.07704961e-01 -3.74449760e-01 2.75090069e-01
6.82381690e-02 3.79853517e-01 1.03892040e+00 -1.11709702e+00
6.37785435e-01 8.87763739e-01 1.37500751e+00 4.55509543e-01
-1.08850265e+00 -4.48054045e-01 -5.52299261e-01 -8.41820054e-03
-1.34439385e+00 -4.19031292e-01 4.88696814e-01 -6.45336986e-01
1.03434718e+00 3.75398636e-01 8.21018040e-01 5.74514866e-01
1.84820145e-01 3.18449259e-01 8.15845966e-01 8.82853046e-02
-1.75886169e-01 -2.82836735e-01 -1.37648821e-01 8.82668495e-01
1.15719751e-01 3.27565551e-01 -3.35946441e-01 -4.75255489e-01
1.06833160e+00 -1.95525020e-01 -1.58646345e-01 -7.59320796e-01
-1.84435022e+00 5.96492887e-01 7.26888239e-01 4.17450368e-01
-3.22617620e-01 5.80105484e-01 2.90071428e-01 -2.56730840e-02
5.51940918e-01 1.64435789e-01 -5.35344839e-01 -2.51750261e-01
-9.17162478e-01 3.40636224e-01 7.86569297e-01 1.12926114e+00
5.53804815e-01 -5.57981096e-02 -2.63543397e-01 9.78854775e-01
4.34681267e-01 9.91218090e-01 4.06097211e-02 -1.37244856e+00
5.29399633e-01 7.91488364e-02 -2.66501546e-01 -8.70122015e-01
-5.53679287e-01 -4.16462421e-01 -8.31794620e-01 1.20444424e-01
6.31241083e-01 2.19104648e-01 -6.49461389e-01 1.26264536e+00
8.34917009e-01 8.10797513e-01 -2.35906944e-01 7.91161597e-01
1.38626051e+00 2.40699142e-01 -6.14979342e-02 4.77080829e-02
1.39780998e+00 -9.50659335e-01 -3.46712530e-01 1.27770171e-01
7.18467116e-01 -1.12519467e+00 8.57004523e-01 -1.27450407e-01
-1.47431421e+00 -3.81402224e-01 -3.74363422e-01 -3.24597746e-01
3.06923613e-02 -1.14024274e-01 5.13992846e-01 4.82540309e-01
-1.23473644e+00 1.17361450e+00 -1.14583325e+00 -5.12442708e-01
1.01672804e+00 6.82200730e-01 -3.58215898e-01 5.28343692e-02
-4.15989757e-01 9.14056480e-01 -5.09404242e-01 1.47110686e-01
-6.28042161e-01 -1.50302708e+00 -5.63481450e-01 -3.47773641e-01
-7.99189210e-02 -1.37121654e+00 1.43407607e+00 -1.67852938e-02
-1.39024174e+00 1.29809344e+00 -2.30544001e-01 -2.75541782e-01
8.55994046e-01 6.58872500e-02 2.65398681e-01 1.91723272e-01
1.70562848e-01 7.27178752e-01 6.73515260e-01 -7.99178600e-01
-9.54924002e-02 -1.97297633e-01 -7.28894114e-01 1.47185564e-01
7.15646744e-02 1.90393776e-02 -4.43105310e-01 -5.38438737e-01
1.86687976e-01 -1.17940319e+00 -3.20215702e-01 6.51809037e-01
-1.23872630e-01 -1.81450471e-01 8.37008178e-01 -5.05266726e-01
3.08274150e-01 -1.64472353e+00 4.73340675e-02 2.11502671e-01
5.46259880e-01 1.42398804e-01 -1.29688472e-01 1.49055019e-01
-1.28654465e-01 1.03091180e-01 -7.72584617e-01 -8.10182154e-01
-7.12495670e-03 3.65115523e-01 -2.27080584e-01 8.90623152e-01
-9.17262435e-02 1.50983751e+00 -7.93873489e-01 -7.42392957e-01
5.05246520e-01 7.46593237e-01 -5.23131490e-01 8.20303410e-02
-4.09003235e-02 1.20011759e+00 -4.44288582e-01 7.54711866e-01
8.58898997e-01 -3.48260492e-01 -3.90166163e-01 -4.10655409e-01
-2.48769850e-01 3.06039572e-01 -8.86389554e-01 1.79606259e+00
-5.27903557e-01 5.23967147e-01 1.91327006e-01 -6.56546831e-01
7.96637893e-01 4.63833541e-01 1.46340477e+00 -2.24508256e-01
1.69978723e-01 5.68830907e-01 -9.29371417e-02 -2.37283677e-01
8.80310759e-02 -2.85035938e-01 3.81545037e-01 4.99692410e-01
-1.77444309e-01 -1.03386855e+00 -2.87686944e-01 -1.74387336e-01
1.13407695e+00 3.06808770e-01 8.81903544e-02 -4.00951207e-01
5.45696676e-01 3.19702506e-01 3.93797338e-01 5.55079579e-01
-4.27980781e-01 1.06579649e+00 -6.24935478e-02 -5.50402761e-01
-1.13703525e+00 -1.22892821e+00 -7.01231778e-01 5.73174655e-01
-2.64682584e-02 -2.28329256e-01 -5.07543206e-01 -5.30691326e-01
4.96703982e-01 1.07006662e-01 -3.43891621e-01 2.31572390e-01
-1.22195804e+00 -7.44748950e-01 7.34388709e-01 6.19742870e-01
3.61849636e-01 -1.07048643e+00 -2.21225992e-01 8.77595544e-02
-3.21897388e-01 -1.40474796e+00 -8.18853080e-01 -3.64523143e-01
-1.38897729e+00 -1.24230242e+00 -8.09426308e-01 -6.97586358e-01
4.92751151e-01 2.72384405e-01 1.41659343e+00 5.38322091e-01
-5.76047540e-01 9.60485101e-01 1.53410673e-01 -2.14820340e-01
-5.77937305e-01 3.38834286e-01 -6.34671971e-02 -4.18589860e-01
-1.67644441e-01 -8.14675987e-01 -6.65966272e-01 6.52302086e-01
-3.15792918e-01 -3.92921269e-01 6.47691265e-02 4.25251305e-01
1.05866992e+00 -6.48492694e-01 9.03073549e-02 -4.83362317e-01
1.68757230e-01 -3.74297798e-01 -6.86939597e-01 -3.14555392e-02
-2.72088826e-01 -1.88630357e-01 3.03491712e-01 -1.93402871e-01
-5.63811302e-01 2.16730699e-01 -4.78542179e-01 -8.78288150e-01
8.03797320e-03 -1.46557376e-01 4.49788749e-01 -9.22924519e-01
6.72528267e-01 -2.77013294e-02 3.20378751e-01 -5.76472998e-01
4.80202645e-01 1.87060952e-01 7.07520843e-01 -1.04108238e+00
1.25754952e+00 9.79413807e-01 7.11170077e-01 -8.93078566e-01
-6.70990229e-01 -7.82918513e-01 -1.44975877e+00 -1.36135906e-01
9.19846654e-01 -6.04347348e-01 -8.20544720e-01 5.33895671e-01
-1.20095634e+00 -8.11364532e-01 -8.28085363e-01 4.76884216e-01
-1.05433297e+00 6.49312437e-01 -5.22835612e-01 -4.38351594e-02
-8.70184124e-01 -1.35136318e+00 1.53599370e+00 -2.24056169e-01
-3.75125371e-02 -1.54145801e+00 4.99421775e-01 5.71379960e-01
6.64224505e-01 4.03451443e-01 5.16721964e-01 -4.38749701e-01
-9.44255292e-01 1.12594932e-01 -1.40151978e-01 1.70165002e-02
1.54728085e-01 2.63313174e-01 -1.01001537e+00 -3.68342608e-01
2.50460841e-02 1.03343442e-01 5.79617798e-01 6.78717315e-01
1.16493261e+00 1.24734044e-01 -5.44097483e-01 1.27122128e+00
1.05953360e+00 -4.63490486e-01 4.74332631e-01 5.94853908e-02
1.26444674e+00 4.37087417e-01 6.08985901e-01 2.19529673e-01
5.16069353e-01 1.13906741e+00 5.45077026e-01 -2.54839122e-01
-6.54841602e-01 1.88339576e-01 -9.99988616e-02 1.29969227e+00
-4.30396289e-01 4.00948495e-01 -1.36454630e+00 4.78806138e-01
-1.73798859e+00 -6.62861586e-01 -7.85245538e-01 2.39503670e+00
6.37946069e-01 -5.26800632e-01 2.45860651e-01 -4.17617112e-01
5.94604850e-01 5.29713370e-03 -3.93507481e-01 -6.80473968e-02
9.70015228e-02 6.87147796e-01 7.05749273e-01 6.30902648e-01
-1.29211032e+00 9.63999212e-01 6.46718264e+00 6.81427836e-01
-1.16214526e+00 5.34405589e-01 2.15506762e-01 8.19310695e-02
-1.77519396e-01 -6.49387091e-02 -6.43466592e-01 1.01435751e-01
8.07176411e-01 -2.02577353e-01 4.14550841e-01 5.76803625e-01
2.49862000e-01 4.49344039e-01 -1.06937742e+00 1.10911369e+00
-1.83203354e-01 -1.68881452e+00 -4.14314091e-01 2.78764248e-01
6.84768796e-01 9.40687001e-01 -2.38341037e-02 -1.05552413e-01
1.04788348e-01 -1.03655195e+00 3.79574746e-01 5.15959144e-01
1.06365740e+00 -1.36097416e-01 4.26817983e-01 4.32512388e-02
-1.68758821e+00 5.71023285e-01 -5.87252140e-01 5.98750234e-01
3.67354214e-01 6.36412382e-01 -1.00350916e+00 7.60463357e-01
7.73640513e-01 1.30315745e+00 -3.75641048e-01 1.51979363e+00
-6.06236840e-03 4.35050726e-01 -8.08869243e-01 5.61548173e-01
-1.09155975e-01 -5.17970920e-01 1.03802049e+00 1.11304617e+00
5.37136853e-01 -2.40538325e-02 1.19894430e-01 1.06662691e+00
-2.23821431e-01 2.30358556e-01 -8.29673290e-01 5.20326197e-01
4.32249427e-01 1.62192702e+00 -6.46052480e-01 -9.86341387e-02
-5.40815480e-02 4.89270121e-01 2.24007621e-01 -1.19954005e-01
-1.06380367e+00 2.92725772e-01 8.26294601e-01 3.71424764e-01
2.34647945e-01 -5.24136841e-01 -5.25302291e-01 -1.01084208e+00
-5.42497523e-02 -4.18025047e-01 4.46175665e-01 -7.48388290e-01
-1.48319805e+00 4.13110316e-01 1.88119203e-01 -1.43489444e+00
9.43830311e-02 -5.36384344e-01 -8.49002719e-01 7.59429991e-01
-1.56148577e+00 -1.32466769e+00 -6.49787188e-01 6.52599812e-01
2.71843165e-01 7.38451853e-02 6.55196905e-01 6.47200704e-01
-3.74622107e-01 3.54038000e-01 1.60045952e-01 5.63479923e-02
7.37025142e-01 -1.26178682e+00 9.66642857e-01 3.99358302e-01
2.70034432e-01 4.12770182e-01 1.84976667e-01 -6.87742472e-01
-1.59928286e+00 -1.31140387e+00 7.01689839e-01 -1.19976270e+00
5.48332036e-01 -1.44368932e-01 -1.02175558e+00 7.93384850e-01
-2.44938880e-01 7.14145958e-01 2.78343529e-01 -3.59409213e-01
2.58070184e-03 -8.28156024e-02 -1.35023916e+00 3.20014656e-01
1.35063577e+00 -3.17641616e-01 -4.38061506e-01 9.55115259e-01
5.18042326e-01 -1.07491398e+00 -1.63054800e+00 6.36620700e-01
4.96750981e-01 -6.51768565e-01 1.61625528e+00 -1.83384851e-01
-1.03607938e-01 -2.84168869e-01 1.06883571e-01 -1.00339532e+00
-2.05077261e-01 -9.35167611e-01 -6.04376663e-04 8.52036536e-01
-5.99746294e-02 -9.46041822e-01 8.48877311e-01 4.76599574e-01
-5.32036543e-01 -8.30489516e-01 -1.52197421e+00 -7.84131765e-01
5.88811457e-01 -7.35837162e-01 6.43933892e-01 9.74236786e-01
-6.27219021e-01 -1.88272730e-01 2.03613073e-01 1.14898048e-01
1.01941168e+00 1.99706376e-01 1.00309193e+00 -1.41315663e+00
1.46258939e-02 -5.08429348e-01 -2.66623855e-01 -9.57684994e-01
2.27935851e-01 -1.64751112e+00 -1.08005881e-01 -1.68719363e+00
-9.20219421e-02 -1.13960564e+00 2.89891452e-01 6.09593451e-01
1.94899157e-01 5.64284623e-01 1.65873572e-01 8.32487166e-01
-2.48823404e-01 5.62593043e-01 1.75797081e+00 6.11490011e-03
-6.20948188e-02 4.49756265e-01 -1.08451687e-01 6.37205839e-01
9.10363734e-01 -7.91388154e-01 6.56184629e-02 -5.30344546e-01
-2.97835499e-01 3.36362496e-02 8.37794423e-01 -1.05654490e+00
3.11082691e-01 -3.13267447e-02 -1.35073185e-01 -8.72346580e-01
5.69462955e-01 -9.36827660e-01 3.76181424e-01 5.75177193e-01
6.33934662e-02 3.14865291e-01 4.34872419e-01 1.79906294e-01
3.37741852e-01 -3.21771391e-02 1.04502928e+00 -5.20318039e-02
-1.17693953e-01 1.10115767e+00 1.72570765e-01 4.62973505e-01
8.34189892e-01 -2.74841070e-01 -3.83134782e-01 -5.29327169e-02
-7.29564071e-01 1.79558635e-01 5.35167873e-01 1.64195895e-01
5.16835868e-01 -1.49263465e+00 -1.00075543e+00 1.86262220e-01
-1.56394899e-01 5.60474157e-01 7.13683590e-02 1.65210855e+00
-9.64467585e-01 2.90270835e-01 5.74163254e-03 -1.41711426e+00
-1.45178545e+00 2.74924608e-03 1.02687109e+00 -2.32623667e-01
-1.18407476e+00 7.14460552e-01 1.05475903e-01 -1.02702773e+00
-1.83350414e-01 -7.11012661e-01 1.95144519e-01 -3.91334355e-01
-2.82776684e-01 7.26725698e-01 4.36239064e-01 -1.05370247e+00
-6.46188140e-01 1.56138813e+00 4.79749382e-01 1.73207030e-01
1.23899055e+00 1.21206127e-01 -4.38210219e-01 7.81372860e-02
1.31889570e+00 8.01510289e-02 -1.04634786e+00 -2.60392457e-01
-2.52885312e-01 -6.03823602e-01 2.02298388e-02 -3.01533222e-01
-1.50105953e+00 1.06332624e+00 5.56373417e-01 -2.22229153e-01
5.32064021e-01 3.58462363e-01 1.17041898e+00 2.99667537e-01
2.94272959e-01 -4.29212928e-01 -4.16134186e-02 5.71890533e-01
1.00517225e+00 -1.15867746e+00 3.44994515e-01 -8.71745050e-01
-3.50129485e-01 1.13549852e+00 2.47141987e-01 -2.24029422e-01
8.65760922e-01 3.97911787e-01 2.01083288e-01 -5.95779181e-01
-4.33204412e-01 -6.95992112e-02 7.80001223e-01 9.12581682e-01
3.95343035e-01 -2.43661627e-02 1.63194224e-01 -2.59907246e-01
-5.15906036e-01 -1.79858238e-01 1.78368747e-01 6.94847226e-01
-7.47389495e-02 -1.40937090e+00 -5.30190647e-01 4.35821742e-01
-3.84827763e-01 -5.09471931e-02 -1.88187748e-01 8.08628857e-01
-1.99813098e-01 3.35147262e-01 1.84574276e-01 1.34305716e-01
5.14827251e-01 -2.75605112e-01 7.51987696e-01 -6.01430714e-01
-9.53940570e-01 7.10416138e-02 -1.03886545e-01 -7.62591362e-01
-6.49504483e-01 -1.02693498e+00 -1.43359518e+00 -4.84225392e-01
-9.84302908e-02 -8.87540057e-02 6.46712661e-01 7.43899345e-01
4.41042483e-01 3.39961916e-01 5.30839860e-01 -1.27659881e+00
-4.94953573e-01 -4.59872156e-01 -6.45901486e-02 4.22805220e-01
2.54990965e-01 -7.33029544e-01 -3.53674442e-01 -1.80332497e-01] | [13.918859481811523, -2.6381893157958984] |
0ac8e078-0dc6-4ce5-9c30-96725d0d325f | unsupervised-person-re-identification-via-2 | 2104.00202 | null | https://arxiv.org/abs/2104.00202v1 | https://arxiv.org/pdf/2104.00202v1.pdf | Unsupervised Person Re-identification via Simultaneous Clustering and Consistency Learning | Unsupervised person re-identification (re-ID) has become an important topic due to its potential to resolve the scalability problem of supervised re-ID models. However, existing methods simply utilize pseudo labels from clustering for supervision and thus have not yet fully explored the semantic information in data itself, which limits representation capabilities of learned models. To address this problem, we design a pretext task for unsupervised re-ID by learning visual consistency from still images and temporal consistency during training process, such that the clustering network can separate the images into semantic clusters automatically. Specifically, the pretext task learns semantically meaningful representations by maximizing the agreement between two encoded views of the same image via a consistency loss in latent space. Meanwhile, we optimize the model by grouping the two encoded views into same cluster, thus enhancing the visual consistency between views. Experiments on Market-1501, DukeMTMC-reID and MSMT17 datasets demonstrate that our proposed approach outperforms the state-of-the-art methods by large margins. | ['Jun Guo', 'Zhanyu Ma', 'Jiyang Xie', 'Siqing Zhang', 'Jiayan Qiu', 'Junhui Yin'] | 2021-04-01 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [-1.56837896e-01 1.61669895e-01 -2.99844235e-01 -7.15707660e-01
-3.09567481e-01 -3.82223874e-01 7.83423960e-01 -1.11275762e-01
-4.81100589e-01 2.85382062e-01 4.75367069e-01 3.37560922e-01
-1.87527657e-01 -2.36682609e-01 -3.78600687e-01 -6.01775527e-01
2.35015526e-01 6.46557152e-01 -1.74603894e-01 4.36136603e-01
1.99182808e-01 6.57287836e-02 -1.42894936e+00 1.22170597e-01
9.34780777e-01 5.90015173e-01 2.09774062e-01 2.43924130e-02
-7.21098706e-02 5.78301251e-01 -2.59436905e-01 -5.91377199e-01
2.87442327e-01 -4.78182018e-01 -8.77106786e-01 6.10097229e-01
5.15580416e-01 -2.20874503e-01 -5.40953219e-01 1.37292552e+00
3.09287667e-01 4.40340072e-01 8.52553666e-01 -1.54041028e+00
-9.98849154e-01 3.89210135e-01 -8.83788824e-01 1.13837406e-01
2.15471372e-01 -2.90625602e-01 8.38777304e-01 -7.70516992e-01
7.22201347e-01 1.50843596e+00 3.67977619e-01 8.80197763e-01
-1.42176569e+00 -7.25093722e-01 5.53636372e-01 4.69377965e-01
-1.52510023e+00 -5.48074067e-01 8.00718784e-01 -6.46064997e-01
5.02217829e-01 2.88903788e-02 3.40699613e-01 1.07302117e+00
-3.98241341e-01 7.20587730e-01 1.13921332e+00 -2.81027049e-01
9.53879654e-02 4.53559756e-01 2.81827599e-01 5.64658940e-01
2.68317640e-01 -7.91665465e-02 -4.93592888e-01 1.44907877e-01
7.08183885e-01 5.77813983e-01 -7.02747703e-02 -6.65385365e-01
-1.11984301e+00 5.73365986e-01 5.28736353e-01 2.70784974e-01
-3.88491736e-03 -9.25988704e-02 2.83632964e-01 1.04971461e-01
4.28414106e-01 5.68314269e-02 2.38281582e-02 2.72730559e-01
-9.07102466e-01 3.92022207e-02 2.80339003e-01 9.97070789e-01
8.19233000e-01 -2.80128926e-01 -2.03617439e-02 1.03825283e+00
6.29907787e-01 3.29713732e-01 7.03909099e-01 -9.85058367e-01
6.01171672e-01 9.24324155e-01 2.71005258e-02 -1.12072134e+00
-2.25791365e-01 -2.35046580e-01 -1.12449408e+00 5.68771511e-02
3.41140568e-01 1.48889378e-01 -9.95108843e-01 1.89246726e+00
3.28805238e-01 2.38811448e-01 1.54247165e-01 1.10659897e+00
7.18674362e-01 3.70947659e-01 2.52811730e-01 -2.13078558e-01
1.30953074e+00 -1.19025266e+00 -7.16681838e-01 -3.20078939e-01
3.29085320e-01 -4.25481796e-01 7.94121742e-01 9.80530754e-02
-9.50404584e-01 -9.59977925e-01 -1.04260433e+00 -1.48794521e-02
-1.97905257e-01 2.96062708e-01 3.67599934e-01 5.09388089e-01
-1.02266479e+00 2.40982011e-01 -6.80998087e-01 -7.28165150e-01
3.20642948e-01 2.19658509e-01 -7.35480249e-01 -3.17990750e-01
-8.34512591e-01 6.08870327e-01 3.96682292e-01 1.21917650e-02
-8.55568349e-01 -4.77799028e-01 -7.79499173e-01 -5.67058427e-03
2.69183338e-01 -5.63277721e-01 7.81112194e-01 -1.07303810e+00
-1.13479376e+00 1.40483439e+00 -4.40124422e-01 -2.66216159e-01
5.91561377e-01 -1.15384147e-01 -6.62864745e-01 4.32573818e-02
5.85229337e-01 8.30023527e-01 8.86128128e-01 -1.71934426e+00
-6.75312281e-01 -7.53000736e-01 -1.46537870e-01 5.13790786e-01
-5.07438660e-01 1.50139653e-03 -9.49732900e-01 -6.54717267e-01
3.45580071e-01 -1.13610852e+00 -2.26591378e-01 -2.66724497e-01
-5.07182598e-01 -3.24374437e-01 8.25130165e-01 -8.96328628e-01
9.48834121e-01 -2.42104006e+00 2.49453992e-01 3.46516371e-01
3.82885516e-01 -1.22104950e-01 -6.50925189e-02 6.60468042e-02
-3.20404559e-01 7.60496557e-02 3.58375050e-02 -9.30243790e-01
-1.10569817e-03 1.04515873e-01 -8.11339393e-02 5.81669271e-01
-3.96288514e-01 7.86664426e-01 -7.25495219e-01 -6.14804387e-01
1.50225803e-01 1.49931431e-01 -3.46821785e-01 4.77724433e-01
1.31079003e-01 7.31217563e-01 -3.06240976e-01 3.66964489e-01
7.35842168e-01 -6.11115932e-01 4.56455886e-01 -3.74643654e-01
6.27050325e-02 -1.15209289e-01 -1.26192474e+00 1.80449128e+00
-1.46315461e-02 3.58342767e-01 -1.83332011e-01 -1.25233519e+00
7.63243020e-01 2.47785836e-01 6.33220255e-01 -7.72561610e-01
-2.15561241e-01 -3.78612906e-01 -4.16103393e-01 -4.15113926e-01
3.96344155e-01 3.57279517e-02 -6.91413805e-02 6.55287266e-01
-1.06981462e-02 7.56521404e-01 2.23288443e-02 3.77706051e-01
3.56168389e-01 9.35639814e-02 1.13139935e-02 -1.18432306e-01
5.23766339e-01 -1.42139211e-01 6.55103683e-01 6.26665652e-01
-3.68067145e-01 7.09290326e-01 5.32327890e-02 -3.00422102e-01
-1.15620291e+00 -1.18635130e+00 1.42655252e-02 1.06305397e+00
7.10990310e-01 -3.58374000e-01 -8.29657555e-01 -9.53323305e-01
-7.40515888e-02 4.91940677e-01 -7.01023936e-01 -4.04678024e-02
-3.97707403e-01 -5.26652038e-01 2.82891810e-01 5.42434394e-01
9.07461822e-01 -6.47783875e-01 8.79137218e-02 -1.49780482e-01
-5.79084396e-01 -1.28475726e+00 -7.65550733e-01 -3.52300107e-01
-7.90942907e-01 -1.12304556e+00 -9.19398129e-01 -1.04691434e+00
1.25290036e+00 8.75547707e-01 7.39474952e-01 8.50164518e-02
-1.64802462e-01 6.25980616e-01 -2.76492774e-01 2.59665877e-01
-1.20477647e-01 -9.00537893e-02 4.43225771e-01 3.33546072e-01
6.05077505e-01 -3.86114627e-01 -6.86910272e-01 6.74099624e-01
-6.29605055e-01 3.56098264e-01 3.45347166e-01 7.36947358e-01
6.91654205e-01 2.07047150e-01 4.78497565e-01 -8.60103190e-01
3.16199243e-01 -4.28532541e-01 -2.88566619e-01 6.27200842e-01
-9.87185657e-01 1.43316746e-01 3.88567686e-01 -4.14982080e-01
-1.51893234e+00 3.60916890e-02 5.04912853e-01 -6.32828355e-01
-2.81831503e-01 2.22846001e-01 -5.15899599e-01 3.22157234e-01
2.50932246e-01 3.57091755e-01 8.81781280e-02 -6.31514490e-01
5.62809646e-01 7.40322292e-01 8.53243828e-01 -4.82930422e-01
8.92423093e-01 8.10711622e-01 -3.50512028e-01 -5.21203578e-01
-8.14570189e-01 -9.01552141e-01 -9.49433982e-01 -1.14694886e-01
1.17164898e+00 -1.31154406e+00 -5.61857879e-01 2.94730604e-01
-8.29341292e-01 5.89689016e-02 2.93295793e-02 4.54062074e-01
-3.53563458e-01 7.86002159e-01 -3.43381017e-01 -6.11406326e-01
-1.50842980e-01 -9.68673050e-01 7.92693853e-01 4.75604892e-01
-1.97295398e-01 -9.32996154e-01 3.53414305e-02 7.57241547e-01
-1.05647303e-01 -4.44494411e-02 7.78774440e-01 -6.58626437e-01
-5.40012658e-01 -3.12661231e-02 -5.54897904e-01 2.06323728e-01
3.62491250e-01 -4.98303622e-01 -9.74189401e-01 -5.66004813e-01
-9.39727873e-02 -9.38509256e-02 9.25597250e-01 2.53515333e-01
1.14368963e+00 -3.92095774e-01 -7.12756574e-01 7.36124039e-01
1.24819446e+00 8.69013444e-02 5.40056348e-01 5.34676075e-01
1.05159509e+00 8.96198630e-01 4.60164756e-01 3.39817405e-01
8.40785086e-01 7.43991017e-01 9.09718648e-02 -1.13928497e-01
-8.68286043e-02 -6.43369019e-01 2.17571333e-01 5.51345885e-01
-1.77335382e-01 -9.89043936e-02 -6.64309978e-01 5.43203115e-01
-2.15437913e+00 -1.07778192e+00 1.36028796e-01 2.24424171e+00
4.54434663e-01 -3.00446302e-01 2.08321363e-01 -1.71621919e-01
1.25524926e+00 1.10176906e-01 -6.63493574e-01 3.48987341e-01
-2.71368884e-02 -7.28690863e-01 3.41143638e-01 2.56295383e-01
-1.20941925e+00 9.78754699e-01 5.49332190e+00 5.98897934e-01
-6.16571426e-01 2.36325279e-01 7.71155238e-01 1.23766392e-01
-2.11580947e-01 5.63729219e-02 -9.60241199e-01 6.08221650e-01
5.16722679e-01 -2.04037488e-01 4.92416918e-01 8.99168193e-01
2.24546731e-01 8.51859301e-02 -1.33979988e+00 1.48653221e+00
4.31264937e-01 -9.18766916e-01 2.29400977e-01 2.79247731e-01
1.08314896e+00 -4.46756929e-01 2.55475640e-01 6.40092269e-02
4.41327989e-01 -1.06891465e+00 5.27990818e-01 7.22707868e-01
8.33720326e-01 -7.18530297e-01 5.04037738e-01 1.85635358e-01
-1.26198578e+00 -1.18670851e-01 -5.30261755e-01 3.44159365e-01
1.34813115e-01 2.03789443e-01 -5.99772573e-01 5.54421186e-01
9.85329568e-01 1.11338782e+00 -9.19104338e-01 7.70635962e-01
-8.60007107e-02 3.37278575e-01 5.71963936e-02 6.34397447e-01
-5.93586154e-02 -4.47361529e-01 2.31187254e-01 9.38098073e-01
1.52499884e-01 2.52886508e-02 4.43722427e-01 8.51814747e-01
-1.11624539e-01 -1.57045156e-01 -4.51483876e-01 6.39347136e-02
4.73683298e-01 1.04600000e+00 -6.34463966e-01 -4.27712470e-01
-5.48622191e-01 1.52523255e+00 5.57693243e-01 6.32961690e-01
-6.50760233e-01 2.10609630e-01 6.61862671e-01 1.92423433e-01
-9.25908908e-02 -2.38930315e-01 -3.22501123e-01 -1.42982566e+00
-6.98287087e-03 -6.80049837e-01 8.68352056e-01 -7.18193412e-01
-1.64147043e+00 3.55264693e-01 1.49731353e-01 -1.35253119e+00
-2.93246627e-01 -1.59799755e-01 -3.80324394e-01 6.77697480e-01
-1.37902117e+00 -1.38300347e+00 -6.54679358e-01 8.76048088e-01
5.49985111e-01 -4.89884257e-01 5.83474874e-01 3.37399602e-01
-8.11233878e-01 9.52178478e-01 3.84403735e-01 4.52884078e-01
1.05241108e+00 -1.21134710e+00 3.39281678e-01 9.80043650e-01
1.51224121e-01 8.26128781e-01 3.43947768e-01 -8.04184496e-01
-8.91904414e-01 -1.21970522e+00 7.23448813e-01 -3.26990396e-01
2.43197232e-01 -3.10195088e-01 -8.85695279e-01 7.73947597e-01
1.08650893e-01 -2.31168509e-01 7.44053006e-01 1.40919298e-01
-6.47740245e-01 -1.23913929e-01 -1.05003595e+00 6.18477643e-01
1.36515760e+00 -5.77588856e-01 -5.76494217e-01 2.19887882e-01
5.19775748e-01 4.81646731e-02 -7.88947105e-01 1.14056163e-01
4.54640180e-01 -8.18023026e-01 1.27385068e+00 -5.58336139e-01
2.24186242e-01 -4.74501431e-01 8.12206976e-03 -1.15941787e+00
-6.24894679e-01 -1.58371314e-01 1.16438963e-01 1.79471672e+00
-1.93680618e-02 -4.74419564e-01 8.97959411e-01 1.12642074e+00
3.09129208e-01 5.51688448e-02 -6.78664029e-01 -7.59610832e-01
-3.28969628e-01 1.02651983e-01 5.41497231e-01 1.29780209e+00
7.14558735e-02 2.70514071e-01 -7.58791745e-01 3.71489882e-01
1.19745314e+00 1.07675754e-01 8.98014307e-01 -1.34376013e+00
-1.02885745e-01 -3.06895435e-01 -4.27009583e-01 -1.20411992e+00
4.56420213e-01 -1.04392135e+00 -1.09204270e-01 -1.46256423e+00
9.89106059e-01 -6.18262947e-01 -3.22602391e-01 3.75687093e-01
-4.10962611e-01 1.25854194e-01 2.06326157e-01 9.67031896e-01
-1.01004088e+00 5.66873670e-01 9.76417124e-01 -3.03047627e-01
-7.99734369e-02 -2.38937467e-01 -7.69802392e-01 6.26888990e-01
6.16176605e-01 -4.30587023e-01 -6.75692976e-01 -4.70981508e-01
-2.54828572e-01 -1.62815869e-01 5.93818069e-01 -9.51620221e-01
5.43991327e-01 -2.12401435e-01 8.67446244e-01 -5.11410177e-01
1.71354398e-01 -9.04599905e-01 3.89992774e-01 5.02560176e-02
-6.35542572e-01 1.43798873e-01 -3.61391366e-01 1.01933444e+00
-2.70311028e-01 -1.73541620e-01 7.83588827e-01 -2.25732908e-01
-1.00925493e+00 5.45322657e-01 -6.80169091e-02 1.13825388e-01
1.02119637e+00 -4.05340314e-01 -2.62621552e-01 -3.62296104e-01
-8.70807528e-01 4.98109668e-01 8.42328489e-01 6.59926713e-01
6.40727878e-01 -1.58606160e+00 -5.55451453e-01 3.46847564e-01
3.92830431e-01 -6.33205473e-02 5.81472516e-01 4.22576100e-01
6.99615255e-02 2.72589117e-01 -2.60190219e-01 -7.28103697e-01
-1.46445811e+00 8.76241684e-01 2.47019291e-01 -2.12703079e-01
-8.51388097e-01 5.13495088e-01 7.59008884e-01 -3.96483094e-01
4.67273474e-01 4.26001072e-01 -5.10523200e-01 -4.86319214e-02
6.60927057e-01 3.90114933e-01 -4.28279132e-01 -9.95135248e-01
-3.34315717e-01 6.24660850e-01 -4.99242067e-01 -2.67697543e-01
1.02374625e+00 -7.99430430e-01 -4.36201654e-02 1.85770676e-01
1.29438746e+00 -4.35127765e-01 -1.38089490e+00 -5.73733032e-01
1.44243091e-01 -6.01312101e-01 -3.18895131e-01 -5.19713461e-01
-1.05828559e+00 6.11061871e-01 8.97815406e-01 -1.46582499e-01
7.97485650e-01 2.03455344e-01 5.69218755e-01 2.62103707e-01
2.16581881e-01 -1.32539403e+00 5.12196720e-01 7.98597559e-02
5.18354475e-01 -1.54036582e+00 3.16156782e-02 -2.69717306e-01
-1.07609999e+00 6.92983806e-01 9.11202133e-01 3.16763148e-02
5.17340422e-01 -4.08515930e-01 5.03777191e-02 -8.56062844e-02
-1.87231794e-01 -1.58705756e-01 5.28037965e-01 8.21228325e-01
1.34231001e-01 1.13216549e-01 3.74595858e-02 7.51568258e-01
4.12420556e-02 -3.05357724e-01 9.67180431e-02 5.02324343e-01
-1.89220518e-01 -1.02541351e+00 -3.94242346e-01 2.94896185e-01
-2.88923830e-01 2.23098010e-01 -2.96831429e-01 5.90061903e-01
9.45443138e-02 1.02287352e+00 1.43729016e-01 -4.38158453e-01
4.27985424e-03 6.05297554e-03 2.67428726e-01 -4.95501608e-01
1.56936906e-02 1.15378119e-01 -1.44120932e-01 -3.66986632e-01
-6.75207794e-01 -7.46468842e-01 -1.12176716e+00 -1.50991634e-01
-1.00030817e-01 3.00741136e-01 4.01843637e-01 1.05401897e+00
3.95997435e-01 1.03594832e-01 7.17334032e-01 -8.01512837e-01
-3.77896905e-01 -8.17777574e-01 -7.08113790e-01 1.09356618e+00
6.30557537e-02 -7.93673456e-01 -3.53994936e-01 5.83476067e-01] | [14.801953315734863, 1.0640575885772705] |
596f4ce8-bb6c-4b9f-a4da-582a3c6f4586 | contrastive-collaborative-filtering-for-cold | 2302.02151 | null | https://arxiv.org/abs/2302.02151v2 | https://arxiv.org/pdf/2302.02151v2.pdf | Contrastive Collaborative Filtering for Cold-Start Item Recommendation | The cold-start problem is a long-standing challenge in recommender systems. As a promising solution, content-based generative models usually project a cold-start item's content onto a warm-start item embedding to capture collaborative signals from item content so that collaborative filtering can be applied. However, since the training of the cold-start recommendation models is conducted on warm datasets, the existent methods face the issue that the collaborative embeddings of items will be blurred, which significantly degenerates the performance of cold-start item recommendation. To address this issue, we propose a novel model called Contrastive Collaborative Filtering for Cold-start item Recommendation (CCFCRec), which capitalizes on the co-occurrence collaborative signals in warm training data to alleviate the issue of blurry collaborative embeddings for cold-start item recommendation. In particular, we devise a contrastive collaborative filtering (CF) framework, consisting of a content CF module and a co-occurrence CF module to generate the content-based collaborative embedding and the co-occurrence collaborative embedding for a training item, respectively. During the joint training of the two CF modules, we apply a contrastive learning between the two collaborative embeddings, by which the knowledge about the co-occurrence signals can be indirectly transferred to the content CF module, so that the blurry collaborative embeddings can be rectified implicitly by the memorized co-occurrence collaborative signals during the applying phase. Together with the sound theoretical analysis, the extensive experiments conducted on real datasets demonstrate the superiority of the proposed model. The codes and datasets are available on https://github.com/zzhin/CCFCRec. | ['Ning Yang', 'Lilin Zhang', 'Zhihui Zhou'] | 2023-02-04 | null | null | null | null | ['collaborative-filtering'] | ['miscellaneous'] | [-2.54878044e-01 -3.68545353e-01 2.20231079e-02 -3.85015279e-01
-2.51222342e-01 -5.68900824e-01 6.20315492e-01 -4.03283566e-01
-1.60504371e-01 1.87706634e-01 6.32807851e-01 -6.36674389e-02
-3.42911452e-01 -8.70429933e-01 -4.30343807e-01 -9.38081563e-01
2.63471305e-01 -1.71959937e-01 -2.40477040e-01 -3.00111264e-01
4.53561358e-02 -2.02952370e-01 -1.29892135e+00 3.21637511e-01
1.11787271e+00 9.02925432e-01 4.34177965e-01 3.41123134e-01
-3.66930187e-01 2.51845330e-01 -5.13826609e-01 -6.00903630e-01
3.45241487e-01 -5.98248959e-01 -6.46047369e-02 -1.52496934e-01
5.36844321e-02 -2.13203430e-01 -4.46376652e-01 9.56428468e-01
5.06611466e-01 5.14685035e-01 4.58284050e-01 -1.03828371e+00
-1.39691865e+00 7.95309722e-01 -3.39598209e-01 1.85706895e-02
3.05180371e-01 3.43700685e-02 1.25993085e+00 -1.50192654e+00
3.28390568e-01 9.58430409e-01 6.83462620e-01 4.38560814e-01
-9.58349347e-01 -9.50607955e-01 7.20992863e-01 2.72024386e-02
-1.35692215e+00 -1.49464682e-01 9.37488973e-01 -3.84726673e-01
2.41834834e-01 1.52131230e-01 9.25193369e-01 1.13375461e+00
1.40460732e-03 8.32719862e-01 7.68604577e-01 -3.49805430e-02
1.67212725e-01 3.78013045e-01 1.69826075e-01 2.98500985e-01
1.00499608e-01 2.89704621e-01 -3.84742826e-01 -1.50165439e-01
6.96074724e-01 9.82395530e-01 -5.03123224e-01 -2.17627928e-01
-9.96269941e-01 1.01551163e+00 7.38726735e-01 5.32528341e-01
-2.60943890e-01 -1.54909343e-01 1.20151497e-01 4.40992594e-01
6.15075529e-01 4.86361802e-01 -2.46382490e-01 3.40959907e-01
-8.70456934e-01 1.80281326e-01 6.79401398e-01 9.28336024e-01
6.86979473e-01 -1.20821208e-01 -3.59852076e-01 1.01936281e+00
7.30069518e-01 4.68567640e-01 5.74913561e-01 -2.09343880e-01
4.13871258e-01 3.07515174e-01 2.43497163e-01 -1.10037172e+00
3.44378799e-02 -1.06757736e+00 -8.86261106e-01 -2.88263828e-01
1.73152626e-01 -4.78045195e-01 -5.37055135e-01 1.71877515e+00
3.75376165e-01 7.64803410e-01 4.00293283e-02 1.23528576e+00
8.45535100e-01 9.91104424e-01 -1.30321249e-01 -3.47596645e-01
1.03366411e+00 -1.14218009e+00 -8.23117316e-01 8.16152319e-02
4.70881611e-01 -9.96559262e-01 1.09875262e+00 3.15420032e-01
-6.88597381e-01 -9.68123555e-01 -1.02373838e+00 7.64360055e-02
-2.98798263e-01 3.22715104e-01 5.46108484e-01 4.91989404e-01
-4.52237636e-01 4.38413084e-01 -2.47785896e-01 8.92910957e-02
2.10072234e-01 8.82127360e-02 1.07508106e-02 -4.00324136e-01
-1.50655293e+00 5.09928107e-01 -1.77114806e-03 5.11333227e-01
-5.90659976e-01 -1.03759086e+00 -4.27228808e-01 1.70476645e-01
4.11062330e-01 -6.96625590e-01 9.23851728e-01 -1.04227483e+00
-1.64898705e+00 -9.45520550e-02 5.40521555e-02 8.52572620e-02
2.61958987e-01 -7.28909373e-01 -8.60258639e-01 -3.84889722e-01
-2.76411682e-01 -2.77124112e-03 1.22331560e+00 -1.32241201e+00
-4.64390993e-01 -8.02975371e-02 4.80359718e-02 3.39344710e-01
-6.35226548e-01 -3.08656275e-01 -5.62155724e-01 -1.19738996e+00
6.48934096e-02 -8.23340058e-01 -2.12180957e-01 -1.59521475e-01
-1.00129828e-01 -8.80388990e-02 7.92621851e-01 -3.45875323e-01
1.51543665e+00 -2.59346199e+00 1.69538468e-01 2.53361583e-01
1.49248391e-01 3.56886804e-01 -5.30151904e-01 6.51842833e-01
1.96661125e-03 -1.25694305e-01 1.72130927e-01 -4.09533054e-01
-2.77718715e-02 1.36941776e-01 -7.78261662e-01 3.20120066e-01
-1.30884023e-02 8.60322058e-01 -1.24951899e+00 7.45743588e-02
2.75646269e-01 7.07463443e-01 -9.58582819e-01 6.05984688e-01
-8.55650380e-02 4.33810145e-01 -5.55640340e-01 7.05643892e-02
9.49251831e-01 -8.96326229e-02 2.12226868e-01 -4.47439998e-01
-5.78503422e-02 2.04723477e-01 -1.23258674e+00 1.84671021e+00
-7.30132937e-01 -2.29656082e-02 -7.09365308e-02 -6.93144619e-01
1.09440112e+00 2.67413884e-01 3.19460571e-01 -5.19839406e-01
1.90303419e-02 1.69755407e-02 -1.11573888e-02 -4.19533968e-01
6.23515308e-01 -2.13625014e-01 3.83153521e-02 6.58339500e-01
2.63967812e-01 1.16113931e-01 -3.18581462e-01 3.18793833e-01
7.10220635e-01 2.85609156e-01 -2.78737247e-01 -3.18709351e-02
4.94258612e-01 -5.62506318e-01 6.77273452e-01 6.38800204e-01
3.49091530e-01 6.60355270e-01 -2.15678677e-01 -2.94526368e-01
-6.93938553e-01 -1.20984900e+00 8.65612105e-02 1.21282268e+00
4.57717329e-01 -8.53094101e-01 -2.61611700e-01 -9.02672589e-01
1.64525285e-02 5.20805955e-01 -6.32244289e-01 -5.00898302e-01
-3.64287883e-01 -6.98193848e-01 -4.42574136e-02 4.26165402e-01
1.90681085e-01 -5.99310875e-01 2.39267066e-01 4.02383059e-01
3.91161479e-02 -4.74606037e-01 -9.85491037e-01 1.23022888e-02
-7.85701334e-01 -7.22879291e-01 -1.12012851e+00 -6.04960144e-01
7.38365412e-01 8.05798709e-01 6.36599004e-01 3.03451300e-01
3.76254737e-01 1.36678249e-01 -8.79909813e-01 -1.67079661e-02
1.02517158e-01 -1.78829744e-01 9.51699689e-02 4.00197476e-01
3.88755202e-01 -6.91199541e-01 -8.97270322e-01 4.80808973e-01
-8.22721422e-01 4.13225219e-02 6.16634965e-01 1.23919082e+00
4.05246168e-01 1.83025561e-02 8.77302229e-01 -1.06117189e+00
8.87573719e-01 -8.11294734e-01 -2.50336528e-01 5.12158014e-02
-7.95827031e-01 -1.45797819e-01 1.24565327e+00 -8.83509219e-01
-1.21329331e+00 -3.51072073e-01 -3.72441202e-01 -6.17584705e-01
2.37926230e-01 6.82241738e-01 -2.53408223e-01 2.73240894e-01
3.68143082e-01 2.64213830e-01 -3.32801193e-01 -9.12281871e-01
9.92779255e-01 7.27522612e-01 2.71517813e-01 -2.29348660e-01
1.08397329e+00 1.52689472e-01 -7.30600417e-01 -3.79054725e-01
-1.11499095e+00 -6.24889612e-01 -3.18957627e-01 -1.48814723e-01
4.65900034e-01 -9.75831211e-01 -1.96862742e-01 2.21612900e-01
-7.93724358e-01 6.61497563e-02 -3.85549843e-01 9.27077591e-01
-1.21791651e-02 3.42115164e-01 -4.87780958e-01 -5.89205444e-01
-3.62205654e-01 -7.95529246e-01 4.59804833e-01 4.94302541e-01
-9.18991957e-03 -1.01198447e+00 3.00139427e-01 1.04750566e-01
5.00264227e-01 -5.31033576e-01 7.88606882e-01 -7.80599892e-01
-3.58223289e-01 -2.04040825e-01 -9.11246464e-02 5.54576278e-01
3.14581960e-01 -4.46183197e-02 -9.68485713e-01 -4.32940662e-01
1.67433485e-01 1.83175400e-01 8.45252693e-01 8.97765085e-02
9.23974395e-01 -3.99706721e-01 -1.92352474e-01 7.14082837e-01
1.41330898e+00 1.87068626e-01 5.41880131e-01 -2.59362310e-01
8.43047440e-01 1.95462495e-01 7.21385777e-01 5.47509253e-01
2.23270029e-01 6.04041398e-01 1.22262955e-01 -1.26571767e-02
-1.10662840e-01 -7.06361175e-01 3.16510320e-01 1.43725002e+00
-1.02753853e-02 -3.35771263e-01 -1.65685877e-01 2.65263706e-01
-1.84910023e+00 -9.55921710e-01 -4.23193164e-02 2.39749312e+00
7.37004340e-01 -8.37720558e-02 -8.91031101e-02 -1.75709818e-02
5.91130495e-01 2.66056269e-01 -2.92406917e-01 -2.07928509e-01
2.63203144e-01 2.25169435e-01 -5.48081063e-02 2.78178811e-01
-8.89537096e-01 6.10443413e-01 5.12149096e+00 7.51524210e-01
-1.19278622e+00 4.06304091e-01 -6.46927655e-02 -3.43216211e-01
-7.69164741e-01 1.40356615e-01 -6.59482718e-01 9.92261291e-01
5.53311586e-01 -1.33396834e-01 3.90147179e-01 8.02392006e-01
2.00448185e-01 2.03011066e-01 -9.42766488e-01 8.84379268e-01
1.08491018e-01 -1.27384520e+00 -1.74842626e-02 -9.18292180e-02
1.00610483e+00 -2.58903652e-01 4.19209719e-01 6.21503711e-01
3.28232855e-01 -6.18659854e-01 5.54089546e-01 9.61239278e-01
4.17256683e-01 -7.88500428e-01 8.54187548e-01 3.01928848e-01
-1.27538180e+00 -4.00167167e-01 -5.85541308e-01 -1.91847868e-02
1.36078358e-01 1.10232270e+00 -1.05761431e-01 9.91547167e-01
5.30627608e-01 1.13805997e+00 -2.57037908e-01 1.25959826e+00
-4.92186576e-01 7.48959541e-01 3.27406935e-02 -1.35543898e-01
7.24436194e-02 -6.85078681e-01 3.52442712e-01 1.16359425e+00
5.07008612e-01 2.06714302e-01 7.54713491e-02 9.71512675e-01
-2.25145563e-01 3.25934619e-01 -2.83172607e-01 -3.67248878e-02
6.74749315e-01 1.34161627e+00 -1.67218149e-01 -3.27505738e-01
-5.84927738e-01 1.07729161e+00 1.58945322e-01 6.29942536e-01
-8.75671864e-01 -5.13175368e-01 7.33430147e-01 -1.91317469e-01
6.33291006e-01 -1.05089523e-01 -1.04689039e-01 -1.53749418e+00
-7.84017146e-02 -7.93501437e-01 1.78655699e-01 -6.42520785e-01
-1.73518157e+00 3.98944229e-01 -3.73307019e-01 -1.69233108e+00
1.43927515e-01 4.02954668e-02 -1.04695928e+00 1.19126093e+00
-1.29398990e+00 -1.01713324e+00 -3.56984407e-01 6.38532817e-01
3.73128921e-01 -5.41954115e-02 7.41671503e-01 5.68635643e-01
-6.16025627e-01 8.58470917e-01 5.89382768e-01 -1.08656704e-01
9.31914151e-01 -9.14466083e-01 9.91998389e-02 8.23437750e-01
4.85914379e-01 1.15484059e+00 4.55721259e-01 -4.69486594e-01
-1.54248309e+00 -1.23635733e+00 5.25197387e-01 -2.88626522e-01
5.44432759e-01 -5.86773396e-01 -1.10111225e+00 4.10662711e-01
2.09264327e-02 1.52115494e-01 1.07943344e+00 5.36248028e-01
-6.28362775e-01 -3.58419389e-01 -7.34927714e-01 5.32933533e-01
9.74693716e-01 -6.85931623e-01 -9.47892964e-01 1.72552139e-01
8.30786526e-01 -3.92168434e-03 -9.48439896e-01 1.68504398e-02
8.01291525e-01 -7.97160864e-01 8.34853470e-01 -3.32290471e-01
3.66186857e-01 -6.67939782e-01 3.73683572e-02 -1.73168314e+00
-7.51283884e-01 -5.12993991e-01 -2.55399168e-01 1.34227240e+00
3.14424932e-01 -5.82109213e-01 3.35553616e-01 1.36330813e-01
-3.64295512e-01 -8.74877930e-01 -3.51510584e-01 -6.30304575e-01
-8.06735456e-03 -3.34209710e-01 7.09123850e-01 1.04332697e+00
2.69740876e-02 3.91643465e-01 -7.14507759e-01 1.36574671e-01
1.67559102e-01 5.93703210e-01 8.24026942e-01 -1.06989300e+00
-7.03364193e-01 -7.98843876e-02 1.08465843e-01 -1.51340485e+00
-3.04724813e-01 -9.68683481e-01 9.75072570e-03 -1.39712548e+00
5.44100255e-02 -8.60632122e-01 -7.37955511e-01 1.63516909e-01
-4.05344188e-01 1.40196502e-01 3.00107092e-01 4.99537766e-01
-3.37274253e-01 8.49088252e-01 1.47418988e+00 3.96177024e-02
-3.41419339e-01 1.33550555e-01 -9.01106954e-01 4.09416676e-01
4.49418128e-01 -4.57713217e-01 -7.73567438e-01 -3.98777217e-01
4.19753760e-01 -3.06328416e-01 -4.50626155e-03 -7.95549870e-01
3.05972189e-01 -1.28136560e-01 5.69914997e-01 -4.17824358e-01
2.62317300e-01 -8.29932868e-01 3.81411195e-01 2.50120819e-01
-4.46229696e-01 -2.90310204e-01 -2.92151004e-01 7.78402627e-01
-2.19739348e-01 -2.32229337e-01 4.83006507e-01 1.21985599e-01
-3.79446656e-01 4.58351135e-01 -7.98237175e-02 -2.36888766e-01
7.48223662e-01 -4.10293452e-02 -2.57574648e-01 -2.95633763e-01
-7.93744802e-01 1.84647009e-01 3.13313246e-01 7.76826918e-01
8.03841352e-01 -1.55964100e+00 -4.58043784e-01 5.80560148e-01
1.47110760e-01 -2.52662003e-01 6.94360375e-01 8.77786994e-01
3.40889037e-01 3.70081700e-02 1.28142461e-01 -8.49935487e-02
-6.95761204e-01 9.93714869e-01 4.21456546e-02 -1.05823606e-01
-6.76199138e-01 1.08162832e+00 4.39481705e-01 -3.02385509e-01
3.96393612e-02 -2.46194988e-01 -2.90856063e-01 2.26176918e-01
7.53688514e-01 8.28285795e-03 -1.55515954e-01 -3.05114299e-01
-1.51152641e-01 5.85836470e-01 -4.08126175e-01 8.72047693e-02
1.33887875e+00 -2.31412530e-01 2.05929130e-01 4.33313817e-01
1.24070060e+00 4.37822938e-01 -1.12614143e+00 -5.76398015e-01
-5.20214677e-01 -7.27763593e-01 1.99916542e-01 -7.83852339e-01
-1.25317192e+00 8.12888324e-01 5.55102646e-01 9.87788886e-02
1.12778139e+00 -2.17235729e-01 9.02583241e-01 1.76522136e-01
2.19188854e-01 -1.04578006e+00 1.83731824e-01 4.15500164e-01
8.18723619e-01 -8.15032005e-01 5.59348986e-03 -3.38256955e-01
-5.13404012e-01 9.25534546e-01 4.61756200e-01 -5.62373936e-01
1.04507792e+00 -2.14361791e-02 3.34770195e-02 1.88447405e-02
-7.87046313e-01 -4.25259285e-02 4.60151404e-01 3.93415034e-01
4.32743609e-01 9.92827564e-02 -5.29940784e-01 1.32585716e+00
2.94179544e-02 1.52952224e-01 2.49564841e-01 6.55213237e-01
-2.19255805e-01 -1.25308871e+00 6.47565676e-03 5.81656992e-01
-1.46420017e-01 -3.76663774e-01 -1.12067319e-01 1.87843308e-01
5.51444829e-01 1.07882535e+00 -8.77994448e-02 -1.01765084e+00
2.16625512e-01 -4.43092994e-02 1.09328933e-01 -9.03188050e-01
-8.46434414e-01 3.36420894e-01 -2.44222745e-01 -3.02538127e-01
-2.56626904e-01 -2.59641439e-01 -8.62833202e-01 -1.87531129e-01
-8.44878435e-01 5.62520921e-01 4.06628877e-01 9.11756277e-01
5.05439579e-01 4.58472610e-01 1.20980442e+00 -7.21177280e-01
-5.01472414e-01 -1.04848397e+00 -7.56128967e-01 5.65101027e-01
1.73024938e-01 -7.23928690e-01 -5.24546862e-01 -1.84878811e-01] | [10.171963691711426, 5.554909706115723] |
c3d4047e-7f97-4666-ba26-55d9636d4afd | sss3d-fast-neural-architecture-search-for | 2304.11207 | null | https://arxiv.org/abs/2304.11207v1 | https://arxiv.org/pdf/2304.11207v1.pdf | SSS3D: Fast Neural Architecture Search For Efficient Three-Dimensional Semantic Segmentation | We present SSS3D, a fast multi-objective NAS framework designed to find computationally efficient 3D semantic scene segmentation networks. It uses RandLA-Net, an off-the-shelf point-based network, as a super-network to enable weight sharing and reduce search time by 99.67% for single-stage searches. SSS3D has a complex search space composed of sampling and architectural parameters that can form 2.88 * 10^17 possible networks. To further reduce search time, SSS3D splits the complete search space and introduces a two-stage search that finds optimal subnetworks in 54% of the time required by single-stage searches. | ['Brett H. Meyer', 'Warren J. Gross', 'Zhuoran Xiong', 'Marihan Amein', 'Olivier Therrien'] | 2023-04-21 | null | null | null | null | ['scene-segmentation', 'architecture-search'] | ['computer-vision', 'methodology'] | [ 1.77114129e-01 2.23798916e-01 -5.57608366e-01 -1.69900030e-01
-4.21119690e-01 -5.93685865e-01 -3.64153475e-01 -3.12245071e-01
-5.46894968e-01 7.25175798e-01 -4.01043415e-01 -6.20493770e-01
-6.24869168e-01 -9.86002624e-01 -3.43981892e-01 -3.68952751e-01
-1.95810422e-01 8.39049220e-01 1.07181466e+00 -1.05569370e-01
1.10795371e-01 9.50405180e-01 -1.30048645e+00 -1.61259532e-01
8.27848971e-01 1.21678007e+00 2.46190056e-01 6.67533636e-01
-2.62318820e-01 -3.56528312e-02 -7.27027476e-01 -4.03255299e-02
6.19378567e-01 1.44650474e-01 -1.15967286e+00 -1.66135311e-01
2.12370515e-01 -5.34377247e-02 -5.12774289e-01 6.95070267e-01
7.79757917e-01 5.75110734e-01 2.33696535e-01 -1.56787932e+00
1.94496006e-01 5.32863259e-01 -7.57052243e-01 4.30033982e-01
1.03809014e-01 5.56970596e-01 9.91837561e-01 -4.51071709e-01
8.31349015e-01 1.51605916e+00 7.93064952e-01 5.31863153e-01
-1.20292807e+00 -8.45797122e-01 1.44366354e-01 -1.74908433e-02
-1.49100220e+00 -1.94160953e-01 6.76417410e-01 3.21125597e-01
1.54556203e+00 6.17725968e-01 1.03477716e+00 5.75530648e-01
-3.59697379e-02 7.96770632e-01 6.27121866e-01 8.82131979e-02
4.39084381e-01 -6.46878123e-01 -8.82598571e-03 8.33092928e-01
4.21784878e-01 -1.15049236e-01 -6.52548611e-01 -2.81495243e-01
1.06521821e+00 -5.75310051e-01 -4.48792204e-02 -2.62204021e-01
-8.95899117e-01 7.49122739e-01 7.08669066e-01 2.23466698e-02
-2.07225889e-01 5.13596654e-01 2.69343346e-01 -8.44417214e-02
5.95131874e-01 8.13999355e-01 -8.23330700e-01 -2.66623110e-01
-1.14106822e+00 4.67858970e-01 8.27886164e-01 9.37960982e-01
5.47280908e-01 -4.84870151e-02 -1.52582740e-02 8.44809949e-01
5.29308438e-01 2.53602505e-01 -3.66426408e-01 -1.72071242e+00
6.39996290e-01 8.28337967e-01 -3.05636138e-01 -8.63310337e-01
-8.85917902e-01 -5.94646811e-01 -6.77457809e-01 3.81177217e-01
3.76192242e-01 -3.92743349e-01 -1.03649521e+00 1.38677776e+00
7.63450742e-01 3.41589719e-01 -2.23258421e-01 8.64183605e-01
6.83509886e-01 6.61214054e-01 -2.16922268e-01 1.00594327e-01
9.48876739e-01 -1.38076782e+00 -7.76561815e-03 -2.92376488e-01
2.36760214e-01 -7.41653979e-01 7.39683270e-01 2.88952440e-01
-1.69409585e+00 -4.36416082e-02 -1.03316116e+00 3.77869196e-02
-3.89270335e-01 -6.55618370e-01 9.06611860e-01 7.25841582e-01
-1.35011673e+00 7.18516052e-01 -8.82899880e-01 -1.30021229e-01
8.91135514e-01 7.99109042e-01 3.94177973e-01 -2.54052967e-01
-8.22789669e-01 6.84534907e-01 5.06674290e-01 -3.91631313e-02
-6.60161793e-01 -1.07538164e+00 -7.06705153e-01 2.73743391e-01
9.46825206e-01 -9.75173950e-01 1.05698228e+00 -1.00799307e-01
-1.48681355e+00 5.23251891e-01 -2.42187023e-01 -2.27531135e-01
4.81384009e-01 4.36317444e-01 -1.91746369e-01 5.31849504e-01
1.67555228e-01 1.11931562e+00 1.78435415e-01 -9.47497845e-01
-5.94290555e-01 -7.76907206e-02 2.41920307e-01 4.38141882e-01
4.70651239e-02 1.58397723e-02 -1.14889348e+00 -5.85650682e-01
4.85450357e-01 -1.01445830e+00 -1.01539850e+00 6.58227146e-01
-7.21793711e-01 -3.44550431e-01 7.75090754e-01 -8.92738774e-02
1.00671923e+00 -1.79089510e+00 8.07881951e-02 9.67402577e-01
5.79921544e-01 2.53917962e-01 -3.65490943e-01 -9.72047225e-02
3.25416595e-01 3.64930332e-01 -2.30144665e-01 -4.18037325e-01
6.50476217e-02 4.52371687e-01 6.89398050e-01 2.93683261e-01
6.68296069e-02 8.69813979e-01 -7.86768496e-01 -6.98640227e-01
4.15447026e-01 3.01124901e-01 -8.36866558e-01 -3.92339289e-01
-2.83732772e-01 -1.27379000e-01 -6.05039716e-01 1.14013648e+00
9.67903674e-01 -6.63352847e-01 7.50110000e-02 -1.70869127e-01
-5.04520424e-02 2.47941613e-01 -1.55131006e+00 2.14457011e+00
-1.59437567e-01 4.97270048e-01 5.13856173e-01 -7.48575211e-01
5.99804223e-01 -2.31715932e-01 9.13435459e-01 -5.38494527e-01
3.31146181e-01 3.47267359e-01 -1.49739727e-01 3.19257602e-02
5.02397537e-01 4.00662422e-01 -7.59596154e-02 3.59376192e-01
6.89328089e-02 -5.81741810e-01 4.84412432e-01 6.39111325e-02
1.47785687e+00 -2.91980028e-01 -3.13110739e-01 -5.76272309e-01
1.72705829e-01 4.29625928e-01 8.81346107e-01 7.97257721e-01
-1.75315723e-01 3.07476670e-01 3.08641165e-01 -3.97047400e-01
-7.86986947e-01 -1.18221521e+00 6.67962730e-02 6.70444965e-01
8.14627588e-01 -1.76858604e-01 -5.68188190e-01 -5.54282725e-01
2.93524325e-01 6.22379363e-01 6.47543669e-02 1.97193369e-01
-4.71719712e-01 -7.49359965e-01 7.42498994e-01 3.01866770e-01
7.03450620e-01 -7.95342386e-01 -7.05022037e-01 3.93439233e-01
5.67458719e-02 -9.93621945e-01 -5.10083258e-01 4.83243197e-01
-1.10295296e+00 -1.45053172e+00 -4.93431747e-01 -7.99903810e-01
8.16521466e-01 3.28672260e-01 1.28968883e+00 3.77154410e-01
-7.17958212e-01 -6.87371269e-02 -5.22973910e-02 -6.93523586e-02
1.17211215e-01 5.36927402e-01 -1.47897199e-01 -9.00040448e-01
-1.34542271e-01 -7.17212617e-01 -7.08252430e-01 6.62712455e-01
-4.83334094e-01 2.71551199e-02 3.04238528e-01 3.52139145e-01
8.19105983e-01 5.68743169e-01 1.07506730e-01 -2.22539023e-01
5.30067086e-01 -3.75092804e-01 -1.05342519e+00 1.88924119e-01
-6.17952406e-01 -1.34385765e-01 4.35181186e-02 -3.12140584e-01
-7.12447584e-01 8.83130878e-02 -1.96212634e-01 -6.18907630e-01
1.29514202e-01 2.47747585e-01 2.04331189e-01 -7.19707131e-01
2.51708955e-01 -3.26563060e-01 3.45314704e-02 -5.77523232e-01
5.34118973e-02 1.61753714e-01 2.49419168e-01 -6.95845366e-01
8.13989937e-01 3.17295790e-01 5.18440902e-01 -6.46688461e-01
-6.62504137e-01 -6.17687047e-01 -3.01880270e-01 -5.80124199e-01
7.84213066e-01 -5.71770668e-01 -1.37032974e+00 5.12875319e-01
-1.19805467e+00 -7.36407578e-01 -4.95907336e-01 1.62405699e-01
-2.07671866e-01 1.48728015e-02 -6.62276447e-01 -6.67477727e-01
-4.21998233e-01 -1.27137518e+00 7.81806290e-01 3.97152930e-01
-3.52090985e-01 -7.34941661e-01 -3.86457145e-01 5.47167301e-01
5.15226364e-01 3.41362059e-01 6.14516556e-01 -2.07119271e-01
-9.25240338e-01 3.70334089e-01 -6.47944212e-01 -4.99321818e-02
-2.59279042e-01 2.18962759e-01 -2.17240959e-01 -2.85301805e-01
-6.42689884e-01 -3.71228382e-02 6.02508247e-01 9.09608245e-01
1.58969426e+00 -9.20993369e-03 -5.95177829e-01 1.04961252e+00
1.57578206e+00 6.16991222e-01 3.21514249e-01 5.24060249e-01
4.31690156e-01 1.43095285e-01 4.34916735e-01 3.90147656e-01
4.05990392e-01 4.15705740e-01 9.26515400e-01 -5.89333892e-01
-1.93419740e-01 -4.06104289e-02 -4.28707957e-01 6.88854516e-01
8.65971670e-02 -7.70409703e-01 -1.12622905e+00 4.51671958e-01
-1.75748694e+00 -5.75543106e-01 -5.83503842e-02 1.38781512e+00
5.95690310e-01 7.21360147e-01 1.70315564e-01 1.08167030e-01
9.07571971e-01 3.10418010e-01 -1.08217072e+00 -3.13346863e-01
-1.58676118e-01 5.39967597e-01 1.11123705e+00 4.58887666e-01
-6.97894871e-01 9.27071631e-01 7.69476509e+00 1.42484033e+00
-7.11698353e-01 -1.61911249e-01 7.74461389e-01 -8.03643465e-01
-2.50687629e-01 -2.74506718e-01 -6.88766718e-01 3.72454613e-01
5.37374139e-01 -1.48591131e-01 7.31542110e-01 8.21229339e-01
3.73821944e-01 -4.76869613e-01 -4.34637040e-01 9.79185462e-01
-2.91094810e-01 -2.11228132e+00 -1.59734040e-01 2.65719861e-01
8.60188603e-01 5.00331342e-01 -2.58481801e-01 -3.27820092e-01
6.09628379e-01 -9.21962202e-01 6.93910480e-01 -8.93835127e-02
8.95289004e-01 -1.08317089e+00 5.52670479e-01 1.94968045e-01
-1.65167010e+00 -1.30756004e-02 -4.63083312e-02 -6.28550127e-02
8.37841213e-01 6.91537976e-01 -6.19591475e-01 3.93085927e-01
1.07699919e+00 3.28020036e-01 3.44302654e-02 1.59027541e+00
8.75836685e-02 3.23027611e-01 -1.21646500e+00 -3.05928528e-01
6.88243508e-01 -2.32904240e-01 9.60076809e-01 7.45058358e-01
2.28221029e-01 3.59730959e-01 3.54125887e-01 1.00731671e+00
-3.02571207e-01 -4.77045596e-01 -2.42223665e-02 3.32849354e-01
9.53731179e-01 8.61610234e-01 -1.46731186e+00 -1.08016886e-01
4.16924246e-02 6.33394957e-01 -1.58751309e-01 2.65596658e-01
-9.36288357e-01 -7.26739407e-01 1.05472553e+00 -1.06254347e-01
3.72527659e-01 -4.50652659e-01 -6.67815387e-01 -4.89303261e-01
-3.09481859e-01 -4.04217750e-01 3.31142306e-01 -8.01556349e-01
-1.00895619e+00 4.55526471e-01 1.69008225e-01 -6.43946230e-01
4.53233778e-01 -5.32360017e-01 -5.91163993e-01 7.75549829e-01
-1.45839882e+00 -6.13748729e-01 -2.85129756e-01 6.39645457e-01
7.94309616e-01 9.60443467e-02 3.18124890e-01 5.71534097e-01
-7.85028696e-01 4.98655349e-01 -1.70624137e-01 -3.68332475e-01
-1.89166307e-01 -8.76343429e-01 7.98473537e-01 6.98578238e-01
-5.22710979e-01 1.46700203e-01 5.11306345e-01 -6.97826028e-01
-1.10189509e+00 -7.51568139e-01 6.47991359e-01 3.50621551e-01
4.97693598e-01 -4.88187745e-02 -2.72783250e-01 6.64179586e-03
-1.48658127e-01 3.19387205e-02 4.77553487e-01 -5.08016311e-02
2.63546646e-01 -2.26691775e-02 -1.79592919e+00 8.31945658e-01
1.79259944e+00 3.16158496e-02 -4.72260043e-02 2.35434249e-01
1.28276873e+00 -9.31614518e-01 -9.25702035e-01 4.01626915e-01
3.44399929e-01 -6.64803207e-01 1.43740964e+00 5.20035326e-02
-4.23759855e-02 -3.24937999e-01 3.55128422e-02 -9.75647628e-01
-2.30413854e-01 -9.09932315e-01 -4.81161922e-02 5.53873122e-01
7.52492189e-01 -8.55532527e-01 1.59379137e+00 7.75083363e-01
-4.28330630e-01 -1.31093252e+00 -1.58809412e+00 -1.02629232e+00
-5.20748496e-01 -6.67819023e-01 1.11042523e+00 6.78579152e-01
-4.84078348e-01 -6.67004287e-02 3.51177663e-01 1.04525372e-01
9.86056328e-01 -1.00182831e-01 3.39063555e-01 -1.31502521e+00
-9.77656916e-02 -7.91195750e-01 -1.08578421e-01 -1.55442524e+00
-1.58096299e-01 -6.41841173e-01 -1.07883386e-01 -1.73992693e+00
-2.89176196e-01 -1.08922529e+00 -4.55230474e-02 6.71961248e-01
4.21735585e-01 3.16700578e-01 4.04831260e-01 -2.04754427e-01
-8.53504479e-01 4.21607792e-01 1.65240538e+00 -1.54278427e-01
-4.26936984e-01 1.32369429e-01 -5.35215080e-01 8.67701173e-01
9.88622904e-01 -5.34935892e-01 -5.73739886e-01 -5.39168477e-01
1.47821099e-01 1.39832184e-01 8.99037495e-02 -1.02793646e+00
3.55630964e-01 -6.39514446e-01 1.53504327e-01 -1.24106002e+00
8.03067029e-01 -1.01463223e+00 2.91366756e-01 6.50912106e-01
9.51862708e-02 1.79939464e-01 5.45179605e-01 2.44705588e-01
1.09441079e-01 -1.41390532e-01 6.35489523e-01 -2.32277855e-01
-7.67102659e-01 5.34817815e-01 -3.87597859e-01 1.70181215e-01
1.21416199e+00 -9.39704776e-01 -2.70283848e-01 1.63697660e-01
-7.25092292e-01 8.58094037e-01 3.37763369e-01 8.83523524e-02
6.30939007e-01 -8.95397365e-01 -5.53307571e-02 -2.16905296e-01
-6.66599870e-01 9.14457023e-01 2.91311085e-01 3.03289980e-01
-1.18033659e+00 2.23448202e-01 -1.77681059e-01 -5.76601923e-01
-1.44925213e+00 -1.98527500e-01 3.37046146e-01 -4.47199821e-01
-4.47794914e-01 1.48888874e+00 -8.30645144e-01 -6.81590855e-01
1.77644923e-01 -2.98073947e-01 3.63768429e-01 -1.33507609e-01
-2.42949307e-01 1.29744768e+00 -1.30569324e-01 -1.89560950e-01
-6.78252041e-01 6.38370574e-01 5.68621039e-01 -1.52385429e-01
1.79716253e+00 -1.67631373e-01 -2.29521170e-01 -3.92033517e-01
1.24096704e+00 -6.67492509e-01 -1.22797716e+00 1.36838844e-02
-2.78985292e-01 -4.70728219e-01 6.70363843e-01 -9.82732117e-01
-1.69902945e+00 -1.63593460e-02 1.73530728e-01 5.75305000e-02
1.23526835e+00 1.19472750e-01 1.28561616e+00 3.21921796e-01
7.02524900e-01 -1.44471192e+00 4.53525148e-02 6.14681065e-01
3.44723225e-01 -6.42374098e-01 2.67840207e-01 -1.17828083e+00
8.99234340e-02 1.09682536e+00 9.24168825e-01 -1.86251163e-01
8.59686673e-01 4.46680099e-01 -5.40262043e-01 -6.93997264e-01
-5.83750546e-01 -2.86734588e-02 1.51215672e-01 5.07080555e-01
-5.35299718e-01 -7.45086148e-02 -3.80772948e-02 2.05045864e-01
-2.00985536e-01 -1.09156340e-01 2.77114101e-02 1.02978754e+00
-3.81688714e-01 -9.22643363e-01 -1.69138178e-01 7.32159019e-01
1.51653979e-02 5.66619039e-02 -5.78820258e-02 7.67782450e-01
2.36291006e-01 1.07078111e+00 3.69123340e-01 -3.10891062e-01
5.12341380e-01 -5.30191422e-01 2.52115816e-01 -3.68572980e-01
-7.72873342e-01 -1.08425088e-01 5.18713653e-01 -1.17220557e+00
-4.06955779e-01 -3.96242201e-01 -1.50151372e+00 -9.87456381e-01
-5.53583145e-01 -6.82347938e-02 9.06378329e-01 7.89467156e-01
5.76779783e-01 8.63617241e-01 4.62366611e-01 -1.26456225e+00
2.15842891e-02 -5.78290261e-02 -4.44417417e-01 -6.51293874e-01
-3.24738808e-02 -8.09246421e-01 -4.47199970e-01 -1.00538242e+00] | [9.259110450744629, -0.6325892210006714] |
7c0c38bd-a60a-4f84-a4d6-a0696ca25045 | learning-to-rank-visual-stories-from-human-1 | null | null | https://aclanthology.org/2022.acl-long.441 | https://aclanthology.org/2022.acl-long.441.pdf | Learning to Rank Visual Stories From Human Ranking Data | Visual storytelling (VIST) is a typical vision and language task that has seen extensive development in the natural language generation research domain. However, it remains unclear whether conventional automatic evaluation metrics for text generation are applicable on VIST. In this paper, we present the VHED (VIST Human Evaluation Data) dataset, which first re-purposes human evaluation results for automatic evaluation; hence we develop Vrank (VIST Ranker), a novel reference-free VIST metric for story evaluation. We first show that the results from commonly adopted automatic metrics for text generation have little correlation with those obtained from human evaluation, which motivates us to directly utilize human evaluation results to learn the automatic evaluation model. In the experiments, we evaluate the generated texts to predict story ranks using our model as well as other reference-based and reference-free metrics. Results show that Vrank prediction is significantly more aligned to human evaluation than other metrics with almost 30% higher accuracy when ranking story pairs. Moreover, we demonstrate that only Vrank shows human-like behavior in its strong ability to find better stories when the quality gap between two stories is high. Finally, we show the superiority of Vrank by its generalizability to pure textual stories, and conclude that this reuse of human evaluation results puts Vrank in a strong position for continued future advances. | ['Lun-Wei Ku', 'Ting-Hao Huang', 'Chacha Chen', 'Kuan-Chieh Lo', 'Vincent Chen', 'Yun-Wei Chu', 'Chi-Yang Hsu'] | null | null | null | null | acl-2022-5 | ['visual-storytelling'] | ['natural-language-processing'] | [ 1.71539381e-01 2.33907416e-01 -1.11147054e-01 -1.36059925e-01
-1.07919812e+00 -8.26236725e-01 1.28462613e+00 3.24575245e-01
-1.13171183e-01 9.50425863e-01 8.74375105e-01 -2.09112629e-01
-1.44732475e-01 -7.68510699e-01 -2.66417414e-01 -1.68375283e-01
1.90441191e-01 7.31004238e-01 1.13746010e-01 -7.15434253e-01
5.18691480e-01 -1.40279785e-01 -1.56129313e+00 6.75197065e-01
8.63917589e-01 5.35197198e-01 1.15456745e-01 9.75723386e-01
7.98443891e-03 1.41526544e+00 -1.10862720e+00 -6.88000262e-01
-4.91449088e-02 -8.80892038e-01 -1.05791152e+00 -4.39626016e-02
3.47004414e-01 -2.69954383e-01 -8.95129517e-02 3.79913419e-01
7.17742801e-01 1.08407862e-01 1.16412175e+00 -1.37986386e+00
-9.43231761e-01 1.19161761e+00 -3.16821337e-01 1.17830537e-01
1.12925100e+00 2.43334785e-01 1.40090919e+00 -1.01631057e+00
1.17582536e+00 1.36094379e+00 5.23553371e-01 4.30292547e-01
-9.32843745e-01 -1.31324545e-01 -1.38372019e-01 3.28979105e-01
-1.23315609e+00 -1.71874344e-01 3.99823517e-01 -8.04691315e-01
1.09752679e+00 5.61114669e-01 7.01179087e-01 1.45044756e+00
4.64709513e-02 1.17669404e+00 1.07633293e+00 -5.70187032e-01
-9.35950726e-02 1.93102732e-01 1.19814426e-01 4.14778113e-01
6.83422908e-02 1.45631999e-01 -7.75328636e-01 8.64506885e-02
5.23372054e-01 -8.79548132e-01 -4.13723797e-01 -6.89671263e-02
-1.82152712e+00 1.01993299e+00 1.59296542e-01 5.72650492e-01
-9.09962654e-02 1.28136367e-01 6.05563223e-01 3.36275727e-01
5.15537679e-01 8.94114017e-01 1.13852330e-01 -5.35631537e-01
-9.40917194e-01 8.26067865e-01 6.56628609e-01 9.63180006e-01
1.13920480e-01 1.31702244e-01 -1.30499136e+00 9.81770635e-01
2.97141424e-03 3.46600890e-01 7.17584908e-01 -8.46016884e-01
5.34464419e-01 6.06939435e-01 1.60404176e-01 -1.10182452e+00
-1.88486010e-01 -2.66670823e-01 -5.89018404e-01 2.30776176e-01
3.70318532e-01 -1.50107630e-02 -4.83693093e-01 1.63463891e+00
-1.11573555e-01 -4.98436719e-01 2.00375482e-01 8.43241572e-01
1.28828418e+00 7.48084486e-01 -2.37537861e-01 -2.52982348e-01
1.18258357e+00 -1.03019571e+00 -7.42067575e-01 1.41800242e-03
8.31132889e-01 -1.01714957e+00 1.55879283e+00 4.64031488e-01
-1.08971441e+00 -6.01536393e-01 -1.35218680e+00 -7.64386207e-02
-4.44151014e-01 2.62858421e-01 5.08155465e-01 4.65852052e-01
-1.21118772e+00 6.56684160e-01 -1.60036221e-01 -5.19778967e-01
1.06932789e-01 -2.81864792e-01 -1.17280245e-01 6.94252178e-02
-1.33716369e+00 1.24474037e+00 4.85499054e-01 -4.38398302e-01
-9.34345186e-01 -5.24286330e-01 -8.55229795e-01 -3.84650916e-01
2.57217020e-01 -8.92325044e-01 1.74957550e+00 -5.56816399e-01
-1.31502438e+00 8.07541072e-01 -5.04448963e-03 -4.38484728e-01
9.33247566e-01 -8.46681297e-02 -3.46454352e-01 6.04134724e-02
5.00903785e-01 7.65634954e-01 3.94183308e-01 -1.46713090e+00
-6.01606250e-01 2.96308190e-01 1.22666426e-01 4.47105974e-01
-2.51656711e-01 1.02220096e-01 -1.72982570e-02 -9.30816948e-01
-4.18211162e-01 -5.21934807e-01 3.26999314e-02 -5.11877835e-01
-6.45341039e-01 -5.79415321e-01 4.33793813e-01 -6.40902758e-01
1.60115623e+00 -1.50706339e+00 1.88383564e-01 -9.31768045e-02
4.85331446e-01 -8.08696002e-02 -2.85221130e-01 7.96590149e-01
4.78128083e-02 4.51892018e-01 -5.85254207e-02 -2.76363373e-01
3.48301321e-01 -2.83203363e-01 -2.77811170e-01 -2.03181818e-01
1.49161264e-01 1.17730081e+00 -1.23161328e+00 -9.04004753e-01
1.07602566e-01 2.69268572e-01 -1.51175454e-01 9.76278111e-02
-3.34368229e-01 1.99841678e-01 -2.02599511e-01 3.47650737e-01
-2.30399165e-02 -2.62600720e-01 -1.87304512e-01 1.44727096e-01
-2.28821352e-01 4.12494928e-01 -7.88295031e-01 1.45453286e+00
-1.32493421e-01 1.33635998e+00 -9.60590839e-01 -2.55233645e-01
9.50214982e-01 5.31253040e-01 4.96870503e-02 -7.44250834e-01
2.51099586e-01 1.18972607e-01 -1.16057701e-01 -3.91377270e-01
1.08641851e+00 -5.02465889e-02 -3.49386781e-01 8.48265648e-01
7.31959864e-02 -6.71892762e-01 9.97006774e-01 5.89331150e-01
1.14069450e+00 1.56022161e-01 4.30148870e-01 -3.15608364e-03
2.73790210e-01 4.19044912e-01 -1.54471532e-01 9.00191247e-01
2.08450750e-01 1.05188644e+00 6.38741612e-01 -6.50148466e-02
-1.14784527e+00 -1.08872688e+00 9.75374728e-02 1.00395131e+00
5.46159735e-03 -1.01853836e+00 -9.80586886e-01 -6.16994560e-01
-2.84117699e-01 1.43340194e+00 -8.02899480e-01 2.68362276e-02
-4.45464581e-01 -6.44890308e-01 6.87356770e-01 3.96810144e-01
1.94680721e-01 -1.34846687e+00 -6.72771633e-01 1.06435351e-01
-6.04269326e-01 -9.11192179e-01 -4.34635818e-01 -3.47050548e-01
-3.90102148e-01 -9.08273518e-01 -1.02143443e+00 -6.29820764e-01
1.79746553e-01 3.22168708e-01 1.60004318e+00 3.61003280e-02
9.27502587e-02 5.09930551e-01 -8.48825216e-01 -6.03258491e-01
-1.00217712e+00 1.71538398e-01 -2.94744760e-01 -4.17350411e-01
1.68500226e-02 -2.92438596e-01 -2.57029772e-01 3.19155037e-01
-8.47503424e-01 6.65073395e-01 2.65878022e-01 6.67248189e-01
2.12609053e-01 -9.35999677e-02 7.07655370e-01 -6.38755202e-01
1.46876764e+00 -2.90883392e-01 -9.00915787e-02 3.38074178e-01
-7.58221030e-01 3.74216288e-02 2.65215099e-01 -3.79395902e-01
-9.83338535e-01 -6.95546687e-01 1.66685507e-01 8.38654265e-02
1.89271495e-01 7.85530388e-01 1.82527900e-01 5.01706362e-01
1.10952818e+00 5.12718819e-02 -3.55508775e-01 -1.45455645e-02
4.82332855e-01 4.61487383e-01 6.43008530e-01 -4.37067091e-01
8.77309859e-01 -1.15503781e-01 -4.30854738e-01 -5.64323843e-01
-7.56571472e-01 -2.59760004e-02 -2.69334912e-01 -8.29388499e-01
8.84344637e-01 -7.27059722e-01 -2.54069477e-01 1.61988273e-01
-1.51485705e+00 -5.64507604e-01 -5.45741796e-01 2.93635398e-01
-8.29524636e-01 2.54626781e-01 -5.63769162e-01 -8.36993396e-01
-4.51966822e-01 -9.16617632e-01 1.21420038e+00 7.48381438e-03
-1.08100295e+00 -9.53800559e-01 4.69535261e-01 5.49798667e-01
1.72093883e-01 6.43921316e-01 9.57651556e-01 -4.30900902e-01
-3.84478480e-01 -2.04865977e-01 -1.86237782e-01 1.10357352e-01
-1.89155176e-01 2.87315100e-01 -7.93786526e-01 7.30828196e-02
-3.46717626e-01 -5.51774561e-01 7.45205998e-01 2.74600208e-01
5.43694496e-01 -1.87622696e-01 -4.73390073e-02 -7.55855143e-02
1.12498152e+00 1.33263856e-01 7.65136898e-01 4.98456210e-01
7.80149877e-01 6.87560618e-01 1.00603926e+00 6.11313581e-01
4.77717817e-01 8.32112014e-01 6.07667789e-02 -2.78003607e-02
-5.52072346e-01 -7.89593637e-01 6.50204241e-01 8.63987446e-01
-4.42417681e-01 -9.71409917e-01 -1.00403929e+00 5.38122356e-01
-1.96658683e+00 -1.23463607e+00 -5.99648237e-01 2.08232737e+00
8.39857340e-01 2.50793546e-01 2.54869372e-01 4.12711084e-01
6.41405106e-01 1.89405769e-01 7.07526878e-02 -4.98970658e-01
-6.82669401e-01 5.15178628e-02 -1.34393230e-01 3.88221264e-01
-6.58748150e-01 9.03726876e-01 7.17103386e+00 1.17855561e+00
-5.00925660e-01 9.25777033e-02 6.54362857e-01 -1.39153227e-01
-6.77732348e-01 -1.59865677e-01 -3.66729230e-01 1.30679637e-01
7.23605990e-01 -9.24121559e-01 2.28175893e-01 7.09944785e-01
4.03992057e-01 -9.61937085e-02 -1.40363312e+00 1.05449212e+00
4.62482542e-01 -1.32140684e+00 3.98361206e-01 -4.94955145e-02
1.09564245e+00 -4.43414986e-01 -1.23542808e-02 5.12611210e-01
5.95459461e-01 -1.25964463e+00 1.44379640e+00 3.91488433e-01
7.22645938e-01 -7.58037567e-01 8.40577543e-01 2.70781726e-01
-1.03674722e+00 2.58301824e-01 -1.04209065e-01 -2.86169112e-01
5.72812617e-01 4.05848593e-01 -1.47015119e+00 5.72449505e-01
2.98217893e-01 6.71093345e-01 -9.84720647e-01 9.09335792e-01
-5.92026532e-01 3.95671070e-01 4.11669910e-01 -4.33244437e-01
1.70825899e-01 1.43811882e-01 8.87117624e-01 1.40510821e+00
7.71913767e-01 -1.79856390e-01 -6.70740008e-02 9.38094616e-01
3.85203143e-03 5.02560437e-01 -9.17013943e-01 -3.09348464e-01
1.81638613e-01 9.84845221e-01 -7.85242260e-01 -5.65895975e-01
5.00613004e-02 9.69093859e-01 1.97286397e-01 1.40836388e-01
-9.68747199e-01 -2.09344625e-01 1.45730674e-01 1.71869338e-01
-1.80866178e-02 7.77262775e-03 -6.21360481e-01 -8.76413763e-01
3.56782638e-02 -1.04802835e+00 2.07671985e-01 -1.54423308e+00
-1.24970269e+00 1.07944560e+00 3.87856752e-01 -1.60203493e+00
-8.85524094e-01 -4.16848630e-01 -7.47428536e-01 5.26645660e-01
-8.09822679e-01 -1.03147340e+00 -4.88349169e-01 2.45877981e-01
8.68622601e-01 -3.49268645e-01 6.87833667e-01 -1.75911516e-01
-2.50953823e-01 6.62554562e-01 -2.42859989e-01 3.08740139e-03
8.04870486e-01 -1.46345162e+00 5.73402166e-01 9.07172263e-01
4.48908150e-01 1.51440531e-01 1.14155567e+00 -7.72328734e-01
-6.20594501e-01 -8.52292418e-01 1.39919770e+00 -1.02270257e+00
6.30171478e-01 -1.99577391e-01 -4.55799639e-01 3.30218464e-01
7.87767112e-01 -1.07615936e+00 5.91658413e-01 -1.06084622e-01
-2.74127692e-01 4.81370717e-01 -7.58586884e-01 1.07975805e+00
1.19238937e+00 -2.69687653e-01 -7.49392390e-01 5.23878634e-01
1.04186463e+00 -3.24186653e-01 -6.65823817e-01 2.61653364e-01
4.40525979e-01 -1.12420559e+00 8.02041650e-01 -2.54383415e-01
1.26944077e+00 -2.76040494e-01 -1.31285921e-01 -1.54352951e+00
-3.46229225e-01 -7.17039526e-01 7.49907941e-02 1.57059371e+00
8.70871067e-01 -1.88190773e-01 4.53019649e-01 1.82311952e-01
-4.05724287e-01 -5.59704363e-01 -5.66823661e-01 -8.24120104e-01
5.25353253e-02 -6.87892079e-01 6.76094890e-01 9.52562809e-01
4.05026406e-01 9.82982278e-01 -3.82921219e-01 -6.01290405e-01
3.56920868e-01 -5.66473641e-02 9.17615056e-01 -1.09014463e+00
-3.99768740e-01 -8.92336249e-01 -3.16958398e-01 -6.42846942e-01
-1.91141739e-01 -1.11818933e+00 2.62052804e-01 -2.09998608e+00
4.85553324e-01 1.14343420e-01 1.12883046e-01 2.13940382e-01
-3.02759111e-01 3.61654580e-01 4.85652655e-01 4.61007431e-02
-7.56935835e-01 6.80580318e-01 1.59782529e+00 -3.80003721e-01
-3.18245113e-01 -2.60744780e-01 -9.48716819e-01 4.69516784e-01
7.43268013e-01 -2.08493218e-01 -7.66557395e-01 -5.63623369e-01
4.92083848e-01 1.29236132e-01 2.68768787e-01 -1.19400573e+00
-1.38668746e-01 -1.20138466e-01 3.17169636e-01 -6.95367515e-01
6.82888106e-02 4.09845151e-02 1.76056698e-01 6.86983317e-02
-7.18210161e-01 4.57323432e-01 -9.80643928e-02 1.01252161e-01
-2.31502876e-01 -4.30191725e-01 2.12974012e-01 2.82394756e-02
-5.33768713e-01 -1.00967750e-01 -6.64014459e-01 2.82275289e-01
8.19903433e-01 -5.71319163e-01 -5.72401404e-01 -8.89458120e-01
-2.78734714e-01 2.11007610e-01 3.79946470e-01 7.85410881e-01
7.56736934e-01 -1.77737987e+00 -1.51478601e+00 -4.81482834e-01
6.64672196e-01 -4.20105606e-01 4.14349660e-02 6.07000530e-01
-5.41722238e-01 4.64881212e-01 5.28281257e-02 -5.10186553e-01
-1.39045954e+00 4.00979251e-01 -1.85482129e-01 -7.10537195e-01
-3.30405682e-01 7.16911435e-01 1.08084448e-01 -1.50789106e-02
2.49717887e-02 -2.43859753e-01 -5.49933851e-01 3.06572020e-01
7.99241364e-01 6.54582798e-01 7.88628310e-02 -7.91676819e-01
1.48567811e-01 1.40387356e-01 3.17631215e-02 -8.30517411e-01
1.01987326e+00 1.43195381e-02 1.31791845e-01 7.31657684e-01
8.15288901e-01 -8.60563219e-02 -6.35353684e-01 9.98688787e-02
8.02483186e-02 -3.17725182e-01 -1.26325741e-01 -1.12436855e+00
-5.74929357e-01 6.18989408e-01 2.38675207e-01 5.33111691e-01
8.04553986e-01 2.38846228e-01 5.38329065e-01 1.98498145e-01
3.41741592e-01 -1.14039719e+00 5.34640610e-01 6.27172291e-01
1.63075376e+00 -1.10299361e+00 -3.75303328e-02 -3.38319957e-01
-1.17923415e+00 8.96386266e-01 6.29678428e-01 3.03702623e-01
-1.84521571e-01 -9.63451341e-02 8.64743218e-02 -2.58246213e-01
-1.07827342e+00 -2.25564733e-01 5.25306165e-01 8.63938510e-01
8.22417438e-01 2.53644317e-01 -6.34794652e-01 5.97723067e-01
-1.13433385e+00 -5.34510352e-02 7.59463191e-01 6.11707389e-01
-4.20300841e-01 -1.09819949e+00 -4.39556390e-01 4.78429943e-01
1.36782350e-02 -1.51331559e-01 -9.44054425e-01 9.70851898e-01
-6.75586313e-02 1.33261240e+00 -2.99503118e-01 -7.30522573e-01
4.79794353e-01 -9.10654590e-02 6.00313306e-01 -7.06532955e-01
-7.64652610e-01 -2.00512692e-01 6.24847710e-01 -1.98978871e-01
-2.44939759e-01 -7.59636462e-01 -9.77442563e-01 -4.69525933e-01
-1.12718590e-01 2.24255025e-01 2.45381311e-01 8.50290954e-01
-1.32158682e-01 6.69103742e-01 4.26339388e-01 -8.22850704e-01
-2.47537881e-01 -1.25316727e+00 -2.45850623e-01 5.29280245e-01
-1.26391947e-01 -6.08788073e-01 -4.01275814e-01 1.64388835e-01] | [11.756668090820312, 8.871672630310059] |
864d78a7-4ea6-4c94-a1f3-70ba4ba7482f | language-embeddings-sometimes-contain | 2301.08115 | null | https://arxiv.org/abs/2301.08115v1 | https://arxiv.org/pdf/2301.08115v1.pdf | Language Embeddings Sometimes Contain Typological Generalizations | To what extent can neural network models learn generalizations about language structure, and how do we find out what they have learned? We explore these questions by training neural models for a range of natural language processing tasks on a massively multilingual dataset of Bible translations in 1295 languages. The learned language representations are then compared to existing typological databases as well as to a novel set of quantitative syntactic and morphological features obtained through annotation projection. We conclude that some generalizations are surprisingly close to traditional features from linguistic typology, but that most of our models, as well as those of previous work, do not appear to have made linguistically meaningful generalizations. Careful attention to details in the evaluation turns out to be essential to avoid false positives. Furthermore, to encourage continued work in this field, we release several resources covering most or all of the languages in our data: (i) multiple sets of language representations, (ii) multilingual word embeddings, (iii) projected and predicted syntactic and morphological features, (iv) software to provide linguistically sound evaluations of language representations. | ['Murathan Kurfali', 'Robert Östling'] | 2023-01-19 | null | null | null | null | ['multilingual-word-embeddings'] | ['methodology'] | [-1.90318003e-01 1.80017740e-01 -5.11760294e-01 -5.92258990e-01
-6.26336515e-01 -9.18911636e-01 1.00659823e+00 4.31621879e-01
-6.98561907e-01 7.83535659e-01 1.01257765e+00 -8.02968800e-01
-1.03104776e-02 -7.96956837e-01 -5.02490640e-01 -1.77603558e-01
-8.75791907e-02 6.60443008e-01 -9.90345478e-02 -6.49795413e-01
4.44919765e-01 3.27973396e-01 -1.30782259e+00 4.39873397e-01
8.24607909e-01 5.70735991e-01 1.85714021e-01 2.62520045e-01
-4.75853264e-01 8.71899962e-01 -3.59040588e-01 -6.01683021e-01
3.09794024e-02 -2.46128619e-01 -1.21532261e+00 -1.71526015e-01
6.99264288e-01 -2.34267525e-02 -2.49801561e-01 9.90623951e-01
2.28331491e-01 8.18106160e-02 9.00824308e-01 -4.84213710e-01
-1.57368529e+00 1.01869667e+00 -2.23726124e-01 6.58212245e-01
1.36849090e-01 -4.00391454e-03 1.39268136e+00 -1.26206946e+00
1.11995459e+00 1.45571363e+00 7.84139276e-01 4.14660364e-01
-1.29977751e+00 -3.67850751e-01 2.18428642e-01 -1.19351540e-02
-1.14790010e+00 -6.29345179e-01 4.47462291e-01 -6.43225849e-01
1.46006143e+00 -5.84396832e-02 4.08544093e-01 1.17778778e+00
3.42029810e-01 4.07491744e-01 1.33181274e+00 -8.26981246e-01
-2.11557478e-01 4.11544234e-01 3.80483508e-01 8.32116902e-01
4.39317733e-01 3.36763524e-02 -3.69291574e-01 -4.05380756e-01
5.28910220e-01 -4.11792040e-01 -3.55068222e-02 -2.05654755e-01
-1.47975290e+00 1.17742550e+00 3.96915317e-01 8.68555069e-01
-9.88822952e-02 3.55108157e-02 6.96774781e-01 4.94089425e-01
6.92671418e-01 6.80781662e-01 -8.53714705e-01 1.00834519e-01
-5.70923507e-01 2.02872843e-01 7.94688404e-01 8.27305675e-01
8.66286814e-01 2.97253877e-01 3.55788320e-01 1.36804008e+00
2.32493445e-01 3.38215113e-01 7.60673523e-01 -8.46549511e-01
5.38195431e-01 4.79181319e-01 -1.63335174e-01 -1.19185591e+00
-5.64801097e-01 1.63131971e-02 -4.23436612e-01 -1.83728084e-01
4.35726106e-01 -1.89922184e-01 -4.86115277e-01 1.94596589e+00
-2.65904099e-01 -7.84990132e-01 6.57635555e-02 6.28312171e-01
1.02559948e+00 6.83620512e-01 3.86088789e-01 2.03280207e-02
1.51901984e+00 -4.15090173e-01 -4.38492388e-01 -4.02726561e-01
8.53468001e-01 -9.88687515e-01 1.17168546e+00 -3.83270308e-02
-1.05311859e+00 -5.68173289e-01 -8.06634128e-01 -5.18050373e-01
-7.95554459e-01 1.00381644e-02 1.00305009e+00 4.60342199e-01
-1.13260710e+00 3.89064968e-01 -7.13177741e-01 -9.96955514e-01
-2.27618460e-02 8.21628720e-02 -4.68035609e-01 2.39376947e-02
-1.38811481e+00 1.38642299e+00 6.47381961e-01 -2.15614498e-01
-4.11551654e-01 -3.73808920e-01 -1.18771446e+00 -2.99868196e-01
6.43431842e-02 -5.47807515e-01 1.13579798e+00 -1.19088793e+00
-8.70118201e-01 1.39397478e+00 -1.27269715e-01 -2.45674700e-01
-1.27471268e-01 1.41669229e-01 -6.62782907e-01 -1.00733839e-01
3.86114955e-01 8.74750853e-01 -4.82174605e-02 -1.14810216e+00
-5.68316400e-01 -3.69991839e-01 1.77928373e-01 1.51124120e-01
-6.23237669e-01 4.99116391e-01 2.32129078e-02 -1.10768080e+00
-1.11858003e-01 -8.11214983e-01 -1.07273422e-01 -1.81582332e-01
-1.03522517e-01 -6.29505932e-01 -1.55393910e-02 -1.06712604e+00
9.83458161e-01 -1.85757220e+00 4.70957048e-02 -1.34045318e-01
7.97263533e-02 -5.85534312e-02 -3.76334548e-01 6.51299834e-01
-1.15415841e-01 6.01480842e-01 -5.74379824e-02 3.05824634e-03
1.67369112e-01 6.11029923e-01 -4.95995641e-01 6.42206609e-01
3.81279171e-01 9.22509909e-01 -9.52868164e-01 -4.12792444e-01
4.68857884e-02 3.51040423e-01 -6.01103246e-01 -4.01951998e-01
-2.46158630e-01 7.41858780e-02 -2.26886764e-01 5.79496741e-01
1.64918423e-01 2.37183347e-01 4.65282112e-01 5.06771496e-03
-4.03184623e-01 1.09021986e+00 -6.02024317e-01 1.63690102e+00
-5.96546054e-01 7.54239500e-01 -1.42732874e-01 -1.15872908e+00
1.09137046e+00 3.49640608e-01 2.75313966e-02 -7.45924890e-01
1.47040114e-01 5.16360402e-01 3.18153679e-01 -5.08678555e-01
8.14694583e-01 -3.20922196e-01 -4.25743163e-01 6.68970823e-01
5.80266535e-01 1.39412194e-01 3.95348370e-01 -3.29329893e-02
6.96237147e-01 1.69972986e-01 5.54810226e-01 -1.00518119e+00
2.43224338e-01 3.99074733e-01 6.54665351e-01 3.96933585e-01
-3.68118212e-02 2.22847164e-01 3.02019745e-01 -9.25995648e-01
-1.44110620e+00 -9.52724755e-01 -5.97880065e-01 1.66549218e+00
-4.53999966e-01 -3.37966800e-01 -3.78676087e-01 -2.64358640e-01
1.79029778e-02 8.24262321e-01 -7.19558835e-01 2.02951059e-01
-7.55468726e-01 -8.60289693e-01 8.27221692e-01 7.80378938e-01
-1.12487644e-01 -1.43187511e+00 -3.02896470e-01 1.85519442e-01
-4.08154279e-02 -8.86085987e-01 -1.92693561e-01 4.71262813e-01
-9.00024056e-01 -9.00654256e-01 -4.51682508e-01 -1.38842618e+00
4.00625259e-01 1.56882703e-02 1.36596489e+00 1.29164994e-01
-8.90803933e-02 2.29753420e-01 -4.52005535e-01 -3.30921590e-01
-8.39152038e-01 2.93595493e-01 3.51917833e-01 -7.82315075e-01
1.04768717e+00 -4.41743523e-01 8.68887976e-02 -8.98974761e-02
-8.30708385e-01 -4.07425493e-01 4.12558228e-01 8.64871144e-01
3.51310313e-01 -4.13175911e-01 6.67760849e-01 -9.86036122e-01
9.57997143e-01 -6.95666134e-01 -3.24767709e-01 1.62716851e-01
-4.12171841e-01 5.94617687e-02 7.16126323e-01 -1.74402133e-01
-8.69252145e-01 -4.53193814e-01 -2.37564787e-01 2.63239771e-01
-2.97415584e-01 9.00717914e-01 1.73095450e-01 3.42773855e-01
1.14146912e+00 1.13963529e-01 -1.44933522e-01 -7.48683631e-01
7.86788225e-01 7.77089715e-01 5.42727053e-01 -1.05353081e+00
5.98247468e-01 1.23819277e-01 -6.15115881e-01 -1.11867499e+00
-7.07693517e-01 -1.23424940e-01 -9.73322809e-01 4.65713918e-01
7.95227647e-01 -9.62141991e-01 -2.32820243e-01 -1.67232260e-01
-1.24007177e+00 -2.02721015e-01 -9.90870073e-02 5.04084349e-01
-5.43313861e-01 2.85013974e-01 -1.00225377e+00 -3.95189524e-01
-8.46177191e-02 -9.08441186e-01 6.76483095e-01 -2.24322081e-01
-7.74350882e-01 -1.60207093e+00 3.59873861e-01 1.52099088e-01
4.44972306e-01 1.48780972e-01 1.63028872e+00 -1.01007831e+00
-1.97188295e-02 1.93071604e-01 -1.56780392e-01 3.12536627e-01
1.67860270e-01 9.27110016e-02 -6.82969689e-01 -2.87368178e-01
-1.99762508e-01 -7.41452515e-01 9.73434865e-01 8.23732391e-02
7.16749966e-01 -4.78151977e-01 -2.19959561e-02 3.56664777e-01
1.40216434e+00 4.82597761e-02 1.00070417e-01 5.81254900e-01
4.41191912e-01 8.62510085e-01 9.44868848e-03 2.11591963e-02
6.38420463e-01 4.47300673e-01 -2.02064767e-01 3.16629671e-02
-2.01106489e-01 -4.12229687e-01 6.52810454e-01 1.47110641e+00
-1.17898636e-01 -7.11769387e-02 -1.35796547e+00 9.72850323e-01
-1.46644807e+00 -1.01933682e+00 8.96077752e-02 1.76225436e+00
9.58995998e-01 -1.47203384e-02 1.71404272e-01 -2.50197977e-01
5.52664876e-01 3.85368466e-01 -2.30285034e-01 -1.09558260e+00
-3.20169032e-01 3.58241528e-01 4.18784231e-01 5.82483888e-01
-1.00423276e+00 1.43557537e+00 7.64679813e+00 4.49475676e-01
-8.50679219e-01 2.28737757e-01 4.68262136e-01 2.90868908e-01
-8.19690764e-01 8.56681839e-02 -7.40311086e-01 -6.08284120e-03
1.04621434e+00 -3.31147015e-01 5.46427310e-01 7.11711049e-01
-2.03479111e-01 2.61774242e-01 -1.30044949e+00 3.77042562e-01
3.00175875e-01 -1.35455751e+00 2.67753869e-01 8.95047337e-02
5.32055140e-01 6.32431865e-01 -5.80667071e-02 5.49963653e-01
8.47383738e-01 -1.26161563e+00 7.51235664e-01 2.26956218e-01
9.66090381e-01 -7.27834821e-01 5.19969463e-01 3.43912214e-01
-9.06495512e-01 -1.67681072e-02 -9.35765982e-01 -3.38256627e-01
-5.87247349e-02 1.74112409e-01 -5.99727213e-01 3.10879529e-01
6.17807686e-01 7.54922867e-01 -8.38945329e-01 4.04230863e-01
-3.33973467e-01 7.42032945e-01 -1.41100183e-01 -3.22290301e-01
5.74066401e-01 1.12241521e-01 3.93396765e-01 1.58536255e+00
6.31110594e-02 -2.26519719e-01 2.73738116e-01 9.36657488e-01
-1.85526833e-01 6.20099664e-01 -1.03769147e+00 -4.19272274e-01
4.87632573e-01 1.01562881e+00 -6.67823315e-01 -1.94040954e-01
-9.27876532e-01 4.20204848e-01 8.83130372e-01 3.02790195e-01
-2.37120122e-01 -3.55931073e-01 6.93328381e-01 7.81465843e-02
-1.05975140e-02 -5.91280162e-01 -2.24610627e-01 -1.37098634e+00
-1.52251989e-01 -8.88016701e-01 4.60382402e-01 -6.87913716e-01
-1.91871667e+00 8.01844060e-01 1.09873503e-01 -5.83823621e-01
-4.80907798e-01 -1.08782172e+00 -3.33992362e-01 9.39634740e-01
-1.11112082e+00 -1.05640638e+00 5.50011218e-01 2.98353076e-01
5.88303983e-01 -6.72125816e-01 1.22365582e+00 1.66369870e-01
-1.80162355e-01 5.56453109e-01 1.65294036e-01 6.34843051e-01
7.50583827e-01 -1.21004510e+00 6.71067953e-01 4.12736088e-01
7.06207335e-01 1.13671279e+00 5.36090791e-01 -4.78930920e-01
-1.05063796e+00 -8.10366631e-01 1.47534716e+00 -8.54169667e-01
1.22616112e+00 -5.46006560e-01 -1.01875186e+00 1.32055581e+00
7.55248189e-01 -2.32356921e-01 9.39146519e-01 7.42560267e-01
-4.66662318e-01 3.26293290e-01 -8.22626412e-01 6.26059055e-01
1.14224470e+00 -6.94628417e-01 -1.47443020e+00 5.28856039e-01
7.62207448e-01 1.27620548e-01 -9.55670714e-01 1.85487449e-01
5.73270857e-01 -6.34887099e-01 7.38890171e-01 -1.11191976e+00
7.01259136e-01 9.56449807e-02 -6.06644094e-01 -1.43223405e+00
-6.24145508e-01 1.04903214e-01 5.67779899e-01 1.51126754e+00
7.61232197e-01 -7.09188998e-01 2.43098646e-01 2.14173749e-01
-2.37218529e-01 -6.44781232e-01 -9.27990139e-01 -7.28601217e-01
1.17274058e+00 -4.44066674e-01 2.84396619e-01 1.47516608e+00
3.12219501e-01 6.65995061e-01 -3.84950964e-03 -2.75262535e-01
1.94615841e-01 2.76441574e-02 3.55612725e-01 -1.26367307e+00
-6.21474348e-02 -7.65699089e-01 -5.35523057e-01 -6.67889774e-01
6.19572163e-01 -1.74163759e+00 -2.44328693e-01 -1.55524993e+00
3.69365156e-01 -4.64162529e-01 -2.58269101e-01 6.51723981e-01
2.23180190e-01 2.88037777e-01 -2.08696499e-02 3.72614980e-01
-1.64315656e-01 3.40375304e-01 8.23136687e-01 -1.30697861e-01
1.47642210e-01 -7.80216694e-01 -1.19104052e+00 1.15064871e+00
7.14233696e-01 -5.06922424e-01 -1.34043932e-01 -1.04494214e+00
4.44250673e-01 -3.35394472e-01 4.13562283e-02 -5.08761525e-01
-2.35930398e-01 -3.21758747e-01 6.94968581e-01 -2.15226218e-01
3.53068002e-02 -4.02062029e-01 -4.31007206e-01 3.25738668e-01
-6.80746317e-01 5.89533091e-01 3.42774421e-01 8.51138607e-02
-2.68883526e-01 -2.85488665e-01 6.59409225e-01 -5.02147615e-01
-1.03110611e+00 9.79324803e-03 -8.31807554e-01 4.07532722e-01
6.34455979e-01 1.15091681e-01 -4.14986372e-01 -1.22675136e-01
-7.83314705e-01 4.55253907e-02 7.57440567e-01 8.91464233e-01
1.00951165e-01 -1.52758729e+00 -8.35272253e-01 -3.48701626e-02
4.72044855e-01 -5.58941007e-01 -4.93358374e-01 3.69142979e-01
-5.25596380e-01 6.93607926e-01 -2.82513350e-01 -1.39726967e-01
-7.05287635e-01 7.57249355e-01 1.42550692e-01 1.70238665e-04
-6.88092649e-01 6.08001053e-01 1.12936005e-01 -1.05976236e+00
-3.93610038e-02 -4.52934772e-01 -3.62074405e-01 2.55028546e-01
3.37790668e-01 -3.86931598e-02 -1.21690303e-01 -1.31358469e+00
-3.98168832e-01 4.93605882e-01 -1.72553048e-01 -1.72492698e-01
1.57138646e+00 -2.68078670e-02 -4.12263393e-01 9.57425296e-01
1.19132793e+00 2.64159918e-01 -3.61754954e-01 -5.28016329e-01
3.99918735e-01 -8.77432451e-02 -2.85039485e-01 -8.31117272e-01
-5.50861120e-01 9.99774098e-01 2.41406411e-01 1.38800636e-01
5.61440110e-01 3.00338030e-01 7.01380134e-01 6.84191823e-01
4.31456536e-01 -1.29564977e+00 -3.42988640e-01 9.25568223e-01
9.82735753e-01 -9.82298374e-01 2.58752517e-02 7.23758042e-02
-5.74691474e-01 1.25923097e+00 2.96906322e-01 -4.52281564e-01
6.67566121e-01 1.99678659e-01 2.59359628e-01 -2.16857448e-01
-9.69736338e-01 -1.03709958e-01 1.76391885e-01 6.19650722e-01
1.23714769e+00 1.62398964e-01 -7.43863046e-01 5.34661055e-01
-6.84103131e-01 -2.79955119e-01 4.43962336e-01 6.60858631e-01
-6.50968134e-01 -1.17835522e+00 -2.64394641e-01 4.02847230e-01
-6.54672801e-01 -4.53820765e-01 -6.63572311e-01 1.19834828e+00
2.61929393e-01 6.71341062e-01 2.63078392e-01 -1.29877731e-01
2.71808773e-01 4.08991337e-01 3.95026743e-01 -1.01598895e+00
-5.64402103e-01 -1.39984384e-01 4.73005265e-01 -1.12862453e-01
-4.72244322e-01 -7.62154400e-01 -1.05492830e+00 -4.71375287e-01
1.18216500e-01 2.46130694e-02 5.59065044e-01 9.80899692e-01
5.09938225e-03 3.33780125e-02 5.84591664e-02 -7.84024119e-01
-6.26194894e-01 -1.23157334e+00 -5.98099530e-01 4.27190572e-01
-8.95766690e-02 -5.24899721e-01 -1.92811668e-01 -3.72518064e-03] | [10.807023048400879, 9.733755111694336] |
031b2243-1499-45f8-9859-a917db08062d | improved-differential-neural-cryptanalysis | 2301.11601 | null | https://arxiv.org/abs/2301.11601v1 | https://arxiv.org/pdf/2301.11601v1.pdf | Improved Differential-neural Cryptanalysis for Round-reduced Simeck32/64 | In CRYPTO 2019, Gohr presented differential-neural cryptanalysis by building the differential distinguisher with a neural network, achieving practical 11-, and 12-round key recovery attack for Speck32/64. Inspired by this framework, we develop the Inception neural network that is compatible with the round function of Simeck to improve the accuracy of the neural distinguishers, thus improving the accuracy of (9-12)-round neural distinguishers for Simeck32/64. To provide solid baselines for neural distinguishers, we compute the full distribution of differences induced by one specific input difference up to 13-round Simeck32/64. Moreover, the performance of the DDT-based distinguishers in multiple ciphertext pairs is evaluated. Compared with the DDT-based distinguishers, the 9-, and 10-round neural distinguishers achieve better accuracy. Also, an in-depth analysis of the wrong key response profile revealed that the 12-th and 13-th bits of the subkey have little effect on the score of the neural distinguisher, thereby accelerating key recovery attacks. Finally, an enhanced 15-round and the first practical 16-, and 17-round attacks are implemented for Simeck32/64, and the success rate of both the 15-, and 16-round attacks is almost 100%. | ['Chao Li', 'Zilong Wang', 'Jinyu Lu', 'Liu Zhang'] | 2023-01-27 | null | null | null | null | ['cryptanalysis'] | ['miscellaneous'] | [-8.8827491e-02 -3.4062776e-01 -6.4428180e-02 8.0964074e-02
-9.3857580e-01 -1.1158069e+00 1.5630925e-01 1.6660941e-01
-6.9145751e-01 4.7836336e-01 -4.4745591e-01 -1.2604643e+00
1.0124594e-01 -9.5744687e-01 -6.1788225e-01 -9.2111051e-01
-7.8522491e-01 -1.0095711e-01 -4.0284269e-02 -6.2433487e-01
5.2242112e-01 4.9460018e-01 -9.3252403e-01 6.6594768e-01
4.7855049e-01 1.3511174e+00 -7.3367000e-01 1.4141902e+00
3.4493464e-01 5.0833440e-01 -9.5001197e-01 -7.9658359e-01
1.0075073e+00 -2.4373528e-01 -8.6467254e-01 -1.2706257e+00
4.5164457e-01 -9.7936517e-01 -9.4755226e-01 1.2284850e+00
8.1188196e-01 -5.1094079e-01 2.7049154e-01 -1.2220492e+00
-4.6497528e-02 1.2260007e+00 1.1732570e-02 1.5236115e-01
2.7649373e-01 4.8944455e-01 1.0954568e+00 -6.0506940e-01
1.8505751e-01 1.0700040e+00 9.9259394e-01 3.4175149e-01
-9.9042332e-01 -1.3147531e+00 -4.6302819e-01 4.8405176e-01
-1.7509303e+00 -2.4296133e-01 2.9226699e-01 1.7054234e-01
1.0304995e+00 4.8403293e-01 5.2259922e-01 6.4499396e-01
4.9860546e-01 5.9404510e-01 1.0819125e+00 -2.9716593e-01
2.2044710e-03 -1.4929439e-01 4.3179539e-01 3.3516529e-01
5.6909716e-01 6.6147602e-01 -3.4403324e-02 -6.7963272e-01
4.9565694e-01 -6.3362792e-02 -5.4313451e-01 3.2936707e-01
-1.0817310e+00 8.3721340e-01 4.2259842e-01 2.4860615e-01
-2.2824724e-01 5.1416206e-01 5.9640253e-01 9.1103446e-01
-2.7001819e-01 6.6850716e-01 -6.1703122e-01 -2.8643337e-01
-1.0491935e+00 5.2012360e-01 1.3421414e+00 6.1704564e-01
1.0152317e+00 2.7601171e-01 -6.8200149e-02 -3.6846498e-01
-1.6704148e-01 7.4280053e-01 3.3252782e-01 -7.6037991e-01
6.4950055e-01 2.7071849e-01 -4.7059783e-01 -7.9923004e-01
-5.0692755e-01 -4.3656978e-01 -1.1582059e+00 5.1860404e-01
8.9373285e-01 -6.1802500e-01 -6.8577427e-01 1.4909284e+00
-1.0919492e-01 1.4299215e-01 6.3745964e-01 5.9483486e-01
5.8082247e-01 6.0710096e-01 -6.0667664e-01 2.1656813e-01
1.3056043e+00 -3.8071051e-01 -3.1326264e-01 1.6945755e-01
1.3580750e+00 -8.4344417e-01 3.1545383e-01 5.3676385e-01
-1.1985402e+00 -3.8906464e-01 -1.5189215e+00 1.4778149e-01
-3.1380484e-01 -1.5937793e-01 7.9424435e-01 1.0324323e+00
-1.1940422e+00 7.3315585e-01 -3.3355540e-01 6.4559239e-01
1.1114897e-02 1.0459143e+00 -3.6956269e-01 8.2503691e-02
-1.8079684e+00 5.6370986e-01 7.1329218e-01 7.4734688e-02
-7.6445317e-01 -7.2591370e-01 -4.9728152e-01 2.3159662e-01
8.0334343e-02 -1.8148005e-01 1.2886084e+00 -7.2405744e-01
-1.5353253e+00 6.5094888e-01 2.1234863e-01 -1.1432896e+00
2.2786237e-01 3.0495036e-01 -3.8724315e-01 2.1712920e-01
-5.7451713e-01 3.1222868e-01 2.6756510e-01 -7.9318964e-01
-7.1675360e-01 -1.2120506e-01 4.3631849e-01 -1.7925176e-01
3.6430005e-02 -2.6191953e-01 1.1490738e-01 -4.5944110e-01
3.2267261e-02 -1.1494855e+00 -2.8197807e-01 -6.6971767e-01
-6.7648995e-01 1.6186216e-01 5.0549787e-01 -6.1491793e-01
1.1988311e+00 -2.1403487e+00 -6.3378543e-01 9.6435410e-01
3.8785705e-01 9.2779571e-01 -1.5049823e-01 6.0133368e-01
-2.4377991e-01 -4.8562557e-02 2.8439060e-02 1.3180958e-01
2.0577323e-02 -1.2959191e-01 -7.1785676e-01 7.0099401e-01
-1.0509466e-01 7.6230001e-01 -7.4273020e-01 3.9635646e-01
-4.3873033e-01 1.8263273e-01 -8.0480599e-01 1.5663551e-01
2.7561361e-01 -3.9964430e-02 -1.4312769e-01 5.9428394e-01
1.3483028e+00 2.1058042e-03 4.3870074e-01 -1.8699400e-01
-5.7218969e-02 8.8023496e-01 -1.2486966e+00 1.1411455e+00
-2.3810759e-01 8.4758705e-01 2.9193668e-02 -5.4656446e-01
8.5361540e-01 3.6347625e-01 -1.4660317e-01 -8.5784745e-01
3.9816636e-01 6.4259666e-01 7.2758722e-01 2.9265720e-01
1.1209605e+00 1.6116275e-01 -3.6313704e-01 7.2427046e-01
-2.0249195e-01 -8.4291592e-02 -9.7038388e-02 4.0634811e-01
9.9905854e-01 -6.6054821e-01 1.5240531e-01 -1.2573379e-01
7.0057303e-01 -3.5526860e-01 5.4793322e-01 1.2137420e+00
-1.7798785e-03 5.4761142e-01 7.9541826e-01 -5.4123890e-01
-6.1586505e-01 -1.0010244e+00 -5.7288248e-02 9.0989524e-01
3.7307027e-01 -5.6057876e-01 -8.5726655e-01 -4.2372498e-01
2.5561917e-01 7.9096198e-01 -3.9603978e-01 -5.0674903e-01
-7.9364061e-01 -8.0133754e-01 2.0953248e+00 5.1744014e-01
8.1134516e-01 -3.1034568e-01 -2.5248909e-01 5.2322052e-02
1.1938355e-01 -8.8150883e-01 -3.7922600e-01 7.6288569e-01
-5.0817066e-01 -1.4206992e+00 -3.0124098e-01 -7.1220404e-01
3.8958618e-01 1.3618210e-01 6.3384193e-01 2.6410741e-01
2.0590878e-01 -1.5047079e-01 2.4702435e-03 -7.6687150e-02
-8.3484364e-01 2.8474972e-02 2.3523639e-01 -4.0072963e-01
3.2122335e-01 -5.1252866e-01 -6.7534560e-01 3.5031971e-01
-9.5305568e-01 -5.2395117e-01 7.1122330e-01 1.0007228e+00
5.7336152e-02 -1.3316362e-01 9.9068902e-02 -5.7280564e-01
5.8528221e-01 -1.4715564e-01 -9.0782773e-01 -2.5934208e-02
-9.7868800e-01 3.4636813e-01 1.1139065e+00 -8.1332338e-01
-3.7615862e-01 -4.4199607e-01 -5.2999914e-01 -4.0643662e-03
4.9385053e-01 8.6291611e-02 -3.7377436e-02 -8.7404478e-01
1.1056105e+00 3.8519019e-01 -5.9331398e-02 -6.6032581e-02
4.3747020e-01 8.0894262e-01 9.7387034e-01 -6.8667555e-01
1.0781789e+00 1.4382404e-01 1.2062777e-01 -1.0645906e-01
9.3693603e-03 -1.4470972e-01 8.5198425e-02 5.6066382e-01
5.6515880e-02 -1.1288434e+00 -1.6438931e+00 1.2177703e+00
-1.3465319e+00 -3.9182559e-01 -6.7863181e-02 5.4585040e-01
9.4141275e-02 7.9874307e-01 -1.1103815e+00 -5.2832454e-01
-8.8258278e-01 -1.5886921e+00 3.9311612e-01 2.9491311e-01
-1.3809188e-01 -9.6549022e-01 -3.2742840e-01 -3.8960707e-01
5.5697697e-01 -1.1512383e-01 9.1952670e-01 -1.3975973e+00
-7.3013866e-01 -6.4765579e-01 -3.3976695e-01 8.2833749e-01
-3.5809830e-01 -2.9039180e-01 -1.0466536e+00 -5.5839771e-01
1.6193563e-01 -2.6476732e-01 8.2290208e-01 -1.1092944e-01
8.2712609e-01 -5.5718291e-01 3.3726662e-02 1.5053282e+00
1.0475957e+00 3.6250499e-01 1.0612155e+00 5.3188694e-01
5.2157634e-01 -5.1061038e-02 5.5296976e-02 5.6155539e-01
5.2927709e-01 1.6279767e-01 3.3240673e-01 -2.0273710e-03
4.3052047e-01 5.8643047e-02 7.2928488e-01 7.5706732e-01
7.9450710e-03 -1.9797821e-02 -6.8106574e-01 -7.3861048e-02
-8.4020358e-01 -8.5257173e-01 -3.4932414e-01 2.6750922e+00
1.2949641e+00 3.1132188e-01 -3.3285388e-01 7.1822506e-01
4.7307259e-01 6.5796636e-02 -2.5750515e-01 -9.8403519e-01
-4.4450986e-01 8.9859259e-01 1.5052723e+00 4.8852885e-01
-7.9447144e-01 8.0369163e-01 6.6590395e+00 1.2095865e+00
-1.6958978e+00 -5.0321531e-01 7.1338171e-01 2.9435277e-01
-4.5482862e-01 3.3786809e-01 -1.2343030e+00 3.4369400e-01
1.2711338e+00 -2.9578319e-01 7.5939465e-01 7.2946620e-01
-5.3568721e-01 -8.0983147e-02 -1.2688904e+00 9.3552011e-01
-2.8309575e-01 -1.4176729e+00 -1.3557644e-01 2.5357822e-01
6.8096936e-01 -1.2521935e-01 3.5814598e-01 6.4276099e-01
2.2650810e-01 -9.1078413e-01 4.5570758e-01 -1.3476187e-01
1.3352947e+00 -1.1401731e+00 1.1139913e+00 1.1417174e-01
-8.7814724e-01 -1.1986642e-01 -2.6421627e-01 -2.9178911e-01
-5.6984121e-01 5.1648062e-01 -1.1628705e+00 5.0335282e-01
1.5933022e-01 -3.4889960e-01 -2.7467009e-01 8.6166388e-01
-4.7477403e-01 8.2573313e-01 -4.7951543e-01 -1.9968657e-01
5.6443858e-01 2.1404940e-01 4.6255127e-01 1.4300137e+00
3.1280339e-01 -5.9570454e-02 -2.6685709e-01 4.3018931e-01
-3.9545166e-01 -3.9029491e-01 -7.1311034e-02 1.5344696e-01
7.3947209e-01 1.1951845e+00 -1.9455893e-01 -5.1795864e-01
3.0316815e-01 7.9577774e-01 -2.5164494e-01 3.0081037e-01
-6.3845581e-01 -1.1582342e+00 1.0822030e+00 -3.6086467e-01
2.5504249e-01 -2.4533576e-01 -5.1591229e-01 -9.5653278e-01
-2.0198315e-01 -1.7421614e+00 2.4969606e-01 1.4788374e-01
-7.6702064e-01 5.2452278e-01 -1.9204211e-01 -1.0004543e+00
-7.2016901e-01 -7.4642324e-01 -8.9784974e-01 1.2188450e+00
-1.5722007e+00 -5.7386798e-01 1.3478009e-01 6.1998147e-01
-6.5246987e-01 -2.8960484e-01 1.0234612e+00 2.2076236e-01
-5.7491839e-02 1.7412224e+00 6.2355161e-01 6.1023730e-01
7.6540309e-01 -1.0967664e+00 1.1359054e+00 9.4458097e-01
-3.9402121e-01 1.1338270e+00 4.1658741e-01 -3.3414367e-01
-2.1448638e+00 -5.8334196e-01 5.5436128e-01 2.4099480e-01
6.0730708e-01 -2.8466189e-01 -8.2760519e-01 5.4957461e-01
-2.1180560e-01 -8.6774878e-02 5.9845304e-01 -1.4768966e-01
-8.7789208e-01 -2.4036416e-01 -1.3179237e+00 5.5283827e-01
4.2869654e-01 -8.8699698e-01 -2.1266335e-01 -8.0299437e-02
4.7516686e-01 -1.0801247e+00 -6.4958984e-01 1.9711843e-01
8.0593866e-01 -1.2193527e+00 1.1319184e+00 -1.3684186e-01
-1.3040480e-01 -1.0191418e-01 -2.4612717e-01 -9.8554361e-01
-2.0972822e-02 -1.4500058e+00 -3.7148541e-01 5.9401947e-01
2.1675412e-01 -1.2258244e+00 8.3375680e-01 1.7682530e-01
-3.6787607e-02 -6.0755247e-01 -1.1213045e+00 -6.3893074e-01
5.4930276e-01 -6.2781221e-01 9.3945378e-01 6.6135740e-01
-1.6354485e-01 -3.1128585e-01 -3.5093573e-01 4.4521129e-01
3.2912967e-01 8.5506290e-02 1.1703175e+00 -7.7058947e-01
-7.2763151e-01 -6.1406326e-01 -6.5165079e-01 -1.3772070e+00
-9.9860728e-02 -1.0835053e+00 -2.3832646e-01 -3.8123280e-01
-3.1202450e-01 -6.1524343e-01 -4.5730269e-01 4.6421185e-01
-2.4767412e-01 3.5329303e-01 7.1420312e-02 -1.6197439e-01
-5.1610261e-02 -2.0089464e-01 7.7636671e-01 -8.4886834e-02
-1.7813705e-01 1.5902863e-01 -8.5651195e-01 4.9431133e-01
6.9620621e-01 -4.1836166e-01 2.3696424e-01 2.3005553e-02
3.7170321e-01 1.3274327e-01 2.8487667e-01 -1.1058005e+00
4.9838346e-01 3.2000408e-01 3.0355662e-01 -5.7529020e-01
1.6572137e-01 -2.1070476e-01 -1.3569088e-01 1.0385277e+00
-9.1376893e-02 3.6129218e-01 2.8663591e-01 -1.2728198e-01
-8.6349308e-02 -8.8734694e-02 7.5745654e-01 3.7281579e-01
-5.7973135e-01 2.6353994e-01 -4.5830458e-01 3.8539463e-01
5.1734608e-01 -2.4153808e-01 -6.3125741e-01 -3.5152587e-01
8.9273512e-02 8.1115015e-02 3.5782436e-01 -3.3318812e-01
5.8467466e-01 -1.1018630e+00 -7.0475048e-01 6.8351644e-01
-2.6712596e-01 2.2992762e-02 8.1447400e-02 6.9242346e-01
-9.5321912e-01 4.2666939e-01 -1.9427218e-01 -2.6998663e-01
-1.2668825e+00 1.2614799e-01 4.8189923e-01 -7.9189461e-01
-1.3680007e-01 1.0907222e+00 8.9348981e-04 -3.7954265e-01
3.3376393e-01 -6.2835503e-01 6.6020167e-01 -1.4439632e-01
9.5022839e-01 5.8147860e-01 2.5579748e-01 -1.7954852e-01
-4.2813319e-01 3.1523255e-01 -1.1540399e-01 3.6636397e-02
5.8717722e-01 4.0716022e-01 -3.0362263e-01 -1.2145922e-01
1.7472363e+00 3.3887222e-01 -5.7134974e-01 -3.4741566e-01
-3.1281227e-01 7.2093390e-02 -1.2065827e-01 -5.7907522e-01
-9.4141096e-01 9.8276317e-01 5.7073015e-01 2.4388129e-02
1.2157501e+00 -1.1054614e+00 1.5671999e+00 1.0758168e+00
4.3552086e-01 -7.5719947e-01 -5.8040541e-01 1.2631123e+00
-7.6617233e-02 -7.1232158e-01 -6.3444927e-02 8.1838124e-02
-8.0125690e-02 1.7945454e+00 -1.7509481e-02 -1.9867857e-01
2.8454927e-01 8.8131797e-01 2.3666354e-01 4.9276128e-01
-6.4819312e-01 3.0763334e-01 3.3977103e-01 2.7903491e-01
1.0373343e-02 1.0997712e-01 -2.9555258e-01 9.3062615e-01
-5.9134132e-01 -1.9144809e-01 8.1627834e-01 6.8576807e-01
-6.2659323e-02 -1.4463369e+00 -7.6263899e-01 6.9028705e-02
-1.0723668e+00 -7.4327987e-01 -1.7726918e-01 7.7504051e-01
2.8777154e-02 7.1864426e-01 -7.4636422e-02 -1.3100936e+00
-2.3066950e-01 -7.3067336e-03 2.4150462e-01 2.2510779e-01
-1.3979431e+00 -4.3386954e-01 1.5381944e-01 -5.9794253e-01
7.0057487e-01 -1.0095823e-01 -1.6825730e+00 -1.1983138e+00
-4.8413154e-01 4.4178692e-01 8.8905138e-01 7.2452354e-01
3.3783296e-01 -3.5758805e-02 1.2866203e+00 -8.3505744e-01
-1.3307061e+00 -4.6888304e-01 -5.7498056e-01 -1.2912878e-01
6.2798512e-01 2.3206164e-01 -7.9236835e-01 -7.0096219e-01] | [5.9021077156066895, 7.344762802124023] |
ada00aef-96b7-47e2-aeea-f8b60ab38ed8 | bias-analysis-and-mitigation-in-the | null | null | https://aclanthology.org/P19-1634 | https://aclanthology.org/P19-1634.pdf | Bias Analysis and Mitigation in the Evaluation of Authorship Verification | The PAN series of shared tasks is well known for its continuous and high quality research in the field of digital text forensics. Among others, PAN contributions include original corpora, tailored benchmarks, and standardized experimentation platforms. In this paper we review, theoretically and practically, the authorship verification task and conclude that the underlying experiment design cannot guarantee pushing forward the state of the art{---}in fact, it allows for top benchmarking with a surprisingly straightforward approach. In this regard, we present a {``}Basic and Fairly Flawed{''} (BAFF) authorship verifier that is on a par with the best approaches submitted so far, and that illustrates sources of bias that should be eliminated. We pinpoint these sources in the evaluation chain and present a refined authorship corpus as effective countermeasure. | ['Janek Bevendorff', 'Benno Stein', 'Matthias Hagen', 'Martin Potthast'] | 2019-07-01 | null | null | null | acl-2019-7 | ['authorship-verification'] | ['natural-language-processing'] | [ 9.37968418e-02 -7.00291712e-03 -4.88034002e-02 1.99062750e-02
-7.45748341e-01 -8.01675498e-01 8.53903651e-01 3.25363636e-01
-6.79849029e-01 8.48665714e-01 1.64083112e-02 -3.45295995e-01
-3.25949699e-01 -2.93805927e-01 -5.61708093e-01 -4.51090544e-01
1.41856223e-01 6.34014130e-01 3.61484826e-01 -7.43463337e-02
9.87447739e-01 6.09066427e-01 -1.38726473e+00 1.39239490e-01
5.69151640e-01 3.96529794e-01 -4.48521942e-01 7.57825971e-01
3.67123224e-02 1.00979483e+00 -1.14306331e+00 -1.47342598e+00
9.09545049e-02 -2.12045610e-01 -1.23338640e+00 -5.23209274e-01
8.06662917e-01 -3.79414678e-01 -7.52109066e-02 9.27033961e-01
8.35494995e-01 -3.99075598e-01 3.53631467e-01 -1.63770568e+00
-6.96596265e-01 1.03447747e+00 -6.44018292e-01 4.65201825e-01
3.67942750e-01 2.91006505e-01 1.19294310e+00 -6.29836321e-01
1.16681838e+00 9.14322913e-01 9.83471751e-01 4.86592054e-01
-9.73158777e-01 -7.94999838e-01 -3.33320469e-01 2.77008861e-01
-1.18704641e+00 -5.10354042e-01 6.60852551e-01 -6.00370169e-01
8.27109993e-01 2.67815202e-01 1.57235205e-01 1.85336506e+00
-1.06168896e-01 7.58256137e-01 1.38656557e+00 -5.60208619e-01
-4.87805232e-02 1.18887007e-01 6.39025629e-01 5.19590735e-01
1.00688934e+00 7.32461736e-03 -9.09178376e-01 -6.34405732e-01
8.24154839e-02 -5.41805089e-01 -1.36888936e-01 -1.01296917e-01
-1.24880588e+00 6.19419813e-01 -3.18780839e-01 7.28064537e-01
-9.65159908e-02 3.98620784e-01 9.61243391e-01 3.92761290e-01
1.60520971e-01 7.08905697e-01 -1.66871548e-01 -5.97252190e-01
-1.47323406e+00 5.23731589e-01 1.05352080e+00 8.04740369e-01
2.70109594e-01 -8.14977214e-02 -2.50070930e-01 2.01991707e-01
1.42784223e-01 3.28988105e-01 1.50593773e-01 -1.22747886e+00
4.45160747e-01 2.12452114e-01 2.80621290e-01 -1.14119792e+00
-1.10950572e-02 -5.23504496e-01 -4.11809385e-01 3.20690632e-01
8.30403268e-01 1.33202657e-01 2.12692954e-02 1.37907231e+00
-1.67128146e-01 1.49150208e-01 -3.23914707e-01 6.03549242e-01
5.16542017e-01 -1.71530172e-01 6.92541897e-02 -5.17001413e-02
1.52002752e+00 -6.91286087e-01 -8.08117092e-01 2.00986803e-01
3.98071140e-01 -9.35692251e-01 1.07543957e+00 8.58423471e-01
-1.10318053e+00 -1.78071111e-01 -9.55926239e-01 -1.43189684e-01
-4.26621974e-01 -4.72847838e-03 5.43321073e-01 1.26279533e+00
-8.34668219e-01 1.06475496e+00 -3.63267630e-01 -3.68100852e-01
6.30932868e-01 7.87598118e-02 -4.33285236e-01 4.68746632e-01
-1.06781769e+00 8.13535392e-01 8.82486627e-02 -4.04371679e-01
-9.62671638e-01 -6.58406794e-01 -7.90927857e-02 -5.28575443e-02
4.95745391e-01 -3.12007546e-01 1.20201743e+00 -5.55916727e-01
-1.37511921e+00 1.45866573e+00 1.62699178e-01 -6.83639765e-01
1.20474422e+00 -4.34608966e-01 -3.77099156e-01 1.80773303e-01
2.07355917e-01 -1.33923545e-01 8.23474944e-01 -1.31389868e+00
-4.16201085e-01 -1.45829409e-01 -4.03196752e-01 -7.34729826e-01
-4.81264323e-01 9.11346555e-01 -8.82430524e-02 -9.75328207e-01
-6.91232085e-01 -5.46490908e-01 2.45466754e-01 -1.51986048e-01
-6.04359508e-01 -2.71586001e-01 6.74042165e-01 -7.76258230e-01
1.58131897e+00 -2.00489140e+00 -9.35986862e-02 1.63904667e-01
6.03383660e-01 4.88215774e-01 8.56183916e-02 7.30859816e-01
1.47661969e-01 7.44696498e-01 -1.62546396e-01 -9.22687888e-01
5.48495233e-01 -1.92674503e-01 -6.28625572e-01 1.00395095e+00
-2.51286328e-01 8.72919440e-01 -1.05494213e+00 -6.70417249e-01
-1.06562547e-01 1.46785766e-01 -2.29684696e-01 1.35036424e-01
-1.10781891e-02 1.13319613e-01 -3.24202627e-01 7.35543013e-01
5.86138248e-01 -3.45208287e-01 3.10101241e-01 1.14452451e-01
-2.51152784e-01 2.66865671e-01 -1.23584545e+00 1.35070121e+00
2.30565339e-01 1.03833568e+00 2.05187321e-01 -4.61480290e-01
8.91166747e-01 3.51425081e-01 2.84239113e-01 -5.68155110e-01
4.59900588e-01 7.67692685e-01 -7.27630183e-02 -4.22489613e-01
8.73705566e-01 1.98366806e-01 -1.12588108e-01 1.12005448e+00
1.23125851e-01 4.89710808e-01 2.12111875e-01 3.35841835e-01
1.28760886e+00 2.07834348e-01 1.81468502e-01 -2.97178537e-01
6.64403379e-01 -2.19529673e-01 4.10984993e-01 1.22575200e+00
-6.71264052e-01 5.87618113e-01 8.58570099e-01 -2.15723857e-01
-1.35352921e+00 -6.65225804e-01 -1.48324743e-01 9.57943916e-01
-8.34819302e-02 -6.94123387e-01 -1.02712345e+00 -9.54167247e-01
1.43004864e-01 6.30342245e-01 -7.66344130e-01 4.89195436e-01
-8.37622464e-01 -5.43088257e-01 1.72267699e+00 2.78370708e-01
2.98866421e-01 -1.20683801e+00 -8.94561589e-01 -4.45082895e-02
-1.88597158e-01 -1.05086374e+00 -9.79034454e-02 -1.99750632e-01
-6.11723900e-01 -1.65490317e+00 -7.19125807e-01 -1.17486157e-01
1.04022563e-01 3.37155424e-02 1.33805466e+00 5.57667077e-01
-3.39257032e-01 2.17781916e-01 -4.27271515e-01 -3.06674421e-01
-7.02515781e-01 2.38643020e-01 -1.66299328e-01 -6.83380961e-02
3.81884962e-01 -4.56297249e-01 -3.11553091e-01 3.12221318e-01
-7.22169280e-01 -7.39284039e-01 4.61137772e-01 7.15644896e-01
8.26586504e-03 -3.57486248e-01 3.37795407e-01 -1.22578323e+00
9.25846934e-01 -4.43952888e-01 -4.52979445e-01 2.82694817e-01
-1.09784532e+00 -2.40267068e-01 5.46515882e-01 -4.07874845e-02
-7.16682732e-01 -8.07664454e-01 6.76863790e-02 -4.43753511e-01
-7.96905980e-02 -5.93029372e-02 -9.64501351e-02 -2.64570624e-01
8.69482636e-01 -1.95934959e-02 1.39181633e-02 -1.04456675e+00
1.56154722e-01 7.09137142e-01 6.95136964e-01 -9.81315374e-01
8.69131863e-01 6.23263180e-01 -5.93534410e-02 -3.01765233e-01
-3.28815430e-01 -1.71014220e-01 -5.66743910e-01 2.54152939e-02
2.95154423e-01 -3.35498214e-01 -1.13932574e+00 6.87024415e-01
-1.53127253e+00 -1.41994059e-01 -3.28343421e-01 -3.93503010e-02
-3.74677867e-01 1.15001106e+00 -8.27710032e-01 -9.36478853e-01
-6.07821465e-01 -9.85714734e-01 8.67909372e-01 -2.10823223e-01
-6.66642010e-01 -9.09137726e-01 3.62627327e-01 5.80463588e-01
4.32310849e-01 5.32777905e-01 4.40854818e-01 -1.21805108e+00
-2.64260858e-01 -4.29218858e-01 -4.30590719e-01 3.13323528e-01
-5.32480121e-01 5.78381836e-01 -1.18695474e+00 -3.86584342e-01
-2.95403779e-01 -2.07434267e-01 6.12992465e-01 -3.79543424e-01
1.10034215e+00 -3.12437356e-01 -3.37861270e-01 4.54069942e-01
1.23630607e+00 -4.65883374e-01 7.35859752e-01 8.99401665e-01
4.24039066e-01 5.57428896e-01 3.55653107e-01 7.12727129e-01
1.86972052e-01 7.68731296e-01 3.08405012e-01 4.35451359e-01
-3.81047070e-01 -2.76972473e-01 2.40826234e-01 4.12470490e-01
-4.61949050e-01 -7.50626922e-01 -1.03294289e+00 4.54916477e-01
-1.79351699e+00 -1.35278153e+00 -7.40681469e-01 2.18030095e+00
5.33912122e-01 1.71843067e-01 4.44105566e-01 4.83749479e-01
9.08889115e-01 2.60157883e-01 1.38793871e-01 -5.56522906e-01
-5.44729590e-01 3.95070583e-01 5.74635446e-01 2.31991574e-01
-8.10421646e-01 8.27529073e-01 7.31482220e+00 9.75614965e-01
-8.10906351e-01 5.71939468e-01 6.90484047e-02 8.78369510e-02
-3.42479169e-01 2.36026764e-01 -8.17417681e-01 9.94636118e-01
1.00765312e+00 -2.86901176e-01 1.34190246e-01 6.80040181e-01
-1.68576255e-01 2.24650130e-01 -1.00321102e+00 9.41012204e-01
3.80803317e-01 -1.49676907e+00 -1.08357735e-01 5.65308392e-01
2.55466461e-01 -4.85768169e-03 -1.32477526e-02 -2.21979953e-02
4.97384459e-01 -1.06233943e+00 1.22330606e+00 5.60866356e-01
6.15775764e-01 -3.58612239e-01 9.87609982e-01 1.58861563e-01
-3.06807309e-01 -3.38666067e-02 -2.67550442e-02 4.38768929e-03
2.28229657e-01 5.90012968e-01 -2.69426644e-01 8.26018572e-01
7.90700912e-01 5.41936100e-01 -9.82151747e-01 1.08666599e+00
-4.80234355e-01 8.99100721e-01 -1.00757614e-01 -6.05185702e-02
1.03321369e-03 1.02999114e-01 9.27412212e-01 1.60533857e+00
2.23805934e-01 -2.68084824e-01 -4.94600326e-01 8.62627089e-01
-5.42966068e-01 1.37677655e-01 -3.94347787e-01 -1.85402498e-01
7.42578983e-01 1.22945142e+00 -8.32917452e-01 -1.73005611e-01
2.74534281e-02 1.03490376e+00 3.83663952e-01 -1.74169689e-01
-8.65549326e-01 -3.98672491e-01 2.20510244e-01 2.72098094e-01
1.08700976e-01 -9.75065529e-02 -5.72743535e-01 -1.12252355e+00
1.03733748e-01 -1.14480197e+00 6.87234282e-01 -4.51442003e-01
-1.50198328e+00 4.95496303e-01 -3.23095977e-01 -9.34210777e-01
3.86427492e-02 -5.94022512e-01 -6.60995483e-01 6.37341261e-01
-1.59172285e+00 -1.15379012e+00 -1.77782282e-01 3.89796317e-01
-1.62510633e-01 -5.50241709e-01 6.17223620e-01 6.11809015e-01
-7.82722235e-01 1.28434944e+00 1.40239000e-01 3.40360969e-01
1.29597914e+00 -1.23652911e+00 5.67159116e-01 1.06157541e+00
5.88712282e-02 8.86869848e-01 9.57327366e-01 -7.88156569e-01
-1.28153229e+00 -4.64843482e-01 1.36548257e+00 -1.21663082e+00
1.12121272e+00 -2.83354521e-01 -9.77675259e-01 4.72004920e-01
4.64151919e-01 -7.85229281e-02 7.82233775e-01 1.76316142e-01
-9.67496216e-01 1.91204429e-01 -1.33636653e+00 2.38943115e-01
1.13055313e+00 -5.80686092e-01 -6.65378094e-01 2.28474036e-01
8.97411350e-03 -7.70314857e-02 -8.05339634e-01 -1.45607024e-01
7.56808102e-01 -1.50785553e+00 7.80154705e-01 -6.86218798e-01
5.73722601e-01 2.58329250e-02 -4.95568197e-03 -4.10787225e-01
-8.23891610e-02 -1.23454726e+00 -3.11622769e-01 1.90638089e+00
-1.12290911e-01 -7.89867401e-01 9.44852173e-01 2.35301629e-01
7.11010098e-02 -2.47250557e-01 -1.13218093e+00 -1.03885615e+00
4.54774857e-01 -3.75077635e-01 6.82235718e-01 1.24645901e+00
-1.30518656e-02 -1.14870220e-01 -5.57907283e-01 -2.29905054e-01
1.18528640e+00 -7.82050416e-02 1.02310789e+00 -1.51278770e+00
-3.67388368e-01 -9.60018337e-01 -3.42113972e-01 -2.82455951e-01
3.15717191e-01 -7.72153080e-01 -4.69835490e-01 -9.45197821e-01
5.24941802e-01 -5.27085006e-01 -2.26908967e-01 4.32726115e-01
-1.57388970e-01 5.23843646e-01 3.97717923e-01 8.15313458e-01
-8.02406609e-01 -1.16659664e-01 7.90744185e-01 1.77810371e-01
5.69785118e-01 -3.32406461e-01 -8.50485682e-01 4.55952466e-01
5.10370076e-01 -7.55822122e-01 4.37005609e-01 -3.87964696e-01
6.50246561e-01 -4.23646271e-01 8.58943343e-01 -6.76319718e-01
3.89801800e-01 2.11202234e-01 -1.27763808e-01 -3.17543119e-01
-1.67197138e-01 -4.84337837e-01 1.13263428e-01 5.41861415e-01
-1.93254799e-01 1.86350569e-01 -1.68222830e-01 4.09199983e-01
1.41960261e-02 -5.25322020e-01 6.52028799e-01 -1.20362118e-01
-4.34722155e-01 2.63453685e-02 -6.70026168e-02 3.46906573e-01
8.57439637e-01 -5.48731208e-01 -9.78939176e-01 4.84532043e-02
-3.64154242e-02 -6.03695437e-02 1.06881857e+00 2.87430465e-01
-2.36357581e-02 -6.80091202e-01 -9.62633371e-01 -3.91673595e-01
3.49062830e-02 -9.56421435e-01 3.54735069e-02 9.39965367e-01
-7.75318325e-01 2.61129260e-01 -3.92372221e-01 -1.50796315e-02
-1.35834897e+00 5.58388710e-01 1.85294107e-01 -4.79760230e-01
-6.52861655e-01 5.36347330e-01 -8.35196018e-01 -1.05877273e-01
3.69443119e-01 5.93449235e-01 -3.27930450e-02 2.52884239e-01
6.28995717e-01 1.20128620e+00 2.09803149e-01 -6.63003922e-01
-6.36210799e-01 1.84600189e-01 2.77498714e-03 -1.95344523e-01
1.35423601e+00 7.22558051e-02 -3.93359274e-01 3.20770919e-01
8.32515121e-01 7.60919094e-01 -6.97319269e-01 -6.52271360e-02
5.39780915e-01 -8.14752758e-01 -3.08476359e-01 -9.82328951e-01
-7.99606979e-01 1.09374142e+00 3.70989799e-01 4.27194715e-01
3.64691615e-01 -1.88337043e-01 7.83696234e-01 9.01363641e-02
5.04314423e-01 -1.17778802e+00 1.09589569e-01 2.50261486e-01
5.55911064e-01 -9.30780232e-01 4.84339595e-01 -2.59666532e-01
-4.29287523e-01 1.30646181e+00 1.87918067e-01 -2.83837229e-01
1.65589616e-01 3.72582436e-01 -9.39648822e-02 -5.16947329e-01
-3.33968371e-01 2.34124199e-01 -2.84706682e-01 4.75422800e-01
5.30266702e-01 -2.05520213e-01 -8.57759774e-01 9.58673120e-01
-1.70372158e-01 1.83162093e-01 6.86593831e-01 8.53726745e-01
-8.09233114e-02 -1.60458100e+00 -6.69956326e-01 7.73953274e-02
-1.06409454e+00 1.35582924e-01 -1.01181197e+00 1.12057936e+00
3.11788350e-01 7.47927368e-01 -3.76124233e-01 -2.97480553e-01
3.32259804e-01 -1.56658061e-03 5.20432234e-01 -1.87367857e-01
-1.57283604e+00 -5.11856556e-01 3.42434913e-01 -4.37724024e-01
-1.59904569e-01 -1.07862341e+00 -6.02715313e-01 -1.15412474e+00
-3.01332176e-01 4.26587969e-01 5.49678266e-01 8.07158470e-01
4.96916234e-01 1.75598636e-01 2.92025924e-01 -5.79786599e-01
-7.09309816e-01 -7.74053335e-01 -5.33686161e-01 6.63124025e-01
5.91029832e-03 -6.40961707e-01 -6.46441102e-01 -1.74294189e-01] | [9.532045364379883, 10.594413757324219] |
42e6a7bb-fe6b-4374-96e2-83f87ba06e47 | evaluating-machine-translation-quality-with | 2306.01549 | null | https://arxiv.org/abs/2306.01549v1 | https://arxiv.org/pdf/2306.01549v1.pdf | Evaluating Machine Translation Quality with Conformal Predictive Distributions | This paper presents a new approach for assessing uncertainty in machine translation by simultaneously evaluating translation quality and providing a reliable confidence score. Our approach utilizes conformal predictive distributions to produce prediction intervals with guaranteed coverage, meaning that for any given significance level $\epsilon$, we can expect the true quality score of a translation to fall out of the interval at a rate of $1-\epsilon$. In this paper, we demonstrate how our method outperforms a simple, but effective baseline on six different language pairs in terms of coverage and sharpness. Furthermore, we validate that our approach requires the data exchangeability assumption to hold for optimal performance. | ['Patrizio Giovannotti'] | 2023-06-02 | null | null | null | null | ['prediction-intervals'] | ['miscellaneous'] | [ 5.07308170e-02 3.98681283e-01 -5.71114182e-01 -4.56019819e-01
-1.68003368e+00 -8.96519005e-01 4.83300656e-01 5.07854223e-01
-7.54626319e-02 1.43418574e+00 -4.41299528e-02 -7.96217144e-01
-1.87083289e-01 -7.28423953e-01 -9.61602032e-01 -1.27466917e-01
-7.54892603e-02 8.24432433e-01 5.30690849e-02 1.36415005e-01
3.02299023e-01 1.67678624e-01 -9.73473728e-01 9.65521261e-02
1.42445934e+00 9.19476986e-01 -1.76772028e-01 6.02632642e-01
2.87445396e-01 1.14397123e-01 -7.92553484e-01 -9.59411085e-01
3.35026681e-01 -5.19189596e-01 -6.40377343e-01 -5.29019535e-01
7.08339512e-02 -1.09810077e-01 4.19983715e-01 1.38686585e+00
1.69854552e-01 -3.07153851e-01 1.08831465e+00 -1.29833889e+00
-5.53577006e-01 8.22686970e-01 -3.83213103e-01 2.64381263e-02
5.33340752e-01 -3.04501593e-01 1.28931975e+00 -9.16327655e-01
5.48597157e-01 8.05767417e-01 8.63691390e-01 1.62716527e-02
-1.42797005e+00 -6.92454517e-01 -3.21276069e-01 -5.04921973e-01
-1.51074135e+00 -4.85235989e-01 8.15204754e-02 -3.72364134e-01
8.34217310e-01 2.62993336e-01 2.69213915e-01 6.71932995e-01
9.19202864e-01 2.42771462e-01 1.22610319e+00 -7.50763178e-01
3.63178164e-01 4.77844864e-01 -1.75057977e-01 3.97047877e-01
5.58783948e-01 3.46616685e-01 -5.58025002e-01 -5.80082953e-01
6.23828888e-01 -5.47369480e-01 -1.75144494e-01 -6.03335425e-02
-1.29889703e+00 8.49699318e-01 -8.52713212e-02 1.47483706e-01
-1.43968329e-01 2.54926056e-01 8.55712593e-02 4.68493879e-01
8.72594416e-01 4.71690804e-01 -5.48465073e-01 -5.56007147e-01
-1.07200027e+00 5.98685928e-02 9.00669932e-01 1.46793234e+00
3.23981434e-01 -4.45965081e-01 -4.54676956e-01 4.07339364e-01
2.68916994e-01 8.25708330e-01 -5.87473437e-02 -8.29082012e-01
7.09087372e-01 1.90225109e-01 7.00594485e-01 -6.91799879e-01
6.48427308e-02 -4.73292351e-01 -4.27400827e-01 -1.21583432e-01
4.67907161e-01 -3.13299268e-01 -6.26264274e-01 1.77603853e+00
-1.20144747e-01 -2.15275586e-01 1.26295745e-01 6.28350556e-01
-6.21017516e-02 7.94568121e-01 -2.61386216e-01 -7.23195672e-01
1.19805062e+00 -3.41559917e-01 -6.80534422e-01 1.01488799e-01
5.13200819e-01 -1.06509399e+00 1.08431196e+00 3.83619905e-01
-1.23713255e+00 -2.07468107e-01 -1.09542871e+00 2.50543833e-01
2.31369361e-01 5.66132143e-02 5.29702008e-01 9.29325879e-01
-1.01784766e+00 6.65423274e-01 -6.50651097e-01 -1.03041413e-04
1.19293056e-01 3.39127034e-01 -1.00704484e-01 7.49545470e-02
-1.25599253e+00 1.00361061e+00 2.67577499e-01 -2.57282227e-01
-5.14987171e-01 -5.98942041e-01 -5.22612453e-01 8.64321515e-02
-1.06955739e-02 -3.37954462e-01 1.27719963e+00 -7.15981245e-01
-1.21315920e+00 3.87070626e-01 -2.64268041e-01 -5.47088742e-01
5.96744537e-01 -1.01950549e-01 -3.68098259e-01 -8.66328329e-02
3.47143292e-01 3.02159190e-01 2.95978546e-01 -1.07501245e+00
-5.89101076e-01 -1.94601804e-01 -1.69787988e-01 1.01978913e-01
2.33187480e-03 2.66589493e-01 -5.75111568e-01 -6.19672477e-01
1.62549749e-01 -9.51226532e-01 1.32075362e-02 -4.45166856e-01
-2.40025729e-01 -1.66532502e-01 -1.92324251e-01 -8.40379715e-01
1.38482499e+00 -1.74785089e+00 4.48810384e-02 6.27858698e-01
-2.49980927e-01 -3.63241464e-01 6.35873452e-02 4.68960077e-01
2.93911695e-01 4.68230605e-01 -3.85760009e-01 -1.01198204e-01
1.15067743e-01 6.46248227e-03 -4.43448424e-01 6.56739771e-01
3.11663806e-01 7.84191549e-01 -8.70506525e-01 -5.49460113e-01
-3.80355090e-01 2.43071750e-01 -4.62324470e-01 -2.98298579e-02
-3.99285406e-01 -3.80503386e-02 -1.15391023e-01 6.68726265e-01
5.68020344e-01 -1.29489258e-01 4.79312271e-01 4.22886431e-01
-6.18223175e-02 4.68986809e-01 -9.48777914e-01 1.39510965e+00
-4.32953060e-01 5.97025394e-01 -2.70514250e-01 -4.93416458e-01
1.18651390e+00 3.19905579e-01 1.00700833e-01 -4.41301018e-01
-2.01782454e-02 5.83287954e-01 -1.60022691e-01 2.47268155e-01
5.52198708e-01 -5.08024096e-01 -2.50654727e-01 6.75050914e-01
-1.71723366e-01 -3.69997799e-01 -1.25515848e-01 -7.12943226e-02
7.37682641e-01 4.58850227e-02 1.85887337e-01 -8.13356698e-01
-4.91359308e-02 6.63502738e-02 8.79162550e-01 7.00550139e-01
-2.10331053e-01 5.98731458e-01 9.58390355e-01 1.11807212e-01
-1.40062380e+00 -1.36396122e+00 -5.25663912e-01 6.66956961e-01
1.30868509e-01 -3.35284531e-01 -8.43150258e-01 -9.10263419e-01
-3.80264148e-02 1.06180608e+00 -4.97328192e-01 1.79536656e-01
-8.31900015e-02 -7.88432002e-01 5.89513659e-01 4.44042236e-01
1.92598447e-01 -4.28208381e-01 -4.27797318e-01 -4.19947989e-02
-4.44920152e-01 -6.66532695e-01 -6.66765273e-01 1.30758375e-01
-8.73724878e-01 -6.67523324e-01 -6.96345985e-01 -6.78183019e-01
5.98569691e-01 -1.86389402e-01 1.38915718e+00 -2.63837904e-01
4.36740577e-01 -1.23494536e-01 -2.63365030e-01 -4.44287598e-01
-7.24339008e-01 -7.33671710e-02 1.19206518e-01 -7.16427743e-01
4.38250065e-01 -2.21183598e-01 -3.75469029e-01 4.49188232e-01
-5.98668635e-01 -2.78416455e-01 4.87169683e-01 1.00366020e+00
7.92127907e-01 1.92402452e-01 7.99785137e-01 -6.63914800e-01
9.72856760e-01 -4.56574738e-01 -9.32880759e-01 7.10421920e-01
-1.21945906e+00 2.56775409e-01 4.18310970e-01 -6.96084648e-02
-7.81592309e-01 -3.02754164e-01 1.12491742e-01 -1.01539016e-01
3.51038128e-01 7.84436405e-01 1.36532322e-01 2.88634717e-01
7.69263446e-01 -1.10663638e-01 5.39116599e-02 -1.50271952e-01
2.35252693e-01 9.11150396e-01 3.31757367e-01 -8.03167224e-01
5.30142725e-01 4.19540405e-02 -2.88069695e-02 -1.28440589e-01
-3.25279176e-01 -7.83968791e-02 -3.36722553e-01 6.20027352e-03
3.59405309e-01 -9.16915953e-01 -4.89009231e-01 -2.19509244e-01
-1.14672148e+00 -6.56578094e-02 -1.52734771e-01 8.18943024e-01
-8.70307148e-01 2.21888915e-01 -4.68307793e-01 -1.11547029e+00
-3.04261088e-01 -1.17353129e+00 1.09247661e+00 -1.26023278e-01
-4.04595107e-01 -9.30773258e-01 9.84063148e-02 8.80373418e-02
2.63241947e-01 2.02060550e-01 1.04592919e+00 -7.06085682e-01
-2.98682928e-01 -4.13978547e-01 -3.07049632e-01 1.73752427e-01
-1.26540428e-03 2.18039185e-01 -4.99013305e-01 -3.22776765e-01
-1.77522734e-01 -1.66392773e-01 4.17760313e-01 6.27152681e-01
6.85400605e-01 -4.73493785e-01 -3.06039363e-01 5.11901192e-02
1.33165383e+00 3.08620602e-01 6.46353424e-01 8.34790841e-02
-2.90725470e-01 5.24633169e-01 9.95869040e-01 4.07267451e-01
1.68180585e-01 7.65907049e-01 -6.68108836e-02 2.51576781e-01
3.10772657e-01 -4.35983896e-01 3.19189638e-01 6.94577754e-01
-9.64128692e-03 -3.98152053e-01 -9.33743179e-01 6.86864853e-01
-1.76836765e+00 -7.13431060e-01 1.48227841e-01 2.86910748e+00
1.14532447e+00 5.19373775e-01 1.00979842e-01 -2.38001838e-01
7.25851774e-01 -5.41448712e-01 -2.29257479e-01 -9.00074720e-01
-8.43303055e-02 1.76163271e-01 7.16772258e-01 8.78485680e-01
-5.85500836e-01 7.06267059e-01 8.05577183e+00 7.70154893e-01
-6.46678090e-01 6.19035475e-02 1.01059663e+00 1.07434161e-01
-9.31184351e-01 1.52798370e-01 -7.30162561e-01 5.18040657e-01
1.28214693e+00 -6.96469843e-01 1.87688380e-01 6.19272828e-01
1.52635515e-01 -3.31165165e-01 -1.05122900e+00 4.40074593e-01
-7.88790286e-02 -1.39295673e+00 -2.14831948e-01 2.50018239e-01
9.23213542e-01 -1.18049003e-01 1.76277950e-01 1.08207561e-01
5.42947471e-01 -1.12730753e+00 6.92236602e-01 5.47437847e-01
1.25953650e+00 -1.23200059e+00 1.13632011e+00 4.17206168e-01
-6.14498556e-01 2.20861375e-01 -4.36987609e-01 -5.36553413e-02
1.22674696e-01 1.12384093e+00 -1.13011837e+00 8.53749633e-01
4.50441211e-01 1.21927470e-01 -5.04321977e-02 8.03497791e-01
-1.90378174e-01 6.12949967e-01 -3.27205271e-01 -1.89447477e-01
6.07191212e-03 -3.49691153e-01 2.19847202e-01 1.21397984e+00
9.00528073e-01 5.59778139e-02 -8.22087526e-02 8.75281453e-01
-2.52168834e-01 1.25458181e-01 -6.71421528e-01 -2.66801029e-01
9.73777711e-01 3.47955942e-01 -6.00467205e-01 -2.11671770e-01
-2.21920118e-01 8.06615829e-01 2.25764886e-01 1.62690803e-01
-8.61526668e-01 -5.02540290e-01 4.54623967e-01 -2.67464429e-01
1.54351324e-01 -1.78991824e-01 -9.44283009e-01 -1.07404327e+00
3.02120715e-01 -9.26828146e-01 2.97167838e-01 -4.88150209e-01
-1.15193427e+00 6.66986883e-01 1.30364999e-01 -1.30572450e+00
-6.37677610e-01 -4.42933410e-01 -2.65269995e-01 1.39464378e+00
-9.68254030e-01 -6.10330820e-01 6.04699194e-01 -9.63093713e-02
1.16620243e-01 -1.18921213e-01 9.99652863e-01 -2.14489967e-01
-1.92771062e-01 1.04302478e+00 4.21514899e-01 -2.42788061e-01
7.98223078e-01 -1.17989576e+00 3.77931923e-01 1.07215989e+00
1.04743607e-01 7.19590604e-01 1.10600460e+00 -9.89815533e-01
-1.17133665e+00 -8.18811893e-01 1.55453002e+00 -6.51629269e-01
5.68274856e-01 -2.46618167e-01 -5.14793694e-01 7.32253671e-01
1.15999982e-01 -4.00718987e-01 7.20914185e-01 5.95263541e-01
-3.56499881e-01 -1.80469994e-02 -1.56027007e+00 4.37520981e-01
7.87548721e-01 -4.91831601e-01 -6.87953949e-01 5.35391688e-01
9.26072836e-01 -4.00749177e-01 -1.35854578e+00 6.60950899e-01
6.69784784e-01 -8.90325129e-01 4.08968508e-01 -5.53122938e-01
5.99092543e-01 -2.47878730e-01 -6.65022075e-01 -1.41435432e+00
-7.43106082e-02 -6.80219829e-01 1.73499882e-01 1.05611598e+00
1.13086236e+00 -6.47071540e-01 5.39156854e-01 8.58977258e-01
6.00976795e-02 -8.18787277e-01 -1.32716119e+00 -1.08403778e+00
6.30166888e-01 -7.70872891e-01 7.79704213e-01 6.35688841e-01
5.79100370e-01 8.49740207e-02 -3.40756565e-01 2.93509841e-01
6.18672371e-01 3.45090061e-01 3.18561554e-01 -8.23870540e-01
-5.24090469e-01 -3.42438161e-01 -4.34642136e-01 -8.09548438e-01
-7.21158041e-03 -6.68736756e-01 2.33497292e-01 -1.43306077e+00
4.84245241e-01 -7.52090454e-01 -2.83264965e-01 4.85667214e-02
-1.97172448e-01 3.92441601e-01 -1.75674990e-01 1.49621099e-01
-2.73860484e-01 2.38161832e-01 7.66825676e-01 1.00001432e-01
-6.62878826e-02 2.15391874e-01 -9.55608785e-01 3.35611492e-01
7.89453030e-01 -3.48509073e-01 -4.31834787e-01 -3.68568212e-01
7.01532364e-01 5.37876964e-01 -8.08269680e-02 -4.80770737e-01
-1.38729975e-01 -2.85102308e-01 3.08633238e-01 -6.65713131e-01
-6.25934601e-02 -5.43248057e-01 4.42284226e-01 4.15818065e-01
-5.15460432e-01 5.47269702e-01 2.70771116e-01 5.53113282e-01
-9.39341187e-02 -2.46544763e-01 3.73672485e-01 3.67689937e-01
9.32441279e-02 -9.08218250e-02 1.05109578e-02 3.41979750e-02
9.58397686e-01 1.53607115e-01 -3.08672786e-01 -5.17472684e-01
-3.41739148e-01 2.33316272e-01 8.05016398e-01 1.90823853e-01
4.62489575e-01 -1.43076932e+00 -1.04581594e+00 7.96901807e-02
3.88072073e-01 -5.08287847e-01 -3.89179736e-01 7.16811895e-01
-4.05921936e-01 6.14426434e-01 1.55818209e-01 -6.82381094e-01
-8.87461901e-01 3.42182875e-01 7.06309825e-02 -1.86653614e-01
-1.37311488e-01 7.70641923e-01 -3.16873819e-01 -3.75012696e-01
1.85194924e-01 -4.57440764e-01 5.04631519e-01 -3.27998400e-01
4.79952395e-01 2.55775928e-01 2.35946998e-01 -2.80749321e-01
-4.73129541e-01 2.07598671e-01 2.12585419e-01 -8.54120433e-01
7.81329215e-01 -9.44161937e-02 -1.15048587e-01 5.17403543e-01
9.20042932e-01 5.35557747e-01 -1.11751616e+00 5.36673889e-02
1.92232668e-01 -6.94160998e-01 -1.33887708e-01 -1.35336721e+00
-2.71540076e-01 4.08151448e-01 3.81505311e-01 1.30801767e-01
9.24533665e-01 5.26683964e-02 3.14095825e-01 5.02813701e-03
7.07791209e-01 -1.11991346e+00 -5.50226629e-01 2.15870216e-01
8.22079778e-01 -1.35131133e+00 8.35698098e-02 -4.67774510e-01
-7.63480246e-01 8.49874973e-01 -6.41068518e-02 1.83885351e-01
5.16083121e-01 4.86832649e-01 -1.42851934e-01 3.89397711e-01
-9.39971447e-01 3.28319401e-01 4.27915961e-01 5.63631833e-01
8.69291782e-01 6.25261068e-01 -9.68172014e-01 6.12676561e-01
-3.45087826e-01 1.04646273e-01 3.81767124e-01 6.63165271e-01
-5.58343887e-01 -1.07872677e+00 -3.79900813e-01 5.44736445e-01
-4.86556232e-01 -4.07943457e-01 -3.57323378e-01 7.28476226e-01
-1.70346588e-01 1.23410606e+00 2.79442072e-01 -4.47090805e-01
1.45496219e-01 1.76706627e-01 6.81776524e-01 -3.64518672e-01
-1.16194323e-01 2.08969697e-01 5.65551639e-01 -2.14616016e-01
-1.66260481e-01 -8.53299320e-01 -1.15156281e+00 -6.68514073e-01
-4.99145806e-01 7.02953398e-01 6.48511112e-01 9.21977878e-01
4.18987900e-01 -8.61608312e-02 8.15569460e-01 6.42242059e-02
-1.01489055e+00 -1.01858771e+00 -4.42432404e-01 -2.23693058e-01
6.99432120e-02 -5.63196838e-01 -3.12404305e-01 -1.95088744e-01] | [7.920725345611572, 4.493966102600098] |
bb4166bf-3a1d-484b-abb6-e60bc0ebf5f7 | perceiving-music-quality-with-gans | 2006.06287 | null | https://arxiv.org/abs/2006.06287v2 | https://arxiv.org/pdf/2006.06287v2.pdf | Perceiving Music Quality with GANs | Several methods have been developed to assess the perceptual quality of audio under transforms like lossy compression. However, they require paired reference signals of the unaltered content, limiting their use in applications where references are unavailable. This has hindered progress in audio generation and style transfer, where a no-reference quality assessment method would allow more reproducible comparisons across methods. We propose training a GAN on a large music library, and using its discriminator as a no-reference quality assessment measure of the perceived quality of music. This method is unsupervised, needs no access to degraded material and can be tuned for various domains of music. In a listening test with 448 human subjects, where participants rated professionally produced music tracks degraded with different levels and types of signal degradations such as waveshaping distortion and low-pass filtering, we establish a dataset of human rated material. By using the human rated dataset we show that the discriminator score correlates significantly with the subjective ratings, suggesting that the proposed method can be used to create a no-reference musical audio quality assessment measure. | ['Carl Thomé', 'Agrin Hilmkil', 'Anders Arpteg'] | 2020-06-11 | null | null | null | null | ['audio-generation'] | ['audio'] | [ 4.16236192e-01 -3.04367065e-01 1.02361694e-01 -1.73651323e-01
-1.23774385e+00 -8.40707123e-01 1.96438447e-01 -3.89433801e-02
-1.91535980e-01 6.44233644e-01 6.34032607e-01 2.21207812e-01
-2.53174037e-01 -6.50095344e-01 -3.60119462e-01 -6.67781770e-01
-6.55369163e-02 7.59544969e-02 1.45302907e-01 -2.26866499e-01
1.31639689e-01 9.20860693e-02 -1.81884277e+00 4.57606494e-01
7.45786428e-01 1.18607962e+00 2.11309567e-01 9.82905865e-01
4.51010793e-01 3.59635264e-01 -1.33179832e+00 -4.50065821e-01
5.17993927e-01 -1.02760625e+00 -4.85492259e-01 -1.79480612e-01
4.62171495e-01 -2.14137048e-01 -1.65477782e-01 1.11742556e+00
1.08505118e+00 3.19566876e-01 7.22753763e-01 -9.93137479e-01
-3.96830648e-01 6.84034407e-01 -4.65900339e-02 8.76946896e-02
7.84712374e-01 7.61353597e-02 1.20615101e+00 -5.62678516e-01
4.43917632e-01 1.02402341e+00 8.35632920e-01 4.80674863e-01
-1.39580035e+00 -8.09976637e-01 -5.84122658e-01 2.45364517e-01
-1.21179855e+00 -6.39204800e-01 9.74148452e-01 -4.10090119e-01
4.89358634e-01 4.87859368e-01 6.98546648e-01 1.00821054e+00
1.21470168e-02 3.38620603e-01 1.08518231e+00 -5.10318577e-01
3.61940563e-01 -2.80308351e-02 -6.10732257e-01 1.79242045e-02
-1.37794018e-01 4.55568165e-01 -8.71077538e-01 -1.04312733e-01
7.18904316e-01 -6.88283145e-01 -5.82264662e-01 -1.06082775e-01
-1.20919132e+00 5.26942670e-01 3.58357757e-01 5.28542817e-01
-1.93337515e-01 3.57855633e-02 4.57508773e-01 9.00034308e-01
3.19253981e-01 9.40657794e-01 -2.18777657e-01 -5.21191180e-01
-1.33346486e+00 2.16865897e-01 6.89106047e-01 5.63623965e-01
1.77830547e-01 3.95259917e-01 -3.10218096e-01 1.28081381e+00
-4.47396822e-02 4.52753872e-01 7.88707852e-01 -1.36470163e+00
2.94330060e-01 -8.63018110e-02 2.89206535e-01 -1.09511912e+00
-5.55236638e-02 -6.55297458e-01 -8.14251065e-01 6.37908459e-01
4.59834009e-01 -1.11096509e-01 -5.29974639e-01 1.75898767e+00
-1.28179982e-01 4.06897627e-02 -2.86007370e-03 1.03934693e+00
4.91369873e-01 5.92920959e-01 -4.02967125e-01 -3.32160711e-01
8.98943901e-01 -6.58910215e-01 -8.02412748e-01 2.23262131e-01
-4.19384474e-03 -1.20656371e+00 1.48858273e+00 1.10404277e+00
-1.42842197e+00 -1.05671930e+00 -1.31948781e+00 1.69805288e-01
9.38040242e-02 -1.03519209e-01 6.58441260e-02 9.51408446e-01
-1.18656087e+00 9.20790434e-01 -3.23137134e-01 6.62715659e-02
8.97744372e-02 2.56078869e-01 -1.00490212e-01 2.10720941e-01
-1.26003444e+00 6.07788205e-01 1.90158397e-01 -2.26835653e-01
-1.10823596e+00 -8.00115585e-01 -4.31405306e-01 9.65756848e-02
-1.83023781e-01 -5.83018005e-01 1.39087117e+00 -1.06935763e+00
-1.92973173e+00 5.60206711e-01 2.53406554e-01 -3.53064299e-01
5.74714899e-01 -3.71073812e-01 -8.75183702e-01 2.81050742e-01
-1.77037120e-02 4.15648758e-01 1.24728942e+00 -1.20171523e+00
-4.52849209e-01 1.01356022e-01 -3.94808091e-02 2.88196325e-01
-3.85476768e-01 1.34320594e-02 -6.69594556e-02 -1.13616574e+00
7.90745318e-02 -5.88639736e-01 2.28105575e-01 -1.78111568e-02
-1.27577916e-01 3.61346573e-01 4.49410796e-01 -8.40039253e-01
1.27226853e+00 -2.52329946e+00 1.47459954e-01 2.98103660e-01
-1.64385721e-01 2.10586652e-01 -3.99668723e-01 4.23344076e-01
1.41444197e-02 9.21862274e-02 -2.09207818e-01 -4.26246449e-02
1.62167326e-01 -2.30175853e-01 -3.20634693e-01 1.49021134e-01
-1.13860190e-01 2.63865650e-01 -1.02397513e+00 -3.62724841e-01
5.87080158e-02 5.99511087e-01 -7.97575057e-01 4.85713899e-01
8.32524076e-02 5.28630078e-01 7.38376230e-02 3.85089308e-01
3.26391220e-01 4.13695633e-01 -1.96441233e-01 -3.57229948e-01
2.75829196e-01 5.59978604e-01 -1.31503236e+00 1.90361691e+00
-6.99431658e-01 7.77606666e-01 2.39381209e-01 -4.90465403e-01
1.27164412e+00 7.14628100e-01 3.30555528e-01 -7.65669227e-01
-3.21688727e-02 4.37612265e-01 3.18314701e-01 -2.81495005e-01
4.09693897e-01 -5.60327590e-01 6.82989731e-02 3.80324095e-01
2.82993197e-01 -7.43424118e-01 1.78949133e-01 -1.82679266e-01
1.27650356e+00 -1.19112758e-02 7.43971318e-02 -8.60155672e-02
4.05553132e-01 -5.41891336e-01 3.99633497e-01 6.34192407e-01
-1.85863689e-01 1.03000188e+00 1.66761205e-01 9.76859629e-02
-1.19648695e+00 -1.41616738e+00 -1.36683717e-01 1.17074776e+00
-1.20713800e-01 -5.14173388e-01 -7.77616024e-01 -3.26905809e-02
-2.93158770e-01 6.26468539e-01 -1.42467245e-01 -3.95930409e-01
-3.21016490e-01 -2.31879726e-01 9.01047945e-01 2.90521979e-01
3.90563190e-01 -1.29507923e+00 -4.32818204e-01 3.71991158e-01
-3.86842787e-01 -5.61089158e-01 -6.34774566e-01 1.21742122e-01
-8.62531841e-01 -7.53873825e-01 -9.39929664e-01 -7.12132037e-01
7.61855543e-02 -1.71548977e-01 1.21618021e+00 -2.05382973e-01
-9.37048346e-02 3.11499864e-01 -4.84335214e-01 -2.14265794e-01
-6.96930170e-01 -9.75897238e-02 3.61075670e-01 -4.45165336e-02
-3.53089750e-01 -1.09741473e+00 -7.75701106e-01 6.61010027e-01
-9.18406367e-01 -3.47012788e-01 4.55332726e-01 8.45727384e-01
5.07388234e-01 4.93803293e-01 9.55638528e-01 -3.61478567e-01
1.14974856e+00 -1.78933665e-02 -1.74634188e-01 -2.67638654e-01
-5.90408027e-01 -1.74552545e-01 7.61436343e-01 -6.20471239e-01
-1.00616968e+00 -4.25514698e-01 -3.88006270e-01 -3.18252593e-01
-1.37957126e-01 4.22234625e-01 -4.38991338e-01 1.14389867e-01
1.06965005e+00 -6.16619065e-02 -2.28971511e-01 -7.11778283e-01
3.39803606e-01 7.94432878e-01 1.06422126e+00 -5.03091097e-01
9.18974698e-01 -4.22266573e-02 -1.27593666e-01 -7.18296826e-01
-6.04376137e-01 -3.38748425e-01 -4.53044057e-01 -3.97971004e-01
4.60817367e-01 -7.29539752e-01 -4.14221048e-01 3.47246706e-01
-7.38134027e-01 -4.66394186e-01 -6.78041101e-01 7.88892627e-01
-1.03287756e+00 2.14425668e-01 -6.56302750e-01 -7.77574539e-01
-3.10563922e-01 -9.23118770e-01 7.62155712e-01 -6.22429103e-02
-6.39377773e-01 -7.24297941e-01 3.04419726e-01 1.98381469e-01
5.78978419e-01 2.14915976e-01 9.09695148e-01 -1.30460143e-01
-8.52764472e-02 -2.88888037e-01 3.28715235e-01 8.29896569e-01
4.98023242e-01 -2.94452161e-01 -1.20862341e+00 -4.41480637e-01
2.18902826e-01 -5.44962347e-01 4.85134423e-01 3.57873738e-01
1.05478907e+00 -3.49840969e-01 4.78790283e-01 5.92041969e-01
1.04224539e+00 3.19950670e-01 9.73850787e-01 2.39351302e-01
2.29687154e-01 3.45883936e-01 5.95355392e-01 3.89446050e-01
-5.59504867e-01 8.91567707e-01 5.99980168e-02 -4.52123620e-02
-5.69063902e-01 -4.65939581e-01 6.45828664e-01 1.38878250e+00
-3.07715416e-01 -2.30642155e-01 -4.34018701e-01 4.80754256e-01
-9.99700546e-01 -1.23984027e+00 3.90234679e-01 2.47466016e+00
1.24276435e+00 2.82826602e-01 4.63798732e-01 1.02088904e+00
7.10313439e-01 7.25004524e-02 -5.28019845e-01 -4.52727050e-01
-1.00156121e-01 7.10269928e-01 -4.20734882e-02 3.52740735e-01
-7.14559197e-01 3.38916838e-01 7.53920507e+00 8.64014268e-01
-1.12265968e+00 1.20800193e-02 2.34016165e-01 -3.30676883e-01
-2.67205834e-01 -3.07363510e-01 -4.87100482e-02 4.26456153e-01
1.11570156e+00 -3.92859221e-01 7.56293416e-01 5.69779277e-01
4.51758653e-01 1.99365467e-01 -1.08158743e+00 1.13923550e+00
4.09628171e-03 -7.15575337e-01 -1.15594409e-01 -1.16432205e-01
7.68510342e-01 -4.92564589e-01 4.56630290e-01 9.35206264e-02
-1.04093505e-02 -1.12131894e+00 9.68299270e-01 6.52673304e-01
1.19531965e+00 -8.43185306e-01 6.83217645e-01 6.35146797e-02
-1.13315284e+00 2.98563186e-02 -2.36690357e-01 -1.44228667e-01
1.39480129e-01 6.19859695e-01 -8.69545877e-01 4.57494855e-01
6.41943276e-01 4.96577889e-01 -4.01348144e-01 1.46416199e+00
-1.72931433e-01 1.00309384e+00 -1.36763185e-01 3.76135588e-01
-3.62734348e-01 -4.20498289e-02 6.79139555e-01 1.12583840e+00
8.04772615e-01 -2.41925791e-01 -2.27717906e-01 6.83529139e-01
-1.47099718e-01 2.63424486e-01 -3.60109031e-01 -6.11938611e-02
5.17028689e-01 8.11250329e-01 -4.71377909e-01 -2.72678807e-02
2.70890612e-02 9.48983371e-01 -4.13498998e-01 2.57756412e-01
-6.10816360e-01 -6.92200005e-01 5.46788514e-01 1.52098373e-01
1.85324639e-01 -1.10489331e-01 -2.30465710e-01 -8.37108552e-01
-2.72400230e-02 -1.41314900e+00 1.74707979e-01 -1.07702088e+00
-1.37720215e+00 8.00927222e-01 -1.40516356e-01 -1.87381542e+00
-5.37246644e-01 -3.06543559e-01 -5.60456157e-01 9.90115464e-01
-8.50571871e-01 -5.24805129e-01 -2.91984886e-01 6.46376967e-01
4.65030253e-01 -3.85118157e-01 1.01148832e+00 6.30363762e-01
2.94377744e-01 8.49327385e-01 2.50771083e-02 -1.26873299e-01
1.07725072e+00 -1.20075500e+00 8.06721598e-02 5.40942311e-01
4.52462673e-01 3.29970777e-01 1.03001440e+00 -3.16811323e-01
-8.93741667e-01 -9.37966645e-01 4.17925477e-01 -1.16285115e-01
5.36242485e-01 -2.01001048e-01 -8.66039753e-01 -7.22336620e-02
1.80418029e-01 -3.61666530e-01 9.99313414e-01 4.07224037e-02
-2.88983166e-01 -4.40505683e-01 -1.26239502e+00 2.91314900e-01
1.09553361e+00 -9.21165228e-01 -7.23254979e-01 -1.58208787e-01
6.55537665e-01 -1.31517708e-01 -1.16009235e+00 3.12642336e-01
7.85436630e-01 -1.16951346e+00 9.42823529e-01 -1.19634658e-01
4.62161571e-01 -5.63365459e-01 -3.27879995e-01 -1.72785187e+00
-4.57878113e-01 -8.29002559e-01 8.99348557e-02 1.50613236e+00
5.23141801e-01 -2.13458017e-01 3.70181799e-01 -1.26323789e-01
-2.36736819e-01 -1.69785470e-01 -8.90915692e-01 -1.07689500e+00
-1.39622120e-02 -5.55004656e-01 6.45047247e-01 6.08864486e-01
9.80323479e-02 2.70934284e-01 -6.06084645e-01 -2.01801673e-01
4.63897556e-01 -1.37932599e-01 7.22935498e-01 -1.29246414e+00
-7.58685589e-01 -4.89308387e-01 -7.09042609e-01 -8.04500043e-01
-2.06360549e-01 -7.49961734e-01 2.69933552e-01 -1.16469264e+00
-2.52188206e-01 -3.27387482e-01 -6.66875899e-01 -3.69322002e-02
1.49116427e-01 7.76289821e-01 2.92809725e-01 2.92665213e-01
-1.51731893e-01 6.22645676e-01 1.37573445e+00 -3.46971691e-01
-3.74691844e-01 2.80863583e-01 -5.06784379e-01 6.51993930e-01
8.43133032e-01 -5.26186824e-01 -5.72767615e-01 -7.48323575e-02
3.52500737e-01 2.37883016e-01 2.43162230e-01 -1.79025793e+00
-1.54310241e-01 1.79866970e-01 3.51458281e-01 -3.16725105e-01
5.46207249e-01 -7.03684866e-01 5.97991168e-01 2.43687585e-01
-7.75951982e-01 -5.64904474e-02 4.14162166e-02 4.95652139e-01
-7.07909763e-01 -3.79071087e-01 9.78238225e-01 5.91660850e-02
-2.00409755e-01 -1.78489849e-01 -1.56104669e-01 1.77258566e-01
2.81083733e-01 -2.88390189e-01 1.07620887e-01 -9.17100072e-01
-9.42829013e-01 -6.72293007e-01 5.56116819e-01 2.44792014e-01
6.83751404e-01 -1.57024312e+00 -7.84830570e-01 1.96266368e-01
-2.54373737e-02 -5.51448405e-01 3.05575076e-02 3.40210557e-01
-3.60723495e-01 -2.39617117e-02 -5.96820235e-01 -4.89534974e-01
-1.14054084e+00 4.71317202e-01 2.67767847e-01 -1.34016818e-03
-4.18084085e-01 6.43654525e-01 -1.08532652e-01 5.18388450e-02
3.90999138e-01 -3.12314272e-01 -1.41008928e-01 2.49381978e-02
6.55487657e-01 5.20503700e-01 3.01425844e-01 -5.57323337e-01
9.80966315e-02 5.66763580e-01 6.30884707e-01 -7.60507405e-01
1.13212895e+00 -7.91919529e-02 1.55839339e-01 1.04288220e+00
1.06072617e+00 4.47006792e-01 -9.54428494e-01 8.60740915e-02
-2.33221889e-01 -7.63932645e-01 7.35104531e-02 -9.66788113e-01
-1.04137743e+00 7.54974961e-01 9.65489149e-01 5.46048760e-01
1.70769322e+00 -2.79355258e-01 7.16880679e-01 1.03738017e-01
5.95839143e-01 -1.08723783e+00 4.54378366e-01 2.05467612e-01
1.20890558e+00 -6.45035446e-01 -2.30909958e-01 -1.56872228e-01
-3.98613989e-01 9.51181114e-01 1.39945731e-01 -9.80687737e-02
5.56028128e-01 2.19046190e-01 3.84191602e-01 2.24884450e-01
-4.86199021e-01 -2.32648719e-02 6.21794224e-01 9.76022303e-01
6.91849947e-01 6.06103465e-02 -2.97825128e-01 5.40126264e-01
-1.12322855e+00 8.80013555e-02 4.07970160e-01 4.61294770e-01
-3.19273680e-01 -1.27565551e+00 -5.75325191e-01 4.67889458e-01
-6.37333393e-01 -5.67068495e-02 -2.86144495e-01 2.74274707e-01
3.22317064e-01 1.30918622e+00 -1.93024248e-01 -6.21001065e-01
5.46109974e-01 1.66745722e-01 6.24986589e-01 -4.78618860e-01
-6.45642221e-01 2.61126190e-01 2.86641903e-02 -3.84996325e-01
-5.34033060e-01 -4.04917866e-01 -6.85240328e-01 -6.48619533e-02
-3.36481839e-01 3.65803212e-01 4.95834321e-01 4.36861724e-01
1.24542983e-02 7.03983307e-01 7.98131227e-01 -9.65667129e-01
-5.27751923e-01 -1.23434496e+00 -9.77084041e-01 9.51382697e-01
3.42793494e-01 -6.09977424e-01 -6.07766926e-01 5.32337189e-01] | [15.51504898071289, 5.720449924468994] |
32c1c117-eaa9-4e7e-b6be-711a9aba5d75 | feature-based-selection-of-dependency-paths | null | null | https://aclanthology.org/P13-1050 | https://aclanthology.org/P13-1050.pdf | Feature-Based Selection of Dependency Paths in Ad Hoc Information Retrieval | null | ['K. Tamsin Maxwell', 'W. Bruce Croft', 'Oberl', 'Jon er'] | 2013-08-01 | null | null | null | acl-2013-8 | ['ad-hoc-information-retrieval'] | ['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] | [-7.404953479766846, 3.6708667278289795] |
b0870622-a2c7-4adc-b164-3189e6167d31 | automated-essay-scoring-using-transformer | 2110.06874 | null | https://arxiv.org/abs/2110.06874v1 | https://arxiv.org/pdf/2110.06874v1.pdf | Automated Essay Scoring Using Transformer Models | Automated essay scoring (AES) is gaining increasing attention in the education sector as it significantly reduces the burden of manual scoring and allows ad hoc feedback for learners. Natural language processing based on machine learning has been shown to be particularly suitable for text classification and AES. While many machine-learning approaches for AES still rely on a bag-of-words (BOW) approach, we consider a transformer-based approach in this paper, compare its performance to a logistic regression model based on the BOW approach and discuss their differences. The analysis is based on 2,088 email responses to a problem-solving task, that were manually labeled in terms of politeness. Both transformer models considered in that analysis outperformed without any hyper-parameter tuning the regression-based model. We argue that for AES tasks such as politeness classification, the transformer-based approach has significant advantages, while a BOW approach suffers from not taking word order into account and reducing the words to their stem. Further, we show how such models can help increase the accuracy of human raters, and we provide a detailed instruction on how to implement transformer-based models for one's own purpose. | ['Steffen Brandt', 'Kerstin Eilers', 'Christopher Hansen', 'Christian Mayer', 'Sabrina Ludwig'] | 2021-10-13 | null | null | null | null | ['automated-essay-scoring'] | ['natural-language-processing'] | [ 1.15497850e-01 1.61702812e-01 -8.19658786e-02 -5.19477308e-01
-9.25913393e-01 -6.10779524e-01 3.30821335e-01 8.76459897e-01
-6.48734868e-01 7.03262508e-01 4.66836005e-01 -6.66220546e-01
-4.28028643e-01 -7.30387390e-01 -5.22237904e-02 -5.65934777e-01
6.32283330e-01 5.95750272e-01 1.44791916e-01 -5.27627647e-01
7.61260211e-01 1.42489582e-01 -1.55937564e+00 3.72294843e-01
1.06311810e+00 6.77272260e-01 2.64796671e-02 7.69449651e-01
-4.85928476e-01 1.05519795e+00 -9.65398610e-01 -1.02132630e+00
-9.51249376e-02 -3.00713867e-01 -8.01760077e-01 -4.53027695e-01
7.15417445e-01 -1.97988958e-03 3.82913575e-02 7.23831177e-01
4.28145170e-01 1.31012112e-01 1.03522038e+00 -9.56317246e-01
-5.75633824e-01 4.40686017e-01 -1.52619585e-01 -7.50440881e-02
7.83066273e-01 -2.03584343e-01 1.45362747e+00 -5.04749477e-01
3.25467795e-01 9.39705968e-01 9.07013118e-01 4.78473008e-01
-1.07899785e+00 -6.72292292e-01 -1.70313820e-01 3.98716807e-01
-6.61919177e-01 -7.19803721e-02 6.97931111e-01 -6.67085767e-01
8.40536118e-01 3.59171152e-01 7.84836352e-01 7.78593540e-01
1.11951135e-01 9.05085921e-01 1.62667239e+00 -1.07325423e+00
5.76329753e-02 5.14560580e-01 7.73179948e-01 7.66298354e-01
3.55609030e-01 -2.98738182e-01 -6.28154635e-01 -2.78784305e-01
1.75036103e-01 -5.40771961e-01 -6.77919611e-02 -3.24403107e-01
-8.55991244e-01 1.17737579e+00 2.20194638e-01 4.01113629e-01
-2.17133015e-01 -1.83457077e-01 6.11069560e-01 7.80311227e-01
4.63241130e-01 9.07272339e-01 -6.27319753e-01 -5.52051306e-01
-1.10534716e+00 4.40338999e-01 1.11548591e+00 5.88931799e-01
3.51189673e-01 -1.67858809e-01 -3.99993777e-01 1.12058091e+00
2.80602902e-01 1.89379781e-01 9.25334930e-01 -3.97832751e-01
4.42337304e-01 1.00615728e+00 -1.93634883e-01 -1.06686962e+00
-4.17284340e-01 -1.77490696e-01 -2.88736224e-01 1.73399568e-01
8.49211335e-01 1.16492271e-01 -6.10966444e-01 1.47271764e+00
-1.38039626e-02 -4.96197343e-01 -2.39083432e-02 5.74172854e-01
1.17421818e+00 5.92453003e-01 1.86588511e-01 -2.98750848e-01
1.69351864e+00 -9.91125762e-01 -7.72942483e-01 -3.11933570e-02
1.19583261e+00 -9.95271623e-01 1.49686348e+00 8.76028299e-01
-1.04496384e+00 -1.24421895e-01 -1.13771272e+00 -2.13350862e-01
-5.47368109e-01 7.29993358e-02 4.31318909e-01 1.30534649e+00
-9.27459180e-01 4.45825040e-01 -2.24561363e-01 -2.98341304e-01
1.10811539e-01 3.77821565e-01 -4.16707575e-01 1.71573430e-01
-1.18392515e+00 1.41272843e+00 -1.99771523e-01 -5.45388937e-01
1.55333072e-01 -5.78165710e-01 -9.31298018e-01 3.84839654e-01
-9.79645699e-02 -2.65307933e-01 1.68222547e+00 -7.23817110e-01
-1.79445791e+00 1.12806046e+00 -2.01273233e-01 -2.08269954e-01
3.89685452e-01 -1.42400162e-02 8.99792463e-02 -1.41881788e-02
-1.62632123e-01 1.89074263e-01 5.60818493e-01 -6.77374721e-01
-4.81500864e-01 -2.23865405e-01 3.97956260e-02 7.58989304e-02
-8.87484014e-01 3.45456541e-01 2.72008061e-01 -7.45870709e-01
2.92357188e-02 -8.98618758e-01 2.16566548e-01 -3.43163729e-01
3.52886736e-01 -9.09451187e-01 3.22692692e-01 -7.22336233e-01
1.81012261e+00 -1.71266866e+00 -3.15043069e-02 1.91378176e-01
1.08713061e-01 5.26788533e-01 8.78251642e-02 4.91824239e-01
-6.57454059e-02 2.16483697e-02 -7.09674060e-02 -3.45706820e-01
3.54986161e-01 -6.82452396e-02 -1.72365397e-01 2.50149041e-01
-1.01659000e-01 6.65533543e-01 -8.10855925e-01 -5.49729705e-01
1.39979452e-01 2.85928130e-01 -5.75131536e-01 2.33472571e-01
1.11373521e-01 -3.13790500e-01 -6.17423132e-02 3.64487886e-01
1.96533620e-01 1.16238773e-01 8.10401365e-02 3.42131436e-01
-1.88716665e-01 1.02736485e+00 -8.46433520e-01 1.08157396e+00
-8.78569782e-01 1.06517315e+00 -1.65773749e-01 -8.77193868e-01
1.36569250e+00 4.54121560e-01 1.29915908e-01 -7.49048829e-01
6.31711259e-02 5.11539698e-01 2.57515103e-01 -4.81008142e-01
5.48122168e-01 -3.15571040e-01 -3.55729729e-01 8.96268487e-01
4.06975579e-03 -5.89500070e-01 3.37664902e-01 2.02143878e-01
1.11030972e+00 -5.42517453e-02 7.09010184e-01 -4.76609409e-01
9.06143367e-01 2.84413956e-02 1.23345673e-01 5.63908815e-01
-2.95812279e-01 3.34637135e-01 7.43570864e-01 -5.02272546e-01
-8.06798220e-01 -4.16355729e-01 -3.11358362e-01 1.48927355e+00
-5.89872301e-01 -7.31003046e-01 -9.30222034e-01 -7.23263860e-01
5.16001657e-02 9.74513650e-01 -5.57644010e-01 -1.11824773e-01
-6.02480412e-01 -6.90798461e-01 4.80152518e-01 1.79836288e-01
9.55244824e-02 -1.22831070e+00 -8.98484349e-01 2.65318125e-01
-2.56052256e-01 -5.95109463e-01 -2.26841658e-01 2.87973017e-01
-7.62535274e-01 -9.15417731e-01 -7.17105746e-01 -7.96708345e-01
3.91462713e-01 2.72463560e-01 1.25717926e+00 4.13817376e-01
-9.20288786e-02 5.73743761e-01 -6.12562001e-01 -9.36254025e-01
-6.09902799e-01 4.19421107e-01 -5.03931828e-02 -4.47049707e-01
1.09766376e+00 -3.81707728e-01 -1.84144557e-01 1.06554985e-01
-6.37471259e-01 -1.30882367e-01 4.27340090e-01 1.01261652e+00
-3.96795481e-01 -3.31782520e-01 7.56184101e-01 -1.00697935e+00
1.26791894e+00 -1.42872944e-01 -4.76857930e-01 2.72359908e-01
-1.05618107e+00 1.21247336e-01 3.63666087e-01 -3.63629609e-01
-8.10855091e-01 -1.45478278e-01 -3.18667740e-01 3.80780965e-01
-9.32018757e-02 5.24070382e-01 3.11223865e-01 -4.10082698e-01
8.51979613e-01 3.88704576e-02 1.46378860e-01 -3.63542944e-01
-2.00754732e-01 1.09408069e+00 -4.35720645e-02 -6.36702776e-01
5.07433474e-01 -4.73605454e-01 -1.06644534e-01 -6.81027710e-01
-1.09717095e+00 -8.04112494e-01 -6.94528103e-01 -4.27242875e-01
4.88259822e-01 -5.92036009e-01 -1.06683803e+00 2.79178828e-01
-1.04297125e+00 -2.92812169e-01 -5.04906513e-02 5.02632856e-01
-7.20631540e-01 3.27471137e-01 -7.41306424e-01 -9.64056551e-01
-4.62974787e-01 -1.09725976e+00 6.80388391e-01 3.25690866e-01
-1.05497360e+00 -1.13421345e+00 3.49240124e-01 1.00015891e+00
5.13701558e-01 -3.54818612e-01 1.44831359e+00 -1.05111015e+00
2.43921936e-01 -5.53121567e-01 -7.19119385e-02 2.21165448e-01
-2.18871430e-01 -6.84053823e-02 -1.04418468e+00 2.40582358e-02
1.85803607e-01 -5.67519605e-01 6.96492434e-01 1.88048422e-01
8.90339136e-01 -4.59312856e-01 3.47062230e-01 1.30415484e-01
1.08132136e+00 -5.17865457e-02 5.39703250e-01 9.40176725e-01
3.46869439e-01 1.10062206e+00 7.10965514e-01 1.08166367e-01
6.19427145e-01 7.20369518e-01 1.32589772e-01 2.03374133e-01
3.47817987e-02 -9.95070860e-02 5.61771274e-01 1.18170178e+00
-1.25569925e-02 -1.10822417e-01 -1.06536531e+00 5.07898450e-01
-1.66302907e+00 -8.24096441e-01 -6.27445698e-01 2.32356954e+00
1.12616944e+00 2.36848667e-01 3.76292735e-01 8.00241888e-01
3.71856809e-01 1.36577319e-02 3.62258196e-01 -1.26419437e+00
1.52741313e-01 7.54090011e-01 2.44050786e-01 8.64908159e-01
-7.49726772e-01 8.69248271e-01 6.50819588e+00 7.81752646e-01
-8.98659170e-01 -1.10420184e-02 2.52937645e-01 1.19657062e-01
-1.85617730e-01 -2.93208688e-01 -6.58715189e-01 2.89155632e-01
1.03624701e+00 -9.57908481e-02 2.22290624e-02 8.43492866e-01
-3.13516473e-03 -2.15805858e-01 -9.02023792e-01 7.84853458e-01
2.66305178e-01 -8.43658209e-01 1.80945955e-02 -1.32625461e-01
4.90102887e-01 -6.33301973e-01 -1.69833079e-01 6.78909242e-01
1.20001063e-01 -1.13561082e+00 6.00017667e-01 2.30053574e-01
4.26340431e-01 -6.71409845e-01 1.03326726e+00 3.98004144e-01
-6.21672153e-01 -1.77829519e-01 -3.10789138e-01 -7.52653182e-01
-3.45611274e-01 3.30301166e-01 -1.09776187e+00 3.17559540e-02
5.11282682e-01 2.43935972e-01 -7.74862468e-01 1.08075345e+00
-5.58630705e-01 1.09625328e+00 9.20248702e-02 -9.30601716e-01
2.94048250e-01 -2.54306912e-01 3.68628919e-01 1.56599212e+00
2.80086935e-01 1.50276929e-01 -4.11891371e-01 2.48507038e-01
3.01079780e-01 8.10907006e-01 -3.47148299e-01 1.11413613e-01
2.92309254e-01 1.40071845e+00 -5.46876848e-01 -3.30056727e-01
-4.30384874e-01 5.28923929e-01 4.67866153e-01 -1.97535127e-01
-3.91077578e-01 -6.60161257e-01 3.83451223e-01 3.11436683e-01
-1.01781867e-01 -1.08144328e-01 -8.10716271e-01 -8.97240877e-01
-3.07996906e-02 -1.19986212e+00 5.37885427e-01 -8.31972480e-01
-1.23307717e+00 4.06556159e-01 -1.05562031e-01 -1.01418567e+00
-1.99604839e-01 -1.17312908e+00 -6.29986703e-01 8.35255444e-01
-1.31220460e+00 -7.36573040e-01 -3.72308791e-01 2.32847050e-01
6.07854724e-01 -3.13216418e-01 1.08936357e+00 2.30339885e-01
-3.05590928e-01 8.45281541e-01 -1.34041637e-01 1.01696476e-02
1.24167526e+00 -1.72339237e+00 -1.41709581e-01 2.50973195e-01
2.21305624e-01 5.01433492e-01 9.33005750e-01 -3.03896338e-01
-8.43397737e-01 -2.54654467e-01 1.83303249e+00 -6.39889240e-01
7.73617446e-01 -2.12646738e-01 -1.08201194e+00 2.63791740e-01
3.52335036e-01 -7.02009261e-01 1.23987699e+00 4.07315940e-01
-4.51697588e-01 6.04087599e-02 -1.05721128e+00 5.43407023e-01
3.66245210e-01 -3.47467035e-01 -1.22877872e+00 4.39772904e-01
1.65132254e-01 -1.68334216e-01 -7.94526458e-01 1.47277415e-01
8.09274614e-01 -1.08222032e+00 6.55461550e-01 -8.06796908e-01
7.93830752e-01 2.38466695e-01 2.38205090e-01 -1.50987637e+00
-4.60479856e-01 -4.92084056e-01 1.10791385e-01 1.05669045e+00
4.35406387e-01 -6.24346197e-01 9.00585592e-01 7.75183439e-01
-3.94424163e-02 -8.80677342e-01 -7.60904670e-01 -4.94836599e-01
5.09292841e-01 -3.25967729e-01 2.99537003e-01 1.11044157e+00
8.49700809e-01 6.39957547e-01 -4.70236950e-02 -5.57016313e-01
1.54586688e-01 -1.28011152e-01 8.27204406e-01 -1.71511209e+00
-2.14869324e-02 -1.08338630e+00 -5.03010273e-01 -7.05572128e-01
2.13865340e-01 -8.36037993e-01 4.39159460e-02 -1.43277824e+00
1.53087646e-01 -4.42718297e-01 -1.34861439e-01 5.40041745e-01
-3.82144213e-01 3.54082108e-01 1.65076032e-01 -4.30847444e-02
-2.36812279e-01 1.76539302e-01 8.33790481e-01 2.28026621e-02
-1.44078642e-01 2.40141168e-01 -8.00105214e-01 7.10008442e-01
8.19298148e-01 -7.36462533e-01 -1.66491136e-01 -7.66286850e-02
6.76241934e-01 -4.34818938e-02 -8.73475745e-02 -6.76555395e-01
4.39900756e-01 -9.66251418e-02 1.04103200e-01 -3.70388106e-02
1.60303056e-01 -6.77887797e-01 -5.62936664e-01 5.13570607e-01
-6.74276173e-01 2.41266906e-01 2.97623008e-01 -1.14383474e-02
-3.45792145e-01 -1.10588181e+00 6.11785412e-01 -7.13197663e-02
1.13108552e-04 -4.54901993e-01 -8.57472122e-01 -2.10278377e-01
9.40729797e-01 -3.27959299e-01 -2.53765643e-01 -7.33881772e-01
-3.19329768e-01 -2.68814024e-02 2.33146295e-01 3.18475068e-01
2.96998322e-01 -1.06100392e+00 -8.41511607e-01 1.74486607e-01
2.16943532e-01 -4.18548346e-01 -1.91158205e-01 1.01034963e+00
-8.69160116e-01 8.66009951e-01 -2.28986710e-01 -2.78258324e-01
-1.98743474e+00 9.49718431e-02 -3.87904197e-02 -7.59343266e-01
-2.80599892e-01 8.14824283e-01 -2.52543002e-01 -6.36573970e-01
3.69959861e-01 -2.17616767e-01 -8.52577925e-01 6.27219319e-01
5.57855844e-01 5.08645236e-01 5.95958948e-01 -6.54499829e-01
-4.07532696e-03 5.54185510e-01 -1.72546744e-01 -2.33655408e-01
1.30792165e+00 2.84858435e-01 -2.10521266e-01 6.25601292e-01
7.94214725e-01 5.17030060e-01 -3.32083941e-01 -1.64668560e-01
4.77868259e-01 -4.33821976e-01 1.84113443e-01 -9.35458422e-01
-3.68685484e-01 1.14737248e+00 -9.76390875e-05 5.80761194e-01
9.41006362e-01 -4.54737782e-01 5.23997068e-01 6.05199397e-01
1.12505123e-01 -1.17128277e+00 1.22052722e-01 8.10250163e-01
8.19668889e-01 -1.20096171e+00 3.23331118e-01 -4.07318354e-01
-6.55003786e-01 1.72472560e+00 5.33310413e-01 -5.54508828e-02
3.16108584e-01 7.20471293e-02 3.36137563e-01 -1.08262554e-01
-8.73493254e-01 6.60377145e-02 6.39671922e-01 2.75287777e-01
1.07558739e+00 1.11420326e-01 -9.39304709e-01 8.27292919e-01
-9.67238724e-01 -2.57032692e-01 8.95589471e-01 8.55473161e-01
-6.84972525e-01 -1.37423098e+00 -6.10760391e-01 6.97427928e-01
-7.11893857e-01 -3.20510328e-01 -8.26090872e-01 7.68607080e-01
-3.67968053e-01 9.52629745e-01 -9.43484977e-02 -2.65457988e-01
3.73707354e-01 6.54515207e-01 6.14366293e-01 -8.19676876e-01
-1.26311409e+00 -4.61276233e-01 2.73288280e-01 1.52989656e-01
-2.21068382e-01 -9.80946958e-01 -7.34447956e-01 -4.48690951e-01
-5.79410136e-01 4.14816350e-01 8.16001058e-01 1.00978422e+00
-2.88533241e-01 2.83358246e-01 3.37259889e-01 -3.43850493e-01
-1.04351223e+00 -1.29539311e+00 -2.77438939e-01 1.94904342e-01
2.31896684e-01 -5.64978421e-01 -6.72994018e-01 -2.65068710e-01] | [11.27468204498291, 9.325878143310547] |
e1071354-6081-446e-8304-946742111bc0 | tinyhd-efficient-video-saliency-prediction | 2301.04619 | null | https://arxiv.org/abs/2301.04619v1 | https://arxiv.org/pdf/2301.04619v1.pdf | TinyHD: Efficient Video Saliency Prediction with Heterogeneous Decoders using Hierarchical Maps Distillation | Video saliency prediction has recently attracted attention of the research community, as it is an upstream task for several practical applications. However, current solutions are particularly computationally demanding, especially due to the wide usage of spatio-temporal 3D convolutions. We observe that, while different model architectures achieve similar performance on benchmarks, visual variations between predicted saliency maps are still significant. Inspired by this intuition, we propose a lightweight model that employs multiple simple heterogeneous decoders and adopts several practical approaches to improve accuracy while keeping computational costs low, such as hierarchical multi-map knowledge distillation, multi-output saliency prediction, unlabeled auxiliary datasets and channel reduction with teacher assistant supervision. Our approach achieves saliency prediction accuracy on par or better than state-of-the-art methods on DFH1K, UCF-Sports and Hollywood2 benchmarks, while enhancing significantly the efficiency of the model. Code is on https://github.com/feiyanhu/tinyHD | ['Kevin McGuinness', 'Concetto Spampinato', 'Morteza Moradi', 'Giovanni Bellitto', 'Federica Proietto Salanitri', 'Simone Palazzo', 'Feiyan Hu'] | 2023-01-11 | null | null | null | null | ['saliency-prediction'] | ['computer-vision'] | [ 3.02301019e-01 1.31049514e-01 -4.23717439e-01 -3.64977151e-01
-8.89148176e-01 -2.06592813e-01 3.62811089e-01 1.39056250e-01
-4.64838177e-01 6.49247050e-01 5.42659760e-01 -1.98968261e-01
1.10910445e-01 -3.13799977e-01 -9.29387152e-01 -3.66183490e-01
1.25297025e-01 1.91373862e-02 7.28433967e-01 -1.95385128e-01
3.30263168e-01 -9.79053006e-02 -1.72823715e+00 3.01508516e-01
1.02510893e+00 1.09279549e+00 6.12430871e-01 6.16054893e-01
3.01891565e-01 9.55800951e-01 -7.38633350e-02 -6.58695579e-01
1.55113056e-01 -1.42267466e-01 -1.00242996e+00 -2.04848081e-01
6.59097731e-01 -3.80738467e-01 -5.45266151e-01 1.05225897e+00
6.22837186e-01 6.01128414e-02 2.82698989e-01 -1.24680173e+00
-7.21959233e-01 8.41288626e-01 -7.07399964e-01 4.26803291e-01
-6.99244365e-02 9.18309763e-02 1.12995625e+00 -9.16326225e-01
4.08374369e-01 9.79051471e-01 5.82263708e-01 4.94777322e-01
-1.21973550e+00 -6.84449315e-01 3.65488350e-01 7.08498001e-01
-1.23780763e+00 -5.69066763e-01 4.67862040e-01 -6.22417703e-02
7.99113035e-01 1.68910027e-01 4.92142379e-01 1.08954167e+00
-1.94979757e-01 1.56763744e+00 9.43299413e-01 -2.84802228e-01
3.26566026e-02 7.76364058e-02 1.41186520e-01 7.01876879e-01
-2.56094476e-03 -2.09702432e-01 -9.61541951e-01 1.69585094e-01
6.82112575e-01 1.17510654e-01 -3.86737138e-01 -4.80013281e-01
-1.30184519e+00 6.52961552e-01 6.25556648e-01 -1.01700760e-02
-2.42835522e-01 3.64807755e-01 4.69682992e-01 1.01540036e-01
5.55704653e-01 2.48489946e-01 -6.33090019e-01 -3.75105679e-01
-1.20873713e+00 3.75276744e-01 3.70450586e-01 1.21238172e+00
7.05235839e-01 -6.54906705e-02 -3.89580011e-01 6.04135990e-01
4.10398841e-02 5.49723208e-01 4.92555559e-01 -1.01130044e+00
4.11744177e-01 3.70856971e-01 1.54293543e-02 -7.61277854e-01
-3.68961543e-01 -6.20310068e-01 -6.13328040e-01 -1.71125963e-01
3.63677979e-01 -4.95184883e-02 -1.00053120e+00 1.67852008e+00
2.01258987e-01 7.45471001e-01 -5.40015362e-02 1.23722875e+00
9.62269723e-01 4.16545808e-01 2.53268957e-01 3.71815979e-01
1.24540758e+00 -1.63026762e+00 -3.11476201e-01 -4.11705822e-01
4.76936847e-01 -6.46142900e-01 1.25876057e+00 2.62378275e-01
-1.23778033e+00 -4.88408118e-01 -8.47913265e-01 -5.74536145e-01
-1.37565151e-01 2.80462772e-01 8.10680091e-01 2.91057318e-01
-1.26822484e+00 5.16445577e-01 -9.10124421e-01 -3.50652099e-01
8.83126020e-01 3.99747908e-01 -1.92033201e-02 -1.60395133e-03
-1.00046480e+00 8.69552851e-01 2.81295538e-01 -1.66335866e-01
-1.05352569e+00 -9.70263779e-01 -8.28543425e-01 3.69673908e-01
6.33262455e-01 -5.70997417e-01 1.68288982e+00 -1.01433253e+00
-1.30850744e+00 6.50615036e-01 -2.81283319e-01 -8.31646979e-01
4.56871003e-01 -5.26140332e-01 -1.55040950e-01 1.82983726e-01
1.55433446e-01 1.19959617e+00 8.83788824e-01 -8.55384886e-01
-1.07102156e+00 -1.34173185e-01 1.29482120e-01 4.04257089e-01
-4.81360495e-01 4.60991403e-03 -7.78461576e-01 -6.58871412e-01
-1.60561442e-01 -9.64651346e-01 -3.96077782e-01 -8.35463032e-02
-5.42878449e-01 -6.87031522e-02 8.16191077e-01 -6.77566171e-01
1.17967236e+00 -2.26363373e+00 1.11423105e-01 -1.81358337e-01
4.62018579e-01 3.25225025e-01 -2.30692670e-01 -9.92585570e-02
1.13589853e-01 -2.14917988e-01 -2.06638202e-01 -4.66799855e-01
1.71725601e-02 -2.18312815e-01 -3.24459553e-01 2.42583692e-01
2.73026377e-01 1.18992698e+00 -1.04330015e+00 -4.45391506e-01
9.40272510e-02 5.14040709e-01 -6.44925296e-01 -7.10392743e-02
-2.18602136e-01 2.87295043e-01 -3.82583469e-01 7.65962720e-01
4.72545326e-01 -7.53275633e-01 -4.46235798e-02 -5.12674786e-02
-1.59778118e-01 6.55310512e-01 -9.79170740e-01 2.03858948e+00
-3.06262970e-01 8.80655527e-01 -5.86570986e-02 -1.10132492e+00
4.55063254e-01 9.16225389e-02 2.08203509e-01 -9.30308521e-01
1.47265032e-01 2.00081244e-01 -1.38460696e-01 -1.51947185e-01
8.51506352e-01 4.84148830e-01 1.12049937e-01 1.79879457e-01
2.39827245e-01 3.09274018e-01 8.30447003e-02 3.90563577e-01
1.16105878e+00 2.77597040e-01 5.18781431e-02 -3.84980887e-01
6.59452006e-02 1.39195085e-01 5.25908232e-01 6.83599472e-01
-3.79661381e-01 8.47740233e-01 4.86975670e-01 -2.11253002e-01
-9.56113935e-01 -1.04769790e+00 8.06889161e-02 1.43903208e+00
6.44686997e-01 -4.29060429e-01 -7.99437463e-01 -6.51908457e-01
-3.41865048e-02 5.68586886e-01 -6.27936184e-01 -1.25879392e-01
-4.06619698e-01 -5.05787909e-01 5.05695581e-01 6.92376435e-01
6.44229710e-01 -9.88651633e-01 -9.95965958e-01 1.89042762e-01
-2.66524017e-01 -1.23792219e+00 -4.90963221e-01 2.67277390e-01
-8.14616501e-01 -8.71027112e-01 -9.91546333e-01 -8.58347714e-01
3.21946383e-01 8.33284259e-01 1.33994782e+00 1.38355121e-01
-2.19641432e-01 1.56247914e-01 -3.96746010e-01 -6.25790656e-01
2.16613621e-01 5.47469079e-01 -2.91311741e-01 -2.63779432e-01
4.30075556e-01 -4.47870135e-01 -8.89837325e-01 1.52118519e-01
-7.07854927e-01 6.81497633e-01 7.66832709e-01 8.35552096e-01
6.52146935e-01 -4.11763102e-01 6.86034858e-01 -8.58979523e-01
1.53899074e-01 -4.75646675e-01 -5.28012037e-01 9.36952606e-02
-5.53496718e-01 3.24432924e-02 4.01594132e-01 -2.25412637e-01
-1.04267228e+00 1.53538093e-01 -3.15174386e-02 -5.12664258e-01
-2.01879859e-01 1.75497487e-01 1.97154239e-01 -1.00544982e-01
4.81598526e-01 3.60757202e-01 -7.42437616e-02 -4.33191627e-01
3.56625825e-01 5.30865133e-01 5.57664454e-01 -1.52023345e-01
5.87454677e-01 4.56513882e-01 -2.92163432e-01 -6.26729846e-01
-1.25737035e+00 -4.84658420e-01 -4.48679507e-01 4.64295521e-02
7.05991924e-01 -1.51759171e+00 -5.60187995e-01 4.74174231e-01
-9.54311371e-01 -7.12359428e-01 -8.48171040e-02 3.87203544e-01
-4.56877649e-01 1.42191410e-01 -7.00217545e-01 -2.73713917e-01
-5.52550495e-01 -1.29610872e+00 1.16305065e+00 5.14023125e-01
-6.46573976e-02 -6.83489859e-01 -4.20703709e-01 4.99470502e-01
5.79731882e-01 -2.25270823e-01 6.92225993e-01 -5.01389086e-01
-9.91311610e-01 2.23790079e-01 -6.20590270e-01 1.44960225e-01
-2.50115424e-01 -3.55276167e-01 -1.10604584e+00 -1.05018228e-01
-5.70902944e-01 -4.87600774e-01 1.25702536e+00 5.92044652e-01
1.49429691e+00 -1.10586718e-01 -4.43530232e-01 6.05203986e-01
1.29634261e+00 -2.66533852e-01 5.77640116e-01 4.84638035e-01
8.37817967e-01 1.94790259e-01 6.81684494e-01 4.05831307e-01
9.73812997e-01 7.52166212e-01 6.02244198e-01 -3.31612885e-01
-3.13437909e-01 -2.34784514e-01 2.98618257e-01 5.26680648e-01
-3.41987051e-02 -1.39070861e-02 -8.36284518e-01 1.00859094e+00
-2.18512392e+00 -9.50382531e-01 -2.25575328e-01 2.07823491e+00
8.44917953e-01 2.24516466e-01 3.75820369e-01 -5.92037514e-02
6.88616753e-01 2.61134624e-01 -6.56357169e-01 -1.66841105e-01
-7.45227709e-02 1.98050424e-01 8.20446908e-01 1.51875243e-01
-1.27899230e+00 1.23563349e+00 5.82797623e+00 9.64510202e-01
-1.31559527e+00 3.03989381e-01 9.98929381e-01 -4.12312359e-01
-2.65455186e-01 -4.83455881e-02 -8.84947479e-01 7.23150015e-01
7.61766016e-01 -9.64263733e-03 4.53793585e-01 8.96645784e-01
-6.04273693e-04 -3.13752562e-01 -9.41285908e-01 9.17621970e-01
7.56300986e-02 -1.57466578e+00 -2.40487441e-01 -3.42508614e-01
9.50947106e-01 7.02254891e-01 4.01867837e-01 3.85396898e-01
2.71385849e-01 -8.93065333e-01 9.17892933e-01 1.71524048e-01
5.53528011e-01 -6.99701667e-01 5.82350135e-01 2.11137503e-01
-9.92769480e-01 -1.23920962e-01 -3.97920877e-01 -2.38314092e-01
-2.00649328e-03 6.05577707e-01 -6.79703772e-01 1.49668485e-01
1.13155973e+00 8.68429899e-01 -9.37919438e-01 1.29407871e+00
-3.37589324e-01 7.76509404e-01 -2.11538717e-01 -8.10477957e-02
4.05626565e-01 2.40965307e-01 2.89107710e-01 1.25574720e+00
2.96106458e-01 -8.24427754e-02 -1.85102634e-02 6.32868826e-01
-2.88467705e-01 -3.75225954e-02 -3.30554783e-01 3.30927938e-01
5.16534209e-01 1.33761060e+00 -6.53798342e-01 -4.27136570e-01
-8.27049434e-01 9.81279373e-01 5.08210599e-01 2.76671737e-01
-1.27571702e+00 -2.99402505e-01 8.15116465e-01 4.44908179e-02
7.28267848e-01 4.59434763e-02 -6.14116132e-01 -1.21877372e+00
2.60699652e-02 -7.29448676e-01 1.31750837e-01 -7.57836223e-01
-5.84699512e-01 3.13061327e-01 -3.25342566e-01 -1.14995122e+00
1.46045744e-01 -2.43166402e-01 -5.81419170e-01 5.97419441e-01
-2.07910848e+00 -1.21813047e+00 -3.87368679e-01 5.46169281e-01
7.36838341e-01 3.30235176e-02 4.09070641e-01 4.68973607e-01
-5.91776490e-01 7.93718994e-01 1.20550781e-01 1.90248881e-02
7.70373166e-01 -1.24918950e+00 7.09093988e-01 1.04460084e+00
2.92688727e-01 6.34845570e-02 6.27969980e-01 -2.29717314e-01
-1.27988243e+00 -1.18583035e+00 9.79237020e-01 -3.12912375e-01
6.85307086e-01 -3.94134909e-01 -8.96055758e-01 7.45680809e-01
3.86993647e-01 1.20048396e-01 4.09041196e-01 1.43485323e-01
-2.40940720e-01 1.45800188e-02 -7.01634884e-01 9.23089445e-01
1.19698334e+00 -4.00493473e-01 -2.64190137e-01 2.05824599e-01
8.77921641e-01 -6.90048158e-01 -4.67459172e-01 3.82987320e-01
4.67776805e-01 -1.07859838e+00 9.94705677e-01 -4.36974376e-01
8.58884335e-01 -2.84754455e-01 2.11447384e-02 -1.23371518e+00
-5.21900713e-01 -4.51450258e-01 -3.27470154e-01 9.61723804e-01
6.56386495e-01 -1.00206599e-01 1.03982437e+00 6.50003433e-01
-4.13983792e-01 -1.07523704e+00 -7.86789954e-01 -5.10801554e-01
-3.71521145e-01 -3.13405275e-01 4.61447924e-01 7.10558295e-01
2.24751271e-02 4.73335028e-01 -5.49267054e-01 2.22226530e-01
5.12237072e-01 4.12964106e-01 7.23760486e-01 -9.17118907e-01
-2.12628469e-01 -5.64964473e-01 -4.57774669e-01 -1.50497699e+00
1.80156425e-01 -8.46298575e-01 6.85801059e-02 -1.42701530e+00
4.09170717e-01 -4.03805107e-01 -6.28012896e-01 8.01095247e-01
-4.28688943e-01 6.02922380e-01 4.62487400e-01 4.55402723e-03
-1.25611424e+00 5.88888824e-01 1.27317739e+00 -1.58579305e-01
-3.84774394e-02 -1.16572842e-01 -1.10285378e+00 7.64985442e-01
7.74468720e-01 -3.32889527e-01 -3.53781730e-01 -9.75470662e-01
-2.96588168e-02 -6.03347719e-02 6.53731167e-01 -1.22087932e+00
3.10404181e-01 -4.47737798e-02 3.52997899e-01 -4.54254568e-01
3.14004123e-01 -6.70092106e-01 -3.81409317e-01 4.13319081e-01
-5.72721899e-01 -8.93119574e-02 2.96381056e-01 4.55442548e-01
-2.90592521e-01 9.41852406e-02 7.61121631e-01 1.33221924e-01
-1.30379224e+00 4.18595284e-01 -2.44411193e-02 1.83522284e-01
9.52029943e-01 -1.78608205e-02 -5.57726860e-01 -4.05136377e-01
-4.10705179e-01 4.53933716e-01 5.14257729e-01 4.79427248e-01
5.71606159e-01 -1.09269881e+00 -6.23354137e-01 -5.69375232e-02
1.27561882e-01 2.01920122e-01 3.65707397e-01 1.26136196e+00
-3.70779216e-01 4.08919036e-01 -2.68310875e-01 -6.07518137e-01
-1.26702535e+00 4.27413940e-01 -1.05402038e-01 -5.41679449e-02
-6.54266536e-01 1.22055888e+00 4.07729954e-01 -6.96409196e-02
4.05042052e-01 -3.35191280e-01 -1.33790776e-01 -2.06437200e-01
5.22989154e-01 4.24739778e-01 6.33062646e-02 -5.07553935e-01
-3.13962609e-01 1.83288485e-01 -2.87959576e-01 3.99803966e-01
1.35111415e+00 -2.79685438e-01 3.93872380e-01 7.48146772e-02
9.96476054e-01 -2.76869804e-01 -1.65881538e+00 -4.78339642e-01
5.93624711e-02 -4.43340003e-01 3.06004107e-01 -8.26344311e-01
-1.25776386e+00 9.83133674e-01 5.38557708e-01 -4.34618928e-02
1.41599548e+00 1.58064455e-01 1.04947937e+00 3.03593308e-01
2.81853974e-01 -1.24375892e+00 -2.84865424e-02 6.24139607e-01
4.41295087e-01 -1.65213585e+00 -8.78690556e-02 -5.40082812e-01
-9.64819789e-01 5.39151788e-01 7.75138021e-01 -2.02949047e-01
4.85600889e-01 1.39142171e-01 -3.49508189e-02 6.53903335e-02
-9.78127122e-01 -5.64577341e-01 2.80473739e-01 3.62227052e-01
4.99637276e-01 7.02642873e-02 -1.05859362e-01 6.24273121e-01
-5.78983240e-02 8.37889910e-02 4.89068121e-01 6.66736603e-01
-4.27709281e-01 -6.90214276e-01 1.75338626e-01 6.85446739e-01
-6.38865113e-01 -5.93535781e-01 1.16934329e-01 5.90343237e-01
-8.36258158e-02 6.89541101e-01 1.33698091e-01 -3.85386288e-01
1.11163341e-01 -1.98342770e-01 2.05241337e-01 -5.28015256e-01
-4.68683064e-01 -7.68521056e-02 1.45766092e-02 -7.97955513e-01
-5.20787656e-01 -6.95043266e-01 -1.24876571e+00 -3.20958763e-01
-2.14342713e-01 -1.15262300e-01 5.19369721e-01 1.03934014e+00
8.50357413e-01 5.23102403e-01 3.08165729e-01 -9.49772894e-01
-3.85885239e-01 -8.05289030e-01 -4.70688343e-01 3.05124134e-01
3.83461654e-01 -7.27102697e-01 5.84028028e-02 1.87138572e-01] | [9.7319974899292, 0.08183370530605316] |
2fe44a56-7bd0-47a4-9a3f-22aab500d3cc | towards-theory-based-moral-ai-moral-ai-with | 2306.11432 | null | https://arxiv.org/abs/2306.11432v1 | https://arxiv.org/pdf/2306.11432v1.pdf | Towards Theory-based Moral AI: Moral AI with Aggregating Models Based on Normative Ethical Theory | Moral AI has been studied in the fields of philosophy and artificial intelligence. Although most existing studies are only theoretical, recent developments in AI have made it increasingly necessary to implement AI with morality. On the other hand, humans are under the moral uncertainty of not knowing what is morally right. In this paper, we implement the Maximizing Expected Choiceworthiness (MEC) algorithm, which aggregates outputs of models based on three normative theories of normative ethics to generate the most appropriate output. MEC is a method for making appropriate moral judgments under moral uncertainty. Our experimental results suggest that the output of MEC correlates to some extent with commonsense morality and that MEC can produce equally or more appropriate output than existing methods. | ['Kenji Araki', 'Rzepka Rafal', 'Masashi Takeshita'] | 2023-06-20 | null | null | null | null | ['ethics', 'philosophy'] | ['miscellaneous', 'miscellaneous'] | [ 2.16391549e-01 7.06244051e-01 -2.38309912e-02 -7.27825463e-01
1.27326205e-01 -2.68373966e-01 4.77353334e-01 -1.85627211e-02
-9.34433818e-01 9.06237125e-01 3.84040624e-01 -3.09840173e-01
-3.15141320e-01 -7.35265672e-01 -1.13548495e-01 -5.36996663e-01
5.63804567e-01 2.73267150e-01 -3.41718525e-01 -4.22619522e-01
8.03071976e-01 -2.67419294e-02 -1.37545455e+00 1.25233710e-01
1.46095514e+00 8.04817736e-01 -4.56419498e-01 3.10238630e-01
4.79764909e-01 1.05448747e+00 -3.97540897e-01 -1.06448579e+00
2.66033292e-01 -7.39435315e-01 -8.63161922e-01 -6.04075849e-01
-2.88026631e-02 -6.71096504e-01 4.07346636e-01 1.39309561e+00
2.38640472e-01 3.91453272e-03 1.11579406e+00 -1.62985873e+00
-1.00987375e+00 9.87900198e-01 5.45207746e-02 -3.27169120e-01
3.16107094e-01 5.07669032e-01 1.20206523e+00 -2.97175050e-02
4.91182566e-01 1.65031087e+00 2.26318464e-01 1.15389752e+00
-1.12762535e+00 -5.49864590e-01 -1.64867327e-01 3.46240908e-01
-7.07176208e-01 -1.79999411e-01 6.96348190e-01 -5.08128881e-01
6.39838099e-01 2.87840426e-01 8.48902524e-01 1.10112071e+00
7.97498345e-01 4.76468146e-01 1.56476653e+00 -3.11047345e-01
6.83148205e-01 3.11007768e-01 -6.78917691e-02 -2.05349848e-01
8.88952792e-01 4.84047681e-01 -2.08274648e-01 -5.36041558e-01
5.02607763e-01 -4.52345580e-01 1.28478914e-01 -7.53108412e-02
-1.23757350e+00 1.07386744e+00 1.41405731e-01 1.58969641e-01
-6.76515162e-01 4.60581720e-01 2.32873723e-01 4.52249110e-01
3.82105149e-02 1.00925970e+00 -1.72656015e-01 -2.93306112e-01
-3.51102680e-01 6.32555008e-01 8.76549304e-01 3.56417120e-01
1.27640232e-01 -8.11974183e-02 -1.32758364e-01 5.20302653e-01
8.52921903e-01 6.40833616e-01 -9.34668630e-03 -2.02214599e+00
-8.73434171e-02 7.23028839e-01 5.35321295e-01 -1.07230711e+00
-2.41254643e-01 1.47019312e-01 -3.28760028e-01 8.27475846e-01
4.35213327e-01 -3.17514688e-01 -7.51950294e-02 1.87438750e+00
1.99364990e-01 -7.16599286e-01 4.70815897e-01 9.47138488e-01
1.66601390e-01 3.42854500e-01 6.64498925e-01 -1.78611159e-01
1.23177397e+00 -1.98600203e-01 -9.45880413e-01 -4.30695891e-01
6.56033218e-01 -4.38089430e-01 1.04553282e+00 6.02967083e-01
-8.50290656e-01 4.27255332e-01 -8.02778006e-01 -3.72231565e-02
1.77021906e-01 -4.51666296e-01 7.29641974e-01 8.31123352e-01
-8.65546286e-01 7.88858891e-01 -5.46265999e-03 -3.82878244e-01
4.83487040e-01 -5.66390418e-02 -1.30335465e-01 3.55160892e-01
-1.53957617e+00 1.51657724e+00 3.19564670e-01 1.09460637e-01
-3.71794432e-01 -8.03328604e-02 -6.61739945e-01 4.31745164e-02
3.36334229e-01 -6.75947189e-01 1.19397485e+00 -1.48495591e+00
-1.43623197e+00 9.84555185e-01 4.93010253e-01 -5.99554598e-01
8.57336223e-01 -3.21378440e-01 -1.83963299e-01 4.45709899e-02
2.19691440e-01 8.15260768e-01 3.12947899e-01 -1.09776735e+00
-4.93174046e-01 -3.59945118e-01 3.05616707e-01 5.30833721e-01
-1.44694403e-01 4.40021008e-01 1.03021669e+00 -3.34473968e-01
-3.32135886e-01 -9.43664968e-01 -3.41657072e-01 1.99128807e-01
-2.32015371e-01 -7.29336500e-01 -1.99692175e-02 -1.84475914e-01
1.18434846e+00 -1.96698534e+00 -2.26113930e-01 1.39100000e-01
-6.94287336e-03 -1.52067304e-01 2.72956520e-01 3.86571556e-01
3.51636678e-01 4.76039082e-01 -3.78106564e-01 4.86175656e-01
6.88913047e-01 2.06500709e-01 -2.18450084e-01 4.27163392e-01
-3.84778157e-02 5.62473655e-01 -9.66150224e-01 -8.95090818e-01
1.49651289e-01 1.01933412e-01 -8.96300614e-01 1.90267265e-01
-1.19779505e-01 6.59240559e-02 -6.15519941e-01 3.79602522e-01
4.29656774e-01 7.81010613e-02 4.41107899e-01 4.24843997e-01
1.41167436e-02 3.38215560e-01 -9.10780251e-01 6.72369599e-01
2.45209739e-01 5.26269376e-01 -8.16153437e-02 -6.15620017e-01
9.01278973e-01 4.04527515e-01 -6.35159612e-02 -5.64025223e-01
7.68151581e-01 4.49776590e-01 4.53707427e-01 -7.33570278e-01
4.61622953e-01 -7.59176731e-01 -3.91954780e-01 9.42099214e-01
-6.78577363e-01 -5.38879633e-01 -1.86469108e-02 1.00629583e-01
5.76855004e-01 5.70656359e-01 7.06085503e-01 -6.35337293e-01
3.07364732e-01 2.13070109e-01 1.13927102e+00 6.20202899e-01
-9.48933542e-01 -5.70986792e-02 9.12993312e-01 -6.31302834e-01
-9.30472493e-01 -7.98109293e-01 -3.32228124e-01 1.00187290e+00
3.34021986e-01 3.27828944e-01 -1.22312033e+00 -4.89829957e-01
2.06054285e-01 2.14564705e+00 -5.86487949e-01 -4.10905272e-01
-3.41775596e-01 -7.02785432e-01 5.73393404e-01 2.64934003e-01
2.94542938e-01 -1.58321202e+00 -1.51389217e+00 -6.25284910e-02
-2.42979929e-01 -7.81491518e-01 -1.80186017e-03 -1.30454242e-01
-6.95352077e-01 -1.02377379e+00 -9.26388875e-02 -7.50015453e-02
7.26297855e-01 -2.80995965e-01 8.55900407e-01 2.88362205e-01
2.41795614e-01 2.16324881e-01 -3.21914256e-01 -7.51661837e-01
-8.95278096e-01 -8.22970986e-01 2.83320516e-01 -2.24182039e-01
9.39054489e-01 -4.26238865e-01 -6.31974399e-01 2.03025952e-01
-8.42791140e-01 8.66253376e-02 2.78375119e-01 3.43497634e-01
-2.79983431e-01 -2.59333372e-01 1.01466799e+00 -8.30855012e-01
1.32927954e+00 -6.10223472e-01 -3.38723153e-01 1.83558136e-01
-1.23439181e+00 1.69732705e-01 5.91068089e-01 -8.07375014e-02
-1.38609838e+00 -5.80186903e-01 1.96019143e-01 5.24514854e-01
-3.84594440e-01 -9.27666854e-03 -1.20390475e-01 2.87189335e-01
8.14455509e-01 -4.40886796e-01 1.19749956e-01 9.11318809e-02
2.78192937e-01 1.03002238e+00 1.07312508e-01 -9.27805305e-01
3.94070715e-01 1.65214956e-01 -3.71485688e-02 -2.58030087e-01
-6.76724315e-01 4.78445292e-01 -1.70396298e-01 -7.19923615e-01
1.07494903e+00 -3.73659790e-01 -1.27400970e+00 2.05332533e-01
-9.23169732e-01 -1.58794507e-01 -2.37523153e-01 7.77116358e-01
-9.24309850e-01 5.58266997e-01 -3.33066255e-01 -1.24648380e+00
-4.10752356e-01 -1.04122376e+00 2.21956998e-01 3.86352420e-01
-9.31076586e-01 -7.15277612e-01 -1.04578316e-01 6.03531122e-01
3.59677017e-01 4.41275477e-01 1.01602793e+00 -7.51707792e-01
-2.17708535e-02 -1.58712983e-01 1.06796384e-01 5.61223149e-01
-3.20327103e-01 2.93203324e-01 -6.78626239e-01 4.36634928e-01
3.27234685e-01 -5.41767359e-01 2.20413312e-01 4.04879004e-01
4.40036565e-01 -7.46939421e-01 1.07537568e-01 7.41105676e-02
1.23157310e+00 6.57350898e-01 8.65311861e-01 8.72813702e-01
-1.58986360e-01 1.40204132e+00 9.88120735e-01 5.32070994e-01
8.70792091e-01 2.50400662e-01 6.03381515e-01 6.52659893e-01
8.04258883e-01 -1.30596012e-02 4.65126783e-01 -2.85488879e-03
-4.40856755e-01 -2.38660141e-03 -1.01108718e+00 4.49935943e-01
-2.06129575e+00 -1.42762566e+00 -1.05926422e-02 2.06172371e+00
1.10196352e+00 3.72260064e-02 -9.40085128e-02 -5.93986874e-03
7.97249138e-01 -2.16085136e-01 -4.11385059e-01 -1.45359242e+00
-8.18605125e-02 -6.71123862e-01 1.83596820e-01 4.56676871e-01
-6.74238026e-01 9.42106128e-01 7.02033758e+00 1.78036943e-01
-3.01073104e-01 -2.83302605e-01 1.00226176e+00 -1.38898522e-01
-1.01209140e+00 3.85640740e-01 -1.95150778e-01 6.45179749e-01
5.87543309e-01 -6.96076035e-01 2.68713564e-01 1.12448311e+00
2.69863278e-01 -6.59815848e-01 -1.36329758e+00 4.59252268e-01
-2.18947291e-01 -5.52896082e-01 -1.39088675e-01 1.30514473e-01
5.61017513e-01 -9.23795998e-01 -6.97667971e-02 5.78784123e-02
8.71305645e-01 -1.30369520e+00 1.10091460e+00 6.06573820e-01
2.49014497e-01 -8.85300517e-01 1.14631927e+00 6.02249086e-01
7.23663792e-02 -2.27124736e-01 -5.52056193e-01 -7.36018121e-01
1.78620264e-01 3.69892657e-01 -4.36408937e-01 -1.80463031e-01
7.41719723e-01 2.29552150e-01 2.16298234e-02 4.69882309e-01
-7.13216782e-01 3.14642102e-01 -1.69159412e-01 -6.39669776e-01
2.95511454e-01 -7.07628071e-01 4.71116066e-01 7.91525960e-01
6.61231130e-02 3.64522398e-01 -8.05324078e-01 1.42358363e+00
2.62741566e-01 1.55994013e-01 -7.70783961e-01 -6.74693882e-02
6.91160798e-01 9.81350839e-01 -5.12859285e-01 -8.49560723e-02
3.26381736e-02 5.35854995e-01 -7.79248476e-02 1.40134513e-01
-7.70830989e-01 -2.22464696e-01 7.26277828e-01 -3.14604849e-01
-6.14693224e-01 3.80281031e-01 -9.33691740e-01 -8.00788462e-01
-1.64072677e-01 -9.47414756e-01 3.17319959e-01 -7.00963080e-01
-1.25525582e+00 1.51120991e-01 2.17188135e-01 -8.37964416e-01
-4.25500721e-01 -6.89748287e-01 -5.82510710e-01 5.16454816e-01
-9.27762866e-01 -5.70246100e-01 1.54820189e-01 -9.05609652e-02
-2.45703578e-01 1.27172992e-01 8.15623403e-01 -4.70300049e-01
-2.38440722e-01 1.50717691e-01 -1.25881806e-01 -6.16159588e-02
9.27919865e-01 -1.20879841e+00 8.70072395e-02 5.12605548e-01
-1.01901734e+00 7.30009079e-01 1.04884458e+00 -7.24210918e-01
-9.44516242e-01 -1.58469945e-01 1.38324034e+00 -4.67079312e-01
5.15676200e-01 1.30855620e-01 -5.21871090e-01 6.08938634e-01
5.93617558e-01 -6.86446309e-01 1.03071225e+00 -2.13628650e-01
-3.11757863e-01 -1.14191538e-02 -1.88845718e+00 1.21893585e+00
8.51392388e-01 -1.72769085e-01 -1.46538591e+00 -8.94278511e-02
5.61428368e-01 4.22305971e-01 -7.81584322e-01 2.52360523e-01
8.69421899e-01 -1.29179001e+00 6.19860590e-01 -5.80071151e-01
7.07410932e-01 -1.04782119e-01 -3.49781275e-01 -1.13446987e+00
-2.80716866e-01 -5.81304848e-01 7.63233125e-01 1.00126386e+00
2.67949164e-01 -1.07867002e+00 2.97961682e-01 1.48674881e+00
3.94669801e-01 -6.34797752e-01 -1.16790962e+00 -4.60121691e-01
6.93952084e-01 -4.04423177e-01 8.06410253e-01 1.13213646e+00
7.07391798e-01 1.77157924e-01 -5.18873274e-01 -3.99267882e-01
1.15760589e+00 -9.73126292e-03 3.26787591e-01 -1.45664489e+00
3.14581156e-01 -8.45142543e-01 -1.31366864e-01 -8.48108083e-02
2.05763355e-01 -6.47458792e-01 2.52033085e-01 -1.80224693e+00
3.20323795e-01 -2.94540137e-01 -1.98099896e-01 4.43359822e-01
-3.82397592e-01 -3.09319943e-01 4.46617693e-01 1.69351071e-01
-4.99714822e-01 5.92177212e-01 1.21685874e+00 3.97385448e-01
7.18460083e-02 -4.23377454e-01 -1.44518661e+00 1.19487584e+00
1.24439275e+00 -5.78229129e-01 -1.85854465e-01 -1.41001835e-01
1.03627706e+00 -1.47192672e-01 4.37385917e-01 -5.37559748e-01
-1.53808728e-01 -1.29560852e+00 6.42022118e-02 1.36931926e-01
-1.53315425e-01 -9.67026174e-01 1.59661755e-01 6.72873974e-01
-8.36789846e-01 -1.64555937e-01 -3.01406920e-01 8.34286883e-02
2.68898070e-01 -6.70504689e-01 1.02693141e+00 -1.91742644e-01
-4.26967114e-01 -4.50874984e-01 -8.44221950e-01 2.66956538e-01
1.41093767e+00 -3.59353542e-01 -5.81672132e-01 -4.83563274e-01
-1.30126730e-01 4.13086891e-01 1.01035142e+00 1.75282225e-01
5.72516441e-01 -1.30253065e+00 -8.91835392e-01 -5.33663034e-01
7.61896223e-02 -6.01009667e-01 3.49281393e-02 6.95769310e-01
-7.53141999e-01 3.01002115e-01 -6.93886042e-01 1.76869422e-01
-7.66095877e-01 3.59249562e-01 5.60945392e-01 4.12562639e-01
-4.44283694e-01 6.27749562e-01 3.22381824e-01 -4.82672513e-01
-4.24846299e-02 2.19375521e-01 -6.22148812e-01 -6.89120218e-02
4.03858989e-01 7.54573524e-01 -1.08908904e+00 -7.09848940e-01
-5.45173109e-01 4.82972950e-01 3.90456647e-01 -3.60735357e-01
1.22872448e+00 -5.30455634e-02 -6.02377534e-01 4.15767014e-01
1.43099517e-01 -2.32139558e-01 -9.99416530e-01 3.99272531e-01
2.90352017e-01 -6.52544796e-01 -2.00398922e-01 -1.07005847e+00
-3.27922702e-01 7.71160185e-01 1.80116609e-01 2.65478134e-01
9.26750302e-01 -4.20739412e-01 2.38089219e-01 4.98060077e-01
6.42360628e-01 -2.05270362e+00 -2.80793250e-01 9.81850028e-02
8.80583763e-01 -1.22207427e+00 8.81179199e-02 -6.64467663e-02
-1.52140725e+00 1.03447318e+00 8.18456113e-01 -1.96826994e-01
2.82419860e-01 7.25337565e-02 4.22720313e-01 -4.08611819e-03
-1.00296390e+00 2.49069318e-01 -1.67520121e-01 5.98715365e-01
6.11185014e-01 7.40561545e-01 -1.44684100e+00 7.81100273e-01
-4.25533682e-01 1.91015974e-01 9.83022153e-01 9.30776477e-01
-7.79076159e-01 -9.66193199e-01 -5.80251813e-01 4.01575059e-01
-7.79855132e-01 -6.19149879e-02 -8.96709979e-01 5.55781066e-01
8.21600854e-02 1.47892845e+00 -2.04658031e-01 -2.52659917e-01
2.57502645e-02 -1.66479610e-02 6.63429871e-02 -1.74715325e-01
-8.52200687e-01 -2.71808296e-01 4.08550858e-01 -7.40848601e-01
-5.43957889e-01 -7.53417134e-01 -1.50672913e+00 -7.59123147e-01
4.67033572e-02 2.20553324e-01 2.81226963e-01 8.98456752e-01
1.38574168e-01 -8.40179920e-02 3.30946594e-01 -1.73951402e-01
-1.13723457e+00 -7.52418637e-01 -6.60181046e-01 6.24414504e-01
-1.35714740e-01 -5.68625093e-01 -8.16724658e-01 -2.84774214e-01] | [9.024393081665039, 6.226876735687256] |
ba0461e4-d04a-45fa-8459-585db2dd33b8 | integration-of-feature-selection-techniques | 2303.02467 | null | https://arxiv.org/abs/2303.02467v1 | https://arxiv.org/pdf/2303.02467v1.pdf | Integration of Feature Selection Techniques using a Sleep Quality Dataset for Comparing Regression Algorithms | This research aims to examine the usefulness of integrating various feature selection methods with regression algorithms for sleep quality prediction. A publicly accessible sleep quality dataset is used to analyze the effect of different feature selection techniques on the performance of four regression algorithms - Linear regression, Ridge regression, Lasso Regression and Random Forest Regressor. The results are compared to determine the optimal combination of feature selection techniques and regression algorithms. The conclusion of the study enriches the current literature on using machine learning for sleep quality prediction and has practical significance for personalizing sleep recommendations for individuals. | ['Venkat Tummala', 'Sai Rohith Tanuku'] | 2023-03-04 | null | null | null | null | ['sleep-quality-prediction', 'sleep-quality-prediction-1'] | ['medical', 'medical'] | [-1.55836076e-01 -3.12103003e-01 -9.26077962e-01 -8.67684186e-01
-3.00215453e-01 3.07724684e-01 -3.15491289e-01 3.56975496e-01
-4.31535631e-01 1.14780486e+00 6.73625231e-01 -1.69714570e-01
-5.33900440e-01 -6.13978863e-01 3.05592716e-01 -5.26214719e-01
1.06197812e-01 2.70605326e-01 -4.94449824e-01 -8.54168907e-02
4.91010219e-01 4.08323199e-01 -1.71072698e+00 -1.09762259e-01
1.13103008e+00 7.73669779e-01 -1.92910939e-01 4.22811270e-01
1.47602633e-01 4.58064735e-01 -6.95335805e-01 -1.88544124e-01
2.06173122e-01 -4.66848940e-01 -4.11733866e-01 -3.40522856e-01
2.46837527e-01 4.62555028e-02 -7.88788870e-02 4.04956341e-01
7.03676522e-01 6.25006318e-01 4.52689826e-01 -1.25257969e+00
-7.66452193e-01 4.36236680e-01 6.74677566e-02 9.29544568e-01
7.58222938e-01 1.61752284e-01 9.79335427e-01 -3.39316130e-01
3.70528810e-02 8.92341614e-01 8.84600699e-01 5.84589601e-01
-1.41323364e+00 -1.05732763e+00 -2.53152281e-01 5.25552690e-01
-1.45489419e+00 -8.68132949e-01 5.45659363e-01 -2.51048386e-01
1.38296258e+00 6.82102621e-01 1.28037155e+00 6.52371407e-01
5.93277872e-01 -2.96506099e-02 1.57286155e+00 -4.24858451e-01
2.99564779e-01 4.39092845e-01 9.64462876e-01 8.02563846e-01
6.22750878e-01 4.33943808e-01 -1.01708567e+00 -4.51362103e-01
7.06571341e-02 3.64902437e-01 2.57960588e-01 4.41860974e-01
-5.19400001e-01 1.14147019e+00 2.60492355e-01 1.79227978e-01
-2.86224961e-01 -2.69279391e-01 3.18312109e-01 5.11614442e-01
7.93023825e-01 6.30206108e-01 -5.74035466e-01 -2.38055840e-01
-1.70808244e+00 1.36340648e-01 7.65760064e-01 2.88987219e-01
6.63388550e-01 2.44375825e-01 -4.73973155e-01 9.11065876e-01
6.39223039e-01 6.82287574e-01 7.84165621e-01 -9.19207990e-01
-3.80066857e-02 8.49587440e-01 1.86046496e-01 -9.56490040e-01
-1.15415084e+00 -3.09235930e-01 -5.23770809e-01 -2.02571332e-01
-1.92308709e-01 -2.05680296e-01 -2.88693309e-01 9.38121438e-01
-2.42596809e-02 -7.84842223e-02 -3.15344363e-01 5.18972158e-01
1.17794311e+00 1.60545304e-01 3.33032846e-01 -8.56213212e-01
1.08707249e+00 -1.07643199e+00 -9.55381274e-01 -2.35322610e-01
2.90233254e-01 -3.76617283e-01 1.09239900e+00 5.44795692e-01
-1.15418136e+00 -5.40074527e-01 -8.01604390e-01 -6.54563233e-02
-3.98209125e-01 2.64077455e-01 8.79889727e-01 1.30805850e+00
-9.61373210e-01 6.35117531e-01 -9.24195886e-01 -5.65986574e-01
3.98140311e-01 9.87781465e-01 -7.99929723e-02 2.80212224e-01
-7.61814535e-01 1.05590451e+00 -3.25493008e-01 -1.16568573e-01
-2.34803841e-01 -8.20915997e-01 -7.84416318e-01 -7.79424161e-02
-3.45203042e-01 -1.27204251e+00 5.88044226e-01 -1.03211141e+00
-1.51530993e+00 6.28925681e-01 -8.97570133e-01 -4.60021764e-01
-2.75201082e-01 -3.96081954e-01 -8.20899606e-01 -3.65632385e-01
3.78922641e-01 -4.81967488e-03 7.22399414e-01 -3.41100603e-01
-5.94872653e-01 -8.61111104e-01 -5.68371177e-01 2.80027036e-02
-1.26019463e-01 4.12441522e-01 4.30916518e-01 -4.13804144e-01
-3.38137269e-01 -8.80330741e-01 -4.83394921e-01 -6.36675358e-01
-3.71399820e-01 -6.43389106e-01 -1.18219899e-02 -5.75109720e-01
1.96294653e+00 -2.02421284e+00 -5.35567626e-02 2.31200203e-01
6.38393909e-02 -3.08787465e-01 4.02667552e-01 2.82062024e-01
2.42831632e-01 3.67118210e-01 2.46035904e-01 -5.83768129e-01
-2.67300457e-01 3.95770639e-01 3.36722851e-01 7.37970829e-01
-3.14507753e-01 7.22083271e-01 -4.30692375e-01 -4.52937275e-01
4.16992188e-01 3.89959335e-01 -6.25681579e-01 3.35658967e-01
5.79076827e-01 4.44906414e-01 -4.98641491e-01 7.63817430e-01
4.97814715e-02 -8.95852372e-02 -2.15582043e-01 3.56948860e-02
-7.33489692e-01 8.86086464e-01 -7.05077171e-01 1.14893997e+00
-3.65137696e-01 5.17204463e-01 -4.55657035e-01 -5.80784976e-01
1.24893153e+00 -1.84206292e-01 5.44166565e-01 -6.04986310e-01
5.82635701e-02 -2.31210530e-01 -7.37655833e-02 -6.89895511e-01
3.35328996e-01 -5.10893285e-01 1.44430131e-01 3.28489393e-01
-5.94951846e-02 1.94541901e-01 1.55335097e-02 -4.48570430e-01
9.70514894e-01 -1.75548688e-01 8.87014747e-01 -4.06099349e-01
6.49891496e-01 -1.64749801e-01 9.83886838e-01 6.46423042e-01
-3.83212209e-01 6.71692491e-02 1.65806841e-02 -6.13378048e-01
-3.15110445e-01 -7.14249611e-01 -3.96265954e-01 1.37763965e+00
-3.37195903e-01 -1.01279843e+00 -8.98403674e-02 -7.30947375e-01
1.05012909e-01 9.97767925e-01 -1.04218102e+00 -6.14454091e-01
-1.51610197e-02 -1.07473481e+00 3.12750697e-01 4.41193074e-01
1.39945179e-01 -9.28678751e-01 -7.60916412e-01 -2.30003409e-02
-8.11951011e-02 -2.48933390e-01 -2.40047157e-01 4.63867992e-01
-1.26070642e+00 -1.03922701e+00 3.05663079e-01 -1.45792454e-01
3.74553591e-01 4.71326202e-01 1.48157310e+00 5.90377212e-01
-2.52873272e-01 4.74177629e-01 -4.61614996e-01 -5.10646164e-01
1.04865648e-01 2.11743206e-01 5.15842497e-01 -4.41052556e-01
1.08203197e+00 -8.87871385e-01 -8.55184376e-01 3.84448737e-01
-1.25985578e-01 -7.10872471e-01 4.16163296e-01 5.88093758e-01
6.91763461e-01 2.60384232e-01 7.43687212e-01 -8.48330259e-01
6.37989402e-01 -8.16763818e-01 -3.66970748e-02 -4.05328311e-02
-1.66598332e+00 -2.00580090e-01 3.38377088e-01 -1.13893703e-01
-7.44208753e-01 -1.78709060e-01 -1.85202032e-01 2.42606312e-01
-2.84803480e-01 1.78456739e-01 8.52878764e-02 -1.49456635e-01
7.31838644e-01 2.55921394e-01 -4.63738516e-02 -5.86072862e-01
-1.25076205e-01 6.40500665e-01 -2.24696055e-01 1.67893559e-01
5.38316190e-01 2.86698610e-01 1.59290910e-01 -1.18103755e+00
-1.10406721e+00 -7.90308297e-01 -8.68372440e-01 4.41610441e-02
9.15767610e-01 -7.73483813e-01 -6.07190847e-01 -3.62679929e-01
-1.55072426e-03 -3.43251318e-01 -4.81003284e-01 6.20693207e-01
-4.19909805e-01 5.70717379e-02 8.00355226e-02 -8.99904609e-01
-1.00865126e+00 -9.13115203e-01 5.15169740e-01 5.09692967e-01
-7.34750211e-01 -1.07630980e+00 5.96136570e-01 9.33482289e-01
4.75344688e-01 -1.33869603e-01 8.04508030e-01 -8.47491860e-01
3.31256092e-01 1.40718132e-01 4.42775100e-01 6.25707135e-02
4.50646788e-01 2.92634875e-01 -6.39398217e-01 -9.86097828e-02
1.03737712e-02 -1.98818818e-01 6.56453073e-01 1.00475323e+00
7.70400047e-01 -4.96000946e-01 -3.65410358e-01 7.41441131e-01
1.22490203e+00 -7.52194002e-02 5.19611180e-01 7.91298926e-01
9.95613262e-02 4.28088993e-01 6.94589972e-01 8.29723001e-01
6.97595775e-01 7.21802115e-01 -1.12156697e-01 1.20363511e-01
-8.10685232e-02 1.45150796e-01 4.84879971e-01 3.86130929e-01
-3.51142257e-01 4.77120489e-01 -2.53137201e-01 -9.64550599e-02
-1.48032260e+00 -9.92455661e-01 -2.05499321e-01 2.27849436e+00
4.20475572e-01 -1.01683423e-01 1.23865998e+00 1.52366385e-01
2.16385107e-02 -1.15001835e-01 -3.70348245e-01 -1.00759196e+00
6.22760653e-02 4.99165773e-01 4.34551716e-01 9.38487127e-02
-8.22197735e-01 5.94436347e-01 8.46875191e+00 -7.84690827e-02
-7.94700503e-01 1.95812747e-01 3.49244773e-01 -7.24069238e-01
5.51103242e-02 -2.00553298e-01 -1.08925378e+00 6.71109617e-01
1.80490637e+00 -1.26842111e-01 7.45830953e-01 7.66535163e-01
1.24556279e+00 -2.22426593e-01 -7.64222145e-01 7.53940761e-01
7.55696073e-02 -9.02342796e-01 -4.86591548e-01 1.19911760e-01
3.47575456e-01 1.23416327e-01 9.90986601e-02 5.45785844e-01
1.66366547e-01 -1.28871715e+00 1.37189940e-01 9.08933818e-01
4.85059261e-01 -6.36861205e-01 8.21754992e-01 -1.15893662e-01
-7.61476815e-01 -6.37930155e-01 -6.17572308e-01 -6.75419748e-01
-1.43828899e-01 5.02343357e-01 -8.25058997e-01 1.57171488e-01
1.13842165e+00 1.11803663e+00 -1.26548123e+00 1.27494848e+00
4.36087102e-02 1.22725022e+00 -8.87600705e-02 -2.21990690e-01
-4.18740332e-01 -4.72696871e-01 1.68735951e-01 1.20781136e+00
3.56798023e-02 1.43794253e-01 -3.30088362e-02 7.23795831e-01
4.68300700e-01 5.96523762e-01 -3.96823049e-01 3.20650548e-01
4.39629793e-01 1.13559377e+00 -4.98589337e-01 2.64520675e-01
-9.66832519e-01 4.86060679e-01 2.39724562e-01 -1.33372322e-01
-3.52665335e-01 2.26989761e-01 7.16655791e-01 5.24340153e-01
6.81405813e-02 -4.29520868e-02 -1.00262260e+00 -8.80842566e-01
-7.43618906e-01 -7.85563648e-01 8.81108463e-01 -4.25278008e-01
-1.42237544e+00 1.52937144e-01 -9.41634774e-02 -7.28749573e-01
3.18168074e-01 2.46159166e-01 -9.58766460e-01 6.32824421e-01
-1.18967450e+00 -1.01960135e+00 -3.31668794e-01 6.88192904e-01
6.18240416e-01 -3.73138070e-01 9.81738031e-01 1.14935711e-01
-1.22049773e+00 7.01448202e-01 1.60107110e-02 -5.89443505e-01
7.82761991e-01 -1.34426141e+00 -2.92401433e-01 2.02138752e-01
-2.74060331e-02 9.74418640e-01 7.14086294e-01 -8.78051579e-01
-1.16078854e+00 -1.05372632e+00 1.02221489e+00 -5.98778367e-01
8.77487436e-02 2.35958099e-01 -4.56198543e-01 6.95350826e-01
4.63894755e-02 -4.28327650e-01 2.00736761e+00 9.71928179e-01
2.63273895e-01 -6.09496653e-01 -1.50234342e+00 -2.77691968e-02
4.32827562e-01 7.70609751e-02 -6.85671508e-01 1.06679671e-01
2.57423550e-01 3.57112974e-01 -1.35141218e+00 1.24132350e-01
9.05586243e-01 -1.30893528e+00 1.07817960e+00 -9.44489062e-01
1.46396562e-01 1.02690943e-01 -1.77109987e-02 -1.08761692e+00
-1.00757086e+00 -5.00195384e-01 -2.09343895e-01 1.24119163e+00
7.02860475e-01 -4.52303201e-01 8.79236162e-01 1.29446328e+00
3.50574148e-03 -9.47434008e-01 -1.13907540e+00 -3.29341620e-01
-9.12249386e-02 -9.81724337e-02 6.41448438e-01 5.95361948e-01
1.62438273e-01 6.97074294e-01 -5.30605674e-01 -3.68824713e-02
2.98462063e-01 1.14660092e-01 7.67915905e-01 -1.56595850e+00
-4.59432267e-02 -2.37153098e-01 -4.76091146e-01 5.03148474e-02
4.86006916e-01 -6.70482457e-01 -4.89271134e-01 -1.52653611e+00
4.41901058e-01 -4.99801427e-01 -7.82842159e-01 5.97319782e-01
-4.91462767e-01 1.69366241e-01 -9.20559391e-02 3.14475536e-01
-6.81605756e-01 4.55719382e-01 5.95752835e-01 7.58193657e-02
-1.14230204e+00 1.06856978e+00 -1.17927635e+00 5.33154964e-01
1.15309262e+00 -8.25287044e-01 -3.97810012e-01 3.90078306e-01
2.82435834e-01 8.72471184e-03 -9.27686706e-05 -8.20657611e-01
-2.75338050e-02 -5.19620776e-01 7.64014184e-01 -2.93566257e-01
2.97347397e-01 -8.52309704e-01 2.15624198e-01 5.08270442e-01
-5.24500787e-01 4.77112710e-01 -1.33471927e-02 5.32755256e-01
5.98830163e-01 -3.22008729e-01 6.19304717e-01 3.13366554e-03
-1.34673417e-01 -1.23917103e-01 -7.05026150e-01 -1.97049275e-01
6.93452597e-01 -4.44328487e-01 -5.16411103e-02 -3.65906954e-01
-1.24615073e+00 1.44970283e-01 3.49042922e-01 1.86672598e-01
8.62693369e-01 -9.14342344e-01 -5.76792657e-01 4.92902696e-01
-1.26376927e-01 -9.42421496e-01 -1.40683413e-01 1.27078891e+00
3.99866402e-02 5.33473909e-01 -5.01916111e-01 6.92895874e-02
-1.75960004e+00 6.26023293e-01 4.50753868e-01 -1.40491173e-01
-5.13108909e-01 5.97366869e-01 -8.17308724e-01 -4.00330760e-02
8.80751684e-02 -1.23733871e-01 -6.57158792e-01 -1.33348927e-01
6.15631461e-01 1.55746698e+00 4.46972884e-02 -5.97744286e-01
-6.80540621e-01 4.02853668e-01 2.98654050e-01 5.56708455e-01
1.58276272e+00 -5.62515140e-01 -5.01560390e-01 8.82570028e-01
7.99299300e-01 4.31962937e-01 -4.23566431e-01 3.96810293e-01
9.07515287e-02 -6.06183529e-01 3.37555200e-01 -9.98116732e-01
-8.90157163e-01 2.94417232e-01 1.19903576e+00 2.43871093e-01
1.30559182e+00 -2.42680252e-01 4.27430928e-01 3.54925007e-01
-2.02104077e-02 -9.15470302e-01 -3.57644320e-01 -1.32656127e-01
4.18788850e-01 -1.48967862e+00 8.76608610e-01 -7.29561895e-02
-7.28430748e-01 1.17489004e+00 4.82910633e-01 -2.99085379e-01
9.59282398e-01 -2.39278048e-01 -2.43069202e-01 -3.38932812e-01
-8.09677660e-01 -1.22461282e-01 5.74266195e-01 8.88333201e-01
7.81739473e-01 3.09795231e-01 -9.22439933e-01 1.12303483e+00
-8.61976624e-01 2.86389053e-01 4.38179165e-01 4.63563114e-01
-6.36238515e-01 -1.34043109e+00 -2.61407018e-01 1.37713182e+00
-9.56368029e-01 -2.76149511e-01 -7.16673017e-01 5.14003932e-01
2.97920048e-01 1.80617213e+00 -1.45634845e-01 -4.74368036e-01
3.72856915e-01 2.59190410e-01 7.42450133e-02 -9.38266277e-01
-1.33487296e+00 -1.24340154e-01 4.29062575e-01 -7.22080827e-01
-7.28915155e-01 -1.16689694e+00 -8.81892681e-01 -2.96942145e-01
-5.61587632e-01 6.97058499e-01 2.94463128e-01 1.06265116e+00
4.06918198e-01 4.12260771e-01 7.81155646e-01 -3.43857706e-01
-7.31852800e-02 -9.93198991e-01 -7.54080355e-01 -1.55344710e-01
6.42834485e-01 -9.09435451e-01 -2.85504192e-01 -2.18127713e-01] | [13.564269065856934, 3.4629387855529785] |
c80dedf6-c62d-4503-9843-7cf72f6a8607 | a-parameter-free-graph-reduction-for-spectral | 2302.13165 | null | https://arxiv.org/abs/2302.13165v1 | https://arxiv.org/pdf/2302.13165v1.pdf | A parameter-free graph reduction for spectral clustering and SpectralNet | Graph-based clustering methods like spectral clustering and SpectralNet are very efficient in detecting clusters of non-convex shapes. Unlike the popular $k$-means, graph-based clustering methods do not assume that each cluster has a single mean. However, these methods need a graph where vertices in the same cluster are connected by edges of large weights. To achieve this goal, many studies have proposed graph reduction methods with parameters. Unfortunately, these parameters have to be tuned for every dataset. We introduce a graph reduction method that does not require any parameters. First, the distances from every point $p$ to its neighbors are filtered using an adaptive threshold to only keep neighbors with similar surrounding density. Second, the similarities with close neighbors are computed and only high similarities are kept. The edges that survive these two filtering steps form the constructed graph that was passed to spectral clustering and SpectralNet. The experiments showed that our method provides a stable alternative, where other methods performance fluctuated according to the setting of their parameters. | ['Masahiro Takatsuka', 'John Stavrakakis', 'Mashaan Alshammari'] | 2023-02-25 | null | null | null | null | ['graph-clustering', 'spectral-graph-clustering', 'graph-partitioning'] | ['graphs', 'graphs', 'graphs'] | [-3.55484396e-01 -9.37045664e-02 3.05774193e-02 -2.20085904e-01
-1.39317229e-01 -5.33096552e-01 3.04863095e-01 5.99058390e-01
-4.44836915e-01 3.53764027e-01 -2.88706869e-01 -1.44665567e-02
-4.84267652e-01 -1.21306026e+00 -2.63192534e-01 -8.72613728e-01
-5.07095993e-01 5.78180730e-01 7.57604718e-01 4.90598530e-02
4.14210230e-01 5.51452816e-01 -1.40444028e+00 -1.58147335e-01
1.07789171e+00 6.49173677e-01 2.04690948e-01 2.76242703e-01
-4.46150899e-01 9.00097862e-02 -3.95222276e-01 -7.74230212e-02
4.86666471e-01 -6.53268933e-01 -4.76539910e-01 2.11598828e-01
-8.99208114e-02 4.41936672e-01 -1.34396240e-01 1.30274343e+00
3.88346046e-01 4.50445801e-01 7.29912460e-01 -1.25345480e+00
-3.65589797e-01 8.51423085e-01 -1.05079162e+00 -2.38711387e-02
3.74765366e-01 -5.42915240e-02 7.84420609e-01 -6.72157824e-01
6.28825188e-01 1.15998924e+00 7.75942683e-01 9.71848145e-02
-1.23663092e+00 -6.09520733e-01 2.48574063e-01 2.36939952e-01
-1.94421375e+00 -4.01776396e-02 9.19679046e-01 -3.60646993e-01
5.39194465e-01 2.85152048e-01 8.39782357e-01 1.24193862e-01
-3.18869859e-01 -3.45935151e-02 8.90962303e-01 -5.53988516e-01
4.29451436e-01 2.03831196e-01 2.09109008e-01 8.05670261e-01
4.48410064e-01 -6.15213752e-01 7.73729011e-03 -2.81948149e-01
3.96522284e-01 -1.16786910e-02 -2.94705600e-01 -7.63010621e-01
-1.05732787e+00 9.37526524e-01 6.41007066e-01 8.45328450e-01
-2.72517294e-01 5.11014387e-02 1.78854421e-01 5.09036593e-02
2.31910750e-01 2.19671309e-01 -9.27740056e-03 2.49238044e-01
-9.89990890e-01 -2.02440813e-01 7.54251897e-01 8.23859632e-01
1.24860120e+00 -3.43887925e-01 4.03175443e-01 9.50142980e-01
4.18601424e-01 3.19012254e-01 3.17952454e-01 -7.13658392e-01
2.03015730e-01 1.26130164e+00 -2.64708579e-01 -1.84068644e+00
-5.70233822e-01 -1.08457528e-01 -1.06816983e+00 3.29116881e-02
3.50531846e-01 1.18720415e-03 -8.79772544e-01 1.43148470e+00
6.11292243e-01 3.65015090e-01 -3.07957411e-01 8.24819505e-01
7.70018041e-01 5.87588012e-01 4.41057086e-02 -5.40131927e-01
1.00275445e+00 -4.79712218e-01 -5.97065032e-01 2.73867518e-01
5.87622285e-01 -8.76421630e-01 1.01192629e+00 1.77186802e-01
-1.03381765e+00 -4.09994155e-01 -7.61268973e-01 6.81646943e-01
-5.44546068e-01 -6.71418980e-02 4.51675951e-01 8.69913995e-01
-1.38205850e+00 6.82900965e-01 -7.82582343e-01 -6.15478337e-01
4.71321233e-02 5.20489991e-01 -2.76477218e-01 1.99368641e-01
-8.79985750e-01 4.62350190e-01 7.10244596e-01 -2.50080507e-02
-2.50274301e-01 -2.10778341e-01 -6.99663341e-01 1.34892687e-01
4.08778042e-01 -4.22364950e-01 3.15195322e-01 -9.26423192e-01
-1.11328363e+00 7.49031663e-01 -2.08003163e-01 -2.79086381e-01
3.23258698e-01 5.68288565e-01 -4.17609066e-01 2.98375964e-01
3.00246980e-02 2.41609856e-01 6.57570720e-01 -1.49225867e+00
-4.93248969e-01 -2.49829322e-01 -3.15711915e-01 1.92241713e-01
-4.01816934e-01 -8.50951001e-02 -8.13556194e-01 -2.68216342e-01
7.92742610e-01 -8.75795722e-01 -5.69764316e-01 -5.16160786e-01
-5.29799581e-01 -4.16303724e-01 9.83733654e-01 -4.17193510e-02
1.54838121e+00 -2.32639909e+00 7.88080469e-02 1.16965449e+00
4.98883903e-01 1.11555129e-01 6.30348772e-02 6.28246665e-01
-5.22222817e-02 3.77848983e-01 -3.17393541e-01 -9.19229314e-02
-2.24846125e-01 -8.14223066e-02 4.02191848e-01 7.89238632e-01
-3.79243821e-01 2.53939062e-01 -1.02119505e+00 -9.57314074e-01
4.95577365e-01 2.13186264e-01 -5.44705868e-01 -1.16876014e-01
4.41313498e-02 5.25363237e-02 -5.25569797e-01 2.84124821e-01
9.05288398e-01 -2.85866588e-01 4.45662469e-01 -2.35802755e-01
-2.50599504e-01 -3.11223239e-01 -1.94475424e+00 1.21383333e+00
1.40745118e-01 1.52632162e-01 2.73878604e-01 -1.37226009e+00
1.24753785e+00 6.48049489e-02 1.05375290e+00 -2.17196062e-01
2.00814575e-01 2.54308820e-01 1.92018718e-01 -2.31357381e-01
3.26674461e-01 1.24120615e-01 1.33640379e-01 4.75268632e-01
-2.98626840e-01 -2.43020847e-01 6.38405204e-01 5.30911684e-01
1.03096104e+00 -4.36488360e-01 1.25426918e-01 -4.88134265e-01
8.47379148e-01 1.47431031e-01 6.61932707e-01 5.02731800e-01
-6.40700087e-02 5.33752441e-01 3.69872570e-01 -2.13445604e-01
-8.83966565e-01 -1.04558086e+00 -2.55413703e-03 7.11588502e-01
5.26704371e-01 -7.35007405e-01 -1.10237253e+00 -5.65207720e-01
-1.97695624e-02 2.15434402e-01 -4.44320083e-01 -3.20439115e-02
-4.30129677e-01 -8.83386016e-01 1.19070206e-02 -1.89247757e-01
4.48076040e-01 -9.56917346e-01 -2.41738304e-01 1.96670473e-01
-3.19066606e-02 -6.48778081e-01 -5.55739522e-01 -4.44810688e-02
-8.90316963e-01 -1.57079959e+00 -5.11394083e-01 -1.01073635e+00
1.23302019e+00 6.51524782e-01 8.99243176e-01 4.28591251e-01
-1.41481951e-01 2.90433824e-01 -6.92528725e-01 4.72030453e-02
-2.04717234e-01 3.26464593e-01 3.10911145e-02 1.84184760e-01
4.38528091e-01 -7.49011934e-01 -6.68587863e-01 6.24673128e-01
-6.45246387e-01 -4.82207537e-01 3.68407339e-01 2.90974200e-01
8.78284693e-01 7.01484144e-01 3.92308503e-01 -8.30970287e-01
8.99983346e-01 -5.07951438e-01 -6.64204538e-01 1.12940967e-01
-5.84218800e-01 -1.28340989e-01 8.65440309e-01 -3.30277026e-01
-5.85874856e-01 3.51768225e-01 2.37133950e-01 -4.14188147e-01
-1.21423498e-01 5.33974588e-01 -2.23607391e-01 -1.79206207e-01
6.58407450e-01 1.19285963e-01 6.90625608e-02 -2.49425083e-01
5.15952408e-01 4.65043277e-01 1.76803440e-01 -1.86063707e-01
9.91533041e-01 6.22753799e-01 1.13119408e-01 -9.13433433e-01
-9.57744345e-02 -7.94938028e-01 -8.32176268e-01 -4.10027683e-01
8.33661318e-01 -3.65361214e-01 -9.13923740e-01 2.60246754e-01
-7.00381279e-01 -3.50026526e-02 -1.80359781e-01 5.53901315e-01
-2.54333317e-01 8.52588296e-01 -3.51835728e-01 -8.91089082e-01
-8.12823623e-02 -8.69611859e-01 3.35967034e-01 2.75038928e-01
-1.42151147e-01 -9.76144791e-01 2.93321013e-01 -2.28514984e-01
1.24744475e-01 2.32933164e-01 8.48915100e-01 -5.56739032e-01
-3.76046360e-01 -8.71152282e-02 -1.88154057e-01 -4.55970056e-02
4.43153381e-01 5.05889297e-01 -2.27796465e-01 -4.88608897e-01
-1.80799186e-01 3.38932008e-01 7.22903550e-01 6.73146486e-01
1.19464707e+00 -1.41853541e-01 -8.10894489e-01 6.47211790e-01
1.52800560e+00 5.05335510e-01 5.47964215e-01 6.89717159e-02
7.71477878e-01 5.23390651e-01 3.08402926e-01 2.04685554e-01
2.61302948e-01 3.59593898e-01 4.41172838e-01 -2.88558066e-01
1.51201040e-01 3.47130671e-02 -4.42398451e-02 1.09845030e+00
-2.79703438e-01 -2.70503908e-01 -1.10143411e+00 6.84883058e-01
-1.96519530e+00 -1.16025972e+00 -7.19964802e-01 2.37575555e+00
4.27223742e-01 2.02870846e-01 4.11601901e-01 4.57806081e-01
1.16623235e+00 2.04133280e-02 -3.05992424e-01 -1.63995385e-01
-4.36775908e-02 1.24253318e-01 6.15402579e-01 5.42726398e-01
-9.84673858e-01 9.98941422e-01 6.45652723e+00 8.24628949e-01
-7.25704372e-01 -1.84307963e-01 4.88803744e-01 1.36105731e-01
-3.34583998e-01 1.24015883e-01 -4.62995648e-01 5.80052078e-01
6.32031679e-01 -2.38412440e-01 3.59522164e-01 8.74193847e-01
4.77818102e-01 -4.43888247e-01 -4.31843817e-01 1.02361178e+00
-2.09898070e-01 -1.00083542e+00 -1.43572584e-01 1.59365207e-01
7.35581517e-01 -1.70950428e-01 -2.93348283e-01 -1.38666138e-01
7.27677464e-01 -9.00478125e-01 1.34762987e-01 4.93166417e-01
4.04831469e-01 -1.17950344e+00 5.52137792e-01 2.69278049e-01
-1.79469705e+00 1.77208558e-01 -5.29262125e-01 2.10532293e-01
1.60209462e-01 1.07066178e+00 -8.28930140e-01 6.88166320e-01
7.92487442e-01 5.27011275e-01 -4.52263176e-01 1.27168167e+00
2.25152627e-01 4.77261394e-01 -6.45593643e-01 -2.24200130e-01
3.12397182e-01 -9.75550950e-01 3.98048252e-01 1.19136846e+00
5.76420963e-01 1.63417369e-01 4.79449630e-01 7.29391098e-01
2.26587385e-01 7.12960958e-01 -7.16295600e-01 2.31202081e-01
8.62632096e-01 1.40461862e+00 -1.49504328e+00 -3.47097367e-01
-3.81268412e-01 6.55698717e-01 2.37447158e-01 1.25449792e-01
-8.06616902e-01 -6.56275213e-01 3.67196262e-01 6.15124702e-01
2.47149631e-01 -4.47075188e-01 5.09177782e-02 -6.92561150e-01
-2.70973384e-01 -5.31375945e-01 6.55499101e-01 -3.90276462e-01
-1.32522655e+00 5.99099517e-01 -2.48227660e-02 -1.16532207e+00
3.37028839e-02 -9.01038274e-02 -7.02447295e-01 6.97894394e-01
-8.03675950e-01 -5.75226188e-01 -5.19492507e-01 1.09369361e+00
-4.19425219e-02 5.21616563e-02 5.06921768e-01 3.42621982e-01
-4.35570359e-01 2.00010166e-01 2.05570728e-01 2.12392151e-01
5.41458964e-01 -1.31342864e+00 -1.53160896e-02 8.35712671e-01
7.45233521e-02 7.14582443e-01 5.91648638e-01 -9.71846998e-01
-1.05242765e+00 -9.42969322e-01 7.05392063e-01 1.29019335e-01
5.65872252e-01 -2.79294670e-01 -1.05998313e+00 2.92304724e-01
1.87964350e-01 -2.33010873e-02 5.54858446e-01 1.29777998e-01
1.98305119e-03 -2.20048547e-01 -1.24129844e+00 5.43852568e-01
1.16230381e+00 -6.24022819e-02 -2.37629578e-01 2.67020732e-01
2.68860340e-01 1.79310471e-01 -8.56848836e-01 2.63098925e-01
6.55543283e-02 -1.27816164e+00 9.03787434e-01 -3.13861929e-02
-3.85213792e-01 -8.34463716e-01 2.32573658e-01 -1.55883634e+00
-6.12728775e-01 -6.11854196e-01 5.38031161e-01 1.26678479e+00
3.29226136e-01 -7.10512221e-01 9.35608685e-01 4.90735658e-02
1.05658203e-01 -3.69179904e-01 -7.31472671e-01 -8.22788179e-01
-2.38207415e-01 -1.61134899e-02 6.34968579e-01 1.31477165e+00
3.26160103e-01 3.01777810e-01 4.63217087e-02 1.12447456e-01
8.85623336e-01 9.29495543e-02 7.94406712e-01 -1.65467668e+00
1.61046356e-01 -7.09437430e-01 -4.72358406e-01 -5.24251401e-01
-5.60604669e-02 -8.97566259e-01 -3.80177088e-02 -1.67561924e+00
1.94698334e-01 -1.00942397e+00 -9.82778743e-02 1.39039993e-01
-1.98731482e-01 2.57186800e-01 -1.69605371e-02 3.80849630e-01
-7.08717287e-01 1.76759854e-01 9.61095989e-01 3.90664339e-02
-7.86157489e-01 -1.28153250e-01 -2.90710151e-01 9.05142963e-01
9.82293010e-01 -4.32762384e-01 -5.84100723e-01 8.01046267e-02
1.21948712e-01 -2.33282804e-01 -2.31330752e-01 -1.10549963e+00
4.71949995e-01 -2.61347204e-01 7.61943310e-02 -7.76441395e-01
-1.49414405e-01 -1.15521514e+00 6.97589397e-01 4.87945259e-01
1.63341150e-01 1.71427175e-01 -1.95000485e-01 5.59166312e-01
-2.09074020e-01 -2.30527818e-01 1.05575371e+00 -3.30789059e-01
-5.17785788e-01 3.47931623e-01 -4.99436468e-01 -9.83855277e-02
1.53065503e+00 -6.42149091e-01 1.51449203e-01 -4.31120038e-01
-1.05106568e+00 3.62605095e-01 6.51045680e-01 -7.08103403e-02
4.86812592e-01 -1.25790262e+00 -4.60079223e-01 -3.15384977e-02
-7.26892799e-02 1.62036657e-01 5.36047444e-02 8.44177544e-01
-6.53463066e-01 1.27128690e-01 -2.80624330e-02 -7.63109326e-01
-1.52089190e+00 9.15172160e-01 4.18397993e-01 -1.69947445e-02
-5.71551025e-01 4.93499428e-01 -2.45418344e-02 -3.95655215e-01
1.61414266e-01 -2.03409493e-01 -4.17300075e-01 4.08058912e-01
1.48909822e-01 4.80054349e-01 -5.56890368e-02 -7.87989080e-01
-5.59155703e-01 9.82780695e-01 2.69036025e-01 4.04675379e-02
1.28584623e+00 -3.60147983e-01 -4.08393115e-01 2.55622864e-01
1.06787217e+00 3.98519218e-01 -6.68449461e-01 1.22630158e-02
1.79014266e-01 -6.15813851e-01 -1.96855873e-01 3.98356318e-02
-1.49846506e+00 3.11382204e-01 4.79567647e-01 9.70801651e-01
1.23258471e+00 2.10916817e-01 4.29388136e-01 1.02337249e-01
4.14440811e-01 -1.27396083e+00 -4.13956307e-02 2.76376098e-01
2.82854617e-01 -8.40470314e-01 7.96361491e-02 -9.29466605e-01
-3.03084433e-01 9.63149250e-01 4.77392286e-01 -3.93412292e-01
9.35051501e-01 1.67745456e-01 -1.02520801e-01 -5.04061222e-01
-9.52423960e-02 -5.67365348e-01 1.33409664e-01 6.57347441e-01
3.16480190e-01 4.07890230e-02 -8.24360788e-01 4.94985422e-03
-2.97047943e-01 -3.80074084e-01 3.44814897e-01 5.05352199e-01
-8.17228019e-01 -1.08886898e+00 -6.45967245e-01 4.33245003e-01
-8.81348625e-02 1.17073618e-01 -6.62050664e-01 1.01141536e+00
3.60983223e-01 1.13404977e+00 2.85681695e-01 -5.12015224e-01
4.27463710e-01 -8.58397558e-02 1.47925809e-01 -5.19093812e-01
-6.37180090e-01 2.96221077e-01 -2.52659559e-01 -5.92711568e-01
-6.55946612e-01 -6.57052517e-01 -1.56061280e+00 -5.72319448e-01
-4.98328269e-01 7.53319621e-01 3.77293617e-01 3.26037556e-01
2.31101766e-01 1.48888826e-01 9.43657458e-01 -4.71128076e-01
1.98251635e-01 -6.21014535e-01 -9.63459492e-01 6.01083755e-01
-2.58200288e-01 -5.58021307e-01 -7.42218196e-01 6.59245104e-02] | [7.562347412109375, 4.681687831878662] |
ca3d039e-c41f-4a38-9c79-ec294fd28e26 | bi-linear-value-networks-for-multi-goal | null | null | https://openreview.net/forum?id=LedObtLmCjS | https://openreview.net/pdf?id=LedObtLmCjS | Bi-linear Value Networks for Multi-goal Reinforcement Learning | Universal value functions are used to score the long-term utility of actions to achieve a goal from the current state. In contrast to prior methods that learn a monolithic function to approximate the value, we propose a bi-linear decomposition of the value function. The first component, akin to a global plan models how the state should be changed to reach the goal. The second component, akin to a local controller selects the optimal action to actualize the desired change in state. We simultaneously learn both components. Such decomposition enables both the global and local components to make efficient use of interaction data and independently generalize. The consequence is superior overall generalization and performance of our system on a wide range of challenging goal-conditioned tasks in comparison to the current state-of-the-art. | ['Pulkit Agrawal', 'Zhang-Wei Hong', 'Ge Yang'] | 2021-09-29 | null | null | null | iclr-2022-4 | ['multi-goal-reinforcement-learning'] | ['methodology'] | [ 9.81719717e-02 3.01110208e-01 -4.91732270e-01 -3.27974677e-01
-8.88045192e-01 -6.44537985e-01 7.39344299e-01 9.75660309e-02
-5.20129204e-01 1.06265020e+00 3.95436704e-01 -2.46294867e-02
-3.38810116e-01 -7.14170992e-01 -7.05320776e-01 -7.62413979e-01
-2.53591806e-01 5.90130806e-01 2.54805416e-01 -2.15565562e-01
2.42068663e-01 2.80126363e-01 -1.29358315e+00 1.21109352e-01
6.82020843e-01 1.02148128e+00 1.68707475e-01 5.57129979e-01
2.86720127e-01 9.60382342e-01 -3.09366286e-01 1.45454526e-01
2.78066427e-01 -7.06797481e-01 -8.88607323e-01 1.20416030e-01
1.51455671e-01 -6.13039851e-01 -3.86071831e-01 9.88989234e-01
3.22995931e-01 5.48372626e-01 5.06728649e-01 -1.17788267e+00
-3.30432057e-01 7.05240965e-01 1.62447616e-01 -1.29767001e-01
4.02730316e-01 8.26201200e-01 1.22499382e+00 -1.25316775e-03
7.04663873e-01 1.31222939e+00 2.90281087e-01 6.33363485e-01
-1.71311951e+00 -1.80891350e-01 6.50454998e-01 -2.99971476e-02
-9.81387973e-01 -4.21318948e-01 4.52037752e-01 -3.56762528e-01
1.17765296e+00 -1.45410120e-01 8.53924096e-01 1.13584435e+00
4.80435699e-01 6.78628683e-01 1.18674922e+00 -6.89570904e-02
8.00708890e-01 -3.31166625e-01 -1.45763636e-01 7.42373347e-01
-6.59812167e-02 6.14419997e-01 -4.50403959e-01 -1.68831483e-01
8.02959561e-01 1.01614362e-02 -1.59260005e-01 -7.81948984e-01
-1.02895248e+00 4.85411435e-01 5.20224333e-01 6.37233406e-02
-7.41863906e-01 8.70383739e-01 2.19112635e-01 5.35705805e-01
-3.60011756e-02 7.73460686e-01 -5.43966413e-01 -5.69629669e-01
-6.35784805e-01 7.43670642e-01 7.66698182e-01 5.78189731e-01
7.54325032e-01 5.74742481e-02 -6.77753150e-01 2.81149656e-01
1.99652284e-01 3.92603755e-01 2.89240509e-01 -1.47396445e+00
4.35967952e-01 7.12894976e-01 5.94323099e-01 -1.56080619e-01
-3.24409723e-01 -5.86009204e-01 -1.22084863e-01 8.49342346e-01
5.39824188e-01 -3.31259489e-01 -1.01400542e+00 2.23684740e+00
1.83680311e-01 2.05689386e-01 1.81783307e-02 7.21321106e-01
-1.31026968e-01 4.04226989e-01 7.23404363e-02 -1.26621619e-01
8.16324592e-01 -8.13174367e-01 -5.18463910e-01 -6.80920601e-01
5.59259772e-01 -1.30260782e-02 1.07926548e+00 3.18676472e-01
-1.24152982e+00 -3.27867419e-01 -9.72901583e-01 2.71957338e-01
-7.81343058e-02 -2.26737291e-01 6.63372457e-01 1.24968894e-01
-1.14459348e+00 1.13576603e+00 -1.28927767e+00 -1.56914517e-01
3.82773817e-01 4.51075047e-01 -2.56379843e-01 5.23957908e-02
-1.01570714e+00 1.25206387e+00 6.40350997e-01 -2.77068555e-01
-1.70117342e+00 -6.26747251e-01 -6.90626383e-01 3.92603099e-01
7.78290451e-01 -8.49643588e-01 1.68052733e+00 -6.06737554e-01
-2.02692199e+00 4.48514998e-01 7.89391175e-02 -6.65773809e-01
4.07088041e-01 -2.16375798e-01 3.24235290e-01 -2.01369122e-01
2.80271377e-02 5.55576444e-01 7.25856483e-01 -9.64533269e-01
-6.99349046e-01 -3.96835655e-01 5.05654752e-01 4.71527398e-01
-1.88174084e-01 -5.08684516e-01 -3.81517828e-01 -8.77579302e-02
-1.95115641e-01 -1.13813138e+00 -6.03498340e-01 -3.89286503e-02
-1.56860441e-01 -2.69822299e-01 4.15488303e-01 -2.58447587e-01
1.31725502e+00 -1.73095679e+00 8.89045835e-01 -4.34105508e-02
-4.91461670e-03 -6.25248104e-02 -1.97425947e-01 7.63743937e-01
1.86292142e-01 -1.72354877e-01 -1.77065805e-01 -3.73998642e-01
2.57776827e-01 2.21334338e-01 -2.72110075e-01 3.57241124e-01
-6.69790059e-03 1.03996360e+00 -1.04911888e+00 1.42010421e-01
2.43749663e-01 9.00303125e-02 -8.40999901e-01 4.88355964e-01
-7.46477425e-01 6.99646473e-01 -8.31627548e-01 1.92524478e-01
1.45494360e-02 -7.95123577e-02 5.27453721e-01 3.55550706e-01
-9.93840918e-02 6.79318905e-01 -1.09303081e+00 1.86268151e+00
-4.72305387e-01 -4.93207388e-02 2.89554000e-01 -9.75995362e-01
6.18491590e-01 3.56455803e-01 5.06686211e-01 -5.95035672e-01
3.41675252e-01 -1.12557635e-01 6.24338770e-03 -2.44813919e-01
2.20294163e-01 -2.71247983e-01 -4.16716635e-01 4.51531738e-01
2.97081828e-01 -4.24226135e-01 2.67187119e-01 6.36208653e-02
1.34131324e+00 7.81055331e-01 6.13304257e-01 -2.89216757e-01
3.54620039e-01 5.31092547e-02 5.97190678e-01 8.66505325e-01
-3.21015686e-01 4.65726480e-02 9.40459132e-01 -3.95390570e-01
-6.30239904e-01 -1.21050942e+00 4.86742795e-01 1.28275275e+00
1.16332091e-01 -4.18285280e-01 -7.06986666e-01 -7.33505249e-01
1.65949240e-01 1.06328893e+00 -6.41521335e-01 -5.73776960e-01
-5.92010379e-01 -4.95723635e-02 7.99091682e-02 6.78852081e-01
4.17514712e-01 -1.18101096e+00 -9.99404728e-01 2.48654976e-01
1.39411502e-02 -7.03044236e-01 -7.23221183e-01 4.07634109e-01
-1.04606450e+00 -9.65040147e-01 -8.29227418e-02 -2.21961826e-01
3.68961155e-01 -4.08528000e-01 1.18319798e+00 -2.26824597e-01
1.94521323e-01 5.18949389e-01 4.66939509e-02 -3.81096341e-02
-3.36744279e-01 -9.31418389e-02 1.14648700e-01 -9.83844176e-02
-5.42315096e-02 -7.63337553e-01 -6.37893558e-01 1.50722172e-02
-5.66543341e-01 -8.31744596e-02 5.86846232e-01 8.32129061e-01
4.94100630e-01 3.91245447e-02 3.86243224e-01 -6.04969621e-01
9.74487901e-01 -2.54867971e-01 -8.06307793e-01 1.11933626e-01
-8.91940176e-01 7.24258661e-01 7.35145271e-01 -4.74272937e-01
-1.09516871e+00 2.48851314e-01 1.34452969e-01 -3.07686269e-01
-1.34304047e-01 6.40259087e-01 -1.27233326e-01 4.10996079e-01
5.91795325e-01 3.36808532e-01 1.10681787e-01 -4.49412405e-01
5.70771933e-01 7.51363933e-02 6.51413083e-01 -1.03137767e+00
4.45036560e-01 1.97225466e-01 1.15977250e-01 -1.23966366e-01
-7.24269032e-01 -1.31166041e-01 -3.89472634e-01 -2.44977444e-01
8.71294677e-01 -7.78768063e-01 -1.10593390e+00 3.60778719e-01
-7.36138761e-01 -1.05547118e+00 -6.88755989e-01 3.29528719e-01
-1.17963505e+00 -2.14304075e-01 -3.69515002e-01 -8.94778728e-01
3.14501822e-02 -1.28994870e+00 1.08037698e+00 2.49094173e-01
-2.20845252e-01 -9.41169739e-01 3.42239827e-01 -6.31003529e-02
5.00422001e-01 4.09915924e-01 9.12225366e-01 -3.01055580e-01
-7.65055120e-01 -1.97240114e-01 4.09383148e-01 2.97004402e-01
5.58985807e-02 -6.13231778e-01 -4.59631592e-01 -6.49614632e-01
6.82386244e-03 -6.72519565e-01 8.36302757e-01 4.57003653e-01
7.49519289e-01 -3.62370282e-01 -3.46416473e-01 3.50842535e-01
1.35825360e+00 4.00498420e-01 5.91911197e-01 2.14936182e-01
2.49448210e-01 1.23362087e-01 6.69116497e-01 6.14024460e-01
3.64564955e-01 9.58216369e-01 6.82481349e-01 4.86347884e-01
-1.35631531e-01 -6.78837419e-01 8.34332466e-01 -5.85203245e-02
-1.56917676e-01 -3.14386636e-02 -6.96171701e-01 4.89614815e-01
-2.27085280e+00 -1.07244503e+00 7.07008541e-01 2.55449843e+00
9.43161428e-01 5.52640021e-01 3.28552663e-01 -4.11855221e-01
1.23826690e-01 3.31029296e-01 -1.13890433e+00 -2.01408580e-01
6.90361857e-01 1.15970559e-01 4.50183481e-01 7.91705012e-01
-9.32588995e-01 1.15554476e+00 7.29965973e+00 6.33243799e-01
-1.05598378e+00 5.38756587e-02 3.64069700e-01 -4.86629635e-01
-1.58820465e-01 3.09780866e-01 -7.32591510e-01 3.04634929e-01
1.09531379e+00 -2.23595351e-01 1.12023675e+00 9.15056467e-01
3.53935301e-01 -1.11258425e-01 -1.49597967e+00 3.47262591e-01
-4.95595545e-01 -1.11272550e+00 -3.72963786e-01 8.89412761e-02
6.61026895e-01 8.07495117e-02 1.55644277e-02 7.57585287e-01
1.11539984e+00 -9.24635351e-01 1.02105343e+00 7.65823066e-01
6.97786987e-01 -6.35839999e-01 2.16699451e-01 7.67400384e-01
-1.13182104e+00 -5.38795769e-01 4.54011038e-02 -5.60109258e-01
1.61409616e-01 8.01335741e-03 -6.41476333e-01 1.30076379e-01
2.89372444e-01 5.01570702e-01 -2.48921424e-01 5.99259317e-01
-6.95881724e-01 4.87757653e-01 -1.98817879e-01 -8.49566236e-02
5.37109256e-01 -4.29000854e-01 6.42733276e-01 5.69947243e-01
1.73196122e-02 -7.63336793e-02 7.93290973e-01 1.03462136e+00
1.25681803e-01 -4.36489165e-01 -6.79070115e-01 -3.44426602e-01
3.58984321e-01 9.08948481e-01 -3.70946109e-01 -2.81461835e-01
-3.31617109e-02 8.76384437e-01 9.47998345e-01 4.70303357e-01
-7.79232323e-01 -1.06407359e-01 1.05457008e+00 -7.31805712e-02
4.32725817e-01 -4.64507461e-01 -2.83847213e-01 -1.05403256e+00
-6.63121268e-02 -9.10636485e-01 4.14198905e-01 -5.05733788e-01
-7.51959264e-01 3.36390018e-01 2.61183560e-01 -1.16269350e+00
-7.71932483e-01 -4.09804136e-01 -7.32201934e-01 9.23774540e-01
-1.15777659e+00 -8.83260846e-01 8.02073479e-02 3.93350661e-01
3.90171647e-01 -1.36088476e-01 9.56628442e-01 -4.50538933e-01
-4.83352453e-01 2.95611560e-01 3.92999351e-02 -3.76775086e-01
4.86079186e-01 -1.42788517e+00 2.32191503e-01 8.61234546e-01
-2.90272534e-01 7.71310151e-01 8.43458116e-01 -6.68948650e-01
-1.54792476e+00 -7.37310529e-01 3.92020881e-01 -6.17193043e-01
6.66444063e-01 -2.25508377e-01 -4.32790697e-01 9.82853532e-01
1.44544870e-01 -8.48917738e-02 7.44298473e-02 3.05670381e-01
-4.33184415e-01 -2.49671251e-01 -1.13376367e+00 8.70724678e-01
1.14484799e+00 -3.80050182e-01 -6.76264763e-01 8.37483555e-02
6.22448742e-01 -4.17930514e-01 -8.34788024e-01 9.41402689e-02
6.29247308e-01 -1.04096222e+00 9.55286384e-01 -1.13978958e+00
4.16333616e-01 -2.28472412e-01 -2.07161248e-01 -1.76209390e+00
-5.97883165e-01 -9.76050913e-01 -5.17157316e-01 5.25237083e-01
3.32128793e-01 -7.23941267e-01 8.22281837e-01 1.02630746e+00
-3.45268548e-01 -1.01452541e+00 -1.06208134e+00 -9.24311101e-01
1.93281591e-01 -1.37407497e-01 4.41550314e-01 3.28369617e-01
4.17509526e-01 6.19102597e-01 -5.69482148e-01 -1.71967596e-02
4.58244532e-01 3.89461637e-01 6.81598365e-01 -9.18883622e-01
-7.96335399e-01 -7.09134042e-01 -2.20114931e-01 -1.29848194e+00
2.69146711e-01 -1.02122164e+00 4.09371048e-01 -1.67054391e+00
2.02721104e-01 -1.28772289e-01 -5.44946432e-01 8.91284287e-01
-2.16277391e-01 -4.55185533e-01 4.42378223e-01 -5.74037284e-02
-8.22898865e-01 8.20226014e-01 1.42993045e+00 -4.28040251e-02
-7.01787770e-01 3.25517841e-02 -9.43671763e-01 5.63473701e-01
8.29932749e-01 -5.66736341e-01 -6.97298229e-01 -3.06377202e-01
-6.80563971e-02 7.04512060e-01 1.27080187e-01 -1.03796232e+00
2.37180635e-01 -9.37867761e-01 1.92973372e-02 2.00744476e-02
5.66077888e-01 -7.64021873e-01 5.85224554e-02 8.68529797e-01
-8.18251014e-01 -1.57784566e-01 1.09913118e-01 9.21864510e-01
4.03826907e-02 1.30447827e-03 8.13409805e-01 -2.68087119e-01
-6.86558723e-01 3.61880124e-01 -3.61826420e-01 1.04435282e-02
1.11114681e+00 1.72942206e-01 -2.18367726e-01 -6.86524034e-01
-9.14124727e-01 5.90357363e-01 4.54412907e-01 4.43500221e-01
3.69712681e-01 -1.39379382e+00 -5.76951563e-01 1.09203961e-02
-1.02320142e-01 -3.21604937e-01 6.51477054e-02 7.22597420e-01
7.90416747e-02 4.08051282e-01 -4.36342180e-01 -1.64997205e-01
-7.49047518e-01 5.29691339e-01 4.88515407e-01 -9.28327978e-01
-5.57964563e-01 5.69441974e-01 4.35142107e-02 -4.44592059e-01
2.88227379e-01 -3.02995294e-01 6.25769570e-02 -3.22160423e-01
4.47359174e-01 3.09456795e-01 -3.18407744e-01 -1.16862565e-01
-2.78334111e-01 1.53962627e-01 1.36037752e-01 -5.02883792e-01
1.36490119e+00 4.84620407e-02 -1.82819217e-02 4.75769937e-01
8.97532344e-01 -5.62218964e-01 -2.05280566e+00 -1.93290606e-01
6.45710975e-02 -4.13116157e-01 2.98016697e-01 -1.25254691e+00
-6.68610573e-01 4.69817221e-01 5.02026260e-01 6.18107617e-02
9.79768991e-01 -1.30643740e-01 4.99897212e-01 7.82981277e-01
7.77291298e-01 -1.17043948e+00 2.17062369e-01 7.47499585e-01
9.88119185e-01 -9.88252580e-01 -9.59425867e-02 1.92433596e-01
-9.65535879e-01 7.95435786e-01 6.43834531e-01 -3.61895025e-01
2.91594177e-01 3.63301158e-01 -2.75117636e-01 -9.76506621e-02
-1.24672961e+00 -4.51729894e-01 1.72242284e-01 6.09854102e-01
2.11711437e-01 1.82256132e-01 -1.58319160e-01 4.50288713e-01
1.15394816e-01 2.52675951e-01 3.64463091e-01 1.13067055e+00
-6.51612282e-01 -1.32962990e+00 -6.26407266e-02 4.40582126e-01
-8.88947025e-02 2.83730626e-01 -3.48077744e-01 5.25918126e-01
-2.57561088e-01 8.82880270e-01 -1.46986455e-01 -4.16288435e-01
4.36086208e-01 4.82252419e-01 6.82551086e-01 -8.24316561e-01
-6.00006521e-01 1.00482255e-02 1.76030308e-01 -1.38622367e+00
5.10683469e-02 -6.80791080e-01 -1.36779654e+00 -3.85190733e-02
1.28196970e-01 -9.26414803e-02 3.70016873e-01 8.81323397e-01
5.03467083e-01 7.34794497e-01 3.40635061e-01 -1.06986153e+00
-1.37092328e+00 -9.10645008e-01 -4.22744840e-01 5.71992934e-01
3.84672821e-01 -9.48198259e-01 -2.00216740e-01 -1.90765455e-01] | [4.148192405700684, 1.6997172832489014] |
aff3dd71-0a10-40f1-91ad-99728da7c673 | malicious-or-benign-towards-effective-content | 2305.15551 | null | https://arxiv.org/abs/2305.15551v1 | https://arxiv.org/pdf/2305.15551v1.pdf | Malicious or Benign? Towards Effective Content Moderation for Children's Videos | Online video platforms receive hundreds of hours of uploads every minute, making manual content moderation impossible. Unfortunately, the most vulnerable consumers of malicious video content are children from ages 1-5 whose attention is easily captured by bursts of color and sound. Scammers attempting to monetize their content may craft malicious children's videos that are superficially similar to educational videos, but include scary and disgusting characters, violent motions, loud music, and disturbing noises. Prominent video hosting platforms like YouTube have taken measures to mitigate malicious content on their platform, but these videos often go undetected by current content moderation tools that are focused on removing pornographic or copyrighted content. This paper introduces our toolkit Malicious or Benign for promoting research on automated content moderation of children's videos. We present 1) a customizable annotation tool for videos, 2) a new dataset with difficult to detect test cases of malicious content and 3) a benchmark suite of state-of-the-art video classification models. | ['Gita Sukthankar', 'H. M. Umer Qaisar', 'Muhammad Junaid Khan', 'Syed Hammad Ahmed'] | 2023-05-24 | null | null | null | null | ['video-classification'] | ['computer-vision'] | [ 5.93399145e-02 2.10157305e-01 -2.48098105e-01 7.34668821e-02
-3.99814069e-01 -1.02712393e+00 5.11002183e-01 2.87036270e-01
-2.27563307e-02 2.72918284e-01 4.21085745e-01 -7.08349943e-02
4.87450093e-01 -2.94340044e-01 -7.48721600e-01 -4.90008116e-01
-2.34585360e-01 -3.57208073e-01 7.12827623e-01 2.69207414e-02
3.97029430e-01 5.93844578e-02 -2.02690244e+00 6.63164437e-01
5.65215111e-01 4.69969600e-01 -5.96628010e-01 1.28280580e+00
2.75098205e-01 1.80854928e+00 -1.04184198e+00 -9.70397830e-01
-4.96889390e-02 -3.32161814e-01 -4.60712314e-01 -3.94866988e-02
1.31725836e+00 -1.20044255e+00 -7.32520282e-01 1.41825783e+00
4.10845816e-01 -2.51788914e-01 2.07216173e-01 -1.74715042e+00
-7.86770165e-01 1.06366920e+00 -5.09395421e-01 6.80392623e-01
9.23471868e-01 2.70405978e-01 4.62619901e-01 -5.42097129e-02
7.30770886e-01 1.14088523e+00 6.76634192e-01 9.12269652e-01
-1.10592830e+00 -1.09401059e+00 -9.50733796e-02 1.63849711e-01
-1.15638685e+00 -6.47355020e-01 7.35380054e-01 -8.50253284e-01
7.80244112e-01 6.75853491e-01 1.10846555e+00 1.88317060e+00
-1.23141304e-01 8.94605756e-01 5.59989393e-01 8.45993683e-02
1.63145393e-01 2.18111277e-01 -2.80406862e-01 5.87853849e-01
3.27105045e-01 -5.81784904e-01 -6.20052397e-01 -5.39553046e-01
-3.52985295e-03 -1.73314810e-01 -2.49066055e-01 1.80605173e-01
-5.37434876e-01 9.90575910e-01 -2.91126043e-01 2.76235104e-01
1.54345825e-01 4.55021024e-01 8.22204888e-01 3.77621025e-01
8.35955679e-01 4.93652403e-01 -9.55270827e-02 -1.06133366e+00
-9.72297728e-01 3.85419637e-01 8.26538980e-01 8.37768972e-01
2.49700233e-01 2.74925917e-01 3.03268284e-01 7.04435349e-01
2.14365989e-01 3.81007016e-01 4.14097697e-01 -1.33295548e+00
3.37165207e-01 1.45310014e-01 -2.00552389e-01 -1.47421443e+00
1.52825534e-01 2.53157675e-01 2.54946724e-02 -4.47828285e-02
3.18374187e-01 -2.52152592e-01 -4.41382200e-01 1.38646448e+00
5.62287748e-01 7.99075842e-01 -6.40345693e-01 5.00758052e-01
1.35437739e+00 3.75065058e-01 1.07823133e-01 -1.51100755e-01
9.87978101e-01 -8.17130685e-01 -7.04747498e-01 2.68569916e-01
6.17510200e-01 -1.01802945e+00 7.46820509e-01 8.47070575e-01
-1.47920144e+00 2.23625034e-01 -8.94987583e-01 -5.03682010e-02
-4.35157716e-01 -8.55550289e-01 4.72654283e-01 1.56461632e+00
-1.13537788e+00 8.53928506e-01 -6.01720929e-01 -9.51916128e-02
7.27298915e-01 7.24943429e-02 -5.08889139e-01 4.06563729e-02
-7.77763486e-01 5.32433033e-01 -3.74493361e-01 -6.96658492e-01
-1.29123855e+00 -1.16077781e+00 -8.90689552e-01 -2.03875154e-01
1.87923878e-01 4.95261759e-01 1.48127949e+00 -1.59241712e+00
-1.32009947e+00 1.15609097e+00 6.04172409e-01 -6.66135363e-03
5.64731121e-01 -5.68266630e-01 -5.97508729e-01 9.47012186e-01
3.61727059e-01 4.13909942e-01 1.76582122e+00 -6.79397881e-01
-4.59594518e-01 1.52987242e-01 2.85076439e-01 -3.87153894e-01
-1.01618052e+00 9.24348712e-01 -1.84093833e-01 -1.02251995e+00
-4.53933001e-01 -6.54657125e-01 4.36397731e-01 -1.34543851e-01
-5.38642049e-01 2.07451522e-01 1.35494721e+00 -9.53389585e-01
1.51883638e+00 -2.20334220e+00 -2.94061035e-01 -1.44033000e-01
6.32058144e-01 5.46238363e-01 -8.94880965e-02 4.32121813e-01
-2.45023668e-01 7.24596143e-01 8.28088820e-01 -2.17134636e-02
2.85843504e-03 -3.11286718e-01 -1.93705991e-01 8.86838377e-01
-2.01013982e-01 2.39641324e-01 -1.20789528e+00 -3.39191616e-01
-3.68935466e-02 5.67042351e-01 -1.06752717e+00 2.90676951e-01
-1.64088294e-01 1.64318070e-01 -2.67555356e-01 1.01691949e+00
7.09468126e-01 1.83406532e-01 -2.09575340e-01 4.97472137e-01
-2.09258810e-01 5.83599687e-01 -6.06436074e-01 8.39860976e-01
2.83175200e-01 1.34561861e+00 5.42935848e-01 -4.40885574e-01
3.40041742e-02 5.56303263e-01 6.90154135e-01 -1.37364358e-01
4.55762625e-01 -1.61501676e-01 -1.35129988e-01 -1.16697550e+00
5.37021756e-01 7.07940698e-01 1.24819510e-01 6.22482061e-01
-1.85233802e-02 -1.01318017e-01 1.19523540e-01 7.05331564e-01
1.97393095e+00 -1.39455244e-01 -7.93928206e-02 1.93676248e-01
8.52055177e-02 -3.41227591e-01 2.18537539e-01 6.21583402e-01
-6.31006539e-01 6.10169470e-01 9.93030131e-01 -3.12507987e-01
-1.11281109e+00 -7.64475822e-01 2.03691795e-01 1.72958505e+00
-2.94940323e-01 -8.84516180e-01 -1.34703732e+00 -7.65180707e-01
1.60452709e-01 4.26562279e-01 -5.38471520e-01 -1.12383015e-01
-3.19681108e-01 -1.72423478e-02 1.15850008e+00 -8.21274072e-02
9.46680531e-02 -7.65318513e-01 -4.48420435e-01 3.93648632e-02
1.55312881e-01 -1.16839492e+00 -7.61271477e-01 -6.79496765e-01
-2.25595921e-01 -1.29536617e+00 -5.14255881e-01 -5.03344536e-01
4.41547364e-01 5.93525827e-01 9.93518412e-01 6.79818153e-01
-4.80949730e-01 9.39032137e-01 -7.91661084e-01 -3.06596160e-01
-8.83074522e-01 -4.09114242e-01 1.66808724e-01 -1.73467979e-01
4.01547194e-01 -8.65019679e-01 -3.20227563e-01 1.29652470e-01
-9.99716461e-01 -1.37832418e-01 -4.92762089e-01 4.14104052e-02
-5.10915220e-01 1.59133092e-01 8.84007961e-02 -9.28921044e-01
2.02895209e-01 -1.04612517e+00 -3.61458153e-01 -4.29650724e-01
2.43227519e-02 -1.02120531e+00 6.34319365e-01 -1.31001878e+00
-5.08761287e-01 -4.28422779e-01 -1.10957809e-01 -8.29920590e-01
-5.26063025e-01 -4.12709564e-01 9.55579281e-02 -5.44926584e-01
5.63592196e-01 -2.04045922e-01 -2.31280014e-01 -4.12671447e-01
1.20774135e-01 6.92788064e-01 4.32449043e-01 -1.31523639e-01
9.74495053e-01 4.54156995e-01 -5.33270538e-01 -1.18654323e+00
-5.62977016e-01 -3.40438545e-01 -3.30640674e-02 -8.45926642e-01
9.07675624e-01 -1.01977897e+00 -1.04976022e+00 9.99285817e-01
-1.03562844e+00 -3.06628644e-01 4.32937115e-01 3.08616161e-01
-2.58952193e-02 6.50787234e-01 -1.35048270e+00 -5.42264581e-01
-1.14508249e-01 -9.88271534e-01 4.41653818e-01 3.17858696e-01
-3.55527461e-01 -7.31131554e-01 1.96483284e-01 8.71336877e-01
4.08311158e-01 5.15797734e-01 4.77455497e-01 -9.02898908e-01
-3.33347470e-01 -4.20951664e-01 1.01806477e-01 4.37486053e-01
-1.70169070e-01 1.10455871e+00 -1.29720926e+00 -3.19431931e-01
-1.93012476e-01 -6.89016879e-01 2.25654677e-01 1.32700056e-02
1.14380193e+00 -9.88951445e-01 7.90721998e-02 4.23490077e-01
9.08012450e-01 8.47586021e-02 6.09063625e-01 2.22401723e-01
9.98842061e-01 5.46444058e-01 1.22600392e-01 7.07318068e-01
2.90335298e-01 3.27570319e-01 9.09427047e-01 6.10829592e-01
-9.90559533e-03 -5.43542564e-01 8.36930215e-01 8.92586708e-01
-1.65503416e-02 -2.83353299e-01 -6.98179841e-01 5.88580489e-01
-1.42614686e+00 -1.53807843e+00 -2.91300088e-01 1.79013908e+00
7.69348383e-01 3.18981051e-01 5.19484222e-01 1.79468960e-01
9.56463277e-01 4.32990491e-01 -2.07875192e-01 -8.41457605e-01
4.38796699e-01 1.69706307e-02 4.88683134e-01 1.15448914e-01
-1.50044405e+00 8.40871334e-01 6.68626118e+00 9.48559999e-01
-1.16962743e+00 4.96943146e-01 5.33768237e-01 -7.51997590e-01
-3.73869091e-01 -3.77032578e-01 -5.75201392e-01 1.23672664e+00
9.39935803e-01 2.84998655e-01 6.51238203e-01 1.41534150e+00
3.15871984e-01 -7.51897544e-02 -1.12291908e+00 9.97505844e-01
6.83381200e-01 -1.51439476e+00 -4.84436661e-01 -1.13678016e-01
9.57099557e-01 6.34939549e-03 3.11003804e-01 2.28811860e-01
1.95704013e-01 -1.03893054e+00 1.06385541e+00 -1.89786136e-01
6.40883982e-01 -5.41501343e-01 2.39068307e-02 -2.18320265e-02
-6.50978148e-01 -1.29447669e-01 -1.18572302e-01 -3.05004895e-01
-4.62130517e-01 2.16952100e-01 -5.35013735e-01 -8.51134241e-01
1.22594202e+00 1.03968215e+00 -9.46238935e-01 1.13439357e+00
-2.89208323e-01 1.20977378e+00 -2.06480637e-01 -3.22944343e-01
1.76479682e-01 3.56381953e-01 9.48166132e-01 1.40440500e+00
1.06895596e-01 7.26018921e-02 -5.32581449e-01 2.89273709e-01
-3.02352875e-01 -7.83969462e-02 -9.81102884e-01 -4.98603344e-01
6.13200903e-01 1.30843973e+00 -7.75769472e-01 -2.67215610e-01
-7.60674596e-01 6.72219038e-01 1.25855878e-01 -9.76447091e-02
-1.20306385e+00 -2.01186627e-01 1.15060365e+00 5.39107740e-01
4.35758173e-01 1.84169456e-01 2.42081702e-01 -1.33943045e+00
-2.77377069e-01 -1.48032749e+00 1.39842167e-01 -8.94919813e-01
-1.06355000e+00 -1.20983668e-01 -1.29460871e-01 -1.06249821e+00
-1.92732021e-01 -1.81032494e-01 -1.15762413e+00 -4.72685814e-01
-5.70935786e-01 -7.92748272e-01 -1.60423353e-01 5.73421001e-01
5.43307960e-01 -2.20716894e-01 3.01807553e-01 7.10110307e-01
-7.17933953e-01 9.60062563e-01 -2.18659058e-01 3.45438063e-01
7.49240756e-01 -9.58585858e-01 1.35537207e-01 8.42329204e-01
-2.15616167e-01 1.85368568e-01 1.22983336e+00 -8.04282904e-01
-1.67878306e+00 -7.73666799e-01 4.33691472e-01 -7.66293824e-01
1.47474980e+00 -5.08819580e-01 -8.00092697e-01 7.00267971e-01
4.35522527e-01 -2.99721599e-01 9.74987805e-01 -2.60295898e-01
-9.53897655e-01 2.40238518e-01 -1.28381228e+00 6.07013404e-01
8.15306187e-01 -5.92647374e-01 -2.66031682e-01 6.70606613e-01
7.83225656e-01 -2.55328745e-01 -6.61070585e-01 -2.03481779e-01
8.11671674e-01 -1.37609422e+00 5.94095767e-01 -6.84864879e-01
1.04984391e+00 2.33100861e-01 2.12732390e-01 -8.12540174e-01
-5.77303767e-02 -1.38964379e+00 -3.48967373e-01 1.67713594e+00
-2.29316447e-02 -1.81962579e-01 1.03255320e+00 1.05962074e+00
7.16736764e-02 -1.31514132e-01 -7.34771609e-01 -3.48204285e-01
-2.25634009e-01 -7.58438408e-01 2.51463741e-01 1.42724812e+00
3.71392787e-01 -3.09181899e-01 -7.14861274e-01 1.65644616e-01
6.03844166e-01 -8.72695029e-01 7.75120497e-01 -5.25901914e-01
-4.50163364e-01 -7.45892942e-01 -9.89156604e-01 -1.42812133e-01
-6.15966432e-02 -1.81873560e-01 -4.84036684e-01 -5.18783987e-01
4.39240724e-01 1.79843724e-01 4.00384277e-01 3.66940677e-01
2.45471627e-01 8.44134212e-01 2.86681205e-01 -1.49581343e-01
-1.13306665e+00 -3.77101451e-01 7.66409039e-01 -1.56162322e-01
8.30666423e-02 1.37950443e-02 -5.11209011e-01 1.30377805e+00
3.93116981e-01 -7.78520346e-01 -4.67358768e-01 -2.99829543e-01
8.26234281e-01 -2.09870368e-01 3.67394477e-01 -7.22143412e-01
-1.25005275e-01 -1.21920332e-01 1.79621339e-01 -3.66271973e-01
3.18691194e-01 -4.97539550e-01 5.07026203e-02 4.15033966e-01
-3.78920972e-01 -1.09349541e-01 -1.18351439e-02 2.24246785e-01
1.74180090e-01 -5.26354909e-01 8.60824704e-01 3.14491764e-02
-3.51012588e-01 7.82116205e-02 -1.31452608e+00 4.72520202e-01
1.34865487e+00 -5.83029054e-02 -1.06566119e+00 -1.07657862e+00
-4.91043299e-01 -8.91727284e-02 9.28531170e-01 5.62716961e-01
6.18640125e-01 -1.00403976e+00 -5.01695633e-01 1.14666469e-01
-3.87508720e-01 -7.73943901e-01 3.06865901e-01 6.46400511e-01
-1.13149440e+00 -2.82391816e-01 -1.88879460e-01 -1.81600586e-01
-2.02360177e+00 8.55016351e-01 -2.13602290e-01 6.45121396e-01
-6.14648104e-01 1.24571931e+00 -1.97398365e-01 4.43582505e-01
6.52005613e-01 -2.67318226e-02 -3.27676892e-01 2.64208436e-01
1.22260916e+00 9.46545124e-01 -3.25353920e-01 -7.80561447e-01
-4.05637532e-01 -8.74736086e-02 -2.88601041e-01 3.68996054e-01
1.23803425e+00 1.23725265e-01 -3.26966345e-01 -6.12920262e-02
1.57825696e+00 7.43150592e-01 -1.09296441e+00 5.93462765e-01
-3.50415766e-01 -1.07299924e+00 -3.12132180e-01 -1.84184298e-01
-1.25068820e+00 5.30229867e-01 7.59802386e-02 1.29475558e+00
8.50374997e-01 1.28518865e-01 8.96638691e-01 -2.15001315e-01
5.05567193e-02 -1.18106902e+00 6.60829961e-01 1.35191947e-01
5.26180029e-01 -1.03620827e+00 3.66334230e-01 -4.92123455e-01
-4.69494760e-01 8.77607822e-01 7.53715158e-01 -2.78168917e-01
7.17264533e-01 4.69989240e-01 8.95915146e-04 -1.70445591e-01
-1.12214506e+00 5.17835259e-01 -3.04409415e-02 7.71696627e-01
2.45950356e-01 2.66695060e-02 -7.17381090e-02 3.89340073e-01
-3.61401081e-01 -5.66403985e-01 1.35186040e+00 9.37096536e-01
-5.62295556e-01 -3.83457899e-01 -6.25677764e-01 4.67898637e-01
-1.50155926e+00 -1.10985123e-01 -7.49384582e-01 4.77088779e-01
4.46974814e-01 1.30183423e+00 -1.36267534e-02 -7.06874132e-01
-4.83467162e-01 -3.86966020e-01 4.17020589e-01 -6.36898756e-01
-9.28884327e-01 2.00793818e-01 5.23246527e-01 -9.51805830e-01
-2.40891218e-01 -8.96340966e-01 -4.55089539e-01 -1.05282688e+00
-3.54640745e-02 -1.80164173e-01 8.23619366e-01 5.85934579e-01
-3.02396882e-02 -2.17954054e-01 7.69721210e-01 -1.01414216e+00
1.08438835e-01 -5.95207095e-01 -3.70961547e-01 6.75338626e-01
6.21231079e-01 -4.64513361e-01 -7.32022583e-01 1.36090681e-01] | [12.486841201782227, 1.1725817918777466] |
2a545a65-8b63-4de2-9407-acb3d62bf45d | unified-autoregressive-modeling-for-joint-end | 2107.01549 | null | https://arxiv.org/abs/2107.01549v1 | https://arxiv.org/pdf/2107.01549v1.pdf | Unified Autoregressive Modeling for Joint End-to-End Multi-Talker Overlapped Speech Recognition and Speaker Attribute Estimation | In this paper, we present a novel modeling method for single-channel multi-talker overlapped automatic speech recognition (ASR) systems. Fully neural network based end-to-end models have dramatically improved the performance of multi-taker overlapped ASR tasks. One promising approach for end-to-end modeling is autoregressive modeling with serialized output training in which transcriptions of multiple speakers are recursively generated one after another. This enables us to naturally capture relationships between speakers. However, the conventional modeling method cannot explicitly take into account the speaker attributes of individual utterances such as gender and age information. In fact, the performance deteriorates when each speaker is the same gender or is close in age. To address this problem, we propose unified autoregressive modeling for joint end-to-end multi-talker overlapped ASR and speaker attribute estimation. Our key idea is to handle gender and age estimation tasks within the unified autoregressive modeling. In the proposed method, transformer-based autoregressive model recursively generates not only textual tokens but also attribute tokens of each speaker. This enables us to effectively utilize speaker attributes for improving multi-talker overlapped ASR. Experiments on Japanese multi-talker overlapped ASR tasks demonstrate the effectiveness of the proposed method. | ['Shota Orihashi', 'Tomohiro Tanaka', 'Akihiko Takashima', 'Mana Ihori', 'Naoki Makishima', 'Daiki Okamura', 'Ryo Masumura'] | 2021-07-04 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [-1.69037312e-01 5.05233519e-02 -8.57111067e-03 -8.02682281e-01
-1.27399695e+00 -1.48269922e-01 1.75146312e-01 -2.45748729e-01
-2.64263660e-01 3.33596319e-01 5.58559954e-01 -2.95187205e-01
2.75321186e-01 -1.88580230e-01 -4.08537328e-01 -7.47365057e-01
3.85740787e-01 7.05931544e-01 -3.77661884e-01 -1.98821694e-01
-3.26872557e-01 2.55413294e-01 -1.27805829e+00 2.18798712e-01
9.27078366e-01 8.57011318e-01 5.49895406e-01 9.63982582e-01
-2.59686619e-01 6.88619554e-01 -9.46091056e-01 -5.07702768e-01
-1.67485774e-01 -1.29390717e-01 -4.28498834e-01 1.33026272e-01
3.13781530e-01 -4.92178202e-01 -6.16675138e-01 6.19576216e-01
9.51555371e-01 2.89343715e-01 6.53934121e-01 -1.03643405e+00
-5.50027072e-01 1.42096388e+00 -6.50322497e-01 3.90256345e-02
1.45930037e-01 -4.23547655e-01 9.83984053e-01 -1.18200779e+00
-2.88015276e-01 1.75426805e+00 3.62034440e-01 8.69108796e-01
-8.60630333e-01 -9.93225157e-01 5.83346069e-01 6.65680319e-02
-1.62045848e+00 -9.18888152e-01 8.27066779e-01 -1.84670150e-01
6.94342732e-01 2.90184855e-01 2.36580014e-01 1.08456981e+00
-4.38762456e-01 1.23112464e+00 7.04262197e-01 -4.37790573e-01
-2.13140681e-01 1.79100960e-01 4.22004521e-01 2.39514396e-01
-5.90078592e-01 -1.91497102e-01 -7.90605962e-01 -5.80963194e-02
5.26796043e-01 2.69317888e-02 3.28179896e-02 5.17205358e-01
-1.07215655e+00 6.62135780e-01 -1.97622120e-01 4.07136887e-01
-3.47962588e-01 2.89788470e-02 4.13265914e-01 2.52064466e-01
8.26715648e-01 -1.64867178e-01 -6.13952160e-01 -3.08140188e-01
-1.06009734e+00 6.93415925e-02 5.01507342e-01 1.11171067e+00
4.64734346e-01 7.06505299e-01 -3.26185793e-01 1.62625790e+00
4.85048234e-01 7.44788051e-01 5.99605560e-01 -5.41563570e-01
5.86219430e-01 4.43669483e-02 -1.97727725e-01 -2.63602376e-01
-6.64822012e-02 -5.83441257e-01 -1.02990484e+00 -5.22769630e-01
2.60660201e-01 -2.92919636e-01 -1.15227413e+00 1.77489579e+00
8.83497894e-02 5.11231981e-02 3.57489258e-01 5.64182520e-01
9.45121527e-01 1.00822461e+00 2.91723579e-01 -6.08890116e-01
1.56064153e+00 -1.03267145e+00 -1.13042152e+00 -2.25035220e-01
5.68776786e-01 -9.26963210e-01 8.96870255e-01 3.92356575e-01
-1.33847511e+00 -7.53992260e-01 -6.90367997e-01 1.58006568e-02
6.13163039e-02 7.02522874e-01 2.45959297e-01 9.48868275e-01
-8.86862159e-01 -7.39544779e-02 -6.32648408e-01 1.13540441e-01
-9.80685279e-02 7.12010443e-01 -1.44935608e-01 1.03070535e-01
-1.27248597e+00 4.39055353e-01 4.99793403e-02 2.35996440e-01
-8.22854578e-01 -5.66343546e-01 -7.64254212e-01 2.58176148e-01
2.42773235e-01 -5.26802659e-01 1.73337698e+00 -9.28772092e-01
-1.78566742e+00 5.54936588e-01 -9.45550621e-01 -3.98666680e-01
2.30411679e-01 -3.06952924e-01 -5.62003613e-01 -1.99556112e-01
-4.11448956e-01 2.57393152e-01 8.54951382e-01 -1.10024977e+00
-6.73620522e-01 -8.94019961e-01 -5.01951635e-01 4.81966227e-01
-6.96522534e-01 6.33993924e-01 -4.30017501e-01 -8.29209208e-01
2.96383947e-01 -7.36758113e-01 -6.44377545e-02 -8.06301415e-01
-3.96786213e-01 -6.28411770e-01 6.92892671e-01 -1.19457150e+00
1.44669247e+00 -2.20978665e+00 1.09872699e-01 -8.40052664e-02
6.66000545e-02 1.56306848e-01 6.02418147e-02 8.86886865e-02
-1.12784982e-01 -8.52821097e-02 2.87609845e-01 -1.00541186e+00
-9.47416425e-02 2.06164941e-02 -4.54727530e-01 5.81456907e-02
-3.93868953e-01 3.50873739e-01 -3.63825768e-01 -5.32644510e-01
1.11340322e-01 8.24052870e-01 -1.41782954e-01 6.67337596e-01
1.60339534e-01 5.75740039e-01 -3.70176226e-01 6.86024547e-01
6.10346735e-01 5.20418465e-01 -1.28376707e-01 -1.42601550e-01
1.11153945e-02 3.89251649e-01 -1.07366335e+00 1.22329068e+00
-1.02513754e+00 4.59017605e-01 5.01225829e-01 -8.78966928e-01
1.48693871e+00 9.75064397e-01 3.18231940e-01 -3.07286024e-01
1.65147364e-01 2.76614904e-01 6.33974075e-02 -2.35089943e-01
8.49906683e-01 -2.78188199e-01 -1.84541911e-01 9.96104479e-02
4.10777424e-03 -5.57657657e-03 -2.27762520e-01 1.64978318e-02
3.59919280e-01 -5.62693298e-01 1.49242610e-01 2.28320852e-01
7.35085964e-01 -8.29720497e-01 7.29644656e-01 6.52629137e-01
-1.72842681e-01 6.57837987e-01 -7.75700212e-02 -4.21762243e-02
-1.01863432e+00 -1.03696537e+00 2.18849972e-01 1.74201000e+00
-4.37205374e-01 -2.87603050e-01 -7.23824143e-01 -3.17268789e-01
-3.67905021e-01 9.25797105e-01 -2.33373493e-01 2.74371989e-02
-6.91786289e-01 -4.96364057e-01 7.49826074e-01 7.84650624e-01
2.94286609e-01 -6.51485741e-01 4.31861609e-01 4.06267792e-01
-6.84324503e-01 -1.30729413e+00 -8.82595658e-01 1.08139860e-02
-7.25251794e-01 -1.10048920e-01 -1.29981101e+00 -9.64501321e-01
5.02869010e-01 2.62211442e-01 6.77190065e-01 -3.35531682e-01
2.19789311e-01 3.45375717e-01 -3.71464938e-01 -6.23004079e-01
-6.96412444e-01 2.44438201e-01 5.34221768e-01 3.66971701e-01
3.01508307e-01 -4.48795497e-01 -1.93988591e-01 6.45236552e-01
-2.61353552e-01 -2.55722161e-02 6.20998621e-01 8.88369977e-01
3.29689294e-01 -6.97118789e-02 1.01469207e+00 -5.20773768e-01
5.42019784e-01 -2.80597806e-01 -1.66159317e-01 4.96374369e-01
-1.63031563e-01 -2.51593515e-02 7.34855056e-01 -8.69610071e-01
-1.45627511e+00 1.16176695e-01 -5.80622435e-01 -7.16164827e-01
-1.40853465e-01 4.01261181e-01 -6.82767212e-01 6.34964466e-01
4.74691167e-02 6.51491344e-01 -1.05837017e-01 -7.05492616e-01
2.37957269e-01 1.64252675e+00 6.64889753e-01 -6.91950858e-01
6.04890108e-01 -2.05286995e-01 -6.52688205e-01 -1.25197732e+00
-4.14833575e-01 -7.20370948e-01 -2.86559016e-01 -1.41747549e-01
6.72127306e-01 -1.52713621e+00 -9.92750525e-01 9.45340216e-01
-1.28217423e+00 -7.92026967e-02 3.45953852e-01 9.13632751e-01
-4.72333401e-01 1.94751620e-01 -7.28054106e-01 -1.64355266e+00
-6.54329598e-01 -1.32113242e+00 9.99389112e-01 1.79143906e-01
-2.13183552e-01 -7.25694716e-01 -4.68173981e-01 8.20828378e-01
3.56325299e-01 -8.23673368e-01 7.04258978e-01 -1.05435193e+00
-1.20851398e-01 -1.31720304e-01 1.67283610e-01 4.53990519e-01
2.17555761e-01 -2.16108654e-02 -1.08629060e+00 -2.75212914e-01
-2.13516653e-02 -1.39562309e-01 6.14868283e-01 7.67528117e-01
1.26905024e+00 -3.17586035e-01 -1.94242314e-01 2.41553962e-01
5.85056722e-01 4.59719330e-01 4.32104766e-01 -4.38261449e-01
9.47449446e-01 8.07136357e-01 6.67009115e-01 5.56561291e-01
6.68784380e-01 8.54912937e-01 -1.86549231e-01 -6.57409504e-02
9.14490744e-02 -3.03850919e-01 7.59577930e-01 1.48979867e+00
-1.14848167e-01 -2.73217708e-01 -7.38836467e-01 7.81199217e-01
-1.54859459e+00 -7.82581866e-01 -5.61586767e-02 2.39667416e+00
8.36455941e-01 -1.26752287e-01 3.85229111e-01 7.34171495e-02
1.12869704e+00 3.36042121e-02 -3.00975680e-01 -5.73329031e-01
-1.10362947e-01 -2.52866685e-01 3.04552853e-01 5.61509550e-01
-6.67921782e-01 9.15306449e-01 5.67788839e+00 1.31268990e+00
-9.40003812e-01 1.80263981e-01 1.01144373e+00 -1.38488352e-01
-1.70755699e-01 -3.08857292e-01 -1.33953393e+00 3.17861408e-01
1.23414660e+00 -4.03553367e-01 4.66053426e-01 1.00087690e+00
5.90825260e-01 2.20556393e-01 -1.05885923e+00 1.35109973e+00
2.58724719e-01 -5.51961601e-01 9.82741863e-02 4.04063687e-02
2.74259478e-01 -3.50760072e-01 4.77726310e-01 5.83325446e-01
2.52613008e-01 -9.21824992e-01 7.91890264e-01 2.37147082e-02
9.13532078e-01 -1.04649401e+00 5.49917459e-01 4.31836277e-01
-1.47862899e+00 -3.31057280e-01 -1.51835218e-01 1.50085554e-01
2.48376280e-01 5.18523455e-01 -1.19017661e+00 5.71662664e-01
3.97582889e-01 1.64074436e-01 -7.89486319e-02 6.24542356e-01
-3.90343554e-02 9.03940380e-01 -1.65205672e-01 -2.47683711e-02
-1.81259736e-01 1.37389109e-01 6.50346041e-01 1.19573975e+00
7.09908009e-01 2.03987956e-01 1.21068954e-01 4.01887208e-01
-1.18489861e-01 4.66370404e-01 -3.97888184e-01 -1.21867180e-01
9.51148331e-01 1.01497483e+00 -2.19384998e-01 -5.67121089e-01
-2.62215078e-01 8.13227117e-01 1.38647199e-01 5.74601173e-01
-6.28422499e-01 -1.76809311e-01 5.66889226e-01 1.26194432e-01
1.85550332e-01 -4.11122710e-01 -2.11554796e-01 -1.03547812e+00
-1.66670024e-01 -1.03037667e+00 2.35925004e-01 -7.63420582e-01
-1.14343011e+00 8.22877049e-01 -2.59192158e-02 -9.11425292e-01
-4.65723962e-01 -1.92870349e-01 -5.85845411e-01 1.14097273e+00
-1.10838115e+00 -1.64702809e+00 1.51783258e-01 6.59964681e-01
1.29058063e+00 -5.95414519e-01 7.82197654e-01 4.95629758e-01
-6.94556296e-01 1.26437986e+00 1.08931892e-01 2.53703386e-01
8.78554463e-01 -1.10385442e+00 2.70937890e-01 7.18212068e-01
5.84625714e-02 8.36295962e-01 8.52348328e-01 -4.98725116e-01
-1.10638869e+00 -9.32451129e-01 1.24105012e+00 -5.87282330e-02
2.89396554e-01 -7.38801718e-01 -9.25668597e-01 8.52727234e-01
6.14944398e-02 -5.69657087e-01 8.29216897e-01 8.13601673e-01
-2.29390368e-01 -4.41134870e-01 -6.76370740e-01 6.29087210e-01
5.13370454e-01 -6.37272537e-01 -5.37677646e-01 5.39585054e-02
1.02016163e+00 -3.69580805e-01 -7.83524334e-01 2.30444223e-01
6.24196947e-01 -4.83052671e-01 8.64078522e-01 -2.33031586e-01
1.54505074e-01 1.34244487e-02 -4.20927852e-01 -1.49412620e+00
1.16524830e-01 -8.48051429e-01 -1.79692641e-01 1.88911974e+00
4.59880114e-01 -3.90523225e-01 5.78366280e-01 8.23448300e-01
-5.06267130e-01 -6.64985538e-01 -8.32154393e-01 -7.80011535e-01
-3.73134539e-02 -5.72812736e-01 7.48523116e-01 5.02281785e-01
-1.24751650e-01 5.55156350e-01 -8.69453907e-01 4.60124463e-01
5.22549808e-01 -3.69409591e-01 7.79916286e-01 -9.86809134e-01
-3.90707523e-01 -1.76251724e-01 -3.95816900e-02 -1.46214223e+00
5.52018404e-01 -4.57443714e-01 3.00252974e-01 -1.25650001e+00
2.17958897e-01 -3.87010604e-01 -2.68190861e-01 4.59878415e-01
-4.77517158e-01 -3.44431669e-01 1.10910565e-01 5.40418848e-02
-2.64750302e-01 9.61092651e-01 8.92191648e-01 -1.89372435e-01
-5.67478836e-01 7.19617605e-01 -5.92973232e-01 6.21905684e-01
8.06500614e-01 -3.98213297e-01 -3.60398173e-01 -4.33008790e-01
-4.59479034e-01 5.85957348e-01 -1.47080407e-01 -6.68191373e-01
3.62161666e-01 5.74410111e-02 1.55512974e-01 -9.75718141e-01
8.12004030e-01 -6.49014771e-01 -9.51933935e-02 4.77548875e-02
-5.14601946e-01 -7.80135989e-02 -2.48932838e-02 3.35601270e-01
-4.61180121e-01 2.62674373e-02 6.02281094e-01 8.65308270e-02
-2.30638385e-01 3.38089556e-01 -7.82365561e-01 -2.63720900e-01
6.17288113e-01 -1.82798281e-01 2.30327457e-01 -8.99034679e-01
-1.04832196e+00 3.17961127e-01 -2.72593588e-01 5.56718469e-01
7.41792738e-01 -1.30267775e+00 -1.10443234e+00 6.32027015e-02
3.42110768e-02 -1.49350300e-01 9.34523582e-01 6.85664237e-01
3.79010797e-01 3.51075858e-01 2.61201888e-01 -4.92924452e-01
-2.11562490e+00 2.90380508e-01 2.55599499e-01 -3.01008850e-01
3.72967646e-02 1.10613096e+00 5.13001740e-01 -5.32557905e-01
5.97316921e-01 -8.13809931e-02 -3.52397472e-01 1.34567007e-01
6.32884562e-01 3.77376258e-01 -6.86069438e-03 -9.97194827e-01
-1.65387318e-01 3.83253902e-01 -5.66352487e-01 -4.94335234e-01
1.07882965e+00 -6.81924284e-01 2.07027301e-01 9.52332675e-01
1.11205459e+00 1.97040066e-01 -8.63599479e-01 -5.64913571e-01
-6.35799587e-01 -3.33232045e-01 1.77770004e-01 -6.11330211e-01
-1.04419053e+00 1.24725819e+00 3.39129508e-01 -2.25831032e-01
1.05260372e+00 7.88224041e-02 9.78218853e-01 2.47860834e-01
1.92937523e-01 -1.41537166e+00 6.42925128e-02 6.60146296e-01
8.44207168e-01 -1.12075591e+00 -4.34145272e-01 -3.97501916e-01
-9.84101772e-01 1.03807080e+00 8.05443883e-01 7.61683047e-01
3.81673098e-01 9.31033418e-02 2.76146084e-01 4.42327350e-01
-7.55929351e-01 -1.06334545e-01 1.93684146e-01 5.81073225e-01
7.90692270e-01 4.62907642e-01 2.28061043e-02 1.20444322e+00
-5.52248657e-01 -5.24804950e-01 4.29391235e-01 2.25684375e-01
-4.37564373e-01 -1.19027221e+00 -8.92826617e-01 3.13455909e-01
-7.49517202e-01 -2.96480536e-01 -1.12449326e-01 4.41959947e-02
-2.82469243e-01 1.50064862e+00 4.34867181e-02 -6.02799714e-01
2.21834615e-01 3.69695276e-01 2.04364955e-01 -5.27339399e-01
-5.78319192e-01 6.51345551e-01 2.74054289e-01 2.85471231e-01
4.66849692e-02 -6.49423778e-01 -1.18192935e+00 -1.40321746e-01
-5.59974432e-01 3.40551108e-01 1.02923977e+00 9.33739424e-01
1.18341982e-01 6.99455619e-01 1.10300338e+00 -6.38303459e-01
-7.75810719e-01 -1.52101791e+00 -9.51933861e-01 8.26698821e-03
4.37867135e-01 -4.37742859e-01 -3.03944826e-01 1.07841402e-01] | [14.63017463684082, 6.349122047424316] |
635f0ad8-1758-4f97-b86b-91bec7f60b5c | quant-quantum-annealing-with-learnt-couplings | 2210.08114 | null | https://arxiv.org/abs/2210.08114v1 | https://arxiv.org/pdf/2210.08114v1.pdf | QuAnt: Quantum Annealing with Learnt Couplings | Modern quantum annealers can find high-quality solutions to combinatorial optimisation objectives given as quadratic unconstrained binary optimisation (QUBO) problems. Unfortunately, obtaining suitable QUBO forms in computer vision remains challenging and currently requires problem-specific analytical derivations. Moreover, such explicit formulations impose tangible constraints on solution encodings. In stark contrast to prior work, this paper proposes to learn QUBO forms from data through gradient backpropagation instead of deriving them. As a result, the solution encodings can be chosen flexibly and compactly. Furthermore, our methodology is general and virtually independent of the specifics of the target problem type. We demonstrate the advantages of learnt QUBOs on the diverse problem types of graph matching, 2D point cloud alignment and 3D rotation estimation. Our results are competitive with the previous quantum state of the art while requiring much fewer logical and physical qubits, enabling our method to scale to larger problems. The code and the new dataset will be open-sourced. | ['Vladislav Golyanik', 'Michael Moeller', 'Zorah Lähner', 'Edith Tretschk', 'Maximilian Krahn', 'Marcel Seelbach Benkner'] | 2022-10-13 | null | null | null | null | ['3d-rotation-estimation', 'graph-matching'] | ['computer-vision', 'graphs'] | [ 5.50863326e-01 5.48606068e-02 -2.98961878e-01 -1.40734360e-01
-1.00446558e+00 -8.15168679e-01 4.46033001e-01 -4.18310314e-02
-4.71372962e-01 7.30500281e-01 -4.93561715e-01 -5.19195855e-01
-5.34182250e-01 -7.74021506e-01 -8.18088055e-01 -9.38703060e-01
4.18409705e-02 6.57369971e-01 -2.40345463e-01 -3.43505383e-01
8.77796650e-01 5.88845849e-01 -1.26424026e+00 -3.92647862e-01
6.64591789e-01 1.17956078e+00 3.78600471e-02 9.60939705e-01
1.82541624e-01 5.16883492e-01 -3.44888896e-01 -3.51188570e-01
6.70462608e-01 -4.09502476e-01 -9.22983170e-01 -4.75246273e-02
7.72125542e-01 -4.64891940e-02 -6.40023291e-01 1.34597635e+00
6.67839110e-01 -4.88377474e-02 5.37144542e-01 -1.15780842e+00
-1.06900275e+00 2.08361119e-01 -3.61382216e-01 2.23067049e-02
2.81184554e-01 3.90358001e-01 1.56202495e+00 -4.49141055e-01
7.45672047e-01 1.00926805e+00 6.51649177e-01 6.28414750e-01
-1.37965322e+00 -4.54441339e-01 -4.92902339e-01 5.90173841e-01
-1.55152261e+00 -6.38254523e-01 8.00124168e-01 -5.57336211e-02
1.12221241e+00 3.58649909e-01 6.19058311e-01 6.17637336e-01
4.27019686e-01 3.19905549e-01 1.05505276e+00 -6.71208382e-01
4.16220278e-01 -2.92218596e-01 -2.75599901e-02 9.75756109e-01
3.21335018e-01 3.20469171e-01 -7.17884898e-01 -1.52706847e-01
6.07610285e-01 -6.24985158e-01 -3.04395314e-02 -7.75463939e-01
-1.34966314e+00 9.80137527e-01 7.44126320e-01 -3.04478824e-01
-6.59636557e-02 8.52738559e-01 -1.88335001e-01 4.99130279e-01
-2.83869416e-01 9.81774092e-01 -1.24137543e-01 -2.85109967e-01
-5.90247869e-01 4.99147117e-01 8.96658182e-01 1.16863799e+00
1.15124166e+00 -2.03539759e-01 1.61483452e-01 2.31002465e-01
3.82149160e-01 9.02566910e-01 -1.75773814e-01 -1.43606663e+00
3.90103340e-01 5.14948294e-02 1.83129326e-01 -1.07474482e+00
-4.33421046e-01 -3.61186713e-01 -5.82276940e-01 1.40024498e-01
5.89639783e-01 -1.11020558e-01 -8.53325605e-01 1.53378046e+00
3.31359327e-01 -7.17067495e-02 9.30324793e-02 1.16815150e+00
6.49853885e-01 5.81893623e-01 -8.25788915e-01 6.71428209e-03
1.18037689e+00 -8.44537377e-01 -5.83258629e-01 -4.29618567e-01
6.11838043e-01 -8.47966254e-01 6.05223060e-01 5.11706650e-01
-1.03520656e+00 -5.77130094e-02 -1.33727098e+00 -5.21411717e-01
4.07350436e-03 -4.58338857e-01 1.27757680e+00 9.59188521e-01
-1.33190763e+00 6.86760724e-01 -7.97632515e-01 -1.53195009e-01
3.40274930e-01 9.45125699e-01 -6.28204346e-02 -1.33962780e-01
-9.70726848e-01 1.07458067e+00 4.86807585e-01 2.58962870e-01
-3.44296604e-01 -3.55533421e-01 -7.76728153e-01 -1.45705700e-01
4.88935679e-01 -9.91144121e-01 1.25745201e+00 -4.93830174e-01
-1.96920967e+00 8.68583322e-01 -2.79595368e-02 -4.52387661e-01
1.40125796e-01 4.65032637e-01 1.27175376e-01 2.70547122e-01
-2.21563712e-01 7.95868516e-01 8.38099182e-01 -6.09356344e-01
-1.06255166e-01 -3.74832958e-01 4.75425243e-01 1.96960315e-01
7.77521506e-02 -1.52074039e-01 -4.76942569e-01 1.13210358e-01
5.37577212e-01 -1.31220114e+00 -4.11962658e-01 1.86459616e-01
-2.94195116e-01 3.12022828e-02 4.03190762e-01 5.51177748e-03
7.67204821e-01 -1.80298102e+00 7.18687475e-01 2.52168924e-01
3.30786973e-01 -5.85105158e-02 -1.50734589e-01 3.47535640e-01
2.11569086e-01 2.75745653e-02 -2.76131004e-01 -2.53989231e-02
4.25875843e-01 4.36103672e-01 -6.94169626e-02 9.93227303e-01
3.27175289e-01 1.34433234e+00 -1.09496915e+00 -4.15978700e-01
2.43320137e-01 1.18909448e-01 -9.07525778e-01 -3.20153713e-01
-2.42624611e-01 4.87863183e-01 -6.48013413e-01 6.93750858e-01
7.40128160e-01 -5.37070692e-01 2.04848543e-01 -2.49941245e-01
-1.88955322e-01 6.55887187e-01 -1.23278880e+00 2.08176804e+00
-1.54330477e-01 7.23471522e-01 2.53238112e-01 -1.36828494e+00
7.34787583e-01 -2.25446776e-01 4.90773410e-01 -8.97491693e-01
3.37112993e-01 4.47555304e-01 3.48081768e-01 -3.12509775e-01
6.92036092e-01 -2.57881522e-01 -1.60138115e-01 3.43302399e-01
1.40336350e-01 -9.77176309e-01 2.24129915e-01 3.67483571e-02
1.03753483e+00 1.66181892e-01 3.21430445e-01 -1.74965516e-01
1.72061846e-01 9.45840329e-02 3.63266289e-01 1.06733322e+00
-3.29708666e-01 6.11775875e-01 4.85789746e-01 -2.36381486e-01
-1.13926589e+00 -9.66102839e-01 -6.33897901e-01 3.98547351e-01
7.68283546e-01 -5.65392017e-01 -4.15693641e-01 -1.49102166e-01
1.72090679e-01 3.93441945e-01 -1.02837101e-01 -8.10847357e-02
-6.68621957e-01 -1.08451915e+00 6.16751492e-01 -9.82526168e-02
1.89627290e-01 -6.00937188e-01 -6.12802267e-01 1.40001059e-01
1.17613398e-01 -1.35407019e+00 -1.43228829e-01 4.09738958e-01
-7.55648434e-01 -1.00722277e+00 -3.36016268e-01 -5.57843924e-01
5.32718539e-01 1.93784922e-01 9.31387305e-01 -3.70790064e-02
-4.85800743e-01 2.61802703e-01 -9.09610912e-02 -1.44179702e-01
-3.21836114e-01 2.36264467e-01 9.42831635e-02 -3.32308531e-01
2.32708126e-01 -5.71186364e-01 -4.05163169e-01 -1.35272041e-01
-4.93840992e-01 -2.24400050e-04 5.30162692e-01 9.54714835e-01
7.17959166e-01 -1.23814307e-01 -8.34789947e-02 -5.02670348e-01
3.61343920e-01 -1.32362247e-01 -1.17441535e+00 1.52548701e-01
-7.79484689e-01 6.92167521e-01 6.01242006e-01 -3.74090336e-02
-2.48357862e-01 1.09984517e-01 6.12494648e-02 -4.89555262e-02
2.72212744e-01 4.17433858e-01 2.09201619e-01 -1.14163160e+00
7.01290071e-01 1.43576954e-02 -3.56755145e-02 6.86976016e-02
5.38451970e-01 6.81347191e-01 5.00849664e-01 -7.75315285e-01
9.02992845e-01 4.27497387e-01 7.16859937e-01 -8.84535730e-01
-8.04200470e-01 -3.12383026e-01 -5.19309044e-01 4.45509888e-02
7.62458265e-01 -4.81049061e-01 -1.16569591e+00 3.95375609e-01
-1.22061396e+00 -4.83097397e-02 -1.29076149e-02 6.02372229e-01
-8.91135931e-01 5.86646736e-01 -4.52333272e-01 -9.13440645e-01
-1.83861971e-01 -1.59221792e+00 1.27012730e+00 2.62912214e-01
1.86150447e-01 -7.65057027e-01 -5.74853010e-02 5.47276914e-01
2.32367665e-01 2.37042353e-01 7.38839507e-01 3.41367126e-01
-1.42369032e+00 -9.85560343e-02 -3.84504378e-01 -7.94475898e-02
-1.90205544e-01 -1.11680716e-01 -7.95462847e-01 -3.81420612e-01
3.30730490e-02 -6.65931761e-01 8.01378787e-01 2.08178937e-01
9.99047339e-01 -1.57475457e-01 -5.62054478e-02 1.13473845e+00
1.36748040e+00 -2.32094005e-01 5.24396002e-01 4.53324139e-01
7.49967635e-01 1.26576707e-01 3.62873733e-01 1.27132729e-01
5.75641334e-01 1.03064847e+00 7.54551947e-01 4.42558438e-01
-7.77761340e-02 2.10234210e-01 2.43080750e-01 1.05885887e+00
-2.66167596e-02 -1.05086528e-01 -1.02555048e+00 2.22365960e-01
-1.88719964e+00 -8.35546494e-01 -1.17645681e-01 2.03250790e+00
7.41088092e-01 9.66194198e-02 -3.08527142e-01 -6.17807452e-03
4.49421495e-01 3.33229691e-01 -9.26542103e-01 -6.62491441e-01
-1.96940258e-01 6.63963377e-01 1.21346784e+00 5.97850025e-01
-1.00666296e+00 1.01023018e+00 6.83680677e+00 7.84502447e-01
-1.19677138e+00 -3.34765203e-02 6.57154247e-02 -2.17148095e-01
-4.48904991e-01 4.81323361e-01 -6.76760435e-01 7.60017633e-02
8.02431524e-01 1.00370601e-01 1.30565834e+00 2.86328137e-01
-3.33091050e-01 -8.32163021e-02 -1.24193811e+00 1.68913531e+00
-3.40958089e-02 -1.68912303e+00 -3.41096967e-01 3.50739360e-01
7.69029737e-01 5.73737085e-01 1.73338115e-01 -1.33301720e-01
1.93210483e-01 -1.24975252e+00 8.05937648e-01 4.09744799e-01
8.54965746e-01 -5.49898982e-01 3.93713444e-01 2.67612487e-01
-5.96116424e-01 -1.45830214e-01 -6.48025692e-01 -4.95304674e-01
5.50012216e-02 2.57101893e-01 -5.57648182e-01 8.07671905e-01
4.41010714e-01 6.29442930e-01 -4.57484037e-01 1.05828643e+00
-2.96704143e-01 4.35144365e-01 -8.82648647e-01 -6.21091545e-01
5.80014229e-01 -5.93978405e-01 7.19521880e-01 6.57160461e-01
2.08049431e-01 4.80516590e-02 -5.57810478e-02 1.11801708e+00
-9.37167108e-02 -1.89678669e-01 -4.13005650e-01 -5.48764050e-01
4.05628383e-01 1.21394300e+00 -6.29873991e-01 3.45026642e-01
-1.33455202e-01 7.87849605e-01 5.24959505e-01 2.71956831e-01
-7.43532717e-01 -6.03288352e-01 5.65504372e-01 -4.97134864e-01
5.95798135e-01 -7.29056776e-01 -6.72547936e-01 -1.53034639e+00
1.73859030e-01 -8.79862845e-01 -5.21336757e-02 -5.44122398e-01
-1.14530718e+00 -4.62855026e-02 -2.69274652e-01 -8.69439662e-01
-2.85348266e-01 -1.15952659e+00 -1.05785377e-01 7.74763465e-01
-1.68790293e+00 -8.65768075e-01 1.08712681e-01 2.13629063e-02
-3.18433046e-01 1.43066272e-01 1.01351321e+00 1.71490788e-01
-3.93072814e-01 5.84125161e-01 3.82719874e-01 -7.68467560e-02
3.13226074e-01 -1.33587086e+00 5.64380467e-01 8.25599432e-01
5.27309477e-01 7.48303354e-01 7.83064306e-01 -1.25832409e-01
-2.65358663e+00 -2.72114336e-01 6.02850497e-01 -7.34313667e-01
1.09337795e+00 -6.64993882e-01 -2.29738250e-01 3.71122658e-01
-1.64519101e-02 6.22544140e-02 3.40869993e-01 2.86183506e-01
-4.31772172e-01 -3.96840051e-02 -1.05885732e+00 4.31972682e-01
1.27819300e+00 -7.41462946e-01 -3.37970436e-01 6.76090300e-01
4.38912064e-01 -9.66220975e-01 -7.23176599e-01 2.29236946e-01
5.79699039e-01 -6.65423453e-01 1.18801844e+00 -5.00189722e-01
2.06028640e-01 -5.24844289e-01 -3.48370075e-01 -8.20826769e-01
-1.94244936e-01 -1.29561913e+00 -4.19054002e-01 2.94877172e-01
2.91635066e-01 -7.59314001e-01 8.86065841e-01 5.43858647e-01
-7.17915446e-02 -7.34098256e-01 -1.42685950e+00 -6.54144883e-01
2.60963470e-01 -6.31242514e-01 6.56968534e-01 9.47028100e-01
-1.00410217e-02 5.34916520e-01 -4.57061887e-01 4.37278926e-01
1.02855122e+00 5.22591233e-01 8.21800470e-01 -7.67094791e-01
-6.99312389e-01 -7.07513392e-01 -9.09292817e-01 -1.48570609e+00
1.12191014e-01 -1.29106927e+00 -1.33421468e-02 -1.26261199e+00
-6.86210692e-02 -7.53667831e-01 -1.39933527e-01 3.37343633e-01
3.96789983e-02 4.77006972e-01 2.94342786e-01 1.28399789e-01
-7.90798903e-01 6.88734412e-01 1.53309441e+00 -4.05272603e-01
1.38413325e-01 -2.67264843e-01 -6.56038523e-01 3.25945377e-01
6.27610624e-01 -5.87833703e-01 -1.38846517e-01 -8.11852634e-01
9.08716857e-01 -1.54196033e-02 5.81087768e-01 -8.75961483e-01
5.44034183e-01 -3.36803347e-01 -2.67637912e-02 -4.28642303e-01
6.50807261e-01 -4.17834461e-01 -7.93763176e-02 4.28598851e-01
-5.38773015e-02 -1.18269734e-01 -5.34923039e-02 3.58720809e-01
2.20161527e-01 -5.94787121e-01 6.62421167e-01 -1.14919707e-01
-6.24599576e-01 3.38406712e-01 1.52839702e-02 2.28767514e-01
7.40476191e-01 -3.88020486e-01 -3.95314455e-01 -4.51586664e-01
-1.49443507e-01 3.89937818e-01 7.31834173e-01 1.93406325e-02
3.61657590e-01 -1.17421448e+00 -4.80641067e-01 5.58395348e-02
1.46097362e-01 2.36845896e-01 -2.15615481e-01 8.82728338e-01
-1.06206036e+00 8.15660119e-01 -9.52143595e-02 -7.91475534e-01
-6.85503423e-01 4.66193110e-01 5.40632427e-01 1.49041265e-01
-4.55067873e-01 9.59627032e-01 -4.88311917e-01 -9.41947222e-01
-4.53791432e-02 -2.90626049e-01 5.04568458e-01 -4.90003407e-01
1.36627004e-01 1.59988314e-01 6.27795160e-02 -7.10754335e-01
-5.21601737e-01 8.12216163e-01 2.37530261e-01 -1.77895248e-01
1.32013822e+00 -2.99942959e-02 -4.30241823e-01 1.51937738e-01
1.30515325e+00 -3.15242708e-02 -1.07297814e+00 -1.01895981e-01
2.69485004e-02 -3.14253092e-01 2.51158535e-01 -3.72818470e-01
-8.10971558e-01 8.73341978e-01 3.70585352e-01 6.45304173e-02
7.78470755e-01 1.32855758e-01 9.51332808e-01 1.25104821e+00
7.95070827e-01 -9.59747910e-01 -2.33352453e-01 7.57396996e-01
3.84559274e-01 -1.19066679e+00 3.31239671e-01 -3.73737752e-01
1.10189572e-01 1.34835327e+00 6.19650520e-02 -3.34479243e-01
2.30753317e-01 3.11004892e-02 4.72145854e-03 -4.18926805e-01
-3.75426084e-01 -1.47970468e-01 4.76160496e-01 4.89187539e-01
5.06683551e-02 3.03339213e-02 -2.46112436e-01 -3.09133768e-01
-8.17453325e-01 -1.97241023e-01 6.19980693e-01 9.73958790e-01
-9.39619765e-02 -1.54621589e+00 -4.36711013e-01 2.26492986e-01
-1.94914535e-01 -2.76059061e-01 -4.36364442e-01 5.34930229e-01
-3.50939017e-03 8.51548016e-01 -1.93972901e-01 -3.18201959e-01
-4.94300872e-02 -2.63514638e-01 1.44430327e+00 -5.99233925e-01
-1.36022642e-01 -3.73014301e-01 1.99847315e-02 -6.78227961e-01
-5.71087301e-01 -6.99454248e-01 -1.36889648e+00 -4.59579110e-01
-7.56613612e-01 -3.89669538e-02 8.62815082e-01 1.09013462e+00
5.02888799e-01 2.38734335e-01 3.51632625e-01 -8.53495896e-01
-1.17405999e+00 -3.60268861e-01 -1.93655327e-01 -7.56071806e-02
6.40233040e-01 -8.01131368e-01 -1.41860008e-01 -4.57268596e-01] | [5.620166301727295, 4.901275634765625] |
a485c124-61d1-43aa-b970-9052602f13d9 | a-vector-based-representation-to-enhance-head | 2010.07184 | null | https://arxiv.org/abs/2010.07184v2 | https://arxiv.org/pdf/2010.07184v2.pdf | A Vector-based Representation to Enhance Head Pose Estimation | This paper proposes to use the three vectors in a rotation matrix as the representation in head pose estimation and develops a new neural network based on the characteristic of such representation. We address two potential issues existed in current head pose estimation works: 1. Public datasets for head pose estimation use either Euler angles or quaternions to annotate data samples. However, both of these annotations have the issue of discontinuity and thus could result in some performance issues in neural network training. 2. Most research works report Mean Absolute Error (MAE) of Euler angles as the measurement of performance. We show that MAE may not reflect the actual behavior especially for the cases of profile views. To solve these two problems, we propose a new annotation method which uses three vectors to describe head poses and a new measurement Mean Absolute Error of Vectors (MAEV) to assess the performance. We also train a new neural network to predict the three vectors with the constraints of orthogonality. Our proposed method achieves state-of-the-art results on both AFLW2000 and BIWI datasets. Experiments show our vector-based annotation method can effectively reduce prediction errors for large pose angles. | ['Yingjie Chen', 'Dongfang Liu', 'Zongcheng Chu', 'Zhiwen Cao'] | 2020-10-14 | null | null | null | null | ['head-pose-estimation'] | ['computer-vision'] | [-3.16770166e-01 1.11026041e-01 -2.12969452e-01 -7.69122839e-01
-5.42955995e-01 1.24867819e-01 1.67377204e-01 -1.53905362e-01
-7.31019914e-01 8.23420227e-01 5.58016300e-01 1.42008841e-01
2.41013050e-01 -4.92322266e-01 -5.68783879e-01 -5.86634517e-01
-1.53195709e-01 3.82517457e-01 1.07073158e-01 -5.00018179e-01
1.54042050e-01 6.01970077e-01 -1.60063517e+00 -2.80209512e-01
5.88172257e-01 7.85293102e-01 -2.68169582e-01 2.05021709e-01
1.74309596e-01 2.48674169e-01 -8.16450715e-01 -2.73861051e-01
2.79066443e-01 8.02487805e-02 -7.11147487e-01 -1.66575462e-01
5.24237573e-01 -5.13395905e-01 -2.64145106e-01 9.77320492e-01
1.12368166e+00 2.79974900e-02 6.17764711e-01 -1.49849284e+00
-2.45938003e-01 4.82586354e-01 -5.64621508e-01 1.21455947e-02
8.48424375e-01 -3.40576589e-01 6.19689167e-01 -9.20415938e-01
5.32926023e-01 1.01292372e+00 1.13533282e+00 4.29507881e-01
-4.67212290e-01 -8.05231750e-01 8.29416141e-02 6.20500922e-01
-1.84663820e+00 -4.18625236e-01 8.40052903e-01 -2.73585707e-01
9.29355979e-01 3.92278135e-01 9.95910525e-01 8.67973268e-01
1.98257148e-01 7.56962240e-01 7.69550204e-01 -5.13231635e-01
4.12422307e-02 -6.51687458e-02 3.78103644e-01 5.56790292e-01
5.44124067e-01 -2.81803966e-01 -5.87223887e-01 -9.07989889e-02
4.33315217e-01 -3.09180886e-01 -6.64712369e-01 -4.99492660e-02
-1.03199053e+00 8.03656101e-01 4.01275307e-01 1.24164805e-01
-2.84178972e-01 -3.44582915e-01 5.10364592e-01 -1.19975343e-01
3.35264862e-01 2.03226909e-01 -5.58925092e-01 -1.24018304e-01
-8.66053641e-01 3.98651361e-01 9.24966455e-01 9.83439207e-01
3.47713023e-01 2.76448309e-01 7.87599683e-02 9.70683336e-01
6.81578457e-01 3.46059382e-01 9.60595906e-01 -5.30615747e-01
4.68112767e-01 4.34539735e-01 -2.28661904e-03 -1.08856285e+00
-1.10550034e+00 -2.96743274e-01 -7.60565639e-01 -6.40138239e-02
4.34896857e-01 -2.75239736e-01 -8.74575436e-01 1.74692237e+00
6.16453648e-01 -1.90729961e-01 -3.00655663e-01 1.15968180e+00
1.06623530e+00 3.85536313e-01 -9.13212895e-02 -4.53695208e-01
1.42697537e+00 -7.48283863e-01 -1.29536319e+00 -1.25979379e-01
9.52988863e-01 -7.82808483e-01 6.75416827e-01 4.41277683e-01
-8.64815950e-01 -4.16981667e-01 -1.35236657e+00 2.10317131e-02
-4.88031775e-01 4.27843660e-01 4.57932681e-01 9.66284811e-01
-9.89984632e-01 2.93348044e-01 -7.65325725e-01 -4.26180869e-01
-2.02900559e-01 7.10152090e-01 -6.40942693e-01 5.75516403e-01
-1.42419434e+00 1.27261853e+00 4.24678802e-01 5.68306386e-01
1.28464714e-01 -1.58365086e-01 -1.03971219e+00 -2.65164316e-01
-1.26664117e-01 -2.20426410e-01 1.33695602e+00 -3.99589151e-01
-1.70415187e+00 5.69845915e-01 -1.21617101e-01 -7.31264502e-02
5.57670236e-01 -6.45638824e-01 -5.01309812e-01 -5.32235384e-01
-2.62566675e-02 6.45768523e-01 3.26532185e-01 -1.13489985e+00
-5.61527073e-01 -6.70484900e-01 -4.00924355e-01 3.60792279e-01
-4.49105561e-01 -8.08017328e-02 -6.67416334e-01 -5.53705752e-01
7.32311368e-01 -1.24272180e+00 1.23516418e-01 -1.58858448e-01
-5.24802029e-01 -5.70671618e-01 6.02715254e-01 -9.78131652e-01
1.53332341e+00 -1.59382284e+00 -7.60253444e-02 1.99751556e-01
4.97519076e-02 3.53938162e-01 3.04402322e-01 1.34191941e-02
-3.79531860e-01 -2.69038618e-01 2.14375988e-01 -4.02838796e-01
-9.52187404e-02 2.60233402e-01 7.40466714e-02 9.28589761e-01
-2.65508920e-01 3.36766005e-01 -3.86232823e-01 -5.84056675e-01
8.74969214e-02 6.65050447e-01 -6.71865284e-01 2.87339866e-01
5.11656821e-01 2.26886183e-01 -1.48891896e-01 5.34974694e-01
1.05288386e+00 3.91193002e-01 3.06441963e-01 -9.22643185e-01
-9.97052267e-02 1.73794746e-01 -1.72445989e+00 1.26392889e+00
-6.88298717e-02 7.04848588e-01 -4.03313816e-01 -7.98710525e-01
8.95753205e-01 4.36252177e-01 5.24284601e-01 -3.57169420e-01
5.07057667e-01 1.64096445e-01 5.01719601e-02 -7.47866392e-01
6.66464984e-01 1.78076282e-01 1.45199805e-01 1.43333137e-01
6.28465191e-02 1.41353384e-01 -1.23543158e-01 -4.16281611e-01
3.72827739e-01 2.77015150e-01 5.71332872e-01 -1.69427484e-01
6.57816947e-01 -5.56606174e-01 9.42919254e-01 2.02355444e-01
-4.42973524e-01 8.16754043e-01 2.03881040e-01 -7.03073144e-01
-9.98721659e-01 -5.52987635e-01 -4.56226259e-01 1.15345073e+00
-2.18502149e-01 -5.10503829e-01 -9.04531419e-01 -3.45292449e-01
-1.37308985e-01 5.93933344e-01 -5.83817363e-01 4.05694768e-02
-1.00494850e+00 -1.07896662e+00 8.05854559e-01 9.78175044e-01
3.92580211e-01 -6.33994937e-01 -5.60642183e-01 -8.71808454e-03
-4.25268382e-01 -1.19812059e+00 -4.32982266e-01 3.36901844e-02
-6.07215762e-01 -9.05851722e-01 -8.15025568e-01 -8.64276886e-01
7.17062533e-01 -4.05070782e-02 8.05804491e-01 6.38013473e-03
1.66817233e-01 -6.93667531e-02 -3.60276967e-01 -7.08987772e-01
2.54525691e-01 2.38370910e-01 6.33323014e-01 -3.36138368e-01
5.32030165e-01 -3.40904742e-01 -4.77051467e-01 3.50707829e-01
-3.37403476e-01 -1.67086139e-01 3.56101811e-01 7.10395634e-01
2.09389046e-01 -5.40919960e-01 3.99933606e-01 -5.02291620e-01
8.10787857e-01 -2.82321334e-01 -4.42885160e-01 1.25382408e-01
-7.37021029e-01 2.95248955e-01 2.43663877e-01 -4.89614904e-01
-7.68613875e-01 1.97574183e-01 -6.60446823e-01 -8.00688285e-03
-2.02687323e-01 5.80370486e-01 -2.73965150e-01 -7.17418641e-02
4.43492919e-01 9.53966975e-02 -3.18328571e-03 -5.30823112e-01
5.58148548e-02 9.97279704e-01 6.04265571e-01 -3.26101959e-01
4.48503554e-01 2.56378739e-03 4.53297757e-02 -9.27300990e-01
-5.24519563e-01 -5.74962258e-01 -9.22147572e-01 -2.12821826e-01
7.65426159e-01 -9.66074049e-01 -1.05566359e+00 7.34540701e-01
-1.39461374e+00 4.36853141e-01 4.29912537e-01 9.73999083e-01
-3.23199511e-01 6.53353572e-01 -4.63371396e-01 -8.21068466e-01
-7.09168196e-01 -1.44352198e+00 9.87950206e-01 2.22940907e-01
-4.23734337e-01 -5.04176080e-01 1.58252358e-01 9.42921937e-02
2.48352587e-01 1.98578969e-01 4.12877172e-01 -7.89371490e-01
1.64106861e-01 -6.53095961e-01 6.66976571e-02 3.77053730e-02
-7.51576945e-02 -2.47345702e-03 -9.38043714e-01 -4.24456090e-01
9.91013274e-02 -2.11213559e-01 1.56719804e-01 3.68969828e-01
1.05003083e+00 -4.41858023e-01 -2.73944378e-01 7.48653829e-01
1.07520652e+00 2.89429218e-01 7.22272873e-01 7.49124706e-01
8.06015015e-01 4.78807330e-01 4.15737033e-01 6.64615512e-01
8.56692195e-01 1.21550882e+00 2.26449698e-01 1.74321041e-01
2.14295059e-01 -5.38480431e-02 2.63092935e-01 1.38315582e+00
-6.73556864e-01 -1.03638001e-01 -9.69790637e-01 2.60883659e-01
-1.69236410e+00 -6.84652269e-01 -2.07598001e-01 2.39496660e+00
9.61691916e-01 8.49886611e-02 2.85941988e-01 6.91681027e-01
6.15710437e-01 1.69512499e-02 -1.42996788e-01 -3.07859451e-01
1.99359328e-01 -2.89743654e-02 6.98256552e-01 4.89861071e-01
-1.15767336e+00 6.73894346e-01 6.53348398e+00 3.02691638e-01
-1.49161422e+00 6.18291050e-02 1.94074344e-02 1.76746652e-01
3.59461755e-01 -5.54899454e-01 -1.28804767e+00 3.53952736e-01
7.51950681e-01 1.85739562e-01 -1.07050627e-01 1.15289092e+00
5.88479377e-02 6.81281313e-02 -9.19986606e-01 1.24192321e+00
6.13869607e-01 -5.60930669e-01 -8.59972686e-02 -9.01934505e-02
2.49899298e-01 6.60616010e-02 -1.92417338e-01 3.75400752e-01
-4.24201727e-01 -9.92460012e-01 7.87080646e-01 3.50255579e-01
7.40362763e-01 -6.40179038e-01 1.25156355e+00 2.50944704e-01
-1.33346820e+00 2.36738220e-01 -3.85298043e-01 -7.71783739e-02
1.78565994e-01 -3.41468789e-02 -1.14589715e+00 4.03913468e-01
6.78284109e-01 2.44616315e-01 -6.43593907e-01 1.31332767e+00
-2.91985929e-01 3.98630470e-01 -8.04543853e-01 -2.26080224e-01
-1.76011235e-01 1.40480399e-01 4.81789917e-01 1.14752841e+00
3.98424804e-01 -4.22236249e-02 4.83842101e-03 9.22343228e-03
2.54687816e-01 4.40613925e-01 -4.74940807e-01 6.99935734e-01
6.06245935e-01 9.55602646e-01 -3.45402271e-01 -1.13071978e-01
-4.45020497e-01 6.13101184e-01 2.76452363e-01 1.26585335e-01
-9.36461210e-01 -5.08895338e-01 4.63303626e-01 6.71702698e-02
-1.59564108e-01 -4.27592903e-01 -3.28875631e-01 -1.18241990e+00
1.33247048e-01 -7.95456052e-01 1.92602143e-01 -7.41955161e-01
-7.97507465e-01 8.14423501e-01 3.58348906e-01 -1.42922771e+00
-5.50942361e-01 -8.35235357e-01 -5.04470944e-01 5.82909822e-01
-1.07859707e+00 -1.17026794e+00 -6.44469440e-01 4.08646315e-01
4.97731179e-01 -1.80119932e-01 9.82576549e-01 6.79692805e-01
-8.57713759e-01 1.25673962e+00 -2.33589113e-01 4.81103450e-01
9.68269944e-01 -1.09744167e+00 2.15101302e-01 5.20625174e-01
-1.30464301e-01 8.10255766e-01 1.15100312e+00 -5.02473116e-01
-1.33657122e+00 -6.53906465e-01 1.17417181e+00 -3.44112128e-01
-1.53916506e-02 4.86364849e-02 -8.22898209e-01 1.02328575e+00
1.61394943e-02 9.28287487e-03 7.41995275e-01 3.33545744e-01
-1.71558276e-01 -5.31550571e-02 -1.03410292e+00 4.32902426e-01
8.39373887e-01 -1.13038637e-01 -8.53812277e-01 3.51165891e-01
3.96630973e-01 -8.68779302e-01 -1.00192690e+00 7.64040351e-01
1.16131115e+00 -7.15629518e-01 9.45386887e-01 -4.46725279e-01
-1.60937414e-01 -3.22667658e-01 -2.39710048e-01 -1.29270673e+00
-1.55482754e-01 -1.04840614e-01 -3.90904732e-02 9.69793379e-01
2.53963113e-01 -7.33877242e-01 8.46945822e-01 7.35131204e-01
-9.54926759e-02 -9.22698736e-01 -1.18517911e+00 -4.31085050e-01
-2.23888699e-02 -4.34481055e-01 8.79999101e-01 9.33773279e-01
1.88705474e-01 3.94256234e-01 -6.09408796e-01 1.93714306e-01
5.02597749e-01 -6.77508831e-01 8.97719145e-01 -1.27050412e+00
3.61209273e-01 -2.04104990e-01 -9.45875704e-01 -9.38546896e-01
2.26399958e-01 -3.84632051e-01 8.19210559e-02 -1.24008358e+00
-6.25225976e-02 -5.88015951e-02 -5.18157557e-02 4.86861736e-01
-1.88241065e-01 1.97577968e-01 1.66406244e-01 1.74439445e-01
-1.67115942e-01 6.06211066e-01 8.61586690e-01 1.12277910e-01
-3.43179762e-01 1.76642120e-01 -1.04351535e-01 1.23362803e+00
6.79653764e-01 -2.73457736e-01 -1.15066640e-01 -4.46597129e-01
1.91314846e-01 -3.06707472e-02 -2.10754871e-01 -1.34742129e+00
3.96838933e-01 1.06768049e-01 8.14247787e-01 -1.02612519e+00
5.65753043e-01 -8.01575124e-01 -1.69903152e-02 4.53801364e-01
3.52660790e-02 5.95556855e-01 3.82837206e-02 1.21986963e-01
-1.28537640e-01 -2.25544274e-01 6.47443056e-01 2.39547923e-01
-7.01606214e-01 8.05938989e-02 -3.03468972e-01 -2.99627244e-01
7.23556399e-01 -4.28488225e-01 -2.19602123e-01 -6.16508603e-01
-7.44773448e-01 2.20645256e-02 8.53530020e-02 4.11613375e-01
5.45529902e-01 -1.61603463e+00 -5.64289510e-01 4.92102444e-01
1.63585052e-01 -8.10827166e-02 -9.22471136e-02 8.64732563e-01
-8.65083158e-01 5.84676623e-01 -4.61692989e-01 -5.34431219e-01
-1.44279742e+00 -7.13809431e-02 4.30614352e-01 -1.89320184e-02
-3.18505049e-01 8.38174701e-01 -3.97279680e-01 -7.56494880e-01
6.39622569e-01 -3.45945150e-01 -1.03990948e+00 4.33923781e-01
3.86191249e-01 4.04166818e-01 3.30392569e-01 -1.33294022e+00
-5.40451527e-01 7.50026226e-01 -7.58136585e-02 -9.16086063e-02
1.30120373e+00 -1.00438260e-01 2.43592598e-02 1.99646816e-01
1.36354899e+00 -1.47341881e-02 -4.70659792e-01 -9.06189084e-02
-2.36346591e-02 -2.52282590e-01 7.06100166e-02 -3.67380857e-01
-1.00855613e+00 7.23384917e-01 1.25806880e+00 -3.10223013e-01
9.06231582e-01 -5.53993702e-01 8.57810199e-01 5.91273844e-01
3.99284035e-01 -1.42223096e+00 -3.16479504e-01 7.91835904e-01
1.03340769e+00 -1.27874303e+00 3.68407160e-01 -4.77261573e-01
-6.41364694e-01 1.36657536e+00 9.62875783e-01 1.62389949e-01
9.41701949e-01 2.20414966e-01 3.18635553e-01 2.07034498e-01
-2.22967207e-01 -1.27648994e-01 5.90948820e-01 5.56803942e-01
1.04610717e+00 1.71021432e-01 -9.92329836e-01 7.19436407e-01
-8.74423802e-01 1.16371483e-01 4.95855242e-01 1.07229781e+00
-5.09431839e-01 -1.00614274e+00 -7.57530034e-01 2.38362804e-01
-5.23373127e-01 2.47371837e-01 1.01305909e-01 9.62062597e-01
2.39829734e-01 6.26955867e-01 -5.61864339e-02 -7.93378830e-01
5.61728656e-01 2.36550584e-01 5.21652341e-01 -1.85918242e-01
-1.95871890e-01 5.68902120e-02 1.18200727e-01 -2.98397601e-01
-3.71506065e-01 -4.25381571e-01 -1.14123714e+00 -2.33826265e-01
-6.92653775e-01 1.58445805e-01 9.64440525e-01 1.05920792e+00
-9.42769870e-02 1.53205648e-01 3.96490961e-01 -1.16219234e+00
-7.81821430e-01 -1.50133133e+00 -4.60544795e-01 5.36552072e-01
2.00415015e-01 -1.01163745e+00 -1.71843827e-01 -7.98367038e-02] | [13.679656982421875, 0.26634207367897034] |
8ce76bfa-2811-450c-b3d6-00c28b52ed84 | learning-efficient-point-cloud-generation-for | 1706.07036 | null | http://arxiv.org/abs/1706.07036v1 | http://arxiv.org/pdf/1706.07036v1.pdf | Learning Efficient Point Cloud Generation for Dense 3D Object Reconstruction | Conventional methods of 3D object generative modeling learn volumetric
predictions using deep networks with 3D convolutional operations, which are
direct analogies to classical 2D ones. However, these methods are
computationally wasteful in attempt to predict 3D shapes, where information is
rich only on the surfaces. In this paper, we propose a novel 3D generative
modeling framework to efficiently generate object shapes in the form of dense
point clouds. We use 2D convolutional operations to predict the 3D structure
from multiple viewpoints and jointly apply geometric reasoning with 2D
projection optimization. We introduce the pseudo-renderer, a differentiable
module to approximate the true rendering operation, to synthesize novel depth
maps for optimization. Experimental results for single-image 3D object
reconstruction tasks show that we outperforms state-of-the-art methods in terms
of shape similarity and prediction density. | ['Chen Kong', 'Chen-Hsuan Lin', 'Simon Lucey'] | 2017-06-21 | null | null | null | null | ['3d-object-reconstruction', 'point-cloud-generation'] | ['computer-vision', 'computer-vision'] | [ 2.54245792e-02 4.69510317e-01 4.21380550e-01 -4.75148380e-01
-4.48524386e-01 -4.07017648e-01 9.85304713e-01 -3.32585961e-01
4.87605155e-01 3.55453551e-01 3.41640487e-02 -2.38390148e-01
2.44098321e-01 -1.46063221e+00 -1.25635946e+00 -3.45771402e-01
3.00520658e-01 1.27544045e+00 -4.15196903e-02 -5.51014394e-02
2.65034854e-01 1.30421925e+00 -1.58339369e+00 6.29758835e-01
6.50433660e-01 1.24662995e+00 2.41017371e-01 5.59455037e-01
-7.15880513e-01 5.90541482e-01 -1.64265662e-01 -5.94675720e-01
5.64547479e-01 -3.23566124e-02 -5.80753028e-01 4.58050311e-01
6.10801876e-01 -8.42493057e-01 -3.54340732e-01 6.16894484e-01
3.43473166e-01 -7.55490363e-02 1.04201758e+00 -1.07524550e+00
-1.12458444e+00 7.10622594e-02 -2.08827421e-01 -6.16356432e-01
3.65327418e-01 1.75356686e-01 6.31272614e-01 -1.54868937e+00
8.06480765e-01 1.60802352e+00 8.26755345e-01 6.54280305e-01
-1.43652546e+00 -3.43574166e-01 1.88648701e-04 -3.53769422e-01
-1.21754050e+00 -1.30975217e-01 1.14046443e+00 -6.48380816e-01
1.34195459e+00 1.10777698e-01 1.02569342e+00 9.08090472e-01
2.16133118e-01 7.37024605e-01 8.52674305e-01 -8.92782137e-02
1.98444590e-01 6.28610253e-02 -5.45900047e-01 9.03964877e-01
-5.82274422e-02 2.97286719e-01 -5.06562889e-01 -1.96897641e-01
1.52094984e+00 2.69432336e-01 -8.48859698e-02 -8.73125732e-01
-9.25620675e-01 8.28751981e-01 6.52284980e-01 -3.85285616e-01
-6.07195020e-01 5.11344433e-01 -2.11518735e-01 -1.03891790e-02
1.08096242e+00 3.94227445e-01 -3.39003891e-01 1.50365010e-01
-6.28577828e-01 7.80751705e-01 8.35347295e-01 1.45192850e+00
9.61722255e-01 4.28868353e-01 -9.25018042e-02 5.05738556e-01
6.18122756e-01 5.82564771e-01 -1.93128183e-01 -1.27252162e+00
2.43743807e-01 8.96346688e-01 2.56226212e-01 -6.86715066e-01
-4.29902785e-02 -3.65477055e-01 -7.97584891e-01 7.61695325e-01
-2.02524960e-01 3.49247873e-01 -1.14121413e+00 1.00097346e+00
5.71652532e-01 2.47249961e-01 -8.32996219e-02 8.25448334e-01
1.30920541e+00 6.88111186e-01 -3.14046353e-01 2.64945745e-01
6.24537706e-01 -6.81667686e-01 -1.22211695e-01 8.07778793e-04
2.07311675e-01 -7.27327228e-01 9.93305564e-01 1.13386810e-01
-1.71090734e+00 -6.17528081e-01 -7.00818956e-01 -5.29030323e-01
-9.31965038e-02 -2.14807808e-01 7.88417459e-01 3.25735927e-01
-1.28193796e+00 9.42935944e-01 -9.38990712e-01 2.32386544e-01
1.07638514e+00 2.11406663e-01 -1.64575100e-01 -1.02399722e-01
-3.97825867e-01 8.70393276e-01 1.56197578e-01 -1.48315206e-01
-1.30421126e+00 -1.41997933e+00 -8.55737388e-01 2.07895681e-01
-3.58425140e-01 -1.42377532e+00 1.30224323e+00 -4.04649556e-01
-1.65805233e+00 1.29845369e+00 -1.38678744e-01 -2.42107362e-01
6.98340714e-01 -1.97895959e-01 4.60072577e-01 -1.13464743e-01
-2.19744369e-01 9.40705717e-01 7.32442498e-01 -1.86230469e+00
-4.06154506e-02 -4.67561722e-01 1.48994684e-01 2.79915959e-01
4.08535242e-01 -6.77346289e-01 -9.58858356e-02 -4.98867840e-01
6.47455215e-01 -4.67381656e-01 -4.24687326e-01 5.98089576e-01
-5.04186511e-01 -1.15938880e-01 8.41794491e-01 -4.73305255e-01
-2.75007058e-02 -1.74824810e+00 2.26769447e-01 7.90645927e-02
5.08175373e-01 -7.98989385e-02 7.84393959e-03 3.24292600e-01
2.39941962e-02 2.33371511e-01 -8.81608054e-02 -1.14080501e+00
3.57312471e-01 2.84397691e-01 -8.98919702e-01 2.89865345e-01
5.69492817e-01 1.35817969e+00 -6.87766910e-01 -2.17650816e-01
5.34952462e-01 9.44274008e-01 -9.32000220e-01 6.36559725e-01
-8.16401601e-01 6.89671695e-01 -4.87451553e-01 6.76412463e-01
1.00426173e+00 -4.46045876e-01 -3.83335829e-01 -2.42293701e-01
-7.81839788e-02 6.21240914e-01 -6.38768435e-01 2.02530408e+00
-5.98693430e-01 3.93647969e-01 -3.41548353e-01 -8.18862081e-01
1.34394968e+00 3.02645057e-01 4.98857051e-01 -2.64126390e-01
1.26489773e-01 1.62323833e-01 -7.79363751e-01 -2.08477527e-02
5.26200414e-01 -5.07239461e-01 3.78715068e-01 3.60063940e-01
2.05579594e-01 -1.33213019e+00 -6.72333062e-01 -1.33674331e-02
6.23882055e-01 9.19553101e-01 -1.12898350e-02 2.77851112e-02
-6.60113692e-02 8.32983181e-02 6.54363409e-02 4.50686961e-01
7.38427281e-01 9.81531978e-01 3.77150536e-01 -9.53198373e-01
-1.64240265e+00 -1.65918767e+00 -6.98354021e-02 1.92490801e-01
1.85483173e-01 -1.79644153e-01 -4.49503899e-01 -1.88231915e-01
2.23530948e-01 1.04250515e+00 -5.79729438e-01 -2.15247609e-02
-6.98737979e-01 -2.82859683e-01 7.54383430e-02 6.26761675e-01
3.12379450e-01 -1.12322998e+00 -5.07653177e-01 1.74285620e-01
2.47190073e-01 -1.06066728e+00 1.23295225e-01 -8.48370045e-02
-1.45388341e+00 -6.89464867e-01 -6.11416578e-01 -6.48349583e-01
8.17820132e-01 2.31604651e-01 1.63090038e+00 3.07266489e-02
-1.64331362e-01 4.33735162e-01 3.73114608e-02 -7.15741396e-01
-7.57230699e-01 -2.90423721e-01 -4.02551740e-01 -2.83357769e-01
5.80607466e-02 -1.24487746e+00 -5.30264199e-01 -7.11639002e-02
-6.23405457e-01 7.88275778e-01 3.15920204e-01 5.20995259e-01
1.32088399e+00 -4.56241697e-01 9.54606906e-02 -7.85104871e-01
2.07165018e-01 -3.40459108e-01 -8.23822200e-01 2.93307216e-03
-2.73305774e-01 1.46707147e-01 5.25787711e-01 -3.38546485e-01
-1.22705519e+00 2.66277939e-01 -3.95722896e-01 -1.26393926e+00
-3.46716762e-01 -6.34512305e-02 -1.08541645e-01 -2.05545351e-01
5.71632206e-01 5.26830316e-01 -8.98132324e-02 -7.38235950e-01
4.53940302e-01 -3.94115373e-02 3.24395329e-01 -6.39641702e-01
1.02098680e+00 8.72160256e-01 4.84143764e-01 -5.62731147e-01
-7.58381486e-01 2.57889658e-01 -7.88180768e-01 -2.28490427e-01
8.22350621e-01 -1.01222098e+00 -7.17016935e-01 3.85172695e-01
-1.78912354e+00 -5.63597679e-01 -8.54095638e-01 1.72078446e-01
-1.34649754e+00 -7.95050487e-02 -5.18111825e-01 -7.44673729e-01
-4.65087652e-01 -1.08766270e+00 1.85293221e+00 1.36495400e-02
6.42848611e-02 -8.82056832e-01 -1.60193279e-01 1.20471656e-01
3.54066908e-01 6.47959054e-01 1.13957620e+00 4.16654833e-02
-1.56110406e+00 -9.09260754e-03 -2.96178371e-01 1.53819486e-01
-1.02032982e-01 -1.22994401e-01 -1.25123048e+00 1.68744087e-01
1.61216855e-01 -2.02897623e-01 5.08318961e-01 5.68466842e-01
1.76353049e+00 -2.58316845e-01 -3.27879131e-01 1.30600727e+00
1.50490272e+00 -2.12178268e-02 7.80883431e-01 -3.07762086e-01
9.92283702e-01 3.76672208e-01 7.23309666e-02 7.96257317e-01
5.06676972e-01 5.12149513e-01 1.03765297e+00 -8.21822360e-02
-4.95936900e-01 -8.28006923e-01 -1.37429237e-01 5.49179733e-01
-5.08923829e-01 -2.04517677e-01 -8.16324413e-01 1.96572945e-01
-1.42193639e+00 -7.53992438e-01 -2.54352301e-01 1.98563313e+00
4.73160595e-01 -2.48260945e-02 -3.25101644e-01 -3.54036450e-01
1.59632489e-01 -3.48206088e-02 -6.60825253e-01 -3.55953544e-01
-3.44176069e-02 6.67126894e-01 4.68596220e-02 5.31187654e-01
-5.53802311e-01 1.05373073e+00 6.33471298e+00 4.98042166e-01
-9.71549273e-01 1.33597150e-01 6.89944685e-01 -2.32001469e-01
-1.13956928e+00 -1.23579755e-01 -8.37062120e-01 -9.53746438e-02
4.15720165e-01 -2.86175720e-02 5.70927203e-01 1.08312011e+00
-5.59364632e-02 5.44689536e-01 -1.50576401e+00 1.23171782e+00
4.26134802e-02 -2.10598087e+00 6.75350547e-01 1.25167266e-01
1.02765071e+00 1.42522424e-01 9.72581729e-02 5.51087782e-02
5.57717741e-01 -1.23528123e+00 1.13229394e+00 1.01997411e+00
8.87627184e-01 -5.89218080e-01 5.80863319e-02 6.48046017e-01
-9.57175314e-01 6.00581050e-01 -7.00626373e-01 1.57617163e-02
5.54819763e-01 6.22919619e-01 -1.12485969e+00 3.75410616e-01
6.95968151e-01 6.93363488e-01 -4.34421701e-03 9.07305062e-01
-2.72185177e-01 -5.37897535e-02 -3.91573548e-01 2.59493031e-02
5.82704358e-02 -2.49286383e-01 7.04233289e-01 6.11267209e-01
7.93785751e-01 2.40408197e-01 -7.70694166e-02 1.91255438e+00
-3.68462026e-01 -1.06310755e-01 -1.04751396e+00 1.25559866e-01
2.40029812e-01 8.14067841e-01 -4.83900219e-01 -3.55248868e-01
-1.57331556e-01 9.96769130e-01 5.31645060e-01 3.22408766e-01
-8.08106482e-01 3.08624208e-01 7.99723148e-01 5.37558079e-01
4.25713122e-01 -5.63550830e-01 -7.90772021e-01 -1.03227711e+00
1.41878892e-02 -3.86802591e-02 -5.03055692e-01 -1.52698076e+00
-1.48842335e+00 5.75103581e-01 3.86409536e-02 -1.31264055e+00
-3.13609809e-01 -8.15164268e-01 -5.55225492e-01 1.24331379e+00
-1.50221562e+00 -1.47296083e+00 -5.38882315e-01 4.49105412e-01
6.58767581e-01 5.16936556e-02 9.83685076e-01 -1.25517726e-01
5.75374246e-01 1.03336468e-01 -3.68520707e-01 -3.29113811e-01
-1.12854846e-01 -1.11105549e+00 9.84475255e-01 3.31199199e-01
2.07827181e-01 2.60269552e-01 4.14990604e-01 -7.08456874e-01
-1.65436625e+00 -1.24560833e+00 6.32388771e-01 -7.11877644e-01
-3.97849493e-02 -5.46798110e-01 -8.87742817e-01 8.75941575e-01
-2.17301399e-01 2.04731852e-01 3.78477514e-01 -3.50118369e-01
-2.86210299e-01 3.99860531e-01 -1.32513869e+00 6.28202140e-01
1.66827559e+00 -4.89490777e-01 -3.61135989e-01 6.64611101e-01
1.20761549e+00 -1.04326653e+00 -1.07106340e+00 4.71621275e-01
4.77156132e-01 -1.13560593e+00 1.52371669e+00 -7.31762946e-01
1.00554633e+00 -1.20181695e-01 -5.56950331e-01 -1.30924571e+00
-3.49695981e-01 -3.73161733e-01 -5.20697534e-01 6.57783985e-01
1.30646631e-01 -4.09358025e-01 1.14106524e+00 6.75060272e-01
-9.06018376e-01 -1.06770527e+00 -7.45936871e-01 -5.45002818e-01
2.78578430e-01 -6.05574012e-01 1.26570165e+00 6.76335454e-01
-9.08709466e-01 1.75220578e-03 -2.07260042e-01 2.92795271e-01
7.92881548e-01 8.57395351e-01 1.08715463e+00 -1.42946303e+00
-3.19854826e-01 -4.70675141e-01 -4.57337618e-01 -1.49716675e+00
2.43205115e-01 -1.03895831e+00 6.55345526e-03 -1.91907859e+00
-2.17122123e-01 -6.66425705e-01 5.61217070e-01 6.59118891e-02
3.30998003e-01 3.22529525e-01 3.30671817e-02 1.85832053e-01
1.38992652e-01 1.12439764e+00 1.86122060e+00 -1.02933273e-01
-2.26924121e-01 8.15327540e-02 -4.08529520e-01 8.09708357e-01
6.60293519e-01 -2.16346994e-01 -4.86099958e-01 -9.87691522e-01
3.30588192e-01 2.47870624e-01 8.14495921e-01 -7.28030503e-01
-2.08571270e-01 -3.49788487e-01 8.54688764e-01 -1.36897755e+00
1.08073676e+00 -7.69835711e-01 7.22569942e-01 1.54129580e-01
-1.11109450e-01 -1.27874687e-01 1.41046882e-01 3.59600931e-01
1.56091049e-01 4.59214626e-03 6.79749846e-01 -5.20629883e-01
-3.81956518e-01 1.10624027e+00 3.13538879e-01 -2.98705757e-01
7.64021635e-01 -4.32075769e-01 -2.37463713e-02 -3.02885592e-01
-8.47692192e-01 -3.24871331e-01 8.65717769e-01 2.29448631e-01
1.27848411e+00 -1.72666001e+00 -7.90545881e-01 4.80695367e-01
-1.66499570e-01 8.85206759e-01 3.51047814e-01 7.24407881e-02
-9.40924406e-01 2.49739990e-01 -2.63851792e-01 -8.78426611e-01
-6.18157864e-01 4.49650943e-01 6.26916409e-01 -5.63158877e-02
-1.07870543e+00 1.09229565e+00 8.09209347e-01 -8.08599174e-01
-1.47565141e-01 -5.72934031e-01 3.92390251e-01 -5.78178167e-01
2.26065412e-01 4.17341292e-02 1.14513919e-01 -4.33585018e-01
1.55509174e-01 6.09868050e-01 4.86902773e-01 1.37758613e-01
1.76713884e+00 2.76814610e-01 -9.23986882e-02 3.72500509e-01
1.16621149e+00 -5.20542800e-01 -1.63679266e+00 -2.03523189e-01
-7.84068525e-01 -9.00666893e-01 5.20267822e-02 -4.56207901e-01
-1.10702026e+00 1.35276628e+00 1.97453454e-01 -2.07310040e-02
7.12475896e-01 3.96932870e-01 6.76846087e-01 2.70489186e-01
6.64858222e-01 -3.05124700e-01 1.28082454e-01 6.19738698e-01
1.52897823e+00 -9.44445014e-01 -4.16172259e-02 -8.42672527e-01
-1.18069023e-01 1.21262074e+00 7.80581176e-01 -7.27087736e-01
9.04939115e-01 2.89914399e-01 -5.18771470e-01 -6.20639801e-01
-7.10051000e-01 3.34150851e-01 4.85452712e-01 8.92452896e-01
2.49599800e-01 2.04630032e-01 5.26810408e-01 2.95355976e-01
-5.78659236e-01 -8.12556967e-02 2.20075160e-01 4.99194741e-01
-2.17535660e-01 -8.30569029e-01 -2.03396231e-01 4.86184984e-01
1.61363602e-01 -1.57229245e-01 -2.46173933e-01 5.39039314e-01
9.87868533e-02 -2.51762327e-02 5.44201136e-01 -2.34559193e-01
4.11136955e-01 6.06861301e-02 1.05694771e+00 -8.27469230e-01
-1.99495763e-01 -9.32990834e-02 -2.44697928e-01 -7.91367054e-01
-2.69952863e-01 -4.34383571e-01 -1.23162055e+00 -4.49762791e-01
5.85998073e-02 -5.58593035e-01 8.32143664e-01 6.55204594e-01
6.45518720e-01 5.31662107e-01 5.11789858e-01 -1.75605595e+00
-3.97227883e-01 -6.90816700e-01 -4.86200243e-01 2.73414165e-01
1.25504285e-01 -9.23301399e-01 -8.84967297e-02 1.74735591e-01] | [8.821029663085938, -3.590196371078491] |
5db5f362-33e0-4c72-9d12-63d6a48b69f0 | investigation-of-multimodal-features | 1809.06225 | null | http://arxiv.org/abs/1809.06225v1 | http://arxiv.org/pdf/1809.06225v1.pdf | Investigation of Multimodal Features, Classifiers and Fusion Methods for Emotion Recognition | Automatic emotion recognition is a challenging task. In this paper, we
present our effort for the audio-video based sub-challenge of the Emotion
Recognition in the Wild (EmotiW) 2018 challenge, which requires participants to
assign a single emotion label to the video clip from the six universal emotions
(Anger, Disgust, Fear, Happiness, Sad and Surprise) and Neutral. The proposed
multimodal emotion recognition system takes audio, video and text information
into account. Except for handcraft features, we also extract bottleneck
features from deep neutral networks (DNNs) via transfer learning. Both temporal
classifiers and non-temporal classifiers are evaluated to obtain the best
unimodal emotion classification result. Then possibilities are extracted and
passed into the Beam Search Fusion (BS-Fusion). We test our method in the
EmotiW 2018 challenge and we gain promising results. Compared with the baseline
system, there is a significant improvement. We achieve 60.34% accuracy on the
testing dataset, which is only 1.5% lower than the winner. It shows that our
method is very competitive. | ['Jian-Hua Tao', 'Ya Li', 'Zheng Lian', 'Jian Huang'] | 2018-09-13 | null | null | null | null | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [ 4.31717150e-02 -2.94073462e-01 1.88748851e-01 -6.15951836e-01
-8.81095171e-01 -4.38225240e-01 2.02184066e-01 -1.19493783e-01
-8.19922984e-01 6.01921499e-01 1.83170125e-01 3.52811247e-01
2.33731061e-01 -6.75486401e-02 -3.87100071e-01 -7.08454609e-01
-2.78571963e-01 3.37993610e-03 -2.34651029e-01 -3.00690889e-01
-9.96258855e-02 1.84295163e-01 -1.82154906e+00 8.30670655e-01
5.38537502e-01 1.82302785e+00 -5.92266440e-01 9.80432749e-01
2.33301580e-01 1.13326085e+00 -5.23027956e-01 -6.87391341e-01
-1.75937995e-01 -2.72522151e-01 -8.21117282e-01 -2.57964194e-01
4.03038949e-01 -3.11182201e-01 -1.08389169e-01 1.02749765e+00
1.01219881e+00 4.03091162e-01 6.41938269e-01 -1.81079388e+00
-1.30487069e-01 3.46437037e-01 -6.52499735e-01 -1.29004732e-01
4.83607858e-01 -1.79721653e-01 8.73506308e-01 -1.23422134e+00
7.09660769e-01 1.09276676e+00 7.86165178e-01 8.01810324e-01
-6.07472181e-01 -9.90180194e-01 1.45337969e-01 8.60198140e-01
-1.26295471e+00 -7.72744238e-01 7.98046112e-01 -1.81798637e-01
1.09846759e+00 4.10877228e-01 7.62625813e-01 1.68634021e+00
2.99509931e-02 1.15203929e+00 1.03304052e+00 -2.90979236e-01
2.76495576e-01 3.19567323e-01 2.37529166e-02 2.60208964e-01
-9.36522007e-01 -2.49488264e-01 -1.10909128e+00 -1.73670650e-01
-2.05803052e-01 -2.26887599e-01 -3.76147032e-01 5.16948663e-02
-8.59091938e-01 6.55066371e-01 2.84802169e-01 2.23894954e-01
-7.03607619e-01 -7.79210404e-02 1.06986010e+00 6.95996940e-01
5.82489431e-01 -3.59877391e-04 -7.47888327e-01 -7.68826962e-01
-7.17081785e-01 5.83910756e-03 8.12546611e-01 4.41968411e-01
3.30880046e-01 9.09361169e-02 -3.15304220e-01 1.06548941e+00
3.16607510e-03 2.85284013e-01 5.70677936e-01 -1.02584779e+00
-8.50855336e-02 2.33460799e-01 1.65004089e-01 -1.00473797e+00
-4.63571221e-01 1.99106000e-02 -8.76931667e-01 2.06694037e-01
1.61327608e-02 -7.57998586e-01 -7.39913821e-01 1.93710887e+00
3.92418683e-01 3.53602916e-01 2.00929523e-01 9.69232500e-01
1.09263623e+00 9.18295205e-01 2.63888568e-01 -3.94480705e-01
1.33702958e+00 -1.16045856e+00 -1.20703614e+00 -2.09677266e-03
5.65556705e-01 -9.31882203e-01 8.79354954e-01 8.92906606e-01
-9.87944424e-01 -3.53357315e-01 -9.72950697e-01 1.68157831e-01
-5.79673469e-01 3.60327572e-01 5.55830538e-01 6.35764837e-01
-1.07184362e+00 4.58877951e-01 -5.73865771e-01 -4.20789123e-01
2.03101069e-01 4.54986006e-01 -7.01400280e-01 1.14839889e-01
-1.52376807e+00 8.91325116e-01 1.37339175e-01 4.57879871e-01
-7.60370493e-01 -5.44157147e-01 -8.89487207e-01 4.60178927e-02
2.54043549e-01 -4.37707268e-02 1.30612767e+00 -1.80070293e+00
-1.96586442e+00 8.72007787e-01 -3.81996661e-01 -2.49372512e-01
1.96102783e-01 -5.47334552e-01 -7.80637383e-01 3.78599495e-01
-5.27718782e-01 9.31601465e-01 1.00381351e+00 -7.97401190e-01
-5.98480105e-01 -2.25086421e-01 -2.74593025e-01 1.99684814e-01
-8.39861155e-01 7.24596083e-01 -3.16446275e-01 -3.68152410e-01
-3.83256733e-01 -1.02466524e+00 2.49964342e-01 -2.27489918e-01
-3.59622240e-01 -5.11687994e-01 9.09628272e-01 -8.17657888e-01
1.17004240e+00 -2.44267750e+00 2.62334287e-01 3.65801901e-02
-5.68916500e-02 8.20993483e-02 -2.82445669e-01 2.15010867e-01
-6.20143354e-01 -7.63260499e-02 2.34103024e-01 -5.88274419e-01
3.63282889e-01 -2.60510772e-01 -2.92705506e-01 3.88444275e-01
4.46503907e-01 8.50033820e-01 -4.44378495e-01 -4.47335780e-01
-1.89916622e-02 7.88000107e-01 -5.63785076e-01 4.00524944e-01
2.82890111e-01 6.28727600e-02 3.91321722e-03 8.55435550e-01
7.51793265e-01 4.83385205e-01 5.75474799e-02 -5.97573161e-01
4.42412831e-02 -7.37022460e-02 -1.09399545e+00 1.56853127e+00
-3.90077204e-01 9.07125235e-01 4.07165587e-01 -1.01346815e+00
8.27169180e-01 8.58706594e-01 5.49166203e-01 -5.51867723e-01
5.95665097e-01 -4.43998128e-02 8.09030980e-03 -7.72629082e-01
4.08544540e-01 -2.77147144e-01 -3.44394714e-01 1.79837421e-01
7.41348386e-01 2.37143874e-01 -1.24267772e-01 2.12638736e-01
1.06887519e+00 2.50339448e-01 -8.08598325e-02 -1.52480155e-01
5.15813291e-01 -5.18387377e-01 6.49211645e-01 4.40522313e-01
-7.56572664e-01 4.77560967e-01 8.37317705e-01 -3.69840175e-01
-4.52662498e-01 -7.71256745e-01 2.64278024e-01 1.53168583e+00
-2.21957698e-01 -7.40884066e-01 -7.17675865e-01 -5.85835874e-01
-5.38384199e-01 5.10067046e-01 -9.80483651e-01 -6.01329803e-01
-5.25438301e-02 -6.56798899e-01 5.52432358e-01 4.76459891e-01
4.23291177e-01 -1.26697910e+00 -6.13081694e-01 -1.89720362e-01
-6.08333528e-01 -1.17440975e+00 -3.27835083e-01 2.93427587e-01
-2.20960796e-01 -5.09247482e-01 -7.65715122e-01 -5.55057704e-01
7.73149878e-02 -3.72537971e-01 9.41858530e-01 -3.77222180e-01
-3.69265705e-01 7.10048676e-01 -5.70548952e-01 -5.87506592e-01
2.10359305e-01 -1.50205687e-01 1.85128748e-01 5.99620581e-01
5.74052453e-01 -4.90028501e-01 -5.80968380e-01 1.38056234e-01
-5.62528729e-01 -3.95312995e-01 1.81855470e-01 9.38414872e-01
1.87016562e-01 -1.76198818e-02 7.00141191e-01 -1.17953621e-01
6.77152872e-01 -4.12627339e-01 7.31656849e-02 2.39874661e-01
7.46302232e-02 -5.38060248e-01 3.42270285e-01 -8.70320976e-01
-1.27861130e+00 3.26336622e-01 -4.51027751e-01 -7.59516537e-01
-3.35415035e-01 4.74870265e-01 -2.44229063e-01 2.07228493e-02
4.05256003e-01 -2.58459836e-01 -2.69394696e-01 -2.92256363e-02
1.85972035e-01 8.46944749e-01 5.44602692e-01 -4.67165262e-01
-1.41771771e-02 2.67595619e-01 -4.85510141e-01 -7.92730808e-01
-9.18405533e-01 -4.09646004e-01 -3.84996861e-01 -8.95334899e-01
9.89266932e-01 -1.05880761e+00 -1.42084825e+00 6.98340476e-01
-1.12470138e+00 -2.31784359e-01 -6.32221177e-02 6.27157450e-01
-3.97948056e-01 2.95633506e-02 -9.33773756e-01 -1.42024028e+00
-5.48134565e-01 -9.02112126e-01 1.05917442e+00 2.34312564e-01
-5.95911503e-01 -4.66773361e-01 -3.18502076e-02 7.28535233e-03
5.47421336e-01 3.85624141e-01 4.02878642e-01 -7.97606289e-01
4.45012778e-01 -4.95066851e-01 -1.01235896e-01 5.66082299e-01
-5.00146508e-01 3.80514026e-01 -1.36060250e+00 -2.79807914e-02
2.11456135e-01 -1.34129310e+00 1.05087614e+00 2.49652445e-01
1.14529562e+00 4.08950299e-02 1.99760228e-01 3.98482502e-01
9.38191950e-01 2.96089888e-01 8.66681278e-01 8.95997137e-02
1.79044649e-01 8.96614134e-01 7.49490976e-01 7.53819287e-01
2.77616203e-01 5.23494065e-01 4.67345655e-01 -5.29563949e-02
5.04255354e-01 2.83922076e-01 9.21792924e-01 7.09894538e-01
-1.17729329e-01 -3.33285928e-01 -7.70229340e-01 4.72708553e-01
-1.92976487e+00 -1.16110909e+00 9.37013254e-02 1.83330870e+00
7.36455560e-01 -5.39903641e-02 2.21069619e-01 2.48382270e-01
6.01099253e-01 -1.22282747e-02 -4.45913047e-01 -9.23707485e-01
-2.70010531e-01 4.79561061e-01 -2.92758346e-01 3.93384427e-01
-1.54848313e+00 8.52652848e-01 5.77059555e+00 1.03933966e+00
-1.49672377e+00 1.21295944e-01 9.28140223e-01 -6.86296701e-01
3.35983276e-01 -6.20474994e-01 -3.46396953e-01 2.28735566e-01
1.43713975e+00 -2.36233743e-03 3.96183580e-01 7.71616399e-01
-7.32023343e-02 -2.68974870e-01 -9.84502017e-01 1.46481490e+00
4.51214284e-01 -5.46100080e-01 -4.06402320e-01 -5.20484388e-01
4.70830858e-01 4.60736640e-03 3.22651833e-01 9.54248548e-01
-1.52140081e-01 -1.12626183e+00 7.87945807e-01 5.77368557e-01
5.28420329e-01 -1.22721839e+00 9.87582088e-01 -5.99193834e-02
-9.36458230e-01 -1.05283365e-01 -9.51188952e-02 -1.24034770e-01
2.18596429e-01 4.92729336e-01 -6.28044978e-02 4.16158676e-01
1.24227500e+00 7.04317033e-01 -1.68161377e-01 7.05126226e-01
-5.50212227e-02 8.30761611e-01 -4.21754479e-01 -2.45622218e-01
1.69820338e-01 1.53277040e-01 3.02785188e-01 1.43020332e+00
5.43492496e-01 1.65025339e-01 -1.40064776e-01 3.39046329e-01
-3.33440363e-01 3.62699538e-01 -2.83750862e-01 -9.32173878e-02
3.99959721e-02 1.85892999e+00 -4.63210553e-01 -4.15188938e-01
-9.62979347e-02 1.58667445e+00 3.77952218e-01 3.38783264e-01
-1.24559987e+00 -7.45414436e-01 7.08175600e-01 -8.71514380e-01
2.19523713e-01 2.79658198e-01 2.48225302e-01 -1.36642146e+00
-1.67370290e-02 -9.55114007e-01 5.52040875e-01 -1.12440133e+00
-1.29935384e+00 1.00449729e+00 -4.87384290e-01 -9.95849073e-01
-2.63909668e-01 -7.57134616e-01 -7.73252428e-01 5.77007949e-01
-1.13956785e+00 -8.30946803e-01 -2.62407452e-01 8.23212504e-01
2.62190938e-01 2.05465574e-02 1.13668406e+00 6.94430113e-01
-7.19366312e-01 8.38055372e-01 -1.93240136e-01 3.12985033e-02
1.37577462e+00 -9.65218663e-01 -4.51754004e-01 3.30190390e-01
-2.23012373e-01 1.89953133e-01 7.47132182e-01 -1.96911633e-01
-1.30621731e+00 -7.90937483e-01 9.43112671e-01 -2.39691451e-01
6.67510271e-01 -6.51070535e-01 -6.59597635e-01 4.21838701e-01
8.27238321e-01 -3.34239006e-02 1.09777904e+00 4.12211388e-01
-5.25984228e-01 -7.40761831e-02 -1.10029268e+00 5.67305446e-01
5.35148084e-01 -5.08663118e-01 -2.45736763e-01 1.90837514e-02
6.40181243e-01 -1.94117829e-01 -9.61399496e-01 6.86998427e-01
9.75280762e-01 -1.24010742e+00 6.93525076e-01 -9.71969604e-01
8.77492368e-01 1.87182263e-01 -3.65030646e-01 -1.42816865e+00
2.98963338e-02 -6.86947405e-01 -5.65416412e-03 1.49831998e+00
3.52022678e-01 -2.52456844e-01 4.90233570e-01 8.16754580e-01
9.85675901e-02 -8.25313628e-01 -1.15821087e+00 -3.99462670e-01
-1.83907434e-01 -8.01088035e-01 4.49968986e-02 1.19312286e+00
6.52348280e-01 6.51686907e-01 -9.76587713e-01 -2.62475908e-01
2.25192279e-01 -5.54793328e-02 3.73386770e-01 -1.02457750e+00
8.65605772e-02 -4.35454547e-01 -3.26971114e-01 -3.65069777e-01
5.83952844e-01 -5.13148725e-01 1.12983696e-01 -8.44224036e-01
2.38266453e-01 5.03239572e-01 -7.79045045e-01 6.56411052e-01
-2.17869747e-02 5.47317445e-01 2.19029874e-01 -6.67833686e-01
-1.23016274e+00 1.05506814e+00 6.04752243e-01 -1.95622310e-01
-2.12665834e-02 -3.64790976e-01 -5.00990868e-01 8.82764757e-01
9.31101859e-01 -2.01954827e-01 3.36091332e-02 -7.55292922e-02
3.72792274e-01 4.91941273e-02 2.19585776e-01 -8.03707421e-01
2.94398457e-01 2.50060707e-01 5.45503855e-01 -6.88432813e-01
1.01634371e+00 -8.86000752e-01 -1.22577369e-01 1.55397311e-01
-4.31555808e-01 2.88085360e-02 5.89508235e-01 1.72748923e-01
-5.69590986e-01 -6.32996410e-02 6.70778334e-01 4.05086309e-01
-6.52765214e-01 9.83406901e-02 -7.11990058e-01 -1.68493912e-01
9.55671787e-01 2.26943001e-01 -8.92197788e-02 -9.21268940e-01
-1.22906494e+00 3.48227471e-01 -1.75944537e-01 4.76773590e-01
7.23186910e-01 -1.73700786e+00 -6.86437666e-01 -2.73458343e-02
1.62391409e-01 -8.21012497e-01 9.90118504e-01 1.31348670e+00
2.73218572e-01 4.43469919e-02 -4.30596620e-01 -4.27131712e-01
-1.63243783e+00 4.72572356e-01 4.61204320e-01 -6.24293759e-02
1.37961462e-01 1.16607094e+00 -1.23230666e-01 -5.11945188e-01
6.56027377e-01 8.28924701e-02 -3.74176443e-01 6.60490811e-01
8.22175503e-01 4.14673030e-01 9.59219262e-02 -7.86693573e-01
-5.28668821e-01 4.04970378e-01 6.54744655e-02 -3.52713704e-01
1.38119614e+00 1.31373666e-02 -3.79018962e-01 7.74230659e-01
1.56067121e+00 -1.83807015e-01 -8.34970593e-01 8.70763808e-02
-2.10205927e-01 -5.24731912e-02 1.65739805e-01 -9.95361328e-01
-1.11146998e+00 1.24583673e+00 9.78720069e-01 9.20814052e-02
1.62828469e+00 -2.96707600e-01 7.31417894e-01 4.82974738e-01
-1.44374713e-01 -1.46340692e+00 1.51036963e-01 8.03818882e-01
1.07083941e+00 -1.21420300e+00 -5.51531732e-01 -3.86181176e-02
-1.11632192e+00 1.14758730e+00 8.63476098e-01 -3.48653272e-02
8.74948084e-01 5.33615470e-01 2.62138784e-01 -1.57600895e-01
-1.45982146e+00 -2.93572787e-02 3.53298217e-01 3.41310859e-01
6.74996972e-01 -1.84794396e-01 -4.05147970e-02 1.28607285e+00
4.82566766e-02 8.20185989e-02 1.78281680e-01 6.29652739e-01
-1.72107235e-01 -6.83836699e-01 -2.08269194e-01 2.79511392e-01
-8.61725986e-01 -5.84302917e-02 -6.12264514e-01 4.56679523e-01
6.02965355e-02 1.10700405e+00 -5.84102869e-02 -8.60577762e-01
4.47232842e-01 6.98046207e-01 3.66733342e-01 1.83303922e-01
-8.22738409e-01 2.73437709e-01 3.92355114e-01 -9.52643514e-01
-5.98192513e-01 -5.73164880e-01 -9.94349301e-01 -1.20162167e-01
-2.30928734e-01 2.89314210e-01 5.97390771e-01 6.35555267e-01
4.37349886e-01 5.26029110e-01 6.87119067e-01 -1.29294264e+00
-1.73405111e-01 -1.11271000e+00 -5.51261604e-01 5.14552474e-01
2.44864717e-01 -6.10598505e-01 -6.23513699e-01 1.02321163e-01] | [13.314289093017578, 5.078708648681641] |
d304438a-7726-4cce-a2bd-f9de5ab68154 | dynamics-based-3d-skeletal-hand-tracking | 1705.07640 | null | http://arxiv.org/abs/1705.07640v1 | http://arxiv.org/pdf/1705.07640v1.pdf | Dynamics Based 3D Skeletal Hand Tracking | Tracking the full skeletal pose of the hands and fingers is a challenging
problem that has a plethora of applications for user interaction. Existing
techniques either require wearable hardware, add restrictions to user pose, or
require significant computation resources. This research explores a new
approach to tracking hands, or any articulated model, by using an augmented
rigid body simulation. This allows us to phrase 3D object tracking as a linear
complementarity problem with a well-defined solution. Based on a depth sensor's
samples, the system generates constraints that limit motion orthogonal to the
rigid body model's surface. These constraints, along with prior motion,
collision/contact constraints, and joint mechanics, are resolved with a
projected Gauss-Seidel solver. Due to camera noise properties and attachment
errors, the numerous surface constraints are impulse capped to avoid
overpowering mechanical constraints. To improve tracking accuracy, multiple
simulations are spawned at each frame and fed a variety of heuristics,
constraints and poses. A 3D error metric selects the best-fit simulation,
helping the system handle challenging hand motions. Such an approach enables
real-time, robust, and accurate 3D skeletal tracking of a user's hand on a
variety of depth cameras, while only utilizing a single x86 CPU core for
processing. | ['Leonid Keselman', 'Sterling Orsten', 'Stan Melax'] | 2017-05-22 | null | null | null | null | ['3d-object-tracking'] | ['computer-vision'] | [ 5.52533902e-02 -4.64247584e-01 -1.20785341e-01 7.97088370e-02
-2.51607388e-01 -8.57102692e-01 5.41286990e-02 -5.87357759e-01
-3.80523205e-01 4.22628790e-01 -1.84889615e-01 -1.78732827e-01
1.80728026e-02 -2.73243397e-01 -2.94080645e-01 -4.83834118e-01
2.42824793e-01 6.86098516e-01 4.48663175e-01 -1.78948477e-01
2.87651122e-01 9.11834896e-01 -1.33749270e+00 -3.68075162e-01
4.16014850e-01 6.71398878e-01 2.86104769e-01 8.51229012e-01
1.18052788e-01 3.38772498e-02 -4.62014019e-01 -1.70870200e-01
6.23538494e-01 -3.06174136e-03 -3.65050286e-01 2.61585981e-01
6.15150273e-01 -7.38407314e-01 -1.51685581e-01 7.90103495e-01
9.75979388e-01 4.23531562e-01 2.06140533e-01 -1.09892917e+00
1.71304002e-01 -2.31243268e-01 -7.11608291e-01 -1.05919622e-01
1.12849092e+00 2.61609107e-01 4.59758192e-01 -8.84120047e-01
9.13179696e-01 1.35921204e+00 9.54662144e-01 8.31857860e-01
-1.25071549e+00 -5.27349412e-01 2.97807038e-01 -6.26864672e-01
-1.52914560e+00 -2.23985702e-01 7.83872008e-01 -5.65519750e-01
9.87992585e-01 5.77909112e-01 1.18392265e+00 1.06636751e+00
3.95590931e-01 4.71081287e-01 7.10326433e-01 -5.15299857e-01
1.60439074e-01 -9.05396342e-02 5.72350323e-02 8.35172892e-01
2.51521200e-01 -3.26601416e-02 -5.26915729e-01 -5.26606679e-01
1.37785685e+00 2.46759832e-01 -2.99447894e-01 -7.65224814e-01
-8.90465319e-01 3.31496656e-01 -4.31182683e-02 -1.66835248e-01
-2.95474231e-01 1.11749865e-01 -3.04004818e-04 -2.97529012e-01
-1.58435497e-02 1.40115365e-01 -5.13881803e-01 -3.27170432e-01
-8.02402377e-01 8.43493342e-01 1.09215081e+00 1.21937191e+00
2.06438288e-01 -8.20017979e-02 1.12967841e-01 2.95762777e-01
7.72026598e-01 6.77825749e-01 -1.02379911e-01 -1.01304531e+00
4.99272525e-01 5.51422656e-01 5.01310527e-01 -9.91535306e-01
-5.02647460e-01 -1.10676169e-01 -3.56683373e-01 6.97012365e-01
6.72604203e-01 -3.61733258e-01 -1.00312829e+00 1.46275246e+00
1.00214517e+00 -1.10927835e-01 -5.21229029e-01 1.39635968e+00
2.83239245e-01 2.01521292e-01 -2.78498679e-01 -2.28370622e-01
1.27126825e+00 -4.10567760e-01 -7.27752090e-01 -1.43506333e-01
6.14626929e-02 -1.19916713e+00 8.16832602e-01 5.57596624e-01
-1.39958787e+00 -4.46271151e-01 -8.63828301e-01 -3.50323021e-02
1.73299551e-01 -2.07854748e-01 5.42988360e-01 7.47386694e-01
-7.95446336e-01 5.11367917e-01 -1.37597203e+00 -4.10419673e-01
-1.25112236e-01 9.85315204e-01 6.98344782e-02 8.37896615e-02
-6.75160587e-01 1.27639139e+00 -1.92432866e-01 2.44672328e-01
-3.08726639e-01 -5.73100269e-01 -5.32192051e-01 -4.39919025e-01
3.41796398e-01 -9.83344913e-01 1.19431520e+00 -2.39456445e-01
-1.86036479e+00 6.99701607e-01 -3.14863563e-01 4.07017082e-01
7.87710488e-01 -4.09589887e-01 1.36027932e-01 -1.18411072e-01
-3.21551949e-01 2.33843982e-01 9.59680736e-01 -1.16206908e+00
-1.87633395e-01 -7.61511445e-01 4.31810655e-02 6.98150933e-01
-1.02385089e-01 1.40130803e-01 -1.00211442e+00 -6.02336466e-01
4.47534382e-01 -1.51865304e+00 -4.07925636e-01 3.79705220e-01
-3.56965005e-01 1.51439831e-01 9.56758380e-01 -8.29894066e-01
1.25650096e+00 -1.95975435e+00 5.47568560e-01 7.42198110e-01
1.20588273e-01 1.07701458e-01 2.96905100e-01 1.72202736e-01
4.35901344e-01 -3.54549825e-01 1.72234178e-01 -4.34259772e-01
-1.89241744e-03 1.89441428e-01 2.17245799e-02 4.84707266e-01
-3.95916671e-01 4.44745898e-01 -5.60177326e-01 -5.24872184e-01
2.95392156e-01 8.37100744e-01 -6.04211390e-01 2.08126470e-01
-1.55795425e-01 5.42558789e-01 -7.39383399e-01 8.58061612e-01
5.94006538e-01 -9.69062187e-03 3.31997126e-01 -2.44426280e-01
-2.69963533e-01 4.78636436e-02 -1.96506512e+00 1.74107444e+00
-1.40394345e-01 -1.35367475e-02 6.85297430e-01 -9.05034691e-02
7.85740554e-01 3.31097215e-01 8.15944910e-01 3.36446464e-02
4.57378268e-01 1.79912955e-01 -1.12070233e-01 -3.37422341e-01
2.74205208e-01 -2.12479997e-02 2.82960504e-01 5.47444642e-01
-4.08560902e-01 -4.60407376e-01 -2.04080001e-01 -1.81051031e-01
9.97303486e-01 7.29075849e-01 1.70486838e-01 -1.45026714e-01
1.26655459e-01 1.96700230e-01 5.80232143e-01 3.98864090e-01
-1.12940138e-02 7.17042089e-01 -2.28322819e-01 -4.26839799e-01
-1.00103295e+00 -9.60849404e-01 -8.75575468e-02 7.97813952e-01
2.82015532e-01 -1.65781751e-01 -8.22197735e-01 -6.76439181e-02
5.97169876e-01 -9.76566374e-02 -1.35145441e-01 2.10801110e-01
-9.14841712e-01 -4.08131033e-01 1.21666439e-01 7.57751882e-01
-6.25690967e-02 -5.96026063e-01 -1.27595925e+00 4.47269678e-01
2.39413708e-01 -9.23078060e-01 -8.66942346e-01 -1.25240115e-02
-1.09729576e+00 -1.13357043e+00 -9.18838680e-01 -5.88223875e-01
9.08945560e-01 2.67184466e-01 7.26766348e-01 2.50400126e-01
-6.89393699e-01 7.47631848e-01 -6.74944669e-02 -2.95911282e-01
9.59173739e-02 -2.22859845e-01 6.06243372e-01 -5.58528960e-01
5.70067391e-02 -3.95176679e-01 -5.70025921e-01 5.50840795e-01
-2.36455321e-01 -1.50819868e-01 5.95183633e-02 5.58113575e-01
7.54290581e-01 -1.60884470e-01 -3.41701180e-01 -1.27278447e-01
5.85110188e-01 1.89631879e-01 -9.58891034e-01 7.55775943e-02
-2.69961149e-01 -1.51418313e-01 1.78918719e-01 -1.06991744e+00
-1.09725165e+00 7.74794757e-01 1.52772972e-02 -8.53715539e-01
2.59972751e-01 1.42558485e-01 -2.32475325e-01 -5.91932893e-01
4.99635577e-01 -2.11854324e-01 3.68808478e-01 -5.38789928e-01
1.75829872e-01 4.26555008e-01 4.94028807e-01 -8.21196139e-01
8.09777796e-01 4.47389632e-01 1.71323702e-01 -8.91168654e-01
-5.60757697e-01 -4.09128308e-01 -1.11844075e+00 -3.35634559e-01
8.13442707e-01 -5.78963220e-01 -1.28191996e+00 3.97742331e-01
-1.20994556e+00 -1.80136636e-01 1.28476713e-02 7.46308804e-01
-4.39282686e-01 5.13837576e-01 -5.16685545e-01 -1.27743006e+00
-3.48666102e-01 -1.20616627e+00 1.02138877e+00 2.81697035e-01
-8.98657024e-01 -7.83834219e-01 1.68469802e-01 1.84847921e-01
6.83407160e-03 4.69020724e-01 3.25331151e-01 8.87588263e-02
-6.42986357e-01 -5.88853955e-01 3.99626613e-01 -2.31595889e-01
1.83451220e-01 2.96440840e-01 -5.12667239e-01 -8.01625133e-01
-1.35058137e-02 1.89671144e-01 -8.02205354e-02 5.68338215e-01
6.29169762e-01 -1.39337495e-01 -6.68838024e-01 6.07574344e-01
1.19107008e+00 3.87503892e-01 8.19214359e-02 1.88974336e-01
8.56529295e-01 3.34404051e-01 7.10627377e-01 7.37913966e-01
2.44882837e-01 1.01237249e+00 4.31073695e-01 1.29787087e-01
1.49987921e-01 1.17624432e-01 1.91689745e-01 4.65254992e-01
-6.95855379e-01 8.17041621e-02 -9.42823648e-01 -3.16800848e-02
-1.70589340e+00 -7.42137074e-01 -1.62186965e-01 2.54205084e+00
7.08246112e-01 1.71980992e-01 5.13941705e-01 1.48478016e-01
6.52504683e-01 -3.39110792e-01 -9.10706162e-01 -1.07389346e-01
4.07959640e-01 3.63429159e-01 5.74268818e-01 7.77569652e-01
-8.05793166e-01 8.12266588e-01 7.07892275e+00 -1.21121004e-01
-1.05705905e+00 -2.13288948e-01 -4.52393889e-01 -5.27425289e-01
1.27864769e-02 -9.45695490e-02 -1.23866165e+00 3.86855036e-01
9.46088657e-02 1.75269723e-01 6.17017150e-01 8.09981823e-01
2.03784227e-01 -1.76291659e-01 -1.02301240e+00 1.07661045e+00
5.06984070e-02 -8.77833128e-01 -3.78191680e-01 3.21461648e-01
4.84526634e-01 -3.89201790e-01 -3.37583534e-02 -3.47442031e-01
2.64973909e-01 -6.22094452e-01 8.00342262e-01 5.85600615e-01
6.59683704e-01 -4.56715375e-01 2.09385335e-01 5.30039370e-01
-1.29748654e+00 3.13255675e-02 -1.08366810e-01 -3.63288879e-01
5.53300858e-01 1.92767769e-01 -6.06115341e-01 5.23398519e-02
7.37871408e-01 9.88878608e-02 1.38810977e-01 8.82444978e-01
1.28241420e-01 -9.62082949e-03 -8.83445323e-01 -1.98968396e-01
-3.46458495e-01 -2.44939923e-01 8.68570626e-01 8.81506741e-01
2.87103802e-01 6.86935484e-01 5.82861602e-01 7.18368769e-01
5.20562291e-01 -2.42521212e-01 -4.79000777e-01 3.76543790e-01
6.20983779e-01 1.02022219e+00 -7.40736008e-01 -3.31258252e-02
-2.03461826e-01 1.14354455e+00 -1.66303739e-01 2.39733353e-01
-6.43521726e-01 -1.20979743e-02 8.34731817e-01 4.53223825e-01
-2.46348046e-02 -9.02915895e-01 -5.06823242e-01 -1.13614786e+00
3.97869617e-01 -7.98626661e-01 1.91749856e-01 -6.11616373e-01
-9.19394195e-01 1.74477011e-01 1.43300341e-02 -1.22761416e+00
-3.25705171e-01 -7.68082142e-01 -4.15675998e-01 1.20986521e+00
-6.99716330e-01 -9.50796783e-01 -5.46869397e-01 8.35664272e-01
3.53285581e-01 1.56818867e-01 8.84767294e-01 1.93705663e-01
-5.32920957e-01 4.58024800e-01 -4.19659913e-01 -1.12233646e-01
6.99777663e-01 -1.01186311e+00 3.40685546e-01 6.06217563e-01
-4.85653192e-01 1.06661046e+00 7.64712334e-01 -1.14914536e+00
-2.28924704e+00 -3.49299043e-01 5.11705041e-01 -9.67502773e-01
2.71000296e-01 -2.87321746e-01 -7.66291797e-01 7.31315315e-01
-4.46521670e-01 9.95720327e-02 4.54161286e-01 1.39724255e-01
1.17115200e-01 3.14254612e-01 -1.30684686e+00 7.19400108e-01
1.30145621e+00 -2.67380208e-01 -5.40239573e-01 2.21531004e-01
1.45626336e-01 -1.46000671e+00 -1.05777991e+00 1.66090369e-01
1.23348153e+00 -4.49344993e-01 1.30502295e+00 -4.94873941e-01
-4.72781360e-01 -5.09818077e-01 -7.53334388e-02 -7.11458325e-01
-4.42184478e-01 -1.07793319e+00 -6.41183913e-01 8.75928819e-01
-7.16059506e-02 -1.76569283e-01 1.22182691e+00 1.42643917e+00
2.08372250e-01 -7.58711994e-01 -8.23949575e-01 -7.72368789e-01
-4.42781091e-01 -2.60915369e-01 2.93904960e-01 8.50266695e-01
7.53017962e-02 7.11668730e-02 -3.62528533e-01 4.09165919e-01
9.58820224e-01 3.33447680e-02 9.67673838e-01 -1.46161091e+00
-2.77722806e-01 -2.40836844e-01 -2.42244616e-01 -1.21822846e+00
-1.79497391e-01 -4.04470205e-01 6.63152412e-02 -1.34379041e+00
-6.57357424e-02 -5.58457017e-01 4.03420150e-01 4.26727593e-01
-7.89351314e-02 9.67961624e-02 3.82361323e-01 2.92718560e-01
-4.82727848e-02 -1.18358128e-01 1.32702208e+00 3.38189423e-01
-8.18262279e-01 3.81739169e-01 -8.59655812e-02 1.03148961e+00
4.37717617e-01 -9.59004834e-02 -9.44012851e-02 -6.39420748e-01
1.46476492e-01 4.67408955e-01 4.87824947e-01 -7.42794633e-01
4.86442327e-01 -3.56161177e-01 7.99575210e-01 -7.88442850e-01
7.69930005e-01 -1.17076898e+00 5.79057038e-01 7.90824950e-01
2.10116744e-01 3.86200309e-01 1.14881277e-01 4.81842123e-02
4.70962763e-01 -1.14673167e-01 7.78983831e-01 -3.05249393e-01
-2.49656677e-01 3.51928771e-01 -2.82652587e-01 -3.25066298e-01
1.13239539e+00 -9.33994949e-01 4.99308258e-01 -5.05370684e-02
-1.12441838e+00 1.35757014e-01 8.20587516e-01 4.33229089e-01
6.40389025e-01 -1.21027613e+00 -1.39515683e-01 2.63157576e-01
-6.18678749e-01 3.63133997e-01 -7.75063504e-03 6.20830119e-01
-5.98529339e-01 1.56360775e-01 -2.21719638e-01 -8.44627440e-01
-1.73174429e+00 3.33793908e-01 3.43948275e-01 2.09718838e-01
-8.67730796e-01 8.10980618e-01 -5.94739079e-01 -4.10118878e-01
5.24123311e-01 -3.01457494e-01 2.73617297e-01 -9.05088335e-02
3.01471561e-01 8.14566672e-01 -9.54792574e-02 -5.80868721e-01
-6.97500765e-01 1.25384271e+00 2.89832413e-01 -3.00081730e-01
9.45822239e-01 -1.41114786e-01 1.20041959e-01 2.05248091e-02
7.20831692e-01 1.62293747e-01 -1.55287623e+00 3.36129358e-03
-2.37386689e-01 -7.02210844e-01 -1.94980890e-01 -5.68595350e-01
-6.74119532e-01 5.90838969e-01 6.50974393e-01 -2.34702975e-01
6.64806187e-01 -3.14027637e-01 7.90757298e-01 2.67791927e-01
6.83209479e-01 -1.20176923e+00 -9.34137590e-03 6.58220649e-01
9.03305292e-01 -7.50416875e-01 4.25013602e-01 -7.07391798e-01
-3.42367589e-01 1.09680641e+00 8.48277271e-01 -2.00742081e-01
6.28370523e-01 9.37091649e-01 1.14149287e-01 2.84136683e-02
-1.05852969e-01 3.32784206e-02 3.82501572e-01 5.80478728e-01
4.35950011e-01 -1.21130303e-01 -9.56788510e-02 2.76276082e-01
-2.85935104e-01 1.50747716e-01 4.38483246e-02 1.53113723e+00
-3.06630939e-01 -1.25395262e+00 -1.04822230e+00 2.44315937e-01
-2.41929576e-01 4.33544010e-01 -3.77038211e-01 8.94557595e-01
-7.09818453e-02 5.41988969e-01 -4.83191535e-02 -3.06282103e-01
7.95583427e-01 2.30238568e-02 1.11357987e+00 -6.80069804e-01
-6.50590420e-01 6.46875083e-01 -2.47234374e-01 -6.99846387e-01
-1.60622135e-01 -1.01667190e+00 -1.59040630e+00 -4.15410727e-01
-6.43394113e-01 -3.48522365e-01 7.86155343e-01 7.68814445e-01
3.48273128e-01 2.50399739e-01 2.20952749e-01 -1.61187148e+00
-8.60899389e-01 -4.52484131e-01 -3.53804111e-01 8.84845331e-02
2.03808904e-01 -1.13386989e+00 -8.93342867e-03 -4.67135757e-03] | [6.620931148529053, -0.9464094638824463] |
11fd3571-1d46-43bb-8ecf-5241615e692d | bootstrapping-objectness-from-videos-by | 2304.08025 | null | https://arxiv.org/abs/2304.08025v1 | https://arxiv.org/pdf/2304.08025v1.pdf | Bootstrapping Objectness from Videos by Relaxed Common Fate and Visual Grouping | We study learning object segmentation from unlabeled videos. Humans can easily segment moving objects without knowing what they are. The Gestalt law of common fate, i.e., what move at the same speed belong together, has inspired unsupervised object discovery based on motion segmentation. However, common fate is not a reliable indicator of objectness: Parts of an articulated / deformable object may not move at the same speed, whereas shadows / reflections of an object always move with it but are not part of it. Our insight is to bootstrap objectness by first learning image features from relaxed common fate and then refining them based on visual appearance grouping within the image itself and across images statistically. Specifically, we learn an image segmenter first in the loop of approximating optical flow with constant segment flow plus small within-segment residual flow, and then by refining it for more coherent appearance and statistical figure-ground relevance. On unsupervised video object segmentation, using only ResNet and convolutional heads, our model surpasses the state-of-the-art by absolute gains of 7/9/5% on DAVIS16 / STv2 / FBMS59 respectively, demonstrating the effectiveness of our ideas. Our code is publicly available. | ['Stella X. Yu', 'Zhirong Wu', 'Long Lian'] | 2023-04-17 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Lian_Bootstrapping_Objectness_From_Videos_by_Relaxed_Common_Fate_and_Visual_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Lian_Bootstrapping_Objectness_From_Videos_by_Relaxed_Common_Fate_and_Visual_CVPR_2023_paper.pdf | cvpr-2023-1 | ['object-discovery', 'unsupervised-object-segmentation', 'motion-segmentation', 'video-object-segmentation', 'video-semantic-segmentation', 'unsupervised-video-object-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 2.79639065e-01 9.17959213e-02 -2.72173017e-01 -2.03987971e-01
-1.97461993e-02 -9.81205821e-01 3.91387194e-01 -1.23597212e-01
-3.72926205e-01 4.33884948e-01 -7.70991109e-03 -7.87664428e-02
-7.15953186e-02 -5.43438077e-01 -1.03277826e+00 -7.69510210e-01
-1.89499810e-01 5.24982929e-01 9.05298948e-01 3.22501622e-02
1.24745406e-01 5.45522809e-01 -1.44999099e+00 1.84194967e-01
6.45096958e-01 9.25856590e-01 2.33901069e-01 1.08598530e+00
-4.70035002e-02 1.18092155e+00 -3.18524420e-01 -1.25497729e-01
6.01714492e-01 -5.03829420e-01 -1.50202239e+00 5.18640995e-01
1.10189903e+00 -6.31225705e-01 -5.43099999e-01 9.63469088e-01
-1.89929962e-01 4.77904737e-01 9.55269277e-01 -1.26671636e+00
-6.55719459e-01 6.51979387e-01 -7.41043985e-01 5.41957676e-01
1.07183896e-01 4.71271664e-01 1.15995002e+00 -4.03118879e-01
9.24638927e-01 1.25739622e+00 3.83542180e-01 5.82161903e-01
-1.29779899e+00 -2.44956657e-01 4.44199145e-01 7.24629164e-02
-1.14158595e+00 -3.06129634e-01 6.18102551e-01 -6.99696600e-01
6.34442985e-01 3.61399114e-01 7.56058156e-01 6.30135953e-01
8.05876106e-02 1.16119707e+00 6.81721032e-01 1.93603970e-02
1.39172167e-01 -2.44991362e-01 3.80929321e-01 8.83446991e-01
3.51482987e-01 -1.37621373e-01 -1.71358168e-01 3.42538238e-01
1.03759682e+00 -2.91757826e-02 -3.68206024e-01 -4.20445979e-01
-1.49544013e+00 3.45349371e-01 6.02355838e-01 2.95574546e-01
-2.10668206e-01 6.42139792e-01 6.26714230e-02 5.26493527e-02
2.17721269e-01 2.95979112e-01 -5.56661963e-01 -4.45224121e-02
-1.28238332e+00 1.78601131e-01 7.26583064e-01 1.08625615e+00
9.82111990e-01 9.55750700e-03 -3.78749110e-02 1.70736596e-01
3.91651720e-01 4.67511088e-01 3.94321293e-01 -1.43679702e+00
8.34049657e-02 6.16299391e-01 1.61926985e-01 -9.06001329e-01
-4.34932232e-01 -2.05362439e-01 -7.56601870e-01 2.84426332e-01
1.01811278e+00 -8.51646140e-02 -1.40549219e+00 1.55547643e+00
5.17305315e-01 5.85366488e-01 -1.13846868e-01 1.26876032e+00
9.60397601e-01 7.09432483e-01 1.38813108e-01 1.49050038e-02
1.38640451e+00 -1.11523402e+00 -2.50572234e-01 -2.43214101e-01
6.16703093e-01 -7.40797222e-01 7.63835073e-01 3.24730426e-01
-1.15364254e+00 -9.00964856e-01 -9.28252339e-01 -2.44346276e-01
3.86775993e-02 -2.22343042e-01 7.52684355e-01 6.69370830e-01
-9.76330519e-01 9.85650718e-01 -9.57105041e-01 -3.04944932e-01
7.59133875e-01 3.55131626e-01 -4.45886642e-01 1.26099467e-01
-6.20650113e-01 3.60976666e-01 5.06518841e-01 5.88349737e-02
-1.05464423e+00 -7.93783367e-01 -5.84137022e-01 -2.14769900e-01
3.13496202e-01 -8.79955828e-01 9.98220384e-01 -1.49134433e+00
-1.31714964e+00 8.11894178e-01 -8.95357877e-02 -5.41833699e-01
7.92402208e-01 -4.30983871e-01 -3.59165967e-02 5.92192590e-01
6.49908707e-02 1.18643582e+00 9.28501904e-01 -1.41536427e+00
-8.30173612e-01 -1.84692040e-01 3.06300253e-01 1.68999627e-01
2.19120562e-01 -1.58269972e-01 -5.97818196e-01 -4.70781237e-01
3.95095348e-01 -1.16923034e+00 -2.01961815e-01 2.74487764e-01
-5.30460358e-01 -1.84493944e-01 1.14764011e+00 -6.73806727e-01
8.48641515e-01 -2.07733727e+00 1.88836902e-01 -1.47120319e-02
4.22032177e-01 1.63129568e-01 -1.63587362e-01 -2.37032816e-01
-4.48217662e-03 3.58410567e-01 -4.37460482e-01 2.96046156e-02
-2.97847360e-01 1.85751960e-01 -2.67937154e-01 9.91625547e-01
2.84755290e-01 1.12878537e+00 -1.09604537e+00 -6.53862000e-01
3.65497023e-01 1.50727019e-01 -6.10779166e-01 6.92078024e-02
-3.62038016e-01 6.05922699e-01 -4.50733006e-01 5.99449515e-01
7.55444765e-01 -5.81001043e-01 -1.68752119e-01 -3.21621656e-01
-8.77350196e-02 -1.92307845e-01 -1.58607209e+00 1.43829751e+00
1.97060168e-01 9.39221919e-01 -2.30994094e-02 -9.50782180e-01
3.42660666e-01 5.05827516e-02 8.26388121e-01 -2.23230794e-01
1.11766405e-01 8.32631718e-03 2.94050187e-01 -7.62421668e-01
5.33174217e-01 1.88540429e-01 3.53306890e-01 5.41620374e-01
2.05329601e-02 -3.31787914e-01 4.11101103e-01 3.86743098e-01
9.43412066e-01 5.06024599e-01 -9.51965060e-03 -5.02243757e-01
3.50232542e-01 7.89207220e-02 4.58564818e-01 8.33689392e-01
-4.40584570e-01 1.01217604e+00 3.76648575e-01 -3.92498791e-01
-1.02115846e+00 -1.29632521e+00 -1.01858154e-01 1.22177279e+00
9.55311120e-01 1.03607051e-01 -7.36372411e-01 -7.96362758e-01
-1.42140901e-02 2.97651768e-01 -8.11589837e-01 2.37865090e-01
-7.98471570e-01 -3.83828402e-01 4.65611875e-01 6.62277102e-01
5.62658012e-01 -1.08509779e+00 -1.03335989e+00 1.25592962e-01
-3.48439127e-01 -1.21852481e+00 -7.09322393e-01 -2.12101396e-02
-9.93542135e-01 -1.32723355e+00 -8.51664960e-01 -7.98912048e-01
7.74767339e-01 6.22317314e-01 1.16596651e+00 4.03767943e-01
-4.74189669e-01 4.75974530e-01 -3.23558569e-01 1.07608043e-01
-2.38252074e-01 -9.29399207e-02 -5.59157953e-02 1.72973722e-01
-3.05608332e-01 -3.23459804e-01 -1.08111060e+00 5.27909577e-01
-1.03932130e+00 1.20505810e-01 3.97424638e-01 2.82120913e-01
4.75943327e-01 -9.56563354e-02 -4.70191054e-03 -7.31320858e-01
-3.47042114e-01 -4.17042077e-01 -3.62198383e-01 1.42678410e-01
3.46674211e-02 3.09636164e-02 1.73352957e-01 -6.53189898e-01
-9.93900239e-01 3.21601361e-01 3.31736714e-01 -4.86363083e-01
-4.29790884e-01 -2.23154575e-01 -4.24953038e-03 1.06022926e-02
5.91508985e-01 1.24132901e-01 -1.31836951e-01 -7.56811202e-02
8.50662053e-01 2.02196553e-01 7.32742727e-01 -4.84019935e-01
8.91401589e-01 1.18602335e+00 1.20395944e-02 -1.17903352e+00
-6.97329938e-01 -9.35050786e-01 -9.80603814e-01 -5.43843627e-01
1.42324567e+00 -6.99447513e-01 -7.23395348e-01 5.58113575e-01
-1.16903865e+00 -6.71206415e-01 -4.61636394e-01 5.14689624e-01
-6.27397239e-01 5.73635995e-01 -6.12496793e-01 -5.30320585e-01
-6.62346259e-02 -1.12492621e+00 1.05571759e+00 5.34933031e-01
-2.82263637e-01 -9.25348103e-01 -2.83766150e-01 4.51865166e-01
5.47309853e-02 3.75628799e-01 5.15038610e-01 -4.97529268e-01
-1.19882715e+00 3.18221271e-01 -4.35618073e-01 3.61889124e-01
3.42535853e-01 5.35814643e-01 -7.79398084e-01 -1.39056191e-01
-3.33834916e-01 -9.30930153e-02 1.12335253e+00 8.40645850e-01
1.05751550e+00 -3.45512986e-01 -4.28654999e-01 7.98305452e-01
1.37824941e+00 2.55728494e-02 8.16207767e-01 1.03528462e-01
1.18665648e+00 6.57276094e-01 3.26078951e-01 1.08213827e-01
2.02252403e-01 3.03203821e-01 5.99666297e-01 -2.63911456e-01
-5.29451430e-01 2.45888501e-01 3.70945007e-01 2.53515959e-01
-5.62019587e-01 -4.35952157e-01 -8.16806674e-01 6.94742978e-01
-1.71178472e+00 -1.11103940e+00 -7.85018981e-01 1.92265773e+00
5.62601507e-01 2.56088704e-01 3.44181627e-01 -1.29383355e-01
5.95198572e-01 1.55296504e-01 -5.75011611e-01 7.20732659e-02
-2.12080032e-01 -7.81106576e-02 8.48595381e-01 5.02074659e-01
-1.33554506e+00 1.15959442e+00 5.78102541e+00 5.54688394e-01
-1.24172020e+00 -8.52127373e-02 9.73516643e-01 1.18768923e-01
-1.27808124e-01 2.36651242e-01 -7.13772416e-01 3.28999937e-01
3.26112777e-01 9.53659341e-02 2.67772257e-01 5.66432238e-01
1.55019477e-01 -3.34518701e-01 -1.21950424e+00 5.64045787e-01
-1.05605386e-01 -1.39927018e+00 7.86007643e-02 -8.18418059e-03
1.06434751e+00 1.30395621e-01 -8.29563662e-03 -1.43630773e-01
2.66599774e-01 -9.78039861e-01 1.23531258e+00 6.21033907e-01
3.12674969e-01 -2.23380581e-01 4.24041808e-01 1.14604704e-01
-1.47530735e+00 1.90192491e-01 -1.66971758e-01 1.50861621e-01
2.42200598e-01 2.64705479e-01 -6.40778422e-01 4.69219238e-01
7.49748349e-01 7.93451548e-01 -6.50246382e-01 1.09734535e+00
-5.34442775e-02 7.67123103e-01 -4.30376023e-01 1.71719253e-01
4.50064480e-01 -2.34283447e-01 7.12746680e-01 1.21276283e+00
-1.31798774e-01 3.29755306e-01 2.86415339e-01 1.00543869e+00
2.08730418e-02 -1.26662537e-01 -1.45035505e-01 -3.35002579e-02
-1.37144653e-02 1.28691280e+00 -1.58728051e+00 -5.47908008e-01
-3.78959090e-01 1.06297457e+00 -1.23016357e-01 6.22393250e-01
-1.02707469e+00 -1.57819465e-01 6.21512890e-01 3.98035139e-01
6.47741079e-01 -4.15118158e-01 -3.02424077e-02 -9.60319817e-01
-1.74618647e-01 -4.44158942e-01 1.95734739e-01 -6.70739770e-01
-1.11317098e+00 3.97153497e-01 5.31062298e-02 -1.37805343e+00
6.28526360e-02 -5.92991948e-01 -6.08753502e-01 2.94112533e-01
-1.12254894e+00 -1.06983781e+00 -4.65910435e-01 4.65298384e-01
7.57671893e-01 4.51651275e-01 4.02485952e-03 -8.32019001e-02
-2.42458597e-01 1.05476581e-01 -3.25469561e-02 4.43426281e-01
3.91879201e-01 -1.29454744e+00 4.92719501e-01 8.97617400e-01
4.91237819e-01 4.84957993e-01 7.46451795e-01 -7.02145934e-01
-1.23799384e+00 -1.26934469e+00 2.55879551e-01 -6.71522081e-01
6.81779802e-01 -1.53856814e-01 -1.08499169e+00 7.23426580e-01
1.86397240e-01 3.68459880e-01 6.96235597e-02 -3.74099970e-01
-2.24353507e-01 1.91970229e-01 -8.11578035e-01 6.74603701e-01
1.25817895e+00 -4.46569249e-02 -6.11746371e-01 1.97985038e-01
8.43814433e-01 -4.21251446e-01 -6.39930964e-01 4.35996771e-01
6.15072310e-01 -1.06276298e+00 1.12764847e+00 -7.79552042e-01
5.49190700e-01 -6.28029943e-01 6.52566832e-03 -7.91777551e-01
-3.45694870e-01 -7.36936986e-01 -1.60928100e-01 9.72972929e-01
1.59246907e-01 -2.49955103e-01 1.00008619e+00 4.72268254e-01
-5.04019372e-02 -4.56603825e-01 -7.94327199e-01 -8.78052354e-01
6.78560287e-02 -4.95819449e-01 1.74763292e-01 8.19904208e-01
-5.69918692e-01 3.69140469e-02 1.00907423e-02 2.72837937e-01
7.38450348e-01 2.12445751e-01 9.22764778e-01 -1.12592328e+00
-2.44195655e-01 -6.97875679e-01 -7.05253899e-01 -1.54657388e+00
6.62104040e-02 -7.85749257e-01 1.83170572e-01 -1.60194349e+00
3.14688325e-01 -4.40068036e-01 3.62774655e-02 1.85636178e-01
-2.17498794e-01 5.59176564e-01 3.62951398e-01 4.54428256e-01
-9.06694114e-01 8.24573860e-02 1.69704449e+00 -4.35887307e-01
-3.85229588e-01 -3.86821441e-02 -3.63651037e-01 1.02645683e+00
5.74144483e-01 -2.88176745e-01 -3.33695948e-01 -3.53801191e-01
-1.67323016e-02 -1.99165374e-01 7.07972229e-01 -1.15130389e+00
4.52092066e-02 -3.33481103e-01 4.91615325e-01 -4.90497351e-01
9.50209945e-02 -7.45543301e-01 1.83674231e-01 5.93462348e-01
-3.09888661e-01 -2.60990888e-01 1.60036936e-01 8.29497397e-01
2.20100433e-01 -2.94560820e-01 9.38125432e-01 -3.48718286e-01
-1.11107397e+00 4.92281497e-01 -4.48808849e-01 4.42772597e-01
1.14126003e+00 -6.02372229e-01 -4.06040251e-01 -2.85530955e-01
-7.97826648e-01 1.36945277e-01 5.56125939e-01 4.35532779e-01
4.99194264e-01 -9.14782047e-01 -8.15794289e-01 -6.19048364e-02
-1.45901129e-01 3.82929772e-01 3.36084753e-01 8.99025321e-01
-9.65883732e-01 1.16188966e-01 -1.76116660e-01 -1.23408163e+00
-1.24255574e+00 4.18182284e-01 4.68489230e-01 3.88900280e-01
-8.54041934e-01 1.21863842e+00 8.24175537e-01 2.65930593e-01
8.33050162e-02 -8.49771202e-01 -8.31311494e-02 3.81056927e-02
4.12429184e-01 4.91929799e-01 -4.09600049e-01 -9.46056068e-01
-4.74079967e-01 9.44513202e-01 -5.65514192e-02 -1.01569118e-02
9.96522844e-01 -2.55321413e-01 -4.83965054e-02 4.77261633e-01
1.18503070e+00 -4.84831929e-02 -1.77971840e+00 3.66308056e-02
-1.39626458e-01 -5.44118285e-01 -1.70209140e-01 -2.74923056e-01
-1.34067023e+00 5.83094358e-01 4.25106317e-01 4.84228343e-01
8.38025808e-01 5.11159003e-01 6.84113741e-01 1.82860240e-01
1.82858333e-01 -9.51740324e-01 4.54219520e-01 4.16013360e-01
6.35430336e-01 -1.23241913e+00 2.34288350e-01 -4.78125036e-01
-6.38622999e-01 1.14549363e+00 6.40719533e-01 -4.43564355e-01
6.28425658e-01 -1.88733861e-02 -1.02521330e-01 -1.60455957e-01
-3.23148876e-01 -4.01431113e-01 7.03256547e-01 4.56671625e-01
3.29650134e-01 4.37315255e-02 1.14087775e-01 -4.32715714e-02
-9.29062590e-02 -3.64245057e-01 5.67143977e-01 6.97749734e-01
-5.95912039e-01 -3.61007571e-01 -2.68733412e-01 3.82042021e-01
-3.80592495e-01 8.13918263e-02 -2.19922930e-01 1.01602077e+00
3.99785519e-01 8.50040376e-01 4.94502097e-01 -1.98445633e-01
6.81442469e-02 -3.98541600e-01 8.02084029e-01 -4.04542446e-01
-3.09017688e-01 7.51822442e-02 -2.83665866e-01 -7.57610619e-01
-8.86427462e-01 -8.61955464e-01 -1.76879799e+00 -2.03148514e-01
-4.33509231e-01 -3.16342264e-01 2.58692414e-01 1.10021782e+00
-1.13579996e-01 6.01902425e-01 4.18887407e-01 -1.15616798e+00
-4.32512835e-02 -6.97968066e-01 -4.62957531e-01 7.39406526e-01
6.11577868e-01 -4.70021725e-01 -6.12402141e-01 9.13888872e-01] | [9.04663372039795, -0.3241814374923706] |
fd117b0c-d79b-4999-b8c1-f8f2daec212b | adaptive-decision-making-at-the-intersection | 2207.11724 | null | https://arxiv.org/abs/2207.11724v1 | https://arxiv.org/pdf/2207.11724v1.pdf | Adaptive Decision Making at the Intersection for Autonomous Vehicles Based on Skill Discovery | In urban environments, the complex and uncertain intersection scenarios are challenging for autonomous driving. To ensure safety, it is crucial to develop an adaptive decision making system that can handle the interaction with other vehicles. Manually designed model-based methods are reliable in common scenarios. But in uncertain environments, they are not reliable, so learning-based methods are proposed, especially reinforcement learning (RL) methods. However, current RL methods need retraining when the scenarios change. In other words, current RL methods cannot reuse accumulated knowledge. They forget learned knowledge when new scenarios are given. To solve this problem, we propose a hierarchical framework that can autonomously accumulate and reuse knowledge. The proposed method combines the idea of motion primitives (MPs) with hierarchical reinforcement learning (HRL). It decomposes complex problems into multiple basic subtasks to reduce the difficulty. The proposed method and other baseline methods are tested in a challenging intersection scenario based on the CARLA simulator. The intersection scenario contains three different subtasks that can reflect the complexity and uncertainty of real traffic flow. After offline learning and testing, the proposed method is proved to have the best performance among all methods. | ['Jianwei Gong', 'Zirui Li', 'Chao Lu', 'Lin Yang', 'Xianqi He'] | 2022-07-24 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [-1.81251332e-01 -4.86209206e-02 -2.52739072e-01 -3.72494578e-01
-6.54118001e-01 -3.54629159e-01 5.30742288e-01 -8.54914263e-02
-5.40591955e-01 1.21854067e+00 -1.23149060e-01 -5.17540038e-01
-2.44797722e-01 -8.17044079e-01 -7.10051298e-01 -7.11872399e-01
-7.26385564e-02 4.98438716e-01 9.02363062e-01 -6.29725993e-01
3.18721116e-01 5.01309097e-01 -2.00202107e+00 -1.04080765e-02
1.13197541e+00 5.99149883e-01 6.20906889e-01 4.44564492e-01
-1.75654575e-01 7.12188482e-01 -2.71343172e-01 1.34801656e-01
3.65406096e-01 3.18840779e-02 -6.38105452e-01 -5.02610654e-02
-1.95879132e-01 -4.71668601e-01 -4.85157102e-01 7.39057779e-01
4.34357882e-01 5.70968628e-01 4.59905416e-01 -1.79714191e+00
3.66128743e-01 4.49091226e-01 -2.50111043e-01 2.21969500e-01
1.76576734e-01 5.92872024e-01 4.18264508e-01 -6.34542525e-01
3.33734006e-01 1.46305430e+00 3.75475168e-01 5.40368557e-01
-6.68340921e-01 -9.06876385e-01 6.94875777e-01 1.01566303e+00
-1.25196350e+00 -3.85841638e-01 5.88379443e-01 -3.16603750e-01
6.57137394e-01 -9.90524739e-02 6.20725930e-01 8.61599267e-01
4.09106433e-01 9.54478920e-01 1.16702831e+00 7.68676773e-02
4.96958196e-01 2.87035167e-01 4.40459475e-02 3.94178063e-01
4.64209586e-01 5.31504750e-01 -8.93446058e-02 1.84227571e-01
9.87834260e-02 -3.24098803e-02 7.25928023e-02 -4.29192275e-01
-1.00299227e+00 6.99314594e-01 9.21299011e-02 6.14200011e-02
-2.37643570e-01 6.61175326e-02 3.68640125e-01 2.73709536e-01
-4.48212594e-01 1.41956881e-01 -3.86913061e-01 -2.54450142e-01
-5.39864659e-01 7.09842503e-01 5.55111468e-01 1.20539224e+00
1.07010961e+00 3.11030865e-01 -2.89271683e-01 4.43169773e-01
4.76870656e-01 5.31550407e-01 4.11698312e-01 -9.11575317e-01
5.35403728e-01 4.10792410e-01 4.62300479e-01 -9.79428649e-01
-6.87133372e-01 -2.72406846e-01 -6.50886893e-01 5.23559928e-01
1.76451117e-01 -3.40867341e-01 -8.33607674e-01 1.52391613e+00
5.90926051e-01 7.47361481e-01 2.97068894e-01 5.68320215e-01
6.95619524e-01 8.98968637e-01 2.95905560e-01 -3.30384701e-01
7.40202248e-01 -9.13566411e-01 -9.64060128e-01 -2.32858539e-01
4.19296980e-01 -5.86895883e-01 7.18193889e-01 4.66330588e-01
-7.87215650e-01 -9.37805057e-01 -1.42083514e+00 5.81813574e-01
-5.57490230e-01 -3.42368901e-01 4.58970100e-01 5.18802702e-01
-8.18525434e-01 4.61045653e-01 -4.87942427e-01 -6.15854003e-02
7.47700362e-03 4.45503026e-01 -7.25875199e-02 -2.50109076e-01
-1.72184372e+00 1.08386624e+00 7.79692829e-01 3.71119738e-01
-1.34448028e+00 -3.65445256e-01 -9.43379819e-01 -1.96434125e-01
7.55362630e-01 -3.59053999e-01 1.42930400e+00 -4.93173957e-01
-1.85290921e+00 -1.29523173e-01 -4.79742400e-02 -4.92079258e-01
5.59336066e-01 -1.00958653e-01 -6.18818223e-01 -2.35677317e-01
-9.86849330e-03 5.17779648e-01 8.05435181e-01 -1.75417483e+00
-1.28386593e+00 1.34957790e-01 2.84275651e-01 3.44575226e-01
3.47817242e-01 -6.68934762e-01 -3.12952727e-01 -1.29567340e-01
-1.29212990e-01 -1.11887133e+00 -5.34024835e-01 -5.72024286e-01
1.64577425e-01 -3.45030099e-01 1.08375192e+00 -2.20187679e-01
1.33556187e+00 -2.04009986e+00 -2.47250959e-01 3.13953906e-01
-2.16419712e-01 5.20891130e-01 -2.80320439e-02 4.48286682e-01
4.01039153e-01 -2.65593767e-01 -2.12211069e-02 1.08029209e-01
8.91173184e-02 6.08821690e-01 -1.50089934e-01 -4.74524796e-02
-6.38011992e-02 6.52519286e-01 -1.07474840e+00 -8.02926421e-01
5.99295914e-01 -8.92694071e-02 -4.38999385e-01 2.89530993e-01
-3.51712555e-01 7.33453035e-01 -7.87002683e-01 3.98901045e-01
9.81907547e-01 4.76411253e-01 -4.85730432e-02 -4.42084012e-04
-3.86890650e-01 -2.40860149e-01 -1.79651201e+00 1.38125181e+00
-6.19446218e-01 3.80350560e-01 2.72658188e-02 -1.13204122e+00
9.14199829e-01 1.96923196e-01 3.79902661e-01 -7.98540473e-01
6.01071678e-02 1.41005114e-01 1.68791845e-01 -1.04263842e+00
5.29856086e-01 -4.01093246e-04 -2.09756404e-01 -6.64763898e-02
-2.67391294e-01 -3.93496841e-01 1.65286481e-01 -4.07255851e-02
1.13215661e+00 4.02153373e-01 1.75509542e-01 -6.78889528e-02
1.03260028e+00 2.03403115e-01 1.11214447e+00 8.51266742e-01
-6.19909108e-01 -1.89303815e-01 8.97197202e-02 -5.96873581e-01
-5.56237936e-01 -9.97782826e-01 1.47238359e-01 7.71211863e-01
7.89176941e-01 -1.11384459e-01 -4.44028795e-01 -6.42754734e-01
4.61177528e-03 1.11538088e+00 -2.13964432e-01 -5.17868161e-01
-6.29516721e-01 -5.16869247e-01 1.64356798e-01 2.00879157e-01
9.09431100e-01 -1.03064835e+00 -6.61816418e-01 5.23917973e-01
-2.12380141e-01 -1.26141822e+00 -9.63682458e-02 -2.52563536e-01
-6.23999417e-01 -1.09261978e+00 -3.26027796e-02 -6.47940040e-01
3.05483103e-01 5.40111661e-01 6.37670994e-01 4.95905504e-02
1.00765362e-01 3.41561288e-01 -4.38343614e-01 -6.31494224e-01
-4.71224904e-01 6.96269497e-02 1.68029532e-01 1.56245515e-01
1.18347600e-01 -3.82400125e-01 -5.29174507e-01 7.81693637e-01
-8.29498708e-01 4.55929665e-03 9.31645155e-01 7.30014503e-01
5.36038399e-01 6.32495105e-01 1.00489724e+00 -5.79495966e-01
5.98767638e-01 -6.37935579e-01 -7.73317814e-01 2.07006395e-01
-7.09267497e-01 3.26896071e-01 7.41036057e-01 -3.14084500e-01
-1.41495788e+00 1.72582462e-01 -1.99182749e-01 -8.47808495e-02
-5.24855196e-01 3.06254417e-01 -5.72628319e-01 -8.93255398e-02
3.02488565e-01 1.50697902e-01 -2.52561182e-01 8.10926259e-02
1.69663966e-01 6.22659743e-01 1.96367323e-01 -6.45655870e-01
9.40667212e-01 1.08919486e-01 3.17928404e-01 -6.28873348e-01
-5.18154025e-01 -4.32550251e-01 -5.52882195e-01 -7.66777396e-01
6.79640293e-01 -8.74264240e-01 -1.04569519e+00 3.54712039e-01
-8.86429489e-01 -5.55526555e-01 -2.54532415e-02 7.38383532e-01
-8.29419792e-01 3.15190405e-01 9.15068910e-02 -8.84026706e-01
3.37714761e-01 -1.56136107e+00 6.00609183e-01 4.01017517e-01
3.17351520e-01 -6.28349781e-01 2.00426430e-02 2.61838466e-01
5.13660014e-01 4.34547216e-01 7.98887849e-01 -3.00276250e-01
-9.51296926e-01 -6.14222772e-02 1.47792473e-01 7.15495832e-03
9.99275073e-02 -8.83438140e-02 -6.80091202e-01 -1.51364103e-01
-1.99080288e-01 -1.55527592e-01 6.89913273e-01 1.36934489e-01
1.04931533e+00 -2.05387414e-01 -5.28186142e-01 8.84187408e-03
1.25874448e+00 7.54057229e-01 7.54703164e-01 6.01234376e-01
2.87914336e-01 7.62531579e-01 1.40280163e+00 3.99178743e-01
9.16724801e-01 6.54386699e-01 5.77974975e-01 3.60529035e-01
1.49387628e-01 -3.02734643e-01 4.16607678e-01 6.95748985e-01
8.06164369e-02 -4.04350869e-02 -9.08320665e-01 4.98851657e-01
-2.45828247e+00 -1.21399355e+00 -4.01355401e-02 2.20655918e+00
5.48727512e-01 7.73908257e-01 5.54145798e-02 2.87768066e-01
5.68237245e-01 -1.14451967e-01 -7.70585537e-01 -4.14383620e-01
2.50993013e-01 -2.68314689e-01 5.18214345e-01 7.39815295e-01
-8.50853026e-01 1.02015209e+00 5.82652664e+00 1.01406074e+00
-8.27841341e-01 4.50108610e-02 2.42780477e-01 3.53385657e-01
-1.78784251e-01 1.47008076e-01 -1.06223524e+00 4.77427661e-01
1.03456068e+00 -3.01923007e-01 2.56537855e-01 9.00640249e-01
6.38089240e-01 -5.30941665e-01 -8.80943835e-01 9.14917052e-01
-3.39475960e-01 -9.26885426e-01 1.41374348e-02 -4.64691073e-01
7.41735280e-01 -2.57652402e-01 -3.78462076e-02 1.11520076e+00
5.03177464e-01 -8.93125057e-01 3.47385585e-01 9.19019222e-01
1.43137187e-01 -1.11586356e+00 1.07112241e+00 6.97233021e-01
-1.33066130e+00 -4.37259287e-01 -3.92647773e-01 -5.20131504e-03
2.26816922e-01 2.09283933e-01 -9.35735047e-01 8.52534175e-01
5.73361099e-01 5.25981605e-01 -4.97119993e-01 1.37043476e+00
-9.05861706e-02 4.42810595e-01 -8.88622031e-02 -2.71630704e-01
5.23202538e-01 -3.00970823e-02 5.23613691e-01 9.14565086e-01
2.71079242e-01 1.74033925e-01 6.18440390e-01 2.70092130e-01
5.08743703e-01 4.85907421e-02 -8.51481140e-01 6.45318925e-01
4.62355584e-01 1.17459333e+00 -4.24035668e-01 -2.95082629e-01
-3.16202402e-01 2.38165334e-01 1.12748235e-01 4.32216406e-01
-1.06449187e+00 -3.56298983e-01 5.28277099e-01 1.31426215e-01
1.55265257e-01 -5.47189713e-01 1.91949204e-01 -6.13203824e-01
-1.97768405e-01 -8.27647448e-01 3.06143194e-01 -6.92005694e-01
-8.09621930e-01 4.88991112e-01 5.52904844e-01 -1.54409981e+00
-3.24945986e-01 -3.89739424e-01 -6.11589849e-01 3.18077296e-01
-1.97605038e+00 -8.83175552e-01 -5.86224973e-01 8.39954138e-01
1.10861158e+00 -3.34907025e-01 3.08572918e-01 4.03160512e-01
-4.69723850e-01 3.73164356e-01 2.69870553e-02 -5.28420925e-01
6.57061696e-01 -8.16233039e-01 -7.65316188e-02 6.08535171e-01
-6.17459476e-01 -1.40828630e-02 9.36507642e-01 -6.90065682e-01
-1.44008493e+00 -1.31832659e+00 3.46888989e-01 -5.18010631e-02
3.31842810e-01 -1.10869780e-01 -8.31630170e-01 5.07986367e-01
1.11638844e-01 -2.15426043e-01 2.25993648e-01 -3.16774189e-01
2.70724714e-01 -6.63949430e-01 -1.14542055e+00 8.15422893e-01
8.52097809e-01 2.39416763e-01 -5.83112478e-01 4.43809154e-03
7.55640030e-01 -4.35593307e-01 -3.81607443e-01 8.93196046e-01
3.65721732e-01 -8.38508248e-01 6.46670878e-01 -4.66110766e-01
-4.81527776e-01 -8.60430598e-01 -7.48141333e-02 -1.39767933e+00
-2.20413551e-01 -6.82437778e-01 -5.09575270e-02 9.89761412e-01
5.15920281e-01 -6.80920482e-01 5.95128417e-01 6.05421960e-01
-3.36108774e-01 -6.52953684e-01 -1.07305372e+00 -9.33438897e-01
-1.19300112e-01 -7.38096297e-01 9.37647939e-01 4.32902455e-01
-9.26284492e-02 2.16871127e-01 -4.20673490e-01 3.97407562e-01
5.36383450e-01 -1.21285737e-01 1.11334646e+00 -1.10269749e+00
7.49708936e-02 -2.26496130e-01 -5.70516288e-01 -8.74016941e-01
4.45794821e-01 -4.97416943e-01 5.23840129e-01 -1.59705198e+00
-3.60656351e-01 -8.21229458e-01 -3.11455607e-01 2.27576241e-01
-1.16828360e-01 -7.26476252e-01 1.02875173e-01 -8.18163157e-02
-9.85441983e-01 9.36814427e-01 1.32510328e+00 -3.66958320e-01
-4.90044832e-01 6.12534761e-01 -2.40917251e-01 7.42671788e-01
1.26196337e+00 -2.65228599e-01 -8.98141384e-01 -6.75382763e-02
1.01776935e-01 3.38918120e-01 5.35422973e-02 -1.57179463e+00
6.01774693e-01 -7.96081185e-01 -3.66587862e-02 -1.06641424e+00
2.19928846e-01 -1.26252794e+00 1.83000252e-01 5.78732193e-01
-5.72993606e-02 2.44552210e-01 6.20628774e-01 8.02816093e-01
-2.65826344e-01 -3.25357020e-01 6.24484539e-01 -1.12675525e-01
-1.38659525e+00 4.41393316e-01 -9.04952645e-01 -7.83036649e-03
1.64048731e+00 -1.75385222e-01 1.05779484e-01 -5.77926874e-01
-7.25667953e-01 1.04252934e+00 -3.08019400e-01 7.08919644e-01
9.36716497e-01 -1.35313845e+00 -5.75082183e-01 1.17150329e-01
7.26541877e-02 1.35484070e-01 4.57313746e-01 6.72097921e-01
-2.77972370e-01 3.79865378e-01 -4.50882047e-01 -5.56812346e-01
-9.04381633e-01 7.77553618e-01 2.67507702e-01 -3.36120069e-01
-2.69892514e-01 4.46975455e-02 5.32386974e-02 -5.41541278e-01
3.85636806e-01 -2.74534017e-01 -6.12105548e-01 -1.68530241e-01
4.09747690e-01 6.59599066e-01 -1.65196005e-02 -6.13888741e-01
-1.00691587e-01 6.00686967e-01 6.92898929e-02 -1.83630437e-01
9.03614700e-01 -5.72669029e-01 4.93636012e-01 6.28350437e-01
5.23690283e-01 -3.28425646e-01 -1.29919052e+00 -9.41461399e-02
1.51766673e-01 -4.40264493e-01 -1.31298661e-01 -6.13337159e-01
-7.27347732e-01 7.89734960e-01 7.44782269e-01 -8.40698108e-02
9.48357463e-01 -5.18521726e-01 8.55124056e-01 7.81002939e-01
9.13865507e-01 -1.52058721e+00 -3.19917649e-02 6.83444738e-01
9.00946736e-01 -1.48934269e+00 -2.93990392e-02 -3.35564524e-01
-8.68417084e-01 1.07097435e+00 1.22168899e+00 2.23907456e-01
1.05600441e+00 4.39371839e-02 -4.42030355e-02 2.09750861e-01
-8.84335577e-01 -5.43623567e-01 -6.83015445e-03 7.07360864e-01
-4.79279637e-01 -2.39990074e-02 -4.92246687e-01 5.23576081e-01
3.33903618e-02 1.25671819e-01 6.18424654e-01 9.09008324e-01
-8.69621098e-01 -1.03714216e+00 -3.63876164e-01 1.30248785e-01
2.85272956e-01 6.41248763e-01 3.82652342e-01 1.04402363e+00
6.48122489e-01 1.42638206e+00 -3.43501002e-01 -5.69856703e-01
7.00785220e-01 -3.56340483e-02 1.08486764e-01 -3.12808901e-01
-6.73964769e-02 -3.63472909e-01 1.35195225e-01 -6.34222925e-01
-4.53252733e-01 -6.97442532e-01 -1.56724238e+00 -2.40899131e-01
-1.05208643e-01 3.07241350e-01 5.06585836e-01 1.11273777e+00
1.94689229e-01 6.10598624e-01 9.72526371e-01 -7.19869912e-01
-3.80282313e-01 -5.87060690e-01 -2.17114240e-01 7.27125779e-02
3.19692671e-01 -1.19675374e+00 -1.80900246e-01 -3.05260479e-01] | [5.2926740646362305, 1.346612811088562] |
214cdaf3-9dc8-4f20-8b4e-3bef463308d6 | satisfiability-and-containment-of-recursive | 2108.13063 | null | https://arxiv.org/abs/2108.13063v2 | https://arxiv.org/pdf/2108.13063v2.pdf | Satisfiability and Containment of Recursive SHACL | The Shapes Constraint Language (SHACL) is the recent W3C recommendation language for validating RDF data, by verifying certain shapes on graphs. Previous work has largely focused on the validation problem and the standard decision problems of satisfiability and containment, crucial for design and optimisation purposes, have only been investigated for simplified versions of SHACL. Moreover, the SHACL specification does not define the semantics of recursively-defined constraints, which led to several alternative recursive semantics being proposed in the literature. The interaction between these different semantics and important decision problems has not been investigated yet. In this article we provide a comprehensive study of the different features of SHACL, by providing a translation to a new first-order language, called SCL, that precisely captures the semantics of SHACL. We also present MSCL, a second-order extension of SCL, which allows us to define, in a single formal logic framework, the main recursive semantics of SHACL. Within this language we also provide an effective treatment of filter constraints which are often neglected in the related literature. Using this logic we provide a detailed map of (un)decidability and complexity results for the satisfiability and containment decision problems for different SHACL fragments. Notably, we prove that both problems are undecidable for the full language, but we present decidable combinations of interesting features, even in the face of recursion. | ['Fabio Mogavero', 'George Konstantinidis', 'Paolo Pareti'] | 2021-08-30 | null | null | null | null | ['formal-logic'] | ['reasoning'] | [ 1.57210365e-01 5.66643417e-01 -1.46374390e-01 -3.92331660e-01
-2.48279143e-02 -9.09789920e-01 4.91188049e-01 4.25956547e-01
8.24777409e-03 6.52755201e-01 4.09489870e-02 -5.95938742e-01
-7.10962951e-01 -1.25832748e+00 -6.59460723e-01 -2.01680765e-01
-2.99296498e-01 6.32558644e-01 8.26563179e-01 -4.15431410e-01
-2.16836885e-01 5.76947212e-01 -2.01271510e+00 3.50378126e-01
6.88275754e-01 1.11702073e+00 -2.90091693e-01 2.68887728e-01
-1.11082315e-01 6.45024180e-01 6.60599396e-02 -5.90380251e-01
1.61565691e-01 -2.08318368e-01 -1.17551017e+00 -6.58791466e-03
3.25184971e-01 1.51440501e-01 2.92960167e-01 1.14377904e+00
-1.28351077e-01 -2.89647192e-01 1.91382006e-01 -1.85696495e+00
-8.44871625e-02 9.64092314e-01 9.27475393e-02 -4.88522947e-01
8.81518424e-01 -2.66410440e-01 1.52519798e+00 -2.92278916e-01
9.66375709e-01 1.06415725e+00 6.11880302e-01 5.04492760e-01
-1.32233906e+00 -1.68372273e-01 1.07204489e-01 3.73545885e-01
-1.52920604e+00 -2.87801474e-01 4.37130868e-01 -3.70210499e-01
1.02857888e+00 9.71897066e-01 7.19607353e-01 3.89001995e-01
-1.33826569e-01 5.55315793e-01 1.18481040e+00 -6.97366714e-01
3.63589048e-01 4.30729806e-01 6.08985603e-01 4.82163072e-01
1.03801799e+00 2.73402222e-02 -2.40498692e-01 -3.59466404e-01
3.00759405e-01 -6.01483345e-01 -2.79890925e-01 -9.06372905e-01
-7.98856676e-01 8.39894295e-01 -2.00253665e-01 5.68482697e-01
3.61442775e-01 -1.12055413e-01 6.13967896e-01 5.40481925e-01
-2.80695558e-01 1.90356955e-01 -6.22675121e-01 2.04106241e-01
-6.26133859e-01 5.90941608e-01 1.58826411e+00 1.35860264e+00
6.01316690e-01 -2.83832878e-01 1.53475761e-01 5.53445667e-02
6.02567077e-01 5.63659549e-01 -5.20876467e-01 -9.14778948e-01
4.41150963e-02 8.63357306e-01 3.08460951e-01 -8.04896295e-01
-4.50525463e-01 -2.00946093e-01 -4.20592517e-01 1.85881332e-01
4.92970467e-01 3.05971503e-01 -6.10441677e-02 1.76387036e+00
3.12932044e-01 -2.50628471e-01 2.01728433e-01 6.31041110e-01
6.16986692e-01 1.54886842e-01 -1.20103285e-02 -5.38194180e-01
1.24377298e+00 -2.13358432e-01 -6.60542667e-01 3.92991275e-01
7.90578723e-01 -2.24100798e-01 6.48653507e-01 6.50213718e-01
-1.37450671e+00 1.49182320e-01 -1.26619947e+00 -4.12242748e-02
-6.87538028e-01 -4.22239721e-01 8.38006973e-01 1.00817811e+00
-1.13924682e+00 3.30876887e-01 -4.82518017e-01 -3.84839863e-01
1.88771281e-02 5.45220554e-01 -5.03060818e-01 -1.73198864e-01
-1.49981856e+00 1.01169813e+00 5.28288364e-01 -4.49424684e-02
-1.11727871e-01 -5.13199627e-01 -1.18940353e+00 1.13264591e-01
9.85851765e-01 -3.91474545e-01 1.12708700e+00 -6.60206437e-01
-1.24610543e+00 1.02944148e+00 -5.13624176e-02 -5.78956366e-01
6.54292703e-01 4.68999565e-01 -8.35878253e-01 -1.11353658e-01
-3.64758633e-02 -6.21683300e-02 -1.42234311e-01 -1.11885679e+00
-6.51921511e-01 -3.91892672e-01 8.49740207e-01 -4.39619213e-01
3.78129482e-01 3.50557357e-01 -3.37061465e-01 -1.46831274e-01
5.33382483e-02 -6.52971685e-01 -1.62690934e-02 -4.71423045e-02
-2.29518786e-01 -4.22858983e-01 3.33488286e-01 5.42957969e-02
1.71221817e+00 -1.79860258e+00 2.45327413e-01 9.01694417e-01
1.56401135e-02 2.12736815e-01 1.66709796e-01 7.84068227e-01
-1.46080539e-01 3.77557218e-01 -4.31959718e-01 2.94890255e-01
8.12751293e-01 5.86352468e-01 -3.21734816e-01 7.01106191e-01
1.74772426e-01 9.81576204e-01 -6.65626168e-01 -4.56949443e-01
2.46043488e-01 -7.16762198e-03 -7.56888807e-01 -2.96036541e-01
-8.59326601e-01 -1.65335447e-01 -2.18232006e-01 6.19743943e-01
1.12055933e+00 -1.01411723e-01 1.06205666e+00 -1.77338675e-01
-4.36552197e-01 1.16519526e-01 -1.92013907e+00 1.39324951e+00
-1.40485466e-01 -1.66376919e-01 2.44959563e-01 -8.88875961e-01
5.64703345e-01 2.30903566e-01 3.72401983e-01 -7.62106121e-01
-6.17663981e-03 5.04580617e-01 -1.31414145e-01 -6.00741267e-01
2.43539587e-01 -3.35138232e-01 -5.43366969e-01 6.05558097e-01
-3.25938195e-01 -4.81482223e-02 7.83764005e-01 8.66221935e-02
9.58316624e-01 2.62263089e-01 6.00775540e-01 -5.77230752e-01
1.06092346e+00 1.24924146e-01 7.29226828e-01 5.90264499e-01
2.96481371e-01 2.76137322e-01 9.62535024e-01 -3.14338565e-01
-8.07992697e-01 -8.75489295e-01 -5.81045568e-01 5.55031896e-01
2.57803977e-01 -1.14692736e+00 -3.82123858e-01 -6.06209040e-01
3.54369074e-01 7.42305934e-01 -4.84432429e-01 1.10573687e-01
-5.11376619e-01 -2.67821372e-01 7.04153836e-01 3.77698809e-01
3.69874649e-02 -7.20598996e-01 -8.39422524e-01 8.51761475e-02
1.50663853e-01 -1.36498713e+00 2.99389005e-01 -3.74208316e-02
-4.81446534e-01 -1.81553757e+00 4.49006885e-01 -3.40873599e-01
3.93128127e-01 -3.93570006e-01 1.25293791e+00 6.19730294e-01
9.82776377e-03 4.88438249e-01 -4.86147553e-01 -3.33374590e-01
-5.10004103e-01 -1.63071290e-01 -2.32493669e-01 -1.58378482e-01
4.41487879e-01 -3.46557766e-01 3.03067595e-01 5.18623471e-01
-1.48699284e+00 1.63626559e-02 -9.60718468e-02 2.66258836e-01
5.28841913e-01 2.75874674e-01 1.90568432e-01 -1.42749798e+00
3.55566144e-01 -1.96690738e-01 -1.18834567e+00 6.67806983e-01
-9.57331777e-01 4.55899835e-01 5.74955285e-01 3.64118397e-01
-5.26095212e-01 -1.04630724e-01 -1.89471871e-01 3.04142416e-01
-1.94341615e-01 9.91916180e-01 -8.39771211e-01 -1.04435161e-01
1.20638944e-01 -4.17392887e-02 -1.71343282e-01 -3.02252889e-01
1.91749096e-01 3.66476387e-01 3.66024226e-01 -1.18319440e+00
7.61404634e-01 6.70102775e-01 6.57232761e-01 -3.74655336e-01
-5.26920557e-01 -3.14004421e-01 -5.09976566e-01 5.74494787e-02
4.49693561e-01 -3.57966244e-01 -1.19448090e+00 7.19509423e-02
-1.02082086e+00 -1.67531624e-01 -7.12085366e-01 -1.72378004e-01
-7.73823202e-01 5.65448582e-01 -1.79446533e-01 -1.15337586e+00
1.03123210e-01 -1.13900292e+00 8.18461299e-01 -3.85390669e-01
-3.50311339e-01 -8.56579542e-01 3.60583812e-02 -1.75881669e-01
3.94513994e-01 5.16116381e-01 1.33263779e+00 -6.67414546e-01
-4.43556964e-01 -1.68918550e-01 -2.68047273e-01 1.67666510e-01
-3.96900982e-01 1.41330555e-01 -5.63050747e-01 -1.70987248e-01
-2.77368754e-01 7.31194839e-02 2.70627409e-01 -1.37001634e-01
7.53908873e-01 -2.41813466e-01 -1.65980130e-01 2.42715701e-01
1.93608737e+00 -2.43906021e-01 9.67287838e-01 3.71913642e-01
-9.30436328e-02 7.81842351e-01 6.64681017e-01 4.86510009e-01
6.08980775e-01 9.19895768e-01 5.91449022e-01 4.52860951e-01
2.22712234e-01 -5.37769087e-02 -4.05398756e-02 3.88464749e-01
-3.93807024e-01 -7.32407942e-02 -9.23865676e-01 2.61457652e-01
-2.06125116e+00 -1.03659439e+00 -7.59004712e-01 2.40730000e+00
7.91684806e-01 8.80260915e-02 2.76166290e-01 7.37731755e-01
4.50206935e-01 -8.52686539e-02 1.05521277e-01 -5.94863892e-01
-5.23501098e-01 6.56170130e-01 5.53833842e-01 8.98111045e-01
-8.58329415e-01 5.18085122e-01 6.58111191e+00 3.26285332e-01
-7.39614189e-01 9.74562392e-02 -4.99569118e-01 1.67937040e-01
-8.92547250e-01 5.33554435e-01 -7.25155950e-01 1.27493247e-01
9.70918298e-01 -4.38034832e-01 4.20628756e-01 5.53962052e-01
1.75989807e-01 -2.48805285e-01 -1.42527568e+00 5.82781911e-01
-4.52492898e-03 -1.33329642e+00 -1.74925216e-02 3.27997692e-02
4.10431772e-01 -4.15715247e-01 -6.14195645e-01 1.31238326e-01
1.67707115e-01 -1.02810538e+00 1.23349500e+00 5.33503234e-01
9.74091470e-01 -6.54527843e-01 8.51493597e-01 1.39892399e-01
-1.21952105e+00 -1.05659775e-01 -1.51699007e-01 -3.07830900e-01
2.21619561e-01 6.50267065e-01 -3.91862355e-02 1.23827708e+00
5.04526198e-01 6.01708770e-01 -3.27927798e-01 1.10564303e+00
-2.92175382e-01 1.66911468e-01 -5.86175859e-01 1.61238208e-01
-1.72785997e-01 -1.74637422e-01 5.59288681e-01 1.24998426e+00
1.34092808e-01 6.73249885e-02 -6.31524920e-02 8.41171622e-01
2.76862651e-01 -1.14107631e-01 -2.62761265e-01 1.48314342e-01
1.99345246e-01 6.37535214e-01 -4.74435717e-01 -1.39622509e-01
-6.41324461e-01 1.32316306e-01 2.35221833e-02 2.08379164e-01
-8.86953950e-01 -1.95339054e-01 5.04909694e-01 6.77820385e-01
3.60594809e-01 -1.50421411e-01 -2.57251799e-01 -1.20749593e+00
4.03895289e-01 -9.24696565e-01 7.89930165e-01 -5.08031309e-01
-1.10421658e+00 3.18956912e-01 5.97647965e-01 -9.53208387e-01
-2.20359862e-01 -8.87396932e-01 -1.24894932e-01 6.60740614e-01
-1.69487906e+00 -1.15275002e+00 -3.00051961e-02 6.68837786e-01
-3.44446480e-01 4.65665907e-01 1.13506877e+00 4.03230429e-01
-3.11704546e-01 3.18590313e-01 -4.68208760e-01 -3.40223014e-01
2.56018132e-01 -1.39138865e+00 -2.36697823e-01 8.51830721e-01
-3.92269701e-01 7.77537584e-01 1.05278027e+00 -5.97452044e-01
-1.96020019e+00 -8.54408860e-01 1.44006658e+00 -2.87223428e-01
8.49570334e-01 -4.23170298e-01 -6.83983266e-01 9.00883436e-01
-4.05497141e-02 2.06050515e-01 5.82724988e-01 5.31969629e-02
-6.30180478e-01 -2.85861194e-01 -1.31144822e+00 2.44232267e-01
1.27663672e+00 -5.77135563e-01 -5.23615718e-01 -7.35118520e-03
4.88166302e-01 -3.25540304e-01 -1.06072819e+00 7.30149984e-01
8.81191015e-01 -1.27925003e+00 6.66188359e-01 -7.48502553e-01
-1.20076716e-01 -9.04280305e-01 -5.87605774e-01 -3.68850321e-01
-1.32123813e-01 -4.71072495e-01 -3.47414583e-01 1.10858512e+00
4.92792994e-01 -8.86163354e-01 5.13857543e-01 1.08637691e+00
-3.14166360e-02 -7.31792927e-01 -1.04754102e+00 -8.91662776e-01
1.58239491e-02 -1.13354671e+00 1.10513878e+00 8.65801692e-01
5.94564438e-01 -1.23269320e-01 -3.28321047e-02 1.94099173e-01
3.57230186e-01 5.42620182e-01 4.36967283e-01 -1.63588047e+00
-3.65542799e-01 -5.00196397e-01 -7.36354351e-01 -3.11321944e-01
1.86665833e-01 -1.19581425e+00 -3.93397838e-01 -1.60435438e+00
-3.71047370e-02 -7.29744613e-01 1.59315526e-01 6.18458390e-01
8.57903838e-01 -1.73585583e-02 1.90330446e-01 -1.58137217e-01
-9.06388223e-01 -1.31358147e-01 8.29352558e-01 -2.13280156e-01
2.36297160e-01 1.08408496e-01 -7.02373981e-01 4.89472985e-01
3.30315202e-01 -2.08363056e-01 -3.76147270e-01 -1.98184237e-01
1.34652793e+00 -4.59523834e-02 4.23793316e-01 -8.79958689e-01
3.11138868e-01 -6.77634835e-01 -5.76564848e-01 -3.54539186e-01
-2.41010249e-01 -1.40672851e+00 1.04612863e+00 3.84272575e-01
-2.12649956e-01 -1.32684603e-01 1.59900725e-01 1.50583282e-01
-1.66810706e-01 -7.35810697e-01 4.34785098e-01 -8.11997708e-03
-7.59284556e-01 1.27267987e-01 -1.41281173e-01 6.44794181e-02
1.31727374e+00 -2.14699954e-01 -2.28528336e-01 -8.05285200e-02
-1.01219249e+00 3.38076502e-01 9.24429655e-01 1.62718892e-02
3.19817543e-01 -1.17593658e+00 -5.22924364e-01 2.47255132e-01
3.96975905e-01 9.63736698e-02 -1.03907689e-01 1.03837895e+00
-6.35086358e-01 6.90296352e-01 4.98154573e-02 -2.58557737e-01
-1.12949538e+00 6.44780397e-01 2.47734010e-01 -4.03895438e-01
-3.37426901e-01 -9.60087329e-02 -3.56681257e-01 -2.78397322e-01
1.66083395e-01 -7.15481699e-01 -1.94402203e-01 -9.13895294e-02
5.04548848e-01 3.18520933e-01 4.47845906e-01 -6.68480337e-01
-7.85449147e-01 5.64998567e-01 5.73246896e-01 5.73160872e-02
1.44400692e+00 -1.21530920e-01 -7.92923987e-01 1.76811576e-01
5.54917455e-01 4.85165238e-01 -1.68737635e-01 -5.37546091e-02
4.48670924e-01 -7.79235289e-02 -4.25488859e-01 -8.76551211e-01
-7.60090053e-01 2.83928454e-01 2.31223758e-02 6.89391792e-01
1.15074861e+00 3.26777041e-01 3.30783397e-01 2.29980960e-01
8.81498456e-01 -9.04370248e-01 -1.23601961e+00 5.74371219e-01
9.56056654e-01 -5.50797403e-01 1.92546025e-01 -1.17268753e+00
-1.44981593e-01 1.34863651e+00 1.90436840e-01 2.14818940e-01
5.39837956e-01 9.49607193e-01 -5.81784129e-01 -2.47219414e-01
-6.46737516e-01 -5.72088122e-01 -1.80419590e-02 6.06729805e-01
2.85516500e-01 2.06313804e-01 -9.71554399e-01 1.11227620e+00
-1.64136916e-01 5.17991662e-01 7.59870350e-01 1.26913321e+00
-3.09340060e-01 -1.89265645e+00 -2.84589857e-01 2.24861845e-01
-3.51771086e-01 1.49644971e-01 -4.01212215e-01 1.06591368e+00
5.04252553e-01 9.30701077e-01 -2.38607273e-01 -2.17353895e-01
6.45466208e-01 2.14379523e-02 9.67787147e-01 -5.98921061e-01
-5.10057449e-01 -4.17660177e-01 7.37254918e-01 -5.11473715e-01
-8.65154803e-01 -7.11269379e-01 -1.50605881e+00 -4.95237380e-01
-4.05204743e-01 5.14839590e-01 3.38538349e-01 1.07337224e+00
-1.12583973e-01 6.16537817e-02 -3.10599362e-03 5.11852652e-02
-4.13584948e-01 -1.86686471e-01 -9.66406584e-01 3.68691683e-01
-5.42374887e-02 -5.83034456e-01 -2.24024147e-01 -2.33913645e-01] | [8.662885665893555, 6.802210330963135] |
c273f8e5-1882-4c86-9ff1-e9026357bc4d | a-mathematical-theory-of-super-resolution-and | 2211.15208 | null | https://arxiv.org/abs/2211.15208v2 | https://arxiv.org/pdf/2211.15208v2.pdf | A mathematical theory of super-resolution and diffraction limit | This paper is devoted to elucidating the essence of super-resolution and deals mainly with the stability of super-resolution and the diffraction limit with respect to the signal-to-noise ratio. The goal is to determine the number, locations, and amplitudes of point sources from noisy Fourier data. The first contribution in this paper is two location-amplitude identities characterizing the relations between source locations and amplitudes in the super-resolution problem. These identities allow us to directly derive the super-resolution capability for number, location, and amplitude recovery in the super-resolution problem and improve state-of-the-art estimations to an unprecedented level to have practical significance. The nonlinear inverse problems studied in this paper are known to be very challenging and have only been partially solved in recent years. However, thanks to this paper, we now have a clear and simple picture of all of these problems. The second crucial result of this paper is the theoretical proof of a two-point diffraction limit in spaces of general dimensionality under only an assumption on the noise level. The two-point diffraction limit is given by \[ R = \frac{4\arcsin \left(\left(\frac{\sigma}{m_{\min}}\right)^{\frac{1}{2}} \right)}{\Omega} \] for $\frac{\sigma}{m_{\min}}\leq\frac{1}{2}$, where $\frac{\sigma}{m_{\min}}$ represents the inverse of the signal-to-noise ratio ($SNR$) and $\Omega$ is the cutoff frequency. This solves the long-standing puzzle and debate about the diffraction limit for imaging (and line spectral estimation) in general setting. Our results also show that, for the resolution of any two point sources, when $SNR>2$, one can definitely exceed the Rayleigh limit $\frac{\pi}{\Omega}$. We also find the optimal algorithm that achieves the optimal resolution when distinguishing two sources. | ['Habib Ammari', 'Ping Liu'] | 2022-11-28 | null | null | null | null | ['common-sense-reasoning'] | ['reasoning'] | [ 6.05985641e-01 1.89063340e-01 4.31330055e-01 1.84558943e-01
-1.00703442e+00 -3.55209470e-01 2.98231822e-02 -3.11242521e-01
-5.35714149e-01 8.78906667e-01 -2.02099845e-01 -2.39444911e-01
-9.96236444e-01 -6.86570585e-01 -3.75113785e-01 -1.24851823e+00
-4.71151501e-01 6.52426109e-02 2.12552682e-01 -2.71609217e-01
1.37946323e-01 5.54881632e-01 -1.78378272e+00 -2.83178121e-01
7.97826707e-01 1.35854256e+00 3.32170635e-01 7.27554619e-01
2.28802174e-01 3.62830788e-01 -6.60119891e-01 4.40715961e-02
2.43236527e-01 -8.94877672e-01 -3.85110229e-01 -1.82440951e-01
-4.80251163e-02 -2.24750280e-01 -6.92485794e-02 1.41951776e+00
5.66862881e-01 1.05022743e-01 7.72712648e-01 -5.77059090e-01
-2.22843036e-01 2.45763078e-01 -8.60808551e-01 6.16565228e-01
2.21793815e-01 -2.27634281e-01 5.60157120e-01 -7.14997828e-01
2.68601418e-01 4.54237312e-01 5.10466337e-01 1.99104011e-01
-8.44428420e-01 -7.38158703e-01 -4.68044579e-01 -3.82524543e-02
-1.59591234e+00 -2.90053457e-01 5.59799671e-01 -4.05932546e-01
2.81010270e-01 5.41003704e-01 5.69433384e-02 3.28102678e-01
-1.37043670e-01 -2.25417137e-01 1.27282882e+00 -7.90935159e-01
1.80656061e-01 7.64660537e-02 -2.12389547e-02 6.64014578e-01
3.66698384e-01 1.78334922e-01 -4.42251027e-01 6.44620582e-02
1.44647765e+00 -5.89290082e-01 -7.38995731e-01 2.23355979e-01
-1.05860949e+00 7.44177103e-01 4.85189781e-02 9.44622457e-01
-2.44296640e-01 -1.22570954e-02 -3.45552444e-01 1.83021113e-01
3.35854053e-01 5.20873666e-01 2.44463142e-02 -1.09318607e-01
-8.75509381e-01 1.70510128e-01 5.93634963e-01 8.24454486e-01
5.89610398e-01 2.00101063e-01 5.75740226e-02 7.25701034e-01
-8.00969005e-02 8.97547722e-01 -6.20126799e-02 -1.42690599e+00
3.04639131e-01 -1.16157018e-01 4.66547698e-01 -8.31296802e-01
-3.95530343e-01 -1.01878130e+00 -9.84968841e-01 3.67579997e-01
9.50217545e-01 -3.76242518e-01 -1.69887081e-01 1.71797287e+00
3.37935146e-03 2.61021815e-02 1.04084887e-01 9.87823009e-01
4.42085773e-01 6.29262269e-01 -4.91338342e-01 -1.06777024e+00
1.48710442e+00 1.23307459e-01 -5.15970111e-01 -5.95802963e-02
2.04493508e-01 -8.49410117e-01 7.07924306e-01 5.56404054e-01
-1.31393468e+00 -3.92255485e-01 -1.12619185e+00 5.16771257e-01
4.24767792e-01 -1.42033637e-01 7.58427531e-02 6.98940873e-01
-8.61151814e-01 6.27161443e-01 -4.55292314e-01 1.87454090e-01
-1.22524239e-02 1.42360538e-01 -3.93948928e-02 -8.61066952e-02
-9.55279946e-01 6.81258917e-01 -1.31335214e-01 5.89119270e-02
-1.04786597e-01 -8.01465690e-01 -3.11665505e-01 1.36752889e-01
5.04374087e-01 -2.15792984e-01 8.72810006e-01 -4.94804800e-01
-1.22968638e+00 5.60891986e-01 -1.64821863e-01 -1.85023233e-01
3.01033944e-01 6.25679418e-02 -5.33346295e-01 7.31266856e-01
3.32346678e-01 -2.59791106e-01 5.60058475e-01 -1.35737252e+00
-5.05416155e-01 -6.89619124e-01 -2.47814119e-01 -8.75853822e-02
4.81855124e-02 2.90944070e-01 -1.63201898e-01 -4.72466081e-01
6.56290889e-01 -5.50630331e-01 -6.55036867e-02 -4.03530568e-01
-1.33991092e-01 1.83262482e-01 5.00935793e-01 -4.84579951e-01
1.12092328e+00 -2.07338452e+00 6.89479038e-02 3.80106777e-01
2.18693987e-01 2.00844646e-01 3.73940080e-01 3.67312372e-01
-2.47000039e-01 3.12144365e-02 -5.15519142e-01 2.38567173e-01
-4.49932218e-01 -2.76576310e-01 -1.20446391e-01 7.65857339e-01
-2.56413788e-01 1.28830567e-01 -6.27173483e-01 -2.39196971e-01
8.58818516e-02 6.74766541e-01 -1.06842831e-01 -1.53778538e-01
1.72884583e-01 8.37865770e-01 -6.65287435e-01 2.45799556e-01
8.88231695e-01 -3.14864069e-01 -6.11487217e-02 -2.49461621e-01
-7.81244159e-01 -2.73807138e-01 -1.76930726e+00 1.00051022e+00
-1.36670604e-01 3.39599878e-01 6.78499997e-01 -1.22382021e+00
8.21421206e-01 3.69867563e-01 7.74282336e-01 -1.02127922e+00
1.20740548e-01 4.30109411e-01 -1.58226967e-01 -5.01143754e-01
1.84502304e-02 -8.21910918e-01 4.60033976e-02 5.03896594e-01
-2.37223312e-01 -1.86343804e-01 1.61222994e-01 -2.20241427e-01
1.07996964e+00 -1.85319573e-01 2.62754530e-01 -5.14834344e-01
5.22602379e-01 -2.65424430e-01 3.83777708e-01 7.58173406e-01
4.38180566e-02 6.12784982e-01 8.28574955e-01 1.01573199e-01
-1.04799163e+00 -9.58680391e-01 -5.93312025e-01 5.68069696e-01
2.34409347e-01 -4.28288169e-02 -8.85201454e-01 2.88262755e-01
-4.71064955e-01 7.04842389e-01 -2.90796131e-01 3.46464038e-01
-8.10664892e-01 -1.06968272e+00 5.56399107e-01 2.37490147e-01
7.38599837e-01 -6.89974785e-01 -1.06583941e+00 -1.20681673e-01
-4.26036417e-01 -1.37420356e+00 1.55681640e-01 1.99052975e-01
-5.32583952e-01 -1.11601841e+00 -9.25674796e-01 -2.52269477e-01
5.92279851e-01 2.50783861e-01 6.78355396e-01 -1.19690947e-01
-1.59616128e-01 3.78841132e-01 -3.02489012e-01 -3.00338775e-01
-2.85739690e-01 -4.96494055e-01 2.27372795e-01 -1.34250373e-01
-1.00360140e-01 -1.02005446e+00 -8.12769234e-01 5.44150174e-01
-9.43792284e-01 -3.55658025e-01 4.62737083e-01 4.53075677e-01
6.38809919e-01 4.61288452e-01 6.65991545e-01 -2.17968717e-01
3.69471014e-01 -1.52994052e-01 -9.99833226e-01 -2.27276921e-01
-3.05188268e-01 1.56822503e-01 7.84880698e-01 -1.02655077e-02
-8.05895269e-01 -3.33369076e-01 -1.17274664e-01 -1.34154320e-01
-2.35410646e-01 3.09799522e-01 -1.86221153e-02 -1.52265832e-01
7.65716434e-01 4.48964655e-01 -9.09948125e-02 -6.45776868e-01
8.58907402e-02 3.97354990e-01 9.60870147e-01 -3.63256991e-01
9.24408257e-01 6.59434080e-01 5.83782315e-01 -1.32856750e+00
-8.79432976e-01 -2.69770414e-01 -3.43994200e-01 -6.93882555e-02
7.26176500e-01 -5.13323247e-01 -1.21700990e+00 -2.01981864e-03
-7.35285342e-01 3.37448567e-02 -4.70886439e-01 8.47559571e-01
-7.49706089e-01 7.26014793e-01 -2.96910435e-01 -1.31100738e+00
-6.67626336e-02 -8.20349336e-01 6.23065412e-01 3.70324790e-01
8.75792056e-02 -6.69966578e-01 -3.05190861e-01 3.90215844e-01
6.04624212e-01 4.80130166e-01 9.06644821e-01 5.72054163e-02
-6.02646112e-01 -1.27165779e-01 -4.64671224e-01 3.90682191e-01
-1.20505735e-01 -5.84938228e-01 -6.78735793e-01 -1.83318183e-01
8.33546877e-01 2.93493330e-01 6.61214888e-01 6.98129237e-01
7.99552143e-01 -2.42693990e-01 -2.12892473e-01 5.09258628e-01
1.72352326e+00 1.81020692e-01 7.44781017e-01 3.06092836e-02
-1.70732215e-02 5.70767522e-01 3.15077782e-01 6.51008248e-01
-3.24161917e-01 8.00570428e-01 5.06120443e-01 1.87600836e-01
-2.13730261e-01 4.88459319e-01 -1.51474625e-02 3.41244817e-01
-5.41478455e-01 -9.32136253e-02 -5.89102507e-01 4.38337624e-01
-1.12119853e+00 -1.09828389e+00 -4.56083447e-01 2.62623477e+00
5.46948671e-01 1.60765678e-01 8.90190452e-02 4.85780209e-01
7.79551625e-01 1.38043966e-02 -8.83336589e-02 5.08875698e-02
-4.15003926e-01 8.02111804e-01 6.76039279e-01 6.28259063e-01
-6.30073428e-01 7.07602575e-02 5.06854630e+00 9.19752240e-01
-1.07203782e+00 1.51898772e-01 2.54835665e-01 -2.36080244e-01
-2.57377267e-01 -6.42684773e-02 -7.50090420e-01 7.55065560e-01
8.48168552e-01 -1.32960796e-01 3.56612027e-01 2.26300985e-01
2.04359382e-01 -7.41157591e-01 -6.68168128e-01 9.80943620e-01
3.82009186e-02 -1.07609522e+00 -5.13041615e-01 2.94926435e-01
5.45256615e-01 -4.52276289e-01 2.69637972e-01 -2.60810196e-01
-1.34276748e-01 -1.12585056e+00 3.61287147e-01 4.31098968e-01
1.38673031e+00 -7.58834124e-01 5.64349294e-01 8.44456017e-01
-1.11618388e+00 -1.34947166e-01 -1.56028077e-01 -5.24433255e-02
4.34915334e-01 1.09472036e+00 -2.66615301e-01 9.36468780e-01
7.35242724e-01 -2.78279305e-01 3.29471201e-01 7.97580063e-01
-1.50399461e-01 3.81236941e-01 -6.99942887e-01 2.79844165e-01
8.94768536e-02 -4.05982792e-01 9.98497546e-01 8.59543383e-01
8.12116802e-01 1.00615275e+00 -2.40034506e-01 1.07381165e+00
1.71970621e-01 -1.89653203e-01 -1.17620870e-01 3.83622259e-01
3.92579108e-01 8.40200782e-01 -9.16569054e-01 3.39245170e-01
-2.87988663e-01 3.66768181e-01 -2.77186483e-01 6.00165725e-01
-5.97722888e-01 -6.63722932e-01 4.40883875e-01 7.85452247e-01
5.32958627e-01 -5.53006649e-01 -4.05844092e-01 -5.98617792e-01
3.86453807e-01 -3.22745353e-01 1.36957154e-01 -5.38235664e-01
-8.26712906e-01 3.76445830e-01 5.54745197e-01 -1.19385099e+00
-3.60117078e-01 -4.67248350e-01 -2.30025828e-01 1.25628114e+00
-1.35025120e+00 -3.73275787e-01 -1.65090352e-01 5.02731860e-01
-1.11120410e-01 1.96016550e-01 8.11280429e-01 2.81076819e-01
-2.55806763e-02 1.87379122e-01 5.38777530e-01 2.06716493e-01
7.25030079e-02 -7.30846882e-01 -4.17480737e-01 9.50757444e-01
-4.10531640e-01 3.59505415e-01 1.21641779e+00 -3.81399840e-01
-1.24760449e+00 -4.43509132e-01 6.71470106e-01 -1.86308190e-01
4.01219308e-01 -9.86812785e-02 -7.30546117e-01 1.79923683e-01
-1.14656240e-01 1.43336490e-01 5.09879827e-01 -4.60876077e-01
2.34317686e-02 -1.62200466e-01 -1.32205009e+00 1.76620021e-01
7.71002293e-01 -2.04043418e-01 -1.91263735e-01 9.85327512e-02
2.60357797e-01 -4.15053695e-01 -9.99282002e-01 6.12004817e-01
4.00529027e-01 -1.43839860e+00 9.06009674e-01 6.23057894e-02
2.13802114e-01 -2.49428287e-01 -4.86720920e-01 -7.67934442e-01
-1.57608300e-01 -7.33948052e-01 1.56914592e-01 7.84204125e-01
2.90668607e-01 -7.25250125e-01 5.52311361e-01 2.40262941e-01
5.18532619e-02 -6.79810286e-01 -1.69804740e+00 -7.95992315e-01
3.00054342e-01 -3.29394102e-01 -1.19945258e-01 7.32514262e-01
1.21694589e-02 -2.22834554e-02 -1.04821794e-01 6.62231147e-01
1.09243155e+00 -4.03528325e-02 2.06294246e-02 -1.30322373e+00
-6.19803429e-01 -3.11553895e-01 -1.24206148e-01 -1.06805372e+00
-5.24142027e-01 -2.65799224e-01 -2.70681679e-01 -1.44835329e+00
2.01257747e-02 -6.50705338e-01 -1.10433765e-01 -2.03767136e-01
3.26592445e-01 2.69429207e-01 -3.55324373e-02 2.66057074e-01
-9.33399945e-02 -6.68168589e-02 1.16732669e+00 4.43907768e-01
-1.53759733e-01 3.83821696e-01 -7.74282217e-01 7.63329208e-01
4.90921557e-01 -2.25637570e-01 -2.80766904e-01 -2.93276578e-01
6.18643105e-01 8.61695468e-01 5.80423832e-01 -1.14283264e+00
2.55958974e-01 -8.10343102e-02 1.88909277e-01 -4.07245308e-01
6.05758131e-01 -5.52912593e-01 3.35577101e-01 2.29702611e-02
-3.49101536e-02 -3.37385923e-01 -8.21360648e-02 3.94379705e-01
1.62461117e-01 -5.06984234e-01 1.16107678e+00 -4.24790293e-01
-3.59714106e-02 -3.54672849e-01 -2.80691832e-01 3.16827521e-02
9.26341951e-01 -2.83534199e-01 -4.42854464e-01 -7.53363311e-01
-8.46281886e-01 -1.90291256e-01 1.69293135e-01 -3.03262800e-01
2.74160981e-01 -8.51532757e-01 -1.02652943e+00 2.48805463e-01
-3.98760498e-01 3.60275060e-02 5.93474388e-01 1.16000581e+00
-3.75464708e-01 4.70685154e-01 7.36886933e-02 -5.93180478e-01
-9.11941409e-01 3.38948458e-01 6.73147500e-01 -1.66729212e-01
-4.20242906e-01 9.54834938e-01 -1.14413060e-01 5.80944419e-01
8.34595263e-02 -4.97888327e-02 -1.32614095e-02 -1.24996774e-01
6.92919552e-01 8.38474452e-01 1.29601523e-01 -7.65477002e-01
-1.59855023e-01 1.00939274e+00 4.59776849e-01 -4.29895788e-01
1.13557065e+00 -3.20419192e-01 -1.78338706e-01 2.78554976e-01
1.12440085e+00 3.26441348e-01 -1.04908121e+00 -1.62429810e-01
-3.98954183e-01 -2.57832348e-01 -1.30911902e-01 -6.88533902e-01
-7.33994365e-01 8.19012880e-01 5.88501871e-01 5.96013010e-01
1.18460751e+00 4.39274728e-01 4.22546029e-01 -1.35449827e-01
4.39861298e-01 -1.01537669e+00 3.58336195e-02 3.41723621e-01
7.70691454e-01 -5.07275879e-01 2.17210308e-01 -7.07911611e-01
-1.81779191e-02 9.89675820e-01 1.37684662e-02 -7.19130561e-02
6.97383285e-01 4.23260421e-01 -3.06365967e-01 -1.60444900e-01
1.26336321e-01 -1.72189787e-01 -4.31851819e-02 4.99511868e-01
4.76129711e-01 -5.56419082e-02 -5.36256731e-01 7.35589504e-01
-4.30823743e-01 -1.93950851e-02 5.87164819e-01 7.09886849e-01
-1.08455110e+00 -8.34457457e-01 -7.77463317e-01 1.33001000e-01
-8.57794046e-01 1.87251374e-01 2.17078835e-01 7.42060542e-01
4.16443497e-01 1.19532907e+00 -7.59249777e-02 2.73797423e-01
4.50140357e-01 -2.53031570e-02 7.58720756e-01 -1.83762103e-01
6.04553558e-02 4.19308424e-01 -1.98280197e-02 -2.13408336e-01
-5.06084919e-01 -5.23220003e-01 -1.52121472e+00 -1.44572496e-01
-1.88038334e-01 6.15513742e-01 5.41263402e-01 9.10071313e-01
-4.69439216e-02 3.77100736e-01 5.60832560e-01 -6.07253969e-01
-6.12461209e-01 -8.05363119e-01 -1.12590277e+00 -1.48454338e-01
4.31463391e-01 -5.63950717e-01 -7.35103369e-01 -1.73159748e-01] | [6.483432292938232, 4.497196674346924] |
2e73a25a-6bc3-457a-9e9c-dce5636aa11f | uniform-subdivision-of-omnidirectional-camera | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Kang_Uniform_Subdivision_of_Omnidirectional_Camera_Space_for_Efficient_Spherical_Stereo_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Kang_Uniform_Subdivision_of_Omnidirectional_Camera_Space_for_Efficient_Spherical_Stereo_CVPR_2022_paper.pdf | Uniform Subdivision of Omnidirectional Camera Space for Efficient Spherical Stereo Matching | Omnidirectional cameras have been used widely to better understand surrounding environments. They are often configured as stereo to estimate depth. However, due to the optics of the fisheye lens, conventional epipolar geometry is inapplicable directly to omnidirectional camera images. Intermediate formats of omnidirectional images, such as equirectangular images, have been used. However, stereo matching performance on these image formats has been lower than the conventional stereo due to severe image distortion near pole regions. In this paper, to address the distortion problem of omnidirectional images, we devise a novel subdivision scheme of a spherical geodesic grid. This enables more isotropic patch sampling of spherical image information in the omnidirectional camera space. Our spherical geodesic grid is tessellated with an equal-arc subdivision, making the cell sizes and in-between distances as uniform as possible, i.e., the arc length of the spherical grid cell's edges is well regularized. Also, our uniformly tessellated coordinates in a 2D image can be transformed into spherical coordinates via one-to-one mapping, allowing for analytical forward/backward transformation. Our uniform tessellation scheme achieves a higher accuracy of stereo matching than the traditional cylindrical and cubemap-based approaches, reducing the memory footage required for stereo matching by 20 %. | ['Min H. Kim', 'Chong-Min Kyung', 'Jungeon Lee', 'Hyeonjoong Jang', 'Donghun Kang'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['stereo-matching-1'] | ['computer-vision'] | [-1.50301168e-02 -2.21863195e-01 2.73316741e-01 -2.08499193e-01
-1.83581591e-01 -8.20540547e-01 5.42538047e-01 -4.91723031e-01
-4.57479030e-01 6.27345324e-01 1.70486629e-01 -4.24155504e-01
6.95529357e-02 -1.06193006e+00 -6.19102120e-01 -6.66728556e-01
4.16244745e-01 1.91746265e-01 3.94737959e-01 5.52449934e-02
5.66089869e-01 6.13387585e-01 -1.64515197e+00 -2.70688802e-01
9.80948448e-01 8.46265495e-01 4.63903069e-01 5.97937942e-01
1.55886963e-01 9.71300602e-02 -2.06122667e-01 -4.29512888e-01
4.91084009e-01 3.94255966e-02 -3.51851255e-01 8.98431838e-02
4.35268581e-01 -5.58943450e-01 -6.24757946e-01 1.13158822e+00
3.27710539e-01 -2.15292517e-02 4.92698789e-01 -9.31288302e-01
-4.82108235e-01 -3.12761068e-01 -7.61495590e-01 -1.82606086e-01
6.57031953e-01 -3.44344735e-01 4.50491637e-01 -9.58629012e-01
5.31534314e-01 8.99576306e-01 6.41174555e-01 2.18463004e-01
-8.08282018e-01 -5.85848808e-01 -2.07872599e-01 1.31997630e-01
-1.86708355e+00 -3.01119477e-01 5.87224662e-01 -2.26003468e-01
5.62596440e-01 6.48602962e-01 8.22029114e-01 3.48608077e-01
3.85542482e-01 1.83144912e-01 1.00941265e+00 -4.98351246e-01
1.83721662e-01 8.26176032e-02 -4.37198222e-01 4.14443374e-01
3.56236875e-01 1.90483481e-02 -1.21110067e-01 -8.95280316e-02
1.41788435e+00 2.38003373e-01 -5.81771612e-01 -8.07810307e-01
-1.19357097e+00 5.94393790e-01 4.40442294e-01 3.68075594e-02
8.11748728e-02 -1.26535386e-01 2.89120097e-02 -2.67418218e-03
3.09341311e-01 4.73274529e-01 1.15128933e-02 -1.25692353e-01
-5.43493330e-01 7.93869570e-02 4.68828231e-01 1.35531521e+00
8.73848319e-01 -1.03536546e-01 5.97107530e-01 9.95873451e-01
8.91919583e-02 5.92935741e-01 3.00240546e-01 -1.43015277e+00
4.43585128e-01 4.75991160e-01 2.78436214e-01 -1.31951475e+00
-3.76361012e-01 -2.11128801e-01 -9.01276588e-01 3.27404976e-01
4.47620332e-01 4.77568284e-02 -4.15648878e-01 1.11895597e+00
4.55262959e-01 1.05455788e-02 1.22956354e-02 1.02817345e+00
5.00609219e-01 6.85916841e-01 -8.47641349e-01 3.83888232e-03
1.22491336e+00 -6.56379104e-01 -2.50988632e-01 1.36199025e-02
4.32373524e-01 -9.39078033e-01 1.05201542e+00 4.25718009e-01
-1.02211750e+00 -1.73745677e-01 -8.78657460e-01 -1.96629390e-01
-1.21087335e-01 -1.53672770e-01 4.33591276e-01 6.72489762e-01
-1.18398333e+00 -1.26801029e-01 -4.59947079e-01 -1.70858905e-01
-1.74586236e-01 2.13746563e-01 -5.43376207e-01 -3.03247362e-01
-7.83380806e-01 6.44267857e-01 -1.13735490e-01 -4.26720046e-02
-1.66471332e-01 -7.06583917e-01 -1.12128472e+00 -4.99549508e-02
7.21953288e-02 -7.23198652e-01 9.87829745e-01 -4.02984798e-01
-1.48969841e+00 9.14640665e-01 -5.79663455e-01 2.03274796e-03
5.48207939e-01 3.08754265e-01 -1.24863304e-01 1.56252742e-01
2.34151542e-01 6.07365608e-01 3.15250814e-01 -1.34326696e+00
-7.48537838e-01 -6.54714406e-01 4.91626859e-01 8.80639553e-01
-2.11844847e-01 -1.52491078e-01 -8.02688479e-01 -3.43493730e-01
6.62269413e-01 -9.14086521e-01 -2.82120019e-01 2.62257665e-01
-1.74780995e-01 2.48104900e-01 8.41683924e-01 -3.23615819e-01
1.16431475e+00 -2.33379674e+00 -3.24887067e-01 2.29353815e-01
9.04610083e-02 -2.72365212e-01 2.95164913e-01 2.59604961e-01
1.69738472e-01 -1.69270873e-01 1.58712476e-01 -3.25563908e-01
-4.13755924e-01 2.67003685e-01 -2.86792181e-02 8.09420705e-01
-8.09016585e-01 4.22795445e-01 -8.78920019e-01 -3.79451275e-01
5.11926830e-01 4.76489455e-01 -7.66032577e-01 1.12482429e-01
5.83863676e-01 4.43978965e-01 -2.25387082e-01 5.89505434e-01
1.21050322e+00 1.43400282e-01 -1.82354003e-01 -8.76077041e-02
-8.76574874e-01 2.13063359e-01 -1.20794213e+00 1.44010532e+00
-8.28288794e-01 6.49682462e-01 1.31519273e-01 -5.61062574e-01
1.11354995e+00 1.17151283e-01 4.24441725e-01 -9.76037920e-01
-2.45798394e-01 2.96318203e-01 -5.03795624e-01 -5.18310666e-02
7.42461741e-01 1.63627550e-01 2.34595723e-02 1.51622749e-03
-7.99288213e-01 -7.02998161e-01 -1.84961278e-02 -2.09029227e-01
5.26550889e-01 -2.45616421e-01 4.07428861e-01 -4.42589045e-01
5.05767345e-01 2.99157016e-02 6.85988009e-01 3.61904293e-01
4.03572649e-01 1.33218122e+00 6.63414747e-02 -7.36493051e-01
-1.44924605e+00 -1.04807925e+00 -6.88440204e-01 1.77904874e-01
9.92340803e-01 -2.70854652e-01 -8.66443694e-01 1.97173670e-01
-1.54738396e-01 4.50528920e-01 -2.66106337e-01 2.31597155e-01
-6.13964438e-01 -3.93376857e-01 2.44525194e-01 3.94600779e-01
8.89132202e-01 -2.48138890e-01 -7.93514550e-01 -1.00559942e-01
-2.30910555e-01 -9.72442985e-01 -9.04160559e-01 -4.07895923e-01
-8.68395746e-01 -1.17109537e+00 -1.02220058e+00 -8.24354589e-01
1.07878268e+00 1.25473356e+00 6.99134111e-01 -3.21056664e-01
-4.00719699e-03 3.82642984e-01 -1.18736565e-01 2.54416559e-03
2.23386630e-01 -3.72996658e-01 1.68434381e-01 -6.66874424e-02
3.15297395e-01 -5.61092913e-01 -1.17926109e+00 1.06571281e+00
-6.97953939e-01 4.48119134e-01 -2.50424687e-02 6.36776090e-01
7.40286827e-01 8.24186057e-02 -3.14855039e-01 -2.88076937e-01
1.06879696e-01 -3.85425746e-01 -1.10443854e+00 -1.57931790e-01
-3.52663517e-01 -4.95002061e-01 8.03198934e-01 7.75655135e-05
-1.03691256e+00 -1.07976727e-01 -3.20470482e-02 -2.16220945e-01
7.55024552e-02 1.99070990e-01 -2.45978504e-01 -5.65923929e-01
3.06021422e-01 4.51912433e-01 -3.85006331e-02 -2.78674513e-01
-3.50490920e-02 1.04365706e+00 7.62273967e-01 -1.74318552e-01
5.99553168e-01 9.62849855e-01 6.77903369e-02 -1.29843116e+00
-1.29988670e-01 -6.33505881e-01 -4.94384289e-01 -3.94288003e-01
7.83396900e-01 -1.02627754e+00 -8.54515195e-01 6.71637475e-01
-1.23393106e+00 9.69515461e-03 1.60165504e-02 6.91818655e-01
-6.46557152e-01 5.84407985e-01 -3.16614091e-01 -6.48716331e-01
1.35888562e-01 -1.23243070e+00 1.07986557e+00 5.43781996e-01
-9.98861417e-02 -1.04695106e+00 9.37004089e-02 3.25552911e-01
3.73561770e-01 -5.88374771e-02 5.89092314e-01 4.10544127e-01
-8.17027748e-01 -6.36310503e-02 -3.78225565e-01 -4.14548665e-02
2.67563969e-01 -1.20182626e-01 -7.40302563e-01 -2.40623504e-01
2.75148213e-01 2.72884876e-01 1.71702787e-01 6.31635845e-01
1.29055464e+00 -1.51414678e-01 -3.79099250e-01 1.29073977e+00
1.40546060e+00 7.31148183e-01 1.03637183e+00 8.07141244e-01
5.40675879e-01 3.60481083e-01 7.44500041e-01 5.05852818e-01
8.20643067e-01 1.00612366e+00 5.71000993e-01 -2.63345450e-01
1.75748572e-01 -3.08309883e-01 4.01717909e-02 7.78214514e-01
-2.56284207e-01 -1.49963930e-01 -8.04704368e-01 4.65476930e-01
-1.39422572e+00 -7.17068851e-01 -3.85424137e-01 2.67202616e+00
5.58289766e-01 -3.50588709e-01 -3.04265469e-01 -1.54737560e-02
7.85393000e-01 -1.32016703e-01 -3.23502660e-01 -5.54547310e-01
-2.45017603e-01 -5.11365950e-01 1.00948584e+00 9.91680980e-01
-7.77842104e-01 5.35535157e-01 6.30632448e+00 6.50348842e-01
-1.11509109e+00 -3.13515812e-01 4.05301481e-01 7.43396357e-02
-7.14155674e-01 7.16010928e-02 -8.74635398e-01 6.17767334e-01
2.31205240e-01 -3.43530059e-01 5.42701602e-01 8.38417530e-01
4.86839771e-01 -6.17422283e-01 -5.30827403e-01 1.44456923e+00
1.33004516e-01 -1.30535543e+00 1.44681362e-02 3.33644480e-01
9.04398799e-01 -1.68027040e-02 7.43674189e-02 -4.53786612e-01
6.36796877e-02 -7.82995999e-01 6.53999925e-01 2.51562119e-01
1.37922084e+00 -7.40787745e-01 4.82476503e-01 5.27592182e-01
-1.28448939e+00 8.84202495e-02 -9.29329395e-01 -2.99994081e-01
2.05931306e-01 4.80068117e-01 -6.53685629e-01 4.22403395e-01
8.69552851e-01 4.61153626e-01 -2.16132626e-01 1.51991355e+00
3.27933282e-01 -2.36519352e-01 -5.21845937e-01 1.23523541e-01
1.96577966e-01 -1.03905606e+00 5.63877702e-01 8.02505910e-01
8.82354498e-01 2.19446346e-01 -1.71957597e-01 4.46752042e-01
3.17448705e-01 -3.67452800e-02 -1.19575322e+00 9.41953003e-01
7.90184975e-01 8.74658346e-01 -6.26664162e-01 -5.30448072e-02
-8.26302648e-01 9.50101316e-01 -1.29996583e-01 4.50988948e-01
-6.54132724e-01 -6.74459577e-01 8.25262308e-01 5.28279960e-01
6.57762438e-02 -5.45917928e-01 -5.64317822e-01 -1.30510628e+00
1.72586590e-01 -5.33487737e-01 -5.42876236e-02 -1.13600302e+00
-5.46039164e-01 5.97181976e-01 2.63992092e-03 -1.56007326e+00
-1.79305106e-01 -6.04222953e-01 -4.76239204e-01 1.02781975e+00
-1.37835622e+00 -7.45455980e-01 -7.31304824e-01 7.90229023e-01
5.31253874e-01 1.31968960e-01 7.96637535e-01 2.34738603e-01
4.58785631e-02 4.74206477e-01 6.30790532e-01 -6.04167692e-02
5.24358869e-01 -9.78948176e-01 3.27499211e-01 9.03782368e-01
-2.95712978e-01 8.06170702e-01 4.98585433e-01 -3.61443609e-01
-1.53872573e+00 -9.64945972e-01 7.70061016e-01 -2.01587617e-01
2.17644751e-01 -3.75965118e-01 -6.40852153e-01 5.66867948e-01
9.25888792e-02 -6.38955981e-02 3.67011040e-01 -4.71336156e-01
-1.76596150e-01 -3.91344458e-01 -1.17674482e+00 1.09821641e+00
1.22318447e+00 -4.80125725e-01 -1.32185578e-01 2.30338946e-02
3.40546131e-01 -7.81760633e-01 -7.61073351e-01 4.24015552e-01
8.03225338e-01 -1.40161514e+00 9.91740346e-01 4.23632830e-01
1.90954044e-01 -6.78355336e-01 -3.85933816e-01 -1.35787666e+00
-3.41946602e-01 -5.91772258e-01 7.13700473e-01 8.38664770e-01
1.20931625e-01 -1.10000253e+00 9.36353028e-01 5.22273004e-01
-4.15144145e-01 -6.72872186e-01 -1.20340776e+00 -6.21811032e-01
-2.29689375e-01 -1.02568366e-01 4.94345397e-01 7.43238747e-01
2.36757532e-01 -1.21335149e-01 -1.21514119e-01 3.41302633e-01
6.75271690e-01 1.91336915e-01 9.62513447e-01 -9.84259427e-01
9.71146673e-02 -2.81230450e-01 -6.27781868e-01 -1.94923627e+00
-2.69921303e-01 -4.79593754e-01 -2.74399787e-01 -1.32733297e+00
1.39026821e-01 -8.73536766e-01 6.66204751e-01 -3.34325403e-01
2.27549374e-01 7.40646660e-01 -1.96162894e-01 3.24459642e-01
-8.38763826e-03 3.73107970e-01 1.45359063e+00 2.04836845e-01
-3.84543762e-02 1.07906513e-01 -5.60020566e-01 9.13588345e-01
6.94639862e-01 5.98402284e-02 -6.56516075e-01 -9.22067583e-01
4.01253879e-01 2.55457342e-01 1.65280521e-01 -8.03573549e-01
5.57053030e-01 -2.40561888e-01 1.81696057e-01 -7.39028275e-01
6.01293504e-01 -9.19060409e-01 6.29896224e-01 1.39651552e-01
2.66535670e-01 2.74989128e-01 -9.17774662e-02 2.27677062e-01
-5.18191934e-01 -3.30054462e-01 9.68699992e-01 -2.23476738e-01
-4.88472909e-01 2.95734644e-01 -4.38296765e-01 -2.25413471e-01
1.17256951e+00 -1.09728813e+00 -2.00358555e-01 -7.48826325e-01
-4.83886413e-02 1.15793943e-01 1.48973179e+00 3.66759859e-02
9.05982912e-01 -1.47250092e+00 -3.37953299e-01 5.78089297e-01
9.81503278e-02 4.45276439e-01 2.33058169e-01 7.41036415e-01
-1.27626240e+00 8.61650765e-01 -1.76506758e-01 -8.29237759e-01
-1.35357881e+00 3.46351087e-01 5.23952007e-01 3.71391594e-01
-7.48886228e-01 4.44664568e-01 9.48696375e-01 -5.34407973e-01
-2.71970462e-02 -2.33004272e-01 -1.88571692e-01 -5.09434700e-01
5.87493300e-01 6.36996806e-01 3.77461724e-02 -6.81341469e-01
-2.66042739e-01 1.33000410e+00 2.27359101e-01 -1.47309884e-01
9.61896181e-01 -6.93165064e-01 -2.51126081e-01 8.63638967e-02
1.13317609e+00 6.37752593e-01 -1.27501488e+00 2.46683925e-01
-7.16151178e-01 -1.15682662e+00 -8.81134719e-02 1.14978395e-01
-7.61460841e-01 7.05495417e-01 1.05249465e-01 2.02969119e-01
1.08207703e+00 -3.04747880e-01 5.83372712e-01 8.78275633e-02
1.01344860e+00 -7.39849091e-01 -5.66115201e-01 6.30662143e-01
7.56660700e-01 -7.57896900e-01 -2.99270153e-01 -8.40382755e-01
-3.03705126e-01 1.40923178e+00 6.49255812e-01 1.27497446e-02
5.74868262e-01 5.01397252e-01 1.08878128e-01 1.87252194e-01
-7.43206665e-02 4.11491007e-01 2.98425998e-03 7.38322198e-01
2.08610535e-01 -6.49966002e-02 -3.60985368e-01 3.09656709e-02
-5.30351520e-01 -2.04532385e-01 8.28881025e-01 5.59163153e-01
-4.89458382e-01 -8.07943881e-01 -8.46664071e-01 6.59078881e-02
-9.22935084e-02 -8.77314657e-02 1.45587832e-01 6.33809865e-01
-1.07127689e-01 8.71899664e-01 6.72594011e-01 -1.68619305e-01
4.14323390e-01 -7.64660358e-01 3.60296756e-01 -3.47236276e-01
3.81724328e-01 -7.87097439e-02 -4.41465639e-02 -5.62834084e-01
-2.80169602e-02 -6.57041788e-01 -1.05733418e+00 -5.13764739e-01
-1.42605901e-01 3.74034733e-01 7.25730777e-01 4.51316684e-01
3.50521147e-01 -3.78672570e-01 1.03261447e+00 -1.06511807e+00
-1.39143229e-01 -3.89116824e-01 -6.94369018e-01 -3.23655307e-02
6.31759688e-02 -6.59830749e-01 -4.92579997e-01 -1.81487828e-01] | [9.016045570373535, -2.4690775871276855] |
15ed20ba-ef73-45e4-b0ed-d1d511c67a99 | deep-learning-with-a-classifier-system | 2103.01118 | null | https://arxiv.org/abs/2103.01118v1 | https://arxiv.org/pdf/2103.01118v1.pdf | Deep Learning with a Classifier System: Initial Results | This article presents the first results from using a learning classifier system capable of performing adaptive computation with deep neural networks. Individual classifiers within the population are composed of two neural networks. The first acts as a gating or guarding component, which enables the conditional computation of an associated deep neural network on a per instance basis. Self-adaptive mutation is applied upon reproduction and prediction networks are refined with stochastic gradient descent during lifetime learning. The use of fully-connected and convolutional layers are evaluated on handwritten digit recognition tasks where evolution adapts (i) the gradient descent learning rate applied to each layer (ii) the number of units within each layer, i.e., the number of fully-connected neurons and the number of convolutional kernel filters (iii) the connectivity of each layer, i.e., whether each weight is active (iv) the weight magnitudes, enabling escape from local optima. The system automatically reduces the number of weights and units while maintaining performance after achieving a maximum prediction error. | ['Larry Bull', 'Richard J. Preen'] | 2021-03-01 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 3.78412813e-01 8.76112953e-02 4.09669019e-02 -1.53570980e-01
6.22985780e-01 -2.65756518e-01 3.53810102e-01 3.24162662e-01
-1.03767073e+00 8.49951982e-01 -4.65460628e-01 -4.79278937e-02
-2.20946446e-01 -1.10707533e+00 -6.32077754e-01 -1.23014927e+00
-2.30581403e-01 4.29510117e-01 6.09785914e-01 -1.55640140e-01
3.68697852e-01 8.33447874e-01 -1.89494359e+00 2.19465509e-01
6.80044055e-01 1.37224722e+00 1.29904717e-01 8.59545112e-01
7.38472417e-02 5.50482273e-01 -9.69033062e-01 -5.30201308e-02
9.06433389e-02 -5.38936079e-01 -4.79415596e-01 1.48015305e-01
-7.49839693e-02 2.99679130e-01 -1.28253534e-01 9.71360624e-01
4.37601566e-01 1.32873625e-01 4.84508961e-01 -9.81361568e-01
-3.75521392e-01 7.49417126e-01 1.80227101e-01 4.53141093e-01
-4.83663350e-01 3.36973161e-01 4.22483385e-01 -3.14482659e-01
3.25588882e-01 8.62178564e-01 5.19332051e-01 8.05939615e-01
-1.33528376e+00 -4.27068532e-01 2.12648869e-01 -7.66728073e-02
-1.27391875e+00 -1.99184015e-01 2.82030672e-01 -1.45046085e-01
1.48820841e+00 1.03608318e-01 1.19063032e+00 4.14341182e-01
7.19476700e-01 3.10298264e-01 5.83660007e-01 -5.84402561e-01
1.08869374e+00 1.85263798e-01 -3.72714899e-03 9.61114824e-01
4.28814828e-01 2.14724243e-01 -5.60969353e-01 -1.93279088e-02
7.95713007e-01 -3.60463738e-01 5.89973107e-02 -3.18123698e-01
-6.34498358e-01 7.48412073e-01 5.55391312e-01 5.66467822e-01
-6.85959995e-01 3.51167470e-01 3.64737958e-01 4.37344640e-01
-2.67118085e-02 9.67165530e-01 -8.19768906e-01 -1.61455333e-01
-8.11575294e-01 -5.02955765e-02 7.23685324e-01 9.55682844e-02
9.27185416e-01 4.90527928e-01 -2.32228547e-01 8.34827662e-01
3.32422070e-02 -3.71300280e-02 1.13509405e+00 -6.88504934e-01
-5.95481507e-02 1.35372710e+00 -4.95599180e-01 -4.81010109e-01
-3.69182974e-01 -6.14185929e-01 -8.97624016e-01 7.96813488e-01
1.76660076e-01 -7.00859249e-01 -1.14081621e+00 1.64259362e+00
1.29561096e-01 -1.03825733e-01 2.52166420e-01 2.79481262e-01
1.79177046e-01 4.15741056e-01 2.91615963e-01 -3.73639226e-01
9.96056616e-01 -7.48130858e-01 -3.16775739e-01 -4.25621599e-01
3.73070002e-01 -1.40528455e-01 3.15064192e-01 2.89881915e-01
-1.30415428e+00 -8.17455828e-01 -1.38615966e+00 8.00476074e-01
-7.47845471e-01 2.73610324e-01 6.87936485e-01 8.57348084e-01
-1.08254516e+00 8.91457617e-01 -1.05465746e+00 2.37383763e-03
4.83294427e-01 8.74714434e-01 3.05559095e-02 3.39699954e-01
-1.07888627e+00 8.19736779e-01 9.69975054e-01 1.90254405e-01
-8.52933049e-01 -3.63504529e-01 -5.92477918e-01 3.45466286e-01
-3.30642104e-01 -6.08119845e-01 5.86456954e-01 -1.88666272e+00
-1.73555982e+00 6.99957728e-01 2.43838891e-01 -8.26895475e-01
3.11978489e-01 3.87409121e-01 -3.43540013e-01 -1.53856680e-01
-6.18457556e-01 1.10466540e+00 1.17485785e+00 -7.46219695e-01
-8.31832647e-01 -4.63417619e-01 -2.91976810e-01 2.21476465e-01
-8.17685246e-01 -2.82091081e-01 -2.74366230e-01 -6.37313187e-01
2.15398952e-01 -9.99734163e-01 -3.81861985e-01 -1.02203429e-01
6.51697293e-02 -1.70211002e-01 7.39080489e-01 -1.40287504e-01
1.34533036e+00 -2.12862420e+00 5.82651019e-01 5.05164385e-01
-2.14378759e-01 5.41628242e-01 -1.18758336e-01 -1.20329425e-01
-2.66057134e-01 -1.63179457e-01 -6.39952242e-01 1.76206902e-01
-4.13119078e-01 4.01578367e-01 2.67290831e-01 1.38489366e-01
7.01813042e-01 6.91233635e-01 -4.35829550e-01 -7.69960508e-02
-1.48399556e-02 5.83952665e-01 -6.35665238e-01 5.17740101e-02
-2.58063078e-01 1.63142458e-01 -3.41746241e-01 3.71065289e-01
2.71357268e-01 -5.70404455e-02 2.74378449e-01 2.52807468e-01
-3.23309839e-01 -5.21987081e-02 -1.29290164e+00 1.20133126e+00
-3.62186551e-01 6.08013690e-01 -2.19491720e-01 -9.61153328e-01
1.42992377e+00 3.53572443e-02 1.48897290e-01 -6.54680550e-01
2.94353902e-01 1.45488068e-01 6.23750031e-01 1.60326324e-02
3.53679150e-01 3.65011334e-01 2.30699912e-01 3.41455102e-01
3.03109586e-01 -6.22901917e-02 5.20799220e-01 -3.57264042e-01
1.29202890e+00 1.63622964e-02 1.03081897e-01 -3.55985433e-01
7.22862184e-01 -2.10505813e-01 6.77468956e-01 6.56100571e-01
-1.44982338e-02 2.10982785e-01 6.08356893e-01 -8.78700495e-01
-1.14491546e+00 -6.83523476e-01 -2.74135649e-01 1.35522461e+00
-3.44484478e-01 1.90249383e-01 -1.09297967e+00 -2.95112073e-01
1.18859932e-01 8.15336943e-01 -1.05859125e+00 -9.06892896e-01
-5.90518653e-01 -1.21777833e+00 5.56242824e-01 5.31035841e-01
6.32022619e-01 -1.54429114e+00 -1.47629380e+00 4.20786113e-01
7.56538808e-01 -2.14012370e-01 1.05940595e-01 1.18018758e+00
-1.38716030e+00 -7.34895229e-01 -5.13003886e-01 -9.60490763e-01
1.22901380e+00 -9.75401521e-01 8.95843208e-01 6.40426993e-01
-7.09608018e-01 -1.14525124e-01 7.55685940e-03 -3.13756526e-01
-4.76267874e-01 3.71668696e-01 -1.17806092e-01 1.78532198e-01
1.14380740e-01 -5.14641285e-01 -3.58137637e-01 2.32197091e-01
-8.72933626e-01 -1.91272900e-01 5.19758165e-01 9.62635398e-01
7.50532568e-01 5.25859833e-01 3.02529812e-01 -4.05701131e-01
8.26771080e-01 -7.80625492e-02 -7.14670181e-01 5.07857084e-01
-8.64888191e-01 4.92069006e-01 3.87428552e-01 -7.20061481e-01
-9.48806405e-01 3.40101331e-01 1.24632254e-01 -3.55654513e-03
8.75446349e-02 2.16127411e-01 9.89486575e-02 -2.16868520e-01
7.69864976e-01 4.69826847e-01 -1.43747464e-01 9.70959067e-02
-1.20045915e-01 2.65184045e-01 3.73943508e-01 -2.88096756e-01
8.65143761e-02 -7.29739964e-02 -2.21088752e-02 -4.43512559e-01
-4.07654904e-02 4.68532443e-01 -9.10416961e-01 -4.61264551e-01
7.54208624e-01 -3.24184030e-01 -7.01440990e-01 9.77509558e-01
-6.10334873e-01 -7.15759635e-01 -7.02949405e-01 2.19642624e-01
-1.54767632e-01 -4.91041452e-01 -5.90336263e-01 -8.30382824e-01
-5.08048296e-01 -9.69138443e-01 2.31150538e-01 7.86128700e-01
-2.79752403e-01 -1.04849088e+00 1.08048350e-01 -3.53651106e-01
5.56650341e-01 1.00521438e-01 1.18698525e+00 -5.08563876e-01
-3.99091952e-02 -4.29675311e-01 2.28014171e-01 7.29056239e-01
-2.26727918e-01 3.70887727e-01 -5.87605238e-01 -4.64695632e-01
-4.01838154e-01 -2.99221367e-01 1.01259482e+00 4.62037683e-01
1.14572394e+00 -2.38048196e-01 -3.10816020e-01 5.55436194e-01
1.14347208e+00 5.84312201e-01 7.71062434e-01 7.08239794e-01
6.64500743e-02 3.40521902e-01 -2.94270851e-02 7.47408867e-01
-2.88927138e-01 2.10135564e-01 5.01743078e-01 2.31111333e-01
-2.20977142e-01 2.64480501e-01 2.08020031e-01 3.93752515e-01
-1.84600636e-01 -2.04862639e-01 -9.56921637e-01 3.02224189e-01
-1.54848921e+00 -6.28700495e-01 4.82501656e-01 2.26791072e+00
9.87300813e-01 5.56490719e-01 -3.42853606e-01 2.53564954e-01
8.15266013e-01 -2.17788950e-01 -1.15716255e+00 -7.72221148e-01
-4.36215192e-01 5.98953187e-01 6.53837979e-01 2.12242320e-01
-5.92786014e-01 8.86726618e-01 6.02452707e+00 2.62149304e-01
-1.43716812e+00 -4.13331509e-01 9.53035116e-01 -3.37116063e-01
9.60068703e-02 -1.02218233e-01 -9.40439105e-01 5.98683536e-01
9.76763546e-01 2.21322581e-01 8.54945183e-01 8.37675035e-01
-5.11441052e-01 -4.30237204e-01 -8.37819517e-01 4.42865252e-01
-8.06746334e-02 -1.36705947e+00 1.00605667e-01 -1.15161344e-01
1.08835411e+00 4.36907113e-02 1.93015561e-01 1.26794174e-01
4.19862509e-01 -9.60738301e-01 9.12573755e-01 7.42196977e-01
4.35365796e-01 -1.22004318e+00 7.47427106e-01 2.55776793e-01
-8.17572534e-01 -8.58972073e-01 -2.94096440e-01 -7.90410768e-03
-4.37578052e-01 3.46604764e-01 -3.67656529e-01 -3.29701066e-01
8.41500580e-01 4.01862189e-02 -9.78920102e-01 1.02589405e+00
-4.22878653e-01 3.48396808e-01 -3.04443508e-01 -3.77537817e-01
3.98073107e-01 4.96374480e-02 2.05888793e-01 1.05144441e+00
2.21351460e-01 2.52853930e-02 -2.62003124e-01 6.20709479e-01
-1.20232068e-01 -3.37209374e-01 2.39259541e-01 -1.45999625e-01
6.93356812e-01 9.67727900e-01 -1.16726983e+00 -2.93671787e-01
5.71252644e-01 9.90732551e-01 5.79545736e-01 2.79788673e-01
-2.79357582e-01 -6.06335282e-01 4.89087015e-01 -1.38787478e-01
6.31866455e-01 8.82480759e-03 -5.18875599e-01 -4.30969417e-01
-3.12965125e-01 -5.88299215e-01 4.52853769e-01 -4.43033248e-01
-6.05114460e-01 9.27070558e-01 -5.81920564e-01 -4.44607645e-01
-5.09397745e-01 -6.15630448e-01 -8.29960167e-01 8.69522393e-01
-1.11127496e+00 -3.66080135e-01 -1.88692525e-01 4.69924480e-01
2.45756686e-01 -8.90750110e-01 9.79073644e-01 -2.29627177e-01
-1.13780344e+00 5.95483422e-01 1.76731691e-01 7.66136423e-02
-7.06084073e-02 -8.75612020e-01 2.14464560e-01 5.83356619e-01
-1.69001833e-01 4.03154731e-01 3.66914630e-01 -5.75248420e-01
-1.04280829e+00 -9.43048298e-01 8.00115466e-01 3.06446135e-01
1.97979167e-01 -2.32013777e-01 -9.37515855e-01 2.54771590e-01
-2.64107324e-02 1.31039456e-01 4.85500664e-01 -1.53109699e-01
1.62334576e-01 -3.36819410e-01 -1.34812784e+00 4.16173458e-01
4.81439620e-01 9.22522023e-02 -2.29999408e-01 -3.12647521e-02
2.31834620e-01 -4.26232189e-01 -6.55930877e-01 2.71858901e-01
5.29783905e-01 -9.00543511e-01 5.09866416e-01 -5.61994076e-01
1.30591184e-01 -2.16781259e-01 1.11764260e-01 -1.33872867e+00
-5.83941877e-01 -1.58784539e-01 -2.38335878e-01 9.06481683e-01
8.43375504e-01 -7.76712418e-01 9.35890555e-01 8.10440719e-01
4.35922816e-02 -9.66283023e-01 -1.27559757e+00 -3.86505961e-01
-1.79858021e-02 5.83372638e-02 5.03809333e-01 6.29801512e-01
-3.96862060e-01 1.02787651e-03 2.44867891e-01 -4.66550849e-02
1.52255476e-01 -4.42649037e-01 -1.06153145e-01 -1.13775408e+00
-4.63012010e-01 -8.75833452e-01 -7.10973561e-01 -2.36142501e-01
7.16641033e-03 -5.31911910e-01 -1.11814708e-01 -9.09182250e-01
-1.59856185e-01 -5.40836692e-01 -5.79476833e-01 7.42787659e-01
7.17222691e-02 8.94533768e-02 4.68896143e-02 -6.65985867e-02
-1.90086171e-01 4.03857470e-01 6.29289925e-01 -1.19753465e-01
-6.33676052e-01 1.04747750e-01 -1.16162032e-01 7.66257167e-01
9.97921407e-01 -5.01455963e-01 -2.28976980e-01 -4.40510094e-01
3.89128655e-01 -2.83600956e-01 9.52361971e-02 -1.51481664e+00
6.09430492e-01 1.47989497e-01 1.18437159e+00 5.49195036e-02
5.06119654e-02 -5.93032598e-01 2.39433214e-01 1.37257230e+00
-6.87756300e-01 4.16178644e-01 6.09078467e-01 2.99558222e-01
8.06046068e-04 -7.53790915e-01 1.14120555e+00 -2.71707326e-01
-7.66056120e-01 -2.32127346e-02 -8.29300821e-01 -2.52138317e-01
1.21645880e+00 -6.18680179e-01 1.89701974e-01 3.55573505e-01
-8.53498340e-01 1.81272462e-01 2.26242855e-01 5.56052804e-01
6.35678887e-01 -9.97087777e-01 -6.41729534e-01 6.84823990e-01
-2.42581755e-01 -3.81875038e-01 5.20363711e-02 1.73615187e-01
-6.07593656e-01 1.85465030e-02 -7.97004819e-01 -3.22452247e-01
-1.25677526e+00 3.11269432e-01 1.10549533e+00 -1.53212681e-01
-5.20213768e-02 1.29343653e+00 -4.52486485e-01 -2.36019462e-01
6.60331845e-01 5.03132865e-02 -4.94660139e-01 1.25926509e-01
5.99374533e-01 3.81779671e-01 2.58217394e-01 -2.84661204e-01
-4.85143334e-01 2.99773604e-01 -1.15867533e-01 1.87266376e-02
1.49738121e+00 3.50112230e-01 -4.60470706e-01 2.61744857e-01
7.41520166e-01 -8.18621278e-01 -1.73111737e+00 2.60624811e-02
-5.68170883e-02 1.70135230e-01 4.67469692e-01 -1.15897429e+00
-1.43850088e+00 5.22074163e-01 1.07430184e+00 2.21934821e-02
1.41947806e+00 -5.03857374e-01 1.15039401e-01 6.89023435e-01
1.70690164e-01 -1.66930890e+00 2.24809170e-01 7.02516139e-01
7.50526547e-01 -5.92431307e-01 5.09129949e-02 5.35285711e-01
-4.76375520e-01 1.60930967e+00 8.35837007e-01 -1.74833804e-01
6.12834513e-01 5.09715438e-01 -2.46898979e-01 -6.66757748e-02
-1.08887541e+00 1.03964299e-01 1.16997480e-01 4.30386215e-01
2.74279684e-01 -1.33067053e-02 -1.94309905e-01 9.38360691e-02
-7.49887824e-02 -1.93319172e-01 3.40436921e-02 1.12925804e+00
-8.97069693e-01 -8.04378211e-01 -1.59333229e-01 2.86409080e-01
-5.90678081e-02 -9.33830142e-02 -4.22828108e-01 3.48863482e-01
8.00457656e-01 5.03970444e-01 7.36325383e-01 -1.46549433e-01
2.68735975e-01 7.03783557e-02 6.54105723e-01 -4.54834729e-01
-1.26722670e+00 -5.48096359e-01 -3.30607712e-01 -9.44834352e-02
6.13105707e-02 -7.79729486e-01 -1.56640232e+00 -8.82916451e-02
-3.29360127e-01 2.47124061e-01 9.61771250e-01 6.56160593e-01
2.31556714e-01 1.05287492e+00 4.54314262e-01 -6.58514917e-01
-4.80297267e-01 -7.87467539e-01 -5.05174637e-01 -2.54139751e-01
-3.94339785e-02 -3.67325902e-01 -1.03102840e-01 1.46325469e-01] | [8.293109893798828, 3.190380573272705] |
0cc76bc1-b772-43d7-9182-8ba2d854f7da | a-study-of-neural-matching-models-for-cross | 2005.12994 | null | https://arxiv.org/abs/2005.12994v1 | https://arxiv.org/pdf/2005.12994v1.pdf | A Study of Neural Matching Models for Cross-lingual IR | In this study, we investigate interaction-based neural matching models for ad-hoc cross-lingual information retrieval (CLIR) using cross-lingual word embeddings (CLWEs). With experiments conducted on the CLEF collection over four language pairs, we evaluate and provide insight into different neural model architectures, different ways to represent query-document interactions and word-pair similarity distributions in CLIR. This study paves the way for learning an end-to-end CLIR system using CLWEs. | ['James Allan', 'Puxuan Yu'] | 2020-05-26 | null | null | null | null | ['cross-lingual-information-retrieval'] | ['natural-language-processing'] | [-3.65484744e-01 -3.60565037e-01 -3.41570377e-01 -5.43473423e-01
-8.80050659e-01 -6.57506287e-01 9.60833073e-01 5.30523121e-01
-1.13565350e+00 1.09416001e-01 4.62994844e-01 -4.68338072e-01
-4.26105440e-01 -3.42268854e-01 -2.80642509e-01 -6.38898164e-02
-1.94998547e-01 9.77058530e-01 -1.51767388e-01 -6.44277751e-01
1.19128443e-01 4.06487405e-01 -1.50907946e+00 5.87236524e-01
3.34157556e-01 5.74353218e-01 3.09620500e-02 7.95392215e-01
-2.20946983e-01 3.31482202e-01 -3.24026525e-01 -8.27372253e-01
9.40881744e-02 -2.45373826e-02 -8.15761030e-01 -8.90126765e-01
9.18778181e-01 1.75879225e-02 -5.07920444e-01 7.51922488e-01
9.09093320e-01 6.40786648e-01 9.96969283e-01 -1.09261000e+00
-1.76683402e+00 8.62030447e-01 -1.31744742e-01 3.61947149e-01
2.42541760e-01 -3.70762199e-01 1.63570869e+00 -1.27393413e+00
8.27088594e-01 1.48098218e+00 5.60213387e-01 8.91055822e-01
-1.05585194e+00 -4.39929783e-01 -2.35888839e-01 5.56478024e-01
-1.71283758e+00 -3.39976490e-01 5.33682585e-01 -1.90729603e-01
1.53350639e+00 2.45095417e-01 2.42606252e-02 9.87230599e-01
3.49015832e-01 1.11403787e+00 2.50371397e-01 -8.73590410e-01
-3.46866041e-01 3.41710478e-01 9.52286303e-01 2.88499951e-01
2.56200701e-01 4.82614875e-01 -3.79391432e-01 -4.62756664e-01
2.03829601e-01 -1.01630099e-01 -6.34471923e-02 -4.04190630e-01
-8.67998302e-01 1.13728428e+00 6.33026719e-01 9.05787468e-01
-1.10278234e-01 2.10153714e-01 8.57322454e-01 5.56998074e-01
4.80755031e-01 9.51963007e-01 -4.66143280e-01 2.99341649e-01
-3.21437269e-01 3.86734128e-01 4.80602592e-01 6.22761846e-01
5.01806140e-01 -4.68846262e-01 -6.12556577e-01 1.50783563e+00
5.32231510e-01 2.60643125e-01 9.68677998e-01 -3.94753367e-01
-8.78093764e-03 7.20831081e-02 -2.23397851e-01 -8.11575890e-01
-2.66294420e-01 -1.21984445e-01 -4.87049252e-01 -3.63262355e-01
-1.45503670e-01 1.60878703e-01 -2.99302548e-01 1.70996785e+00
-2.48817429e-01 -2.51277298e-01 4.47375983e-01 7.48204589e-01
1.15133190e+00 4.68699187e-01 2.99201787e-01 1.97945580e-01
1.50686228e+00 -1.08579826e+00 -9.16149378e-01 -2.17837662e-01
1.34813273e+00 -9.78973329e-01 1.08563328e+00 6.42868280e-02
-1.28142655e+00 -8.80853951e-01 -9.08167362e-01 -7.68807054e-01
-1.14177048e+00 1.07364304e-01 6.69574797e-01 1.23782791e-01
-1.36731875e+00 3.61336499e-01 -2.47019947e-01 -5.51380396e-01
-3.38966578e-01 1.19898885e-01 -3.94057155e-01 -4.82716262e-01
-1.80901742e+00 1.34238601e+00 5.63349783e-01 -5.30382395e-02
-1.42119780e-01 -9.40994978e-01 -1.09644902e+00 -8.03903444e-04
-5.12507141e-01 -4.36703593e-01 1.25847113e+00 -5.93009830e-01
-8.03505480e-01 1.26133299e+00 -3.16715121e-01 -3.91074181e-01
-1.48050144e-01 -1.59654468e-01 -7.96433210e-01 -5.88500679e-01
-2.53152579e-01 1.07975745e+00 -3.37747373e-02 -1.05403483e+00
-1.84513390e-01 -4.15852785e-01 -2.32264400e-01 4.16587681e-01
-6.52059734e-01 5.79485714e-01 -6.20211899e-01 -5.60116172e-01
-5.95728993e-01 -8.19369555e-01 1.11524519e-02 -2.87722468e-01
1.04175761e-01 -1.32246351e+00 5.31895876e-01 -5.21346271e-01
1.42430580e+00 -2.08771038e+00 1.12640455e-01 -9.41769928e-02
8.42446089e-02 5.56725442e-01 -9.51654553e-01 1.02228975e+00
-2.94959217e-01 1.44686967e-01 4.63362545e-01 -6.55125022e-01
4.14371371e-01 1.03659309e-01 -9.24634486e-02 1.19253825e-02
-2.27596723e-02 1.51363778e+00 -8.20585966e-01 -4.00773466e-01
-3.70242484e-02 8.16932321e-01 -6.28746867e-01 6.36355042e-01
8.12057182e-02 -5.66231370e-01 -1.44282907e-01 2.03277886e-01
6.32096350e-01 2.52614319e-01 2.49207586e-01 -4.95915055e-01
7.23761991e-02 2.60883301e-01 -2.26456538e-01 2.00996614e+00
-8.32778692e-01 9.95409131e-01 -3.36784303e-01 -8.56358051e-01
1.00005376e+00 2.11122230e-01 2.06986368e-01 -1.36137319e+00
3.19472998e-01 1.45806000e-01 8.69164243e-02 -4.98772591e-01
1.10560668e+00 3.16264391e-01 -2.80466914e-01 7.66938448e-01
3.56753737e-01 3.53141427e-01 4.60987657e-01 3.33263338e-01
8.70730639e-01 -2.07400903e-01 1.27637044e-01 -4.96255159e-01
4.13036525e-01 -1.73517257e-01 -3.85917485e-01 7.36471951e-01
-2.41408601e-01 1.58573046e-01 -4.23587531e-01 -6.53517067e-01
-9.03096199e-01 -1.16276884e+00 -3.75157505e-01 1.71279538e+00
3.09519209e-02 -5.16703486e-01 -3.33823413e-01 -4.36327159e-01
1.90648496e-01 9.69765365e-01 -7.94216275e-01 -6.68518305e-01
-4.69350636e-01 -3.22248936e-01 7.66909420e-01 5.06692290e-01
-1.79370880e-01 -1.27740681e+00 1.83016911e-01 1.50764495e-01
6.31964579e-02 -9.43195999e-01 -1.03044951e+00 3.80189925e-01
-3.66093934e-01 -7.90194213e-01 -5.68292022e-01 -1.39547801e+00
2.52431273e-01 5.60232460e-01 1.93746686e+00 1.95367798e-01
-7.35261798e-01 5.99708438e-01 -4.32177722e-01 -3.39434654e-01
-3.04351568e-01 8.15205500e-02 3.90054494e-01 -6.14516973e-01
1.65284133e+00 9.82326921e-03 -2.94260949e-01 2.08240539e-01
-9.37329650e-01 -7.30431557e-01 4.15626734e-01 9.99196351e-01
2.59594709e-01 -3.65103334e-01 3.73203546e-01 -5.00879765e-01
1.71583843e+00 -4.53022301e-01 -4.46144044e-01 8.59659731e-01
-7.30898798e-01 3.66650850e-01 2.93721497e-01 -6.99254513e-01
-7.90879726e-01 -6.85370266e-01 -4.34673913e-02 -5.64894795e-01
-9.00253206e-02 8.32964122e-01 3.63429844e-01 1.73420504e-01
1.00727367e+00 4.79752347e-02 -1.84557408e-01 -5.89483857e-01
9.42927539e-01 1.07899618e+00 3.96692753e-01 -8.34438443e-01
1.51887909e-01 -3.27285975e-01 -7.12147236e-01 -6.95325077e-01
-6.02822304e-01 -9.32182729e-01 -6.39063537e-01 2.57743955e-01
1.14107835e+00 -9.25939023e-01 -8.02296996e-01 -2.53761616e-02
-1.49925661e+00 4.37696576e-02 1.01231806e-01 5.18538535e-01
-3.79351736e-03 7.61143491e-02 -7.78128743e-01 -6.38942480e-01
-6.76468551e-01 -8.96757126e-01 1.12180018e+00 -1.23438962e-01
-7.32156098e-01 -1.66317856e+00 9.65049684e-01 3.87581468e-01
6.26924515e-01 -8.21194708e-01 1.08890831e+00 -1.28687203e+00
8.70173350e-02 -3.70095462e-01 -4.07467157e-01 4.46000665e-01
-1.73076272e-01 -1.59304589e-02 -8.69261205e-01 -4.92465794e-01
-7.67546117e-01 -7.22187221e-01 7.88861156e-01 1.32990718e-01
9.83524740e-01 -9.85627249e-03 -3.99135590e-01 2.39495218e-01
1.41340649e+00 2.19906375e-01 5.31138122e-01 1.26993239e-01
6.54518068e-01 8.81579816e-01 3.63650173e-01 -1.58650517e-01
4.94429260e-01 1.03838301e+00 -2.79982775e-01 -3.30223218e-02
-2.60546088e-01 -3.70938271e-01 3.06413293e-01 1.15984464e+00
6.11538649e-01 -6.06746316e-01 -8.81737590e-01 8.07061672e-01
-1.66832292e+00 -7.67099917e-01 3.98789719e-02 1.89622033e+00
9.34009314e-01 -5.94016969e-01 -3.49894434e-01 -6.86802745e-01
5.09122670e-01 -2.23693363e-02 -2.61890888e-01 -1.16987157e+00
1.88348908e-02 5.50823450e-01 3.65849197e-01 8.36643398e-01
-8.55296135e-01 1.24110889e+00 6.66480780e+00 9.86310422e-01
-7.71777332e-01 5.62924035e-02 1.38830274e-01 1.91868752e-01
-4.62183326e-01 -5.80966175e-01 -9.93622780e-01 -6.48259446e-02
1.19660223e+00 -3.92756760e-01 4.96929944e-01 6.76131666e-01
-3.08972120e-01 3.89797807e-01 -1.63632464e+00 1.08390892e+00
5.83044827e-01 -1.16426647e+00 3.97245020e-01 -1.22548990e-01
3.30953419e-01 6.14875555e-01 -3.69114876e-02 8.75712693e-01
8.01715970e-01 -1.07271469e+00 -1.81695804e-01 2.16011792e-01
6.62343323e-01 -8.50156605e-01 8.46066833e-01 -5.20669334e-02
-9.97018158e-01 1.38902679e-01 -7.45803416e-01 3.31015795e-01
-2.74965852e-01 -1.91103995e-01 -7.82559156e-01 4.74294633e-01
6.75009847e-01 8.04390788e-01 -6.92880511e-01 5.10039151e-01
9.14908201e-02 -1.55829415e-01 1.25336453e-01 -3.78793210e-01
4.32473481e-01 -9.09317359e-02 2.77650636e-02 1.55814219e+00
2.24154200e-02 -3.19504291e-01 1.77480578e-02 7.07528472e-01
-3.86645108e-01 4.70373690e-01 -1.00030816e+00 -4.27730739e-01
6.79242909e-01 1.28792834e+00 2.83535689e-01 -2.55674124e-01
-2.92630374e-01 1.07796609e+00 8.86208415e-01 3.69003505e-01
-2.59281039e-01 -5.09763718e-01 1.20762932e+00 -3.57886016e-01
-2.15516463e-01 -1.48774743e-01 1.10838532e-01 -1.10119486e+00
-2.29983494e-01 -8.06558549e-01 7.47368217e-01 -1.06305277e+00
-2.20036817e+00 9.01506543e-01 1.60575107e-01 -9.35408950e-01
-6.61298156e-01 -9.36103880e-01 -4.83826131e-01 1.00476360e+00
-1.60933888e+00 -1.18302822e+00 3.56389314e-01 6.08378589e-01
5.10575950e-01 -6.96758747e-01 1.31819820e+00 6.74517155e-01
-2.25704864e-01 1.39289737e+00 5.61240196e-01 4.75729853e-01
1.02004838e+00 -9.02800024e-01 6.21201396e-01 2.96582758e-01
7.26201773e-01 1.18419456e+00 3.40714008e-01 -1.04459763e-01
-1.58082116e+00 -8.92663062e-01 1.68272507e+00 -7.98664749e-01
9.23030019e-01 -4.87094909e-01 -1.07608724e+00 6.28954589e-01
9.64612663e-01 -2.43945226e-01 1.11901712e+00 7.53550947e-01
-9.43730056e-01 1.68879062e-01 -8.07090461e-01 8.73404026e-01
9.68594968e-01 -1.11911082e+00 -8.44827175e-01 4.51629162e-01
9.99080300e-01 1.58623934e-01 -9.93351638e-01 3.18093866e-01
6.91149235e-01 -4.33570743e-01 1.43607986e+00 -1.26774287e+00
3.41857582e-01 3.06328893e-01 -5.27843237e-01 -1.80936134e+00
-5.80774605e-01 8.15698728e-02 1.95239782e-01 1.03452253e+00
3.51521194e-01 -3.86572987e-01 1.06095999e-01 6.33018076e-01
-6.07097782e-02 -5.15299737e-01 -7.20703244e-01 -8.85989904e-01
7.92303264e-01 -5.69598258e-01 2.51966000e-01 1.33698523e+00
3.68270278e-01 8.06742668e-01 -1.31708907e-06 -1.18882515e-01
4.33214843e-01 -3.56438279e-01 4.72908020e-01 -1.19839919e+00
-3.66692543e-01 -8.78174186e-01 -3.58154953e-01 -7.75928855e-01
1.01008022e+00 -1.52819335e+00 -8.48499686e-02 -1.27668154e+00
3.51002008e-01 -3.11722517e-01 -7.82432854e-01 3.34613770e-01
-2.76735991e-01 3.64094943e-01 -6.37840405e-02 5.01338318e-02
-6.70017421e-01 5.33917427e-01 6.54578507e-01 -4.21856046e-01
1.38424680e-01 -8.69530380e-01 -5.87746143e-01 3.33112739e-02
6.34496570e-01 -3.56403053e-01 -5.01908958e-01 -1.18356121e+00
3.30084503e-01 -3.13401014e-01 3.54830967e-03 -2.37626433e-01
3.04032654e-01 3.57412726e-01 7.48570934e-02 -5.91583371e-01
3.95858765e-01 -7.66822517e-01 -3.85263026e-01 1.01382516e-01
-1.22730398e+00 5.80029488e-01 3.68584186e-01 1.00809850e-01
-5.78632414e-01 -3.33751827e-01 4.15895700e-01 9.93448645e-02
-8.00773561e-01 2.23367855e-01 -3.01817596e-01 2.75480986e-01
4.19696063e-01 4.85403121e-01 -3.74286354e-01 -2.41258726e-01
-4.23030227e-01 4.59112436e-01 1.02329135e-01 1.45499766e+00
4.70599204e-01 -1.85143113e+00 -9.55931783e-01 4.18710500e-01
1.17111456e+00 -9.89888370e-01 2.99377274e-02 7.61210993e-02
4.11496125e-02 1.01723790e+00 1.34019822e-01 -2.47088626e-01
-1.67909598e+00 8.59181464e-01 4.06649619e-01 -5.32159567e-01
2.17975989e-01 1.15368056e+00 5.40320612e-02 -1.10825801e+00
3.97957563e-01 2.39028394e-01 -4.08571810e-01 1.99033648e-01
7.43896902e-01 -3.99873331e-02 3.37719917e-01 -7.03615785e-01
-3.42168063e-01 5.77425003e-01 -6.65233254e-01 -1.83579341e-01
8.71365666e-01 -1.58747211e-02 -3.24205458e-01 4.81717229e-01
2.19935298e+00 -4.27966177e-01 2.30654612e-01 -7.19487071e-01
3.07060182e-01 -2.38721624e-01 4.42209452e-01 -8.34161878e-01
-7.93426454e-01 1.16312659e+00 9.73719776e-01 -1.12124249e-01
5.32370031e-01 1.35089979e-01 9.10966516e-01 1.16333365e+00
3.64014246e-02 -1.12550247e+00 -7.35885128e-02 8.48217845e-01
9.21107471e-01 -1.37466085e+00 -4.16585594e-01 4.50966328e-01
-5.45496881e-01 1.03632641e+00 6.11847281e-01 -1.49289459e-01
9.76563990e-01 -3.25339064e-02 5.23187280e-01 -4.19173717e-01
-1.15158165e+00 -2.30153963e-01 9.23096299e-01 6.61439836e-01
1.29896975e+00 3.61726321e-02 -7.95180440e-01 3.52118552e-01
7.39158615e-02 -1.43849447e-01 -8.58298987e-02 7.16171563e-01
4.69732210e-02 -1.70379746e+00 1.53202593e-01 9.01958719e-02
-2.76725441e-01 -7.63530076e-01 -7.63155282e-01 6.08901858e-01
-2.34111041e-01 8.41039658e-01 6.66520715e-01 -3.96626711e-01
4.04834479e-01 4.00638431e-01 4.42804128e-01 -3.83829325e-01
-9.50293064e-01 -4.47290212e-01 2.80709684e-01 -5.72252929e-01
-8.02761763e-02 -2.65186161e-01 -7.92313457e-01 -1.52820036e-01
-2.49644190e-01 5.44416189e-01 8.57504845e-01 6.74205601e-01
7.11281359e-01 3.05652767e-01 5.12818396e-01 -6.18249595e-01
-4.78994280e-01 -1.29147661e+00 -3.50965410e-01 7.59033799e-01
-7.76206926e-02 -3.39509994e-01 -3.96169305e-01 -3.92473608e-01] | [11.274304389953613, 9.802224159240723] |
4197edb6-279e-4755-bdd0-e29624b27ddc | masked-feature-prediction-for-self-supervised | 2112.09133 | null | https://arxiv.org/abs/2112.09133v2 | https://arxiv.org/pdf/2112.09133v2.pdf | Masked Feature Prediction for Self-Supervised Visual Pre-Training | We present Masked Feature Prediction (MaskFeat) for self-supervised pre-training of video models. Our approach first randomly masks out a portion of the input sequence and then predicts the feature of the masked regions. We study five different types of features and find Histograms of Oriented Gradients (HOG), a hand-crafted feature descriptor, works particularly well in terms of both performance and efficiency. We observe that the local contrast normalization in HOG is essential for good results, which is in line with earlier work using HOG for visual recognition. Our approach can learn abundant visual knowledge and drive large-scale Transformer-based models. Without using extra model weights or supervision, MaskFeat pre-trained on unlabeled videos achieves unprecedented results of 86.7% with MViT-L on Kinetics-400, 88.3% on Kinetics-600, 80.4% on Kinetics-700, 39.8 mAP on AVA, and 75.0% on SSv2. MaskFeat further generalizes to image input, which can be interpreted as a video with a single frame and obtains competitive results on ImageNet. | ['Christoph Feichtenhofer', 'Alan Yuille', 'Chao-yuan Wu', 'Saining Xie', 'Haoqi Fan', 'Chen Wei'] | 2021-12-16 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wei_Masked_Feature_Prediction_for_Self-Supervised_Visual_Pre-Training_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wei_Masked_Feature_Prediction_for_Self-Supervised_Visual_Pre-Training_CVPR_2022_paper.pdf | cvpr-2022-1 | ['self-supervised-image-classification'] | ['computer-vision'] | [ 6.17536791e-02 -1.47063032e-01 -4.18892890e-01 -4.70925778e-01
-5.20068288e-01 -4.58859205e-01 6.59110785e-01 -4.71185505e-01
-6.54728293e-01 4.88201737e-01 8.71955529e-02 -1.29721165e-01
4.77611870e-01 -5.05008638e-01 -1.33666134e+00 -6.06290281e-01
-2.16668665e-01 1.93430454e-01 6.59397125e-01 -1.91333488e-01
3.48984040e-02 4.16192085e-01 -1.60062826e+00 8.11748207e-01
1.82232618e-01 1.36295855e+00 3.93637002e-01 8.71494174e-01
3.49740326e-01 1.23027301e+00 -1.92027673e-01 -2.02361912e-01
4.71617252e-01 -3.25074404e-01 -7.62341857e-01 3.93410563e-01
8.67348850e-01 -5.58070719e-01 -8.71896505e-01 7.91639864e-01
2.82687157e-01 -1.27870649e-01 6.42427325e-01 -1.46195531e+00
-4.39823896e-01 2.83767492e-01 -5.15771568e-01 7.00920463e-01
-4.12227474e-02 6.82429016e-01 8.53181362e-01 -1.29037404e+00
7.71427333e-01 9.27196622e-01 4.94517088e-01 5.40314734e-01
-1.17932212e+00 -6.45863831e-01 4.01692502e-02 6.04184031e-01
-1.42296028e+00 -6.81765676e-01 1.30007610e-01 -5.68280637e-01
1.31855404e+00 7.99808353e-02 7.46085107e-01 1.07886982e+00
4.04173940e-01 8.10925663e-01 1.13552630e+00 -3.20716143e-01
5.33412211e-02 8.99543092e-02 -1.84505999e-01 1.09673059e+00
-1.58606783e-01 4.17810917e-01 -8.65029693e-01 3.73549461e-01
1.02469468e+00 8.31956118e-02 -3.45352352e-01 -5.45395553e-01
-1.35200536e+00 7.08218634e-01 4.76751655e-01 -3.87356728e-02
-2.24593118e-01 4.45628822e-01 4.54382539e-01 5.95489144e-01
2.33964622e-01 1.03515640e-01 -5.20264804e-01 -1.32362500e-01
-1.11348653e+00 -3.12032569e-02 5.16418040e-01 1.10828340e+00
9.28036571e-01 4.32953030e-01 -3.02458584e-01 5.57056129e-01
6.60964102e-02 7.74913073e-01 4.80062813e-01 -1.19557440e+00
2.74507642e-01 1.72319219e-01 -2.11242866e-02 -7.92303145e-01
-2.30518520e-01 -3.40383112e-01 -7.57348001e-01 4.13061798e-01
3.75825465e-01 8.14984590e-02 -1.28295040e+00 1.64111269e+00
-6.73144162e-02 2.37918973e-01 -5.78830987e-02 1.01712430e+00
8.60866010e-01 8.16984296e-01 8.57593305e-03 -1.35621026e-01
1.28945816e+00 -1.49507892e+00 -2.37655997e-01 -4.70076084e-01
4.16600645e-01 -5.35505235e-01 1.01482439e+00 4.29699779e-01
-1.10234451e+00 -8.81569862e-01 -1.07808328e+00 8.35444406e-02
-2.67411828e-01 2.00903267e-01 4.88213331e-01 4.67052311e-01
-1.49347067e+00 6.74697697e-01 -8.55813563e-01 -4.16757017e-01
5.12681067e-01 6.07715905e-01 -6.85503423e-01 -2.62831777e-01
-8.09568048e-01 7.72607028e-01 1.79588392e-01 -1.75416574e-01
-1.49290407e+00 -5.65390766e-01 -7.89363086e-01 1.63664058e-01
2.41995394e-01 -3.92763704e-01 1.22949994e+00 -1.18863928e+00
-1.26155126e+00 1.00946856e+00 -3.33817273e-01 -8.08744967e-01
6.49664819e-01 -9.67626274e-02 -4.93089974e-01 5.13194799e-01
5.75289801e-02 1.36010611e+00 1.18661284e+00 -8.26813400e-01
-7.86961019e-01 -4.90756407e-02 -2.12615833e-01 5.62954694e-02
-4.56847399e-01 1.49088368e-01 -8.29114735e-01 -4.68203634e-01
-1.87634572e-01 -9.15534139e-01 -1.26512989e-01 5.01437150e-02
8.65049940e-03 8.08869228e-02 9.37149405e-01 -6.68081522e-01
8.25168312e-01 -1.96620131e+00 -1.80973917e-01 1.31845430e-01
3.68957102e-01 2.93728322e-01 -3.77189726e-01 2.66864657e-01
-7.04563633e-02 -1.73671901e-01 2.12682262e-01 2.90153064e-02
-2.42279887e-01 2.21924067e-01 -3.71401042e-01 5.40039480e-01
3.99863571e-01 1.22435486e+00 -6.82345331e-01 -4.57497388e-01
3.55126053e-01 3.40990514e-01 -8.08163643e-01 3.48114043e-01
-9.73675400e-02 1.34665206e-01 8.02174658e-02 5.97227454e-01
4.64788228e-01 -7.10729659e-01 2.54162163e-01 -4.60614204e-01
-1.07882805e-01 -1.11894131e-01 -7.45548725e-01 1.55565894e+00
-6.72455579e-02 1.05112660e+00 -2.01958045e-01 -1.09180081e+00
7.26175964e-01 2.62571812e-01 4.21691358e-01 -9.73993897e-01
1.07581623e-01 5.01830243e-02 1.33329509e-02 -5.45095026e-01
2.34235287e-01 1.41776666e-01 2.82897949e-01 2.36184642e-01
6.96714938e-01 2.69593060e-01 2.63330966e-01 2.42222592e-01
1.40437365e+00 5.50793894e-02 1.26133993e-01 -4.50190216e-01
3.80076498e-01 1.59307141e-02 3.70720208e-01 8.81928682e-01
-3.51680577e-01 7.58284152e-01 2.67597169e-01 -6.84204042e-01
-1.35918045e+00 -1.00974464e+00 1.46508440e-01 1.43325818e+00
5.34920581e-02 -4.99349952e-01 -6.65720403e-01 -8.22945774e-01
3.81614119e-02 4.27065156e-02 -6.30230010e-01 -1.98456720e-01
-5.95675409e-01 -4.12208587e-01 4.22972679e-01 9.47772861e-01
6.31664455e-01 -1.04842520e+00 -6.73320234e-01 1.17086984e-01
6.29557576e-03 -1.40483057e+00 -6.16766632e-01 5.12309968e-01
-7.62264550e-01 -1.00685513e+00 -6.48626566e-01 -1.08632660e+00
5.58610737e-01 5.61495245e-01 1.15844846e+00 1.10588074e-01
-2.83542812e-01 3.43625396e-01 -2.46758044e-01 8.21885765e-02
-2.22414657e-01 -1.00695878e-01 1.84061155e-01 4.84078228e-02
3.41727972e-01 -4.70835745e-01 -7.15197444e-01 6.77941561e-01
-6.89757168e-01 1.71556622e-01 6.53501689e-01 1.06153786e+00
4.99830425e-01 -4.85999703e-01 2.32963488e-01 -5.31390369e-01
-3.64999026e-01 -3.33478659e-01 -5.17340839e-01 1.97195217e-01
-5.82222998e-01 4.43296731e-02 6.59399629e-01 -5.21889687e-01
-6.53302193e-01 4.21392083e-01 -4.90614725e-03 -1.03176105e+00
-1.23589955e-01 -7.76574388e-02 5.77947162e-02 -4.49408442e-01
7.49460518e-01 4.98040974e-01 1.69422567e-01 -2.31376275e-01
2.69971907e-01 4.90622699e-01 7.94935703e-01 -3.42524141e-01
7.84804881e-01 7.45917916e-01 -2.43766874e-01 -8.14374089e-01
-6.81869984e-01 -4.70174640e-01 -5.59014738e-01 -3.51375550e-01
8.88443708e-01 -1.35555696e+00 -7.56135285e-01 3.64747733e-01
-8.73644590e-01 -7.03287005e-01 -2.14522630e-01 4.72552538e-01
-7.91789234e-01 1.23171844e-01 -9.30445731e-01 -2.16816634e-01
-2.97754109e-01 -1.16075110e+00 9.25456047e-01 4.38125953e-02
-3.29727866e-02 -6.72181368e-01 -2.54590750e-01 3.61573398e-01
5.13970613e-01 -2.03189552e-01 4.93012428e-01 -4.90690261e-01
-8.36031616e-01 -1.55127412e-02 -5.40416658e-01 7.47783124e-01
-1.89869091e-01 -1.81532770e-01 -1.20172679e+00 -6.13961458e-01
-2.16506585e-01 -8.30790877e-01 1.08066189e+00 2.98429132e-01
1.12217581e+00 -3.16022843e-01 -2.74525583e-01 8.44185293e-01
1.38800454e+00 1.96391106e-01 7.54478157e-01 2.84093946e-01
7.60217369e-01 1.66967824e-01 3.84472221e-01 3.34084392e-01
3.35783780e-01 6.40522361e-01 3.37325186e-01 -2.21040115e-01
-3.56713980e-01 -4.13015366e-01 7.95905709e-01 7.69818842e-01
-3.11833382e-01 -1.61833972e-01 -6.96734667e-01 4.11435068e-01
-1.71649718e+00 -1.16456890e+00 2.12466121e-01 1.88010669e+00
6.50746942e-01 2.24064738e-01 2.94926405e-01 -1.60214022e-01
7.18607068e-01 3.31632972e-01 -5.18212020e-01 -7.65309408e-02
-3.83711815e-01 2.85405695e-01 9.62081850e-01 2.64714748e-01
-1.35709143e+00 1.24506688e+00 7.19011831e+00 9.29043412e-01
-1.34854996e+00 3.13440859e-01 8.62957537e-01 -2.70057499e-01
2.21650824e-01 -5.28395735e-02 -7.72571325e-01 3.73459011e-01
1.04330242e+00 4.97987755e-02 4.49304819e-01 9.19911325e-01
3.75320427e-02 6.75840601e-02 -1.36774445e+00 1.19481611e+00
2.37257510e-01 -1.68052936e+00 1.07928686e-01 1.24454685e-01
8.50579500e-01 5.67937255e-01 1.30133629e-01 3.74378771e-01
2.82583326e-01 -1.25385809e+00 1.01870370e+00 1.95110679e-01
1.16099644e+00 -3.69366050e-01 6.09100044e-01 7.73093104e-02
-1.12150395e+00 -4.67082672e-02 -5.82533658e-01 -1.64319500e-01
-5.05827852e-02 5.64846061e-02 -7.62031138e-01 -8.16526040e-02
1.03147125e+00 9.35125768e-01 -7.12842524e-01 8.70798945e-01
-9.42512378e-02 8.42212975e-01 -3.14458042e-01 2.18554229e-01
4.62861627e-01 3.64374906e-01 3.53756063e-02 1.46057343e+00
1.47174254e-01 -8.15669149e-02 3.28422487e-01 3.79217684e-01
-2.89195031e-01 3.51855978e-02 -5.06897330e-01 5.63457943e-02
1.10627167e-01 1.16439593e+00 -7.12675810e-01 -7.39902437e-01
-6.58724427e-01 1.31541705e+00 4.14446443e-01 4.45253819e-01
-1.05599439e+00 7.91027322e-02 4.98346895e-01 2.78795779e-01
9.09192979e-01 -6.67671561e-02 2.73887992e-01 -1.42255771e+00
-7.68007059e-03 -9.85363305e-01 3.10645789e-01 -8.77344906e-01
-9.21754837e-01 6.66355610e-01 -3.15847188e-01 -1.46509778e+00
-2.35290438e-01 -8.96997154e-01 -4.79413033e-01 3.03062141e-01
-1.37702739e+00 -1.03297460e+00 -5.79229355e-01 7.58103311e-01
7.33867466e-01 -3.70970577e-01 5.64622641e-01 3.66881847e-01
-3.48157078e-01 7.65320301e-01 1.94104642e-01 3.33320081e-01
6.84454024e-01 -9.04327929e-01 6.74169421e-01 8.07537317e-01
4.95211184e-01 1.22258384e-02 5.65847278e-01 -4.10202503e-01
-1.55060732e+00 -1.12752247e+00 5.75028598e-01 -3.65400583e-01
7.17598021e-01 -4.16326910e-01 -7.60505438e-01 8.22692037e-01
3.65810454e-01 7.00274885e-01 2.59753495e-01 -4.46281105e-01
-8.08404803e-01 -2.80009240e-01 -8.68438363e-01 3.48227769e-01
1.47752833e+00 -5.66340506e-01 -2.39250034e-01 4.65095341e-01
5.97588956e-01 -3.73367280e-01 -7.88079619e-01 4.36836094e-01
6.48096621e-01 -1.16774595e+00 9.64166343e-01 -9.40295696e-01
5.54012477e-01 -1.60708949e-01 -4.86965925e-01 -1.05679893e+00
-5.98601758e-01 -4.26989794e-01 -2.48578072e-01 6.45316124e-01
3.23497474e-01 -3.16067338e-01 1.12333882e+00 2.00890258e-01
-5.14132641e-02 -6.43921912e-01 -6.74958467e-01 -1.13689911e+00
-1.81773856e-01 -4.03682828e-01 1.30720049e-01 7.48378038e-01
-1.19392700e-01 2.80996948e-01 -7.98145115e-01 -2.10076012e-02
5.72778702e-01 -3.79189998e-02 8.46143842e-01 -6.68512285e-01
-4.36076969e-01 -2.26003006e-01 -8.47711384e-01 -1.33652711e+00
5.78973107e-02 -7.30187476e-01 4.34037019e-03 -1.01730180e+00
4.83135819e-01 -7.54103959e-02 -5.19076526e-01 7.17128694e-01
1.90224782e-01 9.21769440e-01 4.15905595e-01 3.65709513e-01
-9.50487912e-01 4.60687995e-01 1.05699039e+00 -2.67638922e-01
2.06365243e-01 -5.27347267e-01 -3.33764941e-01 6.68836713e-01
7.23849773e-01 -3.28204393e-01 -1.95767984e-01 -4.30624515e-01
-3.46890479e-01 -2.26370580e-02 6.02306664e-01 -1.23730004e+00
2.54148990e-01 -1.60009369e-01 8.22199881e-01 -3.56938899e-01
3.97561103e-01 -6.76929533e-01 -5.05434871e-02 5.94335616e-01
-1.92230970e-01 3.12854052e-01 1.54131532e-01 4.53011870e-01
-3.04373920e-01 2.73758262e-01 7.85781384e-01 -1.86009496e-01
-1.44586933e+00 3.95601660e-01 -4.83832896e-01 9.58914086e-02
9.87830639e-01 -3.25373858e-01 -4.66708213e-01 -4.36485291e-01
-6.68715835e-01 2.35929385e-01 6.04389489e-01 4.75791574e-01
8.26528609e-01 -1.29997861e+00 -6.99345589e-01 4.60676491e-01
2.38107011e-01 -6.17443264e-01 3.15773755e-01 8.07497859e-01
-6.07182562e-01 5.27295232e-01 -6.27770662e-01 -8.59901071e-01
-1.27818656e+00 6.53135598e-01 2.16927916e-01 -5.77232195e-03
-6.91833496e-01 9.22064602e-01 4.26329225e-01 1.71392918e-01
3.35763603e-01 -1.34597793e-01 9.59745571e-02 -1.79821581e-01
7.05311239e-01 1.22942902e-01 7.21692443e-02 -7.96209633e-01
-5.35092473e-01 5.16075194e-01 -1.99453443e-01 -7.68463090e-02
1.29296148e+00 7.80943632e-02 2.58836329e-01 1.02161959e-01
1.61398947e+00 -3.23440403e-01 -1.84780812e+00 -3.17443997e-01
-2.39970744e-01 -3.95934463e-01 5.26485629e-02 -5.98499179e-01
-1.46047950e+00 9.23526824e-01 8.06433976e-01 -2.92114675e-01
1.08230472e+00 1.20802872e-01 6.81940377e-01 5.23523271e-01
4.37069535e-01 -1.03554678e+00 4.36495125e-01 6.18349791e-01
7.29072869e-01 -1.42772555e+00 -1.61343783e-01 -3.00517291e-01
-8.51394475e-01 9.64386880e-01 8.34258676e-01 -3.77849668e-01
4.29943264e-01 3.70259047e-01 6.82877228e-02 -1.06795073e-01
-1.18100572e+00 -2.92878032e-01 3.52014482e-01 4.70547110e-01
1.60700828e-01 -1.50598630e-01 1.85696930e-01 2.32765496e-01
-8.49418622e-03 1.68836668e-01 3.54840040e-01 9.63111818e-01
-6.50052726e-01 -5.51811874e-01 -1.65076241e-01 5.00283420e-01
-2.80944377e-01 -2.90325671e-01 3.73470626e-04 7.76927590e-01
9.72480252e-02 6.32139623e-01 2.80593038e-01 -8.69169950e-01
1.70983434e-01 1.60447493e-01 6.08470321e-01 -3.48636031e-01
-4.77428764e-01 2.00974256e-01 1.80384666e-02 -9.85826254e-01
-3.87526780e-01 -2.91641921e-01 -1.05790544e+00 -4.40458536e-01
-2.27686465e-02 4.62459587e-02 4.08448339e-01 7.61276960e-01
3.97864312e-01 1.66730240e-01 6.79836988e-01 -1.11937714e+00
-4.75297749e-01 -7.78973162e-01 -4.61359113e-01 4.23360199e-01
3.56528074e-01 -6.55712366e-01 -3.18672210e-01 5.71733177e-01] | [9.31340217590332, 1.0131551027297974] |
fe06c621-0432-4322-a152-2cf331ca02ba | video-propagation-networks | 1612.05478 | null | http://arxiv.org/abs/1612.05478v3 | http://arxiv.org/pdf/1612.05478v3.pdf | Video Propagation Networks | We propose a technique that propagates information forward through video
data. The method is conceptually simple and can be applied to tasks that
require the propagation of structured information, such as semantic labels,
based on video content. We propose a 'Video Propagation Network' that processes
video frames in an adaptive manner. The model is applied online: it propagates
information forward without the need to access future frames. In particular we
combine two components, a temporal bilateral network for dense and video
adaptive filtering, followed by a spatial network to refine features and
increased flexibility. We present experiments on video object segmentation and
semantic video segmentation and show increased performance comparing to the
best previous task-specific methods, while having favorable runtime.
Additionally we demonstrate our approach on an example regression task of color
propagation in a grayscale video. | ['Varun Jampani', 'Raghudeep Gadde', 'Peter V. Gehler'] | 2016-12-16 | video-propagation-networks-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Jampani_Video_Propagation_Networks_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Jampani_Video_Propagation_Networks_CVPR_2017_paper.pdf | cvpr-2017-7 | ['video-propagation'] | ['computer-vision'] | [ 4.20483738e-01 6.03015311e-02 4.69324738e-02 -6.14765167e-01
-4.76271123e-01 -6.68178797e-01 4.06509548e-01 7.36650229e-02
-8.66014123e-01 5.68195581e-01 1.00207977e-01 -1.60474181e-01
-2.23691016e-02 -5.54750502e-01 -7.93187082e-01 -5.07986248e-01
-3.11965972e-01 3.98199022e-01 1.28377128e+00 -3.08527797e-02
3.13393861e-01 5.71218967e-01 -1.34193015e+00 6.20215476e-01
4.23507750e-01 1.02146149e+00 3.91063511e-01 1.08934224e+00
-3.25137794e-01 1.24547684e+00 -2.25539237e-01 -2.13891149e-01
3.39964420e-01 -3.12785894e-01 -1.33785450e+00 4.36313212e-01
4.90405172e-01 -3.46671462e-01 -1.16502166e-01 9.77410376e-01
1.34018725e-02 5.53083420e-01 3.70337903e-01 -1.17230880e+00
-5.02364635e-01 6.73047245e-01 -5.49558818e-01 5.67177176e-01
5.04052103e-01 5.75900916e-03 6.86190546e-01 -8.35262299e-01
1.30092573e+00 1.31660724e+00 8.74674022e-01 6.13386869e-01
-1.34304130e+00 -2.39680469e-01 7.77205825e-01 5.33992887e-01
-1.09555769e+00 -4.74378705e-01 4.05869216e-01 -6.39287591e-01
9.29353952e-01 1.89878404e-01 8.39158297e-01 5.57687461e-01
-1.64643433e-02 1.00034797e+00 8.40278208e-01 -2.92477757e-01
2.68160254e-01 -4.50377539e-02 1.01376712e-01 1.00009263e+00
-4.10304368e-01 -1.50539875e-02 -5.69886267e-01 1.66227803e-01
8.89163852e-01 -4.21225131e-02 -3.08080405e-01 -3.85312587e-01
-1.06771600e+00 5.08136988e-01 3.96106839e-01 4.42307144e-01
-3.68724853e-01 4.40514445e-01 1.40604511e-01 4.42548335e-01
7.69640386e-01 2.93789878e-02 -7.14475036e-01 -9.58934203e-02
-1.54070663e+00 1.74401522e-01 8.58339548e-01 8.87988627e-01
1.04082632e+00 -1.13542013e-01 -2.03421786e-01 5.07511258e-01
5.16711652e-01 4.56731953e-02 1.92144915e-01 -1.67687857e+00
2.09231794e-01 2.03083068e-01 2.14041308e-01 -1.11854160e+00
-4.96442556e-01 2.01704383e-01 -2.78769732e-01 6.17325485e-01
5.19746304e-01 -2.79304832e-01 -1.28113556e+00 1.56069291e+00
5.32076240e-01 4.48985636e-01 -3.25465798e-01 9.62653577e-01
6.03633165e-01 8.98598909e-01 2.24807844e-01 -2.13936165e-01
9.53860521e-01 -1.29773748e+00 -6.19381189e-01 1.05026096e-01
2.19036072e-01 -7.83252299e-01 3.64357829e-01 5.56303144e-01
-1.58496404e+00 -6.51360273e-01 -4.98725355e-01 -2.41732970e-01
-5.37289023e-01 -2.95695603e-01 5.16924739e-01 5.35745502e-01
-1.86441898e+00 1.02051282e+00 -1.06968677e+00 -5.08301914e-01
4.76983815e-01 5.94812632e-01 -2.92546302e-01 9.75987017e-02
-8.34250748e-01 5.68039060e-01 5.41857302e-01 1.74410954e-01
-8.96967769e-01 -7.80477345e-01 -6.03195488e-01 -5.51317967e-02
5.52432120e-01 -1.04658747e+00 1.36724186e+00 -1.68603826e+00
-1.74405587e+00 5.59502304e-01 -2.47054741e-01 -5.97597182e-01
9.74173367e-01 -3.21622372e-01 -4.06043194e-02 6.39796555e-01
-1.45310283e-01 1.22216666e+00 1.26419783e+00 -1.12409735e+00
-1.27466142e+00 -5.70275486e-02 3.87522757e-01 2.02657491e-01
-1.60315961e-01 2.37662107e-01 -1.13447273e+00 -8.17211092e-01
3.34164612e-02 -9.08004165e-01 -6.20631039e-01 3.63393247e-01
-1.21258877e-01 -9.53473002e-02 1.06175923e+00 -8.37994039e-01
1.00180078e+00 -2.14381409e+00 5.05897701e-01 5.79602957e-01
3.38610172e-01 -1.41049758e-01 -1.40563279e-01 3.17256786e-02
9.39808134e-03 1.22478753e-01 -4.13231730e-01 -4.96984303e-01
-2.38080919e-01 9.88811105e-02 9.94924679e-02 4.81869310e-01
1.78155050e-01 7.26171672e-01 -9.19010401e-01 -6.56018913e-01
1.07246399e-01 5.00622392e-01 -7.45283782e-01 5.34375682e-02
-5.44489503e-01 4.03617799e-01 -2.67579108e-01 4.31051254e-01
4.89840746e-01 -3.94093812e-01 -5.08097336e-02 -1.81978077e-01
-4.54438776e-01 -4.37844455e-01 -1.45716834e+00 1.78837132e+00
-7.10674822e-02 8.36503804e-01 6.40274167e-01 -7.26220489e-01
4.22212452e-01 1.84451789e-01 8.78187835e-01 -2.13395044e-01
3.65249030e-02 -1.98464528e-01 -5.00705361e-01 -5.21038949e-01
7.32388616e-01 1.20347992e-01 3.53803158e-01 4.31305021e-01
1.98729590e-01 -1.36576332e-02 6.25660837e-01 5.95028400e-01
1.13429165e+00 6.48566961e-01 -2.92890817e-01 -3.86154950e-01
6.54781878e-01 1.92088440e-01 4.59466577e-01 7.96390235e-01
-1.19942173e-01 6.59508169e-01 2.13735595e-01 -3.66938591e-01
-7.56368697e-01 -1.00564075e+00 4.05214250e-01 1.69464540e+00
3.59312385e-01 -4.64897305e-01 -9.29464340e-01 -7.31871307e-01
-1.89575464e-01 1.86604992e-01 -7.17259288e-01 4.11818445e-01
-7.19405830e-01 -3.68662447e-01 -4.65194695e-02 7.97076464e-01
3.49451214e-01 -9.82428074e-01 -6.64097726e-01 3.28194022e-01
-2.05772873e-02 -1.22281003e+00 -6.89202070e-01 -2.60632783e-02
-1.01753402e+00 -1.17569733e+00 -8.21848989e-01 -8.44476700e-01
7.94966698e-01 1.57030657e-01 1.02933037e+00 2.82458752e-01
-2.08738357e-01 1.01919973e+00 -3.90513450e-01 1.76580414e-01
-3.22176039e-01 3.46558765e-02 -3.59011471e-01 2.24524304e-01
-1.67190164e-01 -1.98766112e-01 -7.51582026e-01 1.19756982e-01
-1.00786972e+00 2.28111342e-01 2.80585855e-01 4.53308970e-01
5.33028722e-01 8.15511346e-02 -9.35433656e-02 -1.35043728e+00
2.69485533e-01 -3.58617812e-01 -8.45528305e-01 2.57650375e-01
-3.97798777e-01 -1.44875377e-01 3.76158506e-01 -1.99603602e-01
-1.47353840e+00 5.80003738e-01 -9.01042894e-02 -4.23756987e-01
-1.82648361e-01 2.85671651e-01 4.54893261e-01 -4.82812583e-01
3.24688971e-01 -1.36739060e-01 -3.80716398e-02 -4.87009615e-01
8.32637548e-01 -9.84699205e-02 5.54381728e-01 -3.99912268e-01
7.62107432e-01 7.64420331e-01 -9.10745114e-02 -8.28429580e-01
-2.74455756e-01 -6.93483055e-01 -8.89328897e-01 -6.61853194e-01
1.26949167e+00 -6.07058167e-01 -5.86360514e-01 4.45665777e-01
-1.23616135e+00 -7.80346990e-01 -4.06890661e-01 3.17169398e-01
-7.97137856e-01 4.95647758e-01 -1.11839247e+00 -4.51740921e-01
-2.65587736e-02 -1.11386740e+00 8.28535914e-01 2.68516302e-01
-1.23230062e-01 -1.35403574e+00 -1.66543946e-01 6.20426610e-02
5.55035055e-01 -4.81579751e-02 4.74447221e-01 -3.28158975e-01
-1.16922987e+00 4.10465211e-01 -3.90688360e-01 2.63866395e-01
-7.33482093e-02 5.48606992e-01 -7.22773194e-01 -1.64399117e-01
-3.85188192e-01 2.96804588e-02 1.23173654e+00 7.18395770e-01
1.29303110e+00 -2.25990221e-01 -4.95213807e-01 7.29358077e-01
1.36988819e+00 1.84669301e-01 4.37187672e-01 3.19320142e-01
9.23585951e-01 7.82070458e-01 4.71949875e-01 2.75344580e-01
3.99509102e-01 4.65056658e-01 1.29600823e-01 -3.71171057e-01
-3.56797040e-01 1.71986863e-01 3.03898990e-01 5.95544517e-01
-5.70346117e-01 -1.72742754e-01 -7.34560430e-01 4.58475918e-01
-2.02261519e+00 -1.02066064e+00 -2.45269284e-01 1.57890344e+00
5.73801875e-01 1.22079097e-01 3.99071872e-01 -1.40248150e-01
7.27715433e-01 5.97527251e-02 -2.72660822e-01 -2.83937216e-01
1.62806913e-01 1.61143035e-01 6.92117214e-01 9.76683497e-01
-1.42522442e+00 1.30117953e+00 7.57445955e+00 5.23542047e-01
-8.79743755e-01 3.39815021e-01 6.64509773e-01 -4.84373681e-02
-2.69477606e-01 -4.95072901e-02 -6.15901887e-01 3.30277503e-01
6.83450699e-01 7.05729052e-02 4.22683179e-01 5.56533217e-01
3.25091273e-01 -2.67566562e-01 -1.05442894e+00 7.43818939e-01
1.17868492e-02 -1.62382770e+00 -1.40538672e-02 -5.87003171e-01
7.41315246e-01 2.33246721e-02 7.29903160e-03 -2.01585680e-01
5.72374344e-01 -4.79701728e-01 1.08137488e+00 8.51083279e-01
3.54135990e-01 -4.77606088e-01 2.14315280e-01 -1.60930604e-02
-1.35801399e+00 -2.24308699e-01 1.23737991e-01 1.75451845e-01
4.90825593e-01 2.82389492e-01 -5.62535584e-01 2.39592195e-01
1.22858119e+00 9.69501197e-01 -6.65060639e-01 1.29165971e+00
-1.95719916e-02 3.54919970e-01 -4.35739696e-01 1.66563243e-01
4.95283484e-01 -2.83855766e-01 4.67813671e-01 1.72017205e+00
1.21493392e-01 2.22467944e-01 4.88126487e-01 5.72427928e-01
6.54421821e-02 9.55207571e-02 -1.74358442e-01 3.00602406e-01
1.07515985e-02 1.18081748e+00 -1.51313150e+00 -6.42573774e-01
-4.94201571e-01 1.35094166e+00 7.24038035e-02 7.36567616e-01
-7.57914186e-01 -3.49820554e-02 2.69367635e-01 1.02937259e-01
7.71258771e-01 -3.58606011e-01 1.03295416e-01 -8.79214942e-01
-3.19750756e-01 -3.59920651e-01 5.81779599e-01 -8.61811042e-01
-1.06494474e+00 6.23504758e-01 1.67037994e-01 -7.37422109e-01
-2.41554886e-01 -6.57663822e-01 -1.44505247e-01 4.05572087e-01
-1.51715875e+00 -9.37346339e-01 -3.37193787e-01 1.08526266e+00
8.83652210e-01 2.02658296e-01 2.13709548e-01 4.11484450e-01
-2.44391426e-01 1.19455894e-02 -2.97737978e-02 -6.00954592e-02
4.29296613e-01 -1.46560168e+00 3.19525033e-01 1.06161320e+00
1.77594483e-01 2.97452062e-01 6.62339211e-01 -6.70553327e-01
-1.00916028e+00 -1.16408849e+00 5.66507161e-01 -2.92154610e-01
7.56757379e-01 -1.78478122e-01 -8.82118165e-01 8.62440944e-01
4.53429580e-01 4.44328822e-02 3.97008568e-01 -1.39561191e-01
1.11110955e-01 -1.79557294e-01 -1.02670407e+00 4.45426404e-01
1.21679139e+00 -2.00683370e-01 -4.19591039e-01 5.12002349e-01
6.61313295e-01 -5.16564429e-01 -6.24240220e-01 8.07726160e-02
4.66134906e-01 -9.38432932e-01 1.07907236e+00 -4.24487472e-01
1.19208440e-01 -4.82147157e-01 2.00618416e-01 -1.03289938e+00
-5.95096290e-01 -9.02752340e-01 -6.57176822e-02 9.41971302e-01
5.48926830e-01 -2.18354270e-01 9.41376984e-01 8.69578779e-01
-1.54815922e-02 -1.90720677e-01 -6.71412587e-01 -5.28117180e-01
-5.11250675e-01 -6.02866232e-01 3.81009988e-02 6.79003596e-01
-3.03352416e-01 -5.59547134e-02 -2.51519501e-01 1.70306891e-01
6.07620716e-01 1.05785243e-02 3.17766756e-01 -1.07296598e+00
-3.59347492e-01 -6.03853881e-01 -4.35393959e-01 -1.32550657e+00
1.59931704e-01 -6.65028751e-01 2.98139542e-01 -1.67014194e+00
9.01255235e-02 -3.67287397e-01 -3.65560114e-01 4.88788098e-01
6.29179552e-02 5.12306035e-01 5.41319907e-01 5.16377725e-02
-1.16417038e+00 -2.11919583e-02 1.00429916e+00 -1.46660388e-01
-3.65309149e-01 -6.99052960e-02 -9.47219357e-02 9.97901857e-01
5.67507625e-01 -4.69518095e-01 -4.84488815e-01 -6.79741502e-01
1.39740795e-01 4.49667647e-02 3.36185247e-01 -8.71307969e-01
7.59846807e-01 -1.03191063e-01 4.97946173e-01 -4.83873904e-01
3.58757555e-01 -1.15216041e+00 2.94077456e-01 3.68458211e-01
-4.09492344e-01 2.60880500e-01 7.56186023e-02 8.04074228e-01
-2.43598506e-01 -2.37192839e-01 7.13780344e-01 -4.27926362e-01
-1.53445971e+00 5.17098606e-01 -8.86177897e-01 -2.46333927e-01
1.31389678e+00 -4.39071059e-01 2.85807271e-02 -3.38748515e-01
-1.51087606e+00 2.74571627e-01 4.91684347e-01 4.34500277e-01
7.90208459e-01 -9.38802838e-01 -4.61884409e-01 1.51468441e-01
-4.98411059e-01 -1.97117284e-01 3.34611297e-01 9.04671431e-01
-1.01828718e+00 -2.38967072e-02 -1.05405480e-01 -7.95342326e-01
-1.55689752e+00 7.23502457e-01 2.90650427e-01 2.33692244e-01
-8.74098361e-01 1.16418445e+00 9.91893113e-02 3.52035850e-01
3.11750472e-01 -5.04805744e-01 -4.44710255e-01 2.76046038e-01
6.42298102e-01 5.67643821e-01 -1.99833915e-01 -5.73258579e-01
-3.94568771e-01 7.10412502e-01 -2.96732187e-02 -4.68587339e-01
1.34593093e+00 -6.91915929e-01 -3.51124227e-01 4.37325299e-01
9.60969448e-01 -1.74939945e-01 -1.82960892e+00 -9.22200233e-02
2.99912214e-01 -6.75742269e-01 1.42775923e-01 -6.05032742e-01
-1.59784746e+00 5.00682592e-01 6.51176035e-01 5.02696991e-01
1.34381592e+00 1.02031097e-01 6.34178162e-01 1.93962857e-01
2.30822667e-01 -1.45876086e+00 1.16080701e-01 4.64976519e-01
4.65719998e-01 -9.60581601e-01 -1.47567708e-02 -7.44677365e-01
-3.49390745e-01 1.51466286e+00 4.03556526e-01 -1.85489357e-01
9.87955093e-01 5.45278609e-01 1.50783703e-01 -1.09527081e-01
-7.44726658e-01 -3.70575547e-01 2.98964530e-01 5.61842442e-01
2.27881610e-01 -3.87495041e-01 -2.04548970e-01 -8.76795128e-02
3.72586131e-01 3.35816354e-01 4.20816213e-01 9.94436681e-01
-5.50282836e-01 -9.37736154e-01 -2.98764408e-01 1.73691034e-01
-6.14308357e-01 -9.89618897e-02 -2.09639654e-01 6.77008271e-01
2.66260564e-01 8.66601467e-01 2.29601502e-01 -8.21854174e-02
9.53321755e-02 -4.84057367e-02 6.54512882e-01 -4.52002943e-01
-6.35142684e-01 3.16499591e-01 2.42577288e-02 -1.16998422e+00
-1.22369933e+00 -8.21914494e-01 -1.27491474e+00 -1.37070045e-01
-3.54167446e-02 4.61772829e-02 4.86582756e-01 1.00636792e+00
1.34207919e-01 6.70671999e-01 2.98364431e-01 -1.10434234e+00
4.03820872e-01 -3.84359479e-01 -4.43156451e-01 5.16567647e-01
4.35065538e-01 -4.21861261e-01 -4.13828976e-02 8.67186844e-01] | [9.079484939575195, -0.26917916536331177] |
898eda53-e17c-4acc-a748-924527ecd568 | residual-shuffle-exchange-networks-for-fast | 2004.04662 | null | https://arxiv.org/abs/2004.04662v4 | https://arxiv.org/pdf/2004.04662v4.pdf | Residual Shuffle-Exchange Networks for Fast Processing of Long Sequences | Attention is a commonly used mechanism in sequence processing, but it is of O(n^2) complexity which prevents its application to long sequences. The recently introduced neural Shuffle-Exchange network offers a computation-efficient alternative, enabling the modelling of long-range dependencies in O(n log n) time. The model, however, is quite complex, involving a sophisticated gating mechanism derived from the Gated Recurrent Unit. In this paper, we present a simple and lightweight variant of the Shuffle-Exchange network, which is based on a residual network employing GELU and Layer Normalization. The proposed architecture not only scales to longer sequences but also converges faster and provides better accuracy. It surpasses the Shuffle-Exchange network on the LAMBADA language modelling task and achieves state-of-the-art performance on the MusicNet dataset for music transcription while being efficient in the number of parameters. We show how to combine the improved Shuffle-Exchange network with convolutional layers, establishing it as a useful building block in long sequence processing applications. | ['Kārlis Freivalds', 'Matīss Apinis', 'Agris Šostaks', 'Emīls Ozoliņš', 'Andis Draguns'] | 2020-04-06 | null | null | null | null | ['music-transcription', 'lambada'] | ['music', 'natural-language-processing'] | [ 6.14014745e-01 -3.40553105e-01 1.14418916e-01 -2.04013512e-01
-6.90990686e-01 -5.47767937e-01 3.85207266e-01 1.69502556e-01
-9.34971809e-01 6.37829840e-01 9.42372009e-02 -4.41629738e-01
-8.28686059e-02 -4.01993394e-01 -7.66916394e-01 -8.87317121e-01
-2.34209165e-01 4.24389571e-01 2.68088758e-01 -3.51149112e-01
2.79387742e-01 5.08741975e-01 -1.61546171e+00 5.78125358e-01
5.26749611e-01 9.34762120e-01 6.45333886e-01 1.10994232e+00
-9.83269438e-02 9.25276041e-01 -4.81004059e-01 -2.64172226e-01
9.19527337e-02 -6.00635707e-01 -8.22033048e-01 -3.53586465e-01
3.43319237e-01 2.17380792e-01 -4.32938561e-02 7.02865660e-01
1.00661969e+00 3.62422645e-01 1.80884361e-01 -6.44594789e-01
-4.24655825e-01 9.96147633e-01 -3.48090053e-01 6.75419390e-01
6.44870326e-02 -1.79410160e-01 1.28035724e+00 -7.41860986e-01
4.96856868e-01 9.66254294e-01 9.06190157e-01 4.18884873e-01
-1.32270181e+00 -6.81773067e-01 7.16902018e-02 4.06511217e-01
-1.33863294e+00 -5.25775075e-01 4.25972253e-01 -2.37931371e-01
1.62808681e+00 3.01266909e-01 5.01620531e-01 1.07645762e+00
7.94601291e-02 9.77222621e-01 7.58269608e-01 -5.64378858e-01
-1.68417692e-02 -5.04347086e-01 2.19640225e-01 4.23567712e-01
-1.16519123e-01 -2.98939049e-02 -7.29064941e-01 -3.36822309e-02
8.33475053e-01 -2.06999168e-01 -4.01105165e-01 -7.44484141e-02
-1.36163175e+00 6.28923774e-01 5.54834723e-01 6.75426602e-01
-4.70874429e-01 4.97933298e-01 7.78091431e-01 6.51035964e-01
3.74885231e-01 4.29007798e-01 -5.30538559e-01 -6.03660524e-01
-1.22904813e+00 2.40102187e-01 7.55545318e-01 6.93441451e-01
3.61340135e-01 2.42446601e-01 -1.51821122e-01 9.46438849e-01
-2.89639104e-02 1.85927358e-02 7.05632389e-01 -4.71530944e-01
4.78755325e-01 2.15161487e-01 -3.00406009e-01 -6.87435150e-01
-7.49363363e-01 -8.81763339e-01 -1.23705077e+00 1.28866658e-01
2.91188717e-01 5.67917153e-02 -8.29647660e-01 1.86539459e+00
-2.23973319e-01 6.06303632e-01 -1.61777511e-01 8.15290213e-01
5.83785713e-01 7.28797853e-01 -2.12211944e-02 -2.19984278e-01
1.17725837e+00 -1.20043802e+00 -6.03764772e-01 -1.50972217e-01
6.54148221e-01 -8.09550881e-01 1.11575592e+00 6.44042552e-01
-1.28738546e+00 -7.12728381e-01 -1.13457966e+00 -2.69140989e-01
-2.37773150e-01 4.33447994e-02 6.60382807e-01 5.39535999e-01
-1.38384044e+00 1.03907371e+00 -7.18683839e-01 -3.45273465e-01
3.77270468e-02 7.82918692e-01 -3.01222116e-01 3.53059798e-01
-1.20861340e+00 7.11314678e-01 6.74575269e-01 4.75434363e-01
-4.96968806e-01 -6.41780198e-01 -7.64651954e-01 3.80192369e-01
2.00824931e-01 -7.01677263e-01 1.22111738e+00 -9.09660220e-01
-1.78267610e+00 6.46245956e-01 -2.20768139e-01 -9.50351834e-01
3.55839670e-01 -3.28525513e-01 -2.40853786e-01 -1.15818664e-01
-4.91760731e-01 5.38120449e-01 7.98064649e-01 -4.45735842e-01
-4.82849240e-01 -2.31391191e-01 -2.45822355e-01 9.64612067e-02
-3.85294147e-02 2.10793376e-01 -4.00930524e-01 -1.02481937e+00
-1.31377056e-01 -1.13881779e+00 -4.02070165e-01 -5.83691478e-01
-7.23811314e-02 -1.34819001e-01 4.88341749e-01 -6.21149540e-01
1.49000371e+00 -2.09700012e+00 3.98912489e-01 1.74982548e-01
-1.20002471e-01 6.83727622e-01 -4.08548176e-01 5.89110851e-01
-4.00268614e-01 -4.52860113e-04 -5.33681273e-01 -4.52804208e-01
-6.67720810e-02 2.60152668e-01 -2.85808265e-01 3.29553634e-01
1.76070362e-01 1.10479975e+00 -7.15047657e-01 1.32534206e-01
1.36245608e-01 7.53368914e-01 -6.84900165e-01 5.88369854e-02
-9.35269892e-02 5.46964586e-01 1.56992063e-01 5.26928380e-02
4.28077310e-01 -2.88597614e-01 1.60793468e-01 2.36464545e-01
-2.04242319e-01 6.15960956e-01 -1.09307873e+00 2.25655127e+00
-6.63554788e-01 7.81714022e-01 2.82779634e-02 -9.48204398e-01
1.12652111e+00 4.81758654e-01 3.30239177e-01 -4.83131945e-01
2.56973058e-01 4.81532037e-01 3.85850459e-01 -3.84348154e-01
4.75891739e-01 -2.41835967e-01 -4.32204567e-02 5.07312536e-01
3.32608938e-01 1.81527495e-01 4.72611308e-01 -1.28786519e-01
1.33238769e+00 2.67316818e-01 3.39267701e-01 -4.48731720e-01
7.98753262e-01 -7.49481618e-01 4.22771990e-01 8.27220440e-01
1.13164820e-01 7.28773654e-01 4.97153044e-01 -6.26905501e-01
-1.13764060e+00 -6.16497159e-01 1.69479594e-01 1.49185693e+00
-4.09121066e-01 -5.73362529e-01 -5.94129264e-01 -2.07815439e-01
-2.71211684e-01 4.53986675e-01 -5.59704304e-01 3.91815677e-02
-9.74932432e-01 -9.39171255e-01 8.51165056e-01 7.26733029e-01
3.22777420e-01 -1.61896133e+00 -6.65446281e-01 8.04398835e-01
-2.86419779e-01 -9.48063195e-01 -4.71820593e-01 6.85751855e-01
-8.57691705e-01 -7.30555236e-01 -7.64261186e-01 -8.64922762e-01
-1.06342440e-03 -2.90908128e-01 1.33043444e+00 6.40748292e-02
-2.15920195e-01 -2.41639823e-01 -5.05010962e-01 -3.11371803e-01
-2.61275470e-01 8.69132161e-01 -2.15188354e-01 -5.86438924e-02
3.28544140e-01 -8.91680896e-01 -4.09972876e-01 -2.57076841e-04
-1.02936399e+00 -2.98623405e-02 7.61215270e-01 1.14057851e+00
4.37929064e-01 -3.19373250e-01 9.57139969e-01 -1.01427400e+00
6.92543507e-01 -4.35865000e-02 -4.97884661e-01 2.24930607e-02
-4.34229583e-01 3.34760159e-01 7.12887585e-01 -3.42266113e-01
-7.08793104e-01 2.82883167e-01 -7.88679421e-01 -2.25246266e-01
2.91414987e-02 5.82918167e-01 7.95020312e-02 -1.14618994e-01
3.95526916e-01 2.17425242e-01 -1.57139406e-01 -8.78032982e-01
2.34057665e-01 5.33141494e-01 6.67661846e-01 -1.78579703e-01
3.21668535e-01 9.97023731e-02 -7.20864721e-03 -9.04792428e-01
-6.08426571e-01 -6.84762061e-01 -8.70779216e-01 2.19662532e-01
7.85621166e-01 -6.04982436e-01 -1.02036822e+00 5.83417952e-01
-1.17651200e+00 -3.99111509e-01 -1.68584645e-01 5.51556766e-01
-7.73396969e-01 2.24661067e-01 -1.03573978e+00 -7.25937247e-01
-6.74328983e-01 -8.88314784e-01 1.02790761e+00 -1.23840712e-01
-1.80963829e-01 -8.76415074e-01 2.07047313e-01 -9.44572315e-02
6.19001746e-01 -7.04406276e-02 9.50175285e-01 -7.37004578e-01
-4.35149491e-01 -6.58043399e-02 -1.86159819e-01 4.93400931e-01
-3.57571036e-01 -3.80398005e-01 -1.14760721e+00 -5.04365265e-01
-1.13415092e-01 -1.60292745e-01 1.39002025e+00 2.91173667e-01
1.30059826e+00 2.71206740e-02 2.42151037e-01 9.36953723e-01
1.39065075e+00 3.43522243e-02 8.56532991e-01 4.26060468e-01
5.89383006e-01 2.40518615e-01 6.66542500e-02 3.39821845e-01
-9.65468287e-02 5.95234096e-01 3.48071277e-01 -1.52113169e-01
-1.44141555e-01 -3.56132872e-02 3.43517989e-01 1.41245782e+00
-4.40732002e-01 -3.54828238e-01 -8.28696191e-01 4.86849815e-01
-2.07524514e+00 -1.08596921e+00 -1.92286849e-01 2.22248125e+00
6.63494885e-01 2.49536306e-01 2.71098893e-02 5.17820418e-01
4.85956043e-01 4.24326032e-01 -1.85912102e-01 -1.03405273e+00
-2.93726534e-01 7.79096782e-01 5.92195272e-01 5.23952186e-01
-1.05323231e+00 9.17485893e-01 6.78417492e+00 8.90466154e-01
-1.16217709e+00 1.31443620e-01 2.11339414e-01 -1.76915482e-01
1.31433219e-01 -3.71631205e-01 -8.57239783e-01 2.95446783e-01
1.39997625e+00 1.39506847e-01 5.23467779e-01 4.86128747e-01
9.46296453e-02 2.58256793e-01 -1.02037847e+00 1.12932074e+00
2.74870336e-01 -1.34892154e+00 -3.40870349e-03 -1.07851759e-01
4.40016240e-01 4.34333593e-01 -2.26383626e-01 3.23523790e-01
-1.00228123e-01 -1.15009046e+00 4.97313797e-01 3.06674004e-01
6.31907463e-01 -1.15398407e+00 1.00095677e+00 4.31217313e-01
-1.39077330e+00 -2.14012533e-01 -4.70004857e-01 -3.74155432e-01
3.77512366e-01 5.42958260e-01 -8.57615650e-01 4.94221449e-01
6.79336429e-01 7.82431126e-01 -5.37695169e-01 1.19174230e+00
-6.76998273e-02 6.27674222e-01 -2.20445126e-01 -2.18056124e-02
5.27318597e-01 -1.78175017e-01 4.60011542e-01 1.81478894e+00
4.47102457e-01 -2.63223380e-01 -6.35849237e-02 3.04002285e-01
-1.70082718e-01 2.50698149e-01 -4.33920681e-01 -1.13255545e-01
-9.35159922e-02 1.10495186e+00 -7.34343469e-01 -3.43105197e-01
-3.06125075e-01 1.08996522e+00 5.17658293e-01 2.42355451e-01
-7.14198887e-01 -7.17755616e-01 5.70174158e-01 -1.12348825e-01
9.91518319e-01 -3.23146731e-01 -1.25223592e-01 -1.06674314e+00
-5.12035675e-02 -8.25106323e-01 3.69623929e-01 -6.37999713e-01
-1.12016547e+00 1.13007104e+00 -4.67035383e-01 -8.88246536e-01
-6.21001959e-01 -8.10368598e-01 -2.42076054e-01 1.04891884e+00
-1.67576003e+00 -9.53801274e-01 -2.97998614e-03 5.14152050e-01
5.83695590e-01 -1.75187606e-02 1.14084017e+00 6.21295810e-01
-3.36900711e-01 5.86415768e-01 1.90690652e-01 4.09313813e-02
5.13866723e-01 -1.25920248e+00 8.79464626e-01 7.60727882e-01
5.31850040e-01 5.33372402e-01 6.09393895e-01 -2.47750938e-01
-1.04123032e+00 -9.95576680e-01 1.31109250e+00 -5.42266108e-02
7.44323850e-01 -7.27060139e-01 -1.25218880e+00 7.37934589e-01
4.73797292e-01 -8.61718580e-02 7.86764503e-01 2.17883468e-01
-3.69023710e-01 -2.29537673e-02 -5.10500014e-01 4.22000676e-01
1.26081955e+00 -6.73279524e-01 -5.65209627e-01 -8.90564099e-02
6.06875718e-01 -3.68626505e-01 -6.62533402e-01 5.51598489e-01
6.22839689e-01 -1.04622519e+00 7.86908329e-01 -4.98865753e-01
-9.82056651e-03 -2.81257123e-01 1.13070130e-01 -1.22780347e+00
-4.96412724e-01 -1.04799318e+00 3.64142768e-02 9.91692007e-01
5.76622128e-01 -5.86136222e-01 7.03977585e-01 -3.15008998e-01
-2.63993591e-01 -7.31215537e-01 -1.02278364e+00 -8.90541494e-01
-9.10014138e-02 -5.63094378e-01 4.58304882e-01 6.86527550e-01
-1.53806433e-01 5.42950988e-01 -6.33961439e-01 -2.04260826e-01
2.90021390e-01 1.58196732e-01 4.29244637e-01 -1.11565685e+00
-5.92828870e-01 -6.47354782e-01 -6.72450006e-01 -1.33081043e+00
2.72777468e-01 -1.13747895e+00 2.30096146e-01 -1.25973904e+00
-4.53635678e-02 -2.14133874e-01 -7.59057164e-01 3.05941582e-01
-4.96391766e-02 6.50711596e-01 4.41135466e-01 -1.50968805e-02
-4.47713971e-01 3.51253748e-01 8.38775218e-01 1.26035675e-01
-2.44766623e-01 9.56187621e-02 -3.61020684e-01 6.48125470e-01
8.28513384e-01 -4.11176413e-01 -2.35410735e-01 -5.96357226e-01
5.00136137e-01 -2.94324040e-01 1.09652214e-01 -1.10776079e+00
3.78990203e-01 6.07052386e-01 2.79401124e-01 -7.58739173e-01
3.20410192e-01 -6.01148903e-01 1.87432855e-01 4.23154265e-01
-6.16548657e-01 5.30088603e-01 3.81354004e-01 3.72466832e-01
-4.14995909e-01 -3.83812487e-01 7.06660688e-01 -2.63379782e-01
-7.95547485e-01 1.66193262e-01 -3.88264388e-01 -2.18540758e-01
5.52970052e-01 -1.66098699e-02 5.33421822e-02 -2.20361963e-01
-9.82888818e-01 -1.20671727e-01 1.20386861e-01 5.11218190e-01
3.21499050e-01 -1.14635086e+00 -9.51406360e-01 5.38805008e-01
-1.76788062e-01 -1.32597357e-01 1.91421926e-01 9.83033836e-01
-6.67065799e-01 7.10296035e-01 -2.05854967e-01 -6.77767694e-01
-1.40459037e+00 5.79164505e-01 1.24048717e-01 -7.68934309e-01
-8.21178973e-01 1.11239898e+00 3.28928232e-02 -4.06774342e-01
4.46602553e-01 -5.11161149e-01 -2.09770933e-01 1.57724068e-01
9.52249944e-01 2.94553369e-01 5.22585809e-01 -6.26379430e-01
-3.96591544e-01 6.59893692e-01 -2.63498761e-02 -1.32869408e-01
1.56277680e+00 -1.15567952e-01 -3.57494563e-01 7.81793177e-01
1.20735908e+00 -1.35239810e-01 -1.04735959e+00 -1.29187226e-01
5.19605935e-01 -1.84297472e-01 -1.19862124e-01 -6.81954145e-01
-8.84863555e-01 1.16311824e+00 4.52361792e-01 1.48513868e-01
1.35637522e+00 -1.94335282e-01 8.64236057e-01 4.45497870e-01
2.82191515e-01 -7.90144444e-01 -3.49906653e-01 1.05193591e+00
1.05054927e+00 -8.05092275e-01 -3.47359866e-01 -1.49407193e-01
-3.53973508e-01 1.20160973e+00 1.68624848e-01 -3.85637760e-01
5.21539569e-01 5.63953042e-01 8.68007913e-02 1.06346875e-01
-1.03871584e+00 -3.55083913e-01 1.67377785e-01 3.05673540e-01
9.39379692e-01 -1.31803244e-01 -4.13410455e-01 3.24562013e-01
-3.46253753e-01 2.20841035e-01 3.02456141e-01 7.59316087e-01
-3.18816990e-01 -1.39841712e+00 -7.89687503e-03 2.57590622e-01
-7.71528304e-01 -5.36343634e-01 -1.43028483e-01 6.78988814e-01
6.38919994e-02 5.60073316e-01 1.68033436e-01 -2.48869359e-01
3.30637217e-01 4.00023550e-01 5.79752445e-01 -4.97668028e-01
-1.20854986e+00 3.36076736e-01 2.59283576e-02 -6.77858770e-01
-6.94211841e-01 -5.00262082e-01 -1.11340845e+00 -2.46621132e-01
-2.95808554e-01 1.32883579e-01 7.20992923e-01 7.89926469e-01
4.91834491e-01 8.94211352e-01 1.70861155e-01 -1.11179113e+00
-3.88741523e-01 -1.18351495e+00 -6.55771673e-01 3.38404894e-01
5.14387131e-01 -1.75773144e-01 -3.63317803e-02 1.10760564e-02] | [10.90872859954834, 6.537520885467529] |
654cb2f1-9ac0-4b8b-88b1-cfcb229525e5 | negative-sample-matters-a-renaissance-of | 2109.04872 | null | https://arxiv.org/abs/2109.04872v2 | https://arxiv.org/pdf/2109.04872v2.pdf | Negative Sample Matters: A Renaissance of Metric Learning for Temporal Grounding | Temporal grounding aims to localize a video moment which is semantically aligned with a given natural language query. Existing methods typically apply a detection or regression pipeline on the fused representation with the research focus on designing complicated prediction heads or fusion strategies. Instead, from a perspective on temporal grounding as a metric-learning problem, we present a Mutual Matching Network (MMN), to directly model the similarity between language queries and video moments in a joint embedding space. This new metric-learning framework enables fully exploiting negative samples from two new aspects: constructing negative cross-modal pairs in a mutual matching scheme and mining negative pairs across different videos. These new negative samples could enhance the joint representation learning of two modalities via cross-modal mutual matching to maximize their mutual information. Experiments show that our MMN achieves highly competitive performance compared with the state-of-the-art methods on four video grounding benchmarks. Based on MMN, we present a winner solution for the HC-STVG challenge of the 3rd PIC workshop. This suggests that metric learning is still a promising method for temporal grounding via capturing the essential cross-modal correlation in a joint embedding space. Code is available at https://github.com/MCG-NJU/MMN. | ['Gangshan Wu', 'TianHao Li', 'Tao Wu', 'LiMin Wang', 'Zhenzhi Wang'] | 2021-09-10 | null | null | null | null | ['video-grounding'] | ['computer-vision'] | [-8.89118984e-02 -1.82006985e-01 -5.98087132e-01 -4.12603408e-01
-1.19838977e+00 -3.72568876e-01 6.25304997e-01 4.27975953e-02
-3.41777831e-01 7.92009756e-02 4.46827501e-01 2.06977203e-01
-1.61142007e-01 -5.97614586e-01 -6.19030237e-01 -4.53544348e-01
-3.58839184e-01 1.83406442e-01 2.51311123e-01 -1.44276783e-01
1.17777407e-01 -5.12850098e-02 -1.63167191e+00 6.97851539e-01
3.26180756e-01 1.13715756e+00 2.00594455e-01 5.93634784e-01
-8.87585133e-02 9.89864945e-01 -6.59272224e-02 -4.80804801e-01
3.30771148e-01 -6.72924519e-01 -8.30978036e-01 -1.58967569e-01
5.85717261e-01 -1.21481106e-01 -7.81132162e-01 1.03598797e+00
3.58573079e-01 2.81035215e-01 3.77913177e-01 -1.77966928e+00
-4.67460752e-01 5.63031197e-01 -6.69044614e-01 2.58903861e-01
7.67456472e-01 -9.56488922e-02 1.56873178e+00 -1.13570833e+00
8.55637670e-01 1.19967687e+00 6.44315124e-01 4.60474879e-01
-1.11876261e+00 -5.67367554e-01 1.90924391e-01 8.56908143e-01
-1.59182560e+00 -2.47866809e-01 9.33247507e-01 -6.55603290e-01
7.82065034e-01 3.80650073e-01 7.83355117e-01 1.18808699e+00
-2.41318010e-02 1.25561404e+00 8.21146667e-01 -2.58392930e-01
-1.03808396e-01 -8.57670456e-02 -1.39549494e-01 9.19741571e-01
-3.91688079e-01 1.46407366e-01 -1.27848780e+00 -1.87458575e-01
4.35179800e-01 2.13797435e-01 -3.59507084e-01 -8.22556078e-01
-1.92985260e+00 8.74394000e-01 4.15898323e-01 6.42654002e-01
-1.92672852e-02 1.94914311e-01 4.05654490e-01 5.42070746e-01
4.20973986e-01 4.39205915e-01 -3.90249610e-01 -2.06668153e-01
-1.10959566e+00 2.65525460e-01 5.84437728e-01 9.65087712e-01
1.12059200e+00 -5.76365769e-01 -5.00062406e-01 5.21151900e-01
5.09179533e-01 2.86396176e-01 5.19156873e-01 -1.08694744e+00
6.85419798e-01 7.21651495e-01 -6.30982444e-02 -1.48237336e+00
-2.30647564e-01 -6.57615587e-02 -5.97452700e-01 -1.65579453e-01
3.40437174e-01 1.50255293e-01 -4.48920757e-01 1.99562263e+00
3.51238310e-01 7.59501278e-01 -1.81444988e-01 9.89538610e-01
8.71763706e-01 5.01425266e-01 -2.46765286e-01 -9.85880718e-02
1.20928049e+00 -9.84820127e-01 -6.72207832e-01 1.24907661e-02
1.04222965e+00 -8.67951334e-01 9.86660361e-01 2.28290737e-01
-7.82477081e-01 -5.22601902e-01 -1.07421851e+00 -7.93254450e-02
-3.98396492e-01 1.34298224e-02 4.89263177e-01 5.81608862e-02
-1.06494331e+00 7.38753736e-01 -8.76991332e-01 -6.05120063e-01
1.16623826e-01 1.72273576e-01 -6.86351895e-01 -1.81297332e-01
-1.38194442e+00 6.60443544e-01 2.49378353e-01 -7.82123357e-02
-6.07826948e-01 -7.53370643e-01 -1.21586573e+00 -4.56503391e-01
3.09649199e-01 -7.48268545e-01 9.76537943e-01 -6.57002151e-01
-9.73480880e-01 1.30816150e+00 -3.50994915e-01 -5.40193439e-01
5.12281895e-01 -3.08727652e-01 -6.42409682e-01 3.55234534e-01
3.50184590e-01 8.65764081e-01 7.64373362e-01 -9.53711629e-01
-7.12363303e-01 -3.15978050e-01 8.26788694e-02 1.71945289e-01
-4.82121557e-01 1.66543182e-02 -8.31969500e-01 -9.24763978e-01
5.14476120e-01 -9.46379364e-01 -8.65127519e-02 1.11316644e-01
-2.75454491e-01 -3.76195639e-01 9.95671451e-01 -5.71227789e-01
1.53583384e+00 -2.18221092e+00 5.82669020e-01 2.83809692e-01
3.59707773e-01 -3.62443686e-01 -2.21961915e-01 5.51830113e-01
-1.08749762e-01 -4.20985632e-02 6.42416477e-02 -5.83004951e-01
3.04157287e-01 4.43883874e-02 -2.60875374e-01 7.05951214e-01
1.01907551e-01 9.12882447e-01 -1.28504908e+00 -7.57902026e-01
2.70510912e-01 3.83460581e-01 -5.99921882e-01 5.41424870e-01
-1.16314292e-01 4.20141876e-01 -1.31866992e-01 7.53719747e-01
3.04181367e-01 -3.60727519e-01 1.79859906e-01 -8.82040441e-01
1.95452243e-01 2.13403394e-03 -1.19154704e+00 2.56100440e+00
-1.20474204e-01 8.35666001e-01 -3.47997040e-01 -1.11605322e+00
7.06362844e-01 3.48943919e-01 1.12615848e+00 -8.33832085e-01
-1.22433253e-01 1.40031904e-01 -3.80234957e-01 -6.72231734e-01
5.26299536e-01 9.87966731e-02 -1.87477738e-01 4.74427134e-01
5.50323308e-01 9.87881273e-02 3.43174011e-01 5.36425233e-01
1.09642351e+00 3.60503793e-01 3.32546979e-02 -1.07547073e-02
4.14381653e-01 -3.10896993e-01 7.80676782e-01 4.91042405e-01
-3.46633971e-01 8.26484382e-01 4.28003162e-01 -3.42364520e-01
-8.17032576e-01 -1.08934748e+00 2.05964521e-01 1.25108182e+00
4.61498111e-01 -1.19526374e+00 -3.84324193e-01 -8.34201455e-01
-1.29272074e-01 2.36363143e-01 -9.81301546e-01 -3.32717985e-01
-4.81105059e-01 -2.53076017e-01 3.82978976e-01 3.50679278e-01
2.85192162e-01 -7.08884418e-01 -2.65848130e-01 -9.70553458e-02
-8.00499380e-01 -1.24876380e+00 -8.25312555e-01 -3.18257213e-02
-6.81007862e-01 -1.32880509e+00 -6.41119301e-01 -7.62381554e-01
1.22438379e-01 4.50432450e-01 1.30310190e+00 -2.44140327e-02
-2.76212871e-01 9.02900219e-01 -5.20826161e-01 7.79403970e-02
4.30251546e-02 -1.05896883e-01 1.11537114e-01 4.03478354e-01
7.33592331e-01 -5.65859199e-01 -7.90560603e-01 5.45523465e-01
-7.18512833e-01 1.83723152e-01 2.34594062e-01 7.62088954e-01
7.87352324e-01 -4.65792269e-01 8.22342187e-02 -2.99065769e-01
6.44183531e-02 -7.77541697e-01 -4.01138753e-01 3.99538517e-01
-3.28202099e-01 1.94169149e-01 -1.36862531e-01 -4.28752214e-01
-4.95940089e-01 6.14049993e-02 2.34460279e-01 -9.57715988e-01
1.66958153e-01 6.13370776e-01 -1.04144074e-01 1.33329421e-01
5.07264733e-01 1.23323977e-01 -9.13168713e-02 -3.20497692e-01
5.32140851e-01 1.62901267e-01 5.76224387e-01 -6.13900185e-01
8.67118478e-01 7.72588968e-01 -2.95971725e-02 -4.00992721e-01
-1.06259108e+00 -1.18821073e+00 -7.20310032e-01 -4.44230109e-01
1.02644038e+00 -1.07661045e+00 -5.12390971e-01 1.19652264e-01
-9.76017356e-01 -2.38250077e-01 -3.27529699e-01 9.00750399e-01
-9.21321213e-01 6.13833129e-01 -4.26790744e-01 -3.69931370e-01
-1.65786520e-02 -9.35119629e-01 1.35515666e+00 -2.04775870e-01
-4.83502418e-01 -1.09688830e+00 5.94318151e-01 4.72455353e-01
-1.03168460e-02 3.73634100e-01 4.23619241e-01 -5.37328124e-01
-6.00266814e-01 -1.85871571e-01 -2.35116500e-02 1.92313761e-01
8.65615085e-02 -6.99864700e-02 -9.10233736e-01 -3.86522174e-01
-1.49654135e-01 -4.55078840e-01 1.07166696e+00 1.71514675e-01
9.88968015e-01 -1.67459354e-01 -3.41495007e-01 8.50446463e-01
1.25213754e+00 -3.42931062e-01 5.19065797e-01 4.45891321e-01
7.87272573e-01 7.00691044e-01 1.12391984e+00 7.47703850e-01
5.94394147e-01 9.56536770e-01 6.43681228e-01 6.50001690e-02
1.00389577e-01 -4.38470185e-01 6.40625358e-01 8.42121959e-01
8.77049789e-02 4.05481681e-02 -9.01780367e-01 7.61467576e-01
-2.36568809e+00 -1.44522703e+00 1.34431601e-01 2.13289309e+00
6.76099539e-01 -2.17789531e-01 3.23987037e-01 1.32463023e-01
7.74226189e-01 5.05248070e-01 -1.22314468e-01 3.36418182e-01
-3.68984252e-01 -1.68015614e-01 1.17542699e-01 3.33601415e-01
-1.42206812e+00 7.30059445e-01 5.24125195e+00 8.49528253e-01
-8.71988356e-01 3.74491334e-01 3.11759830e-01 -2.88800061e-01
-3.68685782e-01 6.31259233e-02 -5.77070236e-01 4.33126122e-01
8.16881001e-01 -2.03105599e-01 2.69112051e-01 5.61341763e-01
1.06738381e-01 2.07936332e-01 -1.48050332e+00 1.53191924e+00
2.09260374e-01 -1.63210797e+00 -1.28760678e-03 -5.62322363e-02
7.74120152e-01 3.86553258e-01 1.66503876e-01 2.61428595e-01
-9.47113708e-03 -7.63184428e-01 9.41476941e-01 8.46831203e-01
5.61916053e-01 -5.43519497e-01 6.49032891e-01 1.02335460e-01
-1.69849443e+00 1.48623377e-01 -7.95693099e-02 2.61442453e-01
1.82795003e-01 4.46612686e-01 -4.61915821e-01 7.01345026e-01
9.90046740e-01 1.41034448e+00 -6.55698359e-01 1.02130342e+00
1.32442802e-01 2.87952513e-01 -2.48338267e-01 4.86579299e-01
3.19226652e-01 -1.30500168e-01 6.09337449e-01 1.43548679e+00
6.33872747e-01 -2.61465520e-01 3.32803279e-01 6.89405084e-01
-5.98170348e-02 2.37962380e-01 -7.86928415e-01 -7.93400258e-02
2.29264915e-01 1.30027807e+00 -4.06475991e-01 -2.82966793e-01
-6.76713228e-01 1.20310485e+00 2.62062430e-01 2.72382855e-01
-1.02961266e+00 -8.11933652e-02 8.47026944e-01 4.68266793e-02
2.69469202e-01 -1.59082398e-01 1.82647482e-01 -1.64092159e+00
3.72068584e-01 -8.85943890e-01 8.86325717e-01 -7.36350715e-01
-1.32650399e+00 4.13312972e-01 -3.65391672e-02 -1.89777792e+00
-4.35220450e-01 -4.58783120e-01 -3.79650384e-01 3.94668102e-01
-1.31576216e+00 -1.37608266e+00 -4.40534830e-01 1.20315075e+00
3.01650405e-01 -3.11962098e-01 7.44912565e-01 5.42685688e-01
-1.22424506e-01 7.22057939e-01 -2.24012151e-01 2.73183674e-01
1.06901634e+00 -1.27292335e+00 -4.89508696e-02 8.05870175e-01
8.58321428e-01 4.85613316e-01 5.19930005e-01 -2.74515241e-01
-1.54379117e+00 -1.17630243e+00 1.03481734e+00 -7.15864122e-01
1.04854429e+00 -3.60564560e-01 -7.68882215e-01 6.26067221e-01
2.22377598e-01 4.31187958e-01 1.05123937e+00 5.76450266e-02
-8.34378123e-01 -1.92094281e-01 -7.61618912e-01 5.14771819e-01
1.27005315e+00 -1.14196777e+00 -4.30723846e-01 5.62388122e-01
8.35214317e-01 -2.69693345e-01 -1.13251889e+00 5.50671697e-01
6.24162614e-01 -1.19480693e+00 9.70913291e-01 -7.38934636e-01
2.67938167e-01 -5.96007824e-01 -7.25283086e-01 -8.72612000e-01
-1.49688333e-01 -9.35365677e-01 -6.40533447e-01 1.16776502e+00
2.42410317e-01 9.52324271e-02 8.37162435e-01 2.60270923e-01
1.64028592e-02 -8.12054276e-01 -1.05622566e+00 -9.61004674e-01
-4.07031596e-01 -9.77927923e-01 4.36353445e-01 1.27252269e+00
4.47730780e-01 2.17028767e-01 -7.17444360e-01 1.19835198e-01
7.50009239e-01 2.44087785e-01 7.97754943e-01 -8.60346317e-01
-4.14553910e-01 -4.18980211e-01 -1.00619876e+00 -1.04825282e+00
3.09896797e-01 -1.04920757e+00 -1.11142226e-01 -1.32107830e+00
4.32701230e-01 3.66917439e-02 -8.10218990e-01 2.19338834e-01
3.32278460e-02 4.62125659e-01 3.35366666e-01 2.82384515e-01
-1.37456584e+00 5.99033237e-01 8.70427966e-01 -4.45503354e-01
2.28635091e-02 -1.44319594e-01 -2.81206578e-01 7.77642190e-01
4.93256092e-01 -3.52902919e-01 -4.30916846e-01 -4.30984825e-01
4.96917576e-01 6.52237386e-02 5.58279932e-01 -1.01299357e+00
5.16388416e-01 -1.60179153e-01 -6.79585487e-02 -8.30214977e-01
4.46653903e-01 -9.02048826e-01 2.10995823e-01 2.12246969e-01
-3.26937228e-01 2.64972895e-01 -1.36565343e-01 8.08457971e-01
-8.24536800e-01 1.33823022e-01 5.26822567e-01 4.15984988e-02
-1.18164635e+00 6.94439530e-01 1.34913638e-01 2.06195638e-01
1.05336750e+00 -2.15952992e-01 7.92217702e-02 -5.51868558e-01
-1.09011889e+00 3.19874734e-01 4.20911163e-01 7.68335164e-01
8.53915811e-01 -1.95881045e+00 -5.72152674e-01 -8.72507021e-02
6.90492630e-01 -5.01150787e-01 3.69231582e-01 1.21836364e+00
-4.63245884e-02 3.35108608e-01 5.24442494e-02 -9.98186171e-01
-1.38510370e+00 6.54144526e-01 3.76223296e-01 -5.03971636e-01
-3.29102278e-01 9.99729693e-01 1.83987170e-01 -3.60150695e-01
4.13598865e-01 -2.49059528e-01 -7.05216601e-02 4.55853283e-01
5.08600414e-01 3.51485729e-01 -7.19995564e-03 -9.53121960e-01
-5.24092078e-01 6.59555137e-01 2.38822281e-01 -3.00668925e-01
1.19426620e+00 -2.92151004e-01 -2.06422433e-01 6.97213352e-01
1.84466052e+00 -2.05232576e-01 -1.13454461e+00 -7.90028751e-01
3.52700412e-01 -6.43343806e-01 -1.35180905e-01 -6.58594295e-02
-1.13925552e+00 7.15347767e-01 7.20769584e-01 -5.05725443e-02
1.05633354e+00 4.84243721e-01 7.19183147e-01 4.74052876e-01
2.85124749e-01 -1.10293174e+00 6.63065255e-01 4.02912289e-01
1.03099513e+00 -1.63680041e+00 -1.44722208e-01 -8.80798697e-02
-4.56980735e-01 1.01688874e+00 6.19791627e-01 -2.16408167e-02
9.84211385e-01 -1.30249858e-01 -1.69583037e-01 -4.75420833e-01
-8.51121366e-01 -4.30610418e-01 8.22005153e-01 4.33975071e-01
5.42783260e-01 -1.04388341e-01 -3.55288200e-02 4.97988790e-01
-3.56534608e-02 -1.73335493e-01 -8.67908746e-02 8.65322471e-01
-1.19523734e-01 -1.24799955e+00 -7.91848823e-02 2.59172589e-01
-2.94660360e-01 -8.29395354e-02 -3.12580645e-01 6.60946012e-01
2.91096091e-01 8.98080289e-01 7.09147602e-02 -9.40233588e-01
2.05382943e-01 -8.45163316e-02 6.30640328e-01 -6.20707214e-01
-2.16929212e-01 9.21100900e-02 -1.27639920e-01 -1.34441316e+00
-1.00830245e+00 -8.28822196e-01 -8.98977578e-01 -2.74894238e-01
-1.15880288e-01 1.77505761e-01 1.02515988e-01 8.07082236e-01
3.96850526e-01 2.65502911e-02 6.00658178e-01 -9.00772929e-01
-2.23078087e-01 -7.72008598e-01 -5.93952119e-01 8.74177814e-01
3.25136513e-01 -7.40712225e-01 -3.42585534e-01 2.47526526e-01] | [9.969358444213867, 0.7467935681343079] |
8627bdd5-67ae-49ee-9c24-454181a87ee1 | bphigh-tamilnlp-acl2022-effects-of-data | null | null | https://aclanthology.org/2022.dravidianlangtech-1.22 | https://aclanthology.org/2022.dravidianlangtech-1.22.pdf | BpHigh@TamilNLP-ACL2022: Effects of Data Augmentation on Indic-Transformer based classifier for Abusive Comments Detection in Tamil | Social Media platforms have grown their reach worldwide. As an effect of this growth, many vernacular social media platforms have also emerged, focusing more on the diverse languages in the specific regions. Tamil has also emerged as a popular language for use on social media platforms due to the increasing penetration of vernacular media like Sharechat and Moj, which focus more on local Indian languages than English and encourage their users to converse in Indic languages. Abusive language remains a significant challenge in the social media framework and more so when we consider languages like Tamil, which are low-resource languages and have poor performance on multilingual models and lack language-specific models. Based on this shared task, “Abusive Comment detection in Tamil@DravidianLangTech-ACL 2022”, we present an exploration of different techniques used to tackle and increase the accuracy of our models using data augmentation in NLP. We also show the results of these techniques. | ['Bhavish Pahwa'] | null | null | null | null | dravidianlangtech-acl-2022-5 | ['abusive-language'] | ['natural-language-processing'] | [-6.40461206e-01 1.39875365e-02 -4.03855622e-01 1.37612462e-01
-8.39966476e-01 -6.81684852e-01 8.62446606e-01 5.52267075e-01
-8.95041764e-01 8.06805432e-01 6.50048673e-01 -2.68374860e-01
5.16476750e-01 -5.38211226e-01 -1.75029263e-01 -4.14806381e-02
1.73650280e-01 3.64964634e-01 7.94525817e-02 -6.42270923e-01
6.52655512e-02 1.93414614e-01 -1.01778114e+00 3.68249357e-01
9.54002440e-01 2.12224677e-01 -1.06075823e-01 3.37358773e-01
-5.83905637e-01 9.84795511e-01 -4.71498847e-01 -6.66703999e-01
8.34242180e-02 -3.30737472e-01 -5.94292343e-01 -2.61394799e-01
5.53647697e-01 1.57986328e-01 -2.01201975e-01 1.04276502e+00
5.86835086e-01 -9.55447853e-02 4.68881160e-01 -9.65998113e-01
-5.61403155e-01 1.26567566e+00 -8.39354813e-01 1.41979426e-01
6.86886489e-01 -1.90253794e-01 1.29738772e+00 -1.13225031e+00
1.15915501e+00 1.30135000e+00 7.04379022e-01 2.60291845e-01
-1.05900347e+00 -8.92816603e-01 1.80924028e-01 -4.39580977e-02
-1.40041411e+00 -3.45636129e-01 7.56087184e-01 -5.45566440e-01
8.69603395e-01 2.94864565e-01 5.29272974e-01 1.61158669e+00
3.33871208e-02 9.57756639e-01 1.33299816e+00 -4.64394271e-01
-1.39151379e-01 5.98248422e-01 -1.13028012e-01 4.41023439e-01
9.84930396e-02 -6.90296173e-01 -7.61534989e-01 -2.69576162e-01
-9.21702981e-02 -1.97555184e-01 -8.55788812e-02 1.58409223e-01
-1.07556486e+00 1.36709487e+00 2.57468343e-01 7.87640154e-01
-1.48250267e-01 -4.14882511e-01 9.35290217e-01 3.19785982e-01
1.18420470e+00 5.03341973e-01 -4.13829535e-01 -2.92014807e-01
-8.45227540e-01 4.56384867e-01 1.00121832e+00 7.71118641e-01
5.10268092e-01 -1.64992705e-01 2.61265635e-01 1.47087491e+00
5.09620488e-01 5.62199771e-01 4.92925256e-01 -2.72083759e-01
7.34410822e-01 8.92993033e-01 -2.41802052e-01 -1.26072884e+00
-5.54180384e-01 -3.74072880e-01 -6.54112279e-01 -2.69223303e-01
5.57624102e-01 -4.21811610e-01 -3.41467410e-01 1.52119875e+00
2.61255741e-01 -3.44387591e-01 -9.87849683e-02 5.17696619e-01
7.46541500e-01 8.35307717e-01 3.27725470e-01 -3.03451508e-01
1.19603276e+00 -7.10300863e-01 -7.46024430e-01 -4.36665297e-01
1.12225139e+00 -1.27342486e+00 1.29071558e+00 3.32117647e-01
-8.66781175e-01 -7.09239990e-02 -8.37834120e-01 -2.37150088e-01
-9.71259177e-01 -2.72265971e-01 5.92738807e-01 8.60903859e-01
-7.76245594e-01 8.86351690e-02 -3.12202364e-01 -8.30815375e-01
3.00056309e-01 -2.01802120e-01 -6.07438803e-01 -7.18716756e-02
-1.47290599e+00 1.18047059e+00 2.24967375e-01 -3.22264761e-01
-2.12866127e-01 -7.67702699e-01 -9.92727518e-01 -6.17369294e-01
4.11638945e-01 2.25946084e-01 8.61722767e-01 -1.18530738e+00
-1.26889098e+00 1.39682496e+00 1.72778979e-01 -3.68710160e-01
8.77081871e-01 -4.75252509e-01 -7.79310584e-01 -2.79783815e-01
3.49862128e-01 5.53953528e-01 4.29873019e-01 -8.67098749e-01
-5.77581346e-01 -1.40936986e-01 -2.92585697e-02 -3.64321135e-02
-7.47898877e-01 8.02891374e-01 -2.89457262e-01 -6.68859780e-01
-1.05893768e-01 -1.16359627e+00 -5.87168634e-02 -4.63467419e-01
-6.11724138e-01 -2.87809044e-01 1.09739482e+00 -1.08396089e+00
1.51959932e+00 -2.06691408e+00 1.20050520e-01 2.87903070e-01
3.53985786e-01 2.44619161e-01 1.35996670e-01 9.69924510e-01
3.09400469e-01 6.49829686e-01 -1.51488394e-01 -5.49436748e-01
-1.87621713e-01 4.51786779e-02 -1.35009363e-01 4.37359631e-01
-1.50722295e-01 5.74090242e-01 -1.02313733e+00 -6.16907358e-01
-1.04290500e-01 5.74001312e-01 -6.20539486e-01 -2.56332308e-01
-8.33655894e-02 5.63426971e-01 -1.11070953e-01 3.97843450e-01
6.36874557e-01 1.96439922e-02 2.90049076e-01 2.29906291e-01
-7.58198380e-01 4.57640380e-01 -7.54293740e-01 1.39276278e+00
-6.99812889e-01 9.93534327e-01 1.74439609e-01 3.40432022e-03
8.89291704e-01 1.36676520e-01 5.78426480e-01 -5.12635529e-01
4.75306809e-01 5.39874792e-01 1.76117018e-01 -2.95806438e-01
8.36052656e-01 2.52690017e-02 -3.53977531e-01 4.25982445e-01
-9.39636603e-02 -2.05339253e-01 4.05061185e-01 6.65141284e-01
6.94733977e-01 -9.32535976e-02 6.98977649e-01 -5.12364149e-01
6.28029704e-01 -6.36851639e-02 4.16297495e-01 4.45531994e-01
-3.12875479e-01 4.56236929e-01 4.87590373e-01 -8.46507326e-02
-1.13151383e+00 -5.31102836e-01 -2.25804687e-01 1.28622675e+00
-4.66299236e-01 -8.38092506e-01 -5.41448176e-01 -6.45598352e-01
-8.35676864e-02 7.15354264e-01 -4.18825120e-01 3.56764168e-01
-5.98643064e-01 -7.75752783e-01 5.45286000e-01 -1.80221319e-01
6.89054608e-01 -1.12580705e+00 3.11602112e-02 2.43403465e-01
-4.25724834e-01 -1.39768565e+00 -2.72486478e-01 -1.33168519e-01
-3.17317992e-01 -9.33789074e-01 -7.06264973e-01 -6.45707071e-01
1.68506935e-01 -8.49255696e-02 1.02512085e+00 1.18505890e-02
5.68967983e-02 3.06655556e-01 -6.67901754e-01 -7.99753845e-01
-9.38543081e-01 7.84906387e-01 1.50102153e-01 2.45765924e-01
5.55854797e-01 -3.68643969e-01 2.97808982e-02 -1.30466335e-02
-7.21367478e-01 2.11088359e-01 -8.10538083e-02 2.39128038e-01
-1.41217276e-01 -6.09391093e-01 8.12174797e-01 -1.61556172e+00
6.72420144e-01 -1.08823729e+00 -2.24367067e-01 -3.32595110e-01
-1.16322011e-01 -5.39479077e-01 6.09195113e-01 -5.50139666e-01
-8.13978016e-01 -3.40513885e-01 -4.54028636e-01 1.15940772e-01
1.10950999e-01 9.40902054e-01 3.74240167e-02 1.44264713e-01
7.31877863e-01 -2.90233523e-01 -3.83500159e-02 -6.35454178e-01
2.07386956e-01 1.07525694e+00 -2.16378078e-01 -1.59020230e-01
8.20217967e-01 3.19948077e-01 -6.66505635e-01 -1.57431972e+00
-9.89448249e-01 -6.55322611e-01 -5.51027179e-01 -4.42004085e-01
9.67345238e-01 -1.04094517e+00 -5.35603940e-01 7.21231461e-01
-1.13612449e+00 -2.59105861e-01 2.06997737e-01 3.59816253e-01
1.20664477e-01 1.43706188e-01 -9.38686430e-01 -8.54926467e-01
-8.17158818e-02 -7.51908123e-01 3.98524046e-01 -4.15447056e-02
-7.18979418e-01 -1.37782156e+00 3.75306040e-01 3.60131443e-01
5.49517274e-01 3.66510332e-01 9.04071331e-01 -1.13672936e+00
7.05849081e-02 -2.78396755e-01 -2.51105875e-01 4.30428535e-01
1.09529488e-01 1.68260545e-01 -9.55951095e-01 -3.45936149e-01
-4.26995575e-01 -5.48171759e-01 4.17569190e-01 -1.59251496e-01
4.91129935e-01 -2.95927584e-01 -2.11500719e-01 -1.31716967e-01
1.26848412e+00 -3.22200835e-01 4.71936047e-01 4.80600208e-01
7.92429388e-01 5.46580851e-01 3.11312407e-01 5.32855213e-01
5.69850862e-01 5.35269797e-01 2.22360775e-01 -7.23316967e-02
-3.36296856e-02 -5.62730730e-01 7.53669858e-01 1.58868670e+00
1.26649216e-01 -1.59073055e-01 -1.27021289e+00 6.10854268e-01
-1.64457762e+00 -7.02841640e-01 -5.35170555e-01 2.03220797e+00
9.58658218e-01 1.85562626e-01 3.43619913e-01 1.30764330e-02
4.91506964e-01 5.22055030e-01 2.29889285e-02 -5.37327170e-01
-5.26507497e-01 -3.23096961e-02 4.76388633e-01 8.30811322e-01
-1.01078832e+00 1.36198509e+00 6.05667448e+00 8.18991482e-01
-1.33418632e+00 5.90641022e-01 5.14765322e-01 1.29121989e-01
-1.30605578e-01 -1.66690558e-01 -1.13699448e+00 4.84290719e-01
9.99602795e-01 -9.41314176e-02 2.37287968e-01 7.14510381e-01
2.94470578e-01 -2.28420362e-01 -6.91752017e-01 8.52155149e-01
4.51558828e-01 -1.18747616e+00 -1.89525738e-01 1.15976207e-01
9.30091918e-01 7.30961442e-01 6.63433298e-02 6.96789920e-01
1.78406283e-01 -7.69126713e-01 8.06892931e-01 -8.47697929e-02
4.73399699e-01 -6.65499985e-01 5.42073071e-01 5.22053540e-01
-6.98208690e-01 9.79527235e-02 -2.39318594e-01 -3.19650650e-01
3.72675419e-01 5.18952191e-01 -7.98041701e-01 5.62111326e-02
6.71916783e-01 9.09096360e-01 -7.99866796e-01 7.63469756e-01
-3.05426449e-01 1.00729489e+00 -3.95913303e-01 -4.26353425e-01
5.17266035e-01 -2.28774711e-01 9.83527422e-01 1.57744515e+00
7.04071224e-02 -5.12686908e-01 5.43971121e-01 4.12503064e-01
-2.76942044e-01 1.04380429e+00 -9.99711275e-01 -4.55517650e-01
3.68156970e-01 1.37026083e+00 -6.41194284e-01 -2.74713375e-02
-9.33909416e-01 7.06072867e-01 6.25712097e-01 9.53508392e-02
-8.05752158e-01 -1.07255898e-01 6.17967248e-01 6.21532321e-01
-5.23734629e-01 -6.00477636e-01 1.05540209e-01 -1.36405528e+00
-3.40229452e-01 -1.32058954e+00 5.32410085e-01 -4.05219883e-01
-1.59863865e+00 6.54233038e-01 -1.36438414e-01 -1.06229401e+00
-1.79629102e-01 -5.36133230e-01 -4.14410606e-02 6.71965361e-01
-1.26206076e+00 -1.37878478e+00 2.44733006e-01 5.39398193e-01
6.69846416e-01 -3.55092764e-01 7.64400840e-01 7.62314677e-01
-4.39021021e-01 5.50826073e-01 -9.34112743e-02 2.58458167e-01
1.01825619e+00 -8.68542016e-01 1.44359633e-01 8.20086896e-01
2.39349321e-01 5.00558794e-01 8.08007836e-01 -9.01064694e-01
-9.12424266e-01 -9.42942321e-01 1.57626402e+00 -7.59346366e-01
1.40849233e+00 -7.91597486e-01 -8.30433786e-01 9.95263815e-01
5.81151068e-01 -5.27949572e-01 9.18239415e-01 5.86983442e-01
-6.63309395e-01 4.45318013e-01 -1.00623453e+00 1.16473687e+00
8.65796387e-01 -6.68424964e-01 -1.61575049e-01 5.95915318e-01
4.65638310e-01 -2.29962721e-01 -6.11625314e-01 -1.80518299e-01
2.05576926e-01 -8.26497316e-01 3.84168804e-01 -5.83532572e-01
5.16965091e-01 1.48850352e-01 -1.60287961e-01 -1.46598089e+00
-3.56549770e-02 -9.11586046e-01 4.34295982e-01 1.58892858e+00
9.43794131e-01 -9.43906605e-01 5.17929375e-01 3.66972387e-01
2.24645287e-01 -9.58151072e-02 -7.63256550e-01 -3.58274251e-01
4.55325872e-01 -7.06409454e-01 -9.06913579e-02 1.58863318e+00
2.38002986e-01 6.45709217e-01 -6.46282673e-01 -2.56357789e-01
1.40835077e-01 -5.77024043e-01 8.83950710e-01 -1.24491870e+00
-1.05299801e-02 -3.71915758e-01 -2.39125341e-01 -5.11070549e-01
2.83146113e-01 -1.32184887e+00 -3.67343843e-01 -1.06110084e+00
2.36397281e-01 -5.54111123e-01 2.06695914e-01 3.77118617e-01
1.50620326e-01 7.37949550e-01 5.88489294e-01 7.06466362e-02
-5.21378100e-01 6.20210497e-03 1.02716732e+00 -8.67444128e-02
-5.05486548e-01 -3.08987796e-01 -7.71810889e-01 9.95876372e-01
6.45326912e-01 -5.65627337e-01 1.28949344e-01 -1.52772203e-01
1.23652387e+00 -4.67138559e-01 -1.65940911e-01 -8.50583136e-01
-1.56595647e-01 -2.85043623e-02 -9.60253626e-02 -3.12547833e-01
2.09995687e-01 -7.58214533e-01 -1.56353474e-01 3.82071555e-01
-4.36065704e-01 1.87429324e-01 2.76992798e-01 2.19483040e-02
-2.52689332e-01 -8.26038495e-02 9.63752687e-01 5.10607362e-02
-3.97588253e-01 1.47277847e-01 -1.01301432e+00 5.04699230e-01
8.45152259e-01 8.67636204e-02 -1.65888354e-01 -7.49344170e-01
-8.18512619e-01 1.28145918e-01 4.11261618e-01 9.61639702e-01
3.42979841e-02 -1.07299995e+00 -1.28882504e+00 -6.89138547e-02
3.58861804e-01 -6.99838638e-01 8.58808085e-02 1.05096674e+00
-6.48979247e-01 9.33264419e-02 -1.37743697e-01 -2.82666653e-01
-1.20727801e+00 8.81761685e-02 7.31513128e-02 -5.46101511e-01
-5.60050607e-01 5.38091004e-01 -2.47765914e-01 -8.94609332e-01
-1.27406389e-01 7.57727623e-02 -5.90972900e-01 7.57005990e-01
5.17305493e-01 3.78423303e-01 -3.22063826e-02 -1.12458909e+00
-3.59045923e-01 -3.56875546e-02 -3.68901640e-01 -3.25774610e-01
1.23830295e+00 -1.81976587e-01 -4.98999029e-01 1.08881748e+00
1.35276163e+00 1.12205744e+00 -3.27029735e-01 -1.41667098e-01
3.25245857e-01 -3.78818423e-01 -5.29520921e-02 -8.37891042e-01
-7.75266230e-01 5.68987608e-01 1.67141333e-01 4.68962491e-01
2.78641731e-01 3.11105344e-02 8.07696164e-01 -8.57040228e-04
3.63603264e-01 -1.27281129e+00 -9.19635892e-02 1.17400122e+00
1.14086378e+00 -1.27972257e+00 2.80550681e-02 -6.29442990e-01
-6.76982224e-01 6.94039822e-01 4.58272338e-01 -1.52319754e-02
9.70564783e-01 2.15973318e-01 5.21302223e-01 2.78550852e-02
-3.80589694e-01 -7.81845227e-02 -2.52722260e-02 2.60774106e-01
1.14907074e+00 1.46847829e-01 -9.09656882e-01 4.82174814e-01
-5.00751734e-01 -4.59088475e-01 8.84275436e-01 6.96803451e-01
-9.39881280e-02 -1.27842379e+00 -2.31880426e-01 3.63169312e-01
-9.24062431e-01 -3.06579739e-01 -7.61909366e-01 1.24344313e+00
4.38941568e-02 8.77734601e-01 -3.20313200e-02 -2.26162568e-01
2.03959588e-02 3.06541204e-01 5.27185872e-02 -8.59970510e-01
-1.21126902e+00 2.43999735e-02 5.81085682e-01 -1.02106884e-01
-2.09497839e-01 -9.34090197e-01 -9.75110948e-01 -6.92113698e-01
-1.54359147e-01 5.29807471e-02 8.80045593e-01 8.30397069e-01
1.22655734e-01 1.35508785e-02 3.89255196e-01 -7.10733593e-01
9.31832343e-02 -1.27044296e+00 -5.06891131e-01 4.35239822e-01
-1.10750504e-01 -2.34147012e-01 -5.17192364e-01 -3.34619224e-01] | [9.107645988464355, 10.447866439819336] |
17c37c03-b4fa-430d-99b5-3342b3075826 | interactions-of-fungi-with-concrete | 1708.01337 | null | https://arxiv.org/abs/1708.01337v2 | https://arxiv.org/pdf/1708.01337v2.pdf | Interactions of Fungi with Concrete: Significant Importance for Bio-Based Self-Healing Concrete | The goal of this study is to explore a new self-healing concept in which fungi are used as a self-healing agent to promote calcium mineral precipitation to fill the cracks in concrete. An initial screening of different species of fungi has been conducted. Fungal growth medium was overlaid onto cured concrete plate. Mycelial discs were aseptically deposited at the plate center. The results showed that, due to the dissolving of Ca(OH)2 from concrete, the pH of the growth medium increased from its original value of 6.5 to 13.0. Despite the drastic pH increase, Trichoderma reesei (ATCC13631) spores germinated into hyphal mycelium and grew equally well with or without concrete. X-ray diffraction (XRD) and scanning electron microscope (SEM) confirmed that the crystals precipitated on the fungal hyphae were composed of calcite. These results indicate that T. reesei has great potential to be used in bio-based self-healing concrete for sustainable infrastructure. | ['Congrui Jin', 'Ning Zhang', 'Guangwen Zhou', 'David G. Davies', 'Hui Zhou', 'Jada Crump', 'Xiaobo Chen', 'Jing Luo'] | 2017-08-04 | null | null | null | null | ['x-ray-diffraction'] | ['miscellaneous'] | [ 3.43863696e-01 -1.04317017e-01 3.09772164e-01 7.45933473e-01
3.60784560e-01 -2.23204300e-01 4.44017380e-01 1.73092410e-01
-1.18598729e-01 1.01098871e+00 1.33635536e-01 -1.23761393e-01
1.37122095e-01 -1.05169892e+00 -1.24668799e-01 -1.31163466e+00
6.74692765e-02 2.88753361e-01 5.16098917e-01 -2.91510850e-01
5.44903398e-01 7.38250136e-01 -1.64435112e+00 -9.60103422e-02
7.76024044e-01 1.54140562e-01 9.54518139e-01 4.18091536e-01
-1.03639886e-01 4.57611233e-01 -7.40461767e-01 -5.52972481e-02
-1.54661313e-01 -7.45915398e-02 -5.85927963e-01 4.89500999e-01
-9.95765030e-01 -2.46978536e-01 1.65679768e-01 6.29183173e-01
-6.29037470e-02 3.97802591e-02 6.63186848e-01 -9.49328303e-01
-5.98821819e-01 3.82573873e-01 -7.51208186e-01 -7.54052848e-02
4.28219855e-01 1.13308944e-01 3.31376225e-01 -1.20613742e+00
7.84535766e-01 8.66562188e-01 4.04903106e-02 7.84530640e-01
-8.35899532e-01 -3.40665430e-01 -4.32167202e-01 2.89475501e-01
-1.24985301e+00 -1.21566586e-01 2.83462763e-01 -3.94383699e-01
9.22762275e-01 4.03999507e-01 1.19188249e+00 6.30112648e-01
9.30072188e-01 1.86868578e-01 1.46342874e+00 -5.64939618e-01
9.19691205e-01 -1.02140456e-01 -4.60986912e-01 3.49828213e-01
8.50594401e-01 3.85671929e-02 -2.79209942e-01 -1.05817206e-01
8.64543200e-01 2.53045354e-02 -4.86569762e-01 4.26305272e-02
-5.71840882e-01 5.75076103e-01 4.48847562e-01 8.36514890e-01
-5.00520825e-01 -9.86307561e-02 1.06773764e-01 -1.96508124e-01
5.90119921e-02 3.06292653e-01 -9.81840938e-02 -1.44825548e-01
-4.06407863e-01 -4.88742143e-02 8.78414571e-01 4.92826626e-02
4.14824992e-01 2.08803087e-01 6.83856189e-01 9.89569962e-01
4.71329957e-01 4.78658170e-01 4.10924792e-01 -8.46417129e-01
-3.00974071e-01 6.15099788e-01 1.39976129e-01 -7.78326213e-01
1.30378559e-01 -1.68510526e-01 -7.43804276e-01 4.26962763e-01
3.24297160e-01 2.45193645e-01 -8.67583632e-01 1.09662998e+00
4.19601053e-01 5.98964877e-02 1.88693941e-01 9.56797481e-01
4.10880506e-01 9.20777977e-01 3.73540074e-01 -4.99004185e-01
1.32096672e+00 -6.53693378e-01 -7.39449680e-01 6.54136613e-02
2.45707169e-01 -5.79966843e-01 1.10603666e+00 5.59825957e-01
-1.16233373e+00 1.66809857e-01 -1.45470178e+00 1.04174542e+00
-3.22471619e-01 -4.59124267e-01 6.43185616e-01 7.86565900e-01
-8.44091594e-01 5.22360146e-01 -1.18731809e+00 -7.49794006e-01
-1.98219597e-01 1.50013864e-01 -4.09518540e-01 -2.94943720e-01
-7.76341677e-01 7.86171257e-01 -1.12057962e-01 2.07844332e-01
-7.61848092e-01 -1.53740838e-01 -2.17201665e-01 -1.90961435e-02
-9.24876928e-02 -5.75005770e-01 7.32885540e-01 -1.73143953e-01
-1.91647184e+00 7.65140474e-01 -1.51796475e-01 1.54639959e-01
-4.56536701e-03 -1.01193495e-01 -4.18816477e-01 6.08974576e-01
4.08266395e-01 -1.87871590e-01 1.42562136e-01 -1.86872983e+00
-4.57069576e-01 -5.43994963e-01 -1.82199910e-01 -2.59001523e-01
-2.44206324e-01 1.12049706e-01 8.09251890e-02 -5.57797730e-01
6.67887926e-01 -9.01470900e-01 -4.86262858e-01 -5.19461989e-01
-3.91621023e-01 -5.05924039e-02 8.25634480e-01 -5.92852056e-01
6.75594211e-01 -1.86926055e+00 6.75025731e-02 3.52332383e-01
5.68183102e-02 1.71662420e-01 2.48867258e-01 1.01512623e+00
4.13466766e-02 3.51677775e-01 -3.10958773e-01 5.77703118e-01
-6.56264603e-01 3.31128955e-01 3.46071541e-01 3.92851651e-01
1.12981580e-01 -3.38973440e-02 -8.89374554e-01 -2.52800345e-01
-9.27854404e-02 5.45133054e-01 -1.58885583e-01 -1.30001470e-01
9.17577296e-02 3.58708262e-01 -5.50224841e-01 1.24726570e+00
1.00449955e+00 8.02311115e-04 2.86942869e-01 2.89380699e-01
-4.27936077e-01 -1.07988209e-01 -1.07642949e+00 1.02087367e+00
-1.25010088e-01 -9.63582937e-03 7.19199240e-01 -4.83898699e-01
1.08738017e+00 5.96391976e-01 7.83404529e-01 -3.12463284e-01
-1.20589763e-01 6.94302201e-01 -2.16222212e-01 -6.56294823e-01
1.49025261e-01 -5.69164693e-01 7.15106368e-01 3.92837226e-01
-7.61671484e-01 -4.10611212e-01 6.23354077e-01 2.71090746e-01
1.26572692e+00 -1.63581356e-01 1.50338963e-01 -4.40823168e-01
8.12215209e-01 2.54073083e-01 5.66711664e-01 -1.98502913e-01
1.87953040e-01 1.83468878e-01 -1.19457006e-01 1.38004079e-01
-9.19967592e-01 -1.38961780e+00 -2.83162624e-01 5.31405151e-01
2.40508690e-01 -1.12003706e-01 -8.09098423e-01 5.80920458e-01
-1.83631822e-01 5.38263619e-01 -3.98166537e-01 4.67648394e-02
-3.21115702e-01 -8.40905428e-01 1.39870375e-01 9.65311900e-02
7.37056553e-01 -9.04629707e-01 -5.88247299e-01 8.33227932e-01
-1.58868313e-01 -4.10044730e-01 4.01091784e-01 2.24257216e-01
-1.08610117e+00 -1.12695098e+00 -7.90531516e-01 -7.60569274e-01
4.17957872e-01 4.70776558e-01 4.59544331e-01 8.17359686e-01
-5.69160163e-01 2.38905251e-01 -7.51268148e-01 -2.58664995e-01
-6.38378024e-01 -5.78873456e-01 -2.80811405e-03 -5.80612242e-01
4.08077419e-01 -9.14370418e-01 -5.88058949e-01 2.71250665e-01
-8.37079823e-01 -2.38035828e-01 3.35784018e-01 6.59641683e-01
6.06359065e-01 6.36622906e-01 6.51725769e-01 -6.47633553e-01
6.67763114e-01 -4.81597662e-01 -6.62675500e-02 1.27826720e-01
-7.91397393e-01 -3.83555502e-01 2.55188525e-01 -3.21269602e-01
-1.17450476e+00 -4.95100111e-01 1.51428074e-01 4.42136854e-01
7.48411566e-02 8.55656087e-01 -2.51034826e-01 2.83899605e-01
6.71541870e-01 2.24022046e-01 7.22588673e-02 -2.32791379e-01
-4.09696758e-01 6.90436184e-01 7.85732627e-01 -5.12146354e-01
7.55473316e-01 6.81075156e-01 1.73378214e-01 -9.31318104e-01
-1.53361276e-01 -3.87862325e-01 -5.87440841e-02 -7.91556239e-01
1.19213343e+00 -5.37487268e-01 -7.59015977e-01 7.18881488e-01
-7.30298102e-01 -2.39905804e-01 7.41847754e-02 7.94823349e-01
-1.96658671e-01 4.56509352e-01 -1.43002105e+00 -9.19133842e-01
-4.14670259e-01 -7.24121988e-01 6.96311444e-02 2.27979079e-01
-2.79845204e-02 -6.97699547e-01 6.31377459e-01 6.16011202e-01
5.50535917e-01 7.52144098e-01 9.31812108e-01 3.56757075e-01
-4.84487027e-01 -1.47042200e-01 3.63571763e-01 -1.46471679e-01
5.45067668e-01 7.02457666e-01 -6.91791773e-01 -1.71351895e-01
4.32051092e-01 1.76828764e-02 5.90549409e-01 2.43118927e-01
2.18878001e-01 -1.13457710e-01 -6.31753743e-01 1.76626042e-01
1.78382409e+00 8.36874843e-01 1.01674020e+00 1.23005641e+00
1.63408294e-01 7.87628472e-01 8.50893974e-01 5.99022090e-01
-2.16293007e-01 -7.82219917e-02 8.17481995e-01 -1.37682697e-02
-7.26188645e-02 7.07505122e-02 4.42064881e-01 8.54114771e-01
-1.12380338e+00 -1.95239156e-01 -1.03108990e+00 5.11126697e-01
-9.40868855e-01 -1.03226483e+00 -7.84022510e-01 2.04986811e+00
1.09776258e+00 -1.01718947e-01 2.22579911e-02 9.13020134e-01
8.37346315e-01 -8.42887938e-01 -2.30412871e-01 -3.30622345e-01
-1.23254322e-01 7.01289713e-01 2.67589480e-01 4.69356358e-01
-1.89039096e-01 5.51905990e-01 5.89161825e+00 -1.19036688e-02
-1.45648181e+00 -3.63025814e-01 -2.07694873e-01 1.15858771e-01
-4.27615106e-01 3.72856230e-01 -1.45415574e-01 2.18988553e-01
7.26236999e-01 -4.88538086e-01 3.72335762e-01 5.46774745e-01
4.50162798e-01 -6.81403637e-01 -4.39121723e-01 2.11373657e-01
-6.10487700e-01 -1.41714799e+00 -3.39038432e-01 4.42570299e-01
5.62568605e-01 -2.59563953e-01 -3.51631016e-01 -3.29164028e-01
4.34086174e-01 -7.70541191e-01 4.25601840e-01 1.15612142e-01
5.31932712e-01 -8.62542152e-01 6.78999066e-01 1.72986090e-01
-1.28426218e+00 -9.74838734e-02 -4.16443765e-01 -3.66521478e-01
4.52481925e-01 9.10735130e-01 -1.07378972e+00 3.70014787e-01
1.04068446e+00 2.49763831e-01 -3.51702690e-01 1.03602386e+00
-3.43331665e-01 4.84679341e-01 -4.07260925e-01 -3.68975997e-02
1.97254624e-02 -6.55348361e-01 3.29113245e-01 6.27630115e-01
5.03053248e-01 4.16381299e-01 -1.74902737e-01 6.82075858e-01
7.53994882e-01 4.19251621e-01 -1.39341071e-01 -2.45445803e-01
9.79632378e-01 6.99307501e-01 -1.42696393e+00 -1.35550365e-01
-1.62470520e-01 5.67143202e-01 -3.39940816e-01 3.19566131e-01
-3.63119602e-01 -2.74460435e-01 2.44199693e-01 5.76202214e-01
1.57008320e-01 -4.65634555e-01 -4.81201857e-01 -6.30862594e-01
-3.62773746e-01 -4.66295987e-01 2.04675615e-01 -8.73095453e-01
-9.83669758e-01 2.37573981e-01 -2.05948040e-01 -6.93711460e-01
1.63871154e-01 -4.39165354e-01 -8.49096298e-01 6.95315659e-01
-5.99642932e-01 -7.25074232e-01 -5.36230505e-01 3.13681990e-01
3.57304037e-01 -2.87664011e-02 1.10917115e+00 -3.56619090e-01
-7.33416140e-01 -3.62248719e-01 3.56522977e-01 -6.30105734e-01
3.01232874e-01 -8.13712716e-01 -4.23779249e-01 8.18259120e-01
-7.18439221e-01 8.02635431e-01 1.07286251e+00 -1.24510264e+00
-1.35364127e+00 -6.74047768e-01 9.87015545e-01 5.91701984e-01
5.38622797e-01 2.38163367e-01 -1.00589430e+00 8.03905502e-02
4.98195469e-01 -1.08938134e+00 1.10638082e+00 -3.63613725e-01
1.22536451e-01 6.17471874e-01 -1.58154857e+00 6.88984036e-01
3.81963134e-01 1.63643524e-01 -2.73407996e-01 5.35530627e-01
3.04877050e-02 -1.43819591e-02 -1.33518207e+00 2.67526746e-01
3.86384189e-01 -7.41790295e-01 1.00148761e+00 -4.41246405e-02
7.58819103e-01 -4.74134237e-01 -3.15907568e-01 -8.21830094e-01
-6.85045004e-01 -1.23523109e-01 3.88992041e-01 1.27710235e+00
3.77140641e-02 -6.26292586e-01 7.61615574e-01 5.65891385e-01
-1.66882247e-01 -4.88943875e-01 -6.72091365e-01 -8.17446470e-01
6.13817684e-02 4.80627835e-01 3.04837376e-01 9.06159282e-01
7.28376925e-01 4.59311903e-01 1.68395281e-01 2.48994716e-02
6.73538864e-01 -2.07821175e-01 4.73528266e-01 -1.51598632e+00
-2.37260684e-01 -1.71390660e-02 -3.19630615e-02 -8.52070227e-02
-2.05017731e-01 -8.03399563e-01 -2.53402799e-01 -2.17442656e+00
2.28424281e-01 -4.32332337e-01 1.76956818e-01 1.60578281e-01
-5.38636073e-02 -1.52748451e-01 1.70020908e-01 5.32541275e-01
7.40542352e-01 2.17463166e-01 1.30247235e+00 -3.05766501e-02
-5.57264447e-01 -2.47917082e-02 -4.11960930e-01 3.73625457e-01
1.27456951e+00 -4.20873851e-01 -5.07496953e-01 -7.69759640e-02
3.75410467e-02 4.16263312e-01 2.90631060e-03 -9.05899882e-01
-1.56433344e-01 -8.35752308e-01 -6.71758354e-02 -6.12252712e-01
3.85113984e-01 -1.05022621e+00 1.11371982e+00 1.30140460e+00
3.51392061e-01 -1.91585377e-01 -1.15795143e-01 6.04567468e-01
-4.77417745e-03 -5.43400347e-01 8.60216856e-01 -2.74301231e-01
-2.32578322e-01 -3.75500441e-01 -1.57428861e+00 -6.78013861e-01
1.33722770e+00 -1.06152904e+00 -2.23145470e-01 -1.97418351e-02
-1.13542438e+00 -2.16137618e-01 1.11622608e+00 -4.90023196e-01
6.06534243e-01 -1.39153028e+00 -5.23589432e-01 -2.21824527e-01
-3.26634467e-01 -2.13552043e-01 1.76388234e-01 8.58537078e-01
-1.50353158e+00 7.88772106e-02 -8.24547470e-01 -5.23857117e-01
-1.26939881e+00 5.29369593e-01 1.92105211e-02 -1.19503655e-01
-6.66322529e-01 8.94171536e-01 -3.14484835e-02 3.83754015e-01
-1.33885980e-01 -1.19166687e-01 -4.68632400e-01 -5.41564107e-01
4.18318480e-01 7.96397150e-01 3.90845351e-02 -1.42469883e-01
-6.48309708e-01 2.60261655e-01 3.33074123e-01 -3.90181750e-01
1.96530759e+00 -5.86304702e-02 -6.26944542e-01 3.59296054e-01
5.73428154e-01 2.93981731e-01 -9.80101943e-01 3.69780630e-01
1.01711631e-01 -6.48853123e-01 -4.36799228e-01 -4.81793910e-01
-1.26633775e+00 2.20479533e-01 7.95430690e-02 3.08616519e-01
1.12212729e+00 -1.10039741e-01 5.65668106e-01 1.41008273e-01
4.67163742e-01 -1.54751551e+00 6.10065460e-01 1.26845375e-01
9.93087173e-01 -1.19157866e-01 -5.12544671e-03 -9.58015800e-01
-2.32621565e-01 1.17647088e+00 5.70770204e-01 -3.27655971e-01
6.03305817e-01 5.78359783e-01 -2.90788226e-02 -3.62453133e-01
-8.09779704e-01 2.55253524e-01 -8.97205412e-01 1.09382510e+00
8.88164043e-01 9.17614028e-02 -1.15782201e+00 3.59982789e-01
-4.04020756e-01 1.97349176e-01 1.10765100e+00 1.44336486e+00
-1.11984956e+00 -1.16763854e+00 -1.05760288e+00 2.27166489e-01
-3.49719018e-01 3.88794690e-01 -3.63907397e-01 8.41983378e-01
-1.79467648e-01 1.43676126e+00 -3.32669139e-01 -1.90231696e-01
9.05314237e-02 1.31673336e-01 5.25863409e-01 -2.90711582e-01
-3.00027221e-01 5.86627901e-01 5.52604273e-02 -1.38891190e-02
-7.65527844e-01 -8.08478773e-01 -1.90613818e+00 -5.43248832e-01
-4.85405296e-01 5.16123235e-01 1.08710706e+00 7.42204428e-01
-5.75316191e-01 4.51397210e-01 8.03663373e-01 -3.38776618e-01
3.40057462e-02 -8.78575861e-01 -1.37211382e+00 1.34566933e-01
-4.07259345e-01 -8.29122126e-01 -7.24065900e-01 6.12545907e-01] | [4.9648919105529785, 4.859765529632568] |
42f3c128-7fa4-4ef7-a263-50f4d26e8b96 | human-mesh-recovery-from-multiple-shots | 2012.09843 | null | https://arxiv.org/abs/2012.09843v1 | https://arxiv.org/pdf/2012.09843v1.pdf | Human Mesh Recovery from Multiple Shots | Videos from edited media like movies are a useful, yet under-explored source of information. The rich variety of appearance and interactions between humans depicted over a large temporal context in these films could be a valuable source of data. However, the richness of data comes at the expense of fundamental challenges such as abrupt shot changes and close up shots of actors with heavy truncation, which limits the applicability of existing human 3D understanding methods. In this paper, we address these limitations with an insight that while shot changes of the same scene incur a discontinuity between frames, the 3D structure of the scene still changes smoothly. This allows us to handle frames before and after the shot change as multi-view signal that provide strong cues to recover the 3D state of the actors. We propose a multi-shot optimization framework, which leads to improved 3D reconstruction and mining of long sequences with pseudo ground truth 3D human mesh. We show that the resulting data is beneficial in the training of various human mesh recovery models: for single image, we achieve improved robustness; for video we propose a pure transformer-based temporal encoder, which can naturally handle missing observations due to shot changes in the input frames. We demonstrate the importance of the insight and proposed models through extensive experiments. The tools we develop open the door to processing and analyzing in 3D content from a large library of edited media, which could be helpful for many downstream applications. Project page: https://geopavlakos.github.io/multishot | ['Angjoo Kanazawa', 'Jitendra Malik', 'Georgios Pavlakos'] | 2020-12-17 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Pavlakos_Human_Mesh_Recovery_From_Multiple_Shots_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Pavlakos_Human_Mesh_Recovery_From_Multiple_Shots_CVPR_2022_paper.pdf | cvpr-2022-1 | ['human-mesh-recovery'] | ['computer-vision'] | [ 1.32499963e-01 -4.20020856e-02 5.55407023e-03 -9.78942961e-02
-4.84363139e-01 -4.19177115e-01 4.00125980e-01 -1.75726101e-01
1.40956298e-01 3.95829111e-01 4.88514990e-01 3.10288161e-01
-2.39875317e-02 -5.99949896e-01 -8.95413578e-01 -5.05615354e-01
-1.98275000e-01 2.48129219e-01 4.84823555e-01 -4.10780370e-01
-6.36408255e-02 4.71192896e-01 -1.81081188e+00 4.96080935e-01
2.92392254e-01 8.65350246e-01 2.97929794e-01 6.93191290e-01
7.21498504e-02 8.85410190e-01 -2.08581969e-01 -4.34633642e-01
5.17200768e-01 -4.35689151e-01 -6.84252858e-01 6.62591279e-01
5.48467815e-01 -6.55728221e-01 -6.25776947e-01 8.39238882e-01
4.16742742e-01 3.43408793e-01 2.44747624e-01 -9.68467057e-01
-1.55277271e-02 6.20777458e-02 -7.89225280e-01 2.05239490e-01
9.87826169e-01 1.27338290e-01 7.42259443e-01 -9.41782176e-01
1.23607278e+00 1.36897218e+00 6.31599605e-01 3.69367808e-01
-9.98354077e-01 -2.75987357e-01 1.54354140e-01 4.36975151e-01
-1.16801500e+00 -8.69809985e-01 1.09472191e+00 -5.95991969e-01
5.11334002e-01 2.52607167e-01 9.88730729e-01 1.47704279e+00
3.28298993e-02 9.10849273e-01 7.31785655e-01 -4.09200996e-01
-7.95158446e-02 -5.92513867e-02 -1.27987489e-01 9.57945764e-01
-1.35365665e-01 1.16742767e-01 -9.15497005e-01 -4.68498059e-02
1.04579389e+00 4.03281927e-01 -5.43223023e-01 -5.97760797e-01
-1.40313053e+00 4.22405511e-01 1.66016430e-01 1.50972366e-01
-4.72967833e-01 -4.08618376e-02 3.62444609e-01 4.31094021e-01
6.99796736e-01 -4.83382978e-02 -1.48351923e-01 -3.59337628e-01
-9.24676836e-01 3.64749014e-01 7.68223822e-01 9.43029761e-01
5.81989527e-01 3.00651491e-02 1.03207037e-01 7.20166624e-01
1.89656508e-03 3.84974748e-01 -4.68034595e-02 -1.46532404e+00
3.47737342e-01 2.40242049e-01 2.27079853e-01 -1.42300069e+00
-1.02359891e-01 1.75864436e-02 -7.25779176e-01 1.49301901e-01
5.22252142e-01 7.76094049e-02 -6.57580435e-01 1.60520577e+00
8.04895341e-01 2.92166948e-01 -4.16231960e-01 1.07904673e+00
8.05399001e-01 5.81696451e-01 -4.93712425e-01 -3.96735430e-01
1.37394345e+00 -8.73946965e-01 -8.58215332e-01 -1.28245026e-01
8.74435827e-02 -7.59327114e-01 8.88146698e-01 3.11422616e-01
-1.45052981e+00 -6.11686826e-01 -7.54946828e-01 -1.83410719e-01
1.13476098e-01 -5.21708429e-01 5.29246747e-01 7.19188601e-02
-6.85998559e-01 8.16526651e-01 -1.02903950e+00 -5.68118453e-01
4.00343955e-01 -4.57387343e-02 -6.73752129e-01 -4.76442039e-01
-9.17311072e-01 6.49955392e-01 -1.38237581e-01 7.90518969e-02
-9.21987891e-01 -7.61601508e-01 -9.79226530e-01 -2.10673466e-01
8.87322485e-01 -8.07261646e-01 1.19006014e+00 -9.94927764e-01
-1.32458246e+00 9.76775646e-01 -3.47344458e-01 -4.59727161e-02
9.26706493e-01 -4.39347655e-01 -1.34414241e-01 5.92110574e-01
1.71169519e-01 3.07539344e-01 9.50549304e-01 -1.27962232e+00
-4.08804059e-01 -3.45011324e-01 2.31218487e-01 2.11931944e-01
2.45168805e-02 3.90021279e-02 -8.39368403e-01 -7.85123944e-01
9.52948928e-02 -7.78183222e-01 -1.21177547e-01 4.78348434e-01
-2.33329877e-01 5.00464261e-01 7.14365125e-01 -8.60853076e-01
9.30809736e-01 -2.25134087e+00 5.27379692e-01 -1.62956834e-01
2.81060725e-01 -1.64188266e-01 7.42488205e-02 6.92645967e-01
1.13974608e-01 -2.35289827e-01 -5.97906411e-02 -5.30251145e-01
-3.74492764e-01 2.27291822e-01 -2.77088493e-01 6.43438935e-01
-1.44182798e-02 7.62854576e-01 -9.91471589e-01 -5.83716214e-01
5.31408012e-01 5.82156718e-01 -6.67005479e-01 4.69664037e-01
-1.55136287e-01 9.04518187e-01 -4.28030640e-01 8.01945210e-01
3.98037970e-01 -3.80448252e-01 -2.87432373e-02 -4.89477128e-01
-2.84346063e-02 -1.10803314e-01 -1.42055285e+00 2.10477233e+00
-1.34868339e-01 4.63086039e-01 2.99774706e-01 -9.26592529e-01
3.38414669e-01 4.83232647e-01 8.80225301e-01 -5.61143398e-01
2.11105615e-01 -1.39767185e-01 -3.61281484e-01 -9.11329448e-01
4.22946364e-01 -2.66664207e-01 6.62103221e-02 3.44389081e-01
3.87441106e-02 -6.45285752e-03 1.56085536e-01 3.32835019e-01
1.06323063e+00 3.51326078e-01 2.32698381e-01 1.23222925e-01
1.41617328e-01 -9.85683277e-02 6.34239018e-01 5.28162420e-01
-1.39755756e-01 1.01205671e+00 2.82897383e-01 -5.66442788e-01
-1.24140131e+00 -1.02722490e+00 1.03048407e-01 8.40089083e-01
3.98653120e-01 -5.43752730e-01 -4.43679273e-01 -2.24926159e-01
-2.51780987e-01 1.51221335e-01 -6.21526897e-01 9.73961577e-02
-7.71031320e-01 -2.64901936e-01 7.09183887e-02 2.65101790e-01
3.54959637e-01 -7.98663139e-01 -8.10125291e-01 2.00948045e-01
-6.97254896e-01 -1.49464047e+00 -6.61009610e-01 -2.85618186e-01
-8.98198187e-01 -1.27029228e+00 -8.71003926e-01 -5.00237465e-01
4.85926598e-01 6.42762899e-01 1.07832074e+00 1.02555409e-01
-4.80746120e-01 8.15867126e-01 -5.09237111e-01 -1.29720289e-02
-3.87968659e-01 -5.97675741e-01 1.46181181e-01 4.45264071e-01
-3.10893089e-01 -9.63964820e-01 -6.89869165e-01 3.94059122e-01
-7.44998991e-01 4.73493844e-01 2.59961206e-02 7.16325223e-01
5.85923493e-01 -1.51788652e-01 -3.45215052e-02 -8.14869940e-01
-6.67234510e-02 -5.40474892e-01 -2.20424011e-01 6.67436570e-02
1.11876756e-01 -1.93114758e-01 3.97569150e-01 -6.14815235e-01
-1.13639379e+00 3.81335430e-02 4.76636030e-02 -1.06038344e+00
-8.08120072e-02 1.40504688e-01 -5.79287559e-02 9.73328650e-02
3.72564435e-01 9.85015854e-02 1.90795437e-01 -7.07497537e-01
3.61132085e-01 1.29703164e-01 6.74361408e-01 -6.02825701e-01
6.38907790e-01 1.01204789e+00 4.29579392e-02 -1.09650159e+00
-9.52848434e-01 -6.01697266e-01 -7.93538213e-01 -5.91711402e-01
7.44928181e-01 -1.06479931e+00 -4.34571415e-01 3.52604359e-01
-1.16027617e+00 -2.87430465e-01 -5.25366306e-01 3.86879057e-01
-7.41866410e-01 7.32862353e-01 -9.28088665e-01 -7.66708851e-01
1.74613539e-02 -1.02770853e+00 1.26294100e+00 -6.34666160e-02
-3.48188907e-01 -9.77238417e-01 8.46739337e-02 5.81238091e-01
-1.42722055e-01 7.50689626e-01 5.74466825e-01 3.41230771e-03
-7.70268381e-01 -1.43384516e-01 2.68356472e-01 1.19975805e-01
1.76414609e-01 6.25685304e-02 -8.12440872e-01 -2.88742453e-01
2.51489758e-01 -2.20245168e-01 6.28900290e-01 6.01466715e-01
8.48601401e-01 -2.31948853e-01 -2.98616618e-01 7.53452718e-01
1.21976161e+00 -2.01493368e-01 5.54857135e-01 5.96365370e-02
8.73320758e-01 8.37866902e-01 7.58057058e-01 7.55429864e-01
2.22383291e-01 9.56311345e-01 3.29273611e-01 3.15996632e-02
-3.47940445e-01 -3.87992471e-01 4.83998597e-01 1.08919859e+00
-6.10859334e-01 -1.65172786e-01 -6.98882401e-01 6.23007774e-01
-1.95364928e+00 -1.36681783e+00 -1.31950036e-01 2.13612962e+00
5.50526798e-01 -1.06562942e-01 2.05226392e-01 1.26804352e-01
7.02315032e-01 4.76803750e-01 -3.91172558e-01 2.27423340e-01
-3.69308852e-02 3.78394406e-03 3.31891365e-02 4.85909194e-01
-7.37018108e-01 6.76075816e-01 5.85148144e+00 6.94404006e-01
-8.14325035e-01 2.16847703e-01 2.92780399e-01 -5.42420924e-01
-9.93139222e-02 2.34198123e-02 -3.90007436e-01 4.26772416e-01
5.81167221e-01 -1.79104686e-01 4.12798554e-01 5.29319048e-01
5.74702084e-01 -1.11136205e-01 -1.29348719e+00 1.33506525e+00
1.60434440e-01 -1.42065871e+00 -6.89065307e-02 1.44649312e-01
7.23870158e-01 -1.58133984e-01 -3.03211957e-01 -1.48174688e-01
5.56908585e-02 -5.55555284e-01 1.08247697e+00 7.67984509e-01
8.13888669e-01 -4.49579149e-01 2.78146923e-01 4.73242223e-01
-1.27812219e+00 6.25473633e-02 -2.47553736e-01 -2.58316696e-01
8.75172079e-01 7.76215434e-01 -3.21177512e-01 7.86631107e-01
8.58688116e-01 1.17899024e+00 -2.96059668e-01 8.28073382e-01
5.59328608e-02 2.13142514e-01 -3.20675373e-01 5.25074363e-01
-2.68023491e-01 -2.98997849e-01 1.03411901e+00 9.02951121e-01
3.68143201e-01 4.87764746e-01 3.49283814e-01 6.52514875e-01
9.69468057e-02 -1.08764611e-01 -9.02217686e-01 1.35124505e-01
2.25990891e-01 1.01640403e+00 -7.68058121e-01 -2.97331184e-01
-6.36726439e-01 1.33945477e+00 2.00223476e-01 3.71332109e-01
-8.84919941e-01 2.58099169e-01 6.33910418e-01 6.09562457e-01
3.50534827e-01 -3.20746422e-01 2.58092374e-01 -1.66938472e+00
3.46123576e-01 -1.04491937e+00 4.39707518e-01 -8.31933200e-01
-1.22250772e+00 3.08873117e-01 2.00901493e-01 -1.47787464e+00
-3.00893843e-01 -2.86879122e-01 -5.09632647e-01 1.67494074e-01
-9.95978773e-01 -1.12308681e+00 -5.88162661e-01 7.98208773e-01
1.06464195e+00 1.50056690e-01 4.27884549e-01 4.77136195e-01
-3.54443073e-01 1.28903478e-01 -4.22020555e-02 1.97280552e-02
7.61115074e-01 -7.61839867e-01 2.44098306e-01 8.68595123e-01
3.56890827e-01 3.19427133e-01 9.16161895e-01 -7.69741893e-01
-1.78349686e+00 -7.39763081e-01 4.07173872e-01 -5.61310410e-01
5.05369008e-01 -4.55243111e-01 -9.11149025e-01 7.66054273e-01
-8.58935639e-02 4.21911851e-02 5.11347592e-01 -7.67504796e-02
-6.28847703e-02 1.83850244e-01 -8.53671193e-01 6.44575059e-01
1.47772598e+00 -5.20299137e-01 -5.90630352e-01 2.93479741e-01
6.85506344e-01 -6.52694881e-01 -8.71508360e-01 2.91176707e-01
7.72427440e-01 -1.31352758e+00 1.27385008e+00 -5.23915887e-01
7.94884741e-01 -9.92339849e-02 -3.40506196e-01 -1.07666898e+00
-1.44367173e-01 -7.90383697e-01 -5.74692070e-01 1.09036803e+00
-3.47869456e-01 -9.77645665e-02 6.09022975e-01 6.32546246e-01
9.80649889e-03 -6.75428748e-01 -8.10743570e-01 -5.99128604e-01
-4.63899612e-01 -6.45211220e-01 2.09092498e-01 9.38683093e-01
-1.93160642e-02 1.88576117e-01 -9.33937967e-01 6.25340715e-02
8.39075685e-01 2.59594083e-01 1.03595650e+00 -1.02210379e+00
-6.25384450e-01 9.06556621e-02 -4.66191232e-01 -1.23506248e+00
-7.69981295e-02 -5.53664982e-01 -5.72375953e-02 -1.21238697e+00
3.10479999e-01 -7.05493465e-02 2.68450826e-01 4.07064632e-02
-2.12088060e-02 2.44722292e-01 3.39791596e-01 3.92650366e-01
-6.27139151e-01 6.63311183e-01 1.42752981e+00 7.90679380e-02
-9.21015516e-02 -6.51210696e-02 -2.13428408e-01 9.93004560e-01
1.30535871e-01 -3.68358165e-01 -3.52650076e-01 -5.48115015e-01
5.15568890e-02 6.04316175e-01 7.17978656e-01 -8.52610290e-01
1.91349000e-01 -1.82926372e-01 3.02103102e-01 -3.96908015e-01
8.15386057e-01 -8.28914940e-01 7.11464047e-01 1.77396342e-01
-3.56434658e-02 -5.37777022e-02 -7.64062703e-02 9.22467470e-01
-2.29105040e-01 -2.72296816e-02 6.48130238e-01 -6.12474918e-01
-7.88866997e-01 5.76065540e-01 -1.08379878e-01 2.28382394e-01
9.19874728e-01 -5.50376594e-01 2.92226136e-01 -7.03235567e-01
-9.62678075e-01 1.63413689e-01 9.04086471e-01 4.93261009e-01
6.29050016e-01 -1.28148937e+00 -6.55866921e-01 2.40870014e-01
-5.41448891e-02 8.96150842e-02 7.49684989e-01 9.51808155e-01
-4.53985333e-01 -2.26178378e-01 -3.29628170e-01 -8.17144454e-01
-1.46062768e+00 5.89379430e-01 2.17302710e-01 5.06433696e-02
-1.32084668e+00 5.84350109e-01 5.34431577e-01 5.61921559e-02
1.91794232e-01 -9.21654254e-02 5.60262464e-02 1.35819912e-01
7.74536610e-01 5.71909130e-01 -1.06403902e-01 -8.87346506e-01
-5.86036667e-02 9.11377430e-01 1.18328758e-01 -1.67219087e-01
1.70901215e+00 -5.53027689e-01 1.72525853e-01 8.93437803e-01
1.15350592e+00 2.04632506e-01 -1.63373077e+00 -2.45875657e-01
-4.05064732e-01 -9.81858552e-01 -1.75926030e-01 -1.32835329e-01
-1.07036269e+00 8.41189802e-01 2.47989938e-01 -7.31347799e-02
9.94053185e-01 3.07654470e-01 1.01298273e+00 5.59113547e-02
7.12721348e-01 -7.89505720e-01 3.39534193e-01 2.11578369e-01
9.42677796e-01 -1.27127469e+00 2.18465656e-01 -7.30277836e-01
-6.27318740e-01 1.08826506e+00 2.38954008e-01 -9.30390358e-02
5.57451308e-01 2.06419691e-01 -2.19014257e-01 -5.63991547e-01
-8.42424810e-01 -1.57495081e-01 1.73837528e-01 5.85201085e-01
5.84937781e-02 -2.73033947e-01 1.37639776e-01 3.52787316e-01
-7.18064755e-02 7.00278906e-03 7.10267603e-01 1.02862322e+00
-1.70915157e-01 -8.73861134e-01 -4.94235575e-01 1.23295844e-01
-4.32076782e-01 1.94890454e-01 -1.21168569e-01 7.53270805e-01
-1.62440035e-02 7.16356933e-01 -6.72645792e-02 -2.44303793e-01
5.55854976e-01 -4.91372868e-02 8.86781991e-01 -5.75247169e-01
-1.27239808e-01 2.85363078e-01 1.95272118e-01 -9.39028859e-01
-7.26185262e-01 -8.93802047e-01 -9.04303908e-01 -5.57042181e-01
-7.41386414e-02 -2.17524245e-01 5.03931753e-02 8.64967048e-01
3.64567399e-01 3.81009132e-01 5.99555075e-01 -1.38000107e+00
-5.14865518e-02 -6.08048856e-01 -6.98256016e-01 9.46818054e-01
5.49789488e-01 -9.89147365e-01 -3.73173535e-01 6.74838603e-01] | [7.340637683868408, -0.7916744947433472] |
36667834-dbf0-4289-8179-25233d5de28c | generating-fluent-fact-checking-explanations | 2112.06924 | null | https://arxiv.org/abs/2112.06924v1 | https://arxiv.org/pdf/2112.06924v1.pdf | Generating Fluent Fact Checking Explanations with Unsupervised Post-Editing | Fact-checking systems have become important tools to verify fake and misguiding news. These systems become more trustworthy when human-readable explanations accompany the veracity labels. However, manual collection of such explanations is expensive and time-consuming. Recent works frame explanation generation as extractive summarization, and propose to automatically select a sufficient subset of the most important facts from the ruling comments (RCs) of a professional journalist to obtain fact-checking explanations. However, these explanations lack fluency and sentence coherence. In this work, we present an iterative edit-based algorithm that uses only phrase-level edits to perform unsupervised post-editing of disconnected RCs. To regulate our editing algorithm, we use a scoring function with components including fluency and semantic preservation. In addition, we show the applicability of our approach in a completely unsupervised setting. We experiment with two benchmark datasets, LIAR-PLUS and PubHealth. We show that our model generates explanations that are fluent, readable, non-redundant, and cover important information for the fact check. | ['Isabelle Augenstein', 'Pepa Atanasova', 'Shailza Jolly'] | 2021-12-13 | null | null | null | null | ['extractive-summarization'] | ['natural-language-processing'] | [ 3.70654017e-01 8.40673685e-01 -4.58733857e-01 -5.53205848e-01
-9.13231671e-01 -6.77203894e-01 7.48273909e-01 8.88901532e-01
-3.78569663e-02 1.16651261e+00 7.56371319e-01 -4.52789307e-01
-1.39359176e-01 -4.24359024e-01 -7.48877525e-01 -1.44832492e-01
2.97651529e-01 5.05545914e-01 -8.82180333e-02 -9.01436508e-02
7.24451959e-01 5.10771200e-02 -1.26351798e+00 6.97752416e-01
1.70759320e+00 2.55461782e-01 -3.55643272e-01 6.86674953e-01
-6.37083426e-02 1.30423200e+00 -8.64263177e-01 -1.16366780e+00
-2.48100400e-01 -1.05074430e+00 -1.14263749e+00 2.31603608e-01
4.24301893e-01 -1.61020696e-01 2.33018640e-02 1.21168232e+00
2.41781771e-01 -1.31966397e-01 6.20954216e-01 -1.30514455e+00
-8.28278720e-01 1.30522835e+00 -3.23437572e-01 8.35126564e-02
8.09110165e-01 -1.93380669e-01 1.14506924e+00 -7.12337494e-01
1.00212514e+00 1.01935875e+00 5.15797019e-01 4.78590786e-01
-9.34532344e-01 -4.37710822e-01 3.89911793e-02 3.94649953e-01
-8.04755628e-01 -5.75749934e-01 5.73613524e-01 -3.71892959e-01
7.32938111e-01 8.01921904e-01 6.45498574e-01 1.13884830e+00
3.60806972e-01 6.67213619e-01 1.07425046e+00 -5.29406726e-01
3.39846522e-01 4.20239329e-01 4.44776535e-01 8.21622789e-01
7.57353604e-01 -5.38369238e-01 -8.65366101e-01 -6.08432651e-01
2.00388283e-01 -8.99761841e-02 -7.38706887e-01 2.05025554e-01
-1.18652701e+00 1.14873791e+00 1.40206113e-01 1.96769148e-01
-5.62857985e-01 -1.42247662e-01 3.43608409e-01 4.45725471e-01
7.86038399e-01 7.70529926e-01 -3.14487755e-01 -2.21206725e-01
-1.14333057e+00 2.43356481e-01 1.11362898e+00 9.39721704e-01
3.07291567e-01 -5.08145034e-01 -5.34531355e-01 7.01401174e-01
-4.55144644e-02 3.48439842e-01 4.46826547e-01 -1.14267206e+00
3.56931716e-01 6.92651212e-01 3.56902152e-01 -1.40413594e+00
-2.51685590e-01 -5.67276776e-01 -9.77751255e-01 -2.37327084e-01
9.85129476e-02 -1.54667059e-02 -5.33925116e-01 1.45598745e+00
3.18777055e-01 -1.68469518e-01 9.58605558e-02 8.74894977e-01
1.04036510e+00 2.70535111e-01 -2.41413385e-01 -7.16435254e-01
1.40090024e+00 -1.06549382e+00 -1.22805643e+00 -3.45103368e-02
7.32112169e-01 -7.76033878e-01 9.57436383e-01 5.63295484e-01
-1.12500036e+00 6.63654432e-02 -9.38704550e-01 -2.57731408e-01
8.89193714e-02 2.90988445e-01 5.44987261e-01 3.52194697e-01
-6.98434889e-01 8.04160476e-01 -5.10079980e-01 -3.89988810e-01
5.34042060e-01 -6.79422766e-02 -5.20944834e-01 4.65751998e-02
-1.15084231e+00 8.04582179e-01 2.62491763e-01 -3.59393388e-01
-2.71438181e-01 -5.99346936e-01 -7.06492722e-01 1.37803838e-01
6.61592782e-01 -1.02462554e+00 1.25894630e+00 -1.10747409e+00
-1.29956865e+00 7.59318769e-01 -4.62048590e-01 -6.90129459e-01
8.12746763e-01 -3.32372397e-01 -4.38676625e-01 5.75712144e-01
4.18160379e-01 2.36558661e-01 7.60128379e-01 -1.20804060e+00
-4.47423726e-01 8.21313933e-02 6.54858202e-02 -4.32343073e-02
-1.41104549e-01 1.12627856e-01 -2.16054410e-01 -6.91021442e-01
2.36524716e-01 -7.52458990e-01 -1.61868826e-01 -2.55188316e-01
-1.07510257e+00 -1.90649912e-01 1.82559133e-01 -1.07384944e+00
1.50542819e+00 -1.76721358e+00 -1.29802585e-01 1.94499105e-01
7.66902030e-01 -6.37059212e-02 1.87542617e-01 5.40197194e-01
8.41860995e-02 6.58507586e-01 -5.31246364e-01 -1.94270357e-01
-1.02170125e-01 7.90105313e-02 -5.31061828e-01 3.25822294e-01
3.73330154e-02 8.88635278e-01 -1.19608521e+00 -8.03654730e-01
-3.16640407e-01 1.14738330e-01 -7.98636019e-01 -9.19020548e-02
-3.09490055e-01 3.57971489e-01 -3.49141598e-01 4.57697660e-01
4.92440969e-01 -6.33180976e-01 2.17269957e-01 4.03622612e-02
8.10449570e-02 8.36826384e-01 -7.28170931e-01 1.40677357e+00
-1.97004020e-01 6.77734673e-01 -2.86330491e-01 -5.77917993e-01
6.46507382e-01 2.83391476e-01 6.64131045e-02 -2.69673824e-01
1.49011731e-01 4.11895722e-01 -2.59526461e-01 -7.47045755e-01
8.47693205e-01 -2.39751562e-01 -2.13628858e-02 8.66490066e-01
-1.89249054e-01 -1.13380603e-01 3.65023196e-01 1.07380164e+00
1.26652718e+00 -3.85056734e-01 8.16273987e-01 -7.13398084e-02
4.19774294e-01 4.00028616e-01 5.31099558e-01 1.12007987e+00
1.57317087e-01 8.09579313e-01 9.46000397e-01 -3.26691270e-01
-9.49667573e-01 -4.17242140e-01 1.20889090e-01 2.97852814e-01
4.20927107e-02 -9.09394681e-01 -9.21452820e-01 -1.11186934e+00
-9.69366506e-02 1.45770276e+00 -7.50647545e-01 -1.49148598e-01
-1.89464882e-01 -3.84712040e-01 5.10887682e-01 -6.78765075e-03
3.04027528e-01 -7.49500632e-01 -7.31975615e-01 2.87465245e-01
-8.94068539e-01 -9.05062318e-01 -5.44356585e-01 -2.10672632e-01
-6.75478041e-01 -1.34674084e+00 -2.95783132e-01 -1.19173415e-01
1.06936729e+00 2.31104329e-01 1.18367207e+00 7.77467251e-01
2.29879022e-01 8.39175209e-02 -6.69835150e-01 -3.77353728e-01
-9.30732667e-01 -6.17760718e-02 -4.56327572e-02 -9.04870406e-02
8.03970620e-02 -4.11263496e-01 -3.68517607e-01 1.06736146e-01
-9.99518812e-01 4.34794158e-01 4.55003619e-01 9.51567948e-01
5.24460077e-01 -1.44384310e-01 5.74806988e-01 -1.73843026e+00
1.03461456e+00 -5.36796451e-01 2.16963161e-02 5.32303274e-01
-7.30164289e-01 2.70765305e-01 8.13600719e-01 -2.45832399e-01
-1.10577118e+00 -2.02163845e-01 2.74583902e-02 -6.42527640e-02
1.91171080e-01 7.32637346e-01 1.63662359e-01 3.97029668e-01
9.21702623e-01 1.69681720e-02 -1.41653895e-01 -2.80527413e-01
6.23249650e-01 7.87455738e-01 7.16722429e-01 -1.79867655e-01
9.33697522e-01 5.27064621e-01 -4.76140290e-01 -4.24344927e-01
-1.36431754e+00 -2.37159759e-01 -3.53866935e-01 -2.05308840e-01
3.47142249e-01 -6.06219709e-01 -5.98337591e-01 -3.52013528e-01
-1.61128390e+00 2.47137487e-01 -3.98484260e-01 3.67295742e-01
-3.40494692e-01 7.25315273e-01 -5.06307721e-01 -7.15461075e-01
-6.03769839e-01 -5.74894726e-01 7.45369852e-01 1.07747149e-02
-9.45307314e-01 -6.52921021e-01 1.24915615e-01 5.66708207e-01
4.67468575e-02 5.13671458e-01 8.34878206e-01 -1.08981478e+00
-2.86816716e-01 -4.56410974e-01 4.31719385e-02 -7.65307546e-02
1.08866848e-01 2.80595988e-01 -6.48088455e-01 1.97318032e-01
1.12062491e-01 -2.90011674e-01 1.01254570e+00 1.52501851e-01
1.05331194e+00 -1.38734496e+00 -3.21069837e-01 8.15787837e-02
1.08411527e+00 -3.74178141e-01 4.71808434e-01 1.09175153e-01
3.76219630e-01 6.24458969e-01 6.79732680e-01 5.80778599e-01
4.33171332e-01 4.45699662e-01 8.43423009e-02 1.68220088e-01
-2.40370515e-03 -5.53378165e-01 2.99993217e-01 1.07443297e+00
-4.60544936e-02 -5.30995488e-01 -5.61605573e-01 5.62240362e-01
-2.01005912e+00 -1.38433266e+00 -6.61844909e-01 1.96032858e+00
1.06699598e+00 2.62048483e-01 -2.02989772e-01 3.18629175e-01
6.79985642e-01 -1.80431768e-01 -1.91021562e-01 -6.23099089e-01
-3.46518874e-01 -1.29939497e-01 2.69729316e-01 7.50677049e-01
-6.71713173e-01 7.14705288e-01 5.81615877e+00 5.49637496e-01
-5.63090742e-01 3.36401999e-01 5.62752664e-01 -2.16698095e-01
-1.08195674e+00 2.19497561e-01 -2.27497652e-01 5.06939530e-01
8.50050271e-01 -4.30588752e-01 3.30752544e-02 8.00763607e-01
5.30231237e-01 -2.96781480e-01 -1.00622129e+00 6.56047165e-01
4.47255343e-01 -1.73063374e+00 2.33035281e-01 -2.55403221e-01
9.65239763e-01 -4.77279246e-01 -3.58425796e-01 -1.49301320e-01
3.89130354e-01 -8.35506022e-01 1.04121435e+00 5.03554404e-01
6.81827188e-01 -5.84795952e-01 1.05904877e+00 3.87170136e-01
-2.76828468e-01 1.21121302e-01 -2.34784469e-01 -2.48147994e-02
4.86548632e-01 1.31309760e+00 -8.94515216e-01 8.27963889e-01
2.03967646e-01 7.72554457e-01 -6.63781464e-01 9.78393614e-01
-9.91736293e-01 7.13724017e-01 -4.70482819e-02 -2.41425693e-01
-1.07377663e-01 1.06707126e-01 6.95374548e-01 1.21433270e+00
3.96986634e-01 5.85562229e-01 -1.15955226e-01 8.86570573e-01
-6.41699076e-01 2.90518165e-01 -4.63826388e-01 -2.03076035e-01
5.57621062e-01 1.13265443e+00 -8.98399055e-01 -8.45080614e-01
5.02746329e-02 1.20166707e+00 3.19438636e-01 -5.61769269e-02
-9.69792902e-01 -2.55289644e-01 1.50507778e-01 7.56669492e-02
-9.20857415e-02 2.46108010e-01 -6.47338033e-01 -1.49568784e+00
2.11960331e-01 -1.23013127e+00 5.46245873e-01 -8.71307254e-01
-1.13110828e+00 7.80610681e-01 -2.39694521e-01 -1.21953988e+00
-2.22375095e-01 2.01135114e-01 -6.51501298e-01 3.10488880e-01
-1.40860999e+00 -5.53813040e-01 -2.99613982e-01 1.58891484e-01
6.20460689e-01 3.07770282e-01 5.54362535e-01 3.88943292e-02
-4.49005336e-01 5.05258620e-01 -2.00925857e-01 -1.14537366e-01
8.83159995e-01 -1.28229785e+00 2.49754384e-01 1.10699475e+00
4.19373125e-01 1.02708125e+00 1.32991445e+00 -1.09773123e+00
-7.76934087e-01 -9.68732417e-01 1.85695958e+00 -6.35241985e-01
4.70183432e-01 -2.95559596e-02 -1.08898652e+00 5.96799791e-01
3.99336398e-01 -6.03488207e-01 8.51989686e-01 1.39130294e-01
-4.34788525e-01 4.81975406e-01 -1.12810445e+00 5.80377042e-01
1.05190206e+00 -4.07192528e-01 -1.05853391e+00 8.95255744e-01
8.87628257e-01 -4.68495607e-01 -2.95413435e-01 -1.16258010e-01
3.00467670e-01 -1.02906632e+00 3.87581915e-01 -7.85634518e-01
1.10935915e+00 -4.37904924e-01 4.63064909e-01 -1.34854698e+00
-2.22641855e-01 -9.13569152e-01 -3.80022638e-02 1.20519876e+00
8.30471158e-01 -4.38341260e-01 3.17100167e-01 6.45853281e-01
-2.70422637e-01 -3.88057858e-01 -7.52882004e-01 -5.46418071e-01
-4.85055476e-01 -2.35688075e-01 5.02665877e-01 1.28674936e+00
8.25999141e-01 3.89853537e-01 -6.13023400e-01 7.87560716e-02
5.13017237e-01 2.67471641e-01 5.60783505e-01 -1.31653214e+00
-3.21996719e-01 -2.32021317e-01 1.09088294e-01 -6.40955150e-01
1.40755445e-01 -9.09943879e-01 1.16945684e-01 -1.84126842e+00
7.88257360e-01 5.51512241e-02 3.02327484e-01 6.85312331e-01
-5.29320717e-01 5.81858344e-02 -8.64206180e-02 5.46349645e-01
-9.26768363e-01 4.70293403e-01 9.82707441e-01 -6.96449578e-02
-1.84583902e-01 -8.58809650e-02 -1.18367815e+00 8.55700850e-01
8.65793288e-01 -1.01527643e+00 -2.50960588e-01 -1.93721965e-01
6.65480018e-01 7.78832333e-03 3.46246064e-01 -4.92652297e-01
3.93508166e-01 -2.27916747e-01 -3.36138532e-02 -4.47889328e-01
-3.15288454e-01 -3.80245268e-01 3.25002044e-01 5.76159716e-01
-8.19979548e-01 3.73795033e-02 -3.41968060e-01 7.11750627e-01
-2.67733693e-01 -3.66864681e-01 5.45374274e-01 -1.32276058e-01
1.10037833e-01 -1.34852007e-01 -4.64201748e-01 1.90757319e-01
6.31847501e-01 -1.99552439e-02 -7.53830194e-01 -8.72008026e-01
-3.61363411e-01 7.78758228e-02 6.22785211e-01 1.53702721e-01
8.01025391e-01 -1.09146094e+00 -1.23743188e+00 -3.47058117e-01
2.12920964e-01 -4.56861764e-01 1.41666561e-01 1.10959804e+00
-5.64971089e-01 3.75858158e-01 9.29269493e-02 -7.66168907e-02
-1.34909439e+00 5.57020724e-01 -1.96542010e-01 -2.36722231e-01
-9.55060124e-01 6.34966969e-01 -3.03741425e-01 -1.85956419e-01
-5.28236404e-02 -4.17474926e-01 -3.71234119e-01 8.53620097e-02
8.84104311e-01 3.55358541e-01 1.17511235e-01 -3.76399755e-01
-3.75536233e-01 -8.43259618e-02 -3.25346217e-02 -2.16625616e-01
1.36521816e+00 -2.80628890e-01 -3.71987551e-01 2.25230813e-01
6.86189592e-01 6.57320678e-01 -6.08402431e-01 -2.87469826e-03
2.65083611e-01 -4.84409004e-01 -3.18853825e-01 -1.05696130e+00
-7.35210001e-01 3.82937521e-01 -5.68788111e-01 5.69771111e-01
8.49906504e-01 2.00715110e-01 6.21974647e-01 2.96524554e-01
-1.64444372e-02 -1.05578530e+00 -2.52846535e-02 1.79446802e-01
1.24029422e+00 -1.12812757e+00 3.79149318e-01 -8.19155812e-01
-9.46301818e-01 1.11272824e+00 9.50785056e-02 4.11956459e-01
4.94919345e-02 -2.19560415e-01 -4.27762568e-02 -5.67615151e-01
-1.07040906e+00 3.25460523e-01 4.14480597e-01 1.56748533e-01
5.90417027e-01 9.20768231e-02 -1.11094761e+00 9.84112859e-01
-6.91651940e-01 -3.45161855e-02 1.38524437e+00 6.18559718e-01
-5.45629978e-01 -7.79127240e-01 -2.94348061e-01 6.81234896e-01
-5.89727283e-01 -3.77997577e-01 -1.04798245e+00 4.28617775e-01
-1.73688591e-01 1.44979417e+00 -4.90937442e-01 -1.92400560e-01
1.45784721e-01 1.05289303e-01 9.80043262e-02 -6.31875813e-01
-7.86033034e-01 -3.19855094e-01 6.75123870e-01 -4.72462863e-01
-6.21193409e-01 -6.31071985e-01 -1.42096198e+00 -6.55138433e-01
-5.77322900e-01 7.23141789e-01 3.98313254e-01 1.22732174e+00
4.86829758e-01 3.05006593e-01 4.02741045e-01 1.14916898e-01
-4.31844413e-01 -8.20904911e-01 -2.25767523e-01 6.77134931e-01
3.10208887e-01 -3.04985642e-01 -6.24700010e-01 3.77464920e-01] | [12.150188446044922, 9.304431915283203] |
ba37d5be-b4f2-4caf-a4d3-de931a0ea683 | a-hierarchical-hybrid-learning-framework-for | 2303.12274 | null | https://arxiv.org/abs/2303.12274v3 | https://arxiv.org/pdf/2303.12274v3.pdf | A Hierarchical Hybrid Learning Framework for Multi-agent Trajectory Prediction | Accurate and robust trajectory prediction of neighboring agents is critical for autonomous vehicles traversing in complex scenes. Most methods proposed in recent years are deep learning-based due to their strength in encoding complex interactions. However, unplausible predictions are often generated since they rely heavily on past observations and cannot effectively capture the transient and contingency interactions from sparse samples. In this paper, we propose a hierarchical hybrid framework of deep learning (DL) and reinforcement learning (RL) for multi-agent trajectory prediction, to cope with the challenge of predicting motions shaped by multi-scale interactions. In the DL stage, the traffic scene is divided into multiple intermediate-scale heterogenous graphs based on which Transformer-style GNNs are adopted to encode heterogenous interactions at intermediate and global levels. In the RL stage, we divide the traffic scene into local sub-scenes utilizing the key future points predicted in the DL stage. To emulate the motion planning procedure so as to produce trajectory predictions, a Transformer-based Proximal Policy Optimization (PPO) incorporated with a vehicle kinematics model is devised to plan motions under the dominant influence of microscopic interactions. A multi-objective reward is designed to balance between agent-centric accuracy and scene-wise compatibility. Experimental results show that our proposal matches the state-of-the-arts on the Argoverse forecasting benchmark. It's also revealed by the visualized results that the hierarchical learning framework captures the multi-scale interactions and improves the feasibility and compliance of the predicted trajectories. | ['Bo Tao', 'Linzhen Nie', 'Xu Zhu', 'Chunyuan Lei', 'Zhishuai Yin', 'Mingze Miao', 'Yujun Jiao'] | 2023-03-22 | null | null | null | null | ['trajectory-prediction', 'motion-planning'] | ['computer-vision', 'robots'] | [-3.41925383e-01 -4.97395843e-02 -3.75411332e-01 -1.09457821e-01
-5.26330590e-01 -1.77561894e-01 7.68294454e-01 3.63033824e-02
-1.50089309e-01 8.63935828e-01 2.41269380e-01 -2.91282058e-01
-3.18729222e-01 -9.93958116e-01 -8.66273820e-01 -8.82100582e-01
-4.43098009e-01 6.82007253e-01 3.96424949e-01 -4.96236205e-01
1.27509544e-02 8.00595105e-01 -1.46500480e+00 1.24586381e-01
1.16196156e+00 8.23183239e-01 5.40014148e-01 6.40551567e-01
-3.02245747e-02 1.19869626e+00 6.77297339e-02 -1.79536641e-01
2.63832957e-01 -1.68645859e-01 -4.51726973e-01 1.90252721e-01
-5.47658838e-02 -3.69559914e-01 -7.67177939e-01 6.82998300e-01
2.04120815e-01 4.86623555e-01 5.64166427e-01 -1.53732741e+00
-4.79928823e-03 3.68508488e-01 -4.45183486e-01 9.47678536e-02
8.07745159e-02 7.26862788e-01 8.89955044e-01 -5.36681116e-01
7.03272998e-01 1.38314283e+00 4.72501844e-01 3.04250956e-01
-8.62421870e-01 -3.90201598e-01 6.72978997e-01 6.71784639e-01
-1.04344296e+00 -1.63886145e-01 9.33027804e-01 -5.31792462e-01
9.36296344e-01 1.48603186e-01 6.73392713e-01 1.19817042e+00
8.25527251e-01 9.80676889e-01 6.13614738e-01 2.06548944e-01
3.65967661e-01 -1.60290211e-01 -2.46114105e-01 7.81796753e-01
-1.31309435e-01 5.08747816e-01 -1.65619701e-01 -1.20408699e-01
5.55676818e-01 2.17902377e-01 1.03861064e-01 -5.21163344e-01
-1.20454335e+00 7.47657120e-01 7.17819035e-01 -4.72675450e-03
-9.31912839e-01 3.02380592e-01 4.03869033e-01 -2.02140436e-02
2.96696723e-01 8.04056451e-02 -3.02647322e-01 -1.53572291e-01
-6.00866437e-01 6.27459645e-01 6.80032492e-01 9.44939613e-01
8.79040778e-01 3.10509920e-01 -2.96371609e-01 4.47411060e-01
3.32009465e-01 4.58615690e-01 4.97793667e-02 -1.10850036e+00
5.41033387e-01 6.16947234e-01 3.87221426e-01 -1.27309644e+00
-6.27353966e-01 -5.84893942e-01 -1.06573081e+00 3.14817935e-01
3.10275108e-01 -3.26867759e-01 -6.68563068e-01 1.70473385e+00
5.82000554e-01 7.35809088e-01 1.47165015e-01 9.62819219e-01
2.85601705e-01 1.04310250e+00 1.76991105e-01 -2.45058760e-01
7.88837135e-01 -1.25636709e+00 -4.93106991e-01 -1.20368049e-01
7.95721948e-01 -2.34647200e-01 6.33473933e-01 -4.52713966e-02
-1.00841987e+00 -7.64405489e-01 -6.68356478e-01 3.69884491e-01
-2.58163691e-01 -3.40959313e-03 4.21972245e-01 -1.05315074e-03
-1.02288342e+00 4.94962364e-01 -9.83309627e-01 -1.73084900e-01
3.47206801e-01 2.76374131e-01 2.66800933e-02 -1.19523719e-01
-1.11743987e+00 9.48317826e-01 1.68130517e-01 2.98397332e-01
-1.39530671e+00 -6.59867823e-01 -7.19601870e-01 4.25074324e-02
5.46479225e-01 -6.87539756e-01 9.02102053e-01 -6.14788890e-01
-1.51666343e+00 -2.01553404e-02 -2.60462880e-01 -6.53301179e-01
9.19027209e-01 4.07067090e-02 -3.52737904e-01 -3.99501808e-02
1.53735995e-01 6.08670890e-01 6.42342329e-01 -1.44657540e+00
-1.07452059e+00 2.56784167e-02 3.00059110e-01 5.43730974e-01
2.28761420e-01 -5.51910579e-01 -4.44389462e-01 -3.78965259e-01
-3.13242614e-01 -1.25736988e+00 -9.00168896e-01 -3.91098320e-01
-2.70605505e-01 -2.76821226e-01 1.04993236e+00 -4.55386609e-01
9.57606733e-01 -1.90908849e+00 3.65136802e-01 2.37275466e-01
3.05247128e-01 1.52573943e-01 -3.38046402e-01 8.43885243e-01
4.83978271e-01 -3.90540481e-01 2.80583594e-02 -4.11897629e-01
1.74481526e-01 4.25851166e-01 -3.98963600e-01 4.71236676e-01
-3.92086990e-02 1.03501046e+00 -1.13169992e+00 -2.67121941e-01
5.32239914e-01 3.28038812e-01 -5.25599778e-01 3.57305527e-01
-6.53169036e-01 9.26136911e-01 -9.50064957e-01 4.39700454e-01
6.28297925e-01 -1.48698077e-01 1.59716025e-01 -3.55580710e-02
-4.00647998e-01 -1.91117048e-01 -9.21201169e-01 1.34400094e+00
-4.72412676e-01 4.74258155e-01 1.26703367e-01 -1.00253892e+00
7.78757513e-01 1.82663292e-01 8.48565340e-01 -7.93976247e-01
7.10022077e-03 1.18581109e-01 -8.24607983e-02 -6.34140849e-01
6.99922621e-01 2.66078115e-01 -4.32327092e-02 9.27550718e-02
-4.82158065e-01 1.46611914e-01 1.85999021e-01 1.66589588e-01
1.11046314e+00 3.03328693e-01 3.82391713e-03 -1.24002054e-01
6.79431975e-01 4.04839128e-01 7.80372620e-01 6.51093364e-01
-2.51353174e-01 -1.02543337e-02 3.97876292e-01 -7.96869814e-01
-1.08151889e+00 -7.24054277e-01 4.57671344e-01 9.62905824e-01
6.26856983e-01 -1.31449372e-01 -5.06610930e-01 -6.05103076e-01
-1.58641059e-02 7.11748421e-01 -4.72852528e-01 -9.78348404e-02
-9.28833008e-01 -6.85042500e-01 8.39638337e-03 3.57844204e-01
5.16182959e-01 -1.13527274e+00 -7.05564082e-01 7.05131769e-01
-1.76950887e-01 -1.28050923e+00 -3.17812771e-01 -3.71094763e-01
-4.77294505e-01 -1.00945199e+00 -3.95072818e-01 -5.27019441e-01
5.43220401e-01 3.56162935e-01 6.25802398e-01 2.86986735e-02
1.05539471e-01 3.53052795e-01 -4.26460654e-01 -1.23381324e-04
-5.63998580e-01 6.79023117e-02 1.22495249e-01 3.72416675e-01
-4.22924310e-02 -7.59365261e-01 -6.76564038e-01 4.01672304e-01
-6.13248825e-01 3.34430516e-01 8.04432511e-01 7.76724756e-01
4.68447268e-01 2.98074424e-01 6.39342964e-01 -4.60600495e-01
3.25057179e-01 -9.98051941e-01 -7.10780144e-01 8.62067044e-02
-3.18349481e-01 -7.73863941e-02 1.07290721e+00 -4.36971635e-01
-1.11115634e+00 2.16826022e-01 -1.44736648e-01 -5.72676957e-01
-2.95589358e-01 4.96819824e-01 -7.77885243e-02 -3.78882065e-02
1.42703623e-01 3.52934688e-01 -1.16807371e-01 7.60186389e-02
3.69474858e-01 1.23827301e-01 3.18595201e-01 -6.20915473e-01
8.31869781e-01 5.73665202e-01 2.79873163e-01 -6.63141012e-01
-3.12154561e-01 -1.52476162e-01 -4.43636656e-01 -8.20880830e-01
9.38394666e-01 -9.52676952e-01 -1.13852406e+00 5.10945559e-01
-1.17380583e+00 -7.99972117e-01 -1.24120213e-01 6.99156523e-01
-8.75447512e-01 1.72080368e-01 -5.63412488e-01 -9.04369056e-01
1.90604448e-01 -1.50807178e+00 8.75529051e-01 2.27629960e-01
2.06412315e-01 -1.13556623e+00 1.82287529e-01 1.74443945e-01
3.80012184e-01 5.57000935e-01 9.48034525e-01 -2.10295320e-01
-1.22776389e+00 -8.17952752e-02 -2.10931432e-02 -3.48363549e-01
-2.35506833e-01 -6.09907992e-02 -4.57487226e-01 -3.92740428e-01
-2.77406603e-01 -2.75313482e-02 7.05800653e-01 6.30867600e-01
8.60171020e-01 -5.55392146e-01 -7.11168468e-01 4.66306359e-01
1.33284032e+00 4.43727702e-01 4.71305370e-01 3.36479604e-01
9.63934302e-01 7.88177848e-01 8.74733150e-01 5.76552391e-01
1.12279034e+00 8.05674493e-01 9.74059463e-01 -4.90211807e-02
1.16850182e-01 -3.84249419e-01 4.84520048e-01 6.24038219e-01
-3.62117141e-02 -5.92700660e-01 -1.06150675e+00 6.90221190e-01
-2.51893306e+00 -1.23790395e+00 -2.58899212e-01 1.72718501e+00
7.44922534e-02 1.48106769e-01 2.61850566e-01 -4.58485246e-01
5.62336683e-01 5.35453737e-01 -7.50382841e-01 -1.38628006e-01
4.98121679e-02 -7.63517141e-01 6.19057596e-01 7.68225193e-01
-8.60556901e-01 1.06402075e+00 5.25296307e+00 9.88481462e-01
-1.11091590e+00 -5.78082213e-03 7.54618883e-01 6.34229779e-02
-3.85901749e-01 -4.57209386e-02 -8.84311616e-01 5.79279363e-01
9.13139701e-01 -1.07392576e-02 4.83702391e-01 8.19892883e-01
9.02222991e-01 -1.03286564e-01 -8.54735494e-01 5.30807793e-01
-4.78290558e-01 -1.70135641e+00 1.92459431e-02 2.67757684e-01
1.01702476e+00 3.90910119e-01 1.30031809e-01 5.41038156e-01
7.01920927e-01 -8.79971206e-01 7.48419762e-01 7.97496378e-01
8.40630904e-02 -9.63929534e-01 6.24594271e-01 8.62254679e-01
-1.63390303e+00 -5.32379746e-01 -1.74020141e-01 -1.70235619e-01
6.10634029e-01 2.60739744e-01 -7.95205116e-01 9.41785455e-01
3.92198980e-01 1.05782235e+00 -1.09204426e-01 9.44860637e-01
1.73462167e-01 4.25366074e-01 -1.52481481e-01 -1.43050507e-01
9.84891474e-01 -6.24841452e-01 9.65206504e-01 9.15309548e-01
3.84275168e-01 1.13324113e-01 7.43334174e-01 7.58547783e-01
3.38666528e-01 -1.18815996e-01 -8.08469713e-01 2.86620229e-01
4.44313526e-01 1.16172087e+00 -5.56142509e-01 -3.97423714e-01
-3.94896895e-01 5.13058960e-01 4.46103930e-01 7.58674979e-01
-1.15589833e+00 3.81546527e-01 8.81175518e-01 7.24509358e-02
4.20136005e-01 -5.85176110e-01 -8.74496624e-02 -8.90424371e-01
-9.66924429e-02 -6.17078662e-01 9.70769965e-04 -3.77382547e-01
-1.08324218e+00 6.59229159e-01 -2.62936875e-02 -1.47371864e+00
-4.75933909e-01 -1.51144192e-01 -9.57556009e-01 5.98149776e-01
-1.71614134e+00 -1.48846364e+00 -2.46194839e-01 5.65257370e-01
7.99597025e-01 -2.11769104e-01 2.46262789e-01 1.68439388e-01
-7.30320752e-01 2.47596726e-02 3.34132642e-01 -3.00215900e-01
-4.22660373e-02 -9.42090213e-01 3.74167442e-01 8.43240857e-01
-3.80548626e-01 -4.95182425e-02 8.16128254e-01 -8.32430661e-01
-1.61326265e+00 -1.53210032e+00 6.88279867e-01 -1.65441692e-01
8.16646397e-01 -1.09505147e-01 -7.02937007e-01 6.03326619e-01
-6.90573379e-02 1.02812849e-01 1.68374944e-02 -3.66182715e-01
2.28169113e-01 -2.98103750e-01 -8.18057775e-01 8.57502997e-01
9.51385796e-01 -1.93372518e-01 -3.89764681e-02 3.17610592e-01
6.11054003e-01 -5.25862873e-01 -5.52980065e-01 5.88904917e-01
4.27950382e-01 -8.63632023e-01 8.28852177e-01 -5.62483191e-01
3.35045487e-01 -3.60160321e-01 -8.57017487e-02 -1.38684118e+00
-5.24147391e-01 -6.47149682e-01 -2.35656947e-01 8.65561664e-01
1.94029599e-01 -5.89924872e-01 9.11339760e-01 5.65409541e-01
-5.49762785e-01 -1.06055641e+00 -1.07097375e+00 -6.48443162e-01
-1.19853735e-01 -4.44446146e-01 7.50850916e-01 6.84840202e-01
-3.13741982e-01 3.75155434e-02 -7.53085434e-01 6.06489301e-01
6.56161964e-01 2.62853175e-01 1.07246399e+00 -8.13280702e-01
-1.75500110e-01 -5.81086755e-01 -2.82394469e-01 -1.24968612e+00
5.04498780e-01 -6.21635139e-01 1.61533117e-01 -1.55525208e+00
-5.34902662e-02 -6.50530040e-01 -1.17871918e-01 4.29177582e-02
-9.84724239e-02 -3.70991766e-01 3.25506777e-01 3.56697738e-01
-7.92079389e-01 1.07413447e+00 1.41185427e+00 -1.66597202e-01
-3.38996589e-01 2.69749433e-01 1.87766831e-02 6.18709743e-01
7.49892771e-01 -2.80908465e-01 -6.85779274e-01 -3.92739058e-01
-1.74790435e-02 6.30755842e-01 5.85869074e-01 -1.09607708e+00
5.70289195e-01 -7.47162998e-01 -1.09380171e-01 -1.02168250e+00
4.71041858e-01 -8.70914102e-01 4.18527603e-01 6.08568490e-01
-3.47077936e-01 2.25491196e-01 5.37734292e-02 1.00668848e+00
-2.09296733e-01 4.96036828e-01 5.95918000e-01 -3.98544408e-02
-1.06219769e+00 9.49002564e-01 -6.95258379e-01 -1.82876229e-01
1.43346858e+00 -5.23221865e-02 -2.02772364e-01 -5.31801641e-01
-6.11854494e-01 8.77870500e-01 3.50540012e-01 4.18216348e-01
5.43556035e-01 -1.39088750e+00 -7.43278921e-01 -6.35302365e-02
-1.69156432e-01 2.91260686e-02 8.24306369e-01 1.09704256e+00
-4.59216356e-01 4.50649112e-01 -3.14188838e-01 -6.14564121e-01
-7.00568795e-01 5.36229014e-01 3.58113199e-01 -7.08362818e-01
-7.87247419e-01 4.78589892e-01 4.74170983e-01 -4.19269651e-01
1.35558113e-01 -2.27139369e-01 -3.84768426e-01 -1.07106157e-01
2.56429076e-01 6.81552529e-01 -4.17064995e-01 -9.67860341e-01
-1.05468556e-01 3.90830487e-01 6.00124374e-02 9.66497511e-02
1.38033962e+00 -4.83697116e-01 2.48194590e-01 2.88466543e-01
9.54198897e-01 -1.89053729e-01 -2.07976961e+00 -1.18987404e-01
-1.49228558e-01 -1.81638569e-01 -6.28917366e-02 -4.23694849e-01
-1.02546382e+00 6.59662008e-01 2.14039981e-01 1.21524036e-01
7.48434305e-01 -2.17214838e-01 1.25635540e+00 2.88071126e-01
5.83012044e-01 -1.03675699e+00 2.31250431e-02 7.46224642e-01
8.30554068e-01 -1.11974311e+00 -3.71685088e-01 -2.48749897e-01
-8.72043014e-01 9.76081610e-01 8.75482321e-01 -3.57701898e-01
4.95485932e-01 2.64792945e-02 -2.03669101e-01 -1.32736504e-01
-1.10854423e+00 -2.49374598e-01 1.25614360e-01 4.81271267e-01
-4.48001504e-01 2.05709875e-01 -4.47767526e-02 2.34631330e-01
1.14219435e-01 -2.46816382e-01 3.57950926e-01 6.48618281e-01
-4.70707625e-01 -8.72555435e-01 -8.63949060e-02 2.38341093e-01
2.62718618e-01 3.98311347e-01 1.75897628e-01 7.81445622e-01
2.39391491e-01 9.75381553e-01 6.44650757e-02 -5.87702155e-01
8.60682204e-02 -5.80477715e-01 -5.79945631e-02 -1.17237613e-01
-3.89652997e-01 -4.67192940e-02 1.02344617e-01 -1.04766679e+00
-2.86546022e-01 -7.23536670e-01 -1.37209260e+00 -6.86470687e-01
2.51704365e-01 1.74403876e-01 3.33356798e-01 1.22465658e+00
6.24048471e-01 6.63315892e-01 9.36109245e-01 -1.42101502e+00
-3.69770646e-01 -5.10855436e-01 -2.63611406e-01 2.13323832e-01
6.47198260e-01 -7.88408160e-01 -1.00119568e-01 -3.30439508e-01] | [5.907289028167725, 0.856507420539856] |
96cd1c3b-dd5e-4d6f-9898-7c857ef0c2e8 | makadi-a-large-scale-human-labeled-dataset | null | null | https://aclanthology.org/2022.wildre-1.12 | https://aclanthology.org/2022.wildre-1.12.pdf | Makadi: A Large-Scale Human-Labeled Dataset for Hindi Semantic Parsing | Parsing natural language queries into formal database calls is a very well-studied problem. Because of the rich diversity of semantic markers across the world’s languages, progress in solving this problem is irreducibly language-dependent. This has created an asymmetry in progress in NLIDB solutions, with most state-of-the-art efforts focused on the resource-rich English language, with limited progress seen for low resource languages. In this short paper, we present Makadi, a large-scale, complex, cross-lingual, cross-domain semantic parsing and text-to-SQL dataset for semantic parsing in the Hindi language. Produced by translating the recently introduced English language Spider NLIDB dataset, it consists of 9693 questions and SQL queries on 166 databases with multiple tables which cover multiple domains. This is the first large-scale dataset in the Hindi language for semantic parsing and related language understanding tasks. Our dataset is publicly available at: Link removed to preserve anonymization during peer review. | ['Nisheeth Srivastava', 'Shashwat Vaibhav'] | null | null | null | null | wildre-lrec-2022-6 | ['text-to-sql'] | ['computer-code'] | [-1.50116369e-01 3.68222088e-01 -3.58523101e-01 -9.27216828e-01
-1.27551961e+00 -1.08840764e+00 2.35446423e-01 4.09001887e-01
-3.07794511e-01 9.44305897e-01 4.59242433e-01 -5.80899060e-01
-8.45648870e-02 -8.86559844e-01 -7.54647255e-01 2.70144671e-01
3.98015440e-01 1.30109930e+00 4.23824102e-01 -4.69091922e-01
1.10340528e-01 3.37075561e-01 -1.04503965e+00 6.55916393e-01
7.82596469e-01 8.95941496e-01 2.69364828e-04 3.79069597e-01
-7.90566504e-01 8.26046348e-01 -5.41111887e-01 -1.02338922e+00
2.36570969e-01 -2.43529409e-01 -1.52050090e+00 -5.93405664e-01
4.13151562e-01 8.97214636e-02 7.73598766e-03 9.98672307e-01
4.22726870e-01 -5.52177787e-01 -1.59222670e-02 -1.36210144e+00
-6.42617881e-01 8.57482135e-01 -2.49397427e-01 -1.71378300e-01
7.38989532e-01 -4.09472227e-01 1.38331461e+00 -2.26495206e-01
1.26015580e+00 1.55578911e+00 5.45125306e-01 5.66736877e-01
-1.09469223e+00 -6.95440352e-01 -3.04044217e-01 -5.38579114e-02
-1.10913479e+00 -2.07916573e-01 4.58382815e-01 -2.46477544e-01
1.24338078e+00 2.39767954e-01 -1.27163127e-01 9.83718932e-01
-3.62059802e-01 6.85317934e-01 1.16569042e+00 -5.20652831e-01
4.75076474e-02 6.16890728e-01 3.50429207e-01 6.90733671e-01
4.16073352e-01 -3.78344506e-01 -5.97206295e-01 -4.77464586e-01
1.29452676e-01 -4.35023934e-01 3.53723615e-01 -5.10398209e-01
-9.00569201e-01 1.23334503e+00 -3.64189893e-02 6.60551265e-02
7.49153048e-02 -3.58498663e-01 6.78632915e-01 5.26226461e-01
2.47114986e-01 3.52914900e-01 -1.04586256e+00 -3.86141360e-01
-4.21501577e-01 6.65744424e-01 1.61541045e+00 1.46441460e+00
9.28408325e-01 -7.74918139e-01 5.77376544e-01 1.09656167e+00
2.55182266e-01 4.92533565e-01 4.41062152e-01 -1.11424983e+00
1.16541839e+00 1.02623892e+00 -1.12498164e-01 -8.97038281e-01
-3.69395286e-01 5.71485102e-01 -2.02723652e-01 -9.91043299e-02
8.62691820e-01 2.81742681e-02 -1.83444053e-01 1.84641469e+00
5.08544683e-01 -1.09243238e+00 5.31042039e-01 6.83229566e-01
9.69775736e-01 3.29724759e-01 2.12305635e-01 5.47057033e-01
1.94292617e+00 -5.22041380e-01 -4.34716523e-01 -4.10638183e-01
8.40692103e-01 -9.37437832e-01 1.36398482e+00 1.91892505e-01
-9.27209258e-01 -1.97693229e-01 -7.88376987e-01 -7.24435747e-01
-1.13857996e+00 -2.77150571e-01 7.42931426e-01 9.48869348e-01
-8.37736070e-01 -3.02496850e-02 -5.90358853e-01 -1.09582829e+00
5.07247388e-01 4.86990176e-02 -8.80000532e-01 -5.19057870e-01
-1.33598435e+00 7.31576562e-01 5.63119888e-01 -4.85771298e-01
-2.96800435e-01 -7.58201838e-01 -8.21067750e-01 -3.61241370e-01
8.51907194e-01 -4.35490578e-01 1.04292941e+00 -3.46760929e-01
-9.95827496e-01 1.59780264e+00 -2.01151192e-01 -5.69357038e-01
6.25795960e-01 -3.23690712e-01 -3.99716914e-01 -4.83469199e-03
5.37455499e-01 7.59226859e-01 1.16154864e-01 -9.79673922e-01
-5.94332218e-01 -1.09549284e+00 1.58486769e-01 -7.98875764e-02
8.19427967e-02 6.46223545e-01 -4.76430297e-01 -5.33973932e-01
3.28836516e-02 -9.23353732e-01 9.67057198e-02 -1.63075879e-01
-5.28057218e-01 -2.37170801e-01 6.94546282e-01 -1.26542652e+00
1.20173919e+00 -2.12781262e+00 -2.42151514e-01 -6.96243495e-02
-1.37786090e-01 9.10715200e-04 7.32745603e-02 8.35403502e-01
2.53338199e-02 4.24299419e-01 -5.13594687e-01 -1.83736570e-02
3.06692153e-01 5.80819666e-01 -6.47379637e-01 -3.95540334e-02
2.14032903e-01 1.06786287e+00 -6.16182685e-01 -8.16942811e-01
-3.09980869e-01 -6.95658401e-02 -8.53673041e-01 2.62672722e-01
-5.60405135e-01 1.26817048e-01 -7.39734113e-01 9.09269214e-01
6.49688423e-01 -2.13576388e-02 4.91350889e-01 5.70558645e-02
8.33081082e-02 5.91230452e-01 -1.05925000e+00 2.11841202e+00
-3.71949524e-01 3.01669717e-01 2.24530488e-01 -1.13183594e+00
1.07056451e+00 5.02828136e-02 3.90596896e-01 -7.44745553e-01
-3.33640873e-01 5.58050692e-01 -5.76197147e-01 -6.40722692e-01
4.57677066e-01 -1.30860597e-01 -1.00793529e+00 4.47365224e-01
1.06126759e-02 -4.21131074e-01 2.01219395e-01 4.38128233e-01
9.99648988e-01 2.79798418e-01 4.73856241e-01 -3.81963640e-01
5.47094166e-01 7.94398963e-01 6.01079941e-01 6.34837210e-01
-1.08073719e-01 4.23967570e-01 8.58209074e-01 -4.65021521e-01
-9.94099379e-01 -1.09634864e+00 -2.62995631e-01 1.16402721e+00
-1.10679775e-01 -4.51802641e-01 -1.11438715e+00 -8.89342189e-01
3.39126050e-01 8.23335230e-01 -3.28659743e-01 1.36999369e-01
-8.02994907e-01 -5.24552584e-01 1.16063118e+00 2.35506594e-01
6.53129220e-01 -1.20154154e+00 -4.02784765e-01 3.01915824e-01
-5.58306873e-01 -1.60139287e+00 -1.03463501e-01 6.49179518e-02
-7.22473979e-01 -1.28940642e+00 -1.96618110e-01 -9.72829938e-01
1.09379187e-01 -3.74844432e-01 1.69204485e+00 -4.95866597e-01
-7.32169390e-01 3.86557430e-01 -2.96978354e-01 -7.33817935e-01
-6.65898204e-01 3.92709553e-01 -6.45088673e-01 -6.20055974e-01
1.16737902e+00 -1.26846775e-01 -1.16104051e-01 3.72181147e-01
-1.09608161e+00 -3.51945251e-01 -1.44634275e-02 5.38880765e-01
5.38145244e-01 -1.62450373e-01 5.40985584e-01 -1.66560125e+00
5.32596290e-01 -5.88510811e-01 -9.03496146e-01 6.92172110e-01
-3.86340559e-01 2.56007373e-01 6.75546885e-01 3.88414294e-01
-1.32079601e+00 -1.24177150e-01 -1.34648025e-01 5.31721473e-01
-5.80179691e-01 5.02464056e-01 -7.29610026e-01 3.10656577e-01
4.92164493e-01 -2.47664094e-01 2.52426174e-02 -8.56164932e-01
7.16756582e-01 9.21354413e-01 5.69445252e-01 -9.77487803e-01
3.97997767e-01 5.28213739e-01 -3.52960080e-01 -8.58376920e-01
-7.69953072e-01 -3.87105495e-01 -7.05906510e-01 5.25118709e-01
1.27527189e+00 -1.03875744e+00 -5.43269694e-01 5.02027214e-01
-1.01538849e+00 -1.47749931e-01 -2.92988509e-01 -4.06098226e-03
-6.11198962e-01 3.08324099e-01 -3.85838330e-01 -3.27400506e-01
-2.20450640e-01 -1.00350714e+00 1.16248989e+00 -2.67178476e-01
-6.82770669e-01 -9.92092311e-01 2.20934719e-01 1.03855562e+00
1.39828011e-01 8.16534162e-02 1.67284238e+00 -1.20922899e+00
-6.41763330e-01 -2.89498687e-01 -4.24627393e-01 1.48154005e-01
-6.99140057e-02 -4.14754450e-01 -6.36321247e-01 2.27213521e-02
7.74655119e-02 -9.39849854e-01 1.91565126e-01 -1.09956851e-02
8.98688495e-01 -3.15321296e-01 -1.04796685e-01 7.11771071e-01
1.73992682e+00 2.32734486e-01 6.30090117e-01 6.46480381e-01
6.11792445e-01 1.17631161e+00 7.69830287e-01 4.45916623e-01
9.85103786e-01 4.86886173e-01 -8.38152412e-03 1.01228721e-01
6.68410137e-02 -6.53319657e-01 -1.02590591e-01 5.78460038e-01
7.12610304e-01 -7.64673799e-02 -1.26157367e+00 4.94725883e-01
-1.54498672e+00 -5.68435729e-01 2.67366581e-02 2.04825664e+00
1.04210174e+00 -1.88529402e-01 1.71300560e-01 -4.51963127e-01
4.68400896e-01 -1.76367592e-02 -5.97391665e-01 -7.37482071e-01
-4.30364609e-01 4.40038592e-01 7.13416815e-01 5.98385513e-01
-9.70737576e-01 1.36018801e+00 5.98052406e+00 5.92075109e-01
-7.05936074e-01 1.14100426e-01 6.16443574e-01 2.02988550e-01
-5.52501738e-01 2.86460608e-01 -1.14242566e+00 4.32018638e-01
1.08995044e+00 -3.11702967e-01 7.26527035e-01 9.66417611e-01
-2.39513755e-01 -2.08892882e-01 -8.79783571e-01 8.04258347e-01
-7.57615222e-03 -1.26913583e+00 -9.02938284e-03 -1.06847703e-01
1.53298452e-01 2.88921565e-01 -3.32387537e-01 3.32464308e-01
8.50826621e-01 -9.47944880e-01 7.23724663e-01 3.68411541e-02
9.70634699e-01 -6.64874732e-01 5.36609530e-01 2.27763265e-01
-8.32900643e-01 -9.85246971e-02 -4.95863736e-01 4.76506203e-01
2.19276726e-01 3.57637674e-01 -5.45375049e-01 6.35980189e-01
1.16852391e+00 5.50266564e-01 -8.53070617e-01 9.41957254e-03
6.53589517e-02 4.19957668e-01 -4.32906479e-01 -1.13245264e-01
1.36492014e-01 -5.04092991e-01 1.52903378e-01 1.22179985e+00
1.76373780e-01 1.83484659e-01 1.90878332e-01 8.19452286e-01
-2.22038344e-01 3.81494582e-01 -1.00486791e+00 -4.11061585e-01
5.17981827e-01 8.53954017e-01 -5.43136060e-01 -4.13225383e-01
-8.50485802e-01 8.84848237e-01 4.28210407e-01 2.00654358e-01
-6.77146435e-01 -5.45372486e-01 8.50798428e-01 1.63426965e-01
1.62727848e-01 -5.99700101e-02 -5.79682231e-01 -1.08972657e+00
4.21124399e-01 -1.40875995e+00 1.25662816e+00 -7.06844687e-01
-1.35469842e+00 5.82272470e-01 1.59356743e-01 -4.75723594e-01
-2.99137354e-01 -7.29641676e-01 3.02548915e-01 1.10584462e+00
-1.47485387e+00 -1.25546944e+00 -1.28282487e-01 7.66435206e-01
2.12068677e-01 -5.66618085e-01 1.20753658e+00 4.81520325e-01
-2.55315691e-01 6.55147731e-01 1.78698197e-01 3.78135711e-01
1.05450773e+00 -1.22437215e+00 7.60768950e-01 4.47006524e-01
4.82228324e-02 8.37150693e-01 3.48271102e-01 -7.67037630e-01
-1.53909528e+00 -9.27571535e-01 1.52392733e+00 -9.04779077e-01
8.26918304e-01 -9.37353551e-01 -1.13219285e+00 9.46822286e-01
4.26774807e-02 -2.18402296e-01 8.56343150e-01 -6.67958558e-02
-7.96005785e-01 -2.26296782e-01 -1.54947627e+00 5.47648147e-02
1.04425788e+00 -1.02311969e+00 -7.63584435e-01 4.92909849e-01
9.27262604e-01 -2.53206193e-01 -1.11001933e+00 1.43154219e-01
3.67082536e-01 -9.49613690e-01 1.05713010e+00 -1.07199287e+00
5.17721951e-01 1.01510938e-02 -5.93386650e-01 -5.70718765e-01
3.78186345e-01 -5.25765240e-01 6.92106247e-01 1.65504241e+00
6.80538356e-01 -1.09664321e+00 1.15292656e+00 1.07579231e+00
3.35764617e-01 -2.83552378e-01 -9.34814394e-01 -6.96207404e-01
5.41901171e-01 -4.21175569e-01 9.04159188e-01 1.01662791e+00
7.10305050e-02 3.03178489e-01 3.53420004e-02 -7.53294975e-02
5.29245377e-01 4.31347549e-01 9.76501942e-01 -1.18070328e+00
-1.84680462e-01 4.81391698e-02 -3.74904692e-01 -7.03994095e-01
2.89683700e-01 -1.19807184e+00 -2.37747312e-01 -1.33302891e+00
1.10658869e-01 -7.16532469e-01 4.32070494e-01 6.62643611e-01
1.83863997e-01 -2.72430778e-01 -6.32504448e-02 1.75809562e-01
-4.66785908e-01 -1.05280288e-01 4.51785982e-01 1.13581792e-01
1.43943250e-01 -4.26406562e-01 -1.16820955e+00 5.14883399e-01
7.06390440e-01 -7.49114096e-01 -5.67172527e-01 -7.19920635e-01
2.64561683e-01 1.25262931e-01 2.00383198e-02 -6.18743658e-01
-4.82984185e-02 -1.07110649e-01 -1.76867649e-01 -6.53936744e-01
-1.23054206e-01 -8.31089497e-01 2.75035739e-01 2.90023535e-01
-5.41115522e-01 5.08275211e-01 2.32205138e-01 3.49749386e-01
-6.44309521e-01 -2.19147786e-01 7.40705371e-01 -6.97163999e-01
-1.03440607e+00 1.78906932e-01 1.10639565e-01 1.36789012e+00
8.30756605e-01 -1.64214913e-02 -5.30172765e-01 -2.34169424e-01
-3.10210884e-01 1.44433916e-01 6.80574417e-01 6.00171864e-01
3.89390849e-02 -7.55253673e-01 -6.17189586e-01 3.71352553e-01
5.89936018e-01 3.24801989e-02 -2.33573038e-02 2.16052160e-01
-9.42046046e-01 8.27881217e-01 -2.42915913e-01 -2.36032248e-01
-1.12051177e+00 6.00539386e-01 -2.07815871e-01 -3.28656942e-01
-4.90847439e-01 8.48083496e-01 -1.55752301e-02 -1.30726314e+00
1.86667442e-01 -1.11531749e-01 2.14606538e-01 7.06208870e-02
2.76835769e-01 2.24869296e-01 1.48715675e-01 -5.30740321e-01
-6.50001585e-01 4.65641052e-01 -1.03463763e-02 -1.99117720e-01
1.46127486e+00 -2.41942450e-01 -6.61687553e-01 2.78408885e-01
1.72866023e+00 1.84447542e-02 -4.46110159e-01 -2.55335003e-01
7.51850307e-01 -4.12955225e-01 -7.51004219e-01 -1.06055713e+00
-7.38578141e-01 8.78180504e-01 1.69304341e-01 1.17293201e-01
8.77158642e-01 4.10241127e-01 1.15335357e+00 5.49994230e-01
7.28304088e-01 -1.29981029e+00 -4.28925157e-01 7.09411442e-01
7.32694149e-01 -1.56578684e+00 -1.85540557e-01 -6.24116540e-01
-9.03694332e-01 8.15328419e-01 4.21181053e-01 4.49960828e-01
6.66768312e-01 3.31379414e-01 6.77371919e-01 -3.28597128e-01
-2.49953926e-01 1.77916571e-01 -3.39051187e-01 8.10233414e-01
4.34533358e-01 7.04459548e-02 -1.98558182e-01 8.28750074e-01
-7.69492328e-01 2.15408858e-02 1.60682693e-01 1.19051659e+00
5.94079904e-02 -1.63624942e+00 -1.77773014e-01 1.54495344e-01
-7.48037398e-01 -2.48079076e-01 -7.92799711e-01 1.20455575e+00
-3.78464460e-01 8.98688793e-01 -3.25701162e-02 2.23614544e-01
6.45183265e-01 4.47195381e-01 1.32623047e-01 -5.63066840e-01
-5.64471245e-01 -6.91895068e-01 5.77869296e-01 -8.38559330e-01
-8.46834481e-02 -8.19737673e-01 -1.32013535e+00 -4.56820041e-01
4.55877006e-01 3.08247685e-01 1.00659239e+00 4.41445172e-01
6.65766537e-01 -2.98769712e-01 5.57811856e-02 3.39285403e-01
-6.98860943e-01 -2.87844479e-01 -6.20355308e-01 9.69232857e-01
-3.45120072e-01 -6.21112436e-02 -3.98240574e-02 1.02148771e-01] | [9.90280532836914, 7.909300804138184] |
25cdd076-49c1-4f07-b006-99d277e6dd62 | transition-based-dependency-parsing-as-latent | null | null | https://aclanthology.org/W16-2403 | https://aclanthology.org/W16-2403.pdf | Transition-based dependency parsing as latent-variable constituent parsing | null | ['Mark-Jan Nederhof'] | 2016-08-01 | null | null | null | ws-2016-8 | ['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] | [-7.404524326324463, 3.792227268218994] |
7f5eebde-8bb5-4ae3-b6c8-23eb53f53e6f | gaze-prediction-in-dynamic-360a-immersive | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Xu_Gaze_Prediction_in_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Xu_Gaze_Prediction_in_CVPR_2018_paper.pdf | Gaze Prediction in Dynamic 360° Immersive Videos | This paper explores gaze prediction in dynamic $360^circ$ immersive videos, emph{i.e.}, based on the history scan path and VR contents, we predict where a viewer will look at an upcoming time. To tackle this problem, we first present the large-scale eye-tracking in dynamic VR scene dataset. Our dataset contains 208 $360^circ$ videos captured in dynamic scenes, and each video is viewed by at least 31 subjects. Our analysis shows that gaze prediction depends on its history scan path and image contents. In terms of the image contents, those salient objects easily attract viewers' attention. On the one hand, the saliency is related to both appearance and motion of the objects. Considering that the saliency measured at different scales is different, we propose to compute saliency maps at different spatial scales: the sub-image patch centered at current gaze point, the sub-image corresponding to the Field of View (FoV), and the panorama image. Then we feed both the saliency maps and the corresponding images into a Convolutional Neural Network (CNN) for feature extraction. Meanwhile, we also use a Long-Short-Term-Memory (LSTM) to encode the history scan path. Then we combine the CNN features and LSTM features for gaze displacement prediction between gaze point at a current time and gaze point at an upcoming time. Extensive experiments validate the effectiveness of our method for gaze prediction in dynamic VR scenes. | ['Yanyu Xu', 'Zhengzhong Sun', 'Shenghua Gao', 'Yanbing Dong', 'Jingyi Yu', 'Zhiru Shi', 'Junru Wu'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['eye-tracking'] | ['computer-vision'] | [ 2.15180501e-01 -2.56139129e-01 -5.12450263e-02 -3.81269991e-01
-2.07237005e-01 -1.83289051e-01 -8.46112072e-02 -1.50314584e-01
-1.98540956e-01 3.53676736e-01 2.11577147e-01 8.25625882e-02
-1.63805783e-01 -5.96113324e-01 -9.83110487e-01 -5.61129570e-01
-9.01321769e-02 -6.39131427e-01 5.28902829e-01 -2.68415660e-01
6.86855912e-01 9.94016752e-02 -2.02631450e+00 2.66781390e-01
6.63953185e-01 1.31450331e+00 9.18961287e-01 6.39292955e-01
1.87229693e-01 1.02350056e+00 -2.75265872e-01 8.09785724e-02
-1.94183692e-01 -4.29602951e-01 -6.40684605e-01 -1.25656696e-02
6.85351253e-01 -3.62588555e-01 -4.60693419e-01 1.10655165e+00
2.11223751e-01 4.44756329e-01 -2.90493132e-03 -1.18844819e+00
-8.78949821e-01 3.35585773e-01 -8.97017717e-01 1.01139224e+00
6.55581892e-01 1.78774074e-01 8.82227242e-01 -8.25913250e-01
9.08257306e-01 9.32151556e-01 9.79454368e-02 4.00099158e-01
-5.62476933e-01 -7.23756433e-01 7.47649133e-01 7.75452077e-01
-1.34544814e+00 -3.59975547e-01 1.01312923e+00 -4.11565721e-01
6.84757650e-01 3.66421700e-01 8.29005837e-01 8.93960118e-01
4.98056471e-01 8.38703930e-01 8.36234987e-01 4.37209904e-02
-3.74756777e-03 1.09125204e-01 -1.64784953e-01 6.87146425e-01
-5.22736132e-01 -5.44673391e-03 -8.29450130e-01 4.53487575e-01
8.76592338e-01 2.86637783e-01 -7.18942344e-01 -2.27455631e-01
-1.37903225e+00 5.54810286e-01 1.08625782e+00 3.80610913e-01
-2.34847605e-01 1.16156012e-01 7.44739100e-02 7.36283660e-02
3.98070604e-01 8.60352814e-02 -5.52034080e-01 -6.48635551e-02
-8.47382665e-01 -3.37917823e-03 -3.69889699e-02 8.30060244e-01
7.67001927e-01 6.16398379e-02 -1.31302759e-01 4.89385277e-01
5.24343610e-01 6.47442222e-01 5.15676618e-01 -7.81590104e-01
4.64726567e-01 4.90485877e-01 1.66121483e-01 -1.53580964e+00
-3.53848696e-01 -3.72776270e-01 -4.46908921e-01 7.22959191e-02
-7.01903552e-03 -1.42454192e-01 -6.75615489e-01 1.85301363e+00
3.51459563e-01 6.14839733e-01 -1.59590974e-01 1.48846400e+00
1.19675863e+00 7.43453383e-01 3.32500897e-02 -5.66789508e-01
1.44847369e+00 -9.00173903e-01 -7.71612704e-01 -3.55907530e-01
2.64361560e-01 -6.55129492e-01 1.09777629e+00 1.68187886e-01
-1.01764536e+00 -8.04809570e-01 -9.22918618e-01 -1.23779684e-01
-7.29335397e-02 5.99783175e-02 1.39463693e-01 9.07413587e-02
-1.24455082e+00 2.76353627e-01 -4.95495081e-01 -3.46322864e-01
1.19356535e-01 3.71965230e-01 -3.62236015e-02 2.58366495e-01
-1.19505441e+00 6.74492061e-01 -6.44939840e-02 3.91527086e-01
-6.84439600e-01 -4.81181860e-01 -1.07864201e+00 1.18218951e-01
5.56461990e-01 -4.24480408e-01 9.36975300e-01 -1.45683813e+00
-1.19618106e+00 6.52766407e-01 -8.39025617e-01 -6.92514181e-02
9.29423049e-02 -1.23506404e-01 -6.71559274e-01 1.56530574e-01
1.46620110e-01 7.48488188e-01 1.12831199e+00 -1.02851880e+00
-1.07300234e+00 -5.32553613e-01 2.61082292e-01 6.28964365e-01
-2.07146749e-01 2.74973333e-01 -6.07513726e-01 -2.74298817e-01
3.26817602e-01 -8.03030729e-01 9.43267941e-02 -1.52864039e-01
-2.72153258e-01 -2.26205990e-01 1.05397558e+00 -6.16112590e-01
1.41989362e+00 -2.21236277e+00 3.93359721e-01 -7.76513964e-02
5.05215526e-01 -1.65901214e-01 -1.97468456e-02 -2.71342248e-01
-3.01552057e-01 -7.30586126e-02 2.90506601e-01 -1.81597788e-02
-5.39465785e-01 -3.71475518e-01 -3.98493648e-01 4.28483367e-01
-2.35386449e-03 1.08810341e+00 -9.41274107e-01 -6.30317807e-01
3.99784684e-01 4.47949320e-01 -2.76771992e-01 1.18427470e-01
-1.37764946e-01 6.63529873e-01 -5.83876967e-01 5.27394831e-01
5.91345251e-01 -5.27285755e-01 -2.83782065e-01 -1.03726275e-01
-4.89548236e-01 -5.89250699e-02 -9.12681639e-01 1.92287457e+00
-3.10835987e-01 1.19565666e+00 -3.95297468e-01 -4.05128241e-01
7.60127723e-01 2.51268782e-03 3.09899211e-01 -1.21909249e+00
4.28148180e-01 -3.18487585e-02 -7.72769004e-02 -9.10186946e-01
8.24120343e-01 3.19057167e-01 1.61530390e-01 4.43630517e-01
-1.39397025e-01 5.96273720e-01 -3.20095941e-02 -1.41841928e-02
4.96088207e-01 1.12936758e-01 -9.80502740e-03 2.94451714e-02
7.12876737e-01 -4.10581410e-01 4.77757871e-01 1.58860445e-01
-3.97224933e-01 8.43000948e-01 5.20742953e-01 -3.63622487e-01
-5.29210389e-01 -8.19472432e-01 7.19260424e-02 1.59979022e+00
1.06306422e+00 -1.02299511e-01 -6.71445668e-01 -4.60562527e-01
-4.75260317e-01 5.49983203e-01 -1.10725117e+00 -1.86894730e-01
-7.20590591e-01 -7.32627288e-02 -4.24012303e-01 3.27639222e-01
3.36483717e-01 -1.63681650e+00 -1.17289019e+00 -2.66912222e-01
-4.04527664e-01 -8.68342698e-01 -9.04497623e-01 -5.34504592e-01
-6.26972973e-01 -1.07834542e+00 -7.41827786e-01 -1.06345975e+00
4.93834883e-01 7.52760470e-01 8.44131470e-01 3.61265652e-02
1.51733443e-01 7.25049302e-02 -3.09449404e-01 -7.39437938e-02
5.40113568e-01 -7.02169240e-02 -6.73512518e-02 2.88215548e-01
3.63793522e-01 -3.00446659e-01 -1.06149435e+00 3.17559302e-01
-4.57939297e-01 2.09449515e-01 2.35294521e-01 3.02906126e-01
7.64515042e-01 -1.60058662e-01 1.44875338e-02 -3.66449773e-01
3.46165210e-01 -6.68451190e-01 -5.30654967e-01 3.13135147e-01
-1.93887338e-01 -4.47087198e-01 1.78703934e-01 -3.82699966e-01
-9.52044129e-01 -1.78831995e-01 6.37914836e-02 -8.98575783e-01
4.85410877e-02 5.08580863e-01 -6.43009618e-02 2.50118822e-02
3.26861858e-01 4.38469768e-01 -2.20209345e-01 -7.06920400e-03
1.07167870e-01 5.74901521e-01 3.11551332e-01 3.84570539e-01
4.04869288e-01 4.36672747e-01 -3.06084126e-01 -7.60132730e-01
-9.64047492e-01 -1.52085707e-01 -5.08723557e-01 -7.33267426e-01
1.21347535e+00 -9.55861747e-01 -1.21971214e+00 3.44915122e-01
-9.47134614e-01 -1.04170725e-01 4.69108634e-02 5.55444956e-01
-4.57753539e-01 -2.51385719e-02 -2.35723436e-01 -6.61348820e-01
-3.12037110e-01 -1.47025013e+00 1.12980294e+00 9.49682951e-01
1.25832587e-01 -8.75595093e-01 -2.10876733e-01 1.40699938e-01
5.60522527e-02 1.70555785e-01 4.52906549e-01 5.58358394e-02
-1.00061500e+00 1.88665375e-01 -4.87424016e-01 -2.41018564e-01
-3.45781185e-02 2.19326779e-01 -9.56982255e-01 -2.59970605e-01
1.84661105e-01 1.38500586e-01 6.16004825e-01 1.07763529e+00
1.26609957e+00 -1.03701442e-01 -5.08435965e-01 8.41105342e-01
1.46919680e+00 4.16368276e-01 6.79554820e-01 2.88480222e-01
1.00349975e+00 6.63694084e-01 1.14757931e+00 2.58359849e-01
7.26858974e-01 7.26426601e-01 8.60794306e-01 5.98087646e-02
2.25030079e-01 -2.18880355e-01 3.24219465e-01 5.96730888e-01
-2.20465496e-01 -2.44161546e-01 -6.84238791e-01 5.43432713e-01
-1.73527765e+00 -1.19719791e+00 -2.60749459e-01 2.19649076e+00
3.05934221e-01 -1.94588993e-02 -3.39120743e-03 -3.60952199e-01
8.67033362e-01 4.54296768e-01 -9.22727823e-01 -2.89789110e-01
2.73219161e-02 -1.59278393e-01 1.86383933e-01 3.07855994e-01
-9.08121407e-01 9.20738399e-01 5.27979279e+00 5.30045509e-01
-1.77519691e+00 9.53345299e-02 7.53972173e-01 -6.44945145e-01
-3.61776382e-01 -2.01842010e-01 -7.61151612e-01 8.41710925e-01
6.85604751e-01 -1.48918703e-01 4.63012993e-01 8.66898835e-01
3.66381228e-01 -4.10282969e-01 -7.85962224e-01 1.23299015e+00
6.43676758e-01 -1.25548983e+00 -3.06985855e-01 -5.97123727e-02
6.15923166e-01 1.79871947e-01 8.15232992e-01 9.78190526e-02
-4.34191585e-01 -1.09272671e+00 9.71692562e-01 8.29027474e-01
9.07406330e-01 -8.20436239e-01 3.85401219e-01 2.39077300e-01
-1.52087116e+00 -4.96215403e-01 -3.41375083e-01 -2.84214597e-02
2.67956674e-01 -6.13647588e-02 -1.52861714e-01 2.04133272e-01
1.33080053e+00 1.17748499e+00 -8.63719046e-01 8.52301478e-01
-1.19950041e-01 -1.40652895e-01 1.59009412e-01 -2.11334676e-01
7.93406069e-02 -1.72352999e-01 5.80468059e-01 4.17737126e-01
6.16802156e-01 3.54424179e-01 -2.72665709e-01 9.14993823e-01
1.38218865e-01 -2.55578943e-02 -4.99208838e-01 4.02840823e-01
3.51479679e-01 1.08352315e+00 -6.30037904e-01 -9.69370008e-02
-3.82896006e-01 9.75950897e-01 2.31916979e-01 6.62117779e-01
-1.14307165e+00 -3.89512271e-01 6.48088217e-01 2.55461991e-01
7.03793705e-01 9.68670994e-02 -1.30300999e-01 -1.28850830e+00
1.92737430e-01 -3.70402694e-01 2.08413586e-01 -1.60551298e+00
-6.27491057e-01 1.01545358e+00 -3.98267537e-01 -1.60744631e+00
-1.52838826e-01 -2.46268953e-03 -7.01964259e-01 9.50367928e-01
-1.64896166e+00 -9.48651969e-01 -6.97232485e-01 8.34621429e-01
7.97827780e-01 1.63245663e-01 2.93093801e-01 -7.20148608e-02
-7.13015318e-01 4.38495338e-01 -1.28608987e-01 -7.83644989e-02
4.88559633e-01 -7.70601332e-01 1.33116037e-01 8.39953363e-01
1.21615209e-01 6.44062221e-01 6.72868490e-01 -5.78225374e-01
-1.11917341e+00 -9.03502584e-01 8.80855441e-01 -6.40656233e-01
5.37921607e-01 -1.72476500e-01 -9.35654521e-01 7.74796128e-01
3.06165129e-01 2.03532517e-01 4.14433539e-01 1.07262522e-01
1.22248873e-01 -1.03868186e-01 -7.66094804e-01 6.22436941e-01
1.14653695e+00 -6.89167678e-01 -4.97906476e-01 -7.87017345e-02
1.05510068e+00 -9.18030620e-01 -5.64736485e-01 4.12205964e-01
6.39446020e-01 -1.33737946e+00 9.55306053e-01 -2.51158655e-01
7.61752307e-01 -3.86679411e-01 6.47075772e-02 -1.01145744e+00
-3.10495079e-01 -2.23950803e-01 -1.77671418e-01 7.83298612e-01
1.19657062e-01 -2.15799868e-01 6.33109272e-01 4.80591327e-01
6.90839905e-03 -9.77507532e-01 -8.36635172e-01 1.30328372e-01
-5.41236937e-01 -3.05357963e-01 7.17322648e-01 7.66785979e-01
8.03058967e-03 3.98560077e-01 -4.28477734e-01 2.86139756e-01
2.23204687e-01 6.00976288e-01 5.10639310e-01 -1.08800340e+00
2.07431838e-01 -2.94993788e-01 -3.83279145e-01 -1.32361603e+00
3.36341001e-02 -2.88077116e-01 9.29057747e-02 -1.26969874e+00
1.96134686e-01 -8.75029787e-02 -6.90107524e-01 -7.40227476e-02
-5.62586069e-01 3.59978229e-01 3.11174452e-01 2.27038577e-01
-9.85556066e-01 5.25628924e-01 1.70389807e+00 1.46454915e-01
-6.12993717e-01 6.54774718e-03 -6.48721814e-01 6.56439424e-01
4.79802340e-01 -6.56007975e-02 -6.30305052e-01 -6.99660718e-01
4.55777913e-01 5.69308281e-01 4.77437913e-01 -9.51513469e-01
3.68925124e-01 -3.65101576e-01 5.71426570e-01 -8.69522095e-01
4.76223499e-01 -6.48686826e-01 -1.10342100e-01 2.55026162e-01
-2.97295779e-01 1.58376262e-01 7.47138485e-02 7.85100222e-01
-1.86535642e-01 2.38447338e-01 5.69304168e-01 1.45166233e-01
-1.16803372e+00 6.82911694e-01 -1.14360340e-01 -1.90714702e-01
1.25688219e+00 -5.72926342e-01 -3.09713751e-01 -4.14819926e-01
-8.37174416e-01 4.19610083e-01 5.92925608e-01 9.28473532e-01
1.02227640e+00 -1.13053262e+00 -2.16592118e-01 5.17609179e-01
1.85850397e-01 3.73209789e-02 1.01012981e+00 1.21598458e+00
-2.43269786e-01 3.63486618e-01 -5.13575017e-01 -9.85362887e-01
-1.58844948e+00 8.10904562e-01 3.24104369e-01 4.00845587e-01
-4.93733972e-01 1.30593157e+00 6.10856056e-01 3.14012945e-01
-4.97543067e-02 -4.95779842e-01 -1.22250009e+00 1.73075810e-01
7.78934419e-01 -3.06211300e-02 -3.75662297e-01 -1.36970317e+00
-2.91418314e-01 1.04840541e+00 -5.06014451e-02 1.76097348e-01
1.17754447e+00 -8.84678245e-01 -5.20827062e-02 7.88388252e-01
1.23738706e+00 -1.96333099e-02 -1.49942946e+00 -2.87751138e-01
-5.13981938e-01 -9.86596763e-01 7.77125880e-02 -3.28875393e-01
-1.56877005e+00 9.65716541e-01 1.01500165e+00 9.39143300e-02
1.37165701e+00 1.49925485e-01 8.67114544e-01 -1.17432393e-01
4.16534007e-01 -7.89275289e-01 1.93967968e-01 4.58106309e-01
6.86566591e-01 -1.39564049e+00 -2.27808252e-01 -2.83319533e-01
-1.07068622e+00 7.31189311e-01 1.07588911e+00 -1.82112381e-01
6.89781189e-01 -6.39821053e-01 1.22442670e-01 -4.95852619e-01
-6.40309274e-01 -2.77641028e-01 5.21074355e-01 4.60985363e-01
1.45771623e-01 -1.97918743e-01 1.77890763e-01 4.58961636e-01
-4.06218380e-01 -8.37808102e-02 5.56461334e-01 5.40760696e-01
-5.45795679e-01 -1.97588786e-01 -2.02097684e-01 4.87131238e-01
-2.31140092e-01 -1.08864963e-01 1.21285364e-01 3.74698371e-01
3.65866691e-01 1.03366792e+00 5.19962728e-01 -7.66261220e-01
1.02595232e-01 -4.73852158e-01 2.58654535e-01 -5.18047094e-01
-3.66416454e-01 5.52927516e-02 -5.93333542e-01 -9.38165963e-01
-5.78057766e-01 -7.72747040e-01 -1.38980317e+00 -3.75018209e-01
-3.85389447e-01 -6.33625612e-02 4.99077559e-01 9.39346433e-01
4.66516972e-01 8.15707743e-01 8.46680880e-01 -9.07814682e-01
4.28834409e-01 -8.33234310e-01 -7.30590403e-01 2.14607581e-01
6.19418323e-01 -8.46584737e-01 -1.55314490e-01 8.44851285e-02] | [9.821004867553711, -0.2572791278362274] |
a884b93b-7fc2-4dd9-8580-0d06bf4c1fa7 | palgan-image-colorization-with-palette | 2210.11204 | null | https://arxiv.org/abs/2210.11204v1 | https://arxiv.org/pdf/2210.11204v1.pdf | PalGAN: Image Colorization with Palette Generative Adversarial Networks | Multimodal ambiguity and color bleeding remain challenging in colorization. To tackle these problems, we propose a new GAN-based colorization approach PalGAN, integrated with palette estimation and chromatic attention. To circumvent the multimodality issue, we present a new colorization formulation that estimates a probabilistic palette from the input gray image first, then conducts color assignment conditioned on the palette through a generative model. Further, we handle color bleeding with chromatic attention. It studies color affinities by considering both semantic and intensity correlation. In extensive experiments, PalGAN outperforms state-of-the-arts in quantitative evaluation and visual comparison, delivering notable diverse, contrastive, and edge-preserving appearances. With the palette design, our method enables color transfer between images even with irrelevant contexts. | ['Yu Qiao', 'Jing Shao', 'Lu Qi', 'Menghan Xia', 'Yi Wang'] | 2022-10-20 | null | null | null | null | ['colorization'] | ['computer-vision'] | [ 2.18209669e-01 -2.86000162e-01 8.39973390e-02 -1.77562222e-01
-6.71811342e-01 -7.59668112e-01 4.16205913e-01 -3.67949992e-01
-2.33515739e-01 7.20463932e-01 3.32977995e-02 -7.15800282e-03
1.15481846e-01 -6.45843387e-01 -6.40391052e-01 -8.17185521e-01
4.41436172e-01 2.89371431e-01 -9.31302905e-02 -2.47608617e-01
3.33102465e-01 4.14291173e-01 -1.13251543e+00 9.47225764e-02
1.24435318e+00 8.62822592e-01 -1.82833821e-02 5.63813210e-01
-2.35266745e-01 3.30490500e-01 -5.55600524e-01 -5.74181795e-01
3.64337951e-01 -6.64578319e-01 -4.49678361e-01 3.05691481e-01
6.90877318e-01 -2.37639308e-01 6.73853457e-02 1.19497967e+00
3.36538285e-01 1.14230469e-01 7.54561186e-01 -1.43108988e+00
-1.28444278e+00 3.14911216e-01 -1.26345181e+00 -3.31972688e-01
2.02143371e-01 1.95801690e-01 1.01258397e+00 -9.18190122e-01
4.95123297e-01 1.40802479e+00 3.21215540e-01 5.63753724e-01
-1.63573444e+00 -7.45874524e-01 3.58956426e-01 1.79503769e-01
-1.34435725e+00 5.38080223e-02 1.18157136e+00 -2.07482040e-01
1.52325612e-02 4.15675700e-01 7.10556507e-01 9.45683181e-01
2.54464298e-02 9.04233694e-01 1.61639273e+00 -5.18714964e-01
2.77226716e-01 -1.40122578e-01 -3.54732573e-01 8.73690009e-01
2.31849432e-01 2.48398874e-02 -4.00198907e-01 -2.61711590e-02
8.50645781e-01 5.52015146e-03 -3.33262950e-01 -5.65648735e-01
-1.28803933e+00 4.56403732e-01 5.33142090e-01 1.42302224e-02
-3.36319596e-01 4.47698027e-01 -1.55446917e-01 -1.47638887e-01
3.75902325e-01 5.20729899e-01 -1.34069487e-01 -3.20838508e-03
-9.14728522e-01 -1.35368854e-01 4.69596535e-01 1.06781089e+00
1.07064581e+00 1.94644883e-01 -5.61039150e-01 8.02223802e-01
3.53317380e-01 9.09768283e-01 -9.00344700e-02 -1.03766668e+00
3.02525699e-01 5.78799546e-01 7.90350288e-02 -1.06344795e+00
-2.32898250e-01 -2.34970853e-01 -1.07079947e+00 6.12654805e-01
5.18816829e-01 -2.30367437e-01 -1.37983239e+00 1.87995410e+00
3.12889397e-01 6.79286495e-02 -1.36568740e-01 1.12307119e+00
3.34362715e-01 5.70199668e-01 3.47937077e-01 4.55073938e-02
1.57930589e+00 -1.07826591e+00 -7.73050010e-01 -2.31856957e-01
-2.69839883e-01 -1.00764227e+00 1.32615161e+00 4.08259153e-01
-1.08228219e+00 -4.31203604e-01 -1.03432703e+00 -1.37123421e-01
-1.91983879e-01 3.09427142e-01 5.98572969e-01 9.18976903e-01
-1.18103302e+00 2.35188648e-01 -7.43273258e-01 -1.69348761e-01
3.80146384e-01 3.33800018e-02 -1.68397605e-01 -3.21051657e-01
-8.86662364e-01 6.07885182e-01 2.28621289e-01 2.70455569e-01
-7.30145574e-01 -6.94838762e-01 -7.80692697e-01 1.28181934e-01
2.57649034e-01 -9.74500716e-01 6.73196018e-01 -1.28556931e+00
-1.96730697e+00 5.96045136e-01 -5.24029024e-02 2.60142416e-01
7.14096725e-01 -1.82309553e-01 -3.65474761e-01 1.90191209e-01
-6.95977062e-02 8.70918810e-01 1.16849339e+00 -1.92469549e+00
-5.75957954e-01 -8.46367329e-02 -6.04000352e-02 2.91771740e-01
-2.81772882e-01 -2.54627317e-01 -1.17943907e+00 -7.98130393e-01
1.17151149e-01 -9.84222710e-01 -1.56436145e-01 1.56994984e-01
-8.26102793e-01 2.62840152e-01 7.64790177e-01 -7.12899387e-01
1.04549611e+00 -2.20735788e+00 4.67244834e-01 5.95061660e-01
3.59905273e-01 -2.68511735e-02 -4.27478582e-01 2.58816570e-01
1.14673346e-01 1.40707031e-01 -4.98984575e-01 -5.30388892e-01
2.77178228e-01 2.79441644e-02 -1.53459609e-01 4.16215003e-01
3.77911806e-01 9.33747888e-01 -8.59808147e-01 -6.16986036e-01
3.16389740e-01 8.38319838e-01 -5.84357798e-01 2.64749557e-01
-2.49213099e-01 5.68786502e-01 -4.31686789e-01 9.63926256e-01
1.31273603e+00 -2.60223616e-02 2.68684655e-01 -6.76787734e-01
6.50937781e-02 -8.81252050e-01 -1.08151591e+00 1.94167757e+00
-4.71640319e-01 5.82001567e-01 1.06299527e-01 -3.15528929e-01
8.16349506e-01 -2.14656875e-01 2.69510239e-01 -6.81952000e-01
1.99465171e-01 -3.95311452e-02 -3.02322567e-01 -1.91247940e-01
7.25753367e-01 -1.07199892e-01 -6.47421703e-02 4.22860324e-01
-2.83160746e-01 -3.09950143e-01 1.97005033e-01 2.45169520e-01
5.80550075e-01 4.73741472e-01 -2.79874921e-01 -1.77105099e-01
4.70757693e-01 -3.57733995e-01 5.66470861e-01 4.85596031e-01
-3.40371206e-02 1.05334902e+00 7.64701486e-01 -1.36650041e-01
-1.03754783e+00 -1.33300686e+00 2.04074621e-01 1.26206124e+00
6.43365383e-01 -1.56753704e-01 -9.70006883e-01 -7.47579694e-01
3.99742927e-03 7.54826844e-01 -9.35417712e-01 -6.49819598e-02
-5.46998203e-01 -7.25111187e-01 2.16661975e-01 4.66363937e-01
6.31932080e-01 -7.39912570e-01 -2.54691958e-01 -2.08299845e-01
-2.38977075e-01 -8.76843870e-01 -1.09098005e+00 -2.79267430e-01
-3.99157435e-01 -1.00794530e+00 -1.13390839e+00 -6.53580010e-01
9.98732626e-01 1.75590023e-01 1.18698525e+00 -1.41873151e-01
-3.75499099e-01 6.09416604e-01 -4.07193124e-01 2.68476997e-02
-1.60484940e-01 -2.66335662e-02 -4.72538263e-01 5.29226243e-01
-2.18716711e-01 -5.13026118e-01 -1.12867188e+00 3.38541567e-01
-1.20450234e+00 4.12636399e-01 9.67797518e-01 8.93150210e-01
6.80776536e-01 -3.05892080e-01 1.05700344e-01 -9.90071118e-01
5.29721022e-01 -1.43411174e-01 -7.17201352e-01 7.36062825e-01
-6.88822210e-01 4.03255582e-01 5.40175140e-01 -3.94518524e-01
-1.40516424e+00 4.80663031e-02 2.79954612e-01 -5.30332208e-01
1.36666512e-02 6.11992553e-02 -4.18465793e-01 -3.09131593e-01
2.96761185e-01 1.09634288e-01 -9.33633447e-02 -3.84491563e-01
1.10685027e+00 1.60795897e-01 6.22488737e-01 -8.67348254e-01
1.19558656e+00 7.79119909e-01 1.71745277e-03 -4.05833066e-01
-4.91752535e-01 6.15862273e-02 -5.30193686e-01 -3.09951067e-01
1.11726558e+00 -7.16415107e-01 -7.95662344e-01 4.76902366e-01
-1.04593015e+00 -3.05287153e-01 -1.24731004e-01 2.28200853e-01
-3.02387804e-01 4.39163804e-01 -4.95801896e-01 -7.07595348e-01
-2.68196702e-01 -1.02708423e+00 1.32108772e+00 5.16966820e-01
2.15379685e-01 -8.55272055e-01 1.54505834e-01 1.76322833e-01
5.86934507e-01 5.89900792e-01 9.58770752e-01 3.87536883e-01
-8.57769012e-01 -1.39723737e-02 -8.27543437e-01 2.54757524e-01
1.90350205e-01 3.06627810e-01 -8.70542169e-01 -2.80384332e-01
-6.93820179e-01 1.67286955e-02 1.08296847e+00 3.12845707e-01
1.10839629e+00 -9.21615884e-02 -1.06849489e-04 1.02044094e+00
1.64068210e+00 1.55303150e-01 7.66835690e-01 2.75724560e-01
1.14532602e+00 4.26161468e-01 3.59841675e-01 4.95102018e-01
3.83099854e-01 4.89426821e-01 5.63097596e-01 -8.23063850e-01
-4.32011604e-01 -3.33668798e-01 9.87348184e-02 4.29924607e-01
-1.34602129e-01 -4.17256385e-01 -5.56129873e-01 2.05093697e-01
-1.66635561e+00 -5.39272487e-01 2.59187520e-01 1.87034059e+00
7.84703791e-01 -2.59166479e-01 -7.25116581e-02 -2.99906433e-01
9.94677901e-01 1.91368788e-01 -6.14421606e-01 -1.07730597e-01
-4.47307408e-01 1.87103301e-01 5.95796406e-01 4.76099402e-01
-9.04436171e-01 1.08613229e+00 6.25877523e+00 9.37153220e-01
-1.19759762e+00 3.83814122e-03 1.03196895e+00 1.68204919e-01
-8.61513972e-01 6.56413361e-02 -1.52927548e-01 5.16590178e-01
-1.02996610e-01 2.67575890e-01 7.55129755e-01 4.43134099e-01
-7.56918192e-02 -2.13218689e-01 -7.87403107e-01 1.15167797e+00
3.32567751e-01 -1.03485930e+00 2.60452509e-01 -9.69963446e-02
1.03767514e+00 -5.44859529e-01 6.99062467e-01 -3.74587551e-02
5.19656241e-01 -8.44356835e-01 1.04372275e+00 7.92183042e-01
1.19433010e+00 -8.68798792e-01 2.80700773e-01 -6.04721546e-01
-1.23262906e+00 1.24562345e-01 -5.43417372e-02 4.92248833e-01
2.38154858e-01 3.69038582e-01 -2.24987596e-01 7.79038012e-01
4.41764772e-01 4.92720634e-01 -8.52758825e-01 9.73925531e-01
-6.75637603e-01 3.42749923e-01 -4.63832170e-02 9.00556669e-02
9.88958254e-02 -7.17897296e-01 3.13117772e-01 1.22786331e+00
3.57501864e-01 -3.80764268e-02 1.26920879e-01 1.30805898e+00
-1.05366543e-01 -7.57674221e-03 2.16852054e-01 1.02445431e-01
3.83525699e-01 1.68060386e+00 -9.79181767e-01 -2.11151779e-01
-8.19124728e-02 1.56236613e+00 1.44711241e-01 1.00275612e+00
-1.09276521e+00 -5.46310842e-01 5.51947057e-01 -2.17938229e-01
2.50783652e-01 -2.58696407e-01 -4.42484289e-01 -1.20911372e+00
-6.40615374e-02 -6.75010324e-01 2.44022861e-01 -1.04382944e+00
-1.40712321e+00 7.22936690e-01 -3.50138433e-02 -1.25084126e+00
2.26713657e-01 -5.62403023e-01 -7.97198832e-01 7.23126173e-01
-1.61700618e+00 -1.64412999e+00 -6.00327075e-01 4.92281914e-01
8.99465978e-02 1.03429757e-01 4.24608558e-01 3.50744694e-01
-7.94322789e-01 6.83161259e-01 3.57793510e-01 8.70985724e-03
1.10227239e+00 -1.54028702e+00 2.34360591e-01 9.94129539e-01
4.61337827e-02 3.62175852e-01 6.18965507e-01 -3.12364161e-01
-1.72296023e+00 -1.03030360e+00 -3.41110453e-02 -1.65596083e-01
3.64884108e-01 -3.09455425e-01 -5.96459627e-01 2.32859805e-01
6.28310204e-01 -2.26057068e-01 4.71430004e-01 -8.41980129e-02
-6.48935914e-01 -3.65161479e-01 -1.02812147e+00 7.74258256e-01
7.59711862e-01 -3.78688872e-01 -5.73286973e-02 -1.06531627e-01
5.36233366e-01 -4.05774444e-01 -4.40254658e-01 9.48512927e-02
6.73648655e-01 -8.35712612e-01 9.39145327e-01 -1.88098595e-01
2.91912884e-01 -6.70340061e-01 1.88251249e-02 -1.42471504e+00
-4.51495677e-01 -7.36800969e-01 4.10371035e-01 1.60009706e+00
4.06917691e-01 -4.36328530e-01 6.80914819e-01 7.61037886e-01
-9.30105895e-02 -3.11726362e-01 -6.97413445e-01 -5.20504236e-01
-2.03489251e-02 -1.28244922e-01 6.90542042e-01 8.55423450e-01
-3.46329272e-01 8.53376985e-02 -7.80445516e-01 6.36737496e-02
9.90011632e-01 5.40912688e-01 7.52542973e-01 -7.20811963e-01
-2.85116076e-01 -7.08194971e-01 -1.49931774e-01 -6.27344310e-01
7.54766315e-02 -4.73014832e-01 1.18290022e-01 -1.46140420e+00
3.94076794e-01 -3.62084687e-01 -5.01286983e-01 5.25242090e-01
-4.79575187e-01 9.06701744e-01 5.05831659e-01 -4.71682623e-02
-8.37680578e-01 8.19078565e-01 1.62289190e+00 -4.55036610e-01
-1.39288381e-01 -5.54768205e-01 -9.40703094e-01 4.60121065e-01
6.45978391e-01 -1.52415827e-01 -2.04597741e-01 -6.36860490e-01
3.02192628e-01 -1.74832210e-01 2.94258893e-01 -8.84452462e-01
1.62431430e-02 -3.88534933e-01 7.71243930e-01 -3.22127968e-01
3.42807829e-01 -8.36191356e-01 2.87251174e-01 2.96573699e-01
-1.77894920e-01 1.24271162e-01 2.59121656e-01 8.69231999e-01
-1.59144595e-01 3.44210654e-01 9.73241806e-01 2.58773506e-01
-7.88219750e-01 3.53796244e-01 -1.27399817e-01 1.51635915e-01
9.98673916e-01 -1.48562044e-01 -4.24472600e-01 -3.54459673e-01
-6.19490623e-01 1.60966933e-01 7.76120126e-01 2.85235167e-01
6.16830945e-01 -1.62528563e+00 -8.35153937e-01 3.63620192e-01
2.24480942e-01 -3.04263651e-01 6.11640215e-01 7.65537918e-01
-7.53194749e-01 -2.84891486e-01 -3.72204274e-01 -5.57793319e-01
-8.99154186e-01 6.12866163e-01 1.17053144e-01 6.32344112e-02
-2.99819887e-01 8.02060127e-01 7.15156436e-01 5.18233366e-02
-3.20042670e-02 -3.08126390e-01 9.74111166e-03 -6.92920536e-02
1.84976831e-01 3.31666291e-01 -3.09794337e-01 -5.46130896e-01
-2.61345983e-01 1.04676092e+00 6.42595962e-02 -4.66704696e-01
9.42739308e-01 -4.62932795e-01 -2.17132375e-01 6.27798513e-02
1.06906581e+00 3.69761467e-01 -1.70262682e+00 -7.20812157e-02
-3.56309026e-01 -9.05429423e-01 -8.81926641e-02 -1.06033838e+00
-1.57345343e+00 6.87547088e-01 9.25650358e-01 -5.98520972e-02
1.51244128e+00 -2.04424798e-01 7.40294933e-01 -1.95593446e-01
-5.65167926e-02 -1.06550336e+00 3.46520096e-01 1.31718606e-01
7.98390031e-01 -1.23587966e+00 -6.91081211e-02 -4.13470656e-01
-9.39695418e-01 9.98878896e-01 6.84482396e-01 -1.02762781e-01
3.22240949e-01 7.97973573e-02 3.75783652e-01 1.55829620e-02
-2.40244940e-01 -2.09348843e-01 7.62804925e-01 6.48620486e-01
2.84445494e-01 2.33238339e-01 -1.96122572e-01 2.06251115e-01
2.69566149e-01 -4.98233080e-01 3.26079100e-01 5.34010708e-01
-1.23305231e-01 -1.26451409e+00 -5.74128270e-01 -2.19253972e-01
-2.11958021e-01 -4.05889332e-01 -3.44251633e-01 6.79244578e-01
1.75356969e-01 7.47126639e-01 7.01531693e-02 -4.22369897e-01
2.89455831e-01 -2.81132996e-01 5.38536966e-01 -1.06422462e-01
-2.84974337e-01 5.60164511e-01 -3.05857152e-01 -6.99090958e-01
-2.85904199e-01 -2.14763716e-01 -9.29299533e-01 -3.77736032e-01
-2.37231359e-01 9.23008025e-02 6.94429278e-01 4.75066364e-01
4.54625040e-01 6.88249648e-01 7.39316821e-01 -9.99816895e-01
-1.06917126e-02 -7.35706985e-01 -8.20875823e-01 7.41916060e-01
2.67111719e-01 -5.17091095e-01 -3.76486301e-01 1.06407456e-01] | [11.413628578186035, -1.0266081094741821] |
444f66c2-a629-4966-96f8-bcaafab08d34 | overview-of-the-first-shared-task-on-multi | null | null | https://aclanthology.org/2022.sdp-1.32 | https://aclanthology.org/2022.sdp-1.32.pdf | Overview of the First Shared Task on Multi Perspective Scientific Document Summarization (MuP) | We present the main findings of MuP 2022 shared task, the first shared task on multi-perspective scientific document summarization. The task provides a testbed representing challenges for summarization of scientific documents, and facilitates development of better models to leverage summaries generated from multiple perspectives. We received 139 total submissions from 9 teams. We evaluated submissions both by automated metrics (i.e., Rouge) and human judgments on faithfulness, coverage, and readability which provided a more nuanced view of the differences between the systems. While we observe encouraging results from the participating teams, we conclude that there is still significant room left for improving summarization leveraging multiple references. Our dataset is available at https://github.com/allenai/mup. | ['Michal Shmueli-Scheuer', 'Tirthankar Ghosal', 'Guy Feigenblat', 'Arman Cohan'] | null | null | null | null | sdp-coling-2022-10 | ['scientific-article-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [-8.34826031e-04 3.20890963e-01 -4.61613715e-01 -2.23635182e-01
-1.50485837e+00 -1.00713825e+00 8.55729103e-01 8.42524469e-01
-1.87851116e-01 1.01154029e+00 1.19486189e+00 -9.48655307e-02
-1.55427847e-02 -2.73733079e-01 -4.30093825e-01 -1.57613069e-01
1.70579359e-01 3.93126756e-01 -1.45024896e-01 -1.12575687e-01
1.10937035e+00 1.76455975e-01 -1.00502908e+00 7.18173563e-01
1.28061926e+00 2.00465366e-01 4.08854149e-02 1.23407853e+00
3.57971415e-02 9.12178516e-01 -1.14584529e+00 -4.32707399e-01
-2.20713004e-01 -4.37547624e-01 -1.17211688e+00 -4.28286582e-01
8.68903041e-01 2.88723148e-02 -1.91304356e-01 6.49724960e-01
7.80928493e-01 1.08553395e-01 7.11734533e-01 -1.09674203e+00
-9.18600261e-01 1.03744435e+00 -6.26377821e-01 7.46381521e-01
7.21604645e-01 2.50274301e-01 1.32732320e+00 -7.27921903e-01
1.11509418e+00 1.20252872e+00 5.72898865e-01 5.43727100e-01
-1.13246465e+00 -4.52168524e-01 -1.29786462e-01 1.34076104e-01
-8.17583501e-01 -1.04536414e+00 1.91633463e-01 -5.02938628e-01
1.06056118e+00 7.17886508e-01 3.94200861e-01 1.14931071e+00
3.24017763e-01 7.28898108e-01 8.39952409e-01 -1.62426218e-01
1.87798776e-02 4.63440344e-02 6.19000614e-01 2.44667247e-01
7.01106668e-01 -7.30203211e-01 -1.00856400e+00 -5.37163377e-01
9.43225063e-03 -4.42174971e-01 -6.01004779e-01 4.80796725e-01
-1.57177711e+00 4.83145416e-01 3.48680536e-03 3.66048902e-01
-2.58030176e-01 -6.10940196e-02 8.05322409e-01 1.01742491e-01
7.98117936e-01 1.30543184e+00 -2.78085142e-01 -6.19470716e-01
-1.40674174e+00 6.99892223e-01 1.15371585e+00 1.04864371e+00
1.76345706e-01 -2.90579498e-01 -9.67467785e-01 7.84500599e-01
-1.33637831e-01 5.14307022e-01 1.47265404e-01 -1.44501877e+00
6.14133418e-01 4.16989386e-01 2.23957449e-01 -1.05261552e+00
-3.93641204e-01 -6.34835184e-01 -7.06906080e-01 -4.20416325e-01
1.69633776e-01 -2.58546591e-01 -4.73692983e-01 1.49543250e+00
-2.15998709e-01 -3.25616062e-01 6.07786588e-02 3.46037507e-01
1.58766365e+00 8.36560428e-01 -1.30683616e-01 -3.77734572e-01
1.35849929e+00 -1.11908603e+00 -1.03011644e+00 -1.11501701e-02
6.88786626e-01 -1.19239712e+00 9.94799674e-01 5.02322733e-01
-1.61816871e+00 -2.72959828e-01 -1.13243032e+00 -4.98736709e-01
-1.32628545e-01 2.28987783e-01 1.90584719e-01 1.50372654e-01
-1.38014317e+00 8.12609911e-01 -7.70219624e-01 -6.72514141e-01
4.58990067e-01 -3.32369059e-01 -2.53409505e-01 9.77878124e-02
-6.32401466e-01 9.42639410e-01 4.34549749e-02 -3.61187071e-01
-6.21403754e-01 -1.09081745e+00 -4.32387143e-01 1.62054747e-01
2.26168573e-01 -9.25912321e-01 1.44762814e+00 1.45321503e-01
-1.13775814e+00 7.77187765e-01 -3.96234870e-01 -4.35813993e-01
5.66427350e-01 -3.72001827e-01 -1.30622640e-01 1.57680526e-01
2.98988342e-01 4.27221179e-01 -2.26734444e-01 -1.07578707e+00
-5.07188916e-01 -2.64526695e-01 -3.49267125e-01 1.79209575e-01
-2.46481374e-01 3.86983335e-01 -4.12335783e-01 -5.47034204e-01
-3.95835549e-01 -6.15433156e-01 6.74087033e-02 -5.86402655e-01
-9.14806128e-01 -5.59271216e-01 3.20765108e-01 -1.14243209e+00
1.50971758e+00 -1.49511182e+00 3.10148448e-01 -3.49794567e-01
5.91297209e-01 -4.56782356e-02 -2.94253439e-01 1.02443862e+00
2.68089622e-01 8.14562261e-01 -1.36122093e-01 -6.59407258e-01
-1.66352183e-01 -5.12927115e-01 -5.68106651e-01 2.71645010e-01
5.27828261e-02 1.01794410e+00 -1.12515187e+00 -4.41203624e-01
-3.90619963e-01 7.36067370e-02 -1.93806902e-01 7.56904483e-02
-2.75705397e-01 5.25285959e-01 -3.16561818e-01 5.52474499e-01
4.15449947e-01 -4.58370835e-01 -2.04118073e-01 -5.21092005e-02
-4.59834993e-01 6.46900237e-01 -4.42539364e-01 1.83106160e+00
-1.84929464e-02 1.06843448e+00 -6.07469045e-02 -3.97481471e-01
7.96905160e-01 3.22843909e-01 3.29053819e-01 -1.89362228e-01
-1.25697240e-01 1.49647921e-01 -9.07793194e-02 -3.23720038e-01
1.12900281e+00 5.42531431e-01 -2.41320550e-01 8.93225968e-01
2.00735062e-01 -4.87112880e-01 8.65240574e-01 9.80335355e-01
1.47783494e+00 -1.13209955e-01 3.63540649e-01 -6.42182767e-01
3.09046626e-01 4.17726845e-01 4.90442932e-01 1.13368559e+00
2.25251261e-02 9.05530155e-01 8.38532388e-01 -1.34229437e-01
-1.27453518e+00 -6.41873777e-01 -8.56049210e-02 8.82283926e-01
-3.40676159e-01 -1.23332477e+00 -5.99115014e-01 -3.43373001e-01
-1.72876015e-01 1.13424563e+00 -6.76279485e-01 2.01718360e-02
-3.66715819e-01 -6.94367349e-01 9.81507242e-01 3.10223848e-01
8.51576496e-03 -8.94209445e-01 -5.04340231e-01 -1.13935292e-01
-7.23350108e-01 -9.38117683e-01 -8.37265790e-01 -2.97951579e-01
-9.31812525e-01 -1.03503036e+00 -7.52383232e-01 -1.47060081e-01
2.23896444e-01 4.59581494e-01 1.46297538e+00 4.06694785e-02
-2.12299734e-01 4.43758994e-01 -3.44266564e-01 -8.05707037e-01
-7.07069218e-01 4.50751424e-01 -1.55178040e-01 -9.81245399e-01
4.38461863e-02 -3.55220884e-01 -6.75285757e-01 -1.05768926e-01
-7.36201882e-01 3.47631544e-01 4.43651706e-01 5.49160957e-01
3.74086440e-01 -7.86524832e-01 8.20971906e-01 -1.05840409e+00
1.41614103e+00 -4.78183657e-01 1.69341806e-02 5.29918551e-01
-6.75113976e-01 -1.01869598e-01 4.34859991e-01 2.24056870e-01
-1.05154192e+00 -1.01593053e+00 1.53595865e-01 2.38092586e-01
8.59848559e-02 7.80849874e-01 3.53919894e-01 3.85456413e-01
1.01761019e+00 4.07952480e-02 -6.16166666e-02 -6.78475380e-01
4.93409842e-01 7.72029221e-01 7.86342263e-01 -6.69713736e-01
2.38513023e-01 1.37936816e-01 -4.18858737e-01 -7.88314879e-01
-1.01057410e+00 -6.03740931e-01 -2.25436106e-01 -2.75458116e-02
5.38732708e-01 -9.71011460e-01 -5.74675858e-01 1.60717830e-01
-1.53985620e+00 -1.91441298e-01 -4.37424093e-01 6.87962398e-02
-3.49219739e-01 6.10642076e-01 -8.21166873e-01 -5.24554789e-01
-1.24104857e+00 -7.26862848e-01 8.91166806e-01 4.44293797e-01
-1.03449047e+00 -8.79718065e-01 6.25552297e-01 7.40063190e-01
4.77316469e-01 4.04538423e-01 7.20905900e-01 -1.06072676e+00
-1.10953093e-01 -2.17435330e-01 -1.78145438e-01 3.11638438e-03
-1.90575477e-02 7.43961155e-01 -7.93620527e-01 -3.26891154e-01
-1.82382375e-01 -5.30119896e-01 1.22916734e+00 4.90934879e-01
1.08925784e+00 -5.93452275e-01 -3.05325776e-01 2.32295066e-01
8.99691641e-01 -2.00679436e-01 5.19911647e-01 3.18503618e-01
5.27668834e-01 5.44219077e-01 4.46688592e-01 8.00507247e-01
6.27270579e-01 3.67177457e-01 -2.48994082e-01 3.43599141e-01
-4.04527873e-01 -2.04631016e-01 2.14779407e-01 1.31488943e+00
-2.18245596e-01 -9.15744662e-01 -1.19678175e+00 6.62314594e-01
-1.97628760e+00 -1.04392266e+00 -3.42232823e-01 2.09175873e+00
1.01139247e+00 6.46917000e-02 1.16429232e-01 -3.76005143e-01
6.03351057e-01 3.33880037e-01 -4.41552937e-01 -7.33198524e-01
-4.24371600e-01 -1.77796185e-01 2.77753204e-01 4.30313110e-01
-5.02642453e-01 5.37550151e-01 7.33223438e+00 9.04676020e-01
-7.18069017e-01 7.21720904e-02 7.77639031e-01 -4.80151415e-01
-7.24318802e-01 4.98775877e-02 -7.90428638e-01 5.68683386e-01
1.24670780e+00 -1.15069175e+00 -4.11612913e-02 2.62267768e-01
4.56208825e-01 -2.99101651e-01 -1.19770551e+00 3.77232641e-01
2.82647133e-01 -1.80184722e+00 2.25362077e-01 -4.71245535e-02
9.07760441e-01 3.98698747e-01 -1.30218983e-01 2.16051251e-01
5.83650708e-01 -9.43408847e-01 5.69173217e-01 8.41047823e-01
6.68217421e-01 -4.01599497e-01 7.29083538e-01 3.95094186e-01
-2.48774976e-01 5.12754142e-01 -3.55462343e-01 -5.85459135e-02
1.55785605e-01 7.73988724e-01 -8.78604650e-01 7.62263119e-01
4.48117882e-01 1.00281072e+00 -9.65871096e-01 1.13550842e+00
-9.85777155e-02 6.67656898e-01 -2.08816249e-02 -1.15734659e-01
-2.17993289e-01 -4.00541723e-02 1.02029276e+00 1.70527840e+00
4.18750823e-01 -6.32205326e-03 -1.26230910e-01 8.34810972e-01
-5.50975680e-01 2.39849046e-01 -5.31719625e-01 -4.77960408e-01
7.85195291e-01 1.52412534e+00 -6.25791728e-01 -6.56226635e-01
7.54345674e-03 4.41538751e-01 3.93682957e-01 2.04885602e-01
-4.97701705e-01 -4.46671396e-01 2.02990904e-01 -2.04477817e-01
-2.96908259e-01 -1.31014749e-01 -9.58063245e-01 -1.41584957e+00
-5.99152036e-02 -9.21799481e-01 5.63338935e-01 -9.82910514e-01
-1.08856988e+00 5.24127603e-01 5.53419022e-03 -9.04400766e-01
-6.45491108e-02 1.04856662e-01 -1.00199509e+00 8.36196601e-01
-1.07341504e+00 -7.60329902e-01 -3.73287499e-01 -2.85787344e-01
7.09359884e-01 -3.99072580e-02 6.89826190e-01 -2.96302825e-01
-7.84901857e-01 5.33461630e-01 3.60122800e-01 -4.57401723e-01
1.49161589e+00 -1.35754895e+00 7.42867291e-01 7.63187289e-01
-2.20358069e-03 8.96372139e-01 1.23292708e+00 -8.38428080e-01
-1.31417751e+00 -7.26913810e-01 1.22763777e+00 -1.00400305e+00
5.86509228e-01 2.50169579e-02 -8.94692421e-01 5.25391757e-01
1.01065552e+00 -8.03449690e-01 1.11180496e+00 4.87390578e-01
-1.97989658e-01 3.79845232e-01 -8.71633589e-01 5.65046966e-01
9.55772519e-01 -1.18699595e-01 -7.08493292e-01 6.34629309e-01
7.06779659e-01 -5.22475421e-01 -1.18150556e+00 2.58948237e-01
4.66585904e-01 -6.90337539e-01 5.73478818e-01 -5.65457523e-01
1.28946459e+00 -4.16559689e-02 3.14658910e-01 -1.65777218e+00
-5.34946561e-01 -6.54966056e-01 -2.78867453e-01 1.50195920e+00
7.31835008e-01 -3.16479892e-01 4.37271506e-01 7.59940982e-01
-5.00092030e-01 -7.78003514e-01 -5.03850162e-01 -5.49527168e-01
4.17745948e-01 1.02043077e-01 3.14013213e-01 8.86748314e-01
6.67772233e-01 7.30994582e-01 -1.77924454e-01 -5.92083871e-01
4.67352957e-01 1.98235735e-01 8.92315626e-01 -1.20945489e+00
3.64832282e-02 -8.44283760e-01 1.69747964e-01 -6.96299791e-01
-1.00901760e-01 -1.07602823e+00 -1.30968213e-01 -2.20615077e+00
9.68270421e-01 2.17025250e-01 -1.88313589e-01 4.22390461e-01
-4.97551054e-01 9.21531767e-02 2.74753690e-01 5.81552923e-01
-1.21074128e+00 4.10706818e-01 1.27192926e+00 -1.49626508e-01
-1.09845877e-01 -3.10573786e-01 -1.56890082e+00 3.11114043e-01
9.80826437e-01 -2.72131592e-01 -5.12976572e-02 -6.76280081e-01
4.72651213e-01 1.47463962e-01 -2.92195603e-02 -9.08690631e-01
6.10521317e-01 -2.23980367e-01 3.19190592e-01 -8.21175516e-01
-7.83502385e-02 5.32269061e-01 4.33858577e-03 1.03765249e-01
-9.58902657e-01 2.63189375e-01 4.35916722e-01 3.25513333e-01
1.37484754e-02 -9.55510810e-02 3.23227078e-01 -2.44708225e-01
8.03480390e-03 -1.02503344e-01 -3.48310858e-01 5.82183301e-01
3.60234112e-01 1.92956015e-01 -1.29210603e+00 -5.31929910e-01
-2.26873711e-01 7.24969506e-01 7.32275784e-01 3.99897039e-01
3.65642846e-01 -7.95931876e-01 -1.45185018e+00 -6.65699363e-01
2.56631523e-01 -1.79911017e-01 3.40837687e-01 9.40453947e-01
-5.51551223e-01 7.53000259e-01 -2.48369053e-01 -2.61645257e-01
-1.36724114e+00 1.29605815e-01 -2.94831336e-01 -5.11189163e-01
-5.19528925e-01 7.59745419e-01 -1.80015177e-01 -2.79423803e-01
-1.95350200e-02 -1.30611897e-01 -4.63283390e-01 3.01499277e-01
9.45093453e-01 1.05273163e+00 2.81095922e-01 -2.96748787e-01
-2.97295451e-01 9.30780694e-02 -4.97925252e-01 -2.79260874e-01
1.51432383e+00 -1.80762053e-01 -4.63713735e-01 6.66844785e-01
9.16140854e-01 4.85715359e-01 -6.45202100e-01 -1.29815444e-01
1.81829691e-01 -2.11607099e-01 -3.31647135e-02 -1.25126922e+00
-3.49575281e-01 6.43468618e-01 -4.02962416e-01 2.49801293e-01
5.87753654e-01 7.51344860e-02 6.43112779e-01 5.24550617e-01
-5.58922924e-02 -1.06034267e+00 -2.49887258e-02 6.18201256e-01
1.44198155e+00 -1.11204195e+00 7.29152024e-01 -1.47301108e-01
-7.79161394e-01 1.13949049e+00 5.61800897e-01 2.18464598e-01
-6.78215176e-02 1.40781820e-01 -7.99899846e-02 -4.17718381e-01
-1.53293931e+00 6.42313123e-01 4.52522457e-01 8.33386853e-02
1.15630066e+00 4.68243919e-02 -8.89905930e-01 8.00941110e-01
-5.85555136e-01 3.90817784e-02 1.30895972e+00 8.10709119e-01
-5.25314152e-01 -7.78214276e-01 -2.30998173e-01 8.59134138e-01
-6.54005229e-01 -1.26990601e-01 -8.71716261e-01 1.95204735e-01
-5.66576064e-01 1.22607851e+00 -1.21320315e-01 -1.68915331e-01
3.70240897e-01 -1.02156624e-01 4.57585484e-01 -9.27972555e-01
-7.80262589e-01 -1.12632558e-01 5.17395973e-01 -3.48589003e-01
-1.17605112e-01 -9.04582918e-01 -7.76032627e-01 -8.41876745e-01
7.82239735e-02 6.57294214e-01 4.90129590e-01 4.51696754e-01
9.23261821e-01 7.74131477e-01 2.22115576e-01 -6.88872695e-01
-5.85208476e-01 -1.41435730e+00 -1.92546502e-01 1.92218274e-01
2.69653887e-01 -1.90502778e-03 -4.05477196e-01 1.08345360e-01] | [12.395326614379883, 9.537493705749512] |
ea45b78d-64a2-4dd9-9447-a6cdb79b575c | probabilistic-deep-learning-to-quantify | 2112.02622 | null | https://arxiv.org/abs/2112.02622v1 | https://arxiv.org/pdf/2112.02622v1.pdf | Probabilistic Deep Learning to Quantify Uncertainty in Air Quality Forecasting | Data-driven forecasts of air quality have recently achieved more accurate short-term predictions. Despite their success, most of the current data-driven solutions lack proper quantifications of model uncertainty that communicate how much to trust the forecasts. Recently, several practical tools to estimate uncertainty have been developed in probabilistic deep learning. However, there have not been empirical applications and extensive comparisons of these tools in the domain of air quality forecasts. Therefore, this work applies state-of-the-art techniques of uncertainty quantification in a real-world setting of air quality forecasts. Through extensive experiments, we describe training probabilistic models and evaluate their predictive uncertainties based on empirical performance, reliability of confidence estimate, and practical applicability. We also propose improving these models using "free" adversarial training and exploiting temporal and spatial correlation inherent in air quality data. Our experiments demonstrate that the proposed models perform better than previous works in quantifying uncertainty in data-driven air quality forecasts. Overall, Bayesian neural networks provide a more reliable uncertainty estimate but can be challenging to implement and scale. Other scalable methods, such as deep ensemble, Monte Carlo (MC) dropout, and stochastic weight averaging-Gaussian (SWAG), can perform well if applied correctly but with different tradeoffs and slight variations in performance metrics. Finally, our results show the practical impact of uncertainty estimation and demonstrate that, indeed, probabilistic models are more suitable for making informed decisions. Code and dataset are available at \url{https://github.com/Abdulmajid-Murad/deep_probabilistic_forecast} | ['Gavin Taylor', 'Kerstin Bach', 'Frank Alexander Kraemer', 'Abdulmajid Murad'] | 2021-12-05 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [-4.37364906e-01 -3.30334038e-01 6.90791309e-02 -7.89538443e-01
-1.08192837e+00 -6.37058258e-01 7.34208703e-01 4.69466187e-02
-9.18420553e-02 1.15117431e+00 2.45262682e-01 -6.93680704e-01
-4.05749291e-01 -1.14503932e+00 -9.10576284e-01 -7.61585236e-01
-9.70863998e-02 5.46106577e-01 -6.76386058e-03 1.35173753e-01
1.14780236e-02 2.01603442e-01 -1.32406962e+00 -1.86810293e-03
1.08939183e+00 1.25601339e+00 -3.85221392e-01 7.18155563e-01
-1.75064117e-01 6.33616149e-01 -5.96845210e-01 -4.99556810e-01
1.67298913e-01 3.26303989e-02 -3.74374062e-01 -7.01773942e-01
2.92279035e-01 -6.12568796e-01 -1.94290787e-01 1.07349551e+00
7.89919019e-01 4.46604013e-01 8.10561538e-01 -1.32167840e+00
-6.30507469e-01 1.01356888e+00 -8.00268352e-02 2.70611435e-01
-1.00403465e-01 3.22957903e-01 8.00038517e-01 -7.26515532e-01
-2.84098148e-01 1.23598814e+00 9.66695011e-01 4.66188222e-01
-1.06459606e+00 -1.13231504e+00 3.30978274e-01 -6.16558082e-02
-1.30286694e+00 -3.61041933e-01 4.43708539e-01 -7.17748046e-01
7.89926112e-01 2.83081979e-01 4.58682887e-02 1.39834714e+00
2.25230023e-01 3.76286775e-01 1.33188891e+00 1.12450279e-01
7.43174434e-01 3.00555468e-01 8.30921829e-02 8.72929171e-02
2.95360476e-01 9.18595552e-01 -2.31481418e-01 -4.30161387e-01
3.34460020e-01 2.44259045e-01 -1.86098263e-01 1.43415675e-01
-8.36260557e-01 9.14563119e-01 5.89043975e-01 1.94765205e-04
-3.11383337e-01 6.67179644e-01 1.45329967e-01 -1.82043947e-02
1.16642475e+00 1.83319315e-01 -7.34640896e-01 -2.92772114e-01
-1.33200049e+00 5.50609529e-01 7.73362041e-01 7.05709636e-01
2.65213102e-01 3.53148222e-01 -4.25935209e-01 3.64445627e-01
6.56501949e-01 1.08575845e+00 5.26937358e-02 -8.77391338e-01
5.09511471e-01 -2.04167694e-01 7.59102583e-01 -9.38501716e-01
-4.87334609e-01 -5.08523047e-01 -9.49560285e-01 4.49021637e-01
3.67830664e-01 -7.17519403e-01 -9.06218231e-01 1.63473976e+00
2.29501083e-01 7.37956464e-01 6.22766279e-02 8.71970296e-01
8.11582029e-01 1.00026917e+00 2.69251645e-01 4.15142238e-01
1.07007921e+00 -6.08039081e-01 -6.43462121e-01 4.58761454e-02
1.81705549e-01 -4.63210195e-01 7.32775211e-01 2.99628109e-01
-7.80549943e-01 -5.99279344e-01 -1.06627047e+00 6.38588607e-01
-8.45092475e-01 -1.58056453e-01 4.57915485e-01 1.22342801e+00
-7.25135565e-01 1.00832105e+00 -1.08214784e+00 3.71512920e-01
3.80698830e-01 1.31255060e-01 1.63736135e-01 -2.33649136e-03
-1.85596132e+00 8.39970171e-01 3.93231571e-01 2.66737252e-01
-1.22743607e+00 -1.09869289e+00 -8.21384192e-01 1.98400646e-01
1.24045290e-01 -8.13740551e-01 1.50172067e+00 -3.63821536e-01
-1.41910434e+00 -1.79682493e-01 1.24903880e-01 -7.90307224e-01
5.53264976e-01 -4.66535568e-01 -8.32999468e-01 -3.72766137e-01
-3.63034010e-01 5.89895427e-01 8.73805165e-01 -1.17000353e+00
-7.49909341e-01 -1.34822920e-01 1.05440125e-01 -3.06554854e-01
2.92586687e-04 5.90945706e-02 3.57215941e-01 -4.93995428e-01
-4.71871406e-01 -8.37880790e-01 -4.04601455e-01 -1.77485228e-01
-4.22196418e-01 -5.28021390e-03 4.40817773e-01 -7.77209640e-01
1.38111711e+00 -1.94452715e+00 -4.29316431e-01 3.00716758e-01
4.72822562e-02 3.40590566e-01 1.26602978e-01 4.13730264e-01
1.04180001e-01 5.98854721e-01 -6.10608041e-01 -5.13784468e-01
5.70160031e-01 1.89732939e-01 -7.46109307e-01 2.54859239e-01
3.34403396e-01 8.10901880e-01 -9.09752309e-01 3.20589952e-02
5.95319152e-01 8.48679245e-01 -2.00453237e-01 1.81446329e-01
-6.03287458e-01 7.53041923e-01 -4.42065001e-01 3.41025472e-01
1.04518223e+00 -1.42175004e-01 -3.06216627e-01 4.26683128e-02
-2.80876532e-02 3.99951220e-01 -1.33871174e+00 1.15345788e+00
-7.57330835e-01 3.33244830e-01 -2.01181158e-01 -4.90448087e-01
6.39876783e-01 5.36861241e-01 3.51420790e-02 -8.22941810e-02
1.52556747e-01 1.68551907e-01 -1.39603436e-01 2.06005070e-02
4.32468385e-01 -1.04081936e-01 2.75585521e-02 3.92870069e-01
-1.60216019e-01 -4.98896033e-01 -1.85048491e-01 -1.05446801e-01
5.95456779e-01 2.22963616e-01 -1.75998718e-01 3.02028805e-02
2.46685684e-01 -2.46373996e-01 4.36596900e-01 9.00135994e-01
-3.45159531e-01 7.71711946e-01 6.74898326e-02 -4.39298362e-01
-8.90432537e-01 -1.25718009e+00 -5.43468118e-01 9.45611537e-01
-3.18227232e-01 1.40245678e-02 -6.38729811e-01 -5.92752397e-01
4.34712499e-01 1.48263645e+00 -5.70638239e-01 8.38633478e-02
6.85011595e-02 -1.23172045e+00 5.08264601e-01 7.18787372e-01
3.97673696e-01 -6.39353275e-01 -1.04066044e-01 2.08940595e-01
1.87426373e-01 -9.22022641e-01 -2.00004801e-01 -1.07337274e-01
-7.93291986e-01 -7.03521073e-01 -7.13212252e-01 4.38413620e-01
2.83885986e-01 -1.75964594e-01 1.36026061e+00 -5.91039181e-01
3.23828250e-01 4.82303679e-01 -5.59023358e-02 -9.60132599e-01
-5.17977953e-01 -1.70413807e-01 3.98751348e-01 -1.96832106e-01
3.98552418e-01 -7.15790868e-01 -7.27555215e-01 3.17321986e-01
-9.78344500e-01 -4.93984550e-01 2.80036211e-01 6.23509526e-01
4.66946781e-01 4.91141260e-01 7.68174231e-01 -5.61016440e-01
8.72807384e-01 -7.46521115e-01 -1.05520320e+00 5.34111299e-02
-9.78928864e-01 1.51194692e-01 5.38933992e-01 6.40963018e-02
-1.33030307e+00 -3.74285012e-01 -4.25603837e-01 -4.54273731e-01
-4.44342941e-01 6.35699451e-01 2.63832152e-01 3.17708552e-01
7.29991376e-01 2.68720277e-02 -3.22421610e-01 -6.25064075e-01
5.28939307e-01 7.91273952e-01 2.30688035e-01 -7.61714697e-01
7.52547026e-01 3.31839204e-01 -2.34883055e-01 -1.32077187e-01
-1.23715138e+00 -1.19839519e-01 -3.15316111e-01 -6.64795712e-02
7.89847970e-01 -1.28008878e+00 -5.50044537e-01 3.73500943e-01
-1.08025312e+00 -2.20187441e-01 -2.63415456e-01 9.25624847e-01
-1.35899633e-01 1.26967952e-01 -4.40195411e-01 -1.43404472e+00
-6.64112508e-01 -1.19968510e+00 9.06851232e-01 2.40088329e-01
1.20458556e-02 -1.33038175e+00 3.36727083e-01 8.20736513e-02
1.11660981e+00 3.18757206e-01 5.62243879e-01 -8.40777576e-01
-5.18388569e-01 -2.97801018e-01 -2.78026760e-01 8.09652090e-01
-8.02152976e-02 3.61266047e-01 -1.30935705e+00 -1.06543303e-01
-1.71054274e-01 -2.22021118e-01 9.92011368e-01 8.61536086e-01
1.21023393e+00 -3.96410525e-01 -1.44594416e-01 3.88556123e-01
1.22049296e+00 3.76635231e-03 4.86025959e-01 -2.73001082e-02
3.04838568e-01 3.50907058e-01 3.94168973e-01 6.03665590e-01
3.55666667e-01 2.14115903e-01 7.15218604e-01 2.42324606e-01
2.15231732e-01 -2.73631483e-01 4.10586059e-01 6.37022555e-01
-1.30427003e-01 -5.49383104e-01 -9.60865974e-01 4.21458155e-01
-1.88942194e+00 -1.03858197e+00 -1.43230231e-02 2.43255877e+00
9.77471113e-01 2.26021349e-01 -1.64054006e-01 -2.37263918e-01
6.39524043e-01 2.99093008e-01 -6.22524679e-01 -3.84087056e-01
2.36899748e-01 2.33464181e-01 7.29754865e-01 6.06847286e-01
-1.50344384e+00 5.37552059e-01 6.64321184e+00 9.83940363e-01
-8.85329127e-01 3.59980047e-01 1.06392646e+00 -2.95837730e-01
-5.44618309e-01 -2.56963402e-01 -1.00018644e+00 9.53718722e-01
1.69361830e+00 7.81232044e-02 3.35042387e-01 1.07325423e+00
3.97230804e-01 5.19554727e-02 -9.85327244e-01 6.17357433e-01
-5.97697258e-01 -1.45691240e+00 -1.87655315e-01 -2.77568698e-01
1.14758039e+00 5.05766749e-01 5.32624781e-01 5.21433532e-01
1.03293073e+00 -1.41098213e+00 5.47379434e-01 1.01813400e+00
5.82495987e-01 -9.88018274e-01 1.28516281e+00 3.29933554e-01
-9.04296458e-01 -1.03728287e-02 -6.14798486e-01 -4.78292741e-02
3.21773440e-01 1.41056716e+00 -4.56214756e-01 7.09510148e-01
9.23943460e-01 3.68512571e-01 -1.21039517e-01 9.05627131e-01
-3.24529916e-01 1.05758762e+00 -5.51138520e-01 -5.33988141e-02
4.56363559e-01 -1.57490402e-01 5.12517571e-01 1.17017889e+00
7.42482364e-01 -2.07552627e-01 -1.63114220e-01 1.36182070e+00
2.33731400e-02 -3.31685215e-01 -6.37152433e-01 -1.48675874e-01
6.35783911e-01 9.15728807e-01 5.44346608e-02 -2.95106202e-01
-4.18080747e-01 2.07710236e-01 -1.16643965e-01 5.19847631e-01
-1.25368142e+00 -1.80092692e-01 9.83615100e-01 -2.21534878e-01
3.70571911e-01 -3.03422391e-01 -2.13498771e-01 -8.52690995e-01
-1.67508423e-01 -7.12201416e-01 2.54730850e-01 -7.74371505e-01
-1.81424510e+00 6.82996154e-01 3.69423300e-01 -1.17624354e+00
-4.94985878e-01 -7.26005793e-01 -7.58525491e-01 1.32533848e+00
-1.66103172e+00 -9.28399742e-01 -6.23682961e-02 1.43582851e-01
1.15098603e-01 -2.32172221e-01 9.86590326e-01 1.53136969e-01
-5.90935290e-01 3.45649004e-01 8.08076620e-01 -3.42432037e-02
5.86528063e-01 -1.32356715e+00 8.18898559e-01 8.33681941e-01
-1.17078088e-02 3.59139472e-01 8.68170679e-01 -6.27105594e-01
-8.29653800e-01 -1.49759591e+00 5.38130641e-01 -8.60369265e-01
8.38889241e-01 -1.83372468e-01 -7.27570355e-01 5.62654316e-01
2.76691347e-01 9.64784473e-02 7.17844188e-01 1.88434556e-01
-4.66355503e-01 -1.95978582e-01 -1.45585835e+00 2.93744087e-01
4.23126757e-01 -4.62180465e-01 -3.89372259e-01 6.01515949e-01
1.05731535e+00 -3.70320976e-01 -1.17524505e+00 5.40294826e-01
4.80267614e-01 -1.17362797e+00 1.05869710e+00 -7.44739413e-01
5.39460421e-01 -4.82225895e-01 -3.17955613e-01 -1.79327917e+00
-1.72846332e-01 -3.02866518e-01 -4.67372626e-01 1.31717622e+00
8.30440283e-01 -1.03738379e+00 3.97061080e-01 1.23982322e+00
3.32763530e-02 -6.34859383e-01 -1.00022721e+00 -9.30059910e-01
4.82540846e-01 -1.18938076e+00 1.20746660e+00 6.80709183e-01
-6.14558101e-01 -4.23271239e-01 -5.84269941e-01 7.82748342e-01
6.35000229e-01 5.87297790e-03 4.33371782e-01 -1.17457533e+00
-2.84071386e-01 -4.38938648e-01 -1.16034560e-01 -6.42135620e-01
8.84212255e-02 -4.90669847e-01 1.06735773e-01 -1.45031655e+00
-2.38572627e-01 -5.56332111e-01 -7.58419812e-01 2.14949444e-01
-2.77700394e-01 -1.26698971e-01 -1.28538802e-01 -9.93122160e-02
-3.59307706e-01 8.15214515e-01 7.70472229e-01 -3.89364153e-01
2.19322015e-02 5.29355288e-01 -5.31256795e-01 5.83630502e-01
1.13784909e+00 -6.85769677e-01 -4.33340281e-01 -4.34079885e-01
3.30153912e-01 -1.72424287e-01 5.04271567e-01 -1.21699762e+00
1.77498877e-01 -2.20565870e-01 6.00396693e-01 -8.32136393e-01
3.46205890e-01 -8.26028407e-01 4.83264208e-01 2.01754153e-01
-3.79006147e-01 5.31819090e-02 5.40275991e-01 9.19777811e-01
-2.81721503e-01 -2.44283274e-01 5.26681960e-01 -2.04618633e-01
-3.78065884e-01 5.63390315e-01 -2.82913804e-01 -1.86292864e-02
6.39931738e-01 5.24325907e-01 -3.29877913e-01 -7.62566447e-01
-7.09999502e-01 4.42076176e-01 -7.47633800e-02 2.21616030e-01
4.28998917e-01 -1.34511638e+00 -1.00656343e+00 -1.30305111e-01
-1.44114226e-01 3.63422297e-02 6.02425277e-01 5.30968666e-01
-2.61955947e-01 8.30102980e-01 3.26954246e-01 -6.18595660e-01
-3.41242701e-01 5.27232766e-01 7.17453778e-01 -3.55888218e-01
-5.13341315e-02 9.82129157e-01 -5.79236425e-04 -9.80413139e-01
3.58772576e-01 -8.33509028e-01 1.52242631e-01 -1.75225928e-01
8.71270418e-01 5.83434403e-01 1.39690757e-01 -1.25280604e-01
-3.33732277e-01 1.66976124e-01 3.24338377e-01 -2.19423622e-01
1.17984569e+00 -8.88124853e-02 1.85487390e-01 5.32403588e-01
6.94196105e-01 -8.21409300e-02 -1.54977787e+00 -1.51471540e-01
-2.01061308e-01 -4.31247979e-01 4.83245760e-01 -1.24707520e+00
-1.14487672e+00 1.24556756e+00 9.49110866e-01 2.68012136e-01
7.59686053e-01 -4.89775777e-01 7.51493692e-01 3.40301841e-01
1.83333591e-01 -9.39050734e-01 -7.49903798e-01 4.67335105e-01
9.26326692e-01 -1.68615603e+00 -1.66756555e-03 1.52854845e-01
-6.96239054e-01 7.87272096e-01 2.96453118e-01 8.31273496e-02
1.40776753e+00 2.69901454e-01 1.68202430e-01 7.11618513e-02
-8.00521731e-01 -1.01400137e-01 6.19173169e-01 7.31952667e-01
3.77773196e-01 5.22899032e-01 2.76262790e-01 1.03487003e+00
-1.64378971e-01 6.80290014e-02 8.72124359e-02 4.09635365e-01
-3.64353925e-01 -7.63859868e-01 -6.21339321e-01 5.09333670e-01
-5.57483435e-01 -5.62424660e-01 1.60967276e-01 4.70132709e-01
-1.52716637e-01 1.23248088e+00 -8.03530496e-03 -3.69759470e-01
1.54605746e-01 3.96378785e-01 -1.04247659e-01 -2.27822229e-01
-4.90190238e-01 -4.56578434e-01 2.17100054e-01 -4.10009891e-01
-3.43350321e-01 -2.88054228e-01 -8.76991630e-01 -7.57453561e-01
-1.17048264e-01 3.04158270e-01 1.04639995e+00 8.24198365e-01
6.72587395e-01 6.80325210e-01 5.12158453e-01 -1.06500149e+00
-1.21131825e+00 -1.16068697e+00 -7.04744279e-01 2.26864535e-02
3.27547759e-01 -7.39313781e-01 -7.96276867e-01 -4.21255618e-01] | [7.033151149749756, 3.359851121902466] |
013b80c4-b9ed-454a-b6e7-ca576d71355d | rasat-integrating-relational-structures-into | 2205.06983 | null | https://arxiv.org/abs/2205.06983v2 | https://arxiv.org/pdf/2205.06983v2.pdf | RASAT: Integrating Relational Structures into Pretrained Seq2Seq Model for Text-to-SQL | Relational structures such as schema linking and schema encoding have been validated as a key component to qualitatively translating natural language into SQL queries. However, introducing these structural relations comes with prices: they often result in a specialized model structure, which largely prohibits using large pretrained models in text-to-SQL. To address this problem, we propose RASAT: a Transformer seq2seq architecture augmented with relation-aware self-attention that could leverage a variety of relational structures while inheriting the pretrained parameters from the T5 model effectively. Our model can incorporate almost all types of existing relations in the literature, and in addition, we propose introducing co-reference relations for the multi-turn scenario. Experimental results on three widely used text-to-SQL datasets, covering both single-turn and multi-turn scenarios, have shown that RASAT could achieve state-of-the-art results across all three benchmarks (75.5% EX on Spider, 52.6% IEX on SParC, and 37.4% IEX on CoSQL). | ['Zhouhan Lin', 'Quanshi Zhang', 'Xinbing Wang', 'Chenghu Zhou', 'Yu Cheng', 'Xiangpeng Wan', 'Ziwei He', 'Jingyao Tang', 'Jiexing Qi'] | 2022-05-14 | null | null | null | null | ['text-to-sql'] | ['computer-code'] | [-5.21474108e-02 3.45777810e-01 -6.06398225e-01 -7.01165617e-01
-9.64913845e-01 -8.42007816e-01 4.51240838e-01 1.60623312e-01
-2.09052324e-01 4.14301842e-01 5.05006075e-01 -7.42938340e-01
-2.65103560e-02 -1.28813469e+00 -1.41784668e+00 1.34316459e-01
5.00401035e-02 7.93618679e-01 1.88216954e-01 -7.36757636e-01
-2.19279677e-01 1.37564361e-01 -1.33024228e+00 7.52639055e-01
1.07588017e+00 1.09944010e+00 -2.45530963e-01 5.44992089e-02
-6.34336829e-01 1.21603274e+00 -3.45708698e-01 -9.95961905e-01
2.37740532e-01 -1.06094209e-02 -1.23354495e+00 -6.86068892e-01
5.58076859e-01 -2.35492080e-01 -4.49736625e-01 8.22262466e-01
2.68760383e-01 -6.50008991e-02 -3.52904536e-02 -8.90456080e-01
-9.41025853e-01 1.45464373e+00 -3.66756469e-01 5.48237516e-03
4.75979209e-01 1.16564333e-01 1.56604230e+00 -6.69887662e-01
8.18957508e-01 1.58390963e+00 6.03167236e-01 2.22189784e-01
-1.34406483e+00 -6.31371021e-01 3.79302762e-02 1.29480720e-01
-1.24009633e+00 -5.01529634e-01 4.76683915e-01 -1.20166596e-02
1.66896212e+00 5.22146583e-01 2.41680697e-01 1.02517128e+00
8.39860663e-02 8.02079618e-01 8.47858906e-01 -2.70475837e-04
-3.05679530e-01 -5.20250294e-03 2.13113233e-01 6.96904540e-01
1.90769196e-01 1.86337516e-01 -3.77993494e-01 5.68791665e-02
4.38020200e-01 -1.88774958e-01 3.55205499e-02 -3.36446792e-01
-1.23875189e+00 5.66464961e-01 9.77417767e-01 -4.65538986e-02
-2.51916230e-01 2.07814053e-01 8.23230982e-01 6.15115881e-01
2.72106647e-01 8.00220251e-01 -7.79298306e-01 -1.79040939e-01
-2.98547775e-01 3.51220518e-01 9.35550451e-01 1.54452312e+00
8.27434480e-01 -1.84872925e-01 -2.81100273e-01 8.94427717e-01
-2.15783026e-02 6.43285155e-01 4.83002095e-03 -9.12514031e-01
1.21216607e+00 1.26607907e+00 -3.18506837e-01 -7.47013569e-01
-3.12128514e-01 -4.40673888e-01 -7.31354713e-01 -4.72981244e-01
1.30143970e-01 1.99641109e-01 -8.02620411e-01 1.80659556e+00
2.03928366e-01 -2.29776233e-01 3.28218162e-01 6.60693109e-01
1.14623785e+00 5.27843952e-01 -2.20778659e-02 4.40237150e-02
1.42994595e+00 -1.07679868e+00 -5.61437964e-01 -3.73836219e-01
8.63021612e-01 -4.44999665e-01 1.71221876e+00 2.45630089e-02
-1.32623875e+00 -7.23255992e-01 -9.36317682e-01 -8.07630360e-01
-5.53372920e-01 -1.95898354e-01 9.19447362e-01 3.55577350e-01
-7.64079809e-01 3.50180715e-01 -8.41656208e-01 -4.86899614e-01
3.75789762e-01 1.90453738e-01 -3.75520289e-01 -2.82585651e-01
-1.43416703e+00 8.74375939e-01 5.76136231e-01 7.89186880e-02
-4.85081583e-01 -1.34828317e+00 -7.82102942e-01 2.69201279e-01
1.10108519e+00 -7.59743750e-01 1.13489497e+00 -1.11078404e-01
-1.24282348e+00 7.20528603e-01 -1.88743591e-01 -6.72358215e-01
4.03101563e-01 -6.06811166e-01 -5.79192638e-01 -2.74163693e-01
1.31255299e-01 6.11524642e-01 -2.27853879e-01 -9.42374110e-01
-2.69575745e-01 -3.55373532e-01 6.14952385e-01 3.54517326e-02
-1.29417330e-02 6.09521382e-03 -9.07081246e-01 -5.07897496e-01
2.19661510e-03 -7.67659545e-01 9.40257013e-02 -3.37273031e-01
-7.88963020e-01 -4.55958426e-01 3.25224429e-01 -5.41505575e-01
1.35477781e+00 -1.96312010e+00 1.87326491e-01 2.44211536e-02
1.66714951e-01 1.70276359e-01 -2.53978193e-01 7.37057686e-01
-1.77587926e-01 4.03312922e-01 -9.34211314e-02 1.14595562e-01
1.45380720e-01 3.46736789e-01 -6.98895633e-01 -4.04575318e-01
5.77363551e-01 1.49882126e+00 -6.85159862e-01 -5.09659052e-01
-2.31393576e-01 1.50363728e-01 -9.23266351e-01 2.83901364e-01
-8.50506246e-01 -1.40484378e-01 -2.72833019e-01 7.19792545e-01
6.82487547e-01 -4.07555670e-01 5.05454183e-01 -6.07278824e-01
1.55548617e-01 1.00258827e+00 -7.96965837e-01 1.93349338e+00
-6.69397116e-01 7.45525733e-02 -3.52774143e-01 -6.46511495e-01
8.98189962e-01 -2.34639514e-02 2.95816183e-01 -1.31088018e+00
-2.75513828e-01 1.33830935e-01 7.85839111e-02 -6.42530918e-01
6.52309716e-01 1.56729162e-01 -3.66933405e-01 3.67297381e-01
-1.10185862e-01 -3.18034962e-02 4.20804024e-01 4.57205564e-01
1.19947433e+00 2.28800803e-01 -3.50984633e-02 -2.11153761e-01
2.24401996e-01 5.38195968e-02 6.17690086e-01 5.96962094e-01
5.50612152e-01 2.57402867e-01 1.03940988e+00 -5.73241711e-01
-7.98646390e-01 -1.08358467e+00 5.76611981e-02 1.25182688e+00
5.55755757e-02 -8.98354173e-01 -4.43804443e-01 -7.76506066e-01
2.49417290e-01 1.01381230e+00 -6.68345094e-01 -5.04193008e-01
-8.77200663e-01 -6.07838094e-01 9.27550852e-01 9.25626516e-01
5.68984330e-01 -1.03422070e+00 -3.78453463e-01 3.37754697e-01
-4.54205215e-01 -1.39806640e+00 -3.79040241e-01 3.59029382e-01
-7.44618773e-01 -1.11015666e+00 3.19211096e-01 -2.92419821e-01
1.70489237e-01 -1.67045947e-02 1.83084190e+00 -5.86529039e-02
9.31353197e-02 -2.20347911e-01 -2.72721350e-01 -1.97603852e-01
-4.30738747e-01 5.54629803e-01 -3.99587184e-01 -4.28708166e-01
4.07596827e-01 -3.81412476e-01 -1.25433952e-01 3.72893631e-01
-1.02451408e+00 1.58903018e-01 6.27534151e-01 7.81202257e-01
7.23832726e-01 -1.03316620e-01 5.37807584e-01 -1.62761343e+00
4.30640399e-01 -5.56097209e-01 -7.26509571e-01 7.64991760e-01
-9.65315819e-01 4.76433337e-01 7.08972156e-01 -7.73776248e-02
-1.04275918e+00 -4.47395533e-01 -9.02871788e-02 -2.40568012e-01
3.16230744e-01 1.10442746e+00 -6.53972030e-01 4.51322615e-01
7.84589887e-01 -1.04487374e-01 -4.43077497e-02 -5.90314806e-01
7.51416206e-01 2.13056684e-01 6.93620503e-01 -1.14839590e+00
7.83426523e-01 9.44160894e-02 -1.63849264e-01 -5.05976826e-02
-1.07070720e+00 1.74824707e-02 -3.09717387e-01 5.41490734e-01
6.35121405e-01 -9.73475814e-01 -8.75534058e-01 1.28396571e-01
-1.12807488e+00 -4.39257830e-01 -3.81413460e-01 -7.20783696e-02
-3.01546186e-01 -4.15371172e-02 -7.07081020e-01 -6.46090284e-02
-6.07610285e-01 -1.39104891e+00 9.05928552e-01 -1.27408817e-01
-1.48209244e-01 -6.94736242e-01 -1.54691070e-01 5.27910113e-01
5.57899833e-01 -3.81485596e-02 1.75117588e+00 -6.69217408e-01
-9.58282888e-01 9.80413929e-02 -6.57886684e-01 2.08710462e-01
1.77683875e-01 -5.35862185e-02 -6.72026753e-01 -9.25476551e-02
-4.55148786e-01 -7.61844039e-01 8.08661342e-01 -3.62827539e-01
1.31478012e+00 -3.75820756e-01 -1.82055891e-01 1.01666820e+00
1.47117233e+00 2.04676658e-01 8.54362965e-01 2.82852054e-01
9.97806370e-01 4.88330454e-01 6.56528950e-01 -2.98964418e-02
9.86837327e-01 8.18161249e-01 5.94923079e-01 -5.72564527e-02
-1.38145432e-01 -6.98088169e-01 2.13196829e-01 8.89845073e-01
2.81253487e-01 -2.60519773e-01 -1.23092318e+00 3.04968566e-01
-1.57855964e+00 -6.45668089e-01 -2.00263649e-01 1.97007489e+00
1.33438599e+00 3.24210912e-01 -2.11733300e-02 -1.81455493e-01
2.00825587e-01 1.94148108e-01 -7.53711581e-01 -3.01127076e-01
-3.97302300e-01 5.10918796e-01 4.28187162e-01 2.99515277e-01
-8.49281073e-01 1.31703389e+00 5.90663958e+00 5.31249106e-01
-1.21968067e+00 -2.23109588e-01 4.03011203e-01 -2.58073866e-01
-7.85279274e-01 1.35634288e-01 -8.54369640e-01 2.53561944e-01
1.09848058e+00 -3.00858140e-01 6.75087333e-01 6.57371044e-01
-3.14539909e-01 2.88858801e-01 -1.56539416e+00 5.71851254e-01
-4.32171464e-01 -1.53534067e+00 3.55121672e-01 -1.17550634e-01
4.34755415e-01 3.27219754e-01 2.75868061e-03 9.25769687e-01
5.74058771e-01 -1.15201306e+00 7.12947071e-01 4.59130675e-01
1.16749239e+00 -6.86771631e-01 8.04356992e-01 -5.45860939e-02
-1.14430380e+00 1.40620723e-01 -3.73030037e-01 3.45599800e-01
-1.19445391e-01 4.15335685e-01 -7.08926916e-01 1.11315560e+00
9.29463446e-01 7.25338459e-01 -9.38854039e-01 4.03632969e-01
-1.99288074e-02 5.20085692e-01 -4.73400801e-01 1.56169012e-01
1.10269897e-01 4.41900231e-02 3.38823721e-02 9.61783588e-01
1.00233607e-01 -4.33892272e-02 4.20842227e-03 1.18495631e+00
-6.04104638e-01 1.45816822e-02 -4.54063982e-01 -2.72059143e-01
8.22528660e-01 8.23566496e-01 -2.21260041e-02 -3.78380954e-01
-6.65713966e-01 4.49485779e-01 7.78657079e-01 3.92465323e-01
-1.04156411e+00 -4.40018535e-01 7.23815262e-01 2.18939573e-01
3.24531764e-01 2.26947963e-01 -4.02631372e-01 -1.31209981e+00
6.09384000e-01 -1.43369019e+00 6.21820629e-01 -7.08699048e-01
-1.16168809e+00 8.12664926e-01 1.94438577e-01 -6.80606008e-01
-1.66634053e-01 -2.22795472e-01 -3.10216993e-01 9.11979377e-01
-1.45298266e+00 -1.52173138e+00 -3.31836939e-01 3.64877731e-01
2.14439750e-01 -1.37026712e-01 8.24827790e-01 6.62157357e-01
-7.43771493e-01 1.11157763e+00 -2.89352208e-01 2.50624090e-01
8.08562636e-01 -1.37514889e+00 1.08409727e+00 6.52527273e-01
9.48724449e-02 1.34357262e+00 3.06668282e-01 -5.98928928e-01
-2.10921621e+00 -1.39889443e+00 1.03715599e+00 -6.68765008e-01
7.57613003e-01 -5.54926336e-01 -1.40293646e+00 1.19067693e+00
1.78414196e-01 1.24038860e-01 3.42679173e-01 6.86018407e-01
-1.07834387e+00 -6.90780759e-01 -7.22202837e-01 6.72095835e-01
1.42793119e+00 -8.57357562e-01 -4.79406774e-01 1.09749056e-01
1.42957950e+00 -7.55481184e-01 -1.41836977e+00 8.06853294e-01
4.27398980e-01 -7.00180888e-01 1.11795878e+00 -1.12096679e+00
8.48256469e-01 -1.27754226e-01 -4.57014650e-01 -9.93261158e-01
-2.08446175e-01 -6.12004578e-01 -3.10181230e-01 1.66080105e+00
7.81211913e-01 -6.18383348e-01 7.65562296e-01 6.35512233e-01
-2.34104455e-01 -1.10626650e+00 -6.21856928e-01 -8.78200054e-01
2.64509261e-01 -4.83415008e-01 1.36135185e+00 9.39830303e-01
-9.99932960e-02 8.22144032e-01 -8.09368268e-02 3.51036759e-03
2.23925918e-01 3.76135141e-01 9.39386249e-01 -8.82607520e-01
-3.38951766e-01 -4.41197544e-01 1.51682109e-01 -1.24539554e+00
2.94691265e-01 -1.37473106e+00 -3.24838340e-01 -1.44426429e+00
1.03411272e-01 -9.31105852e-01 -2.75982618e-01 8.13945353e-01
-3.03948015e-01 -4.90007162e-01 6.70359209e-02 -1.02363070e-02
-4.59713906e-01 5.59033632e-01 1.03244531e+00 -3.96516770e-01
1.25454664e-01 -2.28792757e-01 -1.24302602e+00 -3.82580571e-02
4.89956409e-01 -4.36585248e-01 -8.77277553e-01 -1.08202517e+00
7.93451488e-01 3.69868815e-01 2.19768882e-01 -6.77044868e-01
1.96218863e-01 -2.68368423e-01 -2.92187631e-01 -7.68903673e-01
1.29724786e-01 -6.72099411e-01 2.92377293e-01 1.30929112e-01
-7.45041907e-01 5.49857974e-01 3.82277161e-01 3.64717573e-01
-4.34994221e-01 2.86672980e-01 5.34080207e-01 3.95285301e-02
-5.60004294e-01 2.68686235e-01 5.66769600e-01 8.17917466e-01
5.16347170e-01 5.17555714e-01 -9.96622443e-01 -1.16587348e-01
-9.46043283e-02 6.03983641e-01 3.63489300e-01 7.68020928e-01
2.88251996e-01 -1.35552204e+00 -6.53588533e-01 3.02451938e-01
5.10862887e-01 5.21627486e-01 1.49699394e-02 6.26760066e-01
-5.60310721e-01 7.76380718e-01 -1.00557186e-01 -3.97670627e-01
-7.19377756e-01 9.81072962e-01 3.13950449e-01 -6.50251925e-01
-4.60858226e-01 8.46394539e-01 1.80743560e-01 -1.07824218e+00
1.99339539e-01 -1.05552959e+00 2.24351734e-01 -9.97642651e-02
5.45257665e-02 1.53479457e-01 4.48975742e-01 -1.31853789e-01
-4.02390182e-01 3.41347307e-01 -4.71148908e-01 2.85876632e-01
1.34727454e+00 3.32743347e-01 -5.20674944e-01 3.16442698e-01
1.35917866e+00 4.09526154e-02 -6.83921099e-01 -7.66767561e-01
3.70987177e-01 -1.45895928e-01 -2.97080308e-01 -1.35512364e+00
-1.39242697e+00 8.69628310e-01 -1.51158452e-01 1.02598049e-01
1.01809084e+00 9.59385335e-02 1.00128663e+00 8.56413782e-01
5.54293633e-01 -6.73701048e-01 -1.39905453e-01 7.14636803e-01
1.08361149e+00 -1.23427033e+00 -1.14160500e-01 -7.24086642e-01
-6.76959157e-01 7.35129952e-01 1.05273592e+00 2.09034458e-01
3.37774843e-01 4.86301631e-01 1.15019269e-01 -3.35393310e-01
-1.46629179e+00 -4.77960221e-02 1.96807712e-01 2.39288330e-01
6.93282366e-01 2.64219251e-02 6.72538728e-02 7.43308663e-01
-4.53215241e-01 -6.78876862e-02 1.33714259e-01 8.63708913e-01
3.09321463e-01 -1.35985839e+00 2.99423397e-01 5.82738101e-01
-4.00514275e-01 -4.81209785e-01 -3.40165675e-01 9.99118686e-01
-2.19228640e-01 6.81264281e-01 1.07047066e-01 -6.54041469e-01
1.02385890e+00 5.65734729e-02 3.89787078e-01 -5.42242944e-01
-1.05591881e+00 -3.30549449e-01 6.29598260e-01 -9.65191305e-01
1.46627545e-01 -4.86059397e-01 -1.30070436e+00 -6.20334387e-01
-5.48374392e-02 1.03021927e-01 2.46435359e-01 4.44624424e-01
8.55689526e-01 8.44219625e-01 2.29066864e-01 3.92513663e-01
-8.50311160e-01 -9.89471316e-01 1.73930284e-02 6.17445469e-01
-4.93402034e-02 -4.85776603e-01 2.81784713e-01 -2.18757868e-01] | [9.841968536376953, 7.867154121398926] |
06177f9b-b04b-4699-9500-164bee10afb7 | empirical-analysis-of-ai-based-energy | 2212.09154 | null | https://arxiv.org/abs/2212.09154v1 | https://arxiv.org/pdf/2212.09154v1.pdf | Empirical Analysis of AI-based Energy Management in Electric Vehicles: A Case Study on Reinforcement Learning | Reinforcement learning-based (RL-based) energy management strategy (EMS) is considered a promising solution for the energy management of electric vehicles with multiple power sources. It has been shown to outperform conventional methods in energy management problems regarding energy-saving and real-time performance. However, previous studies have not systematically examined the essential elements of RL-based EMS. This paper presents an empirical analysis of RL-based EMS in a Plug-in Hybrid Electric Vehicle (PHEV) and Fuel Cell Electric Vehicle (FCEV). The empirical analysis is developed in four aspects: algorithm, perception and decision granularity, hyperparameters, and reward function. The results show that the Off-policy algorithm effectively develops a more fuel-efficient solution within the complete driving cycle compared with other algorithms. Improving the perception and decision granularity does not produce a more desirable energy-saving solution but better balances battery power and fuel consumption. The equivalent energy optimization objective based on the instantaneous state of charge (SOC) variation is parameter sensitive and can help RL-EMSs to achieve more efficient energy-cost strategies. | ['Yuanjian Zhang', 'Jingjing Jiang', 'Quan Zhou', 'Dezong Zhao', 'Zhuoran Hou', 'Jihao Li', 'Yang Lin', 'Jincheng Hu'] | 2022-12-18 | null | null | null | null | ['energy-management'] | ['time-series'] | [-3.97983015e-01 3.52218114e-02 -5.74554741e-01 -3.62796560e-02
-4.90546614e-01 -4.69454885e-01 4.42220628e-01 3.54931712e-01
-3.22778553e-01 1.18210518e+00 -3.87087435e-01 -3.06963980e-01
-5.19432664e-01 -9.78211939e-01 -3.88796479e-01 -1.18026936e+00
8.35173484e-03 2.64251173e-01 -1.86091498e-03 -1.24899194e-01
2.85140872e-01 6.40991688e-01 -1.82833076e+00 -5.61514914e-01
1.00908267e+00 9.63606179e-01 6.15207255e-01 4.88290042e-01
3.37180912e-01 4.09508139e-01 -5.15404642e-01 3.11582744e-01
5.47982603e-02 -6.28510237e-01 -3.25327545e-01 -3.60999703e-01
-8.84269059e-01 -1.22362070e-01 2.40956023e-01 9.18100655e-01
7.21840739e-01 5.02935410e-01 7.21057177e-01 -2.04900622e+00
2.50154674e-01 2.74232507e-01 -2.59418070e-01 2.60478944e-01
-2.51868993e-01 4.60438013e-01 3.03130507e-01 -1.82674527e-01
1.15249790e-01 9.77713287e-01 7.30371997e-02 5.78787148e-01
-9.32295501e-01 -6.26259089e-01 -2.10836306e-01 8.20986450e-01
-1.33487535e+00 -2.35501856e-01 8.82955074e-01 -1.22482069e-02
1.41846311e+00 3.96730423e-01 1.36287057e+00 2.42436767e-01
6.86884582e-01 6.65041327e-01 1.33101118e+00 -5.23083568e-01
8.79626870e-01 4.69551921e-01 -1.70042768e-01 1.52973637e-01
6.37601495e-01 3.98292661e-01 4.34392923e-03 6.38943836e-02
2.32713819e-02 -6.15901709e-01 1.41766801e-01 -6.33264124e-01
-3.87354106e-01 9.37163591e-01 1.11414947e-01 4.22936380e-01
-6.77559197e-01 3.41540486e-01 3.35501254e-01 4.98851724e-02
-7.01694340e-02 5.87754369e-01 -4.34783161e-01 -4.28712577e-01
-8.53406847e-01 1.98191673e-01 6.28355682e-01 6.23047709e-01
7.00722456e-01 6.25612557e-01 -2.89935321e-01 6.00377381e-01
4.49039400e-01 8.02254915e-01 4.24475223e-01 -1.27975631e+00
-1.32277146e-01 6.22161567e-01 5.65338075e-01 -3.21803540e-01
-7.05923796e-01 -1.75174326e-01 -4.72172797e-01 7.54272461e-01
-2.52106041e-01 -6.92210615e-01 -3.90957326e-01 1.52948141e+00
3.98298532e-01 -4.62044865e-01 5.02999723e-01 5.38244188e-01
2.68413484e-01 1.08481205e+00 5.11620224e-01 -8.84188771e-01
1.27598429e+00 -7.97478497e-01 -1.36265194e+00 -1.17922202e-01
5.39466441e-01 -1.17368497e-01 3.06926936e-01 -2.78537385e-02
-1.46556151e+00 -3.98280650e-01 -1.60040879e+00 5.32827258e-01
-8.56103837e-01 1.06894448e-01 1.13697618e-01 7.75639415e-01
-1.17160404e+00 4.07159746e-01 -7.08824694e-01 -2.81973571e-01
3.72366272e-02 4.81013864e-01 4.21379894e-01 5.42203605e-01
-1.35039139e+00 1.63011062e+00 6.94402039e-01 -4.45024706e-02
-7.03133345e-01 -6.13239586e-01 -7.89858758e-01 9.99458954e-02
4.81165200e-01 -3.98959279e-01 1.38689029e+00 -7.55712152e-01
-2.06524396e+00 2.76499074e-02 -8.58325288e-02 -6.10693038e-01
3.91571313e-01 3.24191421e-01 -5.44519961e-01 1.57133609e-01
-3.55587959e-01 6.82831526e-01 4.86156642e-01 -1.33507049e+00
-1.02415943e+00 6.82346225e-02 -4.41418320e-01 4.77722019e-01
-8.57448950e-02 -3.90849918e-01 4.96127963e-01 4.07925844e-02
-8.51166010e-01 -9.15991485e-01 -1.16097718e-01 -6.11365259e-01
1.59532577e-01 -9.92994666e-01 1.04592144e+00 -3.40237886e-01
1.16921461e+00 -1.77858222e+00 7.24817216e-02 2.54158676e-01
-5.23872554e-01 4.38923985e-01 1.84452906e-01 6.14528656e-01
2.95958638e-01 -7.33073652e-02 1.11761011e-01 1.52580753e-01
3.58908027e-01 4.74576443e-01 3.14396888e-01 2.49574080e-01
7.37576857e-02 9.89373863e-01 -9.43279564e-01 -3.50856423e-01
8.35071564e-01 3.32255781e-01 -5.65829091e-02 3.06290030e-01
-3.26754332e-01 -1.34595767e-01 -4.75370616e-01 3.41947705e-01
6.15981579e-01 2.56347150e-01 3.30737710e-01 -3.36530119e-01
-6.05790377e-01 -4.95260686e-01 -9.85539496e-01 1.01484430e+00
-6.28979266e-01 7.85091758e-01 3.22384030e-01 -1.22651923e+00
7.83368051e-01 1.38109431e-01 8.93999815e-01 -1.45200598e+00
3.64032954e-01 1.82112277e-01 -3.09517067e-02 -5.73236108e-01
6.23205543e-01 -1.17009416e-01 1.01294719e-01 2.92096492e-02
-7.01448470e-02 -2.24022448e-01 5.15196681e-01 -4.57852751e-01
7.94358969e-01 1.90244675e-01 4.47151065e-01 -8.43376696e-01
7.31711328e-01 1.78880408e-01 7.62755930e-01 7.50916451e-02
-3.70553076e-01 -9.42433894e-01 2.82077998e-01 2.83339411e-01
-1.00220132e+00 -7.87272871e-01 -1.40233815e-01 8.11951339e-01
7.56082296e-01 4.61283624e-02 -8.06212664e-01 -1.68139726e-01
9.44564864e-02 1.77270675e+00 -9.92495567e-02 -5.57326615e-01
-4.67192888e-01 -9.34952557e-01 -1.34514794e-01 4.64964062e-01
6.41396582e-01 -7.74222612e-01 -1.63095522e+00 4.17133152e-01
4.53186259e-02 -9.39008415e-01 1.71001758e-02 5.06995678e-01
-6.23762369e-01 -9.11440134e-01 -4.18700427e-01 -4.31099445e-01
5.90384364e-01 5.45219369e-02 8.78277600e-01 -1.48602590e-01
-3.01262853e-03 6.23021543e-01 -1.17454708e-01 -9.76715028e-01
-6.85172737e-01 -2.04266757e-01 6.18623458e-02 -3.76471847e-01
5.55343568e-01 -4.55405451e-02 -8.23018014e-01 3.97060573e-01
-5.02094030e-01 2.07550395e-02 6.26146019e-01 5.70531845e-01
7.77345538e-01 9.32437241e-01 1.23288047e+00 4.79396991e-02
8.58798444e-01 -5.61905622e-01 -1.09953129e+00 3.84057790e-01
-1.43747246e+00 3.30226481e-01 5.81565976e-01 -1.14053398e-01
-1.08753037e+00 1.80377923e-02 -8.74259397e-02 -2.42181003e-01
-6.35811463e-02 -1.93390489e-01 -3.72296333e-01 -1.30105197e-01
-2.60236502e-01 4.55819547e-01 3.35850209e-01 -9.69329178e-02
-7.77769610e-02 3.78489822e-01 3.58848363e-01 -4.18058127e-01
6.00750327e-01 -2.36553162e-01 6.27041698e-01 -7.03277707e-01
2.54257679e-01 -2.02277556e-01 -9.33990106e-02 -8.06626379e-01
9.02753472e-01 -7.10035980e-01 -1.25483239e+00 3.82280976e-01
-5.62907100e-01 -6.59489810e-01 -4.25289333e-01 4.23704356e-01
-8.63345683e-01 -1.96007147e-01 1.34983093e-01 -1.32522333e+00
-4.92861301e-01 -1.24747372e+00 3.64880502e-01 9.04070437e-01
2.28648260e-02 -1.00411558e+00 1.56694114e-01 -1.17565721e-01
6.76655650e-01 5.65032840e-01 1.07624006e+00 -1.33355990e-01
-4.02771503e-01 2.17263043e-01 2.20161274e-01 4.07460421e-01
-1.40699983e-01 -5.08169606e-02 -7.00926542e-01 -7.56408572e-01
-1.77395329e-01 1.28570637e-02 2.16563389e-01 6.59065783e-01
9.59737062e-01 -3.59169990e-01 -5.37474751e-01 1.08119078e-01
2.21848607e+00 1.23338604e+00 5.43862581e-01 7.20919311e-01
-2.03863010e-01 5.13540447e-01 1.00622928e+00 5.79797506e-01
6.18534744e-01 6.87330902e-01 7.63883650e-01 -1.31009519e-01
1.53868735e-01 -3.19894147e-03 4.79994804e-01 3.08349758e-01
-3.21158245e-02 -5.31069756e-01 -4.54591244e-01 4.19915587e-01
-1.77015686e+00 -8.58147919e-01 1.55418098e-01 2.03049755e+00
4.86435443e-01 3.14333215e-02 3.67066771e-01 4.02718514e-01
5.67739546e-01 -2.44607523e-01 -9.92759287e-01 -1.05248487e+00
-1.69659946e-02 -1.65322334e-01 1.00524652e+00 2.25257188e-01
-3.19527268e-01 2.07126275e-01 6.36934090e+00 1.06663978e+00
-8.87870133e-01 1.46043047e-01 4.96324003e-01 -3.34786981e-01
-3.16747904e-01 -2.81399816e-01 -6.95531368e-01 8.58978033e-01
1.59841979e+00 -9.57771361e-01 7.31678843e-01 6.50980890e-01
8.00788879e-01 -9.83043253e-01 -1.10954905e+00 8.93507898e-01
-3.12892944e-01 -1.20179117e+00 -4.58399892e-01 2.03870565e-01
7.02141464e-01 -2.79672861e-01 -5.41515052e-01 5.67759812e-01
4.41221744e-02 -7.21109629e-01 8.86925995e-01 7.51395285e-01
3.75744462e-01 -1.58795559e+00 1.01308632e+00 4.76600200e-01
-1.24751675e+00 -6.57350898e-01 -1.10980034e-01 2.26116836e-01
3.19684356e-01 1.89281434e-01 -2.78086126e-01 6.40656650e-01
7.37307370e-01 8.85690153e-02 -3.80661368e-01 8.50375116e-01
2.65840530e-01 2.98669547e-01 -3.83562118e-01 -9.40738559e-01
1.80146560e-01 -4.78756279e-01 5.46071589e-01 9.42971528e-01
3.91349286e-01 2.36922473e-01 -1.03496037e-01 7.69974768e-01
4.88248467e-01 -7.99970478e-02 -3.95668179e-01 -2.54601594e-02
8.14923167e-01 1.51256537e+00 -8.74222338e-01 -3.24016780e-01
-1.17056265e-01 2.32562631e-01 -4.27104324e-01 3.46407443e-01
-1.02196777e+00 -5.56300223e-01 7.03318119e-01 6.35835305e-02
2.10711017e-01 7.49803260e-02 -1.13760620e-01 -1.14590071e-01
-2.37969115e-01 -6.89213797e-02 1.58128366e-01 -7.48260796e-01
-7.52268374e-01 7.94226006e-02 7.72906005e-01 -9.37549710e-01
-5.89228332e-01 -5.37344396e-01 -7.28704572e-01 5.36240518e-01
-2.04264379e+00 -4.74397480e-01 -2.50228465e-01 1.92817092e-01
7.22246647e-01 -3.25049073e-01 4.58305597e-01 2.96161294e-01
-9.71917033e-01 3.14424068e-01 8.09297204e-01 -7.58813381e-01
8.98443684e-02 -1.28195083e+00 -7.91633964e-01 6.53225064e-01
-1.26076114e+00 -5.17853975e-01 1.09194934e+00 -2.64625192e-01
-2.10200596e+00 -1.01740205e+00 2.41595477e-01 2.84243554e-01
2.86660254e-01 1.34734735e-01 -4.81870443e-01 -1.67085230e-01
1.15926600e+00 -5.41067183e-01 2.80689657e-01 -8.71465802e-01
1.06549180e+00 -5.67485571e-01 -1.62211466e+00 4.71897900e-01
3.51152062e-01 9.10147727e-02 -5.06597795e-02 -1.23313256e-01
1.67137846e-01 -1.54294118e-01 -9.04726267e-01 7.28396535e-01
3.96294713e-01 -6.22721136e-01 5.48144042e-01 2.05987450e-02
-2.82953918e-01 -4.06353265e-01 9.62892994e-02 -1.82677209e+00
-3.41183662e-01 -6.24637604e-01 -6.68649733e-01 1.36443615e+00
1.59072772e-01 -9.58046973e-01 3.39926690e-01 4.42425698e-01
-2.05118299e-01 -1.17508709e+00 -1.25921261e+00 -1.10631382e+00
1.57176852e-01 -7.25450888e-02 7.86813617e-01 4.16013718e-01
1.42412633e-01 1.32579934e-02 2.38478974e-01 1.79727241e-01
8.97912920e-01 -1.07297324e-01 2.78863817e-01 -7.75013328e-01
3.67545813e-01 -8.73040378e-01 -8.64001736e-02 -3.27483527e-02
3.50567430e-01 -5.93096972e-01 3.77202928e-01 -2.07011700e+00
3.05503104e-02 -2.01717362e-01 -5.76400816e-01 4.29821372e-01
1.18198693e-01 -3.40911686e-01 2.18075350e-01 -3.13177943e-01
-4.26946372e-01 9.00541723e-01 9.61669981e-01 -1.51045501e-01
-2.82899350e-01 -2.02040765e-02 -2.68176675e-01 4.14043069e-01
1.26365721e+00 -3.56999785e-01 -9.35709119e-01 3.43680620e-01
1.58853650e-01 3.64229470e-01 2.23606139e-01 -1.20461583e+00
2.32573137e-01 -6.25945389e-01 3.41946483e-01 -9.63276744e-01
4.06725928e-02 -1.22898710e+00 7.41713345e-01 9.88172591e-01
6.28730888e-03 2.52638310e-01 5.04632294e-01 5.06975532e-01
1.82832584e-01 -3.55238706e-01 1.15614724e+00 3.38653803e-01
-1.13488591e+00 -3.33221674e-01 -1.08855414e+00 -4.01131094e-01
2.02525520e+00 -4.67617840e-01 -3.17380279e-01 8.45610276e-02
-1.78723514e-01 9.79857147e-01 3.13029885e-01 5.20620584e-01
1.79465577e-01 -1.47936475e+00 -3.58862817e-01 2.66662799e-03
-2.76655704e-01 -5.79246461e-01 2.40645617e-01 6.26297414e-01
-2.83120036e-01 5.96441031e-01 -6.57521188e-01 -4.47676152e-01
-1.13210237e+00 4.42411065e-01 7.50821710e-01 -3.12569678e-01
-2.49540433e-01 -1.52513534e-01 -5.29438734e-01 2.28871211e-01
7.41166100e-02 -1.90533698e-01 -4.10506576e-01 3.19986492e-01
9.68840942e-02 1.28131771e+00 -1.46080060e-02 -3.86058211e-01
-4.61170256e-01 4.91602510e-01 8.93269122e-01 2.19534826e-03
1.14241779e+00 -3.67487669e-01 1.66247934e-01 4.27958935e-01
9.58074093e-01 -5.52742720e-01 -1.47239888e+00 5.35986245e-01
1.86764579e-02 9.62000713e-02 7.68422663e-01 -1.14266860e+00
-1.09796846e+00 2.22903237e-01 1.17032492e+00 6.15413666e-01
1.59806371e+00 -3.88926923e-01 5.64857900e-01 2.43106142e-01
4.26499337e-01 -2.15333152e+00 -2.03214899e-01 7.74375424e-02
7.36656606e-01 -9.19763327e-01 7.55184144e-02 3.11267078e-01
-5.28456926e-01 1.18076181e+00 8.87738526e-01 2.09203824e-01
6.52580142e-01 5.56161940e-01 -4.95175481e-01 1.70234650e-01
-9.11999524e-01 -1.25774443e-01 -5.48730753e-02 5.62160432e-01
-2.64153123e-01 4.21620876e-01 -7.11687565e-01 4.62433040e-01
2.24679261e-01 7.23035680e-03 4.20580983e-01 1.17329156e+00
-7.55490065e-01 -8.97971809e-01 -3.70908141e-01 3.31825852e-01
-3.58924195e-02 8.30639422e-01 3.86318773e-01 1.04173529e+00
3.60899657e-01 1.22026598e+00 2.35660046e-01 -4.59300354e-02
5.88648856e-01 9.51802060e-02 4.02755141e-01 2.40787476e-01
-2.06064925e-01 -6.91927820e-02 2.38976344e-01 -5.22855103e-01
-7.13130593e-01 -6.73550189e-01 -1.85738480e+00 -4.69986647e-01
-4.00140256e-01 6.77472293e-01 1.32350683e+00 1.10046089e+00
4.84200150e-01 1.06519639e+00 1.19264436e+00 -8.92458439e-01
-3.70943815e-01 -5.37946701e-01 -7.49255240e-01 -3.86758864e-01
2.89522439e-01 -9.07403946e-01 -5.59934676e-01 -4.91719007e-01] | [5.549935340881348, 2.2465460300445557] |
726b2782-6438-4d96-906a-1c5614757659 | diagnostic-classification-of-lung-nodules | 1803.07192 | null | http://arxiv.org/abs/1803.07192v1 | http://arxiv.org/pdf/1803.07192v1.pdf | Diagnostic Classification Of Lung Nodules Using 3D Neural Networks | Lung cancer is the leading cause of cancer-related death worldwide. Early
diagnosis of pulmonary nodules in Computed Tomography (CT) chest scans provides
an opportunity for designing effective treatment and making financial and care
plans. In this paper, we consider the problem of diagnostic classification
between benign and malignant lung nodules in CT images, which aims to learn a
direct mapping from 3D images to class labels. To achieve this goal, four
two-pathway Convolutional Neural Networks (CNN) are proposed, including a basic
3D CNN, a novel multi-output network, a 3D DenseNet, and an augmented 3D
DenseNet with multi-outputs. These four networks are evaluated on the public
LIDC-IDRI dataset and outperform most existing methods. In particular, the 3D
multi-output DenseNet (MoDenseNet) achieves the state-of-the-art classification
accuracy on the task of end-to-end lung nodule diagnosis. In addition, the
networks pretrained on the LIDC-IDRI dataset can be further extended to handle
smaller datasets using transfer learning. This is demonstrated on our dataset
with encouraging prediction accuracy in lung nodule classification. | ['Zhongjie Lu', 'Yi Hong', 'Raunak Dey'] | 2018-03-19 | null | null | null | null | ['lung-nodule-classification'] | ['medical'] | [ 9.94433532e-04 2.58953273e-01 -4.73437458e-01 -3.22532684e-01
-8.86465490e-01 -2.99067706e-01 4.61321682e-01 -5.03616214e-01
-4.12062496e-01 3.28186333e-01 1.78625360e-01 -7.19616234e-01
-1.65236786e-01 -7.48931885e-01 -3.12227964e-01 -7.04625905e-01
4.89504747e-02 1.13630319e+00 5.08809566e-01 4.30322856e-01
-5.58189273e-01 8.15493882e-01 -9.13572490e-01 5.29994130e-01
4.47310477e-01 1.30648172e+00 3.10305357e-01 6.60718501e-01
3.36112715e-02 1.02376068e+00 1.58688605e-01 -1.25145078e-01
3.03149730e-01 -5.09443164e-01 -1.17415178e+00 8.17414001e-02
3.50369036e-01 -6.34458840e-01 -5.12007356e-01 6.87541306e-01
6.91138327e-01 -4.98864830e-01 1.03196263e+00 -8.68597150e-01
-5.44276834e-01 3.40673566e-01 -2.79553145e-01 3.78175616e-01
-3.50035757e-01 2.28794768e-01 7.64730453e-01 -1.02578938e+00
2.73667574e-01 8.75501633e-01 8.43743503e-01 8.60075057e-01
-4.76399839e-01 -6.16828978e-01 -4.31943119e-01 3.05433869e-02
-1.19386303e+00 3.85870874e-01 9.20410454e-02 -6.29316211e-01
7.18901575e-01 2.92286783e-01 7.06846893e-01 1.04276168e+00
2.25077018e-01 6.63239956e-01 8.30840886e-01 4.45160381e-02
-2.35634491e-01 -5.34125837e-03 -3.30084920e-01 1.28869390e+00
4.52653170e-01 3.41239870e-01 3.14593017e-01 -1.65267155e-01
1.10078776e+00 5.75871706e-01 -3.47493023e-01 -3.59846175e-01
-1.54774523e+00 9.49337602e-01 1.21242428e+00 5.51665664e-01
-3.10366929e-01 3.36365789e-01 3.31833363e-01 -6.20315596e-02
4.68746364e-01 2.68598825e-01 -3.40697497e-01 4.91616994e-01
-6.47948265e-01 -3.39477719e-03 8.69421482e-01 5.59533179e-01
1.25223532e-01 -2.93281019e-01 -8.76682401e-01 7.04371750e-01
3.04049969e-01 4.65980142e-01 1.04013515e+00 -4.54126656e-01
3.20945829e-01 5.92210650e-01 -2.79985219e-01 -1.60791203e-01
-8.53044689e-01 -7.89440036e-01 -1.32398713e+00 3.28868516e-02
2.08495826e-01 6.82585016e-02 -1.31805325e+00 1.10476243e+00
2.68267393e-01 2.85710514e-01 -9.98210460e-02 1.03791010e+00
1.49885488e+00 1.22050114e-01 2.65680207e-03 2.96799839e-01
1.52869308e+00 -1.39813650e+00 -2.80158430e-01 -1.07006386e-01
9.18254137e-01 -3.42666149e-01 8.52045298e-01 -3.22080135e-01
-8.42582464e-01 -5.17927349e-01 -6.03526592e-01 -1.99472830e-01
-9.36849341e-02 6.85987473e-01 3.46930712e-01 5.47557414e-01
-1.14054573e+00 4.17258680e-01 -1.05495906e+00 -5.91143548e-01
7.82487035e-01 6.25365257e-01 -2.43344024e-01 -3.48955750e-01
-9.57988620e-01 9.63077426e-01 5.14599085e-01 -1.61002487e-01
-1.33918166e+00 -9.01459277e-01 -3.73709172e-01 1.77240640e-01
3.78913522e-01 -1.37204254e+00 1.79364228e+00 -4.99784410e-01
-1.19112313e+00 1.27589643e+00 4.11460906e-01 -4.27593440e-01
7.35065401e-01 3.12312722e-01 -4.76377904e-02 2.48780519e-01
3.57603937e-01 9.35531855e-01 6.57164216e-01 -6.26909614e-01
-8.08507442e-01 -1.94364235e-01 -3.83368820e-01 2.73061216e-01
-2.98646063e-01 -3.34921509e-01 -5.39749920e-01 -5.44739246e-01
1.54199436e-01 -1.19790041e+00 -4.19604540e-01 4.39917833e-01
-6.25340164e-01 -4.45076287e-01 1.06906354e+00 -4.78473425e-01
6.80080056e-01 -1.84959447e+00 7.70701049e-03 1.09992018e-02
8.04575026e-01 2.83286870e-01 6.71027899e-02 -3.22984427e-01
-2.73193717e-01 4.21947539e-01 -1.16962992e-01 3.55230644e-02
-2.01842830e-01 2.10890442e-01 4.96684879e-01 3.98588181e-01
2.89020449e-01 1.55149400e+00 -7.13870466e-01 -7.31090963e-01
1.45280585e-01 3.12348068e-01 -4.36862051e-01 5.25358796e-01
-1.13402382e-01 6.09116852e-01 -7.19255209e-01 8.05552721e-01
2.71152377e-01 -1.15961206e+00 -1.57623351e-01 -1.94450855e-01
2.10997015e-01 2.83885241e-01 -1.79359302e-01 1.26237774e+00
-4.35347587e-01 2.60808200e-01 -1.61277369e-01 -6.58978879e-01
5.10738134e-01 8.40591073e-01 8.30751359e-01 -4.03762430e-01
2.64552832e-01 5.48756897e-01 4.54682291e-01 -6.76475227e-01
-4.10944432e-01 -3.77736866e-01 9.42256153e-02 6.22432947e-01
-1.26185760e-01 -4.54740942e-01 -1.44355476e-01 -1.13732167e-01
1.54846191e+00 -6.73812449e-01 5.00781536e-01 -2.56730020e-01
5.28134286e-01 1.07649438e-01 2.06350714e-01 5.78828752e-01
-2.50198901e-01 9.91288006e-01 4.23772693e-01 -5.78615546e-01
-9.26624000e-01 -1.02387035e+00 -2.53054261e-01 5.53969145e-01
-3.20641041e-01 2.32284218e-01 -3.60305458e-01 -1.44453275e+00
2.39378601e-01 1.68152720e-01 -9.35733438e-01 -1.32519081e-01
-6.69214547e-01 -6.71618462e-01 6.58679664e-01 9.83954132e-01
9.04351175e-01 -1.08137500e+00 -4.56918091e-01 -1.32913232e-01
-2.22954988e-01 -1.09657812e+00 -5.81014931e-01 4.41268981e-01
-1.09988391e+00 -1.45680702e+00 -1.17971265e+00 -1.19768178e+00
7.21558571e-01 4.15645361e-01 1.28000367e+00 2.84777164e-01
-5.47311842e-01 3.70959997e-01 -5.98410480e-02 -3.31346184e-01
-6.38490736e-01 5.66865981e-01 -3.79853189e-01 -5.33509076e-01
1.08969487e-01 -4.54045795e-02 -7.50654101e-01 5.07830739e-01
-8.21195424e-01 2.08323628e-01 1.34799945e+00 9.50260758e-01
8.89555573e-01 3.51960049e-03 3.99202764e-01 -1.16433966e+00
3.65274012e-01 -8.26875806e-01 -1.20979458e-01 1.27976790e-01
-3.47298771e-01 -1.81226358e-01 6.00771308e-01 -2.13193372e-01
-8.49726558e-01 2.87748009e-01 -4.34564263e-01 -7.07357347e-01
-1.06508300e-01 4.47867543e-01 2.36685812e-01 -3.07339728e-01
7.43214369e-01 8.01208336e-03 3.00787035e-02 -2.94865340e-01
-7.43195415e-03 6.40440166e-01 4.30038422e-01 1.93930313e-01
1.04518652e+00 5.10934711e-01 3.37091178e-01 -3.84356648e-01
-1.39431906e+00 -7.44850457e-01 -9.12346244e-01 -8.12906623e-02
1.37794471e+00 -1.07436347e+00 -1.00310653e-01 1.86801165e-01
-7.70558536e-01 -3.98634404e-01 -4.68001902e-01 7.55176008e-01
-4.89238560e-01 -9.88224819e-02 -7.55293071e-01 1.73866272e-01
-5.89861989e-01 -1.20971286e+00 1.12870312e+00 8.24067146e-02
8.72789621e-02 -1.16046476e+00 9.59121808e-02 3.47983062e-01
6.45560503e-01 4.59312908e-02 1.28082275e+00 -1.09637702e+00
-6.17877543e-01 -4.42228049e-01 -7.36079156e-01 4.04640019e-01
3.88757348e-01 -3.68945807e-01 -8.20148528e-01 -4.62411195e-01
5.92442835e-03 -6.73958540e-01 1.24311864e+00 5.72520852e-01
1.72799170e+00 -6.53233007e-02 -8.98937941e-01 9.04874325e-01
1.15021348e+00 3.10577860e-04 2.25270450e-01 7.37507129e-04
1.03923047e+00 -1.18893042e-01 1.58063158e-01 6.84939697e-02
7.62113333e-02 1.94984093e-01 1.00239062e+00 -5.91443360e-01
-6.12402797e-01 -3.19781750e-02 -3.59147161e-01 5.88770390e-01
-1.25547960e-01 -4.03228164e-01 -1.18382096e+00 3.93573701e-01
-1.39407945e+00 -6.10223889e-01 -2.14838654e-01 1.74225593e+00
4.51793402e-01 -5.16618490e-02 -3.88197117e-02 -1.68753877e-01
5.85399270e-01 1.10680036e-01 -7.90390491e-01 3.38888943e-01
4.26727086e-01 3.51522505e-01 6.59036636e-01 -2.82504588e-01
-1.64901769e+00 3.20860952e-01 6.23416519e+00 7.38306642e-01
-1.26873338e+00 5.46297848e-01 9.67519104e-01 -1.97647233e-02
1.14403166e-01 -7.22318470e-01 -6.20023787e-01 -3.80523391e-02
6.74938202e-01 9.66094509e-02 -2.03733355e-01 1.14254534e+00
-1.05309226e-02 3.48102480e-01 -1.39291501e+00 8.26633573e-01
-7.86609203e-02 -1.42472732e+00 7.46693760e-02 3.41070145e-01
8.65971327e-01 6.60607100e-01 3.19735616e-01 4.95883971e-01
2.16948614e-01 -1.49006820e+00 1.35467350e-02 2.67512023e-01
1.32821858e+00 -4.51850146e-01 1.10070515e+00 5.42736471e-01
-1.19637048e+00 -8.83827135e-02 -3.89929205e-01 5.68544090e-01
-2.86009550e-01 6.02377295e-01 -1.64593589e+00 3.61112148e-01
8.89499426e-01 9.43224907e-01 -7.78825283e-01 1.31950569e+00
-1.39552459e-01 7.17134237e-01 -1.37074098e-01 -1.74765572e-01
5.39433837e-01 3.81554216e-01 6.56500682e-02 1.00829864e+00
6.89703405e-01 2.06622794e-01 4.24511701e-01 9.94877756e-01
-6.07101023e-01 -2.08474562e-01 -5.68158507e-01 -8.96212831e-02
3.08410097e-02 1.57489061e+00 -8.96476805e-01 -2.82900184e-01
-5.09920597e-01 8.43729913e-01 1.98123977e-01 -1.77241102e-01
-9.96383250e-01 2.50022769e-01 1.52108729e-01 4.35777277e-01
4.40904021e-01 4.67742115e-01 7.72456303e-02 -8.35306287e-01
-4.17859733e-01 -6.02599561e-01 5.52647531e-01 -6.43338144e-01
-1.65600359e+00 7.83944845e-01 -1.68922588e-01 -1.47645736e+00
-3.75019103e-01 -1.01829410e+00 -8.73230934e-01 7.43227959e-01
-1.61088216e+00 -1.23614299e+00 -7.93411374e-01 6.86303675e-01
4.28052902e-01 -2.75915891e-01 7.70522296e-01 2.26073891e-01
-2.80318350e-01 2.95893669e-01 6.31357431e-02 5.37469745e-01
5.83000839e-01 -1.40581822e+00 3.58213484e-01 1.44613683e-01
-2.34158203e-01 -5.18303931e-01 -4.07049149e-01 -5.83939433e-01
-9.69315410e-01 -2.11073828e+00 6.77160084e-01 -4.49337721e-01
4.19637173e-01 1.85815945e-01 -8.69704843e-01 8.33274424e-01
-1.50077328e-01 7.24600792e-01 6.03077888e-01 -6.01748526e-01
-1.16590776e-01 2.28161201e-01 -1.20390737e+00 1.47478044e-01
1.07462430e+00 -3.34870726e-01 -1.26219809e-01 1.00345099e+00
8.33392918e-01 -6.06138766e-01 -9.76901114e-01 8.86129379e-01
2.55074590e-01 -9.69660461e-01 1.19822001e+00 -5.76381922e-01
5.85070968e-01 1.03732511e-01 2.39607647e-01 -1.11308062e+00
-9.94287193e-01 3.68142009e-01 1.82398617e-01 2.14748099e-01
6.06537640e-01 -4.06386703e-01 1.19577098e+00 1.13749661e-01
-3.78457516e-01 -1.23543954e+00 -8.72038960e-01 -6.62213564e-01
4.50051904e-01 -8.30941275e-02 1.79405108e-01 6.60443664e-01
-8.16315234e-01 2.17749730e-01 1.82279199e-01 -1.62287708e-02
3.32668036e-01 6.03136886e-03 1.66585222e-01 -1.39016199e+00
-3.45536590e-01 -6.97693527e-01 -2.21075013e-01 -9.73200202e-01
-6.12625778e-02 -1.54102540e+00 -2.44481601e-02 -1.90085089e+00
7.78155684e-01 -4.55490440e-01 -2.80893296e-01 5.23935735e-01
-8.51830468e-02 5.20191908e-01 -5.83145209e-03 5.78887820e-01
-5.80237567e-01 3.15121800e-01 1.96311891e+00 -3.98909271e-01
2.99273401e-01 7.61548579e-01 -3.66375476e-01 7.77464211e-01
8.03807318e-01 -7.45267391e-01 -2.92172611e-01 -2.52428532e-01
-2.36433908e-01 2.68423110e-01 6.97355330e-01 -1.23602486e+00
1.09362394e-01 7.95835331e-02 7.91160941e-01 -1.14778566e+00
1.40102893e-01 -8.99413824e-01 3.75687107e-02 1.24911940e+00
-3.34133923e-01 -3.14976484e-01 -1.03017196e-01 6.30379260e-01
-2.13831082e-01 -3.22674781e-01 1.01216733e+00 -6.71501279e-01
-2.46189818e-01 1.08018374e+00 -4.50253844e-01 8.93669855e-03
1.21267927e+00 9.72767100e-02 -2.05119774e-01 -9.46340859e-02
-7.21906781e-01 2.59679139e-01 7.53053976e-03 2.68033110e-02
7.28329539e-01 -1.66739297e+00 -8.80331337e-01 1.85857996e-01
8.88650790e-02 6.39049590e-01 -1.92789454e-02 1.23451829e+00
-8.45053673e-01 9.70256209e-01 8.14649090e-02 -9.84526396e-01
-1.24612868e+00 3.24657202e-01 1.20622647e+00 -1.01769567e+00
-5.64257622e-01 1.15147364e+00 6.11527681e-01 -7.98400998e-01
3.34398746e-01 -9.10145342e-01 -1.56559095e-01 -4.75457847e-01
2.58075800e-02 7.71984756e-02 3.90284210e-01 -1.73062608e-01
-2.94835001e-01 3.45281482e-01 -2.40955368e-01 4.67476636e-01
1.09600139e+00 4.63288397e-01 -1.10555336e-01 7.52250180e-02
1.57583261e+00 -7.17179418e-01 -8.17618310e-01 -3.81696522e-01
-7.07778409e-02 -5.12511171e-02 2.36433417e-01 -9.58360493e-01
-1.58658457e+00 9.13909137e-01 9.50742245e-01 1.07895590e-01
9.98484313e-01 5.54211318e-01 8.87017071e-01 7.83640504e-01
-2.22106695e-01 -1.59681663e-01 6.96263552e-01 5.20280659e-01
7.78191745e-01 -1.70512140e+00 -6.81006163e-02 -4.83145982e-01
-5.33951938e-01 1.22590673e+00 9.30853128e-01 -7.88567886e-02
1.11831069e+00 1.26626790e-01 2.28674505e-02 -4.34489846e-01
-8.43829811e-01 -5.89885712e-01 6.12769067e-01 4.29351538e-01
6.50166988e-01 1.99108481e-01 2.17284605e-01 6.79901600e-01
1.15378343e-01 1.28519416e-01 1.01004474e-01 8.07713747e-01
-6.78965509e-01 -7.84139633e-01 -2.14605734e-01 1.17841768e+00
-5.68835676e-01 -5.24009205e-02 -7.15322614e-01 1.26338279e+00
1.19929686e-01 2.04257816e-01 1.99196652e-01 -1.70045868e-01
2.29534626e-01 -8.12342092e-02 2.84719795e-01 -1.05623031e+00
-8.08637321e-01 1.27286181e-01 -2.13654876e-01 -2.59995550e-01
-4.52941954e-01 -2.78363228e-01 -1.24076629e+00 2.46537738e-02
-3.41831088e-01 -5.37559129e-02 3.56419057e-01 7.41298854e-01
4.87001166e-02 1.09512794e+00 6.50379539e-01 -7.58574665e-01
-9.27364886e-01 -1.04407060e+00 -5.32533169e-01 4.64342833e-02
1.85998753e-01 -3.62658501e-01 -4.00961131e-01 -3.01179171e-01] | [15.38630485534668, -2.1398448944091797] |
22b1d669-37a1-4a9e-b7b9-f1c4a7385d49 | visually-verifiable-textual-entailment-a | null | null | https://aclanthology.org/W15-2801 | https://aclanthology.org/W15-2801.pdf | Visually-Verifiable Textual Entailment: A Challenge Task for Combining Language and Vision | null | ['Jayant Krishnamurthy'] | 2015-09-01 | null | null | null | ws-2015-9 | ['prepositional-phrase-attachment'] | ['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] | [-7.168886184692383, 3.719616651535034] |
eb856c8c-84d7-432d-be33-73663eea8bd4 | deep-domain-adaptation-for-polyphonic-melody | 2210.12532 | null | https://arxiv.org/abs/2210.12532v2 | https://arxiv.org/pdf/2210.12532v2.pdf | Deep domain adaptation for polyphonic melody extraction | Extraction of the predominant pitch from polyphonic audio is one of the fundamental tasks in the field of music information retrieval and computational musicology. To accomplish this task using machine learning, a large amount of labeled audio data is required to train the model that predicts the pitch contour. But a classical model pre-trained on data from one domain (source), e.g, songs of a particular singer or genre, may not perform comparatively well in extracting melody from other domains (target). The performance of such models can be boosted by adapting the model using some annotated data in the target domain. In this work, we study various adaptation techniques applied to machine learning models for polyphonic melody extraction. Experimental results show that meta-learning-based adaptation performs better than simple fine-tuning. In addition to this, we find that this method outperforms the existing state-of-the-art non-adaptive polyphonic melody extraction algorithms. | ['Vipul Arora', 'Kavya Ranjan Saxena'] | 2022-10-22 | null | null | null | null | ['melody-extraction', 'music-information-retrieval'] | ['music', 'music'] | [ 3.10136408e-01 -3.39150131e-01 -2.74983346e-01 -5.95178157e-02
-1.03577197e+00 -7.43848264e-01 1.81194544e-01 2.74364769e-01
-3.80874097e-01 5.27883828e-01 3.09289306e-01 3.16970259e-01
-1.92723945e-01 -4.56656784e-01 -3.19406986e-01 -6.09659851e-01
-3.13863643e-02 4.39980209e-01 3.88126463e-01 -4.34659719e-01
5.05244792e-01 1.38755322e-01 -1.70730603e+00 5.33999920e-01
2.81416714e-01 9.99861240e-01 2.82755464e-01 9.79317844e-01
-4.78721634e-02 4.15949255e-01 -8.47464323e-01 -1.67925190e-02
2.76987791e-01 -8.89975429e-01 -7.17855811e-01 -2.39215285e-01
2.03626961e-01 2.06010714e-01 1.73209324e-01 1.02119815e+00
7.46696293e-01 1.28543288e-01 5.93879640e-01 -7.50092328e-01
1.64100289e-01 1.10740769e+00 -2.65190274e-01 3.54381621e-01
3.74515802e-01 -2.60460287e-01 1.19527602e+00 -8.65160644e-01
3.88971269e-01 8.23993444e-01 8.12402606e-01 4.29178625e-01
-1.25523794e+00 -8.30837727e-01 -2.62176126e-01 1.94541633e-01
-1.46933651e+00 -5.06077468e-01 1.40372169e+00 -3.72427374e-01
5.18343508e-01 3.06045413e-01 8.37247431e-01 7.32069969e-01
-1.60813585e-01 6.87556326e-01 9.97086108e-01 -8.93737555e-01
2.47262478e-01 9.32088643e-02 -3.60005528e-01 2.12608725e-02
-4.18201804e-01 9.64630693e-02 -1.03921342e+00 -3.08230519e-01
6.99978888e-01 -7.06826866e-01 -1.99357405e-01 -6.10638075e-02
-1.07534301e+00 7.68680334e-01 -6.33008778e-02 7.32033670e-01
-4.31464791e-01 -7.78774470e-02 7.54445076e-01 5.18892884e-01
4.19166476e-01 1.11668634e+00 -7.73977101e-01 -6.32106304e-01
-1.52286780e+00 4.47477788e-01 8.98837388e-01 4.72810477e-01
4.88089770e-01 2.47807279e-01 -4.83489633e-02 1.23342359e+00
-3.15178856e-02 7.68075213e-02 7.26665974e-01 -8.87037098e-01
2.69720197e-01 2.77121395e-01 -2.48482246e-02 -6.09143436e-01
-4.14384305e-01 -6.37918949e-01 -3.50889623e-01 8.02353304e-03
6.08796954e-01 -1.64184257e-01 -5.29408693e-01 1.68260336e+00
3.07659805e-01 3.23816985e-01 -1.03401057e-01 8.36457849e-01
6.53840959e-01 8.26936245e-01 -5.20294718e-02 -6.65349543e-01
1.14590347e+00 -8.18486631e-01 -6.46902621e-01 -1.13620840e-01
-5.66250132e-03 -1.39909422e+00 1.17176282e+00 8.38370681e-01
-1.18259060e+00 -8.93774807e-01 -1.22229600e+00 2.72023231e-01
-4.44051437e-02 1.80875182e-01 3.55622679e-01 6.31001890e-01
-2.92711526e-01 8.04821014e-01 -4.32753175e-01 -4.15709764e-02
-9.19813886e-02 4.39895719e-01 3.86928134e-02 5.41314483e-01
-1.21228075e+00 4.33370650e-01 7.86992848e-01 -4.15502846e-01
-7.93892443e-01 -9.15634990e-01 -3.87080073e-01 3.82276662e-02
3.14198524e-01 -2.29339153e-01 1.68409216e+00 -1.31481874e+00
-1.75384617e+00 7.00214744e-01 2.37370491e-01 -6.05208933e-01
5.17829880e-02 -4.14102882e-01 -5.90710700e-01 2.74312675e-01
-1.07871816e-01 5.13591170e-01 1.36037195e+00 -9.51328039e-01
-8.93940568e-01 -1.11131445e-01 -2.38397181e-01 2.56985307e-01
-5.46333551e-01 2.55301327e-01 -3.64762843e-01 -1.12904501e+00
-3.34494226e-02 -1.08829629e+00 1.69843230e-02 -7.09184706e-01
-1.73658326e-01 -1.46878347e-01 7.18892395e-01 -6.24637067e-01
1.74362838e+00 -2.32934737e+00 2.28504062e-01 2.87837267e-01
-3.30628604e-01 3.64609927e-01 -1.38460184e-02 5.41706443e-01
-2.90203430e-02 -2.97960877e-01 -5.14782295e-02 4.59583439e-02
-9.20645893e-02 -4.85160649e-02 -3.82884711e-01 8.33987966e-02
-1.14094667e-01 4.09464180e-01 -8.41134310e-01 -5.95269084e-01
1.05433941e-01 4.07771885e-01 -6.92709684e-01 2.17683211e-01
-2.75635481e-01 7.53479540e-01 -5.53322099e-02 3.74261558e-01
1.14804335e-01 1.61435366e-01 1.48629367e-01 -2.79732108e-01
-9.38094705e-02 7.28914797e-01 -1.42172003e+00 1.94717765e+00
-5.87638497e-01 6.73625171e-01 -1.45011768e-01 -8.46168935e-01
1.11010206e+00 7.13884234e-01 8.58393490e-01 -3.62679482e-01
2.23634541e-02 5.84757447e-01 5.21564901e-01 -4.44277614e-01
6.41427755e-01 -5.02001464e-01 -3.43999624e-01 2.59416252e-01
4.57243204e-01 -3.14142376e-01 3.58551025e-01 -5.23531556e-01
7.93890774e-01 4.64826375e-02 6.49722993e-01 -1.15156047e-01
6.70616686e-01 1.32663295e-01 6.74599349e-01 4.69022155e-01
-9.04262532e-03 6.33484602e-01 -3.39616351e-02 -2.94350564e-01
-1.28173268e+00 -7.66039252e-01 -1.17985457e-01 1.58571577e+00
-3.88731509e-01 -9.51026022e-01 -8.66008878e-01 -1.97567001e-01
-2.57755965e-01 3.90476435e-01 -1.54991597e-01 -8.82866383e-02
-7.77074754e-01 -5.70346534e-01 7.64896333e-01 2.96998829e-01
2.20238283e-01 -1.35108960e+00 -6.90136790e-01 7.01705217e-01
-2.95569509e-01 -7.81886339e-01 -5.78103304e-01 3.68055463e-01
-9.12129104e-01 -8.33595157e-01 -7.08045304e-01 -8.82125080e-01
-2.24475220e-01 -2.55380809e-01 1.20717454e+00 -3.80943090e-01
-1.41973108e-01 2.60369241e-01 -6.82486117e-01 -8.57005358e-01
-6.97035313e-01 6.49304032e-01 1.72890529e-01 9.25489068e-02
4.81729746e-01 -8.66599977e-01 -4.91083950e-01 2.58814186e-01
-8.41876447e-01 -2.21274942e-01 4.32610840e-01 7.09421217e-01
7.87697375e-01 5.49609244e-01 1.01219833e+00 -8.03873479e-01
8.96968901e-01 -4.15674560e-02 -3.11260372e-01 -1.76295370e-01
-3.35525274e-01 -2.54528165e-01 8.09222877e-01 -1.02869284e+00
-7.05824018e-01 4.52858865e-01 -2.73845673e-01 -3.33454728e-01
-1.80748597e-01 7.31581330e-01 -2.94498559e-02 9.56658199e-02
9.59287345e-01 6.54080138e-02 -4.22248960e-01 -9.29149032e-01
1.97806299e-01 8.46516252e-01 7.76833415e-01 -4.61513937e-01
7.04173803e-01 1.32178571e-02 -1.44527927e-01 -1.12535667e+00
-1.02518141e+00 -7.90591359e-01 -7.94501543e-01 -3.99269462e-01
5.94374299e-01 -7.15792775e-01 -1.92119271e-01 2.04412565e-01
-7.73203194e-01 -1.78220034e-01 -7.15378284e-01 7.87316084e-01
-8.93815875e-01 -2.96705253e-02 -4.63928610e-01 -9.59235013e-01
-3.91201288e-01 -5.68068922e-01 9.63510811e-01 2.60777861e-01
-7.90247500e-01 -7.71355569e-01 7.53543019e-01 1.46487474e-01
2.25021839e-01 -3.94404419e-02 8.38636994e-01 -8.48537445e-01
1.52019747e-02 -2.89341331e-01 7.04297483e-01 4.81524944e-01
3.55389178e-01 -1.97426200e-01 -1.32910502e+00 -1.55923545e-01
2.19241738e-01 -3.36139619e-01 7.31208682e-01 5.35545528e-01
9.87506270e-01 -3.56562734e-01 2.28487805e-01 3.53704751e-01
1.12846684e+00 3.36732447e-01 4.34529305e-01 3.46273124e-01
2.43326470e-01 5.11235595e-01 1.00432265e+00 4.94837433e-01
-1.68752581e-01 9.59818363e-01 -1.70634985e-02 2.30073407e-01
-3.51315945e-01 -5.45005143e-01 5.03322184e-01 1.37590265e+00
-2.45477811e-01 2.91781962e-01 -8.10241759e-01 7.58843899e-01
-1.71798849e+00 -1.10236931e+00 2.55547941e-01 2.36889577e+00
1.28001285e+00 3.77910346e-01 8.03616107e-01 1.08268273e+00
5.93873739e-01 -7.33236643e-03 -3.11061084e-01 -4.71954823e-01
1.84506066e-02 7.28228271e-01 1.61592942e-02 2.01961890e-01
-1.38200808e+00 9.77024078e-01 6.51461983e+00 1.24214709e+00
-1.29084361e+00 1.46514580e-01 1.90754868e-02 -2.58553654e-01
1.75231323e-01 2.01737210e-02 -7.07170069e-01 2.78564721e-01
1.20730495e+00 -1.71048164e-01 6.72816038e-01 8.41083407e-01
2.78958142e-01 1.70900702e-01 -1.15925515e+00 1.25627851e+00
7.65314177e-02 -1.09384680e+00 -1.08165272e-01 -1.77181035e-01
8.73050869e-01 -3.26021910e-01 -5.61675541e-02 4.09224868e-01
-4.77241725e-01 -8.57444227e-01 7.30102599e-01 3.44292045e-01
5.81117034e-01 -1.03537846e+00 5.67535222e-01 5.02964735e-01
-1.51751053e+00 -7.91692808e-02 -2.46579170e-01 -1.48245245e-01
8.83567929e-02 4.31974441e-01 -1.11633086e+00 3.33968133e-01
6.26986444e-01 4.58543032e-01 -5.06518245e-01 1.40968740e+00
-1.62315562e-01 1.08672011e+00 -3.42218697e-01 9.55589265e-02
-4.40502129e-02 7.42208138e-02 8.44136059e-01 1.38614154e+00
3.11410844e-01 -2.00442746e-01 2.79695421e-01 4.74697053e-01
5.60341328e-02 6.26527846e-01 -4.17704195e-01 -3.48136067e-01
4.82653350e-01 1.23095751e+00 -7.03235745e-01 -2.30338231e-01
-1.61313534e-01 6.38654709e-01 -2.75471359e-01 3.04194819e-02
-5.32952726e-01 -4.00494397e-01 2.32606977e-01 1.78509176e-01
3.62264872e-01 -1.88764960e-01 -2.29283392e-01 -7.94243753e-01
-3.07506025e-01 -1.14302635e+00 5.76133668e-01 -5.12515306e-01
-1.32400060e+00 6.40842199e-01 -9.22623649e-02 -1.74212945e+00
-8.81911814e-01 -3.71131033e-01 -4.19697374e-01 6.04897499e-01
-1.12039125e+00 -1.02610934e+00 1.49739832e-01 5.44859052e-01
9.55010116e-01 -4.71985102e-01 1.12675285e+00 4.87581342e-01
1.69069737e-01 3.42782825e-01 -2.74661839e-01 1.22800268e-01
9.79521155e-01 -1.37353504e+00 6.60370141e-02 4.87887353e-01
1.10649073e+00 2.66617179e-01 1.00048602e+00 -4.20783103e-01
-9.89874244e-01 -8.34697366e-01 1.00435150e+00 -1.87044322e-01
5.99141598e-01 -1.04078196e-01 -8.87863159e-01 1.49995789e-01
7.42319748e-02 -4.17643458e-01 1.24590719e+00 3.57585430e-01
-1.42992750e-01 -3.60762298e-01 -5.45941532e-01 4.23609853e-01
5.16023099e-01 -9.23634708e-01 -9.39103186e-01 -2.41728909e-02
3.55345577e-01 -4.18167114e-01 -1.10908747e+00 4.32273030e-01
8.84551346e-01 -7.86300600e-01 9.83403921e-01 -6.09785199e-01
2.16749147e-01 -4.41842109e-01 -3.05462092e-01 -1.38002181e+00
-1.11648493e-01 -8.42787147e-01 -2.64312267e-01 1.38752759e+00
5.54885864e-01 2.27885902e-01 6.90532327e-01 -1.64803222e-01
-1.25355674e-02 -3.21605951e-01 -8.66061866e-01 -8.44291270e-01
-6.83630258e-02 -8.33492398e-01 4.60611016e-01 8.46072316e-01
1.96485937e-01 7.84650683e-01 -6.35253489e-01 -1.28483012e-01
4.08399612e-01 4.76011217e-01 9.60320413e-01 -1.51898682e+00
-7.23739862e-01 -4.45114642e-01 -4.47819561e-01 -6.15290642e-01
-3.34205963e-02 -7.46097505e-01 -3.82030685e-03 -9.27722633e-01
-9.42444801e-02 -3.02996725e-01 -7.78369009e-01 3.43122542e-01
6.49295375e-02 7.40673184e-01 4.70615804e-01 3.72192055e-01
-2.90740430e-01 2.70747095e-01 1.10713685e+00 -1.58814296e-01
-8.76773775e-01 6.40166700e-01 -4.09282476e-01 1.14268553e+00
1.04384780e+00 -7.74451792e-01 -4.09012705e-01 3.15470576e-01
4.08151448e-01 1.25127867e-01 3.08194328e-02 -1.27059138e+00
2.18707085e-01 5.37099503e-02 2.45597467e-01 -5.26516616e-01
6.69549108e-01 -6.57179475e-01 1.38332635e-01 2.66916513e-01
-5.26116073e-01 -1.68747574e-01 3.44414145e-01 2.69254357e-01
-6.45120442e-01 -5.83008647e-01 8.05039108e-01 -2.05022648e-01
-5.46973288e-01 -1.09534472e-01 -3.46228808e-01 2.85365105e-01
3.02116066e-01 -6.57172054e-02 3.27937782e-01 -4.35867906e-01
-9.09986496e-01 -6.64267778e-01 1.27025470e-01 5.95048785e-01
4.53601748e-01 -1.41055834e+00 -6.80382073e-01 9.56481174e-02
1.30663365e-01 -4.75857139e-01 -2.20009964e-03 5.93190193e-01
-1.65262252e-01 3.31957847e-01 -3.70300710e-01 -5.85106075e-01
-1.46336031e+00 3.90035957e-01 9.32163596e-02 -3.10441047e-01
-4.33124751e-01 5.83072364e-01 -1.25246286e-01 -1.86270431e-01
3.16040695e-01 -1.38983324e-01 -5.01639485e-01 3.21954012e-01
5.60075939e-01 1.86961025e-01 1.74580738e-01 -6.83636606e-01
-1.82744309e-01 6.75711215e-01 1.94664672e-01 -6.76924884e-01
1.33553874e+00 1.12701513e-01 1.02408893e-01 1.18917143e+00
7.30700970e-01 5.69210529e-01 -9.03376043e-01 -1.54166788e-01
3.97642165e-01 -3.74783337e-01 7.40569085e-02 -7.03945756e-01
-6.36403084e-01 8.24607432e-01 5.50992727e-01 4.49426532e-01
1.58280683e+00 -1.46551713e-01 6.77363396e-01 1.96938306e-01
3.73877972e-01 -1.55171061e+00 2.43374676e-01 3.74424785e-01
8.77330422e-01 -8.44085455e-01 -1.14298753e-01 -1.52621567e-01
-5.97637773e-01 1.28451347e+00 1.81192219e-01 -3.46799403e-01
8.91025722e-01 1.10019289e-01 2.41077006e-01 6.64412677e-02
-5.50684631e-01 -3.94827753e-01 9.49956477e-01 3.70131969e-01
6.77298605e-01 4.47330289e-02 -4.01349276e-01 8.94794822e-01
-9.05086160e-01 -4.16019373e-02 2.21923202e-01 7.34524369e-01
-6.51384175e-01 -1.62922788e+00 -5.72514594e-01 1.84071630e-01
-8.70134890e-01 -7.45228752e-02 -7.48814046e-01 7.00887203e-01
3.80283922e-01 1.00040460e+00 -1.82992354e-01 -4.65564132e-01
2.41868839e-01 4.28882688e-01 6.85940325e-01 -8.19636703e-01
-1.06063735e+00 1.01488590e+00 1.26616135e-01 -7.86523446e-02
-8.96744430e-01 -6.52770281e-01 -9.16227996e-01 3.10342699e-01
-4.15829867e-01 5.29835343e-01 6.63265109e-01 7.91904628e-01
-1.12120405e-01 6.77045941e-01 6.86610162e-01 -9.96428609e-01
-3.91450435e-01 -1.22281039e+00 -8.30433011e-01 3.51859540e-01
2.45759785e-01 -4.08953786e-01 -1.32539645e-01 5.10399640e-01] | [15.898159980773926, 5.284669399261475] |
f18fb92a-2dec-4351-86b7-b7cc7e6e6641 | joint-multisided-exposure-fairness-for | 2205.00048 | null | https://arxiv.org/abs/2205.00048v1 | https://arxiv.org/pdf/2205.00048v1.pdf | Joint Multisided Exposure Fairness for Recommendation | Prior research on exposure fairness in the context of recommender systems has focused mostly on disparities in the exposure of individual or groups of items to individual users of the system. The problem of how individual or groups of items may be systemically under or over exposed to groups of users, or even all users, has received relatively less attention. However, such systemic disparities in information exposure can result in observable social harms, such as withholding economic opportunities from historically marginalized groups (allocative harm) or amplifying gendered and racialized stereotypes (representational harm). Previously, Diaz et al. developed the expected exposure metric -- that incorporates existing user browsing models that have previously been developed for information retrieval -- to study fairness of content exposure to individual users. We extend their proposed framework to formalize a family of exposure fairness metrics that model the problem jointly from the perspective of both the consumers and producers. Specifically, we consider group attributes for both types of stakeholders to identify and mitigate fairness concerns that go beyond individual users and items towards more systemic biases in recommendation. Furthermore, we study and discuss the relationships between the different exposure fairness dimensions proposed in this paper, as well as demonstrate how stochastic ranking policies can be optimized towards said fairness goals. | ['Xue Liu', 'Fernando Diaz', 'Chen Ma', 'Bhaskar Mitra', 'Haolun Wu'] | 2022-04-29 | null | null | null | null | ['exposure-fairness'] | ['adversarial'] | [ 1.56351894e-01 3.76105398e-01 -3.32879603e-01 -5.25594831e-01
-1.81197912e-01 -7.98128664e-01 5.49177289e-01 5.63546777e-01
-5.47600210e-01 4.19336349e-01 6.60318255e-01 -4.54962850e-01
-5.06904960e-01 -9.55197632e-01 -1.14774361e-01 -1.79728031e-01
1.48676619e-01 -4.99474667e-02 -3.66627693e-01 -2.95855343e-01
7.82708824e-01 3.57394546e-01 -1.59277856e+00 1.90542743e-01
1.15921259e+00 7.35471487e-01 -3.00519079e-01 2.09267691e-01
-5.43678105e-02 8.29354882e-01 -5.32242715e-01 -9.17534053e-01
8.17009687e-01 -3.69075030e-01 -6.94214702e-01 -2.10140824e-01
9.16500330e-01 -6.83268249e-01 -1.20241009e-02 1.38342249e+00
4.91034389e-01 3.69850248e-01 8.50781560e-01 -1.31361282e+00
-1.34531415e+00 1.07070780e+00 -6.37914479e-01 1.39497936e-01
1.49866685e-01 -1.39526159e-01 1.16439664e+00 -4.56611037e-01
3.70674342e-01 1.49097586e+00 3.37150186e-01 4.95181739e-01
-1.35158896e+00 -8.22559357e-01 4.71429288e-01 -1.66657388e-01
-1.01722085e+00 -4.17058438e-01 1.03168875e-01 -8.30342829e-01
2.89413750e-01 9.00040448e-01 3.57192010e-01 5.48424304e-01
8.43827203e-02 2.12501153e-01 1.36239791e+00 -3.82066041e-01
1.91828519e-01 5.09598970e-01 6.85871780e-01 -5.10042533e-02
9.98046517e-01 2.96225250e-01 -3.52410048e-01 -5.27513146e-01
2.34627083e-01 1.05338603e-01 -1.94041923e-01 -1.89535886e-01
-4.32972848e-01 1.05550623e+00 3.97508681e-01 -1.98606566e-01
-5.52374482e-01 -8.31539631e-02 2.57057875e-01 4.32962269e-01
8.25191140e-01 8.24389279e-01 -8.01463500e-02 5.02053857e-01
-6.92126453e-01 6.96849287e-01 8.56847048e-01 7.74434268e-01
6.46438003e-01 -4.16425347e-01 -8.71809781e-01 6.08615696e-01
5.92593253e-01 3.55208665e-01 -1.73383534e-01 -1.07429290e+00
4.44322526e-01 3.30594599e-01 5.82415342e-01 -1.37000203e+00
-6.21210113e-02 -4.39144164e-01 -3.12955141e-01 6.02169275e-01
7.91429162e-01 -3.46951276e-01 -2.00427264e-01 2.09680820e+00
4.17772889e-01 -7.48278975e-01 -3.69552553e-01 1.19105101e+00
3.67561907e-01 2.66906440e-01 7.77483821e-01 -3.11780214e-01
1.45197606e+00 -5.99911392e-01 -8.15192103e-01 -4.35618348e-02
3.57568324e-01 -7.47170448e-01 1.05207598e+00 1.94321290e-01
-1.45274699e+00 -1.64121330e-01 -6.40969872e-01 -1.38476491e-01
-3.10567409e-01 -5.31274378e-01 3.85578424e-01 1.48598206e+00
-7.80746818e-01 7.10054517e-01 1.22866057e-01 -4.54831839e-01
6.05172455e-01 7.83610046e-02 5.31879485e-01 2.18642518e-01
-1.42437327e+00 1.12581396e+00 -3.59529614e-01 -2.38539696e-01
-4.97409284e-01 -1.42584038e+00 -2.40241364e-01 4.57936168e-01
5.15578687e-01 -9.22600091e-01 1.15877175e+00 -1.36099398e+00
-9.92840409e-01 5.15823603e-01 5.93203723e-01 -1.40457869e-01
8.18523586e-01 -5.89134753e-01 -3.11187744e-01 -4.61157680e-01
2.40350083e-01 1.90785244e-01 4.46970671e-01 -1.35489726e+00
-8.08794200e-01 -7.03011632e-01 6.92183912e-01 6.27917588e-01
-6.58947349e-01 6.71554446e-01 4.52159405e-01 -7.75640547e-01
-6.02174163e-01 -6.64948106e-01 -3.90785158e-01 2.27284104e-01
-4.06087250e-01 4.58022431e-02 4.37314987e-01 -5.97954214e-01
1.49204767e+00 -1.96005511e+00 -2.87148923e-01 3.92879605e-01
3.45353037e-01 -1.03394000e-03 8.87169912e-02 4.64494705e-01
1.74098551e-01 6.93692625e-01 2.64152646e-01 3.45234238e-02
4.15270209e-01 -2.79744565e-01 -3.28240871e-01 6.88012600e-01
-3.53027374e-01 2.00140893e-01 -8.57918024e-01 1.60493031e-02
3.24874111e-02 2.12071270e-01 -8.82232547e-01 1.76933080e-01
1.33834198e-01 8.23193565e-02 -3.21973592e-01 1.11368664e-01
1.07654643e+00 2.87072539e-01 3.15671504e-01 1.95249822e-02
-4.89003122e-01 5.31707048e-01 -1.13308084e+00 7.21076369e-01
-3.60232443e-01 1.50114715e-01 2.45866507e-01 -4.38710928e-01
3.56665701e-01 -1.01904318e-01 2.91633129e-01 -8.66608381e-01
6.43914491e-02 -3.55740450e-02 2.21460149e-01 -2.56248355e-01
1.05254245e+00 -4.34179008e-01 -2.61660278e-01 1.11590588e+00
-4.69203740e-01 3.47903609e-01 -2.04712469e-02 4.65316027e-01
3.72565895e-01 -2.67497748e-01 4.90121126e-01 -1.02170861e+00
2.97549605e-01 -2.71074414e-01 5.72631717e-01 1.16755772e+00
-4.94372666e-01 1.30601168e-01 5.51965535e-01 1.61964834e-01
-9.52952802e-01 -8.84080529e-01 -2.73365080e-01 1.76753151e+00
6.80901349e-01 -5.88362068e-02 -9.54257965e-01 -7.57044375e-01
4.49992836e-01 1.38327754e+00 -7.04251587e-01 -2.23330006e-01
1.11734375e-01 -6.74962163e-01 3.28739703e-01 1.47277519e-01
1.45576879e-01 -5.77978909e-01 -1.02261746e+00 -2.98053592e-01
1.55526623e-01 -2.55438715e-01 -9.50348556e-01 -4.49968219e-01
-4.91829515e-01 -1.03827691e+00 -4.99766618e-01 1.36332929e-01
5.27494907e-01 4.76030976e-01 1.17513096e+00 1.84652373e-01
-1.70171991e-01 5.15879929e-01 -3.06524992e-01 -9.42671895e-01
-2.45431915e-01 -4.60943133e-01 1.71997830e-01 6.49073571e-02
2.38896519e-01 -1.07105240e-01 -1.16098166e+00 6.29977226e-01
-8.56143534e-01 -2.04072788e-01 -1.21436022e-01 1.51342034e-01
-4.45890762e-02 -2.84665763e-01 9.18463051e-01 -1.58500946e+00
1.20687175e+00 -1.03063226e+00 -4.04956311e-01 3.25244635e-01
-1.35109162e+00 -4.96332526e-01 4.70359325e-01 -2.56271750e-01
-1.74665594e+00 -1.06320941e+00 4.87187892e-01 4.74813074e-01
1.07673049e-01 3.70923370e-01 -4.57282066e-01 -7.10478500e-02
7.66422629e-01 -9.26096976e-01 -1.18136145e-01 -4.44470704e-01
9.17502344e-01 8.16946268e-01 -1.42671794e-01 -7.46603429e-01
3.46547067e-01 4.88229990e-01 -4.13152665e-01 -3.93704414e-01
-9.46700215e-01 -1.77014634e-01 8.27517062e-02 -4.22696024e-01
6.05595827e-01 -6.54820323e-01 -6.53762639e-01 1.14929102e-01
-5.99133372e-01 -1.06466763e-01 -6.50728226e-01 3.23342472e-01
-2.33103856e-01 4.79676604e-01 -4.29056555e-01 -1.20886898e+00
-3.11140358e-01 -7.70646811e-01 2.38994166e-01 3.25775623e-01
-6.28677428e-01 -8.57972145e-01 -3.01785395e-03 3.85536611e-01
8.15141976e-01 7.17252865e-02 9.47826445e-01 -7.66788006e-01
-3.98540080e-01 2.89630052e-02 -3.58282000e-01 1.62969470e-01
1.29444554e-01 -4.78987470e-02 -8.12738657e-01 -2.97387034e-01
1.58497363e-01 -7.89505839e-02 4.38157618e-01 5.65200031e-01
1.05165577e+00 -9.24822688e-01 -8.06989372e-02 1.81316525e-01
1.45566356e+00 2.10158199e-01 4.61754918e-01 3.05496335e-01
3.70471388e-01 1.43449724e+00 9.33621466e-01 8.30720127e-01
3.77285987e-01 6.79284990e-01 4.33215171e-01 -1.57869026e-01
1.33896798e-01 -1.60155147e-01 8.54218081e-02 -1.80980265e-02
-1.80722311e-01 -5.37982523e-01 -3.05716157e-01 3.88718218e-01
-1.78006291e+00 -1.13831353e+00 -3.79392028e-01 2.59285092e+00
3.67205977e-01 -3.76755923e-01 5.10630190e-01 -3.55828106e-01
1.05460262e+00 6.15551583e-02 -3.74746859e-01 -1.01962841e+00
1.78301156e-01 -2.25828841e-01 8.63821745e-01 6.72729969e-01
-5.98346293e-01 3.58587652e-01 6.54258823e+00 5.41276693e-01
-4.67108399e-01 2.23829344e-01 1.04181838e+00 -6.42740488e-01
-1.14832950e+00 2.61668891e-01 -5.26340723e-01 6.09286785e-01
8.58311951e-01 -9.15347636e-01 3.89605671e-01 5.18338382e-01
5.46780586e-01 3.27117853e-02 -1.15565598e+00 2.24350318e-01
-2.16626301e-01 -7.35614181e-01 2.84861065e-02 4.26375985e-01
9.92076337e-01 -6.26946330e-01 6.05079114e-01 1.66107610e-01
7.58843839e-01 -9.13512230e-01 1.28287315e+00 7.25608706e-01
4.00127560e-01 -8.14452410e-01 3.23595792e-01 2.33465165e-01
-5.50883830e-01 -4.54636097e-01 -5.37793994e-01 -5.69440901e-01
2.25974858e-01 7.47811854e-01 3.88949811e-02 3.47605258e-01
6.24758959e-01 3.48518267e-02 -2.69077361e-01 8.08426678e-01
1.43311933e-01 3.10936421e-01 1.71901390e-01 5.87614626e-02
-1.03191569e-01 -4.58846927e-01 3.48603129e-01 9.68059123e-01
3.52567732e-01 4.10612643e-01 -2.15383053e-01 1.24476743e+00
-2.16041356e-01 6.06023073e-01 -4.18903619e-01 1.10110506e-01
7.33810127e-01 1.32017922e+00 -3.07649434e-01 -2.63204366e-01
-5.38895309e-01 2.76113242e-01 2.98952341e-01 3.81356418e-01
-6.63566470e-01 -8.52003470e-02 1.17678499e+00 3.83881092e-01
-4.12506968e-01 4.69186872e-01 -7.72243261e-01 -8.22185457e-01
-4.42489803e-01 -1.00367737e+00 6.66785896e-01 -1.18068211e-01
-1.77712333e+00 -2.69566357e-01 1.48135513e-01 -7.03937829e-01
5.29594660e-01 -2.36644134e-01 -7.05203295e-01 1.43141031e+00
-1.21383393e+00 -6.73368514e-01 -1.21567175e-01 1.96699888e-01
5.67416735e-02 2.11423397e-01 3.93240571e-01 4.27142888e-01
-4.90845621e-01 8.28192770e-01 2.03971922e-01 -7.53168941e-01
9.38569546e-01 -1.17119670e+00 7.85561129e-02 8.24206471e-01
-5.08030057e-01 1.05829370e+00 8.66675317e-01 -7.77023017e-01
-8.68770897e-01 -1.11266255e+00 9.56753016e-01 -6.06068134e-01
2.27135599e-01 -1.75512195e-01 -6.69803143e-01 4.60428447e-01
3.26705366e-01 -7.35076427e-01 1.33665001e+00 5.83214700e-01
-4.99264508e-01 3.84609513e-02 -1.66967857e+00 7.97766745e-01
1.28241813e+00 -5.59183538e-01 -1.41209766e-01 1.39727905e-01
4.66438085e-01 1.73174769e-01 -9.77693141e-01 5.22463135e-02
9.57423925e-01 -1.35967731e+00 9.31591868e-01 -9.22247767e-01
6.82027698e-01 8.59315544e-02 -3.16889763e-01 -1.30498743e+00
-8.99827719e-01 -3.85545939e-01 3.07920158e-01 1.55662858e+00
3.27819556e-01 -6.91629171e-01 3.33954930e-01 1.61295927e+00
6.34566844e-02 -2.83497751e-01 -2.31763899e-01 -4.83401835e-01
5.23126662e-01 3.77132930e-02 9.44574952e-01 1.22313011e+00
2.12157577e-01 -1.34495981e-02 -4.58815008e-01 2.19001085e-01
8.36141169e-01 1.21272720e-01 5.54493248e-01 -1.13766551e+00
-1.61689118e-01 -7.93277264e-01 2.93473363e-01 -4.66525376e-01
-1.63645849e-01 -7.80144632e-01 -1.14193305e-01 -1.31268120e+00
5.84717274e-01 -7.09492862e-01 -5.30683815e-01 -1.20218284e-01
-4.37037140e-01 -6.85097277e-02 6.38244510e-01 2.48612404e-01
-2.39661008e-01 1.24324694e-01 1.09798455e+00 1.97855473e-01
-3.75234671e-02 -1.36076473e-02 -2.04423308e+00 6.46942616e-01
5.57990134e-01 -5.35808861e-01 -5.67844868e-01 -2.92487115e-01
6.76340520e-01 -6.20168820e-02 2.84312725e-01 -2.30192035e-01
-4.71590385e-02 -8.73043418e-01 2.76486669e-02 1.10794239e-01
-4.88927007e-01 -9.98854280e-01 5.09544075e-01 4.01142538e-01
-1.12626195e+00 -2.84011483e-01 -2.45987475e-01 7.00358033e-01
4.06808555e-01 -4.94176567e-01 8.28024268e-01 -1.49755582e-01
2.01836303e-02 2.71538019e-01 -4.99138892e-01 1.63599119e-01
1.18782699e+00 -1.20722048e-01 -8.33501101e-01 -5.44349015e-01
-4.52538043e-01 3.03500682e-01 7.97261298e-01 4.02105242e-01
-8.24354216e-02 -1.21804869e+00 -7.57459998e-01 -4.92887884e-01
2.22277164e-01 -8.05732548e-01 5.90705693e-01 5.48217952e-01
-1.59341678e-01 1.80651188e-01 -5.06329179e-01 3.67921263e-01
-1.13961041e+00 6.74838543e-01 3.57731760e-01 9.20059308e-02
5.31823672e-02 8.20397437e-01 9.25342798e-01 -2.57920474e-01
6.15853854e-02 1.98299050e-01 -4.45180714e-01 5.02505064e-01
6.07938528e-01 1.16839468e+00 -2.98287958e-01 -5.26339889e-01
-2.82923639e-01 8.85351673e-02 -1.56821445e-01 -7.81648457e-02
8.28227162e-01 -7.37929583e-01 -1.43753201e-01 -1.09089129e-02
7.76847601e-01 5.04231870e-01 -1.03853762e+00 6.60148039e-02
-1.28135890e-01 -1.23942649e+00 9.45040733e-02 -1.01500475e+00
-8.39962125e-01 7.36894965e-01 5.94803154e-01 7.92305052e-01
9.94575441e-01 -3.75977039e-01 1.68997288e-01 -3.38919461e-01
2.40915194e-01 -1.34525371e+00 -2.46821821e-01 -2.68072486e-01
8.11842620e-01 -7.91459322e-01 2.46647432e-01 -5.64316034e-01
-7.49449432e-01 5.35009682e-01 6.89588368e-01 -5.11575118e-02
6.75263464e-01 -2.60848254e-01 -1.42506380e-02 -1.94561020e-01
-4.18671578e-01 -1.08234128e-02 3.94743592e-01 4.38367307e-01
9.51051652e-01 5.16988575e-01 -1.29943871e+00 8.73444438e-01
2.72623403e-03 -1.90604657e-01 8.75367343e-01 6.57279789e-01
-4.57124084e-01 -1.09549332e+00 -3.09752077e-01 8.53611946e-01
-1.00093794e+00 -1.90255687e-01 -6.20194674e-01 4.95618284e-01
4.13139999e-01 1.18514812e+00 1.74748138e-01 5.13714254e-02
5.67748487e-01 -3.83093834e-01 2.64917701e-01 -7.90638506e-01
-1.16608202e+00 -1.41967192e-01 5.34651577e-01 -5.35808563e-01
-5.73671877e-01 -9.60793614e-01 -4.64471906e-01 -7.47667372e-01
-6.39364183e-01 1.36683390e-01 3.34665805e-01 5.24361134e-01
3.62282813e-01 2.76649684e-01 8.72167885e-01 -3.87258887e-01
-1.07614636e+00 -6.34852529e-01 -1.06723154e+00 7.70380080e-01
-2.43662335e-02 -5.65449834e-01 -5.67915022e-01 -4.23721492e-01] | [9.533177375793457, 5.659571647644043] |
d4afeaa4-9770-4756-b133-37f64f414ba6 | using-referring-expression-generation-to | null | null | https://aclanthology.org/2021.nlp4dh-1.8 | https://aclanthology.org/2021.nlp4dh-1.8.pdf | Using Referring Expression Generation to Model Literary Style | Novels and short stories are not just remarkable because of what events they represent. The narrative style they employ is significant. To understand the specific contributions of different aspects of this style, it is possible to create limited symbolic models of narrating that hold almost all of the narrative discourse constant while varying a single aspect. In this paper we use a new implementation of a system for narrative discourse generation, Curveship, to change how existents at the story level are named. This by itself allows for the telling of the same underlying story in ways that evoke, for instance, a fabular or parable-like mode, the style of narrator Patrick Bateman in Brett Easton Ellis’s American Psycho, and the unusual dialect of Anthony Burgess’s A Clockwork Orange. | ['Alan Y. Zhu', 'Joanne Yuan', 'Ardalan SadeghiKivi', 'Nick Montfort'] | null | null | null | null | nlp4dh-icon-2021-12 | ['referring-expression-generation', 'referring-expression'] | ['computer-vision', 'computer-vision'] | [ 1.34677634e-01 3.78485829e-01 -1.58310294e-01 -1.22425117e-01
-1.21923387e-01 -1.02211273e+00 1.39553392e+00 4.04243395e-02
2.21012929e-03 1.11291862e+00 1.13436699e+00 -4.22359943e-01
-3.71839702e-01 -9.29703236e-01 -3.04302275e-01 -3.79524976e-01
2.08217919e-01 2.63350040e-01 3.20307970e-01 -1.01934755e+00
5.14176369e-01 5.90738475e-01 -1.17010367e+00 8.06188285e-01
3.12055796e-01 -4.86694947e-02 8.46831575e-02 8.20128083e-01
-5.38344204e-01 1.65506661e+00 -1.09816349e+00 -4.80232388e-01
-4.82883662e-01 -1.11706042e+00 -9.75228548e-01 5.67271039e-02
-8.01437348e-02 -1.51489779e-01 -3.24087620e-01 3.59344989e-01
2.05926970e-01 -1.15107529e-01 6.53822482e-01 -9.13365960e-01
-4.32416379e-01 1.42465627e+00 -2.21886203e-01 3.05279642e-01
7.93389738e-01 -4.92180325e-02 5.17383158e-01 -5.82842946e-01
1.42312300e+00 1.31265092e+00 6.86381459e-01 3.94551754e-01
-1.40545583e+00 -3.37786704e-01 -3.49658094e-02 5.54627478e-02
-1.08576989e+00 -4.58591312e-01 7.98738241e-01 -6.19040012e-01
7.82280505e-01 7.96772420e-01 1.23703671e+00 1.38285017e+00
3.77655387e-01 2.70253479e-01 1.19189429e+00 -4.31781739e-01
-6.54291082e-03 3.49157095e-01 -8.96882173e-03 2.12320406e-02
-3.04586701e-02 -1.16982371e-01 -6.84860051e-01 -1.31114557e-01
7.05713093e-01 -2.95902282e-01 -2.46650949e-01 2.97137678e-01
-1.47740293e+00 8.12465847e-01 7.76362121e-02 1.15014124e+00
1.63750269e-03 2.95320928e-01 6.51742876e-01 3.26196015e-01
2.61598863e-02 7.69196391e-01 6.71725273e-02 -7.36519575e-01
-8.47276926e-01 9.66010153e-01 1.13720572e+00 7.77391970e-01
5.09519354e-02 3.21397595e-02 -1.02865919e-01 3.89161557e-01
9.53461137e-03 -3.85038406e-02 2.42296949e-01 -1.04602969e+00
7.35521913e-02 3.35982054e-01 2.58884639e-01 -1.17747772e+00
-3.38407457e-01 -1.75132811e-01 -3.45958956e-02 3.74285966e-01
4.79020387e-01 -1.64387539e-01 -1.09625228e-01 1.49632633e+00
1.72805697e-01 -4.11055028e-01 1.82347059e-01 5.82482159e-01
1.16709054e+00 8.52217495e-01 4.22126614e-02 -4.43164766e-01
1.31818223e+00 -5.63556969e-01 -1.18823969e+00 8.16206932e-02
8.25221360e-01 -1.09905398e+00 1.24052119e+00 2.33639494e-01
-1.56031263e+00 3.86380143e-02 -1.32511592e+00 -2.78887987e-01
-3.22940201e-01 -5.02756298e-01 4.80503678e-01 5.48701882e-01
-5.14267921e-01 6.34009123e-01 -4.02183861e-01 -6.20160282e-01
2.18166977e-01 -5.21634281e-01 -1.71357840e-01 5.20503223e-01
-9.54481065e-01 1.12844789e+00 7.11493254e-01 -3.45231146e-01
-3.32756609e-01 -9.08038259e-01 -4.90724444e-01 -1.08813889e-01
5.06230652e-01 -4.41295236e-01 1.29972041e+00 -1.13575304e+00
-1.50881600e+00 1.10801852e+00 9.21280012e-02 -3.36634666e-01
8.79450679e-01 -1.08215533e-01 -6.88409805e-01 1.56219617e-01
4.85314317e-02 5.49343452e-02 2.59527296e-01 -1.36951709e+00
-3.73301595e-01 8.33158344e-02 4.67051417e-01 -3.16968933e-02
1.89820781e-01 7.54920483e-01 5.94972074e-01 -9.10930037e-01
-6.87461421e-02 -2.65136302e-01 8.83135870e-02 -2.16268435e-01
-5.85063040e-01 1.19951122e-01 1.04084146e+00 -6.42139375e-01
1.79319394e+00 -2.19520473e+00 1.22307561e-01 -3.57859209e-02
3.67801666e-01 -3.14146936e-01 5.62838078e-01 1.43329740e+00
-2.60700267e-02 5.31664431e-01 -2.31766596e-01 3.59660774e-01
1.66742221e-01 4.27860469e-01 -4.40977782e-01 7.02982843e-02
9.59930792e-02 8.42472374e-01 -8.38582516e-01 -4.89640206e-01
-2.19088033e-01 4.66833383e-01 -2.50635028e-01 -2.14223459e-01
-1.44434705e-01 4.36097682e-01 -3.51392537e-01 1.60146311e-01
2.31143653e-01 -1.78070679e-01 4.84345317e-01 2.11448878e-01
-9.08471763e-01 7.20753789e-01 -1.09002507e+00 1.49434090e+00
-7.63979703e-02 1.18874383e+00 -2.08304346e-01 -4.23825443e-01
8.38020205e-01 5.64793527e-01 -2.15212509e-01 -3.95124137e-01
2.62168705e-01 2.21546292e-01 2.52596468e-01 -7.58479178e-01
9.88400698e-01 -6.51605964e-01 -4.10524338e-01 7.90031672e-01
-4.80534226e-01 -5.05564511e-01 4.34505820e-01 4.71231461e-01
1.01082456e+00 4.16739523e-01 8.51593971e-01 -2.80213386e-01
1.27082586e-01 5.65873027e-01 1.62687570e-01 5.06373286e-01
3.19369793e-01 7.56113291e-01 9.78593588e-01 -5.09316564e-01
-1.52977931e+00 -1.06560349e+00 -5.11411309e-01 6.66053891e-01
-8.39140713e-02 -8.84256959e-01 -6.84144080e-01 -6.39583468e-02
-4.96547222e-01 1.55782688e+00 -7.44255781e-01 1.88318923e-01
-9.67781484e-01 -4.50786293e-01 7.02920616e-01 6.48352653e-02
2.05100968e-01 -1.23419023e+00 -1.48675299e+00 3.98768693e-01
-9.08642411e-02 -8.26499641e-01 4.01436724e-02 1.24866301e-02
-4.26198304e-01 -7.50327826e-01 -4.73505318e-01 -3.95699471e-01
8.26360881e-02 -3.42279077e-01 9.45078254e-01 1.12638660e-02
-9.16726887e-02 -1.42906025e-01 -6.08499050e-01 -4.81651157e-01
-1.23545289e+00 -2.42781173e-02 -6.82845771e-01 -3.98864836e-01
-1.46649063e-01 -9.47156310e-01 1.35646611e-01 -2.03295112e-01
-1.12609100e+00 6.93826854e-01 -2.59696752e-01 5.85137367e-01
5.26849292e-02 -2.29887962e-01 6.24637842e-01 -1.35119092e+00
6.34636939e-01 -6.34710252e-01 3.12851697e-01 5.61045893e-02
-9.22874138e-02 -2.72484213e-01 4.43846434e-01 -5.54474950e-01
-1.27440488e+00 -8.47903788e-01 -1.06339119e-01 4.38232332e-01
1.28466859e-02 5.43292224e-01 4.06882688e-02 5.34691930e-01
9.90474403e-01 -1.13850363e-01 -1.21053241e-01 -2.50353813e-01
4.06043530e-01 4.37852830e-01 7.09150255e-01 -4.02657270e-01
8.70100975e-01 5.82045555e-01 -9.26016793e-02 -7.23473728e-01
-4.09496754e-01 3.38636875e-01 -5.88302910e-01 -6.61760509e-01
8.57120097e-01 -5.01997352e-01 -5.97365499e-01 1.03952467e-01
-1.40611839e+00 -5.61644018e-01 -1.11793709e+00 1.32061362e-01
-7.70330966e-01 -2.79274046e-01 -5.49793065e-01 -6.59200430e-01
2.67464489e-01 -4.70603704e-01 2.21580595e-01 4.74114597e-01
-1.24643219e+00 -1.09513736e+00 7.10924715e-02 -1.50609583e-01
3.65912646e-01 1.34419978e+00 1.25040793e+00 -6.35583341e-01
-1.82118475e-01 7.72150382e-02 3.71450782e-02 -4.22755897e-01
1.27094626e-01 5.05398333e-01 -5.60573161e-01 4.15477872e-01
9.29645598e-02 -1.03394620e-01 7.79717416e-02 -2.50866741e-01
2.87174761e-01 -7.90985882e-01 -2.38116369e-01 3.19308847e-01
1.55194807e+00 7.00177729e-01 1.01569581e+00 1.00755203e+00
3.34445208e-01 7.23486960e-01 2.94463336e-01 6.99257135e-01
3.27786684e-01 6.21793807e-01 -1.49907738e-01 2.29133412e-01
-3.55816811e-01 -4.70783770e-01 2.90216118e-01 6.23432815e-01
-2.72034764e-01 -5.65060005e-02 -9.15833473e-01 6.89404368e-01
-1.73732233e+00 -1.81421971e+00 -8.98703814e-01 1.46439528e+00
1.12205720e+00 3.73648435e-01 1.19538136e-01 3.42804819e-01
3.00607145e-01 6.96278036e-01 2.04014197e-01 -1.07346547e+00
-6.09009087e-01 1.36039972e-01 -6.43175915e-02 5.79264045e-01
-3.03032696e-01 6.85617208e-01 7.32549191e+00 6.74199343e-01
-8.36315572e-01 1.37929738e-01 2.20286593e-01 -4.18109089e-01
-9.50290263e-01 3.63027573e-01 -3.92255157e-01 4.37214851e-01
6.87723160e-01 -8.73671949e-01 2.73851961e-01 2.55076647e-01
5.23563385e-01 -5.61015487e-01 -1.01808560e+00 4.23761338e-01
2.55386740e-01 -2.09250379e+00 -2.06456948e-02 -2.61508614e-01
6.83212101e-01 -9.44739103e-01 -2.73837984e-01 -2.60661125e-01
1.99421793e-01 -1.21199751e+00 1.57402813e+00 7.39692569e-01
6.49620593e-01 -9.10352468e-01 1.50284380e-01 3.46296519e-01
-5.26236296e-01 7.30943605e-02 2.46048018e-01 -5.69919884e-01
7.02908754e-01 -7.34067932e-02 -5.44445217e-01 3.67645204e-01
5.41382134e-01 3.13728869e-01 -1.95938617e-01 4.75799471e-01
-4.33625251e-01 7.57712841e-01 -2.29549438e-01 -2.82041252e-01
3.74673873e-01 -1.09423280e-01 1.27979064e+00 1.58311486e+00
2.50024259e-01 6.54105067e-01 -4.42914426e-01 1.07283688e+00
3.52851808e-01 6.14064746e-02 -7.47131228e-01 -3.08459938e-01
6.86617315e-01 9.37571228e-01 -8.76015902e-01 -2.57057339e-01
-2.22160835e-02 6.04428947e-01 -3.83630991e-01 3.21371764e-01
-8.43084812e-01 -3.39609534e-01 2.57812142e-01 8.95447135e-01
1.09413907e-01 -2.32721105e-01 -7.73634911e-01 -5.87362587e-01
-9.06841531e-02 -8.50783825e-01 -7.69160166e-02 -1.00557053e+00
-7.67954886e-01 5.17634034e-01 5.52803576e-01 -6.57168448e-01
-4.88519967e-01 1.48895994e-01 -1.35469377e+00 7.69403160e-01
-5.76184034e-01 -1.34976935e+00 -5.70080653e-02 1.45293787e-01
3.42830926e-01 -3.19832154e-02 7.52541363e-01 -1.34743795e-01
-8.62826854e-02 3.36981803e-01 1.56097203e-01 -1.15042120e-01
4.36501265e-01 -1.01259792e+00 3.10410768e-01 5.77532828e-01
-2.34806389e-02 3.92869353e-01 1.55452418e+00 -4.99211818e-01
-7.69029319e-01 -1.77283108e-01 1.36850238e+00 -4.02321577e-01
1.18567801e+00 -2.36038253e-01 -1.00271666e+00 8.39188933e-01
8.36852074e-01 -1.04796314e+00 6.52959228e-01 -1.52285233e-01
-2.63073593e-01 4.10346568e-01 -1.16550148e+00 1.06333292e+00
1.13784635e+00 -2.88702279e-01 -1.09590507e+00 1.13190740e-01
9.60214615e-01 -4.44490105e-01 -8.45510364e-01 -3.16649437e-01
8.67263138e-01 -1.23781586e+00 6.50329888e-01 -6.78031385e-01
1.06711805e+00 -3.65259856e-01 -1.85586549e-02 -9.69927430e-01
-9.81086120e-02 -1.15218127e+00 4.13889796e-01 1.78928828e+00
4.96622205e-01 -7.08908439e-01 2.47389674e-01 3.82661700e-01
-3.01922828e-01 -5.19009292e-01 -7.95442522e-01 -4.89777595e-01
4.06541139e-01 -2.90109426e-01 8.24743569e-01 1.49956298e+00
6.32119298e-01 3.05546075e-01 -9.19711590e-02 -6.85067654e-01
8.08371902e-02 -1.59829363e-01 8.45549345e-01 -9.78261352e-01
-3.98509353e-01 -6.46375954e-01 -2.56897181e-01 -2.56308496e-01
-4.31381702e-01 -9.33993220e-01 -4.99976307e-01 -1.63769162e+00
1.52828053e-01 -2.34238952e-01 6.94702148e-01 -3.05147022e-02
4.79418755e-01 3.59655954e-02 7.30980873e-01 2.67894685e-01
1.79931611e-01 3.66939120e-02 1.47450101e+00 2.69059837e-01
-5.93099892e-01 -5.97928762e-01 -9.37441409e-01 1.02573943e+00
7.77570128e-01 -7.31970251e-01 -4.06488985e-01 -1.81030221e-02
8.78471076e-01 2.69953877e-01 5.25303125e-01 -6.73626184e-01
1.05147228e-01 -6.30317509e-01 2.60565400e-01 -5.30839682e-01
1.89663261e-01 -3.64294469e-01 1.27729726e+00 4.65930581e-01
-5.74652195e-01 9.86903757e-02 2.51219839e-01 -1.89027458e-01
-1.35540530e-01 -5.77194691e-01 7.29466379e-01 -2.74775833e-01
-2.79309034e-01 -7.11138129e-01 -8.96073580e-01 2.77338177e-01
1.27650642e+00 -7.92282224e-01 -7.92075694e-01 -3.24566305e-01
-8.05681646e-01 -1.74084902e-01 8.31450105e-01 1.79199606e-01
3.33195329e-01 -1.54703522e+00 -1.10762858e+00 -3.87695104e-01
-9.87530574e-02 -1.23187117e-01 2.77975023e-01 7.61055231e-01
-1.07185602e+00 -2.59677589e-01 -6.43115580e-01 -5.99828735e-02
-1.11952496e+00 2.40817606e-01 2.76747942e-01 -1.32321954e-01
-1.31077075e+00 5.38528323e-01 1.28354132e-01 2.95746982e-01
-6.10556722e-01 5.25311306e-02 -5.77047467e-01 6.08372152e-01
9.52478886e-01 5.71721256e-01 -6.92883968e-01 -9.31185305e-01
-9.56800804e-02 1.93053454e-01 -3.85154188e-02 -7.10224152e-01
1.33871961e+00 -3.13717276e-01 -3.65830213e-01 1.37996197e+00
7.11288691e-01 4.84842777e-01 -8.48402321e-01 2.89698631e-01
4.45949212e-02 -5.62028050e-01 -5.53690255e-01 -1.08799970e+00
1.64706856e-01 5.07971048e-01 -1.97096005e-01 8.31916869e-01
7.98756897e-01 3.42657596e-01 6.17059767e-01 -1.44204035e-01
1.50570437e-01 -1.03270280e+00 -1.97823599e-01 3.63910913e-01
1.56303763e+00 -2.03107283e-01 1.70757324e-01 -1.99942321e-01
-7.78937995e-01 1.48661387e+00 -4.30480130e-02 -1.72994703e-01
3.01567554e-01 7.35178590e-01 9.53906626e-02 -4.52028066e-01
-8.84228647e-01 2.40681186e-01 -4.05263186e-01 6.34831786e-01
6.98213100e-01 2.41959393e-01 -1.34722805e+00 7.18653500e-01
-1.00094044e+00 -5.14886640e-02 1.34213412e+00 1.03511715e+00
-4.92020398e-01 -9.59302366e-01 -9.00521874e-01 -4.70983572e-02
-6.81164086e-01 2.39648953e-01 -7.76697338e-01 1.56269753e+00
3.31325561e-01 7.92836010e-01 1.66779011e-01 -1.14791252e-01
3.89743745e-01 2.51138151e-01 6.82550848e-01 -4.09747601e-01
-1.22828782e+00 1.62998497e-01 9.67044413e-01 -6.23363592e-02
-5.24837434e-01 -1.00122404e+00 -1.77087438e+00 -1.15693414e+00
2.10279971e-01 7.50936866e-02 3.62392366e-01 1.05276442e+00
-3.46194774e-01 8.07618499e-01 2.93093681e-01 -5.45343697e-01
1.10782988e-01 -9.92939651e-01 -8.17035139e-01 1.24529116e-01
1.45243913e-01 -3.50870192e-01 -5.90092838e-02 3.65298629e-01] | [11.24326229095459, 0.9449816942214966] |
edbc6b8a-5d9f-49e3-97e1-ba73418e73be | graph4code-a-machine-interpretable-knowledge | 2002.09440 | null | https://arxiv.org/abs/2002.09440v3 | https://arxiv.org/pdf/2002.09440v3.pdf | A Toolkit for Generating Code Knowledge Graphs | Knowledge graphs have been proven extremely useful in powering diverse applications in semantic search and natural language understanding. In this paper, we present GraphGen4Code, a toolkit to build code knowledge graphs that can similarly power various applications such as program search, code understanding, bug detection, and code automation. GraphGen4Code uses generic techniques to capture code semantics with the key nodes in the graph representing classes, functions, and methods. Edges indicate function usage (e.g., how data flows through function calls, as derived from program analysis of real code), and documentation about functions (e.g., code documentation, usage documentation, or forum discussions such as StackOverflow). Our toolkit uses named graphs in RDF to model graphs per program, or can output graphs as JSON. We show the scalability of the toolkit by applying it to 1.3 million Python files drawn from GitHub, 2,300 Python modules, and 47 million forum posts. This results in an integrated code graph with over 2 billion triples. We make the toolkit to build such graphs as well as the sample extraction of the 2 billion triples graph publicly available to the community for use. | ['Jamie McCusker', 'Ibrahim Abdelaziz', 'Julian Dolby', 'Kavitha Srinivas'] | 2020-02-21 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [-4.85484600e-01 3.90919447e-01 -4.69820589e-01 -2.21328929e-01
-3.85586053e-01 -9.96009409e-01 3.60907972e-01 6.90441787e-01
3.49482536e-01 2.97937661e-01 5.91905653e-01 -8.95950794e-01
5.58272451e-02 -1.10151303e+00 -8.31925392e-01 4.50052530e-01
-1.93745241e-01 -2.16934420e-02 5.56480229e-01 -8.66334364e-02
5.88677526e-01 -2.10357025e-01 -1.59241259e+00 3.79655927e-01
8.76279354e-01 4.16504472e-01 1.56793967e-01 5.99787474e-01
-8.08840334e-01 1.02847648e+00 -6.16954923e-01 -5.85619926e-01
-3.17125916e-01 -1.21714965e-01 -1.15101969e+00 -5.00663698e-01
3.40314090e-01 5.04578650e-03 6.44267499e-02 1.28152502e+00
6.96502477e-02 -4.48155671e-01 -1.54126585e-01 -1.47721457e+00
-5.85170329e-01 1.26609218e+00 -3.86881769e-01 2.34675650e-02
8.17054808e-01 -1.57941447e-03 1.10979533e+00 -7.02612281e-01
1.05692005e+00 1.13207591e+00 7.56900251e-01 5.14895320e-01
-1.01233518e+00 -4.26983535e-01 -3.80444169e-01 6.43276423e-02
-9.29088593e-01 -4.22386050e-01 1.90879151e-01 -9.52838898e-01
1.43676579e+00 4.02789921e-01 7.70196974e-01 5.61514914e-01
1.24841034e-02 2.29825079e-01 4.85036820e-01 -4.84532446e-01
9.27645713e-02 1.78055674e-01 7.76570559e-01 1.46483421e+00
5.01476228e-01 -4.67347622e-01 -5.54196656e-01 -1.01111066e+00
1.96591273e-01 -6.27784580e-02 -3.67342055e-01 -2.97183245e-01
-1.02463162e+00 5.67250669e-01 3.82607639e-01 1.21902153e-01
4.18683529e-01 8.65590930e-01 6.71780586e-01 3.46348077e-01
5.36515191e-02 4.62872416e-01 -6.33219957e-01 -5.71954131e-01
-4.64941412e-01 2.42295101e-01 1.52893758e+00 1.44062746e+00
1.42436814e+00 -8.14360827e-02 2.30336487e-01 6.52638376e-01
6.81044161e-01 2.74437100e-01 2.80782104e-01 -8.94923985e-01
5.00309408e-01 1.55820036e+00 -7.79865682e-02 -8.41527998e-01
-2.14938074e-01 -5.28418981e-02 4.81270909e-01 1.46516773e-04
1.27319917e-01 2.23979816e-01 -6.05121315e-01 1.20018494e+00
3.41836601e-01 -2.46717140e-01 -1.56809464e-01 2.44515046e-01
1.34804070e+00 8.94107744e-02 -1.99141335e-02 5.61888099e-01
1.53628480e+00 -8.25364411e-01 -3.18022221e-01 -3.18266660e-01
1.29018664e+00 -6.89647317e-01 1.18907845e+00 -1.23545378e-01
-6.74478531e-01 9.25867110e-02 -7.69341230e-01 -2.11923286e-01
-7.47324646e-01 -3.79825652e-01 8.63836765e-01 7.28414357e-01
-1.30373704e+00 4.39741760e-01 -1.00444043e+00 -6.94613516e-01
4.83633965e-01 -2.73745391e-03 -3.17497432e-01 6.38004169e-02
-5.20445347e-01 5.39583147e-01 4.81899738e-01 -6.63632929e-01
-8.03292036e-01 -1.20243967e+00 -1.17675507e+00 3.81917536e-01
5.22984028e-01 -8.41233313e-01 1.31311154e+00 -6.22743487e-01
-7.06634641e-01 9.77299511e-01 -1.06599219e-01 -1.12997971e-01
-8.28133970e-02 -1.29925266e-01 -3.50986779e-01 -4.37441796e-01
4.80522454e-01 -9.31753963e-02 1.96762502e-01 -9.77995515e-01
-4.31203216e-01 -4.91012245e-01 6.68056190e-01 -5.23196220e-01
-2.86108553e-01 5.13906717e-01 -7.11175323e-01 -6.75206482e-02
-3.01069885e-01 -6.32880092e-01 2.48184055e-01 -1.70594126e-01
-5.65190494e-01 -2.28428736e-01 8.34005833e-01 -9.13678765e-01
1.75725627e+00 -2.18408823e+00 3.08651421e-02 4.08510804e-01
6.30260885e-01 -2.89764285e-01 1.76349074e-01 8.61574829e-01
-1.37114957e-01 9.63655472e-01 -1.79407656e-01 4.01311249e-01
4.78926748e-01 6.72598630e-02 -1.05814070e-01 1.60238519e-01
-1.96706280e-01 9.82035458e-01 -1.12919974e+00 -4.32242930e-01
-1.26543075e-01 1.78181469e-01 -8.29389095e-01 -1.51489660e-01
-8.90558481e-01 -1.10449649e-01 -7.39998877e-01 9.50453222e-01
3.19499016e-01 -7.41061211e-01 5.64787924e-01 4.72598672e-02
-2.86879212e-01 7.22951233e-01 -7.47891307e-01 1.92194045e+00
-6.24620259e-01 7.22584665e-01 8.66069049e-02 -2.26132423e-01
8.09399784e-01 7.72633925e-02 7.30033740e-02 -3.51522654e-01
-4.28580314e-01 1.74335659e-01 -3.30438942e-01 -7.12894976e-01
3.79979849e-01 6.34352684e-01 -4.57691103e-01 8.75180364e-01
-6.06819913e-02 -2.12636262e-01 5.11110961e-01 6.61100090e-01
1.89774215e+00 5.43865025e-01 5.53943276e-01 -5.12790561e-01
2.88767248e-01 5.81750989e-01 3.49783927e-01 4.32209313e-01
5.89557290e-01 -1.58766836e-01 1.18875468e+00 -4.73042667e-01
-9.06096876e-01 -9.00933266e-01 1.14887245e-01 1.23092353e+00
-1.97582524e-02 -1.77070200e+00 -8.08727324e-01 -7.74537444e-01
1.54459760e-01 7.36049116e-01 -3.86662781e-01 -8.72983970e-03
-6.96084082e-01 -2.61494666e-01 7.18742013e-01 5.54902971e-01
2.84418285e-01 -8.63539279e-01 -6.81293488e-01 -4.14766334e-02
-9.56683531e-02 -6.98242188e-01 -3.95068645e-01 -3.12510997e-01
-6.68696165e-01 -1.77872026e+00 2.78253317e-01 -5.62491894e-01
6.64196670e-01 -7.93872476e-02 1.97690403e+00 9.75891650e-01
-5.22749305e-01 8.64221990e-01 -3.72560978e-01 -1.94824874e-01
-6.94235027e-01 -2.95147933e-02 -8.67535532e-01 -1.01685178e+00
3.47080141e-01 -5.97198367e-01 -2.84724891e-01 2.50322044e-01
-9.89516675e-01 9.87769514e-02 -4.23989832e-01 4.27777201e-01
7.63212368e-02 -2.54801065e-01 1.64163169e-02 -1.43394244e+00
6.38207793e-01 -9.84748363e-01 -1.14933705e+00 4.92021710e-01
-5.81172168e-01 3.00026089e-01 5.67473233e-01 4.34276253e-01
-7.15148568e-01 -2.41716728e-01 7.58915320e-02 6.72269166e-02
3.60448003e-01 1.15488589e+00 -2.05189511e-02 -1.60347745e-01
9.94918764e-01 -1.30592987e-01 -1.39850408e-01 -5.97868979e-01
5.65059245e-01 4.91694778e-01 3.45076442e-01 -1.01203525e+00
8.02813053e-01 1.83137074e-01 -3.82310480e-01 -4.05702978e-01
-1.63446814e-01 -5.18263996e-01 -9.62216035e-02 8.62291977e-02
4.57873195e-01 -7.18852460e-01 -8.35035682e-01 -8.41595456e-02
-1.11400342e+00 -5.50261199e-01 -1.07821569e-01 -2.73844689e-01
-3.14278156e-01 3.65703523e-01 -3.88254374e-01 -1.85757428e-01
-3.91680211e-01 -1.26464200e+00 1.03550506e+00 2.49808908e-01
-4.56526488e-01 -1.18762052e+00 2.26229757e-01 3.06600720e-01
5.80728352e-01 3.82324100e-01 1.64457667e+00 -3.59088033e-01
-1.12191033e+00 3.08671948e-02 -3.30407828e-01 -1.69209138e-01
1.08583815e-01 5.57294130e-01 -5.27018845e-01 4.45365906e-02
-7.87588000e-01 -4.52859588e-02 2.36301079e-01 -4.23882097e-01
1.21524370e+00 -4.13631678e-01 -8.07324111e-01 7.10695267e-01
1.78721011e+00 -8.53622258e-02 8.15342486e-01 5.94151735e-01
1.00923884e+00 3.53848279e-01 -7.06943572e-02 4.55812246e-01
7.29347527e-01 4.70126331e-01 7.60797799e-01 3.75142604e-01
-3.40976119e-01 -3.83331001e-01 3.26593965e-01 8.26131761e-01
1.34161785e-01 -1.08074658e-02 -1.52283955e+00 6.57639086e-01
-1.89604127e+00 -7.70507276e-01 -6.46626770e-01 2.21866393e+00
8.12631607e-01 -4.06664163e-01 -1.18247539e-01 -4.94973749e-01
5.42834163e-01 -1.36589736e-01 -4.24430311e-01 -4.16919738e-01
3.76822889e-01 3.05850565e-01 5.41875482e-01 5.65471292e-01
-3.44423771e-01 8.99603367e-01 6.21944427e+00 3.25443059e-01
-6.93383753e-01 2.04888999e-01 -3.84558499e-01 3.43457609e-01
-1.05816829e+00 9.51493740e-01 -7.53138363e-01 5.20173073e-01
1.03046513e+00 -1.03630829e+00 9.78220403e-01 1.16868520e+00
-8.87901038e-02 -2.08763257e-01 -1.25052547e+00 5.93931556e-01
-2.97182500e-01 -1.70012689e+00 -1.45186707e-01 -3.31638716e-02
6.12717211e-01 3.81950378e-01 -6.22702956e-01 2.85452276e-01
9.09110904e-01 -1.00114489e+00 7.74915099e-01 5.56296289e-01
8.60197365e-01 -1.60278261e-01 2.06391796e-01 -5.99172339e-02
-1.36612904e+00 1.53074302e-02 -3.22091281e-02 1.20934881e-01
-3.61339122e-01 6.88394129e-01 -1.22315526e+00 7.64752328e-01
1.02597594e+00 8.57893944e-01 -1.02346444e+00 1.10175920e+00
-2.82624662e-01 5.35528600e-01 -2.90622443e-01 -1.86496466e-01
-4.09410030e-01 8.33806172e-02 3.49310488e-01 1.43593812e+00
3.38927835e-01 -5.65096438e-01 4.74791378e-02 1.25389028e+00
-4.61778432e-01 1.54938579e-01 -7.81902432e-01 -5.10904670e-01
6.94365621e-01 1.22082365e+00 -6.36732638e-01 -5.47103524e-01
-8.62543643e-01 2.80866414e-01 4.67619240e-01 4.85024959e-01
-7.54310906e-01 -8.29202652e-01 8.40569377e-01 4.26636964e-01
1.38266772e-01 -2.39959195e-01 -1.17276255e-02 -1.31148016e+00
4.03270602e-01 -9.79262710e-01 4.26270574e-01 -1.00923657e+00
-8.87257040e-01 2.66243726e-01 1.86297387e-01 -5.46437860e-01
-2.98900664e-01 -4.78617936e-01 -7.71696210e-01 8.40159833e-01
-8.99245262e-01 -6.89881146e-01 -7.39133060e-01 4.46993798e-01
-4.73377667e-02 1.42147332e-01 9.05167460e-01 4.58589435e-01
-3.02297980e-01 1.32889032e-01 -5.10873348e-02 3.23243290e-01
2.65475184e-01 -1.46908259e+00 1.01315832e+00 7.79956460e-01
5.48824221e-02 1.34804118e+00 3.76089156e-01 -8.99449408e-01
-1.85846245e+00 -9.56065655e-01 6.08307123e-01 -7.68257499e-01
1.28852093e+00 -5.90104759e-01 -9.98405874e-01 1.17919922e+00
1.13995291e-01 2.28654325e-01 7.05166101e-01 1.88023150e-01
-9.02560413e-01 7.54649788e-02 -8.22216392e-01 4.92424309e-01
1.43597484e+00 -9.50529337e-01 -3.51110220e-01 6.18653476e-01
6.98661208e-01 -6.86956465e-01 -1.31676555e+00 -1.36086062e-01
4.99115944e-01 -1.00393581e+00 5.89072108e-01 -5.97851574e-01
5.17684937e-01 -7.03332067e-01 -1.21998928e-01 -1.20644546e+00
1.88963771e-01 -8.54310036e-01 -3.34282666e-01 1.19776857e+00
7.30968773e-01 -9.23145413e-01 4.76546705e-01 7.49058843e-01
-4.29748476e-01 -2.39220157e-01 -2.83498794e-01 -4.35045302e-01
-3.95455211e-01 -4.48250264e-01 1.07005060e+00 1.16039228e+00
6.39477670e-01 1.29396424e-01 4.38796490e-01 -2.79341061e-02
3.36671650e-01 3.79419804e-01 1.19501865e+00 -1.23426735e+00
-6.02076113e-01 -4.02970195e-01 -5.62951326e-01 -6.52259648e-01
1.63809597e-01 -1.46319556e+00 -4.72273231e-01 -2.05886865e+00
4.04436916e-01 -5.50555885e-01 5.60023248e-01 1.13944888e+00
-4.14342508e-02 -3.79782110e-01 -9.09462422e-02 2.59523839e-01
-5.49794018e-01 -1.24605566e-01 6.75642431e-01 -2.47785315e-01
9.97220129e-02 -3.70898515e-01 -9.19846356e-01 6.60459042e-01
6.44198239e-01 -7.60804117e-01 -3.29267740e-01 -7.35009074e-01
9.78544474e-01 1.28452748e-01 5.79463005e-01 -7.27757037e-01
2.96908051e-01 1.11306988e-01 -3.41571599e-01 3.93530689e-02
-2.93197513e-01 -6.20291650e-01 8.90769839e-01 4.36320424e-01
5.92500828e-02 2.26364687e-01 5.13981223e-01 3.73821855e-01
-1.21090353e-01 -7.10372627e-01 1.27671519e-02 -8.49581420e-01
-1.01417458e+00 -8.71615950e-03 1.72622904e-01 4.57062095e-01
8.91483068e-01 -1.84830800e-01 -1.44962466e+00 9.79130492e-02
-5.32466233e-01 5.17287493e-01 1.31329668e+00 7.46236861e-01
2.02282354e-01 -1.00653839e+00 -3.05999994e-01 1.91562027e-02
6.18409872e-01 -4.17702705e-01 -3.10664773e-01 6.29461408e-01
-9.99713898e-01 6.97919028e-03 7.91905597e-02 -5.11300087e-01
-1.34749198e+00 3.63764733e-01 3.13845426e-01 7.32471794e-02
-6.62406206e-01 1.90778166e-01 -6.43562302e-02 -7.55842626e-01
-3.72728318e-01 -4.82486576e-01 2.77752101e-01 -5.67813277e-01
6.05667174e-01 5.05483270e-01 2.46932209e-01 2.78227404e-02
-7.91220903e-01 6.11912489e-01 3.08461696e-01 3.34724098e-01
1.31345153e+00 2.95738310e-01 -1.01279700e+00 4.51826066e-01
1.16799307e+00 7.09905326e-01 -7.41476536e-01 9.97844189e-02
4.73135978e-01 -6.31997287e-01 -3.70386243e-01 -9.44864094e-01
-9.71002400e-01 3.57217997e-01 -7.98191950e-02 8.27277422e-01
5.04412949e-01 6.16305709e-01 2.89012939e-01 3.49950016e-01
1.03172779e+00 -3.86797816e-01 -3.01125497e-01 4.60075915e-01
5.86951315e-01 -7.57144928e-01 1.19743101e-01 -6.96940362e-01
2.29359165e-01 1.47023439e+00 4.47477579e-01 2.91055888e-01
5.13792694e-01 5.28848886e-01 -2.01911584e-01 -8.84335995e-01
-7.65498757e-01 -3.09348386e-02 -1.01883456e-01 6.02911890e-01
8.68378937e-01 6.22684695e-02 -2.16347426e-01 2.85554737e-01
-1.69554114e-01 2.63717890e-01 9.40735221e-01 1.27847838e+00
-3.12569767e-01 -1.43709290e+00 -3.50085437e-01 7.21992314e-01
-6.27389431e-01 -5.31300724e-01 -5.17451108e-01 9.56476271e-01
-1.42289385e-01 7.44947970e-01 -1.38047889e-01 -2.73169994e-01
3.41200173e-01 1.02886647e-01 5.43743193e-01 -1.19051492e+00
-8.89679611e-01 -7.15995610e-01 7.12375104e-01 -1.00126934e+00
-1.78049818e-01 -1.65210173e-01 -1.73080862e+00 -4.71201062e-01
-2.32538924e-01 3.12847048e-01 8.90294194e-01 4.81603384e-01
8.88135195e-01 5.02287328e-01 -2.29810029e-01 2.86270231e-02
9.05304626e-02 -6.71566963e-01 -9.31243524e-02 3.49079490e-01
-1.09546989e-01 -5.05984128e-01 -2.97924250e-01 4.45968539e-01] | [7.61470365524292, 7.8877644538879395] |
0ad9f17e-1d8c-491f-a590-5485c582ec67 | two-stream-binocular-network-accurate-near | 1804.10160 | null | http://arxiv.org/abs/1804.10160v1 | http://arxiv.org/pdf/1804.10160v1.pdf | Two-Stream Binocular Network: Accurate Near Field Finger Detection Based On Binocular Images | Fingertip detection plays an important role in human computer interaction.
Previous works transform binocular images into depth images. Then depth-based
hand pose estimation methods are used to predict 3D positions of fingertips.
Different from previous works, we propose a new framework, named Two-Stream
Binocular Network (TSBnet) to detect fingertips from binocular images directly.
TSBnet first shares convolutional layers for low level features of right and
left images. Then it extracts high level features in two-stream convolutional
networks separately. Further, we add a new layer: binocular distance
measurement layer to improve performance of our model. To verify our scheme, we
build a binocular hand image dataset, containing about 117k pairs of images in
training set and 10k pairs of images in test set. Our methods achieve an
average error of 10.9mm on our test set, outperforming previous work by 5.9mm
(relatively 35.1%). | ['Hengkai Guo', 'Huazhong Yang', 'Cairong Zhang', 'Xinghao Chen', 'Yi Wei', 'Guijin Wang'] | 2018-04-26 | null | null | null | null | ['fingertip-detection'] | ['computer-vision'] | [-2.02366337e-01 -4.87970412e-01 -1.45288324e-02 -2.79240817e-01
-2.26358071e-01 -7.24878788e-01 1.66112736e-01 -6.82073236e-01
-5.58642030e-01 4.46540385e-01 -8.63132551e-02 -1.58478573e-01
2.58514643e-01 -5.71609080e-01 -7.37776041e-01 -3.78991812e-01
1.91169336e-01 -1.88088287e-02 7.21320271e-01 1.56780295e-02
6.78522885e-01 8.67210567e-01 -1.42173684e+00 3.71645063e-01
2.87339538e-01 1.16258609e+00 4.65266049e-01 9.62610364e-01
2.35359311e-01 4.02065426e-01 -5.95370352e-01 -3.39218736e-01
7.40607560e-01 1.97924435e-01 -5.65573514e-01 -1.22259073e-01
6.70298219e-01 -1.10300493e+00 -9.19682622e-01 7.96806991e-01
1.20149100e+00 -3.37451428e-01 4.87503052e-01 -1.14993751e+00
-5.64935684e-01 9.56724286e-02 -8.33225846e-01 -9.43288300e-03
5.74380696e-01 4.85069424e-01 2.95257241e-01 -1.22966158e+00
4.02147591e-01 1.31294477e+00 5.58323324e-01 5.45317352e-01
-5.89865327e-01 -8.76477122e-01 6.63906932e-02 5.21789007e-02
-1.51568317e+00 -2.21671060e-01 7.04358935e-01 -4.26563203e-01
9.64820564e-01 1.81808427e-01 6.19898975e-01 9.76841331e-01
2.42688671e-01 1.04718208e+00 1.15847540e+00 -3.50655913e-01
-4.41047698e-01 -1.08159296e-02 -1.97972078e-02 5.83217025e-01
1.59879118e-01 1.91316396e-01 -5.35255730e-01 2.80013472e-01
1.67896879e+00 2.90493757e-01 -2.07204163e-01 2.98634730e-02
-1.23426700e+00 9.66855325e-03 6.95341587e-01 7.51540298e-04
-3.81642163e-01 5.08932360e-02 8.14892501e-02 1.58637762e-01
-1.65926993e-01 3.80710885e-02 -5.35459042e-01 -3.09545159e-01
-3.99333149e-01 4.46255803e-01 3.85706633e-01 1.41143525e+00
2.41008505e-01 -5.85615396e-01 -5.40354729e-01 7.44974673e-01
4.79535490e-01 7.44271874e-01 -5.09997755e-02 -9.18358922e-01
9.59491968e-01 5.54012418e-01 4.12487477e-01 -8.64281297e-01
-2.69353271e-01 -1.23237796e-01 -5.97354352e-01 4.57110584e-01
7.77778745e-01 -2.88488626e-01 -9.15171385e-01 1.18896794e+00
-1.56323742e-02 -3.54368359e-01 -6.08350635e-01 1.34882140e+00
7.07523346e-01 1.13528155e-01 -5.03405094e-01 3.83560330e-01
1.30246377e+00 -9.61639822e-01 -4.37888235e-01 -4.25514504e-02
-1.64773405e-01 -1.16698337e+00 1.36825502e+00 7.98906446e-01
-1.13744712e+00 -8.78462076e-01 -7.31465936e-01 -2.97216117e-01
-2.67713487e-01 5.86623013e-01 6.82070434e-01 5.90183616e-01
-9.43024158e-01 3.42831850e-01 -5.54874182e-01 -4.79775779e-02
4.39828992e-01 8.06224227e-01 -3.20075363e-01 -1.48786763e-02
-9.14027214e-01 5.16222894e-01 -3.60722840e-02 4.25944507e-01
-3.72233361e-01 -3.46900731e-01 -5.02797961e-01 -2.36883700e-01
1.03948995e-01 -5.06695747e-01 1.43315709e+00 -2.24216525e-02
-1.66061103e+00 9.96322811e-01 -6.71819210e-01 2.86070734e-01
8.25760901e-01 -4.22146618e-01 7.99229294e-02 5.73410504e-02
-1.30140521e-02 7.75179327e-01 8.50865722e-01 -1.15285778e+00
-8.08456361e-01 -9.11885262e-01 2.35204950e-01 -7.24979397e-03
-2.40460217e-01 1.97678715e-01 -1.05139995e+00 -6.09123170e-01
3.42974544e-01 -8.34008634e-01 1.71313614e-01 2.27907166e-01
-8.06982160e-01 -5.00891864e-01 6.66048467e-01 -6.76312804e-01
1.03223860e+00 -1.59416187e+00 -6.52325600e-02 1.59380063e-01
5.56298077e-01 3.21757585e-01 -7.48216137e-02 1.03117459e-01
2.87490875e-01 -2.76145667e-01 4.21534836e-01 -5.56673527e-01
-2.67737024e-02 -2.70032197e-01 -1.44373149e-01 -1.16076395e-02
6.57182783e-02 1.05004942e+00 -6.36937320e-01 -5.60484171e-01
5.56502521e-01 7.43420184e-01 -4.11082536e-01 3.14434499e-01
3.92879337e-01 4.80767787e-01 -2.33660355e-01 1.12474096e+00
1.22009206e+00 -5.24274483e-02 -2.47191906e-01 -4.38580453e-01
-4.07249898e-01 1.27896845e-01 -9.82866406e-01 1.75826645e+00
-4.06616718e-01 7.71651447e-01 -1.44228250e-01 -5.84232472e-02
7.86735654e-01 2.01404035e-01 2.22448334e-01 -6.34730697e-01
5.12615561e-01 1.16275474e-01 6.42453879e-02 -5.56912243e-01
2.92113096e-01 4.67281252e-01 2.76485711e-01 2.84693450e-01
-1.27272561e-01 1.48018211e-01 -3.99631917e-01 -1.60808757e-01
6.91904187e-01 5.84246442e-02 -3.33599299e-01 2.29039282e-01
4.55025554e-01 -5.42560339e-01 1.18496291e-01 8.68996322e-01
-4.99262661e-01 1.12502599e+00 5.38916469e-01 -4.10953313e-01
-1.14675653e+00 -1.12590706e+00 -1.86639249e-01 7.13818312e-01
3.23339731e-01 -2.40335390e-01 -8.66682291e-01 -4.46883351e-01
4.60172802e-01 -4.83009368e-01 -5.73991537e-01 3.75777751e-01
-7.08793759e-01 -1.49207070e-01 4.76655424e-01 1.19303799e+00
1.01083219e+00 -1.04322970e+00 -6.03273571e-01 -1.94438919e-01
1.07399642e-01 -1.21145737e+00 -8.49764705e-01 -3.33586842e-01
-5.92403889e-01 -1.35004032e+00 -1.37713802e+00 -9.56440151e-01
6.59120500e-01 5.59196174e-01 5.68880916e-01 -2.77745306e-01
-5.31916440e-01 3.16939466e-02 -1.06238797e-01 -6.88101828e-01
6.60412312e-01 4.53164339e-01 2.37721592e-01 -2.82764286e-01
6.61904454e-01 -5.78150272e-01 -1.19295049e+00 5.94436467e-01
-5.79864345e-02 -7.73361400e-02 7.27063656e-01 6.39972210e-01
2.73963064e-01 -5.74981630e-01 9.27241370e-02 5.71363643e-02
8.09172630e-01 4.15353388e-01 -6.74707353e-01 1.29266605e-01
-3.28365624e-01 -1.93858117e-01 3.89043510e-01 -4.88003790e-01
-8.02189350e-01 2.90392458e-01 -2.21360326e-01 -5.85078835e-01
-1.65378496e-01 -3.17127764e-01 -1.80994347e-01 -4.09581691e-01
4.54345703e-01 2.09215730e-01 6.13828115e-02 -7.43260562e-01
-5.33430986e-02 1.50934958e+00 7.00488746e-01 -3.53301078e-01
5.67198932e-01 3.89806926e-01 -4.18622494e-02 -5.45318842e-01
-4.27350193e-01 -3.75785023e-01 -1.17949796e+00 -3.37545067e-01
5.19155264e-01 -6.58377469e-01 -1.86996281e+00 1.22305763e+00
-1.65959060e+00 -1.35685861e-01 4.98470694e-01 5.69342494e-01
-2.41208747e-01 3.34646702e-01 -7.43210495e-01 -1.09009612e+00
-4.74964917e-01 -1.28595567e+00 1.58180535e+00 3.13109189e-01
6.94321841e-02 -4.32737112e-01 -4.09349769e-01 1.93986341e-01
1.77620366e-01 -1.77329764e-01 2.43859708e-01 2.87903100e-01
-6.02519929e-01 -7.11244106e-01 -8.72284949e-01 3.08439970e-01
5.51836312e-01 -1.67119771e-01 -1.27937686e+00 -3.10958058e-01
-3.81317377e-01 -1.05101287e-01 6.61210239e-01 6.30841076e-01
1.62566590e+00 2.68737555e-01 -4.45801347e-01 5.33180654e-01
9.67328072e-01 3.19108695e-01 8.69606555e-01 2.84074664e-01
9.81719434e-01 5.66858411e-01 5.23687661e-01 5.61430991e-01
5.26481450e-01 8.54473591e-01 3.54556620e-01 -7.52346292e-02
-1.85959920e-01 -2.07811564e-01 -2.43811291e-02 4.02987331e-01
-8.54418993e-01 1.13021649e-01 -9.40737605e-01 1.02779470e-01
-1.62531781e+00 -6.39458299e-01 -1.51776344e-01 2.25353694e+00
6.99146211e-01 1.99888021e-01 5.84221780e-01 3.93418312e-01
7.81460762e-01 -2.39559814e-01 -8.32019091e-01 7.72442222e-02
3.50491405e-02 3.47631395e-01 5.98882556e-01 5.85454702e-01
-1.13372302e+00 1.09697676e+00 6.41402435e+00 2.38229796e-01
-1.19849467e+00 -3.68637413e-01 1.49779528e-01 -2.92734653e-01
3.18421543e-01 -4.71800804e-01 -1.15083766e+00 7.37331867e-01
-9.66389552e-02 4.85080272e-01 5.58609962e-01 7.27294743e-01
1.62499398e-01 -5.77267259e-02 -1.07192016e+00 1.64184380e+00
4.81888875e-02 -8.21138859e-01 -1.69895932e-01 1.89878851e-01
3.36843371e-01 -9.69672203e-02 2.89598584e-01 -1.01671048e-01
-3.72267276e-01 -1.21976531e+00 4.63549674e-01 7.32578158e-01
1.10249245e+00 -6.75491095e-01 8.76155674e-01 4.11950707e-01
-1.34998298e+00 1.07418476e-02 -4.74392474e-01 -4.89632487e-01
-1.59093633e-01 2.43390664e-01 -7.22978711e-01 3.46881077e-02
1.01569903e+00 6.32442653e-01 -6.07006490e-01 9.95017469e-01
-2.54832566e-01 -2.09601119e-01 -3.14607590e-01 -1.87436327e-01
-1.66265562e-01 2.83231437e-01 2.27278143e-01 1.01560450e+00
-3.20377350e-02 -8.46947655e-02 -2.34361455e-01 8.44498217e-01
3.40269500e-04 -4.17720705e-01 -4.90183085e-01 2.85212368e-01
8.13785434e-01 9.99557078e-01 -2.13530675e-01 -2.02646419e-01
-2.81374067e-01 1.64420176e+00 -2.89180670e-02 3.06101203e-01
-4.92450327e-01 -1.37359011e+00 7.92161942e-01 1.15941815e-01
1.81424528e-01 -5.78463435e-01 -6.15116298e-01 -1.29487073e+00
8.29025030e-01 -5.43544769e-01 -3.69453222e-01 -9.21654820e-01
-1.19138062e+00 5.26923120e-01 -4.67030793e-01 -1.34537733e+00
-1.48726687e-01 -1.47512472e+00 -4.29381847e-01 1.48896682e+00
-1.49799049e+00 -1.21484327e+00 -1.15796924e+00 9.47496355e-01
4.77966547e-01 -3.27263832e-01 5.53312600e-01 2.53180414e-01
-4.73210067e-01 1.13903892e+00 -1.99629441e-01 7.28097022e-01
1.07947814e+00 -1.22097051e+00 1.03614128e+00 4.13830817e-01
-4.63575423e-01 1.21159613e+00 8.61530676e-02 -4.09434199e-01
-1.56533539e+00 -5.35764515e-01 8.27127814e-01 -9.94911075e-01
1.54707981e-02 -6.71256483e-01 -2.99872607e-01 6.17021084e-01
-5.57703450e-02 1.58918902e-01 1.11442767e-01 5.02421893e-02
-4.80122983e-01 -1.12554140e-01 -1.17780113e+00 6.66583061e-01
1.68141448e+00 -8.13606620e-01 -4.09490526e-01 1.42978922e-01
4.16129231e-01 -8.88804018e-01 -8.40731978e-01 4.30201590e-01
1.70181453e+00 -1.16163945e+00 1.19584978e+00 -4.02341306e-01
3.78389001e-01 -1.62879601e-01 -1.19221814e-01 -7.30091751e-01
-2.79947460e-01 -4.34890598e-01 -2.80248851e-01 7.34638870e-01
-5.46201430e-02 -7.36627221e-01 1.25622046e+00 4.66901541e-01
2.55244404e-01 -1.00935483e+00 -5.57955980e-01 -8.40787053e-01
1.40814841e-01 -4.00148094e-01 7.37935007e-01 2.57109463e-01
1.55874684e-01 1.24669522e-02 -2.65266031e-01 1.67325810e-01
9.01888609e-01 1.17190495e-01 1.05863094e+00 -1.30999434e+00
-3.32956053e-02 -4.95659888e-01 -6.77690208e-01 -1.88090289e+00
-2.73705035e-01 -2.64711678e-01 -2.03161225e-01 -1.30275941e+00
2.89353937e-01 -2.31249392e-01 -2.23392263e-01 3.81586283e-01
-8.75603259e-02 6.16399467e-01 1.91390425e-01 3.36294442e-01
8.77217427e-02 2.90338583e-02 1.82865095e+00 7.92358536e-03
-2.38352418e-01 3.08307350e-01 -4.08279926e-01 5.78907609e-01
9.54739690e-01 3.31011593e-01 -1.33901522e-01 -7.52057552e-01
-2.64586478e-01 3.65328160e-03 8.91246855e-01 -8.92535448e-01
4.28711265e-01 -5.43913208e-02 1.01514447e+00 -1.12295103e+00
3.89569581e-01 -4.87770289e-01 -7.82759726e-01 4.57956970e-01
-7.78513998e-02 1.86498463e-02 1.44674391e-01 -1.41142637e-01
-3.38176861e-02 2.15738788e-01 4.03620541e-01 -1.16945915e-01
-6.28233016e-01 6.48635387e-01 -5.42284921e-02 -5.48615754e-01
7.33573556e-01 -7.44549274e-01 -3.09136003e-01 -3.37004185e-01
-5.21301031e-01 1.41752288e-01 2.44994044e-01 6.92990899e-01
1.02281344e+00 -1.42826664e+00 -2.83984542e-01 5.83967626e-01
6.99546305e-04 2.27667287e-01 5.96008301e-02 7.44942725e-01
-4.89227802e-01 1.06778598e+00 -5.28641462e-01 -7.83149600e-01
-1.37123919e+00 1.62763640e-01 5.02025187e-01 6.06048763e-01
-5.01843095e-01 1.26929617e+00 1.77044738e-02 -4.77494150e-01
7.70015717e-01 -6.59136057e-01 1.26898661e-01 -4.91302729e-01
6.45218670e-01 4.18625474e-01 -2.49505360e-02 -1.99329719e-01
-4.78885829e-01 1.14805448e+00 -2.55922735e-01 -1.72551021e-01
9.14528906e-01 1.26411021e-01 -1.06867902e-01 2.10312128e-01
1.27122068e+00 9.44191962e-02 -1.57500553e+00 -4.24463749e-02
-4.57021415e-01 -1.16858506e+00 -1.57135025e-01 -9.41027403e-01
-9.21759963e-01 1.53902435e+00 1.00364161e+00 -3.54941726e-01
9.67055559e-01 -6.35489151e-02 9.77167249e-01 5.32307208e-01
7.04243124e-01 -1.00314510e+00 2.95199275e-01 7.09487140e-01
1.24552560e+00 -1.57049704e+00 -4.23012942e-01 -6.33080304e-01
-2.02872232e-01 1.30183816e+00 1.13137174e+00 -1.79590419e-01
5.78486323e-01 5.85635066e-01 8.79988670e-02 -7.64400586e-02
-7.39846975e-02 -2.30735883e-01 4.31970567e-01 7.60920286e-01
6.25247955e-01 9.63286683e-03 7.52719715e-02 7.06163883e-01
-3.20412248e-01 3.10332507e-01 -1.99643999e-01 1.07398999e+00
-2.71863699e-01 -1.22636449e+00 -4.10163760e-01 4.08685565e-01
2.94866320e-02 -8.43503699e-02 -7.23968327e-01 7.71520019e-01
2.32535318e-01 7.39484131e-01 2.59860188e-01 -1.29408681e+00
7.26514399e-01 -2.61930048e-01 1.33473074e+00 -4.73865211e-01
-4.12972987e-01 -8.72100983e-03 -6.34547710e-01 -5.90727389e-01
-1.39305750e-02 -3.97570193e-01 -9.55079556e-01 -5.13045788e-01
-2.28251040e-01 -4.67644900e-01 6.30572498e-01 7.17901409e-01
4.54883844e-01 3.61903012e-01 7.74155378e-01 -1.40867937e+00
-5.05724072e-01 -1.20366454e+00 -6.41734719e-01 -1.59946054e-01
4.05897141e-01 -9.26398695e-01 5.93668856e-02 -3.56258303e-01] | [6.504888534545898, -0.48654162883758545] |
ae7a69ab-6c83-4a21-810c-176ac18f165a | stir-siamese-transformer-for-image-retrieval | 2304.13393 | null | https://arxiv.org/abs/2304.13393v2 | https://arxiv.org/pdf/2304.13393v2.pdf | STIR: Siamese Transformer for Image Retrieval Postprocessing | Current metric learning approaches for image retrieval are usually based on learning a space of informative latent representations where simple approaches such as the cosine distance will work well. Recent state of the art methods such as HypViT move to more complex embedding spaces that may yield better results but are harder to scale to production environments. In this work, we first construct a simpler model based on triplet loss with hard negatives mining that performs at the state of the art level but does not have these drawbacks. Second, we introduce a novel approach for image retrieval postprocessing called Siamese Transformer for Image Retrieval (STIR) that reranks several top outputs in a single forward pass. Unlike previously proposed Reranking Transformers, STIR does not rely on global/local feature extraction and directly compares a query image and a retrieved candidate on pixel level with the usage of attention mechanism. The resulting approach defines a new state of the art on standard image retrieval datasets: Stanford Online Products and DeepFashion In-shop. We also release the source code at https://github.com/OML-Team/open-metric-learning/tree/main/pipelines/postprocessing/ and an interactive demo of our approach at https://dapladoc-oml-postprocessing-demo-srcappmain-pfh2g0.streamlit.app/ | ['Sergey Nikolenko', 'Aleksei Tarasov', 'Aleksei Shabanov'] | 2023-04-26 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [ 9.56292897e-02 -4.09578562e-01 -1.69597015e-01 -4.38865185e-01
-1.39596510e+00 -4.90343392e-01 7.73975313e-01 4.87836868e-01
-4.39648330e-01 2.71358430e-01 1.53301567e-01 -1.72047660e-01
-5.49817622e-01 -7.35436797e-01 -5.68746448e-01 -6.66732252e-01
-1.22019447e-01 4.56158340e-01 2.69834548e-01 -2.30270416e-01
4.46852982e-01 4.37569499e-01 -1.79765093e+00 5.65915525e-01
5.20805240e-01 1.21824467e+00 2.86617398e-01 8.98643911e-01
1.62332505e-02 6.10406876e-01 -3.46001953e-01 -3.86397600e-01
7.20519662e-01 -2.78058141e-01 -8.12200010e-01 -3.99967343e-01
6.42162263e-01 -4.80567217e-01 -4.81708735e-01 1.02118909e+00
7.90798068e-01 -2.06082836e-01 6.80091381e-01 -1.39308918e+00
-8.12892199e-01 5.68204939e-01 -5.51632762e-01 1.65355191e-01
2.85080940e-01 5.18177524e-02 1.44591141e+00 -1.06905234e+00
6.13802135e-01 1.24363780e+00 4.28379774e-01 1.04279153e-01
-1.30351388e+00 -6.07063830e-01 -1.19751431e-01 6.88435733e-01
-1.39627147e+00 -3.52542996e-01 6.10472381e-01 -3.41388136e-01
8.20773005e-01 3.35952699e-01 2.33460605e-01 9.14197505e-01
1.20336853e-03 1.23423994e+00 1.14515662e+00 -4.69628364e-01
-2.30261218e-03 2.83879098e-02 4.53771561e-01 8.45493674e-01
-1.05853770e-02 8.92317966e-02 -6.10667050e-01 -2.20088467e-01
3.78729254e-01 1.67974249e-01 -2.56431550e-01 -6.08508348e-01
-1.50203276e+00 1.05682337e+00 6.03865147e-01 3.91472578e-01
-3.60318959e-01 3.78752440e-01 5.87771058e-01 8.15317392e-01
4.78494734e-01 4.18983281e-01 -3.93516481e-01 1.54272467e-01
-1.41666460e+00 4.02285695e-01 6.92811191e-01 8.65185142e-01
9.14642096e-01 -5.32471180e-01 -3.89633179e-01 7.82109678e-01
3.25167894e-01 2.71800220e-01 6.63597763e-01 -9.34065580e-01
2.23365843e-01 4.03851002e-01 -3.92633677e-02 -1.00783575e+00
-4.40252349e-02 -3.02947372e-01 -3.58320922e-01 2.51170546e-01
1.95401952e-01 3.09225887e-01 -9.43200707e-01 1.26005018e+00
1.16414800e-01 -1.77982107e-01 -2.08359331e-01 1.07673109e+00
6.96466088e-01 7.00425982e-01 -2.27993265e-01 9.90812629e-02
1.17071307e+00 -1.28093123e+00 -3.34331900e-01 2.52598733e-01
6.76235557e-01 -1.17905331e+00 1.22375786e+00 6.95036292e-01
-1.04327059e+00 -5.26349247e-01 -1.45615733e+00 -4.78814214e-01
-8.48903835e-01 1.76121429e-01 6.72819853e-01 3.61715674e-01
-1.28931451e+00 1.10101342e+00 -7.64407396e-01 -5.60852945e-01
4.64452505e-01 3.02841395e-01 -2.71480203e-01 -3.14711660e-01
-1.07927048e+00 7.92592824e-01 2.77042925e-01 -1.78227827e-01
-1.06609595e+00 -7.22571790e-01 -6.60438836e-01 5.84271550e-02
2.44461447e-01 -6.38304293e-01 1.26606321e+00 -9.07250941e-01
-1.36067545e+00 1.02311528e+00 1.65412083e-01 -6.81759238e-01
7.14941919e-01 -5.75051785e-01 -1.43553421e-01 4.16625500e-01
1.38865128e-01 1.04094625e+00 9.02933002e-01 -9.21096742e-01
-5.82927763e-01 -3.97742152e-01 1.04017131e-01 1.37606353e-01
-4.16107506e-01 1.47459507e-01 -5.87328374e-01 -6.04610085e-01
3.94023731e-02 -8.51009548e-01 -7.90180638e-03 4.86002058e-01
-2.72114247e-01 -2.81019777e-01 7.21136093e-01 -3.62472355e-01
1.25757539e+00 -2.08231783e+00 7.67888576e-02 1.69447809e-01
1.13265201e-01 1.92685083e-01 -5.53607881e-01 8.52744579e-01
-2.09169731e-01 4.53881845e-02 -1.73741803e-01 -4.70345974e-01
3.87495875e-01 -1.81653813e-01 -2.83468276e-01 7.46440947e-01
6.93074092e-02 9.32664752e-01 -1.04826748e+00 -6.83610916e-01
3.07465523e-01 4.69462752e-01 -3.43246847e-01 2.98061576e-02
-2.53856093e-01 -3.01642865e-01 -3.04986268e-01 9.06645179e-01
6.20749354e-01 -3.60833555e-01 -2.70994246e-01 -3.68316263e-01
-2.45011061e-01 3.74780953e-01 -1.29046547e+00 2.19370604e+00
-3.86013150e-01 7.24814832e-01 -3.78786474e-02 -1.05956209e+00
7.17195034e-01 2.52838172e-02 7.86246061e-01 -7.73003817e-01
9.41253603e-02 3.94287914e-01 -3.62540752e-01 -3.48152608e-01
7.18948781e-01 4.78068918e-01 -9.85863991e-03 6.13450110e-01
2.63893396e-01 2.45532533e-03 4.56205547e-01 4.40671235e-01
1.34356654e+00 5.17897546e-01 2.21965939e-01 -2.71941900e-01
3.46158803e-01 1.67169854e-01 1.31476626e-01 8.14010859e-01
-1.17613517e-01 9.17621017e-01 4.14992332e-01 -1.17213704e-01
-1.14520466e+00 -1.22249484e+00 -2.88995504e-01 1.43085098e+00
2.16745213e-02 -7.06025064e-01 -4.54340070e-01 -6.24738991e-01
2.30075106e-01 4.26035762e-01 -4.55447793e-01 -1.63207781e-02
-1.78516686e-01 -6.16031945e-01 4.83670890e-01 1.51148632e-01
3.76377881e-01 -1.01136494e+00 -5.66903472e-01 -1.39880078e-02
1.55316040e-01 -6.12726629e-01 -4.89635527e-01 3.33705008e-01
-8.14519107e-01 -1.06643033e+00 -9.62022662e-01 -5.61598539e-01
4.61062670e-01 3.23779255e-01 1.12686992e+00 5.98046817e-02
-7.09496558e-01 4.29954320e-01 -6.57722354e-01 -3.66221845e-01
-8.80774781e-02 2.93542355e-01 -1.90083355e-01 1.02871591e-02
5.74887812e-01 -4.17650580e-01 -1.08665395e+00 2.46199682e-01
-1.07651269e+00 -1.58557057e-01 7.92276800e-01 1.00103259e+00
6.93417430e-01 -2.70884961e-01 1.97666883e-01 -8.18372190e-01
6.71372294e-01 -5.60164988e-01 -8.19413364e-01 3.04671019e-01
-9.72622752e-01 3.30620289e-01 2.18509197e-01 -1.89664170e-01
-4.53884125e-01 1.48140013e-01 9.64716598e-02 -7.74883091e-01
4.01932970e-02 5.39418399e-01 1.98605388e-01 2.04099581e-01
8.32614183e-01 1.62577495e-01 5.75488470e-02 -6.32957101e-01
6.06969535e-01 7.43430853e-01 1.42802998e-01 -3.09582084e-01
9.10769045e-01 5.75172067e-01 -2.42316365e-01 -4.82620507e-01
-6.42453730e-01 -9.79472160e-01 -5.42015731e-01 1.41951501e-01
6.80426359e-01 -7.77801275e-01 -4.77629662e-01 2.95091957e-01
-8.89488220e-01 -2.28742614e-01 -3.98912132e-01 4.75889444e-01
-6.64778233e-01 5.21741748e-01 -6.73328757e-01 -4.65016127e-01
-5.92704952e-01 -1.15043092e+00 1.33315349e+00 -1.13299593e-01
-1.84967741e-02 -5.97790122e-01 1.96747944e-01 2.28955030e-01
6.82814658e-01 -4.31779101e-02 7.33770788e-01 -8.47966731e-01
-9.07426536e-01 -2.04002440e-01 -4.57380861e-01 5.99116623e-01
-1.48654684e-01 5.09120226e-02 -1.09210336e+00 -2.90332854e-01
-3.85289252e-01 -5.55767357e-01 1.33366358e+00 2.45113820e-01
1.07565570e+00 -8.46449360e-02 -2.23054424e-01 6.03091896e-01
1.68821907e+00 -3.34352612e-01 7.40952849e-01 5.59240401e-01
2.47012734e-01 5.59367955e-01 9.06166017e-01 3.73261482e-01
2.01660007e-01 6.37934983e-01 5.07649541e-01 -8.83295611e-02
-2.11312473e-01 -2.42510680e-02 6.34363651e-01 7.24076927e-01
4.15365010e-01 -2.75931984e-01 -6.73161805e-01 8.49602878e-01
-2.09309936e+00 -9.65895295e-01 9.06404629e-02 2.28113437e+00
7.39401460e-01 -1.23328418e-02 -5.79231381e-02 2.33142748e-01
4.27319616e-01 3.75431240e-01 -3.67286474e-01 -3.25510651e-01
3.23272571e-02 4.43003446e-01 7.44565964e-01 6.13366067e-01
-1.22442901e+00 9.24894512e-01 4.65859318e+00 1.01080477e+00
-1.06571853e+00 3.31773192e-01 4.00168091e-01 -3.92157078e-01
-2.61124521e-01 1.31989285e-01 -7.40281940e-01 3.70130956e-01
8.98031235e-01 -1.12242840e-01 2.62218148e-01 8.06072950e-01
4.12464105e-02 -2.44379714e-01 -1.33428919e+00 1.15005350e+00
1.28741816e-01 -1.23658645e+00 1.18048266e-01 3.97023745e-02
3.71213704e-01 5.58206975e-01 2.97488093e-01 2.48444602e-01
1.76996619e-01 -7.55282640e-01 6.89934313e-01 6.35489285e-01
6.99187875e-01 -3.60223860e-01 6.31856740e-01 -7.29706213e-02
-8.94343734e-01 -6.23313114e-02 -4.88947153e-01 3.66182357e-01
3.17073539e-02 6.58666193e-01 -8.17732811e-01 5.81847250e-01
8.87143314e-01 8.38778436e-01 -7.97916234e-01 1.20110071e+00
7.92140961e-02 3.53334248e-01 -5.74118018e-01 8.18964466e-02
3.09900939e-01 -1.38401622e-02 5.62004507e-01 1.42940533e+00
4.50303853e-01 -4.70853269e-01 -2.88746003e-02 5.89684606e-01
-1.00967713e-01 3.44474196e-01 -6.71707928e-01 -6.94391951e-02
2.51664251e-01 1.45639682e+00 -5.76058984e-01 -2.81224340e-01
-3.66902560e-01 1.22586119e+00 1.23555653e-01 1.86536431e-01
-7.70737827e-01 -6.45163655e-01 5.62384427e-01 3.07611674e-01
5.16255677e-01 -1.94892779e-01 1.89728558e-01 -1.09163630e+00
1.06883854e-01 -8.36687386e-01 6.16875112e-01 -8.21591020e-01
-1.44144690e+00 4.40775335e-01 8.32555741e-02 -1.41151404e+00
-2.21064717e-01 -6.93040133e-01 -2.07093105e-01 6.85338140e-01
-1.86066818e+00 -1.15042496e+00 -2.70731539e-01 6.83992088e-01
6.89615786e-01 -3.01487803e-01 9.65388417e-01 5.84025502e-01
-2.38847006e-02 5.90461969e-01 3.28530699e-01 -3.50634754e-02
1.13619506e+00 -1.46142268e+00 1.49146035e-01 5.85338116e-01
5.34387231e-01 4.64950174e-01 5.77193499e-01 -2.29876533e-01
-1.45655632e+00 -8.17823410e-01 9.34896171e-01 -3.61830652e-01
9.64750648e-01 -4.48553771e-01 -5.81650436e-01 3.79729599e-01
5.74948490e-01 1.30616337e-01 5.40667534e-01 5.46619110e-02
-5.83341777e-01 -4.53821540e-01 -8.83848608e-01 4.18770671e-01
7.30347872e-01 -6.71250105e-01 -3.59690875e-01 6.37903392e-01
5.64522922e-01 -5.55734038e-02 -8.57929766e-01 2.31650397e-01
6.87362313e-01 -1.08046138e+00 9.79786098e-01 -2.35197544e-01
4.84529793e-01 -4.34441537e-01 -5.77656090e-01 -9.13357139e-01
-1.92071572e-01 -7.96820939e-01 -5.93844093e-02 1.13796043e+00
5.16094804e-01 -4.27643716e-01 6.72067881e-01 -3.75635549e-02
-3.56762633e-02 -9.28002596e-01 -6.87182426e-01 -7.57976830e-01
-6.52288496e-02 -5.32790124e-01 2.78524101e-01 6.85758412e-01
-1.83971047e-01 1.98218063e-01 -1.21506758e-01 6.71309754e-02
7.90292382e-01 2.61563390e-01 6.48475289e-01 -1.11518693e+00
-4.81002092e-01 -5.64359367e-01 -5.48968971e-01 -8.81664217e-01
-1.51307300e-01 -1.27669311e+00 -1.40742198e-01 -1.65886724e+00
2.20834777e-01 -4.54683870e-01 -8.29634726e-01 4.86486852e-01
3.52471471e-01 5.40982604e-01 4.02758539e-01 3.97652715e-01
-9.53149796e-01 4.35490906e-01 6.80035591e-01 -3.91117215e-01
2.55035963e-02 -1.69369519e-01 -5.21550238e-01 2.52567351e-01
7.38140225e-01 -7.17849493e-01 -3.75845969e-01 -5.75238109e-01
3.09048593e-01 -2.52712816e-01 5.40874004e-01 -9.60888624e-01
4.15676773e-01 5.20579040e-01 2.63092309e-01 -6.48018718e-01
4.10398304e-01 -7.45378792e-01 -1.98008254e-01 3.42914075e-01
-5.29546916e-01 5.11588573e-01 -1.61585331e-01 5.05760849e-01
-3.88074607e-01 -4.37799692e-01 5.82852304e-01 -2.72326648e-01
-6.83365881e-01 4.05235112e-01 -3.58629739e-03 -3.08927327e-01
8.86342287e-01 -1.16875887e-01 -3.88291895e-01 -3.01767647e-01
-6.97121680e-01 4.19449419e-01 4.85991061e-01 6.75112724e-01
5.43121696e-01 -1.18517888e+00 -9.05094206e-01 -5.69250025e-02
3.41410846e-01 -4.49545264e-01 -3.12065315e-02 9.10775244e-01
-7.66598940e-01 5.29691577e-01 -1.56461880e-01 -6.75348163e-01
-1.43315887e+00 6.02234781e-01 1.88457742e-01 -5.65693617e-01
-5.45537114e-01 7.77033985e-01 -6.63954690e-02 -1.93964884e-01
3.70571196e-01 -1.35430008e-01 6.38687909e-02 5.05492926e-01
4.88494039e-01 3.78951252e-01 3.83492351e-01 -3.27526569e-01
-3.59740615e-01 3.75523150e-01 -4.51488107e-01 -2.41144925e-01
1.46553457e+00 -1.79672204e-02 -1.08520128e-01 4.26609069e-01
1.76658189e+00 -2.87442982e-01 -9.79377508e-01 -2.62825280e-01
2.79498339e-01 -5.24367034e-01 3.43903989e-01 -7.57004142e-01
-1.05392337e+00 1.02090275e+00 1.26392972e+00 5.75872660e-02
1.15798998e+00 1.37644991e-01 7.24502146e-01 4.73498255e-01
2.49135613e-01 -1.12398279e+00 2.10020587e-01 1.90583229e-01
1.03685546e+00 -1.41300881e+00 4.26387221e-01 1.64030910e-01
-3.70486319e-01 1.05450833e+00 1.27012640e-01 -3.56512159e-01
9.55801725e-01 1.41643226e-01 7.69006088e-02 -4.25011337e-01
-9.02172685e-01 -5.51396370e-01 2.71860331e-01 2.37465158e-01
5.38269281e-01 -1.01957723e-01 -4.75880176e-01 -1.80841107e-02
-5.48574701e-02 -2.58467300e-03 1.42715350e-01 1.03780544e+00
-1.73573568e-01 -1.43926334e+00 -9.72134396e-02 5.79453588e-01
-6.42938137e-01 -5.42270184e-01 -2.48500764e-01 6.67264104e-01
-4.82252538e-02 5.84329903e-01 -4.28404212e-02 -2.57786900e-01
1.76056698e-01 1.01418465e-01 6.01746500e-01 -6.09971941e-01
-6.03301108e-01 1.63718790e-01 -1.55850753e-01 -9.39354181e-01
-4.63111520e-01 -8.70559990e-01 -7.97953904e-01 -7.57536814e-02
-2.20723584e-01 1.05621750e-02 8.54038537e-01 3.65313441e-01
5.71290493e-01 4.97189723e-02 5.61048687e-01 -1.01716423e+00
-7.41903186e-01 -1.05763078e+00 -4.88505810e-01 3.00695658e-01
3.72543961e-01 -4.33444768e-01 -3.57733339e-01 -8.42158799e-04] | [10.694287300109863, 0.7674407958984375] |
55dcfa6f-252e-4420-b918-6da2fbae77c3 | faces-a-la-carte-text-to-face-generation-via | 2006.07606 | null | https://arxiv.org/abs/2006.07606v2 | https://arxiv.org/pdf/2006.07606v2.pdf | Faces à la Carte: Text-to-Face Generation via Attribute Disentanglement | Text-to-Face (TTF) synthesis is a challenging task with great potential for diverse computer vision applications. Compared to Text-to-Image (TTI) synthesis tasks, the textual description of faces can be much more complicated and detailed due to the variety of facial attributes and the parsing of high dimensional abstract natural language. In this paper, we propose a Text-to-Face model that not only produces images in high resolution (1024x1024) with text-to-image consistency, but also outputs multiple diverse faces to cover a wide range of unspecified facial features in a natural way. By fine-tuning the multi-label classifier and image encoder, our model obtains the vectors and image embeddings which are used to transform the input noise vector sampled from the normal distribution. Afterwards, the transformed noise vector is fed into a pre-trained high-resolution image generator to produce a set of faces with the desired facial attributes. We refer to our model as TTF-HD. Experimental results show that TTF-HD generates high-quality faces with state-of-the-art performance. | ['Teng Zhang', 'Tianren Wang', 'Brian Lovell'] | 2020-06-13 | null | null | null | null | ['text-to-face-generation'] | ['computer-vision'] | [ 5.18049717e-01 -1.03205973e-02 1.83384195e-01 -7.99871802e-01
-9.32957232e-01 -3.53893697e-01 6.80188656e-01 -8.12419593e-01
1.56043902e-01 6.72968745e-01 1.67644411e-01 2.24739477e-01
2.19055682e-01 -8.45196784e-01 -8.64029944e-01 -7.66165137e-01
4.91208553e-01 5.94344735e-01 -3.39907318e-01 1.10124446e-01
-2.10915402e-01 4.96787310e-01 -1.99294388e+00 5.26110291e-01
4.53214735e-01 1.21627986e+00 1.00271516e-01 5.22182703e-01
-1.54881701e-01 6.76506460e-01 -6.14282489e-01 -7.29641378e-01
2.69342959e-01 -6.27322614e-01 -3.74507189e-01 4.17425334e-01
9.03161049e-01 -3.69179159e-01 -9.81693491e-02 1.22468710e+00
4.43676770e-01 -1.31232217e-01 1.06279624e+00 -1.39227331e+00
-1.03686976e+00 1.07468843e-01 -6.81295395e-01 -5.01509428e-01
3.62948596e-01 -8.51550996e-02 6.32931650e-01 -1.30295217e+00
8.05288553e-01 1.91094911e+00 2.42733106e-01 1.03802645e+00
-1.31463003e+00 -1.16091323e+00 -1.28312230e-01 -1.30748615e-01
-1.61320162e+00 -8.27771783e-01 4.71327662e-01 -3.65510076e-01
4.32918757e-01 1.50911674e-01 3.11337084e-01 1.37905681e+00
-4.52129543e-02 2.60378450e-01 1.11188567e+00 -3.04714650e-01
-1.57606415e-02 2.31294125e-01 -7.79423773e-01 9.12874103e-01
8.91376063e-02 -4.30206805e-02 -4.70647842e-01 -7.41260797e-02
1.06186044e+00 -9.64289829e-02 -7.29045644e-02 3.72499004e-02
-9.52246487e-01 8.71580660e-01 3.20029482e-02 -7.93029496e-04
-3.07366669e-01 1.26902163e-01 2.42441282e-01 2.00398460e-01
4.17844862e-01 2.03703232e-02 -1.43774971e-01 2.17241347e-01
-7.50513673e-01 3.41853231e-01 4.84023720e-01 1.24902260e+00
8.63695621e-01 2.98219919e-01 -2.74132818e-01 1.21959269e+00
3.01701039e-01 9.00462985e-01 2.39923805e-01 -1.10378230e+00
2.71106422e-01 2.30213016e-01 4.26412001e-03 -8.76152277e-01
7.38188475e-02 4.86798435e-02 -1.30961084e+00 3.63261640e-01
1.36076987e-01 -3.88091430e-02 -1.03204453e+00 1.87620389e+00
4.33074862e-01 5.51128797e-02 1.48169354e-01 7.93993592e-01
9.11029458e-01 9.34734881e-01 3.26595604e-02 -2.26953328e-01
1.42262876e+00 -7.68208265e-01 -8.34873259e-01 -2.99473405e-01
4.01987843e-02 -8.86378706e-01 1.08562291e+00 1.62576392e-01
-9.18745816e-01 -7.36377060e-01 -6.60826087e-01 -6.28317147e-02
-2.67591160e-02 5.41537166e-01 9.62427557e-02 5.62422872e-01
-1.03671134e+00 -1.43906595e-02 -1.79744869e-01 -1.44732296e-01
8.60724807e-01 2.96233922e-01 -8.07980359e-01 -4.10381109e-01
-8.53180349e-01 6.07049584e-01 1.11314341e-01 -1.75238237e-01
-1.04912233e+00 -6.34578586e-01 -1.06436872e+00 -3.99528556e-02
3.43652248e-01 -6.97189391e-01 1.12906837e+00 -1.11518466e+00
-1.57514536e+00 9.65821445e-01 -4.91443545e-01 1.25166774e-01
4.61602539e-01 1.52404368e-01 -4.50206578e-01 2.24988237e-01
4.45705682e-01 1.06986439e+00 1.62707043e+00 -1.43935204e+00
-6.04709864e-01 -3.17772418e-01 -5.17701983e-01 4.79112789e-02
-3.16786170e-01 3.58530313e-01 -6.03304744e-01 -7.87496746e-01
-3.07637930e-01 -8.72970402e-01 1.00663252e-01 4.13283288e-01
-4.37407166e-01 -1.92460343e-01 1.05961180e+00 -2.04753861e-01
7.68365502e-01 -2.29882264e+00 1.68850303e-01 5.23085557e-02
1.49077669e-01 1.61688790e-01 -4.86256182e-01 2.32651271e-02
-1.06815793e-01 2.28880137e-01 -1.86728001e-01 -5.20625949e-01
-1.71745673e-01 1.82131335e-01 -2.05585524e-01 3.45328152e-01
6.82919204e-01 7.96460032e-01 -5.46021581e-01 -7.73829460e-01
2.36086220e-01 7.42804646e-01 -4.84370917e-01 5.10839462e-01
-4.08265829e-01 4.90354866e-01 -5.08151591e-01 7.64652491e-01
7.87984431e-01 -2.83872068e-01 -2.78325230e-02 -3.14337850e-01
2.46189862e-01 -5.83831787e-01 -9.98915792e-01 1.29183590e+00
-5.47076643e-01 4.00849760e-01 1.45793125e-01 -5.99563122e-01
1.25025094e+00 3.44419718e-01 2.10352391e-01 -5.97326100e-01
2.31434032e-01 1.99345857e-01 -3.04457784e-01 -5.63256443e-01
1.83709621e-01 -4.78208542e-01 -1.89295769e-01 2.91498899e-01
2.73492157e-01 -2.19051719e-01 1.42478243e-01 -8.64838660e-02
5.24270356e-01 1.12549216e-01 5.80670238e-02 -8.89526755e-02
7.08766818e-01 -5.42682171e-01 6.70399189e-01 2.67893285e-01
1.48729682e-01 9.83868480e-01 4.49965239e-01 -3.26193362e-01
-1.41962159e+00 -9.68108833e-01 -2.15611875e-01 8.74328613e-01
-2.65404046e-01 -2.69810796e-01 -9.82691109e-01 -5.14417231e-01
-7.04362988e-02 3.18412006e-01 -7.80991733e-01 -1.49463071e-02
-3.19385707e-01 -4.25496340e-01 5.58622301e-01 2.98298329e-01
5.90167344e-01 -1.20823348e+00 -9.90505219e-02 -3.25273275e-02
-2.05542639e-01 -1.41290855e+00 -7.36469567e-01 -3.94624442e-01
-2.13034570e-01 -7.86569118e-01 -8.12878132e-01 -1.13246894e+00
1.05737925e+00 7.71766007e-02 9.48437274e-01 -2.31388509e-02
-5.32635152e-01 4.58245203e-02 -2.40455389e-01 -3.28763813e-01
-7.31927633e-01 -3.81539106e-01 1.79930627e-01 7.37086296e-01
1.92515746e-01 -3.07648599e-01 -2.41449893e-01 2.35665098e-01
-1.10278249e+00 1.77787080e-01 7.02486694e-01 1.03325689e+00
8.50423992e-01 1.24125682e-01 7.72800565e-01 -1.01225293e+00
3.59821469e-01 -3.22588563e-01 -7.08881915e-01 3.16073179e-01
-2.86657661e-01 1.34603292e-01 9.95031595e-01 -6.96953952e-01
-1.16023815e+00 3.44173700e-01 -2.04496801e-01 -8.89696360e-01
-3.14119965e-01 -6.08938523e-02 -6.84754848e-01 -1.13374218e-02
4.34315592e-01 3.51972997e-01 2.67762333e-01 -2.01351628e-01
5.56350112e-01 8.44898343e-01 7.86151886e-01 -6.73389077e-01
1.03743291e+00 3.92804027e-01 1.54986233e-01 -9.10300970e-01
-8.23090971e-01 2.63102561e-01 -5.41290224e-01 -3.46179664e-01
1.21049082e+00 -1.11408329e+00 -5.87403834e-01 6.87362790e-01
-1.18610191e+00 -1.17463760e-01 -1.97973493e-02 6.28666133e-02
-6.79927409e-01 -1.37861982e-01 -4.38031346e-01 -8.38309050e-01
-2.98118979e-01 -1.34716403e+00 1.58282197e+00 3.38332981e-01
3.52562740e-02 -4.32975590e-01 -4.06994700e-01 3.10464323e-01
2.92216629e-01 4.02615130e-01 1.17595339e+00 -1.42563283e-01
-6.64856732e-01 3.48299444e-02 -5.50987244e-01 4.18353409e-01
2.44568616e-01 1.50760695e-01 -1.07337618e+00 -2.07528129e-01
-8.80127996e-02 -8.51730168e-01 4.72739697e-01 2.52194345e-01
1.57247245e+00 -5.88572145e-01 -1.42885298e-01 8.15268338e-01
1.52602875e+00 1.29909767e-02 7.50340044e-01 -1.12274930e-01
8.11050057e-01 8.02384257e-01 6.51285887e-01 5.44308484e-01
3.51119816e-01 6.90899193e-01 2.81135738e-01 -6.25610575e-02
-3.26166630e-01 -4.15913284e-01 2.68966109e-01 3.25446784e-01
2.89849639e-01 -1.36780158e-01 -4.18300360e-01 3.32574993e-01
-1.40794289e+00 -1.07584536e+00 2.06492528e-01 1.87256777e+00
1.09600914e+00 -3.14624488e-01 5.24391308e-02 -1.06501482e-01
1.00784743e+00 -4.22672704e-02 -6.07732534e-01 -3.45989585e-01
-1.69965044e-01 2.19927773e-01 6.08379245e-02 3.59904796e-01
-9.60436046e-01 1.07389987e+00 5.79034662e+00 1.07569134e+00
-1.12760699e+00 -5.64555787e-02 1.04268241e+00 -1.25643089e-01
-3.15173477e-01 -4.52261388e-01 -1.07832992e+00 4.57473457e-01
9.65890169e-01 -1.70591310e-01 5.73004365e-01 7.77246594e-01
-7.09514990e-02 3.15177739e-01 -1.07326376e+00 1.36103952e+00
4.48129922e-01 -1.28489459e+00 6.07810616e-01 1.28459051e-01
8.39444041e-01 -5.99982381e-01 3.33252281e-01 1.40788555e-01
1.09815203e-01 -1.54204559e+00 9.27018225e-01 5.06608665e-01
1.87342179e+00 -9.23454463e-01 2.55450487e-01 8.75050128e-02
-1.20588791e+00 -1.90047443e-01 -6.10325038e-01 5.24798036e-01
-5.54778762e-02 5.58780611e-01 -6.33666158e-01 2.76587486e-01
7.43570149e-01 6.71672523e-01 -4.07726765e-01 2.42465928e-01
-1.57962588e-03 1.52596921e-01 -1.02009431e-01 4.41004001e-02
-1.40837431e-01 -4.03252840e-01 7.21250400e-02 9.81181324e-01
7.33489692e-01 1.78374052e-01 1.09337956e-01 1.06908178e+00
-5.48630178e-01 1.41270071e-01 -9.15981472e-01 -1.16794415e-01
7.15843141e-01 1.50560439e+00 -2.67992586e-01 -4.04810518e-01
-6.34288728e-01 1.08255863e+00 3.63600522e-01 2.05101013e-01
-7.18649566e-01 -3.36893409e-01 7.60108888e-01 2.18085557e-01
5.17264068e-01 1.89296305e-01 4.70016152e-02 -9.72566068e-01
1.61701202e-01 -1.18772840e+00 -2.02358887e-03 -8.91330838e-01
-1.42304170e+00 1.10853159e+00 -1.71804503e-01 -1.19873321e+00
-5.18279374e-01 -6.06727481e-01 -2.94930160e-01 1.01308799e+00
-1.43628001e+00 -1.60184276e+00 -4.70234722e-01 7.44197369e-01
6.00963295e-01 -5.85864246e-01 1.12794030e+00 3.64967674e-01
-5.89296699e-01 9.51164961e-01 -3.46980020e-02 2.66834736e-01
8.97414148e-01 -7.63512194e-01 3.78732532e-01 5.72777867e-01
-1.33944884e-01 1.47011489e-01 3.18207026e-01 -5.66090107e-01
-1.38391173e+00 -1.85035443e+00 7.89991200e-01 -3.04056138e-01
1.50890023e-01 -6.76493168e-01 -7.57358074e-01 5.35339952e-01
-1.22311376e-02 4.34885591e-01 5.19465029e-01 -5.22062063e-01
-6.97318435e-01 -2.46444777e-01 -1.44736922e+00 6.20765388e-01
9.67567861e-01 -6.28844321e-01 4.53274511e-03 2.00612560e-01
5.05549371e-01 -2.97204733e-01 -9.15640950e-01 2.71195292e-01
6.00782514e-01 -8.47035587e-01 9.53756928e-01 -3.13536376e-01
8.60225260e-01 -2.32835442e-01 -3.42394471e-01 -1.14157403e+00
-3.72164041e-01 -4.63862062e-01 4.16610569e-01 1.58372939e+00
3.18634778e-01 -3.60388219e-01 5.39519608e-01 3.65189552e-01
1.69709817e-01 -8.16879332e-01 -9.05399859e-01 -5.76950967e-01
6.63754493e-02 4.25088853e-02 9.33092654e-01 6.59880042e-01
-6.60136104e-01 4.67479378e-01 -7.08645701e-01 7.92931095e-02
9.17393684e-01 1.49111301e-01 8.09581757e-01 -1.24258244e+00
9.28214267e-02 -2.26385623e-01 -3.27322423e-01 -7.49347985e-01
6.58791542e-01 -9.31937754e-01 2.80315667e-01 -1.09359837e+00
3.99282962e-01 -4.40157294e-01 4.59721148e-01 4.15661156e-01
-1.24338135e-01 6.95023179e-01 1.48151308e-01 -2.22121309e-02
-2.25913703e-01 7.45206177e-01 1.56652892e+00 -4.60195839e-02
3.32819462e-01 -3.28828365e-01 -9.21772301e-01 4.81479466e-01
5.66153347e-01 -4.83985394e-01 -3.70500475e-01 -4.69108492e-01
-2.19134048e-01 3.05001438e-01 3.12171131e-01 -8.18755567e-01
-2.11832479e-01 -3.82257313e-01 8.35070670e-01 -3.20318460e-01
7.12861240e-01 -7.43011057e-01 2.74869442e-01 -9.19426307e-02
-3.75263125e-01 -3.92417498e-02 -2.28660963e-02 3.51966888e-01
-2.61984527e-01 -1.05636299e-01 1.29507875e+00 -1.22536793e-01
-2.44097412e-01 7.66156256e-01 -2.11683363e-01 7.68939406e-02
1.13311052e+00 -2.69858167e-02 -3.34789693e-01 -6.04716003e-01
-3.62594187e-01 -6.08271658e-02 6.03594899e-01 7.20598817e-01
9.71809804e-01 -1.83131278e+00 -1.18675983e+00 8.79088163e-01
2.79974103e-01 1.97697148e-01 1.53738335e-01 -3.86830568e-02
-1.58988416e-01 1.72867194e-01 -4.35174435e-01 -6.49988115e-01
-1.40847063e+00 3.87979746e-01 2.28973776e-01 3.18259269e-01
-5.96116006e-01 8.49870324e-01 5.33215642e-01 -4.02329773e-01
1.30307153e-01 1.72573656e-01 9.83854942e-03 -1.84127036e-02
9.04367387e-01 -1.75410360e-01 -2.65311182e-01 -1.18084931e+00
-2.27959931e-01 9.44375157e-01 6.59819134e-03 -2.11897239e-01
1.28610599e+00 3.86151448e-02 -2.78160900e-01 -1.77843366e-02
1.58075929e+00 -1.95929393e-01 -1.36719930e+00 -2.49809504e-01
-4.50779289e-01 -7.98178554e-01 -2.71640360e-01 -5.31166911e-01
-1.33431983e+00 8.85021746e-01 5.05839288e-01 -2.55471021e-01
1.27284098e+00 1.42074749e-01 7.65964150e-01 1.05312765e-01
2.92589664e-01 -8.01847756e-01 3.27598542e-01 2.64749110e-01
1.22362983e+00 -1.21689320e+00 -3.28436673e-01 -5.50584197e-01
-8.85541320e-01 1.09647763e+00 9.24960017e-01 4.37483117e-02
4.84080881e-01 5.61226308e-01 2.85662323e-01 3.06495111e-02
-9.06696141e-01 -1.25958081e-02 1.32599026e-01 9.80299234e-01
1.84298813e-01 -3.58970650e-02 3.80519241e-01 5.11925220e-01
-4.14214492e-01 5.66747189e-02 4.95301515e-01 3.45066875e-01
-3.02593857e-01 -1.26613688e+00 -5.72547615e-01 7.58720219e-01
-4.42937106e-01 1.11041263e-01 -1.42077148e-01 3.32373321e-01
1.65739447e-01 9.30052638e-01 1.35352626e-01 -4.76688236e-01
2.41440356e-01 2.36604989e-01 6.51980937e-01 -7.44832397e-01
-1.48467764e-01 2.04731718e-01 3.21669690e-03 -3.27014774e-01
-1.71818867e-01 -7.25935042e-01 -1.22981870e+00 -1.38935104e-01
-1.03769256e-02 -1.65100187e-01 6.77407622e-01 7.49642491e-01
4.13761348e-01 3.66071433e-01 9.84311104e-01 -9.57494378e-01
-5.04841387e-01 -9.48830783e-01 -5.15372276e-01 7.18905687e-01
3.97005081e-01 -6.82272255e-01 -1.64086968e-01 4.24320161e-01] | [12.53962516784668, -0.15461592376232147] |
70432a80-3114-4bcf-ab46-a81a24f527fd | semeval-2016-task-9-chinese-semantic | null | null | https://aclanthology.org/S16-1167 | https://aclanthology.org/S16-1167.pdf | SemEval-2016 Task 9: Chinese Semantic Dependency Parsing | null | ['Yanqiu Shao', 'Yu Ding', 'Ting Liu', 'Wanxiang Che'] | 2016-06-01 | null | null | null | semeval-2016-6 | ['semantic-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] | [-7.415567874908447, 3.685319662094116] |
428b2960-f95d-42a1-9e3f-d58aa88bb4f5 | identification-of-significant-permissions-for | 2103.00643 | null | https://arxiv.org/abs/2103.00643v1 | https://arxiv.org/pdf/2103.00643v1.pdf | Identification of Significant Permissions for Efficient Android Malware Detection | Since Google unveiled Android OS for smartphones, malware are thriving with 3Vs, i.e. volume, velocity, and variety. A recent report indicates that one out of every five business/industry mobile application leaks sensitive personal data. Traditional signature/heuristic-based malware detection systems are unable to cope up with current malware challenges and thus threaten the Android ecosystem. Therefore recently researchers have started exploring machine learning and deep learning based malware detection systems. In this paper, we performed a comprehensive feature analysis to identify the significant Android permissions and propose an efficient Android malware detection system using machine learning and deep neural network. We constructed a set of $16$ permissions ($8\%$ of the total set) derived from variance threshold, auto-encoders, and principal component analysis to build a malware detection engine that consumes less train and test time without significant compromise on the model accuracy. Our experimental results show that the Android malware detection model based on the random forest classifier is most balanced and achieves the highest area under curve score of $97.7\%$, which is better than the current state-of-art systems. We also observed that deep neural networks attain comparable accuracy to the baseline results but with a massive computational penalty. | ['Mohit Sewak', 'Ritvik Rajvanshi', 'Sanjay K. Sahay', 'Hemant Rathore'] | 2021-02-28 | null | null | null | null | ['android-malware-detection'] | ['miscellaneous'] | [ 1.80707708e-01 -3.36657941e-01 -4.83568996e-01 -6.94817454e-02
-4.00476813e-01 -2.45551765e-01 4.57353801e-01 -3.03445667e-01
-2.18962505e-01 3.97740901e-01 -5.47218204e-01 -9.70950603e-01
1.08651489e-01 -6.23265982e-01 -5.81015944e-01 -4.75892395e-01
-3.00670773e-01 1.11739179e-02 2.29903057e-01 -1.14329115e-01
3.62834364e-01 1.94916934e-01 -1.51561606e+00 3.78152460e-01
8.77840161e-01 1.52461755e+00 -7.35031441e-02 7.35450745e-01
8.57497454e-02 7.16166556e-01 -9.21896100e-01 -6.96618617e-01
3.41840118e-01 3.34828421e-02 -2.93013126e-01 -5.98488688e-01
1.85563326e-01 -7.88123012e-01 -2.87390620e-01 1.17382336e+00
2.99022138e-01 -4.34480697e-01 6.66036963e-01 -1.47722900e+00
-5.72592795e-01 3.95491242e-01 -7.28625238e-01 3.73651773e-01
1.14574008e-01 1.21967539e-01 5.60896993e-01 -5.11239469e-01
1.44692495e-01 8.89747679e-01 1.08294547e+00 5.93662322e-01
-7.37068176e-01 -1.06323028e+00 -1.26626208e-01 3.61995101e-01
-1.22679162e+00 -3.40751350e-01 8.47056150e-01 -5.96308947e-01
1.32193744e+00 3.56736571e-01 7.02102780e-01 1.61820281e+00
8.23984802e-01 5.22522628e-01 1.05791640e+00 6.87970221e-02
1.50950059e-01 4.45468307e-01 4.12796557e-01 9.61843312e-01
7.35862195e-01 1.44601524e-01 -7.16686025e-02 -7.09457994e-01
-5.01825809e-02 6.02418542e-01 4.96836752e-02 -5.33264771e-04
-4.50374067e-01 1.05757189e+00 1.17910989e-01 2.57392135e-02
-2.33955115e-01 1.24505520e-01 7.87407517e-01 -5.18727079e-02
3.05793434e-01 4.26628113e-01 -8.35289001e-01 -5.32238483e-01
-9.00660694e-01 -7.75702298e-02 8.98588896e-01 5.14701366e-01
5.97654641e-01 3.70000273e-01 2.63811469e-01 6.23391926e-01
4.64422852e-01 6.26522958e-01 8.60595882e-01 -8.38117003e-01
3.20344791e-02 9.54648554e-01 -4.49111760e-01 -1.38321292e+00
-7.37887528e-03 -4.29276168e-01 -8.08164716e-01 3.77687439e-02
-4.42076139e-02 -2.68790931e-01 -7.49755740e-01 1.24549711e+00
-3.88856903e-02 3.06692451e-01 -1.75759718e-01 -2.02954501e-01
2.86121905e-01 6.01172328e-01 -9.59230438e-02 -3.32885683e-01
1.36432636e+00 -7.26945341e-01 -6.58769548e-01 -3.60193640e-01
5.00286758e-01 -3.42994034e-01 1.30837917e+00 6.47155404e-01
-3.38447034e-01 -3.42501581e-01 -1.37035084e+00 2.70414889e-01
-7.07674801e-01 1.55973881e-01 8.54977727e-01 1.44074166e+00
-8.10691178e-01 3.62590194e-01 -1.05454826e+00 -1.66008677e-02
9.26160932e-01 6.99122906e-01 1.24881431e-01 4.46666121e-01
-8.56106222e-01 4.39122885e-01 1.33954793e-01 -2.02254161e-01
-1.31561923e+00 -4.37971652e-01 -5.61113894e-01 3.82377952e-02
5.73794305e-01 -5.42950481e-02 1.18629110e+00 -6.70943320e-01
-1.44094658e+00 4.35523868e-01 -6.05064966e-02 -7.84109056e-01
7.77528733e-02 -4.01845366e-01 -6.81785345e-01 -1.76052034e-01
-3.10142152e-02 2.55255193e-01 1.31568277e+00 -1.12510073e+00
-6.74539983e-01 -5.77680528e-01 -1.97813854e-01 -5.41734993e-01
-1.08242226e+00 6.49435595e-02 -2.86388218e-01 -2.14390770e-01
-5.82693458e-01 -1.21058774e+00 1.05983987e-01 -7.98575342e-01
-4.96473849e-01 6.62254393e-02 1.70985591e+00 -1.08636725e+00
1.63541651e+00 -2.12007952e+00 -5.76191127e-01 1.47353634e-01
4.97211158e-01 8.05820823e-01 3.40001225e-01 -5.82028963e-02
2.20446661e-01 4.27787811e-01 -3.37902784e-01 -1.71112999e-01
-1.23993650e-01 8.29799194e-03 -5.29201567e-01 3.41110826e-01
-5.54235913e-02 8.59302938e-01 -4.51324642e-01 -2.36782536e-01
-5.51899255e-04 6.88274920e-01 -6.76584423e-01 5.65431453e-02
-2.02976435e-01 -1.82371393e-01 -2.84099519e-01 1.18184662e+00
5.97088635e-01 -1.85761735e-01 4.31375355e-01 1.29969284e-01
3.62473220e-01 3.23167711e-01 -3.39595407e-01 9.14507270e-01
-4.91571277e-01 8.54384780e-01 6.47724122e-02 -7.87696958e-01
8.71521652e-01 -2.60130495e-01 5.35074532e-01 -5.56891382e-01
4.75416809e-01 3.30736190e-01 3.84466201e-01 -4.62689072e-01
4.49464232e-01 6.04578316e-01 -1.60421669e-01 3.37102771e-01
-3.67127180e-01 2.54191637e-01 -4.35482204e-01 4.03985046e-02
1.50263536e+00 -2.00214833e-01 4.13399339e-01 -1.03126377e-01
7.23731279e-01 -2.88133323e-01 4.81721669e-01 5.35369933e-01
-6.74148560e-01 -4.52360243e-01 8.48133624e-01 -4.20008987e-01
-5.69342315e-01 -7.80730844e-01 9.72124860e-02 1.26251495e+00
-9.30193812e-02 -6.93750560e-01 -1.30964553e+00 -1.21423507e+00
-3.79359443e-03 5.19262135e-01 -5.57748079e-01 -5.19711137e-01
-6.95027173e-01 -9.16033685e-01 1.01517832e+00 3.28189582e-01
7.65078485e-01 -7.80327797e-01 -9.05385435e-01 -3.16225499e-01
2.39715844e-01 -1.00473714e+00 -2.94380844e-01 2.94384241e-01
-7.84609556e-01 -1.33497036e+00 8.35808273e-03 -5.75641096e-01
3.27174425e-01 2.77835190e-01 5.91577053e-01 1.43580781e-02
-3.03785264e-01 6.17944114e-02 -8.89121890e-02 -6.78849280e-01
-5.43946981e-01 2.97970414e-01 5.18360317e-01 -2.10052192e-01
8.64589751e-01 -4.71043378e-01 -5.26597500e-01 2.18402341e-01
-4.94066030e-01 -8.63815069e-01 7.86448121e-01 7.63525486e-01
3.92873675e-01 6.72544837e-01 4.70585734e-01 -8.08725119e-01
1.06826615e+00 -6.87659919e-01 -5.63162923e-01 -2.82554775e-01
-1.25757468e+00 -3.15444529e-01 7.64208853e-01 -7.28972375e-01
-7.27205515e-01 1.16932705e-01 -3.69604796e-01 -5.12880743e-01
1.53247595e-01 -1.75061524e-02 -4.87569004e-01 -2.06469491e-01
7.47692168e-01 4.06257808e-01 3.18403393e-01 -2.16023624e-01
1.27408495e-02 1.27666795e+00 2.06723556e-01 7.04017514e-03
4.73119050e-01 3.66201133e-01 -2.23923430e-01 -1.04049623e+00
-2.69200001e-02 -1.59445494e-01 -7.15966672e-02 7.35175982e-02
6.73744440e-01 -6.90469563e-01 -1.22969270e+00 7.58385062e-01
-9.29229677e-01 -2.22243264e-01 1.93819329e-01 -3.85168120e-02
-1.61846578e-01 6.35714591e-01 -6.12397790e-01 -1.02183867e+00
-7.32329249e-01 -1.71493411e+00 9.75004554e-01 -1.58377022e-01
-2.96885282e-01 -6.15534961e-01 -8.15839767e-02 4.68658775e-01
5.79359531e-01 1.47151023e-01 9.16592240e-01 -1.12372935e+00
-2.42635965e-01 -5.30416191e-01 -1.83150068e-01 8.44160199e-01
2.76449054e-01 2.32777134e-01 -1.17208838e+00 -3.07281733e-01
7.04858482e-01 2.06400141e-01 9.04651523e-01 4.45942491e-01
1.79592967e+00 -8.95554006e-01 -8.36350083e-01 7.27979958e-01
1.05567479e+00 9.70906556e-01 5.86478829e-01 3.09375495e-01
9.69660640e-01 3.95823151e-01 2.22959757e-01 2.14804560e-01
2.47551546e-01 5.32122612e-01 8.16262245e-01 5.55205345e-01
3.28883678e-01 -1.83949083e-01 8.95142972e-01 5.84092140e-01
2.01633766e-01 1.85601022e-02 -1.10282111e+00 1.29329711e-01
-1.28362048e+00 -8.05678368e-01 7.52151608e-02 2.09815764e+00
6.12861037e-01 6.59137845e-01 4.28934604e-01 3.35392237e-01
4.81510311e-01 1.88667595e-01 -7.18142509e-01 -8.10198963e-01
4.71282095e-01 1.72026262e-01 7.53806949e-01 3.44128191e-01
-1.28754520e+00 8.60507429e-01 6.72963381e+00 1.05978906e+00
-1.49482071e+00 4.89290535e-01 8.15606236e-01 -1.44510642e-01
1.19031452e-01 -4.79472280e-01 -9.73024189e-01 1.03104365e+00
1.69766963e+00 2.79279441e-01 5.37642896e-01 1.67439914e+00
-1.20618030e-01 1.72612414e-01 -6.65217578e-01 1.20491147e+00
3.76659125e-01 -1.06112099e+00 -1.91451341e-01 9.62064087e-01
4.39188331e-01 1.21269412e-01 5.92368901e-01 6.70547843e-01
8.91574398e-02 -1.14619279e+00 1.92647815e-01 -6.96455594e-03
9.30113733e-01 -8.65156531e-01 8.13804388e-01 3.34864467e-01
-7.90705502e-01 -9.48759198e-01 7.92530775e-02 -1.71112999e-01
-4.18960452e-01 4.82889265e-01 -1.30120993e+00 -1.96725652e-01
9.18022633e-01 5.92847586e-01 -8.74260783e-01 2.50631601e-01
8.03469121e-02 9.07710612e-01 -5.53571023e-02 -5.25642157e-01
-1.64650053e-01 2.02019125e-01 3.75772744e-01 1.17248130e+00
4.37387139e-01 -5.72514355e-01 -3.78157236e-02 5.05792558e-01
-2.02795878e-01 -1.70322135e-01 -9.71808076e-01 -4.89479959e-01
4.89096195e-01 1.19682837e+00 -5.75803995e-01 -3.94606203e-01
-1.64128631e-01 8.70226562e-01 -1.01004422e-01 -1.62541047e-01
-1.20170486e+00 -4.44522917e-01 1.14283466e+00 3.58277202e-01
2.99658626e-01 -2.91241378e-01 -5.80966353e-01 -8.92473698e-01
-8.76838192e-02 -1.32056165e+00 -7.57038742e-02 1.14364615e-02
-8.86364818e-01 7.22057283e-01 -2.12556928e-01 -9.05295551e-01
-5.33186495e-01 -1.17649019e+00 -4.09938037e-01 2.30388254e-01
-9.13206041e-01 -8.69994164e-01 -2.06052572e-01 3.57428849e-01
4.89831746e-01 -7.02810168e-01 6.97354138e-01 3.67481053e-01
-9.64501381e-01 8.85118783e-01 1.46425560e-01 -6.12315023e-03
1.88872322e-01 -9.96317208e-01 5.36689878e-01 8.64836872e-01
-1.63423926e-01 1.04006076e+00 3.48114222e-01 -1.09602582e+00
-1.82408071e+00 -1.16578579e+00 2.60338038e-01 -8.42365921e-01
6.54248238e-01 -6.86747551e-01 -7.13682950e-01 6.74349904e-01
1.25909194e-01 -3.84940624e-01 9.86972630e-01 -1.23474672e-01
-6.37785912e-01 -4.53976661e-01 -1.26620448e+00 4.65833247e-01
7.80689538e-01 -7.58685052e-01 2.50586700e-02 3.85135449e-02
8.72202098e-01 -8.80541354e-02 -5.02622426e-01 6.38262570e-01
9.58005369e-01 -1.21067202e+00 9.73233402e-01 -2.50081569e-01
2.18563437e-01 1.15536563e-01 -3.90831709e-01 -6.81121111e-01
-1.51324933e-02 -7.21237063e-01 -1.13302541e+00 7.03089714e-01
3.94906372e-01 -8.21485996e-01 1.05377293e+00 6.17827140e-02
-8.21913481e-02 -1.34005439e+00 -7.94847131e-01 -8.58751774e-01
-1.57482535e-01 -8.42988670e-01 7.79592097e-01 7.82974243e-01
-2.48983219e-01 2.24345163e-01 -4.40508693e-01 -9.02751610e-02
3.94697905e-01 -6.07507586e-01 8.45994055e-01 -1.19382977e+00
-4.10189688e-01 -7.56385565e-01 -4.20725375e-01 -5.77595174e-01
2.74183244e-01 -4.86288995e-01 -2.29578093e-01 -7.84281433e-01
3.48344415e-01 -5.46456352e-02 -2.86564708e-01 6.09287798e-01
-2.73634703e-03 2.70582408e-01 -2.79123515e-01 2.83455253e-01
-2.46846348e-01 1.18738115e-01 4.19411600e-01 -4.15405244e-01
-4.21717018e-01 1.64429709e-01 -1.13760328e+00 1.04229593e+00
1.05895865e+00 -2.70334780e-01 -5.90791106e-01 2.37618387e-02
3.37337404e-02 -6.58378482e-01 2.52872705e-01 -1.10109985e+00
-2.82756388e-01 9.66929793e-02 4.03519988e-01 -3.82224858e-01
5.22593200e-01 -8.48571658e-01 -9.51964185e-02 1.09451807e+00
2.02566564e-01 1.59493953e-01 3.57944816e-01 6.83275759e-01
3.42224836e-01 1.19883507e-01 6.35209918e-01 1.58414394e-01
-6.57464147e-01 1.74701393e-01 -5.18220067e-01 -4.13881868e-01
1.34338236e+00 -4.53172773e-01 -6.64093196e-01 -1.07417703e-01
-2.14032903e-02 -5.30556858e-01 5.15209854e-01 5.08241057e-01
7.70090640e-01 -8.65749836e-01 1.27785336e-02 5.54943383e-01
-1.93097219e-01 -6.29275501e-01 3.44930589e-02 6.47707880e-01
-5.77517509e-01 6.81029975e-01 -2.49927446e-01 -5.47539532e-01
-1.49898565e+00 8.36202860e-01 2.38275707e-01 -4.08471137e-01
-8.43473151e-02 6.84487462e-01 -3.15280288e-01 -2.85253674e-01
3.23341906e-01 -2.10151732e-01 -2.59777367e-01 -6.36828169e-02
7.10104644e-01 8.81885231e-01 1.34253368e-01 -6.82110846e-01
-8.12776864e-01 1.96074620e-01 -3.23540270e-01 3.42018247e-01
8.55792820e-01 4.08223242e-01 -2.61813343e-01 3.45558189e-02
1.54278982e+00 4.72043961e-01 -8.86294961e-01 5.61351955e-01
-4.41945391e-03 -4.74127740e-01 -2.26970911e-02 -8.27578187e-01
-1.04118276e+00 9.42749500e-01 1.11197591e+00 7.97279835e-01
1.25701404e+00 -2.92842448e-01 1.29663277e+00 4.34760958e-01
2.62942672e-01 -7.99948931e-01 2.63535827e-01 4.57175851e-01
2.96027660e-01 -1.30956519e+00 1.86008029e-02 -3.01937222e-01
-4.72824514e-01 5.99918842e-01 8.65583658e-01 -4.54988554e-02
9.50771928e-01 5.62518656e-01 -3.17956060e-01 -2.46326134e-01
-5.45402646e-01 4.48702723e-01 1.29815534e-01 9.66227710e-01
6.08447604e-02 4.50502157e-01 -1.99556991e-01 9.53377724e-01
-3.77179503e-01 -2.63336003e-01 2.36383051e-01 8.08863103e-01
-5.67596912e-01 -1.05807674e+00 -1.88635960e-01 1.11064303e+00
-9.86184061e-01 -1.79797217e-01 -7.95641422e-01 4.91620570e-01
5.19280076e-01 9.74239171e-01 -1.08734876e-01 -1.48052061e+00
-9.65899974e-02 5.14661111e-02 -8.91890451e-02 -3.76270294e-01
-4.64335978e-01 -2.07489297e-01 -4.59468774e-02 -7.45908976e-01
3.90892535e-01 -4.52505499e-01 -9.79308128e-01 -4.28679824e-01
-2.99329847e-01 -1.56426683e-01 1.15386224e+00 6.27820075e-01
9.41938758e-01 5.44929147e-01 6.60736084e-01 -7.34166145e-01
-6.67734325e-01 -9.92533028e-01 -1.19840726e-01 -2.35091433e-01
3.63451749e-01 -7.24079490e-01 -4.73559648e-01 -2.27289021e-01] | [14.423982620239258, 9.68143081665039] |
28fbe664-de51-4aa8-96d6-d9883eff89d6 | reducing-computational-costs-in-sentiment | 2306.09705 | null | https://arxiv.org/abs/2306.09705v1 | https://arxiv.org/pdf/2306.09705v1.pdf | Reducing Computational Costs in Sentiment Analysis: Tensorized Recurrent Networks vs. Recurrent Networks | Anticipating audience reaction towards a certain text is integral to several facets of society ranging from politics, research, and commercial industries. Sentiment analysis (SA) is a useful natural language processing (NLP) technique that utilizes lexical/statistical and deep learning methods to determine whether different-sized texts exhibit positive, negative, or neutral emotions. Recurrent networks are widely used in machine-learning communities for problems with sequential data. However, a drawback of models based on Long-Short Term Memory networks and Gated Recurrent Units is the significantly high number of parameters, and thus, such models are computationally expensive. This drawback is even more significant when the available data are limited. Also, such models require significant over-parameterization and regularization to achieve optimal performance. Tensorized models represent a potential solution. In this paper, we classify the sentiment of some social media posts. We compare traditional recurrent models with their tensorized version, and we show that with the tensorized models, we reach comparable performances with respect to the traditional models while using fewer resources for the training. | ['Joe Kaul', 'Anna Nguyen', 'Gabriel Lopez'] | 2023-06-16 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-1.96147442e-01 -2.35591590e-01 -3.65501851e-01 -4.42043096e-01
-4.25828904e-01 -4.96004254e-01 3.84192735e-01 4.74746764e-01
-5.48667967e-01 5.10428667e-01 3.80913377e-01 -4.68527138e-01
3.02735567e-01 -7.67397761e-01 -2.39471883e-01 -5.95888853e-01
1.03852265e-01 7.62401670e-02 -9.58093181e-02 -6.08443558e-01
3.09726179e-01 2.95560628e-01 -1.24329531e+00 6.39910460e-01
5.65806270e-01 1.36969388e+00 -2.31070504e-01 5.50133467e-01
-6.68879986e-01 1.20984495e+00 -3.35780233e-01 -7.10948706e-01
-1.10423878e-01 -9.22458842e-02 -8.33113730e-01 -1.56163588e-01
-7.89633617e-02 1.27843216e-01 5.95536828e-02 1.13968253e+00
3.14763218e-01 2.58172214e-01 3.22245359e-01 -8.76080632e-01
-5.71664214e-01 8.17949355e-01 -4.97770280e-01 2.76358902e-01
8.16983506e-02 -4.18136954e-01 1.43240070e+00 -1.20339549e+00
2.39038795e-01 1.14116645e+00 6.73847973e-01 3.55161816e-01
-8.95756721e-01 -5.02068818e-01 4.95092601e-01 1.25553817e-01
-9.72380638e-01 -3.02317619e-01 9.43466306e-01 -4.25463527e-01
1.12285984e+00 4.04503137e-01 6.73425555e-01 1.12868202e+00
3.86273324e-01 9.74196076e-01 8.96421432e-01 -3.25998664e-01
1.34615034e-01 2.80427873e-01 5.67941666e-01 4.91181731e-01
-1.82341531e-01 -5.36253452e-01 -4.98925269e-01 -2.34848499e-01
1.24570385e-01 3.98709476e-01 9.28544346e-03 2.48873159e-02
-1.09610868e+00 1.14312851e+00 4.32154745e-01 7.48717964e-01
-5.66052914e-01 3.05298939e-02 8.45777988e-01 5.37553191e-01
1.01520395e+00 5.44437826e-01 -6.32621109e-01 -2.16934562e-01
-7.50252068e-01 1.75798554e-02 7.89218366e-01 4.00503844e-01
7.28368700e-01 1.55230299e-01 1.53617099e-01 9.53644097e-01
2.64354169e-01 3.17823172e-01 1.01776481e+00 -3.93612593e-01
5.37362695e-01 1.01084113e+00 -7.00113401e-02 -1.55189204e+00
-5.90560973e-01 -5.78769267e-01 -1.04562223e+00 -4.06499922e-01
1.31397545e-01 -4.32244658e-01 -6.16256773e-01 1.33569098e+00
9.15655568e-02 -1.73120916e-01 1.55539662e-01 8.84811044e-01
7.76033401e-01 8.98783863e-01 4.77285571e-02 -2.51535833e-01
1.44462645e+00 -9.22341108e-01 -9.84041393e-01 -4.61019576e-01
1.10463941e+00 -1.07089102e+00 1.15169430e+00 4.28787589e-01
-8.35041642e-01 -1.44827083e-01 -8.40822756e-01 -2.50907809e-01
-7.28932858e-01 2.61519700e-01 9.52967227e-01 7.32554138e-01
-1.00302994e+00 5.86079180e-01 -6.48299277e-01 -3.50861609e-01
1.69573538e-02 4.36022520e-01 -1.08370334e-01 4.63214040e-01
-1.44216931e+00 7.26798475e-01 -1.70073435e-01 6.54429257e-01
-8.53766426e-02 -2.13071093e-01 -7.15264797e-01 1.19455755e-01
1.41011611e-01 -1.67315304e-01 1.38172793e+00 -1.25131202e+00
-1.82164717e+00 7.40579367e-01 -3.09252411e-01 -5.34220815e-01
1.50752634e-01 -3.50668758e-01 -5.48905432e-01 -2.77226627e-01
-1.25594810e-01 8.20924193e-02 8.91392171e-01 -4.53110188e-01
-2.92308033e-01 -3.78628165e-01 9.64856818e-02 2.10571215e-01
-9.80607808e-01 4.98001605e-01 -1.29617825e-01 -6.22531116e-01
2.07257196e-01 -1.07006216e+00 -6.87848985e-01 -5.65238416e-01
-3.93015146e-01 -5.01199126e-01 6.82099938e-01 -4.70802158e-01
1.36743748e+00 -2.22483730e+00 9.24039781e-02 3.14711720e-01
3.32758158e-01 3.02650154e-01 -6.06923699e-02 5.29819071e-01
-2.39954013e-02 4.30896759e-01 3.45497191e-01 -3.30918461e-01
-1.14759646e-01 -7.82476142e-02 -6.27690792e-01 2.46052697e-01
1.33655161e-01 7.35294044e-01 -6.15451336e-01 -3.73369336e-01
-9.01046917e-02 5.09658158e-01 -4.38543797e-01 -4.23207209e-02
-1.84306622e-01 1.48786873e-01 -8.40561807e-01 3.85838985e-01
1.27202511e-01 -5.87653518e-01 4.48732227e-01 -1.75012246e-01
-3.41692388e-01 6.50165141e-01 -8.11122954e-01 1.12858188e+00
-6.09264076e-01 6.98717535e-01 -9.38792899e-02 -1.22560585e+00
1.11723948e+00 4.75463688e-01 4.20446008e-01 -6.96405232e-01
6.38298571e-01 3.29158962e-01 -1.51143596e-01 -5.86517274e-01
1.00497317e+00 -3.31780702e-01 -2.79557228e-01 6.90087259e-01
-3.60724419e-01 1.45198345e-01 2.76351243e-01 2.16066062e-01
6.61534786e-01 -4.73623067e-01 1.20178513e-01 -1.81066826e-01
6.17674470e-01 -2.37726733e-01 6.76083803e-01 3.48952442e-01
-1.03218541e-01 3.32182407e-01 8.19231808e-01 -8.22619200e-01
-8.10504973e-01 -8.34048316e-02 2.11613938e-01 1.50407720e+00
-3.20734739e-01 -5.94245315e-01 -4.01353419e-01 -3.21905524e-01
-3.50733817e-01 2.87978023e-01 -7.07342386e-01 6.66113645e-02
-5.76049030e-01 -1.07652986e+00 2.68828660e-01 5.02772391e-01
2.14699805e-01 -1.03706515e+00 -2.77302921e-01 2.58076280e-01
-4.79981989e-01 -1.19613755e+00 -3.65235180e-01 1.96595475e-01
-9.78951037e-01 -6.42795682e-01 -6.40143991e-01 -7.62326062e-01
4.10716385e-01 4.52462465e-01 9.86724854e-01 2.38769371e-02
3.20533484e-01 -6.21688403e-02 -6.75577343e-01 -4.74108011e-01
-1.79907933e-01 4.27576035e-01 2.69211084e-01 5.87502420e-01
6.18779361e-01 -2.80226618e-01 -5.27201355e-01 2.36773506e-01
-8.43282700e-01 -2.23093876e-03 4.85476464e-01 6.66346967e-01
1.58523262e-01 -1.08148828e-01 6.20975673e-01 -9.70389247e-01
1.03775227e+00 -4.82660145e-01 -2.84612656e-01 2.37755403e-01
-4.61823642e-01 -3.73340957e-02 7.69589245e-01 -5.62546730e-01
-8.58362257e-01 -2.76212722e-01 -3.52086425e-02 -1.92555025e-01
4.73252058e-01 1.21570528e+00 4.14985627e-01 7.60974586e-02
6.11113787e-01 -7.91627988e-02 1.93331726e-02 -4.46621329e-01
1.82040021e-01 6.86746061e-01 -3.15903753e-01 -2.73066431e-01
2.49618292e-01 4.33716506e-01 -4.06688511e-01 -8.98253977e-01
-1.21885276e+00 -5.38793266e-01 -3.12344551e-01 -2.92346835e-01
6.19666815e-01 -8.90826285e-01 -9.69197869e-01 3.66272241e-01
-1.19765627e+00 -2.40348540e-02 4.58354764e-02 7.87056684e-01
2.92685311e-02 3.26791674e-01 -9.67556477e-01 -8.56956542e-01
-6.65182948e-01 -1.00484276e+00 6.00501776e-01 1.03041530e-01
-4.07632440e-01 -1.12122285e+00 8.30409303e-02 4.48663175e-01
6.68134272e-01 -9.96017084e-02 8.76872957e-01 -6.82101011e-01
8.17546770e-02 -5.98360896e-01 -2.11724833e-01 5.74021935e-01
-7.00636581e-03 9.03065428e-02 -8.35444450e-01 -2.31123850e-01
3.10755134e-01 -5.48094153e-01 8.01110923e-01 2.93303281e-01
1.00884211e+00 -4.39920962e-01 1.42290697e-01 6.31242767e-02
9.61013854e-01 -2.06446480e-02 3.17652673e-01 3.30464602e-01
7.98120499e-01 9.00883555e-01 5.58552861e-01 4.85213578e-01
4.44684058e-01 2.46623993e-01 9.01069120e-02 -2.32175261e-01
6.08992517e-01 1.60112247e-01 7.72794604e-01 1.52812469e+00
-1.11723013e-01 -1.83006525e-01 -1.00747633e+00 5.62947273e-01
-2.05799556e+00 -8.29372168e-01 -3.08971256e-01 1.87085974e+00
6.96823061e-01 1.97927862e-01 2.61496808e-02 3.75527322e-01
7.06564486e-01 4.57331926e-01 -4.48973387e-01 -1.02054918e+00
-2.93704331e-01 5.69590144e-02 3.60823989e-01 3.01849246e-01
-1.05605006e+00 1.09558320e+00 6.13260889e+00 5.59945285e-01
-1.76716256e+00 8.56573358e-02 7.84609675e-01 -2.34221086e-01
-2.61025310e-01 -2.04739273e-01 -7.16400802e-01 1.26475856e-01
1.25545073e+00 -1.61423609e-01 1.82107374e-01 8.86109710e-01
4.72712159e-01 2.37441286e-01 -7.61438012e-01 9.60270643e-01
1.57420307e-01 -1.26114821e+00 -7.82340392e-03 -6.45773998e-03
6.27585351e-01 4.14170206e-01 4.01721597e-01 5.02134860e-01
1.07838742e-01 -7.58378804e-01 6.14123523e-01 1.39064535e-01
2.63854116e-01 -6.91003680e-01 9.82862175e-01 1.33529514e-01
-9.74158823e-01 -1.06814481e-01 -3.90240222e-01 -5.69299400e-01
1.14304554e-02 9.92556393e-01 -6.78487659e-01 2.03940988e-01
7.80196726e-01 8.28184426e-01 -4.75689113e-01 2.73727298e-01
-7.74552254e-03 6.59718812e-01 -2.42704302e-01 -7.31021643e-01
5.99048138e-01 -3.13115060e-01 1.46955043e-01 1.02374423e+00
2.70151973e-01 -9.52154920e-02 1.94918409e-01 2.98642278e-01
-2.97629714e-01 7.24650621e-01 -5.22158146e-01 -5.85997760e-01
-1.57786921e-01 1.43646324e+00 -7.38856792e-01 -4.88722861e-01
-5.86743295e-01 5.78201413e-01 4.22479331e-01 3.88382465e-01
-5.63794971e-01 -2.97309428e-01 5.79956532e-01 -8.65467712e-02
6.41577244e-02 -2.51266658e-01 -3.44711989e-01 -1.66811097e+00
4.72296178e-02 -9.66732085e-01 3.74423146e-01 -4.48428929e-01
-1.39950204e+00 9.83859956e-01 -7.35115707e-01 -1.27301264e+00
-3.25352192e-01 -7.52676606e-01 -4.53540266e-01 5.90683460e-01
-1.53455532e+00 -1.05687129e+00 2.25849837e-01 6.08869970e-01
4.31716710e-01 -6.06392771e-02 8.83667946e-01 4.26205128e-01
-9.90875363e-01 3.66325080e-01 9.14261416e-02 2.93964416e-01
6.41508102e-01 -9.35798585e-01 2.68276513e-01 6.87193513e-01
-3.15580703e-02 7.90393353e-01 7.49865413e-01 -2.40277007e-01
-1.35156775e+00 -8.99200797e-01 1.59374928e+00 -1.20035872e-01
1.19585824e+00 -3.63112241e-01 -8.02668393e-01 7.10132003e-01
1.86701700e-01 -4.68372107e-02 1.20712662e+00 9.31256056e-01
-4.88732010e-01 -4.17516902e-02 -6.18641019e-01 8.69403780e-01
2.59020627e-01 -8.64054263e-01 -4.09826428e-01 4.12098259e-01
7.37012029e-01 1.51763465e-02 -8.57473910e-01 1.75858200e-01
7.39929438e-01 -8.52561891e-01 4.35073137e-01 -7.17442513e-01
4.51495141e-01 -8.20453987e-02 -1.69195399e-01 -1.13112330e+00
-2.17408031e-01 -6.82763457e-01 1.81022942e-01 9.37278092e-01
8.38255107e-01 -8.27077448e-01 9.55870986e-01 8.53756130e-01
1.52329922e-01 -8.80766392e-01 -5.86291313e-01 -4.04701948e-01
1.47369564e-01 -7.06889570e-01 3.13837022e-01 1.30970085e+00
5.22524536e-01 8.60141695e-01 -6.15796745e-01 -1.80285022e-01
-4.83123399e-02 3.47458899e-01 5.64221382e-01 -1.19389725e+00
1.37713060e-01 -5.57256639e-01 -2.01576561e-01 -9.44553077e-01
1.77390277e-01 -7.15842605e-01 -3.03426385e-01 -1.13442540e+00
-2.09964290e-02 -3.66333842e-01 -6.61208391e-01 4.21443164e-01
1.04568556e-01 3.03330570e-01 4.72566225e-02 3.02247167e-01
-6.28119349e-01 6.16335869e-01 1.00507009e+00 -2.21281797e-01
-3.79182756e-01 5.01133464e-02 -8.34301054e-01 8.46270800e-01
1.16257250e+00 -4.15496111e-01 -1.88106030e-01 -5.51251411e-01
1.30538154e+00 -1.45637527e-01 -9.95032340e-02 -4.20955449e-01
3.00161839e-01 -1.09742880e-01 -9.45913792e-03 -6.11894906e-01
4.27217364e-01 -4.64977235e-01 -3.57554406e-01 2.92962343e-01
-5.45895815e-01 4.33220357e-01 -8.83125663e-02 3.78190100e-01
-6.53311610e-01 -3.72826345e-02 4.34768170e-01 -1.15791060e-01
-3.70545030e-01 2.60592401e-01 -8.56505394e-01 -3.42198312e-01
5.61440349e-01 9.42766294e-02 -2.03966811e-01 -4.68302816e-01
-6.66929901e-01 1.57215774e-01 1.35737717e-01 7.06187963e-01
6.80320084e-01 -1.13066626e+00 -5.43642402e-01 1.11540958e-01
1.22213393e-01 -2.27753982e-01 3.43787163e-01 1.09539390e+00
-4.26093191e-01 5.52011490e-01 2.18017906e-01 -3.19362909e-01
-1.21797895e+00 4.96051252e-01 8.59071165e-02 -6.12634301e-01
-3.53047132e-01 7.90454149e-01 5.51221482e-02 -5.97200215e-01
7.10231513e-02 -5.25374174e-01 -7.51973629e-01 6.85610831e-01
5.81036210e-01 1.07088327e-01 3.39561969e-01 -7.95867324e-01
-2.82321244e-01 5.01710296e-01 -4.35754389e-01 -1.16580494e-01
1.29272377e+00 -5.67823723e-02 -5.04162431e-01 1.02862942e+00
1.41247356e+00 1.66796874e-02 -1.62676707e-01 -6.07714713e-01
2.26514742e-01 -4.75071184e-02 3.76303911e-01 -2.32229963e-01
-1.15845740e+00 1.12289453e+00 2.35006139e-01 6.48341179e-01
8.07346821e-01 -3.55673373e-01 1.04429948e+00 8.57345581e-01
8.46394617e-03 -1.45612526e+00 1.09018162e-01 9.32291806e-01
7.45092332e-01 -1.30485392e+00 -4.47081290e-02 -9.43870246e-02
-8.84178281e-01 1.16624105e+00 2.50289142e-01 -1.42834544e-01
8.86989176e-01 -1.31194264e-01 5.57033777e-01 -3.52142185e-01
-1.05312681e+00 -5.37591614e-02 1.81860536e-01 -3.18675280e-01
1.00554609e+00 1.71237692e-01 -5.29307842e-01 4.42770749e-01
-3.45897704e-01 -2.95654625e-01 6.42717659e-01 8.03412318e-01
-2.35315919e-01 -1.08309615e+00 -2.48798788e-01 5.10474384e-01
-1.12606156e+00 -2.77131766e-01 -4.41131920e-01 3.39963064e-02
-4.39806104e-01 1.31179821e+00 -4.83769365e-02 -6.04102135e-01
7.03729466e-02 3.83446105e-02 -2.58543581e-01 -6.29394114e-01
-1.11470664e+00 1.94801629e-01 2.99641669e-01 -5.25602162e-01
-7.05330908e-01 -5.40253758e-01 -1.04881549e+00 -5.66223919e-01
-6.12346470e-01 4.90019739e-01 1.06613243e+00 9.64059889e-01
5.00264168e-01 3.78357798e-01 1.03018308e+00 -6.80866182e-01
-5.17283440e-01 -1.04526436e+00 -4.41111475e-01 2.21824422e-01
3.56510878e-01 -2.93415666e-01 -5.17558515e-01 -1.56608745e-01] | [11.180615425109863, 6.984334945678711] |
b9638db8-8d10-4cf3-9206-9f0377c4c8a2 | gated-relational-graph-attention-networks | null | null | https://openreview.net/forum?id=v-9E8egy_i | https://openreview.net/pdf?id=v-9E8egy_i | Gated Relational Graph Attention Networks | Relational Graph Neural Networks (GNN) are a class of GNN that are capable of handling multi-relational graphs. Like all GNNs, they suffer from a drop in performance when training deeper networks, which may be caused by vanishing gradients, over-parameterization, and oversmoothing. Previous works have investigated methods that improve the training of deeper GNNs, which include normalization techniques and various types of skip connection within a node. However, learning long-range patterns in multi-relational graphs using GNNs remains an under-explored topic. In this work, we propose a novel GNN architecture based on the Graph Attention Network (GAT) that uses gated skip connections to improve long-range modeling between nodes and uses a more scalable vector-based approach for parameterizing relations. We perform an extensive experimental analysis on synthetic and real data, focusing explicitly on learning long-range patterns. The results indicate that the proposed method significantly outperforms several commonly used relational GNN variants when used in deeper configurations and stays competitive to existing architectures in a shallow setup. | ['Asja Fischer', 'Denis Lukovnikov'] | 2021-01-01 | null | null | null | null | ['long-range-modeling'] | ['natural-language-processing'] | [-2.38305535e-02 3.79768699e-01 -2.06757590e-01 -4.56409305e-01
-7.20662437e-03 -2.06511065e-01 7.30438709e-01 4.11812454e-01
-3.79500270e-01 4.39122289e-01 1.37723640e-01 -5.46811879e-01
-3.78474891e-01 -1.16950119e+00 -9.67582583e-01 -4.43895906e-01
-4.55601037e-01 7.56145895e-01 5.33789515e-01 -5.54205596e-01
-2.87456643e-02 9.21831727e-01 -1.11453843e+00 1.34294137e-01
4.42296028e-01 7.09639013e-01 -1.73656985e-01 8.14314127e-01
-2.69388974e-01 1.25290799e+00 -7.01495409e-01 -6.15696490e-01
3.78479585e-02 -1.18194945e-01 -8.23016822e-01 -3.76867294e-01
5.97957790e-01 7.12765083e-02 -5.84733486e-01 8.57167959e-01
4.00078148e-01 3.67062479e-01 3.42849731e-01 -1.22409105e+00
-9.17040467e-01 1.22853994e+00 -4.41764355e-01 3.67252082e-01
3.67210270e-03 -2.01980338e-01 1.46448588e+00 -6.54653966e-01
5.62072873e-01 1.46784806e+00 1.15844321e+00 2.96473652e-01
-1.32605040e+00 -4.28417653e-01 4.06175107e-01 1.75526798e-01
-1.54533911e+00 -2.00274400e-02 8.07433784e-01 -9.10416022e-02
1.52261782e+00 8.99825171e-02 6.04570687e-01 1.07368648e+00
2.42079437e-01 4.69641417e-01 3.72168243e-01 -4.25589025e-01
-2.14770600e-01 -8.13244656e-02 2.75877804e-01 7.43230581e-01
5.89440465e-01 -1.66423023e-01 -2.24592820e-01 -1.92594659e-02
1.02956522e+00 2.46475041e-02 -4.90456112e-02 -5.42808831e-01
-1.02722180e+00 9.76669610e-01 1.24870181e+00 6.84956193e-01
-3.24523747e-01 5.00321031e-01 4.58714306e-01 3.67291838e-01
5.54535687e-01 5.78881919e-01 -4.58192766e-01 2.28381202e-01
-6.12739742e-01 1.93981882e-02 9.26909804e-01 8.08511317e-01
8.04094672e-01 1.72999203e-01 -2.22296193e-01 9.68337536e-01
2.51234978e-01 -3.09329689e-03 3.88388246e-01 -2.67694861e-01
6.92611635e-01 1.11359644e+00 -7.50299513e-01 -1.31727493e+00
-8.76597464e-01 -7.17189491e-01 -1.21374309e+00 -1.37181655e-01
2.83844322e-01 -2.69937422e-02 -1.17481327e+00 1.66238999e+00
5.92843592e-02 1.16117522e-01 4.68133315e-02 6.66612446e-01
1.26189923e+00 4.72735405e-01 2.31396407e-01 4.09848988e-01
8.41779113e-01 -1.30163920e+00 -6.11122072e-01 -3.08396101e-01
8.92171800e-01 -3.17682117e-01 1.10566628e+00 1.39952987e-01
-9.46016550e-01 -5.92736900e-01 -1.02441859e+00 -3.54465127e-01
-9.24503982e-01 -8.33614469e-02 1.08375013e+00 5.52191079e-01
-1.59971130e+00 1.00207877e+00 -7.62354136e-01 -7.21212685e-01
4.49113399e-01 7.02011585e-01 -5.44996083e-01 1.61307122e-04
-1.28932631e+00 8.90102446e-01 5.87221384e-01 4.58191484e-01
-4.86934185e-01 -6.34668708e-01 -1.09976816e+00 4.30191040e-01
5.65657854e-01 -5.57652175e-01 7.86063552e-01 -6.99367881e-01
-1.35440326e+00 6.89450979e-01 4.63413984e-01 -6.37460709e-01
1.09226406e-01 -2.14794904e-01 -3.47114801e-01 -1.06331602e-01
-3.50006819e-01 4.87187356e-01 3.22397232e-01 -1.05827570e+00
1.29639551e-01 -1.91287041e-01 3.81406933e-01 8.39588270e-02
-4.38133121e-01 -1.34787560e-01 -6.24265194e-01 -7.54991293e-01
1.44418776e-01 -8.65754306e-01 -4.62760597e-01 -5.11467934e-01
-8.36974025e-01 -4.64225799e-01 6.74287915e-01 -2.03777894e-01
1.19108260e+00 -1.85993040e+00 2.44956449e-01 5.30012369e-01
3.96596670e-01 6.29376769e-01 -4.04151440e-01 6.93817854e-01
-4.28635150e-01 3.16985369e-01 3.30478810e-02 -4.18730527e-01
-1.87073022e-01 5.81042767e-01 2.09142372e-01 2.38281608e-01
3.46935898e-01 1.24794912e+00 -6.13240600e-01 -3.27091336e-01
1.45474747e-01 8.05694878e-01 -4.96324480e-01 1.54319108e-01
-3.68251503e-01 -8.67784545e-02 -8.16886798e-02 4.81064558e-01
5.83495617e-01 -7.05594301e-01 4.04865801e-01 -2.15095237e-01
3.23925585e-01 3.09991628e-01 -9.65823293e-01 1.52142417e+00
-4.22796816e-01 6.14555776e-01 -2.26356730e-01 -1.23063600e+00
1.09342885e+00 1.56911612e-02 2.15129480e-01 -5.15241623e-01
3.17281783e-01 -7.07512870e-02 3.75675648e-01 -3.12023520e-01
6.48054659e-01 1.66979998e-01 1.75877139e-01 2.64654040e-01
4.90349948e-01 1.29829079e-01 3.86643589e-01 4.30044770e-01
1.36672068e+00 -7.57385716e-02 1.32221028e-01 -2.33442441e-01
5.29156864e-01 -3.45337421e-01 9.28722396e-02 9.76948082e-01
3.23258907e-01 4.77105170e-01 9.64394152e-01 -7.15697825e-01
-8.93641949e-01 -6.26060367e-01 2.66789347e-01 1.38073575e+00
-3.34523618e-02 -7.16327608e-01 -4.59724963e-01 -7.76570618e-01
-1.42942993e-02 4.34491336e-01 -9.02768910e-01 -2.51739115e-01
-7.77486205e-01 -9.78225589e-01 6.07911825e-01 8.50872338e-01
2.06661746e-01 -1.37517726e+00 5.86924255e-02 2.95558780e-01
4.10625041e-01 -1.37681937e+00 4.37868908e-02 5.20041168e-01
-9.55486953e-01 -1.05793846e+00 -5.54043472e-01 -7.66345859e-01
6.15932822e-01 5.25611965e-03 1.51925242e+00 6.00277424e-01
-5.65822162e-02 7.81037584e-02 -3.74151647e-01 -1.63417980e-01
-3.33159745e-01 9.27835763e-01 -3.61268729e-01 -1.17536493e-01
2.48321056e-01 -8.11022699e-01 -1.35284156e-01 1.12020463e-01
-8.42070818e-01 -1.71243057e-01 9.66920853e-01 7.72397757e-01
3.05035114e-01 8.82138610e-02 3.26263815e-01 -1.68817043e+00
9.37373579e-01 -5.82719862e-01 -6.12616241e-01 3.23431700e-01
-7.15961754e-01 4.18406487e-01 8.43232334e-01 -5.39814889e-01
-6.61197960e-01 -3.45881313e-01 -3.60124499e-01 -4.47151363e-01
1.99158061e-02 7.69634664e-01 -5.22446446e-02 -4.68879580e-01
7.17371047e-01 -3.64109486e-01 -1.72131315e-01 -4.22913998e-01
4.83759552e-01 -2.60186065e-02 2.02396363e-01 -4.59046483e-01
8.53619635e-01 1.12597480e-01 4.80517000e-01 -7.88439274e-01
-7.36200511e-01 -2.70492345e-01 -7.40170896e-01 6.06584586e-02
5.62581778e-01 -5.86305320e-01 -7.05377996e-01 4.40513581e-01
-1.04115713e+00 -8.84392917e-01 -1.87463462e-01 2.95570761e-01
-4.27331701e-02 1.65212676e-01 -1.01842773e+00 -4.00528103e-01
-3.21384162e-01 -9.01309490e-01 8.22643816e-01 2.48080894e-01
-5.28923199e-02 -1.54696000e+00 5.80891371e-02 -9.89398211e-02
9.02855396e-01 2.37837121e-01 1.07070160e+00 -1.15437150e+00
-6.22275412e-01 -2.10803226e-01 -5.90230405e-01 1.90685958e-01
-1.79553330e-01 1.70042753e-01 -7.84893036e-01 -2.64503598e-01
-7.26537526e-01 -3.46287936e-01 1.10213101e+00 2.88815707e-01
1.05647206e+00 -1.67145818e-01 -4.61882800e-01 8.40548098e-01
1.54905522e+00 -1.78544521e-01 7.04781413e-01 3.61448169e-01
1.34129548e+00 5.05206883e-01 -6.49055615e-02 -1.18323348e-01
5.81207514e-01 4.45906132e-01 8.81995440e-01 -4.71633166e-01
-2.03092650e-01 -1.48787439e-01 -1.04186274e-01 7.24225223e-01
-4.11401540e-01 -6.61299765e-01 -1.00429010e+00 5.09783864e-01
-1.95664942e+00 -5.38407207e-01 -2.34715953e-01 1.70398438e+00
3.72890383e-01 4.97880697e-01 1.57780036e-01 1.98593080e-01
6.29437864e-01 3.85370761e-01 -3.46327931e-01 -7.07427084e-01
-2.54014313e-01 6.10309482e-01 6.97463751e-01 4.47026700e-01
-1.00653088e+00 1.27013350e+00 6.23996496e+00 4.72616076e-01
-1.22955930e+00 -6.24841899e-02 4.61016208e-01 2.36607328e-01
-2.78366387e-01 -2.01518849e-01 -8.13378572e-01 -2.70686328e-01
1.16767383e+00 2.87470609e-01 2.39527583e-01 8.70862544e-01
-4.58911777e-01 3.08549851e-01 -1.02699780e+00 7.02615619e-01
1.48654029e-01 -1.46314406e+00 2.04898059e-01 -8.02287646e-03
6.24899864e-01 3.75446916e-01 -3.77832353e-02 5.97635627e-01
7.05947220e-01 -1.33752215e+00 2.21921623e-01 3.81049722e-01
4.03382659e-01 -7.12614894e-01 1.16924787e+00 -1.26603052e-01
-1.32938457e+00 1.98579617e-02 -6.20407879e-01 -1.32988498e-01
-1.55996248e-01 5.67146838e-01 -9.98938918e-01 7.44609535e-01
7.45294511e-01 8.47557783e-01 -1.09837198e+00 8.80650818e-01
-3.45094591e-01 4.79068041e-01 -4.00828600e-01 -3.58593851e-01
5.47487676e-01 -9.05844867e-02 2.42798969e-01 1.33752763e+00
1.03798226e-01 -2.74193615e-01 -5.24641462e-02 7.73976862e-01
-4.28228885e-01 2.33647525e-01 -8.33901942e-01 -2.27497384e-01
1.24477006e-01 1.55547607e+00 -9.38280642e-01 -1.23882107e-01
-5.34853518e-01 6.10927880e-01 1.02874136e+00 4.87317741e-01
-7.63271213e-01 -5.06965995e-01 2.73988634e-01 2.88610399e-01
5.72238386e-01 -1.67904675e-01 -2.71070618e-02 -8.60957086e-01
-8.43438357e-02 -5.54845870e-01 6.91586018e-01 -7.01400340e-01
-1.34160364e+00 1.23382568e+00 5.35840122e-03 -4.54960406e-01
-1.42547026e-01 -8.66809666e-01 -5.58688045e-01 6.90266013e-01
-1.47935581e+00 -1.46474278e+00 -3.85758758e-01 6.22364640e-01
-7.29157552e-02 -1.54991642e-01 7.55317271e-01 3.80722761e-01
-6.22102916e-01 8.16355765e-01 -3.52961004e-01 3.90264452e-01
4.25316334e-01 -1.40364194e+00 7.20082283e-01 8.29671085e-01
3.89138818e-01 7.46214390e-01 4.78268772e-01 -4.95549917e-01
-1.14686728e+00 -1.38989258e+00 9.84099746e-01 -1.02885999e-01
9.05377626e-01 -6.59199238e-01 -1.28908670e+00 1.30826879e+00
2.43383557e-01 4.92642313e-01 4.19525325e-01 7.28384078e-01
-5.37577152e-01 -2.53115058e-01 -8.88681233e-01 6.19635642e-01
1.25414228e+00 -3.86221826e-01 -1.71585679e-01 3.80599111e-01
8.86145413e-01 -4.93035734e-01 -1.14828205e+00 5.63710153e-01
1.25962004e-01 -1.05404329e+00 1.12762916e+00 -5.69496214e-01
2.23079994e-01 8.20753053e-02 1.48652673e-01 -1.41579473e+00
-5.40943563e-01 -5.40321946e-01 -1.12390265e-01 1.15303576e+00
6.16720915e-01 -8.17420363e-01 1.04143417e+00 2.71885842e-01
-1.65821314e-01 -9.35640097e-01 -4.25388455e-01 -6.34925067e-01
2.26972392e-03 -2.22388029e-01 6.71833754e-01 9.65774298e-01
-2.47763813e-01 8.09310138e-01 -3.69030714e-01 2.26913229e-01
3.14767331e-01 -2.17606828e-01 8.49141181e-01 -1.37924492e+00
-2.14291483e-01 -7.09022641e-01 -7.93499827e-01 -8.49386156e-01
3.91494542e-01 -1.05942857e+00 -1.83198005e-01 -1.75847960e+00
-9.61767882e-02 -6.37806892e-01 -3.84781420e-01 6.76171184e-01
-1.81381002e-01 3.51089031e-01 3.44637223e-03 -2.33119741e-01
-6.41277075e-01 3.50089639e-01 1.14045835e+00 -8.91195089e-02
-1.87593579e-01 1.72125623e-02 -6.54018462e-01 5.57151437e-01
8.40863407e-01 -4.27967936e-01 -5.45584083e-01 -4.59137857e-01
7.34881997e-01 -1.99137539e-01 2.87008286e-01 -1.02833200e+00
4.25489664e-01 2.83277065e-01 2.50643015e-01 -6.38957083e-01
3.15874338e-01 -7.94964910e-01 1.83190629e-01 2.62929976e-01
-3.72387052e-01 4.81187284e-01 3.15370321e-01 4.95040685e-01
-2.52566963e-01 -9.69989225e-02 6.33094549e-01 -1.63109854e-01
-5.41018903e-01 4.51766759e-01 -7.44337887e-02 -4.88075428e-02
7.10151911e-01 -2.08482649e-02 -4.35534567e-01 -4.83088970e-01
-7.71796703e-01 1.61836237e-01 1.00038610e-01 4.08656478e-01
4.50144768e-01 -1.26207900e+00 -4.11951721e-01 3.71194959e-01
-1.63306426e-02 3.66874367e-01 -1.67286713e-02 6.84536278e-01
-7.74434805e-01 5.05493581e-01 -7.40255639e-02 -5.11693358e-01
-1.20073259e+00 7.02192843e-01 5.31223655e-01 -9.39133644e-01
-7.56461501e-01 1.23889077e+00 -1.37202039e-01 -8.14679384e-01
3.92018944e-01 -6.66952312e-01 -5.99656284e-01 -1.12629913e-01
-7.94554316e-03 1.30602881e-01 4.24582899e-01 -6.70575619e-01
-3.06060076e-01 5.47645807e-01 -3.00781280e-01 4.66123968e-01
1.66162121e+00 2.41060331e-01 -4.01054263e-01 4.76549566e-01
1.17169273e+00 -2.40004301e-01 -7.30751932e-01 -4.75772381e-01
3.34192574e-01 3.92405391e-02 -2.07848206e-01 -5.14250398e-01
-1.56979036e+00 8.51653397e-01 2.58752201e-02 6.81750596e-01
8.62632871e-01 2.07651705e-01 5.24810076e-01 5.73951423e-01
-4.50518541e-02 -6.22733355e-01 2.46412739e-01 8.17828119e-01
8.23210239e-01 -1.16578591e+00 1.13536187e-01 -5.14130712e-01
-3.02934855e-01 1.27902067e+00 9.25149500e-01 -4.73297179e-01
8.43541741e-01 2.01657459e-01 -4.13905531e-02 -6.35928214e-01
-8.09204400e-01 -2.66867578e-01 3.77211958e-01 5.80174029e-01
6.17093325e-01 -1.02704540e-01 9.32014883e-02 1.29435882e-01
-2.19724894e-01 -3.63612235e-01 4.80098784e-01 6.59515142e-01
1.30826414e-01 -1.22535217e+00 7.03019127e-02 4.93496299e-01
-5.37582755e-01 -4.06868517e-01 -4.59853053e-01 1.34167814e+00
-1.11784071e-01 6.29416645e-01 1.72635913e-01 -5.15277088e-01
5.64876914e-01 -2.02258244e-01 5.42731583e-01 -8.26464176e-01
-1.03138244e+00 -2.99588591e-01 3.25106144e-01 -7.88017988e-01
-5.29402614e-01 -1.81320295e-01 -1.11314046e+00 -4.40257996e-01
-4.66892898e-01 9.39763933e-02 4.68082666e-01 7.07130015e-01
2.85496294e-01 1.10055327e+00 -3.04421335e-02 -7.98052728e-01
-1.24022238e-01 -1.31234896e+00 -4.79612768e-01 5.62549710e-01
2.90186346e-01 -6.65452719e-01 -2.71331578e-01 -6.13084257e-01] | [6.952338695526123, 6.254016399383545] |
d1c58303-bf92-4b19-a685-231f9395b1f6 | end-to-end-joint-entity-extraction-and | 1812.05270 | null | https://arxiv.org/abs/1812.05270v5 | https://arxiv.org/pdf/1812.05270v5.pdf | Joint Entity Extraction and Assertion Detection for Clinical Text | Negative medical findings are prevalent in clinical reports, yet discriminating them from positive findings remains a challenging task for information extraction. Most of the existing systems treat this task as a pipeline of two separate tasks, i.e., named entity recognition (NER) and rule-based negation detection. We consider this as a multi-task problem and present a novel end-to-end neural model to jointly extract entities and negations. We extend a standard hierarchical encoder-decoder NER model and first adopt a shared encoder followed by separate decoders for the two tasks. This architecture performs considerably better than the previous rule-based and machine learning-based systems. To overcome the problem of increased parameter size especially for low-resource settings, we propose the Conditional Softmax Shared Decoder architecture which achieves state-of-art results for NER and negation detection on the 2010 i2b2/VA challenge dataset and a proprietary de-identified clinical dataset. | ['Busra Celikkaya', 'Parminder Bhatia', 'Mohammed Khalilia'] | 2018-12-13 | joint-entity-extraction-and-assertion | https://aclanthology.org/P19-1091 | https://aclanthology.org/P19-1091.pdf | acl-2019-7 | ['negation-detection'] | ['natural-language-processing'] | [ 4.10503626e-01 3.62675756e-01 -2.96825409e-01 -5.63448548e-01
-1.41349530e+00 -4.02805179e-01 2.75282800e-01 5.57312727e-01
-1.03890574e+00 9.23011303e-01 3.17354679e-01 -5.38401783e-01
2.17351735e-01 -4.17800754e-01 -5.96963465e-01 -3.36516470e-01
6.65910766e-02 4.90748137e-01 -1.00138821e-02 4.74826284e-02
-2.20240369e-01 -6.04776405e-02 -8.14700603e-01 8.28639090e-01
6.15947723e-01 8.27744842e-01 -7.67965391e-02 9.88311529e-01
4.77283560e-02 1.31909573e+00 -6.09483123e-01 -8.29985023e-01
-2.78398484e-01 -2.11827129e-01 -1.13295126e+00 -5.30201554e-01
5.12351394e-02 -4.26794916e-01 -1.91650286e-01 1.00373530e+00
1.08907366e+00 -1.50283039e-01 4.94836777e-01 -6.03754342e-01
-7.55135775e-01 7.48437047e-01 -3.17107350e-01 2.24664196e-01
1.68817505e-01 -1.16152473e-01 9.83323097e-01 -7.24673808e-01
8.91280174e-01 8.74586225e-01 9.38731015e-01 9.61651504e-01
-9.46799159e-01 -5.31051040e-01 -1.96280211e-01 -3.90639305e-02
-1.49596286e+00 -5.18251538e-01 1.65436357e-01 -3.57092440e-01
1.59326315e+00 2.43184362e-02 9.26907510e-02 1.29642785e+00
4.67714012e-01 1.03128946e+00 5.30158341e-01 -1.70712739e-01
1.32061228e-01 8.27357993e-02 7.25213513e-02 7.14595377e-01
2.84333348e-01 -3.20077598e-01 -3.54734421e-01 -4.36255485e-01
2.12486431e-01 2.02151685e-04 -3.20775092e-01 4.30033505e-01
-1.19821632e+00 7.65607297e-01 1.85927913e-01 3.74908686e-01
-6.98171556e-01 5.97983710e-02 7.67336249e-01 1.92084253e-01
3.76385778e-01 2.53298283e-01 -1.06792915e+00 -6.89062700e-02
-1.19303703e+00 -9.98590887e-02 1.14410436e+00 7.78478801e-01
-9.95748788e-02 -3.89203161e-01 -5.79713106e-01 8.74303102e-01
3.88049841e-01 2.19386905e-01 6.05517924e-01 -1.37453064e-01
5.04492760e-01 5.35512090e-01 -1.53432041e-01 -5.28453410e-01
-9.37475920e-01 -5.55315077e-01 -1.03747976e+00 -3.01209301e-01
-1.25759887e-02 -7.99956441e-01 -1.27759457e+00 1.91604495e+00
2.20901445e-01 1.37440622e-01 5.04654706e-01 5.28006196e-01
1.47762799e+00 3.45880926e-01 5.80509245e-01 7.00660571e-02
1.85438192e+00 -8.74559581e-01 -1.13808322e+00 -1.81065276e-01
1.12136745e+00 -6.01717293e-01 7.16267824e-02 3.06448728e-01
-1.12264752e+00 -2.06081439e-02 -8.87522757e-01 -5.72254896e-01
-6.47542119e-01 5.72266936e-01 4.13552374e-01 5.20131707e-01
-1.08084595e+00 2.73048252e-01 -1.17427123e+00 -1.44332901e-01
5.81667066e-01 6.79557800e-01 -5.75830996e-01 2.07913026e-01
-1.49650085e+00 1.10956442e+00 6.45231962e-01 3.59527439e-01
-6.91873372e-01 -7.21199691e-01 -1.07000184e+00 2.29756624e-01
4.32920814e-01 -9.78870332e-01 1.46546268e+00 -3.83232772e-01
-1.08170426e+00 1.21935034e+00 -2.33574912e-01 -6.12348616e-01
2.81873375e-01 -1.84738383e-01 -6.89385414e-01 -3.22506875e-01
1.20805964e-01 6.29351854e-01 1.33665323e-01 -6.01815164e-01
-8.06510091e-01 -1.35185748e-01 -3.55241120e-01 -1.74676836e-01
-1.89782575e-01 3.93538505e-01 -6.30329907e-01 -4.67297822e-01
-5.38664281e-01 -6.07485771e-01 -5.33501744e-01 -2.42653802e-01
-1.02678370e+00 -3.97962600e-01 1.93591624e-01 -9.38079000e-01
1.39945483e+00 -2.05533671e+00 2.66289990e-02 -1.13504246e-01
5.83977163e-01 5.79091191e-01 -5.85179478e-02 1.02044031e-01
-3.63990992e-01 4.85983640e-01 -2.86190063e-01 -5.31128883e-01
-1.26635298e-01 -2.45633982e-02 3.72346342e-01 2.40928352e-01
7.94826329e-01 1.01921582e+00 -8.68197501e-01 -5.03199875e-01
-4.00839448e-01 9.18372810e-01 -6.09143674e-01 2.51526743e-01
-7.00585768e-02 1.85541883e-01 -2.70488292e-01 7.86544144e-01
6.03626311e-01 -7.03657866e-01 5.21387160e-01 -2.64948487e-01
1.44505650e-01 6.35928035e-01 -1.08993363e+00 1.55167568e+00
-2.19420820e-01 3.76894742e-01 4.45225000e-01 -6.53781474e-01
4.76536930e-01 8.77837956e-01 4.08685476e-01 -4.98140693e-01
3.36512715e-01 5.17137349e-01 4.65673991e-02 -8.12990248e-01
2.93912530e-01 -4.74411935e-01 -2.90130973e-01 -9.64743346e-02
4.49054241e-01 4.83917326e-01 5.15213758e-02 5.62497275e-03
1.70822597e+00 -1.63749143e-01 8.09713304e-01 5.83116896e-02
6.20932579e-01 -1.22336425e-01 1.18355477e+00 7.29200780e-01
-3.06367934e-01 4.70499009e-01 7.12905705e-01 -1.43830150e-01
-7.00191021e-01 -5.59284031e-01 -4.08551365e-01 9.09699142e-01
-6.07102931e-01 -5.86572230e-01 -4.02836949e-01 -1.03352284e+00
4.08470817e-02 4.91515368e-01 -8.47681880e-01 -1.45493686e-01
-5.41898370e-01 -1.28657627e+00 1.18382430e+00 7.25624144e-01
1.57199487e-01 -9.94723439e-01 -4.42300916e-01 4.88107979e-01
-2.70705611e-01 -1.36750340e+00 -4.02283639e-01 1.03907776e+00
-3.70112240e-01 -1.17445683e+00 -9.04580772e-01 -9.74445939e-01
3.71073306e-01 -7.65641212e-01 1.30990720e+00 -1.45211771e-01
-3.75505477e-01 5.77211939e-02 -2.21171841e-01 -7.57300377e-01
-5.18681049e-01 3.65520388e-01 -2.62504995e-01 -3.31004500e-01
8.41218829e-01 1.40031263e-01 -5.90512455e-01 -3.16124707e-01
-1.08082926e+00 2.05519889e-02 1.07819188e+00 1.24340475e+00
8.23533297e-01 -2.75289506e-01 8.81120265e-01 -1.31441236e+00
8.60201061e-01 -7.51047850e-01 -1.85460269e-01 5.99334896e-01
-3.24174136e-01 2.15299949e-01 5.29339254e-01 -1.19534418e-01
-1.05427134e+00 2.22469375e-01 -7.89634228e-01 3.87955345e-02
-2.73633361e-01 9.35603380e-01 -7.37866238e-02 3.81792367e-01
3.60197872e-01 8.22411627e-02 -3.69812429e-01 -6.18433416e-01
1.01185009e-01 1.11962378e+00 6.27575636e-01 1.61560610e-01
1.06127091e-01 1.19280450e-01 -9.77020860e-02 -5.55539131e-01
-1.12759948e+00 -6.38559282e-01 -4.78182524e-01 5.16311407e-01
1.41869497e+00 -1.24167693e+00 -8.83275449e-01 2.93817818e-01
-1.43745816e+00 -2.51231700e-01 -4.45692390e-02 6.00382566e-01
-1.49290815e-01 2.01390535e-01 -1.11976576e+00 -6.00264668e-01
-9.45445299e-01 -1.27982223e+00 1.26508605e+00 9.60897505e-02
-2.48584673e-01 -1.05480731e+00 3.96033585e-01 2.04079077e-01
2.95900553e-01 3.29575658e-01 1.04709232e+00 -1.41199958e+00
7.77486116e-02 -3.41236532e-01 -2.44047731e-01 3.40541184e-01
-5.89799583e-02 -3.50349277e-01 -8.70549321e-01 -2.28451714e-01
-4.11410294e-02 -5.18158138e-01 1.21225703e+00 2.42676288e-01
1.01824558e+00 -1.05961919e-01 -6.96501732e-01 8.26132238e-01
1.43530452e+00 1.43256009e-01 3.69915158e-01 2.41882771e-01
5.80414057e-01 5.49992695e-02 2.18448741e-03 3.54631782e-01
6.20846808e-01 2.99478173e-01 1.01524681e-01 -5.23995101e-01
-1.93220191e-02 1.05485998e-01 1.26863807e-01 6.17155790e-01
3.27914268e-01 -6.50914967e-01 -1.25556171e+00 7.85312235e-01
-1.68850887e+00 -5.35172522e-01 -1.24558330e-01 1.71986568e+00
1.40361595e+00 9.10174325e-02 -2.96758503e-01 -2.68585175e-01
6.13822758e-01 -2.13892207e-01 -5.29404759e-01 -5.28705657e-01
-2.60778397e-01 4.31310743e-01 6.32191896e-01 6.47658929e-02
-1.53427839e+00 6.34418726e-01 6.52651644e+00 5.61871886e-01
-1.10940647e+00 4.06042516e-01 7.43852079e-01 -3.40327293e-01
8.29507560e-02 -5.98460972e-01 -1.14393675e+00 3.27861398e-01
1.41732693e+00 3.77316982e-01 -3.25192690e-01 5.30802608e-01
-1.05086714e-01 2.39265397e-01 -1.34482503e+00 9.23577726e-01
8.95112306e-02 -1.44379461e+00 -2.02609330e-01 -1.11012720e-01
4.81931239e-01 7.35100269e-01 -1.34237185e-01 6.39467657e-01
4.24761176e-01 -1.23676944e+00 1.98859066e-01 6.45580471e-01
1.14228296e+00 -4.69307959e-01 1.41253829e+00 1.68186739e-01
-8.24146092e-01 -4.75318059e-02 2.36328654e-02 7.73642182e-01
3.89505595e-01 8.09567392e-01 -9.50563967e-01 6.03655815e-01
5.95967293e-01 5.78923583e-01 -3.67349297e-01 1.10664082e+00
-2.02527434e-01 6.80728316e-01 -3.15868080e-01 2.91974619e-02
1.73855901e-01 6.81327522e-01 3.35097909e-01 1.93996513e+00
-1.91914793e-02 7.86303952e-02 5.86858802e-02 7.36772895e-01
-7.93138683e-01 1.60283580e-01 -2.27079973e-01 -2.07191646e-01
1.29381478e-01 1.32162058e+00 -5.65845191e-01 -4.97697294e-01
-8.74151051e-01 1.00607967e+00 5.61051369e-01 4.09835204e-02
-9.35852349e-01 -7.15063035e-01 3.33441585e-01 -4.50406671e-01
7.25439429e-01 2.85011798e-01 -2.15126157e-01 -1.42740870e+00
-1.22473277e-01 -1.00843680e+00 9.23245549e-01 -2.87792951e-01
-1.31248009e+00 8.18998039e-01 -6.84384942e-01 -9.17684853e-01
-2.71605223e-01 -9.01181340e-01 -2.32278779e-01 9.60003078e-01
-1.88416731e+00 -1.03554535e+00 2.29380026e-01 4.38793749e-01
2.94900294e-02 1.10482603e-01 1.25395572e+00 8.85401785e-01
-9.60022628e-01 1.03211176e+00 2.36191079e-01 7.85892785e-01
1.09306896e+00 -1.42964721e+00 2.72645652e-01 6.55071914e-01
-4.02959794e-01 5.84764063e-01 1.35190994e-01 -8.55780065e-01
-1.17277467e+00 -1.29802215e+00 1.66959596e+00 -4.85871017e-01
3.33389938e-01 -3.95100087e-01 -9.72491384e-01 8.58430088e-01
3.20927262e-01 1.30551666e-01 1.31384671e+00 3.13637406e-01
-1.41484410e-01 4.01615888e-01 -1.31730199e+00 2.47739583e-01
6.18006587e-01 -6.15109384e-01 -6.77467167e-01 3.66863400e-01
6.60528362e-01 -5.68226814e-01 -1.27756965e+00 5.78575313e-01
1.68630615e-01 -3.83505672e-01 7.01377571e-01 -1.03774703e+00
4.40197676e-01 -6.26538470e-02 4.04895283e-02 -1.09255612e+00
-2.33036324e-01 -5.74672997e-01 -3.04320127e-01 1.09342253e+00
1.06637204e+00 -4.22914624e-01 5.88774025e-01 6.25220299e-01
-1.53153175e-02 -9.52941239e-01 -9.16389525e-01 -1.29964247e-01
8.86229426e-02 -2.80812413e-01 2.21433416e-01 1.12138343e+00
2.48004004e-01 6.61509097e-01 -2.04056442e-01 1.80254430e-01
2.03154668e-01 -4.00249362e-01 -1.47142515e-01 -1.13409877e+00
-2.99443781e-01 -3.26609403e-01 -3.65752757e-01 -7.27118969e-01
-1.36458734e-02 -1.32897675e+00 2.74119467e-01 -1.78966749e+00
6.06952906e-01 -4.74347360e-02 -6.97616577e-01 1.09890556e+00
-4.64083403e-01 1.02287009e-01 -3.23108137e-01 -1.29278794e-01
-9.94844317e-01 1.89195573e-01 5.77660978e-01 -1.69106379e-01
-2.33465195e-01 -2.43597016e-01 -8.80617738e-01 5.91929734e-01
4.01883721e-01 -9.48669910e-01 2.47742712e-01 -7.34940469e-01
6.33415520e-01 3.59537542e-01 1.24673881e-01 -6.18314445e-01
5.85171044e-01 3.33157331e-01 5.72804689e-01 -7.49681056e-01
-1.64152086e-02 -5.01725495e-01 -3.58668268e-01 6.44581854e-01
-7.00152636e-01 2.04317309e-02 4.35839385e-01 7.02005506e-01
-3.42349201e-01 -1.27118602e-01 6.07134402e-01 -7.21652657e-02
-4.15568560e-01 1.31343916e-01 -6.88816011e-01 2.55840033e-01
7.95605421e-01 3.20151925e-01 -4.03744608e-01 1.41871711e-02
-1.14908075e+00 6.34583831e-01 -4.40763474e-01 2.23793685e-01
5.75012624e-01 -8.49770010e-01 -1.24926877e+00 -2.35834830e-02
9.22629386e-02 1.80987984e-01 4.46621090e-01 1.28831840e+00
-2.15731636e-01 9.16236699e-01 2.09619194e-01 -3.37430656e-01
-1.38369107e+00 4.55434322e-01 7.30451226e-01 -9.94043052e-01
-4.27309722e-01 1.17418730e+00 -1.27628446e-01 -7.80471146e-01
3.38153303e-01 -5.04578948e-01 -4.77670431e-01 1.44565195e-01
6.66742742e-01 -4.99606095e-02 6.29853427e-01 -3.70291293e-01
-8.54958534e-01 -1.41490772e-01 -5.39667249e-01 3.20628494e-01
1.67734218e+00 3.68312865e-01 -1.23133108e-01 -9.34235156e-02
1.42057502e+00 -2.15642184e-01 -5.26687682e-01 -3.44502807e-01
3.49849135e-01 6.26906395e-01 4.40954328e-01 -1.37697971e+00
-9.39562142e-01 6.58023596e-01 5.35868049e-01 -1.71984091e-01
1.14903617e+00 2.32365075e-02 9.25366163e-01 7.21742034e-01
-3.11666518e-01 -1.04783368e+00 -6.47812068e-01 7.05117643e-01
3.01695228e-01 -1.52468109e+00 -3.36629212e-01 -1.38909176e-01
-5.58620095e-01 1.02132988e+00 3.44104350e-01 3.96499395e-01
7.95829296e-01 8.80980790e-01 7.82260597e-02 -4.11416650e-01
-1.27289951e+00 -2.47784242e-01 4.79599088e-01 2.38195524e-01
1.15958166e+00 2.88910288e-02 -3.96239996e-01 1.39823031e+00
3.19444448e-01 5.68127394e-01 3.47305298e-01 1.15385127e+00
4.40751649e-02 -1.09479702e+00 1.05803542e-01 7.22565174e-01
-1.46929002e+00 -6.36778951e-01 -3.41303527e-01 5.34381866e-01
2.68955022e-01 9.78877962e-01 -5.02969362e-02 -3.45249653e-01
6.11505091e-01 4.94660467e-01 1.95569184e-04 -8.10624063e-01
-1.22721446e+00 9.38021913e-02 5.39728284e-01 -4.33832198e-01
-2.69639611e-01 -3.73478860e-01 -1.51985812e+00 9.77774262e-02
-4.73458797e-01 8.31098035e-02 5.69247901e-01 9.17856336e-01
8.54153872e-01 1.18148303e+00 -1.06702298e-01 9.63657349e-02
-7.42668986e-01 -9.45932448e-01 -2.74812371e-01 3.52165431e-01
5.96244752e-01 -1.73771560e-01 6.80114850e-02 -7.05516711e-02] | [8.566598892211914, 8.76343059539795] |
0d8b6dd8-8d68-401e-a0df-a0dbee862ea5 | privacy-preserving-deep-learning-computation | 2001.02932 | null | https://arxiv.org/abs/2001.02932v1 | https://arxiv.org/pdf/2001.02932v1.pdf | Privacy-Preserving Deep Learning Computation for Geo-Distributed Medical Big-Data Platforms | This paper proposes a distributed deep learning framework for privacy-preserving medical data training. In order to avoid patients' data leakage in medical platforms, the hidden layers in the deep learning framework are separated and where the first layer is kept in platform and others layers are kept in a centralized server. Whereas keeping the original patients' data in local platforms maintain their privacy, utilizing the server for subsequent layers improves learning performance by using all data from each platform during training. | ['Jong-Kook Kim', 'Kwangsoo Kim', 'Junhui Kim', 'Joongheon Kim', 'Joohyung Jeon', 'Aziz Mohaisen'] | 2020-01-09 | null | null | null | null | ['privacy-preserving-deep-learning', 'privacy-preserving-deep-learning'] | ['methodology', 'natural-language-processing'] | [-4.92873937e-01 5.26413202e-01 -4.34643954e-01 -4.47492033e-01
-3.72593850e-01 -6.63372517e-01 -2.14528248e-01 5.16270101e-01
-7.50084341e-01 6.83608830e-01 2.25894839e-01 -3.64384025e-01
2.00724766e-01 -7.87630677e-01 -4.90302801e-01 -1.01866400e+00
-3.20276469e-01 -7.06711262e-02 1.23191327e-01 4.33926463e-01
-4.24068749e-01 6.00301981e-01 -9.04312611e-01 9.07205164e-01
3.45022023e-01 1.19649625e+00 -4.72950637e-01 3.43458265e-01
-5.54915927e-02 9.96064246e-01 -6.73903167e-01 -6.42168462e-01
7.46738195e-01 -2.97202077e-02 -5.48335969e-01 -5.17952025e-01
-1.17767148e-01 -9.03089762e-01 -5.09833276e-01 1.21957469e+00
8.18040073e-01 -5.22256374e-01 -3.99126977e-01 -1.37110341e+00
-3.34772408e-01 4.28819567e-01 -2.86571085e-01 -1.44380495e-01
-2.20541060e-01 2.39922553e-02 4.72453415e-01 7.73920584e-03
6.43256247e-01 4.87178385e-01 9.37754452e-01 9.51049864e-01
-6.17060781e-01 -1.06934714e+00 -9.95264500e-02 -3.39315444e-01
-1.42469442e+00 -5.74809670e-01 5.53760648e-01 -2.95637518e-01
4.60707545e-01 6.07506931e-01 4.81987864e-01 6.98781490e-01
6.84712708e-01 6.93608761e-01 7.40124881e-01 -1.21205756e-02
3.40628117e-01 6.81793869e-01 3.60996753e-01 7.09474564e-01
8.23770046e-01 7.37762228e-02 -2.69760311e-01 -7.97099233e-01
4.50461358e-01 7.24691629e-01 -2.39310458e-01 -5.72260916e-01
-6.53342605e-01 8.03299367e-01 6.03946924e-01 1.81328103e-01
-4.72920746e-01 -5.37436157e-02 9.48582709e-01 4.41900015e-01
4.53808993e-01 -1.43375754e-01 -8.44481111e-01 4.59810883e-01
-8.28661025e-01 8.09842944e-02 9.26563323e-01 1.04930198e+00
7.14446664e-01 -5.32128036e-01 -2.20606655e-01 1.71783753e-02
2.19700336e-01 -3.04998428e-01 8.66900384e-01 -4.31340545e-01
4.84449297e-01 9.38451588e-01 -1.09381275e-02 -8.94071579e-01
-2.75781542e-01 -5.94057679e-01 -1.07092977e+00 1.90095574e-01
-8.76266360e-02 -8.29468846e-01 -5.05321741e-01 1.48784590e+00
6.49213791e-01 6.49556443e-02 3.39880824e-01 5.25585949e-01
7.94226468e-01 1.46296307e-01 2.39084914e-01 -2.44506612e-01
1.52319539e+00 -9.71033335e-01 -9.71117616e-01 2.57708877e-01
1.02301860e+00 -2.17587084e-01 3.16138834e-01 1.71812609e-01
-1.09477651e+00 -3.85919847e-02 -1.13379192e+00 -3.46089095e-01
-7.21728265e-01 1.63370594e-01 6.52668417e-01 7.89788604e-01
-1.33870995e+00 3.83921325e-01 -1.03640544e+00 1.34389117e-01
1.13738263e+00 8.21719646e-01 -7.74859369e-01 1.04885548e-01
-1.22932398e+00 2.12532535e-01 3.18922818e-01 8.44146088e-02
-9.00553107e-01 -6.35809004e-01 -7.25891352e-01 1.08339451e-01
-2.08447993e-01 -8.95482957e-01 1.17523229e+00 -7.72419453e-01
-1.15435791e+00 1.11160314e+00 3.61601442e-01 -6.42090261e-01
8.92515361e-01 -1.25926554e-01 -4.51488256e-01 -9.28575732e-03
-2.37961456e-01 1.32792667e-01 4.44404840e-01 -1.21259785e+00
-7.93008208e-01 -7.78679550e-01 -3.62430036e-01 -2.50376672e-01
-8.04804742e-01 -5.23715019e-02 -4.01618302e-01 -1.86397910e-01
-1.69702694e-01 -7.75012493e-01 -5.55529714e-01 4.83229935e-01
-5.76622367e-01 3.56606930e-01 1.21221507e+00 -9.92977917e-01
1.25264013e+00 -2.41980147e+00 -6.71994627e-01 4.43389475e-01
8.29632640e-01 7.01608419e-01 3.18757355e-01 3.02537262e-01
2.02016905e-02 1.63968951e-01 8.57507661e-02 -6.18789613e-01
-3.17524731e-01 3.54791075e-01 -2.26103980e-02 7.07899332e-01
-3.77988905e-01 9.17944551e-01 -5.03500581e-01 -8.18255067e-01
-3.71531546e-02 6.39077961e-01 -8.10853362e-01 1.92114502e-01
1.98783875e-01 3.30991864e-01 -9.58346665e-01 4.43695724e-01
1.29858208e+00 -1.87284052e-01 5.70624173e-01 2.41051465e-01
-5.64490408e-02 1.57342330e-01 -9.01946545e-01 1.55833280e+00
-6.46001995e-02 5.19783236e-02 8.93413782e-01 -3.96699578e-01
6.65854812e-01 7.67961025e-01 8.16036880e-01 -2.23597318e-01
3.64548564e-01 -2.36000285e-01 -2.95757174e-01 -6.66442871e-01
1.75693527e-01 -1.54603019e-01 -1.91649154e-01 4.49236482e-01
-3.70453268e-01 9.66012895e-01 -1.01365066e+00 -2.01462824e-02
1.08902478e+00 -5.62178314e-01 1.61370441e-01 -2.04250723e-01
4.93070692e-01 -2.66183645e-01 9.28771138e-01 4.33059603e-01
-7.27273941e-01 1.50869414e-01 5.97090185e-01 -9.38739598e-01
-7.13836312e-01 -8.23656678e-01 -4.82613817e-02 8.90323460e-01
-3.27771962e-01 -6.21375740e-01 -9.45175648e-01 -1.08730710e+00
3.07612330e-01 7.20064268e-02 -9.56444561e-01 -3.09917748e-01
-2.97299564e-01 -3.65356654e-01 9.30915356e-01 4.04068500e-01
8.28180552e-01 -6.77051127e-01 -8.24541867e-01 8.69187862e-02
1.31279573e-01 -5.19031584e-01 -6.40135467e-01 1.70080513e-01
-1.01585400e+00 -1.19875252e+00 -2.51895249e-01 -7.56547034e-01
8.15150857e-01 8.77355970e-03 5.78751743e-01 1.45124376e-01
-3.09442222e-01 -1.34554312e-01 9.83848348e-02 -9.47472692e-01
-4.11400765e-01 2.06323415e-01 -2.73592025e-01 1.89539105e-01
5.26894927e-01 -3.99031609e-01 -9.40436244e-01 -2.52900958e-01
-1.15154922e+00 -3.91988128e-01 2.88332909e-01 3.78598183e-01
5.02038300e-01 3.40093046e-01 6.36716709e-02 -1.37995028e+00
6.44693792e-01 -7.25695014e-01 -5.40645599e-01 3.03155690e-01
-5.96986353e-01 -2.21251532e-01 7.21209347e-01 5.16368523e-02
-6.57821298e-01 5.80816269e-01 1.32747395e-02 -4.38475847e-01
2.63852160e-02 1.80512324e-01 -5.88367701e-01 -1.24204151e-01
3.06169420e-01 8.15123096e-02 6.02657259e-01 -8.09691310e-01
-1.07185170e-02 9.49194729e-01 2.93557346e-01 1.68079720e-03
2.77314544e-01 6.23055875e-01 -1.98720708e-01 -1.49586096e-01
-3.97253454e-01 -2.32826456e-01 -4.43448693e-01 3.03134084e-01
9.63551402e-01 -1.31441271e+00 -1.28664684e+00 4.81542021e-01
-1.01558077e+00 3.41237098e-01 -3.61788958e-01 4.41685855e-01
2.48487696e-01 1.43918470e-01 -8.50965440e-01 -6.28663540e-01
-1.03155744e+00 -9.36172843e-01 6.89379990e-01 1.61567897e-01
1.38697252e-01 -1.15367198e+00 4.41186488e-01 2.11152658e-01
4.78779137e-01 6.30926251e-01 6.64134145e-01 -1.13943112e+00
-4.43551183e-01 -9.04338360e-01 2.27916256e-01 6.47367001e-01
4.42716777e-01 -2.78608263e-01 -1.24898350e+00 -8.44092488e-01
7.11607039e-01 -3.06363702e-02 5.32759666e-01 2.09430411e-01
1.59340501e+00 -1.12597156e+00 -7.21752167e-01 1.00535274e+00
1.64279127e+00 9.90447178e-02 7.56686151e-01 2.86977798e-01
5.34376919e-01 5.17337024e-01 -1.12921715e-01 7.10823298e-01
2.75878608e-01 -1.21987030e-01 6.29478514e-01 -5.41611254e-01
5.76190889e-01 -2.70501226e-01 1.47443011e-01 5.43974936e-01
4.20346290e-01 6.52304143e-02 -7.34315813e-01 3.07324320e-01
-1.82931781e+00 -6.46966457e-01 8.61519799e-02 2.39917850e+00
9.54696298e-01 -3.73257190e-01 1.13851465e-01 -3.35080475e-01
6.51847899e-01 1.26731589e-01 -7.14754462e-01 -2.83824712e-01
7.79327601e-02 1.27461433e-01 9.97446239e-01 1.93363249e-01
-1.07649219e+00 5.39612532e-01 7.20902967e+00 6.30365014e-02
-1.28592789e+00 4.25371736e-01 8.24694455e-01 -5.76890290e-01
-1.72169879e-01 -2.18073487e-01 -5.74563742e-01 6.27512991e-01
8.04468095e-01 -1.55118436e-01 -2.77062654e-01 1.30644083e+00
9.61762387e-03 3.74758422e-01 -9.51427937e-01 7.74743259e-01
-5.56482255e-01 -1.51990211e+00 -8.45483616e-02 6.11432552e-01
3.90996397e-01 1.50201917e-01 2.15270743e-01 -1.58305645e-01
6.05660677e-01 -8.90146732e-01 2.55402401e-02 4.58569914e-01
7.33744264e-01 -1.10855865e+00 1.07447302e+00 5.28271914e-01
-7.83306777e-01 -2.56764799e-01 -5.04705787e-01 1.49619013e-01
-5.05527079e-01 6.66084766e-01 -9.13808465e-01 9.11872029e-01
9.21387076e-01 3.73358190e-01 -4.51663554e-01 7.11191118e-01
3.00581068e-01 3.91951442e-01 -1.91714644e-01 4.40474749e-01
2.37478629e-01 -1.00180343e-01 -4.61879708e-02 1.03355944e+00
5.05755059e-02 2.72119455e-02 3.07856470e-01 2.61443645e-01
-4.85423714e-01 3.43342274e-01 -8.52354527e-01 2.53498644e-01
4.15954709e-01 1.28198969e+00 -3.91508564e-02 -1.98471159e-01
-4.20217425e-01 1.01480591e+00 2.54612535e-01 -6.46954998e-02
-6.12737238e-01 -4.97769445e-01 1.34370208e+00 3.34419429e-01
1.00941159e-01 2.37656221e-01 -3.65593940e-01 -9.24642086e-01
1.73306957e-01 -6.87915206e-01 1.11091113e+00 -1.14512525e-01
-1.17005587e+00 6.51798487e-01 -6.19488835e-01 -1.23303044e+00
1.27660125e-01 -3.94641221e-01 -4.19069320e-01 1.00359523e+00
-1.49738622e+00 -1.40891862e+00 9.17341374e-03 1.50923955e+00
-8.14485610e-01 -3.93735349e-01 1.26062107e+00 3.74341220e-01
-9.01180208e-01 1.32284427e+00 4.42104667e-01 7.22809494e-01
6.27326190e-01 -6.25086486e-01 -1.74591213e-01 7.27617443e-01
-5.08311093e-01 1.14865124e+00 1.80134550e-02 -7.92127252e-01
-1.34055698e+00 -1.53127551e+00 9.58137691e-01 -2.15448424e-01
-4.18803133e-02 -6.24325693e-01 -1.03582680e+00 1.20831275e+00
2.42015049e-01 4.83975708e-01 1.51466036e+00 -9.41122547e-02
-5.05047321e-01 -5.79771817e-01 -1.88668013e+00 1.50406852e-01
1.38760939e-01 -7.50824392e-01 -8.88944492e-02 4.78497624e-01
8.20761800e-01 -4.28614318e-01 -1.02584660e+00 -2.17832997e-01
4.54096615e-01 -8.19634557e-01 7.17659831e-01 -1.21048105e+00
-3.21710180e-03 -2.24103555e-01 3.02366465e-01 -6.17700815e-01
-2.02697426e-01 -1.08041835e+00 -2.95804814e-02 1.05672562e+00
2.83727080e-01 -1.07170570e+00 1.56810975e+00 1.35785985e+00
1.30793467e-01 -5.18052578e-01 -1.25474048e+00 -2.69421875e-01
2.66247243e-01 1.89500481e-01 1.11004639e+00 1.38280749e+00
3.61413658e-02 -2.67042369e-01 -4.97412384e-01 5.55312097e-01
4.42826807e-01 -3.31355512e-01 6.84419572e-01 -9.58959520e-01
-2.39164364e-02 2.74976134e-01 -3.95105243e-01 -9.15767625e-02
-1.55106485e-01 -1.05507851e+00 -4.50796962e-01 -1.25265408e+00
3.15492332e-01 -4.95005310e-01 -1.02334952e+00 1.15741944e+00
4.92387041e-02 -3.02377790e-01 3.96157466e-02 3.42044413e-01
-4.88133222e-01 -4.87197610e-03 9.74804580e-01 2.04280749e-01
-3.40849966e-01 2.90627509e-01 -1.13171935e+00 4.11725134e-01
1.00146770e+00 -9.26041484e-01 -5.14418542e-01 -6.65249228e-01
-2.28196874e-01 8.80453289e-02 3.14096987e-01 -9.24596786e-01
7.26137221e-01 1.51682213e-01 7.47049391e-01 -5.21479130e-01
-2.15474516e-01 -1.50978339e+00 6.42083883e-01 1.16486573e+00
-3.41632545e-01 1.09532095e-01 3.70028108e-01 4.39606965e-01
-1.15194902e-01 3.47655386e-01 7.50253618e-01 -2.16345876e-01
1.01704389e-01 9.39042628e-01 -1.31996632e-01 -5.80014884e-01
1.44907773e+00 -1.40271679e-01 -2.76804417e-01 -2.65834741e-02
-8.46943080e-01 4.85067129e-01 7.11994648e-01 1.85480062e-03
5.56788385e-01 -1.05226362e+00 -5.28043091e-01 4.53433692e-01
-1.03761451e-02 2.95398712e-01 6.84435368e-01 4.33450192e-01
-6.47410333e-01 3.14412206e-01 -4.46243286e-01 -1.60155892e-02
-1.54675925e+00 1.14087033e+00 5.61220586e-01 -4.21401769e-01
-7.98767090e-01 8.62413824e-01 3.67094874e-01 -3.94268364e-01
5.42750418e-01 -2.09619105e-01 1.57704204e-01 -8.36784393e-02
9.45740759e-01 1.21270470e-01 2.92586207e-01 -6.57504052e-02
-6.17613733e-01 -1.56550571e-01 -6.72940195e-01 1.62461057e-01
1.48500347e+00 2.92835813e-02 -5.94085455e-01 -1.21147498e-01
1.80255771e+00 3.42725366e-01 -1.16201055e+00 -1.54672667e-01
-3.21647823e-01 -3.99119765e-01 1.87781975e-01 -8.49175751e-01
-1.72832966e+00 6.69827044e-01 8.53017986e-01 -1.77386597e-01
1.33177733e+00 -3.80710781e-01 9.94479001e-01 1.15860201e-01
2.73393065e-01 -9.17787611e-01 -6.55421436e-01 -2.01161802e-01
2.37547353e-01 -8.41631174e-01 2.21668497e-01 -2.36171514e-01
-7.61635602e-01 7.84627318e-01 2.72137612e-01 9.47904140e-02
1.17451167e+00 6.65365636e-01 4.70229119e-01 -3.26492876e-01
-6.97788835e-01 7.19096720e-01 -4.54543501e-01 6.66107833e-01
1.69022143e-01 2.00110450e-01 2.91403085e-02 1.26840675e+00
-8.29390585e-02 5.01413584e-01 1.34195432e-01 1.53665972e+00
4.73021790e-02 -1.40790129e+00 -5.14230505e-02 3.08394760e-01
-9.97592568e-01 1.14838883e-01 -4.47846115e-01 5.95866680e-01
5.82224965e-01 4.04242128e-01 4.93499711e-02 -4.91280437e-01
3.46526474e-01 3.24681818e-01 -3.56424004e-01 -4.20644194e-01
-1.52488148e+00 -1.85440809e-01 -2.04195380e-01 -8.35986912e-01
1.73417613e-01 -6.10627830e-01 -1.24299157e+00 -7.21021175e-01
5.38664274e-02 4.74480420e-01 8.68370831e-01 9.47582200e-02
1.12093341e+00 3.94296736e-01 1.03184891e+00 3.26981157e-01
-5.91515839e-01 -2.07225382e-01 -8.72689068e-01 3.58427502e-02
9.23421621e-01 4.42877680e-01 -5.18684275e-02 6.30954019e-05] | [6.149084568023682, 6.622735500335693] |
2a8dc3e3-95bd-4fc5-bac4-b850c8cee91d | a-physics-informed-machine-learning-model-for | 2101.05605 | null | https://arxiv.org/abs/2101.05605v1 | https://arxiv.org/pdf/2101.05605v1.pdf | A Physics-Informed Machine Learning Model for Porosity Analysis in Laser Powder Bed Fusion Additive Manufacturing | To control part quality, it is critical to analyze pore generation mechanisms, laying theoretical foundation for future porosity control. Current porosity analysis models use machine setting parameters, such as laser angle and part pose. However, these setting-based models are machine dependent, hence they often do not transfer to analysis of porosity for a different machine. To address the first problem, a physics-informed, data-driven model (PIM), which instead of directly using machine setting parameters to predict porosity levels of printed parts, it first interprets machine settings into physical effects, such as laser energy density and laser radiation pressure. Then, these physical, machine independent effects are used to predict porosity levels according to pass, flag, fail categories instead of focusing on quantitative pore size prediction. With six learning methods evaluation, PIM proved to achieve good performances with prediction error of 10$\sim$26%. Finally, pore-encouraging influence and pore-suppressing influence were analyzed for quality analysis. | ['Xiaoli Zhang', 'Sen Liu', 'Rui Liu'] | 2021-01-13 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 1.93837985e-01 4.46011871e-02 -2.95594752e-01 -9.18343663e-03
-4.25232410e-01 -2.59329975e-01 2.87735105e-01 6.06348455e-01
-4.41545155e-03 6.03479028e-01 -1.48882300e-01 -5.82382977e-01
-7.42074192e-01 -1.36949778e+00 -8.53715777e-01 -6.10689223e-01
2.03614607e-01 8.06343436e-01 4.74560142e-01 -1.39657676e-01
7.76793659e-01 7.38802373e-01 -1.62334824e+00 1.54942259e-01
9.17668581e-01 8.80828857e-01 1.56764850e-01 5.13006568e-01
-3.20626318e-01 6.06808722e-01 -2.99017578e-01 -1.22513389e-02
-1.35126501e-01 -4.04946238e-01 -8.41697037e-01 7.11081456e-03
-2.46498898e-01 -1.29020497e-01 -1.31872192e-01 4.53041494e-01
2.40560174e-01 -1.52680382e-01 1.24504125e+00 -7.12348878e-01
-5.16544461e-01 8.76615703e-01 -1.50560334e-01 -8.67039524e-03
3.49018812e-01 1.85344517e-01 8.09655130e-01 -8.09156358e-01
3.82816166e-01 9.31790471e-01 3.19017351e-01 3.21991563e-01
-1.39408886e+00 -3.04444999e-01 -3.65365952e-01 3.85698825e-02
-9.79595244e-01 -9.73177627e-02 7.93365717e-01 -8.46332192e-01
7.56415188e-01 4.02271926e-01 9.57422793e-01 5.39736569e-01
7.91203260e-01 2.41238385e-01 1.13758624e+00 -6.55909538e-01
5.88614047e-01 1.35935694e-01 1.28101304e-01 5.44487119e-01
3.41021955e-01 4.14357901e-01 -3.90242428e-01 1.92556828e-01
1.23317420e+00 -4.29026902e-01 -2.28744492e-01 -5.73752224e-01
-7.08605468e-01 7.12981105e-01 3.17310363e-01 1.42223343e-01
-2.75412500e-01 -5.33712246e-02 -7.48544037e-02 2.92819172e-01
1.15228131e-01 1.07406664e+00 -8.14691961e-01 -2.51964092e-01
-8.61960053e-01 3.44682306e-01 1.08141196e+00 7.43256807e-01
9.10580575e-01 -3.31350975e-02 -2.37068489e-01 8.84582043e-01
6.94043577e-01 5.94082713e-01 5.75110205e-02 -7.35873938e-01
2.05420330e-01 6.85091436e-01 1.49174020e-01 -7.30118215e-01
-5.49007714e-01 1.28864050e-01 -3.36412996e-01 3.03952605e-01
4.16279167e-01 -7.08620101e-02 -1.32544911e+00 8.52437079e-01
1.98550135e-01 -6.37693405e-01 -1.10100105e-01 9.12149429e-01
8.12642395e-01 3.40205222e-01 2.77996451e-01 -1.10459492e-01
1.15925741e+00 -6.49328172e-01 -4.57592696e-01 2.50680409e-02
5.97396433e-01 -1.02540779e+00 1.13830435e+00 6.50402427e-01
-1.20558119e+00 -3.71390671e-01 -1.23485410e+00 4.19924766e-01
-4.80194867e-01 1.72161475e-01 8.57369006e-01 6.63706362e-01
-4.74868059e-01 1.23279524e+00 -8.54334593e-01 -2.18998879e-01
2.78434128e-01 5.44716358e-01 1.01793841e-01 1.48133427e-01
-1.03360748e+00 1.01502776e+00 4.31555688e-01 -4.78631258e-02
-6.52824283e-01 -1.07284963e+00 -4.79560316e-01 5.02458215e-02
2.02280372e-01 -6.45091534e-01 1.13346457e+00 -3.94090302e-02
-2.14927101e+00 4.29930806e-01 2.90728599e-01 -1.72976658e-01
4.10738915e-01 -4.76500243e-01 -4.09757942e-01 1.35647252e-01
-2.47037485e-01 2.50130743e-01 6.58270597e-01 -1.69968414e+00
-1.35844693e-01 -6.19874038e-02 -6.69046715e-02 -1.27660513e-01
-7.93688148e-02 -4.76918072e-01 -3.52249771e-01 -7.78360590e-02
6.33080244e-01 -6.49082124e-01 -3.56882185e-01 -1.58014223e-02
-5.07566094e-01 -1.08426496e-01 6.72327518e-01 -5.87705314e-01
1.16737103e+00 -1.85758138e+00 4.55412362e-03 6.44739926e-01
-9.30842757e-03 -2.47566327e-02 2.96403885e-01 6.68233991e-01
-3.88459070e-03 3.15824002e-01 -1.64075986e-01 3.28161567e-01
-2.33456239e-01 1.31015137e-01 1.15210682e-01 2.77061820e-01
4.69795018e-01 7.68517613e-01 -5.75974226e-01 -6.68674648e-01
9.55323219e-01 2.32516676e-01 -7.66652405e-01 -9.58062336e-03
-3.76536906e-01 2.60356009e-01 -8.01129639e-01 9.18257833e-01
7.93323994e-01 -1.39193848e-01 -1.17358848e-01 -5.22170782e-01
-1.42151237e-01 -1.43099017e-02 -1.02995968e+00 1.18712866e+00
-6.90047026e-01 -1.81283459e-01 1.08911209e-01 -6.30756021e-01
1.39598477e+00 5.96820526e-02 8.09803426e-01 -8.57246518e-01
4.47257638e-01 3.88185799e-01 5.78915961e-02 -7.37402201e-01
4.65927660e-01 -3.13148290e-01 1.36079103e-01 -6.78433776e-02
4.60768193e-02 -1.00838041e+00 1.85415611e-01 -1.23981699e-01
1.09576106e+00 2.00969368e-01 -3.08251411e-01 -4.29705650e-01
4.06854004e-01 2.11352035e-01 1.90947354e-01 4.79924589e-01
2.13921323e-01 8.80296290e-01 3.81686330e-01 -6.44409209e-02
-1.29477215e+00 -1.27884126e+00 -5.64181805e-01 5.03219664e-01
6.90472126e-01 -2.59362221e-01 -6.21646047e-01 -1.32364258e-01
3.93233865e-01 8.15373957e-01 -3.93180251e-01 -5.51023424e-01
-4.01538789e-01 -6.78885877e-01 -7.44362324e-02 4.73010123e-01
2.29845643e-01 -9.98516679e-01 -4.33738828e-01 5.43256998e-01
2.65037566e-01 -7.48903036e-01 3.64501566e-01 5.39300382e-01
-1.35837340e+00 -1.44921339e+00 -3.35048348e-01 -6.96964636e-02
3.87119353e-01 -3.14994395e-01 1.29726028e+00 2.53279150e-01
-3.89753282e-01 4.79450315e-01 -6.63908720e-01 -5.06205142e-01
-7.07988083e-01 1.41903475e-01 -1.26064032e-01 -5.91580153e-01
-2.96885166e-02 -6.37687087e-01 -6.73590422e-01 6.70686543e-01
-9.18224335e-01 -9.38313827e-03 1.02410972e+00 5.66889286e-01
9.24298465e-01 4.29483503e-01 9.76530984e-02 -8.15312445e-01
4.93961364e-01 -2.06063882e-01 -2.78690636e-01 2.14092389e-01
-9.51112807e-01 1.90912902e-01 3.82169396e-01 -2.46098578e-01
-8.24907482e-01 -1.69525191e-01 -2.54931986e-01 -3.19185227e-01
-1.86454698e-01 5.76360166e-01 -4.25242722e-01 3.05927079e-02
6.40062213e-01 -1.03063807e-01 6.48793727e-02 -5.99030733e-01
-6.05026260e-02 4.07864422e-01 2.74939895e-01 -1.09678185e+00
8.59756410e-01 1.24126539e-01 4.87542897e-02 -9.68429029e-01
-5.44553757e-01 -2.76673973e-01 -6.25164509e-01 -4.14379179e-01
6.81156337e-01 -4.41356003e-01 -7.15030611e-01 3.20630282e-01
-6.27969861e-01 -5.30511618e-01 -5.02438188e-01 6.71205282e-01
-7.01928258e-01 4.94088344e-02 -5.62863708e-01 -9.14165616e-01
-3.02371323e-01 -1.15113747e+00 1.05933142e+00 4.67795342e-01
-2.51289815e-01 -8.54676485e-01 -1.65245220e-01 5.74635923e-01
4.13467765e-01 1.45533592e-01 1.48209918e+00 -2.06156507e-01
-4.94908422e-01 -1.54021457e-01 -1.52144553e-02 1.45890892e-01
2.02665284e-01 5.28825879e-01 -5.88135183e-01 -2.20105052e-02
-2.23306164e-01 -2.10300699e-01 6.15405142e-01 5.10852814e-01
1.39429593e+00 1.63152456e-01 -5.79312623e-01 2.47645840e-01
1.69561636e+00 4.58874367e-02 9.58123326e-01 5.75300634e-01
4.01570678e-01 4.22343403e-01 8.36942017e-01 4.05232996e-01
-4.38303232e-01 5.52274108e-01 7.05398262e-01 -8.94118398e-02
-1.75016120e-01 -5.60546815e-01 -4.54582609e-02 4.80425626e-01
-4.57577050e-01 -3.05881590e-01 -1.00104344e+00 1.15866028e-01
-1.17861950e+00 -5.55254936e-01 -5.38263321e-01 2.27292013e+00
7.03717053e-01 4.73339856e-01 -1.13219522e-01 6.29402578e-01
6.25901282e-01 -4.80188757e-01 -3.55070859e-01 -4.29399848e-01
1.79625973e-01 5.19109964e-01 8.19373965e-01 2.23513693e-01
-7.66369104e-01 7.46843994e-01 7.11787844e+00 6.85632408e-01
-1.09165895e+00 -4.56666946e-01 3.39469284e-01 2.32940204e-02
-5.29833913e-01 2.01177329e-01 -5.96169114e-01 3.79456669e-01
8.89806688e-01 2.43230745e-01 2.90887475e-01 7.47691512e-01
2.99117446e-01 -6.40822947e-01 -1.04633927e+00 5.79079628e-01
-5.28302729e-01 -1.22776556e+00 1.13840155e-01 2.64892042e-01
3.78437310e-01 -2.40921035e-01 -1.94797814e-01 2.59213895e-01
2.32933283e-01 -9.78253186e-01 9.26811159e-01 8.39771271e-01
6.29159212e-01 -5.89242458e-01 7.68246830e-01 -6.95459098e-02
-8.95985067e-01 -1.45931348e-01 -3.82411182e-01 6.29502162e-02
3.66697967e-01 1.16984367e+00 -6.99790001e-01 8.64625752e-01
7.87438810e-01 1.73627004e-01 -4.55932707e-01 1.13377464e+00
-1.01890318e-01 6.82161331e-01 -6.28373742e-01 -4.52967733e-02
-2.71738291e-01 -2.42442876e-01 2.13936478e-01 8.60675693e-01
4.31805938e-01 1.96058437e-01 -3.23853604e-02 1.21622813e+00
3.74780118e-01 2.21816316e-01 -2.44089827e-01 -2.95356721e-01
5.95103264e-01 9.81568635e-01 -1.05696869e+00 2.91056156e-01
3.06966659e-02 4.30581659e-01 -1.22056544e-01 2.09821656e-01
-9.60310817e-01 -1.56145558e-01 2.29983091e-01 7.64790535e-01
3.18866074e-01 -3.14829111e-01 -6.74619377e-01 -3.53987515e-01
-1.41886741e-01 -2.44851530e-01 -2.58528054e-01 -6.32991612e-01
-1.27639139e+00 -3.61812115e-01 1.18675448e-01 -1.09821844e+00
4.35529977e-01 -1.08218670e+00 -8.00029039e-01 6.27917707e-01
-1.28370690e+00 -8.05122435e-01 -1.22076474e-01 1.60154048e-02
3.29562843e-01 1.56849533e-01 5.90549231e-01 1.40457660e-01
-5.85624754e-01 3.47731590e-01 1.35299385e-01 -1.79500744e-01
4.88440752e-01 -1.19251537e+00 -3.60610634e-01 1.16041429e-01
-2.66916335e-01 2.52183676e-01 1.11964846e+00 -7.56482899e-01
-1.89699173e+00 -6.85129821e-01 4.14556950e-01 -4.02422458e-01
5.94982445e-01 -2.33496204e-02 -9.99030173e-01 -2.60239601e-01
-4.64759618e-01 -2.31401145e-01 4.73837435e-01 3.55433136e-01
2.16655895e-01 9.41409692e-02 -1.14749551e+00 2.20084503e-01
8.14080596e-01 -1.88562050e-01 -3.91805172e-01 1.94924444e-01
3.71713251e-01 -7.69509971e-01 -1.64966583e+00 8.33378732e-01
5.73903799e-01 -8.98752987e-01 9.70136821e-01 -3.26507211e-01
1.10429049e+00 8.53271130e-03 -2.53016446e-02 -9.96751308e-01
-4.19439942e-01 1.16005845e-01 -1.29505172e-01 1.36145401e+00
9.49895978e-01 -3.39335620e-01 1.05975795e+00 9.70711291e-01
-4.99133140e-01 -1.19392717e+00 -6.66894436e-01 -7.00751543e-01
4.09647226e-01 -2.75144815e-01 6.31956935e-01 4.00373220e-01
2.19383761e-02 9.45953950e-02 2.47478545e-01 2.87996054e-01
7.10323304e-02 -4.97079864e-02 5.84117532e-01 -1.56718636e+00
-3.32045645e-01 -5.01657307e-01 -2.82526523e-01 -5.91062486e-01
-3.55781823e-01 -4.00762230e-01 3.06802213e-01 -1.75589824e+00
4.93720872e-03 -9.59367871e-01 -3.69615436e-01 1.88162580e-01
-1.07998345e-02 -2.46393561e-01 -1.61841407e-01 3.31584454e-01
8.65565389e-02 7.49144673e-01 1.64465475e+00 -4.87067401e-01
-5.85132360e-01 7.35179782e-02 -2.84399271e-01 4.08057898e-01
8.55929017e-01 -1.73851803e-01 -3.74200314e-01 -1.68446943e-01
2.64877528e-01 2.92310327e-01 4.83412713e-01 -1.53493881e+00
-2.41668597e-01 -4.49760020e-01 6.41637504e-01 -6.13183737e-01
1.58206046e-01 -9.77723420e-01 2.18483523e-01 5.81987977e-01
2.34733820e-02 -4.89135832e-01 4.19969738e-01 5.19018471e-01
-1.06423639e-01 -4.75715250e-01 6.85666919e-01 9.64049697e-02
-3.26271296e-01 2.13456690e-01 -4.41248417e-01 -3.93218905e-01
9.11281407e-01 -7.34664381e-01 -1.98919028e-02 3.16281766e-01
-9.93948936e-01 1.54350013e-01 6.72784805e-01 2.10195899e-01
4.47728395e-01 -9.97598946e-01 -3.67128253e-01 1.28252283e-01
1.06854491e-01 1.76407441e-01 6.44342363e-01 5.82208514e-01
-7.86111295e-01 3.68011683e-01 3.75510417e-02 -8.22107613e-01
-8.93987954e-01 3.78687620e-01 3.94341558e-01 -1.71094447e-01
-4.87652034e-01 8.87707233e-01 -4.17221785e-01 -4.26305383e-01
-3.78153443e-01 -4.25747603e-01 -2.85494953e-01 -3.09592932e-01
-2.12019578e-01 4.16489065e-01 5.13011754e-01 1.60900634e-02
-3.48354541e-02 8.00903976e-01 1.83757767e-01 8.96971002e-02
1.38266003e+00 8.40905532e-02 2.19715446e-01 5.24921179e-01
6.26689017e-01 -1.67332113e-01 -1.32716393e+00 2.35335439e-01
-6.26615882e-02 -1.82837188e-01 2.32230380e-01 -9.04033244e-01
-1.07374740e+00 5.77615976e-01 6.25725389e-01 2.09662333e-01
8.59325945e-01 2.12821975e-01 4.18316573e-01 2.28618756e-01
4.38500702e-01 -1.56291962e+00 1.19617425e-01 3.72663677e-01
8.66638601e-01 -1.00400913e+00 3.67725223e-01 -9.19471562e-01
-1.33388713e-01 1.23178446e+00 8.64873171e-01 1.62979797e-01
1.01440704e+00 4.25420761e-01 -1.81962997e-01 -5.30567527e-01
-3.38848412e-01 1.36306837e-01 3.19420956e-02 4.38152760e-01
6.68961108e-01 1.97671428e-01 -5.02801955e-01 4.90530193e-01
-5.58933496e-01 2.97128499e-01 2.27486163e-01 1.22747958e+00
-7.93074071e-01 -1.34083641e+00 -6.06432557e-01 1.03048539e+00
6.31429702e-02 2.92159528e-01 -2.39444792e-01 1.17921245e+00
1.29427850e-01 8.63423288e-01 -3.00769936e-02 -6.05819821e-01
8.49589646e-01 1.00730367e-01 8.00524116e-01 -7.35206068e-01
-1.88116655e-01 -1.63550124e-01 7.59706870e-02 -4.07723576e-01
-2.90098101e-01 -6.02939785e-01 -1.35731387e+00 -3.14606130e-01
-8.37941349e-01 2.54657477e-01 7.69177258e-01 1.07326066e+00
-8.01781192e-02 9.36741114e-01 6.19013071e-01 -7.72158682e-01
-3.73371392e-01 -1.04176271e+00 -1.04293609e+00 3.14871222e-01
-4.48794007e-01 -1.21281457e+00 -4.28441018e-01 -6.93309456e-02] | [6.329172611236572, 3.2744860649108887] |
2687c7f0-ab79-4926-bc3a-9450648acad5 | zero-shot-aspect-based-sentiment-analysis | 2202.01924 | null | https://arxiv.org/abs/2202.01924v3 | https://arxiv.org/pdf/2202.01924v3.pdf | Zero-Shot Aspect-Based Sentiment Analysis | Aspect-based sentiment analysis (ABSA) typically requires in-domain annotated data for supervised training/fine-tuning. It is a big challenge to scale ABSA to a large number of new domains. This paper aims to train a unified model that can perform zero-shot ABSA without using any annotated data for a new domain. We propose a method called contrastive post-training on review Natural Language Inference (CORN). Later ABSA tasks can be cast into NLI for zero-shot transfer. We evaluate CORN on ABSA tasks, ranging from aspect extraction (AE), aspect sentiment classification (ASC), to end-to-end aspect-based sentiment analysis (E2E ABSA), which show ABSA can be conducted without any human annotated ABSA data. | ['Hu Xu', 'Bing Liu', 'Jiahua Chen', 'Lei Shu'] | 2022-02-04 | null | null | null | null | ['aspect-extraction', 'aspect-based-sentiment-analysis'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.37691939e-01 3.25418741e-01 3.00786807e-03 -8.43261898e-01
-1.20422971e+00 -6.97459996e-01 8.92952502e-01 1.13516860e-01
-3.64669949e-01 5.49395382e-01 -4.71365376e-04 -5.83132327e-01
2.87727654e-01 -1.04947889e+00 -5.73391378e-01 -1.25119403e-01
4.75471437e-01 8.33363056e-01 2.53024744e-03 -5.70799470e-01
4.84437160e-02 -5.13623096e-02 -1.11206114e+00 5.00630081e-01
7.44874477e-01 1.02760601e+00 -2.07640588e-01 7.71496892e-01
-5.77507138e-01 3.94108325e-01 -7.82272398e-01 -1.14225841e+00
1.44196749e-01 -4.53586549e-01 -9.62240636e-01 -1.02263866e-02
7.04742000e-02 -3.81513424e-02 6.69369340e-01 7.15146244e-01
4.35012907e-01 1.65224493e-01 8.07617188e-01 -1.37774312e+00
-8.87499154e-01 6.00596070e-01 -5.76088488e-01 -3.04898560e-01
2.18036622e-01 2.23710299e-01 1.32579505e+00 -1.19229841e+00
6.42572463e-01 1.26001608e+00 7.80807555e-01 6.11090899e-01
-8.18592548e-01 -5.76999366e-01 3.35076004e-01 -2.74438381e-01
-6.78073883e-01 -1.78713545e-01 8.90661299e-01 -3.29312056e-01
1.43323517e+00 9.01454464e-02 8.87378931e-01 9.64949727e-01
-5.55876754e-02 1.23035705e+00 1.20019114e+00 -4.95020419e-01
5.70048928e-01 3.42001021e-01 6.32538378e-01 4.80984151e-01
3.52512956e-01 -3.06855112e-01 -8.76250923e-01 -3.83736454e-02
7.45468140e-02 -2.44907096e-01 3.22614431e-01 -1.11769825e-01
-1.02677989e+00 8.93887639e-01 4.86351810e-02 -1.27109617e-01
-2.67986119e-01 -2.64598727e-01 7.80796885e-01 8.84985805e-01
9.27002728e-01 1.06162262e+00 -1.30917430e+00 -2.84750849e-01
-7.79465973e-01 1.80546552e-01 1.12414527e+00 1.27816141e+00
1.10224378e+00 1.52465612e-01 -2.27594748e-01 9.11125243e-01
2.05215305e-01 8.82926583e-01 5.78165770e-01 -4.11553890e-01
3.32904071e-01 9.84037519e-01 -6.52469844e-02 -3.41801792e-01
-2.17401892e-01 -6.48777634e-02 -5.29815376e-01 5.01428545e-01
1.38059363e-01 -4.66108561e-01 -1.10988557e+00 1.32113159e+00
4.32825685e-01 -3.09158027e-01 3.87077272e-01 5.01597345e-01
8.63346457e-01 7.10959673e-01 7.66863227e-02 9.02330503e-02
1.83017254e+00 -1.38692760e+00 -6.54085219e-01 -6.08914316e-01
6.61984622e-01 -7.19720423e-01 1.78383815e+00 4.49225664e-01
-8.61619830e-01 -3.84273201e-01 -1.16764855e+00 -4.15991277e-01
-1.09748769e+00 1.00150136e-02 8.93941581e-01 5.85473955e-01
-8.93809080e-01 1.09016053e-01 -7.49755025e-01 -3.57924551e-01
7.65313923e-01 1.77531511e-01 -3.77275467e-01 -9.95316133e-02
-1.27158248e+00 6.39344394e-01 8.99657756e-02 -1.19711183e-01
-8.67254436e-01 -1.10185635e+00 -1.30952907e+00 -6.59245178e-02
6.47002935e-01 -1.02469540e+00 1.64373398e+00 -1.44544315e+00
-1.86229146e+00 1.12945879e+00 -3.14091951e-01 -3.46457511e-01
8.44948962e-02 -6.42979503e-01 -2.11093128e-01 -2.19071999e-01
3.79015774e-01 5.50572217e-01 1.11496782e+00 -1.00812507e+00
-5.16641915e-01 -4.54518467e-01 4.94567215e-01 1.91033736e-01
-4.43839937e-01 3.45513582e-01 -2.58447260e-01 -7.99227715e-01
-7.19491780e-01 -8.31002533e-01 -3.17378640e-01 2.01913547e-02
-3.03243071e-01 -5.38775682e-01 8.21453512e-01 -3.89919966e-01
8.38954866e-01 -1.98348296e+00 -1.16492137e-01 -7.79928565e-02
1.59132550e-03 5.89489937e-01 -6.79511011e-01 1.89609870e-01
-1.05678514e-02 -4.95611764e-02 -7.25717962e-01 -5.70953548e-01
1.46756217e-01 1.49828782e-02 -4.94488060e-01 -3.05242509e-01
7.92010009e-01 1.34484613e+00 -1.06867051e+00 -2.27028787e-01
-1.34739682e-01 2.37786338e-01 -6.86583459e-01 5.73361933e-01
-7.06414640e-01 4.39019129e-02 -4.91120994e-01 9.50969040e-01
6.35982335e-01 -2.47498795e-01 -2.24912092e-01 -1.31041273e-01
1.55360252e-01 5.36339343e-01 -6.67064309e-01 1.95130515e+00
-1.26307106e+00 6.72124565e-01 -1.70980841e-01 -7.71915019e-01
1.14525151e+00 2.49474570e-01 1.26112863e-01 -5.43277502e-01
3.19233358e-01 7.39205405e-02 -1.87058493e-01 -2.25681216e-01
6.65814400e-01 -5.15841246e-01 -5.65083027e-01 8.71385276e-01
6.25526667e-01 -8.22025180e-01 3.91787976e-01 1.76680297e-01
9.87439454e-01 1.92114472e-01 7.18642592e-01 -3.47624302e-01
5.74005723e-01 4.38015163e-01 2.80848205e-01 4.86776561e-01
-6.33948520e-02 6.49691641e-01 7.33490765e-01 -6.31975234e-01
-1.01073623e+00 -9.20605779e-01 7.41893277e-02 1.48155236e+00
-3.57437044e-01 -7.02042818e-01 -7.06289887e-01 -1.21395063e+00
-2.10888550e-01 9.45099473e-01 -8.05804968e-01 -1.23014063e-01
-1.02529399e-01 -6.91256762e-01 2.94360012e-01 6.47979081e-01
6.25132859e-01 -1.41105580e+00 -3.30081254e-01 6.55353367e-02
2.12658897e-01 -1.36092460e+00 -4.92615700e-01 2.84938514e-01
-4.87986445e-01 -8.75700593e-01 -7.19446063e-01 -7.12420344e-01
6.33539319e-01 1.96434066e-01 1.66050470e+00 -4.73849177e-01
-5.65629229e-02 4.50487405e-01 -6.50726795e-01 -1.11961544e+00
-3.58445972e-01 6.22325301e-01 -2.44746119e-01 -6.86141849e-02
9.54435527e-01 -5.97483575e-01 -4.46307451e-01 8.01533088e-02
-9.98886704e-01 9.51262638e-02 8.39708984e-01 6.97546065e-01
7.18045652e-01 -4.46994662e-01 8.78789127e-01 -1.69989657e+00
9.16715145e-01 -3.69735807e-01 -6.15597904e-01 3.65217060e-01
-8.09177756e-01 -7.29299858e-02 9.78797972e-01 -1.99045077e-01
-1.27254081e+00 -9.45610777e-02 -3.12560797e-01 -5.51425405e-02
-1.81875661e-01 8.36809039e-01 -4.34310347e-01 4.61743951e-01
7.05157518e-01 4.91454266e-02 -5.90575188e-02 -2.64324069e-01
7.16438055e-01 8.55757356e-01 1.92695931e-01 -4.80948895e-01
7.74346888e-01 3.74692500e-01 -2.03340098e-01 -8.34496617e-01
-1.75974321e+00 -5.42294621e-01 -7.87468314e-01 1.24434441e-01
9.33721721e-01 -1.15350318e+00 -7.64716864e-02 6.22641802e-01
-1.09389150e+00 -5.14327049e-01 -7.14944661e-01 2.82262191e-02
-4.85577166e-01 -5.09918705e-02 -3.20952356e-01 -5.09824216e-01
-1.14104140e+00 -9.02267098e-01 1.28161502e+00 2.85309345e-01
-5.73091030e-01 -1.10311413e+00 4.50486660e-01 6.27369404e-01
4.95419532e-01 -1.63730100e-01 7.46436954e-01 -1.05142570e+00
-3.36525559e-01 -3.89633358e-01 -1.06411904e-01 7.64760137e-01
2.78030425e-01 -1.20482072e-01 -1.32572699e+00 -7.13176131e-02
-4.06122729e-02 -7.19551265e-01 5.48461735e-01 1.27726346e-01
8.08272839e-01 -3.20322901e-01 2.93337047e-01 5.67120850e-01
1.11621904e+00 -1.88343719e-01 3.80932361e-01 5.59171855e-01
7.24301159e-01 4.65825617e-01 1.23940670e+00 3.21491241e-01
4.35294777e-01 1.69677064e-01 -8.68096054e-02 -8.91256183e-02
-1.96781963e-01 -8.16862956e-02 7.50374258e-01 1.26833606e+00
3.29049379e-02 -3.98061872e-02 -8.98552537e-01 1.01280093e+00
-1.50227845e+00 -3.04174542e-01 1.06001303e-01 1.85523593e+00
1.20099640e+00 2.63776898e-01 2.37377062e-02 -3.49861711e-01
-1.01945465e-02 2.97991782e-01 -7.70284116e-01 -9.65685546e-01
-2.72024795e-02 6.98102057e-01 -1.90023571e-01 5.69756806e-01
-1.29377258e+00 1.35040355e+00 5.40601349e+00 7.42316365e-01
-8.83751512e-01 3.34303677e-01 5.45256793e-01 3.13630030e-02
-7.79440105e-01 2.31721729e-01 -6.66396081e-01 -4.83384728e-02
1.07515025e+00 -4.57053423e-01 3.63005809e-02 1.45585251e+00
-3.96793634e-01 4.46229130e-02 -1.12823093e+00 5.76749444e-01
2.81911135e-01 -1.06772590e+00 6.76938951e-01 -6.26203179e-01
1.02833855e+00 1.34998336e-01 -2.61077769e-02 1.04440153e+00
6.06233895e-01 -7.64064133e-01 5.71582615e-02 9.37933773e-02
9.46119487e-01 -7.00323820e-01 9.42567050e-01 -7.82333091e-02
-1.14767492e+00 5.02875745e-01 -3.96316946e-01 -9.69114974e-02
1.16491966e-01 8.09931219e-01 -7.79028594e-01 6.14757419e-01
5.86187601e-01 1.10341382e+00 -7.04306662e-01 4.52083319e-01
-7.34492302e-01 7.23840654e-01 2.26103440e-02 -3.71501416e-01
3.15582842e-01 -3.60227555e-01 6.50826395e-01 1.23418415e+00
1.04533829e-01 -1.65129781e-01 -1.57794148e-01 6.40532851e-01
-2.74433881e-01 1.93797514e-01 -6.69820905e-01 -3.18371505e-01
-1.32199287e-01 1.63906050e+00 -4.67602521e-01 -7.04679310e-01
-9.02345240e-01 1.18238497e+00 3.10598582e-01 2.99956352e-01
-4.55134660e-01 -8.05618465e-01 8.52824330e-01 -3.97845544e-02
4.02458042e-01 1.75191760e-01 -4.80810702e-01 -1.50645792e+00
1.06924810e-02 -1.22975492e+00 4.50941145e-01 -9.48673129e-01
-1.79039693e+00 9.33094501e-01 -1.59798145e-01 -1.29970133e+00
-3.18942547e-01 -9.49981034e-01 -7.78435588e-01 8.51767778e-01
-1.95388913e+00 -1.66149640e+00 -1.77130297e-01 4.72113460e-01
9.77347434e-01 -5.07218182e-01 1.08413124e+00 2.16570884e-01
-2.14860290e-01 7.24307358e-01 -4.60414529e-01 2.06837475e-01
9.76729095e-01 -1.66758168e+00 9.11272883e-01 7.26652503e-01
-9.42221284e-03 5.04701257e-01 5.58019102e-01 -4.93320435e-01
-1.25230479e+00 -1.46517229e+00 1.12052619e+00 -8.21616232e-01
1.12559807e+00 -6.01318896e-01 -9.76061225e-01 9.48601961e-01
5.90534985e-01 5.06962799e-02 1.05283451e+00 7.60074139e-01
-6.99041784e-01 -3.76808733e-01 -7.94504225e-01 5.95826864e-01
5.26265502e-01 -8.20118427e-01 -8.46957743e-01 1.92109585e-01
1.22859049e+00 1.01651303e-01 -7.06538141e-01 4.43280995e-01
3.53239805e-01 -3.59272510e-01 5.62339604e-01 -9.77285087e-01
1.04718041e+00 -2.78808892e-01 1.89987555e-01 -1.67344320e+00
3.64184827e-01 -5.80096364e-01 2.10156236e-02 1.42014718e+00
1.02883160e+00 -5.93535423e-01 3.89656961e-01 7.17025876e-01
-1.87021405e-01 -8.24023426e-01 -5.83142400e-01 -6.13572419e-01
3.28595757e-01 -5.84279954e-01 8.04125130e-01 1.08468378e+00
8.52017552e-02 1.30079401e+00 -4.06369269e-02 -8.68190303e-02
2.99318194e-01 6.24443650e-01 1.18126261e+00 -1.01585734e+00
-4.23518360e-01 -4.02762473e-01 -1.64826572e-01 -6.82716429e-01
2.14293540e-01 -9.70839620e-01 1.66892231e-01 -1.73543704e+00
3.02955389e-01 -1.41471460e-01 -3.61172140e-01 7.61474311e-01
-6.27114356e-01 1.27444163e-01 -6.15762211e-02 -3.54062110e-01
-9.83594596e-01 9.96212959e-01 1.43218768e+00 -3.82767558e-01
-1.83529913e-01 3.14125955e-01 -9.24143374e-01 9.73801315e-01
7.84046113e-01 -4.90486562e-01 -6.09889746e-01 -4.03610408e-01
8.50205302e-01 -3.34119350e-01 -2.56813556e-01 -5.82722187e-01
-1.77998945e-01 8.10444430e-02 -1.07817821e-01 -5.58889508e-01
1.77214473e-01 -6.82853520e-01 -9.27225232e-01 -3.29479389e-02
-2.98233092e-01 1.31625906e-01 5.42595744e-01 3.81327033e-01
-6.93595648e-01 -3.22122782e-01 4.64690834e-01 -1.15949064e-01
-7.14521587e-01 3.22737753e-01 -3.62447172e-01 5.18904805e-01
8.09808493e-01 3.81321013e-01 -2.18610793e-01 -3.66700321e-01
-5.93236148e-01 2.03531489e-01 2.06451342e-01 3.30476761e-01
4.57403719e-01 -9.61750269e-01 -6.69207931e-01 1.49253234e-01
6.96237803e-01 4.54076588e-01 2.26072058e-01 3.34116399e-01
-2.43874192e-01 1.07443720e-01 -5.29614277e-02 -1.59014046e-01
-1.17152381e+00 5.75617909e-01 -6.31979993e-03 -9.16158617e-01
-3.70020151e-01 8.81030440e-01 2.84530163e-01 -1.27599716e+00
-3.13996166e-01 -3.18584144e-01 -2.13022619e-01 3.00366521e-01
5.92466772e-01 -2.31783330e-01 1.55915558e-01 4.62990887e-02
-1.14928737e-01 5.46027601e-01 -4.50227171e-01 -2.74947554e-01
1.48332846e+00 -1.88610926e-02 -2.73738027e-01 8.09077084e-01
1.13429511e+00 2.91611217e-02 -9.57764447e-01 -2.87208587e-01
-9.48464498e-02 -1.05896331e-01 -1.15841515e-01 -1.29205060e+00
-9.45757866e-01 1.20964181e+00 1.08901067e-02 7.67314434e-02
1.22949958e+00 1.03750303e-01 1.00370026e+00 9.61419046e-01
1.11470357e-01 -1.06589317e+00 1.78141475e-01 9.16796565e-01
9.46339130e-01 -1.68968272e+00 3.96088734e-02 -2.91018814e-01
-1.27699995e+00 7.50654697e-01 7.71750450e-01 -2.57624835e-01
1.03720760e+00 5.04342020e-01 5.77749074e-01 -3.20754141e-01
-1.02505636e+00 -3.74892652e-01 3.84614080e-01 6.56379461e-01
3.20842713e-01 8.70909467e-02 -9.17509422e-02 1.29233134e+00
-5.67770660e-01 6.85734153e-02 4.52149093e-01 9.18124676e-01
1.39162382e-02 -1.18356323e+00 3.17057878e-01 5.11531711e-01
-4.41646934e-01 -5.66341162e-01 -4.96213585e-01 7.34707892e-01
-4.70243037e-01 6.70558155e-01 -9.62850302e-02 4.44614477e-02
4.95556176e-01 4.18076605e-01 6.86635748e-02 -1.04091966e+00
-8.16659510e-01 -1.37375817e-01 4.02697444e-01 -3.85560453e-01
-5.94116509e-01 -5.07103086e-01 -1.01198804e+00 1.90826789e-01
-6.31946102e-02 1.74530119e-01 6.60893917e-01 1.20777822e+00
5.25636554e-01 8.19937646e-01 4.98780638e-01 -1.62669212e-01
-1.26552328e-01 -1.22384179e+00 -3.89479488e-01 3.91892135e-01
3.40143859e-01 -1.03762314e-01 -2.30397046e-01 2.20265970e-01] | [11.460530281066895, 6.688036918640137] |
d85b682f-cdbc-4eea-8559-f954efc0273f | supervised-contrastive-learning-to-classify | 2209.01937 | null | https://arxiv.org/abs/2209.01937v1 | https://arxiv.org/pdf/2209.01937v1.pdf | Supervised Contrastive Learning to Classify Paranasal Anomalies in the Maxillary Sinus | Using deep learning techniques, anomalies in the paranasal sinus system can be detected automatically in MRI images and can be further analyzed and classified based on their volume, shape and other parameters like local contrast. However due to limited training data, traditional supervised learning methods often fail to generalize. Existing deep learning methods in paranasal anomaly classification have been used to diagnose at most one anomaly. In our work, we consider three anomalies. Specifically, we employ a 3D CNN to separate maxillary sinus volumes without anomalies from maxillary sinus volumes with anomalies. To learn robust representations from a small labelled dataset, we propose a novel learning paradigm that combines contrastive loss and cross-entropy loss. Particularly, we use a supervised contrastive loss that encourages embeddings of maxillary sinus volumes with and without anomaly to form two distinct clusters while the cross-entropy loss encourages the 3D CNN to maintain its discriminative ability. We report that optimising with both losses is advantageous over optimising with only one loss. We also find that our training strategy leads to label efficiency. With our method, a 3D CNN classifier achieves an AUROC of 0.85 while a 3D CNN classifier optimised with cross-entropy loss achieves an AUROC of 0.66. | ['Anna Sophie Hoffmann', 'Alexander Schlaefer', 'Christian Betz', 'Bastian Cheng', 'Marvin Petersen', 'Florian Jansen', 'Elina Petersen', 'Dennis Eggert', 'Dirk Beyersdorff', 'Marcel Bengs', 'Finn Behrendt', 'Benjamin Tobias Becker', 'Debayan Bhattacharya'] | 2022-09-05 | null | null | null | null | ['anomaly-classification'] | ['computer-vision'] | [-8.11559111e-02 3.67842197e-01 -5.51377423e-02 -4.83884722e-01
-5.91924310e-01 -2.76436836e-01 4.56182361e-01 7.30114877e-01
-6.27259552e-01 4.08693314e-01 3.11800325e-03 -2.61668116e-01
-1.99455664e-01 -5.42778313e-01 -2.76421249e-01 -8.36332858e-01
-3.56151909e-01 5.61460912e-01 6.73615113e-02 1.71868771e-01
3.86705473e-02 9.27635968e-01 -1.43679368e+00 -2.62707919e-02
8.22819054e-01 1.36619627e+00 -1.97524816e-01 3.69598895e-01
-1.22221500e-01 4.37904388e-01 -5.33906162e-01 -1.74725056e-01
5.21288574e-01 -2.54327774e-01 -9.12284911e-01 7.90270716e-02
3.16741824e-01 -2.32944742e-01 -1.63008422e-01 9.54433620e-01
7.12076366e-01 1.52389884e-01 9.90488827e-01 -1.17363679e+00
-4.18424517e-01 3.50121856e-02 -6.74919784e-01 7.36945271e-01
1.70892943e-02 4.22263406e-02 1.19574332e+00 -7.33839154e-01
2.59023696e-01 8.81293356e-01 7.56719232e-01 5.98124325e-01
-1.33183098e+00 -5.56202233e-01 1.15054742e-01 -2.17269018e-01
-1.30986083e+00 -1.73257723e-01 5.86165845e-01 -5.90877712e-01
8.62433732e-01 2.03705266e-01 7.61386275e-01 6.28301144e-01
4.03742909e-01 6.36078238e-01 1.04865801e+00 -2.02352107e-01
1.47937864e-01 2.01669723e-01 -1.77648619e-01 8.58852148e-01
1.90393910e-01 2.68216252e-01 -1.30558595e-01 -3.00425708e-01
4.82370734e-01 2.40384907e-01 1.53561411e-02 -4.24213201e-01
-7.81552434e-01 1.11772358e+00 7.19447672e-01 4.53241795e-01
-3.41297597e-01 -1.53478295e-01 4.39620167e-01 3.35448533e-01
5.47166467e-01 6.20468140e-01 -3.52534384e-01 1.96854547e-01
-9.97957587e-01 1.90929741e-01 4.55127537e-01 2.70498216e-01
4.64315355e-01 -6.55806391e-03 -2.21232660e-02 1.26266479e+00
4.01157290e-01 -1.40751153e-02 9.50991809e-01 -4.97387886e-01
1.41369198e-02 6.13675833e-01 -5.77535987e-01 -7.53745794e-01
-8.58720422e-01 -6.40985847e-01 -7.15659738e-01 3.36006790e-01
2.38266468e-01 -1.49417818e-02 -1.21723080e+00 1.74625027e+00
3.69350761e-01 -2.02988386e-01 -8.81639868e-02 6.62905931e-01
7.71615565e-01 1.32528394e-01 4.09232229e-02 9.31150168e-02
1.21959376e+00 -7.55125761e-01 -4.35022533e-01 -1.91494465e-01
9.30831730e-01 -7.80313790e-01 9.04647171e-01 2.22415477e-01
-1.14648592e+00 -2.84676608e-02 -1.05837548e+00 4.39847372e-02
-3.99317592e-01 -1.45217180e-01 4.36923027e-01 5.86645961e-01
-1.12660801e+00 6.48846090e-01 -1.06846929e+00 -1.05555996e-01
7.95490682e-01 7.07435131e-01 -4.95696723e-01 2.24837452e-01
-7.67668426e-01 1.02377403e+00 4.76735950e-01 -1.66916192e-01
-4.82382804e-01 -6.58510447e-01 -9.44401741e-01 4.77447584e-02
-1.25843838e-01 -6.28120244e-01 1.18687105e+00 -5.44938505e-01
-1.18364644e+00 1.17651320e+00 1.68986574e-01 -5.67784250e-01
4.45637614e-01 2.81627625e-01 -4.06040907e-01 2.53784567e-01
1.00293368e-01 8.35445106e-01 8.26165378e-01 -1.09655213e+00
-7.30589330e-01 -6.70958221e-01 -1.08035728e-01 6.50997907e-02
-3.09620857e-01 -2.03326777e-01 1.11100413e-01 -7.86360264e-01
7.01297462e-01 -9.34647918e-01 -3.07113737e-01 2.66061276e-01
-3.83949041e-01 -2.23198533e-01 7.40742862e-01 -5.95605135e-01
1.02507985e+00 -2.17573547e+00 -1.92594171e-01 5.69576204e-01
8.76565814e-01 1.78708762e-01 1.45423517e-01 -1.92443967e-01
-3.40186656e-01 2.01698497e-01 -6.78888917e-01 -4.88522828e-01
-1.28415227e-01 4.20868307e-01 2.37724394e-01 6.57653213e-01
4.26967949e-01 5.35183012e-01 -7.17876196e-01 -4.18386012e-01
1.62456945e-01 5.03154397e-01 -1.00186169e+00 -1.03473179e-02
2.48378307e-01 5.35752833e-01 -1.51701704e-01 6.87668800e-01
7.64902234e-01 -3.87187064e-01 -2.14001775e-01 2.91906446e-01
2.50999361e-01 2.76939243e-01 -8.38832617e-01 1.25380468e+00
-7.48460472e-01 4.31436539e-01 6.55968636e-02 -1.20764208e+00
1.00403595e+00 2.97404379e-01 7.95534730e-01 -6.84564471e-01
1.48800597e-01 7.49747872e-01 4.04519707e-01 -3.39534730e-01
6.84793890e-02 -7.06299126e-01 1.76572904e-01 4.89806026e-01
2.33034536e-01 -2.01945931e-01 -1.16272040e-01 -1.28372058e-01
1.13350487e+00 -7.11439729e-01 3.51012766e-01 -1.75128609e-01
6.13275290e-01 -5.88872194e-01 4.15427297e-01 5.29471278e-01
-5.50704420e-01 8.05161476e-01 6.28142297e-01 -5.61316669e-01
-9.50521350e-01 -1.23316479e+00 -8.10529590e-01 6.59852684e-01
-2.56971031e-01 -2.30880491e-02 -4.18929219e-01 -1.12658536e+00
2.46410504e-01 3.59540522e-01 -6.71917856e-01 -3.67790014e-01
-5.38529038e-01 -9.63005543e-01 5.73026419e-01 6.30925119e-01
2.88657755e-01 -9.04773295e-01 -4.15628165e-01 -5.80360182e-02
2.31977105e-01 -7.84649789e-01 -5.11770666e-01 4.91582453e-01
-1.02200878e+00 -1.07969773e+00 -7.45649457e-01 -9.89513934e-01
9.34689879e-01 -3.75569940e-01 8.06526244e-01 4.19704646e-01
-5.14282763e-01 2.54561186e-01 -9.19117108e-02 -4.19817060e-01
-3.09219569e-01 1.05959348e-01 2.40130246e-01 -1.03937052e-01
4.35951561e-01 -7.46676862e-01 -7.85681963e-01 1.85669139e-01
-1.05507934e+00 -7.86150873e-01 4.46874917e-01 1.06950772e+00
7.67321527e-01 -2.90251486e-02 4.87464458e-01 -7.26435065e-01
6.64639950e-01 -6.42802417e-01 -2.88459212e-01 -2.11905122e-01
-9.23268676e-01 1.31423876e-01 4.07618821e-01 -2.19493836e-01
-5.68468094e-01 1.85519122e-02 -6.29401326e-01 -3.21857870e-01
-1.23937525e-01 3.50414813e-01 3.35188568e-01 -2.05820069e-01
6.76016688e-01 6.33032694e-02 3.43327492e-01 -3.64954114e-01
-6.55068904e-02 6.48439348e-01 2.17483982e-01 -3.58376920e-01
5.57370484e-01 7.27594852e-01 1.20490655e-01 -7.88387597e-01
-9.46473539e-01 -6.39640987e-01 -6.44268870e-01 8.16471055e-02
9.25358534e-01 -4.73095208e-01 -4.69027042e-01 2.53550440e-01
-6.26005352e-01 1.24454305e-01 -7.43962228e-01 8.56998622e-01
-5.67233324e-01 3.32457244e-01 -5.38948894e-01 -6.74938440e-01
-3.34739268e-01 -1.39576566e+00 1.05173039e+00 -2.72249449e-02
-4.12044078e-01 -1.32446766e+00 1.76183924e-01 1.26893416e-01
5.03209233e-01 3.49741161e-01 1.08704937e+00 -1.34928632e+00
7.92041719e-02 -3.98752630e-01 -1.50582165e-01 6.54787421e-01
4.13358241e-01 -2.79388547e-01 -1.09300518e+00 -4.57981467e-01
2.71484435e-01 -1.71377197e-01 1.11858332e+00 6.30293429e-01
1.68913150e+00 -2.98935890e-01 -2.52427533e-02 8.31766307e-01
1.04107630e+00 4.06363428e-01 3.05430502e-01 4.92301911e-01
4.23660964e-01 7.01276362e-01 -1.93739254e-02 3.66145164e-01
1.03298269e-01 4.42769945e-01 5.10051906e-01 -1.67394027e-01
1.52358525e-02 -3.28793488e-02 -1.88427135e-01 8.45867753e-01
6.20969385e-02 8.12717825e-02 -1.03860211e+00 5.79933405e-01
-1.23915744e+00 -5.91783524e-01 4.20107007e-01 2.28356075e+00
6.52857423e-01 3.46176147e-01 3.04908156e-01 3.47265750e-01
5.52763641e-01 1.91514075e-01 -4.44467485e-01 -7.11203456e-01
-4.60451543e-02 5.53822696e-01 3.19182366e-01 4.72094417e-01
-1.22359776e+00 4.93321121e-01 6.31862259e+00 7.45576680e-01
-1.45366120e+00 7.97659457e-02 5.78088582e-01 -3.30856383e-01
-3.65503997e-01 -3.52676094e-01 -6.34736001e-01 5.47930539e-01
7.21278250e-01 1.07887655e-01 4.68387641e-03 7.15322912e-01
-6.67725727e-02 1.87028512e-01 -1.13085043e+00 9.42917347e-01
-2.18956862e-02 -1.08110380e+00 9.11027119e-02 5.34073055e-01
5.51335275e-01 1.15613736e-01 5.26916862e-01 2.33106151e-01
-8.33347589e-02 -1.26763761e+00 2.45464876e-01 1.40703186e-01
5.55656016e-01 -8.12864661e-01 8.78466368e-01 8.95413235e-02
-9.76888239e-01 -2.43245363e-01 -1.25603899e-01 2.76833922e-01
2.21309941e-02 5.97785413e-01 -1.21752763e+00 5.32924160e-02
7.14710116e-01 4.69889730e-01 -2.78003126e-01 1.41471350e+00
-1.36388578e-02 2.84334213e-01 -5.64372301e-01 2.49941230e-01
5.68579376e-01 -2.47913569e-01 6.64431572e-01 8.84819865e-01
3.63695949e-01 -8.16090852e-02 1.80952638e-01 7.50820339e-01
-1.76308155e-01 3.77907008e-01 -7.41987288e-01 1.87018916e-01
1.67577922e-01 8.74242067e-01 -8.25755239e-01 -7.47161433e-02
-3.04737389e-01 5.70175111e-01 1.83527574e-01 -1.32074282e-01
-3.85837972e-01 -2.28409931e-01 6.87223375e-01 3.37237775e-01
3.23624521e-01 1.18006080e-01 -5.04841030e-01 -8.61757398e-01
1.16023935e-01 -5.25481582e-01 7.25371063e-01 -1.46382928e-01
-1.53050864e+00 8.52201343e-01 -1.28951982e-01 -1.16104650e+00
-2.11169958e-01 -7.12339580e-01 -6.13112092e-01 6.56386197e-01
-1.62509167e+00 -7.43575633e-01 3.70151037e-03 5.01833737e-01
2.46199429e-01 -4.39087331e-01 7.74435222e-01 5.09745657e-01
-4.26252037e-01 9.08744574e-01 1.69626232e-02 2.10812971e-01
4.67044234e-01 -1.58560097e+00 -3.98043217e-03 3.33573997e-01
7.08743483e-02 5.77072740e-01 3.74651760e-01 -3.91187280e-01
-2.06232011e-01 -9.88563895e-01 9.34762776e-01 -2.48222891e-02
5.94534755e-01 5.71234897e-02 -1.12209594e+00 7.23450601e-01
-1.96076840e-01 4.85808074e-01 9.82905567e-01 1.29192799e-01
-4.71191525e-01 4.18331735e-02 -1.65908945e+00 2.99532890e-01
6.93900585e-01 -5.65027118e-01 -6.77704871e-01 6.78349435e-01
4.26019549e-01 -4.66571629e-01 -1.15827465e+00 5.85514903e-01
3.90563935e-01 -8.54177058e-01 9.27472353e-01 -7.92107224e-01
3.70652109e-01 1.36337712e-01 -8.37589502e-02 -1.38243830e+00
1.17657324e-02 -2.23188519e-01 8.23181495e-02 4.50400770e-01
7.15944290e-01 -1.06575906e+00 7.79528797e-01 4.49115962e-01
-3.53959560e-01 -1.13205969e+00 -1.26983035e+00 -9.73517716e-01
5.56218982e-01 -4.00315464e-01 4.85583872e-01 1.05342829e+00
-1.81265399e-01 -1.15391165e-01 1.02354474e-01 2.02346578e-01
4.57559377e-01 -1.59833297e-01 1.63532555e-01 -1.46196330e+00
-4.34541404e-02 -9.18366134e-01 -9.07825589e-01 -7.77591407e-01
1.37250721e-01 -1.32961249e+00 -2.28237063e-01 -1.26644456e+00
1.61405608e-01 -7.42666662e-01 -5.49999356e-01 5.24143577e-01
1.74898386e-01 2.87232935e-01 -1.55296579e-01 1.89602375e-01
4.66545066e-03 5.66475749e-01 1.26964939e+00 -1.26130462e-01
-3.81849021e-01 4.03039306e-01 -4.85683352e-01 1.03759408e+00
9.39929783e-01 -7.26242959e-01 -1.95715636e-01 -1.02650404e-01
-3.93977799e-02 -3.71830285e-01 3.02590311e-01 -8.15525830e-01
-8.76585096e-02 4.08230811e-01 3.95865619e-01 -4.73621666e-01
5.09599447e-01 -7.22409248e-01 -5.26822269e-01 6.96185946e-01
-4.40616220e-01 1.11072935e-01 3.16265434e-01 4.56053317e-01
-5.42502880e-01 -4.42255646e-01 1.27072299e+00 -1.59960046e-01
-1.43253192e-01 5.85942864e-01 -3.03744256e-01 3.64967793e-01
9.87936258e-01 -3.15541893e-01 1.01262636e-01 -2.80276626e-01
-1.29757440e+00 1.33292213e-01 2.15287223e-01 1.60792217e-01
8.23174357e-01 -1.35262704e+00 -6.67002499e-01 5.92099369e-01
2.99434721e-01 4.89506386e-02 1.68982312e-01 1.20114994e+00
-8.30429256e-01 3.89559001e-01 -2.72177935e-01 -1.00313282e+00
-1.25480366e+00 1.89592615e-01 8.49259496e-01 -3.47007781e-01
-8.64973485e-01 1.23804903e+00 2.42847532e-01 -9.00828838e-01
3.30842316e-01 -4.24240351e-01 -2.49354646e-01 2.09508747e-01
3.97743225e-01 9.26262513e-02 6.02550924e-01 -6.57289445e-01
-3.91727507e-01 4.94920969e-01 -5.24712503e-01 9.63761359e-02
1.45373809e+00 2.08345562e-01 3.82309109e-02 1.27427787e-01
1.68892372e+00 2.70647034e-02 -7.53060460e-01 -4.33200032e-01
4.25577164e-03 -5.45566559e-01 4.01666760e-01 -4.20418054e-01
-1.48036909e+00 8.74407291e-01 1.01415312e+00 4.65414107e-01
1.17375958e+00 2.58732051e-01 9.27598596e-01 1.59414828e-01
-3.33399117e-01 -1.10457695e+00 2.63179064e-01 4.35021818e-01
7.70373821e-01 -1.51170552e+00 -7.07063004e-02 -6.83195367e-02
-5.88817894e-01 9.46118653e-01 4.15099412e-01 -1.99853420e-01
1.18418717e+00 2.15797812e-01 7.73785189e-02 -6.26712620e-01
-3.57394159e-01 -2.44953483e-01 4.48287815e-01 5.56077182e-01
3.91091108e-01 2.65533011e-02 -3.02982301e-01 2.17400402e-01
-5.01380563e-01 -5.41428626e-01 3.95138413e-01 8.66678655e-01
-4.34054136e-01 -9.35241103e-01 -1.71151832e-01 1.03112042e+00
-8.38618338e-01 -1.01363838e-01 -1.95021182e-01 1.02671683e+00
1.82298735e-01 5.14570296e-01 5.09451985e-01 -2.70043075e-01
3.92853796e-01 4.96678315e-02 4.37249035e-01 -7.75367558e-01
-5.41579127e-01 1.77631192e-02 -2.87407219e-01 -3.32989216e-01
-3.26563716e-01 -8.02201390e-01 -1.36819327e+00 -2.27265894e-01
-3.49259108e-01 2.07641665e-02 6.99722290e-01 1.00899231e+00
-8.34752619e-02 4.22576457e-01 7.47629702e-01 -3.99543941e-01
-8.30218792e-01 -6.17812216e-01 -9.17508960e-01 3.55021268e-01
8.81491542e-01 -1.05347896e+00 -9.28382456e-01 -5.96628249e-01] | [14.734060287475586, -2.2840497493743896] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.